Link between type 2 diabetes mellitus and hepatocellular carcinoma

Abstract

Type 2 diabetes mellitus (T2DM) and hepatocellular carcinoma (HCC) have a strong bidirectional relationship. T2DM increases the risk of developing HCC, mainly through the nonalcoholic steatohepatitis pathway, but a significant proportion of patients develop HCC without developing cirrhosis. The identification of HCC in T2DM patients is difficult considering the low incidence of HCC and the high prevalence of T2DM. However, considering the alarming increase in the incidence of diabetes mellitus in the global population, effective strategies are urgently needed to identify patients at high risk. Nonetheless, various classes of drugs, such as sodium-glucose cotransporter-2 inhibitors and incretin analogs, may be promising for reducing the risk of nonalcoholic steatohepatitis and HCC development in T2DM patients in the future. In this review, we discuss all these facets of the relationship between HCC and T2DM, and we summarize future directions.

Key Words: Cancer; Hepatocellular carcinoma; Nonalcoholic fatty liver disease; Steatohepatitis; Type 2 diabetes mellitus

Core Tip: Type 2 diabetes mellitus increases the hepatocellular carcinoma (HCC) risk and also impacts the treatment response of HCC. The molecular pathways of diabetes, development of its complications are interconnected with HCC. However, only a subset of HCC cases in type 2 diabetes mellitus are driven by nonalcoholic fatty liver disease, suggesting the involvement of additional pathogenic pathways. Newer generation anti-diabetic and anti-obesity drugs like sodium-glucose cotransporter-2 inhibitors and incretin analogues can hold promises in the future.

INTRODUCTION

Type 2 diabetes mellitus (T2DM) is becoming a major global metabolic problem, which is principally fueled by the increasing trend in obesity. On the other hand, the incidences of various cancers that are related to dysmetabolic states are also increasing. There is an intriguing relationship between T2DM and cancer, especially hepatocellular carcinoma (HCC), the most common primary liver malignancy in individuals with T2DM. This relationship has been known for more than three decades[1], and recent data have indicated further intricacies. The prognosis of HCC is often guarded, and early detection and a multimodal therapeutic approach are necessary. The relationship between T2DM and HCC is bidirectional in that T2DM increases the risk of HCC and HCC often increases the risk of T2DM. The liver plays a central role in glucose metabolism, and the pathophysiological link between T2DM and HCC has already been established. T2DM predisposes patients to nonalcoholic fatty liver disease (NAFLD), and as NAFLD progresses to nonalcoholic steatohepatitis (NASH) and cirrhosis, pathways related to the development of HCC become linked[2]. On the other hand, NAFLD is a high risk factor for the development of T2DM[3].

EPIDEMIOLOGY OF HCC AND T2DM: A STRONG RELATIONSHIP

Recent data support a strong relationship between HCC and T2DM. Among patients with T2DM, approximately 7% may progress to HCC through the NAFLD-NASH-cirrhosis pathway over a period of 6.5 years[3]. A recent review has further established that the risk of HCC in patients with T2DM is almost 2.5 times greater than that in normoglycemic individuals, thus confirming earlier findings[4]. Indeed, T2DM has remained an independent predictor for the development of both decompensated liver disease and HCC in recent studies[5]. A recent study from Italy revealed that the incidence of HCC development was more than three times higher in T2DM patients than in the general population, with the incidence rate being higher than predicted and HCC associated with a significantly increased risk of mortality in T2DM patients[6]. Not only diabetes but also the prediabetes stage is linked to an increased incidence of HCC, suggesting a continuum of the same spectrum[7]. Interestingly, as the duration of diabetes increases, the risk for the development of HCC also increases, and as comorbidities such as dyslipidemia, hypertension, and obesity add up, the risk of HCC further increases[8]. Moreover, diabetes is an independent risk factor for HCC in patients with hepatitis B and C infection-related cirrhosis, even after treatment[9,10]. These data suggest that not only NAFLD-driven HCC development but that of all types of HCC is influenced by the presence of diabetes.

