Differentially expressed proteins underlying childhood cortical dysplasia with epilepsy identified by iTRAQ proteomic profiling

Cortical dysplasia accounts for at least 14% of epilepsy cases, and is mostly seen in children. However, the understanding of molecular mechanisms and pathogenesis underlying cortical dysplasia is limited. The aim of this cross-sectional study is to identify potential key molecules in the mechanisms of cortical dysplasia by screening the proteins expressed in brain tissues of childhood cortical dysplasia patients with epilepsy using isobaric tags for relative and absolute quantitation-based tandem mass spectrometry compared to controls, and several differentially expressed proteins that are not reported to be associated with cortical dysplasia previously were selected for validation using real-time polymerase chain reaction, immunoblotting and immunohistochemistry. 153 out of 3340 proteins were identified differentially expressed between childhood cortical dysplasia patients and controls. And FSCN1, CRMP1, NDRG1, DPYSL5, MAP4, and FABP3 were selected for validation and identified to be increased in childhood cortical dysplasia patients, while PRDX6 and PSAP were identified decreased. This is the first report on differentially expressed proteins in childhood cortical dysplasia. We identified differential expression of FSCN1, CRMP1, NDRG1, DPYSL5, MAP4, FABP3, PRDX6 and PSAP in childhood cortical dysplasia patients, these proteins are involved in various processes and have various function. These results may provide new directions or targets for the research of childhood cortical dysplasia, and may be helpful in revealing molecular mechanisms and pathogenesis and/or pathophysiology of childhood cortical dysplasia if further investigated.


Introduction
Cortical dysplasia is a common cause of epilepsy and accounts for at least 14% of epilepsy cases [1], among whom more than 40% were refractory epilepsy [2]. It happens mostly in childhood [3]. Although previous researches revealed several genetic and acquired causes of childhood cortical dysplasia (CCD) and the mechanisms of its epileptogenesis [4], our a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 understanding of molecular mechanisms and pathogenesis underlying CCD with epilepsy is still limited.
In previous studies, only a few analyzed the proteomics of epilepsy patients using brain tissues or cerebrospinal fluid [5,6,7]. However, the proteomics or transcriptomics of CCD with epilepsy has not been analyzed, especially in brain tissues of CCD patients. Isobaric tags for relative and absolute quantitation (iTRAQ) is a comparative proteomic approach that can analyze up to 8 samples in one experiment, and is widely used in proteomic researches in different diseases [8,9]. Moreover, This study screened the differentially expressed proteins in brain tissues of CCD patients with epilepsy compared to traumatic intracranial hypertension (TIH) patients using iTRAQ-based tandem mass spectrometry and selected several proteins that are differentially expressed or unreported associated with CCD previously for validation using real-time quantitative polymerase chain reaction (qPCR) analysis, immunoblotting and immunohistochemistry. Our result suggests that 153 out of 3340 proteins were diffrentially expressed in patients with CCD compared to controls, and these proteins are mainly involved in mechanisms of catalytic activity, binding, molecule-structuring activity, transporter activity, and enzyme regulation activity. Among these 153 proteins, 8 proteins that have not been associated with CCD, but participate in CCD-related biological processes or have CCD-related molecular functions according to Gene Oncology, including NDRG1, FSCN1, FABP3, DPYSL5, PSAP, MAP4, CRMP1, and PRDX6 were selected and validated.

Patients and tissue preparation
The study protocol was approved by the Ethics Committee of the Second Affiliated Hospital of Chongqing Medical University (2013-026), and the study was conducted according to the principle expressed in the Declaration of Helsinki. Written consents were obtained from patients and/or their legal guardians. No author but S. L. and H. Y. had access to information that could identify individual participants during data collection, while no author had access to such information after data collection.
All the included CCD patients were recruited from 2013 August to 2014 December, and were preoperatively assessed with detailed history, neurological examination, neuropsychological test, ictal and interictal electroencephalography and MRI together with intraoperative electrocorticography for diagnosis and localization of lesions. After surgery, brain tissues from CCD patients were diagnosed by neuropathologists according to consensus of International League Against Epilepsy [10]. All included TIH patients underwent surgery had no history of neurological diseases, and the resected brain tissues were histologically normal. The resected tissues were immediately immersed into liquid nitrogen and strored at -80˚C. Brain tissues of sex-and age-matched 8 CCD patients and 8 controls were selected for iTRAQ, brain tissues from the rest 15 CCD patients and 15 controls were used for qPCR, immunoblotting and immunohistochemistry.

