Association between leptin and NAFLD: a two-sample Mendelian randomization study

Over the past two decades, nonalcoholic fatty liver disease (NAFLD) has progressed from a relatively unknown disease to the leading cause of chronic liver disease worldwide [1]. Its global frequency is quickly increasing, reaching up to 25% in developed countries like the United States [2]. NAFLD is a degenerative disease caused by the buildup of intracellular lipid droplets in liver cells, which can induce inflammation, cell death, and even more advanced stages such as nonalcoholic steatohepatitis (NASH) (with or without fibrosis), cirrhosis, and liver cancer [3, 4]. Currently, pharmacological options for NAFLD are limited. Treatment cornerstones are a healthy lifestyle and weight loss. There is still an unmet therapeutic need [5].

NAFLD is bidirectionally associated with components of the metabolic syndrome [6], a cluster of alterations that includes centripetal obesity, decreased HDL cholesterol concentrations, increased triglyceride concentrations, arterial hypertension, and hyperglycemia [7‐9]. This syndrome has become one of the epidemics of the twenty-first century. Causative factors include insulin resistance, leptin, lipocalins, microbiota alterations, and epigenetics [10, 11]. Among these, leptin is a molecule secreted primarily from adipose tissues. The circulating levels of leptin are proportional to abrupt changes in percent body fat mass or caloric intake. It is a key regulator of metabolism and energy homeostasis [12]. A potential dual role has been shown between leptin and NAFLD, with leptin possibly exerting an anti-teratogenic effect while also having a pro-inflammatory and pro-fibrotic impact [13‐16]. At the same time, observational clinical studies have shown a relationship between persistent hyperleptinemia and the development of steatosis, fibrogenesis, and hepatocellular carcinoma, suggesting that hyperleptinemia is an independent predictor of the presence or development of NAFLD [12, 17, 18]. According to one study, leptin modulated the function of target cells (hepatocytes and macrophages) and controlled their pyroptosis-like cell death via CD8+ T lymphocytes. The interference of leptin and immune cell-related pathways may provide promising strategies for the treatment of NAFLD [19]. However, the relationship between leptin and NAFLD still needs to be clarified.

In epidemiological studies, the presence of confounders dramatically perturbs causality inferences between exposures and outcomes because causality inferences in observational studies are often challenged by potential confounding biases and reverse causality. Additionally, randomized controlled trials (RCTs) have limitations related to ethical issues, observation time, and resources and costs [20]. Mendelian randomization (MR) is an approach that uses genetic variants associated with specific exposures of interest to study the causal effects of modifiable exposures (potential risk factors) on health, social, and economic outcomes [20‐24]. In recent years, genome-wide association studies (GWAS) have accumulated millions of data points on associations between genetic variants and complex diseases or phenotypes [25, 26]. Two-sample Mendelian randomization (TSMR) analysis is an optimized extension of the one-sample Mendelian randomization (OSMR) analysis, in which aggregated statistics from published GWAS are used instead of individual-level data, with distinct samples for exposure variables and outcome markers. It allows for the evaluation of the causal impact of exposure factors on outcomes without the need for additional studies, reducing research expenditures and improving the bioinformatic application [23, 27, 28].

To the best of our knowledge, no MR investigations have been conducted to investigate the potential causative link between leptin and NAFLD. In this study, to provide some basis for the prevention and treatment of NAFLD and lower the incidence and disease burden, TSMR analysis was utilized to analyze the possible causal relationship between leptin and NAFLD from a genetic perspective using GWAS summary statistics for leptin and NAFLD with leptin-related gene polymorphisms as instrumental variables (IVs).

We performed a TSMR using publicly available summary-level data from several large-scale cohorts. An overview of this study design is shown in Fig. 1.

In this study, we used TSMR to assess the association between leptin and NAFLD. The OR values obtained using IVW (OR 0.6729; 95% CI 0.4907–0.9235; P = 0.0142) and WM methods (OR 0.6549; 95% CI 0.4373–0.9806; P = 0.0399) provide strong evidence that elevated levels of leptin are causally associated with a reduced risk of NAFLD, suggesting that leptin may act as a protective factor for NAFLD to some extent. Therefore, even if the exact mechanism is unknown, we believe that increasing the level of leptin in NAFLD patients can lower the risk of NAFLD development. Furthermore, our study initially explored a causal relationship between leptin-related phenotypes and body composition. However, no strong evidence was found to support a causal relationship between leptin and body composition and adipose tissue.

