Overt IL-32 Isoform Expression at Intestinal Level during HIV-1 Infection is Negatively Regulated by IL-17A

Abstract

Objectives:

Untreated HIV infection was previously associated with IL-32 overexpression in gut epithelial cells (IEC). Here, we explored IL-32 isoform expression in the colon of people living with HIV (PLWH) receiving antiretroviral therapy (ART) and IL-32 triggers/modulators in IEC.

Design:

Sigmoid colon biopsies (SCB) and blood were collected from ART-treated PLWH (HIV+ART; n=17; mean age: 56 years; CD4 counts: 679 cells/μl; time on ART: 72 months) and age-matched HIV-uninfected controls (HIV^neg; n=5). The IEC line HT-29 was used for mechanistic studies.

Methods:

Cells from SCB and blood were isolated by enzymatic digestion and/or gradient centrifugation. HT-29 cells were exposed to TLR1–9 agonists, TNF-α, IL-17A, and HIV. IL-32α/β/γ/D/ε/θ and IL-17A mRNA levels were quantified by real-time RT-PCR. IL-32 protein levels were quantified by ELISA.

Results:

IL-32β/γ/ε isoform transcripts were detectable in the blood and SCB, with IL-32β mRNA levels being predominantly expressed in both compartments and at significantly higher levels in HIV+ART compared to HIV^neg. IL-17A transcripts were only detectable in SCB, with increased IL-17A levels in HIV^neg compared to HIV+ART and negatively correlated with IL-32β mRNA levels. IL-32β/γ/ε isoform mRNA were detected in HT-29 cells upon exposure to TNF-α, Poly I:C (TLR3 agonist), Flagellin (TLR-5 agonist) and HIV. IL-17A significantly decreased both IL-32 β/γ/ε mRNA and cell-associated IL-32 protein levels induced upon TNF-α and Poly I:C triggering.

Conclusions:

We document IL-32 isoforms abundant in the colon of ART-treated PLWH and reveal the capacity of the Th17 hallmark cytokine IL-17A to attenuate IL-32 overexpression in a model of inflamed IEC.

Introduction

Antiretroviral therapy (ART) transformed the Human Immunodeficiency Virus 1 (HIV-1) infection into a controlled chronic disease and increased the quality of life of people living with HIV (PLWH) [1, 2]. However, HIV-1 remains a major health challenge, with 38.0 million PLWH and 1.7 million new infections in 2019 [3]. The main barriers to a cure include the persistence of HIV reservoirs during ART [4–6] and the non-restoration of intestinal mucosal barrier functions, despite early ART initiation [7–13]. CD4⁺ T-cells in the gut-associated lymphoid tissues (GALT) are major sites of HIV replication/persistence [8, 13–21]. The depletion of CD4⁺ T-cells, principally Th17 cells [20–22], results in dramatic alterations in intestinal integrity, which facilitates microbial translocation [19, 21, 23–28]. The state of chronic inflammation facilitates the occurrence of non-AIDS comorbidities [16, 29–34]. Indeed, compared to uninfected individuals, ART-treated PLWH present greater inflammation/immune activation and premature ageing [13, 35–39]. Studies by our group and others revealed the contribution of the interleukin (IL)-32 to this state of inflammation, and linked IL-32 overproduction to cardiovascular risk in ART-treated PLWH [40–45].

