30 July 2009. The old drug methylene blue generated instant curiosity among AD researchers when it burst on the scene as RemberTM with Phase 2 trial results at last year's International Conference on Alzheimer's Disease (ICAD) in Chicago, and was subsequently featured on CNN and other national media (for background, see ARF related news story). Many scientists took it as the first hopeful signal that—maybe, just maybe—inhibiting tau aggregation might actually work in people with Alzheimer disease. So what has happened with methylene blue research in the year since then? This story presents a reporter's update from this year's ICAD conference, held 11-16 July in Vienna, Austria, together with an extended Q&A with the developer of RemberTM, Claude Wischik of the University of Aberdeen, U.K., and the biotech company TauRx. Alongside this update, the Alzforum is pleased to bring to the table an outside perspective from an expert on the biochemistry and pharmacology of methylene blue, and its use as a treatment for other indications. Heiner Schirmer is professor at the biochemistry center at Heidelberg University in Germany. Besides having coauthored a leading textbook on protein structure, Schirmer has published extensively on the structure and biochemistry of parasitic flavoenzymes, some of which bind methylene blue. Schirmer is presently in Burkina Faso. There he is working on a field study sponsored by the German government on using methylene blue as an inexpensive and readily available treatment for malaria in rural Africa (for a recent report, see Meissner et al., 2006). The Alzforum editors would like to thank Schirmer, as well as Eckhard Mandelkow at the Max-Planck Unit for Structural Molecular Biology in Hamburg, who invited Schirmer to brief Alzheimer disease researchers on methylene blue's 120-year history as a drug. See Schirmer essay.
Top image: Green crystals of glutathione reductase (a yellow enzyme) with bound methylene blue. Bottom image: Crystal structure of same. Image credit: Heiner Schirmer
At ICAD in Vienna, Wischik discussed further analysis of the same Phase 2 trial of RemberTM, his biotech company's patented formulation of methylene blue. Wischik noted in his talk that the company had patented a new form called leuco-methylthioninium or LMT, which is no longer blue and renders the drug more bioavailable and less toxic at higher doses. For their part, Schirmer's group had characterized a reduced white version of methylene blue (called leucoMB or methylene white) as a possibly superior form of this drug for use in a colorless drug syrup to treat malaria (see Buchholz et al., 2008 and Schirmer essay). TauRx's new formulation is presently undergoing preclinical studies. For a detailed first-person account of what happened with the high dose of RemberTM in the Phase 2 trial and related topics, see Q&A with Wischik below.
Intrigued by the apparent effect in humans, a growing number of research groups are now exploring methylene blue, a phenothiazine compound, in various model systems. Of those, few are ready to report at meetings quite yet, but here is a summary of what's public so far. In Vienna, Christian Haass of the Ludwig Maximilians University in Munich, Germany, presented data straight off the bench from his group's new transgenic zebrafish model system for neurodegeneration. As reported this past March (see ARF related Prague story, play neuron death clip [Acridine uptake and death of P301L-expressing tau neuron] from end of story, and see Paquet et al., 2009), fish embryos expressing mutant human tau rapidly develop signs of neurodegeneration. In this model, Dominic Paquet and Bettina Schmid in Haass' group characterized a series of readouts that range early on from abnormal (PHF1) tau phosphorylation followed by defective neurite outgrowth, a motor phenotype, widespread neuronal death, and tangle pathology. First signs of toxicity show up 32 hours after fertilization, individual tau-laden neurons die by 60 hours after fertilization, and extensive neuron loss is visible after six days. The researchers found tangles later, by five weeks of age. At ICAD, Haass described tests of methylene blue in this model.
Frauke van Bebber in the Haass lab treated the embryos from four hours to 20 hours after fertilization with non-toxic concentrations of 10 to 100 micromolar methylene blue, some of which ends up being absorbed by the fish. The fish turned blue, hence the study was not blinded. “Any investigator studying methylene blue knows precisely what he or she is analyzing,” Haass noted in his talk. In this study, the drug did not alter the abnormal PHF1 phosphorylation of mutant human tau that occurs in the untreated transgenic fish. The scientists take PHF1 phosphorylation as an indirect measure of aggregation and tangle formation, in part because testing for tangles directly with Gallyas staining in older fish is too late and elaborate for drug studies, Haass noted by e-mail. The German scientists next checked whether methylene blue rescued either tau-induced cell death, stunted neurite growth, or the sluggish swimming of the transgenic fish. It did not, Haass said. The model itself is responsive to drug effects, however. For example, some of a series of new GSK3β inhibitors developed by AstraZeneca do reduce abnormal PHF1 phosphorylation by up to 70 percent, Haass reported.
