The latest candidate in ARIAD Pharmaceuticals, Inc.'s (ARIA) early-stage pipeline, AP26113, is a drug that inhibits both ALK and mutated EGFR kinases. Targeting these two kinases has immense clinical potential for 290,000 lung cancer patients that are diagnosed each year, as well as other cancers like anaplastic large cell lymphoma and neuroblastoma. Our analysis of preclinical and early Phase 1 data shows that AP26113 has potential to outclass other second-gen ALK and EGFR drugs in development for lung cancer.
Insights into Lung Cancer: EGFR and ALK mutations
Lung cancer is the leading cause of cancer mortality worldwide with over 1 million deaths a year, and a survival rate of only 16.3% after five years. Incidence of lung cancer is approximately 220,000 cases per year in the US and 1.6 million per year worldwide. Of the two major sub-groups of lung cancer, small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC), NSCLC comprises approximately 90% of the total diagnoses. There are currently no approved techniques for early identification of NSCLC in patients, thus the majority of cases are diagnosed at advanced or metastatic stages.
However, there has been rapid progress in identifying genetic abnormalities that drive progression of NSCLC. This personalized genomics approach holds hope in improving the progression free survival (PFS) and overall survival (OS) of NSCLC patients compared to the standard chemotherapeutic agents used to treat NSCLC, such as cisplatin and docetaxel. Out of over 15 genes identified as drivers of NSCLC, epidermal growth factor receptor (EGFR) has received the most attention. Driver mutations in EGFR account for 10% US cases and 30% Asian cases in NSCLC, or 290,000 patients worldwide. And EGFR is overexpressed in 62% of NSCLC patients.
Another important genetic driver mutation of NSCLC is anaplastic lymphoma kinase (ALK), a receptor tyrosine kinase in the insulin receptor/IGF1R superfamily. Activating fusions are detected in up to 70% of anaplastic large cell lymphoma (ALCL), and in 4% of NSCLC in the form of EML4-ALK (40,000 patients/year worldwide). Activating mutations in ALK are also found in 10-15% of neuroblastoma, the most common childhood cancer which accounts for 15% of childhood cancer deaths.
Unlike EGFR, ALK is largely absent in normal tissues after birth, but it is thought to play a role during neural development. ALK knockout mice appear normal and viable with no apparent developmental, anatomical or locomotor defects - though some studies suggest that ALK deficiency may also alter alcohol-consumption related behaviors and prevent depression. These genetic studies suggest that a drug inhibitor specific against ALK will have minimal side-effects, and might even exert beneficial anti-depressant effects.
AP26113 Potential in Targeting ALK-driven Cancers
In 2010, ARIAD announced results of preclinical studies on AP26113 showing potent inhibition of ALK, including mutant forms that are resistant to the 1st-gen ALK inhibitor, crizotinib developed by Pfizer (PFE). ARIAD scientists presented these data at the annual meeting of the American Association for Cancer Research (AACR). In another study, ARIAD performed pre-clinical trials comparing AP26113 to crizotinib in a series of ALK-dependent cell culture and in vivo models. In all models, AP26113 was at least ten-fold more potent than crizotinib. In addition, AP26113 exhibited approximately 100-fold selectivity for ALK-positive cell lines compared to the 10-fold selectivity by crizotinib, and demonstrated excellent pharmacokinetics.
In yet another independent study performed by researchers at the Massachusetts General Hospital, scientists found that both AP26113 from ARIAD and NVP-TAE684 from Novartis (NVS) could effectively inhibit the mutated ALK (L1196M) that leads to crizotinib-resistance within 1 year of treatment in lung cancer patients. Furthermore, the scientists found that AP26113 was extremely specific for ALK with an IC50 of 0.62 nM. The study also independently validated AP26113's lack of activity against native EGFR and most other kinases (IC50 > 100 nM). However, AP26113 does have some activity against the ROS1, FER, FLT3, FES, FAK, BRK, STK22D and CHK2 kinases (IC50 < 10 nM). Interestingly though, a new subset of NSCLC patients with ROS1 translocations were recently identified, and are being treated with crizotinib too. It will be interesting to see if AP26113 performs better than crizotinib in ROS1+ tumors too.
