New-onset diabetes in COVID-19: The molecular pathogenesis

Summary Diabetes mellitus (DM) is still a challenging metabolic disease worldwide. In the current situation, the world is facing a COVID-19 pandemic due to SARS-CoV-2 infection. DM is one of the comorbid conditions that can worsen the severity of the COVID-19 condition. Surprisingly, SARS-CoV-2 infection can induce new-onset diabetes, a condition in which acute hyperglycemia occurs and may develop into a complication in nondiabetic patients. Angiotensinconverting enzyme 2 (ACE2) is a crucial entry factor for SARS-CoV-2 infection. ACE2 will bind to the spike protein of SARS-CoV-2, potentially initiating a damaging process in many tissues in the human body, including metabolic tissues. This mechanism suggests a potential role of ACE2 in the pathogenesis of diabetes since ACE2 has been proven to localize in essential metabolic tissues, one of which is the acini and islets part of the pancreas. This interrelated ACE2 in COVID-19 and DM is thought of as the mechanism that induces new-onset diabetes in COVID-19 patients. This review will thoroughly describe the current findings and theories regarding the molecular mechanism of SARS-CoV-2-induced new-onset diabetes and the possible therapeutic intervention.


Background
D iabetes Mellitus (DM) is still a major metabolic problem and the leading cause of morbidity and mortality worldwide. In 2014, World Health Organization (WHO) reported that 422 million adults had diabetes and its prevalence increases yearly [1]. In 2020, we were also challenged by the COVID-19 pandemic. COVID-19 is a disease due to the infection of Severe Acute Respiratory Syndrome-Coronavirus-2 (SARS-CoV-2) [2]. WHO reported that 83 million with COVID-19 confirmed positive cases and about 1.8 million confirmed death [3]. COVID-19 symptoms vary from mild to severe; about 90% of patients showed more than one symptom, and the three most prevalent symptoms are fever, fatigue, and cough [4]. Diabetes has been identified as a risk factor for many infection cases [5,6]; thus, it is suggested as a comorbidity that increases the severity of the COVID-19 infection [7e10].
Uniquely, some cases have been reported with newly diagnosed DM in COVID-19 positive cases without any history of DM [11,12]. Alsadhan et al. reported five patients admitted to the hospital with diabetic ketoacidosis (DKA) and positive result on real-time reverse transcription-polymerase (RT-PCR) COVID-19. Three of them had a DM history, and the others were diagnosed with new DM after being admitted to the hospital with a high level of HbA1c [11]. Another case report reported three patients, one diagnosed with DKA and positive for COVID-19. Interestingly, the other two patients showed classic DM symptoms like polydipsia and polyuria post-infected with COVID-19 [12]. In this condition, COVID-19 infection is aggravated by the acute hyperglycemia onset, which, if not treated properly, could potentially lead to fatal complications in the patient without a history of diabetes. With such unpredictable and rapid disease progression, this phenomenon has become quite a unique and urgent concern that needs to be unveiled. This review will discuss the new onset of DM in COVID-19 infection more profoundly and show how COVID-19 and DM interact with each other in the molecular aspect.

The role of ACE2 receptors in SARS-CoV-2 infection
Angiotensin-converting enzyme 2 (ACE2) has long been a key receptor for the SARS coronavirus. Becoming the first homolog of ACE with a homology sequence of 42%, ACE2 was first found in human heart failure [13]. SARS-CoV and SARS-CoV-2 utilize ACE2 as an entry receptor by binding it with surface protein S and may partly explain the pathogenesis and predilection of COVID- 19 [14e18]. The identification of ACE2 as an entry receptor for SARSCoV-2 was primarily facilitated by its similar role in SARS-CoV, which was revealed in 2003. Using the fusion protein technique, Li et al. unveiled that SARS-CoV efficiently bound ACE2 through S1 protein, and the soluble ACE2 blocked S1 domain-ACE2 interaction [19e22]. However, soluble ACE1 did not produce a similar result [21]. It also revealed that the anti-ACE2 antibody blocked SARS-CoV replication in Vero E6 cells from African green monkeys, but antiACE1 had no effect [21]. These results provided a strong foundation for identifying the SARSCoV-2 receptor.
