|Home | About | Journals | Submit | Contact Us | Français|
Reports of cognitive decline, symptom worsening and brain atrophy in bipolar disorder (BD) suggests that the disease progresses over time. The worsening neuropathology may involve excitotoxicity and neuroinflammation. We determined protein and mRNA levels of excitotoxicity and neuroinflammatory markers in postmortem frontal cortex from 10 BD patients and 10 age-matched controls. The brain tissue was matched for age, postmortem interval and pH. The results indicated statistically significant lower protein and mRNA levels of the NMDA receptors, NR-1 and NR-3A, but significantly higher protein and mRNA levels of IL-1β, the IL-1 receptor, myeloid differentiation factor 88 (MyD88), nuclear factor-kappa B subunits and astroglial and microglial markers (glial fibrillary acidic protein (GFAP), inducible nitric oxide synthase (iNOS)), c-fos and CD11b) in postmortem frontal cortex from BD compared with control subjects. There was no significant difference in mRNA levels of tumor necrosis factor alpha (TNFα) or neuronal nNOS in the same region. These data demonstrate the presence of excitotoxicity and neuroinflammation in BD frontal cortex, with particular activation of the IL-R cascade. The changes may account for reported evidence of disease progression in BD, and be a target for future therapy.
Bipolar disorder (BD) is a severe psychiatric disease characterized by repeated manic and depressive episodes. Reports of symptom worsening, cognitive decline and progressive brain atrophy suggest that the disease is progressive and have a neurodegenerative component. Progression is consistent with reports of structural, metabolic, and signaling abnormalities in postmortem brain from BD patients (1–6), and with evidence of excitotoxicity and neuroinflammation in BD (7–10). Animal studies have shown that both processes are associated with increased levels of pro-inflammatory cytokines, reactive oxygen radicals, and nitric oxide in brain, and may be accompanied by upregulation of brain arachidonic acid signaling markers (11–13). Expression levels of enzymes involved in arachidonic acid metabolism, including an arachidonate-selective cytoplasmic phospholipase A2 (cPLA2), secretory (sPLA2), and cyclooxygenase (COX)-2, were found elevated in postmortem frontal cortex from BD patients compared with controls (14).
A number of markers have been shown to be altered by neuroinflammation and excitoxicity. For example, chronic administration of subconvulsive doses of N-methyl-D-aspartate (NMDA) decreased rat brain expression of the NMDA receptor (NR) subunits, NR-1 and NR-3A (11), and upregulated mRNA and activity of cytosolic phospholipase A2 (cPLA2) and of one of its transcription factors, activator protein-2 (AP-2) (12). Chronic NMDA also upregulated rat brain protein and mRNA levels of neuroinflammatory markers interleukin-1 beta (IL-1β), tumor necrosis factor-alpha (TNFα), glial fibrillary acidic protein (GFAP) and inducible nitric oxide synthase (iNOS) (12).
Binding of IL-1 to the IL-1 receptor (IL-1R) recruits interleukin receptor associated kinase (IRAK) to the receptor complex via its association with the IL-1R accessory protein (13) and with adaptor protein myeloid differentiation factor (MyD88) (15, 16). Upon recruitment, IRAK is highly phosphorylated and subsequently dissociates from the receptor complex to interact with tumor necrosis factor receptor-associated factor 6 (17), which in turn is involved in activation of nuclear I kappa kinase and nuclear factor-kappa B (NF-κB) activation (18) (Fig. 1). TNFα can regulate cPLA2 and cyclooxygenase-2 (COX-2) expression in a mitogen activated protein kinase-dependent manner in various cell types (19–21). Activation of IL-1R by IL-1 triggers activation of its cascade, increasing expression of cPLA2 (13, 15, 16), secretory sPLA2 (16). COX-2 (22), GFAP (19–21) and iNOS via activation of transcription factors AP-2 or NF-κB transcription factors in mouse astrocytes and other cell types. In response to an excitotoxic insult, iNOS (23) and the immediate early gene, c-fos, are expressed and considered as specific biomarkers of excitotoxicity (24).
