© Copyright 2004 by Humana Press Inc.
All rights of any nature, whatsoever, reserved.
0163-4984/04/0000–0000 $25.00
Absorption of the Biomimetic
Chromium Cation Triaqua-µ3-oxo-µhexapropionatotrichromium(III)
in Rats
BUFFIE J. CLODFELDER, CHRISTINE CHANG,
AND JOHN B. VINCENT*
Department of Chemistry and Coalition
for Biomolecular Products, The University of Alabama,
Tuscaloosa, AL 35487-0336
Received August 19, 2003; Accepted September 8, 2003
ABSTRACT
The cation [Cr3O(O2CCH2CH3)6(H2O)3]+ has been shown in vitro to
mimic to the oligopeptide chromodulin’s ability to stimulate the tyrosine
kinase activity of insulin receptor and shown in healthy and type 2 diabetic
model rats to increase insulin sensitivity and decrease plasma total and
low-density lipoprotein cholesterol and triglycerides concentrations. However, the degree to which the complex is absorbed after gavage administration to rats had not been previously determined. The biomimetic cation
at nutritional supplement levels is absorbed with greater than 60% efficiency, and at pharmacological levels, it is absorbed with greater than 40%
efficiency, an order of magnitude greater absorption than that of CrCl3, Cr
nicotinate, or Cr picolinate, currently marketed nutritional supplements.
The difference in degree of absorption is readily explained by the stability
and solubility of the cation.
Index Entries: Absorption; chromium; propionate; rats; insulin
resistance.
INTRODUCTION
The element chromium in the +3 oxidation state is believed to be an
essential trace element for mammals (1). This has led to the development of
*Author to whom all correspondence and reprint requests should be addressed.
Biological Trace Element Research
1
Vol. 97, 2004
2
Clodfelder, Chang, and Vincent
chromium complexes as nutritional supplements; however, recently the
National Academy of Sciences of the United States has set the adequate
intake (AI) of chromium at 35 µg daily for men and 25 µg daily for women
(2). Consequently, nearly all Americans consume a chromium-adequate
diet, and the consensus of studies on the supplementation of the diet of
healthy individuals indicates that supplementation has no beneficial effects
(3–6). A similar lack of effects has been noted in healthy rats (e.g., ref. 7),
with the exception of one compound, [Cr3O(O2CCH2CH3)6(H2O)3]+. This
trinuclear Cr(III) cation has been found to have striking in vivo effects, lowering plasma triglycerides, total cholesterol, and LDL cholesterol in healthy
rats when administered intravenously at a level of 20 µg Cr/kg body mass
daily for 3 mo (8). When administered intravenously at pharmacological
levels of 5–20 µg Cr/kg body mass for 6 mo, the cation has also been shown
to reduce lower triglycerides, total and low-density lipoprotein (LDL) cholesterol, and insulin concentrations and increase insulin sensitivity in
healthy and type II diabetic model rats (9). This ability presumably arises
from the cation’s ability to mimic the naturally occurring, bioactive
oligopeptide chromodulin. Chromodulin is believed to function as part of
a unique autoamplification mechanism for insulin signaling (10). The trinuclear cation activates the insulin-dependent tyrosine kinase activity of the
insulin receptor in a fashion very similar to chromodulin. The kinase activity of a recombinant fragment of the receptor, for example, is stimulated
approximately threefold with a dissociation constant of 1.00 nM by a synthetic complex (11).
Whereas chromodulin is rapidly cleared from the bloodstream into
the urine (12) and does not affect plasma variables when administered
intravenously to rats (8), the biomimetic trinuclear complex is stable in
the blood and can enter cells intact, in the form able to activate the
insulin receptor (13,14). The biomimetic complex is very soluble in
water and stable in dilute mineral acid (it can be recrystallized from
dilute acid) such that it might survive in the digestive tract. Indeed,
recent studies in which healthy and type II diabetes model rats were
gavaged with aqueous solutions of the complex found that the complex
increased insulin sensitivity while lowering plasma total and LDL cholesterol and triglycerides concentrations (Clodfelder et al., unpublished
results). However, the degree to which the complex is absorbed when
given orally has not been determined previously and is the focus of the
work presented herein.
MATERIALS AND METHODS
Materials and Instrumentation
51CrCl
in
1.0
M
HCl
was
obtained
from
ICN.
3
[Cr3O(O2CCH2CH3)6(H2O)3]NO3 was made by the method of Earnshaw et
al. (15) using a tracer amount of 51CrCl3. All procedures were performed
Biological Trace Element Research
Vol. 97, 2004
Cr(III) Absorption in Rats
3
using doubly-deionized water unless otherwise noted. Gamma-counting
was performed on a Packard Cobra II auto-gamma counter.
Animals
Male Sprague–Dawley rats (between ~450 and 775 g) were obtained
from Charles River Laboratories; rats were housed at least 1 wk after
arrival before use. Rats were maintained on a 12-h light/dark cycle. The
rats were gavaged with an aqueous solution of the biomimetic cation
(either 9.77 mg or 10.26 µg in 100 µL H2O) and placed into metabolic cages
for collection of urine and feces. After administration of the cation, the
rats were allowed to feed on a commercial rat chow and drink ad libitum.
Three rats were utilized for each quantity of the biomimetic cation at each
time. After appropriate time intervals (from 30 min to 24 h), blood samples were taken from tail snips. The rats followed for 24 h after treatment
were maintained on the 12-h light/dark cycle. Rats were sacrificed by carbon dioxide asphyxiation, and tissue samples (stomach, small intestine,
large intestine, heart, liver, testes, spleen, kidneys, epididymal fat, right
femur, pancreas, and muscle [muscolus triceps surae] from the right hind
leg) were harvested and weighed. Subcellular liver fractions were
obtained by differential centrifugation according to established procedures (13,14). A portion of the liver from each rat was diced, and the
pieces were rinsed with 0.25 M sucrose. All subsequent steps were performed with the same solution. The pieces were ground in a tissue
grinder. The nuclear, mitochondrial, lysosomal, and microsomal fractions
were obtained as pellets from centrifugation at 30g, 3300g, 25,000g, and
100,000g, respectively. Blood and muscle were assumed to comprise 6%
and 30% of the body mass, respectively. The University of Alabama Institutional Animal Use and Care Committee approved all procedures
involving the use of rats.
Statistical Treatment
Each data point in the figures represents the average value for three
rats; error bars in the figures denote standard error.
RESULTS AND DISCUSSION
Male Sprague–Dawley rats were gavaged with either a high (9.77 mg,
~2 mg Cr) or low (10.26 µg, ~2 µg Cr) dose of the biomimetic cation
[Cr3O(O2CCH2CH3)6(H2O)3]+. Given the body mass of the rats, this corresponds to approx 3 mg (high dose) and 3 µg (low dose) Cr/kg body mass.
Assuming an average body mass of a human being of 65 kg, the low dose
corresponds roughly to a human receiving 200 µg Cr daily. Commercial
Cr-containing nutritional supplements generally contain 200–600 µg Cr;
Biological Trace Element Research
Vol. 97, 2004
4
Clodfelder, Chang, and Vincent
Fig. 1. Distribution of 51Cr from the biomimetic cation in the gastrointestinal
tract as a function of time. Left: 2 mg Cr; right, 2 µg Cr. Solid circles-stomach; open
circles-small intestines; solid triangles-larges intestines.
thus, the low dose is equivalent to that of a human taking a nutritional
supplement. In studies in which healthy or diabetic model rats were gavaged with the biomimetic complex, rats received up to 1 mg Cr daily
(Clodfelder et al., unpublished results); the high dose approximates this
pharmacological quantity of Cr.
With time, the 51Cr from the cation can readily be followed through the
gastrointestinal tract (see Fig. 1). At least 90% of the administered dose clears
the stomach in 30 min (low dose) to 2 h (high dose). Anderson and Polansky
(16) have examined the fate of CrCl3 given by gavage administration to rats;
in 30 min, approx 90% of the Cr had passed through the stomach and the
first 15 cm of the small intestine. After 1 h, almost 100% of the Cr was in the
lower portion of the small intestine and the large intestine. After 24 h, about
55% of the label was still in the lower portion of the intestine and the large
intestine (16); the quantity of Cr in the feces was not reported. In the current
study, the content of the label in the small intestine reaches a maximum in 1
(low dose) to 2 h (high dose). For both doses, the maximum content of the
label in the large intestine is reached at 6 h. This passage through the gastrointestinal tract is accompanied by increased loss of the label in the feces
with time (see Fig. 2). For both doses, between 30% and 40% of the administered label is lost in the feces within 24 h of treatment.
Adding together the Cr content of the gastrointestinal tract and feces
allows the percentage of Cr absorbed to be calculated (see Fig. 3). For the
lower dose of Cr, absorption occurs rapidly during the first 30 min. This
Biological Trace Element Research
Vol. 97, 2004
Cr(III) Absorption in Rats
5
Fig. 2. Distribution of 51Cr from the biomimetic cation in the urine, feces, and
blood plasma as a function of time. Left: 2 mg Cr; right: 2 µg Cr. Solid circlesblood plasma; open circles-feces; solid triangles-urine.
Fig. 3. Percent absorbed 51Cr. Solid circles-2 mg Cr; open circles-2 µg Cr.
Biological Trace Element Research
Vol. 97, 2004
6
Clodfelder, Chang, and Vincent
suggests that absorption of Cr from the administered biomimetic cation is
rapid from the stomach. The drop with time after 30 min and the subsequent recovery with time in the % absorption suggest a loss from and reentry of Cr into the gastrointestinal tract. This can be explained by a
movement of a portion of the rapidly absorbed Cr into the bile and then
movement of the bile into the duodenum. However, Cr in the small intestines subsequently is absorbed until 60% of the administered dose is
absorbed 6 h after treatment; absorption appears to be complete after 6 h.
Thus, Cr absorption appears to stop as Cr enters the large intestine. Consequently, Cr absorption is essentially limited to the stomach and small
intestine under these conditions. For the higher dose of Cr, a distinct maximum in absorption does not occur after 30 min, although the general
shape of the % absorption curve is similar to that of the lower dose. The
initial rapid absorption of Cr via the stomach, which was appreciable at
the low dose, appears to be saturated and thus to comprise a smaller percentage of the administered dose. Absorption continues during the entire
24 h after treatment, suggesting that significant absorption can take place
in the large intestine if levels of Cr reach a certain level.
The appearance of Cr in bile has been noted previously. Rats treated
with an oral dose of CrCl3 have very low concentrations of Cr in their bile
(~0.2% of the dose), suggesting that biliary excretion is not an important
route for loss of Cr when given as the chloride (17). Rats treated with the
biomimetic cation by intravenous injection lost about 2–3% of the administered dose in the feces within 24 h (13,14), indicating that once absorbed
into the bloodstream, Cr can move into the bile. However, these intravenous experiments did not measure the Cr content of the gastrointestinal
tract, such that the full extent of Cr movement as bile in the tract cannot be
estimated.
The magnitude of the absorption of the cation is most notable. In
humans, dietary Cr is absorbed with an efficiency of 0.4–2% (18). In rats,
Olin et al. (19) and Anderson et al. (20) have shown that Cr chloride, Cr
nicotinate, and Cr picolinate [popular forms of Cr(III) in nutritional supplements] are absorbed to similar extents (0.5–1.3% of the gavaged dose of
0.14 or 0.15 µg Cr, respectively, after 24 h). For the biomimetic cation, 2%
(low dose) or 17% (high dose) of the Cr is lost in the urine alone within 24
h. When the Cr in the carcass (except the gastrointestinal tract) is taken into
consideration, over 40% (high dose) or 60% (low dose) of the Cr is
absorbed over 24 h (see Fig. 3), easily 20–30 times greater than with Cr chloride, Cr nicotinate, or Cr picolinate. These numbers fail to take into consideration the Cr in the gastrointestinal tract or in the feces that re-entered
the gastrointestinal tract from the bile. In the case of the nutritional dose of
the cation, this could increase the total by an additional 30%, given the
magnitude in the drop in the % absorption curve for this dose in Fig. 3
after the 30-min time-point. The enormous increase in absorption efficiency for the biomimetic cation compared to the Cr complexes used currently as nutritional supplements can readily be explained.
Biological Trace Element Research
Vol. 97, 2004
Cr(III) Absorption in Rats
7
Fig. 4. Distribution of 51Cr from the biomimetic cation in body tissues as a
function of time. Left: 2 mg Cr; right: 2 µg Cr. Solid circles-liver; open circlesspleen; solid inverted triangles-pancreas; open inverted triangles-kidney; solid
squares-epididymal fat; open squares-testes; solid diamond-heart; open diamondfemur; solid triangle-muscle.
Commercially available Cr chloride, “CrCl3,” is actually the salt trans[CrCl2(H2O)4]Cl. The compound is susceptible to oligomerization, especially at basic pH’s, leading to the formation of numerous multinuclear
hydroxo-bridged chromic species. Thus, Cr in the complex upon exchange
of the aquo ligands can bind to biomolecules and be carried by large biomolecules through the gastrointestinal tract or can undergo oligomerization and polymerization to form species of limited solubility, especially at
the alkaline pH of the intestines. Chromium nicotinate is poorly characterized and has limited solubility; the nicotinate ligands are relatively labile,
generating forms of Cr susceptible to bind to biomolecules and olation
(21). Chromium picolinate has very limited solubility in water (0.6 mM)
(22) and in other common solvents and is not particularly lipophilic (23).
In contrast, [Cr3O(O2CCH2CH3)6(H2O)3]+ is extremely soluble in water
(versus Cr nicotinate and picolinate) and appears to be able to maintain its
integrity in vivo (14), not breaking down in the gastrointestinal tract (versus CrCl3 and Cr nicotinate).
The distribution of Cr in the tissues and fluids after the administration
of the biomimetic cation is similar to that when the complex is administered intravenously (13,14) or when other Cr complexes are administered
Biological Trace Element Research
Vol. 97, 2004
8
Clodfelder, Chang, and Vincent
Fig. 5. Distribution of 51Cr from the biomimetic cation in the subcellular components of hepatocytes as a function of time. Left: 2 mg Cr; right: 2 µg Cr. Solid
circles-nucleus; open circles-mitochondria; solid triangles-lysosomes; open triangles-cytosols; solid squares-microsomes.
orally (16,19,20). However, the levels of Cr in the tissues are much larger
for the biomimetic complex. For both the low and high dose of the cation,
the tissue with the greatest total amount of Cr at all times is the muscle (see
Fig. 4). For the low dose, over 1% of the administered dose is present in the
muscle 6 h after administration, whereas approx 20% of the low dose is in
the muscle 24 h after administration. The liver in both cases contains the
next greatest quantity of Cr, at least for the first 6 h. When the cation was
administered intravenously, liver possessed the greatest amounts of the
administered Cr (13,14) (the content of skeletal muscle was not examined).
For gavage-administered Cr chloride and Cr nicotinate, muscle contained
the greatest quantity of the administered dose at all time-points up to 24 h
after administration, followed by the liver (19). The situation for Cr picolinate is similar except that the kidneys contain more Cr than the liver for
the first 3 h after administration (19).
Finally, the distribution of Cr from the cation in hepatocytes was
examined for both the high and low dose. For the pharmacological dose,
the distribution of Cr was essentially constant (see Fig. 5); this distribution
is the amount of Cr in the subcellular compartment divided by the total Cr
in the hepatocyte and does not directly reflect the change in total liver conBiological Trace Element Research
Vol. 97, 2004
Cr(III) Absorption in Rats
9
tent with time. Most of the Cr in the hepatocytes was localized to the
nucleus and cytosol, followed in order by the mitochondria, microsomes,
and lysosomes. The relatively stable distribution over time may suggest
that Cr intake and initial distribution of the Cr may be substantially faster
than a later step in the processing of Cr required for expulsion of Cr from
the cells. This distribution differs substantially from when a single intravenous dose of the cation is administered to rats (14). In this case, Cr is
found to rapidly enter the microsomes (reaching a maximum in 1 h) and
to be lost almost as rapidly, after which time, most of the Cr is distributed
in the cytosol and nucleus. After daily intravenous administration for 2
wk, most of the Cr from the cation in hepatocytes is localized in the
nucleus (13). Consequently, administration of Cr intravenously, which
results in the cation reaching the hepatocytes intact in a brief interval of
time, gives rise to a different distribution of Cr than oral administration,
where absorption from the gastrointestinal tract over time generates a time
averaging of the results.
For the nutritional dose, the distribution of labeled Cr changes significantly with time (see Fig. 5). Unfortunately, the data on the distribution of
Cr are accompanied by an appreciable degree of error, resulting from the
very small number of counts in the samples. Yet, in the first hour, most of
the Cr is located in the mitochondria, after which, Cr appears to accumulate in the lysosomes and cytosol and decrease in the other subcellular
compartments. The small quantity of Cr in these samples does not allow
for information to be obtained on the chemical nature of the Cr in the subcellular compartments as a function of time, limiting any further interpretation of the data.
In conclusion, the biomimetic cation [Cr3O(O2CCH2CH3)6(H2O)3]+ is
absorbed by rats to a far greater degree than previously reported for any
other Cr complex, including the forms of Cr most commonly used in nutritional supplements. Although the need for Cr nutritional supplements is
questionable, recent studies have suggested that Cr(III) species at pharmacological levels may have beneficial effects on the symptoms of type 2 diabetes (8,9,24). The biomimetic cation does not appear to have the potential
to act as a catalyst to generate reactive oxygen species (25) and should not
possess the potential to act as a mutagen (26,27) and generate other deleterious effects as recently observed with Cr picolinate (5). Thus, the cation
triaqua-µ3-oxo-µ-hexapropionatotrichromium(III) is worthy of continued
investigation as a potential pharmaceutical agent.
ACKNOWLEDGMENTS
Funding was provided by a grant from the National Institutes of
Health (DK62094-01) to J.B.V. C.C. was supported in part by the McNair
Scholars Program of The University of Alabama.
Biological Trace Element Research
Vol. 97, 2004
10
Clodfelder, Chang, and Vincent
REFERENCES
AU:
Vol. number
please.
1. J. B. Vincent, The bioinorganic chemistry of chromium(III). Polyhedron 20, 1–26 (2001).
2. P. Trumbo, A. A. Yates, S. Schlicker, et al., Dietary reference intakes: vitamin A, vitamin
K, arsenic, boron, chromium, copper, iodine, iron, manganese, molybdenum, nickel, silicon, vanadium, and zinc. J. Am. Diet. Assoc. 101, 294–301 (2001).
3. M. H. Pittler, C. Stevinson, and E. Ernst, Chromium picolinate for reducing body
weight: meta-analysis of randomized trials. Int. J. Obes. 27, 522–529 (2003).
4. S. L. Nissen and R. L. Sharp, Effect of dietary supplements on lean mass and strength
gains with resistance exercise: a meta-analysis. J. Appl. Physiol. 94, 651–659 (2003).
5. J. B. Vincent, The potential value and potential toxicity of chromium picolinate as a
nutritional supplement, weight loss agent, and muscle development agent. Sports Med.
33, 213–230 (2003).
6. M. D. Althius, N. E. Jordan, E. A. Ludington, et al., Glucose and insulin responses to
dietary chromium supplements: A meta-analysis. Am. J. Clin. Nutr. 76, 148–155 (2002).
7. R. A. Anderson, N. A. Bryden, and M. M. Polansky, Lack of toxicity of chromium chloride and chromium picolinate in rats. J. Am. Coll. Nutr. 16, 273–279 (1997).
8. Y. Sun, K. Mallya, J. Ramirez, et al., The biomimetic [Cr3O(O2CCH2CH3)6(H2O)3]+
decreases cholesterol and triglycerides in rats: towards chromium-containing therapeutics. J. Biol. Inorg. Chem. 4, 838–845 (1999).
9. Y. Sun, B. J. Clodfelder, A. A. Shute, et al., The biomimetic [Cr3O(O2CCH2CH3)6(H2O)3]+
decreases plasma insulin, cholesterol and triglycerides in healthy and type II diabetic
rats but not type I diabetic rats. J. Biol. Inorg. Chem. 7, 852–862 (2002).
10. J. B. Vincent, Elucidating a biological role for chromium at a molecular level. Acc. Chem.
Res. 33, 503–510 (2002).
11. C. M. Davis, A. C. Royer, and J. B. Vincent, Synthetic multinuclear chromium assembly
activates insulin receptor kinase activity: functional model for low-molecular-weight
chromium-binding substance. Inorg. Chem. 36, 5316–5320 (1997).
12. O. Wada, G. Y. Wu, A. Yamamoto, et al., Purification and chromium-excretory function of
low-molecular-weight, chromium-binding substances. Environ. Res. 32, 228–239 (1983).
13. A. A. Shute and J. B. Vincent, The stability of the biomimetic cation triaqua-µ-oxohexapropionatotrichromium(III) in vivo in rats. Polyhedron 20, 2241–2252 (2001).
14. A. A. Shute and J. B. Vincent, The fate of the biomimetic cation triaqua-µ-oxohexapropionatotrichromium(III) in rats. J. Inorg. Biochem. 89, 272–278 (2002).
15. A. Earnshaw, B. N. Figgis, and J. Lewis, Chemistry of polynuclear compounds. Part VI.
Magnetic properties of trimer chromium and iron carboxylates. J. Chem. Soc. A
1656–1663 (1966).
16. R. A. Anderson and M. M. Polansky, Dietary and metabolic effects of trivalent
chromium retention and distribution in rats. Biol. Trace Element Res. 50, 97–108 (1995).
17. M. L. Davis-Whiteneck, M. S. Bernice, B. O. Adeleye, et al., Biliary excretion of
51chromium in bile-duct cannulated rats. Nutr. Res. 16, 1009–1015 (1996).
18. R. A. Anderson and A. S. Kozlovsky, Chromium intake, absorption and excretion of
subjects consuming self-selected diets. Am. J. Clin. Nutr. 41, 1177–1183 (1985).
19. K. L. Olin, D. M. Stearns, W. H. Armstrong, et al., Comparative retention/absorption of
51chromium (51Cr) from 51Cr chloride, 51Cr nicotinate and 51Cr picolinate in a rat model.
Trace Elements Electrolytes 11, 182–186 (1994).
20. R. A. Anderson, N. A. Bryden, M. M. Polansky, et al., Dietary chromium effects on tissue chromium concentrations and chromium absorption in rats. J. Trace Elements Exp.
Med. 9, 11–25 (1996).
21. K. F. Kingry, A. C. Royer, and J. B. Vincent, Nuclear magnetic resonance studies of
chromium(III) pyridinecarboxylate complexes. J. Inorg. Biochem. 72, 79–88 (1998).
22. G. W. Evans and D. J. Pouchnik, Composition and biological activity of chromium–
pyridine carboxylate complexes. J. Inorg. Biochem. 49, 177–187 (1993).
Biological Trace Element Research
Vol. 97, 2004
Cr(III) Absorption in Rats
11
23. N. E. Chakov, R. A. Collins, and J. B. Vincent, A re-investigation of the electronic spectra of chromium(III) picolinate complexes and high yield synthesis and characterization of Cr2(µ-OH)2(pic)4·5H2O (Hpic=picolinic acid). Polyhedron 18, 2891–2897 (1999).
24. W. T. Cefalu, Z. Q. Wang, X. H. Zhang, et al., Oral chromium picolinate improves carbohydrate and lipid metabolism and enhances skeletal muscle Glut-4 translocation in
obese, hyperinsulinemic (JCR-LA corpulent) rats. J. Nutr. 132, 1107–1114 (2002).
25. J. K. Speetjens, A. Parand, M. W. Crowder, et al., Low-molecular-weight chromiumbinding substance and biomimetic [Cr3O(O2CCH2CH3)6(H2O)3]+ do not cleave DNA
under physiologically-relevant conditions. Polyhedron 18, 2617–2624 (1999).
26. D. M. Stearns, S. M. Silveira, K. K. Wolf, et al., Chromium(III) tris(picolinate) is mutagenic at the hypoxanthine (guanine) phopshoribosyl transferase locus in Chinese hamster ovary cells. Mutat. Res. 513, 135–142 (2002).
27. D. D. D. Hepburn, J. Xiao, S. Bindom, et al., Nutritional supplement chromium picolinate causes sterility and lethal mutations in Drosophila melanogaster. Proc. Natl. Acad.
Sci., USA 100, 3766–3771 (2003).
Biological Trace Element Research
Vol. 97, 2004
J Biol Inorg Chem (2002) 7: 852–862
DOI 10.1007/s00775-002-0366-y
O R I GI N A L A R T IC L E
Yanjie Sun Æ Buffie J. Clodfelder Æ Amanda A. Shute
Turkessa Irvin Æ John B. Vincent
The biomimetic [Cr3O(O2CCH2CH3)6(H2O)3]+ decreases plasma
insulin, cholesterol, and triglycerides in healthy and type II
diabetic rats but not type I diabetic rats
Received: 26 October 2001 / Accepted: 14 March 2002 / Published online: 19 April 2002
SBIC 2002
Abstract The in vivo effects of administration of the
synthetic, functional biomimetic cation [Cr3O(O2
CCH2CH3)6(H2O)3]+ to healthy and type I and type II
diabetic model rats are described. In contrast to current
chromium-containing nutrition supplements, which only
serve as sources of absorbable chromium, the trinuclear
cation has been shown in in vitro assays to interact with
the insulin receptor, activating its kinase activity, presumably by trapping the receptor in its active conformation. Thus, treatment of rats with the trinuclear
cation would be expected to result in changes in lipid
and carbohydrate metabolism related to insulin action.
After 24 weeks of intravenous administration (0–20 lg
Cr/kg body mass), the cation results in a concentrationdependent lowering of levels of fasting blood plasma
LDL cholesterol, total cholesterol, triglycerides, and
insulin and of 2-h plasma insulin and glucose levels after
a glucose challenge; these results confirm a previous 12week study examining the effect of the synthetic cation
on healthy rats and are in stark contrast to those of
administration of other forms of Cr(III) to rats, which
have no effect on these parameters. The cation has little,
if any, effect on rats with STZ-induced diabetes (a type I
diabetes model). However, Zucker obese rats (a model
of the early stages of type II diabetes) after 24 weeks of
supplementation (20 lg/kg) have lower fasting plasma
total, HDL, and LDL cholesterol, triglycerides, and insulin levels and lower 2-h plasma insulin levels. The
lowering of plasma insulin concentrations with little effect on glucose concentrations suggests that the supplement increases insulin sensitivity.
Y. Sun Æ B.J. Clodfelder Æ A.A. Shute Æ T. Irvin Æ J.B. Vincent (&)
Department of Chemistry and
Coalition for Biomolecular Products,
The University of Alabama, Tuscaloosa,
AL 35487-0336, USA
E-mail: [email protected]
Tel.: +1-205-3489203
Fax: +1-205-3489104
Keywords Chromodulin Æ Chromium Æ Rats Æ
Cholesterol Æ Insulin
Abbreviations HDL: high-density lipoprotein Æ
LDL: low-density lipoprotein Æ SD: Sprague-Dawley Æ
STZ: streptozotocin Æ ZKL: Zucker lean Æ ZKO: Zucker
obese
Introduction
In the last 15 years, nutritional studies have appeared
which have suggested that chromium(III) may have an
essential role in mammals [1, 2, 3]. Hence, a search has
been underway to identify the biologically active form of
chromium, that is, the biomolecule which naturally
binds chromium(III) and possesses an intrinsic function
associated with insulin action in mammals [4, 5].
Recently, the naturally occurring oligopeptide chromodulin, also known as low-molecular-weight chromium-binding substance (LMWCr), has been proposed as
a candidate for the biologically active form of chromium
[1, 2, 6, 7, 8, 9]. Chromodulin appears to be part of an
insulin signal amplification mechanism [1, 2, 8, 10]. The
oligopeptide binds to the insulin-activated insulin receptor, stimulating its tyrosine kinase activity up to
eight-fold with a dissociation constant of approximately
100 pM [6]. Spectroscopic studies indicate that at least
the majority of the four chromic ions per molecule are
arranged in a multinuclear chromic assembly where the
chromic centers are bridged by anionic ligands (presumably oxide and/or hydroxide). Carboxylate groups
from aspartate and glutamate residues link the oligopeptide to the assembly [9].
The existence of a multinuclear Cr(III) carboxylate
assembly in an active biomolecule has spurred an interest in the synthesis and characterization of multinuclear oxo(hydroxo)-bridged chromium(III) carboxylate
assemblies [11, 12, 13, 14]. Well-characterized, watersoluble assemblies have been tested for the ability to
stimulate the insulin receptor’s tyrosine kinase ability in
853
a fashion similar to chromodulin. The trinuclear cation
[Cr3O(O2CCH2CH3)6(H2O)3]+ was found in vitro
to mimic the ability of chromodulin to stimulate this
activity [7]; thus, the cation is a functional biomimetic
of chromodulin.
Consequently, both chromodulin and the trinuclear
cation have been proposed as potential therapeutic
agents to increase insulin sensitivity. Both have been
shown not to lead to DNA cleavage (as observed with the
popular nutritional supplement chromium picolinate)
[15, 16]. The synthetic complex has several potential
benefits over the natural material: it is inexpensive to
synthesize and can be readily prepared in bulk. Chromodulin is susceptible to hydrolysis, especially in the
presence of acid, whereas the synthetic material can be
recrystallized from dilute mineral acid [17] and could
potentially survive oral ingestion. After the insulin-signaling event, chromodulin is released in the urine [18, 19,
20, 21, 22], and the body appears to target the material
for excretion rather than absorption. For example, the
mean tubular reabsorption rate of chromodulin is very
low [22], suggesting that intravenously or orally administered oligopeptide should be cleared from the bloodstream very efficiently. However, if the biomimetic cation
remains intact in vivo, the complex could potentially
bind to the insulin-activated insulin receptor, stimulating
its kinase activity; hence, the biomimetic could have an
excellent potential to serve as a therapeutic agent to increase insulin sensitivity. Also the biomimetic, but not
chromodulin, has been shown to lower fasting serum
triglyceride and total cholesterol levels when given intravenously to healthy rats for 12 weeks at a level
equivalent to 20 lg Cr/kg per mass per day [23].
Over 120 million people are estimated to have diabetes mellitus, with approximately 16 million of these in
the United States [24]. Between 90 and 95% of the U.S.
cases are of type II diabetes (also called adult-onset or
non-insulin-dependent diabetes). Type II diabetes is the
form responsible for the rapid increase in diabetes in the
last few decades, and the number of cases is rapidly
rising in third world countries. Obesity is a risk factor
for type II diabetes; unlike type I diabetes, which is an
autoimmune disorder, type II diabetes results from insulin resistance (or insensitivity). The cause(s) of the
disease at a molecular level has only been elucidated in a
tiny percentage of cases. Consequently, any mechanism
by which insulin signaling could be stimulated has possible value in the treatment of the symptoms of type II
diabetes via increasing insulin sensitivity.
Herein is reported the effects of administration of the
biomimetic trinuclear cation to healthy and diabetic
male rats for 24 weeks. Zucker obese rats and rats with
streptozotocin (STZ)-induced diabetes were chosen to
model type II and type I diabetes, respectively. The
Zucker obese rats are a genetic model for the early stages
of type II diabetes, characterized by insulin insensitivity.
The STZ-induced diabetic rats are unable to produce
normal quantities of insulin because of damage to the
beta cells of the pancreas, simulating the loss of beta cell
function in type I diabetes and resulting in elevated
serum glucose concentrations. These studies indicate
that the trinuclear cation has significant effects in healthy and type II diabetic rats on insulin, triglycerides, and
cholesterol levels, suggesting an increase in insulin sensitivity, and no acute toxic effects.
Materials and methods
[Cr3O(O2CCH2CH3)6(H2O)3]NO3
The nitrate salt of the cation was prepared as described in the
literature [25], although it was originally incorrectly formulated as
a hydrate of [Cr3(OH)2(O2CCH2CH3)6]NO3 (for a discussion, see
[9]). All operations were performed with doubly deionized water
unless otherwise noted and performed with plasticware whenever
possible.
Animals
All rats were obtained from Charles River Laboratory. The fiveweek old male Sprague-Dawley (SD) rats and five week old Zucker
lean (ZKL) and Zucker obese (ZKO) rats were allowed to feed ad
libitum on a commercial rat food (Harland Tekland Certified LM485 Mouse/Rat Sterilizable Diet) and tap water. The rat food
contained 2.04 mg Cr/kg diet; the Cr concentration of the tap water
was 0.073 lg/g. The rats are, consequently, receiving a Cr adequate
diet. SD rats with STZ-induced diabetes, type I model rats [26], had
the diabetes induced by a tail vein injection of an aqueous solution
containing 55 mg/kg body mass STZ; after 2 days, all rats had a
plasma glucose concentration far exceeding 250 mg/dL except two
with values of 206 mg/dL. Rats were raised in standard plastic and
stainless steel cages on a 12-h light-dark cycle. Solid food intake
and body mass were monitored at 3-day intervals. Thirty-two SD
rats were divided randomly into four groups of eight. The first
group was injected daily in the tail vein with an aqueous solution
containing the biomimetic cation to give a total amount of chromium equivalent to 5 lg Cr per kilogram body weight. The second
group received an aqueous solution of the cation to give 10 lg Cr/
kg body mass; the third group received an aqueous solution of the
cation to give 20 lg Cr/kg body mass. The last group was injected
with an equal volume of doubly deionized water daily and served as
the control. Sixteen STZ-induced diabetic rats, ZKL rats, and ZKO
rats were each split into two groups of eight rats. For each, one
group of eight received doubly deionized water and served as a
control while the other group received an aqueous solution of the
cation to give 20 lg Cr/kg body mass. Rats were injected with
trimer-containing solutions rather than fed food with added trimer
to eliminate differences that might be associated with absorption of
the compound. Based on the average food intake and body mass
of rats after 12 weeks of this study, the Cr content of an injection of
cation corresponding to 20 lg Cr/kg body weight is approximately
equivalent to the Cr content of the food of a rat consuming a diet
containing 300 lg Cr/kg diet, a pharmacological dosage. After
24 weeks, the healthy SD and Zucker animals were sacrificed by
carbon dioxide asphyxiation; because of health concerns, the STZinduced diabetic SD rats were sacrificed after 22 weeks. Liver,
kidney, heart, spleen, pancreas, testes, and epididymal fat were
quickly harvested and weighed on plastic weighboats. Part of the
largest lobe of the liver and a kidney were placed into plastic screwtop containers and frozen and stored for further analysis. The other
kidney and another portion of the lobe of the liver were used for
histology studies (vide infra). One rat had to be eliminated from the
ZKO supplemented group, the SD STZ-induced diabetic control
group, and the SD STZ-induced diabetic supplemented group because of technical problems; thus, statistics for these groups are
based on seven rats rather than eight. The University of Alabama
Institutional Animal Use and Care Committee approved all
experiments involving rats.
854
Blood chemistry
Blood (1.5 mL) was collected from tail snips into polypropylene
tubes after 4, 8, 12, 16, 20, and 24 weeks of cation or H2O administration. Prior to blood collection, animals were fasted 12–
15 h. Immediately after blood removal, 0.5 mg/mL heparin and
10 mg/mL NaF were added to the blood. The blood was next
immediately centrifuged; the blood plasma was tested for glucose,
total cholesterol, triglycerides, low-density lipoprotein (LDL)
cholesterol (week 24 only), and high-density lipoprotein (HDL)
cholesterol using Diagnostic Kits from Sigma Chemical (St. Louis,
Mo.) and insulin using antibody-coated kits from ICN Biomedicals
(Costa Mesa, Calif.). The accuracy of the assays is within ±5%
except for the insulin assay where the accuracy is within ±<10%.
Glucose tolerance tests
After 6, 10, 14, 18, and 22 weeks of cation or H2O administration,
rats were injected under the skin with aqueous solutions of glucose
(1 mg/mL) such that each rats received 1.25 g glucose per kilogram
body mass [27]. After 2 h, blood (0.5 mL) was collected from tail
snips and handled as described above. The plasma insulin and
glucose concentrations were determined as described above.
Metal analyses
Samples of equal mass from three randomly chosen livers and
kidneys from each group were dried. All vessels used in the drying
were acid-washed. Iron concentrations were determined by the
method of Fish [28]. Chromium concentrations were determined by
graphite furnace atomic absorption spectroscopy utilizing the
method of Miller-Ihli [29]. For chromium analysis, dried tissue
samples were treated with 100 lL of 6% Mg(NO3)2.6H2O, dried,
and heated in a furnace overnight at 480 C. Samples were then
treated with 1 mL of nitric acid and returned to the furnace; this
was repeated until the entire sample was ashed. Ashed samples were
dissolved in nitric acid. Sample preparation blanks were analyzed,
and all data were blank corrected. Calibration was performed using
dilutions of a standard K2Cr2O7 solution. The detection limit
(3.29r) was approximately 0.25 lg/L Cr.
Histology of liver and kidney samples
A kidney and a portion of the largest lobe of the liver were preserved in 10% buffered formalin phosphate. The organs from three
randomly selected rats from each group were used for further analyses. Samples were stained with hematoxylin and eosin for analyses. Samples were examined for cell types, distribution, and
general morphology.
Statistical analyses
Statistical analyses were performed by analysis of variance. Numerical values in the tables and text are presented as
mean±standard deviation, unless otherwise indicated; error bars in
figures represent SEM to keep overlaps to a minimum for presentation purposes. P values were calculated using standard deviations.
Results
The daily food intake (not shown) and the percentage
mass gain (average mass gain/average mass on day one
·100%) of the control groups and the corresponding
trinuclear cation-administered groups (Figs. 1, 2, 3)
Fig. 1. Percentage body mass increase of healthy male SpragueDawley rats supplemented with the biomimetic cation. No
significant differences between groups were found
were statistically equivalent throughout the 24-week
period. All animals appeared normal throughout (the
STZ-induced diabetic rats displayed signs of their diabetes), and no visible differences were observed between
the administered groups and their controls. Interestingly
for all but the ZKO rats, the trinuclear cation-administered rats had lower (although not statistically lower)
percent body mass gains; in fact the more cation the
healthy SD rats received, the lower their percent body
mass gain. The lack of a statistical difference in body
mass with Cr complex administration is consistent with
numerous other studies [30, 31, 32, 33, 34, 35].
Blood plasma variables were determined for all 10
groups after 4, 8, 12, 16, 20, and 24 weeks of administration (Tables 1, 2, 3) (except the STZ-treated rats
which were sacrificed after 22 weeks). Some striking
differences are apparent between the healthy SD rats and
the ZKO rats receiving the Cr complex and their respective controls. For the SD rats, after 24 weeks of
administration of trimer corresponding to 20 lg Cr
daily, plasma total cholesterol, triglycerides, insulin,
LDL cholesterol, and HDL cholesterol levels are all
significantly lower than those of the control. The reduction in insulin concentration is first observed after
just 4 weeks of administration and is also observed after
12, 16, and 20 weeks of administration. After week 20,
855
Fig. 2. Percentage body mass increase of male Sprague-Dawley
rats with STZ-induced diabetes supplemented with the biomimetic
cation (20 lg Cr). No significant differences between control and
supplemented group were found
the plasma insulin levels are clearly dependent on the
level of trimer supplementation. Plasma triglycerides
levels become significantly lower after week 12 and are
significantly lower as well after week 20. After week 24,
plasma triglycerides levels become lower with increasing
amount of trimer administration. Plasma total cholesterol levels for rats receiving 20 lg Cr/ kg body mass
first become significant lower at week 16 and remained
lower throughout the rest of the study. Plasma HDL
cholesterol levels are only significantly lower after
24 weeks for rats receiving the greatest amount of Cr
complex; lowering of plasma HDL levels would be a
matter for concern. It is interesting to note that the
group receiving 10 lg Cr had significantly lower glucose
levels towards the end of the study, after weeks 20 and
24. The effects on triglycerides and cholesterols with the
accompanying lowering of insulin levels, but with little
effect on glucose concentrations, suggests that the
complex is significantly increasing insulin sensitivity in
these rats; however, statistically significant results require the greater quantities of the complex.
The dependence of the blood variables with time was
also analyzed. While Table 1 clearly shows a trend for
plasma insulin, total cholesterol, and triglycerides levels
of rats administered the trimer to decrease with time
when compared to levels in control rats, most of these
Fig. 3. Percentage body mass increase of male Zucker lean and
Zucker obese rats supplemented with the biomimetic cation (20 lg
Cr). No significant differences between control and supplemented
group were found
decreases with time are not statistically significant. At
the 95% confidence limit, only the decrease in the
plasma total cholesterol levels of the rats receiving 20 lg
Cr and the triglycerides levels of the rats receiving 10 lg
and 20 lg Cr daily displayed a significant time dependence when compared to the levels of the control group.
In contrast to the healthy rats, the trimer has virtually
no effect in SD rats with STZ-induced diabetes. After
week 12, plasma HDL cholesterol levels are higher for
rats receiving the trimer than those of the control group;
after 16 weeks, total cholesterol levels are lower. The
only effect observed at multiple times was a lowering of
plasma insulin levels at the extremes of the study, after
week 4 and week 20. Similarly, few effects are observed
on ZKL rats. The trimer-receiving group possessed
lower plasma glucose levels after weeks 4 and 12 and had
lower plasma insulin levels after weeks 16 and 24.
Very large effects of Cr complex administration are
seen for the ZKO rats. Plasma insulin was lower for the
trimer groups after each 4-week period (i.e., after 4, 8,
12, 16, 20, and 24 weeks). Triglycerides were lower after
weeks 8, 12, 20, and 24, as was total cholesterol. HDL
cholesterol was lower after weeks 4, 16, 20, and 24. This
lowering of HDL cholesterol is actually a movement
toward restoring HDL levels to normal, as ZKO rats
856
Table 1. Effect of biomimetic cation on plasma variables of healthy male Sprague-Dawley (SD) rats after 4, 8, 12, 16, 20, and 24 weeks of
supplementation. Values are mean±standard deviation
Glucose (mg/dL)
Week 4
SD control
SD+5Cr
SD+10Cr
SD+20Cr
Week 8
SD control
SD+5Cr
SD+10Cr
SD+20Cr
Week 12
SD control
SD+5Cr
SD+10Cr
SD+20Cr
Week 16
SD control
SD+5Cr
SD+10Cr
SD+20Cr
Week 20
SD control
SD+5Cr
SD+10Cr
SD+20Cr
Week 24
SD control
SD+5Cr
SD+10Cr
SD+20Cr
Total cholesterol
(mg/dL)
Triglycerides
(mg/dL)
Insulin (lIU/mL) LDL cholesterol
(mg/dL)
HDL cholesterol
(mg/dL)
117±11
117±7
105±17
99±9
88.0±19.1
83.7±15.7
89.8±20.3
88±12.6
68.5±18.7
68.9±27.6
61.2±27.3
68.4±23.7
35.3±10.6
35.0±7.3
32.4±12.6
23.5±8.0a
NDc
ND
ND
ND
77.2±25.6
68.9±23.2
43.3±5.9a
53.4±10.4a
99.4±8
96.7±10
88.8±18
100.8±9
65.8±11.3
78.1±14.0
81.3±7.7a
76.2±12.9
70.3±33.7
95.7±37.2
58.6±38.1
72.0±26.9
42.0±7.0
53.0±5.2a
40.0±11.2
41.0±9.2
ND
ND
ND
ND
56.0±14.2
55.5±13.1
57.8±17.7
50.7±8.4
75.8±7.2
74.6±10.7
84.9±21.4
74.8±10.6
90.9±14.8
89.8±19.0
92.9±23.2
80.9±12.2
56.2±16.7
44.0±8.3
46.6±12.0
36.3±7.2a
48±8.0
49±3.8
36±14.1a
32±7.3a
ND
ND
ND
ND
63.5±7.2
74.8±19.3
56.2±14.8
64.7±15.7
139.9±22.0
153.3±24.0
151.8±10.6
159.9±27.1
105.9±13.4
106.5±11.1
109.7±16.8
85.3±12.8a
70.5±14.2
61.2±12.0
58.7±20.5
60.8±14.0
74±10.4
51±6.5a
53±13.6a
42±12.0a
ND
ND
ND
ND
63.4±10.7
68.9±11.3
71.6±9.1
70±12.3
147.7±10.8
154.5±17.3
126.4±12.6a
142.6±11.2
101.6±15.6
87.5±4.0a
93.5±13.0
85.4±8.4a
77.6±18.1
59.8±17.5a
53.4±20.2a
52.1±15.0a
165±9.2
116±8.9a
111±11.6a
86±6.5b
ND
ND
ND
ND
64.5±21.6
54.3±17.3
57.4±18.4
46.0±17.6
116.8±18.8
130.6±20.6
98.9±12.0a
105.5±13.0
126.2±26.2
110.5±13.3
102.0±17.6a
89.3±14.4a
188.3±32.8
169.2±39.0
122.0±11.8a
93.8±18.5b
69±12.3
58±8.4a
54±9.7a
49±9.0a
108.9±20.5
106.9±15.5
85.5±18.4a
76.7±16.8a
60.5±9.3
61.8±10.8
52.3±12.2
45.0±10.7a
a
P<0.05 for comparison against control
P<0.05 for comparison against control and other Cr concentrations
c
ND=not determined
b
Table 2. Effect of biomimetic cation (20 lg Cr) on plasma variables of male SD rats with STZ-induced diabetes after 4, 8, 12, 16, 20 and
22 weeks of supplementation. Values are mean±standard deviation
Week 4
SD diabetic control
SD diabetic+Cr
Week 8
SD diabetic control
SD diabetic+Cr
Week 12
SD diabetic control
SD diabetic+Cr
Week 16
SD diabetic control
SD diabetic+Cr
Week 20
SD diabetic control
SD diabetic+Cr
a
Glucose (mg/dL)
Total cholesterol
(mg/dL)
227.6±134
254.5±152.2
108.2±41.6
88.3±27.5
257.0±133.6
284.6±46.7
Triglycerides
(mg/dL)
Insulin (lIU/dL)
HDL cholesterol
(mg/dL)
82.5±39.2
69.9±35.5
35.1±8.7
17.7±6.4a
52.6±9.3
53.8±15.1
127.4±53.8
97.7±13.8
149.3±118.9
72.4±45.7
20.0±11.1
21.0±3.9
49.3±5.7
52.3±5.3
322.4±195.5
377.8±133.1
106.6±40.8
90.4±19.9
297.9±370.6
112.0±99.8
26±18.6
12.5±27.3
62.0±13.8
76.9±4.9a
426.8±230.6
533.7±207.4
188.3±48.7
109±30.4a
157.5±167.3
104.7±67.8
66±7.4
60±5.1
67.8±12.1
66±16.3
306.4±165.9
422.2±259.5
109.6±31.9
87.9±17.7
177.1+223.5
142.6±102.5
61±8.9
53±4.3a
55.9±25.9
66.9±17.5
P<0.05
have elevated HDL levels [36]. The LDL cholesterol
level after 24 weeks is especially notable, being only 36%
of the level in controls. While not lower after 24 weeks
of treatment, glucose levels were higher after week 8 and
week 16. [Effects on fasting glucose concentrations are,
thus, inconsistent between the different groups of
857
Table 3. Effect of biomimetic cation (20 lg Cr) on plasma variables of male Zucker lean rats (ZKL) and Zucker obese rats (ZKO) after 4,
8, 12, 16, 20, and 22 weeks of supplementation. Values are mean±standard deviation
Week 4
ZKL control
ZKL+Cr
ZKO control
ZKO+Cr
Week 8
ZKL control
ZKL+Cr
ZKO control
ZKO+Cr
Week 12
ZKL control
ZKL+Cr
ZKO control
ZKO+Cr
Week 16
ZKL control
ZKL+Cr
ZKO control
ZKO+Cr
Week 20
ZKL control
ZKL+Cr
ZKO control
ZKO+Cr
Week 24
ZKL control
ZKL+Cr
ZKO control
ZKO+Cr
a
Glucose
(mg/dL)
Total cholesterol
(mg/dL)
Triglycerides
(mg/dL)
Insulin (lIU/mL)
LDL cholesterol
(mg/dL)
HDL cholesterol
(mg/dL)
137.3±11.8
115.8±8.0a
178.0±29.4
160.7±26.9
102.5±24.7
91.7±5.3
183.5±25.0
180.6±50.9
69.1±21.9
63.1±15.2
225.4±55.0
206.4±63.8
32.0±4.0
28.0±8.4
249.0±2.3
175.0±9.5a
NDb
ND
ND
ND
62.8±5.8
58.6±4.2
114.8±20.5
91.7±10.0a
96.7±15.6
92.6±12.2
125.8±12.4
163.7±51.2a
87.6±12.1
79.8±10.0
178.6±30.8
140.5±24.4a
53.5±17.9
57.1±16.4
366.9±60.0
245.4±76.9a
43.0±3.8
38.5±6.3
295.0±2.6
278.0±6.9a
ND
ND
ND
ND
80.8±11.1
80.8±16.9
133.4±29.1
116.5±23.4
149.9±7.7
138.9±9.8a
148.8±22.1
163.8±36.3
85.8±13.2
79.3±10.3
201.1±65.8
154.4±43.1a
24.8±5.5
28.3±8.4
219.1±40.4
143.1±47.4a
36±7.0
37±7.4
310±4.0
250±4.5a
ND
ND
ND
ND
82.2±16.7
79.5±8.0
193.2±67.7
169.8±60.5
114.5±12.9
111.4±4.5
124.9±10.5
143.2±22.7a
89.3±10
102.5±17.9
337.7±157.1
221.5±96.9
66.2±23.1
61.4±20.3
513±332.8
516.7±297
67±8.5
43±7.1a
231±4.1
170±6.5a
ND
ND
ND
ND
68.2±9.0
74.3±15.2
238.6±101.2
160.3±37.0a
96.1±19.9
108.8±14.6
136.6±26.6
147.3±12.1
96.7±13.8
101.5±16.0
335.6±161.7
190.1±74.1a
90.2±28.1
70.0±16.8
802.4±223.7
557.3±179.9a
54±11.8
50±14.5
238±6.4
204±7.6a
ND
ND
ND
ND
52.2±11.9
42.9±7.9
121.8±28.9
78±24.7a
140.1±13.3
135.4±28.4
154.9±42.5
150.0±27.9
146.4±28.9
138.2±47.9
568.4±172.0
301.1±107.9a
164.5±51.0
129.6±47.3
596.0±33.3
522.4±29.9a
113.7±18.3
123.7±15.0
228.3±161.8
83.3±48.3a
81.2±13.9
73.7±24.3
200.4±47.6
130.9±29.8a
91.0±7.6
61.0±9.0a
286.0±7.0
235.0±5.0a
P<0.05
ND=not determined
b
groups. Healthy rats receiving 10 lg Cr and ZKL rats
have reduced concentrations at two of the six time
points, while no effects were observed for the type I diabetes models and the concentrations were statistically
higher at the two time points indicated above for the
type II diabetes models.] Overall, ZKO rats seem to have
significantly increased insulin sensitivity in response to
treatment with the trimer in a fashion similar to that of
the healthy SD rats.
