HIV, Sexual Orientation and Gut Microbiome Interactions

Abstract

Recent studies have raised interest in the possibility that dysbiosis of the gut microbiome (i.e. the communities of bacteria residing in the intestine) in HIV-infected patients could contribute to chronic immune activation, and, thus, to elevated mortality and increased risk of inflammation-related clinical diseases (e.g. stroke, cardiovascular disease, cancer, long-bone fractures and renal dysfunction) found even in those on effective antiretroviral therapy. Yet, to date, a consistent pattern of HIV-associated dysbiosis has not been identified. What is becoming clear, however, is that status as a man who has sex with men (MSM) may profoundly impact the structure of the gut microbiota, and that this factor likely confounded many HIV-related intestinal microbiome studies. However, what factor associated with MSM status drives these gut microbiota-related changes is unclear, and what impact, if any, these changes may have on the health of MSM is unknown. In this review, we outline available data on changes in the structure of the gut microbiome in HIV, based on studies that controlled for MSM status. We then examine what is known regarding the gut microbiota in MSM, and consider possible implications for research and the health of this population. Lastly, we discuss knowledge gaps and needed future studies.

INTRODUCTION

Interest in rectal and colonic pathogens in men who have sex with men (MSM) rose in the late 1970’s, (just prior to the start of the HIV epidemic in the United States), with the publication of a case series by Kazal et al. describing a “gay bowel syndrome” in MSM presenting with anorectal complaints.[1] The term, which is now outdated and considered pejorative,[2] did not truly describe a syndrome, but rather a list of conditions, including anorectal ulcers and proctitis/colitis, which were observed with increased frequency in MSM as compared to men who have sex with women only (MSW).[3] Further testing revealed that these conditions were primarily caused by sexually transmitted infections (STIs) (including rectal gonorrhea, chlamydia, syphilis, herpes simplex virus and human papillomavirus) as well as enteric pathogens (e.g. Shigella, Entamoeba histolytica, Giardia, Campylobacter), whose transmission was related to sexual activity in MSM.[4, 5]

With the advent of the HIV epidemic, organisms such as Cryptosporidium spp. became increasingly associated with gastrointestinal (GI) complaints in HIV-positive (HIV+) MSM.[6] Furthermore, the high prevalence of GI symptoms, GI-related opportunistic infections, as well as intestinal immune dysfunction and enteropathy observed in HIV+ patients focused attention on gut microbes in both MSM and heterosexual populations.[7] As the epidemic progressed, the development of antiretroviral therapy (ART) revolutionized care of HIV+ patients, but mortality in HIV+ individuals remained higher than in the HIV-uninfected (HIV−) population.[8] In this context, studies describing intestinal barrier disruption and increased bacterial translocation from the intestine in HIV+ individuals[9] fueled renewed interest in gut bacteria. Specifically, it was hypothesized that dysbiosis of the gut microbiome (i.e. the communities of bacteria residing in the intestine) could contribute to chronic immune activation, and thus ultimately to the increased risk of inflammation-related clinical diseases (e.g. stroke, cardiovascular disease, cancer, long-bone fractures and renal dysfunction) seen even in patients on effective ART.[10, 11]

A rapid proliferation of studies utilizing new molecular techniques to examine the association between the gut microbiota and HIV status ensued. Yet, to date, a consistent pattern of gut dysbiosis associated with HIV infection has not been identified.[12] What is becoming clear, however, is that MSM status may profoundly impact the structure of the gut microbiota, (perhaps more strongly than HIV), and that this factor likely confounded many HIV-related intestinal microbiome studies.[12–20] Early analyses suggest that the gut microbiome of at least a subset of MSM is marked by characteristic changes, including increases in the relative abundance of the genus Prevotella,[13] but what leads to this distinct microbial signature is unclear. In this review, we will briefly outline available data on changes in the structure of the gut microbiome in HIV based on studies that controlled for MSM status. We will then focus on what is known regarding the gut microbiota in MSM, and consider possible implications for research and the health of MSM. Lastly, we will discuss knowledge gaps and needed future studies.

