Biodel F2Q14 – Concentrated ultra-rapid BIOD-531 moves to phase 2 meal studies; FDA feedback on phase 3 for BIOD-123 – May 15, 2014

Biodel held its F2Q14 financial results update earlier this week in a very organized call led by CEO Dr. Errol De Souza. Management’s enthusiasm was clearly highest for BIOD-531 (“a gamechanger for Biodel”), the company’s phase 2 concentrated ultra-rapid-acting insulin candidate. Below, we enclose the call’s top six highlights, followed by a pipeline summary (insulin and glucagon) and Q&A.

1. Biodel has started dosing patients in the first of two planned phase 2 meal studies of BIOD-531 (U400 ultra-rapid-acting concentrated human insulin). Topline data is expected in 4Q14.

2. Biodel received very helpful, detailed FDA feedback on the proposed phase 3 program for BIOD-123 (ultra rapid-acting recombinant human insulin). Most notably, the FDA will not require a CVOT. BIOD-123 will move to phase 3 only after Biodel finalizes a partnership deal.

3. Biodel is discussing its analog-based ultra-rapid-acting insulin portfolio (both aspart and lispro-based formulations) with potential partners, and the interest level “is very good.” These would be bundled into a portfolio deal with BIOD-123. Biodel has identified viable lispro-based ultra-rapid-acting formulations with the appropriate stability to move into the clinic.

4. Biodel expects a “late 2015” NDA submission for its dual-chamber auto-reconstitution device for severe hypoglycemia. An IND will be submitted in 3Q14 and a clinical trial is expected to start in 4Q14.

5. Biodel will present four abstracts at ADA 2014: two late-breaking abstracts on BIOD-531 (preclinical pump data and the previously shared phase 1 data), an oral presentation on the previously shared phase 2 data for BIOD-123, and one abstract on the company’s glucagon formulation for use in the GEM device (bioequivalence to Lilly).

6. As of March 31, Biodel had cash and cash equivalents of $28.7 million, enough to fund operations until at least the end of 2Q15.

Top Six Highlights

1. Biodel has started dosing patients in the first of two planned phase 2 meal studies of BIOD-531 (U400 ultra-rapid-acting concentrated human insulin). Topline data is expected in 4Q14. Management’s excitement over this compound continues to be palpable (“a gamechanger for Biodel”), and we’d note that this timing represents a very rapid advancement (positive phase 1 data was shared just a few months ago on February 12) and right on par with development plans from the F1Q14 call. The single blind, four-way crossover studies will compare BIOD-531 (given before and after a meal) to Lilly’s Humulin R U500 and Lilly’s Humalog Mix 75/25. One trial will enroll 12 patients with type 2 diabetes who use 50-200 units per day (the target population that uses premix insulin). The other study will enroll 12 patients with type 2 diabetes who use >200 units of insulin per day (the target population that uses U500 insulin). The goal of these phase 2 studies is to “unambiguously” show lower postprandial glucose excursions with BIOD-531 relative to the comparators.

• Biodel has submitted questions to the FDA regarding the regulatory path for BIOD-531; feedback is expected in 3Q14. The key questions relate to the regulatory requirements for approving a concentrated insulin, since Lilly’s Humulin R U500 never went through pivotal studies (it was “grandfathered in”). Once the feedback comes in, management “could talk about moving [BIOD-531] into late stage clinical trials.”
• Notably, Biodel has conducted preclinical insulin pump studies of BIOD-531 through an NIH SBIR grant. Management specifically mentioned potential application to the artificial pancreas. The faster onset of action and more concentrated profile would of course benefit closed-loop algorithms and pump reservoir size, respectively – the key unanswered question is to understand the PK/PD profile of BIOD-531 at lower doses. At the higher doses being tested in the aforementioned type 2 studies, BIOD-531 does have quite a long duration of action that might complicate closed-loop control; however, this remains to be further investigated. Biodel will have a poster at ADA on use of BIOD-531 in pumps in diabetic swine. The other pump application of BIOD-531 related to the growing segment of type 2s taking U500 in a pump, something Insulet is working on with Lilly.
• Management reiterated that based on the phase 1 data shared in February, BIOD-531 could compete with both insulin premixes (e.g., Humalog 75/25, Novolog 75/25;) and Lilly’s concentrated Humulin R U-500 – these represent a $1.5 billion annual market in the US for premix and a $400 million annual market for Humulin R U500 (2014 projection). Management believes Biodel could commercialize the U500 opportunity on its own, as marketing to endocrinologists is possible with a small sales force (and in line with the company’s glucagon program). Regarding the premix opportunity, Biodel would likely look for a partner to take BIOD-531 to a primary care audience.

