Perspectives in Diabetes
Hyperglycemic Pseudohypoxia and
Diabetic Complications
JOSEPH R. WILLIAMSON, KATHERINE CHANG, MYRTO FRANGOS, KHALID S. HASAN,
YASUO IDO, TAKAHIKO KAWAMURA, JENS R. NYENGAARD, MARIA VAN DEN ENDEN,
CHARLES KILO, AND RONALD G. TILTON
Vasodilation and increased blood flow are characteristic
early vascular responses to acute hyperglycemia and
tissue hypoxia. In hypoxic tissues these vascular
changes are linked to metabolic imbalances associated
with impaired oxidation of NADH to NAD+ and the
resulting increased ratio of NADH/NAD+. In
hyperglycemic tissues these vascular changes also are
linked to an increased ratio of NADHINAW, in this case
because of an increased rate of reduction of NAD+ to
NADH. Several lines of evidence support the likelihood
that the increased cytosolic ratio of free NADH/NAD+
caused by hyperglycemia, referred to as pseudohypoxia
because tissue partial pressure oxygen is normal, is a
characteristic feature of poorly controlled diabetes that
mimics the effects of true hypoxia on vascular and
neural function and plays an impohant role in the
pathogenesis of diabetic complications. These effects of
hypoxia and hyperglycemia-induced pseudohypoxia on
vascular and neural function are mediated by a
branching cascade of imbalances in lipid metabolism,
increased production of superoxide anion, and possibly
increased nitric oxide formation. Diabetes 42:801-13,
1993
From the Departments of Pathology, Pediatrics, and Internal Medicine,
Washington University School of Medicine, St. Louis, Missouri; the Department of Internal Medicine, Chubu Rosai Hospital. Nagoya, Japan; the
Stereological Research Lab, Aarhus University, Aarhus, Denmark; and the
Department of Internal Medicine, M~ddelheimHospital, Antwerp, Belgium.
Address correspondence and reprint requests to Dr Joseph R. Williamson,
Department of Pathology, Box 81 18, 660 South Euclid Avenue, St. Louis, MO
631 10.
Received for publication 2 February 1993 and accepted in revised form 12
March 1993
PO, part~alpressure of oxygen, SDH, sorbitol dehydrogenase; LDH, lactate
dehydrogenase;GAP, glyceraldehyde 3-phosphate: 1.3-DPG, 13-bisphosphoglycerate: AR, aldose reductase; ARI, aldose reductase Inhibitor; snGBP, s n
glycerol 3-phosphate; DHAP, dihydroxyacetone phosphate; DAG. 1,2-diacylsnglycerol; GFR, glomerular filtrat~onrate; PKC, prote~nkinase C, CDP-DAG,
cytidine diphosphate-I ,2-diacyl-snglycerol,PG, prostagland~n;Ptdlns, phosphat~dylnositol;
PIP, phosphatidyl~nositol-bisphosphate;
LCA-CoA, long-chain
acyl ester of coenzyme A, LCA-C, long-cha~nacyl ester of carn~tine;FFA, free
fatty acid; O
,; superoxide anion; NO, nitrlc oxide; GSH, glutathione, G-6-P,
glucose-6-phosphate;6-PG, 6-phosphogluconate; 6-PG-DH, 6-phosphogluG-6-P-DH, glucose-6conate-dehydrogenase; R-5-P, r~bulose-5-phosphate;
phosphate-dehydrogenase,DPH, d~phenhydramine.
DIABETES, VOL. 42, JUNE 1993
v
asodilation and increased blood flow are
among the earliest, if not the earliest, vascular
changes associated with diabetes (see APPENDIX, note 1) and with acute hyperglycemia of 4to 5-h duration induced by infusion of glucose in nondiabetic humans and animals (10-18). Vasodilation and
increased blood flow also are characteristic vascular
responses to tissue hypoxia (19-22) (see APPENDIX, note
2). In hypoxic tissues these vascular changes are closely
linked to an increase in NADH/NAD+ (because of impaired oxidation of NADH to NAD+) and associated
metabolic imbalances. An increase in cytosolic-free
NADH/NAD+ also is observed in tissues exposed to
elevated glucose levels (in vivo or in vitro) at normal
tissue PO, (Table 1) (14,23-28). Linkage of this redox
imbalance to increased metabolism of glucose via the
sorbitol pathway has been documented for the asterisked tissues in Table 1. In contrast to hypoxia, the redox
imbalance induced by elevated glucose levels is largely
the result of increased oxidation of sorbitol to fructose
coupled to reduction of NAD+ to NADH in the second
step of the sorbitol pathway (Fig. 1). The considerable
number of cells and tissues listed in Table 1, together
with several lines of evidence discussed in this perspective, suggests that this hyperglycemia-induced redox
imbalance is a characteristic feature of poorly controlled
diabetes that mimics the effects of true hypoxia on
vascular and neural function and plays an important role
in the pathogenesis of diabetic complications.
A comprehensive discussion of all metabolic and functional consequences of an increased ratio of NADHI
NAD+ and of other current hypotheses for the
pathogenesis of diabetic complications is beyond the
scope and space limitations of this perspective. Attention
will be drawn, however, to evidence and hypotheses
consistent with the role of hyperglycemia-induced
pseudohypoxia in mediating diabetic complications. In
80 1
TABLE 1
Tissues manifesting hyperglycemic pseudohypoxia
Cornea*
Endoneurium*
Erythrocytes*
Glomeruli*
Granulation tissue*
1-
Hyperglycemic Pseudohypoxia
tSorbitol Oxidation
Lens'
Retina*
Islets of Langerhans
Liver
Skeletal muscle
'Hyperglycemic pseudohypoxia is linked to increased flux of
glucose via the sorbitol pathway.
j
addition, explanations will be suggested for observations
discordant with the thesis of this perspective. Recent
reviews of other hypotheses as well as supporting data
relevant to this perspective have been reported previously (1 4,26,27-42).
Several parallels between functional abnormalities associated with an increased NADHINAD' in diabetic
tissues and in hypoxic or ischemic myocardium are
depicted in Fig. 2 (albiet the metabolic imbalances that
mediate these abnormalities may differ in hypoxic and
hyperglycemic tissues). This redox imbalance in tissues
of diabetic animals appears to result largely from an
increased rate of oxidation of sorbitol to fructose by SDH.
In hypoxic and ischemic myocardium, the same redox
imbalance results from impaired mitochondrial oxidation of NADH to NAD' because of decreased PO,.
Functional consequences associated with this redox
imbalance in isolated ~erfusedhearts include 1 ) electrophysiological dysfunction (arrhythmias), 2) impaired
myocyte contractile function, 3) increased vascular permeability, and 4) increased blood flow during reflow after
mild hypoxia or brief ischemia but decreased blood flow
after prolonged hypoxia/ischemia. Corresponding functional changes in diabetic tissues include 1 ) electrophysiological dysfunction in many tissues, best characterized
in peripheral nerve and retina; 2) impaired contractile
function of heart, skeletal muscle, and vascular smooth
muscle; 3) increased vascular permeability; and 4) increased blood flow early after the onset of diabetes but
decreased blood flow later in the course of diabetes. All
of these early changes in tissues of diabetic animals are
prevented by inhibitors of AR (13,26,27,43-48) (see
APPENDIX, note 3).
An important feature of this glucose-induced redox
imbalance is that it provides an explanation for the
increased susceptibility of diabetic subjects to hypoxic
-
-
6PG-
R5P
SORBITOL
w G6P
G6PDH
6PGDH
n
SDH
AR
D-GLUCOSE
D-FRUCTOSE
w
LACTATE-PYRUVATE
LDH
FIG. 1. Reduction of glucose to sorbitol and oxldatlon of sorbitol to
fructose in the sorbitol pathway. Reduction of glucose to sorbltol by
AR Is coupled to oxidatlon of NADPH to NADP'. NADP' Is reduced
to NADPH by the hexose monophosphate pathway. Oxidation of
sorbltol to fructose by SDH is coupled to reduction of NAD' to NADH.
The cytosolic ratio of free NADHINAD' is in equlllbrlum wlth lactate
and pyruvate.
Dysfunction
Dysfunction
I
:
1.
1 2.
j 3.
j 4.
!
Electrical
Mechanical
t Vascular Permeability
t
8 Blood Flow
-
1.
2.
3.
4.
5.
Electrical
Mechanical
t Vascular Permeability
t
4 Blood Flow
Vascular Sclerosis
-
FIG. 2. Parallels between functional consequences of an increased
cytosollc NADHINAD' llnked to hyperglycemlc pseudohypoxla in
diabetic tissues and hypoxla or ischemla In myocardlai tissue.
and ischemic injury (32,35,49,50) (see APPENDIX,note 4).
Relatively mild hypoxic or ischemic episodes insufficient
to cause dysfunction in nondiabetic subjects, when superimposed on preexisting pseudohypoxia induced by
hyperglycemia, would result in a higher cytosolic NADHI
NADf that would cause tissue dysfunction and injury in
diabetic subjects.
IMBALANCES IN GLUCOSE METABOLISM CONTRIBUTING
TO PSEUDOHYPOXIA
The ratio of free NADH to NAD' modulates the activity of
many metabolic pathways (54). Because the cytosolic
pool of free NADH and NADf is in equilibrium with
cytosolic lactate and pyruvate, and because oxidation of
NADH to NADf by LDH is coupled to reduction of
pyruvate to lactate (Fig. I ) , an increased ratio of free
NADHINAD' will be reflected by an increased lactate1
pyruvate ratio. Indeed, the tissue lactatelpyruvate ratio is
a more reliable parameter of the cytosolic ratio of free
NADH/NADf than measurements of NADH and NADf
in tissue extracts because it is not possible to distinguish
between mitochondrial and cytosolic pools and between
free and bound nucleotides in tissue extracts (23). Thus,
for simplicity, in this perspective, changes in tissue
lactatelpyruvate ratios are referred to as changes in
cytosolic NADH/NADf. In like manner, an increase in
mitochondrial NADHINAD' is reflected by an increase in
the ratio of p- hydroxybutyrate/acetoacetate (23).
