A Novel Ubiquitin Complex Regulates Aneuploid Epithelial Tumors by Moderating an Integrated Stress Response

In recent years, the rapid accumulation of large-scale patient genomic data has catalyzed a deeper understanding of the elaborate pathways that contribute to tumor progression and metastasis. At the same time, the advent of genome-wide and focused functional screens (using CRISPR or siRNA) has helped to ascribe key functions to a multitude of genes or protein complexes, further refining the landscape of the steps to malignant progression and treatment resistance. Together, these advances have multiplied the number of potential new individual or ancillary pharmacologic targets for cancer treatment (1).

In this issue of Cancer Discovery, Cervia and colleagues (2) aimed to broaden the therapeutic options in the treatment of malignancies by expanding the repertoire of candidate druggable genes beyond those being ascribed to an oncogene function. For this, they took an approach that focused on tumor dependency gene interactions and aimed to recognize common molecular expression pathways that are critical for cancer cell survival and could serve as novel therapeutic targets. Their approach relied on utilizing massive datasets derived from genome-wide CRISPR/Cas9-based screens in 1,086 cell lines and cluster genes based on their essentiality for cancer cell survival. This analysis revealed a new, four-member complex of coessential genes, all of which encode for ubiquitin ligase proteins: UBA6 (E1), BIRC6 (E2), KCMF1, and UBR4 (E3 ligases; Fig. 1). The authors designated this the BIRC6 module after its dominant regulatory member. Interestingly, the correlation of expression among these four ligases was similarly robust in additional gene expression data derived from RNAi-based genome-wide screens in 707 cell lines, solidifying the hypothesis that this new complex must be preferentially essential for the malignant phenotype. Providing further insight was the finding that whereas KCFM1 and UBR4 were universally essential, UBA6 and BIRC6 essentiality was confined to specific cancer subtypes. Notably, epithelial cell lines showed a stronger dependency on expression of BIRC6 compared with mesenchymal ones. Based on this, the authors clustered the cell lines used to mine these data into BIRC6 “dependent” and “nondependent” types to facilitate the deciphering of the downstream mechanisms in the context of essentiality.

The authors proceeded to validate the selective nature of BIRC6 essentiality by depleting its expression and analyzing in vitro cell viability and clonogenicity. Loss of BIRC6 led to cell-cycle disruption and activated apoptosis in the dependent cell clusters. To validate these results in vivo, the authors injected dependent or nondependent cell types in mice, knocked down BIRC6 in an inducible manner and monitored tumor progression. In line with their hypothesis, tumor volumes were strikingly decreased only in the tumor models characterized by BIRC6 dependency. BIRC6 has two distinct domains: a Baculovirus IAP Repeat (BIR) domain, which is a sequence common among the members of the inhibitors of apoptosis family (IAP) and is important for binding to proapoptotic signaling factors, and a C-terminal Ubiquitin Conjugating (UBC) domain, which is essential for interaction with ubiquitin target proteins (3). Inactivating mutations in these domains and downstream functional assays showed that the UBC domain is the essential one for tumor fitness. Furthermore, elegant biochemical interaction and coimmunoprecipitation experiments showed that BIRC6 directly binds to the other members of the complex, validating the formation of a functional ubiquitin ligase complex essential for the survival of a subset of cancers.

Transcriptome analysis of dependent or nondependent cell lines after BIRC6 ablation revealed changes in pathways that regulate apoptosis and promotion of a G₂–M arrest. However, another intriguing finding was the upregulation of genes of the integrated stress response pathway (ISR). The ISR constitutes an evolutionarily conserved pathway that enables cells to adapt to conditions of stress, including nutrient/amino acid deprivation, unfolded proteins in the endoplasmic reticulum, heme deficiency, and viral infection (4). Each stress is transduced to the translation initiation factor eIF2α by the phosphorylation of its ser51 subunit by four distinct corresponding kinases: GCN2 (amino acid deprivation), PERK (endoplasmic reticulum unfolded proteins), PKR (viral or cellular double-stranded RNA), and HRI (heme deficiency). The phosphorylation of eIF2α has two main consequences. The first is a rapid downregulation of CAP-dependent translation, which constitutes the bulk of proteins being synthesized in the cell. The second is the selective upregulation of the translation of specific mRNAs due to highly structured 5′-UTRs and to a lesser extent 3′-UTRs. A key transcription factor that is regulated in this manner is ATF4, which upregulates genes involved in ER stress recovery, amino acid biosynthesis and uptake, autophagy, and redox regulation. The authors showed that BIRC6 module downregulation leads to the selective activation of the eIF2α/ATF4 axis only in the dependent cell lines. To further dissect the mechanistic basis of cell fitness after BIRC6 ablation, they blocked the signaling downstream of eIF2α phosphorylation with a small chemical inhibitor, ISRIB, which is widely used in the field, and observed rescued loss of viability caused by the BIRC6 module, confirming ISR involvement. Interestingly though, ATF4 appeared to be dispensable for the cell-cycle and apoptotic effects elicited by BIRC6 depletion, indicating that some other downstream target of eIF2α phosphorylation was responsible for this effect.

