Bionomics (ASX : BNO) is a company I have watched for a long time as it has gone through many stages of re-invention. It is also a company that I have struggled to motivate any real enthusiasm for, either to praise or criticise, mainly because the company is just so mediocre. Let’s face it the world mostly consists of mediocre companies and therefore this is insufficient reason alone to make a comment. As such, this post has existed in one form or another since February and finally after many requests for commentary (including a good amount of goading and ridicule), I have elected to bite the bullet and capitulate.
In order to keep this short and sweet, I will start by briefly enumerating the “warning signs” that this company exhibits that are seldom discussed by analysts (obviously, because we only have sell-side analysts). I will then write a brief commentary on three of BNO’s product pipeline that I think captures most of the current “value” but also piques most of my skepticism about this company.
Here are the warning signs:
- Basic drug discovery is expensive and under-capitalised drug discovery companies seldom perform well. This company is not a Forma, it has a fairly ad hoc collection of discovery assets. Although for an early-stage ASX-listed biotech it is relatively well-capitalised, in a “hot” discovery area like GCPRs or ion channels (going through a bit of a renaissance at the moment) I would argue that this company lacks the resources to differentiate itself from its much more grown-up competition.
- By extension, valuation doesn’t come from platform technologies, it comes from the clinical development of good medicines. It is my personal opinion that BNO’s articulation of its drug discovery platform(s) is not a differentiation, and mostly just a way that the company can continue to capture feeble investor mindshare with “new toys” popping out on a periodic basis, rather than strong clinical results for its established pipeline (which it does not have, as I will discuss). A pipeline, incidentally, that greatly exceeds its financial resources.
- Small cap companies like this have no business being “vertically integrated”, as is BNO’s claim. It’s just not a meaningful differentiator in an era where almost every facet of drug development can be outsourced, with quality. In BNO’s case, this vertical integration doesn’t seem to have a particularly untoward effect on its burn rate, which probably also infers that its “verticality” is probably not much practical added value, and perhaps mostly just marketing spin.
- BNO has a track record of subsuming mediocre assets through M&A activities. While I generally have respect for management teams that consider growth strategies other than “organic growth” or “financed growth”, as BNO clearly does, an acquisition “process” is not an accretive event in its own right. I just don’t see any sign of quality in the assets that BNO has acquired. What I have observed of BNO’s M&A strategy is not much more sophisticated than “bigger is better”. Just because an asset has some fleas on it, doesn’t mean its bad. But by the same token, just because an asset once loitered in a pharma lab, or a luminary university, doesn’t mean that buying it (for cheap) automatically creates value for the acquirer.
- BNO publishes very little of its clinical data. VERY little. Some of the more recent public disclosures of technology (i.e. ASCO posters) do not constitute peer-reviewed publications, though there is an implication in BNO’s communication strategy that these presentations substantiate the science through “luminary attendance” at conferences. Most of the posters that BNO has produced are more like marketing documents than scientific communications, little more than a placeholder at a conference for doing business development. That’s fine by the way, but I just want to be clear about it.
- BNO’s clinical trial designs are incredibly conservative, verging on scientifically marginal. For example, for its oncology asset currently in Phase II (BNC105) it has never had, as far as I can see, any stand-alone clinical assessment of efficacy other than an initial safety study, which showed no observable anti-tumour effect, and has only ever been meaningfully evaluated in a combination therapy setting. This is notwithstanding that the purported mechanism of action supposedly has potent “dual action” anti-cancer effects in its own rights.
- Related to the above point, BNO is not especially transparent or clear about the mechanisms of action behind its drugs. The company tends to rely on the use of sensationalist concepts like “attacking cancer stem cells” rather than really scientifically justifying the investment thesis behind its pipeline. It needs to do far better if it wants to earn credibility as a diversified biopharma company.
