On March 31, the Australian company Prana Biotechnology announced negative top-line results of its only Phase 2 study of the anti-amyloid drug PBT2. Called IMAGINE, the trial had enrolled 42 people with prodromal or mild Alzheimer’s disease who had a positive brain amyloid scan.

This trial evaluated how a one-year course of a once-daily, 250 mg PTB2 capsule would affect brain amyloid deposition as measured by positron emission tomography with the amyloid imaging tracer Pittsburgh compound B (PiB). Based on draft results, the trial missed its primary endpoint, which was defined as a statistically significant reduction in plaque levels. Overall PiB retention in PBT2-treated patients did go down, but it went down in the placebo group as well.  Of the 42 enrolled patients, 27 received the drug, 15 the placebo. 

The market’s response was swift. Within hours, the company’s stock plummeted after it had risen temporarily in response to a February 18 announcement of top-line results of a separate Phase 2 trial in Huntington’s disease (see Feb 2014 news story). 

In addition to amyloid load, the IMAGINE AD trial had five secondary outcomes. The first confirmed the results of the recent Huntington’s trial in that PBT2 was both safe and tolerable, with equivalent adverse event profiles between groups, according to Prana’s announcement. The second secondary outcome was neuronal function as measured by fluorodeoxyglucose (FDG) PET. On that measure, the company reported no benefit.

The third secondary outcome was PBT2’s effect on brain volume as measured by magnetic resonance imaging of cortical gray matter, hippocampus, and ventricles. On that, the company reported no statistical significance but a trend toward less hippocampal atrophy in the treated group. The fourth secondary outcome was cognition as measured by the neuropsychological test battery (NTB) and the mini-mental state examination (MMSE); PBT2 was negative on both. The final secondary outcome sought to quantify how patients function by way of the Alzheimer’s Disease Cooperative Study-Activities of Daily Living (ADCS-ADL); PBT2 had no effect on this measure.

Prana scientists offered their interpretation of the results in the company press release. The company posted an audio recording of a March 31 investor call, in which company and associated scientists discussed the results and took questions. 

PBT2 is a second-generation drug growing out of prior research on clioquinol. Both are so-called metal-protein interaction-attenuating compounds. MPACs are thought to reduce amyloid aggregation by interfering with the interaction of copper and zinc with beta amyloid.

For details of the IMAGINE trial design, see Australian New Zealand Clinical Trials Register. An extension study is still ongoing. The trial was partially funded by the Alzheimer’s Drug Discovery Foundation in New York City.—Gabrielle Strobel


  1. The big picture here is that Prana’s only prior Phase 2 trial with PBT2 — completed in 2007 with 78 mild AD patients treated for 12 weeks — showed lower CSF Aβ42 and nominally significant effects on verbal fluency and working memory. A reasonable development program would have called for a larger and longer follow-on efficacy trial and confirmation of the CSF Aβ42 outcome. That trial could have been a Phase 2/3 pivotal trial, for 12 months, and included from 400 to 1,200 patients. A second, Phase 2b/3 trial could have been overlapped to ensure that about 1,600 patients would be exposed in order to provide a minimum data base for regulatory approval.

    This didn’t happen with PBT2, while drugs from other sponsors with weaker stories, and less efficacy and safety evidence, were advanced to Phase 3 (where, thus far, they’ve failed).  Instead, this 42-patient prodromal AD PiB trial was launched five years after the 2007 results. By its design, it could not answer relevant questions.  It was more a Hail Mary pass than a hypothesis test.  If the trial had shown a nominal statistically significant result on its PiB primary outcome, or even on one of its MRI volumes, then it might have attracted those investors who chase P values.

    However, as a function of the trial’s small size, made even more inefficient by making the placebo group smaller than the drug group, any nominal P < .05 they might have found would have come with an implausibly large effect size and been controversial nevertheless.  Indeed, Prana reported a “trend” favoring less hippocampus volume loss with PBT2; and one of their experts mused that if they had called hippocampal volume the primary outcome then they would have been talking success.

