Identification of CX3CR1+ mononuclear phagocyte subsets involved in HIV-1 and SIV colorectal transmission.

The difficulty to unambiguously identify the various subsets of mononuclear phagocytes (MNPs) of the intestinal lamina propria has hindered our understanding of the initial events occurring after mucosal exposure to HIV-1. Here, we compared the composition and function of MNP subsets at steady-state and following ex vivo and in vivo viral exposure in human and macaque colorectal tissues. Combined evaluation of CD11c, CD64, CD103, and CX3CR1 expression allowed to differentiate lamina propria MNPs subsets common to both species. Among them, CD11c+ CX3CR1+ cells expressing CCR5 migrated inside the epithelium following ex vivo and in vivo exposure of colonic tissue to HIV-1 or SIV. In addition, the predominant population of CX3CR1high macrophages present at steady-state partially shifted to CX3CR1low macrophages as early as three days following in vivo SIV rectal challenge of macaques. Our analysis identifies CX3CR1+ MNPs as novel players in the early events of HIV-1 and SIV colorectal transmission.

Introduction

Human immunodeficiency virus type 1 (HIV-1) transmission across the colorectal mucosa is the most common mode of acquisition of the virus in western countries (Baldwin and Baldwin, 2000; Misegades et al., 2001). A better understanding of the early target cells and mechanisms that allow HIV-1 passage across the colorectal mucosa could guide the development of prevention strategies aimed at blocking viral entry and spread.

Research on simian immunodeficiency virus (SIV) in nonhuman primates has provided invaluable insights on the pathogenesis of HIV-1 infection. However, the interpretation of results from pathogenic studies and vaccine trials may be biased by possible differences between simian and human mucosa in terms of the phenotype, frequency, and function of target cells present at the site of viral entry. Consequently, the cross-validation of relevant findings in both species is required.

The mononuclear phagocyte (MNP) population that resides in the lower intestinal tract of humans and macaques includes numerous subsets of dendritic cells (DCs) and macrophages (Mφs), shown to be among the first target cells encountered by the virus (Cavarelli et al., 2013; Cavarelli and Scarlatti, 2014). Indeed, CD11c+ HLA-DR+ DCs from the human jejunal mucosa were shown to rapidly transmit HIV-1 infection (Shen et al., 2010) and those from the human colonic mucosa to sample luminal R5 HIV-1 through an envelope-CCR5-mediated mechanism (Cavarelli et al., 2013) in ex vivo tissue models. Consistent with these observations, CD11c+ HLA-DR+ cells of the cynomolgus macaque colonic lamina propria similarly respond to SIVmac251 stimulation ex vivo (Cavarelli et al., 2021), further demonstrating that rapid recruitment of these target cells inside the intestinal epithelium is involved in the early steps of viral dissemination. Recently, two cell subsets in the human colonic mucosa, CD14+ CD1c+ monocyte-derived DCs and langerin-expressing conventional DC2, were reported to be involved in HIV-1 uptake and transmission ex vivo (Rhodes et al., 2021).

Although tissue resident Mφs have been less studied than DCs, human jejunum Mφs have been reported to support HIV-1 replication (Li et al., 1999; Smith et al., 1997a, 1997b) and Mφs from the rectum to constitutively express CCR5 (McElrath et al., 2013), thus being potential HIV-1 target cells. However, the distinction between Mφs and DCs in the intestinal lamina propria based on major surface markers is not always conclusive, and the inability to clearly differentiate between them makes it difficult to define the relative role of each subset in HIV/SIV acquisition and pathology. Indeed, aside from classic Mφs, the gut also contains an atypical population of Mφs that express CD11c, a marker previously specifically associated with DCs (Bradford et al., 2011). Human intestinal CD11c+ Mφs have been recently shown to be immature, loosing CD11c expression with maturation (Bujko et al., 2018). Moreover, DCs can be distinguished from other MNPs, such as monocytes and monocyte-derived Mφs, by their lack of expression of the high-affinity IgG receptor FcγRI, CD64 (Tamoutounour et al., 2012). In contrast to DCs, CD64+ MNPs do not express the integrin αE chain CD103 and include all chemokine CX3C receptor (CX3CR)1high cells, which are tissue-resident Mφs (Bain et al., 2012; Persson et al., 2013; Tamoutounour et al., 2012). The CD64+ population also contains a subset of CX3CR1int/low cells, representing an intermediate stage in the local differentiation of tissue-resident Mφs (Desalegn and Pabst, 2019; Tamoutounour et al., 2012). Of note, intestinal CX3CR1high Mφs have been shown to be abundant in the steady-state and to have an anti-inflammatory phenotype, whereas CX3CR1low Mφs are more abundant in the inflamed intestine and adopt a pro-inflammatory phenotype (Bain et al., 2012; Desalegn and Pabst, 2019). Finally, a discrete population of DCs expressing low levels of CX3CR1 has also been identified in both the mouse and human gut (Bujko et al., 2018; Persson et al., 2013).

Mouse studies have shown that CD103+ and CX3CR1+ cells play different roles in antigen sampling. CD103+ cells are considered to be bona fide DCs, able to migrate to mesenteric lymph nodes upon Toll-like receptor (TLR) stimulation (Varol et al., 2009), priming naive T cells (Schulz et al., 2009). Conversely, CX3CR1+ Mφs are able to shuttle across the epithelium to collect bacterial antigens (Niess et al., 2005) but unable to migrate to mesenteric lymph nodes (Schulz et al., 2009). Although abundant in the colorectal mucosa, CX3CR1+ MNPs have been mostly overlooked in studies on HIV-1 transmission. Moreover, the composition and function of the intestinal MNP subset is still a subject of debate (Chiaranunt et al., 2021) and needs to be clarified in the context of mucosal HIV/SIV infection.

Here, we extensively characterized the MNP subsets that reside in the resting colorectal mucosa of both humans and cynomolgus macaques and analyzed their dynamics following ex vivo and in vivo viral exposure. We found that CD11c+ DCs and Mφs expressing CX3CR1, but not CD11c+ CD103+ DCs, insert themselves in between intestinal epithelial cells to sample luminal virions following ex vivo and in vivo viral exposure. In vivo, we show that the abundance of these cells in the intestinal lamina propria decreased as soon as three days post-SIVmac251 exposure. Among them, CX3CR1+ DCs appeared to accumulate in the draining lymph nodes, whereas Mφs remained in situ while switching from a tissue-resident CX3CR1high to a pro-inflammatory CX3CR1low phenotype. Our results reveal a novel role for CX3CR1+ MNPs in the early events of HIV/SIV transmission.

