Tolcapone-treated mice exhibit DA elevation, hypomyelination, mitochondrial dysfunction in brain cells, and behavioral abnormalities
To determine if DA elevation influences white matter development and impairs mitochondrial functions of brain cells while inducing behavioral anomalies, we administered TOL to C57BL/6 mice intraperitoneally at the dose of 0, 15, 30, or 60 mg/kg, for consecutive 14 days starting on postnatal day (PD) 22. TOL is a brain penetrant selective inhibitor of COMT devoid of psychostimulant properties [25]. The COMT enzyme plays a pivotal role in DA metabolism, specifically in PFC [26]. The four groups of mice were referred to as VEH, TOL15, TOL30, and TOL60 (n = 17/group), respectively. One day after the last TOL administration, the animals were subjected to OFT measuring their locomotor activity and anxiety-like behavior, the Y-maze test assessing spatial working memory, NOR test evaluating recognition memory, and POT examining general cognition and executive function, in the order. Compared to VEH group, the groups TOL30 and TOL60 showed higher anxiety levels indicated by significantly lower CD/TD (p < 0.01, p < 0.05, respectively, Fig. 1A). The TOL60 group also presented spatial working memory impairment indicated by significantly lower spontaneous alternation in the Y-maze test (p < 0.05, Fig. 1B). In addition, the TOL60 group showed a significantly lower RI as compared to the VEH group (p < 0.05, Fig. 1C), suggesting a recognition memory impairment. Moreover, TOL30 and TOL60 groups showed impairment in problem solving ability and executive function revealed by POT in which mice in TOL30 and TOL60 groups took longer duration to complete the task in T5 and T8 (p < 0.001, p < 0.05, respectively, Fig. 1D).
One day after the last behavioral test, five mice in each group were euthanized and the brain regions of PFC, CPu, and hippocampus were isolated and used for HPLC analysis to measure levels of DA, norepinephrine (NE), and 5-hydroxytryptamine (5-HT). As for DA levels, one-way ANOVA revealed significant effects of TOL on DA level in PFC (p < 0.01). Post-hoc comparisons showed significantly higher DA level in PFC (p < 0.05) of the mice in TOL60 group, compared to the VEH group (Fig. 1E). Regarding NE level, one-way ANOVA revealed a significant effect of TOL on this index in the hippocampus (p < 0.01). Post-hoc comparisons showed significantly higher NE levels in hippocampus of the mice in TOL30 and TOL60 groups (p < 0.05 in both comparisons, Fig. 1F). Also, TOL30 group had a higher level of 5-HT in hippocampus compared to the VEH group (p < 0.05, Fig. 1G).
Immunohistochemical staining with the primary antibody to APC (adenomatous polyposis coli gene clone CC1 used as the biomarker of mature OL cell body) showed dose-dependent decreases in the number of APC+ cells in PFC of TOL-treated mice (Fig. 2A & B). Relevantly, Western blot analysis revealed dose-dependent decreases in protein levels of CNP, MAG, MBP, and MOG in PFC of TOL groups compared to VEH group (Fig. 2C & D). To provide further evidence for the hypomyelination in TOL-treated mice, PFC samples of mice in VEH and TOL60 groups were prepared for TEM analysis. The two groups look seemingly different in numeral density of myelinated axon and myelin sheath thickness as shown in Fig. 2E. Indeed, quantitative data revealed a lower numeral density of myelinated axon in the TOL60 group (p < 0.001, Fig. 2F), a higher G-ratio (the ratio of the inner to the outer diameter of the myelin sheath of a myelinated axon) in the TOL60 group (p < 0.001, Fig. 2G), indicating thinner myelin sheath of the myelinated axons in this group. But, the two groups were comparable in axon diameter (p = ns, Fig. 2H). Moreover, the TOL60 group showed a smaller slope of the correlation between axonal diameter and G-ratio (p < 0.001, Fig. 2I) compared to the VEH group, suggesting that the smaller axons were more susceptible to DA elevation in PFC of the TOL-treated mice.
