Replication competence of virions induced from CD4+ lymphocytes latently infected with HIV

HIV DNA integrated into host cell DNA of resting CD4⁺ T lymphocytes can be induced to produce infectious virus upon cellular activation. Effective inhibition of virus replication by antiretroviral therapy (ART) of HIV-infected individuals cannot cure HIV infection because of this latent HIV reservoir [1,2,3]. The quantitative viral outgrowth assay (QVOA) measures the frequency of latently infected cells by expanding replication competent virus after activation of terminally diluted CD4 lymphocytes, with most assays using resting CD4 cells [4]. Therefore, in this assay, replication competent virus must propagate in uninfected cells, resulting in ongoing amplification of viral protein (p24) or RNA. Approximately one in a million CD4⁺ T cells is latently infected using standard QVOA assays; however, intact replication competent proviruses may exceed those that can are measured with the standard QVOA assay by on average 64-fold in one study [5] and 25 to 27-fold in a subsequent study [6, 7]. An important question is how much this discrepancy is attributable either to provirus integrated into sites not amenable to reactivation or to the inefficiencies of the conditions of in vitro assays compared to in vivo conditions. Repeated reactivation of cells can disclose additionally replication competent proviruses, but do not increase QVOA values more than twofold [8].

Recently efforts to improve QVOA sensitivity have been implemented by replacing the standard HIV p24 ELISA to detect the production of virus induced by activated cells in culture supernatants using either a PCR assay for cell-free HIV RNA or a more sensitive digital ELISA with increased sensitivity for HIV p24 (at femtogram levels) [7, 9, 10]. The main concern with these assays is that their sensitivity for virion components in the culture supernatant may also detect low levels of induced extracellular virions that are not replication competent. To ascertain how much virion production, as measured with PCR for cell-free HIV RNA, represented the induction of replication competent virus from latently infected cells, we performed parallel QVOA assays in the presence or absence of the integrase inhibitor raltegravir, and measured virus production frequently over the course of 2 weeks. These assays were performed with the CD4⁺ T cells from 5 subjects effectively treated with ART for more than 1 year, as detailed in Methods. This approach permitted a sense of the kinetics of virion induction without ongoing propagation in the presence of raltegravir compared to a measurement of the frequency of latently infected cells that were replication competent in the absence of raltegravir.

The time course of production and persistence of supernatant virions in the presence of raltegravir, which prevents ongoing viral propagation can be seen in the left panels of Fig. 1a. The time course of production and persistence of supernatant virions in the absence of raltegravir, thus permitting viral propagation, can be seen in the right panels of Fig. 1a. In the assays shown there are 6 replicates at each of the threefold dilutions starting with 1 million cells per well in dilution 1.

Parallel QVOA results in the presence or absence of raltegravir with CD4 lymphocytes from subject 219. Six replicates were performed for each serial threefold dilution of CD4 cells in the QVOA assay with (a, left row) or without (a right row) raltegravir in the culture medium throughout the assay shown on a log₁₀ scale. Culture supernatants were assayed for HIV RNA by real time (RT)-PCR. The asterisks on the lines of the replicates without raltegravir on day 14 indicate values for each well without raltegravir that exceeded the mean plus 5 standard deviations of the 6 replicates performed with raltegravir. b Displays the same data for day 14 to indicate visually that virions are induced in the presence of raltegravir (upper figure) and that amplification occurs in almost as many wells (lower figure)

Full size image

As can be seen in Fig. 1, cells can be induced to produce a substantial number of virions even when ongoing replication is inhibited with raltegravir. Values tend to diminish with each dilution as the number of cells with HIV proviruses that can be induced become fewer. It is not possible to clearly ascertain the number of cells carrying proviruses in each well that contribute to cell free HIV RNA, although the likelihood of just a single cell contributing increases with each dilution of input CD4+ cell number. In the presence of raltegravir virions are usually detectable 3 days after induction; however, in general either new cells begin to produce virus after 3 days or the cells progressively increase the production of virus by day 7 after activation. Virion production persists and remains in the supernatant even while the CD4+ MOLT-4/CCR5 cells propagate in the culture wells. Continuing production is likely, because virions at a level of 1000 copies of HIV RNA when placed in a cell-free well on day 0 do not persist and are no longer detectable after 1 or 2 days (data not shown).

