Neuroscience 2008—MGluR Agonists and Schizophrenia: 2-D or Not 2-D?
In his previous SfN meeting reports, our correspondent Tony Altar assessed cholinergic approaches (see SRF related news story) to schizophrenia and non-cholinergic approaches (see SRF related news story) to treating cognition in the disorder. He also whiled away the hours at the "Wall of D Partial Dopamine Agonists" (see SRF related news story). In this report, he explores a budding controversy regarding whether D2 receptor partial agonism characterizes the Lilly metabotropic glutamate receptor agonists.
7 January 2009. As we know from the landmark publication of Phase 2 results by Patil and colleagues at Eli Lilly and Co., the oral prodrug of LY404039, LY2140023, improves positive and negative symptoms in schizophrenic patients (see SRF related news story and SRF news story). LY2140023 produced about a 21 percent reduction in the PANSS total scale, compared to a 27 percent reduction with olanzapine, both significant (p <0.001) versus placebo (Table 1, Patil et al., 2007). For those schizophrenologists among us, one of the most intriguing issues at SfN this year was the dopamine-D2-receptor-or-not controversy regarding this class of metabotropic glutamate agonists. A potent, direct D2 receptor partial agonism of the group 2 metabotropic glutamate receptor agonists LY354740 and LY379268 was reported by the inventor of the radioligand D2 binding assay, Phil Seeman, and his colleagues (Seeman et al., 2008; Seeman and Guan, 2008; Seeman, 2008). These two molecules are close congeners to the Lilly clinical prodrug and its active metabolite. Seeman obtained half maximal effective concentration (EC50) values of 16 to 48 nM for D2 partial agonist effects for these glutamate agonist congeners and showed that they inhibit prolactin release from cultured lactotrophs (Seeman et al., 2008). According to the classic correlation between dissociation constants at D2 receptors and daily clinical antipsychotic doses (Seeman, 2006), the dissociation constants of 20 to 43 nM they measured at D2 for the two glutamate agonists predicts a clinical dose of ~100 mg/day (Seeman, J. Psychiatry Neurosci., 2009, in press). Remarkably, Patil and colleagues administered 80 mg of LY2140023 per day to treat schizophrenia. Thus, according to Seeman, the D2 partial agonist affinity of the LY compounds predicts their clinically effective dose. Maybe so, or perhaps the creation of more mGluR agonist compounds will reveal a correlation between mGluR receptor affinity and daily antipsychotic dose. That would be an innovative graph for psychopharmacology!
Of course, resolving this issue is crucial. A D2 mechanism for the metabotropic agonists from Lilly would question one of the most exciting new targets for schizophrenia in decades (Patil et al., 2007), or at least require a revision that the efficacy of metabotropic receptor agonist treatment of schizophrenia includes a concomitant partial D2 agonism that reduces endogenous dopamine transmission (Seeman and Guan, 2008). At the Neuroscience meeting, teams from both AstraZeneca and Lilly presented important and thorough evaluations that weighed in against the D2 component.
The AstraZeneca team, led by John Zysk, used mass-spec-validated LY354740 and LY379268 to confirm the potent agonist activities for mGluR2 in a [35S]GTPγS assay, but not at hD2long receptors, using [3H]domperidone and CHO cells, as Seeman's group had done. Both compounds failed to displace [3H]-raclopride, [3H]-PHNO, or the important [3H]-domperidone, which more readily reveals agonist occupancy of the D2 high site in filter binding assays (see Seeman and Guan, 2008). Zysk and colleagues found inhibition of [3H]-domperidone binding only at 30 µM, and not at lower concentrations, while Seeman and colleagues found that LY 354,740 occupied D2 and mGluR2 at 14-75 nM (Seeman, 2008). But even using [3H]domperidone binding in normal or hypotonic sodium, which enhances agonist displacement of antagonist radioligand binding, Zysk and colleagues showed that neither LY404039 nor LY2140023 displayed antagonist or partial agonist activity in the [35S]GTPγS assay. Those data are also consistent with the report by Patil and colleagues (2007) that patients treated with the oral prodrug LY2140023 showed no D2 effects, including no change in serum prolactin and no extrapyramidal symptoms.
The evidence against D2 agonism of the Lilly mGluR agonists was strengthened by the abstract from Matthew Fell and colleagues of Lilly (abstract 155.7). Using the very same Lilly compounds used by Seeman and colleagues, as well as compounds with D2 agonism, they found that aripiprazole = domperidone > dopamine >> LY379268 = LY354740 = 0 for in vitro D2 receptor occupancy, using [3H]raclopride. They also found no occupancy of striatal D2 receptors labeled by non-radioactive raclopride (using a clever and sensitive mass spectrometry-based detection method) in the striata of rats dosed with either LY compound, whereas our well-characterized partial D2 agonist aripiprazole occupied D2 receptors at about the expected oral dose of 10 mg/kg. This would appear to be the final blow to the D2 occupancy argument, since a lower potency and lower degree of partial agonism is proposed for the LY compounds than for aripiprazole. So perhaps the ability of Seeman and Guan to show that LY 379,268 inhibited the binding of [3H]PHNO by 50 percent at 1.5 mg/kg, as measured ex vivo, reveals a subtle action of this compound in this particular agonist-preferring assay. In other in vivo assays conducted by Fell and colleagues, neither LY compound altered striatal dopa accumulation after gamma-butyrolactone, in contrast to aripiprazole, 3PPP or quinpirole, all D2 agonists of various persuasions, which lower L-dopa accumulation in many other laboratories.
