Use of iTRAQ‐based quantitative proteomic identification of CHGA and UCHL1 correlated with lymph node metastasis in colorectal carcinoma

Abstract Metastatic dissemination of colorectal cancer (CRC), the third most common cancer type, is responsible for CRC deaths. Understanding the transition of lymph node metastasis (LNM) from Stage II to Stage III is beneficial in the prognosis and intervention of CRC. In this study, a quantitative proteomic survey was conducted to investigate the LNM‐associated proteins and evaluate the clinicopathological characteristics of these target proteins in CRC. By using the LC–MS/MS iTRAQ technology, we analysed the proteomic changes between LMN II and LMN III. Fresh tumours from the CRC specimens consisting of 12 node‐negative (Stage II) and 12 node‐positive (Stage III) cases were analysed by LC–MS/MS iTRAQ proteome analysis. Subsequently, tissue microarray with immunohistochemistry staining was conducted to access the clinicopathological characteristics of these proteins in 116 paraffin‐embedded CRC samples, each for non‐LNM and LNM CRC. To study the effects of the differentially expressed proteins on the potential mechanism, Boyden chamber assay, flow cytometry and shRNA‐based assessments were conducted to examine the role of the epithelial–mesenchymal transition (EMT) and the invasiveness of CRC cells and others in vivo xenograft mouse model experiments. Forty‐eight proteins were found differentially expressed between non‐LNM and LNM CRC tissues. Protein abundances of chromogranin‐A (CHGA) and ubiquitin carboxyl‐terminal hydrolase isozyme L1 (UCHL1) were observed in node‐positive CRC (p < 0.05). Knockdown of CHGA and UCHL1 significantly regulate cancer behaviours of HCT‐116, including inhibition of cell migration, invasiveness, cell cycle G1/S arrest and reactive oxygen species (ROS) generation. Mechanistically, the CHGA and UCHL1 inactivation displayed decreased levels of UCH‐L1, chromogranin A, β‐catenin, cyclin E, twist‐1/2, vimentin, MMP‐9, N‐cadherin and PCNA through the activation of the Rho‐GTPase/AKT/NFκB pathways. Histone modification of H3K4 trimethylation of CHGA and UCHL1 promoter were increased to activate their transcription through the signalling transduction such as Rho‐GTPase, AKT and NFκB pathways. Our results indicated that UCHL1 and chromogranin A are novel regulators in CRC lymph node metastasis to potentially provide new insights into the mechanism of CRC progression and serve as biomarkers for CRC diagnosis at the metastatic stage.


