Characterization of Butyrate-Resistant Colorectal Cancer Cell Lines and the Cytotoxicity of Anticancer Drugs against These Cells

Colorectal cancer (CRC) is the third most common cancer worldwide. The gut microbiota plays a critical role in homeostasis and carcinogenesis. Butyrate, a short-chain fatty acid produced by the gut microbiota, plays a role in intestinal homeostasis and acts as an anticancer agent by inhibiting growth and inducing apoptosis. However, microbiota studies have revealed an abnormally high abundance of butyrate-producing bacteria in patients with CRC and indicated that it leads to chemoresistance. We characterized butyrate resistance in HCT-116 and PMF-K014 CRC cells after treatment with a maximum butyrate concentration of 3.2 mM. The 50% inhibitory concentration of butyrate was increased in butyrate-resistant (BR) cells compared with that in parental (PT) cells. The mechanism of butyrate resistance was initially investigated by determining the expression of butyrate influx- and drug efflux-related genes. We found the increased expression of influx- and efflux-related genes in BR cells compared with that in PT cells. Proteomic data showed both identical and different proteins in PT and BR cells. Further analysis revealed the crossresistance of HCT-116 cells to metformin and oxaliplatin and that of PMF-K014 cells to 5-fluorouracil. Our findings suggest that the acquisition of butyrate resistance induces the development of chemoresistance in CRC cells, which may play an important role in CRC development, treatment, and metastasis.


Introduction
Colorectal cancer (CRC) is the third most common cancer and second leading cause of cancer-related deaths worldwide. In Thailand, CRC is the third most common cancer, and 11% of the cancer-related deaths are attributed to CRC [1]. According to a World Health Organization report in 2020, the death rate associated with CRC accounted for 9% of the incidence. Several risk factors are associated with CRC, including host genetic and epigenetic alterations, dietary lifestyle, environment, and microbial community imbalance [2][3][4][5]. The gut microbiota plays a key role in homeostasis and carcinogenesis, producing metabolites to maintain intestinal barrier integrity and immune homeostasis. An imbalance in the microbiome leads to carcinogenesis [6][7][8]. Currently, CRC is treated using surgery and chemoradiotherapy. Conventional chemotherapy targets rapidly dividing cells and is the main treatment strategy for improving the mortality rate of patients with CRC. In Thailand, a combination of 5-fluorouracil (5-FU) and oxaliplatin (Oxa) (FuOx) is the chemotherapeutic treatment of choice for CRC [9]. The main mechanism of action of 5-FU is the inhibition of thymidylate synthase [10]. Oxa covalently binds DNA, leading to the formation of platinum-DNA adducts that induce a prolonged G2 arrest and inhibit growth, resulting in apoptotic cell death [11]. Metformin (Met) is typically used to treat diabetes mellitus (DM); however, its use is also associated with the reduction of cervical, endometrial, lung, and colon cancer risk in patients with type 2 DM [12][13][14][15][16]. Met inhibits mitochondrial respiration, leading to an imbalance in the AMP:ATP ratio, which is monitored by-or activates-AMP-activated protein kinase [17]. Decreased cellular invasion and increased adhesion to collagen are correlated with a reduction in cell motility in human astrocytoma (brain tumor) cell lines [18]. This information suggests the potential of Met as an anticancer drug that may kill cancer cells including butyrate-resistant (BR) CRC cells. However, currently, 90% of the chemotherapy failures occur due to the invasion and metastasis of cancer cells [19].
Butyrate, a short-chain fatty acid produced by the gut microbiota, plays a role in intestinal homeostasis by inducing the proliferation and differentiation of cells of the normal colonic epithelium. However, as an anticancer agent, butyrate inhibits the growth and induces the apoptosis of cancer cells [6,20,21]. Previous studies have shown that butyrate inhibits the growth of human endometrial and ovarian cancer cells [22] and induces the apoptosis of breast cancer cells [23]. However, recent microbiota studies have revealed a higher abundance of butyrate-producing bacteria in patients with CRC than in the non-CRC population [2,24,25]. Furthermore, a correlation exists between butyrate resistance and chemoresistance [26]. Acquired resistance to butyrate induces the development of a malignant phenotype, such as the one that decreases cell death under glucosedeprivation conditions in BR colon adenocarcinomas [27,28]. A study on a BR CRC (HCT-116/BR) cell line indicated that resistance to butyrate resulted in the development of resistance against chemotherapies, such as paclitaxel, 5-FU, and doxorubicin [26]. Moreover, stem cell markers, such as OCT4 and ATP-binding cassette (ABC) transporter (ABCG2), are also highly expressed in the HCT-116/BR cell line [29]; this may result in treatment failure. Therefore, understanding the mechanisms underlying the regulation of butyrate resistance may enable us to develop better treatment strategies to eliminate cancer cells and/or improve the quality of life of patients with CRC.
In this study, we established a BR CRC cell line and subsequently evaluated the characteristics of the resistant cells, including cell morphology, butyrate sensitivity, and expression of butyrate-and drug efflux-related genes. Proteomic analyses were performed to determine the differences in parental (PT) and BR cells. Cell migration was used as an indicator of aggressiveness. Finally, anticancer drugs, including 5-FU, Oxa, and Met, were used to evaluate the cellular response. The experimental design is shown in Figure 1.