HIV and the Gut Microbiome: No Consistent HIV-linked Dysbiosis Identified, MSM Status as a Confounder

The increasing availability of ART has shifted the focus of HIV care in many settings from treating opportunistic infections to managing the chronic complications of long-term survivorship with the disease. Unfortunately, the life expectancy of HIV+ individuals, even on effective ART, remains lower than that of HIV− individuals.[8, 21] Much of this discrepancy appears driven by an increased risk in HIV+ patients for inflammation-related clinical diseases.[8, 9, 22] Why low level, usually asymptomatic inflammation and immune activation persists despite ART is unknown.[23] However, some studies have linked intestinal barrier dysfunction and markers of gut bacterial translocation to inflammation and long term chronic complications in HIV+ patients.[9, 24–26] The advent of new molecular techniques and rapid expansion of research on the gut microbiome as a mediator of immune function,[27] coupled with these observations, have led to interest in the possibility that gut microbial dysbiosis may contribute to chronic immune activation in HIV.[10, 11] Significant excitement has surrounded the prospect of improving the health of HIV+ patients through manipulation of the gut microbiota. Indeed, several probiotic trials or trials of fecal microbiota transplant (FMT) in HIV+ patients have been conducted or are underway [28–49], though initial results have been mixed.[34, 38]

Yet, despite a proliferation of recent papers in the field,[12–16, 18, 19, 32, 38, 50–77] a consistent pattern of gut dysbiosis in HIV+ patients has not been defined. Most papers utilize two types of analyses: alpha-diversity (e.g., operational taxonomic units (OTUs), Chao1, Shannon and/or inverse Simpson measures), and individual bacterial taxa changes (e.g., at the phylum, family, genus and less often species level) to characterize the fecal microbiome in differing populations. Alpha-diversity measures provide a sense of the complexity (OTUs or number of different species in the stool) and/or the evenness (frequency) of the detected OTUs. Increased alpha-diversity is generally considered to be a marker of health, whereas decreased diversity has been associated with a variety of disease states.[78–81] Many gut microbiome-related studies in HIV+ individuals have been variable in design, utilizing overlapping, small and/or heterogeneous cohorts. For example, sampled populations in single studies have been diverse in composition by gender, sexual orientation and stage of HIV infection (e.g., ranging from treatment-naïve viremic to virally suppressed on ART). Further technical aspects that can impact the results also vary including sequencing techniques, and statistical analyses. Perhaps as a consequence, and not surprisingly, results have been inconsistent.

Potentially most importantly, some studies have been confounded by failing to control for MSM status.[82] Many papers (most of which did not control for MSM status) have demonstrated increases in the relative abundance of the genus Prevotella in HIV+ as compared to HIV− patients.[51, 57, 63, 64, 66, 68, 72, 76, 83–85] However, a study by Noguera-Julian et al. published in 2016[14] showed that Prevotella predominance was associated with MSM rather than HIV status, a finding that has been since corroborated by several other studies.[13, 18, 19] Importantly, the largest study to date found that MSM status had a much more profound effect on the gut microbiota than HIV status. Namely, in Principal Coordinates Analysis (PCoA), samples clustered overall by MSM, rather than HIV status.[13] Finally, in an individual level meta-analysis, compiling and re-analyzing data from 1032 individuals, we also identified distinct influences of MSM status on the overall structure of the microbiome: alpha-diversity was decreased in HIV+ as compared to HIV− patients, but only amongst women and MSW, not in the MSM population. Furthermore, alpha-diversity was higher in women than MSW, controlling for HIV status.[86] Similar to others, we also found that Prevotella predominance associates with MSM, rather than HIV status [data in preparation].

In a literature search up to June 2019, 16 studies on adult patients (excluding our meta-analysis[86]) compared the gut microbiome in HIV+ versus HIV− participants (utilizing stool samples, rectal swabs, rectal sponge collection or mucosal biopsies) while controlling for MSM status, whether through statistical methods or by including only MSM or only women (presented in Supplementary Tables 1–4).[12–14, 16, 18, 19, 55, 59, 65, 70–75, 77] An additional study controlled for MSM status but was primarily focused on CMV and EBV [87] and is not included in our analyses herein. Overall, these studies are highly variable. Most were based on samples from populations in the US and Europe[12–14, 16, 18, 19, 55, 59, 70, 71, 73, 75], with one study (a second analysis also utilized the same dataset) from Nigeria [65, 77], and two studies from Brazil and Malaysia [72, 74]. Study sizes ranged from 21–383 patients; only 7 studies examined populations ≥100 individuals ([12–14, 65, 70, 73, 75], see Supplementary Tables 1–4). Studies variably assessed HIV+ patients on and off of effective ART (See Supplementary Tables 1–4). Notably, these studies controlling for MSM status account for only 33% (N=16) of the total studies (N=48) comparing HIV+ to HIV− patients published to date[29, 50–54, 56–58, 60–64, 68, 76, 83–85, 88–100]. Several other studies variably controlled for MSM status but had no HIV− control group[15, 30, 32, 34, 38, 54, 66, 67, 69, 101–108]. Thus, the majority of published studies to date have been insufficiently controlled for the critical variable of MSM status, limiting interpretation of the data. Even as more studies on the impact of HIV status on the gut microbiome controlled for MSM status have recently appeared in the literature, results remain mixed with 6 cohorts suggesting little to no impact [16, 18, 19, 65, 71, 73, 77] and 8 cohorts suggesting a role for both HIV and MSM status as gut microbiome modifiers.[12–14, 55, 59, 70, 72, 74, 75] Concerns about residual confounding and/or, as yet, unrecognized or uncaptured variables impacting the gut microbiome remain. Importantly, precisely what factor (e.g., unprotected receptive anal or oral intercourse, substance use, rectal enemas or other considerations) related to MSM status impacts the gut microbiome and, therefore, needs to be controlled for is unknown. Further, few studies have carefully integrated data on medication use and/or diet exposures, known to be major gut microbiome modifiers. Finally, early data suggest that there may be features of the gut microbiota contributing to inflammation or immune activation in HIV+ individuals that are incompletely captured by 16S rRNA gene amplicon sequencing.[16, 19]

