Features Synthetic lethality: beating cancer at its own game Jane de Lartigue, PhD he primary focus for targeted cancer agents “back-up” pathway. It is this dependency that could has typically been to counteract the onco- be targeted by SL-directed anticancer drugs, which Tgenic signaling that results from genetic would aim to knock out the secondary, back-up sys- defects. A new strategy is emerging that actually tem to kill the cancer cell. Two synthetically lethal seeks to exploit the oncogenic features of tumor genes can be part of a single linear signaling path- cells rather than overcome them. Synthetic lethal- way, part of 2 parallel pathways that direct a com- ity (SL) is a situation in which 2 nonlethal muta- mon cellular process, or part of 2 independent cell tions become lethal to a cell when they are present survival pathways that compensate for each other, simultaneously. If SL were to be exploited for anti- each one acting as a salvage pathway in the absence cancer therapy, it could lead to the development of of the other (Figure 1).2 highly selective, less toxic drugs, while expanding Te most signifcant advantage of a SL strategy therapeutic targets to include those that have, until for cancer therapy is that it ofers exquisite speci- now, proven pharmaceutically intractable. Here, we fcity because it should kill only cells that harbor a discuss the idea of SL and how it can be applied to certain genetic mutation. It also ofers the opportu- cancer therapy. nity to exploit targets that, until now, have proven challenging, such as tumor-suppressor proteins Exploiting SL in cancer treatment that are not necessarily readily amenable to current Tere are several issues facing current cancer thera- drug development. Finally, it provides the poten- pies. Traditional therapies, that seek to target rap- tial to treat more advanced, metastatic disease that idly proliferating cells, are extremely toxic because has developed multiple mutations and may perhaps they kill both cancer cells and normal cells fairly have become refractory to other treatments.3 indiscriminately. Newer therapies, though more targeted to cancer cells, face the signifcant issue of Identifying SL interactions resistance, which limits their utility. SL interactions are typically identifed by per- SL was frst observed in genetic studies in yeast forming a screen in which, against a background and occurs when 2 separate genes allow for a viable of mutated gene A, a variety of candidate SL part- cell when mutated individually but are lethal to the ner genes (gene B) are mutated to determine which cell when present simultaneously. Te 2 genes are cause cell death. Historically, there have been 3 said to be SL partners. In the late 1990s, research- major approaches to screening (Figure 2). Initially, ers began to examine whether the concept of SL it was performed predominantly in model organ- might be applied to cancer, which is fundamentally isms, such as yeast, worms, and fruit fies, because a disease driven by genetic mutations. Tanks to they were amenable to simple, rapid screens. Te improvements in genome sequencing technology, disadvantage was that an SL interaction demon- many of the genetic abnormalities underlying can- strated in a model organism may not necessarily cer are now known. Te theory was that if SL part- translate into human cells, even when homologues ners for genes that were mutated in cancer could of the genes involved exist. be identifed, then they might present a therapeutic In 2001, the demonstration that RNA interfer- strategy that would specifcally kill cancer cells that ence (RNAi) – the major technique by which gene harbored those mutated genes.1-3 function can be inactivated experimentally – was SL predominantly occurs because many of the also feasible in mammalian cells, meant that SL molecular pathways that control cellular functions screens became possible in mammalian cancer cells. overlap with one another, so that perturbations in Tere are 2 very diferent approaches to testing SL 1 pathway can lead to a dependency on another, interactions in mammalian cells. Knowledge-based JCSO 2014;12:35-39. ©2014 Frontline Medical Communications Inc. DOI 10.12788/jcso.0008. Volume 12/Number 1 January 2014 ■ THE JOURNAL OF COMMUNITY AND SUPPORTIVE ONCOLOGY 35 Features (A) Partial ablation of two enzymes located on one essential pathway Pre-existing Model knowledge organisms XE1 E2 XE3 E4 ABCDessential product (B) Ablation of two enzymes located on parallel pathways leading to a Knowledge-based Cross-species Whole-genome common essential product direct tests candidate approaches approach XE1 ABC essential product XE2 Candidates interactors silenced XYZ (siRNA or shRNA) (C) Ablation of two enzymes on independent survival pathways leading to synthetic lethality XE1 Wild-type (control) cells Experimental cells ABC product 1 E2 X 1. Knockout (eg, tumor suppressor gene) product 2 XYZ 2. Amplification (eg, oncogene) Enhanced killing in experimental cells FIGURE 1 Three modes of synthetic