Abnormal neurogenesis and cortical growth in congenital heart disease

Abstract

Long-term neurological deficits due to immature cortical development are emerging as a major challenge in congenital heart disease (CHD). However, cellular mechanisms underlying dysregulation of perinatal corticogenesis in CHD remain elusive. The subventricular zone (SVZ) represents the largest postnatal niche of neural stem/progenitor cells (NSPCs). We show that the piglet SVZ resembles its human counterpart and displays robust postnatal neurogenesis. We present evidence that SVZ NSPCs migrate to the frontal cortex and differentiate into interneurons in a region-specific manner. Hypoxic exposure of the gyrencephalic piglet brain recapitulates CHD-induced impaired cortical development. Hypoxia reduces proliferation and neurogenesis in the SVZ, which is accompanied by reduced cortical growth. We demonstrate a similar reduction in neuroblasts within the SVZ of human infants born with CHD. Our findings demonstrate that SVZ NSPCs contribute to perinatal corticogenesis and suggest that restoration of SVZ NSPCs' neurogenic potential is a candidate therapeutic target for improving cortical growth in CHD.

Fig. 1. The piglet SVZ is structurally similar to its human counterpart and displays robust postnatal neurogenesis

(A) 3D model of the pig brain at 7 weeks (w) of age. (B–E) Photographic images on the coronal plane of pig brain tissue at 7 weeks of age. The SVZ is divided into four rostrocaudal domains: (B) anterior end (AE), located at the rostral extremity of the anterior horn; (C) anterior horn (A); (D) middle (M), spanning the body of the ventricle; and (E) posterior (P), near the atrium of the lateral ventricle (LV). (F) Magnified image from the location marked by inset in (C). DAPI stain; L-, lateral; DL-, dorsolateral; scale bar, 1 mm. (G-I) GFAP expression in the four distinct layers of L-SVZ: (I) ependymal layer; (II) Dcx⁺ neuroblasts; (III) GFAP⁺ astrocyte ribbon; (IV) transitional zone. Magnified images from the location marked by inset in (F). (G) GFAP expression; (H) GFAP expression and DAPI stain; (I) GFAP, DAPI, and Dcx immunostains; scale bars, 50 μm. (J-M) Immunostains at 7 weeks of age. (J) Magnified image from the location marked by inset in €. Vimentin (Vim) immunostain; I-, inner; O-, outer; WM, white matter; scale bar, 100 μm. (K-M) Magnified images from the location marked by the inset in (J) showing: (K) Vim and Pax6 expression; (L) Vim and Sox2 expression; (M) Vim and Ki67 expression; scale bars, 50 μm. Quantification of (N) Sox2⁺ and (O) Sox2⁺Ki67⁺ cells in the A- and P-SVZ. Quantification of (P) Sox2⁺ and (Q) Sox2⁺Ki67⁺ cells in I- and O-SVZ (n=6 animals). Quantification of (R) Sox2⁺ and (S) Sox2⁺Ki67⁺ cells at 1w, 7w, and 15w (n=5). (T) Neurospheres isolated from the A- and P-SVZ at p2 and p14 and cultured for 1w; scale bars, 100 μm. (U) Number of neurospheres generated (n=3 animals, p2; n=5–6 animals, p14) and (V) distribution of neurosphere diameter. Data expressed as the mean±SEM. *p < 0.05, **p < 0.01, ***p < 0.001, unpaired Student’s t-test (N-Q); ^#p < 0.05, analysis of variance (ANOVA) with Bonferroni post-hoc test (R,S,U); ordinal logistic with likelihood ratio test (LRT) (V).