Diabetes is not only a strong risk factor for HCC development but also influences its prognosis. The presence of diabetes is associated with significantly poorer overall survival and recurrence-free survival after surgery in patients with HCC. Diabetes also increases mortality and treatment failure risks in HCC patients[11]. Planning for curative therapy for HCC may be difficult because T2DM and NASH-driven HCC are associated with the advanced stage of HCC. Diabetes even confers poorer outcomes following transarterial chemoembolization in HCC patients[12]. A Japanese study revealed that HCC was the leading cause of cancer-related death in T2DM patients [standardized mortality ratio: 3.57, 95% confidence interval (CI): 2.41-5.10][13]. Similar findings were obtained earlier in larger populations, suggesting that HCC is a major cause of cancer-related mortality in diabetes patients[14].

HCC SCREENING IN T2DM PATIENTS: FINDING THE NEEDLE IN THE HAYSTACK

Thus far, we have discussed the already established epidemiological relationship between HCC and T2DM. However, from a clinical point of view, perhaps the most important question is whether we can predict which T2DM patients will develop HCC. Considering the low prevalence of HCC in T2DM patients, systematic screening is impossible at this time; for this reason, routine surveillance for HCC in T2DM patients without cirrhosis is not suggested in any of the international guidelines[15,16]. However, once cirrhosis develops, screening is recommended by guidelines. Furthermore, since NASH-driven HCC occurs without cirrhosis, a separate strategy is needed for T2DM-related HCC screening. However, this issue has been addressed in few studies. Population-based studies have identified increased age, male gender, hypertension, and microalbuminuria as clinical risk factors for advanced liver disease, including HCC[17], indicating that stricter surveillance is warranted if combinations of these factors are present. In the Edinburgh type 2 diabetes study, T2DM patients without liver disease were followed for 11 years. However, none of the common noninvasive risk scoring systems [fibrosis-4 (FIB-4) index, aspartate aminotransferase-to-platelet ratio index, NAFLD fibrosis score, and fatty liver index] reliably identified those who had developed HCC, clearly indicating the need for better prediction models[18]. The addition of hyaluronic acid to the FIB-4 index might improve the ability of this scoring system to predict those who will develop HCC[19], but again, its clinical utility must be further validated. Artificial intelligence tools can also contribute to the development of HCC risk prediction tools for T2DM patients[20].

T2DM AND HCC PATHOPHYSIOLOGY: THE NAFLD-NASH PATHWAY IS NOT SUFFICIENT

The common understanding of the development of HCC is that T2DM increases the risk of NAFLD; thus, NASH and HCC develop following NASH-driven cirrhosis. Metabolic alterations play crucial roles in HCC development in patients with T2DM. Chronic hyperglycemia and insulin resistance lead to increased de novo lipogenesis, accumulation of toxic lipid intermediates, and mitochondrial dysfunction. These processes result in oxidative stress, endoplasmic reticulum (ER) stress, and chronic inflammation, all of which contribute to hepatocyte injury, DNA damage, and fibrogenesis. Moreover, hyperinsulinemia promotes cell proliferation via the activation of the phosphoinositide 3-kinase (PI3K)/protein kinase B (AKT)/mammalian target of rapamycin (mTOR) and mitogen-activated protein kinase pathways. Collectively, these metabolic disturbances create a protumorigenic hepatic microenvironment that facilitates the progression from NASH to cirrhosis and, ultimately, to HCC (Figure 1). This pathway partially explains the relationship between HCC and T2DM. The reported global prevalence rates of NAFLD, NASH, and advanced fibrosis in T2DM patients in 2018 were 55.5%, 37.3%, and 17.0%, respectively[21]. These estimates are significantly greater than those for the general population. In a recent study of a biopsy-driven diagnosis of liver pathology in 360 patients, the prevalence rates of NASH, advanced fibrosis, and cirrhosis were 58%, 38%, and 10%, respectively, which are again much higher than those reported in earlier meta-analyses, possibly reflecting the difference between noninvasive and invasive diagnostic methods[22]. Nevertheless, the prevalence of advanced fibrosis is quite significant in T2DM patients. Furthermore, even a lower serum alanine transaminase level of 20 IU/mL did not predict the presence of fibrosis, suggesting that alanine transaminase is a poor marker for NASH and advanced fibrosis in T2DM patients. Routine screening with established markers of fibrosis, such as FIB-4, and transient elastography, either alone or in combination, is warranted in diabetes patients. Studies in which magnetic resonance imaging methods were used revealed that the prevalence rates of NAFLD, advanced fibrosis, and cirrhosis were 65%, 14%, and 6%, respectively, again suggesting that the prevalence may be underestimated[23]. Another puzzling fact is that approximately 50% of HCC tumors in NAFLD patients develop without cirrhosis, which is called noncirrhotic HCC[24]. Indeed, the combination of all these factors leads to NAFLD-driven HCC being present in an advanced stage for which curative treatment options are limited.