| INTRODUC TI ON
Colorectal cancer (CRC) is frequently categorized as a leading cause of cancer-related deaths as the one of the most common cancers in the world. 1 It arises from epithelial cells and causes death because of uncontrolled metastasis. 2 Although there have been significant improvements over the recent decades in the treatment of CRC, including new surgical, radiotherapy techniques and chemotherapy, the overall survival rate of patients with CRC has not remarkably changed. 3 One of the major factors for this poor outcome in CRC treatment is lymph node metastasis. Therefore, it is critical to advance the early diagnosis of CRC prior to the occurrence of distant organ metastasis. 4 Unfortunately, no early, effectively and accurately diagnostic methods for metastatic CRC is currently available. 5 Since the presence of positive lymph nodes separates Stage II from Stage III CRC as a key factor in patient management, targeting on the lymph node metastasis-associated protein biomarkers could gain in developing the early detection and monitoring markers of CRC metastasis.
The prognosis of patients with CRC is clearly and relatively dependent on the presence or absence of lymph node involvement and metastasis. 6 However, how the lymph node metastasis (LNM) develops in CRC remain unclear. In fact, multiple steps, including altered expression of many different proteins, are involved and required to develop LNM. 7 Several clinical and experimental studies demonstrate that the cellular event of epithelial-mesenchymal transition (EMT), such as upregulation of N-cadherin, participates in cancer migration and invasion and decreases patient survival rate. 7,8 In addition to the N-cadherin-mediated adherens junctions via activation of Rho-GTPase/AKT/NFκB pathways, aberrant Wnt/β-catenin activation regulates several transcription factors to trigger tumorigenesis, including members of the SNAIL family, Twist 1/2 family. [9][10][11][12] These transcription factors promote EMT of CRC malignancy through regulation of expression level of vimentin and matrix metalloproteinase-9 (MMP-9) and activity of E-cadherin. [13][14][15] However, transition mechanisms from EMT induction to the development of CRC metastasis remain to be fully understood. Meanwhile, LNM-associated proteins for early prognosis are treated as an urgent issue in CRC to identify reliable candidate markers.
Among the identified LNM-associated proteins in CRC, reliable candidate markers are yet to be produced. In examining the different functions related to varied protein expression profiles associated with LNM CRC involved in cell migration and invasiveness, proteomic analysis was applied to process and identify differential protein profiles using iTRAQ-based (isobaric tags for relative and absolute quantitation) LC-MS/MS, followed by tissue microarray. 16,17 In this study, we investigated whether experimental manipulation of ubiquitin carboxyl-terminal hydrolase isozyme L1 (UCH-L1) and chromogranin A (CHGA) expressions can influence invasion, survival and EMT of the CRC cell line. We found that intracellular signalling cascades involved in the upregulation of CHGA and UCHL1, including Rho-GTPase, AKT and NFκB pathways. Further, the results showed that these signalling pathways cause a synergistic function in histone H3 lysine 4 methylation (H3K4me3) and transcription activation of the CHGA and UCHL1 promoters. Altogether, the identification of UCH-L1 and CHGA, as the novel biomarkers and the new molecular targets of LNM respective for the early prognosis and treatment of cancer growth, migration and invasion in CRC.

| Human CRC samples and tissue microarrays (TMAs)
Collection of the CRC specimens and the corresponding normal tissue samples was from the Tissue Bank of Chang Gung Memorial Hospital-Kaohsiung Medical Center Cancer, Taiwan. Twelve paired patients with CRC (12 node-negative, 12 node-positive), who underwent surgery but had no presurgical chemotherapy or radiation therapy from 2015 to 2018, were included (Table 1). This study (IRB 104-5165B/CGMH) was approved by the institutional review boards of Chang Gung Memorial Hospital. Meanwhile, all patients informed consent. The tissues were snap-frozen in liquid nitrogen and stored at −80°C after surgery and confirmed by two independent pathologists.
TMAs were surgically resected and included 116 cases of eligible CRC specimens (60 node-negative and 56 node-positive) from the tissue bank at Chang Gung Memorial Hospital-Kaohsiung Medical Center, Taiwan ( Table 2). This protocol was approved by the Ethics Committee of Chang Gung Memorial Hospital with the patient's written informed consent. These human cancers were collected within 1 h of surgery and confirmed by the pathologist for further analysis. Anti-CHGA and UCHL1 antibodies were used in immunohistochemistry staining to detect their protein level in TMA in duplicate. 18 Assessment of immunostaining quantification was independently proceeded by a double-blinded manner. The score of CHGA and UCHL1 staining was represented as the intensity (on a scale of 0-2: negative = 0, low = 1, high = 2) and the percentage (on a scale of 0-3: 0 = zero, 1 = 1%-25%, 2 = 26%-50%, 3 = 51%-100%). 19
The antibodies were respectively obtained by Abcam Technology

| Protein extraction, protein digestion and iTRAQ labelling
The samples were immediately immersed in liquid nitrogen and resuspended in the Lysis buffer (iNtRON Biotechnology, PRO-PREP™ Protein Extraction Solution). After tissue protein samples were desalted and quantified respectively by Amicon® Ultra-15 (Millipore) and BCA protein assay (Thermo Fisher Scientific), using iTRAQ 4-plex kits to label peptides, and then reconstituted in 0.5 M TEAB (TEAB; pH 8.5). Through the iTRAQ reduction buffer (tris-2-carboxyethyl phosphine, TCEP, and iodoacetamide), the protein samples were reduced at 60°C for 30 min and alkylated at 37°C for 30 min in the dark. Based on the manufacturer's instructions (Applied Biosystems Inc.), using a SpeedVac, the iTRAQ dissolution buffer and iTRAQ labelling reagents to dry, reconstitute and label respectively, after digestion of sequencing-grade modified trypsin (Promega).

