Long-term effect of temporary ART initiated during primary HIV-1 infection on viral persistence

Categories: Disease & Virus

July 31, 2025

HIV-1 persistence and immunological response dynamics during early ART

All 64 participants of the Primo-SHM study who completed the trial in the Academic Medical Center of the University of Amsterdam (AMC) were included in this study. Baseline and treatment characteristics of the study participants are shown in Table 1 and extensive characteristics are published elsewhere²⁷. Participants received no treatment (n = 12), 24 weeks (n = 23), or 60 weeks (n = 29) of early ART (Fig. 1). We started with the longitudinal quantification of US RNA and total DNA during early ART in participants who were randomized to receive 24 or 60 weeks of early ART (n = 52). Parameters were measured at baseline (at PHI, before early ART initiation) and every 12 weeks on ART until TI. Thus, participants of the 24-week arm were assessed at 12 and 24 weeks of early ART, while participants of the 60-week arm were additionally assessed at 36, 48, and 60 weeks. We also assessed plasma viral loads, CD4 + T-cell counts, CD4/CD8 ratios, and calculated HIV-1 transcription levels per provirus (US RNA/total DNA ratios) at the same time points.

Table 1 Baseline and treatment characteristics of the study participants (early ART)

As expected, no differences in the longitudinal dynamics of any parameter between the 24-week and 60-week arms were observed (Fig. S1), therefore these two arms were pooled together for further analysis. Figure 2a shows the levels of the measured parameters at baseline and under early ART, while Fig. 3a shows the longitudinal dynamics of these parameters fitted to a two-phase segmentation model. Fig. S2A shows the significance levels of the differences between the time points for all markers. Plasma viral load demonstrated a two-phase decay with a change point at 12 weeks of early ART. By week 24, 92.3% of participants (and by week 36, all remaining participants) achieved virological suppression in plasma to <50 copies/ml and maintained it throughout the therapy period, apart from occasional isolated “blips” that remained below 150 copies/ml. CD4+ count and CD4/CD8 ratio increased between baseline and week 12 but did not significantly change afterwards (Fig. S3A shows the relative increases from baseline of CD4+ counts and CD4/CD8 ratios). Similar to the plasma viral load, US RNA dramatically decreased between baseline and week 12 (median decrease (interquartile range, IQR), 49 (17–389) fold) and remained relatively stable afterwards, although some fluctuations were observed (US RNA levels at 36 weeks were lower than at 24 and 48 weeks ART). In contrast, total DNA showed a continuous decrease until 36 weeks, after which it remained stable. The decay of total DNA during the first phase was much slower than that of US RNA or plasma viral load (T_1/2 = 12.9 vs. 2.11 vs. 1.17 weeks, respectively). Similar differences in the decay of total HIV-1 DNA and US RNA upon ART initiation were reported by us and others previously^8,28,29. The gradual decrease in total DNA reflects the elimination of rare cells with intact proviruses by the host immunity in the presence of the large background of defective proviruses that show a much slower decay on ART^30,31,32, whereas the biphasic kinetics of US RNA reflects the steep decline of productively infected cells upon ART initiation, followed by a quasi-steady state that is fuelled by stochastic reactivation of latently infected cells. Interestingly, the US RNA/total DNA ratio demonstrated a sharp decrease between baseline and week 12 (T_1/2 = 2.69 weeks), followed by a significant, albeit much slower, increase (T_d = 16.8 weeks).

**Fig. 2: Levels of HIV-1 persistence and immunological response parameters.**

**Fig. 3: Longitudinal dynamics of HIV-1 persistence and immunological response.**

To assess the possible influence of the ART regimen on the plasma viral load, HIV persistence markers, and immune restoration, we stratified the participants according to the ART class (Fig. S4A) or nucleoside analogue reverse transcriptase inhibitor (NRTI) backbone (Fig. S4B) they received at every time point of early ART. No differences for any marker have been observed between different NRTI backbones. However, plasma viral load and total HIV-1 DNA levels were higher at 12 weeks ART in the participants treated with the triple-class, four-drug ART regimen that included a nonnucleoside reverse transcriptase inhibitor (NNRTI) combined with a protease inhibitor (PI), compared to those treated with the three-drug NNRTI-based regimen (Fig. S4A). The reason for this difference is that, by the study protocol, almost all participants started the early ART on the four-drug regimen (NNRTI + PI) and were advised to discontinue the fourth drug when their plasma viral load became undetectable^20,27. Consequently, at 12 weeks ART, most of participants whose plasma viral loads were not yet suppressed (and who, therefore, had higher total HIV-1 DNA levels) were still on the four-drug regimen, while most of those with suppressed viral loads have already discontinued the fourth drug. No other differences by the ART class were observed for any marker.

