Rapid Publication

Increased Atherosclerosis in Streptozotocin-induced Diabetic Mice

Vidya V. Kunjathoor, Deborah L. Wilson, and Renée C. LeBoeuf
Department of Medicine, University of Washington, Seattle, Washington 98195

Abstract heavily to risk of atherosclerosis. (J. Clin. Invest. 1996. 97:

1767–1773.) Key words: hyperglycemia • genetics • fatty
Premature and extensive atheroscleroses involving renal, streaks • diabetic complications • vascular disease
peripheral, and cardiovascular sites remain major compli-
cations of diabetes mellitus. Controversy exists as to the Introduction
contribution of hyperglycemia versus elevated local or sys-
temic concentrations of insulin to atherosclerosis risk. In Premature and extensive atheroscleroses involving renal, pe-
this report, we developed the first murine model susceptible ripheral, and cardiovascular sites remain major complications
to both atherosclerosis and diabetes to determine which dia- of diabetes mellitus (1–4). About 70–80% of deaths in diabetic
betogenic factors contribute to vascular disease. C57BL/6 patients are due to vascular disease. Since hypertension and
and BALB/c mice were treated with multiple low-dose hypercholesterolemia, well-known atherosclerosis risk factors,
streptozotocin (STZ) or control citrate buffer and fed rodent explain just a minor part of the excess incidence of vascular
chow or an atherogenic-promoting (Ath) diet for 12–20 wk. disease among diabetic patients, diabetogenic factors them-
STZ treatment resulted in sustained hyperglycemia (250– selves must contribute to the development of arterial disease.
420 mg/dl) and a modest reduction in plasma insulin levels In particular, hyperglycemia, the primary clinical manifesta-
for both strains regardless of diet. Citrate-treated C57BL/6 tion of diabetes, is thought to contribute to diabetic complica-
mice fed the Ath diet showed extensive oil red O–staining tions by altering vascular cellular metabolism, vascular matrix
fatty streak aortic sinus lesions (20,53762,957 mm2), the molecules, and circulating lipoproteins. For instance, hypergly-
size of which did not differ for Ath-fed mice treated with cemia increases diacylglycerol levels and activates protein ki-
STZ (16,83662,136 mm2). In contrast, hyperglycemic BALB/c nase C activity in the aorta of streptozotocin (STZ)1-induced
mice fed the Ath diet showed a 17-fold increase in athero- diabetic rats (5) and dogs (6). Thickening of the basement
sclerotic lesion area (7,99262,096 mm2) as compared with membranes in renal glomeruli and peripheral capillaries has
citrate-treated mice fed the Ath diet (4676318 mm2). Corre- been observed in STZ-induced diabetic rats (7) and diabetic
lations between lesion size and plasma glucose levels were patients (4). Hyperlipidemia is a feature of drug-induced dia-
significant for BALB/c (r 5 0.741, P , 0.009), but not betes in rats (8) and rabbits (9, 10), as well as poorly controlled
C57BL/6 (r 5 0.314, P , 0.3) mice. Lesion size correlated diabetes in humans (11). Alterations in lipoprotein–cell inter-
significantly with plasma cholesterol for C57BL/6 (r 5 actions are also seen in vitro upon glycation of circulating lipo-
0.612, P , 0.03) but not BALB/c (r 5 0.630, P , 0.1) mice. proteins (12). The role of each of these mechanisms in the
Immunohistochemistry showed that aortic sinus lesions pathogenesis of macro- and microangiopathy needs to be clari-
from both strains contained macrophages, but smooth mus- fied.
cle cells were clearly present in lesions of BALB/c mice. In In addition to hyperglycemia, systemic or local elevations
summary, we present the first small animal model showing in insulin may contribute to aberrant lipid metabolism (13) and
accelerated atherosclerosis in response to hyperglycemia. vascular wall function (14). Imperfect normalization of glucose
Fatty streaks resembled those of human type II lesions in metabolism by replacement insulin therapy may alter the con-
that both macrophages and smooth muscle cells were evi- centrations and compositions of potentially atherogenic lipo-
dent. In addition, our results support the concept that hy- proteins (13). Changes in the ratio of apolipoproteins A-I to
perglycemia as opposed to hyperinsulinemia contributes A-II in HDL have been observed (15), possibly interfering
with the protective role of these lipoproteins in vascular dis-
ease (16). At the vascular wall, insulin may contribute directly
Part of this work was presented in abstract form (1995. FASEB [Fed. to increasing the levels of cellular cholesterol via its ability to
Am. Soc. Exp. Biol.] J. 9:766a). increase cellular sterol synthesis, induce LDL receptors, and
Address correspondence to Dr. Renée C. LeBoeuf, Department inhibit HDL-mediated cholesterol removal (17). Although
of Medicine and Nutritional Sciences, Box #353410, University of population studies have shown positive correlations between
Washington, Seattle, WA 98195. Phone: 206-543-5208; FAX: 206-685- hyperinsulinemia and increased incidence and mortality rates
1696; E-mail: [email protected]
of cardiovascular disease (3), it remains to be established
Received for publication 10 November 1995 and accepted in re-
vised form 24 January 1996.
whether the hyperinsulinemia per se is most detrimental to

J. Clin. Invest.
© The American Society for Clinical Investigation, Inc.
0021-9738/96/04/1767/07 $2.00 1. Abbreviations used in this paper: Ath, atherogenic diet; STZ, strep-
Volume 97, Number 7, April 1996, 1767–1773 tozotocin; TBS, Tris-buffered saline.

