|Home | About | Journals | Submit | Contact Us | Français|
Depletion of CD4+ T cells from the gut occurs rapidly during acute HIV-1 infection. This has been linked to systemic inflammation and disease progression as a result of translocation of microbial products from the gut lumen into the bloodstream. Combined antiretroviral therapy (cART) substantially restores CD4+ T cell numbers in peripheral blood, but the gut compartment remains largely depleted of such cells for poorly understood reasons. Here, we show that a lack of recruitment of CD4+ T cells to the gut could be involved in the incomplete mucosal immune reconstitution of cART-treated HIV-infected individuals. We investigated the trafficking of CD4+ T cells expressing the gut-homing receptors CCR9 and integrin α4β7 and found that many of these T cells remained in the circulation rather than repopulating the mucosa of the small intestine. This is likely because expression of the CCR9 ligand CCL25 was lower in the small intestine of HIV-infected individuals. The defective gut homing of CCR9+β7+ CD4+ T cells — a population that we found included most gut-homing Th17 cells, which have a critical role in mucosal immune defense — correlated with high plasma concentrations of markers of mucosal damage, microbial translocation, and systemic T cell activation. Our results thus describe alterations in CD4+ T cell homing to the gut that could prevent efficient mucosal immune reconstitution in HIV-infected individuals despite effective cART.
The immune responses to the antigens encountered along the intestinal mucosa surfaces are mainly initiated in inductive sites of the gut-associated lymphoid tissue (GALT), Peyer’s patches, and mesenteric lymph nodes. The lymphocytes primed in these secondary lymphoid organs then express high levels of gut-homing receptors, integrin α4β7 and CCR9, to subsequently direct their migration from the blood to the effector sites of the gut mucosa, the lamina propria and epithelium (1, 2). The ligand of α4β7 integrin, mucosal addressin cell adhesion molecule-1 (MAdCAM-1), is expressed by endothelial cells of the lamina propria and associated lymphoid tissues along the whole intestine (3). By contrast, the ligand of CCR9, the chemokine CCL25, is expressed only by small intestine endothelial and epithelial cells (4, 5). Thus, the combined expression of CCR9 and α4β7 delineates a T cell subset prone to migrate to the small intestine mucosa.
Most of the CD4+ T cells in the gut mucosa are activated effector memory cells that express the HIV-1 entry coreceptor CCR5, thus providing a large pool of HIV-1 target cells (6, 7). The CD4+ T cells in the gut mucosa are rapidly and deeply depleted during acute HIV-1 infection, due to the direct killing of target cells by the virus and bystander apoptosis (8, 9). The ability of HIV-1 envelope glycoprotein (gp120) to bind to and signal through α4β7 could contribute to the selective tropism of HIV-1 for CD4+ T cells in the gut mucosa (10, 11).
The α4β7hi CD4+ T cell subset includes most gut-homing Th17 cells (12). These specialized cells play a critical role in gut mucosal immune defense (13, 14). Their depletion in the gut mucosa after HIV-1 and pathogenic SIV infections could compromise the integrity of the gut mucosal barrier (15). The subsequent translocation of microbial products from the gut lumen into the bloodstream has been associated with systemic inflammation and disease progression in HIV-infected individuals and SIV-infected macaques (16–20). By contrast, the Th17 cell subset in the gut mucosa is selectively preserved in natural SIV hosts, like sooty mangabeys and African green monkeys (21). This could contribute to the absence of microbial translocation and systemic inflammation in nonpathogenic SIV infections (20). Thus, the restoration of an efficient mucosal immune barrier in HIV-infected individuals receiving combined antiretroviral therapy (cART) would be critical for reducing systemic inflammation.
Gut CD4+ T cells are restored much more slowly than those in the peripheral blood of treated HIV-infected individuals (22, 23). Many HIV-infected individuals still have effector sites of the gut mucosa that are severely lacking CD4+ T cells despite sustained effective cART, in marked contrast to the significant restoration of CD4+ T cells in their peripheral blood and immune inductive sites (24). This could be because HIV-1 replication is incompletely suppressed in the gut despite cART. A lack of recruitment of CD4+ T cells to the gut could also contribute to this incomplete restoration of CD4+ T cells in the gut mucosa. However, we know little of how CD4+ T cells migrate to repopulate the gut during immune reconstitution in response to cART.
