https://orcid.org/0000-0001-7302-460X
Abstract
Bipolar disorder (BD) has been associated with impaired cellular resilience. Recent studies have shown abnormalities in the unfolded protein response (UPR) in BD. The UPR is the cellular response to endoplasmic reticulum (ER) stress. Mesencephalic astrocyte-derived neurotrophic factor (MANF), a trophic factor, decreases ER stress by modulating the UPR. The objective of this study is to investigate the MANF-ER stress pathway in BD and major depressive disorder (MDD) compared to healthy controls (HC).
MANF protein concentration and MANF and GRP78 gene expression were assessed in peripheral blood from individuals with BD, MDD, and HC (protein: 40 BD, 55 MDD, 55 HC; gene expression: 52 BD, 61 MDD, 69 HC). MANF protein and gene expression along with GRP78 gene expression were also analyzed in postmortem brain tissue (20 BD, 20 MDD, 19 HC). MANF protein was quantified using an ELISA assay while quantitative polymerase chain reaction was used for MANF and GRP78 gene expression.
Peripheral MANF protein levels were reduced in individuals with BD in a depressive state compared to controls (P = .031) and euthymic BD participants (P = .013). No significant differences in MANF or GRP78 gene expression were observed in BD irrespective of mood state, or MDD compared to HC (all P > .05). No differences were observed regarding MANF/GRP78 protein or gene expression levels in postmortem tissue (P > .05).
Individuals with BD who were in an acute depressive phase were found to have reduced peripheral MANF levels potentially signifying abnormal UPR and supporting the notion that BD is associated with increased ER stress.
Bipolar disorder (BD) and major depressive disorder (MDD) are complex mood disorders that have been associated with reduced cellular resilience. Recent evidence suggests endoplasmic reticulum (ER) stress can harm cellular resilience and is involved in mood disorder pathophysiology, highlighting the need to further understand the cellular stress responses. Stress within the ER can lead to cellular apoptosis and may contribute to the emergence of symptoms observed in mood disorders. Here, we focused on peripheral and central mesencephalic astrocyte-derived neurotrophic factor (MANF) protein concentration and gene expression in individuals with BD and MDD. We found reduced peripheral MANF protein concentration in individuals with BD who were in a depressive state. Given that BD is often associated with ER stress, MANF was expected to be increased as MANF has been found to alleviate ER stress; however, MANF was found to be reduced, signifying a possible systemic blunted response to ER stress.
INTRODUCTION
Bipolar disorder (BD) is characterized by alternating episodes of mania, depression, and hypomania.1 BD impacts approximately 2%–4% of the world population2 and is often comorbid with other diseases, impacting the quality of life and psychosocial functioning.3 Neuroanatomical studies have consistently demonstrated deficits in brain morphology, such as reduced gray matter volume in brain regions such as the hippocampus4 and white matter deficits, which have been associated with cognitive deficits.5,6 Studies have also found that deficits in brain morphology are correlated with blood markers. For example, catabolites of the kynurenine pathway in the blood of individuals with BD, which is the primary catabolic route for tryptophan, are associated with diffusion tensor imaging markers of white matter damage in several brain regions such as the amygdala and corpus callosum.7,8 In addition, increased proinflammatory cytokines such as tumor necrosis alpha have been reported to be correlated with white matter volume in several brain regions, including the frontal and temporal lobe, in individuals with BD.9,10 Alterations in homeostasis caused by markers such as increased proinflammatory cytokines may significantly impact brain morphology in BD due to impaired cellular resilience. Several studies investigating cellular resilience in peripheral blood mononuclear cells, such as the lymphocytes of individuals with BD, found that in response to stress, more cells were found to be in the early stages of apoptosis and more likely to die compared to cells collected from healthy individuals.11,12 These findings suggest that there may be underlying abnormalities in the cell’s ability to appropriately respond to changes in their environment, leading to increased cellular apoptosis and potentially being associated with the deficits observed in neuroanatomical studies. This stresses the need to understand better the pathways and cellular organelles responsible for maintaining cellular homeostasis in BD to elucidate our knowledge of the pathophysiology.
The endoplasmic reticulum (ER) is a cellular organelle responsible for protein synthesis, folding, and transport, as well as lipid metabolism.13 Recent studies investigating alterations in ER function and ER stress proteins in mood disorders including BD and major depressive disorder (MDD) have reported abnormality in gene expression related to ER function and ER stress response in cells such as lymphoblastoid and leukocyte-derived RNA samples.14–16 The ER is highly sensitive to perturbation in ER homeostasis, which has led to evolutionary adaptations to maintain ER functionality and cellular integrity. One such mechanism is the unfolded protein response (UPR). ER integrity can be compromised by a cascade of factors such as hypoxia, Ca2+ depletion, or altered glycosylation, which leads to an accumulation of misfolded proteins, and activation of the UPR. UPR functions through three principal transmembrane proteins: Inositol-Requiring Enzyme 1 (IRE1α), Protein Kinase RNA-like ER Kinase (PERK), and Activating Transcription Factor (ATF)6; it is further regulated by chaperone proteins such as immunoglobulin-binding protein (BiP/GRP78).17,18 When there is an accumulation of unfolded proteins, GRP78 is released into the ER lumen from IRE1α, PERK, and ATF6 receptors, which activates distinct signaling pathways that begin a cascade of events to attempt to restore homeostasis.17,19 The release of GRP78 from IRE1α causes X-box-binding protein 1 (XPB1) mRNA to splice, resulting in the production of spliced XBP1 (sXBP1), which is an influential transcription factor that upregulates other proteins involved in protein folding. However, the overactivation of IRE1α can also lead to additional interactions that activate the c-Jun N-terminal kinase (JNK) pathway, which is involved in cellular apoptosis. Aside from IRE1α, PERK activation reduces protein translation by phosphorylating eukaryotic translation initiation factor 2α (eIF2α) but also increases the translation of ATF4, which upregulates genes involved in autophagy. Like IRE1α, chronic activation of ATF4 leads to apoptosis by increasing the expression of C/EBP homologous protein (CHOP). CHOP is a transcription factor that functions by downregulating anti-apoptotic proteins within the B-cell lymphoma-2 (Bcl-2) family and upregulating pro-apoptotic proteins such as death receptor 5 and tribbles homolog 3 (TRB3), which results in cell death. Lastly, the transmembrane protein ATF6 moves to the Golgi after GRP78 is released from it, where it cleaves specific proteases that are then released into the cytosol. ATF6 functions as a transcription factor that increases the expression of ER chaperone proteins such as GRP78, further increasing the cell’s ability to manage misfolded proteins.20 Through these proteins, UPR attempts to remedy alterations in function through the reductions of protein translation, protein folding, and the removal of misfolded proteins. However, as discussed above, if UPR is chronically activated, it may lead to cellular apoptosis.
