Introduction

The widespread use of high-fructose corn syrup (HFCS) in the food industry has significantly contributed to excessive fructose consumption in modern diets, posing a growing health concern. HFCS, a sweetener found in a variety of processed foods and beverages, has been linked to an array of health issues, including obesity1, metabolic syndrome2, and colorectal cancer3,4,5. Under physiological conditions, fructose in moderate amounts (less than 1 g/kg) is absorbed through the small intestine via GLUT5 (Glucose Transporter)- 5, allowing for its clearance within the small intestine itself, thereby protecting the liver from fructose-induced liver-related pathologies6,7,8. Excess fructose can spill over to the liver and colon based on the small intestine’s capacity to absorb fructose, which varies from 5 grams to 50 grams, contingent on the saturation level of GLUT5 and individual differences5,9,10,11. As fructose is identified as a poorly absorbed carbohydrate and classified under the FODMAPs group (Fermentable, Oligo-, Di-, Mono-saccharides, And Polyols), an excessive consumption of fructose significantly elevates the likelihood of fructose interacting with colonocytes7,11.

High consumption of simple sugars and sugar-sweetened beverages, especially those containing added fructose, during adolescence, has been significantly associated with an increased risk of conventional adenoma in adulthood, with rectal adenoma being notably prevalent4. Additionally, clinical trial showed that a higher intake of sugar-sweetened beverages was linked to a significantly increased risk of cancer recurrence and mortality in stage III colon cancer patients12. However, the pathogenic mechanisms behind this association remain unclear. In the alimentary tract, cancerous cells distinctively utilize two pathways for nutrient absorption, sourcing nutrients from both the blood supply and the lumen. A recent study demonstrated that in patients with colorectal cancer (CRC), the expression of GLUT5 at the mRNA level in CRC tissue samples is significantly elevated compared to that in samples from healthy marginal colon mucosa This underscores the critical role of fructose in the hypoxic tumor microenvironment of colorectal cancer cells13. In colorectal cancer experimental models, previous studies have shown that fructose can promote tumor growth and survival in adenomatous polyposis coli (APC) mutant mice and hypoxic HCT-116 cells. This is achieved through its conversion to fructose-1-phosphate, which leads to the activation of glycolysis5,14. These findings underscore the critical role of fructose in the hypoxic tumor microenvironment of colorectal cancer cells.

Cancer cells have evolved to thrive in the harsh conditions of the tumor microenvironment, characterized by limited oxygen and nutrients. These adaptations include bioenergetic changes that not only fuel tumor growth but also enable resistance to chemotherapy. A key adaptation is the Warburg effect, where cancer cells favor glucose breakdown via anaerobic glycolysis over oxidative phosphorylation despite sufficient oxygen, which serves as a cornerstone of cancer metabolism15,16. In our previous studies on colorectal cancer cell lines, we have shown that glycolytic pyruvate not only reduces hypoxia-induced receptor-interacting protein (RIP)-dependent necroptotic cell death by scavenging of mitochondrial reactive oxygen species (ROS) but also plays a key role in driving chemoresistance. Necroptosis, also known as programmed necrosis, is regulated by RIP kinases, and its hallmark is the intracellular formation of the RIP1-RIP3 complex. Glycolytic pyruvate accomplishes this by prompting a shift in the cell cycle towards a quiescent state, underscoring its dual function in inhibit cell necroptosis and fostering chemotherapy resistance17,18,19. However, compared to glucose, dietary fructose is more likely to be present in the colon. Being a non-essential nutrient, it holds greater potential as a therapeutic target. Yet, whether fructose contributes to necroptotic death resistance in colorectal cancer (CRC) cells remains unknown. The Caco-2, HT29, and T84 human colorectal carcinoma cell lines are crucial for exploring the complex functions of intestinal epithelial cells across a range of physiological and pathological scenarios. Our research will employ various colorectal cancer cell lines to investigate the role of fructose in conferring resistance to Hx-induced necroptosis. The findings will offer potential therapeutic avenues in colorectal cancer treatment by influencing the colonic microenvironment via dietary modifications.

