| |
HAART & Heart Disease: Antiretroviral therapy in HIV-positive men is associated with increased apolipoprotein CIII in triglyceride-rich lipoproteins
|
| |
| |
HIV Medicine
September 2005
D Rimland1,2, JL Guest1, I Hernández2,3, C del Rio2,3, NA Le2,4 and WV Brown1,2,4
1Atlanta VA Medical Center, 2Emory University School of Medicine, 3Emory University Center for AIDS Research, and 4Emory Lipid Research Laboratory, Atlanta, GA, USA
".....we found that either PI or NNRTI therapy was associated with elevated levels of apoCIII, which were disproportionate to the elevated levels of TGs. These findings suggest impaired metabolism of TG-rich lipoproteins. On a background of an older population with the additional risk factors of smoking and diabetes, these findings portend future atherosclerotic events in these patients.....
.....Patients treated with either PIs or NNRTIs were more likely to have higher apoCIII levels (P<0.05 for each) than patients on no therapy and controls. apoB and apoE levels were higher in patients on PIs or NNRTIs compared with controls. ApoA-I levels were higher in patients on NNRTIs and patients on no therapy compared with controls. LDL particle size was statistically lower in the PI and NNRTI groups compared with the controls..... There was a strong positive correlation between plasma apoCIII and plasma TG concentrations in the non-HIV-infected controls, the HIV-positive patients treated with either PI or NNRTI, and the treatment-naive patients. While this was expected, for comparable TG levels, HIV-positive patients had significantly higher apoCIII levels.....
...... Multivariate analysis was used to assess the individual contributions of specific treatment class regimens and univariately significant demographic and clinical groups using the individual lipid measurements as the outcome variable (Table 3). PI but not NNRTI therapy was independently associated with elevated total cholesterol concentration (P=0.044). Treatment with PI or NNRTI was independently associated with elevated TG concentration (P=0.020). NNRTI but not PI therapy was associated with low HDL and high LDL level (P=0.009 and 0.064, respectively). Neither PI nor NNRTI therapy was independently associated with elevations of apoB; interestingly, regimens containing stavudine were independently associated with high apoB levels . PI therapy was associated with significant elevations in plasma apoCIII. Elevations in apoCIII were also observed with NNRTI therapy, although statistical significance was not reached (P=0.053). Neither PI nor NNRTI therapy was associated with changes in apoE or LDL particle size. Importantly, black race was independently associated with lower TG, HDL and apoCIII concentrations and smaller LDL particle size......"
BACKGROUND
Dyslipidaemia was recognized as a common problem in HIV disease even before the advent of potent antiretroviral therapy. HIV infection alone has been associated with decreased concentrations of high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol and increased concentrations of triglycerides (TGs) [1]. Antiretroviral therapy, especially protease inhibitor (PI) therapy, has been associated with multiple metabolic abnormalities including fat redistribution, insulin resistance and dyslipidaemia [2-4]. Abnormalities have included elevated TG [1,5,6], total cholesterol [1,3,5] and LDL [1,7-9] concentrations. The pathogenic mechanisms for these changes are not well understood but several mechanisms have been postulated [10-13].
Lipoprotein subclasses, as defined by particle concentration and size, have different atherogenic risk. Further elucidating the changes produced by antiretroviral therapy will help to define the mechanisms underlying the lipoprotein changes and the potential for increased atherosclerotic risk of these drugs. Apolipoprotein CIII (apoCIII) is of particular interest because increasing the synthesis of this protein in transgenic mice markedly increases plasma TG levels [14] and the absence of functional apoCIII produces a dramatic reduction in TGs associated with enhanced clearance in both mice [15] and humans [16]. Fibrate drugs reduce apoCIII synthesis and this appears to be a major factor in the TG-lowering effect of these agents [17]. Higher concentrations of apoCIII in apoB-containing lipoproteins [very low-density lipoproteins (VLDLs), intermediate density lipoproteins (IDLs) and LDLs] have been predictive of increased incidence of coronary events in clinical trials [18] and more rapid progression of coronary lesions measured by angiography [19,20]. Few data are available on antiretroviral-associated changes in newer markers of lipid metabolism, such as specific apolipoproteins (apoB, apoE and apoCIII) and lipoprotein particle size [9]. These data may help in predicting the potential long-term impact of various forms of drug therapy in HIV disease.
ABSTRACT
Objectives: Dyslipidaemia has become a common problem in HIV disease, especially in patients on current antiretroviral therapy. However, the pathogenic mechanisms involved are not well understood or documented using conventional lipid measurements.
Methods: Using a cross-sectional design, the prevalence of abnormal standard lipid measurements and novel biomarkers for abnormal lipid metabolism was determined in 271 HIV-positive men from two HIV clinics in Atlanta, GA, USA.