Mesencephalic astrocyte-derived neurotrophic factor (MANF) is a trophic factor belonging to a group of neurotrophic proteins responsible for regulating neuronal survivability.21 It has been observed to protect neurons within the midbrain in animal models of Parkinson’s disease (PD)22,23. The first pivotal study demonstrating MANF’s role in the UPR was in MANF knockout mice displaying severe pancreatic dysfunction. These knockout mice developed severe diabetes due to progressive postnatal reduction of β-cell mass caused by the inhibition of MANF leading to apoptosis. In addition, this study demonstrates that the inhibition of MANF in vivo resulted in increased ER stress and chronic UPR activation, as indicated by increased levels of CHOP.24 MANF has also been found to have positive impacts on other systems, including the immune, nervous, and cardiac systems, where MANF is shown to regulate macrophage polarization and inflammatory signaling, along with improving the survivability of neurons and hepatocytes.20,23 In studies across different diseases, MANF has been shown to be increased compared to controls. For example, individuals with PD have been shown to have significantly higher MANF concentrations compared to controls.25 Similarly, individuals experiencing acute intracerebral hemorrhage displayed significantly elevated serum MANF compared to controls.26 These studies consistently demonstrate a pattern of increased peripheral MANF concentration during active disease, potentially signifying activation of the MANF pathway in these individuals.
The structure of MANF may underlie its observed role in diseases. The N-terminal of MANF contains a saposin-like domain, which is hypothesized to allow for interactions with the cell membrane and lipid binding. The C-terminal domain of MANF has a carboxyl group and RTDL motif, which aids with MANF retention within the ER via KDEL receptor binding. The current literature suggests that MANF plays a vital role in the UPR. Specifically, much of the evidence posits that MANF interacts with GRP78 by binding to its nucleotide-binding domain, which assists its activities during increased ER stress. While the exact mechanism is still debated, it is hypothesized that MANF aids GRP78 in identifying misfolded proteins and provides cysteine bonds to allow GRP78 to correct them. MANF has also been shown to have neuroprotective roles by blocking the nuclear factor kappa B pathway (NF-Kb) and inhibiting pro-apoptotic BAX proteins. Lastly, MANF also regulates intracellular calcium concentration by interacting with neuroplastin.20,27
In a recent study from our group investigating the relationship between MANF and lithium use in rat models, we found that lithium treatment increased mRNA MANF expression in the brain tissue dissected from the prefrontal cortex and striatum,28 revealing another potential pathway that may be involved in the mood stabilizing and neuroprotective effects of lithium. However, no studies have examined the MANF pathway in BD, highlighting the need and the novelty to further investigate MANF in individuals with BD and potentially inform new treatment pathways for drug discovery in BD.
The present study investigated the gene expression of MANF and GRP78 and the concentration of MANF protein levels in human blood samples and in postmortem brain tissue from individuals diagnosed with BD across different mood states (euthymic, depressed, manic/hypomanic), MDD, and healthy individuals. Given that ER stress has been previously reported in BD,29 we hypothesize that there is abnormally reduced MANF and GRP78 gene expression in BD and MDD compared to healthy individuals. Furthermore, we hypothesize significantly lower MANF protein concentration in BD and MDD compared to healthy individuals.
METHODS
Subjects
A group of young adults nested from a large population-based study based in Pelotas, a city in Southern Brazil, was used for this study. For this nested investigation, we selected BD and MDD participants (serum analysis n = 40 BD and 55 MDD, gene expression n = 52 BD and 61 MDD) who were recruited during the second wave of the population-based study conducted between 2012 and 2014.30 Since MDD shares some symptomatic characteristics with BD and has also been shown to be associated with ER stress, MDD was investigated.15,16 The inclusion criteria for BD and MDD groups included individuals that were diagnosed with BD or MDD, according to the DSM-IV.31,32 Lastly, healthy control (HC) individuals were also recruited (serum analysis n = 55, gene expression n = 69). The inclusion criteria used for HC included (1) no mood disorders diagnosis (MDD or BD), (2) no anxiety disorder diagnosis (agoraphobia, social phobia, specific phobia, panic disorder, or Generalized Anxiety Disorder), (3) no diagnosis of Obsessive Compulsive Disorder, (4) no diagnosis of Posttraumatic Stress Disorder, (4) no diagnosis of Attention Deficit Hyperactivity Disorder, (5) no current suicide risk, (6) did not present any lifetime psychotic symptoms, (7) not currently abusing or depending on illicit drugs.
Manic and depressive symptoms were assessed using the Young Mania Rating Scale33 and The Montgomery–Åsberg Depression Rating Scale,34 respectively. Master’s and PhD-level trained psychologists conducted all psychological assessments. All participants provided written consent prior to inclusion and the Research Ethics Committee of the Universidade Católica de Pelotas (UCPel) approved this study under protocol number 2008/118.
Postmortem human brain tissue from the hippocampus and the prefrontal cortex from 20 participants with BD, 20 participants with MDD, and 19 HC were also assessed. The tissue sample was acquired from the Douglas Bell Canadian Brain Bank in Montreal, Canada. BD and MDD diagnosis was determined by psychological autopsies using the Structured Clinical Interview for the DSM-IV axis 1. HC participants included individuals who had died due to natural causes or accidents and had no prior psychiatric conditions. This study was approved by the Hamilton Integrated Research Ethics Board, and written consent was provided by the next of kin and the Douglas Institute Research Ethics Board (#11267).
Tissue Preparation
The tissue samples were dissected and stored at −80 °C until analysis. Before protein analysis, ~100 mg of the prefrontal cortex and the hippocampus were subdivided into sections weighing ~40–60 mg and placed on ice. For protein extraction, 400 μL of cold N-PER lysis buffer was used per the manufacturer’s protocol. Samples used for protein analysis were sonicated and centrifuged at 10 000 g for 10 minutes at 4 °C to extract the supernatant while excluding cellular debris. Protein concentration was determined using the Pierce 660 nm assay with bovine serum albumin standards (ThermoFisher Scientific, Inc.).
