BPX-501 — Value Proposition Explained
Risk / Reward is Widely Misunderstood
See disclosure at the end of this blog
This post was designed to explain why Bellicum’s BPX-501 augmented HSCT is superior to existing alloHSCT approaches and why BPX-501 stands to demonstrate strong adoption once approved. Clearing up misconceptions around BPX-501 and α/β T cell depleted haploHSCT as well as eliminating the noise around the inferior PostCy regimen is the first step in assessing the unmet need. The data we evaluate in this post argue that nearly all alloHSCT patients would benefit from BPX-501.
- There is significant room for improvement in HSCT outcomes across donor types, including matched unrelated (MUD), unmatched (haploidentical) and even matched related (MRD, or matched sibling) as well as disease types, from orphan disorders to malignancies
- PostCy / John’s Hopkins Protocol for haploidentical HSCT is inferior by nearly all measures to the recently developed α/β T cell depleted haploidentical HSCT procedure — even without BPX-501
- BPX-501, when combined with the novel α/β T cell depleted haploHSCT procedure, prevents lethal infections and demonstrates superior outcomes over α/β T cell depleted haploHSCT alone
- Early results suggest BPX-501 augmented α/β T cell depleted haploHSCT has the potential to be standard of care in the majority of the existing alloHSCT markets and to expand the market to populations formerly seen as an opportunity for gene therapy (beta thalassemia / sickle cell disease)
- The risk to the success of BPX-501 in alloHSCT for malignancies is not what you think. Prior data show there is a low risk of failure — it is really a question of when and how you measure success.
BPX-501 — A Deeper Explanation
Bellicum continues to march forward toward potential commercialization of its lead program, BPX-501, in Europe in late 2017 for treating patients undergoing alloHSCT to cure Orphan Blood Disorders and eventually for patients undergoing alloHSCT for malignancies.
There remains significant confusion about the unmet need addressed by BPX-501. Many investors think BPX-501 is for GvHD. While BPX-501 can eliminate GvHD via activation of its CID iCasp9 switch with rimiducid, ameliorating GvHD is not the purpose of BPX-501, per se. BPX-501 prevents lethal infection. It does so while eliminating GvHD risk.
Throughout ASH, investors watched closely for GvHD rates and the use of the CID switch to eliminate alloreactive T-cells. Viewing acute GvHD rates as the primary measure of success has resulted in the misconception that post-transplantation cyclophosphamide (PostCy or PtCy), also called the John’s Hopkins Protocol, is comparable to BPX-501 enabled haploidentical HSCT. As we will show, the PostCy regimen is inferior to the underlying procedure with which BPX-501 is used (α/β T cell depleted haploHSCT) — even when BPX-501 is not administered. When BPX-501 is added, those outcomes are further improved. In fact, the data argue that the use of PostCy puts patients at significant additional risk of transplant related mortality, high grade acute GvHD and a high rate of chronic GvHD.
To be clear, GvHD will never be a problem with BPX-501. The switch works that well. The primary question is whether using BPX-501 reduces transplant related mortality due to viral infection and, specifically in malignancies, whether it also lowers relapse rates. As we will show later, there is no rationale why relapse rates would be worse with BPX-501. However, it will be a while before we have enough relapse rate data to assess BPX-501 in malignant patients.
In the meantime, we will focus on the more near term measures for which we have data when evaluating BPX-501, including GvHD and transplant related mortality (driven by lethal viral infection rates). These are the key measures of risk outcomes for the orphan blood disorder indications, the area of BPX-501’s first potential approval.
A Review of AlloHSCT
Allogeneic HSCT is the primary curative measure for malignancies such as leukemia and lymphoma as well as orphan blood disorders such as beta thalassemia, sickle cell disease, severe combined immunodeficiency (SCID, bubble boy disease), certain anemias and other primary immunodeficiencies. In the US, approximately 90% of allogeneic HSCTs were for malignant diseases and 10% for non-malignant diseases (mostly orphan blood disorders).