- BNO’s claims of efficacy are mostly unsubstantiated. In some cases, ad hoc retrospective analysis of very small clinical trials has yielded statements that are intriguing from a marketing perspective, but are statistically indefensible. In other instances, product positioning claims have been made in advance of sufficient clinical assessment to be able to substantiate those claims. I will address this generalised criticism in the context of individual development programs (below).
- BNO has a huge intellectual property portfolio that consists of IP that is either very very old, or very very recent. Much of the recent intellectual property seems to reflect sensationalist opportunities to position the company into areas that have a high level of investor and public engagement such as immuno-oncology. There is nothing wrong with this per se, it is BNO’s prerogative to position itself how it sees fit. However, the combination of anti-mitotics with checkpoint inhibitors has been patented to death, and therefore yet another microtubule inhibitor + checkpoint inhibitor patent filing (disclosed as a price-sensitive event no less) is clearly a communication strategy that is not targeted at the sophisticated investor or pharma partner. I also shudder to think at the cost of prosecuting what is a large but probably marginally exciting IP portfolio.
Of course, there are some quite positive things about BNO. They are are, notwithstanding the tendency to “rat pack” mediocre assets for cheap, clearly a transactional company. By any metric, the Merck deal looks pretty darned good for a early-stage collaboration and the company got real up-front money, a true rarity in this space. Of course, the touted numbers had a lot of biodollars tagged on the end, but still – I was impressed. I will not deny it.
However, by the same token there are collaborations that seem to be fairly dormant – like Genmab, where it is hard to see that there a strong fit with Genmab’s current focus. Then there are collaborations like the Ironwood partnership (now terminated, though even the positive spin on termination was impressive), which also had big biodollars involved. Now BNO seems to owe Ironwood a royalty for their “work”, who didn’t actually seem to significantly advance the BNC210 program. If I am mistaken about this, I am happy to be corrected.
But these are all annoying generalisations, now to get specific:
BNC210 – Anxiety/Depression
Anxiolytic/anti-depressant drugs are incredibly hard to develop. I have previously waxed lyrical about the challenges of CNS drug development and will not repeat it here. Plus, with the recent acquisition of Spinifex as a perfect example, there is no reason why a small biopharma company can’t develop an excellent CNS drug if the underlying science is there.
In the case of BNC210, the science simply isn’t there, in my opinion. To date, the company has only published animal data and the quality is not very good. Despite over 160 normal patients being dosed across (apparently) 5 clinical trials, there is no peer-reviewed publication – as far as I can ascertain – of the results. All of the historical trials were mostly small studies, usually single site, and really could only be described as “proof-of-concept”. The currently planned clinical studies (in UK/Europe) have some very sophisticated metrology, such as fMRI, so therefore running a small number of sites is completely understandable. But a male-only population is a bit weird, especially since we know from prior development anxiolytic drugs, that drug metabolism is considerably different between men and women.
When I first started reading about BNC210 I obsessed too much about the biology, I cared too much about the potential of nicotine receptors as a target for depression/anxiety management. But the problem with BNC210 is not the biology, it’s the chemistry. BNC210 is L-Isoleucyl-L-tryptophan, essentially nothing more than a dipeptide of isoleucine and tryptophan, two essential amino acids. Both isoleucine and tryptophan-containing dipeptides are known metabolites of peptide digestion (i.e. “breaking down” of food) and are believed to play all kinds of roles in biochemical signalling (tryptophan-containing dipeptides, for example, may have an important role in regulating obesity and diabetes). The problem with most dipeptides in actual biological systems, is that they are usually a short half-life intermediate metabolites and are very easily cleaved into their fundamental amino acids (i.e. just become cell “food”). In fact, isoleucyl-tryptophan has never been “detected” as an intermediate metabolite in biological systems, which sort of substantiates the idea that this molecule is more likely to become a tasty snack for cells, than hang around and demonstrate a lot of drug potency. As an academic comment, this is probably a good example of where Lipinski rules for drug oral bioavailability probably fall on their face.
