HIV and the Microbiome Volume 26 Issue 3 September 2018

Perspective The Human Microbiome: Implications for Health and Disease, Including HIV Infection

Our increased understanding of the human microbiome has brought insight The National Institutes of Health into the role it plays in health and disease, including HIV infection. Studies Human Microbiome Project has pro- have shown that the gut microbiome is less diverse in individuals with HIV duced a number of insights regarding infection than in noninfected control subjects. Efforts to modify the micro- the microbiome over the past 15 years, biome to bolster immune reconstitution in people with HIV infection have so beginning with the recognition that far been unsuccessful. The vaginal microbiome affects risk of HIV acquisition, each individual is not a single ecosys- with Lactobacillus dominance being protective compared with vaginosis tem but numerous ecosystems with characterized by larger populations of Gardnerella. The vaginal microbiome differences among body sites. Further, might also affect efficacy of topical tenofovir disoproxil fumarate preexposure whereas OTUs in the GI tract are very prophylaxis. This article summarizes a presentation by Robert T. Schooley, different from OTUs in the oral cavity MD, at the IAS–USA continuing education program held in San Francisco in a given individual, OTUs from these sites resemble each other across indi- in May 2018. viduals. In addition, in a given individ- Keywords: microbiome, HIV, gut, vaginal, Lactobacillus, Gardnerella ual, the microbial diversity at particular sites changes over time, although the difference in this regard is not as great The human microbiome is the collec- greater than 97% DNA sequence com- as the difference between individuals tion of organisms that lives in and on monality in the rRNA gene. This group- at a given site. us. They live on the skin, in the gas- ing may or may not reflect what has With regard to the gut microbiome trointestinal (GI) tract, and on virtually been traditionally considered a species in particular, we have identified a num- every other surface of the body. Each of microorganism. Diversity in microbial ber of key characteristics that affect human is composed of approximately communities is described by the metrics health and disease. (1) The microbi- 37 trillion cells and we carry some 100 of richness (number of OTUs), evenness ome is established in early life and dif- trillion microbiologic organisms. Our (relative abundance of different OTUs), fers between infants born by cesarean increasing knowledge of the microbi- dominance (emergence of a single OTU), delivery and by vaginal birth. (2) In ome has been made possible through and diversity index (calculated index of murine models, transfer of an “obese” use of metagenomics and the decreas- complexity based on richness and even- microbiome from obese adult animals 1 to a sterile infant mouse affects the ing costs of genetic sequencing, obviat- ness). Antibiotic treatment narrows the ing the need to grow microorganisms growth characteristics of the mouse for in culture for identification purposes. Table. Terms Describing Diversity in life. (3) The microbiome is drastically 1 The best handle we have on under- Microbial Communities. altered acutely by antibiotic adminis- standing the microbiome is 16s ribo- tration. (4) In farm animals (and mice), Term Definition somal RNA (rRNA). Every microbe has antibiotic exposure early in life pro- this region of RNA, including portions Richness Number of OTUs motes microbiota associated with high that are conserved across all of our caloric efficiency. (5) There is some evi- Evenness Relative abundance of OTUs microbiota, because it is the common dence that early exposure to antibiotics feature that allows organisms to trans- Dominance Emergence of single OTU affects humans the same way. (6) After late messenger RNA into protein. Am- cessation of antibiotics, the antibiotic Diversity Index Calculated index of complexity “footprint” remains—ie, the microbi- plification of variable regions permits based on richness and evenness identification of organisms by families ome reverts toward but does not reach Abbreviation: OTU is operational taxonomic pre-antibiotic richness and diversity for and phyla, for example, by evaluating units. relationships among sequences. a prolonged period. A number of concepts have evolved rich microbiome (rich diversity of OTUs) The 4 major phyla in the gut micro- to characterize microbial communities. in the human gut. Conditions such as biome are: firmicutes—Gram-positive One is the operational taxonomic unit obesity and inflammatory bowel dis- organisms (eg, Staphylococcus, Micro- (OTU), a group of organisms that share ease are characterized by low diversity, coccus, Streptococcus, and Lacto- whereas bacterial vaginosis is charac- bacillus), including sporulating Gram- Dr Schooley is Professor and Vice Chair of terized by an unhealthy high diversity positive rods; Bacteroidetes—3 large the Department of Medicine at University of in which lactobacilli that are dominant classes of Gram-negative, non-spore form- California San Diego. He is a member of the in the healthy vagina are replaced by ing, anaerobic or aerobic, rod-shaped Board of Directors of the IAS–USA. a wide variety of other organisms. bacteria (eg, Bacterioides); Actinobacteria