There were roughly 8,000 allo-HSCTs performed in the US and 15,000 in the EU in 2012. Of the total HSCT procedures, roughly 25% are from a matched sibling (MRD), considered to be the best option, if available. Nearly 65% are from a matched unrelated donor (MUD) from a registry, considered to be the next best thing. MUDs can take months to find and many patients never find a match. An estimated 40% or more of patients looking for a donor are unable to find one. Finally, ~10% of alloHSCTs are from haploidentical donors, usually parents but can be other relatives, which are partial HLA matches (sometimes called “unmatched” donors.) This is often a last option due to increased GvHD risk.
While there are many meaningful measures of HSCT success, the most significant risk outcome measures include:
- Transplant Related Mortality / TRM (Due to mostly to viral infection but sometimed uncontrolled GvHD)
- Relapse Rates (in malignant disease only)
- Acute GvHD (typically within the first 100 days post HSCT)
- Chronic GvHD (occurring after 100 days, typically)
Rates of chronic GvHD typically require 12 months to accurately assess and relapse rates can take 2+ years. However, it takes three months to get a clear view on acute GvHD and mostly clear view of infection risk/TRM. High infection rates and resulting mortality (TRM) have been the main trade-off in measures to reduce GvHD. Overcoming this trade-off has persistently eluded transplant physicians and surgeons.
Therefore, BPX-501 should be evaluated not on GvHD rates and switch activations, but on overall outcomes and particularly its ability to reset the balance of lethal infection vs. GvHD.
Essential Background: The haploHSCT Procedure Landscape BEFORE BPX-501
The basic HSCT donor alternatives include (1) matched sibling (MRD), if available (2) MUD, and absent those options (3) a haploidentical HSCT. Much work has been done to improve haploHSCT to address the unmet need — particularly by reducing GvHD, yet lethal infection rates have been the inevitable consequence.
For several years the main “advance” in haploHSCT that has been adopted (primarily in the US) is post transplantation cyclophosphamide — sometimes called the Johns Hopkins Protocol, which I explain here. In brief, this decades old chemotherapy is used post transplantation to ablate mature T cells to prevent GvHD, since cyclophosphamide is selective for these mature cells.
Several years ago, however, a team in Italy developed a novel procedure to deplete the haploidentical stem cell transplants of mature α/β T cells. It is a more elegant way to accomplish eliminate the more alloreactive cells to reduce incidence of GvHD.
Due to its recent development, the published long term data on the α/β T cell depleted haploHSCT procedure have been sparse. However, Alice Bertaina from Bambino Gesù Children’s Hospital in Rome, where this procedure was pioneered, presented a fairly robust data set at ASH in December of 2015 that presents a compelling case for α/β T cell depleted HSCT over conventional HSCT methods, such as PostCy. (In the next section, we will discuss how the addition of BPX-501 makes this case even stronger, but for now will focus on α/β T cell depleted haploHSCT alone.)
The Bertaina paper evaluated a cohort of pediatric patients with leukemia that underwent either matched sibling HSCT (n=41), MUD HSCT (n=51) or α/β T cell depleted haploHSCT (n=80). What is useful to mention is that this is a clean study in that all patients were treated in the same way at the same institution with differences varying primarily by donor type. Detailed methods can be found here in the ASH presentation.
The results strongly suggest that α/β T Cell Depleted HSCT are equivalent or superior to MUDs and even matched sibling transplants on virtually all measures.
α/β T cell depleted haploHSCT is superior to MUDs and even matched sibling donors in all GvHD measures; these results were statistically significant for high grade (III-IV) acute GvHD (p<0.001) and all chronic GvHD (P<0.02). None of the 80 α/β T cell depleted haploHSCT patients had any form of high grade GvHD, either acute or chronic.
The cumulative incidence (CI) of acute GvHD was 30% for α/β T cell depleted haploHSCT, 41% for HLA identical sibling and 42% for MUD. Remarkably, all children who underwent the α/β T cell depleted haploHSCT who developed acute GvHD had a skin-only involvement, while 17% of matched sibling and 16.3% of the MUD group had either gut or liver involvement (p<0.001).
The CI of chronic GvHD was significantly lower in children undergoing α/β T cell depleted haploHSCT than those with matched sibling or unrelated donor (p=0.02). None of the α/β T cell depleted haploHSCT patients experiencing chronic GvHD had the extensive form of the disease, while the CI of extensive chronic GvHD of those with matched sibling or unrelated donor was 8% and 14%, respectively (p=0.01). — Alice Bertaina, MD PhD.
