Aberrant extracellular signal-regulated kinase (ERK)1/2 signalling in suicide brain: role of ERK kinase 1 (MEK1)

Abstract

Extracellular signal-regulated kinase (ERK)1/2 signalling plays a critical role in synaptic and structural plasticity. Recent preclinical and human brain studies suggest that depression and suicidal behaviour are associated with aberrant ERK1/2 signalling. MEK, is a dual-specificity kinase, which is the immediate upstream regulator of ERK1/2. Two isoforms of MEK (MEK1 and MEK2) exist. By phosphorylating at Ser and Thr residues, MEK activates ERK1/2, which then phosphorylates cytoplasmic and nuclear substrates. On the other hand, MEK itself is regulated through phosphorylation by upstream Raf kinases. Recently, we demonstrated that activation of ERK1/2 and B-Raf was attenuated in the brains of suicide subjects. To further investigate the regulation of ERK1/2 signalling, we examined the expression and activation of MEKs, the interaction of MEK with ERKs, MEK-mediated activation of ERK1/2, and ERK1/2-mediated activation of nuclear substrate Elk-1 in the prefrontal cortex and hippocampus of suicide subjects. In addition, in order to investigate whether MEK is regulated by B-Raf, we examined the B-Raf and MEK interaction. No significant changes were observed in expression levels of MEK1 or MEK2; however, the catalytic activity of only MEK1 (not MEK2) was decreased in both the prefrontal cortex and hippocampus of suicide subjects. The interaction of MEK1 with ERK1 and ERK2 was increased along with decreased phosphorylation and catalytic activity of ERK1/2. In addition, we found decreased phosphorylation of MEK1 and less interaction of B-Raf with MEK1. Our results demonstrate abnormalities in MEK1 at multiple levels and suggest that these abnormalities in MEK1 are crucial for aberrant ERK1/2 signalling in suicide brain.

Depression, ERK kinase, ERK1/2 signalling, post-mortem brain, suicide

Introduction

Mitogen-activated protein kinase (MAPK) signalling plays an important role in many brain functions, including synaptic and structural plasticity. Of various MAPKs, signal transduction through extracellular signal-regulated kinase (ERK)1/2 is crucial for neurotrophin/growth factor-mediated neuronal responses. In this signalling pathway, dimerization and activation of tropomycin receptor kinases cause the recruitment of coupling proteins, such as ShC, growth factor receptor-bound protein, and the coupling protein son of sevenless. This, in turn, causes the activation of a small G protein, Ras (Nimnual et al.1998). Ras then recruits another kinase, Raf, to the membrane, where it is phosphorylated and activated by Ser/Thr/Tyr kinases. This leads to the activation of ERK kinase (known as MEK). MEK then phosphorylates and activates ERK1/2 on Ser/Thr residues (English et al.1999). After activation, both ERK1 and ERK2 can phosphorylate regulatory targets in the cytosol (Lin et al.1993) or translocate to the nucleus and phosphorylate common substrates, including extracellular proteins, effectors, and transcription factors, which results in a variety of responses, such as cell proliferation, differentiation, gene expression, and cell cycle response of neurons to neural activity (Grewal et al.1999; Kolkova et al.2000; Pulverer et al.1991).

Recently, ERK1/2 signalling has been the focus of intense investigation for its possible role in mood disorders and suicide. For example, it has been reported that brain-derived neurotrophic factor ameliorates depression by activating the ERK1/2-MAPK pathway (Shirayama et al.2002). Duman et al. (2007) demonstrated that short-term blockade of ERK1/2 signalling produces a depressive-like phenotype and blocks the behavioural actions of antidepressants (ADs). Recently, Qi et al. (2006) found a correlation between decreased phosphorylation of ERK1/2 and depressive-like behaviours in rats. Even more recently, these investigators showed that fluoxetine increases the activity of ERK1/2 signalling and alleviates depression in rats exposed to swim stress (Qi et al.2008). Moreover, long-term corticosterone treatment to rats decreased phosphorylation of ERK1/2 in the hippocampus (Gourley et al.2008), suggesting a possible role of ERK1/2 in stress and depression. We were the first to demonstrate that expression and activation of ERK1/2 were lower in the post-mortem brain of depressed suicide subjects (Dwivedi et al.2001). Recently, we demonstrated a significant down-regulation of upstream ERK1/2 regulator, B-Raf, in the brains of suicide subjects (Dwivedi et al.2006), suggesting abnormalities at multiple levels in ERK1/2 signalling in suicide brain.

One of the most appealing properties of ERK1/2 activation is that it must be phosphorylated on both Thr and Tyr residues to exhibit full enzymatic activity. This requirement is fulfilled only by MEK, which is a dual-specificity kinase sequentially phosphorylating ERK1 and ERK2 at two sites (initially Tyr185, followed by Thr183) (Haystead et al.1992), thus showing unique selectivity towards ERK1/2. Because of this selective property, MEK inhibitors have been extensively used as pharmacological tools to examine the roles of ERK1/2 in neuronal functions and behaviour (Atkins et al.1998; Duman et al.2007; Einat et al.2003; Huang & Lin, 2006; Kelly et al.2003; Ribeiro et al.2005).

MEK constitutes an evolutionary conserved group of three highly homologous mammalian isoforms: MEK1, its alternatively spliced inactive form MEK1b, and MEK2 (Seger et al.1992; Zheng & Guan, 1993). The phosphorylation of MEK1 and 2 is essential for their activation and subsequent dual phosphorylation of ERK1/2. At the upstream level, MEKs are directly phosphorylated by Raf kinases in their activation loop at Ser218 and Ser222 residues located in the Ser-Xaa-Ala-Xaa-Ser/Thr motif (Alessi et al.1994).

