Organizing pains

Update

TRENDS in Neurosciences Vol.27 No.6 June 2004

| Research Focus

Organizing pains Emer M. Garry and Susan M. Fleetwood-Walker Centre for Neuroscience Research, Division of Veterinary Biomedical Sciences, The Royal (Dick) School of Veterinary Studies, The University of Edinburgh, Summerhall, Edinburgh EH9 1QH, UK

Chronic pain is sustained by central neuronal sensitization, with many similar characteristics irrespective of the type of injury incurred. Nevertheless, pain arising from nerve injury (neuropathic pain) is resistant to centrally acting analgesics, whereas inflammatory pain responds well. New research indicates that the role of spinal NMDA receptors in chronic pain depends on adaptor proteins of the membrane-associated guanylate kinase (MAGUK) family and raises the possibility that complexes of different composition might contribute differentially to different pain states. Neuropathic pain involves widespread plastic changes in the anatomy, neurochemistry and function of the nervous system. This occurs via heterosynaptic facilitation from low-frequency activation of diverse inputs involving different transmitters. In contrast to inflammatory pain, many elements of neuropathic pain are intractable to current analgesics. However, each involves the NMDA receptor, which forms a complex of proteins with signalling and structural functions and is crucial in neuronal plasticity both in spinal cord and in other parts of the CNS [1]. Peripheral inflammatory mediators such as prostaglandins represent effective therapeutic targets for treatment of inflammatory pain. No distinct targets have been identified for central neuropathic sensitization and, although tricyclic antidepressants, carbamazepine and gabapentin can be partially effective, their mechanisms of action are unclear. Any distinctions at the molecular level in how these two different types of pain are processed within spinal neurons could point to novel therapeutic targets. NMDA receptor interactions in chronic pain The NMDA subtype of glutamate receptor has two main subunit types: NR1, the key functional subunit, and NR2 subunits, which determine the subtype-specific channel characteristics [2]. The intracellular C terminus of NR2 subunits mediates interactions with membrane-associated guanylate kinase (MAGUK) proteins PSD-95/SAP90, Chapsyn-110/PSD-93, SAP102 and hDlg/SAP97. These MAGUKs contain three N-terminal PDZ domain repeats, a Src homology 3 (SH3) domain and a guanylate kinase (GK)-like domain [3] with 70 –80% sequence homology [4]. Two MAGUKs, PSD-95 and PSD-93, have been the focus of recent research in the pain field. The complex formed by the NMDA receptor and Corresponding author: Susan M. Fleetwood-Walker ([email protected]). www.sciencedirect.com

MAGUK proteins [5] involves an array of signalling and/or scaffolding proteins, in addition to cytoskeletal elements and regulatory kinases and phosphatases that contribute to NMDA receptor regulation, stabilization, presynaptic – postsynaptic apposition and function. Such protein – protein interactions are likely to play key roles in neuronal plasticity, but these differ according to the composition of complexes in different CNS regions and between (or perhaps even within) individual neurons. It might be expected that MAGUKs are important in spinal NMDA receptor function. Experiments designed to disrupt NR2– MAGUK interactions gave variable results in ischaemic and long-term potentiation (LTP) models in hippocampus [6,7]. By contrast, studies on disruption of equivalent AMPA receptor interactions in pain states have shown clear antinociceptive effects [8,9]. New studies using both knockdown and genetic mutant or knockout techniques have examined PSD-95 and PSD-93, both of which can be present in postsynaptic NMDA receptor complexes and emphasize the importance of NMDAreceptor– MAGUK interactions in chronic pain [10,11]. These MAGUKs appear to contribute differentially to the plasticity that underlies chronic pain without affecting normal acute sensory responses, but it is not clear whether their roles are the same in all types of chronic pain. There is no information yet on whether NMDA-receptor– MAGUK interactions are dynamically altered in pain states. Different MAGUKs, different pain? Mutant mice expressing PSD-95 terminated after the first two PDZ domains [12] completely lack the reflex sensitization to light mechanical stimuli that is expected following sensory sciatic nerve injury (neuropathic pain), whereas the sustained phase pain response to formalin remains normal [10]. In PSD-93-knockout mice [13], sensitized inflammatory responses (e.g. responses to complete Freund’s adjuvant, CFA) are completely prevented, whereas part of the sensitization due to spinal nerve injury (SNL) is maintained [11]. So, are different MAGUKs committed to organizing different pains? One important consideration is that these molecular interventions differ, making direct comparison difficult. Nevertheless, knockdown studies with antisense oligonucleotides for PSD-95 [14] and PSD-93 [15] yield similar results. Knockdown of PSD-95 results in reduced sensitivity in a neuropathic pain model (SNL) [14), whereas knockdown of PSD-93 reduced both inflammatory (CFA-induced) and neuropathic (SNL-induced) pain behaviours [15]. Thus, two different nerve injury models, with

