Molecular links between long COVID and newly diagnosed diabetes: biological relationships and current ideas

The term long COVID, or postacute sequelae of SARS-CoV-2 infection (PASC), refers to as the association of SARS-CoV-2 infection with persistent health effects, with new symptoms occurring or relapsing after four weeks of acute infection (19). Although definitions of long COVID vary and many individuals report unexplained symptoms, research shows that 0–35% of COVID-19 survivors experience ongoing symptoms such as fatigue, cognitive impairment, or mental health issues including posttraumatic stress disorder (PTSD) and anxiety (20 –22). Despite millions continuing to suffer from long COVID (5, 23), no treatments with proven efficacy and no diagnostic tests or therapeutic solutions are currently available.

Long COVID is a complex, multisystem disorder affecting nearly every organ in the body, including the endocrine system (24, 25), cardiovascular system (15, 24), nervous system (25, 26), gastrointestinal system (27), musculoskeletal system (28), immune system (5).

Although there are potentially conflicting data, research has indicated sex differences in acute COVID-19 and long COVID outcomes. For instance, some authors reported that, during acute infection, males experience greater disease severity and mortality, whereas a larger percentage of females develop long COVID (21). Furthermore, epidemiological evidence and retrospective studies documented that COVID-19 was more likely to affect older males with comorbidities such as malignant tumors and diseases of the endocrine, digestive, respiratory, cardiovascular, cerebrovascular, and nervous systems (29, 30).

Such differences can be explained in terms of innate and adaptive immune responses that vary by sex. For example, Hamlin et al. uncovered multiple sex-specific immune pathways associated with long COVID (31). In pediatric patients, long COVID can cause unexpected health consequences such as multisystem inflammatory syndrome in children (MIS-C) (32, 33).

The host immune response at the time of acute illness plays a significant role in the pathogenesis of COVID-19. Marked immune dysregulation, including elevated expression of inflammatory mediators, is identified in acute-phase infections (34, 35). Similarly, researchers have discovered genes associated with coagulopathy, lung fibrosis, multi-organ damage, and long COVID-19 (36, 37).

It is now established that long COVID is not confined to respiratory tissues. Over 200 symptoms have been observed across nearly all organ systems in the human body (38), evidencing the complexity of this multisystem disease. Extensive research using immunological, transcriptomic, clinical, and virological data shows that the mechanisms of long COVID are largely non-exclusive. They may involve host cell dysfunction, sex-specific immune dysregulation, and persistent inflammation. For example, excessive immune activation may lead to several downstream pathologies, such as mitochondrial dysfunction, metabolic dysfunction, excessive inflammatory responses, neuronal inflammation, microglial activation, and other abnormalities (Figure 1). As a consequence, multiorgan damage may occur with diverse clinical manifestations. Collectively, viral antigen persistence and immune dysregulation are two major mechanisms that might drive long COVID (38 –43).

Indeed, the persistence of SARS-CoV-2 RNA in different tissues is strongly associated with the development of long COVID (44). For these reasons, tissue viral persistence could lead to long-term immunological perturbations, neurodegenerative diseases, and other chronic health complications (45).

Viruses of the family Coronaviridae share the same general genome organization of a single-stranded, positive-sense RNA genome with length ranges from 26 to 32 kilobases (46, 47). Thus, SARS-CoV-2 is an enveloped, positive sense, single-stranded RNA virus, that shares (about 79%) sequence identity with SARS-CoV-1 in their full-length genome sequences (48, 49). SARS-CoV-2 RNA (+) genome encodes at least 29 proteins, including nonstructural proteins, accessory proteins, and four structural proteins: the spike (S), envelope (E), membrane (M), and nucleocapsid (N) proteins (50) (Figure 2). Their large genome favors coronaviruses to tolerate a wide range of ecological niches, and various hosts as well. As the viral tropism relies on viral and host proteins, the evolution of viral proteins plays an indispensable role in coronaviruses, from host recognition to genome replication and evasion of the host immune infrastructure. Among the structural proteins, the S protein is produced as a dimer or trimer. It performs two major functions: attachment to host cell receptors and facilitation of virion–host membrane fusion with host cell membranes (51). While the M protein functions in viral assembly and drives morphogenesis, while the E protein encapsulates viral RNA (52). Another multidomain N protein performs several crucial functions, including binding, compacting, and packaging the viral genome (53). A concise summary of the biological roles of SARS-CoV-2 nonstructural proteins is provided in Supplementary Table 1.