Results

Characterization of mononuclear phagocyte subsets of the human and macaque colonic lamina propria

Colonic lamina propria MNPs were identified among living CD45+ cells as HLA-DR+ lineage− cells (lineage defined as CD3−CD56−CD19−CD16− for human tissue samples and CD3−CD8−CD20− for macaques) by flow cytometry (Figures 1A and 1B). Among the CD45+ cells, which represented approximately 50% of the recovered living cells from both species (Figure 1C), the frequency of HLA-DR+ lineage− cells was 3.6 ± 1.8% in the human and 5.8 ± 2% in the macaque colon tissue (Figure 1D). Analysis of CD11c and CD64 showed three distinct populations in both the human (Figures 1A and 1E) and macaque (Figures 1B and 1F) mucosa: i.e. CD11c+ CD64− myeloid DCs, CD11c+ CD64+ Mφs, and CD11c− CD64+ Mφs.

Figure 1

Phenotypic characterization of lamina propria mononuclear phagocytes in the human and macaque colon

(A and B) Representative flow cytometry plots of MNPs in untreated and morphologically preserved human (A) and macaque (B) colon. Leukocytes are identified among live cells as CD45+. Three subsets of MNPs were identified based on the expression of CD11c and CD64 within single, viable, CD45+ HLA-DR+ lineage− cells: CD11c+ CD64− DCs, CD11c+ CD64+ Mφs, and CD11c− CD64+ Mφs.

(C and D) The frequency of (C) CD45+ cells and (D) HLA-DR+/lineage− cells (E, F) are shown. Symbols: DCs and Mφs in humans = round symbols; macaques = squared symbols; colon sample number: n = 12 for human and n = 10 for macaque tissues. The mean ± the SD is plotted. Mann-Whitney tests were used to compare the frequency of cells between human and macaque colon, whereas Friedman tests with FDR correction were used to compare DC and Mφ numbers within the same colon donor. p-values < 0.05 were considered significant (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001).

In the human colon, the abundance of CD11c+ CD64+ immature Mφs was significantly higher than that of CD11c− CD64+ mature Mφs (1.5 ± 0.8% and 0.7 ± 0.6% CD45+ cells, respectively, p = 0.009), whereas we observed mostly mature Mφs in the macaque tissue (0.2 ± 0.1% and 0.7 ± 0.8% CD45+ cells, respectively, p = 0.0018). The frequency of immature Mφs was significantly lower than that of CD11c+ CD64− DCs in the macaque tissue (p = 0.0019). CD11c+ CD64− DCs represented 0.9 ± 0.4% of CD45+ cells in the human and 0.34 ± 0.15% in the macaque lamina propria (Figures 1E and 1F).

Assessment of CD103 and CX3CR1 expression by CD11c+ CD64− DCs, CD11c+ CD64+ Mφs, and CD11c− CD64+ Mφs was performed, and gate was positioned based on the fluorescence minus one (FMO) controls (Figures 1A, 1B and S1).

Comparisons of DC subsets showed CD103+ CX3CR1− DCs to be similarly distributed in both species (36.9 ± 15.7% of DCs in humans and 32.1 ± 10.2% in macaques, Figure 2A). However, CD103− CX3CR1− DCs represented 60.7 ± 14.9% of total DCs in humans and 33.5 ± 18.9% in macaque tissues. CD103− CX3CR1+ DCs represented a discrete population of cells that was rare in the human colon but more common in that of the macaques (2.1 ± 2.2% versus 19.3 ± 7.7% of DCs). We did not observe cells expressing both CD103 and CX3CR1 in either species. CX3CR1 levels were lower on DCs relative to both CD11c+ and CD11c− CD64+ Mφs (Figure 2B), as previously described in the mouse lamina propria (Persson et al., 2013).

Figure 2

Definition of DC and Mφ subsets

(A) Relative proportion of CD103+ CX3CR1−, CD103− CX3CR1−, and CX3CR1+ CD103− CD11c+ CD64− DCs in human and macaque colonic mucosa.

(B) Human (upper panel) and macaque (lower panel) CD11c+ CD64− CX3CR1+ DCs expressing less CX3CR1 relative to CD11c+ CD64+ Mφs and CD11c− CD64+ Mφs are shown. The shaded histogram is the isotype control.

(C and D) Human (upper panel) and macaque (lower panel) CD11c+ CD64+ Mφs (C) and CD11c− CD64+ Mφs (D) were divided into CX3CR1high and CX3CR1low subsets relative to the fluorescence minus one (FMO) control (gray).

(E and F) The relative proportion of CD103+ CX3CR1−, CD103− CX3CR1low, and CD103- CX3CR1high cells among the CD11c+ CD64+ Mφs and CD11c− CD64+ Mφs in human (E) and macaque (F) colonic mucosa is displayed. Results were obtained from tissues from12 human donors and 10 macaques. The mean ± the SD is plotted. Friedman tests with FDR correction were applied to the data (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.001).

Further analysis showed varying levels of CX3CR1 for both CD11c+ and CD11c− colonic Mφs, which allowed us to differentiate, among CD103− cells, two subsets: CX3CR1high and CX3CR1low (Figures 2C and 2D). In the human colon, we detected 83.7 ± 17.4% CD103−CX3CR1high CD11c+ Mφ and 83.6 ± 13.7% CD103− CX3CR1high CD11c− Mφs (Figure 2E). Approximately 15% of both CD11c+ and CD11c− Mφ were CD103- CX3CR1low cells. As in human tissues, in macaque the Mφs were mostly CX3CR1high, independently of CD11c expression (94.5 ± 5.9% of CD11c+ Mφs and 85.4 ± 13.1% of CD11c− Mφs, Figure 2F). In contrast to DCs, CD103 was virtually absent from all colonic Mφs of both species (Figures 2E and 2F).

Overall, these results show that, despite differences between the immature/mature Mφ ratio, the same subsets of DCs and Mφs are present in human and macaque colonic mucosa.

Similar distribution of mononuclear phagocyte subsets at steady state in human and macaque colonic lamina propria by microscopic analysis

We analyzed at confocal microscope the tissue distribution of the various cell subsets identified by flow cytometry, combining CD11c with CD103, CX3CR1, or CD64 labeling (Figure 3).