The incidence graph below shows that if chronic GvHD occurs it generally does so primarily in the first 12 months. Interestingly, only the MUD group had a later occurrence.
α/β T cell depleted HSCT was equivalent or superior to all alloHSCTs of all donor types except the sibling match group in terms TRM(though the difference was not significant) and this is where BPX-501 comes in later. Note below that the two were extremely close.
α/β T cell depleted haploHSCT also trends better than both MUD and sibling match on 3 year probability of event free survival. So already this makes the case for a haplo-HSCT (specifically an α/β T cell depleted haploHSCT) as a viable option in lieu of a MUD or Sibling match.
In fact, in all measures, α/β T cell depleted haploHSCT either handily outperforms or is equivalent to all other alloHSCT donor types in this study, MUD or MRD.
PostCy is inferior to α/β T cell depleted haploHSCT (even without BPX-501 incorporated)
The below table summarizes (1) the data that Alice Bertaina published in pediatric patients with malignancies, (2) the Locatelli published results (at 2015 ASH from the same institution) for α/β T cell depleted haploHSCT in non-malignant disorders and (3) aggregated data from the literature on Post-haploHSCT cyclophosphamide that was collected in the last installment of my blog. (No patients in this data set received BPX-501, that comes later).
You will notice that the ONLY measure that is comparable is total (Grade II-IV) acute GvHD. This is likely the reason people think PostCy is competitive. While overall acute GvHD rates are comparable, it should be noted that a substantial portion of the acute GvHD in the PostCy group (21%, or 6.2% overall) was higher grade while there were no such cases for the α/β T cell depleted HSCT group.
For the 113 patients that underwent α/β T cell depleted haploHSCT there were no cases of higher Grade III-IV acute GvHD, no cases of extensive chronic GvHD (roughly 80% reduced incidence of overall chronic GvHD) and a 6.2% rate of TRM, which is about half the rate of PostCy. PostCy patients also must remain on immunosuppressive a post HSCT — BPX-501 + α/β T cell depleted haploHSCT patients do not.
The data indicate that PostCy is an inferior option for patients — exposing them to greater risk of high grade acute GvHD, chronic GvHD (especially the more severe Extensive cGvHD) and TRM.
The Need for BPX-501
As shown, α/β T cell depleted haploHSCT is an option with significant advantages already and is likely to become much more prevalent in the US based on these improved outcomes (outcomes are how institutions are ultimately measured.) Yet there remains a need for improving transplant related mortality, which is driven primarily by viral infection, without re-introducing GvHD risk.
BPX-501 is an elegantly designed solution to add back mature T cells 7–14 days after α/β T cell depleted haploHSCT to prevent lethal infection risk (lowering TRM) that also allows for the complete elimination of GvHD in the event that it is occurs and is uncontrolled. For orphan blood disorders, this approach addresses the key morbidity and mortality risks.
Notably, BPX-501 modified T cells have been shown to continue to express the CID iCasp9 switch well beyond the 100 day mark when chronic GvHD becomes a concern and potentially up to a year or more. Theoretically, activation of the safety switch with the dimerizing agent could address chronic GvHD as well as acute GvHD. Thus far, we are not aware of such a use in chronic GvHD.
Also important, the switch only eliminates reactive T cells, not quiescent ones and therefore does not appear to eliminate immunity in patients when the switch is activated.
The data from ASH comparing α/β T cell depleted haploHSCT to the same procedure with BPX-501 suggests improved outcomes with BPX-501 in transplant related mortality (TRM) and similar or better rates of acute GvHD. Chronic GvHD is included in the table, however it is too early to assess in the BPX-501 group (though it could be addressed by the switch, as mentioned.)
Thus far, it would appear that BPX-501 does have the potential to address the lethal infection risk and that it will be a fundamental aspect of any α/β T cell depleted haploHSCT. More detailed data presented at ASH show that this early immune reconstitution also results in the control of existing viral infections and reduced morbidity as well as hospital stays and readmission rates. This is discussed in detail in this prior post that discusses immune reconstitution, among other things.
