P2X7 Receptor Inhibition Improves CD34 T-Cell Differentiation in HIV-Infected Immunological Nonresponders on c-ART

Peripheral CD4+ T-cell levels are not fully restored in a significant proportion of HIV+ individuals displaying long-term viral suppression on c-ART. These immunological nonresponders (INRs) have a higher risk of developing AIDS and non-AIDS events and a lower life expectancy than the general population, but the underlying mechanisms are not fully understood. We used an in vitro system to analyze the T- and B-cell potential of CD34+ hematopoietic progenitor cells. Comparisons of INRs with matched HIV+ patients with high CD4+ T-cell counts (immune responders (IRs)) revealed an impairment of the generation of T-cell progenitors, but not of B-cell progenitors, in INRs. This impairment resulted in the presence of smaller numbers of recent thymic emigrants (RTE) in the blood and lower peripheral CD4+ T-cell counts. We investigated the molecular pathways involved in lymphopoiesis, focusing particularly on T-cell fate specification (Notch pathway), survival (IL7R-IL7 axis) and death (Fas, P2X7, CD39/CD73). P2X7 expression was abnormally strong and there was no CD73 mRNA in the CD34+ cells of INRs, highlighting a role for the ATP pathway. This was confirmed by the demonstration that in vitro inhibition of the P2X7-mediated pathway restored the T-cell potential of CD34+ cells from INRs. Moreover, transcriptomic analysis revealed major differences in cell survival and death pathways between CD34+ cells from INRs and those from IRs. These findings pave the way for the use of complementary immunotherapies, such as P2X7 antagonists, to restore T-cell lymphopoiesis in INRs.


Introduction
Combined antiretroviral treatment (c-ART) has greatly improved the outcome of HIV infection. The key objective of c-ART is to suppress viral replication and to induce the production of sufficient numbers of CD4+ T cells to prevent AIDS-defining (CD4+ T-cell counts below 200 cells/mm 3 ), and non-AIDS-defining (CD4+ T-cell counts below 500 cells/mm 3 ) severe events [1]. Immunological failure is defined as an inability to reach these levels of CD4+ T cells on c-ART (200 or 500 cells/mm 3 , depending on the type of event considered). In large cohort of patients displaying viral suppression, immunological success seemed to be largely timedependent, as the number of CD4+ T cells seemed to increase steadily, even after seven years [2]. CD4+ T-cell restoration may be hindered by mechanisms related to HIV infection and its consequences, or modulated by host factors, both of which may affect T-cell homeostasis in the periphery or through effects on T-cell production. Demographic factors (age, sex, ethnic group [3][4][5]) affect CD4+ T-cell levels and, thus, immune restoration. The characteristics of HIV infection in the patient (CD4+ T-cell nadir, peak viral load, duration of infection and viral control on c-ART [4,[6][7][8]) are also key determinants of CD4+ T-cell recovery. Increases in immune activation [9,10] and inflammation [11,12] are currently considered to be the principal mechanisms underlying poor immunological responses on c-ART. Such alterations affect the homeostasis of the T-cell pool, modifying both peripheral and thymic T-cell levels [13]. Specific host genetic factors, including polymorphisms of genes of the inflammation/apoptosis pathway [14] or genes involved in T-cell development, such as IL7R [15], are also associated with poor CD4+ T-cell recovery.
In this study, we observed a specific decrease in the T-cell potential of circulating CD34+ progenitors from patients displaying virological suppression on long-term c-ART but with poor CD4 + T-cell restoration. We also showed that CD34+ cells from INRs were extremely sensitive to the extracellular ATP pathway, as inhibition of the ATP receptor, P2X7, restored T-cell differentiation.

