Comparative Proteomic Study Shows the Expression of Hint-1 in Pituitary Adenomas

Pituitary adenomas (PAs) can be unpredictable and aggressive tumors. No reliable markers of their biological behavior have been found. Here, a proteomic analysis was applied to identify proteins in the expression profile between invasive and non-invasive PAs to search for possible biomarkers. A histopathological and immunohistochemical (adenohypophyseal hormones, Ki-67, p53, CD34, VEGF, Flk1 antibodies) analysis was done; a proteomic map was evaluated in 64 out of 128 tumors. There were 107 (84%) invasive and 21 (16%) non-invasive PAs; 80.5% belonged to III and IV grades of the Hardy–Vezina classification. Invasive PAs (n = 56) showed 105 ± 43 spots; 86 ± 32 spots in non-invasive PAs (n = 8) were observed. The 13 most prominent spots were selected and 11 proteins related to neoplastic process in different types of tumors were identified. Hint1 (Histidine triad nucleotide-binding protein 1) high expression in invasive PA was found (11.8 ± 1.4, p = 0.005), especially at high index (>10; p = 0.0002). High Hint1 expression was found in invasive VEGF positive PA (13.8 ± 2.3, p = 0.005) and in Flk1 positive PA (14.04 ± 2.28, p = 0.006). Hint1 is related to human tumorigenesis by its interaction with signaling pathways and transcription factors. It could be related to invasive behavior in PAs. This is the first report on Hint expression in PAs. More analysis is needed to find out the possible role of Hint in these tumors.


Introduction
Pituitary adenomas (PAs) are the most common alteration of the pituitary gland. At the National Institute of Neurology and Neurosurgery in Mexico City, they are the second most treated tumor after meningeal tumor (22.35%) [1] and account for 91% of sellar and parasellar tumors and 10-15% of intracranial neoplasms with prevalence 2.7-24% in autopsy investigations [2][3][4]. They have monoclonal origin and are classified according to their hormonal type (lactotrophs, gonadotrophs, somatotrophs, corticotrophs, and thyrotrophs). Prolactinomas and adrenocorticotropic hormone (ACTH) PAs are the most common in women, while non-functioning pituitary adenomas (NFPAs) and growth hormone PAs are the most frequent in men [5][6][7][8]. They are classified into functioning or results of this work show the search for proteins participating in the neoplastic process that are possible candidates for diagnostic, prognostic, and/or treatment markers.

Materials and Methods
The project was approved by the ethics committee of the National Institute of Neurology and Neurosurgery (101- 17), and the study was conducted in accordance with the Declaration of Helsinki. Informed consent for the use of the samples for research purposes was obtained from the patients.
A total of 128 PA samples were collected and processed for histopathological analysis. The PAs were classified according to their clinical manifestations and radiological (MRI) and histopathological analysis ( Figure 1). The aim of this study was to analyze proteomic profiles in invasive and non-invasive PAs to search for proteins associated with the pathophysiology of these tumors. We analyzed proteomic profiles from PA tissues by 2DE, protein spots were selected and identified by mass spectrometry, and Hint1 protein expression was assessed in PA. The results of this work show the search for proteins participating in the neoplastic process that are possible candidates for diagnostic, prognostic, and/or treatment markers.

Materials and Methods
The project was approved by the ethics committee of the National Institute of Neurology and Neurosurgery (101- 17), and the study was conducted in accordance with the Declaration of Helsinki. Informed consent for the use of the samples for research purposes was obtained from the patients.
A total of 128 PA samples were collected and processed for histopathological analysis. The PAs were classified according to their clinical manifestations and radiological (MRI) and histopathological analysis ( Figure 1).

Histopathological Processing
Surgical specimens for the patient's diagnosis of PAs were collected during surgery. The surgical procedure performed for tumor resection was transnasal-transsphenoidal and transcranial. PA tissue intended for histopathological analysis was fixed with 10% formaldehyde, dehydrated in an automatic tissue processor using alcohol (Histokinette 2000, Reichert-Jung, American Optical Buffalo, New York, NY, USA), and embedded in paraffin. Four micrometer-thick sections were stained with hematoxylin-eosin and observed under a wide-field Nikon photo microscope (Nikon Co., Tokyo, Japan). Neuropathologist experts analyzed the biopsies and determined the histopathological diagnostic. PA were characterized by immunohistochemistry with adenohypophyseal hormones, proliferation, and angiogenic markers.