For the ZKO rats, only the time-dependent decrease
in plasma total cholesterol for the trimer-receiving rats
compared to controls was statistically significant at the
95% confidence level. In fact, the ratio of the total
cholesterol levels of the trimer-receiving rats compared
to those of the control rats decreases almost linearly with
time (r2=0.945).
Similar trends are observed in the results from glucose tolerance tests (Tables 4, 5, 6). For healthy SD rats,
2-h glucose and insulin levels of rats receiving the trimer
are consistently lower than those of controls. Almost no
effects are observed for the SD rats with STZ-induced
diabetes. ZKL rats receiving the trimer had lower plasma 2-h glucose levels only after week 6 but had lower
plasma insulin levels after weeks 6, 10, 14, and 18. ZKO
rats receiving the trimer had lower plasma glucose levels
after only 22 weeks of treatment but had lower insulin
Table 4. Glucose tolerance test blood variables after 2 h for
healthy male SD rats
Week 6
SD control
SD+5Cr
SD+10Cr
SD+20Cr
Week 10
SD control
SD+5Cr
SD+10Cr
SD+20Cr
Week 14
SD control
SD+5Cr
SD+10Cr
SD+20Cr
Week 18
SD control
SD+5Cr
SD+10Cr
SD+20Cr
Week 22
SD control
SD+5Cr
SD+10Cr
SD+20Cr
a
Glucose (mg/dL)
Insulin (lIU/mL)
178.7±42.9
172.3±37.3
136.1±14.9a
157.9±32.3
100.0±6.1
75.0±6.4a
73.0±7.7a
101.0±8.2
140.8±13.5
136.7±18.4
127.5±8.9a
139.3±9.3
90.0±5.0
96.0±5.2
66.0±10.2a
70.0±5.5a
96.7±17.5
82.7±19.3
76.7±16.2a
78.2±10.3a
162.5±5.5
160.0±3.3
96.0±3.1b
125.0±5.4a
135±12
124.1±17.2a
96.3±13.4a
84.0±5.7a
107.3±7.8
114±5.9
87.5±7.7a
88±6.6a
108.0±13.6
90.7±12.1a
80.7±9.6a
66.8±11.7a
158.3±6.6
161.1±4.9
107.0±8.6b
125.0±8.2a
P<0.05 for comparison against control
P<0.05 for comparison against other Cr concentrations
b
858
Table 5. Glucose tolerance test blood variables after 2 h for male
SD rats with STZ-induced diabetes
Week 6
SD diabetic control
SD diabetic+Cr
Week 10
SD diabetic control
SD diabetic+Cr
Week 14
SD diabetic control
SD diabetic+Cr
Week 18
SD diabetic control
SD diabetic+Cr
Weeek 22
SD diabetic control
SD diabetic+Cr
a
Glucose (mg/dL)
Insulin (lIU/mL)
348.1±104.4
428.7±75.6
47.5±20.2
35.0±13.7
332.2±77.1
333.7±92.0
46.0+19.8
26.0±17.0a
160.9±69.8
205.2±64.6
39±17.9
30±17.1
187±75
371±125.5a
245.7±100.4
263.3±99.8
34.3±13.5
27.1±9.8
36.14±19.7
48.5±18.8
P<0.05 for comparison against control
Table 6. Glucose tolerance test blood variables after 2 h for male
ZKL and ZKO rats
Week 6
ZKL control
ZKL+20Cr
ZKO control
ZKO+20Cr
Week 10
ZKL control
ZKL+20Cr
ZKO control
ZKO+20Cr
Week 14
ZKL control
ZKL+20Cr
ZKO control
ZKO+20Cr
Week 18
ZKL control
ZKL+20Cr
ZKO control
ZKO+20Cr
Week 22
ZKL control
ZKL+20Cr
ZKO control
ZKO+20Cr
a
levels after 6, 10, 14, and 18 weeks of treatment. Thus, in
all groups receiving the Cr complex, except the diabetic
SD rats, significant effects on 2-h insulin levels were
observed. These studies are also consistent with the
ZKO and healthy SD rats displaying increased insulin
sensitivity, and the ZKL rats would also appear to display increased sensitivity to insulin in these studies.
Comparison of organ masses after 24 weeks of administration revealed some statistically significant variations from those of the controls (Tables 7, 8, 9);
however, no trends are apparent between groups. For
example, for the healthy SD rats, those receiving the
trimer had larger hearts, larger testes, and larger spleens
than the control group. Yet, ZKO rats receiving the
complex had smaller hearts, smaller testes, and no difference in spleen mass compared to controls. In this
laboratory’s previous study with healthy SD rats receiving the trinuclear cation for 12 weeks, rats receiving
the trimer had on average a larger pancreas mass and a
lower testes mass compared to those of controls [23];
neither was observed in this study. No organs were
visibly different for any of the groups.
Histopathological analyses detected no differences in
tissue samples from the kidneys or livers of any of the
groups receiving the trimer compared to those of their
controls.
Chromium has been suspected of potentially adversely affecting iron metabolism [3]. Thus, the iron and
chromium levels of liver and kidney tissue from rats of
each group were examined. No differences in the iron
content of the liver and kidney of the rats (Table 10)
Glucose (mg/dL)
Insulin (lIU/mL)
168.9±32.1
134.5±19.2a
266.1±76.7
241.7±84.6
123.0±9.5
112.0±4.5a
475.0±2.4
460.0±4.2a
122.3±13.9
122.2±10.8
166.0±36.8
187.8±45.2
62.0±2.9
73.0±5.8a
400±2.5
396.0±2.7a
144.3±39.2
164.5±41.5
128.5±22.2
136.6±12.4
46.0±10.3
116.0±5.0a
392.0±3.4
388.0±2.6a
Table 8. Percent relative organ masses (tissue mass/body mass ·
100%) of male SD rats with STZ-induced diabetes after 22 weeks
of supplementation with the biomimetic cation (20 lg Cr). Values
are mean±standard deviation
192.8±15.5
211.8±31.5
251.8±83.9
209.6±20.4
71±6.2
77±6.3a
350±4.8
270.8±4.7a
Tissue
SD diabetic control
SD diabetic+20 lg Cr
112.6±14.5
124.3±27.2
128.9±14.0
102.4±14.6a
108.3±14.8
116.7±7.8
288.9±7.2
294.5±5.6
Heart
Liver
Kidney
Pancreas
Testes
Epididymal fat
Spleen
0.36±0.06
4.38±0.44
1.12±0.13
0.26±0.14
0.75±0.06
1.57±1.08
0.19±0.06
0.38±0.06
4.90±1.19
1.36±0.28a
0.26±0.11
0.86±0.26
1.18±1.05
0.24±0.07
a
P<0.05 for comparison against control
Table 7. Percent relative organ
masses (tissue mass/body
mass·100%) of healthy male
SD rats after 24 weeks of supplementation with the biomimetic cation (20 lg Cr). Values
are means±standard deviation
P<0.05 for comparison against control
Tissue
SD control
SD+5 lg Cr
SD+10 lg Cr
SD+20 lg Cr
Heart
Liver
Kidney
Pancreas
Testes
Epididymal fat
Spleen
0.50±0.05
2.95±0.29
0.90±0.08
0.40±0.03
0.75±0.04
2.59±0.42
0.36±0.03
0.55±0.03a
3.14±0.55
0.96±0.09
0.42±0.05
0.83±0.04a
2.43±0.49
0.41±0.05a
0.59±0.04a
2.78±0.20
0.96±0.07
0.47±0.05a
0.83±0.07a
2.12±0.63
0.48±0.03a
0.56±0.04a
3.02±0.24
0.91±0.04
0.41±0.05
0.87±0.07a
2.33±0.51
0.46±0.07a
a
P<0.05 for comparison against control
859
Table 9. Percent relative organ
masses (tissue mass/body
mass·100%) of male ZKL and
ZKO rats after 24 weeks supplementation with the biomimetic cation (20 lg Cr). Values
are mean±standard deviation
Tissue
ZKL control
ZKL+20 lg Cr
ZKO control
ZKO+20 lg Cr
Heart
Liver
Kidney
Pancreas
Testes
Epididymal fat
Spleen
0.61±0.03
3.13±0.32
1.04±0.09
0.60±0.09
0.96±0.13
2.23±0.62
0.59±0.39
0.66±0.06a
3.37±0.50
1.08±0.13
0.57±0.05
0.96±0.14
2.04±0.45
0.49±0.03
0.42±0.02
4.45±0.63
0.85±0.11
0.47±0.22
0.63±0.03
2.90±0.20
0.32±0.03
0.39±0.02a
3.41±0.49a
0.76±0.04a
0.38±0.04
0.55±0.10a
3.11±0.73
0.34±0.02
a
P<0.05 for comparison against control
Table 10. Iron and chromium levels of liver and kidney tissues from male SD healthy rats (24 weeks), male SD rats with STZ-induced
diabetes (22 weeks), and male ZKL and ZKO rats (24 weeks). Values are mean±standard deviation
SD control
SD+5Cr
SD+10Cr
SD+20Cr
SD STZ control
SD STZ+20Cr
ZKL control
ZKL+20Cr
ZKO control
ZKO+20Cr
a
Liver Fe (lg/g dry mass)
Kidney Fe (lg/g dry mass)
Liver Cr (lg/g dry mass)
Kidney Cr (lg/g dry mass)
341.52±56.00
333.86±25.41
371.13±70.83
359.66±103.39
343.51±81.13
296.36±51.67
278.13±41.53
374.23±30.70a
328.99±70.81
313.22±36.46
254.36±25.96
315.74±74.56
303.23±107.18
270.06±64.47
246.45±37.50
309.93±66.00
385.53±75.28
257.83±36.72a
390.89±116.48
338.89±101.72
0.11±0.09
0.34±0.39
1.62±0.40a
0.78±0.70
0.19±0.19
0.48±0.46
0.51±0.03
1.76±0.19a
0.50±0.09
1.18±0.43a
0.78±0.76
1.29±0.87
2.07±0.20a
1.81±0.76
1.80±0.17
1.36±0.40
0.77±0.67
1.69±0.16a
0.53±0.22
1.89±0.20a
P<0.05 for comparison against control
were observed except for the ZKL rats receiving the
trimer, which had higher liver iron concentrations and
lower kidney Fe concentrations that their controls. Not
surprisingly, treatment with the trimer results in an accumulation of Cr in the liver of some of the rats. Both
ZKL and ZKO rats possessed significantly elevated liver
Cr levels. For the healthy SD rats, all the trimer-supplemented groups had elevated Cr levels, but the increase was only statistically significant for the group
receiving 10 lg Cr/ kg body mass. [Interestingly, the Cr
concentrations of the organs from the control rats are
somewhat high (for example, compared to those of [32]);
however, the rat’s diet contained more chromium than
found for other commercial diets (e.g. [32]), which may
explain the elevated Cr levels in the control rats.] Yet, no
statistically significant increase in liver Cr was observed
for the type I diabetic model SD rats. This raises the
question of whether the lack of an effect of the trimer on
plasma parameters could be related to the low Cr levels;
for example, could the diabetic rats have a different
degree of cellular absorption of the trimer?
Discussion
Dietary chromium(III) is a nutrient, not a therapeutic
agent [37]. Consequently, an individual who is not
deficient in chromium would not be expected to benefit
from the intake of additional chromium. Most recent
studies on the effects of Cr(pic)3 or other chromium
supplements on healthy individuals observed no beneficial effects from supplementation [2, 38, 39, 40, 41], as
one would expect if such individuals are not chromium
deficient. The only potential demonstration of chromium deficiency in humans is for patients receiving
parenteral nutrition [42]. Recently, the dietary guidelines for chromium intake have been lowered from
200 lg per day for an adult to 35 lg for an adult male
and 25 lg for an adult female [43], suggesting nearly all
Americans have an adequate daily dietary intake of
chromium.
Similarly for rats, observations of effects of diet
supplementation with chromium on rats requires strict
environmental control, such as the removal of any
stainless steel objects, to guarantee chromium deficiency
[33]. The only consistent effect of chromium deficiency
appears to be higher plasma insulin levels in glucose
tolerance tests [33, 44]. Rats on a Cr-deficient and high
fat diet may have higher plasma insulin levels and, in
glucose tolerance tests, higher triglycerides areas [45].
Hence, supplementation of the diet of healthy rats on a
normal diet with absorbable sources of chromium
should have no effect, as has been observed previously
[30, 31, 32, 34]. For example, in the largest such study,
feeding rats a diet containing from 5 to 100 lg Cr/kg
diet as chromium picolinate or chromium chloride for
24 weeks had no effect on serum glucose, cholesterol, or
triglycerides concentrations [32]. Thus, the increased
insulin sensitivity observed in this study arises from the
trinuclear Cr propionate complex, not from chromium(III) itself.
The observation that dietary chromium(III) serves as
a nutrient and not a therapeutic agent is easily rationalized, based on the proposed mechanism of chromium
860
action [1, 2]. The biologically active form of chromium,
chromodulin, is maintained in insulin-sensitive cells in
its apo (metal-free) form. In response to insulin, chromium concentrations in the blood decrease as chromium
is moved by transferrin to insulin-sensitive cells [21, 46,
47, 48]. Chromodulin has a large chromic ion binding
constant and becomes loaded with the metal ion. The
holochromodulin is consequently primed so that, in the
presence of insulin, insulin receptor tyrosine kinase is
activated and presumably held in its active form; consequently, chromium is proposed to act as part of an
insulin signaling autoamplification mechanism [1, 2].
The levels of chromodulin are apparently kept under
homeostatic control [21]. If sufficient chromium is
maintained in the blood and other body stores, chromium is mobilized properly, resulting in the generation
of sufficient quantities of holochromodulin and increased insulin signaling. Increased Cr intake results in
increased urinary chromium loss [1]. Hence, supplemental chromium should have no beneficial effect.
This is consistent with the present work. In contrast
to chromodulin, the biomimetic [Cr3O(O2CCH2CH3)6
(H2O)3]+ can bind and stimulate the insulin receptor (in
a fashion similar to chromodulin) while not being under
homeostatic control; hence, to the extent the cation remains intact, the functional mimetic could potentially
trap the insulin-stimulated insulin receptor in vivo in its
active conformation beyond its normal levels, resulting
in increased insulin signaling and subsequent cellular
action (i.e., increased insulin sensitivity). Insulin-deficient STZ-induced diabetic rats had no significant effects
from the administration of the biomimetic, consistent
with the biomimetic only being able to stimulate the
kinase activity of the insulin receptor previously activated by insulin.
As the biomimetic degrades in vivo, propionate is
likely to be released. Propionate has been proposed to be
able to lower plasma cholesterol and glucose concentrations; however, these results are quite controversial
[49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63,
64]. Propionate occurs naturally in mammals, although
it is actually generated by bacteria. Cecal and colonic
fermentation of dietary fiber results in the production of
short-chain fatty acids, including acetate, propionate,
and butyrate. These agents may be responsible in part
for the action of dietary fiber to reduce plasma cholesterol levels. The results of studies with rats given propionate have recently been reviewed [65]. Effects from
propionate were generally observed in healthy rats or
genetically obese rats consuming diets composed of 0.5–
5% propionate, while other researchers found no effects
for diets containing up to 10% propionate. Similarly
studies with perfused rat liver and isolated rat hepatocytes have generated conflicting results.
Fortunately, the maximum amount of propionate
provided daily in this study and the previous study of the
effects of the trimer in rats (56 lg/kg body mass) [23] is
far below the amounts required to observe effects in the
above studies. Similarly, the lowering of plasma insulin,
cholesterol, and triglycerides levels without affecting
plasma glucose levels by the trimer is inconsistent with
the studies of the effects of propionate in rats, even if
sufficient levels of propionate were generated. Similarly,
the time required for effects to be observed on HDL,
triglycerides, and total cholesterol levels for the healthy
SD rats in this study is different from those in the previous studies, which observed effects in hours or days.
Propionate has been reported not to effect triglycerides
levels [65], unlike the effects observed in the current
work.
Consequently, the effects of the trimer observed herein
and previously observed in rats are best attributed to the
cation and not its degradation products. A recent study
has shown that the trimer has a lifetime of less than 24 h
in vivo [65]. Recent cell culture studies using human liver
carcinoma (HepG2) have shown that the trimer is actively transported into cells (by a yet unknown mechanism), unlike CrCl3 or Cr(pic)3; two hours after
intravenous injection of rats with 51Cr-labeled trimer,
90% of 51Cr in liver cells from the labeled trimer is
localized to the microsomes, indicating selective transport into these organelles [66]. The molecular weight of
the species in the microsomes is similar to that of the
trimer, suggesting it may reach these organelles intact.
The effects seen in rats injected daily with chromium
presumably then must arise from only a small period of
time in which the complex remains intact. The effects of
the chromium cation when given in smaller amounts
every 4 or 6 h rather than every 24 h and when given
orally need to be examined. This laboratory has also
given rats sodium propionate intravenously in amounts
equivalent to the propionate contained in the largest
quantity of trimer used in the current study. Blood
plasma levels were measured after 4, 8, 2, 16, 20, and
24 weeks and 2-h glucose tolerance tests were performed
after weeks 6, 10, 14, 18, and 22. The propionate had no
effect on fasting plasma glucose, insulin, triglycerides,
and total or HDL cholesterol levels; similarly, the propionate had no effect on plasma glucose or insulin levels
in response to glucose administration (Clodfelder and
Vincent, unpublished results). Thus, the effects were in
stark contrast to those from the trimer (where, for example, lowering of insulin levels in response to glucose
was significant at weeks 10, 14, 18, and 22).
In the last decade, vanadium complexes have also
been tested as potential therapeutic agents for the
treatment of the symptoms of diabetes [67]. The proposed mechanism of action of these complexes is distinctly different that than proposed above for the
trinuclear chromium cation, as the vanadium complexes
are believed to act as phosphatase inhibitors. The effects
of the vanadium complexes and chromium complex on
STZ-treated and ZKO rats are distinctly different.
Administration of the vanadium complexes results in
increased insulin sensitivity and lower cholesterol levels
for both the type I and type II diabetes models [67],
in contrast to effects on only the ZKO rats being
observed with the chromium complex.
861
Conclusion
As demonstrated in Tables 1 and 3, the functional biomimetic has a striking effect on plasma triglycerides,
total cholesterol, LDL cholesterol, and insulin levels
after 24 weeks of supplementation in healthy SD and
ZKO rats. The effects by the biomimetic compound on
insulin, cholesterol, and triglycerides with continuing
administration suggest that the trinuclear complex
serves not as simply a chromium source but possesses
an intrinsic activity, in contrast to other sources of
chromium previously examined. Also, no toxic effects
were observed for supplementation with the trinuclear
cation. These results suggest that the compound may
have potential as a therapeutic agent. According to the
proposed mechanism for the mode of action of
LMWCr, the biomimetic should increase the signaling
of insulin and result in increased insulin sensitivity. In
this regard, it is notable that the decrease in triglycerides, LDL cholesterol, and total cholesterol is consistent with the effects of insulin on adult-onset diabetic
patients [68].
Acknowledgements The authors wish to thank Dion Hepburne,
Bryan Gullick, Jason Hatfield, and Kiyan McCormick and
Dr. James A. Neville and the staff of The University of Alabama
Animal Care Facility for assistance with the rat studies. Prof.
Joseph Thrasher kindly allowed the authors access to a graphite
furnace atomic absorption spectrometer. Funding was provided by
the American Diabetes Association (J.B.V.).
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
Vincent JB (2001) Polyhedron 20:1–26
Vincent JB (2000) Acc Chem Res 33:503–510
Lukaski HC (1999) Annu Rev Nutr 19:279–301
Vincent JB (2000) J Nutr 130:715–718
Vincent JB (2000) Nutr Rev 58:67–72
Davis CM, Vincent JB (1997) Biochemistry 36:4382–4385
Davis CM, Vincent JB (1997) Inorg Chem 36:5316–5320
Yamamoto A, Wada O, Suzuki H (1988) J Nutr 118:39–45
Davis CM, Vincent JB (1997) Arch Biochem Biophys 339:335–
343
Vincent JB (1994) J Nutr 124:117–118
Harton A, Terrell K, Huffman JC, MacDonald C, Beatty A, Li S,
O’Connor C, Vincent JB (1997) Inorg Chem 36:4875–4882
Donald S, Terrell K, Robinson K, Vincent JB (1995) Polyhedron 14:971–976
Ellis T, Glass M, Harton A, Folting K, Huffman JC, Vincent
JB (1995) Inorg Chem 33:5522–5527
Nagi M, Harton A, Donald S, Lee Y-S, Sabat M, O’Connor
CJ, Vincent JB (1995) Inorg Chem 34:3813–3820
Speetjens JK, Collins RA, Vincent JB, Woski SA (1999) Chem
Res Toxicol 12:483–487
Speetjens JK, Parand A, Crowder MW, Vincent JB, Woski SA
(1999) Polyhedron 18:2617–2624
Johnson MK, Powell DB, Cannon RD (1981) Spectrochim
Acta Part A 37:995–1006
Anderson RA, Polansky MM, Bryden NA, Roginski EE,
Patterson KY, Reamer DC (1982) Diabetes 31:212–216
Yamamoto A, Wada O, Ono T (1984) J Inorg Biochem 22:91–
102
Wu GY, Wada O (1981) Sangyo Ikgaku 23:505–512; (1981)
Chem Abstr
21. Clodfelder BJ, Emamaullee J, Hepburn DDD, Chakov NE,
Nettles HS, Vincent JB (2001) J Biol Inorg Chem 6:608–
617
22. Wada O, Wu GY, Yamamoto A, Manabe S, Ono T (1983)
Environ Res 32:228–239
23. Sun Y, Mallya K, Ramirez J, Vincent JB (1999) J Biol Inorg
Chem 4:838–845
24. American Diabetes Association (2001) Diabetes: 2001 vital
statistics. American Diabetes Association. Alexandria, Va
25. Earnshaw A, Figgis BN, Lewis J (1966) J Chem Soc A 1656–
1663
26. Kraszeski JL, Wallach S, Verch RL (1979) Endocrinology
104:881–885
27. Schwarz K, Mertz W (1959) Arch Biochem Biophys 85:292–
295
28. Fish WW (1988) Methods Enzymol 158:357–364
29. Miller-Ihli NJ (1996) J Food-Comp Anal 9:290–300
30. Hasten DL, Hegsted M, Keenan MJ, Morris GS (1997) Nutr
Res 17:283–294
31. Hasten DL, Hegsted M, Keenan MJ, Morris GS (1997) Nutr
Res 17:1175–1186
32. Anderson RA, Bryden NA, Polansky MM (1997) J Am Coll
Nutr 6:273–279
33. Striffler JS, Law JS, Polansky MM, Bhathena SJ, Anderson
RA (1995) Metabolism 44:1314–1320
34. Morris GS, Guidry KA, Hegsted M, Hasten DL (1995) Nutr
Res 15:1045–1052
35. O’Flaherty EJ, McCarty CP (1978) J Nutr 108:321–328
36. Sparks JD, Shaw WN, Corsetti JP, Bolognino M, Pesek JF,
Sparks CE (2000) Metabolism 49:1424–1430
37. Anderson RA (1989) Sci Total Envir 86:75–81
38. Campbell WW, Joseph LJO, Davey SL, Cyr-Campbell D,
Anderson RA, Evans WJ (1999) J Appl Physiol 86:29–39
39. Walker LS, Bemben MG, Bemben DA, Knehans AW (1998)
Med Sci Sports Exerc 30:1730–1737
40. Lukaski HC, Bolonchuk W, SidersWA, Milne DB (1996) Am J
Clin Nutr 63:954–965
41. Anderson RA (1998) Nutr Rev 56:266–270
42. Anderson RA (1995) Nutrition 11:83–86
43. Trumbo P, Yates AA, Schlicker S, Poos M (2001) J Am Dietetic Assoc 101:294–301
44. Striffler JS, Polansky MM, Anderson RA (1999) Metabolism
48:1063–1068
45. Striffler JS, Polansky MM, Anderson RA (1998) Metabolism
47:396–400
46. Morris BW, Gray TA, MacNeil S (1993) Clin Chem 84:477–
482
47. Morris BW, MacNeil S, Stanley K, Gray TA, Fraser R (1993)
J Endocrinol 139:339–345
48. Morris BW, Blumsohn A, MacNeil S, Gray TA (1992) Am J
Clin Nutr 55:989–991
49. Chen W-JL, Anderson JW, Jennings D (1984) Proc Soc Exp
Biol Med 175:215–218
50. Nishina PM, Freeland RA (1990) J Nutr 120:668–673
51. Beaulieu KE, McBurney MI (1992) J Nutr 122:241–245
52. Berggren AM, Nyman EM, Lundquist I, Bjorck IM (1996) Br
J Nutr 76:287–294
53. Kishimoto Y, Wakabayashi S, Takeda H (1995) J Nutr Sci
Vitaminol 41:73–81
54. Hara H, Haga S, Kasai T, Kiriyama S (1998) J Nutr 128:688–
693
55. Hara H, Haga S, Aoyama Y, Kiriyama S (1999) J Nutr
129:942–948
56. Levrat MA, Favier ML, Moundras C, Remesy C, Demigne C,
Morand C (1994) J Nutr 124:531–538
57. Illman RJ, Topping DL, McIntosh GH, Trimble RP,
Storer GB, Taylor MN, Cheng B-Q (1988) Ann Nutr Metab
32:97–107
58. Wright RS, Anderson JW, Bridges SR (1990) Proc Soc Exp
Biol Med 195:26–29
59. Demigne C, Morand C, Levrat MA, Besson C, Moundras C,
Remesy C (1995) Br J Nutr 74:209–219
862
60. Lin Y, Vonk RJ, Slooff MJ, Kuipers F, Smit KJ (1995) Br J
Nutr 74:197–207
61. Anderson JW, Bridges SR (1984) Proc Soc Exp Biol Med
177:372–376
62. Boillot J, Alamowitch C, Berger AM, Lou J, Bruzzo F, Bornet
FR, Slama G (1995) Br J Nutr 73:241–251
63. Chan TM, Freedland RA (1972) Biochem J 127:539–543
64. Blair JB, Cook DE, Lardy HA (1973) J Biol Chem 248:3608–
3614
65. Shute AA, Chakov NE, Vincent JB (2001) Polyhedron
20:2241–2252
66. Shute AA, Vincent JB (2002) J Inorg Biochem (in press)
67. Poucheret P, Verma S, Grynpas MD, McNeill JH (1998) Mol
Cell Biochem 188:73–80
68. Brunzell JD, Chait A (1990) In: Rifkin H, Porte D Jr (eds)
Ellenberg and Rifkin’s diabetes mellitus: theory and practice.
Elsevier, New York, pp 756–767
Polyhedron 20 (2001) 2241– 2252
www.elsevier.com/locate/poly
The stability of the biomimetic cation
triaqua-m-oxohexapropionatotrichromium(III) in vivo in rats
Amanda A. Shute, Nicole E. Chakov, John B. Vincent *
Department of Chemistry and Coalition for Biomolecular Products, The Uni6ersity of Alabama, Tuscaloosa, AL 35487 -0336, USA
Received 5 February 2001; accepted 26 April 2001
Abstract
The synthetic biomimetic triaqua-m-oxohexapropionatotrichromium(III) nitrate when given intravenously has been shown
previously to lower fasting blood plasma triglycerides and cholesterol concentrations in rats; thus, the cation has the potential to
serve as a therapeutic agent. Its ability to function in vivo presumably is dependent on its ability to mimic the action of the
natural, bioactive, chromium-binding oligopeptide chromodulin in stimulating insulin receptor kinase activity. For this to happen,
the cation presumably should be incorporated into insulin-sensitive cells intact. Examination of the distribution of 51Cr and 14C
from 51Cr- and [1-14C]-propionate labeled trinuclear cation in rats after injection with the trinuclear complex for 2 weeks suggests
that the cation enters cells intact; however, the cation appears to degrade within 24 h after injection. Chromium from the complex
is excreted in the urine as chromodulin. © 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Chromium; Chromodulin; Propionate; Diabetes; Rat
1. Introduction
The oligopeptide chromodulin (also known as lowmolecular weight chromium-binding substance or
LMWCr) may be the biologically active form of
chromium in mammals [1 – 4]. The naturally occurring
oligopeptide has been proposed to function as part of a
unique autoamplification system for insulin signaling
[1,2]. In this mechanism, apochromodulin is stored in
insulin-sensitive cells. In response to increases in blood
insulin concentrations (as would result from increasing
blood sugar concentrations after a meal), insulin binds
to its receptor bringing about a conformation change
which results in the autophosphorylation of tyrosine
residues on the internal side of the receptor. This
transforms the receptor into an active tyrosine kinase
and transmits the signal from insulin into the cell. In
response to insulin, chromium is moved from the blood
to insulin-sensitive cells. Here, the chromium flux results in the loading of apochromodulin with chromium.
The holochromodulin then binds to the receptor, pre* Corresponding author. Tel.: +1-205-348-9203; fax: + 1-205-3489104.
E-mail address: [email protected] (J.B. Vincent).
sumably assisting to maintain the receptor in its active
conformation, amplifying its kinase activity.
Recent research on the bioactive oligopeptide chromodulin has inspired efforts to synthesize new
chromium –carboxylate assemblies. Known assemblies
with nuclearity greater than two (but less than eight)
possess four types of cores: symmetric or ‘basic carboxylate’ [5] and unsymmetric [6] Cr3O, Cr3(OH)2 [7],
and Cr4O2 [8–10]. Numerous examples of the type
containing the symmetric Cr3O core (Fig. 1) have been
well characterized, and interest in these complexes dates
back to the late 19th century [5]. The other cores have
only been prepared during the last decade using the
symmetric trinuclear complexes as starting materials.
Fig. 1. Structure of chromium ‘basic carboxylate’ cations. For compound 1, R =CH2CH3 and L= H2O.
0277-5387/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 2 7 7 - 5 3 8 7 ( 0 1 ) 0 0 8 2 2 - 1
2242
A.A. Shute et al. / Polyhedron 20 (2001) 2241–2252
Given the novel role in the autoamplification of
insulin signal transduction for chromodulin and its
rather simple composition (carboxylate-rich oligopeptide binding four chromic ions), attempts have been
made to identify a functional model for chromodulin.
Such a biomimetic would be required to be soluble and
stable in aqueous solution. Few of the known trinuclear
and tetranuclear Cr(III) oxo(hydroxo)-bridged carboxylate assemblies are soluble in water. On the basis of
these requirements, two assemblies have been examined: [Cr3O(O2CCH3)6(H2O)3]+, and [Cr3O(O2CCH2CH3)6(H2O)3]+ (1) [11]. Both possess the symmetric
basic carboxylate-type structure comprised of a planar
triangle of chromic ions with a central m3-oxide. The
acetate complex does not activate the tyrosine protein
kinase activity of the active site fragment of insulin
receptor or of adipocytic membrane fragments in the
presence of insulin and actually inhibits the activity. In
stark contrast, the propionate analogue activates the
kinase activities in a fashion very similar to chromodulin. The kinase activity of the isolated receptor fragment, for example, is stimulated approximately
threefold with a dissociation constant of 1.00 nM [11].
The complex appears to be a functional biomimetic for
chromodulin and supports the existence of a multinuclear chromic assembly in chromodulin.
The propionate biomimetic has been found to have
striking in vivo effects, lowering plasma triglycerides,
total cholesterol, LDL cholesterol, and HDL cholesterol levels after 12 weeks of supplementation in rats at
a level of 20 mg Cr per kg body mass daily and
potentially lowering body mass and fat content [12]. No
acute toxic effects were observed for supplementation
with the compound, and it does not give rise to DNA
damage in in vitro studies as observed with the popular
nutritional supplement chromium picolinate [13,14].
Thus, the propionate complex could have potential as a
therapeutic agent. These results are also consistent with
the proposed mechanism. In contrast to chromium
supplements that only provide a source of chromium
and have no intrinsic activity, a functional biomimetic
(if it entered insulin-sensitive cells intact) could trap
insulin receptor in its active conformation, amplifying
insulin signaling and subsequent cellular activity.
However, the stability of the trinuclear propionate
complex in vivo is not known. As the complex is
biologically active in vivo, it presumably stays intact in
vivo for some period of time. If the complex is extremely stable, it could accumulate in cells and with
time shown to have detrimental side effects. However, if
the complex has an intermediate level of stability, it
could remain intact long enough to be active but ideally
also decompose with time such that its components or
other products could be eliminated from the body. In
this work efforts to determine the distribution and
concentration of the degradation products of the
biomimetic in rats are reported.
2. Materials and methods
51
CrCl3·6H2O in 0.5 M HCl was obtained from ICN;
sodium [1-14C]-propionate was obtained from Moravek
Biochemicals.
Labeled
derivatives
of
[Cr3O(O2CCH2CH3)6(H2O)3]+ (1) were made by the method
of Earnshaw et al. [15] except that either sodium [114
C]-propionate was used in place of propionic acid or
a tracer amount of 51CrCl3 (1–5 mCi) was mixed with
the initial aqueous solution of CrCl3. Male Sprague
Dawley rats (between 550 and 650 g) were maintained
for 15 days in metabolic cages. During the first 14 days,
the rats were injected in the tail vein each morning with
150 ml of an aqueous solution of a radiolabeled derivative of 1 containing 276 mg of the cation. Urine and
feces samples were collected every 12 h. On the morning
of day 15, the rats were sacrificed by carbon dioxide
asphyxiation, and tissue samples were harvested and
weighed. Subcellular liver fractions were obtained by
differential centrifugation according to established procedures [16–18]. Components of the urine and nuclear
hepatocyte fractions were separated by G-15 column
chromatography or Shodex OH-PAK HPLC. Before
being applied to chromatography columns, the nuclear
fractions were homogenized and then clarified by centrifugation. Chromatography columns were run with 50
mM NH4OAc buffer, pH 6.5; 0.1 M NaCl was used as
the mobile phase for HPLC experiments. For scintillation counting, feces and tissues were homogenized in
water with a Waring blender. Aquasol-2 (Packard) was
used as the scintillation fluid. Experiments with 51Cr
were performed in duplicate.
UV –Vis measurements were made on a Hewlett–
Packard 8453 spectrophotometer. Gamma counting
was performed on a Packard Cobra II auto-gamma
counter. Scintillation counting was performed on a
Packard 1900 TR scintillation counter. All procedures
were performed using doubly deionized water unless
otherwise noted and performed with plasticware
whenever possible. The University of Alabama Institutional Animal Use and Care Committee approved the
procedures involving the use of rats.
3. Results
In order to determine the stability and distribution of
biomimetic 1 in rats, rats were treated daily for 2 weeks
with 51Cr- or 14C-labeled 1, allowing the fate of the
cation, the chromium, and the organic component to be
followed. The complex was administered by injection to
allow the amount entering the bloodstream to be
known and avoiding other problems potentially associated with absorption from the gastrointestinal tract.
The amount of cation injected daily (276 mg) is six times
A.A. Shute et al. / Polyhedron 20 (2001) 2241–2252
Fig. 2. Urinary and fecal 51Cr loss during the 2-week period of treatment with
injection of an aliquot of a solution of labeled compound 1.
greater than that used previously in studies to examine
the effect of the cation in rats [12]. The amount of Cr
received daily corresponds to 64 mg of chromium. This
quantity was chosen to increase the chances of observing intact cation; if this large quantity can readily be
degraded, then smaller amounts, corresponding to normal human dietary supplementation, should also readily be degraded. For comparison, human nutritional
supplements generally provide 200– 600 mg of Cr per
day. Assuming an average human body mass of 75 kg,
this is equivalent to 1.6– 4.8 mg of Cr per day for a 0.6
kg rat. Thus, the rats are receiving 10 – 40-fold more
Cr if given an amount comparable to humans. Because
approximately 0.5–5% of dietary chromium is absorbed
(depending on the form of Cr), the injection given to
the rats represents an even greater excess of chromium
51
2243
Cr-labeled compound 1. Time zero represents the time of first
when compared to that given to humans orally. Anderson et al. have fed rats diets containing up to 15 mg Cr
per kg body mass for 24 weeks without observing any
acute toxic effects [19]. Assuming 5% absorption, this
would correspond to 450 mg of Cr entering the bloodstream of a 0.6 kg rat; thus, the dosage of Cr in the
current work was not expected to have acute toxic
effects. No ill effects were observed from the injections;
all tissues appeared normal when harvested.
Urine samples collected during the 2 weeks of treatment with 51Cr-labeled trimer reveal a distinct pattern
of Cr excretion (Fig. 2). Loss of chromium is maximal
in the first 12 h after injection and significantly reduced
in the next 12 h. The pattern of chromium lost in the
feces is less distinct; however, chromium loss is generally greater 12–24 h after injection than in the period
2244
A.A. Shute et al. / Polyhedron 20 (2001) 2241–2252
immediately following injection. Urinary chromium
greatly exceeds fecal loss by up to approximately tenfold. Total daily fecal and urinary chromium loss
throughout the 2-week period was rather constant and
corresponded to approximately 30% of the daily quantity of chromium injected. Another less obvious feature
of the urinary chromium loss profile is that the
chromium loss 12–24 h after injection slowly rises from
day 1 to day 14. Previously, a three-compartment
model has been proposed to explain the kinetics of
chromium tissue exchange and distribution in studies
with rats and humans [20– 22]. Plasma chromium was
proposed to be in equilibrium with three pools of
chromium: a small pool with rapid exchange (T1/2 B 1
day), a medium pool with a medium rate of exchange
(days), and a large, slowly exchanging pool (months).
Chromium lost within 12 h of injection primarily then
reflects this small pool in rapid equilibrium; the slow
increase in chromium levels 12– 24 h after injection may
reflect chromium exchanging with the medium pool
that undergoes exchange over a period of days. To
determine the form(s) of chromium lost in the urine,
urine from several of the 12-h periods was subjected to
Shodex OH-PAK HPLC (Fig. 3); this type of size
exclusion chromatography revealed that regardless of
the time the samples were collected their elution profiles
were almost identical. 51Cr eluted in a single band
(other features in Fig. 3 are within the noise) eluting
just ahead of the major components that absorb UV
light at 210 nm. This band does not correspond to the
molecular weight of the cation; the band co-elutes with
chromodulin when it is applied to the column. Thus,
urinary chromium corresponds to chromium from the
degradation of 1 that subsequently binds to the
oligopeptide chromodulin, the only naturally occurring
species of this molecular weight known to bind
chromium. This is consistent with other studies that
have shown that chromium from other sources including Cr–transferrin are ultimately lost in the urine
bound to chromodulin [18,23]. Fecal chromium loss
presumably comes from chromium in the bile. Previous
work has shown that only minor amounts of chromium
Fig. 3. Elution profile from Shodex OH-PAK HPLC column of rat urine sample from day 2 (0 – 12 h after injection). Solid line and circles
represent the absorbance at 210 nm; the dotted line and open circles represent the cpm of 51Cr (scaled by dividing by 26).
A.A. Shute et al. / Polyhedron 20 (2001) 2241–2252
Fig. 4.
51
Cr content of tissues after 2 weeks of treatment with
51
2245
Cr-labeled compound 1. Error bars indicate standard deviation.
are lost in the bile (for example, see Refs. [24,25]). In
bile, chromic ions occur as part of a low-molecular
weight organic complex [26]. The molecular weight of
this species unfortunately was not determined, although
the authors postulated that this complex might be
involved in passage of chromium from the liver to the
bile. The low concentration of chromium in the feces
prevented further studies on the form of the chromium
in the current work. The form of chromium in bile of
rats treated with various forms of chromium needs to
be determined.
After 2-week period of treatment with the trimer, the
distribution of radiolabeled chromium in the tissues
was examined (Fig. 4). Of the tissues examined, the
greatest quantity of chromium was located in the liver,
followed by the spleen and kidneys. Only small
amounts of the label were found in the heart, testes,
pancreas and epididymal fat. The distribution of
chromium is similar to that when other forms of
chromium have been given to rats [18,26–28], although
the level in the spleen is generally lower than in the
current work. As the amount of chromium in liver was
greater than in any other tissue, liver tissue was examined further. Examination of the subcellular distribution of chromium in hepatocytes (Fig. 5), however,
reveals a quite distinct distribution of chromium than
found in studies using other forms of chromium [18,26].
Chromium from the trimer is located primarily in the
nuclear and mitochondrial fraction of the hepatocytes.
Curiously this correlates with the DNA content of the
subcellular components. Approximately seven times
more Cr is located in the nuclear fraction than in the
soluble fraction. In contrast, chromium when injected
in similar quantities as CrCl3 or Cr –transferrin is distributed generally from roughly evenly between the
nuclear and soluble fractions of the cells to three times
A.A. Shute et al. / Polyhedron 20 (2001) 2241–2252
2246
more chromium being present in the nuclear fraction,
with little Cr found in other fractions [18,26]. (Cr from
CrCl3 is bound tightly by transferrin in the blood
plasma [18,27,29–31], such that both sources are probably equivalent.) If the trinuclear complex were decomposing in the blood, then the released chromic ions
would be expected to be bound readily by transferrin,
resulting in a subcellular distribution similar to that in
earlier studies. This strongly suggests that the trimer
enters the cells intact and remains intact for some
period of time resulting in the different subcellular
distribution; this may also account for the large degree
of incorporation of the label into the spleen. In order to
elucidate which species in the liver might be binding
chromium or whether the trinuclear cation was maintaining its integrity, the nuclear fraction containing the
most radiolabel was applied to a G-15 size exclusion
Fig. 5. Subcellular distribution of
deviation.
51
column. The elution profile (Fig. 6) indicated that
nearly all the chromium eluted with the solvent front
indicating the chromium was bound to species with
molecular weights greater than 1500. No evidence was
present for the existence of the intact trinuclear cation.
Thus, the cation appears to fall apart (at least in the
liver) within 24 h of injection.
Treatment of rats with 14C-labeled trimer results in a
significantly smaller portion of the label appearing in
the feces and urine (Fig. 7). As seen with chromium in
the urine, 14C appears in the urine primarily in the first
12 h after injection, although the pattern breaks down
after the first week of injections. Fecal 14C loss follows
no rigorous pattern. Daily loss of the label through
both pathways is relatively stable and accounts for only
3% of the label. This is an order of magnitude less
retention than for the chromium. When the compo-
Cr in hepatocytes after 2 weeks of treatment with
51
Cr-labeled compound 1. Error bars indicate standard
A.A. Shute et al. / Polyhedron 20 (2001) 2241–2252
Fig. 6. Elution profile of nuclear fraction of hepatocytes from
nents of the urine are separated by G-15 chromatography (Fig. 8), most of the labeled carbon elutes with the
solvent front indicating that is part of molecules with
molecular weights greater than 1500. One small band
elutes at a lower molecular weight, although this band
also does not correspond to cation 1. Again there is no
evidence for the cation to reach the urine intact.
Among tissues, the liver retains a comparatively large
relative portion of the label. Only the spleen also seems
to contain an appreciable quantity of carbon from the
propionate ligands (Fig. 9). The distribution of the
label in hepatocytes is distinctly different from that of
chromium (Fig. 10). The different subcellular distributions between the inorganic and organic portion of the
cation suggest that the trimer dissociates in the tissue,
consistent with the other results of this study.
The distribution of 14C is less informative than that
of chromium unfortunately. Propionate occurs natu-
2247
51
Cr-labeled compound 1-treated rat from G-15 column.
rally in mammals and is normally transported to the
liver by the blood where it is metabolized. Therefore,
propionate generated by the degradation of the complex should readily be metabolized, unless the concentrations of propionate generated cannot be handled by
the liver (which is extremely unlikely for the levels in
this study (vide infra)). Hence, the distribution of products containing 14C from propionate probably only
reflects the normal distribution of propionate
metabolism products in the tissues and hepatocyte compartments. This, however, is not the case with
chromium. The rats are receiving a pharmacological
dose of chromium. Humans appear to be capable of
maintaining their chromium balance on approximately
30 mg of Cr3 + [1,32]; hence, a 0.5 kg rat receiving 60
mg of Cr3 + is easily receiving its essential quantity of
chromium when compared to an approximately 75 kg
human receiving 30 mg.
2248
A.A. Shute et al. / Polyhedron 20 (2001) 2241–2252
4. Discussion
The relatively rapid degradation of the synthetic
compound in vivo raises the question of whether previous in vivo studies that examined the effects of the
complex on blood variables in rats were observing
effects from chromic ions or from propionate. The
effects of chromic ions on rats have been reviewed
recently [1]. Supplementation of the diet of rats with
chromic complexes (other than potentially compound
1), which serve only as sources of Cr3 + , has no effects;
consequently, a role for simple inorganic chromium can
be ruled out unless the rats were chromium deficient.
Generation of chromium deficiency in rats is extremely
difficult requiring strict environmental control (such as
preventing access to stainless steel) and a high sugar
diet which stimulates urinary chromium loss [1,33,34].
Chromium deficiency is not a concern for the rats in
this study as the animals consumed a commercial rat
chow and had access to stainless steel (e.g. cage and
water bottle components).
Propionate has been proposed to be able to lower
plasma cholesterol concentrations; however, these results are quite controversial [35– 46]. Propionate occurs
naturally in mammals although it is actually generated
by bacteria. Cecal and colonic fermentation of dietary
fiber results in the production of short-chain fatty acids
including acetate, propionate, and butyrate. These
agents may be responsible in part for the action of
dietary fiber to reduce plasma cholesterol levels.
In rat studies, Chen et al. found that a cholesterolfree diet containing 0.5% propionate fed to rats for 14
days had no effect on fasting serum cholesterol, HDL,
triglycerides, or glucose concentrations and also no
Fig. 7. Urinary and fecal 14C loss during the 2-week period of treatment with
injection of an aliquot of a solution of labeled compound 1.
14
C-labeled compound 1. Time zero represents the time of first
A.A. Shute et al. / Polyhedron 20 (2001) 2241–2252
2249
Fig. 8. Elution profile from G-15 column of rat urine sample from day 2 (0 – 12 h after injection). Solid line and circles represent the absorbance
at 210 nm; the dotted line and open circles represent the cpm of 14C.
effect on liver cholesterol or triglycerides [35]. However,
when rats were fed a cholesterol-containing diet, lower
levels of plasma cholesterol (13%) and liver cholesterol
and triglycerides were found. Illman et al., using a diet
comprised of 5% propionate for 10 days, found that the
rats had lower plasma cholesterol levels and unchanged
triglycerides levels [43]. In a similar study, Levrat et al.
fed a diet containing 2% propionate to rats for 21 days
[42]. Propionate from the diet was found to be metabolized entirely by the liver; no depression of plasma
cholesterol was observed. Kishimoto et al. gave 0.01–
10 mg of sodium propionate to 6-week-old rats by
intravenous injection [39]. After injection (3 h) of 1 mg
of sodium propionate, serum total cholesterol was
found to be lower ( 15%) and to remain below normal for 24 h. The reduction of plasma cholesterol levels
was dependent on propionate dosage over the range
0.01–10 mg. Rats were also given either 0.12 or 1.2 mg
of propionate daily for 14 days intraperitoneally;
cholesterol levels were maximally lower for the higher
dosage at day 7 (29%). Hara et al. fed rats a fiber-free
diet with propionate (approximately 10%) for 14 days.
No effect on plasma HDL or total cholesterol was
observed [40].
Obese, hyperinsulinaemic rats (fa/fa) fed a diet containing 1 g propionate per day or infused rectally with
150 mg propionate per day for 19 days have been
examined by Berggren et al. [38]. The diets fed to both
groups of rats were high in cholesterol and saturated
fat. Liver cholesterol concentrations were unchanged
although the livers of the rats receiving propionate were
reduced in mass compared to those of controls. Fasting
plasma glucose and urinary glucose excretion were reduced in propionate fed rats. The authors suggested
that the manner of propionate intake may affect the
effects of propionate.
