2023 Impact Factor
2023 Impact Factor
Microtubules (MTs), along with actin microfilaments and intermediate filaments, make up the cytoskeleton, which is important for the structure and dynamics of cells [1]. Organization and remodeling of the MT network is essential for developing neurons to form axons, dendrites, and synapses. In mature neurons, MTs continue to maintain the structure of axons and dendrites, and serve as intracellular trafficking pathways that allow motor proteins to deliver specific cargo within the cell [2]. Cytoskeleton dysfunction has been implicated in the pathology of several neuropsychiatric diseases, such as schizophrenia, major depressive disorder, bipolar disorder [3,4], and various neurodevelopmental disorders including autism spectrum disorders and intellectual disabilities [5,6].
Social defeat stress (SDS) is a type of social stress induced by exposure to a dominant conspecific subject. The SDS paradigm has been used widely as an animal model of depression, anxiety disorders [7], and possibly schizophrenia [8,9]. However, few studies have investigated the effects of SDS on changes related to MTs. Eskandari Sedighi et al. [10] reported that chronic social stress decreased MT protein network dynamicity and polymerization in rats. However, no study has investigated the effects of SDS on MT polymerization. The expression of tubulin and its posttranslational modification (PTM) are involved in MT structure and function [11]. Isolation [12], restraint stress [13], and chronic mild stress [14] induce alterations in α-tubulin and acetylated tubulin expression. To our knowledge, no studies have assessed the effects of SDS on tubulin or its PTM. MTs interact with MT-associated proteins (MAPs), which influence MT polymerization, stability, and organization [15]. MT-associated protein 2 (MAP2) is the predominant cytoskeletal regulator within neuronal dendrites. SDS changes MAP2 expression in the hippocampus (HIP) of susceptible (Sus) mice [16]. However, the expression of MAP2 in the prefrontal cortex (PFC) of mice following SDS remains to be explored. Stathmin (STMN1), an MT-destabilizing protein and the phosphorylation are involved in the formation and disassembly of MTs [17]. We previously reported no change in STMN1 expression in the PFC, HIP, amygdala (AMY), or dorsal striatum; but decreased expression of phospho stathmin1 serine 16 (p-STMN1 [Ser16]) in the PFC, after SDS [18]. Stathmin2 (STMN2), another MT-destabilizing protein, is highly expressed in the developing nervous system [19]. The effects of SDS on STMN2, and the phosphorylation thereof, remain unknown. Lastly, emerging evidence indicates connections between the cytoskeleton and endoplasmic reticulum (ER) [20]. ER stress in mouse embryonic carcinoma P19 cells markedly reduced dendrite length and MAP2 expression levels [21]. We also reported alterations in ER stress proteins, such as 78-kDa glucose-regulated protein (GRP-78) and CCAAT/enhancer binding protein (C/EBP)-homologous protein (CHOP), after SDS [22,23]. However, no studies have attempted simultaneous measurement of MT-regulating proteins and ER stress proteins, or analysis of their correlation, after SDS.
The aim of the present study was to examine the expression levels of α-tubulin and its PTM, as well as MAP2, STMN1, and STMN2 and their phosphorylated forms, and GRP-78 and CHOP, in mouse PFC and HIP after SDS. MT stability was investigated using MT polymerization assay. In addition, the associations of ER stress proteins with MAP2 were explored.
At 7 weeks of age, adult male C57BL/6J strain mice weighing 18−25 g were housed in groups of five mice per cage. Retired male CD-1 mice, weighing 40−44 g and aged 15 weeks, were placed in separate cages. The animal room temperature was set at 22°C, and the dark/light cycle interval was 12 hours (lights on at 6 a.m. and off at 6 p.m). All experimental procedures were carried out in compliance with the National Institutes of Health’s Laboratory Animal Care and Use guidelines for experimental animals (NIH, Bethesda, MD, USA), and were approved by the IACUC (cuh-IACUC-151027-32) of Jeonbuk National University Medical School (Care and Animals 1986).
Following the 1-week habituation period, C57BL/6J mice (n = 60) were subjected to chronic social defeat for 10 consecutive days. After excluding mice wounded (n = 11) and outlied in SA test (n = 4), defeated mice (n = 45) were categorized into Sus (n = 22) and unsusceptible (Uns) (n = 23) groups based on their performance in the social avoidance (SA) test. After 1 day, the mice were sacrificed and brain tissues were obtained for Western blot (Sus 13 and Uns 11) and tublin polmerization assay (Sus 9 and Uns 12) (Fig. 1).