Figure 3

Distribution of mononuclear phagocyte subsets in human and macaque colonic mucosa at steady state

(A–L) Sections from untreated human (A–F) and macaque (G–L) colonic tissues fixed with 4% PFA and cryopreserved are shown. DC and Mφ subsets were identified by labeling with CD11c-FITC (green) in combination with either CX3CR1-PE, CD103, or CD64, followed by goat-anti-mouse Alexa Fluor 594 (red). DAPI (blue) was used to stain the nuclei. The bright field image (BF, gray) was superimposed with the fluorescence images to visualize the contour of the intestinal barrier. The lower panels (D–F and J–L) are magnifications of the boxed area in the upper panel. Scale bars indicate the magnification.

We observed CD11c+ CD103− MNPs, CD11c− CD103+ lymphocytes, and CD11c+ CD103+ DCs in both human (Figures 3A and 3D) and macaque mucosal tissues (Figures 3G and 3J). Among them, cells positive for one marker only were mostly apical, whereas those expressing both markers were localized more deeply inside the lamina propria.

CD11c+ CX3CR1+ cells, i.e., Mφs, and, to a lesser extent, DCs, as well as CD11c+ CX3CR1− and CD11c− CX3CR1+ cells, were detected. However, the two latter subsets were more frequently found (Figures 3B and 3E shows human cells and Figures 3H and 3K macaque cells).

Finally, both CD11c+ CD64+ immature and CD11c− CD64+ mature Mφs and CD11c+ CD64− DCs were detected throughout the mucosa (Figures 3C and 3F shows human cells and Figures 3I and 3L macaque cells). In line to what previously described (Han et al., 2021; Sagaert et al., 2012), Mφs were often detected in the sub-epithelial region of the intestinal mucosa.

The HIV-1 coreceptor CCR5 is preferentially expressed by CD11c+ CX3CR1+ mononuclear phagocytes rather than by CD11c+ CD103+ DCs

We showed that the interaction between the HIV-1 envelope and CCR5 coreceptor induced the migration of lamina propria CD11c+ cells to the epithelium to capture luminal virions (Cavarelli et al., 2013). Therefore, we analyzed CCR5 expression by colonic DC and Mφ subsets in tissues from five macaques by flow cytometry. Both CD11c+ and CD11c− Mφs showed higher CCR5 expression than DCs (p = 0.0039 and 0.0085, respectively, Kruskal-Wallis test with FDR correction), with CD64− DCs being predominantly CCR5− cells (Figure 4A). This analysis was limited to macaques due to the relatively more abundant colonic tissue available for cell isolation.

Figure 4

HIV-1 coreceptor expression by mononuclear phagocytes

Immune phenotyping was performed as described in STAR Methods and Table S1.

(A–D) Proportion of CD11c+ CD64− DCs, CD11c+ CD64+ Mφs, and CD11c− CD64+ Mφs expressing the CCR5 receptor in the colonic mucosa from five macaques CCR5 and CXCR4 expression by (B) total CD11c+ HLA-DR+ lineage- MNPs, (C) CD11c+ CD103+ MNPs, and (D) CD11c+ CX3CR1+ MNPs in human (n = 5) and macaque (n = 6) colon samples is presented. The mean ± the SD is plotted. Wilcoxon rank pairs tests were applied in panel A, whereas Friedman tests with FDR correction were used in panels B–D. p-values < 0.05 were considered significant (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001).

The analysis of the HIV-1 co-receptors CCR5 and CXCR4 in tissues from six donors of both species showed that 49.7 ± 14.7% of human and 40.9 ± 20.7% of macaque HLA-DR+ lineage−CD11c+ cells expressed both co-receptors (Figure 4B and gating strategy shown in Figure S2). Single-positive CCR5+ cells were more abundant in human (18.3 ± 9.6%) than macaque (2.9 ± 2.04%) colon, whereas proportion of CCR5+ CXCR4+ cells were similar in both species. CCR5 was less expressed by CD11c+ CD103+ (Figure 4C) than CD11c+ CX3CR1+ cells (Figure 4D), consistent with the predominant expression of CCR5 by Mφs relative to DCs observed for macaque cells.

These observations suggest that the CCR5+ CD11c+ CX3CR1+ populations of DCs and Mφs may be more prone to respond to viral stimulation by actively migrating across the epithelium and sampling virions.

CD11c+ CX3CR1+ mononuclear phagocytes, but not CD11c+ CD103+ DCs, migrate to inside the colonic epithelium following exposure to HIV-1 and SIV

We investigated whether exposure of the colorectal mucosa to CCR5 tropic-virus induces differential recruitment of MNP subsets toward the luminal epithelium using the established ex vivo polarized explant culture system that mimics the biology of HIV/SIV sexual transmission and the natural route of viral entry into the mucosa (Cavarelli et al., 2013, 2021). Confocal microscopy of human and macaque colonic tissues exposed to cell-free HIV-1 AD8 (Figures 5A–5C) or SIVmac251 (Figures 5D–5F), respectively, showed both viruses to induce a comparable migration pattern of DCs and Mφs. Indeed, CD11c+ CD64+ Mφs (Figures 5A and 5D) and CD11c+ CD64− DCs (Figures 5B and 5E) were detected in between cells of the intestinal epithelium following 30 min of viral exposure relative to control tissue (Figures 5G and 5J), consistent with the timing of our previous study (Cavarelli et al., 2013). We did not observe intraepithelial migration of CD11c− CD64+ Mφs. Further characterization showed that migrating cells were CD11c+ CX3CR1+ Mφs and CD11c+ CX3CR1+ DCs (Figures 5C and 5F). CX3CR1+ cells in control tissues are shown in Figures 5H and 5K for comparison.

Figure 5

Pattern of intraepithelial mononuclear phagocyte recruitment following viral exposure of the colonic mucosa

(A–L) Human (A–C, G–I) and macaque (D–F, J–L) colonic tissues incubated for 30 min either in medium with HIV-1 AD8 (A–C, with brightfield image superimposed to the fluorescence) or SIVmac251 (D–F), or medium only (G–L), fixed with 4% PFA and cryopreserved are shown. Labeling of CD11c-FITC (green) in combination with either CD64 (A, B, D, E) followed by goat-anti-mouse Alexa Fluor 594 (red) or CX3CR1-PE (C, F) was performed. Arrows point to subsets of CD11c+ CD64+ (A and D), CD11c+ CD64− (B and E), and CD11c+ CX3CR1+ (C and F) cells that migrate from the lamina propria to in between the epithelial cells following viral exposure. A magnification of the intraepithelial cells is shown in panels A–F (the yellow signal identifies double-positive cells). Magnification of selected regions (indicated by arrows) is shown in the boxed areas of H, J, and K panels. DAPI was used to stain the nuclei. Scale bars indicate the magnification.