Figure 1 Metabolic alteration that leads to hepatocellular carcinoma. PI3K: Phosphoinositide 3-kinase; AKT: Protein kinase B; mTOR: Mammalian target of rapamycin; ER: Endoplasmic reticulum; MAPK: Mitogen-activated protein kinase.

SIGNALING INVOLVED IN HCC

Hyperglycemia, hyperinsulinemia, and insulin-like growth factor (IGF) signaling pathways are closely associated with the development of HCC through multiple oncogenic mechanisms, as shown in Figure 2.

Figure 2 Pathways relating diabetes and hepatocellular carcinoma. Interleukin-6 from cancer-associated fibroblast leads to insulin-like growth factor 1 (IGF-1) generation via signal transducers and activators of transcription 3 pathway. The IGF-1 acts on liver cell by autocrine action. Hyperglycemia facilitates acetylation of β-catenin, promoting its nuclear entry and help to form lymphoid enhancer factor 1-β-catenin complexes, displacing transcription factor 7-like 2-corepressor complexes. Thereby Wnt gene transcription enhances. Acid-labile subunit binds peroxisome proliferator-activated receptor gamma and stabilizes it through deubiquitination. Peroxisome proliferator-activated receptor gamma upregulates HMG-CoA synthase 2 that supress epithelial-mesenchymal transition. Mutated p53 increases IGF-1 receptor expression. β-arrestin 2 translocates IGF-1 receptor to endoplasmic reticulum in hepatocellular carcinoma cell. IGF-1 interacts with sarco endoplasmic reticulum Ca²⁺-ATPase which loads Ca into the ER and creates ER stress. Unfolded protein response sets in as an adaptive mechanism to restore ER homeostasis. Unfolded protein response facilitating cell survival. Activation of the protein kinase B/mammalian target of rapamycin pathway, leading to enhanced glycolysis. Mammalian target of rapamycin upregulates glucose-6-phosphate dehydrogenase which convert glucose-6-phosphate to 6 phosphogluconolactone. Pentose phosphate pathway results in DNA synthesis. Metformin activates adenosine monophosphate-activated protein kinase which inhibit mammalian target of rapamycin complex thereby inhibits cell anabolism and stimulatesglycogen synthase kinase 3 beta which inhibit glycolysis. IGF: Insulin-like growth factor; IL: Interleukin; AKT: Protein kinase B; mTOR: Mammalian target of rapamycin; GLUT: Glucose transporter; ATP: Adenosine triphosphate; AMP: Adenosine monophosphate; PPP: Pentose phosphate pathway; SERCA2: Sarco endoplasmic reticulum Ca²⁺-ATPase; ER: Endoplasmic reticulum; UPR: Unfolded protein response; EMT: Epithelial-mesenchymal transition; LEF: Lymphoid enhancer factor; HMGCS2: HMG-CoA synthase 2; STAT3: Signal transducers and activators of transcription 3; G-6-P: Glucose-6-phosphate; CAF: Cancer-associated fibroblast; IGFALS: Insulin-like growth factor acid labile subunit; PPAR-γ: Peroxisome proliferator-activated receptor gamma; TCF7 L2: Transcription factor 7-like 2.