Similar to the findings in PFC, TOL administration decreased the number of APC+ cells (Supplemental Fig. 1A & B) and expression levels of CNP, MBP, and MOG in the mouse hippocampus (Supplemental Fig. 1C & D). As for CPu, TOL30 and TOL60 groups showed fewer APC+ cells relative to VEH group (Supplemental Fig. 1E & F), but the four groups were comparable in protein levels of CNP and MBP in this brain region (Supplemental Fig. 1G & H).
The coexistence of DA elevation, APC+ cell decrease, and hypomyelination in the TOL-treated mice raised a question, namely, how did DA elevation lead to mature OL decrease and hypomyelination in the mouse brain? We hypothesized that DA elevation impaired mitochondrial function thus inhibited OL maturation and axonal myelination as intact mitochondria in both neurons and OLs are essential for axonal myelination [27, 28]. To test this hypothesis, we assessed mitochondrial functions of brain cells of the mice in this experiment. Western blot analysis showed significantly lower levels of mitochondrial complexes I, II and IV in neural cells of mouse PFC in TOL30 and TOL60 groups relative to the VEH group, while the other two complexes (III & V) were comparable across the groups (Fig. 3A & B). In addition, TOL groups show lower levels of NAT8L (N-acetyltransferase 8-like) in mouse PFC (Fig. 3C & D). This enzyme catalyzes the synthesis of NAA from aspartate and acetyl-CoA in neuronal mitochondria. The neuronal NAA is then transported to the cytoplasm of OLs, where aspartoacylase (ASPA) cleaves the acetate moiety of NAA for use in the synthesis of fatty acid and steroid, the building blocks for myelin lipid synthesis [29, 30]. Furthermore, the TOL30 and TOL60 groups had significantly lower levels of ATP compared to VEH group (Fig. 3E). However, no difference was found between VEH and TOL groups in neuron number in the PFC (Supplemental Fig. 2), suggesting that neurons are more tolerable to mitochondrial dysfunction than OLs, the number of which decreased in TOL-treated mice as described above. This statement is consistent with a recent study reporting that neurons among the neural cells in mouse brain are the most tolerable to mitochondrial damage by cuprizone [31], which is a copper chelator and toxic to mitochondria [31, 32].
COMT -ko mice exhibit dopaminergic changes, hypomyelination, mitochondrial dysfunction in brain cells, and behavioral abnormalities
To substantiate the aforementioned changes in TOL-treated mice, we assessed and compared dopaminergic measurements of wild type (wt) and COMT-ko mice by means of various techniques. The COMT gene is located in a fragment of chromosome 22q11 which when deleted results in a complex syndrome including the psychiatric manifestations such as schizophrenia. As such, the COMT gene has been placed near the top of a list of plausible candidate genes for schizophrenia [33] and the gene variants may be involved in the pathogenesis of psychotic symptoms, and associated especially with negative symptom in schizophrenia [34, 35].
We assessed protein levels of the enzymes relevant to DA metabolism including monoamine oxidase-A (MAO-A), MAO-B, and Dopa decarboxylase (DDC), in addition to COMT in wild type and COMT-ko mice (n = 7/group). Western blot analysis detected the presence of COMT protein in both PFC and CPu of the wt mice, but the absence in COMT-ko mice (Fig. 4A). Compared to the wt mice, COMT-ko mice showed lower MAO-A level in CPu (p < 0.05, Fig. 4B & C). Interestingly, COMT-ko mice showed MAO-B level changes in opposite directions in PFC and CPu, i.e. higher level in PFC (p < 0.05) but lower level in CPu (p < 0.01) as compared to wt mice (Fig. 4D & E). The two groups were comparable in DDC levels in both PFC and CPu (Fig. 4F & G). HPLC results showed that COMT-ko mice had a significantly lower DA level in CPu compared to the wt mice (p < 0.05), whereas the two groups were comparable in DA level in PFC (Fig. 4H). The RT-qPCR analysis showed a higher level of DR1 mRNA in PFC (p < 0.01), but lower levels of DR2 (p < 0.05) and DAT (p < 0.01) mRNAs in CPu of COMT-ko mice compared to the wt mice (Fig. 4I).
