Role of Dopamine D2 Receptor in Stress-Induced Myelin Loss

Dopaminergic systems play a major role in reward-related behavior and dysregulation of dopamine (DA) systems can cause several mental disorders, including depression. We previously reported that dopamine D2 receptor knockout (D2R−/−) mice display increased anxiety and depression-like behaviors upon chronic stress. Here, we observed that chronic stress caused myelin loss in wild-type (WT) mice, while the myelin level in D2R−/− mice, which was already lower than that in WT mice, was not affected upon stress. Fewer mature oligodendrocytes (OLs) were observed in the corpus callosum of stressed WT mice, while in D2R−/− mice, both the control and stressed group displayed a decrease in the number of mature OLs. We observed a decrease in the number of active β-catenin (ABC)-expressing and TCF4-expressing cells among OL lineage cells in the corpus callosum of stressed WT mice, while such regulation was not found in D2R−/− mice. Administration of lithium normalized the behavioral impairments and myelin damage induced by chronic stress in WT mice, and restored the number of ABC-positive and TCF4-positive OLs, while such effect was not found in D2R−/− mice. Together, our findings indicate that chronic stress induces myelin loss through the Wnt/β-catenin signaling pathway in association with DA signaling through D2R.


Results
Chronic stress-induced myelin loss in WT and D2R −/− mice. WT and D2R −/− mice assigned to "stressed" groups were subjected to restraint stress by immobilization in a restrainer for 2 h daily for 14 d, as described previously 14 (Fig. 1A). Control groups of mice were left undisturbed. After stress exposure, WT and D2R −/− mice subjected to chronic stress were evaluated in the forced swim test. In this test, immobility time is considered an indicator of a depressive state. The stress-induced increase in immobility time in D2R −/− mice was significantly longer than that of WT mice (stress × genotype interaction: F 1,30 = 5.43, P = 0.027; Fig. 1B). This finding is in accordance with previous data showing increased levels of depressive behaviors in D2R −/− compared to WT animals following stress 14 .
We determined whether chronic stress affects myelination in WT and D2R −/− mice. Luxol fast blue (LFB) myelin staining was used for measuring the extent of myelin loss 23 . As a myelin loss control, we used the copper chelator cuprizone (CPZ), which can induce severe demyelination with oligodendrocyte damage in animals after 5 to 6 weeks of CPZ feeding 24,25 (Fig. 1C). We observed severe myelin loss of the corpus callosum of stressed WT mice (WT-St) compared to non-stressed control WT mice (WT-NSt) (Fig. 1C,D). By analyzing the density of LFB staining, the white matter of the corpus callosum spanning from bregma 1.10 mm to −2.30 mm was evaluated. We found that most of the corpus callosum between 0.14 mm and −0.82 mm was significantly demyelinated in the WT-St group mice as compared to the control group. This loss of white matter was comparable with the cuprizone (CPZ)-fed (for 5 weeks) animals, where the LFB staining intensity was decreased up to 55% in the corpus callosum. Interestingly, a low level of LFB staining density was already observed in non-stressed D2R −/− mice (D2R −/− -NSt) (51% versus WT-NSt group) and stressed D2R −/− mice (D2R −/− -St) group displayed a similar level of LFB staining density (Fig. 1C,D, stress × genotype interaction: F 1,18 = 5.65, P = 0.029). In parallel, we have also analyzed the representative myelin-specific protein, Myelin Basic Protein (MBP) 26 in the corpus callosum area by immunohistochemistry and observed a significant decrease of MBP expression in the WT-St mice compared to the WT-NSt control group (71% of WT-NSt group). In D2R −/− mice, the non-stressed control group showed a low level of MBP which was not affected by stress subjection (78% of control WT mice, stress × genotype interaction: F 1,20 = 11.33, P = 0.003; Fig. 1E,F). These findings suggest that myelin is decreased by stress in WT mice while the basal myelin level is low in D2R −/− mice, and that there is no change of myelin status by chronic stress in these mice.

Stress-induced myelin loss involves a decrease in mature oligodendrocytes in WT mice.
