Rectal mucosal inflammation, microbiome, and wound healing in men who have sex with men who engage in receptive anal intercourse

Abstract

Mucosal injury is common during consensual intercourse and induces an inflammatory response that could contribute to pathogen transmission including HIV. Here, we compared mucosal immune and microbiome responses to experimentally induced mucosal injury between men who have sex with men engaging in receptive anal intercourse (MSM-RAI) and men who do not engage in RAI (controls), all without HIV. Rectal mucosal secretions were collected from adult MSM-RAI (n = 19) and controls (n = 6) via anoscopy before and up to eight days after experimentally induced injury. Mucosal healing was evaluated by repeated injury surface area measurements with digital imaging. MSM-RAI demonstrated overall significantly higher concentrations of pro-inflammatory cytokines and a distinct rectal microbiome compared with controls. Wound healing was numerically faster in MSM-RAI but did not meet statistical significance (p = 0.09). Different cytokine injury response patterns were observed between MSM-RAI and controls; however, IL-6 and IP-10 were important mediators in both groups. Microbial guilds, particularly from the Lachnospiraceae and Prevotellaceae families, were associated with rectal mucosal inflammation. This work is the first experimental study of rectal mucosal injury and the immune environment in healthy humans and provides a more nuanced understanding of rectal mucosal inflammation after injury, which can inform our understanding of HIV transmission.

Keywords: Bacterial guilds; Microbiome; Mucosal healing; Mucosal immunology; Receptive anal intercourse.

Fig. 1

Rectal mucosal cytokine responses to experimentally induced injury in MSM engaging in RAI and controls. Twelve rectal mucosal cytokines were measured at baseline (day 0) and after experimental rectal mucosal injury (days 2, 5, and 8). Nine cytokines were classified as primarily proinflammatory (IL-1β, IL-17A, IP-10, IL-8, IFN-γ, TNF-α, IL-12p70, IL-6, MCP-1), and three cytokines were classified as primarily anti-inflammatory (IL-4, IL-10, TGF-β1). (A) Concentrations were compared between MSM who have frequent RAI (blue dots) and control men who have never had RAI (red squares) using LDM. (B) A 12-cytokine inflammation score (12C IS) was calculated incorporating the nine proinflammatory and three anti-inflammatory individual cytokines for each participant at each visit. Scores were again compared between MSM who have frequent RAI and controls using LDM. Error bars indicate 95% confidence intervals.

Fig. 2

Rectal mucosal healing time after experimentally induced injury among MSM engaging in RAI and controls. (A) Images of experimentally induced rectal mucosal injuries were collected at baseline (day 0) and at days 2, 5, and 8, and (B) surface area in pixels was measured using ImageJ by outlining the wound boundary in yellow. Surface area over time was compared between MSM who have frequent RAI (blue dots) and control men who have never had RAI (red squares) using a linear mixed effects model. (C) depicts wound healing trajectories for all participants. The MSM-RAI and control splines are represented by thick lines.

Fig. 3

The composition of the rectal mucosal microbiome among MSM engaging in RAI and controls. Rectal microbiome composition was measured at baseline (day 0) and after experimental rectal mucosal injury (days 2, 5, and 8). (A) Bray-Curtis and Jaccard measures of beta diversity, measured by PERMANOVA and visualized with PCoA plots, both demonstrated significant differences between the overall microbiome composition of MSM-RAI (blue dots) and controls (red dots). (B) LDMs assessing relative abundance and probability of presence both demonstrated highly significant (relative abundance p = 0.007; presence-absence p = 0.003) differences between MSM-RAI and controls. The taxa listed were significant (FDR < 20%) contributors to these differences. Taxa were sequenced to the genus level when possible; those sequenced to the family (f) or order (o) level are indicated.

Fig. 4

Associations between the microbiome composition and inflammation as measured by 12 cytokine inflammation score (12C IS). Pearson scores were calculated for each of the 26 taxa that were significantly correlated (FDR < 20%) with the 12C IS. The twenty taxa on the left side of the x-axis had a negative association between RA and 12 C IS. The six taxa on the right side had a positive association between RA and 12 C IS. BOUTH analysis at all available phylogenetic levels identified the Lachnospiraceae and Prevotellaceae families (p < 0.001 for both) as the strongest contributors to the relative abundance-12C IS LDM enrichment signal. Twelve of the 26 inflammation-associated taxa were in the Lachnospiraceae family (eight taxa, in green) or Prevotellaceae family (four taxa, in purple). SparCC network analysis identified nine of these 12 taxa as nodes of inflammation-associated guilds (*).

Fig. 5

Inflammation-associated bacterial guilds among controls and MSM engaging in RAI. Flourish plots were created to visualize representative inflammation-associated guilds by mapping SparCC correlation coefficients > 0.7, represented by thick lines, between the inflammation associated nodes (*) and other bacterial taxa. Circles represent bacterial taxa and are color-coded by family. (A) represents a 5-taxa guild that is associated with increased RM inflammation score (12C IS) and is in decreased abundance in controls. (B) represents a 13-taxa guild that is associated with decreased 12C IS and is in decreased abundance in MSM-RAI.