Excess BAFF Alters NR4As Expression Levels and Breg Function of Human Precursor-like Marginal Zone B-Cells in the Context of HIV-1 Infection

We have reported excess B-cell activating factor (BAFF) in the blood of HIV-infected progressors, which was concomitant with increased frequencies of precursor-like marginal zone (MZp) B-cells, early on and despite antiretroviral therapy (ART). In controls, MZp possess a strong B-cell regulatory (Breg) potential. They highly express IL-10, the orphan nuclear receptors (NR)4A1, NR4A2 and NR4A3, as well as the ectonucleotidases CD39 and CD73, all of which are associated with the regulation of inflammation. Furthermore, we have shown MZp regulatory function to involve CD83 signaling. To address the impact of HIV infection and excessive BAFF on MZp Breg capacities, we have performed transcriptomic analyses by RNA-seq of sorted MZp B-cells from the blood of HIV-infected progressors. The Breg profile and function of blood MZp B-cells from HIV-infected progressors were assessed by flow-cytometry and light microscopy high-content screening (HCS) analyses, respectively. We report significant downregulation of NR4A1, NR4A2, NR4A3 and CD83 gene transcripts in blood MZp B-cells from HIV-infected progressors when compared to controls. NR4A1, NR4A3 and CD83 protein expression levels and Breg function were also downregulated in blood MZp B-cells from HIV-infected progressors and not restored by ART. Moreover, we observe decreased expression levels of NR4A1, NR4A3, CD83 and IL-10 by blood and tonsillar MZp B-cells from controls following culture with excess BAFF, which significantly diminished their regulatory function. These findings, made on a limited number of individuals, suggest that excess BAFF contributes to the alteration of the Breg potential of MZp B-cells during HIV infection and possibly in other situations where BAFF is found in excess.


Introduction
Currently, over 37 million individuals are living with the human immunodeficiency virus (HIV) [1]. Although over 66% of these individuals are successfully treated with antiretroviral therapy (ART), inflammation persists [1,2]. Consistently, we have previously reported that expression levels of the tumor necrosis factor (TNF) family member B-cell Activating Factor (BAFF, TNFSF13B) are in excess in the blood of HIV-infected progressors from the Montreal Primary HIV Infection (PHI) cohort. This upregulation is already present in the acute phase of infection and is still observed 9-12 months post-ART [3]. Moreover, our recent studies suggest that excess BAFF perseveres in progressors on ART for more than 15 years of the Canadian HIV and Aging Cohort Study (CHACS) [4]. Excess BAFF is associated with viral factors, such as Nef, and non-viral factors, such as elements of microbial translocation, and correlates with hypergammaglobulinemia and dysregulation of the B-cell compartment [5][6][7]. BAFF is an important growth factor that helps shape the marginal zone (MZ) B-cell pool. As such, we found that excess BAFF is associated with increased frequencies of a CD19 + CD1c + CD21 lo IgM hi CD27 + CD10 + B-cell population, sharing features of transitional immature and MZ B-cells, we designated "Precursor-like" Marginal Zone (MZp) B-cells, in the blood of PHI progressors [3], and more recently in the blood of CHACS participants [4]. We have reported similar observations with HIV-Transgenic (Tg) mice [8], SIV-infected macaques [9] and with HIV-infected Beninese commercial sex workers [10,11], suggesting that excess BAFF and dysregulated MZp B-cells are reliable markers of inflammation and disease progression in the context of HIV.
BAFF binds to three receptors among B-cell subsets: TNF receptor superfamily 13C (TNFRSF13C, BAFFR), TNF receptor superfamily 13B (TNFRSF13B, TACI) and TNF receptor superfamily 17 (TNFRSF17, BCMA), the second of which is highly expressed by MZ populations [12]. MZ are first-line B-cells that respond quickly to blood-borne antigens via their polyreactive BcR (IG + CD79) and numerous Pattern Recognition Receptors (PRR) such as C-type lectins and Toll-like Receptors (TLR). MZ B-cells have been shown to classswitch their immunoglobulins (IG) from IgM to IgG or IgA in a BAFF-dependent manner via TACI [12]. Importantly, MZ B-cells were shown to bind to the envelope (Env) proteins gp120 and gp41 of HIV, via C-type lectins and TLR10, respectively [13,14]. Whether IG produced by MZ B-cells confer some level of protection or contribute to hyperglobulinemia and autoimmune manifestations in excessive BAFF contexts, such as with HIV, is currently under investigation.
NR4A transcription factors are involved in the regulation of the inflammatory response and their expression is increased upon signaling via the TcR (TR + CD3) or BcR [19], amongst others. They are important regulators of differentiation, proliferation and apoptosis of immune cells [20]. The importance of NR4A transcription factors in immune regulation is highlighted by their importance for Tregs, as NR4A transcription factors directly promote and maintain the expression of FOXP3 [17]. Also, increased expression of NR4A1 in myeloid cells is associated with a diminished T-cell activation profile [21]. Furthermore, NR4A1-3 knockout mice develop systemic autoimmune diseases [20,22].
Given the importance of BAFF for MZ B-cell activity, we have herein assessed the impact of the HIV infection and excess BAFF on the Breg profile and function of MZp B-cells. We show that NR4A1-3 as well as CD83 gene and protein expression levels are severely downregulated in blood MZp B-cells from PHI progressors when compared to controls and HIV-elite controllers (ECs), whose plasma viral load remain undetectable in the absence of ART. Importantly, this downregulation correlates with reduced MZp Breg function. Moreover, this loss of Breg function could be related to the exhaustion profile we describe for this population. We also find that excess BAFF, both as soluble and membrane forms, directly downregulate NR4A1-3 and CD83 expression levels by MZp from the tonsils of control donors in vitro. Total tonsillar membrane BAFF expression levels were also found to correlate with MZp Breg efficiency, as donors expressing relatively low BAFF levels at baseline had a stronger Breg function than donors who had higher BAFF levels at baseline. Strikingly, adding soluble BAFF was found to impede MZp Breg function in low BAFF-expressing tonsillar donors. Altogether, these results suggest that in circumstances of excess BAFF, alteration of MZp Breg profile and function may constitute a major threat to "immune surveillance". As such, the group of Nus et al. described that depletion of NR4A1 in MZ B-cell populations exacerbated atherosclerosis, thus alluding to the MZ B-cell populations' capacity of "immune surveillance" to prevent atherosclerosis, and possibly other co-morbidities associated with chronic inflammatory conditions [23,24].