Update

TRENDS in Neurosciences Vol.27 No.6 June 2004

genetic [10] or chemical [14] disruption of PSD-95, concur and underline its importance in neuropathic pain. A thorough analysis of different inflammatory pain models under the same conditions will be necessary to ascertain whether the distinctive observations with formalin can be generalized. Similarly, PSD-93 is implicated in both inflammatory and neuropathic pain by genetic and chemical methods. Clarification of whether PSD-95 and PSD-93 play differential roles in particular chronic pain states requires systematic analysis of equivalent mutants, knockouts and antisense-knockdown animals in identical pain models. MAGUK PDZ1 and PDZ2 domains directly bind NR2 subunits and are implicated in receptor localization and synaptic stabilization. NMDA receptor ion channels of PSD-95-knockout mice respond normally to acute stimulation with NMDA [10,12], whereas PSD-93-mutant mice show reduced NMDA-receptor-mediated excitatory transmission [11], so the PSD-93 pain phenotype could result from disruption of NMDA receptor function or PSD-93mediated signalling. The apparent differences in neuropathic and inflammatory sensitization seen between these mouse strains could be partially due to impaired cell surface expression of NR2A and NR2B subunits in the PSD-93-knockout mice [11]. In addition, PSD-95 might restrict NR2B internalization [16]. Again, direct comparisons are difficult at this stage because of the different nature of the molecular defect. Downstream partners of different MAGUKs The NMDA-receptor– PSD-95 complex incorporates Ca2þ/calmodulin-dependent kinase II (CaMKII), which interacts directly with NR2 subunits [10,17]. The sensitivity of NMDA-receptor– PSD-95-dependent neuropathic sensitization to CaMKII inhibitors suggests that this enzyme could play a key role in sensitization downstream of the complex. Indeed, CaMKII can mediate spinal sensitization caused by the activator of nociceptive afferents capsaicin [18], and its expression is increased in an inflammatory pain model [19]. CaMKII coimmunoprecipitation with NR2A and NR2B subunits is specifically increased in a neuropathic pain model, whereas CaMKII antagonists prevent sensitization [10]. In PSD-95-mutant mice, CaMKII association with NR2A and NR2B is reduced [10], perhaps underlying the functional deficit in neuropathic sensitization. It is not known whether PSD-93 can contribute to the same kind of interactions. Association of PSD-95 with NR2A is reported to reduce the receptor interaction with CaMKII [20], so in PSD-95-mutant mice, CaMKII binding would be expected to increase. However, the higher affinity interaction of NR2B with CaMKII might not show the same properties [21]. It is important to establish the conditions under which CaMKII can dock to NMDA receptors in pain models, and what the precise roles of MAGUKs are in assembling or modulating these interactions. Identification of a comprehensive list of PSD-93-binding partners, as for PSD-95, is necessary to facilitate an understanding of how these MAGUKs might be differentially involved in neuropathic and inflammatory pain. The differences are likely to be subtle, because analysis of PDZ domain affinities in PSD-95 and SAP102 for different www.sciencedirect.com