At baseline and at every time point during early ART, we assessed pairwise correlations between the measured parameters (Fig. S5A). Positive correlations were observed between the virological markers, as well as between the CD4+ count and CD4/CD8 ratio, whereas the virological markers negatively correlated with the CD4+ count and CD4/CD8 ratio. At baseline, correlations were generally stronger than during ART. Both at baseline and under ART, virological markers correlated stronger with the CD4/CD8 ratio than with the CD4+ count. Strong positive correlations were also observed for CD4+ count, CD4/CD8 ratio, and total DNA between the time points, whereas these correlations were weaker for US RNA and especially for the US RNA/total DNA ratio (Fig. S5B). These differences reflect more stable dynamics of total DNA under ART as compared to US RNA, which is known to fluctuate in time^33,34.

HIV-1 persistence and immunological response dynamics during CHI ART

According to the trial protocol, participants had to (re)start ART in the case of two consecutive CD4+ counts <350 cells/mm³, diagnosis of symptomatic HIV-1 disease, or insistence on (re)starting ART by the physician or participant. 63 out of 64 study participants (re)started ART during CHI (one participant from the 60-week arm chose not to interrupt early ART). Baseline and CHI ART characteristics are shown in Table 2. Median times without treatment between the interruption of early ART (for the 60-week and 24-week arms) or randomization (for the no-treatment arm) and the start of CHI ART were 115.6 (IQR, 50.6–200.6) weeks for all participants, 134.3 (93.7–231.1) weeks for the 24-week arm, 123.4 (51.0–199.5) weeks for the 60-week arm, and 83.4 (33.2–113.4) weeks for the no-treatment arm. The time without treatment was significantly shorter for the no-treatment arm compared to the 24-week arm (Fig. S6), confirming the earlier results of Grijsen et al. ²⁰. During CHI ART, the vast majority of participants were receiving tenofovir/emtricitabine as an NRTI backbone and an NNRTI or a ritonavir-boosted PI as a third drug, although a small minority were receiving an integrase strand transfer inhibitor (INSTI)-based regimen (Table 2). No differences by the ART class were observed for any marker during CHI ART (Fig. S7A) and neither NRTI backbone nor ART class was significantly different between the study arms (Fig. S7B).

Table 2 Baseline and treatment characteristics of the study participants (CHI ART)

We longitudinally quantified the same virological and immunological parameters during CHI ART as described above for the early ART period. Plasma viral loads, CD4+ counts, and CD4/CD8 ratios were measured in all 63 participants treated with CHI ART. US RNA and total DNA were measured in 39 participants, for whom longitudinal PBMC samples were available (16 participants from the 60-week arm, 12 from the 24-week arm, and 11 from the no-treatment arm). Figure 2b shows the levels of the measured parameters at baseline (the time point before CHI ART initiation) and during the first 96 weeks of CHI ART, while Fig. 3b shows the longitudinal dynamics of these parameters fitted to a two-phase segmentation model. Fig. S2B shows the significance levels of the differences between the time points for all markers. In general, the longitudinal dynamics were similar to that observed during early ART, with all parameters fitting to a two-phase model. However, some differences were observed. In contrast to the early ART, where both CD4+ count and CD4/CD8 ratio remained stable after the initial increase, at CHI ART both these parameters continued to increase during the second phase as well, albeit much more slowly than during the first phase. Fig. S3B shows the relative increases from baseline of CD4+ counts and CD4/CD8 ratios during CHI ART. At CHI ART, the US RNA/total DNA ratio remained stable after the initial decrease, whereas an increase in this ratio was observed during the second phase at the early ART.