Of note, CD103+ cells physiologically present beneath/within the epithelium were mainly CD3+ lymphocytes and not DCs, as they lacked the expression of CD11c (Figures 5I, 5L, 6A, and 6D). Following viral exposure, intraepithelial CD3+ CD103+ lymphocytes were still detected (Figures 6B and 6C), whereas CD11c+ CD103+ DC recruitment was not induced, neither by HIV-1 (Figures 6B and 6E) nor by SIV (Figures 6C and 6F). Intraepithelial CD11c+ CD103− MNPs were detected as well (Figure S3).

Figure 6

Intraepithelial CD103+ cells are CD3+ T lymphocytes and not CD11c+ DCs

(A–F) Distribution of CD3+ CD103+ T lymphocytes (yellow cells in the upper panels) and CD11c+ CD103+ DCs (yellow cells in the lower panels) and following 30 min of incubation with medium, HIV-1 AD8, or SIV, respectively, are shown. DAPI was used to stain the nuclei (blue). Brightfield images (BF) were superimposed to visualize the intestinal barrier contour. Arrows indicate the intraepithelial cells (or cells partially interposed to epithelium), i.e. CD3+ CD103+ (A–C), CD3+ CD103− (D–F) and CD11c− CD103+ (D–F). Scale bars indicate the magnification.

In summary, HIV-1 and SIV induced intraepithelial recruitment of both CD11c+ Mφs and DCs co-expressing the CX3CR1 receptor, which might act as an initial target of the virus. Neither virus induced recruitment of tolerogenic CD11c+ CD103+ DCs. We also confirmed previous observations that CD3+ lymphocytes are strategically localized to favor viral transmission to the lamina propria (McElrath et al., 2013).

In vivo challenge with SIVmac251 induces a rapid shift from CX3CR1high to CX3CR1low phagocytes in the macaque colonic mucosa

We investigated whether the identified MNP subsets populating the colonic lamina propria are differentially affected in the early phases of SIV infection by performing an in vivo study. Three cynomolgus macaques were intrarectally exposed to SIVmac251 and euthanized 72h later. MNP subset migration patterns and relative cell frequency were assessed by confocal microscopy and flow cytometry, respectively, and compared with those of ten untreated animals (those characterized earlier).

Analogous to what we observed in ex vivo-treated tissues, we detected the presence of intraepithelial MNPs, specifically CD11c+ CD64+ and CD11c+ CX3CR1+ cells (Figures 7A–7C). Intraepithelial CD11c+ CD103+ cells were not observed. However, in contrast to our ex vivo observations after 30 min of viral exposure, in vivo, after 72h of viral exposure, the intraepithelial cells were mostly found in the crypts and rarely inside the luminal epithelium.

Figure 7

Dynamics of colonic mononuclear phagocyte subsets in cynomolgus macaques three days after in vivo rectal challenge with SIVmac251

(A–I) Colonic tissues from macaques collected after three days of exposure to SIV, fixed, cryopreserved, and subjected either to immunolabeling for confocal microscopy (A–C) or to collagenase digestion for immune-cell phenotyping by flow cytometry (D–I) are presented. (A–C) Labeling by anti-CD11c-FITC (green) in combination with either CD64 (A), CD103 (B) antibodies followed by goat-anti-mouse Alexa Fluor 594 (red), or CX3CR1-PE (C) was performed. Arrows point to subsets of CD11c+ CD64+ (A) and CD11c+ CX3CR1+ (C) cells that migrate from the lamina propria to in between the epithelial cells following viral exposure. DAPI (blue) was used to stain the nuclei. The bright field image (BF, gray) was superimposed with the fluorescence images. Scale bars indicate the magnification. (D) Comparable percentages of CD11c+ CD64+ Mφs, CD11c− CD64+ Mφs, and CD11c+ CD64− DCs within the total number of CD45+ cells between SIV exposed (n = 3, open symbols) and not exposed (n = 10, filled symbols) macaques are shown. The relative proportion of CX3CR1high and CX3CR1low cells within the CD11c+ CD64+ Mφs (E) and CD11c− CD64+ Mφs (F); CX3CR1+, CD103+, and CD103- within CD11c+ CD64− DCs (G); CD14+ cells within CD11c+ CX3CR1high and CD11c + CX3CR1low (H); and CD11c− CX3CR1high and CD11c− CX3CR1low (I) cell populations are plotted. The mean ± the SD is plotted. The Mann-Whitney test was used. p-values < 0.05 were considered significant (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001).

As expected, we did not observe overt signs of systemic infection at such an early time point. Virus remained undetectable in blood of all except one animal, which had 227 copies/mL. The number of colonic lamina propria CD4+ and CD8+ lymphocytes in SIV-exposed macaques was comparable to controls (Figures S4A and S4B), whereas, CD11c− CD123+ plasmacytoid DCs were significantly depleted (0.07 ± 0.05 versus 0.007 ± 0.003%, p = 0.0280, Mann-Whitney test, Figure S4C).

The total number of lamina propria DCs and Mφs (both CD11c+ and CD11c− subsets) did not change following viral exposure (Figure 7D), but the relative proportions of the various MNP subsets did. Specifically, there was a significant lower proportion of CX3CR1high CD11c+ and CD11c− Mφs in SIV-exposed animals compared with controls (94.5 ± 5.9% versus 57.4 ± 4.5% p = 0.007 and 85.4 ± 13.1 versus 50.4 ± 6.8% p = 0.0140, respectively) (Figures 7E, 7F, and S5). These changes were paralleled by a statistically significant higher proportion of CX3CR1low cells in both subsets in exposed versus unexposed animals (4.2 ± 3.7 vs 36.4 ± 2.8% p = 0.007 for the CD11c+ subset and 8.7 ± 7.4 vs. 34.5 ± 4.8% for the CD11c− subset, p = 0.007) (Figures 7E and 7F), suggesting an accumulation of pro-inflammatory Mφs. We also observed a decrease in the proportion of CX3CR1+ DCs (p = 00,088), whereas that of both CD103+ and CD103− DC subsets did not change (Figures 7G and S5). No changes in the total number of DCs or Mφs or in the relative proportion of cell subsets were observed in the draining colon lymph nodes (Figure S6). However, lymph nodes from SIV-exposed macaques showed a higher, although not statistically significant, proportion of CX3CR1+ DCs (0.9 ± 0.5% vs 1.5 ± 0.9%, Figure S6D), suggesting possible specific recruitment of these cells from the intestinal mucosa.

