Autophagy induction sensitizes cancer cells to anti-cancer drugs

ABSTRACT

Resistance to anti-cancer therapy is a major challenge for cancer treatment. Many studies revealed that macroautophagy/autophagy inhibition can overcome autophagy-mediated therapy resistance, but these efforts have not yet led to the success of clinical applications. In a recent paper, we established a 37-gene autophagy signature to estimate the autophagy status of approximately 10,000 tumor samples across 33 cancer types from The Cancer Genome Atlas, and muti-omics characterization reveals that autophagy induction may also sensitize cancer cells to anti-cancer drugs. These findings provide a comprehensive resource of molecular alterations associated with autophagy and highlight the potential to utilize drug sensitivity induced by autophagy to overcome the resistance of cancer therapy.

Autophagy, a conserved lysosomal degradation catabolic pathway that supports nutrient recycling and metabolic adaptation, has been implicated as a significant process that regulates cancer. Much evidence demonstrates that autophagy inhibition can inhibit the growth of established tumors and sensitize tumor cells to anti-cancer drugs. Based on this knowledge, many clinical trials involving autophagy inhibitors, e.g., chloroquine or hydroxychloroquine, are ongoing in combination with other drugs, such as the MTOR inhibitor temsirolimus, the proteasome inhibitor bortezomib, and the histone deacetylase inhibitor vorinostat. Unfortunately, these efforts have so far been unsuccessful in clinical applications. Our recent work utilizing multi-omics analysis reveals that autophagy induction may also sensitize cancer cells to anti-cancer drugs [1].

Estimation of autophagy level is a great challenge for autophagy research. Autophagy was mainly estimated by LC3-based assays and/or SQSTM1/p62-based assays and direct observation of autophagy-related structures by electron microscopy. However, none of these approaches are currently feasible under human physiological and pathological conditions. Therefore, we identified a 37-gene set and utilized the single-sample gene set enrichment analysis (ssGSEA), to estimate autophagy status and confirmed its robustness in 6 public datasets across 5 different cancer types for which samples with high autophagy levels show significantly higher autophagy scores than samples with low autophagy levels. Based on the 37-gene set, we calculated the autophagy score for approximately 10,000 tumor samples across 33 cancers from The Cancer Genome Atlas (TCGA), which is the first comprehensive resource for autophagy-related features across cancers.

We depicted the global landscape of autophagy scores to elucidate autophagy-related molecular features across cancers. We observed that significantly varied genes in the autophagy score-high group versus the autophagy score-low group are enriched in autophagy-related pathways, including the PI3K-AKT, the MTOR, and the MAPK signaling pathways. We further used genes targeted by autophagy-related miRNAs and transcriptional factors (TFs) to construct the miRNA-Target regulatory network and TF-Target network, which provide biological implications for further investigations of these target genes. Additionally, the majority of genes in these autophagy-related pathways are highly associated with the sensitization to anti-cancer drugs, and some varied somatic copy number alterations/SCNVs harbor clinically actionable genes/CAGs. These results suggest the possibility to explore therapeutic strategies via targeting these genes.

We further examined the correlations between the autophagy score and imputed drug data of 138 drugs in the TCGA to elucidate the associations of drug response and autophagy status. We found that autophagy score-high cancer samples are more sensitive to certain drugs. This finding contrasts with the traditional view that autophagy will lead to drug resistance, suggesting complicated effects of autophagy on the drug response. In vitro and in vivo experiments further confirm that the melanoma tumor cells are more sensitive to etoposide and BMS536924 in rapamycin-induced or starvation-induced autophagy conditions. RNA-seq of melanoma cancer cell lines treated with rapamycin plus etoposide or BMS536924 underline that DDIT4 expression is downregulated by this combination treatment, and the overexpression and knockdown of DDIT4 show that DDIT4 may be a crucial gene to mediate drug sensitivity under autophagy-inducing conditions.

In summary, in an attempt to elucidate the complicated role of autophagy in cancer, we established a 37-gene autophagy signature to estimate autophagy status of tumor samples and provided comprehensive landscapes of molecular features and drug sensitivity across cancers, which would be an important resource for further autophagy research. More significantly, autophagy may sensitize cells to an anti-cancer drug in contrast with the traditional view that autophagy will lead to drug resistance, which provides a novel strategy for cancer therapy.

Funding Statement

The author(s) reported there is no funding associated with the work featured in this article.

Disclosure statement

No potential conflict of interest was reported by the authors.