Acrolein contributes to human colorectal tumorigenesis through the activation of RAS-MAPK pathway

Colorectal cancer (CRC) is one of the most well-known malignancies with high prevalence and poor 5-year survival. Previous studies have demonstrated that a high-fat diet (HFD) is capable of increasing the odds of developing CRC. Acrolein, an IARC group 2A carcinogen, can be formed from carbohydrates, vegetable oils, animal fats, and amino acids through the Maillard reaction during the preparation of foods. Consequently, humans are at risk of acrolein exposure through the consumption of foods rich in fat. However, whether acrolein contributes to HFD-induced CRC has not been determined. In this study, we found that acrolein induced oncogenic transformation, including faster cell cycling, proliferation, soft agar formation, sphere formation and cell migration, in NIH/3T3 cells. Using xenograft tumorigenicity assays, the acrolein-transformed NIH/3T3 clone formed tumors. In addition, cDNA microarray and bioinformatics studies by Ingenuity Pathway Analysis pointed to the fact that RAS/MAPK pathway was activated in acrolein-transformed clones that contributed to colon tumorigenesis. Furthermore, acrolein-induced DNA damages (Acr-dG adducts) were higher in CRC tumor tissues than in normal epithelial cells in CRC patients. Notably, CRC patients with higher levels of Acr-dG adducts appeared to have better prognosis. The results of this study demonstrate for the first time that acrolein is important in oncogenic transformation through activation of the RAS/MAPK signaling pathway, contributing to colon tumorigenesis.


Results
Acrolein treatment induced cell proliferation, anchorage-independent activity, spheroid formation ability and cell migration capacity. To determine the potential role of acrolein in oncogenic transformation, we treated NIH/3T3 cells with a low dose of acrolein (7.5 μM, IC 10 ) for one month and selected NIH/3T3 Acr clones #1-#7 ( Supplementary Fig. 1A). The soft agar colony formation activity of these seven clones was analyzed, and the results showed that NIH/3T3 Acr-clones #3, #4 and #6 formed more colonies than the others (Supplementary Fig. 1B). Cell proliferation analysis showed that NIH/3T3 Acr-clone #4 (doubling time = 31.0 h) had faster proliferation than parental cells (doubling time = 39.4 h); however, NIH/3T3 Acr-clone #3 (doubling time = 55.0 h) or NIH/3T3 Acr-clone#6 (doubling time = 42.9 h) showed the opposite phenomenon (Fig. 1A, Supplementary Fig. 1C). Therefore, we selected NIH/3T3Acr-clone #4 for subsequent analysis. Consistently, cell cycle analysis showed that the ratio of NIH/3T3 Acr-clone#4 cells in S phase was markedly higher than that in parental cells (Fig. 1B, Supplementary Fig. 1D), indicating that acrolein promotes S-phase DNA synthesis and accelerates cell proliferation. Anchorage-independent activity (Fig. 1C) in NIH/3T3 Acr-clone#4 cells was also increased compared to parental NIH/3T3 cells using a soft agar colony formation assay. Spheroid formation ability on ultralow attachment plates of NIH/3T3 Acr-clone#4 was also enhanced (Fig. 1D). In addition, NIH/3T3 Acr-clone#4 cells showed enhanced migration capacity compared with NIH/3T3-mock cells (Fig. 1E) using a Transwell assay. However, the drug sensitivity of NIH/3T3 Acr-clone#4 toward chemotherapeutic agents such as oxaliplatin and 5-FU was similar to that of parental NIH/3T3 cells ( Supplementary Fig. 2). These results suggest that acrolein increases the cell cycle rate, proliferation, colony formation activity, spheroid formation ability and cell migration capacity.
