Cytokine Dysregulation in MECP2- and CDKL5-Related Rett Syndrome: Relationships with Aberrant Redox Homeostasis, Inflammation, and ω-3 PUFAs

An involvement of the immune system has been suggested in Rett syndrome (RTT), a devastating neurodevelopmental disorder related to oxidative stress, and caused by a mutation in the methyl-CpG binding protein 2 gene (MECP2) or, more rarely, cyclin-dependent kinase-like 5 (CDKL5). To date, it is unclear whether both mutations may have an impact on the circulating cytokine patterns. In the present study, cytokines involved in the Th1-, Th2-, and T regulatory (T-reg) response, as well as chemokines, were investigated in MECP2- (MECP2-RTT) (n = 16) and CDKL5-Rett syndrome (CDKL5-RTT) (n = 8), before and after ω-3 polyunsaturated fatty acids (PUFAs) supplementation. A major cytokine dysregulation was evidenced in untreated RTT patients. In MECP2-RTT, a Th2-shifted balance was evidenced, whereas in CDKL5-RTT both Th1- and Th2-related cytokines (except for IL-4) were upregulated. In MECP2-RTT, decreased levels of IL-22 were observed, whereas increased IL-22 and T-reg cytokine levels were evidenced in CDKL5-RTT. Chemokines were unchanged. The cytokine dysregulation was proportional to clinical severity, inflammatory status, and redox imbalance. Omega-3 PUFAs partially counterbalanced cytokine changes, as well as aberrant redox homeostasis and the inflammatory status. RTT is associated with a subclinical immune dysregulation as the likely consequence of a defective inflammation regulatory signaling system.

MECP2 mutations are known to cause up to 90-95% of cases with typical clinical presentation (MECP2-RTT) which include early neurological regression followed by loss of acquired cognitive, social, and motor skills in a typical 4-stage neurological regression, together with development of autistic behavior [4]. The typical disease, manifesting above 6-18 months almost exclusively in girls, represents the 2 Oxidative Medicine and Cellular Longevity second most common cause of severe intellective disorder in the female gender [1] and its frequency is approximatively 1 : 10,000 with a random worldwide distribution [5]. MECP2 mutations have also been reported in a milder syndrome phenotype (i.e., preserved speech variant), as well as other non-RTT disorders, such as Angelman syndrome, mild learning disability in females, neonatal encephalopathy (largely predominant in males), and X-linked intellectual disability [6].
Although MeCP2 has been extensively studied in the central nervous system [7], MeCP2 is a ubiquitous nuclear protein [8]. However, the molecular mechanisms by which MeCP2 deficiency drives pathology in RTT remain elusive. MeCP2 is currently considered a multifunctional regulator of gene transcription, involved in RNA splicing and chromatin remodelling [7]. A close biochemical and genetic relationship exists between MeCP2 and CDKL5 in that the encoded CDKL5 protein is known to be a kinase able to mediate MeCP2 phosphorylation [3] and that CDKL5 has been reported to be a MeCP2-repressed target gene [9]. CDKL5-related RTT (CDKL5-RTT) is known as the earlyonset seizures variant (ESV) and is associated with severe early-onset seizures and partially different clinical RTT-like features [10], so that recently the CDKL5 disorder has been proposed as an independent clinical entity rather than an additional RTT variant [11].
Our prior studies indicate that (1) MeCP2 is involvedeither directly or indirectly-in the regulation of the redox balance in RTT [12][13][14][15][16][17][18]; (2) a direct evidence of the link between MeCP2 deficiency, oxidative stress (OS), and RTT pathology exists in murine models of the disease [18]; (3) a redox imbalance has also been evidenced in CDKL5-RTT [16]. Proofs for a direct relationship between the MeCP2 protein and OS are lacking to date. However, we have previously demonstrated that OS in brains of mouse with Mecp2 loss-of-function mutations precedes and accompanies symptoms of the disease and is rescued by selective Mecp2 gene reexpression [18]. Whether mitochondria are the main, or the only, major intracellular source for abnormal redox status in RTT [19][20][21][22][23] or other "MECP2-pathies" is still to be ascertained, given that no major morphological alterations in mitochondria have been evidenced by our group at the transmission electron microscopy of primary cultures of skin fibroblasts from MECP2-RTT, whereas major changes were observed in the endoplasmic reticulum, along with cytoplasmic multilamellar bodies [17].
MECP2-RTT appears to be intimately related to inflammation [24,25]. There is prior evidence that MeCP2 associates with CpG elements of the regulatory regions of Foxp3 gene, a known transcription factor needed for the generation of T regulatory (T-reg) cells [26]. This observation supports the concept that MeCP2 could play a regulatory role in T cells' resilience to inflammation [27]. Furthermore, a very recent report demonstrates that MeCP2 acts as a regulator of the response to inflammatory stimuli in the microglia and macrophages of Mecp2-null mice [28].
Preliminary findings by Ariani suggest that a reduced frequency of HLA-B39 is present in MECP2-RTT [39], whose reasons and biological consequences need to be clarified. Some disrupting mutations in chemokines and chemokines receptors have been reported by Grillo et al. in the preserved speech variant of RTT, thus suggesting this combination of mutations along with the MECP2 loss-of-function may modulate the immune system [40].
An association between MECP2 gene polymorphisms and autoimmune disease, such as systemic lupus erythematosus (LES) [41,42] and primary Sjögren's syndrome [43], has been reported. High levels of serum anti-N (Glc) autoantibodies of the IgM fraction have been recently reported in RTT [44]. A direct relationship between MeCP2 and the immune system has been recently evidenced, since MeCP2 has been found to be indispensable for the differentiation of naïve CD4+ T cells into Th17 cells and for the commitment of naïve CD4+ T cells to the Th1 lineage [30]. Moreover, MeCP2 has been shown to play a critical role in promoting multiple cytokine-mediated signaling pathway through a MeCP2-miR-124-suppressor of cytokine signaling 5 (SOCS5) axis known to be indispensable for the activation of signal transducer and activator of transcription 3 (STAT3) and STAT1 in CD4+ T cells [30]. This signaling pathway appears to be indispensable for the activation of signal transducer and activator of transcription (i.e., STAT3) in naïve CD4+ T cells, with consequent generation of Th17 cells [30].
In the present study, we evaluated systemic cytokine patterns, oxidative stress markers, and inflammatory status in both typical MECP2-RTT and CDKL5-RTT as a function of a 12-month supplementation with -3 polyunsaturated fatty acids (PUFAs).  [45].

