Update on new treatments for fragile X---Part 4: MMP-9 inhibitors

In 2007, Iryna and Doug Ethell of UC Riverside made the seminal finding of excessive activation of the enzyme Matrix Metalloproteinase-9 (MMP-9) in fragile X. They showed that excessive MMP-9 activity maintained dendritic spines in an immature state, and that this contributed to the behavioral abnormalities seen in the fmr1 KO mouse. Most importantly, they showed that the known inhibitor of MMP-9, the available antibiotic minocycline, could rescue all the abnormalities they found in the mouse model. This made minocycline an attractive off-the-shelf treatment for fragile X. Shortly thereafter, I started experimenting with the judicious off-label prescription of minocycline in a small number of carefully selected patients; the results were impressive, to say the least! Indeed, the effect of minocycline was too good, and too rapid to be solely the result of normalized developmental trajectory. Minocycline appears to have an early psychotropic effect which is independent of its fragile X-specific mechanism; this has been utilized in several studies of minocycline as a treatment for schizophrenia and obsessive-compulsive disorder (which leads to the intriguing possibility that minocycline could be an excellent treatment for autism spectrum disorders in general.) Keep in mind that, in the mice, treatment was a very long-term affair---the equivalent of years in human terms.

After conferring with Carlo Paribello, of the Fragile X Research Foundation of Canada, we quickly began organizing a clinical trial. This first trial was an open trial, designed to examine the pattern of responses in high and low dose minocycline treated subjects. The trial achieved excellent results, with substantial decreases in aberrant behaviors over 8 weeks of treatment. Long-term follow-up of these subjects was also done, and results of that phase of the trial are now being analyzed. Very few side effects were seen in either high or low dose groups, and the therapeutic effect seemed similar (though the number of subjects could not discriminate subtle differences.) This is not unexpected, since studies by surgeons have shown that typical antibiotic doses of minocycline (100 mg/d) inhibit MMP-9 levels by 50-75%. That is just about exactly what we want. One unexpected result from the study was that ¼ of all the subjects followed for a year had an increase in Anti-Nuclear Antibodies (ANA), a rather non-specific indicator of autoimmune reactions. None of these subjects had any actual symptoms, though serious autoimmune reactions to minocycline are seen (fortunately, only rarely.) We’re still not entirely sure what to make of this finding, as it’s not known just what percentage of people in the general population develop elevated ANA levels with chronic minocycline treatment. However, it could be a useful indication of impending autoimmune problems, and may be worth checking as a screening test. Interestingly, none of the patients that I treat directly have developed this kind of reaction to minocycline, though I have talked to a number of parents around the country who have encountered this problem.

A larger, placebo controlled trial is under way at the MIND Institute to follow up on these initial findings, and we look forward to these results. In the meantime, another bit of confirmatory evidence came from the lab of Kendal Broadie at Vanderbilt. His group has published results showing dramatic rescue of fragile X fruit fly abnormalities with minocycline---all the more impressive because the fly model is essentially a knockout of 3 genes (fmr1, fxr1, and fxr2.) This group also showed that genetic reduction of the fly equivalent of MMP-9 yielded similar results; this is significant, because minocycline does several different things, but MMP-9 inhibition does seem to be the active ingredient.

So, we have an excellent off-the-shelf treatment in the form of minocycline, a cheap generic medication with a long history and fairly benign safety profile. But some people, perhaps as many as ¼, have trouble taking it, and it’s not recommended for kids under 8 because of dental discoloration (which is not directly related to MMP-9 inhibition.) We at FRAXA are continuing to explore more specific alternatives to minocycline; there is interest in pharmaceutical companies in drugs which can inhibit MMP-9 specifically without some of the other effects of minocycline. However, this is a niche market, and only a few companies have active programs in this area, so it may be a while before we find anything better than minocycline in this area.