The existence of sequential activation of various kinases in ERK1/2 signalling is essential for the amplification and tight regulation of the transmitted signals. Given the role of MEKs as direct and specific regulators of ERK1/2, the present study was undertaken to examine whether MEK plays any role in the observed abnormalities in ERK1/2 activation in the brain of suicide subjects. We investigated the expression, catalytic activity, phosphorylation, and, therefore, activation of MEK. We also examined whether there is any abnormality in the interaction of MEK with ERK1 and ERK2. In addition, we investigated the regulation of MEKs upstream by examining the interaction of B-Raf with MEK and B-Raf-mediated activation of MEK. Furthermore, to examine the role of MEK at the functional level, we determined MEK-mediated phosphorylation of ERK1 and ERK2 and catalytic activity of ERK1/2 in nuclear fractions. These studies were performed in prefrontal cortex (PFC) and hippocampus of suicide subjects. PFC plays a major role in mood regulation and has been implicated in the pathophysiology of affective disorders and suicide (George et al.1994). On the other hand, the hippocampus is involved in cognition (Sweatt, 2004) and is the primary brain area affected by stress (Sala et al.2004), one of the major factors in suicidal behaviour (Clayton, 1985; Monk, 1987).

Methods

Subjects and diagnostic methods

Brain tissues were obtained from the Maryland Psychiatric Research Center, Baltimore, MD. The study was performed in the PFC (Brodmann's area 9) and hippocampus (containing CA1-4 and dentate gyrus) obtained from suicide subjects (n=28) and non-psychiatric normal controls (n=21). The brains were examined for evidence of gross neuropathology at the time of autopsy. Toxicology and presence of ADs were examined by analysis of urine and/or blood samples from these subjects. The pH of the brain was measured in the cerebellum (Harrison et al.1995). Case reports for each normal control and suicide subject were prepared by means of psychological autopsy using Diagnostic Evaluation after Death (Salzman, 1983) and the Structured Clinical Interview for DSM-IV (SCID) instruments (Spitzer et al.1995) as discussed earlier (Dwivedi et al.2005, 2009). This study was approved by the institutional review board of the University of Illinois at Chicago.

mRNA quantitation

RNA concentration and purity were determined by OD_260/280 ratio. All the samples showed an absorbance ratio >1.8 and exhibited strong 28S and 18S ribosomal RNA bands on 1% (w/v) agarose gels. In addition, all the samples showed an RNA integrity number >7 (Agilent 2100; Agilent, USA), an excellent value for mRNA studies.

The mRNA levels of MEK1 and MEK2 and of housekeeping genes [neuron-specific enolase (NSE) and cyclophilin] were determined using internal standards. Detailed cloning and synthesis of internal standards have been described in a previous publication (Dwivedi et al.2002). The sequences of external and internal primers are given in Table 1. To ensure that amplified sequences match with the corresponding sequences reported in GenBank, the internal standards were sequenced using M13 primer. Quantitative analysis of mRNA was performed by competitive reverse transcription–polymerase chain reaction (RT–PCR), as described earlier (Dwivedi et al.2002). Decreasing concentrations of MEK1 or MEK2 (25-0.75 pg) internal standard complementary RNAs (cRNAs) and 1.5 µCi of phosphorus 32-labelled deoxycytidine triphosphate were added to 1 µg total RNA. The PCR mixture was amplified for 28 cycles. After amplification, aliquots were digested with XhoI in triplicate and run by 1.5% agarose gel electrophoresis. The results were calculated as the counts incorporated into the amplified cRNA standard divided by the counts incorporated into the corresponding mRNA amplification product vs. the known amount of MEK1, MEK2, NSE, or cyclophilin internal cRNA standard added to the test sample. The results are expressed as attomol/µg total RNA.

Table 1

External and internal primer sequences of MEK1, MEK2, NSE, and cyclophilin for amplification

F, Forward; R, reverse.

The underlined bases indicate the restriction sites (XhoI), whereas boldfaced and italicized bases indicate the mutation sites. For cyclophilin, the internal standard was prepared by deleting 65 bp.

MEK activity assays and levels of total and phosphorylated (p)-MEK1 and MEK2 and p-ERK1/2

Tissue preparation

Tissues were homogenized on ice with a glass Teflon homogenizer in lysis buffer (50 mm Tris–HCl (pH 7.5), 100 mm NaCl, 1% Triton X-100, 50 mm sodium fluoride (NaF), 10 mm sodium phosphate, 5 mm EDTA, 1 µg/ml aprotinin, 1 µg/ml leupeptin, 1 mm sodium orthovanadate, and 1 mm 4-[2-aminoethyl] benzenesulfonyl fluoride hydrochloride). The homogenates were centrifuged at 14 000 g for 10 min. The protein content in the supernatant was determined by the Bradford method (Bio-Rad, USA).

MEK1 and MEK2 activity assays

Supernatant containing 100 µg protein was incubated with antibodies against p-MEK1 or p-MEK2 (Santa Cruz Biotechnology, USA) for 2 h at 4°C. The samples were added to a suspension of protein A sepharose beads (Amersham, UK) in Tris-buffered saline and incubated at 4°C for 2 h. The pellet was collected by centrifugation at 500 g for 30 s at 4°C and washed four times with Tris-buffered saline containing 0.5 mm sodium orthovanadate (Na₃VO₄) and 0.01 mm phenylmethane sulfonyl fluoride (PMSF). After the final wash, the supernatant was aspirated and discarded. MEK activity was assayed using an assay kit (Upstate Biotechnology Inc., USA). The MEK assay used recombinant mouse inactive MAPK–glutathione-S-transferase (GST), which was phosphorylated and activated by human MEK1 or MEK2. The activated MAPK-GST then phosphorylated a specific substrate, myelin basic protein (MBP). For MAPK activation, 10 µl magnesium adenosine triphosphate (ATP) cocktail (75 mm magnesium chloride and 500 µm ATP), 5 µl of supernatant containing 50 µg protein (from immunoprecipitated MEK1 or MEK2), and 4 µl of inactive ERK2 were added to assay dilution buffer [20 mm 3-(N-morpholino)propanesulfonic acid (pH 7.2), 25 mm β-glycerol phosphate, 5 mm ethylene glycol tetraacetic acid, 1 mm sodium orthovanadate, and 1 mm dithiothreitol]. The mixture was incubated at 30°C for 30 min. For the activity assay, 4 µl reaction mixture was added to 10 µl MBP stock solution (2 mg/ml in assay dilution buffer) and 10 µl [γ-³²P]ATP (3000 cpm/pmol), and incubated for 10 min at 30°C. The reaction was terminated by adding 10 µl of 5×SDS sample buffer. Reaction products were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and visualized by autoradiography.