IGF signaling pathway

IGF-1 is a key ligand that is highly expressed in the liver and is mitogenic for various cancer types. Interestingly, studies have shown that HCC patients with higher IGF-1 levels tend to have better survival rates[25]. However, cirrhosis may act as a confounding factor, as advanced liver disease impairs hepatic synthetic function, potentially contributing to poorer survival outcomes. Crosstalk between cancer-associated fibroblasts and tumor cells plays a crucial role in HCC progression. These fibroblasts promote HCC cell proliferation and inhibit apoptosis by secreting interleukin-6, which induces autocrine IGF-1 production in HCC cells[26]. Thus, while systemic IGF-1 Levels in the blood may be low, local IGF-1 Levels remain essential for tumorigenesis. Additionally, the loss of function of certain tumor suppressors can lead to increased IGF-1R expression, particularly in liver cells. Under normal conditions, p53 suppresses IGF-1R transcription, but p53 mutations, which are more common in the liver than in other organs, disrupt this regulation, resulting in frequent dysregulation of IGF/IGF-1R signaling[27].

IGF-1R is internalized via clathrin-coated vesicles, which leads to its degradation through two primary mechanisms, one involving a tyrosine-based motif and the other relying on ubiquitin-based signaling with β-arrestins. The fate of IGF-1R is determined by specific β-arrestin isoforms, with β-arrestin 2 mediating IGF-1R translocation into the ER, a process unique to cancer cells and absent in normal liver cells. In HCC cells, IGF-1R interacts with sarco endoplasmic reticulum Ca²⁺-ATPase, a calcium pump responsible for transferring calcium from the cytoplasm into the ER, thereby increasing ER calcium levels (Ca²⁺ ER). This calcium imbalance induces ER stress, activating the unfolded protein response as an adaptive mechanism to restore ER homeostasis[28]. Consequently, the protein kinase R-like ER kinase/eukaryotic translation initiation factor 2 alpha/activating transcription factor 4 signaling pathway is triggered and facilitates cell survival under these conditions.

Insulin signaling pathway

Alternative splicing of the insulin receptor (IR) α subunits results in two isoforms: IR-A and IR-B. IR-A is the predominant isoform in fetal tissues and various cancers, including breast, hepatocellular, and thyroid cancer, whereas IR-B is expressed primarily in differentiated insulin target cells. Compared with IR-B, IR-A is associated with stronger proliferative responses to insulin and IGF-I. As a result, disruptions in the IR-A/IR-B balance may influence the development of HCC[29]. Additionally, IR activation can trigger the unfolded protein response to mitigate ER stress, contributing to the inherent multidrug resistance of HCC. Insulin receptor substrate-1 activity is also increased as a result of increased insulin receptor signaling. Notably, insulin receptor substrate-1 is overexpressed in 90% of HCC tumors and is strongly correlated with tumor growth.

IGF-II

Excessive IGF-II production leads to the activation of IGF-IR, driving proliferation, whereas IGF-IIR counteracts this process by binding to IGF-II. Downregulation of IGF-IIR expression increases IGF-II levels, promoting HCC tumor growth[30].

Inflammation and the nuclear factor kappa B pathway

Obesity-induced adipose tissue releases adipokines and proinflammatory cytokines, causing chronic inflammation. Lipotoxicity-induced hepatocyte death activates Kupffer cells to secrete interleukin-6 and tumor necrosis factor[31], which promote tumor growth via signal transducers and activators of transcription 3 signaling.

Wnt/β-catenin signaling

Wnt/β-catenin signaling is frequently triggered in HCC[32,33]. High glucose levels promote cancer-associated Wnt/β-catenin signaling[34]. In the absence of high glucose levels, Wnt signals alone can mobilize β-catenin and lead to its accumulation in the cytosol, but nuclear accumulation does not occur, and there is no induction of Wnt target gene expression. Hyperglycemia facilitates the acetylation of β-catenin, promoting its nuclear entry. Additionally, high glucose levels facilitate the formation of lymphoid enhancer factor 1-β (LEF1-β)-catenin complexes, displacing transcription factor 7-like 2 (TCF7 L2)-corepressor complexes and inhibiting SIRT1, which is essential for the nuclear accumulation of the LEF1-β-catenin complex.

