Skip to main content

Regulatory Affairs

Not Approved Until It’s Approved

For some years now, Biomarin has been working on a gene therapy for hemophilia A. That’s the form of the disease caused by a deficiency in Factor VIII, which is a necessary part of the blood-clotting cascade. Like many such conditions, it comes in a wide range of phenotypes. One group of patients (around 10%) makes a nonfunctional version of the protein, while the majority just make insufficient amounts of the functional one (all the way down to making none of it at all). The majority of cases can be traced back to inheritance within a family, but about 30% of new cases seem to be truly new idiopathic mutations.

If the disease is not too severe, it can be treated with the vasopressin analog desmopressin, which releases stored Factor VIII and can take patients back into a more normal clotting time. Otherwise, the main thing to do is to provide Factor VIII protein itself, either recombinant or isolated from human plasma, with the amounts and dose schedule tailored to the patient’s level of remaining function. Biochemically this works fine, but the practical difficulties are (as you would figure) around the repeated i.v. administration. A permanent venous port can make this easier, but these can be prone to infection. So although this is not an untreatable disease, better options would be welcome.

In 2017, a trial funded by Biomarin reported results of a AAV5 viral-vector gene therapy approach to splice in a functional Factor VIII gene into nine volunteers. Two lower-dose patients did not appear to respond, but 6 out of the 7 higher-dose patients showed definitely improved Factor VIII levels that persisted out to at least a year, with a corresponding sharp drop in bleeding incidents. All of the responding patients were able to discontinue dosing with extra Factor VIII itself. These results were certainly encouraging, and got a good amount of press at the time.

The company has continued to monitor these patients, and the most recent data on them has good news and bad news. The good news is that six of the seven patients still showed no bleeding incidents in the fourth year after the single dose, and all of them remained off prophylactic doses of Factor VIII. The bad news was that the effects of the gene therapy seemed to be wearing off: the amount of circulating Factor VIII had already gone down by year 3 from what was seen in the first year, and the drop in year 4 was even greater. So while the patients were (mostly) still at useful levels, it seemed likely that they would soon be dropping back into some level of hemophilia.

And that is what the FDA now seems to be worried about. This morning, the company announced that they had received a Complete Response Letter from the agency, which many readers will know means, roughly, “We Are Completely Not Approving Your Drug Right Now”. The FDA wants to see two more years of data to see how far the Factor VIII levels drop, a primary endpoint of “annualized bleeding rate”, and safety monitoring as well. You can tell from the Biomarin announcement of this that they’re pretty upset with the news: “The Agency first informed the Company of this recommendation in the CRL having not raised this at any time during development or review

Well, yeah – but the agency found out at the same time as Biomarin that the levels had dropped even more in year 4, so it’s not like they were going to recommend this based on the earlier data. And the company must have (should have) been thinking about the possibility that they would ask for further monitoring. Biomarin’s investors don’t seem to have been thinking about it, though – BRMN stock is down about 36% as I write.

This decision can be argued about, although the FDA is going to win that argument in the end. Their point is that they’d like to know what they’re approving: is this a cure, a four-or-five-year abatement in hemophilia, or what? It would seem that the theoretical option of getting re-gene-therapied is a long shot, since you would expect these patients to have primed an immune response to the AAV5 vector. More realistically, how do these patients do once they have to go back to taking extraneous Factor VIII – are they back to status quo ante, or does what might be a fairly brief excursion into health end up hurting them more in the long run? We really don’t know until we see the data, and the FDA is saying that they want to see it. The belief seems to be that the disease already has direct treatment options without gene therapy, so a completely new mode of therapy like this one needs to prove its worth before it gets approved. Not everyone is going to be happy about this, but the agency does have a case.

In the end, it might still be worth it for people to have a few years off of Factor VIII treatment (although that’s going to be a call for insurance companies to make, for the most part, isn’t it?) Or it might turn out to be a bad idea. We’ll revisit the topic in two years, by which time everyone involved will really be ready to argue.

32 comments on “Not Approved Until It’s Approved”

  1. David E. Young, MD says:

    Sounds like something Yogi Berra would say “It ain’t over ’till it’s over.”

  2. Eugene says:

    This does not strike me as any different than other approved medications that lose effectiveness over a number of years or side effects start cropping up as a patient changes (ages). Chemo-therapy comes to mind. I would be very interested in the underlaying mechanism for the decline. Is there a limited population of secreting cells that age and die (Senescence?) or is there an immune effect that attacks these cells.