Western blot

Samples were resolved onto 10% (w/v) SDS–polyacrylamide gel and blotted on an enhanced chemiluminescence (ECL) membrane (Amersham) (Dwivedi et al.2002, 2005, 2009). Membranes were incubated with human-specific antibodies (MEK1, MEK2, p-MEK1, p-ERK1, p-ERK2, or B-Raf) overnight at 4°C. The dilution for each antibody was as follows: B-Raf, 1:200; MEK1 and 2, 1:1500; p-MEK1, 1:2000; and p-ERK1/2, 1:2000. The ECL membranes were then incubated with horseradish peroxidase-linked secondary antibody (anti-mouse IgG or anti-rabbit IgG, 1:1000) for 5 h at room temperature and exposed to ECL autoradiography film. Where required, membranes were stripped and reprobed with β-actin antibody (Sigma Chemical Co., USA) and ODs of the bands were quantified.

Immunoprecipitation and protein interaction assays

Tissue preparation and immunoprecipitation were performed exactly as previously described (Dwivedi et al.2009). Immunoprecipitation was performed by incubating tissue lysates (0.5 mg protein) for 1 h with 50 µl protein A sepharose. The beads were removed by centrifugation (for 10 min at 3000 g), and the precleared supernatants were added to 50 µl protein A beads, preincubated (overnight at 4°C) with either the antibody (1–4 µg) for proteins to be tested for interaction (ERK1, MEK1, p-MEK1, and B-Raf) or rabbit IgG (2 µg). After 16 h at 4°C, the beads were collected by centrifugation (5 min at 3000 g), washed with immunoprecipitation buffer [125 mm potassium acetate, 0.1% (w/v) Triton X-100, 20 mm Tris–HCl (pH 7.2), 2 µg/ml pepstatin, and 2 µg/ml aprotinin], and resuspended in sample buffer. Immunoprecipitates were then analysed by Western blot using antibodies for ERK1 or MEK1 (for MEK1/ERK1 interaction) or p-MEK1 or B-Raf (for p-MEK1/B-Raf interaction). No specific bands were detected in the presence of rabbit IgG.

ERK1/2 activity assay

ERK1/2 activity was determined in nuclear fraction. We followed exactly the same procedure for preparation and characterization of nuclear fractions, as well as ERK1/2 activity assay as previously described (Dwivedi et al.2001). Transcription factor Elk-1 protein was used as a substrate and ERK1/2-mediated phosphorylation of Elk-1 was determined by Western blot using p-Elk-1 antibody (1:1000 dilution).

Statistical analysis

Data analyses were performed using SPSS (SPSS Inc., USA). All values are the mean±standard deviation (s.d.). The differences in various measures between suicide subjects and normal controls were analysed using the independent-sample t test; p values were two-tailed. Since we did not predict the changes in MEK measures, the t test was exploratory in nature. The overall group differences in measures of MEK between normal controls, depressed and suicide subjects with other psychiatric disorders were evaluated by one-way ANOVA followed by post-hoc (Tukey's) comparisons between groups. Statistical significance was assumed at p<0.05. The relationships between levels of p-MEK1 and catalytic activity of MEK1; catalytic activity of MEK1 and levels of p-ERK1 and p-ERK2; ERK1 and ERK2 catalytic activities and ERK1 and ERK2 interactions with MEK1; and B-Raf-MEK1 interaction and MEK1 were analysed by Pearson product-moment analysis. Similar analyses were performed for correlations between measures of MEK, ERK, and B-Raf and post-mortem interval (PMI), age, and brain pH. The effects of gender and race were determined by an independent-sample t test. Similarly, an independent-sample t test was used to compare depressed subjects who showed AD toxicity with depressed subjects who did not.

Results

The detailed demographic characteristics of suicide subjects and normal control subjects are provided in Table 2. There were no significant differences in age (t=0.63, d.f.=47, p=0.53) or PMI (t=0.11, d.f.=47, p=0.91) or brain pH (t=1.00, d.f.=47, p=0.32) between suicide subjects and normal controls.

Table 2

Characteristics of subjects in the suicide and control groups

ASCVD, Atherosclerotic cardiovascular disease; CO, carbon monoxide; GSW, gunshot wound; n.a., not available; PMI, post-mortem interval.

mRNA levels of MEK1 and MEK2

Competitive RT–PCR for MEK1 and MEK2 mRNA levels revealed amplification product arising from the MEK1 and MEK2 mRNA template at 357 and 319 bp, respectively, and the corresponding digestion products from the cRNA at 184+172 and 164+155 bp, respectively (Fig. 1a, b). The points of equivalence represent the absolute amounts of MEK1 and MEK2 mRNA present (Fig. 1c, d). The amounts of MEK mRNAs (attomol/µg total RNA) in the PFC and hippocampus of normal controls were as follows. PFC (MEK1: 93.4±9.0; MEK2: 25.8±4.8); hippocampus (MEK1: 96.2±13.6; MEK2: 37.3±9.8). As can be seen, the expression of MEK1 was much greater than the expression of MEK2 in both the PFC and hippocampus. On the other hand, their levels were almost the same within PFC and hippocampus.

Fig. 1

Representative gel electrophoreses showing competitive polymerase chain reaction (PCR) analysis for (a) MEK1 or (b) MEK2 messenger RNA (mRNA) contents in prefrontal cortex obtained from one normal control subject. Decreasing concentrations of internal standard complementary RNA (cRNA) (25-0.75 pg) were added to a constant amount (1 µg) of total RNA. The mixtures were reverse-transcribed and PCR amplified in the presence of trace amounts of [³²P]dCTP; aliquots were electrophoresed on 1.5% agarose gel. The higher molecular size band corresponds to the amplification product arising from the mRNA, whereas the lower bands arise from cRNA generated from the internal standard. Data derived from the agarose gel are plotted as the counts incorporated into the amplified (c) MEK1 or (d) MEK2 cRNA standard divided by the counts incorporated into the corresponding mRNA amplification product vs. the known amount of internal standard cRNA added to the test sample. The point of equivalence represents the amount of the respective mRNA.

