COL4A1 mutations cause ocular dysgenesis, neuronal localization defects, and myopathy in mice and Walker-Warburg syndrome in humans

Abstract

Muscle-eye-brain disease (MEB) and Walker Warburg Syndrome (WWS) belong to a spectrum of autosomal recessive diseases characterized by ocular dysgenesis, neuronal migration defects, and congenital muscular dystrophy. Until now, the pathophysiology of MEB/WWS has been attributed to alteration in dystroglycan post-translational modification. Here, we provide evidence that mutations in a gene coding for a major basement membrane protein, collagen IV alpha 1 (COL4A1), are a novel cause of MEB/WWS. Using a combination of histological, molecular, and biochemical approaches, we show that heterozygous Col4a1 mutant mice have ocular dysgenesis, neuronal localization defects, and myopathy characteristic of MEB/WWS. Importantly, we identified putative heterozygous mutations in COL4A1 in two MEB/WWS patients. Both mutations occur within conserved amino acids of the triple-helix-forming domain of the protein, and at least one mutation interferes with secretion of the mutant proteins, resulting instead in intracellular accumulation. Expression and posttranslational modification of dystroglycan is unaltered in Col4a1 mutant mice indicating that COL4A1 mutations represent a distinct pathogenic mechanism underlying MEB/WWS. These findings implicate a novel gene and a novel mechanism in the etiology of MEB/WWS and expand the clinical spectrum of COL4A1-associated disorders.

Figure 1. Mislocalization of retinal ganglion cells in Col4a1^+/Δex40 eyes.

Representative position-matched transverse sections from Col4a1^+/+ (A–D) and Col4a1^+/Δex40 (E–L) eyes at E18 labeled with anti-ISL1 (green in A, E, and I), anti-laminin (red in B, F and J) and DAPI (blue in C, G, and K). Anti-ISL1 labeling revealed an increased number of RGCs in the outer retinas (arrows in D, H and L) of the mutant eyes. Labeling with anti-laminin antibody revealed focal disruptions in the inner limiting membrane in Col4a1^+/Δex40 mice (asterisks in F) and demonstrates that the retinal vasculature is adjacent to the inner limiting membrane in Col4a1^+/+ eyes but not in Col4a1^+/Δex40 eyes (arrowheads in B and F). Notably, in one extreme case, we detected no hyaloid vasculature and markedly reduced retinal thickness in the eye shown in I–L. Scale bar: 50 µm.

Figure 2. Increased apoptosis of Islet-1 positive retinal ganglion cells in Col4a1^+/Δex40 eyes.

Representative eye sections from E18 Col4a1^+/+ mice (A–C) and Col4a1^+/Δex40 mice (D–F) co-labeled with anti-ISL1 (green in A and D) and anti-activated Caspase-3 (red in B and E). Co-labeling for ISL1 and Caspase-3 revealed an increase in the number of apoptotic retinal ganglion cells in Col4a1^+/Δex40 eyes. Col4a1^+/Δex40 eyes were significantly smaller than Col4a1^+/+ eyes at all ages examined (H; E14 p = 0.01; E16 p = 0.04; E18 p<10⁻⁵ comparing at least 6 eyes for Col4a1^+/+ (blue) and Col4a1^+/Δex40 (red) at each age). After correcting for size, there was a significant increase in the number of apoptotic retinal ganglion cells in Col4a1^+/Δex40 eyes at E18 (G; p = 0.02 comparing at least 6 eyes for Col4a1^+/+ (blue) and Col4a1^+/Δex40 (red) at each age). In G and H, data are presented as mean +/− SEM. Scale bar: 50 µm.

Figure 3. Col4a1^+/Δex40 mice have focal and variable cortical malformations.

Representative cresyl violet stained coronal brain sections from adult Col4a1^+/+ (A) and Col4a1^+/Δex40 (B–H) revealed cortical malformations characteristic of cobblestone lissencephaly in all mutant brains examined (n = 6). In contrast to the well-defined cortical lamination observed in Col4a1^+/+ brains, Col4a1^+/Δex40 brains had variable abnormalities including disorganized cortical lamination (bracket in B), ectopias (C, enlarged in D), heterotopic regions (E and F), enlarged ventricles (G) and, occasionally, severe cortical malformations (H). Scale bars: A–C and E–H, 800 µm; d, 200 µm.

Figure 4. Col4a1^+/Δex40 mice display cortical neuronal localization defects.

Representative images of NeuN-labeled coronal brain sections from adult Col4a1^+/+ (A and C) and Col4a1^+/Δex40 (B and D–F) mice revealed heterotopic regions (asterisks in B and E) and ectopias (arrows in B and D) in all mutant brains examined. NeuN labeling of ectopic cells located in the marginal zone (F, enlarged from box in E) confirmed their neuronal identity. Scale bars: A–E, 500 µm; F, 125 µm.

Figure 5. Focal and variable cortical neuronal localization defects are developmental.