Socio-Demographic and Data Characteristics of the Cohorts Used in This Study
Characterizations of the samples from the PHI cohort selected for this study show that there are significant differences for plasma viral loads and blood CD4 + T-cell counts between progressors before and after ART treatment, as expected. No significant difference for age, CD4 + T-cell counts and CD8 + T-cell counts were found between the progressors (5-8 months) and EC (Table 1). Table 1. Socio-demographic and data characteristics of the cohorts used for the RNAseq and flow cytometry analyses. Statistical analyses were done between samples selected from progressors (HIV+) and elite controllers (EC), and between HIV+ untreated (ART−) and treated (ART+) individuals. Statistical significance of difference in clinical data between progressors before and after ART was assessed with the Wilcoxon matched-pairs test. Statistical significance of difference between progressors and EC was assessed with the Mann-Whitney U test. 1 Undetectable viral loads were found in the three ECs.

RNA-Seq Analysis
Flow-Cytometry Analysis

Blood MZp B-Cells from Progressors Show An Altered Transcriptomic Profile
We have previously shown that progressors from the Montreal PHI cohort presented higher levels of both BAFF and MZp B-cell frequencies in their blood, despite 1 year of ART [3]. To further assess the impact of HIV infection and associated excessive BAFF condition on the dysregulation of MZp, we FACS-sorted MZ and MZp B-cell populations from the blood of three PHI progressors (5-8 months), three ECs and three controls selected from the above-mentioned study [3], and proceeded to transcriptomic analyses by RNAseq. As depicted in Figure 1, PCA analyses of the top 5000 varying genes show that the gene expression profiles of blood MZ and MZp B-cell populations from the progressors present a distinct pattern when compared to that of both EC and controls, who share greater similarity ( Figure 1A). Given this similarity between EC and controls and the scope of this manuscript, we then generated a heat map of the top 100 differently expressed genes (DEG) between blood MZp B-cells of the progressors and controls. We show that among those, Interferon Stimulated Genes (ISG) such as 2 -5 -oligoadenylate synthetase 1 (OAS1) (p = 3 × 10 −6 ), SAM and HD domain-containing deoxynucleoside triphosphate triphosphohydrolase 1 (SAMHD1) (p = 5 × 10 −6 ) and SP110 nuclear body protein (SP110) (p = 3 × 10 −10 ) are significantly upregulated in MZp from the progressors, whereas genes that have a role in regulation are severely downregulated in MZp from these same individuals ( Figures 1B,C and 2). The DEG were also analyzed by a volcano plot, showing that 709 genes are significantly downregulated and 392 are significantly upregulated in blood MZp B-cells from the progressors ( Figure 1C) when compared to controls. Gene Set Enrichment Analyses (GSEA) show that both PI3K-AKT-MTOR and CREB/cJUN signatures are downregulated in blood MZp B-cells from the progressors ( Figure 1D,E, and Supplementary Figure S2). Accordingly, blood MZp B-cells from the progressors express higher levels of gene transcripts for exhaustion markers such as T-box transcription factor 21 (T-bet, TBX21) (p = 0.1), integrin subunit alpha X (CD11c, ITGAX) (p = 0.009), FCRL5 (p = 0.003), and leukocyte immunoglobulin-like receptor B1 (CD85j, LILRB1) (p = 0.007), as well as the negative regulators CD22 (p = 0.05) and CD72 (p = 0.003) ( Figure 1F-K) when compared to blood MZp B-cells from controls. Altogether, these results suggest that HIV disease progression greatly solicits the activities of MZ B-cell populations, which leads to changes in their transcriptomic profile and possibly functional impairment, especially in the MZp B-cell sub-population, which presents signs of exhaustion.

Blood MZp B-Cells from Progressors Express Lower Levels of Regulatory Markers
Consistent with the severely altered profiles described above, we found highly significant downregulation of gene transcripts for the NR4A family members in blood MZp B-cells of the progressors. Indeed, gene expression levels of NR4A1 (p = 0.01), NR4A2 (p = 3 × 10 −9 ), NR4A3 (p = 1 × 10 −36 ), as well as genes that they regulate directly such as CD83 (p = 3 × 10 −9 ) and interleukin 6 (IL-6) (3 × 10 −6 ) (Figure 2A-E), were very significantly downregulated in blood MZp B-cells from the progressors when compared to that of controls. Given the similarity observed with NR4As gene expression levels of controls and EC, and the scope of this manuscript, further characterization of MZp Breg capacities were conducted with progressors and controls. We next sought to assess whether the transcriptomic downregulation of Breg markers was suggestive of their protein downregulation. To this end, blood samples from 15 controls and 10 progressors (before and after ART) were selected, including samples from the individuals selected for the RNA-seq analyses. We have measured the protein expression levels of the Breg markers NR4A1, NR4A3, CD83, CD39 and CD73 by multicolor flow cytometry on MZp B-cells from the blood of these individuals. As expected, protein expression levels of NR4A1, NR4A3, CD83 and CD73 were downregulated by blood MZp B-cells from progressors when compared to those of controls ( Figure 2G-K). Strikingly, ART did not restore expression levels of these Breg markers and some of them, such as NR4A3 and CD83, were further downregulated despite ART ( Figure 2L-P). Of note, consistent with previous reports, membrane BAFF levels were found to be in excess in samples from these progressors and were not restored by ART ( Figure 2Q-R). Also, as expected, MZp frequencies in the blood of these progressors, without and with ART, were significantly higher compared to that of controls ( Figure 2S). However, no significant correlation between MZp frequencies and BAFF levels could be observed for progressors (R: 0.1; p-value 0.95) or after ART treatment (R: 0.3; p-value 0.6833). Of note, the small numbers of samples in our groups did not allow us to reach significance for correlations between blood BAFF levels and MZp frequencies; however, as stated earlier, we have previously reported excess BAFF to be concomitant with increased MZp frequencies in several cohorts and study systems. Our data suggest that MZp Breg capacities may be altered in progressors, regardless of therapy. These data also suggest that excess BAFF might contribute to altering the Breg profile of MZp.  and CD73 (P) expression levels between primary infection and following ART for the same donors. Membrane BAFF expression levels were assessed on total PBMCs of 10 controls and 5 progressors before and after ART (Q). Comparison between membrane BAFF expression levels on total PBMCs of progressors before and after ART for the same donors (R). Relative frequencies of MZp in the blood of 10 controls and 10 progressors before and after ART (S). Protein expression levels were assessed by Geometric Mean of Fluorescence Intensity (GeoMFI). * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. The Wald Test with Benjamini-Hochberg correction was used for RNAseq analysis (A-F). Normality was assessed with the Shapiro-Wilk test.
A One-Way ANOVA with a Tukey post-hoc test was used for testing statistical differences between groups in (I,J). A Kruskal-Wallis nonparametric test with Dunn's post-hoc test was used to assess statistical differences between groups in (G,H,K,Q,S). A paired t-test was used in assessing differences between individuals before and after ART in (L-P,R).

