SkinTE and the FDA's 361 Pathway

Doctors who perform wound treatments have limited options for skin repair. A new entrant into the wound treatment space is SkinTE, a product sold by PolarityTE. SkinTE has already been used to treat a first group of patients and has received early plaudits. The company describes SkinTE as a patient-specific, single application product created from a small piece of the patient’s own skin. As described in PolarityTE’s 2017 10K SEC filing, SkinTE is available for:

the repair, reconstruction, replacement and regeneration of skin (i.e., homologous uses) for patients who have suffered from wounds, burns or injuries that require skin coverage over both small and large areas of the body.
— from 2017 Form 10K

SkinTE is described as an impressive, all-encompassing product. The technology behind this product appears to be the company’s “TE” platform based on:

… our ability to create minimally polarized functional units, or MPFUs, which contain polarizing multi-cellular aggregates capable of expanding, proliferating and synthesizing those cells, materials, factors and/or systems we believe are necessary for integrative full-thickness three-dimensional tissue regeneration, not simply two-dimensional cell sheets. Instead of starting with artificial materials, synthetic factors and/or altered cell suspensions, our platform begins with the patient’s own (autologous) tissue and those components, appendages and substrates we believe are necessary for the development of an expandable and self-propagating complete system.
— From 2017 Form 10K

More specifically to SkinTE, the company describes the procedure and strategy:

With SkinTE, often within 24-48 hours of the initial skin harvest, our product is applied to the patient, whose body then provides a receptive environment and nutrients for controlled healing. By both preserving the tissue’s natural microenvironment and using the patient’s body as an intrinsic bioreactor—which means using the body’s own natural biological healing process rather than a manufactured or engineered environment to support the regenerative process—we believe a patient’s own tissue can be regenerated, along with its natural coloring and texture, layers and structure, hair and appendages.
— From 2017 Form 10K

This description is accompanied with a graphic overview of the SkinTE procedure that, admittedly, appears very simple to perform (right).

The company asserts that the TE platform has potential to transform tissue regeneration in multiple tissue types such as bone, muscle, fat, cartilage, nerves, and blood vessels in addition to skin. Given these strong claims, the technology behind SkinTE is certainly of interest. However, finding a detailed protocol for the technology and preparation behind SkinTE is not a trivial task.

Fortunately, Dr. Denver Lough, Chairman of the Board, President, Chief Executive Officer, and Chief Scientific Officer of PolarityTE has filed patent applications numbered 14/954,335 and 15/650,656 both titled “Methods for development and use of minimally polarized function cell micro-aggregate units in tissue applications using LGR4, LGR5 and LGR6 expressing epithelial stem cells.” This invention is largely summarized as:

[0014] The invention provides in a first embodiment a minimally polarized micro-aggregate multi-cellular composition including isolated living LGR expressing cells and a multi-dimensional support selected from the group consisting of scaffolding, collagen, matrix, particle, and fiber.
[0015] The invention provides in a further embodiment to the previous embodiment a minimally polarized micro-aggregate multi-cellular composition including isolated living LGR expressing cells and a multi-dimensional support selected from the group consisting of scaffolding, collagen, matrix, particle, and fiber where the LGR expressing cells are supplemented with growth factors and where the LGR expressing cells are selected from the group consisting of LGR4, LGR5 and LGR6.
[0016] The invention provides in a further embodiment to any of the previous embodiments a minimally polarized micro-aggregate multi-cellular composition including isolated living LGR expressing cells and a multi-dimensional support selected from the group consisting of scaffolding, collagen, matrix, particle, and fiber where the LGR expressing cells are supplemented with migratory/recruiting analytes and the LGR expressing cells being selected from the group consisting of LGR4, LGR5 and LGR6.

In effect, the invention appears to focus on a special population of LGR4, LGR5 and LGR6 expressing cells potentially supported by a matrix and supplemented with growth factors. Data behind these claims is contained in two papers:

Lough et al, 2013 Plast Reconstr Surg 132: 1159-1171 (link)
Lough et al, 2016 Plast Reconstr Surg 137: 495-507 (link)

This can be verified by side-to-side comparisons of sample figures in the above two papers and figures presented in the patent application. For example, in the first comparison below, Figures 12A and 12B from the patent application are the data published in Figure 6 of Lough et al (2013). In the second comparison, the data in Figures 5A-D of the patent application are the data published in Figure 5 of Lough et al (2016).

Patent Application 14/954,335 Figures 12A, B

Patent Application 14/954,335 Figures 5A-D

Lough et al (2013) Figure 6

Lough et al (2016) Figure 5

In Lough et al (2016), the isolation procedure for the cell population is described as generally following the protocol of Snippert et al (2010 Cell 143: 134-144), specifically maintained in:

Dulbecco’s Modified Eagle Medium/F12 (Sigma Chemical Co., St. Louis, Mo.), 10% fetal bovine serum (Atlas Biologicals, Fort Collins, Colo.), 10 ng/ml epidermal growth factor (Invitrogen, Waltham, Mass.), 20 ng/ml basic fibroblast growth factor (Invitrogen), 20 nM hydrocortisone (Sigma), 5 μg/ml insulin (Sigma), and 100 U/ml penicillin/streptomycin (Hyclone Laboratories, Logan, Utah) at 37°C in 5% carbon dioxide until passage 1.

The cells were validated for LGR6 mRNA expression, and further cultured as described prior to testing in preclinical wound models:

The scaffold/matrices were cut into 1-mm^2 pieces and transferred aseptically to 12 × 75-mm polystyrene tubes (BD Falcon, Franklin Lakes, N.J.) and 500 μl of the cells was added (1 × 10^6 total cells). The caps on the tubes were loosened to allow for gas exchange and were placed in a 37°C, 5% carbon dioxide incubator for 3 hours. The matrices were agitated every 15 minutes by gentle shaking of the tube to resuspend any cells that may have settled out. After 3 hours, the matrices were removed aseptically from the tubes and placed individually into wells of a 24-well plate containing growth media. After 24 hours, the matrices were examined under fluorescent microscopy for the adherence of the cells to the matrices. Culture medium was changed every 3 days.

In short, the procedure described by Lough et al (2016) calls for isolation of an LRG6 expressing epithelial cell population that is then cultured in a growth hormone containing medium and allowed to adhere to a matrix prior to placement in the murine wound model. As a result, the authors conclude:

In this article, we present a new method of providing skin to those wound beds that are intrinsically incapable of healing themselves.

FDA Registration Pathway

Typically, new treatments for use in the clinic must go through the FDA and be assessed for their safety and efficacy in clinical trials. However, pathways do exist to market products without safety or efficacy data in humans. One such pathway is to register with the FDA as a 361 HCT/P (Human Cells, Tissues, and Cellular And Tissue-Based Products) designation as described in Section 361 of the Public Health Service Act and 21 CFR Part 1271 (link to pdf). This is indeed the pathway used by PolarityTE. In the 2017 10K filing, PolarityTE describes SkinTE as a registered 361 HCT/P with the FDA:

This can further be confirmed via the FDA’s Human Cell and Tissue Establishment Registration portal wherein SkinTE is confirmed as an HCT/P under 21 CFR 1271.10 (screenshot right):




As described by the FDA, an HCT/P is regulated solely under section 361 of the PHS Act and 21 CFR Part 1271 if it meets all of the following criteria (21 CFR 1271.10(a)):

 1) The HCT/P is minimally manipulated;

 2) The HCT/P is intended for homologous use only, as reflected by the labeling, advertising, or other indications of the manufacturer’s objective intent;

 3) The manufacture of the HCT/P does not involve the combination of the cells or tissues with another article, except for water, crystalloids, or a sterilizing, preserving, or storage agent, provided that the addition of water, crystalloids, or the sterilizing, preserving, or storage agent     does not raise new clinical safety concerns with respect to the HCT/P; and

4) Either:

  1. The HCT/P does not have a systemic effect and is not dependent upon the metabolic activity of living cells for its primary function; or

  2. The HCT/P has a systemic effect or is dependent upon the metabolic activity of living cells for its primary function, and a) Is for autologous use; b) Is for allogeneic use in a first-degree or second-degree blood relative; or c) Is for reproductive use.

The descriptions in the patent application and published papers regarding PolarityTE’s technology raise the question of how SkinTE meets the first, second, and third criteria listed above.

Minimally Manipulated

Under the FDA guidance, skin is considered a structural tissue and serves to “physically support or serve as a barrier or conduit, or connect, cover, or cushion in the donor.” This refines the minimally manipulated criterion to ask if the processing of the tissue “alters an original relevant characteristic of the tissue relating to the tissue’s utility for reconstruction, repair, or replacement as structural tissue.”

As noted in a previous PolarityTE presentation and shown in their overview of the procedure, the SkinTE product starts as a piece of excised, full thickness skin and is then processed to yield an applied paste. This processing would appear to exceed the description of minimally manipulated, and an analogous method is specifically described as such in the FDA guidance under Example 10-4b in 21 CFR 1271.10 (right; pdf link)

Further to this point, Example 11-3 in the FDA guidance makes further clarification (right). As described in PolarityTE’s presentations and in the relevant literature, the procedure behind SkinTE no longer allows the product to maintain the “original relevant characteristics of skin related to its utility as a protective covering.” It is no longer a protective covering but rather a paste due to the processing, nor does the product retain its “dense, strong, and flexible connective tissue layer.”

