FGF21 in Obesity and Cancer: New Insights

Abstract

The endocrine FGF21 was discovered as a novel metabolic regulator in 2005 with new functions bifurcating from the canonic heparin-binding FGFs that directly promote cell proliferation and growth independent of a co-receptor. Early studies have demonstrated that FGF21 is a stress sensor in the liver and possibly, several other endocrine and metabolic tissues. Hepatic FGF21 signals via endocrine routes to quench episodes of metabolic derangements, promoting metabolic homeostasis. The convergence of mouse and human studies shows that FGF21 promotes lipid catabolism, including lipolysis, fatty acid oxidation, mitochondrial oxidative activity, and thermogenic energy dissipation, rather than directly regulating insulin and appetite. The white and brown adipose tissues and, to some extent, the hypothalamus, all of which host a transmembrane receptor binary complex of FGFR1 and co-receptor KLB, are considered the essential tissue and molecular targets of hepatic or pharmacological FGF21. On the other hand, a growing body of work has revealed that pancreatic acinar cells form a constitutive high-production site for FGF21, which then acts in an autocrine or paracrine mode. Beyond regulation of macronutrient metabolism and physiological energy expenditure, FGF21 appears to function in forestalling the development of fatty pancreas, steato-pancreatitis, fatty liver, and steatohepatitis, thereby preventing the development of advanced pathologies such as pancreatic ductal adenocarcinoma or hepatocellular carcinoma. This review is intended to provide updates on these new discoveries that illuminate the protective roles of FGF21-FGFR1-KLB signal pathway in metabolic anomalies-associated severe tissue damage and malignancy, and to inform potential new preventive or therapeutic strategies for obesity-inflicted cancer patients via reducing metabolic risks and inflammation.

1. Introduction to the hallmarks in FGF21 biology

FGF21 was identified in 2000 as a novel FGF family member preferentially expressed in the liver [1]. However, it was unexpected when FGF21 was first revealed, in 2005, as a potent metabolic rather than overt growth regulator [2], even though there had been prior signs, in particular, in an identical cellular mitogenic activity assay, showing that FGF21 is unable to stimulate cell proliferation as observed for most other FGF family members [3]. The proclaimed metabolic activities of FGF21 in this earliest study include stimulating glucose uptake in adipocytes, blocking obesity development in transgenic animal, and pharmacologically curbing hyperglycemia and hyperlipidemia in both ob/ob and db/db mice, but lacking mitogenecity. For the first time, the authors advocated FGF21 as a novel, effective therapeutic strategy for type 2 diabetes. Interestingly, in 2006, a liver-specific FGF21 transgenic mouse model was developed, which consistently showed that excess FGF21 might antagonize the development of chemical-induced liver tumors, which contradicts the popular portrait for a growth factor that should promote cell growth [4]. In 2007, two independent groups revealed an important physiological activity of FGF21 in the liver [5, 6]. Both groups showed that FGF21 is a PPAR alpha-regulated hepatic hormone in response to fasting, starvation, or ketogenic diets and acts to stimulate lipolysis in adipose tissue, promote fatty acid oxidation and ketogenesis in the liver, and clear hepatic and systemic lipids. Around the same time, molecular mechanistic studies identified Klotho beta (KLB), a pseudo member of the transmembrane beta-glycosidase family that includes also the Klotho (Klotho alpha or KL) as a co-receptor for FGF23 and the Klotho gamma and GBA3 of unknown functions, as a specialized obligatory accessory receptor to the IIIC type isoforms of the FGFR family, in particular, the FGFR1, for mediating FGF21 signal [7, 8]. The expression of KLB is spatially confined to only several endocrine and metabolic tissues, including the liver, adipose tissues, pancreas and certain domains of the CNS, setting a tissue-specific nature for the pleiotropic metabolic actions of FGF21 [7, 9]. In brief, it appears that these early studies from 2005 to 2007 laid the foundation for the FGF21 field.