Immunohistochemistry
Sections of each case were deparaffinized, rehydrated, and rinsed in phosphate-buffer saline (PBS); antigenic retrieval was performed heating the sections in a pressure cooker in a commercial solution (Reveal, Biocare Medical, Concord, CA, USA) for 7 min, and rinsed again in PBS. Afterwards, endogenous peroxidase was blocked with 0.25% H2O2/distilled water for 20 min, rinsed in PBS at first, and then in 0.1%-Triton

Histopathological Processing
Surgical specimens for the patient's diagnosis of PAs were collected during surgery. The surgical procedure performed for tumor resection was transnasal-transsphenoidal and transcranial. PA tissue intended for histopathological analysis was fixed with 10% formaldehyde, dehydrated in an automatic tissue processor using alcohol (Histokinette 2000, Reichert-Jung, American Optical Buffalo, New York, NY, USA), and embedded in paraffin. Four micrometer-thick sections were stained with hematoxylin-eosin and observed under a wide-field Nikon photo microscope (Nikon Co., Tokyo, Japan). Neuropathologist experts analyzed the biopsies and determined the histopathological diagnostic. PA were characterized by immunohistochemistry with adenohypophyseal hormones, proliferation, and angiogenic markers.

Immunohistochemistry
Sections of each case were deparaffinized, rehydrated, and rinsed in phosphate-buffer saline (PBS); antigenic retrieval was performed heating the sections in a pressure cooker in a commercial solution (Reveal, Biocare Medical, Concord, CA, USA) for 7 min, and rinsed again in PBS. Afterwards, endogenous peroxidase was blocked with 0.25% H 2 O 2 /distilled water for 20 min, rinsed in PBS at first, and then in 0.1%-Triton X-100 PBS. The sections were incubated in a wet chamber, in primary antibodies against adenohypophyseal hormone antibodies (Prl, GH, LH, FSH, ACTH, TSH), proliferation marker antibodies (Ki67, p53), angiogenic marker antibodies (CD34, VEGF, Flk1), and Hint1 protein (Table 1) at 4 • C overnight. Slides were washed after incubation with the primary antibody, and the reaction was detected by streptavidin-biotin system with the Peroxidase Mouse & Rabbit kit (Diagnostic BioSystems, Pleasanton, CA, USA) and revealed with diaminobenzidine Diagnostics 2021, 11, 330 4 of 17 using a 2 Component DAB Pack kit (BioGenex, Carpinteria, CA, USA) according to the manufacturer's instructions; by the last, sections were hematoxylin counterstained. Normal human pituitary tissue sections were used as a positive control of the immunohistochemical technique, and sections incubated without primary antibody were used as negative control. Immunodetection was analyzed by two experts, under a wide-field photo microscope Nikon (Nikon Co. Tokyo, Japan). The assessment of immunohistochemical reaction (adenohypophyseal hormones, p53, VEGF, Flk1) was reported as positive in tissues with constant brown stain (nuclear or cytoplasmic) throughout the histological section, ruling out a scant or sporadic mark. Ki67i and Hint1 proteins were evaluated by quantifying the number of positively stained nuclei in five to 10 high power fields at ×400 per case and were reported as mean value of positive nuclei. For the evaluation of microvascular density (CD34d), blood positive vessels were quantified in the three most vascularized areas (hot spots) of the tumors at ×400 magnification (×40 objective lens) per case, and the average of each slide was obtained [44].
All data were presented as mean ± SD. To analyze the variables, the Kolmogorov-Smirnov test was performed, and the data were analyzed by one-way ANOVA (IBM SPSS Statistics v. 25.0; GraphPad Prism 5). To identify whether gender, age, MRI characteristics, or hormone expression on tissue were associated with the expression of Ki-67i, p53, CD34, VEGF, Flk1, and Hint1, Student's t-test, analysis of variance, and chi-square test were used; p < 0.05 was considered statistically significant.

