C-terminus of HSC70-interacting protein (CHIP) encoded by the gene STUB1 is a co-chaperone and E3 ligase that acts as a key regulator of cellular protein homeostasis. Mutations in STUB1 cause autosomal recessive spinocerebellar ataxia type 16 (SCAR16) with widespread neurodegeneration manifesting as spastic-ataxic gait disorder, dementia and epilepsy. CHIP−/− mice display severe cerebellar atrophy, show high perinatal lethality and impaired heat stress tolerance. To decipher the pathomechanism underlying SCAR16, we investigated the heat shock response (HSR) in primary fibroblasts of three SCAR16 patients. We found impaired HSR induction and recovery compared to healthy controls. HSPA1A/B transcript levels (coding for HSP70) were reduced upon heat shock but HSP70 remained higher upon recovery in patient- compared to control-fibroblasts. As SCAR16 primarily affects the central nervous system we next investigated the HSR in cortical neurons (CNs) derived from induced pluripotent stem cells of SCAR16 patients. We found CNs of patients and controls to be surprisingly resistant to heat stress with high basal levels of HSP70 compared to fibroblasts. Although heat stress resulted in strong transcript level increases of many HSPs, this did not translate into higher HSP70 protein levels upon heat shock, independent of STUB1 mutations. Furthermore, STUB1(−/−) neurons generated by CRISPR/Cas9-mediated genome editing from an isogenic healthy control line showed a similar HSR to patients. Proteomic analysis of CNs showed dysfunctional protein (re)folding and higher basal oxidative stress levels in patients. Our results question the role of impaired HSR in SCAR16 neuropathology and highlight the need for careful selection of proper cell types for modeling human diseases.

Disease modeling requires appropriate cellular models that best mimic the underlying pathophysiology. Human origin and an adequate expression of the disease protein are prerequisites that support information from a model to be meaningful. In this study we investigated expression profiles of (i) PBMCs and (ii) fibroblasts as patient derived cells as well as (iii) lymphoblasts and (iv) induced pluripotent stem cells (iPSC) as immortalized sources, and (v) iPSC-derived cortical neurons to assess their aptitude to model motor neuron diseases (MNDs) including hereditary spastic paraplegia (HSP), amyotrophic lateral sclerosis (ALS) and spinal muscular atrophy (SMA). We generated all five different cell types from two healthy donors and performed RNA sequencing to display expression patterns in MND-related genes. For the ten most common HSP genotypes we validated gene expression by qPCR. To verify the results on protein level, proteome analysis of fibroblasts, iPSCs and cortical neurons was performed. Depending on the specific MND gene we found largely different expression patterns. Out of 168 MND-related genes, 50 had their highest expression in iPSC-derived cortical neurons, 41 were most strongly expressed in fibroblasts, 26 in lymphoblasts, 22 in iPSCs, and 14 in PBMCs. Pathophysiologically related MNDs like HSPs associated with axonal transport deficits shared highest expression in cortical neurons. 15 MND-related genes were not detectable in any of the analyzed cell types. This may reflect the critical dependency of motor neurons on support of other cell types like oligodendrocytes which express myelin proteins like L1CAM (SPG1), PLP1 (SPG2) and MAG (SPG75) which are lacking in neurons but cause MNDs if mutated. This study provides comprehensive information on expression of genes associated with a large spectrum of MNDs. Expression profiles can be used to inform on appropriate cell models for genotype specific motor neuron research.