We found 15.4 ± 7.6% of CD11c− and 6.6 ± 3.9% of CD11c+ colonic macaque Mφs being CD14+ CD16− cells in the steady-state condition (Figure S7) and thus, likely unresponsive to LPS/bacteria stimulation, as previously described in the human intestine (Caër and Wick, 2020; P. Smith et al., 1997a, 1997b; Smythies et al., 2005). Of note, we observed a significant increase in the proportion of CD14+ cells among CD11c+ CX3CR1low Mφs in SIV-exposed animals (7.61 ± 3.9 at 3 dpi versus 5.7 ± 15%, p = 0.0350, Mann-Whitney test, Figure 7H), substantiating the hypothesis of a proinflammatory phenotype of the CX3CR1low cells. No significant changes were, however, observed in the CD11c− CX3CR1low subset (Figure 7I), although cells from one of three animals showed a strong increase in CD14 expression.

Overall, our results indicate that the CX3CR1+ MNP populations of the colonic mucosa were rapidly and profoundly affected by in vivo exposure to SIV, before evident changes of the CD103+ cells.

Discussion

Here, we investigated the role that colorectal lamina propria MNP populations play in HIV-1 and SIV transmission. We began by thoroughly identifying and comparing the MNPs present in the colorectal mucosa of humans and cynomolgus macaques, determining five main MNP subsets in both species: three populations of CD64− DCs (tolerogenic CD103+, nontolerogenic CD103-, and CX3CR1+ DC) and two of CD64+ Mφs (immature CD11c+ and mature CD11c−, both expressing different levels of CX3CR1). The distinction between DC and Mφ phenotypes is challenging due to a proportion of intestinal Mφs expressing CD11c and high levels of MHC class II, markers classically used to identify DCs (Cerovic et al., 2014; Joeris et al., 2017). Moreover, the differentiation between intestinal DCs and Mφs exclusively based on the expression of CD103 and CX3CR1 has been proven to be incorrect, as not all intestinal DCs express CD103 and some also express CX3CR1 (Bujko et al., 2018; Persson et al., 2013). We found that CX3CR1 is expressed at higher levels on Mφs than DCs and that CD64 is specifically expressed by Mφs. These results confirm previously reported human data (Bernardo et al., 2016; Bujko et al., 2018; Persson et al., 2013) and extend these observations to nonhuman primates, in which such a detailed characterization of MNP subsets is novel.

Our study led to an alignment of the analytical strategy for human and macaque intestinal MNP characterization, highlighting both similarities and differences between species. We thus found that CD64, CD103, and CX3CR1 expression are a shared feature of intestinal MNPs, whereas immature CD11c+ Mφs were relatively more abundant than mature Mφs in the human mucosa but rare in the macaque mucosa, which is mostly populated by mature CD11c− Mφs. CD11c+ monocytes have been reported to be recruited to the intestine to undergo conditioning by the microenvironment, leading to an anti-inflammatory and anergic state and differentiation toward CD11c− Mφs (Bujko et al., 2018). The observed interspecies difference may thus reflect different turnover of intestinal Mφs from blood monocytes.

A major finding of our study was the fundamental difference between the CX3CR1+ and CD103+ subsets in their ability to respond to ex vivo viral exposure. Our data show that CD11c+ CX3CR1+ cells, including CD64− CD11c+ DCs and CD64+ CD11c+ Mφs, are those devoted to extend trans-epithelial dendrites or migrate across the tight intestinal epithelium in response to viral stimulation. Thus, here we confirm our previous results but further refine the phenotypic characteristics of these CD11c+ MNPs capable of capturing luminal virions and transfer them to CD4+ T lymphocytes (Cavarelli et al., 2013). The extension of protrusions by migrating cells involved the interaction between their CCR5 receptor and the viral gp120 envelope protein (Cavarelli et al., 2013). Here, we show that both human and macaque CD11c+ CX3CR1+ MNPs expressed higher levels of CCR5 than CD11c+ CD103+ cells, consistent with the observation that the latter ones did not migrate into the epithelium after viral stimulation. The CD11c+ CX3CR1+ MNPs included both DCs and Mφs. Further analysis, including CD64 detection, indicated that macaque CD64+ Mφs showed higher CCR5 levels than CD64− DCs. Whether the predominance of CX3CR1+ Mφs over CX3CR1+ DC and the higher CCR5 expression might favor Mφs recruitment to the epithelium is puzzling. Nevertheless, CX3CR1+ DC were detected at intraepithelial level ex vivo in both species, and the same migration pattern was observed in macaques exposed to SIV in vivo.

Given their long lifespan, intestinal Mφs play a key role in HIV-1 persistence. Indeed, although less susceptible to HIV-1 and SIV infection than vaginal Mφs or memory CD4 T cells, intestinal Mφs facilitate HIV-1 entry and carry viral DNA (Moore et al., 2012; Shen et al., 2009), even in HAART-treated patients (Zalar et al., 2010), are resistant to the cytopathic effects of HIV-1 infection/replication, and are able to disseminate the virus (Coleman and Wu, 2009; Ho et al., 1986; Kumar et al., 2014). It is tempting to speculate that the CD11c+ CX3CR1+ Mφs that capture the luminal virions act as viral reservoirs, as recently demonstrated in urethral macrophages (Ganor et al., 2019).

During homeostasis, circulating CD14+ monocytes enter the mucosa to replenish Mφ populations and differentiate into functionally plastic CD11c+ immature cells that rapidly respond to pathogens through phagocytosis and cytokine secretion without inducing inflammation (Bain and Mowat, 2011; Caër and Wick, 2020). It is feasible that HIV-1/SIV not only hijack this response by using CD11c+ cells as carriers but also exploit the sub-epithelial position of CD11c+ CX3CR1+ MNPs to gain access to the intestinal lamina propria. Our observation that such a mechanism is favored by CD11c+ CX3CR1+ MNPs but not CD11c+ CD103+ DCs has profound implications on the successive events that lead to dissemination of the virus. It has been reported that CX3CR1+ Mφs do not migrate to draining lymph nodes, whereas they may transfer antigens to migrating DCs in humans and in animal models (Bain and Mowat, 2011; Myoungsoo et al., 2018). Consistent with these data, we did not observe a reduction in the total number of colonic Mφs or CX3CR1+ Mφs accumulation in the draining lymph nodes, which suggest that Mφs remained in the lamina propria. In contrast to the behavior of CX3CR1+ Mφs, the number of CX3CR1+ DCs appeared to significantly decrease in the lamina propria and showed a tendency to accumulation in the draining lymph nodes, where they likely facilitate infection of CD4+ T cell.