lncreased sorbitol pathway metabolism. lncreased
metabolism of glucose via the sorbitol pathway is probably the most important mechanism by which hyperglycemia increases cytosolic NADH/NADf (Fig. 1 ) . Because the K, of AR for glucose is very high (70 mM), the
rate of reduction of glucose to sorbitol increases with
increasing glucose levels in tissues that do not require
insulin for glucose uptake (33). lncreased sorbitol levels,
in turn, will increase the rate of oxidation of sorbitol to
fructose coupled to a reduction of NADf to NADH. At
elevated glucose levels, glucose metabolism via the
sorbitol pathway accounts for 33% of glucose consump-
DIABETES, VOL. 42, JUNE 1993
The likelihood that increased sorbitol pathway flux is
the more important of these two mechanisms for increasing cytosolic NADHINAD+ in diabetes is supported by
evidence
that inhibitors of AR prevent this glucosef
La.m&
t Sorbitol
induced
redox
imbalance in tissues in which glycolysis is
* SDHi
f FSP
increased as well as in those in which it is not
N4DH
NAD'
Pyruvate
f Fructose
(14,24,26,28),and decrease sorbitol levels and NADHI
NAD+ in tissues exposed to normal glucose levels. The
t FDP
t
latter finding implies that flux of glucose via the sorbitol
pathway modulates cytosolic NADHINAD+ even at
physiological glucose levels. Furthermore, exposure of
human erythrocytes to elevated sorbitol levels (at normal
t sew, FA
glucose levels) in vitro markedly increases cytosolic
NADHINAD+ (59). The importance of oxidation of sorbisnG3P
t LCA-C and LCA-CoA
to1 to fructose in mediating this diabetes-induced redox
imbalance and associated vascular and neural dysfuncFA
Phosphatidic a c ~ d
tion
also is supported by recent investigations with a new
J
t
inhibitor of SDH. Inhibition of SDH attenuates diabetesCDP-DAG
toaG
induced vascular dysfunction in retina, peripheral nerve,
and aorta and the increased cytosolic NADHINAD+ in
f PKC
4 Na+K+-ATPase
retina (the only tissue in which it has been examined at
this time), despite a four- to sixfold increase in tissue
DPH
sorbitol levels above those in untreated diabetic rats (60,
FIG. 3. Imbalances in glucose, sorbitol, and lipid metabolism linked to
60a). All of these observations force the conclusion that
hyperglycemic pseudohypoxla and vascular and neural dysfunction.
sorbitol pathway-linked vascular and neural dysfunction
*Pharmacological interventions that prevent glucose-Induced vascular
dysfunction include ARls, SDH inhlbltors, acetyl-L-carnltlne, PKC
are more likely the consequence of an increased rate of
inhlbitors, pyruvate, myelnositol, and the antlhistamine DPH.
oxidation of sorbitol to fructose than to putative osmotic
Substances that mimic glucose effects on vascular function are
italicized and underlined: lactate, sorbltol, DAG, and LCA-C.
stress associated with increased sorbitol levels Der se or
to metabolic imbalances linked to reduction of giucose to
sorbitol (33,36).
Many studies in diabetic animals in which ARls have
tion by the lens and 10% by human erythrocytes (24,55).
The extent to which tissue sorbitol levels are increased by failed to prevent early vascular or neural dysfunction may
elevated glucose levels is influenced by several factors be attributable to drug underdosing, i.e., failure to comincluding the ratio of AR to SDH and the availability of pletely inhibit the sorbitol pathway, despite normalization
cofactors for both enzymes (56). Thus, tissue sorbitol of tissue sorbitol levels (13,45,47). The same problem
levels are roughly proportional to the ratio of AR to SDH may explain some reports that ARls fail to prevent
metabolic imbalances (linked to early vascular and neuin diabetic rats (56).
lncreased glycolysis. An increased rate of glycolysis ral dysfunction) in cultured cells and incubated tissues.
also can alter the cytosolic redox state of NADHINAD+. The importance of documenting normalization of fructose
lncreased glucose metabolism to lactate by human levels (in addition to sorbitol levels) as a parameter of
erythrocytes (induced by elevating intracellular inorganic glucose flux via the sorbitol pathway cannot be overemphosphate, which activates phosphofructokinase) and in phasized (45,47) (see APPENDIX, note 5). Both Ao et al.
isolated perfused hearts (by increasing the work load), (45) and Cameron and Cotter (47) have reported that
even under aerobic conditions and at constant glucose doses of ARls that normalized neural fructose as well as
levels, is associated with an increased cytosolic NADHI sorbitol levels also normalized nerve conduction in diaNAD+ (57,58). Under these conditions of markedly ac- betic rats, whereas doses that normalized sorbitol but not
celerated glycolysis, oxidation of GAP to 1,3-DPG by fructose levels failed to normalize electrophysiological
GAP dehydrogenase (Fig. 3) appears to become the function. In addition, Tilton et al. (13) reported that in
rate-limiting step in glycolysis (57,58). This reaction is diabetic rats treated with two different doses of an ARI,
coupled to reduction of NAD+ to NADH and is inhibited vascular dysfunction was normalized only by the higher
by NADH. The mechanism by which an increased rate of dose despite normalization of tissue sorbitol levels by
glycolysis increases free cytosolic NADH/NAD+ is not both doses.
It is unclear whether vascular and neural dysfunction in
fully understood; it appears to result from a dysequilibrium between the rate of oxidation of GAP to 1,3-DPG chronically diabetic animals can be prevented by comand the rate of reduction of pyruvate to lactate (coupled plete inhibiton of AR or by other interventions. The
to oxidation of NADH to NAD+) (57,58). These observa- disappointing effects of ARls on complications in diations raise the possibility that the increased rate of betic humans and in chronic studies in animals may be
glycolysis in retina, glomeruli, endoneurium, and granu- explained by the fact that the studies in human subjects
lation tissue (but not lens or erythrocytes) exposed to have all been intervention trials, and it is now evident that
elevated glucose levels may contribute to an increase in even subclinical changes are very poorly reversible even
cytosolic NADHINAD+.
by normalization of glucose levels, and the doses of ARls
t Glucose
1
'
1c
3
t
\
/-
1
-3
-
DIABETES, VOL. 42, JUNE 1993
used in diabetic humans and animals may not have
completely inhibited metabolism of glucose via the sorbitol pathway.
NADH generated in the cytosol by glycolysis (coupled
to oxidation of GAP to 1,3-DPG)or by oxidation of sorbitol
to fructose must be reoxidized to NADf, or glycolysis will
be blocked at the level of GAP dehydrogenase as
indicated above. In the cytosol, NADH is oxidized to
NADf by LDH (coupled to reduction of pyruvate to
lactate) and by snG3P dehydrogenase (coupled to reduction of DHAP to snG3P) (Fig. 3). Cytosolic reducing
equivalents also are transported (by the snG3P and/or
the malate-aspartate shuttle) into mitochondria for oxidation by the electron transport chain. The limited capacity
of many cells and tissues to oxidize cytosolic NADH by
the sum of these mechanisms is well documented as
discussed above.
This discussion has emphasized hyperglycemia-induced changes in cytosolic NADHINAD' because of
evidence linking this redox imbalance to increased production of reducing equivalents by cytosolic enzymes.
An increase in cytosolic NADH/NAD+ may or may not be
accompanied by corresponding increases in mitochondrial NADH/NAD+, because in the liver of diabetic rats,
the ratio is increased in the cytosol but decreased in
mitochondria (23). The same discordance between cytosolic and mitochondrial ratios of NADH/NAD+ is observed in isolated perfused hearts subjected to an
increased work load at normal PO, and constant glucose
levels (58). In hypoxic tissues (and in cyanide-poisoned
cells) the redox imbalance originates in mitochondria but
also affects the cytosol because reducing equivalents
generated in the cytosol can no longer be oxidized in the
mitochondria.
If an increase in cytosolic NADH/NAD+ plays an
important role in mediating vascular and neural dysfunction associated with diabetes and increased sorbitol
pathway metabolism, the linkage between cytosolic ratios of lactatelpyruvate and NADH/NAD+ (depicted in
Figs. 1 and 3) suggests that glucose-induced vascular
dysfunction might be prevented simply by raising tissue
pyruvate levels sufficiently to drive oxidation of NADH to
NAD' as rapidly as NAD+ is reduced by oxidation of
sorbitol to fructose. Conversely, elevation of lactate levels
should mimic the effects of hyperglycemia on vascular
function. Both of these predictions have been confirmed
experimentally. Elevation of tissue pyruvate levels prevents glucose-induced vascular dysfunction in the tissue
chamber model as well as in nondiabetic rats with acute
hyperglycemia of -5-h duration induced by glucose
infusion (61, 61a). Pyruvate also prevents sorbitol-induced vascular dysfunction in skin chamber granulation
tissue (62). In contrast to inhibitors of AR and SDH,
pyruvate has little impact on flux of glucose via the
sorbitol pathway (63). Although pyruvate can function as
a free-radical scavenger and can be oxidized in mitochondria via the citric acid cycle, these reactions would
not decrease cytosolic NADH/NAD+ or normalize cytosolic redox-linked metabolites. In the skin chamber granulation tissue model, addition of lactate (at normal
glucose levels) increases blood flow and vascular per-
804
meability. In nondiabetic and diabetic human subjects
infusion of lactate increases renal blood flow and GFR
independent of glucose levels (64). It is of interest that
cyanide and ethanol (both of which increase tissue
NADH/NADf at normal glucose levels) also are glucomimetic in their effects on vascular function in the skin
chamber model.
The implications of these observations are that 1)
pseudohypoxia of any cause, i.e., elevated levels of
glucose, sorbitol, lactate, cyanide, or ethanol, induces
vascular responses, i.e., dysfunction (see APPENDIX, note
6), like those caused by true hypoxia; and 2) an important determinant of the susceptibility of cells/tissues to
sorbitol pathway-mediated injury is whether they can
reoxidize NADH to NADf rapidly enough to keep pace
with the increased rate of reduction of NAD+ to NADH
associated with increased flux of glucose via the sorbitol
pathway. Although it remains unclear why some cells and
tissues (i.e., retina, kidney, and nerve) appear to be more
susceptible to injury by the diabetic milieu, their vulnerability may reflect unusual metabolic, structural, and/or
functional characteristics in addition to a possible higher
content of sorbitol pathway enzymes (resulting in more
pronounced pseudohypoxia in response to elevated glucose levels). In tissues containing fructokinase (in addition to AR and SDH) fructose, produced from oxidation of
sorbitol, may be phosphorylated and enter glycolysis
directly (bypassing phosphofructokinase, which normally
limits metabolism of glucose via glycolysis).
METABOLIC CONSEQUENCES OF AN INCREASE IN
CYTOSOLIC NADHINA D'
An increase in cytosolic NADH/NAD+ may impact on the
activity of numerous cytoplasmic and mitochondrial enzymes that use NADH and NAD+ as cofactors and/or
are inhibited or activated by NADH or NAD+. Thus, an
important feature of this redox imbalance is that it has
the potential to explain many metabolic imbalances
associated with the diabetic milieu as well as the increased susceptibility of diabetic subjects to vascular
and neural injury by factors independent of the diabetic
milieu. Several imbalances in lipid metabolism linked to
hyperglycemic pseudohypoxia have been implicated as
mediators of diabetes-induced vascular and neural dysfunction. The roles of these imbalances in mediating
vascular and neural dysfunction may vary in different
cells and tissues depending on differences in their capacity for lipid uptake, synthesis, storage, and/or oxidation.
lncreased de novo synthesis of DAG. lncreased tissue
levels of DAG and/or evidence of glucose- and diabetesinduced increased de novo synthesis of DAG (and
associated activation of PKC) have been reported in
retina, isolated glomeruli, heart, cultured retinal capillary
endothelial cells, cultured aortic endothelial and smooth
muscle cells, and granulation tissue (14,26,27,61,6569). On the other hand, DAG levels were unchanged or
decreased in peripheral nerve and aorta of diabetic rats
(70,71), although PKC activity was increased in cultured
aortic smooth muscle cells exposed to elevated glucose
levels (72). An increased cytosolic NADH/NAD+ favors
DIABETES, VOL. 42, JUNE 1993
increased de novo synthesis of DAG by two mechanisms. The increase in NADH favors reduct~onof DHAP
to snG3P, which is the first step in one pathway for de
novo synthesis of DAG (Fig. 3) (61). It also inhibits GAP
dehydrogenase (Fig. 3), thereby impairing the oxidation
of GAP to 1,3-DPG,which increases availability of DHAP
(which is in equilibrium with fructose 1,6-bisphosphate
and GAP) as substrate for DAG synthesis. Activation of
PKC, presumably by DAG and/or by LCA-C (discussed
below), has been linked to many metabolic and functional vascular and neural changes in diabetes, i.e.,
decreased Na+K+-ATPase activity, increased PG synthetic activity, and vascular dysfunction in glomeruli,
aorta, peripheral nerve, and granulation tissue (61,65,
68,69).
