The collapse of Vertex's (NASDAQ:VRTX) hepatitis business has knocked the share price down about 30% from its recent high of $89.96 for a current enterprise value of $13 billion. Although the hepatitis business is not recoverable, Vertex is expecting results from pivotal phase 3 clinical trials on a new treatment for cystic fibrosis in April. If Vertex's new cystic fibrosis drug succeeds only in patients with two copies of the delta-F508 mutation, cash flows from lumacaftor alone will be $1.45 billion, for a risk-adjusted estimated share price of $80. Should Vertex's new drug also receive approval for use in patients with just one copy of the delta-F508 mutation, cash flows could easily be $3.42 billion, yielding an expected share price of $185. Given the powerful and incontrovertible outcomes seen in phase 2 trials, I estimate that Vertex has a greater than 90% likelihood of receiving approval for patients with two copies of the delta-F508 mutation. There is not enough information available at this time to determine how likely it is that Vertex will hit the jackpot and receive approval for patients with single copies of the mutation as well. Still, the market is offering investors a pretty good deal: a likely 30% return with a decent chance at 200%.
Cystic fibrosis is one of the most common deadly genetic diseases in the world. Predominantly affecting Caucasians, cystic fibrosis is the result of mutations in the cystic fibrosis transmembrane conductance regulator, or CFTR. The CFTR is a chloride channel whose function is to secrete chloride into mucus. When there is not enough chloride in the mucus, the mucus lacks the osmotic pressure to pull water into the carbohydrate components and it becomes too sticky. This results in a plethora of clinical manifestations:
- Sticky mucus in the lungs, resulting in recurring infections, lung abscesses, bronchiectasis (dilation of the bronchi), pulmonary hypertension, fungal invasion, pneumonias, and bleeding into the lung from expanded bronchial arteries.
- Sticky secretions in the pancreas, resulting in premature pancreatic failure. The pancreas becomes replaced by fatty tissue and the CF patients experience malnutrition from lack of pancreatic digestive enzymes. They also develop diabetes.
- Sticky secretions in the liver, resulting in early liver failure. It is common for cystic fibrosis patients to need both liver and lung transplants.
- Sticky mucus in the intestines, resulting in bowel obstructions by inspissated bowel contents (distal intestinal obstruction syndrome)
There are many mutations in the CFTR that can result in cystic fibrosis. Here they are, with percentage prevalence:
Source: Human Mutation (2002) 19: 575-606.
To manifest clinically with cystic fibrosis, patients need two mutations in the CFTR gene. The probability of having mutations with fractional prevalence p and q is given by: fraction with two P mutations = p2; fraction with 1 P and 1 Q mutation = 2pq; and fraction with 2 Q mutations = q2. Thus, among all patients, 90% will have at least one delta-F508 mutation, 47% will be homozygous for delta-F508 (0.47 = 0.682), 43% will have only 1 delta-F508 mutation, and 4.2% will have at least one G551D mutation.
Each mutation has slightly different biology. In the case of the G551D mutation, the mutant receptor is properly trafficked to the cell membrane, but then fails to function properly. Vertex's drug ivacaftor binds to the mutant CFTR with the G551D mutation and restores its function. Patients with G551D mutation who are treated with ivacaftor have dramatic responses, with some patients even normalizing their sweat chloride levels. The market response to these extraordinary results has been breathtaking: a year after FDA approval of ivacaftor, essentially all patients in the developed world with the G551D mutation had been placed on ivacaftor therapy for quarterly revenues of $100 million. Everybody in the cystic fibrosis world knows about this drug and Vertex Pharmaceuticals. If their next clinical trial succeeds, we can expect immediate and universal adoption of the drug.
The biology of the delta-F508 mutation is slightly more complicated than the G551D mutation. The deletion of phenylalanine-508 results in a protein that fails to traffic properly to the cell membrane and also fails to function properly if it gets there. Fortuitously, ivacaftor rescues the function of properly-trafficked CFTR with the delta-F508 mutation. The problem, therefore, is to overcome the trafficking problem.