Competitive landscape for ALK inhibitors
In the field of ALK inhibitors, the most effective approved inhibitor currently is Crizotinib, a triple ALK, ROS1 and c-MET inhibitor developed by Pfizer. It received accelerated approval from FDA for ALK-positive locally advanced or metastatic NSCLC patients on Aug 26, 2011. The approval was based on two single-arm trials in 136 patients and 119 patients, respectively, as a second/third line therapy. The most common adverse reactions in both studies were vision impairment, gastrointestinal toxicity (vomiting, diarrhea, and constipation) and edema in >25% of patients. Grade 3-4 adverse events included liver dysfunction and neutropenia in >4% of patients. Fatal treatment-related pneumonitis was observed in 1.6% of patients.
However, most importantly, Crizotinib-treated patients showed a significant improvement in OS compared historically to patients treated with standard 2nd-line therapies (55 % vs. 12 % survival at 2 years, P=0.004). Since the FDA label did not specify prior therapies, Crizotinib can be prescribed as a 1st-line therapy for newly diagnosed patients with advanced ALK+ NSCLC. Currently a 1st-line study (PROFILE 1014) comparing Crizotinib to a platinum/pemetrexed combination is ongoing, and the Phase 1 trial reported 80% response rate in 15 newly diagnosed ALK+ patients.
Unfortunately, within 1 year of treatment, resistance to Crizotinib develops in ALK+ NSCLC patients due to acquisition of several mutations, including L1196M. Only a few second-gen ALK inhibitors like ARIAD's AP26113, Chugai Pharmaceuticals' CH5424802 and Xcovery's X-396 can demonstrably overcome mutated ALK (L1196M). Phase 1 results of CH5424802 reported partial responses in all 7 patients, but showed mild to moderate toxicity of grade 3 hypophosphatemia (2 out of 15), neutropenia (2 out of 15), blood CPK increases (1 out of 15) and hypermagnesemia (1 out of 15). Preclinical studies of X-396 also show 10-fold more potency than Crizotinib, like AP26113, but Xcovery only started its Phase 1 trial in June 2012. And although LDK378 from Novartis has shown robust activity in a Phase 1 trial, it showed relatively severe gastrointestinal toxicity, and has not demonstrated activity against the L1196M mutation.
Another class of drugs that can address Crizotinib-resistance are HSP90 inhibitors, such as AUY922 (Novartis) and Ganetespib (STA9090). Oncogenic proteins like EGFR and ALK form tight complexes with HSP90 chaperone proteins, and therefore HSP90 represents another target to indirectly inhibit ALK. For example Ganetespib developed by Synta Pharmaceuticals (SNTA) is a potent HSP90 inhibitor with 50-fold higher potency than Crizotinib in H3122 NSCLC cells, and effective for all 15 tested crizotinib-resistant NPM-ALK/BaF3 mutations with IC50 ranging between 14-23 nM. Data from the Phase 2b/3 GALAXY study with Ganetespib was presented at ESMO recently, showing promising safety profiles but an insignificant benefit in PFS (2.8 vs 4.2 months, P=0.076), and an undetermined improvement in OS.
AP26113 Potential in Targeting EGFR-driven Cancers
In 2011, ARIAD announced preclinical studies showing that AP26113 effectively inhibited EGFR activated by T790M, and L858R or exon-19 deletion mutations, both in cell culture and in mouse models. These mutant EGFRs were resistant to Tarceva (erlotinib), developed by OSI/Genentech (OTCQX:RHHBY). When tested against native (unmutated) EGFR activated by endogenous growth factors, AP26113 lacked activity, indicating a specific selectivity for mutated EGFR. These data were presented at the 14th World Conference on Lung Cancer.
Mutations that lead to an increase in EGFR abundance or activity have been associated with a number of cancers, including lung cancer, head and neck cancer, colorectal cancer, and glioblastoma. However, EGFR is a receptor tyrosine kinase that is present and active throughout human adult body tissues. EGFR knockout in mice leads to embryonic death. If born, mice with EGFR deficiency show widespread impairment of epithelial growth in multiple organs, e.g. the gut, lungs, brain, skin, eyes, kidneys, and the liver. These genetic studies suggest that any drug inhibitor specific against native EGFR will have severe side-effects. Thus only a drug specific for mutated EGFR, like AP26113, is likely to be well-tolerated in NSCLC patients.
1st-Gen EGFR inhibitors
The most widely used EGFR inhibitors include the small molecule tyrosine kinase inhibitors Iressa (gefitinib) developed by AstraZeneca (AZN) and Tarceva developed by OSI/Genentech, approved in 2003 and 2004. Pivotal trials for these drugs were carried out before activating mutations in EGFRs were discovered. Therefore, these drugs were initially used in NSCLC patients without regard for their genetic heterogeneity. As a result, clinical trial results were disappointing, and Iressa has not been prescribed as treatment to new patients in the US since 2005. And although Tarceva was granted approval in the US as a second/third line treatment for NSCLC in late 2004, results from other trials that compared Tarceva to platinum-based therapy in patients with progressed NSCLC were negative. Overall, while these EGFR inhibitors showed some efficacy, improvement in OS was not dramatic.
