Session 59Poster Presentations Viral Genetic Diversity Session Day and Time: Thursday 1:30 - 3:30 pm Room: Hall A
498 A Bioinformatic Predictor of Coreceptor Usage Correlates with Markers of Disease Progression and Supports the Gradual Evolution of X4 Virus via R5X4 Intermediates M. Jensen*1, G. Gottlieb1, A. van 't Wout1, F.-S. Li2, D. Nickle1, D. Shriner1, K. Wong1, H.-X. He1, S. McLaughlin1, R. Shankarappa3, J. Margolick4, J. Mullins1 1Univ of Washington, Seattle; 2Primal Inc, Seattle, WA; 3Allegheny-Singer Res Inst, Pittsburgh, PA; and 4Johns Hopkins Sch of Hygiene and Publ Hlth, Baltimore, MD
Background: HIV in primary infection is usually R5 in phenotype, e.g., it uses only CCR5 as a co-receptor. Often, the virus undergoes a genetic shift, leading to the outgrowth of CXCR4-using X4 virus, associated with increased pathogenicity and faster disease progression. We previously validated a bioinformatic predictor of X4 phenotype. This predictor scores a env V3 loop sequence based on its similarity to known X4 or dual-tropic (R5X4) viruses. The score can be interpreted as the propensity to use CXCR4: known R5 viruses had low scores; R5X4 viruses, intermediate scores; and X4 viruses, high scores. We apply this method to viral sequences from 30 patients (pts) sampled over time. We conjectured that the score may be a better indicator/predictor of pathogenesis than methods using the presence of basic residues at V3 sites. We also address the hypothesis that X4 evolution within pts is gradual, involving an intermediate R5X4-prevalent phase.
Methods: Viral sequences were sampled over time from: 9 intermediate progressors (IP; seroconversion (SC) to death/HAART: 6-13 yrs); 2 long-term (>13 yrs) nonprogressors (LTNP); and 19 rapid progressors (RP; SC to death: < 3 yrs), and scored by a position-specific scoring matrix method. For the IP, viral phylogenies were calculated and V3 ancestors estimated and scored. We examined average score vs viral load and CD4 count and time of peak average score vs CD3+ homeostasis failure, using linear regression.
Results: IP acquiring high scoring (X4) virus (8/9) exhibited a gradual transition from low- to high-scoring virus, involving putative R5X4 virus at the ancestral nodes. X4s were often lost over time both by extinction and reversion. For 6 IP, mean X4 score was inversely correlated with CD4 count; for 3, this association persisted upon correction for viral load increase. Peak mean score was a better predictor of CD3+ decline than presence of basic residues. The LTNP and fastest IP never evolved high scores, despite the occurrence of basic mutations. Only 2/19 RP had predicted X4 virus within 1 yr of SC.
Conclusions: Our analysis supports gradual X4 evolution involving R5X4 intermediates. X4s may be more likely to influence disease in IP than in RP or LTNP. In IP, mean X4 score followed the course of immune decline. The association in some pts of score with CD4 count, after correcting for viral load, suggests that high scoring virus may have increased pathogenicity in vivo.