Results: A total of 147 men were treated with protease inhibitors (PIs) for >3 months (54%), 84 were treated with nonnucleoside reverse transcriptase inhibitors (NNRTIs) for >3 months (31%) and 40 had not received antiretroviral therapy in the past 3 months (15%). Patients being treated with a PI had higher total cholesterol and triglyceride (TG) levels than patients on no therapy (P<0.05 for each). Patients in the NNRTI group had higher TG, lower high-density lipoprotein (HDL) levels, and higher low-density lipoprotein (LDL) levels than patients on no therapy (P<0.05 for each). Patients treated with either PIs or NNRTIs were more likely to have higher apolipoprotein CIII (apoCIII) levels (P<0.05 for each) than patients on no therapy. Elevated TG was associated with disproportionably elevated apoCIII levels in both treatment groups.
Two hundred and seventy-one men were enrolled in this cross-sectional study. One hundred and forty-seven were treated with PIs for >3 months (54%), 84 were treated with NNRTIs for >3 months (31%) and 40 had not received antiretroviral therapy in the past 3 months (15%). Most patients were African-American (56%), were >40 years of age (66%), had CD4 counts >200 cells/muL (76%) and had viral loads <10 000 HIV-1 RNA copies/mL (72%). Forty-one per cent of the patients smoked and 10% had diabetes. Twelve per cent were on lipid-lowering agents, with all but one of these patients being treated with a PI or a NNRTI. The mean body mass index (BMI) was 25.0, with 41.0% of the participants having a BMI greater than 25.
The most commonly prescribed NNRTI in these patients was efavirenz (82%). There was more variation in the prescribed PIs, but the two most common were nelfinavir (34%) and lopinavir/ritonavir (25%); 54% of the patients were on a boosted PI-containing regimen. The majority of the patients had lamivudine or stavudine as one of the NRTIs in their regimen (75% and 53%, respectively). Stavudine was the only consistently significant specific drug associated with lipid abnormalities and was retained in the multivariate analysis for all lipid measurements.
Conclusions: In this cross-sectional study of HIV-infected men, either PI or NNRTI therapy elevated levels of TG and apoCIII. Higher concentrations of apoCIII in apoB-containing lipoproteins [very low-density lipoproteins (VLDLs), intermediate density lipoprotein (IDL) and LDLs] have been predictive of an increased incidence of coronary events in clinical trials and more rapid progression of coronary lesions measured by angiography. These findings, on a background of an older population with additional risk factors of smoking and diabetes, portend future atherosclerotic events in these patients.
Patient population
HIV-infected men were selected from two out-patient HIV clinics: the Atlanta VA Medical Center and the Infectious Disease Program of the Grady Health System. The Emory Institutional Review Board approved the protocol. After obtaining informed consent, patients were selected if they fitted one of the following three classifications of antiretroviral therapy: (1) no therapy for at least 3 months; (2) treatment with a PI for at least 3 months, or (3) treatment with a nonnucleoside reverse transcriptase inhibitor (NNRTI) for at least 3 months. Patients on a PI- or NNRTI-based regimen were also taking combination nucleoside reverse transcriptase inhibitors (NRTIs). Information was recorded on demographics, family history, smoking habits, antiretroviral treatment, use of lipid-lowering agents and pertinent laboratory results, including recent CD4 count and HIV viral load. The control values for all measurements were derived from a database of 145 healthy HIV-negative men with no vascular diseases and <=1 risk factor for cardiovascular disease (except age) used for validation of instruments measuring a number of cardiovascular risk factors; equal numbers of men were included from each decade from 20 to 80 years.
RESULTS
There were statistically significant differences in lipid levels in the three groups when compared with the HIV-negative controls by analysis of mean values (Table 2). Patients being treated with a PI had higher total cholesterol than patients on no therapy and controls; higher TGs than patients on NNRTIs, patients on no therapy and controls; and lower HDL than patients on NNRTI and controls (P<0.05 for all comparisons). Patients in the NNRTI group had higher total cholesterol than patients on no therapy and controls; higher TGs than controls; higher HDL than patients on no therapy but lower HDL than controls; and higher LDL than patients on no therapy. Patients treated with either PIs or NNRTIs were more likely to have higher apoCIII levels (P<0.05 for each) than patients on no therapy and controls. apoB and apoE levels were higher in patients on PIs or NNRTIs compared with controls. ApoA-I levels were higher in patients on NNRTIs and patients on no therapy compared with controls. LDL particle size was statistically lower in the PI and NNRTI groups compared with the controls.