The SARS-CoV-2 infection cycle requires host proteins (including a cohort of enzymes) and viral proteins. It mainly involves host cell recognition and entry into the host cytoplasm, viral genome replication and transcription, protein maturation, and release of viral particles (Figure 3). Viral gene expression and RNA synthesis are highly programmed and complex processes that occur in a coordinated spatiotemporal manner (Figure 3, Supplementary Table 1).

To shed light on the mechanistic links between viral tropism and pancreatic involvement, it is important to characterize pancreatic expression of ACE2, serine proteases, and other gateways for SARS-CoV-2 entry into different islet cell subtypes. In this section, we summarize expression profiles of the most common SARS-CoV-2 receptors in the pancreas and highlight gene and protein expression patterns associated with pancreatic β-cell pathophysiology.

A failure to maintain immune tolerance (i.e., the state in which the immune system is unresponsive to self-antigens) is the underlying cause leading to the development of autoimmune disorders, including T1DM. Thus, regulation of immune tolerance is critical to immune homeostasis (109). Regulatory immune cells, particularly CD4⁺CD25⁺FoxP3⁺ regulatory T cells (Treg cells), play a crucial role in maintaining and stabilizing immune tolerance via the production of powerful inhibitory cytokines (110). Of note, in the pancreatic islet niches, Tregs act as peacekeepers of immune tolerance to protect β-cells by inhibiting autoreactive β-cells and suppressing the differentiation and function of autoantigen-specific cytotoxic CD8⁺ T cells (111). Critically, viral infection can compromise immune self-tolerance. There is decades of evidence supporting the link between Epstein–Barr virus infection and multiple sclerosis. Similarly, compelling evidence suggests that SARS-CoV-2 compromises immune self-tolerance through mechanisms shared with other viruses, such as depleting Treg cell numbers or impairing their function via repression of FoxP3 expression (i.e., a key transcription factor of Treg cells) (112). Importantly, loss of the protective function of CD4⁺CD25⁺FoxP3⁺ Treg cells has been tied to T1DM (111, 113). Notably, Anindya et al. demonstrated that SARS-CoV-2 infection leads to dysregulation of immune homeostasis, including compromised self-tolerance, diminished Treg function, and Treg instability, which represents a potential root of T1DM onset (114). In line with this, several autoimmune mechanisms—including molecular mimicry, epitope spreading, and bystander activation—have been hypothesized as potential mechanisms by which SARS-CoV-2 might trigger T1DM (115).

Another crucial point is the potential mechanistic link between virus-induced IFN signaling and T1DM. Although the link is complex, both conditions share a chronic inflammatory state. Studies have revealed impaired type I interferon (IFN-I) responses in the blood of severe and critical COVID-19 patients. This delayed or exhausted interferon production may contribute to desensitization or a refractory state in IFN signaling, promoting ongoing viral replication and hyperinflammation (116). Consequently, this exacerbated inflammatory response, accompanied by high levels of IL-6 and TNF-α, may induce insulin resistance and hyperglycemia. Similarly, elevated IFN-I signaling may trigger both the initiation and progression of autoimmune diabetes. It is strongly hypothesized that IFN-α and IFN-γ exert direct effects on β-cells, leading to β-cell toxicity. Moreover, virus-induced IFN-γ can directly cause insulin resistance (117).

Collectively, this evidence indicates that SARS-CoV-2 infection may act as an environmental trigger for T1DM by directly damaging β-cells and disrupting the protective role of Treg cells within the pancreatic islet niche. Therefore, virus-induced autoimmune responses, systemic inflammation, immunologic injury, exhaustion of pancreatic β-cells, and stress-induced hyperglycemia may contribute to the progression of new-onset T1DM and T2DM.