The Breg Function of Blood MZp B-Cells from Progressors Is Altered, despite ART
Given the important downregulation of Breg markers by blood MZp B-cells from progressors, and despite ART, we next sought to investigate whether the Breg function of MZp B-cells is altered in these individuals. In order to offset the low cellular yields of MZp B-cells in blood, High-Content Screening (HCS) by fluorescence in a confocal mode was used to assess the Breg function of sorted MZp B-cells on activated autologous CD4 + T-cells ( Figure 3A-D). We observed for controls that, while cell count, as determined by the amount of positive CD4 + cells in each analyzed well (indicator of proliferation), seemed to be the same when CD4 + T-cells were cultured alone or co-cultured with either MZp or CD1c − B-cells (negative control) ( Figure 3E), mortality rates of CD4 + T-cells were strikingly higher in the MZp: T-cell co-culture, suggesting regulation by controlled cell death ( Figure 3F). As such, induction of apoptosis is a common regulatory function and has been described for Tregs [25]. Accordingly, we found that MZp B-cells strongly express CD274 (PD-L1, B7H1) when compared to other B-cell subtypes (Supplementary Figure S3). Furthermore, the addition of an anti-PD-1 blocking antibody to the MZp:T-cell co-culture diminished the mortality rates of CD4 + T-cells, thus, supporting a role for the PD-1 (PDCD1)/PD-L1 pathway in MZp Breg function ( Figure 3H). However, when assessing the Breg function of blood MZp from progressors, our results rather suggest that there is a lower induction of CD4 + T-cell mortality in presence of MZp, and, strikingly, despite ART ( Figure 3G). Altogether, our data suggest that HIV infection affects the Breg function of blood MZp B-cells, and this is not restored by ART.
CD1c − B-cells (negative control) ( Figure 3E), mortality rates of CD4 + T-cells were strikingly higher in the MZp: T-cell co-culture, suggesting regulation by controlled cell death ( Figure  3F). As such, induction of apoptosis is a common regulatory function and has been described for Tregs [25]. Accordingly, we found that MZp B-cells strongly express CD274 (PD-L1, B7H1) when compared to other B-cell subtypes (Supplementary Figure S3). Furthermore, the addition of an anti-PD-1 blocking antibody to the MZp:T-cell co-culture diminished the mortality rates of CD4 + T-cells, thus, supporting a role for the PD-1 (PDCD1)/PD-L1 pathway in MZp Breg function ( Figure 3H). However, when assessing the Breg function of blood MZp from progressors, our results rather suggest that there is a lower induction of CD4 + T-cell mortality in presence of MZp, and, strikingly, despite ART ( Figure 3G). Altogether, our data suggest that HIV infection affects the Breg function of blood MZp B-cells, and this is not restored by ART.

Excess BAFF Directly Contributes to the Altered Breg Potential of MZp B-cells
Our RNA-seq analyses show that gene expression levels of BAFF-R are significantly downregulated by MZp from the blood of progressors compared to controls, while those of TACI and BCMA are not significantly affected (Supplementary Figure S4), albeit a trend for BCMA upregulation by blood MZp from progressors could be observed. These are consistent with our observations that expression levels of BAFF-R, but not those of TACI, are decreased on B-cells following culture with soluble BAFF (Supplementary Figure S5). CD4 + T-cell count on each analyzed well (E). CD4 + T-cells mortality following culture with Medium Alone or when co-cultured (3:1 ratio) with CD1c-B-cells (negative control) or MZp (F). CD4 + T-cells mortality when co-cultured (3:1 ratio) with MZp B-cells from controls and progressors (with and without ART) (n =3) (G). CD4 + T-cells mortality when co-cultured with MZp from controls with and without the addition of an anti-PD-1 blocking antibody at 2 µg/mL (H). Mortality was assessed as the relative frequency of Live or Dye cells when compared to the total CD4 + T-cells. ** p < 0.01. Statistical differences between groups were assessed with the Kruskal-Wallis test with post-hoc Dunn test. Normality was assessed with the Shapiro-Wilk test.