Our findings suggest that the involvement of CX3CR1+ Mφs in HIV pathogenesis goes beyond their ability to extend protrusions. Indeed, we identified two populations of Mφs that express CX3CR1 at different levels, both involved in the early events of HIV-1/SIV transmission. The CX3CR1high versus CX3CR1low subsets have never been investigated in the macaque intestine, although they have been described in the mouse (Bain et al., 2012; Desalegn and Pabst, 2019; Kayama et al., 2012) and, to a lesser extent, in the human intestine (Bernardo et al., 2018). In addition, no other studies examined their role in HIV-1 or SIV transmission. In accordance with the literature, we show that the CX3CR1high subset, known to play an anti-inflammatory role during intestinal homeostasis (Bain et al., 2012; Desalegn and Pabst, 2019; Mowat and Agace, 2014; Weber et al., 2011), predominates in the mucosa at steady-state. Of note, we observed a high prevalence of CX3CR1low Mφs in SIV-exposed macaques in vivo. This observation is consistent with the rapid and abundant accumulation of inflammatory CD14+ CD11c+ monocytes differentiating into Mφs, but expressing low levels of CX3CR1, which occurs during intestinal inflammatory events (Bain et al., 2014; Grainger et al., 2017; Weber et al., 2011; Zigmond et al., 2014). Such CX3CR1low Mφs were shown to be highly responsive to stimulation by microbial antigens, subsequently producing pro-inflammatory cytokines (see Li et al., 2021 for a recent review). It has been proposed that the gut microenvironment is unable to commit infiltrating monocytes and differentiating Mφs toward an anti-inflammatory phenotype under pathological conditions (Hine and Loke, 2019). Conflicting results have been reported concerning CD14 expression by intestinal Mφs, which may possibly reflect varying conditions under which tissues were obtained. A number of studies described colonic Mφs being mostly CD14− cells (Caër and Wick, 2020; P. Smith et al., 1997a, 1997b; Smythies et al., 2005), whereas other groups detected an abundant population of CD14+ Mφs (Bernardo et al., 2016, 2018; Bujko et al., 2018), which include tissue resident autofluorescent Mφs (Doyle et al., 2021; Haniffa et al., 2009). Here we show that CD14− Mφs outnumbered the CD14+ ones in the macaque colonic mucosa at steady-state. An accumulation of pro-inflammatory CD14+ cells has been reported in the colon of AIDS patients (Cassol et al., 2015), and the inflammatory Mφs in HIV-1 infection have been shown to be associated with increased inflammation and tissue damage (Porcheray et al., 2006). Extending these observations, here, we show that CD14 expression specifically increases in CD11c+ CX3CR1low cells in the earliest phase of SIV infection, likely identifying cells of recent gut recruitment. Further studies, possibly including gene expression profiling of FACS-sorted CX3CR1high and CX3CR1low cells, are needed to unravel their functional properties in the context of HIV-1 mucosal infection and to provide useful information for therapeutic intervention for mucosal HIV-1 transmission. Previous studies in humans have indicated that intestinal damage and inflammatory changes begin soon after initial HIV-1 exposure (Appay and Sauce, 2008; Hunt et al., 2014). In nonhuman primate models, the epithelial damage occurs 3 to 14 days post-SIV infection but prior to widespread immune dysfunction (e.g. Th17 depletion) (Hensley-McBain et al., 2018). The loss of CX3CR1high MNPs in our model was observed three days post-SIV exposure and before CD4+ T cell depletion from the colonic lamina propria. Given that CX3CR1 is a gatekeeper for intestinal barrier integrity (Medina-Contreras et al., 2011; Schneider et al., 2015), our results highlight CX3CR1+ MNPs as new key players in the pathogenesis of acute HIV-1 infection, possibly contributing to breakdown of the intestinal barrier. Delineating the role of MNP subsets during viral transmission will lead to the fine-tuned characterization of host-virus interactions during the earliest events of colorectal HIV-1 transmission, important for the design of therapeutic approaches. Moreover, it will be of interest to investigate whether the intestinal accumulation of CX3CR1low Mφs increases during chronic HIV infection and whether antiretroviral therapy can restore the CX3CR1high phenotype, similarly to therapies resolving intestinal inflammation.

Limitations of the study

This study has some inherent limitations related to the human and macaque populations. Specifically, the human tissues were obtained from patients undergoing colonic ablation because of cancer, and although the fragments used were morphologically preserved, we cannot fully exclude the persistence of residual inflammation, which could explain the higher abundance of immature CD11c+ Mφs in the human mucosa samples than in nonhuman primates.

We have also provided a proof of concept showing that as soon as three days after in vivo viral exposure, the population of colonic CX3CR1high Mφs is partially replaced by cells expressing low levels of the receptor. Further studies in larger cohorts are warranted to confirm the observed phenotypic shift. In addition, the investigation of the functional profile of CX3CR1high versus CX3CR1low Mφs in homeostatic and infection conditions, while overcoming the purposes of the present study, represents an envisaged future development.

STAR★Methods

Key resources table

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact Mariangela Cavarelli (mariangela.cavarelli@cea.fr).

Materials availability

This study did not generate new unique reagents. The authors declare that all data supporting the findings of this study are available within the article and its supplemental information files or are available from the authors upon request.

Experimental model and subject details

Ethical statement and animal care

The study protocol for human mucosal tissues was approved by the institutional Ethics committee of IRCCS Ospedale San Raffaele (MUCHIV protocol Ethical Committee approval #35 06/02/2014 and extension #40 09/11/2015). Patients were followed at the Department of General and Gastrosurgery, IRCCS Ospedale San Raffaele, and signed the informed consent form.

