A progressive approach to drug development (the 5Rs) is one of the elements of the story of the development of osimertinib. By adhering to the 5Rs (selecting the right target, patient, tissue, safety and commercial potential), the expert team that developed osimertinib achieved a very fast clinical development time for an oncology drug.
With the emergence of new knowledge about the T790M mutation from the clinical research community, chemists saw the importance of drawing on AstraZeneca’s expertise in tyrosine kinase inhibitors. The details on what it takes to usher in a promising therapy through clinical trials quickly include the decision to develop AstraZeneca’s first oncology drug that irreversibly binds to its therapeutic target.
Attaining the right safety of the molecule was particularly important and one of the challenges the team had to overcome before the selection of osimertinib.
Mitigating the safety risks pre-clinically contributed to the rapid pace of the registrational trials. The FDA granted drug approval for the treatment of anti-EGFR resistant non-small cell lung cancer on Friday 13 November 2015.
“I believe this is among the shortest time periods that a cancer drug programme has ever gone from first dose in man to drug approval,” says Mireille Cantarini, Executive Medical Science Director at AstraZeneca. What she is referring to is an R&D programme involving an AstraZeneca oncology molecule, osimertinib, for treatment-resistant non-small cell lung cancer.
Designed to target the T790M mutation of EGFR, osimertinib defines a new generation of targeted epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs). Growth factors are effective therapeutic targets for treating cancer because when mutated they drive the uncontrolled cell division that helps to cause cancer. “The therapeutic benefits of first generation EGFR-TKIs last for a year or so. Then cancer returns in a form that is resistant to currently available treatments,” explains Dr Cantarini.
That is because over time some lung cancer cells can develop another mutation to circumvent the anti-cancer activity of the first generation EGFR inhibitors. “It in effect changes into a different cancer,” Dr Cantarini describes. “In patients who have developed resistance to EGFR-TKIs, we have tested to see what might be causing the resistance. The T790M mutation is the one most commonly identified.”
The emergence of the T790M mutation meant treatment options for lung cancer patients treated with first generation EGFR-TKIs were extremely limited. The launch of osimertinib could change that and is an achievement in terms of speed of development and design of the drug. Did the introduction of a new framework for drug development help improve AstraZeneca’s approach?
Based on a review of its drug development track record from 2005 to 2010, AstraZeneca characterised the qualities associated with the successes and the pitfalls of drug failures. The effort yielded the 5R framework, devised to guide successful and efficient drug development.
The 'R' stands for ‘right’ — determining the right therapeutic target, designing the drug to reach the right tissue, selecting the right patients, achieving the right level of safety, and determining the right commercial potential.
“When we proposed the original ‘EGFRm’ project which led to the discovery of osimertinib in 2009, the 5Rs strategy was still being developed. However, some of the learnings from the 5Rs were consistent with the strengths of the T790M-targeted opportunity — we were clear from the outset on what the appropriate activity profile for an effective therapy would be and who the appropriate patients would be,” outlines Richard Ward, Principal Scientist and Computational Chemist.
The framework was implemented in time to support osimertinib in the later stages of preclinical development. This meant the project was provided with extra resources after being given ‘must-win’ status around lead generation under the 5R framework. “The additional resource helped the programme move more quickly,” comments Dr Ward.
Discussing the 5Rs' influence on the other end of the drug development path, Sue Ashton, Team Leader comments, “The innovation I saw stemming from the introduction of the new framework was the clinical team engaging much earlier in the process. We worked alongside the other teams and that meant we could be more innovative in refining the clinical plans and get them approved much more quickly.”
While a framework can go a long way towards integrating and aligning the different aspects of drug development, osimertinib’s route from discovery to approval also includes a few twists.
The osimertinib story begins with exacting chemistry. "For me, the need and opportunity to target T790M emerged from internal discussions with then Chemistry Director, Andy Barker, and ICC's Xiaolin Zhang around how we could use our existing knowledge in tyrosine kinase biology,” says Darren Cross, the lead bioscientist on osimertinib. “An important conversation was triggered during our internal ‘big science day’,” elaborates Dr Ward. “Myself and Darren had worked together on TKIs and started talking about how we could use our existing expertise to target the T790M form of EGFR.”
TKIs, or tyrosine kinase inhibitors, define the specific drug class to which osimertinib belongs. Tyrosine kinases are enzymes that serve as the on/off switch for many cell functions. Their link to cancer cell growth and expansion and many other cell processes make tyrosine kinase receptors a therapeutic target for treating disease1,2. AstraZeneca was one of the first pharmaceutical companies to explore the development of TKIs. Consequently, the company has an extensive chemical library of TKIs.
