Keywords: Glycated Albumin; Type 2 Diabetes Mellitus; Ig: immunoglobulin; Colorectal Cancer; Biomarkers; EpCAM; Galectin-3. IGF-1R: insulin growth factor-1 receptor;

Abbreviations:

JNK: c-jun kinase;

AGE: advanced glycation end product;

MAAPster: Microarray Analysis Pipeline;

AP: activator protein;

MAPK: mitogen-activated protein kinase;

BD: Becton Dickinson; mCRC: metastatic colorectal cancer;

BSA: bovine serum albumin;

MG: methylglyoxal;

cDNA: complementary DNA; mRNA: messenger RNA;

CEL: cell intensity file, mTOR: mammalian target of rapamycin;

CRC: colorectal cancer;

NF-kB: nuclear factor-kappa B;

CTC: circulating tumor cell;

O.D.: optical density;

DMEM: Dulbecco’s modified Eagle’s medium;

PBS: phosphate-buffered saline;

EMT: epithelial-mesenchymal transition;

PI3K: phosphatidylinositol 3-kinase;

EpCAM: epithelial cell adhesion molecule;

PL: phospholipase;

ERK: extracellular signal-regulated kinase; p: phospho;

FBS: fetal bovine serum; p70S6K1: protein 70 S6 kinase beta-1;

GA: glycated albumin;

PTEN: phosphatase and tensin homolog;

Gal-3: galectin-3;

RAGE: receptor for AGEs;

GAPDH: glyceraldehyde 3-phosphate dehydrogenase;

RMA: robust multi-chip average;

GEO: Gene Expression Omnibus;

RT-qPCR: reverse transcription-quantitative polymerase chain reaction;

GSE: GEO series;

SD: standard deviation;

SDS: sodium dodecyl sulfate;

SPM: serum-poor medium;

STAT: signal transducer and activator of transcription;

t: total;

T2DM: type 2 diabetes mellitus; 

Introduction

Colorectal cancer (CRC) ranks second in cancer-related deaths worldwide and is the third most frequently diagnosed cancer [1]. With the highest incidence in developed countries, the number of CRC patients globally is predicted to increase to 2.5 million by 2035 as a result of hereditary, environmental factors, and unhealthy dietary habits [2, 3]. Mainly caused by insulin resistance and pancreatic b-cell dysfunction, high blood sugar (also known as hyperglycemia) and long-term complications of hyperglycemia can result in chronic consequences, such as type 2 diabetes mellitus (T2DM), diabetic retinopathy, neuropathy, nephropathy, impaired wound healing leading to foot injury, cardiovascular disease, and cancer, particularly CRC [4-6]. The onset, development, and progression of CRC stimulated substantial attention to the advanced glycation end products (AGEs), especially glycated albumin (GA), which are primarily produced under hyperglycemic conditions [7-9].

Food is the primary exogenous source of AGEs (also called dietary AGEs) which are generated inside the body during digestion, absorption, and metabolism, and contributing to the body’s total AGEs pool [10]. Physiological endogenous AGEs represent a minor component; however, reprogramming cancer cell metabolism shifting to aerobic glycolysis (i.e., the Warburg effect) leads to increased production of methylglyoxal, a highly reactive a-dicarbonyl compound that enhances the production of AGEs [4, 11, 12]. Generated AGEs are toxic products formed during spontaneous non-enzymatic reactions between free amino groups of proteins, nucleotides, or lipids and reactive carbonyl groups of reducing sugars [13]. AGEs and their main cell surface receptor RAGE are implicated in inflammation and in the pathogenesis of several chronic diseases such as T2DM and cancer, including breast and colon cancer [14-18]. RAGE is weakly expressed in most of the tissues, including the small intestine and colon, but is strongly expressed in the developing embryo, adult brain, and lungs [19-21]. Excessive glucose, external stress, and AGEs can cause cancer cells to overexpress RAGE [22]. Numerous oncogenic signaling pathways are triggered by the AGEs-RAGE axis. These pathways include phosphatidylinositol 3-kinase (PI3K)/Akt and mitogen-activated protein kinase (MAPK), such as extracellular signal-regulated kinase (ERK)1/2, p38 and c-jun kinase (JNK), which in turn initiates the activation of transcription factors like nuclear factor-kappa B (NF-kB), activator protein (AP)1 and signal transducer and activator of transcription (STAT)-3 [18].

