Liraglutide Impacts Iron Homeostasis in a Murine Model of Hereditary Hemochromatosis

Abstract

Classic hereditary hemochromatosis (HH) is an autosomal recessive iron-overload disorder resulting from loss-of-function mutations of the HFE gene. Patients with HH exhibit excessive hepatic iron accumulation that predisposes these patients to liver disease, including the risk for developing liver cancer. Chronic iron overload also poses a risk for the development of metabolic disorders such as obesity, type 2 diabetes, and insulin resistance. We hypothesized that liraglutide, GLP1 receptor agonist, alters iron metabolism while also reducing body weight and glucose tolerance in a mouse model of HH (global HFE knockout, HFE KO) and diet-induced obesity and glucose intolerance. The total body HFE KO and wild-type control mice were fed high-fat diet for 8 weeks. Mice were subdivided into liraglutide and vehicle-treated groups and received daily subcutaneous administration of the respective treatment once daily for 18 weeks. Liraglutide improved glucose tolerance and hepatic lipid markers and reduced body weight in a mouse model of HH, the HFE KO mouse, similar to wild-type controls. Importantly, our data show that liraglutide alters iron metabolism in HFE KO mice, leading to decreased circulating and stored iron levels in HFE KO mice. These observations highlight the potential that GLP1 receptor agonist could be used to reduce iron overload in addition to reducing body weight and improving glucose regulation in HH patients.

Iron homeostasis is maintained by a robust and complex network of proteins that act both in individual cells and enable coordination between organs. Iron is absorbed from the diet through apical transporters on duodenal enterocytes and then released from the basolateral surface through the only known mammalian iron exporter, ferroportin (Fpn) (1, 2). Iron is found in 3 main states in serum: iron bound to transferrin, heme bound to hemopexin, or hemoglobin bound to haptoglobin (1, 2). Iron is an essential cofactor for enzymes present in all mammalian tissues, but the largest destination of iron in humans is heme bound to hemoglobin within erythrocytes. Iron metabolism is regulated through the protein hepcidin, a hormone that is primarily synthesized and secreted by hepatocytes (3). Hepcidin suppresses iron flux by binding to ferroportin and decreasing ferroportin expression and activity on enterocytes, hepatocytes, and macrophages (3, 4). Therefore, hepcidin is a key gatekeeper of systemic and tissue iron load by regulating the absorption of iron from the intestine, the release of stored iron from the hepatocytes, and the recycling of iron by macrophages. Hepcidin synthesis and excretion is stimulated by high iron levels (3). One regulator of hepcidin expression is the protein HFE, a major histocompatibility complex class I-like protein that is expressed in various tissues (5). Mutations in HFE lead to decreased serum hepcidin levels and iron accumulation, causing hereditary hemochromatosis (6, 7).

Classic hereditary hemochromatosis (HH) is an autosomal recessive iron-overload disorder associated with a mutation of the HFE gene. Approximately 85% to 90% of patients who have inherited forms of iron overload are homozygous for the C282Y mutation in HFE, with a small minority who are compound heterozygotes (8). HH is diagnosed by increased transferrin saturation (<45%) and elevated ferritin levels (8). Although iron is an essential mineral for health, if HH is left untreated, severe iron overload in liver, heart, and pancreas results in potentially fatal organ damage. Patients with HH exhibit excessive iron accumulation in the liver that predisposes them to hepatic liver disease, including the risk of developing liver cancer (9-12). Chronic iron overload also poses a risk for the development of metabolic disorders such as obesity, type 2 diabetes, and insulin resistance (13-19). Serial therapeutic phlebotomy, the standard of care for HH, is usually effective in reducing stores of both plasma iron and tissue iron and can prevent the progression of liver disease in these patients (20).