To gain insights into the progression of T1DM and T2DM, it is important to understand the structural and functional abnormalities of pancreatic tissues before and after SARS-CoV-2 infection. Studies at the molecular level—including expression of antiapoptotic genes, endoplasmic reticulum stress–related genes, viral recognition pathways, and innate immune response genes in β-cells and α-cells—could also provide therapeutic strategies.

Mounting evidence shows that SARS-CoV-2 enters pancreatic cells and establishes infection, thereby associating with morphological, transcriptional, and functional changes of the pancreas. An early study by Yang et al. demonstrated SARS-CoV-2 infection of human pancreatic α and β cells (118). In this regard, pancreatic culture systems containing endocrine and exocrine cells, such as human induced pluripotent stem cell (iPSC)-derived cultures, have been widely used. Shaharuddin et al. used iPSC-derived pancreatic cultures to demonstrate deleterious effects of SARS-CoV-2 on normal molecular and cellular phenotypes. SARS-CoV-2 infection of the pancreas has also been confirmed in postmortem tissues from COVID-19 patients (119).

Pancreatic histopathology and immunofluorescence analyses revealed the presence of SARS-CoV-2 spike and nucleocapsid proteins in the islets and acinar epithelium of the acini (65, 105, 120). SARS-CoV-2 infection of islets causes both morphological and functional changes. Several studies reported increased numbers of insulin- and glucagon-positive cells, with a profound reduction in insulin-secreting granules in β-cells, suppressed insulin gene transcription, and impaired insulin secretion (16, 88, 120). Normal pancreatic cells exhibit ductal, acinar, or endocrine architecture, whereas SARS-CoV-2–infected cells show pancreatic tissue hyperplasticity, islet shrinkage, degeneration, and β-cell polyploidy. Mild lymphocytic infiltration of the endocrine and exocrine pancreas has been documented in some SARS-CoV-2–infected individuals with SARS-CoV-2 infection (65, 105). Such studies reported evidence supporting significant long-term morphological defects of the pancreas upon SARS-CoV-2 infection. For instance, histologic examination of pancreatic sections from postmortem tissues and infected pancreatic tissues showed thrombotic lesions, fibrosis, necroptotic cell death, immune cell infiltration, chronic pancreatitis, and focal acute pancreatitis (55, 65, 121). As demonstrated by the work of Steenblock et al., both intact islets and impaired islets were evident in COVID-19 patients without a previous diabetes diagnosis (65). In addition, examination of postmortem pancreatic tissue (from newly hyperglycemic patients with COVID-19 showed mild lymphocytic infiltration of pancreatic islets and pancreatic lymph nodes, presence of SARS-CoV-2–specific viral RNA, and several immature insulin granules (proinsulin) (122). Furthermore, crystalline morphology of β-cells, β-cell death, a substantial decrease in islet size, and a slight reduction in insulin expression have been observed (122). These findings highlight major structural and functional alterations—including β-cell degeneration, altered proinsulin processing, and β-cell hyperstimulation—among others that could eventually lead to disturbances in glucose homeostasis.

SARS-CoV-2 infection induces cellular trans-differentiation. For instance, while β-cells showed downregulated expression of insulin, higher expression of glucagon in α-cells and trypsin as a pancreatic acinar cell marker were observed upon SARS-CoV-2 infection (88). Similar findings by Wu et al. confirmed reduced pancreatic insulin levels and secretion, as well as β-cell apoptosis triggered by SARS-CoV-2 infection. Interestingly, phosphoproteomic pathway results showed similar apoptotic β-cell signaling between islets infected with SARS-CoV-2 and T1DM (16).

Cellular stress and increasing chemokine responses are hallmarks of SARS-CoV-2 infection. One notable mechanism through which SARS-CoV-2 perturbs key inflammatory responses is by hijacking the ribosomal machinery. For instance, as demonstrated in iPSC-derived pancreatic cultures, perturbed inflammatory responses—such as upregulation of the cytokine stromal cell–derived factor 1 (SDF-1) or CXC motif chemokine 12 (CXCL12)—were observed in the pancreas upon SARS-CoV-2 infection (119).