Accordingly, ACE2 also became the primary receptor of SARS-CoV-2, which was revealed through extensive in vivo studies [23e27]. Transgenic mice with ACE2 deficiency had much less viral load and viral replication than the control mice [23,25]. These mice also experienced milder pulmonary alterations compared to the wild-type mice. Additionally, mice with human ACE2 overexpression also developed a higher rate of severe symptoms, which resemble human patients. Interestingly, the symptoms worsened when the mice were only injected with SARS-CoV-2 spike protein [27]. Consistently, the administration of recombinant soluble ACE2 effectively blocked the interaction between spike (S) protein and ACE2, highlighting its therapeutic potential for both SARS and COVID-19 [26]. Together, all of this evidence suggests the pivotal role of ACE2 in SARS-CoV-2 infection and, possibly, pathogenesis (see Fig. 1, illustration adapted from Pang et al. [26]).

Spike (S) protein and cellular proteases mediate SARS-CoV-2 entry
The spike (S) proteins of SARS-CoV and SARS-CoV-2 are similar to 76.5% similarity in amino acid sequences [28]. SARS-CoV-2's S protein has 1273 amino acids with two crucial domains, referred to as S1 and S2. Other important parts of the S protein are the 19 AA Nterminal, which serves as a signal peptide, and the C-terminal's short cytoplasmic and short transmembrane domains. The S1 can be divided into an N-terminal domain (NTD) and a Cterminal domain (CTD); both serve as receptorbinding domains. In both SARS-CoV and SAR-SCoV-2, these domains recognize ACE2, which serves as the virus's entry receptor. Despite the difference between the S protein of both viruses, their 3D structure is similar, which underlies the function and receptor's similarity [14,15,29].
In order to facilitate its entry, the S protein is primed by cellular proteases such as endosomal cysteine proteases (cathepsin B and L) and transmembrane serine protease 2 (TMPRSS2) [15,30e33]. This process is very similar to the cellular entry of SARS-CoV. Proteolysis is essential for SARS-CoV-2 entry, and both S1 and S1 proteins need to be cleaved to initiate the viral entry process [34]. Interestingly, the cleavage site of SARS-CoV-2 is slightly different from SARS-CoV with a new, conserved insertion sequence between S1 and S2, which is recognized by Furin, a kexin-like subfamily of proprotein convertases [35]. Another difference is the arginine residues found in the S1/S2 cleavage site, but the importance of these differences needs to be investigated [36].
The importance of endosomal cysteine proteases is reported in several studies that showed that modification of endosomal pH inhibited SARS-CoV-2 entry, which is likely due to the inactivation of endosomal proteases [15]. On the other hand, camostat mesylate, a TMPRSS2 inhibitor, only partly inhibited SARS-CoV-2-S cellular entry. Finally, the viral entry is entirely blocked if TMPRSS2 and endosomal cysteine proteases are inhibited [15].
TMPRSS2 is important for viral fusion protein activation during cellular entry, specifically by cleaving and activating subunit S1, facilitating viral attachment to the target cell plasma membrane [15,30,31]. Pathologically, both TMPRSS2 and ACE2 are expressed in the lungs. TMPRSS2 is expressed mainly in sub-segmental bronchial branches and lung tissue, while ACE2 is mainly found in subsegmental bronchial branches by transient secretory cell types [32,37]. Their colocalization is important and highlights the propensity of SARS-CoV-2 in this region. Colocalization of TMPRSS2 and ACE2 is also essential for effective infection of SARS-CoV-2, which enhances the efficiency of viral cellular entry due to proteolysis of the ACE2-protein S complex [37]. The importance of TMPRSS2 has been demonstrated in vivo studies in which.
TMPRSS2 deficiency in mice reduced viral particles in the lungs [38]. Also, it was reported that ADAM17 could also cleave ACE2, and it competed with TMPRSS2 [39]. Therefore, this evidence indicated the potential protective effect of ADAM17. However, with TMPRSS2, ADAM17 also regulates the ectodomain shedding of ACE2, which supports SARS-CoV-2 entry through endocytosis [40]. Therefore, further investigations are needed to delineate the exact role of ADAM17 in SARS-CoV-2 infection.