“The target was known,” explains Dr Ward. The emergence of a T790M mutation had been anticipated from lab experiments, and was then confirmed in DNA from the tumour tissue of relapse patients4. From the outset, we were focused on what the profile of a suitable compound would be, specifically a 'mutant-selective profile'. In other words, a compound more active against the mutant form of EGFR than the wild-type.”
Previous experience with the development of EGFR-TKIs also provided a steer. “The advantage of the EGFR programme was we knew the limitations of first generation compounds,” explains, Lead Medicinal Chemist, Ray Finlay. Diarrhoea and a skin rash are among the most common side effects. “Over time, we understood that the toxicity associated with older EGFR inhibitors was due to them binding wild type [non-mutated] EGFR receptor. The profile of the required compound was clear: inhibit the mutant receptor and have a margin to wild-type.”
Assays to screen for wild-type binding versus mutated EGFR were used to obtain more information about the possibility. “We picked 40 compounds from our corporate collection, which contains over 1 million compounds. When we got the data back, it was clear that there was something there. Two compounds stood out as having a mutant-selective profile, the seeds from which osimertinib was discovered,” says Dr Ward.
“Being able to initially test such a small number of compounds was only possible due to our historic in-house experiences with kinase-targeted drug discovery projects. Without this experience we would have required a more conventional approach, potentially involving the testing of over a million compounds and delay of up to a year,” Dr Ward added.
Overcoming the drug resistance conferred by the T790M mutation warrants an alternative approach – covalent or irreversible binding of the drug rather than reversible binding. “We showed early on we needed to go down the irreversible road,” explains Dr Cross.
A step forward in the project was when we managed to evolve the early reversible mutant-selective inhibitors into irreversible (covalent) inhibitors – helped in part by computer modeling and knowledge of the shape of the adenosine triphosphate (ATP) pocket of EGFR. “We designed a handful of targeted compounds and saw an increase in potency of up to 100-fold, whilst maintaining a margin to wild-type in cellular systems, this was a really exciting moment” says Dr Ward.
“The search for potential candidates had moved quickly but the 'Eureka' Moment came in converting compounds from our earlier programme to covalent, irreversible inhibitors using all of that knowledge we already had. It was pretty exciting when we saw that potent cellular activity along with selectivity over the wild type form,” adds Dr Cross.
The Eureka Moment came in converting compounds from our earlier programme to covalent, irreversible inhibitors using all of that knowledge we already had
This was the first oncology project within AstraZeneca to intentionally design irreversible covalent inhibitors.
“The high ATP affinity of the T790M mutation is a big barrier for an ATP-like competitive compound to overcome.” The threshold to which Dr Cross is referring relates to how the T790M mutation changes the chemistry of the receptor and confers resistance to the first generation of anti-EGFR therapies.
The previous generations of EGFR inhibitors kill cancer cells by preventing ATP (an essential energy source for living cells) from binding to a specific binding region for ATP on the epidermal growth factor receptor. This keeps the receptor in an inactive state, preventing it driving cancer cell growth and survival. In chemistry terms this is known as competitive inhibition. The T790M mutation edges out the ATP-competitive inhibitors by increasing the receptor’s affinity for ATP.
“We decided on an irreversible mode of action in order to overcome the high ATP affinity of the T790M mutated EGFR,” Dr Cross comments. The greater affinity for ATP is conferred when the mutation substitutes a methionine (M) molecule for the amino acid threonine (T) at position 7903.
“The patient experience is a critical part of safety,” explains Mark Anderton, Discovery Safety Specialist. “Working with covalent binding, we were mindful of the potential safety concerns. It can theoretically increase the likelihood of an idiosyncratic toxicity.” Potential drug candidates undergo rigorous pre-clinical assessment to ensure they assert the desired activity in a way that is likely to be safe when introduced into humans.
“We were trying to drive for a very specific profile that hit not just the resistance mutation [T790M], but also that hit the sensitising EGFR mutation and had a big margin of selectivity from the wild-type receptor.”
“We had to understand the profile of every compound across three cell types,” Dr Anderton emphasises. The triple assay assessment demonstrated the potential for promising anti-tumour activity against mutant receptors while sparing wild-type EGFR that can cause skin rash and diarrhoea.
While testing the medicinal chemistry team’s shortlist of drug candidates, the safety team identified cross-reactivity against the insulin receptor (InsR) and insulin-like growth factor receptor (IGF1R). Dr Cantarini elaborates, “The toxicity associated with InsR and IGF1R inhibition is the development of diabetes in patients.”