Although the beneficial effects of myeinositol supplementation on glucose- and diabetes-induced vascular
and neural dysfunction have been attributed to normalization of Ptdlns metabolism and Na'K'-ATPase activity
(34,40), they may also (or instead) accrue from concomitant decreases in DAG levels and PKC activity, because
elevated myeinositol levels favor incorporation of DAG
(via CDP-DAG into Ptdlns [Fig. 31). This view is supported
by evidence that 1) agonist-induced increases in DAG
and CDP-DAG (resulting from hydrolysis of PIP,) (Fig. 3)
are prevented by myeinositol (73-76), 2) glucose-induced vascular dysfunction in skin chamber granulation
tissue (which is associated with an increase in DAG
mass) is attenuated by an inhibitor of PKC (61), and 3)
decreased neural Na'K'-ATPase
activity in diabetic
mice is prevented by inhibitors of PKC (68). The prevention of glucose- and diabetes-induced decreased
Na+K+-ATPaseactivity and imbalances in Ptdlns metabolism in aorta and nerve by ARls also is consistent with
the possibility that these changes are linked to sorbitol
pathway-mediated pseudohypoxia (34,40).
Inhibition of fatty acid oxidation. The second step in
@-oxidationof long-chain fatty acids is inhibited by an
increase in mitochondrial NADH/NAD+ (31). As discussed earlier, mitochondrial NADH/NAD+ may be increased by impaired oxidation of NADH to NAD+ in the
mitochondrdia (hypoxia) or by increased transport into
the mitochondria of reducing equivalents generated in
the cytosol (i.e., by oxidation of sorbitol to fructose); it
also can result from increased oxidation of fatty acids
(77,78), In hypoxic and ischemic myocardium, longchain fatty acids accumulate as esters of CoA and of
carnitine (Fig. 3) (14,26,31,32,49). These long-chain acyl
esters, like DAG, are amphipathic molecules and are
potent modulators of many enzymes whose activities are
altered by hypoxia and by diabetes, i.e., Na+K+-ATPase
(inhibition), PKC (activation), and Ca++-ATPase (inhibition) (31,32).Accumulation of these esters in hypoxic and
ischemic hearts and in hypoxic cultured myocytes is
associated with impaired myocyte contractile function as
well as electrophysiological dysfunction (31,49,79,80).
These changes are more pronounced in hearts from
diabetic animals and are made worse by elevating FFA
levels in the perfusate (14,26,32,49).All of these changes
are attenuated by addition of L-carnitine or short-chain
DIABETES. VOL. 42, JUNE 1993
esters of L-carnitine (acetyl-L-carnitine, propionyl-L-carnitine).
Exposure of skin chamber granulation tissue vessels to
palmitoyl-L-carnitine (an LCA-C) in vivo mimics the effects of elevated glucose levels and of hypoxia on blood
flow and vascular permeability (14,26,27,81); exposure
of cultured myocytes to palmitoyl-L-carnitine also induces electrophysiological dysfunction like that caused
by hypoxia (80). These effects of palmitoyl-L-carnitine on
vascular and electrophysiological dysfunction are prevented by coadministration of L-carnitine, acetyl-L-carnitine, and/or propionyl-L-carnitine. Acetyl-L-carnitine
also prevents vascular and neural (electrophysiological)
dysfunction in diabetic rats (82,83). Note that plasma and
myocardial levels of L-carnitine are decreased by diabetes (84). At this time, it is unclear how the beneficial
effects of L-carnitine, acetyl-L-carnitine, and propionylL-carnitine are mediated. Several mechanisms have
been proposed that include scavenging of free radicals, increasing the availability of free CoA for energy
metabolism, and compensating for detergent effects of
long-chain acylcarnitines on membranes. Recent observations in diabetic rats' indicate that acetyl-L-carnitine
prevents diabetes-induced decreases in endoneurial
Na'K'-ATPase and PKC activity (84a)
lncreased PG synthetic activity. PG synthetic activity is
increased in hypoxic and ischemic cells and tissues
(85-87) as well as in glomeruli and aorta of diabetic and
nondiabetic animals exposed to elevated glucose levels
in vitro (48,88-92). The particular PGs synthesized (and
the balance between vasodilator versus vasoconstrictor
prostanoids) varies in different vascular cells and tissues
and with the severity and duration of hypoxia and hyperglycemia (85,88). The increased GFR observed early
after the onset of poorly controlled diabetes in experimental animals is associated with increased production
of vasodilatory PGs (88-90); a decreased ratio of vasodilator/vasoconstrictor PGs is associated with decreased
blood flow and GFR later in the course of diabetes (88).
Impaired contractile function of aortic rings from diabetic
and nondiabetic rabbits exposed to elevated glucose
levels in vitro is associated with increased production of
vasoconstrictor prostanoids (91,92).
lncreased PG synthetic activity in glomeruli from diabetic rats is associated with increased de novo synthesis
of DAG (67), activation of PKC (65), increased activity of
some phospholipase A, isozymes (93), and increased
GFR (13,14,88). All of these phenomena are attenuated
to varying degrees by inhibitors of AR (13,14,26,89).A
plausible linkage of these observations is depicted in Fig,
4.
EFFECTS OF AN INCREASED NADHINAD' ON
FREE-RADICAL PRODUCTION
lncreased free-radical production associated with hypoxia and ischemia. The importance of increased production of 0; and oxygen reactive species derived from
it, i.e., hydrogen peroxide and hydroxyl radical, in mediating vascular injury during reperfusion after hypoxia and
ischemia is well known (30,31,94,95). 0; production also
is increased in cultured endothelial cells subjected to
805
scavengers, ARls, and inhibitors of cyclooxygenase
(48,101) as noted earlier. At this time, it is unclear
4
tsorbitol pathway activity
whether NO production is increased by diabetes or
1
a~vcarion
whether the effects of NO synthase inhibitors on vascular
dysfunction associated with diabetes reflect correction of
an imbalance in vasodilators versus vasoconstrictors
(inhibition of NO production could compensate for in4
creased vasodilator prostanoid production and/or der - *Protein kinase C1
creased responsiveness to endothelin-1, a potent
vasoconstrictor). Because plasma levels of endothelin-1
~ P h o s p h ~ i p a s eA 2 +Na+K+-ATPase
have been reported to be increased in diabetic humans,
note that 1 ) endothelin-1 plasma levels are increased by
insulin (105), 2) the biochemical effects of endothelin-1
on cultured pericytes are blunted by exposure to elevated glucose levels (106), and 3) low-affinity endothelin-1 receptors are downregulated in glomeruli from
diabetic rats (107). This downregulation is prevented by
an inhibitor of PKC given in vivo.
VASCULAR AND NEURAL DYSFUNCTION
Mechanisms of increased 0,' production by NADH.
lncreased 0; production resulting from an increased
FIG. 4. Tentative scenario tor linkage of metabolic Imbalances that
mediate the effects of hyperglycemia and pseudohypoxla (increased
rate of reduction of NAD+ to NADH may be mediated by
cytosolic NADH/NAD+) on vascular and neural dystunctlon.
several different reactions. The likelihood that an increased NADH/NAD+ will increase 0,- production via
hypoxia (96,97). On the other hand, increased NO (en- increased PG synthetic activity is supported by evidence
dothelium-derived relaxing factor) (98,99) production discussed earlier linking increased PG synthetic activity
plays an important role in mediating hyperemia during in diabetes to pseudohypoxia, together with evidence of
reperfusion after mild hypoxia and brief ischemia (21,22). 0,- production coupled to PGH synthase activity
0; and NO both cause vasodilation and increased blood (100,108,109). Vascular dysfunction induced by topical
flow under conditions of normoxemia and euglycemia application of arachadonic acid is prevented by super(98,100); synthesis of both of these free radicals is oxide dismutase and cyclooxygenase inhibitors (100).
oxygen dependent and therefore is markedly inhibited by Reduction of PGG, to PGH, by PG hydroperoxidase
produces 0,- via a sidechain reaction that uses NADH
anoxia and severe hypoxia.
lncreased free-radical production associated with el- (and NADPH but not GSH) as a reducing cosubstrate
evated glucose levels. Numerous reports of elevated (100). Subsequent autoinactivation of prostacyclin synlevels of lipid peroxidation and glycoxidation products in thase by 0;
(produced by PG hydroperoxidase)
diabetic humans and animals attest to diabetes-induced (108,109), would explain diabetes-induced decreased
increased oxidative stress (see APPENDIX, note 7). It is not production of prostacyclin and other vasodilatory prosknown, however, to what extent these parameters of tanoids reported by some investigators. This would result
increased oxidative stress reflect increased production in a decreased ratio of vasodilator/vascoconstrictor prosof free radicals and/or impaired free-radical scavenger tanoids because thromboxane synthase is resistant to
function (29). Recent studies in animal models of diabe- inactivation by 0,. (108,109). Higher levels of oxygen
tes are consistent with the likelihood that increased free radicals also inactivate PG cyclooxygenase activity.
production of 0, and relative or absolute increases in Note that NO has been reported to inhibit lipid oxidation
NO (and/or related monoxides [99]) formation mediate by lipoxygenase and cyclooxygenase (110). If increased
vascular and neural dysfunction induced by acutely PG synthetic activity associated with diabetes is indeed
elevated glucose levels (in vitro and in vivo) and by triggered by increased availability of arachadonic acid,
diabetes of short duration (17,18,42,101-104). Increases then metabolism of arachadonic acid and associated
also
in blood flow and vascular albumin permeation induced formation of 0,. by microsomal cytochrome ,,P,
by acute (-5- h duration) hyperglycemia in nondiabetic may be increased (30,111).
