Nerve agents have profound effects on brain neurotransmission, mediated via direct effects on ligand-gated receptors and voltage-gated ion channels that are controlled by muscarinic-type and nicotinic-type cholinergic receptors and by indirect modulation of multiple downstream intracellular neuronal signaling pathways (Shih & McDonough, 1997
). Here, we have examined the coordinated effects of the nerve agent DFP on multiple signaling pathways in mouse hippocampus and striatum in vivo
using CNSProfile to monitor the state of phosphorylation of neuronal signaling proteins. Common patterns of protein phosphorylation changes were evident after DFP treatment of two different mouse strains (C57BL/6 and FVB) with different sensitivities to the nerve agent. Because these changes correlate with onset of CNS symptoms of nerve agent toxicity they represent important signaling targets for nerve agents that will be useful for the development of more effective treatments to block or attenuate short-term and long-term nerve agent effects.
Female C57BL/6 mice displayed dose-dependent increases in seizure-like behavior in response to DFP within 5 min after nerve agent administration, often culminating in death within 20–30 min. Male FVB mice of a similar age and body weight also developed seizure-like behaviors rapidly (within 5–10 min) after DFP injections, but exhibited sustained seizure-like symptoms for several hours with lower overall lethality. In both mouse strains, DFP exposure elicited comparable site- and region-specific effects on phosphorylation of several signaling phosphoproteins in the brain that correlated with the onset of the most severe seizure-related behaviors. Phosphorylation site changes were typically observed by 15 min in the female C57BL/6 mouse brains, whereas most phosphorylation changes in the brains of male FVB were most pronounced at 2h after nerve agent exposure.
A major effect of DFP exposure in mice is the alteration of the state of phosphorylation of regulatory residues on glutamate receptors, including S897 of the NR1 NMDA receptor subunit (Tingley et al., 1997
). These data are consistent with reports that nerve agents induce a sequential activation of distinct neurochemical systems in the brain resulting in a delayed recruitment of glutamatergic neurons (Shih & McDonough; 1997
; Shih et al., 2003
). A rapid reduction was seen in the level of NR1 phosphorylated at the S897 residue in mouse striatum at the earliest time point monitored after DFP exposure (15 min in female C57BL/6 mice and 30 min in male FVB mice). Previous work from our laboratory (Snyder et al., 1998
) has shown that the phosphorylation state of S897 on NR1 in striatum is under the control of a PKA-dependent signaling cascade that is reciprocally regulated by both dopamine and glutamate neurotransmission. Phosphorylation of NR1 S897 accentuates NMDA receptor signaling, increasing gene transcription involving CREB (Dudman et al., 2003
) and reducing receptor removal from the plasma membrane (Scott et al., 2003
). We interpret the profound dephosphorylation of striatal S897 NR1 observed after DFP exposure as a signal subsequent to elevated glutamatergic activity which occurs as the delayed response to the nerve agent. Dephosphorylation of this site in response to glutamate overactivity could be anticipated to dampen glutamate effects by attenuating gene expression effects via CREB, and reducing receptors in the plasma membrane.
In contrast, NR1 phosphorylation in hippocampus was upregulated after DFP exposure. S897 phosphorylation was elevated by 75% in hippocampus, relative to vehicle-treated control mice. The biochemical basis for the bi-directional regulation of S897 phosphorylation in these two brain regions is unclear. The increase in NR1 phosphorylation state was delayed until 2 h after DFP treatments, compared to striatal NR1, which was significantly dephosphorylated at 30 min after DFP exposure. One plausible hypothesis for the difference in NMDA receptor phosphorylation in these two brain regions is that rapid activation of protein phosphatase-1, governed by the DARPP-32 cascade, leads to striatal dephosphorylation of NR1. Since DARPP-32 is enriched in striatal medium spiny-type neurons, but is essentially undetectable in hippocampus, this would explain the distinct hippocampal pattern (Ouimet et al., 1984
) Future studies comparing NR1 phosphorylation in response to DFP in wildtype versus DARPP-32 knockout mice will be useful in testing this hypothesis. In addition to the effects of NR1 phosphorylation, increased phosphorylation of GluR1 residues is believed to enhance AMPA receptor activity (Derkach et al., 1997). A trend toward an increase in phosphorylation at S831 and S845 of the GluR1 AMPA receptor subunit (Barria et al., 1997
; Derkach et al., 1997), but did not reach statistical significance.