Overall, evidence shows the pivotal role of protein S priming by host proteases and inhibiting these proteases may hold a clue for COVID-19 therapy. Protein S cleavage allows viral fusion and entry to the cells, initiating infection and the viral reproduction process. These processes are also crucial in the pathogenesis of COVID-19-induced new-onset diabetes, which will be explained in the next section of this review.

ACE2: The link between SARS-CoV-2 and the key metabolic tissues
ACE2 is a unique, newly found enzyme that plays an important role as a compensatory enzyme in the pathogenic process of diabetes. Besides its wellknown expression in the respiratory tract, ACE2 also presents in essential metabolic tissues such as the pancreas, liver, adipose tissue and kidney. ACE2 is localized in the acini and islets part of the pancreas, similar to ACE distribution [41]. A recent study suggested that ACE2 expression is slightly higher in the pancreas than in the lungs. Additionally, its expression occurs in both the exocrine and endocrine glands of the pancreas [42]. In the liver, ACE2 presents in hepatocytes, where it was found to be elevated in hepatic fibrosis and hypoxic condition of the liver, indicating its compensatory part for such fibrogenic diseases [43,44]. While in the kidney, ACE2 collocates with ACE on the apical surface of the proximal tubules and glomerulus [45]. Its presence in the vital metabolic tissues contributes to the normal regulation of the tissue renin-angiotensin system (RAS) signaling pathway and maintenance of the metabolism homeostasis (see Fig. 2, illustration adapted from Battle et al. [46]).
ACE2 inhibits angiotensin-II upregulation by degrading angiotensin-II into the primary products Ang-(1e7), which could counter the debilitating effects of RAS hyperactivity such as hyperglycemia, hypertension, cardiac dysfunction and fibrosis [47]. While angiotensin-II has vasoconstriction, pro-oxidant and inflammation effects, the products of ACE2 act on the Mas receptor to counteract such effects by inducing vasodilation, prostaglandin release and inhibition of norepinephrine secretion [48e50]. The vasodilatory effect is thought to result from the modulation of NO release by the Akt pathway by the angiotensin-(1e7). The NO modulation leads to compensatory impacts such as increased blood flow of the islet vessels during demanding functional conditions such as obesity, diabetes or merely high blood glucose peak [47].
Moreover, ACE2 also has an inhibitory effect on damaging islet factors, reactive oxygen species (ROS) and TGF-b. Previous studies imply that during the state of hyperglycemia, ROS has produced in pancreatic b-cells through the activation of NAD(P) H oxidase (NOX). The activation of NOX is induced by angiotensin-II and AT1 receptor interaction, eventually leading to pancreatic b-cells dysfunction [51]. Through its degrading angiotensin-II mechanism, ACE2 could prevent this damaging process and maintain the pancreatic b-cells morphology [52]. In line with that, ACE2 also preserves islet structure through the blockade of TGF-b by RAS inhibition. All these protective and compensatory mechanisms supposedly prevent islet fibrosis and function loss. Despite its lack of effect on basal insulin secretion, RAS inhibition by ACE2 could protect b-cells from damaging factors, thus improving insulin synthesis and secretion [47].
The evidence of the presence of ACE2 in essential metabolic tissues, especially the pancreas and its key role in SARS-CoV-2 infection shows a strong link that may lead to pancreatic injury hyperglycemia episodes or, even worse, new-onset diabetes. However, its pathogenesis of pancreatic damage remains controversial, and it is essential to uncover possible therapeutical intervention purposes. The established theory of ACE2 as a key entrance of SARSCoV-2 indicates a more complicated possibility of the exact role of ACE2 in diabetes since it could act as a double-edged sword [47e53]. On the one hand, ACE2 expression is favorable for its protective mechanism in acute lung injury and compensatory effects in diabetes. However, on the other hand, its elevated expression may also facilitate more coronavirus entry into the host cells.
