Markers linked to sudden pancreas inflammation in people with BMI over 30 using genetic and gene activity analysis

Acute pancreatitis (AP) triggers a significant inflammatory response caused by a wide range of factors. This sequence of events leads to the activation of pancreatic enzymes within the pancreas, resulting in autodigestion, tissue swelling, bleeding, and potentially, pancreatic tissue necrosis. Key triggers include cholelithiasis and alcohol consumption. Statistical evidence underscores the substantial public health impact of AP, with an incidence rate of 33.74 cases per 100,000 person-years [1]. In certain countries, the incidence rate of AP has risen to 72 cases per 100,000 person-years [2]. Remarkably, the overall mortality rate associated with this variant of pancreatitis is 15%. When organ failure becomes a prominent feature of the disease, the mortality rate surges to a staggering 35% [3]. In Sweden, the annual financial burden of AP is 38,500,000 euros, with the average treatment cost per patient reaching approximately 10,000 euros [4]. The evaluation of the disease's impact underscores the critical need for proactive management strategies and heightened awareness among healthcare providers and the public. Furthermore, this study highlights the essential need for comprehensive and prompt diagnostic and therapeutic interventions. However, the complex pathogenesis of AP at the microlevel presents a formidable challenge, with many aspects yet to be uncovered.

The World Health Organization defined a Body Mass Index (BMI) of 30 or higher as indicating obesity. Obesity substantially affects the development and progression of AP, exerting multiple adverse effects on this condition. Research indicates that obesity can exacerbate the inflammatory response associated with AP, leading to detrimental outcomes [5, 6]. A study conducted in 2014 corroborated these findings, highlighting the strong link between obesity and the exacerbation of the systemic inflammatory response in acute pancreatitis [7]. Individuals with a BMI of 23 or higher face a significantly increased risk of developing severe acute pancreatitis compared to those with a normal BMI [8]. Obesity is acknowledged as a major risk factor for AP, introducing additional complexity to the disease etiology. Considering the established link between obesity and AP, unraveling the potential molecular mechanisms that connect these two conditions is crucial. Gaining such insights is essential for enhancing our understanding of AP pathogenesis and could pave the way for innovative therapeutic approaches for individuals affected by this condition. Furthermore, lipid metabolism has been identified as a key element in the pathogenesis of various diseases, including obesity, AP, cancer, immune disorders, and neurodegenerative diseases [9–12]. The complex interplay between lipid metabolism and disease pathogenesis highlights its importance in understanding disease mechanisms and developing personalized treatment strategies.

Bioinformatics offers a systematic approach to deciphering complex biological processes, aiding in the identification of molecular signatures that underpin disease pathophysiology. Machine learning, a subset of artificial intelligence, bolsters bioinformatics through the use of algorithms capable of identifying patterns and relationships within vast datasets. The integration of machine learning into modern precision medicine is attributed to its ability to accurately process and manage large quantities of data, enhancing the development and application of personalized treatment strategies [13]. By integrating the capabilities of bioinformatics and machine learning, it becomes possible to efficiently explore the complex network of molecular interactions and pinpoint potential biomarkers with clinical significance.

Despite the limited research exploring the causal relationship between BMI and AP and the scarcity of studies identifying shared diagnostic biomarkers for high BMI status and AP, this study sought to address these gaps. Initially, hypothesizing a distinct causal link between BMI and AP complicated by genetic factors, this investigation represents a pioneering effort to combine Mendelian randomization (MR), bioinformatics analysis, and machine learning algorithms to examine the BMI-AP connection. Moreover, this study aimed to identify key genes implicated in the progression of AP in individuals with a BMI >30.

All the statistical analyses and visualizations were performed using R software (version 4.3.1). Comprehensive descriptions of the statistical tests employed can be found in the corresponding bioinformatics methods section and figure legends.

