Identification of aldolase A as a potential diagnostic biomarker for colorectal cancer based on proteomic analysis using formalin-fixed paraffin-embedded tissue

Colorectal cancer (CRC) is one of the most common cancers worldwide, and many patients are already at an advanced stage when they are diagnosed. Therefore, novel biomarkers for early detection of colorectal cancer are required. In this study, we performed a global shotgun proteomic analysis using formalin-fixed and paraffin-embedded (FFPE) CRC tissue. We identified 84 candidate proteins whose expression levels were differentially expressed in cancer and non-cancer regions. A label-free semiquantitative method based on spectral counting and gene ontology (GO) analysis led to a total of 21 candidate proteins that could potentially be detected in blood. Validation studies revealed cyclophilin A, annexin A2, and aldolase A mRNA and protein expression levels were significantly higher in cancer regions than in non-cancer regions. Moreover, an in vitro study showed that secretion of aldolase A into the culture medium was clearly suppressed in CRC cells compared to normal colon epithelium. These findings suggest that decreased aldolase A in blood may be a novel biomarker for the early detection of CRC.


Introduction
Colorectal cancer (CRC) is one of the most common cancers worldwide. If the tumor is limited to the mucosa or submucosa, CRC can be completely cured by endoscopic or surgical therapy; however, many patients are already at an advanced stage when they are diagnosed. Therefore, an early detection method is needed. Diagnostic blood tests based on detection of carcinoembryonic antigen (CEA) are now widely used for CRC, but the sensitivity of this biomarker in early-stage cancer is only 5-10 % [1,2]. The identification of novel candidate molecules that are secreted by cancer cells may lead to the development of early detection methods and improved prognosis.
In this study, we performed shotgun liquid chromatography (LC)/mass spectrometry (MS)-based global proteomic analysis using proteins from FFPE CRC tissue to identify proteins whose expression levels are modified in cancer regions. We identified aldolase A as a candidate protein with potential as a novel biomarker for early detection of CRC.

Patients and tissue specimens
The tissues used in this study were from 45 patients who underwent surgical resection for CRC at Nippon Medical School Hospital between October 2007 and August 2014. None of the patients received chemotherapy or radiation therapy prior to surgery and none had inflammatory colorectal disease such as colitis or infectious diseases. The pathological diagnosis and clinicopathological stage were determined according to the criteria of the World Health Organization [18]. After a histopathological analysis, 10 of these 45 cases were selected for proteomic analysis. Immunohistochemical analyses were performed on all 45 cases. Paraffin-embedded specimens were prepared for proteomic and immunohistochemical analyses. This study was carried out in accordance with the principles embodied in the Declaration of Helsinki, 2013, and the Japanese Society of Pathology Ethics Committee. Informed consent for the use of colorectal tissues was obtained from all patients.
Protein extraction from FFPE tissue FFPE CRC tissues from ten patients were used for proteomic analysis. Pathological and clinical information are shown in Table 1. Following histological examination of hematoxylin and eosin (H&E) sections, we separated the cancer regions from the non-cancer regions, which maintained normal structures. Sections (10 μm) were deparaffinized in xylene and rehydrated through a series of graded alcohols (100, 90, 80, and 70 %). After staining with Mayer's hematoxylin for 5 min, cancer and non-cancer regions were manually dissected under a microscope (Fig. 1). Proteins were extracted from both cancer and non-cancer regions using lysis buffer (6 M guanidine-HCl, 40 mM Tris-HCl pH 8.2, and 65 mM DTT) according to a previous report [17]. Protein concentration was measured by the Bradford method.

In-solution trypsin digestion
A gel-free digestion was performed according to the protocol described by Bluemlein et al. [19], with slight modifications. Briefly, 10 μg of protein was extracted from each sample, reduced with 45 mM DTT and 20 mM TCEP, and then alkylated using 100 mM IAA. After alkylation, samples were digested with proteomics-grade trypsin (Agilent Technologies Inc., Santa Clara, CA, USA) at 37°C for 24 h. Digests were purified using PepClean C-18 Spin Columns (Thermo, Rockford, IL, USA) according to the manufacturer's protocol.