The effects of propionate on isolated rat hepatocytes
have also been examined, but these studies also suffer
to some degree from a lack of consistent results.
Nishina and Freedland found that 1 mM propionate
had no effect on sterol synthesis but inhibited fatty acid
synthesis [36]. Wright et al. looked at the effects of a
A.A. Shute et al. / Polyhedron 20 (2001) 2241–2252
2250
range of propionate concentrations on cholesterol and
fatty acid synthesis [44]. Statistically significant inhibition of cholesterol synthesis from labeled acetate was
observed at [propionate]]1.0 mM while inhibition was
significant at 2.5 mM greater concentrations if labeled
H2O or mevalonate was used. Fatty acid synthesis was
not inhibited by propionate when labeled water or
mevalonate was used to follow the synthesis; however,
] 2.5 mM concentrations of propionate inhibited synthesis from labeled acetate. In contrast, Lin et al. found
50% inhibition of cholesterol and triglycerides synthesis
from labeled acetate at 0.1 mM propionate, although
human hepatocytes required 10–20 mM concentrations
for the same level of inhibition to be achieved [46]
Demigne et al. found that 0.6 mM and higher propionate concentrations inhibited fatty acid and cholesterol synthesis [45]. To put the concentrations in
perspective, Illman et al. [43] found that their rat diet
(5% propionate) led to an increase in the hepatic portal
Fig. 9.
14
vein concentration of propionate from 0.22 mM in
control rats to 0.38 mM. Thus, the isolated hepatocyte
studies other than that of Lin et al. required propionate
concentrations which should be difficult to generate in
rats; the results of Lin et al. are difficult to reconcile as
they observe substantial inhibition at propionate concentrations less than that of the hepatic portal vein
concentration of rats fed a standard commercial diet.
Illman et al. have also examined the effects of propionate on perfused rat liver. Cholesterol synthesis was
not affected by 1 mM propionate and inhibition required 10 mM propionate [43].
Propionate has also been proposed to have effects on
glucose production [47–50]. In a recent study using
healthy rats, Boillet et al. fed rats a diet containing
0.78% propionate for 3 weeks [47]. The propionate
resulted in significantly lower fasting plasma glucose
but no difference in plasma insulin and basal hepatic
glucose. No difference in hepatic glucose production or
C content of tissues after 2 weeks of treatment with
14
C-labeled compound 1.
A.A. Shute et al. / Polyhedron 20 (2001) 2241–2252
Fig. 10. Subcellular distribution of
14
C in hepatocytes after 2 weeks of treatment with
utilization was observed at two different insulin concentrations. Thus, both healthy and obese rats fed propionate have been observed to possess reduced plasma
glucose.
Given the conflicting results in propionate supplementation studies, the expectation of any effects in rats
from propionate resulting from the decomposition of 1
are at best difficult to ascertain. However, the amount
of propionate provided daily in the previous study of
the effects of 1 in rats (56 mg kg − 1 body mass) [12] are
far below the amounts required to observe an effect in
the above studies. Similarly the lowering of plasma
cholesterol and triglycerides levels without affecting
plasma glucose levels by 1 is inconsistent with the
studies of the effects of propionate in rats, even if
sufficient levels of propionate were generated. Conse-
2251
14
C-labeled compound 1.
quently, the effects of 1 previously observed in rats are
best attributed to the cation and not its degradation
products. The effects seen in rats injected daily with
chromium presumably then must arise from only a
small period of time in which the complex remains
intact. Further studies are thus warranted for examining the effects of alternative methods of delivering the
cation such as adding the trimer to drinking water
resulting in the rats receiving smaller doses over more
extended period of time.
5. Conclusions
The biomimetic cation has a limited lifetime in vivo,
decomposing in less than 24 h. Chromium from the
A.A. Shute et al. / Polyhedron 20 (2001) 2241–2252
2252
cation is lost primarily in the urine; the form of
chromium in the urine is chromodulin. The tissue distribution and hepatocyte subcellular distribution of
chromium from the cation is different from that when
rats are given other forms of chromium including CrCl3
and Cr –transferrin, suggesting that some percentage of
the trimer enters cells intact. Propionate released from
the trimer is insufficient to account for the previously
observed cholesterol- and triglycerides-lowering effects
of the complex when given to rats intravenously.
Acknowledgements
Funding was provided by the American Diabetes
Association to J.B.V.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
J. Vincent, Polyhedron 20 (2001) 1.
J. Vincent, J. Nutr. 130 (2000) 715.
J. Vincent, Nutr. Rev. 58 (2000) 67.
J. Vincent, Acc. Chem. Res. 33 (2000) 503.
R.D. Cannon, R.P. White, Prog. Inorg. Chem. 36 (1988) 195.
M. Nagi, A. Harton, S. Donald, Y.-S. Lee, M. Sabat, C.J.
O’Connor, J.B. Vincent, Inorg. Chem. 34 (1995) 3813.
A. Harton, K. Terrell, J.C. Huffman, C. MacDonald, A. Beatty,
S. Li, C. O’Connor, J.B. Vincent, Inorg. Chem. 36 (1997) 4875.
A. Bino, R. Chayat, E. Pedersen, A. Schneider, Inorg. Chem. 30
(1991) 856.
S. Donald, K. Terrell, K. Robinson, J.B. Vincent, Polyhedron 14
(1995) 971.
T. Ellis, M. Glass, A. Harton, K. Folting, J.C. Huffman, J.B.
Vincent, Inorg. Chem. 33 (1994) 5522.
C.M. Davis, A.C. Royer, J.B. Vincent, Inorg. Chem. 36 (1997)
5316.
Y. Sun, K. Mallya, J. Ramirez, J.B. Vincent, J. Biol. Inorg.
Chem. 4 (1999) 838.
J.K. Speetjens, A. Parand, M.W. Crowder, J.B. Vincent, Polyhedron 18 (1999) 2617.
J.K. Speetjens, R.A. Collins, J.B. Vincent, S.A. Woski, Chem.
Res. Toxicol. 12 (1999) 483.
A. Earnshaw, B.N. Figgis, J. Lewis, J. Chem. Soc. A (1966)
1656.
C. de Duve, B.C. Pressman, R. Gianetto, R. Wattiaux, F.
Appelmans, Biochem. J. 60 (1955) 604.
G.L. Rowin, Methods in Molecular Biology, Marcel Dekker,
New York, 1974 (pp. 90 –109).
B.J. Clodfelder, J. Emamaullee, D.D. Hepburn, N.E. Chakov,
H.S. Nettles, J.B. Vincent, J. Biol. Inorg. Chem. 6 (2001), in
press.
R.A. Anderson, N.A. Bryden, M.M. Polansky, J. Am. Coll.
Nutr. 16 (1997) 273.
.
[20] W. Mertz, E.E. Roginski, R.C. Reba, Am. J. Physiol. 209 (1965)
489.
[21] C. Onkelinx, Am. J. Physiol. 232 (1977) E478.
[22] T.H. Lim, T. Sargent III, N. Kusubov, Am. J. Physiol. 244
(1983) R445.
[23] G.Y. Wu, O. Wada, Jpn. J. Ind. Health 23 (1981) 505.
[24] L.L. Hopkins Jr., Am. J. Physiol. 209 (1965) 731.
[25] M.L. Davis-Whitenack, B.O. Adeleye, L.L. Rolf, B.J. Stoecker,
Nutr. Res. 16 (1996) 1009.
[26] L. Manzo, A. Di Nucci, J. Edel, C. Gregotti, E. Sabbioni, Res.
Commun., Chem. Pathol. Pharmacol. 42 (1983) 113.
[27] M.L. Davis-Whitenack, B. Adeleye, B.J. Stroecker, Biol. Trace
Elem. Res. 68 (1999) 175.
[28] S.M. Kamath, B.J. Stoecker, M.L. Davis-Whitenack, M.M.
Smith, B.O. Adeleye, S. Sangiah, J. Nutr. 127 (1997) 478.
[29] L.L. Hopkins Jr., K. Schwarz, Biochim. Biophys. Acta 90 (1964)
484.
[30] F. Borguet, R. Cornelis, J. Delanghe, M.C. Lambert, N.
Lamiere, Clin. Chem. Acta 238 (1995) 71.
[31] F. Borguet, R. Cornelis, N. Lamiere, Biol. Trace Elem. Res.
26 – 27 (1990) 449.
[32] E.G. Offenbacher, H. Spencer, H.J. Dowling, F.X. Pi-Sunyer,
Am. J. Clin. Nutr. 44 (1986) 77.
[33] J.S. Striffler, M.M. Polansky, R.A. Anderson, Metabolism 48
(1999) 1063.
[34] J.S. Striffler, J.S. Law, M.M. Polansky, S.J. Bhathena, R.A.
Anderson, Metabolism 44 (1995) 1314.
[35] W.-J.L. Chen, J.W. Anderson, D. Jennings, Proc. Soc. Exp. Biol.
Med. 175 (1984) 215.
[36] P.M. Nishina, R.A. Freeland, J. Nutr. 120 (1990) 668.
[37] K.E. Beaulieu, M.I. McBurney, J. Nutr. 122 (1992) 241.
[38] A.M. Berggren, E.M. Nyman, I. Lundquist, I.M. Bjorck, Br. J.
Nutr. 76 (1996) 287.
[39] Y. Kishimoto, S. Wakabayashi, H. Takeda, J. Nutr. Sci. Vitaminol. 41 (1995) 73.
[40] H. Hara, S. Haga, T. Kasai, S. Kiriyama, J. Nutr. 128 (1998)
688.
[41] H. Hara, S. Haga, Y. Aoyama, S. Kiriyama, J. Nutr. 129 (1999)
942.
[42] M.A. Levrat, M.L. Favier, C. Moundras, C. Remesy, C.
Demigne, C. Morand, J. Nutr. 124 (1994) 531.
[43] R.J. Illman, D.L. Topping, G.H. McIntosh, R.P. Trimble, G.B.
Storer, M.N. Taylor, B.-Q. Cheng, Ann. Nutr. Metab. 32 (1988)
97.
[44] R.S. Wright, J.W. Anderson, S.R. Bridges, Proc. Soc. Exp. Biol.
Med. 195 (1990) 26.
[45] C. Demigne, C. Morand, M.A. Levrat, C. Besson, C. Moundras,
C. Remesy, Br. J. Nutr. 74 (1995) 209.
[46] Y. Lin, R.J. Vonk, M.J. Slooff, F. Kuipers, M.J. Smit, Br. J.
Nutr. 74 (1995) 197.
[47] J. Boillot, C. Alamowitch, A.M. Berger, J. Lou, F. Bruzzo, F.R.
Bornet, G. Slama, Br. J. Nutr. 73 (1995) 241.
[48] T.M. Chan, R.A. Freedland, Biochem. J. 127 (1972) 539.
[49] J.B. Blair, D.E. Cook, H.A. Lardy, J. Biol. Chem. 248 (1973)
3608.
[50] J.W. Anderson, S.R. Bridges, Proc. Soc. Exp. Biol. Med. 177
(1984) 372.
Biochemistry 1996, 35, 12963-12969
12963
A Biologically Active Form of Chromium May Activate a Membrane
Phosphotyrosine Phosphatase (PTP)†
C. Michele Davis, K. Heather Sumrall, and John B. Vincent*
Department of Chemistry, UniVersity of Alabama, Tuscaloosa, Alabama 35487-0336
ReceiVed February 12, 1996; ReVised Manuscript ReceiVed May 9, 1996X
ABSTRACT: Chromium is essential for proper carbohydrate and lipid metabolism in mammals, although
the mechanism of this action has previously proved elusive. Low-molecular-weight chromium-binding
protein (LMWCr), a biologically active form of chromium in mammals, potentiates the effect of insulin
on the conversion of glucose into lipid and into carbon dioxide in isolated adipocytes. Kinetics studies
indicate that LMWCr isolated from bovine liver activates phosphotyrosine phosphatase (PTP) activity in
adipocyte membranes while having no intrinsic phosphatase activity. This activation is directly proportional
to the amount of added LMWCr. The pattern of inhibition of this activity in the presence of a number
of known phosphatase inhibitors suggests the involvement of a membrane phosphotyrosine phosphatase
similar to PTP1A′ or PTP1B. We propose that chromium plays a biological role in the activation of a
membrane phosphotyrosine phosphatase.
The first-row transition elements from vanadium to zinc
are each essential for some form of life (Kaim & Schwederski, 1994; Lippard & Berg, 1994; Frausto da Silva &
Williams, 1991). For each metal except chromium, at least
one (and usually many) metallobiomolecule containing an
ion of that element has been well-characterized in terms of
its function and mode of action. Since its inception nearly
forty years ago (Schwarz & Mertz, 1959), chromium
biochemistry has proved to be an enigma (Vincent, 1994a).
In the mid 1950’s, rats fed a Torula yeast diet [which proved
to be Cr-deficient (Anderson, et al., 1978)] developed glucose
intolerance which could be reversed only by addition of Cr
to the diet (Schwarz & Mertz, 1957, 1959; Mertz & Schwarz,
1959). A Cr-rich material extracted from Brewer’s yeast,
named glucose tolerance factor (GTF),1 was found to be
especially effective in reversing the glucose intolerance
(Schwarz & Mertz, 1959). Unfortunately, the isolation of
this material involved procedures such as an 18-h reflux in
5 N HCl which would have destroyed any protein or similar
components (Toepfer et al., 1977); the isolation and characterization of this material have not proved to be reproducible in some laboratories (Haylock et al., 1983; GonzalezVergara et al., 1982; Shepherd et al., 1992). Additionally,
recent analysis of kinetics studies on the biological activity
of GTF indicates that it serves only as a readily absorbable
form of Cr and that this yeast material has no intrinsic
function in mammals (in fact, it may inhibit insulin in nonCr-deficient cells) (Vincent, 1994b). [Biological activity is
defined as the ability to potentiate the effects of insulin on
† This work was funded by the American Heart Association, No.
94011190.
* Author to whom correspondence should be addressed.
X Abstract published in AdVance ACS Abstracts, September 15, 1996.
1 Abbreviations: GTF, glucose tolerance factor; LMWCr, lowmolecular-weight chromium-binding substance; PTP, phosphotyrosine
phosphatase; EDTA, ethylenediaminetetraacetic acid; BSA, bovine
serum albumin; Hepes, N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic
acid; p-NPP, para-nitrophenyl phosphate; PP1, phosphoprotein phosphatase 1; PP2A, phosphoprotein phosphatase 2A; LAR, leukocyte
antigen related phosphotyrosine protein phosphatase.
S0006-2960(96)00328-5 CCC: $12.00
conversion of glucose into carbon dioxide or into lipid by
epididymal fat tissue or isolated adipocytes (Mertz &
Roginski, 1963; Anderson et al., 1978).] Recent interest has
also been directed to chromium picolinates (Evans &
Bowman, 1992; Hasten et al., 1992; Evans & Pouchnik,
1993); however, these materials also appear to serve only
as readily absorbable sources of Cr (McCarty, 1993). Also
because of the picolinate ligands, they cause chromosome
damage (Stearns et al., 1995).
Nevertheless, medical studies clearly indicate that Cr is
required for normal carbohydrate and lipid metabolism
(Anderson, 1985, 1986). Cr deficiency in humans and other
mammals results in symptoms comparable to those associated
with adult-onset diabetes and cardiovascular disease: decreased glucose tolerance (Mertz et al., 1965; Woolliscroft
& Barbosa, 1977; Hopkins et al., 1968; Schroeder, 1965),
increased concentration of circulating insulin (Schroeder &
Balassa, 1965), decreased insulin receptor number (Anderson
et al., 1987), elevated cholesterol and triglyceride levels
(Riales & Albrink, 1981), and reduced high-density-lipoprotein (HDL) cholesterol levels (Riales & Albrink, 1981;
Anderson, 1987). Improvement in glucose tolerance after
supplementation of the diet with chromium has been
documented many times (Schwartz & Mertz, 1959; Anderson
et al., 1983, 1987, 1991a; Glinsman & Mertz, 1966; Levine
et al., 1968; Hopkins, 1965; Gurson & Saner, 1971; Martinez
et al., 1985; Mahdi & Naismith, 1991), as well as improvements in other symptoms (Anderson, 1985, 1986; Riales &
Albrink, 1981; Madhi & Naismith, 1991). Cr is additionally
released from storage pools and resultantly excreted in the
urine in response to certain stresses including high sugar
intake (Martinez et al., 1985; Anderson et al., 1982). In
particular, carbohydrates that alter circulating insulin levels
also effect urinary Cr losses, as glucose tolerance decreases,
the mobilization of Cr and resulting Cr loss have also been
shown to decrease (Anderson et al., 1990). Yet as much as
90% of the American population and half of the population
of developed nations daily intakes are less than the recommended safe and adequate quantities of Cr (Anderson &
Kozlovsky, 1985; Anderson, 1994).
© 1996 American Chemical Society
12964 Biochemistry, Vol. 35, No. 39, 1996
At present, the only viable candidate for the biologically
active form of Cr is low-molecular-weight Cr-binding
substance (LMWCr) (Yamamoto et al., 1987; Davis and
Vincent, submitted for publication). LMWCr is a mammalian polypeptide of circa 1500 Da which binds four
chromic ions in a multinuclear assembly; this laboratory has
recently developed a procedure for isolating milligram
quantities of this polypeptide (Davis and Vincent, submitted
for publication). Kinetics studies on LMWCr indicate that
it has an intrinsic postreceptor role (i.e., after the binding of
insulin to the external surface of insulin receptor) in the
insulin-dependent activation of conversion of glucose into
carbon dioxide and lipid (Vincent, 1994b; Yamamoto et al.,
1988) and that this activation is directly dependent on the
Cr content of LMWCr (Yamamoto et al., 1989). In addition,
LMWCr is a major form of chromium in urine and bile
(Manzo et al., 1983; Wada et al., 1983) and likely represents
the form of Cr released in response to stress, such as
carbohydrate stress. Because Cr(III) complexes are substitutionally inert, it is unlikely that LMWCr has a direct
enzymatic function (Vincent, 1994a). The fact that insulin
functions by activating a series of cascade systems involving
multiple phosphorylation/dephosphorylation events (Saltiel,
1994) suggests that LMWCr may be involved in the
regulation of these phosphorylation steps. Herein are
reported kinetics studies which indicate that LMWCr activates a membrane PTP (phosphotyrosine phosphatase).
MATERIALS AND METHODS
LMWCr and ApoLMWCr. LMWCr was available from
previous work (Davis and Vincent, submitted for publication). ApoLMWCr was prepared by heating a solution of
LMWCr in 3.5 mM EDTA and a catalytic amount of sodium
cyanoborohydride adjusted to pH 3 at 60 °C for 12 h. The
apoprotein was separated from the EDTA and reductant by
passing over a column of G-15 Sephadex. ApoLMWCr
contains approximately 0.3 Cr per polypeptide. Chromium
was assayed using the diphenylcarbizide procedure (Marczenko, 1986) utilizing the method of standard addition to
minimize any potential matrix effects. Oligopeptide concentrations were assayed by the fluorescamine procedure of
Undenfriend and co-workers (1972). The concentration of
holoLMWCr in certain experiments is given in terms of
chromium concentration as LMWCr; preparations of LMWCr in this work contained 3.6 chromium per oligopeptide.
Preparation of Isolated Adipocytes. Fat cells from male
Sprague-Dawley rats were isolated by modifications of the
procedure of Rodbell (1964). Three rats (not kept on a Crdeficient diet) were sacrificed by decapitation, and their
epididymal fat pads were removed. Subsequent operations
followed Anderson et al. (1978) with minor modification,
and 2% bovine serum albumin (BSA) media was changed
to 1% bovine serum albumin and solutions were not gassed
with O2.
Adipocyte Membrane and Cytosol Preparation. Rat
adipocytes were isolated as above except they were washed
with 1% bovine serum albumin, 50 mM Hepes, pH 7.4 buffer
containing 10 µg of leupeptin/mL and 5 µg of aprotinin/L.
Cells were homogenized with a manual Teflon homogenizer
and frozen and thawed five times. The lipid layer was
removed, and the cell homogenate was centrifuged for 1 h
at 40 000g. The supernatant was removed and used as a
Davis et al.
source of cytosol phosphatase activity. The pellet was
suspended in Hepes buffer and used as a source of membrane
phosphatase activity. Protein concentrations were determined
using the BCA method (Pierce Chemical Co.) with BSA as
standard.
Phosphotyrosine Phosphatase and Phosphoserine/Phosphothreonine Phosphatase ActiVity. PTP activity was estimated using p-nitrophenyl phosphate (p-NPP) by the method
of Li et al. (1995). The assay used 5 mM substrate in 0.05
M Tris, pH 7.5, unless otherwise noted. Hydrolyses were
allowed to proceed 1 h at 37 °C. The extent of hydrolysis
was determined at 404 nm ( ) 1.78 × 104 M-1 cm-1). PTP
activity was also determined using a phosphotyrosine assay
kit (Boehringer Mannheim), which uses a fragment of human
gastrin (amino acids 1-17) phosphorylated on tyrosine-12
and a synthetic fragment of hirudin (53-65 C-terminal
fragment) phosphorylated on tyrosine-63 as substrates.
These assays were also performed in 0.05 M Tris, pH 7.5,
and hydrolysis reactions were allowed to proceed 1 h at 37
°C unless otherwise noted. When the kit was used,
membranes (which contain phosphotyrosine themselves)
were removed after the reaction using Microcon 30 microconcentrators (Amicon); 0.8 mM piceatannol was used as a
kinase inhibitor. Bovine intestinal mucosa alkaline phosphatase (Sigma), Yersinia PTP fragment (Boehringer Mannheim), catalytic subunit of the γ form of human protein
phosphatase 1 (PP1) (Boehringer Mannheim), human transmembrane leukocyte antigen related tyrosine phosphatase
(LAR) (Calbiochem), [NH4][O2CMe], LMWCr, and apoLMWCr were assayed for PTP activity using 5 mM p-NPP in
0.05 M Tris, pH 7.5, at 37 °C for 1 h except the reaction
with alkaline phosphatase which was allowed to proceed 15
s at room temperature, the reaction with LAR (10 units)
which was allowed to proceed for 45 min at 37 °C, and the
reaction with PP1 (0.05 milliunits) which was allowed to
proceed for 2 h at 37 °C. For experiments utilizing
monoclonal antibodies whose epitope is within the catalytic
domain of human PTP 1B (and which reacts with rat PTP
1B) (Calbiochem), PTP activity of rat adipocyte membranes
was assayed as above except membranes (corresponding to
5 µg of protein) were incubated for 2 h at 4 °C with 1.09 µg
of antibody before addition of p-NPP; the hydrolysis
reactions were allowed to proceed 75 min at 37 °C.
Phosphoserine/phosphothreonine phosphatase activity was
measured using 0.83 mM phosphvitin as substrate in 0.05
M Tris, pH 7.5, at 37 °C by following phosphate release
after 1 h by the method of Zhang and VanEtten (1991).
Curve-fitting was performed using SigmaPlot (Jandel Scientific).
Inhibitor and Metal Reconstitution Studies. Metals were
used as the following salts: NiSO4‚6H2O (J. T. Baker), Zn(NO3)2‚6H2O (Mallinckrodt), Cu(SO4)‚5H2O (Fisher), CoCl2
(J. T. Baker), [NH4]VO3 (Fisher), FeCl3‚6H2O (J. T. Baker),
Cr(NO3)3‚6H2O (Fisher), CrCl3‚6H2O (Fisher), MnSO4‚H2O
(Mallinckrodt), and Na2MoO4‚2H2O (Fisher). Metals and
inhibitors were added to the reaction mixture and incubated
for 10 min at 37 °C prior to reaction initiation by addition
of substrate. Microcystin LR was obtained from Calbiochem.
Miscellaneous. All visible spectroscopic measurements
were made with a Hewlett-Packard 8451A diode array
spectrophotometer. EPR were collected on a Varian E-12
spectrophotometer equipped with an Oxford ESR 900
Bioactive Chromium May Activate Membrane PTP
Biochemistry, Vol. 35, No. 39, 1996 12965
cryostat. For apoLMWCr, solutions in 0.050 M NH4OAc
were rapidly frozen in liquid N2. Integrations were performed as previously described using hexaaquochromium(III) as a standard (Davis and Vincent, submitted for
publication); the broad EPR signal of pseudooctahedral
mononuclear Cr3+ requires the use of a standard with a
similarly broad EPR signal for accurate integrations. Aqueous solutions of [Cr(H2O)6]3+ were rapidly frozen in liquid
N2; the integrity of the sample was checked by UV/visible
spectroscopy after the sample was thawed at the conclusion
of the EPR experiment. Fluorescence measurements were
obtained with a Perkin Elmer 204 fluorescence spectrophotometer. All kinetics experiments were performed in triplicate and reproduced at least once. Errors are presented
throughout including all tables and graphs as the standard
deviations (1σ) of the triplicate analyses. Similarly, all Cr
and LMWCr concentration determinations were made in
triplicate. Doubly deionized water was used in all operations.
RESULTS AND DISCUSSION
Phosphatase ActiVation
LMWCr functions in a manner in insulin-sensitive cells
such that the action of insulin is potentiated. In insulin
dose-response studies, the degree of incorporation of 14C
or 3H from labeled glucose into carbon dioxide or total lipids
is enhanced while the concentration of insulin for halfmaximal activity is unaffected (Yamamoto et al., 1988; Davis
and Vincent, submitted for publication). This indicates that
LMWCr has an intrinsic role “inside the cell” which is
stimulated by the action of insulin while not affecting the
interaction of insulin with its receptor. Given the unlikely
possibility that LMWCr acts as a catalyst and that insulin
action is propagated by series of phosphorylation/dephosphorylation events, a role for LMWCr in the activation or
inhibition of kinases or phosphatases associated with insulin
action seemed plausible. Therefore, the effects of LMWCr
on rat adipocyte phosphatases were examined. Rat adipocytes were homogenized and separated into three portions:
lipids, soluble, and membrane/particulate. LMWCr had no
effect on phosphatase activity toward p-NPP in either the
lipid or soluble portions (not shown); however, a distinct
effect was observed in the phosphatase activity associated
with the membrane fragments (Figure 1). The ammonium
acetate buffer in which LMWCr is stored displayed no ability
to activate the phosphatase activity, and LMWCr in the
absence of the membranes also demonstrated no catalytic
activity. Over the concentration range 0.1 nM to 10 µM (in
terms of Cr), the addition of LMWCr activates the arylphosphate phosphatase activity of the membranes in a concentration dependent fashion. Fitting the hyperbolic phosphatase
activation behavior to the Henri-Michaelis-Menten equation yields a LMWCr dissociation constant of 4.4 nM.
Similar results over the same Cr concentration range have
also been obtained utilizing adipocytes isolated from adipose
tissue trimmed from bovine liver. To determine whether this
arylphosphate phosphatase activity corresponded to phosphotyrosine phosphatase activity, the ability of LMWCr to
activate the hydrolysis of phosphotyrosine-containing polypeptides was examined. Using human gastrin (amino acids
1-17) phosphorylated on tyrosine-12 and a synthetic fragment of hirudin (53-65 C-terminal fragment) phosphorylated
on tyrosine-63 as substrates and the adipocyte membrane
FIGURE 1: Activation of rat adipocytic membrane phosphotyrosine
phosphatase activity using 5 mM p-NPP as substrate by LMWCr
(solid circles) and rat adipocyte membrane phosphoserine phosphatase activity using 0.83 mM phosvitin as substrate (shaded
triangles). 125 µL of a rat membrane suspension corresponding
to 136.8 µg of protein/mL was utilized. Activation of phosphotyrosine phosphatase activity of Yersinia phosphotyrosine phosphatase fragment (19.5 milliunits) using 5 mM p-NPP as substrate
by LMWCr (open squares). The line is the best fit hyperbolic curve
giving a LMWCr dissociation constant of 4.4 nM. Inset: Activation
of rat adipocyte membrane phosphotyrosine activity using 0.75 µM
human gastrin fragment (open squares) or hirudin fragment (solid
circles) as substrate. 125 µL of a rat membrane suspension
corresponding to 129 µg of protein/mL was utilized. The line is
the same best fit curve as above. All Cr concentrations are
presented in units of nM.
fragments, LMWCr activated hydrolysis in a concentration
dependent manner (Figure 1), nearly identical to that of
p-NPP by the membrane fragments. For example, 72%
activation of hirudin dephosphorylation was observed at an
LMWCr concentration corresponding to 50 µM chromium;
63% activation of gastrin dephosphorylation was observed
at the same LMWCr concentration. These phosphotyrosinecontaining peptide assays were performed in the presence
of the kinase inhibitor piceatannol, suggesting that phosphatase activation and not kinase inactivation is responsible
for this activity. When the phosphoserine-containing protein
phosvitin was used as a substrate, no activation of dephosphorylation this protein was observed over the same range
of LMWCr concentrations (Figure 1). The activation
potential of LMWCR is therefore likely to be directed toward
phosphotyrosine phosphatase activity.
Inhibition Studies
Rat adipocyte membranes possess a number of phosphatases including phosphoserine/phosphothreonine phosphatases and phosphotyrosine phosphatases (Ding et al.,
1994; Liao et al., 1991; Begum, 1995). PP1 comprises 80%
of phosphoserine/phosphothreonine phosphatase activity in
12966 Biochemistry, Vol. 35, No. 39, 1996
Davis et al.
Table 1: Inhibition of Membrane PTP Activity and
LMWCr-Activated Membrane PTP Activity Using 5 mM p-NPP as
Substratea
% activity
inhibitor
none
EDTA (5 mM)
microcystin LR (1 µM)
spermine (2 mM)
spermidine (2 mM)
zinc nitrate (100 µM)
apoLMWCr (12.5 µM)
apoLMWCr (12.5 µM) + CrCl3 (50 µM)
CrCl3 (50 µM)
(- LMWCr) (+ LMWCr)
100(4)
148(10)
110(3)
115(5)
109(2)
77(1)
133(2)
197(7)e
102(10)
191(8)c
149(5)
119(4)c
126(5)d
122(2)c
86(2)c
-
a Standard deviations are given in parentheses. b 50 µM Cr. c Differs
significantly (P < 0.002) from the value of the assay without LMWCr
(Student’s unpaired t-test). d Differs significantly (P < 0.01) from the
value of the assay without LMWCr. e Differs significantly (P < 0.0002)
from the value of the assay without CrCl3.
rat adipocyte membranes, while PP2A is apparently restricted
to the cytosol (Begum, 1995); rat adipocytes contain virtually
no PP2B activity and only low levels of PP2C activity (Wood
et al., 1993). To determine which type of phosphatase was
activated by LMWCr, the effects of the addition of specific
phosphatase inhibitors on the activation by LMWCr were
examined, as was the effect of LMWCr on isolated phosphatases. In the same concentration range used with the
membrane fragments, LMWCr had no effect on alkaline
phosphatase or phosphoprotein phosphatase 1 (PP1). In
contrast, LMWCr activates the hydrolysis of p-NPP by the
catalytic fragment of Yersinia phosphotyrosine phosphatase;
the concentration dependence of the activation of the
fragment is essentially identical to that of the membrane
fragments (Figure 1), indicating that LMWCr must be
activating a PTP. The use of the isolated phosphatase clearly
suggests that inactivation of a kinase is not responsible for
the observed apparent dephosphorylation.
Rat adipocytes contain a number of PTP’s including LRP/
RPTPR, PTP 1B, SH-PTP2/Syp, LAR (Ding et al., 1994),
HA1 and HA2 (Liao et al., 1991). [HA2, however, is a
homolog of PTP 1B (Liao & Lane, 1995).] LAR, LRP/
RPTPa, HA1, and HA2 are membrane proteins. PTP’s can
be broadly distinguished by their behavior in the presence
of a variety of inhibitors (Pot & Dixon, 1992). For example
rat LAR is inhibited by EDTA, spermine, and spermidine
and almost unaffected by zinc cations (Pot et al., 1991); the
human analogue of RPTPR is inhibited by zinc, EDTA, and
spermine (Wang & Pullen, 1991). In contrast, rat adipocyte
membrane phosphotyrosine phosphatase activity demonstrates a different pattern of inhibition (Table 1). The PTP
activity is activated by EDTA, spermine, and spermidine but
inhibited by zinc cations. The activity is not consistent with
the dominate phosphatase activity being LAR, RPTPa, or
even alkaline phosphatase (inhibited by EDTA as it possesses
two zinc ions at its active site) (Vincent & Crowder, 1995).
Similarly, microcystin LR, a specific inhibitor of PP1 and
PP2A at nanomolar concentrations (Honkanen et al., 1990),
does not inhibit overall phosphatase activity. However,
microcystin does inhibit the activation of phosphatase activity
by LMWCr; the origin of this effect is uncertain. The
inhibition/activation pattern, however, corresponds to that
of two types of membrane PTP’s isolated from human
placenta (Tonks et al., 1988c), P1A′ and P1B. P1B, for
example, is activated toward hydrolysis of phosphotyrosyl
RCM lysozyme 166% by 5 mM EDTA, 40% by 2 mM
spermine, and 138% by 2 mM spermidine and is inhibited
94% by 100 µM Zn2+. The striking similarity between the
P1A′ and P1B enzymes and the major PTP of the adipocyte
membranes suggests that they are probably related. HA1
and HA2 are completely inhibited by 1 mM Zn2+ (Liao et
al., 1991). HA2, which is a PTP 1B homolog, possesses
approximately the same molecular weight as isolated P1B
(Liao et al., 1991; Tonks et al., 1988b, 1991) and may, then,
represent the major PTP activity in the membrane fragments.
The activation of the membrane PTP activity by LMWCr is
inhibited by the addition of the other activators (Table 1).
This suggests that LMWCr activates HA2 or a PTP with
similar behavior toward EDTA and the polycationic compounds. LMWCr could activate other PTP’s, but their
contribution cannot be elucidated by these studies. Curiously, microinjection of soluble placental PTP 1B into
Xenopus oocytes is antagonistic to insulin action (Cicirelli
et al., 1990); however, the enzyme seems to associate with
internal membrane fragments (rather than the plasma membrane) where it could block specific physiological responses
to insulin (Tonks et al., 1988a). However, insulin results in
the stimulation of the phosphatase activity of SH2-PTP2 (also
known as syp, PTP2C, or PTP1D) and SH2-PTP1 (also
known as PTP1C, HCP, or SHP) (Sugimoto et al., 1994;
Uchida et al., 1994). It is conceivable that the ability of the
reagents in Table 1 to inhibit activation by LMWCr of PTP
activity could be by interaction with LMWCr (making
conclusions about the identity of the PTP difficult); however,
as each of these reagents is an inhibitor or activator of the
PTP activity in the absence of LMWCr and, thus, known to
bind to the PTP already, this possibility is unlikely.
Studies of the activation of the membrane PTP by LMWCr
as a function of LMWCr and substrate concentration reveal
that not all of the phosphatase activity associated with the
membranes is activated by LMWCr (Figure S1 in supporting
information), as indicated by the curvatures of the lines in
the Lineweaver-Burk plot, distinctive of curvilinear (complex) activation. This is to be expected as the membranes
certainly must contain PP1 and other phosphatases, capable
of dephosphorylating p-NPP. Differential activation of the
various membrane PTP’s such as LAR, RPTPR, etc. may
also add to the curvature. As the curve is linear in the
absence of LMWCr, an apparent Km of 632 µM can be
estimated for the overall hydrolysis. This value is similar
to the Km reported for p-NPP hydrolysis catalyzed by PTP’s
[for example for rat LAR, Km ) 420 µM (Pot et al., 1991)].
Thus, the major component of the rat adipocyte membrane
PTP activity is affected by addition of LMWCr and has a
pattern of inhibition and activation by certain reagents similar
to that of PTP1A′ and PTP2B; this suggests the that
activation of PTP activity by LMWCr involves of a
membrane PTP similar to PTP1A′ or PTP2B.
Incubation of rat adipocyte membranes with monoclonal
antibodies whose epitope is the catalytic domain of human
PTP1B and which react with rat PTP1B results in the
reduction of the ability of LMWCr to stimulate PTP activity;
when 50 µM Cr is used as LMWCR, the antibodies result
in a loss of 78% of the oligopeptide’s ability to potentiate
PTP activity. Hence, the PTP(s) activated by LMWCr must
be PTP1B and/or closely related phosphotyrosine protein
phosphatases. This is also supported by studies using the
Bioactive Chromium May Activate Membrane PTP
FIGURE 2: Chromic ion titration of the ability of apoLMWCr to
activate the hydrolysis of p-NPP by rat adipocyte membrane
fragments. The concentration of apoLMWCr was 12.5 µM. 125
µL of a rat membrane suspension corresponding to 326 µg of
protein/mL was utilized.
isolated phosphotyrosine protein phosphatase LAR, which
has distinct inhibition properties from PTP1A′ and PTP1B;
no activation of this enzyme’s activity towards the dephosphorylation of p-NPP was observed over a range of LMWCr
concentrations (Cr concentrations varying from 5 nM to 50
µM).
Metal Reconstitution
Apoprotein (which contains circa 0.3 Cr) activates PTP
activity slightly (∼33%), while addition of apoprotein and
four equivalents of chromic ions completely restores the
activation potential of the polypeptide (Table 1). The
activation by high concentrations of apoprotein suggests that
only a tiny fraction of the protein still binding chromium is
present in the form containing four Cr centers while the
remainder probably contain inactive mononuclear chromium
centers. Integration of EPR spectra of apoprotein [using
hexaaquochromium(III) as standard] indicates that essentially
all of the chromium exists as mononuclear centers. Titration
of apoprotein with chromium(III) reveals that 3.89 Cr/protein
are required for complete restoration of the phosphatase
activation activity (Figure 2). This is consistent with reports
that LMWCr isolated from mammalian liver possesses four
chromic ions per polypeptide (Yamamoto et al., 1987).
[However, addition of more than 10 equiv results in
inhibition of the activation activity.] Chromic ions by
themselves are ineffective. PTP activation is, thus, dependent
on the Cr content; insulin potentiation by LMWCr is similarly
dependent on the Cr content (Yamamoto et al., 1989),
suggesting that the PTP activation is related to insulin
potentiation. Hence, the metal would appear to be important
in maintaining the proper conformation of the polypeptide.
With the exception of ferric ions, transition metal ions
commonly associated with biological systems (other than Cr)
Biochemistry, Vol. 35, No. 39, 1996 12967
FIGURE 3: Activation of rat adipocyte membrane phosphotyrosine
phosphatase activity using 5 mM p-NPP as substrate by apoLMWCr
(25 µM) and metal ions (100 µM). Bars marked with “*” differ
significantly (P < 0.0002) from the control assays (i.e., in the
absence of added metal) (unpaired Student t-test), and those marked
with “**” differ significantly (P < 0.0002) from the apoLMWCr
assays in the absence of added metal.
are ineffective in potentiating the ability of apoprotein to
activate the PTP activity (Figure 3).
Manganous ions have no effect on apoprotein, while all
others except ferric ions are inhibitory. Even activation by
ferric ions is extremely small although significant statistically
(P < 0.0002). LMWCr would appear then to be specific
for Cr3+. [In the absence of LMWCr, these transition metal
ions with the exception of Co, Fe, and Mn display inhibition
of phosphatase activity as in the presence of LMWCr. Co,
Mn, and Fe had no effect on phosphatase activity in the
absence of LMWCr (Figure 3). None of the metals examined
in the absence of apoprotein resulted in activation of the
phosphatase activity. The effect of the addition of cobalt to
apo-oligopeptide is most interesting as it differs appreciably
from the effect in the absence of apoLMWCr; this suggests
cobalt may bind appreciably to LMWCr and that the resulting
Co-LMWCr complex (although inactive) could serve as a
new spectroscopic probe of LMWCr.] The ferric ions
probably associate with the apoprotein because of their
similar charge to size ratio to chromic ions, giving a
conformation similar to that of the native polypeptide; note
that in Vitro chromic ions compete for ferric ions in ironbinding proteins such as transferrin (Aisen et al., 1969). How
selectivity for chromium versus iron is achieved in ViVo is
currently under investigation.
CONCLUSIONS
If LMWCr activates a PTP in response to insulin, the
question arises as to what aspect of insulin action results in
“turning on” LMWCr to activate the PTP. A clue may come
from the observation that within 90 min of ingestion of
glucose by humans, plasma insulin levels increase with a
12968 Biochemistry, Vol. 35, No. 39, 1996
parallel increase in urinary chromium loss (Anderson et al.,
1990); this suggests a mobilization of chromium in response
to insulin. As a low-molecular-weight Cr-containing species
may represent the form of Cr(III) in urine (Manzo et al.,
1983), LMWCr may somehow be mobilized or apoLMWCr
loaded with Cr in response to insulin. In order to address
these issues, studies are in progress to examine the distribution of LMWCr in rat hepatocytes before and after treatment
with insulin.
In conclusion, LMWCr activates a membrane phosphotyrosine phosphatase in a manner dependent on the Cr
content of the oligopeptide. Indirect evidence suggests that
this activation may play a role in the potentiation of insulin
action by LMWCr. Establishing whether this activation
activity has a direct role in insulin potentiation will require
in ViVo studies now in preparation. The previous assays for
chromium biological activity all involved the use of 14C- or
3H-labeled substrates; this assay may provide a simple, nonradioactive alternative for determining this activity.
SUPPORTING INFORMATION AVAILABLE
Lineweaver-Burk plot of LMWCr concentration and
p-NPP substrate concentration dependence of rat adipocyte
membrane PTP activity (1 page). Ordering information is
given on any current masthead page.
REFERENCES
Aisen, P., Aasa, R., & Redfield, A. G. (1969) J. Biol. Chem. 244,
4628-4633.
Anderson, R. A. (1985) in Trace Elements in Health and Disease,
pp 110-124, Norstedt, Stockholm.
Anderson, R. A. (1986) Clin. Physiol. Biochem. 4, 31-41.
Anderson, R. A. (1987) in Trace Elements in Human and Animal
Nutrition, Vol. 1, pp 225-244, Academic Press, Orlando, FL.
Anderson, R. A. (1994) in Risk Assessment of Essential Elements
(Mertz, W., Abernathy, C. O., & Olin, S. S., Eds.) pp 187-196,
ISLI Press, Washington, DC.
Anderson, R. A., & Kozlovsky, A. S. (1985) Am. J. Clin. Nutr. 41,
1177-1183.
Anderson, R. A., Brantner, J. H., & Polansky, M. M. (1978) J.
Agric. Food Chem. 26, 1219-1221.
Anderson, R. A., Polansky, M. M., Bryden, N. A., Roginski, E.
E., Patterson, K. Y., and Reamer, D. C. (1982) Diabetes 31, 213216.
Anderson, R. A., Polansky, M. M., Bryden, N. A., Roginski, E.
E., Mertz, W., & Glinsmann, W. (1983) Metabolism 32, 894899.
Anderson, R. A., Polansky, M. M., Bryden, N. A., Bhathena, S. J.,
& Canary, J. J. (1987) Metabolism 36, 351-355.
Anderson, R. A., Bryden, N. A., Polansky, M. M., & Reiser, S.
(1990) Am. J. Clin. Nutr. 51, 864-868.
Anderson, R. A., Polansky, M. M., Bryden, N. A., & Canary, J. J.
(1991) Am. J. Clin. Nutr. 54, 909-916.
Begum, N. (1995) J. Biol. Chem. 270, 709-714.
Cicirelli, M. F., Tonks, N. K., Diltz, C. D., Weiel, J. E., Fischer,
E. H., & Krebs, E. G. (1990) Proc. Natl. Acad. Sci. U.S.A. 87,
5514-5518.
Ding, W., Zhang, W.-R., Sullivan, K., Hashimoto, N., & Goldstein,
B. J. (1994) Biochem. Biophys. Res. Commun. 202, 902-907.
Evans, G. W., & Bowman, T. D. (1992) J. Inorg. Biochem. 46,
243-250.
Evans, G. W., & Pouchnik, D. J. (1993) J. Inorg. Biochem. 49,
177-187.
Frausto da Silva, J. R., & Williams, R. J. P. (1991) Biological
Chemistry of the Elements, Oxford University Press, Oxford,
England.
Glinsmann, W. H., & Mertz, W. (1966) Metabolism 15, 510-520.
Davis et al.
Gonzalez-Vergara, E., Hegenauer, J., & Saltman, P. (1982) Fed.
Proc. 41, 286.
Gurson, C. T., & Saner, G. (1971) Am. J. Clin. Nutr. 24, 13131319.
Hasten, D. L., Rome, E. P., Franks, B. D., & Hegsted, M. (1992)
Int. J. Sports Med. 2, 343-350.
Haylock, S. J., Buckley, P. D., & Blackwell, L. F. (1983) J. Inorg.
Biochem. 19, 105-117.
Honkanen, R. E., Zwiller, J., Moore, R. E., Daily, S. L., Khatra,
B. S., Dukelow, M., & Boynton, A. L. (1990) J. Biol. Chem.
265, 14901-14904.
Hopkins, L. L., Jr. (1965) Am. J. Physiol. 209, 731-735.
Hopkins, L. L., Jr., Ransome-Kati, O., & Majaj, A. S. (1968) Am.
J. Clin. Nutr. 21, 203-211.
Kaim, W., & Schwederski, B. (1994) Bioinorganic Chemistry:
Inorganic Elements in the Chemistry of Life, John Wiley & Sons,
Chichester, England.
Levine, R. A., Streeten, D. H. P., & Doisy, R. J. (1968) Metabolism
17, 114-125.
Li, J., Elberg, G., Gefel, D., & Shechter, Y. (1995) Biochemistry
34, 6218-6225.
Liao, K., & Lane, M. D. (1995) J. Biol. Chem. 270, 12123-12132.
Liao, K., Hoffman, R. D., & Lane, M. D. (1991) J. Biol. Chem.
266, 6544-6553.
Lippard, S. J., & Berg, J. M. (1994) Principles of Bioinorganic
Chemistry, University Science Books, Mill Valley, CA.
Mahdi, G. S., & Naismith, D. J. (1991) Ann. Nutr. Metab. 35, 6570.
Manzo, L., Di Nucci, A., Edel, J., Gregotti, C., & Sabbioni, E.
(1983) Res. Commun. Chem. Pathol. Pharmacol. 42, 113-125.
Marczenko, Z. (1986) Spectrophotometric Determination of the
Elements, Ellis Horwood, Chichester, England.
Martinez, O. B., MacDonald, A. C., Gibson, R. S., & Bourn, O.
(1985) Nutr. Res. 5, 609-620.
McCarty, M. F. (1993) J. Optim. Nutr. 2, 36-53.
Mertz, W., & Schwarz, K. (1959) Am. J. Physiol. 196, 614-618.
Mertz, W., & Roginski, E. E. (1963) J. Biol. Chem. 238, 868872.
Mertz, W., Roginski, E. E., & Reba, R. C. (1965) Am. J. Physiol.
209, 489-494.
Pot, D. A., & Dixon, J. E. (1992) Biochim. Biophys. Acta 1136,
35-43.
Pot, D. A., Woodford, T. A., Remboutsika, E., Haun, R. S., &
Dixon. (1991) J. Biol. Chem. 266, 19688-19696.
Riales, R., & Albrink, M. J. (1981) Am. J. Clin. Nutr. 34, 26702678.
Rodbell, M. (1964) J. Biol. Chem. 239, 375-380.
Saltiel, A. R. (1994) FASEB J. 8, 1034-1040.
Schwartz, K. & Mertz, W. (1957) Arch. Biochem. Biophys. 72,
515-518.
Schwartz, K., & Mertz, W. (1959) Arch. Biochem. Biophys. 85,
292-295.
Schroeder, H. A. (1965) Life Sci. 4, 2057-2062.
Schroeder, H. A., & Balassa, J. (1965) Am. J. Physiol. 209, 489494.
Shepherd, P. R., Elwood, C., Buckley, P. D., & Blackwell, L. F.
(1992) Biol. Trace Element Res. 32, 109-113.
Stearns, D. M., Wise, J. P., Sr., Patierno, S. R., & Wetterhahn, K.
E. (1995) FASEB J. 9, A451.
Sugimoto, G., Wandless, T. J., Shoelson, S. E., Neel, B. G., &
Walsh, C. T. (1994) J. Biol. Chem. 269, 13614-13622.
Toepfer, E. W., Mertz, W., Polansky, M. M., Roginski, E. E., &
Wolf, W. R. (1977) J. Agric. Food Chem. 25, 162-166.
Tonks, N. K., Cicirelli, M. F., Diltz, C. D., Krebs, E. G., & Fischer,
E. H. (1988a) Mol. Cell. Biol. 10, 458-463.
Tonks, N. K., Diltz, C. D., & Fischer, E. H. (1988b) J. Biol. Chem.
263, 6722-6730.
Tonks, N. K., Diltz, C. D., & Fischer, E. H. (1988c) J. Biol. Chem.
263, 6731-6737.
Tonks, N. K., Diltz, C. D., & Fischer, E. H. (1991) Methods
Enzymol. 201, 427-442.
