Deletions in CWH43 cause idiopathic normal pressure hydrocephalus

Abstract

Idiopathic normal pressure hydrocephalus (iNPH) is a neurological disorder that occurs in about 1% of individuals over age 60 and is characterized by enlarged cerebral ventricles, gait difficulty, incontinence, and cognitive decline. The cause and pathophysiology of iNPH are largely unknown. We performed whole exome sequencing of DNA obtained from 53 unrelated iNPH patients. Two recurrent heterozygous loss of function deletions in CWH43 were observed in 15% of iNPH patients and were significantly enriched 6.6-fold and 2.7-fold, respectively, when compared to the general population. Cwh43 modifies the lipid anchor of glycosylphosphatidylinositol-anchored proteins. Mice heterozygous for CWH43 deletion appeared grossly normal but displayed hydrocephalus, gait and balance abnormalities, decreased numbers of ependymal cilia, and decreased localization of glycosylphosphatidylinositol-anchored proteins to the apical surfaces of choroid plexus and ependymal cells. Our findings provide novel mechanistic insights into the origins of iNPH and demonstrate that it represents a distinct disease entity.

Keywords: CWH43; GPI-anchored protein; hydrocephalus; normal pressure hydrocephalus.

Figure 1. iNPH‐associated deletions disrupt the ability of Cwh43 to regulate the localization of GPI‐anchored receptors

1. Sanger DNA sequencing data from two iNPH patients confirming the presence of heterozygous CWH43 deletions. Arrows identify location of the deletion.

2. Diagram of the domain structure of the Cwh43 protein illustrating the location of the damaging CWH43 deletions. Lower panel details the disruptive effect of the k696Asnfs mutation on the C‐terminal endoplasmic reticulum export signal (YF), replacing it via a frameshift with a 23 amino acid sequence that contains an endoplasmic reticulum retention signal (KKKS).

3. Fluorescence micrographs of HeLa cells transfected with plasmids encoding human GFP‐Cwh43, GFP‐Cwh43‐k696fs, or GFP‐Cwh43‐Leu533ter fusion proteins. To label the ER, HeLa cells were transduced with a baculovirus encoding a fluorescently labeled RFP‐calreticulin fusion protein containing an ER retention signal (RFP‐calreticulin‐KDEL). To label the Golgi apparatus, cells were immunostained for golgin‐97 (red). When compared with wild‐type GFP‐Cwh43, GFP‐Cwh43‐k696fs, and GFP‐Cwh43‐Leu533 showed decreased association with the Golgi apparatus (but not the ER) and were diffusely distributed throughout the cytoplasm. Scale bar is approximately 5 µm.

4. Western blot analysis of total membrane, aqueous and lipid (GPI‐anchor‐containing) Triton X‐114 extracts derived from wild‐type HeLa cells, and two independent CRISPR CWH43 knockout (KO) HeLa cell lines in which a mutated CWH43 gene encodes a protein that is truncated near Leu533 and CWH43 mRNA and protein are markedly reduced. Cells were transfected to overexpress a control GFP plasmid, a plasmid encoding human wild‐type Cwh43 with GFP fused to the N‐terminus, or a plasmid encoding human CWH43 harboring the iNPH‐associated mutation (Lys696AsnfsTer23) with GFP fused to the N‐terminus. The Western blot was stained using an antibody directed against CD59, a GPI‐anchored protein.

Source data are available online for this figure.

Figure 2. Expression of Cwh43 mRNA and protein in the mouse brain

1. mRNA in situ hybridization images showing expression of Cwh43 mRNA in the mouse brain. Enclosed areas containing potions of the ventricle, hippocampus, and dorsal thalamus are shown at higher magnification on the right. Arrowheads point to choroid plexus. Scale bar is approximately 400 µm.

2. Fluorescence immunohistochemistry of the ependymal surface of the lateral ventricle of the mouse brain. Cilia are visualized using an antibody for acetylated alpha tubulin (green). Cwh43 is visualized using a specific anti‐Cwh43 antibody (red). Nuclei are counterstained using DAPI (blue). Arrowheads point to motile cilia and scale bar is approximately 5 µm.