Blood Collection
For the peripheral protein analysis, an anticoagulant-free vacuum tube was used to draw 10 mL of blood from each participant between 8 am and 11 am. Serum was extracted within 2 hours by centrifugation at 4000 g for 15 minutes. The serum was stored at −80 °C until analysis.
MANF Protein Assay
A high-sensitivity enzyme-linked immunosorbent assay (ELISA) kit (AB215417; ABCAM) was used to determine MANF concentration in serum and postmortem tissue according to the manufacturer’s protocol. In summary, 100 μL of diluted serum/supernatant or standard diluent was added into a 96-well plate pre-coated with an antibody specific to MANF and incubated for 1 hour on a microwell plate shaker at 31.42 rad/s to allow for binding. After 1 hour, the standards and samples were discarded and washed 4 times using 300 μL of provided buffers. This was followed by the addition of 100 μL of MANF enzyme conjugate working solution in each well, which was then incubated again for 1 hour at 31.42 rad/s on a microwell plate shaker. After 1 hour, the solution was discarded and 4 more cycles of washing using 300 μL of the provided wash buffer were performed. Afterward, 100 μL of substrate solution was added and incubated for 20 minutes on a microwell plate shaker at 31.42 rad/s while protected from light, followed by the addition of the stop solution. MANF levels were determined by absorbance at 450 nm using a microplate reader (EPOCH 2 microplate spectrophotometer, BioTek), converting absorbance to concentration using a standard curve that demonstrated a linear relationship between optical density and MANF concentration (R2 > 0.95). The samples were run in duplicates and the intra-assay coefficient was <10%. The results are expressed in ng/mL.
RNA Extraction and Quantitative Real-Time PCR for Postmortem Tissue
RNA concentration and purity were determined using a NanoDrop ONE spectrophotometer. All samples used for cDNA synthesis had an A260/A280 and A260/230 ratio between 1.8 and 2.2. QScript (95048; Quantabio) was used to synthesize equal amounts of cDNA from DNAse-treated RNA samples as per the manufacturer’s instructions. Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) reactions using SYBR green master mix (BIORAD) were performed to determine the relative mRNA expression of MANF with β-actin (ACTB) and glyceraldehyde-3-phosphate dehydrogenase as housekeeping genes (Table 1). The cDNA was run in duplicates and amplified for 40 cycles using the CFX96 Real-Time System (BioRad). Relative gene expression was then calculated using the 2–∆∆Ct method described by Schmittgen and Livak.35
List of human oligo-dT primer sequences used for RT-qPCR (5’–3’).
| Genes | Forward primer | Reverse primer |
|---|---|---|
| ACTB | ACA GAG CCT CGC CTT TG | CCT TGC ACA TGC CGG AG |
| TBP | CAG CAA CTT CCT CAA TTC CTT G | GCT GTT TAA CTT CGC TTC CG |
| GAPDH | ACA TCG CTC AGA CAC CAT G | TGT AGT TGA GGT CAA TGA AGG G |
| GRP78 | GTG CCT ACC AAG AAG TCT CAG | CCA GTC AGA TCA AAT GTA CCC A |
| MANF | TCA CAT TCT CAC CAG CCA CT | CAG GTC GAT CTG CTT GTC ATA C |
| Genes | Forward primer | Reverse primer |
|---|---|---|
| ACTB | ACA GAG CCT CGC CTT TG | CCT TGC ACA TGC CGG AG |
| TBP | CAG CAA CTT CCT CAA TTC CTT G | GCT GTT TAA CTT CGC TTC CG |
| GAPDH | ACA TCG CTC AGA CAC CAT G | TGT AGT TGA GGT CAA TGA AGG G |
| GRP78 | GTG CCT ACC AAG AAG TCT CAG | CCA GTC AGA TCA AAT GTA CCC A |
| MANF | TCA CAT TCT CAC CAG CCA CT | CAG GTC GAT CTG CTT GTC ATA C |
Abbreviations: ACTB, β-actin; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; MANF, mesencephalic astrocyte-derived neurotrophic factor; TBP, TATA-box binding protein.
List of human oligo-dT primer sequences used for RT-qPCR (5’–3’).
| Genes | Forward primer | Reverse primer |
|---|---|---|
| ACTB | ACA GAG CCT CGC CTT TG | CCT TGC ACA TGC CGG AG |
| TBP | CAG CAA CTT CCT CAA TTC CTT G | GCT GTT TAA CTT CGC TTC CG |
| GAPDH | ACA TCG CTC AGA CAC CAT G | TGT AGT TGA GGT CAA TGA AGG G |
| GRP78 | GTG CCT ACC AAG AAG TCT CAG | CCA GTC AGA TCA AAT GTA CCC A |
| MANF | TCA CAT TCT CAC CAG CCA CT | CAG GTC GAT CTG CTT GTC ATA C |
| Genes | Forward primer | Reverse primer |
|---|---|---|
| ACTB | ACA GAG CCT CGC CTT TG | CCT TGC ACA TGC CGG AG |
| TBP | CAG CAA CTT CCT CAA TTC CTT G | GCT GTT TAA CTT CGC TTC CG |
| GAPDH | ACA TCG CTC AGA CAC CAT G | TGT AGT TGA GGT CAA TGA AGG G |
| GRP78 | GTG CCT ACC AAG AAG TCT CAG | CCA GTC AGA TCA AAT GTA CCC A |
| MANF | TCA CAT TCT CAC CAG CCA CT | CAG GTC GAT CTG CTT GTC ATA C |
Abbreviations: ACTB, β-actin; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; MANF, mesencephalic astrocyte-derived neurotrophic factor; TBP, TATA-box binding protein.