PI3K/AKT signaling

Oncogenic signal transduction pathways, including the PI3K, AKT, and mTOR pathways, strengthen the Warburg effect in tumors. HCC is frequently associated with the activation of the AKT/mTOR pathway, leading to increased glycolytic activity. Glucose is catabolized by two parallel metabolic pathways, i.e., glycolysis and the pentose phosphate pathway. Glucose flux through the glycolytic pathway is shifted to the pentose phosphate pathway, which leads to DNA synthesis and cell proliferation[35].

Peroxisome proliferator-activated receptors

The liver is the primary site of acid-labile subunit (ALS) expression, where it plays a key role in stabilizing circulating IGFs by forming ternary complexes. However, ALS functions extend beyond IGF regulation. ALS binds to peroxisome proliferator-activated receptor gamma (PPARγ) and stabilizes it through deubiquitination. This ALS-PPARγ interaction upregulates the expression of HMG-CoA synthase 2, promoting ketogenesis and inhibiting HCC by suppressing epithelial-mesenchymal transition[36].

Gut microbiota

Emerging evidence suggests that alterations in the gut microbiota play a role in the development of T2DM and NAFLD. Gut dysbiosis leads to the production of toxic secondary bile acids and bacterial metabolites, which activate toll-like receptors, triggering fibrosis, inflammation, and cancer progression through hepatic stellate cells. Studies of advanced HCC have revealed a decrease in Bacteroidetes abundance and increases in Proteobacteria and Fusobacteria abundances, highlighting the potential link between the gut microbiota and hepatocarcinogenesis[37].

Genetic risk factors

A recent bioinformatics study identified several genes as links between HCC and T2DM[38]. Furthermore, lipotoxicity often contributes to the development of noncirrhotic HCC. Recent studies have indicated that the polygenic risk score (PRS) related to lipid metabolism [patatin-like phospholipase domain-containing protein 3, transmembrane 6 superfamily member 2, and membrane-bound O-acyltransferase domain-containing 7 (MBOAT7)] can identify individuals at high risk for HCC, particularly those with T2DM[39]. TCF7 L2 mutations are considered the strongest risk factors for type 2 diabetes. Recent studies have also shown that TCF7 L2 mutations can affect the progression of cancer via the Wnt/β-catenin pathway, which is one of the most frequently mutated pathways in HCC[40]. The PRS could identify HCC in both patients with and without cirrhosis and had a very good specificity but a low sensitivity in both groups. Certainly, the PRS might be a promising new tool for identifying HCC cases related to dysmetabolic states in the future.

BIOMARKERS FOR DIABETES-RELATED HCC

In NAFLD-associated HCC, alpha-fetoprotein (AFP) remains the only biomarker that is currently recommended for surveillance. However, its sensitivity is only 51.9%, and its levels are often lower in NAFLD-related HCC than in viral-related HCC. AFP also lacks specificity, as it can be elevated in chronic liver disease, pregnancy, and malignancies other than HCC. AFP-L3 is a glycosylated isoform of AFP. It is typically measured as a percentage of total AFP, with a cutoff value of > 10%, indicating enhanced sensitivity over AFP alone. Des-gamma-carboxy-prothrombin, an abnormal prothrombin protein that lacks γ-carboxy residues, is associated with a high tumor burden. Given the limitations of individual markers, combining AFP, AFP-L3, and Des-gamma-carboxy-prothrombin can improve the detection of HCC. The GALAD score, which integrates these three markers with age and sex, has demonstrated superior diagnostic performance, particularly in NAFLD-related HCC.

