Interview conducted by April Cashin-Garbutt, BA Hons (Cantab)
What are kinase inhibitors and what functions do they perform?
Kinase inhibitors are molecules that block the activity of kinases. Kinases are a specific class of enzymes. They are extremely important in signal transduction processes in the human body meaning that they actually regulate most of the physiological processes that take place in the body.
There are about 518 different kinases in the body, which is a very large number, so this is one of the larger protein classes. Together they compose what is called the human kinome.
They are extremely important therapeutic targets – drug targets for the pharmaceutical industry - because they belong to a group called druggable targets, meaning that you can interfere or moderate them by using both antibody approaches and small molecule approaches.
I think up to today the estimation is about 20% of the R&D funding in pharmaceuticals is spent on kinases. This is an enormous number.
There are two major ways of approaching kinases; one of them is with monoclonal antibodies. This is mainly for kinases that also have an extracellular part, because only this part is accessible to antibodies, so these are more or less what are called tyrosine kinase receptors. These are molecules that have, in addition to the kinase function, also a receptor function.
The second approach is by small molecules. This is what we do at Oncodesign. Kinases activate or modulate other proteins by transferring a phosphate group. They do this by binding ATP (adenosine triphosphate) at their active site or catalytic site. When their substrate binds, which are also proteins, they transfer one of these phosphate groups from ATP to that substrate. By doing that the substrate, in most cases, changes confirmation and becomes active or inactive. So it is really a way to trigger a response from proteins in the human body.
It is known today that the deregulation of kinases have been linked to over 400 disease indications. This is an enormous number. The deregulation of kinases takes place in two ways. It can happen by mutations in the kinase that lead to a hyperactive kinase or on the other hand a kinase that does no longer function. Or it can be the fact that these kinases are expressed at very high levels, which is typically the case in certain forms of cancer.
Today there are, I believe, 15 kinase inhibitors on the market - all but one in cancer. There are hundreds of molecules that are in pre-clinical development or in clinical development as kinase inhibitors, but mainly in cancer indications.
Are there different types of kinase inhibitor?
Yes. There are two big classes. There are the monoclonal antibodies which are mainly active on kinase receptors. Then there are the small molecule inhibitors, which is what we pursue.
Within these classes there are multiple types. There are 3 major types that exist and have been described largely and then there are Nanocyclix which I would like to introduce as a fourth class.
All of these molecules at least in part bind to the site where ATP binds in the kinases. It is important to understand that this site is very highly conserved among all of the kinases because all of them bind the same molecules – namely ATP.
This leads us to the prime challenge in kinase inhibitory design which is obtaining good selectivity. If one doesn’t have this selectivity then off-target effects can happen. One hits kinases that are not necessarily intended to be blocked and this can lead to major safety issues.
So the different types of small molecule kinase inhibitors are defined as type 1, type 2 and more recently an intermediate type which is called type 1.5.
Type 1 are compounds that only bind to the ATP binding sites in the catalytically active form of kinases. Kinases have a form which is called the active form and also the inactive form. You can think of kinases in a mechanical type of way. So a kinase that is active is one where the active site is open and a kinase that is inactive is where there is a loop of the kinase that actually blocks access to the site.
So the first type of inhibitors binds to the site which is open and only to the ATP binding site and in general selectivity for these molecules is very hard to achieve because it binds to the site which is highly conserved.
There is also a type 2 inhibitor class which, in contrast to type 1, binds to the catalytically inactive form of the kinases which is also called the DFG out conformation. DFG stands for three different amino acids – aspartate, phenylalanine, and glycine. This is a very specific sequence in the kinase that is highly conserved and the fact that it is in or out makes the kinase active or inactive.
So type 2 binding molecule inhibitors also extend in to an additional site which is a hydrophobic site that is also called the allosteric site - by doing that, these molecules tend to be much more selective. They also can have different binding kinetics to the kinase but they have the disadvantage that in order to bind in this way they have to be bigger.
Here we come into a real problem because on the one hand these molecules are potent and selective but on the other hand because they are bigger they have problems with their PK properties. This means it is very difficult to convert them into drugs. Typically these molecules have large molecular weights of between 500 and 600 which is really borderline on what can be used as drug molecules.
The third type, which is type 1.5, is a mix between type 1 and type 2. Type 1.5 bind to the active kinase – so they bind to the ATP pocket – but also to an additional hydrophobic pocket. This is a more recent class that has been identified and there are a number of groups working on these.