When compared, mRNA expression levels of both MEK1 and MEK2 were not different between normal controls and suicide subjects either in the PFC (MEK1: t=0.12, d.f.=47, p=0.90; MEK2: t=0.36, d.f.=47, p=0.72) or the hippocampus (MEK1: t=0.85, d.f.=45, p=0.40; MEK2: t=0.76, d.f.=45, p=0.45) (Fig. 2).

Fig. 2

Messenger RNA levels of MEK1 and MEK2 in the prefrontal cortex (PFC) and hippocampus (Hipp) of suicide subjects (S) and normal controls (C). PFC samples were from 21 normal controls and 28 suicide subjects, and hippocampus samples were from 21 normal controls and 21 suicide subjects. Values are given as mean±s.d. The suicide group was compared with the control group.

We used NSE and cyclophilin as housekeeping genes. As previously reported (Dwivedi et al.2005, 2009), mRNA levels of cyclophilin (attomol/µg total RNA) did not show any significant change between normal controls and suicide subjects, either in the PFC (control vs. suicide group: 776.6±112.5 vs. 801.5±117.3; t=0.75, d.f.=47, p=0.46) or the hippocampus (control vs. suicide group: 783.5±110.1 vs. 768.3±102.8; t=0.29, d.f.=45, p=0.77). Similarly, no significant differences were observed in mRNA levels of NSE in the PFC (control vs. suicide group: 360.2±47.7 vs. 344.0±43.9; t=1.2, d.f.=47, p=0.22) or the hippocampus (control vs. suicide group: 349.8±38.2 vs. 345.9±81.4; t=0.46, d.f.=45, p=0.65) between the two groups. When mRNA levels of MEK1 and MEK2 were calculated as ratios to cyclophilin or NSE, similar results were noted, such that both MEK1 and MEK2 mRNA levels were not different between normal controls and suicide subjects in the PFC (ratio to cyclophilin: MEK1, t=0.74, d.f.=47, p=0.46; MEK2, t=0.71, d.f.=47, p=0.48; ratio to NSE: MEK1, t=0.77, d.f.=47, p=0.45; MEK2, t=0.19, d.f.=47, p=0.85) or the hippocampus (ratio to cyclophilin: MEK1, t=0.52, d.f.=45, p=0.61; MEK2, t=0.47, d.f.=45, p=0.64; ratio to NSE: MEK1, t=0.33, d.f.=45, p=0.74; MEK2, t=0.06, d.f.=45, p=0.95).

Protein levels of total MEK1/2 and p-MEK1 and p-MEK2

The protein levels of total MEK1/2 and p-MEK1 and p-MEK2 were determined in total lysates using human-specific antibodies. The specificity of MEK antiserum was checked by using a 100-fold excess of blocking peptide (relative to the molarity of the antiserum) corresponding to the epitope used to generate MEK1 and MEK2. We also examined the antibodies by including positive cells (HeLa, CTLL-2, Jurkat, or NIH3T3 whole-cell lysates) along with human frontal cortex and hippocampus for Western blot and observed that the bands in the PFC and hippocampus were of the same size as observed in cell lines (data not shown). In addition, to validate our data, we initially determined the immunolabelling of MEK1 and MEK2 in the PFC and hippocampus of suicide and control subjects using five different concentrations of protein (5–50 µg). It was observed that the OD of the band increased linearly with increased concentration of protein and that the curve shifted towards the right when a decrease in immunolabelling was observed (data not shown). Western blots of MEK1 and MEK2 in the PFC of three normal controls and three suicide subjects are given in Fig. 3a. β-Actin was used as a housekeeping protein, and ratios of MEK1 or MEK2 to β-actin were calculated. Immunolabelling of β-actin was not different between normal controls (1.1±0.3 AU) and suicide subjects (1.2±0.2 AU). Comparison analysis revealed that there were no differences in the immunolabelling of MEK1 and MEK2 between normal controls and suicide subjects in both the PFC (MEK1: t=1.2, d.f.=47, p=0.22; MEK2: t=1.0, d.f.=47, p=0.32) and the hippocampus (MEK1: t=0.12, d.f.=45, p=0.40; MEK2: t=0.83, d.f.=45, p=0.41) (Fig. 3b).

Fig. 3

Western blots showing the immunolabelling of total (a) MEK1 and MEK2 and (c) phosphorylated MEK (p-MEK)1 and p-MEK2 in the prefrontal cortex (PFC) of three normal controls and three suicide subjects. Protein samples from tissue lysates were subjected to 10% polyacrylamide gel electrophoresis and transferred to enhanced chemiluminescence-nitrocellulose membranes, which were then incubated with primary antibody specific for MEK1, MEK2, p-MEK1, p-MEK2, or β-actin and corresponding secondary antibody. The bands were quantified as described in the Materials and methods section. Ratios of the optical densities of MEKs to that of β-actin were calculated. The mean±s.d. of (b) MEK1 and MEK2 and (d) p-MEK1 and p-MEK2 immunolabelling from normal controls (C) and suicide subjects (S) is given. The PFC samples were from 21 normal controls and 28 suicide subjects, and the hippocampus (Hipp) samples were from 21 normal controls and 21 suicide subjects. The suicide group was compared with the control group (* p<0.001).