Protein Pituitary Adenoma Extraction
During neurosurgery, a dry tumor sample was collected in a sterile Eppendorf tube and kept on ice immediately to avoid degradation. The sample was evaluated by a neuropathologist to separate tumor tissue and stored at −80 • C until process. PA proteins were analyzed by 2-DE using immobilized pH gradient strips [45]. Tumor tissue samples were homogenized in lysis buffer (7 M urea, 2 M thiourea, 4% CHAPS, 2% IPG buffer pH 3-10 (GE, Healthcare, Piscataway, NJ, USA), 40 mM DTT) with protease inhibitor cocktail (Complete, Roche Diagnostics, Indianapolis, IN, USA) and phosphatase inhibitors (PhosStop, Roche Diagnostics, Penzberg, Germany) followed by three freeze/thaw cycles in liquid nitrogen. Then, the samples were centrifuged at 14,000 × g at 4 • C for 30 min, and the supernatant was recovered and subsequently precipitated with acetone at −20 • C with methanol/ chloroform. Finally, the proteins were cleaned with a 2-D Clean-Up kit (GE Healthcare, Piscataway, NJ, USA) according to the manufacturer's instructions. The samples were solubilized in rehydration buffer (Destreak Rehydration Solution, GE Healthcare, USA) with protease and phosphatase inhibitors. Protein quantification was done using a 2-D Quant Kit (GE Healthcare, Biosciences, USA) according to the supplier's instructions. One hundred fifty µL samples (5.153 µg/10 µL ± 1.784; 77 ± 26.8 µg protein) were applied to IPG strips pH 3-10, 7 cm (GE Healthcare, Sweden), and rehydrated at room temperature for 16 h. Afterwards, IEF was carried out in a Protean IEF Cell (Bio-Rad, Hercules, CA, United States), following the manufacturer's protocol. Second dimension was performed in 15% SDS-PAGE, and gels were stained with Bio-Safe Coomassie (Bio-Rad, USA).

Image and Data Analysis of Gel
Images of 2-DE gels were obtained using the Fusion FX 6 Edge V. 070 imaging system (Vilber Lourmat, Collégien, France), and the proteomic profile of each biopsy was analyzed, using Bio1D EvolutionCapt (Vilber Lourmat, Collégien, France). Protein spots were quantified in each 2-DE gel and compared in invasive and non-invasive PA; differential protein spots (different between invasive and non-invasive PAs) were selected, considering those that showed outstanding intensity and a better definition. The quantification number of 2-DE spots was analyzed by Student's t-test.

Nanoflow LC-MS/MS
Protein identification was done at PlanTECC National Laboratory of the Center for Research and Advanced Studies at the National Polytechnic Institute, Campus Irapuato, Guanajuato Mexico. All experiments were performed on a nanoACQUITY nano-flow liquid chromatography (LC) system, coupled to an LTQ velos linear ion trap mass spectrometer (Waters, Thermo Fisher Scientific, Bremen, Germany) equipped with a nanoelectrospray ion source.
The selected spots were manually excised from Coomassie blue 2-DE gels under sterile conditions; proteins were extracted and trypsin was digested [46]. Immediately afterwards, 3 µl digested proteins were resuspended in solvent A (0.1% formic acid) and bound to a pre-column (Symmetry ® C18, 5 µm, 180µm × 20 mm, Waters). Subsequently, the flow was then switched to a 10-cm capillary UPLC column (100 µm ID BEH-C18 1.7µm particle size). The column temperature was controlled at 35 • C. The peptides were separated by a 60-min gradient method at a flow rate of 400 nL/min. The gradient was programmed as follows: 3-50% solvent B (100% acetonitrile in 0.1% formic acid) over 30 min, 50-85% B over 2 min, 85% B over 4 min, and 3% B over 22 min. The peptides were eluted into the mass spectrometer nano-electrospray ionization source through a standard coated silica tip (NewObjective, Woburn, MA, USA). The mass spectrometer was operated in data-dependent acquisition mode in order to automatically alternate between full scan (400-1600m/z) and subsequent Top 5 MS/MS scans on the linear ion trap. Collisioninduced dissociation was performed using helium as collision gas at a normalized collision energy of 35% and 10 ms activation time. Data acquisition was controlled using Xcalibur v2.3 (Thermo Fisher Scientific).

Automated Data Evaluation Work-Flow
Tandem mass spectra were extracted in Proteome Discoverer v1.4 and searched against a database on a Sequest HT engine. Searches were executed with the following parameters: 2 Da parent MS ion window, 1 Da MS/MS ion window, and two missed cleavages allowed. The iodoacetamide derivative of cysteine (carbamidomethylcysteine) was specified on Sequest as a fixed modification and oxidation of methionine as a variable modification.