There are some additional but relatively smaller alternative approaches. People look to block the substrate site for example. People look to block special pockets that only exist in certain kinases -a company that is doing that is ArQule. Again the goal there is to reach selectivity in molecules.
Finally I think we can say that there are the Nanocyclix, which are Oncodesign’s macrocycles. These molecules bind to the binding site in a very unique way because their selectivity is not coming from specific interaction but it is really because in an ideal way they fill up this active pocket in kinases.
They are composed out of 3 interaction sites. They bind to the hinge region which is also where the adenine part of ATP is binding. They bind to what is called the gatekeeper which is an essential residue in kinases. And, this is quite unique, they also bind to the sugar binding pocket of ATP. So their selectivity, and we have seen this in experimental X-ray structures, is really because the shape of these molecules has a 3D confirmation which ideally fills up the pocket of kinases.
What current clinical uses are there for kinase inhibitors?
Up to the end of 2012 it has mainly been cancer. Cancer is obviously a life threatening disease so due to the fact that it has been very difficult to obtain very high specificity for kinase inhibitors there has been some higher risk for safety issues than in what I would call non-kinase type of drugs. This is why the main indication has been in cancer where some level of toxicity is much more acceptable than in other therapeutic areas.
This has led to a number of breakthrough cancer treatments. I think the first kinase inhibitor Imatinib, commercial name Gleevec, is a very good example of that. People have even used the name of magic bullets in cancer for kinase inhibitors. They also called them targeted cancer agents because these are molecules that really converted the way of treating cancer from more or less broad cytotoxic approaches towards molecular targeted approaches.
So kinase inhibitors have had a major impact and in some cases the lack of selectivity of these molecules has been used as an argument in favour of them because in cancer in some cases one needs to block multiple pathways in order to get an effect. This is because a deregulated cell, like in cancer, very often has redundant pathways so a very specific agent might not be sufficient.
As I mentioned, kinases are involved in all major physiological processes. This means actually they have major potential in all major therapeutic areas. So oncology is clear but there is clear potential in inflammation, metabolic disorders, CNS, cardiovascular. Even in antivirals and antibacterials there is huge potential because there are specific kinases also in bacterial genomes and in viral genomes which we can target by specific kinase approaches.
The first non-cancer kinase inhibitor has been approved by the FDA in November last year. This was JAK 3 inhibitor by Pfizer called Tofacitinib with the commercial name Xeljanz. It is an inhibitor in rheumatoid arthritis.
There is a lot of research activity going on outside of oncology and this was really the proof of concept that a molecule can actually be commercialised also outside of cancer - in this case rheumatoid arthritis. This compound is also currently tested in other inflammatory diseases in clinical development, for example in psoriasis and IBD.
It is clearly expected that in years to come there will be major advances in other areas and cancer, but one needs to address this big challenge of selectivity which is of major importance - especially for chronic diseases – and of addressing appropriate PK. A good example there is how to find kinase inhibitors that can get into the brain. For example, if one thinks about Parkinson’s disease or Alzheimer’s disease.
These for me are the two big challenges – selectivity and appropriate PK for drug use. We believe we are starting get very good evidence that Nanocyclix technology can address those issues.
How are kinase inhibitors designed and synthesized in the lab?
There is a very strong use of structure-based drug design and pharmacophore based modelling on these molecules. There are hundreds of experimental X-ray structures out there for kinase inhibitor complexes. So we have a very good view on how these small molecule kinase inhibitors block kinases. This also provides very good guidance in optimizing these compounds.
The way that Oncodesign finds and designs novel kinase inhibitors is quite different to what is done in big pharma. What is typically done in pharma is that first a new or interesting kinase target is identified – so they start by identification of the target. They do that by using multiple methods in biology. This can be based on the existence of kinase mutations or over-expression in specific disease indications.
It is very often followed by some type of experiment often using siRNA approaches in cells. This can also be done in vivo for example in a transgenic animal which over-expresses or knocks out a specific kinase.
At that stage there is sufficient confidence in pharma companies to actually start the chemistry approach on that specific target. This is typically done by submitting that target on high throughput screening using, in most cases, hundreds of thousands of compounds – so this is typically the large pharma compound library. It can also be a library that is specifically targeting kinases that has already been enriched in molecules that are expected to bind kinases.
