Hepatic Lipase mRNA, Protein, and Plasma Enzyme Activity Is Increased in the Insulin-Resistant, Fructose-Fed Syrian Golden Hamster and Is Partially Normalized by the Insulin Sensitizer Rosiglitazone

Male Syrian golden hamsters (Mesocricetus auratus) were purchased from Charles River (Quebec, Canada). All animals were housed in pairs and were given free access to food and water. Hamsters were fed a normal diet for 7 days to allow acclimatization to the new environment and recovery from the stress of shipping. They were then placed on one of three feeding protocols as previously described (29): 1) normal diet for 5 weeks (DIET), 2) high-fructose diet (hamster diet with 60% fructose; Dyets, Bethlehem, PA) for 5 weeks (FRUC), and 3) high-fructose diet for 5 weeks with rosiglitazone (20 μmol · kg⁻¹ · day⁻¹) (GlaxoSmithKline, Philadelphia, PA) diluted in water and given once daily by gavage for the last 3 weeks of the fructose feeding period (FRUC+RSG). The hamsters’ weights were monitored every week. At the end of the 5 weeks, 14 h after removal of food, the animals received an intraperitoneal injection of heparin for determination of postheparin plasma HL activity as described below. Additional animals did not receive heparin but were killed for fast-protein liquid chromatography (FPLC) determination of plasma lipoproteins and HL mRNA and protein mass on snap-frozen liver biopsy samples. Plasma HL enzymatic activity was also determined in the absence of heparin administration in four normal diet–fed hamsters to determine whether HL can be detected in preheparin plasma, as has been suggested by others (30). In the present study, comparative data between groups are reported for postheparin samples because the administration of heparin allowed us to measure both LPL and HL activities. In addition, the administration of heparin ensured that we were measuring the total HL activity (i.e., bound plus unbound), not just the unbound HL.

We have found that retro-orbital plexus blood sampling from Syrian golden hamsters induces considerable stress and results in an elevation of plasma glucose and free fatty acid (FFA) concentrations. Because hamsters do not have a tail, it is also not possible to obtain blood by tail vein sampling as it is in other rodents. Consequently, measurements of fasting plasma insulin, triglyceride, glucose, and FFAs that are reported in Table 1 represent cumulative data from our laboratory from DIET, FRUC, and FRUC+RSG hamsters that had placement of femoral and arterial catheters for in vivo studies, as previously described by us (29). All animal protocols were approved by the Animal Ethics Committee of the University Health Network.

Following the 5-week treatment period, 50 units heparin was injected intraperitoneally into the hamster, and after 30 min, blood was collected under anesthesia from the abdominal aorta into chilled EDTA tubes on ice to prevent ongoing lipolysis of the samples in the test tubes (DIET, n = 20; FRUC, n = 19; and FRUC+RSG, n = 21). Samples were spun at 3,000 rpm for 15 min, and plasma was stored at −70°C for later analysis of LPL and HL activity.

For measurement of LPL and HL activity, the total plasma triglyceride lipase activity was determined as previously described (31) in pre- and postheparin plasma. In brief, plasma was diluted 1:10 with Krebs-Ringer phosphate buffer, pH 7.4, incubated with a triolein/phosphatidylcholine/albumin emulsion, and trace labeled with glycerol tri[1-¹⁴C]oleate (Amersham, Arlington Heights, IL) in 0.178 mol/l Tris-HCl, 0.11 mol/l NaCl, buffer, pH 8.5, containing 55 mg/ml albumin and 0.01 mg/ml heparin, resulting in final incubation pH of 8.2. The incubation was carried out for 60 min at 37°C, and the resultant FFAs were extracted and [¹⁴C] content determined. The inclusion of a monoclonal antibody specific for LPL allowed for the determination of LPL activity and HL activity (32). The LPL-specific mAb 5D2 antibody used to block LPL activity in this study recognizes multiple species of LPL. Although this mAb was originally produced against bovine LPL, it cross-reacts strongly with human LPL and with LPL from numerous animals tested (33,34). The antibody is effective against hamster LPL because there was substantial inhibition of total postheparin lipase activity, in some cases inhibiting 70% of the activity. A bovine skim milk LPL standard was included in each assay and used to correct for interassay variation. Human postheparin control plasma was assayed to monitor interassay variation. The intra-assay coefficient of variation was 7% for human LPL and 6% for human HL. The interassay coefficient of variation was 8% for LPL and 10% for HL.