IL-32, identified by Dinarello’s group [46], is a pro-inflammatory cytokine highly expressed in activated T and NK cells [47]. The il32 gene codes for multiple isoforms (α, β, γ, δ, ε, θ, ζ, η, small) generated by alternative splicing [45, 48], exerting pro-inflammatory and antiviral functions [49–53]. In addition, a new IL-32D isoform (NM_001012636.2; distinct from IL-32delta/δ, NCBI #AY495333.1) was identified by our group (Zaidan et al., 2019). The human IL-32 receptor remains unknown [54]. The absence of il32 gene in mice explains limitation in understanding the mechanism of IL-32 action (e.g., signaling pathways, target cells). Nevertheless, the exogenous administration of human IL-32γ in mice was documented to cause joint swelling [55], indicating that the IL-32 receptor is expressed on murine cells. IL-32 promote the expression of TNF-α [46], thus positioning IL-32 upstream in the inflammation cascade hierarchy. In humans, IL-32 overexpression is associated with auto-immune and inflammatory diseases [56], and is potentially linked to cardiovascular disease [43]. IL-32 expression is also upregulated during HIV infection [57, 58]. Our group showed that exposure to HIV induces the expression of IL-32 in peripheral blood mononuclear cells (PBMC) and that IL-32 levels are not normalized by ART [41, 45]. IL-32 overt expression in plasma and PBMCs of PLWH is associated with CD4⁺ T-cell decline, viral rebound, and inflammation in individuals with a history of spontaneous disease control [41]. More recently, our group identified the IL-32β as the dominantly expressed isoform in the PBMCs of ART-treated PLWH [45]. While multiple studies documented IL-32 overexpression during HIV infection in plasma and PBMCs [41, 45, 57–62], studies on IL-32 expression in peripheral tissues such as the intestine are limited. Of note, intestinal epithelial cells (IEC) produce IL-32 during various inflammatory conditions in vitro and in vivo [63–69], including HIV, as previously reported in the intestine of untreated PLWH [58].

Herein, we aimed to determine the contribution of the intestinal environment to IL-32 overexpression during ART-treated HIV-1 infection, and we sought to identify triggers/modulators of IL-32 production in IEC. We quantified the expression of IL-32 isoforms (α, β, γ, D, ε, θ) and Th17 hallmark cytokine IL-17A in the colon and PBMCs of ART-treated PLWH. Also, we investigated the capacity of IL-17A to modulate the production of specific IL-32 isoforms by IEC in response to inflammatory and bacterial/viral stimuli in vitro.

Material and Methods

Study participants

Study participants were recruited at McGill University Health Centre, as previously described [70]: ART-treated PLWH with undetectable plasma viral load (<40 copies HIV RNA/ml, median age: 56 years; CD4: 679 cells/µl; CD4/CD8 ratio: 0.71; time since infection: 20 years; time on ART: 7 years) and HIV-uninfected participants (median age: 59 years; CD4: 750 cells/µl; CD4/CD8 ratio: 2.11). Colonoscopies were performed in the context of a colorectal cancer screening. Detailed clinical information on study participants are included in Supplemental Table 1.

Sample processing and phenotypic analysis

Sigmoid colon biopsies (SCB) were processed using Liberase DL (Roche Diagnostics). Matched blood was collected and PBMCs)were isolated by gradient density centrifugation. Cells were stained with the following fluorochrome-conjugated antibodies: CD3-Pacific-Blue (UCHT1), CD4-Alexa-Fluor 700 (RPA-T4; BD Biosciences), CD45RA-APC-eFluor780 (HI100) (eBioscience), CD3-Alexa700 (UCHT1), CD326-BV650 (9C4), CD8-PerCP-Cy5.5 (RPA-T8), CD19-PerCP-Cy5.5 (HIB19), CD66b-PerCP-Cy5.5 (G10F5; Biolegend), CD4-FITC (SFCI12T4D11; Beckman Coulter), CD64-APC (10.1.1; Miltenyi). Aqua Vivid Dead Cell Stain Kit (Life Technologies) was used to exclude dead cells. Phenotypic analysis was performed by flow cytometry using BD-LSRII cytometer, BD-Diva (BD Biosciences) and FlowJo software (Tree Star). Detailed sample processing informations are included in our previous publication [70].

Ethics statement

This study was conducted in compliance with the Declaration of Helsinki and received approval from the Institutional Review Board of the McGill University Health Centre and CHUM-Research Centre. Participants signed a written informed consent.