Enrollment and characteristics of the patients
We selected, from our cohort of HIV+ patients, those with poor immunological CD4+ T-cell restoration (i.e. CD4+ T-cell count <500/mm 3 and a CD4/CD8 ratio <1) and a plasma viral load that had remained below the detection threshold for more than eight years. These patients are referred to hereafter as "immunological non-responders" (INRs). Patients with high levels of immunological CD4+ T-cell recovery (i.e. with values close to those of the general population of uninfected individuals: >900 CD4+ T cells/mm 3 and a CD4/CD8 ratio >1 [40]), referred to hereafter as "immunological responders" (IR), were selected and matched with INRs for factors predictive of immune recovery on c-ART, including estimated date of infection, treatment duration, periods with a sustained undetectable viral load (i.e. <50 HIV-1 RNA copies/mL), CD4+ cell nadir and pretreatment CD4+ counts (Table 1). Uninfected control individuals were matched with HIV-positive patients for age. Median (IQR) absolute CD4+ T-
The impairment of T-cell differentiation from CD34+ cells in INRs is not associated with altered responses to Notch or a particular genotype of IL7RA The IL7/IL7R and Notch pathways are the two principal pathways of T-cell differentiation [43][44][45][46][47]. We therefore investigated whether perturbations of these two pathways could account for the lower T-cell potential of CD34+ cells in INRs. We assessed the prevalence in our patients of SNPs of the IL7R gene associated with decreases in T-cell production: SNPs present in the promoter region (rs7701176), exon 6 (rs6897932), intron 6 (rs987106) and the 3' region (rs10491434) [15] (Fig 3A). There was no clear difference in the distribution of polymorphisms between the two groups of HIV-infected patients. Moreover, soluble IL7RA (S) and membrane-bound IL7RA  We assessed the functionality of the Notch1 receptor by analyzing the expression of Notch target genes in CD34+ cells after incubation with a recombinant Notch ligand, Delta-like 4 (hDLL4-Fc), in the presence of IL-7 [45]. HES1 mRNA levels increased rapidly after incubation with DLL4 alone or together with IL-7, in cells from HIV-negative and HIV-infected individuals (P<0.05; Fig 3B). However, no difference in the expression of HES1 was observed between IRs and INRs. These results suggest that alterations in Notch signaling and differences in IL7RA genetic background cannot account for the impaired lymphopoiesis in INRs.

P2X7 is strongly expressed on CD34+ cells in INRs and its inhibition restores the potential for T-cell differentiation
We then asked whether the impairment of CD34+ cell differentiation into T cells resulted from changes in death pathways due to persistent immune activation in INRs. We observed no difference in the FAS expression of CD34+ cells between HIV-uninfected subjects and HIVinfected IRs and INRs (Fig 5A, P = NS). Extracellular nucleotides and purinergic receptors modulate CD34+ cell homeostasis [49][50][51][52][53]. Among P2 family receptors ex vivo purified peripheral CD34+ express P2X1, P2X4 and P2X7 [54]. It is well documented that the binding of ATP to its receptor, P2X7, induces the assembly of a cytoplasmic multipartner complex, the inflammasome, and caspase-1 activation, leading to secretion of the proinflammatory cytokines IL1ß and IL-18 [55,56]. We observed that P2X7 was markedly more strongly expressed in INRs than in IRs and HIV-uninfected subjects (P<0.05; Fig 5B). Extracellular ATP may also be hydrolyzed by the ectoenzymes CD39/CD73 [57][58][59][60][61][62]. Particularly CD73 expression was undetectable in all the INRs studied (P<0.05; Fig 5C).
We investigated the role of P2X7 in poor immune restoration in HIV patients, in LDAs involving irreversible antagonism with PPAD. We showed that P2X7 inhibition significantly improved the potential of CD34+ cells from INRs to differentiate into T cells

Microarray analysis reveals a downregulation of cell survival pathways and an upregulation of apoptosis in CD34+ cells from INRs
We explored the mechanisms underlying the functional alterations of CD34+ cells in INRs, by performing transcriptomic analysis in ex vivo-purified CD34+ cells from IRs and INRs ( Fig  6A). The gene expression profiles of these two groups were very similar, with only 210 genes differentially expressed. These genes were grouped by biological function, and the top five were identified on the basis of activation z-score (Fig 6B). This score predicts the activated (positive z-score) or inactivated (negative z-score) state of genes from the same functional group. Cells from IR patients displayed an upregulation of genes from the following categories