Therapeutic phlebotomy is effective in reducing stores of both plasma iron and tissue iron in HH patients and generally poses no risk of anemia. However, some patients are intolerant or have low acceptance of and compliance with life-long treatment and therefore are good candidates for new therapies for restoration of iron levels. Liraglutide is a GLP1 receptor agonist (GLP1RA) used clinically for the management of type 2 diabetes and obesity (21, 22). Liraglutide is a derivative of GLP-1 and shares 97% amino acid sequence homology with its parent molecule (23). This peptide-based drug has numerous effects all based on the physiology of GLP-1. It improves glucose control through glucose-dependent insulin secretion and glucagon lowering as well as reduced gastric emptying; it lowers body weight through reduction of hunger and increased satiety and control of eating, leading to lowered energy intake (23). Liraglutide also has cardiovascular benefits (24). Specifically related to this work, liraglutide has potent effects on obesity-related liver pathology, improving liver enzyme levels and decreasing hepatic steatosis and fibrosis (25-27). Aberrations in iron metabolism have been proposed to be involved in the pathogenesis of obesity-related metabolic disease including in patients with type 2 diabetes mellitus without HH (3, 4, 14, 17, 19, 20). An intriguing clinical case report demonstrated the successful use of liraglutide to lower glycated hemoglobin and maintain normal ferritin levels without phlebotomy in a patient with diabetes mellitus and previously phlebotomy-dependent hereditary hemochromatosis (28, 29). A recent study in a genetic mouse model of diabetes noted reduced hepatic iron levels after liraglutide treatment (30). We hypothesized that liraglutide may alter iron metabolism while also reducing body weight and improving glucose tolerance in a mouse model combining HH and diet-induced obesity (DIO) and glucose intolerance. Mice deficient in Hfe (HFE knockout [KO]) recapitulate important aspects of the human HH phenotype, most notably iron accumulation in the liver (31, 32). In this study, whole-body HFE KO mice and their littermate (wild-type [WT]) controls were fed a high-fat diet to induce obesity. These mice were then dosed with either liraglutide or vehicle (saline). Our data demonstrate that liraglutide improves obesity-related metabolic disease, including glucose tolerance, hepatic lipid markers, and body weight in HFE KO mice just as in WT controls. Importantly, however, our data show that liraglutide alters iron metabolism in HFE KO mice. Liraglutide treatment of HFE KO mice for 18 weeks decreased serum and hepatic iron levels. Together, these data show that liraglutide alleviates iron overload in a mouse model of HH.

Materials and Methods

Study Approval

All studies were approved by the University of Michigan Institutional Animal Care and Use Committee (Ann Arbor, MI) and were in accordance with National Institute of Health guidelines.

Animals and Diet

Total body HFE KO mice were purchased from The Jackson Laboratory (B6. 129S6-Hfe^tm2Nca/J; strain # 017784). Male mice HFE KO and their WT littermates were maintained on regular chow diet until 2 to 3 months of age. All mice were then fed 60% HFD from Research Diets, Inc. (New Jersey, US; Catalog D12492) for the duration of the study (26 weeks). Mice were fasted for 4 hours and humanely euthanized using CO₂. Tissues were snap frozen in liquid nitrogen. Whole blood and serum were collected and analyzed for hematologic parameters.

Dosing Protocol

Male diet-induced obese mice were randomized by weight per genotype and assigned to receive either liraglutide or saline vehicle for 18 weeks. Mice were injected once daily with liraglutide or saline as follows and as we have previously reported (33): 100 μg/kg on day 0, 300 μg/kg on day 1, 600 μg/kg on day 2, and 1000 μg/kg on days 3 to end of study (18 weeks). All injections were given subcutaneously.

Intraperitoneal glucose tolerance test was performed on week 14 of dosing. Briefly, mice were fasted for 4 hours. Glucose (2 g/kg) was administered IP and circulating glucose levels were measured at baseline, and 30, 60, and 120 minutes using the Biosen Glucose Analyzer (EKF Diagnostics).