Atherosclerosis in Diabetic Mice 1767

vascular health in diabetes. The alternative possibility is that determined using a colorimetric kit (diagnostic kit No. 236691; Boeh-
hyperglycemia is associated with atherosclerosis. ringer Mannheim, Indianapolis, IN) with cholesterol standards (Pre-
In this report, we test the hypothesis that hyperglycemia ciset No. 125512; Boehringer Mannheim) as described (19). HDL
can directly contribute to increased risk of large vessel disease. cholesterol values were measured after selective precipitation of
VLDL/LDL by polyethylene glycol (20). Triglyceride concentrations
We used STZ to induce hyperglycemia in mice which normally
in plasma were determined after removal of free glycerol (diagnostic
exhibit resistance (BALB/c) or susceptibility (C57BL/6) to kit No. 450032; Boehringer Mannheim). Plasma glucose concentra-
high fat diet-induced atherosclerosis (18). We show that STZ- tions were determined colorimetrically (glucose trinder kit No. 315;
induced hyperglycemia alone did not induce fatty streak le- Sigma Diagnostics, St. Louis, MO). Hepatic lipids were extracted us-
sions in either mouse strain. C57BL/6 mice showed atheroscle- ing the method of Folch et al. (21), modified to contain Triton X-100
rotic lesions upon feeding a high fat/high cholesterol diet, but as described by Carr et al. (22). Insulin was quantified using a rat in-
fatty streak lesion formation was independent of hyperglyce- sulin radioimmunoassay kit as described by the manufacturers (No.
mia. However, BALB/c mice exhibited significant aortic fatty RI-13K; Linco Research Inc., St. Charles, MO).
streak lesions upon combined treatment with STZ and the Aortic sinus lesion area. Quantification of atherosclerotic fatty
atherogenic diet. Further, the fatty streak lesions were compli- streak lesions was done by evaluation of lesion size in the aortic sinus
as described (23) with modifications. Briefly, the heart and upper sec-
cated, containing both macrophages and smooth muscle cells.
tion of the aorta were removed from animals, cleaned of peripheral
These data suggest that both dietary and genetic factors can in- fat under a dissecting microscope, and sectioned directly under and
fluence lesion formation as a result of hyperglycemia. Since parallel to the atrial leaflets. The upper section was embedded in
these mice were not hyperinsulinemic, our results support the OCT medium and frozen. Every other section (10 mm thick) through-
idea that hyperglycemia alone is a risk factor for cardiovascu- out the aortic sinus (400 mm) was taken for analysis. The distal por-
lar disease in diabetes. tion of the aortic sinus is recognized by the three valve cusps which
are the junctions of the aorta to the heart. Sections were evaluated for
fatty streak lesions after staining with oil red O and counterstaining
Methods with hematoxylin and eosin. Lesion areas per section were counted
and areas estimated using a Compaq 286 computer (Compaq Com-
Animals and diets. Female BALB/cByJ (BALB/c) and C57BL/6J puter, Houston, TX) equipped with an FG-100 image acquisition
(C57BL/6) mice from 6 to 8 wk of age were obtained from The Jack- board, a high resolution video camera, and a Sony video monitor.
son Laboratory (Bar Harbor, ME). Mice were fed a pelleted rodent Area measurements were done using the Optimas Image Analysis
chow diet (Wayne Rodent BLOX 8604; Teklad Test Diets, Madison, Software Package (BioScan Inc., Edmonds, WA).
WI) for 1–2 wk before initiation of studies. Mice were maintained in a Evaluation of pancreata. Pancreata were embedded in paraplast
temperature-controlled (258C) facility with a strict 12-h light/dark cy- and 10-mm sections were stained with eosin and counterstained with
cle and given free access to food and water. Blood was collected every hematoxylin. Nine sections per pancreas were examined.
4 wk from the retroorbital sinus into tubes containing anticoagulant Immunohistochemistry. Immunohistochemical staining of cry-
(1 mM EDTA) and plasma was stored at 2708C before analysis. Mice ostat cut sections (10 mm thick) of the aortic sinus of the heart was