Further mechanistic studies in mouse models identified both the white and brown adipose tissues [10–12] and possibly certain domains of the hypothalamus and hindbrain [13, 14] that express a KLB and FGFR1c binary complex as the predominant tissue and molecular targets of endocrine and pharmacological FGF21. The contribution of the liver as a direct target of endocrine/hepatic FGF21 as claimed by several studies has yet to be clearly dissected from that of these major peripheral targets. Through these inter-tissue and molecular signal pathways, FGF21 exerts a myriad of metabolic functions by promoting the offset of metabolic derangements and thus, the homeostasis of lipid, glucose and energy metabolism. These studies allow a better understanding of the basic molecular mechanism of FGF21 signaling and metabolic effects, providing new critical insights for the development of FGF21 signal-based therapies in terms of both the molecular and tissue targets for a number of metabolic diseases. Consistently, subsequent pharmacological studies with FGF21, FGF21 analogs or FGFR1-KLB agonists confirmed that the activation of FGF21 signal pathways improves common metabolic diseases such as obesity, diabetes, hepatosteatosis, steatohepatitis, hyperlipidemia, and hyperglycemia in mouse models [11, 15–23]. However, it should be noted that the anti-diabetic and glucose-lowering effects of pharmacological FGF21 were not as significant in clinical patients as those in mice to merit an effective clinical management strategy for human patients with diabetes mellitus [24, 25]. Continued efforts elucidated the molecular and atomic structures of the aggregated extracellular soluble domains of the ternary complex of FGF23, KL and FGFR1c, and of the binary complex of KLB and FGF21 C-tail [26, 27]. In brief, upon binding to the cell surface, the C-tails of FGF21 and FGF23 that extend from their barrel-like cores latch onto the pseudo catalytic pockets located at the junctions of the two glycosidase-like domains of KLB and KL, respectively. Meanwhile, a stretch of amino acids on the membrane-proximal domain of KLB and KL protrudes like an arm from their respective cores, which extends and grips onto the extracellular membrane-proximal Ig-like domain III of FGFR1. The installation of these two new structures from both FGF21/23 and KL/KLB in the complex tethers the individual constituents into a stable ternary structure, on which a high order of oligomers capable of transducing FGF21/23 signal into intracellular mediators then occur. The details of these novel complex structures provide new insights into the molecular mechanisms of endocrine FGF signaling and gain new ground towards a platform for the structure-based novel drug design against common metabolic disorders [28].

There has been a growing appreciation of the expanding roles of FGF21 in tissue-specific pathologies associated with metabolic disorders (e.g. obesity) and meta-inflammation. In pursuit of uncovering potential unknown function of FGF21 and developing FGF21 system-based therapies for other possible disease states, several groups recently made seminal discoveries in the roles and pharmacological effects of FGF21 on pathologies of the pancreas and the liver that are in particular pertinent to chronic high fat consumption and obesity [29–35]. These studies essentially credit FGF21 signaling pathways for establishing a tissue-defensive barrier against metabolic abnormalities, meta-inflammation, and the resultant tissue damage and neoplastic transformation, as manifested typically in the spectrum of diseases associated with fatty liver and fatty pancreas. These studies highlight possible preventive roles of FGF21 in linking metabolic anomalies and meta-inflammation to tissue pathogenesis, e.g., tumorigenesis (Fig. 1). Furthermore, at least two studies raised a new, potentially important possibility of FGF21 as a bona fide downstream metabolic target of oncogene or as a corrective metabolic suppressor in the path toward neoplastic lesions or cancers associated with obesity or meta-inflammatory insults, which echoes the revelation of the antagonistic effect of FGF21 overexpression on hepatocarcinogenesis in 2006 as mentioned earlier [32, 35]. This review will focus on this new direction of development in the FGF21 biology. For a more general view of FGF21 on physiological roles, basic molecular mechanisms, and pharmacological effects on other common metabolic diseases, including but not limited to, obesity, diabetes, and hyperlipidemia, readers are encouraged to explore several other excellent reviews [28, 36–40].

Fig. 1. FGF21 activates a tissue-defensive and metabolic tumor-preventive mechanism.

Metabolic abnormality that fuels malignant transformation of cells, Which has been touted as cancer metabolism, is a hallmark of cancer. The underlying mechanisms are multifaceted. Evidently, the metabolic and endocrine tissues/organs are prone to the development of metabolic abnormalities and associated cancers, e.g. under chronic challenge of obesity. Obesity is known to manifest in the liver and the pancreas as non-alcoholic fatty liver disease (NAFLD) and non-alcoholic fatty pancreas disease (NAFPD), respectively. Chronic NAFLD and NAFPD can manifest into inflammatory and fibrotic states, e.g. steatohepatitis (NASH) and steatopancreatitis (NASP), all of which appear to be risk factors for the terminal development of neoplastic lesions as hepatocellular carcinoma (HCC) and pancreatic ductal adenocarcinoma (PDAC). FGF21 is a potent inhibitor of obesity and metaflammation. In the liver, FGF21 is highly induced in response to metabolic and pathogenic perturbations, and acts to quench lipid overload and steatosis, prohibiting the development of steatohepatitis and fibrogenic injury that could lead to the chronic HCC stage. In the pancreas, FGF21 is maintained at a high basal level but is also inducible in response to insults, and acts to dismount lipid excess, thus keeping the development of fatty pancreas and NASP that can lead to PDAC at bay. However, the oncogenically-activated KRAS silences the expression of acinar cell FGF21, allowing the damaging effects of obesity and inflammation to take maximal toll on the pancreas, leading to rapid development of mutant KRAS-mediated invasive PDAC with high penetrance. In this sense, FGF21 acts as an intrinsic tissue defensive barrier against metabolic deterioration, metaflammation, and fibrotic damage, thus suppressing neoplastic transformation. The biphase change in FGF21 levels in the perturbed liver, which recede at the late stage of steatosis when irreversible damage starts, has been demonstrated; however, whether a similar silencing mechanism as dictated by oncogenic KRAS in the pancreas that autonomously decreases FGF21 levels, and thus, dismounting the FGF21-rendered defense barrier in the path toward HCC exists in the liver remains an interesting question. Nevertheless, current evidence supports FGF21 as an important immanent defensive mechanism against cancer metabolism and as a novel adjuvant therapeutic agent for obese patients suffering from further development to HCC or PDAC.