Materials and methods

Cell Culture Protocols

The cultivation of human colonic carcinoma cell lines, including Caco-2, HT29, and T84, utilized Dulbecco’s Modified Eagle Medium (DMEM) provided by (Invitrogen, part of Thermo Fisher Scientific, Waltham, MA, USA), which contained a concentration of 5 mM glucose and was formulated without the inclusion of pyruvate. This growth medium was additionally enriched with 10% fetal bovine serum, 15 mM HEPES buffer, penicillin at a concentration of 100 U/ml, and streptomycin at 0.1 mg/ml. Specifically, the growth medium for Caco-2 cells was augmented with an extra 10 µg per liter of holo-transferrin, with all drugs and reagents sourced from Sigma (St. Louis, MO, USA). For cell seeding, densities were set at 10^5 cells per well in 96-well plates and 10^6 cells per well in 24-well plates, both supplied by Costar (Corning, NY, USA). Cultivation conditions were maintained at 37 °C in an atmosphere of 5% CO2 and 96% relative humidity until the cells reached full confluency, typically within a week. The cells employed in the research were consistently chosen from passages 21 through 27.

Hypoxic challenge and fructose administration

Cells underwent oxygen and glucose deprivation as outlined in previous studies17. The hypoxic (Hx) challenge was implemented using a modular incubator chamber (Billups-Rothenberg, Del Mar, CA, USA) through the infusion of 5% CO2 and 95% N2 at a flow rate of 10 l/min for 5 min; meanwhile, normoxic (Nx) controls were maintained in conditions of 5% CO2 and 95% air.

In subsequent experiments, cells were cultured in a modified DMEM from Invitrogen, which was devoid of glucose and pyruvate but supplemented as previously mentioned, with an addition of 0–25 mM fructose. To elucidate the contribution of glycolysis to death resistance, cells received pre-treatment with metabolic pathway inhibitors such as IA (Iodoacetate, 1 mM, targeting glyceraldehyde-3-phosphate dehydrogenase, GPD, as a glycolytic inhibitor) and UK5099 (UK, 10 mM, acting as a mitochondrial pyruvate carrier inhibitor), or with vehicle controls, prior to the hypoxic challenge accompanied by fructose exposure. All reagents utilized were procured from Sigma-Aldrich (St. Louis, MO, USA).

LDH Leakage Assay

The release of the intracellular enzyme lactate dehydrogenase (LDH) into the extracellular space serves as an indicator of plasma membrane rupture, a characteristic sign of cell necrosis. Following the hypoxic challenge, cell culture supernatant was collected for the assessment of LDH activity. In brief, a reaction mixture containing 0.2 mM NADH and 0.36 mM sodium pyruvate was prepared in Krebs-Henseleit (K-H) buffer, which also included 2% bovine serum albumin. The composition of the K-H buffer is as follows: 118 mM NaCl, 4.8 mM KCl, 1.2 mM MgSO4, 1.25 mM CaCl2, 1.2 mM KH2PO4, and 24 mM NaHCO3, with a pH of 7.4. For the assay, 10 µl of cell supernatant were combined with 190 µl of the reaction mixture in 96-well plates. This setup was then subjected to spectrophotometric kinetic analysis. Due to the distinct absorption spectra of NADH and its oxidized form, NAD+, alterations in NADH levels could be monitored at a wavelength of 340 nm. The reduction in absorbance, recorded every minute for a duration of 10 min, corresponds to LDH activity. LDH activity is defined as the amount required to oxidize 1 mmol of NADH per minute, and the activity levels in the cell supernatant were expressed in Units per liter (U/L)17. Each group consists of 8 independent samples, with each sample tested in duplicate.

