Letter to the Editor: Tumor necrosis factor-α-mediated β-cell death in acute pancreatitis-addressing etiology, cytokine crosstalk, and pyroptosis

Abstract

The study by Lizzi and Massimi published in the recent issue of the World Journal of Gastroenterology, commented on our recent study of tumor necrosis factor-α signaling and β-cell apoptosis following acute pancreatitis (AP). Their model places this pathway within a broader network of cellular stress responses. We suggest that the etiology of AP, interactions among inflammatory mediators, and non-apoptotic mechanisms should also be considered. Addressing these questions may clarify why post-pancreatitis diabetes mellitus develops. However, whether these pathways have therapeutic relevance remains to be determined.

Key Words: Tumor necrosis factor-α; Etiology; Acute pancreatitis; Inflammatory factors; Cell death pathways; Post-pancreatitis diabetes mellitus

Core Tip: This letter discusses three issues arising from the model proposed by Lizzi and Massimi: Etiology-specific differences in acute pancreatitis, cytokine crosstalk, and the potential involvement of pyroptotic pathways in β-cell injury.

TO THE EDITOR

We appreciate the thoughtful comments on our recent article[1] from Lizzi and Massimi[2]. This response addresses two central questions: Whether tumor necrosis factor-α (TNF-α)-mediated β-cell death in acute pancreatitis (AP) may contribute to abnormal glucose metabolism (AGM) and whether this process is influenced by AP etiology, cytokine crosstalk, and pyroptotic pathways. Lizzi and Massimi[2] proposed that TNF-α-mediated β-cell injury cannot be explained by apoptosis alone. Their model links TNF-α signaling to nuclear factor kappa B (NF-κB) activation, oxidative stress, endoplasmic reticulum (ER) stress, and cytokine interactions. Focusing specifically on AP, Alam and Abbas[3] proposed TNF-α-driven pancreatic β-cell apoptosis as a potential link between inflammation and metabolic dysfunction. We emphasize that this letter is conceptual and does not present new original data. We therefore invite further consideration of whether this integrated model could be extended to account for etiology-specific differences, cytokine crosstalk, and pyroptotic pathways in AP-related β-cell injury.

First, the proportions of AP cases attributable to biliary/gallstone, alcoholic, and hypertriglyceridemic (HTG) etiologies vary[4-6], and the prevalence of HTG-AP is increasing[4]. HTG-AP also has distinct clinical features, including a poorer prognosis and higher mortality, and therefore warrants greater clinical attention[7]. Compared with patients with AP of other etiologies, those with HTG-AP had significantly higher levels of inflammatory markers, including white blood cell, lymphocyte, neutrophil, and platelet counts, as well as C-reactive protein[8]. Additionally, higher levels of neutrophils and C-reactive protein were associated with the development of post-pancreatitis diabetes mellitus (PPDM), suggesting a possible link between heightened inflammation and subsequent islet dysfunction[8].

Different AP etiologies involve distinct pathophysiological mechanisms. Obesity can increase the risk and severity of AP through mechanisms including gallstone formation, hypertriglyceridemia, and lipotoxicity[9]. In obesity-related AP, serum exosomes promote M1 macrophage polarization, thereby increasing the release of interleukin (IL)-1, IL-6, and TNF-α[10]. Elevated free fatty acids can induce ER stress in pancreatic β cells by enhancing calcium influx, thereby contributing to β-cell dysfunction and apoptosis[11]. Tauroursodeoxycholic acid (TUDCA), an endogenous hydrophilic bile acid, is significantly reduced in AP models, and its supplementation alleviates pancreatic injury[12]. TUDCA also alleviates ER stress, reduces β-cell apoptosis, and promotes cell survival[13]. Studies of alcohol-related pancreatitis have shown that, in mice with impaired nuclear factor erythroid 2-related factor 2 function, ethanol exposure induces severe pancreatitis and pancreatic β-cell injury, accompanied by increased insulin secretion and/or leakage that may contribute to hypoglycemia[14]. These etiological differences may result in distinct patterns of β-cell injury and differential susceptibility to TNF-α-mediated apoptosis.

