Compartment-specific distribution of human intestinal innate lymphoid cells is altered in HIV patients under effective therapy

Abstract

Innate lymphocyte cells (ILCs), a novel family of innate immune cells are considered to function as key orchestrators of immune defences at mucosal surfaces and to be crucial for maintaining an intact intestinal barrier. Accordingly, first data suggest depletion of ILCs to be involved in human immunodeficiency virus (HIV)-associated damage of the intestinal mucosa and subsequent microbial translocation. However, although ILCs are preferentially localized at mucosal surfaces, only little is known regarding distribution and function of ILCs in the human gastrointestinal tract. Here, we show that in HIV(-) individuals composition and functional capacity of intestinal ILCs is compartment-specific with group 1 ILCs representing the major fraction in the upper gastrointestinal (GI) tract, whereas ILC3 are the predominant population in ileum and colon, respectively. In addition, we present first data indicating that local cytokine concentrations, especially that of IL-7, might modulate composition of gut ILCs. Distribution of intestinal ILCs was significantly altered in HIV patients, who displayed decreased frequency of total ILCs in ileum and colon owing to reduced numbers of both CD127(+)ILC1 and ILC3. Of note, frequency of colonic ILC3 was inversely correlated with serum levels of I-FABP and sCD14, surrogate markers for loss of gut barrier integrity and microbial translocation, respectively. Both expression of the IL-7 receptor CD127 on ILCs as well as mucosal IL-7 mRNA levels were decreased in HIV(+) patients, especially in those parts of the GI tract with reduced ILC frequencies, suggesting that impaired IL-7 responses of ILCs might contribute to incomplete reconstitution of ILCs under effective anti-retroviral therapy. This is the first report comparing distribution and function of ILCs along the intestinal mucosa of the entire human gastrointestinal tract in HIV(+) and HIV(-) individuals.

Fig 1. Characterisation of human ILC subsets in intestinal biopsy specimen.

(A) Representative FACS plots showing the gating strategy from singlet lymphocytes to the ILC1 (blue), ILC2 (green), and ILC3 (violet) populations. (B) Gating strategy used for the identification of CD103(+)ILC1 (red) and conventional NK (cNK) cells (orange). (C) Expression of the transcription factors T-bet, RORγt, GATA, and Eomes within intestinal CD127(+)ILC1 (blue), CD103(+)ILC1 (red), ILC2 (green), ILC3 (violet), and cNK cells (orange). (D) Production of IFN-γ, IL-13, and IL-22, respectively, in PMA/ ionomycin activated intestinal CD127(+)ILC1 (blue), CD103(+)ILC1 (red), ILC2 (green), ILC3 (violet), and cNK cells (orange). Due to low frequencies pooled samples of three donors had to be studied in the case of ILC2.

Fig 2. Compartment-specific distribution and function of intestinal ILCs in HIV(-) individuals.

(A) Frequency of ILCs (black dots, left panel) and cNK cells (white dots; right panel) expressed as percentage of CD45(+) lymphocytes in the different parts of the human GI tract (ILCs: oesophagus, n = 10; stomach n = 15; duodenum, n = 21; ileum, n = 13; colon: n = 13; cNK: oesophagus, n = 12; stomach n = 15; duodenum, n = 21; ileum, n = 14; colon: n = 13; grey columns) in comparison to peripheral blood (n = 16; white columns). (B) Distribution of CD103(+)ILC1 (upper left panel), CD127(+)ILC1 (upper right panel), ILC2 (lower left panel), and ILC3 (lower right panel) in the human intestinal tract (black columns) and the circulating blood (white columns). (C) Composition of CD127(+)ILC1 (blue), intraepithelial ILC1 (red), ILC2 (green) and ILC3 (violet) in the peripheral blood and the different segments of the gastrointestinal tract. (D) Distribution of NCR(-) (left panel) and NCR(+) ILC3 (right panel) in the peripheral blood and the different segments of the gastrointestinal tract. (E) Percentage comparison between NCR(-) and NCR(+) ILC3 subsets in those compartments. For functional characterisation ILCs derived from PBMC or biopsy samples (oesophagus, stomach, duodenum, Ileum and colon) obtained from HIV(-) individuals, were stimulated with PMA/ionomycin and then tested for IFN-γ production in CD103(+)ILC1 (F), CD127(+)ILC1 (G),or IL-22 production in ILC3 (H), respectively. * p≤0.05; ** p≤0.01; *** p≤0.001 (FDR-adjusted P values, parametric ANOVA).

Fig 3. Local IL-7 levels and composition of the intestinal ILC pool in HIV(-) individuals.

