Role of Intestinal Myofibroblasts in HIV-Associated Intestinal Collagen Deposition and Immune Reconstitution following Combination Antiretroviral Therapy

Abstract

Objective

To investigate the potential role of mucosal intestinal myofibroblasts (IMFs) in HIV and associated fibrosis in GALT.

Design

Profibrotic changes within the secondary lymphoid organs and mucosa has been implicated in failed immune reconstitution following effective cART. Microbial translocation is believed to be sustaining these systemic inflammatory pathways. IMFs are non-professional antigen-presenting cells with both immunoregulatory and mesenchymal functions that are ideally positioned to respond to translocating microbial antigen.

Methods

Duodenal biopsies obtained from patients naïve to cART underwent trichrome staining and examined for TGF-β expression. Combined immunostaining and second harmonic generation-analysis was used to determine IMF activation and collagen deposition. Confocal microscopy was performed to examine for IMF activation and TLR4 expression. Finally, primary IMF cultures were stimulated with LPS to demonstrate expression of inflammatory biomarkers.

Results

The expression of the fibrosis-promoting molecule, TGF-β1, is significantly increased in duodenal biopsies from HIV patients naïve to cART and negatively correlated with subsequent peripheral CD4 recovery. The TGFβ1 increases coincided with an increase in collagen deposition in duodenal mucosa in tissue area adjacent to IMFs. We also observed that IMFs expressed TLR4 and had an activated phenotype since they were positive for fibroblast activation protein. Finally, stimulation of IMFs from HIV patients with TLR4 resulted in significantly increased expression of profibrotic molecules, TGF-β1 and IL-6.

Conclusions

Our data support the hypothesis that activated IMFs may be among the major cells contributing to the profibrotic changes and thus, the establishment and maintenance of systemic inflammation interfering with immune reconstitution in HIV patients.

Introduction

Effective combination antiretroviral therapy (cART) has dramatically impacted the morbidity and mortality associated with HIV infection [1–3]. Among those who achieve CD4⁺ T-cell counts in the normal range, HIV has become a manageable chronic disease with mortality approaching uninfected matched populations [4]. However, despite undetectable plasma HIV loads (pVL), up to 20% of patients fail to realize complete immune reconstitution in the peripheral vascular compartment as measured by CD4⁺ T-cell count [5–7]. This subgroup of patients is at highest risk for residual morbidity from HIV infection [8–10], with increased frequency of AIDS and non-AIDS defining malignancies and complications of hepatic, metabolic and cardiovascular diseases [10–14].

Incomplete immune reconstitution is an important challenge facing patients and providers with no effective interventions available. Treatment intensification with additional or more potent cART, cytokine therapy and therapeutic vaccination has not had a clinically meaningful impact up to the present [15–19]. Age, especially over 40–45 years old, and lower naïve CD4⁺ T-cells prior to initiating cART, are the most reliable baseline characteristic associated with blunted CD4⁺ T-cell rise [20, 21]. Since most of the HIV viral replication occurs within lymphoid tissue, reconstitution of the T-cells within this compartment would be essential for the complete recovery from the disease. Thus, an understanding of the mechanisms preventing this process would be essential for the development of successful full immune reconstitution and functional cure. Recent studies point out that development of fibrosis due to the abnormal collagen deposition in the secondary lymphoid tissue is a major factor preventing successful immune restoration even during long term cART [22, 23]. The increase in the amount of collagen within the T zone of secondary lymphoid tissue is suggested to be a prognostic marker for poor CD4⁺ T-cell reconstitution and development of AIDS [24]. It has been proposed that collagen deposition in secondary lymph nodes (LN) perturbs the T-cells and fibroblastic reticular cells (FRC) network leading to the disruption of CD4⁺ T-cell migration and preventing T-cell access to the cytokines and growth factors (e.g. IL-7, lymphotoxin) essential for their homeostasis [24, 25].

