Blood biomarkers of expressed and inducible HIV-1

Abstract

Objective: To define the relationships between molecular measures of viral persistence in blood (i.e., plasma viremia, cellular HIV-1 DNA, and mRNA) and expressed or inducible virus from resting CD4 T cells of individuals on suppressive antiretroviral therapy.

Design: We compared molecular measurements of HIV-1 in plasma and in uncultured peripheral blood mononuclear cells (PBMCs) to the levels of virions produced by either unstimulated or phorbol myristate acetate and ionomycin (PMA/iono)-stimulated PBMC or resting CD4 T cells from 21 donors on suppressive antiretroviral therapy.

Results: We found that unstimulated virion release from cultured resting CD4 T cells was positively correlated with the levels of plasma viremia in vivo (Spearman rho = 0.67, P = 0.0017). We also found that levels of both cellular HIV-1 DNA and unspliced HIV-1 mRNA per million uncultured PBMC were positively correlated with the levels of inducible virion release from both PMA/iono-stimulated PBMC (total HIV-1 DNA: rho = 0.64, P = 0.0017; unspliced HIV-1 RNA: rho = 0.77, P < 0.001) and PMA/iono-stimulated resting CD4 T cells (total HIV-1 DNA: rho = 0.75, P < 0.001; unspliced HIV-1 RNA: rho = 0.75, P < 0.001).

Conclusion: These results show for the first time that there are strong associations between in-vivo measures of HIV-1 persistence and ex-vivo measures of spontaneous and inducible virus production from cultured PBMC and resting CD4 T cells. Findings from this study provide insight into the biology of HIV-1 persistence and suggest methods to guide the evaluation of clinical strategies to reduce the size of the viral reservoir.

Fig. 1. Correlogram showing the interrelationships between the variables studied

Each of the 13 variables studied is listed on the x- and y-axes. The bottom left-hand portion of the correlogram shows scatter plots between the individual variables, and the top right-hand portion of the plots shows a heat map of correlations between variables. The key on the right-hand side shows the colors corresponding to strong positive correlations (red), strong negative correlations (blue), and neutral correlations (white). Asterisks indicate comparisons that were statistically significant by Spearman’s correlation after correcting for multiple comparisons by maintaining a false discovery rate of 5%. In total, 18 of the 78 comparisons were statistically significant. Highly interrelated variables included total cellular HIV-1 DNA, cellular unspliced HIV-1 RNA, and virion release from stimulated resting CD4+T-cells.

Fig. 2. Unstimulated virus release from cultured PBMC and resting CD4+T-cells correlates with persistent viremia on ART

(a) Unstimulated virus release from 15×10⁶ PBMC following 5 days of culture correlated with levels of plasma viremia (HIV-1 RNA) measured by large-volume iSCA (rho=0.50, p=0.098). (b) Similarly, unstimulated virus release from 5×10⁶ resting CD4+T-cells after 7 days of culture correlated with the levels of plasma viremia by large-volume iSCA (rho = 0.67, p=0.0016). In both (a) and (b), open symbols denote samples that had undetectable HIV-1 RNA by large-volume iSCA (interpolated at 50% of the limit of detection based on plasma volume assayed) or CAP/CTM qRT-PCR of culture supernatant (interpolated at 50% of the limit of detection, or 10 copies per milliliter of culture supernatant assayed).

Fig. 3. The frequency of infected cells (cellular HIV-1 DNA) is correlated with HIV-1 transcriptional activity in PBMC (cellular HIV-1 RNA)

Levels of total cellular HIV-1 DNA were significantly correlated with basal levels of unspliced HIV-1 RNA in PBMC that were cryopreserved after processing the leukapheresis product. The open symbol denotes a sample with undetectable HIV-1 RNA, interpolated at 50% of the limit of detection, or 0.5 copies of HIV-1 RNA per million PBMC.

Fig. 4. Spontaneous virion release is correlated with the frequency of infected cells and their transcriptional activity in PBMC

(a) The frequency of infected cells (cellular HIV-1 DNA in PBMC) is correlated with the spontaneous release of virions from unstimulated cultured PBMC (rho=0.55, p=0.010). (b) The level of unspliced cellular HIV-1 RNA transcription in PBMC is correlated with spontaneous virion release from PBMC (rho=0.64, p=0.0016). (c) The frequency of infected cells in PBMC is correlated with spontaneous virion release from resting CD4+T-cells (rho=0.69, p<0.001). (d) The level of unspliced cellular HIV-1 RNA transcription in PBMC is correlated with spontaneous virion release from resting CD4+ T cells (rho=0.55, p=0.010). In (a), (b), (c), and (d), open circles represent samples that were undetectable by CAP/CTM qRT-PCR for spontaneous virion release, or one sample that had undetectable levels of cellular unspliced HIV-1 RNA. In all cases of non-detection, values were interpolated as 50% of the limit of detection of the assay (i.e., 10 copies per milliliter of supernatant for CAP/CTM and 0.5 copies of unspliced cellular HIV-1 RNA per million PBMC).

Fig. 5. The frequency of infected cells and their transcriptional activity in PBMC are correlated with the level of inducible virion release from PBMC and resting CD4+T-cells

(a) The frequency of infected cells in PBMC is correlated with the level of inducible virion release from cultured PBMC treated with PMA/iono for 5 days (rho=0.64, p=0.0017). (b) The level of cellular unspliced HIV-1 RNA transcription is correlated with the level of inducible virion release from cultured PBMC (rho=0.77, p<0.001). (c) The frequency of infected cells in PBMC was correlated with the level of inducible virion release from cultured resting CD4+T-cells treated with PMA/iono for 7 days (rho=0.75, p<0.001). (d) The level of cellular unspliced HIV-1 RNA in PBMC is correlated with the level of inducible virion release from resting CD4+T-cells stimulated with PMA/iono (rho=0.75, p<0.001). In (b) and (d), one sample (shown as an open circle) had undetectable levels of cellular unspliced HIV-1 RNA. The value of this sample was interpolated as 50% of the limit of detection of the assay, or 0.5 copies per million PBMC.