Taking all these results into

Taking all these results into account, we propose a model for the evolution of HIV DNA forms in blood during the natural history of HIV infection (Fig. 4a). During PHI, the labile linear and episomal HIV DNA forms are particularly abundant and mostly constitute evidence of active replication in recently infected activated RU 58668 [17]. This observation might be explained by the simultaneous cytokine storm and high immune activation detected at this stage. Later during the first year following infection, while the infection is evolving towards a steady state, integrated HIV DNA becomes the major component of the total HIV DNA in most patients. Both HIV DNA markers continue to increase over the subsequent years of follow-up (Fig. 4a). The expansion of HIV reservoirs can be linked to several factors: the infection of new cells due to continuous viral replication, the persistence of long-lived infected cells, and the proliferation of infected cells [[33], [34], [35]]. The last two factors might explain the increasing contribution of integrated HIV DNA forms, which are transmitted from a mother cell to all daughter cells, while unintegrated forms are diluted during cell division. The models in Fig. 4, depicting the evolution of the different viral forms over time, propose an explanation for why very early treatment is highly beneficial. Such early treatment has been proven to be more effective in reducing total HIV DNA levels and restoring immune functions [11,12,36,37]. The rapid and steep initial decline of total HIV DNA (Fig. 4b) could be explained mainly by the elimination of the labile unintegrated forms. Early cART initiation can also limit the establishment of persistent integrated HIV DNA forms and prevent the progressive increase of blood reservoirs observed in the natural history of the infection. Moreover, previous studies have shown that HIV-infected blood cells during PHI are mostly short-lived T-cells [38], which is an additional factor contributing to reservoir instability during this stage. Treating HIV-infected individuals as soon as the PHI stage allows partial preservation of the most long-lived memory T-cells from infection and the maintenance of a predominant contribution of the short-lived ones [[36], [37], [38], [39], [40]]. The slower but continued decrease in total HIV DNA at later times following cART initiation might be explained by the death of these short-lived infected cells [11,12]. In contrast, in patients treated during the chronic stage (Fig. 4c), total HIV DNA decreases only shortly after cART initiation, due to the elimination of infected activated cells [41]. Integrated HIV DNA, which is highly predominant in latently infected cells, persists at high levels thereafter. Long-lived highly proliferative central-memory T-cells have been shown to be the main contributors to HIV blood reservoirs during the chronic stage and in patients treated since the chronic stage, contributing to more stable reservoirs [35,40]. Taken together, these results suggest that during PHI, the existing reservoirs, which are mainly composed of short-lived cells infected with unstable viral forms, are easily eliminated through the early initiation of cART. In contrast, in the chronic phase, the blood reservoirs are mainly composed of stable proviruses in long-lived quiescent cells and are unaffected by cART. This could explain the greater impact of treatment when initiated during PHI compared to the chronic phase, which is an additional reason for recommending the treatment of all HIV-infected individuals, including the initiation of treatment as soon as possible in those diagnosed during PHI. We acknowledge that in the present study we quantified total and integrated HIV DNA on PBMC and not CD4+ T cells. Nevertheless, in a previous study on patients from the same ANRS-SEROCO cohort, we showed a high correlation of total HIV DNA levels expressed either by copies/106 PBMC, or copies/106 CD4+ T-cells or copies/mL of whole blood [42]. The predictive value level on disease progression was similar whatever the expression of the total HIV DNA load. Besides, we only investigated blood reservoirs. Despite the continuous circulation of lymphocytes between the blood and tissues, the dynamics of integrated HIV DNA and total HIV DNA in infected CD4+ T-cell subsets and their relative contributions could differ in lymphoid tissues because of different inflammation and cellular activation levels, which should therefore also be explored.