Discussion
This study provides insight into the major mechanisms driving immune recovery in HIVinfected patients with long-term virological success on c-ART regimens. The design of this study differs from that of most other studies of poor immune recovery. First, we selected patients with opposite and extreme immunological profiles. Second, the patients enrolled in this study had been treated for at least eight years, and for up to 16 years in some cases, whereas most previous studies were carried out two to four years after c-ART initiation [4,40,[63][64][65][66], with only a few investigating immune recovery mechanisms after more than five years of treatment [6,[67][68][69]. Finally, we applied strict criteria for patient selection, and most of the parameters predictive of immune recovery described in previous studies were characterized in our patients (age [3][4][5], sex [5,70], and ethnic origin [70], durations of infection and treatment, duration of time for which the virus was undetectable, nadir and pre-therapy CD4+ counts [4,[6][7][8]).
Low levels of de novo lymphocyte production and peripheral dysfunction are considered to be major barriers to efficient immune restoration in HIV+ individuals [66,[71][72][73][74][75]. We showed, by assessing CD4+ RTE frequency, that lymphopoiesis was severely compromised in INRs. Our results strongly suggest that T-cell recovery is influenced by CD34+ cell impairment. We We cannot exclude the possibility that CD34+ cells from INRs also display impaired differentiation into other hematopoietic lineages. However, we observed no differences in B-cell generation between groups. Furthermore, we found that the potential of CD34+ progenitors to generate T cells was strongly correlated with the degree of peripheral T-cell restoration, implying that the maintenance of adequate lymphocyte levels is highly dependent on efficient de novo T-cell lymphopoiesis.
Some authors have suggested that CD34+ cells or more differentiated colonies may be directly infected [26, 28, 30, 36], regardless of the population studied, but there is little evidence to support this view. Several viral proteins (gp120 [76] or Nef [77]) have been shown to impair CD34+ differentiation. We quantified HIV RNA in culture supernatants, but detected no ongoing viral replication. We cannot rule out the possibility of latent infection, but the incidence of such infection would be too low to account for such specific changes in T-cell differentiation.
We investigated the molecular abnormalities underlying impaired lymphopoiesis, focusing, in particular, on the key factors for T-cell lymphopoiesis, Notch and IL7R. Some IL7RA SNPs have been reported to be associated with impaired immune recovery [14,78,79], but the prevalence of these SNPs was not higher in INRs than in other subjects, and no difference in the mRNA levels for soluble and membrane-bound IL7R were identified between groups. Target gene expression after a short period of Notch activation was similar in INRs and HIV-uninfected subjects. However, the timing of transcriptional programs downstream from Notch signaling is highly variable [80]. We cannot, therefore, exclude the possibility that some differences in the response to Notch ligands become visible only after longer periods of stimulation, although this would be difficult to test because survival issues affect mRNA quality at later time points in feeder cell-free culture systems.
Consistent with previous reports, we observed abnormal immune activation in INRs despite long-term c-ART. We also found that sCD14 levels remained high in treated HIV+ patients, as reported in a previous study [81]. Analyses of plasma samples collected from the HIV+ patients over the last few years showed that levels of soluble markers of inflammation remained stable and high over time. The observed inflammation would probably affect CD34+ cell survival. The Fas receptor did not seem to be involved in this process. We therefore investigated other cell death pathways involving ATP signaling. The availability of extracellular ATP is regulated by the CD39/CD73 ectoenzymes [57][58][59][60][61][62]. In the absence of ectonucleotidase activity, high ATP concentrations trigger the low-affinity P2X7 receptor to induce a massive release of proinflammatory cytokines, leading to cell death by pyroptosis [82]. We provide several lines of evidence suggesting that T-cell differentiation is impaired due to the extremely high sensitivity of CD34+ HPCs to extracellular nucleotides. CD34+ cells from INRs displayed P2X7 upregulation and no CD73 expression, suggesting a greater susceptibility of these cells than of those from other subjects to ATP-induced cell death. Consistent with this hypothesis, a P2X7 antagonist restored the T-cell differentiation of CD34+ cells in INRs. ATP has been reported to be involved in stem cell metabolism [49][50][51][52][53]. In murine hematopoietic and human neural progenitors, extracellular ATP causes rapid cell death and an increase in the frequency of apoptotic features [53,83]. It is spontaneously released in cultures of human mesenchymal stem cells, inhibiting cell proliferation, and it appears to be a key regulator of early lineage commitment [84,85]. Consistently, transcriptomic analysis revealed that, unlike the non-cycling cells of INRs, CD34+ cells from IRs underwent mitosis. The greater immune activation observed in the peripheral blood of INRs may also occur in the bone marrow. IL-6 and sCD14 are markers of the activation of monocytes/macrophages, the principal supporting cells of the bone marrow niche. LPS, through sCD14, may trigger TLR4-promoted cell death and initiate an auto-amplification loop in which dying cells release their contents, thereby creating a highly inflammatory environment for neighboring differentiating progenitors. In this setting, other P2X7-expressing cells may also contribute to the chronic inflammatory state. We observed no modification of the caspase-1 activation profile in hematopoietic progenitors ex vivo, suggesting that no pyroptosis was occurring. However, further studies are undoubtedly required to define the precise mechanism of CD34+ cell turnover in inflammatory conditions. It also remains unclear why B-cell differentiation is unaffected. The role of IL7 in this process may provide an explanation. B-cell lymphopoiesis is dependent on IL7 in mice, but not in humans [86][87][88]. We have shown that the transcription of P2X7 is increased by IL7 [89]. We did not measure plasma IL7 levels, but they are likely to be high in INRs, as either a cause or a consequence of the low CD4+ T-cell levels. We suggest that, at early stages of lymphopoiesis, IL7 potentiates extracellular nucleotide signaling, thereby inducing cell death and regulating stem cell metabolism. At late stages, the expansion of maturing progenitor populations is highly dependent on IL7 [45]. IL7 thus plays a dual role in T-cell development.
Finally, microarray comparisons of IRs and INRs revealed that CD34+ cells from INRs lacked transcripts associated with proliferation and survival. Transcript levels for PKP4 and SEPT11, both of which are involved in cytokinesis during cell division [90,91], and those for PRKCZ, an important target of PI3K signaling [92], DEF6, regulating lymphocyte survival [93], and GADD45B, involved in FOXO signaling, oxidative stress resistance and DNA repair [94], were also downregulated. Conversely, CD34+ cells from INRs displayed an upregulation of genes involved in cell death, such as RNASEL and BIRC2 (encoding cellular inhibitor of apoptosis, cIAP) [95,96]. Further characterization of the genes identified in microarray analyses should help us to determine which factors limit the differentiation into T cells of CD34 HPCs from INRs.
Our findings suggest that patients with limited immune recovery on c-ART could be given complementary treatments, such as anti-inflammatory compounds, and, more specifically, P2X7 inhibitors [97], which have been shown to have a good safety profile in clinical trials [98].
These findings provide compelling evidence that successful immune restoration in HIVinfected patients on c-ART involves the regeneration of new T cells from CD34+ cells. It may therefore be possible to identify potential targets for the enhancement of T-cell lymphopoiesis in patients with an incomplete restoration of CD4 T-cell counts on c-ART.

