Comparative analysis of measures of viral reservoirs in HIV-1 eradication studies

Abstract

HIV-1 reservoirs preclude virus eradication in patients receiving highly active antiretroviral therapy (HAART). The best characterized reservoir is a small, difficult-to-quantify pool of resting memory CD4(+) T cells carrying latent but replication-competent viral genomes. Because strategies targeting this latent reservoir are now being tested in clinical trials, well-validated high-throughput assays that quantify this reservoir are urgently needed. Here we compare eleven different approaches for quantitating persistent HIV-1 in 30 patients on HAART, using the original viral outgrowth assay for resting CD4(+) T cells carrying inducible, replication-competent viral genomes as a standard for comparison. PCR-based assays for cells containing HIV-1 DNA gave infected cell frequencies at least 2 logs higher than the viral outgrowth assay, even in subjects who started HAART during acute/early infection. This difference may reflect defective viral genomes. The ratio of infected cell frequencies determined by viral outgrowth and PCR-based assays varied dramatically between patients. Although strong correlations with the viral outgrowth assay could not be formally excluded for most assays, correlations achieved statistical significance only for integrated HIV-1 DNA in peripheral blood mononuclear cells and HIV-1 RNA/DNA ratio in rectal CD4(+) T cells. Residual viremia was below the limit of detection in many subjects and did not correlate with the viral outgrowth assays. The dramatic differences in infected cell frequencies and the lack of a precise correlation between culture and PCR-based assays raise the possibility that the successful clearance of latently infected cells may be masked by a larger and variable pool of cells with defective proviruses. These defective proviruses are detected by PCR but may not be affected by reactivation strategies and may not require eradication to accomplish an effective cure. A molecular understanding of the discrepancy between infected cell frequencies measured by viral outgrowth versus PCR assays is an urgent priority in HIV-1 cure research.

Figure 1. HIV-1 persistence assessed by six different assays performed on the indicated cell or tissue obtained from patients starting HAART during acute/early infection (open symbols) or chronic infection (closed symbols).

Geometric mean values are indicated by a horizontal black line. Long red lines indicate the limit of detection (LOD) for the relevant assay with a standard input sample size. Short red lines indicate patient-specific limits of detection in cases where the blood or tissue sample was smaller than normal. Negative assays are indicated by samples plotted below the red LOD lines.

Figure 2. Correlation between assays for HIV-1-infected cells in patients on HAART.

(A) Correlation between infected cell frequencies measured in the viral outgrowth assay on purified resting CD4⁺ T cells and the droplet digital PCR assay for HIV-1 DNA in unfractionated PBMC. (B) Correlation between the viral outgrowth assay on purified resting CD4⁺ T cells and the droplet digital PCR assay for HIV-1 DNA in purified resting CD4⁺ T cells. (C) Correlation between the droplet digital PCR assay for HIV-1 DNA and the Alu PCR assay for integrated HIV-1 DNA in PBMC. (D) Correlation between the viral outgrowth assay on purified resting CD4⁺ T cells and the Alu PCR assay for integrated HIV-1 DNA in PBMC. (E) Correlation between the viral outgrowth assay on purified resting CD4⁺ T cells from the blood and the PCR assay for HIV-1 DNA in rectal CD4⁺ T cells. (F) Correlation between the viral outgrowth assay on purified resting CD4⁺ T cells from the blood and the single copy assay for residual viremia. Patients starting HAART during acute/early or chronic infection are indicated by open or closed symbols, respectively. See Figure 1 for patient ID key. The red lines indicate samples that fell below the limit of detection of the relevant assay. These samples were not used in the calculation of the correlation coefficients.

Figure 3. Ratio of infected cell frequencies determined by droplet digital PCR for HIV-1 DNA or by viral outgrowth assay.

Analysis was done on the same sample of purified resting CD4⁺ T cells. * indicates maximum values in cases in which the HIV-1 DNA level was below the limit of detection (2 copies/ml).