Mice with deletions of mGluR2/3 receptors show a loss of LY404039 effects against PCP-hyperactivity (Patil et al., 2007). Knockout of the mGluR2 but not the mGlur3 receptor prevents the ability of LY404039 to block amphetamine or PCP hyperactivity; (Fell et al., 2008). The selective D2/3 blocker, raclopride (Fell et al., SfN), and clozapine or risperidone (Fell et al., 2008) are still active in these mutant mice, again excluding a D2 role for the LY compounds.
Both the AstraZeneca and Lilly SfN abstracts concluded that there appears to be a lack of direct dopamine D2 receptor effects of these important mGluR2/3 agonists. Similar findings by other groups, and the preponderance of evidence, seem to support this. Still, it would be prudent to ascertain these non-D2 effects with the actual clinical players in this issue, LY404039 and its prodrug, LY2140023. One is struck with the impressive nM potencies of the LY compounds in the in vitro displacement isotherms in the Seeman papers, the agonist-like decreases of prolactin from lactotrophs, and decreases in in vivo striatal binding of [3H]PHNO by 1.5 mg/kg doses of LY379268 (Seeman et al., 2008a, b). I suggest that someone conduct a human PET study to determine whether [11C]PHNO binding to D2 receptors is competed for by LY 2140023 or aripiprazole, a positive control for this experiment.
Everything else mGluR
Although it was presented on the dreaded afternoon of the last day of this king-sized meeting, when most attendees were trudging home or, more commonly, to a local DC bar, a poster by Phil Seeman and colleagues, University of Toronto, rewarded the stalwarts congregating there with a new perspective on dopamine-glutamate interactions (abstract 823.22). Phil's poster reported evidence that mice lacking either mGlu2 or mGlu3 receptors are supersensitive to dopamine receptor agonists. Hey, isn't that an alternate aspect of the controversy we just reviewed? The Seeman lab again measured the proportion of D2High receptors, labeled by the D2 antagonist [3H]domperidone, which can be potently displaced by the full D2 agonist, dopamine. The proportion of D2High in striatum was increased from 14 percent in wild type mice to 53 percent in mGluR2 or 3-/- mice. Even more impressive was the enhanced dopamine effect in the absence of mGluR2 or 3. The potency of the D2 agonist (+)PHNO to stimulate [35S]GTP-γ-S incorporation in striatal homogenates from mGlu2 receptor knockout mice was increased by 67-fold compared to wild type, and by only 17-fold in mGluR3 knockout mice. We like these kinds of effect sizes! These data suggest that glutamate signaling at mGluR2 and mGluR3 receptors may normally lessen dopamine affinity at D2 receptors. This result suggests that mGlu2/3 receptor agonists could promote antipsychotic activity by decreasing excessive dopamine tone at D2 receptors. Excessive dopamine tone is a long-held view of schizophrenia, based on the consistent ability of D2 antagonism to treat schizophrenia, and by findings of greater D2 occupancy by dopamine in schizophrenia patients through in vivo PET D2 receptor binding studies. That left the few of us still standing at the posters to entertain the possibility that mGluR2 and 3 receptors could play a role in schizophrenia etiology, and its treatment, by augmenting dopamine signaling at D2 receptors when signaling is low, and by lessening it when mGlur2/3 tone is restored, say, by LY404039.
Another receptor action of interest in schizophrenia is alpha-adrenergic antagonism. Continuing from findings first reported in Science for raclopride (Hertel et al., 1999), Torgney Svensson and colleagues proposed that adjunctive α2 adrenoceptor blockade used with risperidone may contribute to a more advantageous effect profile in schizophrenia, particularly in treatment-resistant patients (abstract 55.17). This should allow for improved efficacy due to reduced D2 receptor occupancy and reduction of extrapyramidal side effects. Consistent with this, the addition of the α2 adrenoceptor idazoxan to risperidone, which has little α2 interaction, enhanced suppression of the conditioned avoidance response (CAR), increased dopamine release in the prefrontal cortex, and facilitated the augmentation of prefrontal NMDA transmission by low dose risperidone. A study for the future is whether idazoxan can potentiate the antipsychotic efficacy by mGluR agonist drugs.
Moving to mGluR5 agonism, Laura Gray from Duke University showed that clozapine was still capable of reversing schizophrenia-related behaviors in the mGluR5 knockout mouse (abstract 761.14). The NMDA antagonists, PCP- and MK-801, produce a psychotic-like state in humans and disrupt PPI and decrease cognitive functions. Gray and colleagues identified that a component of these deficiencies may be due to mGluR5 receptor dysfunction. Mice with deletions of the mGluR5 receptor showed increased potency of the selective NMDA antagonist MK-801 to induce hyperactivity. These mice also show abnormal locomotor patterns, reduced prepulse inhibition, and deficits on performance of a short-term spatial memory task on the Y-maze. In this study, chronic treatment of these mGlurR5-less mice with clozapine lessened their hyperactivity and PPI deficits, but did not alter their poorer cognitive performance on the Y maze task. Interestingly, clozapine elevated NMDA-R binding labeled with [3H]MK 801 without altering dopamine D2 ([3H] YM-09151), serotonin 2A ([3H] ketanserin), or muscarinic M1/M4 receptor ([3H]pirenzepine) binding. Clozapine may treat schizophrenia via effects on NMDA receptors, similar to those produced by glutamate activation of mGluR5.
The 2008 SfN meeting revealed that the metabotropic 2, 5, and 3 receptor agonist flags are planted a little deeper into the heart of what may create, and treat, schizophrenia. It's promising for future combination therapies that these approaches are consistent with the D2 antagonist and D2 partial agonist approaches, which to date remain the mainstay of antipsychotic pharmacology.—C. Anthony Altar.