Excess BAFF Directly Contributes to the Altered Breg Potential of MZp B-cells
Our RNA-seq analyses show that gene expression levels of BAFF-R are significantly downregulated by MZp from the blood of progressors compared to controls, while those of TACI and BCMA are not significantly affected (Supplementary Figure S4), albeit a trend for BCMA upregulation by blood MZp from progressors could be observed. These are consistent with our observations that expression levels of BAFF-R, but not those of TACI, are decreased on B-cells following culture with soluble BAFF (Supplementary Figure S5). To assess the relative contribution of excess BAFF on the dysregulation of MZp Breg capacities, blood MZp B-cells from controls have been sorted and cultured in the presence or absence of soluble BAFF at 50 ng/mL or 500 ng/mL overnight. Cells were then harvested for RNA extraction and a one-step qPCR for NR4A1, NR4A3 and CD83 mRNA quantification was performed, with GAPDH as a housekeeping gene. Our data show that at high concentrations (500 ng/mL), BAFF-treated blood MZp B-cells express lower NR4A1, NR4A3 and CD83 mRNA levels compared to unstimulated MZp B-cells, similar to that of sorted MZp B-cells from the blood of progressors ( Figure 4A-C). Likewise, we enriched total B-cells from the tonsils of control donors and cultured them overnight with soluble BAFF at increasing concentrations. We found that tonsillar MZp B-cells treated with high BAFF concentrations show lower expression levels of NR4A1 and CD83 proteins compared to those treated with lower BAFF concentrations or medium alone ( Figure 4D-G). Interestingly, we found that, as soluble BAFF concentrations increase, IL-10 expression levels decrease ( Figure 4H,I). Of note, we had previously shown that IL-10 is highly expressed by ex vivo unstimulated blood MZp B-cells when compared to other B-cell populations ( [26], and Supplementary Figure S6 Figure S7). We observe a trend for the downregulation of NR4A1 and NR4A3 expression levels by these blood MZp B-cells. Although this finding will need further investigation, it is tempting to speculate that DC-like and MoDC-like populations in secondary lymphoid tissues, which are known to interact with B-cell populations such as MZ B-cells [11], may influence MZp Breg activity through membrane and/or released BAFF expression levels.