HT-29 experiments

HT-29 cells (ATCC^® HTB38™) were cultured using McCoy’s medium (ATCC® 30–2007™, 10% of heat-inactivated fetal bovine serum (FBS), 1% penicillin/streptomycin (GIBCO), at 37 °C and 5% CO2, seeded in 75 cm² flasks, and harvested at confluence using 0.05% of trypsin/EDTA. Experiments were performed from passage 3 to 10. Cells were stimulated with the following toll-like receptor (TLR) ligands (InvivoGen): TLR1/2 (Pam3CSK4, 1 µg/ml), TLR2 (HKLM, 10 µg/ml), TLR3 [Poly(I:C), High Molecular Weight (HMW) and Low Molecular Weight (LMW), 10 µg/ml], TLR4 (LPS E. coli K12, 1 µg/ml), TLR5 (Flagellin from Salmonella typhimurium, 1 µg/ml), TLR6/2 (FSL1, 1 µg/ml), TLR7 (Imiquimod, 1 µg/ml), TLR8 (ssRNA40/LyoVec, 1 µg/ml) and TLR9 (ODN2006, 0.26 µg/ml). In parallel, HT-29 cells were exposed to recombinant human (rh) TNF-α (10 ng/ml) and rhIL-17A (100, 10, 1 ng/ml; R&D Systems), or the NL4.3BaL and transmitted/founder THRO HIV (100, 50, 10 ng/ml HIV-p24). NL4.3BaL is a replication-competent NL4.3-based provirus expressing the BaL envelope provided by Dr. Michel Tremblay (ULaval, Québec), originating from Dr. Roger Pomerantz (Thomas Jefferson University, Philadelphia, USA). THRO is a replication-competent molecular clone NIH-HIV Reagent Program, Division of AIDS, NIAID: HIV-1 pTHRO.c/2626, ARP-11745, by Dr. John Kappes and Dr. Christina Ochsenbauer (100, 50 and 10 ng HIV-p24/ml). HIV-1 stocks were produced by 293T cell transfection, as previously described [71], and passaged once on CD3/CD28-activated memory CD4⁺ T-cells. HIV-p24 was quantified by ELISA. All investigations were performed in triplicates in three independent experiments.

IL-32 and IL-17A real-time RT-PCR quantification

RNA from PBMCs/SCB was isolated using DNA/RNA dual extraction kit (Qiagen), and from HT-29 dry pellets using RNeasy kit (Qiagen), following manufacturer’s instructions. IL-32α/β/γ/D/ε/θ isoforms and IL-17A mRNA were quantified using one-step SYBR Green real-time PCR, using the LightCycler 480 II (Roche) and Qiagen reagents. Real-time RT-PCR was performed in duplicates using 25 ng RNA/reaction for IL-32 and 70 ng RNA/reaction for IL-17A. Negative controls (no sample) were included for each gene/isoform. IL-32 expression was normalized to the housekeeping gene β-glucuronidase. Primers for IL-32α/β/γ/D/ε/θ and reference gene were previously described [45]. Of note, we tested for IL-32D (NM_001012636.2 → NP_001012654), not IL-32delta/δ. IL-17A mRNA (Qiagen QuantiTect Primers) was normalized to 28S rRNA reference gene (IDT), as previously described [71, 72]. Graphs were plotted as the relative target gene quantification (delta CT).

ELISA

Levels of human proteins in HT-29 cell culture supernatants and plasma of study participants were quantified using ELISA kits for total IL-32, TNF-α, CCL20, sCD14, I-FABP (R&D Systems) and IL-17A (eBiosciences), according to manufacturer’s protocol. In parallel, HT-29 dry pellets were lysed in 1X RIPA buffer (Cell Signaling Technology) supplemented with protease inhibitors and by sonication. Total protein levels in cell lysates were quantified using the DC protein assay kit (Biorad). Total IL-32 levels were quantified by ELISA (R&D Systems) and normalized to total protein levels (pg/1 μg total protein).