Colorectal cancer is a heterogeneous disease classified in clinical stratifications, biological and molecular subtypes, including adenocarcinoma and metastatic CRC [23, 24]. Epidemiological data and meta-analysis studies indicated that the colorectal adenoma incidence increases in patients diagnosed with T2DM compared to their non-diabetic counterparts [25, 26]. Clinical studies reported an increased detection of AGEs, including GA, in the blood collected from diabetic patients [27-29]. The clinical significance of the AGEs-RAGE axis in colorectal carcinogenesis was described after the detection of increased AGEs and RAGE expression levels in patient’s CRC tissues compared to adjacent non-tumor tissues [30]. The main therapy for CRC patients with T2DM is hypoglycemic agents such as metformin, the first-line antidiabetic drug with anticancer effects, as well as insulin growth factor-1 receptor (IGF-1R) and RAGE inhibitors [2]. Due to the lack of obvious clinical symptoms affecting the early diagnosis of CRC, the primary cause of high morbidity [31], it is urgent to study the oncogenic effects of GA, the main circulating AGEs, and to identify new biomarkers, which could be used for the early diagnosis of CRC in T2DM patients.

During this study, the oncogenic effects of GA on colorectal adenocarcinoma HT29 and metastatic CRC (mCRC) SW620 cell lines were evaluated using cell-based assays. In addition, the biological impact of GA on the expression levels of signaling phospho-proteins and cancer-related proteins was investigated. Bioinformatics analysis was performed on the relevant cancerrelated protein expressions induced by GA to confirm any upregulation of their gene expression levels in tissues of T2DM CRC patients compared to their non-diabetic counterparts, which could reveal potential biomarkers.

Materials and Methods

Generation of Glycated Albumin

Glycated albumin (GA), generated from bovine serum albumin (BSA) and highly reactive metabolite methylglyoxal (MG), as well as non-glycated albumin were prepared as previously described [14].

Cell Culture and Treatment

The human colorectal adenocarcinoma (advanced primary tumor) HT29 (#HTB-38) and lymph node metastasis-derived CRC SW620 (#CCL-277) cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, VA). The cells were cultured in complete medium composed of Dulbecco’s modified Eagle’s medium (DMEM from Gibco® containing 4.5 g/l glucose, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS from Gibco®), 100 UI/ml penicillin, and 100 µg/ml streptomycin (Gibco®). The cultured HT29 and SW620 cells exhibited the characteristic cobblestone epitheliallike morphological features and ovoid shape with the formation of aggregates, respectively. The cells were maintained in an incubator at 37°C in a 5% CO2 humidified atmosphere. When the cells reached 70-80% confluence, they were enzymatically harvested using Gibco® TrypLE™ Express Enzyme solution and then split in a 1:2 ratio. Between passages 3 and 7, the cells were used for downstream applications.

After cell seeding in complete medium, the next day, the medium was replaced with serum-poor medium (SPM), composed of DMEM containing low (1 g/l) glucose and supplemented with 2.5% FBS. After 24 h of incubation, the HT29 and SW620 cells were treated with various concentrations (25, 50, 100, and 200 mg/ml) of GA (reported from previous studies as MG-BSA-AGEs [14, 15]) and non-glycated albumin in SPM, for various incubation times. The untreated cells were the controls.