Discussion
There is a limitation of our study in the number of participants, and we believe adding further subjects in the future will strengthen the interpretation of our data. Nevertheless, the present study sheds some light on a better understanding of the important Breg potential of MZp B-cells, and its dysregulation in the context of chronic inflammation involving excess BAFF, such as is encountered in HIV infection. Importantly, MZp Breg potential was not restored by therapy. Several factors can be nourishing chronic inflammation and impacted immunocompetence. For example, we have shown that factors such as Nef persist in blood populations and directly contribute to BAFF over-expression, as do factors of microbial translocation, such as LPS [6]. These observations further stress the importance of very early treatment in cases of PHI to avoid the establishment of a chronic inflammatory status and its impact on immunocompetence [5].
We have previously shown that excess BAFF in the blood of progressors was concomitant with the dysregulation of the blood B-cell compartment, notably with increased frequencies of MZp B-cells [3], despite ART. In this study, we report that blood MZ populations, especially MZp B-cells from progressors, present a different transcriptomic profile compared to that of ECs and controls, suggestive of their being highly solicited and functionally impaired, with signs of exhaustion. Importantly, differences include the significant downregulation of gene transcripts for the molecules NR4A1, NR4A2, NR4A3 and CD83, which we have previously associated with the Breg potential of MZp B-cells in controls [15]. Furthermore, the protein expression levels of these molecules as well as those of CD73, also associated with regulatory capacities [27], are downregulated in blood MZp B-cells of progressors, despite ART. Even more importantly, we found that the Breg function of blood MZp B-cells is impaired in progressors and ART-treated individuals. Consistent with our previous observations [3], levels of BAFF were found to be in excess in the blood of progressors and ART-treated individuals selected for this study when compared to controls. Importantly, in vitro studies show that high levels of BAFF impact the expression of regulatory proteins by MZp B-cells from the tonsils of control donors and have a direct effect on their Breg function.
Our findings are suggesting that, in the context of HIV, which we characterize as a persistent excessive BAFF environment, MZ populations may be highly solicited, selected and expanded to produce antibodies at the expense of their Breg competence. As such, BAFF levels may dictate MZ B-cell activities, from Breg to antibody production, which in the context of HIV may be expedited, bypassing the Breg capacity, for tentative eradicative purposes. As to whether MZ B-cells antibody responses are beneficial or a nuisance to the host in the context of HIV remains to be established. As such, given the facts that HIV-Transgenic and BAFF-Transgenic mice presented a breakage of tolerance and autoimmunity [8,[28][29][30], and that excess BAFF and MZ B-cell dysregulations have also been reported in the context of autoimmune diseases such as Systemic Lupus Erythematosus (SLE) and Sjögren Syndrome (SS) [31], it would be sound to bear in mind that MZ B-cell antibody responses in the context of excess BAFF may be nourishing polyclonal, possibly auto-reactive, low-affinity antibody responses, whose pertinence in the battle against HIV have yet to be determined.
Notably, individuals with SLE or SS present a strong IFN signature, which promotes BAFF expression, reminiscent of what was observed in the context of chronic infections such as with HIV [32,33]. Thus, unsurprisingly, we found that genes that are associated with proinflammatory outcomes and ISG are upregulated in blood MZp B-cells of progressors. As such, we found that transcripts for genes involved in the regulation of IFN responsiveness such as OAS1, SP110 and IRF7 are highly upregulated in MZp B-cells of progressors when compared to ECs or controls. These data suggest that MZp B-cells are in a hyperactive state in the context of HIV, in a similar manner to that observed in autoimmune diseases such as SLE or SS.
Our GSEA analyses show that the major PI3K-AKT-MTOR pathway is downregulated in blood MZp B-cells from progressors. Mechanistic target of rapamycin kinase (MTOR) has been shown to be important for normal MZ B-cell function, as TACI signaling converges towards MTOR in order to lower the MZ activation threshold [34], which is in accordance with their role as first-line innate-like cells that respond quickly to blood-borne pathogens [12]. Interestingly, the transcript for the gene coding for BAFF-R is significantly downregulated by blood MZp B-cells of progressors, while the transcript for TACI remains highly expressed (Supplementary Figure S4). Unfortunately, we did not assess BAFF-R nor TACI protein expression levels by blood MZp in these individuals, but they are also likely to be modulated as we found that excess soluble BAFF in vitro downregulates cell surface expression levels of BAFF-R but not those of TACI on blood MZp (Supplementary Figure S5). It is therefore possible that, in progressors, reduced MZp expression of BAFF-R reflect ongoing exposure to increased amounts of its cognate ligand BAFF. Moreover, our data also suggest that BAFF signals may be mostly mediated via TACI during HIV infection. Importantly, the binding of BAFF to TACI allows the recruitment of TRAFs such as TRAF3 [35], which has been shown to inhibit the CREB pathway in B-cells [36], known to induce the expression of the NR4As [37,38], and which we find is also downregulated in MZp B-cells from progressors. Of note, a trend of low TRAF6 gene transcripts was found in progressors (p = 0.1), but not TRAF3 (Supplementary Figure S4), which could mean that TRAF3 could see a higher recruitment rate due to lower competition for TACI signals. These data suggest that the excess BAFF found during HIV infection may lead to CREB and NR4As downregulation via excess TACI signaling, and, thus, excess TRAF3 recruitment. Given the importance of CREB and NR4As in controlling cell activation [39], it is possible that their downregulation allows for MZp B-cells to maintain a hyperactive state despite MTOR downregulation. The fact that gene expression levels of BCMA show a trend for upregulation by blood MZp from progressors when compared to controls, could suggest differentiation towards an early plasmablast stage, as they also significantly upregulated CD38 gene transcripts (Supplementary Figure S4L). Importantly, NR4A1, NR4A2 and NR4A3 gene transcripts were downregulated in a highly significant manner in MZp B-cells from the blood of progressors, with p values reaching orders of magnitude of 0.01, 10 −9 and 10 −36 , respectively. In addition, other important regulatory molecules that are directly modulated by NR4As such as CD83 and IL-6 [18] were also significantly downregulated. Consistently, blood MZp B-cells from progressors presented significantly lower protein expression levels of NR4A1, NR4A3 and CD83, and this was not restored by ART. The importance of NR4A transcription factors in immune regulation has been highlighted by the fact that NR4A1-3 doubleknockout mice develop systemic autoimmune diseases [40]. Furthermore, in humans, downregulation of NR4As has been observed in the context of autoimmunity [40] and acute myeloid leukemia [41] whereby synthetic regulation of NR4A expression is currently used for eliminating cancerous cells [42]. Additionally, since the expression of NR4As is necessary for the maintenance of FOXP3 expression by Tregs [17], there have been reports of several studies where those genes are targeted to modulate Treg responses in cancer [42]. Moreover, increasing NR4A1 expression in myeloid cells led to a diminished T-cell activation profile [21]. It is, thus, reasonable to think that NR4A modulation could be envisaged in view of restoring MZp Breg capacities, possibly in conjunction with therapies reducing excessive BAFF levels, already FDA-approved in the treatment of SLE.
We found similar observations for the ectonucleotidase CD73, the limiting molecule of the 2-aminoethanethiol dioxygenase (ADO) pathway [43]. Others have reported that the diminution of CD73 + B-cells and CD39 + CD73 + B-cells and the diminution of ADO correlated with a loss of CD4 + T-cells and disease progression during HIV infection [44,45]. Interestingly, the HIV context is also associated with downregulated CD39 and CD73 expression levels in Tregs, despite ART [46]. In other chronic infections, such as hepatitis, a loss of CD73 and CD39 has also been observed and associated with viral load and high levels of inflammation [47]. It is also important to remember that ADO has been found to modulate NR4As expression levels via purinergic receptors such as A2 A [48]. Therefore, the loss of CD39 and CD73 can contribute to a reduction of extracellular ATP conversion into ADO, which could promote a pro-inflammatory environment caused by the accumulation of extracellular ATP and a decrease in NR4A expression due to a decrease in the interaction between ADO and its receptors. Since the MTOR pathway is upregulated by the binding of ADO to its receptors [49], it is possible that the lower levels of ADO production by regulatory cell populations (MZp B-cells and others) contribute to the downregulation observed for the MTOR pathway.
Interestingly, we noticed that gene transcripts for Breg markers, such as IL-10 and the IL-10 regulatory transcription factor aryl hydrocarbon receptor (AHR), recently shown to be an important Breg marker in mice [50], were not affected in blood MZp from progressors, whereas others, such as transcription factor basic helix-loop-helix family member E40 (BHLHE40), also involved in IL-10 regulation, were significantly diminished when compared to controls (Supplementary Figure S4).
Consistent with their loss of several Breg markers, we show that blood MZp B-cells from progressors and ART-treated individuals lose their Breg function. We have previously shown that MZp B-cells from controls can control CD4 + T-cell proliferation and that this involves signals delivered via CD83 [15]. Here, we show that the control of MZp B-cells on the activation of autologous CD4 + T-cells may also involve controlled cell death and PD-1/PD-L1 signaling. Accordingly, the reduced Breg function observed for blood MZp from progressors, without and with ART, is concomitant with their lowered expression levels of PD-L1 (Supplementary Figure S3). Of note, the differences observed between PD-L1 gene transcripts, the expression levels of which were not affected in blood MZp from progressors, and PD-L1 protein expression levels, which were decreased, suggests that protein expression of PD-L1 could be affected by post-transcriptional mechanisms. Further studies are, thus, needed to confirm the relative involvement of all markers engaged in the MZp Breg function. Importantly, we show that excess BAFF can directly impact on MZp Breg capacities. Indeed, gene and protein expression levels of the regulatory markers NR4A1, NR4A3 and CD83 were downregulated in the presence of high-soluble and membrane BAFF levels, which also had an impact on IL-10 production, which could contribute further to the loss of Breg function in these cells.
When performing in vitro studies with tonsils from control donors, we noticed that BAFF levels varied between donors and this influenced in vitro MZp Breg capacities. Variations of BAFF levels between each donor appear to be especially attributable to levels expressed by myeloid cells such as DC-like and MoDC-like populations. As such, DC are an important source of BAFF, and have been shown to interact with B-cells in T-independent manners [12], and could have, thus, modulated the tonsillar MZ B-cell environment. Of note, in HIV-Tg mice, we have observed that myeloid DCs accumulated in the expanded MZ of these mice [51]. We have previously published that HIV infection increases BAFF levels on DC and MoDCs via Nef [7]. Likewise, elements of microbial translocation, such as LPS were also shown to increase BAFF expression by MoDCs ( [7], Supplementary Figure S7). Our findings point to BAFF as being a major contributor to the loss of MZp Breg capacities during HIV infection. Similar to that observed for SLE and other rheumatic disorders, there is growing evidence that BAFF is in excess in contexts of cancer, and could also play a role in the pathogenesis of several inflammatory chronicity, such as multiple sclerosis and HBV infection [32], suggesting that excess BAFF and dysregulated MZp may be important indicators of deterioration of immunocompetence.
We found relatively higher gene expression levels of exhaustion markers, such as FCRL5 and CD85j, and negative regulators, such as CD22 and CD72, in blood MZp B-cells of progressors when compared to controls, which suggests a higher activation index. Additionally, MZp B-cells express both IL-21R and TLR7 (Supplementary Figure S4), which signals are important to drive B-cell differentiation towards an "age associated"-like exhausted B-cell profile. As such, blood MZp B-cells from progressors also possess higher gene expression levels of T-bet and CD11c, a characteristic related to age-associated B-cells, and shared by a similar heterogeneous population found to be expanded in autoimmunity, chronic infections, antibody-mediated rejection in organ transplants [52][53][54] and critically ill patients with SARS-CoV-2 [55]. Interestingly, B-cells that express T-bet and CD11c have been associated with their accumulation at extra-follicular sites and the production of immunoglobulins (IG) of poor affinity [56].
To date, many groups have identified MZ-like populations with Breg capacities vs. T-cells, in non-cognate fashions, in vitro, in murine models and humans (recently reviewed by [57]). Herein, our data suggest that such Breg capacity of MZp is greatly reduced in progressors, and not restored by ART, and this possibly involves the excessive BAFF levels reported for these individuals. As to whether MZp can exert their Breg functions through cognate interactions with CD4+ T-cells, needs further experimentation and is beyond the scope of our study but deserves to be addressed. Mouse MZ B-cell populations have been shown to migrate to T-cell zones of secondary lymphoid organs (SLO), and activate CD4+ T-cells in a HEL-dependent manner [58]. This, and the fact that MZ B-cell populations have the capacity to shuttle to follicular areas [59], and that MZ B-cells were shown to possess an atheroprotective role attributed to their NR4A1 expression [23,24], all point to the capacity of MZ populations to be recruited to outer MZ areas of SLO, in given situations. Furthermore, MZ B-cells are able to present lipidic antigens to invariant natural killer T-cells (iNKT), in the context of CD1, which confer activation via, notably, the CD40-CD40 ligand (CD40L) pathway [60]. All of which support their influence on several cell types. The capacity of MZ B-cell populations to perform antigen presentation and Breg activity, or to differentiate towards Ig production, may be dictated by the overall situation and environment (reviewed in [61]).