Homologous Use

The FDA also mandates that products registered under section 361 meet the definition of homologous use. This generally requires the product to adequately perform its same basic function:

Under this guideline, SkinTE would have to be a product wherein its “basic” function to regenerate full thickness skin in the recipient is obvious and commonly understood, and should not require any laboratory, preclinical or clinical studies. This does not apply to the SkinTE product. In effect, if the biological function of SkinTE to regenerate full thickness skin (footnote 1) was well understood and did not require any preclinical or clinical examination, scientific papers attempting to explain and demonstrate this effect would be unnecessary. In effect, in the publication by Lough et al (2016) the authors clearly acknowledge that the group “seeks to determine” whether the LGR6 cell population isolated by fluorescent-activated cell sorting is effective for translational use (caption right). If the function of SkinTE were “basic” as described under FDA guidelines, such experiments to clarify the function of this cell population would be unnecessary.

As it stands, PolarityTE’s description that SkinTE “regenerates full-thickness hair-bearing skin” is a claim quite beyond a basic function under homologous use but rather a putatively significant treatment advancement in need of clinical validation.

Combination of Another Article

The FDA further notes that products registered under 361 may not involve “the combination of the cells or tissues with another article, except for water, crystalloids, or a sterilizing, preserving, or storage agent, provided that the addition of water, crystalloids, or the sterilizing, preserving, or storage agent does not raise new clinical safety concerns with respect to the HCT/P.” Under this guidance, products are not allowed to be mixed with a carrier during manufacture, nor are they allowed to be treated with supplements such as growth hormones during manufacture.

As shown below, both previously referenced publications by Lough in 2013 and 2016 clearly note that the fundamental research behind the PolarityTE patents involves culturing with fetal bovine serum (which contains a collection of growth factors) as well as endothelial growth factor and basic fibroblast growth factor. The conditions also call for the addition of insulin and hydrocortisone, two additional factors that would lie outside of the FDA’s guidance.

Lough et al (2013)

Lough et al (2016)

Lough et al (2016)

It’s important to note that these additives  are not limited to growth factors, nor is citation of their use confined to the published literature. For example, Lough et al (2016; caption right) as well as the cited patent filing (excerpt below) mention the use of a scaffold to support the cells in addition to the aforementioned growth factors.


In fact, relevant excerpts from a PolarityTE patent filing run counter to the FDA’s guidance:

  • [0014] The invention provides in a first embodiment a minimally polarized micro-aggregate multi-cellular composition including isolated living LGR expressing cells and a multi-dimensional support selected from the group consisting of scaffolding, collagen, matrix, particle, and fiber.

  • [0015] The invention provides in a further embodiment to the previous embodiment a minimally polarized micro-aggregate multi-cellular composition including isolated living LGR expressing cells and a multi-dimensional support selected from the group consisting of scaffolding, collagen, matrix, particle, and fiber where the LGR expressing cells are supplemented with growth factors and where the LGR expressing cells are selected from the group consisting of LGR4, LGR5 and LGR6.

  • [0016] The invention provides in a further embodiment to any of the previous embodiments a minimally polarized micro-aggregate multi-cellular composition including isolated living LGR expressing cells and a multi-dimensional support selected from the group consisting of scaffolding, collagen, matrix, particle, and fiber where the LGR expressing cells are supplemented with migratory/recruiting analytes and the LGR expressing cells being selected from the group consisting of LGR4, LGR5 and LGR6.

  • [0017] The invention provides in a further embodiment to any of the previous embodiments a minimally polarized micro-aggregate multi-cellular composition including isolated living LGR expressing cells and a multi-dimensional support selected from the group consisting of scaffolding, collagen, matrix, particle, and fiber where the LGR expressing cells are supplemented with LGR specific binding elements selected from the group consisting of ligand families, R-spondin, EDGF, PDGF, Wnt, VEGF, and antimicrobial peptides and where the LGR expressing cells are selected from the group consisting of LGR4, LGR5 and LGR6.

As clearly described in the bolded segments, the patent filing by PolarityTE consists of claims which include combinations with other components such as scaffolding and growth factor receptor ligands. These additives are not allowed for products registered under the 361 pathway. It remains possible that PolarityTE intentionally removed both the scaffold and growth factors from the marketed version of SkinTE. At a recent investor conference, the company mentions the use of “crystalloids” as a carrier rather than scaffolds. But if they have removed both scaffolds and growth factors, does the product used to treat patients meet the company’s own efficacy claims as published in the scientific literature and related patent filings? What is the final SkinTE product?


In sum, PolarityTE’s registration of SkinTE as a 361 HCT/P has allowed the company to rapidly bring their product to market for sale and use in the clinical setting. However, the FDA guidelines regarding the standards for a 361 HCT/P designation and the SkinTE product as described in whole or in part through PolarityTE presentations, patent filings, and scientific publications authored by the CEO raises questions. Are the production methods behind SkinTE and the final product offered by PolarityTE compliant with the FDA’s 361 standards? Have the scaffolds and growth factors described in the scientific publications and patent filings been omitted from SkinTE? Absent appropriate growth factors and culturing conditions, how do the cells in the excised skin maintain their differentiated state and pattern of LGR receptor expression? Are specific LGR receptor expressing cells isolated via fluorescent-activated cell sorting for use in SkinTE? Has the FDA effectively examined products like SkinTE to assure the public that the standards outlined by 21 CFR 1271 are met and adhered to? In this specific case, PolarityTE is marketing a product that has not been tested in randomized controlled clinical trials. Whether this treatment choice meets the FDA guidelines under section 361, and whether this choice is clinically safe and efficacious are questions in the public’s interest that should be further examined.



Footnote 1: PolarityTE makes a clear claim that “SkinTE Regenerates Full-Thickness Hair-Bearing Skin.” Screenshot below taken from the PolarityTE website.


Incyte's Great Expectations

At some point, all successful companies end up being the target of scrutiny. Wall Street is a fickle place, with (silly) ramping quarterly financial expectations and an army of analysts publishing analyses with the vacuousness and accountability of a Tom Friedman column. In that regard, Incyte is now moving into the stage wherein deserved (and undeserved) criticism is directed at the company.

In general, management has done a rather competent job in developing ruxolitinib to the point where it is now a blockbuster. Outside of a hiccup wherein they chased an LDH subgroup through various trials, they’ve done well to expand indications for the drug. Expansion into polycythemia vera came methodically, and further testing into essential thrombocytopenia and graft versus host disease (where the drug is already receiving anecdotal use) may incrementally increase sales. The next challenge for the company is the proverbial diversification of the revenue stream. Baricitinib was on track to aid in this process, but suffered a recent setback. Nonetheless, the company is working on a broad pipeline, with some agents for well-defined targeted indications (FGFR1-3 inhibitor INCB054828) and some with no clear early direction (PIM Kinase inhibitor INCB053914). There are even a series of programs from the Merus collaboration that remain to be publicized, and a cMET molecule being developed by Novartis (INCB028060).

Outside of ruxolitinib, the major value driver for the company appears to be the IDO inhibitor epacadostat (INCB024360). I’ve written about it previously, as the molecule engenders every assessment from strong support to borderline mockery from the biotech community. Fueling this broad range of sentiment are clinical results to date that have ranged from encouraging (melanoma) to ambiguous (nonsmall cell lung cancer) to inactive (colorectal cancer) depending on the indication. Investors should be mindful to treat upcoming clinical data releases with rigour and skepticism, as is required for any compound. Further, hedges and protection going into the pivotal ECHO-301 melanoma data release are always wise safeguards for long term investors.

But arguably, some of the skepticism towards the company relates to the perceived valuation of epacadostat within Incyte’s market capitalization. However, this frustration appears misdirected. Investors are free to buy and sell shares on the open market, so that valuation is out of management’s hands. And despite management’s unappreciated piece-meal trickle of epacadostat data, overall they’ve not been unduly promotional or unrealistically boastful in their expectations for this or any other program. Nonetheless, moves by management to raise money through public offerings while selling their own shares simply add to the skepticism. All of these data points should be assimilated into an investor’s view of the company.

Conversely, a point in favour of management in the IDO discussion is perhaps one of opportunity and execution. With ruxolitinib, Incyte management have shown themselves to be competent in expanding indications once a molecule is on the market. If they’re able to demonstrate efficacy for epacadostat in ECHO-301 and receive marketing approval, a 1-2 year head start and a methodical approach by management may be able to expand use for the drug. And although those indications may not be front line NSCLC and may not appease valuation critics overnight, steady revenue building along with a developing pipeline may make long term investors happy despite the gnashing of teeth amongst the analyst class.

In my experience, the best companies become richly valued early and remain richly valued for a long time. The key is to assess new data as skeptically as possible, keep a level-headed view of the long term picture, and if warranted, dispassionately jump ship before the captain sounds the horn.