NIH/3T3 Acr-clone#4 formed tumors in xenograft nude mice. Our in vitro results indicate that acrolein can transform normal mouse NIH/3T3 fibroblasts into malignant cells. To confirm its tumorigenic potential, we performed in vivo studies of tumor xenografts in nude mice using parental NIH/3T3 cells as the negative control. NIH/3T3 Acr-clone#4 and parental NIH/3T3 cells were injected subcutaneously into the right axillary fossa (5 × 10 6 cells/animal). Three weeks after injecting NIH/3T3 Acr-clone#4 into nude mice, nodular neoplasms could be observed, while tumors were obvious at 10 days, whereas the parental NIH/3T3 cells failed to form any tumors ( Fig. 2A). Tumors formed by NIH/3T3 Acr-clone#4 cells were observed, and their volumes and growth curves were calculated for 4 weeks after the tumors could be observed (Fig. 2B,C). These data further indicate that acrolein leads to oncogenic transformation in vivo.
Acrolein induced the RAS/MAPK signaling pathway in CRC tumorigenesis using ingenuity pathway analysis (IPA). To determine the underlying mechanisms by which acrolein induced oncogenic transformation, cDNA microarray analysis with IPA was performed in NIH/3T3 Acr-clone#4 cells (Fig. 3A). The results showed that four genes (Rnd1, Rras2, myc and PI3Kcb) involved in the RAS/MAPK signaling pathway were upregulated in acrolein-transformed clone #4 (NIH/3T3 Acr-clone #4) (Fig. 3B, Supplementary Table 1). These results were confirmed using Western blot analysis (Fig. 3C, Supplementary Fig. 3A). Furthermore, we also found that acrolein activated the RAS/MAPK signaling pathway and increased c-myc in NIH/3T3 cells and the human normal colon epithelium CCD-841CoN (Fig. 3D,E, Supplementary Fig. 3B,C). Intriguingly, acrolein induced cell proliferation, colony formation activity and cell migration capacity in CCD-841CoN cells (Fig. 4A,D). Activation of the RAS/MAPK signaling pathway was also observed in the CCD-841CoN Acr clone Immunohistochemistry analysis of acrolein-DNA (Acr-dG) levels in human colon cancers. Acrolein can react with DNA-inducing modifications, which, if not repaired, can result in mutations and lead to cancer development. Acrolein has been shown to produce propano-2'-deoxyguanosine (Acr-dG) adducts in human cells 15,30,31 . Acr-dG adducts are mutagenic and induce mainly G to T and G to A mutations 15,[31][32][33][34][35][36][37][38][39][40] . To further investigate whether acrolein contributes to colon cancer formation, we analyzed Acr-dG adduct expression in CRC tissues and normal epithelial cells adjacent to tumor tissues using immunohistochemical (IHC) staining. The results showed that Acr-dG adduct levels were mainly located in the nucleus and were higher in CRC tumor tissue than in normal epithelial cells in 18 CRC patients ( Fig. 5A-C). Based on our cDNA microarray data, c-myc was upregulated in acrolein-transformed cell clones (Fig. 3C). We further analyzed c-myc levels in CRC tissues and normal epithelial cells in the same patients using IHC staining ( Supplementary Fig. 5). Similar to Acr-dG adduct levels, higher c-myc levels were observed in CRC tumor tissues than in normal epithelial cells.
Higher Acr-dG expression is associated with improved survival in CRC patients. We further evaluated the effect of Acr-dG expression on CRC characteristics and patient survival. The demographic data are shown in Table 1 Fig. 5D). In addition, high expression of Acr-dG was inversely correlated with clinical stages and grades using chi-square analysis ( Table 1). The impact of confounders on Acr-dG was analyzed using Cox proportional haz- The cell migration activity of NIH/3T3 Acrclone #4 cells was analyzed using Transwell migration analysis. Scale bar: 100 µm. NIH/3T3 Acr-clone #4 had the highest cell transformation activity. Student's t tests were used to determine statistical significance, and twotailed p-values are shown. *p < 0.05, **p < 0.01, ***p < 0.005 compared with NIH/3T3 parental cells.