Since it is not possible to measure the induction and propagation of HIV after activation in the same cells both with and without raltegravir, parallel experiments were performed with cells from the same participant. The results of these comparisons performed as shown in Fig. 1 were performed in 4 other study subjects (Additional file 1: Fig. S1). Different definitions of replication competence in the culture wells without raltegravir were then explored to ascertain a sense of the difference between total induced virions (with raltegravir) compared to the individual culture wells without raltegravir that fulfill a definition of propagation of replication competent virus. Two different criteria of amplification in the absence of the inhibitor were examined in Table 1.

The results of these experiments indicated that the detection of induced virion production from latently HIV-infected cells is a close reflection of the number of cells harboring replication competent virus, although for most subjects on average virus amplification is lower than the number of cells generating virions on most days of culture (Fig. 2). The values for IUPM with the QVOA without raltegravir on day 14 ranged between 1 and 11-fold higher than those in the presence of raltegravir (Table 1). When the 2 rigorous criteria for amplification of replication competent virus were examined, the IUPM calculations were not substantially diminished. The results from this study indicated that there are more CD4⁺ T cells that produce virions after activation than those that produce replication competent virus; nevertheless; by examining a rigorous criterion for amplification of replication competence at day 14, virions that are induced by activation of CD4⁺ T cells are in general replication competent. A recent publication using the digital p24 assay in the QVOA from 1 subject showed 8 induced cells generated viral antigen that amplified and 9 that did not, consistent with the results from this study [11].

IUPM values with or without raltegravir for each day of culture for the 5 study subjects. IUPM values (with 95% CI) are graphed for each study subject for each day of supernatant collection. Positive values were defined as > 100 copies HIV RNA/mL. Also indicated for each subject with the day 14 results are the IUPM values using the 2 criteria for amplification (replication competence): day 14 criterion (▲) and AUC criterion (▼)

Full size image

There are a number of sources of imprecision and variability in these measurements. First parallel assays needed to be performed, because there is no way to have an individual well with the same cells be both exposed and not exposed to raltegravir. Second, as indicated with the 95% confidence intervals, there is inherent imprecision in any terminal dilution assay that contains a realistic number of dilutions and replicates per dilution, considering the cells available from study subjects. Third, it is well established that after prolonged ART, the proportion of clonally expanded CD4 cells with integrated HIV DNA progressively increases [8, 12, 13]. The composition of these clones, which can be either competent or defective proviruses will certainly vary between subjects. Nevertheless, the detection of HIV antigen or RNA can represent a relatively accurate surrogate measurement of the true replication competent latent reservoir. A 3-day assay of virion induction (cell-free RNA) was recently shown to produce measurements within twofold of the increased sensitivity of using HIV RNA to measure propagation of HIV after a 14-day QVOA as used in these experiments [7, 9]. The replication competence of the virions in that study was not determined, although the values generated were within twofold of the IUPM values. In the current study parallel assays with and without raltegravir confirmed that induction of virions generates similar values to those determined to be replication competent. The possibility is thus raised that a shorter and easier measurement of induced virus may prove of similar utility; however, validation will require comparing these assays when an effective intervention becomes available.

Table 1 indicates that measuring statistically significant amplification without raltegravir generates IUPM values in a similar range to the IUPM values using RNA produced in the presence of raltegravir. Moreover, the detection of cell-free virions in these assays that do not propagate may be important targets for candidate HIV “cure” strategies. First, failure to demonstrate propagation of a virion in a culture well may not reflect the capacity of that virion to ignite ongoing replication in the lymphoid tissue of an infected individual where conditions and cell density may be much more conducive to replication [14]. Second, even defective virions, which do compose some of the low-level viremia during effective ART, may still contribute to the immune activation and the non-AIDS related complications of well treated HIV infection. In summary, a high proportion and often the majority of cells that can be induced to generate virions upon activation are producing replication competent virions. This information should support the rationale to utilize more sensitive and shorter assays to measure the replication competent latent HIV reservoir.