Subventricular zone stem cell niche injury is associated with intestinal perforation in preterm infants and predicts future motor impairment

Abstract

Brain injury is highly associated with preterm birth. Complications of prematurity, including spontaneous or necrotizing enterocolitis (NEC)-associated intestinal perforations, are linked to lifelong neurologic impairment, yet the mechanisms are poorly understood. Early diagnosis of preterm brain injuries remains a significant challenge. Here, we identified subventricular zone echogenicity (SVE) on cranial ultrasound in preterm infants following intestinal perforations. The development of SVE was significantly associated with motor impairment at 2 years. SVE was replicated in a neonatal mouse model of intestinal perforation. Examination of the murine echogenic subventricular zone (SVZ) revealed NLRP3-inflammasome assembly in multiciliated FoxJ1⁺ ependymal cells and a loss of the ependymal border in this postnatal stem cell niche. These data suggest a mechanism of preterm brain injury localized to the SVZ that has not been adequately considered. Ultrasound detection of SVE may serve as an early biomarker for neurodevelopmental impairment after inflammatory disease in preterm infants.

Keywords: cerebral palsy; cranial ultrasound; ependymal; inflammasome; intestinal perforation; necrotizing enterocolitis; neonatal brain injury; neural stem cell; neurodevelopmental outcome; niche; preterm infant; subventricular zone; white matter injury.

Figure 1.. SVE is associated with intestinal perforation and motor impairment.

(A) Initial coronal (left) and sagittal plane (right) images through the caudothalamic groove in ELBW infant. Images were obtained on day of life (DOL) 13 and show no intraventricular hemorrhage or abnormal echogenicity. (B) Coronal (left) and sagittal plane (right) images through same levels in the same infant at 3 months of age following a history of NEC and intestinal perforation depict SVE (arrowheads). (C) Multinomial regression testing for association between common preterm risk factors and combined SVE score (in ELBW cohort) after adjusting for confounders (subset of risk factors presented, see table S2 for full analysis). (D to G) Time course of SVE development in a single patient born at 24 weeks gestation prior to surgical NEC after 4 weeks of life. (D) normal coronal CUS on DOL 6 without hemorrhage or SVE. (E) Developing SVE (arrowheads), 14 days after intestinal perforation. (F) 1 month after intestinal perforation depicts more robust SVE (arrowheads). (G) Development of bilateral microcystic evolution in the subventricular zone region of the germinal matrix (asterisks). (H) Intestinal perforation cohort (N = 143), 62 infants were excluded due to death or other neurologic sequelae leaving 81 patients available for neurodevelopmental assessment. (I) Histogram from SVE-positive cases (N = 49, Intestinal perforation cohort) showing when SVE was observed (days) in retrospective CUS relative to intestinal perforation event (day 0). (J) Ordinal logistic regression testing (intestinal perforation NDI subset) for association between 2 year Bayley scales for motor and cognitive domains and average SVE Likert after adjusting for gestational age (complete case analysis). There was a significant association with motor impairment at 2 years of age.

Figure 2.. SVE Induction Following MIP in Mice.

(A and B) Human SVE obtained by CUS before and after perforation for comparison to mouse studies. (C) MIP was induced on p5. On p8–9, mice were anesthetized for CUS. Region of interest (ROI) is shown. (D) Mouse undergoing CUS. (E) Representative images from control littermates (left) and MIP cohort (right). (F) Mean gray echogenicity value was obtained in Image J from the ROI depicted in (C). Analysis of control (n=5) and MIP (n=8) mice reveals an increased mean gray value (echogenicity). Mean ± SEM. Unpaired t-test. ****P < 0.0001.

Figure 3.. MIP induces subventricular zone NFκB-linked inflammatory injury.