2. FGF21 in preserving pancreatic function

As soon as field studies revealed FGF21 as a potential anti-obesity and anti-diabetic metabolic regulator in animals, its roles and potential effects in different metabolic tissues or organs were placed under spotlight. In as early as 2005–2006, the potential insulinogenic, insulinotropic or insulin-independent glucose-lowering effect of FGF21 was first postulated. The pancreas is both an exocrine and an endocrine organ and is pivotal for digestive juice production and secretion into guts to control nutrition breakdown, as well as for insulin and glucagon production and secretion into circulation to regulate glucose homeostasis. It was actually the first organ together with the liver studied for a potential role or a metabolic effect of FGF21 [2, 41]. FGF21 was shown to protect islet cells from glucolipotoxicity-induced apoptosis, promote beta-cell survival and function without driving proliferation, and increase the number of insulin-positive islets in db/db mice, thereby contributing to the maintenance of glucose homeostasis and the prevention of hyperglycemia. Later studies indicated more of an insulinotropic or insulin-independent effect rather than a direct insulinogenic effect per se, in particular in humans [24], although this point is still under debate.

Consistent with Wente’ s early study, the deficiency of FGF21 in the islets was later reported to cause islet dysfunction-associated insulin resistance and compensatory non-malignant islet hyperplasia in response to the urge of increase in insulin production [42, 43], possibly due to an enhanced de-repression of growth hormone (GH) signaling in islets, as both GH-treated normal mouse islets and FGF21-deficient mouse islets displayed enhanced basal insulin secretion but impaired insulin secretion in response to glucose challenge. These results indicate an important role of FGF21 in maintaining islet function and insulin sensitivity while limiting islet cell growth and thus the size of normal islets and the level of insulin production. To a different dimension of experiments in mice, within the first three days of a syngeneic islet transplantation, which is a critical period for effective islet engraftment, FGF21 treatment was found to significantly suppress islet graft loss, possibly by increasing insulin sensitivity and inhibiting apoptotic death of beta cells [44]. In 2012, KLB was shown to be expressed in some portions of β-cells in the isolated pancreatic islets, suggesting that FGF21 could also directly act on endocrine pancreas [45]. Interestingly, results from quantitative autofluorescence imaging of living pancreatic islets in a microfluidic device under controlled supply of nutrients and pharmaco-agents suggest that the insulin-sensitizing and hyperglycemia-curbing effect of FGF21 is due to an improvement of islet survival as a result of promoting fatty acid oxidation and limiting glucose-stimulated mitochondrial NADH generation, membrane potential, and production of damaging ROS [45, 46]. High glucose or hyperglycemia in type 2 diabetes reduces pancreatic KLB expression, which may result in FGF21 resistance in islets [47]. However, other studies showed that FGF21 appeared to have no direct effect on insulin or glucagon secretion in pancreatic islets, and the glucose-lowering and insulin-sensitizing effects of FGF21 were thus likely associated with its metabolic actions in the liver and adipose tissues [48, 49]. In insulin-deficient mice with diabetogen STZ-induced diabetes, FGF21 still prevented increases in glycaemia concurrently with lipid-lowering effects [50]. There was also a report showing that FGF21 was expressed in portions of both β-cells and α-cells of pancreata, and significantly increased in the compensatory non-malignant hyperplastic α-cells in response to the loss of glucagon receptor (Gcgr^−/−). The Gcgr^−/− mice lacking glucagon action were protected from developing diet-induced obesity and STZ-induced insulin-deficient diabetes with enhanced glucose disposal, due to an insulinomimetic effect of the elevated FGF21 that is additive, but not synergistic, to that of the simultaneously elevated Glucagon-like peptide 1 (GLP-1) [51].

In 2009, a study showed that FGF21 gene expression was responsive to pancreatitis and mechanic or chemical pancreatic injury, while FGF21 protein could directly act on pancreatic acinar cells, suggesting that FGF21 is an immediate response factor in invoking mechanisms for protection of pancreatic acini from pancreatitis and overt damage [52]. By contrast, loss of Fgf21 gene markedly increased intra-acinar triglyceride vacuole accumulation and accentuated edema and necrosis, along with vascularity and fibrosis resulting from chronic pancreatitis in the exocrine pancreas. Pancreatic injury and increased susceptibility to caerulein (a cholecystokinin (CCK) analogue)-induced pancreatitis resulting from MIST1 deficiency was thought to be attributable to the loss of FGF21 expression, depending at least in part on an epigenetic mechanism [53]. These early studies indicate possible roles of FGF21 in protecting pancreatic acini from metabolic insults, pancreatic inflammation, and damage.

