The present study is the first to analyze levels of the NMDA receptor subunits NR1, NR2A, NR2B, and their anchoring protein PSD-95 in the anterior PFC from subjects diagnosed with MDD. Significant reductions in the expression of NR2A, NR2B, and PSD-95, were observed in depressed subjects as compared to psychiatrically healthy controls. In contrast, the level of the NR1 subunit was unchanged in depressed subjects.
Abnormalities in the NMDA receptor system have been previously observed in postmortem brain tissue from major depressives and suicide victims. Previously, we have reported elevated protein expression of the NR2A subunit in the amygdala (Karolewicz et al., 2008b
) and the NR2C subunit in the locus coeruleus (Karolewicz et al., 2005
) in depressed subjects. Now we demonstrate that NR2A and NR2B proteins are reduced in the anterior PFC in depression. In line with our present study is the recent observation by Beneyto and Meador-Woodruff (2008)
who reported reduced expression of the NR2A transcript in the dorsolateral PFC in depression. Additionally, the same group found a reduction in expression of both NR2A and NR2B transcripts in the perirhinal cortex in depression (Beneyto et al., 2007
). Moreover, specific binding of the NMDA receptor antagonist CGP-39653 was significantly reduced in the anterior PFC from suicide victims (Nowak et al., 1995
) suggesting dysfunction of NMDA receptors in frontal cortices from suicide victims. Taken together, these studies highlight the specific regional distribution of NMDA receptor pathology and support the hypothesis that NMDA receptor signaling is altered in depression.
The abnormal expression of NR2 subunits can change the function of the receptor. It is known that different combinations of specific NR2 and NR1 subunits result in NMDA receptor/channel complexes with different characteristics. For example, NR2A or NR2B-containing receptors exhibit higher responsiveness to glutamate and higher fractional Ca2+
current than do heteromers containing the NR2C or NR2D subunits (Monyer et al., 1994
; Yamakura and Shimoji, 1999
). The majority of the NMDA receptors in the PFC are expressed as heteromeric complexes composed of NR1/NR2A, NR1/NR2B, or NR1/NR2A/NR2B subunits. NR1 subunits are expressed in large excess, and are rapidly degraded when unassembled with NR2 partners (Huh and Wenthold, 1999
). Hence the total number of functional NMDA receptors appears to be controlled by expression of the NR2 subunits. Since lower levels of NR2A and NR2B protein could be translated into reduced numbers of functional receptors, one would speculate that depression is associated with hypofunction of the NMDA receptor in the anterior PFC.
Despite a robust reduction in both NR2A and NR2B immunoreactivity, the NR1 subunit was unchanged in the PFC in depression. This is congruent with our previous observation in the amygdala, locus coeruleus and cerebellum where no changes in the expression of NR1 subunit were observed in depression (Karolewicz et al., 2005
; Karolewicz et al., 2008b
). Similarly, Toro and Deakin (2005)
reported unchanged NR1 immunoreactivity in the orbitofrontal cortex and in subregions of the hippocampal dentate gyrus in depressed subjects. Since the specific properties of NMDA receptors such as binding affinities for agonists and antagonists and differences in conductance properties are shaped by the combination of NR1 with NR2 subunits (Lopez de Armentia and Sah, 2003
; Prybylowski et al., 2002
), the unchanged NR1 immunoreactivity found in this study does not rule out abnormal NMDA receptor expression and/or function in depression.
Our present observation that the PSD-95 is markedly reduced in the anterior PFC in depression is not surprising as this protein plays a crucial role in the trafficking, membrane targeting, and internalization of NMDA receptor complexes. Thus, lower levels of PSD-95 may reflect reduced communication/coupling of the NMDA receptors as well as non-NMDA receptors with intracellular signaling cascades. In contrast, previously we have reported a marked elevation in the levels of PSD-95 protein in the amygdala in depression (Karolewicz et al., 2008b
). Interestingly, the difference between the present findings in the PFC and our previous studies in the amygdala suggests that depression-associated pathology of NMDA/PSD-95 is regionally specific. Other reports that have measured PSD-95 in different brain regions have not described changes in the expression of PSD-95. Unchanged levels of PSD-95 mRNA were previously reported in the dorsolateral PFC and striatum (Kristiansen and Meador-Woodruf, 2005
; Beneyto and Meador-Woodruf, 2008
) or in the hippocampus and orbitofrontal cortex in depression (Toro and Deakin, 2005
). Differences between experimental techniques and subjects characteristics (e.g. medication exposure) are key factors that may be associated with discrepancies between postmortem studies.