Immunoprecipitation of the RIP1–RIP3 Complex Assay

Cell lysates from 3 independent samples were collected and mixed for a single immunoprecipitation reaction. The sample mixture subjected to immunoprecipitation overnight with anti-human RIP1 antibody (BD Biosciences, Franklin Lakes, NJ, USA), then incubated with protein G agarose beads for 1 h at 4 °C, followed by centrifugation. The resulting pellet was resolubilized in electrophoresis sample buffer and subjected to heat denaturation. The immunocomplexes were then analyzed using reducing SDS-PAGE. For immunoblotting, the membranes were probed with anti-RIP1 antibody (1:1000, BD Biosciences) or polyclonal rabbit anti-RIP3 antibody (1:1000, Abcam, Cambridge, UK)17. Each blotting band represents a mixture of 3 independent samples.

Immunofluorescent staining of ZO-1

Immunofluorescent staining was performed to visualize tight junction structures. Cells were exposed to Nx or Hx for 16 h, then fixed with 4% paraformaldehyde on ice for 1 h and quenched with 50 mM NH4Cl in PBS at room temperature for 10 min. After blocking with 0.1% bovine serum albumin in PBS for 1 h, the monolayers were incubated with a polyclonal rabbit anti-human ZO-1 antibody (1:100, Invitrogen) in a permeabilizing buffer containing 0.05% saponin and 0.1% bovine serum albumin in PBS for 1 h. Subsequently, cells were incubated with Alexa 488-conjugated goat anti-rabbit IgG secondary antibodies (1:1000, Invitrogen) for 1 h in the dark, followed by staining with Hoechst dye to visualize cell nuclei. The slides were then mounted with aqueous mounting media and observed under a Zeiss fluorescence microscope.

For quantifying ZO-1 expression signals, the fluorescence intensity was calculated as the average intensity within the region of interest (ROI). Eight independent samples were measured for each group. The quantitative analysis was conducted using MetaXpress (Molecular Devices, CA, USA).

Determination of intracellular pyruvate, ATP, and lactate levels

Commercially available assay kits were employed to determine the levels of intracellular pyruvate (Biovision, Milpitas, CA, USA), ATP (Invitrogen), and lactate (Biovision) in cell lysates. Pyruvate concentrations were measured by oxidizing pyruvate with pyruvate oxidase, leading to a colorimetric reaction, with the absorbance read at 570 nm using a plate reader. ATP levels were assessed using a luciferase-based assay, which requires ATP to emit light, quantified by a luminometer. Lactate was quantified using a chromogenic assay, where lactate dehydrogenase oxidizes cellular lactate, and the resulting color change was measured at an absorbance of 450 nm with a plate reader. Each group consists of 8 independent samples, with each sample tested in duplicate.

RNA extraction and polymerase chain reaction

Total RNA was extracted using Trizol reagent (Invitrogen), with subsequent cDNA synthesis conducted in accordance with the manufacturer’s guidelines. The quantitative assessment of gene expression was performed using the CFX Connect Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA). For the PCR, a reaction mixture was prepared containing 50 ng of cDNA, 10 µl of Power SYBR Green PCR Master Mix (Roche, Basel, Switzerland), and 500 nM of each primer, resulting in a final volume of 20 µl. The human gene-specific primer pairs were employed as detailed. The PCR cycling conditions included an initial denaturation at 95 °C for 3 min, followed by 40 cycles of 95 °C for 3 s for denaturation, and 60 °C for 30 s for annealing and extension. The procedure enabled the calculation of relative gene expression levels. The primer sequences for PKM1 were CAGCCAAAGGGGACTATCCT (forward) and GAGGCTCGCACAAGTTCTTC (reverse), for PKM2 were CTATCCTCTGGAGGCTGTGC (forward) and GTGGGGTCGCTGGTAATG (reverse), for ALDOB (Fructose-bisphosphate aldolase B) were GCAACCCAGGAGGCTTTTAT (forward) and ACCCGTGTGAACATACTGTC (reverse), and for the reference gene GAPDH, TCAAGGCTGAGAACGGGAAG (forward) and CGCCCCACTTGATTTTGGAG (reverse). Each group consists of 6–8 independent samples, with each sample tested in duplicate.