We then assessed the longitudinal decay dynamics of HIV-1 markers (plasma viral load, total DNA, and US RNA) by comparing their log-transformed relative changes from baseline at early and CHI ART. At all time points, the change in total DNA was the smallest, followed by the change in US RNA, while the change in plasma viral load was the most prominent (Fig. S8). Pairwise comparisons between the relative changes of different markers from baseline revealed significant differences, both during early and CHI ART. When visually comparing early and CHI ART, the decay dynamics of total DNA and US RNA were similar, whereas plasma viral load decreased by a larger magnitude on early ART than on CHI ART (Fig. S8). This reflects higher plasma viral loads at PHI (the baseline of early ART) than at CHI (Fig. 2).

Finally, we assessed pairwise correlations between the measured parameters at baseline and at every time point during CHI ART (Fig. S9A). Similarly to the early ART, we observed positive correlations between the virological markers, as well as between the CD4+ count and CD4/CD8 ratio, and negative correlations between virological markers and CD4+ count or CD4/CD8 ratio. However, the correlations at CHI and during CHI ART were weaker than the corresponding correlations at PHI and during early ART (Figs. S5A and S9A). As for early ART, positive correlations were observed for CD4+ count, CD4/CD8 ratio, and total DNA between the time points, whereas these correlations were weaker for US RNA and the US RNA/total DNA ratio (Fig. S9B).

Comparisons between early and CHI ART in different participants

We then proceeded to compare the CD4+ count, CD4/CD8 ratio, and HIV-1 persistence markers between early and CHI ART. First, we compared these parameters in different participants: parameters measured during early ART were compared with the corresponding parameters measured during the first 60 weeks of CHI ART in the no-treatment arm (Fig. 4). Separate comparisons were performed for baseline (PHI vs. CHI) and on-ART values. For the latter comparisons, we used mixed-effects modelling to account for the correlations of longitudinal measurements within participants. The CD4+ count was significantly higher both at PHI than at CHI and during early ART than during CHI ART. In contrast, no difference between PHI and CHI was observed for the CD4/CD8 ratio at baseline but this ratio was significantly higher during early ART than during CHI ART. Accordingly, no difference between early and CHI ART was observed in the relative increases of the CD4+ count from baseline, but for the CD4/CD8 ratio this difference was significant, with a more prominent increase during early ART (Fig. S10A). For both total DNA and US RNA, we observed nonsignificant trends towards lower values at PHI compared to CHI at baseline, and both these markers were significantly lower during early than during CHI ART (Fig. 4).

Comparisons of immunological response to early and CHI ART in the same participants

Next, we compared the CD4+ count and CD4/CD8 ratio during early and subsequent CHI ART periods in the same participants. As above, the CD4+ count was higher both at PHI than at CHI and during early ART than during CHI ART (Fig. 5a). The CD4/CD8 ratio was slightly higher at PHI than at CHI, but the difference between CD4/CD8 ratios during early and CHI ART periods was much larger than at baseline (Fig. 5a). Relative increases of the CD4+ count and CD4/CD8 ratio from baseline confirmed this, with a significant difference between early and CHI ART for the CD4/CD8 ratio but not for the CD4+ count (Fig. S10B). The difference in the rates of CD4/CD8 ratio restoration was the most prominent during the first 12 weeks of ART, as the CD4/CD8 ratio slopes during early ART were significantly steeper than those during CHI ART in the same participants in the first 12 weeks of ART (Fig. 5b) but not thereafter (Table S1).

We then compared the immunological response to early and CHI ART in the same participants. Normalization of the CD4/CD8 ratio to more than 1 is considered an important measure of immunological response to ART, and a low CD4/CD8 ratio is a prognostic marker for both opportunistic infections and non-AIDS morbidity and mortality^35,36. Therefore, we compared the normalization of the CD4/CD8 ratio in the first year of early and CHI ART in the same 19 participants, for whom the CD4/CD8 ratio measurements at baseline and 48 weeks of both early and CHI ART were available (only the 60-week arm was included in this analysis). No difference between PHI and CHI was observed in the proportions of participants with the normalized CD4/CD8 ratio at baseline, as this ratio was <1 in 18 of 19 participants at PHI and all 19 participants at CHI (Fig. 5c). However, by week 48 of early ART, 12 of 19 participants (63%) achieved the CD4/CD8 ratio of >1, whereas this threshold was achieved by only 3 of the same 19 participants (16%) by week 48 of CHI ART (p = 0.0069) (Fig. 5c). Thus, in the same PWH, early ART was superior to the subsequent CHI ART in restoration of the CD4/CD8 ratio.

