Differential Contribution of Cadm1-Cadm3 Cell Adhesion Molecules to Peripheral Myelinated Axons

Abstract

Cell adhesion proteins of the Cadm (SynCAM/Necl) family regulate myelination and the organization of myelinated axons. In the peripheral nervous system (PNS), intercellular contact between Schwann cells and their underlying axons is believed to be mediated by binding of glial Cadm4 to axonal Cadm3 or Cadm2. Nevertheless, given that distinct neurons express different combinations of the Cadm proteins, the identity of the functional axonal ligand for Cadm4 remains to be determined. Here, we took a genetic approach to compare the phenotype of Cadm4 null mice, which exhibit abnormal distribution of Caspr and Kv1 potassium channels, with mice lacking different combinations of Cadm1-Cadm3 genes. We show that in contrast to mice lacking the single Cadm1, Cadm2, or Cadm3 genes, genetic ablation of all three phenocopies the abnormalities detected in the absence of Cadm4. Similar defects were observed in double mutant mice lacking Cadm3 and Cadm2 (i.e., Cadm3^-/-/Cadm2^-/-) or Cadm3 and Cadm1 (i.e., Cadm3^-/-/Cadm1^-/-), but not in mice lacking Cadm1 and Cadm2 (i.e., Cadm1^-/-/Cadm2^-/-). Furthermore, axonal organization abnormalities were also detected in Cadm3 null mice that were heterozygous for the two other axonal Cadms. Our results identify Cadm3 as the main axonal ligand for glial Cadm4, and reveal that its absence could be compensated by the combined action of Cadm2 and Cadm1.SIGNIFICANCE STATEMENT Myelination by Schwann cells enables fast conduction of action potentials along motor and sensory axons. In these nerves, Schwann cell-axon contact is mediated by cell adhesion molecules of the Cadm family. Cadm4 in Schwann cells regulates axonal ensheathment and myelin wrapping, as well as the organization of the axonal membrane, but the identity of its axonal ligands is not clear. Here, we reveal that Cadm mediated axon-glia interactions depend on a hierarchical adhesion code that involves multiple family members. Our results provide important insights into the molecular mechanisms of axon-glia communication, and the function of Cadm proteins in PNS myelin.

Keywords: Schwann cells; SynCAM; axon-glia interaction; cell adhesion; myelin; node of Ranvier.

Figure 1.

Genetic deletion of Cadm4, but not of other family members, results in axonal organization abnormalities. Immunolabeling of teased sciatic nerve fibers isolated from 2.5-month-old wild-type mice (WT), or mice lacking Cadm1 (Cadm1^−/−), Cadm2 (Cadm2^−/−), Cadm3 (Cadm3^−/−), and Cadm4 (Cadm4^−/−), using antibodies to Kv1.2 potassium channels to label the juxtaparanodal region and Caspr to label the paranodal junction. Images of the merged channels are shown on the right panels. Arrowhead mark the nodes of Ranvier. Note the abnormal labeling of Caspr (asterisks) and Kv1.2 (arrows) away from the paranodal junction and the juxtaparanodal region, respectively. Scale bar: 10 μm.

Figure 2.