When we re-probed the membranes with p-MEK1 or p-MEK2 antibody, in which MEK1 and MEK2 were immunolabelled, we found that the level of p-MEK1 was significantly lower in both the PFC (t=6.9, d.f.=47, p<0.001) and hippocampus (t=8.5, d.f.=45, p<0.001) of suicide subjects (Fig. 3d). However, the level of p-MEK2 was not different between these two groups either in the PFC (t=0.74, d.f.=47, p=0.46) or hippocampus (t=0.59, d.f.=45, p=0.56) (Fig. 3c)

Catalytic activities of MEK1 and MEK2

Figure 4(a, b) shows p-MEK1- and p-MEK2-mediated phosphorylation of MBP in the PFC and hippocampus, respectively, of normal controls and suicide subjects. Similar to our findings of the levels of p-MEK1 and p-MEK2, we observed that the catalytic activity of MEK1 was significantly decreased in both the PFC (t=8.6, d.f.=47, p<0.001) and the hippocampus (t=8.0, d.f.=45, p<0.001) of suicide subjects, whereas no significant differences were observed in the catalytic activity of MEK2 in either the PFC (t=0.84, d.f.=47, p=0.40) or hippocampus (t=0.05, d.f.=45, p=0.96) between normal controls and suicide subjects (Fig. 4c).

Fig. 4

The catalytic activities of MEK1 and MEK2 in post-mortem brains of suicide subjects and normal controls. MEK1 and MEK2 catalytic activities were determined as described in the Materials and methods section using recombinant mouse inactive mitogen-activated protein kinase (MAPK)–glutathione-S-transferase (GST), which was phosphorylated and activated by human MEK1 or MEK2. The activated MAPK–GST then phosphorylated a specific substrate, myelin basic protein (MBP). The reaction was performed in the presence of MBP and [γ-³²P]ATP (3000 cpm/pmol) for 10 min at 30°C. The reaction was terminated by adding 5× sodium dodecyl sulfate (SDS) sample buffer. Reaction products were separated by SDS–polyacrylamide gel electrophoresis and visualized by autoradiography. Data on MEK1- and MEK2-mediated phosphorylation of MBP in (a) the prefrontal cortex (PFC) and (b) hippocampus (Hipp) of the suicide and normal control subjects are given. (c) The mean±s.d. of MEK1 and MEK2 activity in the PFC and hippocampus from normal controls (C) and suicide subjects (S) is also given. PFC samples were from 21 normal controls and 28 suicide subjects, and hippocampus samples were from 21 normal controls and 21 suicide subjects. The suicide group was compared with the control group (* p<0.001).

Correlation between p-MEK1 level and MEK1 catalytic activity

To examine whether the decreased catalytic activity of MEK1 is associated with p-MEK level, we determined the interrelationships between the decreases in protein level and in catalytic activity of MEK1. We observed significant correlations between level of p-MEK1 and catalytic activity of MEK1 in both the PFC (r=0.56, p<0.001) and hippocampus (r=0.59, p<0.001).

Levels of p-ERK1 and p-ERK2

Because ERKs are specific substrates of MEKs and MEKs are capable of phosphorylating both ERK1 and ERK2 with equal efficiency, we determined protein levels of p-ERK1/2. Western blots of p-ERK1/2 in the PFC and hippocampus of suicide subjects and normal controls are given in Fig. 5(a, b), respectively. We observed that levels of p-ERK1 and p-ERK2 were significantly decreased in the PFC (p-ERK1: t=8.7, d.f.=47, p<0.001; p-ERK2: t=5.9, d.f.=47, p<0.001) and hippocampus (p-ERK1: t=8.0, d.f.=45, p<0.001; p-ERK2: t=4.2, d.f.=45, p<0.001) of suicide subjects compared to normal controls (Fig. 5c).

Fig. 5

Western blots showing the immunolabelling of phosphorylated extracellular signal regulated (p-ERK) 1/2 in (a) the prefrontal cortex (PFC) and (b) hippocampus (Hipp) of three normal controls and three suicide subjects. Protein samples from tissue lysates were subjected to 10% polyacrylamide gel electrophoresis and transferred to enhanced chemiluminescence-nitrocellulose membranes, which were then incubated with primary antibody specific for p-ERK1/2 or β-actin and corresponding secondary antibody. (c) The mean±s.d. of p-ERK1/2 immunolabelling in the PFC and hippocampus from normal controls (C) and suicide subjects (S) is given. PFC samples were from 21 normal controls and 28 suicide subjects, and hippocampus samples were from 21 normal controls and 21 suicide subjects. The suicide group was compared with the control group (* p<0.001).

Catalytic activity of ERK1/2 in nuclear fractions

Because phosphorylation and activation of ERK1/2 lead to their translocation to the nucleus, we immunoprecipitated p-ERK1/2 and subsequently determined the catalytic activity of ERK1/2 in nuclear fractions. ERK1-mediated phosphorylation of Elk-1 in the PFC and hippocampus of normal controls and suicide subjects is given in Fig. 6a. We observed that p-ERK1/2–mediated phosphorylation of Elk-1 was significantly decreased in the PFC (t=9.4, d.f.=47, p<0.001) and hippocampus (t=8.9, d.f.=45, p<0.001) of suicide subjects compared to normal controls (Fig. 6b).

Fig. 6

(a) Western blots showing extracellular signal regulated (ERK)1/2 catalytic activity in the prefrontal cortex (PFC) and hippocampus (Hipp) of control and suicide subjects. Protein samples (15 µg) were used to immunoprecipitate active ERK1/2 by incubating with immobilized phosphorylated ERK1/2 monoclonal antibody overnight at 4°C. The samples were centrifuged and washed, first with lysis buffer and then with kinase buffer, followed by suspension in kinase buffer supplemented with 20 µm adenosine triphosphate and Elk-1 fusion protein, as detailed in the Materials and methods section. They were then incubated for 30 min at 30°C. The reaction was terminated by adding 3×sodium dodecyl sulfate (SDS) sample buffer. The samples were boiled and subjected to 10% SDS–polyacrylamide gel electrophoresis and electrotransferred to enhanced chemiluminescence-nitrocellulose membranes. The membranes were then incubated with phosphorylated Elk-1 monoclonal antibody (1:1000) overnight at 4°C. (b) The mean±s.d. of ERK1/2 catalytic activity in the PFC and hippocampus from normal controls (C) and suicide (S) subjects is given. PFC samples were from 21 normal controls and 28 suicide subjects, and hippocampus samples were from 21 normal controls and 21 suicide subjects. The suicide group was compared with the control group (* p<0.001).