Solving Clinical Trial Equations

Biotech investors are often left to solve a variation of the following equation:

a(OR) + b(PFS) + c(DOR) = OS


  • OR is the magical objective response rate, in %*
  • PFS is the mystical progression free survival, in months
  • DOR is the mercurial duration of response, in months
  • OS is the elusive overall survival, in months (or years if you’re brave)


No problem! A few hard numbers to define coefficients a through c, some reliable OR, PFS, DOR values, and you can triumphantly tap Return to cue Excel’s internal canine to fetch your answer. Unfortunately, none of the cases that matter are ever so easy. And like your high school physics teacher, biotech companies know to present the easiest examples in class, but send you home with problems that bear no relation.

And so we’re left with relatively confusing situations, like Incyte’s epacadostat development program. This program got off the ground with OR data in combination with pembrolizumab for the treatment of advanced melanoma, presented in late 2015. It was quickly signaled that a phase 3 trial would soon begin, hoping to extend PFS in patients receiving combination therapy. But as the data from this small trial matured, its fit within the comparator landscape did not necessarily become clearer.

Screen Shot 2017-08-22 at 10.13.02 AM.png

The above table attempts to provide some clarity to the comparisons that await epacadostat in this disease setting. Although the OR for the epacadostat combinations appear numerically higher than PD-1 monotherapy, the epacadostat phase 3 trial is measuring PFS, not OR. PFS for the combination therapy has been updated only once, at SITC 2016 (right). Inexplicably, the company did not update this melanoma dataset at all for ASCO 2017. 

That leaves us with the immature data presented in 2016 as the bellwether for the phase 3 ECHO-301 trial. But due to the small patient numbers in the ECHO-202 dataset, neither the OR or PFS values are comfortably clear of the PD-1 monotherapy ranges. Attempting to better estimate the PFS data through DOR is not helpful either, as the latter are too immature to provide any meaningful insight. Incyte has hinted these data may be updated by year end, but will that provide the necessary confidence? Given their reluctance to be transparent with the accrued data, it’s fair for investors to be skeptical about the quality of the developing dataset, and the extent of curation it seems to require.

But fear not: Incyte has suggested the consistently high OR across disease indications should give investors all the confidence they need. Below are a stack of tables that attempt to position the epacadostat combination data in various indications versus the results from PD-1 and PD-L1 monotherapy trials. One theme is common: although the OR trend higher in the epacadostat combinations, there are no PFS data available, and the DOR data are too immature to remove any uncertainty about outcomes.

This uncertainty around these data have been compounded by Incyte’s strategy towards their release. At ASCO, the company provided a data snapshot that was months out of date. A convoluted explanation was weaved that included a desire to focus on more “mature” datasets, as well as calls for consistency. In effect, they admitted to having more data, but chose not to disclose it. This triggered a cardinal rule for assessing company communications: If the data aren’t disclosed, they’re likely not favourable.

Although this rule is not infallible, it does capture the essence of most public company disclosure strategies. A company is much more likely to keep the veil in front of poor datasets than to obscure strong data they have in hand. That places Incyte’s ASCO disclosure in a rather poor light, and tempered initial enthusiasm garnered at the time of abstract release.

So are Incyte’s actions around their IDO data disclosures necessarily a bad omen? Not entirely. The immunotherapy doublet landscape is far from formed, and it would not be unusual for a company to limit public dissemination of data on competitive grounds. For instance, pembrolizumab and nivolumab are engaged in a head-to-head battle for NSCLC, and the competitive / strategic choices by their sponsors have had large consequences. In that regard, it was notable Bristol-Myers did not release data from the NSCLC cohort of the ECHO-204 study testing nivolumab plus epacadostat. It’s not controversial to suggest there were NSCLC enrollees in the trial, since Incyte and Bristol-Myers have already announced intentions for a phase 3 trial. Therefore, the lack of disclosure in this instance was likely for competitive reasons.**

On the other hand, it is very likely that Incyte management does feel some pressure to diversify and increase revenues. Although ruxolitinib sales continue to grow, the difficulty with baricitinib’s FDA application has definitely increased the focus on epacadostat. And although the series of phase 3 trial announcements for epacadostat may signal confidence, it is not unreasonable for companies to take steps to more quickly increase their revenues. In that way, the phase 3 trial announcements could be viewed as a calculated risk in a competitive landscape rather than an outward show of confidence. After all, phase 3 trials can always be stopped.

Regardless of the hopes investors may have for this entire development program, a dispassionate look at the totality of the data doesn’t reveal unequivocal promises of success. Risk minimization and protection are prudent, as it wouldn’t surprise if the phase 3 readout next year was the first of multiple disappointments.



* Why is this termed a “rate” when it is presented as a percentage? Nonsensical conventions are maddening. Someone needs to rename it.

** This could be the ultimate bluff by Bristol-Myers to goad Merck into spending its money. At which point, well played.

Reata's Bits and Pieces

Reata is attempting to develop two plant triterpenoid derivatives, bardoxolone (RTA-402, CDDO-Me) and omaveloxone (RTA-408), for a broad variety of diseases. These two derivatives are follow-on candidates from an earlier failed candidate, RTA-401, also referred to as CDDO. These derivatives tend to be worked on in an effort to improve the weak anti-inflammatory properties of the parent triterpenoids. The general consensus is these triterpenoids disrupt the interaction between Keap1 and Nrf2, releasing Nrf2 to upregulate a variety of anti-inflammatory / anti-oxidant genes including thioredoxin, glutathione reductase and superoxide dismutase.

The interaction of CDDO and its derivatives with its target proteins is mediated by relatively nonspecific targeting of exposed thiol groups on proteins. In effect, this isn’t a specific targeting mechanism, and there is even data to suggest that thiol-containing antioxidants (such as glutathione) can alter the binding selectivity of CDDO and its derivatives. Further underlying this poor selectivity is the relatively long list of “target” proteins for CDDO and its derivatives. In addition to Keap1/Nrf2, there are data demonstrating effects on the NF-kB regulatory mechanism (via interference with IK-kB), PPAR, JAK/STAT, TGF, PTEN and mTOR. Rather than providing confidence these derivatives will have an effect, the large number of targets lessens enthusiasm that the drugs have a concerted impact on disease.


2. Clinical Trials

2.1. Phase I Trials

RTA-401 (CDDO)

This parent compound was tried in two phase 1 trials, one aimed at AML and the other enrolling those with solid tumours. In the AML trial, there were no responses. This was no different in the solid tumour trial, as no anti-tumour activity was observed. However, the trial was closed at 7 accrued patients when the 6th and 7th enrollees both developed pulmonary emboli. These two events were in addition to a pulmonary embolism in patient 5 and a deep vein thrombosis event in patient 4. In the publication, all events were deemed related to the study drug. In effect, RT-401 could not be safely dosed in an oncology indication and was abandoned.


RTA-402 (Bardoxolone, CDDO-Me)

This compound is a methylated version of the parent CDDO triterpenoid. This drug was used in a 44 patient phase I trial enrolling patients with solid tumours or lymphomas. Unsurprisingly, and similar to the parent compound, the drug had little to no meaningful efficacy. Two responses were observed, one complete response in a mantle cell lymphoma patient and a partial response in an anaplastic thyroid tumour patient. Despite these two responses, it should be noted there is no follow-up trial in either patient population, suggesting that the company was unimpressed with the trial outcome.

Nonetheless, this trial has the fortuitous observation that patients on drug had an increase in estimated glomerular filtration rate via measurements of serum creatinine. This, along with the disappointing anti-tumour activity, appears to have led to Reata moving the drug to a chronic kidney disease (CKD) population.


RTA-408 (Omaveloxone)

There was a phase I in melanoma called REVEAL. The trial enrolled 11 patients, but no data have been disclosed.


2.2. Phase II Trials

2.2.1. RTA - 402 / Bardoxolone


RTA-402 - CKD / Type 2 Diabetes

The phase II trials of RTA-402 were focused on Stage 3/4 CKD and type 2 diabetes patients. The results of these are not relevant given the results from the Phase III Beacon trial described later. However, the important point to remember is that these trials did not reveal the safety signal observed in the phase 3 Beacon trial.

RTA-402 - Pulmonary Hypertension

Arguably this is the most well developed, active program at this time. They believe the myriad of preclinical mechanisms on display by bardoxolone, including anti-reactive oxygen species (ROS), anti-inflammation and anti-fibrotic, will all contribute to improving performance in pulmonary hypertension (PH) patients. To this end, they’re presenting data mostly pertaining to the connective tissue disease (CTD) classification of pulmonary arterial hypertension (PAH), which is part of WHO group I of PH patients. The various cohorts of the LARIAT trial are as described below, and particular emphasis is on the CTD-PAH group that forms the basis of the now enrolling CATALYST phase 2/3 trial.


The cohorts from the LARIAT trial are generally as below:

Cohorts 1 and 2

These two cohorts are basically identical expect that Cohort 1 is enrolling PAH patients (encompassing both idiopathic or CTD subgroups of PAH) that have pre-entry 6 meter walk distance (6MWD) of 150m>x>450 and a less sick cohort that enrolled those with entry 6 MWD of >450m (Cohort 2).