Reconstitution of CD4⁺ T-cells during cART is incomplete in the gut-associated lymphoid tissue (GALT), in particularly within the mucosal lamina propria effector compartment [26–29]. Failure of GALT immune reconstitution not only contributes to the development of HIV-associated enteropathy, but might also be involved in the maintenance of the viral reservoir in the gut mucosa [26]. Further, this limited reconstitution of CD4⁺ T-cells in GALT correlates with an early and extensive collagen deposition in the terminal ileal lamina propria and Peyer Patches to a greater extent than that which occurs in LN [27]. Although the mechanisms of HIV-associated GALT fibrosis are unknown, attention has turned to potential consequences of sustained immune activation driven by microbial antigen translocation [30–32]. Further, fibrosis in the setting of HIV/SIV-induced chronic immune responses may be due to the unbalanced activation of the gut innate immune system toward resident gut microbiota [24, 33, 34]. Despites the advance in the understanding of the influence of gut microbiota to the immune reconstitution in HIV, the role of the mucosal cell network in profibrotic changes within the GALT in HIV is understudied.

We, and others demonstrated that intestinal CD90⁺ stromal cells (myofibroblasts and fibroblasts, IMFs) are non-professional antigen presenting cells that are located beneath the epithelial basement membrane [35–38] and play a major role in the innate GI immune responses via signaling through TLR1-9 [39, 40]. These cells are important contributors to the wound healing process [41, 42]. However, uncontrolled proinflammatory activation of IMFs leads to hardwired functional changes in these cells that may contribute to the gut fibrosis via activation of profibrogenic signaling [43]. This may result in increased collagen deposition in response to the proinflammatory cytokine milieu (e.g. IL-1β, IL-6, TGF-β) and also may contribute to the increase/modulation of these proinflammatory molecules [44, 45]. Characterization of these processes has been well advanced in Crohn’s disease, where chronic inflammation leads to the burst of profibrogenic stimuli from professional immune cells and stromal cells leading to the downstream activation of collagen deposition by IMFs [46]. Thus, in this present work we have extended our study and have analyzed the profibrotic changes in the duodenum from HIV⁺ subjects and determined whether CD90⁺ stromal cells are activated and thus may contribute to the profibrotic changes within the GI mucosa in HIV.

Materials and Methods

Clinical Trial Design and Tissue Processing

Treatment-naive HIV patients underwent upper endoscopy before and 9 months after starting cART. Duodenal specimens from HIV-negative volunteers were used as a control group. The negative control subjects tend to be of similar age and lifestyle of the HIV-infected subjects enrolled in our GALT cohorts. All participants signed an informed consent form approved by the UC Davis Institutional Review Board prior to initiation of study procedures. Duodenal tissue was snap frozen for RNA extraction, embedded for trichrome stain or immunofluorescence antibody assay (IFA) or digested to single-cell suspension for FACS for T-cell subsets, lymphocyte activation levels, and IMF isolation. Trichrome stain slides were read as negative or positive (mild to severe) by a single, blinded pathologist (RS). Peripheral blood mononuclear cells (PBMC) had the same FACS analysis using a previously described 10-color panel and three-laser LSRII flow cytometer [26]. Single-cell suspensions of 10⁶ cells of duodenal tissue were pelleted for HIV proviral DNA quantification as previously described [47].

Real-Time RT-PCR

RT-PCR was selected as the most robust measure of chronic TGF-β tissue up regulation [48]. Total cellular RNA was isolated from duodenal cells using RNeasy® RNA isolation kit (Qiagen, Valencia, CA) according to the manufacturer’s instructions. RT real-time PCR was performed according the Applied Biosystems’s two-step RT real time PCR protocol (Applied Biosystems, Foster City, CA). Human GAPDH was used as the housekeeping gene. The primers for human TGF-β1 and procollagen type I (α1) were reported previously [49]. The endpoint used in real-time PCR quantification, CT, defined as the PCR cycle number that crosses the signal threshold. CT values range from 0 to 40, with the latter number assumed to represent no product formation.

Quantification of induced cytokine gene expression from primary IMF cultures was performed using the comparative CT method (Sequence Detector User Bulletin 2; Applied Biosystems) and reported as the fold-difference relative to the human housekeeping gene, 18S rRNA. In order to calculate the fold change (increase or decrease), the CT value of 18S rRNA was subtracted from CT value of the target cytokine gene to yield the CT. Change in the expression of the normalized target gene as a result of experimental conditions was expressed as 2-CT where CT=CT experimental samples - CT biological control.

sCD14 measurement

Soluble CD14 levels in plasma samples were quantified by ELISA with the Quantikine Human sCD14 Immunoassay (R&D Systems, Minneapolis, Minnesota, USA) according to the manufacturer’s protocol in duplicate.

Primary IMF cultures

Primary cultures of IMFs were generated as we have reported previously [50]. IMFs were passaged at least three times in culture in order to eliminate residual macrophages and other adherent cells. Purity of isolated cells was analyzed by flow cytometry analysis as we previously described [35].