LC-MS/MS analysis for protein identification
Approximately 2 μg of purified peptide samples were injected onto a peptide L-trap column (Chemicals Evaluation and Research Institute, Tokyo, Japan) using an HTS PAL autosampler (CTC Analytics, Zwingen, Switzerland) and further separated thorough an Advance-nano UHPLC (AMR Inc., Tokyo, Japan) using a reverse-phase C18-column (Zaplous column α, 3-μm diameter gel particles and 100 Å pore size, 0.1 × 150 mm; AMR). The mobile phase consisted of solution A (0.1 % formic acid in water) and solution B (acetonitrile). The flow rate was 500 nL/min, with a concentration gradient of acetonitrile from 5 % B to 35 % B over 120 min. Gradient-eluted peptides were analyzed using an amaZon ETD ion-trap mass spectrometer (Bruker Daltonics, Billerica, MA, USA). Data were acquired in a data-dependent manner, in which MS/MS fragmentation was performed on the ten most intense peaks of every full MS scan.
All MS/MS spectra data were searched against the SwissProt Homo sapiens database with Mascot (ver-sion_2.3.01; Matrix Science, London, UK). Search criteria were as follows: enzyme, trypsin; allowance of up to two missed cleavage peptides; mass tolerance ±0.5 Da and MS/ MS tolerance ±0.5 Da; and modifications of cysteine carbamidomethylation, methionine oxidation, and N-formylation including formyl (K), formyl (R), and formyl (N terminus).

Spectral counting analysis of identified proteins
To compare protein expression across all tissue samples, we used the spectral counting method. In this analysis, when not specific spectral peak could be identified, the expression level of that protein was taken as zero, as described in a previous report [20]. Fold-changes of expressed proteins in the base 2 logarithmic scale were calculated with Rsc based upon spectral counting [21]. Relative amounts of identified proteins were also calculated with the normalized spectral abundance factor (NSAF) [22]. Candidate proteins with modified expression levels in the cancer region were chosen to ensure that the Rsc satisfied >1 or <−1, which corresponded to fold changes of >2 or <0.5.

Bioinformatics
Functional annotations for the identified proteins with modified expression levels in the cancer region were processed using the Database for annotation, visualization, and integrated discovery (DAVID) version 6.7 (http://david. abcc.ncifcrf.gov/home.jsp) [23][24][25].

Quantitative RT-PCR
Total RNA was extracted from FFPE CRC tissue with the RNeasy FFPE Kit (QIAGEN, Valencia, CA, USA) and from CRC cell lines with the GenElute Mammalian Total RNA Miniprep Kit (Sigma). cDNA was synthesized using the SuperScript VILO cDNA Synthesis Kit for FFPE tissue and the High Capacity cDNA Reverse Transcription kit for the CRC cell line according to the manufacturer's protocols (Life Technologies Japan, Tokyo, Japan). To measure expression of aldolase A, cyclophilin A, and annexin A2 mRNA, quantitative reverse transcription PCR (qRT-PCR) was performed with the 7500 system (Applied Biosystems, Foster City, CA, USA). Primers and TaqMan probes for aldolase A (Hs00765620_m1), cyclophilin A (Hs04194521_s1), annexin A2 (Hs00743063_s1), and 18S rRNA (Hs03928990_g1) were used with the TaqMan Gene Expression Assay. qRT-PCR results were expressed relative to an internal standard concentration as a ratio of target/18S rRNA. Gene expression was measured in triplicate.
Immunohistochemistry FFPE CRC tissues from 46 patients were used for a validation analysis. Pathological and clinical information are

CRC cell lines
CRC cell lines DLD-1, SW480, and SW620 and normal human colon epithelial cell line CCD 841 CoN were purchased from the American Type Culture Collection (Manassas, VA, USA). All cells were cultured in RPMI 1640 medium supplemented with 10 % fetal bovine serum (FBS) (Gibco, Carlsbad, CA, USA) in an atmosphere containing 5 % CO 2 .

Protein preparation
CRC cells were plated at a density of 5 × 10 5 cells per dish in a 100-mm dish and grown in culture medium. After 72 h, cells were solubilized in urea lysis buffer (7 M urea, 2 M thiourea, 5 % CHAPS, and 1 % Triton X-100). To quantify secreted protein, the culture medium of CRC cells was collected after 72 h.