Statistical analysis

Unless otherwise indicated in figure legends, all values include independent measures (biological replicates) indicates in sample numbers (N) and shown as mean ± SEM. Data were compared by one-way ANOVA followed by a Student-Newman-Keul test or ANOVA followed by Fisher’s least significant difference test using Prism (GraphPad Software, San Diego, CA, USA). Statistical significance was set at P < 0.05.

Results

Fructose inhibited hypoxic-induced necroptosis in human colorectal carcinoma cells with a dose dependent manner

Three human colorectal carcinoma cell lines (Caco-2, HT29, and T84) were subjected to a hypoxic environment for 4-24 h. Hypoxia-induced necrosis in these three cell lines was evidenced by a time-dependent increase in lactate dehydrogenase (LDH) leakage (Fig. 1a–c). Apical administration of fructose inhibited LDH leakage in a dose-dependent manner in both Caco-2 and HT-29 cells, but not in T84 cells (Fig. 1d–f). Hypoxia challenge simultaneously induced the formation of the RIP1-3 complex, which is a hallmark of necroptosis in both Caco-2 and HT-29 cells (Fig. 2a, b). Cell necroptosis was accompanied by the disruption of cellular integrity, as indicated by an increase in transepithelial resistance (TER) and structural disruption of ZO-1 (Fig. 3) (with minimal TER changes observed in HT-29 cells, Supplementary Fig. 2). RIP-dependent necroptosis was inhibited by apical fructose in Caco-2 and HT29 cells (Fig. 2). Additionally, apical administration of fructose also ablated hypoxia-induced tight junctional disruption (Fig. 3). In T84 cells, neither glucose nor fructose could reverse the LDH leakage caused by hypoxia (Fig. 1f and Supplementary Fig. 1). Therefore, subsequent experiments will primarily focus on Caco-2 and HT29 cells.

Fig. 1: Dose-dependent mitigation of hypoxia-induced necroptosis by fructose in human colorectal carcinoma cells.
figure 1

Human colorectal carcinoma cells Caco-2, HT29, and T84, after being deprived of sugars (glucose and fructose), were exposed to normoxic (Nx) or hypoxic (Hx) conditions for durations of 4, 8, 12, 16, and 24 h (panels a–c). Additionally, cells were cultured with varying concentrations of fructose (0–25 mM) administered apically and subjected to normoxic (Nx) or hypoxic (Hx) environments for 16 h (panels d–f). Following these treatments, apical culture supernatants were collected, and lactate dehydrogenase (LDH) activity was measured to assess the extent of cell necrosis. Under hypoxic conditions, LDH activity in the culture medium of Caco-2 (a) and HT-29 (b) cells increased over time, but the addition of fructose mitigated the release of LDH induced by hypoxia. However, in T84 cells (c), the addition of fructose did not prevent the release of LDH into the culture medium caused by hypoxia. The introduction of fructose to Caco-2 (d) and HT29 (e) cells reduced the LDH release due to hypoxic stress in a concentration-dependent manner. Nonetheless, a fructose concentration of 25 mM failed to ameliorate the hypoxia-stimulated LDH release in T84 cells (f). (a–c: N = 8 samples/group, *P < 0.05 versus respective Nx groups; (d–f): N = 8 samples/group, *P < 0.05 versus respective Nx groups, #P < 0.05 versus ‘Hx+0’ group).

