Is LMTM the Norwegian Blue of Alzheimer's Therapy?

19 March 2018

Treatment and interpretation of LMTM trial data by Wilcock and colleagues [1] is reminiscent of the Monty Python sketch in which the demise of an obviously dead parrot is disputed by a desperate shop owner who describes it as merely resting, stunned or pining for the fjords. This phase 3 trial of 18 months' treatment in patients with mild Alzheimer's disease (AD) followed a 15-month study in mild to moderate patients that showed no treatment benefits for LMTM [2]. The latest trial was also apparently negative. Faced with the task of how to best sell their own dead parrot, the authors present a case for the efficacy of LMTM-but only in AD patients not additionally treated with a cholinesterase inhibitor or memantine. In their introduction, the authors say: "As the originally intended analysis was unlikely to achieve its intended purpose, we modified the primary analyses and treatment comparisons in the TRx-237-005 Statistical Analysis Plan prior to database lock and unblinding to investigate whether the monotherapy differences could be confirmed as observational cohort comparisons defined as primary outcomes with strong control of family-wise type 1 error in the second independent study". The intention of any clinical trial will always be to test whether or not pre-specified drug-placebo differences achieve statistical and clinical significance. Once objectivity and respect for the conventions of clinical trial design are lost, so too is the value that can be placed upon one of Medicine's most trusted and powerful research tools. The statement of intent in the introduction is followed by an inventive analysis approach involving counter-intuitive, non-randomised, unmatched and unblinded participant group comparisons, which are themselves much more likely to have generated any apparent benefits than the unsupported theory that LMTM works to slow AD, but acetylcholinesterase inhibitor treatment sabotages this by impairing the clearance of tau monomers.

Monty Python's dead parrot was, of course, a Norwegian Blue. A challenge for those who design trials involving compounds based on methylene blue is that administration of these agents discolours participants' urine and faeces, quickly unblinding treatment allocation. Of all the claims made in the current report, the most remarkable is that the dose of 4mg of LMTM twice daily, chosen to act as a urine-colouring placebo, turned out to be as effective as the active intervention dose of 100mg twice daily, selected on the basis of earlier dose-ranging studies. It is worth bearing in mind that this group previously justified post-hoc exclusion of data from participants allocated to the highest dose of methylthioninium on the basis of unconvincing and unsupported mechanisms to generate an apparently positive outcome [3], but claims of efficacy based on data from placebo participants represent a whole new level of audacity.

The authors acknowledge some of the limitations of their analysis, but the overall message of the paper is misleadingly positive about LMTM's potential as an AD therapy. For example, there is no mention in the abstract of the overall non-significance of the originally intended primary outcome data and these are not presented in the paper. Convincing failure of LMTM in two large and well-conducted AD trials should have been sufficient to finally close down investigation of this agent. Conduct of future AD trials involving participants who are not receiving cholinesterase inhibitors or memantine should be acknowledged as hugely problematic, even by those who don't share our scepticism about the use of these latest data. Will potential participants be persuaded to forgo evidence-based treatments by a flimsy promise of disease modification based on the data? Will only patients who can't tolerate licensed dementia drugs (and very few are likely to be intolerant of memantine) be eligible for such a trial, in which case recruitment will be slow and participants highly unrepresentative? Even, in the unlikely event of demonstration of a degree of disease-modification in such a trial, is there a realistic market for an agent that cannot be given alongside the current symptomatic treatments?

We hope that the authors will quickly publish the results of the original full per protocol analysis. The search for AD-modifying treatments must now move on to other and more promising agents, despite what look like increasingly ridiculous efforts to nail LMTM's feet to its perch.

Robert Howard, Professor of Old Age Psychiatry, Division of Psychiatry, University College London, 149 Tottenham Court Road, London W1T 7NF, UK. E-mail: robert.howard@ucl.ac.uk.

Alex Berry, Core Trainee in Psychiatry, Camden and Islington NHS Trust, 4 St Pancras Way, London NW1 0PE, UK.

