Description

“Precision medicine” is becoming a ubiquitous buzzword and yet I have been challenged to find a small cap pure-play public company that is creating direct value from genetics, especially outside of oncology.  If AEVI Genomic Medicine (GNMX) did not exist, someone would have to create it.  The company is matching genes to drugs.  With only a $65 million market cap, GNMX is punching far above its weight, especially from a managerial perspective.  Its CEO previously ran a $2.6 billion revenue business inside of Shire which included the largest drug in the first market that GNMX is pursuing: ADHD.  Its Chairman was the longtime CEO of Celgene and is currently Chairman of Teva.  Its partner and largest shareholder after a 2017 $23.5 million hard dollar investment is the Children’s Hospital of Philadelphia (CHOP), who is among the thought leaders in academic genetics.        

This opportunity exists because GNMX stumbled out of the gates with a trial that did not meet its clinical endpoint in April 2017.  The stock got annihilated (down 70%) and has drifted downwards another 30% from there.   As the company has continued to genetically define its target population more narrowly for its first indication (again, its ADHD in case you have it!), it has looked like other companies that, without its genomic tools, have used sub-group analyses as an attempt to justify continued development after a failed trial.  This strategy has had limited success so the market is skeptical.  I believe the company is doing exactly the right thing and has an excellent rationale for its current path that is backed by limited but convincing data. If its strategy was better understood I believe the stock could nearly double with the opportunity for further upside in the longer term.    Because there is a basket of independent programs, a failure in its first indication (which would surprise me) would suggest that the risk of permanent capital impairment is low. 

Business Description:

GNMX has aligned with a clinically annotated biobank (CHOP) and is generating info gleaned from DNA to match drugs to diseases.  Who else is set up like this? Amgen and Regeneron.  In addition, GNMX is targeting the genetics of pediatric diseases, some of which are classified by the FDA as “rare” which offer attractive economics and could come with re-sellable vouchers if the drugs are approved.  These vouchers, designed as an incentive to study neglected small diseases, give their holders the right to priority review with the FDA.  These vouchers have recently sold for over $100 million (see:  https://www.ptcommunity.com/news/20180213/rare-pediatric-disease-priority-review-voucher-sells-130-million as a recent example). Furthermore early-onset diseases are more likely to have a strong genetic component.  Along those lines, it is unusual and very encouraging that CHOP invested $23.5 million in October 2017; this happened after three years of working with GNMX and after the aforementioned trial failure earlier that year.  (The CHOP Foundation previously invested in Spark Therapeutics’ Series B, today a $3 billion company-nice job!)

More detail on the company can be found here:  https://aevigenomicmedicine.gcs-web.com/static-files/aa1640bb-f7be-4513-986f-f6aa3d8643f2

While most of the well-known gene/drug combos involve cancer, which is not GNMX’s focus, there are certainly examples elsewhere.  Amgen has linked PCK9 to the mechanism of Repatha for example in cardiovascular disease.  Vertex has also grown its business by understanding and drugging the nuances of mutations that cause cystic fibrosis.  Of course there are also gene therapies like Bluebird but development is expensive and in the case of gene editing, still controversial. (And I bet VIC members are not interested in another 200+ comment thread on exon skipping companies)  So, by addressing causes and not symptoms, small molecule and biologic drugs that are linked to genetic pathways have a better chance of helping patients.   I also believe they will have a better chance of succeeding which leads me back to GNMX’s glutamatergic neuromodulator (say that five times fast) called AEVI-001 that it brought in house when it acquired CHOP affiliate Neurofix in September 2015.  Neurofix had identified a failed Japanese Phase III compound that had a mechanism of action that could be linked to its genetic insights in subgroups of pediatric ADHD (and autism) patients.  An important differentiator is that this ADHD medication would be a non-stimulant so it would not have the same side-effect profile as current options like adderall.

For brevity, I am going to gloss over the pre-AEVI-001 era when the company was called Medgenics as well as the larger trial history of AEVI-001 in ADHD (discussed in filings) to focus on the fact that 6 patients in its most recent phase II shared a common mutation (CNTN4) and a more severe set of symptoms.  All 6 of the CNTN4 mutation carriers responded to GNMX’s drug, a 100% response rate.  This response was statistically significant even in a small trial (p value of 0.03 which is less than the important 0.05 cutoff that is widely used).  There were eight other genes found to be linked to response in a statistically significant way but CNTN4 represented the majority of the responders.  CNTN4 is also linked to autism and is found in kids who present with both autism and ADHD at the same time.  A 100% response rate, even in just six patients, is extraordinary but this fact seems to have been lost by the previous trial missing its overall endpoint.  