It is worth noting that a group of experts has proposed altering the nomenclature of NAFLD to metabolic-dysfunction-associated fatty liver disease (MAFLD), signaling a shift in the paradigm and underlying etiology toward a more broad term that does not specifically address NAFLD in recent years [48, 49]. Instead of being a disease in and of itself, fatty liver is a histological change in the liver that is a reflection of anomalies in the human metabolic system. As a result, the focus of the disease's treatment is on metabolic management, although in practical practice, there are still a number of bottlenecks to treating NAFLD [9]. To begin with, although there are better recognized and established methods for monitoring NAFLD in traditional high-risk categories (diabetes, hypertension, and overweight), increasing numbers of studies have indicated that NAFLD occurs more commonly in adults with normal BMI (i.e., visceral obesity) [50]. For such patients, especially nonobese people with intermittent transaminase abnormalities, there is still a lack of sufficiently effective biological markers for the diagnosis of NAFLD, as well as a lack of effective techniques for predicting the risk of development. It is due to the fact that liver aspiration biopsy is still a rare procedure, and current guidelines merely prescribe regular follow-up, which does not allow for early treatment [51, 52]. Second, according to the existing guideline recommendations, lifestyle changes, exercise, and diet control are important nonpharmacological options for people with established NAFLD [53]. Still, these nonpharmacological therapies lack sufficient quantifiable indicators, particularly for weight control (guidelines recommend a 5–10% reduction), which is not feasible in lean nonalcoholic fatty liver disease. In people with NAFLD with normal liver enzymes, there is a lack of evaluable serological indicators [54]. Third, due to the complex pathological mechanism of NAFLD itself and the fact that this disease is often only the "tip of the iceberg" of metabolic system diseases, the current treatment of fatty liver lacks sufficiently targeted and recognized effective treatment options. Although, in recent years, there have been advances in treatment options, including the use of a combination of drugs such as semaglutide, firsocostat (ACC inhibitor), and cilofexor (FXR agonist), their reliability and effectiveness need to be confirmed by clinical trials [55]. Therefore, although the treatment of fatty liver is a classic and relatively old topic, developing its specific drugs is still a virgin territory to be explored, and the discovery of more therapeutic targets is of great value for drug development. Based on a review of the literature, leptin regulates food intake, energy balance/body weight, and some metabolic functions [56]. In this regard, leptin should have an anti-steatosis impact on hepatocytes [15]. However, no therapeutics for NAFLD are directed at this target. As a result, the finding that leptin and NAFLD are correlated may be useful for assessing disease risk, preventing NAFLD, combining existing therapy regimens for potentiation, and identifying prospective targets for novel drug development. Our study, based on the literature, did find a significant association between increased leptin levels and reduced incidence of NAFLD, and our findings not only coincide with previous literature but also validate the hypothesis in the literature through database analysis of real-world case sources.

Leptin signals through binding to its receptors, mainly Lep Rb, which is a long stretch of extracellular structure, a transmembrane region and an elongated intracellular extension. As Lep Rb does not possess intrinsic kinase activity, the conformational change of Lep Rb upon leptin binding to Lep Rb induces the activation of Janus kinase (JAK2) phosphorylation, which phosphorylates three tyrosine residues (Y985, Y1077 and Y1138) in the intracellular extension of Lep Rb. These phosphorylated tyrosine residues then recruit proteins containing the SH2 phosphorylation recognition domain for downstream signaling. Currently, the most studied leptin signaling is the JAK/signal transducer and activator of transcription (STAT) pathway. Leptin and NAFLD exert their effects mainly through the JAK2/STAT3 pathway [57, 58]. An important role of leptin is to direct the storage of triglycerides in adipocytes and prevent their deposition in non-adipose tissues such as the liver, thus preventing hepatocyte lipotoxicity and apoptosis. Leptin also inhibits the production of hepatic glucose and the formation of new hepatic fat, acting as an insulin-like agent to prevent the development of NAFLD. Studies have shown that chronic central leptin infusion can reduce hepatic lipid synthesis gene expression and triglyceride levels by stimulating hepatic sympathetic activity and that this effect of leptin is associated with the PI3K signaling pathway, blocking which can specifically induce hepatic steatosis without causing obesity. In addition, leptin promotes fatty acid oxidation in the liver and increases fatty acid consumption in the liver [59]. In addition to its direct effects on the liver, leptin also affects hepatic glucose metabolism indirectly through its central regulation. Leptin infusion into the ventricles of type 1 diabetes mice inhibited the expression of glucagon, consistent with the phenotype of peripheral hyperleptinemia [60]. Specific expression of Lep Rb in the arcuate nucleus of the rat hypothalamus by adenoviral transfection improves peripheral insulin sensitivity and reduces hepatic gluconeogenesis in leptin receptor-deficient Koletsky rats [61]. The regulation of hepatic glucose by leptin may be related to the effect of its phosphatidylinositol 3-kinase (PI3K), which increases insulin signaling and decreases the expression of glucose synthesis genes such as glucose-6-phosphatase (G-6-P) and phosphoenolpyruvate carboxykinase (PEPCK) [61]. In addition, the effects of leptin can also be mediated by central neural regulation, e.g., selective severance of the hepatic vagus nerve can prevent hypothalamic leptin from regulating hepatic insulin sensitivity.