Immunohistochemical staining showed decreased number of mature OLs in PFC of COMT-ko mice compared to wt mice (p < 0.05, Fig. 5A & B), but comparable MBP-immunostaining intensity between the two groups (p = ns, Fig. 5C & D). Western blot analysis revealed lower level of MBP protein in the same brain region relative to that of wt mice (p < 0.05, Fig. 5E & F). Biochemical analysis with the PFC tissue revealed a marginal significant difference (p = 0.07) in ATP level between the wt and COMT-ko mice (Fig. 5G), but no difference in ROS levels (p = ns, Fig. 5H). The changes in CPu of COMT-ko mice are similar to those in PFC, except that both ATP and ROS levels were significantly lower in COMT-ko mice compared to the wt mice (Supplemental Fig. 3).
We also assessed behavioral performances of the mice in the second animal experiment. Compared to wt mice, the COMT-ko mice showed behavioral anomalies indicated by a higher level of locomotor activity and a higher CD/TD in OFT (Supplemental Fig. 4A & B), longer duration on open arms and the central zone, but shorter duration in the closed arms of the EPM (Supplemental Fig. 4C) while visiting both the closed arms and central zone more frequently (Supplemental Fig. 4D). In the SIT, COMT-ko mice were unable to tell an empty cage from an identical cage with a novel conspecifics (Supplemental Fig. 4E & F).
DA inhibits the development of cultured OLs and induces OLs apoptosis via inhibiting mitochondrial functions of the cells
All data from the above animal experiments strongly suggest a neurobiological mechanism in which DA elevation impairs mitochondrial function in brain cells thus inhibiting OL development/myelination process. To substantiate this neurobiological mechanism, we did in vitro experiments in which purified oligodendrocyte precursor cells (OPCs) were cultured in the absence or presence of DA at the concentrations of 50, 100, or 200 µM starting on DIV (day in vitro) 12 and continuing for 48 hrs. Dual immunofluorescent staining with the primary antibodies to Olig 2 (O2) and O4, O2 and CNP (2',3'-cyclic nucleotide phosphodiesterase), or CNP and MBP, was done to label immature and mature OLs while nuclei of cells were stained with DAPI dye. As shown in Fig. 6A, all differentiated OLs at specific developmental stages appear in distinctive morphology (size and shape) and labeling (color). Cell counting revealed: 1) no difference between VEH and DA groups in numbers of DAPI+ cell nuclei and of cells labeled by the antibody to O2 (Fig. 6B), which is a sustained marker of OLs and expressed in all stages of OL development, from OPC to mature OL; 2) the ratio O4+/O2+ cells (expressed as percentage, the same below) decreased in DA groups in a concentration-dependent manner (p < 0.001; Fig. 6C), indicating that DA inhibited the differentiation of OPC into immature OLs; 3) the ratios CNP+/O2+ (p < 0.001) cells and MBP+ cells/DAPI+ nuclei (p < 0.001) decreased in DA groups, but the ratio MBP+/CNP+ cells did not change across VEH and all DA groups (p = ns, Fig. 6C). These data demonstrate that DA elevation retards the maturation of O4+ cells into CNP+ and MBP+ cells but has no impact on the further maturation from CNP+ cells to MBP+ cells.
Furthermore, we analyzed effects of DA on mitochondrial functions of the cultured OLs. The mitochondrial membrane potential (ΔψM) of cultured OLs from the VEH and DA groups was assessed using an assay kit with JC-1. Compared to VEH group, JC-1 aggregates decreased in DA groups in a concentration-dependent manner indicated by gradual decreases in red signal which almost completely disappeared in the cells treated with the highest concentration of DA at 200 µM. In contrast, JC-1 monomers increased in DA-treated cells relative to VEH group indicated by green signal increase which was strongest in cells treated with DA at 100 µM (Fig. 6D). Therefore, ΔψM (FL590/FL530) values decreased in a DA concentration-dependent manner (p < 0.001, Fig. 6E). These data demonstrate the disruption of mitochondrial membrane potential in OLs of DA groups. In contrast, DA increased intracellular level of reactive oxygen species (ROS) in cultured OLs in a concentration-dependent manner (p < 0.001, Fig. 6F & G).