Confocal microscopy

The immunostaining followed by confocal microscopy was performed as described previously [51]. Frozen tissue sections of intestinal mucosa, were fixed in 2 % paraformaldehyde for 30 minutes, blocked with normal mouse and goat serum (1:10 in PBS) for 30 min at room temperature. Samples were then incubated with the mixture of following antibodies: 1 μg/mL of anti-fibroblast marker ER-TR7 rat monoclonal mAbs (clone ER-TR7, Santa Cruz Biotech Inc., CA) conjugated to AF-546, anti-α-SMA monoclonal Abs (clone 1A4, Sigma), anti-CD90 (Clone 5E10, eBioscience) conjugated with AF-488, or TLR 4 was detected using anti-TLR4 mAbs (clone HTA125, eBioscience) for 30 min at room temperature. Each staining step was followed by three washes with PBS. Zenon mouse IgG and Apex Antibody Labeling kits (Invitrogen, CA) were used to conjugate mouse and rat mAbs. Microscopy was performed with confocal and multiphoton microscopes Zeiss LSM510 META, a (Carl Zeiss, Thornwood, NY) and Olympus FV1000 MPE (Olympus, Center Valley, PA), as well as modified Zeiss LSM 410 confocal microscope as described previously [52]. Combination of immunostaining and second harmonic generation (SHG)-analysis followed by confocal and two-photon excitation and second harmonic generation (SHG) microscopy was used to determine collagen A deposition within the intestinal mucosa as previously described by our group [52, 53]. The physical origin of SHG within the connective and muscle tissues is attributed to the laser interaction with dipolar protein structures that is enhanced by the intrinsic chirality of the protein helices. Collagen has been shown to give rise to SHG [54]. The source of SHG signal in the mucosa arises mostly from fibrillar collagen, which is mostly located in the mucosal stroma [53].

Stimulation of cultured IMFs with TLR-4 agonists

IMF primary cultures were generated as described above. IMFs were stimulated with LPS, a TLR4 ligand, 1 mg/mL, or buffered saline for 24h (InvivoGen, San Diego, CA). The RNA was extracted and gene expression was measured using real-time RT-PCR as described above.

Statistical analysis

Nonparametric statistical methods were used throughout. Wilcoxon matched-pairs signed rank test and Mann–Whitney test was used as appropriate. The Spearman rank correlation coefficient was used to study correlations between parameters. All values are represented as the median (interquartile ranges). Analyses were performed with GraphPad Prism Software V6.01 (GraphPad Software, Inc., La Jolla, CA).

Results

Baseline characteristics and virologic/immunologic outcome after nine months of ART

The samples used for this pilot study were obtained during a randomized clinical trial where HIV positive volunteers naïve to ART underwent upper and lower endoscopy prior to and nine months after being randomized to either a non-nucleoside reverse transcriptase inhibitor or raltegravir, both in combination with tenofovir/emtricitabine [26]. Overall, 16 HIV patients and 5 controls underwent upper endoscopy and a subset underwent flexible sigmoidoscopy. This report focuses on the 14 study participants for whom the relevant samples were available for the assays reported herein at both baseline and at 9 months post cART and negative control volunteers who underwent the same interventions. The control subjects were family and friends of HIV patients at low risk for HIV infection. The demographic profiles and results from clinical responses are reported in Table 1, as well as the peripheral blood and duodenal tissue virologic and immunologic results. Importantly, the median CD4⁺ T-cell count in HIV patients rose from a median of 328 cells/mL (235–417) at baseline to 562 cells/mL (388–651) among those who completed 9 months of treatment (n=14). This represented a median increase of 181 cells/mm² (145–295) in the peripheral blood absolute CD4⁺ T-cell counts. All patients became undetectable after nine months by standard commercial assays. The single copy HIV RNA and proviral DNA assay results are listed in Table 1 for peripheral blood and tissue compartments. HIV proviral DNA declined from 208 (31–1189) to 110 (14–694) in the peripheral blood and from 37 (3.9–193.4) to 15.4 (2.6–105.6) in the duodenal tissue single cell suspensions of patients (P=0.005 and P=0.32, respectively). As expected both plasma sCD14 concentration and CD8⁺ T-cell percentage with the activated phenotype decreased after the introduction of cART (P=0.01 and P=0.03, respectively). Although we observed significant increases in the tissue CD4⁺ T-cell count (p=0.024), they did not achieve the average level observed in the lamina propria of normal controls (data not shown).