There was a strong positive correlation between plasma apoCIII and plasma TG concentrations in the non-HIV-infected controls, the HIV-positive patients treated with either PI or NNRTI, and the treatment-naive patients (Fig. 1). While this was expected, for comparable TG levels, HIV-positive patients had significantly higher apoCIII levels (Fig. 2; partial F-test; P<0.0001).
Multivariate analysis was used to assess the individual contributions of specific treatment class regimens and univariately significant demographic and clinical groups using the individual lipid measurements as the outcome variable (Table 3). PI but not NNRTI therapy was independently associated with elevated total cholesterol concentration (P=0.044). Treatment with PI or NNRTI was independently associated with elevated TG concentration (P=0.020). NNRTI but not PI therapy was associated with low HDL and high LDL level (P=0.009 and 0.064, respectively). Neither PI nor NNRTI therapy was independently associated with elevations of apoB; interestingly, regimens containing stavudine were independently associated with high apoB levels. PI therapy was associated with significant elevations in plasma apoCIII. Elevations in apoCIII were also observed with NNRTI therapy, although statistical significance was not reached (P=0.053). Neither PI nor NNRTI therapy was associated with changes in apoE or LDL particle size. Importantly, black race was independently associated with lower TG, HDL and apoCIII concentrations and smaller LDL particle size.
DISCUSSION
Dyslipidaemia was recognized as a common problem in HIV disease before the advent of antiretroviral therapy. Chronic infections, including HIV infection, have been associated with decreasing HDL and LDL cholesterol concentrations and increased TG concentration [1]. Antiretroviral therapy, especially PI therapy, has been found to be associated with multiple metabolic abnormalities including fat redistribution, insulin resistance and dyslipidaemias [2-4]. Several authors have described protease-associated lipid abnormalities in both cross-sectional and longitudinal studies. The prevalence and degree of lipid abnormalities are probably related to the class of drugs, the specific agent and the duration of therapy studied. Changes in lipid levels have variously been compared with baseline levels, with levels in HIV-infected patients not on therapy, and with levels in a control group of uninfected patients. In addition, other factors including nutritional status, HIV disease stage and concomitant presence of lipodystrophy and insulin resistance are certainly important. Lipid abnormalities have been found to include elevated concentrations of TGs [1,5,6], total cholesterol [1,3-5] and LDL [1,6]. More recent studies have described elevations in concentrations of lipoprotein a [8,12], apoA [21], apoB [9] and apoCIII [22] and changes in lipoprotein particle size [9]. Lipoprotein subclasses, as defined by particle concentration and size, have different atherogenic risk. Although little is known about these changes in HIV-infected patients on antiretroviral therapy, a recent report on 24 patients showed that 1 month of lopinavir/ritonavir use enhanced TG and apoCIII levels and reduced LDL size, suggesting an increase in atherogenicity [21]. In a small study of five patients on antiretrovirals, the ratio of large, buoyant LDL1 over small, dense LDL2 was significantly reduced from 2.0 to 0.8 [10].
Very few reports have suggested that NNRTIs could also produce elevated lipid levels. In the large Data Collection and Adverse Events of Anti-HIV Drugs (DAD) study, Friis-Moller et al. recently reported that current NNRTI therapy was associated with elevated concentrations of total cholesterol and TGs [23]. In contrast, in a cross-sectional analysis of patients on their first NRTI and NNRTI regimen, there was no association between elevated concentrations of TGs and cholesterol and any specific NRTI or NNRTI [24]. In fact, nevirapine was recently shown to elevate HDL, which may be associated with a decreased incidence of coronary artery disease [25]. In our study, NNRTI therapy was clearly associated with elevated concentrations of TGs. Importantly, most of the patients on an NNRTI were taking efavirenz.
NRTIs alone have not been associated with lipoprotein abnormalities, although stavudine-based regimens, but not tenofovir-based regimens, are associated with early and significant increases in TG and total cholesterol levels [26]. We found no lipid abnormalities independently associated with specific nucleosides, except that stavudine was associated with higher apoB levels.
