The crosstalk between fibroblast growth factor 21 (FGF21) system and substance use

Rationale

Substance use disorder (SUD) remains a leading cause of morbidity and mortality and has a devastating impact on the individual and society as a whole.^1,2,3 Of further concern is the paucity of effective treatment options for SUDs, which is in part due to the heterogeneity of responses to treatment. Accordingly, the need to develop new medications for SUDs is critical.³ Addictive substances that have the potential to lead to SUDs and addiction act on both the central nervous system (CNS) as well as the peripheral nervous system (PNS) and peripheral organs (e.g., liver, pancreas, and gut). These substances disrupt the interplay between the CNS and the periphery,⁴ such as the neuroendocrine signaling that mediates feeding and satiety. Neuroendocrine pathways are not only crucial for the regulation of metabolism, energy homeostasis, feeding behavior, and consumption of natural rewards like food but are also involved in the regulation of substance use, and growing evidence suggests their role in the development and maintenance of addictive disorders.^5,6,7,8,9,10 As such, neuroendocrine pathways may be suitable targets to develop novel pharmacotherapeutics of SUDs. The fibroblast growth factor 21 (FGF21) system is one such example that is gaining more traction in the SUD field based on growing evidence from preclinical and clinical research.^11,12,13,14

FGF21 is a 208 amino acid endocrine hormone that is one of many in a family of fibroblast growth factors that have broad mitogenic and cell survival effects.¹⁵ The FGF21 gene is located on chromosome 19 (19q13.33) in humans.¹⁵ FGF21 signals through a co-receptor complex composed of FGF receptor (FGFR subtype 1–4) and β-Klotho receptor (KLB).¹⁶ Once FGF21 is secreted, it binds the FGFR-KLB receptor complex, initiating autophosphorylation of FGF receptors, which triggers subsequent intracellular signaling pathways including mitogen-activated protein kinases (MAPK signaling),^17,18 the mechanistic/mammalian target of rapamycin complex 1 (mTORC1) signaling,¹⁷ AMP-activated protein kinase (AMPK) signaling,¹⁹ and inhibition of the phosphatidylinositol 3-kinase-Akt kinase-mechanistic/mammalian target of rapamycin (PI3K-Akt-mTOR) pathway.²⁰ These signaling pathways in general promote cell proliferation, cell growth, energy metabolism, and cell survival. FGF21 is expressed in hepatocytes (greatest expression), pancreatic beta and alpha cells, skeletal muscle cells, white and brown adipose cells, heart cells, intestinal cells, and neurons in brain regions such as the thalamus, hypothalamus, choroid plexus, putamen, amygdala, and hippocampus.^21,22 However, not all these cell types show constitutive expression under normal physiological conditions.¹⁸ The FGF21 co-receptor complex of FGFR1 and β-Klotho has similar expression patterns as FGF21, with most expression in the liver, pancreas, and adipose tissue, but also some expression throughout the brain (similar to FGF21 expression in brain).^21,22 However, there are several brain regions which express β-Klotho, but not FGF21, and vice versa. FGF21 serves as a metabolic regulator by stimulating hepatic fatty acid oxidation and ketogenesis, as well as suppressing lipogenesis.^15,23 FGF21 has been shown to produce beneficial health effects by maintaining whole-body energy homeostasis by promoting cellular glucose metabolism and reducing inflammation. FGF21 also protects the liver from the detrimental effects of excessive lipid production and storage by promoting the breakdown of lipids stored in the liver.^{15,18,23,24,25}

Due to the known role FGF21 plays in energy homeostasis, it is not surprising that serum FGF21 levels are increased in patients with type 2 diabetes and obesity.²⁶ Daily administration of FGF21 to rhesus monkeys with diabetes led to reductions in fasting serum glucose, triglyceride, and insulin levels, in addition to significant improvements in lipoprotein profiles.²⁷ Therefore, FGF21 holds promise as a therapeutic target for metabolic disorders.²⁸

Growing literature suggests a potential role of FGF21 in SUDs, including alcohol use disorder (AUD). Of note, AUD is associated with a host of metabolic changes, such as damage to pancreatic β cells, diminished insulin sensitivity and altered glucose metabolism, liver damage, and increased triglycerides and cholesterol levels.²⁹ Alcohol produces profound damage to peripheral organs, especially the liver, and to significantly alter immune function.^30,31,32,33 FGF21 may counteract some of those changes.³⁴

To date, there are a few narrative reviews on FGF21 signaling that touch on aspects of alcohol-related outcomes and FGF21 interplay.^12,18,35,36 However, there is no comprehensive scoping review on the interplay between substance-related outcomes and the FGF21 system. Therefore, a comprehensive and unbiased overview of the current literature has the potential to identify novel themes and gaps in the knowledge base that may guide subsequent research.