The global prevalence of AP substantiates its significance as a common gastrointestinal emergency, necessitating urgent attention and intervention. Although substantial advancements have been made in the clinical treatment and diagnostic laboratory parameters of AP in recent years, the intricate anatomical positioning of the pancreas, coupled with the subtle presentation of this type of pancreatitis, poses significant challenges in the realm of early diagnosis. Delayed detection of the disease has profound ramifications, including heightened susceptibility to complications and elevated mortality rates. Failing to promptly diagnose and institute appropriate therapeutic measures has led to severe pancreatitis in a substantial portion of patients, such as necrotizing pancreatitis or even organ failure. Consequently, the elucidation of diagnostic biomarkers for this condition remains a pressing concern. Mounting evidence points to a conspicuous correlation between high BMI and the onset and progression of AP, as highlighted in the literature [19, 20]. An increased prevalence of comorbidities is frequently observed in obese patients, predisposing them to an elevated risk of developing severe pancreatitis [5, 21]. Furthermore, lipids, which are essential elements of cellular architectures, play a pivotal role in forming phospholipid bilayers, which are fundamental to cell membrane integrity and function. In addition to serving as vital sources of energy and key players in cellular metabolic activities, rigorous scientific investigations have revealed the complex dynamics interlinking lipid metabolism with obesity and AP. These studies have shed light on the profound and intricate connections underpinning these biological phenomena, offering new insights into their interdependencies [7, 12, 22, 23]. In the pursuit of tailored diagnostic and therapeutic approaches for AP, it has become crucial to investigate the cumulative effect of BMI > 30, AP, and lipid metabolism from the standpoint of gene expression.

This study broke new ground by utilizing available GWAS data through a dual-sample MR approach to clarify the causal link between BMI and the risk of AP. The results decisively demonstrated a direct correlation between higher BMI and increased vulnerability to AP. In an effort to explore the molecular foundations of AP, especially among individuals with a BMI > 30, within the context of lipid metabolism, this research undertook a series of critical analytical efforts. The “limma” R package offers a robust framework for analysing gene expression data [15]. In this study, 1372 DEGs were observed among the BMI > 30 and BMI < 30 groups using the “limma” R package. In this study, 1233 significant module genes were identified, 698 ORDEGs were identified from the intersection of DEGs and module genes. Subsequently, the same methods were used to analyse the AP dataset, resulting in 4828 APDEGs. The intersection of ORDEGs and APDEGs, along with LMRGs, revealed 21 common risk genes associated with AP in patients with a BMI > 30. This foundational phase of this study enabled the identification of genes that undergo significant expression changes, potentially playing pivotal roles in the pathogenesis of AP among individuals with a BMI > 30. Despite the recognition of these genes, the precise mechanisms through which they contribute to AP regulation in obese patients remain elusive. Moreover, the biomarkers initially identified, while numerous, proved to be impractical for clinical application. The objective of this investigation was to refine the search for diagnostic biomarkers with increased specificity and accuracy, necessitating a more detailed examination of the expression of these genes.

Initially, to elucidate the regulatory mechanisms of the identified genes within the human body, this study performed an enrichment analysis of signaling pathways and biological functions linked to the 21 identified genes. KEGG analysis revealed that these CDEGs were predominantly enriched in pathways and functions associated with infection and inflammation. These pathways included the cytosolic DNA-sensing pathway, cytokine‒cytokine receptor interaction pathway, and NOD-like receptor signaling pathway. These findings indicate that these terms are closely related to inflammatory processes and the body's response to them. Prior research has underscored the pivotal role of cytosolic DNA sensing in tissue damage and inflammation across a variety of diseases [24–27]. Obesity-induced mitochondrial DNA (mtDNA) release initiates an increase in chronic sterile inflammatory responses in adipose tissue via this pathway [28]. Additionally, it plays a role in inflammation associated with AP, where acinar cell death activates interferon (IFN) signaling through the STING pathway in macrophages. This highlights the essential role of IFNs in AP through various innate immune-sensing pathways [29]. Cytokine‒cytokine receptor interactions and the MAPK signaling pathway are associated with consistent cytokine expression throughout different stages of severe acute pancreatitis, indicating their involvement in the regulation and progression of the inflammatory response during the disease [30]. NOD-like receptors constitute a broad and intricate group of signaling regulators. These proteins consolidate both favorable and unfavorable signals and subsequently activate additional signaling regulators implicated in inflammatory responses, tumorigenesis, cellular senescence, and stem cell characteristics [31]. Moreover, inhibition of the NLRP3 inflammasome has been shown to reduce the degree of experimentally induced AP in obese mice [32]. These findings shed light on the roles of these genes in contributing to the intricate regulatory networks that oversee cellular functions and systemic responses.