Body Composition Analysis

Mice underwent body composition analysis to measure lean and fat mass using EchoMRI (EchoMRI, LLC) at 3 time points: before placement on high-fat diet, following 8 weeks of high-fat diet and before initiation of liraglutide/saline dosing, and 18 weeks after initiation of liraglutide/saline.

Hematological Analysis

Peripheral blood from the animals was subjected to complete blood cell count analysis. Blood was collected in EDTA-coated Greiner Bio-One MiniCollect Capillary Blood Collection System Tubes. Hemoglobin concentration, hematocrit, erythrocyte count, and other indices (mean corpuscular volume, mean corpuscular hemoglobin [MCH], mean corpuscular hemoglobin concentration, and reticulocytes) were analyzed by the University of Michigan In Vivo Animal Core.

Iron Analysis

Peripheral blood from the animals was collected at the end of studies by cardiac puncture, and serum was separated using Serum Greiner Bio-One MiniCollect Capillary Blood Collection System Tubes. Serum iron, total iron-binding capacity (TIBC), and ferritin saturation was analyzed by the Cornell University College of Veterinary Medicine Animal Health Diagnostic Center, Ithaca, NY. Transferrin saturation was calculated by dividing serum iron levels by TIBC. Iron levels in liver, duodenum, and cecal contents were quantified as described previously (34). Briefly, samples were first homogenized in trace metal analysis grade water (Aristar Ultra), then the homogenates were mixed (1:1 ratio) with a 1 M HCl and 10% (wt/vol) trichloroacetic acid solution and heated for 1 hour at 95 °C. The samples were centrifuged at 10 000 relative centrifugal force at room temperature for 15 minutes, followed by incubating the supernatants with an iron assay reagent (1:1, 1 mM ferrozine: 3 M sodium acetate, pH 5.5) plus 1% mercaptoacetic acid) at 1:1 ratio until the color turns into purple. Iron quantitation was done at 562 nm and compared with iron standard solution (1 mg/mL Fe in 2% HNO₃).

Quantitative Real-Time PCR

RNA was extracted from tissue samples using RNeasy isolation kit (Qiagen). cDNA was synthesized by reverse transcription from mRNA using the iScript cDNA Synthesis Kit (Bio-Rad). Gene expression was performed by quantitative real-time reverse transcriptase PCR using Taqman gene expression assay and using the StepOnePlus detection system (Applied Biosystems) with a standard protocol. Relative abundance of each transcript was calculated by a standard curve of cycle thresholds and normalized to RL32.

Statistics

Statistical analyses were performed with Prism (GraphPad Software v9). Statistical tests performed are 2-way ANOVA with Tukey's multiple comparisons post hoc test. Data are presented as mean ± SEM. Statistical significance was set at P < .05 for all analyses. The number of animals for each experiment is listed in figures and figure legends.

Results

We designed experiments to test whether liraglutide alters iron metabolism in addition to improving body weight and glucose tolerance in a mouse model combining both HH (Hfe−/− mice (31, 32, 35)) and DIO and glucose intolerance. The total body HFE KO and WT control mice were subdivided into liraglutide and saline vehicle groups and received daily subcutaneous administration of the respective treatment once daily for 18 weeks (experimental scheme in Fig. 1A). HFE KO (Hfe−/−) mice and their WT littermates were fed a high-fat diet before initiation of liraglutide/saline treatment to induce obesity, as seen in the increase in body weight and fat mass (Fig. 1B and C). Liraglutide reduced body weight, fat mass, and lean mass similarly in both WT and HFE KO mice (Fig. 1B-D). Liraglutide also improved glucose tolerance in both WT and HFE KO mice to similar degrees after 14 weeks of treatment (Fig. 1E). These results reassured us that liraglutide treats obesity-related metabolic disease in HFE KO mice similar to its effects on WT mice. They also suggest the function of the drug itself is not directly affected by elevated iron levels.

Figure 1.