The mechanism of leptin in NAFLD has been supported by a large body of experimental data, and clinical studies on leptin and NAFLD have focused on the association of leptin or leptin receptor levels with NAFLD. The findings on circulating leptin levels in NAFLD patients are not very consistent, with some studies reporting high leptin expression in NAFLD patients [62, 63] and others finding no difference in leptin levels in NAFLD patients compared to non-NAFLD populations [64, 65]. Clinical studies of leptin and Lep R gene expression and SNPs in the NAFLD population have also been reported sporadically. Two small clinical studies showed no expression of the Lep R gene in liver tissue, while in peripheral leukocytes and abdominal adipose tissue Lep gene expression did not differ significantly between NAFLD patients and healthy populations [66, 67]. In another study, immunohistochemical staining of liver tissue for leptin showed that leptin expression was higher in patients with NAFLD than in the healthy population, consistent with altered circulating leptin levels [63]. Some of the Lep R gene SNP studies have also shown a positive association with the development of NAFLD, even if this association is not dependent on the presence of obesity. Given the complexity of clinical studies and the multilevel nature of clinical data, it is difficult to obtain direct evidence that leptin resistance causes NAFLD from the available clinical research data, which need to be interpreted with caution.

Our study has several advantages. First, the TSMR analysis method is based on the principle of Mendelian randomization-free segregation and combination, which excludes the influence of acquired factors (social environment and natural environment) on the study results at the genetic level. In order to successfully compensate for the vulnerability to confounding factors and reverse causality interference in traditional observational studies for inferring the etiology of complex disorders, the genes must arise prior to the disease with a precise causal time sequence [68, 69]. Second, this study uses publicly available GWAS summary statistics with a large sample size to obtain more precise estimates and greater statistical power, saving research costs and improving the utilization of biological information while limiting the study population mainly to individuals of European ancestry, reducing some of the bias that may arise due to population stratification. Finally, the value of this study lies in establishing an association between leptin levels and the incidence of NAFLD using a database of real-world sources. Based on our findings, it is reasonable to believe that leptin levels may be used for the assessment of NAFLD, including the screening of people who are traditionally at high risk of developing NAFLD (e.g., those with comorbid diabetes and those who are overweight), and, more importantly, for the assessment of the risk of developing lean NAFLD in people with normal BMI. In addition, it can be used as an indicator to evaluate the improvement potential of NAFLD. Since there are no drugs that target leptin, we believe that, based on the current state of research, leptin can reduce the incidence of NAFLD without duplicating the mechanism of action of other existing drugs for NAFLD and can be used as a complement to existing treatment regimens. Additionally, there is evidence that leptin regulation may have favorable effects on a variety of other factors, including weight loss, reducing blood sugar levels, and controlling intestinal functions [70]. Therefore, modulation of leptin levels may be used in multiple aspects of metabolic disorders and may have a wider range of potential applications. However, there are some limitations to this study. First and foremost, the majority of these GWAS data are from European populations. It needs to be determined if the findings we described would hold in other people. Second, this study lacks a multidimensional stratification of the heterogeneity of patients with NAFLD. In the future, a multicenter prospective cohort study is needed to fully consider the heterogeneity of NAFLD, integrate demographic characteristics, lifestyle, genetics and other factors to accurately identify high-risk groups for NAFLD, and develop targeted and individualized body mass control strategies, with a view to achieving accurate prevention and control of NAFLD.

In conclusion, our study reveals a causal relationship between leptin and NAFLD. It thus provides further insight into the factors that may be associated with a reduced risk of NAFLD development. Additionally, from a systems biology standpoint, it aids researchers in better understanding the connections between diverse diseases.