Evaluation of the Expression of Butyrate-Related Genes
Using Quantitative Reverse-Transcription Polymerase Chain Reaction (qRT-PCR). To determine the expression of butyrate-related genes, total RNA was isolated from PT and BR cells using the TRIzol Reagent (Invitrogen, Waltham, USA) in accordance with the manufacturer's instructions. RNA was then quantified using absorbance measurements, 1.5% agarose gel electrophoresis, and a Nanodrop (Thermo Fisher Scientific). RNA samples (1 μg) of sufficient quality were reverse-transcribed into complementary DNA (cDNA) [30,31] using an iScript™ cDNA synthesis kit (Bio-Rad Laboratories, Hercules, USA) in accordance with the manufacturer's instructions. GAPDH was used as the internal control. Differential expression of butyrate-related genes and ABC transporters was determined. The primer sequences are listed in Table S1. Quantitative reverse-transcription PCR (qRT-PCR) was performed in duplicates with three independent experiments. The delta-delta Ct (2 -(ΔΔCt) ) method was used to calculate the relative gene expression levels.
2.5. Proteomic Analysis. Mass spectrometry was performed at the Functional Proteomics Technology Laboratory, National Center for Genetic Engineering and Biotechnology (BIOTEC), Thailand. The PT and BR proteins were reduced, alkylated, digested with trypsin, and analyzed by liquid chromatography-mass spectrometry (MS) (Impact II, Bruker, Billerica, USA). Differentially expressed proteins were quantified and identified using the DeCyder MS differential analysis software 2.0 (GE Healthcare, Chicago, USA) and MASCOT search engine (Matrix Science, London, UK) based on the NCBI human protein database. Proteins differentially expressed among cell types were analyzed for pathway enrichment using Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis by employing R (version 4.1.1; R Foundation for Statistical Computing, Vienna, Austria). Uniquely expressed proteins in each cell type were categorized using PANTHER (version 16.0) and UniProt.

Cell Migration Assay.
To determine the migration of each cell, cell migration or wound healing assays were conducted as previously described [32]. Briefly, a 90-95% confluent cell monolayer was scratched using a pipette tip to generate a wound. Images were captured 0, 6, 12, and 24 h after wound scratching. The ImageJ (version 1.34) was then used to calculate the wound field. 2.7. Anticancer Drug Sensitivity Assay. The cytotoxic effects of anticancer drugs, including 5-FU, Oxa, and Met, were evaluated using the MTT assay. PT and BR cells were seeded in 96-well plates and incubated with various concentrations of the drugs for 72 h. After incubation, MTT was added to the cells and incubated for 2 h. To detect the MTT results, DMSO (Sigma-Aldrich) was added and incubated for 30 min. The absorbance of formazan was recorded using a microplate spectrophotometer system (SpectraMax190, Molecular Devices). The results were analyzed using Soft-MaxPro (version 2.2.1) and presented as the percentage of inhibition compared with control values.

Cellular Morphology of BR Cells.
Human colon cancer HCT-116 and PMF-K014cell lines were used to generate BR cells: HCT-BR and PMF-BR cells, respectively. To develop butyrate resistance, the cells were exposed to butyrate for three months. The morphology of the BR cells was altered slightly relative to that of the PT cells as shown in Figure 2. The HCT-116 PT cells displayed a sharp-pointed shape, while the HCT-BR cells were more rounded and expanded. The PMF-K014 PT cells displayed a polygonal epithelial structure indicating robust cell-cell interactions, while the BR cells showed a decrease in cell-cell interactions compared to the PT cells. Vacuolization was also observed in BR cells (indicated by black arrows).