In response to that evidence, more researches are ongoing to elucidate a more explicit pathogenic process. Liu et al., in their cohort research, showed that 1e2% of nonsevere (without comorbidity and asymptomatic) and 17% of severe (with comorbidities and presenting symptoms) COVID-19 patients  had a pancreatic injury [42]. Immunohistochemical staining in their study showed that both endocrine and exocrine glands of the pancreas expressed ACE2 quite significantly in COVID-19 patients. It was also suggested that patients in the study might have been experiencing pancreatic injury even before admission, indicating the coronavirus's rapid attack on the pancreatic cells. This acute damaging process leads to the acute onset of hyperglycemia, which becomes one possible reason for the higher risk of death in SARS-CoV-2 infection [42]. This result is also in line with Yang et al. report that SARS-CoV-infected patients presented with hyperglycemia, which might be resulted from the damaging process of the pancreatic islets through ACE2 [53]. The idea of COVID-19-induced newonset diabetes is still very novel, and evidencebased theories are still limited. Nonetheless, further research has been undertaken better to understand the unique mechanisms and well-established therapeutical interventions.

SARS-CoV-2-induced new-onset diabetes
New-onset diabetes has been long established for more than a decade [54], and in recent days, it is coming to the surface again due to COVID-19. According to the term itself, new-onset means the symptoms occur for the first time in people without any history of diabetes. If left untreated, the symptoms manifest so acutely that they may develop into fatal complications such as ketoacidosis and hyperosmolarity [55,56]. Several conditions can induce this condition, including organ transplantation [57], the use of hypertensive drugs, thiazide diuretics and beta blockers [58], as well as a severe infection [53,55,56]. In the case of SARS-CoV-2 infection, however, diabetes and COVID-19 have a bidirectional relationship. Diabetes could exist previously as a comorbid and increase the risk of severe COVID-19, but it could also present for the first time as an acute onset. The latter case will require aggressive treatment, and its disease progression depends on the patient's clinical status [59].
Several studies illustrate a possible link between new-onset hyperglycemia and the severe coronavirus disease 2019 (COVID-19). Interestingly, this new-onset hyperglycemia is not associated with other risk factors, such as obesity, prediabetes, diabetes mellitus, or corticosteroid use [60]. Another finding by Li et al. states that COVID-19 patients who develop new-onset diabetes are known to have a higher mortality risk than COVID-19 patients who have had a history of diabetes or hyperglycemia [61]. Also, evidence regarding a high prevalence of diabetic ketoacidosis and hyperosmolarity has been documented in patients with COVID-19. Case reports suggest that COVID-19 can accelerate diabetic ketoacidosis (DKA) in subjects with new-onset hyperglycemia (diabetes) or pre-existing diabetes mellitus [62]. Early identification of DKA symptoms is needed to improve the prognosis of DKA related to COVID-19 [62]. Table 1 [63e65] summarizes the characteristics of new-onset diabetes in patients with COVID-19 that have been reported in several case reports.
However, the specific metabolic complications of COVID-19 are still not well defined. Therefore, an international diabetes research group initiated the CoviDIAB Project to conduct global records of diabetes patients related to Covid-19 (covidiab.edendrite.com). The purpose of recording this data is to define the phenotype of new-onset diabetes in patients with COVID-19. This condition is determined based on hyperglycemia, confirmed COVID-19, previous negative diabetes history, and a history of normal HbA1c levels. The registry will also be expanded to allow records of patients with preexisting diabetes who later present with severe acute metabolic disorders. So, this data is expected to discover the epidemiology and pathogenesis of diabetesrelated to COVID-19 and obtain instructions on the right treatment choice for patients [59].

Diabetic ketoacidosis as a possible complication of new-onset diabetes in COVID-19
Diabetic Ketoacidosis (DKA) is a complication that can cause morbidity and mortality in people with diabetes mellitus. DKA generally occurs due to decreasing insulin levels in the blood, which causes a decrease in glucose use and uncontrolled lipolysis, which in turn causes an excessive increase in ketone bodies and acidosis. This insulin deficiency condition occurs due to decreased secretion by pancreatic beta cells or increased insulin requirements triggered by infectious stressors and sepsis. The study results by Ahuja et al. stated that the strongest predisposing factor for acute DKA attacks was infection compared to other predisposing factors such as an inadequate insulin regimen, early presentation, or other unknown reasons [66].
