Research ArticleDEVELOPMENTAL NEUROSCIENCE

Neuroinflammatory signals drive spinal curve formation in zebrafish models of idiopathic scoliosis

See allHide authors and affiliations

Science Advances  12 Dec 2018:
Vol. 4, no. 12, eaav1781
DOI: 10.1126/sciadv.aav1781
  • Fig. 1 Loss of Ptk7 in the head is sufficient to cause scoliosis.

    (A) Schematic of transgene-based conditional ptk7 loss-of-function strategy. (B and C) Representative 5-dpf experimental larvae following trunk-specific (B and B″) and head-specific (C and C″) heat shock at 2.5 dpf. GFP to mCherry Tg(ubi:Switch) conversion reports localized Cre activity and the site of Tg(floxed-foxj1a::ptk7) recombination. (D) Experimental timeline of heat shock treatment. (E to G) Lateral views of adult ptk7 mutant experimental fish following no heat shock (E), localized trunk heat shock (F), or head heat shock (G). (H) Lateral view of control ptk7/+ adult fish following head heat shock. (E to H) Photo Credit: C. W. Boswell, The Hospital for Sick Children. (I and J) Violet highlighter ink injected into the ventricle of 14-dpf juvenile animals reports movement of CSF down the spinal canal of ptk7/+ fish [white arrowheads; n = 11 of 12 (G)], which is absent in ptk7 mutants [n = 12 of 12 (H)]. Photo Credit: J. L. M. Van Gennip, The Hospital for Sick Children. Scale bars, 500 μm (B, I, and J), 200 μm (C), and 1 mm (E to H).

  • Fig. 2 Spinal curve progression is associated with immune and neuroinflammatory responses.

    (A and B) Top 20 significantly (P < 0.05) up-regulated (A) and down-regulated (B) genes [sorted by log2 fold change (LFC)] identified in ptk7 mutant fish with severe spinal curvatures versus ptk7/+ control siblings. padj, Benjamini-Hochberg adjusted P value. (C) Gene Ontology (GO) term enrichment analysis of biological processes on all significantly up-regulated genes. Teal, complement response genes; yellow, other immune responses; blue, ptk7. (D) Heatmap depicting expression of immune and inflammatory response genes significantly (P < 0.05) up-regulated in ptk7 mutants at curve onset or during mild and severe curve progression versus ptk7 + Tg(foxj1a::ptk7) control siblings. (E) Quantitative reverse transcription polymerase chain reaction (qRT-PCR) verifying differential expression of selected genes in ptk7 mutants with severe curves versus ptk7 + Tg(foxj1a::ptk7) siblings. *P < 0.05, **P < 0.01. n.s., not significant.

  • Fig. 3 Macrophages accumulate within the spinal cord of ptk7 mutants.

    (A and B) Whole-mount fluorescent images of Tg(mpeg1:eGFP) macrophage reporter expression in ptk7/+ [n = 10 of 10 (A)] and ptk7 mutant fish [n = 14 of 15 (B)]. Dashed boxes indicate magnified regions (A′ and B′). (C and D) Transverse sections through 18-dpf ptk7/+ and ptk7 mutant fish demonstrating GFP+ macrophage accumulation within the spinal cord of ptk7 mutants. Dashed boxes indicate magnified regions of spinal cord (C′ and D′). (E) Plot demonstrating significant increase in the percentage of sections with greater than two macrophages within the spinal cord for ptk7 mutant (n = 260 sections from 10 fish) versus ptk7/+ controls (n = 151 sections from 5 fish) (****P < 0.0001). (F) No significant difference (P = 0.4273) between the percentage of sections with greater than seven macrophages within the muscle area for ptk7/+ (n = 146 sections from 5 fish) and ptk7 (n = 248 sections from 10 fish) mutants. Note that two and seven represent the average number of macrophage observed in control spinal cord and muscle tissues, respectively (fig. S5). Scale bars, 1 mm (A and B), 0.5 mm (A′ and B′), 100 μm (C and D), and 50 μm (C′ and D′).

  • Fig. 4 Proinflammatory cues within the CNS are sufficient to induce scoliotic curves.

    (A) Schematic of Tg(βact::LGSL mCherry-irg1l) or Tg(βact::LGSL ifng1-2) and the mosaic transposon approach used to generate clones of irg1l-expressing cells within foxj1a+ lineages. (B) qRT-PCR demonstrating functionality of mCherry-irg1l fusion construct. Significant up-regulation of the Irg1l target gene mmp9 is observed upon overexpression of equimolar amounts of both irg1l (P < 0.0004, two-tailed t test) and mCherry-irg1l (P < 0.0284, two-tailed t test) mRNA. Error bars represent the SE for the expression level fold change. ** P < 0.01, ***P < 0.001. (C and C″) Representative Tg(foxj1a::iCre) embryo at 72 hours post-fertilization (hpf) injected at the one-cell stage using mosaic labeling strategy. Arrowheads indicate clonal populations of mCherry-Irg1l–expressing cells within the foxj1a+ lineage (C′). (D and D′) Juvenile experimental fish develop spinal curvatures with mCherry-Irgl1+ cells visible at the site of curvature. Red rectangle indicates the area shown in higher magnification (D′). Arrowheads indicate clonal populations of mCherry-Irg1l–expressing cells within the foxj1a+ lineage. (E and F) Wild-type (WT) sibling controls injected with Tg(βact::LGSL mCherry-irg1l) develop normally (E), whereas Tg(foxj1a::iCre) animals develop scoliosis (F). (G and H) Wild-type sibling controls injected with Tg(βact::LGSL ifng1-2) develop normally (G), whereas Tg(foxj1a::iCre) fish develop scoliosis (H). Scale bars, 200 μm (C, C″, and D′) and 1 mm (D and E to H). (D to H) Photo Credit: C. W. Boswell, The Hospital for Sick Children.