Cell Viability Assay

The HT29 and SW620 cells (1 × 10⁵/ml) were seeded in a 96-well plate (Nunc™, Thermo Fisher Scientific) in 100 ml of complete medium and conditioned as described. After 72 h of incubation in the presence or absence of various concentrations (25-200 mg/ml) of GA and non-glycated albumin, the cell viability was determined using the crystal violet assay method [32]. Briefly, the cells were gently washed with warm phosphate-buffered saline (PBS). After carefully removing the PBS solution, 0.2% crystal violet solution (containing 2% ethanol) was added to fix and stain the cells. After two washes in tap water, removing the non-viable cells, the staining of the viable adherent cells was solubilized after addition of 1% sodium dodecyl sulfate (SDS). The optical density (O.D.) of the solubilized solution was measured using SoftMax Pro 6.21 software on the SpectraMax Plus 384 spectrophotometer (Molecular devices, Wokingham, UK) with a reading at 570 nm. The percentage of cell viability was calculated as follows: [O.D. (treated cells) – O.D. (blank)]/ [O.D (untreated cells) – O.D.

(blank)] × 100.

Directional Cell Migration and Invasion Assays

The Boyden chamber system composed of Transwell® inserts in 24-well plates (NuncTM) with an uncoated porous membrane to assess directional cell migration, and a porous membrane coated with diluted Becton Dickinson (BD Biosciences, Franklin Lakes, NJ) growth factor-reduced Matrigel® matrix to assess cell invasion, as previously described [14]. The untreated and treated HT29 and SW620 cells (10⁴/100 ml) were seeded in SPM in the upper chamber of the insert while the lower chamber (corresponding to the well) contained SPM in the presence or absence of various concentrations (50-200 mg/ml) of GA and non-glycated albumin

(100 mg/ml, corresponding to the efficient GA concentration). Each condition was performed in duplicate and each experiment repeated independently four times. After 24 h of incubation, migrated and invaded cells were fixed, stained and counted as previously described [14].

Protein Extraction and Western Blotting Analysis

The HT29 and SW620 cells (0.5×10⁵/ml) were seeded in a 24well plate (NuncTM, Fisher Scientific, Loughborough, UK). The untreated cells and cells treated with GA or non-glycated albumin were incubated for different incubation times (in minutes, to explore signaling pathways and in hours, to monitor cell surface molecule expression levels). The cells were lysed in cold NP40 lysis buffer (Invitrogen, Paisley, UK) for protein extraction and the protein extracts in the cell lysates were kept on ice. From the protein concentration determination, separation by 12% SDS-polyacrylamide gel electrophoresis to protein transfer onto polyvinylidene fluoride membranes, Western blot procedures were performed as previously described [33] for the detection of phospho-ERK (p-ERK)1/2 and total ERK1/2 (t-ERK1/2), RAGE, epithelial cell adhesion molecule (EpCAM) and galectin-3 (Gal3) using mouse monoclonal anti-p-ERK1/2 (#sc-7976, 1:1,000 dilution, clone E4, Tyr204 of ERK1), polyclonal anti-t-ERK1/2 antibody (#sc-93, 1:1,000), rabbit monoclonal anti-RAGE antibody (#MA-29007, 1:250) from Invitrogen, and mouse monoclonal anti-Gal-3 antibody (#sc-32790, clone B2C10, 1:1,000) and anti-EpCAM antibody (#sc-66020, clone EBA-1, 1:1,000) were provided by Santa Cruz Biotechnology (Dallas, TX). Rabbit antiglyceraldehyde-3-phosphate dehydrogenase (GAPDH, #ab9485,

Abcam, 1:2,000) and anti-a-Tubulin (#ab18251, Abcam, 1:2,000) antibodies were used as loading controls for the detection of these housekeeping proteins. Infrared fluorescent IRDye® 680RD (red)conjugated goat anti-rabbit or IRDye® 800RD (green)-conjugated goat anti-mouse secondary antibodies (LI-COR Biosciences, Lincoln, NE, USA) were used to detect primary antibodies. These secondary antibodies were diluted (1:5,000 dilution) in Odyssey® blocking buffer and continuously stirred for 1 h at room temperature. Following washing, the blots were scanned, viewed using a LI-COR Odyssey CLx Scanner (LI-COR Biosciences, Lincoln, NE, USA), and analyzed using ImageJ software (http:// rsbweb.nih.gov/ij/index.html).