were killed with Nembutal (80 mg/kg) given intraperitoneally. Hearts done as described previously (24) with some modifications. Briefly,
were perfusion-fixed with 4% buffered formalin. Livers were directly sections were fixed in cold acetone for 10 min, air-dried, and washed
placed into liquid nitrogen and stored at 2708C. Pancreata were fixed two times in Tris-buffered saline (TBS). Endogenous peroxidase ac-
in 10% neutral buffered formalin. This project was approved by the tivity was blocked by incubating slides in a solution of 3% hydrogen
Animal Care and Use Committee of the University of Washington peroxidase for 5 min. All incubations were performed at room tem-
(Protocol No. 2140-02). perature in humidified chambers. Slides were incubated with the ap-
Diets and drug treatment. Two diets used in this study were pel- propriate primary antibody overnight, followed by 4–6 h of incuba-
leted rodent chow and an “atherogenic” (Ath) diet known to elicit tion with secondary horseradish peroxidase–labeled goat anti–rat
fatty streak lesions in strain C57BL/6 (18). The rodent chow diet con- immunoglobulin G antibody (1:200 dilution in TBS). Sections were
tained z 4% fat, 24% protein, and 4.5% crude fiber. The Ath diet visualized by chromogenic detection (AEC substrate system No.
provided 30% kcal from fat (primarily cocoa butter) and contained K0697; Dako Corp., Carpinteria, CA). AEC was used as a chro-
1.25% cholesterol and 0.5% sodium cholate (18). mogen. After treatment, sections were washed with TBS and coun-
BALB/c and C57BL/6 mice were randomly divided into four terstained with hematoxylin. Smooth muscle cells were identified
groups which were mice treated with STZ (mixed anomer; Sigma using a mouse monoclonal antibody against human alpha-actin di-
Chemical Co., St. Louis, MO) and fed either rodent chow or the Ath rectly coupled to horseradish peroxidase (antibody dilution was 1:6;
diet, or mice treated with the vehicle, citrate buffer (0.05 M sodium No. U7033; Dako Corp.). Macrophages were identified using a rat
citrate, pH 4.5), and fed rodent chow or the Ath diet. STZ was dis- monoclonal antibody to the mouse and human Mac-1 antigen (No.
solved in sterile citrate buffer and injected intraperitoneally into mice 1118-129; Boehringer Mannheim). VCAM-1 was detected using a rat
(40 mg/kg, z 20 ml) within 5 min of preparation. STZ or citrate buffer monoclonal antibody against murine VCAM-1 (antibody dilution
was administered for five consecutive days during the first week of was 1:100, No. 1510-01; Southern Biotechnology Associates, Inc., Bir-
this study. 4 wk later, half of the mice in each treatment group were mingham, AL). Anti–Mac-1 and anti–VCAM-1 antibodies bound to
placed on the Ath diet for the duration of the study (12–20 wk). To tissue were visualized after the binding of anti–rat immunoglobulin
maintain hyperglycemia in the STZ treatment groups, the multiple coupled to horseradish peroxidase. Control slides were incubated
low-dose STZ treatment was repeated during week 7, and control with nonspecific primary antisera, or in some cases without the pri-
mice were injected with citrate buffer. Drug and diet treatments were mary antibody. In no case did control slides show a positive signal.
staggered to allow mice to adjust to the Ath diet. Statistics. Data are reported as mean6SEM. Statistical differ-
Body weight. During the 24 wk of this study, mice were healthy ences were determined by ANOVA using SYSTAT for the Macin-
as evidenced by coat conditions and body weight gain. Initial body tosh (SYSTAT, Inc., Evanston, IL). Pearson’s correlation coefficients
weights for BALB/c and C57BL/6 mice were 2161 and 1861 grams were used to assess correlations. Three-way ANOVA was used to de-
(mean6SEM), respectively. Final body weights for all treatment termine interactions between strain, diet, and drug treatment with
groups were z 25–26 grams for BALB/c and 23–24 grams for C57BL/ diet as the covariant. Post-hoc analyses of significance were made using
6 mice. Tukey’s test for additivity. The Student’s t test was used to compare
Analytical tests. Total and HDL cholesterol concentrations were independent means. P , 0.05 was accepted as statistically significant.

