Diabetes Accelerates Smooth Muscle Accumulation in Lesions of Atherosclerosis: Lack of Direct Growth-Promoting Effects of High Glucose Levels

Diabetes was induced in male Yorkshire swine 8–10 weeks of age (18 ± 3 kg) by administering 50 mg/kg streptozotocin daily into an ear vein for 3 days, as previously reported in preliminary studies (10). This dose of streptozotocin induces hyperglycemia without the need for insulin treatment. Three days after the inducement of diabetes, diabetic pigs and age-matched nondiabetic pigs were fed a cholesterol-rich diet (1.5% cholesterol and 19.5% lard [% weight]) for 20 weeks to induce atherosclerosis. The combination of diet and streptozotocin resulted in four treatment groups of animals: nondiabetic/normolipemic, diabetic/normolipemic, nondiabetic/hyperlipemic, and diabetic/hyperlipemic. After 20 weeks, body weights were 55 ± 6 kg for nondiabetic/normolipemic animals, 46 ± 6 kg for diabetic/normolipemic animals, 52 ± 9 kg for nondiabetic/hyperlipemic animals, and 39 ± 2 kg for diabetic/hyperlipemic animals (n = 10 animals/group).

In the present study, thoracic aortas from nondiabetic hyperlipemic animals and diabetic hyperlipemic animals were compared. In addition, thoracic aortas from nondiabetic normolipemic and diabetic normolipemic pigs were investigated as controls. Fresh segments of the thoracic aortas (∼2 g) were used. The segments were taken from the same anatomical site for each animal (between the third and fourth intercostal artery) and were ∼3 cm long. After cleaning the segments of adventitial fat, two arterial rings (∼5 mm in length) were taken from each side of the 3-cm specimen from each animal. These rings were fixed in methyl Carnoy’s fixative for immunohistochemical studies. The remaining tissue was used to generate segments for studies of SMC proliferation in isolated arterial tissue (organ culture), to isolate primary SMCs, and to establish cell cultures under normal and high-glucose conditions, as further described.

Blood glucose levels were measured daily during the first 2 weeks of the study and then weekly thereafter, using an Ames Glucometer II (Meditend Medicare Services, Cape Town, South Africa). All blood glucose measurements were completed after an 18-h fast. Blood for this and cholesterol measurement was obtained without sedation by pricking an ear vein. Plasma lipids were measured every 2 weeks. Total plasma cholesterol and triglycerides were measured by standard enzymatic assay kits (Sigma, St. Louis, MO). Lipoproteins were isolated and quantified by ultracentrifugation using the method of Mahley et al. (11). Fasting insulin levels were measured in serum using a radioimmunoassay (Diagnostics Products, Los Angeles, CA).

Fixed tissue was used for immunohistochemical identification of SMCs, macrophages, and proliferating cells expressing proliferating-cell nuclear antigen (PCNA) (DNA polymerase δ-binding protein). Tissues were embedded in paraffin and cut into 5-μm sections. The sections were deparaffinized in three changes of Histoclear (National Diagnostics, Atlanta, GA) and rehydrated in 100, 95, and 75% ethyl alcohol. Endogenous peroxides were then blocked by a 10-min incubation with 0.3% hydrogen peroxide/1% sodium azide. The primary antibodies used were a mouse monoclonal anti-smooth muscle α-actin antibody (DAKO , Carpenteria, CA) at a 1:200 dilution, a mouse monoclonal anti-porcine macrophage antibody (ATCC HB 142.1; American Type Culture Collection, Rockville, MD) at 3–30 μg/ml, a mouse monoclonal anti-PCNA antibody (Santa Cruz Biotechnology, Santa Cruz, CA) at a 1:400 dilution, or purified mouse IgG2a or IgG2b as negative controls (Zymed Labs, South San Francisco, CA). The negative controls did not result in the staining of the specimens. Porcine small intestine was used as a positive control for the PCNA and macrophage antibodies. All antibodies were diluted in phosphate-buffered saline (PBS)/3% rabbit serum and were incubated with the specimens overnight at 4°C. The sections were washed three times in PBS and then incubated with a biotinylated rabbit anti-mouse IgG2A secondary antibody at a dilution of 1:200–1:2,000 for 30 min at room temperature. The slides (smooth muscle α-actin and PCNA) were developed using the streptavidin–horseradish-peroxidase complex (Vector Laboratories, Burlingame, CA) at a 1:5,000 dilution for 30 min at room temperature and a subsequent 10-min incubation with diaminobenzidine and a 2–5-min counterstaining with methyl green. This procedure results in a brown reaction product. In some experiments, the sections were double-stained with the smooth muscle α-actin antibody and the PCNA antibody. In these experiments, the α-actin antibody was developed as previously described, whereas the PCNA was developed in the presence of NiCl, using the FAST DAB kit with metal enhancer (Sigma), which results in a black reaction product localized to the cell nucleus. The sections stained with the porcine anti-macrophage antibody were developed using the Vector Red Alkaline Phosphatase Substrate Kit I (Vector Laboratories), which results in a red reaction product.