Standard SDS was induced in mice as described in our previous studies [18,24]. All male CD-1 mice were screened for aggressiveness by measuring the latency to attack a naive C57BL/6J mouse and were chosen for aggressiveness based on two criteria: first, the CD-1 mouse had to attack in at least two of the three daily 180-second screening sessions; second, the latency to initial aggression recorded during each session had to be less than 60 seconds [25]. Social defeat sessions were conducted under dim light on 10 consecutive days, and daily SDS was conducted between 14:00−17:00 hours. On day 1, each SDS C57BL/6J mouse was removed from its home cage and placed singly in the compartment of a CD-1 mouse. Behavior was observed, and the duration of each physical attack was timed for 5 minutes. The SDS mouse remained in the compartment in which it had been attacked, and the CD-1 mouse was placed in the opposite compartment, allowing continuous olfactory, visual, and auditory contact during the following 24 hours. Moreover, we checked the wounds every time after a social defeat bout. The mice with wound sizes greater than 1 cm were supposed to be removed based on the recommendation of a previous study [25]. To avoid contaminating data due to habituation with the aggressor’s cages, on a subsequent day, the C57BL/6J mouse was exposed to a new resident CD-1 aggressor mouse. The social defeat procedure lasted for 10 consecutive days. As a Con group, C57BL/6J mice were placed into equivalent cages with members of the same strain, which were changed daily.
Following completion of the social defeat procedure, SA test was performed on day 11 of the study, to categorize the defected mice into Uns and Sus groups. The defeated mouse was placed in an interaction box (42 × 42 cm) with an empty wire mesh cage (10 × 4.5 cm) located at one end. The first session was performed without the CD-1 mouse in the wire-mesh cage. The movement of the defeated animal was tracked for 2.5 minutes. After a 1-minute interval, a novel CD-1 mouse was introduced into the wire mesh cage, and the same defeated animal from the first session was placed into the box and tracked for another 2.5 minutes. The total time spent by the experimental mouse in an 8 cm-wide corridor surrounding the wire mesh cage (interaction zone) was automatically measured by SMART software (Panlab). The interaction ratio was defined as 100 × (interaction time with a target mouse present) / (interaction time without a target mouse present). An interaction ratio of 100 was used as the cut- off value, where scores < 100 were defined as “Sus” and scores ≥ 100 as “Uns” for both wild-type based on previous studies [26,27].
After brain extraction, target regions were immediately dissected out on an ice plate according to the Bregma zero coordinates: prefrontal cortex (PFC, +1.54 to +1.98 mm), hippocampus (HIP, −1.5 to −2.18 mm) (Supplementary Fig. 1; available online). The tissues (PFC, 10−12 mg; HIP, 18−22 mg) were quickly cryopreserved in liquid nitrogen and stored at −80°C until assay. The tissue samples were homogenized following standard pro-tocols. The resulting supernatant fractions were analyzed to estimate the protein concentrations using protein assays (Bio-Rad Laboratories). Protein samples (15 mg/15 ml) underwent gradient gel (4−15%) polyacrylamide gel electrophoresis (SMOBIO). After transfer, the membranes were treated with 0.25% glutaraldehyde at room temperature for 10 minutes in 0.2% Tween 20/Tris-buffered saline (TTBS) to crosslink tissue proteins that might destroy or mask immunogenic epitopes within the tissue [28,29]. The membranes were then blocked with 5% skimmed milk and incubated with a mouse monoclonal antibody against tubulin (1:50,000; Sigma Aldrich), acetylated- tubulin (1:50,000; Sigma Aldrich), tyrosinated-tubulin (1: 1,000; Sigma Aldrich), MAP2 (1:1,000; Sigma Aldrich), and CHOP (1:1,000; Sigma Aldrich); a rabbit polyclonal antibody against p-STMN1 (Ser16) (1:1,000; Cell Signaling Technology), phospho stathmin1 serine 25 (p-STMN1 [Ser25]) (1:500; Invitrogen,), phospho stathmin1 serine 38 (p-STMN1 [Ser38]) (1:1,000; Cell Signaling Technology) and phospho stathmin 2 serine 73 (p-STMN2 [Ser73]) (1: 1,000; Invitrogen); a rabbit monoclonal antibody against STMN1 (1:50,000; Abcam) and STMN2 (1:10,000; Abcam); or a rabbit polyclonal antibody against GRP-78 (1:1,000; Cell Signaling Technology) and β-actin (1:1,000; Cell Signaling Technology) at 4°C overnight. After washing five times with TTBS, the membranes were diluted (1:5,000) and incubated with peroxidase-labeled goat anti-rabbit IgG (H + L) (Vector Laboratories) at room temperature for 2 hours.