We then compared functional HIV-specific T-cell responses between early and CHI ART periods in the same participants. To this end, we selected 22 participants, for whom paired PBMC samples obtained at early and CHI ART with similar times from ART initiation were available (median (IQR) difference in times on ART, 4.9 (1.8–11.6) weeks) and performed cross-sectional pairwise comparisons of CD4+ and CD8 + T-cell proliferative responses, reactivity (activation-induced marker (AIM) assay), and IFN-γ release upon stimulation with Gag peptide pools. Robust CD4+ and CD8 + T-cell responses were observed at both early and CHI ART. A significant difference between the ART periods was observed for Gag-specific CD8 + T-cell reactivity that was higher during CHI ART, accompanied by a nonsignificant trend towards higher IFN-γ release during CHI ART (Fig. 5d).

Finally, we compared levels of plasma soluble markers of systemic inflammation (IL-6), intestinal damage (I-FABP), and monocyte activation (sCD163 and sCD14) between early and CHI ART periods in 41 participants with paired plasma samples from early and CHI ART periods (median (IQR) difference in times on ART, 6.0 (2.9–11.1) weeks). Levels of IL-6 and I-FABP were significantly higher, while both monocyte activation markers were significantly lower, on early ART compared to CHI ART in the same participants (Fig. 5e).

Comparisons of HIV-1 persistence markers between early and CHI ART in the same participants

Next, we compared total DNA and US RNA during early and subsequent CHI ART periods in the same participants (n = 28). Remarkably, in contrast to the differences observed above for these markers between early ART and CHI ART in the no-treatment arm, we measured no significant differences for any of these markers in the same participants (Fig. 6a). We then performed a sensitivity analysis, in which, for each time point, we only retained paired early and CHI ART measurements of CD4+ count, CD4/CD8 ratio, total DNA, and US RNA (Fig. S11). This analysis did not change the conclusions as the results were very similar to those shown in Figs. 5a and 6a. Furthermore, we observed strong correlations between the average levels of total DNA or US RNA measured during early and CHI ART (Fig. 6b). These results demonstrate that, upon reinitiating ART during CHI, PWH who have been temporarily pre-treated during PHI achieve similar levels of HIV-1 persistence markers to those achieved during early ART.

We then asked whether HIV-1 sequence diversity differs between early and CHI ART in the same participants. To determine this, we used single-genome sequencing (SGS) to measure the nucleotide diversity of the 1.4 kb p6 – protease (PR) – reverse transcriptase (RT) region^37,38 in the total HIV-1 DNA isolated from paired PBMC samples of 8 participants obtained at the same time points during early and CHI ART (12 weeks ART, n = 6; 24 weeks ART, n = 2). HIV-1 diversity, calculated as average pairwise distance (APD), was significantly higher during CHI ART than during early ART (p = 0.023) (Fig. 6c). This difference likely reflects the (on average) 2.5-year period without treatment between the early and CHI ART, during which HIV-1 nucleotide substitutions are expected to accumulate due to the unsuppressed virus replication³⁹.

To obtain a deeper insight into possible differences in the persistence of HIV-1 reservoir between early and CHI ART in the same participants, we used intact proviral DNA assay (IPDA) to perform a cross-sectional comparison of levels of intact and defective proviruses in 22 participants with paired samples obtained at early and CHI ART, in whom we also measured T-cell responses (see above). Similarly to what was observed for US RNA and total DNA, no significant difference in the levels of intact proviruses between the therapy periods was measured in the same participants (Fig. 6d). However, nonsignificant trends towards higher levels at CHI ART were observed for both 3’ defective and 5’ defective proviruses, and significantly higher levels of total defective (3’ defective + 5’ defective) proviruses were observed at CHI ART (Fig. 6d). In spite of this difference, total (intact + defective) proviruses did not significantly differ between the two therapy periods (Fig. 6d). We also observed strong positive correlations between the two therapy periods for all proviral forms measured: intact, 3’ defective, and 5’ defective proviruses (Fig. 6d). Taken together, these results strongly suggest that in these participants, upon reinitiation of ART during CHI, intact HIV-1 reservoirs returned to their early-ART levels despite the long TI. At the same time, an increase in defective proviruses at CHI ART compared to early ART was observed, which may be explained by clonal expansion of cells harbouring defective proviruses⁴⁰.

