HIV-associated changes in the enteric microbial community: potential role in loss of homeostasis and development of systemic inflammation

Supplemental Digital Content is available in the text

Abstract

Purpose of review

Despite HIV therapy advances, average life expectancy in HIV-infected individuals on effective treatment is significantly decreased relative to uninfected persons, largely because of increased incidence of inflammation-related diseases, such as cardiovascular disease and renal dysfunction. The enteric microbial community could potentially cause this inflammation, as HIV-driven destruction of gastrointestinal CD4⁺ T cells may disturb the microbiota–mucosal immune system balance, disrupting the stable gut microbiome and leading to further deleterious host outcomes.

Recent findings

Varied enteric microbiome changes have been reported during HIV infection, but unifying patterns have emerged. Community diversity is decreased, similar to pathologies such as inflammatory bowel disease, obesity, and Clostridium difficile infection. Many taxa frequently enriched in HIV-infected individuals, such as Enterobacteriaceae and Erysipelotrichaceae, have pathogenic potential, whereas depleted taxa, such as Bacteroidaceae and Ruminococcaceae, are more linked with anti-inflammatory properties and maintenance of gut homeostasis. The gut viral community in HIV has been found to contain a greater abundance of pathogenesis-associated Adenoviridae and Anelloviridae. These bacterial and viral changes correlate with increased systemic inflammatory markers, such as serum sCD14, sCD163, and IL-6.

Summary

Enteric microbial community changes may contribute to chronic HIV pathogenesis, but more investigation is necessary, especially in the developing world population with the greatest HIV burden (Video, Supplemental Digital Content 1, which includes the authors’ summary of the importance of the work).

INTRODUCTION

The enteric ‘microbiome’ consists of a diverse collection of trillions of Bacteria, Archaea, Eukarya, and viruses [1–5], with a large aggregate genome, referred to as the ‘metagenome’, that contributes to normal immune development [6] and a number of pathological processes [7–9]. The host immune system acts as an essential curator for this luminal enteric microbial community, serving to shape and control the structure and function of this diverse collection of organisms [10–12]. HIV infection leads to the widespread destruction of host immune function [13,14], including the rapid and profound depletion of CD4⁺ T cells within gut-associated lymphoid tissue [15,16]. As could be predicted from the loss of mucosal immune cells, a wide range of changes in the enteric microbial community have been reported during HIV infection [17,18^▪▪,19,20^▪,21^▪▪,22^▪,23^▪▪,24^▪–27^▪,28^▪▪–32^▪▪,33^▪] (Fig. 1 and Table S1). In light of the role of the enteric microbiome in many inflammation-associated pathologies such as diabetes, obesity, and inflammatory bowel disease (IBD) [7–9], it has been proposed that HIV-associated microbiome shifts could contribute to inflammation-related noncommunicable diseases (NCD) that are responsible for a large share of the increased mortality observed during chronic HIV infection [34–39]. Investigation of microbial changes associated with HIV infection has the potential to aid in the development of therapeutic interventions that could improve many of the pathologic consequences of chronic HIV infection. 

FIGURE 1.

Alterations in the gut microbiota reported in HIV microbiome studies. Black and gray boxes, respectively, indicate reported increases and decreases in relative abundance. Bracketed letter indicates taxonomic level of classification (e.g., [p] indicates phylum), parenthetical taxa indicate subtaxa that commonly drive significance of parent taxa. ^aDysbiosis observed in HIV-positive subjects with less than 200 CD4⁺ T cells/μL. Black dashed box outlines the only study to date that investigated the relationship between gut microbiota and HIV in a sub-Saharan African cohort. ^bFecal microbiome data only obtained for HIV seronegative unexposed and exposed infants. ^cProteobacteria class indicated by corresponding Greek letter (e.g., β for Betaproteobacteria). Enterobacteriaceae are Gammaproteobacteria [c] within Proteobacteria [p]. ART, antiretroviral therapy.