75 IAS–USA Topics in Antiviral Medicine

● No HIV infection inflammation. The recently reported 200 ● HIV infection Promaltia study, performed by Serrano- Villar and colleagues, evaluated whether 150 the gut microbiome could be altered in ● ● ● HIV infection to improve immune re- ● 3 ● constitution. In the placebo-controlled 100 ● ● ● ● ● ● study, patients with advanced HIV ● ● ● infection were randomly assigned to ● ● ● 50 receive treatment with a placebo (skim ● ● ● milk) or an immunonutrition mixture Number of Observed Species of prebiotics, probiotics, oligonutrients, ● 0 ● 0 100 200 300 400 500 600 700 800 900 1000 docosahexaenoic acid, eicosapentaenoic Number of Sequences acid, gamma linolenic acid, and amino acids for 48 weeks, each in combination with initial antiretroviral therapy. The Figure 1. Reduced richness of the gut microbiome with HIV infection compared with no HIV infection. Adapted from Mutlu et al.2 mixture represented a collation of sub- stances used in prior studies aimed at —primarily Gram-positive bacteria (in- Some of these differences may be producing changes in the microbiome. cluding those that are acid fast; eg, due to changes in the innate and adap- At 48 weeks, there were no significant Mycobacteria and Gardnerella); and pro- tive immune response at the level of differences in microbiome diversity teobacteria—primarily Gram-negative the gut, one of the first organs attacked between the treatment and placebo rods (eg, Escherichia coli, Salmonella). during acute HIV infection. Some dif- groups. Some enrichment of unclassi- The firmicutes, in particular, are highly ferences may reflect use of antibiotics fied bacteria from the Lachnospiraceae adept at metabolizing food to help us in individuals with HIV infection. Stud- and Victivallaceae families and deple- better absorb calories. ies that have attempted to evaluate tion of Blautia species were observed in the change in microbiome over time as the treatment group compared with the HIV infection progresses suggest that placebo group. No differences in inflam- Gut Microbiome in HIV Infection some of the organisms found more fre- matory markers were found between quently in HIV-infected microbiomes treatment and placebo groups, with An initial study investigating the gut are more associated with inflamma- no significant differences in changes microbiome in HIV infection performed tion and become enriched as immune in CD4+ or CD8+ cell counts being by Mutlu and colleagues demonstrated response declines. observed between groups over time. differences between individuals with The microbiome is important from Thus, the test of whether or not HIV and without HIV infection. The study the standpoint of maintenance of the immune reconstitution can be achieved involved 56 samples from individu- gut barrier. A considerable amount of by changing the microbiome has not als with HIV infection (mean CD4+ work has been done over the past really been documented, since the cell count of 425/µL; plasma HIV-1 RNA decade to ascertain whether trans- treatment in this study did not change level <75 copies/mL in 17 of 21 par- location of gut bacterial components the microbiome. There is more work to ticipants) and 65 samples from 22 across the GI tract have an impact on be done in this area. age- and sex-matched noninfected control participants undergoing elec- HIV infection No HIV infection tive colonoscopy (primarily for cancer 20 20 screening).2 The microbiome was found to be markedly less diverse in 15 15 individuals with HIV infection (Figure 1). In addition, the microbiomes of individuals with HIV infection and non- 10 10 infected control participants contained different bacterial species, with nonin- 5 5 fected control participants harboring Number of Participants significantly more Bacteroides (P = .0002) and Clostridiales-unclassified 0 0 (P = .002), and participants with HIV 0 1 2 3 0 1 2 3 infection harboring more Enterobac- Species Diversity (using the Shannon Diversity Index) teriaceae-unclassified P( = .0002) and Campylobacter (P = .0034), as well as Figure 2. Increased risk of HIV infection with greater vaginal bacterial diversity. Adapted from more fusobacteria and Prevotella. McClelland et al.5