We have therefore investigated CD4+ T cells trafficking between the blood and gut compartments in HIV-infected individuals on sustained effective cART, focusing on CCR9+α4β7hi CD4+ T cells as a traceable phenotype for cells that home to the small intestine mucosa.
We assessed CD4+ T cell reconstitution in the peripheral blood and small intestine mucosa of treated HIV-infected individuals by flow cytometry and immunohistochemistry. The frequency of CD4+ T cells remained lower in the peripheral blood of treated HIV-infected individuals than in uninfected controls (40.7% vs. 64.8%, P < 0.0001; Figure Figure1A).1A). CD4+ T cell depletion also persisted in the small intestine mucosa of HIV-infected individuals compared with that in uninfected controls, as analyzed by flow cytometry (median frequency of 28.9% vs. 48.4%, respectively, P = 0.002; Figure Figure1A)1A) and immunohistochemistry (median absolute number of 362 cells/mm2 vs. 660 cells/mm2, respectively, P = 0.0001; Figure Figure1B).1B). The frequency of CD4+ T cells in the gut was positively correlated with the frequency in the peripheral blood (correlation coefficient [ρ] = 0.67, P = 0.0001; Figure Figure1C).1C). Almost all the CD4+ T cells in the gut mucosa were memory CD45RO+CCR5+ CD4+ T cells, 88% of which had an effector phenotype (Supplemental Figure 1; supplemental material available online with this article; doi: 10.1172/JCI59011DS1).
Activated memory CCR5+ CD4+ T cells in the gut could be highly permissive to HIV-1 infection. We therefore quantified the HIV-1 DNA and RNA in samples of jejunum mucosa. The median HIV-1 DNA load in CD4+ T cells in the gut mucosa was 6-times greater than that in CD4+ T cells in the peripheral blood (P = 0.013; Figure Figure1D),1D), and their numbers in the 2 compartments were positively correlated (ρ = 0.54, P = 0.013; data not shown). HIV-1 RNA was also detected in the jejunum mucosa of all the individuals tested (n = 16), albeit at low frequency (mean of 2.5 copies per mg tissue; data not shown).
The magnitude of CD4+ T cell restoration in the small intestine mucosa of patients on cART varied greatly from one individual to another. Good and poor immunological responders were divided on the basis of whether the percentage of CD4+ T cells in their gut mucosa was above or below the median. The good immunological responders had lower gut DNA viral loads than the poor immunological responders (P = 0.038). They also had a higher frequency of CD4+ T cells (P = 0.019) and a lower frequency of activated HLA-DR+ CD8+ T cells (P = 0.023) in the blood than the poor immunological responders (data not shown).
We assessed CD4+ T cells trafficking between the peripheral blood and small intestine mucosa by measuring CD4+ T cells expressing gut-homing receptors, α4β7 and CCR9, in the 2 compartments (Figure (Figure2A2A and Supplemental Figure 2, A–C). The frequencies of CCR9+β7hi in CD4+ T cells in the peripheral blood and small intestine mucosa were inversely correlated in HIV-infected individuals (ρ = –0.53, P = 0.017) and in uninfected controls (ρ = –0.75, P = 0.019; Figure Figure2B).2B). But the distribution pattern of CCR9+β7hi CD4+ T cells between the blood and gut compartments of HIV-infected individuals was the inverse of that in uninfected controls. There were more CCR9+β7hi in CD4+ T cells in the peripheral blood (P = 0.002) and fewer cells in the small intestine mucosa (P = 0.014) of HIV-infected individuals than in those of the controls (Figure (Figure2C).2C). Thus, there appear to be significantly more CCR9+β7hi CD4+ T cells circulating in HIV-infected individuals than are localizing in the gut mucosa, in contrast to those in uninfected controls.