For isolating tubulin protein from mouse brain tissue, we referred to a previous method [30]. Particularly, the sacrificed mouse brain was taken out of the blood clot with paper, cleaned in lysis buffer, and then weighed. After that, we put them into tubes while simultaneously adding 500 ml of low depolymerization buffer (DB). The mixture was homogenized twice for about 30 seconds each time with a tiny probe on a POLYTRON® homogenizer PT1200E (Thomas Scientific), then centrifuged at 13,000 rpm for 60 minutes at 4°C in a TLA 100.3 Beckman desktop ultracentrifuge rotor. Equal amounts of high-molarity PIPES buffer (HMPB) enriched with ATP (1.5 mM final) and GTP (0.5 mM final), as well as anhydrous glycerol, were added and mixed to the supernatant. Before adding HMPB and glycerol to the supernatant, they were pre-heated to 37°C. This mixture was incubated for 1 hour in a 37°C water bath before being centrifuged for 30 minutes at 37°C on a TLA 100.3 Beckman rotor at 45,000 rpm. The supernatant was removed after centrifugation, and the MT pellet was dissolved in cold DB buffer and incubated on ice for 30 minutes. The depolymerized tubulin was then centrifuged for 30 minutes at 4°C in a TLA 100.3 at 30,000 rpm (first cold spin). The supernatant was collected after centrifugation, and HMPB (ATP + GTP) and glycerol (1/3 of the final volume) were added in the same manner as previously reported. The mixture was incubated for 30 minutes at 37°C in a water bath, and the polymerized tubulin was centrifuged for 30 minutes at 45,000 rpm in a TLA 100.3 Beckman rotor at 37°C. The MT pellet was resuspended in cold BRB80 after centrifugation, and then incubated on ice for an additional 10 minutes. Tubulin was centrifuged at 50,000 rpm for 30 minutes at 4°C in a TLA 100.3 Beckman benchtop ultracentrifuge rotor after this depolymerization procedure (second “cold spin”). The supernatant of high-purity tubulin was collected and flash-frozen in liquid nitrogen.
The purity of the tubulin protein was validated using the Coomassie blue staining method [31], and the tubulin concentration was calculated using the bicinchoninic acid test. Tubulin collected by phosphocellulose column chromatography were validated using sodium dodecyl sulfate gel electrophoresis (SDS-PAGE) to measure the purity of tubulins. However, only a limited quantity of these tubulin proteins were isolated, insufficient to carry out the tubulin polymerization assay (Supplementary Fig. 2; available online).
To obtain a sufficient amount of tubulin, we decided to extract it from the pooled brains of three mice. Then, after the final depolymerization, we discovered that the amount of high-molarity tubulin in the supernatant was sufficient to perform the tubulin polymerization experiment (Supplementary Fig. 3; available online).
TPA was performed using an HTS-Tubulin Polymeriza-tion Assay Kit (BK004P; Cytoskeleton, Inc.) according to the manufacturer’s instructions. For reaction analysis, tubulin was dissolved at a concentration of 4 mg/ml with general tubulin buffer plus 1 mM guanosine-5’-triphosphate (GTP) (G-PEM) + 5% glycerol buffer, 80 mM PIPES buffer (pH 6.9), 0.5 mM EGTA, 2 mM MgCl2, 1 mM GTP, and 5% glycerol (provided in the kit). A half-area 96-well plate was pre-warmed to 37°C for 30 minutes prior to the reaction. The test proteins, paclitaxel (10 mM; enhancer control) and nocodazole (5 mM; negative control) were dispensed at various concentrations (10 × final assay concentration; 10 ml/well) in the half-area 96-well plates preheated to 37°C. After 5 minutes, 100 ml of tubulin (4 mg/ml in G-PEM + 5% glycerol) was added and the kinetics were recorded at 340 nm every 30 seconds for 60 minutes using a plate reader (Molecular Devices).