Objectives

The objective of this scoping review is to comprehensively examine and synthesize the current evidence on the interaction between the FGF21 system and SUD in both human studies and animal models.

Narrative summary of publications

Bi-directional relationship between the FGF21 system and alcohol-related behaviors

Experimental manipulation of the FGF21 system impacts alcohol-related behaviors, and alcohol consumption (or administration) impacts endogenous FGF21 signaling. Several animal models demonstrate that increasing FGF21 signaling through administration of human recombinant FGF21 (hrFGF21)^11,39,44,45 administration of the FGF21 mimetic PF-05231023,¹¹ and genetic overexpression of FGF21⁴⁴ all result in decreased alcohol consumption. Likewise, decreasing or blocking FGF21 signaling through FGF21¹¹ or KLB³⁹ genetic knock out increases alcohol preference/consumption, suggesting that FGF21 signaling and alcohol consumption are inversely associated. Furthermore, alcohol intake or administration consistently increases endogenous FGF21 signaling, both in the blood as well as its hepatic expression, in animal models and humans.^{11,34,46,47,48,49,50} These associations between alcohol and FGF21 suggest that endogenous increases in FGF21 following alcohol exposure promote homeostasis to counteract alcohol-induced biological changes.

Talukdar et al. found that FGF21-transgenic mice (overexpression) exhibited reduced alcohol preference ⁴⁴. The same group showed that neuron-specific knock out of KLB (KLB^Camk2a) produced increased alcohol preference, in a two-bottle choice drinking paradigm, compared to wild-type (WT) mice.³⁹ rFGF21 administration by osmotic minipump suppressed alcohol preference in the control but not in the KLB^Camk2a KO mice. Alcohol metabolism did not differ between KLB^Camk2a KO and control mice. This work suggests that the effect of FGF21 on suppressing alcohol consumption depends on FGF21 signaling in neurons.³⁹

In another comprehensive study, Flippo et al.¹¹ first showed that knockout of liver-specific FGF21 gene (FGF21-LivKO) in C57BL/6J mice blocked the circulating (blood) upregulation of FGF21 due to alcohol administration (3.5 g/kg, i.g.), demonstrating that alcohol-induced increases in FGF21 primarily originate from increased production and secretion of FGF21 in the liver. Researchers next showed that FGF21-LivKO resulted in increased alcohol consumption and preference in a two-bottle choice paradigm. Having demonstrated this bi-directional relationship between liver-produced FGF21 signaling and alcohol consumption, the authors proposed that FGF21 functions as a homeostatic mechanism to decrease alcohol consumption. Congruent findings were observed following intraperitoneal administration of FGF21 or the FGF21 mimetic PF-05231023 (long-acting FGF21 analogue), which both resulted in decreased alcohol consumption in male and female mice using a two-bottle choice paradigm and the Lieber-DeCarli liquid alcohol diet. Administration of intravenous PF-05231023 also reduced alcohol, but not water, consumption in alcohol preferring vervet monkey. Neural mechanisms underlying the effects of FGF21 signaling on reduced alcohol consumption was also investigated in this work. To start, neurons expressing KLB (part of the receptor complex for FGF21 signaling) were fluorescently labeled, and the basolateral amygdala (BLA) demonstrated high expression for KLB. Next, administration of FGF21 or PF-05231023 specifically into the BLA was sufficient to decrease alcohol consumption. In vitro electrophysiological recordings from KLB-expressing neurons in the BLA demonstrated that FGF21 bath application increased measures of neuronal excitability and increased the frequency of action potential firing, demonstrating that FGF21 acts directly on neurons. To determine the projection site of the BLA KLB+ (KLB^BLA) neurons, viral anterograde tracing techniques were used showing that KLB^BLA neurons project to the nucleus accumbens (NAc). Further electrophysiology experiments demonstrated that in vivo systemic administration of FGF21 (1 mg/kg intraperitoneal injections 3 days prior to recording) increased neuronal excitability specifically in the subpopulation of KLB^BLA neurons that also project to the NAc.¹¹ Notably, BLA→NAc circuitry and dopaminergic signaling in the NAc are involved in reward-seeking and consummatory behaviors,^51,52,53,54 which are the result of excitatory projections onto medium spiny neurons (MSNs) in the NAc. Researchers also found that in vivo FGF21 treatment results in increased spontaneous post-synaptic currents (sEPSC) frequency on dopamine D2 receptor (D2)-expressing MSNs, but not D1-expressing MSNs. Downregulation of D2 receptors in the striatum (the NAc is part of the striatum) are associated with SUDs.⁵⁵ As such, the finding that FGF21 impacts KLB^BLA projections to D2 MSNs in the NAc may underlie decreased alcohol consumption by reducing motivation for drinking behavior. Finally, researchers demonstrated that FGF21 signaling in KLB^BLA→NAc neurons is both necessary and sufficient to produce decreased alcohol consumption, highlighting neural circuitry that contributes to the effects of FGF21 on suppression of alcohol consumption.¹¹

