Fluoxetine increases brain MeCP2 immuno-positive cells in a female Mecp2 heterozygous mouse model of Rett syndrome through endogenous serotonin

Motor skill deficit is a common and invalidating symptom of Rett syndrome (RTT), a rare disease almost exclusively affecting girls during the first/second year of life. Loss-of-function mutations of the methyl-CpG-binding protein2 (MECP2; Mecp2 in rodents) gene is the cause in most patients. We recently found that fluoxetine, a selective serotonin (5-HT) reuptake inhibitor and antidepressant drug, fully rescued motor coordination deficits in Mecp2 heterozygous (Mecp2 HET) mice acting through brain 5-HT. Here, we asked whether fluoxetine could increase MeCP2 expression in the brain of Mecp2 HET mice, under the same schedule of treatment improving motor coordination. Fluoxetine increased the number of MeCP2 immuno-positive (MeCP2+) cells in the prefrontal cortex, M1 and M2 motor cortices, and in dorsal, ventral and lateral striatum. Fluoxetine had no effect in the CA3 region of the hippocampus or in any of the brain regions of WT mice. Inhibition of 5-HT synthesis abolished the fluoxetine-induced rise of MeCP2+ cells. These findings suggest that boosting 5-HT transmission is sufficient to enhance the expression of MeCP2 in several brain regions of Mecp2 HET mice. Fluoxetine-induced rise of MeCP2 could potentially rescue motor coordination and other deficits of RTT.

and enhanced MeCP2 levels, assessed by western blotting, in the hippocampus of Ts65Dn mice, a model of Down syndrome 14 . Therefore, it is not clear whether the motor effect of FLX is mediated solely by its ability to enhance 5-HT, to boost the expression of MeCP2 or both. It is not known whether FLX enhances the expression of MeCP2 in experimental models of RTT. The present study therefore aims to provide evidence that FLX can enhance brain expression of MeCP2 in Mecp2 HET mice and that 5-HT is involved.

Results
Fluoxetine increases the number of MeCP2 + cells in various brain regions of Mecp2 HET mice except the CA3. The drug's effect on the expression of MeCP2, assessed by immunohistochemistry (IHC) in WT and Mecp2 HET mice is shown in Figs. 1 and 2. In Mecp2 HET mice there were significantly fewer MeCP2 + cells than in WT mice in all brain regions examined, reaching 46% of WT levels in the prefrontal cortex (PFC), 51% in the M1, 47% in M2 and 42% in the CA3 (Fig. 1). Two weeks of treatment with FLX partially compensated the reduced number of MeCP2 + cells. In the PFC, FLX restored MeCP2 + cells to 70% of those in WT mice treated with FLX (WT + FLX). Restoration of MeCP2 + cells reached 72% in the M1 and 64% in the M2. FLX had no effect in the CA3 (Fig. 1).
In striatal sub-regions of Mecp2 HET mice, MeCP2 + cells were reduced to 37-38% of WT values (Fig. 2). FLX partially restored the number of MeCP2 + cells reaching 59-62% of those in WT mice treated with FLX. Two-way ANOVA showed significant effects of genotype and FLX but not their interaction for all brain regions examined except CA3 where FLX had no effect (Table 1). FLX did not significantly affect MeCP2 expression in any brain regions of the WT mice (Figs. 1, 2).
Densitometric analysis showed that FLX increased MeCP2 staining density in the PFC of Mecp2 HET mice reaching 69% of that in WT-FLX mice. In the M1 of FLX-treated Mecp2 HET mice MeCP2 density reached 67% of that in WT-FLX mice ( Supplementary Fig. S6). FLX increases staining density and MeCP2 + cell number to a similar extent, in both brain regions.
Correlation between rotarod performance and number of MeCP2 + cells in various brain regions. Using the Pearson coefficient statistics to correlate the number of MeCP2 + cells and latency to fall On day 15 mice were euthanised and their brain processed for MeCP2 immunohistochemistry. Scale bars, 100 µm. (C) Histograms, generated with the Prism software (version 7.02; URL link), show the mean number (± SEM) of MeCP2 + cells in the PFC, M1 and M2 and CA3 of WT and HET mice (6-14 mice/group). Black dots and squares indicate individual data points. Blue and red columns indicate treatment with vehicle and FLX, respectively. ***P < 0.001 versus WT-H2O; #P < 0.01 versus HET-H2O; §P < 0.002, § §P < 0.0001 versus WT-FLX (Sidak's test).