Clinical Characteristics
Demographic and clinical dates of PA are presented in Table 2. Recurrences have no relation with gender (p = 0.437), hormonal PA types (p = 0.116), or between invasive and non-invasive PAs (p = 0.983). These results are in accord with those previously reported. Although PA invasiveness has been associated with recurrence, studies have concluded that the invasive behavior itself is not a significant factor in predicting recurrence. Furthermore, suprasellar extension and/or cavernous sinus invasion are also not associated with tumor recurrence and, although larger tumors are found to recur more frequently, no statistically significant differences were found. [47].
No statistically significant difference was found between hormonal PA types and Ki-67 (p = 0.182), p53 (p = 0.733), CD34 (p = 0.620), VEGF (p = 0.138), Flk1 (p = 0.266), and gender (p = 0.095). Angiogenic grade in solid tumors, as PA, is commonly evaluated by microvascular density in which the number of vessels in a certain area are quantified. The CD34d ( Figure 3A) was 5.1 ± 3.9 (range, 1-33.3), while no statistical difference was found between invasive (CD34d 5.3 ± 3.1) and non-invasive PAs (CD34d 6.3 ± 6.6) (p = 0.274). VEGF expression was found in 49 (38.3%) cases ( Figure 3B), out of which 41 (83.7%) were invasive PA; still no statistically significant difference was found between invasive and non-invasive PAs (p = 0.563). The expression of Flk1 was positive in 36 (28.1%) PAs, out of which 33 (91.7%) were invasive ( Figure 3C). There was no statistically significant difference between invasive and non-invasive PAs (p = 0.147). An inverse relationship was found between the positive expression of Flk1 and Ki67i (p = 0.030) such that a lower Ki67i was observed in positive Flk1 PAs. Additionally, a relation between Flk1 and positive p53 expression was found (p = 0.031). No relation was found between recurrence with CD34 (p = 0.374), VEGF (p = 0.696), and Flk1 (p = 0.972).  These results show that even though 80.5% of the cases studied are invasive, they do not show aggressive behavior. It has been shown that the term invasiveness in PA is not synonymous with aggressiveness. Aggressive PAs show a Ki-67 index >3%, extensive p53 positivity, and a high rate of mitosis, recurrence, and resistance to treatment. These factors were not observed in the cases studied; still, it is necessary to follow up these patients, since this was the first surgery for most of them [16,18].

Proteomic analysis
Demographic and clinical data of PAs analyzed by proteomics are shown in Table 3.  These results show that even though 80.5% of the cases studied are invasive, they do not show aggressive behavior. It has been shown that the term invasiveness in PA is not synonymous with aggressiveness. Aggressive PAs show a Ki-67 index >3%, extensive p53 positivity, and a high rate of mitosis, recurrence, and resistance to treatment. These factors were not observed in the cases studied; still, it is necessary to follow up these patients, since this was the first surgery for most of them [16,18].

Proteomic Analysis
Demographic and clinical data of PAs analyzed by proteomics are shown in Table 3.  Sixty four samples out of 128 cases were collected and analyzed by 2-DE. The availability of tissue for proteomic analysis depended on the amount of tumor tissue obtained during surgery. The tissue was divided into two parts: one sample was used in histopathology to make the diagnosis and the other in proteomic analysis. In the proteomic analysis, 103.4 ± 42.3 spots (range, 32-260) were found distributed in the area of pH 4.0-8.0 and mass 10-100 kDa. Invasive PAs (n = 56) showed 105 ± 43 spots, while 86 ± 32 spots were observed in non-invasive PA (n = 8). No statistical difference was found between invasive and non-invasive PAs (p = 0.226). A master 2-DE gel was selected for invasive and non-invasive Pas, and the most representative and differential spots of each group were selected (Figure 4). Thirteen differential spots were selected, and 11 proteins were identified in the proteomic profile (Table 4).