Fig. 2. Diverse patterns of neuroblast cell populations are present throughout the developing piglet brain

(A) MRI of piglet brain at 1 week (w) of age. (B-C) Magnified images of insets corresponding to (A). Immunofluorescence analysis of Dcx and PSA-NCAM expression in the (B) cortex and (C) subventricular zone (SVZ). Dcx⁺PSA-NCAM⁺ cells extend (magnified in inset) from the SVZ (C) to the cortex (B,C); scale bars, 50 μm (B), 200 μm (C). (D) 3D reconstruction of a tiled immunostain for Dcx in 3 mm thick, optically cleared brain tissue encompassing the SVZ and prefrontal cortex; curved dashed line marks the boundaries of the surface of the lateral ventricle (LV) and the SVZ. Reconstruction is a magnification of the inset denoted in (A). (E-H) 3D segmentation of Dcx⁺ cells within select regions of interest demarcated in panel D illustrating multiple patterns of neuroblast cell distribution (E) adjacent to the LV, (F) within and near the SVZ, (G) in the subcortical white matter (WM), and (H) in the cortex (ctx); individual and groups of cells are assigned random colors. Note: segmentation is throughout the entire depth (2.4 mm after optical clearing) of the tissue and therefore not all segments are visible in the 3D reconstruction shown in D, particularly segments from panel F near the SVZ where Dcx⁺ cells are highly dense.

Fig. 3. Focally labeled SVZ-derived cells are primarily of the neuronal lineage in piglets

(A) Schematic of CTG injection (green) into the A-DL-SVZ of the 1 week old piglet brain. (B) Schematic corresponding to the AE-SVZ noted in (A) denoting regions of interest identified in higher magnification in (D,J,L,M). CTG⁺ cells in the (C) A-SVZ (region noted in (A) and (D) AE-SVZ (region noted in A), 1 week post injection (wpi); scale bars, 1 mm (C), 500 μm (D), 50 μm (inset in C). (E) Higher magnification of CTG⁺Dcx⁺ in the AE-SVZ denoted in (D), 1 wpi; scale bar, 50 μm. (F) Quantification of the number of CTG⁺ cells near the injection site at 2 weeks of age (n = 3 animals). (G-I) Quantification of the cell fate of CTG⁺ cells in the (G) DL-Med-SVZ, (H) DL-Mid-SVZ, and (I) L-SVZ (n = 3 animals); GFAP represents astrocytes and radial glia, Olig2 represents oligodendrocytes, and Iba1 represents microglia. CTG⁺Dcx⁺ cells in the (J) rostral migratory stream, (K) AE-SVZ and periventricular white matter (PVWM), 1 wpi. (L) CTG⁺ cells within the subcortical white matter (SCWM) and (M) CTG⁺NeuN⁺ neurons in the prefrontal cortex, 1wpi; scale bars, 50 μm (E,J-M). DAPI counterstain (blue) is shown in (C,E,J,K,L,M). Panels on the right in (K) and (M) identify single channels and merge images for insets. CTG, cell tracker green; A, anterior; AE, anterior end. LV, lateral ventricle; DL, dorsolateral; L, lateral; Med, medial; Mid, middle; CC, corpus callosum. Data expressed as the mean±SEM. *p<0.05, **p<0.01 compared with CC-Mid, CC-L, PVWM, and Caudate (F) by one-way ANOVA with Bonferroni post-hoc comparisons. ***p<0.0001, compared with GFAP, Olig2, and Iba1 (G-I) by 1-way ANOVA with Bonferroni post-hoc comparisons. ^#p<0.05, GFAP compared with Dcx, Olig2, and Iba1 (H) by one-way ANOVA with Bonferroni post-hoc comparisons.

Fig. 4. SVZ-derived cells migrate to the frontal cortex and differentiate into neurons in piglets