Blood was drawn from five hamsters in each of the four treatment groups for FPLC determination of plasma lipoproteins. The animals were then killed, and the liver tissue was frozen in liquid nitrogen at −70°C for later analysis of HL mass and mRNA. HL mass was determined from 150 μg of liver tissue lysate from each animal using chemiluminescent Western blots, as previously described (35), with an affinity-purified anti-rat HL antibody provided by Dr. A. Bensadoon (Cornell University, Ithaca, NY). Purified human HL was used as the positive control.

For measurement of HL mRNA, total RNA was extracted from frozen samples using the RNeasy Mini kit (Qiagen, Valencia, CA). Five micrograms of total RNA was converted to single-stranded cDNA using SuperScript II Reverse Transcriptase (Invitrogen, Carlsbad, CA). Ten percent of this first-strand reaction was used as the template for subsequent PCR. Twenty-four cycles of PCR (95°C for 30 s, 55°C for 60 s, and 72°C for 90 s) were performed using primers specific for HL or β-actin. The primer pairs were as follows: HL: 5′-atg cca att ttg tgg atg cca t-3′, and 5′-ctg ttt tcc cac ttg aac ttg a-3′; and actin: 5′-cct ttt cca gcc ttc ctt c-3′, and 5′-tac tcc tgc ttg ctg atc c-3′.

Fasting plasma was separated by FPLC. Briefly, 100 μl of fasting plasma was applied to a Superose 6HR column, and the samples were eluted in PBS at a rate of 0.4 ml/min in 60 fractions of 500 μl. Fractions corresponding to VLDL, LDL, and HDL were pooled and lipid composition quantified by enzymatic assays using commercially available kits for triglycerides and total cholesterol (Roche Diagnostics, Laval, Quebec, Canada). Apolipoprotein (apo)B was determined using an in-house electroimmunoassay.

All the values are reported as means ± SE. For comparison of HL activity, HL mRNA and protein mass, and VLDL, LDL, and HDL lipid composition between DIET, FRUC, and FRUC+RSG hamsters, one-way ANOVA was used followed by post hoc analysis with Tukey’s test. P < 0.05 was considered to be significant. The 95% CIs were calculated and illustrated for difference plots of FPLC profiles.

There were no significant differences among the groups for any of the measured baseline parameters, although there was a trend for FRUC hamsters to have higher FFA, insulin, and triglyceride levels.

VLDL-triglyceride in the FRUC group was significantly higher than in the other two groups (P < 0.05). LDL cholesterol was higher in FRUC (P < 0.01) and FRUC+RSG (P < 0.05) hamsters than in DIET. HDL triglyceride was higher in FRUC (P < 0.01) and FRUC+RSG animals compared with DIET (P < 0.05), and HDL cholesterol was also higher in the FRUC group than in the DIET group (P < 0.05).

HL activity was increased ∼40% in fructose-fed hamsters compared with control hamsters (112 ± 7 nmol FFA · ml⁻¹ · min⁻¹ in FRUC vs. 66 ± 4 in DIET, P < 0.001) (Fig. 1A). Treatment of fructose-fed hamsters with rosiglitazone resulted in partial normalization of the elevated HL activity (84 ± 7 nmol FFA · ml⁻¹ · min⁻¹ in FRUC+RSG, P < 0.001, versus FRUC, P = 0.2, for FRUC+RSG vs. DIET). There were no significant differences in plasma postheparin LPL activity between the various treatment groups (Fig. 1B).

Preheparin HL activity from the four normal diet–fed hamsters was 71.2 ± 9.1 nmol FFA · ml⁻¹ · min⁻¹ (range 60–80), similar to that detected in postheparin plasma from normal diet–fed hamsters. LPL, on the other hand, was lower in preheparin plasma (16 ± 6.3 nmol FFA · ml⁻¹ · min⁻¹) than in postheparin plasma.

Changes in HL protein mass in liver paralleled the changes in HL activity in plasma. HL mass was increased in fructose-fed hamsters (P < 0.001 for FRUC vs. DIET) and was partially normalized with rosiglitazone treatment (P < 0.05 for FRUC+RSG vs. FRUC). There were no significant differences between the FRUC+RSG and DIET groups (Fig. 2).