Butyrate Sensitivity.
MTT assays were used to determine the 50% inhibitory concentration (IC 50 ) to evaluate the butyrate sensitivity of PT and BR cells. The viability of BR cells was significantly higher than that of their PT cells. Cell survival is shown in Figure 3, and the IC 50 of each cell line is shown in Table 1. The IC 50 value of the HCT-BR cells was 5.38-fold higher than that of the HCT-PT cells. For PMF-BR cells, the IC 50 was 19.72, which was 3.00-fold higher than that of their PT cells.

Expression of Butyrate-Related Genes and Drug Efflux
Pumps. The expression of butyrate-related genes was determined using GAPDH as an internal control.     To evaluate the expression of the efflux pump, the expression of the ABC transporter, which plays an important role in drug transport in cancer cells, was evaluated. BR cells showed higher expression of this drug efflux gene than PT cells but showed different patterns of expression for the efflux genes. We found that the ABC-A5 and ABC-C5 were significantly upregulated in both BR cells compared with that in the corresponding PT cells. The expression of ABC-B1was low in both BR cell types. HCT-BR cells showed higher expression of ABC-B6, ABC-C2, ABC-C5, and ABC-F2 than HCT-PR cells. Simultaneously, PMF-BR cells showed higher expression of ABC-C1, ABC-C3, and ABC-G2 than PMF-PT cells. However, the expression of ABC-B6 was not clear. We proposed a mechanism for resistance in Figure 6.  and 80% (2,285 proteins) of the proteins in HCT and PMF cells, respectively. We analyzed the unique proteins according to cell type and resistance status and found that 3.84% (112 proteins) and 5.56% (162 proteins) were found in HCT-PT and HCT-BR cells, respectively. Unique proteins in PMF cells accounted for 11.10% (313 proteins) and 7.70% (217 proteins) in PT and BR cells, respectively. The differential expression of proteins is shown in Figure 9, and the protein names are listed in Tables 2 and 3. We identified differential expression of proteins between the HCT and PMF cells, and 2,916 and 2,816 proteins were differentially expressed in HCT and PMF cells, respectively. Notably, the expression of coatomer complex subunit beta 2(COPB2), an essential protein required for Golgi budding and vesicular trafficking [33], was decreased in both HCT-BR and PMF-BR cells, compared with that in the corresponding PT cells.
A pie chart showing the biological process annotations of the proteins in each cell type is shown in Figure 10. HCT-BR cells expressed more biological processes than their PT cells did, including localization, response to stimuli, signaling, biological adhesion, and locomotion. For the PMF cells, PMF-BR expressed fewer biological processes than their PT cells did, except the interspecies interaction between organisms, with PMF-PT cells expressing greater reproductive and multiorganism processes, reproduction, biological adhesion, and locomotion. We then used a bar graph to illustrate KEGG pathway enrichment ( Figure 11). The pathways that were enriched in both PT cell lines were protein digestion and absorption, the phosphatidylinositol 3-kinase/Akt pathway, and pathways involved in cancers. For HCT-BR cells, pathways involved in homologous recombination and mismatch repair and signaling pathways were enriched. For PMF-BR cells, adherent junctions, cytokines, and receptors were enriched.

Cell Migration.
To examine the migration capacity of PT and BR cells, we conducted wound healing assays. The field containing the wound area after 24 h is shown in Figure 12, and the percentage of migrated cells is shown in Figure 12(b)and Table S2. We found a significant difference between the migration rates of PT and BR cells. The migration rate in HCT-PT cells (45.81%) was higher than in BR cells(33.29%). However, the migration rate in PMF-BR cells (37.96%)was higher than in their PT cells (25.39%).
3.6. Anticancer Drug Sensitivity. To test the anticancer drug sensitivity of the cells, the MTT results were used to calculate the IC 50 . The IC 50 of each cell was determined using data from three independent experiments and analyzed using

