Growth differentiation factor 6 as a putative risk factor in neuromuscular degeneration

Abstract

Mutation of Glass bottom boat, the Drosophila homologue of the bone morphogenetic protein or growth/differentiation factor (BMP/GDF) family of genes in vertebrates, has been shown to disrupt development of neuromuscular junctions (NMJ). Here we tested whether this same conclusion can be broadened to vertebrate BMP/GDF genes. This analysis was also extended to consider whether such genes are required for NMJ maintenance in post-larval stages, as this would argue that BMP genes are viable candidates for analysis in progressive neuromuscular disease. Zebrafish mutants harboring homozygous null mutations in the BMP-family gene gdf6a were raised to adulthood and assessed for neuromuscular deficits. Fish lacking gdf6a exhibited decreased endurance (∼ 50%, p = 0.005) compared to wild type, and this deficit progressively worsened with age. These fish also presented with significantly disrupted NMJ morphology (p = 0.009), and a lower abundance of spinal motor neurons (∼ 50%, p<0.001) compared to wild type. Noting the similarity of these symptoms to those of Amyotrophic Lateral Sclerosis (ALS) model mice and fish, we asked if mutations in gdf6a would enhance the phenotypes observed in the latter, i.e. in zebrafish over-expressing mutant Superoxide Dismutase 1 (SOD1). Amongst younger adult fish only bigenic fish harboring both the SOD1 transgene and gdf6a mutations, but not siblings with other combinations of these gene modifications, displayed significantly reduced endurance (75%, p<0.05) and strength/power (75%, p<0.05), as well as disrupted NMJ morphology (p<0.001) compared to wild type siblings. Bigenic fish also had lower survival rates compared to other genotypes. Thus conclusions regarding a role for BMP ligands in effecting NMJ can be extended to vertebrates, supporting conservation of mechanisms relevant to neuromuscular degenerative diseases. These conclusions synergize with past findings to argue for further analysis of GDF6 and other BMP genes as modifier loci, potentially affecting susceptibility to ALS and perhaps a broader suite of neurodegenerative diseases.

Figure 1. Zebrafish harboring homozygous mutations in gdf6 can be raised to adulthood and do not display overt skeletal defects.

A. gdf6^−/− fish are viable into adulthood, exhibit variably penetrant microphthalmia and normal body morphology. B. gdf6^−/− fish exhibit somewhat decreased survival compared to gdf6^+/+ siblings (n = 11 gdf6^−/− fish; n = 4 gdf6^+/+ siblings). C, D. gdf6^−/− fish lack overt skeletal phenotypes, as revealed by (C) clearing and staining or by (D) microCT analysis. The latter is further represented as Supplemental Movie S1. Scale bars are 5 mm. A variety of fin morphologies were present in the fish examined, but these were neither different between experimental groups (genotypes) nor a significant covariant with swim performance (see Results).

Figure 2. Zebrafish harboring homozygous mutations in gdf6 exhibit decreased endurance.

A. gdf6^−/− fish have a lower endurance compared to gdf6^+/+ siblings as measured by increasing water velocity in a swim channel to determine the U _crit i.e. ‘critical swimming speed’ which is taken to be the highest speed that a fish can swim at for a period of several minutes before exhaustion. Each fish tested at 9 months is plotted. B. The same data in panel A (9 month) plotted along with 18 month old siblings. gdf6^−/− fish have approximately 50% lower endurance compared to gdf6^+/+ siblings at each age (*p = 0.005, **p<0.01). Endurance trends towards being decreased in older fish of each genotype, but this difference is not significant. C. Open field tests of average swim velocity during 10 minutes of each hour through a circadian cycle, in tanks replicating lifetime husbandry conditions, show a lower mean movement at night (in either genotype, **p<0.01) but no difference between genotypes in any measure during day (p = 0.247) or night (p = 0.814).

Figure 3. Zebrafish harboring homozygous mutations in gdf6 exhibit altered neuromuscular junctions and fewer spinal motor neurons.