1768 Kunjathoor et al.

Results Plasma lipids. Previous studies have shown that plasma
triglyceride levels decrease for mice fed the Ath diet (19) and
Hyperglycemia occurs in STZ-treated mice. Our goal was to that STZ leads to hypertriglyceridemia in rodents and rabbits
test the role of hyperglycemia on atherosclerosis development (8, 9). In the current study, the Ath diet decreased plasma tri-
without the complications of concomitant administration of in- glyceride levels markedly in citrate-treated BALB/c mice and
sulin to maintain animal viability. Further, hyperglycemia had in C57BL/6 mice regardless of drug treatment (Table I). Tri-
to be maintained for at least 14 wk to provide time for fatty glyceride levels tended to be higher for hyperglycemic mice as
streak formation in C57BL/6 mice (19). Thus, our strategy was compared with citrate-treated animals, but the differences did
to use a multiple low-dose injection regimen for STZ with a re- not reach statistical significance. The Ath diet caused marked
peat series of injections to ensure 18 wk of hyperglycemia in increases in plasma total (greater than threefold) and VLDL/
both strains of mice. By repeating the 5-d injection regimen LDL (greater than eightfold) cholesterol levels. Cholesterol
during week 7, STZ-treated mice maintained hyperglycemia levels were not significantly different between strains, and in
for at least 16 of the total 24 wk of study regardless of diet (Fig. general hyperglycemic mice showed greater levels of plasma
1). During weeks 8–24, plasma glucose levels for BALB/c mice and VLDL/LDL cholesterol than citrate-treated mice. Changes
ranged from z 250–360 mg/dl as compared with z 150–190 in HDL levels due to diet and drug treatments were strain de-
mg/dl for citrate-injected controls. C57BL/6 mice showed pendent. For citrate-treated mice, the Ath diet tended to de-
plasma glucose levels ranging from 310 to 420 mg/dl after STZ crease HDL levels for C57BL/6 but not BALB/c mice as seen
treatment as compared with 160–230 mg/dl for citrate-treated previously (26). STZ given to chow or Ath diet-fed mice
mice. Overall, plasma glucose levels nearly doubled for mice tended to increase HDL cholesterol levels, although changes
treated with STZ as compared with citrate-treated animals. did not reach statistical significance.
A three-way ANOVA showed a main effect of drug treat- Hepatic lipids. A three-way ANOVA in terms of strain,
ment on plasma glucose (P , 0.001) and that the three param- diet, and drug treatment showed main effects of diet (P ,
eters (strain, diet, treatment) accounted for 85% of plasma 0.001) and strain (P , 0.002–0.023) on hepatic lipid levels. For
glucose levels (r2 5 0.859). Thus, STZ treatment alone and instance, mice fed the Ath diet exhibited significant elevations
not diet or strain influenced glucose levels. Plasma glucose in triglyceride and cholesterol levels as compared with rodent
levels for mice treated with STZ were significantly higher chow–fed mice, and triglyceride levels for C57BL/6 were con-
than citrate-treated mice at each time point (P , 0.001– sistently higher than levels for BALB/c. Hepatic lipid levels
0.005). Plasma glucose and lipid values within each diet and were unaffected by STZ treatment. Thus, changes in plasma
strain group were not significantly different during weeks 12–24. lipids due to STZ treatment were not driven by STZ effects on
Thus, for clarity of analysis and presentation, plasma glucose gross hepatic lipid metabolism.
and lipid levels are given as mean values evaluated during Atherosclerosis development. It is well established that
weeks 12–24 (Table I). C57BL/6 but not BALB/c mice develop fatty streaks and more
Hyperglycemia was accompanied by a marked reduction in advanced atherosclerotic plaques in the aortic sinus within 16
pancreatic islet number of STZ-treated BALB/c and C57BL/6 wk of feeding diets rich in fat and cholesterol (18, 27, 28). We
mice as compared with control animals, which is consistent tested the possibility that the combination of STZ-induced hy-
with previous reports (25). The loss of islets most likely con- perglycemia and diet-induced atherosclerosis susceptibility
tributed to the decreased plasma insulin levels observed for would lead to accelerated lesion formation in C57BL/6 mice.
STZ-treated mice. Upon STZ treatment, insulin levels for In our study, citrate-treated C57BL/6 mice fed the Ath diet
BALB/c mice decreased from 1.5060.01 to 0.760.08 ng/ml (P , for 20 wk developed extensive aortic fatty streaks with a wide
0.001), and values for C57BL/6 mice decreased from 0.9660.01 range among animals of 3,300–66,500 mm2, with a mean lesion
to 0.6560.02 ng/ml (P , 0.001). Insulin levels for both strains area of 20,53762,957 mm2 (6SEM) (Fig. 2). In contrast, no sig-
were unaffected by diet. nificant lesions were seen for BALB/c mice fed the Ath diet

Figure 1. Plasma glucose concentration

in response to diet and drug treatments
for BALB/c and C57BL/6 mice. Animals
were treated with STZ (S, filled symbols)
(40 mg/kg, 5 d) or citrate buffer (C, open
symbols) during weeks 0 and 7. Mice
from both treatment groups were fed ei-
ther an atherosclerosis-promoting diet
(Ath, circles) rich in saturated fat (30%
calories primarily as cocoa butter) and
cholesterol (1.25% by weight) starting at
week 4, or were continually fed rodent
chow (Chow, squares). Plasma glucose
was determined colorimetrically as de-
scribed in Methods. Values are reported
as mean6SEM for n 5 9–18 mice.

