Gene therapy – M&A picking up, imminent crucial update from Biomarin
In Q1, the gene therapy space saw one big acquisition (Roche/ Spark (ONCE)) and several smaller deals including Biogen/Nightstar (NITE), Pfizer/Vivet and J&J/MeiraGTx (MGTX). These deals demonstrate the industry’s appetite for gene therapies with an emphasis on liver and ophthalmology as validated domains. CNS (primarily AAV9) and muscle (primarily AAVrh74) are the two other popular domains
What I find interesting in these deals is the fact they weren’t done from a position of strength (as opposed to the Novartis/Avexis deal, for example). Spark was struggling with its HemA program and did not have near term catalysts with other programs. Nightstar was trading around its IPO price with initial XLRP data that were hard to interpret at higher doses. MeiraGTx’s stock also hasn’t performed well and the company was facing an imminent fundraising.
The next 12 months will be crucial for GTx with tens of clinical readout that should justify or disprove the current sentiment. The single most important readout will be a 3-year follow up for Biomarin’s (BMRN) HemA program (ValRox), expected mid-2019. This program is probably the most prominent gene therapy in development (rivaled only by Sarepta’s (SRPT) DMD program). It is in pivotal trials, it clearly works and market potential is huge (peak sales >$3B).
Biomarin’s 3-year update will potentially have dramatic implications for the entire liver-targeting gene therapy space because it will start to answer the durability question. AAV-based products are hailed as single lifetime treatments but actual durability is an open question. Investors are rightfully nervous about ValRox and other liver-targeted programs because in contrast to other tissues (brain and retina), cell divisions clearly occur in adult livers, potentially leading to transgene dilution.
Results to date with ValRox show a peak in FVIII levels at 6 months followed by a gradual decline up to 2 years (last data point). Will the trend continue or can levels reach a plateau that can be maintained for years? If the 3-year update demonstrates a significant decline towards the clinically meaningful threshold (12%), this will cast a shadow over ValRox’s value proposition. Stabilization in FVIII levels will reinforce the notion of a long term treatment although actual persistence will have to be demonstrated. The market would like to see 5 years at minimum but ideally 10+ years of durability.
Biomarin’s update should have limited read-across to CNS and ophthalmology (retinal) programs given the limited cell turnover in these tissues but one area that should be affected is muscle diseases where vector dilution due to cell division is potentially the biggest issue. In diseases like DMD and XLMTM, relevant patients are children or infants who add muscle mass as they grow. So far, initial biomarker data from Sarepta’s DMD and LGMD2 programs are impressive but functional data are limited. Audentes (BOLD) presented encouraging results with longer follow up in XLMTM but more follow is needed to understand durability.
Antibody drug conjugates – DS-8201 gets center stage
As far as investors are concerned, ADCs are still in the penalty box given the limited success seen to date with the class. Out of >100 programs, only a handful of ADCs made it to the finish line, predominantly due to a narrow therapeutic window. As most large biopharma companies like Pfizer and Roche de-prioritized ADCs as a class, next-gen ADCs are being developed by smaller, specialized companies like Zymeworks (ZYME) who are trying to design safer ADCs.
While it is too soon to assess the next wave of ADCs, Q1/2019 had news regarding four late stage ADC programs (2 positives and 2 negatives). Daiichi Sankyo announced a massive deal with Astrazeneca for its HER2 ADC and Astellas/Seattle Genetics’ (SGEN) Nectin4 program reported positive results in a pivotal P2, whereas Immunomedics (IMMU) encountered a major regulatory setback with its Trop2 ADC and Immunogen (IMGN) reported the failure of its P3 FRa study.
The economics of the Daiichi/Astra are definitely unusual ($1.35B upfront and $5.55B in future milestones for a 50/50 profit share ex-Japan). DS-8201’s efficacy appears best-in-class and since the targeting antibody is trastuzumab, the differentiated profile can be attributed only to the linker and payload. In HER2+ breast cancer, this ADC generated a 59.5% response rate and a 20.7 month durability of response in heavily pretreated (8th line!) patients. These figures are almost double what Kadcyla reported in a less heavily-pre-treated population.
DS-8201 also generated positive data in populations for which Kadcyla is not approved for such as HER2+ gastric and lung cancer patients. This extended market potential was likely the reason behind the rich deal terms.
Daiichi’s ADC technology is unique as it utilizes a topo-1 inhibitor (most programs utilize tubulin inhibitors [DM1, MMAE] or DNA binders [PBD, IGN]) and a high DAR (drug/antibody ratio) ratio of 7-8 (vs. 3-4 in most ADC programs). Immunomedics’ ADC technology is also similar in terms of payload and DAR, so this may represent a new ADC class.
While Daiichi’s ADC platform clearly delivers on the efficacy front, safety profile is challenging. Activity comes at a heavy price of side effects including a high rate of nausea/vomiting and lung toxicity (which was fatal in some cases). Going forward, Daiichi hopes to mitigate lung toxicity with active monitoring and dose reductions in ongoing P3 trials.
Daiichi is one of the few pharmas with an active ADC pipeline (figure below). At its R&D days, the company shared data with a second ADC program targeting HER3 with encouraging P1 data to date (42.9% response rate, 8.3 months PFS). Safety profile is too early to assess but lung toxicity was also reported, making this a potentially class effect.
Seattle Genetics’ positive P2 results in bladder cancer got muted reaction from the market despite of a 44% response rate and a ~7-month durability of response in PD-1 failures. To put this number in perspective, PD-1 inhibitors were approved based on a 20-23% response rate in a similar setting. Safety profile isn’t benign but the overall clinical profile still looks strong.
Immunogen’s data for mirvetuximab in ovarian cancer was clearly negative but it is hard to ignore the signal in the FRa-high subset with a strong overall survival benefit (HR=0.62, p=0.033). The study was designed with PFS as an endpoint so the company would not be able to use it for approval but in contrast to most retrospective subset analyses, the signal in the FRa-high subset looks real. It implies that FRa-medium patients in the study had a slightly shorter survival with mirvetuximab vs. chemo, which isn’t unusual as mirv was tested as monotherapy. To me, this is one of these rare cases where a second, relatively small (250 patients) study is warranted. Mirvetuximab’s is in multiple ongoing P2 combination trials which may be amended to include only FRa-high patients.
NASH – P3 disappointments, activity continues
NASH is as popular as ever on Wall Street, evidenced by NGM (NGM) Bio’s recent IPO and $1B valuation. I am more cautious on NASH as I view it as a high risk indication: It isn’t commercially validated (no approved drugs yet), disease is asymptomatic in most patients and the bar for safety is very high.
So far, P3 readouts have been disappointing with a mixed data set for Intercept’s (ICPT) Ocaliva and a complete failure for Gilead’s (GILD) Ask1 program. Ocaliva met the primary endpoint of the study and demonstrated a clear anti-fibrotic effect but the benefit was marginal which, coupled with a problematic side effect profile, may prevent wide adoption of the drug. Ocaliva’s data serves as a proof of concept for more selective next-generation non-bile acid FXR inhibitors which should hopefully be safer and more efficacious.
I still plan to keep my NASH exposure limited to Madrigal (MDGL) and Viking (VKTX) with their THRb programs despite the lack of near-term catalysts. What I like most about the two is the mutual validation each drug provides, validating the strong efficacy data. Biggest risk is still long term safety but so far both MGL-3196 and VK2809 appear to be safe (especially in light of the recent 5mg data for VK2809).
Portfolio holdings – Apr 28, 2019