Correlations between MEK1 catalytic activity and p-ERK1 and p-ERK2 levels and between ERK1/2 catalytic activity and p-ERK1 and p-ERK2 levels

We observed significant correlations between MEK1 catalytic activity and levels of p-ERK1 and p-ERK2 (PFC: pERK1, r=0.64, p<0.001; pERK2, r=0.52, p<0.001; hippocampus: pERK1, r=0.49, p<0.001; pERK2, r=0.47, p<0.001). Similarly, a significant correlation between p-ERK1/2 level and ERK1/2 catalytic activity was observed (PFC: pERK1, r=0.67, p<0.001; pERK2, r=0.59, p<0.001; hippocampus: pERK1, r=0.69, p<0.001; pERK2, r=0.49, p<0.001).

ERK1 and MEK1 interaction

To further examine whether decreased ERK1/2 catalytic activity in suicide subjects was associated with alteration in interaction between ERK and MEK1, we co-immunoprecipitated ERK1 and MEK1, followed by Western blots using MEK1 or ERK1 antibodies. Western blots for these interactions in PFC and hippocampus are given in Fig. 7(a, b), respectively. We observed a significant increase in interaction between MEK1 and ERK1 whether the interaction was examined using MEK1 or ERK1 antibodies in both the PFC (MEK1: t=7.6, d.f.=47, p<0.001; ERK1: t=9.6, d.f.=47, p<0.001) and the hippocampus (MEK1: t=8.0, d.f.=45, p<0.001; ERK1: t=10.7, d.f.=45, p<0.001) of suicide subjects (Fig. 7c).

Fig. 7

MEK1 and ERK1 interaction in post-mortem brains of normal controls and suicide subjects. MEK1 and ERK1 were co-immunoprecipitated using specific antibodies followed by Western blot using MEK1 or ERK1 antibody. The optical densities of the bands were calculated. The interaction of MEK1 and ERK1 in (a) the prefrontal cortex (PFC) and (b) the hippocampus (Hipp) of three normal controls and three suicide subjects is given. (c) The mean±s.d. of the MEK1/ERK1 interaction in the PFC and hippocampus from normal controls (C) and suicide (S) subjects is given. PFC samples were from 21 normal controls and 28 suicide subjects, and hippocampus samples were from 21 normal controls and 21 suicide subjects. The suicide group was compared with the control group (* p<0.001).

Correlation between MEK1–ERK1 interaction and ERK1/2 catalytic activity

Significant correlations between decreased catalytic activity of ERK1/2 in nuclear fraction and increased ERK1 and MEK interaction were observed when immunoprecipitates were immunolabelled with ERK1 or MEK1 antibodies (MEK1: PFC, r=0.67, p<0.001; hippocampus, r=0.69, p<0.001; ERK1: PFC, r=0.66, p<0.001; hippocampus, r=0.59, p<0.001).

Interaction between B-Raf and p-MEK1

To examine whether MEK is regulated by upstream kinases, we examined the interaction of B-Raf with p-MEK1 and B-Raf-mediated catalytic activity of MEK1. For this, B-Raf and p-MEK1 were co-immunoprecipitated and analysed by Western blot using B-Raf or p-MEK antibodies. Western blots showing p-MEK1 and B-Raf interactions in the PFC and hippocampus of normal controls and suicide subjects are given in Fig. 8(a, b), respectively. We observed a significant reduction in this interaction in both the PFC and hippocampus of suicide subjects compared to normal controls whether co-immunoprecipitates were subjected to B-Raf or p-MEK antibody (PFC: B-Raf: t=12.4, d.f.=47, p<0.001; p-MEK1: t=10.6, d.f.=47, p<0.001; hippocampus: B-Raf: t=7.4, d.f.=45, p<0.001; p-MEK1: t=8.6, d.f.=45, p<0.001) (Fig. 8c).

Fig. 8

p-MEK1 and B-Raf interaction in post-mortem brains of normal controls and suicide subjects. p-MEK1 and B-Raf were co-immunoprecipitated using specific antibodies followed by Western blot using p-MEK1 or B-Raf antibody. The optical densities of the bands were calculated. The interaction of p-MEK1 and B-Raf in (a) the prefrontal cortex (PFC) and (b) the hippocampus (Hipp) of three normal controls and three suicide subjects is given. (c) The mean±s.d. of the p-MEK1 and B-Raf interaction in the PFC and hippocampus from normal controls (C) and suicide (S) subjects is given. PFC samples were from 21 normal controls and 28 suicide subjects, and hippocampus samples were from 21 normal controls and 21 suicide subjects. The suicide group was compared with the control group (* p<0.001).

Correlation between MEK1 catalytic activity and B-Raf/p-MEK1 interaction

We found significant correlations between MEK1 catalytic activity and B-Raf/p-MEK1 interaction in both the PFC and hippocampus whether the interaction assay was determined using B-Raf or p-MEK1 antibody (PFC: B-Raf, r=0.58, p<0.001; p-MEK1, r=0.59, p<0.001; hippocampus: B-Raf, r=0.49, p<0.001; p-MEK1, r=0.47, p<0.001).

Correlation between level of p-MEK1 and B-Raf/p-MEK1 interaction

Protein level of p-MEK1 and B-Raf/p-MEK interaction were significantly correlated in both the PFC and hippocampus when B-Raf (PFC: r=0.64, p<0.001; hippocampus: r=0.50, p<0.001) or p-MEK1 (PFC: r=0.59, p<0.001; hippocampus: r=0.71, p<0.001) antibody was used for interaction assays.

Effects of confounding variables

The effects of potential confounding variables were evaluated for various measures in which we had found differences between normal controls and suicide subjects. No significant effect of age, PMI, or brain pH was observed on any of the measures between normal controls and suicide subjects (see Supplementary Table 1, available online). Similarly, no significant effect of gender or race was noted in various measures of MEK in which we found changes in suicide subjects, except in the hippocampus, where we observed a slight but significant difference between males and females in interaction between ERK1 and MEK1, when ERK1 antibody was used (p=0.007) (see Supplementary Table 2, available online).