As it turns out, the company has known since at least 2008 that the half-life of BNC210 is roughly 6 hours, at least in animal studies. By the way, this particular data point is really …er … tough to find in a peer-reviewed journal. So far I have only been able to find it in a poster given at a conference in Barcelona in 2008, and the “footers” of old press releases (no longer archived by BNO). Usually murine PK is faster than humans, so it may well be that the half-life is somewhat more prolonged in humans. But if it isn’t, then BNO has a major problem because it doesn’t support the proposition of daily oral dosing. By way of comparison, most selective seratonin reuptake inhibitors (SSRIs) have a half-life of between 13 and 45 hours in people (and mice), mostly driven by gender differences in metabolism.
The truth is, all this doesn’t matter. BNO has finished playing with normal volunteers to understand the basic safety of their drug. Unsurprisingly, a dipeptide that is probably naturally – and transiently – produced when you eat a good (or even bad) piece of steak is neither here nor there in terms of safety. BNO is now, approximately 7 years later, doing the proper PK/PD work in humans so eventually I am sure we will see a high-quality peer-reviewed publication of all the great data obtained to date (though someone reading the hype in 2011 would be forgiven for having expected to see it by now). The issue I have is that the management team is already making claims about the drug that they cannot substantiate on the basis of the data they have today. This is a “table” from a recent CEO newsletter:
Firstly, the inclusion of valium (diazepam) in this table is a bit obnoxious because although it is commonly prescribed for anxiety, it is not the “competitive” product for BNC210, SSRIs are (Prozac a.k.a. fluoxetine is only one such example) – despite previous “published” science to the contrary. Valium is just “table filler” to make BNC210 look good. Not only does BNC210 have insufficient PK/PD data from patients to make virtually all of these claims, I would argue that some of the competitive arguments are also misleading. For example, most SSRIs are also fast-acting (what does that actually mean???), it’s just that many patients are put on a dose-escalation schedule in order to appropriately monitor and risk-manage a patient. SSRIs are also incredibly safe and have drug interaction profiles that are not only excellent, but increasingly well understood. In terms of CYP450 interactions, fluoxetine is probably at the more complex end of the scale compared with, for example sertraline, but any drug interaction issues can be managed proactively and this is usually a low prescription risk.
So to repeat, this table is just a huge generalisation and doesn’t accurately reflect SSRIs as a class. As for an issue like drug withdrawal, until BNO does a long-term usage study, the company has no basis to make this claim. Unless, of course, it already knows that the drug has negligible effect. In which case, that may be a very safe marketing claim to make in the absence of clinical data. I have already addressed the the once-a-day oral dosing issue previously, again insufficient data has been ascertained to make this claim. I would argue until their current Phase Ib/II studies are complete, I don’t see how the company can, in good faith, position the drug in this way. It also doesn’t help that sell-side analysts, such as Tenushree Jain at Bell Potter are willing to throw terms like ‘Blockbuster’ out there without an ability to understand the basic science (though I note, by all accounts, she is a very bright and talented individual and no unkindness is intended).
So I am going to go out on a limb here.
Here is my opinion of this drug. I think it is long in the tooth, and any time I see a company take forever to publish clinical data, I am immediately skeptical. BNO has had plenty of opportunity to substantiate its marketing claims. It hasn’t chosen to do so. It makes claims about the drug that are not reflective of any robust public-domain, peer-reviewed evidence. I am sure that the company knows much more that we do, but when it starts to make serious marketing claims for a drug that has yet to complete PK/PD in patients, I get antsy. But my really controversial opinion about this “drug” is that when it undergoes “gastric metabolism” (i.e. you eat it and your stomach does its biz) it is simply going to become a source of L-tryptophan. Don’t get me wrong, tryptophan is critical for serotonin synthesis and therefore has a fundamental role to play in mood, behavior and cognition. But what it means is this : if BNC210 is to be proven to be an efficacious drug, it should be compared with oral tryptophan.
