Genomic Mechanisms of Metabolic Complications & Management, 13th CROI 2006
13th CROI: Conference on Retroviruses and Opportunistic Infections
Poster Discussion: Genomic Mechanisms of Metabolic Complications
AT CROI results from several studies examined the relationships between genetics of patients (genetic polymorphisms) and changes in lipids and fat distribution demonstrating differences in the risk of metabolic complications on therapy. Since we are in the early stages of this research some of the studies could be performed better and I think in the future they will be.
Developments in Improving Patient Management of Lipids, Heart Disease, and Glucose or Diabetes.
As the Swiss study concludes below: "Genetic profiling may identify patients at risk for ART-associated dyslipidemia." We are not ready for such testing. But perhaps in the future we will be able to perform a simple blood test whose results may identify patients have a genetic predisposition for metabolic abnormalities and this would help in selecting a safer and more effective HAART regimen and designing followup for the patient. Below are studies from the 13th CROI demonstrating what the future may hold. At CROI several studies presented suggested that on the whole clinicians may be reversing the trend of increasing risk for heart disease and diabetes. Studies showed the perhaps the incidence of cardiovascular disease is leveling off and in some cases decreasing due to better clinician management. The risk for heart disease can be better managed by comprehensive evaluation before HAART. If a patient appears to have a predisposition for metabolic abnormalities (lipids, glucose), this information can be utilized in selection of HAART and future patient management. The use of lipid-lowering agents appears to be increasing and helping in management of lipids and heart disease. Screening for glucose abnormalities, insulin resistance, and diabetes may be improving overall as well, but I suspect that proper screening is not widespread. Diet and diabetes have become a national problem in the overall US population, and moreso among HIV+ individuals. All patients should be screened and evaluated for potential glucose abnormalities. This information should be utilized in patient management. Patients should be adequately followed for the development of glucose abnormalities, insulin resistance, and diabetes. Diabetes is associated with the development of heart disease. Diet, exercise, not smoking, and therapeutic interventions can be utilized to treat and perhaps prevent diabetes. Studies have shown that abacavir and tenofovir appear to be more lipid friendly than other NRTIs. The protease inhibitor atazanavir appears more lipid friendly. For patients without glucose abnormalities already research suggests atazanavir may be more friendly regarding the development of glucose abnormalities. And NNRTIs appear more lipid friendly. In sum, over the past several years these improvements in patient management regarding metabolic complications fortell better outcomes for patients if clinicians and patients utilize this information. On a controversial note, two studies, SMART interruption study and HOPS Analysis of Cardiovascular Risk Factors found stopping HAART may worsen risk factors for heart disease. At CROI studies found that pioglitazone provided a significant improvemt in body fat, although it wasn't a large improvement. Recent research found pravastatin significantly improved body fat, although again not by that much. Previous research found that patients who switched from AZT or d4T to abacavir or tenofovir improved body fat. Taken together, these interventions perhaps could have additive benefits in improving body fat loss, although the potential for additive benefit from combining these interventions have not been studied.
A Single Nucleotide Polymorphism in the Resistin Gene Is Associated with Adverse Metabolic Changes on HAART: An Exploratory Pharmacogenetic Association Study of A5005s, the Metabolic Sub-study of ACTG 384
Koustubh Ranade*1, R Parker1, L Ploughman1, R Parker2, P Tebas3, W Powderly4, S Grinspoon5, K Mulligan6, M Noor1,7, and M Dube8
1Bristol-Myers Squibb, Princeton, NJ, US; 2Statistical and Data Analysis Ctr, Boston, MA, US; 3Univ of Pennsylvania, Philadelphia, US; 4Univ Coll Dublin, Ireland; 5Harvard Sch of Publ Hlth, Boston, MA, US; 6Univ of California, San Francisco, US; 7Bristol-Myers Squibb, Plainsboro, NJ, US; and 8Indiana Univ, Indianapolis, US
Background: Dyslipidemia and insulin resistance can be associated with HAART, but the incidence and severity vary widely suggesting a genetic component. We sought single nucleotide polymorphisms in select candidate genes associated with higher risk for adverse metabolic changes after initiating HAART.
Methods: We genotyped 189 HIV+ subjects (mean age 38 years, 88% male, 56% Caucasian, 28% black, and 14% Hispanic) from ACTG 5005s for whom DNA samples and biochemical profiles prior to HAART and on therapy were available with 285 single nucleotide polymorphisms in 137 candidate genes. Principal components and cluster analyses of metabolic variables at week 32 were employed to identify a sub-group of patients at risk for developing adverse metabolic changes. Single nucleotide polymorphisms association with group membership was analyzed by Fisher's exact test and differences in metabolic variables between groups was analyzed by Kruskal-Wallis test.
