The evaluation of pancreas β-cell autoantibodies in non-diabetic COVID-19 patients

ABSTRACT Objective: This study aims to evaluate potential pancreas endocrine damage due to SARS-CoV-2 by measuring β-cell autoantibodies in COVID-19 patients. Subjects and methods: Between June and July 2020, 95 inpatients with a positive COVID-19 test result after polymerase-chain-reaction (PCR) and who met the inclusion criteria were enrolled in our study. Laboratory parameters that belong to glucose metabolism and β-cell autoantibodies, including anti-islet, anti-glutamic acid decarboxylase, and anti-insulin autoantibodies, were measured. β-cell autoantibodies levels of the patients were measured during COVID-19 diagnosis. Positive results were reevaluated in the 3rd month of control. Results: In the initial evaluation, 4 (4.2%) patients were positive for anti-islet autoantibody. Only one (1.1%) patient was positive for anti-glutamic acid decarboxylase autoantibody. No patient had positive results for anti-insulin autoantibody. FPG, HbA1c, and C-peptide levels were similar in patients who were split into groups regarding the initial positive or negative status of anti-islet and anti-GAD autoantibodies (p>0.05). In the 3rd month after the initial measurements, anti-islet autoantibody positivity of 2 (50%) of 4 patients and anti-glutamic acid decarboxylase positivity of 1 (100%) patient were persistent. Finally, 3 (3.1%) patients in the whole group were positive for anti-islet autoantibody in the 3rd month of control. No difference was determined between the initial and the 3rd month of parameters of glucose metabolism. Conclusion: Following an ongoing autoantibody positivity in the present study brings the mind that SARS-CoV-2 may be responsible for the diabetogenic effect. Clinicians should be aware of autoantibody-positive DM as a potential autoimmune complication in patients with SARS-CoV-2.


INTRODUCTION
S evere acute respiratory syndrome coronavirus-2 (SARS-CoV-2) pandemic that was called Coronavirus Disease 2019 (COVID-19) by the World Health Organization (WHO) first emerged in December 2019 in Wuhan, China (1,2).The pancreatic islet is a target tissue of the SARS-CoV-2 driven by ACE2 expression (3).The entry and propagation of this virus depend on the binding of its spike glycoprotein to the ACE2 receptor present in several host organs (4).
Several studies in the field suggest that diabetes mellitus (DM) is a significant risk factor for COVID-19 and has an association with poor prognosis; the influence of COVID-19 on glycemic parameters remains unclear (5)(6)(7).Previous case reports have been presented to the literature showing the possible impact of the direct cytotoxic effect of SARS-CoV-2 on pancreatic β-cells and leading to the development of autoantibody negative insulin-dependent DM.However, the importance of ACE2 on the intra-islet COVID-19 and pancreas autoantibodies Arch Endocrinol Metab.2022;66/4 paracrine mechanism is debatable, and there is no clear data available showing the association between SARS-CoV-2 infection and insulin-dependent DM (8).Additionally, the increased incidence of diabetic ketoacidosis in children and adolescents reveals a possible association between COVID-19 and newlyonset type 1 DM (9).SARS-CoV infection damages pancreatic islets and also causes subsequent acute DM (3).Moreover, it also leads to significant changes in the whole metabolism, including glucose, fat, and protein metabolism (10).Chronic inflammation affects systemic glucose homeostasis and contributes to hyperglycemia (11).
As is known, various viruses such as enteroviruses, coxsackie B virus, retroviruses, rubella, mumps, cytomegalovirus, Epstein-Barr, and varicella-zoster virus have played a role in the development of DM (12).Viral infections cause type 1 DM by triggering the production of cross-reactive antibodies due to molecular mimicry or by activating cross-reactive T cells (13).This situation has not been defined for coronavirus yet.According to a study conducted in a tertiary hospital in the United States, a significantly increased incidence of new-onset type 1 DM is observed during the COVID-19 pandemic compared to the previous years (14).Likewise, an increased type 1 DM incidence was established in a German populationbased study during the COVID-19 pandemic, which may be explained by β-cell autoimmunity due to the COVID-19 virus (15).The aim of this study is to draw attention to potential pancreas endocrine injury due to SARS-CoV-2 by measuring β-cell autoantibodies levels in COVID-19 patients.

Study design
This study was designed as a prospective observational study.The Ethics Committee of our institute approved this study regarding the principles of the Declaration of Helsinki (Date: 08.06.2020/Number: 89/02).Written informed consent of all patients was obtained before inclusion.