Uchida, T., Matozaki, T., Noguchi, T., Yamao, T., Horita, K.,
Suzuki, T., Fujioka, Y., Sakamoto, C., & Kasuga, M. (1994) J.
Biol. Chem. 269, 12220-12228.
Bioactive Chromium May Activate Membrane PTP
Undenfriend, S., Stein, S., Bohlen, P., Dairman, W., Leimgruber,
W., & Weigele, M. (1972) Science 178, 871-872.
Vincent, J. B. (1994a) in Encyclopedia of Inorganic Chemistry
(King, B., Ed.) Vol. 2, pp 661-665, John Wiley & Sons, New
York.
Vincent, J. B. (1994b) J. Nutr. 124, 117-118.
Vincent, J. B., & Crowder, M. W. (1995) Phosphatases in Cell
Metabolism and Signal Transduction, R. G. Springer-Verlag,
New York.
Wada, O., Wu, G. Y., Yamamoto, A., Manabe, S., & Ono, T. (1983)
EnViron. Res. 32, 228-239.
Wang, Y., & Pallen, C. J. (1991) EMBO J. 10, 3231-3237.
Biochemistry, Vol. 35, No. 39, 1996 12969
Wood, S. L., Emmison, N., Borthwick, A. C., & Yeaman, S. J.
(1993) Biochem. J. 295, 531-535.
Woolliscroft, J., & Barbosa, J. (1977) J. Nutr. 107, 1702-1706.
Yamamoto, A., Wada, O., & Ono, T. (1987) Eur. J. Biochem. 165,
627-631.
Yamamoto, A., Wada, O., & Suzuki, H. (1988) J. Nutr. 118, 3945.
Yamamoto, A., Wada, O., & Manabe, S. (1989) Biochem. Biophys.
Res. Commun. 163, 189-193.
Zhang, Z.-Y., & VanEtten, R. L. (1991) J. Biol. Chem. 266,
1516-1525.
BI960328Y
Birth Defects Research (Part B) 83:27–31 (2008)
& 2007 Wiley-Liss, Inc.
Original Article
Comparison of the Potential for Developmental
Toxicity of Prenatal Exposure to Two Dietary
Chromium Supplements, Chromium Picolinate
and [Cr3O(O2CCH2CH3)6(H2O)3]1, in Mice
M.M. Bailey,1 J. Sturdivant,1 P.L. Jernigan,1 M.B. Townsend,1 J. Bushman,1 I. Ankareddi,2 J.F. Rasco,1
R.D. Hood,1,3 and J.B. Vincent4
1
Department of Biological Sciences, The University of Alabama, Tuscaloosa, Alabama
Department of Chemical and Biological Engineering, The University of Alabama, Tuscaloosa, Alabama
3
Ronald D. Hood and Associates, Toxicology Consultants, Tuscaloosa, Alabama
4
Department of Chemistry and Coalition for Biomolecular Products, The University of Alabama, Tuscaloosa, Alabama
2
BACKGROUND: Chromium(III) is generally thought to be an essential trace element that allows for proper glucose
metabolism. However, chromium(III) picolinate, Cr(pic)3, a popular dietary supplement form of chromium, has been
shown to be capable of generating hydroxyl radicals and oxidative DNA damage in rats. The cation
[Cr3O(O2CCH2CH3)6(H2O)3]1, Cr3, has been studied as an alternative supplemental source of chromium. It has been
shown to increase insulin sensitivity and lower glycated hemoglobin levels in rats, making it attractive as a potential
therapeutic treatment for gestational diabetes. To date, no studies have been published regarding the safety of Cr3
supplementation to a developing fetus. METHODS: From gestation days (GD) 6–17, mated CD-1 female mice were fed
diets delivering either 25 mg Cr/kg/day as Cr(pic)3, 3.3 or 26 mg Cr/kg/day as Cr3, or the diet only to determine if Cr3
could cause developmental toxicity. Dams were sacrificed on GD 17, and their litters were examined for adverse effects.
RESULTS: No signs of maternal toxicity were observed. No decrease in fetal weight or significantly increased incidence
of skeletal defects was observed in the Cr3 or Cr(pic)3 exposed fetuses compared to the controls. CONCLUSION:
Maternal exposure to either Cr(pic)3 or Cr3 at the dosages employed did not appear to cause deleterious effects to the
developing offspring in mice. Birth Defects Res (Part B) 83:27–31, 2008.
r 2007 Wiley-Liss, Inc.
Key words: chromium; picolinate; Cr3; fetus; developmental toxicity; mice
INTRODUCTION
Chromium(III) is generally thought to be an essential
trace element, enhancing glucose tolerance, protein
and lipid metabolism, and serum cholesterol (Abraham
et al., 1992; Anderson et al., 1983, 1991; Glinsmann
and Mertz, 1966; Gurson and Saner, 1971; Hopkins et al.,
1968; Press et al., 1990; Shinde et al., 2004). Dietary
chromium is not readily absorbed (about 0.5–2%),
making the coupling of chromium to a suitable ligand
necessary to increase its bioavailability. Identifying
the biologically active form of chromium—the biomolecule that naturally binds chromium(III) and possesses
an intrinsic function associated with insulin action—
has been of interest in recent years (Sun et al., 2002).
While chromium picolinate, Cr(pic)3, was once thought
to be the biologically active form of chromium, that is
unlikely because in vivo levels of chromium and
picolinic acid make its natural biosynthesis unlikely,
and it has not been shown to be biologically active
(Vincent, 2000a, 2001). More recently, low-molecular-
weight-chromium-binding substance, or chromodulin,
has been proposed to be the biologically active form
of chromium (Sun et al., 2002; Vincent, 2000a, 2001).
Nevertheless, chromium picolinate has traditionally
been one of the most popular forms of chromium being
sold as a dietary supplement, with sales in the billions of
dollars.
Chromodulin is an oligopeptide that has been proposed to be part of a unique autoamplification system for
insulin signaling through its effect on the tyrosine kinase
activity of insulin-activated insulin receptors, resulting in
an increase in insulin sensitivity (Davis et al., 1997;
Vincent, 2000a,b). In this model, apochromodulin is
stored in insulin-sensitive cells. Increases in blood insulin
*Correspondence to: Ronald D. Hood, R. D. Hood & Associates,
Toxicology Consultants, Box 870344, Tuscaloosa, AL 35487-0344.
E-mail: [email protected]
Received 24 August 2007; Accepted 22 October 2007
Contract grant sponsor: Howard Hughes Medical Institute.
Published online in Wiley InterScience (www.interscience.wiley.com)
DOI: 10.1002/bdrb.20140
28
BAILEY ET AL.
concentration result in an increase in insulin binding in
sensitive cells. The binding of insulin brings about a
conformation change that results in the autophosphorylation of tyrosine residues on the internal side of the
receptor, transforming the receptor into an active tyrosine
kinase and transmitting the signal from the insulin to the
cell. In response to insulin, chromium is moved into the
insulin-sensitive cells from the blood, where the apochromodulin is loaded with chromium. The resulting
holochromodulin subsequently binds to the insulin
receptor, which is presumed to aid in the maintenance
of the receptor in its active conformation, amplifying its
kinase activity. A drop in blood insulin levels facilitates a
relaxation of the conformation of the receptor, and the
holochromodulin exits the cells into the blood and is
excreted in the urine (Vincent, 2004). Chromium is
essential to the function of chromodulin, as apochromodulin alone has not been shown to increase receptor
activity (Davis and Vincent, 1997).
A chromodulin biomimetic has been described by
Vincent (2004), and it was found to stimulate insulin
receptor’s tyrosine kinase ability in a manner similar to
that of chromodulin. This biomimetic, the trinuclear
cation [Cr3O(O2CCHCH3)6 (H2O)3]1, or Cr3, is a multinuclear oxo-bridged chromium(III) carboxylate assembly
(Sun et al., 2002). In contrast to Cr(pic)3, Cr3 is absorbed
with much greater efficiency (40–60%, vs.r2%) (Clodfelder et al., 2004). Intravenous and oral administration of
Cr3 has been shown to lower plasma triglycerides, total
cholesterol, low-density lipoprotein (LDL), and insulin
concentrations and to increase insulin sensitivity in
healthy, obese, and type 2 diabetic rat models (Sun
et al., 1999, 2002). Additionally, the biomimetic has been
shown to lower fasting plasma glycated hemoglobin
levels in obese and type 2 diabetic rats (Clodfelder et al.,
2005). Thus, administration of this compound might be
useful in improving blood parameters in type 2 diabetics
and in individuals at risk for type 2 diabetes because of
obesity.
Stearns and coworkers (1995a) were among the first to
report deleterious effects of Cr(pic)3 when they demonstrated that it was clastogenic and later that it was
mutagenic in Chinese hamster ovary cells (Stearns et al.,
2002). Chromium from Cr(pic)3 accumulates in cells and
has nuclear affinity, implying that supplemental levels
over long periods of time might be harmful (Stearns
et al., 1995b). While Cr(pic)3 was not found to be
mutagenic in Salmonella typhimurium, it was found to
induce mutagenic responses in the L5178Y mouse
lymphoma mutation assay (Whittaker et al., 2005).
Recent studies have failed to demonstrate that commercially prepared Cr(pic)3 was clastogenic or mutagenic
(Gudi et al., 2005; Slesinski et al., 2005). However, these
apparently conflicting results have recently been reconciled, as the studies in which damage was not observed
used dimethyl sulfoxide, a radical trap that could quench
reactive oxygen species, as a solvent for Cr(pic) 3 (Coryell
and Stearns, 2006). Chromium picolinate was also found
to be capable of cleaving DNA under physiologically
relevant conditions (Speetjens et al., 1999a), and was
recently found by our laboratory to cause significant
increases in cervical arch defects in offspring of exposed
pregnant mice (Bailey et al., 2006). Conversely, both
chromodulin and Cr3 were found in a recent study not to
cleave DNA under physiologically relevant conditions.
That finding supports the possibility that Cr3 may
present an alternative to Cr(pic)3 as a nutritional
supplement, although the effects of long term supplementation with Cr3 have not yet been reported (Speetjens
et al., 1999b).
Pregnancy may increase urinary chromium loss,
making pregnant women susceptible to chromium
deficiency (Morris et al., 1995). Moreover, the results of
a study by Jovanovic et al. (1999) indicate that chromium
supplementation may be useful in the treatment of
gestational diabetes, a common pregnancy complication.
The increased bioavailability of Cr3 and its ability to
increase insulin sensitivity and decrease glycated hemoglobin levels indicate that Cr3 may be an effective aid in
the treatment of gestational diabetes; however, no studies
to date have tested its safety or efficacy in a mammalian
pregnancy. Thus, the current study evaluated the developmental toxicity potential of Cr3 in comparison with
that of Cr(pic)3 in a mammalian model.
MATERIALS AND METHODS
Animals and Husbandry
Male and female CD-1 mice, obtained from Charles
River Breeding Laboratories, International (Wilmington,
MA) were housed in an AAALAC-approved animal
facility in rooms maintained at 22721C, with 40–60%
humidity and a 12-hr photoperiod. Animals were bred
naturally, two females with one male. Observation of a
copulation plug was designated GD 0. Mated females
were individually housed in shoe-box-type cages
with hardwood bedding and were given Harlan-Teklad
LM-485 rodent diet and tap water ad libitum.
Test Chemicals
Chromium(III) picolinate, Cr(pic)3, was synthesized
according to the methods of Press et al. (1990). Cr3 was
synthesized according to the methods of Earnshaw et al.
(1966). The authenticity of both was established by high
resolution electron impact mass spectrometry (Chakov
et al., 1999; van den Bergen et al., 1993). Picolinic acid
was purchased from Fisher Scientific (Pittsburgh, PA).
LM-485 milled rodent diet was purchased from Harlan
Teklad (Madison, WI). Either Cr(pic)3 or Cr3 was added
to milled rodent chow in sufficient quantities to achieve
the appropriate concentration of the test compound. All
calculations were based on data from previous studies,
which indicated that pregnant CD-1 mice consume an
average of 7 g diet/day. Extensive stability studies
indicate that chromium test compounds are extremely
stable and that no degradation in the diet would be
expected (Chakov et al., 1999).
Because the purpose of this study was to determine the
effects of pharmaceutical levels of chromium, no special
measures were taken to prevent exposure of the mice to
small amounts of chromium that may be introduced into
the diet through methods of feed preparation or from the
cage hardware. The diet purchased, Teklad LM-485
(7012), contained added chromium in the form of
chromium potassium sulfate (0.48 mg/kg of diet).
Treatments
Mated females were randomly assigned to one of four
treatment groups: (1) control, untreated rodent diet,
Birth Defects Research (Part B) 83:27–31, 2008
COMPARISON OF CHROMIUM SUPPLEMENTS
(2) 200 mg/kg/day Cr(pic)3 (providing 25 mg Cr/kg/
day), (3) 15 mg/kg/day Cr3 (providing 3.3 mg Cr/kg/
day), or (4) 120 mg/kg/day Cr3 (providing 26 mg Cr/
kg/day). All test diets were administered from GD 6 to
GD 17. Food consumption was measured for the
intervals GD 6–9, 9–13, and 13–17, and clinical signs
were recorded daily. The dosage of chromium picolinate
was based on results from a previous study, in which
25 mg Cr/kg/day as Cr(pic)3 was associated with a
significant increase in the incidence of cervical arch
defects compared to control animals (Bailey et al., 2006).
Dosages of Cr3 were thus chosen to either deliver an
equivalent dosage of chromium based on bioavailability
(15 mg/kg/day) or to deliver approximately the same
mass of chromium (120 mg/kg/day) as the 200 mg/kg/
day dose of Cr(pic)3. Cr nutritional supplements generally provide 200 to 1,000 mg Cr, while recent human
clinical trials have tended to use 1,000 mg Cr daily.
Assuming the average human body weight is 65 kg, this
range corresponds to B3 to 15 mg Cr/kilogram body
weight/day.
Data Collection
On GD 17, mated females were euthanized by a CO2
overdose. The uteri of pregnant dams were exposed, and
the numbers of resorptions and dead or live fetuses were
recorded. Live fetuses were removed from the uterus,
weighed individually, and examined for gross malformations. Maternal body weight, minus the gravid uterine
weight, was then obtained. Litters were initially fixed in
70% ethanol and then cleared and stained by a doublestaining technique (Webb and Byrd, 1994). The fetuses
were subsequently examined for skeletal anomalies.
Data Analysis
The litter was used as the experimental unit for
statistical analysis. The data from each study replicate
were calculated independently, tested for homogeneity of
variance by the Levene statistic using SPSS (SPSS Inc.,
Chicago, IL), and then the replicates were pooled and
analyzed together to give the reported results. All tabular
data are presented as the mean7SEM, and the mean
value for each parameter was calculated as the mean of
the litter means. Data were analyzed by one-way analysis
of variance (ANOVA), followed by an LSD post-hoc test
to determine specific significant differences (Pr0.05).
29
RESULTS
Maternal Data
Maternal weight gain was not affected by the administration of Cr(pic)3 or Cr3 (Table 1). No signs of maternal
toxicity were observed for dams in any of the groups.
Food consumption was virtually identical among the
treatment groups, and average food consumption was
approximately 7 g diet/day.
Fetal Data
Fetal weight and percentage of resorbed or dead
fetuses did not differ among treatment groups (Table 1),
and no gross malformations were observed in any of the
fetuses. No difference was found in the number of
implantations per litter in treated groups as compared to
controls; however, the numbers of implantations in litters
from the Cr3 15 mg/kg/day dosage group were significantly lower than those of either the Cr(pic)3 or the
Cr3 120 mg/kg/day groups. As this was clearly not a
dose-related occurrence, and because maternal exposures
did not commence until after implantation should have
occurred, the authors feel that this was an anomaly that
is not an adverse effect of treatment. No skeletal defects
in fetuses from Cr(pic)3- or Cr3-treated dams differed in
incidence from the control value.
DISCUSSION
While chromium(III) is generally believed to be much less
harmful than chromium(VI), a known carcinogen and
mutagen (Bagchi et al., 2002), both forms are capable of
generating the same oxidation states during metabolism.
Chromium(V) and chromium(IV) intermediates are believed
to be the active mutagens and carcinogens that result from
chromate exposure (Dillon et al., 2000; Levina et al., 1999;
Sugden and Wetterhahn, 1997). Chromium(VI) is reduced to
chromium(V), and chromium(IV), and finally to chromium(III) during its residence in the body (O’Flaherty et al.,
2001). Chromium(III) is generally more stable but can be
susceptible to reduction, depending on the coupling ligand.
Picolinate is an aromatic, bidentate ligand that coordinates via pyridine-type nitrogens and is thought to make
chromium(III) more susceptible to reduction in the presence
of biological reductants by shifting the redox potential of the
chromic center (Speetjens et al., 1999a). Chromium(III) is
reduced to chromium(II), which then enters the Fenton
Table 1
Maternal and Fetal Findings Following Exposure to Cr(pic)3 or Cr3 in CD-1 Mice
Treatment and dose (mg Cr/kg/day)
Fetuses/litters examined
Maternal weight gain (g7SEM)
Fetal weight (g7SEM)
Implantations (mean7SEM)b
Resorbed or dead fetuses (No.7SEM)b
Cervical arch defectsc (%7SEM)b
a
Control (0)
Cr(pic)3 (25)
Cr3 (3.3)
Cr3 (26)
332/27
12.3970.37
1.0270.02
12.6470.50
2.7471.07
4.6571.14
369/29
11.9670.36
1.0570.02
13.1870.35
3.2970.86
6.2671.63
275/26
11.4370.48
1.0870.03
11.0070.92a
3.4870.95
5.1871.58
342/24
13.0170.42
1.0270.03
13.7970.55
1.2970.64
3.9871.33
Significantly different from Cr(pic)3 (25 mg Cr/kg/day), and Cr3 (26 mg Cr/kg/day) values.
The mean value for each parameter was calculated as the mean of the litter means.
c
Cervical arch defects refer to a distal split in the first or second cervical vertebral arch.
b
Birth Defects Research (Part B) 83:27–31, 2008
30
BAILEY ET AL.
and/or Haber-Weiss cycles, where the binding of dioxygen
to the chromium center occurs concomitant with the
oxidation of chromium(II) to chromium(IV) and reduction
of dioxygen to hydrogen peroxide. The reduced dioxygenderived species can undergo O-O bond cleavage, generating
hydroxyl radicals that are able to react with DNA, causing
oxidative DNA damage. The Cr(pic)3 complex may enter
cells intact, allowing it to generate these byproducts in vivo
and damage cellular DNA (Hepburn and Vincent, 2003a).
Chromium(III) has definite potential to be genotoxic under
the right conditions, given that it has been shown to
accumulate intracellularly (Stearns et al., 1995b). Speetjens
and coworkers (1999b) demonstrated that while Cr(pic)3
increased the rate of DNA cleavage as measured by the
appearance of circular DNA in the presence of physiologically relevant concentrations of ascorbate, Cr3 did not
increase these rates. The lack of apparent toxicity of Cr3
stems from its chemical structure. Cr3 lacks any nitrogenbased ligation, and thus it is not susceptible to reduction by
ascorbate or thiols.
Previously, offspring of Cr(pic)3-treated dams were
shown to have a significantly increased incidence of
cervical arch defects (Bailey et al., 2006), but those results
were not replicated in the current study, mainly because of
an increased incidence of arch defects in the control
offspring. The incidence of cervical arch defects in
offspring exposed to Cr(pic)3 in this study (6.26%) was
in fact very similar to the results of the previous study
(5.79%). Historical control data from previous studies
indicate that the controls in this study may have displayed
a higher than normal incidence of arch defects (historical
range was 0–2.09%), but the reason for the increase in the
current study remains unknown. Although the offspring
exposed to Cr(pic)3 in utero still had a slightly greater
incidence of cervical arch defects than either the control or
Cr3-exposed offspring, the difference was too small to be
definitively attributable to the Cr(pic)3 treatment.
The previous study proposed that either picolinic acid
or the chemical combination of the picolinic acid and
chromium, but not the chromium solely, was responsible
for the deleterious effects observed, as offspring of
CrCl3-treated dams did not show an increased incidence
of cervical arch defects. Offspring of the Cr3 (26 mg Cr/
kg/day)-treated dams did not have an elevated incidence
of skeletal or other defects, despite the fact that they were
presumably exposed to a much higher dosage of
chromium than the offspring of the Cr(pic)3 exposed
dams (10.4 mg bioavailable Cr vs. 1.25 mg bioavailable in
Cr(pic)3, assuming 40 and 5% bioavailability, respectively). However, the relative availability of Cr3 to the
embryo and fetus has not yet been determined, so their
relative Cr exposure is as yet only speculative.
The enhancement of insulin sensitivity and lowering of
glycated hemoglobin levels, combined with its apparent
low potential as a developmental toxicant, indicate that
Cr3 may be a useful chromium-containing therapeutic in
the treatment of gestational diabetes. However, additional research would be required to fully assess its
possible benefits, as well as the potential risk of harm.
ACKNOWLEDGMENTS
This work was supported in part by a Howard Hughes
Medical Institute Undergraduate Biological Sciences
Education Program grant to The University of Alabama.
REFERENCES
Abraham A, Brooks B, Eylath U. 1992. The effects of chromium
supplementation on serum glucose and lipids in patients with and
without non-insulin dependent diabetes. Metabolism 41:768–771.
Anderson R, Polansky M, Bryden N, Roginski E, Mertz W, Glinsmann W.
1983. Chromium supplementation of human subjects: effects on
glucose, insulin, and lipid parameters. Metabolism 32:894–899.
Anderson R, Polansky M, Bryden N, Canary J. 1991. Supplementalchromium effects on glucose, insulin, glucagon, and urinary
chromium losses in subjects consuming controlled low-chromium
diets. Am J Clin Nutr 54:909–916.
Bagchi D, Stohs S, Downs B, Bagchi M, Preuss H. 2002. Cytotoxicity and
oxidative mechanisms of different forms of chromium. Toxicology
180:5–22.
Bailey M, Boohaker J, Sawyer R, Behling J, Rasco J, Jernigan J, Hood R,
Vincent J. 2006. Exposure of pregnant mice to chromium picolinate
results in skeletal defects in their offspring. Birth Defects Res B
77:244–249.
Chakov N, Collins R, Vincent J. 1999. A re-investigation of the electronic
spectra of chromium(III) picolinate complexes and high yield
synthesis and characterization of Cr2(m-OH)2(pic)4 5H2O (Hpic 5 picolinic acid). Polyhedron 18:2891–2897.
Clodfelder B, Chang C, Vincent J. 2004. Absorption of the biomimetic
chromium cation triaqua-m3-oxo-m-hexapropionatotrichromium(III)
in rats. Biol Trace Elem Res 97:1–11.
Clodfelder B, Gullick B, Lukaski H, Neggers Y, Vincent J. 2005. Oral
administration of the biomimetic [Cr3O(O2CCH2CH3)6 (H2O)3]1
increases insulin sensitivity and improves blood plasma variables
in healthy and type 2 diabetic rats. J Biol Inorg Chem 10:119–130.
Coryell V, Stearns D. 2006. Molecular analysis of hprt mutations induced
by chromium picolinate in CHO AA8 cells. Mutat Res 610:114–123.
Davis C, Vincent J. 1997. Chromium oligopeptide activates insulin
receptor tyrosine kinase activity. Biochemistry 36:4382–4385.
Davis C, Royer A, Vincent J. 1997. Synthetic multi-nuclear chromium
assembly activates insulin receptor kinase activity: functional model
for low-molecular-weight chromium-binding substance. Inorg Chem
36:5316–5320.
Dillon C, Lay P, Bonin A, Cholewa M, Legge G. 2000. Permeability, cytotoxicity,
and genotoxicity of Cr(III) complexes and some Cr(IV) analogues in V79
Chinese hamster lung cells. Chem Res Toxicol 13:742–748.
Earnshaw A, Figgis B, Lewis J. 1966. Chemistry of polynuclear
compounds. Part IV. Magnetic properties of trimeric chromium
and iron compounds. J Chem Soc (A):1656–1663.
Glinsmann W, Mertz W. 1966. Effects of trivalent chromium on glucose
tolerance. Metabolism 15:510–519.
Gudi R, Slesinski R, Clarke J, San R. 2005. Chromium picolinate does not
produce chromium damage in CHO cells. Mutat Res 587:140–146.
Gurson C, Saner G. 1971. Effect of chromium on glucose utilization in
marasmic protein-calorie nutrition. Am J Clin Nutr 24:1313–1319.
Hepburn D, Vincent J. 2003a. Tissue and subcellular distribution of
chromium picolinate with time after entering the bloodstream. J
Bioinorg Chem 94:84–93.
Hopkins LJ, Ransome-Kuti O, Majaj A. 1968. Improvement of impaired
carbogydrate metabolism by chromium(III) in sensitizing infants.
Am J Clin Nutr 21:203–211.
Jovanovic L, Gutierrez M, Peterson C. 1999. Chromium supplementation
for women with gestational diabetes mellitus. J Trace Elements Exp
Nutr 12:91–97.
Levina A, Barr-David G, Codd R, Lay P, Dixon N, Hammershoi A,
Hendry P. 1999. In vitro plasmid DNA cleavage by chromium(V) and
(IV) 2-hydroxycarboxylato complexes. Chem Res Toxicol 12:371–381.
Morris B, MacNeil S, Fraser R, Gray T. 1995. Increased urine chromium
excretion in normal pregnancy. Clinical Chemistry 41:1544–1545.
O’Flaherty E, Kerger B, Hays S, Paustenbach D. 2001. A physiologically
based model for the ingestion of chromium(III) and chromium(VI) by
humans. Toxicol Sci 60:196–213.
Press R, Gellar J, Evans G. 1990. The effects of chromium picolinate on
serum cholesterol and apolipoprotein fractions in human subjects.
West J Med 152:41–45.
Shinde U, Sharma G, Xu Y, Dhalla N, Goyal R. 2004. Insulin sensitizing
action of chromium picolinate in various experimental models of
diabetes mellitus. J Trace Elements Med Biol 18:23–32.
Slesinski R, Clarke J, San R, Gudi R. 2005. Lack of mutagenicity of chromium
picolinate in the hypoxanthine phosphoribosyltransferase gene mutation assay in Chinese hamster ovary cells. Mutat Res 585:86–95.
Speetjens J, Collins R, Vincent J, Woski S. 1999a. The nutritional
supplement chromium(III) tris(picolinate) cleaves DNA. Chem Res
Toxicol 12:483–487.
Speetjens J, Parand A, Crowder M, Vincent J, Woski S. 1999b. Lowmolecular-weight chromium-binding substance and biomimetic
[Cr3O(O2CCH2CH3)6 (H2O)3]1 do not cleave DNA under physiologically-relevant conditions. Polyhedron 18:2617–2624.
Birth Defects Research (Part B) 83:27–31, 2008
COMPARISON OF CHROMIUM SUPPLEMENTS
Stearns D, J Wise S, Patierno S, Wetterhahn K. 1995a. Chromium(III)
picolinate produces chromosome damage in Chinese hamster ovary
cells. FASEB J 9:1643–1648.
Stearns D, Belbruno J, Wetterhahn K. 1995b. A prediction of chromium(III) accumulation in humans from chromium dietary supplements. FASEB J 9:1650–1657.
Stearns D, Silveira S, Wolf K, Luke A. 2002. Chromium(III) tris(picolinate)
is mutagenic at the hypoxanthine (guanine) phosphoribotransferase
locus in Chinese hamster ovary cells. Mutat Res 513:135–142.
Sugden K, Wetterhahn K. 1997. Direct and hydrogen peroxide-induced
chromium(V) oxidation of deoxyribose in single-stranded and
double-stranded calf thymus DNA. Chem Res Toxicol 10:1397–1406.
Sun Y, Mallya K, Ramirez J, Vincent J. 1999. The biomimetic
[Cr3O(O2CCH2CH3)6 (H2O)3]1 decreases plasma cholesterol and
triglycerides in rats: towards chromium-containing therapeutics. J
Biol Inorg Chem 4:838–845.
Sun Y, Clodfelder B, Shute A, Irvin T, Vincent J. 2002. The biomimetic
[Cr3O(O2CCH2CH3)6 (H2O)3]1 decreases plasma insulin, cholesterol,
Birth Defects Research (Part B) 83:27–31, 2008
31
and triglycerides in healthy and type II diabetic rats but not type I
diabetic rats. J Biol Inorg Chem 7:852–862.
Van den Bergen A, Colton R, Percy M, West B. 1993. Electrospray mass
spectrometric study of [M3O(RCOO)6L3]1 cations (M 5 chromium,
iron; L 5 H2O, MeOH, py). Inorg Chem 32:3408–3411.
Vincent J. 2000a. Elucidating a biological role for chromium at a
molecular level. Accounts Chem Res 33:503–510.
Vincent J. 2000b. Quest for the molecular mechanism of chromium action
and its relationship to diabetes. Nutr Rev 58:67–72.
Vincent J. 2001. The bioinorganic chemistry of chromium(III). Polyhedron
20:1–26.
Vincent J. 2004. Recent advances in the nutritional biochemistry of
trivalent chromium. Proc Nutr Soc 63:41–47.
Webb G, Byrd R. 1994. Simultaneous differential staining of cartilage and
bone without glacial acetic acid. Biotech Histochem 69:181–185.
Whittaker P, San R, Clarke J, Seifried H, Dunkel V. 2005. Mutagenicity of
chromium picolinate and its components in Salmonella typhimurium
and L5178Y mouse lymphoma cells. Food Chem Toxicol 43:1619–1625.
Polyhedron 18 (1999) 2617–2624
Low-molecular-weight chromium-binding substance and biomimetic
[Cr 3 O(O 2 CCH 2 CH 3 ) 6 (H 2 O) 3 ] 1 do not cleave DNA under
physiologically-relevant conditions
a
b
b
a,
J. Kristopher Speetjens , Ali Parand , Michael W. Crowder , John B. Vincent *,
a
Stephen A. Woski
a
Department of Chemistry and Coalition for Biomolecular Products, University of Alabama, Tuscaloosa, AL 35487 -0336, USA
b
Department of Chemistry and Biochemistry, Miami University, Miami, OH 45056, USA
Received 10 March 1999; accepted 1 June 1999
Abstract
Chromium(III) tris(picolinate), Cr(pic) 3 , is currently a very popular nutritional supplement; however, at physiologically-relevant
concentrations, it has recently been demonstrated to cleave DNA [J.K. Speetjens, R.A. Collins, J.B. Vincent, S.A. Woski, Chem. Res.
Toxicol. 12 (1999) 483]. A number of other chromium-containing compounds have been proposed as substitutes for Cr(pic) 3 . Of
particular interest are low-molecular-weight chromium-binding substance (LMWCr) and [Cr 3 O(O 2 CCH 2 CH 3 ) 6 (H 2 O) 3 ] 1 1. The former
compound has recently been identified as the biologically active form of chromium in mammals, activating the kinase activity of insulin
receptor in the presence of insulin. Complex 1 is a functional biomimetic for LMWCr. Both compounds have been proposed as possible
nutritional supplements and therapeutics for adult-onset diabetes. This work demonstrates that these complexes, unlike Cr(pic) 3 , are poor
DNA-cleaving agents and may represent safer materials for human consumption.  1999 Elsevier Science Ltd. All rights reserved.
Keywords: Low-molecular-weight chromium-binding substance; Chromium(III) tris(picolinate); Chromium(III) complexes; DNA nicking complexes;
Trinuclear complexes
1. Introduction
Chromium is an essential nutrient for proper carbohydrate and lipid metabolism in mammals. Dietary sources of
chromium are poorly absorbed, with an average efficiency
of 0.5% [1]. It is estimated that approximately 90% of
Americans do not consume the daily adequate and safe
quantity of chromium in their diets [2]. Understandably,
the manufacture and sale of chromium(III) tris(picolinate),
Cr(pic) 3 , as a more readily absorbable nutritional supplement has become a multimillion dollar industry. The
compound is available in a variety of forms for human
consumption including pills, sports drinks, and chewing
gums. Although a lack of acute toxicity has been observed
for Cr(pic) 3 fed to rats in doses up to 30 mg per kg body
weight [3], questions have been raised regarding the health
effects of ingesting the compound [4–6]. Wetterhahn et al.
have shown that Cr(pic) 3 can induce chromosome damage
*Corresponding author. Tel.: 11-205-348-9203; fax: 11-205-3489104.
E-mail address: [email protected] (J.B. Vincent)
in cultured Chinese hamster ovary cells [4]. Some of the
clastogenic damage was shown to result from free picolinate, while Cr(pic) 3 had increased clastogenicity compared
to picolinate. The mechanism of DNA damage has recently
been elucidated [7]. The picolinate ligands of Cr(pic) 3
prime the redox potential of the chromic center for
reduction by biological reductants such as ascorbate or
thiols. The reduced Cr(II) reacts with oxygen, entering into
Haber–Weiss and Fenton chemistry [7]. Hydroxyl radicals
generated by this process are potent DNA-damaging
agents. This ability of Cr(pic) 3 to be reduced by biologically relevant reducing agents is consistent with studies
that indicate mutagenic forms of Cr(III) require aromatic,
bidentate ligands coordinated via pyridine-type nitrogens
[8]. Furthermore, Cr(pic) 3 is remarkably stable [7,9]; the
compound is unreactive towards water or common buffers,
does not transfer chromium to transferrin or albumin, and
appears to be unaffected by the presence of fatty acids or
cholesterols [9,10]. Only dissolution in mineral acid solutions (concentration 0.1 M or greater) seems to lead to loss
of picolinate ligands at an appreciable rate [10,11]. The
stability of Cr(pic) 3 may explain its unique ability to be
0277-5387 / 99 / $ – see front matter  1999 Elsevier Science Ltd. All rights reserved.
PII: S0277-5387( 99 )00166-7
J.K. Speetjens et al. / Polyhedron 18 (1999) 2617 – 2624
2618
Nomenclature
Pic
DOS
LMWCr
SOD
AMP
DNA
EDTA
E. coli
kB
LB
TBE
Tris
X-gal
picolinate
Density of states
low-molecular-weight chromium-binding
substance
superoxide dismutase
ampicillin
deoxyribonucleic acid
ethylenediamine tetraacetic acid
Escherichia coli
kilobase
Luria-Bertani
45 mM Tris–45 mM borate–1 mM
EDTA
tris(hydroxymethyl)aminomethane
5-bromo-4-chloro-3-indolyl-b-D-galactopyranoside
1 are much less able to effect this type of damage to DNA
under physiologically-relevant conditions.
2. Experimental
All manipulations were performed under aerobic conditions at room temperature, and all chemical were used as
received unless otherwise noted. Bovine liver LMWCr was
isolated and purified as previously described [16].
[Cr 3 O(O 2 CCH 2 CH 3 ) 6 (H 2 O) 3 ]NO 3 1 [19], [Cr 3 O(OAc) 6(H 2 O) 3 ]Cl [20], and Cr(pic) 3 ?H 2 O [21] were synthesized
as previously reported. pUC19 plasmid DNA was obtained
from Bayou Biolabs and was gel-purified and quantitated
by ultraviolet spectroscopy [22], using a Pharmacia Ultrospec 2000 spectrophotometer, prior to use in reactions.
Superoxide dismutase (SOD) and catalase were obtained
from Sigma.
2.1. DNA cleavage reactions
absorbed but ironically may allow the compound to
accumulate in cells in its DNA-damaging form.
The biologically-active, naturally-occurring form of
chromium, low-molecular-weight chromium-binding substance (LMWCr), activates the insulin-dependent tyrosine
kinase activity of insulin receptor (Kd |250 pM) [12–18]
and is probably responsible for the role of chromium in
carbohydrate and lipid metabolism. LMWCr is an
oligopeptide which binds four chromic ions in a multinuclear assembly, presumably bridged by oxo / hydroxo ligands and supported by carboxylate groups provided by the
oligopeptide. Given this composition, attempts have been
made to identify synthetic anion-bridged Cr(III)-carboxylate assemblies that model the spectroscopic properties
and mode of action of LMWCr. A synthetic, biomimetic
chromium(III) complex, [Cr 3 O(O 2 CCH 2 CH 3 ) 6 (H 2 O) 3 ] 1
1, has been identified which is biologically active, stable
under acidic conditions, and readily and inexpensively
synthesized [18]. Given its stability, it presumably could
be taken orally as a nutritional supplement or therapeutic,
in contrast to LMWCr which is susceptible to hydrolysis in
acidic solution (such as the environment of the stomach)
[16,17]. While LMWCr and compound 1 have been
proposed as potential nutritional supplements, it is important to determine whether they exhibit the same deleterious ability as Cr(pic) 3 to cleave DNA under physiological conditions. Herein are reported studies designed to
examine the DNA-cleaving abilities of LMWCr and 1 in
the presence of biological reductants such as ascorbate and
thiols or in the presence of hydrogen peroxide. The results
described herein indicate that LMWCr and the biomimetic
A typical reaction was carried out mixing aliquots of
pUC19 (in 5 mM Tris, 500 mM EDTA buffer (pH 8.0)),
reductant in H 2 O, LMWCr in 50 mM NH 4 OAc (pH 6.5)
or 1 in H 2 O, phosphate buffer (pH 7.0), and H 2 O to give a
final volume of 15 ml; the final phosphate concentration
was 50 mM. Inhibitors were added as solutions in H 2 O,
solutions in phosphate buffer, or as neat liquids. (Complex
1 has previously been shown to be stable in slightly basic
aqueous solution [18]). Reactions were allowed to proceed
180 min unless otherwise noted. All reactions were
quenched by addition of loading buffer (24% glycerol and
0.1% bromophenol blue). Seven microliter aliquots were
loaded directly onto a 1% agarose gel and electrophoresed
at 60 V. The gel was stained with ethidium bromide and
was photographed on a UV transilluminator. Any timed
assays were performed by adding the appropriate reagents
to the sample tubes, and the addition of DNA was marked
time zero. The reactions were quenched at various time
points by adding loading buffer. Reactions under argon
were performed by bubbling the gas through the reaction
mixture until all the reagents were added and then sealing
the container until the reaction was quenched by addition
of loading buffer.
2.2. Effect of nicked DNA on cell viability
A reaction mixture containing 10 mg of pUC19, 1.5 mM
1, and 0.90 mM H 2 O 2 in 10 mM Tris, pH 8.0 was allowed
to react for 3 h at 258C. Quench buffer was added, and the
sample was immediately loaded onto a 1% agarose gel
containing 0.5 mg ml 21 of ethidium bromide. After
electrophoresis in 13 TBE buffer, the band at 3.5 kB was
excised from the gel, and the DNA was removed from the
J.K. Speetjens et al. / Polyhedron 18 (1999) 2617 – 2624
gel using a Qiagen gel purification kit. The eluted DNA
was gel-quantitated and used without further modification
for transformation.
Electrocompetent DH5a E. coli cells were electroporated using 1 ml of 0.1 mg ml 21 of either trimer-nicked
DNA or supercoiled pUC19 plasmid DNA in 50 ml of
cells. The E. coli cells were rescued by adding 1 ml of
SOC media (20 g l 21 Bactotryptone, 5 g l 21 yeast extract,
9 mM NaCl, 20 mM glucose, pH 7.5), followed by
incubation at 378C for 1 h. Each cell suspension was
diluted 1:100 with SOC media, and 50 ml of each diluted
suspension was spread on separate LB-AMP agar plates
containing 80 mg ml 21 X-gal, 100 mM IPTG, and 100 mg
ml 21 ampicillin. The X-gal and IPTG were added to
facilitate detection of colonies. The plates were placed in a
378C incubator overnight. The transformation efficiency of
relaxed plasmid DNA is comparable to that of supercoiled
pUC19 plasmid DNA under these conditions [23].
2.3. Phosphate mono- and diester hydrolysis reactions
Phosphate mono- and diester hydrolysis reactions were
monitored on an HP8452 diode-array UV-Vis spectrophotometer at 258C, measuring the production of pnitrophenol at 404 nm. Multiple trials were performed for
each set of reactions. Background reactions of phosphate
mono- and diester compound alone, with H 2 O 2 , and with 1
were used to correct any observed activities. Typical
reaction mixtures contained 1.25 mM bis-p-nitrophenyl
phosphate or 2.5 mM p-nitrophenyl phosphate, 1.5 mM 1,
and 900 mM H 2 O 2 in freshly prepared, chelexed 0.010 M
Tris, pH 8.0. Rates were calculated using ´404 for pnitrophenol at pH 8.0 of 17 000 M 21 cm 21 .
2619
3. Results and discussion
3.1. DNA cleavage by complex 1
The reaction of trinuclear Cr(III) propionate complex, 1,
with pUC19 was monitored by observing the conversion of
the supercoiled plasmid DNA (faster migrating species) to
the circular, nicked form (slower migrating species). All
reactions were evaluated by comparison of the amount of
relaxed plasmid to the amount in gel-purified plasmid
controls. In the presence of the biological reductant
ascorbic acid (and atmospheric oxygen), plasmid DNA is
slowly nicked (Fig. 1, lane 2); this is a long known
phenomenon [24,25] and appears to result in part from the
production of hydroxyl radicals. The addition of 1 to the
mixture of ascorbate and DNA, regardless of concentration
(in the range examined: 0.040–120 mM), has no effect the
amount of relaxed DNA (Fig. 1, lanes 3–7). Given that 1
by itself has no effect on the rate of DNA cleavage (Fig. 2,
lane 2), the trimer appears to be ineffective in catalyzing
the cleavage of DNA in the presence of ascorbate and
oxygen. Similar results are obtained for 1 in the presence
of the reductant dithiothreitol: DTT under aerobic conditions produces some cleavage of the plasmid DNA while
addition of 1 does not increase the amount of cleavage (not
shown). The cleavage in the presence of ascorbate (and of
1) is time dependent (not shown) and is inhibited in the
absence of oxygen (Fig. 1, lane 9) the presence of the
radical traps EtOH and t-BuOH (Fig. 1, lanes 10 and 11),
and the presence of superoxide dismutase or catalase.
These results are consistent with the mechanism of DNA
cleaving involving the reduction of dioxygen by ascorbate
to give radical products capable of cleaving DNA.
These results are in stark contrast to those using Cr(pic) 3
Fig. 1. Cleavage of pUC19 by compound 1. The reactions contain pUC19 (39 mM in base pairs) and 50 mM phosphate buffer, pH 7.0; and: lane 2, 5 mM
ascorbate; lanes 3–7, 5 mM ascorbate and 120, 40, 4.0, 0.40, and 0.040 mM 1, respectively; lane 8, 5 mM ascorbate and 1.2 mM Cr(pic) 3 ; lane 9, 5 mM
ascorbate and 120 mM 1 under argon; lane 10, 5 mM ascorbate, 120 mM 1, and 1 M EtOH; lane 11, 5 mM ascorbate, 120 mM 1, and 1 M t-BuOH; lane 12,
5 mM ascorbate, 120 mM 1, and 100 mg ml 21 SOD; and lane 13, 5 mM ascorbate, 120 mM 1, and 100 mg ml 21 catalase. Reactions were allowed to
proceed 60 min before quenching.
2620
J.K. Speetjens et al. / Polyhedron 18 (1999) 2617 – 2624
Fig. 2. Cleavage of pUC19 by compound 1. The reactions contain pUC19 (39 mM in base pairs) and 50 mM phosphate buffer, pH 7.0; and: lane 2, 120
mM 1; lane 3, 215 mM H 2 O 2 ; lanes 4–8, 215 mM H 2 O 2 and 120, 40, 4, 0.4, and 0.04 mM 1, respectively; lane 9, 215 mM H 2 O 2 and 120 mM Cr(pic) 3 ; lane
10, 215 mM H 2 O 2 and 120 mM 1 under argon; lane 11, 215 mM H 2 O 2 , 120 mM 1, and 1 M EtOH; lane 12, 215 mM H 2 O 2 , 120 mM 1, and 1 M t-BuOH;
and lane 13, 215 mM H 2 O 2 , 120 mM 1, and 100 mg ml 21 SOD. Reactions were allowed to proceed 180 min before quenching by addition of loading
buffer.
as a chromium source [7]. As shown in Fig. 1 (lane 8),
Cr(pic) 3 (at a chromium concentration of 1.2 mM, equivalent to that of the trimer in lane 6) is quite effective in
generating relaxed DNA from supercoiled plasmid DNA in
the presence of 5 mM ascorbate. Significantly, these
concentrations of Cr(pic) 3 and ascorbate are physiologically relevant. It has been predicted that a person consuming
5.01 mg of Cr(pic) 3 , equivalent to three pills of some
common supplements, daily for 5 years would have a liver
Cr concentration of 13 mM [5], more than ten times the
Cr(pic) 3 concentration used in this experiment. Ascorbate
concentrations in most tissues, including liver, are in the
millimolar range [26]. The ability of Cr(pic) 3 to catalyze
cleavage of DNA under these conditions stems from the
presence of the picolinate ligands. The aromatic, bidentate
ligands coordinated via pyridine-type nitrogens modify the
redox potential of the central chromic ion such that it can
be reduced by ascorbate or thiols [7]; dioxygen oxidizes
the chromium center back to Cr(III) with concurrent
reduction of dioxygen. This reduced dioxygen-derived
species can undergo O–O bond cleavage, generating OH ?
which may diffuse to react with DNA. The resulting Cr
species may again be reduced and reenter the catalytic
cycle. Consequently, it is not surprising that Cr(pic) 3 has
been found to lead to DNA strand breaks in cultured
Chinese hamster ovary cells [4] and cultured macrophages
[27]. (However, it should be noted that a recent study
failed to find oxidative DNA damage using an antibody
titer to 5-hydroxymethyluracil for ten obese women taking
Cr(pic) 3 for 8 weeks [28]). The trinuclear 1 lacks any type
of nitrogen-based ligation and, thus, is not susceptible to
reduction by ascorbate or thiols. Correspondingly,
chromium acetate hydroxide (h[Cr 3 O(OAc) 6 ]OAcj x , a
polymeric array of oxo-centered trinuclear assemblies
which breaks apart to give the acetate analogue of 1) at
sub-cytotoxic concentrations has been found to fail to
produce DNA strand breaks in peripheral lymphocytes
[29]. The acetate analogue of 1, [Cr 3 O(OAc) 6 (H 2 O) 3 ] 1 ,
has previously been shown to be unable to catalyze DNA
cleavage in the presence of ascorbate [23], consistent with
the results of this study on complex 1 and of Ref. [29].
Previously these laboratories have shown that the trinuclear complex [Cr 3 O(OAc) 6 (H 2 O) 3 ] 1 was capable of
relaxing plasmid DNA in a time- and concentration-dependent fashion in the presence of large concentrations of
hydrogen peroxide [23]; this cleavage appeared to result
from the formation of hydroxyl radical from peroxide
catalyzed by the trinuclear species. Given the strong
similarities between this compound and its propionate
analogue, 1, the ability of complex 1 to cleave DNA in the
presence of peroxide was probed. As shown in lanes 4–8
of Fig. 2, the trimer catalyzes the cleavage of DNA in the
presence of peroxide in a concentration dependent fashion,
although the reaction is extremely slow. The reaction
catalyzed by the acetate analogue of 1 and peroxide in pH
7.0 phosphate buffer results in a degree of cleavage
indistinguishable from produced by complex 1 (not
shown). The small degree of cleavage obtained in these
experiments required a 3-h reaction time. The cleavage of
DNA is also time dependent (not shown). The rate of
cleavage is not affected by removing atmospheric oxygen
but is inhibited by the radical traps EtOH and t-BuOH. The
mechanism for this cleavage probably corresponds to that
proposed previously for the acetate analogue. Hydrogen
peroxide presumably displaces a carboxylate ligand and
bridges between two chromic centers. Homolytic cleavage
of the hydrogen peroxide, catalyzed by the trimers,
produces two equivalents of hydroxyl radical. The OH ?
species diffuse to react with DNA. SOD also has an
inhibitory effect on the reaction, suggesting that superox-
J.K. Speetjens et al. / Polyhedron 18 (1999) 2617 – 2624
ide produced from peroxide may also play a role in the
cleavage reactions. Previously, reactions with the acetate
analogue in Tris buffer (rather than phosphate buffer) have
failed to detect any inhibition from SOD [23]; the Tris may
serve as a suitable trap for superoxide radicals, such that
this side pathway to radical-based cleavage was not
observed.
Cr(pic) 3 has been shown to generate hydroxyl radicals
from hydrogen peroxide in a pathway independent of
added reductant [7]. This pathway is, however, orders of
magnitude less efficient than that in the presence of
reductant, although the rate of cleavage catalyzed by the
nutritional supplement still far exceeds that of trinuclear
complex 1 under identical conditions. This is readily
demonstrated in Fig. 2 (lane 9); in the presence of 215 mM
peroxide, 120 mM Cr(pic) 3 gives rise to a much greater
quantity of relaxed DNA than does 1 at equivalent or
higher Cr concentrations (lanes 5 and 4, respectively).
Generation of hydroxyl radicals from hydrogen peroxide in
a pathway independent of added reductant has also been
observed for other mononuclear Cr(III) complexes
[30,31,32]; however, the efficiency of such pathways are
roughly equivalent to that of Cr(pic) 3 and probably not
responsible for the mutagenic effects of Cr(III), which
require added reductants [8,33].
