Characterization of the dual functional effects of heat shock proteins (HSPs) in cancer hallmarks to aid development of HSP inhibitors.

Heat shock proteins (HSPs), a representative family of chaperone genes, play crucial roles in malignant progression and are pursued as attractive anti-cancer therapeutic targets. Despite tremendous efforts to develop anti-cancer drugs based on HSPs, no HSP inhibitors have thus far reached the milestone of FDA approval. There remains an unmet need to further understand the functional roles of HSPs in cancer. We constructed the network for HSPs across ~ 10,000 tumor samples from The Cancer Genome Atlas (TCGA) and ~ 10,000 normal samples from Genotype-Tissue Expression (GTEx), and compared the network disruption between tumor and normal samples. We then examined the associations between HSPs and cancer hallmarks and validated these associations from multiple independent high-throughput functional screens, including Project Achilles and DRIVE. Finally, we experimentally characterized the dual function effects of HSPs in tumor proliferation and metastasis. We comprehensively analyzed the HSP expression landscape across multiple human cancers and revealed a global disruption of the co-expression network for HSPs. Through analyzing HSP expression alteration and its association with tumor proliferation and metastasis, we revealed dual functional effects of HSPs, in that they can simultaneously influence proliferation and metastasis in opposite directions. We experimentally characterized the dual function of two genes, DNAJC9 and HSPA14, in lung cancer cells. We further demonstrated the generalization of this dual direction of associations between HSPs and cancer hallmarks, suggesting the necessity to more carefully evaluate HSPs as therapeutic targets and develop highly specific HSP inhibitors for cancer intervention. Our study furnishes a holistic view of functional associations of HSPs with cancer hallmarks to aid the development of HSP inhibitors as well as other drugs in cancer therapy.