Progressive Assessment of Ischemic Injury to White Matter Using Diffusion Tensor Imaging: A Preliminary Study of a Macaque Model of Stroke.

BACKGROUND: Previous Diffusion Tensor Imaging (DTI) studies have demonstrated the temporal evolution of stroke injury in grey matter and white matter can be characterized by DTI indices. However, it still remains not fully understood how the DTI indices of white matter are altered progressively during the hyperacute (first 6 hours) and acute stage of stroke (≤ 1 week). In the present study, DTI was employed to characterize the temporal evolution of infarction and white matter injury after stroke insult using a macaque model with permanent ischemic occlusion. METHODS AND MATERIALS: Permanent middle cerebral artery (MCA) occlusion was induced in rhesus monkeys (n=4, 10-21 years old). The brain lesion was examined longitudinally with DTI during the hyperacute phase (2-6 hours, n=4), 48 hours (n=4) and 96 hours (n=3) post-occlusion. RESULTS: Cortical infarction was seen in all animals. The Mean Diffusivity (MD) in lesion regions decreased substantially at the first time point (2 hours post stroke) (35%, p <0.05, compared to the contralateral side) and became pseudo-normalized at 96 hours. In contrast, evident FA reduction was seen at 48 hours (39%, p <0.10) post-stroke. MD reduction in white matter bundles of the lesion area was much less than that in the grey matter during the hyper-acute phase but significant change was observed 4 hours (4.2%, p < 0.05) post stroke . Also, MD pseudonormalisation was seen at 96 hours post stroke. There was a significant correlation between the temporal changes of MD in white matter bundles and those in whole lesion areas during the entire study period. Meanwhile, no obvious fractional anisotropy (FA) changes were seen during the hyper-acute phase in either the entire infarct region or white matter bundles. Significant FA alteration was observed in entire lesion areas and injured white matter bundles 48 and 96 hours post stroke. The stroke lesion in grey matter and white matter was validated by pathological findings. CONCLUSION: The temporal evolution of ischemic injury to the grey matter and white matter from 2 to 96 hours after stroke onset was characterized using a macaque model and DTI. Progressive MD changes in white matter bundles are seen from hyperacute phase to acute phase after permanent MCA occlusion and temporally correlated with the MD changes in entire infarction regions. MD reduction in white matter bundles is mild in comparison with that in the grey matter but significant and progressive, indicating it may be useful to detect early white matter degeneration after stroke.

Temporal evolution of ischemic lesions in nonhuman primates: a diffusion and perfusion MRI study.

BACKGROUND AND PURPOSE: Diffusion-weighted imaging (DWI) and perfusion MRI were used to examine the spatiotemporal evolution of stroke lesions in adult macaques with ischemic occlusion. METHODS: Permanent MCA occlusion was induced with silk sutures through an interventional approach via the femoral artery in adult rhesus monkeys (n = 8, 10-21 years old). The stroke lesions were examined with high-resolution DWI and perfusion MRI, and T2-weighted imaging (T2W) on a clinical 3T scanner at 1-6, 48, and 96 hours post occlusion and validated with H&E staining. RESULTS: The stroke infarct evolved via a natural logarithmic pattern with the mean infarct growth rate = 1.38 ± 1.32 ml per logarithmic time scale (hours) (n = 7) in the hyperacute phase (1-6 hours). The mean infarct volume after 6 hours post occlusion was 3.6±2.8 ml (n = 7, by DWI) and increased to 3.9±2.9 ml (n = 5, by T2W) after 48 hours, and to 4.7±2.2ml (n = 3, by T2W) after 96 hours post occlusion. The infarct volumes predicted by the natural logarithmic function were correlated significantly with the T2W-derived lesion volumes (n = 5, r = 0.92, p = 0.01) at 48 hours post occlusion. The final infarct volumes derived from T2W were correlated significantly with those from H&E staining (r = 0.999, p < 0.0001, n = 4). In addition, the diffusion-perfusion mismatch was visible generally at 6 hours but nearly diminished at 48 hours post occlusion. CONCLUSION: The infarct evolution follows a natural logarithmic pattern in the hyperacute phase of stroke. The logarithmic pattern of evolution could last up to 48 hours after stroke onset and may be used to predict the infarct volume growth during the acute phase of ischemic stroke. The nonhuman primate model, MRI protocols, and post data processing strategy may provide an excellent platform for characterizing the evolution of acute stroke lesion in mechanistic studies and therapeutic interventions of stroke disease.