Finally, acute GvHD is rapidly addressed when it is not controlled by conventional means. We saw an example of the latter case at ASH with the single patient for whom steroids were not deemed appropriate. Administration of the dimerizing agent fully resolved the child’s GvHD within 60 minutes. The pictures are dramatic (apologies if you have seen this before.)
In short, BPX-501 in the context of α/β T cell depleted haploHSCT is producing outcomes that are superior to all prior experience in allogeneic HSCTs of all donor types.
BPX-501 Market Opportunity
Existing alloHSCT Market Demand
The market for BPX-501 is potentially most allogeneic HSCTs but will unfold initially in orphan blood disorders followed by malignant disease. While the data above strongly suggest BPX-501 will be used in all patients who do not have a matched related donor (MRD) / sibling, we can safely assume that the most obvious penetration will be in existing haploidentical procedures as well as patients who do not have a donor but are searching for one. Bellicum has the following market assessment in its corporate deck, which puts the initial market in the US and EU at >37,000 patients — excluding the MUD and MRD donor procedures below.
Even my slightly more conservative numbers, below, suggest that the addressable market is over 25,000 — but I think BPX-501 + α/β T cell depleted haploHSCT will be proven superior to MUD, at a minimum. That puts the population at nearly 50,000 using Bellicum’s sources. Either way, it is plenty large.
Beyond The Current HaploHSCT Market Demand— Orphan Blood Disorders like Beta Thalassemia & Sickle Cell Disease
A safe enough HSCT procedure would also make a viable option for less severe patients who currently are not pursuing a curative HSCT (think beta thalassemia, sickle cell etc.) For example, in Italy all Beta Thalassemia patients who have a matched sibling are transplanted at a very young age. Those who do not have a matched sibling are not.
In a research note dated March 1, 2016, Wedbush Securities published highlights from an interview with Dr. H. Franklin Bunn, a sickle cell blood specialist at Harvard Medical School and Brigham and Women’s Hospital in Boston.
(Dr. Bunn) reiterated commentary that allo-transplants are performed infrequently due to the mortality risk but noted that a safer transplant (i.e. autologous with gene therapy or gene editing) could be attractive, particularly in younger patients.
With BPX-501 this risk to alloHSCT is addressed. Discussed below is strong evidence that gene therapy is inferior to BPX-501 + α/β T cell depleted haploHSCT. As such, BPX-501 has the potential address nearly the entire market that investors once ascribed to bluebird bio’s gene therapy approaches and others.
Market Adoption Expectations for BPX-501
Strong Macro Drivers for Adoption of BPX-501 + α/β T cell depleted haploHSCT
The market is in dire need for a better and safer allogeneic HSCT. The results show that BPX-501+ α/β T cell depleted haploHSCT has the target profile. If we believe doctors just want the best for their (mostly pediatric) patients, the data suggest BPX-501 + α/β T cell depleted haploHSCT is materially safer and would gain rapid adoption. Thankfully, for these children, there are economic incentives too. First, hospitals are paid on a procedure basis and increasingly are risk sharing. Better outcomes means shorter hospital stays and fewer re-admissions, hence more profitability. Hospitals & physicians are also evaluated on quality of care — and outcomes. At some point, not adopting a superior procedure may present a legal liability, not just an ethical one. Awareness of α/β T cell depleted haploHSCT and BPX-501 in the US is low but it is front of mind in Europe. KOL awareness in the US is needed.
Adoption Timing Will Vary by Disease Type
(1) Orphan Blood Disorders (existing market)
Adoption of BPX-501 + α/β T cell depleted haploHSCT will initially be in orphan blood disorders and the data should drive strong uptake. Assuming the market for HaploHSCT+Untransplanted (from the tables above) is 15,000–35,000 (untransplanted being the variable), that pegs the orphan market (10% of alloHSCTs) at approximately 1,500–3,500 patients per year in the US+EU. Wall Street analysts put the price of BPX-501 at $125,000-$250,000 per year, imputing an addressable existing market opportunity in orphan diseases alone at $200M-$875M.
Approval could be as early as late 2017 in Europe and early 2018 in the US. Since there are many more HSCTs performed in the EU, more cases of orphan disorders and greater use of α/β T cell depleted haploHSCT, uptake in the EU should be relatively rapid and substantial.
