Study 1. Autoimmunity to Post Translationally Modified Beta Cell Antigens Abstract Background Type 1 Diabetes (T1D) is a heterogenous autoimmune disorder, whereby the immune system specifically targets and destroys the pancreatic beta cells, leading to little to no insulin production and secretion. T1D is characterized by hyperglycemia, juvenile age of onset, and the presence of one or multiple autoantibodies (AABs) directed towards insulin (INS), glutamate decarboxylase 65 (GAD65), zinc transporter 8 (ZnT8), and islet antigen 2 (IA-2). The gold-standard for diagnosing T1D is via a radio-binding assay detecting the key AABs, with the detection of multiple autoantibodies in genetic susceptible individuals conferring a higher risk of developing T1D. Environmental triggers, such as viral infections, dietary changes, and pollutants, have been implicated in inducing the onset of T1D. However, the current AABs differ from various populations and have a tendency to disappear after initial recognition. Hence there is a need to identify, isolate, and develop more specific biomarkers that can more accurately predict, diagnose and monitor the development of T1D. Although the key instigator of autoimmunity has not been fully elucidated, numerous studies have implicated the role of oxidative stress and reactive oxygen species in initiating an autoimmune response. It has been suggested that increased levels of oxidative stress can induce oxidative post translational modifications (oxPTMs) to native beta cell antigens, forming neo-epitopes. These neo-epitopes appear foreign to the immune system, and thus elicit an immune response. Additionally, the sequence similarities between the neo-epitope and the native epitope induces an immune response towards the native epitope through a mechanism known as epitope spreading. Studies have reported the presence of autoantibodies to oxPTM-INS and hydroxyl modified GAD65 in T1D. Hypothesis The main hypothesis of the study is that oxPTMs of native pancreatic β-Cells antigens may play a role in the development of autoimmunity associated with T1D. Aim 1. The first aim of the study is to deduce the level of reactivity to oxPTM β-cell antigens, focusing on GAD65 and ZnT8, in patients with new-onset T1D. Aim 2. The second aim is to compare the reactivity of AABs to oxPTM antigens to AABs to native antigens in patients with T1D and healthy controls. As well as, to deduce the efficiency of oxPTM β-cell antigens in being potential biomarkers for T1D compared to native antigens. Aim 3. The final aim of the study is to determine whether there are any correlations with AAB development to oxPTM or native antigens to patient demographic and clinical information. This includes genetic information pertaining to HLA-typing, markers of oxidative stress, HbA1c levels, and C-peptide levels. Materials and Methods A total of N= 69 patients with T1D, N= 32 patients with type 2 diabetes (T2D), and N= 39 healthy controls (HC) were recruited from the outpatient clinic of the Endocrinology and Diabetes department of the Policlinico Campus Biomedico Università di Roma. Patient characteristics and demographic data were collected, along with informed consent. In vitro chemical modifications of INS, GAD65, and ZnT8 were performed to generate oxPTMs modification via glycation (GLY-) with 2M D-ribose, the hydroxyl radical (•OH-) via the Fenton reaction using 100 mM hydrogen peroxide and 100 mM copper chloride, and modification with a 1:100 dilution of 1M hypochlorous acid (HOCl-). The induced modifications were monitored and assessed through SDS-PAGE analysis and Western Blotting, with the subsequent gels imaged with the Chemi-Doc imaging system and quantified with ImageJ. A “homemade” ELISA was devised to measure and detect the levels of antibody reactivity to each native (NT-) and oxPTM-beta cell antigen. This was done by coating a 96 well plate with either NT- or modified antigens (10 μg/mL or 2 μg/mL) in 0.05M bicarbonate-carbonate buffer (pH 6), and incubating the plates overnight at 4⁰C. The plates were blocked with either 5% BSA in 1x PBS with 0.1% Tween (PBST) or 5% skimmed milk in PBST. A 1:200 dilution of T1D, T2D or HC serum samples were then prepared in blocking buffer and loaded into the plates. This was followed by the addition of a secondary antibody (rabbit anti-human IgG) also diluted in blocking buffer to the plates. The plates were washed 3 times with PBST and then 3 times with PBS between each step. TMB was then added to the plates to yield a colorimetric reaction, the reaction was stopped with sulphuric acid, and the plates were read on the Tecan M200 plate reader at O.D 450 nm. Carboxymethyl-Lysine was measured with an ELISA assay. Results When comparing demographic patient characteristics such as age, gender distribution, BMI, and HbA1c, the T1D were typically younger compared to T2D (median age in years [IQR],13 [9-26] vs 62.50 [52- 68.25], p-value <0.0001, respectively), BMI was lower in T1D than in T2D (median kg/m2 [IQR], 18 .60 [16.20-20.95] vs 30.40 [24.35-33.05], p-value < 0.0001, respectively). SDS-PAGE analysis showed that when GAD65 is modified in vitro with 2 M D-ribose, there appeared to be an upshift in the molecular weight of GAD65 from bands at 116.8, 54.9 and 22.7 KDa to 117.5, 59.5 and 23.3 KDa in the GLY-GAD65 lane. Similarly, when assessing the band intensity, there was a decrease in band intensity between NT-GAD65 and GLY-GAD65. With •OH-GAD65, there appeared to be 2 faint bands at 60.1 KDa and 53.3 KDa. In comparison to the NT-GAD65, the •OH-GAD65 had a decreased band intensity. Finally, the HOCl-GAD65 showed bands at 117.5, 60.1 and 23.5 KDa. Upon glycation of NT-ZnT8 there was an apparent up-shift in GLY-ZnT8 with distinct bands at 93.1 and 43.2 KDa compared to the NT-ZnT8, which showed clear bands at 83.4 and 36.6 KDa. In comparison, modifying NT-ZnT8 with the hydroxyl radical to yield •OH-ZnT8 showed a similar increase in molecular weight to 88.9, 72.9, and 38.9 KDa compared to the NT-ZnT8 bands at 83.4, 36.6 and 31.0 KDa. HOCl-ZnT8 appeared to have bands at a molecular weight of 77.2 and 31.4 KDa compared to 60.2 and 23.5 KDa seen in the NT-ZnT8 lane. A cut-off for antibody binding positivity was calculated from the mean IgG binding of the healthy controls towards native or oxPTM beta cell antigens (mean O.D+ 2*standard deviations), resulting in an O.D of 0.5 as the cut-off for the homemade ELISAs developed towards NT- and oxPTM- GAD65 and ZnT8. Using this cut-off, the prevalence of antibody binding towards modified or