After a long summer break it is time to review recent events and update the portfolio. As far as clinical readouts go, my portfolio had a brutal summer with one complete P3 failure from Array Biopharma (ARRY), a mixed data set from Aurinia (AUPH) and a win from SAGE (SAGE) that resulted in limited share appreciation. This was offset by strong performance from Exelixis (EXEL), my biggest holding which is up 48% quarter to date.
For the remainder of 2016 I plan to gradually increase exposure to gene therapy, which I hope will become one of the industry’s primary growth drivers in the coming years. In parallel, as I am still pessimistic about the biotech field in general (R&D productivity, pricing, biosimilars…), I intend to keep my short ETFs and a significant cash position.
Data readouts – Array and Aurinia disappoint, SAGE delivers
Array Biopharma – Astrazeneca’s (AZN) selumetinib, originally developed by Array, failed to improve PFS or overall survival in KRAS-mutated lung cancer. This P3 failure demonstrates once again the challenges in corroborating positive P2 data and joins a series of disappointing news for MEK inhibitors. In 2015, selumetinib failed another P3 study in uveal melanoma and earlier this year, Array terminated a P3 study for its wholly-owned MEK inhibitor, binimetinib, in ovarian cancer. Recent P3 data for binimetinib in NRAS+ melanoma was technically a success but benefit was underwhelming (improvement in PFS without a survival benefit).
MEK inhibitors are still being explored in multiple indications that may enable developers to expand their use beyond BRAF+ melanoma. At ASCO 2016, Roche and Exelixis reported an intriguing efficacy signal with Cotellic in KRAS+ colon cancer. Another indication that is not discussed much is neurofibromatosis type 1 where selumetinib demonstrated strong efficacy but timelines and registration strategy are not clear. Later in 2016, Array will report data for binimetinib in BRAF+ melanoma, where two other MEK inhibitors (Mekinist and Cotellic) are already approved. Array hopes to demonstrate a differentiated safety profile but penetrating the BRAF+ market is going to be challenging.
Aurinia – Aurinia reported results from a P2b evaluating two doses of its calcineurin inhibitor (voclosporin) in lupus nephritis. While voclosporin demonstrated superiority over the control arm at the low dose in inducing remission (32.6% vs. 19.3%), there was no effect with the high dose and there was an alarming imbalance in deaths between the low dose cohort (13 cases overall, 10 of which occurred in the lower dose arm). The fact the active arm had such a high mortality obviously raises safety concerns but also casts a shadow on the efficacy signal and trial design.
SAGE –SAGE reported a strong efficacy signal for its lead program, SAGE-547, in a placebo- controlled P2 in postpartum depression (PPD). Despite the typically high placebo response in depression trials, the drug led to a clinically meaningful reduction in depression with a ~12 point reduction from placebo on the HAM-D scale. Importantly, the effect had a quick onset and appears durable almost a month following treatment cessation.
These results are very significant for SAGE-547 because they validate the initial signal observed to date in single arm open label trials or case studies, which are rightfully viewed as unreliable (especially in CNS indications). Despite the small sample size, the strength and consistency of the PPD data imply the drug is active and de-risk the program.
Going forward, SAGE decided to focus on SRSE with SAGE-547 (P3 data pushed out to 1H:2017) while pursuing PPD as well as other CNS diseases (essential tremor, Parkinson’s) with a next-generation oral drug (SAGE-217). This decision makes sense in light of the different clinical settings (SRSE patients are hospitalized and need immediate solution for their seizures whereas PPD may require more prolonged treatment). SAGE-217 appears more potent and selective than SAGE-547 with good pharmacologic profile based on P1, but it is still not validated in the clinic.
Momentum continues to build around gene therapy
There have been some notable news in the gene therapy space including three data readouts and an acquisition. This gradual but consistent progress helps to build confidence around gene therapy and makes it one of the most important segments in the industry.
The combined recent data set from all gene therapy companies still comprises less than 100 patients with limited follow up, but for the first time in more than 25 years, we are starting to get positive answers on three major questions: Efficacy, durability and safety. With tens of active clinical programs, strong data across multiple domains (liver, CNS, retina, bone marrow) and an armamentarium of tissue-specific vectors, gene therapy may be ready for primetime. The field will surely see its share of failures but the path to commercial success with gene therapy is finally visible.
Biomarin (BMRN) made a big splash with its hemophilia A treatment, BMN-270, demonstrating that gene therapy can lead to a functional cure (at least for a period of time). Results from 7 patients treated at the highest dose demonstrated quick and robust expression of factor VIII (hemophilia A is caused by low factor VIII activity). Strikingly, all but one patient achieved normal levels and this was accompanied by a dramatic reduction in bleeding events despite taking patients off prophylactic FVIII treatment. Follow up is still limited (up to 20 weeks) but Factor VIII production appears to be stabilizing or even increasing in some cases.
The study was a major win but there were two safety signals: ALT elevation (marker for liver damage) and an above normal FVIII levels in two patients.
ALT elevations are closely watched in every liver-targeted gene therapy study because earlier products led to ALT elevations (probably an immunogenic response against the viral vector) that led to clinically relevant elevations and were associated with loss of transgene expression. There is still not enough follow up to rule out any chronic liver toxicity but according to Biomarin, the ALT elevations were not clinically meaningful and appear to be manageable with steroids (in some cases patients were taken off steroids and liver functions returned to normal). Fortunately, the ALT elevations do not seem to correlate with loss FVIII expression as production continued to rise or stabilized during these events.