Liraglutide reduces body weight in a diet-induced obese mouse model of hemochromatosis and littermate controls. (A) Experimental timeline: WT and HFE KO mice were fed 60% high-fat diet for 26 weeks. After 8 weeks of HFD, the mice were separated into 4 groups and received either daily saline or liraglutide (1000 μg/kg) for 18 weeks while maintained on HFD. Illustration was created with BioRender. (B) Body weight. (C) Fat mass. (D) Lean mass. (E) Intraperitoneal glucose tolerance test (2 g/kg) performed 15 weeks after initiation of daily saline or liraglutide. Animal number: WT saline (n = 12), WT liraglutide (n = 13), HFE KO saline (n = 14), HFE KO liraglutide (n = 13). Data are shown as means ± SEM. *P < .05 (2-way ANOVA with Tukey's post hoc test).

We then examined the effects of liraglutide on iron levels. After 18 weeks of liraglutide or saline treatment, mice were fasted for 4 hours, and peripheral blood was collected for hematologic assessment. Circulating iron levels were higher in HFE KO + saline mice compared to WT groups, as expected for this mouse model (Fig. 2A) (35). Consistent with our hypothesis, liraglutide treatment significantly decreased serum iron levels in HFE KO mice (Fig. 2A), although it was not able to restore serum iron levels to normal. The TIBC was lower in HFE KO mice than in WT mice, as expected (Fig. 2B). TIBC in both WT and HFE KO mice groups was reduced after liraglutide treatment (Fig. 2B). Also, consistent with human HH patients and previous reports of this mouse model, transferrin saturation (the percent of transferrin bound to iron) was elevated in both HFE KO mice groups, consistent with iron overload (Fig. 2C). Liraglutide treatment did not affect the transferrin saturation. This is consistent with a simultaneous proportional decrease in serum iron and TIBC, which approximates transferrin concentration, balancing one another (transferrin saturation is the ratio of serum iron over TIBC). Ferritin saturation also remained increased in HFE KO mice compared to WT mice and was similarly unaltered by liraglutide (Fig. 2D). These serum iron metabolism measurements were consistent with liraglutide treatment improving iron overload.

Figure 2.

Liraglutide decreases serum iron levels in HFE KO mice. (A) Serum iron levels. (B) Total iron-binding capacity (TIBC). (C) Transferrin saturation %. (D) Ferritin saturation %. (E) Hemoglobin (HGB). (F) Hematocrit (HCT). (G) Red blood cells (RBC). (H) Platelets (PLT). (I) Reticulocytes (%). (J) Mean corpuscular hemoglobin concentration (MCHC). (K) Mean corpuscular hemoglobin (MCH). (L) Mean corpuscular volume (MCV). (M) Red cell distribution width (RDW) %. (N) Lymphocytes (LYM) %. (O) Monocytes (MONO) %. (P) Eosinophils (EOS) %. (Q) Basophils (BAS) %. (R) Neutrophils (NEU) %. Animal number: WT saline (n = 12), WT liraglutide (n = 13), HFE KO saline (n = 14), HFE KO liraglutide (n = 13). Data are shown as means ± SEM. *P < .05 (2-way ANOVA with Tukey's post hoc test).

To further explore these effects of liraglutide, we assessed how various hematologic parameters respond to liraglutide treatment and iron status. The largest pool of iron in the body (nearly 2/3 total iron in humans) is contained in hemoglobin inside erythrocytes, so we first evaluated red blood cell parameters (1). Consistent with a reduction in overall systemic iron, liraglutide mildly reduced hemoglobin (Fig. 2E), hematocrit (Fig. 2F), and red blood cell count (Fig. 2G), although not below the lower limit of normal in mice (36). This effect was present in both WT and HFE KO mice, and liraglutide-dosed HFE KO mice were similar to WT saline-dosed mice, although not as low as liraglutide-dosed WT mice. Platelets and reticulocytes (Fig. 2H and 2I) were unaffected by genotype or drug treatment. The mean corpuscular hemoglobin concentration was also not altered by genotype or treatment (Fig. 2J). As expected in states of iron overload, the MCH (Fig. 2K) and mean corpuscular volume (Fig. 2L) were higher, and the red cell distribution width (Fig. 2M) was lower in both HFE KO groups compared to WT mice. We also assessed leukocytes. Liraglutide slightly increased lymphocytes in HFE KO mice but not in WT mice (Fig. 2N). It correspondingly slightly decreased monocytes in HFE KO mice and showed a trend of decreased monocytes in WT mice (Fig. 2O). Eosinophils, basophils, and neutrophils were not different between genotypes or treatments (Fig. 2P-2R). In total, these data further confirmed that iron metabolism and levels are improved by liraglutide treatment.