Full size image
Fig. 2: Fructose negates hypoxia-induced RIP-dependent necroptosis in Caco-2 and HT29 cells.
figure 2

Fructose administered apically inhibited necroptosis induced by hypoxia (Hx) in colonic carcinoma (a) Caco-2 and (b) HT29 cells. Cells were treated with fructose (0 or 25 mM) on the apical side and were subjected to normoxia (Nx) or hypoxia (Hx) to assess resistance to cell necroptosis. Immunoprecipitation blots showing the formation of RIP1–RIP3 complex in hypoxic cells without supplementation (None). No sign of RIP3 signaling was seen in hypoxic cells added 25 mM of fructose (Fru), and normoxic counterparts with or without fructose. Immunoblot lanes represent samples pooled from three wells per condition. The quantification results were derived from the density of two separate RIP3 blots for each group. (N = 6 samples/group, 3 wells of cells pooled together for one Immunoprecipitation assay. *P < 0.05 versus respective Nx groups, #P < 0.05 versus ‘Hx+ 0 mM fructose’ group).

Full size image
Fig. 3: Fructose reverses hypoxia-induced decline in transepithelial resistance and cellular integrity in Caco-2 cells.
figure 3

Supplementation of apical fructose counteracted the compromise in cellular integrity and transepithelial resistance (TER) caused by hypoxia (Hx) in Caco-2 cells. The cells, under normoxia (Nx) and Hx, were treated without or with fructose for TER assessment and to evaluate the expression of zonula occludens-1 (ZO-1). a The TER in Caco-2 cells was reduced by hypoxic exposure. b Fructose mitigated this hypoxia-driven decline in TER in a concentration-dependent manner. c Immunofluorescence images showed the maintenance of tight junction integrity and ZO-1 staining in hypoxic cells treated with fructose. Fructose prevented hypoxia-induced tight junction disruption and cell detachment (indicated by the arrow). d Fluorescence intensity changes of ZO-1 proteins in each group. a: N = 8 samples/group, *P < 0.05 versus respective Nx groups; (b, d: N = 8 samples/group, *P < 0.05 versus respective Nx groups, #P < 0.05 versus ‘Hx+ 0 mM fructose’ group).

Full size image

Anti- necroptosis to hypoxia stress is associated with anaerobic glycolytic metabolism in Caco-2 and HT29 cells after 16 h hypoxia

In the absence of sugars, hypoxic exposure resulted in a decrease in ATP and pyruvate production in Caco-2 (Fig. 4a, b) and HT-29 cells (Fig. 4d, e). Following the administration of apical fructose, there was a significant increase in ATP levels in hypoxic Caco-2 (Fig. 4a) and HT-29 cells (Fig. 4d) in hypoxic cells. A similar pattern was observed in the production of pyruvate; after the addition of apical fructose, the quantity of pyruvate in hypoxic cells was comparable to or exceeded that in normoxic cells (Fig. 4b, e). In terms of lactate production, fructose caused a substantial increase in lactate in hypoxic Caco-2 cells in compare to cell without fructose treatment (Fig. 4c). However, in hypoxic HT-29 cells, fructose did not further induce lactate production (Fig. 4f). To validate the metabolic process involved in resistance to necroptotic death, cells were pre-treated with iodoacetate (IA, a glycolytic inhibitor of glyceraldehyde-3-phosphate dehydrogenase, GPD) and UK5099 (UK, an inhibitor of the mitochondrial pyruvate carrier, MPC) prior to the hypoxic challenge in the presence of fructose. In Caco-2 cells, under normoxic conditions, IA caused partial cell necrosis, while under hypoxic conditions, it significantly inhibited fructose-mediated death resistance (Fig. 4g). Nevertheless, in HT-29 cells, under normoxic conditions, the extent of cell necrosis modulated by IA was comparable to that observed under hypoxic conditions (Fig. 4h). This suggests that HT-29 cells have a higher reliance on the glycolytic pathway than Caco-2 cells, regardless of the presence or absence of oxygen. The administration of UK in both cell lines had no impact on the fructose-mediated death resistance (Fig. 4g, h), suggesting, as in studies with glucose inhibiting cell necroptosis17, that glycolytic pyruvate plays a crucial role in preventing necrosis in colorectal cancer cells.

