Effect of ritonavir on lipids and post-heparin lipase activities in normal subjects

Background: Intensive therapy of HIV infection with highly active antiretroviral therapy (HAART) dramatically reduces viral loads and improves immune status. Abnormalities of lipid levels, body fat distribution, and insulin resistance have been commonly reported after starting HAART. Whether the lipid abnormalities result from changes in metabolism after an improvement in HIV status or are partly attributable to the effects of protease inhibitor use is unknown.

Methods: Twenty-one healthy volunteers participated in a 2 week double-blind, placebo-controlled study on the effect of the protease inhibitor ritonavir on total lipids, apolipoproteins, and post-heparin plasma lipase activities.

Results: Those taking ritonavir (n = 11) had significantly higher levels of plasma triglyceride, VLDL cholesterol, IDL cholesterol, apolipoprotein B, and lipoprotein (a) compared with placebo (n = 8). HDL cholesterol was lower with therapy as a result of a reduction in HDL₃ cholesterol. Post-heparin lipoprotein lipase (LpL) activity did not change but hepatic lipase activity decreased 20% (P < 0.01) in those taking ritonavirrcompared with placebo. Although all lipoprotein subfractions became triglyceride enriched, most of the increase in triglyceride was in VLDL and not in IDL particles.

Conclusion: Treatment with ritonavir in the absence of HIV infection or changes in body composition results in hypertriglyceridemia that is apparently not mediated by impaired LpL activity or the defective removal of remnant lipoproteins, but could be caused by enhanced formation of VLDL. Long-term studies of patients with HIV infection receiving HAART will be necessary to determine the impact of these drugs and associated dyslipidemia on the risk of coronary artery disease.

Introduction

Before the introduction of highly active antiretroviral therapy (HAART), which includes drugs from the protease inhibitor class, patients with AIDS had been shown to have higher triglyceride levels but lower levels of total, LDL, and HDL cholesterol compared with non-HIV-infected controls ^[1-3] With the use of HAART, HIV viral loads have been dramatically reduced, CD4 cell counts have increased, and opportunistic infections have resolved in those with AIDS ^[4] Rather than an improvement in lipid levels with HAART, however, a cluster of metabolic abnormalities has been described that includes hypertriglyceridemia, abnormal central body fat distribution, and insulin resistance ^[5-9] Because of the association between central obesity, insulin resistance, and dyslipidemia (including increased triglyceride levels, increased apolipoprotein B, and decreased HDL cholesterol) ^[10,11] it is possible that the hyperlipidemia associated with HAART is influenced by changes in body composition or insulin sensitivity. Protease inhibitors may, however, have effects on lipids and lipoproteins independent of other metabolic changes resulting from treatment of HIV infection or AIDS. To answer this question, healthy individuals not known to be infected with the HIV virus were recruited to assess the effects of the protease inhibitor ritonavir on lipoprotein metabolism.

Materials and methods

Twenty-one healthy, normal men and women were recruited from the Seattle area for this study. Individuals were excluded if they had a triglyceride level greater than 400 mg/dl. All medications were discontinued at least one week before entering the trial. Informed consent was obtained before entering into the study and all procedures were approved by the Human Subjects Committee of the University of Washington.

Study design

Subjects were randomly selected in a double-blind manner to receive either capsules containing ritonavir 300, 400, and 500 mg doses every 12 h on study days 1, 2, and 3-14, respectively (n = 13), or a matching placebo (n = 8) (Fig. 1). Two subjects randomly assigned to the ritonavir treatment dropped out during the study because of side-effects from the medication: one subject withdrew on study day 3 after experiencing severe chills, nausea, sweating and dizziness, which resolved within 24 h after leaving the study; the other subject experienced flushing and developed a moderate rash over the neck and face on study day 1. At baseline and on the morning of the 14th day of therapy, subjects underwent the following procedures at the Northwest Lipid Research Clinic in Seattle, WA, USA.