To examine in more detail the altered myelination in stressed WT and D2R −/− mice, we identified OL cell populations in the corpus callosum of each group and determined the number of oligodendrocyte precursor cells (OPCs) and mature OLs among all the OL lineage cells. PDGFRα (platelet-derived growth factor receptor, alpha polypeptide) is a known marker of oligodendrocyte precursors 27 , and it has been shown that the adenomatous polyposis coli (APC) protein is a marker for mature OLs 28 . Oligodendrocyte transcription factor (Olig2) is a basic helix-loop-helix (bHLH) transcription factor encoded by the Olig2 gene and Olig2 labels OPCs and mature OLs, and is thus indicative of all OL lineage cells 29 . The number of OPCs (PDGFRα+), APC-positive cells (APC+, representing mature OLs), and Olig2-positive cells (Olig2+, representing all OL lineage cells) were enumerated in control and stressed animals. By double immunofluorescence analysis, we found that in the corpus callosum of WT-St mice, Olig2+ and APC+ cells were decreased and APC+/Olig2+ cells were decreased by 47% of the WT-NSt control group ( Fig. 2A,C). In the corpus callosum of D2R −/− mice, both D2R −/− -NSt and D2R −/− -St groups of mice displayed a significant decrease of Olig2+ and APC+ cells, and APC+/Olig2+ cells by 48-53% of WT control mice (D2R −/− -NSt mice, 48% of WT-NSt mice; D2R −/− -St mice, 53% of WT-NSt mice, stress × genotype interaction: F 1,24 = 4.55, P = 0.043, Fig. 2A,C). In contrast, the number of PDGFRα+ cells was comparable across the corpus callosum of both WT and D2R −/− mice, and was not changed by stress (stress × genotype interaction: F 1,28 = 0.54, P = 0.47, Fig. 2B,D). These results suggest that chronic stress-induced myelin loss involves a decrease in the number of mature OLs in WT mice and in D2R −/− mice, the latter which already had decreased number of mature OLs and similar level of OLs number is observed in D2R −/− -St mice. . Approximately 1500 axons from 5 mice were investigated in each group. Average myelin thickness (H) and g-ratio (I) in the middle corpus callosum were measured and grouped by axon diameter (J,K) (n = 5 mice for each group). White bars indicate non-stressed animals, black bars indicate stressed animals and gray bars indicate cuprizone-fed animals. Data are presented as mean ± SEM. **P < 0.01, ***P < 0.001 versus NSt mice; # P < 0.05, ## P < 0.01, ### P < 0.001 versus WT-NSt mice; + P < 0.01 versus genotype (Two-way ANOVA post hoc test). § P < 0.05, § § P < 0.01 versus WT-NSt mice (Two-tailed Student's t-test). Wnt/β-catenin-TCF signaling is associated with stress-induced myelin loss. Further, we examined the mechanism by which stress-induced myelin loss is mediated and determine why a low level of myelin is observed in D2R −/− mice despite similar level of OPC cells as WT mice.
Recent studies have shown that the Wnt/β-catenin-TCF/LEF signaling pathway is critical in the regulation of oligodendrogenesis, OL differentiation, and myelination [29][30][31][32][33][34][35] , although the exact role of Wnt signaling in the regulation of myelination appears contradictory. We recently found that the D2R can interact with the Wnt protein, suggesting that Wnt may play an important role in dopaminergic neuronal development through the ERK pathway 36,37 . Based on these findings, we hypothesized that Wnt/β-catenin-TCF/LEF signaling might be involved in the stress-induced myelin loss associated with D2R.
Double immunofluorescence analysis with antibodies against ABC and TCF4 revealed that the number of ABC+/TCF4+ double-labeled cells were significantly decreased in WT-St mice (22 ± 5% of WT-NSt mice, n = 7, Fig. 3B,D). Non-stressed D2R mice, which were previously shown to have a reduced number of ABC+/TCF4+, showed a reduction of ~48% of control WT mice (48 ± 11%, n = 6); this change was nominally affected by stress in D2R −/− mice (42 ± 17% of control WT-NSt mice, n = 6, Fig. 3B,D). These results suggest that the β-catenin/ TCF4 signaling is an important mediator in stress-induced myelin loss and that the absence of D2R blunts this regulation.

Effect of lithium on stress-induced myelin loss.