References
[1] Wilcock GK, Gauthier S, Frisoni GF, Jia J, Hardlund JH, Moebius HJ, Bentham P, Kook KA, Schelter BO, Wischik DJ, Davis CS, Staff RT, Vuksanovic V, Ahearn T, Bracoud L, Shamsi K, Marek K, Seibl J, Riedel G, Storey JMD, Harrington CR, Wischik CM (2018) Potential of low dose leuco-methylthioninium bis(hydromethanesulphonamate) (LMTM) monotherapy for treatment of mild Alzheimer's disease: Cohort analysis as modified primary outcome in a phase III clinical trial. J Alzheimers Dis 61, 435-457.
[2] Gauthier S, Feldman HH, Schneider LS, Wilcock GW, Frisoni GB, Hardlund JH, Moebius HJ, Bentham P, Kook KA, Wischik DJ, Schelter BO, Davis CS, Staff RT, Bracoud L, Shamsi K, Storey JMD, Harrington CR, Wischik CM (2016) Efficacy and safety of tau-aggregation inhibitor therapy in patients with mild or moderate Alzheimer's disease: a randomised, controlled, double-blind, parallel-arm, phase 3 trial. Lancet 388, 2873-2884.
[3] Wischik CM, Staff RT, Bentham P, Murray AD, Storey JMD, Kook KA, Harrington CR (2015) Tau aggregation inhibitor therapy: An exploratory phase 2 study in mild or moderate Alzheimer's disease. J Alzheimers Dis 44, 705-720.

Comments

Professor Howard and Dr Berry argue that LMTM should not be investigated further as a potential treatment of Alzheimer's disease based on inhibition of pathological tau aggregation. We thank the Editor for providing the opportunity to respond to their letter, as the arguments they present are based on a number of factual errors and inaccuracies.

Before detailing these, it is important to highlight the strong scientific case for the potential clinical utility of LMTM. In vitro, it prevents pathological aggregation of tau protein in cell-free and cell-based models [1], and reduces tau pathology and accompanying behavioural deficits in two tau transgenic mouse models [2]. There is a growing body of literature indicating that in addition to inhibition of tau aggregation, the methylthioninium (MT) moiety is effective in vivo in enhancing autophagy [3], inducing Nrf2-mediated proteotoxic clearance pathways [4] and enhancing mitochondrial metabolism [5] at clinically relevant concentrations of MT. Finally, the MT moiety is clinically safe and, as the LMTM form, is well tolerated.

Howard and Berry argue first that both of the Phase 3 trials we have recently completed failed to find drug-placebo differences. This assertion is inaccurate, since neither of the trials had a placebo control. Both trials were actually dose-comparison studies that were based on the assumption that a low dose (4 mg twice a day), intended as a control for urinary discolouration, would be ineffective.

This assumption was based on an earlier placebo-controlled dose-finding monotherapy Phase 2 study using not LMTM, but the oxidised form of the methylthioninium moiety, namely methylthioninium chloride (MTC, commonly known as methylene blue). In referring to this study, Howard and Berry erroneously assert that there was "post hoc exclusion of data from participants allocated to the highest dose". This is simply not the case. The study was analysed as randomised, achieved statistical significance on the primary efficacy endpoint as defined a priori in the Statistical Analysis Plan, and identified 138 mg/day of MTC as the minimum effective dose. This result was confirmed independently using HMPAO-SPECT and FDG-PET in nested sub-studies [6].

Howard and Berry assert that the explanation for a relative lack of effect of a higher dose (210 mg/day) was based on "unconvincing and unsupported mechanisms". They appear not to have read either the paper reporting the study [6], nor an earlier paper reporting a large body of work aimed at discovering why there was dose-dependent impairment in absorption of the MT moiety when delivered as the oxidised species as MTC [7]. It was because of this limitation that a stabilised reduced form of MT (i.e., as LMTM) was taken forward for further development. It was discovered only after the Phase 3 studies had been initiated that in vivo the cellular uptake of MT delivered as LMTM is in fact 20-fold better than for MTC [7]. Ongoing PK studies have shown that the 4 mg twice a day dose of LMTM delivers an estimated brain concentration of MT that would be expected to be therapeutically active. Howard and Berry consider that reporting of this low dose as potentially beneficial "represents a whole new level of audacity".

Howard and Berry assert that "the overall non-significance of the originally intended outcome data … are not presented in the paper". On the contrary, the first sentence of the discussion states that "there was no evidence of any difference on any of the primary or secondary endpoints in the as-randomised analyses defined in the protocol". The title of the paper states clearly that it presents a "cohort analysis as modified primary outcome in a phase 3 clinical trial" [8].