Changes in HL mRNA paralleled the changes in postheparin HL activity and HL mass. HL mRNA was increased in fructose-fed hamsters (P < 0.05 for FRUC vs. DIET) and tended to be lower with rosiglitazone treatment, although the latter was not significantly different from that of FRUC hamsters. There were no significant differences between the FRUC+RSG and DIET groups (Fig. 3).

Fructose feeding resulted in a significant increase in large (FPLC fractions 22–24) and small (FPLC fractions 26–29) LDL (Fig. 4A shows the differences between mean LDL profiles for FRUC and DIET, and 95% CIs that do not cross zero indicate significant differences between FRUC and DIET animals, P < 0.05). The addition of rosiglitazone to a high-fructose diet decreased small LDL (Fig. 4B) (P < 0.005 for FRUC+RSG vs. FRUC) and shifted the peak to larger LDL size.

In the present study we have shown that fructose feeding was associated with increases in postheparin plasma HL enzymatic activity, hepatocyte HL protein mass, and HL mRNA. Rosiglitazone, a PPARγ agonist member of the thiazolidinedione class of insulin-sensitizing therapeutic agents, partially normalized the elevated HL mRNA, protein, and plasma enzyme activity in fructose-fed hamsters. Although we cannot prove from these studies that it is the insulin resistance per se that induces these changes in HL, these findings do suggest that the HL gene is up- and downregulated by some factor that is associated with whole-body insulin resistance and insulin sensitization, respectively. The precise molecular mechanism of HL gene regulation in insulin resistance remains to be determined.

We have previously shown (28,29) that high-fructose feeding for a 3- to 5-week period induced significant hypertriglyceridemia, with increased hepatic VLDL production and the development of whole-body insulin resistance in the Syrian golden hamster. We have also recently shown (29) that these abnormalities were ameliorated with rosiglitazone treatment. As in the present study, we could not determine whether the effects of rosiglitazone on HL activity were mediated by the potent insulin-sensitizing effect of rosiglitazone or by some other effect of PPARγ activation. As one may predict with induction of a hypertriglyceridemic insulin-resistant state, LDL particles shifted from large to small with fructose feeding and back again to large with rosiglitazone treatment. HL is known to play an important role in the metabolic conversion of large, buoyant LDL to potentially more atherogenic, small, dense LDL particles, although triglyceride enrichment of LDL particles per se is also thought to play an important role (24,25,26). We cannot be certain, therefore, whether the changes in LDL particle size occurred as a direct consequence of the changes in plasma HL activity or are attributable to some other factor. One could speculate though that the insulin resistance–induced increase in HL activity could play an important role in the formation of small, dense LDL particles that typically characterize the insulin-resistant state (8,26).

The plasma LDL cholesterol concentration increased significantly with fructose feeding (FRUC group) compared with a normal diet (DIET group) (Table 2). Rosiglitazone treatment of fructose-fed hamsters (FRUC+RSG) was associated with a plasma LDL cholesterol concentration that was significantly elevated above that of DIET hamsters. An elevation of LDL cholesterol with rosiglitazone treatment has been noted previously in diabetic humans (36). Despite the elevation of plasma LDL cholesterol, there was a net shift in LDL size from small to large with rosiglitazone in the present study, as discussed above, similar to the LDL size changes that occur in rosiglitazone-treated diabetic humans (36).

With respect to plasma HDL cholesterol concentrations, there was a significant increase in HDL cholesterol with fructose feeding. Rosiglitazone treatment of fructose-fed hamsters tended to reduce HDL cholesterol back toward the levels seen in the DIET group, but the reduction was not significant. The increase in HDL cholesterol that occurs with fructose feeding of hamsters is not typical of the lowering of plasma HDL cholesterol that is commonly seen in humans with insulin resistance syndromes. We can only speculate on the mechanisms whereby fructose feeding raises HDL in Syrian golden hamsters, since this was not the focus of the present study. First, we point out that others have shown similar elevations of HDL cholesterol with fructose feeding of rats (37) and Syrian golden hamsters (38,39). In the latter study, the elevation of VLDL, intermediate density lipoprotein/LDL, and HDL cholesterol accounted for the increased total cholesterol concentrations, whereas the increase in plasma triglycerides occurred entirely in the VLDL density range. To the best of our knowledge, the mechanism of the fructose-induced elevation of HDL cholesterol has not previously been studied. Although hamsters do have cholesteryl ester transfer protein (38), one major difference between rodents and humans is that rodents carry a greater proportion of their total plasma cholesterol in the HDL rather than in the LDL fraction (40). Human studies (41,42) have shown that nutritional interventions such as high-fat feeding induce insulin resistance with an elevation of HDL cholesterol and apoA-1 production, whereas a reduction of apoA-1 production occurs when the diet is switched from high to low fat. The anticipated lowering of HDL cholesterol in insulin-resistant states may not occur when insulin resistance is induced by short-term high-fructose or high-fat feeding. The exact mechanism of the elevation of HDL cholesterol in the fructose-fed hamster is currently not known and will require further investigation.