The baseline enrollment characteristics of the patients enrolled into Cohorts 1 and 2 are below. Of note, and as detailed in their 2015 annual presentation, the baseline 6MWD in the 2.5 mg group of cohort 1 was significantly higher than in the placebo group (despite not being indicated in this table). Also, patients are allowed to be on background therapies.

The above tables show the enrollment for cohorts 1 and 2 plus the efficacy results from cohort 1. None of the individual dose groups were significant versus placebo, and the only favourable comparison was if the data were pooled. However, two important observations about this pooled comparison:

1. There is absolutely no dose response, and if anything it’s inverted. They note this is likely due to baseline patient BMI, with heavier patients in the 2.5 mg group showing no benefit whereas BMI < 30 mg/m2 showed an ostensible benefit (as shown below). This comparison is high suspect, and likely not even worth seriously considering as the number of patients falls to 3 in most of the groups. More likely, this simply reflects the fact this dose group had higher 6MWD at entry.

2. They’ve omitted data from the bardoxolone 10 mg group. The “Baseline Characteristics” table above describing the dose levels and patients enrolled in Cohorts 1 and 2 clearly show that the bardoxolone 10 mg arm in Cohort 1 had 6 patients. However, when the 6MWD comparisons are being made, two patients are omitted and the arm falls to 4 patients only. It’s reasonable to be suspicious about the reason for this omission. There was no explanation in the 2015 Year end report or any follow up data releases.


The CTD-PAH patients across doses were pooled and demonstrated the below change in 6MWD over baseline. Although the change is statistically significant when compared to each patient’s baseline, the difference is not significant (p=0.13) when placebo-adjusted.


They followed up the analysis of Cohort 1 with a brief look at Cohort 2, which are the patients with >450m 6MWD at baseline. Only 6 patients are available, and only one is classified CTD-PAH. These 6 patients from Cohort 2 all received 5 mg bardoxolone, and were compared to the 6 patients (3 CTD-PAH, 3 idiopathic) receiving 5 mg bardoxolone from Cohort 1. It is unclear what this analysis is meant to show. The problems are:

1. The patients are a mix of CTD-PAH and idiopathic patients. In that regard, disease etiology is mixed and is in different proportions between groups (50% CTD-PAH in Cohort 1 vs 17% in Cohort 2).

2. Importantly, Cohort 2 is the healthier arm (>450m baseline 6MWD) but shows identical change in 6MWD when compared to the sicker cohort. So in addition to a lack of dose response for the drug from Cohort 1, the drug seems to provide a larger % gain in 6MWD over baseline for the sicker patients than the healthier ones. That’s extremely unlikely.


Synopsis of Data from Cohorts 1 and 2:

  • No statistically significant increase in placebo-adjusted 6MWD.
  • No dose-response relationship evident for the drug.
  • There are a total of 7 CTD-PAH patients enrolled in the drug arms of Cohorts 1 and 2, but going forward, only 6 of these patients are used in pooled analyses. The 1 CTD-PAH patient in Cohort 2 is omitted.
  • There was on SAE of pneumonia in Cohort 2. It was judged as possibly drug related by the investigator but the company disagrees.
  • 16 patients were enrolled in Cohort 2, but going forward company has dropped the 20 mg arm of this Cohort due to nausea. Any CTD-PAH patients in this dose group appear to have been excluded from any future analyses, and there is no clear indication of the outcomes for the 8 patients (16 enrolled - 6 bard - 2 placebo = 8).


Cohort 3

This cohort is enrolling PAH patients with baseline 6MWD > 150m. Cohort 3a is focused on CTD-PAH, whereas Cohort 3b is looking at the other PAH etiologies. Eligible patients can be receiving zero to two disease-specific PAH background therapies. Patients in the treatment group are titrated from 5 mg to 10 mg doses based on tolerability. The available data need to be imputed from a recent press release, as the data aren’t explicitly broken down. However, in SEC filings the company has disclosed that 24 total patients have been enrolled in this cohort as of YE2015.

As of October, 2016, the company notes that a total of 22 CTD-PAH patients have been enrolled in Lariat’s Cohorts 1, 2 and 3a, with 15 receiving bardoxolone and 7 on placebo. In the press release, they note 6 of the bardoxolone patients were from Cohort 1, meaning that the 2 placebo patients also carried over. It is therefore unclear where the other 9 bardoxolone patients come from, whether from Cohort 2 or 3a. They’ve only indicated 1 CTD-PAH patient enrolled in Cohort 2, and that person was excluded from the pooled analyses of Cohort 1 and 2. It’s unclear why that patients would be included now but not before. That leaves the balance of the 9 remaining patients to be from Cohort 3a.

Nonetheless, their Cohort 1+2+3 pooled analysis showed a change in 6MWD that was not statistically significant when placebo-adjusted. They subsequently (and retroactively) removed patients with moderate to severe anemia, resulting in 1 bardoxolone and 2 placebo patients being omitted from the analysis, leaving 14 drug and 5 placebo patients. They note that this comparison is statistically significant when placebo corrected. But it’s clear from the table below that these exclusions were motivated with the outcome in mind:


Notably, the exclusions disproportionately decreased the performance of the placebo group while improving the performance of the drug arm. This is especially strange considering these patients with anemia are “excluded… because treatment with iron supplementation or erythropoietin post-randomization can affect 6MWD values independent of study drug effect.” If anything, treatment with iron supplementation or erythropoietin should improve the 6MWD. It is notable that after exclusion, the change from baseline 6MWD for the placebo group decreased, as would be expected if you remove patients that likely improved their 6MWD due to the effects of erythropoietin. Yet, strangely, the drug group’s 6MWD improves when excluding the patient that received iron supplementation or erythropoietin.


Synopsis of Data from Cohort 3:

  • No clear presentation of the Cohort 3 data is made by the company, and rather they opt for a pooled analysis that obscures the data from the cohort. This is historically a bad sign.
  • The post-hoc exclusion of patients with moderate to severe anemia is clearly a ploy to make the data look better than they are.


Cohort 4

The fourth cohort began enrolling in September 2015, consisting of PH-ILD (pulmonary hypertension-interstitial lung disease, WHO group 3) patients. There are no data available from this cohort, although their filings indicate 7 patients enrolled as of YE2015. More concerning, there was one SAE leading to death. The company’s explanation speaks for itself:

We previously reported that a serious adverse event, or SAE, involving a patient death had occurred in cohort 4a and that the investigator had initially reported it as possibly related to study drug. The investigator has since changed his evaluation to unlikely related. In addition, the Protocol Safety Review Committee that oversees safety for the LARIAT trial concluded that the SAE was unlikely treatment-related.


RTA-402 - Chronic Kidney Disease caused by Alpert Syndrome

The company has recently announced a phase 2 trial in patients with Alpert syndrome. These patients are at heavy risk for developing end-stage renal disease, and Reata believes bardoxolone can increase estimate GFR in these patients in order to delay progression. There are no data available, and given their phase 3 experience with bardoxolone, this is a high risk trial.


2.2.2. RTA-408 / Omaveloxone

Omaveloxone is a close structural analog of bardoxolone methyl that was developed to improve tissue distribution. It basically has the same anti-inflammation / anti-oxidant method of action. No solid clinical data are available from this drug, other than these cursory descriptions found in their filings:


RTA-408 in Friedreich’s Ataxia

This is a condition that inflicts progressive nerve damage due to the inheritance of mutations in the frataxin protein. Absence of frataxin can increase reactive oxygen specifies generated by the mitochondria. RTA-408’s putative anti-oxidant method of action is supposedly the draw here. Their filings note that they’ve enrolled 8 patients as of YE2015 in their MOXIe study. No data are provided.

RTA-408 in Mitochondrial Myopathies

These are a series of disorders that compromise mitochondria function and can result in progressive muscle weakness. Their purported method of action is an increase in ATP turnover caused by RTA-408 that will ostensibly cover the deficit from the mitochondria in this condition. Their filings note they’ve enrolled 8 patients as of YE2015 in their MOTOR trial. No efficacy data are provided, but notably one SAE has been observed. As before, the explanations in their filings aren’t fully convincing:

An SAE of tachycardia was submitted by a clinical site and further reviewed by the chair of the DSMB, who is a cardiologist. The DSMB chair concluded that the event should not be filed as an SAE as it was likely a previously undiagnosed and benign tachycardia, which is common in this patient population.

RTA-408 in Melanoma

The company is currently running a phase 1b trial of RTA-408 in combination with nivolumab in melanoma called DISCOVER. There are no data available from this trial. It is a follow-up to a phase I trial of monotherapy RTA-408 in melanoma. Given the high competition in this area, this program is unlikely to be meaningful.

RTA-408 in Cataract Surgery

This drug was also tried in the GUARD trial, testing a topical formulation of omaveloxone in 307 patients undergoing cataract surgery. There was no efficacy demonstrated.