Statistical analysis

Statistical analyses were performed using the GraphPad Prism 8 software (GraphPad Software, Inc.). Friedman with Dunn’s multiple comparisons tests were applied to determine statistical significance for differences between matched values. Mann–Whitney p-values were calculated for comparisons between unmatched groups. Spearman correlation and linear regression models were applied to study correlations between matched values.

Results

IL-32β mRNA overexpression in the colon of ART-treated PLWH.

Matched PBMCs and SCB samples were available from n=17 ART-treated PLWH (HIV+ART) and n=5 HIV-uninfected controls (HIV^neg) [70]. Cells extracted from SCB contained various lineages including IEC and CD3⁺ T-cells (Supplemental Figure 1A–B) [70, 71, 73], with differences in the frequency of IEC and CD3⁺ T-cells, and the IEC/CD3⁺ T-cell ratios between HIV+ART and HIV^neg being similar (Supplemental Figure 1C–E). Results in Figure 1A–B depict the higher expression of IL-32β, γ and ε isoforms in PBMC compared to cells extracted from SCB samples of HIV+ART. Of note, IL-32β mRNA levels were significantly higher in SCB of HIV+ART compared to HIV^neg (Figure 1C). Consistent with our previous reports [41, 45], plasma levels of total IL-32 protein were higher in HIV+ART compared to HIV^neg (p=0.0518; Supplemental Figure 2A). In contrast, differences in plasma levels of the pro-inflammatory chemokine CCL20, microbial translocation (sCD14) and gut permeability (I-FABP) markers were not statistically different between those two groups (Supplemental Figure 2B–D), with TNF-α levels being undetectable (data not shown). IL-32β mRNA levels in the colon did not correlate with IL-32 mRNA levels in PBMCs (data not shown), nor with the plasma IL-32 levels (Supplemental Figure 2E), indicative that IL-32 expression in these compartments is differentially regulated. In addition, IL-32 mRNA levels in SCB did not correlate with plasma CCL20, sCD14 and I-FABP levels (Supplemental Figure 2F–H).

Figure 1: Expression of specific IL-32 isoforms in the blood and colon of ART-treated PLWH.

Total RNA was extracted from matched PBMCs and sigmoid colon biopsies from ART-treated PLWH (HIV+ART; n = 17) and age-matched uninfected individuals (HIV-; n=5) (Supplemental Table 1). RNA was used to measure IL-32 mRNA isoforms (α, β, γ, D, ε, θ) by real-time RT-PCR. Shown are IL-32α, β, γ, D, ε, and θ mRNA expression in blood (A) and colon (B), as well as the statistical analysis of IL-32β, γ, and ε mRNA expression in HIV+ART (n=17) and HIV- participants (n=5). Mann-Whitney p-values are indicated on the graphs.

Together, these results provide a cartography of IL-32 isoforms expressed in cells from colon and blood of ART-treated PLWH, and demonstrate an increased expression of IL-32β mRNA in the colon of HIV+ART compared to HIV^neg.

IL-32β and IL-17A mRNA levels are negatively correlated in the colon of ART-treated PLWH.

The paucity of IL-17A-producing Th17 cells is well documented during ART-treated HIV infection [21, 22, 25, 28]. Therefore, we quantified IL-17A mRNA levels in PBMCs and SCB of HIV+ART and HIV^neg participants by RT-PCR. While IL-17A transcripts were undetectable in PBMCs from both HIV+ART and HIV^neg individuals, IL-17A mRNA expression was readily detectable in the colon of HIV+ART participants (Figure 2A), but at significant lower levels compared to HIV^neg (Figure 2B). Notably, levels of IL-32β and IL-17A transcripts were negatively correlated in SCB (Figure 2C). When SCB IL-32β and IL-17A mRNA levels were correlated with clinical parameters such as age, CD4 counts and time since infection, only SCB IL-32β levels negatively correlated with CD4 counts (Supplemental Table 2), consistent with previous studies in blood [41]. Finally, HIV-DNA load in colon-infiltrating CCR6⁺ T-cells, measured in our previous studies [70], did not correlate with IL-32β and IL-17A mRNA levels in SCB and PBMCs (Supplemental Table 3).

