Atomic structure of an ICSM18 antibody fragment (yellow) binding prion protein (blue). PDB 2W9E [Antonyuk 2009].

Since late last year, the MRC Prion Unit in London has been giving injections of the PRN100 antibody to a small number of patients diagnosed with prion disease. Several families affected by prion disease have emailed me to ask about this, and I have been remiss in not updating the blog — my last post on PRN100 was in 2013, and I do strive to maintain a complete view of all candidate therapeutics being tested clinically. So today’s post will review the background on PRN100, cover the status of human treatments and the legal and regulatory framework under which they are being conducted, and conclude with implications for patients and families eager to learn anything they can about experimental therapeutics.

An important disclaimer is that this blog post reflects what I’ve been able to learn as a third party, and contains some of my own interpretations and speculations. I am not affiliated with MRC Prion Unit and nothing reflects the official view of the MRC Prion Unit. For the official word on these treamtents, please see the announcement on the UCL website.

background on the PRN100 antibody

Interest in antibody-based treatments for prion disease goes back at least a couple of decades. Concurrent studies from two different research groups showed that prion-infected cultured cells could be cured of infectivity if treated with monoclonal antibodies to the normal form of PrP [Enari 2001, Peretz 2001]. Both groups speculated that this was because the antibodies, by grabbing tightly onto normal PrP, blocked the misfolded form of PrP from binding and converting it. At the time, all of the other candidate therapeutics that people were studying for prion disease, such as quinacrine, were existing drugs whose mechanism of action against prions was either indirect or unclear. Antibodies seemed to offer a more rational, targeted approach against the very protein that causes the disease, but with the tradeoff that antibodies are large molecules that are difficult to get into the brain or to distribute across the brain. A news article at the time, foreshadowing perhaps the greatest challenge with this approach, noted that “The hope is that antibodies could eventually be developed into a vaccine, or even a treatment, if they can be shown to cross from the blood to the brain.”

An article later that same year showed that engineering transgenic mice to express, in their blood, an antibody against PrP, made them resistant to infection by prions injected peritoneally (into the body cavity) [Heppner 2001], although the study did not address whether this approach could be effective against prion disease originating in the brain, nor whether treatment of normal, non-genetically engineered, animals with an antibody could be effective.

Two years later, it was shown that treatment with a monoclonal antibody against PrP worked in vivo — in certain contexts [White 2003]. When mice were infected peripherally (i.e. outside of the brain) with prions, and then treated peripherally with an intravenous administration of antibody before the infection had reached the brain, they were completely cured — a stunning result. But if the mice were infected directly into the brain, or if they were infected peripherally but then the antibody treatment was delayed until after the infection had reached the brain, then the antibody did not bring about any detectable benefit. It’s not clear why. One possible explanation is the oft-cited statistic that only ~0.1% of antibody injected into the periphery reaches the brain [Poduslo 1994]. But at the time, the variant CJD epidemic in the U.K. was still near its zenith, and the fear of widespread peripherally acquired prion disease in humans was still prominent in people’s minds, so it seemed that a treatment effective against peripheral prion infection might be critically valuable. The antibody from that study, ICSM18, originally derived from mice, was “humanized”, with the new human antibody being named PRN100 and put into development for eventual clinical use.

In the years since then, there have been fewer and fewer peripherally acquired cases of prion disease in humans, and interest has shifted towards treatment of brain-based prion disease. One study examined whether infusion of antibodies directly into the brain of mice could be effective [Song 2008], and nominally, they obtained positive results, but the reported differences in survival time were all very small (<9%) and of marginal statistical significance (P < 0.05). No other study I am aware of has tried to replicate them. To this day, there is no clear evidence that antibody-based treatments can be effective against prions once the disease has reached the brain.

But developments in other neurodegenerative diseases continue to keep interest in this approach alive. For example, Biogen’s Aβ antibody aducanumab, which was discontinued earlier this year, failed to show a clinical benefit to Alzheimer’s patients in clinical trials, but it did show evidence of engaging its target — binding and clearing Aβ plaques — in the human brain [Sevigny 2016]. The fact that this antibody seemed to have measurable activity in the brain after peripheral dosing raised many people’s hopes that maybe the 0.1% of antibody that crosses the blood-brain barrier could be enough to make a meaningful difference in some diseases, even if that particular antibody for Alzheimer’s disease didn’t work out. Meanwhile, lots of scientists are researching potential ways to deliver greater quantities of antibodies into the brain, such as the so-called “brain shuttle”, though to my knowledge there has not been clear clinical demonstration of the utility of any of these approaches just yet.

