Mash1 and Math3 are required for development of branchiomotor neurons and maintenance of neural progenitors

Abstract

Basic helix-loop-helix (bHLH) transcription factors are known to play important roles in neuronal determination and differentiation. However, their exact roles in neural development still remain to be determined because of the functional redundancy. Here, we examined the roles of neural bHLH genes Mash1 and Math3 in the development of trigeminal and facial branchiomotor neurons, which derive from rhombomeres 2-4. In Math3-null mutant mice, facial branchiomotor neurons are misspecified, and both trigeminal and facial branchiomotor neurons adopt abnormal migratory pathways. In Mash1;Math3 double-mutant mice, trigeminal and facial branchiomotor neurons are severely reduced in number partly because of increased apoptosis. In addition, neurons with migratory defects are intermingled over the midline from either side of the neural tube. Furthermore, oligodendrocyte progenitors of rhombomere 4 are reduced in number. In the absence of Mash1 and Math3, expression of Notch signaling components is severely downregulated in rhombomere 4 and neural progenitors are not properly maintained, which may lead to intermingling of neurons and a decrease in oligodendrocyte progenitors. These results indicate that Mash1 and Math3 not only promote branchiomotor neuron development but also regulate the subsequent oligodendrocyte development and the cytoarchitecture by maintaining neural progenitors through Notch signaling.

Figure 1.

Expression patterns of Mash1 and Math3 in the developing hindbrain. In situ hybridization was performed for Mash1 (A, C, E, H, K, O, N, R), Math3 (B, D, F, I, L, P, N, R), and Phox2b (G, J, M, Q). A-D, The hindbrain was cut along the roof plate and flat mounted. At E9.5 and E10.5, Mash1 was strongly expressed in the ventral hindbrain where trigeminal and facial branchiomotor neurons arise (A, arrowheads, C). Math3 was also expressed in the ventral hindbrain at E9.5 and E10.5 (B, arrowheads, D). E-R, In situ hybridization was performed with transverse sections through r2 (E-G, K-N) and r4 (H-J, O-R) at E9.5 (E-J) and E10.5 (K-R). In r2 and r4, which give rise to trigeminal and facial branchiomotor neurons, respectively, Mash1 was expressed by subsets of ventricular cells as well as by cells migrating out of the ventricular zone (E, H, between dashed lines in K, O). Mash1 was also expressed in postmitotic inner-ear efferent neurons (O, arrowhead), as reported previously (Tiverson et al., 2003). Math3 was expressed by cells in the ventricular zone as well as in the outer layers of r2 and r4 (F, I, between dashed lines in L, P). Phox2b expression was well overlapped with Mash1 and Math3 expression domains in both r2 and r4 (G, J, M, Q). At least some of the ventricular and migrating cells coexpressed Mash1 and Math3 (N, R, between dashed lines). Scale bar, 100 μm.

Figure 2.

Developmental defects of trigeminal and facial branchiomotor neurons in mutant embryos. A-P, In situ hybridization for Islet1 was performed with the wild-type (A, E, I, M), Mash1(-/-) (B, F, J, N), Math3(-/-) (C, G, K, O), and Mash1(-/-);Math3(-/-) (D, H, L, P) hindbrain. The samples were cut along the roof plate and flat mounted. A-D, At E10.5, trigeminal (A, filled arrowhead) and facial (A, open arrowhead) branchiomotor neurons (Islet1⁺) were born in the ventral region of r2-r4 and started migration in the wild type. Similarly, these neurons were born in Mash1(-/-) (B) and Math3(-/-) (C). In Mash1(-/-); Math3(-/-), fewer neurons were born from r2 (D, arrowhead) than the others (A-C). E-H, At E11.5, the majority of trigeminal branchiomotor neurons migrated to the dorsolateral position of r2 in the wild type and Mash1(-/-) (E, F), although some neurons remained in the ventral region in Mash1(-/-) (F, arrowhead). Many of facial branchiomotor neurons migrated from r4 into r5 and r6 in the wild type (E, open arrowhead) and Mash1(-/-) (F). In Math3(-/-), many neurons still remained in r2 (G, filled arrowhead), resulting in a small nucleus (G, open arrowhead). Only subsets of r4 neurons migrated into r5 (G, arrow). In Mash1(-/-);Math3(-/-), most of the r2 neurons did not migrate (H, filled arrowhead), resulting in a very small trigeminal nucleus (H, open arrowhead). Only subsets of r4 neurons migrated into r5. I-L, At E12.5, most trigeminal branchiomotor neurons finished migration, forming the trigeminal nucleus (nV) in the wild type and Mash1(-/-) (I, J). Many of the facial branchiomotor neurons migrated into the dorsal region of r6, forming the facial nucleus (nVII) in the wild type and Mash1(-/-) (I, J). In contrast, most facial branchiomotor neurons still remained in r4 in Math3(-/-) (K) and Mash1(-/-);Math3(-/-), resulting in a very small facial nucleus (K, L, open arrowhead). In Mash1(-/-);Math3(-/-), r4 neurons were intermingled across the midline (L). M-P, At E13.5, most facial branchiomotor neurons finished migration, forming the facial nucleus (nVII) in the wild type and Mash1(-/-) (M, N). In contrast, in Math3(-/-), some facial branchiomotor neurons migrated within r4, forming an elongated nucleus (O, arrowhead and bracket). In Mash1(-/-);Math3(-/-), still many r4 neurons remained in the ventral r4, resulting in a small nucleus (P, open arrowhead). More neurons were intermingled across the midline (P, filled arrowhead). Q-X, Immunohistochemistry against Islet1 was performed with transverse sections of r2 (Q-T) and r4 (U-X) at E10.5. Islet1⁺ cells were significantly reduced in number in Mash1(-/-); Math3(-/-) r2 (T). Scale bar, 100 μm. WT, Wild type.