Howard and Berry assert that the analyses we present are "counter-intuitive, non-randomised, unmatched and unblinded". Our abstract clearly states that the objective of the analyses was to "examine the efficacy of LMTM as monotherapy in non-randomised cohort comparisons". The abstract also states that the statistical analysis plan was modified prior to unblinding. The analyses were not "counter-intuitive" in the light of the findings from the first of the two phase 3 trials [9], the results of which became available prior to database lock and unblinding of the second study. The aim was to see if the results indicating differences in favour of monotherapy from the earlier study [9] were reproduced as primary outcomes in a second independent study with strong control of type I error.

Howard and Berry take the view that the differences in favour of monotherapy that we report can be accounted for simply by inherent cohort differences. We do not dispute that this is a plausible hypothesis. However, we present analyses that exclude obvious baseline differences in severity, extent of brain atrophy, severity of glucose uptake deficit, non-AD diagnosis, APOE ε4 allele frequency and concomitant vascular pathology as explaining the differences in favour of monotherapy. We fully acknowledge, as stated in the final paragraph of the paper, that "non-obvious or unmeasured confounding factors cannot be excluded without a further randomised clinical trial comparing LMTM with true placebo".

Howard and Berry take issue with the hypothesis that anticholinesterase inhibitor treatment may impair the clearance of tau monomers released by LMTM. We have recently found evidence in two different tau transgenic mouse models of a negative interaction between LMTM and an acetylcholinesterase inhibitor of the same kind as the interaction seen in the phase 3 trials. These results have been submitted as abstracts to be followed in due course by full reports. This negative interaction appears, therefore, to be a real neuropharmacological phenomenon, and an extensive body of ongoing work aims to elucidate the underlying mechanism. There is a broader implication that other intended disease-modifying treatments that rely on intact neuronal clearance pathways may also be impacted negatively when administered in combination with approved symptomatic treatments for AD.

Howard and Berry appear to consider that we are not permitted to use a cohort analysis as the basis for generating a hypothesis for a further clinical study. The paper presents the case that there is a reasonable scientific rationale for conducting a further placebo-controlled study. Results that cannot be dismissed lightly include the striking consistency in clinical and imaging outcomes in two independent trials, and the significant within-cohort deceleration in brain atrophy rate seen after 9 months in patients receiving LMTM 8mg/day as monotherapy. We believe, therefore, that there is valid case for conducting further clinical trials, and indeed the first of two planned global studies is already ongoing.

Finally, Howard and Berry consider that there is no place for a disease-modifying monotherapy for AD unless it is on the background of currently approved symptomatic treatments. This is a strange view given the extensive efforts to identify and treat the earliest stages of the AD disease process in multiple ongoing clinical trials. The trial we are currently conducting is in very mild AD at the CDR 0.5 stage. It is entirely plausible that disease-modifying treatments will eventually become first-line if these can be shown to have efficacy in appropriately randomised placebo-controlled trials.

Gordon K. Wilcocka, Serge Gauthierb, Jianping Jiac, Jiri H. Hardlundd, Hans J. Moebiuse, Peter Benthamf, Karin A. Kookg, Bjoern O. Schelterh, Damon J. Wischiki, Charles S. Davisj, Roger T. Staffk, Vesna Vuksanovick, Trevor Ahearnk, Luc Bracoudl, Kohkan Shamsim, Ken Marekn, John Seibyln, Gernot Reidelo, John M.D. Storeyd,p, Charles R. Harringtond,o , Claude M. Wischikd,o (E-mail: cmw@taurx.com)

aNuffield Department of Clinical Neurosciences, University of Oxford, Oxford, UK
bMcGill Centre for Studies in Aging, Alzheimer's Disease Research Unit, and Douglas Mental Health University Institute, Montreal, QC, Canada
cBeijing Institute for Brain Disorders Alzheimer's Disease Centre, Beijing, China
dTauRx Therapeutics, Aberdeen, UK
eMoebius-Consult, Baar, Switzerland
fBirmingham and Solihull Mental Health Foundation Trust, Birmingham, UK
gSalamandra LLC, Bethesda, MD, USA
hInstitute for Complex Systems and Mathematical Biology, University of Aberdeen, Aberdeen, UK
iComputer Laboratory, University of Cambridge, Cambridge, UK
jCSD Biostatistics, Tucson, AZ, USA
kAberdeen Biomedical Imaging Centre, School of Medicine, Medical Sciences and Nutrition, University of Aberdeen, Aberdeen, UK
lBioClinica, Lyon, France
mRadMD, New York, NY, USA
nMNI Imaging, New Haven, CT, USA
oSchool of Medicine, Medical Sciences and Nutrition, University of Aberdeen, Aberdeen, UK
pDepartment of Chemistry, University of Aberdeen, Aberdeen, UK