Atherosclerosis in Diabetic Mice 1769

Table I. Plasma and Hepatic Parameters for BALB/c and C57BL/6 Mice Fed Rodent Chow or Ath Diets and Treated with STZ or
Citrate Buffer
BALB/c C57BL/6

Parameters Treatment Chow Ath Chow Ath

Plasma (mg/100 ml)

Glucose Citrate 163610 17462 178610 203614
STZ 344615 291633 355624 356624
P 0.001 0.005 0.001 0.001

Triglyceride Citrate 46612 1861* 3465 761*‡

STZ 5469 50611* 62614 1565*§
P NS NS NS NS

Total cholesterol Citrate 7263 220620* 6662 236613*

STZ 7962 377683i 7565 290610*
P NS NS NS 0.005

VLDL/LDL cholesterol Citrate 1863 14469* 1764 202616*

STZ 1863 265656* 1864 23767*
P NS NS NS NS

HDL cholesterol Citrate 5463 8068i 49615 3563§

STZ 6262 95615 5763 52610§
P NS NS NS NS

Hepatic lipids (mg/gram tissue)

Triglyceride Citrate 360.6 1161* 1161§ 1561i
STZ 360.3 1462* 761‡ 2162*
P NS NS NS NS

Cholesterol Citrate 360.2 4863* 260.1 6464*

STZ 360.3 52610* 260.3 7568*
P NS NS NS NS

Data are presented as mean6SEM averaged over 12–24 wk for n 5 9–20 mice. P values in each category compare citrate and STZ treatment groups.
*P , 0.005 and i P , 0.05 denote comparisons between groups fed rodent chow (Chow) and Ath diet. § P , 0.005 and ‡ P , 0.05 denote comparisons
between mouse strains.

for up to 20 wk (4676318 mm2). STZ-induced hyperglycemia lesions of C57BL/6 mice reacted only with antibody to Mac-1
combined with the Ath diet did not induce larger lesions in showing that the primary cell type in these fatty streaks was
C57BL/6 mice (16,83662,136 mm2). However, BALB/c mice macrophages. In contrast, oil red O staining regions of BALB/c
treated with STZ and fed the Ath diet showed a 17-fold in- mice contained cells immunoreactive to Mac-1 and alpha-actin.
crease in lesion size (7,99262,096 mm2). No lesions were seen Immunoreactive signal to VCAM-1 was present in lesions
for chow-fed mice regardless of drug treatment. Thus, in- from both strains, and no qualitative differences in location or
creases in lesion size due to feeding the high fat diet and in- signal intensity were seen between strains. Thus, BALB/c fatty
duced hyperglycemia were strain dependent, suggesting that streaks were complex, consisting of at least two cell types, mac-
both diet and genetic background may contribute to acceler- rophages and smooth muscle cells, both of which were lipid
ated atherosclerosis as seen in diabetes. laden. This is the first report of smooth muscle cells appearing
To determine whether levels of plasma cholesterol, glu- early in fatty streak development in mice and suggests that the
cose, or both contributed to lesion formation, correlations be- etiology of lesion development is distinct between these
tween lesion sizes and these plasma parameters among indi- strains.
vidual mice fed the Ath diet were calculated. For STZ-treated
BALB/c mice, plasma glucose (r 5 0.741, P , 0.009) but not Discussion
plasma total cholesterol (r 5 0.630, P , 0.1) levels correlated
with lesion size. In contrast, plasma cholesterol (r 5 0.612, P , In this report, we tested whether hyperglycemia in the absence
0.03) but not plasma glucose (r 5 0.314, P , 0.3) levels deter- of concomitant hyperinsulinemia could contribute to en-
mined lesion sizes for STZ-treated C57BL/6 mice. Thus, hy- hanced atherosclerosis formation in mice. One of the earliest
perglycemia in mice fed the high fat diet is important for lesion features of developing atherosclerotic plaques is the appear-
development in BALB/c mice. ance of fatty streaks in the artery wall. BALB/c mice made dia-
The character of lesions seen in BALB/c and C57BL/6 was betic by STZ treatment and fed a high fat/high cholesterol diet
distinct (Fig. 3). While fatty streaks of both strains showed oil exhibited significant fatty streak formation. Since these mice
red O staining of intra- and extracellular lipid, immunohis- had low levels of circulating insulin, hyperglycemia was the
tochemical staining showed distinct cellularities. Cells within main diabetogenic factor contributing to increased lesion for-

1770 Kunjathoor et al.

 Figure 2. Aortic sinus fatty streak lesion
sizes for BALB/c and C57BL/6 mice. Mice
were treated during weeks 0 and 7 with STZ
(40 mg/kg, 5 d) or citrate buffer as described
in Methods. Mice were fed an atherosclero-
sis-promoting diet rich in saturated fat (30%
calories primarily as cocoa butter) and cho-
lesterol (1.25% by weight) starting at 4 wk
and continuing for 12–20 wk. Aortic sinuses
were collected, sectioned, and stained with
oil red O as described in Methods before
quantification of fatty streak regions. Hori-
zontal lines indicate mean lesion area for
each group.