Effects of high glucose on SMC proliferation in isolated arterial tissues were studied using an organ culture model. A segment of the thoracic aorta from each animal was divided into two equal parts. After removal of the endothelium, the neointima, and the adventitia, the medial smooth muscle layer was cut into 1.5 × 1.5 × 1.5-mm pieces and was placed in tissue culture flasks. The tissue from one part of the aortic segment was subjected to normal glucose levels (5.6 mmol/l), whereas the other part was subjected to high glucose levels (25 mmol/l) in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco-BRL, Rockville, MD) and 10% fetal bovine serum (FBS) (Gibco). The tissue culture medium was changed daily. Following a 5-day incubation, arterial segments were fixed in methyl Carnoy’s fixative for immunohistochemical studies of SMC proliferation using the smooth muscle α-actin/PCNA double-staining method previously described. The 5-day time point was chosen based on our observation that the first SMCs were consistently observed to migrate out from the tissue under both normal and high-glucose conditions at this time, a sign that a process of SMC dedifferentiation has been initiated. Previous studies of porcine SMCs in thoracic aorta organ cultures show that no significant SMC proliferation occurs at earlier times (12).

Human newborn (2 days to 3 months of age) thoracic aortas were obtained from infants following accidental death, death from sudden infant death syndrome, or congenital defects. Porcine SMCs from the four groups of animals and the human SMCs were isolated by the explant method in DMEM containing 5.6 or 25 mmol/l glucose and cultured and characterized as previously described (6,13). The same batch of glucose-free pyruvate-free DMEM was used to make media containing different glucose concentrations. All experiments were performed in DMEM with 1% human plasma-derived serum (PDS) and 80 μmol/l pyruvate, rather than the 1 mmol/l normally included in cell culture media, because plasma levels of pyruvate are in the range of 30–100 μmol/l and because pyruvate can compete with glucose as an energy source. In some experiments, the cells were stimulated for indicated periods of time with human recombinant platelet-derived growth factor (PDGF)-BB (Upstate Biotechnology, Lake Placid, NY), human recombinant IGF-I (Upstate Biotechnology), vehicle (10 mmol/l acetic acid/0.25% bovine serum albumin [BSA]), or 10% FBS. In additional experiments, the DMEM was modified to reduce amino acid concentrations to near human physiological plasma levels, because excess amino acids can be used as metabolic substrates. The amino acid and vitamin composition of this DMEM, referred to as “low–amino acid DMEM,” was fivefold lower than in the previously described DMEM, whereas the salt concentrations were similar. Glucose levels in the cell culture medium were monitored, and the medium was replaced frequently to avoid glucose depletion during experiments.