The different chemistry observed with the trinuclear
complex and with Cr(pic) 3 is readily borne out by their
redox potentials. Oxo-centered trinuclear chromic carboxylates are electrochemically inactive in water [34], such
that they are far too difficult to reduce for ascorbate or
thiols to serve as reducing agents. In contrast, previous
spectroscopic and electrochemical studies place the redox
potential for the metal-centered reduction of Cr(pic) 3 [35]
in the range of those for chromium(III) complexes known
2621
to give rise to oxygen radical-mediated DNA damage, as
well as in the range of the redox potential of biological
reductants such as ascorbate, thiols, and NADH [8].
3.2. DNA cleavage by LMWCr
Spectroscopic studies on LMWCr suggest that its
chromic centers possess primarily, if not solely, oxygenbased ligands [16]; consequently, LMWCr is not expected
to catalyze DNA cleavage in the presence of mild reductants. This is also born out by the technique used to
remove chromium from the oligopeptide to produce the
apo-oligopeptide [16]. Chromium is removed by chelation
with EDTA at acidic pH values at ¯608C; however, for the
chromium to be removed, it must be reduced to the
chromous state. Ascorbate and thiols are ineffective;
cyanoborohydride is required for the reduction.
As shown in Fig. 3 (lanes 3–7), LMWCr in the presence
of 5 mM ascorbate actually does catalyze the cleavage of
DNA in a concentration-dependent fashion to a detectable
extent; however, the degree of cleavage is all but insignificant in comparison with that of Cr(pic) 3 at an
equivalent chromium concentration of 1.2 mM (Fig. 3,
lanes 6 and 8). An effort was made to examine the
inhibition of the LMWCr-catalyzed cleavage by a variety
of inhibitors but the amount of catalyzed cleavage is of the
same order of magnitude of cleavage by ascorbate itself;
consequently, the effects of the inhibitors on the cleavage
reaction inititated by ascorbate and oxygen could not be
separated from effects on the inhibition of the LMWCrcatalyzed process. Cleavage of DNA by hydrogen peroxide
is also catalyzed by LMWCr but to an extent even less
than that of complex 1 (Fig. 4, lanes 4–8). An appreciable
amount of cleavage can be observed only at the highest
Fig. 3. Cleavage of pUC19 by LMWCr. The reactions contain pUC19 (39 mM in base pairs) and 50 mM phosphate buffer, pH 7.0; and: lane 2, 5 mM
ascorbate; lane 3, 90 mM LMWCr and 5 mM ascorbate; lane 4, 30 mM LMWCr and 5 mM ascorbate; lane 5, 3.0 mM LMWCr and 5 mM ascorbate; lane 6,
0.30 mM LMWCr and 5 mM ascorbate; lane 7, 0.030 mM LMWCr and 5 mM ascorbate; lane 8, 5 mM ascorbate and 1.2 mM Cr(pic) 3 ; lane 9, 90 mM
LMWCr and 5 mM ascorbate under argon; lane 10, 90 mM LMWCr, 5 mM ascorbate, and 1 M EtOH; lane 11, 90 mM LMWCr, 5 mM ascorbate, and 1 M
t-BuOH; lane 12, 90 mM LMWCr, 5 mM ascorbate, and 100 mg ml 21 SOD; lane 13, 90 mM LMWCr, 5 mM ascorbate, and 100 mg ml 21 catalase.
Reactions were allowed to proceed 60 min before quenching by addition of loading buffer.
2622
J.K. Speetjens et al. / Polyhedron 18 (1999) 2617 – 2624
Fig. 4. Cleavage of pUC19 by LMWCr. The reactions contain pUC19 (39 mM in base pairs) and 50 mM phosphate buffer, pH 7.0; and: lane 2, 90 mM
LMWCr; lane 3, 215 mM H 2 O 2 ; lane 4, 90 mM LMWCr and 215 mM H 2 O 2 ; lane 5, 30 mM LMWCr and 215 mM H 2 O 2 ; lane 6, 3.0 mM LMWCr and 215
mM H 2 O 2 ; lane 7, 0.30 mM LMWCr and 215 mM H 2 O 2 ; lane 8, 0.03 mM LMWCr and 215 mM H 2 O 2 ; lane 9, 120 mM Cr(pic) 3 and 215 mM H 2 O 2 ; lane
10, 90 mM LMWCr and 215 mM H 2 O 2 under argon; lane 11, 90 mM LMWCr, 215 mM H 2 O 2 , and 1 M EtOH; lane 12, 90 mM LMWCr, 215 mM H 2 O 2 ,
and 1 M t-BuOH; lane 13, 90 mM LMWCr, 215 mM H 2 O 2 , and 100 mg ml 21 SOD. Reactions were allowed to proceed 180 min before quenching by
addition of loading buffer.
LMWCr concentration examined (90 mM, 360 mM Cr).
This is in stark contrast to Cr(pic) 3 (Fig. 4, lane 9) where
at a three-fold lower concentration (in terms of chromium)
nearly all the plasmid DNA is nicked. Oxygen is not
required for the cleavage in the presence of peroxide, while
the cleavage is apparently inhibited to some degree by
EtOH, t-BuOH, and SOD (Fig. 4, lanes 10–13), although
the low levels make this difficult to determine.
shown that transformation of cells with nicked DNA,
produced by prolonged storage of pUC19, results in an
appreciable reduction of colonies present [23]). Thus, the
E. coli DNA repair enzymes could repair the damage
produced by incubation with the trimer and peroxide.
Consequently, the likelihood that the use of complex 1 as a
nutritional agent would lead to irreversible DNA cell
damage would appear remote.
3.3. Effect of DNA nicking on cell viability
3.4. Phosphate ester cleavage reactions
For the cleavage of DNA to have serious deleterious
effects on the cells containing the damage, the damage
must not be readily repairable. The transformation of E.
coli cells with plasmid DNA, before and after incubation
with complex 1, provides a convenient and accurate
measure of whether complex-mediated DNA damage
(although generated at peroxide concentrations far above
physiological levels) was irreversible. As the plasmid
carries an antibiotic resistance gene marker for ampicillin,
only E. coli cells that possess intact plasmid DNA will be
able to grow on media plates containing AMP. Complex
1-nicked DNA was generated by incubating the DNA for 3
h with complex and peroxide. The nicked band was
excised from the gel, gel-quantitated, and used to transform E. coli cells; additionally an equal concentration of
supercoiled pUC19 was used to transform cells. The
transformed cells were then plated on LB plates containing
ampicillin. As shown in Fig. 5, the plate with cells
transformed with nicked DNA contained the same number
of colonies (within the relative error of the gel-quantitation
method) as the positive control. (It has previously been
Previously [Cr 3 O(OAc) 6 (H 2 O) 3 ] 1 has been shown to
hydrolyze both phosphate mono- and diesters in the
presence of hydrogen peroxide while the complex or
peroxide by itself were ineffective [23]. The cleavage was
proposed to be performed by Cr(IV)-hydroxide species (in
equilibrium with Cr(III)-hydroxyl radical species) produced by the homolytic cleavage of peroxide catalyzed by
the trinuclear complex. Consequently, the ability of complex 1 to hydrolyze the phosphate monoester p-nitrophenyl
phosphate and phosphate diester bis-p-nitrophenylphosphate was investigated. Complex 1 does hydrolyze both
phosphate ester compounds with similar rate constants in
the presence of peroxide (Fig. 6). The rate is approximately one-fourth order in phosphate ester, is one-half order in
peroxide, and has a complex dependence on the trimer
concentration.
LMWCr has previously been shown not to possess
phosphatase activity [15]. As the oligopeptide is susceptible to hydrolysis [16,17], no attempt was made to probe
the ability of LMWCr to catalyze the hydrolysis of
phosphate esters in the presence of peroxide.
J.K. Speetjens et al. / Polyhedron 18 (1999) 2617 – 2624
2623
Fig. 5. LB-AMP plates of DH5a E. coli colonies transformed with (A) complex 1-nicked pUC19 and (B) supercoiled pUC19. The plates were
supplemented with 30 mg ml 21 of X-gal and 100 mM IPTG in order to better visualize the colonies.
tibility to reduction and subsequent oxidation by dioxygen.
In this regard, multinuclear oxo(hydroxo)-bridged
chromium carboxylate assemblies such as LMWCr and
[Cr 3 O(O 2 CCH 2 CH 3 ) 6 (H 2 O) 3 ] 1 would appear to be better
candidates as nutritional supplements than Cr(pic) 3 . However, data on the long term effects of supplementation with
LMWCr or complex 1 in animal models and humans are
required before any beneficial claims can be made. Studies
of the effect on LMWCr and trimer 1 given intraperitoneally to rats on body weight, fat content, organ size and
chromium content, and blood plasma variables (such as
glucose, triglyceride, total cholesterol, and insulin concentrations) are currently in progress.
Acknowledgements
Fig. 6. Hydrolysis of bis-p-nitrophenyl phosphate catalyzed by complex
1. Reactions were conducted in 0.010 M Tris, pH 8.0 at 258C. The
reaction (♦) contained 2.5 mM bis-p-nitrophenyl phosphate and buffer;
reaction (^) contained 2.5 mM bis-p-nitrophenyl phosphate, buffer, and
900 mM H 2 O 2 ; reaction (j) contained 2.5 mM bis-p-nitrophenyl
phosphate, 1.5 mM complex, and buffer; and reaction (d) contained 2.5
mM bis-p-nitrophenyl phosphate, 1.5 mM complex, 900 mM H 2 O 2 , and
buffer.
4. Conclusions
The results of this study suggest that care should be used
in the choice of chromium source used as a nutritional
supplement or therapeutic, especially in terms of suscep-
Funding was provided by the NRICGP/ USDA to J.B.V.
Yanjie Sun and Kavita Mallya provided purified LMWCr.
References
[1] R.A. Anderson, in: W. Mertz (Ed.), 5th ed, Trace Elements in
Human and Animal Nutrition, Vol. 1, Academic Press, New York,
1987, p. 225.
[2] R.A. Anderson, A.S. Kozlovsky, Am. J. Clin. Nutr. 41 (1985) 1177.
[3] R.A. Anderson, N.A. Bryden, M.M. Polansky, J. Am. Coll. Nutr. 16
(1997) 273.
[4] D.M. Stearns, J.P. Wise Jr., S.R. Patierno, K.E. Wetterhahn, FASEB
J. 9 (1995) 1643.
2624
J.K. Speetjens et al. / Polyhedron 18 (1999) 2617 – 2624
[5] D.M. Stearns, J.J. Belbruno, K.E. Wetterhahn, FASEB J. 9 (1995)
1650.
[6] J. Cerulli, D.W. Grabbe, M. Malone, M.D. Goldrick, Ann. Pharmacother. 32 (1998) 428.
[7] J.K. Speetjens, R.A. Collins, J.B. Vincent, S.A. Woski, Chem. Res.
Toxicol. 12 (1999) 483.
[8] K.D. Sugden, R.D. Geer, S.J. Rogers, Biochemistry 31 (1992)
11626.
[9] K.F. Kingry, A.C. Royer, J.B. Vincent, J. Inorg. Biochem. 72 (1998)
79.
[10] N.E. Chakov, R.A. Collins, J.B. Vincent, submitted for publication.
[11] D.M. Stearns, W.H. Armstrong, Inorg. Chem. 31 (1992) 5178.
[12] J.B. Vincent, J. Am. Coll. Nutr. 18 (1999) 6.
[13] C.M. Davis, J.B. Vincent, J. Biol. Inorg. Chem. 2 (1998) 675.
[14] C.M. Davis, J.B. Vincent, Biochemistry 36 (1997) 4382.
[15] C.M. Davis, K.H. Sumrall, J.B. Vincent, Biochemistry 35 (1996)
12963.
[16] C.M. Davis, J.B. Vincent, Arch. Biochim. Biophys. 399 (1997) 335.
[17] K.H. Sumrall, J.B. Vincent, Polyhedron 16 (1998) 4171.
[18] C.M. Davis, A.C. Royer, J.B. Vincent, Inorg. Chem. 36 (1997) 5316.
[19] A. Earnshaw, B.N. Figgis, J. Lewis, J. Chem. Soc. A (1966) 1656.
[20] M.K. Johnson, D.B. Powell, R.D. Cannon, Spectrochim. Acta 37A
(1981) 995.
[21] R. Press, J. Geller, G.W. Evans, West. J. Med. 152 (1990) 41.
[22] J. Sambrook, E.F. Fritsch, B. Maniatis (Eds.), Molecular Cloning: A
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
Laboratory Manual, Vol. 1, Cold Spring Harbor Press, Plainview,
NY, 1989.
A. Parand, A.C. Royer, T.L. Cantrell, M. Weitzel, N. Memon, J.B.
Vincent, M.W. Crowder, Inorg. Chim. Acta 268 (1998) 211.
A.R. Morgan, R.L. Cone, T.M. Elgert, Nucleic Acid Res. 3 (1976)
1139.
V.C. Bode, J. Mol. Biol. 26 (1967) 125.
D. Hornig, Ann. NY Acad. Sci. 258 (1975) 103.
D. Bagchi, M. Bagchi, J. Balmoori, X. Ye, S.J. Stohs, Res. Commun.
Mol. Pathol. Pharmacol. 3 (1997) 335.
I. Kato, J.H. Vogelman, V. Dilman, J. Karkoszka, K. Frenkel, N.P.
Durr, N. Orentreich, P. Toniolo, Eur. J. Epidemiol. 14 (1998) 621.
M. Gao, S.P. Binks, J.K. Chipman, L.S. Levy, R.A. Braithwaite, S.S.
Brown, Hum. Exp. Toxicol. 11 (1992) 77.
X. Shi, N.S. Dalal, K.S. Kasprzak, Arch. Biochem. Biophys. 302
(1993) 294.
T.-C. Tsou, C.-L. Chen, T.-Y. Liu, J.L. Yang, Carcinogenesis 17
(1996) 103.
T.-C. Tsou, J.-L. Yang, Chem.-Biol. Interact. 102 (1996) 133.
E.A. Hassoun, S.J. Stohs, J. Biochem. Toxicol. 10 (1995) 315.
R.D. Cannor, L. Montri, D.B. Brown, K.M. Marshall, C.M. Elliot, J.
Am. Chem. Soc. 106 (1984) 2591.
C.K. Vinayakumar, G.R. Dey, K. Kishore, P.N. Moorthy, Radiat.
Phys. Chem. 48 (1996) 737.
Journal of Inorganic Biochemistry 89 (2002) 272–278
www.elsevier.com / locate / jinorgbio
The fate of the biomimetic cation
triaqua-m-oxohexapropionatotrichromium(III) in rats
Amanda A. Shute, John B. Vincent*
Department of Chemistry and Coalition for Biomolecular Products, The University of Alabama, Tuscaloosa, AL 35487 -0336, USA
Received 13 September 2001; received in revised form 16 January 2002; accepted 17 January 2002
Abstract
The synthetic biomimetic triaqua-m-oxohexapropionatotrichromium(III) nitrate when given intravenously to rats lowers fasting blood
plasma triglycerides and cholesterol concentrations; thus, the cation has the potential to serve as a therapeutic agent. Its ability to function
in vivo presumably is dependent on its ability to mimic the action of the natural, bioactive, chromium-binding oligopeptide chromodulin
in stimulating insulin receptor kinase activity. Consequently, the cation should be incorporated into insulin-sensitive cells intact. Thus, the
fate of the 51 Cr-labeled complex during the first 24 h after injection in tissues, blood, urine, and feces was followed. The complex appears
to be readily incorporated into tissues and cells. In hepatocytes, the cation is efficiently transported into microsomes where its
concentration reaches a maximum in approximately 2 h.  2002 Elsevier Science Inc. All rights reserved.
Keywords: Chromium; Chromodulin; Propionate; Rat
1. Introduction
The biomimetic cation, [Cr 3 O(O 2 CCH 2 CH 3 ) 6 (H 2 O) 3 ] 1
(Fig. 1), has been found to have striking in vivo effects,
lowering plasma triglycerides, total cholesterol, and LDL
cholesterol levels after 12 weeks of supplementation in
healthy rats at a level of 20 mg Cr per kg body mass daily
[1]. Recently the cation has also been found to lower
triglycerides and cholesterol concentrations and insulin
concentrations and to increase insulin sensitivity in type II
diabetic model rats [2]. The ability presumably results
from the cation’s ability to mimic the naturally occurring,
bioactive oligopeptide chromodulin; the propionate trimer
activates the insulin-dependent kinase activity of insulin
receptor in vitro in a fashion very similar to chromodulin.
The kinase activity of an isolated active site fragment of
the receptor, for example, is stimulated approximately
threefold with a dissociation constant of 1.00 nM [3]. The
oligopeptide chromodulin (also known as low-molecularweight chromium-binding substance or LMWCr) has been
proposed recently to be the biologically active form of
chromium in mammals [4–7]. The naturally occurring
oligopeptide may function as part of a unique autoamplifi*Corresponding author. Tel.: 11-205-3489-203; fax: 11-205-3489104.
E-mail address: [email protected] (J.B. Vincent).
cation system for insulin signaling [4,5]. In this mechanism, apochromodulin is stored in insulin-sensitive cells.
In response to increases in blood insulin concentrations (as
would result from increasing blood sugar concentrations
after a meal), insulin binds to its receptor bringing about a
Fig. 1. Structure of chromium ‘‘basic carboxylate’’ cations. For compound 1, R5CH 2 CH 3 and L5H 2 O.
0162-0134 / 02 / $ – see front matter  2002 Elsevier Science Inc. All rights reserved.
PII: S0162-0134( 02 )00382-3
A. A. Shute, J.B. Vincent / Journal of Inorganic Biochemistry 89 (2002) 272 – 278
conformation change which results in the autophosphorylation of tyrosine residues on the internal side of the
receptor. This transforms the receptor into an active
tyrosine kinase and transmits the signal from insulin into
the cell. In response to insulin, chromium is moved from
the blood to insulin-sensitive cells. Here, the chromium
flux results in the loading of apochromodulin with
chromium. The holochromodulin then binds to the receptor, presumably assisting to maintain the receptor in its
active conformation, amplifying its kinase activity.
The propionate complex possesses other attributes that
make it an excellent candidate for a therapeutic agent. It is
very soluble in water and stable in dilute mineral acid (it
can be recrystallized from dilute acid) such that it might
survive conditions in the stomach. No acute toxic effects
were observed upon treatment of rats with the compound,
and it does not give rise to oxidative cell damage in in
vitro or in vivo studies as observed with the popular
nutritional supplement chromium picolinate [8,9].
However, the stability of the trinuclear propionate
complex in vivo is not known. As the complex is biologically active in vivo, it presumably stays intact in vivo
for some period of time. A recent study in which rats were
treated daily with the complex by intravenous injection for
2 weeks revealed that the complex had an in vivo lifetime
of less than 24 h but appeared to enter cells intact [10].
The fate of the complex for the first 24 h after injection
was, thus, examined in the current study.
2. Materials and methods
51
CrCl 3 in 1.0 M HCl was obtained from ICN.
[ Cr 3 O(O 2 CCH 2 CH 3 ) 6 (H 2 O) 3 ]NO 3 , compound 1, was
made by the method of Earnshaw et al. [11] using a tracer
amount of labeled CrCl 3 . Male Sprague–Dawley rats
(between 650 and 750 g) were obtained from Charles
River Laboratories; rats were housed at least 1 week after
arrival before use. The rats were injected in the tail vein
with 150 ml of an aqueous solution of 1 containing 276 mg
of the cation and placed into metabolic cages for collection
of urine and feces. After appropriate time intervals (from
30 min to 24 h), the rats were sacrificed by carbon dioxide
asphyxiation, and tissue samples were harvested and
weighed. Subcellular liver fractions were obtained by
differential centrifugation according to established procedures [12,13]. Tissues were diced, and the pieces were
rinsed with 0.25 M sucrose solution. All subsequent steps
were performed using the same solution. The pieces were
ground in a tissue grinder. The nuclear, mitochondrial,
lysosomal, and microsomal fractions were obtained as
pellets from centrifugation at 30, 3300, 25 000, and
100 000 g, respectively. Components of the urine and
nuclear or microsomal hepatocyte fractions were separated
by G-15 column chromatography (44 cm35.1 cm; 1 ml /
min flow rate) or Shodex OH-PAK high-performance
51
273
liquid chromatography (HPLC; 30038 mm; |0.5 ml / min
flow rate). Before being applied to chromatography columns, the nuclear or microsomal fractions were suspended
(in the buffer to be used in the following chromatography
step), homogenized in a tissue grinder to disrupt their
membranes and release their contents, and then clarified by
centrifugation. Chromatography columns were run with 50
mM NH 4 OAc buffer, pH 6.5; 0.1 M NaCl was used as the
mobile phase for HPLC experiments. Experiments performed in duplicate.
Ultraviolet–visible measurements were made on a Hewlett-Packard 8453 spectrophotometer. Gamma counting
was performed on a Packard Cobra II auto-gamma counter.
All procedures were performed using doubly deionized
water unless otherwise noted and performed with plasticware whenever possible. The University of Alabama
Institutional Animal Use and Care Committee approved all
procedures involving the use of rats.
3. Results and discussion
In order to determine the stability and distribution of
biomimetic 1 in rats, rats were treated with 51 Cr-labeled 1,
allowing the fate of the cation or chromium from its
decomposition to be followed. The complex was administered by injection to allow the amount entering the
bloodstream to be known and avoiding other problems
potentially associated with absorption from the gastrointestinal tract. The amount of cation injected (276 mg) is six
times greater than that used previously in studies to
examine the effect of the cation in rats [1,2] but is the
same as that used to follow the fate of the biomimetic after
daily injections for 2 weeks [10]. The amount of Cr
received daily corresponds to 64 mg of chromium. This
quantity was chosen to increase the chances of observing
intact cation; if this large quantity can readily be degraded,
then smaller amounts, corresponding to normal human
dietary supplementation, should also readily be degraded.
For comparison, human nutritional supplements generally
provide 200 to 600 mg of Cr per day. Assuming an average
human body mass of 75 kg, this is equivalent to 1.6 to 4.8
mg of Cr per day for a 0.6 kg rat. Thus, the rats are
receiving |10- to 40-fold more Cr than if given an amount
comparable to humans. Because approximately 0.5–5% of
dietary chromium is absorbed (depending on the form of
Cr), the injection given the rats represents an even greater
excess of chromium when compared to that given humans
orally. Anderson and coworkers have fed rats diets containing up to 15 mg Cr per kg body mass for 24 weeks
without observing any acute toxic effects [15]. Assuming
5% absorption, this would correspond to 450 mg of Cr
entering the bloodstream of a 0.6 kg rat; thus, the dosage
of Cr in the current work was not expected to have acute
toxic effects. No ill effects were observed from the
injections; all tissues appeared normal when harvested.
274
A. A. Shute, J.B. Vincent / Journal of Inorganic Biochemistry 89 (2002) 272 – 278
Generation of Cr deficiency in rats is extremely difficult,
requiring strict environmental control (such as preventing
access to stainless steel) and a high sugar diet that
stimulates urinary chromium loss [4,16,17]. Cr deficiency
is not a concern for the rats in this study as the animals
consumed a commercial rat chow and had access to
stainless steel (e.g., cage and water bottle components).
Urine samples collected over the time intervals from 30
min to 24 h after injection of 51 Cr-labeled trimer reveal a
distinct pattern of Cr excretion (Fig. 2). The rate of loss of
chromium is maximal the first 2 h after injection and
significantly reduced the next 22 h. Fecal Cr loss is small
at all times and may be nearly constant with time. Urinary
chromium greatly exceeds fecal loss by up to approximately 10-fold. Urinary and fecal Cr loss corresponded to
approximately 15 and 1.5%, respectively, of the quantity
of chromium injected. The rapid urinary chromium loss
during the initial 2 h after injection (from the initial
injection corresponding to just under 130 000 cpm 51 Cr to
|5000 cpm after 120 min) may suggest (assuming the
isotope is evenly distributed in the blood during this time)
that the urinary excretion pathway is saturated. To determine the form(s) of chromium lost in the urine, urine
from representative time intervals was subjected to Shodex
OH-PAK HPLC (Fig. 3); this type of size-exclusion
chromatography revealed little change in the identity of the
Cr-containing species in urine occurred with time. After 2
h by which time most of the urinary chromium loss had
occurred, Cr eluted at the same molecular weight as
complex 1 (Fig. 3), indicating that the intact cation can
pass from the blood to the urine and indicating that the
complex does not need to degrade for chromium to reach
Fig. 2. Total urinary and fecal 51 Cr loss with time after injection of
51
Cr-labeled compound 1. Error bars represent standard deviation.
Fig. 3. Elution profile from Shodex OH-PAK HPLC column of rat urine
sample from h after injection of 51 Cr-labeled compound 1. Urine from the
first 2 h after injection is represented by the dotted line; urine from the
first 24 h, by the dashed line. The elution profile of 1 (solid line) is shown
for comparison.
the urine. After 24 h, most of the 51 Cr eluted in a similar
fashion as after 2 h. A small feature appears at an earlier
time and, thus, does not correspond to the molecular
weight of the cation. The band may correspond to
chromodulin (see below). The shape of the elution profile
for the 24-h urine sample is broader and less distinct in
shape, suggesting that small products from the breakdown
of 1 may be present. These results are in contrast with
those of other studies that have shown that chromium from
other sources including Cr-transferrin is ultimately lost in
the urine bound to chromodulin [14,18]; for these complexes, Cr is lost from the original complex. The ‘‘free’’ Cr
can then bind to chromodulin and subsequently be lost in
the urine. Thus, when Cr complexes degrade and release
their chromic ions, the Cr is lost as chromodulin. For
complex 1, which is more stable in vivo, this route is not
operative to a great degree in the first 24 h. The results of
the current study contrast with the results of an earlier
study in which rats were given compound 1 daily for 14
days. Urine samples a few to several days after the start of
administration of the biomimetic complex possesses only
one chromium-containing compound, chromodulin [10].
This indicates that while the first approximately one-sixth
A. A. Shute, J.B. Vincent / Journal of Inorganic Biochemistry 89 (2002) 272 – 278
of chromium injected as 1 is lost rapidly in the urine that
chromium from the breakdown of the |80% of the
complex still in the tissues is lost as chromodulin.
Fecal chromium loss presumably comes from chromium
in the bile. Previous work has shown that only minor
amounts of chromium are lost in the bile (for example
Refs. [19] and [20]). In bile, chromic ions occur as part of
a low-molecular-weight organic complex [21]. The molecular weight of this species unfortunately was not determined, although the authors postulated that this complex
might be involved in passage of chromium from the liver
to the bile. The low concentration of chromium in the feces
prevented further studies on the form of the chromium in
the current work. The form of chromium in bile of rats
treated with various forms of chromium needs to be
determined.
The 51 Cr content of the blood decreases rapidly with
time for the first 30 min after intravenous administration,
after which the Cr content of the blood appears to remain
constant (Fig. 4). As only 15% of the Cr enters the urine
and kidneys, the trinuclear cation appears to enter tissues
extremely rapidly.
The time dependence of the distribution of radiolabeled
chromium in the tissues is shown in Fig. 5. Of the tissues
examined, the greatest total organ quantity of chromium at
every time interval was located in the liver, followed by
the kidneys (except at 12 h when the content of the
epididymal fat is equivalent to that of the kidneys). Only
small amounts of the label were found at any time in the
Fig. 4. Blood 51 Cr content with time after injection of
compound 1. Error bars represent standard deviation.
51
Cr-labeled
Fig. 5. Total 51 Cr content of tissues with time after injection of
labeled compound 1.
275
51
Cr-
heart, testes, spleen, and pancreas. Intermediate amounts
were found in the epididymal fat. The distribution of
chromium is similar to that when other forms of chromium
have been given to rats [14,21–23]. The content after 24 h
is similar to that seen after 2 weeks of daily injection of
the trimer to rats with the exception of the amount in the
spleen being larger in the earlier study [10]; this suggests
the content of the spleen may result as an accumulation
over multiple injections. Notably the Cr content of the
tissues is greatest 1 h after injection with the single
exception of the epididymal fat (where the chromium
content is not significantly different throughout the time
intervals). This indicates that Cr rapidly moves from the
blood into the tissues where it is subsequently lost. This
movement from tissues corresponds nicely with the blood
and urine data. Cr from the trimer was found to rapidly
leave the blood stream. Most of the chromium enters the
tissues within 1 h of injection. After incorporation into
tissues, Cr from the trimer appear to be transferred into the
oligopeptide chromodulin and then lost in the urine where
loss is maximal from 60 to 120 min after injection.
As the amount of 51 Cr in the liver was greater than in
any other tissue, liver tissue was examined further. Examination of the subcellular distribution of Cr in hepatocytes
as a function of time (Fig. 6) reveals a surprising result.
The distribution of Cr after 2 h is quite distinct from than
found after the same period of time in studies using other
276
A. A. Shute, J.B. Vincent / Journal of Inorganic Biochemistry 89 (2002) 272 – 278
Fig. 6. Subcellular distribution of 51 Cr in hepatocytes with time after
injection of 51 Cr-labeled compound 1. Error bars represent standard
deviation.
forms of Cr [14,21]. Cr when injected in similar quantities
as CrCl 3 or Cr-transferrin is generally distributed from
roughly evenly between the nuclear and soluble fractions
of the cells to three times more chromium being present in
the nuclear fraction, with little Cr found in other fractions
[14,21]. (Cr from CrCl 3 is bound tightly by transferrin in
the blood plasma [14,22,24–26], such that both sources are
probably equivalent). If the trinuclear complex were
decomposing in the blood, then the released chromic ions
would be expected to be bound readily by transferrin,
resulting in a subcellular distribution similar to that in
earlier studies. This strongly suggests that the trimer enters
the cells intact and remains intact for some period of time
resulting in the different subcellular distribution. Remarkably, the microsomal fraction contains greater than .90%
of the total chromium in the hepatocytes 2 h after the
injections. This strongly suggests that the trimer is specifically transported into these organelles. The preponderance of the trimer in microsomes may suggest that the
trimer binds to a membrane or membrane-associated
protein and undergoes endocytosis. Six to 12 h after
injection, the Cr disappears from the microsomes; this is
mirrored by a loss of chromium from the whole liver
during this same time period. This suggests that the trimer
in the microsomal fraction may degrade during this time
interval, with the chromium-containing products being lost
from the liver.
After 2 weeks of daily administration of the trimer,
chromium from the trimer 24 h after the final injection is
located primarily in the nuclear and mitochondrial fraction
of the hepatocytes [10]. The results of the current study are
in part consistent with this as after 24 h more Cr is found
in the nuclei than in other subcellular components. A
corresponding increase in mitochondrial Cr is not observed
suggesting than Cr may accumulate in mitochondria only
after multiple injections of compound 1.
In order to elucidate which species in the liver might be
binding chromium or whether the trinuclear cation was
maintaining its integrity, the microsomal fraction 2 h after
injection was applied to a G-15 size-exclusion column.
The elution profile (Fig. 7) indicated that all the chromium
eluted in two bands. Application of an aqueous solution of
compound 1 to the column indicated that essentially all
(.90%) the 51 Cr in the microsomes elutes at fractions
corresponding exactly to that of the trimer solution. The
trinuclear complex (as with all of the trinuclear complexes)
is in equilibrium in solution with one or more dinuclear
complexes [27]; the smaller feature, which elutes after the
major band and corresponds to a smaller molecular weight
species, probably represents one or more of these dinuclear
Cr propionate complexes. However, this elution profile
clearly establishes that compound 1 after injection into the
blood stream can enter cells and microsomes intact. Thus,
the primarily route of movement of the trimer and
Fig. 7. Elution profile of microsomal fraction of hepatocytes (2 h after
injection of 51 Cr-labeled compound 1) and aqueous solution of 51 Crlabeled compound 1 from the G-15 column. The cpm from the aqueous
solution of compound 1 have been scaled for comparison purposes.
A. A. Shute, J.B. Vincent / Journal of Inorganic Biochemistry 89 (2002) 272 – 278
Fig. 8. Elution profile of nuclear fraction of hepatocytes (24 h after
injection of 51 Cr-labeled compound 1) from the G-15 column. The counts
of 51 Cr have been divided by 6520 for scaling purposes.
chromium in the rat can readily be followed. Within 1 h of
injection, the majority of compound 1 introduced into the
blood stream enters tissues intact. In at least the liver and
probably most other tissues (based the similarity of the
time course of buildup and loss of chromium to that of the
liver), the intact trimer is found in the microsomes of cells.
During the next hour the majority of the complex is rapidly
degraded.
After 24 h, the largest portion of the Cr left in the
hepatocytes is located in the nuclei. Separation of the
components of the nuclei on a G-15 column reveals that
the majority of Cr elutes earlier than compound 1, although
there are appreciable counts from 51 Cr in the later fractions
from the column (Fig. 8). The resolution of the lowermolecular-weight species is too poor to allow for determination if any of the intact trinuclear cation is present.
4. Conclusion
The previous observation that the synthetic compound
degraded in less than 24 h in vivo [10] raised the question
of whether studies that examined the in vivo effects of the
complex on blood variables in rats were observing effects
277
from the intact trimer or from chromic ions and / or
propionate. Given the tissue and subcellular distribution of
Cr found after 2 weeks of daily injections of compound 1
in that study, the authors proposed that the trimer entered
cells (at least in part) intact [10]. This was bolstered by
previous studies of the effects of chromic ions on rats
(reviewed in Ref. [4]) that indicate that a role for simple
inorganic chromic ions can be ruled out. Similarly, propionate should not be responsible for the effects on the
cholesterol and triglycerides levels and insulin sensitivity
of the rats given compound 1 [1,2]. The amounts of
propionate provided daily in the previous studies (56
mg / kg body mass or less) [1,2] are far below the amounts
required to observe an effect (reviewed in Ref. [10]).
Consequently, the effects of 1 previously observed in rats
are best attributed to the cation and not its degradation
products. The results of the current study are completely
consistent with this interpretation. Compound 1 enters cells
intact and exists intact in the microsomes for ca. 1 h. The
effects with compound 1 presumably then must arise from
only a small period of time in which the complex remains
intact.
Another possibility, while unlikely, cannot be exhaustively excluded. The active uptake of the trimer by cells
(Shute and Vincent, unpublished results) and the transport
of the complex into microsomes could result in cells
accumulating significant greater concentrations of Cr(III)
about 1 h after injection than possible with other Cr(III)
sources such as CrCl 3 or chromium picolinate, especially if
taken orally. If Cr(III) has a pharmacologically beneficial
effect at these concentrations, then this could account for
the effects on cholesterol levels and insulin sensitivity.
Further studies are thus warranted for examining the
effects of alternative methods of delivering the cation such
as adding the trimer to drinking water resulting in the rats
receiving smaller doses over more extended period of time.
Acknowledgements
Funding was provided by the American Diabetes Association to J.B.V.
References
[1] Y. Sun, K. Mallya, J. Ramirez, J.B. Vincent, J. Biol. Inorg. Chem. 4
(1999) 838–845.
[2] J.B. Vincent, Y. Sun, B.J. Clodfelder, Diabetes 50 (2001) A508.
[3] C.M. Davis, A.C. Royer, J.B. Vincent, Inorg. Chem. 36 (1997)
5316–5320.
[4] J. Vincent, Polyhedron 20 (2001) 1–26.
[5] J. Vincent, J. Nutr. 130 (2000) 715–718.
[6] J. Vincent, Nutr. Rev. 58 (2000) 67–72.
[7] J. Vincent, Acc. Chem. Res. 33 (2000) 503–510.
[8] J.K. Speetjens, A. Parand, M.W. Crowder, J.B. Vincent, Polyhedron
18 (1999) 2617–2624.
278
A. A. Shute, J.B. Vincent / Journal of Inorganic Biochemistry 89 (2002) 272 – 278
[9] J.K. Speetjens, R.A. Collins, J.B. Vincent, S.A. Woski, Chem. Res.
Toxicol. 12 (1999) 483–487.
[10] A.A. Shute, N.E. Chakov, J.B. Vincent, Polyhedron 20 (2001)
2241–2252.
[11] A. Earnshaw, B.N. Figgis, J. Lewis, J. Chem. Soc. A (1966)
1656–1663.
[12] C. de Duve, B.C. Pressman, R. Gianetto, R. Wattiaux, F. Appelmans,
Biochem. J. 60 (1955) 604–617.
[13] G.L. Rowin, in: Methods in Molecular Biology, Marcel Dekker,
New York, 1974, pp. 90–109.
[14] B.J. Clodfelder, J. Emamaullee, D.D. Hepburn, N.E. Chakov, H.S.
Nettles, J.B. Vincent, J. Biol. Inorg. Chem. 6 (2001) 608–617.
[15] R.A. Anderson, N.A. Bryden, M.M. Polansky, J. Am. Coll. Nutr. 16
(1997) 273.
[16] E.G. Offenbacher, H. Spencer, H.J. Dowling, F.X. Pi-Sunyer, Am. J.
Clin. Nutr. 44 (1986) 77.
[17] J.S. Striffler, M.M. Polansky, R.A. Anderson, Metabolism 48 (1999)
1063–1068.
[18] G.Y. Wu, O. Wada, Jpn. J. Ind. Health 23 (1981) 505.
[19] L.L. Hopkins Jr., Am. J. Physiol. 209 (1965) 731.
[20] M.L. Davis-Whitenack, B.O. Adeleye, L.L. Rolf, B.J. Stoecker,
Nutr. Res. 16 (1996) 1009.
[21] L. Manzo, A. Di Nucci, J. Edel, C. Gregotti, E. Sabbioni, Res.
Commun. Chem. Pathol. Pharmacol. 42 (1983) 113.
[22] M.L. Davis-Whitenack, B. Adeleye, B.J. Stroecker, Biol. Trace
Elem. Res. 68 (1999) 175.
[23] S.M. Kamath, B.J. Stoecker, M.L. Davis-Whitenack, M.M. Smith,
B.O. Adeleye, S. Sangiah, J. Nutr. 127 (1997) 478.
[24] L.L. Hopkins Jr., K. Schwarz, Biochim. Biophys. Acta 90 (1964)
484.
[25] F. Borguet, R. Cornelis, J. Delanghe, M.C. Lambert, N. Lamiere,
Clin. Chem. Acta 238 (1995) 71.
[26] F. Borguet, R. Cornelis, N. Lamiere, Biol. Trace Elem. Res. 26–27
(1990) 449.
[27] A.C. Royer, Ph.D. Dissertation, The University of Alabama, Tuscaloosa, AL, 1999.
5316
Inorg. Chem. 1997, 36, 5316-5320
Synthetic Multinuclear Chromium Assembly Activates Insulin Receptor Kinase Activity:
Functional Model for Low-Molecular-Weight Chromium-Binding Substance
C. Michele Davis, April C. Royer, and John B. Vincent*
Department of Chemistry, University of Alabama, Tuscaloosa, Alabama 35487-0336
ReceiVed May 14, 1997X
The biologically-active form of chromium, low-molecular-weight chromium-binding substance (LMWCr), activates
the insulin-dependent tyrosine protein kinase activity of insulin receptor (IR). The site of activation was shown
to be on the active site fragment of the β subunit of IR. As LMWCr previously has been proposed to contain a
multinuclear chromic assembly, the ability of multinuclear chromium assemblies to activate IR kinase activity
has been probed. The trinuclear cation [Cr3O(O2CCH2CH3)6(H2O)3]+ (1) has been found to activate IR activity
in a fashion almost identical to that of LMWCr using rat adipocytic membrane fragments and an active site
fragment of IR, while a variety of other chromic complexes have in contrast been found to be ineffective or to
inhibit kinase activity. The activation of IR kinase activity by complex 1, its stability in aqueous and strongly
acidic solution, and its low molecular weight suggest that it potentially could be used in a treatment for adultonset diabetes.
Introduction
The biologically-active, naturally-occurring oligopeptide lowmolecular-weight chromium-binding substance (LMWCr) has
been found to activate the insulin-dependent tyrosine protein
kinase activity of insulin receptor (IR) approximately 8-fold with
a dissociation constant of circa 250 pM.1 This activity is directly
proportional to the Cr content of the oligopeptide (being
maximal at four chromic ions per oligopeptide), while substitution of chromium with metal ions commonly associated with
biological systems results in inactiving the oligopeptide. Similarly, LMWCr has been reported to activate a membraneassociated phosphotyrosine phosphatase; this activation also
requires four chromic ions per oligopeptide to be maximal, while
chromic ions could not functionally be replaced with other
transition metal ions.2 A role for LMWCr in amplification of
insulin-signaling has been postulated.1,3 Chromium is mobilized
from the blood and taken up by insulin-dependent cells in
response to insulin.4 LMWCr is maintained in its apo form5
but possesses a large chromic ion binding constant(s) as it is
capable of removing chromium from Cr-transferrin.5,6 The
holoLMWCr is then capable of stimulating IR kinase activity,
amplifying the signal of insulin into the insulin-dependent cells.
An association between chromium and insulin-dependent glucose and lipid metabolism has been reported for nearly four
decades;7 however, only recently since procedures for isolation
of quantities of LMWCr suitable for kinetic and spectroscopic
studies have been developed3 has progress been made in
understanding the association on a molecular level.
Abstract published in AdVance ACS Abstracts, October 15, 1997.
(1) Davis, C. M.; Vincent, J. B. Biochemistry 1997, 36, 4382.
(2) Davis, C. M.; Sumrall, K. H.; Vincent, J. B. Biochemistry 1996, 35,
12963.
(3) Davis, C. M.; Vincent, J. B. Arch. Biochem. Biophys. 1997, 399, 335.
(4) Morris, B. W.; Blumsohn, A.; McNeil, S.; Gray, T. A. Am. J. Clin.
Nutr. 1992, 55, 989. Morris, B. W.; Gray, T. A.; MacNeil, S. Clin.
Sci. 1993, 84, 477. Morris, B. W.; MacNeil, S.; Stanley, K.; Gray, T.
A.; Fraser, R. J. Endrocrin. 1993, 139, 339.
(5) Yamamoto, A.; Wada, O.; Ono, T. Eur. J. Biochem. 1987, 165, 627.
(6) Yamamoto, A.; Wada, O.; Ono, T. J. Inorg. Biochem. 1984, 22, 91.
(7) Mertz, W.; Roginski, E. E.; Schwartz, K. J. Biol. Chem. 1961, 236,
318. Mertz, W.; Roginski, E. E. J. Biol. Chem. 1963, 238, 868. Mertz,
W. J. Nutr. 1993, 123, 626. Vincent, J. B. In Encyclopedia of Inorganic
Chemistry; King, B., Ed.; John Wiley: New York, 1994; Vol. 2, pp
661-665.
X
S0020-1669(97)00568-5 CCC: $14.00
An association between the essential nutrient chromium and
adult-onset diabetes has also been postulated.8 Most recently
Anderson and co-workers found improved glycemic control for
180 adult-onset diabetic patients following chromium supplementation,9 while Ravina and Slezack using 138 adult-onset
diabetic patients found reduced insulin requirements.10 Unfortunately, the form of chromium used as a dietary supplement
in these studies, chromium(III) picolinate, has been found to
cause chromosome damage.11 This suggests that a new form
of chromium for use as a dietary supplement and as part of a
potential treatment for adult-onset diabetes is required. LMWCr
would appear to be a possibility. It has a high LD506 and is
biologically active, opposed to chromium picolinate and glucose
tolerance factor (a material isolated from acid-hydrolyzed
Brewer’s yeast extracts) which serve only as sources of readily
absorbable chromium.12 [Despite the apparent significance of
Cr, as much as ninety percent of the American population and
half of the population of developed nations fail to intake the
daily recommended safe and adequate quantities of Cr].13
However, LMWCr is susceptible to hydrolysis under acidic
conditions14 and consequently could not be taken orally without
degradation. Herein is reported a synthetic chromium(III)
complex which is biologically active, stable under acidic
conditions, and readily and inexpensively synthesized and,
therefore, may have potential as a new agent in the treatment
of adult-onset diabetes and associated conditions and as a
nutritional supplement.
Materials and Methods
LMWCr, [Cr3O(O2CCH3)6 (H2O)3]Cl, and [Cr3O(O2CCH2CH3)6
(H2O)3]NO3. LMWCr was available from previous work and contained
(8) Anderson, R. A. Biol. Trace Elem. Res. 1992, 32, 19.
(9) Anderson, R. A.; Cheng, N.; Bryden, N.; Polansky, M.; Cheng, N.;
Chi, J.; Feng, J. Diabetes 1996, 45, Suppl. 2, 124A.
(10) Ravina, A.; Slezack, L. Harefuah 1993, 125, 142.
(11) Stearns, D. M.; Wise, J. P., Jr.; Patierno, S. R.; Wetterhahn, K. E.
FASEB J. 1995, 9, 1643.
(12) McCarty, M. F. J. Opt. Nutr. 1993, 2, 36. Vincent, J. B. J. Nutr. 1994,
124, 117.
(13) Anderson, R. A.; Kozlovsky, A. S. Am. J. Clin. Nutr. 1985, 41, 1177.
Anderson, R. A. In Risk Assessment of Essential Elements; Mertz,
W., Abernathy, C. O., Olin, S. S., Eds.; ISLI Press: Washington, DC,
1994; pp 187-196.
(14) Sumrall, K. H.; Vincent, J. B. Polyhedron, in press.
© 1997 American Chemical Society
Functional Model for LMWCr
3.6-4.0 chromium per oligopeptide.1-3 [Cr3O(O2CCH3)6 (H2O)3]Cl
(2) and [Cr3O(O2CCH2CH3)6 (H2O)3]NO3 (1) were prepared as previously described.15,16 Oligopeptide concentrations were assayed using
the fluorescamine procedure of Undenfriend and co-workers17 with
glycine as a standard. Chromium concentrations were measured using
the diphenylcarbazide method18 and the method of standard additions
to minimize any potential matrix effects. For all kinetic experiments,
solutions of LMWCr, 1, and 2 were prepared by dilutions from more
concentrated stock solutions. The chromium(III)-amino acid mixture
was prepared by mixing chromium(III) nitrate nonahydrate, aspartate,
glutamate, glycine, and cysteine in a 4:2:4:2:2 ratio in water, followed
by heating at 37 °C for 30 min to allow complexes to form. To obtain
proper concentrations of this mixture (in terms of Cr concentration)
for kinetic experiments, serial dilutions were prepared.
Purification of Isolated Adipocytes and Adipocytic Membranes.
Fat cells from male Sprague Dawley rats were isolated using the
procedures of Rodbell19 and Anderson et al.20 with modifications.2 Three
rats (not kept on a Cr-deficient diet) were sacrificed by decapitation
and their epididymal fat pads removed. Rat adipocytes were washed
with 1% bovine serum albumin (BSA), 50 mM Hepes, pH 7.4 buffer
containing 10 µg/mL leupeptin and 5 µg/mL aprotinin. Cells were
homogenized with a manual Teflon homogenizer and frozen and thawed
five times. The lipid layer was removed, and the cell homogenate was
centrifuged for 1 h at 40 000g. The supernatant was removed, and the
pellet was suspended in Hepes buffer and used as the source of
membrane phosphatase and kinase activity. Protein concentrations were
analyzed using the BCA method (Pierce Chemical Co.) with BSA as
standard.
Phosphotyrosine Phosphatase Activity. p-Nitrophenyl phosphate
(p-NPP) was used to determine the amount of PTP activity using the
method of Li et al.21 The assay used 5 mM substrate in 0.05 M Tris,
pH 7.5, unless noted. Activation of PTP activity by LMWCr and other
Cr-containing species was examined as described by Davis et al.2
Solutions of LMWCr, 1, 2, and the Cr-amino acid mixture were
incubated with the enzyme for 15 min at 37 °C before initiation of the
reaction. Hydrolyses proceeded for 1 h at 37 °C. The extent of
hydrolysis was determined at 404 nm ( ) 1.78 × 104 M-1 cm-1).
Phosphotyrosine Kinase Activity. Phosphotyrosine kinase activity
was measured using a protein tyrosine kinase assay kit (Boehringer
Mannheim) which uses an anti-phosphotyrosine antibody to recognize
phosphotyrosine. A fragment of gastrin (amino acids 1-17) which
has been biotinylated so it can be immobilized to streptavidin-coated
microtiter plates (Boehringer Mannheim) was used as the substrate.
The assays were performed in 50 mM Tris, pH 7.4 containing 0.75
µM ATP and 7.5 mM MgCl2 at 37 °C for 75 min as previously
described1 unless otherwise noted. The membrane fragments which
contain phosphotyrosine themselves were removed after the reaction
was terminated with EDTA by Microcon 30 or Microcon 50 microconcentrators (Amicon), and ammonium vanadate was used as a
phosphatase inhibitor. Contributions to the assay by the addition of
metal-containing materials were determined by measuring the background absorbance of the assay in the absence of membranes, and these
contributions were subtracted from all data. Bovine pancreas insulin
was from Sigma. Isolated kinase active site fragment from the β subunit
of human insulin receptor (residues 941-1343) was obtained from
Stratagene and diluted with 50 mM Tris, pH 7.4; the fragment does
not require activation of the kinase activity by added insulin. Five
units of IR fragment were used in kinetic assays; a unit of activity is
defined as the picomoles of (phosphate incorporated/min)/µL of kinase
as received from the manufacturer. Recombinant human insulin-like
growth factor-1 (IGF-1) was obtained from Sigma and reconstituted
with 10 mM HCl.