Several emerging biomarkers, such as Golgi protein 73, glypican-3, high mobility group box 3, dickkopf related protein 1, and spalt-like transcription factor 4, are also under investigation. Additionally, microRNAs such as miR-182, miR-301a, and miR-373 are significantly elevated in patients with NASH-associated HCC compared with those with NASH alone, suggesting their utility as potential early detection markers for HCC[41].

ANTIDIABETIC DRUGS AND HCC

As described previously, insulin is a potent mitogen, and its use has been associated with an increased risk of HCC [odds ratio (OR) = 3.73, 95%CI: 2.52-5.51]. Similarly, insulin secretagogues, such as sulfonylureas, carry a moderate risk of HCC (OR = 1.39, 95%CI: 0.98-1.99), which is further increased with their prolonged use[42]. In contrast, metformin has demonstrated protective effects against HCC (hazard ratio = 0.48)[43]. It inhibits liver cancer cell growth through adenosine monophosphate-activated protein kinase, which suppresses the mTOR pathway and serves as a key regulator of the cell cycle[44]. Additionally, metformin interferes with the energy metabolism of liver cancer cells by activating glycogen synthase kinase 3β. Thiazolidinediones activate PPARγ, which inhibits tumor cell division, inducing G2/M phase arrest. TZD use was associated with overall reduced HCC risk (OR = 0.92), with a more pronounced effect in Asian populations[45]. Dipeptidyl peptidase 4 inhibitors also reduce HCC development by activating lymphocyte chemotaxis in mice[46]. However, alpha-glucosidase inhibitor use was associated with modestly increased HCC risk[45].

Studies suggest that early glucagon-like peptide-1 receptor agonist (GLP-1RA) use may delay liver disease progression. A Swedish study reported a 49% reduction in major liver events over 10 years among adherent users. Preclinical studies, especially those with liraglutide, have revealed reduced liver fat, enzymes, and oxidative stress, potentially preventing HCC by inhibiting hepatocyte apoptosis. A systematic review by Shabil et al[47] also revealed a protective effect of GLP-1RAs against HCC. Compared with bariatric surgery, weight loss is important for reducing the risk of HCC. However, GLP-1RAs may offer benefits beyond weight loss, particularly through improved glycemic control, which helps mitigate hepatic inflammation and fibrosis. Further research is needed to clarify the relative contributions of weight loss, glycemic control, and incretin effects to HCC prevention[47]. Sodium-glucose cotransporter-2 (SGLT2) inhibitors may prevent NASH progression and HCC development by inducing cell cycle arrest, promoting apoptosis, and directly inhibiting SGLT2 in tumor cells. Multiomic and metabolomic analyses have shown that canagliflozin alters mitochondrial oxidative phosphorylation and purine/pyrimidine metabolism, leading to metabolic reprogramming and reduced proliferation of HCC cell lines. In experimental NASH mouse models, canagliflozin was shown to suppress hepatic fibrosis and HCC development[48]. In a xenograft model, Kaji et al[49] demonstrated that canagliflozin directly inhibited the growth of SGLT2-expressing liver cancer via the suppression of glycolytic metabolism. Hendryx et al[50] analyzed National Surveillance, Epidemiology and End Results-Medicare data from 3185 HCC patients and reported that the initiation of SGLT2 inhibitors was associated with a 14%-60% reduction in mortality risk[50].

CONCLUSION

HCC and T2DM are strongly related, and diabetes portends the risk of an unfavorable outcome for patients with HCC. The pathophysiology of this relationship is partially explained by the NASH/liver fibrosis pathway, but a significant number of T2DM patients develop HCC without developing cirrhosis. This peculiarity makes screening programs for HCC difficult, and until further biomarkers or combinations of biomarkers are discovered and validated, simple steatosis markers, such as FIB-4, can continue to guide the identification of high-risk patients, whereas genetic polygenic risk scoring holds promise in the future. Newer antidiabetic drugs, such as SGLT2 inhibitors or incretin analogs, have the potential to modify the disease course and may decrease HCC rates, although the evidence is limited to that from small trials.