Finally, we determined correlations between the HIV-1 proviral forms and HIV-specific T-cell responses at both early and CHI ART. At early ART, intact HIV-1 DNA weakly negatively correlated with CD8 + T-cell reactivity and IFN-γ release but a correlation of 5’ defective HIV-1 DNA with IFN-γ release was stronger (Fig. S12). However, at CHI ART, intact HIV-1 DNA significantly negatively correlated with CD8 + T-cell proliferation, reactivity, and IFN-γ release (Fig. 6e), while correlations of defective HIV-1 DNA with T-cell responses were weaker (Fig. S12). These results suggest that HIV-specific T-cell responses may restrict the intact HIV-1 reservoir during ART, especially during ART initiated at CHI.

Comparisons between pre-treated and not pre-treated participants during CHI ART

Taken together, the above results suggested that temporary early ART had a suppressive effect on the HIV-1 persistence markers but did not affect the CD4+ count or CD4/CD8 ratio measured during ART restarted during CHI. To prove this, we compared the parameters during the first 96 weeks of CHI ART between participants who were or were not pre-treated with early ART. Indeed, while no difference was observed for the CD4+ count or CD4/CD8 ratio, both US RNA and total DNA were significantly lower during CHI ART in those participants who had been pre-treated with temporary early ART (Fig. 7). This indicates that temporary early ART had a long-term suppressive effect on the viral persistence, as revealed during therapy reinitiated after several years. We also observed lower levels of US RNA (but not total DNA) at the CHI ART baseline in the pre-treated participants (Fig. 7). These results are in line with the earlier report of Grijsen et al., who measured a lower plasma viral load set point in those Primo-SHM study participants who were treated with early ART²⁰. Taken together, this evidence indicates that temporary early ART caused a reduction in the levels of both HIV-1 replication and persistence, observed at CHI and during CHI ART, respectively.

To elucidate any possible long-term effects of the duration of early ART, we compared the parameters during CHI ART between participants who had been pre-treated for 24 weeks, 60 weeks, or not at all, during PHI (Fig. S13). No differences between arms were observed for CD4+ count. The CD4/CD8 ratio during CHI ART was significantly higher in the 24-week arm compared to the 60-week arm but no differences with the no-treatment arm were observed. In contrast, US RNA was significantly lower in both 24-week and 60-week arms, compared to the no-treatment arm, while no difference between 24-week and 60-week arms was observed. Total DNA was significantly lower only in the 60-week arm compared to the no-treatment arm.

Finally, we asked whether the long-term effects of early ART on the viral persistence during CHI ART could be mediated through better antiviral immune responses in pre-treated participants. To this end, we selected 40 participants, for whom PBMC samples at CHI ART were available (60-week arm, n = 16; 24-week arm, n = 14; no-treatment arm, n = 10) and performed cross-sectional comparisons of functional HIV-specific T-cell responses (CD4+ and CD8 + T-cell proliferation and reactivity, as well as IFN-γ release) between the arms. Although it was not possible to perfectly match the participants for times from CHI ART initiation, there were no significant differences in times on ART between the arms (Fig. S14A). Robust CD4+ and CD8+ responses were observed in all arms but no significant differences were observed between the arms for any of the measured HIV-specific T-cell response parameters (Fig. S14B). We also measured soluble markers of systemic inflammation, intestinal damage, and monocyte activation in 53 participants (60-week arm, n = 23; 24-week arm, n = 18; no-treatment arm, n = 12) at CHI ART. No significant differences between the arms were observed for any of these markers as well (Fig. S14C).

Source link