Specimen Collection and Clinical Data
Cryopreserved blood specimens from 13 male HIV-1-infected classic progressors were selected from the Montreal PHI cohort specimen bank. For simplicity, the term "HIV" is used throughout this manuscript. Clinical data from these individuals can be found in Table 1. The date of infection was estimated based on clinical and laboratory results, using criteria established by the Acute HIV Infection and Early Disease Research Program (National Institute of Allergy and Infectious Diseases [NIAID], Bethesda, MD). Blood samples had been collected at two time points: the acute phase of infection (i.e., 0-3 months after HIV acquisition) and/or early phase of infection (i.e., 5-8 months after HIV acquisition), as well as 96 weeks after HIV acquisition, with an average of 92 weeks of ART treatment (Table 1). Cryopreserved blood samples from three male EC were obtained from the Montreal Long Term Non-Progressor (LTNP) cohort [26]. Cryopreserved blood samples from 18 male HIV-negative healthy volunteers were obtained from the Montreal PHI cohort, to be used as controls. Plasma viral loads and blood CD4+ T-cell counts were determined as reported previously [62]. None of the subjects had syphilis, hepatitis B, or hepatitis C.
Written informed consent was obtained from all subjects, and the research conformed to the guidelines of, and was approved by, the University of Montreal Hospital Research Center Ethics Review Board (project reference SL05.028). . Sorted live CD19 + CD1c + IgM + CD27 + CD21 hi CD10 − mature MZ and CD19 + CD1c + IgM + CD27 + CD21 lo CD10 + MZp B-cells were >95% pure. Total RNA was extracted using RNeasy Micro Kit (Qiagen) according to the manufacturer's instructions. RNA integrity was validated using an RNA Pico Chip on the Agilent BioAnalyzer 2100, and RNA was sent to the IRIC Genomics Core Facility for RNA-seq transcriptomic profiling and analysis. Libraries were prepared using Clontech Ultra Low RNA SMARTer v4 (Takara) and sequenced on a HiSeq2000. Genes with false discovery rate (FDR) values < 0.05 were considered to be differentially expressed. Gene expression levels were compared using raw read counts and the negative binomial distribution model implemented in DESEq2 [63], a differential expression analysis package developed for R, which adjusts for sample variations with the assumption that the vast majority of genes should have correlating expression levels. More specifically, the regularized log transformation (rlog) implemented in DESeq2 was used to transform raw data into log2 (readcount) for analysis and visualization.