(15) Johnson, M. K.; Powell, D. B.; Cannon, R. D. Spectrochim. Acta 1981,
37A, 995.
(16) Earnshaw, A.; Figgis, B. N.; Lewis, J. J. Chem. Soc. A 1966, 1656.
(17) Undenfriend, S.; Stein, S.; Bohlen, P.; Dairman, W.; Leimgruber, W.;
Weigele, M. Science 1972, 178, 871.
(18) Marczenko, Z. Spectrophometric Determination of the Elements; Ellis
Horwood: Chichester, England, 1986.
(19) Rodbell, M. J. Biol. Chem. 1964, 239, 375.
(20) Anderson, R. A.: Brantner, J. H.; Polansky, M. M. J. Agric. Food
Chem. 1978, 26, 1219.
(21) Li, J.; Elberg, G.; Gefel, D.; Shechter, Y. Biochemistry 1995, 34, 6218.
Inorganic Chemistry, Vol. 36, No. 23, 1997 5317
Figure 1. Activation of protein tyrosine kinase activity of the isolated
active site fragment of the β subunit of insulin receptor by bovine liver
LMWCr (open squares) and [Cr3O(O2CCH2CH3)6(H2O)3]NO3 (solid
circles) using a fragment of gastrin (0.75 mM) as substrate. The line is
the best curve fit giving for LMWCr a dissociation constant of 133
pM and for [Cr3O(O2CCH2CH3)6(H2O)3]+ a dissociation constant of
1.00 nM.
Miscellaneous. All visible spectroscopic measurements were
obtained with a Shimadzu UV-160A diode array spectrophotometer.
Fluorescence measurements were made with a Perkin-Elmer 204
fluorescence spectrophotometer. 1H NMR were obtained using a Bruker
AM-360 spectrometer at circa 23 °C. Chemical shifts are reported on
the δ scale (shifts downfield are positive) using solvent protio impurity
as a reference. Curve-fitting was performed using SigmaPlot (Jandel
Scientific). All kinetic experiments were performed in triplicate. Errors
are presented throughout including all tables and graphs as the standard
deviations (1σ) of the triplicate analyses. Similarly, all Cr and LMWCr
concentration determinations were made in triplicate. Doubly deionzed
water was used in all operations; plasticware was used whenever
possible.
Results and Discussion
The binding of insulin to the R subunit of insulin receptor
results in tyrosine autophosphorylation of the β subunit of the
receptor, transmitting the signal of the hormone insulin into a
cell; autophosphorylation activates the kinase in the β subunit
which catalyzes phosphorylation of other proteins.22 This kinase
activity is potentiated by the oligopeptide LMWCr. Using
isolated IR, potentiation of IR tyrosine protein kinase activity
by LMWCr has been found to require insulin and is prevented
when the insulin binding site of the external R subunit is
blocked.1 However, the binding site on IR for LMWCr is
unknown. However, studies with a catalytically active fragment
(residues 941-1343) of the β subunit of the human enzyme
indicate that the effect of LMWCr on kinase activity is
associated with this fragment. As shown in Figure 1, addition
of LMWCr to the fragment results in an approximately 3-fold
activation of the kinase activity. Fitting the curve to a
hyperbolic equation gives a dissociation constant for LMWCr
of 133 pM, very similar to the dissociation constant found for
the interaction of LMWCr with isolated rat insulin receptor (250
pM).1 The 3-fold activation is significantly less than that
observed with isolated receptor (approximately 8-fold),1 but this
may be associated with small structural differences between the
(22) Lee, J.; Pilch, P. F. Am. J. Physiol. 1994, 35, C319. White, M. F.;
Kahn, C. R. J. Biol. Chem. 1994, 269, 1.
5318 Inorganic Chemistry, Vol. 36, No. 23, 1997
Davis et al.
Figure 3. Structure of oxo-centered trinuclear chromium carboxylate
cations: Compound 1, L ) H2O, R ) CH2CH3; compound 2, L )
H2O, R ) CH3.
Figure 2. Activation of rat adipocytic membrane protein tyrosine
kinase activity using a fragment of gastrin (0.75 mM) as substrate by
LMWCr (solid circles) and by [Cr3O(O2CCH2CH3)6(H2O)3]NO3 (open
squares) in the presence of 100 nM IGF-1. A 25 µL volume of a rat
membrane suspension corresponding to 0.0895 mg of protein/mL was
utilized. The line is the best fit hyperbolic curve giving a dissociation
constant of 507 pM for LMWCr and for [Cr3O(O2CCH2CH3)6(H2O)3]NO3 a dissociation constant of 730 pM.
fragment and the entire receptor protein. The results suggest
that LMWCr may associate with the kinase active site of the
insulin receptor.
Insulin receptor is part of a family of receptor proteins which
includes the insulin-like growth factor receptors.23 All these
receptors are disulfide-bound heterotetramers of R and β
subunits. Ligand (insulin or insulin-like growth factors) presumably cause a conformational change in the preformed
receptors, resulting in receptor activation. To examine the
specificity of LMWCr, the effects of the oligopeptide on IGF-1
receptors were probed. The kinase activity of rat adipocytic
membrane fragments in the presence of 100 nM IGF-1 is more
than doubled by the addition of LMWCr (Figure 2) with a
dissociation constant of 507 pM. In contrast, in the presence
of 100 nM insulin kinase activity has previously been shown
using the same fragment of gastrin as substrate to be increased
three and one-half times by LMWCr with a similar dissociation
constant of 875 pM.1 [In the absence of added hormone,
LMWCr has no detectable effect on the membrane kinase
activity.1] Thus, LMWCr potentiates both members of the IGF
receptor family.
Given this novel role in the amplification of signal transduction for LMWCr and its rather simple composition (carboxylaterich oligopeptide binding four chromic ions),3,5 the possibility
of identifying a functional model for LMWCr was examined.
Such a model would be required to be soluble and stable in
aqueous solution, be well characterized, and contain a carboxylate-supported multinuclear chromic assembly.3 Fortunately, a
review of the literature revealed a number of trinuclear and
tetranuclear Cr(III) carboxylate assemblies;24,25 however, few
(23) Heldin, C.-H. Cell 1995, 80, 213.
(24) (a) Harton, A.; Nagi, M. K.; Glass, M. M.; Junk, J. L.; Atwood, J. L.;
Vincent, J. B. Inorg. Chim. Acta 1994, 217, 171. (b) Ellis, T.; Glass,
M.; Harton, K.; Huffman, J. C.; Vincent, J. B. Inorg. Chem. 1994,
33, 5522. (c) Nagi, M. K.; Harton, A.; Donald, S.; Lee, Y. S.; Sabat,
M.; O’Connor, C. J.; Vincent, J. B. Inorg. Chem. 1995, 34, 3813. (d)
Bino, A.; Chayat, R.; Pederson, E.; Schneider, A. Inorg. Chem. 1991,
30, 856. (e) Donald, S.; Terrell, K.; Vincent, J. B.; Robinson, K. D.
Polyhedron 1995, 14, 971.
were soluble in water. On the basis of these requirements, two
were chosen: [Cr3O(O2CCH2CH3)6(H2O)3]+ (1) and [Cr3O(O2CCH3)6(H2O)3]+ (2). Both of these complexes possess a basic
carboxylate type structure25 comprised of a planar triangle of
chromic ions with a central µ3-oxide (Figure 3). Each set of
two chromic ions is bridged by two carboxylates ligands, while
six coordination about the chromium centers is completed by a
terminal aquo ligand.
The cation [Cr3O(O2CCH2CH3)6(H2O)3]+ (1) is a wellcharacterized species. Its preparation was first described in
1911, although the formula was proposed as a hydrate salt of
[Cr3(O2CCH2CH3)6(OH)2]+.26 A similar synthesis of a variety
of salts of the cation (still with the wrong formulation) was
reported in 1930.27 The cation, which was originally characterized only by its color and elemental analysis, has subsequently
been characterized by variable-temperature magnetic susceptibility measurements,16,28 electronic spectroscopy,28 luminescence
spectroscopy,29 infrared spectroscopy and X-ray crystallography
(of the nitrate salt),30 ESR,31 fast atom bombardment and
electrospray mass spectrometry,32 and NMR.33 The cation
[Cr3O(O2CCH3)6(H2O)3]+ (2) has been more exhaustively
studied and has served as a model upon which theories of the
magnetic interactions between multiple paramagnetic centers
were tested (reviewed in ref 25).
The ability of the synthetic materials to activate membrane
phosphotyrosine protein phosphatase activity and insulindependent membrane tyrosine protein kinase activity was
examined (LMWCr has previously been shown to also activate
a membrane-associated phosphotyrosine phosphatase activity
in rat adipocytic membrane fragments).2 As shown in Figures
4 and 5, the acetate triangle 2 does not activate but rather inhibits
both the membrane phosphatase and kinase activity. In stark
contrast, the propionate analogue results in an actiVation of both
activities in a fashion very similar to LMWCr. The kinase
activity is stimulated approximately 2-fold, while the phosphatase activity is increased nearly 50%. Fitting the curves of
Figures 4 and 5 to a hyperbolic function results in dissociation
constants for the trinuclear species of 2.98 and 30 nM for the
(25)
(26)
(27)
(28)
(29)
(30)
(31)
(32)
(33)
Cannon, R. D.; White, R. P. Prog. Inorg. Chem. 1988, 36, 195.
Weinland, R. F.; Hoehn, K. Z. Anorg. Chem. 1911, 69, 158.
Weinland, R. F.; Lindner, J. Z. Anorg. Chem. 1930, 190, 285.
Szynanska-Buzar, T.; Ziolkowski, J. J. SoV. J. Coord. Chem. 1976, 2,
897. Zelentsov, V. V.; Zhemchuzhikova, T. A.; Rakitin, Yu. V.;
Yablokov, Yu. V.; Yakubov, Kh. M. Koord. Khim. 1975, 1, 194.
Yoshida, T.; Morita, M.; Date, M. J. Phys. Soc. Jpn. 1988, 57, 1428.
Morita, M.; Kato, Y. Int. J. Quantum Chem. 1980, 18, 625.
Antsyshkina, A. S.; Porai-Koshits, M. A.; Arkhangel’skii, I. V.; Diallo,
I. N. Russ. J. Inorg. Chem. 1987, 32, 1700.
Hondo, M.; Morita, M.; Date, M. J. Phys. Soc. Jpn. 1992, 61, 3773.
Nishimura, H.; Date, M. J. Phys. Jpn. 1985, 54, 395.
Fu, G.; Yu, L.; Zhu, Z.; Xie, W.; Zheng, Y.; Zhang, L. Jiegou Huaxue
1990, 4, 278. van den Bergen, A.; Colton, R.; Percy, M.; West, B. O.
Inorg. Chem. 1993, 32, 3408.
Glass, M. M.; Belmore, K.; Vincent, J. B. Polyhedron 1993, 12, 133.
Functional Model for LMWCr
Figure 4. Activation of rat adipocytic membrane protein tyrosine
kinase activity using 0.75 mM gastrin (amino acids 1-17) as substrate
by [Cr3O(O2CCH3)6(H2O)3]Cl (solid squares) and by [Cr3O(O2CCH2CH3)6(H2O)3]NO3 (open circles) in the presence of 100 nM insulin. A
5 µL volume of rat adipocyte membrane suspension corresponding to
0.0856 mg of protein/mL was utilized. The 100% activity corresponds
to insulin-stimulated kinase activity and is typically about 0.338 pmol
of phosphotyrosine/mg of membranes. The line is the best fit hyperbolic
curve yielding a dissociation constant of 2.98 nM for [Cr3O(O2CCH2CH3)6(H2O)3]NO3.
Figure 5. Activation of rat adipocytic membrane protein phosphatase
activity using 5 mM p-NPP as substrate by [Cr3O(O2CCH3)6(H2O)3]Cl
(solid squares) and by [Cr3O(O2CCH2CH3)6(H2O)3]NO3 (open circles).
A 125 µL volume of a rat membrane suspension corresponding to
0.0856 mg of protein/mL was utilized. The line is the best fit hyperbolic
curve yielding a dissociation constant of 30 nM for [Cr3O(O2CCH2CH3)6(H2O)3]+.
kinase and phosphatase activities, respectively. These results
are strikingly similar to those using LMWCr. LMWCr results
in a 250% increase in insulin-dependent tyrosine kinase activity
with a dissociation constant of 875 pM1 (one third that of the
model) and a 100% increase in phosphatase activity with a
dissociation constant of 4.4 nM2 (one-seventh that of the model).
Consequently, 1 is an excellent functional model of LMWCr
but possesses somewhat less activation while requiring slightly
higher concentrations to achieve these affects. To test just how
good a model of LMWCr that complex 1 is, its ability to activate
Inorganic Chemistry, Vol. 36, No. 23, 1997 5319
the active site fragment of the β subunit of IR and the IGF-1
receptor were also examined (Figures 1 and 2). For the IR β
subunit fragment, complex 1 resulted in a circa 60% increase
in kinase activity with a dissociation constant of 730 nM; for
the IGF-1-dependent membrane kinase activity, an increase of
250% was observed with a dissociation constant of 1.00 nM.
In both cases the dissociation constant for the synthetic material
is within 1 order of magnitude of that for LMWCr. Thus, the
trinuclear chromic assembly 1 mimics LMWCr in its ability to
activate adipocytic membrane phosphotyrosine phosphatase
activity, insulin-dependent adipocytic membrane tyrosine protein
kinase activity, insulin-like growth factor-1-dependent adipocytic
membrane tyrosine protein kinase activity, and the tyrosine
protein kinase activity of the active site fragment of the β subunit
of insulin receptor. The ability of LMWCr and complex 1 to
activate both protein tyrosine kinases and phosphotyrosine
phosphatases may seem paradoxical; however, the stimulation
of both types of enzymes appears to be common in complex
signal transduction pathways.34
To guarantee that the trinuclear cation 1 was the actual active
species in solution, the stability of complex in water and in the
buffer had to be ascertained. Paramagnetic NMR has been
demonstrated to be of utility in characterizing antiferromagnetically-coupled chromium(III) assemblies.24a-c,33,35 For acetate and propionate ligands bridging between chromic centers
in these assemblies, the resonances of methyl hydrogens of
acetate and the methylene protons of propionate occur in the
+35 to +45 ppm range.33 The nitrate salt of 1 was dissolved
in D2O and in 50 mM Tris buffer (prepared by dissolving Tris
in D2O and adding a quantity of DCl equivalent to the quantity
of HCl needed to make the same quantity of 50 mM Tris buffer,
pH 7.4 in H2O); 1H NMR spectra of 5 mM solutions of 1 were
collected every 5 min for 2 h. During this period, the
integrations of the propionate methylene resonances (+42 ppm)
were unchanged, and no new signals appeared. Thus, the
triangle appears to be stable in aqueous solution and in the assay
buffer, and the activation activity can be assigned to the
trinuclear cation.
Additionally the components of 1 were examined for any
ability to potentiate membrane phosphatase activity and insulindependent membrane kinase activity (Figure 6). Propionic acid
and nitrate (at concentrations comparable to those if 1 dissociated completely in water) do not potentiate either phosphatase
or kinase activity; previously, mononuclear chromic salts were
also shown to not result in potentiation of either activity.1,2 A
mixture of the components of LMWCr was similarly tested for
its ability to potentiate these activities. The mixture consisted
of chromic ions, aspartate, glutamate, glycine, and cysteine in
a 4:2:4:2:2 ratio, corresponding to the approximate ratio of the
components in isolated bovine liver LMWCr.3 At a chromium
concentration equivalent to that used for LMWCr in the assays,
the mixture actually inhibited kinase and phosphatase activity.
Thus, the multinuclear chromic complexes 1 and LMWCr
appear to be unique in their ability to potentiate membrane
phosphotyrosine phosphatase and insulin-dependent membrane
kinase activity. Unfortunately, the isostructural Mn(III) and Fe(III) analogues of complexes 1 and 2 are unstable in water and
consequently unsuitable for use for comparison in the phosphatase and kinetic assays.
While LMWCr has been proposed for use as a nutritional
supplement and in treatment of adult-onset diabetes and related
conditions associated with improper carbohydrate and lipid
(34) Hunter, T. Cell 1995, 80, 225.
(35) Vincent, J. B. Inorg. Chem. 1994, 33, 5604. Belmore, K.; Madison,
X. J.; Harton, A.; Vincent, J. B. Spectrochim. Acta 1994, 50A, 2365.
5320 Inorganic Chemistry, Vol. 36, No. 23, 1997
Davis et al.
mineral acids;15 consequently 1 might readily survive oral
ingestion unlike LMWCr. Cation 1 also has a molecular weight
of 664 compared to approximately 1480 Da for bovine liver
LMWCr,3 which should facilitate movement of the former across
cell membranes. Curiously, Mirsky and co-workers reported
that addition of cations 1 and 2 and [Cr3O(O2CH)6(H2O)3]+
caused a 15-20% enhancement of carbon dioxide production
by yeast;36 the similar behavior of the three cations does not
agree with the quite different behavior seen for cations 1 and 2
in this work. Unfortunately, experimental details for the studies
with the three cations by Mirsky et al. are lacking, and no data
are presented. As a result, it is difficult to determine how the
cations might be influencing fermentation and to ascertain the
relationship (if any) between the yeast fermentation assay and
human glucose and lipid metabolism.
Figure 6. Influence of chromium complexes and their components
on adipocytic membrane tyrosine protein kinase activity in the presence
of 100 nM insulin (solid bars) and adipocytic membrane phosphotyrosine phosphatase activity (open bars). The 100% activity represents
the activity in the absence of added chromium complexes or their
components. A 25 µL volume of a rat membrane suspension corresponding to 0.0895 mg of protein/mL was utilized. For the kinase
assays, 50 nM concentrations of LMWCr, complex 1, and the Cramino acid mixture (in terms of Cr) were used; 500 nM nitrate and
propionate were used. For the phosphatase assays, 50 µM concentrations
of LMWCr, complex 1, and the Cr-amino acid mixture (in terms of
Cr) were used; 50 µM nitrate and propionate were used. Model )
complex 1; Acid ) propionic acid; AA + Cr ) chromium-amino
acid mixture.
metabolism,1-3 complex 1 may be even more promising for use
in these applications. The synthetic material is prepared from
inexpensive reagents16 (and consequently not requiring a timeconsuming isolation as with LMWCr), is extremely stable in
aqueous solution (LMWCr undergoes a slow hydrolysis),3 and
even stable in acidic solution. The trinuclear basic carboxylates
of chromium(III) can, for example, be recrystallized from dilute
Conclusions
These studies show that LMWCr and the synthetic analogue
[Cr3O(O2CCH2CH3)6(H2O)3]+ activate insulin receptor protein
tyrosine kinase activity by interacting at or near the kinase active
site of the enzyme’s β subunit, while both materials are able to
activate protein tyrosine kinase activity of rat adipocytic
membrane fragments in response to insulin-like growth factor
in addition to insulin and also activate phosphotyrosine phosphatase activity of adipocytic membranes. The similarity
between the activation by LMWCr and complex 1 supports the
proposal that LMWCr possesses a multinuclear chromic assembly similar to that of complex 1.3 The mechanism of the
activation by the chromium complexes is under investigation,
as are the effects of the materials on diabetic animal models.
Acknowledgment. The authors gratefully acknowledge the
support of this work by the American Heart Association (Grantin-Aid 94011190).
IC970568H
(36) Mirsky, N.; Weiss, A.; Dori, Z. J. Inorg. Biochem. 1980, 13, 11.
© Copyright 2006 by Humana Press Inc.
All rights of any nature, whatsoever, reserved.
0163-4984/(Online) 1559-0720/06/11301–0053 $30.00
High-Dose Chromium(III)
Supplementation Has No Effects
on Body Mass and Composition
While Altering Plasma Hormone
and Triglycerides Concentrations
RANDALL BENNETT,1 BOBBI ADAMS,1 AMANDA FRENCH,1
YASMIN NEGGERS,2 AND JOHN B. VINCENT*,1
Departments of 1Chemistry and Coalition for Biomolecular
Products and 2Human Nutrition and Coalition for Biomolecular
Products, The University of Alabama,
Tuscaloosa, AL 35487-0158
Received November 8, 2005; Accepted January 6, 2006
ABSTRACT
Chromium is generally believed to be an essential element and is
often claimed to have value as a weight loss or muscle building agent.
Recent studies in humans and rats have failed to demonstrate effects on
body composition, although recent studies with pharmacological doses of
the cation [Cr(III)3O(O2CCH2CH3)6(H2O)3]+ (or Cr3) (≤1 mg Cr/kg body
mass) in rats have noted a trend toward body mass loss and fat mass loss.
Thus, the effects of large gavage doses of Cr3 (1–10 mg Cr/kg) on body
mass, organ mass, food intake, and blood plasma variables (insulin, glucose, leptin, cholesterol, and triglycerides) were examined over a 10-wk
period using male Sprague–Dawley rats. No effects on body composition
were noted, although Cr3 administration lowered (p < 0.05) plasma
insulin, leptin, and triglycerides concentrations. As Cr3 is absorbed greater
than 10-fold better than commercially available nutritional supplements,
the lack of an effect of the Cr(III) compound at these levels of administration clearly indicates that Cr(III) supplements do not have an effect on
body composition at any reasonable dosage.
Index Entries: Chromium; aspartame; saccharin; insulin; leptin; cholesterol; Cr3; rats; body composition.
* Author to whom all correspondence and reprint requests should be addressed.
Biological Trace Element Research
53
Vol. 113, 2006
54
Bennett et al.
INTRODUCTION
In developed countries, such as the United States, more than half of
the adult population can be classified as overweight or obese (1). Nearly
50 million Americans go on a diet each year, yet only 5% keep off the
weight they lose. Obesity increases the likelihood of developing conditions
such as type 2 diabetes, cancer, and heart disease. In the last 15 yr, nutritional studies have suggested that chromium (Cr) in the +3 oxidation state
has a role in insulin-dependant carbohydrate and lipid metabolism in
mammals (2–4). As Cr’s role might involve potentiation of insulin signaling, increasing insulin sensitivity, Cr has been proposed to elicit an effect
on body composition. Initial studies reported that dietary Cr supplementation could potentially reduce fat mass and increase lean body mass (5–9).
However, well-controlled follow-up studies were unable to support the
effect of Cr on body composition in humans (10–17). A recent comprehensive review (18) and recent meta-analyses (19,20) have clearly established
that Cr supplementation has no effect on body mass or composition of
healthy individuals. However, Cr remains the second largest selling mineral supplement after calcium.
Recently, gavage administration of the trinuclear chromium(III) complex [Cr3O(O2CCH2CH3)6(H2O)3]+ (also known as Cr3) has been shown to
lower plasma insulin, total and low-density lipoprotein (LDL) cholesterol,
and triglycerides levels of healthy rats (21). In association with these
changes in blood variables, a trend toward loss of body mass with increasing Cr dosage and a statistically significant loss of epididymal fat at the
highest Cr3 dosage (1 mg Cr/kg body mass) was noted (21). Given that
Cr3 is absorbed with 40–60 efficiency (22), over 10-fold greater than that of
commercial chromium(III) supplements such as chromium picolinate and
chromium nicotinate (23,24) and the large doses of Cr used in rat studies
compared to human studies, these observations could suggest that high
doses of Cr3 (and potentially other forms of Cr) could effect body composition at high dosages.
Additionally, effects of high doses of Cr(III) compounds have been
reported to lower the concentration of the hormone leptin in blood plasma
or serum (25–27). Leptin is an adipocyte hormone that signals the brain
about the fat content of the body, resulting in changes in appetite. Sun et
al. (25) found that rats on a basal or high-fat diet receiving 1, 2, or 3 mg
Cr/kg diet for 8 wk had lower serum leptin levels (p < 0.05) than those of
controls. Similarly, Wang et al. (26) have found that rats fed a high-fat diet
supplemented with 3 mg Cr/kg for 5 wk possessed lower leptin levels
than those of controls on the high-fat diet. As blood leptin levels might
reflect fat composition with increased percent body fat leading to
increased leptin levels, lowering of leptin by Cr treatment could suggest
the existence of effects on body composition.
Herein are reported the results of a study to test whether Cr3 can have
an effect on body composition at high dosages and to determine whether
Cr supplementation can affect plasma leptin levels.
Biological Trace Element Research
Vol. 113, 2006
Cr(III) Supplements Do Not Affect Body Composition
55
MATERIALS AND METHODS
[Cr3O(O2CCH2CH3)6(H2O)3]NO3, Cr3, and Artificial
Sweeteners
The nitrate salt of the trinuclear Cr(III) cation, Cr3, was prepared as
described in the literature (28). The artificial sweeteners saccharin and
aspartame were obtained from ACROS Organics (New Jersey, USA) and
Sigma Chemical Co. (St. Louis, MO, USA), respectively. All operations
were performed with doubly-deionized water unless otherwise noted and
were performed with plasticware whenever possible.
Animals
All rats were obtained from Charles River Laboratories, Inc. (Wilmington, MA, USA). The 5-wk-old male Sprague–Dawley rats were allowed
to feed ad libitum on a commercial rat food (Harland Tekland Certified
LM-485 Mouse/Rat Sterilizable Diet) and tap water and allowed to acclimate to their surroundings for 1 wk before the initiation of the experiments. The commercial food provides a Cr-adequate diet (21). Rats were
raised in standard plastic and stainless-steel cages on a 12-h light–dark
cycle. Forty-eight Sprague–Dawley rats were divided randomly into six
groups of eight. The first group was gavaged daily with an aqueous solution containing Cr3 to give a total amount of Cr equivalent to 1 mg Cr/kg
body weight. The second group received daily an aqueous solution of Cr3
equivalent to 5 mg Cr/kg body mass. The third group received daily an
aqueous solution of Cr3 equivalent to 10 mg Cr/kg body mass. The fourth
and fifth groups were provided a 0.1% aqueous solution of aspartame or
saccharin instead of normal drinking water. The last group was gavaged
daily with an equal volume of water as the Cr3-treated rats and served as
the control group. Solid food intake and body mass were monitored every
4 d. The amount of water consumed by the control and the amount of artificial-sweetener-treated water consumed by the fourth and fifth groups
were monitored every day. After 10 wk, the rats were sacrificed by CO2
asphyxiation. The liver, kidney, heart, spleen, testes, pancreas, epididymal
fat, perirenal fat, and subcutaneous fat were quickly harvested and
weighed on plastic weigh boats. The University of Alabama Institutional
Animal Use and Care Committee approved all experiments involving rats.
Blood Chemistry
Blood (approx 1.5–2.0 mL) was collected from tail snips into
polypropylene tubes after 5 and 10 wk of administration of the
[Cr3O(O2CCH2CH3)6(H2O)3]+, H2O, saccharin, and aspartame. Prior to
blood collection, the rats were fasted for 12–16 h. The 0.1% aspartame and
saccharin solutions were replaced during this time with water. Immediately after blood removal, 0.5 mg/mL heparin and 10 mg/mL NaF were
Biological Trace Element Research
Vol. 113, 2006
56
Bennett et al.
added to the blood. The blood was then immediately centrifuged; the
blood plasma was tested for glucose, total cholesterol, triglycerides, and
high-density lipoprotein (HDL) cholesterol levels using diagnostic kits
from Pointe Scientific Inc. (Canton, MI, USA), for insulin levels using antibody-coated kits from MP Biomedicals (Orangeburg, NY, USA), and for
leptin levels using precoated microplates from R&D Systems Inc. (Minneapolis, MN, USA). Ultraviolet (UV)–visible absorbance measurements
were made with a Hewlett-Packard 8453 spectrophotometer. Gamma
counting was performed with a Packard Cobra II auto-gamma counter.
Statistical Analyses
Data were stratified by weeks of dietary Cr3 treatment or treatment
with aspartame and saccharin into two groups (5 and 10 wk). In each
group of Sprague–Dawley rats, analysis of variance (ANOVA) was used to
test the difference in mean concentrations of plasma variables: glucose,
insulin, cholesterol, triglycerides, HDL, LDL, and leptin by four levels of
dietary Cr: control or 0, 1, 5, and 10 mg Cr/kg body mass. Differences in
mean organ mass by the four dietary Cr levels were also tested by
ANOVA. Differences in blood variables and organ masses between control
and artificial-sweetener-treated groups were evaluated using a pooled Student’s t-test. The level of significance for all analyses was set at p < 0.05.
Data were analyzed using SAS software (version 9.1). Numerical values in
the tables and the text are presented as mean ± the standard deviation.
RESULTS
Body Composition and Food and Water Intake
There was not a statistically significant difference (p > 0.05) in the percent mass gain (average mass gain/average mass on d 1 × 100%) of the
control group and the treated groups throughout the 10-wk period (Fig. 1).
All of the rats appeared normal throughout, and no visible differences
were observed among any of the groups. As expected from the lack of a
significant body mass difference, the food intakes of the different groups
were also not statistically different throughout the experimental period
(Fig. 2). (The decrease in food intake at d 32 resulted from the first fasting
period before blood sampling, which occurred during this time period.)
Thus, neither Cr3 nor the artificial sweeteners had any effect on body mass
or food intake. In contrast, saccharin treatment resulted in a distinct
increase in the amount of water consumed by the rats compared to that of
the control, and rats receiving aspartame consumed an amount of water
equivalent to that of the control group (Fig. 3). The lack of a statistical difference in body mass and food intake as a function of Cr-complex administration of rats is consistent with numerous other studies (21,29–33),
although these studies involved lower levels of Cr. The increase in water
Biological Trace Element Research
Vol. 113, 2006
Cr(III) Supplements Do Not Affect Body Composition
57
Fig. 1. Percent body mass increase of male Sprague–Dawley rats administered Cr3 or artificial sweeteners. No significant differences between control and
supplemented groups were found.
Fig. 2. Food intake of male Sprague–Dawley rats administered Cr3 or artificial sweeteners. No significant differences between control and supplemented
groups were found.
intake with saccharin but not aspartame at 0.1% concentration is also consistent with previous results (34,35). The results with aspartame on body
composition are in contradiction to those of Beck et al. (36), who provided
male Long–Evans rats a solution of 0.1% aspartame for drinking water for
14 wk. Aspartame treatment resulted in lower body masses (p < 0.05) after
Biological Trace Element Research
Vol. 113, 2006
58
Bennett et al.
Fig. 3. Fluid intake of male Sprague–Dawley rats administered artificial
sweeteners.
4 wk of treatment and an 8% loss (p < 0.02) of mass compared to those of
controls by the end of the experiment. Other than the use of different
strains of rats, no explanation is apparent for the discrepancy between the
results. However, the results of Beck et al. on body mass are also not consistent with the work of other laboratories (37). Additionally, a review of
human studies using aspartame has indicated that aspartame does not
have any effect of food intake or body mass (38). The Long–Evans rats did
not have increased food intake or fluid intake compared to controls (36), in
agreement with the results with the Sprague–Dawley rats.
Tissue Mass
Comparison of the tissue masses after 10 wk of treatment revealed no
statistically significant effects from any of the treatments (Table 1). A lack
of any statistically significant effect of treatment of Cr3 is consistent with
previous results using lower doses of the Cr compound (21,29–33). A statistically significant decrease in epididymal fat mass upon gavage treatment of male Sprague–Dawley rats with 1 mg Cr/kg as Cr3, but not 0.500
or 0.250 mg Cr/kg, for 24 wk was observed (21). This suggested that the
highest dose of Cr3 previously examined might point toward an effect of
large doses of the compound on fat mass. Given the variability of this
result from experiment to experiment, this potentially important observation required reproduction before any true significance could be associated
with it. The lack of any effect in the current work suggests that the previous observation was a statistical anomaly.
Biological Trace Element Research
Vol. 113, 2006
Cr(III) Supplements Do Not Affect Body Composition
59
Table 1
Percentage of Relative Organ Mass (Tissue Mass/Body Mass × 100%)
of Male Sprague–Dawley Rats After 10 wk of Administration of Cr3
and Artificial Sweeteners
Note: Values are means ± standard deviation; eight rats per group. No variables were significantly different between treated rats and the control group at p < 0.05.
Similarly to Cr3, neither artificial sweetener had any effect on tissue
mass. Again, this is not in agreement with the results of Beck et al. (36),
who observed that Long–Evans rats receiving 0.1% aspartame-containing
drinking water had lower epididymal, subcutaneous, and perirenal fat
mass than controls. The mass of the heart, liver, and spleen were not
affected by the treatment (36).
With one exception, all of the tissues appeared normal, and no visible
differences were observed among any of the groups in the current study.
The exception was the livers of the aspartame-treated rats, which possessed a distinct gray color compared to the livers of the other groups.
Blood Variables
A distinct downward trend in insulin concentration was observed as
a function of increased Cr dosage after 5 wk of treatment, but this effect
was not statistically significant. However, as observed previously with
gavage administration of Cr3 to healthy rats (and diabetic model rats) (21),
Cr3 was found to result in significant lowering of blood plasma insulin
concentrations (1, 5, and 10 mg Cr/kg) after 10 wk of treatment, and this
effect was dependent on the Cr dosage (Table 2). Similarly, after 10 wk of
treatment, plasma triglycerides dropped as the Cr dosage was increased,
again consistent with the results of earlier studies using Cr3 given intravenously or orally (21,39,40). In contrast to previous studies with lower
doses of Cr (21,39,40), plasma glucose levels dropped as a function of Cr
dosage after 5 wk of treatment, but the effect was largely ameliorated after
10 wk of treatment. Cr3 had no effect on plasma HDL levels, as has been
noted previously (21,39,40). Although after 10 wk of treatment, the levels
Biological Trace Element Research
Vol. 113, 2006
60
Bennett et al.
Table 2
Effects of the Cr3 and Artificial Sweeteners on Plasma Variables of Male
Sprague–Dawley Healthy Rats After 5 and 10 wk of Administration
Note: Values are means ± standard deviation; eight rats per group. For each variable for rats
receiving Cr, means with different superscripts are significantly different from those of the control (p < 0.05). For rats receiving the artificial sweeteners, asterisks indicate variables significantly different from those of the control (p < 0.05).
of plasma cholesterol and LDL cholesterol appeared to drop with Cr
administration, the effect is not statistically significant. This result is in
contrast to earlier studies with Cr3 in which the decrease has been statistically significant (39,40). Additionally, Cr3 administration led to lower
plasma leptin concentrations in a dose-dependent fashion. The lowering of
leptin levels is also consistent with previous reports using other forms of
Cr (25,26). The lowering of plasma leptin levels without an appreciable
effect on body mass or fat mass indicates that the drop in leptin is not simply a result of less fat being present to produce the hormone.
After 5 wk of administration in the drinking water, aspartame was
found to lower plasma glucose, insulin, and leptin concentrations; however, after 10 wk of administration, these differences were no longer
observed. However, plasma LDL cholesterol levels and triglycerides levels
were lower in the rats receiving aspartame for 10 wk. Beck et al. (36) have
previously observed that Wistar rats receiving 0.1% aspartame in their
drinking water possessed lower leptin levels after 14 wk of treatment; this
lowering of leptin levels could be correlated with a loss of fat mass and
Biological Trace Element Research
Vol. 113, 2006
Cr(III) Supplements Do Not Affect Body Composition
61
body mass. Again, the current study has failed to reproduce the results of
Beck et al. The original design of this study was to utilize the sweetenertreated rats as a positive control based on the results of Beck et al. Saccharin-treated rats possessed significantly lower concentrations of leptin after
5 wk of treatment, but no plasma variables except triglycerides differed
from controls after 10 wk of treatment. Consequently, the sweeteners
appear to have effects at the early stage of the study, but these effects generally disappear by the 10th week of treatment.
DISCUSSION
The in vivo effects of administration of Cr3 to healthy and type 1 and
type 2 diabetic model rats have previously been examined (21,39,40). (The
synthetic cation was initially given intravenously to avoid potential differences in absorption between the healthy and diabetic model rats.) After 24
wk of intravenous administration (0–20 µg Cr/kg body mass) to healthy
male rats, Cr3 resulted in a concentration-dependent lowering of levels of
fasting blood plasma LDL cholesterol, total cholesterol, triglycerides, and
insulin and of 2-h plasma insulin and glucose levels after a glucose challenge (40). These results confirmed the results of previous 12-wk study
examining the effect of the synthetic cation on healthy rats (39) and are in
stark contrast to those from the administration of other forms of Cr(III) to
healthy rats, which have no effects on these parameters (30). The cation
had little, if any, effect on rats with streptozotocin-induced (type 1 model)
diabetes. This might have resulted from the increased variation in blood
variables measurements for these rats compared to those of controls, such
that insufficient power existed to observe any potential effects. However,
Zucker obese rats (an early-stage type 2 diabetes model) after 24 wk of
supplementation (20 µg/kg) had lower fasting plasma total cholesterol,
triglycerides, insulin, and LDL and HDL cholesterol levels and lower 2-h
plasma insulin levels after a glucose challenge. The lowering of plasma
insulin concentrations with little effect on glucose concentrations suggests
that the supplement increases insulin sensitivity (40). No acute toxic effects
were observed for supplementation with the compound (41), and it does
not give rise to DNA damage in in vitro studies as was observed with
chromium picolinate (42).
The effects of oral (gavage) administration of the complex have also
been examined (21). For levels of 250, 500, or 1000 µg Cr/kg body mass,
the treatment at all doses lowered fasting plasma insulin, triglycerides,
total cholesterol, and LDL cholesterol levels of healthy rats while having
no effect on plasma glucose or HDL cholesterol. These levels were lower
after 4 wk of treatment and remained lower for the next 20 wk of treatment. The maintenance of glucose levels with less insulin indicates
increased insulin sensitivity. Both plasma glucose and insulin levels were
lowered in 2-h glucose tolerance tests. Additionally, the healthy rats receivBiological Trace Element Research
Vol. 113, 2006
62
Bennett et al.
ing the largest dose of Cr tended to possess less body mass than controls
and had approx 10% less epididymal fat; this tendency toward less body
mass [also observed in earlier studies with Cr3 (39,40)] and loss of fat suggested the possibility that Cr3 could potentially affect body composition at
the highest doses. In Zucker obese rats, the early-stage type 2 diabetes
model, receiving 1000 µg Cr/kg body mass, the results were similar to
those from intravenous administration. Also in this same study, the effects
of the Cr3 on ZDF rats, a genetic model for type 2 diabetes, were also
examined using 1000 µg Cr/kg body mass. Again, fasting plasma insulin,
triglycerides, total cholesterol and LDL cholesterol levels were all lower
while glucose concentrations were consistently but not statistically lower.
HDL levels were lowered from their very high levels. Two-hour plasma
insulin levels were also lowered. Plasma glycated hemoglobin levels, a
measure of longer-term blood glucose status, were examined in the
healthy, Zucker obese, and ZDF rats after 4, 12, and 24 wk of treatment. No
effect was seen for the healthy rats; however, significant effects were noted
for the diabetic models. For the ZDF rats, glycated hemoglobin was lower
after 12 and 24 wk of treatment, reaching almost a 22% drop compared to
ZDF controls by wk 24; for the Zucker obese rats, glycated hemoglobin
was 27% lower at wk 24. Control studies using an intravenous injection
containing an amount of propionate equivalent to that received in the
largest dose used earlier have not observed similar effects (21).
The effects of Cr3 on healthy and model diabetic rats have also been
examined by Debski and co-workers (43,44). Male Wistar rats were provided a control diet or a diet containing 5 mg Cr/kg diet as the cation for
10 wk. Blood plasma insulin levels were lowered 15.6% by the Cr-containing diet, and glucose transport by red blood cells was increased 9.6% (43).
In another study, this group utilized male Wistar rats with streptozotocininduced diabetes. Using similar diets for 5 wk, the rats given the Cr diet
had lower blood serum glucose levels (26%) and increased HDL levels
(14%) (44).
The effects of Cr3 might result from the intact complex [it has been
proposed to activate the tyrosine kinase activity of insulin-stimulated
insulin receptor (45)] or might result from Cr3 being a particularly efficient
means by which to incorporate Cr into cells. At a nutritionally relevant
level (3 µg Cr/kg body mass) and a pharmacologically relevant level (3
mg/kg), at least 60% and 40% of the compound, respectively, is absorbed
in 24 h when given by gavage administration to rats (22). This represents
an approx 10-fold increase over those of chromium picolinate (marginally
soluble in water, 0.6 mM), CrCl3 (which oligimerizes in water), and
chromium nicotinate [Cr(nic)2(OH), insoluble in water] (23,24). The solubility of the Cr3 and its stability thus allow a unique amount of the material to enter the circulatory system and tissues. During the first 24 h after
intravenous injection, the fate of 51Cr-labeled Cr3 in tissues, blood, urine,
and feces has been followed (46). Remarkably, the complex is readily incorporated into tissues and cells. The complex rapidly disappears from the
Biological Trace Element Research
Vol. 113, 2006
Cr(III) Supplements Do Not Affect Body Composition
63
blood (<30 min) as radiolabeled Cr from the cation appears in tissues. In
hepatocytes, the intact cation is efficiently transported into microsomes,
where its concentration reaches a maximum in approx 2 h (and corresponds to >90% of Cr in the cells from the injected complex); this suggests
that the cation is actively transported into cells via endocytosis; identification of the protein(s) responsible is needed. As the complex is degraded in
hepatocytes and the levels in microsomes rapidly decrease, Cr appears in
the urine as chromodulin (or a similar molecular-weight chromium-binding species). The synthetic complex is degraded before or during its disappearance from the microsomes. Rats have also been given the 51Cr- or
14C-labeled complex by intravenous injection daily for 2 wk (47). Thirty
percent of the injected Cr was lost daily as chromodulin or a similar
species; only very small amounts are lost in the feces. The tissue and subcellular hepatocyte distribution of Cr after 2 wk was examined; no intact
complex could be detected. Only approx 3% of the propionate from each
injection of the complex was lost daily; the tissue and cellular distribution
of derivatized propionate after 2 wk varied greatly from that of Cr. Thus,
the active transport of Cr3 is different from the transport of all other synthetic forms of Cr proposed as dietary supplements; these proposed supplements appear to enter cells passively by diffusion. Hence, Cr3 has a
significantly greater ability to enter cells than other Cr supplements and
can bring about positive changes in carbohydrate and lipid metabolism
unlike other Cr supplements.
In contrast, for example, chromium chloride and chromium picolinate
at comparable oral doses (30) to previous gavage studies with Cr3 (21) not
only had no effects on body mass or body composition but also had no
effects on plasma blood variables (total cholesterol, triglycerides, and glucose). Given this, Cr3 might be more likely to have effects on body composition at higher dosage than other Cr(III) compounds examined to date
because of its effects on blood variables at the lower dosages. However,
this clearly is not the case. Given that a daily dose of 10 mg Cr/kg body
mass for a rat corresponds to an average body mass human daily taking
over a gram of Cr (this is far in excess of the 200 µg Cr generally taken
daily in human nutritional supplements), these data in rats in which no
effects of high-dose Cr supplementation are observed on body mass or
body composition clearly indicate that no such body changes should be
expected in humans. Thus the results of this study are in accord with
recent reviews and meta-analyses (18–20) on the effects of Cr on body
composition in humans.
CONCLUSIONS
The results of this study using very high doses of Cr(III) as Cr3
demonstrate that Cr(III) has no effect on body composition and body mass
in rats. These results combined with recent analyses of data on human subBiological Trace Element Research
Vol. 113, 2006
64
Bennett et al.
jects taking Cr supplements suggests that no effects should be anticipated
on body mass or body composition in humans taking much lower levels of
Cr supplements. Cr3 was found to result in lower leptin levels in blood
plasma after 5 and 10 wk of gavage administration at doses from 1 to 10
mg Cr/kg body mass. The effects of Cr(III) compounds on plasma leptin
is an area worthy of continued investigation.
ACKNOWLEDGMENT
The authors wish to thank Dontarie Stallings, James A. Neville, and
the staff of The University of Alabama Animal Care Facility for assistance
with the rat studies. A.F. was supported by the undergraduate research
component of a Howard Hughes Medical Research Institute Award to The
University of Alabama; B.A. was supported by the McNairs Scholars program at The University of Alabama. J.B.V. is the inventor or co-inventor of
five patents on the use of Cr-oligopeptides or Cr3 as nutritional supplements or therapeutic agents.
REFERENCES
1. K. M. Flegal, M. D. Carrol, C. L. Ogden, and C. L. Johnson, Prevalence and trends in
obesity among US adults 1999–2000, JAMA 288, 1723–1727 (2002).
2. J. B. Vincent, The bioinorganic chemistry of chromium(III), Polyhedron 20, 1–26 (2001).
3. J. B. Vincent, Elucidating a biological role for chromium at a molecular level, Acc. Chem.
Res. 33, 503–510 (2000).
4. H. C. Lukaski, Chromium as a supplement, Annu. Rev. Nutr. 19, 279–301 (1999).
5. G. W. Evans, The effect of chromium picolinate on insulin controlled parameters in
humans, Int. J. Biosoc. Med. Res. 11, 163–180 (1989).
6. G. R. Kaats, J. A. Wise, K. Blum, et al., The short-term therapeutic efficacy of treating
obesity with a plan of improved nutrition and moderate calorie restriction, Curr. Ther.
Res. 51, 261–274 (1992).
7. G. R. Kaats, K. Blum, J. A. Fisher, and J. A. Adelman, Effects of chromium picolinate
supplementation on body composition: a randomized double-masked placebo-controlled study, Curr. Ther. Res. 57, 747–756 (1996).
8. R. Bulbulian, D. D. Pringle, and M. S. Liddy, Chromium picolinate supplementation in
male and female swimmers, Med. Sci. Sport. Exerc. 28(5 Suppl.), S111 (1996).
9. B. Bahadori, S. Wallner, H. Schneider, T. C. Wascher, and H. Topak, Effects of chromium
yeast and chromium picolinate on body composition in obese non-diabetic patients during and after a very low-calorie diet, Acta Med. Austr. 24, 185–187 (1997) (in German).
10. S. P. Clancy, P. M. Clarkson, M. E. DeCheke, et al., Effects of chromium picolinate supplementation on body composition, strength and urinary chromium loss in football
players, Int. J. Sport Nutr. 4, 142–153 (1994).
11. L. K. Trent and D. Thielding-Canel, Effects of chromium picolinate on body composition, J. Sports Med. Phys. Fitness 35, 273–280 (1995).
12. H. C. Lukaski, W. Bolonchuk, W. A. Siders, and D. B. Milne, Chromium supplementation and resistance training: effects on body composition, strength, and trace element
status of men, Am. J. Clin. Nutr. 63, 954–965 (1996).
Biological Trace Element Research
Vol. 113, 2006
Cr(III) Supplements Do Not Affect Body Composition
65
13. M. A. Hallmark, T. H. Reynolds, C. A. DeSouza, C. G. Dotson, R. A. Anderson, and M.
A. Rogers, Effects of chromium on resistance training on muscle strength and body
composition, Med. Sci. Sports Exerc. 28, 139–144 (1996).
14. W. J. Pasman, M. S. Westerperp-Plantenga, and W. H. M. Saris, The effectiveness of
long-term supplementation of carbohydrate, chromium, fibre, and caffeine on weight
maintenance, Int. J. Obes. Related Metab. Disord. 21, 1143–1151 (1997).
15. W. W. Campbell, L. J. Joseph, S. L. Davey, D. Cyr-Campbell, R. A. Anderson, and W. J.
Evans, Effects of resistance training and chromium picolinate on body composition and
skeletal muscle in older men, J. Appl. Physiol. 86, 29–39 (1999).
16. R. I. Press, J. Geller, and G. W. Evans, The effect of chromium picolinate on serum cholesterol and apolipoprotein fractions in human subjects, West. J. Med. 152, 41–45 (1990).
17. R. B. Krieder, R. Klesges, K. Harmon, et al., Effects of ingesting supplements designed
to promote lean muscle tissue accretion on body composition during resistance training, Int. J. Sports Nutr. 6, 234–246 (1996).
18. J. B. Vincent, The potential value and toxicity of chromium picolinate as a nutritional
supplement, weight loss agent and muscle development agent, Sports Med. 33, 213–230
(2003).
19. M. H. Pittler, C. Stevinson, and E. Ernst, Chromium picolinate for reducing body
weight: meta-analysis of randomized trials, Int. J. Obes. 27, 522–529 (2003).
20. S. L. Nissen and R. L. Sharp, Effect of dietary supplements on lean mass and strength
gains with resistance exercise: a meta-analysis, J. Appl. Physiol. 94, 651–659 (2003).
21. B. J. Clodfelder, B. M. Gullick, H. C. Lukaski, Y. Neggers, and J. B. Vincent, Oral administration of the biomimetic [Cr3O(O2CCH2CH3)6(H2O)3]+ increases insulin sensitivity
and improves blood plasma variables in healthy and type 2 diabetic rats, J. Biol. Inorg.
Chem. 10, 119–130 (2005).
22. B. J. Clodfelder, C. Chang, and J. B. Vincent, Absorption of the biomimetic chromium
cation triaqua-µ3-oxo-µ-hexapropionatotrichromium(III) in rats, Biol. Trace Element Res.
98, 159–169 (2004).
23. K. L. Olin, D. M. Stearns, W. H. Armstrong, and C. L. Keen, Comparative
retention/absorption of 51chromium (51Cr) from 51Cr chloride, 51Cr nicotinate and 51Cr
picolinate in a rat model, Trace Elements Electrolytes 11, 182–186 (1994).