Phospho-kinase and oncology protein arrays

The detection of the phosphorylation of 43 human kinases was done using the Proteome Profiler Human phospho-kinase Array kit (#ARY003C, R&D systems, Minneapolis, MN, USA). After 30 min of incubation, the untreated HT29 and SW620 cells and the cells treated with 100 mg/ml (the most effective concentration) of GA and non-glycated albumin were centrifuged at 200 × g for 5 min. The detection of cancer-related proteins in the untreated and treated cell lysates was performed using the Proteome Profiler Human XL Oncology Array kit (#ARY026, R&D systems) after 48 h of incubation. All the steps from the protein extraction to the protein visualization using the ECL Chemiluminescent Kit (BioRad Laboratories) were performed according to the manufacturer’s instructions. The proteins were detected as a black spot on a white background using the LI-COR C-DiGit® Blot Scanner and quantified through densitometry using ImageJ software.

RNA extraction and reverse transcription-quantitative polymerase chain reaction (RT-qPCR) analysis

From the untreated HT29 and SW620 cells and the cells treated with 100 mg/ml of GA and non-glycated albumin for 48 h of incubation, total RNA (1000 ng) was extracted using the RNeasy mini kit (#74104, Qiagen, Hilden, Germany). Due to the low stability of RNA, the RNA extract was converted to complementary DNA using the Superscript® First-Strand Synthesis System (#4368814, Invitrogen, Carlsbad, CA) and performed on a Tetrad2 thermal cycler (Bio-Rad Laboratories, Hercules, CA). Human messenger RNA (mRNA) was monitored for EpCAM, LGALS3, and GAPDH (internal control) gene expression using primer pairs and sequences (Integrated DNA Technologies, Inc., Coralville, IA) as follows: 5’-TGT GGT TGT GGT GAT AGC AGT T-3’

(EpCAM forward); 5’-CCC ATC TCC TTT ATC TCA GCC TTC3’ (EpCAM reverse); 5’-GGC CAC TGA TTG TGC CTT AT-3’

(LGALS3 forward); 5’-TCT TTC TTC CCT TCC CCA GT-3’

(LGALS3 reverse); 5’-TGA TGA CAT CAA GAA GGT GGT GAA G-3’ (GAPDH forward) and 5’-TCC TTG GAG GCC ATG TGG GCC AT-3’ (GAPDH reverse). RT-qPCR was performed using the QuantiTect Reverse Transcription kit containing SYBR™ Green PCR Master Mix (#A25742, Applied Biosystem, Thermo Fisher) and each total reaction (10 ml) was prepared with cDNA (25 ng, 4.5 ml), SYBR™ Green PCR Master Mix (5 ml) and primer pair solution (0.5 ml, 10 mM). All reactions were run in triplicate on 7500 real-time PCR system (Applied Biosystem) using the following cyclin conditions: an initial step at 95ºC for 10 min, followed by 40 cycles of amplification at 95ºC for 15 sec and 60ºC for 1 min. Each targeted gene expression was related to that of GAPDH. The relative fold change in gene expression was calculated using the 2^-ΔΔCt method [34].

Bioinformatics analysis

Data repositories were searched for publicly available gene expression datasets containing T2DM and non-diabetic CRC patients. The raw signal intensity gene expression (cell intensity file, CEL) files were downloaded and pre-processed using the microarray analysis pipeline (MAAPster) with a robust multi-chip average (RMA) normalization method [35]. The normalized data were checked for quality and used for downstream expression analysis.

