Hyaluronidase Impairs Neutrophil Function and Promotes Group B Streptococcus Invasion and Preterm Labor in Nonhuman Primates

Abstract

Invasive bacterial infections during pregnancy are a major risk factor for preterm birth, stillbirth, and fetal injury. Group B streptococci (GBS) are Gram-positive bacteria that asymptomatically colonize the lower genital tract but infect the amniotic fluid and induce preterm birth or stillbirth. Experimental models that closely emulate human pregnancy are pivotal for the development of successful strategies to prevent these adverse pregnancy outcomes. Using a unique nonhuman primate model that mimics human pregnancy and informs temporal events surrounding amniotic cavity invasion and preterm labor, we show that the animals inoculated with hyaluronidase (HylB)-expressing GBS consistently exhibited microbial invasion into the amniotic cavity, fetal bacteremia, and preterm labor. Although delayed cytokine responses were observed at the maternal-fetal interface, increased prostaglandin and matrix metalloproteinase levels in these animals likely mediated preterm labor. HylB-proficient GBS dampened reactive oxygen species production and exhibited increased resistance to neutrophils compared to an isogenic mutant. Together, these findings demonstrate how a bacterial enzyme promotes GBS amniotic cavity invasion and preterm labor in a model that closely resembles human pregnancy.IMPORTANCE Group B streptococci (GBS) are bacteria that commonly reside in the female lower genital tract as asymptomatic members of the microbiota. However, during pregnancy, GBS can infect tissues at the maternal-fetal interface, leading to preterm birth, stillbirth, or fetal injury. Understanding how GBS evade host defenses during pregnancy is key to developing improved preventive therapies for these adverse outcomes. In this study, we used a unique nonhuman primate model to show that an enzyme secreted by GBS, hyaluronidase (HylB) promotes bacterial invasion into the amniotic cavity and fetus. Although delayed immune responses were seen at the maternal-fetal interface, animals infected with hyaluronidase-expressing GBS exhibited premature cervical ripening and preterm labor. These observations reveal that HylB is a crucial GBS virulence factor that promotes bacterial invasion and preterm labor in a pregnancy model that closely emulates human pregnancy. Therefore, hyaluronidase inhibitors may be useful in therapeutic strategies against ascending GBS infection.

Keywords: group B streptococcus; hyaluronidase; immune evasion; neutrophils; pregnancy; preterm labor.

FIG 1

GBS hyaluronidase promotes amniotic cavity invasion and preterm labor. Chronically catheterized pregnant pigtail macaques (Macaca nemestrina) received choriodecidual inoculations of either HylB-proficient WT GBS strain GB37 (n = 5), an isogenic GBS strain lacking HylB (GB37ΔhylB, n = 5), or saline (n = 6) at 118 to 125 days gestation (term, 172 days). Uterine contractions (vertical gray lines), cytokines (IL-1β, IL-6, IL-8, and tumor necrosis factor alpha [TNF-α]), prostaglandins (PGE₂ and PGF_2α), and GBS CFU from the AF are shown from representative animals that received either GB37 (A), GB37ΔhylB (B), or saline (C).

FIG 2

MMP-1 and MMP-3 are elevated in the amniotic fluid and lower uterus of GB37-inoculated NHP. At Cesarean section, amniotic fluid and a tissue segment from the lower uterus were collected from each animal, and each sample was analyzed for MMP-1 (A) and MMP-3 (B) levels by Luminex. Data from GB37 and GB37ΔhylB were compared using a Mann-Whitney test. Although P values between 0.055 and 0.1 are noted due to the small sample size of this nonhuman primate study, P values of <0.05 were considered significant. GB37ΔhylB#5 is designated by an open square.

FIG 3

Microbial invasion of the amniotic cavity coincided with fetal bacteremia and systemic fetal inflammation. At Cesarean section, fetal tissues were harvested from chronically catheterized pregnant NHP that received a choriodecidual inoculation of saline, GB37, or GB37ΔhylB. GB37ΔhylB#5 is designated by an open square. (A) Numbers of GBS CFU obtained from various fetal tissues. Note that GBS was recovered from all fetal tissues of all animals that experienced microbial invasion of the amniotic cavity and preterm birth, including all (5/5) GB37-inoculated animals and one of five GB37ΔhylB-inoculated animals (GB37ΔhylB#5, designated by an open square). Differences in CFU between treatment groups were analyzed using the Mann-Whitney test. *, P < 0.05; **, P < 0.01. (B) Lysates from fetal tissues were examined by Luminex to evaluate differences in levels of fetal cytokines. A Kruskal-Wallis test with Dunn’s multiple-comparison test was performed. *, P < 0.05. P values between 0.05 and 0.1 are noted due to the small sample size of this nonhuman primate study but were not considered significant. Since fetal tissue lysates were not available from historical saline animals (n = 4) reported previously (28), only saline animals performed in the present study (n = 2) were included in these analyses.

