Transcription factor Gbx2 acts cell-nonautonomously to regulate the formation of lineage-restriction boundaries of the thalamus

Abstract

Relatively little is known about the development of the thalamus, especially its differentiation into distinct nuclei. We demonstrate here that Gbx2-expressing cells in mouse diencephalon contribute to the entire thalamic nuclear complex. However, the neuronal precursors for different thalamic nuclei display temporally distinct Gbx2 expression patterns. Gbx2-expressing cells and their descendents form sharp lineage-restriction boundaries delineating the thalamus from the pretectum, epithalamus and prethalamus, revealing multiple compartmental boundaries within the mouse diencephalon. Without Gbx2, cells originating from the thalamus abnormally contribute to the epithalamus and pretectum. This abnormality does not result from an overt defect in patterning or cell-fate specification in Gbx2 mutants. Chimeric and genetic mosaic analysis demonstrate that Gbx2 plays a cell-nonautonomous role in controlling segregation of postmitotic thalamic neurons from the neighboring brain structures that do not express Gbx2. We propose that, within the developing thalamus, the dynamic and differential expression of Gbx2 may be involved in the specific segregation of thalamic neurons, leading to partition of the thalamus into different nuclei.

Fig. 1.

Generation of a Gbx2-CreER-ires-Egfp knock-in allele. (A) Schematic representation of gene targeting strategy. The thick line in the targeting vector represents Gbx2 genomic DNA with an insertion of CreER-ires-Egfp-neo cassette at the 5′ untranslated region of the mouse Gbx2 gene. The selectable marker, neo, is flanked by two Frt sites (red triangles). (B) Southern blot identification of targeted ES cell clones using the 5′ and 3′ probes after digestion of genomic DNA with EcoRV restriction enzyme (RV). (C,D) Dorsal view of E18.5 Gbx2^CreER/+ (C) and Gbx2^CreER/- (D) brains. Arrow indicates the absence of the cerebellum in Gbx2^CreER/- embryos. cb, cerebellum; mb, midbrain. Scale bar: 1.7 mm.

Fig. 2.

Expression of CreER and EGFP recapitulates the endogenous Gbx2 in Gbx2^CreER/+ embryos. (A-C) Expression of Gbx2 (A) and CreER (B) transcripts, and EGFP proteins (C) on adjacent sagittal brain sections of Gbx2^CreER/+ embryos at E12.5. The expression in the thalamus and the medial ganglionic eminence is marked by the asterisk and arrowhead, respectively. The arrow indicates thalamic axonal processes. (D-E″) Double-immunofluorescence analysis of BrdU and GFP on coronal brain sections of Gbx2^CreER/+ embryos at E12.5. The boxed area in D is magnified in E-E″. (F-G″) Confocal images of immunofluorescence of pH3 and GFP on coronal brain sections of Gbx2^CreER/+ embryos at E12.5. The boxed area in F is magnified in G-G″. Note that most of the pH3-positive cells are negative for GFP, whereas two pH3-positive cells that display weak GFP immunoreactivity are marked by the yellow arrowheads. (H-I″) Confocal images of immunofluorescence of GFP and TuJ1 on coronal brain sections of Gbx2^CreER/+ embryos at E11.5. The boxed area in H is magnified in I-I″. ET, epithalamus; mb, midbrain; Ncx, neocortex; PT, pretectum; PTh, prethalamus; TH, thalamus. Scale bars: 1034 μm in A-C; 400 μm in D; 300 μm in F; 600 μm in H; 50 μm in E-E″,G-G″,I-I″.

Fig. 3.

The thalamus is a developmental compartment. (A) Whole-mount in situ hybridization showing that Gbx2 expression demarcates the alar plate of p2 (asterisk) at E10.75. (B) Whole-mount X-gal staining of Gbx2^CreER/+; R26R embryo at E11.5 showing the initial population of Gbx2-expressing cells (asterisk) labeled at E10.5. (C-F) Analysis for β-gal activity in whole-mount (C), horizontal (D), sagittal (E) and coronal (F) sections of Gbx2^CreER/+; R26R embryos at E18.5 after administration of tamoxifen at E10.5. Insets in D-F indicate the sectioning plane; the arrowheads mark the sharp border; the arrow indicates a few marked cells in the nucleus of Darkschewitsch. ET, epithalamus; hb, hindbrain; mb, midbrain; Ncx, neocortex; PT, pretectum; PTh, prethalamus; TH, thalamus. Scale bars: 400 μm in D-F.

Fig. 4.

Loss of Gbx2 disrupts the dorsal and posterior borders of the thalamus. (A-F) After tamoxifen administration at E10.5, X-gal staining of whole-mount (A,B) and sagittal brain sections of Gbx2^CreER/+ and Gbx2^CreER/- embryos (C-F) at different stages, as indicated. (G,H) Nissl analysis of immediate adjacent sections of E and F, respectively. (I-P) After tamoxifen administration at E10.25 (I-L) or E10.5 (M-P), X-gal analysis of coronal brain sections of Gbx2^CreER/+ and Gbx2^CreER/- embryos at E12.5 (I,J) and E18.5 (K-P). The boxed area in M and N is magnified in O and P. (Q,R) Nissl analysis of adjacent sections of O and P, respectively. The arrowheads indicate the lineage-restriction boundaries; the asterisks indicate cells that violate compartment boundaries; the red dashed line indicates the border between the thalamus and lateral habenular nucleus; the arrow indicates marked cells in the nucleus of Darkschewitsch (E) and the lateral habenlar nucleus (O). Note that the HPT is enlarged in Gbx2 mutants (H and L), and that the region under the pial surface contains sparse cells in the mutants (bracket in P and R), probably due to abnormal accumulation of neuritis. ET, epithalamus; HPT, habenular-peduncular tract; mb, midbrain; Ncx, neocortex; PT, pretectum; PTh, prethalamus; TH, thalamus. Scale bars: 265 μm in C,D; 250 μm in E-H; 350 μm in I,J; 450 μm in K-N; 200 μm in O-R.