(A) Mucosal mRNA expression levels of IL-7 (left panel) and IL-15 (right panel) were analysed in the oesophagus (n = 7), the stomach (n = 9), the duodenum (n = 14), the ileum (n = 12), and the colon (n = 10), respectively (* p≤0.05; ** p≤0.01; *** p≤0.001 (FDR-adjusted P values, parametric ANOVA). Cytokine mRNA expression levels are expressed in relation to expression levels of beta-actin. (B) Compartment-specific IL-7mRNA levels (green) in comparison to frequency of total ILCs (orange; left panels; oesophagus (n = 6), stomach (n = 7), duodenum (n = 10), ileum (n = 9), and colon (n = 8)) and cNK cells (brown; right panels; oesophagus (n = 6), stomach (n = 7), duodenum (n = 9), ileum (n = 9), and colon (n = 7)) in the different segments of the GI tract. The right panels depict correlation between local IL-7 mRNA levels and ILC (upper panel) or cNK cells frequency (lower panel), respectively. Pearson correlation coefficient was computed for analysing correlations (right panels).(C) Relation between mucosal IL-15 mRNA expression levels (violet) and frequency of total ILCs (orange; left panels; oesophagus (n = 6), stomach (n = 7), duodenum (n = 10), ileum (n = 9), and colon (n = 8)) and cNK cells (brown; right panels; oesophagus (n = 6), stomach (n = 7), duodenum (n = 9), ileum (n = 9), and colon (n = 7)) in the different segments of the GI tract. (D) Local IL-7 mRNA concentrations (green) in relation to frequency of CD127(+)ILC1 (blue; upper left panel), CD103(+)ILC1 (coppery; upper right panel), ILC2 (lime green; lower left panel), and ILC3 (violet; lower right panel).; oesophagus (n = 6), stomach (n = 7), duodenum (n = 10), ileum (n = 9), and colon (n = 8). Pearson correlation coefficient was computed for analysing correlations. (E) Compartment-specific IL-7mRNA levels (green) in comparison to frequency of ILC3 (violet; left panel; oesophagus (n = 6), stomach (n = 7), duodenum (n = 10), ileum (n = 9), and colon (n = 8) in the different segments of the GI tract. The right panel depicts correlation between local IL-7 mRNA levels and ILC3 frequency. Pearson correlation coefficient was computed for analysing correlations.

Fig 4. Local IL-7 levels and segment-specific characteristics of intestinal ILCs in HIV(-) individuals.

(A) Relation between local IL-7 mRNA concentrations (green) and expression of RORγt in ILC3 (violet). Expression of RORγt is given as mean fluorescence intensity (MFI).oesophagus (n = 5), stomach (n = 7), duodenum (n = 6), ileum (n = 7), and colon (n = 6). (B) In vitro effect of increasing concentrations of recombinant IL-7 (0, 10, 50 ng/ml) expression on RORγt expression in tonsil ILC3. (C) Relation between local IL-7 mRNA concentrations (green) and expression of T-bet (MFI; blue) within intestinal CD127(+)ILC1. oesophagus (n = 5), stomach (n = 6), duodenum (n = 7), ileum (n = 7), and colon (n = 6)(D) Association between CD127 surface expression (MFI) on CD127(+)ILC1 (left panel; blue) and ILC3 (right panel; violet), respectively, and IL-7 mRNA concentrations (green). oesophagus (n = 6), stomach (n = 7), duodenum (n = 10), ileum (n = 9), and colon (n = 8) (E) In vitro effect of increasing concentrations of recombinant IL-7 (0, 10, 50 ng/ml) on surface expression of CD127. (F) CD127 surface expression (MFI) on ILC3 (violet;) compared to that of CD127(+)ILC1 (blue). oesophagus (n = 6), stomach (n = 7), duodenum (n = 10), ileum (n = 9), and colon (n = 8). Pearson correlation coefficient was computed for analysing correlations.

Fig 5. Alterations of the intestinal ILC pool in HIV(+) individuals.