Update on new treatments for fragile X---Part 3: GSK3 inhibitors

After mGluR5 antagonists, the drug class with the most extensive validation as a disease-modifying treatment for fragile X is one called GSK3 inhibitors. Glycogen Synthase Kinase 3 is a ubiquitous enzyme present in a number of signaling pathways throughout the body (a problem which we will re-visit later.) GSK3 beta is the specific version which is excessively active in fragile X (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2707186/?tool=pubmed ), and studies in fragile X mouse, fly, zebrafish, and neural stem cells all show that reducing GSK3 activity, either genetically or with chemical inhibitors, can rescue a very wide range of fragile X phenotypes (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2838793/?tool=pubmed ). FRAXA invested heavily in funding the study of GSK3-based treatment strategies following the original finding that MAP1b was one of the critical overexpressed proteins in fragile X (http://www.ncbi.nlm.nih.gov/pubmed/11733059). Since it is known that activation of MAP1b occurs via GSK3, it was recognized early on that inhibition of GSK3 could be therapeutic for fragile X. Subsequent evidence has shown this to be correct, and perhaps even an underestimate. GSK3 inhibitors appear to have as much therapeutic potential as mGluR5 antagonists, though they may not be quite as well tolerated.

While virtually every major pharmaceutical company in the world is working on improved GSK3 inhibitors for a number of different indications, there is an available drug which is an excellent GSK3 inhibitor. This drug is lithium. Ironically, even though lithium has a reputation for being rather toxic, it may turn out to be less toxic than any of the newer, “cleaner” GSK3 inhibitors in development. Indeed, most of these development programs are in trouble because of toxicity and off-target effects. It is because of the ubiquitousness of GSK3 that these side effects may be unavoidable (in fact, in this light, lithium looks pretty good!), but there are efforts to develop brain-specific GSK3 inhibitors which leave the rest of the body untouched.

Millions of people around the world have been treated safely and effectively with lithium for psychiatric disorders (especially Bipolar Disorder, a.k.a. Manic-Depressive Illness.) However, lithium has fallen out of favor with psychiatrists who now have many other options for treating psychiatric disorders, especially in the form of heavily promoted, brand name drugs like Zyprexa and Depakote. Lithium is a cheap generic, and no one is promoting it. It is also hard to prescribe, and no doctors other than psychiatrists have any experience using it. But lithium does appear to be therapeutic for fragile X, showing disease-modifying properties in animal models (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3102293/?tool=pubmed and http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2810609/?tool=pubmed and http://www.ncbi.nlm.nih.gov/pubmed/21078304 and http://www.ncbi.nlm.nih.gov/pubmed/20705090, among others) and excellent results in a pilot clinical trial in fragile X patients ( http://www.ncbi.nlm.nih.gov/pubmed/18698192 ).

However, it is fair to say that lithium has not become a common treatment for fragile X, probably because it is widely perceived as a dangerous drug (whether this is correct or not, the perception certainly influences acceptance of the treatment by patients, families, and physicians.) Further trials are clearly warranted, however. A larger, controlled trial of lithium in fragile X subjects is a high priority for FRAXA. There is also the possibility of identifying available agents which can enhance the effectiveness of lithium (http://www.fraxa.org/researchTeam.aspx?id=659 ), and perhaps allow for lower dosing. This might eliminate the need for blood testing to determine lithium levels, one of the more unpleasant aspects of lithium treatment for patients and families. It is surprising to me that more psychiatrists are not prescribing lithium for their fragile X patients, especially in case where the psychiatric presentation alone (i.e. aggression or mood lability) might justify a trial of lithium. It may simply be that doctors today have so many choices, it’s easy to avoid lithium.

mGluR5 antagonists for fragile X

The disease-modifying potential of this class of drugs, which block type 5 metabotropic glutamate receptors, has been incredibly well validated in animal models of fragile X. It’s actually so well validated that it’s unprecedented in the history of medicine. There really has never been another example of a small molecule fixing so many facets of a neuropsychiatric disease in multiple animal models. In part, this is because most neuropsychiatric diseases haven’t had very good animal models, at least until recently, but this just reinforces the point that fragile X is such an important model for the study of many other disorders---we have animal models that aren’t just models, they actually have fragile X. It seems like we’ve been talking about the promise of this treatment strategy forever, but it was only 2001 when the “mGluR Theory” of fragile X was conceived, and now we have 3 companies with active programs to develop their mGluR5 antagonists for fragile X (as well as the now-defunct Neuropharm, which may yet rise from the ashes.)

