Analysis of Expression Pattern and Genetic Deletion of Netrin5 in the Developing Mouse

Abstract

Boundary cap cells (BCC) are a transient, neural-crest-derived population found at the motor exit point (MEP) and dorsal root entry zone (DREZ) of the embryonic spinal cord. These cells contribute to the central/peripheral nervous system (CNS/PNS) boundary, and in their absence neurons and glia from the CNS migrate into the PNS. We found Netrin5 (Ntn5), a previously unstudied member of the netrin gene family, to be robustly expressed in BCC. We generated Ntn5 knockout mice and examined neurodevelopmental and BCC-related phenotypes. No abnormalities in cranial nerve guidance, dorsal root organization, or sensory projections were found. However, Ntn5 mutant embryos did have ectopic motor neurons (MNs) that migrated out of the ventral horn and into the motor roots. Previous studies have implicated semaphorin6A (Sema6A) in BCC signaling to plexinA2 (PlxnA2)/neuropilin2 (Nrp2) in MNs in restricting MN cell bodies to the ventral horn, particularly in the caudal spinal cord. In Ntn5 mutants, ectopic MNs are likely to be a different population, as more ectopias were found rostrally. Furthermore, ectopic MNs in Ntn5 mutants were not immunoreactive for NRP2. The netrin receptor deleted in colorectal cancer (DCC) is a potential receptor for NTN5 in MNs, as similar ectopic neurons were found in Dcc mutant mice, but not in mice deficient for other netrin receptors. Thus, Ntn5 is a novel netrin family member that is expressed in BCC, functioning to prevent MN migration out of the CNS.

Keywords: axon guidance; chemorepulsion; dorsal root entry zone; motor exit point.

Figure 1

Ntn5 is expressed in Boundary cap cells (BCC). A dendrogram calculated from the primary amino acid sequences of mouse netrins demonstrates that netrin5 is more similar to netrin1 and netrin3 than is netrin4 (A). Ntn5 expression was assayed by in situ hybridization. Using a riboprobe and tyramide signal amplification, expression was detected at the dorsal root entry zone (DREZ) of the spinal cord (B) and adjacent to the trigeminal ganglia (C) at embryonic day 15.5. By the ViewRNA branched DNA probe technique, expression was verified at the DREZ and detected at the motor exit point (MEP, motor neurons (MNs) labeled for ChAT, E). Ntn5-expressing cells at both locations were Krox20-positive BCC (F,G). Expression at the MEP was strongest between E13.5 and E17.5 (K,L) and was barely detectable at P0 (M). Likewise, labeling at the DREZ peaked between E13.5–E17.5 (H,I) and was diminished by P0 (J). In Ntn1^GT/+ embryos immunostained for βGal—a reporter of Ntn1 expression in these mice—there is no colocalization with Krox20-positive BCC (D). TG, trigeminal ganglion; DH, dorsal horn of the spinal cord; VH, ventral horn. Scale bar is 550 μm in (B,C,E), 300 μm in (F,G), and 400 μm in (D,H–M).

Figure 2

BCC develop in the absence of Ntn5. Ntn5 was deleted without disrupting the Sec1 gene on the opposite strand. A cassette encoding a farnesylated YFP and a FRT-flanked neomycin resistance gene was knocked into the Ntn5 locus so that YFP was fused to the 5th base pair of exon 2, while the remainder of exons 2–6 was deleted. Open boxes indicate untranslated regions of exons (A,B). Subsequent breeding to FLPe-expressing mice removed the neomycin cassette (C). Mice were genotyped using primers to exons 3 and 4 to detect wild type alleles and with primers to intron 1 and GFP to detect the mutant allele, (A,C). Homologous recombination was verified by Southern blot analysis using a probe (indicated in B) to detect a HindIII fragment (sites in B, HIII) shifted from 6–4.5 kb (D). Krox20-positive BCC were still present at the DREZ and MEP in Ntn5 mutant embryos, as detected by immunoflouorescence and in situ hybridization (E13.5, E–J), while Ntn5 transcripts detected by ISH were virtually eliminated (G,J). DH, dorsal horn of the spinal cord; MN, motor neurons; DRG, dorsal root ganglia. Scale bar is 50 μm.