2.3. Phase III Trial


RTA-402 - CKD / Type 2 Diabetes

Their only phase III trial to data has been for the use of bardoxolone to delay end-stage renal disease or death from cardiovascular disease in stage 4 CKD and type 2 diabetes. This trial by de Zeeuw randomized 2185 people to placebo or 20 mg of bardoxolone. The impetus of this trial was the observation that bardoxolone increased eGFR, and this would presumably benefit patients with CKD. They did a series of phase 2 trials and eventually ran a phase 3 trial, Beacon. The trial was terminated early due to an increase of cardiovascular events in the bardoxolone arm. 

The company states the increase in heart failure events was due to the susceptibility of patients to retain fluid. In effect, the drug was observed to increase sodium retention, increased blood pressure, decreased body weight and also increased estimated GFR as expected. The company completed retrospective studies of the data and concluded patients with a prior history of heart failure and baseline BNP > 200 pg/mL were the “at-risk” group for fluid retention leading to SAEs. In effect, they claim that the drug can be safely administered if such patients are excluded from future trials. There is currently a trial in Japan by collaborator Kyowa Hakko Kirin that may answer this question, but no data are available.


3. Conclusion

This looks like a classic case of a biotech that knows it has difficult drug candidates. Their lead drug, bardoxolone, has significant safety concerns. In the Beacon trial, the company notes they’ve identified the patient population at-risk for bardoxolone’s AEs. However, this explanation is tenuous on multiple counts, and presents additional program risk:


1) They’ve retrospectively determined this “at-risk” population, and it may be related but not capture the root cause of the bardoxolone-induced increase in heart failure. In effect, there is no guarantee that they’ve completely removed future AE risk from their trials. 

2) Their thesis that bardoxolone increases fluid retention is valid, but puts the pulmonary hypertension program in jeopardy. Pulmonary hypertension patients are, almost by definition, susceptible to fluid retention. Therefore, any subsequent program in PH that enrolls large numbers of patients will be heavily at risk of showing increased SAEs.

3) The data analyses from their phase 2 trials in PH are evidently cherry-picked. Omission of patients without explanation, retrospective omission of patients at risk of moderate / severe anemia, lack of disclosure of full data sets, and almost uniform lack of statistical significance despite the cherry-picked datasets suggests there is little efficacy to be seen. In sum, as of YE2015, their Lariat phase 2 trial has enrolled 85 total patients across cohorts, but the company is most recently making claims on a 19 patient subgroup (14 drug, 5 placebo) of this population. Nobody hides efficacy.

4) A role of anti-oxidants in lung diseases is not clear. In the PANTHER trial, the anti-oxidant n-acetylcysteine had no physiological or clinical benefit in combination therapy for idiopathic pulmonary fibrosis. This population is being tested in Cohort 4 of the Lariat trial, and suggests the anti-oxidant method of action of bardoxolone is unlikely to provide meaningful benefit.

5) There is no clear path to approval. Although the company has started their phase 3 “Catalyst” trial in pulmonary hypertension, their SEC filings are notably shifty about what this trial will allow. The company explicitly notes that:

“During our interaction with the FDA in October 2015, the FDA noted that CATALYST, together with the Phase 2 data from our LARIAT trial in PAH patients and prior clinical trials with bardoxolone methyl, would provide adequate data for a New Drug Application, or NDA, review of the safety profile of bardoxolone methyl.”

This language is present in multiple SEC filings, and is clear that these data will not be sufficient to demonstrate efficacy of the drug for NDA purposes. In effect, the path to approval in pulmonary hypertension is unclear.

6) Perhaps the quickest path to approval for bardoxolone is in Alpert syndrome. But this has the concern of once again testing the drug in a patient population that is notably at risk for end-stage renal disease, overlapping the population that prematurely terminated their large phase 3 trial, Beacon. Although the drug does increase eGFR, it is impossible to ignore the AE profile and there is no clear indication that their retrospective identification of patients at risk for volume retention is, in fact, valid prospectively.


As for the follow-on drug omaveloxone / RTA-408, its near-identical method of action suggests that it won’t salvage the pipeline. Further, although the company is more than happy to provide small n data for bardoxolone, they’ve not released any data (other than AEs) for RTA-408. This suggests the drug is having little meaningful effect to date. Further indirect proof of this is from the September, 2016 disclosure that AbbVie has opted out of co-developing omaveloxone. Although AbbVie reserves the right to opt back in at a future date, leaving the drug behind at phase I suggests there is little efficacy to be observed.


Finally, these derivatives have had a long list of adverse events that have caused a drug to be abandoned (RTA-401) and a large phase 3 trial to stop prematurely (RTA-402). It seems clear the class has safety issues.

Proteon's PRT-201 (Vonapanitase)

This is an analysis of Proteon's PRT-201 (vonapanitase) that was previously completed.



This is a 26 kDa preparation of elastase enzyme for use at the time of arteriovenous graft (AVG) or fistula (AVF) creation, with the company concentrating initially on AVF. For patients undergoing hemodialysis, AVF, AVG or insertion of a temporary dialysis catheter via the internal jugular / subclavian vein / femoral vein are the choices for allowing vascular access. Of these, AVFs are the generally preferred method.

One of the major complications for AVFs is loss of patency, general short-hand for the reduction of blood flow via the access site. Loss of patency can be due to multiple factors, but most relevant to this drug candidate is adverse remodeling of the vessels involved in the AVF creation.

This figure demonstrates the general pathways of AVF dysfunction as they relate to PRT-201. Following formation of the AVF, vessel diametre of the outflow vein is of primary importance, but can be encroached by adverse remodeling. In essence, remodeling that reduces the diametre of the vein via neointima formation or negative remodeling all challenge AVF function and are causes for primary loss of patency. Proteon believes dropwise administration of PRT-201 to the outside of the vessel for 10 minutes during AVF formation will improve the function of the fistula and reduce primary patency loss.



Throughout multiple papers, Proteon believes PRT-201 will preserve AVF function by various mechanisms:

1) Degradation of elastase along the adventitia layer will increase compliance (1/stiffness) of the vessel, allowing greater lumen diametre immediately at the time of AVF creation, thereby allowing the vessel to achieve higher blood flows (versus placebo). These higher blood flows would presumably require more extensive neointima formation prior to occlusion, thereby allowing primary patency to be maintained for a longer duration.

Best surrogate: increased vein lumen diametre immediately after administration of PRT-201


2) Proteolysis of elastin fibres within the adventitia by elastase will create small elastin fragments that serve as known chemoattractants for smooth muscle cells and monocytes. As the elastin fragments are generated in the outside layers of the adventitia, they will serve as decoys prompting smooth muscle cells to move to the adventitial layer to promote positive outward remodeling rather than towards the lumen to increase neointima formation.

Best surrogate: this mechanism reflects on the primary patency endpoint, but vein lumen diametre at 6 weeks and 3 months or time to hemodynamically significant lumen stenosis (HSS) would be representative.


In effect, these two mechanisms should lead to increased vessel dilation at time of surgery, increased blood flow and decreased HSS. Achievement of all three would rationally allow the treatment to have a chance at increasing primary patency duration, which is the endpoint of the current PATENCY-1 trial.



Proteon has completed a phase I trial (Peden et al, 2013) and a randomized phase II trial (Hye et al, 2014) in AVF. Although the primary endpoint of primary patency was not met in the phase II trial, a subgroup was identified wherein patients receiving radiocephalic fistula (RCF) benefitted more from PRT-201 than patients receiving brachiocephalic fistula (BCF). The follow on phase III PATENCY-1 trial is therefore a double-blind, placebo controlled trial examining the effect of 30 ug PRT-201 on the primary patency of RCF AVF.


This is a classic example of a biotech running a phase III program from a post-hoc subgroup obtained from their phase II trial. There are multiple reasons to be skeptical about the outcome of the PATENCY-1 trial. To that end, we can focus on the proposed mechanism of action and determine if there is indication of PRT-201 activity in the phase I and II trials completed for this patient population.


Phase I Trial

Red Flags

1. In the phase I trial, Peden et al showed an inverse dose-response relationship wherein the primary patency trend favoured the low dose PRT-201 arm (3.3, 10 and 33 ug) whereas the medium (100, 330 and 1000 ug) and high (3, 6, 9 mg) dose arms regressed and where indistinguishable from placebo. 

There is no biological reason to favour low dose over high dose, as the high dose has up to 270x more active enzyme. That should translate to significant vessel dilation at the time of surgery and significant increase in blood flow. However, the paper expressly notes that *none* of the participants showed an increase in outflow diametre of 25%. Further, the percent change in vein diametre pre- and post-surgery also showed an inverse dose-response, with high dose (3.2%) being lower than low (6.1%) and medium dose (5.7%). This significantly raises questions as to whether the 10 minute treatment is doing anything at all.


2. Although the vein diametre increase in the low and medium dose groups where higher than that for placebo (2.1%), this finding likely reflects noise. From personal experience, digestion of veins and arteries by enzyme preparations is strictly dose and time related. For example, when vessels are taken into culture and enzymatically digested to release the smooth muscle cell population, the procedure is necessarily time and activity related. Digestion of increasing amounts of elastin would simply compromise the structural integrity of the adventitial layer, increasing external load on the smooth muscle cells, which would in turn necessarily increase vessel diametre. That this does not happen in the measurements pre- and post-surgery suggests the mechanism is not progressing as intended.