Glucose and lactate levels in the cell culture media were measured using a colorimetric glucose kit (Trinder) and a colorimetric lactate kit (Sigma). Total glucose consumption was measured as the loss of glucose from cell culture media. Initial analysis showed that the loss of glucose is linear for several days and over a wide range (2–25 mmol/l) of glucose concentrations, as long as there is no severe depletion (<2 mmol/l) of glucose (data not shown). The flux of glucose carbon through the citric acid cycle and the pentose phosphate pathway was determined by using d-[1-¹⁴C]glucose and d-[6-¹⁴C]glucose (specific activity 56.0 Ci/mol) (Amersham Pharmacia Biotech, Piscataway, NJ). This determination is possible because the first carbon of glucose is metabolized to CO₂ in the citric acid cycle and in the pentose phosphate pathway, whereas the sixth carbon of glucose is metabolized only in the citric acid cycle. Because the first and sixth carbon of glucose are released as CO₂ in equimolar amounts from the citric acid cycle, the amount of glucose metabolized in the pentose phosphate pathway can be obtained by subtracting 6-¹⁴CO₂ from the 1-¹⁴CO₂ generated (14,15). Human SMCs were plated in 24-well trays (50,000 cells/well) and incubated in the presence of 1% PDS, 80 μmol/l pyruvate, and 5.6 mmol/l glucose, 5.6 mmol/l glucose/19.4 mmol/l mannitol (an osmolarity control), or 25 mmol/l glucose for 6 days. The cells were then incubated during 1 h with 1 μCi/well (medium containing 5.6 mmol/l glucose) or 4.5 μCi/well (medium containing 25 mmol/l glucose), d-[1-¹⁴C]glucose or d-[6-¹⁴C]glucose in 0.5 ml culture medium. The CO₂ liberated by the cells from radiolabeled glucose was collected onto Millipore AP25 filters, which were saturated with 100 μl 3.5 mol/l KOH and attached to rubber stoppers using pins. The amount of released ¹⁴CO₂ is linear for up to 4 h using the conditions previously described. After incubation, the media were acidified by adding sulfuric acid to a final concentration of 0.3 mol/l, and the incubation was allowed to continue for another hour. The amount of radioactive CO₂ bound to the filters was measured in a β-scintillation counter. The radioactive CO₂ formed in parallel incubations without cells was subtracted from the radioactive CO₂ formed in the presence of SMCs.

To investigate the extent to which SMCs use amino acids present in the cell culture medium, CO₂ formation from three different groups of amino acids was measured: alanine, glutamic acid, and leucine. Human SMCs were incubated for 1 h in the presence of 2.5 μCi/0.5 ml l-[U-¹⁴C]alanine (specific activity 152 Ci/mol), l-[U-¹⁴C]leucine (specific activity 315 Ci/mol), or l-[U-¹⁴C]glutamic acid (specific activity 247 Ci/mol) in DMEM, 1% PDS, 80 μmol/l pyruvate, and 5.6 mmol/l glucose, 5.6 mmol/l glucose/19.4 mmol/l mannitol, or 25 mmol/l glucose. In some experiments, the cells were incubated with 2.5 μCi/0.5 ml l-[U-¹⁴C]lactic acid (154 Ci/mol) in the presence of 1 mmol/l unlabeled l-lactic acid to measure the ability of SMCs to utilize lactic acid as a fuel substrate. The concentration of unlabeled lactic acid added to the incubations corresponds to the concentration that is generated during 5 h of spontaneous release of lactic acid from these cells. Radioactive CO₂ was trapped onto Millipore AP25 filters, as previously described.

Expression of the GLUT-1 glucose transporter was studied in human SMCs exposed to the indicated glucose concentrations for 6 days. Membrane fractions were prepared after scraping and sonicating the cells for 2 × 10 s using a Braun-Sonic 2000 sonicator (B. Braun Biotech, Allentown, PA) at 50% output in a buffer containing 50 mmol/l HEPES (pH 7.4), 50 mmol/l NaCl, 1 mmol/l MgCl₂, 2 mmol/l EDTA, 10 mmol/l pyrophosphate, 10 mmol/l NaF, 500 μmol/l Na₃VO₄, 1 mmol/l dithiothreitol, 1 mmol/l benzamidine, 1 mmol/l leupeptin, 1 mmol/l pepstatin, and 1 mmol/l aprotinin. The samples were centrifuged for 10 min at 10,000g at 4°C in a microfuge. The supernatant was then centrifuged for 60 min at 100,000g at 4°C, which resulted in a pellet rich in plasma membranes and a supernatant. The pellets were resuspended in the previously described buffer with the addition of 1% Triton X-100. Protein concentrations in the supernatant and pellet fractions were quantitated either by the bicinchoninic acid protein assay (Pierce, Rockford, IL) or by the Bio-Rad protein assay according to Bradford (15a) (Bio-Rad Laboratories, Hercules, CA). GLUT-1 levels were analyzed using Western blot analysis and a rabbit polyclonal GLUT-1 antibody (MYM antiserum) at a 1:3,000 dilution (Chemicon International , Temecula, CA).