0,- is a normal byproduct of oxidation of NADH to
rats and in rats with short duration diabetes (2-5 wk) are
largely prevented by free-radical scavengers (probucol NAD+ by the electron transport chain in mitochondria
and/or superoxide dismutase) and by inhibitors of NO (30,31). Reduced electron transport chain components
synthase, AR, and PG synthetic activity (13,14,26,27, can undergo spontaneous auto-oxidation with 0,. forma42,103,104,104a; J.R.W., K.C., R.G.T., unpublished ob- tion (97). If cytosolic-reducing equivalents produced by
servations). Compounds that release NO independent of accelerated sorbitol pathway metabolism are transported
NO synthase activity, such as nitroglycerin, sodium nitro- into mitochondria rapidly enough, autoxidation of reprusside, and 3-(4-morpholinyl)sydnone irnine (Sin I ) , all duced electron transport components and 0,. formation
increase vascular permeability as well as blood flow in may be increased.
the skin chamber model (104).
NADH also has been reported to increase 0,. producImpaired contractile function of aortic rings exposed to tion by cultured human fibroblasts exposed to interleuelevated glucose levels is prevented by free-radical kin-1 (112). The nature of the reactions(s) that mediate
(
1
HYPERGLYCEMIA
I
,.?~~~~$d~fa'~~s
lAb
1
1
I
1
DIABETES, VOL. 42, JUNE 1993
this effect of NADH are unclear but may be analogous to
production of 0; by plasma membrane-associated
NADPH oxidase in leukocytes. These observations raise
the possibility that increased availability of NADH may
favor increased 0; formation by other peroxidases, i.e.,
GSH peroxidase (100).
Note that increased oxidative stress and 0; formation
by the above mechanisms may occur in the absence of
an increased cytosolic ratio of NA DH/NADf provided
that the increased rate of reduction of NADf to NADH by
SDH does not exceed the maximal rate at which the cell
can reoxidize NADH to NADf.
Mechanisms of increased NO production by NADH.
lncreased NO production associated with hypoxia and
diabetes may be initiated by an increase in intracellular
calcium (i.e., by increased 0,- formation [ I 13,114]),
which would activate the constitutive isoform of NO
synthase (98), and/or by activation of PKC (by increased
levels of DAG and/or LCA-C and LCA-CoA), which has
been reported to phosphorylate the constitutive isoform
of NO synthase that results in a -40% increase in activity
(1 15,116). Note that a marked increase in NO production
may increase NADH/NADf by inactivating aconitase
and electron transport chain enzymes (98). It also is of
interest that NO synthase, like AR, requires NADPH as a
cofactor (98).
Many (but not all) investigators have reported that
acute hyperglycemia and diabetes impair acetylcholineinduced (NO-mediated) relaxation of vessels in vitro
(46,48,91,92,101,117-120). Most reports suggesting impaired NO production are based on experiments in which
vessels were incubated for several hours in media lacking L-arginine as substrate for NO synthesis. The possibility that impaired NO production by tissues from
diabetic animals (and from nondiabetic animals incubated at elevated glucose levels) in L-arginine-deficient
media may be an artifact of the in vitro milieu is suggested by several lines of evidence: 1) when the concentration of L-arginine (or H, biopterin, another cofactor
required for NO synthesis) is suboptimal, activation of
purified brain NO synthase results in production of hydrogen peroxide rather than NO (121); 2) glucose- and
diabetes-induced decreases in Naf Kf -ATPase activity
in rabbit aortic rings are not observed in tissues incubated in media containing L-arginine (122); and 3 ) impaired acetylcholine-induced relaxation of arteries from
hypercholesterolemic rabbits appears to be an in vitro
artifact of increased free-radical production (catalyzed
by trace metals in the buffer) (123); trace metal-catalyzed free-radical production would be even greater in
hyperglycemic media because of autoxidation of glucose
(29). Whereas glycated proteins and 0; react with (and
inactivate) NO in vitro (98,124), the pathophysiological
significance of such observations is unclear because the
rate of NO quenching versus rates of NO production and
advanced glycation product formation in vivo (in diabetic
subjects) is unknown. Further studies are needed to
resolve these discordant observations regarding the nature and time course of changes in NO production as well
as changes in other vasodilators and vasoconstrictors
induced by diabetes.
DIABETES, VOL. 42, JUNE 1993
Second messenger effects and cell injury linked to
increased 0,' production. Increasing evidence attests
to important roles for both 0; and NO in modulating
normal cellular metabolism and mediating responses to
vasoactive agents and injury (94,98,99,114).Therefore, if
0; production is increased in diabetic subjects, it could
mediate many (otherwise apparently unrelated) diabetes-induced metabolic imbalances and biochemical
changes, including extracellular matrix changes. For
example, ascorbic acid-induced increases in collagen
production and expression of procollagen a1 (1) mRNA
levels in human fibroblasts are mediated by reactive
aldehyde products of lrpid peroxidation (caused by 0);
and are prevented by probucol (125) (which also prevents glucose-induced vascular dysfunction [104]).
Thus, increased 0; production could potentially mediate
glucose-induced increased gene expression (in cultured
vascular and mesangial cells) for extracellular matrix
constituents, i.e., type IV collagen, fibronectin, and p1
laminin (12 6 1 27). lncreased 0; production linked to
hyperglycemic pseudohypoxia also could mediate DNA
damage and impaired cell replication in cultured endothelial cells and tissues exposed to elevated glucose
levels (126,127). Oxygen free radicals have been reported to depress sarcolemmal Naf Kf -ATPase and
Caf -ATPase activity (52), increase intracellular Caf
(1 13,114), release arachadonic acid and stimulate cyclooxygenase activity (1 14,128),and inhibit NO synthase
activity (129).
+
+
GALACTOSE-INDUCED PSEUDOHYPOXIA
Rodents fed galactose-enriched diets develop vascular
and neural dysfunction and early vascular structural
changes identical to those in diabetic animals and which,
like those in diabetic animals, are prevented by inhibitors
of AR (7,26,27,33). Unlike sorbitol (which is readily
oxidized to fructose by SDH, galactitol does not appear
to be further metabolized except in the liver. For this
reason, many (if not most) investigators have attributed
sorbitol pathway-linked diabetic complications to increased intracellular concentration of sorbitol (osmotic
stress) and/or to redox changes and metabolic imbalances linked to reduction of glucose to sorbitol (33,36).
Redox changes and metabolic imbalances associated
with oxidation of sorbitol to fructose were considered
unlikely candidate mechanisms for mediating sorbitol
pathway-linked complications of diabetes.
In view of these considerations, it was of great interest
to find that in human erythrocytes and in rat granulation
tissue exposed to elevated galactose levels (in vitro and
in vivo, respectively) cytosolic NADH/NADf and associated increases in triose phosphates and snG3P equal or
exceed those induced by elevated glucose levels and
are prevented by ARls and by pyruvate (26,27,130;
J.R.W., unpublished observations). Exposure of human
erythrocytes to sorbitol also causes a marked increase in
NADH/NADf that is attenuated by pyruvate (59); galactitol has no effect on NADH/NADf. These observations
suggest the existence of an as yet unidentified polyol
pathway-linked mechanism for increasing cytosolic
807
NADH/NAD+. In any case, these findings are consistent
with the likelihood that increased cytosolic NADH/NAD+
mediates polyol pathway-linked vascular and neural
functional and early structural changes induced by
galactosemia (33) as well as by hyperglycemia. It is of
interest in this regard that elevated galactose levels (like
elevated glucose levels) stimulate de novo synthesis of
DAG in cultured vascular cells (G. King, unpublished
observations).
Depletion of GSH levels and increased susceptibility to
oxidative stress in selected tissues (i.e., lens and erythrocytes) of diabetic and galactose-fed animals have
been widely attributed to competition between AR and
GSH reductase for NADPH cofactor. Even though
NADPH levels are modestly decreased by accelerated
polyol pathway activity, recent studies indicate that the
decrease is not rate limiting for reduction of oxidized
GSH because of the much lower K, of GSH reductase
versus AR for NADPH (131). The likelihood that sorbitol
pathway-linked depletion of reduced GSH in these
tissues reflects increased oxidative stress (i.e., an increased NADH/NAD+) that results from accelerated
oxidation of sorbitol to fructose is consistent with depletion of reduced GSH levels linked to the increased
NADHINAD~ in hypoxic tissues of nondiabetic subjects
(132).
In view of this new evidence of galactose-induced
pseudohypoxia and earlier evidence that nonenzymatic
glycation of lens crystallins by galactose is prevented by
GSH (133), note that ARls prevent galactose-induced 1)
depletion of lens GSH, 2) increased fluorescence (indicative of nonenzymatic glycation) and crosslinking of lens
crystallins (134), and 3) pseudohypoxia. These observations are consistent with the putative importance of
polyol pathway-induced oxidative stress mediated by
pseudohypoxia (i.e., an increased NADH/NAD+) in the
pathogenesis of diabetic cataracts and nonenzymatic
glycation as well as in other complications of diabetes.
IMPACT OF INCREASED CYTOSOLIC NADH/NAD+ ON THE
SUSCEPTIBILITY OF VESSELS AND NERVES TO INJURY
BY FACTORS INDEPENDENT OF DIABETES
The nature of the metabolic imbalances and functional
consequences linked to hyperglycemic pseudohypoxia
suggests new insights for understanding the increased
susceptibility of diabetic subjects to vascular and neural
injury by risk factors independent of diabetes as well as
by other systemic metabolic and biochemical imbalances associated with the diabetic milieu. Included
among such factors are hypoxic and ischemic injury,
elevated plasma levels of atherogenic lipoproteins and
FFAs, hypertension, cigarette smoking, ethanol abuse,
and increased oxygen-derived free radicals associated
with nonenzymatic glycation and glycoxidation. Potential
interactions between hyperglycemic pseudohypoxia and
each of these factors are considered below.
First, the mechanism by which an increase in cytosolic
NADH/NAD+ could increase the susceptibility of tissues
to hypoxic and ischemic injury is depicted in Fig. 2. As
discussed earlier, cells with an increased NADH/NAD+
attributable to hyperglycemic pseudohypoxia will require
808
less severe hypoxia or ischemia to increase NADHI
NAD+ to the same level caused by hypoxia or ischemia
alone.
Second, because albumin permeation into the aorta is
increased two- to threefold by diabetes, it is likely that
vascular permeation by other macromolecules such as
atherogenic lipoproteins also will be increased. Thus,
even in diabetic subjects with normal cholesterol and
lipoprotein profiles, permeation of atherogenic lipoproteins may be increased two- to threefold, which corresponds to the increased frequency of heart attacks in
diabetic subjects (35). Any further increase in plasma
levels of atherogenic lipoproteins would result in a corresponding increased flux into the vessel wall.
The likelihood that elevated plasma FFA levels in
diabetic subjects may accentuate hyperglycemic
pseudohypoxia and contribute to excessive accumulation of LCA-C and LCA-CoA, causing more vascular and
neural injury, is supported by evidence that 1) elevation
of FFAs leads to increased rates of fatty acid oxidation
and reduction of NAD+ to NADH (resulting in an increase in mitochondria1 NADH/NAD+) (77,78), which will
accentuate oxidative stress and pseudohypoxia induced
by hyperglycemia; 2) elevated FFAs increase hypoxic
and ischemic injury to the isolated perfused heart; 3)
levels of cytotoxic LCA-C and LCA-CoA are higher in
hypoxic and ischemic hearts from diabetic than from
nondiabetic subjects; and 4) microalbuminuria in diabetic humans and animals is significantly correlated with
elevated plasma triglycerides (which also correlate with
plasma nonesterified fatty acids) (14).