As with many other diseases, COVID-19 could affect DKA patients by increasing the production of stress hormones and stimulating cytokines. Severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) utilizes binding to the angiotensin-converting enzyme 2 (ACE2) receptor on infected cells' membrane to enter the body's cells as a viral complex. ACE2 is found in many organs, such as the lungs, intestinal tissue, kidneys, heart and pancreas. ACE2 will convert angiotensin II to angiotensin I. The wide expression of ACE2 in these organs may explain the clinical symptoms of SARS. When talking about DKA that occurs in patients with COVID-19, it is known that ACE2 is expressed in the endocrine part of the pancreas. This evidence supports the statement that SARS-CoV-2 can enter the islet of the pancreas using ACE2 as its receptor. Therefore, a possible mechanism that plays a role in the development of DKA is the spread of ACE2 receptors by SARS-CoV-2 during this virusehost interaction, which can cause damage to pancreatic beta cells and subsequently interfere with their function. Furthermore, insulin deficiency can occur, leading to the development of acute diabetes [67,68].
In addition to direct beta cell damage, ACE2 expression on the surface of the pancreas is down regulated along with endocytosis of the ACE2-virus receptor complex. Reduced ACE2 expression can increase angiotensin II concentration, which cannot be converted to angiotensin I. In turn, the condition can inhibit insulin secretion [10,69]. The interaction between the virus that causes COVID-19 and the renin-angiotensin-aldosterone (RAAS) system may explain the pathophysiology that underlies DKA. These two factors are likely the basis of the acute deterioration of pancreatic beta cell function and the trigger for DKA in patients with COVID19. Studies on whether the nature of these changes is permanent or temporary are still to be carried out.

Possible therapeutic intervention
According to the latest studies, there are several recommendations regarding antidiabetic agents in COVID-19 patients. As discussed earlier, patients with COVID-19 can experience acute hyperglycemia. Clinicians must carry out glycemic control quickly, precisely and effectively to deal with this condition. Therefore, it is necessary to be careful in choosing the therapeutic modality based on its potential effectiveness and side effects. In their respective reviews, Lim et al. and Drucker have recommended glucagon-like peptide-1 receptor agonists (GLP-1Ras) for COVID-19 patients with mild to moderate symptoms because these agents can reduce glucose levels as well in outpatients [70,71]. However, the study results still do not support the recommendation to use this modality as a substitute for insulin in critically ill patients with type-2 diabetes mellitus and COVID-19, especially if therapy must be started in severe conditions. GLP-1 receptor agonists (GLP1-RAs) or incretinmimetics provide pharmacological levels of exogenous GLP1, which, analogous to the incretin hormone, have the effect of losing weight, inhibiting the release of glucagon, inhibiting appetite, and slowing gastric emptying [72,73]. GLP-1RAs have broad anti-inflammatory action when studied in animals with inflammation. This agent can also reduce systemic inflammatory biomarkers in human subjects with type 2 diabetes mellitus and obesity [74]. Several studies have shown that GLP-1RAs can reduce lung inflammation, decrease cytokine production and maintain lung function in mice with experimental lung injury [75e77]. GLP-1RAs have been shown to reduce pulmonary type 2 immune cytokine responses and lung damage levels in mice with respiratory syncytial virus (RSV) infection isolated from a hospitalized infant with severe lower respiratory tract infection and bronchiolitis [78]. Liraglutide, a GLP-1Ras, has a good safety and effectiveness profile when used as acute control of perioperative blood glucose in adult subjects undergoing elective cardiovascular surgery [79]. Also, liraglutide has been reported to improve cardiovascular outcomes in diabetic patients. It has a minimal risk of causing hypoglycemia, so it would be great if the administration of this agent could be further investigated in COVID-19 patients with diabetes [80].