Fig. 2: Quantification of IL-17A mRNA in blood and colon of ART-treated PLWH.

RNA was extracted as described in Figure 1 was used to quantify IL-17A mRNA expression by real-time RT-PCR. (A) Shown are the IL-17A mRNA levels in PBMCs and SCB of ART-treated PLWH (A), as well as the statistical analysis of IL-17A mRNA expression HIV+ART (n=17) and HIV- participants (n=5) (B). Finally, the correlation between IL-32β mRNA (depicted in Figure 1B) and IL-17A mRNA levels (depicted in Figure 2B) was studied, using Spearman correlation (SC; p and r values) and linear regression (LR, p and r2 values) models (C).

Together, these results demonstrate that IL-32 overexpression in the colon of ART-treated PLWH is associated with the paucity of colon IL-17A and blood CD4 counts.

Viral and bacterial sensing promotes IL-32 expression by IECs.

A significant fraction of SCB cells are CD326⁺ IEC [70, 71, 73] (Supplemental Figure 1), which are reported to produce IL-32 during inflammatory conditions, including untreated HIV infection [42, 56, 65]. IEC sense microbial and inflammatory products to maintain homeostasis or to initiate immune response to threats [74, 75]. Here, we used the HT-29 colorectal cell line [76] as in vitro model to assess IEC capacity to produce IL-32 isoforms in response to inflammatory, bacterial and viral products. HT-29 cells were stimulated with rhTNF-α, an important inducer of IL-32 [46] and a key pro-inflammatory cytokine during HIV-1 infection [77], or exposed to NL4.3BaL and THRO HIV-1. While TNF-α was the strongest inducer of IL-32 mRNA, mainly the β/γ/ε isoforms, exposure to HIV strains also resulted in the upregulation of IL-32 mRNA in a dose-dependent manner (Figure 3A–B). Consistently, cell-associated total IL-32 protein levels were also upregulated upon IEC exposure to TNF-α and HIV (Figure 3C).

Figure 3: Identification of TLR ligands that trigger IL-32 mRNA and protein expression in the HT-29 IEC line.

The HT-29 IEC were stimulated with a set of TLR ligands (TLR1/2, TLR2, TLR3 HMW, TLR3 LMW, TLR4, TLR5, TLR6/2, TLR7, TLR8 and TLR9) for 24h. Total RNA was extracted from HT-29 cells for the quantification of IL-32 α, β, γ, D, ε, and θ mRNA by real-time RT-PCR. Shown are results from one representative experiment (A), and the statistical analysis of the fold changes in IL-32β, γ and ε mRNA expression in response to TLR triggering using results form 3 different experiments (B). Friedman p-values and Dunn’s multiple comparison statistical significance (versus Medium) are indicated on the graphs. Finally, cell-associated IL-32 protein levels were quantified by ELISA and normalized to total protein levels. Shown are results from one experiment representative of results generated from 3 independent experiments (C).

In addition, we further investigated the impact of TLR1–9 agonists on IL-32 expression by HT-29 cells. Our results demonstrated that TLR3 (Poly I:C HMW and LMW) and TLR5 (Flagellin from Salmonella typhimurium) ligands induced IL-32 mRNA expression, mainly the isoforms β/γ/ε (Figure 4A–B). Exposure to Poly I:C HMW induced the highest levels of cell-associated total IL-32 protein (Figure 4C).

Figure 4: Exposure to HIV trigger IL-32 mRNA expression in IEC.