Third, the finding that blood flow in diabetic rats is
preferentially increased in tissues prone to late complications, in the absence of an increase in systemic blood
pressure, implies dilation of resistance arterioles and
impaired smooth muscle contractile function in the affected tissues. Thus, systemic blood pressure will be
transmitted further downstream causing microvascular
hypertension (Fig. 5). Hypertension is a well- known stimulus for vascular collagen production (and is associated
with increased 0; production) (loo), which will result in
sclerosis and stiffening of the vessel wall (14,26,27).
Impaired contractile function of resistance arterioles also
will permit any increase in systemic blood pressure to be
transmitted further downstream resulting in even more
severe microvascular hypertension. These hemodynarnic changes, coupled with increased vascular permeability, may account for the more severe and widespread
hyalinization of arterioles in diabetic than nondiabetic
subjects. The dilation of resistance arterioles induced by
diabetes is in sharp contrast to the increased peripheral
resistance (attributable to contraction of resistance arterioles) that contributes to hypertension in large arteries
while tending to limit transmission of systemic blood
pressure into the microvasculature distal to resistance
arterioles. Thus, these observations predict that at any
arterial blood pressure, microvascular blood flow will be
higher in these (complication-prone) tissues of diabetic
subjects than in nondiabetic subjects and much higher
than in nondiabetic hypertensive subjects (until blood
flow is reduced by vascular sclerosis).
DIABETES, VOL. 42, JUNE 1993
NEG 4
I
-'i
t Superoxide
t
4Sorbitol pathway flux + 4 reduction of NAD*
4 Free radicals, t Na*K*-ATPase
to NADH
+
+Vascular permeability
+Ratio of vasodilatorlvasoconstrictoragents
Impaired contractile function of resistance arterioles
t
Vasodilation, increased blood flow, microvascular hypertension
+
4 Collagen synthesis -w vascular scarring
t
+
+Vascular compliance + pulsatile blood flow
t
Tissue PO, + +oxidation of NADH to NAD*
t
Progressive vascular sclerosis -w organ failure
FIG. 5. Translation of early reversible metabolic Imbalances
associated with hyperglycemlc pseudohypoxia and vascular
dysfunctlon into self-perpetuating vascular structural changes
culminating in organ failure.
Fourth, cigarette smoking has been implicated as a
significant risk factor for vascular complications of diabetes. lncreased CO levels in smokers could impact on
diabetes-induced vascular injury in several ways. CO,
like NO, reacts with iron-containing electron transport
chain enzymes in the mitochondria, thereby impairing
oxidation of NADH; binding of CO to Hb also would
impair oxygen delivery to tissues. In addition, CO (like
NO) activates guanylate cyclase (129,133, which could
contribute to vasodilation and increased blood flow.
Fifth, hyperglycemic pseudohypoxia also may provide
an explanation for reports that neuropathy is more common and more severe in diabetic subjects who consume
excessive amounts of ethanol (136) (assuming the presence of alcohol dehydrogenase in neural tissue). Oxidation
of ethanol to acetaldehyde by alcohol dehydrogenase, like
oxidation of sorbitol to fructose, is coupled to reduction of
NAD+ to NADH. Thus, the impact of ethanol ingestion on
cytosolic NADH/NAD+ is equivalent to higher glucose
levels and an increased flux of glucose via the sorbitol
pathway.
Sixth, numerous reports attest to increased 0; production by glycated proteins and by autoxidative glycoxidation reactions (29,137-139). Glycated proteins induce
vascular dysfunction (during euglycemia) like that
caused by elevated glucose levels, which also is prevented by free-radical scavengers such as probucol and
superoxide dismutase (104a; R.G.T., K.C., J.R.W., unpublished observations). These observations suggest
that elevated glucose levels initiate production of free
radicals by two independent mechanisms: 1) hyperglycemic pseudohypoxia, and 2) nonenzymatic glycation
and autoxidative glycoxidation. Therefore, increased
production of oxygen free radicals appears to be a
strong candidate mechanism for effecting nonenzymatic
as well as enzymatically mediated glucose effects on
vascular and neural dysfunction and the late complications of diabetes (Figs. 4 and 5).
CONCLUSIONS
A considerable body of evidence 1) indicates that in
many tissues hyperglycemia-induced metabolic imbal-
DIABETES, VOL. 42, JUNE 1993
ances increase the cytosolic ratio of free NADH/NADf
(despite normal tissue PO,) that results in pseudohypoxia; 2) attests to the similarity of vascular, neural, and
redox changes induced by true hypoxia and hyperglycemia-induced pseudohypoxia; and 3) supports the importance of oxidative stress and a branching cascade of
metabolic imbalances linked to pseudohypoxia in mediating vascular and neural dysfunction induced by diabetes. Imbalances in lipid metabolism, increased 0;
production, and perhaps increased NO formation appear
to play important roles in mediating these functional
disorders. Tentative scenarios for translation of these
early metabolic imbalances and vascular dysfunction
into irreversible progressive vascular sclerosis and organ
failure are suggested in Figs. 2-5.
lncreased production of 0; and NO appears to play
an important role in mediating early vascular and neural
dysfunction linked to true hypoxia, hyperglycemic
pseudohypoxia, and nonenzymatic glycation. Clearly,
much more work is needed to delineate the mechanisms
by which elevated glucose levels cause pseudohypoxia
and the role of hyperglycemia-induced pseudohypoxia
(and associated oxidative stress and metabolic imbalances) in mediating complications of diabetes.
ACKNOWLEDGMENTS
This research was supported by grants from the National
Institutes of Health (HL-39934, AM-20579, and EY06600) and by Pfizer-Central Research Division, Sigma
Tau s.p.a. Rome, and the Kilo Diabetes and Vascular
Research Foundation. J.R.N. is supported by the Danish
Diabetes Association, Danish Medical Research Council,
and Novo Foundation.
We thank Drs. John W. Baynes, Peter 6. Corr, Peter J.
Oates, and John W. Turk for critical readings of this
perspective and for thoughtful comments and suggestions. We thank Wanda Allison, Susan Berhorst, Judi
Burgan, Antoinette M. Faller, Joyce Marvel, Eva Ostrow,
Nemani Rateri, and Samuel R. Smith for expert technical
assistance and Maria Himmelmann for preparation of the
manuscript.
APPENDIX
First, because of the putative importance of early vascular dysfunction (and of vascular responses to hypoxia
and pseudohypoxia) in the pathogenesis of diabetic
vascular complications, it is appropriate at the outset to
acknowledge and address discordant reports of decreased retinal and endoneurial blood flow in acutely
diabetic/hyperglycemic rats (1-3). Such reports are derived largely from indirect assessments of blood flow
(transit time of fluorescein-labeled albumin for retina [3],
hydrogen clearance or laser Doppler methods for endoneurium [1,2]) and/or invasive procedures (hydrogen
clearance and laser Doppler methods for endoneurium).
The problem with indirect measures of blood flow is that
they can be influenced by many variables independent of
blood flow, i.e.,changes in tissue blood volume and/or in
vascular permeability. Hydrogen clearance values based
on oxidation of hydrogen by a platinum electrode also
can be affected by tissue edema and by a number of the diabetic milieu (such as hypertension, hypercholesmetabolites, including ascorbic acid and oxygen-reac- terolemia, cigarette smoking, ethanol abuse, etc.) as
tive species (4); thus, redox changes and increased discussed later.
production of oxygen-reactive species associated with
Fourth, in hearts from diabetic animals, basal myocarthe diabetic milieu may affect blood flow values based on dial contractile function is impaired (49,50), and global
hydrogen clearance. The problem with invasive proce- ischemia-induced contractile dysfunction develops more
dures is that vasoreactivity in many tissues is altered by rapidly than in control hearts (49). Paradoxically, diabediabetes (5,6); thus, neural trauma and cooling inherent tes improves recovery of myocyte and vascular smooth
to surgical exposure required for hydrogen clearance muscle contractile function during reperfusion after ischand laser Doppler methods (and insertion of electrodes emia (50,51). Whereas endothelial barrrier functional
into endoneurium for the hydrogen clearance method) integrity in diabetic hearts does not differ from control
may cause a relative or absolute decrease in neural hearts before ischemia, albumin leakage in diabetic
blood flow in diabetic versus control subjects. For these hearts during reflow exceeds that in control hearts by
reasons, indirect assessments of blood flow obtained by two- to threefold (50). These paradoxical effects of diainvasive procedures lack credibility if they are discordant betes may be explained by differential responses of
with more direct measures of blood flow (i.e., appropri- endothelium versus vascular smooth muscle and cardiac
ately sized microspheres) obtained by noninvasive tech- myocytes to the combination of preexisting pseudohyniques.
poxia and impaired Ca++ transport function in the sarBlood flow measurements based on radiolabeled mi- colemma and endoplasmic reticulum (31,32,51,52).
crosphere injection methods are not subject to any of the Impaired Ca++ transport function could limit accumulaabove limitations. The major considerations in the use of tion of Ca++, which is believed to play an important role
microspheres are that the microspheres must be large in mediating hypoxic injury (31,32,51).It is of interest that
enough to be trapped efficiently by the tissue vasculature a similar paradoxical effect of ischemia is observed in
examined, and enough microspheres must be captured peripheral nerve in which resistance to ischemic conducto ensure reliable sampling and counting (this can be a tion failure is increased by diabetes (53).
Fifth, even the combined measurements of tissue frucproblem in small tissues like retina and sciatic nerve of
rats) (7). These potential confounding problems have tose and sorbitol will tend to underestimate sorbitol
been excluded in the experimental models we have used pathway activity because fructose can be further metabby the demonstration that molecular microspheres (i.e., olized and also readily diffuses out of cells to be carried
3H-desmethylimipramine, 266,000 M,) yield blood flow away by the blood. About 66-75% of the fructose provalues identical to those obtained with conventional 10- duced by incubated tissues is recovered in the medium.
Sixth, vasodilation and increased blood flow may be
to 15-pm microspheres in retina and endoneurium of
diabetic rats (8). These findings, coupled with elevated viewed as normal vascular responses to hypoxia and
microvascular hematocrits (a rheological parameter of pseudohypoxia as well as to various vasoactive subincreased blood flow) in retina and sciatic nerve tissue of stances. The point at which these normal vascular rediabetic rats (9), strongly support the conclusion that sponses should be considered vascular dysfunction is
blood flow in these tissues is increased by acute hyper- unclear. As long as thevessels are dilated and blood flow
is increased in response to sustained pseudohypoxia,
glycemia and by diabetes of short duration.