α(2A)-adrenergic receptors filter parabrachial inputs to the bed nucleus of the stria terminalis.

α2-adrenergic receptors (AR) within the bed nucleus of the stria terminalis (BNST) reduce stress-reward interactions in rodent models. In addition to their roles as autoreceptors, BNST α(2A)-ARs suppress glutamatergic transmission. One prominent glutamatergic input to the BNST originates from the parabrachial nucleus (PBN) and consists of asymmetric axosomatic synapses containing calcitonin gene-related peptide (CGRP) and vGluT2. Here we provide immunoelectron microscopic data showing that many asymmetric axosomatic synapses in the BNST contain α(2A)-ARs. Further, we examined optically evoked glutamate release ex vivo in BNST from mice with virally delivered channelrhodopsin2 (ChR2) expression in PBN. In BNST from these animals, ChR2 partially colocalized with CGRP, and activation generated EPSCs in dorsal anterolateral BNST neurons that elicited two cell-type-specific outcomes: (1) feedforward inhibition or (2) an EPSP that elicited firing. We found that the α(2A)-AR agonist guanfacine selectively inhibited this PBN input to the BNST, preferentially reducing the excitatory response in ex vivo mouse brain slices. To begin to assess the overall impact of α(2A)-AR control of this PBN input on BNST excitatory transmission, we used a Thy1-COP4 mouse line with little postsynaptic ChR2 expression nor colocalization of ChR2 with CGRP in the BNST. In slices from these mice, we found that guanfacine enhanced, rather than suppressed, optogenetically initiated excitatory drive in BNST. Thus, our study reveals distinct actions of PBN afferents within the BNST and suggests that α(2A)-AR agonists may filter excitatory transmission in the BNST by inhibiting a component of the PBN input while enhancing the actions of other inputs.

Memory deficits in aging and neurological diseases.

Memory is central to our ability to perform daily life activities and correctly function in society. Improvements in public health and medical treatment for a variety of diseases have resulted in longer life spans; however, age-related memory impairments have been significant sources of morbidity. Loss in memory function is not only associated with aging population but is also a feature of neurodegenerative diseases such as Alzheimer's disease and other psychiatric and neurological disorders. Here, we focus on current understanding of the impact of normal aging on memory and what is known about its mechanisms, and further review pathological mechanisms behind the cause of dementia in Alzheimer's disease. Finally, we discuss schizophrenia and look into abnormalities in circuit function and neurotransmitter systems that contribute to memory impairment in this illness.

Distribution of AMPA receptor subunit glur1 in the bed nucleus of the stria terminalis and effect of stress.

The brain circuitry thought to be involved in stress responses includes several nuclei of the extended amygdala. The bed nucleus of the stria terminalis (BNST) is thought to be involved in the generation of sustained, nonspecific anxiety. Previous behavioral and electrophysiological experiments demonstrate that glutamate systems are involved in anxiety-like behaviors in the BNST. Antagonists for AMPA receptors injected into the BNST decrease anxiety-like behaviors. However, little is known about the role of AMPA receptors and the mechanism by which they act in the establishment of anxiety-like behavior in response to a stressor. We hypothesized that the distribution of AMPA receptors is changed following a paradigm of unpredictable footshock as has been seen in the basolateral amygdala (BLA). We examined the subcellular localization of the GluR1 subunits of the AMPA receptor. We found that the neuropil of the BNST had a lower density of dendritic spines compared to dendritic shafts in the BLA. The majority of elements immunolabeled for GluR1 were dendritic shafts and spines with axonal and glial elements rarely labeled. Compared with controls, no significant effect was observed on days 1, 6, or 14 poststress. However, there was a trend for an increase at 6 and 14 days poststress. These data demonstrate that GluR1 subunits are primarily located on postsynaptic elements in the BNST. Moreover, it was shown that the response of the AMPA GluR1 subunit does not undergo a significant migration into spines from dendrites in response to a stressor as has been demonstrated in the BLA.

Distribution and functional expression of Kv4 family α subunits and associated KChIP ß subunits in the bed nucleus of the stria terminalis.