native beta cell antigens was assessed. This study found that 53.2% (36/69) of patients with new-onset T1D were positive for IgG antibodies to NT-GAD65, 43.49% (30/69) were positive for GLY-GAD65, 17.4% (12/69) were positive for •OH-GAD65 and 39.1% (27/69) were positive to HOCl-GAD65. When determining the binding of oxPTM-GAD65-Ab in patients with new-onset T1D, there was a significant difference in the antibody binding towards •OH-GAD65 compared to NT-GAD65 patients with T1D (mean O.D ± S.D, 1.01 ± 0.40 vs 0.80 ± 0.15, p-value = 0.065, respectively). When comparing the changes in IgG binding towards native or oxPTM-ZnT8 in the 45.3% (24/53) of patients with new-onset patients with T1D positive for ZnT8-Ab, there appeared to be no significant difference between antibody binding to NT- and GLY-ZnT8 (mean O.D ± S.D, 0.7292 ± 0.1873 vs 0.6118 ± 0.2464, p-value= 0.064, respectively) or NT- and HOCl-ZnT8 (mean O.D ± S.D, 0.7292 ± 0.1873 vs 0.6285 ± 0.2505, p-value= 0.1140, respectively). However, there was an apparent significant decrease in IgG binding between NT- and •OH-ZnT8 (mean O.D ± S.D, 0.73 ± 0.19 vs 0.27± 0.14, p <0.0001, respectively). When assessing the antibody response towards NT- or oxPTM-INS, there was a significant decrease with antibody binding towards NT-INS and GLY-INS (mean O.D ± S.D, 0.221 ± 0.108 vs 0.170 ± 0.081, p-value <0.05, respectively). However there was a significant increase in antibody binding seen between NT-INS and HOCl-INS (mean O.D ± S.D, 0.221 ± 0.108 vs 0.363 ± 0.134, p-value <0.0001, respectively), as well as a significant increase between antibody binding towards NT-INS and •OH-INS (mean O.D ± S.D, 0.221 ± 0.108 vs 0.400 ± 0.129, p-value <0.0001, respectively). The correlation of the induced IgG response towards NT- or oxPTM-GAD65 was assessed by determining the degree of overlap in the patients towards each antigen. Within the total cohort of N=69 patients with new-onset T1D, 17% of patients were positive for an IgG antibody response towards NT-, Gly-. •OH- and HOCl-GAD65. With a total of 41% (29/69) appearing positive for GLY-GAD65 and 39% (27/69) towards •OH-GAD65. However, 45% of T1D were negative for antibodies directed towards all antigens (NT- and oxPTM-GAD65). When assessing antibody responses to NT- and oxPTM-ZnT8, 59% (41/69) of patients were negative for any antibody response, whereas 33% (23/69) were positive for antibodies against NT-GAD65. There were only 4.3% (3/69) patients with T1D who were positive to both NT- and all oxPTM-ZnT8, 25% (17/69) were positive for IgG antibodies to GLY-ZnT8 and HOCl-ZnT8, and only 5.8% (4/69) were positive to •OH-ZnT8. Discussion Research has indicated that PTMs may play a role in the progression of autoimmune diseases. PTMs are able to change the secondary and tertiary structure of proteins, impacting self-tolerance and immunogenic mechanisms. Although the majority of PTMs work to better the function of proteins, those that arise spontaneously, as a result of a breakdown in homeostasis, can be attributed to the development of autoimmune disorders. Multiple sclerosis, rheumatoid arthritis, and juvenile idiopathic arthritis are just a few of the autoimmune disorders with post-translationally modified autoantigens being named as contributors to the progression of these diseases. This study suggests that oxPTMs to native β-cell antigens alters the subsequent antigenicity of the targeted proteins and contributes to the progression of clinical T1D in people at-risk of developing T1D. When comparing the reactivity between NT-GAD65 and oxPTM-GAD65, this study showed significantly higher antibody reactivity towards GLY-GAD65, HOCl-GAD65 and NT-GAD65 in patients with new- onset T1D compared to healthy controls and people with T2D. However, there was no significant difference between the IgG antibody reaction towards GLY-GAD65, HOCl-GAD65 and NT-GAD65 in patients with T1D, yet certain patients showed an increased reactivity to GLY-GAD65 compared to NT-GAD65. This suggests the possible role of hyperglycemia or dysregulated glucose in modifying GAD65 in the initial stages before the onset of clinical T1D. In comparison, a significant decrease was seen in the antibody response towards oxPTM-ZnT8 compared to NT-ZnT8, indicating that oxidative modifications to ZnT8 may not impact the immunogenicity of the antigen. Controversially, oxPTMs appeared to increase the immunogenicity of insulin. In fact, patients with T1D developed significantly higher antibody responses towards •OH-INS compared to NT-INS. There was also a significant difference in antibody reactivity to •OH-INS in patients with T1D compared to patients with T2D and HC. These findings may suggest that the oxidative modification of insulin, specifically through the hydroxyl radical, may be an initial target of immunity in patients at-risk of developing T1D. A phenomena known as molecular mimicry has been suggested to play a role in the spread of autoantigenicity. In molecular mimicry, the sequence similarities between the autoantigen instigating the immune response and the native antigen lead to the immune system targeting the native antigen as well. Thus, ensuing epitope spreading and further stimulating the immune system. Upon the initial seroconversion, the immune system may begin to target the remaining beta cell antigens, such as GAD65 and ZnT8, as well as NT-INS due to the sequence similarity between both NT-INS and •OH-INS. However, further studies are required to confirm the findings, with the recruitment of more patients in the initial phases of T1D. Moreover, it is necessary to identify the exact peptides being targeted by the immune system, and those being modified with oxidative radicals in each beta cell antigen. In conclusion, oxPTMs to the beta cell antigens, insulin, GAD65 and ZnT8, alter the immune response in patients with new onset T1D, however further studies are needed in order to determine the role they play in the development of autoimmunity before the onset of clinical T1D. Study 2: DiabeSARS; A Tale of Two Pandemics Abstract Background The COVID-19 pandemic was the major health concern of the last 3 years. The sudden onset, and widespread of the SARS-CoV-2 virus, necessitated the expedition of research to fully understand the pathogenesis of the virus and to develop effective treatments. Diabetic patients have consistently been associated with severe and potentially fatal disease, more so than their non-diabetic counterparts. Hence, there is a clear need to fully elucidate the underlying pathology that leads to