(A) 24 hours after MIP induction, qPCR was used to quantify IL1β, Tnfα, and IL-6 transcripts in spleen. Number of biological replicates used in each analysis is shown. Unpaired T Test. *P<0.05,**P<0.01. (B) qPCR was used to quantify IL1β, Tnfα, and IL-6 transcripts in SVZ tissue. Number of biological replicates used in each analysis is shown. Unpaired T Test. *P<0.05,**P<0.01. (C) Volcano plot of DEGs from snRNAseq data in the macrophage cell cluster. (D) Macrophage GSEA (Hallmark) reveals enrichment in proinflammatory signaling pathways and depletion in respiratory electron transport pathways. (E) Representative coronal sections through the SVZ from control and MIP littermates were analyzed by confocal microscopy for IBA⁺ macrophage/microglia (white). n=3 biological replicates. Bar = 50μm. (F) IBA1⁺ macrophages (white) in representative choroid plexus wholemounts from control (left) and MIP (right) on p6. n=3 biological replicates. Bar = 200μm. (G) At 48 hours, SVZ wholemount preparation from FoxJ1-EGFP mice was imaged in brightfield with region of interest boxed. Representative confocal images of the ventricular surface from control (top panel) and MIP (lower panel) show cilia (acetyl tubulin, white) and ependymal cells (EGFP, green) n=4 biological replicates. Bar = 50μm. (H) Representative SVZ wholemounts from control (n=7) and MIP (n=8) littermates were imaged on a scanning electron microscope. Bar = 500nm. (I) Representative western blot for FoxJ1 in SVZ protein extracts. (J) FoxJ1 protein levels measured in control (n=7) and MIP (n=11) littermates. FoxJ1 bands were normalized to actin. Data presented as means ± SEM. Non-parametric Mann-Whitney testing. *P<0.05.

Figure 4.. Spatial cell transcriptomic analysis 48 hours after MIP induction.

(A) Transcription profiles were used to annotate spatial nodes aligned to auto-segregated cells and visualized in UMAP plot for control (left) and MIP (right) tissue samples. (B) Control (left) and MIP (right) tissue sections were overlayed with color-annotated cells showing alignment of anatomical regions with respect to cortex, corpus callosum, caudate and septum. Dashed box depicts ventricular region of interest shown in (C). (C) Magnification of boxed area in (B). Ependymal layer (green) is continuous in control tissue (left) compared to MIP (right). In control mice (left), macrophage cells (blue) are limited to the choroid plexus (peach). In MIP mice (right), macrophage cells accumulate in the choroid plexus and are seen on the SVZ surface and in the SVZ. (D) GSEA on macrophage cells (CD68⁺Tmem^-) in the choroid plexus and SVZ revealed enrichment in proinflammatory pathways. (E) GSEA on ependymal cells (FoxJ1⁺) revealed enrichment in toll-like receptor (TLR) and inflammatory pathways.

Figure 5.. Long-term disruption of ependymal cells following neonatal MIP induction

(A) Experimental design. FoxJ1 CreER;Rosa26r-TdTomato pups were injected with tamoxifen on p3. Injected pups were randomized into control and MIP groups on p5. Wholemounts of the SVZ were analyzed by confocal microscopy on p35. Alternatively, TdTomato⁺ cells were quantified using flow cytometry on p27. (B) Representative coronal sections immunostained for acetyl-tubulin (green), TdTomato (red), and GFAP (white) from of p25 Control (top) and MIP (bottom). n=3 biological replicates. White bar = 25μm. (C) Representative confocal images of SVZ wholemounts. β-catenin (green) and TdTomato (red) on the ependymal surface of the SVZ in control mice (top, n=4) and MIP mice (n=5). White bar = 25μm. (D) Representative flow cytometry results from control and MIP SVZ single cell suspensions made at p27. (E) Averaged results from TdTomato flow experiments assessing the percentage of TdTomato⁺ cells in control ( n=7) and MIP mice (n=6). Unpaired t-test. **P < 0.01.

Figure 6.. LPS-induced loss of cilia and FoxJ1 expression in primary ependymal cells in vitro.