Metabolic stress is a critical factor that can lead to pancreatic β-cell dedifferentiation, thereby causing downregulation of ACE2 and other key proteins. For instance, high-fat diet (HFD)–induced insulin resistance and glucose intolerance are associated with decreased ACE2 expression. It is noteworthy that under HFD conditions, a higher percentage of dedifferentiated β-cells was detected in ACE2-knockout mice than in wild-type mice (123). Moreover, diet-induced obesity promotes delayed virus clearance, as demonstrated in HFD mouse models (124). Critical metabolic disorders have also been observed in adult nonhuman primates (NHPs) infected with SARS-CoV-2 (63). According to this study, major loss of β-cells and attenuated insulin expression were evident, associated with complex and severe pathological conditions such as islet amyloidosis and necrosis, activation of alpha-smooth muscle actin (α-SMA), and aggravated fibrosis with lower serum collagen. Furthermore, increased pancreatic inflammation and stress markers—intercellular adhesion molecule 1 (ICAM-1) and Ras GTPase–activating protein-binding protein 1 (G3BP1)—were observed (63). In line with previous findings, new-onset DM accompanied by pancreatic damage (e.g., generalized fibrosis with multiple vascular thrombi) was observed in SARS-CoV-2–infected NHPs (62).

Patients without preexisting diabetes can also develop persistent hyperglycemia as a result of viral-mediated pancreatic damage (125, 126). It is also possible that even without developing new-onset diabetes, immune cell infiltration and necroptotic cell death may lead to varying degrees of metabolic dysfunction due to SARS-CoV-2–infected pancreatic β-cells (65).

Genes such as Synaptotagmin 4 (SYT4), Per-Arnt-Sim (PAS) domain-containing protein kinase (PASK), Phospholipase C X domain 3 (PLCXD3), and Peroxisomal biogenesis factor 6 (PEX6) are linked to β-cell physiology or DM (127 –130). Interestingly, an elegant work by Müller et al. revealed significant dysregulation of these genes in uninfected and infected cultured human islets. SARS-CoV-2 infection resulted in downregulation of SYT4, PASK, PLCXD3, and PEX6, while several interferon (IFN)–stimulated genes (ISGs), such as IFITMs, OAS2, IFI27, and ISG15, were upregulated (120). Similarly, in other tissues SARS-CoV-2 elicits higher ISG transcript levels. For example, in an organ-specific manner (ovary, pancreas, and thyroid), downregulation of endocrine-specific genes such as IAPP, HSD3B2, INS, LEP, FOXE1, TSHR, and CRYGD has been identified in COVID-19 patients (131). Many molecules essential to islet function—including CD36 (regulates β-cell response to hyperglycemia and glucolipotoxicity), insulin receptor substrates IRS1 and IRS2, glucose transporters GLUT2, PPARG, and pancreatic and duodenal homeobox 1 (PDX1)—play key roles in the pathogenesis of diabetes (132, 133). Of note, SARS-CoV-2 infection provokes dysregulation of CD36, GLUT2, IRS1, IRS2, PDX1, and PPARG expression in the pancreas (134).

Given that destruction of β-cells is relevant to both T1DM and T2DM, it is important to understand different SARS-CoV-2 targeted cell types involved in β-cells physiology and function. For example, in vitro infection of human pancreatic islets revealed that SARS-CoV-2 targets adventitial cells and pericytes (135). Pericytes are mural cells located within the basement membrane of blood microvessels that supply trophic factors crucial for maturation and proper function of β-cell (136) and regulate islet capillary diameter and blood flow (137, 138). Pericytes in the pancreas and other organs, such as the central nervous system and heart, express ACE2 (61). Examination of living pancreas slices from nondiabetic individuals showed that SARS-CoV-2 recombinant spike S1 proteins impair ACE2 activity and induce islet pericyte and microvascular dysfunction (139). This study offers new insights into molecular mechanisms such as vascular dysfunction, activation of pericytes, and islet capillary constriction upon SARS-CoV-2 infection (139) (Figure 4). This confirms the diabetogenic actions of SARS-CoV-2 through pancreatic islet pericyte dysfunction and disturbance of glucose homeostasis.