In summary, early studies suggest that both the exocrine and endocrine compartments of the pancreas are potential origins and targets of FGF21. The roles of pancreatic action of FGF21 are to protect the exocrine and endocrine functions from glucolipotoxicity-induced deterioration and injury through direct enhancement of fatty acid oxidation and insulin-independent glucose clearance, as well as interplay with other pathways such as the growth hormone and GLP-1 signal pathways to limit beta cell proliferation, islet cell death, hyperglycemia, pancreatic (both the acinus and islet) inflammation, and fibrosis. However, the endocrine effects of hepatic FGF21 on the functions of the pancreas have not been clearly dissected from the local effects of pancreas-derived FGF21.

3. FGF21 in warding off pancreatic meta-inflammation and neoplastic progression

Overweight and obesity are the most prevalent nutritional metabolic disorders, which can manifest in the pancreas as non-alcoholic fatty pancreas disease (NAFPD) [54], ranging from simple pancreato-steatosis, steato-pancreatitis (NASP), and fibrosis to pancreatic cancer development. However, the exact pathogenic mechanism and the impact of NAFPD on clinical practice are still largely unknown. Recurrent or chronic pancreatitis is pathogenic pancreatic inflammation that worsens over time, leading to permanent damage and irreversible pathologies in the pancreas. Epidemiological evidence supports that obesity is among the major modifiable environmental risk factors for chronic pancreatitis while both obesity and chronic pancreatitis are major risk factors for pancreatic cancer, in particular, the most prevalent pancreatic ductal adenocarcinoma (PDAC) [55–57]. Other etiologies of chronic pancreatitis include gallstone formation, alcohol abuse, autoimmune diseases, choledocholithiasis, and genetic variations or defects in cationic trypsinogen (PRSS1), serine protease inhibitor Kazal type 1 (SPINK1), chymotrypsin C (CTRC), and cystic fibrosis transmembrane conductance regulator (CFTR). Since FGF21 is an anti-obesity factor and can potentially act on both exocrine and endocrine compartments of the pancreata under conditions of stress, it is anticipated to impact the obesity sequelae in the pancreas. But how does FGF21 achieve this? Several recent studies have renewed our interest in the roles of FGF21 in pancreatic pathologies associated with NAFPD.

It has been shown in some comparison studies that the pancreas maintains a high basal expression level for FGF21 under normal dietary and physiological conditions, as the liver in certain conditions [29, 32, 42, 58]. It is worthwhile to point out that at least two comparative studies reported the known highest basal level for FGF21 in the pancreas [32, 59]. Furthermore, it is also apparent that FGF21 expression is significantly higher in the acini than in the islets [32, 42]. Like pancreatitis, refeeding and obese conditions further induce while fasting and starvation reduce FGF21 expression in the pancreas [29, 42, 60], which most likely occurs in the acini; however, it has remained an open question why the acinar cell compartment maintains high levels of FGF21. Unlike fasting and starvation, refeeding is accompanied by production and secretion of a mixture of enzymes from pancreatic acinar cells to assist breakdown and uptake of nutrients in the gut. As these enzymes have potentials to cause self-destruction and thus inflammatory damage or even death to pancreatic cells if improperly activated or uncontrolled, as exemplified by hereditary pancreatitis caused by the activating or gain-of-function mutations in PRSS1, SPINK1 and CTRC enzymes, the exocrine process is inducible in nature and is under tight control; however, the underlying mechanism for such a precise control has remained obscure. In 2017, a study found that the induced higher levels of FGF21 in the acini upon refeeding might act as a secretagogue in an autocrine/paracrine mode to stimulate secretion of digestive enzymes into the duodenum [29], suggesting a new regulatory mechanism of pancreatic enzyme secretion in addition to the classic post-prandial regulator CCK and the autonomous vagal nerve system. However, differing from CCK, FGF21 appeared to be blunt for stimulating enzyme production, which is important in preventing excess enzyme/protein accumulation that may cause a proteolytic insult. Deficiency of FGF21 or its co-receptor KLB in acinar cells caused intracellular accumulation of zymogen granules, rendering cells highly susceptible to ER stress that could be reversed by the administration of recombinant FGF21 [29]. These results suggest that high levels of pancreatic FGF21 protect the pancreas by maintaining the acini in proteostasis. Intricately, in a follow-up study, the same group suggested that pancreatitis, which can be a consequence of the unreleased excess enzyme pool, was associated with a loss of FGF21 expression due to the activation of the integrated stress response pathway (ISR) while pharmacological replenishment of FGF21 mitigated ISR, and thus, resolved pancreatitis [31]. Despite of the discrepancy, these studies highlight a critical role of FGF21 in preserving exocrine pancreas function and a therapeutic potential of FGF21 pathway for pancreatitis.