FGF21-induced decrease in alcohol consumption may be driven in part by general decreases in reward or reinforcement motivation. Sucrose, like addictive substances, has rewarding properties which rodents and other mammals are highly motivated to consume. Talukdar et al. showed that increasing FGF21 through transgenic overexpression decreased two-bottle choice preference for a sweet solution (saccharin or sucrose) vs. water. The long-acting FGF21 analog PF-05231023 decreased sweet preference in monkeys. Of note, this effect of FGF21 on decreasing sweet preference was shown to be dependent on expression of KLB in neurons. Finally, the researchers showed that FGF21 administration decreased metabolites of dopamine in the NAc in mice.⁴⁴ Furthermore, a Genome-Wide Association Study (GWAS) that was previously described in the text above showed an association between the FGF21 single nucleotide polymorphism (SNP) rs838133 with both candy intake and alcohol consumption in humans.³⁷ Another study showed that FGF21 suppression of sugar intake and sucrose preference was regulated by hypothalamic glutamate neurons that express the KLB and FGFR1 receptors.⁵⁶ This study found that FGF21 signaling on specific glutamatergic neurons in the hypothalamus was both necessary and sufficient to produce the suppression of sucrose intake. Furthermore, they showed that FGF21 signaling directly to dopamine neurons was not required for the suppressive effects on sucrose consumption.⁵⁶ Collectively, these findings suggest that FGF21 may blunt the rewarding or reinforcing effects of rewards in general. However, there seems to be divergent neural pathways that underlie the suppressing effects of FGF21 on alcohol consumption compared to natural reward (sucrose) consumption.

Human studies have shown that acute or binge alcohol intake increases circulating FGF21 levels,^45,47,57 consistent with findings from animal experiments. Stankevic et al. found that FGF21 levels in both systemic and hepatic blood were elevated at 60- and 180-min post alcohol administration (nasogastric tube – 2.5 mL/kg, 40% EtOH, over 30-min). Interestingly, several immune and inflammatory markers, except IL-6, were downregulated at these time points following alcohol administration.⁵⁸ Two studies found increased serum FGF21 levels in participants engaging in binge drinking at Oktoberfest and Roskilde national festivals.^45,57 Desai et al. conducted parallel human and rodent experiments to elucidate the effects of alcohol on endogenous FGF21 levels over time. In humans, 0.9 g/kg binge-intake of alcohol resulted in a steady increase in serum FGF21 levels starting by 2-h post alcohol that peaked at 6-h post alcohol consumption. FGF21 levels returned nearly to vehicle levels by 10-h post intake. In mice, following a 3.5 g/kg alcohol gavage, serum FGF21 levels were elevated by 3-h, peaked at 6-h, and were identical to that of vehicle by 9-h post alcohol gavage. These findings indicate similar effects of alcohol on circulating FGF21 across mice and humans.⁴⁷ In another study, healthy males showed robust increases in FGF21 levels (9-fold) following intragastric or intravenous alcohol administration, with comparable FGF21 pharmacokinetics across the two routes of alcohol administration. These findings suggest that alcohol-induced stimulation of FGF21 is not dependent on the gut.⁵⁹ Ho et al. observed a positive association between plasma FGF21 levels and recent alcohol use in patients with AUD.³⁸ In a follow-up study, they found that plasma FGF21 levels were positively correlated with plasma GABA levels in patients with AUD, and then used iPSC-derived brain organoids to show that FGF21 signaling strongly impacts gene expression related to the GABAergic synaptic pathway.⁶⁰ Another study performed a secondary analysis of alcohol cue-reactivity, alcohol priming, and alcohol self-administration in a bar-like laboratory in non-treatment seeking individuals with AUD. FGF21 levels did not change after cue-reactivity or the priming alcohol drink, but robustly increased after the 2-h alcohol self-administration session and changes in FGF21 levels were trending toward a positive correlation with the amount of alcohol self-administered.⁶¹ Collectively, elevations in endogenous FGF21 following alcohol exposure have been observed across multiple species, including mouse, rat, monkey, and humans.