Discussion
Butyrate is an anticancer agent against which resistance has been reported in CRC cells. However, the mechanism underlying the development of butyrate resistance in CRC cells remains unclear. Investigation of the mechanism by which butyrate resistance develops in cells may help establish new therapeutic strategies for CRC. HCT-BR and PMF-BR cells were established by continuously exposing the cells to butyrate. BR cells showed slight differences in morphology with respect to their PT cells based on the results of inverted light microscopy. Increasing vascularization and cellular volume were found in BR cells. This result is similar to that of previous studies in BCS-TC2. BR2 cells, a BR human colon adenocarcinoma cell line, showed an increase in vacuolization and cellular volume [27]. The butyrate sensitivity test revealed that BR cells had a higher survival rate than the PT cells in both cell lines. The IC 50 values were also higher in BR cells than in PT cells in both cell lines. These results confirm that the BR cells exhibit greater resistance to butyrate than PT cells. Several mechanisms are involved in the resistance to butyrate, including alteration of the drug target and drug inactivation and efflux. To investigate the underlying mechanism in BR cells, we evaluated the expression of butyraterelated genes, including the butyrate receptors, GPR109A and GPR109B, and butyrate transporter, SLC5A8. We found different patterns of expression of butyrate-related genes in the two cell lines. HCT-BR cells showed an upregulation of SLC5A8, while PMF-BR cells showed an upregulation of GPR109A, GPR109B, and SLC5A8. In a previous study, the expression of butyrate-related genes was upregulated in BR mice compared with that in germ-free mice [34]. We also performed Kaplan-Meier survival analysis to evaluate the correlation between butyrate-related gene expression and survival outcome of patients with cancer treated with 5-FU and/or Oxa using The Cancer Genome Atlas (TCGA) pancancer data portal. Interestingly, the Kaplan-Meier analysis showed that only patients with CRC with high SLC5A8 expression displayed significantly worse survival than patients with low expression ( Figure S1). This finding suggests that high SLC5A8 expression found in BR cells represents a lower survival outcome in patients with CRC. We further investigated drug efflux gene expression. We found that the expression of ABC-B6, ABC-C5, and ABC-F2 was upregulated in HCT-BR cells, whereas that of ABC-C1, ABC-C3, and ABC-G2 was upregulated in PMF-BR  Furthermore, doxorubicin resistance in breast cancer cells is induced by the overexpression of ABC-C1 and ABC-F2 [36,37]. Lastly, cells expressing ABC-A5 and ABC-F2 show stem cell features [38]. We further evaluated the cytotoxicity of anticancer drugs including Met, 5-FU, and The coatomer is a cytosolic protein complex that binds to dilysine motifs and reversibly associates with Golgi nonclathrin-coated vesicles, which further mediate biosynthetic protein transport from the ER, via the Golgi up to the trans Golgi network.

12
BioMed Research International Oxa against PT and BR cells in both HCT and PMF cells.
The results revealed different drug responses in different cells. HCT-BR showed a crossresistance toward Met and Oxa, while PMF-BR showed a crossresistance to 5-FU; similarly, our previous study demonstrated that PMF-BR spheroid cells were also crossresistant to 5-FU [39]. This crossresistance may be due to the upregulation of ABC transporters. Proteomic analysis identified differential expression of proteins in HCT and PMF cells. PCA showed a difference between the BR and PT cell groups. A 14.56% variation in PC1 showed a shared dimension in HCT cells but presented a dissimilarity in PMF cells. Simultaneously, an 11.86% variation in PC2 showed a shared dimension in PMF cells whereas indicated a distinct feature of HCT cells. PCA indicated that there were some differences and similarities between PT and BR cells, and the Venn diagrams correlated with these results. The Venn diagram showed that several proteins were shared between the PT and BR cells; however, there were some unique proteins present in each cell, which resulted in differences in cell characteristics. We used the University of Alabama at Birmingham Cancer (UALCAN) data analysis portal (http://ualcan.path.uab.edu/index.html) to explore the relevant proteins enriched for the differen-tially expressed genes. We found that the expression of COPB2 was decreased in both HCT-BR and PMF-BR cells. Regarding this unique finding in HCT-BR cells, we discovered two proteins from UALCAN that showed increased expression levels: xanthine dehydrogenase (XDH) and formin homology 2 domain-containing protein 1 (FHDC1). In the PMF-BR cells, the expression of FRAS1-related extracellular matrix 1 (FREM1) showed an increase. The COPB2 protein is related to a dilated endoplasmic reticulum (ER) with granular material, prominent rough ER, and vacuoles resulting in intracellular trafficking deficiency [33]. In cancer, COPB2 is highly expressed in glioblastomas and hepatocellular carcinomas, resulting in a worse overall survival [40,41]. Additionally, studies on CRC cells have found that COPB2 plays an essential role in cancer cell proliferation and cell cycle progression [42]. The UALCAN database showed that the expression of COPB2 increases in CRC patients. However, the levels of the phosphorylated form of this protein were found to be decreased in every stage of CRC ( Figure S2). The level and function of COPB2 across different cancers is controversial. Therefore, the function of COPB2 in BR-CRC requires further investigation. In HCT-BR cells, XDH exerts purine oxidation and electron acceptor functions and is highly expressed in uterine