The chemoattraction of CCR9+ lymphocytes is governed by their ligand, CCL25 (5). We therefore assessed whether the abnormal distribution of CCR9+β7hi CD4+ T cells between the blood and gut compartments of HIV-infected individuals could be linked to differences in CCL25 expression in the small intestine mucosa. We measured CCL25 mRNA in small intestine epithelial cells using real-time RT-PCR. The median amount of CCL25 mRNA in HIV-infected individuals was 5.4-times lower than that in uninfected controls (P = 0.0001; Figure Figure3A).3A). We also confirmed the reduced expression of the CCL25 chemokine in the small intestine mucosa of HIV-infected individuals by immunohistochemistry on jejunal tissue (automatic quantification of CCL25 per unit surface area, P = 0.006; Figure Figure3B).3B). It was correlated with the amount of CCL25 mRNA (ρ = 0.45, P = 0.026; data not shown). The homing of CCR9+β7hi CD4+ T cells to the small intestine mucosa appears to be mainly driven by the CCL25-CCR9 axis. The amount of CCL25 mRNA was positively correlated with the frequency of CCR9+β7hi in CD4+ T cells in the small intestine mucosa (ρ = 0.45, P = 0.026, Figure Figure3C)3C) and conversely negatively correlated with the frequency of CCR9+β7hi in CD4+ T cells in the peripheral blood (ρ = –0.63, P = 0.0008; Figure Figure3D).3D). The same positive and negative correlations were obtained when we measured CCL25 using immunohistochemistry (ρ = 0.35, P = 0.064 and ρ = –0.46, P = 0.013, respectively; data not shown). These results strongly suggest that a reduced expression of the chemokine CCL25 in the small intestine mucosa is involved in the altered homing of CCR9+β7hi CD4+ T cells to the gut of HIV-infected individuals.
In contrast to the depletion of CD4+ T cells observed in the jejunum mucosa, the absolute numbers of CD8+ T cells in HIV-infected individuals (783 cells/mm2) and healthy controls (797 cells/mm2) were similar (P = 0.673). Thus, the defect in the CCL25-CCR9 axis seems to affect the CD4+ and CD8+ T cell populations differently in HIV-infected individuals. We found that CD8+ T cells in the jejunum mucosa expressed much higher levels of β7 integrin than the CD4+ T cells (P < 0.0001; Supplemental Figure 3). We assessed the expression of MAdCAM-1 on endothelial cells in the lamina propria using immunohistochemistry. The amounts of MAdCAM-1 in the guts of HIV-infected individuals and healthy controls were similar (P = 0.493; Supplemental Figure 3). The preserved absolute numbers of CD8+ T cells, despite the altered levels of CCL25, could thus be due to CD8+ T cells being more dependent on β7 integrin than on CCR9 in their trafficking to the gut during HIV-1 infection.
Th17 cells play a critical role in the immune defense of the gut mucosa. Their depletion in the gut after HIV-1 infection could compromise the integrity of the gut mucosal barrier (15). The α4β7hi CD4+ T cell subset has been shown to harbor most of the gut-homing Th17 cells (12). We assessed the frequency of Th17 cells in the CCR9+ subset of α4β7hi CD4+ T cells in peripheral blood. The chemokine receptor CCR6 is a marker of Th17 lineage polarization that contributes to the Th17 migration to the gut (25–28). We therefore first measured the frequency of CCR6+ cells among CCR9+β7hi and CCR9–β7hi CD4+ T cell subsets using flow cytometry. There were more CCR6+ cells in CCR9+β7hi T cells than in CCR9–β7hi CD4+ T cells (P < 0.0001; Figure Figure4A).4A). We then looked directly for Th17 and Th1 cells in 5 HIV-infected individuals and 5 uninfected controls by measuring the frequency of IL-17– and IFN-γ–producing cells in the CCR9+β7hi and CCR9–β7hi CD4+ T cell subsets sorted by flow cytometry (Supplemental Figure 4). The CCR9+β7hi CD4+ T cell subset was richer in Th17 cells than the CCR9–β7hi subset (P = 0.006; Figure Figure4B),4B), while both subsets had similar frequencies of Th1 cells (data not shown). Thus, most gut-homing Th17 cells are in the CCR9+β7hi CD4+ T cell subset.