    This story teaches us that business does not always act in harmony with preclinical science. No new information is learned from this study relevant to the future development of PBT2 for AD.  Too bad PBT2 didn’t get a better chance to prove any potential efficacy.   

  2. The community as a whole—independent researchers, colleagues and competitors alike—are eagerly awaiting a successful AD trial, hence there is widespread disappointment about the results of this 52-week study of PBT2 in 42 prodromal AD or mild AD patients.

    Based on what little information has been released, I believe it is not surprising that the trial saw no significant effect of PBT2 on brain amyloid deposition as assessed with PiB-PET. The reason is that annual change in global PiB retention in MCI or AD subjects is quite small (Jack et al., 2009). With a small sample size, it is hard to detect a drug effect unless you have a potent drug favoring plaque clearance, such as gantenerumab (Ostrowitzki et al., 2012). 

    The lack of effect on brain glucose metabolism is also not surprising because the annual rate of change of FDG-PET in prodromal AD is quite limited (Knopman et al., 2013). 

    I do not see as particularly encouraging the trend favoring PBT2 in the rate of hippocampus atrophy compared to placebo (-2.6 percent vs. -4.0 percent), because the mean atrophy rate observed in the placebo group is in the upper range of what one would expect in mild AD subjects (Kaye et al., 2005). Thus, the apparent difference in favor of PBT2 could be by chance due to limited sample size. In addition, the positive trend observed in the hippocampal atrophy rate seems not confirmed for ventricles or cortex.

    In my opinion, a quite negative outcome of the study is the apparent lack of any trend favoring PBT2 in neuropsychological tests. This is disappointing in particular for executive function. Apparent positive effects on the Trail-Making Test B and the Category Fluency Test had been observed in the previous 12-week study in 78 subjects with mild to moderate AD (Lannfelt et al., 2008). A significant improvement on Trail-Making Test B has also been claimed in the recent 26-week Phase 2 study in 104 patients with Huntington’s disease. A possible reason is that the IMAGINE AD study included subjects with prodromal AD, who generally decline less rapidly than mild to moderate AD patients. Another possibility could be that the significant effects on measures of executive function observed in the previous Phase 2 study in mild to moderate AD subjects resulted from not adjusting p values for multiple comparisons.

    The good news of the IMAGINE AD trial is the apparent good safety and tolerability of the drug. Only two patients discontinued the study.  This is reassuring since in the Huntington’s Phase 2 study there was a dose dependence of serious adverse events (1 on placebo, 3 on 100 mg PBT2, 6 on 250 mg PBT2).

    Most of my comments are speculative because numerical data of the primary outcome and most secondary outcomes were not disclosed. Importantly, we do not know the characteristics of the patients at baseline. Potential differences between treatment groups at baseline can strongly influence the results of a study. The only numerical data provided, on hippocampal brain atrophy, are not accompanied by a measure of variability (standard error or standard deviation). A deeper understanding of this trial will emerge once these and other data become available.


    . Serial PIB and MRI in normal, mild cognitive impairment and Alzheimer's disease: implications for sequence of pathological events in Alzheimer's disease. Brain. 2009 May;132(Pt 5):1355-65. PubMed.

    . Mechanism of amyloid removal in patients with Alzheimer disease treated with gantenerumab. Arch Neurol. 2012 Feb;69(2):198-207. Epub 2011 Oct 10 PubMed.

    . Selective Worsening of Brain Injury Biomarker Abnormalities in Cognitively Normal Elderly Persons With β-Amyloidosis. JAMA Neurol. 2013 Aug 1;70(8):1030-8. PubMed.

    . Asynchronous regional brain volume losses in presymptomatic to moderate AD. J Alzheimers Dis. 2005 Sep;8(1):51-6. PubMed.

    . Safety, efficacy, and biomarker findings of PBT2 in targeting Abeta as a modifying therapy for Alzheimer's disease: a phase IIa, double-blind, randomised, placebo-controlled trial. Lancet Neurol. 2008 Sep;7(9):779-86. PubMed.