Box 1.

no caption available

CHRONIC SYSTEMIC IMMUNE ACTIVATION IN HIV INFECTION

Despite advancements in HIV antiretroviral therapy (ART), average life expectancy in HIV-infected individuals on effective treatment is 14% less than that of uninfected persons [35]. The ART Cohort Collaboration Study found that deaths were largely because of inflammation-related clinical diseases, such as stroke, long-bone fractures, cardiovascular disease (CVD), and renal dysfunction [34,35,37,38]. CVD, which constitutes a large proportion of HIV-associated NCD [40–42], is increased in HIV infection [43–45] and associated with systemic immune activation as measured by markers such as serum IL-6, sCD163, and C-reactive protein [46–51]. In HIV-uninfected individuals, microbiota-induced inflammation has been shown to drive CVD pathogenesis [52–55], suggesting that this disorder could occur in HIV as well.

Further supporting the connection between systemic inflammation and chronic pathogenesis, individuals with the highest degree of persistent elevated immune activation while on suppressive ART experience higher overall mortality, even with CD4⁺ T-cell reconstitution more than 500 cells/μl [56,57]. Although systemic immune activation declines after initiation of ART, it remains persistently elevated in the majority of study participants even after years of therapy [58–60] and has also been observed in individuals with undetectable viral loads [61]. In a cohort of HIV-infected Ugandans, a 1.6-fold increased hazard of death was associated with each 10% increase in CD8⁺ T-cell activation following initial viral suppression with ART [62].

POTENTIAL ROLE OF HIV-ASSOCIATED GUT MICROBIAL CHANGES IN HIV DISEASE PROGRESSION

HIV-associated changes to the enteric microbiome may lead to systemic inflammation by disrupting the balance of metabolic functions performed by the microbiota, such as short-chain fatty acid or bile acid metabolism [63], or causing increased translocation of bacterial products into the systemic circulation [64]. Elevated plasma kynurenine, a tryptophan metabolite, has been found to be associated with CD8⁺ T-cell activation [65] and mortality [66] in HIV-infected individuals and HIV infection was associated with the presence of a gut microbial community with both the genetic capacity to metabolize tryptophan into kynurenine and demonstrable kynurenine production in vitro[21^▪▪]. In a similar paradigm, independent of HIV, the enteric microbiome can drive CVD pathogenesis by transforming dietary constituents such as phosphatidylcholine and bile acids into reactive intermediates such as trimethylamine N-oxide that can lead to macrophage and platelet activation, thrombosis, and arterial plaque formation [67–71].

Early in HIV infection the intestinal mucosa is a critical reservoir of viral replication and substantial CD4⁺ T-cell depletion (80–90%) [72–80]. This depletion is attributed to the widespread surface expression of CCR5 on gastrointestinal T cells, which serves as the coreceptor for HIV entry early on in infection [13,72]. Because certain lymphocyte subsets are important to the maintenance of enteric epithelial integrity (e.g., IL-22-producing ILC3 and Th17 cells [12,81]), it is believed that such a loss of immune surveillance may result in impaired intestinal epithelial barrier function, increased gut permeability, and the translocation of gut bacterial products into systemic circulation [16,64]. These products could trigger persistent systemic immune activation and drive turnover of CD4⁺ and CD8⁺ T cells leading to clonal exhaustion and progressive impairment of T-cell function [82]. HIV-associated chronic systemic immune activation, secondary to a loss of gut epithelial homeostasis, may then lead to disease progression in both treated and untreated HIV infection. This model is supported by data showing that circulating levels of lipopolysaccharide (LPS), a bacterial product known to elicit an innate immune response by binding host Toll-like receptors [83], increase significantly as HIV infection progresses [84], though initiation of ART lowers circulating LPS levels [64]. Individuals with both treated and untreated HIV infection, however, exhibit elevated serum levels of bacterial LPS and systemic inflammation relative to uninfected controls [64,85,86].