FGF21 elevations following alcohol consumption may also be involved in counteracting the sedative and intoxicating effects of alcohol.⁶² Choi et al. found that after an intoxicating dose of alcohol (5 g/kg, i.g.), mice lacking the FGF21 gene (global FGF21-KO) required on average 1.5-h more to recover their righting reflex (Loss of Righting Reflex test – LoRR) and spent decreased time on a spinning rotarod compared to WT mice. LoRR is a common method used to assess the sedative and anesthetic effects of a substance and measures the amount of time it takes for an animal that has lost consciousness to orient itself in the upright position. The spinning rotarod evaluates motor coordination and is used as a measure of alcohol intoxication. Alcohol concentrations were measured in the brain, demonstrating no difference between FGF21 KO and WT mice, indicating that differences in LoRR recovery time were not the result of differences in brain alcohol clearance or metabolism. Similar results were observed in liver-specific knockout of FGF21. Next, following a 5 g/kg oral gavage of alcohol, rats were injected with rFGF-21 (i.p.) 1-h later while they were unconscious. rFGF21 administration reduced the time to recover LoRR in both male and female mice – an effect that was dependent on the expression of the KLB gene. Furthermore, rFGF21 administration increased time spent on the spinning rotarod.⁶² As such, FGF21 signaling has bi-directional effects on alcohol sedation/intoxication, similar to alcohol consumption.¹¹ Choi et al. then sought to investigate the neural mechanisms that may be underlying these effects. They first showed that 5 g/kg alcohol gavage resulted in increased expression of c-Fos (marker of neuronal activity) in norepinephrine transporter positive (NET⁺) neurons in the locus coeruleus (LC). This effect was completely absent in the FGF21-KO mice, indicating that alcohol increases the activity of NET+ neurons in the LC through increases in FGF21 signaling. Finally, the authors demonstrate that FGF21 acts directly on norepinephrine neurons in the LC to promote arousal that counteracts the sedative effects of alcohol.⁶² This furthers the idea that FGF21 acts as a homeostatic regulator to the acute effects of alcohol. Interestingly, sensitivity to the sedative effects of alcohol is a predictor of AUD development and associated with other AUD-related outcomes.^{63,64,65,66,67}

Other studies have demonstrated a relationship between FGF21 and alcohol-induced water intake.⁴⁶ Song et al. administered a daily bolus of alcohol (3.5 g/kg, i.g.) over the course of 8 days. Starting on day 3, water consumption began to increase in WT mice treated with alcohol compared to water-treated controls, but this effect was not observed in the FGF21-KO mice. The researchers then used the KLB^Camk2a KO method to knockout KLB in neurons specifically. rFGF21 administration by osmotic minipump increased water consumption in control (KLB^fl/fl) but not in the KLB-KO mice. Researchers then crossed the KLB^fl/fl mice with Sim1-Cre mice, in which the Cre expression was enriched in the paraventricular nucleus of the hypothalamus (PVN). The effects of FGF21 on water intake was lost in the KLB^Sim1 mice. Therefore, they concluded that FGF21 exerts its effects on drinking water in part by acting on Sim1+ neurons in the PVN. As such, FGF21 appears to stimulate the behavioral motivation to consume water following alcohol consumption. This effect of FGF21 on water consumption may play a role in counteracting dehydration due to alcohol consumption.^46,68,69

Existing evidence suggests that increases in FGF21 signaling following alcohol consumption may act as a compensatory mechanism by preventing or decreasing escalated alcohol consumption, attenuating the intoxicating effects of alcohol, and promoting water intake following alcohol consumption. Interestingly, the alcohol-induced increases are largely driven by increased expression of FGF21 in the liver, and the effects of FGF21 on alcohol drinking behavior are dependent on the FGF21 receptor complex expression in neurons in the CNS. Therefore, hepatic increases in FGF21 due to alcohol ingestion may serve to protect the liver from subsequent damage by communicating, through molecular signaling, with the CNS to impact alcohol-seeking behavior and other alcohol-related behaviors.

Effects of FGF21 on alcohol-induced organ (primarily liver) damage

In addition to involvement in alcohol-related behaviors, FGF21 signaling also modulates the effects of alcohol on the liver and hepatocytes, as well as associated damage from oxidative stress and inflammation. The liver shows the highest expression of FGF21 relative to other organs/tissue.⁷⁰ Several studies in humans and animal models point to a link between alcohol-associated liver damage and FGF21 levels, as well as a protective role for FGF21 in alcohol-induced liver damage, oxidative stress, and inflammation.^13,34,48,58 In general, experimental suppression of FGF21 signaling exacerbates the damaging effects of alcohol on the liver, whereas experimental augmentation of FGF21 signaling attenuates these effects.