Methods
Clinical examinations and blood samplings in the MECP2-and CDKL5-RTT patients' groups were performed before and after -3 polyunsaturated fatty acids (PUFAs) supplementation during the routine follow-up study at hospital admission. Given the specific aims of this study, patients with clinically evident inflammatory conditions (i.e., upper respiratory tract infection, pneumonia, urinary infection, stomatitis, and periodontal inflammation), either acute or chronic, were excluded.
None of the patients at the time of enrollment were on anti-inflammatory drugs or undergoing supplementation with other known antioxidants. All the patients were consecutively admitted to the "Rett Syndrome National Reference Centre" (Head Professor J.H.) of the University Hospital of the Azienda Ospedaliera Universitaria Senese (AOUS). The subjects examined in this study were on a typical Mediterranean diet. Rett syndrome clinical severity was assessed using either the total clinical severity score (CSS), a validated clinical rating specifically designed for RTT, or the "compressed" CSS, both based on 13 individual ordinal categories measuring clinical features common in RTT [46]. Healthy female control subjects who have comparable age ( = 24; age: 13.4 ± 4.8 years, range: 2-32) were also included. Blood samplings from the control group were carried out during routine health checks, sports, blood donations, or periodic clinical checks.
The study was conducted with the approval of the Institutional Review Board of the AOUS. All informed consents were obtained from either the parents or the legal tutors of the enrolled patients or directly from healthy adults.

Inflammatory Marker.
We measured erythrocyte sedimentation rate (ESR) as nonspecific marker of inflammation. ESR was assayed on a widely used automated system (i.e., "TEST 1") measuring the aggregation capacity of red blood cells (RBCs) by an infrared ray microphotometer with a light wavelength of 950 nm [47].