“At this point we wondered whether our methylene blue preparation was inactive,” Haass said in his talk, so the team next tested the drug on a different zebrafish strain, one that models Huntington disease. These fish express fluorescent-coupled huntingtin with an expanded polyglutamine repeat and accumulate aggregates of the transgene in their neuronal nuclei. When treated with methylene blue, the aggregates did not form. “Methylene blue wiped out aggregate formation of httQ102GFP,” Haass said. Next, the scientists looked for effects on toxicity, as this particular transgene is highly toxic, killing the fish. Methylene blue did not block the toxicity of httQ102GFP.
Haass interpreted these findings to mean that methylene blue does inhibit aggregation of tau, as published (Wischik et al., 1996), but that this inhibition is unable to prevent tau toxicity. “I believe that soluble oligomers are the toxic species of tau and huntingtin. It appears that this drug affects species that show pathology but may not be the toxic ones,” Haass said. This finding taps into a fluid area of tau research. Some in-vivo multiphoton imaging studies in transgenic mutant tau mouse models point in a similar direction by suggesting that tau tangles are not what cause neurons to die (e.g., de Calignon et al., 2009; for a recent review on tau and neurodegeneration, see Spires-Jones et al., 2009), but tau aggregation inhibitors are nonetheless drawing renewed interest in academia and drug development (e.g., see Bulic et al., 2009).
Another presentation at ICAD took methylene blue research in a different direction. On a poster, Takashi Nonaka of the Tokyo Institute of Psychiatry, Japan, presented his group's latest data on TDP-43. Working with Haruhiko Akiyama and Masoto Hasegawa, Nonaka has been establishing cell-based models of TDP-43 aggregation. This protein has rapidly moved front and center of molecular studies on amyotrophic lateral sclerosis (ALS) and frontotemporal lobar dementia (FTLD), because intracellular aggregates of TDP-43 mark the pathology of these diseases and TDP-43 mutations have turned up in familial ALS and mixed ALS with FTLD-U. Few TDP-43 animal models are quite ready for primetime just yet. In the interim, the Japanese scientists devised two neuroblastoma lines that express either a TDP-43 deletion mutant missing its nuclear localization signal or an aggregation-prone, fluorescent-tagged C-terminal fragment. Both cell lines form intracytoplasmic aggregates that resemble aggregates seen in patients with these diseases in that they are insoluble to detergent and abnormally phosphorylated. The scientists used these cells to screen for drug effects, and they came up with a combination of none other than Dimebon and methylene blue. These are both available drugs whose safety record could accelerate their use in trials in ALS or FTLD-U. (For frontotemporal dementia, ClinicalTrials.gov at present lists active trials for only one drug, memantine, whose effect is known to be modest, at least in AD.)
Nonaka and his colleague Makiko Yamashita found that treatment with 50 nanomolar methylene blue reduced the number of cellular inclusions by about half compared to untreated cultures. (Dimebon had about the same effect at 5 micromolar.) When given together, both drugs reduced aggregates by 80 percent. The effect was concentration-dependent and occurred in both cell lines. Two other phenothiazine compounds, chlorpromazine and perphenazine, which do not affect tau aggregation, did not affect TDP-43 aggregation in these cell lines, either. The transgenic cells do not appear to die from the TDP-43 inclusions, though TDP-43 is abnormally phosphorylated and its normal, unphosphorylated form is somewhat depleted in the nucleus. Each drug, but especially both of them together, reduces this abnormal phosphorylation, Akiyama wrote in an e-mail to ARF. The scientists suggest that a combination of these drugs be considered for clinical testing in ALS and FTLD and have begun early efforts in this direction (Yamashita et al., 2009; Nonaka et al., 2009; Nonaka et al., 2009).
Together, then, current research suggests that methylene blue blocks aggregation of several different proteins involved in neurodegeneration—tau, TDP-43, polyQ huntingtin, and previous research has shown at least in-vitro effects on Aβ and α-synuclein aggregation as well (e.g., Necula et al., 2007). The Japanese drug study measured aggregation, not toxicity or functional outcomes in vivo, hence gives no clues as to whether blocking TDP-43 aggregation might be beneficial in humans. A rat model of TDP-43, and mouse, fly, or zebrafish models, may be amenable to this question. Several labs in the field are buckling down on developing TDP-43 animal models; for the latest status, see ARF related London story. The Haass lab has a zebrafish line expressing TDP-43 at physiological levels, but because this line does not so far form aggregates, the scientists have not tried methylene blue in it, Haass wrote by e-mail.