Fig. 1

Lipids and lipoproteins

Blood was collected in 0.1% ethylenediaminetetraacetic acid after a 12-16 h overnight fast for lipoprotein measurements and post-heparin lipase activity. A heparin bolus of 60 units/kg was given intravenously and blood was collected after 10 min in lithium-heparin tubes for the measurement of post-heparin lipase activity. Blood was immediately centrifuged at 4°C at 1500 g for 15 min. Lipid measurements were made on plasma stored at 4°C within 2 days of collection. Lipase activities were obtained on plasma that had been immediately frozen and stored at -70°C. Lipoproteins were separated by ultracentrifugation ^[12] and HDL precipitation ^[13] and cholesterol and triglyceride levels were analysed enzymatically ^[14] Apolipoprotein B and apolipoprotein A-I were measured by radioimmunoassay ^[15] Liproprotein (a) mass was determined using a double monoclonal antibody-based enzyme-linked immunosorbent assay ^[16]

Density gradient ultracentrifugation

The cholesterol content of lipoprotein subfractions was determined using non-equilibrium density gradient ultracentrifugation (DGUC) in a Sorvall TV-865B vertical rotor (DuPont, Wilmington, DE, USA) ^[17] In this protocol, 1 ml of plasma was adjusted to a density of 1.08 g/ml (total volume 5 ml) and layered below 13 ml of a 1.006 g/ml sodium chloride solution. Samples were then centrifuged at 399 727 g for 90 min at 10°C. Maintaining the temperature, centrifuge tubes were then placed in a tube fractionator (ISCO, Lincoln, NE, USA), pierced, and drained from the bottom using a P-1 peristaltic pump (Pharmacia, Piscataway, NJ, USA) at a flow rate of 1.0 ml/min. A total of 38 0.47 ml fractions were collected. Cholesterol was measured in each fraction by an enzymatic kit (Diagnostic Chemicals, Canada). The between-rotor coefficient of variation (CV) for LDL buoyancy (determined by dividing the fraction number containing the peak level of cholesterol within the LDL range by the total number of fractions, or 38) was 3.5%.

Post-heparin lipase activities

The total lipolytic activity was measured in plasma after a heparin bolus of 60 units/kg ^[18] Tri-1-¹⁴C oleate (Amersham, Arlington Heights, IL, USA) and triolein emulsified with lecithin were incubated with postheparin plasma for 60 min at 37°C and the liberated ¹⁴C labelled free fatty acids were then extracted and counted. Lipoprotein lipase (LpL) activity was calculated as the lipolytic activity removed from the plasma by incubation with the specific 5D2 monoclonal antibody against LpL, and hepatic lipase activity was determined as the activity remaining after incubation with the LpL antibody. Enzyme activity is expressed as nanomoles of free fatty acid released per minute per milliliter of plasma at 37°C. For each assay, a bovine milk LpL standard was used to correct for inter-assay variation and a human postheparin plasma standard was included to monitor inter-assay variation. The intra-assay CV of hepatic lipase is 6%, inter-assay CV is 14%.

Statistical methods

For comparisons between ritonavir and placebo groups at baseline, the two-sample t-test was used. For comparisons of ritonavir and placebo at the end of the dosing regimen (day 14), analysis of covariance was used with the variable from the day before the first dose as the baseline covariate. Within-group changes from baseline were tested using the paired t-test if the data were normally distributed, and the signed-rank test was used if they were not. Significance was ass-igned at P < 0.05. Results are expressed as mean ± SE.

To test the significance of differences in cholesterol distributions in the profiles generated by DGUC between baseline and follow-up visits, a difference plot was generated by subtracting the mean cholesterol value of each fraction measured at baseline from the mean cholesterol value in the same fraction at follow-up with determination of the 95% confidence interval (CI) for this difference. A difference in fractional cholesterol content between the time points is significant (P < 0.05) when the 95% CI does not cross the zero line.

Results

The age of the ritonavir group (30 ± 1 years) was not significantly different from that of the placebo group (31 ± 2 years) The profile of the study is shown in Fig. 1. The groups did not differ significantly with respect to weight, and neither group experienced a significant change in weight during the study period (Table 1). No changes occurred in the levels of aspartate aminotransferase (AST) or alanine aminotranferase (ALT) in the ritonavir treatment group (AST: 16 ± 1.8 versus 19 ± 1.7 units/l, P = 0.76; ALT: 16 ± 4.2 versus 20 ± 2.3 U/l, P = 0.48, baseline and day 14 of treatment, respectively).