SARS-CoV-2 profoundly affects endocrine organs/glands, such as the hypothalamo–pituitary–adrenal axis, thereby disturbing glucose metabolism. It can worsen glycemic control in individuals. Thus, new-onset hyperglycemia, also termed "stress hyperglycemia," observed in patients with severe COVID-19 illness could be significantly influenced by the stress response. The increasing levels of stress hormones such as catecholamines and cortisol can provoke insulin resistance and promote blood glucose elevation, consequently leading to poor outcomes in pre-diabetic individuals (140).

Follow-up studies at the community level would help determine the extent to which insulin resistance and β-cell dysfunction contribute to new-onset diabetes. For instance, one recent observational follow-up study reported a greater risk of new-onset DM due to exacerbated insulin resistance rather than β-cell dysfunction (141). Furthermore, studies focusing on homeostatic maintenance and turnover, regenerative mechanisms, islet maturation, and cell type specification may help dissect the underlying molecular mechanisms and provide new avenues for understanding SARS-CoV-2–associated new-onset diabetes.

The expression of ACE2 can be upregulated by inflammatory cytokines following SARS-CoV-2 infection in airway epithelium (142). The virus itself may also upregulate ACE2 expression, suggesting a positive feed-forward effect that enhances viral infection (143), potentially leading to increased inflammation and multi-organ damage in tissues such as the airways, intestines, pancreas, liver, and vasculature (Figure 5).

In vitro, SARS-CoV-2 infection induces expression of proinflammatory genes such as CCL2, CCL3, CCL5, CCL10, IL-1β, and IL-6 in human small intestinal epithelial cells derived from pluripotent stem cells (144). The inflammatory response in the gut may increase barrier damage and contribute to the gastrointestinal symptoms observed following infection.

Notably, a growing body of literature highlights the role of gut microbiota in various metabolic diseases. It is well established that SARS-CoV-2 infection and viral entry into enterocytes lead to gut microbiome dysbiosis, which can disrupt gut barrier integrity and promote inflammation, along with altering metabolite levels that may be linked to new-onset DM (144, 145). Such gastrointestinal damage may also result in malabsorption of vital nutrients, including vitamins, which play indispensable roles in metabolism.

Persistent gut damage caused by SARS-CoV-2 infection has been shown to induce long-term dysregulation of tryptophan (Trp) absorption from the intestine (146) potentially contributing to the drastic decrease in plasma Trp levels in long COVID patients (147). In addition, viral infection and type I IFN–driven inflammation also cause serotonin reduction through decreased intestinal absorption of the serotonin precursor Trp, platelet hyperactivation, and thrombocytopenia, all of which affect serotonin storage. Notably, serotonin reduction has been associated with PASC (146). Another study indicated decreased peroxisome biogenesis and enhanced peroxisome degradation as a result of increased IFN levels following SARS-CoV-2 infection and inflammation, revealing peroxisomes as essential regulators of macrophage-mediated lung inflammation resolution and tissue regeneration (148). Hence, since peroxisomes play an important roles in lipid metabolism and immunometabolism, dysregulation of their function can disrupt lipid balance, resulting in considerable cellular stress and immune dysfunction.