We found that a significant fraction of CCR9+α4β7hi CD4+ T cells remains circulating rather than repopulating the gut of HIV-infected individuals. We also showed that this abnormally large circulating cell subset contains most of the gut-homing Th17 cells, which play an important role in the immune defense of the gut mucosa. We therefore looked at the association between this defect in gut homing and any weakness in the gut mucosal barrier, microbial translocation, and increased T cell activation in HIV-infected individuals.
We evaluated enterocyte damage by measuring the plasma concentration of intestinal-fatty acid–binding protein (I-FABP), which leaks out of damaged small intestine epithelial cells (29). The median plasma concentration of I-FABP was higher in treated HIV-infected individuals (238.3 pg/ml) than that in uninfected controls (54.5 pg/ml, P = 0.038; Figure Figure5A).5A). The defective homing of CCR9+β7hi CD4+ T cells to the gut was associated with enterocyte damage, as the plasma I-FABP concentration increased as the frequency of CCR9+β7hi in CD4+ T cells in the gut mucosa decreased (ρ = –0.51, P = 0.005; data not shown).
We assessed the translocation of bacterial products through the gut mucosal barrier by measuring the plasma concentrations of LPS, a component of the cell wall of Gram-negative bacteria, and soluble CD14 (sCD14), a marker of monocyte activation after stimulation by LPS (30). The median plasma concentrations of LPS (4.29 pg/ml) and sCD14 (1560 ng/ml) were higher in treated HIV-infected individuals than they were in uninfected controls (3.69 pg/ml for LPS, P = 0.004; 860 ng/ml for sCD14, P = 0.005; Figure Figure5A),5A), suggesting that microbial translocation persists in HIV-infected individuals despite cART. Microbial translocation was associated with enterocyte damage, because the concentrations of sCD14 and I-FABP were positively correlated (ρ = 0.41, P = 0.027; data not shown). The correlations between the concentrations of LPS and sCD14 and the frequency of CCR9+β7hi in CD4+ T cells in the blood (ρ = 0.57, P = 0.002 for LPS; ρ = 0.48, P = 0.008 for sCD14; Figure Figure5B)5B) and the inverse correlations of these concentrations and the frequency of CCR9+β7hi in CD4+ T cells in the gut mucosa (ρ = –0.48, P = 0.011 for LPS; ρ = –0.48, P = 0.008 for sCD14; Figure Figure5C)5C) indicated that microbial translocation increased when more CCR9+β7hi CD4+ T cells were circulating rather than localizing in the gut. This was supported by our finding that the concentrations of LPS and sCD14 in the plasma were both inversely correlated with the amount of CCL25 mRNA in the small intestine mucosa (ρ = –0.58, P = 0.003 for LPS; ρ = –0.56, P = 0.004 for sCD14; Figure Figure5D). 5D).
Lastly, we examined the relationship between the dysregulated gut homing of CCR9+β7hi CD4+ T cells and an increase in T cell activation in HIV-infected individuals. The degree of CD4+ T cell activation in peripheral blood, measured by Ki67 expression, was strongly correlated with the plasma concentration of LPS in HIV-infected individuals (ρ = 0.69, P = 0.001; data not shown). Systemic T cell activation could be favored by the alterations in gut homing of CCR9+β7hi CD4+ T cells, because the frequency of Ki67+ in CD4+ T cells in the peripheral blood was positively correlated with the frequency of circulating CCR9+β7hi in CD4+ T cells (ρ = 0.64, P = 0.003; Figure Figure5E)5E) but inversely correlated with the amount of CCL25 mRNA in the small intestine mucosa (ρ = –0.61, P = 0.007; Figure Figure5F).5F). Thus our data strongly suggest that the defective homing of CCR9+β7hi CD4+ T cells to the gut in HIV-infected individuals is correlated with mucosal damage, microbial translocation, and systemic T cell activation.
We have described alterations in the homing of CD4+ T cells to the gut that could contribute to the lack of mucosal immune reconstitution in HIV-infected individuals despite cART. CD4+ T cells in the small intestine mucosa of the subjects studied remained depleted, despite the subjects being on effective cART for more than 5 years. This is in marked contrast to a substantial restoration of their blood CD4+ T cell count. The persistence of microbial translocation and systemic inflammation in these subjects underlines the importance of restoring an efficient mucosal immune barrier.