Mice lacking all three axonal Cadms exhibit abnormal axonal organization. A, Cross sections of sciatic nerves isolated from wild-type (WT) and the triple Cadm1^−/−/Cadm2^−/−/Cadm3^−/− null mice reveal normal myelin morphology. Toluidine blue (left panels) and electron microscopy (right panels) images are shown for each genotype. B, Immunolabeling of teased sciatic nerve fibers isolated from 2.5-month-old WT mice (WT), or mice lacking Cadm1–Cadm3 (Cadm123^−/−) using antibodies to Kv1.2 and NrCAM or Caspr and NrCAM as indicated. NrCAM was used to mark the location of the nodes of Ranvier. The single Cadm1 (Cadm1^−/−), Cadm2 (Cadm2^−/−), Cadm3 (Cadm3^−/−) null mice were used as additional controls. Arrows and asterisks mark abnormal distribution of Kv1.2 and Caspr, respectively. C, Summary of the abnormalities detected. Teased sciatic nerve fibers from the indicated genotypes were labeled using antibodies to Caspr or Kv1.2. Schematic illustrations describing the distribution of Caspr and Kv1.2 are shown below each panel. Caspr abnormalities include unwinding of the paranodal mesaxonal line, accumulation of Caspr in patches away from the paranodes, and its diffusion along the internodes. Kv1.2 abnormalities include the presence of gaps within the juxtaparanodal area, and the appearance of intermittent clusters as well as focal accumulation of these channels throughout the internodes. Arrowheads and arrows mark the nodes and the various abnormalities, respectively. D, E, Quantitative analysis of the percentage of Kv1.2 (D) and Caspr (E) abnormalities in the different genotypes. Graphs represent mean ± SEM of 300 sites corresponding to the paranodal junction and the juxtaparanodal region per each genotype (n = 3); *p < 0.0001. Scale bars: 10 μm.

Figure 3.

Cadm1-3 null mice phenocopy mice lacking Cadm4. A, Double immunolabeling of teased sciatic nerve fibers isolated from 2.5-month-old mice lacking Cadm1–Cadm3 (Cadm123^−/−), Cadm4 (Cadm4^−/−), or wild-type (WT) mice using antibodies to neurofilament (NFH) and either Kv1.2 or Caspr as indicated. Arrowheads and arrows mark the nodes and abnormal distribution of Kv1.2 and Caspr, respectively. B, C, Quantification of Kv1.2 channels (B) and Caspr (C) abnormalities in the different genotypes, presented as a percentage of total number of nodes. Graphs represent mean ± SEM of 300 sites corresponding to the paranodal junction and the juxtaparanodal region per each genotype (n = 3); *p < 0.0001. ns, not significant. Scale bars: 10 μm.

Figure 4.

Cadm3 is the main functional axonal ligand for Schwann cell Cadm4. A, Immunolabeling of teased sciatic nerve fibers isolated from 2.5-month-old wild-type (WT) animals or mice lacking either Cadm2 and Cadm3 (Cadm2^−/−/Cadm3^−/−), Cadm1 and Cadm2 (Cadm1^−/−/Cadm2^−/−), or Cadm1 and Cadm3 (Cadm1^−/−/Cadm3^−/−) using antibodies to Caspr and neurofilament (NFH). Insets represent higher magnification images of the boxed areas. Arrows mark abnormal distribution of Caspr. B, C, Quantification of Caspr (B) and Kv1.2 (C) abnormalities in the different genotypes, presented as a percentage of total number of nodes. Graphs represent mean ± SEM of 300 sites corresponding to the paranodal junction and the juxtaparanodal region per each genotype (n = 3); *p < 0.0001. Scale bars: 10 μm.

Figure 5.

Cadm-mediated axoglial interaction involves multiple members of the family. A–C, Caspr immunolabeling of teased sciatic nerve fibers isolated from 2.5-month-old mice lacking Cadm3 and one allele of Cadm1 and Cadm2 (Cadm3^−/−/Cadm1^−/+/Cadm2^−/+; A), mice lacking Cadm1 and one allele of Cadm2 and Cadm3 (Cadm1^−/−/Cadm2^−/+/Cadm3^−/+; B), or mice lacking Cadm2 and one allele of Cadm1 and Cadm3 (Cadm2^−/−/Cadm1^−/+/Cadm3^−/+; C). Arrows mark abnormal distribution of Caspr. Insets shows higher magnification of the boxed area in each panel. D, Quantification of Caspr abnormalities in the different genotypes, presented as a percentage of total number of nodes. Graph represents mean ± SEM of 300 from three animals per each genotype (n = 3); *p < 0.0001. Scale bars: 10 μm.