However, it is important to note that gut microbiota may be associated with T2DM either positively or negatively. For instance, a systematic review reported that the genera Bifidobacterium, Bacteroides, Faecalibacterium, Akkermansia, and Roseburia were negatively associated with T2DM (149). A meta-analysis in antibiotic-naïve COVID-19 patients showed reductions in beneficial (symbiotic) bacteria and increased levels of opportunistic pathogens, including Actinomyces viscosus, Clostridium hathewayi, and Bacteroides nordii, compared with controls (150). Recent findings indicate that gut microbiota alterations—due to various causes including SARS-CoV-2 infection—affect the metabolism of peptides and amino acids (e.g., tryptophan, phenylalanine, glutamate, citrulline), short-chain fatty acids (SCFAs), xenobiotics, and other metabolites, which may negatively influence infection rates, inflammation, and glycometabolic abnormalities, including T2DM (150, 151). Additionally, although not COVID-19–specific, microbiota changes in Clostridium species have been linked to predicted plasma metabolites involved in phenylalanine metabolism—such as phenylacetate, phenylacetylglutamate, and phenylacetylglutamine—which correlate with heightened cardiovascular risk and T2DM (152). A recent metabolomics study of 1,167 individuals identified blood metabolites, including lipid- and amino acid–related metabolites associated with impaired glucose control, with most DM patients showing altered gut microbiome interactions involving Hominifimenecus microfluidus and Blautia wexlerae via Hippurate (151).

Notably, long COVID has been linked to gut microbiome dysbiosis, which contributes to insulin resistance and T2DM. These findings highlight the critical effect of SARS-CoV-2 infection on the gut microbiota, leading to a decrease in both the total number of organisms and the variety of species present. Such microbiome dysbiosis and microbial translocation may contribute to metabolic derangements and the progression of new-onset diabetes.

A study conducted on glucose-utilizing tissues such as adipose tissue showed higher ACE2 expression than in lung tissue, which could be linked with ACE2-mediated viral replication and pathogenesis, including oxidative stress and inflammation in such tissues as well (58).

Researchers detected SARS-CoV-2 RNA in adipose tissue specimens (visceral, epicardial, pericardial, and subcutaneous) from autopsy samples of deceased patients with COVID-19 undergoing cardiothoracic surgery, along with inflammatory infiltrates, suggesting that the virus also infects inflammatory adipose tissue–resident macrophages and adipocytes (153). Several other research groups also reported detection of SARS-CoV-2 RNA in abdominal subcutaneous adipose tissue specimens, including 13 of 23 COVID-19 autopsy cases (154), 28 of 59 cases of human adipocytes (155) and experimental infection of adipose tissue in the hamster model of SARS-CoV-2 infection (156).

Bulk RNA-seq analysis revealed that lungs and adipose tissue specimens had various immune cells and inflammatory markers (e.g., IL-6) upregulated in COVID-19–infected obese K18 mice (n=4) as compared with COVID-19–infected normal or non-obese (n=4) control mice (157). The data from this model study support a link between inflammation and increased adipose mass or obesity-related COVID-19 infection, which in turn could be associated with glycometabolic disturbances such as insulin resistance. Interestingly, transcriptome analysis conducted on lung and adipose tissue specimens obtained from deceased COVID-19 patients also found that obese patients (n=3) had increased expression of genes and activation of inflammation-related pathways as compared with normal-weight COVID-19 patient tissues (n=2) (157).

A longitudinal study revealed that SARS-CoV-2 infection provokes systemic metabolic abnormalities, mitochondrial dysfunction of skeletal muscle, and abnormal skeletal muscle response to exercise (including exercise-induced myopathy and tissue infiltration of amyloid-containing deposits and leukocytes) (28).

In individuals with severe COVID-19 illness, long-lasting liver damage is observed, and ACE2 expression in cholangiocytes has been hypothesized as a potential cause of liver damage (158, 159). NRP-1 expression in various body tissues or cell types, including liver-resident cells such as sinusoidal endothelial cells and hepatic stellate cells, may also facilitate SARS-CoV-2 infection, thereby increasing liver damage and decreasing hepatic function (160).

Besides pulmonary and other complications, endothelial injury has been established as a primary finding in patients infected with SARS-CoV-2, and the vasculature has been clinically observed as one of the main trans-organ systems affected by SARS-CoV-2 infection. A postmortem histopathological study revealed viral inclusions in apoptotic endothelial cells, infiltration of inflammatory cells, microvascular endothelitis, and vascular complications (161). Furthermore, SARS-CoV-2 infection leads to complex and multifactorial endothelial cell activation, progressive loss of antithrombotic factors, and promotion of local pro-angiogenesis (162).