To further clarify the possible contributions of neuroinflammation and excitotoxicity in BD, in the present study we measured protein and mRNA levels of a markers of these processes in the postmortem frontal cortex from BD patients and control subjects, matched for age, postmortem interval (PMI) and pH. Among others, we measured expression of NMDA receptor subunits, of IL-1β, IL-1R, and MyD88, and markers of activated glia and astroglia. We chose these markers focus on possible changes in the IL-1 R cascade (Fig. 1). We studied the frontal cortex because previous studies have indicated structural, metabolic, and signaling abnormalities in this area in BD patients (1–6, 14).
The study was approved by the Institutional Review Board of McLean Hospital, and by the Office of Human Subjects Research (OHSR) of NIH (#4380). Frozen postmortem human frontal cortex from 10 BD patients and 10 age-matched controls were provided by the Harvard Brain Tissue Resource Center (McLean Hospital, Belmont, MA) under PHS grant number R24MH068855. Mean age, postmortem interval and pH of the frozen brain samples (measured by the method of Harrison et al. (25), did not differ significantly between the BD and control groups. The age (years, control: 43 ± 3.5 vs BD: 49 ± 7.2) postmortem interval (hours, control: 27 ± 1.5 vs BD: 21 ± 3.0) and brain pH (control: 6.6 ± 0.16 vs BD: 6.7 ± 0.09) did not differ significantly between the two groups, whereas the BD patients were exposed to various psychotropic medications as shown previously (26).
Membrane, cytoplasmic, and nuclear extracts were prepared from frozen tissue samples as described (11). Briefly, the tissue was homogenized in 20 mM Tris HCl (pH 7.5), and 0.2 mM EDTA buffer containing a cocktail of protease inhibitors (Roche, Indianapolis, IN). The suspension was centrifuged at 100,000 g for 1 hr at 4°C and the supernatant (S1), containing mostly cytosolic constituents, was removed. The pellet was re-suspended in the previously mentioned buffer containing 0.1% Triton-X 100, and incubated for 1 hour at 4°C. This mixture was centrifuged at 100,000 g for 1 hour at 4 °C. The resulting supernatant (S2) containing membrane constituents was removed. Protein concentrations of cytosolic and membrane extracts were determined using a Bio-Rad protein reagent (Bio-Rad, Hercules, CA). Each membrane and cytosolic fraction was characterized using cadherin and tubulin antibodies.
Nuclear extracts were prepared from frozen tissue as previously described (27). Nuclear fraction was characterized by using lamin B antibody. Protein concentrations of cytoplasmic and nuclear extracts were determined using Bio-Rad Protein Reagent (Bio-Rad, Hercules, CA).
Protein (50 μg) from the membrane, cytoplasmic and nuclear extract was separated on 4–20% SDS-polyacrylamide gels (PAGE) (Bio-Rad). Following electrophoresis, the proteins were transferred to a nitrocellulose membrane. Membrane protein blots were incubated overnight in Tris-buffered-saline solution, containing 5% nonfat dried milk and 0.1% Tween-20, with specific primary antibodies (1:200 dilution) for the NR-1, NR-2A, NR-2B, NR-3A, IL-1R and cadherin (Cell Signaling, Beverly, MA). Cytosolic protein blots were incubated similarly but with primary antibodies (1:200 dilution) for IL-1β, MyD88, GFAP, iNOS, nNOS and tubulin (Cell Signaling, Beverly, MA). Nuclear NF-κB p50 and NF-κB p65 protein levels were determined using specific (1:200) primary antibodies (Cell Signaling). Membrane, cytoplasmic and nuclear protein blots were incubated with appropriate HRP-conjugated secondary antibodies (Bio-Rad) and visualized using a chemiluminescence reaction (Amersham, Piscataway, NJ) on X-Ray film (Kodak, Rochester, NY). Optical densities of immunoblot bands were measured using Alpha Innotech Software (Alpha Innotech, San Leandro, CA) and were normalized to β–actin (Sigma, St. Louis, MO). Experiments were carried out in duplicate for the 10 controls and 10 BD brain samples. Mean density values were expressed as percent of control.