Although different AP etiologies are associated with distinct inflammatory profiles, it remains unclear whether TNF-α expression or activity varies by etiology and whether such variation influences the development of PPDM. Olesen et al[15] reported that PPDM may develop at any stage of the disease course and that progressive metabolic alterations may be influenced by both the timing and severity of pancreatic injury. Subsequently, Olesen et al[16] emphasized that pancreatitis progression may be discontinuous, suggesting that PPDM need not develop continuously or inevitably. Together, these findings raise the possibility that a prediabetic state may precede the diagnosis of PPDM. We therefore consider AGM to represent an early metabolic phenotype associated with PPDM, and propose that it may provide a potential risk marker and therapeutic window.

Our results showed no significant association between AP etiologies and the risk of AGM (χ² = 4.893, P = 0.180; n = 149, 70, 102, and 48 patients with biliary, alcoholic, HTG, and other etiologies, respectively; Table 1 in our previous report[1]). This negative result should be interpreted cautiously, as it may reflect the relatively small sample size and the single-center design. Because etiology-specific inflammatory profiles may differentially affect β-cell damage, β-cell dysfunction, and PPDM phenotypes, future studies should incorporate etiology-stratified sampling, standardized timing of glucose assessment, adjustment for disease severity, and longitudinal ascertainment of PPDM. We speculate that etiology-associated inflammatory patterns could contribute to risk stratification, help distinguish PPDM subtypes and inform personalized treatment, although these possibilities require further clinical validation. Beyond AP etiology, other plausible contributors to PPDM, such as AP severity, recurrent disease, pre-existing dysglycemia, and pancreatic exocrine dysfunction, should also be considered. These factors may interact with inflammatory responses to contribute to β-cell injury, which may ultimately result in PPDM.

Second, the inflammatory network involved in β-cell injury extends beyond TNF-α. Cytokines such as IL-1β and interferon-γ (IFN-γ) may also contribute to β-cell damage and dysfunction through signaling crosstalk[17-19]. Studies have shown that IL-1β contributes to β-cell apoptosis through activation of NF-κB and mitogen-activated protein kinase (MAPK) pathways[17,20], whereas IFN-γ promotes β-cell death via NF-κB activation and induction of apoptosis[21]. TNF-α may act through signaling pathways distinct from those of these two cytokines. In type 1 diabetes and type 2 diabetes models, IFN-γ synergizes with TNF-α or IL-1β to promote β-cell death[19,22] and can upregulate TNF-α production[23]. Mechanistically, the combination of IFN-γ and IL-1 has been reported to stimulate inducible nitric oxide synthase expression in pancreatic islets, with the resulting increase in nitric oxide production contributing to islet cell destruction[24]. Whether such interactions also occur in AP-related β-cell injury remains to be elucidated. As summarized in Figure 1, solid lines and arrows denote pathways established in AP-related β-cell injury, dashed lines and arrows denote pathways demonstrated in non-β-cells or β cells under non-AP conditions but not yet validated in this condition, and dash-dot lines and arrows denote hypothetical pathways.