Cynomolgus macaques (Macaca fascicularis), originating from Mauritian AAALAC certified breeding centers, were used in this study. All animals were housed in the IDMIT facilities (CEA, Fontenay-aux-Roses) under BSL-3 containment (Animal facility authorization #D92-032-02, Préfecture des Hauts de Seine, France) and in compliance with European Directive 2010/63/EU, French regulations, and the Standards for Human Care and Use of Laboratory Animals. All work related to animals was conducted in compliance with institutional guidelines and protocols approved by the local ethics committee “Comité d’Ethique en Expérimentation Animale du Commissariat à l’Energie Atomique et aux Energies Alternatives” (CEtEA #44). The study was authorized by the “Research, Innovation and Education Ministry” under registration number APAFIS#6058-20 1607121 0017243. Sigmoid colon and draining colon lymph nodes were collected during animal necropsies.

Study design and sample utilization

Human colonic samples were obtained from 20 patients with invasive colon cancer but without HIV-1 infection (55% females, age 63.9 ± 13.3 years [mean ± standard deviation], age range 44 to 89 years). Sections of sigmoid or descending (close to the sigmoid) colon from morphologically normal tissue (distal from the site of the pathology) were obtained from leftover surgical tissue of intestinal resections. Among them, 17 samples were used to carry out immune-cell phenotyping (panel for lamina propria mononuclear cells, summarized in Table S1) and three for ex vivo virus exposure experiments and confocal microscopy analysis.

Twenty-one cynomolgus macaques (67% females, age 4.7 ± 1 years [mean ± standard deviation], weight range 3.8 to 7.5 kg), including 18 uninfected controls and three animals exposed to SIVmac251 in vivo and sacrificed three days later, were used. All animals were negative for antibody responses to SIV, simian retrovirus type D (SRV), and simian T-cell lymphotropic virus (STLV) at the beginning of the study. Eighteen colon samples (n = 15 SIV- and n = 3 SIV-exposed) and six lymph nodes draining the colon (LN, n = 3 SIV- and n = 3 SIV-exposed) were used to carry out immune-cell phenotyping (Table S1). Colon tissue from 3 SIV- donors was used for ex vivo virus exposure experiments and confocal microscopy analysis.

Method details

Virus production

HIV-1 AD8 infectious molecular clone was produced by transfection of 293T cells with the pAD8 plasmid using Fugene 6 (Roche, Indianapolis, IN, USA), as previously described (Cavarelli et al., 2013). The SIVmac251 viral isolate used for the in vivo experiments was kindly provided by Dr A.M. Aubertin (Université Louis Pasteur, Strasbourg, France). The titer, determined after intra-rectal inoculation, was 100 AID50/mL. For ex vivo explant experiments, the SIVmac251 isolate was expanded in CEMx174 cells grown in RPMI medium. Cells (30 × 106) were incubated at a multiplicity of infection (MOI) of 10−2 for 2 h at 37°C and the volume adjusted to 1 × 106 cells/mL. Cultures were maintained for 21 days, changing half of the cells every 3 to 4 days. Viral replication was monitored by measuring p27 levels in culture supernatants by ELISA (RETRO-TEK SIV p27 Ag ELISA kit, Helvetica Health Care). Cell-free viral stock was passed through a 0.2-μm pore-size filter, aliquoted, and stored at −80°C.

Isolation and phenotyping of lamina propria mononuclear and lymph-node cells

Macaque draining colon lymph-node cells were obtained by mechanical dissociation. To isolate lamina propria mononuclear cells (LPMCs), both human and macaque sigmoid colons were cut into small pieces and incubated for 20 min at 37°C in HBSS medium without Ca++/Mg++ (Fisher Scientific, Illkirch, France) supplemented with 5 mM EDTA and 1 mM DTT (Sigma-Aldrich, St Quantin Fallavier, France) to eliminate mucus and epithelial cells. After washing in PBS, the tissue was incubated for 1 h at 37°C with agitation in HBSS medium with Ca++/Mg++ (Fisher Scientific, Illkirch, France) containing collagenase type VIII (0.25 mg/mL, Sigma Aldrich, St Quantin Fallavier, France) and DNase (5 U/mL, Roche, Mannheim, Germany). Undigested pieces were submitted to a second digestion for 30 min. Cell suspensions from lymph nodes or colon were filtered through 70-mm sterile nylon cell strainers (BD Biosciences), washed with complete medium (RPMI supplemented with 10% FCS, 100 U/mL penicillin/streptomycin, 1% glutamine, 1% NEAA, 1% Na-pyruvate, 1% HEPES buffer [1 M]; all from Fisher Scientific, Illkirch, France) and analyzed by flow cytometry. Antibodies used for the staining of human and macaque cells are listed in Table S2. Human cell phenotypes were acquired on a Navios Ex flow cytometer (Beckman Coulters) at Ospedale San Raffaele, and macaque cell phenotypes on an LSR Fortessa (BD Biosciences) at the IDMIT institute. All results were analyzed using FlowJo 9.8.3 (Tristar, USA) software within the singlet viable fraction. Positive and negative gating was set using the fluorescence minus one (FMO) method.

Polarized ex vivo explant culture

Human colonic fragments and macaque sigmoid colon were collected in cold PBS supplemented with 100 U/mL penicillin/streptomycin and 50 mg/mL gentamicin and then processed for ex vivo culture within 30 to 45 min from excision, as previously described (Cavarelli et al., 2013, 2021). Briefly, after abundant washing in the same solution, the specimens were transferred into complete medium and the epithelial surface exposed. For polarized ex vivo culture, tissue explants (diameter 8.0 mm), including the epithelium and sub-mucosa, were cut out with a biopsy punch (Stiefel, Laboratories, Inc. North Carolina) and placed on a sponge support (area = 2.25 cm2, BioOptica, Milan, Italy) with the sub-mucosa facing the sponge. A polystyrene cylinder (I.D. x H 4.7 × 8 mm, Sigma-Aldrich, St. Louis, MO) was sealed with veterinary glue (3M Vetbond, St. Paul, MN) onto the apical surface of the mucosa to avoid leakage. Specimens were placed in a 60-mm center-well organ culture dish (BD Falcon, San Diego, CA) containing 1 mL complete medium.

Mucosal explant cultures were treated apically with 100 μL viral culture supernatant (HIV-1 AD8, SIVmac251) or complete medium as a negative control. After 30 min, the apical stimulus was halted, the tissues washed six times with PBS, and then fixed in 4% paraformaldehyde (PFA) for 4 hours at 4°C, cryoprotected in 10% sucrose over-night at 4°C, embedded in optimal cutting temperature embedding medium (OCT compound, VWR, Milan, Italy), and snap-frozen. One explant was fixed before the establishment of the tissue culture to serve as a time zero control.