The liver is the largest storage site of iron outside of hemoglobin, so we next measured liver iron levels (1). As mentioned previously, patients with HH exhibit excessive iron accumulation within hepatocytes, predisposing them to liver fibrosis and progression to end-stage liver disease (10). HFE KO mice recapitulated iron overload in liver consistent with prior reports (Fig. 3A) (32, 37). Liraglutide treatment in obese mice leads to a reduction in liver mass, mostly related to decreases in steatosis. Therefore, we calculated the total hepatic iron content (total µg) using both the total liver mass and tissue iron content of each mouse (Fig. 3A). The total liver iron content was significantly decreased in both WT and HFE KO mice after liraglutide treatment (Fig. 3A). We also measured iron levels in duodenal tissue and cecal contents; these levels were comparable between both genotypes and treatments (Fig. 3B and 3C). These data suggest that liraglutide did not decrease iron uptake from cecal contents or increase the retention of iron in duodenal tissue. These data demonstrate that liraglutide not only decreases the serum levels of iron, but also alleviates hepatic iron overload in HFE KO mice.

Figure 3.

Liraglutide decreased hepatic iron and lipid stores in WT and HFE KO mice on a high-fat diet. (A) Total liver iron content. (B) Duodenal iron content. (C) Iron levels in cecal contents. Expression levels of genes regulating hepatic iron metabolism. (D) Hamp, (E) Fpn, (F) Dmt1, and (G) Tfrc. Expression levels of genes regulating duodenal iron metabolism. (H) Dmt1 and I. Fpn. (J) Liver weights normalized to body weight. Expression levels of genes regulating hepatic lipogenesis. (K) Scd1, (L) FASN, and (M) Ppar alpha. Animal number: WT saline (n = 12), WT liraglutide (n = 13), HFE KO saline (n = 14), HFE KO liraglutide (n = 13). Data are shown as means ± SEM. *P < .05 (2-way ANOVA with Tukey's post hoc test).

Mechanistically, HH leads to iron overload by decreased expression of the iron-regulatory hormone hepcidin, which causes increased absorption of dietary iron in the duodenum and increased iron transport from the duodenum into the portal circulation, ultimately leading to direct tissue injury induced by iron overload (8). Therefore, we next analyzed the expression of genes regulating iron metabolism in liver. The iron-regulatory hormone hepcidin is a polypeptide primarily produced by the liver (38). Hepcidin's role is to maintain proper iron homeostasis and it does so largely by regulating the expression and function of cell membrane-embedded iron exporter ferroportin (Fpn) (1-3). Hepcidin limits iron efflux into the bloodstream by binding to the iron exporter ferroportin in tissue macrophages (which recycle iron in erythrocytes) and intestinal enterocytes and thus leading to ferroportin internalization and degradation (1-3). Hamp, the gene encoding for hepcidin, was not differentially expressed in liver tissue between the genotypes or after liraglutide treatment (Fig. 3D). Despite the lack of difference in Hamp expression, Fpn showed a trend of increased expression with liraglutide treatment in WT mice and Fpn was significantly higher in HFE KO liraglutide compared to WT saline controls (Fig. 3E). Increased hepatic Fpn implies that sequestration of iron in hepatocytes was decreased by liraglutide treatment. Interestingly, the divalent metal transporter 1 (Dmt1) responsible for free iron import into hepatocytes either at the plasma membrane or in the endolysosomal pathway was also increased in both WT and HFE KO mice with liraglutide treatment (Fig. 3F). Transferrin receptor protein 1, encoded by Tfrc, is a transmembrane glycoprotein responsible for iron import from transferrin (most of the serum iron is bound to transferrin) into the cells. Tfrc had decreased expression in HFE KO mice compared to WT but was not affected by liraglutide (Fig. 3G). Overall, these results suggest that liraglutide treatment leads to decreased hepatic iron levels via a hepcidin-independent mechanism.