Christidis et al. showed that patients with alcohol-associated liver disease (ALD) have significant upregulation of serum FGF21 compared to healthy controls.⁴⁸ This study also performed parallel animal experiments. Mice were exposed to a 10-day Lieber-DeCarli liquid alcohol diet followed by a 4 g/kg alcohol gavage, and blood and liver were collected 7–9 h after the gavage. The alcohol-exposed group showed increased hepatic FGF21 mRNA and protein levels compared to isocaloric liquid diet, alcohol-naïve controls.⁴⁸ Liu et al. showed that FGF21 was six times elevated in patients with alcohol-induced hepatic steatosis compared to controls but those with stable cirrhosis (no fatty liver) had similar FGF21 levels compared to controls.³⁴ Wagner-Skacel et al. found higher FGF21 levels in patients that had recently consumed alcohol compared to those that did not recently consume alcohol (verified by urinary ethyl glucuronide (ETG) levels which indicated alcohol consumption in the last 12–72 h). However, no statistical differences in FGF21 levels were found between the three groups tested based on liver pathology (alcohol liver cirrhosis - ALC, non-alcohol liver cirrhosis - NALC, and healthy controls).¹³ In a recent clinical study, Sak et al. measured serum levels of oxidative stress induced growth inhibitor 1 (OSGIN1) and FGF21 in patients with varying degrees of alcohol-related cirrhosis compared to controls. They found that OSGIN1 and FGF21 levels were elevated in people with ALC, and that levels of OSGIN1 correlated with levels of FGF21. Interestingly, OSGIN1 levels were highest in the cirrhosis group at the Child-Pugh C stage, while FGF21 levels were highest in the combined Child-Pugh A/B group compared to Child-Pugh C.⁷¹ Another study that measured serum FGF21 levels in people with alcohol-associated hepatitis vs. people with decompensated alcohol-associated cirrhosis found elevated FGF21 levels in the hepatitis group compared to cirrhosis.⁷² Together, these data indicate that acute alcohol exposure increases FGF21, and that ALD is associated with elevated FGF21 levels, which may be blunted in more severe forms of ALD in which the liver is cirrhotic and reduced overall functioning.

Many studies demonstrate that the damaging effects of alcohol on the liver are enhanced or exacerbated upon attenuation of FGF21 signaling. The Liu et al. study mentioned earlier found that mice fed a standard Lieber-DeCarli alcohol liquid diet for four weeks showed higher serum and liver FGF21 levels compared to isocaloric liquid diet controls (no alcohol). The alcohol-exposed group also showed increased hepatic triglyceride (TG) levels, histological findings consistent with fatty liver, increased plasma concentrations of the liver enzyme ALT, and increased liver/body weight ratio. These effects were exacerbated in FGF21-KO mice compared to WT controls. Furthermore, FGF21-KO mice showed more severe hepatic steatosis as a result of alcohol exposure compared to WT controls.³⁴ Another study showed that FGF21-KO mice exhibited exacerbated effects of alcohol exposure on hepatic steatosis, lipotoxicity and inflammation.⁷³ A study by Xu et al. used mice knocked out for the FGF receptor 1 gene (FGFR1) specifically in adipose cells (fgfr1^adipoQ-cre) and WT controls. Mice were exposed to the Lieber-DeCarli alcohol liquid diet for 3 weeks, and some mice received WY-14,643 (PPAR-α agonist that induces FGF21) treatment. Liver triglyceride accumulation was induced by chronic alcohol to a greater extent in the fgfr1^adipoQ-cre mice compared to WT mice. Furthermore, liver serum triglyceride accumulation caused by alcohol exposure was blunted by WY-14,643 treatment in the fgfr1^adipoQ-cre mice, but not in WT mice. As such, the FGFR1 receptor expression in adipocytes is necessary for FGF21 signaling to induce its protective effects against alcohol-induced TG accumulation in the liver and serum.⁷⁴ A follow up study replicated these effects using a binge-like alcohol drinking model instead of the Lieber-DeCarli alcohol liquid diet.⁷⁵ Desai et al. showed that alcohol exposure also increased liver expression of the inflammatory markers tumor necrosis factor α (TNFα), interleukin 6 (IL6), and monocyte chemoattractant protein 1 (MCP1). This hepatic pro-inflammatory effect of alcohol was exacerbated in the FGF21-KO group compared to WT controls. Alcohol administration (3.5 g/kg, i.g.) increased serum FGF21; fatty acid synthesis genes were upregulated and fatty acid oxidation genes were downregulated. They also showed that chronic alcohol consumption produces significant liver pathology in FGF21-KO mice compared to WT controls, again suggesting that FGF21 protects the liver against long term alcohol-related consequences.⁴⁷ A separate group showed that FGF21-KO mice had significantly lower blood alcohol concentration (BAC) 0.5-h post alcohol administration (5.0 g/kg, i.g.), slower gastric emptying, and higher glucagon-like peptide-1 (GLP-1) levels compared to WT mice,⁷⁶ suggesting that effects of FGF21 on the liver might be related to attenuated alcohol metabolism. However, Desai et al. study found no differences in blood alcohol clearance between WT and FGF21-KO mice,⁴⁷ suggesting that FGF21 may impact the gastric emptying of alcohol without affecting the activity of alcohol metabolizing enzymes. Another study by Zhao et al. used FGF21-KO mice and Lieber-DeCarli alcohol liquid diet (5% w/v, 12-day exposure), combined with a 5.0 g/kg oral alcohol gavage 6-h prior to euthanasia and biospecimen collection. Alcohol exposure resulted in increased expression of FGF21 in blood, liver, and ependymal white adipose tissue (eWAT). Alcohol exposure also resulted in decreased eWAT/body weight ratio, decreased average adipocyte area, increased plasma glycerol and non-esterified fatty acid (NEFA). These effects of alcohol were blunted or completely blocked in FGF21 KO mice, indicating that alcohol-induced lipolysis is dependent on FGF21 expression.⁴⁹ Another study using FGF21-LivKO and the Lieber-DeCarli alcohol diet found that greater alcohol-induced fatty liver in the FGF21-LivKO compared to WT.⁷⁷ In summary, reducing endogenous FGF21 signaling exacerbates the damaging effects of alcohol on the liver independent from and without affecting alcohol metabolism.