RAGE blockade

To demonstrate that GA acts via RAGE in the CRC cells, a RAGE blockade was performed using an anti-RAGE neutralizing monoclonal antibody (#MA5-29007, Invitrogen) along its isotype control immunoglobulin (Ig)G (#31235, Invitrogen) for a 2-h cell pretreatment followed by 30 min or 48 h of additional treatment, as previously described [15]. The protein extraction and Western blot analysis were also done.

Statistical Analysis

The results are expressed as mean ± standard deviation (SD). The data were obtained from at least three independent experiments. A one-way ANOVA followed by post-hoc Tukey’s test was used for multiple comparison analysis. The mean expression level and twosample independent student’s t-test were used for the comparative analysis. A value of p < 0.05 was considered significant.

Results

GA modulates HT29 and SW620 Cell Proliferation in a DoseDependent Manner

As cell proliferation is the main cellular event contributing to tumor development, we tested the effects of various concentrations

(25-200 mg/ml) of GA as well as non-glycated albumin on the human colorectal adenocarcinoma cell line HT29 and mCRC cell line SW620, after 72 h of incubation. The percentage of cell viability was determined after applying the crystal violet method. Compared to the untreated cells, defined as the control and corresponding to 100%, the addition of 50-100 mg/ml of AGEs resulted in the stimulation of HT29 (1.2-1.3-fold, p < 0.01, Figure 1A) and SW620 (1.4-1.5-fold, p < 0.05, Figure 1B) cell proliferation. Unlike the GA, under all the conditions tested, the non-glycated albumin did not modify the HT29 and SW620 cell proliferation, as compared to the untreated cells (Figure 1).

Figure 1: Effects of GA and non-glycated albumin on colorectal adenocarcinoma HT29 (A) and mCRC SW620 (B) cell viability. HT29 and SW620 cells were treated with various concentrations (25-200 µg/ml) of GA and non-glycated albumin for 72 h of incubation. Cell viability was determined using the crystal violet method and expressed as a percentage of the control, with untreated cell viability corresponding to 100%. Results are presented as mean ± SD of three independent experiments. (*), (**) and (***) signify a statistically significant difference (p < 0.05, p < 0.01 and p < 0.001), compared to the control.

GA stimulates HT29 and SW620 directional cell migration and cell invasion in a dose-dependent manner

Using the Boyden chamber systems, we explored the biological impact of GA on the directional migration of HT29 and SW620 cells and cell invasion through diluted growth factor-reduced Matrigel®, a reconstituted basement membrane. Compared to the control, the addition of 100 mg/ml of GA significantly increased (1.2-fold, p = 0.004, Figure 2A-B) the number of migrated HT29 cells, and when tested at 50 and 100 mg/ml, GA significantly augmented (1.4-fold, p = 0.027 at 50 mg/ml and 1.8-fold, p =

0.00097 at 100 mg/ml, Figure 2C-D) the number of migrated SW620 cells (Figure 2A-D). At the highest concentration (200 mg/ml), the GA had no effect on the directional migration of HT29 (Figures 2A-B) and SW620 (Figure 2C-D) cells, compared to the control (Figure 2C-D). During the cell invasion, the addition of 50100 mg/ml of AGEs significantly increased the number of invaded HT29 cells (1.50-fold, p = 0.034 at 50 mg/ml and 1.9-fold, p = 0.0013 at 100 mg/ml, Figure 2E-F) and SW620 cells (1.39-fold, p

= 0.0052 at 50 mg/ml and 1.59-fold, p = 0.00074 at 100 mg/ml, Figure 2G-H), compared to the control (Figure 2E-H). Once again, the GA had no effect on the HT29 and SW620 cell invasion when tested at 200 mg/ml, compared to the control (Figure 2). Tested at

100 mg/ml, the effective concentration of GA, the non-glycated albumin did not affect the HT29 and SW620 cell migration and invasion, compared to the control (Figure 2). 