In vitro assays demonstrated that lumacaftor (VX-809) potentiates the action of ivacaftor. It has now been demonstrated that lumacaftor succeeds in changing the shape of the CFTR within the cell so that enough escapes from the endoplasmic reticulum to the cell surface for ivacaftor to rescue the function. In other words, lumacaftor solves the trafficking problem while ivacaftor solves the function problem.
This combination therapy has undergone phase 2 clinical trial. The trial was an extremely well-designed randomized double-blind placebo-controlled trial. This is the gold standard in clinical research; better trial design does not exist. The trial enrolled patients had two copies of the delta-F508 CFTR. Patients were assigned to these groups:
- VX-809 for 28 days, followed by VX-809 + ivacaftor for 28 days
- Placebo for 56 days
This trial designed allowed several key evaluations. First, the placebo effect is easily measured. Patients on placebo saw a decrease in lung function over the first 28 days (-0.9% FEV1), followed by a further decrease in lung function over the next 28 days (-2.5% FEV1). (FEV1 is how much air a patient can blow out in 1 second.) This is what you would expect: patients with cystic fibrosis and receiving no treatment experience gradual decline in FEV1. The typical clinical course is not monotonic; patients experience transient improvements in lung function followed by exacerbations, with inexorable decline over a period of years until they can't breathe.
The next comparison is patients who received VX-809 for the first 28 days versus placebo. These patients had a statistically significant decrease in the amount of sweat chloride, indicating a physiologic response to the drug. However, these patients also experienced a decrease in lung function over the first 28 days (-2.9% FEV1), which was not statistically different from placebo alone. In other words, VX-809 by itself has a measurable physiologic effect but not a clinically important effect.
When patients who were receiving VX-809 then also received ivacaftor (days 29-56), there was an immediate improvement in lung function: +6.1% FEV1 compared to their baseline. Compared to placebo, patients had +8.6% FEV1. Clinically, this is huge: on average, with maximum standard therapy, patients lose about 1% FEV1 per year. VX-809 in combination with ivacaftor turned back the clock in disease progression by 6-8 years after only 28 days of treatment. There were no major adverse effects that were different than placebo (although many patients experienced lung problems during the trial, which is the norm for cystic fibrosis patients.) Clinical trial results don't get better than that -- ever. Results were highly statistically significant, with p < 0.001.
Multiple phase 3 clinical trials are ongoing with the ivacaftor-lumacaftor combination. They can be found here. In brief, the company is looking at randomized, double-blind, placebo-controlled trials in delta-F508 homozygotes. The trials are similar to the successful groups in the previous phase 2 trials and are almost certain to succeed (>90% likelihood). The estimated end date is in April 2014 as reported on clinicaltrials.gov. Recruitment is complete and the trials have fixed duration of 24 weeks, so there's no reason to think that results will come out later than April. A phase 2 trial including delta-F508 heterozygotes is currently recruiting patients and is expected to yield results by March 2014.
Vertex has two additional drugs in this space, VX-661 and VX-983. The phase 2 trial with VX-661 was released in April 2013. Trial design was the same as for VX-809 and results were also highly favorable. Clinical data for VX-983 is not yet available.
In terms of chemical structure and synthesis, these drugs all have favorable features.
Structure of ivacaftor.
Ivacaftor is notable for significant ease of synthesis. The molecule is small (molecular weight 392.5), lacks any chiral centers, and contains mildly reactive groups. The molecule is metabolized in the liver by the cytochrome P450 system, so it is possible that drug interactions will occur with other molecules that depend on liver metabolism such as warfarin or cimetidine. Vertex is directly addressing the issue of possible interactions with other medicines used in cystic fibrosis through clinical trials. Ivacaftor is orally available with twice-daily dosing.
Structure of lumacaftor.
Lumacaftor also demonstrates decent chemical qualities, although slightly less positive than ivacaftor. There are no chiral centers. The molecular weight is larger (452.4), but still below the molecular weight cutoff of 500 seen in most commercially viable small-molecule drugs. From a chemical perspective, lumacaftor is somewhat more challenging to synthesize than ivacaftor but it remains fairly straightforward compared to the feats of chemistry seen in the high-end total synthesis literature. Lumacaftor is available orally with once-daily dosing.