Systemic iron level regulation is largely a product of modification of duodenal uptake of iron via DMT1 and secretion into the portal vein via Fpn. Therefore, we investigated liraglutide's effects on intestinal absorption of iron by measuring the transcript levels of duodenal Fpn and Dmt1 (Fig. 3H and 3I). Duodenal Dmt1 transcription increased after liraglutide in WT, but not in HFE KO mice (Fig. 3H). Duodenal Fpn trended toward increased transcript levels after liraglutide treatment in WT, but not in HFE KO, mice (Fig. 3I). These data suggest that the decrease in systemic iron levels in HFE KO mice dosed with liraglutide is not regulated by transcriptional suppression of duodenal iron uptake genes.

Finally, we investigated the effects of liraglutide on liver weight and steatosis. As mentioned earlier, liraglutide decreased liver weight regardless of genotype (Fig. 3H). There was decreased expression of stearoyl coenzyme A desaturase 1 (Scd1). The Scd1 protein is an enzyme involved in lipogenic metabolism and its activity is increased in humans and rodent models of nonalcoholic fatty liver disease and hepatic insulin resistance (4, 39). Deficiency of Scd1 has been reported to increase fatty acid oxidation in the liver and is thus protective against increased body adiposity in mice (39). Furthermore, increased hepatic expression of Scd1 has been reported in hepatic iron overload, suggesting that iron overload may independently stimulate the biosynthesis of fatty acids (40). However, these data show that liraglutide decreased hepatic lipids by decreasing hepatic Scd1, but this mechanism could be either independent of or dependent on iron overload status. The decrease in Scd1 transcription with liraglutide treatment in WT and HFE KO mice also corresponds to the decrease in total liver iron stores (Fig. 3A). Two other important genes involving in fatty acid metabolism, fatty acid synthase (FASN) or peroxisome proliferator-activated receptor alpha (PPARα), were not differentially expressed (Fig. 3J and 3K). These data demonstrate that liver steatosis and weight are effectively dosed by liraglutide in WT and HFE KO mice.

In summary, our findings show that serum iron was decreased by liraglutide both in the setting of iron overload (HFE KO mice) and in normal WT mice (Fig. 4). Importantly, hepatic iron levels were also decreased in iron-overload HFE KO mice (Fig. 4).

Figure 4.

Liraglutide decreases hepatic iron levels in DIO WT mice and reduces hepatic and circulating iron levels in DIO HFE-KO (Hfe−/−) mouse model of hereditary hemochromatosis. Illustration was created with BioRender.

