Musashi and Plasticity of Xenopus and Axolotl Spinal Cord Ependymal Cells

Abstract

The differentiated state of spinal cord ependymal cells in regeneration-competent amphibians varies between a constitutively active state in what is essentially a developing organism, the tadpole of the frog Xenopus laevis, and a quiescent, activatable state in a slowly growing adult salamander Ambystoma mexicanum, the Axolotl. Ependymal cells are epithelial in intact spinal cord of all vertebrates. After transection, body region ependymal epithelium in both Xenopus and the Axolotl disorganizes for regenerative outgrowth (gap replacement). Injury-reactive ependymal cells serve as a stem/progenitor cell population in regeneration and reconstruct the central canal. Expression patterns of mRNA and protein for the stem/progenitor cell-maintenance Notch signaling pathway mRNA-binding protein Musashi (msi) change with life stage and regeneration competence. Msi-1 is missing (immunohistochemistry), or at very low levels (polymerase chain reaction, PCR), in both intact regeneration-competent adult Axolotl cord and intact non-regeneration-competent Xenopus tadpole (Nieuwkoop and Faber stage 62+, NF 62+). The critical correlation for successful regeneration is msi-1 expression/upregulation after injury in the ependymal outgrowth and stump-region ependymal cells. msi-1 and msi-2 isoforms were cloned for the Axolotl as well as previously unknown isoforms of Xenopus msi-2. Intact Xenopus spinal cord ependymal cells show a loss of msi-1 expression between regeneration-competent (NF 50-53) and non-regenerating stages (NF 62+) and in post-metamorphosis froglets, while msi-2 displays a lower molecular weight isoform in non-regenerating cord. In the Axolotl, embryos and juveniles maintain Msi-1 expression in the intact cord. In the adult Axolotl, Msi-1 is absent, but upregulates after injury. Msi-2 levels are more variable among Axolotl life stages: rising between late tailbud embryos and juveniles and decreasing in adult cord. Cultures of regeneration-competent Xenopus tadpole cord and injury-responsive adult Axolotl cord ependymal cells showed an identical growth factor response. Epidermal growth factor (EGF) maintains mesenchymal outgrowth in vitro, the cells are proliferative and maintain msi-1 expression. Non-regeneration competent Xenopus ependymal cells, NF 62+, failed to attach or grow well in EGF+ medium. Ependymal Msi-1 expression in vivo and in vitro is a strong indicator of regeneration competence in the amphibian spinal cord.

Keywords: Axolotl regeneration; Xenopus regeneration; ependymal cells; musashi-1; musashi-2; spinal cord regeneration.

FIGURE 1

Cartoon representing ependymal outgrowth from cranial (Left) and caudal (Right) stumps of regenerating Xenopus and Axolotl spinal cord. (A) Regenerating NF 50–53 Xenopus tadpole cord showing gap regeneration with ciliated epithelial ependymal cells in the stump and the bulb-like ependymal outgrowth. (B) Regenerating adult Axolotl gap regeneration with mesenchymal ependymal outgrowth and several layers (bracket) of epithelial ependymal cells in the stump. The glia limitans, the basal lamina of the cord, is represented as a red line adjacent to ependymal endfeet. All other cell types and extracellular matrix were omitted to simplify the representation.

FIGURE 2

Msi-1 alignment, deduced amino acid sequences. A multispecies alignment including both Xenopus and Axolotl is shown. The RNA recognition motifs (RRMs), also known as mRNA-binding domains, are indicated by the boxed areas labeled RRM1 or RRM2. The RRMs are located near the amino terminus of the protein. A 16 amino acid linking region between RRM1 and RRM2 is identical in Xenopus, Axolotl, mouse, and human. Amphibian and mammalian RRM sequences all have greater than 90% homology at the amino acid level, details in the Section “Results.”

FIGURE 3

Msi-2 multispecies alignment, deduced amino acid sequences including Xenopus and Axolotl results. The RNA recognition motifs (RRMs) are indicated by the boxed areas labeled RRM1 or RRM2. Subcloning of Axolotl Msi-2 transcripts showed the presence of a shortened isoform ending with a truncated RRM2 as well as the long and short Xenopus Msi-2 isoforms described in Figure 4. For Xenopus, the original Msi-2 sequence is labeled xrp-1 as originally published. The C-terminal end is not shown for some species.

FIGURE 4

(A) The nucleotide and deduced amino acid sequences are shown for segments of two isoforms of Xenopus Msi-2 detected by RT-PCR. These isoforms were detected using one set of forward and reverse primers. The arrows under the nucleotide sequence show the position of forward (5′-CTCTATGGAACTGCCAGCCAGGA-3′) and reverse (5′-GATGCCCAAGCAAGACTGGGAG-3′) primers sequences. The longer isoform (A1) yielded a 203 bp band, while the shorter isoform band (A2) had 145 bp. The splicing insert (underlined DNA sequence) in the long isoform (A1) occurred between the greater-than (>>>>>) and the less-than signs (<<<<<). In the shorter isoform (A2), the region underlined in A1 was spliced out and the regions designated between the greater-than the less-than signs, respectively, were directly joined. The bracketed numbers above some amino acids indicate the differing region, and correspond to schematic (A). The asterisks in A1 and A2 show the terminal codons. (B) The schematic for the deduced amino acid sequence region illustrates 100% homology from amino acid 1 to 393, between the longer and shorter band isoforms, with the alternatively spliced region from 394 to 406. The full length amino acid sequence of both isoforms has the same 406 amino acid length. RRM1 and RRM2 indicate, respectively, RNA-recognition motif-1 and -2.