Figure 2: Effects of GA and non-glycated albumin on directional migration (A-D) and invasion (E-H) of colorectal adenocarcinoma HT29 (A-B, E-F) and mCRC SW620 (C-D, G-H) cells using the Boyden chamber. Representative photomicrographs showing HT29 (A) and SW620 (C) cells that migrated across a porous membrane after 24 h of incubation in the absence (Control) or the presence of GA and non-glycated albumin tested at 100 mg/ml. Scale bar = 100 mm. Representative photomicrographs showing HT29 (E) and SW620 (G) cells that invaded across a diluted Matrigel®-coated porous membrane after 24 h of incubation in the absence (Control) or the presence of GA and non-glycated albumin tested at 100 mg/ml. Scale bar = 100 mm.  Quantification of the number of migrated HT29 (B) and

SW620 (D) cells and the number of invaded HT29 (F) and SW620 (H) cells after treatment with 50-200 mg/ml of GA or 100 mg/ml of non-glycated albumin, compared to control, the untreated cells. Results are presented as the mean ± SD of four independent experiments. (*), (**) and (***) represent a statistically significant difference (p < 0.05, p < 0.01 and p < 0.001), compared to the control.

GA effect on ERK1/2 phosphorylation level in HT29 and SW620 cells

To investigate the signaling pathways involved in the stimulatory effects of GA, the phosphorylation level of the crucial oncogenic signaling protein ERK1/2 was monitored in HT29 and SW620 cells, at different incubation time periods (10, 30, and 60 min) in the presence of the most effective concentration (100 mg/ml) of GA. After 30 min of exposure to the AGEs, the highest degree of ERK1/2 phosphorylation was observed for both the HT29 and SW620 cells (Figure 3). At the shortest exposure time (10 min) and beyond 30 min of exposure (i.e., 60 min), the degree of ERK1/2 phosphorylation was significantly lower than that detected after 30 min of exposure in the SW620 cells and was non-significant in the HT29 cells (Figure 3).

Figure 3: Time course of GA effect on phospho (p)-ERK1/2 expression levels monitored in HT29 and SW620 cells. Representative Western blots showing the variation in p-ERK1/2 expression levels in HT29 (left) and SW620 (right) cells stimulated by 100 µg/ml of GA after different incubation periods varying from 10 to 60 min, compared to untreated control cells. Bar graphs show the relative expression level of p-ERK1/2 in HT29 (left) and SW620 (right) cells, calculated as a ratio to total ERK1/2 (t-ERK1/2), the loading control. The results are presented as mean ± SD of three independent experiments. (*), (**) and (***) mean a statistically significant difference (p < 0.05, p < 0.01 and p < 0.001), compared to the control.

After defining the optimal exposure time corresponding to the highest degree of ERK1/2 phosphorylation induced by GA, the HT29 and

SW620 cells were exposed to various concentrations (50-200 mg/ml) of the GA and non-glycated albumin. After 30 min of incubation, a progressive and statistically significant (p < 0.05) increase in ERK1/2 phosphorylation levels was observed in the HT29 and SW620 cells after the addition of 50 and 100 mg/ml of GA, compared to the low ERK1/2 phosphorylation level detected in the untreated cells

(Figure 4). When tested at the highest concentration (200 mg/ml) on the HT29 and SW620 cells, no significant difference was obtained between the level of ERK1/2 phosphorylation detected in GA-treated cells and in untreated cells (Figure 4). In contrast to GA-induced ERK1/2 phosphorylation, no change in the ERK1/2 phosphorylation level was noticed in both the HT29 and SW620 cell lines at all the non-glycated albumin concentrations (Figure 4). 