In the intervening years, a debate has also raged in the prion field over what types of antibodies to PrP could be safely used as therapeutics. Following a report that certain PrP antibodies can cause neurons to die [Solforosi 2004], it was discovered that antibodies to different regions on the surface of PrP can have very different effects [Sonati 2013], with some apparently toxic and others protective. One mouse study found that 2 μg of ICSM18 injected directly into brain tissue was tolerated [Klohn 2012], while another observed toxicity at a slightly higher dose, 6μg, again injected focally into a small region (about 5 mm3) of brain tissue [Reimann & Sonati 2016]. To put these numbers in context, in the aducanumab mouse studies done in preparation for Alzheimer’s disease trials, the highest antibody concentration reached in the brain was just over 1 μg per g of brain tissue [Sevigny 2016]. Assuming that the density of brain is about 1 g/cm3 [Barber 1970], similar to water, this corresponds to about 0.001 μg/mm3 in the brain, whereas the toxicity for ICSM18 was observed at about 1 μg/mm3. It is scientifically interesting that this toxicity for some PrP antibodies appears to be on-target — mediated by binding PrP — but there is not yet clear evidence that this poses a risk at concentrations reached in routine dosing of animals or humans.

status of PRN100 human treatments

After years of developing and testing the PRN100 antibody in preparation for human trials, the MRC Prion Unit and its commercial spinout company, D-Gen Ltd., obtained funding to produce a single clinical-grade batch of PRN100 with the goal of testing it in humans. Subsequently, however, they were unable to obtain funding to actually run the clinical trial to administer this batch of drug to patients. When I last blogged about PRN100 in 2013, they had already begun to think about several details of how a clinical trial could be designed and run, but there was still no firm date for when it could begin. Eventually, the patient community in the U.K. raised enough funds to allow some patients to be treated with the existing supply of antibody.

The first treatment was given in October 2018 [Dyer 2018], with announcements that a second, third, and fourth patient began treatment in the following few months. To my knowledge, all patients so far have been symptomatic with sporadic CJD, and the treatments have been given by intravenous injection.

At the Prion2019 conference, Dr. John Collinge provided a brief update on PRN100 to colleagues attending the CJD International Support Alliance meeting. He said that 5 patients have been treated so far, and that they have enough drug to treat ~8-10 patients total. The decision to treat each patient is being made on a case-by-case basis and only U.K. residents covered by the National Health Service are eligible. Collinge expects that the existing batch of the antibody will be used up by the end of 2019, and at that time, they would need to find more funding if they were to produce more of the antibody and pursue further clinical development.

The PRN100 treatments ended up being done under a particular legal framework that exists in the U.K. known as a “Specials exemption”, rather than as a formal clinical trial. While this distinction sounds boring, it is worth understanding, because it very much frames what we can expect to learn and how we can expect development of PRN100 to proceed.

A typical clinical trial is designed as a research experiment — patients are recruited explicitly for research, they might be assigned to different doses or treatment groups, and lots of monitoring and tests are done to systematically assess the safety and, ultimately, the efficacy of the drug. Running these sorts of clinical trials is very expensive, it usually requires prior approval from regulators (such as the FDA in the U.S., or the MHRA in the U.K.), and in many cases, patients are randomized so that not all of them get the drug. The data generated from a clinical trial can often help to support an application to get the drug approved so that it can be distributed to doctors around the country and made widely available to patients. Most countries have a regulatory framework in which these sorts of experiments to demonstrate safety and efficacy are required before a drug can be broadly administered to humans.

In contrast, treatments authorized under a “Specials exemption” are intended to be delivered as an emergency basis, compassionate use therapy, are not designed as experiments, and are not intended to generate data that would support a drug approval. The drugs involved may not yet have any human evidence of safety or efficacy, nor have any trials underway to evaluate safety or efficacy. The administration of the drug is undertaken as part of clinical care and is often less expensive than a clinical trial. We do not quite have a perfect analogue of a “Specials exemption” in the U.S. regulatory framework — the closest thing might be a treatment IND, but treatment INDs are usually undertaken only after at least one controlled trial has examined safety of a drug. Similarly, the 2018 Right To Try Act, and its equivalents in many states, allow U.S. patients to request access to an experimental drug only after Phase I trials to evaluate safety are completed. In contrast, in the U.K., a “Specials exemption” can actually permit a first-in-human treatment.