Risks to Adoption in Existing OBD Market
- US Adoption May Be Slower: The data will drive the adoption and since HSCT for orphan diseases are fewer, we think that most institutions will be willing to test the procedure. Right now, awareness remains low and needs to be more widely disseminated among KOLs in the US to drive early adoption.
(2) Orphan Blood Disorders (market expansion) — Gene Therapy Is Probably Going to Have to Go Elsewhere
If you have read my posts before, you would know I am bearish on gene therapy for disorders that can be otherwise cured by alloHSCT (the only currently acknowledged cure). This is because BPX-501 + α/β T cell depleted haploHSCT appears to address the primary problem that gene therapy was designed to solve, namely to avoid GvHD and lethal infection from otherwise curative alloHSCT. There is no argument that this was the premise of gene therapy in these hemoglobinopathies, as pointed out on the web site of the leading gene therapy company for hemoglobinopathies:
Currently, the only curative treatment option for transfusion-dependent β-thalassemia is allogeneic hematopoietic stem cell transplant (HSCT, also called bone marrow transplant). The procedure works by transferring a donor’s stem cells into a patient’s body, where they produce a supply of new blood cells with functioning β-globin. However, there are risks associated with the transplant, including serious infection, graft failure and graft- versus-host-disease, some of which can be fatal, so transplants are primarily only offered to pediatric patients with matched sibling donors, which typically occurs in less than 25 percent of all patients.
HSCTs have always produced real cures and the Bellicum efficacy data are unequivocally superior to gene therapy. The example of beta thalassemia (Thalassemia Major, or β0/β0) is perhaps the most visible example of the evaporating need for gene therapy when alloHSCT is already curative but now much safer. Comparing the data below from the JPMorgan Conference corporate presentations of BLCM and BLUE make this difference clear:
By day 90, all BPX-501 patients are near above 9Hb (g/dL) and were transfusion independent post-transplant. Compare this to the data from bluebird’s presentation below, where the median at 90 days was only 3.8 Hb (g/dL) and the trajectory of increase is slow and in some cases faltering altogether. Most remain transfusion dependent.
Notably, the scales of these graphs are different and only 5 of the 13 patients in bluebird’s slide would even show up on the BPX-501 graph above. Only one Lentiglobin patient reaches the 9 HbAT87Q (g/dL) mark that is comparable to the lowest of all of the patients in the BPX-501 slide [9Hb (g/dL) is bluebird’s target for Lentiglobin.] That Lentiglobin patient, however, takes 6 months to reach this level whereas all BPX-501 patients reach it at 90 days or before — all but one does by day 30.
The market for Lentiglobin that created such excitement for bluebird could legitimately be ascribed to the more promising BPX-501 + α/β T cell depleted haploHSCT. It does essentially everything Lentoglobin was designed to do, and investors hoped it would do. BPX-501 should drive significant expansion of the use of HSCTs to these often very sick patients who currently do not seek a cure by alloHSCT.
Risks to Orphan Market Expansion
- Marketing & HSCT Risk Perception in the US: The main risk to expanding to populations like the sickle cell population in the US is getting the word out to patient communities that an acceptable risk/reward is now offered to them. Beta thalassemia in Europe, where it is more prevalent, is an easier market. There is large pent up demand for a cure. Sickle cell disease in the US is a patient population that is under-treated and less well managed. Driving adoption in this population is going to take time, partly due to the fact that alloHSCT outcomes were so poor that treating physicians don’t offer it to sickle cell patients. Historically, trading one chronic disease (sickle cell) for high transplant related mortality risk (15%) and potentially another, often worse chronic disease in cGvHD (20%+) meant that a patient had a 30–35% chance of a worse outcome from an alloHSCT than living with their sickle cell disease. If the data continue to be as positive on these measures, BPX-501 should address the market that bluebird’s gene therapies was once thought to address — but physician and patient education will be the challenge.
(3) Hematologic Malignancies
All research analysts have said that the “big” opportunity for BPX-501 is in hematologic malignancies. Indeed this is a far larger population than the existing (not expanded) market in orphan disorders.