Patient selection
The patients diagnosed with COVID-19 between June and July 2020 were evaluated in terms of eligibility for inclusion in the study.A confirmed case of COVID-19 was defined by a positive result on a polymerase-chain-reaction (PCR) assay of a specimen collected on a nasopharyngeal swab.One hundred and five inpatients over 18 years of age who accepted the invitation were evaluated for suitability to be included in this study.The patients diagnosed with DM before and/or after hospitalization, who were started steroid or tocilizumab/anakinra treatment (drugs which can stimulate immunologic response), and pregnants were excluded.Finally, 95 patients were included in the study.

The evaluation of demographics and laboratory parameters of the patients
Demographic data, additional diseases, and the severity of the clinical condition of the patients with a positive test result for COVID-PCR were recorded.Hemogram, kidney and liver enzymes, fibrinogen, C-reactive protein, ferritin, and D-dimer levels were evaluated.The clinical condition was categorized as asymptomatic, mild (no pneumonia or mild pneumonia), moderate/ severe (dyspnea, respiratory frequency ≥ 30/min, blood oxygen saturation ≤ 93%, partial pressure of arterial oxygen to fraction of inspired oxygen ratio < 300, and/or lung infiltrates > 50% within 24/48 h), and critical (respiratory failure, septic shock, and/or multiple organ dysfunction or failure) (16).
In order to evaluate the glucose metabolism of the patients, fasting plasma glucose (FPG), HbA1c, C-peptide and anti-islet, anti-GAD, and anti-insulin autoantibodies were measured.Peripheral blood samples were collected between 8:00 and 10:00 am after at least 8-hour of overnight fasting in the first 3 days of hospitalization.

COVID-19 and pancreas autoantibodies
Arch Endocrinol Metab.2022;66/4 positive; <1 U/mL: negative; 1-2 U/mL: borderline for anti-GAD autoantibody; <8.2%: negative; >8.2%: positive for anti-insulin autoantibody.Borderline antibody results were analyzed for the second time, and the confirmed antibody statutes of the patients were presented.Antibody levels of those patients with positive or borderline antibody results and parameters related to glucose metabolism were repeated in the 3rd month after the first measurement.

Statistical analysis
Statistical analyses were performed using SPSS software version 21 (Chicago, IL).The variables were assessed through visual (histograms, probability plots) and analytic methods (Kolmogorov-Smirnov/Shapiro-Wilk's test) to determine whether they were normally distributed or not.While the Student's t-test was performed to compare normally distributed parameters, the Mann-Whitney U test was used to compare nonnormally distributed ones.Descriptive analyses were demonstrated using means and standard deviation for normally distributed variables, whereas medians and interquartile ranges (IQR) 25 and 75 percentiles were used for non-normally distributed variables.The Chisquare test or Fisher's exact test, where appropriate, was used to compare proportional data.The Kruskal-Wallis test was performed to compare non-normally distributed variables regarding the presence of autoantibodies.One-way ANOVA was used to compare normally distributed variables among the autoantibody statuses.P-values, as well as p-trend, were calculated when one-way ANOVA was used.Paired Student's t-test was used to compare the measurements at twotime points (baseline and 3 months).A p-value less than 0.05 was considered to show a statistically significant difference.When the overall significance was observed, pairwise posthoc tests were performed.

