Preservation of Gastrointestinal Mucosal Barrier Function and Microbiome in Patients With Controlled HIV Infection

Abstract

Background: Despite successful ART in people living with HIV infection (PLHIV) they experience increased morbidity and mortality compared with HIV-negative controls. A dominant paradigm is that gut-associated lymphatic tissue (GALT) destruction at the time of primary HIV infection leads to loss of gut integrity, pathological microbial translocation across the compromised gastrointestinal barrier and, consequently, systemic inflammation. We aimed to identify and measure specific changes in the gastrointestinal barrier that might allow bacterial translocation, and their persistence despite initiation of antiretroviral therapy (ART).

Method: We conducted a cross-sectional study of the gastrointestinal (GIT) barrier in PLHIV and HIV-uninfected controls (HUC). The GIT barrier was assessed as follows: in vivo mucosal imaging using confocal endomicroscopy (CEM); the immunophenotype of GIT and circulating lymphocytes; the gut microbiome; and plasma inflammation markers Tumour Necrosis Factor-α (TNF-α) and Interleukin-6 (IL-6); and the microbial translocation marker sCD14.

Results: A cohort of PLHIV who initiated ART early, during primary HIV infection (PHI), n=5), and late (chronic HIV infection (CHI), n=7) infection were evaluated for the differential effects of the stage of ART initiation on the GIT barrier compared with HUC (n=6). We observed a significant decrease in the CD4 T-cell count of CHI patients in the left colon (p=0.03) and a trend to a decrease in the terminal ileum (p=0.13). We did not find evidence of increased epithelial permeability by CEM. No significant differences were found in microbial translocation or inflammatory markers in plasma. In gut biopsies, CD8 T-cells, including resident intraepithelial CD103+ cells, did not show any significant elevation of activation in PLHIV, compared to HUC. The majority of residual circulating activated CD38+HLA-DR+ CD8 T-cells did not exhibit gut-homing integrins α4ß7, suggesting that they did not originate in GALT. A significant reduction in the evenness of species distribution in the microbiome of CHI subjects (p=0.016) was observed, with significantly higher relative abundance of the genus Spirochaeta in PHI subjects (p=0.042).

Conclusion: These data suggest that substantial, non-specific increases in epithelial permeability may not be the most important mechanism of HIV-associated immune activation in well-controlled HIV-positive patients on antiretroviral therapy. Changes in gut microbiota warrant further study.

Keywords: CD4; HIV; antiretroviral therapy (ART); gut-associated lymphoid tissues (GALT); microbiome.

Figure 1

Example images of terminal ileal villi exhibiting (A) fluorescein leak and (B) cell junction enhancement (A) Cell junction enhancement, characterised by an area of increased fluorescence from the basal to the apical surface between two epithelial cells. (B) Fluorescein leak, characterised by increased fluorescence in the epithelial cell layer, and a distinct plume of contrast leakage into the lumen.

Figure 2

Patient recruitment flowchart outlining selection of patients for confocal endomicroscopy imaging, tight junction, blood and tissue lymphocyte, plasma marker and faecal microbiome analyses. HUC, HIV-uninfected control; PHI, HIV-positive participant who initiated antiretroviral therapy during primary HIV infection; CHI, HIV-positive participant who initiated antiretroviral therapy during chronic HIV infection.

Figure 3

Images with identified confocal endomicroscopy features (fluorescein leak and cell junction enhancement) as a percentage of total images analysed for the participant, across three groups. Median percentage of images was 0.26% for HIV-uninfected controls (HUC), 0.00% for HIV-positive participants treated in primary infection (PHI), and 0.11% for HIV-positive participants treated in chronic infection (CHI). Statistical analysis was conducted using Kruskal Wallis one-way analysis of variation.

Figure 4

Changes in CD4 panels (A–C) and CD8 (D–F) cell numbers in the terminal ileum and left colon between patient groups as: (A, D) an absolute count from ten biopsies; (B, E) standardised per milligram of tissue; and (C, F) standardized by epithelial cell number HUC, HIV-uninfected controls; PHI, HIV-positive patients treated during primary infection; CHI, HIV-positive patients treated during chronic infection; TI, terminal ileum; LC, left colon, *statistically significant (p ≤ 0.05). Statistical analysis was conducted using Kruskal Wallis one-way analysis of variance with pairwise post-hoc testing using Mann Whitney U.

Figure 5

Activated CD38+HLA-DR+ CD8 T-cells in blood and tissues. (A) Activated CD38+HLA-DR+ as % of CD45RA- CD8 T cells in peripheral blood, by study group. HUC, HIV-uninfected controls; PHI, HIV-positive patients treated during primary infection; CHI, HIV-positive patients treated during chronic infection. The Comparator Groups are (left) normal range for HUC subjects from reference (38) and the 95^th percentile (4.6%) is shown as the dotted horizontal line, and (right) PINT subjects longitudinal observations from reference (38). Statistical analysis between HUC, PHI and CHI groups was conducted using Kruskal Wallis one-way analysis of variance. Statistical analysis between comparator normal range and PINT groups was done using Mann-Whitney test. (B) Comparison of activated CD38+HLA-DR+ as % of CD45RA- CD8 T-cells in peripheral blood, TI biopsies and LC biopsies, respectively by study group. (C) Percentage of activated CD38+HLA-DR+ CD8 T-cells in TI and LC biopsies that are also CD103+. (D) Percentage of activated CD38+HLA-DR+ CD8 T-cells in peripheral blood that are either CD49d+integrin ß7+ gut-homing or CD49d+integrin ß7+-negative non-gut-homing, by study group. (E) Correlation of activated CD38+HLA-DR+ as % of CD8 T-cells in peripheral blood with activated CD38+HLA-DR+ as % of CD8 T cells in TI biopsies. (F) Correlation of activated CD38+HLA-DR+ as % of CD8 T-cells in peripheral blood with activated CD38+HLA-DR+ as % of CD8 T cells in LC biopsies. (G) Correlation of CD103+ as % of activated CD38+HLA-DR+ CD8 T-cells in TI biopsies with activated CD38+HLA-DR+ as % of CD8 T cells in TI biopsies. (H) Correlation of CD103+ as % of activated CD38+HLA-DR+ CD8 T-cells in LC biopsies with activated CD38+HLA-DR+ as % of CD8 T cells in LC biopsies. ***p < 0.0001.

Figure 6

Levels of soluble proteins measured in plasma between the study groups. (A) Levels of soluble CD14 measured in plasma between the study groups. (B) Levels of TNF-α measured in plasma between the study groups. (C) Levels of IL-6 measured in plasma between the study groups. Statistical significance was calculated using the Kruskal Wallis test.

Figure 7

Faecal microbiota composition. (A) Bacterial composition from analysis of faecal samples from study participants prior to bowel preparation. A total of 145 types of taxa are represented, with increased representation of fermentative taxa (i.e., Ruminococcaceae, Faecalibacterium, Prevotella, Bacteroides) compared to other taxa. (B) Species diversity and taxa distribution in study participants as measured by Shannon diversity index (H’) (C) Species diversity and taxa distribution in study participants as measured by Simpson diversity index E (D) Relative abundance of the Spirochaeta genus between groups. (E) Relative abundance of the Acidaminococcus genus between groups.