Table 1

At baseline, there were no significant differences between treatment groups in the levels of lipids, apolipoproteins, or post-heparin lipase activities (Tables 1-3). After 14 days of treatment, those receiving ritonavir experienced increased levels of total cholesterol, triglyceride, lipoprotein (a), and apolipoprotein B compared with baseline and the placebo group (Table 1). This increase in total cholesterol in the ritonavir group resulted from increases in VLDL cholesterol and IDL cholesterol but not LDL cholesterol (Table 2). HDL cholesterol decreased on ritonavir as a result of a decrease in HDL₃ cholesterol. The triglyceride content of all lipid subfractions was increased in the group taking ritonavir compared with baseline and the placebo group (Table 2). Using DGUC to measure the cholesterol content of lipoprotein subfractions, no difference was seen in the placebo group from baseline to day 14 of dosing (Fig. 2). DGUC analysis from the ritonavir group demonstrated an increase in cholesterol in VLDL fractions, IDL fractions, and the most buoyant and most dense LDL fractions (Fig. 3), whereas cholesterol was reduced to the greatest extent in the dense HDL fractions. The buoyancy of the peak LDL particle (Rf) decreased slightly (became more dense) in the ritonavir group (Rf: 0.261 ± 0.008 versus 0.249 ± 0.010, baseline to follow-up, respectively, P = 0.126), but did not change in the placebo group (Rf: 0.274 ± 0.006 versus 0.270 ± 0.004, baseline to follow-up, respectively, P = 0.60).

Fig. 2	Fig. 3	Table 2
Table 3

Post-heparin plasma hepatic lipase activity was decreased in the ritonavir group at day 14 of dosing compared with baseline and compared with the placebo group (Table 3). LpL activity did not change with study drug dosing (Table 3).

Discussion

In this group of healthy, normal subjects, treatment with ritonavir increased triglyceride levels in all lipoprotein fractions, with the greatest increase occurring in VLDL particles. Total cholesterol was increased with therapy as a result of increases in VLDL and IDL cholesterol. These effects on lipids occurred within 2 weeks of therapy while the subjects were weight stable. It is therefore unlikely that these changes in lipid levels resulted from the changes in body composition (i.e. central obesity) that have been reported in patients with HIV receiving HAART. Nor could this hyperlipidemia be the result of an interaction between ritonavir and inflammatory components of the HIV-infected state because these individuals were not known to be infected with the HIV virus. Therefore ritonavir itself is partly a cause of the hypertriglyceridemia and increases in cholesterol that have been described in HIV-positive patients taking this medication.

Increases in triglyceride, VLDL cholesterol, IDL cholesterol, and apolipoprotein B levels could result from increased secretion of VLDL particles, decreased lipolysis by LpL, delayed hepatic clearance of remnant lipoproteins, or some combination of the above. In this study, post-heparin LpL activity was not affected by treatment. In separate studies in our laboratory, ritonavir was also shown not to impair the lipolytic activity of LpL in vitro (unpublished data). Therefore an abnormality in LpL activity can be excluded as a cause of the triglyceride elevation in these individuals. Impaired remnant removal is also not the primary cause of the hypertriglyceridemia because subjects taking ritonavir experienced an increase mainly in VLDL particles, whereas an isolated increase in IDL particles was not seen (Fig. 3). Therefore, the most likely cause of the hypertriglyceridemia with ritonavir therapy was an increased secretion of VLDL particles. Because a small increase in IDL cholesterol was demonstrated in this study, however, some impairment of remnant clearance cannot be ruled out. Studies that measure VLDL secretion and clearance rates are needed to determine the specific abnormality in the processing of triglyceride-rich particles that occurs in patients taking ritonavir.

Hepatic lipase hydrolyses triglyceride and phospholipid in IDL, LDL and HDL particles. Alterations in the activity of this enzyme are also associated with changes in the cholesterol content of LDL and HDL particles ^[17,19,20] Hepatic lipase activity is higher in individuals with central obesity ^[21] and is associated with smaller, more dense LDL particles ^[20] which have been demonstrated to be an independent risk factor for the progression of heart disease ^[22,23] A low hepatic lipase activity, on the other hand, has been shown to be associated with buoyant LDL particles ^[20] An increase in LDL particle buoyancy is associated with the regression of coronary artery disease after aggressive lipid-lowering therapy ^[24]