RNA Extraction and Quantitative Real-Time PCR for Human Blood
Total RNA was extracted from 2 mL of blood collected in Paxgene blood RNA tubes (762165; Qiagen) using the Paxgene blood RNA kit (762164; Qiagen). The extracted RNA was treated with DNase, and its purity was tested using A260/A280 method on a microplate reader (EPOCH 2 microplate spectrophotometer, BioTek). All treated RNA used for cDNA synthesis had an A260/A280 ratio between 1.8 and 2.2. The treated RNA was used to produce equal amounts of cDNA using Qscript (95048; Quantabio) as per the manufacturer’s protocol. RT-qPCR reactions were performed using SYBR green master mix (1725271; BIORAD) to analyze the mRNA expression of 2 genes of interest, MANF and GRP78. TATA-box binding protein and ACTB were used as housekeeping genes. All samples were analyzed using triplicate technical replicates and amplified for 40 cycles using the QuantStudio 7 Pro Real-Time PCR System (Applied Biosystems). The ThermoFisher Cloud Relative Quantification analysis module was used to calculate the relative gene expression of each gene using the 2–∆∆Ct method as described by Schmittgen and Livak.35 The human oligo-dT primer sequences used for RT-qPCR (5’–3’) are described in Table 1.
Statistical Analysis
The GraphPad Prism 9 was used for postmortem tissue analysis, the Statistical Package for the Social Sciences (SPSS) version 26.0 was used for peripheral protein analysis, and Rstudio was used to conduct peripheral gene expression analysis. The continuous variables were tested for normality using the Shapiro–Wilk test, and the Bartlett test was used to test for homogeneity of variances. Differences in protein or mRNA expression in the postmortem tissue between BD, MDD, and HC were determined using a 2-way analysis of variance (ANOVA) with Dunnett’s and Tukey’s post-hoc tests. Tukey’s post-hoc test was used to determine the difference between BD and MDD compared to HC, while Dunnett’s post-hoc test was performed to determine differences between BD and MDD. For all peripheral analysis, the BD group was analyzed as 1 group including all mood states; however, the data were also further analyzed by mood states to determine if differences among and between mood states and HC existed. Logarithmic transformations were used to normalize peripheral MANF protein data since the data did not present a normal distribution. An ANOVA with Bonferroni post-hoc tests was conducted to determine if there were any significant differences in protein concentration between BD, BD mood states, MDD, and HC. Furthermore, a Kruskal–Wallis test with sex as a covariate was used to determine differences in mRNA MANF and GRP78 expression in the periphery between BD, BD mood states, MDD, and HC.
RESULTS
Participants within the BD, MDD, and HC groups were matched by sex, age, body mass index (BMI), and years of education for the peripheral protein analysis; participants within the 3 groups were matched by age, BMI, and years of education for the peripheral gene expression analysis. The participants for the peripheral analysis were from the same sample, except a number of individuals did not have serum available to analyze, which resulted in fewer participants for the serum analysis. The average age for the peripheral analysis was approximately 25 years old, and both MDD and BD groups had participants who were treated with psychotropic medication. The average age for the postmortem analysis was approximately 45–49 years old (see Table 2 for a complete breakdown of participant demographics).
Participant clinical and demographic characteristics.
| Control | BD | MDD | P-value | |
|---|---|---|---|---|
| Blood protein analysis | ||||
| Age | 25.4 (2.3) | 25.9 (2.3) | 25.8 (2.2) | .335 |
| Sex (M/F) | 25/30 | 11/29 | 25/30 | .141 |
| Years of Education | 10.67 (2.76) | 9.70 (2.59) | 9.75 (1.98) | .300 |
| BMI | 26.60 (6.14) | 28.30 (5.75) | 25.66 (4.57) | .136 |
| Clinical Assessment | ||||
| MADRS | 0 (0 -2.00) | 7.5 (3.0–12.0) | 2.0 (0.0–4.00) | <.001 |
| FAST | 4.5 (2.0–8.25) | 12.5 (8.0–24.5) | 7.5 (3.0–12.0) | <.001 |
| BRIAN | 24.11 (4.98) | 38.00 (12.17) | 27.20 (6.75) | <.001 |
| MOCA | 23.31 (3.34) | 22.00 (3.82) | 23.39 (3.27) | .116 |
| Medication | ||||
| Mood stabilizer | – | 2 (9.1) | 5 (15.6) | .938 |
| Antidepressants | – | 16 (72.7) | 18 (56.3) | .565 |