This approach typically leads to Hits which are molecules that actually are not necessarily very potent or selective but they provide a starting point for further synthetic optimization which can actually take years of effort from a large group of chemists to get to the required level of potency, selectivity and the required PK properties.
I think it is fair to say that in many cases, these types of molecules, which are the ideal compromise of potency, selectivity and appropriate PK, are not found in quite a number of cases.
Most pharma companies are fishing in the same pond with this approach. First of all because they all work on the same target and that is why only for about half of the human kinome inhibitor patents exist. This obviously means that about 250 kinases are completely unexplored today by pharma.
They also fish in the same pond because if one looks closely they all have similar compound collections. One should know that these compound collections to a large part have been bought in from compound brokers which have sold their compounds in a non-exclusive way. At least up to a number of years ago there were specific groups in pharma that had a mission to expand the compound collection as widely as they could.
So both aspects, that pharma is quite conservative in the choice of targets and that they screen typically very similar compound collections, obviously leads to the fact that half of the kinome today is unexplored. There are nevertheless very strong indications that these could be very good targets for therapeutic interventions. This has been validated by knockout approaches showing clear pathologic potential of these molecules but also by the fact that in these additional targets that are unexplored, there are actually many mutations present that can directly be linked to specific disease pathology.
Please can you outline Oncodesign’s Nanocyclix technology? How does this technology produce highly selective kinase inhibitors?
Our approach is very different. Oncodesign’s Nanocyclix approach is what can be called a chemigenomics approach or a chemical genomics approach. This means we let the activity of our compounds decide on which target we work.
We first synthesize the molecules and once we have those in hand we test them very broadly on the human kinome. It is only once we have a very potent and a very selective compound from the start that we start to check the biology of that specific target. In case of interest in that biology, that’s actually where we start our project.
This has the very big advantage that in our portfolio of targets we already have very relevant chemical starting points that can already be defined as leads, not as hits, because they are very potent and very selective even before we start the programme.
What are the benefits of highly selective kinase inhibitors and are they difficult to produce?
Our molecules have an advantage as they are very selective and very potent. Very selective actually means the potential for fewer side effects. That immediately links to users outside of oncology. I mentioned Parkinson’s disease which is a chronic disease. Treatment needs to start at a relatively early phase and needs to continue probably for multiple years until the end of life. In order to do that these molecules need to be extremely safe. That is a real challenge for the conventional kinase inhibitors today.
A second big advantage of our molecules is they remain very small. We use the word ‘nano’ in Nanocyclix as the molecules are very small despite the fact that they contain large cycles – they are macrocyclic but still very small molecules. On the other hand as a conception they have nanomolar activity so they are very potent from the start.
For me the selectivity and the small size allow us to use them outside of oncology but also allow us to get to next generation kinase inhibitors that have the potential of much better safety.
Our current programmes are not only in oncology. Our first big collaboration, which we signed at the beginning of 2012, is in the central nervous system on a kinase called LRRK2, which is a kinase in Parkinson’s disease. This really shows why our technology is important. This kinase was identified in 2005. Parkinson’s disease is a very important disease which still has very high unmet need. Nevertheless, many companies have started to work on LRRK2 since its disclosure in 2005 based on a genetic mutation in familial Parkinson’s disease.
The number of patents that are out there after 8 years of specific research in all major companies on LRRK2 is less than a handful. This is obviously very poor productivity and the major reasons for that is that typical kinase inhibitors can be made to be very selective and very potent but these molecules typically will never get in the brain.
This is where the small size of our molecules really makes a difference. Nanocyclix molecules are very potent, very selective and we can get them in the brain in very nice levels which are essential to be able to use them in Parkinson’s disease.
Please can you outline Oncodesign’s latest research collaboration with Sanofi? How did this agreement arise?
I have around 22 years’ experience in pharma. From 2000-2008 I headed up European Medicinal Chemistry at Johnson & Johnson. It was in my position at Johnson & Johnson that this technology was actually developed. In 2010 we were able to in-license this technology at Oncodesign.
After 2 years we now have signed 3 major collaborations. The first one was on Parkinson’s disease with Ipsen – the LRRK2 target. The second and the third ones are with 2 different groups at Sanofi.
The first one was an application of our macrocyclic nanocyclix technology on a number of Sanofi kinase targets. This was announced in Sept 2012. The most recent one that was announced just a few weeks ago is a different group at Sanofi looking at one of their kinase targets in tissue reparation and repair.
