**4. Discussion**

This study shows that CCR5-tropic and CXCR4-tropic strains exhibit different kinetics of replication in MDM and highlights the capability of CXCR4-tropic strains to promote the apoptosis of this important HIV-1 reservoir. HIV-1 reservoirs represent so far the major obstacle for achieving HIV cure.

Our findings are in line with a previous study showing that differences between CCR5-tropic strains and CXCR4 strains in productive infection of MDM occurred during the early stages of HIV-1 life cycle and in particles at levels of reverse transcription and nuclear translocation of viral genomes [32]. Though our study did not consider the phase22s of HIV life cycle, we investigated a relationship between different levels in viral production and MDM homeostasis according to co-receptor usage.

We evaluated different kinetics of replication in MDM of CXCR4-tropic and CCR5-tropic molecular clones, respectively, NL4-3 and 81A, differing only in env variable domains. After a starting boost, the replication of CXCR4-tropic clones in MDM subsequently diminished reaching a status of abortive infection, while the replication of CCR5-tropic clones tended to increase, reaching a plateau after 10 days of infection.

It is important to stress that NL4-3 did not affect HIV-1 productive infection up to the seventh day in MDM, suggesting that the clearance of the CXCR4 strain may not be due to a failure in the entry or in other preintegrational phases (Figure 1), but may be the result of the killing of the host cells during the onset of infection.

These results underline the tendency towards an in vitro disappearance of the most aggressive CXCR4- tropic virus in the course of the HIV-1 infection and the survival of CCR5- tropic strain infected MDM reservoirs as key determinant of HIV-1 persistence in this cellular reservoir.

This evidence provided us the clue to analyze how coreceptor usage may differently modulate MDM homeostasis and particularly apoptosis.

In particular, the MAPK p38 plays a pivotal role in the transmission of signals from cell surface receptors to the nucleus. It is activated by diverse extracellular stimuli that regulate important cellular processes including response to stress factors in many cell types [66]. Our results show a transient but marked induction of the phosphorylated form of the p38 MAPK at 30 after exposure to CXCR4-tropic, but not CCR5-tropic HIV strains in MDM. The role of activation of MAPKs p38 in programmed

death of MDM and T-cells due to CXCR4-tropic strain infection still remains controversial: whereas a role in HIV pathogenicity is already demonstrated [50,67], some studies report no association in Caspase-dependent apoptosis [68] moreover, the p38 activation pathway, in cell reservoirs such as MDM, was attributed to β chemokine secretion rather than apoptosis [49,69]. In this last case, this disagreement with our results in HIV mediated signalling may be attributed to a di fferent experimental approach as we used the whole pure viruses and not the recombinant gp120 and did not consider any serum starvation for exposure of MDM to the di fferent strains, in order to avoid them excessive stress and permit the primary cells to reproduce under more natural physiological conditions.

The role of p38 has been elucidated in the setting of infection of T cells by CXCR4-tropic strains [70,71]. In particular, the role of replication of HIV-1 in human T lymphocytes requires the activation of host cellular proteins [72]. Previous studies have identified p38 mitogen-activated protein kinase (MAPK) as a kinase necessary for HIV-1 replication in T cells [73–76]. Among them, Cohen et al. 1997 have shown that HIV-1 CXCR4 strain infection of both primary human T lymphocytes and T cell lines immediately stimulates the cellular p38 MAPK pathway, which remains activated throughout the experimental conditions. Inclusion of an antisense oligonucleotides to p38 MAPK expressly inhibited viral replication [70,77–79]. Blockade of p38 MAPK by addition of CNI-1493 also inhibited HIV-1 viral replication of primary T lymphocytes in a dose- and time-dependent manner. Stimulation of p38 MAPK activation did not occur with the addition of heat-inactivated virus, suggesting that viral internalization, and not just membrane binding, is necessary for p38 MAPK activation [80,81]. The results of this work show that activation of the p38 MAPK cascade is critical and essential for HIV-1 replication in T cells [81,82].

In consideration of key determinants of HIV persistence in MDM reservoirs, post translation changes of cell and nuclear targets is one of the upstream events due to viral exposure, culminating in absence of cytolitic e ffects. Macrophages provide an ideal environment for the formation of viral reservoirs since they live long and are widely distributed throughout the body [83].

As microarray analysis showed, 10 genes related to the apoptosis pathway were up-regulated in NL4-3 infected MDM compared to 81A infected ones and genes related to the apoptosis pathway, such as Defender against cell death 1(DAD-1) [84] and Cullin 2 (hCUL2), were up-regulated in 81A infected compared to NL4-3 infected MDM.

Our studies demonstrate the up-regulation of genes included the polypeptide chain-releasing factor GSPT1/eRF3 protein, which in the processed form has been shown to promote caspase activation, IAP (inhibitors of apoptosis) ubiquitination and apoptosis [53], Caspase 7 (CASP 7), an apoptosis-related [69] and Cytochrome C whose release has been shown in HIV dependent apoptosis [50]. The activation of such genes related to apoptosis in NL4-3 infected MDM can then have downstream e ffects being responsible for the progressive decrease of p24 production of the CXCR4-using NL4.3.

Our results would not seem in line with in vivo evidence of the emergence of the more aggressive sincytium-inducing (SI) CXCR4- tropic strains in the terminal phases of HIV-1 disease associated with rapid decline of CD4+ and CD8+ T cells [28,85] but this phenomenon almost represents an e ffect of the breakdown of the immune system and the onset of AIDS [86]. We speculate that CXCR4- tropic strains play a minor role in disease progression because dying CXCR4 virus infected reservoirs, cannot provide virus nor continue to directly contribute to the depletion of immune cell system. Indeed, CCR5-using strains are associated with a lower percentage of cell death, suggesting the capability of these strains to promote cell survival as supported also by transcriptome analysis.

On the other hand, there are many reasons to consider a role of CCR5 viruses and their host cells as target for therapeutic strategies: (i) the protective role of the 32-nucleotides ( Δ32) deletion in CCR5 gene in homozygous condition against HIV-1 infection and the more benign pattern of disease progression associated with the deletion in one of the two alleles [87,88] (ii) a logarithmic correlation between CCR5 expression and viremia in patients with disease progression [89–91] (iii) the importance of CCR5-tropic isolates for dissemination outside peripheral blood in compartments considered as "sanctuaries" like the Central Nervous System where macrophages represent more than 90% of the

HIV-1 infected cells [92–95] the capacity of R5 isolates harbored in macrophagic reservoirs to provoke the immune anergy through host-related factors (bystander e ffect) and the emergence of more virulent SI variants and the subsequent AIDS progression (iv) increase of viremia in later stages of HIV disease caused by macrophages during opportunistic infections [96,97] (v) increase of both CCR5 expression on CD4+ T cells and the frequency of memory CD4 T-cells (the target cells of CCR5 virus variants) over the course of infection [97,98].