DARPP-32 phosphorylation in striatum was selectively increased at a single regulatory site in DFP-exposed mice. Phosphorylation of DARPP-32 at T75 was significantly increased in the striatum in both C57BL/6 and FVB mice exposed to DFP. Cdk5, a member of the cyclin-dependent kinase family that exists in neurons as a neuron-specific Cdk5/p35 complex, is the key kinase phosphorylating this site. Cdk5 has been implicated as a mediator with diverse biochemical and cell biological roles in models of drug addiction, learning and memory and neurodegenerative disease (Fienberg et al., 1998
; Greengard, 2001
). Whether the elevation of T75 DARPP-32 phosphorylation is due to an increase in Cdk5 activity or a reduction in activity of the phosphatase (PP2A) that dephosphorylates T75, the increase in phospho-T75 DARPP-32 levels is also seen with other OP nerve agents, including sarin (i.e., G.L. Snyder, T.M. Shih, and J. McDonough, unpublished observations and G.L. Snyder and H. van Helden, unpublished observations). The induction of Cdk5-mediated phosphorylation after DFP (and sarin) represents an early potentially causative factor in long-term neuronal damage. Damage to neurons leads to calpain-induced increases in activated Cdk5, and results in tau phosphorylation reminiscent of Alzheimer’s disease pathology (Ahlijanian et al., 2000
). However, DFP exposure in the present study did not lead to increased expression of well-characterized markers of neuronal damage, including the glial fibrillary acidic protein (GFAP) (J. P. O’Callaghan and D.B. Miller, unpublished observations). The observed increases in T75 DARPP-32 are apparently indicative of early-steps toward neuronal damage. Increases in T75 DARPP-32 phosphorylation levels of similar magnitude to those seen in this study have been found to result from exposure to drugs of abuse (Bibb et al., 1999
; Norrholm et al., 2003
), and appear to accompany a process of synaptic re-organization that occurs after drug exposure. For example, repeated exposure to cocaine, which can lead to long-lasting drug craving, also results in the remodeling of synaptic spines, a process which is dependent upon enhanced activity in Cdk5-dependent signaling pathways (Norrholm et al., 2003
). In fact, excessive stimulation of NMDA receptors (such as that evoked by exposure to nerve agents) has been shown to increase brain levels of the constitutively-active Cdk5 cofactor, p25, and to result in persistent motor in-coordination and impaired learning in mice (Meyer et al., 2008
). It would be tempting to speculate, then, that increased Cdk5 activity, as monitored by elevated phospho-T75-DARPP-32 levels, may be responsible for initiating synaptic changes responsible for the persistent deficits in motor behavior and cognitive performance characteristic of animals and humans surviving nerve agent-related seizures (Shih et al., 2003
; Miyaki et al., 2005
; Suzuki et al., 1997
In addition to DARPP-32, exposure of mice to DFP resulted in a robust increase in phosphorylation of synapsin I, a presynaptic vesicle-associated protein, in both striatum and hippocampus. Synapsin I is enriched in nerve terminals throughout the brain (DeCamilli et al., 1983
). Different functional properties of the protein appear to be mediated via distinct phosphorylation sites that are controlled by different protein kinases/protein phosphatases (Matsubara et al., 1996
; Jovanovic et al., 1996
; Jovanovic et al., 2001
). DFP exposure selectively increased phosphorylation of synapsin I at S549, a Cdk5 substrate (Matsubara et al., 1996
; Yamagata et al., 2002
) that has been shown to affect interactions of the protein with the cytoskeleton mediated via F actin (Jovanovic et al., 2001
) while having no effect on phosphorylation state of a CaMKII-dependent residue, S603 (Yamagata et al., 2002
), which may control interactions of synapsin with synaptic vehicles. Though direct measures of Cdk5 activity were beyond the scope of the present study, the data support the idea that nerve agent exposure may promote increases in Ckd5 in multiple brain regions (e.g., striatum and hippocampus) and in both presynaptic nerve terminals and in specific post-synaptic neuron populations (e.g., striatal medium spiny neurons containing DARPP-32).