Sample Collection.
All samplings from RTT patients and healthy controls were carried out around 8 AM after overnight fasting. Blood was collected in heparinized tubes and used for cytokine evaluations and oxidative stress (OS) markers determinations. In particular, blood samples were centrifuged at 2,400 g for 15 min at room temperature. The platelet poor plasma was saved, and the buffy coat was removed by aspiration. RBCs were washed twice with physiological solution, resuspended in Ringer solution (125 mM NaCl, 5 mM KCl, 1 mM MgSO 4 , 32 mM HEPES, 5 mM glucose, and 1 mM CaCl 2 ), pH 7.4 as a 50% (vol/vol) suspension, and then used for the determination of intraerythrocyte non-protein-bound iron (IE-NPBI). Plasma was used for cytokine, and OS marker (F 2 -isoprostanes: F 2 -IsoPs; F 2dihomo-isoprostanes: F 2 -dihomo-IsoPs; F 4 -neuroprostanes: F 4 -NeuroPs, and plasma non-protein-bound iron: P-NPBI) determinations. All manipulations were carried out within 2 h after sample collection. For all isoprostane determinations, butylated hydroxytoluene (90 M) was added to plasma as an antioxidant and stored under nitrogen at −70 ∘ C until analysis.
For serum preparation (IFN-and TGF-1 evaluations), blood samples were left to coagulate for 15-30 min at RT and centrifuged at RT for 10 minutes at 2000 g. Separated sera were kept in aliquots at −80 ∘ C until the time of assay.
Blood GSH and GSSG levels were determined by an enzymatic recycling procedure according to Tietze et al. [53] and Baker et al. [54].

Data
Analysis. All variables were tested for normal distribution (D' Agostino-Pearson test) and data were presented as mean and standard deviation unless otherwise specified. Statistical analysis for circulating cytokine levels in the groups at baseline and after -3 PUFAs supplementation was carried out using Student's -test and one-way ANOVA test. Bonferroni-corrected significance levels were used for multiple -tests. Associations between variables were tested by univariate linear regression analysis and multiple regression analyses. A two-sided < 0.05 was considered to indicate statistical significance. The MedCalc version 12.1.4 statistical software package (MedCalc Software, Mariakerke, Belgium) was used.
Overall, both Th1-and Th2-related (except for IL-4) cytokines were upregulated in CDKL5-RTT, together with increased levels of IL-22 and T-reg-related cytokines. No changes were observed for chemokine levels.

Relative Changes (Fold Variations).
In order to account for real changes, that is, independently of absolute basal concentrations, values data were submitted to double standardization (i.e., basal values of untreated MECP2-and CDKL5-RTT patients versus control values and treated MECP2-and CDKL5-RTT patients versus untreated). After standardization, some differences lost weight, whereas others gained statistical significance (Supplementary Tables 1  and 2 in the Supplementary Material available online at http://dx.doi.org/10.1155/2015/421624). In particular, changes in IL-10 in MECP2-RTT and IL-6, TGF-1, IP-10, IL-5/TNFratio, IL-6/TNF-ratio, and IL-10/TNF-ratio in CDKL5-RTT were found to be not significant after -3 PUFAs treatment, whereas changes in IL-5, IL-6/IFN-ratio, IL-4/TNF-ratio, and IL-6/TNF-ratio in MECP2-RTT and IFN-, IL-1 , IL-5, IL-37, and IL-4/IFN-ratio, F 4 -NeuroPs, and GSH/GSSG ratio in CDKL5-RTT were found to be statistically significant. activation of Th2 lineage is more evident as compared to that of Th17 line. The Th1 lineage appears to be inhibited with the single exception of TNF-, while no changes were observed regarding the T-reg lineage. Besides, chemokines appear to be inhibited with the single exception of IL-8. IL-33, known to be involved in Th2 differentiation, is increased. IL-37, known to counteract inflammatory status, appears to be increased and likely reflects a proinflammatory status in the untreated MECP2-RTT population (Figure 6(a)). In CDKL5-RTT, the Th1 and T-reg lineages appear to be strongly activated, whereas paradoxical effects were observed regarding Th17. An important feature of untreated CDKL5-RTT appears to be a strongly increased production of IL-10 dominating the Th2 pattern changes. Chemokines and IL-37 show the same pattern observed in MECP2-RTT ( Figure 6(b)).

Summary Effects of Circulating Basal Cytokine
Overall, several of the observed cytokine pattern changes in untreated RTT (either MECP2-or CDKL5-related RTT patients) appear to reflect a likely macrophage dysregulation/dysfunction as suggested by Cronk et al. [28]. In particular, in MECP2-RTT, a strongly increased production of TNF-was evidenced, whereas apart from TNF-a strongly increased release of IL-10 and IL-12p70 was observed in CDKL5-RTT.