Discussion

In this study, we demonstrate an effect of liraglutide in reducing systemic and hepatic iron levels in a mouse model of HH (Figs. 2-4). The mouse model of human hereditary hemochromatosis we used, the HFE KO mouse, is well-characterized and include serum and hematologic markers of hemochromatosis, as well as increased hepatic iron accumulation (31). Our findings show that serum iron was decreased by liraglutide in the setting of iron overload (HFE KO mice) (Figs. 2 and 4). Importantly, hepatic iron levels were decreased in both WT mice and iron-overload HFE KO mice (Fig. 3). Serum iron is a measure of the total iron not incorporated into erythrocytes or leukocytes and generally correlates with overall iron load. While transferrin saturation is elevated, as expected, in HFE KO mice, this value was not altered by liraglutide treatment as we might expect in the setting of overall iron level improvement. However, this is explained by the concomitant decrease in total iron-binding capacity, which is an indirect measure of transferrin concentration. In the setting of decreased serum iron but correspondingly decreased TIBC/transferrin levels, there is minimal change in iron saturation of transferrin. Patients with hemochromatosis, just like this mouse model, often have low normal or below normal TIBC. We might therefore expect improvements (decreases) in iron levels to normalize TIBC, but importantly transferrin is also affected by inflammation. For instance, in anemia of chronic disease, a common anemia found in ill patients with a chronic inflammatory state, despite adequate iron stores, there is low iron availability with concomitant suppression of transferrin expression, and hence low TIBC (41). Obesity and diabetes are pro-inflammatory processes, and liraglutide is known to decrease inflammation (42). Therefore, we speculate that liraglutide-mediated improvements in inflammation might increase transferrin expression, reducing TIBC in the setting of a concomitant decrease in iron. However, these processes merit further exploration.

Ultimately serum laboratory values are proxy measurements for overall iron load, which mediates the actual tissue damage in hemochromatosis. The largest pool of body iron is in hemoglobin stored in red blood cells. We see minimal decreases in hemoglobin, hematocrit, and red blood cell count after liraglutide treatment, although this is not unexpected because the organs most affected by iron excess in patients with HH are the liver, heart, and pancreas (Fig. 2). Therefore, perhaps the most important aspect of our study is the significant decrease in overall iron load in the liver in liraglutide-dosed mice. One question is what the mechanism of this improvement in liver iron overload might be. Based on our transcriptional data, the improvement is not merely a reversal of the low hepcidin state in hemochromatosis as hepatic hepcidin expression is similar. Furthermore, at least in liver tissue, we see increases, albeit not statistically significant changes, in Fpn expression rather than decreases as would be seen with hepcidin activity (Fig. 3). Although an Fpn increase could explain how the liver could unload iron, the trend toward Dmt1 increases is harder to explain because these would in principle increase liver iron uptake (Fig. 3). We hesitate to speculate on a mechanism based on nonstatistically significant differences in transcription. Of note, a recent study of liraglutide in leptin receptor-deficient mice, which are genetically prone to extreme obesity and diabetes, did note improvements in liver iron load just as in our data, but reported slight increases in transferrin receptor, decreased Fpn, and stable Dmt1 protein levels (30). When we examined duodenal transcript levels of the iron uptake genes Dmt1 and Fpn, we observed significant increases in Dmt1 and a trend toward Fpn increase in WT mice but no change in HFE KO mice (Fig. 3). This is surprising because it would be consistent with no duodenal reduction in iron uptake in HFE KO mice, and perhaps even increased iron uptake in WT mice. We speculate that this is the result of reach a plateau by the time treatment was finished. Future work is needed to clarify potential mechanisms mediating liraglutide-induced decreases in both systemic iron levels and liver iron unloading.

Our assessment of serum iron-related laboratory values is consistent with decreases in available iron. Serum iron was decreased by liraglutide in normal WT mice and HFE KO mice with iron overload (Fig. 2). Serum iron is a measure of the total iron not incorporated into erythrocytes or leukocytes, and generally correlates with overall iron load. While transferrin saturation is elevated, as expected, in HFE KO mice, this value was not altered by liraglutide treatment as we might expect in the setting of overall iron level improvement. However, this is explained by the concomitant decrease in total iron-binding capacity, which is an indirect measure of transferrin concentration. In the setting of decreased serum iron but correspondingly decreased TIBC/transferrin levels, there is minimal change in iron saturation of transferrin. Patients with hemochromatosis, similar to HFE KO mice, often have low normal or below normal TIBC. We might therefore expect improvements (decreases) in iron levels to normalize (increase) TIBC. Transferrin levels are also suppressed by inflammatory processes, such as in anemia of chronic disease (41). Obesity and diabetes are pro-inflammatory processes, and liraglutide is known to decrease inflammation (42). Therefore, we might also expect liraglutide-mediated improvements in inflammation to increase transferrin expression, not decrease it. Hence, why TIBC decreases with liraglutide treatment is not clear. Despite this unpredicted effect on transferrin levels, overall serum iron levels decrease with liraglutide dosing in mice, implying that measurement of TIBC should be interpreted with care in patients with HH who may be treated with liraglutide.