Cells were plated at a density of 25,000 cells/well in 24-well trays and incubated in 2 ml DMEM with 80 μmol/l pyruvate, 1% human PDS, and indicated glucose concentrations ranging between 5.6 and 50 mmol/l, and/or mannitol as an osmolarity control. For DNA synthesis measurements, PDGF-BB or IGF-I was added after 6 days of incubation in the different glucose concentrations. The cells were incubated for an additional 18 h and subsequently labeled with [³H]thymidine, as previously described (6).

Cell proliferation was also measured by determining cell number. Cells were incubated in the presence of indicated glucose concentrations with or without 1 nmol/l PDGF-BB, IGF-I, 10% FBS, or vehicle for up to 9 days. Mannitol was used as an osmolarity control. The cells were trypsinized, fixed in Holley’s fixative (3.7% formaldehyde, 86 mmol/l NaCl, and 106 mmol/l Na₂SO₄), and counted using a cell counter (Coulter , Hialeah, FL).

Differences between group mean values were analyzed using analysis of variance, and pairwise comparisons between means were made using two-tailed Student’s t test. Statistical significance was defined as P < 0.05. All values represent means ± SD.

Animals from the four treatment groups were examined with emphasis on comparing the two atherosclerotic (hyperlipemic) groups (diabetic and nondiabetic). The mean blood glucose in the two nondiabetic groups was 116 ± 7 mg/dl (6.4 ± 0.4 mmol/l), and the mean plasma cholesterol in the two normolipemic groups was 81 ± 13 mg/dl (2.1 ± 0.3 mmol/l). The mean blood glucose level in diabetic hyperlipemic animals (339 ± 103 mg/dl; 18.8 ± 5.7 mmol/l) was not significantly different from that in diabetic animals fed the normal diet (328 ± 69 mg/dl; 18.2 ± 3.8 mmol/l). No significant differences in mean plasma cholesterol levels were seen between diabetic and nondiabetic animals during the 20-week study. In the hyperlipemic groups, total plasma cholesterol levels were 584 ± 133 mg/dl (15.1 ± 3.4 mmol/l) in diabetic animals and 519 ± 116 mg/dl (13.4 ± 3.0 mmol/l) in nondiabetic animals at the end of the study. Serum insulin levels after an 18-h fast were measured in a subset of animals (three animals per group). At 20 weeks, insulin levels were 37.8 ± 16 pmol/l in nondiabetic animals fed a normal diet, 29.4 ± 7 pmol/l in diabetic animals fed a normal diet, 35.2 ± 4 pmol/l in nondiabetic animals fed a lipid-rich diet, and 35.5 ± 16 pmol/l in diabetic animals fed a lipid-rich diet. Corresponding blood glucose values in these animals were 6.1 ± 1.2, 12.2 ± 1.7, 6.4 ± 1.3, and 13.8 ± 2.7 mmol/l, respectively. The lack of fasting hypoinsulinemia in diabetic animals at 20 weeks may be because of the partial regeneration of β-cells that occurs over time in this animal model. Thus, ∼90% of the β-cells are lost at 2 weeks after the inducement of diabetes, and ∼80% are still lost at 20 weeks (data not shown).

Thoracic aortas from hyperlipemic animals with and without diabetes for 20 weeks were fixed and used for immunohistochemical determination of SMC involvement and extent of cell proliferation in the atherosclerotic lesions. The thoracic aorta segments used in this study had an even coverage of lesions. It has been previously reported that diabetes results in a twofold increase in total aortic surface area covered by atherosclerotic lesions at 12 weeks using this animal model (10). Our results show that thoracic aortas from diabetic animals fed a cholesterol-rich diet consistently contain advanced lesions of atherosclerosis with thick fibrous plaques consisting of SMCs, as shown by the smooth muscle α-actin staining in the neointima (Figs. 1A and B). In Fig. 1C, these lesions also have a core of macrophage foam cells, as judged by the positive staining using a porcine macrophage antibody. This antibody recognizes an unknown antigen expressed by porcine macrophages (16). There are many PCNA-positive cells that have entered the cell cycle in these lesions (Fig. 1D). A significant number of SMCs in the neointima express PCNA, as shown by PCNA and smooth muscle α-actin double-staining experiments (Fig. 1E). Several SMCs in the macrophage-rich core region of the lesion also express PCNA (Fig. 1F). SMC accumulation and expression of PCNA were only observed in areas invaded by macrophages.