24. R. A. Anderson, N. A. Bryden, M. M. Polansky, and K. Gauteschi, Dietary chromium
effects on tissue chromium concentrations and chromium absorption in rats, J. Trace Elements Exp. Med. 9, 11–25 (1996).
25. C. Sun, W. Zhang, S. Wang, and Y. Zhang, Effect of chromium gluconate on body weight,
serum leptin and insulin in rats, Wei Sheng Yan Jiu 29, 370–371 (2000) (in Chinese).
26. S. Wang, C. Sun, Q. Kao, and C. Yu, Effects of chromium and fish oil on insulin resistance and leptin resistance in obese developing rats, Wei Sheng Yan Jiu 30, 284–286 (2001)
(in Chinese).
27. B. M. Gullick, Ph.D. dissertation, The University of Alabama, Tuscaloosa (2005).
28. A. Earnshaw, B. N. Figgis, and J. Lewis, Chemistry of polynuclear compounds. Part VI.
Magnetic properties of trimer chromium and iron carboxylates, J. Chem. Soc. 1656–1663
(1966).
29. J. S. Striffler, J. S. Law, M. M. Polansky, S. J. Bhathena, and R. A. Anderson, Chromium
improves insulin response to glucose in rats, Metabolism 44, 1314–1320 (1995).
30. R. A. Anderson, N. A. Bryden, and M. M. Polansky, Lack of toxicity of chromium chloride and chromium picolinate in rats, J. Am. Coll. Nutr. 16, 273–279 (1997).
31. D. L. Hasten, M. Hegsted, M. J. Keenan, and G. S. Morris, Effects of various forms of
dietary chromium on growth and body composition in the rat, Nutr. Res. 17, 283–294
(1997).
32. D. L. Hasten, M. Hegsted, M. J. Keenan, and G. S. Morris, Dosage effects of chromium
picolinate on growth and body composition in the rat, Nutr. Res. 17, 1175–1186 (1997).
Biological Trace Element Research
Vol. 113, 2006
66
Bennett et al.
33. G. S. Morris, K. A. Guidry, M. Hegsted, and D. L. Hasten, Effects of dietary chromium
supplementation on cardiac mass, metabolic enzymes, and contractile proteins, Nutr.
Res. 15, 1045–1052 (1995).
34. A. Sclafani and M. Abrams, Rats show only a weak preference for the artificial sweetener aspartame, Physiol. Behav. 37, 253–256 (1986).
35. J. C. Smith, T. W. Castonguay, D. F. Foster, and L. M. Bloom, A detailed analysis of glucose and saccharin drinking in the rat, Physiol. Behav. 24, 173–176 (1980).
36. B. Beck, A. Burlet, J.-P. Max, and A. Stricker-Krongrad, Effects of long-term ingestion of
aspartame on hypothalamic neuropeptide Y, plasma leptin and body weight gain and
composition, Physiol. Behav. 75, 41–47 (2002).
37. K. P. Porikas and H. S. Koopmans, The effect of non-nutritive sweeteners on body
weight in rats, Appetite 11(Suppl. 1), 12–15 (1998).
38. B. J. Rolls, Effects of intense sweeteners on hunger, food intake, and body weight: a
review, Am. J. Clin. Nutr. 53, 872–878 (1991).
39. Y. Sun, K. Mallya, J. Ramirez, and J. B. Vincent, The biomimetic [Cr3O(O2CCH2CH3)6
(H2O)3]+ decreases plasma cholesterol and triglycerides in rats: towards chromiumcontaining therapeutics, J. Biol. Inorg. Chem. 4, 838–845 (1999).
40. Y. Sun, B. J. Clodfelder, A. A. Shute, T. Irvin, and J. B. Vincent, The biomimetic
[Cr3O(O2CCH2CH3)6(H2O)3]+ decreases plasma insulin, cholesterol and triglycerides in
healthy and type II diabetic rats but not type I diabetic rats, J. Biol. Inorg. Chem. 7,
852–862 (2002).
41. J. K. Speetjens, A. Parand, M. W. Crowder, and J. B. Vincent, Low-molecular-weight
chromium-binding substance and biomimetic [Cr3O(O2CCH2CH3)6(H2O)3]+ do not
cleave DNA under physiologically relevant conditions, Polyhedron 18, 2617–2624 (1999).
42. J. K. Speetjens, R. A. Collins, J. B. Vincent, and S. A. Woksi, The nutritional supplement
chromium(III) tris(picolinate) cleaves DNA, Chem. Res. Toxicol. 12, 483–487 (1999).
43. B. Debski, Z. Krejpcio, T. Kuryl, R. Wokciak, and M. Lipko, Biomimetic chromium(III)
complex and frutan supplementation affect insulin and membrane glucose transport in
rats, J. Trace Elements Exp. Med. 17, 206–207 (2004).
44. Z. Krejpcio, B. Debski, R. Wojciak, T. Kuryl, and M. Tubacka, Biomimetic chromium(III)
complex and fructan supplementation improve blood variables in STZ-induced diabetic rats, J. Trace Elem. Exp. Med. 17, 207–208 (2004).
45. C. M. Davis, A. C. Royer, and J. B. Vincent, Synthetic multinuclear chromium assembly
activates insulin receptor tyrosine kinase activity: functional model for low-molecularweight chromium-binding substance, Inorg. Chem. 36, 5316–5320 (1997).
46. A. A. Shute and J. B. Vincent, The stability of the biomimetic cation triaqua-µ-oxohexapropionatotrichromium(III) in vivo in rats, Polyhedron 20, 2241–2252 (2001).
47. A. A. Shute and J. B. Vincent, The fate of the biomimetic cation triaqua-µ-oxohexapropionatotrichromium(III) in rats, J. Inorg. Biochem. 89, 272–282 (2002).
Biological Trace Element Research
Vol. 113, 2006
J Biol Inorg Chem (2005) 10: 119–130
DOI 10.1007/s00775-004-0618-0
O R I GI N A L A R T IC L E
Buffie J. Clodfelder Æ Bryan M. Gullick
Henry C. Lukaski Æ Yasmin Neggers Æ John B. Vincent
Oral administration of the biomimetic [Cr3O(O2CCH2CH3)6(H2O)3]+
increases insulin sensitivity and improves blood plasma variables
in healthy and type 2 diabetic rats
Received: 22 October 2004 / Accepted: 30 November 2004 / Published online: 30 December 2004
SBIC 2004
Abstract The in vivo effects of gavage administration
of the synthetic, functional biomimetic cation
[Cr3O(O2CCH2CH3)6(H2O)3]+ to healthy and type 2
diabetic model rats are described. After 24 weeks of
treatment (0–1,000 lg Cr/kg body mass) of healthy
Sprague Dawley rats, the cation results in a lowering
(P<0.05) of fasting blood plasma low-density lipoprotein (LDL) cholesterol, total cholesterol, triglycerides,
and insulin levels and of 2-h plasma insulin and glucose
concentrations after a glucose challenge. Zucker obese
rats (a model of the early stages of type 2 diabetes) and
Zucker diabetic fatty rats (a model for type 2 diabetes)
after supplementation (1,000 lg Cr/kg) have lower
fasting plasma total, high-density lipoprotein, and LDL
cholesterol, triglycerides, glycated hemoglobin, and
insulin levels and lower 2-h plasma insulin levels in
glucose tolerance tests. The lowering of plasma insulin
concentrations with little effect on glucose concentrations suggests that the supplement increases insulin
sensitivity. The cation after 12 and 22 or 24 weeks of
administration lowers (P<0.05) fasting plasma glycated
hemoglobin levels in the Zucker diabetic and Zucker
obese rats, respectively, and thus can improve the glucose status of the diabetic models. The effects cannot be
attributed to the propionate ligand.
Supplementary material is available for this article at http://
dx.doi.org/10.1007/s00775-004-0618-0.
B. J. Clodfelder Æ B. M. Gullick Æ J. B. Vincent (&)
Department of Chemistry and Coalition for
Biomolecular Products, The University of Alabama,
Tuscaloosa, AL 35487-0336, USA
E-mail: [email protected]
Tel.: +1-205-3489203
Fax: +1-205-3489104
H. C. Lukaski
USDA ARS Grand Forks Human Nutrition Research Center,
POB 9034, Grand Forks, ND 58202-9034, USA
Y. Neggers
Department of Human Nutrition and Hospitality Management,
The University of Alabama, Tuscaloosa, AL 35487-0158, USA
Keywords Chromodulin Æ Chromium Æ Rats Æ
Cholesterol Æ Insulin Æ Type 2 diabetes
Abbreviations HDL: High-density lipoprotein Æ LDL:
Low-density lipoprotein Æ PTP 1B: Protein tyrosine
phosphatase 1B Æ SD: Standard deviation Æ VLDL: Very
low density lipoprotein Æ ZDF: Zucker diabetic
fatty Æ ZKO: Zucker obese
Introduction
The element Cr in its trivalent oxidation state is generally considered to have a role in insulin-dependent carbohydrate and lipid metabolism [1–3]. Rats, for
example, receiving a Cr deficient diet (approximately
30 lg Cr/kg diet) have elevated insulin areas (area under
a plot of insulin concentration versus time) in glucose
tolerance tests, suggesting the development of insulin
resistance [4, 5]. In humans, the only potential cases of
Cr deficiency arise from five subjects on total parenteral
nutrition, before the intravenous fluid was supplemented
with Cr [4]. The Food and Nutrition Board of the National Academy of Sciences (USA) in 2001 established
that the daily adequate intake of Cr was 35 lg for men
and 25 lg for women [6]. These recommendations were
established by using data such as those of Anderson and
Kozlovsky [7], who reported the Cr content of self-selected diets of American men and women. The average
daily Cr intake for men was 33 lg; the average for women was 25 lg. Other studies [8, 9] have shown that
humans consuming 35 lg of Cr daily are Cr-sufficient
(i.e., the Cr quantity consumed in the diet greater than
the Cr quantity lost in the stool and urine); thus, humans
consuming reasonable diets are not likely to be Crdeficient and would receive little if any benefit from
supplementation of the diet with Cr. Recent metaanalyses [10, 11] and reviews of studies of Cr dietary
supplementation [12] support this conclusion.
120
Chromium supplementation may have beneficial effects on individuals with altered carbohydrate and lipid
metabolism; however, the results are not conclusive.
Type 2 diabetes, for example, leads to increased urinary
Cr loss [13]; this increased urinary Cr loss could potentially with time lead to a decrease in Cr status or even Cr
depletion. A meta-analysis [14] of studies with diabetic
subjects revealed that ‘‘A study of 155 diabetic subjects...’’ [15] ‘‘showed that chromium reduced glucose
and insulin concentrations; the combined data from...
the other studies did not’’ and indicated that the results
on the studies with diabetic subjects were inconclusive.
The positive placebo-controlled study performed by
Anderson et al. [15] in China is the largest study of Cr
supplementation of diabetic adults. Subjects received 0,
200, or 1,000 lg Cr as chromium picolinate daily for
4 months. Cr supplementation (1,000 lg/day) improved
fasting serum glucose, insulin, glycated hemoglobin, and
total cholesterol concentrations and insulin and glucose
concentrations 2 h after a glucose challenge (P<0.05)
compared with a placebo. At the 200-lg/day dosage, Cr
supplementation improved fasting insulin and glycated
hemoglobin concentrations and insulin concentrations
2 h after a glucose challenge (P<0.05), compared with a
placebo. (The study results have been thoroughly reviewed by Hellerstein [16].) Thus, pharmacological, but
not nutritional, levels of Cr may have beneficial effects
on diabetic subjects. Anderson [17] has reviewed studies
on the effects of Cr supplementation of type 2 diabetics
and concluded that the amount of supplemental Cr was
important with a threshold of more than 200 lg Cr daily
postulated for beneficial effects.
The studies using large quantities of Cr (more than
200 lg/day) are supported by studies on rats. A trinuclear Cr(III) propionate complex, the cation
[Cr3O(O2CCH2CH3)6(H2O)3]+, has beneficial effects
[lower fasting plasma total and low-density lipoprotein
(LDL) cholesterol, triglycerides, and insulin levels] on
Zucker obese (ZKO) rats, models for early stages of type
2 diabetes [18, 19], while in this case significant effects
were also noted in healthy rats. Rats received 5, 10, or
20 lg Cr intravenously as the cation per kilogram body
mass daily for 12 or 24 weeks. Cefalu et al. [20] have
observed beneficial effects on insulin sensitivity in type 2
diabetic model rats with cardiovascular disease, but not
in healthy rats, from oral chromium picolinate administration (18 lg/kg body mass). Thus, two Cr(III) complexes in pharmacological doses have been shown to
have beneficial effects on genetic rodent models of diabetes, although only the trinuclear cation has been
shown to have an effect on healthy rats. Studies of
healthy rats supplemented with chromium picolinate
and CrCl3 (up to 100 mg Cr/kg diet for 24 weeks, a dose
larger than any of the intravenous doses of the trinuclear
cation, even after absorption is taken into consideration)
found no effects of Cr on fasting glucose, cholesterol, or
triglycerides levels [21].
Hence, only one Cr(III) complex to date has been
shown to have reproducibly significant effects on
blood plasma variables in healthy rats. The trinuclear
chromium propionate cation has other unique properties that may explain these observations. The complex has been shown to mimic the effects of the
naturally occurring, chromium-containing oligopeptide
chromodulin in that it can stimulate the tyrosine kinase activity of insulin-activated insulin receptor [22]
and is thus a functional biomimetic of chromodulin.
When introduced intravenously, the complex can enter
cells intact, i.e., in the form that can potentially increase insulin signaling. Additionally, the complex can
be recrystallized from dilute mineral acid [23] and
could potentially survive oral ingestion intact. The
complex is absorbed very efficiently, with absorption
ranging from greater than 60% at nutritionally relevant doses (approximately 4 lg Cr/kg) to 40% at
pharmacologically relevant doses (approximately 4 mg/
kg) [24]. However, the effects of oral administration of
the complex on blood variables have not previously
been examined.
Herein is reported the effects of oral administration
of the biomimetic trinuclear cation to healthy and diabetic male rats for 24 weeks. ZKO rats and Zucker
diabetic fatty (ZDF) rats were chosen to model type 2
diabetes. The ZKO rat is a genetic model for the early
stages of type 2 diabetes, characterized by insulin
insensitivity. The ZDF rat is a genetic model for type 2
diabetes and possesses elevated glucose levels and other
diabetic symptoms. These studies indicate that the trinuclear cation has significant effects in healthy and type
2 diabetic rats on insulin, triglycerides, and total and
LDL cholesterol levels and in diabetic rats on glycated
hemoglobin levels, suggesting an increase in insulin
sensitivity, and no acute toxic effects.
Materials and methods
Cr3 OðO2 CCH2 CH3 Þ6 ðH2 OÞ3 NO3
The nitrate salt of the cation was prepared as described
in the literature [25]. All operations were performed with
doubly deionized water unless otherwise noted and were
performed with plasticware whenever possible.
Animals
All rats were obtained from Charles River Laboratory.
The 5-week-old male Sprague Dawley rats and ZKO rats
were allowed to feed ad libitum on a commercial rat food
(Harland Tekland Certified LM-485 Mouse/Rat Sterilizable Diet) and tap water. The ZDF rats were allowed to
feed ad libitum on a high-fat commercial rat food (Purina
5008 Diet) and tap water; use of the high-fat food is
necessary to guarantee diabetes. The high-fat diet contained approximately 0.3 g Cr/kg, while the normal diet
contained approximately 0.4 g Cr/kg. Thus, the com-
121
mercial foods provide Cr-adequate diets. Rats were
raised in standard plastic and stainless steel cages on a 12h light–dark cycle. Solid food intake and body mass were
monitored at 4-day intervals. Thirty-two Sprague Dawley rats were divided randomly into four groups of eight.
The first group was gavaged daily with an aqueous
solution containing the biomimetic cation to give a total
amount of chromium equivalent to 250 lg Cr/kg body
weight. The second group received an aqueous solution
of the cation to give 500 lg Cr/kg body mass; the third
group received an aqueous solution of the cation to give
1,000 lg Cr/kg body mass. The last group was gavaged
with an equal volume of doubly deionized water daily
and served as the control. Sixteen ZKO and ZDF rats
were each split into two groups of eight rats. For each,
one group of eight received doubly deionized water and
served as a control, while the other group received an
aqueous solution of the cation to give 1,000 lg Cr/kg
body mass. After 24 weeks, the Sprague Dawley and
ZKO animals were sacrificed by CO2 asphyxiation; because of health concerns, the ZDF rats were sacrificed
after 22 weeks. Liver, kidney, heart, spleen, pancreas,
testes, and epididymal fat were quickly harvested and
weighed on plastic weighboats. Part of the largest lobe of
the liver and a kidney were placed into plastic screw-top
containers and frozen and stored for metal analyses. The
University of Alabama Institutional Animal Use and
Care Committee approved all experiments involving rats.
Blood chemistry
Blood (approximately 1.5 mL) was collected from tail
snips into polypropylene tubes after 4, 8, 12, 16, 20, and
24 (or 22 for ZDF) weeks of cation or H2O administration. Prior to blood collection, animals were fasted
for 12–16 h. Immediately after blood removal, 0.5 mg/
mL heparin and 10 mg/mL NaF were added to the
blood, and a small aliquot (approximately 200 lL) of
the blood was removed. The blood was next immediately
centrifuged; the blood plasma was tested for glucose,
total cholesterol, triglycerides, LDL cholesterol, and
high-density lipoprotein (HDL) cholesterol using diagnostic kits from Sigma Chemical Co. (St. Louis, MO,
USA) and for insulin using antibody-coated kits from
ICN Biomedicals (Costa Mesa, CA, USA). The aliquot
of whole blood was used for glycated hemoglobin assays; glycated hemoglobin was measured using a kit
from Biotron Diagnostics (Hemet, CA, USA) on blood
samples from weeks 4, 12, and 24 (or 22 for ZDF). UV–
vis measurements were made with a Hewlett-Packard
8453 spectrophotometer. Gamma counting was performed with a Packard Cobra II auto-gamma counter.
Glucose tolerance tests
After 6, 10, 14, 18, and 22 weeks of cation or H2O
administration, the rats were injected under the skin
with aqueous solutions of glucose (1 mg/mL) such that
each rat received 1.25 mg glucose/kg body mass [26].
After 2 h, blood (approximately 0.5 mL) was collected
from tail snips and handled as described earlier. The
plasma insulin and glucose concentrations were determined as described earlier.
Metal analyses
Samples of approximately equal mass from three randomly chosen livers and kidneys from each group were
dried. All vessels used in the drying were acid-washed.
Fe concentrations were determined by the method of
Fish [27]. Cr concentrations were determined by
graphite furnace atomic absorption spectroscopy using a
PerkinElmer AAnalyst 80. Samples were prepared utilizing the method of Miller-Ihli [28]. Sample preparation
blanks were analyzed, and all data were blank-corrected.
Analytical accuracy was monitored through periodic
analyses of certified reference materials from the National Institutes of Standards and Technology. The
analysis wavelength was 357.9 nm.
Statistical analyses
Data were stratified by weeks of dietary Cr treatment
into six groups (4, 8, 12, 16, 20, and 22 or 24 weeks). In
each group of Sprague Dawley rats, analysis of variance
was used to test the difference in mean concentrations of
plasma variables: glucose, insulin, cholesterol, triglycerides, HDL, and LDL by four levels of dietary Cr;
control or 0, 250, 500, and 1,000 lg Cr/kg body mass.
Differences in mean organ mass by four dietary Cr levels
were also tested by analysis of variance. A pooled Student’s t test was used in each group of ZKO and ZDF
rats to evaluate the difference in mean concentrations of
plasma variables between control and Cr-treated rats.
Differences in the mean organ mass and Fe and Cr levels
in the liver and kidney tissues were also analyzed by the
pooled Student’s t test. The level of significance for all
analyses was set at P £ 0.05. Data was analyzed using
SAS software (version 8.1).
Numerical values in the tables and text are presented
as mean plus/minus the standard deviation unless
otherwise indicated. The error bars in Figs. 1, 2, and 3
represent the standard error of the mean to keep overlaps to a minimum for presentation purposes.
Results
The daily food intake (not shown) and the percentage
mass gain (average mass gain=average mass on day 1
100%) of the control groups and the corresponding
trinuclear-cation-administered groups (Figs. 1, 2) were
generally statistically equivalent throughout the 24-week
period. However, the more cation the healthy Sprague
122
Fig. 1 Percentage body mass increase of healthy male Sprague
Dawley rats supplemented with the biomimetic cation. No
significant differences between control and supplemented groups
were found
Dawley rats received the lower their percentage body
mass gain tended to be. For the ZKO rats, the only
statistically significant effects were observed at the very
end of the study when rats receiving Cr had greater body
mass gains; why these rats gain weight compared with
controls is unclear. The general lack of a statistical difference in body mass with Cr-complex administration is
consistent with numerous other studies [4, 21, 28–31].
All animals appeared normal throughout, and no visible
differences were observed between the administered
groups and their controls.
Blood plasma variables were determined for all eight
groups after 4, 8, 12, 16, 20, and 24 weeks of administration (Tables 1, 2) (except the ZDF rats which were
killed after 22 weeks). Some striking similarities are
apparent between the healthy Sprague Dawley rats, the
ZKO rats, and the ZDF rats receiving the Cr complex
and their respective controls. For the Sprague Dawley
rats, after 24 weeks of administration of trimer corresponding to 500 or 1,000 lg Cr daily, plasma total
cholesterol, triglycerides, insulin, and LDL cholesterol
levels are all significantly lower than those of the control.
For the 250-lg Cr dose, total cholesterol, triglycerides,
and insulin levels are all lower than for the controls. The
reduction in insulin concentration is first observed for
the higher two doses after just 4 weeks of administration
and is also observed after 8 (1,000 lg only), 12, 16, and
20 weeks of administration. Reductions in insulin levels
Fig. 2 Percentage body mass increase of male Zucker obese (ZKO)
and Zucker diabetic fatty (ZDF) rats supplemented with the
biomimetic cation (1,000 lg/kg). *P<0.05
are also observed for the lowest dose of Cr starting at
week 12. At week 20, the highest Cr dose results in a
significantly lower insulin concentration than the other
Cr doses. Plasma triglycerides levels become significantly
lower after week 4 for the largest quantity of Cr and are
significantly lower after each subsequent 4-week period
for each dose of Cr. At week 8, the highest Cr dose
results in a significantly lower triglycerides concentration
than the other Cr doses. Plasma total cholesterol levels
are not significantly lower for any dose at week 4 but
with two exceptions are lower for all doses at all subsequent times. Plasma LDL cholesterol levels were significantly lower for at least two of the doses of the Cr
complex each time they were examined. The highest two
Cr doses generated lower LDL levels than the lowest Cr
dose at week 20. Plasma HDL levels are essentially unchanged by cation administration. As HDL levels are
unchanged and total cholesterol levels drop with cation
treatment, the total cholesterol-to-HDL cholesterol ratio
drops with treatment. Glucose concentrations for rats
receiving the cation after week 4 tend to be lower than
those of controls; however, the effects are not consistently significant. The effects on triglycerides and total
and LDL cholesterol concentrations with the accompanying lowering of insulin levels with little effect on glucose concentrations suggests that the complex is
significantly increasing insulin sensitivity in these rats.
123
Fig. 3 Cr content of rat liver and kidney after 6 months of
treatment with varying amounts of chromium complexes. a Cr
content versus daily dose of Cr; b Cr content versus daily absorbed
dose of Cr. Cr3—[Cr3O(O2CCH2CH3)6(H2O)3]+. Data for CrCl3
and Cr picolinate are adapted from Ref. [21]
The dependence of the blood variables with time was
also analyzed. At the 95% confidence limit, none of the
blood variables displayed significant time dependence
when compared with the levels of the control group.
Very large effects from Cr-complex administration
are seen for the ZKO and ZDF rats. For the ZKO rats,
plasma insulin was lower for the trimer groups after each
4-week period; the same pattern held for the ZDF rats
except no significant difference in insulin levels was
present at week 4. Triglycerides were lower for both
diabetic models receiving Cr at each time point. LDL
cholesterol was lower for the ZKO rats after weeks 8, 16,
20, and 24 and after weeks 12, 16, 20, and 22 for the
ZDF rats. Total cholesterol was lower for the ZKO rats
after weeks 8, 12, 16, 20, and 24 and after weeks 12, 16,
20, and 22 for the ZDF rats. At week 22 or 24, ZKO and
ZDF rats receiving Cr had 38 and 21% lower total
cholesterol, respectively. For HDL cholesterol, the ZKO
rats treated with cation after weeks 8, 12, 16, 20, and 24
possessed lower levels; and levels were lower after weeks
8, 12, and 22 for the ZDF rats. This lowering of HDL
cholesterol is actually a movement toward restoring
HDL levels to normal, as the diabetic model rats have
elevated HDL levels [32]. For the ZDF rats, the ratio of
total to HDL cholesterol does not appreciably change
compared with that of controls, while the ratio actually
decreases about 15% in the ZKO rats. No consistent
statistically significant effects were observed in the
plasma glucose concentrations. Overall, the diabetic
model rats seem to have significantly increased insulin
sensitivity in response to treatment with the trimer in a
fashion similar to that of the healthy Sprague Dawley
rats. Notably, the ratio of plasma total cholesterol for
the trimer-receiving rats to that of controls decreases
significantly at the 95% confidence level as a function of
time. None of the other blood variables display time
dependence.
Similar trends are observed in the results from glucose tolerance tests (Tables 3, 4). For healthy Sprague
Dawley rats, both glucose and insulin concentrations 2 h
after administration of the biomimetic trimer are consistently lower than those of controls. The effects are
most notable for insulin levels where a distinct dependence on Cr dosage is discernable. The ZKO and ZDF
rats receiving trimer had lower plasma insulin concentrations in the tests from 10 to 22 weeks of treatment,
while the ZKO rats receiving Cr also had lower 2-h
insulin concentrations after 6 weeks. In each test period,
the ZKO and ZDF rats receiving Cr had lower 2-h
plasma glucose concentrations; however, the effect was
never statistically significant. The lowering of plasma
insulin and glucose displays no time dependence at the
95% confidence limit. Thus, in all groups receiving the
Cr complex significant effects on 2-h insulin levels were
observed, indicating less insulin is required to lower
elevated glucose concentrations. Glucose concentrations
also appear to drop faster at these reduced insulin levels.
124
Table 1 Effect of the biomimetic cation on plasma variables of Sprague Dawley healthy rats after 4, 8, 12, 16, 20, and 24 weeks of oral
supplementation
Glucose
(mg/dL)
Total cholesterol
(mg/dL)
Triglycerides
(mg/dL)
Insulin
(lIU/mL)
LDL cholesterol
(mg/dL)
HDL cholesterol
(mg/dL)
Week 4
Control
+250Cr
+500Cr
+1,000Cr
79.9±41.5a
67.8±17.8a
66.3±16.0a
93.9±42.3a
98.0±13.2a
97.1±15.2a
94.4±22.1a
92.8±40.7a
67.9±13.7a
64.8±6.2a
61.8±10.1a
42.8±10.5b
52.9±10.0a
54.5±11.8a
41.4±6.9b
40.9±7.2b
143.4±57.0a
76.9±25.1b
117.8±40.0a,b
83.3±25.3b
51.8±4.9a
54.9±8.6a
55.5±12.3a
70.8±20.3b
Week 8
Control
+250Cr
+500Cr
+1,000Cr
97.4±11.5a
83.3±10.5a
88.6±20.21a
94.1±18.4a
101.4±11.5a
85.3±9.6b,c
89.9±6.2b
79.1±9.4c
85.8±11.2a
56.6±21.1b
69.4±6.4b
43.3±5.5c
69.1±7.3a
65.3±8.3a
51.3±19.2b
48.1±5.7b
104.6±17.0a
81.3±27.1b
72.5±12.4b
82.3±17.2b
62.4±8.4a,b
68.9±6.1a
52.1+6.6c
61.3±4.9b
Week 12
Control
+250Cr
+500Cr
+1,000Cr
80.2±11.5a
76.2±21.3a
70.8±28.5a
71.3±31.5a
92.3±15.3a
81.7±3.4b
79.6±10.1b
75.8±8.8b
79.8±8.2a
61.1±8.9b
58.4±11.6b
56.1±10.5b
63.6±5.9a
55.5±7.1b
56.3±8.4b
52.3±4.7b
121.6±28.4a
90.6±10.7b
92.6±15.3b
90.8±12.3b
58.0±7.2a
52.1±21.2a
56.7±6.9a
59.9±5.4a
Week 16
Control
+250Cr
+500Cr
+1,000Cr
85.6±19.4a
81.0±35.6a
74.0±26.1a
70.0±25.5a
100.5±8.0a
90.8±12.2a,b
89.6±8.3a,b
86.9±13.9b
81.4±10.5a
72.4±11.9a,b
69.8±6.5b
68.2±13.5b
61.8±10.5a
54.7±3.8b
52.3±4.7b
52.3±6.4b
114.7±10.0a
94.5±6.9b
113.4±10.3a
93.9±11.3b
61.8±11.4a
54.6±5.5a
60.0±7.6a
57.0±13.2a
Week 20
Control
+250Cr
+500Cr
+1,000Cr
91.4±17.4a
75.6±20.2a
80.9±23.4a
82.4±28.2a
95.8±11.7a
80.4±5.2b
78.9±6.5b
76.2±8.4b
83.5±12.4a
69.4±8.7a
60.7±10.7b
57.1±14.3b
58.5±2.8a
51.9±4.4b
49.8±3.1b
43.0±9.1c
136.2±19.8a
115.8±17.4b
91.6±14.9c
90.5±13.2c
70.8±21.7a
69.5±18.3a
67.2±20.0a
60.4±17.4a
Week 24
Control
+250Cr
+500Cr
+1,000Cr
89.4±24.6a
82.4±19.9a
78.5±25.0a
79.3±24.3a
98.1±10.2a
81.6±11.7b
75.8±9.1b
74.8±14.3b
82.4±12.8a
71.1±6.2b
65.1±8.7b
60.4±11.7b
59.0±6.1a
52.4±3.8b
50.0±2.4b
48.0±5.7b
121.3±24.7a
110.8±10.3a
90.3±8.6b
89.8±13.7b
68.4±7.4a
60.3±12.7a
64.1±18.7a
62.5±15.6a
Values are means ± the standard deviation (SD); eight rats per group. For each variable, means with different superscripts are significantly
different from each other (P<0.05).
Low-density lipoprotein (LDL), high-density lipoprotein (HDL)
These studies are also consistent with the biomimetic
increasing insulin sensitivity. In this regard, the decrease
in triglycerides, LDL cholesterol, and total cholesterol
notably is consistent with the effects of insulin on type 2
diabetic human patients [33]. These studies are also
consistent with previous studies in which rats were
administered the trinuclear cation intravenously for 12
or 24 weeks [18, 19]. However, some differences occur in
the time dependence of the lowering of plasma variables
between the oral and intravenous studies. For ZKO rats
receiving the biomimetic trimer intravenously, total
cholesterol dropped as a function of time. The difference
based on the method of administration is not currently
understood but may reflect that intravenous administration provides all of the cation to the bloodstream at
essentially one time where oral administration allows the
complex to be absorbed into the blood over a period of
time as the complex passes through the gastrointestinal
tract [24].
To examine whether the administration of the
biomimetic cation was having any effect on the longterm glycemic control of the rats, fasting plasma
glycated hemoglobin concentrations were determined
after weeks 4, 12, and 22 or 24 (Table 5). Glycated
hemoglobin is formed noncatalytically, and thus slowly,
from hemoglobin and glucose; consequently, the percentage of hemoglobin that is glycated provides a ‘‘glycemic history’’ of the previous 120 days, the average
erythrocyte life span. The concentrations of glycated
hemoglobin were not affected in the healthy Sprague
Dawley rats. This is not surprising as insulin levels but
not glucose levels were affected by treatment with
chromium. However, after 12 and 22 weeks the ZDF
rats experienced a significant decrease in the percentage
of glycated hemoglobin, reaching almost a 22% drop
compared with controls by week 22. For the ZKO rats,
the glycated hemoglobin percentage is only significantly
lower at week 24 (27% lower than that of controls). This
is consistent with the tendency of the fasting plasma
glucose levels of the ZKO and ZDF rats receiving the
cation to be lower than those of their controls; (glucose
levels are lower for both types of rats at all but one time
point each and are statistically lower for the ZDF rats
receiving cation at week 16). Thus, the administration of
the biomimetic cation is successfully treating the symptoms of the type 2 diabetes, although alone it is not
125
Table 2 Effect of the biomimetic cation on plasma variables of Zucker diabetic fatty (ZDF) rats and Zucker obese (ZKO) rats after 4, 8,
12, 16, 20, and 24 (or 22 for ZDF) weeks of oral supplementation
Glucose
(mg/dL)
Total cholesterol
(mg/dL)
Triglycerides
(mg/dL)
Insulin
(lIU/mL)
LDL cholesterol
(mg/dL)
HDL cholesterol
(mg/dL)
Week 4
ZDF control
ZDF+1,000Cr
ZKO control
ZKO+1,000Cr
204.7±59.2
225.6±48.4
145.3±71.1
135.2±65.5
161.2±25.3
153.8±24.9
163.7±37.7
127.4±28.6
624.6±102.5
422.9±119.5*
436.1±56.2
343±68.7*
120.6±6.4
135.5±18.0
310.7±10.5
280.3±8.2*
122.4±57.8
150.6±56.3
135.2±46.5
100.8±29.4
102.8±23.1
105.4±21.3
138.4±25.1
143.9±18.8
Week 8
ZDF control
ZDF+1,000Cr
ZKO control
ZKO+1,000Cr
218.4±31.8
196.5±53.7
129.4±64.6
136.6±29.0
178.6±28.9
154.9±40.1
224.5±12.8
156.3±42.5*
606.9±68.2
529.7±71.3*
595.7±42.1
472.4±59.3*
115.8±9.5
103.2±5.8*
296.1±7.9
274.8±6.3*
115.9±39.2
94.6±31.5
163.2±20.4
139.5±15.2*
99.4±10.3
81.7±6.5*
124.1±21.6
103.4±16.0*
Week 12
ZDF control
ZDF+1,000Cr
ZKO control
ZKO+1,000Cr
200.6±48.7
171.3±37.2
139.5±68.3
130.6±44.2
183.4±30.9
160.2±24.7*
234.9±40.1
158.3±20.9*
583.9±61.0
499.7±59.3*
550.1±81.4
420.6±53.7*
118.4±15.9
95.6±10.0*
263.1±9.3
237.1±16.2*
124.6±29.8
99.7±15.4*
150.3±41.9
130.9±16.8
121.4±19.7
100.3±15.2*
140.6±25.8
126.9±20.5*
Week 16
ZDF control
ZDF+1,000Cr
ZKO control
ZKO+1,000Cr
236.1±40.8
198.2±37.1*
160.5±89.1
143.1±58.3
179.5±31.2
158.4±22.6*
227.3±41.6
165.0±27.6*
580.1±73.4
502.3±65.9*
563.5±70.1
499.9±67.8*
140.9±11.4
120.1±9.6*
270.4±16.8
240.5±15.9*
137.3±19.5
109.2±22.6*
160.4±47.0
110.5±21.7*
127.5±34.8
119.6±20.3
148.0±36.1
115.9±16.7*
Week 20
ZDF control
ZDF+1,000Cr
ZKO control
ZKO+1,000Cr
253.7±76.9
220.4±56.1
155.9±61.0
148.7±70.1
184.9±37.5
167.2±15.3*
231.0±48.9
157.4±29.0*
578.5±88.6
500.3±51.4*
570.1±71.3
500.2±69.1
130.1±8.4
115.6±4.2*
226.7±12.8
200.9±17.9*
144.2±26.9
116.7±18.4*
153.6±14.8
139.7±12.3*
134.8±27.6
120.8±24.5
167.9±19.0
124.1±23.8*
Week 22/24
ZDF control
ZDF+1,000Cr
ZKO control
ZKO+1,000Cr
261.5±38.2
234.7±46.5
179.1±25.8
161.9±39.6
188.8±26.4
149.3±24.2*
249.7±31.5
155.1±26.9*
581.0±41.3
512.7±39.6*
575.9±54.2
499.3±47.1*
145.6±15.2
120.9±8.6*
231.9±20.3
198.5±11.7*
139.1±13.8
110.1±12.7*
157.3±9.9
130.6±10.2*
140.5±34.9
100.6±15.0*
175.4±21.2
128.3±18.7*
Values are means±SD; eight rats per group.
*P<0.05
restoring the altered glycated hemoglobin and other
plasma variables to their healthy range. Further studies
looking at the effects of the complex in concert with
insulin therapy and other treatments will be needed to
further examine the potential of the complex to aid in
the treatment of type 2 diabetes.
Comparison of organ masses after 24 weeks of
administration revealed some statistically significant
variations from those of the controls (Tables 6, 7);
however, no trends are apparent between groups. For
example, for the healthy Sprague Dawley rats, rats
receiving the largest amount of the trimer had less epididymal fat than the control group (32%). This loss of
fat might be reflected in the trend toward body mass
reduction in this group compared with controls. Yet,
ZDF rats receiving the complex only had a smaller
pancreas. No effects were observed in the ZKO rats. In
this laboratory’s previous study with healthy Sprague
Dawley rats receiving the trinuclear cation for 12 weeks,
healthy rats receiving the trimer on average had a larger
pancreas mass and a lower testes mass compared with
those of controls [18]; in an earlier 24-week study,
healthy rats had increased heart, testes, and spleen
masses compared with those of controls, while ZKO rats
had smaller hearts, liver, testes, and kidneys [19]. None
of these effects were observed in this study. No organs
were visibly different for any of the groups.
Cr has been suspected of potentially adversely
affecting Fe metabolism [34]. Thus, the Fe and Cr levels
of liver and kidney tissue from rats of each group were
examined. No differences in the Fe content of the liver
and kidney of the rats (Table 8) were observed in the
ZDF rats receiving Cr, which had lower kidney Fe
concentrations than their controls. Surprisingly, the Cr
supplementation has no significant effect on tissue Cr
levels (Table 8). Kidney Cr concentrations are generally
greater than liver concentrations, as previously observed
in the intravenous studies [19] and in studies in which
other Cr complexes were given orally to rats [21].
As the biomimetic degrades in vivo, propionate is
likely to be released. Propionate has been proposed to be
able to lower plasma cholesterol and glucose concentrations; however, these results are quite controversial
[36]. This laboratory has also given rats sodium propionate intravenously in amounts equivalent to the propionate contained in the largest quantity of trimer used
126
Table 3 Glucose tolerance test blood variables after 2 h for healthy
male Sprague Dawley rats
Glucose
(mg/dL)
Insulin
(lIU/mL)
Week 6
Control
+250Cr
+500Cr
+1,000Cr
162.4±17.5a
158.3±25.6a
149.5±38.9a
138.6±41.1a
98.4±6.5a
98.2±15.4a
90.5±5.3b
81.3±8.7b
Week 10
Control
+250Cr
+500Cr
+1,000Cr
169.2±12.4a
142.3±15.8b
138.4±24.4b
130.1±18.6b
Week 14
Control
+250Cr
+500Cr
+1,000Cr
Table 4 Glucose tolerance test blood variables after 2 h for male
ZKO and ZDF rats
Glucose
(mg/dL)
Insulin
(lIU/mL)
Week 6
ZDF control
ZDF+1,000Cr
ZKO control
ZKO+1,000Cr
429.7±112.5
398.5±89.1
206.3±62.4
184.6±51.7
285.6±15.7
295.4±8.6
320.4±10.4
291.5±5.2*
124.5±7.7a
110.3±9.9b
103.2±12.5b,c
94.3±9.5c
Week 10
ZDF control
ZDF+1,000Cr
ZKO control
ZKO+1,000Cr
443.1±106.9
382.4±61.7
243.5±41.9
229.2±20.2
225.1±9.3
208.5±6.3*
304.7±7.4
276.9±8.1*
157.1±18.2a
140.0±15.3b
133.3±12.1b
132.4±15.7b
97.3±5.4a
90.2±7.3b
86.1±6.5b
78.0±4.7c
Week 14
ZDF control
ZDF+1,000Cr
ZKO control
ZKO+1,000Cr
429.8±56.7
390.5±71.0
300.1±88.4
249.7±56.1
233.4±16.7
199.5±12.8*
293.1±14.9
268.2±18.0*
Week 18
Control
+250Cr
+500Cr
+1,000Cr
163.8±19.7a
150.8±13.6a,b
147.2±14.9a,b
143.5±16.0b
95.6±4.9a
88.1±8.3b
80.4±7.9c
71.5±5.2d
Week 18
ZDF control
ZDF+1,000Cr
ZKO control
ZKO+1,000Cr
400.1±48.6
315.2±31.8
367.3±72.9
300.5±43.2
215.8±17.4
184.6±11.5*
287.3±19.1
261.2±15.7*
Week 22
Control
+250Cr
+500Cr
+1,000Cr
160.2±17.3a
143.5±14.9b
140.8±12.6b
138.7±16.0b
99.4±6.1a
87.5±3.2b
81.9±5.8b
75.1±6.7c
Week 22
ZDF control
ZDF+1,000Cr
ZKO control
ZKO+1,000Cr
445.7±56.1
389.4±74.3
325.8±59.5
291.6±80.4
228.6±10.7
181.3±12.2*
289.4±9.9
266.5±14.8*
Values are means±SD; eight rats per group. For each variable,
means with different superscripts are significantly different from
each other (P<0.05).
in the intravenous studies [18, 19]. Blood plasma levels
were measured after 4, 8, 12, 16, 20, and 24 weeks and 2h glucose tolerance tests were performed after weeks 6,
10, 14, 18, and 22 (Tables S1, S2). The propionate had
no significant effect on fasting plasma glucose, insulin,
triglycerides, and total or HDL cholesterol levels; similarly, the propionate had no effect on plasma glucose or
insulin levels in response to glucose administration and
no effects on body mass (Fig. S1), food intake, and organ mass (Table S3). Thus, the effects were in stark
contrast to those from the trimer.
Discussion
In only the last decade, the effects of Cr deficiency and
the effects of administration of chromium to healthy rats
have been established to an appreciable degree; the
interpretation of studies before 1990 is complicated by
methodological problems [1]. Effects from Cr supplements have been observed for otherwise healthy rats fed
an apparently Cr-deficient diet; generation of Cr-deficiency is extremely difficult, requiring strict environmental control, such as the removal of any stainless steel
objects [4, 5]. The only consistent effect of the apparent
Cr deficiency appears to be higher plasma insulin levels
in glucose tolerance tests [4, 5]. Rats on an apparently
Values are means±SD; eight rats per group.
*P<0.05
Cr-deficient and high fat diet may have higher plasma
insulin levels and in glucose tolerance tests, higher triglycerides areas [37]. In previous studies with healthy
rats on a normal diet, receiving an oral dose of Cr had
no effect on the body composition or blood variables of
the rats [4, 21, 28–31]. For example, in the largest such
study, feeding rats a diet containing up to 100 mg Cr/kg
diet as Cr picolinate or CrCl3 for 24 weeks had no effect
on body mass and tissue masses nor on serum glucose,
cholesterol, or triglycerides concentrations [21]. The rats
receiving a 100 mg Cr/kg diet received approximately
15 mg Cr/kg body mass daily [21]. This quantity is 15
times the highest dose used in the present study. Hence,
the two most common forms of Cr used in human
nutritional supplements when given to rats for 6 months
at a dose 15-fold higher than in this study had no effect
on the body composition or serum glucose, cholesterol,
or triglycerides concentrations. The difference in
absorption should also be considered here. Oral CrCl3
and Cr picolinate are absorbed with an efficiency of
approximately 0.5–2% [38, 39]. In contrast, the trinuclear Cr propionate cation is absorbed with 40–60%
efficiency [24]. After correcting for absorption efficiency,
the rats in this study and the earlier study with CrCl3
and Cr picolinate that received the maximum dose of Cr
had approximately the same quantity of Cr entering
their bloodstreams. While the trinuclear cation generated significant reductions in fasting plasma triglyce-
127
Table 5 Percentage of glycated hemoglobin of healthy Sprague Dawley, ZDF, and ZKO rats
Group
Percentage of glycated
hemoglobin (week 4)
Percentage of glycated
hemoglobin (week 12)
Percentage of glycated
hemoglobin (week 22/24)
Sprague Dawley control
Sprague Dawley+250 lg Cr
Sprague Dawley+500 lg Cr
Sprague Dawley+1,000 lg Cr
ZDF control
ZDF+1,000 lg Cr
ZKO control
ZKO+1,000 lg Cr
6.8±1.3
6.9±1.0
7.1±1.0
7.0±0.7
15.6±2.2
15.1±1.8
12.3±2.6
11.8±2.4
6.5±1.3
6.6±1.1
6.4±1.0
6.7±1.2
17.3±1.6
14.8±1.7*
14.1±2.5
12.5±2.7
6.2±0.9
7.3±1.2
6.5±1.1
6.8±1.1
17.4±2.2
14.6±1.8*
14.7±2.4
10.7±2.1*
Values are means±SD.
*P<0.05
Table 6 Percentage of relative organ mass (tissue mass=body mass 100% ) of healthy male Sprague Dawley rats after 24 weeks of
supplementation with the biomimetic cation
Tissue
Control
+250 lg Cr
+500 lg Cr
+1,000 lg Cr
Heart
Liver
Kidney
Pancreas
Testes
Epididymal fat
Spleen
0.304±0.068a
3.533±0.208a
0.723±0.060a
0.165±0.044a
0.600±0.096a
2.627±0.348a
0.179±0.042a
0.286±0.130a
3.342±0.273a
0.665±0.095a
0.172±0.095a
0.659±0.154a
2.711±0.709a
0.133±0.034a
0.333±0.074a
3.265±0.359a
0.662±0.070a
0.196±0.062a
0.628±0.123a
2.817±0.404a
0.205±0.057a
0.315±0.042a
3.219±0.201a
0.643±0.069a
0.188±0.085a
0.564±0.160a
1.787±0.641b
0.157±0.050a
Values are means±SD; eight rats per group. For each variable, means with different superscripts are significantly different from each other
(P<0.05).
Table 7 Percentage of relative organ mass (tissue mass=body mass 100% ) of male ZDF and ZKO rats after 24 weeks (ZKO) or
22 weeks (ZDF) of supplementation with the biomimetic cation
Tissue
ZDF control
ZDF+Cr
ZKO Control
ZKO+Cr
Heart
Liver
Kidney
Pancreas
Testes
Epididymal fat
Spleen
0.292±0.073
6.195±0.357
1.048±0.088
0.202±0.063
0.892±0.092
2.078±0.161
0.159±0.054
0.320±0.036
6.563±0.549
1.088±0.131
0.134±0.021*
0.847±0.065
2.095±0.140
0.146±0.020
0.214±0.025
4.579±0.997
0.646±0.077
0.120±0.020
0.448±0.051
2.988±0.419
0.099±0.016
0.221±0.020
4.467±0.559
0.615±0.041
0.135±0.023
0.435±0.026
3.040±0.352
0.090±0.024
Values are means±SD; eight rats per group.
*P<0.05
Table 8 Iron and chromium levels of liver and kidney tissues from male Sprague Dawley, ZDF, and ZKO rats
Control
+250Cr
+500Cr
+1,000Cr
ZDF control
ZDF+1,000Cr
ZKO control
ZKO+1,000Cr
Liver Fe (lg/g dry mass)
Kidney Fe (lg/g dry mass)
Liver Cr (lg/g dry mass)
Kidney Cr (lg/g dry mass)
219.9±28.0
245.8±31.9
222.5±85.7
187.7±55.8
123.4±35.0
109.7±22.3
139.3±20.2
134.3±25.0
88.17±19.1
115.0±22.4
118.4±23.1
118.7±5.4
119.4±14.6
122.3±12.6
213.1±16.0
164.9±11.9*
0.0419±0.0212
0.1230±0.1148
0.0392±0.0177
0.0420±0.0098
0.1543±0.0807
0.1007±0.1065
0.0298±0.0054
0.0642±0.0339
0.1961±0.0872
0.2143±0.1650
0.1054±0.0231
0.1381±0.0781
0.0915±0.0780
0.1428±0.0865
0.2346±0.0888
0.1416±0.0561
Values are means±SD.
*P<0.05 for ZKO and ZDF; no means were significantly different for the Sprague Dawley rats.
rides, insulin, and LDL cholesterol starting after
4 weeks of treatment and an additional lowering of total
cholesterol after 8 weeks, the two popular supplements
had no effect after 6 months of treatment. Effects on
blood variables were also observed for the smaller doses
of Cr as the biomimetic complex. Thus, the increased
128
insulin sensitivity in the healthy rats observed in this
study would appear to arise from the trinuclear Cr
propionate complex, not from Cr(III) itself. Insulinresistance results in fat cells increasing their intracellular
hydrolysis of triglycerides, releasing fatty acids into the
bloodstream; the increased flux of fatty acids stimulates
the liver to increase the formation and excretion of very
low density lipoprotein (VLDL) cholesterol and triglycerides (generating hypertriglyceridemia) [40]. In the
plasma collisions between VLDL and LDL in the presence of cholesteryl transfer protein result in exchange of
VLDL triglycerides for LDL cholesterols, leading to the
formation of small dense LDL particles. Thus, insulin
resistance results in changes in plasma lipids and increases the level of triglycerides [40]. Increases in insulin
sensitivity have the opposite effect. Hence, the changes
in plasma lipid and cholesterol concentrations also reflect increases in insulin sensitivity.