The immune inductive sites of the GALT are better reconstituted than effector sites in response to cART (24). But unlike the ileum and colon, the mucosa of the upper small intestine is devoid of lymphoid follicles. We have focused on CCR9+α4β7hi CD4+ T cells trafficking to the jejunum mucosa. This allows us to assess the immune reconstitution of the effector sites of the gut during cART, without bias due to the presence of immune inductive sites, as is the case for the ileum or colon mucosa.
We found that a significant proportion of CCR9+β7hi CD4+ T cells are in the circulation of HIV-infected individuals rather than localizing in the gut, as they are in uninfected controls. Similarly, the distribution of CCR9+ T cells between the blood and small intestine compartments appears to be the inverse in inflammatory bowel diseases, as in celiac and Crohn’s diseases, with the frequency of CCR9+ T cells in the peripheral blood being markedly increased and that in the small intestine mucosa being reduced (31, 32). It has been suggested that the preferential apoptosis of CCR9+ T cells or the downregulation of CCR9 expression after T cell activation in the small intestine mucosa may explain the reduced frequency of CCR9+ T cells in the gut mucosa during celiac or Crohn’s diseases. The ex vivo activation of T cells downregulates CCR9 expression (31, 32). However, we found no expansion of the CCR9–β7hi CD4+ T cell subset accompanying the reduction of the CCR9+β7hi CD4+ T cell subset in the gut of HIV-infected individuals (data not shown). Some downregulation of CCR9 on T cells could have occurred during tissue processing, but it cannot account for the difference we found between HIV-infected individuals and healthy individuals. The preferential activation-induced death of CCR9+ T cells could contribute to their depletion in the gut mucosa. But our data better support a mechanism involving a lack of recruitment of CCR9+β7hi CD4+ T cells to the gut, as the amount of CCL25 in the small intestine mucosa of HIV-infected individuals was much lower than that in uninfected controls. Moreover, the amount of CCL25 appears to be positively correlated with the frequency of CCR9+β7hi CD4+ T cells in the gut mucosa but inversely correlated with their frequency in the peripheral blood. The expression of CCL25 is altered in Crohn’s disease, with patchy increases in areas of lymphocyte infiltration but reduced amounts in inflamed and ulcerative areas of the small intestine mucosa (31). By contrast, we found a diffuse reduction in CCL25 in the jejunum mucosa of HIV-infected individuals. We detected CCL25 expression by immunohistochemistry in both the crypt and villous epithelial cells. Some published studies found more CCL25 in crypt epithelial cells (4, 5), while others found more CCL25 in villous cells (33). Ericsson et al. isolated epithelial cells from crypt and villous regions by laser capture microscopy and found high amounts of CCL25 mRNA in both crypt and villous enterocytes (34). A reduction in the amount of CCL25 mRNA has also been observed in macaques lymph nodes during SIV infection (35). Epithelial cells are the main source of CCL25 in the small intestine mucosa. The reduced expression of CCL25 could thus be due to persistent enterocyte damage in HIV-infected individuals despite cART. This, in turn, could impair the CCL25-mediated recruitment of CCR9+β7hi CD4+ T cells to the gut, setting up a vicious circle that prevents efficient mucosal reconstitution.
Gene expression profiles of gut mucosal tissue have revealed increased activity of the genes involved in inflammation and apoptosis in individuals with poor CD4+ T cell restoration in the gut, while the genes involved in mucosal repair and regeneration are more active in those having an efficient immune restoration (36). Residual HIV-1 replication in the gut despite cART could result in the intestine mucosa being persistently inflamed. Some virus proteins might have bystander effects, notably gp120 signaling through GPR15/Bob on the enterocytes could lead to their apoptosis (37, 38). Intraepithelial T cells bearing the Fas ligand could also induce Fas-mediated apoptosis of the enterocytes. The resulting damage to enterocytes could be responsible for the reduced expression of CCL25 in the small intestine mucosa and the associated defective homing of CD4+ T cells to the gut in HIV-infected individuals.