Western blot analysis
The cell extract, culture medium, and commercial human normal serum (ImmunoBioScience Corp., Mukilteo, WA) were subjected to SDS-PAGE under reducing conditions.
The separated proteins were transferred to polyvinylidene fluoride transfer membranes. Membranes were incubated with an anti-annexin A2 rabbit monoclonal antibody, antialdolase A rabbit monoclonal antibody, or anti-cyclophilin A antibody (Cell Signaling Technology Inc., Beverly, MA, USA) at 4°C overnight. Membranes were then washed and incubated with HRP-conjugated anti-rabbit IgG antibody (American Qualex, San Clemente, CA). After washing, blots were visualized by enhanced chemiluminescence and detected using a myECL Imager system (ThermoFisher Scientific). The same membranes were reprobed with anti-β-actin antibody (Sigma) to confirm equal loading of the proteins. All Western blot analyses were performed three times.

Statistical analysis
All data are presented as the mean ± standard error of the mean. Data between two groups were compared with the unpaired t test. Values of *P < 0.05 and **P < 0.001 were considered significant in all analyses. Computations were performed with GraphPad Prism version 5 (GraphPad Software, La Jolla, CA, USA). Expression levels of these 84 proteins were more than two-fold higher or lower in cancer regions compared to non-cancer regions of CRC samples

Protein identification and profiles in cancer and non-cancer regions of CRC tissue
To investigate the molecular profile of proteins expressed in relation to cancer progression, we performed shotgun proteomic analysis using FFPE CRC tissue. We used manual dissection to separate the cancer regions and non-cancer regions from FFPE CRC tissues and examined the protein expression not only in cells but also in the stromal tissue surrounding cells (Fig. 1). We successfully identified proteins in both cancer and non-cancer regions of FFPE tissues (Table 1). Figure 2 shows the Venn map for the proteins identified.

Semiquantitative comparison of proteins identified in cancer and non-cancer regions of CRC tissue
We next performed a label-free semiquantitative method based on spectral counting to identify proteins with modified  Twenty-one proteins that were differentially expressed in cancer regions of CRC samples were classified as extracellular region proteins by gene ontology analysis expression levels in the cancer region. The Rsc value was plotted against the corresponding protein (x-axis) in increasing order from left to right for proteins identified in the cancer and non-cancer regions. Positive and negative Rsc values indicate increased and decreased expression, respectively, in the cancer region. In Fig. 3, the NSAF value (bar) was plotted against the corresponding protein (x-axis) and the NSAF values of the cancer regions (black bar) and non-cancer regions (gray bar). Proteins are indicated above and below the x-axis. Proteins with either a high positive or negative Rsc value were considered as potential early detection markers for CRC. A total of 84 differentially expressed proteins were identified in the cancer regions (Table 3). Representative proteins that were upregulated included known tumor markers such as 60 kDa heat shock protein (HSP60) and carcinoembryonic antigen-related cell adhesion molecule 5 (CEAM5, also known as CEA). By contrast, expression levels of housekeeping proteins such as β-actin and histone H4 did not change.
Functional annotation of differentially expressed proteins in cancer regions of CRC tissue DAVID was used to perform GO analyses for the differentially expressed proteins identified for each cellular component (Fig. 4). Functional annotations were counted by normalizing to the total number of identified proteins. Based on the cellular component classification, we focused on 21 proteins classified in the extracellular region, which could potentially be detected in the blood of patients with CRC (Table 4). qRT-PCR analysis of aldolase A, cyclophilin A, and annexin A2 in cancer and non-cancer regions of CRC tissue Fructose-bisphosphate aldolase A (aldolase A), peptidylprolyl cis-trans isomerase A (also known as cyclophilin A), and annexin A2 were upregulated in cancer regions (Table 3) and consequently selected as candidate diagnostic biomarkers. To confirm the higher expression levels of the candidate proteins in cancer regions compared to noncancer regions, we performed qRT-PCR analysis. Expression levels of the mRNAs for these proteins were significantly higher in cancer compared to non-cancer regions (Fig. 5a-c).

IHC analysis of aldolase A, cyclophilin A, and annexin A2 in cancer and non-cancer regions of CRC tissue
We next performed IHC analysis of the three candidate proteins for validation with tissues from 45 CRC cases (Table 2). IHC analysis revealed that expression levels of all three proteins were significantly higher in cancer regions compared to non-cancer regions (Fig. 6).