Novartis is the furthest along, and their compound AFQ056 just entered Phase IIb/III trials for adults and adolescents with fragile X at a large number of sites around the world (see clinicaltrials.gov for more details.) Trials in children are expected to follow soon. AFQ056 is a modern, advanced compound which appears quite safe, and also appears to be an effective treatment for complications of Parkinson’s disease (see http://www.ncbi.nlm.nih.gov/pubmed/21484867 ). It seems likely it will make it to the market for Parkinson’s, even if things don’t work out for fragile X, which is a reassuring backup plan. However, if the current round of clinical trials is successful, it could be marketed for fragile X as the first indication for any mGluR5 antagonist. The Phase II trial results with AFQ056 for fragile X have been published ( http://www.ncbi.nlm.nih.gov/pubmed/21209411 ) and much has been made of the different response to the drug in full mutation males with “full methylation” vs. full mutation males with “partial methylation” on a proprietary assay of methylation status used in this trial. However, it is important to remember that this was a relatively brief trial in a small number of subjects, so it is not surprising that the best effect was seen in the relatively pure sample of fully methylated (ie non-mosaic, for all practical purposes) subjects. Subjects in this trial received the full dose of AFQ056 for only one week; such brief treatment often results in a high placebo response rate, and that was clearly the case here. Clinical trials in psychiatry have shown over and over that the placebo effect generally wears off over 3-4 weeks, and the current phase of AFQ056 trials will dose subjects for much longer. I would expect that this longer treatment would result in statistically significant improvement in all fragile X subject groups, with decreased placebo effect. In addition, preclinical testing of mGluR5 antagonists in animal models predicts that fragile X patients would require much higher doses of these drugs for optimal effects, and that these doses would be well tolerated, so this trial may not have pushed to dose into the optimal range.

However, this raises an important issue with all the disease-modifying treatments for fragile X. How long does it take to see genuine developmental improvement, rather than simple symptomatic improvement? It stands to reason that longer trials are required to see genuine developmental effects, and that even longer trials are required to see disease modification in older subjects. As I noted in the previous post, having extensive preclinical validation gives us confidence that a particular therapeutic strategy will be disease modifying, even if this is not practical to demonstrate in a clinical trial. It may take years to see real developmental effects in 20 year olds, although clinical trials are rarely more than a few months long. Besides, drug companies aren’t required to show that they can cure fragile X, only make it measurably better. So, I’m expecting that this phase of the trials will show improvement in all subjects on AFQ056, with more severely affected subjects showing the best effect, as well as the least placebo response. But, these things are hard to predict---that’s why they do the trials!

Another Swiss pharmaceutical giant, Hoffmann-LaRoche (or just “Roche” to those of us in the know) also has an advanced mGluR5 antagonist, RO4917523, which has just completed a Phase II trial in fragile X adults. RO4917523 may actually be the most (chemically) sophisticated of all the drugs in development; it is ultra-long-acting and very potent. A tiny dose once a day is all that is required to block mGluR5 quite effectively. Results of this trial are a closely guarded secret, and no announcements have been made about future plans for RO4917523, so we are left in the dark concerning Roche’s intentions.
Seaside Therapeutics also has an mGluR5 antagonist, STX107, which was licensed from Merck. It has reportedly completed Phase I trials (normal volunteers), but has yet to be administered to any fragile X subjects. Phase II trials have yet to be announced. Seaside appears to be devoting much of its energy to the development of arbaclofen (R-baclofen, or STX209) for autism and fragile X; this will be discussed in another post.

As noted above, Neuropharm, a UK-based startup, is now out of business. The company’s fortunes rested on the near-certainty of approval of a proprietary formulation of fluoxetine (aka Prozac) for autism. However, their large, multicenter trial failed to show superiority to placebo, even though everyone still uses fluoxetine and other SSRIs for autism spectrum disorders, because they clearly do work. This serves as an important cautionary tale: trials can fail for any number of reasons! Anyway, their Phase I/II trial of fenobam was quite successful; it was only intended to show safety and tolerability, but also showed significant improvement in pre-pulse inhibition (PPI) in fragile X adults, with many anecdotal reports of behavioral improvement from a single dose. Fenobam, like Neuropharm may yet be resurrected---many companies around the world are looking at possible uses for this off-patent compound.