Other mechanisms are probably involved in this incomplete mucosal immune reconstitution, notably, in the colon, in which the CCL25-CCR9 axis does not play a major role. The chemokines CCL28, CCL20, and CXCL12 and the ligands of CXCR3 could play a role in the homing of T cells to the gut, but little is known regarding a potential dysfunction of this chemokine network in the setting of HIV infection (39, 40). The depletion of CD4+ T cells in treated HIV-infected individuals appears much more pronounced in the small intestine than in the colon (41). Thus, while the defect in the CCL25-CCR9 axis is probably not the sole mechanism involved — as it cannot explain the lack of immune reconstitution of the colonic mucosa — its impairment could contribute significantly to the particular depletion of CD4+ T cells observed in the small intestine mucosa.
The disrupted integrity of the gut mucosal barrier allows translocation of microbial products from the gut lumen into the bloodstream during HIV-1 infection (20, 42, 43), as in inflammatory bowel diseases (44). HIV-infected individuals have been reported to have increased plasma levels of I-FABP and LPS, reflecting ongoing enterocyte death and bacterial translocation. Further, high levels of sCD14, a marker of monocyte activation after stimulation by LPS, have been associated with an increased risk of mortality in HIV-1 infection (18). Th17 cells play a critical role in the immune defenses of the gut mucosa. We found that most gut-homing Th17 cells are within the CCR9+β7hi CD4+ T cell subset and that the defective homing of this subset to the gut is associated with microbial translocation. The lack of recruitment of Th17 cells to the gut could thus be involved in the persistence of a “leaky” intestinal mucosal barrier.
In conclusion, we find that the defective homing of CCR9+β7hi CD4+ T cells to the gut could impair mucosal immune reconstitution in treated HIV-infected individuals. A vicious circle could arise among mucosal inflammation, enterocyte damage, reduced CCL25 expression, and defective homing of CCR9+β7hi CD4+ T cells to the gut. The lack of recruitment of CCR9+β7hi CD4+ T cells to the gut was associated with persistent microbial translocation and systemic T cell activation despite cART. Monitoring the frequency of circulating CCR9+β7hi CD4+ T cells can provide a surrogate marker of poor immune reconstitution of the gut mucosa in treated HIV-infected individuals.
Twenty HIV-1–infected individuals and ten uninfected controls were enrolled in this study at the Toulouse University Hospital, Toulouse, France. All 20 HIV-1–infected individuals initiated cART at the chronic stage of infection, with a median nadir CD4+ T cell count of 185 cells/μl (interquartile range [IQR], 123–221 cells/μl). They had been on cART for a median of 66 months (IQR, 49–67 mo.). All had sustained plasma HIV-1 RNA of less than 20 copies per ml. Their median CD4+ T cell count was 668 cells/μl (IQR, 451–849 cells/μl) at the time of enrollment. Samples (80 ml) of peripheral blood were collected from each participant. Jejunal mucosa (8 biopsies) was obtained from each participant during upper endoscopy. All individuals were free of inflammatory or lymphoproliferative bowel diseases on histopathologic examination. One uninfected control was excluded because of a concomitant angiocholitis.
Jejunal biopsies (n = 5) were digested with 0.5 mg/ml collagenase type II-S (Sigma-Aldrich) in RPMI by incubation at 37°C, with shaking, for 4 periods of 30 minutes. The cells were then filtered through 70-μm gauze, and intestinal T lymphocytes were isolated by positive selection (EasySep Human CD3 Positive Selection Kit, Stemcell Technologies Inc.) and immediately processed.
RNA was extracted from one biopsy that had been snap-frozen on dry ice. Each sample was weighed and homogenized, and RNA was extracted (QIAamp RNeasy Mini Kit, Qiagen) with on-column DNase treatment (RNase-Free DNase Set, Qiagen).
HIV-1 DNA was quantified from sorted CD4+ T cells by real-time PCR on a LightCycler (Roche), as previously described (45). The HIV-1 RNA in gut tissue was quantified by real-time RT-PCR using the SuperScript III One-Step RT-PCR System (Invitrogen) on a LightCycler 480 (Roche). This assay, adapted from a previously published procedure (46), has a sensitivity of 1 copy per reaction. HIV-1 RNA is expressed as copies per milligram of gut tissue.