lethality. Synthetic lethal mu- tations may constitute partial mutations present in one essential 1. Colorimetric assays (eg, MTT) pathway, mutations in components of two parallel pathways 2. Imaging (eg, nuclear counts) that lead to the same essential product, or mutations in two 3. Flow cytometry (eg, total number of events) 4. Clonogenic assays (eg, colony formation) independent survival pathways that compensate for one an- 5. Array comparative genomic hybridization other – each one serving as a salvage pathway in the absence 6. DNA shRNA sequencing of the other. Reproduced with permission from Canaani, D. Application of the concept of synthetic lethality toward anti- Putative interacting gene cancer therapy: A promise fulflled? [Published online ahead Confirmation assays of print September 3, 2013]. Cancer Lett. doi: 10.1016/j. canlet.2013.08.019. Candidate drug target Chemical screen Preclinical study direct testing is essentially an “educated guess” that can be undertaken if there is a signifcant amount Clinical study of pre-existing knowledge about genes being tested and SL is already suspected. FIGURE 2 A depiction of the 3 fundamental experimental approaches Alternatively, advances in high throughput tech- typically used to identify synthetic lethal interactions: knowledge-based nology have meant that we can now perform com- direct tests, an ‘educated guess’ approach that requires some pre-exist- ing knowledge of the functions and interactions of the genes being test- pletely unbiased whole genome-based screening. SL ed; cross-species approaches that use animal models such as yeast, fruit interactions can be identifed and validated using fies, or worms, to test genetic interactions more simply and then infer either an RNAi library (testing candidate genes) or a interactions in their homologous genes in humans; and whole-genome small-molecule compound library (testing candidate approaches in which an entire gene or drug library is tested to identify drugs). On the one hand, the advantage of an RNAi potential ‘hits.’ Either siRNA or shRNA can be used to silence the target genes in these approaches. Reproduced with permission from Sajesh et library screen is that it provides direct identifcation of al. Cancer Lett. 2013;5:739-761. a genetic target, so that we know exactly which genes shRNA, short hairpin RNA; siRNA, small intering RNA; are involved in the SL interaction. However, this may MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazolium bromide. not necessarily lead to the development of a therapeu- tic compound as the target may not be amenable to drug development. On the other hand, a small-mol- ecule compound screen directly identifes drugs that could be J Inactivated tumor suppressor gene plus inactivated second used to generate SL in cancer cells, the disadvantage being gene. 1,3-5 that we may not fully understand the mechanism of action. J Inactivation of 2 components of same signaling network. J Inactivation of a pair of genes involved in DNA repair or Targeting SL for cancer treatment synthesis.4 Tere are several scenarios in which SL can occur in cancer cells (Figure 3): Targeting tumor suppressors and oncogenes J Activated or overexpressed oncogene plus inactivated Te oncogene Ras is estimated to be mutated in about a second gene. quarter of all cancers and therefore represents a signifcant 36 THE JOURNAL OF COMMUNITY AND SUPPORTIVE ONCOLOGY ■ January 2014 www.jcso-online.com Synthetic Lethality target for cancer therapy but, until now, it AB has not proven easily druggable. SL screens Oncogene Gene A Tumor suppressor Gene B are therefore underway to see if this strat- egy could allow us to indirectly target Ras KRAS Activating Inhibition, inactivation Mutation Inhibition, inactivation in tumors. So far, the gene has been mutation mutation (loss of function) mutation found to have SL with genes like cyclin A2, (loss of function) (loss of function) kinesin-like protein 2C, polo-like kinase 1 (PLK1) and the anaphase-promoting com- plex. Since these genes encode proteins that regulate mitotic cell division, this suggests C D that KRAS-mutant cells may be particularly Oncogene Gene C vulnerable to perturbations in mitosis and Kinase D Kinase E could be preferentially killed by drugs that afect this process, such as paclitaxel. PLK1 Overexpression Inhibition, inactivation inhibitors, such as BI2536, are also in clinical mutation (loss of function) Kinase F development and could be tested in KRAS- mutant patients.3 Te epidermal growth factor receptor (EGFR) is already a validated cancer thera- Kinase G Kinase H peutic target, with several EGFR inhibitors E DNA Inhibition, inactivation DNA repair genes/ mutation and monoclonal antibodies approved or in (loss of function) repair genes DNA synthesis- development. An SL screen of the EGFR related genes protein network was recently carried out to Inhibition, inactivation determine if EGFR inhibitors could be used mutation in an SL capacity. SL interactions included (loss of function) protein kinase C and aurora kinase A.