Moreover, GO analysis revealed significant enrichment in processes such as the neuroinflammatory response, regulation of T-helper 2 cell differentiation, regulation of the inflammatory response, and positive regulation of the type 2 immune response. We were surprised to find that the GO terms were predominantly related to immunity and inflammation. To gauge the extent of infiltration by immune elements and understand the proportion of immune cells in AP in patients with a BMI > 30, this study employed the "ssGSEA" algorithm to assess immune infiltration levels. Several immune cell types, such as activated dendritic cells, CD56bright natural killer cells, central memory CD8+ T cells, effector memory CD4+ T cells, eosinophils, and macrophages, exhibited consistent correlations in the BMI>30 and AP datasets. These findings significantly underscore the role of immune factors in the mechanisms triggering AP in individuals with a BMI > 30. These findings suggest the initiation of a prolonged proinflammatory response and the mobilization of inflammatory cells, consistent with the findings of previous studies [33–35]. These findings offer valuable insights, indicating that inflammation plays a crucial role in the pathogenesis of AP in patients with a BMI > 30.

To pinpoint more accurate diagnostic biomarkers for AP in patients with a BMI > 30, focusing on lipid metabolism, this study employed LASSO, RF, and SVM-RFE to conduct further analysis on two disease-specific datasets. LASSO regression, commonly used to filter variables and mitigate the risk of overfitting, was employed. The optimal number of DEGs can be determined using the binomial deviation method [36]. RF is adept at ranking genes and is well suited for managing high-dimensional data, constructing prognostic models, and assessing the significance of individual variables [37]. SVM-RFE has proven to be a convenient tool for eliminating redundant components and retaining outcome-relevant variables, especially in datasets with limited samples [38]. In this study, four genes (MMP9, ABCA3, HADH and ESR1) were identified as potential diagnostic markers by cross-referencing the outcomes derived from three distinct machine learning methodologies in the BMI >30 dataset. The same methodologies were applied to pinpoint five genes (ACSS2, MBOAT2, LRIG1, IL18 and HADH) with potential diagnostic value in the AP dataset. Fascinatingly, the HADA gene has prominently surfaced during the meticulous search for diagnostic biomarkers applicable to both conditions under study. This finding was validated through subsequent colocalization analysis, analysis of GEO external datasets, and qPCR experiments in animal models. This observation led us to propose that the HADA gene has a significant, undeniable influence on the initiation and progression of AP in individuals with a BMI >30.

In the context of recent advancements, an extensive body of work has been devoted to pinpointing biomarkers critical for the diagnosis and progression of AP. A landmark study in 2023 by Zheng Wang et al. revealed severe AP, with a special focus on the role of immunogenic cell death mechanisms. Their groundbreaking findings revealed that LY96, BCL2, and IFNGR1 were instrumental biomarkers for both the emergence and evolution of severe AP [35]. In a hospital-based case‒control investigation, Francisco D'Oliveira Martins and his team proposed that GSTM1 may increase vulnerability to AP [39]. Data from the AP, septic AP, and control groups were collected, and CitH3 levels were meticulously quantified using enzyme-linked immunosorbent assay (ELISA). This method is pivotal for identifying circulating CitH3 as a reliable marker for diagnosing and predicting outcomes in septic AP patients [40]. A subsequent forward-looking, double-blind study identified fatty acid ethyl ester (FAEE) as a precise marker for diagnosing alcohol-related pancreatitis [41]. Further research has highlighted the potential of intercellular adhesion molecule 1, red cell distribution width (RDW), along with urinary trypsinogen-2 and trypsinogen activating peptide (TAP), as informative biomarkers for AP [42–44]. Despite these advances, investigations specifically targeting AP in individuals with a BMI >30 are lacking. Bridging this knowledge gap, our research suggested that HADH is a novel biomarker indicative of AP onset within this specific population.

HADH is located on chromosome 4q25 and is affiliated with the 3-hydroxyacyl-CoA dehydrogenase gene family. It codes 3-hydroxyacyl-CoA dehydrogenase, a pivotal enzyme in the fatty acid beta-oxidation pathway. HADH expression is widespread across various tissues (especially adipose tissue), with notably high enzyme activity observed in the pancreas. Mutations in HADH have been linked to hyperinsulinemic hypoglycemia, a condition characterized by abnormalities in insulin secretion and recognized as a fatty acid oxidation deficiency disease [45–48]. Moreover, reduced HADA expression has been demonstrated to enhance tumor cell migration and invasion by activating the Akt signaling pathway [49]. Research indicates that elevated HADH expression is correlated with an unfavorable prognosis in acute myeloid leukemia patients [50]. Poor clinical outcomes have also been observed in colon cancer patients with high HADH expression [51]. These findings highlight the diverse roles of HADA in cellular processes across different diseases.