In summary, MSM status is a critical confounder which should be controlled for in all studies relating to the gut microbiota in HIV. Data on gut microbiome changes based on HIV status even in studies that controlled for MSM status are, at best, inconsistent. Notably, some studies do not identify any alpha-diversity[13, 65, 70–75] or taxa level changes[14, 18, 71, 73]. To our knowledge, large scale studies comparing ART-treated, virally-suppressed HIV+ individuals with immune activation or inflammation-related morbidities (e.g., myocardial infarction, stroke) to healthy, virally-suppressed HIV+ or HIV− individuals, while controlling for MSM status, have not been published.

The Gut Microbiota of Men Who Have Sex With Men: What Do We Know?

What distinguishes the structure of the gut microbiome in MSM?

Several studies from the HIV literature, and one cohort utilizing samples only from HIV− patients which compared men endorsing receptive anal intercourse (RAI)[17, 20] to those who never had RAI, have compared the gut microbiota in MSM to non-MSM (i.e., women or MSW) (See Table 1). Three studies conducted in the US or Europe found increased alpha-diversity in MSM as compared with non-MSM [13–15], and only one found decreased alpha-diversity.[17] The most consistent finding at the taxa level is that of increases in the abundance of Prevotella spp. in MSM versus non-MSM. In studies reporting species-level data, these increases are specifically in P. copri and P. stercorea.[13, 18, 19] Many studies have found, in parallel, decreased abundance in the genus Bacteroides [13, 14, 16, 18]. Other taxa such as Faecalibacterium prauznitzii [18, 19] or the genus Desulfovibrionaceae [13, 14] have also been found to differ by MSM status, but less consistently across studies. Very little data is available on cohorts recruited outside of the U.S. or Europe. A study conducted in MSM in Nigeria found the genus Prevotella was consistently represented across MSM samples, and in approximately one-third of MSM, Prevotella made up over 40% relative abundance. However no species-level data were available, and comparisons to non-MSM populations were not made.[65, 77]

Table 1:

Do the gut microbiota of MSM and non-MSM differ in studies of HIV-negative men or in studies controlled for HIV status?