Total RNA from brain and lipid-rich tissue was isolated using an RNeasy mini kit (Qiagen, Valencia, CA). The RNA integrity number (RIN) was measured using a Bioanalyzer (Agilent 2100 bioanalyzer, Santa Clara, CA). RIN values were control 6.9 ± 0.4 for the control samples and BD 7.15 ± 0.5 (Mean ± SEM) for the BD samples. Complementary DNA (cDNA) was prepared from total RNA using a high capacity cDNA Archive kit (Applied Biosystems, Foster City, CA). Messenger RNA levels of NR-1, NR-2A, NR-2B, NR-3A, IL-1β, IL-1R, MyD88, TNFα, GFAP, iNOS, nNOS, CD11b, p50, p65, and c-fos were measured by quantitative RT-PCR, using an ABI PRISM 7000 sequence detection system (Applied Biosystems). Specific primers and probes for NR-1, NR-2A, NR-2B, NR-3A, IL-1β, IL-1R, MyD88, GFAP, iNOS, nNOS CD11b, p50, p65, and c-fos were purchased from TaqManR gene expression assays (Applied Biosystems) and consisted of a 20X mix of unlabeled PCR primers and Taqman minor groove binder (MGB) probe (FAM dye-labeled). The fold-change in gene expression was determined by the ΔΔCT method (27). Data were expressed as the relative level of the target gene in the BD tissue normalized to the endogenous control (β-globulin) and relative to the control (calibrator), as previously described (28). All experiments were carried out twice in triplicate with 10–10 independent samples, and data were reported as relative expression.
Frozen tissue, held on a dry ice bed, was sectioned in a cryostat. The cut surface then was immediately refrozen. One section was used for biochemical or molecular analysis, while a corresponding section was used for immunohistochemistry.
For immunohistochemistry, frozen sections were warmed to −20°C from −80°C storage conditions. Each section was gently raised to room temperature and immersion-fixed with 4% paraformaldehyde for 18 hours and cyroprotected in 30% sucrose. The cryostat sections were air dried for 60 min, rinsed in 1x automation buffer (Biomedia Corp, Foster City, CA). Sections were treated with 0.3% hydrogen peroxide to quench endogenous peroxidase activity, and then incubated for 1 hr at room temperature (RT) with a monoclonal human leukocyte antigen–D related (HLA-DR) antibody (1:750; MBL, Woburn, MA). Sections were rinsed, incubated for 30 min at RT with secondary antibody (Vector Elite Kit, xxx) and detected with DAB chromagen (Dako, Carpinteria, CA). An antibody to GFAP was used to detect astrocytes. Staining was conducted on a IHC Omni-UltraMap HRP (Ventana Medical Systems, Tucson, AZ). Sections were incubated with anti-GFAP (1:3500; Dako) for 32 min at RT, rinsed, incubated with OmniMap anti-Rb HRP for 16 min, and counter-stained with modified Harris hematoxylin. Images were collected using an Aperio Scanscope T2 Scanner (Aperio Technologies, Vista, CA) and viewed using an Aperio Imagescope v. 126.96.36.1997. Images were rated by two independent scorers blind to subject classification. Samples were rank-ordered based upon the level of the glial responses then clustered based upon subject classification.
Data are expressed as mean ± SEM. Statistical significance of means was calculated using a two-tailed unpaired t-test at p < 0.05. Pearson correlations were made between age, post-mortem interval, pH of the frontal cortex, and mRNA levels of NR-1, NR-2A, NR-2B, NR-3A, IL-1, IL-1R, Myd88, GFAP, iNOS, NF-κB65, NF-κB 50 and CD11B in post-mortem brains from controls and BD patients, separately. Statistical significance was set at p < 0.05.
The mean protein levels of NR-1, NR-2A, NR-2B and NR-3A in postmortem BD brain were compared with the matched controls. NR-1 and NR-3A protein levels were decreased significantly by 42% (p < 0.05) and 41% (p < 0.01) respectively, in BD compared to controls (Figs. 2A and 2D), whereas the mean NR-2A and NR-2B protein levels did not differ significantly between the groups (Figs. 2B and 2C). mRNA levels of NR-1 and NR-3A were significantly decreased in BD compared with control brain by 0.42-fold (p < 0.01) and 0.46-fold (p < 0.01), respectively (Figs. 2E and 2H), while NR-2A and NR-2B mRNA levels were not significantly different (Figs. 2F and 2G)
As illustrated in Figure 3, compared with controls, there were significantly elevated protein levels of the inflammatory cytokines, IL-1β (41%; p < 0.05), IL-R (41%; p < 0.05) and MyD88 (38%; p < 0.05) in frontal cortex from BD patients (Figs. 3A to to3C).3C). Significant increases also were seen in mRNA levels for IL-1β (2.2 fold; p < 0.01), IL-R (3.6 fold; p < 0.001) and MyD88 (3.6 fold; p < 0.01) (Figs. 3D to to3F3F).