Discussion
Acrolein is the most reactive α, β-unsaturated aldehyde present in tobacco smoke, in ambient air pollution, and in some cooking oils heated to a high temperature 18 . Acrolein was previously evaluated as a group 3 carcinogen by the IARC Working Group in 1995; however, it was re-evaluated as probably carcinogenic to humans (Group 2A) 17 . Our previous studies have also supported that acrolein is associated with oral, lung and bladder cancer 12,13,15,16 . Furthermore, our current studies have shown that individuals could be exposed to acrolein from consuming fried food 24 . Although the association between HFD and CRC risk has been known for quite a while [27][28][29] , the exact mechanisms underlying HFD-induced colon cancer risk and recurrence remain unclear. In the present study, our results showed that acrolein induced oncogenic transformation in NIH/3T3 cells in vitro and in vivo. The underlying mechanism was through activation of the RAS/MAPK pathway, which contributes to www.nature.com/scientificreports/ colon carcinogenesis. Additionally, Acr-dG adducts were higher in CRC tumor tissues than in normal epithelial cells in CRC patients. These results suggest that acrolein may contribute to colon tumorigenesis. Furthermore, slot blot analysis showed increased Acr-dG levels in mouse colon tissues fed a HFD for 24 weeks compared to mice fed a normal diet ( Supplementary Fig. 6). Furthermore, we found that acrolein-protein conjugates (Acr-PC) were increased in colon tissues of mice fed a HFD for 4-24 weeks ( Supplementary Fig. 7), and activation of RAS/MAPK signaling was also observed in colon tissues of mice fed a HFD (data not shown). Acrolein contains a carbonyl group and an olefinic double bond and has been shown to react with cysteine, histidine, and lysine residues of proteins and with nucleophilic sites in DNA [41][42][43] . These results indicate that HFD induced acrolein production in the mouse colon and that acrolein may contribute to HFD-induced colon tumorigenesis. In vivo exposure to acrolein in most situations is quite low, and the effects may differ from those seen at acutely toxic doses 10,42 . Acrolein, the most reactive α,β-unsaturated aldehyde, rapidly binds to and depletes cellular nucleophiles such as glutathione and reacts with proteins and DNA [41][42][43] . This reactivity is the basis for the cytotoxicity evident in all cells exposed to high concentrations of acrolein [44][45][46][47] . On the other hand, acrolein inhibits cell proliferation without causing cell death at low doses [48][49][50] . Previous studies have shown that the peak concentration of daily oral acrolein exposure has been estimated at 30 μg/kg bw 25 , which is approximately 7.5 μM based on a 70 kg adult with a 5 L blood volume. Therefore, we used a sublethal dose of acrolein (7.5 μM, IC10) to expose NIH/3T3 cells for one month to mimic in vivo conditions ( Supplementary Fig. 1A). The results showed that acrolein was able to transform NIH/3T3 cells, and NIH/3T3 Acr-clones #3, #4 and #6 formed more colonies than the others (Supplementary Fig. 1B). However, cell proliferation analysis showed the opposite phenomenon (Fig. 1A, Supplementary Fig. 1C). Furthermore, NIH/3T3 Acr-clone#4 was able to form tumors, whereas no tumors were observed in mice inoculated with NIH/3T3 parental cells or NIH/3T3 Acr-clone#3 (data not shown) using a xenograft mouse model (Fig. 2). The mechanisms underlying acrolein-induced cell transformation may be related to the ability of acrolein to deplete cellular thiols or other nucleophiles and/or to effects on gene activation, either directly or subsequent to effects on redox-regulated transcription factors 10,42 . To explore the possible signaling involved in acrolein-induced oncogenic transformation, we used a cDNA microarray with IPA analysis, and the results showed that the RAS/MPAK pathway was the top of canonical pathway analysis (Fig. 3A).