HRI activates the ISR upon various stress conditions, such as oxidative stress, or osmotic or heat shock, but mainly under heme depletion (4). To test whether BIRC6 downregulation causes ISR activation and loss of viability via HRI, and not any of the other eIF2α kinases (PERK or GCN2 or PKR), Cervia and colleagues blocked each kinase individually and observed rescue of the phenotype only after HRI ablation. Subsequently, the authors analyzed the proteome after BIRC6 complex suppression to identify possible ubiquitin targets that would lead to the activation of ISR-associated genes. To identify which of these proteins are directly affected by BIRC6 (rather than as a secondary effect), they blocked BIRC6 in the presence or absence of ISRIB. In either condition, HRI was strongly upregulated, indicating that this kinase is upregulated by suppression of the BIRC6 module and is directly controlled by the complex. In supporting experiments, they demonstrated coimmunoprecipitation of HRI with the BIRC6 complex components and traced targeted ubiquitination of HRI by the complex. Altogether, these data established that HRI is a direct target of the BIRC6 ubiquitin complex. Lastly, HRI ablation failed to activate ISR components in only the BIRC6-dependent cell lines, suggesting that HRI is constitutively active in the dependent cluster and that BIRC6 inhibition is required to prevent a constitutively active (and likely lethal) ISR (Fig. 1).

Finally, to test the potential physiologic relevance of their findings for human disease, the authors evaluated patient datasets. They uncovered high expression of HRI in tumor versus normal tissues and a strong correlation between HRI and the components of the BIRC6 complex. Interestingly, they identified a strong correlation between BICR6 and the degree of tumor aneuploidy. Aneuploidy, which is the imbalanced complement of chromosomes, is present in ∼90% of all solid tumors and is often associated with chemotherapy and targeted therapy resistance (5). Strikingly, the BIRC6-dependent cell lines had a higher aneuploidy score compared with the nondependent ones, suggesting a possible therapeutic strategy for aneuploid epithelial tumors by targeting this novel ubiquitin complex.

In this study, Cervia and colleagues have identified a new cluster of genes that could serve as nononcogene targets in a selective subset of cancer cells. They have carefully interpreted the collective data from other studies and mechanistically dissected these to map their mechanism and dependency patterns. The present study connects the novel BIRC6 ubiquitin complex with ISR through HRI modulation. One question that remains unanswered by this study is the precise mechanism leading to cell death and loss of clonogenicity in response to BIRC6 depletion in the sensitive cell lines. ATF4, whose dysregulation upon prolonged or unresolved stress has been linked to apoptosis, appears to be dispensable for this effect. Therefore, it would be intriguing to identify other cell-fate mediators downstream of eIF2α phosphorylation.

Targeting the components of the ISR has been extensively reported in the literature (6). It is important to note that inhibition of the ISR and overactivation of ISR signaling have both been shown to reduce cell viability and tumor growth (albeit by distinct mechanisms). The precise role of HRI in cancer is not yet clear. Some studies have shown that HRI depletion blocks the response to chemotherapeutic agents such as proteasome inhibitors in pancreatic tumors (7) and, along with mTOR induction, in multiple myelomas (8). Interestingly, a small-molecule inhibitor of HRI has been developed to increase hemoglobin levels and improve the oxygen-carrying capacity of red blood cells, but has poor bioavailability (9). Other HRI-targeting strategies assume an alternative approach to pharmacologic activation. These HRI activators (i.e., BTdCPU) induce eIF2α phosphorylation in lung, breast, prostate, and melanoma cells without promoting chronic stress or activating other kinases of the ISR (10). In the future, it would be intriguing to test these HRI activators in BIRC6-dependent cell lines, which might have higher levels of HRI and thus sit at a lower apoptotic threshold than BIRC6-dependent cells (Fig. 1). A better understanding of tumor type–specific HRI regulation in combination with BIRC6 module relevance to aneuploidy could establish a new approach to dealing with BIRC6-dependent malignancies.