Under inflammation, pancreatic acinar cells may undergo transdifferentiation to an embryonic progenitor phenotype that expresses ductal markers, the process of which is called acinar-to-ductal metaplasia (ADM). After inflammation recedes, the metaplastic cells may resume original acinar cell phenotype to facilitate pancreatic regeneration. Studies have shown that under inflammation, the presence of oncogenic KRAS hijacks the regenerative process and efficiently redirects metaplastic cells toward irreversible ADM, and thus, the development of pancreatic intraepithelial neoplasia (PanIN), a definitive precursor of PDAC. Therefore, acinar cells are considered the major cellular origin of PDAC [61]. Activating mutations in KRAS are prevalent in more than 95% of PDAC [62], which accounts for 90% of all pancreatic cancer cases. Although mutation-activated KRAS is known to drive irreversible ADM and promote cell growth and proliferation as well as metabolic reprogramming, oncogenic KRAS alone induces mostly pancreatic inflammation, fibrosis, and PanINs, but rarely to the stage of full-blown PDAC, for which the underlying mechanism is unclear. A second hit is proposed to be obligatory for the development to the invasive PDAC stage. Studies showed that when fed an obsogenic HFD, treated with pancreatitis-inducing caerulein or upon inflammogen COX-2 overexpression, mice expressing an endogenous level of Kras^G12D/+ in pancreatic acinar cells developed marked inflammation, advanced PanIN lesions, and invasive PDAC leading to lethality with high penetrance [32, 63–65]. However, the molecular underpinnings for the vulnerability of acinar cells to HFD/obesity or inflammation challenge and for the synergistic interaction between obesity/inflammation and oncogenic KRAS in the path toward malignant neoplastic lesions remain a major knowledge gap. A new report demonstrated that the activation of oncogenic KRAS in acinar cells at even the basal levels led to a significant reduction and eventually silencing of the expression of pancreatic Fgf21 gene [32]. In several mouse models, the drastic downregulation could occur immediately following Tamoxifen-induced KRAS^G12D/+ expression, before any signs of tissue structural abnormalities typical of neoplastigenic process, and along the path toward PanINs and PDAC [32]. Similarly, loss of FGF21 is prevalent in pancreata of human pancreatic cancer patients, of whom 95% are predicted to bear KRAS mutations. Therefore, the shutdown of FGF21 gene appears to be an intrinsic function of KRAS oncoprotein irrespective of any adverse liaison challenge. Studies have suggested the possible involvement of PPARs, RORa and ATF3 in the regulation of pancreatic FGF21 expression [31, 32], yet the exact mechanism of negative control of FGF21 expression by oncogenic KRAS requires a dedicated investigation. Unlike FGF21, the expression of the accessory receptor KLB that switches FGFR1c signaling from promoting cell growth to metabolic homeostasis upon FGF21 binding appears to be not under the control of oncogenic KRAS.

The drastic reduction of FGF21 or its signal pathway makes the acini or pancreas vulnerable to a second hit, e.g., a metabolic or inflammatory challenge, which mounts to boost KRAS oncogenicity and to accelerate invasive PDAC development. Consistent with this notion, chronic administration of FGF21, which compensates for the loss of pancreatic FGF21, strikingly improved pancreatic and systemic inflammation, ameliorated pancreatosteatosis and fibrosis, inhibited PanIN occurrence and severity, thus significantly inhibiting malignant progression to PDAC while blocking liver metastasis induced cooperatively by KRAS^G12D and chronic obesogenic HFD. Notably, some of these effects elicited by FGF21 could be observed quantitatively even under normal dietary conditions. Mechanistically, FGF21 was showed to educe a profound inhibition on a spectrum of inflammatory cytokines, chemokines and their receptors while promoting the anti-inflammatory factors in the pancreata, as well as to a lesser extent in the systemic levels including the liver and white adipose tissue. Inhibition of local pancreatic inflammation resulted from chronic consumption of HFD and oncogenic KRAS by the administered FGF21 was thus considered as an important underlying mechanism. This is in accordance with other reports [42, 52, 66] showing that FGF21 reduced the severity of both acute and chronic pancreatitis while improving acinar cell atrophy and stromal fibrosis, at least in part by regulating macrophage polarization [66]. In this sense, acinar cell FGF21 serves as a tissue-residing stress-defensive barrier in the path towards pancreatitis and neoplastic lesions, by normalizing tumorigenic metabolic abnormalities and preventing inflammation and associated damage, while mutant KRAS breaks this previously unknown protective barrier by forcing FGF21 gene into silence, thus contumaciously allowing the needed metabolic damage to proceed in escalation. This explains at least in part why and how both obesity and inflammation serve as major risk factors for invasive PDAC development (Fig. 1). Therefore, pharmacological anti-obesity FGF21 may stand well as an effective intervention strategy for fatty pancreas and pancreatitis (e.g., NASP) and at least as an adjuvant therapy for subsequent development of pancreatic cancer as a manifestation of obesity.