With methylene blue, it's not all about aggregation, either. As a drug, it is “dirty.” In pharmacology, this is not nearly as iffy a label as in gastronomy. It simply means a drug may bind to many different targets (see Schirmer essay), not all of which scientists necessarily understand equally well. Dimebon, too, falls into this category, as do some NSAIDs. In the past year, research groups interested in oxidative damage and mitochondrial protection have taken an interest in methylene blue and tested it in various animal models. These scientists did not present at ICAD, but briefly, a Brazilian group reported in the journal Neurochemistry International that injecting methylene blue into the striatum protected rats against induced seizures (Furian et al., 2007). And this year, researchers led by Francisco Gonzalez-Lima at the University of Texas, Austin, reported that methylene blue protected the rat striatum and the retina against injections of the Parkinson disease-related neurotoxin rotenone, possibly by way of propping up mitochondrial energy metabolism (Rojas et al., 2009; Rojas et al., 2009).—Gabrielle Strobel.
Q&A with Claude Wischik. Questions by Gabrielle Strobel.
Q: Are people giving methylene blue to their loved ones?
A: Yes, I am worried that this has the potential to become a serious problem. We have received numerous requests for advice regarding the use of methylene blue. People seem to think that if they find a supplier of methylene blue, they could in principle get a hospital or pharmacy to formulate it for them and then administer it either alone or with the help of a physician. We have heard of a case like this in the U.K., and we understand that the police had to be called in to deal with a report arising from the local health authority. But we have also heard of examples in the U.S.
Q: What's the worry?
A: This is not an approved product. People taking it into their hands to give it to their loved ones are acting irresponsibly. The due regulatory processes have been set up for good reasons, to ensure safety and efficacy of pharmaceuticals, and these apply very much in the case of methylene blue for reasons I will now discuss.
Methylene blue is a relatively crude dyestuff which has various industrial applications. The actual pharmacopoeial substance is methylthioninium chloride (MTC), the chloride salt of the methylthioninium moiety. MTC is not available in a pure pharmaceutical grade anywhere for treatment of AD and related disorders apart from what we have produced for clinical trials. There are several distributors of a form of methylene blue who claim to meet U.S. pharmaceutical standards. However, the standard they quote is very old and was defined by 1970s technology, which would not generally be considered appropriate for a pharmaceutical substance today. The batches of this material that we have tested are impure in regard to MTC itself. They contain manufacturing contaminants, and high levels of heavy metals. We therefore do not advocate their use as a pharmaceutical product, however desperate people may be.
TauRx has gone to great lengths to develop processes whereby MTC can be produced at greater than 99 percent purity. We have obviously patented these processes, which are published as WO06/032879. The material produced by us has very low levels of heavy metal contaminants and meets modern standards for a pharmaceutical product. We believe that the material produced by us would be acceptable to regulatory authorities for Phase 3. I believe the commonly available form of methylene blue is unlikely to satisfy current regulatory standards that apply in the EU or U.S. for administration of pharmaceutical substances to humans.
The next issue concerns the formulation. We found in our Phase 2 clinical trial that unless the formulation is right, even high doses of MTC (e.g., 100 mg 3x/day), whilst having little efficacy, can produce side effects, particularly on red blood cells and diarrhea. I spoke at the recent ICAD in Vienna about our analysis of this data. We now understand why the 30 mg and 60 mg doses showed a simple dose-response relationship in terms of efficacy, but the 100 mg dose did not.
Q: In brief, why?
A: The formulation of MTC we used in the RemberTM Phase 2 trial was a semi-solid gelucire suspension in a gelatin capsule in which different dosage strengths were produced by adjusting the fill-weight during manufacture. For the 100 mg dose, the capsules were near to maximum fill, whereas the 30 mg and 60 mg doses had free head space when the suspension set on cooling in the capsule. MTC was found to cross-link gelatin in the region of contact over time. For the 30 mg and 60 mg doses, this cross-linking did not matter because the capsules could be readily breached via regions of the capsule not in contact with MTC. But breach of the 100 mg capsule was delayed, particularly in simulated gastric fluid, but there was an adequate dissolution profile in simulated intestinal fluid. This led to differential dissolution and absorption according to dose. In effect, the 100 mg capsule was a delayed-release formulation, whereas the 30 mg and 60 mg doses were nearer to rapid release formulations.
