1718
Nephrol Dial Transplant (1996) 11: Editorial Comments
rats with the NO-inhibitor L-NAME plus indometh- make this worse, to stimulate endothelin out of proporacin, before the contrast agent isovue was injected. In tion to the other regulators of vascular tone? While we
this model BMS-182 874, a selective endothelin-A have to leave this question open for now, we would
receptor antagonist was partially protective against like to sum up by proposing that endothelin is highly
contrast nephropathy. Following BMS-182 874 isovue likely to contribute to contrast nephropathy, and that
inflicted a lesser decrease in renal blood flow and research of the prophylactic role of endothelin antagglomerular filtration rate upon the kidney than that onists is promising.
which occurred in the absence of endothelin inhibitor.
In addition the extent of necrosis usually observed in
the medullary thick ascending limbs was significantly References
ameliorated after BMS-182 874. In a similar fashion,
1. Kon V, Yoshioka T, Fogo A, Ichikawa I. Glomerular actions
Pollock et al. [8] worked with a model of diatrizoate
of endothelin in vivo. J Clin Invest 1989; 83: 1762-1767
induced contrast nephropathy in the rat. The animals 2. Firth JD, Ratcliffe PJ. Organ distribution of the three rat
had been preconditioned with L-NAME and indoendothelin messenger RNAs and the effects of ischemia on renal
gene expression. J Clin Invest 1992; 90: 1023-1026
methacin. Diatrizoate reduced the creatinine clearance
3. Gellai M, Jugus M, Fletcher T, De Wolf R, Nambi P. Reversal
by 40% and increased the urinary protein excretion
of postischemic acute renal failure with a selective endothelin A
rate five-fold. However FR 139 317, a specific
receptor antagonist in the rat. J Clin Invest 1994; 93: 900-904
endothelin-A receptor antagonist, given 5 min prior to 4. Nambi P, Pullen M, Jugus M, Gellai M. Rat kidney endothelin
diatrizoate, completely prevented the reduction of the
receptors in ischemia induced acute renal failure. J Pharmacol
Exp Ther 1993; 264: 345-349
creatinine clearance after diatrizoate.
5. Iwasaki S, Homma T, Kon V. Endothelin B receptor mediates
Taken together these recent observations from
auto-induction of endothelin-1. J Am Soc Nephrol 1994; 5:
different laboratories imply a multifactorial regulation
717 (abstract)
6. Cantley LG, Heyman SN, Epstein FH. Radiocontrast agents
of the vascular response to radiocontrast. In this
induce endothelin-1 mRNA in bovine endothelial cells. J Am
regulation the stimulation of endothelin is an important
Soc Nephrol 1991; 2: 660 (abstract)
contributor to adverse effects—but endothelin is only
7. Cantley LG, Spokes K, Clark B, Kuhlik A, Epstein FH. An
one of several factors involved, such as nitric oxide,
endothelin receptor antagonist decreases the effect of iothalamate
prostanoids and probably others. In other words endoon renal blood flow. J Am Soc Nephrol 1992; 3: 434 (abstract)
8. Pollock DM, Polakowski JS. Beneficial effect of ETA receptor
thelin appears to manifest deleterious consequences
blockade in radiocontrast-induced nephropathy in the rat. J Am
only when other defence mechanisms have been
Soc Nephrol 1995; 6: 1289 (abstract)
removed or blocked beforehand. It is not unlikely that
9. Bird E, Giancarli M, Barton D, Durham S. Endothelin (ET)-A
these implications derived from the experimental laborreceptors mediate ET effects in radiocontrast nephropathy. J Am
atory bear some analogy to clinical contrast nephroSoc Nephrol 1995; 6: 994 (abstract)
pathy in patients. Just as experimental models could 10. Gross P, Bussemaker E, Kapturczak M, Distler A.