Sample preparation and iTRAQ reagents labeling
Total proteins were extracted with iTRAQ lysis buffer. The concentration of proteins was measured using 2-D Quant Kit (Amersham Biosciences, Uppsala, Sweden). Equal amount of proteins from each group were mixed. The pooled samples were subjected to iTRAQ labeling according to the iTRAQ kit protocol (Applied Biosystems, Framingham, MA, USA). Briefly, 2 μl reducing reagent was added to 200 μg protein and centrifuged, then it was incubated at 37˚C for 1 hour. 1 μl of Cysteine-Blocking Reagent was added for cysteine blocking. Each protein sample was digested into peptide with 4 μg Trypsin overnight at 37˚C. iTRAQ reagents were dissolved in isopropanol, and then mixed with the corresponding sample followed by incubation at room temperature for 3 hours. Samples from CCD patients were labeled with 118 tag and 121 tag, and samples from controls were labeled with 117 tag and 119 tag. All the iTRAQ reagent-labeled samples were then combined. [11,12] Peptide fractionation with Isoelectric Focusing (IEF) The labeled peptides samples were fractionated by IEF on immobilized pH gradient as described previously [13][14][15]. Briefly, the labeled peptides were dissolved in urea and Pharmalyte solution, applied to IPG strips (pH 3-10), and then focused with an IPGphor system (GE Healthcare Life Sciences Amersham Biosciences, Pittsburg, PA, USA) at 68 kVh. The IPG strips were cut into 36 pieces (0.5cm per piece). Peptides in each pieces was extracted with 0.1% formic acid and 2% acetonitrile and lyophilized and desalted with a C18 Discovery DSC-18 SPE column (Sigma-Aldrich). The desalted peptides were lyophilized again and stored at -20˚C for mass spectrometry analysis.

Mass spectrometry and gene oncology analysis
Mass spectrometry was performed with liquid chromatography coupled inline to a QStar mass spectrometer (Applied Biosystems, Framingham. MA, USA). Desalted peptides were reconstituted in a solution containing 0.1% formic acid and 2% acetonitrile, half of which was delivered into a trap column by an online capillary liquid chromatography system (Dionex Ultimate 3000, Amsterdam, The Netherlands). The peptide mixture were automatically separated on a C18-PepMap column (ThermoFisher Dionnex, Sunnyvale, CA, USA) at 0.3 μl/min. The eluent was analysed by OStar Elite Hybrid ESI Quadrupole time-of-flight tandem mass spectrometer (Applied Biosystems, Framingham. MA, USA) in an information-dependent acquisition mode. Mass spectrometer data acquisition was performed in the positive ion mode, with a selected mass range of 300-1800 m/z. A setting of 2 s was used as the total time for MS/MS events. The two charged peptides which were most abundant, with more than 20 counts, were selected for MS/MS and dynamically excluded for 30 s with ± 50 mDa mass tolerance.
Peptide identification and quantification was performed by ProteinPilot software (Applied Biosystems, Framingham, MA, USA). The search was performed using the International Protein Index (IPI) human database v3.87. Cysteine modification by MMTS was specified as a fixed modification [16].
For protein identification, a generally accepted standard which has been widely used in identifying and quantifying proteins with iTRAQ was taken [17][18][19][20]. The protein threshold was set to achieve 95% confidence, False discovery rate (FDR) statistics and 1.3-fold change cut-offs were used to classify the protein expressions as up-regulated (FDR<0.05 and iTRAQ fold-changes above 1.3) or down-regulated (FDR<0.05 iTRAQ and fold-changes below 0.77). For technical variation, while an analysis of repeated iTRAQ experiments established the technical variability to be not more than 30%. The gene oncology of each differentially expressed proteins was searched and classified using PANTHER classification system (www.pantherdb. org). 8 differentially expressed proteins which have not been reported associated with CCD, but may participate in CCD-related biological processes or have CCD-related molecular functions according to previous literature were selected for further validation.

Immunoblotting analysis
Total proteins were extracted with RIPA Lysis Buffer and the concentrations were determined with BCA Kit (Beyotime, Haimen, China). Protein samples were loaded to 10% SDS-PAGE gel for electrophoresis and transferred to PVDF membranes. The membranes were incubated in 0.4% gelatin for 1 hour at room temperature and then incubated in primary antibodies (CRMP1, DPYSL5, FSCN1, NDRG1, PRDX6) (1:1000-1:10000 dilution, Abcam, Cambridge, UK) at 4˚C overnight. After washed with TBST buffer, the membranes were incubated in HRP-conjugated secondary antibody (1:5000 dilution) for 1h at room temperature. The protein bands were visualized using ECL detection system (Millipore, Germany) and analyzed using Quantity One software (Bio-Rad Laboratories, Hercules, CA, USA).

