Characterization of the Gbx1-/- mouse mutant: a requirement for Gbx1 in normal locomotion and sensorimotor circuit development

Abstract

The Gbx class of homeobox genes encodes DNA binding transcription factors involved in regulation of embryonic central nervous system (CNS) development. Gbx1 is dynamically expressed within spinal neuron progenitor pools and becomes restricted to the dorsal mantle zone by embryonic day (E) 12.5. Here, we provide the first functional analysis of Gbx1. We generated mice containing a conditional Gbx1 allele in which exon 2 that contains the functional homeodomain is flanked with loxP sites (Gbx1(flox)); Cre-mediated recombination of this allele results in a Gbx1 null allele. In contrast to mice homozygous for a loss-of-function allele of Gbx2, mice homozygous for the Gbx1 null allele, Gbx1(-/-), are viable and reproductively competent. However, Gbx1(-/-) mice display a gross locomotive defect that specifically affects hindlimb gait. Analysis of embryos homozygous for the Gbx1 null allele reveals disrupted assembly of the proprioceptive sensorimotor circuit within the spinal cord, and a reduction in ISL1(+) ventral motor neurons. These data suggest a functional requirement for Gbx1 in normal development of the neural networks that contribute to locomotion. The generation of this null allele has enabled us to functionally characterize a novel role for Gbx1 in development of the spinal cord.

Figure 1. Generation of Gbx1 mutant mice.

(A) For the construction of the conditional Gbx1- null, loxP sites were inserted flanking exon 2 and a neomycin expression cassette was inserted 5′ to the flanked region as depicted in the schematic. Deletion of exon 2, thereby rendering the endogenous gene null, is mediated by the mating of a mouse containing the targeted allele with a mouse ubiquitously expressing ß-actin Cre. (B) To identify ES cells that properly underwent homologous recombination, the long and short homology arms were screened by Southern Blot of restriction digested genomic DNA. (C) The presence of the WT and targeted alleles in the progeny derived from targeted ES cells were confirmed by PCR of genomic DNA using specific primer sets. Arrow in (A) represents direction of neo transcription; Arrowheads in (A) represent the location of PCR primers used in (C); Gray box indicates exon 2 (E2), which encodes the Gbx1 homeodomain; Purple box indicates the neo cassette flanked by FRT sites (orange circles); White triangles represent the position of loxP sites.

Figure 2. The functional homeodomain of Gbx1 is deleted in Gbx1^−/− mutants.

(A–F) Whole-mount in situ hybridization for Gbx1 full-length or Gbx1 exon 2 mRNA at embryonic (E) day 9.5. (A–C) Lateral view, dorsal is to the left. (D–F) Dorsal view. (A, D) Gbx1 full-length expression in a Gbx1 WT embryo. Strong Gbx1 expression detected within the anterior hindbrain with a lessening gradient as expression extends caudally. Expression not detected within the otic vesicles. (B and E) Gbx1 exon 2 expression in a Gbx1 WT embryo. Expression of sequence encoding the functional HD of GBX1 recapitulates the pattern detected using the full-length RNA probe. Otic vesicle staining observed within the dorsal view is nonspecific (arrow). (C and F) Gbx1 exon 2 expression in a Gbx1^−/− embryo. No specific staining observed throughout the entire embryo, demonstrating deletion of the GBX1 functional domain. Otic vesicle staining observed is nonspecific (arrows). ov, otic vesicle.

Figure 3. Gbx1^−/− mice display a profound locomotive defect severely affecting hindlimb gait.

Photograph depiction of the locomotive phenotype observed in a Gbx1^−/−3-month-old mouse (A) compared to Gbx1^−/+ age-matched control (B).

Figure 4. Deletion of Gbx1 does not affect the patterning of embryonic developmental markers that direct specification of progenitor cell identity within the neural tube.

Immunostaining of E10.5 spinal cords to detect localization of proteins required for the acquisition of distinct spinal neuron progenitor populations: Pax6 [dI4-pMN] and Pax3 [dI1-dI6] (A–B and G–H); Pax2 [dI4, dI6-v1] and Lhx1/5 [dI2, dI4, dI6-v1] (C–D); En1 [v1] and Foxd3 [dI2, v1] (E–F); HB9 [pMN] and ISL1 [pMN] (I–J). The expression domains, including the dorsal and ventral boundaries of the various genetically distinct populations of cells do not appear perturbed in Gbx1^−/− mutants (panels in right column), when compared to Gbx1^−/+ age matched controls (panels in left column).