This may be explained by the biomimetic cation’s
ability to stimulate insulin receptor (in a fashion similar
to the oligopeptide chromodulin); hence, the functional
biomimetic could potentially trap insulin-stimulated
insulin receptor in vivo in its active conformation beyond its normal levels, resulting in increased insulin
signaling and subsequent cellular action (i.e., increased
insulin sensitivity). For this to be significant, the trinuclear cation must remain intact in vivo for an appreciable period of time. Two hours after intravenous
injection of rats with 51Cr-labeled trimer, approximately
90% of 51Cr in liver cells from labeled trimer is localized
to the microsomes, indicating selective transport into
these organelles [41]. The molecular weight of the species
in the microsomes is similar to that of the trimer, suggesting it may reach these organelles intact. However,
the complex has a lifetime of less than 24 h in vivo [36].
The effects seen in rats administered the biomimetic
complex daily presumably then could arise from only the
period of time in which the complex remains intact;
however, this cannot be definitely stated as the unique
ability of the complex to be absorbed from the gastrointestinal tract and then enter cells, such that higher
intracellular Cr levels are possible, could also be
responsible.
Such a case for a unique action for the trinuclear
cation cannot be made in the case of the rat models of
diabetes. Cefalu et al. [20], using JCR(LA)-cp rats (a
genetic type 2 diabetes model with cardiovascular disease) administered chromium picolinate (aqueous solution in a water bottle) at a dose of 18 lg Cr/kg body
mass, observed that Cr administration resulted in lower
fasting plasma insulin and total cholesterol levels and
lower plasma glucose and insulin levels in glucose tolerance tests, similar to the results of the current study.
However, fasting HDL levels increased in the rats
receiving Cr picolinate compared with those of controls,
in contrast to the results with the biomimetic cation
observed here and in earlier intravenous administration
studies [18, 19]. The JCR(LA)-cp rats have greatly elevated plasma HDL cholesterol levels (in a similar fash-
ion to the ZKO and ZDF rats). The disparity between
the results on HDL levels using the two sources of Cr is
difficult to explain at the current time; additionally, the
health consequences of increasing elevated HDL levels
in the JCR(LA)-cp rats with previously existing elevated
HDL levels need to be ascertained.
The safety of Cr-containing nutritional supplements
and potential therapeutic agents has been a matter of
debate [12, 42, 43], although most of the attention has
been focused on potential deleterious effects of Cr
picolinate [44–52]. The complex has been shown to be a
mutagen in cell culture [47] and in Drosophila melanogaster [51] and to give rise to oxidative damage in test
tube studies [49], cell culture studies [46], and at high
doses [52] and nutritional doses [50] in rats. In March
2003, the Expert Group on Vitamins and Minerals
determined that Cr picolinate was a potential carcinogen
and requested that the health supplement industry voluntarily withdraw the products containing the compound while consulting on a ban on the use and sale in
Great Britain [53]. One study looking for oxidative
damage in humans failed to observe any harmful effects
[54]. The potentially deleterious effects appear to be
unique to this complex compared with other supplements [42, 43]. For example, the biomimetic cation does
not give rise to DNA damage in the test tube studies [55]
and does not generate developmental delays and decreases in the number of successful progeny in Drosophila (D. Stallings, J. O’Donnell, and J.B. Vincent,
unpublished results). The lack of accumulation of Cr
from the cation in liver and kidney when given daily for
approximately half a year in large oral doses minimizes
the potential for deleterious effects; accumulation of Cr
from Cr picolinate has been described as a potential
health concern [56].
As shown in Fig. 3a, liver and kidney Cr levels for the
Sprague Dawley rats remain constant within error in
stark contrast to those when Cr picolinate and CrCl3 are
given orally to rats [21]. The Cr intake for these two
supplements was calculated assuming the food consumption rate from Ref. [21] (15 g food daily for a 100-g
rat). However, Fig. 3a does not tell the entire story; the
contrast is even more apparent when the differences in
absorption are considered—2% absorption of Cr from
Cr picolinate and CrCl3 (vide infra), and 40% absorption
of the trinuclear cation [24]. For both Cr picolinate and
CrCl3, Cr accumulates in the kidney and liver in proportion to the dose, with the greater amounts being in the
kidney. For the chloride, the increase in liver Cr levels is
small but statistically significant. The observation that
the Cr complex that is absorbed to the greatest degree
results in the least accumulation of Cr in the kidney and
liver would at first appear to be contradictory. However,
absorption and retention studies may explain this
apparent problem. Anderson and Polansky [35] have
given 51CrCl3 orally to rats and examined the distribution of 51Cr in the tissues and body fluids from 5 min to
24 h after administration. The concentration of 51Cr in
the kidney and the liver (cpm 51Cr/g tissue) was signifi-
129
cantly higher 24 h after administration than at any earlier time. In fact, the concentration was almost 10 times
higher after 24 h than after 1 h for the kidney and almost
5 times higher for the liver [35]. For the biomimetic
complex, the maximum concentrations of Cr in the liver
and kidney after an oral dose comparable to the highest
dose used in this work are reached rapidly (30 min after
injection); the concentration in either organ is several
fold lower 24 h after injection than after 30 min [24].
Consequently, while the biomimetic cation is absorbed to
a greater extent, it is not retained. This difference in 24-h
retention levels between the complexes if retention does
not change over 6 months of daily administration could
easily explain the different Cr accumulations between the
complexes observed in Fig. 3.
Current drugs for treating diabetes (in addition to
insulin treatment) fall into five categories: metformin
(whose mechanism of action is uncertain), thiazolidines
[which activate peroxisome proliferator-activated
receptor gamma (PPARc)], a-glucosidase inhibitors, and
the last two sulfonylureas and non-sulfonylurea insulin
secretagogues (which increase insulin secretion by the
pancreas) [57]. None of their modes of action resemble
those of the trinuclear Cr cation. Ross et al. [57] have
recently reviewed the types of emerging therapeutic
agents for treating type 2 diabetes. Two of these types of
agents are small-molecule insulin receptor mimetics and
inhibitors of protein tyrosine phosphatase 1B (PTP1B).
The first insulin receptor mimic was developed at Merck
Research Laboratories and was a small, nonpeptidyl
fungal metabolite [58, 59]; this compound is active in the
absence of added insulin but has toxic side effects. In
contrast Telik has developed nonpetidyl small molecules
that increase insulin receptor phosphorylation in the
presence of insulin [59, 60]. This mode of action strongly
resembles that proposed for the trinuclear Cr propionate
cation. Complexes of another transition metal have been
suggested as therapeutic agents for treating type 2 diabetes—vanadate complexes. Vanadate complexes fall
into the category of PTP1B inhibitors. Despite the nature of the vanadate complex, the active species in
phosphatase inhibition appears to be ‘‘naked’’ vanadate
[61]. In contrast to the trinuclear Cr complex, vanadate
increases the basal levels of phosphorylation of insulin
receptor and insulin receptor substrate-1 and the activity
of PI3-kinase, but these levels are not further stimulated
by insulin [62]. The trinuclear Cr complex does not inhibit PTP1B activity or the activity of other phosphatases examined to date [63].
concentrations as well were lower in 2-h glucose tolerance tests for rats receiving the cation. The effects of the
biomimetic compound on insulin, cholesterol, and triglycerides on the healthy rats suggest that the trinuclear
complex serves not as simply a chromium source but
possesses an intrinsic activity, in contrast to other
sources of chromium previously examined. The complex
also lowers blood plasma glycated hemoglobin levels in
the rat models of diabetes, indicating the complex can be
used to improve the status of the diabetic animals. Also
no toxic effects were observed for supplementation with
the trinuclear cation. These results suggest that the
compound may have potential as a therapeutic agent.
Acknowledgements The authors wish to thank Christine Chang,
Jelena Hamilton, Allison Pickering, and James A. Neville and the
staff of The University of Alabama Animal Care Facility for
assistance with the rat studies. Funding was provided by the National Institutes of Health (DK62094–01) (J.B.V.).
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Conclusion
As demonstrated in Tables 1 and 2, the functional
biomimetic has a striking effect on fasting plasma triglycerides, total cholesterol, LDL cholesterol, and
insulin levels over 24 weeks of administration in healthy
Sprague Dawley, ZKO, and ZDF rats. Insulin concentrations for all the rats and for the healthy rats’ glucose
22.
23.
24.
25.
26.
Vincent JB (2001) Polyhedron 20:1–26
Vincent JB (2000) Acc Chem Res 33:503–510
Lukaski HC (1999) Ann Rev Nutr 19:279–301
Striffler JS, Law JS, Polansky MM, Bhathena SJ, Anderson RA
(1995) Metabolism 44:1314–1320
Striffler JS, Polansky MM, Anderson RA (1999) Metabolism
48: 1063–1068
Trumbo P, Yates AA, Schlicker S, Poos M (2001) J Am Diet
Assoc 101:294–301
Anderson RA, Kozlovsky AS (1985) Am J Clin Nutr 41:1177–
1183
Bunker VW, Lawson MS, Delues HT, Clayton BE (1984) Am J
Clin Nutr 39:797–802
Offenbacher EG, Spencer H, Dowling HJ, Pi-Sunyer FX (1986)
Am J Clin Nutr 44:77–82
Pittler MH, Stevinson C, Ernst E (2003) Int J Obesity 27:522–
529
Nissen SL, Sharp RL (2003) J Appl Physiol 94:651–659
Vincent JB (2003) Sports Med 33:213–230
Morris BW, MacNeil S, Hardisty CA, Heller S, Burgin C, Gray
TA (1999) J Trace Elem Med Biol 13:57–61
Althius MD, Jordan NE, Ludington EA, Wittes JT (2002) Am
J Clin Nutr 76:148–155
Anderson RA, Cheng NC, Bryden NA, Polansky MM, Cheng
N, Chi J, Feng J (1997) Diabetes 46:1786–1791
Hellerstein MK (1998) Nutr Rev 56:302–306
Anderson RA (1998) J Am Coll Nutr 17:548–555
Sun Y, Mallya K, Ramirez J, Vincent JB (1999) J Biol Inorg
Chem 4:838–845
Sun Y, Clodfelder BJ, Shute AA, Irvin T, Vincent JB (2002) J
Biol Inorg Chem 7:852–862
Cefalu WT, Wang ZQ, Zhang XH, Baldor LC, Russell JC
(2002) J Nutr 132:1107–1114
Anderson RA, Bryden NA, Polansky MM (1997) J Am Coll
Nutr 16:273–279
Davis CM, Royer AC, Vincent JB (1997) Inorg Chem 36:5316–
5320
Johnson MK, Powell DB, Cannon RD (1981) Spectrochim
Acta 37A: 995–1006
Clodfelder BJ, Chang C, Vincent JB (2004) Trace Elem Biol
Res 98:159–170
Earnshaw A, Figgis BN, Lewis J (1966) J Chem Soc A 1656–
1663
Schwarz K, Mertz W (1959) Arch Biochim Biophys 85:292–295
130
27. Fish WW (1988) Methods Enzymol 158:357–364
28. Miller-Ihli NJ (1996) J Food Comp Anal 9:290–300
29. Hasten DL, Hegsted M, Keenan MJ, Morris GS (1997) Nutr
Res 17:283–294
30. Hasten DL, Hegsted M, Keenan MJ, Morris GS (1997) Nutr
Res 17:1175–1186
31. Morris GS, Guidry KA, Hegsted M, Hasten DL (1995) Nutr
Res 15: 1045–1052
32. Sparks JD, Shaw WN, Corsetti JP, Bolognino M, Pesek JF,
Sparks CE (2000) Metabolism 49:1424–1430
33. Brunzell JD, Chait A (1990) Lipoprotein pathophysiology and
treatment. In: Rifkin H, Porte D Jr (eds) Ellenberg and Rifkin’s
diabetes mellitus: theory and practice. Elsevier, New York, pp
756–767
34. Lukaski HC, Bolonchuk W, Siders WA, Milne DB (1996) Am J
Clin Nutr 63:954–965
35. Anderson RA, Polansky MM (1995) Biol Trace Elem Res
50:97–108
36. Shute AA, Chakov NE, Vincent JB (2001) Polyhedron
20:2241–2252
37. Striffler JS, Polansky MM, Anderson RA (1998) Metabolism
47: 396–400
38. Olin KL, Stearns DM, Armstrong WH, Keen CL (1994) Trace
Elem Electrolytes 11:182–186
39. Anderson RA, Bryden NA, Polansky MM, Gautschi K (1996)
J Trace Elem Exp Med 9:11–25
40. Ginsberg HN (2000) J Clin Invest 106:453–458
41. Shute AA, Vincent JB (2002) J Inorg Biochem 89:272–278
42. Vincent JB (2004) Biol Trace Elem Res 99:1–16
43. Vincent JB (2003) Sports Med 33:213–230
44. Stearns DM, Wise JP Sr, Patierno SR, Wetterhahn KE (1995)
FASEB J9: 1643–1648
45. Bagchi D, Bagchi M, Balmoori J, Ye X, Stohs SJ (1997) Res
Commun Mol Pathol Pharmacol 97:335–346
46. Bagchi D, Stohs SJ, Downs BW, Bagchi M, Preuss HG (2002)
Toxicology 180:5–22
47. Manygoats KR, Yazzie M, Stearns DM (2002) J Biol Inorg
Chem 7:791–798
48. Stearns DM, Silveira SM, Wolf KK, Luke AM (2002) Mutat
Res 513: 135–142
49. Speetjens JK, Collins RA, Vincent JB, Woski SA (1999) Chem
Res Toxicol 12:483–487
50. Mahboob L, McNeil L, Toliver T, Odgen L (2000) Toxicol Sci
66(1-S): 32
51. Hepburn DDD, Xiao J, Bindom S, Vincent JB, O’Donnell
(2003) Proc Natl Acad Sci USA 100:3766–3771
52. Hepburn DDD, Burney JM, Woski SA, Vincent JB (2003)
Polyhedron 22: 455–463
53. http://www.foodstandards.gov.uk/news. newsarchive/safetyhighdosesvitsandmins accessed on 08 May 2003
54. Kato I, Vogelman JH, Dilman V, Karkoszka J, Frenkel K,
Durr NP, Orentreich N, Toniolo P (1998) Eur J Epidemiol
14:621–626
55. Speetjens JK, Parand A, Crowder MW, Vincent JB, Woski SA
(1999) Polyhedron 18:2617-2624
56. Stearns DM, Belbruno JJ, Wetterhahn KE (1995) FASEB
J 9:1650–1657
57. Ross SA, Gulve EA, Wang M (2004) Chem Rev 104:1255–
1282
58. Zhang B, Salituro G, Szalkowski D, Li Z, Zhang Y, Royo I,
Vilella D, Diez MT, Pelaez F, Ruby C, Kendall RL, Mao X,
Griffin P, Calaycay J, Zierath JR, Heck JV, Smith RG, Moller
DE (1999) Science 284:974–977
59. Li M, Youngren JF, Manchem VP, Kozlowski M, Zhang BB,
Maddux BA, Goldfine ID (2001) Diabetes 50:2323–2328
60. Pender C, Goldfine ID, Manchem VP, Evans JL, Spevak WR,
Shi S, Rao S, Bajjalieh S, Maddux BA, Youngren JF (2002) J
Biol Chem 277: 43565–43571
61. Peters KG, Davis MG, Howard BW, Pokross M, Rastogi V,
Diven C, Greis KD, Eby- Wilkens E, Maier M, Evdokimov A,
Soper S, Genbauffe F (2003) J Inorg Biochem 96:321–330
62. Goldfine AB, Patti ME, Zuberi L, Goldstein BJ, LeBlanc R,
Landaker EJ, Jiang ZY, Willsky GR, Kahn CR (2000)
Metabolism 49:400–410
63. Goldstein BJ, Zhu L, Hager R, Zilering A, Sun Y, Vincent JB
(2001) J Trace Elem Exp Med 14:393–404
JBIC (1999) 4 : 838±845
SBIC 1999
OR IGIN AL A RT IC L E
Yanjie Sun ´ Kavita Mallya ´ Juliet Ramirez
John B. Vincent
The biomimetic [Cr3O(O2CCH2CH3)6(H2O)3]+ decreases plasma
cholesterol and triglycerides in rats: towards chromium-containing
therapeutics
Received: 27 May 1999 / Accepted: 4 October 1999
Abstract The in vivo effects of administration of the
synthetic, functional biomimetic [Cr3O(O2CCH2CH3)6
(H2O)3]+ 1 and a naturally occurring, biologically
active form of chromium, low-molecular-weight chromium-binding substance (LMWCr), to rats are
described. Given that the complexes are proposed to
function by interacting with insulin receptor, trapping
it in its active conformation, in contrast to current
chromium-containing nutrition supplements, which
only serve as sources of absorbable chromium,
changes in lipid and carbohydrate metabolism would
be expected. After 12 weeks administration (20 mg/kg
body mass), compound 1 results in 40% lower levels
of blood plasma LDL cholesterol, 33% lower levels of
total cholesterol, and significantly lower HDL cholesterol and triglyceride; these results are in stark contrast to those of administration of other forms of
Cr(III) to rats, which have no effect on these parameters. LMWCr, in contrast to 1, has no effect as it
probably is degraded in vivo or excreted. These results
are interpreted in terms of the mechanism of chromium action in response to insulin and the activation
of insulin receptor, and the potential for the rational
design of chromium-containing therapeutics is discussed.
Key words Low-molecular-weight chromium-binding
substance ´ Chromium ´ Rats ´ Cholesterol
Abbreviations GTF: glucose tolerance factor ´
HDL: high-density lipoprotein ´ LDL: low-density
lipoprotein ´ LMWCr: low-molecular-weight chromium-binding substance ´ pic: picolinate
Y. Sun ´ K. Mallya ´ J. Ramirez ´ J.B. Vincent ())
Department of Chemistry and Coalition for Biomolecular
Products, The University of Alabama, Tuscaloosa, AL
35487-0336, USA
e-mail: [email protected]
Tel.: +1-205-3489203
Fax: +1-205-3489104
Introduction
In the late 1950s and 1960s, rats fed a chromium-deficient diet were found to possess a decreased ability to
repress blood glucose concentrations, while chromium(III) ions were shown to increase the efficiency of
insulin action in rat epididymal tissue [1±5]. Since
these observations, a search has been under way to
identify the biologically active form of chromium, that
is, the biomolecule which naturally binds chromium(III) and possesses an intrinsic function associated
with insulin action in mammals [6±8]. Subsequently,
the populations of developed nations have been demonstrated to intake on average less than the recommended safe and adequate amount of chromium in
their daily diet [9, 10]. These results have been touted
by industry to justify the development of chromiumcontaining dietary supplements. Such materials may
also have potential as insulin-potentiating therapeutics
that could possibly see use in the treatment of adultonset diabetes [11]. Determining the structure, function, and mode of action of the biologically active
form of chromium could greatly aid in the rational
design of such potential therapeutics.
The first chromium-containing species proposed to
be biologically active was glucose tolerance factor
(GTF) [1, 12]. GTF was first isolated from acid-hydrolyzed porcine kidney powder, although a similar material was also isolated from yeast [1, 13]. Currently the
term GTF is usually understood to refer to only the
material isolated from yeast. GTF is absorbed better
than simple chromic salts (such as chromic chloride or
chromium alum) and potentiates insulin action in rat
epididymal tissue or isolated rat adipocytes [14]. However, kinetics studies indicate that GTF does not
intrinsically possess biological activity [15]; additionally, the material is apparently a byproduct of the acid
hydrolysis step used in its purification [16]. Low-molecular-weight chromium-binding substance (LMWCr)
has been isolated from porcine kidney and porcine
839
kidney powder; acid hydrolysis of the isolated oligopeptide results in smaller chromium complexes resembling GTF [16].
Yeast GTF was proposed to be composed of
chromic ion, nicotinic acid, and the amino acids glycine, glutamic acid, and cysteine [13]. While these
results have not been reproducible in some laboratories [17±21], this report stimulated an intense interest
in the synthesis of chromic-nicotinate complexes
[22±25], some of which have been patented as nutritional supplements (e.g. US patent 5,194,615). The
proposed identification of nicotinic acid (3-carboxypyridine) also stimulated investigations of complexes of
chromium(III) with the related pyridinecarboxylic
acids picolinic acid (2-carboxypyridine) and isonicotinic acid (4-carboxypyridine) [26±28]. As a result,
chromium(III) tris(picolinate), Cr(pic)3, (US patent
4,315,927) has become a very popular nutritional supplement and is being tested as a therapeutic for the
treatment of symptoms of adult-onset diabetes. It is
available over-the-counter in the form of pills, chewing gums, sport drinks, and nutrition bars. Cr(pic)3 is
also a well-absorbed form of chromium and has been
proposed to be the naturally occurring, biologically
active form of chromium [29]; however, there is no
evidence that Cr(pic)3 occurs naturally in vivo.
In the last decade, a number of investigators have
examined the effects of administering Cr(pic)3 [and in
some cases other forms of chromium(III)] to rats on
regular diets [30±33]. After an initial preliminary
report which suggested beneficial effects on blood variables [30], detailed examinations of the effect of
Cr(pic)3 administration in amounts up to 1500 mg
Cr/kg feed for up to 24 weeks have found no acute
toxic effects [31±33]. However, the compound and
other chromium sources examined [most notably "Cr
nicotinate" and chromium(III) chloride] also had no
effect on body mass, percentage lean or fat content,
tissue size (heart, testes, liver, kidney, muscle, epididymal fat, spleen, and kidney), or blood variables (fasting glucose, insulin, cholesterol, etc.). No differences
in the gross histology of the liver or kidney [organs
where chromium(III) preferentially accumulated] were
found, although chromium did accumulate in these
organs [33]. Another study compared the effects of a
Cr-deficient rat diet with diets supplemented with 10
different sources of chromium, including allowing rats
to live in stainless steel cages. The Cr sources had no
effect on body mass; all but one source decreased
epididymal fat. Testes and liver masses tended to be
lowered, whereas kidney, heart, and spleen masses
were not significantly altered. Supplemental Cr had no
effect on serum triglycerides or cholesterol, and only
one source resulted in lower serum glucose [34]. While
these studies did not manifest any acute toxicity, the
lack of beneficial effects of Cr(pic)3 supplementation
on growth, fat content, or glucose, insulin, or cholesterol concentrations raises questions about its therapeutic potential. Recently, the safety of intaking
Cr(pic)3 has been questioned, especially in regards to
its potential to cause clastogenic damage [35, 36]. At
physiologically relevant concentrations of chromium
(120 nM) and biological reductants such as ascorbic
acid and thiols (5 mM), Cr(pic)3 has been shown in
vitro to catalytically produce hydroxyl radicals which
cleave DNA [35]. This ability stems from the combination of chromium and picolinate; the picolinate ligands prime the redox potential of the chromic center
such that it is susceptible to reduction. The reduced
chromous species interacts with dioxygen to produce
reduced oxygen species including the hydroxyl radical.
These studies are in agreement with earlier studies
which showed that mutagenic forms of chromium(III)
required chelating ligands containing pyridine-type
nitrogens coordinated to the metal [37].
Recently, a naturally occurring oligopeptide,
LMWCr, has been proposed as a candidate for the
biologically active form of chromium [6, 7, 38, 39].
Kinetics studies of insulin action on isolated rat adipocytes suggest that LMWCr possesses an intrinsic function in insulin-sensitive cells [15, 40]. The oligopeptide
appears to be part of an insulin signal amplification
mechanism [6, 7]. The oligopeptide containing four
chromic ions binds to insulin-activated insulin receptor, stimulating its tyrosine kinase activity up to eightfold with a dissociation constant of approximately
100 pM [38]. Spectroscopic studies have shown that
LMWCr possesses a multinuclear chromic assembly
where the chromic centers are bridged by anionic ligands (presumably oxide and/or hydroxide). The
assembly is supported by carboxylate groups from
aspartate and glutamate residues from the oligopeptide [41]. This discovery has spurred an interest in the
synthesis and characterization of multinuclear oxo(hydroxo)-bridged chromium(III) carboxylate assemblies
[42±45]. In 1997, such an assembly, [Cr3O
(O2CCH2CH3)6(H2O)3]+ 1, was found to mimic the
ability of LMWCr to stimulate insulin receptor kinase
activity [39]. Both LMWCr and the biomimetic 1 have
been proposed as potential nutritional supplements
and therapeutics. Both LMWCr and 1 have been
shown not to lead to DNA cleavage [46]. The synthetic complex has several potential benefits over the
natural material: it is inexpensive to synthesize and
can be readily prepared in bulk. LMWCr is susceptible to hydrolysis, especially in the presence of acid,
whereas the synthetic material can be recrystallized
from dilute mineral acid [47] and could potentially
survive oral ingestion. After the insulin signaling
event, LMWCr may be excreted in the urine [48±50],
and it is possible the body might target the material
for excretion rather than absorption.
Herein is reported the effects of administration of
LMWCr and compound 1 to male rats, which indicate
that 1 has significant effects on triglyceride and cholesterol levels in rats and no acute toxic effects.
840
Materials and methods
LMWCr and [Cr3O(O2CCH2CH3)6(H2O)3]+
LMWCr was obtained as previously described [41]. Oligopeptide
concentration was determined by the fluorescamine procedure
of Undenfriend and co-workers [51] using glycine as standard or
by measuring chromium content by the diphenylcarbazide
method [52] using the method of standard addition to minimize
any matrix effects and assuming a chromium to oligopeptide
ratio of 4 : 1. The nitrate salt of compound 1 was prepared as
described in the literature [53], although it was originally incorrectly formulated as [Cr3(OH)2(O2CCH2CH3)6]NO3 (for a discussion see [39]). Its integrity was established by electronic,
NMR, and IR spectroscopic studies [54±56]. For all experiments,
solutions of LMWCr or 1 were prepared by dilutions of more
concentrated stock solutions. Stock solutions of LMWCr were
approximately 1 mg oligopeptide per mL in concentration. All
operations were performed with doubly deionized water unless
otherwise noted and performed with plasticware whenever possible.
Animals
Four-week-old male Sprague-Dawley rats were allowed to feed
ad libitum on a commercial rat food (Harland Tekland Certified
LM-485 Mouse/Rat Sterilizable Diet) and tap water. For the
standard rat food, the estimated chromium content is on the
order of 0.2 mg/kg. Rats were raised in standard plastic and
stainless steel cages on a 12 h light-dark cycle. Solid food intake
and body mass were monitored daily. Twenty four rats were
divided randomly into three groups of eight. The first group was
injected daily in the tail vein with 200 mL of an aqueous solution
containing LMWCr to give a total amount of chromium equivalent to 20 mg Cr per kilogram body weight; the second group
received 200 mL of an aqueous solution of compound 1 to give a
total amount of chromium equivalent to 20 mg Cr per kg body
mass. The last group was injected with 200 mL of doubly deionized water daily and served as the control. Rats were injected
with chromium-containing solutions rather than fed chromiumsupplemented food to eliminate differences which might be associated with absorption of LMWCr or compound 1. Based on the
food intake of rats in this study, 20 mg Cr/kg body weight is
approximately equivalent to the Cr intake of a rat consuming a
diet containing 300 mg Cr/kg diet, a pharmacological dosage.
After 12 weeks, the animals were sacrificed by carbon dioxide
asphyxiation. Liver, kidney, heart, pancreas, testes, and epididymal fat were quickly harvested and weighed on plastic weighboats. Liver and kidney were placed into plastic screw-top containers and stored at ± 80 C for further analysis.
Metal analyses
Samples of equal mass from three randomly chosen livers from
each group were dried. All vessels used in the drying were acidwashed. Determinations of chromium and iron concentrations
were performed by Galbraith Laboratories (Knoxville, Tenn.)
using inductively coupled plasma-optical emission spectroscopy.
The accuracy of the analyses is  10%.
Histology of liver and kidney samples
The right kidney and a portion of the largest lobe of the liver
were preserved in 10% buffered formalin phosphate. The organs
from three randomly selected rats from each group were used
for further analyses. Histopathological analyses were performed
in the laboratory of Professor Thomas Bauman of the Department of Biological Sciences of The University of Alabama. Samples were stained with hematoxylin and eosin for analyses.
Statistical analyses
Statistical analyses were performed by analysis of variance. All
values are presented as mean  SEM. P Values are calculated
using standard deviations.
Results
The daily food intake (not shown) and the daily percentage mass gain (average mass gain/average mass on
day one ” 100%) of the control group and the chromium-supplemented groups (Fig. 1) were statistically
equivalent throughout the 12-week period. All animals
appeared normal throughout, and no visible differ-
Blood chemistry
Blood was collected from tail snips into polypropylene tubes
after 4, 8, and 12 weeks of Cr or H2O administration. Prior to
blood collection, animals were fasted 12±15 h. Immediately after
blood removal, 0.5 mg/mL heparin and 10 mg/mL NaF were
added to the blood. Blood was next immediately centrifuged;
the blood plasma was tested for glucose, total cholesterol, triglycerides, low density lipoprotein (LDL) cholesterol (week 12
only), and high density lipoprotein (HDL) cholesterol (week 12
only) using diagnostic kits from Sigma (St. Louis, Mo.) and insulin using double antibody or antibody coated kits from ICN Biomedicals (Costa Mesa, Calif.).
Fig. 1 Percentage body mass increase of control and rats supplemented with LMWCr and compound 1. The spikes represent
mass losses as a result of fasting before blood samples were
taken. Circles LMWCr, squares compound 1, and triangles control
841
Table 1 Percent relative organ masses (tissue mass/body
mass ” 100%) of control and rats supplemented with LMWCr or
compound 1. Values are means  SEM with eight rats per group
Tissue
Control
LMWCr
Compound 1
Heart
Liver
Kidney
Pancreas
Testes
Epididymal fat
0.317  0.012
4.141  0.22
0.474  0.014
0.147  0.016
0.769  0.023
1.071  0.12
0.318  0.017
3.851  0.17
0.463  0.025
0.270  0.46*
0.674  0.050
0.975  0.120
0.283  0.038
3.881  0.13
0.469  0.017
0.283  0.038*
0.660  0.021*
0.881  0.089
*
P< 0.05 for comparison between control and supplemented rats
ences were observed among the groups. Rats in the
LMWCr-supplemented group had a slightly (but not
statistically) higher percent mass gain over the first
few weeks, but this was a reflection of the group starting with a slightly lower initial average mass (97.5 
5.9 g versus 105.6  5.9 g for the control and 107.8 
5.8 g for compound 1-supplemented rats) and disappeared as maturity was approached. Interestingly,
during the last few weeks of the study, the average
mass of the compound 1-supplemented rats began to
diverge from that of the other groups toward lower
mass. This suggests the need for longer-term studies
to determine whether this trend would continue over
time and become statistically significant. The lack of
difference in body mass with chromium supplementation is consistent with numerous other studies [31±33,
57±59].
Organ masses after 12 weeks of supplementation
did not differ statistically from those of the control for
the heart, liver, kidney, and epididymal fat pads (Table 1). No organs were visibly different for any of the
groups. For both the LMWCr- and compound 1-sup-
plemented groups, the pancreas was enlarged, and the
testes were smaller for the compound 1 group. The
smaller testes are reminiscent of the effect of a chromium-enriched (i.e., chromium-deficient diet plus
added chromium) versus a chromium-deficient diet
[34] and interesting in relationship to a study that
observed increased sperm count and fertility in male
rats fed a Cr-enriched diet versus those on a Cr-deficient diet [60]. The increased pancreas size may also
be of note as Anderson and co-workers [61] have
reported that Cr-deficient diets have an effect on rat
pancreas. Elucidating the significance of the pancreatic hypertrophy and testicular atrophy merits and
will require further study.
Blood variables were determined for all three
groups after 4, 8, and 12 weeks of supplementation
(Table 2). At each time, the plasma glucose concentrations were equivalent among the groups. The only
significant differences observed after 4 or 8 weeks
were the triglyceride concentration for the LMWCrsupplemented group at week 4 and the plasma insulin
concentration for the compound 1-supplemented
group at week 8. However, neither of these apparent
differences is confirmed at week 12. Yet, at week 12,
a major trend is indicated as the triglyceride, total
cholesterol, LDL cholesterol, and HDL cholesterol
concentrations are all low for the compound 1-supplemented group. For example, the LDL cholesterol concentration in this group is nearly 40% less than that of
the control, and total cholesterol concentration of the
group is 33% less than that of the control. The LDL
cholesterol level for the LMWCr-supplemented rats is
also statistically lower than that of the control. The
lack of effect on fasting glucose concentrations is consistent with results of recent studies using other forms
of chromium as supplements for rats [31±33], but the
effects on cholesterol and triglyceride levels are in
Table 2 Effects of LMWCr and compound 1 on plasma variables after 4, 8, and 12 weeks of supplementation. Values are mean 
SEM with eight rats per group
Week 4
Control
LMWCr
1
Week 8
Control
LMWCr
1
Week 12
Control
LMWCr
1
*
a
Glucose
(mg/dL)
Total cholesterol
(mg/dL)
Triglycerides
(mg/dL)
Insulin
(mIU/mL)
LDL cholesterol
(mg/dL)
HDL cholesterol
(mg/dL)
116  23
157  14
168  33
71.4  5.6
78.9  4.9
65.2  3.4
81.1  5.6
65.2  4.3*
83.3  2.7
99.5  3.9
100  3.3
99.0  3.2
NDa
ND
ND
ND
ND
ND
79.4  3.1
76.6  1.9
75.5  2.9
71.6  7.3
86.7  5.7
63.8  3.3
48.5  5.8
42.3  2.7
41.8  2.1
70.0  4.3
63.3  3.5
43.3  2.1*
ND
ND
ND
ND
ND
ND
66.5  3.2
71.4  2.0
68.2  2.8
69.8  9.1
58.7  5.9
46.9  2.8*
64.3  7.8
68.9  4.7
53.2  3.7*
42.0  2.4
40.0  2.5
43.0  1.9
41.9  6.8
29.9  5.2*
25.6  2.4*
15.1  1.6
15.1  0.90
10.7  0.65*
P< 0.05 for comparison between control and supplemented rats
Not determined
842
stark contrast to these studies. The lack of effect on
glucose concentrations is also consistent with recent
studies of the effect of a Cr-deficient diet on rats versus a Cr-enriched diet and a normal diet [59, 62].
Thus, no effect on fasting glucose levels is expected
for supplementation. The lack of an effect of compound 1 supplementation on insulin levels at week 12
may be surprising given the effects on triglycerides
and cholesterol. However, studies on Cr-deficient rats
and rats on a Cr-enriched diet versus those on a normal diet by the same research laboratory [59, 62] have
shown some interesting results. Cr-enriched diets
resulted in increased plasma insulin after 12 weeks,
while the Cr-deficient diet resulted in only a small
increase versus the control; however, after 24 weeks,
all animals had normal fasting insulin concentrations
[59]. After 16 weeks in a second study, rats on the Crenriched diet had normal insulin levels while levels
were raised in Cr-deficient rats [62]. Thus, it is unclear
what should be expected for insulin levels, although
the decreased fasting insulin levels for the compound
1-supplemented group at week 8, which became normal at week 12, becomes more interesting in light of
these previous results. A tendency for Cr deficiency to
raise triglyceride levels has been observed [62].
Histopathological analyses detected no differences
in tissue samples from the kidneys or livers of any of
the three groups.
Chromium has been suspected of potentially
adversely affecting iron metabolism [8]. Thus, the iron
and chromium levels of liver tissue from rats of each
group were examined. In all cases, chromium concentrations in dried liver were less than 2 mg per g dry
mass. Twelve weeks of compound 1 supplementation
had no effect on iron concentration (6.0 ” 102 mg/g dry
mass versus 6.1 ” 102 mg/g dry mass for control livers).
Livers of rats supplemented with LMWCr had a
decreased iron content (3.8 ” 102 mg/g dry weight).
This may be significant and would be a potential concern if LMWCr had shown promise in the above studies as a therapeutic.
Discussion
Chromium(III) is a nutrient, not a therapeutic [63].
Consequently, an individual who is not deficient in
chromium would not be expected to benefit from the
intake of additional chromium. Most recent studies on
the effects of Cr(pic)3 or other chromium supplements
on healthy individuals observed no beneficial effects
from supplementation [8, 64±67], as one would expect
if such individuals are not chromium deficient. The
only conclusive demonstration of chromium deficiency
in humans is in patients receiving parenteral nutrition
[68]. Current dietary guidelines for chromium intake
are too high for both adults and infants, and no specific data are available on which to base recommendations for children and adolescents [69]. Similarly,
observations of effects of diet supplementation with
chromium on rats requires strict environmental control, such as the removal of any stainless steel objects,
to guarantee chromium deficiency [1±5, 58]. Hence,
supplementation of the diet of healthy rats on a normal diet with absorbable sources of chromium should
have no effect, as has been observed previously
[31±34, 57].
The observation that chromium(III) serves as a
nutrient and not a therapeutic is easily rationalized
based on the proposed mechanism of chromium action
[7]. The biologically active form of chromium,
LMWCr, is maintained in insulin-sensitive cells in its
apo (metal-free) form. In response to insulin, chromium concentrations in the blood decrease as chromium is moved to insulin-sensitive cells [70±72].
LMWCr has a large chromic ion binding constant and
becomes loaded with the metal ion. The holo-LMWCr
is consequently primed so that, in the presence of
insulin, insulin receptor tyrosine kinase is activated
and held in its active form; consequently, chromium is
proposed to act as part of an insulin signaling autoamplification mechanism [7]. If sufficient chromium is
maintained in the blood (probably in the form of
chromium transferrin [73]), supplemental chromium
should have no beneficial effect. However, one possible mechanism is available by which chromium could
potentially act with significant effect in healthy individuals. If the concentration of the biologically active
form of chromium, holo-LMWCr, or a functional
mimetic could be increased in insulin-sensitive cells,
this could trap insulin-stimulated insulin receptor in its
active conformation, resulting in increased insulin signaling and subsequent cellular action. The reason for
the current investigation is to test this hypothesis.
As demonstrated in Fig. 2, the functional
biomimetic, compound 1, has a striking effect on
plasma triglycerides, total cholesterol, LDL cholesterol, and HDL cholesterol levels after 12 weeks of supplementation. Interestingly, the average mass of epididymal fat follows the same pattern, being lowest (but
not statistically lower than the control) in the compound 1-supplemented group. However, the fat content is statistically lower at the level of standard error
used by some researchers (rather than the more rigorous standard deviation criterion utilized herein). The
effects by compound 1 on cholesterol and triglycerides
and potentially epididymal fat and body weight with
continuing supplementation suggest that the trinuclear
complex serves not as simply a chromium source but
possesses an intrinsic activity, in contrast to other
sources of chromium previously examined. Also no
toxic effects were observed for supplementation with
compound 1. These results suggest that compound 1
may have potential as a therapeutic; thus, the
approach of identifying and characterizing the naturally occurring, biologically active form of chromium
and using these data to develop and design potential
therapeutics seems to be rational. In contrast, LMWCr
843
Fig. 2 Effect of Cr supplementation on plasma total cholesterol,
HDL cholesterol, LDL cholesterol, and triglycerides and epididymal fat pad weight. *P< 0.05 for comparison between control
and supplemented rats
may dissociate rapidly as it hydrolyzes readily. Additionally, LMWCr is rapidly excreted, several times
faster than other forms of chromium, and has a very
large intraperitoneal LD50 in mice, compared to other
forms of chromium, of 134.9 mg/kg body mass [49],
suggesting that this naturally occurring biomolecule
may be recognized and excreted. This would prevent
its entry into insulin-dependent cells as well, resulting
in the lack of observed effects from supplementation.
According to the proposed mechanism for the mode
of action of LMWCr, the biomimetic should increase
the signaling of insulin and result in effects similar to
insulin©s action. In this regard, it is notable that the
decrease in triglycerides, LDL cholesterol, and total
cholesterol is consistent with the effects of insulin on
adult-onset diabetic patients while the decrease in
844
HDL cholesterol is not [74]. Thus, the mechanism of
these effects in compound 1-treated rats requires further investigation.
Current efforts are devoted to determine the stability of complex 1 in vivo and to test whether it actually
enters insulin-dependent cells intact. Furthermore, the
synthesis and characterization of a second generation
of chromic carboxylate assemblies that more closely
approximate the structural, spectroscopic, and functional properties of LMWCr are under way, in association with studies to determine the three-dimensional
structure of LMWCr. In such a manner, it is hoped
that better therapeutics for the treatment of adult-onset diabetes, certain cardiovascular diseases, and other
conditions can be developed.
Acknowledgements The authors wish to thank Dr. James A.
Neville of The University of Alabama Animal Care Facility for
assistance with the rat studies and Prof. Thomas Bauman for
assistance with the histopathological studies. Funding was provided by NRICGP/USDA 97-35200-4259 (J.B.V.).
References
1. Schwarz K, Mertz W (1959) Arch Biochem Biophys
85 : 292±295
2. Mertz W, Schwarz K (1959) J Physiol 196 : 614±618
3. Mertz W, Roginski EE, Schwarz K (1961) J Biol Chem
236 : 318±322
4. Mertz W, Roginski EE (1963) J Biol Chem 238 : 868±872
5. Mertz W, Roginski EE, Schroeder HA (1965) J Nutr
86 : 107±112
6. Davis CM, Vincent JB (1997) JBIC 2 : 675±679
7. Vincent JB (1999) J Am Coll Nutr 18 : 6±12
8. Lukaski HC (1999) Annu Rev Nutr 19 : 279±301
9. Anderson RA, Kozlovsky AS (1985) Am J Clin Nutr
41 : 768±771
10. Anderson RA (1994) In: Mertz W, Abernathy CO, Olin SS
(eds) Risk assessment of essential elements. ISLI Press,
Washington, pp 187±196
11. Anderson RA (1998) J Am Coll Nutr 17 : 548±555
12. Schwarz K, Mertz W (1957) Arch Biochem Biophys
72 : 515±518
13. Toepfer EW, Mertz W, Polansky MM, Roginski WW, Wolf
WR (1977) J Agric Food Chem 25 : 162±166
14. Anderson RA, Brantner JH, Polansky MM (1998) J Agric
Food Chem 26 : 1219±1221
15. Vincent JB (1994) J Nutr 124 : 117±118
16. Sumrall KH, Vincent JB (1997) Polyhedron 16 : 4171±4177
17. Gonzalez-Vergara E, Hegenauer J, Saltman P (1982) Fed
Proc 41 : 286
18. Haylock SJ, Buckley PD, Blackwell LF (1983) J Inorg Biochem 18 : 195±211
19. Mirsky N, Weiss A, Dori Z (1980) J Inorg Biochem
13 : 11±21
20. Kumpulainen J, Koivistoinen P, Lahtinen S (1978) Bioinorg
Chem 8 : 419±429
21. Votava HJ, Hahn CJ, Evans GW (1973) Biochem Biophys
Res Commun 55 : 312±319
22. Gonzalez-Vergara E, Hegenauer J, Saltman P, Sabat M,
Ibers JA (1982) Inorg Chim Acta 66 : 115±118
23. Gerdom LE, Goff HM (1982) Inorg Chem 21 : 3847±3848
24. Chang JC, Gerdom LE, Baenziger NC, Goff HM (1983)
Inorg Chem 22 : 1739±1744
25. Cooper JA, Anderson BF, Buckley PD, Blackwell LF (1984)
Inorg Chim Acta 91 : 1±9
26. Bradshaw JE, Grossie DA, Mullica DF, Pennington DE
(1988) Inorg Chim Acta 141 : 41±47
27. Stearns DM, Armstrong WH (1992) Inorg Chem
31 : 5178±5184
28. Evans GW, Pouchnik DJ (1993) J Inorg Biochem
49 : 177±187
29. Evans GW, Bowman TD (1992) J Inorg Biochem
46 : 243±250
30. Evans GW, Meyer L (1992) Age 15 : 134
31. Hasten DL, Hegsted M, Keenan MJ, Morris GS (1997) Nutr
Res 17 : 283±294
32. Hasten DL, Hegsted M. Keenan MJ, Morris GS (1997) Nutr
Res 17 : 1175±1186
33. Anderson RA, Bryden NA, Polansky MM (1997) J Am Coll
Nutr 6 : 273±279
34. Anderson RA, Bryden NA, Polansky MM, Gautschi K
(1996) J Trace Elem Exp Med 9 : 11±25
35. Speetjens JK, Collins RA, Vincent JB, Woski SA (1999)
Chem Res Toxicol 12 : 483±487
36. Stearns DM, Belbruno JJ, Wetterhahn KE (1995) FASEB J
9 : 1650±1657
37. Sugden KD, Geer RD, Rogers SJ (1992) Biochemistry
31 : 11626±11631
38. Davis CM, Vincent JB (1997) Biochemistry 36 : 4382±4385
39. Davis CM, Vincent JB (1997) Inorg Chem 36 : 5316±5320
40. Yamamoto A, Wada O, Suzuki H (1988) J Nutr 118 : 39±45
41. Davis CM, Vincent JB (1997) Arch Biochem Biophys
339 : 335±343
42. Harton A, Terrell K, Huffman JC, MacDonald C, Beatty A,
Li S, O©Connor C, Vincent JB (1997) Inorg Chem
36 : 4875±4882
43. Donald S, Terrell K, Robinson K, Vincent JB (1995)
Polyhedron 14 : 971±976
44. Ellis T, Glass M, Harton A, Folting K, Huffman JC, Vincent
JB (1995) Inorg Chem 33 : 5522±5527
45. Nagi M, Harton A, Donald S, Lee Y-S, Sabat M, O©Connor
CJ, Vincent JB (1995) Inorg Chem 34 : 3813±3820
46. Speetjens JK, Parand A, Crowder MW, Vincent JB, Woski
SA (1999) Polyhedron 18 : 2617±2624
47. Johnson MK, Powell DB, Cannon RD (1981) Spectrochim
Acta 37A : 995±1006
48. Anderson RA, Polansky MM, Bryden NA, Roginski EE,
Patterson KY, Reamer DC (1982) Diabetes 31 : 212±216
49. Wada O, Wu GY, Yamamoto A, Manabe S, Ono T (1983)
Environ Res 32 : 228±239
50. Wu GY, Wada O (1981) Sangyo Ikgaku 23 : 505±512
51. Undenfriend S, Stein S, Bohlen P, Dairman W, Leimgruber
W, Weigle M (1966) Science 178 : 871±872
52. Marczenko Z (1986) Spectrophotometric determination of
the elements. Ellis Horwood, Chichester
53. Earnshaw A, Figgis BN, Lewis J (1966) J Chem Soc A
1656±1663
54. Glass MM, Belmore K, Vincent JB (1993) Polyhedron
12 : 133±140
55. Antsyshkina AS, Porai-Koshits MA, Arkhangel©skii IV,
Diallo IN (1987) Russ J Inorg Chem 32 : 1700±1703
56. Szynanska-Buzar T, Ziolkowski JJ (1976) Sov J Coord
Chem 2 : 897±912
57. Morris GS, Guidry KA, Hegsted M, Hasten DL (1995) Nutr
Res 15 : 1045±1052
58. O©Flaherty EJ, McCarty CP (1978) J Nutr 108 : 321±328
59. Striffler JS, Law JS, Polansky MM, Bhathena SJ, Anderson
RA (1995) Metabolism 44 : 1314±1320
60. Anderson RA, Polansky MM (1981) Biol Trace Elem Res
3 : 1±5
61. Striffler JS, Polansky MM, Anderson RA (1993) J Trace
Elem Exp Res 6 : 75±81
62. Striffler JS, Polansky MM, Anderson RA (1998) Metabolism
47 : 396±400
63. Anderson RA (1989) Sci Total Environ 86 : 75±81
64. Campbell WW, Joseph LJO, Davey SL, Cyr-Campbell D,
Anderson RA, Evans WJ (1999) J Appl Physiol 86 : 29±39
845
65. Walker LS, Bemben MG, Bemben DA, Knehans AW
(1998) Med Sci Sports Exerc 30 : 1730±1737
66. Lukaski HC, Bolonchuk W, SidersWA, Milne DB (1996)
Am J Clin Nutr 63 : 954±965
67. Anderson RA (1998) Nutr Rev 56 : 266±270
68. Anderson RA (1995) Nutrition 11 : 83±86
69. Hunt CD, Stoecker BJ (1996) J Nutr 126 : S2441±S2451
70. Morris BW, Gray TA, MacNeil S (1993) Clin Chem
84 : 477±482
71. Morris BW, MacNeil S, Stanley K, Gray TA, Fraser R
(1993) J Endocrinol 139 : 339±345
72. Morris BW, Blumsohn A, MacNeil S, Gray TA (1992) Am J
Clin Nutr 55 : 989±991
73. Yamamoto A, Wada O, Ono T (1987) J Nutr 165 : 627±631
74. Brunzell JD, Chait A (1990) In: Rifkin H, Porte D Jr (eds)
Ellenberg and Rifkin©s diabetes mellitus: theory and practice. Elsevier, New York, pp 756±767