In its current Phase II trial, GNMX is focused on carriers of CNTN4 and eight other related genes but as I value the company, I rely solely on CNTN4 which shrinks the market but looks very likely to succeed.  A point on the current trial is that is it adaptive, meaning it is permitted to make certain modifications called out in advance during the trial.  An independent statistician recommended increasing the size of the placebo arm (per the adaptive protocol) to ensure the trial is sufficiently powered to measure a treatment effect.  Though this is permitted, the market did not like the news (it was announced in mid-June but telegraphed on the Q4 2017 conference call that they may do this) that the trial’s readout was pushed from mid-2018 to the end of 2018 and the stock sold off but has completely recovered. The company also announced that it has improved the molecule itself and called it AEVI-004 which will have patent protection until 2039 and can be used in the Phase III.

If the trial is successful and the response signal as strong as in the previous trial, the next trial would not need to be very large and could be completed relatively quickly.  For example, a rare disease company called Alderya conducted a 6 patient Phase II in which 5/6 were strong responders and all 6 responded.  It is now conducting a Phase 3 with only 39 patients.  Supportive of the concept of a quick trial for GNMX, the measurement period of its current trial is only 7 weeks (with up to 4 weeks needed beforehand to washout other treatments if necessary).

In addition to ADHD, the company will begin a Phase II in autism with AEVI-001, where treatment options are very limited but CNTN4 is important.  It also has begun enrollment for pediatric crohn’s disease with a second unrelated drug candidate called AEVI-002 (also sourced from Japan).  Finally, GNMX has a third pre-clinical program in an undisclosed ultra-rare pediatric indication that I suspect will be voucher-elgible. 

Concerns

• ·         While the company is adequately funded into early next year after it has had  a topline readout of its current ADHD trial (and potential its first pediatric crohn’s trial), like many biotechs more funding will be required. Of course, the valuation of the next capital will be determined by the data but a high response rate in the 9-gene population, led by CNTN4, would put the company in a strong position. I have accounted for dilution in my valuation.  
• ·         Trials are hard to enroll – the precision medicine community is learning that the mutations with the strongest effect that can be linked to distinct genetic subtypes are rare and thus it is very difficult to enroll patients in trials. GNMX actually discontinued its program in 22q syndrome because it could only get 2 patients into the trial.   Fortunately, ADHD and Autism are larger indications so it is far more likely to GNMX will be able to leverage the CHOP patient flow (and other centers) to complete its trials. 
• ·         Relatedly, it isn’t just the trials that will be hard to enroll. Once approved, patients have to know that they are candidates for the drug which can only be determined with a genetic test.  Obviously, penetration will be low without more widespread genetic testing.  GNMX mentions offering its own companion tests but there are a number of rapidly growing testing companies that are better suited to do so.  I believe genetic testing will be much more common  in general five years from now, and in particular for indications where there is a medically actionable step (i.e. a drug), but I have assumed low penetration in valuing GNMX. 
• ·         Finally, the Company could owe up to $450 million in milestones to Neurofix, the company that it acquired to secure AEVI-001.  We don’t have transparency into the breakdown (although it is likely back-end loaded and may include sales milestones) or what needs to be achieved to trigger a milestone (i.e. can success in a smaller population like CNTN4 trigger a full payment of a milestone).  If the deal was structured to match the initial expectations of the lead compound (25% of ADHD patients) and it turns out that it is CNTN4 only (6%), then if the milestones were scaled back commensurately, they would total $108 million. This would be much more palatable for a small company.  Given that the acquired company was a CHOP affiliate, I suspect they will not impair their own investment and the management team will get more clarity on this front (if it hasn’t already) before proceeding further so that it does not drive GNMX off a liquidity cliff. 

Capitalization:

There are 59.3 million primary shares outstanding, 4.1 million warrants with an average strike of $2.88 and out of the money options (6.5m and 2.8m plan/non-plan with average strikes of $4.34 and $4.28 respectively).  Note also that the Company has had a $20 million ATM facility in place since May 23.  If it was 20% of the volume during this time at the closing daily price, it would have sold approx. 1.2 million shares for $1.4 million. 

The company had $26.3 million of cash as of 3/31 and no debt.  It still owes CHOP $6 million in license payments during 2018.  As mentioned above, cash is sufficient to reach a topline data readout and will fund the company into early next year. 