Results: Multivariate analysis revealed a high-risk and a low-risk subgroup of patients. At baseline, lipids and insulin resistance by HOMA-IR were comparable between the 2 groups. At 32 weeks, the low-risk subgroup (n = 141) had lower mean body mass index (25±5 kg/m2), normal mean total cholesterol (180±29 mg/dL), LDL (110±28 mg/dL), triglycerides (143±77 mg/dL), and HOMA-IR (1.6±1.2). Relative to this subgroup, the high-risk group (n = 47) had higher body mass index (29±7 kg/m2, p <0.001), higher total cholesterol (258±47 mg/dL, p <0.001), LDL (175±32 mg/dL, p <0.001), triglycerides (305±236 mg/dL, p <0.001), and HOMA-IR (4±5.6, p <0.001). Race (p = 0.2) and sex (p = 0.3) were not significantly associated with risk of adverse metabolic changes. C T single nucleotide polymorphism in the second intron of the resistin gene was associated with significantly increased risk for developing adverse metabolic changes on HAART (p = 0.001). Heterozygotes (n = 65; odds ratio=2.8; 95%CI 1.4 to 5.7) and homozygotes (n = 5; odds ratio = 19.4; 95%CI 2 to 182) were at increased risk relative to wild type (n = 117) of being in this high-risk group. This polymorphism consistently increased risk of adverse metabolic changes on HAART in all races.
Conclusions: In this exploratory pharmacogenetic study, a single nucleotide polymorphism in the resistin gene was significantly associated with risk of developing adverse metabolic changes on HAART. Confirmation in other cohorts and relationship to individual ART agents is required.
Modeling the Influence of Polymorphisms of Several Genes Involved in Lipid Metabolism on the Risk of ART-associated Dyslipidemia
Mireia Arnedo*1, P Taffe2, H Furrer3, B Hirschel4, M Battegay5, R Weber6, P Vernazza7, E Bernasconi8, A Telenti1, P Tarr9, and Swiss HIV Cohort Study
1Inst of Microbiology, Univ of Lausanne, Switzerland; 2Swiss HIV Cohort Study Data Ctr, Lausanne; 3Univ Hosp, Berne, Switzerland; 4Geneva Univ Hosp, Switzerland; 5Univ Hosp, Basel, Switzerland; 6Univ Hosp, Zurich, Switzerland; 7Kantonsspital, St Gallen, Switzerland; 8Hosp Civico, Lugano, Switzerland; and 9Ctr Univ Hosp Vaudois, Lausanne, Switzerland
Background: Dyslipidemia in HIV-1-infected individuals receiving ART reflects the complex contribution of treatment-specific effects against a background of genetic predisposition.
Methods: We evaluated the contribution of polymorphisms to dyslipidemia in APOE, APOC3, APOA5, LPL, LIPC, LIPG, ABCA1, ADRB2, SCARB1, TNF, and CETP. We used longitudinal modeling of 3300 lipid determinations in 307 ART-treated patients (testing cohort) during a median follow-up period of 4.6 years. Polymorphisms with a significant effect in the testing cohort were reevaluated in an independent validation cohort of 88 patients that contributed 757 lipid determinations during the same duration of ART exposure. Genotyping methods included Taqman allelic discrimination, restriction fragment length polymorphism analysis and direct sequencing.
Results: The effects of variant alleles APOE e2/e4, APOC3 -482C>T/ -455T>C/ 3238C>G, APOA5 64G>C, CETP 279G>A on triglyceride (TG) levels were comparable to those reported in the general population. In addition, ABCA1 2962 A>G was found associated with elevated TG levels. The overall contribution of these variants was confirmed in the validation cohort. The effect of variant allele of CETP -629C>A on plasma HDL cholesterol was comparable to that reported in the general population. In addition, APOA5 64G>C and -1131T>C variants were associated with low HDL cholesterol levels. In patients receiving ritonavir (RTV), the most unfavorable APOE/APOC3/APOA5/CETP/ABCA1 genotype resulted in median TG levels of 4.17 mmol/L, while the absence of RTV, and a favorable genetic profile resulted in a median TG level of 1.64 mmol/L. In patients receiving protease inhibitor (PI)-sparing ART, the favorable CETP/APOA5 genotype resulted in median HDL cholesterol levels of 1.40 mmol/L, while individuals with no ART and an unfavorable genetic profile presented median HDL levels of 0.90 mmol/L.