Value of the market
According to the American Lung Association, there are 30,000 Americans with cystic fibrosis and approximately 1,000 new cases per year in the United States. The disease occurs in 1/2500 Caucasians, 1/13500 Hispanics, 1/15100 African-Americans, and 1/31000-1/100000 Asians. The population of Western Europe multiplied by the US incidence data yields ~40,000 patients in Western Europe. This yields global market sizes of ~70,000 x 0.041 = 2870 for G551D mutation; 70,000 x 0.47 = 32,900 for delta-F508 homozygotes; and 70,000 x 0.43 = 30,100 for delta-F508 heterozygotes. If Vertex's drug works on both homozygotes and heterozygotes, the total global market for lumacaftor is 63,000 patients. We assume that essentially no meaningful non-American, non-European market will exist due to the lower incidence of CF in the rest-of-world populations.
We estimate the dollar size of the delta-F508 market based on the revenues from ivacaftor in the G551D mutation. As we know, approximately 2.1% of all cystic fibrosis alleles are G551D. 0.04% of patients are homozygous for G551D, 4.1% of patients are heterozygous for G551D, and 96% of patients are negative for G551D. Ivacaftor has been adopted by nearly all patients with the G551D mutation in the developed world (according to the most recent 10-Q filing). Revenues are $100 million per quarter, or $139,000 average realized revenue per patient. In the worst-case scenario for lumacaftor, it will work only in the 47% of cystic fibrosis patients who are homozygous for delta-F508. In that case, revenues would be $1.13 billion per quarter. Alternatively, if both heterozygotes and homozygotes respond to the combination therapy, 90% of patients with cystic fibrosis will be eligible for combination therapy and the market will be $2.17 billion per quarter. Because the underlying genetic defect cannot be reversed, cystic fibrosis patients will require Vertex's medications for the durations of their lives.
In the most recent quarter, Vertex took in $187 million in product revenues. SG&A was $87.8 million, royalties were $7 million, and production costs were $20 million. As a loose approximation of product margins, we get (187-114.8)/187 = 39% pre-tax. Assuming a tax rate of 35%, we get a net profit of: (1.13 x 4 x 0.39 x 0.65) = $1.15 billion annually as a lower bound, or $2.20 billion (2.17 x 4 x 0.39 x 0.65) as an upper bound. However, those calculations are with the conservative assumption that there will be no economies of scale and that royalties will also scale proportionately. If we relax those assumptions and instead project that SG&A will increase 3x (instead of growing >11x), we come up with (1.13 - 3 x 87.8 - 11.4 x 27) x 4 x 0.65 = $1.45 billion in the homozygotes-only case or $3.42 billion in the homozygotes + heterozygotes case. At a PE of 15, this gives a target price of $22-$51 billion, or $94-$218 per share. We can discount this for risk by 15%, which yields a value of $80 in the homozygotes-only case to $185 per share in the homozygotes + heterozygotes case. For extra conservatism in our estimate, we have completely discounted both the value of Vertex's pipeline and its existing $400 million/year revenues from ivacaftor used in the treatment of cystic fibrosis patients with the G551D mutation. We have also made the conservative assumption that Vertex will not charge any more for ivacaftor/lumacaftor combination therapy compared to ivacaftor therapy alone.
The company has a strong balance sheet and will have no trouble completing the trials and bringing the drugs to market without further dilution. The company has $1.4 billion in cash and reports positive equity of $1.2 billion. Looking forward, quarterly revenues will be about $100 million from ivacaftor; that revenue is reliable and will be sustained for years. Cost of product will continue to be $20 million/quarter, royalties will continue at $8 million/quarter, R&D will probably increase to $250 million/quarter (largely to fund multiple late-stage clinical trials), and SG&A should decrease to $80 million/quarter. In other words, the company has the cash to make it through about 5-6 more quarters without risk of additional dilution. My expectation is that they will have an approved drug in hand in 4 quarters and will start generating hugely increased revenue in 5 quarters.
If the cystic fibrosis franchise doesn't work out, the company is definitely up a creek. They are developing a rheumatoid arthritis drug that targets Jak3, but there is a plethora of treatments for rheumatoid arthritis ranging from prednisone and methotrexate to biologics like infliximab and etanercept. It's difficult to come up with a huge blockbuster in that space due to the very tight competition that already exists. For that reason, I do not include Vertex's rheumatoid arthritis pipeline in the company valuation, despite extremely promising phase 2 clinical trial results.