Volcano Plot and Gene Set Enrichment Analyses (GSEAs)
The volcano plot was created using Python 3.9.5. First, a comma-separated value (CSV) file containing the raw RNA-Seq gene expression data was imported into a pandas 1.3.1 DataFrame. Each gene was assigned the color blue if significantly upregulated/downregulated (p < 0.05 and fold change > or <1), otherwise it was assigned the color gray. Annotations for the genes were assigned black if significantly upregulated/downregulated beyond the specified p-value and n-fold change. Otherwise, they were assigned red labels regardless of their p-values or n-fold-changes if they appear in a specified list of genes to highlight. Finally, the volcano plot was generated using Matplotlib 3.4.2. Specified p-values and n-fold lines were drawn as blue lines. The total number of upregulated/downregulated genes were displayed in the upper right/left corners, respectively.
Gene Set Enrichment Analyses (GSEAs) were produced using the software application GSEA 4.1.0, developed by the Broad Institute, as previously published [64,65]. Required input files are the expression data, the phenotype labels, and the gene set. The expression data file consists of a table whose first column is gene name, followed by gene description, followed by the genes' expression values for each sample. The phenotype labels file associates a phenotype label (e.g., HIV+ or HIV−) with each sample in the group. Gene sets for this manuscript were downloaded from the Broad Institute's Molecular Signatures Database (MSigDB) v.7.4.; they are KEGG's MTOR pathway (KEGG_MTOR_SIGNALING_PATHWAY) and Transcription Factors' CREB/cJUN pathway (CREBP1CJUN_01) The software generates an analysis for each gene set containing the Net Enrichment Score (NES), false discovery ratio (q-value) and p-value. A gene set is considered significantly modulated if q-value and p-value <0.05.

Multicolor Flow-Cytometry
PBMCs from 15 controls and 10 progressors (without and with ART) were processed for flow cytometry as previously described [15]. Briefly, live/dead exclusion was performed using Aqua-LIVE/DEAD Fixable Stain (Invitrogen Life technologies, Eugene, OR, USA). Non-specific binding sites were blocked using FACS buffer (1 × PBS, 2% hi-FBS, and 0.1% sodium azide) supplemented with 20% hi-FBS, 10 µg mouse IgG . Non-specific binding sites were blocked using 20% hi-FBS. The mAbs used were the PE-conjugated human REA clone anti-mouse NR4A1, the cross-reactivity of which with human has been previously assessed [15], and compared to the PE-conjugated human REA isotype control (Miltenyi Biotech, Bergisch Gladbach, Germany), as well as the PE-conjugated mouse anti-human NR4A3 (Santa Cruz Biotechnology, Dallas, TX, USA). Intra-cellular labeling for the detection of IL-10 was performed using the Intracellular Fixation & Permeabilization Buffer Set (eBioscience, San Diego, CA, USA) and the PE mouse anti-human Anti-IL-10 mAb (Biolegend, San Diego, CA, USA). Cells were kept at 4 • C in 1.25% paraformaldehyde until analysis. Data acquisition was performed with FACSFortessa (BD-Biosciences, Franklin Lakes, NJ, USA) for blood PBMC samples and LSRIIB (BD-Biosciences) for tonsillar samples (described below). Analyses were done with FlowJo 10 software and GraphPad Prism. All stainings were compared to that of fluorescence minus one (FMO) values and isotype controls (see gating strategy in Supplementary Figure S1). Anti-mouse Ig(κ) Compbeads and CS & T Beads were used to optimize fluorescence compensation settings and calibrate the LSRIIB, respectively.

Human Tonsillar B-Cells
Human tonsils from controls, who had undergone surgical tonsillectomy, were mechanically processed and cells were cryopreserved in liquid nitrogen until use, as described above. Cells were thawed and washed in IMDM, and B-cells were negatively enriched >95% by an immunomagnetic-based technology (Dynabeads Untouched Invitrogen Life technologies). Total B-cells were subsequently cultured at a concentration of 10 6 cells/mL in IMDM supplemented with 10 −4 β-2-mercaptoethanol, 10% hi-FBS, and 1% penicillin/streptomycin, in the absence or presence of stimuli (PMA/ionomycin or human recombinant soluble BAFF/BLyS/TNFSF13B Protein (R & D System)) for 18 h at 37 • C and 5% CO 2 . Cells were cultured with Brefeldin A for the last 4 h of incubation prior to staining for intra-cellular IL-10 expression levels. Cells were recovered and processed for flow-cytometry as stated above.

qPCR Characterisation of NR4A1, NR4A3 and CD83 mRNA Expression Levels by Blood MZp B-Cells
MZp B-cells sorted from the blood of controls, and cultured with or without soluble BAFF at 50 ng/mL or 500 ng/mL overnight, were assessed for their expression levels of NR4A1, NR4A3, and CD83 mRNA by qPCR. Briefly, blood MZp B-cells were cell-sorted as stated above and lysed with TRIzol solution (ThermoFisher, Waltham, NJ, USA). mRNA was extracted with the phenol/chloroform technique. The quality of extracted mRNA (260/280 and 260/230 ratios) was assessed with a DS-11 FX Spectrophotometer (DeNovix, Wilmington, DE, USA). RNA was considered pure when the 260/280 ratio was above 1.8 and the 260/2360 ratio was above 2.0. One-step qPCR was done with the QuantiNova SYBR Green qPCR Kit (QIAGEN) by using 20 ng RNA from each donor. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a housekeeping gene. The primers used were: NR4A1 forward: 5 -CCAGCACTGCCAAACTGGACT A-3 ; NR4A1 reverse: 5 -CTCAGCAAAGCCAGGGATCTTC-3 ; NR4A3 forward: 5 -CCCTTTCAGACTATCTGTAC-GGAC-3 ; NR4A3 reverse: 5 -CTCAGTGTTGGAATGGTAAAAGAAG-3 ; CD83 forward: 5 -TCCTGAGCTGCGCCTACAG-3 ; CD83 reverse: 5 -GCAGGGCAAGTCCACATCTT-3 ; GAPDH forward: 5 -GTCTCCTCTGACTTCAACAGCG-3 ; and GAPDH reverse: 5 -ACCACCCTGTTGCTGTAGCCAA-3 . Ct values were determined with the thermocycler Corbett Research RotorGene 6000. Ct is the value of the number of cycles necessary for the number of RNA copies to reach a precise fluorescence threshold. The relative expression of mRNAs was determined by the ∆Ct method, where the Ct values of each gene are subtracted from the Ct value of the GAPDH housekeeping gene. mRNA expression = (2 −∆Ct ) × 10,000.