Conclusions: The additive effect of multiple genetic variants influence the net effect of RTV on triglycerides and the beneficial effect of PI-sparing regimens on HDL-cholesterol. Genetic profiling may identify patients at risk for ART-associated dyslipidemia.
The Effect of Polymorphism of the MDR-1 Gene on the Long-term Risk of Lipodystrophy and Dyslipidemia in HIV-infected Patients Starting ART
Andrea De Luca*, S Di Giambenedetto, J Schwarz, A Marzocchetti, M Colafigli, C Pinnetti, A Bacarelli, M Fantoni, and R Cauda
Catholic Univ, Rome, Italy
Background: MDR-1 gene C3434T polymorphism has been associated with different in vitro expression of the p-glycoprotein on cell membranes and different circulating levels of ART drugs. We investigated whether these polymorphisms could be associated with a different risk of developing morphologic and metabolic alterations on HAART.
Methods: Genomic DNA samples from HIV+ patients on HAART actively followed at the UCSC Rome were analyzed for MDR-1 3435 polymorphisms by polymerase chain reaction (PCR) and restriction fragment length polymorphism (RFLP) analyses. Patients' data were collected from clinical records beginning from initiation of HAART. Lipodystrophy, lipodystrophy in any trunk fat accumulation, and lipoatrophy were detected by objective examination. Time-to-event analyses for the 3, and NCEP-ATPIII definitions of high total cholesterol (TC), triglycerides (TG), LDL, non-HDL low-HDL cholesterol were performed using Kaplan-Meier method and Cox's models.
Results: We analyzed 180 patients, of whom 69% were male, 46% heterosexual, and 23% hepatitis C virus (HCV)+. At HAART initiation, median age was 38 years, CD4 157, and viral load 10,000; 82% started a protease inhibitor (PI)-based and 17% a non-nucleoside reverse transcriptase inhibitor (NNRTI)-based regimen, 45% with stavudine (d4T). During follow-up, high TC (' 240 mg/dL), high non-HDL (' 190 mg/dL), and high TG (' 500 mg/dL) were reached by 43%, 11%, and 25%, respectively. In 154 patients evaluable for the morphologic outcomes, during a median follow-up of 50 months, 50% developed lipodystrophy, 29% lipodystrophy in any trunk fat accumulation, and 36% lipoatrophy. Alleles presented were: 3435 TT (26%), CC (17%), and CT (55%). Hazard of developing lipodystrophy in any trunk fat accumulation was significantly lower in patients with the TT genotype (HR compared to other genotypes 0.30; 95%CI 0.11 to 0.85, HR compared to CC 0.15, 0.03 to 0.65) and higher in women (HR 2.62, 1.45 to 4.73). TT genotype was independently predictive of lipodystrophy in any trunk fat accumulation after adjusting for gender and treatment type (HR 0.32, 0.11 to 0.89). Predictors of high TC were the CT genotype (HR 1.58, 1.02 to 2.45) and HCV+ status (HR 0.54, 0.27 to 1.06). After adjusting for HCV and treatment type, CT genotype remained predictive of high TC (HR 2.23, 1.30 to 3.82). The TT genotype predicted high non-HDL cholesterol levels (HR 3.41, 1.48 to 7.89).
Conclusions: The TT homozygous genotype at position 3435 of MDR-1, which is related with higher expression of p-glycoprotein, was associated with a reduced risk of developing trunk fat accumulation, but a higher risk of high non-HDL cholesterol levels, while the heterozygous genotype was associated to high TC. Knowledge on the effect of genetic determinants on long-term HAART side effects might be useful for individualized treatment decisions.
Relationship of G516T Polymorphism to Lymphocyte Expression of CYP2B6
Andrew Owen*, C Cook, N Liptrott, S Khoo, and D Back
Univ of Liverpool, UK
Background: The G516T and C1459T single nucleotide polymorphisms (SNP) in cytochrome P450 (CYP) 2B6 affect hepatic expression of this key enzyme in non-nucleoside reverst transcriptase inhibitor (NNRTI) metabolism. Furthermore, the G516T SNP is associated with altered pharmacokinetics, and toxicity of NNRTI. CYP3A enzymes have a more minor role for NNRTI, but are important for metabolism of the protease inhibitors (PI) and also exhibit polymorphic expression. Since CYP2B6 and CYP3A4 are expressed in lymphocytes, we have performed a proof-of-concept study to assess the effect of genetic polymorphism on enzyme expression within HIV-replication competent cells.