Since the chronic stress-induced depressive state evoked such significant myelin loss, we next investigated the ability of an antidepressant to modulate myelination.
Effect of D2R activation on stress-induced myelin loss. We next examined the expression of D2R in the oligodendrocytes corpus callosum using fluorescent in situ hybridization (FISH) with D2R-specific probe (RNAscope in situ hybridization). With the same section, we performed the immunohistochemistry with Olig2 antibody to identify D2R expression in the oligodendrocytes. This combined FISH and immunohistochemistry revealed that co-expression of D2R mRNAs (red) and Olig2 proteins (green) in the corpus callosum with approximately 26 percent of Olig2-positive cells found to express D2R (Fig. 6A). As a positive control, same D2R probe  was used to perform FISH with brain sections containing caudate putamen (CPu) area where D2R expression is abundant (Fig. 6A, right panel).
We analyzed whether stress exposure or lithium treatment can modulate the expression of D2R in the corpus callosum, using FISH analysis. We observed that D2R mRNA expression was not significantly changed by chronic stress, nor by the lithium treatment in the corpus callosum (Fig. 6B,C).
Quinpirole treatment effectively prevented the myelin loss in stressed WT mice, as revealed by LFB and MBP staining (Fig. 6F,G). LFB staining within the middle corpus callosum (eight sections, region from bregma 0.02 mm to −0.82 mm, that showed severe myelin loss with LFB and MBP staining, Fig. 1C-F) showed an increase in the intensity of LFB staining in quipirole-treated stressed WT mice, (WT-St mice, n = 6, 46 ± 8% of control WT-NSt mice, n = 6; 96 ± 5%, n = 6 in quinpirole treated WT-St mice, Fig. 6F,G). However, we detected no effect of quinpirole in D2R  Fig. 6H,I). Together, these findings suggested that activation of D2R prevented chronic stress-induced myelin damage, and this effect was not seen in mice lacking D2R.

Discussion
Increasing evidence suggests that abnormalities in myelin are associated with mental illness, pointing to an important role of oligodendrocytes in this disease 16,41 . However, the underlying molecular mechanisms remain unclear. Anatomical or brain imaging studies of patients with severe mental illnesses, including major depressive disorder (MDD), bipolar disorder (BD), and schizophrenia (SZ) have revealed deficiencies in brain myelination 19,[42][43][44][45][46] . Expression profiling studies using post mortem brains of patients with major depressive disorder indicated that depressive disorder might be associated with changes in genes related to myelin, oligodendroglia, and synaptic functions 15,19 . In animal models of depression, reduced glial cell number and morphological changes in OLs are observed [47][48][49] ; these effects are similar to the abnormalities in OLs reported in patients with depression. These alterations in OLs would induce impairment of myelin integrity, which in turn can deteriorate the axon-myelin interactions, and could be key cellular and molecular events implicated in synaptic modification occurring in chronic-stress induced depression. Here, we mainly investigated the corpus callosum, because the corpus callosum is the largest white matter structure in the brain, having high content of myelin. Numerous findings demonstrated that abnormal myelination in the corpus callosum is associated with bipolar disorder [50][51][52][53][54] . Benedetti et al., for example, reported that adverse childhood experiences alter white matter microstructure including corpus callosum, in patients with bipolar disorder 55 . As well, brain imaging studies with major depressive disorder (MDD) patients showed that MDD patients have widespread abnormalities in the corpus callosum [56][57][58] . Therefore, together with other animal studies 59-61 , these observations indicate a strong association between alterations of the myelination of the corpus callosum with chronic stress-induced depressive behaviors.