Figure 3

Netrin5 keeps MNs cell bodies from migrating out of the CNS. Immunolabeling of E13.5 embryos showed ectopic MN cell bodies in the ventral roots of Ntn5^−/− mutants. Ectopias were not observed in controls (A), but cells positive for Islet 1/2 (B) and HB9 (C) were observed in mutants, indicative of MNs. While some MNs were positive for nrp2, (D,E) ectopic MNs in mutant embyros were not (D,F), suggesting that these MNs are a different population from those known to use a semaphorin6A/nrp2 mechanism. Furthermore, there were more ectopias in the rostral half of the embryo than in the caudal (quantified in G) which is the inverse of what was observed in Nrp2 mutant embryos. CNS glia such as GFAP-positive astrocytes did not enter the ventral root (H,I) indicating this aspect of the CNS/PNS boundary was not lost. NF, neurofilament. Scale bar is 50 μm. *indicates P < 0.05 by student’s t-test.

Figure 4

DCC is required to prevent ectopic motor neuron migration. A series of netrin receptor mutant embryos were analyzed at E13.5 for ectopic MNs in the rostral half of the spinal cord. As in controls (A, n = 3), ectopias were not observed in Dscam (B, n = 3), Unc5C (C, n = 3), or Neogenin (D, n = 2) mutants. There were, however, ectopic MNs in DCC mutants (n = 4) positive both for Islet 1/2 (E) and HB9 (F). Quantified in (G). Scale bar in (F) is 50 μm. *indicates P < 0.05 by Tukey’s post hoc test.

Figure 5

Sensory neurons project through the DREZ without Ntn5. There was no obvious disorganization of dorsal roots in Ntn5^−/− mutants visualized by immunolabeling for neurofilament in cryosections from E13.5 embryos (A,B), or in whole embryo labeling at E11.5 (G,H, arrowheads). Nociceptive fibers positive for CGRP and IB₄ projected into the superficial laminae at P3 in mutants and controls (C,D), and parvalbumin-positive Ia afferents targeted to the motor neuron region normally (E,F, arrows). There were also normal numbers of CGRP- and IB₄-positive sensory neurons in mutant DRGs (I,J). Scale bar is 400 μm in (A−F), 100 microns in insets (C,D), and 100 μm in (I,J).

Figure 6

Cranial nerves develop normally in Ntn5^−/− mutants. To visualize cranial nerves, whole embryos were stained for neurofilament at E11.5. The three branches of the trigeminal nerve were positioned normally in both controls (A) and mutants (B). The three branches are labeled as OP (ophthalmic) MX (maxillary) and MD (mandibular). Likewise, the spinal accessory nerve was not visibly reduced and exhibited normal pathfinding in Ntn5^−/− mutants (C,D). Cranial nerve IV, the trochlear nerve, extended dorsally to cross at the back of the head to innervate the contralateral side in controls (E−G) and mutants (H−J). The oculomotor nerve (III) is also visible in these images. All observations were verified in five mutant and five control embryos.

Figure 7

BCC prevent motor neuron exit through multiple mechanisms. (A) BCC express SEMA6A, which signals through PLXNA2/NRP2 receptors on motor neurons (MN; Bron et al., ; Mauti et al., 2007). NTN5 from BCC prevent ectopic MN migration through putative interactions with DCC, probably in concert with an UNC5 receptor. In the absence of NTN5 (B) a subset of MNs exit the CNS. This is a separate subset from that depending on semaphorin signaling for proper positioning. When BCC are completely ablated (C), more MNs enter the ventral root than in Ntn5 or Sema/Plxn/Nrp mutants alone (Vermeren et al., 2003).