To compare groups, one-way analysis of variance and a post hoc test were conducted using SPSS software (version 21.0; IBM Co.). Graphs were constructed using Prism 5 software (GraphPad Software Inc.). Interaction ratio (IR) and Western blot results are presented as the mean ± standard error of the mean. In all cases, p values < 0.05 were considered to indicate statistical significance.
Eleven mice were removed during social defeat stress due to their wound size being more than 1 cm, and 4 mice were excluded from analysis because the time in the interaction zone without CD1 of these mice was ‘0’. There was a significant difference in the IR among the three groups (F[2,67] = 111.4, p < 0.0001) (Supplementary Table 1; avail-able online). Post hoc tests revealed that the IR of the Sus and Uns groups were significantly lower (p = 0.0004) and higher (p < 0.0001) than the Con group, respectively, and the IR of the Sus group was significantly lower than that of the Uns group (p < 0.0001) (Supplementary Table 1, Supplementary Fig. 4; available online).
There was no significant difference in the α-tubulin expression level in the PFC and HIP among the three groups. However, acetylated α-tubulin expression was found to be significantly different among the three groups only in the PFC (F[2,31] = 4.823, p = 0.0153). The post hoc analysis revealed significantly higher expression in the Sus and Uns groups in comparison with the Con group (p = 0.0426 and p = 0.0177, respectively). Regarding the tyrosinated α-tubulin, significant differences were found in both PFC (F[2,31] = 8.362, p = 0.0014) and HIP (F[2,31] = 8.608, p = 0.0012) among the three groups. The post hoc test showed significantly higher expression in the Sus and Uns groups compared to the Con group in the PFC (p = 0.0023 and p = 0.0059, respectively) and HIP (p = 0.0011 and p = 0.0145, respectively).
The MAP2 expressions in the PFC (F[2,31] = 16.06, p < 0.0001) and HIP were significantly different among the three groups (F[2,31] = 19.67, p < 0.0001). In the PFC, post hoc analysis revealed significantly lower expression in the Sus group compared with the Con (p < 0.0001) and Uns groups (p = 0.0074). In the HIP, post hoc analysis revealed significantly lower expression in the Sus (p < 0.0001) and Uns groups (p = 0.0057) compared with the Con group, and the Sus group had significantly lower expression than the Uns group (p = 0.0307) (Table 1 and Fig. 2).
There were no significant differences in STMN1 and STMN2 expression levels in the PFC and HIP among the three groups. Significant decreases in the expression of p-STMN1 (Ser16) in the PFC and HIP were observed among the three groups (F[2,31] = 4.887, p = 0.0154; and F[2,31] = 5.762, p = 0.0076, respectively). Post hoc test results showed significant decreases in PFC expression in the Sus and Uns groups compared to the Con group (p = 0.0351 and p = 0.0231, respectively) and HIP (p = 0.0289 and p = 0.0083, respectively). With respect to p-STMN1 (Ser25) expression, significant differences among the three groups were evident only in the HIP (F[2,31] = 8.542, p = 0.0011). The post hoc test showed significantly higher expression in the Sus and Uns groups compared to the Con group (p = 0.0008 and p = 0.0274, respectively). Regarding p-STMN2 (Ser73), a significant group difference in expression was found only in the HIP (F[2,31] = 7.439, p = 0.0025). The post hoc test showed substantial increases in expression in the Sus and Uns groups compared to the Con group (p = 0.0157 and p = 0.0025, respectively) (Table 1 and Fig. 2).
Significant differences in GRP-78 expression were observed in the PFC and HIP among the three groups (F[2,31] = 5.797, p = 0.0073; and F[2,31] = 5.813, p = 0.0073, respectively). Post hoc analysis revealed significantly higher expression in the Sus than Con group in both the PFC (p = 0.0059) and HIP (p = 0.0055). Regarding CHOP, significant differences were found in both the PFC (F[2,31] = 8.362, p = 0.0031) and HIP (F[2,31] = 8.608, p = 0.0004) among the three groups. In the PFC, post hoc analysis showed significantly higher expression in the Sus (p = 0.0062) and Uns (p = 0.0069) groups compared to the Con group. In the HIP, post hoc analysis revealed significantly higher expression in the Sus than Uns and Con groups (p = 0.0004 and p = 0.0056, respectively) (Table 1 and Fig. 2).