Embryos pulsed-labeled with BrdU at E14 were harvested at birth and BrdU was immunolabeled in position-matched coronal brain sections of Col4a1^+/+ (A and F) and Col4a1^+/Δex40 mice (B–E and G–J). In contrast to Col4a1^+/+ controls, in which BrdU-labeled cells are distributed in a uniform layer in the outer cortex, all Col4a1^+/Δex40 brains had focal neuronal localization defects that ranged in severity from diffuse cell layers to ectopias and severely disorganized cortical lamination (note gradient of severity from B through E and from G through J). Similar results were observed in embryos labeled at E16. Scale bar: 500 µm.

Figure 6. Col4a1^+/Δex40 mice have focal breaches of the pial basement membrane during development.

Representative images of laminin labeled (green) coronal brain sections from P0 Col4a1^+/+ (A–D) and Col4a1^+/Δex40 (E–L) revealed breaches of the pial basement membrane in mutant animals. All sections of Col4a1^+/+ brains examined (A–D) showed intact pial membranes (higher magnifications in D). In contrast, position-matched Col4a1^+/Δex40 brains display multiple focal disruptions of the pial basement membrane (arrows and asterisks in G and K) that are shown in higher magnification in H, and L. Notably, the high magnification image in L shows ectopic cells (DAPI in blue) that have breached the pial membrane adjacent to a disruption. Scale bars: A–C, E–G, and I–K, 500 µm; D, H and L, 50 µm.

Figure 7. Col4a1^+/Δex40 mice have functional, biochemical, histological, and genetically–complex myopathy.

(A) Comparison of the peak grip force (mean +/−SEM) between Col4a1^+/+ (blue; n = 7) and Col4a1^+/Δex40 (red; n = 7) mice at 3 months of age revealed a significant reduction in Col4a1^+/Δex40 mice (** indicates p<0.01). (B) Serum creatine kinase (CK) activity was compared (mean +/−SEM) between Col4a1^+/+ (blue; n = 7) and Col4a1^+/Δex40 (red; n = 7) mice before (Pre) and after (Post) exercise (* indicates p<0.05 vs all other groups). (C–E) Comparison of non-peripheral nuclei (arrows in D and E; mean +/−SEM in C) between Col4a1^+/+ (counting of 17830 fibers in 5 muscles from 4 mice) and Col4a1^+/Δex40 (counting of 13820 fibers in 4 muscles from 4 mice) mice (14 to 16 months old) revealed a significant increase in the number of non-peripheral nuclei in Col4a1^+/Δex40 mice on the C57BL/6J background (control mean = 1.5%; mutant mean = 5.2%; p<0.0005). This difference is modified by the genetic context. There was no difference between Col4a1^+/+ and Col4a1^+/Δex40 CASTB6F1 mice and there was a significant rescue of non-peripheral nuclei in CAST/B6F1 Col4a1^+/Δex40 mice compared to C57BL/6J Col4a1^+/Δex40 mice (p<0.005). Scale bar: 100 µm.

Figure 8. Identification of COL4A1 missense mutations in two human patients.

Direct sequence analysis of patient DNA revealed two heterozygous missense mutations (A, top panels) that were not found in control patients (A, bottom panels). (B) Protein sequence alignments with multiple species indicate a very strong degree of conservation of the altered amino acid for both mutations. Methionine at position 1016 is conserved in all species tested except for C. elegans and Drosophila where this amino acid is a lysine. Glutamine at position 1316 is also highly conserved and is never a glutamate residue. (C) Triple helix thermal stability was calculated by amino acid sequence and is plotted for the length of the COL4A1 protein. Arrows indicate the positions for each mutation and the effects of the mutations on calculated thermal stability are indicated in red. (D) For quantitative analysis of the ratio of secreted to intracellular mutant COL4A1 protein levels, values are expressed as percentage of the wild-type COL4A1 ratio and are presented as mean +/− SEM. Representative Western blot images for secreted and intracellular COL4A1 are shown below the graph. Nine independent Western blot experiments were performed using 6 independent clones per mutation for this analysis, **P<0.01.

Figure 9. The pathogenic mechanism by which Col4a1 mutation causes MEB/WWS phenotypes is independent of dystroglycan glycosylation.

(A) Western blot analysis of total protein lysates from quadriceps biopsy using antibody for dystroglycan precursor protein (bottom) did not reveal a difference in dystroglycan precursor expression between Col4a1^+/+ and Col4a1^+/Δex40 mice. To test for differences in quantity or mobility (glycosylation) of α-dystroglycan and β-dystroglycan, lysates were enriched for glycosylated proteins by WGA binding. WGA-enriched fraction of quadriceps protein extracts did not reveal differences in quantity or glycosylation of either α-dystroglycan (top) or β-dystroglycan (middle) between Col4a1^+/+ and Col4a1^+/Δex40 mice. (B) Immunohistochemical labeling of muscle sections with antibodies against the glycosylated form of α–dystroglycan (red) and β-dystroglycan (green) show similar patterns of α– and β-dystroglycan expression between Col4a1^+/+and Col4a1^+/Δex40 mice.