Importance of EGFR mutations
Interestingly, during the clinical trials for Iressa, it became apparent that one subset of patients responded particularly well. EGFR mutation-positive patients treated with Iressa had significantly improved PFS (9.5 vs. 6.3 months, P = 0.0001). This and other results resulted in approval of Iressa outside the US for first-line treatment of advanced NSCLC patients with EGFR mutations. Similar trials led to approval of Tarceva outside the US for first-line treatment of NSCLC patients with EGFR mutations.
However, despite the statistically significant three-to-seven-month benefit in PFS seen in these trials, it is critical to note that OS was still not significantly improved by Tarceva or Iressa even in patients with EGFR mutations. This could be due to crossover of patients during the trials, so a true effect was masked. Alternatively this could be due to the side-effects of 1st-gen EGFR inhibitors, some of which led to a warning from FDA, and which limit the efficacious dosage.
More importantly, patients only respond to Iressa and Tarceva initially, as most will develop resistance within 1 year. This resistance is now known to be due to emergence of new EGFR mutations, primarily the T790M mutation (~50% of resistance cases). While 2nd-gen EGFR inhibitors are in development, clinical efficacy will still be limited due to toxicity associated with native EGFR inhibition. Hence therapies designed to target activated EGFR and the T790M mutation, but avoiding native EGFR inhibition, are a very promising class of cancer therapeutics.
Competition landscape for second-gen EGFR inhibitors
It should be noted that the EGFR inhibitor field is becoming crowded. There are at least 3 EGFR drugs in Phase 3 trials, while AP26113 is still in the midst of a Phase 1/2 study. Of the second-gen EGFR inhibitors that are advanced in their clinical development, Dacomitinib from Pfizer and Afatinib from Boehringer Ingelheim are the strongest contenders - both being irreversible EGFR, HER2 and HER4 inhibitors with long-lasting action. Pfizer recently reported data from a Phase 2 trial comparing Dacomitinib to Tarceva as second/third line treatment for patients with advanced NSCLC. However, dacomitinib showed only a small increase in PFS (2.86 vs. 1.91 months, P = 0.012), and no significant benefit in OS. Additionally, dermatological and gastrointestinal side effects were more frequent in the dacomitinib-treated patients.
Afatinib is thought to be effective in many mutations that confer resistance to Iressa and Tarceva, including T790M. Data from the Phase 3 LUX-Lung 3 trial showed that Afatinib worked as well as Tarceva and Iressa as there was a statistically significant 4 month improvement in the PFS (11.1 months vs. 6.9 months), when compared to cisplatin/pemetrexed in the 1st-line setting. However, the adverse events in Afatinib-treated patients appeared more severe than Tarceva, including severe rashes, diarrhea, mucositis and dry skin. These side effects are common to native EGFR inhibitors and often cause physicians to reduce the dosage administered. Thus Afatinib's benefit in PFS might be negated by its more severe side effects, perhaps leading to no significant improvement in OS just like Iressa and Tarceva. But for definitive results on Afatinib, one will have to wait for data from the LUX-Lung 7 trial comparing Afatinib directly to Iressa.
The Afatinib trial most relevant to AP26113 is the LUX-Lung 1 trial. This trial aimed to compare Afatinib vs. placebo in patients with advanced metastatic NSCLC after failure of chemotherapy, and Tarceva, Iressa, or both. Afatinib performed slightly worse than placebo in OS, but there was a small increase in PFS compared to placebo (3.3 vs 1.1 months, P<0.0001). However there were no complete responses, only 7% of patients had a partial response to Afatinib (vs. 0.5% for placebo), and Afatinib showed severe side-effects. But in another Afatinib study, combination therapy with the monoclonal EGFR antibody Erbitux developed by ImClone/Eli Lilly (LLY), did show promising preliminary results as 2nd/3rd-line treatment in these late-stage patients.
One noteworthy point is that from ARIAD's preclinical data, Afatinib seems to inhibit mutated EGFR much more potently than AP26113 (10-100 fold, depending on mutation), suggesting Afatinib might be more efficacious than AP26113. But what distinguishes AP26113 is that it does not inhibit native EGFR (IC50 > 3000 nM), whereas Afatinib does so with great potency (IC50 = 10 nM), and the safety profile of AP26113 appears to be far superior to Afatinib (see Phase 1 results below), suggesting that the maximum tolerated dose for AP26113 will be far higher than Afatinib (50mg/day).