Regulation of BNSTALG neuronal firing activity is tightly regulated by the opposing actions of the fast outward potassium current, IA , mediated by α subunits of the Kv4 family of ion channels, and the transient inward calcium current, IT . Together, these channels play a critical role in regulating the latency to action potential onset, duration, and frequency, as well as dendritic back-propagation and synaptic plasticity. Previously we have shown that Type I-III BNSTALG neurons express mRNA transcripts for each of the Kv4 α subunits. However, the biophysical properties of native IA channels are critically dependent on the formation of macromolecular complexes of Kv4 channels with a family of chaperone proteins, the potassium channel-interacting proteins (KChIP1-4). Here we used a multidisciplinary approach to investigate the expression and function of Kv4 channels and KChIPs in neurons of the rat BNSTALG . Using immunofluorescence we demonstrated the pattern of localization of Kv4.2, Kv4.3, and KChIP1-4 proteins in the BNSTALG . Moreover, our single-cell reverse-transcription polymerase chain reaction (scRT-PCR) studies revealed that mRNA transcripts for Kv4.2, Kv4.3, and all four KChIPs were differentially expressed in Type I-III BNSTALG neurons. Furthermore, immunoelectron microscopy revealed that Kv4.2 and Kv4.3 channels were primarily localized to the dendrites and spines of BNSTALG neurons, and are thus ideally situated to modulate synaptic transmission. Consistent with this observation, in vitro patch clamp recordings showed that reducing postsynaptic IA in these neurons lowered the threshold for long-term potentiation (LTP) induction. These results are discussed in relation to potential modulation of IA channels by chronic stress.

The new small-molecule mixed-lineage kinase 3 inhibitor URMC-099 is neuroprotective and anti-inflammatory in models of human immunodeficiency virus-associated neurocognitive disorders.

Human immunodeficiency virus (HIV)-associated neurocognitive disorders (HAND) is a significant source of disability in the HIV-infected population. Even with stringent adherence to anti-retroviral therapy, >50% of patients living with HIV-1 will develop HAND (Heaton et al., 2010). Because suppression of viral replication alone is not enough to stop HAND progression, there is a need for an adjunctive neuroprotective therapy in this population. To this end, we have developed a small-molecule brain-penetrant inhibitor with activity against mixed-lineage kinase 3 (MLK3), named URMC-099. MLK3 activation is associated with many of the pathologic hallmarks of HAND (Bodner et al., 2002, 2004; Sui et al., 2006) and therefore represents a prime target for adjunctive therapy based on small-molecule kinase inhibition. Here we demonstrate the anti-inflammatory and neuroprotective effects of URMC-099 in multiple murine and rodent models of HAND. In vitro, URMC-099 treatment reduced inflammatory cytokine production by HIV-1 Tat-exposed microglia and prevented destruction and phagocytosis of cultured neuronal axons by these cells. In vivo, URMC-099 treatment reduced inflammatory cytokine production, protected neuronal architecture, and altered the morphologic and ultrastructural response of microglia to HIV-1 Tat exposure. In conclusion, these data provide compelling in vitro and in vivo evidence to investigate the utility of URMC-099 in other models of HAND with the goal of advancement to an adjunctive therapeutic agent.

Schizophrenia: causes and treatments.

Schizophrenia is a major mental illness that is characterized by psychosis, apathy, social withdrawal and cognitive impairment. These abnormalities in patients results in impaired functioning in work, school, parenting, self-care, independent living, interpersonal relationships, and leisure. Although the search for the biological correlates of schizophrenia has met with limited success, new advances in genetics and pharmacology are promising. Here, we describe the symptoms, causes, diagnosis, strategies for treatment, and clinical impact of the currently available medications.

Ultrastructure of microglia-synapse interactions in the HIV-1 Tat-injected murine central nervous system.

The destruction of normal synaptic architecture is the main pathogenetic substrate in HIV-associated neurocognitive disorder (HAND), but the sequence of cellular events underlying this outcome is not completely understood. Our recent work in a mouse model of HAND using a single intraparenchymal injection of the HIV-1 regulatory protein trans-activator of transcription revealed increased microglial phagocytosis that was accompanied by an increased release of pro-inflammatory cytokines and elimination of dendritic spines in vivo, thus suggesting that microglia-synapse interactions could be dysregulated in HAND. Here, we further examine the relationships between microglia and synaptic structures in our mouse model, at high spatial resolution using immunocytochemical electron microscopy. Our ultrastructural analysis reveals the prevalence of putative microglial filopodial protrusions, which are targeting excitatory and inhibitory synapses, some of which contain phagocytic inclusions at various distances from their distal extremities to the microglial cell bodies. These observations thus suggest that cell-to-cell contacts mediated by microglial filopodia might be a crucial preliminary step in the elimination of synaptic structures in a neuroinflammatory milieu that occurs in HAND.

