Derisking Cool-Bio’s CLB-100
Ronald Rothman

As an invited panelist at the February 13, 2012 meeting of Pharmaceutical Consulting Consortium International, Inc., Ronald Rothman advised Cool-Bio on derisking their lead therapeutic product to interest potential partners

I like this concept – it’s simple yet clever, a probable solution for an unmet need. It uses a concept that is very well established – temperature-dependent conformational change in structure – as the basis of an on-off switch.

To make a stronger case that this has the potential of becoming a commercial success, I did a gap analysis – identifying experimental results that were inconclusive, deficient or nonexistent.

The mechanism of action (MOA) is demonstrated by inhibition of binding of fibrinogen to its receptor on platelets. This very nicely demonstrates target specificity of the antibody, but I see a few problems that detract from an otherwise very good presentation. First, there is no observed dose response. Perhaps it’s due to receptor saturation; and lower doses may need to be assessed. But showing the higher dose adds no value to the presentation. Also, there’s a significant amount of binding under control conditions. In a prior conversation (with Dawn Bell, president and CEO), there was some confusion about the nature of the control condition – whether it was normal temperature, or quiescent platelets. Significant binding at 37 deg might be a red flag, detracting from your value proposition as a safer alternative to currently available anti-thrombotics. But it might be a consequence of low avidity binding, perhaps characteristic of quiescent platelets; and this low avidity binding may be irrelevant to anti-thrombotic activity of your antibody. In which case, eliminating this distraction may be as simple as repeating the binding assay under more stringent conditions to reduce this low avidity binding. In the meantime, the inclusion of error bars would be beneficial to demonstrate the difference in binding under the 2 different conditions.

In vivo efficacy is demonstrated by an increase in jugular blood flow. I’m unable to comment upon the validity of this biomarker. Although I’m a firm believer in the utility of biomarkers, there are skeptics, as biomarkers are not always predictive of health outcome. What immediately comes to mind is the Vytorin trial a few years ago that demonstrated improvement in the biomarker status for which the trial was designed, yet no improvement in health outcome for which the drug was indicated.

In your study, perhaps improved blood flow is not from prevention of thrombosis, but some unrecognized effect of your drug causing vasodilation, an effect totally unrelated to the proposed anti-thrombotic mechanism. Probably unlikely, but until disproven, this remains a possibility, as the MOA, rather than platelet inactivation (anesthesia); and remains a source of speculative risk that this venture will fail.

So I would prefer demonstrating efficacy with the effect upon a health outcome. Looking at the potential indications, this drug would protect against necrosis from ischemia-reperfusion injury. One possible endpoint would be upon extent of myocardial infarction. Several animal models are available; and analysis involves histopathology and some not too sophisticated imaging.

I would also like to see this in vivo efficacy demonstrated at various lowered temperatures, particularly in the 32 to 34 deg range. There are different strategies for inducing hyperthermia, each resulting in different degrees of cooling. For reasons I will discuss later, a temperature profile demonstrating health outcomes under these more-milder hypothermic temperatures would be very desirable.

And just a bit of presentation advise. Although you want to minimize the number of technical data slides, I’d separate the experimental data into 2 slides, 1 concept per slide. The side-by-side comparison spotlighted a glaring inconsistency: The efficacy study was done at a lower temperature than the binding study. It made me wonder how effective the drug is under mild hypothermic conditions. Which is another reason for the need of a temperature profile of efficacy.

The next few points I will bring up only because they are highly desirable. I suspect they already have been accomplished; but from the presentation, I was uncertain, and they should be more prominently mentioned in the presentation.

Some potential patients for this drug may experience repeated episodes of stroke or cardiac arrest, requiring therapeutic hypothermia. So humanizing the antibody would prevent rejecting the antibody on subsequent administration. Also, a scFv (single chain variable fragment) would eliminate any concerns about unnecessary effector functions – like cytotoxicity – that would be mediated by a full-length antibody.

As I mentioned, I suspect these are already features of your drug. If not, they would be highly desirable. But considering the time required to engineer these constructs, doing so would detract from your primary goal of demonstrating efficacy; and should be considered more of a wish list. Nevertheless, you may wish to consider incorporating them before you get too deeply involved in safety studies, to prevent the expense of a new or bridging study.

The greatest deficiency of this study is a demonstration of safety. Foremost, I’m concerned about the stability of the blocking peptide “switch” – is it susceptible to proteolysis during the time the antibody persists in circulation? This study would need to be done in conjunction with a clearance study. The antibody needs to persist in circulation long enough to be therapeutic. Yet it needs to be cleared rapidly enough before becoming susceptible to proteolysis, losing its off-switch, and essentially becoming a rogue antibody. At which point your drug would lose its safety advantage over commercially available anti-thrombotic therapeutics, some of which are small molecules, which are priced much cheaper.

Also of concern would be any off-targeting effects. This could be mediated by the blocking peptide “switch”, which exists in 2 conformations, depending upon the temperature. Or off-targeting could be mediated by the antibody itself. Even if the antigen is uniquely expressed on platelets, low avidity binding sites may be expressed elsewhere, on other tissues. Alternatively, off-targeting could be a consequence of the antibody preparation – either aggregates or misfolding of the antibody could generate novel binding sites on the molecule. Off-targeting would be evaluated by the tissue distribution of the antibody. Essentially, you want assurance that the antibody does not wind up anywhere that it should not be.

Additionally, is there any immunotoxicity, such as immunosuppression, immunostimulation, or hypersensitivity (immunogenicity or allergenicity)? Getting back to the desirability of the construct features, humanizing the antibody would eliminate some of these concerns, and scFv would address the others.

Also, I have one lingering concern about safety: What is the effect of irreversible binding of this antibody to stimulated platelets during the 1-week period that platelets remain in circulation? Admittedly, only a small fraction of platelets would be affected. And amongst these platelets, only a small fraction of the receptors would be inactivated. But what’s the effect upon bleeding-time subsequent to rewarming days later? Does this therapy essentially result in pharmacologically-induced hemophilia? Nevertheless, the benefit would probably outweigh the risk. And admittedly, this would be a very challenging question to answer in an animal model using a humanized antibody.

As for the last point I wish to make, I don’t wish to go into too much detail, without risk of encroaching upon Deni’s discussion. But in a discussion of derisking, I’d be remiss if I didn’t emphasize the necessity of a Voice-of-the-Customer survey to align performance with clinical need – to derisk the venture.

You should be interviewing everyone, not just surgeons, but also the doctors and nurses in the ER and intensive care units – any practitioner who would treat cardiac or stroke patients with therapeutic hyperthermia – about their needs and how they’re currently being met or, more importantly, not being met.

There are various strategies (internal v external) for inducing hypothermia, each inducing a different degree of cooling. Deep hypothermia (28 deg or less) is common in surgery, but mild hypothermia (32 to 34 deg range) has broader application. Mild hypothermia is commonly used out-of-hospital by paramedics responding to cardiac arrest. But this might be an impractical use of your drug for financial and logistic reasons (i.e. refrigeration, cold chain). However, mild hypothermia is also commonly used in-hospital, in the ER and intensive care units. Potentially, these could be a very large and attractive market to pursue. The question becomes: Where’s the greatest need? And who is more willing to use this drug? So getting back to my previous recommendation of performing a temperature profile for efficacy: It’s important to know the preference of the practitioners; and match their need with the temperature performance of your therapeutic.