The seizure-like behaviors and protein phosphorylation changes evoked by DFP treatment are differentially affected by distinct classes of cholinergic ligands; both outcomes are substantially reversed by a non-selective muscarinic cholinergic antagonist. Pre-treatment of mice with phencynonate hydrochloride (PCH), a muscarinic cholinergic antagonist (Wang et al., 2005a
) reduced seizure-like behaviors and reversed DFP-induced changes in phosphorylation. These data complement and extend a previous study reporting that PCH, at similar dose levels used in the present study, effectively reduces seizure activity in rats elicited by the nerve agent, soman (Wang et al., 2005b
). The salient biochemical effects of DFP exposure, including the striatal and hippocampal changes in the phosphorylation state of S897 NR1 and S549 synapsin, are effectively mimicked in vivo
by treatment of mice with pharmacological doses of an AChE inhibitor therapeutic agent used for Alzheimer’s disease, donepezil, and by a non-selective muscarinic receptor agonist, oxotremorine. The data support the idea that these phosphorylation changes reflect a cellular biochemical response to increased muscarinic receptor activation that is a secondary response to the increased synaptic levels of acetylcholine occurring after DFP exposure. Thus, increased neurotransmission via muscarinic cholinergic receptors appears sufficient to mediate many of the biochemical and behavioral effects of DFP.
The PCH data support a prominant role for muscarinic acetylcholine receptors in the induction of seizure-like behaviors and the associated biochemical changes seen after nerve agent exposure. The data are consistent with previous work that identified M1-subclass of receptors as the relevant muscarinic receptors mediating seizure activity resulting from cholinergic stimulation (Bymaster et al., 2003
). We further explored the importance of M1 receptor-preferring agents by testing whether pretreatment with a high affinity muscarinic receptor allosteric modulator prior to DFP might protect against the effects of the nerve agent. AC-260584, a partial allosteric modulator with partial agonist properties, preferentially binds M1-type muscarinic cholinergic receptors at a binding site distinct from the binding site for acetylcholine (Vanover et al., 2008
). A recent report by Conn and colleagues has demonstrated that allosteric regulators of M1 receptors can preferentially affect some signal transduction responses to M1 receptor stimulation, while having no effect on other M1 receptor pathways (Conn et al., 2008; Marlo et al., 2009
). We reasoned that such agents might stabilize muscarinic receptor activities associated with seizure generation, and thus, preclude the behavioral and/or biochemical effects of DFP. Pretreatment of mice with AC-260584, however, failed to protect against both the appearance of seizure-like behaviors and the protein phosphorylation changes seen in DFP–treated mice, as measured 2h after DFP exposure. Interestingly, mice pretreated with AC-260584 did display an apparent delay in onset of seizure-like behavior (as measured at 15 min after DFP) and a reduced morbidity after DFP, compared with mice treated with DFP alone.
Varenicline, a potent partial agonist at the α4β2 subclass of nicotinic cholinergic receptors (Coe et al., 2005
), which is thought to functionally desensitize and inactivate nicotinic receptors (many of them located on pre-synaptic nerve terminals), was used to test whether manipulation of (presynaptic) nicotinic receptors would attenuate the effects of DFP exposure. Varenicline did not prevent DFP-induced seizure-like behavior at 15 min or 2h after DFP, but, like AC-260584, did appear to reduce morbidity seen with DFP exposure. Taken together these data support the idea that muscarinic antagonism most effectively blocks both DFP-induced seizure-like behaviors and the biochemical cascades recruited by the nerve agent. However, novel pharmaceutical approaches, including the use of partial agonists that may displace acetylcholine from select receptor subclasses or desensitize select pre-synaptic or post-synaptic cholinergic receptor populations confer neuroprotection meriting further investigation.
In summary, these data reveal selective, region-specific, nerve agent-associated effects on intracellular signaling pathways and phosphoproteins that are reversed by the muscarinic receptor antagonism. The approach identifies specific targets for focused evaluation of novel anti-convulsant mechanisms.