FIGURE 5

Xenopus Msi-1 and Msi-2 semi-quantitative RT-PCR gels show strong expression of Msi-1 in intact, regeneration-competent NF 50 spinal cord (gel column 1), and 5-day regenerating NF 50 spinal cord (Reg. 5D NF 50 spinal cord, gel column 3). Both Msi-1 and Msi-2 were reduced from NF 50 levels in intact NF 62 (non-regeneration-competent) and adult spinal cord (compare gel columns 1, 2). Msi-2 specific primers showed bands representing the two isoforms shown in the Figure 4 alignments (arrowheads). The larger isoform was universally present. The smaller isoform (lower arrowhead) was present only in NF 62 and adult frog (compare gel columns 2, 4). GAPDH (glyceraldehyde-3-phosphate dehydrogenase) was used as the housekeeping control gene. The No-RT control (–RT) was performed without reverse transcriptase using the GAPDH primers. Each gene is shown in a separate row. All bands in each row are from one original gel. Negative was shown as for image clarity.

FIGURE 6

Axolotl Msi semi-quantitative RT-PCR gel. Msi-1 was strongly expressed in the pre-hatching/late tailbud stage embryo (BD 42–44; column 1), juvenile brain and spinal cord (gel columns 2, 3), but was absent in intact adult spinal cord (column 4). At 3 weeks post-transection in the adult (3 weeks regenerating adult cord) Msi-1was up-regulated (column 5). Msi-2 expression was stronger in the juvenile brain and spinal cord (columns 2, 3) than in the BD 42–44 embryo (column 1). Msi-2 was still expressed in intact adult spinal cord (column 4), and not up-regulated in regenerating adult cord (column 5). GAPDH was used as the internal housekeeping gene loading control, and (–RT) was performed without reverse transcriptase using the GAPDH primers. Each gene is shown in a separate row. All bands in each row are from one original gel (see Supplementary Figure 2) with limb regeneration sample bands excised here. Negative was shown as for image clarity.

FIGURE 7

mRNA tissue distribution of Xenopus Msi-1 in developing and regenerating spinal cord using section in situ hybridization (blue reaction product). (A) Cross section of intact NF 50 spinal cord. Msi-1 was strongly expressed in the ependymal layer. (B) Shh mRNA localization in intact NF 50 spinal cord in the section adjacent to the one shown in “A.” Shh is expressed in the floorplate (arrow). (C) NF 62 spinal cord cross section showed Msi-1 was still expressed in the ependymal layer, but mostly ventrally located and weaker. (D) NF 62+ Shh mRNA is still expressed in the ventral floorplate, overlapping the Msi-1 expression (arrow). (E) Parasagittal section of NF 52 spinal cord 1-day (1-D) post-transection showed high level of Msi-1 expression throughout the cord, lesion site and caudal stump. (F) Cross section of NF 52 regenerating cord 5-days (5-D) post-transection showed a broad ventral arch of Msi-1 expression in the ependymal outgrowth. The dorsal portion of the notochord is visible below the outgrowth (noto). (G) Cross section of NF 52 regenerating cord 5-days (5-D) post-transection in the section adjacent to the one shown in “F”. Shh localization is largely in a complementary zone of outgrowth ventral to Msi-1 expression. (Combined Msi-1 and Shh expression shown in Supplementary Figure 6.) The dorsal portion of the notochord is visible below the outgrowth (noto). (H) Cross section of lesion site in NF 52 spinal cord 10-days (10-D) post-transection showed the reconstructed central canal with ependymal Msi-1 expression. (I) NF 52 spinal cord 10-days post-transection (10-D) in the section adjacent to the one in “H” showed Shh expression restored to floorplate location. Differential interference contrast optics, transmitted light images.

FIGURE 8

Axolotl mRNA (purple) or Msi antibody (red) tissue localization and DAPI nuclear label (blue). (A) Late tailbud embryo (BD 42–44) cross-section in situ hybridization for Msi-1 mRNA showed localization in ependymal region around the central canal and in gray matter. Differential interference contrast image. (B) Antibody localization in late tailbud embryo sections showed Msi-1 protein was present in the same regions as the mRNA shown in “A.” (C) Juvenile cord maintained Msi-1 expression in the ependymal zone, extending dorsally. (D) Intact adult Axolotl cord showed no Msi-1 protein expression, only DAPI-stained nuclei (blue). (E) Parasagittal section of 2 weeks adult regenerative outgrowth showed strong upregulation of ependymal Msi-1expression. (F) In cross section, at higher magnification than in “E” the mesenchymal nature of the Msi-1-positive 2-week adult regenerated outgrowth is shown. Fluorescence microscope images. All images are dorsal side up.