293

peptides shows only modest differences in specificity [22]. Although neuronal nitric oxide synthase (nNOS) has been proposed as a MAGUK-associated signalling protein with a role in pain processing [14], its contribution to nociceptive signalling in neuropathic pain is questionable [23]. Furthermore, nNOS is phosphorylated by CaMKII to reduce its activity [24]. The wide range of potential downstream influences from spinal NMDA-receptor– MAGUK complexes might differ in relative importance from those in other CNS regions. The links and pathways that are functionally significant in different pain states remain to be determined. Different receptor–MAGUK combinations? PSD-93 and PSD-95 have a very subtle distinction in their expression profiles in spinal dorsal horn, in that PSD-95 is expressed in lamina I and lamina IIouter of the superficial dorsal horn [9], whereas PSD-93 expression is localized to lamina IIinner [14]. Thus, both are potentially placed in proximity to the NR2B subunit, which is present in laminae I and II – the ‘hot spot’ for incoming nociceptive transmission, whereas NR2A shows a more diffuse dorsal horn distribution [25]. NR2B-selective antagonists effectively attenuate nerve-injury-induced nociceptive behaviour [25]. By contrast, no NR2A-selective pharmacological agents are known. It is possible that PSD-93 and PSD-95 predominantly associate with NMDA receptor complexes containing different NR2 subtypes. This could be influenced by their coexistence in neurons within different laminae of the spinal cord as well as by rules of molecular assembly. If such segregation of MAGUKs into different complexes were substantiated by experiment, it could represent a feasible basis for different MAGUKs contributing to different forms of chronic pain. To establish this, it would be necessary to demonstrate differences, first in the colocalization of NR2A and NR2B subunits with either PSD-95 or PSD-93 within individual dorsal horn neurons, and second in their coimmunoprecipitation. The recent results implicating MAGUKs in chronic pain states clearly open the way to many lines of further research, which will be required to establish their precise physiological role they play. In the long term, the clinical implications of this work could be important because chronic pain, particularly that arising from nerve injury, is notoriously difficult to treat. New effective painkillers with minimal side effects are urgently needed. These recent reports could point the way to a new strategy; perhaps we should not simply try to stop the messenger – the smart approach might be to restrain the organizer. Acknowledgements We thank Rory Mitchell and Seth Grant for their advice in the preparation of this review. Our laboratory is supported by The Wellcome Trust.

References 1 Ji, R-R. et al. (2003) Central sensitization and LTP: do pain and memory share similar mechanisms? Trends Neurosci. 26, 696– 705 2 Chen, N. et al. (1999) Subtype-dependence of NMDA receptor channel open probability. J. Neurosci. 19, 6844 – 6854 3 Kennedy, M.B. (1997) The postsynaptic density at glutamatergic synapses. Trends Neurosci. 20, 264 – 268

Update

294

TRENDS in Neurosciences Vol.27 No.6 June 2004

4 Kim, E. and Sheng, M. (1996) Differential Kþ channel clustering activity of PSD-95 and SAP97, two related membrane-associated putative guanylate kinases. Neuropharmacology 35, 993 – 1000 5 Husi, H. et al. (2000) Proteomic analysis of NMDA receptor – adhesion protein signaling complexes. Nat. Neurosci. 3, 661 – 669 6 Aarts, M. et al. (2002) Treatment of ischemic brain damage by perturbing NMDA receptor – PSD-95 protein interactions. Science 298, 846 – 850 7 Lim, I.A. et al. (2003) Disruption of the NMDA receptor – PSD-95 interaction in hippocampal neurons with no obvious physiological short-term effect. Neuropharmacology 45, 738 – 754 8 Li, P. et al. (1999) AMPA receptor– PDZ interactions in facilitation of spinal sensory synapses. Nat. Neurosci. 2, 972 – 977 9 Garry, E.M. et al. (2003) Specific involvement in neuropathic pain of AMPA receptors and adapter proteins for the GluR2 subunit. Mol. Cell. Neurosci. 24, 10 – 22 10 Garry, E.M. et al. (2003) Neuropathic sensitization of behavioral reflexes and spinal NMDA receptor/CaM kinase II interactions are disrupted in PSD-95 mutant mice. Curr. Biol. 13, 321 – 328 11 Tao, Y.X. et al. (2003) Impaired NMDA receptor-mediated postsynaptic function and blunted NMDA receptor-dependent persistent pain in mice lacking postsynaptic density-93 protein. J. Neurosci. 23, 6703 – 6712 12 Migaud, M. et al. (1998) Enhanced long-term potentiation and impaired learning in mice with mutant postsynaptic density-95 protein. Nature 396, 433 – 439 13 McGee, A.W. et al. (2001) PSD-93 knock-out mice reveal that neuronal MAGUKs are not required for development or function of parallel fiber synapses in cerebellum. J. Neurosci. 21, 3085– 3091 14 Tao, Y.X. et al. (2000) Expression of PSD-95/SAP90 is critical for N-methyl-D-aspartate receptor-mediated thermal hyperalgesia in the spinal cord. Neuroscience 98, 201 – 206 15 Zhang, B. et al. (2003) Effect of knock down of spinal cord PSD-93/ chapsin-110 on persistent pain induced by complete Freund’s adjuvant and peripheral nerve injury. Pain 106, 187 – 196