Several studies suggest that boosting the FGF21 system attenuates the damaging effects of alcohol on the liver. One study using transgenic overexpression of FGF21 (mostly in liver due to ApoE promoter driving overexpression for FGF21) were resistant to alcohol-induced fatty liver and metabolic disorders, and had reduced plasma alanine aminotransferase levels, disrupted caspase-3-dependent apoptosis, and blunted pro-inflammatory cytokines.⁷³ Zhu et al. administered 5 g/kg alcohol by oral gavage to male mice every day over the course of 6 weeks, which produced considerable fatty liver and liver injury. After the 6-week alcohol exposure, mice were treated with FGF21 (2 mg/kg) or vehicle every day for the next 3 weeks. Mice were then sacrificed after an overnight fast and liver and blood samples were collected. FGF21 treatment reversed several markers indicative of alcohol-induced liver injury. Data from this study also suggested that FGF21 counteracts fatty acid accumulation in the liver by promoting AMPK-SIRT1 signaling pathway.⁷⁸ Another study by Liu et al. found that following alcohol exposure (Lieber-DeCarli alcohol liquid diet for 4 week (5% alcohol)), hepatic expression and circulating FGF21 levels increased, and that rFGF21 administration (4 mg/kg/day, i.p.) significantly attenuated chronic alcohol-induced hepatic fat accumulation and inflammation in WT mice.³⁴ Another study showed that genetic knockout of the Tgr5 gene protected mice from alcohol-induced accumulation of hepatic triglycerides and increased liver and serum FGF21 levels.⁷⁹ A different study showed that liver-specific knock out of the Zinc finger protein 36 like 1 gene (ZFP36L1) in mice protected against alcohol-induced hepatic steatosis, liver injury, and liver inflammation, and resulted in elevated FGF21 mRNA levels in the liver.⁸⁰ In summary, experimental increases in FGF21 signaling attenuate liver damage produced by alcohol.

In addition to the protective effects of FGF21 on the liver, some studies have shown that FGF21 also protects the heart from alcohol-induced damage. One study examined postmortem cardiac samples from people with a history of heavy alcohol use (≥60 g/day for 10 years) with and without cardiomyopathy, and a healthy control group. FGF21 and β-Klotho levels were elevated in the heavy alcohol groups compared to controls, and protein levels were highest in the group with cardiomyopathy. A subsequent experiment using the Lieber-DeCarli alcohol diet in FGF21-KO and WT mice found that FGF21 levels were elevated in blood, liver, and myocardium after alcohol exposure in WT mice. FGF21-KO mice developed greater cardiac hypertrophy, fibrosis and cardiac dysfunction compared to WT.⁸¹

Oxidative stress plays a key role in mediating the in and inflammatory response and liver damage induced by alcohol.⁸² Excessive reactive oxygen species (ROS) disrupt homeostasis, resulting in oxidative stress, which triggers hepatic damage by inducing irreversible alteration of lipids.⁸² Liver disease-induced oxidative stress can damage extra-hepatic organs, and FGF21 may mitigate this damage, as evidenced in mouse models of ALD where FGF21 suppresses. Preclinical studies in mouse models of ALD show that FGF21 suppresses oxidative stress in peripheral organs such as the liver and cardiac tissue.^78,81,83 Li et al. found that after a high-fat diet and one-week binge alcohol gavage, mice lacking cathelicidin-related antimicrobial peptide (CRAMP), an antimicrobial peptide that helps maintain a balanced microbiota, compared to wild-type mice, had elevated FGF21 levels in epididymal white adipose (eWAT) mass and increased serum adiponectin secretion, which has been shown to alleviate hepatic fat deposition and inflammation.⁸⁴ Given that after alcohol exposure, hepatic Camp mRNA significantly increased in WT mice,⁸⁵ which has been associated with increased reactive oxygen species production, and that FGF21 inhibits oxidative stress through cAMP response element binding protein (CREB) activation,⁸⁶ which regulates adiponectin gene expression,⁸⁷ a working hypothesis is that CRAMP deficiency upregulates FGF21/CREB/adiponectin signaling to limit alcohol-induced oxidative stress.