Table.

	normal control	baseline	cART 9-month	p value
n	5	14	—
male/female	3/2	9/5	—
race/ethnicitya	3/1/1	6/4/4	—
median age years (interquartile range)	42 (33–44)	37 (28–46)	—
Peripheral blood CD4b	—	328 (235–417)	562 (388,651)	<0.0001
plasma viral load cp/mL	—	25,600 (14,770–70,613)	11.5 (11.3, 750)	<0.0001
PBMC proviral DNA cp/106 cells	—	208 (31–1189)	110 14–694	0.005
Duodenal tissue proviral DNA - cp/106 cells	—	37 (3.9–193.4)	15.4 (2.6–105.6)	0.32
sCD14 μg/mL	—	1.9 (1.7–2.4)	1.7 (1.2–1.8)	0.01
CD8+/CD38+/HLA-DR+ PBMCs	—	44.2% (32.2–61.4)	27.7% (18.8–38.5)	0.03
Doudenal lamina propria CD3+/CD4+ cells/mm2	845 (665–1159)	105 (76.5–170.5)	151.5 (129–172)	0.02b
PBMC ∆CD4b (IQR)	—	—	181 (145–295)

^aCaucasian/African-American/Hispanic

^bBaseline versus 9-month (Wilcoxon rank sum test)

^cmedian cells/mL; (interquartile range)

Duodenal Tissue TGF-β Expression is increased in HIV⁺ subjects and correlates with increases in collagen deposition

We analyzed the mRNA levels of TGF-β1 in duodenal LP and correlated it with the level of procollagen type I (α1, procollagen A) expression. TGF-β1 mRNA levels were higher in HIV patients (335.4 (187.6–621.7)) compared to controls (57.6 (2.1–134.6)) P=0.002, respectively (Fig 1a). Procollagen A mRNA levels were also higher in patients (462.0 (189.0–805.5) compared to controls (92.5 (49.7–258.3)) P=0.048)(Fig 1b). As would be expected and as an internal control, TGF-β1 and procollagen A expression levels were positively correlated (r=0.65, P=0.0018). Conversely, tissue duodenal HIV proviral DNA did not correlate with TGF-β expression (r=0.03, P=0.5). Neither monocyte activation as measured by sCD14 levels, nor CD8⁺ T-cell activation percentages correlated with peripheral blood or tissue HIV proviral DNA or HIV plasma viral load (both P>0.5). Interestingly, sCD14 and CD8⁺ T-cell activation only correlated at the 9-month time point (r=0.63, P=0.01), but not at baseline pre-treatment, suggesting perhaps that several possible factors influence monocyte activation prior to suppression of HIV viremia but less so once pVL is controlled by effective cART.

Figure 1. Upregulation of biomarkers of fibrosis are increased in duodenal tissue of untreated HIV-infected subjects and predict subsequent immunologic response to cART.

Duodenal tissue mRNA levels of TGF-β (a) and Procollagen type I (b) are increased in HIV-infected patients compared to controls by quantitative PCR (P=0.002 and P=0.048, respectively). Immune reconstitution (rise in absolute PBMC T-cell counts) correlates negatively with baseline tissue TGF-β expression (c) and positively with peripheral naïve CD4⁺ T-cell proportions (d). Baseline PBMC naïve CD4⁺ T-cells correlate with the naïve CD4⁺ T-cell in the duodenal tissue compartment (e) and negatively with the TGF-β expression in the duodenal tissue (f).

Profibrotic changes in duodenal tissue from HIV⁺ subjects negatively correlates with the reconstitution of CD4⁺ T-cells