Presynaptic muscarinic M(2) receptors modulate glutamatergic transmission in the bed nucleus of the stria terminalis.

The anterolateral cell group of the bed nucleus of the stria terminalis (BNST(ALG)) serves as an important relay station in stress circuitry. Limbic inputs to the BNST(ALG) are primarily glutamatergic and activity-dependent changes in this input have been implicated in abnormal behaviors associated with chronic stress and addiction. Significantly, local infusion of acetylcholine (ACh) receptor agonists into the BNST trigger stress-like cardiovascular responses, however, little is known about the effects of these agents on glutamatergic transmission in the BNST(ALG). Here, we show that glutamate- and ACh-containing fibers are found in close association in the BNST(ALG). Moreover, in the presence of the acetylcholinesterase inhibitor, eserine, endogenous ACh release evoked a long-lasting reduction of the amplitude of stimulus-evoked EPSCs. This effect was mimicked by exogenous application of the ACh analog, carbachol, which caused a reversible, dose-dependent, reduction of the evoked EPSC amplitude, and an increase in both the paired-pulse ratio and coefficient of variation, suggesting a presynaptic site of action. Uncoupling of postsynaptic G-proteins with intracellular GDP-ß-S, or application of the nicotinic receptor antagonist, tubocurarine, failed to block the carbachol effect. In contrast, the carbachol effect was blocked by prior application of atropine or M(2) receptor-preferring antagonists, and was absent in M(2)/M(4) receptor knockout mice, suggesting that presynaptic M(2) receptors mediate the effect of ACh. Immunoelectron microscopy studies further revealed the presence of M(2) receptors on axon terminals that formed asymmetric synapses with BNST neurons. Our findings suggest that presynaptic M(2) receptors might be an important modulator of the stress circuit and hence a novel target for drug development.

Molecular mechanisms of working memory.

Working memory is a process for temporary active maintenance of information and the role of prefrontal cortex in this memory has been known since the pioneering experiments of Fulton in the early 20th century. Sustained firing of prefrontal neurons during the delay period is considered the neural correlate of working memory. Evidence in literature suggests the involvement of areas beyond the frontal lobe and illustrate that working memory involves parallel, distributed neuronal networks. Prefrontal cortex is part of a complex neural circuit that includes both cortical and subcortical components and many of these regions play vital roles in working memory function. In this article, we review the current understanding of the neural mechanisms of memory maintenance in the brain.

Chronic, constant-rate, gastric drug infusion in nontethered rhesus macaques (Macaca mulatta).

As part of a study of antipsychotic drug treatment in monkeys, we developed a technique to provide chronic, constant-rate, gastric drug infusion in nontethered rhesus macaques. This method allowed us to mimic the osmotic release oral delivery system currently used in humans for continuous enteral drug delivery. Rhesus macaques (n = 5) underwent gastric catheter placement by laparotomy. After the catheters were secured to the stomach, the remaining catheter length was exited through the lateral abdomen, tunneled subcutaneously along the back, and connected to a 2-mL osmotic pump enclosed in a subcutaneous pocket. Osmotic pumps were changed every 2 to 4 wk for 1 y and remained patent for the duration of the study. Four complications (including cutting of the catheter, incisional dehiscence at the pump site, and loss of 1 catheter into the abdominal cavity requiring catheter replacement) occurred among the 80 pump changes performed during the year-long study. At necropsy, histopathologic examination of the catheter implant sites revealed mild changes consistent with a foreign-body reaction. Our results indicate that the gastric catheter and osmotic pump system was well tolerated in rhesus macaques for as long as 12 mo after placement and suggest that this system will be an attractive option for use in studies that require chronic, constant-rate, gastric drug infusion in nontethered monkeys.

Distribution of D1 and D5 dopamine receptors in the primate and rat basolateral amygdala.