FIG 4

Histological examination of the placental membranes revealed increased neutrophil infiltration in GB37-infected animals. (A to D) H&E staining of NHP placental sections. Representative H&E-stained sections from NHPs in each group are shown, including saline#3 (A), GB37#1 (B), GB37#2 (C), and GB37ΔhylB#2 (D). (E to H) Representative MPO-stained sections from NHPs in each group, including saline#3 (E), GB37#1 (F), GB37#2 (G), and GB37ΔhylB#2 (H). (I to L) Representative CD68-stained sections from NHPs in each group, including saline#2 (I), GB37#1 (J), GB37#2 (K), and GB37ΔhylB#2 (L). For the GB37 group, tissues in panels B, F, and J were taken 24 h after inoculation and tissues in panels C, G, and K were harvested 48 h after inoculation.

FIG 5

Digital spatial profiling of placental tissues revealed few differential expressions of immune signatures GBS hyaluronidase. (A) Representative placental sections from NHP in each treatment group (GB37#2, GB37ΔhylB#2, and saline#3). We treated each placental section with fluorescent anti-pan cytokeratin (Pan CK, green), anti-fibronectin attachment protein (FAP, yellow), anti-GBS (red), and DAPI (blue) and then identified the decidua, chorion, and amnion within each section as distinct ROIs. Each discrete ROI (i.e., chorion, amnion, and decidua) was analyzed for analyte abundance. (B) Heatmaps showing analyte abundance (normalized by the signal/noise ratio) in the decidua, chorion, and amnion of each animal in the GB37 (n = 5), GB37ΔhylB (n = 5), and saline (n = 4) groups are shown.

FIG 5

Digital spatial profiling of placental tissues revealed few differential expressions of immune signatures GBS hyaluronidase. (A) Representative placental sections from NHP in each treatment group (GB37#2, GB37ΔhylB#2, and saline#3). We treated each placental section with fluorescent anti-pan cytokeratin (Pan CK, green), anti-fibronectin attachment protein (FAP, yellow), anti-GBS (red), and DAPI (blue) and then identified the decidua, chorion, and amnion within each section as distinct ROIs. Each discrete ROI (i.e., chorion, amnion, and decidua) was analyzed for analyte abundance. (B) Heatmaps showing analyte abundance (normalized by the signal/noise ratio) in the decidua, chorion, and amnion of each animal in the GB37 (n = 5), GB37ΔhylB (n = 5), and saline (n = 4) groups are shown.

FIG 6

GBS HylB evades neutrophil killing independently of cell death by interfering with TLR-2/4 signaling. (A) Primary human neutrophils were isolated from fresh adult blood, exposed to GB37 or GB37ΔhylB at MOIs 100, 10, or 1 for 4 h, and then examined for cell death by LDH release. The percent live cells was calculated relative to Triton X-100-treated positive controls (0% live cells) and PBS-treated negative controls (100% live cells). (B) GB37 or GB37ΔhylB was exposed to primary human neutrophils isolated from fresh blood (MOI of 1) for 1 h. The percent killing was calculated as the number of CFU recovered after incubation with neutrophils out of the number of the number CFU recovered after incubation without neutrophils × 100. Differences among groups were determined by a paired t test. (C) Primary human neutrophils were isolated as described above, pretreated with dihydrorhodamine-123 (DHR), and then exposed to GB37 or GB37ΔhylB (MOI of 100). Since GBS capsule can suppress neutrophil ROS generation by blocking Siglec 9 (62), the GB37ΔcpsE was included as a control. The conversion of DHR to fluorescent MHR indicates ROS production in cells and was measured by flow cytometry at 60 min postinfection. Differences among treatment groups were determined by one-way analysis of variance (ANOVA). (D) Filtered supernatants of stationary-phase GB37 or GB37ΔhylB liquid cultures were incubated with HA for 18 h to allow for enzymatic digestion of HA. Meanwhile, primary human neutrophils were pretreated with 10 μg/ml anti-TLR-2 antibody (Invivogen) plus 10 μg/ml anti-TLR-4 antibody (Invivogen) or vehicle control. Neutrophils were treated with DHR and then exposed to the digested HA solutions from each strain for 60 min. As described above, ROS production in cells was measured by detecting fluorescent MHR via flow cytometry. Differences in MHR-positive cells among treatment groups were determined by one-way ANOVA. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; ns, P ≥ 0.05.