HIV-1 CNS in vitro infectivity models based on clinical CSF samples.

Abstract

The concentration of antiretrovirals in CSF is often utilized as a surrogate for CNS drug exposure. This measurement does not consider pharmacodynamic or combinative effects of ART. We have developed a novel endpoint measurement to assess antiretroviral activity of CSF from subjects on ART. CSF samples were obtained from patients receiving tenofovir/emtricitabine (245/200 mg once daily) with either rilpivirine (25 mg once daily) or lopinavir/ritonavir/maraviroc (400/100/150 mg twice daily) and HIV-uninfected controls. Antiviral activity of ART-containing CSF was assessed in cell cultures using PBMCs and neuro-derived glial (U87) and astrocyte (373) cell lines. Infectivity model half-maximal inhibitory concentration (IMIC50) values were calculated and expressed as -log2IMIC50. Results were correlated with CSF antiretroviral concentrations. Compared with controls, CSF from both ART studies demonstrated in vitro antiretroviral activity in all models. CSF antiretroviral activity of patients on lopinavir/ritonavir/maraviroc was significantly greater than that of patients on rilpivirine [-log2IMIC50 (95% CI) 4.82 (4.74-4.89) versus 3.43 (3.33-3.54) in PBMCs, 3.06 (2.98-3.15) versus 2.56 (2.46-2.65) in U87 cells and 6.00 (6.11-5.88) versus 4.90 (5.09-4.72) in 373 cells, respectively]. Positive correlations were observed for individual CSF antiretroviral activity in different cellular models with CSF concentrations of rilpivirine (P = 0.040 in 373 cells) and lopinavir (P = 0.048 in 373 cells), but not maraviroc. Antiviral activity of CSF from patients on ART was successfully calculated and was greater with a regimen containing four active drugs compared with three active drugs. The use of neuro-derived cell lines alongside PBMCs to assess the effect of ART on CSF may act as a useful future clinical research tool.