| Proteomic bioinformatic analysis
The

| Preparation of total cell extracts and immunoblot analyses
The cell protein lysates were obtained by a buffer lysis contained 1% NP-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulphate (SDS) and a protease inhibitor mixture phenylmethylsulfonyl fluoride (PMSF), aprotinin and sodium orthovanadate as previously described. After resolved in SDS-polyacrylamide gel electrophoresis (12% running and 4% stacking gels) and transferred to the PVDF membrane, expression of the protein was blotted using specific antibodies, and measured in the Western light chemiluminescent detection system (Bio-Rad)., Quantitative analysis of the area of the photo images in the immunoblots for in terms of their numbers of pixels were performed using the ImageGauge 3.46 software (Fujifilm, Inc.) as previously described. 25

| Matrigel invasion and scratch assays
After scratched a straight wound line in the monolayer of the transfected cells, the Openlab v3.0.2 image analysis software (Improvision, Coventry, UK) was used analyse to the images of the wound line as described previously. 26 The Boyden chamber for detection of matrigel invasion of tumour cells was performed as previously described. 26,27 After 24 h of incubation and fixed, the cell membrane was stained by the modified Giemsa stain (Sigma-Aldrich). Then, the quantification analysis of average number/field of the lower side of the membrane of the cells was calculated from triplicated wells with five random fields for potted a graph. With sterilized food and water, the animals were maintained under specific pathogen-free (SPF) conditions. After subcutaneous injection of the HCT-116 cells (10 6 cells/0.2 mL) into the flanks of male athymic BALB/c-nu mice with 4-week-old to 6-week-old and then tumour inoculation, these mice were randomly grouped into four groups (n = 6 per group) as described in the main text. These mice were euthanized in 18 days Their tumours and organs were collected for further analysis of the measurement of tumour volumes using callipers, including the liver, lungs and kidneys. After using 4% formaldehyde to fix the tissues and embedding them in paraffin blocks, histochemistry and immunohistochemistry analyses were performed as described previously. 27

| Chromatin immunoprecipitation (ChIP) analysis
After treatment with 1% formaldehyde for 10 min at room tempera-

| Statistical analysis
Data are represented as mean ± SD with three repeats from more than triplicate independent experiments. Using the SPSS software (version 10.0; SPSS), statistically significant differences were evaluated with one-way anova with post-hoc Mann-Whitney U test and Student's paired t-test and established at p < 0.05. 18,29 3 | RE SULTS

| Analysis of differentially expressed proteins in human colorectal cancer with lymph node metastasis by iTRAQ-based quantitative proteomics
We collected CRC samples from 12 node-positive and 12 nodenegative patients (Table 1)  were considered under-expressed) and 1.52 (more than 1.5 were deemed over-expressed). 17,20 Using this, 26 upregulated and 32 downregulated proteins in tumour tissues were found in this selective strategy (Tables 3 and 4) as compared against patients' tumour samples with node-negative II.

| Classification of differentially expressed proteins
We in Figure 2. Figure 3 revealed that The Protein-protein interaction (PPIs) among 26 upregulated proteins can be predicted using the STRING database (https://strin g-db.org/), a helpful tool for eliminating proteins groups involve in a particular pathway. 17 For instance, the network apparently shows that CHGA and UCHL1 are linked together, whose levels are affected, focusing specifically on its potential use as a prognostic and predictive biomarker in oncology. 30 Previous studies showed that the CHGA protein

| Validation of ubiquitin carboxyl-terminal hydrolase isozyme L1 and chromogranin A expression patterns on tissue microarrays
To identify the potential role of UCHL1 and CHGA in CRC, 116 CRC samples of CRCs stratified by node status tumour tissues were evaluated. In particular, CRC samples from 60 node-negative and 56 nodepositive patients ( Table 2) Table 5). 19 These findings suggested that a high expression of UCHL1 and CHGA as a potential therapeutic target might participate in the progression of CRC (p = 0.02 and p = 0.03; Table 5).

