Transgenic studies on homeobox genes in nervous system development: spina bifida in Isl1 transgenic mice

Abstract

To develop in vivo assays for homeobox gene function in neural development, we generated transgenic mice in which the expression of a homeobox gene is altered only within the nervous system, in neurons or neuronal precursor cells. Transgenic expression of Hoxc8 did not result in gross abnormalities, while a Hoxd4 transgene caused death shortly after birth. In neural progenitor cells, the motorneuron-specific homeodomain transcription factor Isl1 induced early developmental defects, including absence of anterior neural structures, profound defects in the neuroepithelium and defective neural tube closure. A fraction of Isl1 transgenic mice exhibited spina bifida. Isl1 transgene expression was also associated with decreased proliferation and increased Pbx1 expression in the ventral neural tube. Our results suggest a function for some homeobox genes in development of the nervous system, and that cell-type- and region-specific transgenic models will be useful to identify the cellular and molecular targets of homeobox transcription factors in nervous system development.

Fig. 1

VP16-driven binary transgenic system for activation of homeobox transgenes. a Schematic overview. The system consists of two parental mouse strains: transactivator mice (TA) and transresponder mice (TR). The transactivator carries the Herpes Simplex Virus VP16 gene under control of a cell-type-specific enhancer/promoter. The transresponder strain carries the transgene of interest linked to a VP16-responsive viral immediate early (IE) gene promoter. This promoter is silent in mice in the absence of VP16. When the parental strains are crossed, double transgenic progeny will inherit both the TA and TR transgene (TA + TR). In these mice, VP16 transactivates the transresponder gene and the transgene of interest (GENE) becomes expressed in cells where the promoter/enhancer of choice is active. b Double transgenic embryo that expressed VP16 under control of the Hoxc8 promoter, and therefore displays VP16-mediated activation of expression of the IE-LacZ transresponder gene in the region where the Hoxc8 promoter is normally active at E10.5. c Section from a paraffin-embedded stained embryo shows β-galactosidase activity in mesodermal derivatives, such as the somites, and the neuroepithelium (Gardner et al. 1996)

Fig. 2

Constructs for transgenes. a The enhancer from the rat nestin gene (Yaworsky and Kappen 1999) was linked to the neurofilament light chain gene promoter (Yaworsky et al. 1997) upstream of the VP16 coding region, an SV40 intron and a polyadenylation site. b, c The rat nestin gene enhancer was ligated in both directions (b and c, respectively) to the minimal promoter region (96 bp) from the neurofilament gene upstream of the same VP16 construct (Yueh et al. 2000) as in a. d The rat nestin enhancer downstream of the VP16 construct in orientation opposite to the minimal NF promoter. e The promoter from the Hoxc8 gene (Gardner et al. 1996) was ligated upstream to the Isl-1 cDNA (Muller et al. 2003), followed by the same SV40/polyA fragment as in a

Fig. 3

Neuron-specific transgene activation by neuron-specific transactivators. Embryos were isolated at E10.5 and genotyped by PCR on genomic DNA isolated from their associated yolk sacs, respectively. Staining for β-galactosidase activity was performed on whole mount specimens. TA-negative embryos did not exhibit any staining in the neural tube; staining of the apical ectodermal ridge is an inherent property of the IE-LacZ strain (Gardner and Kappen 2000). a, b Progeny from a cross of Nestin-VP16 transactivators crossed to IE-LacZ transresponders. Staining with X-gal reveals β-galactosidase expression in the developing nervous system of double transgenic (TA + TR) progeny from transgenic founder animals #73 (a) and #2 (b). c Histological section through an embryo derived from founder #2 demonstrating specific localization of β-galactosidase activity in neuroepithelial cells, predominantly in ventricular zone cells. d Progeny from a cross of an NFL-VP16 transactivator to an IE-LacZ transresponder mouse. e Histological section demonstrating neuron-specific localization of β-galactosidase activity