3. Further, when blood flow was measured immediately pre- and post-treatment, the percent change in blood flow was relatively superior for the placebo group (41%) versus the medium (17%) and high doses (19%), and very similar to the low dose group (43%). This also suggests, along with the vein diametre data, that the enzymatic digestions at the adventitial layer is not having the intended effect.


4. Finally, the investigators report the time to HSS, which was 59 days in the placebo group versus 209, 49 and 43 days in the low, medium and high dose groups, respectively. Again, this suggests an anomalous outcome. Although there is likely to be an optimal treatment time to produce chemoattractant elastin fragments, that would predict an optimal dose that would produce peak results followed by the effect waning off as dose was increased. In this respect, the medium dose group would have been expected to be in between the results of the low dose and the high dose group.


5. Furthermore, since each dose group has 3 dose levels, there is no indication which of the 3.3, 10 or 30 ug doses in the low dose group most contributed to the increase in HSS. If there was a clear indication for 30 ug (the Phase III dose), it would have been noted as it would strengthen the method of action and rationale for the phase II.


In sum, this results from the phase I trial support safety of PRT-201 but provide no meaningful efficacy signal.


Phase II Trial

Red Flags

1. This randomized, placebo controlled trial failed to meet the primary endpoint of primary patency for either the 10 or 30 ug dose group versus placebo. Moreover, the endpoint was missed despite the investigators using a per-protocol (PP) analysis rather than an intent-to-treat (ITT) analysis. Of the 169 participants randomized, data from 151 were included in the analysis. Analyses that are PP favour the sponsor, and the trial’s inability to hit on this lessen hope for a subsequent phase III trial wherein ITT analyses will be necessary.


2. The investigators begin data-mining in order to find a benefit, and determine that patients receiving RCF AVF benefit from treatment whereas BCF AVFs do no benefit at all. This is a dubious finding, mostly because PRT-201 is a pure enzyme preparation and has little reason to act differentially in the adventitia of veins involved in the BCF procedure versus the RCF procedure. 

Note from Phase I: When the investigators examined the data in the phase I trial via a Cox proportional hazard model, only treatment with low dose, white race and age <65 years were associated with decreased risk of patency loss. If the RCF finding in the phase II was real and robust, it would have also appeared in the exploratory analyses of the phase I trial. It did not.


3. The RCF subgroup analysis is very likely unbalanced between arms. For the trial as a whole, the investigators note that multiple prognostic indicators including exposed vein length, shorter arteriotomy length, white race, predialysis status were all associated with decreased primary patency loss. There is no indication that the investigators examined the placebo, 10 ug and 30 ug groups from the RCF subgroups to determine if they were balanced in this regard. But given the low numbers (n < 25 in each group), it is very unlikely that the groups were balanced.


4. For the trial as a whole, HSS at either 6 weeks or 3 months was not significantly impacted. At 6 weeks, HSS was 51, 30 and 39% for placebo, low and high dose, and this changed to 40, 41 and 35% at 3 months. Only the comparison of HSS for placebo versus 10 ug at 6 weeks was significant; this was not maintained at 3 months, likely suggesting it was a chance finding. This is especially likely given that HSS in the placebo group appeared to improve over time; this may reflect a change in the number of evaluable participants, but nonetheless supports a lack of differentiation between placebo and treatment groups.

Interestingly, the HSS data are *not* broken down further into RCF versus BCF. A reliable rule for interpreting peer reviewed papers (and biotech company PR) is to assume that analyses not presented were not statistically significant. In this regard, given the authors’ proposal that the RCF versus BCF analysis is meaningful for primary patency, it would have been consistent if the HSS surrogate was significant in RCF versus the BCF group.


5. Although vein lumen diametre and flow rate data are provided by the investigators, none of the comparisons across groups or over time points are noted to be statistically significant, nor do they appear to adhere to a meaningfully different trend in the treatment versus placebo groups. Once again, this works against one of their proposed mechanisms of action, and provides little evidence that PRT-201 has short term or long term impact on remodeling of the outflow veins or flow characteristics. 



This appears to be a classic case of subgroup data-mining in phase II leading to a phase III program. There is little to no internal consistency between the phase I and II data. Further, there is no mechanistic consistency between their proposed biological mechanism of action and the surrogate measurements that would reflect the action of the drug. Vein diametre, blood flow and HSS data are not different between placebo and PRT-201 dose groups, providing little confidence that primary patency would be maintained for longer durations. Further, the separation of RCF versus BCF subgroups has little biological basis with respect to PRT-201’s mechanism of action, and the purported efficacy in the RCF subgroup is likely due to imbalances in patient population rather than a mechanism driven discovery.


Miscellaneous Technical Aspects

- Preparation of the enzyme is relatively trivial and will not be a limiting factor. Enzymes are supplied as powders in a vial and need to be reconstituted by the team at the time of surgery. Technically this poses a risk for some loss of enzyme activity, as enzymes tend to lose activity over time if not properly maintained.

- Administration of protein always raises the risk of antibody generation, or pre-existing antibody titre. However, there is no retreatment in this protocol, and evidence of very low / non-existant antibody titre is provided.

How to Avoid Heartbreak

Celyad’s Chart-1 trial has failed.

The failure is unambiguous*.


With that out of the way, we can dissect some of what went wrong and how investors can avoid future bouts of Cardio3/Celyad-induced portfolio decompensation.


In the heart failure area, mouse models are not serious efforts at validating the platform. End of story. Companies such as Cardio3 / Celyad, and their academic collaborators, move therapies into the clinic based on efficacy in mouse models without repeating the observations in large animal models. Mouse preclinical models of heart failure designed to rationalize a treatment’s transition to the clinic are simply not useful, not reliable, and not reflective of human pathophysiology. They are necessary as a precursor to testing in a large animal model where the platform’s efficacy should be reproduced, albeit at smaller effect sizes. However, they are not sufficient.

And upon reflection, this point has a very obvious but important element: the company and its collaborators know this as well!

I can assure you the in-house scientists and academic collaborators at companies such as Cardio3 / Celyad know full well efficacy in mouse models is not a serious hurdle for the development of heart failure therapies. They know full well that multiple cell therapy efforts with stunning results in mouse models moved into the clinic and walked away with an HR of ~1. They know full well the importance of the therapy showing reproducible efficacy in large animal models.

Investors can debate the reasons and motivations behind companies and collaborators running through preclinical studies at such a record pace. However, many of the descriptors are unflattering.


Track record matters. Not a novel insight by any means. However, companies based primarily on the work of academic collaborators can offer investors an additional edge in appraising the prospects. Over the last 12-18 months, both Celladon and Cardio3/Celyad have had significant failures in the clinic. Each company relied heavily on work derived from academic collaborators. It is wholly appropriate for investors to examine future start-up efforts from those academics with skepticism and take track records into consideration when doing their due diligence.


Never trust a phase 2 trial that succeeded based on subgroup analyses. This, also, is not novel and may have exceptions. However, in the heart failure field, it is almost axiomatic. Cardio3/Celyad’s phase 2 trial leading up to Chart-1 did not focus on the intent-to-treat (ITT) analysis but kicked out some participants for whom the cell therapy could not be procured / manufactured. That was always a red flag. But it is compounded in heart failure where participants appear to be a much more heterogenous group at all stages. Small (n < 100) trials in heart failure tend to pop up positive even with proper control arms. Perhaps it is because we don’t understand the disease all that well, and we’re still learning how to classify stage / progression. Regardless, small, singular phase 2 clinical trials in this indication are simply poor predictors of success, especially when subgroups begin to pop up.


Never trust companies where the CEO corresponds with investors, directly or indirectly, via a message board. The internet is a big and wonderful place. Much is written, much is fabricated, much is conspiratorial. But if you notice the CEO of your company taking time out to gossip about the merits of a blog post, then s/he is not focused on the right questions. This can sound contradictory, because having the CEO or CMO of a company engage you directly can sometimes feel empowering and reassuring. But it’s a bad sign.


And lastly: Stay away from biotechs who use the word “cure” in their slogans or their treatment brand names. This one is self-explanatory.


And that’s that.



* One way to tell if a trial has failed is the Nouveau Analysis Rule: Did you hear about a specific analysis for the first time on the day of the data release? If so, the trial has failed.

Is Eteplirsen an Exceptional Case for the FDA?

Prospective Duchenne muscular dystrophy (DMD) treatment eteplirsen faced an uphill battle at the April 25th FDA advisory panel (replay here). Earlier release of the briefing documents had underlined the FDA’s persistent concerns about the sponsor Sarepta’s data package, leaving most observers skeptical about the panel’s chance at a positive outcome. The advisory panel itself was a long affair marked by an open public hearing full of resolute and emotional statements made by young boys with DMD, parents of boys with DMD, medical doctors and researchers. 

