Molecular Characterization of Group B Streptococcus (GBS) and Its Association with the Vaginal Microbiome Among Pregnant Women from Low–Middle Income Countries

Abstract

Introduction: Recto-vaginal colonization by Group B streptococcus (GBS) in expectant women is a risk factor for invasive GBS disease in newborns and may predispose to stillbirths and preterm births. The focus of the research included: i. Determine the genetic composition of colonizing GBS sequence types (STs) and clonal complexes (CCs) in pregnant women and their newborns from six sub-Saharan African and two Southeast Asian low and middle-income countries (LMICs); ii. Determine the prevalence of concurrent colonization by more than one GBS serotypes using microarray; iii. Investigate the association between vaginal GBS colonization and the vaginal microbiome in the pregnant women from LMICs. Methods: Group B streptococcus isolates cultured from rectum or vagina of pregnant women (n=816) and skin surface or throat mucosal colonizing isolates from their newborns (n=628) immediately following delivery were whole genome sequenced using next generation sequencing platform (Illumina HiSeq 2000; Illumina, USA) and analyzed using bio-informatic tools, including FastQC for quality control and the short-read sequence typing tool (SRST) to determine the capsular serotypes, sequence types and antimicrobial resistance genes (AMR). Vaginal swabs from pregnant women from Kenya, Ethiopia, Mali, Bangladesh, Nigeria and South Africa were analyzed for vaginal colonization by more than one GBS serotype using microarray, and the vaginal microbiome was investigated using open array quantitative polymerase chain reaction (qPCR) methods. Results: A total of 53 STs in pregnant women and 51 STs in newborns were identified, including five new described STs. There was high concordance at the serotype (90%; 547/608), ST (88%; 533/608) and CC level (89%; 537/608) between the maternal and newborn isolates. Consequently, further detailed whole genome sequencing analyses focused on the maternal colonizing isolates. The CC23 was the most common lineage in India (47.4%; 65/137), Mozambique (40.0%; 44/110) and Kenya (29.1%; 32/110). The lineage CC17 was common in Mozambique (27.3%; 30/110), Kenya (28.2%; 31/110) and South Africa (20%; 21/105) but absent in Southeast Asian countries. The clonal complex 255 (3.4%; 3/89), CC889 (13.5%; 12/889), and CC1212 (4.5%; 4/89) were detected exclusively in Bangladesh, albeit at low frequency. There was within serotype geographic variability in relation to ST and CC between countries. Notably, serotype III GBS isolates were variably associated with the hypervirulent CC17, including high prevalence in Mozambique (97%; 30/31), Kenya (89%; 31/35) and South Africa (77%; 17/22), modest prevalence in Mali (45%; 9/20), Nigeria (35%; 6/17) and Ethiopia (33.3%; 2/6); whilst it was not identified in any of eleven serotype III isolates from the Southeast Asian countries. All serotype III isolates (66.9%; 95/142) positive for the hypervirulent A gene (hvgA gene) were associated with CC17. A rare genotype CC934 (0.2%; 2/816) which also includes the hvgA gene, was identified in Ethiopia. The prevalence of any alpha-like family genes (alp1, alp2/3, alphaC, and rib) ranged from 80.9% (72/89) in Bangladesh to 100% (110/110) in Kenya. The erythromycin resistance gene was modestly prevalent in Bangladesh (31.5%; 8/89) and Mozambique (20.9%; 23/110), whilst the prevalence thereof was less than 5% in Ethiopia, India, Nigeria, Mali, and Kenya. Analysis of the vaginal microbiome was done using open array qPCR-based method, which was targeted to identify 52 organisms. Investigation of the vaginal microbiome was done of vaginal swabs in women from Kenya, Ethiopia and South Africa. Four community state types (CSTs) were identified, including CST I dominated by lactobacillus crispatus (3.9%; n=11/282), CST III (Lactobacillus inners dominant, 7.1%; n=20/282), CST IV (dominated by diverse anerobic bacteria, 87.2%; n=246/282), and CST V (Lactobacillus jensenii dominant, 1.8%; n=5/282). The prevalence of GBS colonization was commonly associated with CST IV-A (56.4%; n=22/39), less common in CST III (30%; n=6/20) but absent in women with CST V (n=5). Colonization with Veillonella parvula (adjusted odds ratio, 2.89; 95% confidence interval, 1.53-5.43), Staphylococcus aureus (aOR, 12.3; 95% CI, 5.31-28.24) and Candida albicans (aOR, 3.98; 95% CI, 2.33-6.79) were positively associated with vaginal GBS colonization, whilst no association was evident for other organisms. Among women with GBS vaginal colonization, more than one serotype was detected using microarray in 10% (n=72/702) of women and ranged from 3.7% (n=2/54) in Bangladesh to 14.4% (n=20/139) in Mali. Conclusion: Differences in the prevalence of GBS STs and CCs between low and middle-income countries of Southeast Asia and sub-Saharan Africa may account for the regional differences in the burden of invasive GBS disease between countries. In particular, the low prevalence of CC17 associated serotype III isolates among Southeast Asian countries (India and Bangladesh), and the high prevalence of CC17 serotype III isolates from African countries (including South Africa, Mozambique, Nigeria, and Kenya) may be contributing to the reported differences in the incidence of early-onset GBS disease between the sub-Saharan African and Southeast Asia countries. Vaginal colonization by more than one GBS serotype is low in pregnant women from LMICs. The association between vaginal GBS colonization and other bacteria colonizing the vagina in pregnant women, suggest that the maternal vaginal microbiome may also influence the risk of invasive GBS disease in their newborns.