Discussion
The pituitary gland is made up of different cell types, each one related (each cell type) to its hormone secretion. Therefore, PAs are a heterogeneous group of tumors due to their monoclonal origin, giving rise to different tumor types. PAs have been classified according to their radiological, clinical, and histopathological characteristics as well as their hormonal content. In this cellular diversity, many molecular processes participate, complicating the study and limiting the understanding of PAs. Despite being considered benign, PAs can be invasive, aggressive, and recurrent, causing deterioration in human health. Markers are necessary to explain their biological behavior and help in their prognosis and treatment. Our study aimed to identify proteins in the expression profile

Discussion
The pituitary gland is made up of different cell types, each one related (each cell type) to its hormone secretion. Therefore, PAs are a heterogeneous group of tumors due to their monoclonal origin, giving rise to different tumor types. PAs have been classified according to their radiological, clinical, and histopathological characteristics as well as their hormonal content. In this cellular diversity, many molecular processes participate, complicating the study and limiting the understanding of PAs. Despite being considered benign, PAs can be invasive, aggressive, and recurrent, causing deterioration in human health. Markers are necessary to explain their biological behavior and help in their prognosis and treatment. Our study aimed to identify proteins in the expression profile between invasive and non-invasive PAs to search for possible biomarkers. By proteomic strategy, differences in proteomic profile between invasive and non-invasive PAs were observed, and 11 proteins were identified. Proteins related to metabolic enzymes (Phosphoglycerate mutase 1, Gamma enolase), cellular signals (14-3-3 protein epsilon), cell structure and mobility (Tropomyosin alpha 3 chain), and energy metabolism (ATP synthase beta chain, mitochondrial) were found, as reported in PAs [30]. Proteins involved in a neoplastic process were identified, and no reports of their expression in PA were found.
Hint1 (Histidine triad nucleotide-binding protein 1) was first identified as a human PKC-interacting protein whose cDNA encodes a 13.7 KDa protein located in 5q31.2 human chromosome. Later, it was found in normal and tumor cell lines, and its function as PCK inhibitor was discarded [48,71]. Hint1 is a protein member of the histidine triad family (HIT; Hist-X-Hist-X-HistX-X; X = hydrophobic amino acid), and it is a part of a binding loop for the α-phosphate of purine nucleotides. It can bind to nucleotides as AMP, ADP, and diadenosine polyphosphates Ap3A and Ap4A. Rabbit Hint1 can also bind to several purine nucleosides and nucleoside '5-phosphates, while rabbit and human Hint1 can hydrolyze ADP in vitro; therefore, Hint1 is considered a purine nucleotide-binding protein. It has been suggested that the biological effects of unusual purine nucleotides can be mediated by Hint1 protein [49,[72][73][74][75].
It has been reported that Hint1 can play a role in transcription regulation that could affect tumorigenesis signaling pathways [50], and it has been observed that Hint1 protein can act as tumor-suppressor [48,49]. Hint1 exerts its tumor suppressor activity by binding to transcription factors, such as MITF and β-catenin, and its suppressive function is in turn regulated by an acetylation-dependent mechanism [51,76,77]. Hint1 negative regulation in TCF/β-catenin transcriptional activity was found, which represses the expression of Wnt signaling pathways target genes such as axin2 and cyclinD1 [52]. The expression of SLC20A1 (phosphate transporter 1) in a study on somatotroph adenomas was associated with the activation of the Wnt/β-catenin signaling pathway. In this study, SLC20A1 expression was related to tumor size, invasive behavior, and tumor recurrence [53]. On the other hand, the MITF-Hint1 interaction can be disrupted by the binding of the second messenger Ap4A, and MITF is activated together with post-translational modifications of Hint1 (acetylation and phosphorylation) [78,79]. It has been reported that MITF expression promotes cell proliferation, invasion, and cell survival in rat prolactinomas. MITF can reverse the antitumor effect of miR-137, which has been correlated with invasive behavior in prolactinomas. miR-137 can upregulate Wtn-inhibitory factor1 and inhibit nuclear translocation of β-catenin [78,80].
Cyclin D1 is a cell cycle regulator and can act as oncoprotein. In PAs, the correlation between cyclin D1 expression with Ki67 and tumor size has been reported, and nuclear accumulation of β-catenin and over expression of cyclin D1 and c-Myc was found in non-functioning PA [81,82]. In corticotroph adenomas, cyclin D1 has been proposed as biomarker of tumor aggressiveness [83].