Second, true hypoxia is defined as decreased tissue they may be considered to be reacting normally. On the
p 0 2 levels (in the absence of restricted blood flow) that other hand, vessels that are dilated in response to
result from a decrease in 0, content of blood (hypox- hyperglycemic pseudohypoxia may dilate and constrict
emia) delivered to tissue. Ischemia is defined as de- less in response to vasodilating and vasoconstricting
creased tissue PO, levels that result from decreased agents than vessels in a euglycemic milieu. Although the
dampened responses of such vessels to vasoactive
volumetric flow of blood with a normal PO, content.
Third, despite these provocative similarities in neural, agents is generally viewed as evidence of vascular
vascular, and myocyte contractile dysfunction induced dysfunction, it may just as well be considered a normal
by diabetes and hypoxia, it is clear that chronic hypoxic response to the combined effects of pseudohypoxia and
conditions (i.e., chronic lung disease, cyanotic heart the particular vasoactive agent examined. In any case,
disease, etc.) are not associated with diabetic complica- as discussed later, chronic vasodilation and increased
tions and vice versa (although retinal capillary microan- blood flow imply microvascular hypertension, which is an
eurysms and diabetic vascular changes in retina, kidney, important risk factor for diabetic retinopathy and neand muscle are not unique to diabetes). Clearly, the phropathy.
combined effects of chronically elevated glucose levels
Seventh, glycoxidation products are formed by reacand pseudohypoxia are essential for development of the tions between sugar-derived autoxidation products
late complications of diabetes. In addition, other risk (such as glucosone) and proteins, lipids, etc. Oxidative
factors such as hypertension appear to play a critical role stress is defined as an increase in steady-state levels of
in the pathogenesis of late complications. Indeed, we oxygen-reactive species (including O,; hydrogen peroxhave proposed that the major impact of the diabetic ide, and hydroxyl radical) that result either from inmilieu on the vasculature and nerves is to induce meta- creased production of precursors of reactive oxygen
bolic imbalances and dysfunction, which make them species or decreased free-radical scavenger activity
more susceptible to injury by risk factors independent of (29).
810
DIABETES, VOL. 42, JUNE 1993
REFERENCES
1. Yasuda H, Sonobe M, Yamashita M, Terada M, Hatanaka I, Huitian
Z, Shigeta Y: Effect of prostaglandin E, analogue TFC 612 on
diabetic neuropathy in streptozocin-induced diabetic rats. Diabetes 38:832-38, 1989
2. Cameron NE, Cotter MA, Low PA: Nerve blood flow in early
experimental diabetes in rats: relation to conduction deficits. Am J
Physiol261:EI-E8, 1991
3. Shiba T, Bursell SE. Clermont A, Sportsman R, Heath W, King GL:
Protein kinase C (PKC) activation is a causal factor for the
alteration of retinal blood flow in diabetes of short duration (Abstract), lnvest Ophthalmol Visual Sci 32:785, 1991
4. Young W: H, clearance measurement of blood flow: a review of
technique and polarographic principles. Stroke 11:552-64, 1980
5. Tomlinson KC, Gardiner SM, Hebden RA, Bennett T: Functional
consequences of streptozotocin-induced diabetes mellitus, with
particular reference to the cardiovascular system. Pharmacol Rev
44:103-50, 1992
6. We~dmannP, Beretta-Piccoli C, Trost BN: Pressor factors and
responsiveness in hypertension accompanying diabetes mellitus.
Hypertension 7 (Suppl. 2):33-42, 1985
7. Tilton RG, Chang K, Wei el C, Kilo C, Williamson JR: Increased
ocular blood flow and '2'l-albumin permeation in galactose-fed
rats: inhibition by sorbinil lnvest Ophthalmol VisualSci29:861-68,
1988
8. Tilton RG, Chang K, Allison W, Williamson JR: Comparable diabetes induced increases in retinal blood flow assessed with conventional versus molecular (3H-Desmethylimiprimine) microspheres
(Abstract), lnvest Ophthalmol Visual Sci 33:1048, 1992
9. Sutera SP, Chang K, Marvel J, Williamson JR: Concurrent increases in regional hematocrit and blood flow in diabetic rats:
prevention by sorbinil. Am J Physiol263:H945-50, 1992
10 Patel V, Rassam S, Newsom R, Wiek J, Kohner E. Retinal blood
flow in diabetic retinopathy. Br Med J305:678-83, 1992
11. Ernest JT, Goldstick TK: Response of choroidal vascular resistance to hyperglycemia. Int Ophthalmol6:119-24, 1983
12. Cunha-Vaz JG, Fonseca JR, deAbreu JRF, Lima JJP: Studies on
retinal blood flow. II. Diabetic retinopathy Arch Ophthalmol 96:
809--11, 1978
13. Tilton RG, Chang K, Pugliese G, Eades DM, Province MA, Sherman WR, Kilo C, Williamson JR: Prevention of hemodynamic and
vascular albumin filtration changes in diabetic rats by aldose
reductase inhibitors. Diabetes 38,1258-70, 1989
14. Tilton RG, Baier LD, Harlow JE, Smith SR. Ostrow E, Williamson JR:
Diabetes-induced glomerular dysfunction: links to a more reduced
cytosolic ratio of NADH/NADt. Kidney Int 41:778-88, 1992
15 Sandeman DD. Shore AC, Tooke JE. Relation of skin capillary
pressure in patients with insulin-dependent diabetes mellitus to
complications and metabol~ccontrol. N Engl J Med 327:760-64,
1992
16 Houben AJHM, Schaper NC, Nieuwenhuijzen Kruseman AC: Acute
effects of local hyperglycaemia on peripheral blood flow in man.
Diabetic Med 10:39-43, 1993
17. Hasan KS, Chang K, Allison W, Faller A, Santiago JV, Tilton RG,
Williamson JR: Glucose-induced increases in regional blood flow
are prevented by aminoguanidine and L-NMMA, inhibitors of nitric
oxide synthase. (Abstract) FASEB J 7(Part I):A105, 1993
18. Hasan KS, Chang K, Allison W, Faller A, Santiago JV, Tilton RG,
Williamson JR: Glucose-induced increases in ocular blood flow are
prevented by aminoguanidine and L-NMMA, inhibitors of nitric
oxide synthase (Abstract), lnvest Ophthalmol Visual Sci 34:1127,
1993
19. Needleman P, Key SL, lsakson PC, Kulkarni PS: Relationship
between oxygen tension, coronary vasodilation and prostaglandin
biosynthesis in the isolated rabbit heart. Prostaglandins 9:123-34,
1975
20. Fallon TJ, Maxwell D, Kohner EM. Retinal vascular autoregulation
in conditions of hyperoxia and hypoxia using the blue field entoptic
phenomenon. Ophthalmology 92:701-705, 1985
21. Kostic MM, Schrader J: Role of nitric oxide in reactive hyperemia of
the guinea pig heart. Circ Res 70:208-12, 1992
22. Park KH, Rubin LE, Gross SS, Levi R: Nitric oxide is a mediator of
hypoxic coronary vasodilatation: relation to adenosine and cyclooxygenase-derived metabolites. Circ Res 71 :992-1001, 1992
23. Williamson DH, Lund P, Krebs HA: The redox state of free
nicotinamide-adenine dinucleotide in the cytoplasm and mitochondria of rat liver. Biochem J 103:514-27, 1967
24. Travis SF, Morrison AD, Clements RS Jr, Winegrad Al, Oski FA:
Metabolic alterations in the human erythrocyte produced by increases in glucose concentration: the role of the polyol pathway. J
Clin lnvest 50:2104-12, 1971
25. Ruderman NB, Schmahl FW, Goodman MN: Regulation of alanine
formation and release in rat muscle in vivo: effect of starvation and
DIABETES, VOL. 42, JUNE 1993
diabetes. Am J Physiol233:E109-14, 1977
26. Pugliese G, Tilton RG, Williamson JR: Glucose-induced metabolic
imbalances in the pathogenesis of diabetic vascular disease.
Diabetes Metab Rev 7:35-59, 1991
27. Williamson JR, Kilo C, Tilton RG: Mechanisms of glucose- and
diabetes-induced vascular dysfunction. In Hyperglycemia, Diabetes, and Vascular Disease Ruderman NB, Williamson JR, Brownlee M, Eds. New York, Oxford Univ. Press, 1992, p. 107-32
28. Van den Enden M, Ostrow E, Williamson JR: The role of "hyperglycemic (metabolic) hypoxia" in the pathogenesis of diabetic
complications (Abstract). Diabetologia 35 (Suppl. 1):A46, 1992
29. Baynes JW: Role of oxidative stress in development of complications in diabetes Diabetes 40:405-12, 1991
30. Freeman BA, Crapo JD: Biology of disease: free radicals and
tissue injury. Lab lnvest 47:412-26, 1982
31. Corr PB, Gross RW, Sobel BE: Amphipathic metabolites and
membrane dysfunction in ischemic myocardium. Circ Res 55:13554, 1984
32. Tahiliani AG, McNeill JH: Diabetes-induced abnormalities in the
myocardium. Life Sci 38:959-74, 1986
33. Dvornik D: Aldose Reductase Inhibition: An Approach to the
Prevention of Diabetic Complications. Porte D, Ed. New York,
McGraw-Hill, 1987
34. Winegrad Al: Does a common mechanism induce the diverse
complications of diabetes? Diabetes 36:396-406, 1987
35. Crepaldi G, Nosadini R: Diabetic cardiopathy: Is it a real entity?
Diabetes Metab Rev 4:273-88, 1988
36. Kinoshita JH, Nishimura C: The involvement of aldose reductase in
diabetic complications. Diabetes Metab Rev 4:323-37, 1988
37. Brownlee M, Cerami A, Vlassara H: Advanced products of nonenzymatic glycosylation and the pathogenesis of diabetic vascular
disease. Diabetes Metab Rev 4:437-51, 1988
38. Ruderman NB, Williamson JR, Brownlee M: Glucose and diabetic
vascular disease. FASEB J 62905-1 4, 1992
39. Ruderman NB, Williamson JR, Brownlee M: Hyperglycemia, Diabetes, and Vascular Disease. New York, Oxford Univ. Press, 1992
40. Greene DA, Sima AAF, Stevens MJ, Feldman EL, Lattimer SA:
Complications: neuropathy, pathogenetic considerations. Diabetes Care 15:1902-25, 1992
41. Misko TP, Moore WM, Kasten TP, Nickols GA, Corbett JA, Tilton
RG, McDaniel ML, Williamson JR, Currie MG: Selective inhibition of
the inducible nitric oxide synthase by aminoguanidine. Eur J
Pharmaco1233:119-25 1993
42. Tilton RG, Chang K, Hasan KS, Smith S, Petrash JM, Misko TP,
Moore WM, Currie MG, Corbett JA. McDaniel ML, Williamson JR:
Prevention of diabetic vascular dysfunction by guanidines: inhibition of nitric oxide synthase versus advanced glycation end
product formation. Diabetes 42:221-32, 1993
43. Cameron NE, Cotter MA, Robertson S: Contractile properties of
cardiac papillary muscle in streptozotocin-diabetic rats and the
effects of aldose reductase inhibition. Diabetologia 32:365-70,
1989
44. Cameron NE, Cotter MA. Robertson S: Changes in skeletal muscle
contractile properties in streptozocin-induced diabetic rats and
role of polyol pathway and hypoinsulinemia. Diabetes 39:460-65,
1990
45. Ao S, Shingu Y, Kikuchi C, Takano Y, Nomura K, Fujiwara T,
Ohkubo Y, Notsu Y, Yamaguchi I: Characterization of a novel
aldose reductase inhibitor, FR74366, and its effects on diabetic
cataract and neuropathy in the rat. Metabolism 40:77-87, 1991
46. Cameron NE, Cotter MA: Impaired contraction and relaxation in
aorta from streptozotocin-diabetic rats: role of polyol pathway.