	MSM Comparisons	Does the study control for HIV status?	Study population	Country	Alpha-diversity in MSM vs. non-MSM	Taxa enriched in MSM vs. non-MSM (p=phylum, o=order, f=family, g=genus, species)	Taxa decreased in MSM vs. non-MSM	Comments
US/Eurooe
Nogue ra-Julian, 2016#[14]	Compared MSM to non-MSM (MSW, women).	Stratified by HIV status for alpha-diversity only.	Spain: (100 MSM vs. 24 MSW, 31 women and 1 transgender), Sweden: (19 MSM vs. 32 MSW, 33 women)	Spain Sweden	Richer and more diverse in MSM by observed OTUs, Chao 1, ACE, Shannon even after FDR. (Shan non in Spain only)	Not controlling for HIV status, in Spain and Sweden: g_Bacteroidales_unclassified g_Phascolarctobacterium g_Desulfovibrio g_Paraprevotella g_RC9_gut_group g_Alloprevotella g_Dorea g_Succinivibrio g_Prevotetta	Not controlling for HIV status, in Spain and Sweden: g_Bacteroides g_Barnesiella g_Odoribacter	Taxa analyses reported in this table are from LEfSe analysis. Only genera that overlap between Spain and Sweden cohorts are reported in this table.
Kelley, 2017*[17]	Compared MSM engaging in CRAI to men who had never engaged in RAI.	All HIV-.	41 MSM engaging in CRAI, 21 control men who had never engaged in RAI. (Micro biota analysis conducted on 34 MSM engaging in CRAI and 17 controls)	USA	Decreased using Shannon metric in MSM engaging in CRAI vs. controls	MSM engaging in CRAI vs. controls (p<0.05): f_Prevotettaceae	MSM engaging in CRAI vs. controls (p<0.05): i_Bacteroidaceae	Analyzed only rectal mucosal swabs. Analysis was at the family level only. Reported p-values not adjusted for multiple comparisons.
Pescatore, 2018*[20]	Compared MSM engaging in CRAI to men who had never engaged in RAI.	All HIV-.	34 MSM engaging in CRAI vs. 20 control men who had never engaged in RAI	USA	Not report ed. (See Kelley, 2017[17])	MSM engaging in CRAI vs. controls (anal swabs) (p<0.05): g_Prevotella MSM engaging in CRAI vs controls (rectal mucosal swabs) (p<0.05) g_Prevotella g_Streptococcus	MSM engaging in CRAI vs. controls (anal swabs) (p<0.05): g_Bacteroides g_Clostridiales_unclassified g_Lachnospiraceae_unclassified MSM engaging in CRAI vs controls (rectal mucosal swabs) (p<0.05) g_Bacteroides and g_Lachnospiraceae_unclassified g_Facalibacterium and g_Parabacteroides	Although the Kelley and Pescatore cohorts overlapped, there were several differences. Pescatore analyzed both rectal mucosal and anal swabs and included a larger number of controls. Additionally, analysis went beyond the family to the genus level. Reported p-values not adjusted for multiple comparisons.
Neff, 2018[19]	Compared MSM to non-MSM (MSW, women)	Stratified by HIV status.	25 HIV+ (11 treated MSM, 7 untreated MSM, 7 treated females), 24 HIV− (13 MSM, 6 MSW, 5 females)	USA	Not reported.	[Genera] HIV− only, enriched in MSM vs. MSW (FDRO.1): o_Bacteroidales f_Desulfovibrionaceae f_Paraprevotellaceae g_Prevotella g_Succinivibrio g_Eubacterium g_Catenibacterium g_Mitsuokella [OTU] HIV-only, enriched in MSM vs. MSW (FDRO.1): f_S24–7 (Bacteroidetes) f_Paraprevotellaceae f_Desulfovibrionaceae f_Clostridiaceae f_Lachnospiraceae f_Coriobacteriaceae g_Phascolarctobacterium g_Diaiister g_Prevotella g_Desulfovibrio g_Catenibacterium g_Sutterella g_Megasphaera g_Succinivibrio s_Prevotella copri s_Eubacterium biforme s_Prevotella sterocorea	[Genera] HIV− only, depleted in MSM vs. MSW (FDR<0.1): f_Erysipelotrichaceae g_Bacteroides g_Akkermansia g_Coprococcus g_Streptococcus g_Blautia g_Anaerostipes g_Eggerthella [OTU] HIV-only, depleted in MSM vs. MSW (FDR<0.1): s_Ruminococcaeceae_unspecified s_Bacteroides uniformis s_Bacteroides_unspecified s_Anaerostipes_unspecified s_Erysipelotrichaceae_unspecific s_Clostridiaceae_unspecified s_Blautia_unspecified s_Clostridales_unspecific s_Faecalibacteriurn prausnitzii s_Sutterella_unspecified s_Coprococcus_unspecified s_Eggerthella lenta s_Dialister_unspecified s_Akkermansia_mucuniphila	Statistical analysis was conducted either on bacterial genera or OTUs to achieve maximum taxonomic resolution. Results from both analyses are shown on the left. Prevotella is most prevalent in MSM. Small sample size (6 HIV− MSW, 13HIV− MSM).