In our next step, we analyzed whether increases in the fibrosis in duodenal lamina propria correlated with the immune reconstitution of the HIV-positive subjects following initiation of ART as has been reported in lymph node parenchyma [22]. Baseline duodenal tissue TGF-β, but not procollagen A expression negatively correlated with subsequent rise in absolute T-cell counts (r=−0.74, P=0.0024)(Fig 1c). Interestingly, the baseline PBMC naïve CD4⁺ T-cell subset (median of 39.6% (22.1, 49.0)) was the only CD4⁺ T-lymphocyte maturational subset to correlate with TGF-β expression at baseline (r=−0.69, P=0.017)(Fig 1d). In parallel, the only subset of CD4⁺ T-lymphocytes to correlate between the two compartments of peripheral blood and duodenal tissue was the naïve CD4⁺ T-cells (r=0.59, P=0.016)(Fig 1e). The absence of a correlation between the peripheral blood absolute CD4⁺ T-cell count and the density of CD4⁺ T-cells in the duodenal tissue (r=0.38, p=0.15) suggests that the correlation between the 2 compartments relative to the naïve subsets is more likely due to systemic conditions for naive T-cell maturation as has been described in the lymph nodes previously [25]. Finally, and as others have reported regarding baseline immunologic parameters that best predict subsequent CD4⁺ T-cell response to effective ART, baseline PBMC percentage of naïve CD4⁺ T-cells correlated significantly with immune reconstitution (r=0.71, P=0.002)(Fig 1f) [5, 55, 56]. These data strongly suggest that TGF-β expression in GALT correlates with subsequent immune reconstitution in the peripheral compartment.

Collagen deposition is increased in the duodenal lamina propria during HIV infection and is associated with the increase in the activation of the IMFs

Next, we examined collagen deposition within the duodenal mucosa using trichrome stain followed by histological analysis. H&E stain was included to identify inflammation which was not seen (Figure 2A). As detected by trichrome staining (Figure 2B, collagen deposition is in blue) an increase in mucosal lamina propria collagen was observed in 10 out of 20 tested HIV⁺ patients compared to 1 out of 5 control patients (p=0.078). Increased collagen deposition was confirmed by SHG-analysis using 2-photon microscopy in a subset of six out of ten tested HIV⁺ patients, but was not observed in a healthy control group (Figure 2C). Taken together with our quantitative mRNA data analysis (Figure 1A), these observations strongly support the conclusion that fibrotic changes occur in the gut during chronic HIV infection. No significant decrease in the collagen deposition was observed in the tissue from subjects after 9 months of cART (data not shown)

Figure 2. Collagen deposition in duodenal lamina propria.

Hematoxylin and eosin staining from representative two HIV+ and an HIV negative control reveals normal architecture (row A). Trichrome stain (row B) of duodenal mucosa was positive from 10/20 HIV positive subjects (representative examples from the same subjects as row A). (C) (SHG)-analysis using 2-photon microscopy was used to determine collagen A deposition (in gray), 40x objective. SHG analysis from duodenal mucosa from two additional HIV+ and one HIV negative control subjects are shown. Arrows point out collagen deposition in panel B and C. Very little staining for collagen is evident for the HIV negative sample, hence the faint image.

Mucosal increase in collagen A deposition is concomitant with the increase in the activation of IMFs

Intestinal fibrosis is classically viewed as an inflammatory process associated with the proliferation and activation of local IMFs leading to abnormal deposition of collagen [44, 57]. We observed that the increase in deposition of collagen A in the duodenal lamina propria of HIV⁺ subjects (Figure 3 in blue) was accompanied with the disruption of normal mucosal fibroblastic network architecture and activation of IMFs, as determined by expression of α-SMA (Figure 3 in green) and Fibroblasts Activation Protein (FAP) (Figure 3, in red). FAP is believed to be increased in fibroblasts in fibrotic tissue and cancer [58]. Thus, colocalization of the α-SMA (Figure 3) and the FAP (formation of the yellow-orange color on merged images, Figure 3) within IMFs from HIV⁺ subjects suggests that abnormal activation of the stromal cells in the intestinal mucosa may contribute to the fibrotic tissue remodeling in HIV-associated immunopathogenesis.

Figure 3. IMFs are source of collagen in HIV disease.

Anti-α-SMA+ mAbs (Clone 1A4) were used to detect myofibroblasts (shown in green), rabbit polyclonal fluorochrome conjugated anti-FAP Abs were used to analyze activation of IMFs (shown in red), collagen A deposition was detected by second harmonic generation (SHG)-analysis (shown in blue) using two-photon microscopy. Increase in a “filament” type fully formed collagen A in HIV⁺ duodenal mucosa is adjacent to the α-SMA⁺ IMFs (in green), which also express marker of activation, FAP, and result in formation of orange-yellow color on merged images.