Immunohistochemistry
Brain tissues from CCD patients and controls were fixed with 4% paraformaldehyde overnight at 4˚C, then routinely embedded in paraffin and sectioned. After dewaxing and rehydration, the sections were boiled in citrate buffer (pH 6.0) in microwave oven for 20 min for antigen retrieval. Endogenous peroxidase activity was quenched by treatment with 3% H 2 O 2 for 10 min. The sections were blocked with goat serum for 30 min and incubated at 4oC overnight with primary antibodies (CRMP1 1:250, DPYSL5 1:100 and FSCN1 1:250). Then sections were incubated in goat-anti-rabbit secondary antibodies (30 min, 37˚C) and visualized using DAB (ZSGB-Bio, Beijing, China). After counterstain with hematoxylin and dehydrated, sections were evaluated under microscope. A semi-quantitative score was applied to the images obtained from the microscope inspection. The staining intensity ranged from 1 to 3 and the percentage of positive cells was measured manually in a range from 0 to 100%. Staining intensity (1-3) was multiplied by the percentage of positive cells (0-100) to obtain a final score ranging from 0 to 300 [21]. This method was repeated three times for the immunohistochemistry of each antibody in the brain tissue samples.

Statistical analysis
Data was expressed as mean±SD, FDR statistics was used to identified differentially expressed proteins, FDR<0.05 was considered statistically significant. Intergroup differences in immunoblotting and immunohistochemistry between the CCD group and the control group were analyzed using t test or rank sum test. p<0.05 was considered statistically significant.