Conversely, in aortas from nondiabetic animals fed the same cholesterol-rich diet, there is only a thin neointima, consisting of foam cells that exhibit typical monocyte/macrophage-like morphology (17) and do not express smooth muscle α-actin (Fig. 1G). There are very few SMCs in these lesions of atherosclerosis, and most of the cells stain positive for the macrophage antigen (Fig. 1H). Furthermore, there is little PCNA-positive staining, as shown in Fig. 1I. Sections from nondiabetic and diabetic animals fed a normal diet were studied for comparison. No intimal thickening or foam cell infiltration is seen in aortas from nondiabetic animals or diabetic animals fed a normal diet at the time point studied (data not shown). As expected, the principal cell type in the media is SMC. In addition, no cell proliferation is observed in these animals (data not shown). Thus, in hyperlipemic animals, diabetes markedly accelerates atherosclerosis and stimulates SMC proliferation and accumulation in the lesions.

To investigate whether high glucose can directly stimulate SMC proliferation in arterial tissue, medial segments from the thoracic aortas of nondiabetic and diabetic hyperlipemic animals were subjected to normal glucose (5.6 mmol/l) or high glucose (25 mmol/l) conditions for 5 days. There is no SMC proliferation in the media of freshly isolated aortas from diabetic or nondiabetic animals, as shown in Figs. 2A and D. After 5 days in organ culture, however, a significant number of medial SMCs are proliferating, as judged by the positive double-staining of PCNA and smooth muscle α-actin (Fig. 2B,C,E, and F). The PCNA-positive SMCs are distributed in streaks throughout the isolated arterial segments. Hyperglycemia does not result in an increased number of PCNA-positive SMCs in arterial segments obtained from diabetic animals (compare Figs. 2B and C) or nondiabetic animals (compare Figs. 2E and F). To quantify SMC proliferation, the number of cells positive for both smooth muscle α-actin and PCNA was determined within four random sample regions selected from each sample. A nonsignificant tendency of reduced SMC proliferation in the tissue subjected to high-glucose conditions was observed (data not shown). Thus, high glucose does not directly stimulate SMC proliferation in isolated arterial tissue from diabetic or nondiabetic hyperlipemic animals.

To determine whether high glucose stimulates primary SMC proliferation or the ability of the SMC to respond to mitogens, outgrowth of SMCs from arterial explants was monitored daily by phase-contrast microscopy. Arterial explants from the nondiabetic and diabetic hyperlipemic animals were placed in cell culture media containing 10% serum and 5.6 mmol/l glucose (normal) or 25 mmol/l glucose (high). The medium was changed daily. The first outgrowth of SMCs from the arterial explants was observed by days 3–5. There was no detectable stimulatory effect of high glucose on outgrowth of primary SMCs from the arterial tissue explants over a 3-week period (data not shown). In addition, high glucose did not stimulate the outgrowth of SMCs from explants obtained from normolipemic animals (data not shown).

To test the possibility that different populations of SMCs with different sensitivity to glucose are present within the arterial wall, SMCs were isolated from each aorta under both normal and high-glucose conditions. The effects of high glucose on the proliferative capacity of these SMCs were studied in subcultured (passages 2–4) SMCs from nondiabetic/hyperlipemic animals and diabetic/hyperlipemic animals isolated under normal or high-glucose conditions and subsequently subjected to 5.6 mmol/l glucose, 5.6 mmol/l glucose/19.4 mmol/l mannitol (an osmolarity control), or 25 mmol/l glucose for 6 days. Nondiabetic/normolipemic and diabetic/normolipemic animals were also studied for comparison. As shown in Figs. 3A and B, high glucose does not stimulate basal cell proliferation (vehicle-treated cells). Furthermore, the ability of PDGF-BB to stimulate cell proliferation is similar under normal and high-glucose conditions. IGF-I does not significantly induce an increase in SMC number (6) under any glucose concentration. There is no marked difference in the mitogenic effects of 10% FBS under the different glucose conditions (Fig. 3A,B). Similar results were obtained for SMCs isolated from the arterial tissue in 5.6 and 25 mmol/l glucose, indicating that SMCs selected under high-glucose conditions do not show a different response to glucose compared with SMCs selected under normal glucose conditions (Fig. 3). Moreover, no significant differences in the responsiveness to high glucose are observed in SMCs isolated from the two groups of hyperlipemic animals (Fig. 3) or in SMCs isolated from the normolipemic animals (data not shown). In one diabetic hyperlipemic animal, the lesion was advanced enough to independently isolate lesion (intimal) SMCs and SMCs from the underlying media. High glucose did not enhance basal or growth factor–stimulated proliferation of SMCs from either the fibrous cap of the lesion or the underlying media (data not shown).

