What is the collective effect of aging and HIV on the gut microbiome?

Abstract

Purpose of review:

Aging and HIV share features of intestinal damage and alterations in the communities of enteric bacteria, termed dysbiosis. The purpose of this review is to highlight the various features of the gut microbiome in 1) aging and in 2) people with HIV (PWH) and to discuss how aging and HIV converge to impact the gut microbiome. The term microbiome reflects the combined genetic material of micro-organisms present including bacteria, viruses, bacteriophages and fungi. To date, the majority of studies investigating the impact of aging and HIV on the gut microbiome have focused on bacteria and therefore, for the purposes of this review, the term “microbiome” is used to reflect enteric bacterial communities.

Recent Findings:

Aging is associated with alterations in the gut bacterial microbiome. Although changes vary by the age of the population, lifestyle (diet, physical activity) and geographic location, the age-associated dysbiosis is typically characterized by an increase in facultative anaerobes with inflammatory properties and a decrease in obligate anaerobes that play critical roles in maintaining intestinal homeostasis and in regulating host immunity. PWH also have dysbiotic gut microbiomes, many features of which reflect those observed in elderly persons. In one study, the age effect on the gut microbiome differed based on HIV serostatus in older adults.

Summary:

HIV and age may interact to shape the gut microbiome. Future studies should investigate relationships between the gut microbiome and age-associated co-morbidities in older PWH populations. Identifying these links will provide new avenues for treatments and interventions to improve the healthspan and lifespan of older PWH.\

Introduction.

The human gut is home to hundreds of different species of bacteria which have co-evolved with the human body and have symbiotic (mutually beneficial) or commensal (of benefit to one without harming the other) relationships with their host (1). A large proportion of enteric bacteria are obligate anaerobes primarily belonging to the Bacteroidetes and Firmicutes phyla. Species within these phyla are known to serve critical roles in host physiology either by interacting directly with host immune cells or providing essential metabolites through digestion of diet-related complex carbohydrates, proteins and fats (2–4). For example, species of Bacteroides directly induce regulatory T cells (5) while several Firmicutes species produce the metabolite butyrate, a short chain fatty acid (SCFA) produced following fermentation of dietary fiber (6). Butyrate not only regulates host immune cell function to dampen inflammation, but also serves as a critical energy source for epithelial cells, regulates the epithelial immune response (e.g., production of anti-microbial peptides) and maintains a low oxygen environment favorable to obligate anaerobes (7).

Given the mutualistic and complex nature of the microbiome-host interactions, it is not surprising that a loss of control of this delicate balance leads to changes in the microbial community, termed dysbiosis. For example, a number of elegant mouse studies have implicated the loss of butyrate-producing bacteria (BPB) and therefore lower butyrate levels in increasing levels of luminal oxygen, thus providing an ideal environment for the expansion of less common species of facultative anaerobes (i.e., anaerobes that can survive in oxygen) (8–10). Despite their commensal nature, many of these bacteria have the potential to induce inflammatory responses and exert pathogenic effects on host immunity and are termed pathobionts (11, 12). Host mucosal inflammatory immune responses may also lead to increases in luminal oxygen, thus further exacerbating dysbiosis (13, 14).

Both HIV and aging are associated with a chronic low-level inflammatory state which has been linked to increased risk of co-morbidities (recently reviewed in (15, 16)). Although there are multiple drivers of chronic inflammation, a breakdown in intestinal homeostasis and the associated ‘leaky gut’ and increased microbial translocation have been linked to systemic inflammation in both HIV (17–20) and aging (21, 22). Given the importance of the enteric bacterial community in maintaining a healthy homeostatic gut microenvironment, understanding the composition and function of the enteric microbiomes associated with physiologic aging and HIV infection have therefore become areas of great scientific interest.

Enteric microbial communities are altered in elderly persons.