That said, the PRN100 treatment proposal did get reviewed by the U.K.’s medical product regulators (the MHRA), but ultimately, the authorization to proceed to human injections of the antibody came from a judge, who in the court proceedings explicitly stated that:

The MHRA has confirmed that the use of PRN100 under the Specials exemption does not constitute a clinical trial. It is important that I emphasise this. This is not a trial; this is not an experiment; it is the provision of a treatment or procedure to an individual patient who is in need.

This distinction is critically important because it changes what we can expect to learn from the PRN100 treatments going on in humans. For example, because PRN100 is to be administered as part of clinical care rather than as an experiment, any data on PRN100’s safety, efficacy, or pharamcological properties will come only from those blood draws, lumbar punctures, brain scans and so on that need to be undertaken for routine clinical care reasons. No such procedures can be performed specifically with the goal of evaluating how PRN100 is performing. Similarly, patients cannot be randomized to drug vs. placebo or to different dose levels or different dosing regimens.

This means that the current round of PRN100 treatments may only be able to tell us about any rather dramatic effects of the antibody. For example, if the few treated patients exhibit a miraculous clinical improvement, that would be so wildly outside the normal course of prion disease that we would be able to tell it was due to the antibody. But if the antibody only slowed their decline, we might or might not be able to discern that from the data. Likewise, if the antibody was so horribly toxic that it killed patients instantly, we would certainly know, but beyond that, we might or might not be able to distinguish side effects of the drug from the background impairment of prion disease.

But in my view, either extreme outcome — miraculous efficacy or deadly toxicity — is at this point fairly unlikely given the preclinical studies already undertaken in animals. If the antibody were so effective as to bring about a clinical improvement in symptomatic patients, then we probably would have already seen at least a whiff of efficacy when mice with brain prion infections were treated, whereas instead, the antibody seemed not to work in mice once the infection was in the brain [White 2003]. Likewise, if the antibody posed an imminent mortal danger to patients, that toxicity would have likely turned up in the mouse studies or in further preclinical toxicity experiments undertaken to date. Although the data have not been published, the court proceedings note that toxicology studies were performed on animals and that MRC Prion Unit did actually undertake the studies that would have been necessary to register a Phase I clinical trial with the MHRA. Such studies usually include non-human primate dosing in addition to rodents.

Therefore, when the current batch of drug is used up at the end of 2019, I am not expecting to see any conclusive evidence of whether the drug works. Instead, my guess is that the question will probably come down to whether the available data provide enough supportive evidence to motivate further fundraising and clinical development of the antibody. If the data show reasonable pharmacological properties, no obvious safety issues, and some hint of possible efficacy, then the MRC Prion Unit might decide it wants to continue developing the antibody, and it might be able to convince the U.K. government, investors, or donors to help fund a more formal clinical trial with the goal of eventually getting to a drug approval. If the data look relatively less promising, the antibody could be backburnered in favor of other therapeutic strategies. Again, these are just my expectations based on what I have been able to learn as an outsider, and does not reflect any official statement from the MRC Prion Unit.

implications for patients

Since the PRN100 treatments began last year, I’ve gotten emails from patients all over the world wanting to know will this treatment be effective, when will it be a drug, and how their loved one can gain access to this experimental treatment tomorrow. And I’m just a blogger and scientist on the other side of the world, so I can only imagine how many emails the MRC Prion Unit is getting about this subject. Patients’ expectations are often rather high, understandably given the desperation for any shred of hope that one feels when going through this disease. My main goal in writing this blog post is to help manage those expectations, so I want to conclude with two summary points.

First, the court order authorizing these treatments only extends to U.K. residents that are covered under their National Health Service. For patients anywhere else in the world, there is simply no way to volunteer to receive this antibody. Remember, this isn’t a clinical trial with enrollment criteria, it’s a small effort with each treatment being undertaken on a case-by-case basis.

Second, the treatments are not part of a clinical trial, and that means that they are unlikely to give us unmabiguous evidence regarding the safety and efficacy of the drug, and cannot directly lead to approval of a drug that would then be more broadly available. At most, the treatments might generate enough data to motivate a more formal clinical trial that could eventually lead to a drug approval.

Everyone who has been touched by prion disease is desperate for there to be an effective treatment. This is the cause to which I have now devoted my life, and I am optimistic about the progress we can bring about in our lifetimes. But biomedical research is a slow, iterative process — getting a candidate therapeutic to its first-in-human dose takes years, and even after that, the path to an effective therapy is not always a straight line.

Update 2019-10-11: UCL has announced that the 6th & final patient has begun receiving PRN100, and that 4 of the 6 patients treated so far have now died.