16 Roche, K.W. et al. (2001) Molecular determinants of NMDA receptor internalization. Nat. Neurosci. 4, 794 – 802 17 Gardoni, F. et al. (1998) Calcium/calmodulin-dependent protein kinase II is associated with NR2A/B subunits of NMDA receptor in postsynaptic densities. J. Neurochem. 71, 1733 – 1741 18 Fang, L. et al. (2002) Calcium-calmodulin-dependent protein kinase II contributes to spinal cord central sensitization. J. Neurosci. 22, 4196– 4204 19 Carlton, S.M. (2002) Localization of CaMKIIa in rat primary sensory neurons: increase in inflammation. Brain Res. 947, 252 – 259 20 Gardoni, F. et al. (2001) Hippocampal synaptic plasticity involves competition between Ca2þ/calmodulin-dependent protein kinase II and postsynaptic density 95 for binding to the NR2A subunit of the NMDA receptor. J. Neurosci. 21, 1501 – 1509 21 Mayadevi, M. et al. (2002) Sequence determinants on the NR2A and NR2B subunits of NMDA receptor responsible for specificity of phosphorylation by CaMKII. Biochim. Biophys. Acta 1598, 40 – 45 22 Lim, I.A. et al. (2002) Selectivity and promiscuity of the first and second PDZ domains of PSD-95 and synapse-associated protein 102. J. Biol. Chem. 277, 21697 – 21711 23 Luo, Z.D. et al. (1999) Neuronal nitric oxide synthase mRNA upregulation in rat sensory neurons after spinal nerve ligation: lack of a role in allodynia development. J. Neurosci. 19, 9201– 9208 24 Watanabe, Y. et al. (2003) Post-synaptic density-95 promotes calcium/calmodulin-dependent protein kinase II-mediated Ser847 phosphorylation of neuronal nitric oxide synthase. Biochem. J. 372, 465 – 471 25 Boyce, S. et al. (1999) Selective NMDA NR2B antagonists induce antinociception without motor dysfunction: correlation with restricted localisation of NR2B subunit in dorsal horn. Neuropharmacology 38, 611 – 623 0166-2236/$ - see front matter q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.tins.2004.03.014

Neuron – glia interactions clarify genetic – environmental links in mental illness Akira Sawa1,2, Mikhail V. Pletnikov1 and Atsushi Kamiya1 1 Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine, 600 N. Wolfe Street, Meyer 2-181, Baltimore, MD 21287, USA 2 Department of Neuroscience, Johns Hopkins University School of Medicine, 600 N. Wolfe Street, Meyer 2-181, Baltimore, MD 21287, USA

How genes and environment interface to generate major psychiatric disorders such as schizophrenia has been puzzling, as are the relative roles of neurons and glia in such disturbances. Tomonaga and colleagues have recently reported striking neurobehavioral abnormalities in mice expressing Borna disease virus phosphoprotein (BDV-P) selectively in glial cells. The study provides a novel approach of linking environmental and genetic factors to behavior by producing genetically engineered mice. The key role for glial BDV-P implicates neuron –glia interactions in the pathogenesis of psychiatric conditions. Psychiatric conditions such as schizophrenia and mood disorders reflect a combination of genetic and environmental Corresponding author: Akira Sawa ([email protected]). www.sciencedirect.com

factors [1]. Recent advances in human genetics have revealed candidate susceptibility genes for schizophrenia. Because even in identical twins the concordance for schizophrenia and mood disorders is only 40 – 50%, a contribution of environmental factors has been long considered likely and infection by viruses and other agents have been implicated in the etiopathogenesis of schizophrenia [2 –6]. For example, cytomegalovirus, herpes simplex virus 1, influenza virus, poliovirus, Toxoplasma gondii, rubella and Borna disease virus (BDV) have been studied as candidate agents using epidemiological, immunological and neuropathological approaches [3– 6]. Genetic factors in schizophrenia can be examined, at least in part, using genetically engineered mice. However, it has been difficult to pinpoint causal links between viral infections and neurobehavioral disturbances. Furthermore, lack of reactive astrocytosis, a sign of viral infection, in