Figure 3.

Increased apoptosis in Mash1(-/-);Math3(-/-) hindbrain. TUNEL assay was performed with transverse sections of r2 (A-D, I, K) and r4 (E-H, J, L) at E10.5. More cells died in r2 and r4 of Mash1(-/-);Math3(-/-) (D, H-L). These apoptotic cells expressed Nkx2.2 (I, J) but not MAP2 (K, L), suggesting that double-mutant cells undergo apoptosis before becoming mature neurons (I-L). Apoptotic cells were also slightly increased in number in Mash1(-/-) r2 (B). Scale bar, 100 μm. WT, Wild type.

Figure 4.

Ectopic gene expression in Math3(-/-) hindbrain. In situ hybridization for c-Ret (A, B), Sim1 (C, D), and Ebf-1 (E, F) was performed with flat-mounted hindbrain (A-D) and transverse (E, F) sections of r4 at the indicated time points. A, B, c-Ret expression was rarely detectable in the wild type although ectopically upregulated in the ventral r4 of Math3(-/-). C, D, Sim1 expression was not detected in the wild type but occurred at a high level in the ventral r4 of Math3(-/-). Thus, branchiomotor neurons of r4 were misspecified in Math3(-/-). E, F, Ebf-1, which is required for migration of branchiomotor neurons, was not significantly affected in Math3 mutant embryos, suggesting that Ebf-1 is not involved in the migratory defects of Math3 mutant facial branchiomotor neurons. Scale bar: (in F) E, F, 100 μm. WT, Wild type.

Figure 5.

Intermingling of r4 neurons across the midline in Mash1(-/-);Math3(-/-). Transverse sections of the ventral r4 were examined. A, B, B′, In the wild type, the inner surface of the hindbrain was covered by MAP2-negative ventricular cells, and neurons (MAP2⁺) were not present in the midline region at E13.5 (A). In contrast, many neurons were intermingled across the midline in Mash1(-/-);Math3(-/-) (B). The boxed region in B is enlarged in B′. Intermingled neurons were exposed to the lumen of the hindbrain. C, D, The adherens junction molecule N-cadherin was highly expressed at the apical side in the wild type (C), whereas it was missing in the double mutants (D, arrowheads). E, F, BrdU uptake was examined at E10.5. BrdU uptake was reduced in the double mutants (F) compared with the wild type (E). G, H, Ki-67 expression was examined by immunohistochemistry at E10.5. Ki-67 expression was reduced in the double mutants (H) compared with the wild type (G). I-L, In situ hybridization of Shh. Shh was expressed in both the wild-type and double-mutant floor plate at E10.5 and E13.5. Scale bars: (in A) A, B, 400 μm; B′, 100 μm; E-L, 100 μm. WT, Wild type; DKO, double knock-out.

Figure 6.

Impairment of Notch signaling in the ventral r4 of Mash1(-/-);Math3(-/-). In situ hybridization for Hes5 (A-H), Dll1 (I-P), and Dll3 (Q-T) was performed with transverse sections of the ventral r4 at E10.5 (A-D, I-L, Q-T) and E11.5 (E-H, M-P). In Mash1(-/-);Math3(-/-), expression of Hes5 (D, H), Dll1 (L, P), and Dll3 (T) was significantly downregulated in the ventral r4, which gives rise to facial branchiomotor neurons (between dashed lines). In Mash1(-/-), expression of Hes5 (B, F), Dll1 (J, N), and Dll3 (R) was slightly downregulated compared with the wild type and Math3(-/-). Scale bar, 100 μm. WT, Wild type.

Figure 7.

The cells that should normally express Mash1 and Math3 remained in Mash1(-/-); Math3(-/-). In situ hybridization was performed with transverse sections of the ventral r4 at E11.5. A, Math3 was expressed in a salt-and-pepper pattern in the ventral r4 of the wild type (arrow). B, Expression of the antisense neomycin mRNA under the control of the Math3 promoter was detected in the ventral r4 of Mash1(-/-); Math3(-/-) (between dashed lines). C, Mash1 was expressed in the ventral r4 of the wild type. D, Expression of the sense neomycin mRNA under the control of the Mash1 promoter was detected in the ventral r4 of Mash1(-/-); Math3(-/-) (between dashed lines). Thus, the cells that should normally express Mash1/Math3 remained in the double mutants. Cells adjacent to the floor plate were decreased in the double mutants (compare arrowheads in A, B). Scale bar, 100 μm. WT, Wild type.

Figure 8.

Impairment of oligodendrocyte development in the ventral r4 of Mash1(-/-); Math3(-/-). In situ hybridization for PDGFRα was performed at E12.5. PDGFRα expression was significantly downregulated in the ventral r4 of Mash1(-/-); Math3(-/-) (D) compared with the others (A-C). Thus, oligodendrocyte progenitors were significantly reduced in number in the double mutants. WT, Wild type.