The MTH1 Story
Fueled by the responses obtained in cancer patients with PARP inhibitors, we and many others were very enthusiastic about identifying novel synthetic lethal approaches for treatment of cancer. However, soon after the initial responses in patients we were also reached by the news that resistance to PARP inhibitors emerged in a similar manner as seen with other types of targeted therapies, aiming to target the cancer genotype. Hence, targeting the genotype by the synthetic lethal concept is poised by the intra-tumour heterogeneity, genetic buffering, and resistance mutations also limiting the efficacy of traditional oncogene targeted therapies. A new approach was needed.
High level of DNA damage in cancer – a novel target
It was for a long time well established that cancer cells had a high background level of markers of DNA damage such as γH2AX, which could be related to genetic instability in cancer. In the very same issue of Nature that we demonstrated the BRCA-PARP concept the labs of Jiri Bartek, Thanos Halazonetis and Vassilis Gorgoulis showed that the high level of DNA damage in cancers triggered the p53 tumour barrier, preventing the cancers to grow out. They contacted us and were on track that oncogenes could cause replication stress as a general mechanism to cause DNA damage and senescence in cancer. Natalia Issaeva in the lab helped them validate the replication stress and we published this soon after in Nature.
These data really demonstrated that cancers have a high load of damage at replication forks, similar to that caused by chemotherapy. We were playing with the thought that what if we can stop the repair of these chemotherapy-like DNA damage, then we can find a treatment that can kill the cancer cells and not kill normal cells. This was not a unique idea in our lab and now there was a race to find the way how to selectively kill cancer cells owing to high load of ‘replication stress’. Here, Oscar Fernandez-Capetillo and others did a fantastic job and demonstrated that cancer cells became dependent on the ATR and Chk1 kinases for survival and that they can be targeted by small molecules.
Birth of MTH1 as a target for cancer treatment
The ATR concept was interesting, but both ATR and Chk1 are essential proteins and inhibitors may have off-target effects on normal cells. Ideally, we would want to have a DNA repair protein that is non-essential, and that is required to repair the replication damage specifically occurring in cancer cells. In our lab, Cecilia Lundin assembled an siRNA library consisting of about 200 DNA repair genes that we used for screening. MTH1 was one of the hits coming out of the screens and what really caught my attention was how interesting this enzyme was. It really ticked all boxes for a perfect DNA repair target for anti-cancer treatments and the publications in the literature really supported the idea that damage on dNTP pool is very important in cancer. Helge Gad in the lab set out to validate the concept.
At the time our lab was a basic molecular biology lab, but I was really keen to pursue this as a target. MTH1 belonged to the Nudix hydrolase family and the biochemical assays described used HPLC, which is not suitable for screening compound libraries. Pyrophosphate is one of the products released from the reaction catalysed by MTH1 and then I came to think about my father’s second cousin, Pål Nyrén, who when bicycling came up with the idea to exploit the release of pyrophosphate as a way of sequencing (this became pyrosequencing). I thought that we could use the same reaction as an effective assay for MTH1 screening, which worked very well.
At this time, we met up with the Chemical Biology team with Thomas Lundbäck and Annika Jenmalm Jensen, who were just in the process of starting up a core facility with equipment donated by Biovitrum AB. They helped with improving the assay to allow screening of larger libraries. To obtain a real MTH1 inhibitor, we would need more chemistry and I hired two fantastic medicinal chemists Martin Scobie and Tobias Koolmeister, and placed them at the organic chemistry department at Stockholm University. It is very unusual for a molecular biology lab to hire its own medicinal chemists, but turned out fantastic and the two of them really made a huge effort into optimising the hits into potent MTH1 inhibitors.
The open innovation approach
We had the protein and also a potent inhibitor, so we approached Pål Stenmark, a structural biologist at Stockholm University, who quickly solved the crystal structure of MTH1 with the inhibitors. To optimise the compounds we needed to know pharmacokinetic properties and started collaborating with Per Artursson at Uppsala University. We now had really promising compounds but needed to know their properties in animals. I managed to persuade Camilla Gokturk, who runs a company In vivo Design AB, to start working with us part time to do PK/PD work. We also needed to measure the compounds in plasma and we started a nice collaboration with the analytic chemists at Stockholm University, Ingrid Granelli and Alonja. At this stage, the project was growing exponentially and a highly skilled project leader, Ulrika Warpman Berglund came onboard the MTH1 project, which made a huge difference. We continued our open innovation approach and even before publications, many many labs around the globe had the MTH1 inhibitors.
For us, the MTH1 project has demonstrated how powerful interdisciplinary collaboration can be and that an open innovation model is preferred.
During our work we were contacted by Kilian Huber and Giulio Superti-Furga in Vienna. They had identified MTH1 when screening for targets of a compound selectively killing cancer cells. As we already had an MTH1 assay up and running, Ann-Sofie Jemth went to Vienna, in an open innovation approach, to test if their compound was indeed an MTH1 inhibitor. It turned out that indeed it was an inhibitor and then we could use all the resources we had to also progress and validate Kilan’s and Giulio’s discovery.
Cancer Phenotypic Lethality
We believe the MTH1 story exemplifies that a normally non-essential enzyme can become acutely required for cancer cell survival, hence the synthetic lethality with the cancer phenotype. This concept could provide many advantages over current therapy, targeting the genetic defects in the cancer (see review for more details). The Cancer Phenotypic Lethality approach denotes targeting a specific enzyme, which is normally non-essential but that becomes generally essential in the cancer phenotype.