The first extensive study to investigate the impact of aging on the gut microbiome utilized a panel of group- and species-specific bacterial 16S ribosomal RNA (rRNA)-targeted oligonucleotides to probe the bacterial taxonomic composition of fecal samples from younger (25–49 years) and older (61–100 years) persons at four European locations (23). Older people, irrespective of geographical location, had higher proportions of the genus Enterobacteria, a facultative anaerobe, belong to the Enterobacteriaceae family. Conversely, aging was associated with lower amounts of the butyrate-producing species Faecalibacterium prausnitzii, but only in elderly persons from Italy and Sweden and not France or Germany. Thus, while differences in aging-associated microbial communities were identified, they differed between study location, highlighting that aging-associated microbial profiles are likely influenced by factors associated with geographical location such as genetic, dietary and environmental conditions (24). These initial observations are now supported by a number of subsequent studies that primarily used high-throughput bacterial 16S rRNA gene sequencing approaches. Despite varying by location (primarily Europe and Asia) and differences in participant age ranges and health status (e.g., institutionalized, adjustments for polypharmacy), generally similar aging-associated changes in fecal enteric microbiomes were noted (recently reviewed in (25)). The dysbiotic profiles typically reflected a greater abundance of facultative anaerobes (e.g., Enterobacteriaceae species) and lower abundances of obligate anaerobes with immune-regulatory and/or anti-inflammatory potential including butyrate-producing species (e.g., Roseburia) and probiotic species such as Bifidobacterium. Other groups have gone beyond taxa identification to highlight potential differences in functional capability of fecal bacteria in older persons. Fecal samples from elderly persons (mean age 86 years) had fewer bacterial genes associated with butyrate production versus younger persons (26). Rampelli and colleagues utilized a metagenomics approach and applied shotgun sequencing to total fecal bacterial DNA (27). In this small study, the aging human fecal bacterial functional profile was characterized by a loss of genes utilized in SCFA production, in starch and sucrose metabolism, and in essential amino acid metabolism, and associated with an enrichment in genes reflective of pathobiont taxa.

A number of recent cross-sectional clinical studies have investigated microbiome profiles associated with extreme aging including super-centenarians (>105 years) (28–35). Although increases in some pathobiont species (e.g., Proteobacteria spp.) and lower abundances of some butyrate-producing bacteria (e.g., Faecalibacterium) were noted in centenarians, increased abundances of other potentially health-promoting bacterial groups (e.g., Akkermansia, Bifidobacterium) were common in extreme aging. Importantly, as noted by Biagi et al., these cross-sectional studies do not delineate whether these health-associated communities are maintained with age and contribute to longevity or are lost and reacquired as new relationships are established between microbiota and host (28). Interestingly, a recent large scale study in China that investigated microbiome changes across the age spectrum (3–100 years) utilizing a compositional data analysis approach with age as a continuous variable (versus other common measures such as Bray-Curtis dissimilarity measures) found only minor variations in taxa between healthy older versus younger persons, even in centenarians (36). This highlights that, in addition to geographical location, the type of analyses undertaken may also be a critical component in how we ultimately define an age-associated dysbiotic profile.

While studies on extreme aging may aid in understanding what constitutes a ‘healthy’ microbiome, investigations into relationships between the gut microbiome and age-associated inflammation and comorbidities may shed light on potentially ‘unhealthy’ microbiome profiles. Alterations in microbiome profiles have been linked to indicators of systemic inflammation. In one study, Proteobacteria species (enriched with age) correlated positively with increased inflammatory cytokine levels (IL-6, IL-8), whereas butyrate-producing bacterial taxa (depleted with age) were inversely associated with these cytokines (29). Numerous studies of aging cohorts (age ranges 42 to >100 years) have observed potential links between gut microbial communities and a decline in physical function with age. Specifically, indicators of frailty have been associated with higher abundance of bacterial taxa in several phyla, including Enterobacteriaceae (Proteobacteria), Eggerthella sp. (Actinobacteria) and Rumminococcus (Firmicutes) and lower abundance of butyrate-producing bacteria (e.g., F. prausnitzii) (37–41). Metagenomic analyses revealed that frailty was associated with genes responsible for bacterial persistence and with Gram-negative cell wall component Lipopolysacchride (LPS) biosynthesis (38, 40). Production of SCFAs, measured in fecal water, were lower in frail persons (37), reflecting the lower abundance of SCFA-producing bacteria. A recent study provided the first causal evidence of a gut-muscle axis whereby transfer of human fecal samples from high-functioning adults led to greater grip-strength in recipient germ-free mice relative to transfer of fecal samples from low-functioning persons (42).