I think it is unlikely to beat it – notwithstanding that 75% of anti-depressant efficacy in drug studies is attributable to placebo effect. Perhaps then, BNC210 is just the “ultimate” placebo – a placebo with a bit of science behind it (does that make it a drug?). Instead of doing studies to look at the impact of food blocking on dose administration (perhaps already a bit of a hint about how Bionomics may really feel about concerns regarding oral availability of this drug), it may need to ask the question whether its drug is any better than tryptophan-rich food.
BNC105 – Oncology (Small Molecule)
I will admit, that my ire rose a couple of weeks ago when I read about BNO’s “combination therapy” patent with checkpoint inhibitors. When that is the most exciting piece of news that a company has, you know that a program isn’t really all that exciting.
Fundamentally, BNC105 isn’t exciting, especially considering where the overall field of cancer is going at the moment. It is yet another microtubule inhibitor or, as the company prefers to call it, a vascular disrupting agent (VDA). Despite claiming monotherapeutic efficacy (a “one-two punch” at endothelial cells and tumor cells, no less), the Phase I clinical (safety) experience with the drug showed no objective anti-tumor efficacy. Now that’s ok to some extent, the study was a very bog-standard dose-escalating safety study, and efficacy isn’t the main goal. But then again, BNO was clearly hoping to see something because they evaluated treatment response with magnetic resonance imaging (MRI), which is great for looking at changes in tumour vasculature. Nothing. Nada. Moreover all subsequent clinical work has been in combination with other drugs, which doesn’t exactly send a highly positive signal that this drug is believed to have efficacy in its own right.
I would have expected to see a stand-alone Phase II efficacy study for a drug that is claiming any kind of effect. The fact that BNO didn’t do this, strongly suggests that it internally lacks confidence that this drug has any monotherapeutic efficacy. I also would have possibly tolerated the shift over to combination therapy studies if the trial outcomes hadn’t been so shamelessly hyped. The Phase I/II study was in a very small number of ovarian cancer patients, in combination with carboplatin and gemcitabine, followed by “maintenance therapy” of BNC105 as if to somehow imply that after whacking the patient with a nasty dose of gemcitabine, the follow-on monotherapy with BNC105 was the magic bullet! In a study design like this, there could be absolutely no claim of efficacy attributed to the drug, because neither the study design nor the patient statistics warrant it. This is awful clinical science.
This is the clinical science equivalent of trialing a gun control case in Texas.
Even worse was the Phase II renal cancer announcement in March of last year. The DisrupTOR-1 trial was intended to compare BNC105 combined with Everolimus (Afinitor), with Everolimus alone. The results presented by the company demonstrated that there was no difference in PFS between treatment arms. The company did take the position that a sub-population analysis hinted at efficacy (though given the number of patients, not with statistical significance) and subsequent biomarker work suggests a patient profile that is likely to be a responder, based on 6 month progression profiles. The challenge is that BNO has no data to substantiate whether this profile predicts a responder to Everolimus or Everolimus+BNC105. To be honest, I no longer really understand where this is going.
To conclude, this is a drug that has been reinvented a number of times. It started off as a VDA, then it became sexy to accentuate an mTOR potentiating effect (in combination) and now it is obviously a potential combo therapy for checkpoint inhibitors (presumably due to some fundamental cytotoxicity, rather than anything cleverer). Despite all these incarnations and guises, the fundamental reality is this: there is simply no evidence that this drug does anything useful, either as a monotherapy or a combination therapy. There is plenty of slick positioning that makes this drug look a lot more exciting than it really is.
BNC101 : Anti-LGR5 mAb
I will try to keep this one short. It’s also probably the yukkiest one from an investor perspective because it is a biologic, and it costs serious $$$s to do the front-end development. We’ve known for a couple of years now that BNO is writing some pretty hefty checks to Lonza. We are also reminded incessantly that this asset basically came from Biogen Idec (via Eclipse and Dyax), which somehow implies a quality mAb/target. That may be true, but it is far from obvious to me. I also don’t particularly like it when a small molecule company takes on a biologic because the development skills required of the “fully vertically integrated” company are rather different.
