Spatial profiling of the placental chorioamniotic membranes reveals upregulation of immune checkpoint proteins during Group B Streptococcus infection in a nonhuman primate model

Abstract

Background: Preterm birth is a leading cause of neonatal mortality, which is often complicated by intrauterine infection and inflammation. We have established a nonhuman primate model of Group B Streptococcus (GBS, Streptococcus agalactiae) infection-associated preterm birth. Immune checkpoints are modulators of the immune response by activating or suppressing leukocyte function and are understudied in preterm birth. The objective of this study was to spatially profile changes in immune protein expression at the maternal-fetal interface during a GBS infection with a focus on immune checkpoints.

Methods: Twelve nonhuman primates (pigtail macaques, Macaca nemestrina) received a choriodecidual inoculation of either: 1) 1-5 X 10⁸ colony forming units (CFU) of hyperhemolytic/hypervirulent GBS (GBSΔcovR, N=4); 2) an isogenic/nonpigmented strain (GBS ΔcovRΔcylE, N=4); or, 3) saline (N=4). A Cesarean section was performed at preterm labor or 3 days after GBS infection or 7 days after saline inoculation. Nanostring GeoMx® Digital Spatial Profiling technology was used to segment protein expression within the amnion, chorion, and maternal decidua at the inoculation site using an immuno-oncology panel targeting 56 immunoproteins enriched in stimulatory and inhibitory immune checkpoint proteins or their protein ligands. Statistical analysis included R studio, Kruskal-Wallis, Pearson and Spearman tests.

Results: Both inhibitory and stimulatory immune checkpoint proteins were significantly upregulated within the chorioamniotic membranes and decidua (VISTA, LAG3, PD-1, CD40, GITR), as well as their ligands (PD-L1, PD-L2, CD40L; all p<0.05). Immunostaining for VISTA revealed positive (VISTA+) cells, predominantly in the chorion and decidua. There were strong correlations between VISTA and amniotic fluid concentrations of IL-1β, IL-6, IL-8, and TNF-α (all p<0.05), as well as maternal placental histopathology scores (p<0.05).

Conclusion: Differential regulation of multiple immune checkpoint proteins in the decidua at the site of a GBS infection indicates a major perturbation in immunologic homeostasis that could benefit the host by restricting immune-driven pathologies or the pathogen by limiting immune surveillance. Protein expression of VISTA, an inhibitory immune checkpoint, was upregulated in the chorion and decidua after GBS infection. Investigating the impact of innate immune cell expression of inhibitory immune checkpoints may reveal new insights into placental host-pathogen interactions at the maternal-fetal interface.

Keywords: Group B Streptococcus; amnion; chorion; decidua; immune checkpoint; placenta; pregnancy.

Figure 1

This figure shows the study design. The x-axis reflects both days with respect to inoculation and is not to scale. The approximate days since conception are indicated in text. There were four nonhuman primates (NHP) in each experimental group: saline, GBSΔcovR, and GBSΔcovRΔcylE.

Figure 2

Segmentation of the amnion, chorion and decidua was performed using the Nanostring GeoMx® Digital Spatial Profiling platform. Row (A) demonstrates the initial field of view using fluorescent microscopy and the regions of interest drawn to segment the amnion, chorion, and decidua. Row (B) demonstrates the segmented tissue regions prior to UV cleavage and capillary collection of the probes from the area of interest.

Figure 3

Differential expression of DSP antigens by Region of Interest for GBSΔcovR versus Saline. The log10 S6 normalized antigen counts for GBSΔcovR (red) and saline controls (blue) are shown on the x-axis by region of the chorioamniotic membranes with the mean (dot) and +/- 1 standard deviation (bar). Although the entire GeoMx panel included 56 antigens, we show only the 24 antigens (y-axis) that were significantly different between the GBSΔcovR and saline controls in either the amnion, chorion or decidua using Kruskal-Wallis (*, p <0.05).

Figure 4

Differential expression of DSP antigens by Region of Interest GBSΔcovR versus GBSΔcovRΔcylE. The log10 S6 normalized antigen counts for GBSΔcovR (red) and GBSΔcovRΔcylE (green) are shown on the x-axis by region of the chorioamniotic membranes with the mean (dot) and +/- 1 standard deviation (bar). Although the entire GeoMx panel included 56 antigens, we show only the 18 antigens (y-axis) that were significantly different between the GBSΔcovR and saline controls in either the amnion, chorion or decidua using Kruskal-Wallis (*, p <0.05).

Figure 5

VISTA correlation and expression by DSP and Immunohistochemistry. (A) shows the relationship between log₁₀DSP antigen counts and percent positive VISTA immunostaining (DAB, brown stain) out of the total area stained (DAB + haematoxylin blue stain) for GBSΔcovR (red) and GBSΔcovRΔcylE (green), and controls (blue) in the decidua. VISTA (B) and CD163 (C) immunostaining of the amnion, chorion and decidua are shown to demonstrate morphologic overlap between the staining profiles in a saline control. Immunohistochemistry targeting VISTA in the decidua is shown for GBSΔcovR (D) and GBSΔcovRΔcylE (E) groups to show the cellular influx of VISTA+ cells in the GBSΔcovR group.

Figure 6

Heatmap of correlations between DSP antigen counts and amniotic fluid cytokine concentrations by Region of Interest. Log10 cytokine concentrations were correlated to log10 S6 normalized counts of immunoproteins (y-axis) for GBSΔcovR, GBSΔcovRΔcylE and controls. All 56 immunoproteins have been included, to evaluate the correlation in either the amnion, chorion or decidua using Spearman rank-order correlation (*, p <0.05; **, p<0.01; ***, p<0.001).

Figure 7

Correlations between DSP antigen counts and Peak 24-hour contraction area. The peak average hourly contraction area per day was correlated to log10 S6 normalized counts of immunoproteins (y-axis) for GBSΔcovR, GBSΔcovRΔcylE and controls. All 56 immunoproteins have been included, to evaluate the correlation in either the amnion, chorion or decidua using Spearman Rank order correlation coefficient. Black dots indicate non-significant (ns) antigens, while colored dots indicate antigens with a statistically significant p-value (light pink, p <0.05).

Figure 8

Correlations between DSP antigen counts and placental redline maternal stage score. Placental Redline Maternal Stage scores were correlated to log10 S6 normalized counts of immunoproteins (y-axis) for GBSΔcovR, GBSΔcovRΔcylE and controls. All 56 immunoproteins have been included, to evaluate the correlation in either the amnion, chorion or decidua using Spearman Rank order correlation coefficient. Black dots indicate non-significant (ns) antigens, while colored dots indicate antigens with a statistically significant p-value (light pink, p <0.05; medium pink, p<0.01; red, p<0.001).