Valuation:

Of course, there is no precision (pun intended!) when valuing a company like this.  Directionally, I will share my assumptions to illustrate how I think about it.  While the current Phase II trial is evaluating patients with CNTN4 and a few other mutations representing 10% of the ADHD population, let’s only use the 6% who have CNTN4 as the basis for valuing the program (even though the company is targeting the full 10%).  The Company has estimated that it will charge approximately $1,000 to $1,500 per annum (based on a $2 to $3 billion market opportunity for the 10% cohort) and thus 6% would be a $800 million to $1.2 billion market.  The valuation is highly sensitive to penetration assumptions and until genetic testing is more routine (as I believe it will be), let’s assume that the company can only get 10% of the market even though it would have an efficacious non-stimulant treatment for ADHD with customized efficacy. 

At the low end of the market, 10% penetration would be $80 million of revenues. Assuming 90% gross margin and $20 million of SG&A (double 2017’s SG&A to include a targeted sales force), EBIT would be approximately $50 million or $40m after-tax.  With a multiple of 10x EBIT or 12.5x Net Income and including potential proceeds from options and warrants of $52 million, the unrisked value of the low-penetration CNTN4 program would be approximately $575 million.  Assuming the shares outstanding double to 145.5 million shares to complete a pivotal trial (even though I have mentioned above that it could be a small trial due to the expected high efficacy) and fund milestones, the unrisked value per share is nearly $4.00.  So what is the right risk factor to use?  I find this analysis quite useful: https://www.bio.org/sites/default/files/Clinical%20Development%20Success%20Rates%202006-2015%20-%20BIO,%20Biomedtracker,%20Amplion%202016.pdf .  It suggests on page 18 a 33.8% chance of success for a biomarker-based Phase 2 to reach approval and a 72.3% chance for a biomarker-based Phase III.  Here I would use an average of the two (53%) as a Phase 2 has already generated positive efficacy data in CNTN4 and the drug is clearly safe and well tolerated.  Thus a “risked” value per share is nearly a double from here.  You could also discount it back 3 years at a 8% discount rate (because already risked) to reduce the targets by 20%. 

Very quickly, the autism CNTN4 program could be worth approximately an additional $1.25 unrisked and $0.50 risked to 33.8% as a Phase II using a 3.3% incidence rate in an autistic population from Roohi et.al. https://www.omim.org/entry/607280  , 9.2 million births per year in the US and Europe, a relatively new guideline of a 1 in 59 overall autism incidence rate (https://www.autismspeaks.org/science/science-news/cdc-increases-estimate-autism%E2%80%99s-prevalence-15-percent-1-59-children).  These assumptions would suggest about 4,400 new addressable cases per year which I multiply times 50 to approximate the treatable population (people born every year for 50 years).  I also assumed a $1,250 annual price point, 10% penetration, 90% gross margin, $10 million of SG&A and a 12.5x P/E. If the ADHD program does not move forward, GNMX could likely charge a lot more for a rare disease drug in autism and it could be voucher eligible. 

It is too early to ascribe any value other than option value to the pediatric crohn’s disease program and my assumptions do not suggest it is worth very much until the company offers more detail.  For reference, 20% of crohn’s is pediatric (http://www.seattlechildrens.org/medical-conditions/digestive-gastrointestinal-conditions/crohns-disease/ ) and the incidence rate is 1 in 5,000 (https://health.clevelandclinic.org/signs-child-might-crohns-disease/) and the target population with releavant mutations would represent approximately 12.5% of that group.  Thus with 15 years worth of patients  in the US and Europe and a $25,000 price point (same as pre-biosimilar Remicade although this would be a rare disease when genetically sub-defined and could potentially have a much higher price point), sales with 10% penetration would be $8.6 million.  That is not very exciting.  Risking the $100 million voucher value to a 25.9% chance of success for a biomarker-based phase I (or II) would add some incremental value.     

Finally, there is a pre-clinical ultra-orphan auto-immune pediatric disease that won’t be ready for the clinic for one to two years.  Very interesting because there clearly is a genetic signal in the CHOP database but it is far too early to ascribe any value yet.  This will certainly not be the last program to come out of the platform. On conference calls, management has mentioned inbound inquiries for new programs as well.

If CNTN4 in ADHD fails then the autism value would also be severely discounted and I would expect the stock to be cut in half as the primary value would be the pediatric crohn’s drug and an unproven rare pediatric disease / voucher generation platform.   Those programs could ultimately become the basis for a successful company but they will take more time and ,of course, capital.  

Disclosure

We make no claims, promises or guarantees about the accuracy, completeness or adequacy of the contents of this document and expressly disclaim liability for errors and omissions in the document.  We have no obligation to update this document.  We may change our position at any time without posting an update.  The views expressed here are merely the opinion of the author.  Readers should do their own research.

Catalyst

Data readout at end of the year for ADHD, progress in other programs or annoucement of new programs