Fig. 4

Nestin-VP16-driven Isl1 transresponder activation. Nestin-VP16 transactivators were crossed to IE-Isl1 transresponder mice, embryos were isolated at E10.5 and genotyped by PCR on yolk sac DNA. a Schematic of an E10.5 embryo for the purpose of size comparison. Note the size bar to the right, representing the length of the embryo in b. b Double transgenic (TA + TR) embryo was much smaller than normal size and lacked head development. c Additional independent double transgenic embryo was also much smaller and anteriorly malformed. d Third double transgenic embryo exhibited substantially retarded development. e Histological section through a fourth double transgenic embryo was stained with Hematoxylin/Eosin and reveals abnormal folding of the neuroepithelium, which is also thinner than normal. Staining of adjacent sections with antibodies specific for Proliferating Cell Nuclear Antigen (PCNA) and nestin (not shown) indicated that the epithelial cells are proliferating neural precursor cells. Magnifications are not equivalent

Fig. 5

Hoxc8-VP16-driven Isl1 transresponder activation. Hoxc8-VP16 transactivators were crossed to IE-Isl1 transresponder mice, in two types of crosses: TA/+ × TR/+, or × TA/+ TR/+. Pregnancies were allowed to go to term, and tail size was determined on neonates, with confirmation of assessment on day 3 after birth, or on the carcass. Genotyping was performed on DNA from the tail of weaning age progeny, or from the carcass on the day of discovery. Bars represent the number of live offspring (green), and of dead neonates (red). TA-only containing newborns had all normal tail length, one pup homozygous for the TA locus was recovered dead. Neonates inheriting both a TA and TR locus exhibited shortened tails, and some deaths. More offspring with shorter tails were found with higher TA transgene dosage, and more deaths, correlated to tail length; all tailless offspring were recovered dead. No abnormalities or deaths were found in progeny that inherited only the TR locus from the first type of cross

Fig. 6

Spina bifida in Isl1-transgenic mice. a Neonate from a cross of Hoxc8-VP16 transactivator to IE-Isl1 transresponder with shortened tail and open back. b Close-up of the open site, with splaying of the neural arches. c Scanning electron micrograph of an E11.5 Isl1-transgenic embryo with open neural tube in the caudal region. d Independent Isl1-transgenic embryo with spina bifida just posterior to the hind limbs

Fig. 7

Staining with PCNA-specific antibody. Sections were cut from paraffin-embedded control and Isl1-transgenic embryos and processed together in the same incubation chamber. a Control embryo at E11.5; section at the hind limb level. Areas for cell counting are marked by a rectangle, and magnified pictures are provided for: dorsal neural tube, ventral neural tube and somite. b Control embryo at E12.5; section behind the hind limb level. c Isl-1 transgenic littermate at E11.5; level comparable to A. d Isl1-transgenic embryo at E12.5; level comparable to B, with spina bifida. e, f Quantification of results in three independent embryos for each area, genotype, and time point, respectively, plotted as % of cells with PCNA positive nuclei (average ± standard deviations). Open bars control embryos; filled bars Isl1-transgenic embryos; **P <0.005. e Quantification of staining for E11.5 embryos in somite, ventral neural tube and dorsal neural tube; f Quantification of staining for E12.5 embryos in somite, ventral neural tube and dorsal neural tube

Fig. 8

Pbx1a as a potential target gene of Islet-1 in motorneurons. Sequence A67 was identified in a differential display screen between wildtype (WT) and Isl1-transgenic embryos. a A67 expression was higher in Isl1-transgenic embryos; incorporation of ³²P into A67 PCR product was normalized to incorporation of ³²P-ATP into PCR product from β-actin mRNA. N = 4 independent embryos for each group. b Immunohistochemistry for Islet-1 demonstrates expression in motorneurons and dorsal root ganglia on a frozen section from a normal E10.5 mouse embryo. c An in situ hybridization probe for A67 detects expression in motorneurons in a normal embryo at E10.5 (the section is from a slightly more posterior level than in panel b. Some expression was also detected in mesoderm at the hindlimb level, an area known to express Isl1 mRNA (Kappen and Salbaum 2009). Sequencing revealed that A67 was the Pbx1a isoform derived from the Pbx1 gene