Potential links between the microbiome and other age-associated co-morbidities such as dementia and cardiovascular disease (CVD) have also been suggested. In elderly Italians with cognitive impairment and brain amyloidosis, higher abundances of fecal Escherichia were observed, and Escherichia abundance positively correlated with blood levels of the inflammatory cytokine IL-1β. Conversely, lower abundance of the butyrate-producing bacteria Eubacterium rectale was linked to cognitive impairment and brain amyloidosis, and abundance of this organism inversely correlated with IL-1β (43). In an elderly Japanese cohort, lower prevalence of Bacteroides was strongly associated with dementia (44). In a USA-based cohort of elderly patients with and without Alzheimer’s disease, Alzheimer’s disease was associated with lower abundance of families of bacteria belonging to Firmicutes (e.g., Rumincoccaceae, Clostridiaceae) and Actinobacteria (e.g., Bifidobacteriaceae), but with higher abundance of Bacteroidetes (e.g., Bacteroidaceae, Rikenellaceae) and Proteobacteria (e.g., Bilophila) (45). Production of the bacterial metabolite trimethylamine-N-oxide (TMAO) has been mechanistically linked to atherosclerosis (46, 47), and enteric dysbiosis has been associated with known risk factors for CVD such as hypertension, obesity, metabolic syndrome and diabetes (48).

Enteric microbial communities are altered in PWH.

Recent studies, including our own, have documented that enteric bacterial communities in PWH are different when compared to individuals without HIV infection (recently reviewed in (49)). In both ART-treated and in treatment-naïve PWH, initial studies described a dysbiotic fecal microbiome in PWH characterized by greater abundance of Prevotella with concurrent depletion of some Bacteroides species, higher abundance of Proteobacteria (including the family Enterobacteriaceae), and lower abundance of obligate anaerobes including butyrate-producing bacteria (49). Of note, a number of subsequent studies have highlighted that certain dysbiotic features, in particular high abundance of Prevotella sp., were likely attributable to sexual practices rather than to HIV infection per se (50–52). Studies suggest that antiretroviral drugs themselves may also contribute to dysbiosis, although studies directly investigating interactions between orally-administered ART drugs and gut microbiota have not been undertaken (reviewed in (53)).

Bacterial taxonomic profiling of gut tissue rather than of fecal samples may provide a better approximation of the communities of bacteria closest to the intestinal epithelium and the underlying immune cells and thus have the greatest potential to directly impact gut health and immunity. In general, profiles in gut tissue samples from PWH reflected those observed in fecal samples from independent studies, with the exception that tissue profiles were notable for observing greater abundance of a greater variety of Proteobacteria genera (e.g., Escherichia, Acinetobacter, Klebsiella) and lower abundances of genera associated with butyrate production (e.g., Roseburia and Faecalibacterium) in PWH (54–58). Of note, our study demonstrated that the total summed abundance of known butyrate-producing bacterial species was significantly lower in colonic tissue of untreated PWH compared to uninfected controls, but this difference was not observed in subject-matched fecal samples (59). Variances between fecal and mucosal tissue taxa profiles, as also noted in studies of healthy individuals (60), may reflect differences in the local environment being conducive to growth of one bacteria over another (e.g., an oxygen gradient from tissue to luminal content) and/or that the fecal samples may contain both luminal bacteria as well as those that are poorly adherent to or shed from the mucosa. Different collection procedures may also be a contributing factor. Clinic-based endoscopic collection of tissue samples likely results in the immediate appropriate storage of tissue samples whereas there is likely more variability in storage of fecal samples collected “at home” by study participants.

Both fecal and tissue-associated dysbiotic profiles have been variably linked to indicators of HIV-1 pathogenesis. Dysbiotic microbial communities in PWH positively associated with blood T cell activation and indicators of systemic inflammation and microbial translocation (54, 56, 57, 61). At an individual taxa level, numerous genera containing butyrate-producing bacteria inversely associated with microbial translocation and immune activation (54, 56, 61, 62).