The excessive FVIII levels in two patients are troubling on paper (especially if production keeps going up) because elevated levels of FVIII are associated with higher risk of thrombotic complications according to several studies, reviewed here. This demonstrates a primary concern with gene therapy – it’s irreversible and cannot be “switched off”. While the company disclosed patients are not experiencing any thrombosis-related events, they will have to be monitored and potentially treated in case FVIII continue to climb or any abnormalities emerge. Going forward Biomarin plans to evaluate a lower dose to avoid such cases.
Overall, Biomarin’s hem A data are impressive and together with data from hemophilia B programs from Spark (ONCE)/Pfizer (PFE) and UniQure (QURE), they lay the groundwork for registration programs in hemophilia potentially next year.
While Biomarin’s gene therapy for hemophilia demonstrated both protein production (FVIII in the blood) and clinical improvement (bleeding incidence), Avexis’ (AVXS) AVXS-101 demonstrated only the latter. Nevertheless, the dramatic clinical signal coupled with the fact that the target tissue are neurons make AVXS-101 one of the most important gene therapies in development (and personally speaking, the most intriguing one).
Last month, Avexis provided an update from a clinical trial evaluating AVXS-101 for SMA1, a fatal disease in infants that affects the motor neurons and is invariably fatal by 2 years of age. Following the previous data readout (which I discussed here), the clinical effect appears to hold with unprecedented functional and survival improvements over historical data. All but one patient on the high dose cohort had a dramatic functional improvement vs. the expected persistent decline that is the hallmark of SMA1. Importantly, none of the patients had an “event” (defined as ventilation or death) despite crossing 8 months of age where a recent study showed an event rate of 50%. Duration of the effect is still unknown due to limited follow up (appears to be at least 1 year).
The trial did not have a control arm but the difference is so dramatic that I find it hard to believe it is just a statistical artifact. The recent breakthrough therapy designation the FDA granted to AVXS-101 is another testament to the robustness of the data. Biogen (BIIB) and Ionis (IONS) reported P3 success for nusinersen in SMA1, which validates the signal observed in a previous single-arm study and this strengthens the notion that single-arm trials in SMA1 can generate a reliable signal. Although nusinersen can theoretically be seen as a competitive threat to AVXS-101, if both drugs are approved, nusinersen will probably be used on top of a one-time gene therapy like AVXS-101.
AVXS-101’s effect, if real, implies that AAV9 vectors Avexis utilizes can cross the blood-brain-barrier and reach the central nervous system. This could have dramatic implications for many rare genetic diseases (and down the road more prevalent diseases like Parkinson’s and Alzheimer’s disease) where the nerve system is damaged but cannot be modulated by existing drugs. In this perspective, companies like REGENXBIO (RGNX) and Abeona (ABEO) are very interesting opportunities as both target neurologic diseases (or ones with neurologic manifestations) with AAV9-based gene therapies.
Spark reported updated results from its pivotal trial for its RPE65 gene therapy (Voretigene neparvovec) for inherited retinal disease. The study was overwhelmingly positive and will likely be the basis of the first ever FDA approval for gene therapy next year. In contrast to other gene therapy programs in development, Spark’s program has up to 3 years of follow-up and so far the effect appears to hold.
RPE65 gene therapy is to retinal diseases what hemophilia gene therapy is to liver diseases – a proof of concept that paves the way for other genetic retinal diseases like XLRS, XLRP, achromatopsia and choroideremia. Spark will have first clinical data for its choroideremia program later in 2016. AGTC (AGTC) is expected to report initial clinical data in XLRS and achromatopsia also in 2016.
Spark also has a Hemophilia A program (SPK-8011), which is about to enter P1. Despite being more than a year behind Biomarin, recent results for Spark’s Hemophilia B program (partnered with Pfizer) were very encouraging and may point to some advantages with Spark’s vectors. These include better potency and lower immunogenicity that could translate to a more compelling clinical profile (i.e achieving the required protein levels with limited immune reaction, and down the road better durability and dosing flexibility).
Pfizer acquires Bamboo
One item that went relatively unnoticed is the acquisition of Bamboo by Pfizer. Bamboo is developing gene therapies for neuronal and neuromuscular indications with a program in P1 for Giant Axonal Neuropathy. Despite the modest deal size ($150M upfront), it is an important milestone in big pharma’s venture into gene therapy, which historically has not been aligned with the pharma business model (one time genetic treatment vs. recurring sales of small molecules or biologics). The title of Pfizer’s press release (“Pfizer aims to become industry leader in gene therapy with aquisition of bamboo therapeutics”) clearly demonstrates this trend.
If gene therapy is indeed going to become a central component of the biopharma industry, the demand for validated platforms with clinical data and CMC capabilities could grow dramatically in the near future. In contrast to cell therapies such as CARs and TCRs, gene therapy may be more palatable to pharma because the end product is still a drug in a vial and not a process.
Given the early stage and inherent risk in gene therapy, I intend to build a diversified portfolio with the goal of starting 2017 with 7-8 stocks that will comprise a third of the portfolio. Today I am starting with Spark, Avexis, REGENXBIO and Abeona (Next on the list is Bluebird going into ASH 2016…)
Lastly, I am initiating a small position in Kura Oncology (KURA) following an early signal in HRAS+ H&N cancer.
Portfolio holdings – September 4th, 2016