RNA was extracted from small intestine epithelial cells (CD3– fraction of the digested cell suspension) using the RNeasy Mini Kit (Qiagen). The primers were designed to amplify a fragment encompassing a spliced region of CCL25 mRNA. RT-PCR was performed on a LightCycler 480 (Roche). Data are given as fold increase in CCL25 mRNA normalized to the GAPDH control.
Fresh tissues were fixed in 4% neutral buffered formalin and embedded in paraffin. Sections (3 μm) were stained with hematoxylin and eosin. Immunohistochemistry was performed using anti-CD3 and CD4 (both from Novacastra); CD8 (Dako); CCL25 (R&D Systems); MAdCAM-1 (Serotec) mAb; and the appropriate secondary antibodies. Quantification was performed using LAS v3.7 (Leica microsystems) and NIS-element (Nikon).
Flow cytometry analyses were performed on a BD LSRII driven by the FACSDiva software (BD Biosciences). Intracellular staining was performed using Cytofix/Cytoperm (BD Biosciences). We used anti–CD3-PECy7, CD4-ECD, and CD45RA-FITC (all from Beckman-Coulter); CD8-Pacific blue, CD27-PECy5, CD45RO-PECy5, α4-PE, β7-APC, Ki67-FITC, and CCR5-PE (all from BD Biosciences); and CCR6-PE, CCR7-PE, and CCR9-FITC (all from R&D Systems) mAbs. Control experiments showed that more than 99% of peripheral CCR9+β7hi CD4+ T cells also expressed the α4 chain (Supplemental Figure 2A). The α4 and β7 chains were closely associated on CCR9+ CD4+ T cells, as demonstrated by FRET experiments (Supplemental Figure 2B). CCR9+β7hi CD4+ T cells were colabeled with the α4β7 mAb produced from the Act-1 clone (obtained through the NIH AIDS Research and Reference Reagent Program from A.A. Ansari, Emory University School of Medicine, Atlanta, Georgia, USA) that recognizes an epitope on β7 that is specific to the α4β7 heterodimer (Supplemental Figure 2C).
CD4+ T cells were stained with anti–CD3-PECy7, CD4-ECD, CCR9-FITC, and β7-APC mAbs. Unfixed cells were sorted on a BD FACSAria (BD Biosciences). Cells were stimulated by incubation with 10 ng/ml PMA and 500 ng/ml ionomycin (Sigma-Aldrich) for 1 hour at 37°C. Monensin was then added (GolgiStop, BD Biosciences), and the cells were incubated for a further 5 hours at 37°C. The cells were then washed, fixed/permeabilized, labeled with anti–IL-17–PE (eBioscience) and IFN-γ–Alexa Fluor 700 (BD Biosciences) mAbs, and analyzed using a BD LSRII (Supplemental Figure 4).
The plasma concentrations of I-FABP and sCD14 were determined by ELISA (Cell Sciences and R&D Systems). Plasma was diluted 50% for I-FABP assays and 0.5% for sCD14 assays. LPS was measured in plasma diluted to 5% and heated at 70°C using the Limulus amoebocyte lysate assay (Lonza). All assays were performed in duplicate.
Quantitative variables were compared using the Wilcoxon rank-sum test. Correlations were estimated by calculating Spearman’s rank correlation coefficients. The Wilcoxon signed-rank test was used for matched pairs. All tests were 2 sided, and P values of less than 0.05 were considered statistically significant. Statistical analyses were performed with Stata 9.2.
The study was approved by the Institutional Review Board CPP Sud-Ouest et Outre-Mer II. All participants provided written informed consent (trial registration no. NCT01038401).
This project was supported by grant EP44 from the French National Agency for Research on AIDS and Viral Hepatitis. We are indebted to the patients who took part in this study. We also thank L. Cuzin, F. Balsarin, S. Lagarrigue, and A. Fooladi for their help in monitoring the study; T. al Saati and F. Capilla for histopathologic analyses; R. Burcelin and C. Garret for LPS and sCD14 assays; F. Barré-Sinoussi for helpful discussions; and O. Parkes for checking the English text.
Conflict of interest: The authors have declared that no conflict of interest exists.
Citation for this article: J Clin Invest. 2012;122(1):62–69. doi:10.1172/JCI59011.