(A) T2* weighted MRI scans displaying hypointense voxels illustrating the presence of SPIO⁺ nanoparticles (magenta arrowheads in insets), 4 weeks post injection (wpi) in the piglet brain. Cortices are color coded as follow: blue, prefrontal; pink, primary (P) somatosensory; green, insular. (B) Quantification of the distribution of hypointense voxels generated by SPIO nanoparticles as measured by T2* weighted MRI, 4 wpi (n=3); Ctx, cortex. (C) Quantification of SPIO⁺ co-labeled cells within the upper (UL) and lower (LL) layers, and subcortical white matter (WM), of the cortices where the majority of hypointense voxels were located in (B), 4wpi (n=3). (D-G) Examples of SPIO⁺ cells in the cortex, 4 wpi. (D) SPIO⁺Calr⁺NeuN⁺ interneurons and (E) SPIO⁺Calr⁻NeuN⁺ neurons in the upper layers (II/III) of the prefrontal cortex. (F) SPIO⁺GFAP⁻Iba⁻ cells in layer I of the prefrontal cortex. (G) SPIO⁺Calr⁺NeuN⁺ interneurons in layers (II/III) of the primary somatosensory cortex. (D-G) DAPI⁺ nuclei are seen in the blue channel; dashed white lines separates layer I from II/III of the cortex; insets represent higher magnification and different combinations of channels to better visualize nanoparticles within specific cell types; Scale bars, 100 μm. (H) Quantification of the distribution of SPIO⁺ co-labeled cells throughout the prefrontal cortex, where the majority of the nanoparticles were found in (B). (I) Quantification of the cell fate of co-labeled cells in the upper layers of the prefrontal cortex, where the majority of the nanoparticles were located (H). (J,K) Neurosphere differentiation potential, isolated from p14 SVZ (n=6, anterior; n=5 animals/group, posterior); scale bars, 100 μm; Tuj1 (Neuron-specific class III beta-tubulin) labels neurons, O4 labels oligodendrocytes. (L) Quantification of Dcx⁺ cells at 1, 7, and 15w (n=4 each). Data expressed as the mean±SEM. ***p < 0.001, prefrontal ctx and p somatosensory cortex vs. all other regions (B); **p<0.01, UL vs. LL and SCWM (H); **p<0.01, Calr⁺NeuN⁺ vs. Calr⁻NeuN⁺ and GFAP⁺ (I); **p<0.01, anterior Tuj1 vs. all markers in anterior and posterior (J); one-way ANOVA with Bonferroni post-hoc comparisons (B, H-K). More data are presented in table S2.

Fig. 5. Chronic hypoxia impairs cortical development and reduces neurogenesis in the SVZ in piglets

(A,B) Reconstruction of MRI scans at p14. (C) Quantification of average brain weight (n=15 animals/group). (D,E) Cortical gray matter volume (green). (F) Quantification of cortical volume (n=4–5 animals/group) in (D,E). (G,H) Perimeter of the total pial surface (red lines, inner) and total perimeter of the brain (yellow lines, outer); scale bars, 1 cm (D,E,G,H). (I) Quantification of cortical folding determined by the gyrification index (n=5 animals/group) expressed as a ratio of the inner vs. outer perimeter traces in (G,H). (J,K) DAPI stains of A-SVZ at p14; scale bars, 500 μm. (L) Quantification of the average width/thickness of the A-SVZ (n=4 animals/group). (M,N) Immunostains of GFAP. (O) Quantification of the average length of GFAP⁺ processes. (P,Q) Immunostains of Sox2⁺ cells. (R) Quantification of the average number of Sox2⁺ cells. (S,T) Immunostains of Sox2⁺Ki67⁺ cells. (R) Quantification of the average number of Sox2⁺Ki67⁺ cells. (V,W) Immunostains of Dcx⁺ cells. (X) Quantification of the average number of Dcx⁺ cells. Scale bars, 50 μm; n=3–4 animals/group (M,N,P,Q,S,T,V,W). LV, lateral ventricle; Nx, normoxia; Hx, hypoxia. Data expressed as the mean ± SEM. *p<0.05, **p < 0.0001, Hx vs. Nx, unpaired Student’s t-test (C,F,I,L,O,R,U,X).

Fig. 6. Structural and cellular alterations/abnormalities of the SVZ are associated with CHD in humans