Radiocontrast agents stimulate endothelin to a comparable
not be generated in normal animal kidneys the occurdegree. Nieren und Hochdruckkrankheiten 1994; 9: 423 (abstract)
rence of the clinical disease is also confined to special 11. Heyman SN, Clark BA, Cantley L, Spokes K, Rosen S, Brezis M,
settings: diabetic nephropathy, renal insufficiency,
Epstein FH. Effects of ioversol versus iothalamate on endothelin
dehydration, and some others. It seems to fit that these
release and radiocontrast nephropathy. Invest Radiol 1993; 28:
313-318
same disturbances are known for a prevailing increase
in the production of endothelin intrarenally even under 12. Oldroyd SD, Haylor JL, Morcos SK. The acute effect of ioversol
on kidney function: role of endothelin. Eur J Radiol 1995;
baseline conditions. Is it possible for radiocontrast to
19: 91-95
The Maillard reaction—from food chemistry to uraemia research
T. Henle1, R. Deppisch2 and E. Ritz2
'Forschungszentrum fiir Milch und Lebensmittel Weihenstephen, Technische Universitat Munchen, Freising; 2Klinikum der
Universitat Heidelberg, Sektion Nephrologie, Heidelberg, Germany
The Maillard reaction in food
Cooking, baking or roasting of foods is essential for
the formation of desirable flavour and colour
Correspondence
and
offprint
requests
to:
T.
Henle.
Forschungszentrum fur Milch und Lebensmittel Weihenstephen,
Technische Universitat Munchen, Vottinger Str. 45, 85350 Freising,
Germany.
(Figure 1). Among the numerous chemical reactions
which occur during this heating process, the so-called
Maillard reaction or 'non-enzymic browning' is of
particular importance. This terms stand for a complex
series of carbonyl-amine reactions, initially described
by the French biochemist Louis-Camille Maillard in
1912, who was thefirstto report that aqueous solutions
consisting of amino acids and reducing sugars turned
progressively brown during heating [1]. For the next
1719
Nephrol Dial Transplant (1996) 11: Editorial Comments
\
Fig. 1. Crackling brown skin of chicken as a hallmark of good cuisine, taken from a famous 19th century cartoon series 'Max and Moritz'
by Wilhelm Busch.
60 years, research in the field of the Maillard reaction
was a domain of food scientists, with the interest
focused mainly on the reactive behaviour of carbohydrates. Based on innumerable model studies, a great
variety of sugar degradation products have been isolated and characterized [2]. Today it is generally
accepted that in the initial stage of the reaction, the
carbonyl group of reducing carbohydrates attacks free
amino groups to form labile glycosylamines, which
afterwards are converted to the more stable 1-amino1-deoxyketoses, the so-called Amadori products.
Depending on time and temperature of the heating or
storage process, aminoketoses are degraded during the
advanced stages of the Maillard reaction to deoxyosones, highly reactive a-dicarbonyl compounds, which
afterwards, in the final stages, can induce the formation
of a large number of low-molecular heterocyclic flavor
compounds or high-molecular brown pigments, the
so-called melanoidines [3].
By comparison, little is known about the formation
of protein-bound Maillard reaction products. In the
early 1970s, the Amadori products fructoselysine la
and lactuloselysine lb (Figure 2) were isolated from
heated milk, representing the first known derivatives
resulting from the reaction of the g-amino group of
protein-bound lysine and lactose or glucose [4]. Such
lysine modification gained attention from the nutritional point of view, because the Amadori products la
and lb are not available as lysine during digestion [5].