The effects of high glucose on DNA synthesis and proliferation of human aortic SMCs were also investigated. As shown in Fig. 4A, there is no difference in basal or PDGF- or IGF-I–stimulated DNA synthesis in human SMCs kept in 5.6 vs. 25 mmol/l glucose for 6 days. Accordingly, 25 mmol/l glucose does not affect basal or PDGF-stimulated human SMC proliferation during a 9-day incubation under the conditions previously described (Fig. 4B). Lowering the glucose level to 12.5 mmol/l or raising it to 50 mmol/l does not result in an increased basal or PDGF-stimulated proliferation when compared with 5.6 mmol/l glucose (data not shown). Futhermore, no differences in cell proliferation between the different glucose conditions are seen when the cells are grown in 10% FBS or when the cultures are kept in different glucose concentrations for as long as two passages (∼3 weeks) (data not shown). Furthermore, no differences in cell proliferation between cells in different glucose concentrations are observed when the experiments are performed in low–amino acid DMEM (data not shown).

Altogether, the results show that high glucose does not directly stimulate proliferation or enhance growth factor–induced proliferation in primary (nonpassaged) porcine SMCs from diabetic or nondiabetic animals, or in early passage porcine or human SMCs.

Next, we wanted to verify that the lack of mitogenic effects of high glucose was not because of a downregulation of SMC glucose metabolism under high-glucose conditions. Therefore, glucose consumption and lactate formation were measured in SMCs subjected to normal or high glucose levels. Porcine and human SMCs consume 0.2–0.8 pmol lactate · h^–1 · cell^–1 and produce 0.4–1.6 pmol lactate · h^–1 · cell^–1, depending on SMC strain. Because 2 mol lactate is formed per mole of glucose consumed, the results show that lactate formation can account for ∼100% of the glucose consumed. High glucose does not markedly alter total glucose consumption or lactate formation in human (Fig. 5A) or porcine SMCs (data not shown), although slight increases are normally observed. Moreover, high glucose does not lead to a decreased expression or plasma membrane association of the main glucose transporter, GLUT-1, in human SMCs (Fig. 5B). These findings are in contrast to the downregulation of GLUT-1 expression observed after a 24-h exposure of rat or bovine SMCs to high-glucose conditions (18,19). Further analyses show that high glucose increases glucose metabolism through the pentose phosphate pathway by 90% and through the citric acid cycle by 240% (Fig. 5C). However, glucose metabolized through the pentose phosphate pathway and citric acid cycle only accounts for 4 and 0.5% of total glucose consumed, respectively, under normal glucose conditions and 8 and 1.5%, respectively, under high-glucose conditions (compare Figs. 5A and C). Thus, although high glucose stimulates glucose metabolism through the pentose phosphate pathway and the citric acid cycle in SMCs, the relative contribution of these pathways to total glucose metabolism is small.

Finally, we wanted to confirm that glucose is indeed the principal source of metabolic fuel for SMCs under the conditions used in this study. To investigate the extent to which human SMCs use amino acids and lactic acid as fuel, we measured CO₂ formation from three amino acids metabolized through different mechanisms. Because SMCs produce large amounts of lactic acid, we also measured the ability of these cells to utilize lactic acid as a substrate. As shown in Fig. 6, CO₂ formation from the different amino acids is very low compared with CO₂ formed from glucose. Leucine is the preferred amino acid of the three amino acids investigated, and CO₂ formation from leucine is approximately fivefold greater than that generated from alanine or glutamate. As expected, SMCs do not use significant amounts of lactic acid as a metabolic fuel (Fig. 6). There are no differences in amino or lactic acid use between SMCs incubated under normal and high glucose levels (data not shown). Thus, glucose is the preferred metabolic fuel for isolated SMCs under the conditions used in this study.