At the end of the ~11 hour panel, those previously unconvinced of the drug’s efficacy were likely in agreement with the negative tallies that accrued for the voting questions. Nonetheless, those hopeful for the drug’s prospects perceived two significant lifelines cast out by medical doctors as well as by prominent FDA staff members.


The Doctors

Medical doctors who had treated boys with eteplirsen voiced a unanimous message: boys with DMD who had received eteplirsen performed unequivocally better than other boys in their practice history. Such comments understandably bolstered the confidence of advocates and families pushing for approval. Although encouraging, the anecdotal comments by the doctors are unlikely to trump the accrued clinical data, for a few reasons:

  1. Doctors should indeed be advocates for their individual patients and no one should want a doctor who isn't. In that regard, it is not unreasonable or out of place to see individual doctors lobby on behalf their patients, even if the drug in question provides an ambiguous outcome.
  2. Clinical trial history has shown anecdotal individual experience, either by a patient or a treating physician, to be no substitute for larger controlled trials. In this case, the statements by the doctors left no room to suggest any overlap between the performance of boys on eteplirsen versus previous boys in their practice. Unfortunately, the clinical data are not indicative of such a uniformly exceptional experience.

Shown below are figures from Sarepta’s briefing documents (Fig. 19) and the FDA’s (Fig. 2).

Perhaps some advocates can reconcile the uniformly exceptional performance cited by the testifying clinicians against the clear overlap of the eteplirsen and control data sets. However, to neutral observers, such graphs are unlikely to mesh with their mental image of an unequivocal effect. It’s understandable that the doctors who participated in the open hearing are advocating for their patients, but this advocacy has little scientific backing.



Ultimately, the decision of approval lies with the FDA. Adam Feuerstein’s recent article builds on this theme and assembles clues to buttress Sarepta’s case. Initially, he notes:

“If Woodcock [FDA Director, Center for Drug Evaluation and Research] agreed with all the bad things said about eteplirsen on Monday night, if she had no plans to overrule or cast aside the recommendations against approval, why did she go out of her way to meet with DMD boys and their families immediately after the meeting ended?
Could the answer be that Woodcock plans to approve eteplirsen?”

An observer could certainly sense positive vibes from this type of post-panel interaction, if they were so inclined. However, this meeting could just as easily be described as normal social behaviour. In fact, Monday’s panel was not the first time that Dr. Woodcock had met these families. In February of 2013, Drs. Woodcock and Robert Temple of the FDA met with the Jett Foundation, a group working to improve the lives of those with DMD. After that meeting, the group wrote:

“Our conversation with the FDA leaves us positive and optimistic. These high ranking FDA officials were extremely supportive, engaging, and receptive to our ideas about accelerated approval. It was our collective impression that the FDA seemed committed to working as expeditiously as possible to grant access to life saving treatments to those who will benefit. We were able to start the dialogue that will affect how Duchenne is viewed by the regulatory agency. We appreciate that the FDA is allowing patients, parents, and advocates to play an integral role in the regulatory process. We will continue our work with the FDA as more drugs begin to reach trial level. Our goal is that all Duchenne patients get treatment as quickly as possible.”

In that regard, Dr. Woodcock is simply being a normal, decent human being. She has met these families before and no doubt cares for their compelling cause. There is little reason to believe she would have ignored these parents even if the agency was on track to reject the application for marketing. To further buttress this interpretation, we can go back to 2013 once again. Despite the positive interaction the Jett Foundation had with Drs. Woodcock and Temple in February, November 2013 saw the FDA push back a potential filingby eteplirsen’s sponsor.


Later on during the panel, a theme emerged regarding the interpretation of the presented data and testimony: are we to believe the graphs and the underlying statistics, or should we give strong weight to the doctor and DMD community testimonies?

Dr. Woodcock explains:

“The standard is adequate and well controlled trials, OK, that's what is in this statute, but we are instructed to have flexibility on how we interpret that based on medical need.”
“We are instructed, as people said, to take the views of the patient community into account, more on the benefit and the risk. So, the statutory standard is more or less as described there but there is flexibility and that's where we should take the views of the community into account.”

As Feuerstein notes, the FDA’s Ellis Unger added this clarification:

“I think with the majority of the [eteplirsen] patients here, we have an incredible advantage in my time at the FDA that is unprecedented. To have all the patients here so that is an important advantage we have. One thing you can try to do is reconcile what you heard from the patients with the data that you've seen presented by the company. We're hearing patients are improving, doing things next year that they didn't do last year, and you have to figure out if you can reconcile that with the actual hard data you've been analyzing today.”

The suggestion is the above two comments tip the FDA’s position towards listening to the patients. Perhaps. But Dr. Unger is very clear the testimony and the actual data are at conflict, and his comments could simply be interpreted as a neutral acceptance the conflict is there and requires resolving by panel members. Woodcock’s comment regarding flexibility is true, but could also neutrally be viewed as a simple recitation of the standard under which the FDA operates.

Subsequently, Feuerstein outlines the following exchange between panel member Dr. Chiadi Onyike and the FDA’s Robert Temple:

Dr. Onyike: The question twice mentions 'well controlled,' and as you've heard repeatedly, some people have said they have trouble with the controls. So, this 'well controlled' phrase in a sense tips or constrains the question.
Dr. Temple: I understand a lot of people don't like historically controlled trials. They're not sure they believe they're well controlled. Our regulations since 1970 have said that a historical controlled trial can be adequately and well controlled study. The question here goes: Under the circumstances do you think it was? Do you think the way they selected patients, the way they analyzed them, was good enough to make it an adequate and well-controlled study? That's the question. Historically, historical-controlled trials have been the basis for approval -- sometimes in sort of obvious cases and sometimes in cases that quite aren't so obvious.

This is interpreted as the FDA “saying that comparing data collected from a single arm study against a historical control -- like what Sarepta did -- is definitely okay.” Absolutely. But this is again a recitation of FDA guidelines. Dr. Temple provides no hint the FDA believes Sarepta’s historically-controlled study is adequate; rather, he simply states the FDA would accept one, provided the sponsor conducted an adequate one, and that panel members should not see the use of a historical control arm to be disqualifying.

In sum, although many of the actions and statements at the panel can be weaved into a thesis for approval of eteplirsen, their simplest explanation supports a view of the FDA as fastidious and thorough in its duties. They asked panel members to reconcile data and testimony, clarified the prevailing statutes, and provided a forum for the DMD community to express their views.


Perhaps the Sarepta story is best summed up by Mr. Benjamin Dupree, a 23 year old with DMD who served as a patient representative on Monday’s panel. Mr. Dupree voted Yes in Vote 7, wherein panelists were asked:

“Do the clinical results of the single historically-controlled study (201/202) provide substantial evidence (ie, evidence from adequate and well-controlled studies or evidence from a single highly persuasive adequate and well-controlled study that is accompanied by independent findings that substantiate efficacy) that eteplirsen is effective for the treatment of DMD?”

When given a chance to comment on his vote, Mr. Dupree was overcome with emotion while saying the following:

“I voted yes. I can’t really reconcile the difference between the testimony that was given suggesting that boys are recovering abilities. I don’t… living with Duchenne I don’t understand how that’s even possible but at the same time the study doesn’t prove from a scientific… it doesn’t provide what i think is adequate evidence to support all this testimony that I am seeing in here.”

It’s not useful or fair to question how Mr. Dupree reconciled his Yes vote with his admitted difficulty in seeing adequate scientific evidence. That he was so emotionally overcome hints, at the least, to a part of Mr. Dupree that wishes the scientific data behind eteplirsen were as strong as the efforts of the DMD community.


Prospectively Defined.

In the next few weeks Celyad (née Cardio3) will announce their results from Chart 1, the first of two phase 3 trials examining their cell therapy for the heart. Previous entries on this site have expressed skepticism about the chances of success for these trials, and that stance remains unchanged. However with data release imminent, intellectual honesty requires that we set some prospective expectations for the dataset. Such expectations are particularly important for biotech investors, because companies can be very industrious and creative during data analyses.

To their credit, Celyad have published a paper prospectively defining their primary endpoint. They define the primary and secondary objectives:


The primary objective of CHART-1 ... is to evaluate the efficacy of cardiopoietic stem cells ... delivered using an endoventricular injection catheter (C-Cathez®) in comparison with a sham procedure on a hierarchical outcome comprising measures of mortality, morbidity, and changes in quality of life, 6MWT distance, functional capacity, and LV structure and function at 39 weeks (9 months) post-procedure.
The secondary objective is to assess safety by comparing the incidence of serious adverse events between study groups at 52 weeks (12 months) and all-cause mortality at 104 weeks (24 months) post-procedure.
Follow-up will occur at 4, 13, 26, 39, 52, and 104 weeks post-procedure. The primary efficacy endpoint is a Finkelstein – Schoenfeld hierarchical composite endpoint compris- ing all-cause mortality, worsening HF (WHF) events, and changes in Minnesota Living with Heart Failure Questionnaire (MLHFQ) score, 6MWT distance, LVESV, and LVEF by transoesophageal echocardiography (TTE) at 39 weeks.