Histopathology analysis, immunofluorescence markers, and confocal microscopy

Tissue preservation was verified under a light microscope (Nikon Eclipse 80i) for all tissues on at least three 5-μm thick cryosections taken 100 μm apart and stained with Hematoxylin/Eosin. For confocal microscopy analysis of colonic tissues, at least 5 to 6 serially cut 10-μm thick sections, sampled 100 μm apart, were obtained. The antibodies used for confocal microscopy phenotyping are listed in Table S3. Primary antibodies were used at a final concentration of 10 μg/mL and were omitted from the methodological negative control sections. Single-channel images from Z-series were collected from at least three representative fields/labelling/sample using a Leica TCS SP8 confocal microscope (Leica Microsystems GmbH, Wetzlar Germany) and the images processed using Image J 1.49v and PhotoshopCS software. Human and macaque tissues were independently processed at the IRCCS Ospedale San Raffaele and at the IDMIT center, respectively.

In vivo infection, sample collection, and quantification of blood viral load of cynomolgus macaques

Three adults male cynomolgus macaques were intrarectally exposed to 50 animal infectious doses 50 (50 AID50) of SIVmac251 isolate and sacrificed three days later by intravenous injection of 180 mg/kg sodium pentobarbital. Blood samples were collected twice in BD VacutainerR Plus Plastic K3EDTA tubes for plasma viral load quantification, i.e., before viral exposure from sedated animals following 5 mg/kg intra-muscular injection of ZoletilR 100 (Virbac, Carros, France) and at necropsy. The sigmoid colon and draining colon lymph nodes were collected at necropsy. Punches of the colonic tissue (6 mm in diameter, n = 2/donor) were immediately fixed in 4% PFA, processed as above, and snap-frozen. The remaining tissue was subjected to collagenase digestion, as described above.

Plasma was isolated from EDTA blood by centrifugation for 10 min at 1,500 × g and cryopreserved at −80°C. Blood viral RNA was obtained from 250 μL cell-free plasma using the Nucleospin 96 RNA kit (Macherey Nagel GmbH&Co KG, Düren, Germany), according to the manufacturer’s instructions. Retro-transcription and cDNA amplification and quantification were performed in duplicate by RT-qPCR using the Superscript III Platinum one-step quantitative RT-PCR system (Invitrogen, Carlsbad, USA). RT-PCR was performed as previously described (Cavarelli et al., 2021). The quantification limit (QL) was estimated to be 111 copies/mL and the detection limit (DL) 37 copies/mL.

Quantification and statistical analysis

All data visualization and statistical analyses were carried out using Prism v9.2.0 software (GraphPad software, La Jolla, USA). The statistical significance of differences between two groups was tested using Mann-Whitney U tests or Wilcoxon signed rank tests. The statistical significance of differences between more than two groups was tested using Friedman tests and p values were corrected for multiple comparisons using the Benjamini, Krieger, and Yekutieli FDR approach. p values ≤0.05 for two-tailed tests were considered significant, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
CD45-V500	BD	Cat# 560779; RRID: AB_1937332
CD45-PerCp	BD	Cat# 558411; RRID: AB_397080
CD3-PER-CP-Cy5.5	BD	Cat# 560835; RRID: AB_2033956
CD3-V500	BD	Cat# 560770; RRID: AB_1937322
CD8-BV650	BD	Cat# 563821; RRID: AB_2744462
CD14-Alexafluor700	BD	Cat# 557923; RRID: AB_396944
CD16-PER-CP-Cy5.5	BD	Cat# 560717; RRID: AB_1727434
CD16-PE-CF594	BD	Cat# 562293; RRID: AB_11151916
CD19-PER-CP-Cy5.5	BD	Cat# 332780; RRID: AB_2868629
CD20-BV711	BD	Cat# 563126; RRID: AB_2313579
CD56-PER-CP-Cy5.5	BD	Cat# 560842; RRID: AB_2033964
HLA-DR APC	BD	Cat# 559866; RRID: AB_398674
HLA-DR APC-H7	BD	Cat# 561358; RRID: AB_10611876
CD11c-PB	Biolegend	Cat# 301625; RRID: AB_10662901
CD11c-APC	BD	Cat# 333144; RRID: AB_2868645
CD123-PE-Cy7	BD	Cat# 560826; RRID: AB_10563898
CD64-PE-CF594	BD	Cat# 565389; RRID: AB_2739213
CD64-V450	BD	Cat# 561202; RRID: AB_10564066
CD103-PE	ebioscience	Cat# 12-1038; RRID: AB_11150242
CX3CR1-FITC	Biolegend	Cat# 341605; RRID: AB_1626276
CXCR4-PE	BD	Cat# 555974; RRID: AB_396267
CXCR4-PECF594	BD	Cat# 562389; RRID: AB_11153132
CCR5-FITC	BD	Cat# 555992; RRID: AB_396278
CCR5-BV421	BD	Cat# 565000; RRID: AB_2739038
CD3	BD	Cat# 551916; RRID: AB_394293
CD11c-FITC	Invitrogen	Cat# MHCD11c01; RRID: AB_10373970
CD64	BD	Cat# 555525; RRID: AB_395911
CD103	AbDSerotec	Cat# MCA708T; RRID: AB_1100549
CX3CR1-PE	Biolegend	Cat# 355704; RRID: AB_2561681
Bacterial and virus strains
HIV-1 AD8	Cavarelli et al., 2013 ; https://doi.org/10.1002/emmm.201202232	N/A
SIVmac251 isolate	Cavarelli et al., 2013; https://doi.org/10.1038/s42003-021-02383-9.	N/A
Biological samples
Human colon samples	IRCCS Ospedale San Raffaele, Milan, Italy	https://www.hsr.it
Macaque colon samples	ICEA, MVA-HB/IDMIT, Fonteney-aux-Roses, France	https://www.idmitcenter.fr
Critical commercial assays
Superscript III Platinum one-step quantitative RT-PCR system	Invitrogen	11732088
RETRO-TEK SIV p27 Ag ELISA kit	Helvetica Health Care	0801169
Software and algorithms
Prism	GraphPad	https://www.graphpad.com
BD FACSDiva	BD	https://www.bdbiosciences.com
FlowJo	BD	https://www.flowjo.com
Image J	NIH	https://imagej.nih.gov/
PhotoshopCS	Adobe	https://www.adobe.com/