Together, these studies demonstrate that multiple molecular signaling effects of FGF21 counteract alcohol-associated organ damage, e.g., on the liver and heart. FGF21 is upregulated in response to alcohol, which is thought to serve as a protective and homeostatic regulator of alcohol-related damage, including oxidative stress and inflammation. FGF21 emerges as a novel and promising target for the treatment of ALD.

FGF21 and other substances

Although we have focused until now on the potential link between FGF21 and alcohol-related outcomes, initial work suggests that FGF21 is also involved in the effects of other substances. One study showed that cigarette smokers had higher circulating FGF21 levels than non-smokers, in both male and female participants.⁴² The effect of cigarette smoking cessation on circulating FGF21 levels was explored in a clinical study where smokers were treated with either varenicline or a nicotine replacement therapy via a transdermal nicotine patch as part of a 12-week smoking cessation program. There was no significant difference in circulating FGF21 in either the group that quit smoking nor the group that did not.⁸⁸ An in vitro study found that cannabidiol (CBD) enhanced the levels of FGF21 mRNA, and other genes involved in browning of white adipocytes in a mouse cell line of adipocytes.⁸⁹ CBD administration in mouse model of alcohol-induced steatosis led to less oxidative stress,⁹⁰ suggesting that part of the protective and anti-inflammatory effects of CBD may be driven by increased FGF21 expression.⁸⁹ In one study, FGF21 transgenic overexpression resulted in lower morphine preference than WT control mice, demonstrating that FGF21 is involved in preference for morphine reward. This study also found that FGF21 overexpression blunted the magnitude and rate of acute morphine antinociceptive tolerance development, and blunted acute and chronic morphine physical dependence.⁹¹

These studies suggest that FGF21 is not only involved in alcohol seeking and other related outcomes such as liver dysfunction but may also interact with other substances. The relative scarcity of research on FGF21 and substances other than alcohol highlights a need for more research on FGF21 and substances other than alcohol. Of note, the basic function of FGF21 signaling in suppressing consumption of calorie rewards to maintain organismal homeostasis suggests that FGF21 plays a role in satiety mechanisms for other drugs and has potential as a therapeutic for other substances by signaling satiety. This mechanistic overlap between reward consumption and satiety signals for food (a “natural reward”) and substances of addiction (chemical compound(s) with the biochemical properties that happen to act on reward consumption and satiety signaling) has been highlighted in the literature for decades.¹⁰

Summary of evidence

Most publications that met eligibility for inclusion in this scoping review investigated the relationship between the FGF21 system and alcohol. Overall, FGF21 appears to produce compensatory and/or protective mechanisms against the deleterious effects of alcohol and other addictive substances. Alcohol directly acts on the liver to upregulate FGF21 and then to impact different cell types. One effect of elevated FGF21 is decreased motivation for alcohol through signaling in neurons that project to dopamine neurons in the nucleus accumbens (part of the striatum).¹¹ As such, hepatic increases in FGF21 from alcohol signals through neurons to attenuate continued alcohol use that protect the liver from additional damage. In addition, FGF21 protects against fat accumulation, inflammation, and damage in the liver, decreases alcohol-related inflammation, and attenuates alcohol-related oxidative stress. Other effects of FGF21 signaling include counteracting the sedative and intoxicating effects of alcohol, and increasing motivation for alcohol-induced water consumption, the latter being a potential compensatory effect. Preliminary work suggests that FGF21 may also be involved in the effects of other addictive substances, hence posing the question whether FGF21 may play a role not only in AUD but also in SUDs at large.

Limitations of the study

Many animal studies use cell-type specific genetic knockout mouse lines combined with pharmacological experiments to precisely elucidate the relationship and mechanisms between FGF21 signaling and alcohol effects. However, the human studies are limited in number and nature (i.e., mostly association studies), hence clinical studies that manipulate FGF21 signaling and measure the effects of alcohol consumption are needed to establish causality. Although preclinical studies provide an important platform to elucidate the relationship between FGF21 and alcohol-related behaviors, such information still needs to be confirmed in randomized clinical trials in people with AUD.

The small number of peer-reviewed, published articles on FGF21 signaling and other substance limits our ability to draw robust conclusions. There is a need for more studies that investigate how FGF21 affects other substance-related outcomes, including but not limited to, opioids, cannabis, stimulants, and sedatives. Information relevant to how FGF21 analogues perform in comparison or in combination with other potential medications that are known to treat AUD and/or other SUDs is also of interest and remains to be explored.