Fig. 4: Anaerobic glycolytic metabolism contributes to anti-necrotic response in Caco-2 and HT29 cells under hypoxia.
figure 4

The resistance to necrotic cell death in human colonic carcinoma Caco-2 and HT29 cells under hypoxic stress for 16 h was linked to their anaerobic glycolytic activity. These cells, when pre-treated with or without fructose, displayed alterations in the intracellular concentrations of glycolytic end products such as ATP, pyruvate, and lactate (panels a–c for Caco-2 and d–f for HT29). To further elucidate the metabolic pathways conferring necrotic resistance, cells were treated with metabolic inhibitors—iodoacetate (IA, 1 mM; an inhibitor targeting GPD in glycolysis) or UK5099 (UK, 10 mM; an inhibitor of the mitochondrial pyruvate carrier, MPC)—prior to hypoxia exposure in the presence of fructose (panels g for Caco-2 and h for HT29). (a–f: N = 8 samples/group, *P < 0.05 versus respective Nx groups, #P < 0.05 versus ‘Hx + None fructose’ group. g, h: N = 8 samples/group, *P < 0.05 versus respective Nx groups, #P < 0.05 versus veh group).

Full size image

Fructose can affect the expression of Pyruvate kinase (PK)-M1/2 mRNA, but it does not impact the expression of GLUT5 and ALDOB mRNA

Our results indicated the presence of GLUT5 mRNA expression in both Caco-2 and HT29 colorectal cancer cell lines, albeit without significant differences between groups (Fig. 5a, e). ALDOB is capable of catalyzing the reversible cleavage of fructose-1-phosphate (F1P) into dihydroxyacetone phosphate (DHAP) and glyceraldehyde within the glycolysis pathway, thereby facilitating the entry of fructose into glycolysis. Our further analysis of ALDOB expression in hypoxic colorectal cancer cells revealed that there were no significant differences at either the mRNA or protein levels among the groups (Fig. 5b, f). We further explored the role of glycolytic enzymes in fructose-mediated death resistance. PKM, encodes the enzyme pyruvate kinase which catalyzes the final, irreversible step in glycolysis. This step involves the transfer of a phosphate group from phosphoenolpyruvate (PEP) to ADP, generating pyruvate and ATP. PKM has two isoforms, PKM1 and PKM2, which are thought to be involved in regulating the anabolic metabolism and tumor growth of cancer cells. At the mRNA level, we observed that under hypoxic conditions without the addition of fructose, PKM1 expression significantly decreased (Fig. 5c, g). However, with the addition of fructose, PKM1 expression significantly increased, showing the same trend in both cell lines (Fig. 5c, g). Regarding PKM2, its expression also significantly decreased under hypoxic conditions without fructose (Fig. 5d, h). Yet, the addition of fructose led to an increase in PKM2 expression only in HT-29 cells (Fig. 5h), not in Caco-2 cells (Fig. 5d). Despite these trends at the mRNA level, there were no significant differences in protein expression among the groups (Supplementary Fig. 3).

Fig. 5: Fructose can influence the pyruvate kinase (PK)-M1/2 mRNA expression, but it has no effect on the GLUT (glucose transport)-5 and ALDOB (Fructose-bisphosphate aldolase B) mRNA expression.
figure 5

Under hypoxic conditions, fructose treatment of human colonic carcinoma cells had no effect on GLUT5 (panels a and d) and ALDOB (panels b and f) mRNA expression. However, in low oxygen scenarios, fructose induced specific regulatory effects on pyruvate kinase isoforms. Notably, PKM1 mRNA levels were elevated in both cell lines (panels c and g), while an upregulation of PKM2 mRNA (panels d and h) was exclusively seen in HT29 cells. In contrast, ALDOB mRNA expression did not vary among the different experimental groups (panels c and f). (N = 6–8 samples/group, *P < 0.05 versus respective Nx groups, #P < 0.05 versus ‘Hx + None fructose’ group).