The animals used in this study are part of a larger study (Gerrity RA, Natarajan R, Nadler JL, Kimsey T, unpublished observations) in which the progression of diabetes-accelerated atherosclerosis has been followed for up to 48 weeks in the recently developed porcine model. The present results are consistent with the findings of this larger study and of preliminary studies (10), in that the thoracic aortic lesions at 20 weeks are more advanced in diabetic hyperlipemic animals than in nondiabetic hyperlipemic animals. An accelerated formation of fatty streaks in diabetic miniature swine fed a cholesterol-rich diet for 12 weeks has also been recently reported (20). Of particular interest in the present study, even as early as 20 weeks, the lesions in the diabetic animals exhibit a well-developed fibrous cap rich in SMCs. In contrast, lesions in nondiabetic animals consist predominantly of monocyte/macrophage foam cells, a finding consistent with our earlier studies in the 20-week time period (17,21–23). Furthermore, proliferation of SMCs in the lesion, measured as expression of PCNA, is markedly increased in diabetic animals. These findings indicate that hyperglycemia may play a role in the accelerated atherosclerosis and smooth muscle accumulation in the diabetic animals. However, hyperglycemia alone is not sufficient to induce significant diabetes-accelerated atherosclerosis, at least not during the time frame studied, because there are no lesions in the arteries of diabetic animals fed a normal diet (this study and Gerrity et al., [unpublished observations]). In addition, the greater severity of lesions in diabetic animals cannot be attributed to higher total serum cholesterol levels induced by diabetes, as these were not statistically different between diabetic and nondiabetic animals and cannot be attributed to differences in total LDL + VLDL-cholesterol or HDL-cholesterol, which also do not differ at 20 weeks (Gerrity et al., [unpublished observations]). However, as in diabetic humans, diabetic swine in this model demonstrate elevated triglyceride levels (Gerrity et al., [unpublished observations]) that are not seen in nondiabetic hyperlipemic animals (23). In addition, it is possible that, even if lipid levels are similar in nondiabetic and diabetic hyperlipemic animals, differences in lipid oxidation (24) may stimulate SMC proliferation. Because (as discussed further) hyperglycemia alone does not, in this model, directly induce SMC proliferation, the accelerating effect of diabetes on SMC accumulation may be mediated by lipid abnormalities or by effects secondary to the hyperglycemic state, which act synergistically with hyperlipemia.

It is possible that hyperglycemia, in combination with hyperlipemia, induces atherogenic alterations in endothelial cells and/or monocytes/macrophages that in turn accelerate SMC accumulation. Interestingly, nonenzymatic glycation of molecules resulting from hyperglycemia or lipid peroxidation (25) has been indicated to contribute to diabetes-accelerated atherosclerosis in apolipoprotein E–deficient mice; blocking the interaction of the receptor for advanced glycation end products (RAGE) with its ligands results in a delayed lesion formation (26). It has also been recently shown that the activation of RAGE can mediate the expression of vascular cell adhesion molecule-1 and intracellular adhesion molecule-1 on endothelial cells that may lead to increased monocyte infiltration into the arterial wall and activation of macrophages (27). Thus, it is possible that an increased accumulation and activation of macrophages with subsequent release of smooth muscle growth regulatory molecules in diabetic lesions may be responsible for the increased SMC accumulation in these lesions.