The Uns group showed similar MT polymerization kinetics to the Con group. However, in the Sus group MT assembly took about twice as long to start, and there was a more gradual plateau and lower rate of tubulin polymerization (Fig. 3).
Brain development relies heavily on proper MT func-tion. A precise understanding of how stressors alter MT function would shed light on the pathophysiology of neuropsychiatric disorders. Extensive behavioral, neuroendo-crinological, neurophysiological, and molecular biological research using SDS as an animal model of depression and anxiety disorders has been conducted. However, few studies have examined the effects of SDS on MTs. By measuring MT-regulating proteins such as MAP2, STMN1 and STMN2, and MT polymerization, we confirmed significant effects of SDS in mice on MTs.
We observed increased levels of acetylated α-tubulin in the PFC of both Sus and Uns groups, but decreased levels of tyrosinated α-tubulin in both PFC and HIP of both groups. Acetylated α-tubulin is considered to reflect a stable, long-lived MT [32]. Several studies have reported that acetylation of α-tubulin increases MT stability [33-35]. However, considering the TPA results in the present study and lower rate of tubulin polymerization in defeated mice, especially in the Sus group, increased acetylated α-tubulin may indicate innate compensatory mechanisms to repair MT damage due to SDS, rather than a stable MT state. The findings regarding tyrosinated α-tubulin are difficult to interpret because in vitro the level of tubulin tyrosination/detyrosination is know to have no effect on tubulin polymerization [36] or the binding of MAPs [37,38]. The decreased level of MAP2 observed in this study is in line with a previous study in which the mRNA expression of MAP2 in the mouse HIP was significantly decreased by chronic SDS [16]. Given that MAP2 is a robust somatodendritic marker [39], and that dendritic shortening of medial PFC pyramidal neurons by SDS has been reported [40], it may be inferred that decreased MAP2 expression may have affected MT dynamic stability and decreased tubulin polymerization. This may be supported by our observation that the level of MAP2 and tubulin polymerization rate were lower in the Sus compared to Con and Uns groups. In relation to this, there is accumulating evidence that MT dysregulation contributes to the reduced dendritic complexity and synaptic density seen in the central neurons of patients with mood disorders and schizophrenia [41,42].
Our finding of no change in STMN1 levels is consistent with some previous studies [18,43], although Han et al. [44] reported decreased STMN1 expression in rat AMY and HIP after a single exposure to immobilization stress. It may be that STMN1 levels decrease after short-term stress, but recover to normal levels after long-term stress. The STMN2, which is highly expressed in the developing nervous system, plays an important role in the dynamic assembly and disassembly of growth cone MTs during axonal elongation [19]. The finding of no change in STMN2 expression suggests that STMN2 itself is not involved in the changes of MT function induced by SDS. We observed a significant decrease and increase of p-STMN1 (Ser16) and p-STMN1 (Ser25) in defeated mice, respectively. Phosphorylation at Ser16 is known to strongly reduce the ability of stathmin to bind to and sequester soluble tubulin, resulting in increased MT stability [45]. Also, p-STMN1 (Ser25) increases the “steady-state catastrophe frequency” at the plus and minus ends of MT as efficiently as unphosphorylated STMN1, and inhibits tubulin polymerization [45]. Therefore, it may be interpreted that decreased p-STMN1 (Ser16) and increased p-STMN1 (Ser25) levels act together to reduce MT stability and possibly to impaire axonal transport and synaptic activity. On the other hand, phosphorylation at Ser73 of STMN2 stabilizes MT in the brain of the mouse embryo [46]. The increased expres-sion of p-STMN2 (Ser73) in the HIP of defeated mice may be a self-repair mechanism against MT damage due to SDS. In summary, the findings regarding MAP2, p-STMN1 (Ser16 and Ser25), and TPA may reflect decreased MT stability, whereas those regarding acetylated α-tubulin and p-STMN2 (Ser73) are associated with compensatory pro-cesses.