HT-29 IEC were exposed to different concentrations (100, 50 and 10 ng HIV-p24/ml) of two CCR5-tropic replication-competent HIV-1 strains, NL4.3BAL and transmitted/founder (TF) THRO, for 72h. Total RNA was extracted from cell pellets for measurement of the IL-32α, β, γ, D, ε, and θ isoforms by real-time RT-PCR. Shown are results from one representative experiment (A), as well as statistical analysis of fold change increase in IL-32β, γ, and ε expression upon exposure to HIV using results from 3 independent experiments (B). Friedman p-values and Dunn’s multiple comparison statistical significance (versus Medium) are indicated on the graphs. Finally, cell-associated IL-32 protein levels were quantified by ELISA and normalized to total protein levels. Shown are results from one experiment representative of results generated from 3 independent experiments (C).

Together, these results demonstrate that IEC respond to pro-inflammatory, bacterial and viral stimuli by producing IL-32β/γ/ε isoforms, which likely fuels inflammation and immune activation during chronic HIV-1 infection.

IL-17A downregulates expression of IL-32 isoforms mRNA and total IL-32 protein in IECs.

The finding that IL-32 and IL-17A mRNA expressions negatively correlated in SCB of ART-treated PLWH (Figure 2C) suggests a potential mechanistic link between IL-17A and IL-32. To test this hypothesis, HT-29 cells were stimulated with rhTNF-α or Poly I:C HMW in the presence/absence of serial dilutions of rhIL-17A (100 to 1 ng/ml). While IL-17A alone did not modify IL-32 mRNA expression, IL-17A acted in a dose-dependent manner to significantly downregulate IL-32 mRNA expression induced upon stimulation with TNF-α and TLR3 triggering (Figure 5A–B). The same decrease was found for cell-associated IL-32 protein (Figure 5C).

Figure 5: IL-17A downregulates IL-32 expression in a dose-dependent manner.

The HT-29 IEC were stimulated with rhTNF-α and TLR3 agonist Poly(I:C) in the presence or the absence of various doses of rhIL-17A (100, 50, 10 and 1 ng/mL) for 24h. Total RNA was extracted from cell pellets and levels of IL-32α, β, γ, D, ε, and θ mRNA were quantified by real-time RT-PCR. Shown are results from one experiment representative of results generated in 3 independent experiments (A), as well as statistical analysis of fold change decrease in IL-32β, γ, and ε expression in response to TNF (B) and TLR3 ligands (C) upon exposure to rhIL-17A, using results from 3 independent experiments. Friedman p-values and Dunn’s multiple comparison statistical significance (versus Medium) are indicated on the graphs. (D) Cell-associated IL-32 protein levels were quantified by ELISA and normalized to total protein levels. Shown are results from one experiment representative of results generated from 3 independent experiments.

Altogether, these results reveal the important role of IL-17A in tuning down IL-32 overexpression in IEC exposed to various stimuli.

rhIL-32α/β/γ fail to trigger chemokine production in HT-29 cells

Other immune cells infiltrating the colon such as and myeloid cells are a source of IL-32 as well [41, 42, 45, 54]. Given the overt IL-32 expression in SCB of HIV+ART individuals (Figure 1), we aimed to determine whether exogenous IL-32 might act on IEC to promote inflammation. For this, we exposed HT-29 cells to rhIL-32α/β/γ (the only commercially available rhIL-32 isoforms), separately or rhTNF-α. Our results showed that the production of MIP-3α/CCL20 (a chemokine attractant for CCR6⁺ cells [78], such as Th17 cells [79] and dendritic cells [80, 81], and involved in the early steps of HIV transmission [80, 81]); IL-8/CXCL8 (a chemokine attracting CXCR1/CXCR2⁺ neutrophils to inflammation sites [82]); and IP-10/CXCL10 (a chemokine attracting inflammatory CXCR3⁺ cell subsets that represent a marker of HIV disease progression [83]) was induced in HT-29 cells upon exposure to rhTNF-α alone, but not affected by the addition of rhIL-32α/β/γ in the presence/absence of rhTNF-α (Supplemental Fig. 3A–C). We additionally measured TNF-α production, a hallmark cytokine induced by IL-32 [46]. While CD4⁺ T-cells produce TNF-α upon exposure to rhIL-32α/β/γ (data not shown), HT-29 cells did not produce TNF-α after IL-32 stimulation, neither separately nor in the presence of LPS (data not shown). Thus, we concluded that, while HT-29 IEC produced IL-32 transcripts and proteins, they are not responsive to the soluble rhIL-32 isoforms, at least for the chemokines measured.