Our results are in accordance with numerous studies (8,12–19), which have shown that an elevation in postheparin plasma HL activity is often associated with insulin-resistant states such as central obesity and type 2 diabetes. These findings have led to the hypothesis that the relation between HL and adiposity reflects the modulating effect of secondary factors associated with increased adiposity and insulin resistance, although the precise factors that regulate HL activity are not known. HL activity is not upregulated in a clear-cut fashion by insulin. Although studies (12–14,16,19) in patients with insulin resistance and type 2 diabetes, conditions that are commonly associated with chronic hyperinsulinemia, have shown higher HL activity than in appropriate control subjects, euglycemic-hyperinsulinemic clamp studies in humans have shown (43) that acute hyperinsulinemia actually reduces plasma HL activity. It is likely that instead of hyperinsulinemia per se it is some other aspect of insulin resistance that induces the increase in HL activity.

Although the polymorphism C-514T in the promoter region of the HL gene, which explains up to 38% of the variability in HL activity (44), has recently been (45) associated with fasting hyperinsulinemia and insulin resistance, the functional significance of this site is not known. HL is a secreted glycoprotein, and although we detected a concomitant increase in HL mRNA with induction of insulin resistance, it is also possible that the increase in plasma HL enzymatic activity could have occurred due in part to posttranslational modification of HL. HL is synthesized in the endoplasmic reticulum, first as a high mannose form (52–55 kDa in the rat) and subsequently acquires sialic acid–containing complex oligosaccharides during transit through the Golgi cisternae (46). The mature HL (57–59 kDa, in the rat) is then rapidly secreted and has a residence half-time of ∼60 min in the hepatocyte. Trimming of the glucose residues off the high-mannose form appears to be a determining step for rat HL secretion (47,48). In humans, besides remodeling of the N-linked oligosaccharides, association of HL with calnexin, a chaperone protein, may increase the efficiency of HL export from the endoplasmic reticulum (49). Further studies are required to determine the precise molecular regulation of HL that occurs in the setting of insulin resistance.

We demonstrated that HL but not LPL enzymatic activity is easily measurable in preheparin plasma in the Syrian golden hamster, as has been previously reported (30), indicating that a significant proportion of the secreted enzyme is not bound to the endothelium by heparan sulfate proteoglycans but circulated freely in the circulation. Activity levels reported in this work, however, are postheparin values because the administration of heparin allowed us to simultaneously measure changes in LPL in response to the various experimental interventions, in addition to the combination of unbound and hepatic-bound HL activity.

In conclusion, we have shown that plasma HL enzymatic activity and hepatocyte HL mass and mRNA are increased in the fructose-fed, nondiabetic, insulin-resistant Syrian golden hamsters and that the increase is partially normalized by the PPARγ agonist insulin sensitizer rosiglitazone. To the best of our knowledge, this is the first study to examine the molecular regulation of HL in vivo with induction of insulin resistance and subsequent insulin sensitization. Elevated HL in insulin resistance may play an important pathophysiological role in the atherogenic dyslipidemia of insulin-resistant states, particularly in the formation of potentially atherogenic, small, dense LDL particles. Further studies are required to confirm these findings in insulin-resistant humans and to determine the molecular regulation of HL in insulin-resistant states.

This work was supported by an operating grant from the Canadian Institutes of Health Research (to G.F.L., MOP43839, and to K.A., MOP53093) and National Institutes of Health Grant (to J.D.B., HL64322). This work was supported in part by a research grant from GlaxoSmithKline Canada (to G.F.L. and K.A.).

G.F.L. holds a Canada Research Chair in Diabetes and is a Career Investigator of the Heart and Stroke Foundation of Canada.