For Assays with Tonsillar Populations
As previously described [14], live autologous MZp B-cells, CD1c − B-cells (negative control) and CD4 + T-cells were sorted (see description above for RNA-Seq) from human tonsillar samples of control donors. CD4 + T-cells were cultured alone or co-cultured with either of the B-cell populations at a ratio of 3:1 for 36 h at 37 • C and 5% CO 2 ; on anti-CD3 (2 µg/mL) (ULTRA-LEAF Biolegend, San Diego, CA, USA)-coated flat-bottomed 96 well plates with soluble anti-CD28 (2 µg/mL) (ULTRA-LEAF Biolegend, San Diego, CA, USA); and with or without human recombinant soluble BAFF at 500 ng/mL. The read-out of the Breg function was based on cell cycle entry of CD4 + T-cells, which was assessed by flowcytometry analyses of intra-nuclear Ki67 expression levels using the PE-mouse anti-human Ki67 mAb (eBioscience, San Diego, CA, USA), as described above.

For Assays with Blood Cell Populations
Because of limited quantities of cells in our populations of interest, we proceeded to High-Content Screening (HCS) assays, the read-out for Breg function of which was based on the mortality of CD4 + T-cells rather than cell cycle entry. As described above, autologous MZp B-cells, CD1c − B-cells (negative control) and CD4 + T-cells were sorted (see description above for RNA-Seq) from human PBMCs of controls and progressors (without and with ART). CD4 + T-cells were cultured alone or co-cultured with either of the B-cell populations at a ratio of 3:1 for 60 h at 37 • C and 5% CO 2 on anti-CD3 (2 µg/mL) (ULTRA-LEAF Biolegend, San Diego, CA, USA)-coated flat-bottomed 384 CellCarrier Ultra microplates (PerkinElmer, Waltham, MA, USA) with soluble anti-CD28 (2 µg/mL) (ULTRA-LEAF Biolegend, San Diego, CA, USA) in the presence or absence of a mouse anti-human PD1 (2 µg/mL) blocking antibody (Biolegend, San Diego, CA, USA). CD4 + T-cells were identified using a mouse-anti-human CD4 (OKT4 clone) primary mAb and FITC goat anti-mouse secondary antibody (Invitrogen, Waltham, MA, USA). The cells were then fixed using 4% paraformaldehyde before the nuclei were stained with Hoescht 33342 (Invitrogen, Waltham, MA, USA). Data acquisition was performed with Operetta and analyses were done using Harmony and GraphPad Prism. The algorithm excluded debris based on Hoescht intensity, then excluded B-cells based on FITC negativity, and dead cells were counted based on the intensity of Live or Dye CellMask signals of a mortality control. To this end, CD4 + T-cells were cultured alone in a flat-bottomed 96-well plate for 60 h in the same conditions and then collected and placed at 56 • C in a water bath for 45 min. The cells were then stained with Live or Dye CellsMask for Live/Dead exclusion.

MoDC: B-Cell Co-Cultures
To assess the impact of membrane-bound BAFF on MZp expression levels of Breg markers, MoDC were generated from blood samples of controls, as previously described [7]. Briefly, cryopreserved PBMCs were thawed and washed in IMDM and CD14 + monocytes were negatively enriched >95% by an immunomagnetic-based technology (Dynabeads Untouched Invitrogen Life technologies). Total CD14 + monocytes were subsequently cultured at a concentration of 5 × 10 5 cells/mL in IMDM supplemented with 10 −4 β-2-mercaptoethanol, 10% hi-FBS, 1% penicillin/streptomycin, 100 ng/mL Human Recombinant IL-4, and 250 ng/mL Human Recombinant GM-CSF (ThermoFisher, Waltham, NJ, USA) for six days at 37 • C and 5% CO 2 . During the third day of incubation, 30% of growth conditions were reinstated. At day 6, MoDC purity was assessed by flow cytometry as stated above (>90% of cells were HLA-DR + CD11c + CD14 +/− ). MoDC were subsequently cultured, with and without LPS (2 µg/mL), overnight at 37 • C and 5% CO 2 . MoDC were then assessed for membrane BAFF expression levels by flow-cytometry as stated above, fixed in 0.25% paraformaldehyde, washed and co-cultured with autologous total B-cells (enriched as described above) at a 1:10 ratio overnight at 37 • C and 5% CO 2 . Lastly, B-cells were harvested and processed for flow cytometry, as stated above.

Statistical Analyses
Statistical significance of differences between groups was assessed with a one-way ANOVA with post-hoc Tukey test for data normally distributed or otherwise with the Kruskal-Wallis test with post-hoc Dunn test. Statistical significance of difference in clinical data between progressors before and after ART was assessed with the Wilcoxon matchedpairs test. Statistical significance of difference between progressors and EC was assessed with the Mann-Whitney U test. Analyses were performed using GraphPad Prism 9.1.1, on Windows. Results were considered significant when p < 0.05.

Conclusions
Our observations are essential to help develop strategies viewed at restoring MZp immune surveillance activities. To this end, existing strategies, such as dihydroergotamine (DHE) viewed to upregulate NR4As expression levels could be envisaged to reinstate MZp Breg capacities. Additionally, BAFF-blocking therapies, such as Belimumab (Benlysta), which is FDA-approved in treating SLE, could be contemplated to try to control BAFF levels in order to lower the inflammatory burden and restore B-cell immune competence.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki, and the research conformed to the guidelines of, and was approved by, the University of Montreal Hospital Research Center Ethics Review Board (project reference SL05.028).
Informed Consent Statement: Written informed consent was obtained from all subjects involved in this study.