Diabetologia 35: 1011-19, 1992
47. Cameron NE, Cotter MA: Dissociation between biochemical and
functional effects of the aldose reductase inhibitor, ponalrestat, on
peripheral nerve in diabetic rats. Br J Pharmacol 107:939-44.
1992
48. Cohen RA, Tesfamariam B: Diabetes mellitus and the vascular
endothelium. In Hyperglycemia, Diabetes, and Vascular Disease
Ruderman NB, Williamson JR, Brownlee M, Eds. New York, Oxford
Univ. Press, 1992, p . 48-58
49. Feuvray D, Idell-Wenger JA, Neely JR: Effects of ischemia on rat
myocardial function and metabolism in diabetes. Circ Res 44:32229, 1979
50. Tilton R, Daugherty A, Sutera SP, Larson KB, Land MP, Rateri DL,
Kilo C, Williamson JR: Myocyte contracture, vascular resistance,
and vascular permeabil~tyafter global ischemia in isolated hearts
from alloxan-induced diabetic rabbits. Diabetes 38:1484-91, 1989
51. Tani M, Neely JR: Hearts from diabetic rats are more resistant to in
vitro ischemia: possible role of altered Ca2+ metabolism. Circ Res
62:931-40, 1988
52. Kaneko M, Singal PK, Dhalla NS: Alterations in heart sarcolemmal
Cast-ATPase and Ca"-binding activities due to oxygen free
811
radicals. Basic Res Cardiol85:45-54, 1990
53. Low PA: Recent advances in the pathogenesis of diabetic neuropathy. Muscle Newe 10:121-28, 1987
54. Lagunas R, McLean P, Greenbaum AL: The effect of raising the
NADf content on the pathways of carbohydrate metabolism and
lipogenesis in rat liver. Eur J Biochem 15:179-90, 1970
55. Cheng HM, Gonzalez RG: The effect of high glucose and oxidative
stress on lens metabolism, aldose reductase, and senile cataractogenesis. Metabolism 35 (Suppl. 1):lO-14, 1986
56. Griffin BW, McNatt LG, York BM Jr: Characterization of aldose
reductase activities from human and animal sources by a sensitive
fluorescence assay. In Enzymology and Molecular Biology and
Carbonyl Metabolism Weiner H, Ed. New York, Liss, 1987, p.
325-40
57. Rose IA, Warms JVB: Control of glycolysis in the human red blood
cell. J Biol Chem 241 :4848-54, 1966
58. Kobayashi K Neely JR: Control of maximum rates of glycolysis in
rat cardiac muscle. Circ Res 44:166-75, 1979
59. Frangos M, Smith S, Santiago J, Kilo C: Sorbitol-induced imbalances in glycolysis in human erythrocytes are reduced by pyruvate
(Abstract). Diabetes 39 (Suppl. 1):274A, 1990
60. Williamson JR, Chang K, Allison W, Burgan J, Tilton RG: Inhibition
of retinal vascular dysfunction in diabetic rats by S-0773, an
inhibitor of sorbitol dehydrogenase (Abstract), lnvest Ophthalmol
Visual Sci 34:712, 1993
6Oa.WilliamsonJR, Chang K, Allison W, Smith SR, Burgan J, Tilton RG:
Sorbitol dehydrogenase inhibition (SDI) attenuates sciatic nerve
vascular dysfunction in diabetic rats (Abstract). Diabetes. In press
61. Wolf BA, Willamson JR, Easom RA, Chang K, Sherman WR, Turk
J: Diacylglycerol accumulation and microvascular abnormalities
induced by elevated glucose levels. J Clin lnvest 87:31-38, 1990
6la.Hasan KS, Santiago JV, Williamson JR: Acute hyperglycemiainduced increases in regional blood flow are prevented by pyruvate and by tolrestat (Abstract). Diabetes. In press
62. Williamson JR, Chang K, Ostrow E, Allison W, Harlow J, Kilo C:
Sorbitol-induced increases in vascular albumin clearance (VAC)
are prevented by pyruvate but not myeinositol (Abstract). Diabetes 38:94A, 1989
63. Nyengaard JR, Van den Enden M, Ostrow E, Burgan JE, Williamson JR: Prevention of glucose-induced increases in rat retinal
NADH/NADf by an aldose reductase inhibitor and pyruvate
(Abstract). lnvest Ophthalmol Visual Sci 34:716, 1993
64. Trevisan R, Nosadini R, Fioretto P, Velussi M, Avogaro A, Duner E,
lori E, Doria A, Merkel C, Valerio A, Crepaldi G: Metabolic control
of kidney hemodynamics in normal and insulin-dependent diabetic
subjects: Effects of acetoacetic, lactic, and acetic acids. Diabetes
36:1073-81, 1987
65. Craven PA, DeRubertis FR: Protein kinase C is activated in
glomeruli from streptozotocin diabetic rats. J Clin lnvest 83:166715, 1989
66. Lee T-S, Saltsman KA, Ohashi H, King GL: Activation of protein
kinase C by elevation of glucose concentration: proposal for a
mechanism in the development of diabetic vascular complications.
Proc Natl Acad Sci USA 86:5141-45, 1989
67. Craven PA, Davidson CM, DeRubertis FR: Increase in diacylglycerol mass in ~solatedglomeruli by glucose from de novo synthesis
of glycerolipids. Diabetes 39:667-74, 1990
68. Hermenegildo C, Felipo V, Minana MD, Romero FJ, Grisolia S:
Sustained recovery of Naf/Kf-ATPase activity in sciatic nerve of
diabetic mice by administration of H7 or callphostin C, inhibitors of
protein kinase C. Diabetes 42:257-62,1993
69. King GL, Shiba T, Feener EP, Nayak R: Cell culture model for the
study of vascular complications of diabetes: the effect of high
glucose levels on metabolism and growth of vascular cells. In
Hyperglycemia, Diabetes, and Vascular Disease Ruderman NB,
Williamson JR, Brownlee M, Eds. New York, Oxford Univ. Press,
1992, p. 162-77
70. Zhu X, Eichberg J: 1,2-diacylglycerol content and its arachidonylcontaining molecular species are reduced in sciatic nerve from
streptozotocin-induced diabetic rats. J Neurochem 55:1087-90,
1990
71. Okumura K, Nishiura T, Awaji Y, Kondo J, Hashimoto H, Ito T:
I,2-diacylglycerol content and its fatty acid composition in thoracic aorta of diabetic rats. Diabetes 40:820-24, 1991
72. Williams B. Schrier RW: Characterization of glucose-induced in situ
protein kinase C activity in cultured vascular smooth muscle cells.
Diabetes 41 : I 464-72, 1992
73. Drummond AH, Raeburn CA: The interaction of lithium with thyrotropin-releasing hormone-stimulated lipid metabolism in GH3 pituitary tumour cells. Biochem J 224:129-36, 1984
74. Downes CP, Stone MA: Lithium-induced reduction in intracellular
inositol supply in cholinergically stimulated parotid gland. Biochem
J234:199-204, 1986
812
75. Berridge MJ, Downes CP, Hanley MJ: Neural and developmental
actions of lithium: a unifying hypothesis. Cell59:411-19, 1989
76. Godfrey PP: Potentiation by lithium of CMP-phosphatidate formation in carbachol-stimulated rat cerebral-cortical slices and its
reversal by myeinositol. Biochem J 258:621-24, 1989
77. Williamson JR, Kreisberg RA, Felts PW: Mechanism for the stimulation of gluconeogenesis by fatty acids in perfused rat liver. Proc
Natl Acad Sci USA 56:247-54, 1966
78. Neely JR, Whitmer KM, Mochizuki S: Effects of mechanical activity
and hormones on myocardial glucose and fatty acid utilization.
Circ Res 38 (Suppl. 1):22-30, 1976
79. Knabb MT, Saffitz JE, Corr PB, Sobel BE: The dependence of
electrophysiological derangements on accumulation of endogenous long-chain acyl carnitine in hypoxic neonatal rat myocytes.
Circ Res 58:230-40, 1986
80. Corr PB, Creer MH, Yamada KA, Saffitz JE, Sobel BE: Prophylaxis
of early ventricular fibrillation by inhibition of acylcarnitine accumulation. J Clin lnvest 83:927-36, 1989
81. Williamson JR, Chang K, Kilo C, Sherman WR, Montenero M,
Arrigoni-Martelli E: Prevention of glucose-induced vascular functional changes by acetyl-L-carnitine. Diabetologia 33 (Suppl.):
A1 75, 1990
82. Lowitt S, Malone JI, Salem A, Korthals J: Acetylcarnitine improves
neuronal function in streptozotocin (STZ) diabetic rats (Abstract).
Diabetes 39 (Suppl. 1):155A, 1990
83. Williamson JR, Chang K, Allison WS, Tilton RG, Orfalian Z, ArrigoniMartelli E: Acetyl-L-carnitine, but not L-carnitine, attenuates diabetes-induced increases in vascular 13'1-BSA clearance, lnvest
Ophthalmol Visual Sci 32:1029, 1991
84. Vary TC, Neely JR: A mechanism for reduced myocardial carnitine
levels in diabetic animals. Am J Physiol243:H154-58, 1982
84a.ldo Y, Rattarasarn M, Allison W, Orfalian Z, Arrigoni-Martelli E:
Effects of acetyl-L-carnitine (AC) on funct~onal and metabolic
abnormalities in diabetic peripheral nerve (Abstract). Diabetes. In
press
85. Farber HW, Center DM, Rounds S: Effect of ambient oxygen on
cultured endothelial cells from different vascular beds. Am J
Physiol253:H878-83, 1987
86. Farber HW, Barnett HF, Differences in prostaglandin metabolism in
cultured aortic and pulmonary arterial endothelial cells exposed to
acute and chronic hypoxia. Circ Res 68:1446-57, 1991
87. Martin LD, Barnes SD, Wetzel RC: Acute hypoxia alters eicosanoid
production of perfused pulmonary artery endothelial cells in culture. Prostaglandins 43:371-82, 1992
88. Craven PA, Caines MA, DeRubertis FR: Sequential alterations in
and thromboxane svnthesis in diabetic
alomerular ,orostaalandin
~ - - rats: relatonship to thehyperfiltration of early dkbetes. Metabolism
36:95-103, 1987
89. Craven PA, DeRubertis FR: Sorbinil suppresses glomerular prostaglandin production in the streptozotocin diabetic rat. Metabolism
38:649-54, 1989
90. Chang WP, Dimitriadis E, Allen T, Dunlop ME, Cooper M, Larkins
RG: The effect of aldose reductase inhibitors on glomerular prostaglandin production and urinary albumin excretion in experimental diabetes mellitus. Diabetologia 34:225-31, 1991
91. Tesfamariam B, Jakubowski JA, Cohen RA: Contraction of diabetic
Am J
rabbit aorta caused by endothelium-derived PGH,-TxA,.