(A) Frequency of ILCs in the peripheral blood of HIV(-) controls (blue; n = 19) in comparison to HIV patients (red; n = 20) under effective cART (ns, not significant). (B) Frequency of total ILCs in the different segments of the GI tract in HIV(-) controls compared to HIV patients (* p≤0.05). oesophagus: HIV(-) n = 12, HIV(+) n = 9; stomach HIV(-) n = 15, HIV(+) n = 15; duodenum: HIV(-) n = 21, HIV(+) n = 19; ileum: HIV(-) n = 15, HIV(+) n = 16; colon: HIV(-) n = 13, HIV(+) n = 17. (C) Frequency of cNK cells (left panel) and CD4(+) T cells (right panel) in the different segments of the GI tract in HIV(-) controls compared to HIV patients. cNK cells: oesophagus: HIV(-) n = 12, HIV(+) n = 10; stomach HIV(-) n = 15, HIV(+) n = 15; duodenum: HIV(-) n = 21, HIV(+) n = 19; ileum: HIV(-) n = 13, HIV(+) n = 16; colon: HIV(-) n = 13, HIV(+) n = 17. CD4(+) T cells: HIV(-) n = 9, HIV(+) n = 9; stomach HIV(-) n = 12, HIV(+) n = 15; duodenum: HIV(-) n = 18, HIV(+) n = 18; ileum: HIV(-) n = 10, HIV(+) n = 17; colon: HIV(-) n = 11, HIV(+) n = 18. (D) Frequency of CD103(+)ILC1 (upper left panel), ILC2 (upper right panel), CD127(+)ILC1 (lower left panel), and ILC3 (lower right panel; in the different segments of the GI tract in HIV(-) controls compared to HIV patients. CD103(+)ILC1: oesophagus: HIV(-) n = 15, HIV(+) n = 10; stomach HIV(-) n = 17, HIV(+) n = 16; duodenum: HIV(-) n = 31, HIV(+) n = 22; ileum: HIV(-) n = 17, HIV(+) n = 15; colon: HIV(-) n = 18, HIV(+) n = 16. CD127(+)ILC1: oesophagus: HIV(-) n = 15, HIV(+) n = 10; stomach HIV(-) n = 16, HIV(+) n = 16; duodenum: HIV(-) n = 31, HIV(+) n = 22; ileum: HIV(-) n = 17, HIV(+) n = 15; colon: HIV(-) n = 18, HIV(+) n = 16. ILC2: oesophagus: HIV(-) n = 15, HIV(+) n = 10; stomach HIV(-) n = 16, HIV(+) n = 16; duodenum: HIV(-) n = 31, HIV(+) n = 22; ileum: HIV(-) n = 17, HIV(+) n = 15; colon: HIV(-) n = 18, HIV(+) n = 16. ILC3: oesophagus: HIV(-) n = 15, HIV(+) n = 10; stomach HIV(-) n = 17, HIV(+) n = 16; duodenum: HIV(-) n = 30, HIV(+) n = 22; ileum: HIV(-) n = 19, HIV(+) n = 15; colon: HIV(-) n = 19, HIV(+) n = 16. (E) Frequency of NCR(+) (upper panel) and NCR(-) ILC3 (lower panel) in the different segments of the GI tract in HIV(-) controls compared to HIV patients. NCR(+) ILC3: oesophagus: HIV(-) n = 15, HIV(+) n = 10; stomach HIV(-) n = 17, HIV(+) n = 16; duodenum: HIV(-) n = 30, HIV(+) n = 22; ileum: HIV(-) n = 19, HIV(+) n = 15; colon: HIV(-) n = 19, HIV(+) n = 16. NCR(-) ILC3: oesophagus: HIV(-) n = 15, HIV(+) n = 10; stomach HIV(-) n = 17, HIV(+) n = 16; duodenum: HIV(-) n = 30, HIV(+) n = 22; ileum: HIV(-) n = 19, HIV(+) n = 15; colon: HIV(-) n = 19, HIV(+) n = 16. (F) Functional activity of PMA/ ionomycin-stimulated ILC subsets (left panel: IFN-γ(+)CD127(+)ILC1; middle panel: IFN-γ(+)CD103(+)ILC1; right panel: IL-22(+) ILC3) isolated from the duodenum, the ileum, and the colon, respectively in HIV(+) patients and HIV(-) individuals. * p≤0.05; ** p≤0.01; *** p≤0.001 (FDR-adjusted P values, parametric ANOVA).

Fig 6. Expression of IL-7mRNA and CD127 in HIV(+) individuals.

IL-15 (A) and IL-7 (B) mRNA expression was studied in the different compartments of the intestinal tract in HIV(+) patients (red) compared to HIV(-) controls (blue). oesophagus: HIV(-) n = 7, HIV(+) n = 6; stomach HIV(-) n = 10, HIV(+) n = 10; duodenum: HIV(-) n = 14, HIV(+) n = 11; ileum: HIV(-) n = 12, HIV(+) n = 15; colon: HIV(-) n = 10, HIV(+) n = 10) (C) Surface expression of CD127, expressed as mean fluorescence intensity (MFI), on ILCs isolated from different parts of the GI tract in HIV(+) patients and HIV(-) individuals duodenum: HIV(-) n = 10, HIV(+) n = 10; ileum: HIV(-) n = 10, HIV(+) n = 10; colon: HIV(-) n = 10, HIV(+) n = 11) (* p≤0.05). (FDR-adjusted P values, parametric ANOVA).

Fig 7. Association between mRNA expression of inflammatory cytokines and frequency as well as functional capacity of the intestinal ILC3 in HIV(+) individuals.

(A) Gene expression levels of CXCL6, IL-18, TGF-β1/3, IL-1β, IFN-γ, and TNF-α were studied in duodenum and colon samples (* p≤0.05). (B) Correlation between IL-1B mRNA expression and frequency of colon ILC3. (C) Correlation between IL-1B expression (upper panel) and IFN-γ expression (lower panel) and IL-22 production of colon ILC3.Pearson correlation coefficient was computed for analysing correlations.

Fig 8. Association between plasma markers of gut epithelial damage/ microbial translocation and colon NCR(+)ILC3 frequency in HIV(+) individuals.

(A) Plasma serum levels of intestinal fatty acid binding protein (I-FABP; ng/ml), a marker for gut barrier defects, sCD14 (μg/ml), sCD163 (μg/ml), and LPS binding protein (LBP, μg/ml) were quantified by ELISA in HIV patients (red) and HIV(-) controls (blue) (* p≤0.05). Next, association between plasma levels of I-FABP (ng/ml) (B) and sCD14 (μg/ml) (C), respectively, and frequency of colon NCR(+)ILC3 (left panel) or colon NCR(-)ILC3 (right panel) were examined. Pearson correlation coefficient was computed for analysing correlations.