2. Biodel received very helpful, detailed FDA feedback on the proposed phase 3 program for BIOD-123 (ultra rapid-acting recombinant human insulin). The FDA weighed in on the design of proposed phase 3 clinical studies to secure approval of BIOD-123 – in Q&A, management called characterized them as “pretty standard” parallel group pivotal trials in line with most insulin products. FDA weighed in on the size in type 1 and type 2 (smaller in type 1, given the phase 2 data), the length of titration/stable dosing periods, and regulatory expectations. Notably, Biodel will not be required to carry out a CVOT for BIOD-123 – we wonder if that’s because there is a general level of comfort with recombinant human insulin. The FDA feedback provides Biodel with a lot of clarity that will now “greatly assist discussions with partners.”

• Moving BIOD-123 into phase 3 will only occur once Biodel finalizes a partnership deal. These deal discussions are ongoing and also include the analog-based ultra-rapid-acting insulin portfolio (see below). Biodel is ideally looking for a “pharma company with adequate resources to commercialize BIOD-123 for type 1 and type 2 diabetes” – we assume that only the established insulin players would meet this criterion, which likely means Sanofi or Lilly (Novo Nordisk does not typically partner and has its own ultra-rapid-acting version of aspart in phase 3).
• At ADA, Biodel will give an oral presentation on the phase 2 BIOD-123 results, which we covered in detail last September.

3. Biodel is discussing its analog-based ultra-rapid-acting insulin portfolio (both aspart and lispro-based formulations) with potential partners, and the interest level “is very good.” As noted above, these would be bundled in a portfolio with the ultra-rapid-acting recombinant-human-insulin-based BIOD-123. Biodel has identified viable lispro-based ultra-rapid-acting formulations with the appropriate stability to move into the clinic, and the company is in active discussions to source GMP supplies of lispro. The company has also formally kicked off an ultra-rapid-acting aspart-based program following April’s signing of a research supply and technology development agreement with China-based HEC Pharm. Under the terms, HEC will supply Biodel with insulin aspart (the active pharmaceutical ingredient in Novo Nordisk’s Novolog), allowing Biodel to add its proprietary excipients (including EDTA, citrate, and magnesium sulfate) and ultimately make an ultra-rapid-acting formulation of insulin aspart. We saw the news as a major positive for Biodel, as an API source of lispro/aspart was one of the gating factors for advancing the analog-based ultra-rapid-acting insulin program forward.

• Looking forward, the regulatory path for this program would be similar to other 505(b)(2) insulin development efforts – that allows referencing of existing data on Novolog and Humalog and potentially moving directly from phase 1 to phase 3 (although the latter is not a given). However, to secure approval and start selling an ultra-rapid version of insulin aspart or lispro, the appropriate patents would need to expire – to our knowledge, 2017 in the US for Novolog (aspart) and June 2014 for Humalog (lispro).
• As a reminder, Biodel reported phase 1 results in January 2013 on BIOD-238 and -250, its ultra-rapid-acting versions of Humalog (insulin lispro). In the 12-patient study, absorption rates of BIOD-238 and -250 were significantly more rapid than that of Humalog, which translated to 35-45% reductions in mean times to half maximal insulin concentrations (an early Tmax of 14 minutes for BIOD-238, 15 minutes for BIOD-250, and 25 minutes for Humalog; p<0.001 and p=0.001 vs. Humalog, respectively). Encouragingly, the key problem with Biodel's phase 3 Linjeta, injection site tolerability, was mitigated with BIOD-250 by adding magnesium sulfate.

4. Biodel expects a “late 2015” NDA submission for its dual-chamber auto-reconstitution device for severe hypoglycemia (what is now being called the “Glucagon Emergency Management” or “GEM” device). We’d note that the previous F1Q14 timing called for a more vague “2015” submission, though this is not necessarily a delay, as the 2014 milestones are running exactly in line with previous expectations: an IND for the GEM device will be submitted in 3Q14 and a clinical trial is expected to start in 4Q14. During this past quarter, Biodel completed the dosing portion of an ongoing GLP toxicology study and executed a commercial manufacturing agreement with Emergent BioSolutions (to fill and finish commercial quantities of the dual-chamber device). Biodel expects to conduct at least two human factors studies in 2014, which will compare the GEM device to existing rescue kits. Management noted that the Emergent commercial manufacturing agreement, combined with the previously announced 15-year supply agreement with Unilife (for the custom dual-chamber device itself), “potentially completes the supply chain requirements” for an NDA submission and scale up for a final commercial product.