| Antipsychotics | – | – | 3 (9.4) | .659 |
| Antianxiolytic | – | 7 (31.8) | 17 (53.1) | .459 |
| 2 + psychiatric intervention | 5 (26.3) | 9 (33.3) | .814 | |
| Blood gene expression analysis | ||||
| Age | 25.99 (2.13) | 25.75 (2.195) | 25.57 (2.24) | .559 |
| Sex (M/F) | 31/38 | 12/40 | 13/48 | .005 |
| Years of Education | 10.91 (3.59) | 9.72(3.16) | 10.50 (3.44) | .161 |
| BMI | 25.61 (5.68) | 28.60 (6.15) | 26.40 (5.74) | .058 |
| Clinical assessment | ||||
| MADRS | 0.0 (0.0-4.0) | 12.0 (2.0 -18.0) | 8.0 (2.0 -16.0) | <.001 |
| FAST | 6.0 (3.0-10.0) | 11.0 (7.0 -23.0) | 12.5 (5.0 -19.75) | <.001 |
| BRIAN | 25.34 (6.67) | 35.43 (11.02) | 33.18 (11.21) | <.001 |
| MOCA | 23.32 (3.68) | 22.12 (3.94) | 22.66 (3.45) | .201 |
| Medication | ||||
| Mood stabilizer | – | 4 (13.3) | 4 (13.3) | .580 |
| Antidepressants | 6 (66.7) | 19 (63.3) | 22 (73.3) | .836 |
| Antipsychotics | – | 5 (16.7) | 3 (10.0) | .675 |
| Antianxiolytic | 3 (33.3) | 16 (53.3) | 15 (50.0) | .778 |
| 2 + psychiatric intervention | 2 (14.3) | 10 (37.0) | 10 (40.0) | |
| Postmortem | ||||
| Age | 49.63 (2.43) | 45.80 (3.36) | 49.85 (1.53) | .23 |
| Sex (M/F) | 16/3 | 9/11 | 17/3 | .006 |
| PMI | 47.83 (4.25) | 46.33 (3.94) | 62.99 (5.99) | .09 |
| Cause of death | ||||
| Suicide | 0 | 15 | 20 | |
| Accident | 5 | 3 | 0 | |
| Natural | 14 | 2 | 0 | |
| Medication (last 3 months) | ||||
| Mood stabilizer | 0 | 2 | 0 | |
| Antidepressants | 0 | 0 | 8 | |
| Antipsychotics | 0 | 3 | 0 | |
| Antianxiolytic | 2 | 0 | 3 | |
| 2 + psychiatric intervention | 0 | 9 | 3 |
| Control | BD | MDD | P-value | |
|---|---|---|---|---|
| Blood protein analysis | ||||
| Age | 25.4 (2.3) | 25.9 (2.3) | 25.8 (2.2) | .335 |
| Sex (M/F) | 25/30 | 11/29 | 25/30 | .141 |
| Years of Education | 10.67 (2.76) | 9.70 (2.59) | 9.75 (1.98) | .300 |
| BMI | 26.60 (6.14) | 28.30 (5.75) | 25.66 (4.57) | .136 |
| Clinical Assessment | ||||
| MADRS | 0 (0 -2.00) | 7.5 (3.0–12.0) | 2.0 (0.0–4.00) | <.001 |
| FAST | 4.5 (2.0–8.25) | 12.5 (8.0–24.5) | 7.5 (3.0–12.0) | <.001 |
| BRIAN | 24.11 (4.98) | 38.00 (12.17) | 27.20 (6.75) | <.001 |
| MOCA | 23.31 (3.34) | 22.00 (3.82) | 23.39 (3.27) | .116 |
| Medication | ||||
| Mood stabilizer | – | 2 (9.1) | 5 (15.6) | .938 |
| Antidepressants | – | 16 (72.7) | 18 (56.3) | .565 |
| Antipsychotics | – | – | 3 (9.4) | .659 |
| Antianxiolytic | – | 7 (31.8) | 17 (53.1) | .459 |
| 2 + psychiatric intervention | 5 (26.3) | 9 (33.3) | .814 | |
| Blood gene expression analysis | ||||
| Age | 25.99 (2.13) | 25.75 (2.195) | 25.57 (2.24) | .559 |
| Sex (M/F) | 31/38 | 12/40 | 13/48 | .005 |
| Years of Education | 10.91 (3.59) | 9.72(3.16) | 10.50 (3.44) | .161 |
| BMI | 25.61 (5.68) | 28.60 (6.15) | 26.40 (5.74) | .058 |
| Clinical assessment | ||||
| MADRS | 0.0 (0.0-4.0) | 12.0 (2.0 -18.0) | 8.0 (2.0 -16.0) | <.001 |
| FAST | 6.0 (3.0-10.0) | 11.0 (7.0 -23.0) | 12.5 (5.0 -19.75) | <.001 |
| BRIAN | 25.34 (6.67) | 35.43 (11.02) | 33.18 (11.21) | <.001 |
| MOCA | 23.32 (3.68) | 22.12 (3.94) | 22.66 (3.45) | .201 |
| Medication | ||||
| Mood stabilizer | – | 4 (13.3) | 4 (13.3) | .580 |
| Antidepressants | 6 (66.7) | 19 (63.3) | 22 (73.3) | .836 |
| Antipsychotics | – | 5 (16.7) | 3 (10.0) | .675 |
| Antianxiolytic | 3 (33.3) | 16 (53.3) | 15 (50.0) | .778 |
| 2 + psychiatric intervention | 2 (14.3) | 10 (37.0) | 10 (40.0) | |
| Postmortem | ||||
| Age | 49.63 (2.43) | 45.80 (3.36) | 49.85 (1.53) | .23 |
| Sex (M/F) | 16/3 | 9/11 | 17/3 | .006 |
| PMI | 47.83 (4.25) | 46.33 (3.94) | 62.99 (5.99) | .09 |
| Cause of death | ||||
| Suicide | 0 | 15 | 20 | |
| Accident | 5 | 3 | 0 | |
| Natural | 14 | 2 | 0 | |
| Medication (last 3 months) | ||||
| Mood stabilizer | 0 | 2 | 0 | |
| Antidepressants | 0 | 0 | 8 | |
| Antipsychotics | 0 | 3 | 0 | |
| Antianxiolytic | 2 | 0 | 3 | |
| 2 + psychiatric intervention | 0 | 9 | 3 |
Abbreviations: BD, bipolar disorder; BMI, body mass index; BRIAN, Biological Rhythms Interview Assessment in Neuropsychiatry; F, female; FAST, Functioning Assessment Short Test; M, male; MADRS, Montgomery–Åsberg Depression Rating Scale; MDD, major depressive disorder; MOCA, Montreal Cognitive Assessment; PMI, Postmortem Interval.
Age, years of education: data are shown as mean ± standard deviation. MADRS, FAST: data are shown as median and interquartile range (25–75 percentiles).
Participant clinical and demographic characteristics.