In the present study, we demonstrated that chronic stress induced-myelin loss was regulated by Wnt/β-catenin signaling, and that this regulation was dependent on the DA system through D2R. Recent findings identified a role of the Wnt/β-catenin signaling in myelination, with initial reports suggesting, rather, an inhibitory role for Wnt signaling in OPC differentiation and OL maturation during early postnatal development 29,31,62 . Other studies using genetic and pharmacological antagonists to inhibit components of Wnt/β-catenin signaling also demonstrated that inhibition promoted OPC differentiation and remyelination in both in vitro and in vivo models [63][64][65][66] . In contrast, numerous reports have shown that activation of Wnt signaling induces myelination. For instance, Tawk et al. 32 , have demonstrated a key role for Wnt/β-catenin signaling in the expression of myelin-related genes and myelin sheath compaction in both in the peripheral and central nervous systems using loss-of-function analyses in zebrafish embryos. Inhibition of Wnt/β-catenin signaling resulted in hypomyelination without affecting Schwann cell and OL generation, or axonal integrity. It has also been reported that in vivo activation or inhibition of canonical Wnt signaling, respectively, increased or decreased the number of Olig2-positive and PDGFRα-positive cells in the adult subependymal zone, thereby showing the contribution of Wnt/β-catenin region from bregma 0.02 mm to −0.82 mm, WT n = 6, D2R −/− n = 4). Data are presented as mean ± SEM. White bars indicate non-stressed animals and black bars indicate stressed animals. NSt: non-stressed mice, St: stress-imposed mice. ***P < 0.001 versus NSt mice; ## P < 0.01, ### P < 0.001 versus genotype. +++ P < 0.001 versus lithium treatment (Two-way ANOVA post hoc test).
Scientific RepoRts | 7: 11654 | DOI:10.1038/s41598-017-10173-9 pathway to oligodendrogliogenesis, myelin gene expression, and myelinogenesis 67 . Furthermore, it was demonstrated that in Tcf4 gene KO mice, there are few changes in the number of OPCs; however, there is severe defects in oligodendrocyte maturation 62,68 . It has been reported that Tcf4 is expressed during remyelination stages in patients with multiple sclerosis and in the cuprizone mouse model of myelin loss, thereby supporting a role for Tcf4 in remyelination 62,[68][69][70] . These contradictory findings indicate that Wnt pathway plays an important role in OL maturation and myelin formation, most likely in a complex manner, which depends on developmental stage and corresponding molecular networks. One possible explanation is that interaction of the Wnt signaling with other pathways can tightly adjust the balance between the negative and positive effects on myelination. In this context, our present finding suggests a possible link between DA and Wnt system through D2R regulation of remyelination, particularly under the stressful stimuli.
Given that lithium, generally known as a glycogen synthase kinase-3β inhibitor, treatment normalized chronic stress-induced Wnt signaling inactivation we also examined whether lithium treatment in the modulation of myelin in corpus callosum would involve glycogen synthase kinase-3β (GSK3β) signaling. However, immunofluorescent histochemistry performed using antibody to GSK3β revealed no signal of GSK3β in the corpus callosum ( Supplementary Fig. 1A) while robust labeling of GSK3β could be detected in the cingulate gyrus (cg), caudate putamen (CPu) and primary somatosensory cortex (S1 cortex) ( Supplementary Fig. 1B,C). Therefore, based on our data, we believe that the regulation we observed here involves rather a link between dopamine/lithium and Wnt/ β-catenin/Tcf4 signaling without direct involvement of GSK3β in this regulation of myelin in the corpus callosum.
Dopamine systems are considered important modulators of stress-related depressive illness which is associated with altered response to reward, thereby contributing to the pathophysiology of stress-evoked depression [2][3][4]6,71 . This deficiency in DA transmission involves not only reduced DA release but also abnormal density and sensitivity of dopamine receptors, resulting in dysfunction of DA receptors and disrupted DA transmission in response to external stimuli 72,73 . How dopamine systems with dysfunctional DA receptors contribute to the pathophysiology of depression is not yet known.
It has been reported that the dopamine system can regulate OL functions. For example, D2R expression was observed in interfascicular OLs around the peak of myelination, from P15 to P20, in rat brain 21 and in differentiated OLs from rat cortical OL cultures 22,74 . In addition, a protective role of D2R against oxidative glutamate toxicity and oxygen-glucose deprivation injury in OLs 22 has been demonstrated as chronic treatment with the D2R antagonist haloperidol reduced expression of myelin protein genes in mice 75 . Disrupted integrity in myelination observed in the corpus callosum of D2R −/− mice in the present study thus raises the possibility that D2R is a potential regulator of OL development and myelination formation. This is consistent with the observations made in this study, in which quinpirole treatment in stressed WT mice protected stress-induced myelin damage, while no such effect is observed in D2R −/− mice.