Figure 5. Deletion of Gbx1 does not affect the generation of ventral spinal cord motor neurons.

Immunostaining to detect HB9+ (A–B) and ISL1+ (C–D) motor neurons (indicated by white brackets) at E11.5 in lumber spinal cord sections. Specification of Gbx1 null motor neurons appears unaffected (B and D) when compared with normal embryos (A and C). Quantification of HB9+ (E) and ISL1+ (F) ventral motor neurons in Gbx1^−/− embryos at E11.5, reveals no significant differences in the number of immunopositive cells when compared to heterozygous littermate controls. Scale bar represents 100 μm. 20X magnification.

Figure 6. Gbx1^−/− embryos display abnormal projection of proprioceptive afferents and decrease in peripherin+ ventral motor neurons.

Peripherin immunolabeling in lumbar spinal cord sections at E14.5 and E15.5. Arrows indicate proprioceptive afferents extending into the spinal cord. Control mice show normal projection of afferents into the intermediate spinal cord (A) and ventral termination zone (C). Many of the proprioceptive afferents fail to project into the ventral spinal cord of Gbx1^−/− mice (B and D). Gbx1^−/− embryos also show a marked decrease in the expression of peripherin+ ventral motor neurons (brackets) in (B and D) when compared to control mice (A and C). Scale bars represent 100 μm. 10X magnification.

Figure 7. Gbx1^−/− embryos continue to display abnormal projection of proprioceptive afferents at late embryonic and early postnatal stages in development.

Parvalbumin immunolabeling in transverse lumbar spinal cord sections at E17.5 and P5. Brackets indicate the innervation of proprioceptive afferents into the intermediate and ventral spinal cord where they are destined to make synaptic connections with their interneuron or motor neuron targets, respectively. Control mice show normal projection of afferents to their intermediate and ventral termination zones (A and C). Many of the proprioceptive afferents fail to fully project to their ventral termination zones in the spinal cord of Gbx1^−/− mice (B and D), while maintaining their proper termination in the intermediate spinal cord. 10X magnification.

Figure 8. ISL1+ and ISL1+/peripherin+ co-expressing motor neurons are reduced in Gbx1−/− ventral spinal cord.

Immunohistochemical analysis for ISL1 (A–D) and ISL1+/peripherin+ co-expressing cells (E–H) in lumbar spinal cord sections at E14.5 and E15.5. Expression of ISL1+ motor neurons (A and C) and ISL1+/peripherin+ co-expressing cells (E and G) in the ventral spinal cord of control embryos. Gbx1^−/− embryos show a significant reduction in the number of ISL1 ventral motor neurons at E14.5 and E15.5 (B and D) and motor neurons coexpressing ISL1/peripherin (F and H). Scale bars represent 100 μm. 20X magnification.

Figure 9. Quantification of ISL1+ and ISL1+/peripherin+ ventral motor neurons in Gbx1^−/− embryos.

Quantification of ISL1 expressing motor neurons (A) and ISL1/peripherin coexpressing motor neurons (B) in the lumbar ventral spinal cord of E14.5 and E15.5 embryos. Each bar represents the average from 10 sections (n = 4) for null and (n = 4) for heterozygotes; *P<0.0001.

Figure 10. Expression analysis of Sox10 in Gbx1^−/− embryos.

(A–D) Whole-mount in situ hybridization for Sox10 expression at E9.5. (A–B) Lateral view, dorsal is to the left. (C–D) Dorsal view. Sox10 expression detected in the r4/r6 hindbrain neural crest streams (asterisks) and within the dorsal root ganglia in the trunk adjacent to the developing spinal cord, is largely unaffected in a Gbx1 ^−/− mutant embryo (B) compared to a littermate control embryo (A). r, rhombomere.

Figure 11. The number of proprioceptive neurons in the DRG is not affected in Gbx1−/−

mice. The distribution of proprioceptive neurons marked by the expression of TrkC antibody in the DRG of embryos at E15.5 (A and B). The proprioceptive neurons show no difference in the expression between the Gbx1^−/− (B) and control embryos (A). Quantification of TrkC expressing neurons in the DRG (C). Each bar represents the average of six DRGs from (n = 3) heterozygous and (n = 3) for Gbx1^−/− embryos; P<0.8107, non-significant; Scale bar represent 100 μm. 20× magnification.