Description
Analysis of the Genital Tract Metaproteome in South African Women: From Bacterial Vaginosis Biomarkers to HIV Prevention
Arghavan Alisoltani1, Matthys Potgieter1, Liam Bell2, Elizabeth Ross2, Arash Iranzadeh1, Zac McDonald, Imane Allali1, Nicola Mulder1, Smritee Dabee1, Shaun Barnabas1, Heather Jaspan1,3, David L. Tabb4, Linda-Gail Bekker5, Jo-Ann Passmore1, Lindi Masson1
1Institute of Infectious Disease and Molecular Medicine (IDM), University of Cape Town; 2Centre for Proteomic and Genomic Research, South Africa; 3Seattle Children's Research Institute, University of Washington; 4Division of Molecular Biology and Human Genetics, University of Stellbosch; 5Desmond Tutu HIV Foundation, South Africa
Background
Untreated bacterial vaginosis (BV) increases the risk of HIV acquisition and reproductive complications in women. Management of BV is limited as many women with asymptomatic infections remain undiagnosed and untreated. This study aimed to identify potential biomarkers of asymptomatic BV using metaproteomics.
Methods
Vaginal metaproteomic profiles of 109 HIV-negative young South African women (aged 16-22 years) were analyzed using liquid chromatography tandaem mass spectrometry. Proteins were identified and quantified using MaxQuant followed by taxonomic and functional analysis. Women were screened for BV using Nugent's scoring.
Results
Key BV-associated bacteria, Gardnerella vaginalis, Megasphaera, Prevotella, Bifidobacteriaceae, and Atopobium were significantly more abundant in women with BV, while Lactobacillus proteins were underabundant. Key underrepresented human proteins in women with BV included keratins and zinc finger protein 185, suggesting impaired epithelial structure and barrier function. Overrepresented human proteins in women with BV included those involved in regulatory and response processes, such as antimicrobial humoral immune responses. The combination of four biomarkers (Gardnerella vaginalis transaldolase, Megasphaera cold-shock DNA-binding domain protein, Atopobium vaginae Elongation factor Tu, and a human EF-hand domain family member) correctly classified 84% of women (cross-validation accuracy: 83%), with 97% sensitivity and 69% specificity.
Conclusion
Microbial populations identified as associated with BV using metaproteomics were similar to those characterized using 16S metagenomics. Four potential diagnostic biomarkers were identified that correctly classified 84% of women according to BV status.