Discussion

In this manuscript, we reveal the abundance of IL-32β/γ/ε isoforms, predominantly IL-32β, in relationship with the paucity of the Th17 hallmark cytokine IL-17A, in the colon of ART-treated PLWH. By using HT-29 cells as IEC model, we also reveal that IL-32β/γ/ε isoform expression is induced by exposure to pro-inflammatory, viral, and bacterial stimuli, and that rhIL-17A inhibits IL-32 expression/production in response to these stimuli. Moreover, IL-32 production was detected in cell lysates, but not in cell culture supernatants, indicating that IL-32 is not secreted by IEC. Finally, while HT-29 IEC produce IL-32, exogenous rhIL-32 α/β/γ isoforms fail to induce the production of inflammatory chemokines or TNF in these cells, raising the possibility that IEC lack the yet unidentified IL-32 receptor(s). Our results support a model in which IEC are a source of IL-32 under inflammatory conditions, with the IL-17A scarceness, likely related to the depletion of Th17 cells [25, 28], fuelling IL-32 overexpression at the intestinal levels during ART-treated HIV infection.

IL-32 is produced by several cell subsets, as NK, T-cells, monocytes and epithelial cells, with an increased expression under inflammatory conditions or infections [42, 54]. Previous reports demonstrated IL-32 overexpression in the plasma and PBMC of ART-treated PLWH when compared to uninfected individuals [41, 45, 57–62]. To our knowledge, our report is the first to reveal the expression of six specific IL-32 isoforms in the colon of ART-treated PLWH. Our finding that IL-32β was significantly overexpressed in the colon of ART-treated PLWH compared to uninfected individuals is consistent with our recent observation in the circulating blood [45]. Of note, IL-32β possesses high pro-inflammatory properties [45, 84–86].

IL-17A-producing Th17 cells are essential for mucosal immunity and epithelial cell barrier integrity [20, 87–90]. This population is rapidly depleted from the GALT during acute HIV infection, a deficit not fully restored by ART [11, 20–28]. Our group and others demonstrated that Th17 cells are highly permissive to HIV infection [21, 70, 91–94]. Consequently, the Th17 paucity in the GALT accounts for defective gut permeability and microbial translocation that stimulate chronic immune activation during HIV infection [18, 25, 28, 95, 96]. The relationship between IL-32 and IL-17A is being studied due to their pathological role in inflammatory diseases [97]. Hence, we quantified IL-17A mRNA in blood and SCB. While IL-17A was not expressed in the blood it was expressed in the colon of ART-treated PLWH; however, IL-17A expression was significantly downregulated in the colon of these individuals compared to uninfected controls. We discovered that these cytokines negatively correlate in the colon, indicative that sustained IL-32 overexpression coincides with IL-17A paucity in the GALT of ART-treated PLWH and might impact epithelial barrier integrity.