Stimulation of TLR4 on intestinal CD90⁺ fibroblast/myofibroblasts derived from HIV+ subjects upregulates the production of profibrotic mediators

The main driver of gut fibrogenesis is believed to be chronic inflammation, which leads to mesenchymal cell recruitment and activation. Recent findings suggest that microbial overgrowth and abnormality in microbial translocation play a critical role in gut fibrogenesis [34, 57]. Lipopolysaccharide (LPS) is known to be a profibrotic agent [59]. Thus, we determined whether IMFs express TLR4, a receptor for LPS, in duodenal mucosa from HIV⁺ subjects. Using immunostaining followed by confocal microscopy of duodenal mucosa tissue sections, we were able to detect compartmentalized TLR4 expression in duodenal mucosa from HIV⁺ patients (Figure 4, showing in red). In duodenal LP, CD90⁺ stromal cells (a.k.a. IMFs) were among the major cell phenotype expressing TLR4 (Figure 4, showing in green). Next we determined how TLR4 stimulation modulates expression of the cytokine and growth factors involved in the fibrosis. Our data demonstrated that stimulation of the HIV-derived IMF primary cultures with LPS (lμg/mL) for 24h resulted in significant upregulation of TGF-β1 and IL-6 mRNA expression (Figure 5). Taken together, these data suggest that during HIV enteropathy, the interaction of IMFs with LPS may contribute to the increase in the soluble mediators of fibrogenesis (IL-6, TGF-β1) and that may directly enhance collagen deposition by IMFs.

Figure 4. IMFs are abundant in intestinal lamina propria of HIV+ subjects and express TLR4.

IMFs are abundant in intestinal lamina propria of both HIV- and HIV+ subjects and express TLR4. Confocal microscopy analysis of longitudinally displayed duodenal sections. DAPI was used to identify nuclei of cells (blue), anti-fibroblast marker CD90 mAbs (clone ASO2)(green) was used to identify mesenchymal stromal fibroblasts/myofibroblasts in duodenal mucosa. TLR 4 was detected using anti-TLR4 mAbs (clone HTA125). Merged images demonstrated co-localization between CD90 and TLR4, yellow-orange coloring (as shown by arrow).

Figure 5. LPS stimulation of cultured IMFs.

Activation of primary culture of HIV-derived IMFs through TLR4 results in upregulation of TGFβ1 and IL-6 expression. Open bars-IMF alone; Solid bars-IMF stimulated with LPS, 1 μg/mL for 24 h, mRNA was analyzed using real time RT-PCR. Relative fold increase was calculated using 18S as a housekeeping gene. The results represent mean of mRNA A fold increase of duplicates ± SE from three donors (n=3), *p<0.05.

DISCUSSION

Incomplete immune reconstitution following effective cART is associated with significant residual morbidity amongst HIV-infected patients. Its understanding and control is a significant unmet need in the management of HIV disease. Considerable debate exists in the literature regarding the underlying pathophysiology of failed immune reconstitution in the gut following effective cART. In this report, we sought to examine the role of events at the interface of the gut mucosa and the systemic immune system to explore which factors influence peripheral CD4⁺ T-cell reconstitution following effective cART. We and others have previously demonstrated that higher percentages of proinflammatory bacterial taxa including the Proteobacteria/Enterobacteriales were associated with changes in the duodenal GALT T-cell subset proportions and T-cell activation [33, 60]. Similarly, Dillon et al. also observed a relationship between altered mucosal microbiome and blood T-cell activation [61]. More recently, we reported the impact of oral serum-derived bovine immunoglobulin (which is broadly reactive against bacterial and viral antigen) on HIV enteropathy, gut mucosal repair and function, and systemic immune activation [62]. Most relevant to the work described herein, oral bovine immunoglobulin administration resulted in a significant increase in duodenal lamina propria CD4⁺ T-cells after only 8-weeks of treatment supporting the hypothesis that bacterial products impact immune reconstitution locally [62]. Klatt et al observed similar results with the administration of probiotics/prebiotics in combination with ART in SIV-infected pigtail macaques further supporting the hypothesis that microbial translocation is causing fibrosis [63].

This report extends that work by exploring pathways of focal tissue inflammation by examining TGF-β expression as an effector molecule and procollagen type I expression as the downstream mediator of fibrosis in duodenal tissue. We observed increased extracellular matrix deposition of collagen as detected by histopathologic measures using trichrome staining in a majority of duodenal samples and dual photon microscopy. Similar observations were previously reported by the Estes group for ileal lamina propria and Peyer Patches [27]. Increased expression/production of procollagen and TGF-βl in mucosal and lymph node tissue have been reported previously in the setting of both HIV-1 and macaque SIV infection in the cervix, lymph nodes and GALT [64–68], demonstrating the systemic nature of HIV-associated mucosal inflammation. We were also able to demonstrate the impact of baseline profibrotic pathways on subsequent peripheral blood CD4⁺ T-cell immune reconstitution. Our results, which focus on events at the GALT-gut lumen interface, identify a portal for systemic inflammation and by extension, immune activation in patients on effective cART.