The mechanisms producing these lipid abnormalities are not well understood, but several mechanisms have been postulated [10-13,27]. Several PIs have been found to increase hepatic TG synthesis in vitro and in a mouse model [28]. Inadvertent effects of PIs may be important. The sterol regulatory element-binding proteins (SREBPs) are membrane-bound transcription factors that play a central role in cellular lipid homeostasis. It has been suggested that PIs inhibit proteases that degrade SREBP [29]. This inhibition of proteases could decrease degradation of apoB100, which might then increase the number of VLDL particles [27]. A careful kinetic analysis of five patients with antiretroviral-induced dyslipidaemia showed that total apoB synthesis was significantly increased and shifted toward TG-rich VLDL1. These data suggest that increased concentrations of TGs are primarily attributable to reduced rates of VLDL conversion to denser lipoproteins, implying a lower rate of lipoprotein lipase-mediated delipidation [10]. Genetic predisposition is probably important, but may account for only a proportion of the lipid abnormalities. HIV-infected patients homozygous for the E2 genotype of apoE2 develop marked hyperlipidaemia when treated with PIs [30,31]. In addition, HIV-infected patients with a specific polymorphism in the apoCIII gene had lower levels of HDL cholesterol and increased levels of TGs [10,32]. The mechanism of NRTI-induced abnormalities may include mitochondrial toxicity, adipocyte damage and lipoatrophy [31].
In our study, apoCIII was not only elevated in total plasma but present in excess of the levels attributable to TGs (Fig. 1). High levels of apoCIII are thought to be associated with elevated TG concentrations because of the impairment of lipolysis, as the activity of lipoprotein lipase is inhibited by apoCIII [33]. A positive correlation between plasma apoCIII and plasma TG has been used to suggest the importance of apoCIII in the inhibition of the activity of lipoprotein lipase [34]. apoCIII also reduces VLDL and chylomicron remnant uptake by the liver [35,36]. Homology between PI and the LDL receptor-like protein (LRP) responsible for the removal of chylomicron remnants may prolong the residence time of these particles in the circulation. These remnants are considered to be atherogenic. Excess apoCIII has also been reported in coronary artery disease (CAD) patients who demonstrated disease progression in spite of efficacious cholesterol-reduction therapy, especially when apoCIII was preferentially retained in the TG-rich lipoprotein fraction [18]. Increased synthesis of apoCIII under the influence of PIs or NNRTIs could explain our findings. The abnormally high apoCIII/TG ratio seems consistent with this hypothesis. The promoter region of the apoCIII gene includes several regulatory elements, including an insulin response element [37], a nuclear factor kB (NFkB) response element [38] and a peroxisome proliferator-activated receptor (PPAR) gamma response element [39]. In addition, the homology of PIs with cis-retinoic acid binding protein 1 (CRABP1) may interfere with the down-regulation of apoCIII expression via the PPAR nuclear factor. Whether the presence of infection with HIV and/or antiretroviral therapy can affect the expression of apoCIII is not clear. Metabolic tracer studies are underway to directly address this hypothesis.
A critical step in the pathogenesis of atherosclerosis is the deposition and retention of apoB- and apoE-containing lipoproteins by intimal macrophages. In persons with normal TGs, most apoCIII is associated with HDL; in persons with elevated TGs there is redistribution of apoCIII to VLDL and a concomitant increase in plasma apoCIII level. Using niacin and colestipol therapy, Blankenhorn et al. [19], in the Cholesterol Lowering Atherosclerosis Study (CLAS), found that individuals with the greatest regression of coronary artery lesions by angiography were those with higher apoCIII in HDL and lower apoCIII in other lipoproteins, primarily VLDL, while on treatment. In a separate study, the same group confirmed the finding using lovastatin [20]. In this case, they reported that high apoCIII levels in plasma VLDL plus LDL were associated with increased progression of angiographically demonstrated lesions. Clinical events observed in the Cholesterol and Recurrent Events Trial (CARE) were strongly associated with increased apoCIII in the apoB-containing lipoproteins [18].
One limitation of this study is the cross-sectional design defining the current class of antiretrovirals as the independent variable. It is likely that previous cumulative history of treatment is important in producing specific lipid abnormalities. In addition, the relatively small numbers of patients on each specific drug combination made it impossible to associate specific drugs with specific abnormalities.
The description of these multiple lipid abnormalities has raised concern about the atherogenic potential of these changes and the subsequent development of serious cardiovascular and cerebrovascular morbidity and mortality. Reports of myocardial infarctions in patients on antiretroviral therapy including PIs heighten this concern [40-42]. Recent studies have also described an increased incidence of surrogates of atherosclerotic disease, including intimal thickening and calcification of the carotid arteries [43,44] and coronary artery calcification [45]. Analyses of large patient databases have led to conflicting data about the risk of thrombotic events after the advent of potent antiretroviral therapy [46-49]. Continuing analyses of patient cohorts including longitudinal studies will be needed to resolve these conflicting findings.
In summary, we found that either PI or NNRTI therapy was associated with elevated levels of apoCIII, which were disproportionate to the elevated levels of TGs. These findings suggest impaired metabolism of TG-rich lipoproteins. On a background of an older population with the additional risk factors of smoking and diabetes, these findings portend future atherosclerotic events in these patients.
|
|
| |
| |
|
|
|
