Exelixis (EXEL) came under a lot of pressure recently and the stock is now trading near its 52 week low following two unrelated events that took place last month. In late June, the company announced the resignation of its CEO, George Scangos, who was appointed as Biogen Idec’s (BIIB) CEO. Earlier that month, the company announced BMS (BMY) decided to give back rights for Exelixis’ flagship product, XL184. Adding to the pressure are interesting but not stellar clinical results for some of the company’s compounds at ASCO and potentially an additional partnership termination.
With respect to future development of XL184, Exelixis is in a delicate position. On the one hand, it would like the drug to have an aggressive development program, which requires a large partner. On the other, it might want to wait until it generates more data, due to the bad sentiment around a compound that has been abandoned twice by two large pharmas (BMS and GSK). XL184 faces some meaningful potential value creation events in the coming 12 months, including data that could support FDA approval in a niche indication. The company will therefore have to find the optimal point in time for a licensing deal that will not slow down the development process while enabling it to increase the value of the compound.
Regardless of XL184’s fate, Exelixis has much more to offer, with a pipeline of 10 drugs in clinical development and one of the top discovery engines in the industry. With ~$150M in the bank, a leaner organization and a flow of potential milestone payments, I wouldn’t give up on the stock.
In some cases, a CEO resignation is seen as deserting a sinking ship but this is unlikely to be the case. Biogen Idec is one of the few independent commercial stage biotechs with a broad pipeline and operational profit so in that sense, Scangos was simply made an offer he can’t refuse.
The more worrying event was getting back the rights for XL184 from BMS. This comes less than two years after BMS licensed XL184 as part of a monster deal in December 2008. Since I could not find any public comments from BMS, the information in the public domain comes exclusively from Exelixis. In its press release, Exelixis claims that the two companies could not agree on what the appropriate clinical program would be. In other words, BMS was not willing to pursue XL184 as aggressively as Exelixis wanted.
In drug development there is always an issue of pipeline prioritization, even in large companies such as BMS. The cost of running a late stage clinical program that includes multiple phase III trials in oncology starts at hundreds of millions. As the winner of ASCO 2010, BMS has its hands full, with multiple agents that are either in phase III testing or expected to enter phase III in the coming years. The acquisition of Medarex added additional projects that need to be financed, including one of the most promising antibodies in development, MDX-1106.
Therefore, XL184 had to compete internally with several promising agents. What made the competition even harder was the fact that BMS is developing Brivanib, which has some overlapping activities with XL184. XL184 inhibits several targets, including MET, VEGFR2 and RET. Brivanib also targets VEGFR2 plus a couple of other kinases.
One could argue that the two compounds should have a different spectrum of activity and their respective development programs are quite different: Brivanib is in 4 registration trials in liver cancer whereas XL184 is in a pivotal trial in thyroid cancer and expected to be in two phase III studies in GBM (brain cancer). In less advanced trials, the two drugs are being evaluated in overlapping indications so the potential for brivanib to compete directly with XL184 clearly exists. At the end of the day, XL184 was not good enough for BMS based on the available clinical data in light of the brivanib data.
Investors should also be worried about the fate of Exelixis’ second MET inhibitor, foretinib (XL880) which was outlicensed to GSK three years ago. The situation with this compound is disturbingly similar to XL184: Foretinib also inhibits both MET and VEGFR2; besides interesting data in a niche indication, results to date are disappointing and, most importantly, GSK has its own approved VEGFR inhibitor. If GSK returns the compound to Exelixis it should come as no surprise.
Promiscuous vs. specific
This story demonstrates the problematic nature of multi-targeted kinase inhibitors, as it is hard to attribute clinical activity to the inhibition of a single target. For example, it is still unclear whether the activity seen with XL184 is a result of MET inhibition, VEGFR inhibition or both. Theoretically, although Exelixis hails XL184 as the most advanced MET inhibitor in the clinic, it might be “just another VEGFR inhibitor”. In that case, it will be one in a long list of approved and investigational agents.