Figure 4: Dose-dependent modulatory effect of GA on p-ERK1/2 expression levels in HT29 and SW620 cells. Representative Western blots showing the effect of various concentrations (50, 100, and 200 mg/ml) of GA and non-glycated albumin on p-ERK1/2 expression levels detected in HT29 (top) and SW620 (bottom) cell lysates, compared to the untreated cells. Bar graphs show the relative expression levels of p-ERK1/2 detected in the HT29 (top) and SW620 (bottom) cell lysates, calculated as a ratio to total ERK1/2 (t-ERK1/2), the loading control. Results are presented as mean ± SD of three independent experiments. (*) and (**) represent a statistically significant difference (p < 0.05 and p < 0.01) compared to the control.

A wider investigation of the signaling phospho-proteins involved in the GA stimulatory effects was done using a phospho-kinase array after 30 min of cellular exposure to 100 mg/ml of GA. Compared to the basal expression level detected in the untreated cells (control), in the HT29 and SW620 cells, a strong increase in pro-tumorigenic phospho-proteins, including Akt1/2/3, phospholipase (PL)-Cg and p70S6K1 was observed while the tumor suppressor phospho-p53 was concomitantly overexpressed in the SW620 cells and, to a lesser extent, in the HT29 cells (Figure 5).

Figure 5: Effect of GA on phospho-protein expression levels in HT29 and SW620 cells. Multi-immunoblots of 39 phospho-proteins expressed in (100 mg/ml) GA-treated HT29 (upper panel) and SW620 (lower panel) cells, compared to the control (untreated cells). Bar graphs show the expression levels of the detected phospho-proteins expressed in pixels. Results are presented as mean ± SD of three independent experiments.

GA upregulates Galectin-3 and EpCAM expression levels in HT29 and SW620 cells

To pinpoint the cancer-related proteins driving the GA-induced colorectal adenocarcinoma and mCRC pathogenesis, human proteome profiler oncology arrays were performed after 48 h of (HT29 and SW620) cell incubation in the presence or absence of 100 mg/ml of GA and non-glycated albumin. Of the 84 proteins listed in the oncology array, only EpCAM and Gal-3 expressions were detected in both the HT29 and SW620 cell lysates (Figure 6A). The upregulation of the detected cancer-related EpCAM and Gal-3 proteins was also confirmed at the gene expression levels, compared to the basal expression level monitored in the untreated cells, the control (Figure 6B). Besides the EpCAM and Gal-3, the protein level of RAGE, the main receptor for AGEs, was also assessed in both the HT29 and SW620 cell lysates. Using Western blot technology, RAGE, EpCAM and Gal-3 protein expression levels were significantly increased in the GA-treated HT29 and SW620 cells, compared to the basal expression level detected in the untreated cells (Figure 6C).

Figure 6: Effect of GA on cancer-related protein and gene expression levels in HT29 and SW620 cells. (A) Representative Oncology array membranes of 84 cancer-related proteins detected in the untreated HT29 (left panel) or SW620 (right panel) cell lysates and the cell lysates after treatment with 100 mg/ml of GA for 48 h of incubation, revealing the detection of Gal-3 and EpCAM. (B) RT-qPCR analyses showing the upregulation of Gal-3 and EpCAM transcripts in both the HT29 and SW620 RNA extracts after 48 h of treatment with 100 mg/ml of GA. (C) Representative Western blots showing GA-induced RAGE, Gal-3 and EpCAM upregulation detected in the HT29 (left) and SW620 (right) cell lysates, compared to the control. Bar graphs show the relative protein expression levels, calculated as a ratio to a-Tubulin. Results are presented as mean ± SD of three independent experiments. (*), (**), (***) and (****) represent a statistically significant difference (p < 0.05, p < 0.01, p < 0.001 and p < 0.0001), compared to the control.