more severe disease in diabetic patients. Studies have previously assessed the role of underlying inflammation and dysregulated glucose in both T1D and T2D in previous viral infections. However, within the COVID-19 pandemic, although diabetic patients only make up a small percentage of the general population, they seem to comprise a similar percentage of critically ill patients, requiring admission to the intensive care unit as their non-diabetic counterparts. Studies have found the dysregulated glucose, regardless of diabetic state, promotes detrimental viral infections. Yet, the role of the endocrine system on immunometabolic outcomes has not been fully elucidated in terms of COVID-19. Increased hyperglycemia in diabetic patients has been associated with increased potential of glycation to native proteins, in fact, glycated hemoglobin (HbA1c) is consistently measured to monitor the development and progress of diabetes. Studies have demonstrated the glycosylation of both the SARS-CoV-2 spike (S) glycoprotein and the angiotensin converting enzyme 2 (ACE2), the main SARS-CoV-2 receptor. The induced glycosylation has been linked to increased virulence of the SARS-CoV-2 virus, as well as increased host-cell infiltration potential. However, the impact of non-enzymatic glycation, a potential outcome due to the elevated levels of glucose in an overly distressed system, has not been fully elucidated. Additionally, the efficacy of the developed SARS-CoV-2 mRNA vaccine (Pfizer, BioNtech) has not been assessed in terms of glucose control in diabetic patients. Hypothesis The main hypothesis is that dysregulated glucose levels may induce glycation of the SARS-CoV-2 S protein, altering the virulence of the disease. With one of the main focuses of this study to determine whether glucose regulation plays a role in vaccine efficacy and, in turn, immune protection against COVID-19. Aim 1. To induce glycation of the S protein and asses ACE2 binding to both native (NT-) and glycated (GLY-) S protein. Aim 2. To determine the impact of antibody binding to NT- and GLY-S protein in previous COVID-19 patients. Aim 3. To evaluate the role of glucose levels and monitoring on the protective immune response in patients following the administration of the SARS-CoV-2 mRNA vaccine (Pfizer-BioNTech, BNT162b2) Aim 4. In order to better understand the role of hyperglycemia and diabetes in the adverse outcomes of COVID-19, the fourth aim of this study is to evaluate the clinical risk of diabetes and glucose levels on mortality in patients with COVID-19. Methods Antibody response to NT- or GLY-S protein in patients with COVID-19 and vaccinated patients COVID-19 patients with and without diabetes were recruited for this study from specifically designated COVID-19 wards. For the vaccine study, patients were screened and recruited from the Endocrinology and Diabetology unit at Policlinico Campus Biomedico di Roma. With n= 26 patients being T1D and n=32 being diagnosed with T2D. The inclusion criteria for the study were patients >18 years old, scheduled to receive the SARs-CoV-2 mRNA vaccine (Pfizer-BioNTech), signing informed consent, having a diagnosis of T1D or T2D for more than 3 months, and using at least two anti-diabetic drugs in the case of T2D according to the vaccine priority criteria. Demographic and clinical data was collected at each time point involved in the study for all the patients recruited. The study covered a total of 6 months with the following timepoints: T0 (baseline): before the administration of SARs-CoV-2 mRNA vaccine (within 3 days of the first dose) T1: 21 days after the first dose (day of the second dose) T2: 35 days from baseline (T0) T3: 90 days from baseline (T0) T4: 180 days from baseline (T0). The SARS-CoV-2 S protein (0.450 mg/mL) was modified via glycation with equal volumes of 0.5 M D-ribose (sigma), this was incubated overnight at 37⁰C and stored short term at 4⁰C. The induced modifications were monitored by SDS-PAGE analysis, whereby 5 μg of NT- or GLY-S protein were loaded into the gel wells with equal volumes of Laemmli loading buffer with or without β-mercaptoethanol to promote reducing conditions. The gels were imaged with the Chemi-Doc imaging system from Bio-Rad and quantified with Imagej software. Changes in ACE2 binding towards NT- or GLY-S protein was determined via a homemade ELISA, whereby increasing dilutions of ACE2 (2 μg/mL-0.25 μg/mL) were measured against the same concentration of NT- and GLY-S protein (1 μg/mL). Moreover, COVID-19 patients with and without diabetes were assessed via a homemade ELISA for the antibody response towards NT- and GLY-S protein. The antibody responses were then analyzed depending on the fructosamine levels, measured via a kit (ab228558, abcam). The antibody response induced after immunization with the SARS-CoV-2 mRNA vaccine was measured via a homemade ELISA devised to analyze the IgG response towards NT- or GLY-S protein. Serum samples were collected and assessed at each timepoint and compared to the results of controls without diabetes. The level of neutralization antibodies was measured via a devised neutralization assay using live SARS-CoV-2 (Vero E6 cells). Devising a Clinical Risk Score to Assess in-hospital Death from COVID-19 Data from patients for devising the clinical risk score was collected retrospectively from 417 COVID-19 patients admitted to Jaber Al-Ahmed Hospital in Kuwait between February 24th and May 3rd, 2020. Due to the emergency state of the COVID-19 pandemic, the need for signed consent was waived by the ethical committee from the Ministry of Health in Kuwait. Setting the primary outcome as in-hospital death, a series of multivariant logistical regression models were performed to identify independent factors that may be prognostic for the primary outcome. The models were created by adding or removing variables individually depending on the results of the previous logistic regressions, with variables showing a p-value<0.1, being retained in the score. The independent predictive variables included in the final model were gender, asthma, glucose categories, and non-Kuwaiti national. Weighted points were assigned to significant risk factors proportional to their beta regression coefficient values. The effectiveness of the risk score to predict mortality in patients with COVID-19 was analyzed via receiver operating characteristic (ROC) curves, with an AUC of 0.5 or less was taken as insignificant Youden’s index was applied to set a cut-off for mortality prediction. Significance was set as a two-tailed