Furthermore, based on Lim et al. [70] and Drucker's [71] recommendation, the use of insulin can be suggested. Insulin has become the glycemic control agent of choice for hospitalized COVID-19 patients, and its use is mandatory for critically ill patients. In its guideline, the American Diabetes Association (ADA) states that basal insulin or basalcorrected insulin regimen plus a bolus is a therapeutic option for hospitalized patients who are not seriously ill. Meanwhile, continuous intravenous insulin infusion is becoming a more recommended treatment for critically ill patients in the ICU. The expected target blood glucose for critically ill and non-critically ill patients ranges from 140 mg/dL to 180 mg/dL (7.8e10.0 mmol/L) [81].
Regardless of the ADA recommendations, although insulin treatment is the choice for diabetic patients with severe COVID-19 [81], a study in Wuhan, China, reported a worse prognosis based on clinical and laboratory data in patients using insulin than with patients using metformin [82]. However, these results should still be examined with caution because of the possible confounding, as insulin treatment is generally used in more severe diabetes patients. Other research supports this hypothesis that insulin infusion is an effective method to achieve the expected glycemic control and could reduce the severity and mortality in diabetic patients with COVID-19 [83].
Several studies have also shown that insulin administration can reduce urine ACE2, kidney ADAM-17 and kidney ACE2 in type-1 diabetes mouse models [84,85]. Insulin is known to act as an immunomodulatory agent and an additional antiinflammatory agent. These roles include blocking the NF Kb signaling pathway, reducing TNF-a levels and disrupting neutrophil chemotaxis [86].
Palermo et al. reviewed the recommendations for DKA treatment in COVID-19 patients [87]. The subcutaneous insulin regimen is the primary modality emphasized in the article. Blood glucose and ketone bodies in COVID-19 patients with hyperglycemia should be monitored regularly [88]. There are no specific guidelines regarding fluid and electrolyte management in patients with COVID-19 and diabetes mellitus. However, several articles can be referred to for management considerations [89,90].
Preclinical studies report the anti-inflammatory role of metformin, wherein metformin can reduce inflammatory biomarkers' levels in the circulation of patients with type-1 diabetes mellitus [91]. In a Chinese study comparing hospital mortality among COVID-19 patients with diabetes, the hospital mortality rate was significantly higher in patients who did not receive metformin than in those who received metformin (12.3% vs.2.9%; P ¼ 0.01) [92]. However, these findings may have a selection bias because patients with severe respiratory problems cannot receive metformin. When discussed from a molecular perspective, 50-AMP-activated protein kinase (AMPK) is the main effector of metformin's pharmacological action. This molecule appears to have a role in regulating the stability and expression of ACE2. Metformin can increase the expression of ACE2 and phosphorylation to the Ser680 residue in HUVEC cells. In addition, through AMPK, metformin also mediates ACE2 phosphorylation, thereby increasing the stability of ACE2. This process occurs through the inhibition of ubiquitination and degradation of its proteasomes. Therefore, theoretically, metformin might increase the amount of ACE2 in the respiratory tract, thereby increasing the chance for SARS-CoV2 to enter cells [93][94][95]. The clinical evidence to prove this theory requires further investigation.
Nonetheless, it cannot be denied that metformin does have protective effects due to its multiple molecular mechanisms in the vascular. Metformin could halt the activation of platelet and the release of mitochondrial DNA and suppress interaction between leukocytes and endothelium, thus reducing endothelial inflammation. These mechanisms prevent vein and artery thrombosis, conferring vascular protection [96]. With its broad protective mechanism, metformin is considered one of the primary choices of infusion medication with micro needles in the newest technology of diabetes treatment [97].

Conclusion
Diabetes mellitus and COVID-19 are the challenging diseases faced in recent days. New-onset diabetes is one problem induced by SARS-CoV-2 infection in nondiabetic patients. ACE2 is the critical key factor that possibly plays a vital role in new-onset diabetes, as DM and COVID-19 are interrelated with ACE2 in molecular pathogenesis. Treatments in acute hyperglycemic conditions are still controversial since there are still discrepancies regarding the results in the field. A careful decision based on the patient's current clinical condition and comorbidities is needed to make rational choices of treatment, which could then provide a precise and good outcome.