Methods: Whole blood was obtained from 20 healthy volunteers by venopuncture and genomic DNA and lymphocytes prepared. All SNP were genotyped by real-time polymerase chain reaction (PCR) allelic discrimination with specific primers and probes. Quantitative real-time PCR was used to examine the effect of CYP2B6 (G516T, C1459T), CYP3A4 (*1B), and CYP3A5 (*3) SNP on lymphocyte expression of the enzymes.
Results: The allele frequencies for CYP2B6 516T, CYP2B6 1459T, CYP3A4*1B, and CYP3A5*3 were 17.5%, 12.5%, 0%, and 5%, respectively. The median (range) expression was 2.8 (2.3 to 22.9) for CYP2B6 and 2.3 (2.3 to 11.7) for CYP3A4. No correlation was observed between 2B6 and 3A4 (r2 = 0.01; p = 0.67). A significant association was observed between G516T and lymphocyte CYP2B6 expression (10.1±2.4 in G homozygotes vs 2.3±0.03 in heterozygotes; p = 0.003, 95%CI 3.4, 12.2; see the figure). No differences in CYP2B6 and CYP3A4 expression were observed for the other SNP.
Conclusions: These data indicate that the previously reported association between CYP2B6 SNP and expression is not limited to hepatic tissue. Metabolism of NNRTI at their site of action may be an additional factor impacting on response.
FASl-670 and APOC3 Polymorphisms as Predictors of Lipoatrophy in Patients Receiving ART
A Cossarizza1, A Riva2, B Zanone Poma2, P Cicconi2, M Nasi1, V Broggini2, M Pinti1, A Cozzi Lepri3, A D'Arminio Monforte2, Massimo Galli*2, and LIPO I.Co.N.A. study group
1Univ Studi di Modena, Italy; 2Univ of Milan, Italy; and 3Royal Free and Univ Coll Med Sch, London, UK
Background: Single-nucleotide polymorphysms in genes involved in apoptosis and in adipocye metabolism may explain why lipoatrophy occurs in some, but not all, ART-treated individuals. The present study aims at evaluating the influence of FASl-670 and APOC3 polymorphisms on ART-associated lipoatrophy.
Methods: We included in the study 274 patients on ART who were enrolled in Lipo.I.Co.N.A. We assessed the distribution of FAS-670 and APOC3-455 polymorphisms and calculated crude and adjusted relative rates of lipoatrophy (time of the first observed fat loss at any site) using Poisson regression.
Results: In our population, 23.3% were female and 36.1% hepatitis C virus (HCV)+; 21.1% of the person-year follow-up was spent on non-nucleoside reverse transcriptase inhibitor (NNRTI)-containing regimen and 42.1% on protease inhibitor (PI); 47.3% was spent on zidovudine/stavudine (AZT/3TC), 20.5% on lamivudine/stavudine (d4t/3TC), and 16.4% on didanosine/stavudine (ddI/d4T). During 778 person-years' follow-up, we observed 55 lipoatrophy events, crude incidence rate 70.7 (54.2 to 92.0) per 1000 person-years follow-up. The distribution of FASL-670 was: AA 95 (35%), AG 113 (41%), and GG 66 (24%); of APOC3, ut was: CT 129 (47%), TT 107 (39%), and CC 38 (14%). After having adjusted for gender, HIV exposure, age, current viral load, HCV serology, NRTI pair/"third drugs" currently used, and months of pre-HAART exposure to NRTI, the relative rate of lipoatrophy comparing APOC3 CT with CC genotypes was 6.3 (95%CI 1.4 to 28.0, p = 0.014) while FASL-670 AG compared with AA was protective against lipoatrophy (RR 0.34, 95%CI 0.15 to 0.76, p = 0.009). Patients on d4T/3TC were at higher risk of lipoatrophy than those on AZT/3TC (adjusted RR 2.79, 95%IC 1.23 to 6.29, p = 0.013). The risk associated with the most unfavorable genetic scenario (overall adjusted RR for APOC3 CT combined with FASL670 = AA vs other possible combinations = 2.28 95%CI 1.13 to 4.57, p = 0.019) was not different in patients currently on d4T/3TC (RR = 3.25, 95%CI 0.95 to 11.1) or on another pair of NRTI (1.95, 95%CI 0.82 to 4.59, for the interaction p = 0.44) after having adjusted for other potential confounders (the same listed above).