mation. Strengthening this conclusion is the observation that engineered mice with severe hypercholesterolemia (35),
plasma glucose but not lipid levels correlated with lesion for- smooth muscle cells are not a general feature of early mouse
mation in BALB/c mice. fatty streaks. The fact that these cells were seen in BALB/c but
The development of fatty streaks in response to hypergly- not in C57BL/6 mice suggests that diabetogenic factors con-
cemia was strain specific, as both STZ and high fat diet treat- tribute to smooth muscle cell proliferative events. Many bio-
ments were required to produce fatty streaks in BALB/c mice, chemical pathways linking hyperglycemia to vascular changes
but C57BL/6 mice showed extensive lesion development with have been recognized, including the increased synthesis of spe-
fat feeding alone. Our results with C57BL/6 mice are compara- cific growth factors (36). Smooth muscle cells appeared early
ble to those of Nishina et al. (29) who showed that C57BL/6 in lesion development in BALB/c mice (16 wk) and thus fac-
and C57BL/Ks strains carrying mutations which predispose tors influencing the proliferation of smooth muscle cells can be
them to diabetes and/or obesity do not show enhanced lesion examined rather quickly in this model.
formation upon high fat/high cholesterol diet feeding. This The role of hyperglycemia in BALB/c lesion development
suggests that diabetes does not contribute to arterial disease in is still unclear. Apart from direct effects due to inflammation,
the C57BL genetic background. lipoproteins may be altered via glycation (31) making them
What could account for the differences in diabetes-acceler- more readily taken up by scavenger receptors on cells of the
ated lesion formation we observed in BALB/c and C57BL/6 artery wall (12, 14). Glycation of matrix molecules occurs,
mice? Previous studies have indicated that among inbred which could provide a stronger trapping network for lipopro-
mouse strains fed the Ath diet, the extent of lesion formation teins penetrating vascular spaces. In addition, a direct role of
is positively correlated with the expression of genes associated STZ on the vasculature cannot be ruled out. Treatment of
with inflammation and oxidative stress (30). In particular, he- mice with STZ resulted in modest or no significant changes in
patic mRNA levels of several inflammatory and oxidative plasma and hepatic lipids as compared with citrate-treated
stress responsive genes were markedly induced in C57BL/6 mice within each diet group. Also, pancreas damage was lim-
mice, but less so in BALB/c mice. Further, oxidized LDL were ited and plasma insulin levels were decreased to approxi-
shown to stimulate the expression of these same genes. Glucose mately half normal levels. Thus, evidence for severe STZ tox-
promotes the oxidation of serum and arterial wall proteins (31, icity is lacking, making it unlikely that STZ was responsible for
32) and thereby contributes to the induction of vascular and li- the acceleration of lesion formation in BALB/c mice. Further
poprotein changes associated with aberrant lipoprotein up- study of these processes and the effects of insulin treatment on
take, monocyte binding, and inflammatory events. We hypoth- lesion development in BALB/c mice is underway.
esize that hyperglycemic BALB/c mice fed the high fat diet Our study provides the first system in which to carefully ex-
experience an increase in oxidized lipoproteins which increase amine the role of hyperglycemia as opposed to hyperinsuline-
oxidative stress at the vascular wall over a threshold needed to mia in atherosclerosis development. Overall, animal models
accelerate the development of early atherosclerotic lesions. simulating the interactions between diabetes and atherosclero-
For C57BL/6 mice, this threshold is surpassed by diet alone. sis have been poorly developed. Still et al. (8) demonstrated
We are currently addressing this hypothesis by examining in- that rats recovered from the diabetic effects of alloxan showed
flammatory gene expression which is expected to increase in increased atherosclerosis in the aorta when fed a high fat diet,
Ath diet–fed BALB/c mice treated with STZ as compared but characterization of lesions and atherogenic mechanisms
with citrate. was not performed. Alloxan-induced diabetic rabbits are
A distinct feature of the aortic fatty streaks in BALB/c largely protected against diet-induced atherosclerosis probably
mice was their complexity, characterized by a mixture of lipid- because of an accumulation of large triglyceride-rich lipopro-
laden macrophages and smooth muscle cells reminiscent of teins which may be excluded from the artery wall (9). Primate
type II lesions in humans (33). Although smooth muscle cells models of diabetes which are susceptible to atherosclerosis
have been seen in aortic sinus lesions of specific strains chroni- may be available but are prohibitively expensive for research.
cally (. 7 mo) fed fat and cholesterol (18, 34), or in genetically Thus, manipulation of BALB/c mice with STZ and high fat di-

Atherosclerosis in Diabetic Mice 1771

 Acknowledgments
We thank Dr. Åke Lernmark and Dr. John Oram (Department of
Medicine, University of Washington) for helpful comments and sug-
gestions when reviewing this manuscript, and Dr. Emil Chi and Dr.
Ying-Tzang Tien (Department of Pathology, University of Washing-
ton) for their excellent technical support.
This work was supported by National Institutes of Health grant
DK-02456.