Redox Status in MECP2-and CDKL5-RTT.
In typical MECP2-RTT an oxidative imbalance was evidenced, thus confirming our prior observations [12][13][14][15]. In particular, a significant increase in P-NPBI and IE-NPBI (i.e., free redox active iron), F 2 -IsoPs (i.e., markers of systemic OS), F 4 -NeuroPs (i.e., markers of brain grey matter oxidative injury), and F 2 -dihomo-IsoPs (i.e., markers of brain white matter oxidative injury) levels (all < 0.05) as compared to healthy subjects is observed. In CDKL5-RTT, similar results were found (Figure 7), confirming our preliminary observations [16]. However, some differences between MECP2-RTT and CDKL5-RTT were observed. While all the examined OS markers were found to be increased in MECP2-RTT, F 2dihomo-IsoPs plasma levels in CDKL5-RTT were found to be comparable to those of the control group. The observed redox imbalance was mirrored by a strongly reduced GSH/GSSG ratio as detectable in both conditions (Figure 7).

Inflammatory Status in MECP2-and CDKL5-RTT.
Elevated ESR values were detectable in MECP2-RTT as compared to healthy subjects (mean value 9.4 ± 5.8 mm/h), thus confirming the presence of a subclinical inflammatory status in MECP2-RTT patients [24]. A slight, albeit not statistically Data are expressed as means ± standard deviations. Asterisks denote significant post hoc pairwise tests ( < 0.05). The dashed line denotes upper range limit as obtained from manufacturer's data sheet or from published literature. For IL-4, normal range is referred to literature [55]. For IL-13, normal range is referred to literature [56]. For IL-33, normal range is not indicated in the manufacturer's data sheet. However, some authors reported also a range [57]. significant, increase in ESR was observed in CDKL5-RTT patients (Figure 8).

Omega-3 PUFAs Effects.
During the study, compliance to the administered -3 PUFAs was optimal, with no side effects requiring the drop-out from the protocol treatment.
Interestingly, -3 PUFAs induced a similar hypersecretion in IL-6 associated with a similar depression in TNF-and IL-8 circulating levels for both MECP2-and CDKL5-RTT.
In supplemented MECP2-RTT, all examined OS markers were significantly decreased ( < 0.05) and GSH/GSSG ratio increased ( < 0.05). Only IE-NPBI levels resulted as normalized, as compared to those of the control subjects ( Figure 7).

Summary Effects of -3 PUFAs Supplementation on Circulating Cytokine Changes in MECP2-and CDKL5-RTT.
In MECP2-RTT, a prolonged high-dosage strongly suppressed TNF-production and regulated the Th1 cytokine pattern and T-reg lineage as well as Th2 with the exception of IL-4. As it concerns chemokines, -3 PUFAs effect seems to be regulatory with a single exception of a persistently inhibited RANTES secretion. IL-37 appears to be inhibited as the likely net result of anti-inflammatory action (Figure 6(c)).
As compared to the MECP2-RTT picture, the effects of -3 PUFAs in CDKL5-RTT appear to be more difficult to be interpreted in a univocal way. For Th1 lineage, -3 PUFAs modulate the examined cytokines in a similar way to that observed in MECP2-RTT. Paradoxical responses within the Th2 and T-reg lineages were observed. As far as chemokines are concerned, -3 PUFAs partially counteract baseline changes, with the exceptions of IP-10 and RANTES. On the other hand, IL-37 levels appear to be unchanged (i.e., elevated levels) (Figure 6(d)).

