The Maternal Embryonic Leucine zipper Kinase (MELK) is an interesting enzyme:
In 2005, MELK was first implicated in cancer by finding elevated levels of MELK RNA expression in many human tumors (REF1). Since the initial article, another 32 articles have been published that specifically implicate MELK as a therapeutic target in human cancer, including ovarian, breast, NSCLC, SCLC, AML, prostate, pancreas, gastric, renal, astrocytoma, glioma, medulloblastoma, colorectal, liver, and rectal cancers.
23 of the published MELK studies reported efficacy of MELK-targeted RNAi in cancer cell lines and/or xenograft tumor models, thereby directly implicating MELK function in cancer cell proliferation or tumorigenic potential. 15 studies made use of the MELK inhibitor OTS167, which is currently being evaluated in human clinical trials, presumably prompted by many of these published pre-clinical findings.
Given all the evidence for a likely role of MELK in a variety of human cancers, a pair of student scientists, Ann Lin and Chris Giuliano first doubted themselves when using a CRISPR/CAS9-based approach to knockout MELK in 13 different cancer cell lines and showed that the treatment has no effect on their growth (REF2).
Importantly, they also designed the following ‘killer experiment’: They compared the sensitivity of wildtype and MELK-/- cells to OTS167. If the antiproliferative activity of OTS167 was primarily due to MELK inhibition, then MELK-/- are expected to exhibit reduced sensitivity to OTS167 treatment. However, the scientists did not observe any differences in cell viability after treatment with OTS167, indicating the observed cytotoxicity following OTS167 treatment reflected an off-target, MELK-independent mechanisms.
Subsequently, new potent MELK inhibitor compounds with a greatly improved selectivity profile have been developed, which showed that acute MELK inhibition did not affect the proliferation or growth of several tested cancer cell lines previously reported to be MELK-dependent (REF3) – implying that the ongoing clinical development of OTS167 as a MELK-targeted therapeutic may be misguided.
What about the previous RNAi experiments showing MELK-targeting siRNA impaired growth? Since some studies were able to rescue the antiproliferative activity observed in MELK knockdown cells by re-expressing shRNA-resistant MELK (REF4), it is highly likely that the potential off-target of shMELK might only manifest its effect in the presence of MELK knockdown, a so-called ‘synthetic lethal’ interaction (a synthetic lethal interaction occurs between two genes when the perturbation of either gene alone is viable but the perturbation of both genes simultaneously results in the loss of viability).
Trying to validate potential drug targets, the described example can serve as a reminder to the research community of the issues that can arise when RNAi or pharmacologic tool compounds are used and critical specificity controls are neglected/missing – even when several independent studies all reach similar conclusions.
In our case studies we typically look at Good Research Practice-related topics but this example reminds us that good research is not only about proper study design but also that high-quality research tools are needed. Moreover, given that cancer patients are currently being treated with OTS167 based on published findings highlights the importance of robust preclinical target as well as tool validation.