As I discussed in my first Exelixis article two years ago, there are basically two approaches for developing kinase inhibitors. Some claim that targeting several kinases with one compound could have a synergistic effect that is more suitable to deal with the complexity and redundancy in cancer. The other camp is led by companies such as Arqule (ARQL) and Plexxikon, which focus on more selective compounds. The main claim they make is that targeting multiple targets with the same molecule limits the ability to reach an adequate dose and to combine it with other drugs due to side effects.
In certain cases, targeting a single kinase has a dramatic effect, as shown with Plexxikon’s BRAF inhibitor (PLX4032), which has demonstrated stellar activity in a predefined subset of melanoma patients harboring BRAF mutations. This is not the case, however, in the majority of patients, whose tumors rely on a complex and dynamic signaling network. For example, the same drug had a far less impressive effect in colorectal cancer patients with mutated BRAF. Therefore, despite the growing understanding in tumor biology and the ability to distinguish between patients based on molecular markers, both sides agree that a single hit is not enough. The difference is whether to combine different selective compounds or target several kinases within a single drug.
Accordng to Plexxikon, it was PLX4032’s selectivity that enabled investigators to push the dose high enough up to the clinically relevant level. At the AACR annual meeting this year, Plexxikon’s chief medical officer, Keith Nolop, explained why PLX4032 succeeded where other RAF inhibitors, particularly Nexavar (a multi-targeted agent), failed in melanoma. The selectivity of PLX4032 was twofold: selective inhibition of BRAF versus other kinases and selectivity for the mutant form of the enzyme over normal BRAF. According to Nolop, this rendered the drug so safe that in animal toxicology studies they could not reach dose limiting toxicities. He concluded by saying that the days of promiscuous multi targeted kinase inhibitors are numbered, which is definitely not something everybody in the industry will agree on.
Exelixis vs. Arqule
In the field of MET inhibitors the differences between the two approaches were illustrated by Exelixis’ XL184 and Arqule’s ARQ 197. As a single agent, XL184 looks more potent in terms of its ability to shrink tumors, which could be a result of inhibiting not only MET but also VEGFR. As its clinical program advanced, XL184’s promiscuity turned out to be a burden with respect to safety.
In the ARQ 197 trial, both ARQ 197 and Tarceva were given at their single agent doses with no meaningful safety signals. In the XL184 trial, doses of either XL184 or Tarceva had to be reduced due to side effects. Even as a single agent in the brain cancer trial, Exelixis had to decrease the dose due to side effects that led to frequent treatment discontinuations. Arqule was also able to combine ARQ 197 with other drugs, including chemotherapy agent Gemzar at the single agent dose, which is another testament for the drug’s good safety profile.
In general, one can conclude that a selective compound enjoys greater flexibility in terms of combination regimens. As the activity spectrum and ratio of these compounds is fixed, it is impossible to “turn off” or modulate a portion of their activity on demand. It is easier to mix and match selective compounds but it is also more expensive and requires more clinical testing. As single agents, multi-targeted agents still have a lot to offer, and sometimes having more than one activity can save a compound from the abyss of attrition.
This was the case with Pfizer’s (PFE) Crizotinib (PF-02341066), which is approaching an almost certain FDA approval. Crizotinib started as a non-selective MET inhibitor with an “add-on” activity against a relatively unexplored target - ALK. A small minority of lung cancer patients have tumors that are highly dependent on ALK, making them extremely sensitive to the drug. Although the mutation is quite rare, crizotinib market opportunity is anywhere between $0.5B to $1B globally.
Gone till November
Although Exelixis’ flagship compound has not lived up to expectations so far, the jury is still out on their long term potential. Later this year at the AACR-EORTC conference which will take place in November, Exelixis will present data from a large randomized discontinuation trial that already generated activity signals from its non-randomized portion. Data from the randomized portion will be crucial in evaluating the drug’s potential going forward as the first placebo controlled data in multiple indications.