13
BioMed Research International corpus endometrial carcinoma and lung and pancreatic cancers( Figure S3). Furthermore, XDH is expressed at lower levels in colon adenocarcinoma tissue than in healthy tissue. Our study was performed in HCT cells which are derived from primary colorectal carcinoma [43]. Therefore, there may be a difference in the expression of XDH. A previous pan-cancer study showed that XDH is involved in proinflammatory and immune stimulation. Additionally, increases in XDH combined with adenine phosphoribosyl transferase, a key enzyme in the purine salvage pathway, result in increased sensitivity to 5-FU [44]. This finding requires further investigation to predict the effects of 5-FU more effectively in patients with CRC and the BR phenotype. Another upregulated protein in HCT-BR cells was FHDC1, a microtubule-associated formin involved in regulating actin and microtubule dynamics. The expression of FHDC1 is increased in various cancers including CRC and at its highest level in stage 1 cancer ( Figure S4). A study in lung adenocarcinoma reported that the high expression of FHDC1 was associated with significantly improved survival outcomes compared with that of low expression [45]. In PMF polygonal epithelial cells derived from the peritoneal dissemination of highly metastatic patients with CRC [46],  14 BioMed Research International we found the higher expression of FREM1 in CRC patients than in healthy patients, with the highest expression in stage 4 patients ( Figure S5). FREM1 plays a role in epidermal differentiation and is required for epidermal adhesion during embryonic development. However, the function of FREM1 in cancer is not well understood. A study in breast cancer showed that the expression of FREM1 was dramatically decreased, which correlated with a lower overall and recurrence-free survival [47]. However, the expression of FREM1 in CRC differs from that in breast cancer. Moreover, FREM1 levels increased in the primary tumor and higher cancer stages. However, the function of FREM1 in CRC remains unclear and requires investigation.
Cell migration was also investigated to evaluate the aggressiveness of BR and PT cells, and differences were found. HCT-BR showed a lower migration rate than HCT-PT. This may be due to the downregulation of the erythropoietin receptor, which is related to the downregulation of the JAK2/STAT5 signaling cascade [48,49]. In contrast, the PMF-BR cells showed a higher migration rate than PMF-PT cells. The protein expression profile showed the upregulation of Notchless protein homolog 1 in PMF-BR cells. This protein plays a role in Notch activity which  is known to promote cell migration and invasion in brain cancer [50].
Taken together, our findings suggest that 5-FU may be used as a potential anticancer agent in HCT-BR cells, representing primary tumor treatment. Meanwhile, Met and Oxa may be effective in PMF-BR cells, which represent metastatic tumors. However, further studies are needed to investigate the effects of Met.

Conclusions
Herein, we investigated the characteristics of butyrate resistance in CRC cells in both primary and metastatic conditions. The mechanisms underlying the development of butyrate resistance involve an increase in the expression levels of various efflux genes. The proteins expressed in the BR cells shared some parental characteristics, while unique proteins showed characteristic of resistance. Our study also confirmed that chemotherapy resistance arises from butyrate resistance in CRC cells. Moreover, Met showed a potential therapeutic effect against PMF-BR cells. Thus, further investigation of Metin, the molecular mechanism underlying butyrate resistance, is needed for its use in clinical settings to enhance the effectiveness of CRC therapies. Lastly, the effects of anticancer drugs on BR CRC cells requires further investigation in animal models.

Data Availability
The data that support the findings of this study are available from the corresponding author (Navakanitworakul, R), upon reasonable request (nraphatp@medicine.psu.ac.th).

Supplementary Materials
Supplementary 1. Table S1: primer sequences used for butyrate receptors and drug efflux pumps. Supplementary 2. Table S2: the migration rate (%) of parental and butyrateresistant cells. Figure Figure S4: expression of formin homology 2 domain-containing protein 1