High glucose has been previously reported to have no effect on the proliferation of cultured arterial SMC (28–31). Other studies have found a stimulatory effect of high glucose on proliferation of cultured SMCs (32–39). Interpretation of these results is often complicated by the facts that these studies were utilizing SMCs from species that do not develop diabetes-accelerated atherosclerosis (e.g., rats and rabbits), that SMCs kept in culture for several passages were studied, and that different cell culture conditions were used. To clarify the issue of direct effects of high glucose on SMC proliferation, we investigated SMCs from arterial tissues in which diabetes markedly accelerated atherosclerosis and SMC accumulation/proliferation (Fig. 1) under both normal and high-glucose conditions. We also took precautions to make certain that the cell culture conditions used would produce reliable results; pyruvate levels were reduced to physiological plasma levels, media were changed frequently, glucose levels were monitored to avoid glucose depletion, and a variety of glucose concentrations and incubation times were investigated. The following chain of results shows that high glucose does not directly stimulate SMC proliferation: 1) diabetic hyperglycemic pigs fed a normal diet do not show increased SMC proliferation in vivo; 2) high glucose does not stimulate proliferation of SMCs in organ cultures of aortas from hyperlipemic nondiabetic or diabetic animals; 3) outgrowth of primary SMCs from arterial tissues from these animals is not stimulated by high glucose levels; 4) high glucose does not stimulate proliferation of early passage porcine SMCs isolated under normal or high-glucose conditions; and 5) high glucose does not stimulate proliferation of human SMCs isolated under normal or high-glucose conditions.

Also, it is unlikely that nonenzymatic glycation, as a result of long-term high glucose levels, significantly stimulates SMC proliferation, because SMCs subjected to high glucose for 20 weeks in vivo and subsequently in organ culture do not show increased proliferative capacity compared with SMCs subjected to normal glucose conditions. Moreover, nonenzymatic glycation of fibronectin (40) or BSA (41) has previously been shown to not stimulate SMC proliferation. Thus, our data support the conclusion that hyperglycemia is not sufficient to stimulate SMC proliferation in vivo.

Smooth muscle glucose metabolism is characterized by a high rate of conversion of glucose to lactate under aerobic conditions (aerobic glycolysis), a phenomenon shared with many tumor cells and retinal cells (42–45). Although high rates of glycolysis in some tissues were discovered in the 1920s (46), the phenomenon is still counter-intuitive, because conversion of glucose to lactate is far less effective in ATP production than conversion of glucose through the citric acid cycle and oxidative phosphorylation. It has been suggested that the high conversion of glucose to lactate is required for cell proliferation of some cell types (47,48) and that glycolysis may protect cells against oxidative stress (48). By generating most of their ATP through glycolysis, cells can avoid generation of reactive oxygen species through mitochondrial respiration and at the same time produce pyruvate, which acts as an antioxidant (48). Our results show a high rate of glycolysis (90–95% of glucose is metabolized to lactate) in isolated porcine and human SMCs under both normal and high-glucose conditions. Conversely, glucose metabolized through the pentose phosphate pathway and citric acid cycle accounts for only 4 and 0.5%, respectively, of consumed glucose under normal glucose conditions and 8 and 1.5%, respectively, under high-glucose conditions. An increased rate of the pentose phosphate pathway appears to be responsible for the activation of diacylglycerol-sensitive protein kinase C (PKC) isoforms observed under high-glucose conditions (49) and has also been suggested to lead to activation of nuclear transcription factor κB and cell proliferation (32,34,50). Thus, we propose that the high rate of glycolysis may be responsible for the relative insensitivity of SMCs to adverse effects of high glucose mediated though the pentose phosphate pathway (i.e., activation of PKC) and the citric acid cycle (i.e., increased generation of oxygen-free radicals through oxidative phosphorylation). In this context, it is interesting that endothelial cells appear to be more sensitive to high glucose than SMCs (51–53). This concept is consistent with the notion that endothelial cells and/or monocytes/macrophages may be the primary targets of hyperglycemia and that SMC proliferation is induced secondary to these changes.

In conclusion, non–insulin-treated diabetes markedly accelerates atherosclerosis and SMC accumulation and proliferation in this new hypercholesterolemic porcine model. However, high glucose levels do not directly stimulate proliferation of isolated SMCs, and hyperglycemia does not result in SMC accumulation in the absence of a cholesterol-rich diet. Thus, the stimulatory effect of diabetes on SMC accumulation and proliferation in vivo is likely to be dependent on plasma lipids and/or on other cell types present in the lesion, rather than on a direct growth-promoting effect of hyperglycemia on SMCs.

This work was supported by a Career Development Award from the American Diabetes Association (to K.E.B.); the Martha Shamberger Medical Research Fund (to R.G.G.); Grants HL62887 (to K.E.B.) and HL55798 from the National Institutes of Health; and Grants 193134 and 995009 (to R.G.G.) from the Juvenile Diabetes Foundation.

We thank Roderick Browne for sharing his expertise on histology and immunohistochemistry.