A number of recent studies have begun to probe the metabolic profiles of enteric bacteria in PWH and to look for bacteria-associated functional pathways involved in HIV pathogenesis and inflammation-related comorbidities. In one study, fecal butyrate levels were significantly lower in untreated PWH with no differences noted in the other major gut-associated SCFAs (acetate or propionate) (63). By contrast, butyrate levels in ART-treated person were not found to be significantly different versus controls in another study (64). Compared to uninfected persons with infectious diarrhea caused by toxigenic Clostridioides difficile, another infection known to induce gut damage, HIV infection uniquely altered the composition of fecal bacterial metabolites (65). Among these PWH, the impact on gut bacterial metabolism was most notable for the apparent inability of gut bacteria to metabolize the amino acids proline, phenylalanine and lysine, which may be a contributing factor to nutritional deficits observed in PWH (65). Characterization of the functional gene content of the fecal microbiota in PWH either via imputed metagenomics (e.g., PICRUSt) or metagenomic sequencing noted enrichment of genes associated with LPS biosynthesis, bacterial translocation and oxidative stress resistance with depletion of genes associated with anti-inflammatory pathways (e.g., butanoate), amino acid metabolism and energy processes (61, 66–68). Other groups have reported alterations in abundances of enteric bacteria capable of tryptophan catabolism and in metabolites associated with this pathway, suggesting that increased bacteria-produced tryptophan catabolites may be a contributing factor to HIV-associated immunosuppression (57, 65). A recent study noted plasma TMAO levels correlated inversely with fecal Bacteroidetes and positively with Firmicutes, but notably, not with any known TMA-producing genera (69).

HIV and age interact to shape the gut microbiome.

Although many similarities exist in the microbiomes of people as they age and among PWH, few studies have directly investigated the collective effect in an individual aging with HIV. We recently undertook a pilot cross-sectional study based in the USA, comparing the fecal microbiome profiles of PWH aged 50–75 years on ART (≥2 years) and viral suppression to those of uninfected age-matched controls using 16S rRNA gene sequencing (70). Our analysis revealed that among older adults, HIV and age were independently associated with distinct changes in the fecal microbial communities. For example, microbiomes of older PWH were characterized by increased abundances of Enterobacter, Paraprevotella and Howardella genera and decreased abundances of Eggerthella, Alistipes, Barnesiella and Roseburia. Evaluation of the impact of age on microbiome profiles independent of HIV serostatus revealed a different profile with increased abundance of Parabacteroides, Lactobacillus, Sneathia and Desulfovibrio with decreased abundance of Butyricimonas, Howardella, Fusobacterium and Enterobacter. Critically, we demonstrated that the age effect on the fecal microbiome differed based on HIV serostatus: the abundance of certain genera (e.g., Bifidobacterium, Bacteroides, Alistipes) increased to a significantly greater extent with age in PWH than in controls. Other genera (e.g., Escherichia, Butyricimonas, Oxalobacter) significantly increased with age in PWH but decreased in controls. No significant differences were observed in fecal SCFA levels (butyrate, acetate, propionate) between the two cohorts. Associations between the top 25 genera based on all study participants and systemic inflammatory biomarkers also revealed different patterns between the two cohorts. As an example, Escherichia, Subdoligranulum and Bifidobacterium positively associated with levels of soluble TNF receptor (sTNFR) in PWH but not in controls. Our study highlights that the age effect on the gut microbiome and associations between microbiota and systemic biomarkers in an older population differ based on HIV serostatus. However, this study was limited by the small sample size preventing adjustment for a number of confounders including differences in sexual behaviors.

In another recent study, the fecal microbiomes of USA-based older (≥55 years) men who have sex with other men (MSM) with long-term controlled HIV (>10 years ART) versus age-matched MSM without HIV were evaluated by Rhoades and colleagues using 16S rRNA gene sequencing (71). The fecal microbiomes of PWH with long-term controlled HIV were enriched in Sneathia, Fusobacterium, Lactobacillus, Helicobater and Atopbium, but had lower abundance of Oxalobacter, Eggerthella and Streptococcus versus controls. Atopbium and Streptococcus are commonly found in the oral cavity, thus the authors conclude that, despite long-term adherence and undetectable viral loads, older PWH with long-term controlled HIV still have alterations in fecal microbial communities in part reflective of a loss of compartmentalization. A cross-sectional study of ART-treated (≥2 years) PWH aged >50 years versus uninfected controls aged >65 years from Mexico quantified specific fecal taxa using qPCR (72). The differences in the lower age limit for recruitment of PWH was explained by the authors as a way to account for the potential “early aging” of the immune system in PWH. In this study, the authors reported a significant increase in Proteobacteria and depletion in Firmicutes in PWH versus controls as well as increased levels of fecal propionate, but decreased butyrate and acetate levels.