Discussion
Our findings indicate, for the first time, a complex cytokine dysregulation in RTT. In particular, in MECP2-RTT, a Th2shifted balance was evidenced, whereas in CDKL5-RTT both Th1-and Th2-related (except for IL-4) cytokines were found to be upregulated.
Almost specular changes were observed regarding IL-22 and T-reg cytokine TGF-1 between these disorders, decreased IL-22 and unchanged T-reg levels in MECP2-RTT versus increased IL-22 and T-reg levels in CDKL5-RTT.
The observed cytokine dysregulation was proportional to clinical severity, inflammatory status, and redox imbalance. Interestingly, the macrophage has been very recently reported as a key target cell for MeCP2 action in a mouse model of Mecp2 loss-of-function [28]. Interestingly, several of the assayed cytokines, whose circulating levels were found to be abnormal in MECP2-RTT girls, included macrophagerelated cytokines, such as TNF-, IL-6, IL-12p70, IL-10, TGF-1, IL-8, and RANTES [60].
Although the characteristic cytokine pattern in RTT appears to be inhomogeneous, it includes high plasma levels of TNF-and IL-5. TNF-is a proinflammatory cytokine and produced in response to inflammatory stimuli. Elevated levels of TNF-have been reported in rheumatoid arthritis, ankylosing spondylitis, irritable bowel disease, and psoriasis [61]. IL-5, a Th2 cytokine, acts on eosinophils, basophils, mast cells, and B cells and elevated levels were found in allergy asthma and hypereosinophilic syndrome [62].
Actually, none of the clinical features observed in these associated pathologies are usually present in RTT, either MECP2-or CDKL5-related disorder.
Our data confirm the presence of a significant proinflammatory status in MECP2-RTT. For the first time, we observed an upregulation of both Th1-and Th2-related cytokines in CDKL5-RTT, which does not translate into increased inflammatory marker levels, likely due to a bulk increase in the anti-inflammatory IL-10 (about 21-fold), a key regulator cytokine of immune response [50].
The clinical translation for the observed changes remains to be elucidated. Nevertheless, cytokine dysregulation in MECP2-RTT appears to be associated with a proinflammatory status, as evidenced by raised ESR values and an acute phase protein response [24]. On the other hand, the final result in CDKL5-RTT remains unclear, given that ESR was found to be slightly higher than control values, although not statistically different. We have previously described in MECP2-RTT patients the presence of respiratory bronchiolitis-associated interstitial lung disease-(RB-ILD-) like features at the high-resolution CT (HRCT) images of the lungs in up to half of patients aged 10 years or more [63]. More recently, histologic evidence for an inflammatory lung disease has been reported by our group in a murine model of the disease [25]. Therefore, it is conceivable that the unexplained persistent inflammatory process in MECP2-RTT could be the possible result of an ineffective chemotaxis, at least in some anatomical districts. No information to this regard is available to date for CDKL5-RTT.
Prior reports have suggested the occurrence of defective Th1 differentiation in RTT [30,34]. Our findings in MECP2-RTT appear to be in line with this suggestion, given that levels of cytokines known to be produced or related to Th1 are generally reduced, with the only notable exception of TNF- (Figure 9) [61,64]. On the contrary, Th1 cytokines were found to be all increased in CDKL5-RTT. To date, no information concerning immunological response, inflammatory status, or defense against infections is available for this condition.
On the other hand, Th2-cytokines were found to be generally increased in RTT. In particular, a significant increase in IL-5, IL-6 was observed in MECP2-RTT whereas in CDKL5-RTT IL-5, IL-10, and IL-13 levels were increased. Thus the Th2-dependent response appears to be upregulated in RTT independently of the mutated gene. This observation is also in line with the hypothesis of a compensatory effect of immune response related to a defective Th1 differentiation [30,34]. Possible clinical consequences of this Th2/Th1 shift imbalance in RTT could be a risk factor for autoimmunity, which is consistent with earlier reports on a link between MeCP2 and autoimmunity [41][42][43].
Although Fiumara et al. [31] did not find anti-neuronal and anti-myelin ganglioside as well as anti-nuclear, antistriated muscle, and anti-smooth muscle autoantibodies, a recent report by Papini et al. [44] evidenced high levels of serum anti-N(Glc) autoantibodies in RTT. The immune regulatory response appears to be also aberrant in RTT in that while the cytokines produced by Treg cells (i.e., IL-10, TGF-1) [50] were unchanged in MECP2-RTT, they appear to be increased in CDKL5-RTT. T-reg cells are critical in regulating the inflammatory process and Mecp2 has been recently shown to be a crucial player in determining T-reg resilience against inflammation [27]. The loss of control of the T-reg function is one of the known causes for the loss of the immune tolerance [65].
