Loss of Gas6 and Axl signaling results in extensive axonal damage, motor deficits, prolonged neuroinflammation, and less remyelination following cuprizone exposure

Abstract

The TAM (Tyro3, Axl, and MerTK) family of receptor tyrosine kinases (RTKs) and their ligands, Gas6 and ProS1, are important for innate immune responses and central nervous system (CNS) homeostasis. While only Gas6 directly activates Axl, ProS1 activation of Tyro3/MerTK can indirectly activate Axl through receptor heterodimerization. Therefore, we generated Gas6^-/- Axl^-/- double knockout (DKO) mice to specifically examine the contribution of this signaling axis while retaining ProS1 signaling through Tyro3 and MerTK. We found that naïve young adult DKO and WT mice have comparable myelination and equal numbers of axons and oligodendrocytes in the corpus callosum. Using the cuprizone model of demyelination/remyelination, transmission electron microscopy revealed extensive axonal swellings containing autophagolysosomes and multivesicular bodies, and fewer myelinated axons in brains of DKO mice at 3-weeks recovery from a 6-week cuprizone diet. Analysis of immunofluorescent staining demonstrated more SMI32+ and APP+ axons and less myelin in the DKO mice. There were no significant differences in the number of GFAP+ astrocytes or Iba1+ microglia/macrophages between the groups of mice. However, at 6-weeks cuprizone and recovery, DKO mice had increased proinflammatory cytokine and altered suppressor of cytokine signaling (SOCS) mRNA expression supporting a role for Gas6-Axl signaling in proinflammatory cytokine suppression. Significant motor deficits in DKO mice relative to WT mice on cuprizone were also observed. These data suggest that Gas6-Axl signaling plays an important role in maintaining axonal integrity and regulating and reducing CNS inflammation that cannot be compensated for by ProS1/Tyro3/MerTK signaling.

Keywords: TAM signaling; axonal damage; demyelination; motor deficits; neuroinflammation.

Figure 1

Myelination is normal in naïve Gas6−/−Axl−/− (DKO) mice. A) DKO genotypes were confirmed by PCR and gel electrophoresis. The first lane in each panel shows the DNA molecular weight marker. Middle lane shows a lower migrating band signifying Gas6 (left panel) or Axl (right panel) knockout alleles; higher migrating bands signify the relevant WT alleles. All PCR products were evaluated on the same gel. B) 5μm sections of 4% paraformaldehyde fixed, paraffin-embedded brain and spinal cord sections were stained with H&E or MBP. Sagittal sections of brain denote the corpus callosum (arrows) of 8-week old WT (left) and DKO (right) and at higher magnification below. Equivalent MBP+ immunoreactivity is observed in cross sections of dorsal spinal cord. C) Equivalent numbers of cytoplasmic Olig1+ oligodendrocytes, and D) myelinated axons/X5000 TEM contact sheets (n=3/group) in groups of mice; naïve male and female values were the same (presented as mean±SEM; student’s t-test, p≥0.05).

Figure 2

DKO mice have less remyelination and more axonal damage at 6-weeks cuprizone + 3-weeks recovery relative to WT mice. Aa–b,B) MBP immunostaining was performed on 4%-paraformaldehyde fixed paraffin-embedded sections. MBP content was determined by immunofluorescence and light microscopy (a,b). B) Mean fluorescence intensity (MFI) was determined using Image J software. DKO mice had a MFI of 4302±168.4 (n=4) and WT mice had a MFI of 4977±192.2 (n=3); student’s t-test, p≤0.05. Ac–d,C) Cytoplasmic (arrows) and nuclear (arrow heads) Olig1+ cells are shown at 20x magnification (c–d). Supplemental Figure 1 shows the single and merged channels for DAPI and Olig1. Olig1+ cells were quantified by immunofluorescence at 60x magnification and translated to cells/mm2. Significantly fewer DAPI+ cytoplasmic Olig1+ cells were seen in the DKO mice (591.3±45, n=11) relative to WT mice (739.7±48.4, n=9; Mann-Whitney U test, p=0.027) and had significantly more DAPI+ nuclear Olig1+ cells (202.4±26.02, n=7) relative to WT (86.0±27.2, n=6; Mann-Whitney U test, p=0.033). Ae–f,D) SMI32+ axonal swellings (≥3μm, arrows) were counted at 60x magnification via immunofluorescence. E) The mean number of SMI32+ axonal swellings (≥3μm, arrows) for DKO mice was 322.4/mm2±74.5 (n=6) and for WT mice was 138 swellings/mm2±32.7 (n=7); student’s t-test, p=0.036. Ag–h,E) APP+ axonal swellings (≥3μm, arrows) indicative of defective axonal transport were quantified. E) The mean number of APP+ axonal swellings for DKO mice was 300.5/mm2±31.84 (n=6) and for WT mice was 156.4 swellings/mm2±39.4 (n=7); student’s t-test, p=0.018. Significance for all counts was determined using GraphPad Prism v7 software with values presented as mean±SEM.

