Deciphering the effect of mutant STUB1 on the heat shock response in SCAR16 patient-derived cells

Abstract

CHIP, encoded by the gene STUB1, is a central component of cellular protein homeostasis. It acts as a co-chaperone of HSC/HSP70 and HSP90 to modulate their activity and as an E3 ligase, tagging chaperone-bound misfolded proteins with ubiquitin and thereby leading to their degradation. Mutations in STUB1 cause the neurological disorder autosomal recessive cerebellar ataxia type 16 (SCAR16), characterized by atrophy of the cerebellum, brain stem and spinal cord, but patients also show widespread neurodegeneration with symptoms of epilepsy, cognitive impairment and hypogonadism. CHIP-/- mice present with cerebellar atrophy manifesting in distinct motor but also cognitive impairment phenotypes and aging-induced cardiac hypertrophy. 20% of CHIP-/- mice die post-natally and 100% die upon thermal challenge. Furthermore, they present with decreased stress tolerance, increased oxidative damage and increased insoluble protein aggregate levels. In vitro data supports this line of evidence by showing a higher vulnerability to stress and an impaired heat shock response (HSR) in STUB1-/- cells. To further understand the pathophysiology of SCAR16 and the effect of mutant STUB1 on the HSR, we first analyzed HSR induction and recovery in patient-derived fibroblasts. In accordance with previously published results from mesodermal cell lines, we saw a trend towards lower nuclear HSF1 levels after heat shock in patient cells compared to control cells which translated into lower HSP transcript levels upon induction. Furthermore, we detected an impaired HSR recovery on protein level, indicated by remaining high levels of HSP70. Interestingly, mutant CHIP did not alter cell viability of fibroblasts upon prolonged heat stress, but a CHIP-independent high susceptibility of fibroblasts to heat-inducible cell death was observed in all lines. As SCAR16 primarily affects the central nervous system, we next attempted to determine the relevance of CHIP in the HSR of cortical neurons. For this purpose and for ideal in vitro disease modeling, we generated induced pluripotent stem cells from fibroblasts of 3 SCAR16 patients and 3 healthy controls. Furthermore, we generated a homozygous knockout by CRISPR/Cas9-mediated genome editing from one isogenic control line by dual cleavage of DNA and deletion of 155 base pairs, thereby generating a premature stop codon at amino acid position 99. The homozygous knockout state was verified on transcript and protein level. All generated iPSC lines were validated for genomic integrity by exclusion of plasmid integration, SNP array analysis, resequencing of the mutation site and STR analysis. To verify the pluripotency of iPSCs, we assessed the expression of pluripotency-related intra- and extracellular proteins and the transcript levels of pluripotency-related genes and examined the potential of iPSCs to spontaneously differentiate into cells of all three germ layers. The generated iPSCs of 3 SCAR16 patients, 3 healthy controls and the STUB1(-/-) line were differentiated into cortical neurons specific for layer V of the neocortex with high homogeneity of CTIP2/TUJ-positive cells. In contrast to our findings in patient-derived fibroblasts, we did not see any distinctive (dys)functional CHIP-related effect on HSF1 translocation from the cytoplasm to the nucleus upon heat shock, as shown by cellular fractionation analysis. Yet, we observed an increased HSR induction on transcript level that, surprisingly, did not translate into any changes on HSP70 protein level. However, unstressed neurons of both patients and controls already expressed high levels of HSP70 compared to fibroblasts. Analyzing cell viability of cortical neurons upon prolonged heat stress, we saw a surprising resistance of neurons to this stress stimulus compared to fibroblasts, again regardless of CHIP mutations. To gain more insights into the effect of dysfunctional CHIP in cortical neurons, we next performed proteomic analysis and observed dysfunctional protein (re)folding and a higher basal oxidative stress level in patients. Our results question the role of impaired HSR in SCAR16 neuropathology and highlight cell-specific differences of the HSR, emphasizing the need of careful selection of proper disease models.