FIGURE 9

Axolotl Shh semi-quantitative RT-PCR. Intact adult Axolotl cord showed low levels of Shh (gel column 1). Expression was upregulated at 3 weeks regeneration (column 2). Intact late tailbud (BD 42–44) Axolotl embryo cord, juvenile cord and brain all showed high levels of Shh expression (columns 3–5). GAPDH was used as the internal housekeeping gene control, and the No-RT (–RT) was performed without reverse transcriptase using the GAPDH primers. Each gene is shown in a separate row. All bands in each row are from one gel, but limb regeneration results were excised and only CNS results that flanked limb samples are shown (see original rows in Supplementary Figure 3). Negative was shown as for image clarity.

FIGURE 10

Axolotl GFAP antibody labeling (red), with DAPI labeled nuclei (blue). GFAP labeling showed bundles of radial ependymal fibers in sections of the intact juvenile (A) and adult (B) Axolotl spinal cord. Arrow in B indicates the bundle of tanycyte radial fibers extending dorsally. There was also cytoplasmic GFAP content at both stages. In the stump region at 3 weeks of regeneration, GFAP was reduced in the radial processes (C) and was at low levels in the regenerative outgrowth (D). Arrows indicate location of the denticulate ligaments associated with meninges surrounding the regenerative outgrowth (D). The bright red spots are red blood cell autofluorescence. Fluorescence microscope images.

FIGURE 11

Xenopus and Axolotl cell proliferation; PCNA antibody (red) and DAPI nuclear stain (blue) are used throughout. (A) PCNA staining in NF 50 spinal cord. The ependymal layer cells are PCNA-positive. The PCNA labeled region coincides with the region of Msi-1-positive cells shown in Figure 7A. (B) NF 62 spinal cord with PCNA antibody and DAPI staining. Some cells in the ependymal layer are still proliferating, but in a reduced area around the central canal. (C) Parasagittal section shows regenerative outgrowth 7-days (7D) post-transection regenerating NF 50 tadpole spinal cord. The zone of outgrowth and lesion site is indicated within the dashed line. The entire ependymal population contains proliferating cells. (D) Intact juvenile Axolotl (∼12 cm; less than 6 months old) spinal cord cross section shows extensive ependymal and gray matter PCNA labeling. There is a substantial dorsal extension of labeled ependymal cells. (E) Intact adult Axolotl (∼25 cm; approximately 2–3 years old) spinal cord cross section shows strong, but reduced PCNA labeling in the ependymal zone. The dorsal plume of PCNA-labeled cells no longer exists and labeling is reduced in the gray matter laterally. (F) Parasagittal section showing PCNA-positive cells in mesenchymal ependymal outgrowth. (Supplementary Figure 8 shows DIC image of the section with the outgrowth and stump region labeling for orientation.) Fluorescence microscope images.

FIGURE 12

Xenopus, Axolotl cell culture in EGF-containing culture medium (+EGF). (A–C) Phase contrast images. (A) Control NF 50 Xenopus ependymal cells in +EGF, 1 day (1D) 6 to 8 large patches of cells adhere to each dish. (B) NF 64 cultures: small, islands of cells, 4–6/dish, exhibit poor attachment and growth even at 8 days (8D) in culture. (C) Control adult Axolotl 2 weeks regenerating ependymal outgrowth explants in EGF at 3 days (3D) culture, patches of cells continue to expand in culture. (D,E) Fluorescent-Phalloidin probes are used to show the organization of F-actin in Xenopus and Axolotl +EGF cultures. (D) Alexa fluor 488-Phalloidin green fluorescence in a +EGF Xenopus ependymal culture shows typical mesenchymal cell F-actin organization, parallel to the long axis. (E) Rhodamine-Phalloidin red fluorescence in a +EGF Axolotl culture shows cell F-actin organization typical of mesenchymal outgrowth. DAPI nuclear stain used in “D,E” (blue). (D,E) Fluorescence microscope images.

FIGURE 13

Xenopus and Axolotl Msi-1, proliferation in vitro. (A) NF 50 Xenopus dish in situ hybridization with Msi-1 riboprobe shows Msi-1 expression is maintained in culture. Bright field image (B) Axolotl anti-Msi-1 antibody localization in 2 weeks regenerating mesenchymal ependymal outgrowth. Msi-1 expression is maintained in culture. (C) Labeling with PCNA antibody showed NF 50 Xenopus ependymal cells are proliferative in culture. (D) Labeling with PCNA antibody showed 2 weeks regenerating adult Axolotl mesenchymal ependymal outgrowth is proliferative. (B–D) Fluorescence microscope images. All of the cultures are in EGF-containing medium.