Discussion
Repeated treatment with FLX, which can rescue the motor coordination deficit in Mecp2 HET mice 20 , increased the number of cells expressing the MeCP2 protein in the PFC, M1 and M2 motor cortices, DSTR, VSTR and LSTR, but not in the CA3 region of the hippocampus. The 5-HT synthesis inhibitor pCPA prevented the effect of FLX in all brain regions examined, demonstrating that the FLX-induced increase of MeCP2 + cell numbers depends on endogenous 5-HT. In line with this, 5-HT enhanced the expression of Mecp2 in cultured PC12 cells 23 , which are capable of taking up 5-HT from extracellular medium and express 5-HT receptors on the cell membrane 24,25 . MeCP2 protein is present in most brain neurons and regional differences in its abundance have been related to specific behavioral impairments [26][27][28] . We found a significant correlation between rotarod performance and the number of MeCP2 + cells in PFC but not in the other brain regions, suggesting that increased MeCP2 in the PFC may underlie FLX's ability to improve rotarod performance in Mecp2 HET mice. In line with this proposal, the role of the PFC in motor coordination is supported by a previous study in C57BL/6 mice showing that the volume of frontal association cortex, an indicator of structural plasticity, was positively correlated with better rotarod performance 29 .
Although we found no significant correlation between the effect of FLX on rotarod performance and the increase of MeCP2 + cells in the other brain regions analyzed, we cannot exclude the contribution of, at least, the striatum and motor cortex. This stems from the observation that the selective deletion of Mecp2 from the striatum replicates the rotarod impairment caused by global Mecp2 deletion, while the selective preservation of striatal Mecp2 expression in otherwise Mecp2 null mice was sufficient to prevent the impairment of rotarod performance 30,31 . In addition, motor cortex is critical for the acquisition, consolidation and retention of motor skill, including rotarod skill 32 .
Previous studies in conditional knockout mice have identified cell types in which the deletion or restoration of the Mecp2 gene affected rotarod performance. Selective expression of Mecp2 in GABAergic or glutamatergic neurons of Mecp2-null mice was sufficient to rescue the rotarod deficit 33,34 . Restoration of Mecp2 expression in GABAergic neurons rescued the rotarod deficit preferentially in male mutants 34 , while the expression of the gene in glutamatergic neurons was sufficient to rescue the rotarod deficit in female Mecp2 HET but had no effect in Mecp2 null mice 33 . As we found that FLX was more effective in restoring rotarod performance in Mecp2 HET Table 1. Two-way ANOVA comparing the effect of FLX on regional MeCP2 + cells in Mecp2 HET and WT mice shown in Figs. 1 and 2. Significant effects are highlighted in bold type.   Table S1). Thus, further investigation is now warranted to examine the requirement for MeCP2 in different brain regions and cell populations in FLX's effect on motor coordination in Mecp2 mutant mice.
FLX increased Mecp2 gene and protein expression in the striatum and PFC of Wistar rats 21 , whereas it had no effect on MeCP2 + cells in any brain region of WT mice. As the FLX dose and duration of treatment were similar in our and the earlier study in the rat, it is likely that species differences influence the ability of FLX to affect MeCP2 expression.
We found that FLX had no effect on the number of MeCP2 + cells in the CA3 region of the hippocampus of Mecp2 HET mice where the expression of the 5-HT transporter (SERT) is selectively reduced 35,36 . Reduced production of the SERT protein impairs the response to SSRIs 37 and may account for the failure of FLX to enhance MeCP2 expression in the CA3.