There were significant increases in nuclear protein and mRNA levels of p50 (45% and 0.8 fold, respectively) (Figs. 4A and C; p < 0.05) and of p65 (62% and 1.4 fold, respectively) (p < 0.01, p < 0.001, Figs. 4B and D) in the BD brain.
A significant increase was observed in the mean protein levels of GFAP (46%; p < 0.05) and of iNOS (37%; p < 0.05), in BD compared with control frontal cortex (Figs. 5A and 5B). The increased protein levels were accompanied by significant increases in mRNA levels of GFAP (0.7 fold; p < 0.01) and iNOS (1.7 fold; p < 0.01) (Figs. 5D and 5E). However, there was no significant difference in the protein or mRNA level of nNOS (Figs. 5C and 5F). c-fos and CD11b protein and mRNA levels were significantly higher in BD compared to control brain (Figs. 6A and 6B; 6C and 6D).
There was no significant difference in TNFα mRNA level in frontal cortex from BD patients (Fig. 6E). The elevation in astrocyte and microglia markers was supported by immunohistochemical staining for both GFAP and HLD-A (Fig. 6F). In control tissue, GFAP+ astrocytes displayed fine fibrous processes, while in the BD tissue, hypertrophic GFAP+ astrocytes were detected. In BD tissue, furthermore, HLA-DR staining detected increased staining for process bearing microglia displaying a thickening of processes.
Pearson correlations between mRNA expression levels in control and BD brains with respect to postmortem interval, age and pH, were all statistically insignificant (p > 0.05) (Table 1). Mean values of the three parameters did not differ significantly between the patient and control groups.
The present study demonstrated a significant increase in excitotoxicity and neuroinflammatory markers in BD compared with control frontal cortex. Decreased mean protein and mRNA levels of the NR1 and NR-3A subunits were accompanied by increased levels of markers of excitotoxicity, c-Fos and iNOS mRNA. In addition, protein and mRNA levels of IL-1β, IL-1 receptor and MyD88 and NF-κB subunits (p50 and p65) in the same region were increased significantly. There was a significant increase in GFAP expression, and also in the level of CD11b mRNA (a marker of astrocyte and microglial activation) in postmortem frontal cortex from BD patients. Increases in protein and mRNA levels were associated with increased staining for GFAP and HLA-DR in the same brain region from BD patients. However, there was no significant difference in nNOS or TNFα expression in postmortem frontal cortex from BD compared with control. In sum, these results are consistent with a marked activation of the IL-1R receptor cascade, which is characteristic of both systemic and local insults (29, 30).
Studies have shown that an excitotoxic insult caused by chronic NMDA administration decreased expression of NMDA receptor subunits and increased levels of neuroinflammatory markers in rat brain (11, 12). Similarly, we observed decreased expression of NR-1 and NR-3 subunits in the postmortem frontal cortex from BD patients. Excitotoxicity (23) in the current study was indicated by the increased protein and mRNA levels of iNOS without a significant change in nNOS expression in the BD brain. Although this observation is not consistent with a previous study (31) that reported an increase in nNOS in postmortem hippocampal region from BD patients, the difference may represent differences in the response between the different brain regions. Further, characterization of the BD frontal cortex revealed a significant increase in c-fos expression, a marker of excitotoxicity (24, 32, 33), suggesting the presence of excitotoxicity in BD.
Consistent with our findings regard the presence of excitotoxicity markers, earlier studies indicated an elevated brain glutamate/glutamine ratio in children and adults with BD (8). Postmortem brains of BD patients displayed decreased levels of the NMDA receptor subunits, NR-1, NR-2A and NR-3A (9, 10). mRNA levels of NR-1, NR-2A, and NR-3A also were found to be decreased in the postmortem BD brain, as were concentrations of the NMDA receptor-associated postsynaptic proteins, PSD-95 and SAP102 (34). Finally, gene variants of NR-1 and NR-2 have been linked to the disease (10, 35, 36). These studies suggest that an increase in glutamate transmission is associated with decreased expression of NMDA receptor subunits.