Alterations in NMDA receptor expression can be a consequence of altered glutamate levels in brains of depressed subjects. In fact, several lines of evidence indicate reduced (Auer et al., 2000
; Hasler et al., 2007
; Michael et al., 2003 a
; Mirza et al., 2004
; Pfleiderer et al., 2003
) or elevated (Hashimoto et al., 2007
; Sanacora et al., 2004
) glutamate or glutamate/glutamine (Glx) levels in various brain regions in depression. Interestingly, reduced Glx levels were recently reported in the anterior PFC in unmedicated depressed patients (Hasler et al., 2007
). Previous postmortem studies have revealed a reduction in cellular size and density in the PFC in depression (Cotter et al., 2002
; Rajkowska et al., 1999
, Rajkowska et al., 2005
). Rajkowska and co-workers (1999)
reported reduction in cortical thickness, and reductions in neuronal size and density in the upper and lower cortical layers of the rostral orbitofrontal cortex (Brodmann’s area 10–47) in depressed subjects. Neuronal pathology detected in cortical layers III, V and VI of the dorsolateral PFC and anterior cingulate cortex in depression (Cotter et al., 2001
; Rajkowska et al., 1999
) is associated with the pathology of glutamatergic pyramidal neurons that express NMDA receptors (Akbarian et al., 1996
). It has been established that activation of synaptic NMDA receptors promotes neuronal survival, and enhanced expression of brain derived neurotrophic factor (Hardingham et al., 2002
). Thus, it is likely that disturbances in the NMDA receptor system in depression may underlie the impairment in cellular plasticity and resilience, and may contribute to cellular pathology consistently detected in the PFC in depression (for review see Rajkowska, 2000
; Rajkowska, 2003
). However, further studies are required to elucidate whether aberrations in the NMDA receptor complex are the reason for or consequence of the cellular changes detected in depression.
Studies have demonstrated that a single dose of the NMDA antagonist ketamine induces a rapid (within hours) antidepressant effect (Berman et al., 2000
; Zarate et al., 2006
). Recent biochemical and behavioral studies support the contention that the fast antidepressant response to ketamine is mediated by increased AMPA to NMDA throughput (Maeng et al., 2008
). Notably, ketamine rapidly increases the release of glutamate (Moghaddam et al., 1997
), a process most likely mediated via NMDA receptors expressed on GABAergic interneurons. Our current and previous postmortem studies demonstrate region specific abnormalities in NMDA receptor expression in MDD. Based on these observations it is tempting to hypothesize that ketamine produces its rapid antidepressant responses by correcting these abnormalities in “here and now”
manner in critical neuronal circuits. Moreover, based on the current observation, it is plausible to hypothesize that optimal levels of NMDA receptor activation are essential for proper PFC function and could be required for antidepressant activity. However, further studies are needed to elucidate whether aberrations in the NMDA receptor complex coexist with the pathology of AMPA receptors in depressed subjects.
A possible limitation of this study is the confounding effect of medication. In fact, previous animal studies have provided evidence that antidepressant drugs produce region-specific reductions of transcripts for NMDA receptor subunits (Boyer et al., 1998
) and produce changes in the radioligand binding properties of the NMDA receptor (for review see Skolnick et al., 1996
). The present study includes four depressed subjects with a reported history of antidepressant medication within four weeks before time of death (). However, antidepressants were not detected in the postmortem toxicology screening. Two of these four subjects had NR2A, NR2B, and PSD-95 levels lower than that of matched controls and comparable to the average level of immunoreactivity of depressive subjects lacking the drug exposure. The other two depressed subjects had NR2A, NR2B and PSD-95 levels nearly unchanged compared to their matched controls. Based on this observation it can be concluded that past antidepressant treatment is unlikely to contribute to the reduction of NR2A, NR2B or PSD-95 levels observed in depression. However, further research is needed to determine the possible influence of antidepressant medication on the expression of NMDA receptor proteins.