• Regarding stable liquid glucagon for use in the ultra-portable, low cost BD Uniject device (prefilled multipacks that last four months each) for severe hypoglycemia, Biodel is now testing several formulations in animal models. Management added that the company has “made solid progress in improving stability.” This was a less specific update than offered in the F1Q14 call, where the hope was to get a formulation worked out in the second half of 2014 (at least six months of real-time stability is needed). As a reminder, this very novel approach to stable glucagon markedly lowers the hurdle for room temperature stability from two years to four months – patients would simply take a new Uniject out of the fridge every four months, but without the need to get a whole new prescription. Management believes it will be “particularly compelling for active people with diabetes and parents/caregivers.” Things are clearly still very early (the program was announced just two quarters ago), so we expect firmer timing as development progresses. For much more detailed information on the Uniject product, see our Biodel F4Q13 report.
• Along with the Uniject device, Biodel “continues to develop” stable glucagon formulations for use in insulin pumps/artificial pancreas. No timing update was provided, though we expect more specifics will materialize in parallel with the Uniject program. These efforts are being funded by an NIH SBIR grant.

5. Biodel will present four abstracts at ADA 2014: two late-breaking abstracts on BIOD-531 (preclinical pump data and the previously shared phase 1 data), an oral presentation on the previously shared phase 2 data for BIOD-123, and one abstract on the company’s glucagon formulation for use in the GEM device (bioequivalence to Lilly).

6. As of March 31, Biodel had cash and cash equivalents of $28.7 million, enough to fund operations until at least the end of 2Q15 – the same runway given in the F1Q14 call in February. Cash burn was $6.5 million during F2Q14, up from $4.6 million in F1Q14. R&D expenses were $4.3 million in F2Q14 vs. $3 million in F2Q13; the increase was primarily due to expenses associated with the development of Biodel’s dual chamber auto-reconstitution GEM device.

Pipeline Summary – Insulin

Pipeline Summary – Glucagon

Questions and Answers

Q: What is the status of the ongoing partnership discussions for the ultra-rapid acting prandial program?

A: As you know from the past, we won't project when a partnership deal might be consummated. But having said that, I think the interest level is very good – and it's really a bundling approach in terms of our discussions with partners to really leverage the portfolio. The portfolio of ultra-rapid-acting prandial insulin candidates currently includes several lispro-based formulations, which would achieve both the stability profile and the PK/PD profile similar to what we showed in the clinic with BIOD-250. It obviously includes the later-stage assets, including BIOD-123, where we not only met the primary endpoint but also showed improvements relative to Humalog in terms of the meal challenge study. And finally, now we are adding to that portfolio a program that was just initiated, which would be to develop aspart-based ultra-rapid-acting insulin formulations. So it's truly a portfolio discussion that we're having with the different players, and the interest level is clearly there.

Q: And now with the visibility from the FDA on the pivotal study requirements, that should assist in your discussions?

A: I think what it does is actually provide clarity, because the obvious question that was always asked related to 123 is, “What would the pivotal trial look like, what toxicology requirements are there, do you need a cardiovascular outcome trial as the FDA has required for several studies?” What our correspondence with the FDA gave us was clarity in terms of the trials. We submitted to them for consideration four protocols. And they gave us detailed feedback, most of which concurred with our plans. So we really have very definite guidance on how big these trials would need to be, how long they would need to last, what sort of expectations they would have, and what sort of regulatory expectations might exist in terms of the output of the trial.

Q: Is there anything you are willing to share with us now?

A: I think these trials are in general pretty standard. And the trials we're talking about now would look very much like the parallel group randomized studies that are done as pivotal trials for most insulin products. We got some definition on how the trial would be divided. Most of these trials, as you know, are six months in treatment duration, and we clarified with the FDA how long we should take to titrate the doses vs. how long we should leave the doses relatively stable, and then measure A1c. Some analytic features were discussed, some statistical features, but in general, the overview of the trial is very similar to other pivotal trials.

As an example, I think there was an acknowledgement in terms of the efficacy demonstrated in phase 2 – the pivotal study in type 1 will be smaller than the pivotal study would be in type 2 patients, as an example. They encouraged us to blind the trial as much as possible based on some of the discussions. So we got a lot of feedback, which is exactly what we were looking for, and it helps us in terms of the discussions that we're currently having.

Q: In terms of the 531 program, now that you’re in the phase 2 meal studies, what should we look for in that data once it’s announced?

A: I would hope that 531, based on its PK and PD profile that we've demonstrated in the initial clamp study, I would hope that we're able to show unambiguously that the post-prandial glucose excursion after a meal is lower with 531 relative to the comparators (Humalog Mix and Humulin R U500). I think that would be a clear demonstration of proof-of-principle that the more rapid absorption and the more rapid early glucose lowering effect translates into better coverage at mealtime.

A key thing is to gain clarity from the FDA in terms of what is the path to NDA submission. As a reminder, Humulin R U-500 never went through doing pivotal studies. It was grandfathered in. So we need that clarity. And as soon as we get that clarity and we anticipate getting that in the third quarter since we've already filed our questions, we could come back and talk about moving into late-stage clinical trials, which really is a gamechanger for Biodel in terms of a late-stage program. So that coupled with the meal studies, gives some very nice value driving milestones in the second half of 2014.