Dopamine, acting at the D1 family receptors (D1R) is critical for the functioning of the amygdala, including fear conditioning and cue-induced reinstatement of drug self administration. However, little is known about the different contributions of the two D1R subtypes, D(1) and D(5). We identified D(1)-immunoreactive patches in the primate that appear similar to the intercalated cell masses reported in the rodent; however, both receptors were present across the subdivisions of the primate amygdala including the basolateral amygdala (BLA). Using immunoelectron microscopy, we established that both receptors have widespread distributions in BLA. The D1R subtypes colocalize in dendritic spines and terminals, with D(1) predominant in spines and D(5) in terminals. Single-cell RT-PCR confirmed that individual BLA projection neurons express both D(1) and D(5) mRNA. The responses of primate BLA neurons to dopamine and D1R drugs were studied using in vitro slices. We found that responses were similar to those previously reported in rat BLA neurons and included a mixture of postsynaptic and presynaptic actions. We investigated the distribution of D1R in the rat BLA and found that there were similarities between the species, such as more prominent D(5) localization to presynaptic structures. The higher affinity of D(5) for dopamine suggests that presynaptic actions may predominate in the BLA at low levels of dopamine, while postsynaptic effects increase and dominate as dopaminergic drive increases. The results presented here suggest a complex action of dopamine on BLA circuitry that may evolve with different degrees of dopaminergic stimulation.

Localization of myocyte enhancer factor 2 in the rodent forebrain: regionally-specific cytoplasmic expression of MEF2A.

The transcription factor myocyte enhancer factor 2 (MEF2) is expressed throughout the central nervous system, where four MEF2 isoforms play important roles in neuronal survival and differentiation and in synapse formation and maintenance. It is therefore somewhat surprising that there is a lack of detailed information on the localization of MEF2 isoforms in the mammalian brain. We have analyzed the regional, cellular, and subcellular expression of MEF2A and MEF2D in the rodent brain. These two MEF2 isoforms were co-expressed in virtually all neurons in the cortex and the striatum, but were not detected in astrocytes. MEF2A and MEF2D were localized to the nuclei of neurons in many forebrain areas, consistent with their roles as transcriptional regulators. However, in several subcortical sites we observed extensive cytoplasmic expression of MEF2A but not MEF2D. MEF2A was particularly enriched in processes of neurons in the lateral septum and bed nucleus of the stria terminalis, as well as in several other limbic sites, including the central amygdala and paraventricular nuclei of the hypothalamus and thalamus. Ultrastructural examination similarly revealed MEF2A-ir in axons and dendrites as well as MEF2A-ir nuclei in the lateral septum and bed nucleus of the stria terminalis neurons. This study demonstrates for the first time extensive cytoplasmic localization of a MEF2 transcription factor in the mammalian brain in vivo. The extranuclear localization of MEF2A suggests novel roles for MEF2A in specific neuronal populations.

Is the loss of thalamostriatal neurons protective in parkinsonism?

Neuronal loss in Parkinson's disease (PD) is more widespread than originally thought. Among the extrastriatal sites in which significant loss of neurons has been reported is the centremedian-parafascicular (CM-PF) complex of the thalamus, which provides one of the three major afferent sources to the striatum. The functional significance of CM-PF loss in PD is unclear. Interestingly, several recent small trials have suggested that deep brain stimulation of the CM-PF improves motor function in PD. We discuss the possible transsynaptic determination of CM-PF loss secondary to nigrostriatal dopamine degeneration, and suggest that expression of the glycoprotein cerebellin1 (Cbln1) in CM-PF neurons may play an important role in striatal synaptic remodeling in parkinsonism.

Dopamine D1 and D5 receptors are localized to discrete populations of interneurons in primate prefrontal cortex.

Working memory (WM) is a core cognitive process that depends upon activation of D1 family receptors (D1R) and inhibitory interneurons in the prefrontal cortex (PFC). D1R are comprised of the D(1) and D(5) subtypes, and D(5) has a 10-fold higher affinity for dopamine. Parvalbumin (PV) and calretinin (CR) are 2 interneuron populations that are differentially affected by D1R stimulation and have discrete postsynaptic targets, such that PV interneurons provide strong inhibition to pyramidal cells, whereas CR interneurons inhibit other interneurons. The distinct properties of both the D1R and interneuron subtypes may contribute to the "inverted-U" relationship of D1R stimulation and WM ability. To determine the prevalence of D(1) and D(5) in PV and CR interneurons, we performed quantitative double-label immunoelectron microscopy in layer III of macaque area 9. We found that D(1) was the predominant D1R subtype in PV interneurons and was found mainly in dendrites. In contrast, D(5) was the predominant D1R subtype in CR interneurons and was found mainly in dendrites. Integrating these findings with previously published electrophysiological data, we propose a circuitry model as a framework for understanding the inverted-U relationship between dopamine stimulation of D1R and WM performance.