Obviously we have a longstanding set of collaborations with Sanofi so Oncodesign runs a dual business model, on the one hand we provide service – pharmacology in oncology – and we have also been able to convince their scientists about the potential of our Nanocyclix approach – they had a very nice buy in in the fact that these type of molecules can actually be breakthrough in certain applications in kinase inhibitors.
What are the main aims of this collaboration?
This is a collaboration on a target that Sanofi has been working on for many years. They have found inhibitors but they have had a hard time bringing these inhibitors to a stage where they can start clinical development.
Our goal is to find, within our collection of nanocyclix inhibitors, starting points for this type of target with the hope to be very efficient and rapid with bringing these compounds to clinical development.
Is it true that Sanofi will have exclusive licensing rights to the molecules that result from the collaboration?
They do but this will only be limited to their specific target of interest so it is clear that our Nanocyclix platform is a very large platform. It is a very precise set of molecules. These molecules are composed out of what could be called an ATP scaffold and a macrocyclic linker. We have made combinations of about 50 of those ATP scaffolds with 200 macrocyclic linkers.
What is licensed to Sanofi is a real small sub-part of that which are specific molecules that inhibit their target of interest.
What impact do you think this research will have?
I think for them the hope is that this will allow them to find inhibitors that can be developed for commercial use, which they have not been able to do with conventional kinase inhibitor technology.
In addition, in this case specifically, in tissue repair and reparation, this is a non-oncology application so they are very interested in getting in to very safe inhibitors. I can’t disclose the target of interest as I don’t have permission, but this is a target which belongs to a small family of very similar kinases. It is quite important to not only have selectivity but also sub-family selectivity. We believe that our Nanocyclix technology could be breakthrough technology for that.
What plans do Oncodesign have for the future?
We have lots of plans. We have already identified, using this broad chemogenomics approach, today we have inhibitors that are first-in-class for 7 targets on which we have actually found hardly any inhibitors out there.
These 7 targets already have a very strong biological rationale and validation but seemingly pharma companies have not been able to find inhibitors in their compound sets.
We are talking to multiple pharma companies because they show a very high interest in these new molecules. So Oncodesign is fishing outside of the pool where everybody is fishing. Nevertheless there is potentially very high value there.
Clearly our focus is on unmet need. We want to bring solutions to patients that today have hardly any viable treatment options. That obviously positions our molecules in the first-in-class or best-in-class type of category in all major therapeutic areas.
The nice thing about having these molecules very early on is that we get a very high degree of interest from biology experts on those targets. Typically we would contact them as soon as we identify those molecules. They would take our molecules and they would validate them in their biology models. They do that because these are unique chemical probes for them to validate their science. While doing that they create value for our molecules.
Today we have these novel molecules in about 10 different expert research groups worldwide. We are working with groups in Oxford, in the US, in Japan, in Singapore. Each time we identify those groups because they are world experts on the biology of these targets.
Some of these targets are very exciting. LRRK2 we already mentioned, we were able to partner this very early with Ipsen on Parkinson’s disease.
Another target where there is a lot of interest is RIP2. It is a major target in inflammatory and autoinflammatory diseases with a clear potential in Crohn’s disease, IBD, asthma, arthritis, sarcoidosis, diabetes type 2. This is validated biology so RIP2 non-selective inhibitors have been used in these models in vivo in order to validate the potential of this target.
Today there’s only Glaxo that have 2 compounds that are RIP2 inhibitors in pre-clinical development. Then there’s Oncodesign and we have 20 RIP2 inhibitors that are already very advanced compounds that are very potent in cells. We are very close to candidate selection for this target RIP2 and clearly these molecules have blockbuster potential. So these are unique and we have multiple discussions on-going in relation to these.
A second target that we are pushing quite strongly is a kinase called SIK2, which is actually a kinase that is stress related – it is upregulated in stressful conditions. For example, oxidative stress, which is strongly linked to neurodegeneration for example Alzheimer’s disease.
Nevertheless, these SIK2 compounds also have potential in cancer. Notably, SIK2 is upregulated in ovarian cancer. There’s also an anticipated role in resistance to targeted therapies, so by blocking SIK2 one could actually sensitise cancers to currently existing targeted therapies. There’s also potential in inflammatory disorders and metabolic disorders.
Again a lot of validated biology but no molecules out there other than the ones we have and again we have a whole bunch of quite advanced compounds with very interesting potentials for further development.
