A study of the genital microbiotas of black South African women and men: associations with human papillomavirus and HIV infections

Onywera, David Harris

A study of the genital microbiotas of black South African women and men: associations with human papillomavirus and HIV infections

Doctoral Thesis

2018

Abstract

Persistent genital infection with oncogenic or high-risk human papillomavirus (HPV) is causally associated with cervical cancer in women and some penile cancers in men. The role of the complex genital microbiota in HPV infection has not been extensively addressed. This study characterised the genital microbiotas of heterosexually-active Black South African women and men, predominantly of the Xhosa ethnicity, recruited from a community in Cape Town, South Africa. The association of the genital microbiotas with prevalent HPV, HIV, demographic, behavioural, and clinical characteristics of the participants was examined. In Chapter 2 the bacterial communities in cervicovaginal samples from 62 HIVseronegative South African women were profiled by Ion Torrent PGM sequencing of the V4 hypervariable region of the bacterial 16S rRNA gene (IT-V4 method). The cervicovaginal microbiotas (CVMs) were found to cluster into three distinct community state types (CSTs): Lactobacillus iners-dominated CVMs (CST I (38.7%, 24/62)), unclassified Lactobacillusdominated CVMs (CST II (4.8%, 3/62)), and diverse CVMs (CST III (56.5%, 35/62)) with an array of heterogeneous bacteria, predominantly the bacterial vaginosis (BV)-associated Gardnerella, Prevotella, Sneathia, and Shuttleworthia. The majority of the women had nonLactobacillus-dominated CVMs. Lactobacilli are recognised as protective against sexually transmitted infections. Among the Lactobacillus species detected in the women, L. iners was the most prevalent and abundant. This species is recognised as the least protective amongst the vaginal lactobacilli. Women in CST I were more likely to be on hormonal contraception compared to women in CST III (relative risk (RR): 2.6 [95% CI 1.3-5.3]; p=0.005). Further research is required to confirm this association and to determine the biological mechanism. Microbiome research methodologies are constantly improving and in Chapter 3 the performance of two bacterial 16S rRNA gene amplicon-based methodologies were compared. The CVMs of 19 women were characterised using the IT-V4 method (Chapter 2) and using the Illumina 16S rRNA metagenomics method (IM-V3/V4 method). The latter method involves sequencing the V3 and V4 hypervariable regions of the 16S rRNA gene on the Illumina MiSeq platform. The two methods showed a high degree of correlation (r=0.89, p< 0.0001) in the average relative abundance of shared bacterial taxa. Procrustes analyses of the weighted UniFrac distances further showed a statistically consistent clustering between the two methods (M 2 =0.3, p< 0.0001). The IM-V3/V4 method proved to have a greater throughput, longer read-length, and lower error rates than the IT-V4 method and was therefore used in the subsequent chapters (4 and 5). In Chapter 4, the CVMs of 87 HIV-seronegative women from the same cohort were examined using the IM-V3/V4 method. The CVMs clustered into eight CSTs. Only 23 women (26.4%) had CVMs dominated by a single Lactobacillus species, this included two women (2.3%) with L. crispatus (CST-1), two (2.3%) with L. jensenii (CST-2), and 19 (21.8%) with L. iners (CST-3). The majority of the women (64.4% (56/87)), however, had diverse and heterogeneous CVMs (CST-8) that were associated with BV (p2, p<0.05). In the final experimental chapter, the penile microbiotas of 238 Black South African men were characterised. This is the first large-scale study of the penile microbiota of South African men. Corynebacteriaceae (47.2%) and Prevotellaceae (6.6%) were found to be the most abundant bacterial families. The penile bacterial communities clustered into six CSTs (designated 1-6). A majority of the men (53.4% (127/238)) had Corynebacterium-dominated microbiotas (CST-1). The remaining CSTs (2-6) had lower relative abundances of Corynebacterium than CST-1 and were colonised with several vaginal bacteria. The prevalences of these CSTs (2-6) in men together with their respective most abundant genera (besides Corynebacterium) were as follows: CST-2 (9.2%; unclassified Clostridiales and Porphyromonas), CST-3 (8.8%; Gardnerella), CST-4 (7.6%; Chryseobacterium and Acinetobacter), CST-5 (18.5%; unclassified Clostridiales and Porphyromonas), and CST-6 (2.5%; Lactobacillus). One hundred and thirty (54.6%) and 102 (42.9%) of the men were positive for HPV and high-risk HPV, respectively, as detected by the Roche Linear Array HPV Genotyping assay. Of the 130 HPV-positive men, 37 (28.5%) and 93 (71.5%) had single and multiple HPV types, respectively. Men in CST-1 were less likely to have high-risk HPV and multiple HPV infections relative to men in CSTs 2-6 (RR: high-risk HPV: 0.8 [95% CI 0.6-1.0]; p=0.027 and multiple HPV: 0.8 [95% CI 0.6-1.0]; p=0.042). LefSe revealed that prevalent HPV infection was strongly associated with higher relative abundances of Sneathia, Porphyromonas, Prevotella, Dialister, and Campylobacter (LDA score >3, p3, p< 0.05). The relative abundances of the latter three bacteria together with Peptoniphilus were strongly associated with high-risk HPV infection (LDA score >3, p <0.05). In our cohort, 88 men (37.0%) were positive for HIV. Of these, 71.6% and 60.2% were positive for HPV and highrisk HPV infection, respectively. Among the HIV-negative men (n=150), 44.7% and 32.7% were positive for HPV and high-risk HPV infection, respectively. Although HIV status did not impact the overall composition of the penile microbiotas, HIV-infected men had higher relative abundances of Staphylococcus, Faecalibacterium, Strenotrophominas, Jonquetella, Ruminococcus, Roseburia, and Lamia (LDA score >2, p<0.05) Men with BV-negative female sexual partners (66.5% (157/236)) had higher relative abundances of Lactobacillus in their penile microbiotas than men with BV-positive female partners (p=0.007). Atopobium, Sneathia, and Saccharofermentans were significantly more prevalent in men with BV-positive female partners than men with BV-negative partners (p<0.020). The main limitations of our study include relatively small sample size of women, insufficient participant information such as host genetics, other STIs (e.g., herpes simplex virus) and abnormal vaginal flora (e.g., aerobic vaginitis), using a less sensitive method to diagnose BV in women, and inherent biases evident in any retrospective study. Moreover, we did not adjust for confounding factors in our analysis due to the small sample size. Despite the underscored limitations, our findings provide insight into the baseline genital microbiotas of the Black South African women and men. The associations identified in this cross-sectional study between specific microbiota members and HPV infection, particularly the association between Sneathia and HPV/high-risk HPV infection, identified in both women and men, are hypothesis-generating and warrant further investigation. The study forms a critical starting point for future longitudinal confirmatory association studies and studies examining these bacteria as potential biomarkers or risk factors for HPV infection.

Keywords

Medicine

Reference:

Collections

PhD / Doctoral

Full item page