627 – HIV Transcripts in Blood May Identify Patients at Lower Risk of Brain Injury in Cure Research

Background

Suppressed HIV patients’ PBMC have residual cell associated HIV-1 RNA transcripts (HRT). Elevated HRT, despite ART, associate with viral blips and with brain injury. Most cure studies aim to reduce HIV DNA reservoir but may require initial viral reactivation, risking brain injury from viral blips due to reservoir activity. Under ART, blips are believed to reflect reservoir size, but whether they could be independently due to reservoir activity is unclear. Current understanding of blips does not allow prediction of viral blips. We hypothesized that HRT levels measured by the Double R assay may predict risk of blips, helping identify low risk patients for cure studies.  

Methods

RNA and DNA were extracted from cells from 5 ml blood samples from 64 patients on suppressive ART at baseline and sequentially for 4y Of these, 32 had 1 or 2 blips (isolated detectable pVLs <200 cp/ml in the 2y prior to baseline), 29 had no blips, and 3 were Elite Controllers (to serve as benchmarks of functional cure). HRT was measured by the Double R assay. Multivariable mixed-effects logistic regression was used to predict any future blip over 4y. A cut-off value of HRT was determined using Youden’s index with Receiver Operator Characteristic (ROC) curves.

Results

Baseline HRT levels in the “blip” group were significantly higher than for the “non-blip” group: median 168 vs 22 copies/10⁶ cells; p<0.001. Multivariable mixed-effects logistic regression showed HRT was strongly associated with increased risk of blip within 4y [OR 3.2, CI=2.1-5.1, p<0.001] per 10-fold increase in HRT. ROC area under the curve was 0.73, (CI 0.66-079) using cut-off of 91 copies/10⁶ cells, showing 71% sensitivity and 67% specificity for future blips. Baseline HIV-1 DNA copies were associated, unadjusted, with increased odds ratio of a future Blip (3.6, CI 2.4-6, p<0.001), but HIV-1 DNA did not significantly predict blips after adjustment for HRT (p=0.60). PBMC from 3 Elite controllers were negative for HRT, but in vitro stimulation of their CD4+ T cells for 1 week with anti-CD3/CD28/CD2 +IL-2 resulted in 1/10, 3/10 and 0/16 replicate cultures being positive for HRT. We also identified those with extremely low-levels of HRT had no significant brain injury in our previous dataset. 

Conclusions

HRT detection by the Double R assay is a scalable tool to quantify reservoir activity, independent of reservoir size, to identify potential patients with low risk of brain injury during cure research.

Table or Figure