The findings regarding GRP-78 and CHOP in this study replicate those of our previous study [23]. Given that GRP-78 is an antiapoptotic protein, and that CHOP is an apoptotic transcription factor [47], increased levels of both proteins imply that two responses occur simultaneously in the brain of defeated mice, i.e., ER stress and a counteracting process designed to overcome any damage caused by ER stress. It is particularly interesting that GRP-78 expression in both PFC and HIP was increased only in the Sus group compared to control, while the CHOP level in the HIP was significantly higher in the Sus group than Uns group. These findings indicate that the detrimental effect of SDS is greater in Sus mice. The molecular and genetic abnormalities underlying this susceptibility need to be further explored in future research. One critical remaining question is why most of the results in Uns group are similar to those in Sus group even though SI ratio in Uns group is greater compared to Sus group. It may be that there could be compensatory or different molecular mechanisms making Uns mice behave resiliently. This might signify that molecular mechanisms affecting resilience should be explored in other pathways than MT associated proteins. In sum, greater reduction in the MAP2 level and increases in GRP-78 and CHOP seem to be molecular markers of SDS susceptibility. Finally, the TPA results were partly in line with a previous study showing that polymerization was decreased under long-term chronic stress [10]. This is the first report of the effects of SDS on polymerization kinetics to use the TPA.
The present study had several limitations. First, we did not conduct a morphological analysis to determine whether changes in MT-regulating proteins induced by SDS can reduce dendrite arborization. Second, TPA was performed in pooled brain tissues, which makes it difficult to determine the specific brain regions affected by MT poly-merization. Nevertheless, this is the first report of the effects of SDS on MT-regulating proteins and MT polymerization kinetics.
In summary, we observed changes in acetylated α-tubulin, tyrosinated α-tubulin, MAP2, p-STMN1 (Ser16 and Ser25), p-STMN2 (Ser73), and GRP78 and CHOP levels in the PFC and/or HIP of defeated mice, and a lower polymerization rate in Sus mice. Notably, alterations of MAP2, GRP-78, and CHOP were greater in the Sus group compared to the Uns and/or Con groups. These findings suggest that SDS undermines MT stability and promotes greater alterations of MAP2, GRP-78, and CHOP; additionally, a lower tubulin polymerization rate could be a molecular marker for susceptibility to SDS. The manipulation of MT-regulating proteins and polymerization ability may lead to novel therapeutic interventions for social stress-related neuropsychiatric diseases.
No potential conflict of interest relevant to this article was reported.
Young-Chul Chung, Thi-Hung Le, Jung-Mi Oh, and Sung-Kun Chun had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Young-Chul Chung. Acquisition, analysis: Thi-Hung Le, Jung-Mi Oh, Fatima Zahra Rami, Ling Li. Statistical analysis: Young-Chul Chung, Thi-Hung Le, Jung-Mi Oh. Drafting of the manuscript: Thi-Hung Le, Young-Chul Chung critically revised the manuscript and approved the final version to be published. Supervision: Young-Chul Chung, Sung-Kun Chun. Project administration: Young-Chul Chung. Funding acquisition: Young-Chul Chung.
Fig. 1. Experimental procedure. A CD-1 mouse is housed in a compartment separated by a transparent acrylic divider containing many holes: (A) C57BL/6J mouse was transferred from its home cage to the CD-1 home cage; (B) social defeat was observed during 5 minutes; and (C) defeated mouse was separated by the divider for 24 hours. The following day, the C57BL/6J mouse was subjected to social conflict with another CD-1 mouse. This sequence of physical and psychological stress was repeated for 10 days to induce a social defeat model.
Fig. 2. Results of one-way analysis of variance (ANOVA) for Western blotting assay of the microtubule and endoplasmic reticulum stress proteins marker among three groups in the medial prefrontal cortex (PFC) and hippocampus (HIP) (*
Fig. 3. Effect of different ligands on tubulin polymerization 
CrossRef (0) Scopus (0) Download (388)
Author ORCID Information Thi-Hung Le Jung-Mi Oh Fatima Zahra Rami Ling Li Sung-Kun Chun Young-Chul Chung
Funding Information National Research Foundation of Korea (NRF) Ministry of Science and ICT, South Korea |
 |