The final one I can mention is a unique compound. It is actually linked to TGF-β which is a major target in inflammation and oncology. There are multiple approaches. There are actually compounds that are non-selective for TGF-β R1 and TGF-β R2 which are receptors that are activated by TGF-β. We are the only ones in the world to have compounds that are completely selective for TGF-β R2 versus R1. Again these molecules can also be targeted in resistance to targeted therapy which is actually an important internal Oncodesign platform also in pancreatic cancer.
Why is Oncodesign’s Nanocyclix technology important?
Nanocyclix really takes us out of the pool in which everyone else is fishing. We are actually, in a very important way, opening the kinase field for applications that have been pursued for a while but where it has been very difficult to find molecules. I mentioned the Pfizer compound in Rheumatoid Arthritis actually has taken close to 15 years of kinase research to bring one single molecule to the market. For the reasons I have mentioned which are mainly the safety of these compounds but also the use outside on an oncology setting. And so we have technology that allows us to do that.
It is actually quite amazing that we started this only two years ago when we in-licensed this technology and we already have signed three major collaborations to date and we have 7 novel targets with already very interesting molecules that we are pursuing internally and obviously where we are looking for an early partner.
Why is Oncodesign an unusual company?
We are a really atypical type of Biotech Company meaning that we are not a company that has investors with really deep pockets – that is also why we look for early partnerships. We are a company that runs a hybrid business model meaning on the one hand we have about 50 people in oncology services. Our drug discovery group internally is operating in a semi-virtual way meaning we have 4 people internally. We have a group of about 20 chemists as CRO.
Then obviously we partner with pharma companies that have complementary expertise to ours. Our molecules not only take us into oncology but to all major therapeutic areas meaning that we are looking for partners that can bring that expertise to us. Ipsen was a good example because they are a company that is well-known in neurodegenerative diseases including Parkinson’s. They bring this knowledge to the project and we bring the nanocyclix molecules to the project.
Where can readers find more information?
I recently gave some talks on the kinase drug and the anti-inflammatory drug at the Drug Discovery Chemistry Meeting at San Diego April 16-18 2013 and I was also present at Bio Chicago 22-25 April 2013.
About Jan Hoflack
Jan Hoflack is Chief Scientific Officer and Head of Drug Discovery at Oncodesign, a leading biotechnology company in Dijon, France. Oncodesign runs a hybrid business model, providing on the one hand advanced oncology pharmacology models as fee-for-service, and on the other hand kinase inhibitor based Drug Discovery as shared risk partnerships.
Before joining Oncodesign in 2008, Jan filled senior executive roles in drug discovery and early development at some of the largest pharmaceutical companies at international locations. His professional career started in 1987 at Merrell Dow (now Sanofi) in Strasbourg, France, as a computational chemist and CNS project leader. Following a short stay at Novartis, Basel, Switzerland, where he headed the computational chemistry and structural biology groups, Jan started the Astra Structural Chemistry Laboratory in Gothenburg, Sweden, which had a major impact on drug discovery programs in Astra and later AstraZeneca. He moved on to the role of Vice-President, Enabling Sciences and Technology in Chemistry with global responsibilities for AstraZeneca’s chemistry support groups. In 2000, Dr. Hoflack joined Johnson & Johnson’s Janssen Pharmaceutica in Beerse, Belgium, leading the European medicinal chemistry and enabling biology teams.
Jan has a strong focus on innovative approaches in drug discovery and is a strong believer in platform-based research and alternative business models. He is the originator of the Nanocyclix kinase inhibitor platform at Oncodesign, which in less than 3 years’ time has led to the discovery of potential first-in-class macrocyclic inhibitors for 6 currently unexplored kinases, the signature of 3 major drug discovery partnering deals and the filing of 15 patent applications. The Nanocyclix technology is also successfully applied to the discovery of novel kinase directed PET tracers as diagnostic tools for cancer patient stratification and treatment follow-up. Jan remains very close to the actual science in the kinase programs and has set up international collaborations with leading experts in kinase biology across the world.
Dr. Hoflack has served as a board member of GNI, an emerging commercial biotechnology company in Tokyo, Japan and Shanghai, China.
Jan holds a PhD in organic chemistry from the State University of Ghent in Belgium. He lives with his family close to Antwerp, Belgium, and is a weekly commuter to Dijon, at the heart of beautiful Burgundy famous for its vineyards and its culinary specialties.