| Effects of differentially expressed proteins ubiquitin carboxyl-terminal hydrolase isozyme L1 and chromogranin A expression on cell cycle checkpoint and release of reactive oxygen species
The functional role of CHGA and UCHL1 in tumour growth, migra- According to these data, CHGA and UCHL1 inactivation induced G1 and S arrest by 83% and 82%, respectively, compared to the scrambled shRNA treated group (shControl) of HCT-116 cells ( Figure 5B). One of the functions of ROS (a class of oxygen-containing and associated active species) is to interfere with tyrosine kinases and tyrosine phosphatases, resulting in increased antioxidant and detoxification capacities and thereby admitting anti-cancer effects. 33 Using the fluorescent probes of H2DCFDA to detect extracellular superoxide release, compared to the invasiveness of the untreated group (Control) of HCT-116 cells, the downregulation of shCHGA and shUCHL1 increased the generation of ROS. Our data suggested the CHGA and UCHL1 involve in the cells that insensitivity to free radicals ( Figure 5B).

| Functional analysis of differentially expressed proteins ubiquitin carboxyl-terminal hydrolase isozyme L1 and chromogranin A on epithelialmesenchymal markers and cell invasion and survival in CRC
EMT, which is a cell  Figure 7A). Furthermore, the Boyden chamber assay showed more than 27% and 24% reduction in the invasion of the cells with the knockdown of UCHL1 and CHGA (* # p < 0.05, Figure 7B). These results indicated that that UCHL1 and CHGA as the pivotal roles in proliferation and metastasis initiation of tumour cell through the regulation of RhoGTPase/Akt/NFκB signalling pathways and EMT-related protein.

| Loss of ubiquitin carboxyl-terminal hydrolase isozyme L1 and chromogranin A expression suppresses tumorigenesis in vivo
To investigate whether the HCT-116 (shCHGA) and HCT-116

| DISCUSS ION
In most countries, CRC is highly prevalent as one of the most aggressive cancers. 35 (Table 1). In the present study, we analysed protein expression profiles in 12 pairs of Stage II/III CRC cancer samples by iTRAQ. We found 26 differentially high proteins expressed in all samples. Then, we selected CHGA and UCHL1 for further study, which indicated that UCH-L1 and CHGA high-level scores expression might contribute to CRC and is correlated with metastatic risk (p = 0.02 and p = 0.03; Figure 3, Table 5).    After normalized signal to the negative control ChIP asΔCT by subtracting the mean CT of the input from that of the individual region among the untreated control group, using the ΔΔCt method to calculate the effect of the specific inhibitors on specific genes. The data are presented as fold change of the untreated group (mean ± SD) of three independent experiments with *p < 0.01.
F I G U R E 1 0 Schematic presentation of the working model of CHGA and UCHL1 participate in promotion of the LNM-associated CRC through endogenous Rho-GTPase/AKT/NFκB signalling pathways-mediated histone modification (H3K4me3) as reliable candidate LNM-associated markers.

| CON CLUS IONS
In summary, our results indicate that the Rho-GTPase/AKT/ NFκB signalling pathway plays a critical role in the upregulation of CHGA-and UCHL1 expression that involved CRC cell survival and aggression through the methylation at H3K4 (Figure 10).
Ultimately, this study suggested the molecular mechanism by which CHGA and UCHL1 mediate the invasive pathway in CRC cells associated with the pathological Stage III lymph node metastasis.

CO N FLI C T O F I NTE R E S T S TATE M E NT
There is no financial/commercial conflict of interest. The authors in the manuscript declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All relevant data are within the paper.

I N S TITUTI O N A L R E V I E W B OA R D S TATE M E NT
The institutional review boards of Chang Gung Memorial Hospital have followed this study (IRB 104-5165B/CGMH).

I N FO R M E D CO N S E NT
Informed consent was obtained from all patients.