Upregulation of Gal-3, EpCAM and p70S6K1 gene expression levels in the tissues of T2DM patients diagnosed with CRC

In the Gene Expression Omnibus (GEO) data repository for publicly available data related to CRC tissues and T2DM patients, a suitable dataset (GEO series, GSE115313) from Del Puerto-Nevado and colleagues [36] was relevant for our current study. The dataset contains formalin-fixed, paraffin-embedded tissue samples collected from 23 (7 females and 16 males) T2DM and 19 (9 females and 10 males) non-diabetic CRC patients [36]. In silico analysis showed that there was no significant difference in the RAGE gene expression levels monitored in the tissues of T2DM CRC patients and their non-diabetic counterparts (Figure 7A). However, the gene expression levels of EPCAM and LGALS3 (i.e., Gal-3) were significantly upregulated in the tissues of T2DM CRC patients compared to their non-diabetic counterparts (Table 1, Figure 7).

Table 1: A comparative analysis on the mean gene expression level for EPCAM and LGALS3 (i.e., Galectin-3) between CRC with T2DM and without T2D.

Figure 7: Boxplots showing the gene expression levels of RAGE (A), EPCAM (B) and LGALS3 (C) in tissues of T2DM CRC patients and their non-(T2DM) diabetic counterparts, retrieved from the dataset published by Del Puerto-Nevado et al [36].

Two enzymes of the phosphokinase pathway (i.e., p70S6K1 and phosphatidylinositol 3-kinase) showed significant upregulation of the gene expression in the tissues of T2DM CRC patients compared to their non-diabetic counterparts (upregulation of RPS6KB1 and PIK3C3 by 0.204 and 0.545-folds; p = 0.0009 and 0.0065, respectively) (Figure 8). Of note, there was a significant downregulation of nitric oxide synthase NOS3 by 0.144-fold (p-value = 0.0179, Figure 8).

Figure 8: Boxplots showing the gene expression levels of RPS6KB1 (A; left, using probe: 16836626; right, using probe: 17121812), PIK3C3 (B), and NOS3 (C) genes monitored in the tissues of T2DM CRC patients and their non-(T2DM) diabetic counterparts. Data retrieved from dataset published by Del Puerto-Nevado et al [36].

Loss of GA-stimulated oncogenic ERK1/2 over-phosphorylation and EpCAM and Gal-3 upregulation after RAGE blockade

To verify the key role of the GA-RAGE axis in GA-induced oncogenic ERK1/2 over-phosphorylation and upregulation of EpCAM and Gal-3 in mCRC and colorectal adenocarcinoma in T2DM conditions, a blockade of the RAGE function was performed by pretreating the cells with a neutralizing anti-RAGE antibody along its isotype IgG, used as a negative control. In the IgG-pretreated HT29 and SW620 cells, the GA induced ERK1/2 over-phosphorylation and increased Gal-3 and EpCAM expression levels, compared to the basal levels detected in the untreated control cells (Figure 9). In the HT29 and SW620 cells pretreated with anti-RAGE antibodies, the addition of GA did not augment the expression levels of p-ERK1/2, Gal-3 and EpCAM (Figure 9). Thus, GA lost its stimulatory effect on the oncogenic ERK1/2 phosphorylation and upregulation of Gal-3 and EpCAM in colorectal adenocarcinoma HT29 and mCRC SW620 cells after the RAGE blockade.

Figure 9: Loss of GA stimulatory effects on ERK1/2 phosphorylation, Gal-3 and EpCAM upregulation in HT29 (A) and SW620 (B) cells after RAGE blockade. Representative Western blot analysis showing ERK1/2 over-phosphorylation (upper panel) and the upregulation of Gal-3 and EpCAM (lower panel) in the IgG-pretreated cells exposed to GA while GA lost its stimulatory effect in the anti-RAGE antibody (Ab)-pretreated cells. Bar graphs show the relative expression of p-ERK1/2, Gal-3, and EpCAM, calculated as a ratio to a-Tubulin. The results are presented as mean ± SD of three independent experiments. (*), (**) and (***) mean a statistically significant difference (p < 0.05, p < 0.01 and p < 0.001), compared to the control.