401
Validation of a V3-based Genetic Biomarker of HIV-1 Neurovirulence
Satish Pillai*1,2, W Pasutti1, A McCutchan3, C Marra4, D Clifford5, B Gelman6, D Simpson7, J McArthur8, I Grant3, and J Wong1,2
1VAMC, San Francisco, CA, US; 2Univ of California, San Francisco, US; 3Univ of California, San Diego, US; 4Univ of Washington, Seattle, US; 5Washington Univ, St Louis, MO, US; 6Univ of Texas Med Branch, Galveston, US; 7Mt Sinai Sch of Med, New York, NY, US; and 8Johns Hopkins Univ, Baltimore, MD, US
Background: Although ART has decreased the incidence
of HIV-associated neurocognitive impairment, the prevalence may be rising due
to the beneficial effects of ART on the life span of HIV-infected individuals.
We previously studied the effects of viral genotype on HIV-1 neurovirulence by
analyzing sequences of the HIV-1 env gene from 18 HIV-infected
individuals with available neuropsychiatric data. This exploratory analysis
revealed that the presence of serine at gp160 position 300 (V3 loop position 5)
was strongly associated with impaired neuropsychological performance. Here we
rigorously evaluate and validate gp160 N300S as a viral genetic biomarker of
neurovirulence, focusing on a separate population of well-characterized
HIV-infected individuals.
Methods: From the CHARTER (CNS HIV Anti-Retroviral
Effects Therapy Research) cohort, 39 subjects were selected for analysis based
on neuropsychological performance, summarized as a global deficit score (GDS): 19
were cognitively impaired (GDS >0.5, median 1.89), and 20 were cognitively
normal (GDS ≤0.5, median 0.11). All subjects were free of CNS
opportunistic infections. Clonal RNA sequences of the HIV-1 env C2-V3
region were generated and analyzed from both cerebrospinal fluid (CSF) and
plasma samples. Logistic regression and machine learning techniques were used
to identify sequence correlates of cognitive impairment.
Results: HIV-1 gp160 N300S was associated with
impaired neuropsychological performance (p <0.05, Fisher's exact),
and was predictive of neurocognitive status independent of CSF RNA, nadir CD4,
glycosylation extent, and coreceptor usage phenotype. Positive predictive value
= 0.86, sensitivity = 0.32, and specificity = 0.94.
Conclusions: These results provide further evidence
that gp160 N300S plays a key role in the genetic signature underlying HIV-1
neurovirulence. The high positive predictive value of this mutation suggests
that genetic features of the HIV-1 envelope could be used to predict the onset
of HIV-associated cognitive impairment in infected individuals, which may
ultimately guide therapeutic decisions in the future (e.g., by optimizing
regimens for maximal CNS penetration). The phenotypic consequences of variation
at position 300 will be investigated as a continuation of this work using in
vitro models of neuropathogenesis and by modeling the influence of C2-V3
sequence variation on three-dimensional conformation of the V3 loop.
|