76 HIV and the Microbiome Volume 26 Issue 3 September 2018

Lactobacillus-dominant Non–Lactobacillus-dominant among those without Lactobacillus 0.40 0.40 dominance and greater frequency of Gardnerella (Figure 3). These investiga- 0.35 Efficacy, 61% 0.35 Efficacy, 18% tors found that when TDF was added 95% CI, 0.11-.084 95% CI, −0.77-0.63 to cell culture in which Gardnerella was 0.30 0.30 TDF TDF grown, the organism degraded tenofo- Placebo Placebo 0.25 vir, metabolizing it to adenine, whereas 0.25 P = .644 no such effect was observed when 0.20 0.20 TDF was put into culture containing Lactobacillus. These findings indicate 0.15 0.15 that Gardnerella metabolizes tenofovir 0.10 P = .013 0.10 before it can get into the cells and pro- Probabillity of HIV Infection Probabillity of HIV Infection tect the cells from becoming infected 0.05 0.05 with HIV; as shown in Figure 4, the HR = 0 .39 (95% CI: 0.20, 0.83) HR = 0 .82 (95% CI: 0.40, 1.65) presence of Gardnerella prevents cellu- 0 0 0 0.5 1.0 1.5 2.0 2.5 0 0.5 1.0 1.5 2.0 2.5 lar uptake of tenofovir and intracellular Years in Study Years in Study conversion to the active diphosphate form. Klatt and colleagues have now Figure 3. Effect of the vaginal microbiome on effectiveness of topical tenofovir disoproxil also reported that dapivirine is also fumarate (TDF) preexposure prophylaxis in Lactobaccillus-dominant and non–Lactobaccillus- metabolized and degraded in women dominant cell cultures. HR is hazard ratio; CI, confidence interval. Adapted from Klatt et al.6 with the more diverse, non-Lactobacil- lus-dominant microbiomes.7 There is Vaginal Microbiome in HIV vaginal bacterial diversity in the context no evidence that bacterial vaginosis of decrease in Lactobacillus dominance affects the efficacy of oral preexposure The vaginal microbiome may be of (Figure 2). These findings provide prophylaxis.8 considerable importance in HIV trans- some evidence that the microbiomic mission. The vaginal microbiome tends signature characterized by Lactoba- Conclusion to be stable over time within indi- cillus dominance and less diversity is viduals. In the healthy state, there is protective against HIV acquisition. The Humans live in concert with an im- minimal diversity, with dominance of study also found that presence of par- mense array of microbial organisms Lactobacillus. However, in bacterial ticular bacterial species was associated from birth to death. The microbiome vaginosis, there is increased diversity, with increased risk of acquiring HIV differs from body site to body site with dominance of Gardnerella. Lac- infection. It is of considerable interest 0.07 tobacillus populations are decreased whether further studies confirm that in the context of sexual intercourse particular species are associated with 0.06 ◆ (assessed by the presence of prostate greater transmission risk and whether ▼■ 0.05 specific antigen[ PSA] in the sample) this might lead to protective measures. and are also decreased in amenorrheic 0.04 ◆▼ women, but inflammatory mediators Vaginal Microbiome and Topical ■ (interleukin [IL]-8, IL-12, and MIP-1b) 0.03 Preexposure Prophylaxis 4 ▲ are increased. The presence of inflam- 0.02 ◆▼ ▲ matory mediators is also increased in The efficacy of topical tenofovir diso- ■ 0.01 the presence of PSA. proxil fumarate (TDF) preexposure Diphosphate (mg/mL) Tenofovir ▲ Studies have now shown that risk of prophylaxis in preventing HIV infec- 0◆▼●■ ● ● ● HIV acquisition is increased with bac- tion has been lower than anticipated 0 15 30 45 60 Time (minutes) terial vaginosis and increased vaginal in clinical trials, with poor adherence ● bacterial diversity. In a case-control largely being considered the culprit. TFV alone ▼ TFV + cells study among African women who had However, a recent study by Klatt and ▲ TFV + cells + Gardnerella vaginalis undergone microbiomic studies and colleagues in the population of the ■ TFV + cells + Lactobaccillus iners were at similar risk of HIV infection, CAPRISA (Centre for the AIDS Program ◆ TFV + cells + Lactobaccillus crispatus McClelland and colleagues found that of Research in South Africa) trial of participants in the control group who topical TDF indicates that the vaginal Figure 4. Depletion of tenofovir (TFV) by did not acquire HIV infection had 6 microbiome affects efficacy. In partic- Gardnerella vaginalis in culture prevents greater dominance of Lactobacillus ular, preventive efficacy was 61% (P = accumulation of intracellular tenofovir in their microbiome than in those .013) versus placebo among women diphosphate. No such effect is observed who acquired infection.5 Women who with Lactobacillus-dominant microbi- with Lactobacillus species. Adapted from acquired HIV infection had greater omes, compared with 18% (P = .644) Klatt et al.6

