Zebrafish models of Mucopolysaccharidosis types IIIA, B, & C show hyperactivity and changes in oligodendrocyte state

Sanfilippo syndrome childhood dementia, also known as mucopolysaccharidosis type III (MPS III), is a rare inherited lysosomal storage disorder. Subtypes of MPS III are caused by deficiencies in one of four enzymes required for degradation of the glycosaminoglycan heparan sulfate (HS). An inability to degrade HS leads to progressive neurodegeneration and death in the second or third decades of life. Knowledge of MPS III pathogenesis is incomplete, and no effective therapies exist. We generated the hypomorphic mutations sgshS387Lfs, nagluA603Efs and hgsnatG577Sfsin the endogenous zebrafish genes orthologous to human SGSH, NAGLU, and HGSNAT that are loci for mutations causing MPS III subtypes MPS IIIA, B and C respectively. Our models display the primary MPS III disease signature of significant brain accumulation of HS, while behavioural analyses support hyperactivity phenotypes. Brain transcriptome analysis revealed changes related to lysosomal, glycosaminoglycan, immune system and iron homeostasis biology in all three models but also distinct differences in brain transcriptome state between models. The transcriptome analysis also indicated marked disturbance of the oligodendrocyte cell state in the brains of MPS IIIA, B and C zebrafish, supporting that effects on this cell type are an early and consistent characteristic of MPS III. Overall, our zebrafish models recapture key characteristics of the human disease and phenotypes seen in mouse models. Our models will allow exploitation of the zebrafish's extreme fecundity and accessible anatomy to dissect the pathological mechanisms both common and divergent between the MPS IIIA, B, and C subtypes. Overall design: To compare brain transcriptomes in our three MPS III zebrafish lines, we generated families of 2nd Filial generation zebrafish originating from G0 in-crosses of parental zebrafish heterozygous for two of the MPS III mutations. We initially aimed to analyse n = 6 fish per genotype from single families of progeny from in-crosses of doubly heterozygous mutant fish. We chose n = 6 as, according to a previous power calculation, this would allow approximately 70% power to detect differentially expressed genes in zebrafish models of Alzheimer's disease. However, due to smaller than expected clutch sizes, and questionable purified RNA quality, we were forced to use multiple families of siblings, or n<6, in some analyses. For the A and B (AB) arm of the experiment, we analysed n = 4 fish per genotype, which also included 4 sgsh S387Lfs/+ fish to explore cellular processes affected in the brains of presumably unaffected carriers of MPS III mutations. For the A and C (AC) arm, we sequenced all homozygous sgsh S387Lfs zebrafish in the family (n = 7) to account for the lower number of fish with this genotype in the AB arm. For the B and C (BC) arm, we could not obtain a large enough family in a single breeding event within our timeframe to be able to compare with the other families. Therefore, we analysed fish spawned by two pairs of parents (P1 and P2, both doubly heterozygous for the nagluA603fs and hgsnatG577fs mutations) in a total of 3 spawning events.