The mechanism by which FLX raises the number of MeCP2 + cells in Mecp2 HET mice was not addressed in the present study. FLX can only increase MeCP2 in cells that already express the wild type allele of Mecp2 or might reactivate the Mecp2 gene in cells in which the wild type allele is silenced because of the X chromosome inactivation (XCI), a physiological phenomenon by which one of the two X chromosomes in female cells becomes transcriptionally silent to compensate differences in gene dosage between sexes 38 . The first mechanism is unlikely as FLX had no effect on MeCP2 + cells in WT mice (present study). The potential effect of FLX on XCI is supported by indirect evidences showing that FLX enhances the levels of the transforming growth factor-β1 (TGF-β1) in cultured astrocytes 39 . TGF-β1 suppresses the expression of the X-inactivating specific transcript (Xist), a long non-coding RNA which plays a key role in XCI 6,38 . Thus, FLX by increasing TGF-β1 may suppress Xist and in turn enhance the expression of Mecp2.
Previous reports that the stimulation of 5-HT neurotransmission with SSRIs and 5-HT receptor agonists ameliorates symptoms in Mecp2 mutant mice [26][27][28][29][30] confirm the importance of the 5-HT system as a potential therapeutic target for RTT 40 . The SSRI citalopram improves respiratory sensitivity to CO 2 in Mecp2 tm1.1Bird mice 41 . The 5-HT 1A receptor agonists 8-OH-DPAT, F15599 and sarizotan-which is also a dopamine D2 receptor partial agonist, reduce apnea and improve breathing regularity [42][43][44] . However, sarizotan failed to reduce apnea in patients with RTT and did not achieve the trial's secondary end-point, which included tests of motor function  www.nature.com/scientificreports/ (press release); this raises questions about the validity of 5-HT 1A receptor stimulation as a therapeutic target for RTT (see Abdala et al. 45 for a review of the role of 5-HT 1A receptors in RTT). LP-211, a 5-HT 7 receptor agonist, improves cognitive functions and reduces anxiety in Mecp2-308 mice 46 . Thus, enhanced 5-HT transmission resulting from blockade of the 5-HT transporter or selective stimulation of 5-HT receptor subtypes is a common feature of serotonergic drugs that improve RTT deficits in mouse models. Mirtazapine, a noradrenergic and serotonergic antidepressant, rescues cortical atrophy in Mecp2 null mice 47 . Although mirtazapine blocks 5-HT 2/3 receptors, the fact that it blocks α 2 -adrenoceptors results in an increased firing rate of 5-HT neurons of the raphe, 5-HT release and tonic activation of hippocampal 5-HT 1A receptors 48 . Currently, information on the effects of FLX and other SSRIs in patients with RTT is limited to a few case reports showing that these drugs may be effective on psychiatric, behavioral and motor symptoms and sleep apneas [49][50][51] . An open study on the effects of FLX in RTT was terminated prematurely (EudraCT Number: 2008-000787-16) and to our knowledge no other clinical assessments of the effects of SSRIs in RTT have been published. Thus, further clinical studies are needed to draw any firm conclusion about the efficacy of these drugs in patients with RTT.
As the FLX-induced improvement of motor coordination and increased number of MeCP2 + cells depend on endogenous 5-HT, the present data suggest that boosting 5-HT neurotransmission may be sufficient to drive both effects. 5-HT levels and TPH2 expression are reduced in the brain of patients with RTT 52 . Thus, FLX may possibly have limited effect on motor coordination in patients. However, stimulation of 5-HT synthesis restored the antidepressant-like effect of SSRIs in mice that did not respond to the SSRI alone because of genetic polymorphism of Tph2 reducing the biosynthesis of 5-HT or following treatment with pCPA [53][54][55] . These findings suggest that boosting 5-HT synthesis may improve the efficacy of SSRIs on motor coordination and increase MeCP2 expression also in subjects with impaired 5-HT synthesis.
The identification of the molecular and cellular mechanisms by which FLX boosts the expression of brain MeCP2 has the potential to provide novel therapeutic targets to counteract RTT deficits.