Another interesting finding of the present study is that in D2R −/− mice, upon chronic stress, Wnt singling pathway-mediated regulation of myelination was blunted, which supports the hypothesis that DA signaling through D2R, in association with the Wnt pathway, is critical for stress-induced myelin loss. Since stress can affect axon-myelin interaction, myelin formation and remyelination in response to stress must be tightly controlled by neurotransmission. Based on our present findings, we suggest that DA, via D2R, can modulate the Wnt pathway, which in turn controls stress-induced myelin loss, thereby contributing to the regulation of white matter integrity in stress-induced neurological disorders. Our findings may explain the complexity of antidepressant treatment resistance, known to be related a hypodopaminergic state, which can result in lack of effect owing to the absence of DA system activation. Indeed, it has been demonstrated that some D2R agonists such as pramipexole and aripiprazole are effective in patients with depression who have failed to respond to previous medications 9, 10,76,77 . Our finding that the effect of lithium treatment was dependent on the presence of D2R may be related to this clinical situation. Consideration of DA signaling would not only be particularly important for synaptic modification in the regulation of myelination associated with the Wnt/β-catenin axis but also be promising as effective antidepressant medication.

Chronic Immobilization Stress.
Behavioral experiments were performed with male D2R −/− and WT control mice at 11-13 weeks of age. Age-matched WT and D2R −/− mice were housed individually and allowed to acclimatize to the cage for 1 week. Mice assigned to stress groups were restrained in a ventilated acrylic restrainer fit to allow the animal to breathe but not to move otherwise. For chronic stress, they were immobilized for 2 h once daily for 14 days. WT and D2R −/− mice were placed in the home cage for restraint stress. Mice assigned to non-stressed groups were untouched in the home cage, whereas their counterparts were subjected to restraint stress. Luxol fast blue stain. Six sections were prepared from specific regions within the middle corpus callosum that showed severe myelin loss in LFB stain (region from bregma 0.14 mm to −0.82 mm, 0.16-mm intervals). Brain sections were dehydrated with ethanol grades, incubated in 0.1% luxol fast blue stain solution (solvent blue 38 0.1 g, ethyl alcohol 95% 100 ml, glacial acetic acid 0.5 ml) at 56 °C overnight, and washed with distilled water. Sections were dipped in 0.05% lithium carbonate 30 sec and 70% regent alcohol for gray and white matter differentiation. The slices were incubated in cresyl violet stain for 1 min, dipped in 70% regent alcohol, dehydrated through 2 changes of absolute alcohol, and mounted with Permount (Fisher Scientific). The myelin loss area and integrated optical density (IOD) were analyzed with MetaMorph 6.1 software (Universal Imaging Corp.; Downingtown, PA).

Electron microscope histological analysis (TEM).
After anesthesia, mice were perfused with 4% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M phosphate buffer. Dissected tissues containing the corpus callosum (bregma −0.46 mm) were prepared from 1 mm parasagittal serial brain sections. The tissues were buffered washed, fixed in 1% osmium tetroxide, dehydrated in graded ethanol series, transitioned in propylene oxide, and embedded in TAAB resin (TAAB Laboratories). Semi-thin sections (1 µm mid-sagittal, lateral 0 mm) were cut, mounted on glass slides, and stained in toluidine blue for general assessment in the light microscope. Subsequently, 50~60-nm ultra-thin sections were cut with a diamond knife (Diatome), mounted on copper slot grids coated with parlodion, and stained with uranyl acetate and lead citrate for examination on an electron microscope (Hitachi; H7650). Ten random images within each field were collected at 12000× magnification. The digitalized images were then analyzed on a MetaMorph image analysis system (Universal Imaging Corp.).