We deliberately opted to examine the effects of liraglutide in HFE KO mice that also had DIO. Our data show that iron overload does not affect the efficacy of liraglutide on obesity and diabetes outcomes (Fig. 1). However, because all mice in this study were maintained on high-fat diet during the studies, we do not know to what extent that diet and diet choice might alter liraglutide's effect to alter tissue-specific iron storage. These studies could be repeated in lean mice to dissect the effects of dietary iron intake and obesity on liraglutide's role in iron metabolism. We also did not address the role of body weight and body weight loss in response to liraglutide in altering iron levels in the HFE KO mouse model of HH. Liraglutide-dosed mice stabilized their body weight at approximately 6 weeks’ posttreatment initiation, 12 weeks before we obtained laboratory results and tissues. Our previous studies show that liraglutide decreases food intake only for the first 4 days of treatment, although body weight remains lower for the duration of treatment (43). Therefore, it is reasonable to expect that the dietary intake is unlikely to have affected iron-related parameters. We have previously shown that neuronal GLP1R receptors mediate the body weight and anorectic effects of liraglutide, but that these receptors are not required for its glucose-lowering effects (33). Further studies will need to dissect whether populations of GLP1R-expressing cells contribute to liraglutide's effect on iron metabolism.

Taken together, these data demonstrate that liraglutide is not only effective in alleviating obesity-related metabolic disease in the context of a mouse model of HH but is also able to modify iron metabolism in HFE KO mice. HH triggers excessive iron accumulation in liver and that predisposes patients with HH to hepatic liver disease including the increased risk of developing liver cancer (9-11). However, many patients with HH are also at an increased risk for developing metabolic syndrome, including obesity and diabetes, and may be treated clinically with GLP1RAs such as liraglutide (13-19). Importantly, these effects are independent of hepcidin transcription, although it remains possible that they modify hepcidin activity in a transcription-independent fashion. We also do not see a clear pattern of iron transporter expression changes that would explain these effects. Nevertheless, these observations highlight the potential that GLP1RAs could be investigated further as therapeutic targets to reduce iron overload in patients with HH, particularly in those with concurrent diabetes and/or obesity.

Abbreviations

DIO

diet-induced obesity

GLP1RA

GLP1 receptor agonist

HFE KO

Hfe knockout

HH

hereditary hemochromatosis

MCH

mean corpuscular hemoglobin

TIBC

total iron-binding capacity

WT

wild type

Funding

This work was supported by NIH P30DK089503, 5T32DK108740 (N.B.K.) and Department of Veterans Affairs IK2BX005715 (N.B.K.).

Author Contributions

R.J.S. conceived and designed the study. N.B.K., J.H.S., N.B.B., C.J., and N.K.D. performed the experiments and analyzed results. Y.S., L.B.K., and J.P.W. helped with data interpretation and discussion. All authors edited manuscript. R.J.S. provided final approval of the submitted manuscript.

Disclosures

R.J.S. has received research support from Novo Nordisk, Fractyl, Astra Zeneca, Congruence Therapeutics, Eli Lilly, Bullfrog AI, Glycsend Therapeutics, and Amgen. R.J.S. has served as a paid consultant for Novo Nordisk, Eli Lilly, CinRx, Fractyl, Structure Therapeutics, Crinetics, and Congruence Therapeutics. R.J.S. has equity in Calibrate and Rewind. J.P.W. and L.B.K. are paid employees of Novo Nordisk. J.H.S. is a paid employee of Amgen. The remaining authors declare no competing interests.

Data Availability

The authors declare that the data supporting the findings of this study are available within the paper. Source data can be shared upon request.