p-value <0.05. The score was built using SPSS (IBM Corp. IBM SPSS Statistics for Windows, Version 21.0. Released 2012. Armonk, NY: IBM Corp.). The score was internally validated by a Kuwaiti COVID-19 cohort of N=923 patients, and externally validated using the CoViDiab cohort from Italy (N= 178). Results Antibody response to NT- or GLY-S protein in COVID-19 patients and vaccinated patients Upon assessing the binding of NT- and GLY-S protein (1μg/mL) to serial concentrations of the SARS-CoV-2 receptor ACE2 (2.0 μg/mL to 0.25 μg/mL) there was diminished binding of ACE2 to NT-S vs GLY-S, however this was not significant (mean ± S.D, 0.1393 ± 0.1732 vs 0.1943 ± 0.2355, p-value= 0.6026, respective). Fructosamine levels, which are a short-term assessment of glucose control, were measured in the total COVID-19 patient cohort (N=46). Upon stratification of fructosamine levels, patients with <563 μmol/L had no significant difference between IgG levels to NT- or GLY-S, yet patients with fructosamine levels >563 μmol/L had a significant difference between NT- and GLY-S (mean O.D ± S.D= 0.7945 ± 0.2564 vs 0.6155 ± 0.2369, p-value= 0.0078, respectively). When correlating fructosamine levels with the total cohort (N=46), there appeared to be an overall negative correlation to both NT- and GLY-S, yet this was not significant (r= -0.1848, p-value= 0.2188 vs r= -0.2203, p-value= 0.1413, respectively). In diabetic patients (n=22) there appeared to be no correlation with fructosamine levels to the IgG response to either NT- or GLY-S protein (r= 0.04029, p-value= 0.8587 vs r= 0.010201, p-value= 0.9640). However, in the case of non-diabetic COVID-19 patients, there was a strong negative correlation between fructosamine levels and the immune response to NT- and GLY-S (r= -0.3824, p= 0.1462 vs r=-0.4042, p= 0.0501, respectively). This study attempted to assess the IgG response towards the COVID-19 Pfizer BioNTech mRNA vaccine in terms of continuous glucose monitoring (CGM) data from N=10 patients with T1D with available CGM profiles. The key measurements assessed were Time-in-range (TIR), which is the percentage of time within a 24 hour period that a patient is within the ideal glucose range, Time-above-range (TAR), which is the percentage of time a patient is above the ideal glucose range in a 24 hour period, and Time-below-range (TBR), which is the percentage of time a patient is below their ideal glucose range. When determining the overall area under the curve (AUC) of the IgG response over all study timepoints (T0-T4) and correlating it with the average TIR, there appeared to be a strong correlation with AUC IgG response and TIR to both NT- and GLY-S protein (r= 0.8082, p-value= 0.0084 vs r= 0.7996, p-value= 0.0097, respectively), with a significant difference between the AUC IgG to NT- and GLY-S protein (p-value= 0.0034). When correlating the AUC IgG of both NT- and GLY-S protein to the average TAR over all study timepoints (T0-T4), there was a strong negative correlation (r= -0.7926, p-value= 0.0108 vs r= -0.7430, p-value= 0.0218, respectively) There was no correlation with HbA1c at T0 or average TBR with AUC IgG response towards native or GLY-S protein. When dividing patients with T1D with CGM data (N=13) based on their recommended glucose targets, (TIR > 70% and TBR <25%), there appeared to be a stronger neutralizing antibody response to the native SARs-CoV-2 spike protein who in patients with T1D had a TIR>70% than those who did not (p<0.0001). Furthermore, when assessing the neutralizing antibody response against TBR measurements, patients who had a TBR<25% were more likely to have a stronger neutralizing antibody response (p=0.008), this was seen regardless of HbA1c levels. Devising a Clinical Risk Score to Assess in-hospital Death from COVID-19 The score was built by assessing the significance of several predictive variables against the primary outcome (in-hospital mortality). The final score included asthma, gender (male), nationality (non-Kuwaiti national), and blood glucose levels (either between 7.0-11.1 mmol/L or >11.1 mmol/L) as independent predictors of mortality in COVID-19. A point system was given to each predictive variable based on the beta coefficients allocated to each variable. The cut-off of the score to predict death was 5.5, showing a specificity of 86.3% and sensitivity of 75% (AUC= 0.901). The clinical risk score requires internal and external validation to assess the potential to predict the primary outcome. Two cohorts were used for internal validation, the initial N=417 Kuwaiti COVID-19 group used to build the score and a separate cohort of N=923 Kuwaiti COVID-19 patients admitted from May 4th to August 26th, 2020, both admitted within one COVID-19 center within Kuwait. External validation was performed using an N=178 CoViDiab Italian cohort. The score was calculated for each patient and then tested against the primary outcome (in-hospital mortality from COVID-19); the score was then plotted as a ROC curve with the AUC calculated. The AUC showed 0.901 ± 0.20 fit for the score for the 417 Kuwaiti cohort, 0.826 ± 0.91 fit for the score for the 923 Kuwaiti cohort, and a 0.687 ± 0.06 fit for the score for the CoViDiab cohort, with respective negative predictive values of 95.4%, 93.9%, and 94.1%. Conclusions Diabetic patients are characterized by hyperglycemia and chronic low-grade inflammation. Upon viral infection, these patients present with more exacerbated immune responses, that may lead to severe disease, ICU admission, and potentially death. The findings in this study suggest that dysregulated fructosamine levels are more strongly correlated with a decreased antibody response towards NT- and GLY-S protein of SARS-CoV-2. Additionally, when assessing the IgG and neutralization antibody response in patients with T1D in association with CGM data, there appeared to be a stronger association with more improved TIR and TAR glucose measurements and a stronger antibody response. This again suggests that better controlled glucose measurements aid in improving the protective immune response towards SARS-COV-2. Finally, the development of the clinical risk score demonstrated that elevated glucose was a stronger predictor of negative outcomes and mortality in COVID-19 related infections than diabetes. In fact, the addition of glucose measurements removed diabetic state as an independent predictor of in-hospital death. In conclusion, maintaining key glucose targets may aid in preventing detrimental outcomes towards not only COVID-19 and other viral infections.