There are two classes of potential competition in the cystic fibrosis space: 1) traditional therapy; and 2) CFTR potentiator therapy. A thoughtful note has already been written on this subject. The closest similar drug in late-stage (phase 3) trails is ataluren (PTC Therapeutics). Ataluren is a nonsense mutation suppressor, which allows the ribosome to read through stop codons. In cystic fibrosis patients where mutations result in an abnormal stop codon appearing in the CF gene, ataluren was essentially no better than placebo (although PTC jiggered the results a little to give one result a p-value of 0.0478).
Mean relative change in %-predicted FEV1 (ITT population)
Ataluren vs. Placebo
All patients at Week 48
All patients - average over all post-baseline visits
Patients not on chronic inhaled antibiotics
Reduction in pulmonary exacerbation rate over 48 weeks (ITT population)
Ataluren vs. Placebo
Patients not on chronic inhaled antibiotics
Ataluren data from here.
Ataluren has no effect on gene expression in the delta-F508 mutation or the G551D mutation. As such, it is not truly a competitor to Vertex's drugs.
More worrisome to investors in Vertex are the drugs being developed at N30 Pharmaceuticals. These drugs are S-nitrosoglutathione reductase inhibitors (GSNORis). This class of drugs has a dual effect that includes reduction of inflammation and also modulation of CFTR activity. There are currently phase 1 trials examining the activity of GSNORi N6022, an injected drug. First dosing occurred in March 2013; no results are yet available. An oral drug with similar activity, N9115, is in preclinical development and is expected to go to phase 1 clinical trial in 1H 2014. The laboratory data on N6022 are quite promising, but not slam-dunk.
Source: N30 Pharmaceuticals.
In the experiment above, N30 studied chloride currents (ISC) in cultured human airway epithelium cells; positive deflections indicate chloride flowing through the CFTR. In normal cells, currents of 30-40 microamps/cm2 could be triggered with the agents forskolin and genistein. In delta-F508 cells, hardly any current could be triggered by those compounds (black line, bottom graph). With the addition of N30's compound "corrector K," there was partial rescue of the chloride flow (green line, bottom graph). However, note the dramatically different scales on the Y-axis. At best, the N30 correctors yield a few percent of normal CFTR activity. See table below for degree of CFTR correction.
Source: N30 Pharmaceuticals.
Is less than 6% correction enough? It's an open question that can only be answered through clinical trial. Right now, there's only in vitro data to go on. Vertex writes this in Van Goor 2011:
The pharmacology of VX-809 was assessed in cultured human bronchial epithelial (HBE) cells isolated from the lungs of seven patients with CF homozygous for the F508del-CFTR mutation (F508del-HBE). Incubation of F508del-HBE with VX-809 for 48 h increased CFTR maturation by approximately eightfold, with an EC50 of 350 ± 180 nM, and enhanced chloride transport by approximately fourfold, from 1.9 ± 0.4 μA/cm2 to 7.8 ± 1.3 μA/cm2, with an EC50 of 81± 19 nM. This corresponded to an increase in chloride transport from 3.4 ± 0.7% to 13.9 ± 2.3% of that measured in HBE isolated from four non-CF donor lungs (56 ± 6 μA/cm2).
If we examine the similar experiments from N30 Pharmaceuticals, their best result was a chloride current of 4 microamps/cm2 in a similar system of bronchial epithelial cells. Because these experiments were not performed in parallel, quantitative comparison is on shaky ground. However, it is my belief that the upshot of the N30 Pharma experiments is that their drug is probably inferior to Vertex's. It is possible that the anti-inflammatory effects of N30's drugs will be important and will add efficacy beyond the CFTR effect. Overall, it looks like N30 has a drug that can give at most 6% CFTR activity, whereas Vertex's drugs boost CFTR activity to 13.9% of normal. Those differences might not be clinically significant, but I think that the available information suggests that Vertex has a leg up here.