Materials and methods
Experimental design and statistical analysis. Breeding pairs consisting of female Mecp2 HET (B6.129P2(C)-Mecp2 tm1.1Bird /J; stock: 003890) and male wild type (C57BL/6 J; stock: 000664) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Details of mice breeding and husbandry are described in Villani et al. 20 . The first test cohort comprised 34 Mecp2 HET and 26 WT littermate controls. They were used to www.nature.com/scientificreports/ assess the ability of FLX to enhance the expression of MeCP2 protein in the brain of Mecp2 HET mice (primary end point of the study). Mice of this cohort were the same as those used to study the motor effects of FLX by Villani et al. 20 and they were 9 weeks old at the start of FLX treatment. At the end of the behavioral assessment they were euthanised, their brain were removed and processed according to the immunohistochemical protocol described below. A second cohort of 21 Mecp2 HET mice, 9 weeks old at the start of treatment, was used to assess the effect of pCPA on FLX-induced rise of MeCP2 expression. All statistical analyses were performed by GraphPad Prism version 7.02 for Windows (GraphPad Software, CA) and all data were expressed as mean ± SEM of the number of MeCP2 immunopositive (MeCP2 + ) cells. Two-way ANOVA, followed by Sidak's test was used to assess the effect of FLX on MeCP2 protein expression in WT and Mecp2 HET mice and the effect of pCPA on FLX-induced rise in the number of MeCP2 + cells in Mecp2 HET mice.
Gaussian distribution was formally tested for the effect of FLX on MeCP2 expression, the primary end-point of the study, with the D' Agostino-Pearson normality test included in the GraphPad Prism 7.02 package 56 . This analysis confirmed the normal distribution of the MeCP2 + cell in all experimental groups of the PFC, M1, M2, DSTR, VSTR, LSTR. Because of the small number of subjects, no normality test was applied to CA3 data. The correlation between the number of MeCP2 + cells and the latency to fall from the rotarod, measured in the same mice 20

Western blotting.
To examine the effect of FLX on MeCP2 levels, brain areas of Mecp2 HET and WT mice were dissected out free-hand. Striatum, hippocampus and cortex were homogenized in 1% Sodium dodecyl sulphate (SDS) and total protein concentration was determined by the bicinchoninic acid protein assay (BCA Protein Assay Kit, Pierce Biotechnology, USA). Twenty µg protein/sample were loaded onto 10% acrylamide gels. The amount of protein was established by running increasing amount of protein on the gel ( Supplementary  Fig. 4). For the MW estimates of target proteins, Precision Plus Protein Dual Color Standards (BIO-RAD, Cat. #1610374) was loaded in one well of each gel. Proteins were separated by SDS polyacrylamide gel electrophoresis (SDS-PAGE) and electrophoretically transferred to a polyvinylidene difluoride (PVDF) membrane (Amersham, GE Healthcare). Membranes were cut approximately in correspondence to the 50 kDa MW. Blots were blocked with 5% non-fat milk in tris-buffered saline (TBST) (20 mM Tris, pH 7.6, 140 mM NaCl, and 0.5% Tween-20) for 1 h at room temperature. Membranes containing proteins heavier than 50 kDa were incubated overnight at 4 °C with primary antibody against MeCP2 (MW = 75 kDa) diluted 1:2000 (Cell Signaling Technology, Cat. # 3456), which detects both isoforms 1 and 2 of the protein (datasheet). Membranes containing proteins lighter than 50 kDa were incubated with the antibody against glyceraldehyde 3-phosphate dehydrogenase (GAPDH; MW = 37 kDa) (Abcam, Cat. # Ab22555) diluted 1:20,000, as a loading control. After washing with TBST (5 × 5 min), anti-rabbit IgG, horseradish peroxidase (HRP)-linked antibody (Cell Signaling Technology, Cat. # 7074) were applied for 1 h at room temperature (1:5000 and 1:30,000 for MeCP2 and GAPDH, respectively). Primary antibodies were diluted in bovine serum albumin (BSA) 2% in TBST while the secondary antibody was diluted in 5% non-fat milk. The immune-positive protein bands were detected with a chemiluminescent home-made enhanced chemiluminescence (ECL) luminol/p-coumaric acid solution. Membranes were exposed to autoradiography films in the dark room (Hyperfilm ECL, Amersham GE Healthcare) then developed. Autoradiographs were scanned and signal intensity assessed with ImageJ (NIH) software. The ratio between MeCP2 and GAPDH for each sample was normalized to the WT mice average. Tissue fixation and sectioning. Twenty-four h after the last dose of FLX or vehicle, Mecp2 HET and WT mice, were anaesthetized with 75 mg/kg ketamine plus 1 mg/kg medetomidine and perfused through the heart using a peristaltic pump (Mityflex, ANKO, Bradenton, FL) with Ice-cold 0.01 M phosphate buffer-saline (PBS Tablets; Sigma, Cat. # 5564) at 10 mL/min for 3 min followed by 4% para-formaldehyde (PAF) solution in PBS for 5 min. The brain was removed and post-fixed in 4% PAF for 4 h, at 4 °C. Brains were then dipped in a 20% sucrose solution and stored at 4 °C overnight. Brains were removed from the sucrose solution, damped with Kleenex, frozen in n-pentane at − 45 °C for 3 min and stored at − 80 °C until sectioning.