Many other companies have their own mGluR5 antagonists, typically developed at vast expense, and now looking for some purpose. It seems to be generally agreed that fragile X and Parkinson’s are good, proven indications for these drugs, so we anticipate that more companies will move into fragile X trials, especially if Novartis continues to have success with AFQ056.

Update on new treatments for fragile X

In the coming weeks, I hope to provide up-to-the-minute updates (along with many frank opinions, since this is a blog!) about the status of all the various new treatments that have the possibility of disease modification in fragile X. I’ll do this in writing, since I do want to be fairly precise, and you may want to be able to copy these things; I’ll also reference much of what I say, except where the references are too numerous to list.

To do this, I’ll need to define a few terms, so let’s start there. First is the term “disease modification”, referenced above. We all are hoping for treatments which actually affect the course and the outcome of this single-gene disease we call fragile X. There are many available psychiatric drugs which can positively affect the behavioral manifestations of fragile X, and I’ve written plenty about them. But there is no indication that they change the long-term outcome in any significant way, as I noted in my last post. Obviously, the better kids do over the course of many years, the better their long-term prognosis, but the same can be said for any number of behavioral and educational interventions. There may very well be some kind of gray area here, but I think most of us can agree that we would like to see something much more specific, treatments which get to the very heart of the dysfunction in fragile X, and actually facilitate much more normal development. This is the mission of FRAXA, and it has become my life’s work, so I don’t consider these trivial distinctions. You’ll have to pardon me if I have some strong opinions in this area, and if I sometimes seem dismissive of ideas which are proposed that just don’t have the research backing to make them potentially “disease modifying” by accepted definitions.

So, how does one demonstrate that a particular treatment strategy might be disease modifying? Well, you could just give the treatment to patients and see what happens, based on some speculative theory of what’s wrong in the fragile X brain. This is the dominant model in autism studies today, and it carries over to some extent in fragile X. It is also quite common in psychiatry, and I’ve had the chance to observe the failure of this model on many occasions. For example, millions of children, adolescents, and adults take stimulant medications every day to enhance their attention (as well as decrease hyperactivity and other disruptive behaviors associated with ADHD.) Most parents hope, and assume, that this treatment will result in improved academic performance over the long term (we certainly see this in fragile X, too.) The drugs certainly work well in the short term, and efficacy is easy to demonstrate. However, most of the studies which have examined the long-term academic performance of kids with ADHD treated with stimulants show little, if any, advantage for the drugs. More recent studies have shown some benefits, but the results are debatable and more subtle than you might imagine (see http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2629512/?tool=pubmed for a more thorough review.) In the fragile X field, we just don’t have the time or money to try this willy-nilly approach, and the number of research subjects is too small to make it work.

A better way, I think, is to develop treatments based on the observed abnormalities in the animal models of fragile X (mainly fruit fly and mouse), then test potential treatments in those animal models. This initial observation would be considered basic science; the attempt to find ways to reverse a defect might be termed “translational research”, and the actual testing of a specific drug in the animal model would be called “pre-clinical validation”, leading to clinical trials in people with fragile X. The “FRAXA Method”, as I like to call it, is to fund this kind of progression from basic research, through translational research, followed by preclinical validation, leading to clinical trials of medications that have genuine potential to change the outcome of fragile X. We didn’t invent this method, and it really isn’t anything new at all---it’s what every pharmaceutical company does every day. But you’d think it was some sort of radical new approach, compared to the half-baked and poorly reasoned trials coming from many areas (not just the autism field, either; many clinical trials in psychiatry, or in the broader rare disease field are little more than wishful thinking, with no real scientific basis.)

In the next few installments, I’ll talk about the results of this approach, then move on to discuss some other developments in clinical trials for fragile X. Stay tuned!