Virulence, phenotype and genotype characteristics of invasive group B Streptococcus isolates obtained from Swedish pregnant women and neonates

Abstract

Group B streptococci (GBS) are bacteria that can cause preterm birth and invasive neonatal disease. Heterogeneous expression of virulence factors enables GBS to exist as both commensal bacteria and to become highly invasive. A molecular epidemiological study comparing GBS bacterial traits, genotype and host characteristics may indicate whether it is possible to predict the risk of perinatal invasive GBS disease and more accurately target intrapartum antibiotic prophylaxis. A total of 229 invasive GBS isolates from Swedish pregnant women or neonates were assessed for virulence and phenotypic traits: hemolysis zone, hemolytic pigment (Granada agar), Streptococcus B Carrot Broth (SBCB) assay, CAMP factor, and hyaluronidase activity. Genes regulating hemolytic pigment synthesis (covR/covS, abx1, stk1, stp1) were sequenced. Of the virulence factors and phenotypes assessed, a Granada pigment or SBCB score ≥ 2 captured more than 90% of EOD isolates with excellent inter-rater reliability. High enzyme activity of hyaluronidase was observed in 16% (36/229) of the invasive GBS isolates and notably, in one case of stillbirth. Hyaluronidase activity was also significantly higher in GBS isolates obtained from pregnant/postpartum individuals versus the stillbirth or neonatal invasive isolates (p < 0.001). Sequencing analysis found that abx1 (g.T106I), stk1 (g.T211N), stp1 (g.K469R) and covS (g.V343M) variants were present significantly more often in the higher (Granada pigment score ≥ 2) versus lower pigmented isolates (p < 0.001, each variant). Among the 203 higher Granada pigment scoring isolates, 22 (10.8%) isolates had 3 of the four sequence variants and 10 (4.9%) had 2 of the four sequence variants. Although heterogeneity in GBS virulence factor expression was observed, the vast majority were more highly pigmented and contained several common sequence variants in genes regulating pigment synthesis. High activity of hyaluronidase may increase risk for stillbirth and invasive disease in pregnant or postpartum individuals. Our findings suggest that testing for GBS pigmentation and hyaluronidase may, albeit imperfectly, identify pregnant people at risk for invasive disease and represent a step towards a personalized medical approach for the administration of intrapartum antibiotic prophylaxis.

Keywords: Early-onset disease; Group B Streptococcus; Late-onset disease; Neonate; Pregnancy; Preterm birth; Preterm labor; Streptococcus agalactiae.

Fig. 1

Scoring examples of Group B Streptococcus (GBS) β-hemolysis, Granada pigment, Strep B Carrot Broth (SBCB) assay and CAMP Factor. Examples of the range of scores are shown that were associated with β-hemolysis (A), pigment on Granada agar (B), orange pigment using the SBCB assay (C) and CAMP factor (D). In panel D, an illustration is shown of the vertical Staphylococcus aureus (S. aureus) streak near the horizontal GBS streak, which can produce the CAMP factor effect that appears as an arrowhead

Fig. 2

Study groups and sources of GBS isolates. The number of invasive GBS isolates and biological source of the isolate is shown for each clinical group

Fig. 3

Virulence factor scoring of the GBS invasive isolates by test type. A dot plot of scores with mean and one standard deviation error baris shown for: A CAMP factor score on sheep blood agar, B Granada pigment score, C β-hemolysin activity on sheep blood agar, and D SBCB assay

Fig. 4

Hyaluronidase enzyme activity (boxplot: median, thick line; mean, large dot) is shown for each isolate by invasive GBS isolate group (maternal, stillbirth, EOD and LOD groups) in comparison to the GB37 wild-type (high hyaluronidase activity) and GB37ΔhylB (low hyaluronidase activity). The hyaluronidase enzyme activities of GB37 WT and GB37ΔhylB were assayed on each plate as a control. All GBS isolates with a pigment score greater than zero are shown as small dots. Non-pigmented GBS isolates are shown as blue triangles (Granada pigment score = 0). Note that 451 units of hyaluronidase activity/mg is equivalent to 1 mg/ml of hyaluronidase. We considered a hyaluronidase enzyme activity of > 10 mg/mL as “high”

Fig. 5

Correlation between hyaluronidase activity and Granada pigment, hemolysis, and SBCB scores

Fig. 6

Assay grade scores (A) and hyaluronidase enzyme activity (B) are shown by GBS capsular serotype (Ia-IX). Serotype VI was not observed in these isolates. The symbols represent the mean assay grade (A) or hyaluronidase activity (B) with error bars indicating 1 standard error of the mean

Fig. 7

Scatter plot of Granada pigment score for maternal and neonatal isolates by amino acid position grouped by frequent non-synonymous mutations in covR, covS, abx1, and stk1/stp1. Isolates are represented with a dot for each non-synonymous mutation in the gene of interest and may be represented more than once

Fig. 8

Venn diagram representing isolates with sequence variants in abx1, stk1, stp1, and covS that also had higher Granada pigment scores (2–4). The red line encircles the abx1 T106I variants, green line the stk1 T211N variants, turquoise line the stp1 K469R variants and the purple line the covS V343M variants