Interestingly, the behaviour of circulating IL-22 was the opposite in MECP2-RTT and in CDKL5-RTT. Although IL-22, a proinflammatory cytokine, is considered to be upregulated in chronic inflammatory diseases, its role seems to be depending on the biological context in that it can be able to favour the antimicrobial defense, regeneration, and protection against damage and induce acute phase reactants, as well as some chemokines [66]. Another proinflammatory cytokine, IL-17A, when compared to our control group, was found to be significantly increased in MECP2-RTT but significantly decreased in CDKL5-RTT. Likewise, when compared to our control group, levels of the anti-inflammatory cytokine IL-37 [49] were found to be significantly elevated in MECP2-RTT as the consequence of a compensatory mechanism.
Overall, the observed cytokine dysregulation does not appear to translate into a primary immunodeficiency, nor in a classical autoimmune disease [44].
Although the reasons behind to date observed cytokine dysregulation in RTT are unknown, possible explanations may include either a defective genetic/epigenetic control on target genes or an aberrant redox imbalance. Our findings, combined with published evidence, suggest that Mecp2 lossof-function mutations could lead to abnormal redox homeostasis via the Jak/STAT3 signaling pathway [67,68]. Our demonstration of a redox imbalance in MECP2-RTT and   CDKL5-RTT confirms and extends prior evidence of OS as a hallmark feature of RTT [16,18]. Oxidation of polyunsaturated fatty acids, with consequent enhanced formation of IsoPs and NeuroPs [69,70], is one of major features of OS. A key difference between MECP2-RTT and CDKL5-RTT appears to reside in the F 2 -dihomo-IsoPs behaviour, in that their levels are comparable to those of control group in CDKL5-RTT, whereas they are significantly increased in MECP2-RTT. To date, no clear explanation exists for these findings; it might reflect the clinical difference between two disorders. Clinical differences between the disorders may account for the observed redox pattern differences, in that MECP2-RTT can be considered a multisystemic disease, whereas CDKL5-RTT is essentially a myoclonic encephalopathy.
Therefore, a complex interplay exists between cytokines, redox homeostasis, and inflammatory status in RTT. Statistically, cytokine levels were found to explain a very consistent fraction of the observed variance for subclinical inflammation and redox abnormality, thus indicating that the aberrant  immune response, as regulated by cytokine signalling, is intimately related to redox imbalance and both are likely responsible for modulating phenotype severity in RTT.
Likely, the more striking result of our study is the evidence of a persistent and unexplained hyper-TNF-status in RTT. Hyperleptinemia in MECP2-RTT has been previously reported by Blardi et al. [71]. The findings of our study would allow a novel data interpretation from a different perspective. Leptin is able to modulate both innate and adaptive immune response [1]. Indeed, the overall leptin action in the immune system is that of a proinflammatory effect, activating proinflammatory cells, promoting Th1 responses, and mediating the production of the other proinflammatory cytokines, such as TNF-, IL-2, or IL-6 [72]. Therefore, it is plausible that, in untreated MECP2-RTT girls, the hyper-TNF-status might be triggered by hyperleptinemia.
EPA and DHA, that is, the major -3 PUFAs contained in fish oil, are known to partly inhibit several aspects of inflammation, including leukocyte chemotaxis, adhesion molecule expression, production of eicosanoids, production of inflammatory cytokines, and T-helper 1 lymphocyte reactivity [77]. The critical interplay between inflammation and OS in the underlying mechanisms leading from gene mutation to disease expression [78] is further supported by the modulatory effect of -3 PUFAs on cytokine patterns in RTT. Our findings demonstrate that -3 PUFAs partially counterbalance cytokine changes, aberrant redox homeostasis, and proinflammatory status.
The present study indicates that RTT is associated with a subclinical immune dysregulation, as a likely consequence of a defective inflammation regulatory signaling system. This abnormal regulation of the inflammatory response appears to be an unrecognized hallmark feature of RTT, intimately related to OS imbalance and likely contributing to disease expression. Library for online bibliographic research assistance; they acknowledge the Medical Genetics Unit of the Siena University (Head: Professor Alessandra Renieri) for MECP2 and CDKL5 gene mutation analysis. They heartily thank professional singer Matteo Setti (http://www.matteosetti.com/) for many charity concerts and continued interest in the scientific aspects of their research. This work was supported by the Regione Toscana (Bando Salute 2009; "Antioxidants ( -3 polyunsaturated Fatty Acids, lipoic acid) supplementation in Rett syndrome: A novel approach to therapy," RT no. 142).