| Control | BD | MDD | P-value | |
|---|---|---|---|---|
| Blood protein analysis | ||||
| Age | 25.4 (2.3) | 25.9 (2.3) | 25.8 (2.2) | .335 |
| Sex (M/F) | 25/30 | 11/29 | 25/30 | .141 |
| Years of Education | 10.67 (2.76) | 9.70 (2.59) | 9.75 (1.98) | .300 |
| BMI | 26.60 (6.14) | 28.30 (5.75) | 25.66 (4.57) | .136 |
| Clinical Assessment | ||||
| MADRS | 0 (0 -2.00) | 7.5 (3.0–12.0) | 2.0 (0.0–4.00) | <.001 |
| FAST | 4.5 (2.0–8.25) | 12.5 (8.0–24.5) | 7.5 (3.0–12.0) | <.001 |
| BRIAN | 24.11 (4.98) | 38.00 (12.17) | 27.20 (6.75) | <.001 |
| MOCA | 23.31 (3.34) | 22.00 (3.82) | 23.39 (3.27) | .116 |
| Medication | ||||
| Mood stabilizer | – | 2 (9.1) | 5 (15.6) | .938 |
| Antidepressants | – | 16 (72.7) | 18 (56.3) | .565 |
| Antipsychotics | – | – | 3 (9.4) | .659 |
| Antianxiolytic | – | 7 (31.8) | 17 (53.1) | .459 |
| 2 + psychiatric intervention | 5 (26.3) | 9 (33.3) | .814 | |
| Blood gene expression analysis | ||||
| Age | 25.99 (2.13) | 25.75 (2.195) | 25.57 (2.24) | .559 |
| Sex (M/F) | 31/38 | 12/40 | 13/48 | .005 |
| Years of Education | 10.91 (3.59) | 9.72(3.16) | 10.50 (3.44) | .161 |
| BMI | 25.61 (5.68) | 28.60 (6.15) | 26.40 (5.74) | .058 |
| Clinical assessment | ||||
| MADRS | 0.0 (0.0-4.0) | 12.0 (2.0 -18.0) | 8.0 (2.0 -16.0) | <.001 |
| FAST | 6.0 (3.0-10.0) | 11.0 (7.0 -23.0) | 12.5 (5.0 -19.75) | <.001 |
| BRIAN | 25.34 (6.67) | 35.43 (11.02) | 33.18 (11.21) | <.001 |
| MOCA | 23.32 (3.68) | 22.12 (3.94) | 22.66 (3.45) | .201 |
| Medication | ||||
| Mood stabilizer | – | 4 (13.3) | 4 (13.3) | .580 |
| Antidepressants | 6 (66.7) | 19 (63.3) | 22 (73.3) | .836 |
| Antipsychotics | – | 5 (16.7) | 3 (10.0) | .675 |
| Antianxiolytic | 3 (33.3) | 16 (53.3) | 15 (50.0) | .778 |
| 2 + psychiatric intervention | 2 (14.3) | 10 (37.0) | 10 (40.0) | |
| Postmortem | ||||
| Age | 49.63 (2.43) | 45.80 (3.36) | 49.85 (1.53) | .23 |
| Sex (M/F) | 16/3 | 9/11 | 17/3 | .006 |
| PMI | 47.83 (4.25) | 46.33 (3.94) | 62.99 (5.99) | .09 |
| Cause of death | ||||
| Suicide | 0 | 15 | 20 | |
| Accident | 5 | 3 | 0 | |
| Natural | 14 | 2 | 0 | |
| Medication (last 3 months) | ||||
| Mood stabilizer | 0 | 2 | 0 | |
| Antidepressants | 0 | 0 | 8 | |
| Antipsychotics | 0 | 3 | 0 | |
| Antianxiolytic | 2 | 0 | 3 | |
| 2 + psychiatric intervention | 0 | 9 | 3 |
| Control | BD | MDD | P-value | |
|---|---|---|---|---|
| Blood protein analysis | ||||
| Age | 25.4 (2.3) | 25.9 (2.3) | 25.8 (2.2) | .335 |
| Sex (M/F) | 25/30 | 11/29 | 25/30 | .141 |
| Years of Education | 10.67 (2.76) | 9.70 (2.59) | 9.75 (1.98) | .300 |
| BMI | 26.60 (6.14) | 28.30 (5.75) | 25.66 (4.57) | .136 |
| Clinical Assessment | ||||
| MADRS | 0 (0 -2.00) | 7.5 (3.0–12.0) | 2.0 (0.0–4.00) | <.001 |
| FAST | 4.5 (2.0–8.25) | 12.5 (8.0–24.5) | 7.5 (3.0–12.0) | <.001 |
| BRIAN | 24.11 (4.98) | 38.00 (12.17) | 27.20 (6.75) | <.001 |
| MOCA | 23.31 (3.34) | 22.00 (3.82) | 23.39 (3.27) | .116 |
| Medication | ||||
| Mood stabilizer | – | 2 (9.1) | 5 (15.6) | .938 |
| Antidepressants | – | 16 (72.7) | 18 (56.3) | .565 |
| Antipsychotics | – | – | 3 (9.4) | .659 |
| Antianxiolytic | – | 7 (31.8) | 17 (53.1) | .459 |
| 2 + psychiatric intervention | 5 (26.3) | 9 (33.3) | .814 | |
| Blood gene expression analysis | ||||
| Age | 25.99 (2.13) | 25.75 (2.195) | 25.57 (2.24) | .559 |
| Sex (M/F) | 31/38 | 12/40 | 13/48 | .005 |
| Years of Education | 10.91 (3.59) | 9.72(3.16) | 10.50 (3.44) | .161 |
| BMI | 25.61 (5.68) | 28.60 (6.15) | 26.40 (5.74) | .058 |
| Clinical assessment | ||||
| MADRS | 0.0 (0.0-4.0) | 12.0 (2.0 -18.0) | 8.0 (2.0 -16.0) | <.001 |
| FAST | 6.0 (3.0-10.0) | 11.0 (7.0 -23.0) | 12.5 (5.0 -19.75) | <.001 |
| BRIAN | 25.34 (6.67) | 35.43 (11.02) | 33.18 (11.21) | <.001 |
| MOCA | 23.32 (3.68) | 22.12 (3.94) | 22.66 (3.45) | .201 |
| Medication | ||||
| Mood stabilizer | – | 4 (13.3) | 4 (13.3) | .580 |
| Antidepressants | 6 (66.7) | 19 (63.3) | 22 (73.3) | .836 |
| Antipsychotics | – | 5 (16.7) | 3 (10.0) | .675 |
| Antianxiolytic | 3 (33.3) | 16 (53.3) | 15 (50.0) | .778 |
| 2 + psychiatric intervention | 2 (14.3) | 10 (37.0) | 10 (40.0) | |
| Postmortem | ||||
| Age | 49.63 (2.43) | 45.80 (3.36) | 49.85 (1.53) | .23 |
| Sex (M/F) | 16/3 | 9/11 | 17/3 | .006 |
| PMI | 47.83 (4.25) | 46.33 (3.94) | 62.99 (5.99) | .09 |
| Cause of death | ||||
| Suicide | 0 | 15 | 20 | |
| Accident | 5 | 3 | 0 | |
| Natural | 14 | 2 | 0 | |
| Medication (last 3 months) | ||||
| Mood stabilizer | 0 | 2 | 0 | |
| Antidepressants | 0 | 0 | 8 | |
| Antipsychotics | 0 | 3 | 0 | |
| Antianxiolytic | 2 | 0 | 3 | |
| 2 + psychiatric intervention | 0 | 9 | 3 |
Abbreviations: BD, bipolar disorder; BMI, body mass index; BRIAN, Biological Rhythms Interview Assessment in Neuropsychiatry; F, female; FAST, Functioning Assessment Short Test; M, male; MADRS, Montgomery–Åsberg Depression Rating Scale; MDD, major depressive disorder; MOCA, Montreal Cognitive Assessment; PMI, Postmortem Interval.