Previous studies have shown that alterations in EGFR-related Ras-Raf-MAPK and PI3K-Akt pathways are involved in the pathogenesis of up to 55% and 15% of CRC, respectively 51 . Upregulation of c-myc protein plays  www.nature.com/scientificreports/ an essential role in tumorigenesis through frequently altered kinase MAPK and RAS pathways in CRC 52 . In this study, we found that acrolein upregulated the RAS/MAPK pathway followed by overexpression of c-myc in both NIH/3T3 and colon cells, CCD-841coN (Figs. 3,4). Consistently, increased c-myc expression was also observed in these CRC tumor tissues ( Supplementary Fig. 5), along with higher Acr-dG adducts in these tumor tissues ( Fig. 5A-C). Acrolein is a highly reactive aldehyde reacting with dG of DNA to form Acr-dG adducts, which were shown to be mutagenic 15,[31][32][33][34][35][36][37][38][39][40] . It is unclear whether acrolein induces mutations in RAS/MAPK pathways. These results showed that acrolein may be involved in colon tumorigenesis and that the underlying mechanism is possible through activation of the RAS/MAPK pathway and upregulation of c-myc. Interestingly, we found that CRC patients with higher Acr-dG expression in tumor tissues had a better prognosis (Fig. 5D, Table 1). A possible explanation is that Acr-dG adducts are involved in the initiation of colon tumorigenesis; however, accumulating high amounts of Acr-dG adducts trigger cellular apoptosis. Acrolein can be produced through lipid peroxidation in fast dividing cells such as cancer cells 53 . Our previous studies have shown that hypoxia induces acrolein production, resulting in cellular apoptosis 54 . Furthermore, an inverse correlation between cell viability and relative Acr-dG adduct levels was observed in acrolein-treated NIH/3T3 cells (Supplementary Fig. 8). In addition, acrolein induced cytotoxicity in the colon cancer cell lines SW480 and HCT116 (Supplementary Fig. 9). This may explain why CRC patients with higher Acr-dG adduct levels were associated with better survival. However, the detailed mechanisms still need further investigation.
The major restriction of this study is that NIH/3T3 is a mouse fibroblast cell model, and the genetic background may not be correlated with epithelia, although NIH/3T3 has been recognized as a cell line for tumorigenesis studies in vitro and in vivo 17,55,56 . We could not observe EGFR expression in NIH/3T3 cells, which is similar to previous studies showing that NIH/3T3 cells lack EGFR 57,58 . Therefore, we tried to use a normal colon cell line, CCD-841CoN, as a model and found that acrolein indeed induced activation of the RAS/MAPK pathway, which was similar to NIH/3T3 (Fig. 3E). In addition, acrolein also induced cell proliferation, colony formation activity and cell migration capacity in CCD-841CoN cells (Fig. 4A-C). Activation of the RAS/MAPK signaling pathway was also observed in the CCD-841CoN Acr-clone (Fig. 4D). Furthermore, we found that acrolein increased the phosphorylation of EGFR, indicating that activation of EGFR results in the downstream RAS/MAPK pathway in CCD-841CoN (Figs. 3E, 4D).
Taken together, we found that acrolein induced oncogenic transformation using NIH/3T3 cells in a xenograft mouse model through upregulation of the RAS/MAPK pathway. In addition, higher acrolein-induced DNA www.nature.com/scientificreports/ damage (Acr-dG adducts) was observed in tumor tissues than in adjacent normal epithelial cells in CRC patients. Interestingly, increased Acr-dG levels were associated with better prognosis in CRC patients. To our knowledge, this is the first study to show that acrolein is important in oncogenic transformation through activation of the RAS/MAPK signaling pathway, contributing to colon carcinogenesis. Thus, this study provides insight for the early detection and prevention of colon cancer in the future.

Materials and methods
Cell culture and acrolein treatment. A mouse fibroblast cell line (NIH/3T3) and human normal colorectal cell CCD 841 CoN (ATCC CRL-1790) were purchased from ATCC and maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% BCS and 15% FBS, respectively. Acrolein stock solution (Sigma-Aldrich) was prepared freshly before use. Cells at 70% confluency were treated with different concentrations of acrolein (0-10 μM) in complete culture medium for 1 months at 37 °C in the dark, and acroleincontaining medium was changed every two days.