Physiol257:H1327-33, 1989
92. Tesfamariam B, Brown ML, Deykin D, Cohen RA: Elevated glucose
promotes generation of endothelium-derived vasoconstrictor prostanoids in rabbit aorta. J Clin lnvest 85:929-32, 1990
93. Craven PA, Patterson MC, DeRubertis FR: Role of enhanced
arachidonate availability through phospholipase A2 pathway in
mediation of increased prostaglandin synthesis by glomeruli from
diabetic rats. Diabetes 37:429-35, 1988
94. Halliwell B: Superoxide, iron, vascular endothelium and reperfusion injury. Free Radical Res Comms 5:315-18, 1989
95. Lefer AM, Tsao PS, Lefer DJ, Ma X: Role of endothelial dysfunction
in the pathogenesis of reperfusion injury after myocardial ischemia.
FASEB J 5:2029-34, 1991
96. Ratvch RE, Chuknviska RS, Bulkley GB: The primary localization of
free radical generation after ano&/reoxygenation-in isolated endothelial cells. Surgery 102:122-31, 1987
97. Block ER, Patel JM, Edwards D: Mechanism of hypoxic injury to
pulmonary artery endothelial cell plasma membranes. Am J Physio1257:C223-31, 1989
98. Nathan C: Nitric oxide as a secretory product of mammalian cells.
FASEB J 6:3051-64, 1992
99. Stamler JS, Singel DJ, Loscalzo J: Biochemistry of nitric oxide and
its redox-activated forms. Science 258:1898-1 902, 1992
100. Kukreja RC, Kontos HA, Hess ML, Ellis EF: PGH synthase and
lipoxygenase generate superoxide in the presence of NADH or
NADPH. Circ Res 59:612-19, 1986
--
~
DIABETES, VOL. 42, JUNE 1993
101. Tesfamariam B. Cohen RA: Free rad~calsmediate endothelial ceil
dysfunction caused by elevated glucose. Am J Physiol263:H32126, 1992
102. Cameron NE, Cotter MA, Maxfield E: Antioxidant treatment prevents peripheral nerve dysfunction in diabetic rats (Abstract).
Diabetologia (Suppl. 1) 35:A150. 1992
103. Corbett JA, Tilton RG, Chang K, Hasan KS, Ido Y, Wang JL,
Sweetland MA, Lancaster JR Jr, Williamson JR, McDaniel ML:
Amnoguanidine, a novel inhibitor of nitric oxide formation, prevents diabetic vascular dysfunction. Diabetes 41 :552-56, 1992
104. Tilton RG, Chang K, Allison W. Williamson JR: Glucose-induced
increases in blood flow and 1251-albuminpermeation: roles of nitric
oxide and superoxide. (Abstract). FASEB J 7:(Part ll):A905, 1993
104a.TiltonRG, Chang K, Faller AM: Probucol prevents vascular protein
leakage induced by diabetes, glucose, and glycated proteins
(Abstract). Diabetes. In press
105. Hu R-M, Levin ER. Pedram A, Frank HJL: Insulin stimulates the
production and secretion of endothelin from bovine endothelial
cells. Diabetes 42351-58. 1993
106. De La Rubia G, Oliver FJ, lnoguchi T, King GL: Induction of
resistance to endothelin-1's biochemical actions by elevated glucose levels in retinal pericytes. Diabetes 41:1533-39, 1992
107. Awazu M, Parker RE, H a ~ i eBR, lchikawa I. Kon V. Downregulation of endothelin-1 receptors by protein kinase C in streptozotocin diabetic rats. J Cardiovasc Pharmacol 17 (Suppl. 7):
S500-5502, 1991
108. Egan RW, Paxton J, Kuehl FA Jr: Mechanism for irreversible
self-deactivation of prostaglandin synthetase. J Biol Chem 251:
7279-35
. --- - - , 1976
.-. 109. Warso MA, Lands WEM: Lipid peroxidation in relation to prostacyclin and thromboxane physiology and pathophysiology. Br Med
Bull 39:277-80, 1983
110 Kanner J. Harel S. Granit R. Nitric oxide. an inhibitor of lioid
oxidation by lipoxygenase, cyclooxygenase and hemoglobin, iipids 27,46-49, 1992
111 Capdevila JH, Falck JR, Estabrook RW: Cytochrome P450 and the
arachidonate cascade. FASEB J 6:731-36, 1992
112. Meier B, Radeke HH, Selle S, Younes M, Sies H, Resch K,
Habermehl GG: Human fibroblasts release reactive oxvaen sDecies in response to interleukin-1 or tumour necrosi6'iacto~-a.
Biochem J 263:539-45, 1989
113. Masumoto N, Tasaka K, Miyake A, Tanizawa 0 : Superoxide anion
increases intracellular free-calcium in human myometrial cells. J
Biol Chem 265:22533-36, 1990
114 lkebuchi Y, Masumoto N, Tasaka K. Koike K, Kasahara K: Superoxide anion increases intracellular pH, intracellular free calcium,
and arachidonate release in human amnion cells. J Biol Chem
266:13233-37, 1991
115 Nakane M, Mitchell J. Forstermann W, Murad F: Phosphowlation
by calcium calmoduiin-dependent protein kinase II and protein
kinase C modulates the activity of nitric oxide synthase. Biochem
Biophys Res Commun 180, 1396-1 402, 1991
116. Bredt DS, Hwang PM, Glatt CE, Lowenstein C, Reed RR, Snyder
SH: Cloned and expressed nitric oxide synthase structurally resembles cytochrome P-450 reductase. Nature (Lond) 351 :714-18,
1991
117. Pieper GM, Gross GJ: Oxygen free radicals abolish endotheliumdependent relaxation in diabetic rat aorta. Am J Physiol255:H82533, 1988
118. Calver A, Collier J, Vallance P: Inhibition and stimulation of nitric
oxide synthesis in the human forearm arterial bed of patients with
insulin-dependent diabetes. J Clin Invest 90:2548-54, 1992
119. Elliott TG. Cockcrofl JR, Groop PH, Earle K, Morocutti A, Viberti
GC, Ritter JM: Vascular responses in the forearm of microalbuminuric type I (insulin-dependent) diabetic patients: evidence of
altered nitric oxide biosynthesis (Abstract). Diabetologia 35
(Suppl. 1):A19, 1992
DIABETES, VOL 42, JUNE 1993
120. Smits P, Kapma J, Jacobs M, Lutterrnan J, Thien T: Endotheliumdependent vascular relaxation in patients with type I diabetes.
Diabetes 42:148-53, 1993
121. Heinzel B, John M, Klatt P, Bohrne E, Mayer B: Ca2+/calmodulindependent formation of hydrogen peroxide by brain nitric oxide
synthase. Biochem J 281 :627-30, 1992
122. Gupta S, Sussman I, McArthur CS, Tornheim K. Cohen RA,
Ruderman NB: Endothelium-dependent inhibition of NaC-KCATPase activity in rabbit aorta by hyperglycemia: possible role of
endothelium-derived nitric oxide. J Clin Invest 90:727-32. 1992
123. Simon BC, Cohen RA: EDTA influences reactivity of isolated aorta
from hypercholesterolemic rabbits. Am J Physiol262:H1606-10,
1992
124. Bucala R, Tracey KJ, Cerami A: Advanced glycosylation products
quench nitric oxide and mediate defective endothelium-dependent vasodilatation in experimental diabetes. J Clin lnvest87.43238, 1991
125. Chojkier M, Houglum K, Solis-HerruzoJ, Brenner DA: Stimulation of
collagen gene expression by ascorbic acid in cultured human
fibroblasts: 1 role for lipid peroxidation? J Biol Chem 264:1695762, 1989
126. Cagliero E, Roth T, Roy S, Maiello M, Lorenzi M: Expression of
genes related to the extracellular matrix in human endothelial cells:
differential modulation by elevated glucose concentrations, phorbol esters, and CAMP. J Biol Chem 266.14244-250, 1991
127. Kreisberg JI: Biology of disease: hyperglycemia and microangiopathy. Direct regulation by glucose of microvascular cells. Lab
Invest 67:416-26, 1992
128. Lawson DL, Mehta JL, Nichols W, Mehta P, Donnelly WH:
Superoxide radical-mediated endothelial injury and vasoconstriction of rat thoracic aortic rings. J Lab Clin Med 115:541-48, 1990
129. Brijne B. Schmidt K-U. Ullrich V: Activation of soluble auanvlate
cyclase by carbon monoxide and inhibition by superoxGe ahion.
Eur J Biochem 192:683-88, 1990
130. Frangos M, Santiago JV, Williamson JR: In human erythrocytes
galactose induces greater imbalances in glycolysis than glucose
(Abstract). Diabetes 40 (Suppl. 1):254A, 1991
131. Lou MF. Dickerson JE Jr. Garadi R. York BM Jr: Glutathione
depletion in the lens of galactosemic and diabetic rats. Exp Eye
Res 46:517-30, 1988
132. Mak IT, Boehme P, Weglicki WE: Antioxidant effects of calcium
channel blockers against free radical injury in endothelial cells:
correlation of protection with preservation of glutathione levels.
Circ Res 70:1099-1103, 1992
133. Huby R, Harding JJ: Non-enzymatic glycosylation (glycation) of
lens proteins by galactose and protection by aspirin and reduced
glutathione. Exp Eye Res 47:53-59, 1988
134. Nagaraj RH, Monnier VM: Non-tryptophan fluorescence and high
molecular weight protein formation in lens crystallins of rats with
chronic galactosemia: prevention by the aldose reductase inhibitor
sorbinil. Exp Eye Res 51 :411-18, 1990
135. Verma A, Hirsch DJ, Glatt CE, Ronnett GV. Snyder SH: Carbon
monoxide: a putative neural messenger. Science 259:381-84,
1993
136. McCulloch DK, Campbell IW, Prescott RJ, Clarke BF: Effect of
alcohol intake on symptomatic peripheral neuropathy in diabetic
men. Diabetes Care 3:245-47, 1980
137. Gillery P, Monboisse JC, Maquart FX, Borel JP. Glycation of
proteins as a source of superoxide. Diabete Metab 14:25-30,
1988
138. Jiang N ,
Woollard ACS, Wolff SP: Hydrogen peroxide production
during experimental protein glycation. FEBS Lett268:69-71, 1990
139. Mullarkey CJ, Edelstein D, Brownlee M: Free radical generation by
early glycation products: a mechanism for accelerated atherogenesis in diabetes. Biochem Biophys Res Commun 173:932-939.
1990