AP26113's Phase 1/2 Trial Results
Last Saturday, September 29th Ariad presented data from 29 NSCLC (14 ALK+, 11 EGFRm) patients in a Phase 1 trial of AP26113. All patients enrolled had advanced stage tumors and were resistant to standard chemotherapies and/or Tarceva or Iressa. The goal of this study was to establish a maximum tolerated dose of AP26113 and characterize anti-tumor activity. Patients were treated in different cohorts with doses ranging from 30 to 240mg/day. AP26113 appeared well tolerated with low rates of gastrointestinal side-effects, including nausea (26%), diarrhea (18%), decreased appetite (12%), and vomiting (12%). Importantly, no patients reported the rash characteristic of EGFR inhibitors, nor the vision impairments typical of crizotinib.
Amongst the 11 EGFR-mutant patients, one had been treated with crizotinib, 10 with an EGFR inhibitor and one with neither. However only three are still ongoing, as six had discontinued due to cancer progression, one discontinued due to a possible adverse event and one patient died suddenly (180mg). Of the 11 EGFR-mutant patients, only one showed a partial response (120mg). This patient had the exon-19 deletion. Of the two patients that showed disease stabilization, one had the T790M mutation and both had exon-19 deletions too. From the investor conference call today, it seemed that ARIAD can and will escalate the dose beyond 300mg for future cohorts of EGFR-mutant patients. The possible adverse event in the discontinued patient also turned out to be unrelated to AP26113, and had already shown a partial response, according to the trial physician Scott Gettinger, so it would have been two partial responses and two disease stabilizations out of 11 late-stage patients. But in general, given the small numbers and the need to escalate the dose, it is still too early to tell if AP26113 is effective in EGFR-mutant patients.
Amongst the 11 evaluable ALK+ patients, 8 patients showed responses to treatment - including those treated with the lowest 60mg dose. There were partial responses in 6/9 crizotinib-resistant patients and in 2/2 crizotinib-naïve patients. One ALK+ crizotinib-resistant patient also showed a response in a pre-existing brain metastasis. Overall, the data on AP26113 seemed very encouraging, especially regarding the responses in ALK+ lung cancer patients and the excellent side effect profile.
During the investor call, Ariad has indicated that it will extend its Phase 1 portion for EGFR-mutant patients and also transition into the Phase 2 portion of its trial. Investors are likely to be most interested in the response of EGFR-mutant patients to even higher dosages of AP26113, given its unprecedented safety profile. But will it have enough funding?
Ariad doubled its quarterly burn rate to $52M in 2Q of 2012, compared to $25M last year. And it has $250M cash left. Given that it is still expanding its Phase 1/2 trial for AP26113 and it is expanding its sales force for Ponatinib, its cash reserves can only last for one more year. We anticipate that it will need to do another stock offering by 3Q 2013. In December 2011, Ariad raised $243M with a market cap under $2B, and without a drop in stock price. Given its market cap of $3.2B right now, it could potentially raise another $300-600M in 2013. Furthermore as of 2Q of 2012, it is still eligible to sell $66M worth of securities at any time. By the time positive news on Ponatinib is released from FDA in 4Q 2012, we believe the stock price will likely peak at about $30, and Ariad will likely do a stock offering again.
To get a hint of what AP26113's sales could be like in ALK+ lung cancer, we can look at Pfizer's crizotinib. Crizotinib is a potential blockbuster projected to bring $500M - $1B in annual sales by 2015. This is partly because, although the ALK mutation is rare (~4%), lung cancer is the most common cancer in developed countries, so even a small market fraction translates to 40,000 patients a year. In addition, as a drug for a niche indication, crizotinib (Xalkori) can be priced at a premium of $9600 per month. Most importantly, every lung cancer patient diagnosed with the ALK+ mutation will probably stay on crizotinib for 8-9 months (before resistance inevitably develops). If AP26113 grabs the second-line market, which has little competition on the horizon, it could do just as well as, if not better than crizotinib. And given the dismal safety and efficacy results from other competing 2nd-gen EGFR inhibitors, the strong preclinical results of AP26113 and its excellent safety profile means AP26113 is at least still in the race for the $2B EGFR market. Consistent with our analysis, various analysts upgraded the stock recently. Bank of America Merrill Lynch raised its price target from $26 to $30 and Brean Murray raised its price target from $19 to $35.