Tumour necrosis factor-alpha gene -238G/A promoter polymorphism associated with a more rapid onset of lipodystrophy

The effect of TNF polymorphisms on the time-to-onset of lipodystrophy was investigated in 191 Caucasian participants in the western Australian HIV cohort. Carriage of the TNF-α-238G/A polymorphism was found in 13.1% of the cohort, and was associated with an earlier onset of lipodystrophy. The presence of TNF-α-238G/A was associated with lipodystrophy progression in Cox proportional hazards regression analyses, without influencing the relative risk and significance of other predictive host and treatment-related variables.

Progressive subcutaneous fat wasting is a component of the antiretroviral therapy-associated lipodystrophy syndrome that is highly prevalent among recipients of highly active antiretroviral therapy (HAART) in white Caucasian populations. The dominant risk factor for the development of fat wasting appears to be the concurrent use of nucleoside analogue reverse transcriptase inhibitors (NRTI) and HIV protease inhibitors (PI), with additional evidence of a differential risk associated with select NRTI drugs [¹]. There is increasing awareness, however, that host factors also significantly modify the susceptibility to the development of fat wasting, as well as its severity [²]. Most notably, older age and white racial origin are factors associated with an increased likelihood of developing subcutaneous fat wasting. In the present study, we sought to investigate a possible role for functional promoter polymorphisms in TNF-α and -β loci, as the expression of this cytokine may be involved in adipose tissue pathology associated with lipodystrophy [³], and the carriage frequency of these polymorphisms is divergent among racial populations [⁴]. In addition, we addressed the hypothesis that these host factors may modify the independent contribution of antiretroviral therapy-specific factors to the progression of lipodystrophy.

In a previous study involving participants in the western Australian HIV Cohort, host and treatment-related factors were found independently to increase the rate of progression to clinically apparent subcutaneous fat wasting in participants on a HAART regimen (n = 230) in a time-to-onset analysis (Cox proportional hazards regression) [⁵]. An increased relative risk (RR) was associated with older age (RR 1.052/year, P < 0.001), white race (RR 3.9, P = 0.023), duration of dual NRTI therapy before the commencement of triple combination antiretroviral therapy (RR 1.021/month, P = 0.005), and increased cumulative time on a regimen containing stavudine compared with regimens incorporating zidovudine therapy (RR 1.085/month, P < 0.001). The use of the non-nucleoside reverse transcriptase inhibitor nevirapine was associated with a reduced rate of progression, compared with regimens incorporating PI therapy (RR 0.943/month, P = 0.022).

This cohort was utilized for the present study, in which stored DNA samples of 220 participants were typed for polymorphisms in the promoter regions of TNF-α (-238G/A and -308G/A), and TNF-β (-250G/C). Results were incorporated into Cox proportional hazards regression analyses (SAS statistical package, SAS Institute, Cary, NC, USA), to assess the effect of variant alleles at these loci in the development of subcutaneous fat wasting, with adjustment for known predictive variables.

Heterozygosity at the -238G/A TNF-α polymorphism was found in 25 out of 220 individuals (11.4%), all of whom were of white racial origin. To characterize whether this TNF polymorphism exerted an effect on lipodystrophy progression, analyses were performed in a dataset limited to 191 white Caucasian participants, thereby excluding potential confounding attributable to the role of the TNF-α-238G/A polymorphism as a surrogate marker of white racial origin. Carriage frequency in this restricted dataset was 25 out of 191 (13.1%). The risk of progression to lipodystrophy was significantly increased in -238G/A hetrozygotes compared with carriers of the wild-type allele (P = 0.014, log rank, Fig. 1). Incorporation of this variable in Cox proportional hazards analyses demonstrated a significant independent effect of TNF-α-238G/A heterozygosity (RR 1.73, P = 0.041), which did not impact on the effects of other predictive variables (Table 1).

Fig. 1

Table 1

Promoter polymorphisms at TNF-α-308G/A and TNF-β-250G/C were found not to contribute significantly to the Cox proportional hazards regressions (P = 0.22 and P = 0.12, respectively). In addition, analyses combining TNF-α-238G/A heterozygosity and treatment-related variables as interaction terms did not suggest synergistic effects of TNF polymorphism and antiretroviral therapy on the progression of fat wasting.

In this study, carriage of the TNF-α-238G/A promoter polymorphism independently increased the risk of lipodystrophy progression in a cohort of white Caucasian HAART recipients. Similar results have been presented by Maher and colleagues [⁶] in a UK case control study involving 96 participants, in which the -238G/A variant allele was more frequent among HAART recipients with lipodystrophy (9/61, 14.7%) than in those with no evidence of lipodystrophy (0/35, 0%, P = 0.01).

The possible explanations for the effect of the TNF-α-238G/A promoter polymorphism on lipodystrophy progression are speculative at present. Altered TNF-α expression may modulate the 'downstream' effects of antiretroviral therapy in adipose tissue, where this cytokine is involved in determining insulin sensitivity and the differentiation of adipocytes [⁷], as well as transducing apoptotic stimuli [⁸]. Consistent with this possibility, significantly elevated adipose TNF-α protein expression has been found in individuals affected with lipodystrophy [³], whereas synergistic antiadipogenic effects of PI therapy and exogenous TNF-α have also been noted in vitro [⁹]. Morphological studies of affected adipose tissue [^10,11] in lipodystrophic HIV-infected individuals have also revealed the presence of macrophage-derived lipogranulomata that appear to be recruited to apoptotic cells, consistent with a role for inflammatory cytokines in tissue pathology. These results suggest that the presence of the -238G/A promoter polymorphism may influence lipodystrophy progression through a stimulatory effect on TNF-α transcription and expression. However, there is a lack of consensus regarding the effect of this polymorphism on TNF regulation, with evidence for both increased and decreased transcriptional activity in restricted cell lines [⁴]. Therefore, it is plausible that the -238 variant allele, which has strong linkage disequilibrium with specific allelic combinations within the major histocompatibility complex [⁴], may be acting as a surrogate marker in these analyses for an unknown genetic factor.

We conclude that the TNF-α-238G/A promoter polymorphism defines a host factor that may influence the progression of lipodystrophy, although the mechanism awaits characterization. The influence of treatment-related variables on lipodystrophy progression was not altered after adjustment for the presence of the TNF-α-238G/A polymorphism, indicating that the effects of therapy are not contingent on the presence of this host factor.