The determination of the amount of unmodified ('available') and modified ('not available' or 'blocked') lysine
is therefore of basic interest for the correct evaluation
of the impact of food processing on the nutritional
quality of food proteins. For this purpose, numerous
analytical methods have been developed [5]. Severe
heating or prolonged storage of foods can lead to an
oxidative cleavage of Amadori compounds, giving rise
to the formation of N-e-carboxymethyllysine (CML)
2, which could be detected in acid hydrolysates of milk
proteins [6]. Only limited information is available with
respect to the main pathways of the degradation
of aminoketoses, i.e. the formation of 1,2- or
COOH
I
H 2 N-CH
(CH2)4
1a:R = H
1b:R = p-gal
NH
I
CH2
C=O
I
HO—C—H
I
H—C-OR
I
H—C-OH
I
CH2OH
COOH
I
H2N—CH
(CH 2 ) 4
NH
I
CH2
COOH
HOH2C
CHO
N
(CH2)4
CH—NH2
COOH
® I T ^-NH-(CH 2 ) 3 -CH-NH 2
NH
I
(CH 2 ) 3
CH-NH 2
COOH
N
I
I
COOH
(CH2)4
CH-NH 2
COOH
Fig. 2. Protein-bound Maillard reaction products: la, lactuloselysine; lb, fructoselysine; 2, N-£-carboxymethyllsine; 3, pyrraline; 4,
N-(5-( 5-methyl-4-oxo-5-hydroimidazol-2-yl )-L-ornithine; 5, pentosidine.
2,3-deoxyosones via enolization reactions, and their
reaction with proteins. Very recently, considerable
amounts of protein-bound pyrraline 3, an acid labile
pyrrole derivative of lysine with mutagenic and antiproteolytic properties, could be quantified in enzymic
hydrolysates of processed foods [7]. The formation of
pyrraline was strongly dependent on the heating conditions, the water content and the carbohydrate composition of the food samples, thus making this lysine
derivative a suitable tool for monitoring advanced
stages of the Maillard reaction.
A basic structure for possible derivatives arising
from the reaction between the guanido group of protein-bound arginine and a-dicarbonyls could be estab-
1720
lished with the isolation of the imidazolinone 4 [8].
During roasting of coffee as well as during baking
processes, up to 30% of the total arginine in the food
samples can be modified to compound 4.
Several studies document that proteins polymerize
when they are heated in the presence of reducing
carbohydrates [3], but the chemistry behind this phenomenon is still not known in detail. To date, only
one protein-bound cross-link amino acid resulting from
this 'sugar-induced' polymerization has been identified
[9]. Very small amounts of this compound with the
common name pentosidine 5 could be quantified in
various food samples [10], indicating that pentosidine
plays only a minor part in cross-linking of food proteins. Consequently, numerous presently unknown
amino acid derivatives resulting from the advanced
Maillard reaction must exist in processed foods.
Nephrol Dial Transplant (1996) 11: Editorial Comments
as well as in multiple tissues of patients with renal
failure, particularly in diabetic patients [16]. The concentration of circulating AGE-modified small peptides
was increased in patients with end-stage renal disease
and correlated with the severity of nephropathy [18].
The first evidence for a direct pathophysiological role
of the Maillard reaction was reported by Miyata et al.
[19]. They found that AGEs were present in /?2microglobulin isolated from amyloid fibrils of longterm haemodialysis patients with haemodialysisassociated amyloidosis. Non-enzymic modification of
proteins by glucose or other carbonyls appears to be
a controlled process and mechanisms exist for the
removal of AGE-modified proteins: a specific scavenger
receptor (RAGE) was found on macrophages, which
is able to trap and remove AGE structures [20]. This
may constitute a defence mechanism, pointing to toxicity of protein-bound Maillard products.
There is no doubt that the Maillard reaction is
The Maillard reaction in vivo
important for the understanding of various aspects of
uraemia and for certain therapeutic considerations.