  • Fig. 5 Aspirin treatment significantly affects spinal curve formation.

    (A) Kaplan-Meier estimator analysis demonstrating significant differences in the incidence and onset of scoliosis in control (n = 87) versus aspirin-treated (n = 111) ptk7 mutant fish (N = 5). (B) Box and whisker plot demonstrating a significant difference in the age of scoliosis onset between control (n = 90) and aspirin-treated (n = 88) ptk7 mutants (P = 1.8 × 10−7, two-tailed t test). *** P < 0.001. (C) Scoliotic fish were fixed at 30 days after phenotype onset to assess curve severity. Quantification of combined Cobb angles in untreated (n = 14) versus aspirin-treated (n = 12) ptk7 scoliotic fish revealed no significant difference in curve severity (P = 0.1779, two-tailed t test).

  • Fig. 6 NAC treatment significantly prevents spinal curve formation and progression.

    (A) Kaplan-Meier estimator analysis demonstrating significant differences in the incidence and onset of scoliosis in control (n = 95) versus NAC-treated (n = 47) ptk7 mutant fish. (B) Schematic representing the treatment at curve onset administration plan. Fish in the control are untreated and collected at 20 days after curve onset to assess curve magnitude. Fish in the treatment at onset cohort are administered NAC (200 mg/liter) when spinal curvatures are observed and then collected at 20 days after curve onset and curve severity are assessed. (C) Quantification of combined Cobb angles in untreated ptk7 fish (n = 24) and ptk7 fish treated with NAC at curve onset (n = 31) displayed a significant decrease in curve severity with NAC treatment (P = 0.0119, two-tailed t test). *P < 0.05.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/4/12/eaav1781/DC1

    Fig. S1. Cre/lox lineage tracing of foxj1a+ cells throughout zebrafish development demarcates all tissues that may be contributing to scoliosis phenotype.

    Fig. S2. Floxed rescue transgene can be efficiently excised in vivo to restore scoliosis phenotype in ptk7 mutants.

    Fig. S3. Volcano plots of differential gene expression in ptk7 mutants with severe curves versus ptk7/+ control siblings.

    Fig. S4. Volcano plots of differential gene expression in ptk7 mutants versus ptk7 + Tg(foxj1a::ptk7) siblings at various time points (curve onset, mild curvature, and severe curvature) confirms scoliosis-associated immune response.

    Fig. S5. Quantification of macrophage within the spinal cord and muscle of ptk7/+ and ptk7 mutant fish expressing the Tg(mpeg1:eGFP) reporter transgene.

    Fig. S6. Proinflammatory signals recruit macrophage, but expression within the vasculature is not sufficient to cause scoliosis.

    Fig. S7. Measuring Cobb angles to determine spinal curve severity.

    Table S1. List of primers used in this study for qRT-PCR analysis.

    Data file S1. RNA-seq differential expression analysis results.

    Movie S1. foxj1a+ cells are broadly present throughout the central and peripheral nervous system.

  • Supplementary Materials

    The PDF file includes:

    • Fig. S1. Cre/lox lineage tracing of foxj1a+ cells throughout zebrafish development demarcates all tissues that may be contributing to scoliosis phenotype.
    • Fig. S2. Floxed rescue transgene can be efficiently excised in vivo to restore scoliosis phenotype in ptk7 mutants.
    • Fig. S3. Volcano plots of differential gene expression in ptk7 mutants with severe curves versus ptk7/+ control siblings.
    • Fig. S4. Volcano plots of differential gene expression in ptk7 mutants versus ptk7 + Tg(foxj1a::ptk7) siblings at various time points (curve onset, mild curvature, and severe curvature) confirms scoliosis-associated immune response.
    • Fig. S5. Quantification of macrophage within the spinal cord and muscle of ptk7/+ and ptk7 mutant fish expressing the Tg(mpeg1:eGFP) reporter transgene.
    • Fig. S6. Proinflammatory signals recruit macrophage, but expression within the vasculature is not sufficient to cause scoliosis.
    • Fig. S7. Measuring Cobb angles to determine spinal curve severity.
    • Table S1. List of primers used in this study for qRT-PCR analysis.
    • Legend for data file S1
    • Legend for movie S1

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    • Data file S1 (Microsoft Excel format). RNA-seq differential expression analysis results.
    • Movie S1 (.mp4 format). foxj1a+ cells are broadly present throughout the central and peripheral nervous system.

    Files in this Data Supplement:

Navigate This Article