Expression of aldolase A, cyclophilin A, and annexin A2 in CRC cell lines
To investigate whether the candidate proteins were suitable as biomarkers, their mRNA and protein expression levels in CRC cell lines were determined by qRT-PCR analysis and western blotting. All tested cell lines expressed the three candidate proteins, and their expression levels in CRC cells tended to be higher than in normal colon epithelial cells (Fig. 7a-d). Aldolase A was not detected in the culture medium of the three CRC cell lines, but it was detected in normal colon epithelial cells (Fig. 7e, upper  panel). Cyclophilin A was not detected in the culture medium of any cell lines (Fig. 7e, middle panel). The level of annexin A2 in the culture medium appeared to depend on the level of annexin A2 in the cell extract for each cell line (Fig. 7e, lower panel). Finally, we detected aldolase A in normal human serum and confirmed that the level depended on the loaded serum volume (Fig. 7f, lane 1-3).

Discussion
In this study, we used a gel-free LC-MS-based proteomics approach to identify potential biomarkers for cancer in FFPE CRC tissues. We used semiquantitative methods based on spectral counting to detect 84 proteins in which expression levels were altered >2-fold in cancer compared to noncancer regions of CRC tissue (Table 3). CEA, a known tumor marker used in diagnostic blood tests for CRC, was among these proteins. Thus, the approach used in this study could potentially be used to identify novel biomarker candidates.
To identify early detection markers for CRC that can be detected by diagnostic blood tests, we focused on proteins in the Bextracellular region^category of cellular components based on the results of GO analysis (Table 4). These proteins are secreted into the extracellular space and thus may potentially be detected in blood. These criteria led us to select aldolase A, cyclophilin A, and annexin A2 as candidates. We did not include HSP60, neutrophil defensin1, or alpha-actinin-1, because previous reports had already suggested that these proteins might be biomarkers for CRC [26][27][28][29].
Validation studies revealed that mRNA and protein expression levels of the three candidate proteins were significantly higher in the cancer region compared to non-cancer regions (Figs. 5 and 6). Moreover, to evaluate whether these proteins were useful as biomarkers, we also investigated their mRNA and protein expression levels in CRC cell lines and secretion of the proteins into the medium. Cyclophilin A was not secreted into the culture medium of all tested CRC cell lines (Fig. 7e). Thus, cyclophilin A is not suitable as a diagnostic blood tumor marker. On the other hand, annexin A2 expression in SW480 cells derived from the primary tumor site of a CRC patient was higher than expression in SW620 cells that were established from a metastatic site in the same patient ( Fig. 7c-e). Decreased annexin A2 expression in cancer cells could play a role in the progression and metastasis of CRC and may be useful for predicting metastasis.
Although mRNA and protein expression of aldolase A were detected in all tested cell lines ( Fig. 7a and d), aldolase A protein was not detected in the culture medium of the three CRC cell lines, while it was clearly detected in the culture medium of normal colon epithelial cells (Fig. 7e). Moreover, aldolase A was detected in normal human serum (Fig. 7f). These results suggest that secretion of aldolase A may be suppressed by dysfunction of aldolase A due to accumulated genetic mutations during the carcinogenesis process of CRC. Therefore, decreased aldolase A levels in blood may be useful as a biomarker for the early diagnosis of CRC.
Aldolase A is a glycolytic enzyme and contributes to various cellular functions related to muscle maintenance, cell shape and mobility regulation, striated muscle contraction, actin filament organization, and ATP biosynthesis [30][31][32][33][34][35][36][37][38][39][40]. There are several reports of elevated expression of aldolase A in the serum of patients with malignant tumors, such as those with lung and renal cancer [41,42]. However, the expression kinetics and functions of aldolase A in CRC are not well understood. To our knowledge, this is the first report of decreased secretion of aldolase A in CRC. Thus, further studies are necessary to clarify the decrease in aldolase A levels in the blood of CRC patients in order to determine if this protein can be used as an early detection biomarker for CRC.

Conclusion
In conclusion, we identified 248 proteins from FFPE CRC tissues using global shotgun proteomics. A label-free semiquantitative method based on spectral counting and GO identified 21 candidate early detection biomarkers for CRC that could potentially be detected in blood. Validation studies revealed that cyclophilin A, annexin A2, and aldolase A expression levels were significantly higher in cancer compared to non-cancer regions. In vitro studies showed that secretion of aldolase A into the culture medium was clearly suppressed in CRC cancer cells to normal colon epithelium. Finally, aldolase A may play an important role in the carcinogenesis process of CRC and may be useful as an early detection biomarker for CRC.