We assumed naively going into the trial that it did not matter whether the MT moiety was absorbed via the stomach or the small intestine. In light of the trial data, we developed a simple mathematical model which appears to explain both the beneficial cognitive effect and the adverse hematological effect on the basis of differential absorption of two species according to dissolution time in vitro. This two-species model fits all of the available clinical data with a correlation coefficient of 0.997. The model suggests that absorption of the MT moiety in the reduced form is the main determinant of clinical efficacy, and that this in turn depends on rapid dissolution. However, delayed breach of the 100 mg dose appears to lead to preferential absorption of the MT moiety in the oxidized form from the small intestine, possibly as an uncharged dimer, which forms readily in vitro at high concentration. This species has low clinical efficacy and a high tendency to oxidize hemoglobin.
Q: You also reported at ICAD that you have developed a new, colorless formulation of MTC?
A: Yes. This follows directly from the mathematical analysis I described above. The formulation is critical as the absorption of MT is a complex process, something we had to learn the hard way from the RemberTM Phase 2 trial. This is fully explained in our published patent WO07/110627. When administered as MTC, the MT moiety is presented as the chloride salt of the oxidized form of MT. In order for MT to be absorbed as the beneficial monomer, it has first to be reduced to the uncharged form (i.e., leuco-MT, or LMT), which is able to cross the gut lining and enter the blood. We believe this reduction process is probably enzyme-mediated, and occurs most favorably at the low pH of the stomach. We believe from our experimental work and analysis of our Phase 2 data that it is the LMT form which carries the activity required to retard the progression of AD. The oxidized form is the predominant form absorbed if the formulation is not right (as, e.g., with our 100 mg capsule), and it has the potential to give low efficacy coupled with increased side effects.
We have discovered a process to produce a pure LMT form that can be administered as a tablet. Yes, it is substantially colorless (actually, a light yellowy-green). We have found that when administered orally to animals, this LMT form enhances availability of the active moiety in the brain while reducing toxicity.
Q: Is there going to be a Phase 3 trial?
A: In the near future, we should be ready to conduct Phase 3 with either of the two pure forms of MT that we have developed, but most likely the new LMT form. The final decision as to which form of MT is to be adopted has to await completion of ongoing studies and further discussions with the relevant regulatory authorities.
TauRx is also planning studies in other closely related neurodegenerative disorders which have similar tau pathology to AD, i.e., progressive supranuclear palsy and corticobasal degeneration. We have also had some preliminary evidence of efficacy in a case of frontotemporal dementia with a primary tau mutation, so a clinical trial in FTD is also a likely option.
Q: When and where?
A: In addition to the ongoing PK studies, the further timeline critical issue is completion of relevant higher species toxicology, and regulatory views on this can differ according to jurisdiction. Studies are ongoing with both forms of MT. We plan to start Phase 3 about mid-2010 after these studies have been completed and reviewed by the regulatory authorities. We are planning to undertake Phase 3 in the EU and U.S. with either form. We are also in the midst of a financing round to support Phase 3 and welcome investment enquiries.
Q: Is a manuscript of the full Phase 2 data in preparation/submitted/in press at a peer-reviewed journal?
A: We do plan to submit the RemberTM Phase 2 trial data for peer-reviewed publication in due course. As I discussed at the recent ICAD in Vienna, we have conducted a relatively large disease-modifying trial by Phase 2 standards, and this provides many useful insights into the problems that are inherent in the conduct of such trials in AD. However, we also have to be aware that the overriding imperative is to complete due regulatory process required for us to be able to conduct a Phase 3 trial to confirm our Phase 2 findings. This trial will necessarily require some subjects to receive either placebo or a relatively ineffective dose. Our expert advice is that early peer-reviewed publication of data, which might appear to support off-label use of an impure and inappropriately formulated methylene blue, could work against our ability to conduct such a trial. As I explained above, this is already a potential problem even after a simple conference presentation. This is an important concern and may require us to delay publication until well into Phase 3. Myriad and Medivation both appear to have followed a similar course. We don't rule out earlier publication, and the matter is under constant review.
There is a high level of skepticism about Phase 2 trial data in general in AD irrespective of publication, and there is a further layer of skepticism towards a tau-based approach arising from widely held β amyloid preconceptions. What the field is really waiting for is robust Phase 3 confirmation of any promising Phase 2 efficacy signal for a new disease- modifying approach. The fundamental TauRx objective is to confirm at Phase 3 that a treatment based on targeting the tau aggregation cascade with RemberTM will achieve control over disease progression in AD.