Brain sections (25 μm) for IHC were cut with a cryostat (CM1850-UV, Leica Microsystems, Buccinasco, Italy) at − 17 °C and mounted on gelatin-coated slides. Consecutive brain slices containing the areas of interest were mounted onto glass slides for 3,3′-diaminobenzidine tetrahydrochloride (DAB) staining. After incubation with the primary antibody, slices were rinsed in PBS and incubated with biotinylated horse anti-rabbit IgG for 1 h. Sections were then incubated with an avidin-biotin-peroxidase complex (Vectastain ABC Elite Kit) for 1 h, rinsed in PBS and reacted for about 12 min with the chromogen DAB and H 2 O 2 . After 3 washes in PBS, sections were coverslipped with DPX New Mounting Medium (Merck, Cat. # 100579). To avoid procedural bias, comparative brain slices of all genotype and treatment groups were processed simultaneously.

Immunohistochemistry.
Staining was observed under a bright field microscope, the Virtual Slide Microscope (VS120) with a 2 × lens for the overview and 20 × lens for the acquisition (Olympus, Italy). MeCP2 + cells were counted in pre-defined areas of the dorsal (DSTR), ventral (VSTR) and lateral (LSTR) striatum, M1 and M2 motor cortices, prefrontal cortex (PFC) and CA3 sub-region of the dorsal hippocampus. Before the analysis, each image has been optimized by improving contrast/brightness and the background has been subtracted to better highlight the labeled protein Cell count biases were avoided by using automatic counting software (Image J) and custom-made macros to apply the same pre-defined functions: "Convert to Mask", "Fill Holes", "Watershed", "Analyze Particles", "size = 100 − Infinity and circularity = 0.00-1.00". The automatic thresholding function of the software was used to distinguish the specific signal generated by MeCP2 from the background.
For each brain region, the MeCP2 + cells count was the mean of 2-5 adjacent slices/mouse. Specificity of the MeCP2 antibody is confirmed by the absence of specific staining and very low background signal in slices obtained from the brain of Mecp2-null mice ( Supplementary Fig. 3).
A densitometric analysis of MeCP2 + cells in the same pre-defined area of the PFC, used to count the number of MeCP2 + cells, was done by the Image J software as described in detail elsewhere 57 . Briefly, images were converted to black and white and analysed by setting "Mean gray value" in the "Set measurement" window. Density was expressed as the average pixel intensity of the digitalized histochemical image.
Drug treatment. Fluoxetine HCl (Casen, Recordati SL, Spain) was dissolved in pyrogen-free water (10 mL/ kg). WT and Mecp2 HET mice were injected intraperitoneally (i.p.) with 10 mg/kg FLX (as salt) once daily for 14 days and 24 h after the last dose of the chronic schedule they were anaesthetized and perfused intracardially with PAF solution as described above.
To assess whether chronic i.p. administration of FLX could have adverse effects on animal's health, we monitored the weight gain during treatment, as this is a reliable indicator of their health. The Mecp2 HET and WT mice, do not differ in weight and FLX did not affect the weight growth curve in either genotype 20 .
To assess the role of 5-HT in the ability of FLX to rescue motor deficits of Mecp2 HET mice, mice were given 100 mg/kg pCPA ethyl ester hydrochloride as free base (Sigma-Aldrich, Milano, Italy) dissolved in pyrogen-free water. pCPA or water was given orally by gavage for three consecutive days simultaneously with FLX or vehicle i.p. (from day 15 to 17 from the first dose of FLX). The dose of pCPA was based on the results of previous studies showing full antagonism of the effect of SSRIs in the forced swimming test 53,54 .