Figure 3

TEM show extensive axonal pathology and fewer myelinated axons at 6-weeks cuprizone + 3- weeks recovery in male and female DKO mice. A) Transmission electron micrographs of WT (a–d) and DKO (e–g) mice show extensive axonal damage and pathology at the midline of the corpus callosum of DKO mice. DKO mice have more swollen axons (white arrows; Ae) containing multivesicular bodies (black arrows; Af,g), autophagosomes (black asterisk; Ag), and autophagolysosomes (white asterisk; Af,g) relative to WT mice. Boxes in Ab and enlarged in Ac,d show a rare enlarged unmyelinated axon in the WT corpus callosum (Ac), and a myelinated axon with vacuoles and damaged myelin (Ad). The number of myelinated axons in female and male DKO mice are significantly less than in WT mice (B,D). Also, the percentage of X5000 contact sheets/mouse containing swollen axons with MVBs and autophagic vacuoles was higher in DKO mice relative to WT mice (C,E). Values are present as mean±SEM. Statistical analysis was performed using a student’s t-test in GraphPad Prism software.

Figure 4

TEM show pronounced zebra bodies in the cell body of microglia and oligodendrocytes of DKO mice at 6-weeks cuprizone+3-weeks. Zebra bodies (black arrowhead) consisting of dark amorphous bodies composed of stacks of membranes were observed in WT (a) and DKO (b) corpus callosum. A large zebra adjacent to the nucleus of a microglia is seen in c; A denotes axon, M denotes microglia. The microglia is in contact with a swollen axon (white box) seen at higher magnification in d. A normal appearing oligodendrocyte with nucleus (lower right), rich cytoplasm and elongated process extending toward and contacting thinly myelinated axons indicative of remyelination is also observed in c (black box, enlarged in e). The arrow in e is indicative of a point of contact between the oligodendrocyte process and an axon at higher magnification. An oligodendrocyte with zebra bodies in the cell body adjacent to the nucleus is shown in f.

Figure 5

DKO mice display defects in debris clearance and more pathological lipid droplets at 6 weeks cuprizone + 3 weeks recovery. A) Frozen sections from DKO and WT mice were incubated in Oil Red O to stain for neutral lipid debris during 5- and 6-weeks cuprizone and 6-weeks cuprizone + 1-week and 3-weeks recovery. Mounted slides were imaged on a Perkin Elmer P250 Panoramic Slide Scanner and images were analyzed in Panoramic Viewer software. B) Lipid droplets appearing as round gray spherical structures depicted by white arrowheads were counted in 5–7 contact sheets/mouse from male and female DKO and WT mice. DKO mice had significantly more lipid droplets at the midline of the corpus callosum following 3- weeks recovery from cuprizone. Values are presented as mean±SEM and analyzed using student’s t-test.

Figure 6

Remyelination in DKO mice lags following 1-week and 3-weeks recovery from cuprizone diet. A) Immunofluorescent staining of OPCs (PDGFRα, left), Olig2+ OLs (center), and CC1+ OLs (right) at 5- and 6- weeks cuprizone diet and 1-week and 3-weeks recovery. No significant differences in the numbers of PDGFRα+ OPCs, total Olig2+ OLs, and CC1+ OLs at the midline of the corpus callosum of WT (black bars) and DKO (grey bars) mice by Mann-Whitney U test. B) Frozen sections from WT and DKO mice after 5- and 6-weeks cuprizone and 6-weeks cuprizone + 1-week and 3-weeks recovery were stained with Myelin Black Gold-II (Millipore). C) Frozen sections from WT (n=9) and DKO (n=6) mice after 6-weeks cuprizone + 6- weeks recovery were stained as in B. Corpus callosum width displayed in micrometers±SEM was measured using Panoramic Viewer software. Measurements were analyzed using in GraphPad Prism v7 software using student’s t-test.

Figure 7

Astrocyte and microglia/macrophage recruitment is not affected by deletion of Gas6-Axl signaling. Immunofluorescent staining for astrocytes (left, GFAP) and microglia/macrophages (right, Iba1) was performed on frozen sections (n=3–9/group/time point). GFAP+ and Iba+ glia were counted and translated to cells/mm2±SEM. Data were analyzed in GraphPad Prism by Mann-Whitney U test.

Figure 8

DKO mice have increased pro-inflammatory cytokine expression at 6-weeks cuprizone treatment and 1-week recovery relative to WT mice. A 2-mm section of corpora callosa was isolated at the midline from DKO (gray thatched bars) and WT (black bars) mice following perfusion with PBS. The isolated region was homogenized in Trizol. Extracted RNA was reverse transcribed and qRT-PCR was performed. All genes were normalized to HPRT expression with naïve (0-weeks) WT mice set as the reference. Differences in expression are shown as 2−ddCt. Data is presented as mean±SEM and analyzed in GraphPad Prism by student’s t-test, p≤0.05 as significant.

Figure 9

Deletion of Gas6 and Axl signaling causes behavioral and motor deficits in DKO mice during cuprizone-mediated toxicity. Two behavioral tests were used to assess motor function during cuprizone diet. A) Negative geotaxis was used to assess proprioceptive pathways and is measured by the latency (seconds) to right when placed head downward on a grid screen. Mice were tested after 5- and 6-weeks cuprizone diet. B) The balance beam test, assessing motor coordination as the number of slips made while crossing a round wooden beam over 3 levels of difficulty (with smaller diameter beams being harder). For A and B, data were analyzed in GraphPad Prism using Mann-Whitney, p≤0.05 as significant. C) Cognitive deficits in the object placement test visuospatial memory were significantly different between cuprizone-treated DKO mice and cuprizone-treated WT mice after 6-weeks when preference for the novel object is expressed as a preference score (left panel, p=0.003) or as the % of subjects passing and failing (X2=6.4, p<0.01).