In the present study, treatment with ritonavir resulted in a decrease in hepatic lipase activity, similar to the effect of estrogen treatment of post-menopausal women ^[25,26] As mentioned above, low hepatic lipase activity is usually associated with less cholesterol in small, dense LDL and more in HDL, primarily in HDL₂. Neither of these lipoprotein changes occurred, in spite of the decrease in hepatic lipase activity that occurred with ritonavir therapy. This lack of expected change in the particle compositions of LDL and HDL₂ is probably due to the off-setting effect of the induced hypertriglyceridemia, which is associated with the formation of small, dense LDL and a decline in HDL₂ cholesterol ^[27,28] In this respect, the effect of ritonavir differs from the effect of estrogen which, in addition to its association with a reduction in post-heparin hepatic lipase activity and an increase in VLDL entry into the circulation, results in a decrease in buoyant LDL, no change in dense LDL, and an increase in HDL₂ cholesterol ^[29,30] An additional mechanism affecting LDL and HDL is cholesteryl ester transfer protein (CETP). CETP transfers triglyceride from apolipoprotein B-containing particles to HDL in exchange for cholesteryl ester, and is thought to play an important role in the formation of dense LDL in hypertriglyceridemia disorders ^[27] An increase in CETP activity could also partly explain the increase in cholesterol in dense LDL fractions and the reduction in HDL cholesterol, although the activity of this enzyme was not measured as a part of this study. The effect of ritonavir on lipoprotein metabolism thus appears to be a hybrid of effects, including those associated with a reduction in post-heparin hepatic lipase activity and an increased production of lipoproteins. This overproduction of lipoproteins results in a lipid profile that is similar to that described in familial combined hyperlipidemia ^[31] including increased levels of triglyceride, VLDL cholesterol, and apolipoprotein B; and a reduction in LDL buoyancy.

The mechanism whereby ritonavir treatment might enhance the secretion of apolipoprotein B-containing particles or reduce hepatic lipase activity is unclear. Ritonavir and other protease inhibitor drugs have been shown to bind to and inhibit the catalytic site of HIV-1 protease. It is also known that protease inhibitors inhibit the cytochrome P450 system ^[4] and have been reported to have limited homology with the LDL receptor-related protein and the cytoplasmic retinoic acid-binding protein type 1 ^[32] None of these properties, however, satisfactorily explains the action of ritonavir on fatty acid synthesis, the hepatic secretion of apolipoprotein B-containing particles, or the alteration of hepatic lipase activity that would result in the lipid changes found in this study.

Lipoprotein (a) levels also increased in the subjects receiving ritonavir but not in those taking placebo. This increase may have been partly responsible for the increase in cholesterol in dense LDL fractions (Fig. 3). In healthy subjects, levels of lipoprotein (a) are determined primarily by genetic factors ^[33] Lipoprotein (a) levels have been shown to be increased in a number of diseases including nephrotic syndrome ^[34] renal failure ^[35] and hypothroidism ^[36] Estrogen deficiency in women also results in an increase in lipoprotein (a) levels that can be reversed with estrogen therapy ^[37] As the pathways responsible for the assembly, secretion, and catabolism of lipoprotein (a) particles remain unclear, how treatment with ritonavir might result in increased levels of this lipoprotein is unknown.

Conclusion

The treatment of healthy, normal individuals with ritonavir results in increases in triglyceride and cholesterol levels that are independent of changes in body composition or interactions with inflammatory conditions that accompany HIV infection. This increase in triglyceride is not caused by the impairment of lipoprotein lipase activity, nor is it solely the result of the impaired clearance of remnant particles. Changes in lipid metabolism with ritonavir treatment are associated with a worsening of the cardiovascular lipid risk profile, including increases in VLDL cholesterol, IDL cholesterol, apolipoprotein B, and lipoprotein (a). Post-heparin hepatic lipase activity decreased with ritonavir treatment, but this was not accompanied by potentially beneficial changes, including an increase in peak LDL buoyancy or HDL₂ cholesterol. Although these effects on lipids occurred independently of changes in body composition, they may be additive to the dyslipidemia and risk of coronary artery disease with the central obesity and insulin resistance described in HIV-positive patients treated with HAART regiments that include ritonavir. Metabolic studies of individuals taking this medication may offer new insights into the regulation of lipoprotein processing and hepatic lipase activity. Long-term studies will be needed, however, to determine the significance of these changes on the risk of coronary artery disease in patients with HIV infection.