(A) GSEA analysis of FoxJ1⁺Ak9⁺ ependymal cells reveal enrichment in TNFα-NFκB pathway (Hallmark) and depletion in oxidative phosphorylation pathways. (B) Nuclear translocation of NFκB-linked p65 (red) in ciliated (acetylated tubulin, green) observed at 2 and 6 hours after treatment with LPS (250ng/ml). n=3 biological replicates. Bar=20μm. (C) IL-1β expression following 18hrs of LPS stimulation. n=3 biological replicates. Mean ± SEM. Unpaired t-test. **P < 0.01 (D) TNFα expression following 18hrs of LPS stimulation. n=3 biological replicates. Mean ± SEM. Unpaired t-test. **P < 0.01 (E) LPS-induced loss of FoxJ1 (purple) expression and disruption of β-catenin (white) borders in primary ependymal cells 2 days following LPS stimulation. n=3 biological replicates. Bar=20μm. (F and G) Representative western blot analysis for ependymal cell culture lysates from control and LPS-treated at 2 and 6 days. n=3 biological replicates. (H) qPCR analysis of Foxj1 gene expression at 2 days. n=3 biological replicates. Unpaired t-test. **P < 0.01. (I) qPCR analysis of IFT gene expression at 2 days. n=3 biological replicates. Unpaired t-test. ***P < 0.001 (J) qPCR analysis of CDCC gene expression at 2 days. n=3 biological replicates. Unpaired t-test. ****P < 0.0001. (K) Representative confocal images of ependymal cell cultures immunostained for acetyl tubulin (multicilia, green) at 6 days following LPS stimulation and compared to unstimulated controls. n=4 biological replicates.

Figure 7.. Inflammasome assembly in ependymal cells precede pyroptosis cell death.

(A) 24 hours after MIP induction, the SVZ was assessed for NLRP3 transcripts in vivo by qPCR in control and MIP. n=8 biological replicates. Unpaired t-test **P < 0.01. (B) Representative ependymal protein lysates probed for NLRP3 in control and LPS-treated cells 24 hours after treatment. (C) NLRP3 bands were normalized to actin and quantified for control and LPS-treated n=3 biological replicates. Unpaired t-test **P < 0.01. (D) Representative ependymal protein lysates probed for IL-1β cleavage (p17) in control, LPS-treated, and LPS-treated with 5μM nigericin (45 minute stimulation). (E) Cleaved IL-1β normalized to actin in control, LPS-treated, and LPS-treated with 5μM nigericin relative to control. n=3 biological replicates. One-way ANOVA and Bonferroni’s multiple comparisons test. **P < 0.01, ***P < 0.001. (F) Ependymal cells were treated with media control, LPS+nigericin, and LPS+nigericin+Caspase-1 inhibitor (YVAD-FMK). At 45 minutes, protein lysates were probed for cleaved IL-1β by western blot. Representative blot shown. (G) Levels of cleaved IL-1β (p17) were compared between LPS+nigericin and LPS+nigericin+Caspase-1 inhibitor. n=3 biological replicates. Unpaired t-test **P < 0.01. (H) Ependymal cells were treated with media control, LPS, and LPS+nigericin. At 45 minutes of stimulation, protein lysates were probed for cleaved GSDMD. Representative blot shown. (I) Cleaved GSDMD was quantified and normalized to actin and presented as fold increase over control. n=3 biological replicates. One-way ANOVA and Bonferroni’s multiple comparisons test. *P < 0.05. (J) LDH release in ependymal cell supernatants for control, LPS, and LPS+nigericin following 45 minutes of nigericin stimulation. n=3 biological replicates. One-way ANOVA and Bonferroni’s multiple comparisons test. **P < 0.01, ***P < 0.001. (K) Ependymal cells from ASC-citrine Tg mice were stimulated with LPS and analyzed for inflammasome specks. (L) Representative images for ASC-citrine specks (green), β-catenin (white) and acetylated tubulin (red). n=4 biological replicates. White bar = 25μm. (M)High magnification of boxed region in L shows ASC-citrine (green) specks in multiciliated ependymal cells. White bar = 10μm.