Q: And you should have that feedback in the third quarter?

A: We anticipate – there's no clock ticking on it. But I've got to say the FDA has been pretty good in terms of the general quarter and a half in terms of feedback, and that's the basis of our projections anticipating feedback in the third quarter, which is not that far away in terms of giving us clarity or whatever we need to do in terms of moving into pivotal trials.

Q: It sounds like you're going to have a relatively significant presence at the ADA this year. Could you give us a sense of what we should focus on related to the glucagon program?

A: At ADA, actually, we were pleased that two of the presentations are actually related to 531, both late-breaking – I think it gives us a really good sense of the interest level in 531 that they were selected after all the abstract deadlines had come about. For glucagon, we're going to publish the PK/PD data just showing bioequivalence to the Lilly product, and then talk about the GEM device. The other thing that we were very pleased with is Alan will have an oral presentation on the phase 2 data related to BIOD-123. So we've got a full plate with four presentations at ADA.

Q: Did I understand correctly that you would not start a phase 3 BIOD-123 trial until a marketing arrangement had been done?

A: Right. In line with the strategy that we've outlined over the last couple of calls, we've got a lot on our plate. And we really need to think about which programs we fund with dollars from cash that we currently have, so that we can drive significant milestones to build shareholder value, and which programs are best developed in the hands of partners that can not only bring the larger financial resources that would be required for a phase 3 study – for example, with BIOD-123 – but equally is important the commercial infrastructure that is required to compete in this market where you have Novo and Eli Lilly dominating this area. So that's the reason for the decision that we've made on the ultra-rapid-acting prandial insulin program – that’s BIOD-123 and the lispro-based approaches – is to really move forward with a partner and not on our own.

Q: Just wanted to follow up on the rapid-acting program and the discussions you had with FDA. Did you, as well as discussing trial design, discuss the potential label claims and really get a sense of how you could at least test questions to get items of clinical differentiation into label?

A: Yes. Actually, we have had discussions with the FDA on what I think is the key component of the label that you may be getting at, and that is, “How do we convey in the label to prescribing physicians that this is a faster insulin than what's on the market now.” And the FDA gave us very specific guidance on a model study type that they've actually published that would achieve that – and specifically, it's a standardized meal challenge study.

And in fact, we used the feedback we received from the FDA on a proposed meal challenge study – that was in the context of BIOD-123, but we used that guidance to help us design these ongoing meal studies for BIOD-531. It's a same principle. In these studies, you take patients and you give them the experimental insulin versus the comparator insulin either right before a meal or right after a meal, and in some of these studies, you sometimes give it at different times with respect to the meals. And you look at postprandial glucose curves as a result of the various timings of the injection with respect to the meal. And so this actually was very, very specific and useful feedback we received and we're actually using that feedback as we speak.

Q: On 531, how do you think about the hypoglycemia risk in patients who are not severely insulin-resistant?

A: Well, that's an interesting question. I would say that in general, even non-severely insulin-resistant type 2 patients are much less prone to hypoglycemia overall than type 1 patients. And so I would imagine that there would be no difference in risk. If anything, I might even hope to see a little bit of a benefit, although that might be hard to achieve, because the incidence rate of hypoglycemia is so low in this patient population. So in order to see a difference statistically, you might need a tremendous sample size. But in general, I certainly wouldn't expect excess hypoglycemia when you're able to match a dose to a meal, like type 1's often do. And when you achieve better prandial coverage, I think that actually allows you to better choose the doses and, if anything, prevent hypoglycemia. But in general, I wouldn't imagine that would be a problem for any of the insulins that I'm aware of in the type 2 populations.

This is a product where we're looking at really leveraging the more rapid absorption with a label advantage, which would be from the meal studies, but then the rest of the pivotal study that frankly we have proposed to the FDA and await their guidance would be a standard non-inferiority. And then we would have the advantage of lower volume injections vs. the mix population, and we would be comparable to the volume injected with concentrated insulin. But there we should have real benefits in terms of the onset of action based on its more rapid absorption. So this is a little bit of a different type of a product where hypoglycemia, frankly, is something we will monitor, but we're not projecting any added benefit there, especially in this population.

If I could add, another thing that might correlate with increased hypoglycemia risk is the overall duration of action of an insulin. In some cases, we can see stacking of insulin that hangs around in the system too long. What we showed in our phase 1 study was actually a lower overall duration of action compared to Humalog Mix or Humulin R U500. Again, I would predict that might come into play particularly overnight when patients are sleeping. So we might be able to show less nocturnal hypoglycemia, but again that awaits multi-dose trials.

--by Adam Brown and Kelly Close