References
1. Pyorala, K., M. Laakso, and M. Uusitupa. 1987. Diabetes and atheroscle-
rosis: an epidemiologic view. Diabetes Metab. Rev. 3:463–524.
2. Bierman, E.L. 1992. George Lyman Duff Memorial Lecture. Atherogen-
esis in diabetes. Arterioscler. Thromb. 12:647–656.
3. Krolewski, A.S., E.J. Kosinski, J.H. Warram, O.S. Leland, E.J. Busick,
A.C. Asmal, L.I. Rand, A.R. Christlieb, R.F. Bradley, and C.R. Kahn. 1987.
Magnitude and determinants of coronary artery disease in juvenile-onset, insu-
lin-dependent diabetes mellitus. Am. J. Cardiol. 59:750–755.
4. Andersen, A.R., J.S. Christiansen, J.K. Andersen, S. Kreiner, and T.
Deckert. 1983. Diabetic nephropathy in type I (insulin-dependent) diabetes: an
epidemiological study. Diabetologia. 25:496–501.
5. Inoguchi, T., P. Xia, M. Kunisaki, S. Higashi, E.P. Feener, and G.L. King.
1994. Insulin’s effect on protein kinase C and diacylglycerol induced by diabe-
tes and glucose in vascular tissues. Am. J. Physiol. 267:E369–E379.
6. Xia, P., T. Inoguchi, T.S. Kern, R.L. Engerman, P.J. Oates, and G.L.
King. 1994. Characterization of the mechanism for the chronic activation of dia-
cylglycerol-protein kinase C pathway in diabetes and hypergalactosemia. Dia-
betes. 43:1122–1129.
7. Olgemoller, B., and E. Schleicher. 1993. Alterations of glomerular matrix
proteins in the pathogenesis of diabetic nephropathy. Clin. Invest. 71:S13–S19.
8. Still, W.J.S., J.M. Martin, and W.H. Gregor. 1964. The effect of alloxan
diabetes on experimental atherosclerosis in the rat. Exp. Mol. Pathol. 3:141–147.
9. Nordestgaard, B.G., S. Stender, and K. Kjeldsen. 1988. Reduced athero-
genesis in cholesterol-fed diabetic rabbits. Giant lipoproteins do not enter the
arterial wall. Arteriosclerosis. 8:421–428.
10. Miller, R.A., and R.B. Wilson. 1984. Atherosclerosis and myocardial is-
chemic lesions in alloxan-diabetic rabbits fed a low cholesterol diet. Arterioscle-
rosis. 4:586–591.
11. The Diabetes Control and Complications Trial Research Group. 1993.
The effect of intensive treatment of diabetes on the development and progres-
sion of long-term complications in insulin-dependent diabetes mellitus. N. Engl.
J. Med. 329:977–986.
12. Hunt, J.V., M.A. Bottoms, K. Clare, J.T. Skamarauskas, and M.J.
Mitchinson. 1994. Glucose oxidation and low-density lipoprotein-induced mac-
rophage ceroid accumulation: possible implications for diabetic atherosclerosis.
Figure 3. Characterization of early atherosclerotic lesions in the Biochem. J. 300:243–249.
proximal sinus. BALB/c and C57BL/6 mice were fed an atherosclero- 13. Bierman, E.L., and J.A. Glomset. 1992. Disorders of lipid metabolism.
sis-promoting diet containing excess fat and cholesterol for 12–20 wk In Williams Textbook of Endocrinology. 8th edition. J.D. Wilson and D.W.
and treated with multiple low doses of STZ as described in Figs. 1 and Foster, editors. W.B. Saunders, Philadelphia.
14. Lyons, T.J. 1992. Lipoprotein glycation and its metabolic complications.
2. A and B are serial sections from one BALB/c mouse which were Diabetes. 41(Suppl. 2):67–73.
stained for neutral lipid with oil red O (A) and treated with antibody 15. Biesbroeck, R.C., J.J. Albers, P.W. Wahl, C.R. Weinberg, M.L. Basset,
to human alpha-actin (B) as described in Methods. The location of and E.L. Bierman. 1982. Abnormal composition of high density lipoproteins in
one of several positively reacting smooth muscle cells is indicated non-insulin-dependent diabetics. Diabetes. 31:126–131.
16. Barbaras, R., P. Puchois, J.C. Fruchart, and G. Ailhaud. 1987. Choles-
with an arrow. C is a section from a C57BL/6 mouse which was
terol efflux from cultured adipose cells is mediated by Lp AI particles but not
treated with antibody to human alpha-actin and illustrates the ab- LpAI:AII particles. Biochem. Biophys. Res. Commun. 142:63–69.
sence of positively reacting cells in lesions from this mouse strain. 17. Oram, J.F. 1995. Can insulin promote atherogenesis by altering cellular
cholesterol metabolism? J. Lab. Clin. Med. 126:229–230.
18. Paigen, B., B.Y. Ishida, J. Verstuyft, R.B. Winters, and D. Albee. 1990.
Atherosclerosis susceptibility differences among progenitors of recombinant in-
ets will allow identification of dietary and genetic factors af- bred strains of mice. Arteriosclerosis. 10:316–323.
19. LeBoeuf, R.C., M. Caldwell, and E. Kirk. 1994. Regulation by nutri-
fecting vascular disease in diabetes. tional status of lipids and apolipoproteins A-I, A-II, and A-IV in inbred mice. J.
In summary, our results support the concept that hypergly- Lipid Res. 35:121–133.
cemia as opposed to hyperinsulinemia contributes heavily to 20. Izzo, C., F. Grillo, and E. Murador. 1978. Improved method for determi-
nation of high-density-lipoprotein cholesterol. I. Isolation of high-density-lipo-
risk of atherosclerosis. Both diet and genetic factors contribute proteins by using polyethylene glycol 6000. Clin. Chem. 27:371–374.
to responsiveness of lesion formation in the environment of di- 21. Folch, J., M. Lees, and G.H. Sloane Stanley. 1957. A simple method for
abetes and crosses of these strains treated with STZ and high the isolation and purification of total lipids from animal tissues. J. Biol. Chem.
226:497–509.
fat diets can be used to identify diabetogenic factors contribut- 22. Carr, T.P., C.J. Andresen, and L.L. Rudel. 1993. Enzymatic determina-
ing to the accelerated atherosclerosis. The hyperglycemic tion of triglyceride, free cholesterol, and total cholesterol in tissue lipid extracts.
BALB/c mouse provides a convenient model in which to iden- Clin. Biochem. 26:39–42.
23. Paigen, B., A. Morrow, P.A. Holmes, D. Mitchell, and R.A. Williams.
tify genetic, dietary, and diabetogenic factors contributing to 1987. Quantitative assessment of atherosclerotic lesions in mice. Atherosclero-
accelerated fatty streak lesions of complex cellularity. sis. 68:231–240.