Several studies have investigated fecal microbial communities in PWH with age-associated co-morbidities. In a China-based study, gut microbial dysbiosis was not independently associated with neurocognitive impairment in younger (mean age 33 years), untreated PWH with and without HIV-associated neurological diseases (HAND) after adjusting for gender, age, education level and CD4 T cell count (73). In a German-based study, no specific bacterial taxa were found to associate with coronary heart disease (CHD) in PWH (mean age 53 years), but co-occurrence of clusters of bacteria were different between PWH with and without CHD suggesting that networks of bacteria may be more pertinent in determining potential links between the gut microbiome and CHD in PWH (74). Of note, to the best of our knowledge, no studies have directly investigated potential relationships between the gut microbiome and frailty in older PWH.

Key questions that remain.

The above studies provide the first insights into how aging and HIV converge to impact the microbiome and suggest that HIV may influence age-associated effects on the gut microbiome. However, a number of questions remain.

1. How does the enteric microbiome contribute to age-associated co-morbidities such as frailty, dementia and CVD in older PWH?

   Additional well-controlled studies that identify both age and HIV effects on gut bacterial communities and their metabolic products in the setting of co-morbidities will help to identify functional pathways that may be modifiable by interventions that alter the existing microbiome (e.g., pre/probiotics, fecal transplant, dietary manipulation, exercise) to supplement current treatment strategies and improve clinical outcomes. Critical to the effectiveness of these studies in contributing to our understanding of the microbiome in HIV and aging will be the undertaking of well-controlled studies accounting for the multitude of confounders that impact the gut microbiome (e.g., diet, gender, sexual behavior, ART) and investigations into microbial communities and function at the tissue level.

2. How do age and HIV impact the virome and mycobiome in older PWH?

   Although techniques for characterizing the viral, bacteriophage and fungal populations that inhabit the human body may be less well established than those used for bacteria, both the virome (viruses, bacteriophages) and mycobiome (fungal) communities influence host physiology in both positive and negative ways (reviewed in (75, 76)) and therefore may impact gut health in older PWH. Indeed, alterations in the fecal virome of PWH (aged <50 years) were linked to features of HIV-associated pathogenesis (77–79). The prevalence of fecal Candida colonization significantly increased with age in ART-treated PWH, (80) and systemic levels of fungal β-D-glucan are known to associate with inflammatory and immune activation markers (reviewed in (81)). Treating or modulating the entire gut microbial community may dramatically improve gut health in aging PWH over simply treating bacterial dysbiosis.

3. How can the mechanisms that drive dysbiosis be elucidated in order to further improve treatment options?

   A number of clinical studies have attempted to durably alter the microbiome or block microbial translocation but thus far have met with limited success (reviewed in (82)). These types of treatments will be drastically improved by a better understanding of mechanisms by which HIV-associated mucosal damage drives dysbiosis as well as how dysbiotic microbial and metabolic profiles influence gut health and host immunity in persons aging with HIV infection.

Key Points:

• A breakdown in intestinal homeostasis and the associated leaky gut and increased microbial translocation are linked to systemic inflammation in both HIV and aging.

• Aging is associated with alterations in the gut bacterial microbiome, typically characterized by an increase in facultative anaerobes with pathogenic or inflammatory properties and a decrease in obligate anaerobes that play critical roles in maintaining intestinal homeostasis and in regulating host immunity.

• People with HIV also have dysbiotic gut microbiomes, many features of which reflect those observed in elderly persons.

• The age effect on the gut microbiome and associations between microbiota and systemic biomarkers in an older population differ based on HIV serostatus.

• Additional well-controlled studies that identify both HIV and age effects on gut bacterial communities and their metabolic products in the setting of co-morbidities will help to identify functional pathways that may be modifiable by interventions that alter the existing microbiome to supplement current treatment strategies and improve clinical outcomes.