Of note, not all anti-obesity medications, such as Phentermine, Lorcaserin and Orlistat, have been found to associate with such a significant anti-inflammation as well as anti-cancer property as FGF21. It is therefore conducive to hypothesize that there may be other mechanisms underlying such additional properties of FGF21 beyond correction of tumorigenic metabolic abnormalities and metaflammation; which cannot be excluded currently and deserve further investigation. As aforementioned, it has been reported that FGF21 could suppress the expression of some growth factors, such as GH and GH-induced IGF-1, and cell proliferation [43, 67]. It was also showed to inhibit ER stress, ROS stress and mitochondrial dysfunction, as well as inflammation that is not from a direct metabolic cause, which are among the known contributing factors to tumorigenesis [29, 68–72]. Liposarcoma is a rare type of cancer that begins in the fat cells that are the main target of FGF21 action as well as the autocrine/paracrine origin of FGF21. A recent study has shown that liposarcoma patients of high expresser for FGF21 have a better prognosis and a lesser chance of recurrence than those with low levels or no expression of FGF21, who are more likely to relapse and die in a shorter period of time [73]. Hence, liposarcoma may present another model to study the roles or effects of FGF21 and associated pathways. All these results underscore the potential of FGF21 as a metabolic and anti-inflammatory therapy, thus indirectly improving cancer treatment outcome. Interestingly, a number of recent studies corroborate an important role of FGF21 in control of general inflammation or in interplay with classic inflammatory pathways [66, 74–78]. The anti-metaflammatory and anti-fibrotic effects of FGF21 may therefore represent a critical mechanism [31, 32, 34, 42, 53] among those mentioned above. Whether FGF21 can directly intervene inflammation pathways is an important future topic. As the inflammation process also helps to recruit immune cells to fight cancer, whether FGF21 treatment causes any downside or risk in immunity, in particular in association with the tumorigenic process, remains to be determined. In addition, future studies should be directed to further exam these effects and mechanisms by forced inactivation or enhancement of FGF21 signal pathway, e.g. by acinar cell-specific gene deletion or overexpression of KLB or FGF21, on the progression of oncogenic KRAS-mediated PDAC in the context of inflammation, obesity-associated metaflammation, and other tumorigenic insults.

Although systemic injection of FGF21 induced ERK1/2 phosphorylation (pERK 1/2) in only a small subset of islet cells compared to in more than 85% of acinar cells that account for a large portion of the whole pancreas, the effects of FGF21 on islets under obesogenic HFD condition were also obvious [42]. In 2016, a study showed that the pancreata of dietary obese mice with whole-body FGF21 deficiency developed significant islet hyperplasia with periductal lymphocytic inflammation and large inflammatory infiltrates, unlike the pancreata of wildtype mice [42]. Consistent with a previous report [43], the islet hyperplasia appeared to be benign and might have been meant to compensate insulin loss as a result of the development of islet dysfunction and insulin resistance. Nevertheless, these studies again suggest that FGF21 acts to summon a defense against pancreatic inflammation that may cause compensatory islet hyperplasia, and thus, to maintain islet homeostasis.

In brief, the inhibitory impact of FGF21 on pancreatic cancer development or islet benign hyperplasia is likely a result of its activities of preventing obesity, metabolic damage and adverse inflammation, which serve to maintain pancreatic cellular and metabolic homeostasis. Even though it has been showed that FGF21 is under strong negative control of oncogenic KRAS, whether FGF21 involves directly in pancreatic tumorigenic process still remains to be determined.

4. FGF21 in fatty liver disease and associated neoplastic progression

The liver has been shown as the principal organ to command endocrine FGF21 signaling pathways. Under chronic stress conditions, such as obesity, diabetes, and inflammation resulting from diverse types of insults, the liver often suffers from nonalcoholic fatty liver disease (NAFLD) that can progress from simple hepatosteatosis, non-alcoholic steatohepatitis (NASH), fibrosis, and cirrhosis to hepatocellular carcinoma (HCC), all of which resemble the aforementioned spectrum of NAFPD. The expression level of FGF21 in the liver is highly sensitive to insults and can increase markedly in response to metabolic perturbation and tissue damage of multitudinous, disparate etiologies [79, 80]. The liver-secreted FGF21 appears to be the major source of circulation levels that command endocrine effects. Early studies focused on understanding the mechanisms that control hepatocyte FGF21 expression and on characterizing the physiological roles and pharmacological effects of hepatic and pharmacological FGF21 on hepatosteatosis as a result of obesity, diabetes or other hepatic pathologies. Different mouse models with transgenic overexpression or inactivation of FGF21 gene, as well as tissue-specific knockout of FGFR1 or KLB were developed to facilitate the investigations [2, 6, 10, 12, 81, 82]. It should be noted that biochemical studies showed that FGF21 activates the FGFR1-KLB but not FGFR4-KLB complex [7, 83]. Although FGFR1 could be detected in the liver, its abundance is several folds less than that of FGFR4, and the significance of a direct autocrine/paracrine effect of hepatic FGF21 on the liver itself has thus been a matter of debate. The pronounced systemic effects of FGF21, in particular on both white and brown adipose tissues, also contribute to the difficulty in dissecting the hepatic effect from systemic effect of hepatocyte-derived FGF21. As in several other tissues, it is highly likely that under certain conditions the hepatic FGF21 plays direct and significant paracrine/autocrine roles in regulating some of the metabolic functions of the liver in addition to its prominent endocrine roles in the peripheral tissues.