(A,B) GFAP⁺ cells within the SVZ underlying the frontal cortex of human infants or neonates born (A) without or (B) with severe or complex CHD; insets represent areas where images were acquired at higher magnification, corresponding to C,D,F,G; dashed lines illustrate the boundary between the lateral ventricle and SVZ; scale bars, 1 mm. (C-H) Immunostains illuminating GFAP⁺ cell processes spanning layer II of the (C-E) DL- and (F-H) L-SVZ; insets in D,G represent higher magnification shown in E,H. (C,F) without CHD and (D,E,G,H) with severe or complex CHD; scale bars, 50 μm (C,D,F,G), 20 μm (E), and 10 μm (H). Quantification of the average length of GFAP⁺ processes within (I) the SVZ, (J) the DL-SVZ, and (K) the L-SVZ (n=5 with no CHD and n=4 with CHD). (L,M) Immunostains illustrating Dcx⁺ cells within the SVZ for a human infant (L) without and (M) with CHD; insets represent areas where images were acquired at higher magnification corresponding to N-S; dashed lines illustrate the boundary between the lateral ventricle and SVZ; Scale bars, 1 mm. Images in panels P and S were acquired in the same regions marked by the insets in M from a different CHD specimen. Dcx⁺ cells within the DL- and L-SVZ of a human infant born (N,Q) without CHD and two different infants born (O,P,R,S) with severe or complex CHD; scale bars, 50 μm. Quantification of the average number of Dcx⁺ cells within (T) the SVZ, (U) the DL-SVZ, and (V) the L-SVZ (n=4 with no CHD and 3 with CHD). CHD, congenital heart disease; LV, lateral ventricle; DL, dorsolateral; L, Lateral; N, normal. Data are expressed as the mean±SEM. *p < 0.05, **p < 0.01, No CHD vs. CHD, unpaired Student’s t-test (I,T). Two-way ANOVA with Bonferroni comparisons (I,T). Pearson correlation (U).

Fig. 7. Chronic hypoxia impedes the contribution of SVZ-derived neurons to cortical development in piglets

(A,B) CTG⁺ cells within the AE-SVZ; insets represent higher magnification shown in (C,D); (A,C) normoxia, (B,D) hypoxia. Scale bars, 500 μm. A Schematic of CTG injection (green) into the A-DL-SVZ is shown at the top of (B). Dashed lines in D outline the SVZ. (E, F, H, I) CTG⁺Dcx⁺ and (G,J) CTG⁺Dcx⁺PSA-NCAM⁺ cells, originating from the A-SVZ, within the AE-SVZ. Quantification of the number of (K) CTG⁺ and (L) CTG⁺Dcx⁺ cells within the AE SVZ at p14, 11 days after hypoxic exposure (n=3 animals/group). Insets show higher magnification labeled cells in (E-J). (M) Quantification of the regional distribution of hypointense voxels from SPIO⁺ labeled cells present in the cortex (n=6 animals/group), expressed as a percentage of the total number of hypointense voxels. LV, lateral ventricle; Nx, normoxia; Hx, hypoxia; dpi, days post injection; Ctx, cortex. Data expressed as the mean ± SEM. *p<0.05, **p < 0.01, unpaired Student’s t-test.

Fig. 8. Chronic hypoxia alters immature and mature neuronal populations in a region-specific manner in piglets

Immunostains for (A) Dcx, (B) Casp3, and (C) Dcx/Casp3 within layers II/III of the prefrontal cortex at p14, normoxia (Nx); scale bars, 50 μm. (D) Quantification of Dcx⁺ cell density within the upper layers (UL) of distinct cortices. Immunostains for (E) Dcx, (F) Casp3, and (G) Dcx/Casp3 within layers II/III of the prefrontal cortex at p14, hypoxia (Hx); scale bars, 50 μm. (H) Quantification of Dcx⁺ cell density within the subcortical white matter at p14. Immunostains of NeuN within the prefrontal cortex under (I) normoxia and (J) after hypoxia at p14; dashed lines border layers II/III of the cortex; scale bar, 500 μm. Quantification of (K) the number of NeuN⁺ cells, (L) layer volume, (M) number of Calr⁺ cells, and (N) Tbr1⁺NeuN⁺ cell density within the upper layers of the cortex, p14. Quantification of (O) layer volume, (P) the number of NeuN⁺ cells, and (Q) the number of Calr⁺ in the lower layers (LL) of the cortex, p14. WM, white mater; Ctx, cortex; PF, prefrontal cortex; SS, primary somatosensory cortex; I, insular cortex; C, cingulate cortex; P, prepyriform area; PV, periventricular white matter. Data expressed as the mean±SEM. n=6 animals per group. *p < 0.05, **p < 0.01, ***p<0.001, Hx vs. Nx, unpaired Student’s t-test (D,H, K-Q). More data are presented in Supplemental Tables 4,5.