BACTERIAL MICROBIOME DIFFERENCES IN HIV INFECTION

Changes in the composition of the enteric microbiome that cause it to deviate from what is considered a healthy baseline state are often collectively referred to as ‘dysbiosis’. Dysbiotic states are thought to underlie many of the pathologies linked with the microbiota [7–9]. Comparative studies between HIV-infected and uninfected individuals have differed in methods and patient populations studied (Fig. 2 and Table S1) and found a wide spectrum of differences associated with HIV infection (Figs. 1 and S1). However, several overarching themes have emerged. HIV-infected individuals often have reduced diversity of their enteric microbiome, which is also observed in a number of pathologies such as IBD [87–89], type 1 diabetes mellitus [90], obesity [91], and Clostridium difficile infection [92,93]. HIV infection also frequently correlates with increased tissue and circulating measures of inflammation (e.g., sCD14, IL-6, CD38⁺HLA-DR⁺CD8⁺ T cells) [64,85,86], as well as increases in traditionally pathogenic bacteria such as Enterobacteriaceae [20^▪,21^▪▪,24^▪,29^▪▪,31^▪▪,32^▪▪,33^▪] and decreases in commensals such as Lactobacillaceae, Lachnospiraceae, and Ruminococcaceae [19,20^▪,21^▪▪,22^▪,23^▪▪,25^▪,26^▪,28^▪▪–30^▪▪,32^▪▪,94,95].

FIGURE 2.

Study design and patient metadata in HIV microbiome studies. ^∗CD4 cell counts (cells/μL) displayed as either median [range (not notated in table)] median (IQR), or average ± SD. ^†Sex composition of clinical study participant groups not specified. ^aThe same patient cohort was used for both Lozupone et al. 2013 and 2014. Displayed is patient information from Lozupone et al. 2014, which included 17 additional subjects. Proportion of males versus females in each clinical group is only described in Lozupone et al. 2013 (male/female: 8/5 seronegative; 11/0 HIV+ untreated; 7/1 HIV+ treated). ^bThe black dashed box outlines the only study to date that investigated the relationship between gut microbiota and HIV in a sub-Saharan African cohort. ^cFecal microbiome data only obtained for HIV seronegative unexposed and exposed infants. ^dThe HIV-infected group has a total of 31 individuals, including 22 males and 9 females, specific clinical group not specified. ^eThirty study participants (22 males, 8 females) originally enrolled but only 24 study participants completed the study, specific clinical group not specified. ^fThe HIV-infected groups of the Barcelona and Stockholm cohorts, respectively, included 101 males and 28 females, and 46 males and 31 females, specific clinical group not specified. ^gImmunological responders (IR) and nonresponders (INR) respectively defined as having at least 350 and less than 350 CD4⁺ T-cell counts/μl after more than 2 years of ART. ART, antiretroviral therapy.

Taxa from four bacterial phyla – Proteobacteria, Fusobacteria, Bacteroidetes, and Firmicutes – have been reported in multiple studies to differ between HIV-infected and HIV-uninfected individuals. Proteobacteria were more abundant in HIV-infected individuals in 10 of the 16 published studies whereas others reported no change [17,18^▪▪,19,20^▪,21^▪▪,22^▪,23^▪▪,24^▪,27^▪,30^▪▪,32^▪▪,33^▪]. The phylum Proteobacteria includes numerous pathogens, such as Shigella, Salmonella, and Helicobacter. Many specific Proteobacteria genera have been shown to be enriched in HIV-infected individuals and are capable of potentially pathogenic activities in this context [96]: Pseudomonas[21^▪▪,95] is an opportunistic pathogen that is capable of impairing host mucus production [97,98]; Desulfovirbrio[17] can produce hydrogen sulfide compounds and inflame the host epithelium [99]; Acinetobacter[23^▪▪] can produce LPS [100] and in vitro is able to induce IL-8 production and neutrophil recruitment that could lead to inflammatory tissue damage [101,102]; and Campylobacter[20^▪] produces multiple toxins [103] that can induce mucosal inflammation [104]. The Proteobacterial family Enterobacteriaceae is associated with inflammation [105,106] and was enriched in HIV-infected individuals in seven of the 16 studies surveyed [20^▪,21^▪▪,24^▪,29^▪▪,31^▪▪,32^▪▪,33^▪] (most frequently driven by its constituent genus Escherichia, which is capable of pathogenic bile acid transformations and degradation of host mucus [98,107]). The phylum Fusobacteria, which is of particular interest because of its previous associations with intestinal inflammation and colorectal cancer [108–110], was enriched (most frequently driven by its constituent genus Fusobacterium) in HIV-infected individuals in four of the 16 previously published studies [19,20^▪,22^▪,32^▪▪].

