Togaram1 is expressed in the neural tube and its absence causes neural tube closure defects

Abstract

Neural tube closure defect pathomechanisms in human embryonic development are poorly understood. Here we identified spina bifida patients expressing novel variants of the TOGARAM gene family. TOGARAM1 has been associated with the ciliopathy Joubert syndrome, but its connection to spina bifida and role in neural development is unknown. We show that Togaram1 is expressed in the neural tube and Togaram1 knockout mice have abnormal cilia, reduced sonic hedgehog (Shh) signaling, abnormal neural tube patterning, and display neural tube closure defects. Neural stem cells from Togaram1 knockout embryos showed reduced cilia and defects in Shh signaling. Overexpression in IMCD3 and HEK293 cells of TOGARAM1 carrying the variant found in the spina bifida patient resulted in cilia defect along with reduced pericentriolar material one (PCM1), a critical constituent of centriolar satellites involved in transporting proteins toward the centrosome and primary cilia. Our results demonstrate the role of TOGARAM1 in regulating Shh signaling during early neural development that is critical for neural tube closure and elucidates potential mechanisms whereby the ciliopathy-associated gene TOGARAM1 gives rise to spina bifida aperta in humans.

Keywords: PCM1; TOGARAM1; cilia; sonic hedgehog signaling; spina bifida.

Graphical abstract

Figure 1

Index patient with TOGARAM1 variant has spina bifida (A) Pedigree of the index patient with spina bifida aperta (SBA). (B) Electropherogram depicting heterozygous TOGARAM1 gene missense variant (NM_001308120: c.A4141T) highlighted by red dashed line box. (C–F) Phenotype of the patient. (C) Postoperative MRI recording demonstrating myelomeningocele had been located at the lower thoracic level (T10) (yellow arrowhead). (D, E) Sagittal T2-TSE MRI recordings showed agenesis/dysgenesis of the corpus callosum (black arrowheads) with slim anterior commissure and cerebellar tonsils extending into the foramen magnum (yellow arrowhead). (F) Chest X-ray illustrated a situs inversus thoracalis, showing that the heart was on the right side (yellow arrowhead). (G) Patient variant in transcript variant 1 (NM_001308120.2) of human TOGARAM1 is shown (red arrow). (H) Protein schematic of TOGARAM1 with patient variant indicated. It has four conserved TOG domains and localization of the amino acid sequence affected by the mutation (NP_001295049:p.N1381Y) in the TOG3 domain (red arrow), HEAT repeats are shown light purple. Amino acid and TOG domains are identified from Latour et al.

Figure 2

The expression panel of Togaram1 mRNA in developing mouse embryos by WISH (A, D, G, J, M, P) Whole mount views and transverse sections at different levels are shown in Togaram1 WISH mouse embryos. Lateral views and transverse sections at hindbrain level (green arrowhead) are shown in E9.5 (A, D), E10.5 (G, J), and E11.5 (M, P) mouse embryos. (B, E, H, K, N, Q) Dorsal views and transverse sections at thoracic spine level (yellow arrowhead) are shown in E9.5 (B, E) and E10.5 (H, K), and at upper level in E11.5 (N, Q) mouse embryos. (C, F, I, L, O, R) Frontal views and transverse sections are shown at lumbar spine level (black arrow) in E9.5 (C, F), E10.5 (I, L), and E11.5 (O, R) mouse embryos. Strong expression of Togaram1 was detected in telencephalon (red arrowhead) and diencephalon (blue arrowhead) from E9.5 to E11.5 (A, G, M). Low expression of Togaram1 in the brain was first observed in the mesencephalon (black arrowhead) and metencephalon (green arrowhead) at E9.5 (A) that was getting stronger at E10.5 (G) and E11.5 (M). The yellow arrowhead and black arrow indicate the posterior extent of Togaram1 expression along the spinal cord at E9.5 (B, C, E, F), showing that the strong expression was progressively restricted to more anterior levels at E10.5 (H, K, I, L) and E11.5 (N, Q, O, R). n ≥ 3 of each genotype per age. Scale bar, 1 mm.

Figure 3

Neural tube defect and disrupted Shh mRNA expression in Togaram1^−/− embryos (A–F) E11.5 WT and Togaram1KO mouse phenotypes are shown. Compared to WT embryos (A and D), Togaram1KO embryos showed exencephaly (white arrowhead in B and C). Some embryos displayed open neural tube defect restricted to a part of the spine with wavy spine (white arrowhead in E) or without wavy spine (F), or in the end of the tail (C). hin = hindbrain. (G–J) WISH for Shh in E10.5 and E11.5 mouse embryos. In E10.5 (G) and E11.5 (I) WT embryos, Shh expression was observed in the basal plate of midbrain and forebrain (outlined by dashed red line) and ZLI (yellow arrowhead), Shh was also present in the spinal neural tube (green arrowhead) in the E11.5 WT embryo. Shh mRNA was highly reduced in the brain and in the spinal neural tube in E10.5 (H) and E11.5 Togaram1KO mouse (J). Shh expressed in the ZLI originates from basal plate in WT (G, I) and Togaram1KO (H). There was no ZLI in the other Togaram1KO embryo (J) (yellow arrowhead). n ≥ 3 of each genotype per age. Scale bar, 1 mm.