Notwithstanding, a great deal of studies have made clear that hepatic FGF21 is a general stress responsive factor, of which gene expression activation reflects an intrinsic response of functional hepatocytes to a broad spectrum of hepatic, and to some extent, peripheral (through inter-organ cross-talk) pathological changes that unload cellular and metabolic stress on the liver, the metabolic hub of our body. The effects of endocrine FGF21 on the liver appear to be twofold, as inferred from studies with FGF21 deficiency or overexpression under different dietary conditions as well as pharmacological administration at either acute or episodic bolus [5, 6, 19, 30, 33, 49, 69, 84]. Firstly, FGF21 directly promotes fatty acid oxidation while inhibiting lipogenesis and gluconeogenesis in the liver, which at least in part contribute to the effective prevention of hepatosteatosis and NASH in disease models, as exemplified by the ob/ob, db/db, and DIO mice. On the contrary, loss of whole-body FGF21 exacerbates hepatosteatosis and steatohepatitis under different dietary challenges, including HFD, methionine and choline deficient diet (MCD), and Lieber DeCarli liquid diet. Secondly, FGF21 improves hyperlipidemia and hyperglycemia by promoting adipose tissue lipolysis, fatty acid oxidation, futile cycling of lipid energy, mitochondrial activity, energy dissipation as heat, and glucose uptake and utilization, thus indirectly impacting the liver. It also acts on different hypothalamus and hindbrain regions to promote energy expenditure and thermogenesis, as well as aversion of sugar and alcohol intake [17, 85, 86]. Such intense anti-obese, anti-diabetic, and energy-consumptive activities evoked by FGF21 signal appear to be the major contributor of the anti-hepatosteatotic nature of FGF21. In line with these direct observations, many subsequent studies found that the up- or down-regulation of hepatic FGF21 mediates the anti- or pro-hepatosteatotic functions, respectively, of many other regulators or pathways upon perturbation, including but not limited to ATG7 autophagy regulator [87], mTORC1 pathway [88, 89], SIRT1 [90], aryl hydrocarbon receptor [91], CREBH [92], PPAR alpha [93], cytochrome P450 enzyme CYP2A5 [94], gluconeogenic and glycogenolytic dual enzyme glucose-6-phosphatase (G6PC) [95], IDH2 [96], CREB/CRTC2 [97], CCR4-NOT deadenylase complex [98], YIPF6 protein trafficking regulator [99], and JMJD3 histone demethylase [100]. Such a broad scope of functional association with a multitude of functionally distinct enzymes, regulatory proteins or key signal pathways across different hierarchy reflects the critical role of hepatic FGF21 in sensing and defending against hepatosteatosis and hepatic inflammation, and in maintaining lipid homeostasis in the liver.

Obesity and diabetes are risk factors for the development of hepatosteatosis and NASH, all of which are also risk factors for liver tumors, in particular, hepatocellular carcinoma. Elevated levels of both hepatic and serum FGF21 correlate with liver fat content and associate with NAFLD and NASH [101–105]. By contrast, such increased levels of hepatic FGF21 expression drop significantly following further extensive inflammatory damage and are completely abrogated with the appearance of fibrosis, cirrhosis and HCC, largely due to transformative cellularity change to non-hepatic cells [80, 106]. However, whether a similar regulatory mechanism exists in the transiting hepatocytes that may silence FGF21 genes autonomously as oncogenic KRAS does in the pancreatic acinar cells remain to be determined. In as early as 2006, it was demonstrated that overexpression of FGF21 in hepatocytes delayed chemically-induced liver tumors including adenoma and HCC, even though no clear mechanistic explanation was forthcoming at the time. As the defensive roles of FGF21 unfolded, its effects on the more advanced stage of fatty liver disease were being subsequently unveiled in recent several years.

Although loss of FGF21 alone in normal conditions does not induce overt hepatosteatosis, steatohepatitis and liver injury (e.g., fibrosis), it predisposes the liver to develop severe liver pathologies under conditions of STZ-induced diabetes, HFD-induced obesity, tunicamycin-induced hepatic ER stress, and even non-obesogenic MCD-induced nonalcoholic steatohepatitis [107–109]. This phenomenon appears to be mainly due to an excess accumulation of non-activated fatty acids that invites peroxidative lipotoxicity, abundant infiltration of inflammatory cells, and further deterioration of metabolic variables. Conversely, administration of recombinant FGF21 or overexpression of FGF21 gene intercepts steatohepatitis and inflammatory stress damage [84, 107, 108], owing to the augmentation of fatty acid activation and oxidation and amelioration of ER stress. Notably, the majority of mice exposed to a long-term obesogenic diet developed HCC [35], suggesting that FGF21 is critical in limiting hepatic stress and the progression from fatty liver to HCC under chronic adverse conditions. Therefore, pharmacological FGF21 may represent an efficacious preventive or therapeutic strategy against fatty liver disease, thus deterring subsequent progression to liver cancer. Recent evidence from the study of hereditary genetic disease also supports such a protective role of FGF21 in the liver. Loss-of-function mutations in the catalytic subunit of G6PC underlie glycogen storage disease type 1a characterized by impaired glucose homeostasis and long-term risks of developing HCC and hepatocellular adenoma. Mice with a restored G6PC activity at various degrees exhibit beneficiary features of calorie restriction and the absence of liver tumors, which is attributed to, at least in part, significant upregulation of both FGF21 and its co-receptor KLB [95]. Recent clinical trials with FGF21 analogs in human patients with non-alcoholic steatohepatitis or with obesity and diabetes predisposed to fatty liver indicate an excellent efficacy in reducing hepatic fat content and improving markers of metabolism and liver fibrosis without treatment-related adverse effects [34, 110]. These favorable clinical outcomes are consistent with previous reports in obese patients with type 2 diabetes, who showed significant improvements in body weight and dyslipidemia even though the liver parameters were not tested [24, 25].