While previous reports have identified lymphoid T-regulatory and double-negative lymphocytes as a potential source for the TGF-β [25, 66], we sought to pursue the potential role of widely-distributed IMFs in the genesis of collagen deposition in GALT of HIV patients. This cell type is of particular interest in this regard as it is of the same lineage as the stellate cell in the liver and is known to play a key role in the fibrogenesis of other inflammatory bowel diseases such as Crohn’s Disease [46]. These stromal cells represent non-professional antigen presenting cells which form a continuous interconnecting network under the epithelial layer of the entire gastrointestinal tract. Thus, they are ideally positioned to respond to microbial antigen translocation. When IMFs are stimulated with bacterial antigen, they increase in number and volume as demonstrated by α-SMA staining [69]. Indeed, we observed increased α-SMA staining in the duodenal tissues of HIV patients compared to control. We were also able to show co-localization of IMF and procollagen, supporting the hypothesis that these cells represent a source for profibrotic pathways and processes in GALT in HIV disease. To further pursue a sequence of events in the proposed pathway, we first identified whether an effector molecular receptor of bacterial antigen LPS signaling, TLR4, was present on IMFs in the tissue samples from HIV patients. It is known from previous reports that there are increased plasma levels of LPS in HIV patients [30]. Although limited, the TLR-4 staining was strongest on IMFs which are located in subepithelial position, a location ideal for signaling microbial antigen translocation. Several pathogen-associated molecular patterns (PAMPs) recognition receptors have been detected previously on IMFs, TLR 2, 5, and 9 [45]. In the final set of experiments, we cultured IMFs from intestinal biopsy specimens and stimulated them with TLR4 agonists in order to assess whether TGF-β1 and IL-6 upregulation could be detected. IL-6 is an important cytokine in the inflammation cascade and contributes to collagen deposition. It has been suggested to be a useful predictive marker for liver fibrosis in HIV-infected patients with alcohol addiction [70]. Consistent with animal models and other disease state reports [71, 72], the IMFs had a significant increased expression of TGFβ-1 and IL-6, further supporting the proposed hypothesis. These data suggests that a complex of profibrotic molecules may be released by IMFs upon microbial stimulation in the HIV leaky gut which will require further investigation to verify.

A strong negative correlation was observed between baseline duodenal TGF-β expression and systemic immune reconstitution following 9 months of effective cART. A similar correlation was observed between baseline duodenal TGF-β expression and systemic naïve CD4⁺ T-cell percentages. Zeng et al. have presented an important set of experiments demonstrating the impact of lymph node fibrosis and loss of normal architecture in the FRC network on IL-7 signaling which is necessary for naïve CD4⁺ T-cell survival. The findings reported herein provide a complimentary sequence of events that might be implicated in the incomplete immune reconstitution seen in up to 20% of patients receiving long-term successful cART. Recognizing that HIV is still detectable in tissue compartments [47], there does not appear to be a correlation between viral loads and failed immune reconstitution as might be expected if HIV-driven inflammation is the nexus for failed immune reconstitution [73]. Conversely, bacterial antigen present in high concentration adjacent to GALT may be altered in its composition favoring pro-inflammatory taxon [33, 74, 75]. These findings reveal a vital interaction between the gut microbiota and the mucosal immune system that calls for further investigation [75–77].

IMFs have been shown to impact T-regulatory cell activation and function [51] and are anatomically ideally situated to respond to microbial antigen translocation in the setting of HIV disease. Perhaps these cells are what links failure of immune reconstitution despite virologic control. While these present data do not prove, nor do they attempt to prove, a causal relationship between IMFs and upregulated systemic profibrotic pathways, understanding the role of IMFs at the front gate of microbial antigen:mucosal immunity interactions may provide a rational strategy for restoring microbial communities and interrupting focal/systemic profibrogenic pathways. These novel findings in the small intestines of HIV patients may provide insights into alternative therapeutic opportunities to improve immune reconstitution and reduce associated morbidity.