The observed differences in expression of the NMDA subunits in the BD brain indicate increased changes NMDA receptor activity. NMDA receptor stimulation by glutamate or by NMDA decreases NR1 expression, which we observed previously (37, 38). Furthermore, in vitro studies indicate that the NR3A subunit co-assembles with other subunits (NR-1, NR-2A or NR-2B) to form NMDA receptors with reduced activity and Ca2+ influx (39, 40), whereas mice lacking the NR-3A subunit have increased NMDA receptor activity (41). Increased NMDA function could increase arachidonic acid signaling (38). Rats given a daily subconvulsive dose of NMDA for 3 weeks demonstrated reduced brain levels of NR-1 and NR-3A, increased arachidonic acid turnover in brain phospholipids, increased protein and mRNA levels of cPLA2 and sPLA2, increased AP-2 DNA binding activity, and increased AP-2α and AP-2β protein in the frontal cortex (38).
IL-1 is a major cytokine responsible for inducing a number of proteins associated with inflammation(42). Many of these responses are induced by the rapid activation of the transcription factor NF-κB following signal transduction caused by IL-1β binding to the type I IL-1 receptor (43). Similarly, the current study demonstrates increased protein and mRNA levels of IL-1β, IL-1R, and MyD88 in BD brain, which may be responsible for the observed upregulated NF-κB transcription factor. The increase in NF-κB may also be responsible for increased expression of iNOS and GFAP, markers for astrocytes. The observed increase in GFAP expression in the current study is not in line with other observations on BD brain (44, 45), but this discrepancy may be due to regional differences or drug exposure. A previous clinical study did indicated an increased in serum TNFα level in BD patients (46). However, in this study we did not see similar change in postmortem frontal cortex region from BD patients.
The observed increases in neuroinflammatory markers in the BD brain (14) may have been induced in part by underlying process of excitotoxicity. Thus, chronic NMDA administration to rats upregulated neuroinflammatory markers and cerebroventricular infusion of high lipopolysaccharide concentrations for weeks to months induced IL-1β, TNFα, and amyloid precursor protein, leading to degeneration of hippocampal CA3 pyramidal neurons (47). Cytokines associated with neuroinflammation can activate both cPLA2 and sPLA2 at astrocytic cytokine receptors (48–51).
None of the mRNA levels in either the BD or control brains was correlated significantly with postmortem interval, brain pH, or subject age, and mean the values of these parameters did not differ significantly between the two groups. Nevertheless, the BD patients had been exposed to a variety of drugs not experienced by the control subjects, which may have confounded the results. Because of their selective exposure, our findings may be related to differences in drug exposure, rather than being specific to the BD trait. Thus, future studies should examine arachidonic cascade markers in brains from patients with schizophrenia (to use a roughly comparable drug exposure as a control), or with unipolar (primary major) depression, or with Alzheimer’s disease (to test for disease specificity) (52).
In conclusion, markers of the excitotoxicity and neuroinflammation are significantly upregulated in postmortem frontal cortex from BD patients. Their upregulation might result in cell death, and account for brain atrophy and cognitive decline that have been reported in BD patients. Our observations suggest that agents that attenuate excitotoxicity or neuroinflammation could prove beneficial for treating BD.
Chronic administration to rats of lithium, valproic acid, carbamazepine or lamotrigine, mood stabilizers that have been approved for treating BD, was reported to decrease the markers of brain arachidonic acid metabolism that had been shown to be upregulated in rat models of excitotoxicity and neuroinflammation, as well as in the BD brain (12, 14). These markers include cPLA2 and COX-2, and some of their transcription factors, AP-2 and NF-κB. Some of the mood stabilizers also prevented NMDA- and lipopolysaccharide-induced increases in arachidonic acid markers in rat brain (12, 53–56). The present study supports presence of excitotoxicity and neuroinflammation may play role BD pathology.
We thank the Harvard Brain Bank, Boston, MA for providing the post-mortem brain samples under PHS grant number R24MH068855. This research was entirely supported by the Intramural Research Programs of the National Institute on Aging and the National Institute of Environmental Health Sciences, National Institutes of Health Bethesda, MD 20892. We thank the National Cancer Institute (NCI), Center for Cancer Research (CCR) Fellows Editorial Board for proofreading the manuscript.