77 IAS–USA Topics in Antiviral Medicine

and from person to person, but is rea- Presented by Dr Schooley in May 2018. First women from sub-Saharan Africa. Sci Rep. draft prepared from transcripts by Matthew 2017;7(1):11974. sonably stable over time unless it is 5. McClelland RS, Lingappa JR, Srinivasan disturbed by antibiotics. Other envi- Stenger. Reviewed and edited by Dr Schooley S, et al. Evaluation of the association be- ronmental factors including food and in July 2018. tween the concentrations of key vaginal bacteria and the increased risk of HIV those with whom we associate also Financial affiliations in the past 12 months: acquisition in African women from five have direct effects on the microbi- cohorts: a nested case-control study. Lan- Dr Schooley has served as a consultant to cet Infect Dis. 2018;18(5):554-564. ome. The microbiome influences gut CytoDyn, Pfizer, and Vir Biotechnology. He 6. Klatt NR, Cheu R, Birse K, et al. Vaginal metabolism, systemic immunity, and has received grant support from Gilead Sci- bacteria modify HIV tenofovir microbi- cide efficacy in African women.Science . susceptibility to disease. It changes in ences, Inc, and holds stock or stock options in 2017;356(6341):938-945. health and disease, including in HIV Antiva Biosciences and CytoDyn. 7. Klatt N. The vaginal microbiome and ac- infection. We are only now beginning quisition of HIV infection [CROI Abstract 64]. In Special Issue: Abstracts From the to distinguish the carts from the horses. 2018 Conference on Retroviruses and Can we alter the microbiome? Yes, References Opportunistic Infections. Top Antivir Med. but currently only with the elegance of 1. Cui L, Morris A, Huang L, et al. The micro- 2018;26(suppl 1):1-483. biome and the lung. Ann Am Thorac Soc. 8. Heffron R, McClelland RS, Balkus JE, et a meat cleaver, as occurs with antibi- 2014;11 Suppl 4:S227-S232. al. Efficacy of oral pre-exposure prophy- otic treatment. Antibiotics are the most 2. Mutlu EA, Keshavarzian A, Losurdo J, et laxis (PrEP) for HIV among women with al. A compositional look at the human abnormal vaginal microbiota: a post-hoc dramatic way we affect the human gastrointestinal microbiome and immune analysis of the randomised, placebo-con- microbiome, usually for the worse, as activation parameters in HIV infected sub- trolled Partners PrEP Study. Lancet HIV. jects. PLoS Pathog. 2014;10(2):e1003829. 2017;4(10):e449-e456. in the case of Clostridium difficile over- 3. Serrano-Villar S, de LM, Vazquez-Castel- growth and Salmonella infection. Other lanos J, et al. Effects of immunonutrition Top Antivir Med. 2018;26(3):75-78. approaches to beneficially change the in advanced HIV disease: a randomized 2018, IAS–USA. All rights reserved placebo controlled clinical trial (Promaltia © microbiome, including probiotics, pre- Study). Clin Infect Dis. 2018;[Epub ahead biotics, fecal transplantation, and bac- of print]. 4. Jespers V, Kyongo J, Joseph S, et al. A teriophage therapy (eg, for C difficile), longitudinal analysis of the vaginal micro- are being investigated. biota and vaginal immune mediators in

78