Age, years of education: data are shown as mean ± standard deviation. MADRS, FAST: data are shown as median and interquartile range (25–75 percentiles).
We found reduced MANF serum levels in BD participants compared to that of HC in agreement with our hypothesis. Specifically, individuals with BD in a depressive state had a lower serum MANF concentration in comparison to HC (P = .031) and euthymic BD participants (P = .013; see Figure 1). In addition, there was no significant increase in peripheral MANF or GRP78 gene expression in BD or MDD compared to HC participants even when controlling for sex (P > .05, Figure 2A and B). Similarly, no significant increase was observed regarding MANF protein or MANF and GRP78 gene expression levels in postmortem tissue (Figure 3A–C, P > .05; Figure 3A).
Serum mesencephalic astrocyte-derived neurotrophic factor (MANF) protein concentration across bipolar disorder, major depressive disorder, and healthy control groups. Error bars represent the standard error of mean.
(A) Comparison of relative quantification of peripheral mesencephalic astrocyte-derived neurotrophic factor (MANF) bipolar disorder (BD), major depressive disorder (MDD), and healthy control groups (HC). (B) Comparison of relative quantification of peripheral GRP78 between BD, MDD, and HC. Error bars represent the standard error of mean.
(A) Mesencephalic astrocyte-derived neurotrophic factor (MANF) protein concentration in the prefrontal cortex (PFC) and the hippocampus in individuals with bipolar disorder (BD), major depressive disorder (MDD), and healthy control groups (HC). (B) MANF gene expression in the PFC and the hippocampus in individuals with BD, MDD, and HC. (C) GRP78 gene expression in the PFC and the hippocampus in individuals with BD, MDD, and HC. Error bars represent the standard error of mean.
DISCUSSION
We found a reduction of peripheral MANF protein concentration in individuals with BD in a depressive episode and no changes in peripheral MANF or GRP78 gene expression compared to HC. In the postmortem brain tissue, there was no change in MANF protein levels and MANF or GRP78 gene expression in BD and MDD compared to HC. These findings may corroborate existing research that has demonstrated abnormalities within the ER stress response in mood disorders.14,36 It is noteworthy that reduced MANF protein concentration was only observed when the BD group was stratified by mood state, with individuals with BD who were in a depressed state showing lower peripheral MANF concentrations compared to HC and those who were euthymic. It is therefore possible that depressive states in particular may be accompanied by alterations in UPR. Given that MANF protein concentration was lower and MANF/GRP78 gene expression remained unchanged, it is feasible that the observed results indicate abnormalities during protein translation in which gene transcription is unaltered but proteins are not being synthesized in sufficient amounts. This can potentially stem from a reduction of ribosomes and other proteins responsible for protein translation and protein modification within the cytoplasm.37 Furthermore, it is possible that certain medications used to treat BD can hinder protein translation by affecting ribosomal function. This explanation is consistent with a recent study investigating the effects of commonly used psychotropic interventions including aripiprazole, clozapine, and lithium on human neuronal-like pluripotent cells, which identified that certain medications used to treat BD can significantly decrease ribosomal gene expression and overall protein synthesis within the cytosol.38 As a result, if there is a reduction in ribosomes and other proteins responsible for protein synthesis, it is possible that this mechanism may reduce MANF protein abundance. Similarly, another possibility regarding this discrepancy may be caused by additional post-transcription regulation mechanisms disrupting protein formation. For example, mRNA stability, transport, and translation efficiency are factors that impact protein levels. Thus, it is possible that while MANF mRNA may not be impacted and therefore is like HC, there may be issues with microRNAs (miRNA) or RNA-binding proteins, that may underlie reductions in MANF protein levels. This would align with previous studies that have shown that miRNA can regulate the expression of neurotrophic factors such as MANF, CDNF, and BDNF.39,40 Lastly, it is possible that increased protein degradation may be a factor that underlies the discrepancy between MANF mRNA and protein concentration, where MANF mRNA is stable, but MANF protein is rapidly degraded. Abnormalities in MANF structure, not assessed by our current study, may be activating certain pathways such as the autophagosome–lysosomal system and the ubiquitin–proteasome system, which are pathways through which proteins are degraded. As a result, this may cause these pathways to target MANF in BD. Previous studies in mood disorders such as schizophrenia have shown increased proteins from the ubiquitin–proteasome system, compared to controls.41 However, these relationships between protein synthesis, gene expression, and protein degradation regarding MANF have not been investigated in BD and are therefore warranted to better understand deficits within the UPR pathway.
A strength of this study is the use of both postmortem brain tissue and peripheral blood to investigate MANF and GRP78 in BD, MDD, and HC. There is much debate regarding the correlation between markers analyzed within the periphery and the brain. For example, a recent meta-analysis reported a weak correlation between the abundance of inflammatory cytokines within the periphery and CSF, although the association was stronger in individuals with autoimmune diseases and those aged 50 years and above. In addition, it has been reported that other signaling proteins such as neurofilament light chain were strongly correlated between the periphery and CSF.42 As a result, it seems that there are varying levels of associations between peripheral and central biomarkers, which is why it is important to analyze biomarkers within both the periphery and the brain. Furthermore, cells within organs such as the pancreas and heart have been shown to produce and release MANF, therefore it is possible that peripheral MANF does not originate from the brain.20 Future studies should investigate the association between peripheral and cerebral MANF concentration.