Armstrong, 2018[13]	Compared MSM to non-MSM (MSW, women).	Adjusted for HIV.	45 HIV+ untreat ed (40 MSM, 1 MSW, 4 women), 67 HIV+ treated (50 MSM, 2 MSW, 15 women), 105 HIV− (35 MSM, 29 MSW, 41 women)	USA	HIV− and HIV+ treated groups only: Increased in MSM vs. MSW by PD and observed OTUs.	All HIV+ and HIV− MSM vs. non-MSM (Bonferroni<0.05): o_Clostridiales f_S24–7 f_Coriobacteriaceae f_Clostridiaceae f_Ruminococcaceae g_Desulfovibrio g_Catenibacterium g_Phascolactoba cterium g_Prevotella g_Clostridium g_Slackia s_Prevotella copri s_Prevotella stercorea s_Eubacterium biforme HIV− MSM vs. non-MSM, (Bonferroni<0.05): g_Desulfovibrio s_Eubacterium biforme s_Prevotella copri	All HIV+ and HIV− MSM vs. non-MSM (Bonferroni<0.05): f_Erysipelotrichaceae f_Lachnospiraceae f_Coriobacteriaceae g_Clostridium g_Adlercreutzia g_Blautia g_Ruminococcus g_Phascolactob acterium s_Bacteroides ovatus s_Bacteroides uniformis HIV− MSM vs. non-MSM, (Bonferroni<0.05): f_Erysipelotrichaceae g_Clostridium g_Adlercreutzia s_Bacteroides uniformis	The most striking difference in MSM compared to non-MSM was the increase in relative abundance of the Prevotella genus. HIV− MSM had 3.9-fold higher relative abundance of Prevotella compared with HIV-non MSM. RAI frequency and diet did not seem to drive the Prevotella difference in MSM, and the women reporting RAI were not Prevotella-predominant. There were few Prevotella-predominant women. Additional taxa were significant if the FDR cutoff was used (not listed here).
Kehrmann, 2019[15]	Compared MSM to non MSM (MSW, women).	All HIV+.	42 MSM, 18 non-MSM (mostly men)	Germany	Elevated in MSM vs. non-MSM by Chao 1, Observed OTUs, Shannon, Simpson (0.048), PD, and Gini index.	MSM status was associated with being in the Prevotella-rich cluster. 14 MSM and 17 non-MSM were in Prevotella-poor cluster (also Bacteroides-rich). 28 MSM and 1 non-MSM were in Prevotella-rich cluster.
Hensley-McBain, 2019[16]	Compared MSM to non-MSM (MSW, women).	Adjusted for HIV.	40 HIV+ ART-suppressed (34 MSM, 6 women); 35 HIV− (11 MSM, 10 non-MSM, 2 male, 12 female)	USA	Not reported.	FDR<0.05: g_Streptococcus	FDR<0.05: f_Barnesiellaceae g_Bacteroides	Before adjusting for age, race, sex and HIV status, g_Prevotella and g_Streptococcus were enriched in MSM vs. non-MSM, which may be due to confounders as the groups were unbalanced with respect to MSM status in HIV+ vs. HIV-.
Guillén, 2019#[12]	Compared MSM to non-MSM (MSW, women).	Stratified by HIV status.	^129 HIV+; 27 HIV− (See Noguera-Julian “Spain” cohort for MSM, MSW and women)	Spain	Not reported.	MSM status correlated with higher-microbial gene count in multivariate regression models controlling for ethnicity and CD4 count	Not reported.	Although cohort overlapped with Noguera-Julian, only the Barcelona Spain cohort was used. Additionally, techniques were different; Noguera-Julian utilized 16S rRNA gene sequencing, Guillen utilized whole fecal metagenome shotgun sequencing and analysis focused on microbial gene richness.
Li, 2019[18]	Compared MSM to MSW.	Stratified by HIV status.	13 HIV+ untreated MSM, 18 HIV− MSW, 19 HIV− MSM	USA	Not reported.	HIV-only, MSM vs. MSW (FDR<0.05): o_Clostridiales f_Coriobacteriaceae f_Ruminococcaceae f_Rikenellaceae f_S24–7 g_Phascolarctobacterium g_Desulfovibrio g_Mogibacterium s_Desutfovibrio desuifricans s_Hoidemaneila biformis s_Howardeiia ureiiytica s_Eubacterium biforme s_Prevoteiia copri	HIV-only, MSM vs. MSW (FDR<0.05): g_Bacteroides g_Sutterlla g_Doprococcus g_Ciostridium g_Biautia s_Ciostridium ieptum s_Bacteroides uniformis s_Fiavonifractor piautii s_Bacteroides caccae s_Ciostridium ciostridioforme s_Faecaiibacterium prausnitzii s_Ciostridium spiroforme s_Ruminococcu s gnavus	Prevotella is noted in original manuscript to be higher in MSM than MSW.
Non-US/Eurooe
NONE