    1. Giannis says:

      Also people on monoclonals/biologics. A friend of mine was on one anti-TNF biologic and after some time it just stopped working. She most probably had developed antibodies against the drug. After taking another anti-TNF monoclonal antibody, everything went back to “normal”. This is a huge problem for diseases where FDA has approved only a single agent and patents are a huge issue that nobody discusses. Another company cannot just create another antibody against the same target, it’s covered by patents!!!

    2. insilicoconsulting says:

      Exactly. Approve but put a timeline of 3 years!

  3. Mister B. says:

    Would it be possible to give a booster / reinject the patient after 5 to 7 years ?

    For a chronic disease to a once-every-5-years injection, that is quite an improvment for the patients, isn’t it ?

    1. KazooChemist says:

      Derek addressed that in the second to last paragraph. It is very likely that the patients would develop antibodies to the vector from the first dose and the immune system would mop the new dose before it could do its job.

      1. Giannis says:

        Ι highly doubt that the antibodies will last for four years. For example Oxford manages to get a boosting effect with their ChAdOx SARS-CoV-2 vaccine after a second dose with the same ChAd vector. If LLPCs aren’t generated, which they usually are not generated for “minor infections” , I doubt that the titres will remain high enough to stop the adenovirus from infecting the cells.

        Now, even if I am completely wrong and after 4-5 years there is sterilizing immunity, you can develop a process to remove the neutralizing antibodies from the patients blood. It’s not easy, but creating CAR-T cells is far far far harder.

        1. Charles H. says:

          This isn’t an infectious agent, though. Even a mild immunity might render re-treatment ineffective, as the agent doesn’t reproduce.

      2. Steve Scott says:

        And this also brings up the question of whether Covid-19 adenovirus vaccines would be effective if a booster is needed after a year.

      3. Mister B. says:

        Thanks for all the answers, even I poorly asked my question.

        It was, indeed, more like ” what would happend if patient are reinjected every 4 to 5 years ?”.
        Giannis gave a perfect answer.

        Possibly, they’ll make a try with this first cohort.

        Thanks again !

    2. Aleksei Besogonov says:

      The fear is that patients might have a primed immune response to viruses used, so the immune system will either just kill off the viruses immediately using antibodies in the blood or interfere with “infected” cells.

      People are trying to work around both issues, but for now it’s assumed that a lot of gene therapies can only be used once.

    3. Brian Hanley, PhD says:

      Yes. This can be done, even with an AAV vector. There are multiple ways to deal with this.

    4. Thoryke says:

      As I recall, the pricing for this treatment was based on the “How much is it worth to patients and society if people are cured of Hemophilia A by a single, one-dose treatment..?” idea. Biomarin was able to justify a very high price on the basis of an incredible benefit. If, in fact, the real scenario is that patients will need ~20 of these treatments, maybe the costs could be split out? But if it also turns out Biomarin has to make 20 different vectors in order to manage the immune issues for successive presentations of the gene…… lots of calculations will need re-doing…

  4. Barry says:

    The various hemophilias (defects anywhere along the clotting cascade) have always been attractive targets for gene therapy; the missing protein just has to make it to the plasma compartment, rather than to a chosen cellular target. But if you’re transforming fully-differentiated liver cells (adenovirus likes hepatocytes) rather than their progenitors, the effects would not be expected to last longer than those transformed hepatocytes

  5. MFruchtman says:

    Reminds me of the Oakland biohacker who dosed himself with a virus edited to only contain the gene for lactase to solve their lactose intolerance. They got a nasty immune response as one would expect, but it worked. Or at least a year and half when it wore off as well. And then he took the crazy step of reinfecting himself with the same virus again.

    The resulting immune response could best be described as atrociously unpleasant and completely immune to the virus. Needless to say, the lactose tolerance did not come back and further attempts with the same method would be extremely dangerous.

    For viral gene therapy, viral immunity is going to be a problem if the treatment wears off.

    1. confused says:

      Do you have a link to read more about this? Wow…

      1. a says:

        How about george church taking homemade vaccine from one of his protoges?