Previous studies reported that IL-32 is weakly expressed by IEC of healthy individuals; however, during chronic inflammation, IL-32 expression augments and correlates with increased disease severity [98]. The recent literature shows the role of IL-32 in gastric inflammatory disorders [42]. IL-32 biology is complex, with various splicing-generated isoforms that play different roles in different onsets [48, 50, 99]. Likewise, IL-32 is upregulated during HIV infection [41, 45, 60] and different isoforms might bring distinct outcomes for the disease [61]. Here, we run a complete TLR screening of six IL-32 isoforms. Our results demonstrate that IEC produced IL-32β/γ/ε mRNA and cell-associated IL-32 protein when stimulated with TLR3 (Poly I:C) and TLR5 (Flagellin) agonists; consistent with previous studies [67, 100–103]. LPS (TLR4 agonist) did not induce IL-32 production; this is consistent with the reported IEC resistance to LPS, to avoid chronic inflammation upon exposition to the commensal flora [104].

We questioned whether HIV exposure could increase gut inflammation via IL-32 production by IEC. Our results showed that HIV induced IL-32 expression in a dose-dependent manner, especially the IL-32β/γ/ε isoforms. The same increase was observed measuring cell-associated IL-32 protein. Our results also revealed that IL-32 is mainly acts intracellularly and poorly secreted by HT-29 cells (data not shown), consistent with previous reports [54, 65]. Although IEC [105–107], including the HT-29 line [76, 108–111], can become productively infected by HIV, via the receptor galactosylceramide, IEC principally contribute to HIV dissemination via transcytosis [105, 112, 113]. Whether integrative infection is required for IL-32 expression in IEC remains to be investigated. Our data strongly indicate that IL-32 is produced in response to viral threats, and microbial products, thus enhancing intestinal inflammation.

One important finding of our work is the negative correlation between IL-32 and IL-17 in the colon of ART-treated PLWH. IL-17A acts on IEC via its receptors IL-17RA/RC to promote mucosal immunity and tissue integrity [90]. Exploring the relationship between these cytokines, we found that IL-17A significantly decreased TNF-α-induced IL-32 production in a dose-dependent manner, mainly IL-32α/β/γ/ε isoforms. The same decrease was observed for TLR3 agonist-stimulated cells. Similarly, IL-17A decreased cell-associated IL-32 protein production. These results indicate the ability of IL-17A in controlling IL-32-mediated inflammation in IEC. Thus, we hypothesize that IL-17A-producing Th17 cell paucity during HIV infection favors IL-32 overexpression upon interaction with microbial products. Our findings are consistent with studies by Koeken and collaborators showing that IL-32 inversely correlated with IL-17A during Mycobacterium tuberculosis infection; in this context where elevated IL-17A is pathogenic, IL-32 expression may control Th17-mediated lung damage during tuberculosis [114]. In contrast, Moon and collaborators found that IL-17 and IL-32 reciprocally influence each other’s expression, act in synergy to induce osteoclast differentiation, and together aggravate rheumatoid arthritis pathogenesis [115]. Thus, there are cell-specific differences in the capacity of IL-32 and IL-17A to mechanistically interact.

One limitation of our study is the fact that we did not identify cellular sources of IL-32 in SCB. Although CD326⁺ IEC represent an abundant fraction of SCB, consistent with our previous studies [70, 71, 73], we were unable to sort viable IEC by flow cytometry. Alternative strategies, such as in situ RNAscope®, should be used in future studies to identify cellular sources of IL-32 in small SCB of ART-treated PLWH. Another limitation of our study was the sample relatively low sample size of SCB from HIV+ART individuals; this limited the ability to link IL-32β and IL-17A mRNA levels in SCB with clinical parameters.

In conclusion, we document the IL-32β overexpression in the colon of ART-treated PLWH and indicate IEC as an important source of IL-32, induced by inflammatory, bacterial and viral triggers. The capacity of IL-17A to decrease IL-32 expression in IEC indicates the deleterious consequence of Th17 paucity, thus fueling IL-32-dependent inflammation in the GALT. Our results highlight the importance of Th17 replenishment during HIV infection to avoid IL-32-mediated non-AIDS comorbidities. Future studies are needed to identify molecular mechanisms of IL-32 downregulation by IL-17A. Such insights will lead towards the identification of new druggable targets for the restoration of mucosal homeostasis in PLWH receiving ART.