Abbreviations: ACE = abundance coverage estimator; ART = antiretroviral therapy; RAI=receptive anal intercourse; CRAI = condomless receptive anal intercourse; FDR = false discovery rate; HIV+ = HIV-infected; HIV− = HIV-uninfected; LEfSe = linear discriminate analysis effect size; MSM = men who have sex with men; MSW = men who have sex with women; OTU = operational taxonomic units; PD = phylogenetic diversity; RAI = receptive anal intercourse; VL = viral load; vs. = versus.

^#Guillen 2019 and Noguera-Julian 2016 utilize overlapping datasets.

^*Kelley 2017 and Pescatore 2018 utilize overlapping datasets.

Importantly, across cohorts, not all MSM are found to have Prevotella-predominant microbiota. To date, however, no large scale studies conducted in MSM have defined the gut microbiota of this population and what proportion of these men may be Prevotella predominant.[13, 15] Nor is it known how stable these microbiota differences are over time, though in a cross-sectional cohort of older (>age 55) MSM, Prevotella was still the most common genus in most individuals.[70] In a meta-analysis that did not control for MSM status, the prevalence of P. copri was recently shown to be 29.6% in a Westernized lifestyle, compared with 95.4% in non-Westernized lifestyle.[109] Lastly, existing studies have not consistently tested the same sample types (some utilize stool, rectal swabs, mucosal swabs, or colorectal biopsies), though one study suggests that there may be some differences in results depending on sample type.[20]

What factors may be leading to differences in the gut microbiota of at least a subset of MSM versus other populations?

This topic is, as yet, under-explored. Diet and medications (especially antibiotics) are well-known to impact the gut microbiome. As discussed below, these factors and others such as MSM-linked behavioral practices are under investigation to understand why differences in the gut microbiota in MSM are observed.

Is diet a key factor?

Prevotella-rich gut microbiota have been reported to distinguish rural African and Amazonian populations from U.S. and European populations.[110–112] This difference has been hypothesized to relate to the very different, high fiber diets consumed in agrarian as compared to industrialized societies. Another very small study found increased Prevotella (including P. copri and P. stercorea) in individuals from India as compared to individuals from China and related this to the higher fiber and lower meat content of the Indian diet.[113] One study in HIV+ individuals found that MSM whose microbiota were Prevotella-rich were more likely to be of high socio-economic status than those who were Bacteroides-rich, and postulated that diet may account for these results.[15] Armstrong et al. collected data on diet in a subset (N=98) of individuals. When the HIV− MSM (N=24) were compared to HIV− non-MSM (unclear if women or MSW, N=45), the study found that MSM ate more lean meat and fewer servings of fruit and grams of dietary fiber. When diet was added to linear models, this did not change the association of the top ten most significantly different OTUs with MSM status.[13] A second paper also found few statistically significant associations after adjustment for multiple comparisons of diet with Prevotella or other microbiome features, though those with Prevotella-predominant gut microbiota ate fewer servings of meat than those with Bacteroides-predominant gut microbiota.[14]

Is drug use a key factor?

Limited data have examined this issue. One study (Fulcher et al.)[69] conducted in HIV+ MSM found that the majority of patients had a high relative abundance of Prevotella in rectal swab samples. They found that marijuana use was associated with increased abundance of Fusobacterium and Anaerotruncus and a decrease in Dorea, while methamphetamine use was associated with increases in Fusobacterium, Granulicatella, and Anaerococcus as well as decreases in Parabacteroides, Collinsella, Paraprevotella and multiple Clostridiales organisms. When propensity score analysis was conducted to control for multiple confounders, marijuana use was associated with decreased Prevotella abundance.[69] To our knowledge, marijuana use has not been examined in any other studies published before June 2019 that controlled for MSM status.

Is RAI or are practices linked to RAI important?

An obvious factor which might distinguish some MSM from other populations is RAI (although a substantial proportion of women report RAI as well)[114, 115]. Inherent in this exposure could be exposure to semen (through unprotected anal intercourse), lubricants and/or rectal enema use prior to sex. One small study of MSM engaging in RAI as compared with men who had never engaged in RAI showed Prevotella predominance (based on rectal mucosal and anal swabs), but it was not clear whether the men who had never engaged in RAI were MSM or MSW.[17, 20] Fulcher et al. found that frequency of RAI and recent RAI were associated with changes in rectal bacterial genera (increased Anaerococcus and Peptostreptococcus and decreases in the family Clostridiaceae), but did not comment on (or identify changes in) Prevotella relative abundance.[69] However, this paper did not explicitly report whether the men had protected or unprotected RAI. Though numbers were small, Armstrong et al. did not find any significant association with fecal taxa (including Prevotella) or alpha-diversity between MSM who engaged in RAI (n=31) and those who did not (n=15), or with those who engaged in <1 time per week versus >1 time per week RAI. Furthermore, the 5 women in the cohort who reported RAI did not have Prevotella predominance.[13]

A very high percentage of MSM (>90%) who engage in RAI utilize lubricants during sex.[116] Many of these lubricants are hyperosmolar and have been linked to epithelial damage in the distal colon as well as to an increased risk of acquisition of rectal gonorrhea and chlamydia.[117, 118] A small study of MSM randomized to apply daily hyperosmolar rectal lubricant (N=20), take daily oral pre-exposure prophylaxis (PrEP) (N=19) or both (N=21) for 7 days found that lubricant users had a decrease in the genus Bacteroides (p=0.01) and a non-significant increase in the genus Prevotella (p=0.09) in rectal mucosal swabs.[119] There is no other available data on this topic to our knowledge.