    2. Charles H. says:

      Hmmmm…..and corona viruses tend to generate an immunity that quickly wears off. So what’s needed is a gene treatment based around a modified corona virus…

      1. Barry says:

        The novel coronavirus is novel (duuh!) but we do know that immunity to SARS and to MERS persists for years. It is reasonable to expect that immunity to this newest coronavirus will be like those earlier versions

  6. Brian Hanley, PhD says:

    Hemophilia AAV5 vector – An open letter.
    Of course factor level’s going to drop. The surprise is when it doesn’t. There’s multiple ways to deal with the issues with viral vectors. First, there’s a patent (not mine, and not terribly new) for including a stoichiometric amount of viral vector capsid proteins (not assembled into capsids) sufficient to soak up low levels of antibodies, either hours prior to, or together with the vector. It’s also quite practical to use immunosuppresants short term to prevent antibody production in the first place, for the future. One could consider also apheresis to remove specific antibodies. And, in many patients, 4 years after a small viral stimulus, there won’t be much antibody present anyway.

    I expect you are already pursuing these options. I just want to make sure that you guys get through this issue. It’s a good product and should succeed. Business-wise, this is a good thing for gene therapy in general. It’s not the pure surgical model. Much as I like the surgical model and try to make it work, if it doesn’t that’s a positive for the business.
    I’d also suggest rewriting your protocol to include multiple dosing to start with. Think about it. Short term immunosuppression and a lower first dose. Do a titration strategy. Price each dose lower, but expect to administer 2 or 3 for average patients, and 4 or 6 for resistant ones. The resistant patients will most likely have cross-reactive antibodies of some kind. (IgG, IgE) So, add to your protocol a vector cross-reactivity test step. Then you can handle every injection the same way. Optional immunosuppressant, depending on judgment of physician. Apheresis or a capsid protein cocktail injection to knock down antibody levels. And then inject, wait 6 weeks, test, and decide if a booster is needed.

    It’s difficult to produce a one-size fits all for gene therapy. Everyone is a bit different, cross-reactive antibodies are a problem, body mass is a factor, and how the nucleus downregulates these genes varies. I’d suggest accepting that, and you’ll have a better product as far as the patients are concerned.

    1. Barry says:

      I must be missing something here. Giving “immunosuppresants short term to prevent antibody production” along with the booster would defeat the point of vaccination entirely. And although one could swamp the host vs. vector Ab response with viral vector capsid fragments, that would be indistinguishable from giving a booster vaccination for exactly the host vs. vector immunity that we don’t want.

      1. Marko says:

        “….would defeat the point of vaccination entirely.”

        It’s not a vaccination. The idea is to deliver the gene therapy agent while bypassing any immune response against the vector used. Once the payload has been delivered , immunity isn’t an issue unless further treatments are required. Then you may need to use immunosuppression , or a different vector altogether , to get the job done again.

  7. Maurice says:

    I appreciate your blog posting which are very well thought out and written. Now the disagreement part 🙂

    The durability issue while certainly a concern is likely NOT the full story in the CRL. The mention of differences between the phase 2 data, which used research grade material and the phase 3, which used commercial material was notable. The factor levels generated from Phase 3 were significantly lower than the phase 2! Biomarin was repeatedly asked about that and said they could not find any differences in any step of the process to account for that. While prophylactic steroids use was later suggested as having some impact it would certainly not account for the majority of the difference. This would seem the likely central point of the CRL. If the materials are not comparable (in FDA opinion) then that would likely seem reason to want more data with the commercial material before approving.

    Why 2 years of durability vs. 1 year, if one already know its declining by month ~9 or so? Why not require 3, 4, 5 years? FDA does not require small molecules or biologics to show durability years out. And some patients have been known to stop responding (development of inhibitors, antibodies, etc.). Could it be that FDA wants some parameters to quantify the decline? This to me is an important question that FDA clarity on would help are gene therapies going to be required to show X years of durability? Does it depend on if other treatments are available? How fast the decline is? Why wouldn’t it just be a label issue?

    1. matt says:

      ” Could it be that FDA wants some parameters to quantify the decline? This to me is an important question that FDA clarity on would help are gene therapies going to be required to show X years of durability? Does it depend on if other treatments are available? How fast the decline is? Why wouldn’t it just be a label issue?”

      I don’t speak for the FDA of course, but if you are trying to get a novel approach approved, isn’t it up to you to solve some of these issues, or at least explain how they will be addressed?

      If it wears off, shouldn’t the company provide some idea of how long it lasts so patients and doctors can assess whether it is right for them? If they plan to re-dose, shouldn’t the company have to provide some data about whether the immune response prevents that, or methods (as outlined above) to get around that and evidence of how well they work? Isn’t this intimately connected to the efficacy of the treatment? The reason gene therapies need to show X years of durability is because they are going to sell it as a permanent fix, right? If it is not, and if you can’t re-take, that seems like an important detail to nail down.