Another possibility is that the drugs will be complementary. Lumacaftor and the GSNORis have different mechanisms of action and it is possible that patients would benefit from receiving both lumacaftor and a GSNORi as combination therapy. The key issue is that we're not looking at another Incivek-type disaster, where the new drug will totally displace Vertex's offering. Because the current N30 drug is injectable instead of oral, it will not be able to compete with ivacaftor/lumacaftor. The next-generation GSNORi (N9115), which will be orally available, has not yet entered clinical trials. N30 will not have any credible oral competition to Vertex's drugs before 2016 at the absolute earliest, and probably not before 2017. I suspect that if N30's early clinical trial results are compelling, Vertex will try to acquire N30. Such a merger would have powerful synergies for both companies and would create massive shareholder value for both Vertex and N30 shareholders because the CFTR potentiator pipeline would be completely locked up, much like Gilead's (NASDAQ:GILD) HIV franchise.
To the best of my knowledge, there are no other remotely close competitors out there. Synthesizing new CFTR correctors from scratch will take years. Because Vertex is offering a combination therapy, that approximately squares the effort needed to duplicate their results. It will be a significant challenge for another company to break into this space within the next 5 years and perhaps even within the next 10.
In terms of standard therapy, these drugs are always advancing and include a variety of antibacterials and drugs that soften mucus such as N-acetylcysteine, tobramycin, aztreonam, DNase, and colistin. These drugs are complementary to CFTR therapy rather than being true competition, and patients will probably be on both groups of drugs. At present, Vertex is performing clinical trials to determine if their therapy will cause drug interactions with select standard therapy drugs. While drug interactions are hard to predict, it is likely that Vertex will have no problem with standard therapy drugs, especially since many are inhaled and have little systemic absorption.
Vertex has stellar science, so much so that I think Vertex's scientists will eventually win the Nobel Prize for their work on cystic fibrosis. Will investors win, too? Their drugs are very late-stage and have already been validated in small trials using the gold standard of trial design. The phase 3 clinical trials are very similar to the phase 2 clinical trials and will almost definitely succeed in patients with two copies of the delta-F508 mutation. There is not yet enough clinical trial data to assess whether Vertex will also capture patients with only one copy of the delta-F508 mutation. The only potential serious competition at this point comes from N30 Pharmaceuticals, which is still several years behind in clinical development and has probably created an inferior class of drugs. If N30 generates significantly promising results, acquisition by Vertex will probably soon follow, benefiting shareholders of both companies. After discounting for risk, the share value attributable to ivacaftor/lumacaftor combination therapy is $80 if the combination works only in patients with two copies of the delta-F508 mutation. If ivacaftor/lumacaftor therapy is also approved for patients with only one copy of the delta-F508 mutation, shares could easily be worth $185. VRTX is a moderate risk and could end up being a near-total loss, but I think the likelihood of failure in the phase 3 trials for delta-F508 homozygotes is only 10%.
Appendix: Clinical Trial Calendar
- TRAFFIC (phase 3 ivacaftor/lumacaftor in delta-F508 homozygotes), 559 enrolled, results due April 2014. Recruitment complete. NCT01807923.
- TRANSPORT (phase 3 ivacaftor/lumacaftor dosing comparison in delta-F508 homozygotes), 563 enrolled, due April 2014. Recruitment complete. NCT01807949.
- Phase 2 ivacaftor/lumacaftor study in delta-F508 homozygotes and heterozygotes, results due March 2014. Currently recruiting. NCT01225211.
- Phase 1 lumacaftor/antibiotic interaction study. Results due any time. NCT01768663.
- Phase 1 study of ivacaftor/lumacaftor pharmacokinetics in children, currently recruiting, results due any time. NCT01897233.
- Phase 1 study of ivacaftor/lumacaftor on QT/QTC prolongation, due March 2014, recruiting. NCT01910415.
- Phase 1 study of ivacaftor/lumacaftor in patients with hepatic impairment, due November 2013. NCT01888393.
- Phase 1 study of N6022, 1 infusion/day for 7 days. Currently recruiting. Aiming for enrollment of 67. Results due January 2014. NCT01746784.
Additional disclosure: Please perform your own due diligence prior to investing in this or any other security.