Taxa in the phylum Bacteroidetes, including the families Prevotellaceae, Porphyromonadaceae, Bacteroidaceae, and Rikenellaceae, exhibited a more heterogeneous pattern of changes in HIV-infected individuals. The family Prevotellaceae (most frequently driven by its constituent genus Prevotella) was enriched in HIV-infected individuals in seven of 16 studies [17,18^▪▪,20^▪,23^▪▪,26^▪,27^▪,28^▪▪,33^▪] and depleted in one study [31^▪▪]. Taxa within the family Prevotellaceae have been associated with inflammation (particularly in the context of autoinflammatory disease [111,112]) and activation of gut dendritic cells [28^▪▪], but a greater abundance of Prevotellaceae is also characteristic of the baseline enteric microbial community of healthy individuals in developing world countries such as Burkina Faso, Venezuela, Malawi, or Papua New Guinea [17,113–117]. Taxa from the family Porphyromonadaceae (most frequently driven by the genera Barnesiella or Odoribacter) were generally depleted in HIV-infected individuals, exhibiting a decrease in nine of 16 studies [17,18^▪▪,20^▪,21^▪▪,22^▪,23^▪▪,30^▪▪–32^▪▪,33^▪] and an increase in two studies [19,24^▪]. Independent of HIV, Porphyromonadaceae exhibit a diverse and complex array of functions, with both positive [110,118,119] and negative [119] associations with colorectal cancer, and negative associations with C. difficile[120–122], Salmonella[123], vancomycin-resistant Enterococcus[124], and Citrobacter rodentium[125,126] infection that imply a putative protective role. The family Bacteroidaceae (mostly driven by the abundance of the genus Bacteroides) is overall depleted in HIV infection, showing a reduced HIV-associated abundance in 10 of 16 studies [17,18^▪▪,21^▪▪,23^▪▪,26^▪,28^▪▪–31^▪▪,33^▪]. This family is generally considered to play an anti-inflammatory role [127–131], with the species Bacteroides fragilis promoting regulatory T-cell differentiation and IL-10 production via secreted bacterial products [127–129]. The family Rikenellaceae (mostly driven by the abundance of the genus Alistipes) is also depleted in HIV infection, with decreased abundance in seven studies [17,18^▪▪,19,20^▪,21^▪▪,23^▪▪,24^▪,28^▪▪] and this bile-tolerant family [132] displays protective properties against C. difficile infection [133] and a negative association with obesity [134] as well as positive associations with both type 1 [135] and type 2 [136] diabetes mellitus.