Autoimmunity and Antibody Response in Diabetes: The Role of Hyperglycemia and Antigen Post Translational Modifications / Ghadeer Alhamar , 2023 Mar 22. 35. ciclo, Anno Accademico 2019/2020.

Autoimmunity and Antibody Response in Diabetes: The Role of Hyperglycemia and Antigen Post Translational Modifications

ALHAMAR, GHADEER
2023-03-22

Abstract

Study 1. Autoimmunity to Post Translationally Modified Beta Cell Antigens Abstract Background Type 1 Diabetes (T1D) is a heterogenous autoimmune disorder, whereby the immune system specifically targets and destroys the pancreatic beta cells, leading to little to no insulin production and secretion. T1D is characterized by hyperglycemia, juvenile age of onset, and the presence of one or multiple autoantibodies (AABs) directed towards insulin (INS), glutamate decarboxylase 65 (GAD65), zinc transporter 8 (ZnT8), and islet antigen 2 (IA-2). The gold-standard for diagnosing T1D is via a radio-binding assay detecting the key AABs, with the detection of multiple autoantibodies in genetic susceptible individuals conferring a higher risk of developing T1D. Environmental triggers, such as viral infections, dietary changes, and pollutants, have been implicated in inducing the onset of T1D. However, the current AABs differ from various populations and have a tendency to disappear after initial recognition. Hence there is a need to identify, isolate, and develop more specific biomarkers that can more accurately predict, diagnose and monitor the development of T1D. Although the key instigator of autoimmunity has not been fully elucidated, numerous studies have implicated the role of oxidative stress and reactive oxygen species in initiating an autoimmune response. It has been suggested that increased levels of oxidative stress can induce oxidative post translational modifications (oxPTMs) to native beta cell antigens, forming neo-epitopes. These neo-epitopes appear foreign to the immune system, and thus elicit an immune response. Additionally, the sequence similarities between the neo-epitope and the native epitope induces an immune response towards the native epitope through a mechanism known as epitope spreading. Studies have reported the presence of autoantibodies to oxPTM-INS and hydroxyl modified GAD65 in T1D. Hypothesis The main hypothesis of the study is that oxPTMs of native pancreatic β-Cells antigens may play a role in the development of autoimmunity associated with T1D. Aim 1. The first aim of the study is to deduce the level of reactivity to oxPTM β-cell antigens, focusing on GAD65 and ZnT8, in patients with new-onset T1D. Aim 2. The second aim is to compare the reactivity of AABs to oxPTM antigens to AABs to native antigens in patients with T1D and healthy controls. As well as, to deduce the efficiency of oxPTM β-cell antigens in being potential biomarkers for T1D compared to native antigens. Aim 3. The final aim of the study is to determine whether there are any correlations with AAB development to oxPTM or native antigens to patient demographic and clinical information. This includes genetic information pertaining to HLA-typing, markers of oxidative stress, HbA1c levels, and C-peptide levels. Materials and Methods A total of N= 69 patients with T1D, N= 32 patients with type 2 diabetes (T2D), and N= 39 healthy controls (HC) were recruited from the outpatient clinic of the Endocrinology and Diabetes department of the Policlinico Campus Biomedico Università di Roma. Patient characteristics and demographic data were collected, along with informed consent. In vitro chemical modifications of INS, GAD65, and ZnT8 were performed to generate oxPTMs modification via glycation (GLY-) with 2M D-ribose, the hydroxyl radical (•OH-) via the Fenton reaction using 100 mM hydrogen peroxide and 100 mM copper chloride, and modification with a 1:100 dilution of 1M hypochlorous acid (HOCl-). The induced modifications were monitored and assessed through SDS-PAGE analysis and Western Blotting, with the subsequent gels imaged with the Chemi-Doc imaging system and quantified with ImageJ. A “homemade” ELISA was devised to measure and detect the levels of antibody reactivity to each native (NT-) and oxPTM-beta cell antigen. This was done by coating a 96 well plate with either NT- or modified antigens (10 μg/mL or 2 μg/mL) in 0.05M bicarbonate-carbonate buffer (pH 6), and incubating the plates overnight at 4⁰C. The plates were blocked with either 5% BSA in 1x PBS with 0.1% Tween (PBST) or 5% skimmed milk in PBST. A 1:200 dilution of T1D, T2D or HC serum samples were then prepared in blocking buffer and loaded into the plates. This was followed by the addition of a secondary antibody (rabbit anti-human IgG) also diluted in blocking buffer to the plates. The plates were washed 3 times with PBST and then 3 times with PBS between each step. TMB was then added to the plates to yield a colorimetric reaction, the reaction was stopped with sulphuric acid, and the plates were read on the Tecan M200 plate reader at O.D 450 nm. Carboxymethyl-Lysine was measured with an ELISA assay. Results When comparing demographic patient characteristics such as age, gender distribution, BMI, and HbA1c, the T1D were typically younger compared to T2D (median age in years [IQR],13 [9-26] vs 62.50 [52- 68.25], p-value <0.0001, respectively), BMI was lower in T1D than in T2D (median kg/m2 [IQR], 18 .60 [16.20-20.95] vs 30.40 [24.35-33.05], p-value < 0.0001, respectively). SDS-PAGE analysis showed that when GAD65 is modified in vitro with 2 M D-ribose, there appeared to be an upshift in the molecular weight of GAD65 from bands at 116.8, 54.9 and 22.7 KDa to 117.5, 59.5 and 23.3 KDa in the GLY-GAD65 lane. Similarly, when assessing the band intensity, there was a decrease in band intensity between NT-GAD65 and GLY-GAD65. With •OH-GAD65, there appeared to be 2 faint bands at 60.1 KDa and 53.3 KDa. In comparison to the NT-GAD65, the •OH-GAD65 had a decreased band intensity. Finally, the HOCl-GAD65 showed bands at 117.5, 60.1 and 23.5 KDa. Upon