Conclusions: Our study suggests that FAS genotype -670 AG is protective against lipoatrophy compared with the AA genotype, while patients with APOC3 genotype CT are at higher risk. Our results confirm previous observations linking nucleoside analogues, particularly d4T, to the emergence of lipoatrophy, even if there is no evidence of a greater detrimental effect on patients with an unfavorable genetic asset. The determination of APOC3 and FAS polymorphisms, a relatively simple and inexpensive assay, could be considered to identify patients at potentially higher risk of lipoatrophy.
Association of Lipid Changes in HAART-treated Individuals with Apolipoprotein Genotypes
Edward D Gometz*1,2, D Grimm1, M King1, R Rode1, G Hanna1, B Da Silva1, and D Katz1
1Abbott Labs, Abbott Park, IL, US and 2Univ of Chicago Sch of Med, IL, US
Background: The mechanism of lipid changes in patients receiving ART has not been delineated. A pharmacogenetic approach can be used to correlate genetic variants with lipid changes observed during treatment, thereby suggesting that the gene product may be part of a relevant pathway affected by a drug regimen. In the general population, apolipoprotein genotypes APOC3 3238C>G and APOA5 56C>G appear to be independent risk factors for increased triglycerides (TG). A recent report described increased TG changes in joint carriers of an APOE variant and APOC3 -482C>T, -455T>C, and 3238C>G variants treated with ritonavir (RTV)-containing regimens.
Methods: Relationships of total cholesterol (TC) and TG with variants of APOE, APOC3, and APOA5 genes were assessed in a phase 3 randomized, double-blind clinical trial (n = 653) comparing the safety and efficacy of regimens containing stavudine (d4T), lamivudine (3TC), and either lopinavir/ritonavir (LPV/r) or nelfinavir (NFV) for the initial treatment of HIV infection. Baseline TC and TG, and maximum and average TC and TG changes from baseline within 48 weeks were analyzed by analysis of co-variance including factors for genotype, sex, race/ethnicity, smoking status and treatment arm; and age, body mass index, viral load, CD4 count, and baseline TC or TG levels as co-variates. Non-significant terms were sequentially dropped from the models.
Results: Genotypes and on-study lipid data were available from 364 subjects. Baseline TC was not associated with any tested genotype. Baseline TG was associated with APOC3 3238C>G (p = 0.04) and APOA5 56C>G (p = 0.03) genotypes. Both maximum (p = 0.03) and average (p = 0.04) TC changes from baseline were inversely associated with APOE e4 genotype. Average TG change from baseline was associated with APOC3 3238C>G genotype (p = 0.03) but not with APOA5 56C>G genotype. No lipid parameters (baseline or changes) were associated with APOE e2, APOC3 -482C>T or APOC3 -455T>C genotypes. TG changes in LPV/r-treated subjects were not associated with joint APOE/APOC3 variant carrier status.
Conclusions: The mean magnitudes of TC changes from baseline in different APOE genotypes (e3/e3 > e 3/e4 > e4/e4) are consistent with increased affinity of the ApoE e4 protein for the low density lipoprotein receptor. Patients with APOC3 3238GG genotype (~4% of the genotyped subjects) had higher co-variate-adjusted average TG changes from baseline. These relationships were not dependent on treatment arm.
62% Steatosis in Coinfected; Abacavir & Nevirapine May Be Protective (03/07/06)
Heart Disease in HIV: Carotid IMT, measuring plaque in arteries (03/02/06) (Not from CROI)
Analysis of Cardiovascular Risk Factors in the HIV Outpatient Study (HOPS) Cohort (02/23/06)
Risk of Heart Disease Appears to Level Off in Kaiser in Northern California (02/23/06)
D.A.D. Study: risk of myocardial infarction (02/22/06)
The Take Home from CROI on Body Shape - Written for NATAP By David Alain Wohl, MD - The University of North Carolina AIDS Clinical Research and Treatment Unit (02/17/06)
Early HAART Associated with Improved Outcomes & Fewer Toxicities (02/14/06)
TESTOSTERONE EFFECT ON FAT MASS & DISTRIBUTION (02/13/06)
Effect of pioglitazone on HIV-1 lipoatrophy : a randomised, double-bind, placebo-controlled trial (ANRS 113) (02/13/06)
Metformin + Rosiglitazone for Lipodystrophy -Combination Study (02/13/06)
Metformin for Lipodystrophy (02/13/06)
Pioglitazone Improved Limb Lipoatrophy (02/10/06)