Similar to the phylum Bacteroidetes, bacterial families within the phylum Firmicutes were in general reduced in abundance in HIV-infected individuals, though this pattern did not hold true for every family within this phylum. The phylum overall behaved in this manner, with decreased abundance in five of 16 studies [19,20^▪,22^▪,23^▪▪,27^▪] and increased abundance in two studies [30^▪▪,33^▪]. The Firmicutes phylum is quite diverse, but broadly can be characterized as associated with developed world individuals [113] as well as obesity and increased energy harvest from diet [137,138]. Within the Firmicutes, the class Clostridia, which was overall depleted in HIV infection with decreased abundance in eight of 16 studies [19,20^▪,21^▪▪,22^▪,23^▪▪,30^▪▪,32^▪▪,33^▪], is characterized by taxa that often function in anti-inflammatory roles by producing butyrate and other short-chain fatty acids (SCFA) [107,139] and shifting T-cell differentiation toward regulatory T cells [140–143]. Within the class Clostridia, the family Lachnospiraceae, which was decreased in abundance in HIV-infected individuals in six of 16 studies [19,20^▪,21^▪▪,23^▪▪,25^▪,30^▪▪], includes members that are commonly found to be uniquely effective metabolizers of complex polysaccharides [144,145] and characterized by the production of SCFA such as butyrate [146] and acetate [147] that are thought to be anti-inflammatory. Also within the class Clostridia, the family Peptostreptococcaceae varied in its HIV-associated shifts, with two of 16 studies [20^▪,32^▪▪] showing a relative decrease in abundance in HIV-infected individuals and one study [19] showing a relative increase. Peptostreptococcaceae have been found to function in mostly a proinflammatory role, with positive associations with C. difficile infection [121,148], viral diarrhea [149], intestinal inflammation [150], and the mucosal [151] and fecal [152] communities of individuals with colorectal cancer. In contrast to the family Peptostreptococcaceae, the Clostridia family Ruminococcaceae was in general decreased in HIV infection, with 10 of 16 studies [19,20^▪,21^▪▪,22^▪,23^▪▪,26^▪,28^▪▪–30^▪▪,32^▪▪] reporting decreased abundance in HIV-infected individuals and only one study [17,18^▪▪] reporting increased abundance. Ruminococcaceae have been associated with both protective and disruptive roles within the gut microbial community, such as the production of anti-inflammatory SCFA [144] or the degradation of host mucus and potential proinflammatory role in IBD [153], and functional effects within this family have been found to be highly species dependent [132,154–156]. The bacterial family Erysipelotrichaceae, which is contained within the separate class Erysipelotrichia, was on the whole found to be increased in association with HIV infection with a greater abundance demonstrated in six of 16 studies [17,18^▪▪,20^▪,21^▪▪,24^▪,26^▪,33^▪] and decreased abundance in only one study [31^▪▪]. Erysipelotrichaceae are described as adhesive and potentially pathogenic [157], and have been found to be positively associated with obesity [158,159] and luminal microbial communities in colorectal cancer [151,160]. Interestingly, this family is also found to be enriched in the enteric communities of the Hadza hunter-gatherers of Tanzania [157].

MICROBIOME DIFFERENCES AND SAMPLING SITE

Some of the variation in the findings among these studies may be attributed to the differences in body site sampled and sampling methodology. The predominant collection method was stool sampling [17,18^▪▪,20^▪,23^▪▪,24^▪–26^▪,28^▪▪–32^▪▪,33^▪,161^▪–164^▪] which is most representative of the luminal microbial community, though anal swabs and washings [19,22^▪,165] or mucosal biopsies [20^▪,21^▪▪,23^▪▪,27^▪,28^▪▪] were also used. Independent of HIV, mucosal and luminal microbial communities have been shown to differ to varying degrees [166–171], though in some contexts, the mucosal and luminal communities correlate strongly and are representative of one another [170]. With regard to functional differences, facultative anaerobes have been found to be more abundant in mucosal-associated environments, whereas obligate anaerobes are more prevalent in the gut lumen [170]. Studies of the HIV-associated microbiota that used both mucosal and luminal-targeted sampling techniques [20^▪,23^▪▪,28^▪▪] found a variety of differences between these techniques. In the most concordant finding, the two sampling sites produced similar conclusions, with most taxa showing the same patterns of HIV-associated enrichment or depletion at both sampling sites (Fig. 1 and Table S1) [20^▪]. However, the HIV-associated reduction in diversity was more pronounced in mucosa than stool samples. Other studies found greater variation in the HIV-associated differences by body site. For instance, Dillon et al.[28^▪▪] detected an HIV-associated increase in abundance of Prevotellaceae at both sampling sites, but only observed HIV-associated decreases in taxa in the phyla Bacteroidetes and Firmicutes and increases in Proteobacteria in the mucosal samples. The study with the greatest discordance between mucosal and luminal findings observed many more HIV-associated differences in mucosal-associated communities than luminal communities [23^▪▪]. In addition, mucosal community composition changes were more closely associated with mucosal cellular immune activation. Mucosal findings may be more sensitive or representative of the causative community, as these microorganisms are in the closest contact with epithelial cells and immune cells. These advantages, however, must be weighed against the relative difficulty of obtaining mucosal samples.