glycation of NT-ZnT8 there was an apparent up-shift in GLY-ZnT8 with distinct bands at 93.1 and 43.2 KDa compared to the NT-ZnT8, which showed clear bands at 83.4 and 36.6 KDa. In comparison, modifying NT-ZnT8 with the hydroxyl radical to yield •OH-ZnT8 showed a similar increase in molecular weight to 88.9, 72.9, and 38.9 KDa compared to the NT-ZnT8 bands at 83.4, 36.6 and 31.0 KDa. HOCl-ZnT8 appeared to have bands at a molecular weight of 77.2 and 31.4 KDa compared to 60.2 and 23.5 KDa seen in the NT-ZnT8 lane. A cut-off for antibody binding positivity was calculated from the mean IgG binding of the healthy controls towards native or oxPTM beta cell antigens (mean O.D+ 2*standard deviations), resulting in an O.D of 0.5 as the cut-off for the homemade ELISAs developed towards NT- and oxPTM- GAD65 and ZnT8. Using this cut-off, the prevalence of antibody binding towards modified or native beta cell antigens was assessed. This study found that 53.2% (36/69) of patients with new-onset T1D were positive for IgG antibodies to NT-GAD65, 43.49% (30/69) were positive for GLY-GAD65, 17.4% (12/69) were positive for •OH-GAD65 and 39.1% (27/69) were positive to HOCl-GAD65. When determining the binding of oxPTM-GAD65-Ab in patients with new-onset T1D, there was a significant difference in the antibody binding towards •OH-GAD65 compared to NT-GAD65 patients with T1D (mean O.D ± S.D, 1.01 ± 0.40 vs 0.80 ± 0.15, p-value = 0.065, respectively). When comparing the changes in IgG binding towards native or oxPTM-ZnT8 in the 45.3% (24/53) of patients with new-onset patients with T1D positive for ZnT8-Ab, there appeared to be no significant difference between antibody binding to NT- and GLY-ZnT8 (mean O.D ± S.D, 0.7292 ± 0.1873 vs 0.6118 ± 0.2464, p-value= 0.064, respectively) or NT- and HOCl-ZnT8 (mean O.D ± S.D, 0.7292 ± 0.1873 vs 0.6285 ± 0.2505, p-value= 0.1140, respectively). However, there was an apparent significant decrease in IgG binding between NT- and •OH-ZnT8 (mean O.D ± S.D, 0.73 ± 0.19 vs 0.27± 0.14, p <0.0001, respectively). When assessing the antibody response towards NT- or oxPTM-INS, there was a significant decrease with antibody binding towards NT-INS and GLY-INS (mean O.D ± S.D, 0.221 ± 0.108 vs 0.170 ± 0.081, p-value <0.05, respectively). However there was a significant increase in antibody binding seen between NT-INS and HOCl-INS (mean O.D ± S.D, 0.221 ± 0.108 vs 0.363 ± 0.134, p-value <0.0001, respectively), as well as a significant increase between antibody binding towards NT-INS and •OH-INS (mean O.D ± S.D, 0.221 ± 0.108 vs 0.400 ± 0.129, p-value <0.0001, respectively). The correlation of the induced IgG response towards NT- or oxPTM-GAD65 was assessed by determining the degree of overlap in the patients towards each antigen. Within the total cohort of N=69 patients with new-onset T1D, 17% of patients were positive for an IgG antibody response towards NT-, Gly-. •OH- and HOCl-GAD65. With a total of 41% (29/69) appearing positive for GLY-GAD65 and 39% (27/69) towards •OH-GAD65. However, 45% of T1D were negative for antibodies directed towards all antigens (NT- and oxPTM-GAD65). When assessing antibody responses to NT- and oxPTM-ZnT8, 59% (41/69) of patients were negative for any antibody response, whereas 33% (23/69) were positive for antibodies against NT-GAD65. There were only 4.3% (3/69) patients with T1D who were positive to both NT- and all oxPTM-ZnT8, 25% (17/69) were positive for IgG antibodies to GLY-ZnT8 and HOCl-ZnT8, and only 5.8% (4/69) were positive to •OH-ZnT8. Discussion Research has indicated that PTMs may play a role in the progression of autoimmune diseases. PTMs are able to change the secondary and tertiary structure of proteins, impacting self-tolerance and immunogenic mechanisms. Although the majority of PTMs work to better the function of proteins, those that arise spontaneously, as a result of a breakdown in homeostasis, can be attributed to the development of autoimmune disorders. Multiple sclerosis, rheumatoid arthritis, and juvenile idiopathic arthritis are just a few of the autoimmune disorders with post-translationally modified autoantigens being named as contributors to the progression of these diseases. This study suggests that oxPTMs to native β-cell antigens alters the subsequent antigenicity of the targeted proteins and contributes to the progression of clinical T1D in people at-risk of developing T1D. When comparing the reactivity between NT-GAD65 and oxPTM-GAD65, this study showed significantly higher antibody reactivity towards GLY-GAD65, HOCl-GAD65 and NT-GAD65 in patients with new- onset T1D compared to healthy controls and people with T2D. However, there was no significant difference between the IgG antibody reaction towards GLY-GAD65, HOCl-GAD65 and NT-GAD65 in patients with T1D, yet certain patients showed an increased reactivity to GLY-GAD65 compared to NT-GAD65. This suggests the possible role of hyperglycemia or dysregulated glucose in modifying GAD65 in the initial stages before the onset of clinical T1D. In comparison, a significant decrease was seen in the antibody response towards oxPTM-ZnT8 compared to NT-ZnT8, indicating that oxidative modifications to ZnT8 may not impact the immunogenicity of the antigen. Controversially, oxPTMs appeared to increase the immunogenicity of insulin. In fact, patients with T1D developed significantly higher antibody responses towards •OH-INS compared to NT-INS. There was also a significant difference in antibody reactivity to •OH-INS in patients with T1D compared to patients with T2D and HC. These findings may suggest that the oxidative modification of insulin, specifically through the hydroxyl radical, may be an initial target of immunity in patients at-risk of developing T1D. A phenomena known as molecular mimicry has been suggested to play a role in the spread of autoantigenicity. In molecular mimicry, the sequence similarities between the autoantigen instigating the immune response and the native antigen lead to the immune system targeting the native