The fact that a non-enzymic glycation of proteins by However, many questions are still unresolved—and in
carbohydrates also occurs within the human body was this context, it may be fruitful to have a look at what
unequivocally established with the discovery of the is going on in food chemistry. First of all, what is the
minor haemoglobin HbAlc. In the blood of diabetics, chemical basis of 'AGE'? Which individual proteinelevated levels of this protein were detected [11]. bound reaction products are formed, and what is their
Following in vitro studies, it could be shown that quantitative contribution? The amino acid derivatives
HbA lc results from the formation of a protein-bound we know today (Figure 2) surely represent only a small
Amadori product between glucose and the aminotermi- proportion of all the AGE compounds formed by
nal valine residue of the haemoglobin /?-chain [12,13]. interaction with body proteins. Most of these comFurther studies showed that 'non-enzymic glycation' pounds have been found in food proteins before their
reactions occur at the aminotermini as well at the s- formation under physiological conditions was estabamino groups of lysine residues of numerous body lished. The job of the food chemist traditionally was
proteins [3]. Based on thesefindings,the determination to isolate and to synthesize reference material and to
of glycosylated haemoglobin and serum albumin now- develop suitable analytical methods for quantification
adays has become a valuable tool for a long-term of such amino acid derivatives. It seems noteworthy
that, for example, in roasted coffee the amount of
glycaemic control of diabetic patients [3].
The Maillard reaction in vivo is not restricted to the pentosidine is roughly similar to what is found in the
early stage and the formation of Amadori products. sera of uraemic patients (Figure 3 a, b). As can be
Certain changes during aging of long-lived proteins seen from these chromatograms, numerous other preslike collagen and eye lenses, for example increasing ently unknown fluorescent derivatives can be detected
fluorescence, cross-linking and impaired digestibility, in food samples as well as in plasma. Almost certainly
could at least in part be attributed to the formation of these compounds represent structures which in the
protein-bound compounds of advanced stages of the future will also be identified among the AGE comMaillard reaction, the so-called 'advanced glycation pounds found in vivo. Can we isolate them from food
end-products' (AGEs) [14]. Carboxymethyllysine 2, matrices as well as from model mixtures, clarify their
pyrraline 3, and pentosidine 5 could be detected tissue chemical nature, andfinallydevelop sensitive analytical
and plasma samples, correlating with the age of the methods to detect them in plasma or tissue samples?
patient and the concentration of blood glucose [14-17].
Secondly, if AGE-modified proteins play a possible
The hypothesis has been proposed that the Maillard pathophysiological role, what about the intake of
reaction is responsible for the pathogenesis of certain Maillard compounds via the daily food? In heated
complications of diabetes and ageing, e.g. atheroscler- food such as bakery products up to 10% of the total
osis and cataract formation [17].
lysine can be modified to pyrraline [7]. Milk powder
may contain about 50% of its lysine in the form of the
Amadori product lactuloselysine [21]. Nothing is
Uraemia research and food chemistry—
known about the physiological fate of such modified
interdependent aspects of the Maillard reaction
food proteins. Are they proteolysed during digestion;
are the amino acid derivatives absorbed; how are they
Very recently Maillard reaction products have also excreted? What about the handling of these Maillard
been implicated in some complications of renal disease. products by the healthy and the uraemic kidney? Can
Elevated levels of pentosidine were found in the plasma we be sure that the AGEs detectable in biological
Nephrol Dial Transplant (1996) 11: Editorial Comments
1721
a>
u
0)
o
in
£
o
3
20
40
60
Time (min)
20
40
Time (min)
60
Fig. 3. Ion-exchange chromatography with direct fluorescence detection (Pen, pentosidine; PI, P2, unknown compounds), a Acid hydrolysate
of a roasted coffee sample (37 pmol pentosidine per mg of protein), b Acid hydrolysate of a human plasma sample (7.5 pmol pentosidine
permg of protein).
samples are exclusively formed in vivo—or might there
be a certain contribution from an external intake?
Questions such as these can only be resolved by
close cooperation between medical and chemical
research. Some months ago [22] it was mentioned that
the Maillard reaction will keep the dialysis community
busy for years to come—it certainly kept the food
chemists busy for more than 80 years. The key to more
complete understanding of the chemistry and consequences of the Maillard reaction, either in food or
in the human body, is a complementary approach,
joining the resources and the acumen of medicine and
food chemistry.