Cell proliferation assay. Cell proliferation was determined using modified 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium (MTT; Sigma, St. Louis, MO) assay 59 . Briefly, cells (1000/well) were seeded in 96-well plates overnight and measured every day for 7 days. The resulting formazan dissolved in DMSO was measured at 570 nm, and the results are presented as the percentage of the control values. All of these experiments were performed in triplicate and were repeated independently at least three times.
Flow cytometry analysis of cell cycle phases. Cells were washed twice in ice-cold PBS and fixed in ice-cold 70% ethanol for 30 min or overnight at 4 °C. Cells were then washed in PBS and digested with DNasefree RNase A (50 U/mL) at 37 °C for 30 min. Before flow cytometry analysis, cells were resuspended in 500 μL of propidium iodide (PI, 10 μg/mL; Sigma) for DNA staining. PI staining was used to measure the cell cycle status using a Becton-Dickinson FACScan instrument and Cell Quest software.

Soft agar colony formation assay.
A soft agar colony formation assay was performed as described previously 60 . Briefly, a 3-mL aliquot of 1.2% agar in culture medium was plated in 60-mm dishes. Then, 1,000 transformed malignant or untransformed cells were mixed with 3 mL of 0.35% agar in medium and plated on solidified bottom agar. When the top agar solidified, the dishes were transferred to an incubator and cultured for 30 days. Two or three drops of the medium were added to each dish three times a week. After culturing for 30 days, the visible cell colonies were photographed and counted.
Tumor sphere culture assay. Acrolein-transformed NIH/3T3 clones were trypsinized and resuspended at 1000 cells/Ultra-Low Attachment 96-well Plate (Corning) in culture medium containing 2 mM l-glutamine, N2 supplement, B27 supplement, 20 ng/mL hrEGF (Sigma), and 20 ng/mL hrbFGF (Sigma) for two weeks. Fresh growth factors were added to the cells twice a week. Cumulative total numbers of cells from the spheroid cultures were calculated.
Cell migration assay. The cell migration assay was performed in vitro utilizing modified Boyden chambers with a Transwell apparatus (polycarbonate membranes with 8-mm pores, Corning) 61 . Parental NIH/3T3 or NIH/3T3 Acr clones (5 × 10 4 in 500 μL of growth medium/well, 6-well plates) were added to the upper chamber, and the lower chamber contained 750 μL of growth medium supplemented with 10% FCS. Cells on the upper membrane surface were wiped with a cotton swab after 24 h of incubation at 37 °C in a 5% CO 2 incubator. Membranes were then fixed and stained with crystal violet, and cells that migrated to the lower membrane surface were counted in nine random fields using a microscope at 200 × magnification. These experiments were performed in triplicate and were repeated at least three times.
Immunoblotting analysis. Cells  www.nature.com/scientificreports/ Slot blot assay for Acr-dG detection. Analysis of Acr-dG adducts in DNA samples was based on previously described methods 12,62 . Briefly, buccal DNA (0.25 μg) was loaded onto PVDF membranes using a Bio-Dot SF microfiltration apparatus (Bio-Rad, Hercules, CA). WesternDot 625 Western blotting kits (Invitrogen) were used for Western blot analysis according to the manufacturer's instructions. The membrane was probed overnight at 4 °C with anti-Acr-dG mouse monoclonal antibodies 63 . Acr-dG adducts were detected using a UVP BioDoc-It imaging system, and band density was quantified with UVP imaging software. Relative Acr-dG levels were calculated by the fluorescence intensity of Acr-dG stained with an anti-Acr-dG antibody normalized to the amount of loaded DNA stained with methylene blue.