antigen as well. Thus, ensuing epitope spreading and further stimulating the immune system. Upon the initial seroconversion, the immune system may begin to target the remaining beta cell antigens, such as GAD65 and ZnT8, as well as NT-INS due to the sequence similarity between both NT-INS and •OH-INS. However, further studies are required to confirm the findings, with the recruitment of more patients in the initial phases of T1D. Moreover, it is necessary to identify the exact peptides being targeted by the immune system, and those being modified with oxidative radicals in each beta cell antigen. In conclusion, oxPTMs to the beta cell antigens, insulin, GAD65 and ZnT8, alter the immune response in patients with new onset T1D, however further studies are needed in order to determine the role they play in the development of autoimmunity before the onset of clinical T1D. Study 2: DiabeSARS; A Tale of Two Pandemics Abstract Background The COVID-19 pandemic was the major health concern of the last 3 years. The sudden onset, and widespread of the SARS-CoV-2 virus, necessitated the expedition of research to fully understand the pathogenesis of the virus and to develop effective treatments. Diabetic patients have consistently been associated with severe and potentially fatal disease, more so than their non-diabetic counterparts. Hence, there is a clear need to fully elucidate the underlying pathology that leads to more severe disease in diabetic patients. Studies have previously assessed the role of underlying inflammation and dysregulated glucose in both T1D and T2D in previous viral infections. However, within the COVID-19 pandemic, although diabetic patients only make up a small percentage of the general population, they seem to comprise a similar percentage of critically ill patients, requiring admission to the intensive care unit as their non-diabetic counterparts. Studies have found the dysregulated glucose, regardless of diabetic state, promotes detrimental viral infections. Yet, the role of the endocrine system on immunometabolic outcomes has not been fully elucidated in terms of COVID-19. Increased hyperglycemia in diabetic patients has been associated with increased potential of glycation to native proteins, in fact, glycated hemoglobin (HbA1c) is consistently measured to monitor the development and progress of diabetes. Studies have demonstrated the glycosylation of both the SARS-CoV-2 spike (S) glycoprotein and the angiotensin converting enzyme 2 (ACE2), the main SARS-CoV-2 receptor. The induced glycosylation has been linked to increased virulence of the SARS-CoV-2 virus, as well as increased host-cell infiltration potential. However, the impact of non-enzymatic glycation, a potential outcome due to the elevated levels of glucose in an overly distressed system, has not been fully elucidated. Additionally, the efficacy of the developed SARS-CoV-2 mRNA vaccine (Pfizer, BioNtech) has not been assessed in terms of glucose control in diabetic patients. Hypothesis The main hypothesis is that dysregulated glucose levels may induce glycation of the SARS-CoV-2 S protein, altering the virulence of the disease. With one of the main focuses of this study to determine whether glucose regulation plays a role in vaccine efficacy and, in turn, immune protection against COVID-19. Aim 1. To induce glycation of the S protein and asses ACE2 binding to both native (NT-) and glycated (GLY-) S protein. Aim 2. To determine the impact of antibody binding to NT- and GLY-S protein in previous COVID-19 patients. Aim 3. To evaluate the role of glucose levels and monitoring on the protective immune response in patients following the administration of the SARS-CoV-2 mRNA vaccine (Pfizer-BioNTech, BNT162b2) Aim 4. In order to better understand the role of hyperglycemia and diabetes in the adverse outcomes of COVID-19, the fourth aim of this study is to evaluate the clinical risk of diabetes and glucose levels on mortality in patients with COVID-19. Methods Antibody response to NT- or GLY-S protein in patients with COVID-19 and vaccinated patients COVID-19 patients with and without diabetes were recruited for this study from specifically designated COVID-19 wards. For the vaccine study, patients were screened and recruited from the Endocrinology and Diabetology unit at Policlinico Campus Biomedico di Roma. With n= 26 patients being T1D and n=32 being diagnosed with T2D. The inclusion criteria for the study were patients >18 years old, scheduled to receive the SARs-CoV-2 mRNA vaccine (Pfizer-BioNTech), signing informed consent, having a diagnosis of T1D or T2D for more than 3 months, and using at least two anti-diabetic drugs in the case of T2D according to the vaccine priority criteria. Demographic and clinical data was collected at each time point involved in the study for all the patients recruited. The study covered a total of 6 months with the following timepoints: T0 (baseline): before the administration of SARs-CoV-2 mRNA vaccine (within 3 days of the first dose) T1: 21 days after the first dose (day of the second dose) T2: 35 days from baseline (T0) T3: 90 days from baseline (T0) T4: 180 days from baseline (T0). The SARS-CoV-2 S protein (0.450 mg/mL) was modified via glycation with equal volumes of 0.5 M D-ribose (sigma), this was incubated overnight at 37⁰C and stored short term at 4⁰C. The induced modifications were monitored by SDS-PAGE analysis, whereby 5 μg of NT- or GLY-S protein were loaded into the gel wells with equal volumes of Laemmli loading buffer with or without β-mercaptoethanol to promote reducing conditions. The gels were imaged with the Chemi-Doc imaging system from Bio-Rad and quantified with Imagej software. Changes in ACE2 binding towards NT- or GLY-S protein was determined via a homemade ELISA, whereby increasing dilutions of ACE2 (2 μg/mL-0.25 μg/mL) were measured against the same concentration of NT- and GLY-S protein (1 μg/mL). Moreover, COVID-19 patients with and without diabetes were assessed via a homemade ELISA for the