As part of our ongoing coverage of the 2009 International Congress on Schizophrenia Research (ICOSR), 28 March to 1 April 2009, in San Diego, California, we bring you a meeting missive from researcher C. Anthony Altar, NeuroDrug Consulting. In his first report from ICOSR, Tony described rough sailing for Lilly's mGluR2/3 agonist, LY2140023 (see SRF related news story). Here he surveys other antipsychotic treatment contenders presented at the Sunday, 30 March, morning symposium on New Developments in Treatment, chaired by S. Charles Schulz.
22 April 2009. Promising results in drug development for schizophrenia were scarce at the ICOSR, except for a standout, the multiple receptor-active drug, lurasidone. It shows potential efficacy for schizophrenia and some of the depressive symptoms associated with this disease.
Marc Laurelle, from GlaxoSmithKline in London, provided no clinical data, but he gave as comprehensive a review as 15 minutes would allow for the use of a glycine transport inhibitor for treating schizophrenia. The type 1 glycine transporter (GLYT1) is present in non-human primate and rodent spinal cord, mesencephaon, and cerebellum. While these are not notable hotspots for schizophrenia etiology or treatment, a GLYT1 radiolabeled PET tracer, GSK931145, is being profiled for its distribution in human brain. However, PET study results for [18F]CFPyPB, already presented at the 2008 Society for Neuroscience meeting (885.15), cast doubt on this approach based on localization. In both monkey and human, GLYT1 was bound in vivo by [18F]CFPyPB and with high, 90 percent, specificity. However, binding was densest in pons, almost as high in cerebellum, and very low or absent in striatum, hippocampus, and neocortex. Still, the ability of the GSK931145 transport blocker to increase neuron firing rates when combined with NMDA block PCP-induced changes in fMRI signals in rat brain, and PCP-induced cognitive deficits in rodents may be positive for therapy. Thus, based on mechanism if not localization, GLYT1 inhibition may overcome NMDA hypofunction in schizophrenia by increasing synaptic glycine levels.
In the same symposium, Uli Hacksell, from Acadia Pharmaceuticals in San Diego, described how their highly selective, 5 nM-affinity 5HT2A inverse agonist, ACP 103 (primavanserin) was tried as an adjunct to risperidone or haloperidol to treat schizophrenia. Primavanserin is a 5HT2A/C inverse agonist that occupies ~60 percent of 5HT2A/C sites in vivo and 30 percent of D2 sites when given to rats at the impressively low dose of 0.1 mg/kg. The six-week clinical combination studies showed no added benefit to the level of improvement obtained with 4 mg risperidone alone, except for a modest increase in the PANSS score in patients treated with 2 mg risperidone. These findings are at odds with the primavanserin block of catalepsy and cognitive deficits produced in rats treated with haloperidol, risperidol, olanzapine, or clozapine.
The conclusion from these studies is that, as shown many times before, animal models can yield false positives for clinical response, particularly for drugs whose targets are yet to have been proven in the clinic. Of course, one hopes true positives are the case when Acadia tests two doses of primavanserin against psychosis in Parkinson's disease in a 240-patient, six-week, Phase 3 study planned at 50 U.S. and international sites. The rationale for these studies is based on the ability of primavanserin to reverse the near-complete loss of pre-pulse inhibition of startle in rats with dopaminergic lesions, and to reverse the increase in amphetamine-hyperactivity in muscarinic receptor type 1 (M1) deletion mice. These studies are supported by other clinical results presented by Hacksell, in which 500 volunteers and Parkinson's patients with or without psychosis have, in Phase 1 and 2 studies, shown good tolerability to primavanserin, and a long duration of action. This was assessed by no change (but thus also no improvement) in Parkinson's symptoms of 60 patients, reflected by UPDRS parts I and II, and by modest and generally not significant improvements of psychosis, particularly hallucinations and delusions.