RESULTS
Out of 105 patients meeting inclusion criteria, 6 patients previously diagnosed with type 2 DM, 2 patients using chronic steroids, and 2 patients diagnosed with DM during hospitalization were excluded.Finally, 95 patients were enrolled in the study.
While 8 (8.4%) patients had a family history of type 2 DM, no patients had relatives with type 1 DM.Demographics, comorbidities, and clinical features of the participants are presented in Table 1.
The median FPG level of the patients was 92 (IQR 25-75; 87-99) mg/dL, and the mean HbA1c level was 5.4 ± 0.45% (Table 1).The median C-peptide level of the patients was 2.29 (IQR 25-75; 1.75-3) ng/mL (Table 1).Other laboratory test results, including hemogram parameters, kidney and liver functions, CK, D-dimer, ferritin, and fibrinogen, are demonstrated in Table 1.In the initial evaluation of the patients, 4 (4.2%) and 12 (12.6%)patients were positive and borderline, respectively, for anti-islet autoantibody.Only one (1.1%)patient was determined to be positive for anti-GAD autoantibody, whereas 4 (4.2%)patients had borderline results for anti-GAD autoantibody.No patient had positive or borderline results for anti-insulin autoantibody.The initial β-cell autoantibody levels of the patients are presented in Table 1.
FPG and HbA1c levels were similar in the patients grouped according to the initial positive or negative status of anti-islet and anti-GAD autoantibodies (p>0.05 for each) (Table 2).Despite being within the normal limits, the initial C-peptide level of the patients with anti-islet or anti-GAD positivity was relatively lower than the negative ones.However, the difference was not statistically significant (p>0.05)(Table 2).There was no linear trend in FPG, HbA1c, and C peptide levels among subgroups of anti-islet and anti-GAD autoantibodies (p>0.05 for each).The clinical status of the patients was similar among negative and positive autoantibody groups (p>0.05 for each).No difference was determined between the initial and the 3rd month of FPG, HbA1c, C-peptide, and anti-islet autoantibody levels in 3 patients who had ongoing antiislet autoantibody positivity.One patient with anti-GAD positivity was also positive for anti-islet autoantibody.
Meanwhile, a significant difference was observed in terms of gender among the groups considering antiislet autoantibody status (p<0.001)(Table 2).Posthoc analysis showed that the number of the males was significantly higher in the patients with borderline antiislet autoantibody levels compared to the subjects with negative results (p<0.001) (Table 2).
In the 3rd month after the initial measurements, anti-islet autoantibody positivity of 2 (50%) of 4 patients and anti-GAD positivity of 1 (100%) patient were persistent.The other two patients with positive initial anti-islet autoantibody levels were observed in borderline.Among 12 patients having borderline results for anti-islet autoantibody at the initiation, 1 (8.3%), 7 (58.4%),and 4 (33.3%)patients were positive, negative, and borderline, respectively, in the 3rd month of control.Final 3 (3.1%)patients in the whole group were positive for anti-islet autoantibody in the 3rd month of control.While 2 (50%) of 4 patients with initial borderline results for anti-GAD remained the same, 2 (50%) of them were observed as negative.Baseline and 3rd-month autoantibody levels of the groups are demonstrated in Figure 1.Demographics and laboratory test results of the patients who had β-cell autoantibody positivity in the 3rd month are presented in Table 3.

DISCUSSION
The present study revealed that anti-islet and anti-GAD autoantibody positivity were 4.2% and 1.1%, respectively, in COVID-19 patients in the initial evaluation.Anti-islet autoantibody positivity decreased by 3.1%, although anti-GAD autoantibody positivity remained the same in the 3rd month.Categorical data were demonstrated with numbers.Normally distributed variables were presented asmeans (standard deviations).Non-normally distributed variables were presented as medians (interquartile ranges .*According to post-hoc analysis, male patients in anti-islet borderline group make the difference.GAD: glutamic acid decarboxylase; FPG: fasting plasma glucose; HbA1c: glycated hemoglobin.A limited number of recent studies have shown the increased incidence of type 1 DM during the pandemic (14,15).The possible reason for this increase is due to pandemic-related restriction precautions rather than the direct effect of SARS-CoV-2 on the pancreas (15).Additionally, immune-mediated and inflammatory responses, stress-related and steroid-induced hyperglycemia were responsible for the increase in newly diagnosed DM (17).
Viruses, one of the environmental factors, are responsible for type-1 DM and cause β-cell damage either by stimulating autoimmune attack or directly with cytotoxicity (18).Serological evidence of infection and isolation of viruses from the pancreas have been reported in a few cases recently diagnosed with DM (12,19).This situation has not been defined for coronavirus yet.However, over a decade ago, it was indicated that SARS-CoV (a cousin of SARS-CoV-2) might cause insulin-dependent DM with acute onset (3).We studied β-cell autoantibody levels in COVID-19 patients based on the idea that SARS-CoV-2 may be one of the potential environmental factors for the development of DM, although coronavirus is not on the list of viruses in type-1 DM etiopathogenesis or latent autoimmune diabetes in adults.That autoantibody positivity has remained the same (3.1%) in some patients caused worry about whether COVID-19 led to a permanent damage in pancreas islet cells when the patients detected a positive autoantibody in the inactive infection period were evaluated three months later.
SARS-CoV-2 receptor, ACE2, which plays a crucial role in the relationship between COVID-19 and hyperglycemia, is expressed in both exocrine glands and pancreas islets; therefore, pancreatic endocrine damage is expected (20).SARS-CoV-2 is likely to cause endocrine damage in the pancreas through immunemediated mechanisms (immune-mediated cellular responses, indirect systemic inflammatory or direct cytopathic effects) due to high ACE2 concentration in the pancreas islets and also to cause insulin-dependent DM with acute onset (20,21).Apart from direct β-cell damage, alterations in self-antigens and subsequent immune-mediated destruction of β-cells could be implicated.

Figure 1 .
Figure 1.The changes in autoantibody levels at the baseline and in the 3rd-month.

Table 2 .
Demographics and glucose metabolism parameters according to initial anti-islet and anti-GAD autoantibodies status

Table 3 .
Demographics and laboratory test results of the patients with positive β-cell autoantibodies in the 3rd month Abbreviation: RR; reference range Abbreviation: RR; reference range.