Figure 1 Proposed conceptual model of inflammatory cytokine-mediated β-cell stress during acute pancreatitis. Etiological factors such as biliary/gallstone, alcohol, and hypertriglyceridemia can lead to trypsinogen activation, thereby inducing acute pancreatitis (AP). AP induces a strong inflammatory milieu that triggers the release of substantial amounts of inflammatory cytokines from acinar cells and/or infiltrating immune cells, including tumor necrosis factor-α (TNF-α), interleukin (IL)-1β, and interferon-γ (IFN-γ). TNF-α can induce pancreatic islet β-cell death through the Bax/Bcl-2/caspase-3 apoptosis pathway and/or the hypothetical caspase-8/caspase-3/GSDME pyroptosis pathway (extrapolated from non-β-cells or β cells under non-AP conditions and not yet validated in AP-related β-cell injury). IL-1β is hypothesized to induce β-cell apoptosis through activation of nuclear factor kappa B (NF-κB) and/or MAPK signaling pathways. IFN-γ is hypothesized to induce β-cell apoptosis through activation of NF-κB. Damage to β cells induced by multiple inflammatory cytokines ultimately results in reduced insulin secretion, leading to abnormal glucose metabolism and increasing the risk of post-pancreatitis diabetes mellitus. Solid lines and arrows indicate direct evidence, dashed lines and arrows indicate extrapolated evidence, and dash-dot lines and arrows indicate hypothetical evidence. MAPK: Mitogen-activated protein kinase; AGM: Abnormal glucose metabolism; PPDM: Post-pancreatitis diabetes mellitus; GSDME: Gasdermin E.

Third, in addition to apoptosis mediated by the Bax/Bcl-2/caspase-3 pathway, pyroptosis has been implicated in pancreatic exocrine cells in AP, including ductal epithelial cells (HPDE6-C7) and acinar cells[25-27], but direct evidence in AP-related β-cell injury is still lacking. TNF-α-induced pyroptosis has also been reported in diseases other than AP. Gao et al[28] reported that TNF-α induces pyroptosis in diabetic liver injury through the caspase-8/caspase-3/GSDME pathway. Tan et al[29] showed that TNF-α promotes skeletal muscle pyroptosis by activating the tumor necrosis factor receptor 1 (TNFR1)/NOD-like receptor family pyrin domain-containing 3 (NLRP3)/caspase-1/gasdermin D (GSDMD) pathway. In mice with type 2 diabetes, excessive ethanol exposure induces β-cell pyroptosis via NLRP3/GSDMD activation[30], indicating that β cells can undergo pyroptosis under certain pathological conditions, although this was observed in type 2 diabetes rather than AP. In addition, Yang et al[31] developed human pluripotent stem cell (hPSC)-derived vascularized macrophage-islet (VMI) organoids and provided mechanistic and functional evidence that proinflammatory macrophages can induce β-cell pyroptosis via the tumor necrosis factor superfamily member 12 (TNFSF12)-tumor necrosis factor receptor superfamily member 12A (TNFRSF12A) pathway. This is an hPSC-derived VMI organoid model, rather than primary human islets or islets from patients with AP. Collectively, these findings indicate that TNF-α can induce pyroptosis in non-β-cells under non-AP conditions. In addition, under non-AP conditions, β cells can undergo pyroptosis via the NLRP3/GSDMD pathway in type 2 diabetes models or via the TNFSF12-TNFRSF12A pathway, neither of which involves the TNF-α/TNFR1 pathway. However, whether TNF-α/TNFR1 directly induces pyroptosis in human primary pancreatic β cells in AP, and whether this pathway intersects with, operates in parallel to, or diverges from TNF-α-driven apoptosis, remain unknown. For now, this represents a working hypothesis rather than a confirmed mechanism in AP-related β-cell injury, and further basic and clinical studies are needed for validation.

Notably, Zhang and Liu[32] acknowledged the involvement of the Bax/Bcl-2/caspase-3 pathway in AGM progression after AP. They also highlighted several unresolved questions, including the upstream regulation of TNF-α, the distinct roles of TNFR1 and tumor necrosis factor receptor 2, and the potential crosstalk with NF-κB and MAPK pathways. These perspectives are consistent with the broader view of Lizzi and Massimi that TNF-α may act as an upstream coordinator of convergent stress pathways, although the precise contributions of NF-κB and MAPK signaling to TNF-α-mediated β-cell injury in AP remain unclear. Based on the above rationale, we propose a hypothetical integrative framework in which AP etiology, cytokine signaling, and cell death pathways may interact to generate distinct PPDM phenotypes.