Under ideal conditions, MANF has been observed to protect neuronal cells from death.43 A recent study investigating the role of MANF in response to amyloid β-peptide (Aβ) in neuronal cells and the brains of APP/PS1 transgenic mice (serving as a model of Alzheimer’s disease), found MANF expression to be increased in response to exposure to Aβ, a protein reported to cause the death of neuronal cells through increased ER stress and a key characteristic of Alzheimer’s disease (AD). The increase in MANF expression and recombination in human MANF treatment reduced Aβ-induced cell death. Furthermore, MANF protection against cytotoxicity caused by Aβ was also associated with the reduction of ER stress, as demonstrated by the inhibition of apoptosis-inducing proteins such as CHOP, caspase-3, and other UPR-related molecules including ATF6, GRP78, phosphorylated-IRE1, phosphorylated-eIF2α, XBP1s, and ATF4. It was also observed that MANF knockout cells had increased levels of these proteins, which suggests that MANF may function by inhibiting these proteins to alleviate ER stress.44 While the study discussed above pertains to AD, MANF is a novel biomarker that is of recent interest in BD, which is why further studies investigating the mechanisms through which MANF functions in BD are encouraged. Additionally, AD and BD share an association with each other, which may provide some insight into the potential mechanism through which MANF should be functioning in BD. For example, a recent genome-wide association study investigating the association between AD and BD reported a polygenic overlap between the 2 diseases, which suggests shared genetic influences that could explain the physiological basis of AD and BD.45 In addition, individuals diagnosed with BD are at high risk of developing AD, which further suggests an association between AD and BD.46,47
While studies in other diseases have demonstrated ideal MANF functioning during ER stress, studies of UPR in BD provide insight into UPR deficits and in turn demonstrate how increased ER stress occurs in the disorder. A study by Pfaffenseller et al.36 analyzed UPR in BD using lymphocytes treated with tunicamycin, which induces ER stress by inhibiting glycosylation. Cells from the control group exposed to tunicamycin showed significant differences in protein levels of the UPR markers GRP78, CHOP, and eIF2α-P; however, there was no change in these markers in the BD group, and post-exposure to tunicamycin led to a 2-fold increase in cell death compared to lymphocytes from HC alluding to a failure/deficit in the pro-survival UPR in response to ER stress in BD.36 Another study investigating the ER stress response in BD reported that the expression of genes responsible for protein folding (eg, XBP1, GRP78, calreticulin [CALR], and CHOP) was altered in BD, leading to an insufficient response to ER stress.14 Based on the findings by Pfaffenseller et al.36 and Hayashi et al.14, there may be some deficits in UPR during times of increased stress, which may occur during a depressive phase. This also coincides with our findings, which also suggest abnormalities in the UPR as there was no significant change in GRP78 gene expression in BD compared to HC and MANF abundance was also either unchanged (peripheral and postmortem gene expression) or lower (peripheral protein levels).
A number of limitations need to be considered when interpreting the results of our study. Firstly, causation cannot be inferred since this is a cross-sectional study. A longitudinal study examining MANF levels across different mood states and time (eg, mood “switches”) would allow for a better understanding of the relationship between MANF and bipolar depression. In addition, longitudinal studies including pre- and post-treatment analysis are required to obtain a better understanding of the possible mechanisms that underlie the observed association between MANF concentration and mood. BD participants were also medicated with a combination of psychotropic medications, which may impact MANF levels given that medications such as lithium have been reported to interact with the MANF pathway.28 Furthermore, we did not have enough samples of individuals with MDD who were in a depressive state as many were in remission, preventing us from stratifying the participants into subgroups to determine if mood states affected MANF and GRP78 levels within MDD. Lastly, the sample was not matched by sex in the peripheral gene expression analysis; however, even when controlling for sex, no significant differences were found between the groups. Despite these limitations, this is the first study using both peripheral and postmortem brain tissue to analyze MANF protein levels and MANF and GRP78 gene expression in a well-characterized population-based sample of young adults from the community.
In conclusion, this study found lower peripheral levels of MANF during the acute depressive state in BD. Longitudinal studies following the same individuals across different mood states are needed to confirm or refute this initial finding. Given recent evidence that lithium increases MANF levels,28 the study of MANF as a potential protective pathway in BD is encouraging.
Acknowledgments
This study was made possible through the financial support of the Baszucki Brain Research Fund and the Milken Institute, a philanthropic organization focused on advancing bipolar disorder research. We would also like to thank the team from Pelotas, Brazil, who provided the blood samples for the peripheral analysis. Lastly, we would also like to thank the Douglas Bell Canadian Brain Bank in Montreal, Canada, for providing the brain tissue.
Author Contributions
Mohammad Ali (Data curation [Lead], Formal analysis [Equal], Investigation [Lead], Methodology [Equal], Visualization [Equal], Writing—original draft [Lead], Writing—review & editing [Lead]), Bianca Wollenhaupt- Aguiar (Data curation [Lead], Formal analysis [Supporting], Investigation [Supporting], Methodology [Equal], Project administration [Lead], Resources [Lead], Writing—review & editing [Equal]), Yifan Wang (Formal analysis [Supporting], Investigation [Supporting], Methodology [Supporting]), Fahed Abu-Hijleh (Data curation [Equal], Formal analysis [Equal], Funding acquisition [Lead], Investigation [Equal], Methodology [Equal], Visualization [Equal], Writing—review & editing [Supporting]), Nicolette Rigg (Data curation [Equal], Formal analysis [Equal], Investigation [Equal], Methodology [Supporting], Writing—review & editing [Supporting]), Taiane de Azevedo Cardoso (Formal analysis [Supporting], Investigation [Equal], Resources [Lead], Writing—review & editing [Equal]), Imran Ahmed (Investigation [Supporting]), Ridhi Gopalakrishnan (Investigation [Supporting]), Luciano Dias de Mattos Souza (Investigation [Supporting], Resources [Lead], Writing—review & editing [Equal]), Karen Jansen (Investigation [Equal], Resources [Lead], Writing—review & editing [Equal]), Ricardo Azevedo da Silva (Investigation [Supporting], Resources [Lead], Writing—review & editing [Equal]), Thaise Campos Mondin (Investigation [Supporting], Resources [Lead], Writing—review & editing [Equal]), Flavio Kapczinski (Resources [Supporting], Writing—review & editing [Equal]), Fernanda Pedrotti Moreira (Formal analysis [Supporting], Investigation [Equal], Resources [Lead], Writing—review & editing [Equal]), Andrew Lofts (Writing—review & editing [Equal]), William D. Gwynne (Writing—review & editing [Equal]), Todd Hoare (Funding acquisition [Lead], Writing—review & editing [Equal]), Ram Mishra (Conceptualization [Equal], Funding acquisition [Lead], Writing—review & editing [Equal]), and Benício Frey (Conceptualization [Equal], Funding acquisition [Lead], Methodology [Equal], Resources [Equal], Supervision [Lead], Writing—review & editing [Equal])
Funding
This work was supported by the Milken Institute—Baszucki Brain Research Fund Bipolar Disorder Grant Award (#BD-0000000133)
Conflict of Interest
None declared.
Data Availability
The data used in this study re not publicly available due to participant privacy consent. However, readers are welcome to contact the corresponding author regarding any additional information on study data.