GEOGRAPHICAL CONTEXT FOR BACTERIAL MICROBIOTA DIFFERENCES

Most current work investigating HIV and the enteric microbiome has focused on populations in the developed world (Fig. 3), as opposed to the developing world where HIV burden is greatest [172]. As the burden of NCD in chronic HIV is growing rapidly in the HIV-infected population in sub-Saharan Africa [173–176], this region could potentially benefit from the deployment of microbiota-directed therapeutics. However, geographic differences in the gut microbiome may make it difficult to translate data derived from developed world subjects to target populations in the developing world.

FIGURE 3.

People living with HIV and locations of HIV-microbiome studies. Locations of HIV-microbiome cohorts overlaid on world map of HIV prevalence. The vast majority of HIV-microbiome studies have been conducted in developed world settings. Map source: Adults and children estimated to be living with HIV 2015, WHO/UNAIDS/UNICEF.

The microbiota differences observed between HIV-infected and uninfected individuals as a whole [17,18^▪▪] mirror some differences seen at baseline between populations in the developing and developed world [113,115,117]. It is currently thought that the differences between developed and developing world microbiota are primarily a result of the corresponding dietary differences, with greater consumption of fat and simple carbohydrates in the developed world and greater consumption of more complex carbohydrates in the developing world [114,132,177]. The observations that a Prevotellaceae-rich community is frequently observed in healthy individuals in developing nations such as Burkina Faso, Venezuela, Malawi, or Papua New Guinea [17,113–117] argue against the simple conclusion that defined HIV-associated taxa changes in the gut microbiota (e.g., an increase in Prevotellaceae and a decrease in Firmicutes) alone are responsible for chronic inflammation and disorder in HIV-infected individuals. Rather, these potentially conflicting observations suggest a more complex relationship wherein a mismatch between the extant taxa and their host context causes inflammation [18^▪▪]. One possible example of this mismatch might be a microbial community that is inappropriate for the host diet and as a result produces inflammatory metabolites from these dietary constituents. This ‘context-dependent’ model of microbial-driven systemic inflammation would further heighten the need to study HIV-associated microbial differences in appropriate populations, as therapeutic interventions constructed from data originating from one population may be poorly suited for application in a population with a different microbial and environmental milieu.

VIROME DIFFERENCES IN HIV INFECTION

The diverse human enteric virome includes a wide array of eukaryotic viruses, bacteriophages, and endogenous retroviruses [178–181] and is largely unexplored, with some estimates that only 1% of the virome has been sequenced [182]. To date, only a handful of studies have examined alterations in the virome during lentiviral infection [29^▪▪,183–186]. New sequencing technologies offer a unique opportunity for unbiased examination of the virome including discovery of potentially novel unculturable viruses [187]. Enteric eukaryotic viruses can drive host pathology that directly causes gastroenteritis, enteritis, or colitis. Bacteriophages, which are the most abundant enteric viruses, can initiate changes in the bacterial community that influence gut health and may also be able to directly interact with the host immune system [178–190]. In IBD, diversity of the bacteriophage virome increases whereas diversity of the bacterial microbiome decreases, suggesting a potential competitive relationship between bacteria and bacteriophages during enteric inflammation [191]. In light of the emerging recognition of the potential role of the enteric virome in gastrointestinal pathology, there is now a growing focus on characterizing the gut virome associated with HIV infection.

Nonhuman primate simian immunodeficiency virus models of HIV infection have shown expansion of the enteric virome, with significant increases in Picornaviridae, Adenoviridae, and Parvoviridae abundance in pathologic simian immunodeficiency virus infection of rhesus macaques [183,184]. In this model, Adenoviridae were associated with development of severe gastrointestinal disease [183,184] and enteric viruses spread to other tissue compartments, with Parvoviridae detected in the serum and Adenoviridae found in the enteric mucosa [184].