Figure 4

E9.5 Togaram1KO mouse exhibit a dorsalized neural tube Neural tube transverse sections (hindbrain, thoracic, and hindlimb level) from E9.5 WT and Togaram1KO embryos stained for designated neural tube markers (n ≥ 3 embryos for each genotype). (A and B) Shh (red) was observed in the notochord (yellow arrow) and the floor plate (white arrow) in the WT; however, Shh was only present in the notochord (yellow arrow) in the Togaram1KOs. (C and D) FoxA2 (white arrow) expressed in the floor plate in the WT but not in the Togaram1KOs. (E and F) Nkx2.2 (red) was missing and Pax6 (blue) domain expanded ventrally in the Togaram1KOs at all levels. Olig2 (green) was missing at the hindbrain level while expanding ventrally below the hindbrain level. (G and H) Nkx6.1 (purple) became more restricted to the ventricular zone at the hindbrain level but expanded dorsally at the hindlimb level in the Togaram1KOs. (I and J) Pax7 (cyan) expressed in the most dorsal part of neural tube in the WT which expanded ventrally in the Togaram1KOs at the hindbrain level, while restriction in the dorsal region was observed at the hindlimb level. Spinal cord is outlined by dashed lines. n ≥ 3 of each genotype. Scale bar, 10 μm.

Figure 5

Togaram1 regulates Gli1 and proteolytic processing of GLI3FL (A) Analysis of Gli1 quantitative real-time PCR results from WT and Togaram1KO embryo heads. Gli1 was significantly decreased in Togaram1KO compared with WT embryos at mRNA basal level (p = 0.04 using unpaired t test, replicates data from three embryos of each genotype). (B) The basal level of Gli1 in Togaram1KO NSCs was significantly lower than in WT NSCs (p = 0.0045). After 24 h of treatment with 100 nM SAG, Gli1 was both upregulated in WT and Togaram1KO NSCs compared with DMSO control (WT DMSO vs. WT SAG, p = 0.03; Togaram1KO DMSO vs. Togaram1KO SAG, p = 0.03, p value of comparison between WT SAG with Togaram1KO SAG is 0.0045). Data were pooled from three independent experiments. p values from two-way ANOVA corrected with Tukey’s multiple comparisons test. ∗∗p < 0.01, ∗p < 0.05. (C and E) Western blot and corresponding quantification of GLI3FL and GLI3R in E10.5 WT and Togaram1KO embryo heads normalized to ACTIN (quantificated individually in more than three embryos per genotype). The ratio of GLI3FL:GLI3R levels were significantly increased in the Togaram1KO compared with the WT (WT vs. Togaram1KO, p = 0.0097 by unpaired t test). (D) Western blot for GLI3 expression in WT and Togaram1KO NSCs stimulated with 100 nM SAG. (F) The basal level of GLI3FL:GLI3R was also significantly higher in Togaram1KO NSCs (WT DMSO vs. Togaram1KO DMSO, p = 0.0052). There was a trend in the ratio of GLI3FL:GLI3R to increase in WT NSCs in the presence of SAG. Compared with DMSO condition, Togaram1KO NSCs showed no significant difference after SAG treatment (Togaram1KO DMSO vs. Togaram1KO SAG, p = 0.66). Statistics was performed by two-way ANOVA test with Tukey’s multiple comparisons test. (G) GLI3R basal level was significantly reduced in Togaram1KO NSCs than in WT NSCs (WT vs. Togaram1KO, p = 0.0069 by unpaired t test). n > 2 independent experiments. ∗∗p < 0.01; ∗p < 0.05; ns, not significant.

Figure 6

Primary cilia defect in Togaram1KO embryo and NSCs (A) Primary cilia extending from pericentrin were seen in nearly all WT NSCs, while very few cilia were seen in Togaram1KO NSCs (white arrow). (B and C) Quantification of ciliation percentage in NSCs (WT n = 926 versus Togaram1KO n = 1270 cell number counted, p < 0.0001 by Welch’s corrected unpaired t test). Cilia length was significantly shorter in Togaram1KO than WT NSCs (WT n = 734 and Togaram1KO n = 97 cilia, p < 0.0001 by Mann-Whitney test). n > 2 independent experiments were pooled. Data information: Each dot represents one field (B, E), and each point represents one cilium (C, F). Cilia marker Arl13b (red), centrosome marker pericentrin (green). All cilia measurements were based on Arl13b staining. Scale bar, 10 μm. ∗∗∗∗p < 0.0001; ∗p < 0.05. (D) Immunostaining of transverse sections at the posterior limb level of E11.5 WT and Togaram1KO embryos are shown. Primary cilium labeled by Arl13b was nearly absent in the Togaram1KO compared with the WT. A higher magnification is displayed in the insets. Intermediate mesoderm area (white arrow) was taken for cilia quantification. (E) Ciliation percentage was quantified for embryos (WT n = 791 and Togaram1KO n = 1245 cell number counted, p = 0.0101 by unpaired t test). (F) Cilia length was strongly reduced in Togaram1KO embryos (WT n = 641 and Togaram1KO n = 467 cilia, p < 0.0001 by Mann-Whitney test), three embryos of each genotype were used for embryo ciliary analysis.