Fig. 5.

Descendents of Gbx2-expresssing cells at different developmental stages populate distinct thalamic nuclei. (A-I) X-gal staining of coronal sections at the rostral, middle and caudal levels of the thalamus of Gbx2^CreER/+; R26R mice at P15 after administration of tamoxifen at E9.5 (A-C), E10.5 (D-F) and E15.5 (G-I). The asterisk indicates some marked cells outside the medial geniculate nucleus. The arrowheads indicate marked cells that originate from Gbx2-expressing cells in the medial ganglionic eminence, in the caudateputamen and globus pallidus. (J-L) Schematic summary of five classes of thalamic nuclei formed by temporally distinct Gbx2-expressing cells. The nuclei marked by light blue, light green and red represent nuclei formed by the initial wave (between E9.5 and E10.5), second wave (between E10.5 and E11.5) and the final wave (E15.5) of Gbx2-expressing cells, respectively. The nuclei in which Gbx2 expression is maintained to postnatal stages are indicated by blue dots. CP, caudateputamen; GP, globus pallidus; HyT, hypothalamus; MGE, medial ganglionic eminence; SC, superior colliculus. See Table 1 for abbreviations of thalamic nuclei. Scale bar: 400 μm.

Fig. 6.

Loss of Gbx2 does not result in obvious defects in patterning of the diencephalon. (A-H) In situ hybridization assay on coronal sections of Gbx2^CreER/+ (A-D) and Gbx2^CreER/- embryos (E-H) at E12.5 with different markers for the thalamus, prethalamus and ZLI as indicated. (I-N) Analysis of markers as indicated for the pretectum on sagittal sections of Gbx2^CreER/+ (I-K) and Gbx2^CreER/- embryos (L-N) at E13.5. The border between the thalamus and the pretectum is demarcated by the cell-free zone, corresponding to the habenulopeduncular tract. The arrowheads indicate the border between the epithalamus and the thalamus; the arrows mark the ZLI. ET, epithalamus; HPT, habenulopeduncular tract; NCx, neocortex; PT, pretectum; PTh, prethalamus; TH, thalamus. Scale bar: 220 μm in A-H; 200 μm in I-N.

Fig. 7.

Gbx2 plays a cell-nonautonomous role in controlling the thalamic lineage boundaries. (A-D) In situ hybridization analysis of Cdh6 and Efna5 expression on the coronal brain sections of Gbx2^CreER/+ and Gbx2^CreER/- embryos at E14.5. The arrowhead indicates the expression in the thalamus; the asterisk marks the diminished Cdh6 and lost Efna5 expression in the mutant embryos. (E-H) X-gal and Fast Red staining of sagittal brain sections of E16.5 chimeric embryos composed of wild-type (blue) and ES-derived cells (pink) of genotype Gbx2^+/- (E) or Gbx2^-/- (F). The arrow indicates aggregates of Gbx2^-/- cells in the cerebellum; the arrowheads indicate the sharp thalamic borders. (G,H) Magnified view of the thalamus in E and F. Note that the Gbx2^-/- cells, like Gbx2^+/-, extensively intermingle with the host cells in the thalamus. (I) Schematic diagram illustrating the generation of genetic mosaics using CreER-mediated deletion of Gbx2 in the thalamus of Gbx2^CreER/F; R26R embryos. The arrows indicate the primers for PCR analysis to detect the floxed (1.7 kb) and deletion (0.4 kb) alleles of Gbx2. (J) PCR analysis of microdissected thalamus of Gbx2^CreER/F; R26R embryos. Note that the thalamus contains both Gbx2^CreER/F and Gbx2^CreER/- cells after tamoxifen administration, but only Gbx2^CreER/F cells without tamoxifen. (K,L) X-gal staining of coronal brain sections of two Gbx2^CreER/F; R26R embryos with different levels of β-gal activity at E18.5 after tamoxifen administration at E10.5. The arrowheads mark the sharp borders of the fate-mapped Gbx2 lineage. (M-P) In situ hybridization of Efna5 on sagittal brain sections of Gbx2^CreER/+ (M), Gbx2^-/-wild-type chimera (N), Gbx2^CreER/FGbx2^CreER/- mosaic (O), and Gbx2^CreER/- (P) embryos at E16.5. Inset in O shows X-gal staining of a sagittal section of the mosaic embryo. The arrows indicate the restored expression of Efna5 in the chimeric and mosaic embryos; the asterisk indicates the absence of Efna1 expression in the thalamus of Gbx2^-/- embryo. ET, epithalamus; Ncx, neocortex; PT, pretectum; PTh, prethalamus; TH, thalamus; WT, wild type. Scale bars: 400 μm in A-D; 450 μm in E,F; 29 μm in G,H; 380 μm in K; 200 μm in L-O.