It should be specially noted that, in accordance with the timeline of disease progression, the 2018 study demonstrated that mice deficient in FGF21 developed excess fatty liver within 16 weeks while on a conventional obesogenic HFD [35]; however, after 52 weeks, the hepatic pathology progressed to significantly worsened fibrosis with 78% of these mice developed HCC, which was in marked contrast to only 6% of wild-type mice having HCC [35]. These results are nearly the mirror image of the development of pancreatic pathologies and PDAC upon endogenous FGF21 loss under similar conditions [32], indicating an important life-long role of FGF21 in limiting the development of metabolic abnormalities and associated organ pathologies that can progress to a long-term neoplastic lesions (Fig. 1).

Nevertheless, studies in models of alcohol consumption and alcoholic fatty liver disease also buttress the tissue-defensive roles of FGF21 in the liver. It was showed that acute, binge or chronic alcohol consumption activates hepatic FGF21 expression and increases its circulation levels [30, 33, 111]. However, FGF21 promotes aversion of alcohol in favor of water without directly affecting the catabolic rate of alcohol [85, 111]. Loss of FGF21 exacerbated chronic alcohol-induced hepatic steatosis and pathologies under normal dietary conditions, and increased mortality under the Lieber-DeCarli diet, a HFD supplemented with ethanol [30, 33], which were accompanied by increase in hepatic lipogenesis and decrease in fatty acid activation and oxidation. These results demonstrate that FGF21 expression is a hepatic intrinsic adaptive response to alcohol-induced lipid dysregulation and liver injury, with an immanent role of protecting against ethanol-induced hepatic damage and pathogenesis. Thus, FGF21 presents a potential therapeutic opportunity for effective prevention and treatment of alcoholic fatty liver disease in addition to the non-alcoholic type, and potentially for the management of alcoholism as well.

5. Conclusion and challenges

Since the discovery, FGF21 has emerged as a sensor of stress of various etiologies and an endocrine regulator of systemic homeostasis of lipid and energy metabolism in contrast to insulin as a direct regulator of glucose metabolism, as supported by voluminous data from many early studies. This is followed by the elucidation of the molecular mechanisms underlying its endocine actions and the potential paracrine/autocrine functions in specific tissues. More importantly, dozens of recent studies shed light on the role of FGF21 as an endogenous regulator of tissue homeostasis by keeping tissue inflammation of both metabolic and non-metabolic origins and subsequent tissue deterioration and damage at bay. As a consequence, FGF21 (indirectly) thwarts neoplastic progression in metabolic and endocrine tissues. In this sense, FGF21 represents a missing link between metabolic abnormalities and neoplastic progression, e.g., oncogenic KRAS-mediated pancreatic tumorigenesis (Fig. 1). However, some questions still remain. Among them, one of the most interesting questions is the generality and the underlying mechanism of FGF21 silencing in such pathologies. Another interesting question is whether and by what mechanism FGF21 prevents tissue inflammation beyond meta-inflammation. Future studies should be directed to determine whether FGF21 alone or in conjunction with other therapies is effective in preventing or treating obesity that as a major risk factor contributes to malignancy of metabolic tissues/organs in obese cancer patients.

Highlights.

• FGF21 is a novel regulator of metabolic homeostasis with potent anti-obesity effect

• Oncogenic KRAS silences FGF21 gene expression to promote obesity-associated pancreatic cancer

• FGF21 loss concurs with the progression of fatty liver disease to hepatocellular carcinoma

• FGF21 signal pathway creates a barrier to metaflammation and tumorigenesis

• FGF21 presents a therapeutic opportunity for cancers associated with metabolic abnormalities.

Abbreviations:

ADM

acinar-to-ductal metaplasia

CCK

cholecystokinin

DIO

diet induced obesity

FGF

fibroblast growth factor

FGFR

fibroblast growth factor receptor

GBA3

glycosidase, beta, acid 3 (cytosolic)

GCGR

glucagon receptor

GH

growth hormone

HCC

hepatocellular carcinoma

HFD

high fat diet

KL

Klotho alpha

KLB

Klotho beta

KRAS

Kirsten rat sarcoma-2 viral (v-Ki-ras2) oncogene homolog

NAFLD

non-alcoholic fatty liver disease

NASH

non-alcoholic steatohepatitis

NAFPD

non-alcoholic fatty pancreas disease

NASP

non-alcoholic steatopancreatitis

PanIN

pancreatic intraepithelial neoplasia

PDAC

pancreatic ductal adenocarcinoma

STZ

Streptozotocin