  3. Our group has been working on longitudinal amyloid PET measurements recently, and there are a number of methodological challenges. Measuring longitudinal amyloid PET is new territory, and there is not yet a clear answer about the best way to do it.

    The field still has work to do to standardize analysis methods and eliminate noise in the data. In some amyloid PET datasets the degree of one-year change falls below the signal-to-noise confidence interval limit.  A critical factor in addressing this noise is the selection of a reference region that is used to detect relative changes in cortex. Unexplained noise in the reference region, perhaps due to white-matter retention or image acquisition inconsistencies, could result in the appearance of increases or decreases in placebo (or even treatment) groups.

    There is also some evidence that the amyloid PET signal decreases over time in AD. There is as yet no scientific consensus on what this means physiologically, or whether it might perhaps even be an artifact due to atrophy or reference region selection. 

    Without knowing exactly how the imaging data was analyzed in the IMAGINE study, it is difficult to speculate why it found a decrease in the placebo group. But it is not surprising that it would be challenging to detect differences in small placebo and treatment groups over a relatively short time period of a year.

  4. The IMAGINE study conducted by Prana Biotechnology was a small (n=42) exploratory trial supported by a grant from the Alzheimer’s Drug Discovery Foundation to see if the novel 8-OH quinoline (PBT2) would have an effect on Aβ-amyloid burden as reported by PiB-PET neuroimaging over 12 months. Secondary objectives included MRI volumetry, FDG-PET, cognitive change and Aβ-related blood biomarkers. The draft preliminary results indicated a surprising fall in the PiB SUVRs in the placebo group, which obscured any statistical separation from the drug treatment group (which also showed a larger decrease).

    I agree with most of the comments from Lon Schneider, Bruno Pietro Imbimbo, and Susan Landau. I would prefer to reserve judgement on the study until all the data have been verified, analysed and publicly released, particularly the effect of the ApoE4 polymorphism and the changes in blood biomarkers (especially Aβ dimers). It's worth noting that the IMAGINE trial is still ongoing in an open-label extension study that will conclude in early 2015, and that the overall safety data from the IMAGINE trial so far is excellent.

    Further independent support for the use of this class of drug has come from unbiased screens of Aβ-toxicity (Matlack et al., 2014). Indeed the scientific argument for this therapeutic approach has become more compelling with published data for PBT2 demonstrating its reduction of tau protein, upregulation of synaptic plasticity markers, and reduction of Aβ-induced synaptotoxicity (Adlard et al., 2008).

    It is far too premature to abandon this clinical approach of proof-of-concept. As Lon comments, a reasonable development program would have entailed a much larger study, but unfortunately the funds for this are yet to be sourced. Lon's further analogy with a Hail Mary pass invites the rebuttal that the effort was more like an Our Father, or even The Credo: “I believe in one principle, the almighty Aβ, cause of all Alzheimer's disease, both clinical and preclinical." 

    For those who appreciate that an n=42 PiB study does not portend the doom of this novel therapeutic, it is hoped that there are others who understand that the real test and opportunity for PBT2 in Alzheimer’s disease lies in an appropriately powered trial to explore clinical benefit.   


    . Clioquinol promotes the degradation of metal-dependent amyloid-β (Aβ) oligomers to restore endocytosis and ameliorate Aβ toxicity. Proc Natl Acad Sci U S A. 2014 Mar 18;111(11):4013-8. Epub 2014 Mar 3 PubMed.

    . Rapid restoration of cognition in Alzheimer's transgenic mice with 8-hydroxy quinoline analogs is associated with decreased interstitial Abeta. Neuron. 2008 Jul 10;59(1):43-55. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Drug Appears Safe in Huntington’s; Experts Split on Functional Benefit

External Citations

  1. stock plummeted
  2. company press release
  3. audio recording
  4. Australian New Zealand Clinical Trials Register

Further Reading


  1. . Safety, efficacy, and biomarker findings of PBT2 in targeting Abeta as a modifying therapy for Alzheimer's disease: a phase IIa, double-blind, randomised, placebo-controlled trial. Lancet Neurol. 2008 Sep;7(9):779-86. PubMed.