There are also some notable potential limitations to FGF21 and FGF21 analogs as therapeutics for AUD and SUDs. Mostly, there are known side effects of FGF21 that need to be taken into account. Primarily, FGF21 may produce a stimulatory effect.⁶² Other side effects include bone loss, increased heart rate, and high blood pressure.²⁸ Like any potential pharmacotherapy, the risk and benefits must be weighed, especially when a new pharmacotherapy is tested in people with different medical conditions. These results will only be obtained through well-controlled clinical trials in people with AUD.

Resource availability

Experimental model and study participant details

As a review article, this work did not use study subjects, but summarizes previously published work in human subjects and animal models.

Method details

Additional resources

Our protocol was pre-registered in Open Science Framework (DOI 10.17605/OSF.IO/VH287).

PubMed = 56 citations

(“Fibroblast growth factor 21”[tiab] OR FGF21[tiab]) OR ("fibroblast growth factor 21"[tiab] AND (analog[tiab] OR analogue[tiab])) OR “fibroblast growth factor 21-receptor agonists”[tiab] OR “fibroblast growth factor 21 receptor analogues”[tiab] OR “fibroblast growth factor 21 analog”[tiab] OR BMS986171[tiab] OR AMG876[tiab] OR BFKB8488A[tiab] OR NGM313[tiab] OR AKR-001[tiab] OR LY2405319[tiab] OR BMS986036[tiab] OR PF-05231023[tiab] OR BFKB8488A[tiab]) AND (“alcohol-related disorders”[Mesh] OR “alcohol disorder∗”[tiab] OR alcoholism[Mesh] OR alcoholism[tiab] OR “alcohol drinking”[Mesh] OR “alcohol drinking”[tiab] OR ethanol[Mesh] OR ethanol[tiab] OR “substance-related disorders”[Mesh] OR “substance related disorder∗”[tiab] OR “substance use”[tiab] OR opioids[tiab] OR tobacco[Mesh] OR tobacco[tiab] OR amphetamine[Mesh] OR amphetamine[tiab] OR cannabis[Mesh] OR cannabis[tiab] OR marijuana[tiab] OR cocaine[Mesh] OR cocaine[tiab] OR “crack cocaine”[tiab] OR morphine[Mesh] OR morphine[tiab])

Embase = 110 citations

('fibroblast growth factor 21':ti,ab OR 'fgf21':ti,ab OR 'fibroblast growth factor 21-receptor agonists':ti,ab OR 'fibroblast growth factor 21 receptor analogues':ti,ab OR 'fibroblast growth factor 21 analog':ti,ab OR 'bms986171':ti,ab OR 'amg876':ti,ab OR 'ngm313':ti,ab OR 'akr 001':ti,ab OR 'ly2405319':ti,ab OR 'bms986036':ti,ab OR 'pf 05231023':ti,ab OR 'bfkb8488a':ti,ab) AND ('alcohol-related disorders'/exp OR 'alcohol-related disorders' OR 'alcohol disorders':ti,ab OR 'alcoholism'/exp OR 'alcoholism' OR 'alcoholism':ti,ab OR 'alcohol drinking'/exp OR 'alcohol drinking' OR 'alcohol drinking':ti,ab OR 'ethanol'/exp OR 'ethanol' OR 'ethanol':ti,ab OR 'substance-related disorders'/exp OR 'substance-related disorders' OR 'substance related disorders':ti,ab OR 'substance use':ti,ab OR 'opioids':ti,ab OR 'tobacco'/exp OR 'tobacco' OR 'tobacco':ti,ab OR 'amphetamine'/exp OR 'amphetamine' OR 'amphetamine':ti,ab OR 'cannabis'/exp OR 'cannabis' OR 'cannabis':ti,ab OR 'marijuana':ti,ab OR 'cocaine'/exp OR 'cocaine' OR 'cocaine':ti,ab OR 'crack cocaine':ti,ab OR 'morphine'/exp OR 'morphine' OR 'morphine':ti,ab)

Web of Science = 40 citations

(TS="Fibroblast growth factor 21" OR TS=FGF21 OR TS="fibroblast growth factor 21-receptor agonists" OR TS="fibroblast growth factor 21 receptor analogues" OR TS="fibroblast growth factor 21 analog" OR TS=bms986192 OR TS=amg176 OR TS=BFKB8488A OR TS=nem316 OR TS=bms986012 OR TS=PF-05231023 OR TS=BFKB8488A) AND (TS="alcohol-related disorders" OR TS="alcohol disorders" OR TS=alcoholism OR TS ="alcohol drinking" OR TS="alcohol drinking" OR TS=ethanol OR TS="substance-related disorders" OR TS="substance use" OR TS=opioids OR TS=tobacco OR TS=tobacco OR TS=amphetamine OR TS=cannabis OR TS=marijuana OR TS=cocaine OR TS="crack cocaine" OR TS=morphine)