References
1. Maillard LC. Action des acides amines sur les sucres. Formation
des melanoidines par voi methodique. C R Acad Sci Ser 1912;
154: 66-68
2. Ledl F, Beck J, Sengl F el al. Chemical pathways of the Maillard
reaction. Prog Clin Biol Res 1989; 304: 23-42
3. Ledl F, Schleicher E. New aspects of the Maillard reaction in
food and in the human body. Angew Chem Int Ed Engl 1990;
29: 565-594
4. Moller AB, Andrews AT. Cheeseman GC. Chemical changes in
ultraheat-treated milk during storage. II. Lactuloselysine and
fructoselysine formation by the Maillard reaction. J Dairy Res
1977; 44: 267-275
5. Mauron J. The Maillard reaction in food. A critical review
from the nutritional standpoint. Prog Food Nutr Sci 1981; 5:
5-35
6. Buser W, Erbersdobler HF. Carboxymethyllysine, a new compound of heat damage in milk products. Milchwiss 1986; 41:
780-785
7. Henle T, Walter AW, Klostermeyer H. Simultaneous determination of protein-bound Maillard products by ion-exchange chromatography and photodiode array detection. In: Labuza TP,
Reineccius GA, Monnier VM, O'Brien J, Baynes JW, eds.
Maillard Reactions in Chemistry, Food, and Health. Royal Society
of Chemistry, Cambridge, 1994; 195-200
8. Henle T, Walter AW, HaeBner R, Klostermeyer H. Isolation
and identification of a protein-bound imidazolone resulting from
the reaction of arginine residues and methylglyoxal. Z Lebensm
Unters Forsch 1994; 199: 55-58
9. Sell DR, Monnier VM. Structure elucidation of a senescence
cross-link from human extracellular matrix. J Biol Chem 1989;
264: 21597-21602
10. Henle T. Proteingebundene Aminosaurederivate aus der
Reaktion von Lebensmittelproteinen und Kohlenhydraten:
Analytische und biochemische Aspekte. Lebensmittekhem 1996;
50: 3-4
11. Rahbar S, Blumenfeld O, Ranney HM. Studies of an unusual
hemoglobin in patients with diabetes mellitus. Biochem Biophys
Res Commun 1969; 36: 838-843
12. Bunn HF, Haney DN, Gabbay KH, Gallop, PM. Further
identification of the nature and the linkage of the carbohydrate
in hemoglobin Ai c . Biochem Biophys Res Commun 1975: 67:
103-109
13. Fliickiger R, Winterhalter KH. In vitro synthesis of hemoglobin
A l c . FEBS Lett 1976; 71: 356-360
14. Bucala R, Vlassara H, Cerami A. Advanced glycosylation end
products. In: Harding JJ, Crabbe MJC, eds. Post-translational
Modification of Proteins. CRC Press, Boca Raton, 1992; 54-79
15. Brownlee M. Glycation and diabetic complications. Diabetes
1994; 43: 836-841
16. Monnier VM, Sell DR. The advanced Maillard reaction in aging
and age-related diseases probed with pentosidine. In: Labuza
TP, Reineccius GA, Monnier VM, O'Brien J, Baynes JW, eds.
Maillard Reactions in Chemistry, Food, and Health. Royal Society
of Chemistry, Cambridge, 1994; 235-242
17. Monnier VM. Toward a Maillard reaction theory of aging. Prog
Clin Biol Res 1989; 304: 1-22
1722
18. Gugliucci A, Bendayan M. Renal fate of circulating advanced
glycated end products (AGE): evidence for reabsorption and
catabolism of AGE-peptides by renal proximal tubular cells.