References
[1] Harrington CR, Storey JMD, Clunas S, Harrington KA, Horsley D, Ishaq A, Kemp SJ, Larch CP, Marshall C, Nicoll SL, Rickard JE, Simpson M, Sinclair JP, Storey LJ, Wischik CM (2015) Cellular models of aggregation-dependent template-directed proteolysis to characterize tau aggregation inhibitors for treatment of Alzheimer's disease. J Biol Chem 290, 10862-10875.
[2] Melis V, Magbagbeolu M, Rickard JE, Horsley D, Davidson K, Harrington KA, Goatman K, Goatman EA, Deiana S, Close SP, Zabke C, Stamer K, Dietze S, Schwab K, Storey JMD, Harrington CR, Wischik CM, Theuring F, Riedel G (2015) Effects of oxidized and reduced forms of methylthioninium in two transgenic mouse tauopathy models. Behav Pharmacol 26, 353-368.
[3] Congdon EE, Wu JW, Myeku N, Figueroa YH, Herman M, Marinec PS, Gestwicki JE, Dickey CA, Yu WH, Duff KE (2012) Methylthioninium chloride (methylene blue) induces autophagy and attenuates tauopathy in vitro and in vivo. Autophagy 8, 609-622.
[4] Stack C, Jainuddin S, Elipenahli C, Gerges M, Starkova N, Starkov AA, Jové M, Portero-Otin M, Launay N, Pujol A, Kaidery NA, Thomas B, Tampellini D, Beal MF, Dumont M (2014) Methylene blue upregulates Nrf2/ARE genes and prevents tau-related neurotoxicity. Hum Mol Genet 23, 3716-3732.
[5] Rojas JC, Bruchey AK, Gonzalez-Lima F (2012) Neurometabolic mechanisms for memory enhancement and neuroprotection of methylene blue. Prog Neurobiol 96, 32-45.
[6] Wischik CM, Staff RT, Wischik DJ, Bentham P, Murray AD, Storey JMD, Kook KA, Harrington CR (2015) Tau aggregation inhibitor therapy: an exploratory phase 2 study in mild or moderate Alzheimer's disease. J Alzheimers Dis 44, 705-720.
[7] Baddeley TC, McCaffrey J, Storey JMD, Cheung JKS, Melis V, Horsley D, Harrington CR, Wischik CM (2015) Complex disposition of methylthioninium redox forms determines efficacy in tau aggregation inhibitor therapy for Alzheimer's disease. J Pharmacol Exp Therap 352, 110-118.
[8] Wilcock GK, Gauthier S, Frisoni GB, Jia J, Hardlund JH, Moebius HJ, Bentham P, Kook KA, Schelter BO, Wischik DJ, Davis CS, Staff RT, Vuksanovic V, Ahearn T, Bracoud L, Shamsi K, Marek K, Seibyl J, Reidel G, Storey JMD, Harrington CR, Wischik CM (2018) Potential of low dose leuco-methylthioninium bis(hydromethanesulphonate) (LMTM) monotherapy for treatment of mild Alzheimer's disease: cohort analysis as modified primary outcome in a phase 3 clinical trial. J Alzheimers Dis 61, 635-657.
[9] Gauthier S, Feldman HH, Schneider LS, Wilcock GK, Frisoni GB, Hardlund JH, Moebius HJ, Bentham P, Kook KA, Wischik DJ, Schelter BO, Davis CS, Staff RT, Bracoud L, Shamsi K, Storey JMD, Harrington CR, Wischik CM (2016) Efficacy and safety of tau-aggregation inhibitor therapy in patients with mild or moderate Alzheimer's disease: a randomised, controlled, double-blind, parallel-arm, phase 3 trial. Lancet 388, 2873-2884.