A–B. Assessment of neuromuscular junction (NMJ) morphology including presynaptic (synaptophysin) and postsynaptic (αBTX) compartments in 9 month old gdf6^−/− and gdf6^+/+ siblings by immunohistochemistry. C. ALS-like increases in motoneuron pre-synaptic volumes are observed in gdf6^−/− fish when normalized to post-synaptic volumes (**p = 0.009, n = 5 fish per genotype). Coefficients of colocalization for pre- and post-synaptic compartments are not altered, as is expected in later-stage zebrafish ALS models. D–E. Motor neuron cell bodies were identified in cross-sections of spinal cord using immunohistochemistry against choline acetyl transferase (ChAT, e.g. arrowheads) in nine month old gdf6^+/+ and gdf6^−/− fish (panels A and B, respectively). Bottom panels affirm motoneuron cell body identification using actin and nuclear counter-stains. F. gdf6^−/− fish have approximately 50% the abundance of spinal motor neurons compared to sibling gdf6^+/+ fish. (Mann-Whitney U Test, *** p<0.001, n>50 sections from 4 fish per genotype, researcher blinded to genotype during quantification). Scale bar = 60 µm in A,B and 50 µm in D,E.

Figure 4. Mutations in gdf6 sensitize SOD1∧G93R zebrafish to develop ALS-like symptoms.

Six genotypes combining gdf6^−/− alleles and SOD1∧G93R alleles were examined in endurance and sprint tests, which primarily measure red and white muscle respectively. The SOD1∧G93R fish mimic many aspects of ALS progression . Swimming performance was measured in all six genotypes (determined by outcrosses and examining progeny) at 4.5 months of age. A. bigenic fish, i.e. gdf6^−/− mutants expressing SOD1∧G93R, had decreased survival compared to all other genotypes, i.e. their siblings ( = gdf6^+/+ or gdf6^+/−) with or without SOD1∧G93R. B. bigenic gdf6^−/− mutants expressing SOD1∧G93R had significantly lower endurance compared to heterozygous siblings also expressing SOD1∧G93R, and to WT and heterozygous siblings without SOD1∧G93R. C. Sprint test demonstrates a significant deficit in white muscle function only when SOD1∧G93R and gdf6^−/− genotypes are combined ( = 25% deficit). (ANOVA and multiple linear regression, p<0.05; sample size indicated below graph).

Figure 5. Disruption of gdf6 function exacerbates neuromuscular junction abnormalities in ALS model zebrafish.

ALS model zebrafish possess disruptions to neuromuscular junctions (NMJ), and loss of gdf6 function exacerbates this by 7 months of age. A. The presynaptic junctions (labeled with synaptophysin antibody) and postsynaptic junctions (labeled with fluorescently tagged αBTX) in ALS model zebrafish expressing the mutant SOD1^G93R show punctate morphology, deviations in presynaptic volume and less overall colocalization compared to WT sibling junctions. Some abnormalities are exacerbated in bigenic siblings expressing the mutant SOD1^G93R that are also gdf6^−/− (scale bar is 40 µm). B. Quantification of these NMJs suggests the presynaptic/postsynaptic volume ratios of SOD1^G93R and bigenic gdf6^−/−;SOD1^G93R zebrafish are larger than those of WT siblings at this age, though these differences do not rise to statistical significance (Kruskall-Wallis ANOVA, p = 0.134; n = 6,4,6 for WT, SOD1^G93R, and bigenic fish respectively). Colocalization coefficients, that measure overall colocalization of presynaptic and postsynaptic junctions, are altered in these fish. The values for SOD1^G93R zebrafish are significantly lower than wild type sibling values, indicating that presynapses and postsynapses overlap less, as characterized previously for this transgenic ALS model . Bigenic SOD1^G93R zebrafish that are also gdf6−/− have a dramatically lower coefficient than either sets of siblings, including being 30% lower than ALS model SOD1^G93R fish with normal gdf6a (*p<0.05; ***p<0.001. Kruskall-Wallis ANOVA with pairwise comparisons).