Diabetologia 1996; 39: 149-160
19. Miyata T, Oda O, Inagi R et al. /?2-microglobulm modified with
advanced glycation end products is a major component of
hemodialysis-associated amyloidosis. / Clin Invest 1993; 92:
1243-1252
Nephrol Dial Transplant (1996) 11: Editorial Comments
20. Neeper M, Schmidt AM, Brett J el al. Cloning and expression
of a cell surface receptor for advanced glycosylation end products
of proteins. J Biol Chem 1992; 267: 14998-15004
21. Finot PA, Deutsch R, Bujard E. The extent of the Maillard
reaction during the processing of milk. Prog Food Sci Nutr 1981;
5: 345-355
22. Ritz E, Deppisch R, Nawroth P. Toxicity of uremia—does it
come from AGE? Nephrol Dial Transplant 1994; 9: 1-2
The set point of calcium—another view
A. J. Felsenfeld1 and M. Rodriguez
Department of Medicine West Los Angeles VA Medical Center and UCLA, Los Angeles, CA, USA, and Department of
Nephrology and Unit of Investigation, Hospital Universitario Reina Sofia, Cordoba, Spain
In a recent editorial in this journal, Drs Goodman and
Salusky questioned whether the set point of calcium,
as denned as the serum calcium at 50% of the maximal
PTH, provides relevant information [1]. Their editorial
concluded by stating that until evidence becomes available that in vivo estimates of alternative methods of
calculating the set point of calcium correspond to data
obtained in parathyroid tissue in vitro using the fourparameter model, the relevance of alternative methods
remains uncertain. It would seem to us that this logic
is not entirely correct since the purpose of in vitro
studies is to reflect in vivo physiology and not the
opposite; considering that the in vitro measurement of
the set point has varied considerably even with normal
parathyroid tissue [2], it is unlikely that in vitro studies
will consistently reflect in vivo PTH responses.
We are pleased to have the opportunity to address
the concerns of Drs Goodman and Salusky about the
methods for calculating the set point of calcium. As
opposed to the four-parameter model, in which the set
point is obtained at the mid-range between the minimal
and maximal PTH [1,3], the set point in our method
is obtained at 50% of the maximal PTH. In general,
we believe that both methods for the calculation of
the set point provide similar information and intend
to present data which demonstrate this conclusion.
However, it is also important to discuss conceptual
differences between the two methods and discuss why
mathematical models may not always provide the most
relevant information with respect to the abnormal
PTH-calcium curve in dialysis patients.
In the four-parameter model, the mathematical formula for calculating the set point and slope elongates
the PTH-calcium curve by setting the minimal PTH
to zero and the set point is then obtained at the 50%
value of the elongated curve. As illustrated in Figure 1,
Correspondence and offprint requests to: Arnold J. Felsenfeld MD,
c/o Unit of Investigation, Hospital Universitario Reina Sofia. Avda
Menendez Pidal S X. 14004 Cordoba. Spain.
1
Written while Visiting Professor. Faculty of Medicine, University
of Cordoba.
120
< 100
V
\
i
f 60
• - • Normal Humans
- o Dialysis Patients
5.
t 40
\
t
;
m Set Point
\
f
m
k
20 -
CO
<
•
1.1
1.2
1.3
1.4
1.5
SERUM IONIZED CALCIUM (mmol/L)
Fig. 1. The figure has been adapted from the editorial by Drs
Goodman and Salusky (reference 1) which was originally published
in reference 6. The PTH-calcium curves were obtained in normal
humans and dialysis patients with marked secondary hyperparathyroidism. The set points (calculated by the four-parameter model) of
the two PTH-calcium curves were stated to be not significantly
different.
with maximal PTH as 100% and minimal PTH at 5%
(normal human), the mid-range would be 52.5%; for
a dialysis patient in whom the minimal PTH is 20%,
the mid-range would be 60%. While mathematically
correct, the question could be asked whether the fourparameter model has the potential to distort biological
reality. As shown in Figure 1 which is reproduced from
the editorial by Drs Goodman and Salusky [1], it can
be observed that the minimal PTH, as a percentage of
maximal PTH, is considerably different between
normal humans and dialysis patients; this has been
shown many times previously [4-6]. Our inspection of
this figure suggests to us that the two PTH-calcium
curves appear to be different; thus it would not surprise
us if the set points, which are supposed to reflect the
PTH-calcium curve, were different for the two curves.
From the results published in the original article of
Drs Goodman and Salusky [6], we have found that
the P value for the comparison of the two set points
to be 0.09. However, the use of the four-parameter