Due to the extent of IGF1R and InsR inhibition, the team decided to abandon development of their early lead compounds. “It was a big decision to make,” Dr Cross recalls.
Once the studies from the discovery safety group made it clear that the inherent IGF1R/InsR activity of this template was an issue, the chemistry team focused on how this activity could be tempered through chemical changes to the inhibitors.
“It was highly challenging as the ATP binding pocket of IGF1R/InsR are similar to EGFR, so trying to pick this apart in such tight timescales was challenging,” commented Dr Ward. “The team, however, did manage to identify areas of the molecule where we could reduce IGF1R/InsR activity while having less impact on mutant EGFR.”
“Based on this knowledge, we designed and synthesised a small set of key examples, and this rapid chemistry effort delivered osimertinib,” reflects Dr Finlay.
“Despite promising efficacy with lead compounds, the team was receptive to the early safety concerns and, despite the momentum in the project, decided to spend additional time in the discovery phase to improve the safety quality," describes Dr Anderton. “It was a collaborative effort between bioscience, safety and chemistry that enabled the identification of compounds without the IGF1R/InsR inhibition and ultimately osimertinib. Identifying and mitigating undesirable safety concerns is exactly what we aim to achieve within the Right Safety criteria of the 5R framework.”
“It was a good call to continue chemistry. Osimertinib was in that later work and was the last compound that made it onto the shortlist,” describes Dr Cross.
This care over the safety profile early in discovery contributed to the speed at which the clinical trials could take place. During the clinical trials, most patients who experienced side effects had mild cases of diarrhoea or rash2. "Other toxicities, such as diabetes, would have been less familiar to oncology clinicians, and may require additional monitoring and drugs to control elevated blood sugar levels. It could also limit how high a dose of drug could be given in patients and ultimately efficacy,” Dr Anderton comments.
In 2013, the US FDA had announced its willingness to approve cancer drugs on the basis of fewer and smaller studies. “The FDA indicated that in ‘areas of high clinical unmet need’, they would consider an initial registration strategy based on single arm studies. They made these comments several weeks before we first planned to discuss the initial clinical data from osimertinib with them,” remarks John Freeman, Director, Clinical Development at AstraZeneca. The regulatory shift, meant to speed up access to promising experimental therapies, was a response to pressure from patient advocacy groups. The change in policy, along with foresight applied when designing the trial, contributed to the speed of osimertinib’s regulatory approval.
“We went straight to the specific patient population that could actually be given some benefit [in order] to answer the key question ‘does the drug work in patients whose tumours harbour the T790M mutation?’” describes Dr Freeman.
“Based on the pre-clinical data, the clinical team were able to be bold in designing a Phase 1 study in the targeted patient population and started at doses that were likely to be efficacious.” The design of the osimertinib AURA Phase 1 took place 15 months prior to the FDA’s decision and the protocol for the registrational, single-arm Phase 2 component of the study was established a few weeks before the FDA announcement.
The decision to conduct the Phase 1 trial on a very specific patient population also raised concerns about patient recruitment. “The question was, if you significantly narrowed down the patient population, would you be able to recruit patients?” elaborates Dr Freeman. The choice to conduct the clinical trials in Asia, concurrently with the Western trials, represented another progressive deviation from the old habits. Traditionally, a bridging study is conducted in Japan after the initial tolerability and pharmacokinetic data has been obtained in Western patients. “In the past, there have been drug development programmes that have revealed differences in the way, for example, Japanese patients metabolise a drug, compared to Western patients. We decided to assume that osimertinib would not be affected by differing ethnicity and ensure both Asian and Caucasian patients were recruited to AURA.” The international quality of the trial also meant the trials proceeded 24 hours a day. “There was always someone working on the study somewhere in the world.”
“In the second patient to ever take osimertinib, we saw a significant reduction after 6 weeks [in tumour load],” states Dr Freeman. “We knew the patient also had the T790M mutation. As it emerged that other patients were having similar responses to osimertinib, the issue for AURA study became a good one – many doctors wanted to put their patients on the study and complained when we couldn’t give them enough slots on the trial for all their patients.”
“I have put many drugs in the clinic and many failed in the development. This marks the high point of my career and this framework should now be the standard way we develop drugs.”
After the trials, osimertinib was granted regulatory approval by the FDA in the US on Friday 13 November 2015.
- Pao W, Miller VA, et al PLoS Med. (2005) Mar: 2(3):e73.
- Blencke S, et al, Chemistry and Biology (2004) May: 691-701; and Kobayashi S, Boggon TJ, et al. N Engl J Med (2005) Feb: 24;352(8):786-92.
- Siegelin M and Borczuk A. Laboratory Investigation(2014) 94:129–137