      And if you haven’t quantified any of this, what are you going to put on the label? And how is that different from “we don’t really know how well this works yet”?

      The pioneers get the arrows. I’m sure in a little while, there will be all sorts of counter-measures, and a company will just be able to say, “we plan to handle this in the standard way, using procedure XYZ,” and the FDA will know how that works, and doctors will know how that plays out, and it will be fine. But for now, if you are the first, you get to set the ropes for the Hillary Step yourself.

  8. idiotraptor says:

    Do these results potentially bode ill for AA5 specifically and AAV in general as gene therapy vectors?

  9. matt says:

    Are there companies out there offering custom-generated viral vector platforms? The non-SARS coronaviruses seem to dodge long-term immune response, is someone offering a scooped-out one of those as a vector? How about influenza–it seems to have an infinite ability to generate new versions that are not intercepted or are very poorly intercepted by the immune system. Why can’t we generate some of our own? It’s not like pharmaceutical chemists don’t know how to vary parameters to reduce binding affinities.

    If something lasted 5 years, then at most 20 variations of viral vector would be a permanent fix, right? Every five years, you move production to the next vector in a series, and after 100 years you’d be back to the first. But the problem would have been solved way before then.

    If neurotransmitter lipoids can get you across the blood-brain barrier, could steroid lipoids deliver something into the nucleus of cells?

  10. Lapsed Chemist says:

    Would the payers be willing to fund a treatment which cost 2-3 million dollars ( and may last 2-3 years, or would they prefer to pay for the standard treatment of care?

    Given the extremely high costs of any Gene Therapy Treatment, patients may have a single shot, so would you take a risk, or wait out for a candidate which may overcome some of these challenges? The FDA may have had one eye on other treatments “in the pipeline” which may overcome some of the limitations listed

    1. Lambchops says:

      Given the value proposition tied to that price tag is being a “one and done” type therapy it would be a struggle to demonstrate cost-effectiveness when there’s evidence of waning.

      Simplistic back of the envelope analysis ahead! A quick google suggests haemophilia treatment costs $300-500k a year. Given there’s no data beyond 3 years and evidence of waning a conservative assumption of no treatment effect at 5 years is a reasonable scenario to consider given the uncertainty. Assuming standard care is 400k then total cost for 5 years of standard care is 2 million dollars. So even at the lower end of the price range for the gene therapy it’s struggling to look like good value.

      Would likely require some heavy discounting, or some sort of outcomes based pricing approach. Even then could be an uphill battle.

      Obviously this is putting aside long term safety concerns and the quite valid point that from a clinical standpoint why go for a new treatment modality that only lasts for a few years when there are other options available? Without some way of identifying in advance a patient population who are likely to benefit for longer might not be worth the risk for many. If this is ever approved that would be a decision for the patient/clinician to make.

      Seems a sensible call from the FDA to ask for more data, here’s hoping they take a similar stance with aducanumab in AD.

  11. not Oscar Ratnoff says:

    The severity of classic Hemophilia A correlates well with the baseline Factor VIII activity of the patient. At 20% of normal, life is fairly benign, but at 5% there are lots of problems. Hemophilia strikes men, and men, especially when young, do not always behave rationally. I cared for one who chose motorcycle racing as a career. One was shot by an angry husband. Another was a musician who got in a fight when someone in the audience seemed too attentive to his girlfriend. A treatment that works for five years, say age 18 to 23, would greatly reduce problems.

    1. Marko says:

      A low-cost option for that risky male group would be chemical castration.

  12. idiotraptor says:

    I was alerted to a relevant point of biology about hemophilia A when discussing the topic of this post with my hematologist wife. Hemophilia A Factor VIII deficiency is a monogenic disease. It is heterogeneous in manifestation and may present in patients as complete absence of Factor VIII or as a truncated or mutated protein. Hemophilia A patients apparently often will make inhibitory antibodies to native Factor VIII administered therapeutically. Presumably this is the result of absent or incomplete negative selection during T and B cell ontogeny. AAV5 directed immunity may not be be a principal cause in the apparent drop in Factor VIII levels.

  13. LÉO CUENCA says:

    I think in 3-4 years (which this treatment gives) scientists will find a way out.
    It’s important not to stop studying this disease.

Leave a Reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload CAPTCHA.

This site uses Akismet to reduce spam. Learn how your comment data is processed.