Description
Distinct cervical microbial diversity and community structures by human papilloma-virus and cervical disease status in HIV-positive women in South Africa
Annavajhala, Medini K^, Saidu, Rakiya+, Tergas, Ana^, Kuhn, Louise^, Denny, Lynette+, Uhlemann, Anne-Catrin^
^Columbia University, New York, + University of Cape Town
Background: Cervical cancer continues to cause significant morbidity and mortality in HIV+ women especially in resource-limited settings. Although HPV infection is a necessary cause of cervical cancer, other biomarkers which could distinguish which women with HPV infection are most likely to have cervical intraepithelial neoplasia (CIN) would be beneficial. The role of the cervical microbiome in progression of HPV-related disease, especially in HIV+ women, remains unclear. Here, we aimed to identify if cervical microbiota in HIV+ women with HPV16 predict CIN.
Methods: We enrolled HIV+ women aged 30 to 65 from Khayelitsha Day Hospital in Cape Town, South Africa. Exclusion criteria were history of anogenital cancer, treatment for cervical dysplasia, or hysterectomy. HIV and HPV testing and histological diagnosis based on biopsy and/or LEEP were performed to define three experimental categories: (1) HPV16+/CIN3 (n=17), (2) HPV16+/No CIN (n=15), and (3) negative for all high-risk HPV types with no CIN (n=19). Cervical swabs were collected during gynecological exams and stored in ThinPrep. We extracted DNA (Qiagen PowerViral) and performed 16S rRNA (V3V4) sequencing. Closed-reference operational taxonomic unit (OTU) picking against the Greengenes 97 database was performed using QIIME (minimum count cutoff: 2,500). Microbiome Chao alpha-diversity and UniFrac beta-diversity were calculated using phyloseq and differential abundance testing was performed using DESeq in R.
Results: After quality-filtering of sequencing data, we were able to analyze 42/51 cervical swabs (n=16 group 1 (HPV16+/CIN3), n=14 group 2 (HPV16+/No CIN), and n=12 group 3 (No hrHPV/No CIN)). We found that samples from women with no HPV had comparable alpha-diversity to those from women with HPV16 and CIN3. However, women with HPV16 and no CIN had significantly lower alpha-diversity (Chao index of 101 compared to 130 (group 3) and 142 (group 1), p<0.05). Overall microbial community structure was not significantly different across all three groups (UniFrac PERMANOVA, P=0.218). Compared to women without hrHPV, those with HPV16 and CIN3 had cervical communities significantly enriched in Lactobacillus iners and Akkermansia muciniphila (DESeq, p<0.05, FDR<0.05). Conversely, women without hrHPV had significantly higher levels of OTUs assigned to Prevotella, Fusobacterium, Megasphaera, Anaerococcus. Furthermore, HPV16-positive women with CIN3 were significantly enriched in L. iners and two OTUs within the Rs-045 family of the TM7 phylum. HPV16-positive women with no CIN were instead enriched in Prevotella.
Conclusions: Our results support distinct microbiota by HPV and CIN status in HIV+ women. Further studies are needed to evaluate their role in the pathogenesis of cervical cancer in HIV+HPV+ women. Our findings demonstrate the potential utility of the cervical microbiome and specific microbiota as adjunct diagnostic tools to improve screening.