Current cancer drugs may be hitting the wrong protein targets, a study has found, giving a possible explanation for the low success rates of those drugs during their early stages. 

The study, recently published in Science Translational Medicine journal and led by cancer geneticist Jason Sheltzer of New York’s Cold Spring Harbor Laboratory, used CRISPR technology to determine how anti-cancer drugs react with malignant cells. While the initial study has some limitations, its findings could change the way pharma experts think about drug targeting and finding patients for clinical trials.

The finding cannot come too soon. A study published in April this year found that only 3% of cancer drugs which made it to clinical trial phase in the 15 years up to 2015 were actually approved for human treatment. 

Sheltzer announced that his lab team’s initial findings were purely coincidental: using CRISPR-Cas9, two of its members mutated a gene controlling production of a protein presumed to be important in breast cancer cell division. To their surprise, the edit had no effect on the cancer’s cell growth. 

Intrigued, Sheltzer’s team then analysed ten drugs, including seven currently in clinical trials, that targeted six different proteins. These drugs, which included major players like Celgene’s HDAC6 inhibitors citarinostat and ricolinstat and Pfizer’s PAK4 inhibitor PF03758309, had previously been used in 29 clinical trials of more than 1,000 patients in total. The protein targets of these drugs were collectively implicated in more than 180 publications as being vital to cancer cell survival and growth, or in one case inducing cell death when activated. 

The evidence for these proteins’ importance, however, was based on an older method of testing: RNA-interference, or RNAi, which turn off certain genes but also unintentionally affect the activity of others. 

Through different experiments, including use of CRISPR, the team found that the proteins targeted by each drug were not necessarily the ones vital for the survival of a cancer cell, and that after they had been “silenced” the cancer cells continued to grow. However, even with the removal of their primary target, the drugs were still able to kill cancer cells when utilised. 

On closer study of one of these drugs, OTS964, it was found that the drug in fact affected the cancer cell through a different protein to its intended protein target: it acted on CDK11, involved in cell division. This makes it the first ever drug found to target CDK11, an understudied protein implicated in several cancer types. 

These results could show why so many cancer drugs fail in clinical trials after promising results from RNAi screening. The current method of target selection often involves determining which proteins prevent cancer growth when blocked through RNAi studies, before finding drugs which target that protein. This study shows what a number of scientists have believed for a while: that RNAi screening’s unfortunate potential to block other additional proteins (in addition to their target) could cloud the truth of just what makes a cancer drug effective. Currently, it is unknown how many drugs are being applied to proteins falsely believed to stop cancer cell growth, wasting years of time and millions of dollars fighting a losing battle through their trial phases.

This theory also shows why the drugs tested were still successful in stopping growth, when the target protein had been snipped out: they were in reality acting on other targets. 

This discovery could also limit the effectiveness of pharmaceutical companies selecting clinical trial participants with molecular markers that indicate the patient will receive above-average benefits from particular therapies, a process sometimes done to better achieve regulatory approval. But with less knowledge around the drug’s real target, a lack of understanding could raise the potential for collateral harm to other cells and dangerous side effects. 

There are a number of limitations to this study which should also be mentioned. The studies were carried out in lab-grown cells, not in human bodies, and as such the targets found may indeed be relevant in vivo. There was also scarce prior data on both the drugs and targets examined, and some of the drugs’ reported targets could in fact be influencing the growth of cancer through other cells, indirectly influencing tumour growth. 

Though it is clear that the drugs examined by Shultzer and his team did work, and had a significant effect on cancer when introduced to the cell, greater investigation into this area could have large implications for drugs moving through the R&D stage into market. 

The research done by the Cold Spring Harbor Laboratory points the way for a rethink in how scientists and pharmaceutical professionals develop cancer drugs in future. With this new knowledge, it is important that a greater emphasis is placed on looking for mutations in cancer cells, in order to avoid false positives that lead to an inaccurate assessment of what a drug’s primary target should be. 

Further study could reveal new vulnerabilities in cancer cells that have not yet been targeted, or open the way for better targeting of patient populations for future clinical trials. 

Finally, it is probable that this information will do much to shift impressions of RNAi screening to a more sceptical tone: it is important to remember, however, that the newer CRISPR technology has also been dogged by similar problems, with scientists only now beginning to uncover methods to prevent it from rogue cutting or editing the wrong target. Rather than a simple shift of trust from one technology to the next, this news should if anything prove the need for constant rigour when investigating new targets, no matter what technology is implemented. Hopefully, the only result will be visible improvement in drug efficacy, and a huge rise in the amount of successful drugs that make it to the people who need them.

Joshua Neil, Editor
Proventa International