Full size image

Discussion

In this study, we demonstrated that the fructose-mediated glycolytic pathway confers resistance to RIP-dependent necroptosis in hypoxic colorectal carcinoma. To the best of our knowledge, this is the first study to provide evidence of anti-necroptotic pathways against hypoxic stress mediated by apical fructose in colorectal cancer cells. Anaerobic glycolysis serves as one of the key strategies for malignant cells to adapt to hypoxic environments and evade cell death. While most research on the Warburg effect in cancer cells has primarily focused on glucose15,16,20, within the digestive tract, the scenario differs significantly. In the alimentary tract, most of the glucose in the lumen can be completely absorbed by the small intestine, whereas fructose is likely to be present in the colon. As an essential nutrient, over 90% of dietary glucose is absorbed by small intestinal epithelial cells via the SGLT (sodium-glucose cotransporter)-1 and GLUT (glucose transporter)-221,22,23. Therefore, dietary glucose is almost entirely absent from the colonic lumen, with colonic cells only able to acquire glucose through systemic circulation. This makes targeting glucose uptake for therapeutic intervention in CRC quite challenging. In contrast, the absorption of fructose in the small intestine is different. The consumption of as little as 5 g of fructose can saturate GLUT59,10,11. When the dietary fructose concentration exceeds the absorption capacity of the small intestine, the excess fructose reaches the colon, leading to gastrointestinal discomfort5,8,9,11,24,25. Fructose, being one of the FODMAP monosaccharides that is poorly digested and can accumulate in the colon, has the potential to significantly affect the gut microbiota and the microenvironment under physiological and pathological conditions7,11. Recent studies have shown that excessive fructose intake is associated with CRC tumorigenesis4,5,12. Therefore, for colorectal cancer, the cancer cells have the opportunity to acquire fructose from lumen side, as well as glucose from the bloodstream. Our current research highlights that apical fructose can inhibit the cell death of colorectal cancer cells induced by hypoxia. Given that fructose is a non-essential nutrient and not a primary energy source for cells, targeting its dietary absorption through restriction or blockade presents a more feasible and non-detrimental approach to life compared to glucose. Despite having comparable chemical structures and equal caloric values, glucose and fructose are processed differently in the liver and in intestinal epithelial cells2,5,26. In line with our findings, recent studies have discovered that an increase in GLUT5 expression in colorectal cancer patient samples and found that the administration of a GLUT5 inhibitor could suppress the proliferation of HT-29 colorectal cancer cells13. In our experiments, GLUT5 mRNA was expressed by both Caco-2 and HT29 cells, yet there were no appreciable variations between the groups. Upon absorption via apical GLUT5 transporters, fructose is enzymatically converted to F1P by fructokinase (also known as ketohexokinase). F1P is subsequently cleaved by ALDOB into DHAP and glyceraldehyde, with DHAP acting as an immediate intermediary in the glycolytic pathway, thereby facilitating the incorporation of fructose into glycolysis. In our research, we found no difference in ALDOB expression across all groups. The specific mechanisms by which fructose supports tumor endurance under hypoxic stress conditions remain inadequately elucidated.