Discussion
In our study, 153 proteins were identified differentially expressed in brain tissues of CCD patients with epilepsy compared to controls using iTRAQ. According to the functional classification using PANTHER, the 153 differentially expressed proteins were divided into 9 categories, which were involved in activities of various biological process, including catalytic activity (38.0%), binding (26.3%), molecule-structuring activity (14.0%), transporter activity (8.8%), and enzyme regulation activity (5.3%). Among the 153 proteins, the expression of FSCN1, CRMP1, NDRG1, DPYSL5, MAP4 and FABP3 was increased in the CCD patients compared to controls, while the expression of PRDX6 and PSAP was decreased in iTRAQ analysis. And these results were validated by realtime PCR, immunoblotting and immunohistochemistry.
FSCN1 is an actin-binding protein and can affect the formation and maintenance of cytoskeleton structure [22]. FSCN1 increases in neurogenesis and can help neurites maintain their normal shape, and it is considered as a candidate gene for developmental brain disorders [23,24]. FSCN1-related pathways mainly participate in the migration of neurons, which was known as a key mechanism of cortical dysplasia [25]. So it is possible that, in our study, the increased level of FSCN1 indicates an abnormally enhanced neurogenesis, neurite outgrowth and neuronal migration, and thus, result in CCD and epilepsy. However, the effect of increased FSCN1 on neurons needs to be further researched.
Collapsin response mediator protein 1 (CRMP1) belongs to the collapsing response mediator protein family (CRMPs) which is involved in the Sema-3A signaling pathway [26][27][28], CRMP1 also regulates migration, neurite outgrouwth, and dendrite orientation of neurons, its loss can retard the radial migration and neurite outgrowth of neurons and lead to abnormal orientation of basal dendrites of neurons [29][30][31]. Similar to FSCN1, the increased level of CRMP may also possibly indicate an abnormally enhanced neuron migration and neurite  outgrowth and abnormal orientation of dendrites, which may have roles in CCD. Interestingly, in temporal lobe epilepsy (TLE) patients and animal models, decreased CRMP1 expression was reported [32]. It is possible that CRMP1 plays different roles in CCD with epilepsy and TLE. The specific role of CRMP1 in CCD with epilepsy needs further evaluation. NDRG1 is upregulated during cell differentiation, and its cellular distribution and molecular assembly changes with postnatal development, which is correlated with the maturation of brain [33]. NDRG1 exists in oligodendrocytes in cerebrum and decreases significantly at the end stage of myelin degradation [34,35], and its mutation is found related to subcortical white matter abnormalities and severe demyelinating neuropathy [36]. Interestingly, in patients with cortical dysplasia, the change of oligodendrocytes and oligodendrocyte precursor cells is conflicting in previous reports [37,38], and some patients with malformation of cortical development have reactive oligodendroglial hyperplasia [37]. These suggest complicated roles of oligodendrocytes and myelin sheath in cortical dysplasia. In our study, we found NDRG1 abnormally increased in CCD patients, which may suggest a possible mechanism of reactive oligodendroglial hyperplasia in CCD. However, whether NDRG1-mediated oligodendroglial change participate in the pathogenensis of CCD needs to be further illustrated.
Increased expression of DPYSL5 can regulate dendritic development by mediating BDNF signaling in the central nervous system and modulate the function of CRMP2 by interacting with tubulin [39,40], thus affect the cytoskeleton remodeling, which is important in CCD with epilepsy. It has been reproted that BDNF, a neurotrophin, plays an important role in dendritic arborzation and synaptic neurotransmission [41][42][43], and CRMP2, a signaling molecule of Semaphoring-3A and a repulsive guidance cue, can induce growth cone collapse and regulate neuronal polarity [28], axon elongation and multiple axon formation [44,45]. These suggest that DPYSL5 may function in CCD with epilepsy via affecting BDNF and CRMP2.
MAP4 exists in brain and many other organs, one of its isoforms was found neural cell specific and it can inhibit the movement of the microtubules in a concentration-dependent manner and reduce microtubule-stabilizing activity [46][47][48]. MAP4 is also known associated with epilepsy [49]. Notably, microtubule-associated proteins were known important in regulating neuronal migration and brain development [50]. Defects of neuronal migration can lead to cortical malformation and consequently cause severe intellectual disability and refractory epilepsy [51]. Therefore, the increase of MAP4, as in our study, may inhibit the movement and activity of microtubles and thus impair neuronal migration which participate in CCD.
FABP3 is considered as a promising and sensitive marker for minor brian injury and Creutzfeldt-Jakob disease [52,53]. FABP3 expression is very low in neonatal brains and gradually increases after birth until adulthood, its expression pattern is correlated with synaptogenesis, myelinogenesis, neurite formation and synapse maturation [54]. FABP3 regulate the incorporation of arachidonic acid into brain, and may also regulate gene expression via controlling the availability of fatty acid ligands required for PPAR and RXR activity [54]. In our study, a increased FABP3 level was found in CCD patients, possibly indicating early maturity of metabolism pattern in CCD patients, which may contribute to the formation of cortical dysplasia. Moreover, FABP3 deficiency in mice showed protective effect against experimental autoimmune encephalomyelitis [55], indicating a possible role of autoimmune inflammation in CCD.
PRDX6 is an antioxidant protein which mainly exists in glia and keeps increasing as growing, it may have important roles in alzheimer's disease and parkinson's disease [56,57]. PRDX6 can clear reactive oxygen species, regulate gene expression in brain and protect against oxidative stress-induced neuronal death [58]. Whether the reduction of PRDX6 in CCD patients is causal or consequential factor of CCD remains to be further illustrated. It is possible that, reduction of PRDX6 is a result of enhanced oxidative stress, which has been reported in previous study [58]. However, it is also possible that reduction of PRDX6 may contribute to the pathogenesis of CCD, because oxidative stress has been associated with developmental brain disorders and epileptogenesis, although the specific role of oxidative stress in the pathogenesis of cortical dysplasia remains to be illustrated [59,60].
PSAP is precursor of saposin and acts as a lysosomal protein and a potent secreted neurotrophic factor, its temporal pattern of expression in perinatal brain indicate its potential role in brain development [61]. Infants with PSAP deficiency presented multifocal myoclonus and cyanotic hypoxemia immediately after birth, grand-mal epilepsy in the following days, and cortical and white matter morphogenetic disorders [62,63]. This deficiency is considered to cause such manifestations via impairing the lipid storage [62,63]. Therefore, abnormally reduction of PSAP in CCD patients may indicate a possible role of PSAP in the pathogenesis of CCD. Moreover, in kainate-induced epilepsy models, PSAP reactively increases and protects against the neurotoxicity [64]. Thus, PSAP reduction in CCD may also participate in the neuronal damage in CCD.
In conclusion, we identified 153 differentially expressed proteins in CCD patients compared to controls. Among these proteins, FSCN1, CRMP1, NDRG1, DPYSL5, MAP4, FABP3, PRDX6 and PSAP were further validated. These proteins have not been related to CCD before. Mechanisms including neuronal migration, neurite growth, cytoskeleton remodeling, inflammation, oligodendroglia hyperplasia, metabolic pattern and lipid storage may be involved in CCD pathogenesis and/or pathophysiology via these proteins, providing potential targets and directions for future researches on cortical dysplasia. Our study also indicate a complicated pathogenetic background of CCD, as these differentially expressed proteins have various cellular distribution and function. Moreover, further study is needed to illustrate the specific effects of these differentially expressed proteins on CCD with epilepsy, considering the limited sample size due to the critical criteria of surgery in CCD patients, especially in children.
Supporting information S1 Table. Gene oncology terms of