Another important event is data readout from a pivotal phase III trial in a subtype of thyroid cancer in the first half of next year. Positive results could lead to FDA approval in this orphan indication, which, despite the small market opportunity, could be an important value creation event. The potential market is probably in the ~$20-40M range.
On top of XL184 and XL880 (assuming GSK will return the rights to Exelixis), the company has plenty of irons in the fire. The most important assets are the two PI3K inhibitors (XL147 and XL765) which Sanofi (SNY) licensed in May 2009. Unlike the MET compounds, Exelixis pursued the selective approach with one of the compounds (XL147) and the non-selective approach with the other compound (XL765). Compounds that target PI3K can be found in the pipelines of several companies including Novartis (NVS), Roche (OTCQX:RHHBY) and GSK.
Exelixis’ compounds, which are the most advanced PI3K inhibitors in development for solid tumors, demonstrated a good safety profile alone or in combination with other drugs with minor signs of efficacy in their phase I trials. This kind of activity is representative of practically all PI3K inhibitors in solid tumors so far (even though Novartis’ drug looks more active at first glance), which imply that PI3K inhibition should not be pursued as monotherapy for unselected patient populations. One exception is Calistoga’s CA-101, a selective for a specific subtype of PI3K that had very impressive activity in lymphoma patients. It will be interesting to see how Calistoga’s approach pans out with other isoform selective inhibitors. Exelixis is developing a similar drug to CAL-101, which is currently in preclinical stage. Importantly, unlike the its MET inhibitors, Exelixis’ PI3K compounds are not in the risk of ending up as me-too drugs.
In light of the mild efficacy profile of PI3K inhibitors, the industry has moved to the next stage of identifying predictive biomarkers for response and finding the right drug combinations. The hottest trend in the industry right now is combining PI3K inhibitors with RAF or MEK inhibitors. This trend was one of the reasons Array (ARRY) managed to get such a lucrative deal from Novartis for its MEK inhibitor. Oddly enough, according to clinicaltrials.gov, Novartis plans to initiate a clinical trial for its PI3K compound in combination with a MEK inhibitor from GSK, even though it has access to Array’s compound. From this angle, Exelixis’ and Sanofi’s position might be inferior, as Sanofi does not have a RAF or MEK inhibitor in its clinical pipeline. Exelixis has developed inhibitors for both RAF (XL281) and MEK (XL518) but both are already out-licensed to BMS and Genentech (DNA), respectively.
In search of the “killer app”
In summary, despite the recent setbacks, Exelixis’ potential is still obvious. It has one of the best discovery engines in the industry with new candidates every year. Exelixis’ pipeline is still very broad, with ten drugs in clinical development by the company or its partners (including 4 compounds in development in the hands of BMS). The only thing missing now is identifying the “killer app” for one or more of these compounds that will help investors assign real value to the drug. The term “killer app” includes three components: A relatively small indication representing a fast route to market, substantial sales potential (at least $100M) and strong activity that makes approval likely.
Until Exelixis finds a new partner, the cost for XL184’s aggressive development program will be borne by the company, which might put some pressure on the balance sheet. The company secured two lines of credit worth up to $160M which could support XL184’s development for the coming year or two. Nevertheless, drawing these funds will create a debt overhang investors rarely feel comfortable with (See Incyte's (INCY) example). In any case, it will be interesting to see when and at what terms Exelixis will strike a new licensing deal for XL184.
Biotech Portfolio updates
It has been a while since the last update in our biotech portfolio, which is still No.2 in the list below with a return of 50.9%. We are selling Isis (ISIS) and adding two new positions in Synta Pharmaceuticals (SNTA) and Onyx (ONXX). Synta, which seems to be making a nice comeback, will be the subject of the next write-up.
Disclosure: See portfolio above