Why I’m Shorting Axovant

The first company I’m going to dive into is Axovant Sciences Ltd. Results from their upcoming phase 3 clinical trial, called Mindset, will lead to a big move in the company’s stock. This trial is evaluating the effects of the 5-HT6 receptor antagonist, Intepirdine, in combination with Donepezil on Alzheimer’s Disease (AD). As of last Friday, options are pricing in a move of ~$24 before the end of 10-2017, so it’s worth investigating if this event is worth attempting to profit from. Results are expected before the end of 09-2017.

AD drug trials have largely been met with disappointment with the exception of some acetylcholinesterase inhibitors (Donepezil, etc) and a NMDA receptor AD antagonist (Memantine). The reasons for this are numerous, but it also stands to reason that the FDA has an interest in approving a new AD drug, if one shows good efficacy.

One of the hallmarks of AD is reduced cholinergic neurons in the brain. This is associated with the dementia symptoms like loss of memory, therefore, the development of safe acetylcholinesterase inhibitors is a direction many drug companies went. Turns out these drugs increase acetylecholine and improve symptoms without reducing the actual progress of the AD. This brings us to the 5-HT6 receptor.

5-HT receptors are serotonin receptors, but happen to increase acetylcholine when they are antagonized. This led researchers to postulate that acetylcholine could be increased in AD patients and improve the symptomology, which Donepezil does. However, most 5-HT6 receptor antagonists by themselves were ineffective at treating AD symptoms to a significant degree (Maher-Edwards G, 2010, Maher-Edwards G, 2011, Maher-Edwards G, 2015). When combined with Donepezil, however, there is preliminary data suggesting that this synergy could increase brain acetylcholine more than each could by themselves, demonstrated in rodents (Fig. 1. adapted from Dawson, LA, 2011, compare the purple line to the red line). This is the working theory behind the effectiveness in the phase II trial that prompted Axovant to conduct phase III clinical studies (Maher-Edwards G, 2015).

Fig. 1. Effects of SB-742457, Donepezil or their combination on brain acetylecholine. Note that Intepirdine is the same as SB-742457, which is the same as RVT-101.  Keep in mind this is data only suggests a short-term effect and I haven’t found any data on long-term brain acetylcholine levels with treatment of Donepezil and Intepirdine.

The Mindset clinical trial (NCT02585934) has two primary outcome measures. Change from baseline on the (1) ADAS-Cog-11 scale and (2) ADCS-ADL scale. Both these scales are commonly used in clinical studies to analyze patient’s cognitive and non-cognitive brain function and both will need to show a statistical significance in order for the FDA to approve this drug. The researchers are comparing RVT-101 (Intepirdine) plus two different doses of Donepezil (5mg/day and 10mg/day) to only Donepezil at those doses. This splits the 1315 patients into 4 groups, lowering the power of the analyses than if they only had two groups (i.e. RVT-101 + 10mg Donepezil vs 10mg Donepezil). This also increases their chance of seeing a positive effect, since the 5mg Donepezil dose leaves much more room for the potential acetylcholine-boosting effects of RVT-101 to improve the primary endpoints of the trial.

Burns et al. evaluated long-term Donepezil treatment in a group similar to those in the Mindset trial and saw that 5mg/day Donepezil led to a substantial improvement in ADAS-Cog scores that drifted towards placebo by 24 weeks, whereas the 10mg/day group still had significant benefits from the treatment (Fig 2. adapted from Burns A, 1999).

Fig. 2. Mean change from baseline (+/- SEM) of ADAS-cog scores in patients receiving 5 and 10mg/day Donepezil and placebo groups.

Patients in the Mindset trial will be somewhere in the treatment regimen of Donepezil (the phase II study had them on a stable dose for 2 months with the first dose given at least 6 months before RVT-101 treatment, Maher-Edwards G, 2015). This leaves the potential of Intepirdine to come in and improve ADAS-Cog-11 and ADCS-ADL at the 24 week time frame. Some confusion will arise if RVT-101 + 5mg Donepezil leads to significant improvement compared to 5mg Donepezil, but not compared to only 10mg Donepezil, since the max benefits could be achieved by providing only one drug. I would argue giving only one drug is better if you don’t get a better response with two, despite the higher dose. Also, when Donepezil is ramped up from 5mg/day to 10mg/day after 4-6 weeks, the cholinergic adverse events are no different than placebo (Burns A, 1999).