Targeting glucose metabolism by leveraging the differential expression of isoenzymes across the glycolytic pathway is a promising strategy. However, the roles of these enzymes in the glycolytic process involving fructose in cancer cells remain unclear. Among these, the isoforms of PKM have been extensively studied. PKM displays functional diversity through its two isoforms: PKM2 and PKM1. PKM2, with its lower activity, is crucial in cancer metabolism, redirecting glucose metabolites to support cell growth and adapt to the tumor environment. In contrast, PKM1, more active, predominates in high-energy-demand tissues, facilitating efficient ATP production through glycolysis. However, as research progresses, the role of these isoforms in cancer metabolism remains a subject of ongoing debate27. It was previously believed, PKM2 expression and reduced pyruvate kinase activity were thought to promote anabolic metabolism, Warburg effect, tumor growth, and suppress oxidative phosphorylation5,27. However, recent studies have challenged this hypothesis, a newer model suggests that PKM1 expression and increased pyruvate kinase activity also support the Warburg effect and anabolic metabolism, leading to tumor growth without necessarily reducing oxidative phosphorylation28,29,30. In our research, at the mRNA level, we observed in both Caco-2 and HT-29 cells that hypoxia leads to a decrease in the expression of both PKM1 and PKM2. However, upon fructose administration, there was a significant increase in PKM1 expression. In the case of PKM2, fructose administration resulted in an increase in HT-29 cells but had no effect on Caco-2 cells. This suggests that the expression of PKM1/2 in different CRC cells may exhibit distinct regulatory mechanisms.

In colon cancer biology and gastrointestinal research, the human colorectal carcinoma cell lines Caco-2, HT29, and T84 provide a critical understanding of the intestinal epithelium’s cellular and molecular dynamics. Caco-2 and HT29 cells, derived from colon adenocarcinoma, exhibit epithelial morphology and mimic intestinal epithelial cell functions. Caco-2 cells are known for their enterocyte-like differentiation, forming tight junctions and expressing transporters and enzymes typical of intestinal absorptive cells, along with high transepithelial resistance (TER), which speaks to their strong barrier function31. These cells also express absorption and metabolism-related genes, including peptide transporters (PEPT1) and P-glycoprotein (MDR1)32,33,34. HT29 cells exhibit a less differentiated state compared to Caco-2 cells, failing to form a uniform enterocyte-like layer, which is reflected in their lower transepithelial resistance (TER). This characteristic aligns with their less differentiated nature and the inclusion of mucus-secreting goblet cells in the culture35,36,37. Genotypically, both Caco-2 and HT29 cells are KRAS and PTEN wild type but possess TP53 mutations. Additionally, HT29 cells also harbor mutations in BRAF and PIK3CA (PI3K catalytic subunit alpha)38. T84 cells, on the other hand, originate from a metastatic site and show genetic markers linked to cancer metastasis and aggressiveness, with alterations in genes related to cell migration, invasion, and epithelial-mesenchymal transition (EMT). Similar to Caco-2 cells in their ability to form a polarized monolayer, T84 cells are characterized by their resistance to tight junction formation and their notable chloride ion transport, regulated by genes coding for ion channels and transporters39. Our current study has revealed that different cell lines exhibit distinct responses to fructose in mitigating hypoxia-induced cell death. Similar to our prior observations regarding glucose’s ability to inhibit hypoxia-induced necroptosis, we found that fructose administration also curtails necroptosis in Caco-2 and HT-29 cells under hypoxic conditions. However, in T84 cells, neither fructose nor glucose ameliorated the necroptosis caused by hypoxic conditions. The primary distinction of T84 cells is their origin from a lung metastasis of a colon carcinoma, which hints that metastatic colonic carcinoma cells may possess distinct sugar requirements and metabolic pathways compared to their primary colon carcinoma counterparts. This distinction suggests a potential divergence in metabolic demands and responses to stress conditions between cells from primary tumors and those from metastatic sites.　In the field of colorectal cancer cell metabolism, particularly in the context of varying genotypes and their sugar absorption and metabolism, merits further research to fully understand the metabolic distinctions and their implications in cancer progression and treatment.

In conclusion, we found that apical fructose-mediated glycolysis inhibits hypoxia-induced RIP-dependent necroptosis in primary colonic cancer cells. The comprehension of the mechanism by which fructose resists hypoxia necrosis holds potential for identifying new therapeutic targets in the treatment of colorectal cancer.