Patient samples
Peripheral blood samples were collected from HIV-negative healthy donors at the Centre regional de transfusion sanguine and from HIV+ patients on c-ART followed at Henri Mondor Hospital (Créteil, France). Ethics committee approval was obtained and the subjects included gave written informed consent, in accordance with the Helsinki Declaration, before the start of the study.

Quantification of HIV-1 DNA and inflammation factors
Cell-associated HIV DNA was quantified by ultrasensitive real-time PCR (Biosentric, Bandol, France), in total PBMCs, as previously described [41]. Plasma samples were used to determine the concentrations of IL6, CRP and sCD14 with Quantikine ELISA kits (R&D Systems, Minneapolis, MN), according to the manufacturer's protocol.

Molecular analysis of mRNA levels
Purified CD34+ cells were exposed overnight to IgG1-Fc or DLL4-Fc (5 μg/mL, generously provided by A.Sakano [99]) in T-cell medium with or without IL-7, as previously described [45][46][47]80]. RNA was isolated in TRIzol (Invitrogen), according to the standard procedure. RT-qPCR was carried out with the SuperScript VILO cDNA synthesis kit (Invitrogen) and Brilliant II SYBR Green Master Mix (Agilent, Santa Clara, CA), in standard conditions, on an MX3005P (Stratagene, La Jolla, CA). The primer sequences used have been described elsewhere [80].

Microarray analysis
RNA was extracted from ex vivo purified CD34+ cells with the RNeasy Micro Kit (Qiagen), and quantified on an ND-8000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). Its integrity was then checked on a 2100 BioAnalyzer (Agilent Technologies). In vitro transcription was performed on 60 ng of RNA (Ambion Illumina TotalPrep RNA Amplification Kits, Applied Biosystems/Ambion), as described elsewhere [100].

Statistical analysis
All statistical analyses were performed with GraphPad Prism software v6 (La Jolla, CA). We used nonparametric Mann-Whitney and Kruskal-Wallis tests to compare continuous variables between two and three groups, respectively. For paired groups, Wilcoxon tests were used. Discrete variables were compared in Fisher's exact tests. Differences were considered non-significant if P>0.05. Microarray data were analyzed by ViroScan3d (Lyon, France), as previously described [100]. Ingenuity pathway analysis (Qiagen) was also conducted, focusing on both canonical pathways and biological functions.