Given everything above, here are my overall suspicions on the Mindset trial:

  1. The chance that the data from the phase II trial will to be fully reproduced is low (Fig. 3. adapted from Maher-Edwards G, 2015).
maher-edwards 2015 important fig mod
Fig. 3. Summaries of Study 1 (SB742457 monotherapy) and Study 2 (SB742457, Donepezil cotreatment from Maher-Edwards G, 2015.
  • Study 1 shows an insignificant effect of donepezil on ADAS-Cog compared to placebo (p=0.821). We would expect that Donepezil treatment leads to a significant improvement in those time points, shown reproducibly in many previous studies. This is not directly relevant because Study 2 is the design of Mindset, but it’s worth noting because something with this cohort is not representative of the greater population.
  • Study 2 shows that the placebo group (Donepezil) gets worse by +1.2 points on the ADAS-Cog scale after 24 weeks (blue box). If we assume the patients had been on Donepezil for months, we would expect a drop in ADAS-Cog of ~0.5 as we see in Burns A. 1999. SB742457 35mg + Donepezil improves ADAS-Cog compared to placebo by -1.5 compared to a group worsening in their disease. This is problematic because if this cohort of people happens to have less Donepezil effectiveness compared to the greater population, a larger study will see a greater effect of Donepezil, leading to a smaller difference between RVT-101 + Donepezil.

2. Previous drugs with the same mechanism as Intepirdine failed numerous phase III clinical trials.

  • Idalopirdine is a 5-HT6 receptor inhibitor that was brought to multiple phase III trials by Lundbeck for its effectiveness in AD in combination with Donepezil. All the trials failed, and sadly we don’t have the publications to investigate deeper.
  • The CEO of Axovant, David Hung, gave an interview with someone from Morgan Stanley. In regards to the upcoming trial, Hung was asked about Lundbeck’s phase III attempts with Idalopirdine. He said that the reason for the failure is due to Lundbeck changing the Idalopirdine dosing from their phase II dosing of 3 times daily (TID) regimen to once daily in their phase III. The once daily dose was higher than the TID dose, so even with a 10-hour half-life, a substantial blood concentration would be present most of the day so I don’t find this argument compelling.

3. GSK was not interested in pursuing RVT-101 further.

  • This is a smaller point, but one that should be addressed. There are many reasons why GSK may not have been interested in taking RVT-101 to phase III. Among those, we can assume that the risk-reward of doing so was not enticing enough.

The prior probability that this trial is a success is 1:1. Looking at clinical trial success data, we see that an average neurological disease phase III clinical trial is a success 58.1% of the time (Clinical Development Success Rates, 2006-2015). If we assign subjective but educated values to the independent claims I make above, the likelihood that the trial succeeds given Suspicion 1 is about 3:5. I give Suspicion 2 a ratio of 1:4 and Suspicion 3 a ratio of 0.99:1. I think these are generous enough, given the data. Multiplying this out leads to a 13% chance that the trial succeeds.

Together, this makes me confident in a short position. A look at the IV of Oct options shows around 400% which means that options prices are greatly inflated. This is an argument for selling premium, but then a look at the bid-ask spread is dissuading since they are all over >0.5. Given this, I’m going to short equity and aim to profit off a drop in the stock price.


So far, the position is down ~6% since AXON has had a nice run-up this past week. I’m going to hold and cover after the news is released in a week or so.

Good luck,



Axovant Sciences’ (AXON) CEO David Hung Presents at Morgan Stanley 15th Annual Global Healthcare Brokers Conference (Transcript)

Burns A, Rossor M, Hecker J, Gauthier S, Petit H, Möller HJ, Rogers SL, Friedhoff LT. The effects of donepezil in Alzheimer’s disease – results from a multinational trial. Dement Geriatr Cogn Disord. 1999.

Clinical Development Success Rates, 2006-2015. BIO, Biomedtracker, Amplion. https://www.bio.org/sites/default/files/Clinical%20Development%20Success%20Rates%202006-2015%20-%20BIO,%20Biomedtracker,%20Amplion%202016.pdf

Dawson LA. Pharmacology of 5-HT6 receptors, Part II. International Review of Neurobiology. 1st Edition. Borsini F. 2011.

Maher-Edwards G, Dixon R, Hunter J, Gold M, Hopton G, Jacobs G, Hunter J, Williams P. SB-742457 and donepezil in Alzheimer disease: a randomized, placebo-controlled study. Int J Geriatr Psychiatry. 2011.

Maher-Edwards G, Waston, C, Ascher J, Barnett C, Boswell D, Davies J, Fernandez M, Kurz A, Zanetti O, Safirstein B, Schronen JP, Zvartau-Hind M, Gold M. Two randomized controlled trials of SB742457 in mild-to-moderate Alzheimer’s disease. Alz Dem: Trans Res Clin Int. 2015.

Maher-Edwards G, Zvartau-Hind M, Hunter AJ, Gold M, Hopton G, Jacobs G, Davy M, Williams P. Double-blind, controlled phase II study of a 5-HT6 receptor antagonist, SB-742457, in Alzheimer’s disease. Curr Alzheimer Res. 2010.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s