To examine whether the method of suicide had any effect on the measures in which we found changes, we compared suicide subjects who died by violent or other means (n=20) with those who died by drug overdose (n=8). No significant differences were observed between these two groups (data not shown).

Effects of major depression

To examine whether the differences in various MEK measures between normal control subjects and suicide subjects were related to depression or were present in all suicide subjects, we examined the effect of the diagnosis of major depression on these measures. For this purpose, we stratified suicide subjects into those who were diagnosed as having major depression and those who were diagnosed as having other psychiatric disorders or those who had no mental illness. Of the 28 suicide subjects, 12 had major depression. In the suicide group with other psychiatric disorders (n=16), there were five with adjustment disorder, two with schizoaffective disorder, two with bipolar disorder, and three with drug/alcohol abuse; there was no diagnosed psychiatric illness in two of these subjects, and the diagnosis was not available in two of these subjects. Hippocampus specimens were available for 26 of the 28 suicide subjects.

A one-way ANOVA followed by pairwise between-group comparisons revealed that none of the variables were different between suicide subjects with major depression and suicide subjects with other psychiatric disorders in both PFC (Table 3) and hippocampus (Table 4). However, the groups of suicide subjects with major depression and suicide subjects with other psychiatric disorders both showed significant differences in these measures in both PFC (Table 3) and hippocampus (Table 4) when compared separately with normal control subjects.

Table 3

Effect of major depression (MDD) on measures of MEK in the prefrontal cortex of subjects in the suicide group^a

a

Data were analysed using univariate analysis of variance (ANOVA). Bonferroni-adjusted p values were considered significant (p=0.05/6=0.008).

b

Group 1 indicates normal controls; group 2, subjects in the suicide group with a history of MDD; group 3, subjects in the suicide group with a history of other psychiatric disorders.

c

Percentage of control.

Table 4

Effect of major depression (MDD) on measures of MEK in the hippocampus of subjects in the suicide group^a

a

Data were analysed using univariate analysis of variance (ANOVA). Bonferroni-adjusted p values were considered significant (p=0.05/5=0.008).

b

Group 1 indicates normal controls; group 2, subjects in the suicide group with a history of MDD; and group 3, subjects in the suicide group with a history of other psychiatric disorders.

c

Percentage of control.

Effects of AD toxicology

To examine whether the observed changes in measures of MEK in the suicide group were related to the presence of AD, we compared the suicide subjects who tested positive for ADs during the screen at the time of death (n=5) with those who did not (n=23). We did not find significant difference in any of the measures between these two groups (see Supplementary Table 3, available online).

Discussion

The physiological functions of ERK1/2 are mediated only when these ERKs are phosphorylated on their regulatory Thr183 and Tyr185 residues in the activation loop. At an unphosphorylated state, ERK1/2 are inactive and the binding of the substrates to the catalytic domain is inhibited by the sequestering activation loop. Monophosphorylation of ERK1/2 is not sufficient to exert their downstream effects. Thus, phosphorylation of ERK1/2 at both Thr and Tyr residues is critical for conformational change and their subsequent binding to various substrates. MEK1 and MEK2 belong to the rare breed of dual-specificity kinases, which phosphorylate both Thr and Tyr residues (Dhanasekaran & Reddy, 1998) and, thus, are responsible for full activation of ERK1 and ERK2. In a previous study, we demonstrated that activation of ERK1 and ERK2 is reduced in the PFC and hippocampus of suicide subjects (Dwivedi et al.2001). To fully investigate this hypoactivation in the present study, we thoroughly examined the status of their direct regulators, MEK1 and MEK2, in the brains of suicide subjects.

In PFC and hippocampus of suicide subjects, we found that neither mRNA nor protein levels of MEK1 or MEK2 were altered; however, the catalytic activity of MEK1, but not MEK2, was significantly lower in both these brain areas of suicide subjects. Because activation of MEKs itself requires phosphorylation at Ser218 and Ser222 residues (Rubinfeld & Seger, 2005; Zheng & Guan, 1994), and elimination of either residue completely inactivates MEK, we examined phosphorylation states of MEK1 and MEK2 using antibodies that recognize phosphorylation of MEKs at Ser218 and Ser222 residues. As with catalytic activity, phosphorylation of only MEK1, but not MEK2, was significantly decreased in the PFC and hippocampus of suicide subjects. Furthermore, there was a significant correlation between decreased catalytic activity of MEK1 and the level of p-MEK1. These results indicate that MEK1 activity is reduced in suicide brain because of less phosphorylation at Ser218 and Ser222 residues. In light of the reports that both MEK1 and MEK2 phosphorylate ERK1 and ERK2 with equal efficiency, our finding of less activation of only MEK1, but not MEK2, is intriguing and renders the possibility that, under certain conditions, the two isoforms of MEK may be differently regulated and thus may have distinct functions. It is interesting to note that although both MEKs are almost identical in their overall structure, there are subtle structural differences between MEK1 and MEK2. For example, unlike the kinase domains, the N-termini of MEK1 and MEK2 show only 40% similarities with each other. On the other hand, genetic studies in Caenorhabditis elegans and Drosophila demonstrate that a single MEK gene is sufficient to induce Trk-mediated signal transduction (Kornfeld et al.1995; Tsuda et al.1993). These studies thus suggest that mammalian MEK1 and MEK2 may have both overlapping as well as distinct functions. In addition, in-vitro studies suggest that upstream regulators, Ras and Raf, form a complex with MEK1 but not with MEK2 for cell cycle progression (Jelinek et al.1994). On the other hand, in several cell lines of non-mammalian origin, the two MEK isoforms have distinct ways to contribute to a regulated ERK activity and cell cycle progression (Ussar & Voss, 2004). For example, the MEK1/ERK module preferentially provides proliferative signals, whereas the MEK2/ERK module induces growth arrest at the G1/S boundary. Whether such functional differences occur under in-vivo conditions are not known and further studies are required to delineate these functional differences.