The foregoing ROS and ΔψM data strongly suggest a cytotoxic effect (lethal effect at the highest concentration) of DA on cultured OLs. To verify this suggestion, the immunofluorescent staining with the antibody to cleaved Caspase-3 was done to label apoptotic OLs while the antibody to O2 was used to label nuclei of all OLs in the cultures without or with DA at the indicated concentrations. As shown in Fig. 6H & I, Caspase-3+ cells increased in DA-treated cultures in a concentration-dependent manner (p < 0.001) and the O2+ nuclei in cells treated with DA at 100 and 200 µM look much smaller than those in the VEH group, indicating the nuclear pyknosis of these damaged OLs and confirming the apoptotic OLs induced by the higher concentrations of DA.
NAC and TCP ameliorate the adverse effects of DA on cultured OLs
The above in vitro data strongly indicate that the inhibitory effects of DA on cultured OLs development and axonal myelination in neuron-OL co-cultures are achieved via impairing mitochondrial function. If so, mitochondrial protection approaches should be able to ameliorate the inhibitory effects of DA. To substantiate this possibility, we did another two in vitro experiments in which NAC (N-acetyl-L-cysteine, a well-established antioxidant) or TCP (trans-2-phenylcyclopropy, an inhibitor of mitochondrial MAOs) was used, respectively. In the TCP experiment, primary OLs were cultured at the indicated concentrations of DA (0, 100 µM, 200 µM) in the absence or presence of TCP (100 µM) during DIV 15–17. At the end, the mature OLs were analyzed for cell viability, ΔψM, and the production of mitochondrial ROS and ATP, in addition to Western blot analysis measuring CNP and MBP levels. As shown in Fig. 8A & B, both 100 µM and 200 µM DA significantly decreased CNP and MBP levels in cultured OLs as compared to CNT group, but these effects were prevented or ameliorated in the presence of 100 µM TCP. 200 µM DA significantly decreased cell viability of cultured OLs as compared to CNT group (p < 0.01), but this effect was effectively ameliorated in the presence of 100 µM TCP (p < 0.05, Fig. 8C). Moreover, 200 µM DA significantly decreased mitochondrial ΔψM compared to CNT group (p < 0.001) and this damaging effect was effectively ameliorated by TCP (p < 0.01, Fig. 8D). Relevantly, DA significantly decreased the production of ATP at both 100 µM (p < 0.01) and 200 µM (p < 0.001) as compared to CNT group, but these decreases were effectively ameliorated by TCP (p < 0.05 and p < 0.01 in the two cases, respectively; Fig. 8E). In contrast, addition of 200 µM DA significantly increased ROS production in cultured OLs relative to CNT group (p < 0.01), and this increase was not seen in the presence of TCP (p < 0.05, Fig. 8F). TCP alone did not impact any of the above measurements. These data indicate that the protection of TCP against the toxic effects of DA on OLs is achieved by inhibiting the catabolism of DA and consequently decreasing ROS production.
In the NAC experiment, 250 or 500 µM NAC (the two concentrations were referred to as NAC-l and NAC-h, respectively) was provided to the primary OLs 4 h before addition of 50 µM DA. Forty-eight hours later, the cell viability of cultured OLs was assessed using the CCK-8 assay kit, in addition to dual immunofluorescent staining with the primary antibodies to CNP and MBP. The other OLs were treated with 50 µM DA or 500 µM NAC alone, and the cells in Control group were treated with the vehicle in the absence of DA and NAC. As shown in Supplemental Fig. 5A, CNP+ and MBP+ cells appear in all groups at various proportions. CCK-8 assay showed that NAC alone had no effect on cell viability of cultured OLs (p = ns), but 50 µM DA significantly decreased cell viability, compared to the Control group (p < 0.001). Importantly, NAC pre-treatment at both 250 and 500 µM concentrations completely prevented the DA-induced cell viability decrease (Supplemental Fig. 5B). Furthermore, DA alone (50 µM) significantly decreased numbers of CNP+ and MBP+ cells compared to VEH group (p < 0.01, p < 0.001, respectively), but these inhibiting effects were significantly ameliorated by NAC-l (p < 0.01) and NAC-h (p < 0.001) (Supplemental Fig. 5C). These data add further evidence for the antioxidative action of NAC which is effective in protecting OLs against the toxicity of DA.