HIV-associated differences in the virome have also been observed in HIV-infected individuals. In a Ugandan cohort, CD4⁺ T-cell counts less than 200 cells/μl were associated with a significant increase in enteric Adenoviridae and Anelloviridae[29^▪▪]. Anelloviridae are small, nonenveloped viruses with circular negative-sense ssDNA genomes often found in human serum [192]. Anelloviridae have not yet been identified as causative disease agents, but have been shown to be increased in the serum of immunosuppressed solid organ transplant patients [193] and (not using modern sequencing methods) HIV-infected study participants [194,195]. Similar alterations to the plasma virome were observed in HIV-infected individuals in the United States and Uganda, with an increase in Anelloviridae[185]. In the United States but not the Ugandan cohort, there was also a significant increase in total plasma viral sequences, mainly attributed to bacteriophages, in HIV-infected individuals with low (< 20 cells/μl) CD4⁺ T-cell counts as compared with individuals with high (> 700 cells/μl) CD4⁺ T-cell counts [185].

CONCLUSION

The role of enteric microbial changes in HIV disease progression has been the focus of increasing investigation. Models regarding the connections between gut microbiome changes and chronic HIV pathogenesis hypothesize a role for gut epithelial damage and systemic immune activation as an intermediate mechanism. Investigations into HIV-associated differences in the gut microbial community have found varied changes, but a few overall consistent patterns have emerged. In general, HIV infection is associated with decreases in many bacterial families within the phyla Bacteroidetes and Firmicutes as well as increases in the family Prevotellaceae and families within the phyla Proteobacteria and Fusobacteria. This community shift implies an overall pathogenic or proinflammatory outcome based on the functions of the differentially abundant microorganisms, but this causative relationship has not been conclusively shown and bacterial behavior can vary widely based on context. The viral component of the gut microbial community has received relatively little attention, but findings have generally shown an increase in the family Anelloviridae, which is frequently found in immunocompromised hosts, and in potentially pathogenic viruses such as the family Adenoviridae.

The heterogeneity of the conclusions drawn to date within this field may be in part because of the variation in study populations and methods employed (Fig. 2 and Table S1). Greater coordination of methodologies would allow more robust analysis of multiple studies and potentially reconcile conflicting conclusions. There is also a need to more finely examine the taxa differentially abundant in HIV infection, as meaningful functional variation can often occur at the species or strain level rather than the family or genera level [196–198]. 16S rRNA gene sequencing cannot provide this taxonomic resolution, creating a role for techniques such as shotgun metagenomic sequencing, which could detect strain or gene content variation. In light of the current findings of HIV-associated Prevotellaceae enrichment and the great abundance of this taxa in developing world populations that constitute the majority of HIV-infected individuals, there is a dire need for further examination of HIV-associated gut microbial differences in developing world populations. It is possible that therapeutic strategies that consider Prevotellaceae enrichment a pathogenic state would be ill suited for HIV-infected individuals in the developing world. In addition, microbiota-targeted probiotic interventions have produced mixed results, with some therapeutics reducing inflammatory markers and others lacking efficacy [162^▪,199–208]. Further understanding of the enteric microbial changes associated with HIV infection, especially among developing world populations that bear the greatest burden of HIV infection, is therefore necessary to design therapeutic strategies that could alleviate the sequelae of systemic inflammation and NCD in chronic HIV infection.

Acknowledgements

The authors would like to acknowledge Dr Herbert W. Virgin, Dr Scott Handley, and Dr Guoyan Zhao for informative discussions regarding the content of this review.

Financial support and sponsorship

Funding received from NIH (NIDDK and NIAID), and the Burroughs Wellcome Fund.

Conflicts of interest

There are no conflicts of interest.

REFERENCES AND RECOMMENDED READING

Papers of particular interest, published within the annual period of review, have been highlighted as:

• ▪ of special interest

• ▪▪ of outstanding interest