Quantification of D1 and D5 dopamine receptor localization in layers I, III, and V of Macaca mulatta prefrontal cortical area 9: coexpression in dendritic spines and axon terminals.

D1 family receptors (D1R) in prefrontal cortex (PFC) are critical for normal cognition and are implicated in pathological states such as schizophrenia. The two D1R subtypes, D1 and D5, cannot be pharmacologically distinguished but have important functional differences. To understand their contributions to cortical function, we quantified their localization in the neuropil of primate PFC. We identified different patterns of distribution for the two receptors that showed variation across cortical laminae. Although D1 was enriched in spines and D5 in dendrites, there was considerable overlap in their distribution within neuronal compartments. To determine whether the D1 and D5 receptors are localized to separate populations of synapses, we employed double-labeling methods. We found the two receptors colocalized and quantified the overlap of their distribution in spines and axon terminals of prefrontal cortical area 9 in the Macaca mulatta monkey. The two receptors are found in partially overlapping populations, such that the D5 receptor is found in a subpopulation of those spines and terminals that contain D1. These results indicate that dopamine activation of the two D1R subtypes does not modulate disparate populations of synapses onto dendritic spines in prefrontal cortical area 9; rather, dopamine can activate D1 and D5 receptors on the same spines, plus an additional group of spines that contains only D1. The implications of these results for the dose-dependent relationship between D1R activation and PFC function are discussed.

Subcellular distribution of the Rho-GEF Lfc in primate prefrontal cortex: effect of neuronal activation.

The strength of synaptic connections in the brain varies with activity, and this plasticity depends on remodeling of the actin cytoskeleton in dendritic spines. Critical to this are the Rho family GTPases, whose activity is controlled by various modulatory proteins, including the Rho-GEF Lfc. In cultured neurons and nonneuronal cells, Lfc has been shown both to bind to microtubules and to regulate the actin cytoskeleton. Significantly, Lfc was found to be concentrated in the dendritic shafts of cultured hippocampal neurons under control conditions but then translocated into spines when neural activity was stimulated. In this study, we used immunohistochemistry and electron microscopy to examine activity-dependent changes in the distribution of Lfc in the neuropil of monkey prefrontal cortex. We found that, although Lfc was concentrated in dendrites, it also had a complex distribution in the neuropil, including being present in spines, axons, terminals, and glial processes. Moreover, Lfc distribution varied in different layers of cortex. By using an in vitro slice preparation of monkey prefrontal cortex, we demonstrated an activity-dependent translocation of Lfc from dendritic shafts to spines. The results of this study support a role for Lfc in activity-dependent spine plasticity and demonstrate the feasibility of studying activity-dependent changes in protein localization in tissue slices.

Analysis of NR3A receptor subunits in human native NMDA receptors.

NR3A, representing the third class of NMDA receptor subunits, was first studied in rats, demonstrating ubiquitous expression in the developing central nervous system (CNS), but in the adult mainly expressed in spinal cord and some forebrain nuclei. Subsequent studies showed that rodent and non-human primate NR3A expression differs. We have studied the distribution of NR3A in the human CNS and show a widespread distribution of NR3A protein in adult human brain. NR3A mRNA and protein were found in all regions of the cerebral cortex, and also in the subcortical forebrain, midbrain and hindbrain. Only very low levels of NR3A mRNA and protein could be detected in homogenized adult human spinal cord, and in situ hybridization showed that expression was limited to ventral motoneurons. We found that NR3A is associated with NR1, NR2A and NR2B in adult human CNS, suggesting the existence of native NR1-NR2A/B-NR3A assemblies in adult human CNS. While NR1 and NR2A could only be efficiently solubilized by deoxycholate, NR3A was extracted by all detergents, suggesting that a large fraction is weakly anchored to cell membranes and other proteins. Using size exclusion chromatography we found that just as for NR1, a large fraction of NR3A exists as monomers and dimers, suggesting that these two glycine binding subunits behave similarly with regard to receptor assembly and trafficking.