G-ratio analysis. Digitized, non-overlapping electron micrographs of corpus callosum were analyzed for axon diameter and g-ratio. Images were analyzed using MetaMorph 6.1 software. The g-ratio measurements were obtained by manually drawing lines across 2 perpendicular diameters for both the axons and for the axons plus myelin sheath. At least 300 axons per brain were measured. The analysis was modified such that the inner diameter of compact myelin (instead of the axon diameter) was divided by the outer diameter of the myelin sheath (because the inner tongue was frequently enlarged in demyelinated animals, but not in controls). Axons with diameters typical of unmyelinated fibers (<0.3 mm) were excluded from analysis. Fibers with prominent outfoldings in the plane of section were also excluded. Fluorescent in situ hybridization (FISH) and Olig2 immunofluorescence-histochemistry. Ani mals were anesthetized and transcardially perfused with 4% paraformaldehyde in PBS. Brains were removed and postfixed 4 h at 4 °C. Dehydrate the tissue in 10%, 20%, 30% sucrose in 1X PBS at 4 °C until the tissue sinks to the bottom of the container. Tissue were frozen at −20 °C in the Optimal Cutting Temperature (OCT) embedding media and 14 μm-thick coronal sections were placed on SuperFrost ® Plus slides (Fisher Scientific #12-550-15) and air dry the slides for 20 min at −20 °C. FISH was performed according to RNAscope ® ISH Tissue instructions (Advanced Cell Diagnostics (ACD), ACD 320850, Hayward, CA) for fixed-frozen tissue. Slides were digested with protease (1:15, diluted in PBS) for 30 min at 40 °C. Sections were incubated for 2 h at 40 °C in a hybridization oven (HybEZ ™ , ACD) with RNAscope ® Probe-mm-Drd2-C3 (ACD 406501-C3) probes to detect mouse D2R mRNA. Ppib probe (ACD 313911-C3), a mouse housekeeping gene, was used as a positive control probe, and a bacterial dapB probe (ACD 310043-C3) was used as a negative control. Sections were then incubated with Amp 1-FL, Amp 2-FL, Amp 3-FL, and 4-FL-Alt A solutions according to instructions. Wash with 0.1 M PBS twice, blocked with a blocking solution (5% bovine serum albumin, and 0.2% Triton X-100 in PBS, pH 7.4) for 1 h at room temperature. Tissue sections were incubated overnight at 4 °C with a-Olig2 (1:500; Millipore, AB9610). After rinsing in PBS, the double-stained sections were incubated at room temperature for 1 h with Alexa Fluor 488 anti-Rabbit (donkey IgG) (1:1000; Life Technologies, A-21206). After counterstaining with DAPI for 30 sec, slides were cover-slipped with Vectashield (Vector Laboratories) to prevent fading. Slides were imaged using confocal laser scanning system, (LSM 700, Zeiss, Berlin, Germany). Probe channel C3 was observed with a filter for Atto 647. To count D2R expressing cells in Olig2-positive cells in WT mice (Fig. 6A), cells were counted in the corpus callosum region within the range from bregma −0. 22 mm to −0.70 mm (4 sections per mouse) and summed data were used for quantification. Approximately 400 cells expressing D2R or Olig2 were counted and only cells with a D2R mRNA puncta fluorescent signals or Olig2 fluorescent signals that filled the cell body were included for analysis. Individual puncta were counted using MetaMorph image analysis system (Universal Imaging Corp.). In a similar way, the expression levels of D2R mRNA transcripts in stress-imposed WT mice with and without lithium treatment (Fig. 6C) was analyzed, puncta were counted in the corpus callosum region within the region from bregma −0.22 mm, −0.46 mm and −0.70 mm (3 sections per mouse) and summed data were used for quantification.

Forced swim test.
Mice were transferred to the experimental room 60 min before the onset of the experiment to allow for habituation and reduce stress (brightness of the experimental room was 70 lux). The forced swim test was performed by dropping a mouse in water and recording its behavior. The apparatus consisted of a plastic cylinder (internal diameter, 10 cm; height, 25 cm) containing a 19 cm column of water maintained at 22 ± 2 °C. The mouse was thus not able to support itself by touching the bottom of the apparatus with its paws or tail. An animal was considered to be immobile when it floated in an upright position and made only small movements to maintain its head above water. The total duration of immobility, which is thought to reflect a depressive state, was determined during a 6-min test session.
Statistical analysis. Data are presented as means ± s.e.m. and were analyzed with one-way or two-way analysis of variance (ANOVA) followed by Bonferroni's post hoc test. A P-value of < 0.05 was considered statistically significant.