3. Confocal fluorescence immunocytochemistry images of a single cultured mouse ciliated ependymal cell. Cilia were visualized using an antibody for acetylated alpha tubulin (green). Cwh43 immunoreactivity was visualized using a specific anti‐Cwh43 antibody (red). Scale bar (left column) is approximately 4 µm. Images in the column on the right represent a Z‐stack reconstruction of confocal images showing localization of Cwh43 immunoreactivity in cilia of a mouse ependymal cell. Scale bar (right column) is approximately 5 µm.

Figure 3. CWH43 mutation increases ventricular volume and causes gait dysfunction in mice

1. Scatter plot comparing ventricular volume from CWH43^WT/WT (wild‐type, WT), heterozygous CWH43^WT/M533, homozygous CWH43^M533/M533, and CWH43^M533/A530 mice at 6 months. Ventricular volume was calculated from T2‐weighted MR images of the mouse brain using a custom automated computer algorithm. Horizontal bars indicate the mean of the measurements in each column. Statistical significance for each mutant mouse line compared to WT was determined using the unpaired t‐test (P = 0.0015 for CWH43^WT/M533; P = 0.0014 for CWH43^M533/M533; P = 0.006 for CWH43^M533/A530).

2. Representative 3D volumetric MR images of mouse brains from 6‐month‐old CWH43^WT/WT, CWH43^M533/M533, and CWH43^M533/A530 mice. LV = lateral ventricle, 3^rd V = third ventricle, 4^th V = fourth ventricle. Scale bar is approximately 1 mm.

3. Quantitative gait analysis at 7 months of age revealed increased sway among homozygous CWH43^M533/M533 mice, (*P = 0.03, n = 5) and compound heterozygous CWH43^M533/A530 (*P = 0.034, n = 4) mice when compared to wild‐type CWH43^WT/WT mice (n = 5). Sway (the distance between the hind paws during walking) was measured repeatedly for individual mice in each group during a constrained unidirectional walk. Data shown are the mean ± SD for each group. Statistical significance was determined using the unpaired t‐test.

4. Box plot showing rotarod performance data for CWH43^WT/WT, CWH43^WT/M533, CWH43^M533/M533, and CWH43^M533/A530 mice at 7 months of age. Data shown are the mean, 1^st quartile, 3^rd quartile, minimum, and maximum for each group of mice. Statistical significance was determined using the unpaired t‐test. When compared to wild‐type CWH43^WT/WT mice (n = 10), balance time on the rotarod was decreased significantly among heterozygous CWH43^WT/M533 mice (*P = 0.04, n = 8), homozygous CWH43^M533/M533 mice (*P = 0.03, n = 7), and compound heterozygous CWH43^M533/A530 mice (*P = 0.03, n = 4).

Source data are available online for this figure.

Figure 4. CWH43 mutation decreases cilia number and alters the distribution of GPI‐anchored proteins in mouse ventricular ciliated epithelia

1. Scanning electron micrographs of the ependymal surface of the lateral ventricle of CWH43 WT and CWH43^M533/M533 mice. Scale bar is approximately 40 µm. The graph on the right quantifies the data from the electron micrographs on the left. Data shown are the mean ± SEM. *P = 0.004, n = 4, unpaired t‐test.

2. Western blot analysis of total membrane, aqueous, and lipid (GPI‐anchor‐containing) Triton X‐114 extracts derived from wild‐type, CWH43^M533/M533, and CWH43^WT/M533 mouse brain or kidney. The Western blot was stained using an antibody directed against CD59, a GPI‐anchored protein.

3. Fluorescence immunohistochemistry for CD59 in the ependymal layer and choroid plexus of the lateral ventricle from CWH43^WT/WT, CWH43^M533/M533, and CWH43^M533/A530 mice. Arrowheads point to apical surfaces of ependymal and choroid plexus cells. Nuclei are counterstained using DAPI (blue). Scale bar is approximately 5 µm.

Source data are available online for this figure.