Arguably, the secondary objective of the trial will not be readily assessable at the time of the upcoming data analysis. The trial enrollment completed in Q3 of 2015, and with the upcoming primary objective analysis due 9 months from then (~Q2 of 2016), safety at 12 and 24 months will not be available. That leaves us with the responsibility of setting prospective expectations for the primary objective. To provide some additional clarity on what exactly the primary endpoint is, we can add the description from the NIH to our notebook:


“Efficacy between groups post-index procedure [ Time Frame: 39 weeks post-index ]
Change between groups from baseline and 39 weeks in a hierarchical composite outcome comprising, from most to least severe outcome, days to death from any cause, number of worsening of heart failure events, change in score for the Minnesota Living with Heart Failure Questionnaire (MLHFQ) (10-point deterioration, no meaningful change,10-point improvement), change in six-minute walk distance (40-m deterioration, no meaningful change, 40-m improvement) and change in left ventricular end systolic volume (15-mL deterioration, no meaningful change, 15-mL improvement), and left ventricular ejection fraction (4% absolute deterioration, no meaningful change, 4% absolute improvement).”


This is corroborated by the description in the published paper, wherein the parameters included in the hierarchical analysis are narrowed down more clearly:

  1. Mortality = days alive out of 39 weeks
  2. Number of WHF events: 0, 1, or ≥ 2
  3. MLHFQ ≥10 point improvement, no meaningful change, ≥10 point deterioration
  4. 6MWT ≥40 m improvement, no meaningful change
  5. LVESV ≥15 mL improvement, no meaningful change
  6. LVEF ≥4% absolute improvement, no meaningful change, ≥4% absolute deterioration

Based on a method outlined by Finkelstein and Schoenfeld here, the company aims to analyze the data in a hierarchical manner that gives weight to item 1, which is the most clinically relevant hard endpoint, but also takes input from the softer, longitudinal measures. The Chart-1 trial planners themselves note that these endpoints are not robust:


The relevance of the specified changes in MLHFQ, 6MWT, LVESV, and LVEF merits discussion. Some parameters, including the 6 MWT and MLHFQ, are subjective and may be particularly influenced by knowledge of the treatment assignment. The absolute change in MLHFQ score that would indicate a clinical meaningful outcome is not certain. The selected 10-point change excludes chance variability, and was associated with substantially increased risks of death and re-hospitalization in patients with advanced chronic HF followed for 18 months on average. Regarding the 6MWT, a 43 m improvement was found to be statistically significant in the COMPANION study and accompanied by reduced risks of death or HF rehospitalization at 6 months. We have therefore considered a 40 m change as meaningful. Lastly, LVESV and LVEF are more objective measures of response, which, when considered in conjunction with more subjective parameters, provide a clinically valuable readout of regenerative impact. However, when taken alone, in a patient population with baseline NYHA class II/III symptoms and LVESV of 200 mL, it has been suggested that a change of 10 mL is clinically meaningful, informing the CHART-1 design.


Whether this hierarchical measurement increases or decreases the chances of a statistically significant outcome is up to the reader. Nonetheless, Finkelstein and Schoenfeld suggest caution when longitudinal measures are included:


If the true treatment effect is to improve a longitudinal measure but make survival worse, then a combined efficacy measure may lead to the conclusion that the treatment with worse survival is superior. The fact that mortality is included in the efficacy measure makes this less likely than an analysis that ignores mortality, but it can still happen, especially if the effect on the longitudinal measure is strong and occurs early in the trial.


All told, my expectations for a successful Chart-1 trial would have to meet these two criteria:

  1. A statistically significant improvement on the composite endpoint when analyzed on an intent-to-treat (ITT) basis and
  2. An accompanying statistically significant reduction in mortality on an ITT basis

The reason that I would like to see both the overall composite *and* the mortality endpoints improve is to ensure that the soft endpoints (#3-6 above) aren’t driving the outcome. Additionally, the analysis must be completed on an ITT basis.


The most important red flag to look for in the data release will be any mention of an analysis based on a per-protocol population rather than an ITT population. Per protocol analyses tend to omit patients who were randomized but unable to undergo the procedure. In this case, it would primarily be patients from whom the cell preparation could not be reliably extracted and/or propagated for reinjection. There is a long history demonstrating the importance of ITT analyses over PP analyses, and this blog won’t try to re-enact those lessons.

An additional red flag will be singular mentions of statistical significance for parameters 2 through 6 listed above, with no mention of (or a trend in) parameter 1. Further still, if the primary endpoint is not explicitly disclosed and there is mention of statistical significance in a subgroup of the Chart-1 trial, that is a red flag and should be accepted as an implicit sign that the primary endpoint has failed.

To readers, this may seem an unnecessarily skeptical view of a data release. As the company has prospectively published their analysis plan for Chart-1, it would seem data release will be a simple matter of completing 9 months of follow up on the last enrollee and publishing the results. However, the company recently noted that a former board of directors member, Prof. William Wijns, will oversee data analysis and dissemination . The company notes that Dr. Wijns is a co-founder of the company.

One has to ask why a prospectively defined statistical analysis plan needs a co-founder and ex-BOD member to be appointed to oversee its execution and dissemination? In any case, being armed with clear prospective expectations of what is and isn’t a success will allow an objective assessment of the technology.

Abbvie to Galapagos: "It's not you, it's me"

Abbvie has decided to pass up the chance to co-develop filgotinib, preferring to advance their own internal candidate ABT-494 into phase 3 trials in rheumatoid arthritis. Admittedly, this came as an initial surprise given the strong efficacy results posted by filgotinib in the DARWIN1 and 2 trials. However, Abbvie has provided 12 week data for ABT-494, allowing a quick comparison of the 12 week data across oral JAK inhibitors in rheumatoid arthritis:

ACR70 Data Across Trials (12 week time point)

Tofacinitib data from 2x5 mg arm of Study IV, 2 mg arms of RA-BUILD and RA-BEACON, 2x100 mg arm of DARWIN1, 100 mg arm of DARWIN2 and 6 mg twice-daily arms of Balances -I and II. Data are a mix of LOCF and NRI reporting.

Given these data, Abbvie’s decision is not surprising. The spread in efficacy across the various JAK inhibitors suggests ABT-494 will be very competitive, and has the added benefit of being internally developed and wholly owned. Abbie also reports that they will be moving ABT-494 into phase 3 by the end of the year, serving their intent to rapidly capitalize on the phase 2 Balance data. This left Galapagos to announce that they’re already courting offers, but that phase 3 studies will begin no sooner than early 2016. Of these three investigational JAK inhibitors, this leaves baricitinib closest to an NDA, followed by ABT-494 and then filgotinib.

Since filgotinib’s future is the most speculative of the three, some general points bear consideration in predicting the type of deal Galapagos can expect to garner in a subsequent partnership. A data package including DARWIN1 and 2 would often be considered a de-risking proposition by potential suitors, allowing Galapagos to command more favourable deal terms. However, the following points may suggest that a future deal for filgotinib may not be as lucrative, or may result in Galapagos assuming more of the financial and regulatory risk than previously:

- Filgotinib will potentially be the 4th JAK inhibitor on the market after tofacitinib, baricitinib and ABT-494. With Abbvie rapidly moving into phase 3 trials, they seem to have done their best to ensure that Galapagos is a step behind. Any potential suitor will have to consider this as a major negative to the economics of the collaboration.

- Galapagos has claimed filgotinib is the most JAK1 selective (over JAK2) of the investigational inhibitors, potentially conferring a safety advantage over predecessors like tofacitinib and baricitinib. This was often underlined by a focus on the anemia side effect profile, a talking point that I suggested was largely overblown. Galapagos claims that filgotinib is 3x more JAK1 selective than ABT-494 and by their own internal data, they claim filgotinib is also 3x more JAK1 selective than tofacitinib. This would place ABT-494 around the same JAK1 selectivity range as tofacitinib. In effect, Abbvie does not appear to be daunted by the proposition that JAK1 selectivity is a de facto advantage. In that regard, this selling point may not resonate as well with potential suitors.

- The FDA and Galapagos agreed to exclude men from the 200 mg daily dose group, centered around a potential concern for side effects to the male fertility system. Even though the daily 100 mg dose groups look strong from an efficacy perspective, the potential suitor may deem that this issue merits consideration and, subsequently, a commitment of time and resources to address.

In sum, I remain of the opinion that filgotinib is an efficacious drug for rheumatoid arthritis and Galapagos will be able to find another collaborator. However, I suspect the new deal will not be as favourable as the previous.



Postscript / Erratum:

A previous version of this entry suggested Eli Lilly and Abbvie had comparable legacies in RA disease modifying agents. That can safely be considered as incorrect and has been edited. Hat tip.

Current positions - 2015/08/19

For the sake of transparency, biotech / pharma stocks I currently hold include (in alphabetical order):


And in the paraphrased words of Morrissey, some positions are bigger than others. Importantly, everyone has a bias and it's worth keeping this in mind when assessing the validity of entries in this blog.

August 19, 2015 Update - The position in SGYP has been liquidated, replaced in very small part by a speculative position in ENUM. Why? Because I like 84,672 well plates.