To further investigate whether reduced MEK1 catalytic activity was associated with decreased activation of its substrates ERK1/2, we examined phosphorylation states of both ERK1 and ERK2. We found that phosphorylation of both of these ERKs was decreased in the PFC and hippocampus of suicide subjects and that there was a significant correlation between decreased catalytic activity of MEK1 and decreased phosphorylation of ERK1/2. This suggests that a defect in MEK1 activation is solely responsible for less phosphorylation and, therefore, less activation of ERK1 and ERK2 in suicide brain. We also observed that ERK1/2-mediated phosphorylation of transcription factor Elk-1, a major substrate of ERK1/2, was reduced in nuclear fractions of the PFC and hippocampus of suicide subjects and that there was a significant correlation between decreased phosphorylation of ERK1/2 and decreased phosphorylation of Elk-1. These results further emphasize that activation and functioning of ERK1/2 is reduced in suicide brain, which is associated with less activation of the selective MEK1 isoform.

Because ERK1/2 and MEK1/2 co-express and interact with each other directly (Fukuda et al.1997), we also examined whether decreased activation of ERK1/2 is associated with alteration in such an interaction. Our co-immunoprecipitation studies demonstrate that ERK1 interacts with MEK1 and that this interaction is significantly greater in the PFC and hippocampus of suicide subjects. This finding is quite relevant to the observed reduction in MEK1-mediated reduced activation of ERK1/2 in suicide brain. It has been shown that ERK1/2 and MEK interact with each other by protein interaction motifs located in each of the proteins. Phosphorylation of the regulatory residues in the activation loop of ERKs by MEKs induces a major conformational change (Canagarajah et al.1997), dramatically increasing the catalytic activity of ERKs and also forcing the detachment of the two proteins (Wolf et al.2001). This dissociation allows the translocation of the two proteins separately to the nucleus; ERK1/2 remain in the nucleus (Pouysségur et al.2002), whereas MEKs are rapidly exported to the cytosol because of the nuclear export signal in their N-terminal region (Fukuda et al.1997; Jaaro et al.1997). Our findings of reduced nuclear catalytic activity of ERK1/2 in suicide brain could be due to less nuclear translocation of these ERKs, which, in turn, could be associated with the observed increased interactions of MEK1 with ERKs.

It has been demonstrated that growth factor stimulation increases B-Raf kinase activity to phosphorylate and functionally activate MEK1 (Jaiswal et al.1994; Reuter et al.1995). Recently, we demonstrated that B-Raf activation is attenuated in the PFC and hippocampus of suicide subjects (Dwivedi et al.2006). Because B-Raf directly interacts with p-MEK1 (Papin et al.1995), to further characterize the regulation of MEK1, we determined the interaction of B-Raf with p-MEK1. We found a significant decrease in B-Raf/p-MEK1 interaction in both the PFC and hippocampus of suicide subjects. We further found that decreased catalytic activity of MEK1 was significantly correlated with a decreased B-Raf/MEK1 interaction. Thus, our results demonstrate that decreased phosphorylation and catalytic activity of MEK1 may be due to less interaction of upstream B-Raf kinase with MEK1 and that B-Raf may be one of the major regulators of MEK1.

As mentioned in the Introduction, several studies have directly and indirectly implicated a role of ERK1/2 in depressive behaviour. In addition, a recent study demonstrated that treatment of rats with an inhibitor of MEK blocks activation of ERK1/2 and induces mood disorder-related behavioural deficits (Einat et al.2003). Peripheral injection of MEK inhibitors also produces helpless-like behaviours and eliminates response to ADs in behavioural despair (Duman et al.2007). On the other hand, depressive-like behaviour in rats is correlated with less activation of ERK1/2 (Qi et al.2006). Because depression is a major factor in suicidal behaviour, we determined whether the observed changes in various measures were related specifically to depression. We found that the measures of B-Raf/MEK/ERK were altered in all suicide subjects irrespective of psychiatric diagnosis, suggesting that these abnormalities may be relevant to suicide rather than any specific psychiatric diagnosis. Further studies in a large population of suicide subjects are required to confirm specific diagnosis-related changes in Raf/MEK1/ERK1/2 signalling.

In conclusion, to our knowledge, this is the first study examining the regulation of the ERK1/2 pathway in depression/suicide. We demonstrate abnormalities in the selective isoform of MEK (i.e. MEK1), a direct upstream regulator of ERK1/2, in the post-mortem brain of suicide subjects. We found that MEK1 is less active, shows increased interaction with ERK1/2, causes decreased activation and functions of ERK1/2, and is a substrate of B-Raf. These results clearly demonstrate that MEK1 is central to the abnormalities in ERK1/2 signalling in the brain of suicide subjects. Our results provide a novel mechanism by which the ERK1/2 signalling cascade is regulated in the brain of suicide subjects. Overall, our present study and the previous findings of less activation of ERK1/2 (Dwivedi et al.2001) and B-Raf (Dwivedi et al.2006) suggest the possibility that less activation of B-Raf/MEK1/ERK1/2 pathway could lead to functional abnormalities in brain circuitry of suicide subjects, which in turn, may be crucial in the pathogenic mechanisms of depressive/suicidal behaviour.

Acknowledgements

This research was supported by grants from NIMH (R01MH68777, RO1MH082802, R21MH081099), NARSAD, and the American Foundation for Suicide Prevention to Dr Y. Dwivedi; NIMH RO1MH48153 to Dr G. N. Pandey; and MH60744 and MH66123 to Dr R. Roberts. We thank Miljana Petkovic, Barbara Brown, and Joy K. Roche for their help in organizing the brain tissues. We also thank the members of the Maryland Brain Collection for their efforts, particularly in family interviews and dissection. We are grateful for the cooperation of the Office of the Chief Medical Examiner in Baltimore, Maryland.

Note

Supplementary material accompanies this paper on the Journal's website.

Statement of Interest

None.