Figure 7

Cilia length defect in IMCD3 cells expressing human p.Asn1381Tyr variant (A and B) Electron microscopy picture of cilium from nasal mucous of patient showed absence of inner dynein arms (white arrows) (B) vs. healthy control (A). Inset is the schematic of the microtubule doublet. (C) Cilia percentage quantification showed no difference (n = 561 for GFP-wtTOGARAM1 vs. n = 319 for GFP-sbTOGARAM1, p = 0.7 by unpaired t test), each dot represents one field. (D) Cilia length was reduced in GFP-sbTOGARAM1 IMCD3 cells (GFP- wtTOGARAM1 n = 136 vs. GFP-sbTOGARAM1 n = 104 cilia, p < 0.0001 by Mann-Whitney test), each data point represents one cilium, cilia analysis from two pooled independent experiments. (E) IMCD3 cell transfected with GFP tagged wtTOGARAM1 and sbTOGARAM1 were stained for cilia (Arl13b, red), centrosome (pericentrin, white), and nuclear (DAPI, blue) after serum starvation for 48 h. Cilia was colocalized with GFP-TOGARAM1 in both GFP-wtTOGARAM1 and GFP-sbTOGARAM1 IMCD3 cells (white arrow). Scale bar, 10 μm. ∗∗∗∗p < 0.0001; ns, not significant.

Figure 8

Expressing p.Asn1381Tyr variant in HEK cells disrupts PCM1 (A) No difference in the α-tubulin colocalization with GFP-wtTOGARAM1 and GFP-sbTOGARAM1 (wt-TOGARAM1 n = 33 vs. sb-TOGARAM1 n = 26 cells analyzed, p = 0.96 by unpaired t test) in TOGARAM1 transfected HEK cells. (B and C) Representative high lattice expression and α-tubulin localization in GFP-wtTOGARAM1 and GFP-sbTOGARAM1 cells. A cell has a localized GFP but no PCM1 and disorganized tubulin lattice (thick yellow arrow), another cell with a dispersed GFP puncta has PCM1 and disorganized tubulin structure (thin yellow arrow). (D) Percentage of GFP-TOGARAM1 localized to PCM1 in transfected HEK cells (wtTOGARAM1 n = 64 vs. sbTOGARAM1 n = 41 GFP⁺ cells analyzed, p = 0.1 by Mann-Whitney test). In the boxplot, 25th and 75th percentiles are depicted, midlines represent the median and whiskers depict range. (E) PCM1 area quantitation results (wt-TOGARAM1 n = 50 vs. sb-TOGARAM1 n = 26, and RFP n = 13 GFP⁺ cells analyzed, p = 0.002 between wtTOGARAM1 and sbTOGARAM1, p > 0.99 between wt and RFP by Kruskal-Wallis test corrected by Dunn’s test). (F) Integrated intensity of PCM1 (wt-TOGARAM1 n = 50 vs. sb-TOGARAM1 n = 26, p = 0.04 by Mann-Whitney test) showed significantly higher in GFP-wtTOGARAM1 compared with GFP-sbTOGARAM1. (G and H) Representative GFP-wtTOGARAM1 cell with low lattice expression that had GFP with PCM1 and pericentrin colocalized (thick arrow). (H) While in a representative GFP-sbTOGARAM1 cell, there was no PCM1 and pericentrin around the bright GFP spot (thick arrow). Scale bar, 10 μm, n = 2 independent experiments. ∗∗p < 0.01; ∗p < 0.05; ns, not significant.

Figure 9

The disruption of PCM1 was partially rescued in Togaram1KO MEFs transfected with wtTOGARAM1 (A) Representative wild-type MEF cells have smaller nuclei and localized punctate Cdk5rap2 and PCM1 staining that overlapped. (B) One representative Togaram1KO MEF cell in the center has non-localized PCM1 and Cdk5rap2. (C) One representative Togaram1KO MEF cell transfected with the GFP-wtTOGARAM1 construct has more punctate localized distribution of both PCM1 and Cdk5rap2 staining. (D) One representative Togaram1KO MEF cell transfected with GFP- sbTOGARAM1 has no PCM1 and Cdk5rap2 staining. Scale bar, 10 μM.