Xenograft mouse model. week-old male Balb/c nude mice weighing 25-30 g were used. All animal experiments were approved by the Institutional Animal Care and Use Committee of National Yang-Ming University, and the study was carried out in compliance with the ARRIVE guidelines (IACUC#1070208rr). Tumors were induced by injecting acrolein-transformed NIH/3T3 cells (5 × 10 6 in 100 μl of PBS per animal) subcutaneously into the right axillary fossa of mice as described previously with slight modification 61 . To generate the tumor growth curve, measurement of tumors was performed twice a week with a digital caliper, and volumes were calculated by (length x width 2 )/2. Body weight was also evaluated twice weekly. Tumor samples were collected after sacrifice. Each sample was cut in half; one half was saved in 10% formaldehyde, and one half was stored at − 80 °C until further use.
RNA isolation and cDNA microarray analysis. Total  , and bioinformatic analysis of the differentially expressed genes (DEGs, fold change ≥ 2) was performed. The IPA identified biological functions that were most significant to the data set. DEGs that were associated with biological functions in the Ingenuity Knowledge Base (Ingenuity Systems) were used for the analysis. Fisher's exact test was used to calculate a p value that determined the probability that each biological function assigned to that network or to the data set was due to chance alone.

Collection of tissue microarray of CRC patients. A total of 236 patients diagnosed with CRC at Taipei
Veterans General Hospital were enrolled. Disease stage was assessed based on the American Joint Committee on Cancer staging system, 6th edition. Clinicopathological staging and clinical course were determined by searching a computer database containing detailed information. The medical residual samples of the patients were acquired from the residual sample bank of Taipei Veterans General Hospital, and this study was approved by the Institutional Review Board of Taipei Veterans General Hospital (VGHIRB, IRB#2020-01-010BC). VGHIRB waived the requirement for the use of informed consent. Patients were classified based on their primary tumor locations, including the right-sided colon (tumors originating in the cecum, ascending colon, hepatic flexure, and transverse colon), left-sided colorectum (tumors originating in the splenic flexure, descending colon, sigmoid colon, rectosigmoid junction, and rectum. Low-grade cancers have cancer cells that are well differentiated or moderately differentiated. High-grade cancers have cancer cells that are poorly differentiated or undifferentiated.

Immunohistochemistry (IHC) analysis of the Acr-dG adduct and c-myc.
For the tissue microarray (TMA), hematoxylin and eosin-stained sections from each paraffin-embedded, formalin-fixed block were used to define diagnostic areas, and a representative 0.6 mm core was obtained from each case and inserted in a grid pattern into a recipient paraffin block 49,50 . IHC analysis was carried out as previously described with slight modification 64 . Briefly, sections (4 µm) were then deparaffinized in xylene and rehydrated in a descending ethanol series. To enhance immunoreactivity, sections were incubated in Tris-EDTA, pH 6.0, and boiled for 12 min. Endogenous peroxidase activity was eliminated by incubation in hydrogen peroxide. Incubation with primary antibodies for Acr-dG antibody (generated in house) and c-myc (Santa Cruz, sc-40) was performed overnight at 4 °C in 1% BSA in phosphate-buffered saline (PBS). Bound antibodies were visualized with DAB (diaminobenzidine) used as a chromogen, and omission of the primary antibody served as a negative control. Positive controls (normal liver) were stained in parallel with each set of TMAs studied. Assessment of Acr-dG and c-myc immunoexpression was performed by light microscopy at × 400 magnification by a pathologist.
Statistical analyses. Descriptive statistics are presented as the mean ± standard deviation or as the number (percentage). Student's t tests were used to determine statistical significance, and two-tailed P-values are shown. A minimum of three independent replicate experiments was performed to justify the use of statistical tests. Survival was analyzed using Kaplan-Meier survival analysis, and the log rank test was used for comparison between the two groups. Multivariate analysis was performed using chi-square analysis or Fisher's exact test. Statistical significance was defined as a p < 0.05. All analyses were performed with the IBM SPSS Statistics software package, version 23.0 or R software (R version 4.1.0).