In a transient transfection experiment with Hint1, apoptosis induction associated with high p53 and Bax expression and decreased Bcl-2 expression was observed [84]. The extensive p53 expression and Ki67i >3% have been associated with tumor recurrence in PA [47]. Our analysis found no relation of p53 positive expression to Ki67i; however, high Hint1i (>10) was observed in p53 positive cases and low Ki67i, which could indicate the tumor suppressor activity of Hint1. This result likely shows a possible relationship between the expression of p53 and Ki-67 and Hint1 expression in invasive PA.
The role of Hint1 in cancer migration and invasion has been analyzed. Downregulated Hint1 expression was found in metastatic lymph nodes cells in hepatocellular carcinoma, involving Hint1 in a migration and invasion process by modulating girdin and AKT expression and phosphorylation [54]. Hint1 gene and mRNA expression were assessed in a family history of gastric cancer (FHGC) cases. Higher Hint1 gene expression levels were found in antrum samples with atrophic changes in FHGC cases, while lower mRNA levels were observed in antrum samples of FHGC patients compared against control samples. The decreased Hint1 mRNA levels in FHGC patient samples could be a predisposing marker to develop gastric cancer [85]. We observed high Hint1i expression in invasive PA, and high Ki67i was found in cases of low Hint1i levels. This may point to a necessary closer follow-up of these patients, since the aggressive and invasive behavior of PA can also be related to tumor recurrence [55]. Hint1 has been proposed as potential biomarker of radiosensitivity and therapeutic target. In the gastric cancer cell line SGC-7901, Hint1 inhibits cell proliferation, arrests the cell cycle in G1 phase, and reduces the DNA damage repair induced by radiation, increasing the radiosensitivity [56]. An increase in Hint1 expression by Taraxasterol inhibits the growth of liver cancer cells and regulates Bax, Bcl2, and cyclin D1 expression in human liver cancer [86].
Angiogenesis is an important factor in cell growth, cell differentiation, and endothelial migration and is regulated by vascular endothelial growth factor (VEGF) and its receptor (Flk1). In PAs, VEGF participates in vascular network formation and PA tumorigenesis, as it is involved in cell proliferation and invasion [87][88][89]. In prolactin, PA angiogenesis, together with Ki67 >3%, p53 positive, mitoses >2, and vascular invasion, has been associated with aggressiveness and is suspicious of malignancy. A high rate of VEGF expression has been found in pituitary carcinomas. This factor has been proposed as a marker of poor outcome after partial tumor resection [90][91][92][93]. The VEGF receptor Flk1 (VEGFR-2) is a signal transducer that participates in signaling cascades that can promote endothelial cell survival, proliferation, and angiogenesis, and it has been associated with aggressiveness [87,94]. Although less vascularization has been observed in PA than in a normal pituitary gland, a greater vascular density has been reported in invasive macroprolactinomas [95,96]. Our study found no relation between CD34d and invasive behavior; however, high CD34d was observed in Hint1 positive cases, while Hint1i expression was elevated in invasive VEGF, positive PA, and Flk1-positive cases. Although a relation between Flk1 expression and p53 positive expression was found, an inverse relationship between the Flk1 positive expression and Ki-67i was observed. No reports of Hint1 participation were found in angiogenesis nor vascularization processes. The results of this analysis suggest the possible participation of HINT1 in some signaling pathway that can be related to angiogenesis mechanism. In our work, significant differences were found between invasive and non-invasive PAs. More research is necessary in order to include a greater number of non-invasive PA cases, which is a limitation in this type of study.

Conclusions
Protein expression analysis by proteomic strategy shows a reference map to assess the cell physiology under a special condition. The heterogeneous nature of pituitary adenomas, given that they are originated by different types of cells, has made it difficult to understand their behavior. Then, it has been necessary to identify biomarkers, tools that allow investigating new pathways involved in their development. Here, we identified the expression of Hint1 protein as related to human tumorigenesis by its interaction with signaling pathways and transcription factors. We found that Hint1 expression is higher in invasive pituitary adenomas, showing a possible relation with the expression of cell proliferation markers and angiogenic factors; it could also be related to invasive behavior. This is a first report on Hint1 expression in pituitary adenomas. Further and more detailed analyses are necessary to understand the signaling pathways in which Hint1 participates and elucidate the possible role of Hint1 in these tumors.  Institutional Review Board Statement: The project was approved by the Ethics committee of the National Institute of Neurology and Neurosurgery (101-17) and the study was conducted in accordance with the Declaration of Helsinki.
Informed Consent Statement: Informed consent for the use of the samples for research purposes was obtained from the patients.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.