Finally, MSM may engage in other sexual practices which could be postulated to lead to gut microbiome changes. A large proportion (over 60%) of MSM engage in rectal douching or enema use prior to RAI, and this has been linked to rectal gonorrhea and chlamydia acquisition.[120] Other practices include fisting, which has been linked to Hepatitis C virus acquisition [121], and anolingus (“rimming”) or using saliva as a lubricant during anal sex which has been linked to rectal gonorrhea and chlamydia.[122, 123] Since Prevotella is known to be common in oral flora, perhaps increased abundance of fecal Prevotella is linked to anolingus. However, P. copri and P. stercorea, the species identified to be increased in the gut microbiome of MSM, appear to be distinct from Prevotella spp. found in microbiome studies of saliva, though saliva studies have not yet been done in MSM.[124]

In summary, changes in several gut microbiome taxa, most consistently an increase in the Prevotella genus and to a lesser extent a decrease in the Bacteroides genus, have been associated with MSM status. However, what is driving increases in Prevotella and other changes in the gut microbiome of MSM, (and what proportion of MSM across risk groups are Prevotella-predominant), is still largely unknown. More research is needed to elucidate why this difference is seen and if it has biologic importance. Importantly, there is no data comparing the gut microbiota in MSM to non-MSM outside of the US and European regions, so whether these differences are consistent in MSM in other settings are unknown.

What are the possible implications of a distinct MSM-associated gut microbiota for the health of this population?

If accumulating data continue to bear out that there is a distinct gut microbiome (perhaps characterized by Prevotella predominance or other features) which defines at least a subset of MSM, then researchers, when designing studies examining the gut microbiome in the context of any given disease process, need to carefully consider whether MSM status could be a confounder and control for this factor as necessary. However, it is less clear what relevance, if any, these gut microbiota characteristics may have to the long-term health of MSM.

Several studies have linked Prevotella predominance, a reduced Lactobacillus:Prevotella ratio[16], “high risk” MSM status (i.e., condomless RAI with multiple partners)[125] or MSM-associated gut microbiota[18] to immune activation or markers of bacterial translocation. On the other hand, in a cohort of HIV+ MSM of similar ages, there was a borderline increased incidence of hyperlipidemia in patients whose microbiome was Prevotella-poor when compared to patients with Prevotella-dominant microbiota (p=0.08).[15] Large population-wide cohort studies have not generally examined MSM as a risk factor for cardiovascular disease or other inflammation-associated chronic diseases, though two smaller studies have pointed to a potential for enhanced risk for cardiovascular biomarkers and risk factors in MSM.[126, 127] Multiple confounders and the failure to capture MSM status in existing cohorts make large population-level studies on the association of MSM status and cardiovascular disease and other chronic disease risk challenging.

What is known about the contribution of Prevotella-predominant microbiota to health or disease? Can we make inferences from studies on Prevotella in other populations?

The genus Prevotella has been associated with both negative and positive health outcomes in the wider human population. Increases in the Prevotella genus have been associated with diets high in plant-based fiber and lower in animal-based fats, as well as with the production of anti-inflammatory fecal short chain fatty acids. As such, gut microbiota dominant in the genus Prevotella may be linked to potential health benefits.[111, 112, 128] In particular, P. copri has been associated in several studies with improved glucose metabolism or overall anti-inflammatory effects of diet.[129–132]

Conversely, however, the genus Prevotella has been associated with inflammation at various body sites[133]. Increases in Prevotella have been associated with obesity,[134] cardiovascular disease[135] and hypertension.[136] In mouse models, P. copri has been associated with elevated trimethylamine-N-oxide (TMAO) levels and transmissible atherosclerosis risk.[137] Other studies found P. copri associated with insulin resistance and glucose intolerance[138]. Similarly, increased relative fecal abundance of P. copri has been reported in predominantly female patients with new-onset rheumatoid arthritis[139] or at risk for rheumatoid arthritis[140]. Notably, transfer of this microbiota into mouse models resulted in enhanced development of arthritis.[141] In a mouse model of gut inflammation, animals colonized with P. copri displayed increased sensitivity to chemically-induced colitis.[139]

Differences in the species of Prevotella present may account for discrepancies in genus-level associations between gut microbiome studies. The predominant Prevotella spp. identified in the gut microbiota in available studies on MSM [13, 18, 19] (P. copri and P. stercorea) are distinct from the predominant Prevotella spp. found in other body sites [e.g., oral samples include predominantly P. intermedia and P. nigrescens[142, 143], and vaginal samples include (though are not restricted to): P. amnii, and P. bivia. [144, 145]]. Yet, even species-level data may be insufficient to understand associations with disease. As demonstrated above, P. copri has been associated with both negative and positive health outcomes. New data suggest that there are strain-level differences in P. copri which associate with diet and also differentiate Western from non-Western populations. Thus, strain-level differences may account for some of the discrepancies observed.[146] As an example, Armstrong et al. found evidence of strain-level variation in P. copri between MSM and MSW.[13] For this and likely other reasons, associations between disease conditions and the Prevotella genus identified in other populations may not be generalizable to MSM populations.[146] Together, these observations indicate that carefully controlled studies, likely going beyond species-level identifications, are needed to understand the health versus disease potential of Prevotella species in different populations.