Kimberly Vanover, from Intra-Cellular Therapies (ITI), described ITI-007 in the New Developments in Treatment session. This 007 agent remains a little mysterious. ITI-007 has 0.5 nM affinity as a 5HT2A antagonist, and potently (0.15 mg/kg) blocks head shakes induced by the 5HT2A agonist, DOI. It has far less affinity (32 nM) as a presumptive D2 "presynaptic partial agonist/postsynaptic antagonist that is mesolimbic selective." It is weaker (52 nM) at the serotonin transporter (SERT), and inactive at 5HT1A. Its inactivity at H1, β2C, β1, or muscarinic receptors is cited with good reason to predict that ITI-007 may not produce sedation or metabolic side effects. ITI-007 has been shown in clinical studies to dose-dependently increase slow wave sleep and hasten waking after sleep onset, and sleep time, in patients with sleep maintenance insomnia. It was also proposed to be a potential therapy for depression and to lack sexual side effects. The occupancy of D2 receptors in human brain, determined by [11C]-raclopride PET studies, attained only 10, 20, and 30 percent with increasing doses of 10, 20, and 30 mg of ITI-007. The highest, 30 percent, occupancy is probably still insufficient to mediate antipsychotic efficacy, unless, for example, if a higher occupancy of limbic D2-like receptors could be shown. A limbic versus striatal selectivity of D2 compounds was supported by second messenger activation in limbic regions.
Unfortunately, such low occupancy for D2 also bodes poorly for the even weaker affinity for the serotonin transporter, which probably requires occupancy of >50 percent to produce antidepressant effects. While a 75-patient and -volunteer study has shown ITI-007 to be "safe and well tolerated over a wide range of doses," neither the PET occupancy studies, nor its claim for mesolimbic and mesocortical dopamine systems, seemed convincing. Perhaps the best evidence for an antipsychotic effect was preclinical activity at a dose of 1.5 mg/kg in the conditioned avoidance response assay. Again, caution comes from the inability of the 5HT2A receptor inverse agonist, primavanserin, to augment antipsychotic efficacy of risperidol or haloperidol. Consistent with the serotonin-dopamine antagonism (SDA) hypothesis for diminished antipsychotic side effects (Altar et al., 1986), primavanserin did diminish the EPS and body weight gain obtained with the low or high (6 mg) dose of risperidone, and did not elevate serum prolactin (see SRF related news story). Such beneficial effects may be predicted for ITI-007 in combination with a typical antipsychotic, but without higher D2 and SERT affinity, the antipsychotic or antidepressant potential of the compound seem less likely. Results of the planned study of 20 mg in schizophrenia will be of great interest.
Lurasidone—new kid on the block
Fresh promise was not completely missing from the realm of new antipsychotics, however. Lurasidone, asenapine, and SB269970 are among the new antipsychotic candidates active in preclinical models of cognition and schizophrenia. Presentations at ICOSR on lurasidone indicated a novel set of receptor mechanisms, smaller placebo responses, and voila! confirmation of Phase 2 efficacy!
Lurasidone potencies at D2, 5HT6, 5HT7, 5HT1A, and β2c receptors and lack of interaction with M1 or H1 receptors may contribute to these preclinical and clinical actions. These properties were summarized in posters presented by Dainippon Sumitomo Pharma America and their collaborators. Hiroyuki Nishikawa (Dainippon Sumitomo Pharma Co., Ltd., Osaka, Japan; Poster 97) and colleagues summarized the potent D2 antagonism, which probably mediates efficacy, and the 5HT2A antagonism, partial 5HT1A agonism, and β2C antagonism that may reduce EPS. The favorable preclinical profile for lurasidone, particularly its non-effects in mouse or rat catalepsy, or bradykinesia tests, versus two typicals and five atypical antipsychotics was supported by considerable behavioral data.
Philip Harvey of Emory University and colleagues (Dainippon/Merck funded; Poster 212) used the Schizophrenia Cognition Rating Scale (SCoRs) to show a bigger effect size of lurasidone compared with ziprasidone. This test appeared to be simpler and more sensitive than the MATRICS-consensus cognitive battery (MCCB) test.
John Guarino (Dainippon America; Poster 83) and colleagues presented relatively low placebo responses and positive effects on symptom remission for 50 patients per group. These effects were confirmed by Josephine Cucchiaro (Dainippon America; Poster 82) and colleagues in an even larger (~150 pts/group) study. They found some of the PANSS positive, negative, and total scores to be superior to ziprasidone during several times in the three-week trial. Unfortunately, there was only a "one- to three-day" placebo run in, with no active placebo comparator. A Phase 3 placebo-controlled study is underway.
Masaaki Ogasa (Dainippon America; Poster 81) and colleagues presented a six-week, double-blind, placebo-controlled randomized trial conducted in the U.S. with 90 patients per group. Lurasidone at 80 mg produced a 15-point improvement in the PANSS-total score, and antidepressant activity determined with the MADRS total score and the PANSS G6 depression items score. Perhaps those many receptors provide a "magic shotgun" for two or more targets at once. There's a lot of anticipation for that Phase 3 study of lurasidone.—C. Anthony Altar.