CD4+ T-cell decline after the interruption of antiretroviral therapy in ACTG A5170 is predicted by differential expression of genes in the ras signaling pathway

Author Department


Document Type

Article, Peer-reviewed

Publication Date



Patterns of expressed genes examined in cryopreserved peripheral blood mononuclear cells (PBMCs) of seropositive persons electing to stop antiretroviral therapy in the AIDS Clinical Trials Group Study A5170 were scrutinized to identify markers capable of predicting the likelihood of CD4+ T-cell depletion after cessation of antiretroviral therapy (ART). A5170 was a multicenter, 96-week, prospective study of HIV-infected patients with immunological preservation on ART who elected to interrupt therapy. Study entry required that the CD4 count was greater than 350 cells/mm(3) within 6 months of ART initiation. Median nadir CD4 count of enrollees was 436 cells/mm(3). Two cohorts, matched for clinical characteristics, were selected from A5170. Twenty-four patients with an absolute CD4 cell decline of less that 20% at week 24 (good outcome group) and 24 with a CD4 cell decline of >20% (poor outcome group) were studied. The good outcome group had a decline in CD4+ Tcell count that was 50% less than the poor outcome group. Significance analysis of microarrays identified differential gene expression (DE) in the two groups in data obtained from Affymetrix Human FOCUS GeneChips. DE was significantly higher in the poor outcome group than in the good outcome group. Prediction analysis of microarrays (PAM-R) identified genes that classified persons as to progression with greater than 80% accuracy at therapy interruption (TI) as well as at 24 weeks after TI. Gene set enrichment analysis (GSEA) identified a set of genes in the Ras signaling pathway, associated with the downregulation of apoptosis, as significantly upregulated in the good outcome group at cessation of ART. These observations identify specific host cell processes associated with differential outcome in this cohort after TI.

Publication ISSN