But I am willing believe that BNO has the talent and capital to make it happen. So let’s park that issue for now.
LGR5 (aka GPCR49), the target that BNC101 “hits”, is part of the family of receptors called G-protein coupled receptors (GPCRs) and there is no doubt that this is a hot target space at the moment. There are many GPCRs that remain unexplored and even undiscovered. Indeed, that this part of the mystique of this target arena. However, we have known about LGR5 for almost two decades, including normal tissue expression of the target. Notwithstanding any potential anti-cancer effect, LGR5 is highly expressed in a wide range of non-cancerous tissues, including mundane and ubiquitous things like muscle tissue, presumably part of the reason why Eclipse never got any traction for this drug. This normal expression profile is problemetic when dosing a patient with a biologic targeting LGR5, because although it might be over-expressed in certain cancer stem cells (it is), you are probably going to have a hell of a protein sink when administering this drug. What this means is that all the normal tissues act like a big sponge and suck up the dose, so in order to deliver the drug to where you want to, it could mean huge doses overall. This in turn means a higher risk of toxicity (biologics are usually safer but dose still matters), shorter circulation half-life (impacts drug administration regimen) and cost (more material = more $$$s).
But even this aside, it’s not immediately evident that LGR5 is a good target for a cancer drug. A biomarker? Yes, almost certainly. A drug? I am not so certain. Part of the problem is that LGR5 expression seems to be the result of a co-regulatory gene in the Wnt / β-catenin (“pronounced “wint”) pathway, a fundamental regulatory pathway that is implied in both stem cell regulation and cancer proliferation. This pathway is incredibly complex, to the extent where anytime I have ever attended a scientific presentation on it, my eyes start to glaze over in about 3 minutes (and I will accordingly try not to bore you here). There is no doubt that LGR5 seems to be a good diagnostic biomarker for cancer stem cells, particularly in certain cancers like colorectal cancer. However there are conflicting views about whether targeting LGR5 is a particularly strong therapeutic concept, with the prevailing view that LGR5 overexpression is part of an important tumour suppressing mechanism and, sort of non-intuitively, may actually increase cell adhesion (i.e. prevent metastasis). Certainly, there are more promising parts of Wnt to hit with a drug than LGR5. Therefore hitting LGR5 with an antibody, thereby suppressing LGR5, might not necessarily give the desired effect in patients.
To be clear, I am not completely negative about BNC101, I just don’t see the right level of science in the decision-making and I am not sold on an asset just because it has some big names in its commercial lineage. There is no doubt that recent poster presentations show an impressive anti-tumour effect, and this deserves acknowledgement. It’s just that these models are sort of naive and might not be capturing the regulatory role of LGR5 signalling in the “intact” tumour environment (let alone the effects of normal tissue expression). To be clear, even a patient-derived xenograft is almost certainly not going to demonstrate native CSC behavior in an animal model.
I have elected not to spend time talking about the melanoma program (BNC420) which is a neat story but seems to just ignore the fact that melanoma is known to have many different metastatic mechanisms. Notwithstanding BNO’s impressive success in engaging with Merck, I am not going to bore you with commentary about the probability of success of an Alzheimer’s drug (0%), let alone a CNS drug (<5%). They could be beautiful molecules but all it means is those particular assets are at the speculative end of the spectrum when trying to work out what this company is really worth.
In short, this is a company that is commercially aggressive and has shown that it can play with the big boys. At first blush, it does a lot of things well, including promoting itself. However the science is fundamentally weak and the lack of visible clinical data to substantiate the company’s claims is a real concern. Let’s hope that this is something that the management team recognises as an area for improvement in the future. Only when that improvement happens, and not before, will the company be able to justify its current market cap.