UCLA

Background: B-cell non-Hodgkin lymphoma (NHL) in HIV populations (AIDS-NHL) has become the leading cause of AIDS-defining cancers. Studies suggested that genetic or serum markers of B-cell activation are related to AIDS-NHL. However, associations between HIV viral load and AIDS-NHL risk have not been explicitly explored with consideration of B-cell activation markers. Furthermore, associations of hepatitis C virus (HCV) infection to AIDS-NHL risk are inconclusive.

Methods: We used two nested case-control studies within the Multicenter AIDS Cohort Study - the genetic association study including 183 AIDS-NHL cases and 533 matched HIV-positive controls, and the serum cytokine study of 179 AIDS-NHL cases and 179 HIV-infected controls - to explore associations between HIV viral load or HCV infection and AIDS-NHL risk. To screen out possible B-cell activation biomarkers, we implemented weighted gene co-expression network analysis (WGCNA) and visANT. We explored three HIV viral load measurements: (1) viral load at set point, indicating a balance between HIV viral replication and clearance, (2) pre-HAART viral load closest and prior to AIDS-NHL diagnosis, and (3) slope and intercept for linear regression on viral load by year. HCV infection was measured at multiple time-points retrospectively using stored serum specimen or prospectively since 2001. We calculated odds ratios (ORs) and 95% confidence interval (CIs), using conditional logistic regression models including potential confounders and additional single nucleotide polymorphisms (SNPs) or serum cytokines levels suggested by the WGCNA.

Results: The WGCNA screened out sCD27, sCD30, and CXCL13 from 328 SNPs and 31 cytokines measured longer than 3-year preceeding AIDS-NHL diagnosis. One-unit increase of log10-HIV RNA at set point was associated with an increased AIDS-NHL risk (SNPs adjusted OR: 2.01; 95% CI: 1.35-2.99, sCD27 and sCD30 adjusted OR: 3.65; 95% CI: 1.52-8.72). Ever HCV infection was possibly associated with AIDS-NHL susceptibility (adjusted OR: 1.77; 95% CI: 0.89-3.50). Association remained after SNP adjustment (OR: 1.75, 95% CI: 0.97-3.13).

Conclusion: Our results suggested a higher viral-load set point was associated with elevated AIDS-NHL risk from models with or without biomarkers. Point-estimates consistently indicated associations between HCV and AIDS-NHL risk.