Most analysts are cautious on the data regarding relapse rates — as they likely feel that this will determine adoption. This is partly true. More likely, all cause mortality will be critical when evaluating BPX-501 + α/β T cell depleted haploHSCT — and it already has an advantage in TRM. If BPX-501 + α/β T cell depleted haploHSCT is no different on relapse rates, it will be seen as better overall by virtue of the reduced infection and GvHD rates.
So what reason do we have to think that BPX-501 + α/β T cell depleted haploHSCT will be at least as good in terms of relapse rates? Once again, we can mine the data of Dr. Bertaina from ASH, which shows that the underlying procedure, α/β T cell depleted haploHSCT without BPX-501, is better than MUD or MRD.
There is no reason to think that adding back T cells will make relapse rates worse; conventional wisdom suggests it might make them better. A careful look shows that at one year, the relapse rates for α/β T cell depleted haploHSCT appear worse than MUD and MRD but then becomes equal or superior at 2.5 years. This underscores that it is crucial to pick the right time frame to evaluate relapse rates, as we will discuss later.
Either way, the point is clear. Over time, relapse rates are better than MRD and equivalent to MUDs. If the hurdle is equivalence — BPX-501 + α/β T cell depleted haploHSCT should be just as good or better than α/β T cell depleted haploHSCT alone.
In summary, with the addition of BPX-501 T cells, which may have additional ant-tumor effects, patients on BPX-501 should perform sufficiently in terms of relapse rates. This would facilitate wide adoption in the larger malignant population undergoing alloHSCT due to a reduction in all cause mortality. Recall above where α/β T cell depleted haploHSCT scores higher than MRD or MUD on 3 year probability of event free survival. BPX-501 should improve this graph based on lowered TRM alone, and potentially improved relapse related mortality.
Risks to Adoption in Malignancy
- What is the Measure of Success? All Cause Mortality? The primary measure to drive adoption is likely to be all-cause mortality. If BPX-501 + α/β T cell depleted haploHSCT is superior on this measure, it will gain adoption. It has a big head start with materially lower TRM than other procedures.
- When and How to Assess Relapse Rates? Unlike Wall Street research analysts, who place significant risk on the actual relapse rates, one must think they did not focus on the results from Bertaina of the underlying procedure, which are quite comforting. However, there are two issues with relapse rates. The first is assessing what it is the correct measure of success. Do we compare BPX-501 to MRD, MUD or other alloHSCTs? Arguably, relapse rates should not be evaluated alone but in the context of a probability of event free survival, which is the ultimate measure of success. Second, the time-frame to evaluate relapse or event free survival needs to also be chosen appropriately. α/β T cell depleted haploHSCT without BPX-501 had relapse rates in the Bertaina presentation that were better than MRD and similar to MUD at 3 years. But it looked slightly worse early on. Waiting three years is a long time but using 1 year or even 1.5 years to benchmark relapse rates (or any measure that incorporates them) might be a mistake, leading us to believe (falsely) worse outcomes.
Near Term BPX-501 Milestones
BPX-501 could be commercial in Europe in late 2017. Things to watch with respect to BPX-501 include:
April 3–6, 2016 — Update on BP-004 Trial: Data update at EBMT. This should give longer term follow-up on the BPX-501 data on orphan blood disorders that were presented at ASH along with some additional patients and would comprise most of the data set to be discussed with regulators in 2Q:2016. We will be looking for infection rates / TRM, Acute GvHD & Chronic GvHD rates. Relapse rates in malignancies still have limited utility, as discussed.
2Q:2016 — Meet with FDA and EMEA. Bellicum will meet with regulators 2Q:2016 to discuss pivotal studies and approval strategies.
- EMEA — Potential to file for approval in orphan blood disorders with commercial potential in H2:2017
- FDA — Likely agreement on Phase III trial protocol, if the trial is necessary for approval
H2:2016 — Commence Two Phase III Trials in BPX-501. Commencement of pivotal trials in Orphan Blood Disorders and another in Hematologic Malignancies (H2:16).
Disclosure: The fund I work for, Aju IB Investment, was a private investor in Bellicum before its IPO and invested further in the IPO. The firm continues to hold all of the shares it has purchased in Bellicum. I wrote this article myself and it expresses my own opinions. It does not reflect the views of my employer. I am not receiving any compensation for this blog post. I have no business relationship with any company whose stock is mentioned in this article. This is not a recommendation to buy any security.