A study leveraging proteomics techniques revealed that, in comparison with individuals of normal weight, obese patients exhibited markedly lower HADH protein levels, averaging merely 45% of the control group's HADH protein levels. Analysis via Ingenuity Pathway Analysis suggested that this decrease in HADH might be associated with inhibited activation of the LXR/RXR pathway [52]. This hypothesis is supported by experimental observations in which HADH-deficient mice displayed a compromised ability to metabolize TG in plasma under cold stress conditions. This impairment led to significant triglyceride and fatty acid accumulation [53], underscoring the role of elevated plasma fatty acid levels in promoting obesity. Fatty acid interferes with the ability of insulin to inhibit lipolysis [54], leading to increased fatty acid circulation and accumulation, which in turn triggers the proinflammatory NF-κB pathway in both animal models and cell culture studies, indicating persistent inflammation [55]. Moreover, HADH is crucial for the differential handling of stored lipids [56], highlighting its significant role in metabolic health and disease progression.

What series of reactions occur in AP patients with a BMI >30 due to alterations in HADH expression levels? In patients exhibiting a high BMI, the downregulation of HADH expression reduces fatty acid beta-oxidation, the primary pathway for fatty acid degradation, causing intracellular fatty acid accumulation. This metabolic alteration has profound implications for cellular physiology: (1) Fatty acid accumulation can lead to mitochondrial dysfunction, selectively inhibiting the active form of mitochondrial complex I. This suppression triggers necrotic cell death by releasing intracellular calcium and disrupting mitochondrial complexes I and V [57, 58]. These mitochondrial dysfunctions have extensive implications, disrupting cellular energy homeostasis and overall functionality. (2) The increase in free fatty acids subsequently triggers an increase in reactive oxygen species (ROS) production. ROS act as potent mediators of mitochondrial damage and tissue inflammation, contributing to the pathogenesis of AP [59]. (3) Elevated levels of fatty acids decrease the production of reduced glutathione, further compromising the cell's capacity to mitigate oxidative stress [60]. (4) The buildup of free fatty acids within cells can induce lipotoxicity, resulting in local and systemic consequences. This plays a role in the inflammatory response, multisystem organ failure, and necrotic acinar cell death in AP among patients with a BMI >30 [58]. Inflammation is a significant consequence of disrupted fatty acid metabolism and mitochondrial dysfunction. The upregulation of inflammatory mediators stimulated by elevated fatty acid levels implies a potential connection between HADH-related pathways and the onset of inflammatory conditions. Additionally, other studies have confirmed that the release of free fatty acids contributes to the exacerbation and severity of AP [61]. Based on previous research findings and our own findings, this study revealed that reduced HADH expression disrupts lipid metabolism, leading to the accumulation of FFAs, potentially playing an essential role in the pathogenesis of AP individuals with a BMI >30. This underscores the importance of conducting further investigations in future studies.

To validate this hypothesis, this study utilized external datasets concerning individuals with a high BMI and AP to corroborate the findings. The results were promising, as the expression patterns of HADH genes aligned with the research findings in both datasets. At this juncture, there are sufficient grounds to consider HADH as a potential biomarker with diagnostic value in the development of AP in patients with a BMI >30.

This study integrated MR with bioinformatics data analysis to explore the crucial role of HADH in patients with AP and a BMI over 30, particularly regarding the immune response. It can facilitate more comprehensive and in-depth genetic research, expedite biological discovery, and enhance personalized medical outcomes for this patient population, thereby laying a theoretical foundation for personalized precision treatment. It is important to acknowledge that these findings are primarily based on computational analyses and existing data. Further validation through animal and cell experiments is imperative to confirm and extend these observations. Animal models, such as rodent models with genetic modifications related to HADH expression or activity, can offer a more direct means of elucidating the causative relationship between HADH and AP in patients with a BMI >30. By manipulating HADH expression levels or activity in vivo, this study revealed the resulting effects on pancreatic function, inflammation, and immune responses. These experiments provide a deeper understanding of the mechanistic interactions and pathways involved. Furthermore, this study only conducted corresponding analyses on the existing dataset, and the specific role of HADH in disease progression and outcomes still needs further research.

In this study, a comprehensive approach was employed, leveraging MR, bioinformatics methods, and various machine learning algorithms to pinpoint HADH as a potential biomarker for AP in patients with a BMI >30, with a focus on lipid metabolism. These findings can empower clinicians to customize treatment strategies according to the diverse genetic profiles of BMI in patients with AP. Moreover, these findings pave the way for the development of medications specifically targeting HADH specifically to improve treatment efficacy and reduce side effects in patients with AP and a BMI exceeding 30.