What poor health outcomes are MSM known to be at risk for and could gut microbiota differences be playing a role?

Given the challenges with trying to draw conclusions from other populations, it may be instructive to identify the health outcomes that MSM are known to be at increased risk for relative to the rest of the population and then consider whether the gut microbiota including Prevotella could play a role in this susceptibility. Known MSM-linked health risks are primarily rectal sexually transmitted infections (including gonorrhea, chlamydia, and human papillomavirus (HPV)/HPV-related anal cancer) and HIV, as well as to some extent, drug-resistant STIs including gonorrhea.[147]

Rectal STIs:

Nowak et al. observed that enrichment in Fusobacterium in MSM rectal microbiota was positively associated with oncogenic anal HPV-16 and negatively associated with oncogenic anal HPV-35,[77] and Serrano-Villar et al. correlated a few (non-Prevotella) taxa with the detection of precancerous anal lesions in MSM.[66] Nowak et al. noted an increase of Prevotella in MSM positive for rectal gonorrhea,[65] and Ceccarani et al. noted some differences in a number of (non-Prevotella) taxa associated with rectal gonorrhea and chlamydia infection[148]. However, all of these data were cross-sectional and failed to control for exposures such as number of sexual partners. Research suggests that Prevotella spp. from the female genital tract may promote chlamydia persistence, but no data is available based on rectal Prevotella strains in MSM.[149]

There is interest, but little data, on potential interactions between MSM-specific gut microbiota and enhanced risk of drug-resistant STIs (e.g., drug-resistant gonorrhea).[147] Given escalating rates of STIs, particularly drug-resistant gonorrhea in MSM populations [150, 151], identifying modifiable microbiota differences in those who acquire these infections may be clinically important, or could lead, in the long term, to non-antibiotic therapies to treat or prevent rectal infection.

HIV:

Although PrEP has provided an important, highly efficacious method of preventing HIV acquisition in MSM [152], there is interest in whether changes in the gut microbiota could enhance risk of HIV transmission or acquisition in MSM. Li et al. found that HIV− MSM had higher measures of systemic immune activation compared with HIV− MSW, and the mice that were recipients of these HIV− MSM fecal specimens also displayed elevated gut immune activation.[18] Additionally, human lamina propria cells stimulated with fecal microbiota isolated from MSM stimulated higher frequencies of HIV-infected cells than fecal microbiota from MSW.[18] Other studies have also found evidence of enhanced CD4 T-cell HIV infection or induction of inflammation associated with experiments utilizing P. copri or P. stercorea. [88, 153]

CONCLUSIONS AND FUTURE DIRECTIONS

Existing studies -- often limited by sample size, heterogeneous study populations and/or sampling techniques -- have failed, thus far, to identify a consistent pattern of gut dysbiosis associated with HIV status. However, large, properly controlled studies comparing HIV+ individuals with clear markers of immune activation or inflammation-related long term morbidities (e.g., stroke, cardiovascular disease) to healthy HIV+ or HIV− individuals have not been completed. MSM status has emerged as a confounder in some HIV-related studies and appears to have a profound influence on the gut microbiota, but many basic questions remain unanswered.

While many questions remain, research gaps of particular interest include:

• What, if anything, differentiates the gut microbiome in MSM across time, geography and risk populations from otherwise similar MSW? If so, what causes these differences?

• In large, longitudinal (ideally collecting samples prior to an adverse outcome) studies properly controlled for MSM status, are there identifiable differences in the gut microbiota in patients with long-term poor outcomes (e.g., CVD, renal dysfunction among others) stratified by HIV status? If so, does this correlate with markers of microbial translocation and/or immune activation and are there relevant interactions with MSM status? Or conversely, does MSM status confer increased risk of long-term poor outcomes when controlled for HIV status?

• Can the gut microbiota be related to increased risk of rectal STI or HIV acquisition in longitudinal (not cross-sectional) studies that control for exposures? Can gut microbiota associations be translated into therapeutics to decrease risk of these infections?

Large scale studies will be required to address these questions and others and ultimately to understand whether the gut microbiota may be a therapeutic target to improve the health of MSM and HIV+ patients.

Biographies

Wei Li Koay

Cynthia Sears

Susan Tuddenham