A small number of heat-shock proteins have previously been shown to act protectively on aggregation of several proteins containing an extended polyglutamine (polyQ) stretch, which are linked to a variety of neurodegenerative diseases. A specific subfamily of heat-shock proteins is formed by the HSPB family of molecular chaperones, which comprises 10 members (HSPB1-10, also called small HSP). Several of them are known to act as anti-aggregation proteins in vitro. Whether they also act protectively in cells against polyQ aggregation has so far only been studied for few of them (e.g. HSPB1, HSPB5 and HSPB8). Here, we compared the 10 members of the human HSPB family for their ability to prevent aggregation of disease-associated proteins with an expanded polyQ stretch. HSPB7 was identified as the most active member within the HSPB family. It not only suppressed polyQ aggregation but also prevented polyQ-induced toxicity in cells and its expression reduces eye degeneration in a Drosophila polyQ model. Upon overexpression in cells, HSPB7 was not found in larger oligomeric species when expressed in cells and-unlike HSPB1-it did not improve the refolding of heat-denatured luciferase. The action of HSPB7 was also not dependent on the Hsp70 machine or on proteasomal activity, and HSPB7 overexpression alone did not increase autophagy. However, in ATG5-/- cells that are defective in macroautophagy, the anti-aggregation activity of HSPB7 was substantially reduced. Hence, HSPB7 prevents toxicity of polyQ proteins at an early stage of aggregate formation by a non-canonical mechanism that requires an active autophagy machinery.

Based on our results, we propose a model for the aggregation-associated pathology of spinocerebellar ataxia type 3: GCS and DNS aggregation likely represents early stages of pathology, which progresses towards formation of p62-positive NNI. A fraction of NNI exhibits UBB⁺¹ staining, implying proteasomal overload at a later stage. Subsequently, the stress-inducible HSPA1A is elevated while DNAJB1 is recruited into NNIs. This indicates that the stress response is only induced late when all endogenous protein quality control systems have failed.

We propose that the upregulation of HSPB8 and BAG3 may enhance the ability of astrocytes to clear aggregated proteins released from neurones and cellular debris, maintain the local tissue homeostasis and/or participate in the cytoskeletal remodelling that astrocytes undergo during astrogliosis.

Small heat shock proteins have been characterized in vitro as ATP-independent molecular chaperones that can prevent aggregation of un- or mis-folded proteins and assist in their refolding with the help of ATP-dependent chaperone machines (e.g., the Hsp70 proteins). Comparison of the functionality of the 10 human members of the small HSPB family in cell models now reveals that some members function entirely differently and independently from Hsp70 machines. One member, HSPB7, has strong activities to prevent toxicity of polyglutamine-containing proteins in cells and Drosophila, and seems to act by assisting the loading of misfolded proteins or small protein aggregates into autophagosomes.

In polyglutamine disorders, the length of the expanded CAG repeat shows a strong inverse correlation with the age at disease onset, yet up to 50% of the variation in age of onset is determined by other additional factors. Here, we investigated whether variations in the expression of heat shock proteins (HSP) are related to differences in the age of onset in patients with spinocerebellar ataxia (SCA)3. Hereto, we analysed the protein expression levels of HSPA1A (HSP70), HSPA8 (HSC70), DNAJB (HSP40) and HSPB1 (HSP27) in fibroblasts from patients and healthy controls. HSPB1 levels were significantly upregulated in fibroblasts from patients with SCA3, but without relation to age of onset. Exclusively for expression of DNAJB family members, a correlation was found with the age of onset independent of the length of the CAG repeat expansion. This indicates that DNAJB members might be contributors to the variation in age of onset and underlines the possible use of DNAJB proteins as therapeutic targets.