1772 Kunjathoor et al.

 24. Hsu, S.M., L. Raine, and H. Fanger. 1981. Use of avidin-biotin-peroxi- scription factor activation in response to an atherogenic diet in mice. J. Clin. In-
dase complex (ABC) in immunoperoxidase techniques: a comparison between vest. 91:2572–2579.
ABC and unlabeled antibody (PAP) procedure. J. Histochem. Cytochem. 29: 31. Mullarkey, C.J., D. Edelstein, and M. Brownlee. 1990. Free radical gen-
577–580. eration by early glycation products: a mechanism for accelerated atherogenesis
25. Bonnevie-Nielsen, V., M.W. Steffes, and Å. Lernmark. 1981. A major in diabetes. Biochem. Biophys. Res. Commun. 173:932–939.
loss in islet mass and B-cell function precedes hyperglycemia in mice given mul- 32. Baynes, J.W. 1991. Role of oxidative stress in development of complica-
tiple low doses of streptozotocin. Diabetes. 30:424–429. tions in diabetes. Diabetes. 40:405–412.
26. LeBoeuf, R.C., M.H. Doolittle, A. Montcalm, D.C. Martin, K. Reue, 33. Stary, H.C., A.B. Chandler, S. Glagov, J.R. Guyton, W. Insull, M.E.
and A.J. Lusis. 1990. Phenotypic characterization of the Ath-1 gene controlling Rosenfeld, S.A. Schaffer, C.J. Schwartz, W.D. Wagner, and R.W. Wissler. 1994.
high density lipoprotein levels and susceptibility to atherosclerosis. J. Lipid Res. A definition of initial, fatty streak, and intermediate lesions of atherosclerosis.
31:91–101. Circulation. 89:2462–2478.
27. Stewart-Phillips, J.L., and J. Lough. 1991. Pathology and atherosclerosis 34. Qiao, J.-H., P.-Z. Xie, M.C. Fishbein, J. Kreuzer, T.A. Drake, L.L. De-
in cholesterol-fed susceptible mice. Atherosclerosis. 90:211–218. mer, and A.J. Lusis. 1994. Pathology of atheromatous lesions in inbred and ge-
28. Nishina, P.M., J. Verstuyft, and B. Paigen. 1990. Synthetic low and high netically engineered mice. Genetic determination of arterial calcification. Arte-
fat diets for the study of atherosclerosis in the mouse. J. Lipid Res. 31:859–869. rioscler. Thromb. 14:1480–1497.
29. Nishina, P.M., J.K. Naggert, J. Verstuyft, and B. Paigen. 1994. Athero- 35. Nakashima, Y., A.S. Plump, E.W. Raines, J.L. Breslow, and R. Ross.
sclerosis in genetically obese mice: the mutant obese, diabetes, fat, tubby, and 1994. ApoE-deficient mice develop lesions of all phases of atherosclerosis
lethal yellow. Metab. Clin. Exp. 43:554–558. throughout the arterial tree. Arterioscler. Thromb. 14:133–140.
30. Liao, F., A. Andalibi, F.C. deBeer, A.M. Fogelman, and A.J. Lusis. 36. Pfeiffer, A., and H. Schatz. 1995. Diabetic microvascular complications
1993. Genetic control of inflammatory gene induction and NF-kB-like tran- and growth factors. Endocrinol. Diabetes. 103:7–14.

Atherosclerosis in Diabetic Mice 1773