Most studies investigating the role of diabetes in cancer progression have focused on the signaling pathways and gene expressions altered by high glucose16–18. Here, we explored the mechanism by investigating the role of fructose rather than glucose per se. Under hyperglycemia, gastric cancer cells acquired an increased potential for migration and invasion with EMT, all of which occurred depending on fructose synthesis via the polyol pathway. Under hyperglycemia, the polyol pathway-derived fructose was sufficient to stimulate the KHK-A signaling pathway. In two different animal models mimicking gastric cancer metastasis, we found that metastasis was significantly increased in diabetic mice bearing AKR1B1-overexpressing tumors. Mechanistically, the polyol pathway-derived fructose triggered the nuclear translocation of KHK-A, which phosphorylated YWHAH and repressed the CDH1 gene by recruiting SLUG to its promoter, thereby inducing EMT. Collectively, we propose that the connection between the polyol pathway and KHK-A signaling pathway plays a crucial role in diabetes-induced gastric cancer metastasis. Based on this mechanism, we also suggest that the enzymes AKRIB1 and KHK-A could be potential targets for lowering metastatic risk in patients with gastric cancer. The polyol pathway and KHK-A signaling for gastric cancer metastasis are summarized in Fig. 8.
Gastric cancer is the fifth most common cancer and the third most common cause of cancer related deaths worldwide2. Several clinical studies have recently reported a significant correlation between diabetes and gastric cancer progression11, 13. In addition, a meta-analysis revealed that hyperglycemia correlates with gastric cancer risk (HR, 1.11; 95% CI, 0.98–1.26)30. The 5-year survival rate of gastric cancer patients with diabetes is significantly lower than that of non-diabetic patients31. Although gastric cancer is highly prevalent, the clinical outcomes of patients without metastasis are favorable32. Indeed, gastrectomy is the best way to eradicate gastric cancers, and patients without the stomach can maintain relatively healthy lives with nutritional supplements. Thus, it is important to prevent metastasis in patients with gastric cancer. Our results provide a theoretical basis for strict control of hyperglycemia in cancer patients with diabetes to prevent metastasis.
According to Lauren’s criteria, gastric cancer can be classified into intestinal and diffuse types33. The intestinal type forms localized masses removable surgically, but the diffuse type infiltrates into the surrounding tissues and shows a worse prognosis33. MKN-28 and SNU-638 were used as representative cells for intestinal and diffuse types, respectively. Considering such distinct properties of these cells, we also adopted two xenograft models: intra-splenic implantation of MKN-28 cells and subcutaneous implantation of SNU-638 cells. Interestingly, MKN-28 and SNU-638 metastasized to the liver in different histological patterns. Within mouse livers, MKN-28 formed large nodules, whereas SNU-638 infiltrated into the liver parenchyma without clumping. Even in tumor xenografts, gastric cancer cells may retain growth properties inherited from their origins.
Several studies have reported the effect of hyperglycemia on gene expression. Hyperglycemia induces MMP2 expression in cholangiocarcinoma by activating STAT334, upregulates MMP9 in lung cancer by inducing HMOX135, and upregulates MMP2/9 in breast cancer36. The upregulation of MMPs could be responsible for hyperglycemia-induced cancer metastasis37. In lung cancer cells, hyperglycemia induces TGF-β secretion, which stimulates EMT and cell migration38. Hyperglycemia also triggers the degradation of a p53 activator HIPK239, inhibiting he p53-dependent apoptosis40. HIF1A, which expresses many hypoxia-induced genes, is also upregulated by hyperglycemia and consequently induces VEGF and HMOX1, thereby promoting angiogenesis and tumor growth41. However, to the best of our knowledge, little is known about the signaling pathways that initiate these hyperglycemic effects. Moreover, it remains unclear how cancer cells sense glucose levels. Here, we suggest that polyol pathway-derived fructose stimulates cancer metastasis, and KHK-A appears to act as a fructose sensor.
ALDOB and ALOX12 have been reported to be involved in fructose-induced cancer metastasis27–29. KHK converts fructose to fructose-1-phosphate and ALDOB converts fructose-1-phosphate to glyceraldehyde and dihydroxyacetone phosphate. Colorectal cancer cells undergo metabolic reprogramming after liver metastasis. ALDOB is transcriptionally induced by GATA6 and promotes the fructose metabolism, which provides metastasizing cancer cells with the energy and materials necessary for increased growth28. The lipoxygenase ALOX12 produces 12-HETE, which induces inflammation and promotes cancer progression29, 42. In breast cancer, fructose upregulates ALOX12 and a corresponding increase in 12-HETE, thereby promoting lung metastasis29. In contrast, KHK-A acts as a nuclear protein kinase upon fructose stimulation and represses CDH1, thereby facilitating breast cancer metastasis27. However, our results showed that fructose-induced cell migration and invasion were not attenuated by silencing either ALDOB or ALOX12, indicating that KHK-A primarily contributed to the pro-metastatic effect of fructose.
Several AKR1B1 inhibitors are in clinical trials as therapeutic drugs for diabetic complications43, 44. According to our results, AKR1B1 inhibitors could be potential drugs for preventing cancer metastasis in diabetic patients. If needed, AKR1B1 inhibitors can be co-administered with conventional anticancer drugs in cancer patients with diabetes. In some cases, AKR1B1 inhibitors could be used for dual purposes to inhibit diabetic complications or cancer exacerbation. On the other hand, KHK-A inhibitors could also have potential therapeutic benefits, and may prevent gastric cancer metastasis in diabetic patients. The KHK inhibitor PF-06835919, which is currently under clinical trials as a therapeutic agent for non-alcoholic steatosis and steatohepatitis45, is an emerging agent for the prevention of cancer metastasis. Theoretically, it is plausible to try a combination therapy using AKR1B1 and KHK inhibitors to lower the risk of cancer metastasis in patients with diabetes.
Despite the high homology, KHK-A and KHK-C have different biochemical functions46, 47. Several studies have investigated the KHK-C-driven fructose flux as an underlying mechanism of fructose-induced cancer progression. High fructose levels have been reported to provide fuel and building blocks necessary for cancer growth and metastasis28, 48. However, it should be noted that KHK-A is predominantly expressed in most cancer cells, whereas KHK-C is rarely expressed. Since KHK-A has poor fructose phosphorylation activity, it is expected that fructose metabolism does not profoundly contribute to progression of most cancers lacking KHK-C. However, the endogenous role of KHK-A remains to be uncovered27. A recent study identified the function of KHK-A as a protein kinase. KHK-A enhances nucleic acid synthesis by phosphorylating and activating PRPS1, augmenting cell proliferation49. KHK-A also phosphorylates YWHAH, and consequently suppresses CDH1 expression, thereby promoting EMT and metastasis27. Based on these reports, KHK-A is likely to act as a protein kinase to facilitate cancer growth and metastasis.