antibody response towards NT- and GLY-S protein. The antibody responses were then analyzed depending on the fructosamine levels, measured via a kit (ab228558, abcam). The antibody response induced after immunization with the SARS-CoV-2 mRNA vaccine was measured via a homemade ELISA devised to analyze the IgG response towards NT- or GLY-S protein. Serum samples were collected and assessed at each timepoint and compared to the results of controls without diabetes. The level of neutralization antibodies was measured via a devised neutralization assay using live SARS-CoV-2 (Vero E6 cells). Devising a Clinical Risk Score to Assess in-hospital Death from COVID-19 Data from patients for devising the clinical risk score was collected retrospectively from 417 COVID-19 patients admitted to Jaber Al-Ahmed Hospital in Kuwait between February 24th and May 3rd, 2020. Due to the emergency state of the COVID-19 pandemic, the need for signed consent was waived by the ethical committee from the Ministry of Health in Kuwait. Setting the primary outcome as in-hospital death, a series of multivariant logistical regression models were performed to identify independent factors that may be prognostic for the primary outcome. The models were created by adding or removing variables individually depending on the results of the previous logistic regressions, with variables showing a p-value<0.1, being retained in the score. The independent predictive variables included in the final model were gender, asthma, glucose categories, and non-Kuwaiti national. Weighted points were assigned to significant risk factors proportional to their beta regression coefficient values. The effectiveness of the risk score to predict mortality in patients with COVID-19 was analyzed via receiver operating characteristic (ROC) curves, with an AUC of 0.5 or less was taken as insignificant Youden’s index was applied to set a cut-off for mortality prediction. Significance was set as a two-tailed p-value <0.05. The score was built using SPSS (IBM Corp. IBM SPSS Statistics for Windows, Version 21.0. Released 2012. Armonk, NY: IBM Corp.). The score was internally validated by a Kuwaiti COVID-19 cohort of N=923 patients, and externally validated using the CoViDiab cohort from Italy (N= 178). Results Antibody response to NT- or GLY-S protein in COVID-19 patients and vaccinated patients Upon assessing the binding of NT- and GLY-S protein (1μg/mL) to serial concentrations of the SARS-CoV-2 receptor ACE2 (2.0 μg/mL to 0.25 μg/mL) there was diminished binding of ACE2 to NT-S vs GLY-S, however this was not significant (mean ± S.D, 0.1393 ± 0.1732 vs 0.1943 ± 0.2355, p-value= 0.6026, respective). Fructosamine levels, which are a short-term assessment of glucose control, were measured in the total COVID-19 patient cohort (N=46). Upon stratification of fructosamine levels, patients with <563 μmol/L had no significant difference between IgG levels to NT- or GLY-S, yet patients with fructosamine levels >563 μmol/L had a significant difference between NT- and GLY-S (mean O.D ± S.D= 0.7945 ± 0.2564 vs 0.6155 ± 0.2369, p-value= 0.0078, respectively). When correlating fructosamine levels with the total cohort (N=46), there appeared to be an overall negative correlation to both NT- and GLY-S, yet this was not significant (r= -0.1848, p-value= 0.2188 vs r= -0.2203, p-value= 0.1413, respectively). In diabetic patients (n=22) there appeared to be no correlation with fructosamine levels to the IgG response to either NT- or GLY-S protein (r= 0.04029, p-value= 0.8587 vs r= 0.010201, p-value= 0.9640). However, in the case of non-diabetic COVID-19 patients, there was a strong negative correlation between fructosamine levels and the immune response to NT- and GLY-S (r= -0.3824, p= 0.1462 vs r=-0.4042, p= 0.0501, respectively). This study attempted to assess the IgG response towards the COVID-19 Pfizer BioNTech mRNA vaccine in terms of continuous glucose monitoring (CGM) data from N=10 patients with T1D with available CGM profiles. The key measurements assessed were Time-in-range (TIR), which is the percentage of time within a 24 hour period that a patient is within the ideal glucose range, Time-above-range (TAR), which is the percentage of time a patient is above the ideal glucose range in a 24 hour period, and Time-below-range (TBR), which is the percentage of time a patient is below their ideal glucose range. When determining the overall area under the curve (AUC) of the IgG response over all study timepoints (T0-T4) and correlating it with the average TIR, there appeared to be a strong correlation with AUC IgG response and TIR to both NT- and GLY-S protein (r= 0.8082, p-value= 0.0084 vs r= 0.7996, p-value= 0.0097, respectively), with a significant difference between the AUC IgG to NT- and GLY-S protein (p-value= 0.0034). When correlating the AUC IgG of both NT- and GLY-S protein to the average TAR over all study timepoints (T0-T4), there was a strong negative correlation (r= -0.7926, p-value= 0.0108 vs r= -0.7430, p-value= 0.0218, respectively) There was no correlation with HbA1c at T0 or average TBR with AUC IgG response towards native or GLY-S protein. When dividing patients with T1D with CGM data (N=13) based on their recommended glucose targets, (TIR > 70% and TBR <25%), there appeared to be a stronger neutralizing antibody response to the native SARs-CoV-2 spike protein who in patients with T1D had a TIR>70% than those who did not (p<0.0001). Furthermore, when assessing the neutralizing antibody response against TBR measurements, patients who had a TBR<25% were more likely to have a stronger neutralizing antibody response (p=0.008), this was seen regardless of HbA1c levels. Devising a Clinical Risk Score to Assess in-hospital Death from COVID-19 The score was built by assessing the significance of several predictive variables against the primary outcome (in-hospital mortality). The final score included asthma, gender (male), nationality (non-Kuwaiti national), and blood glucose levels (either between 7.0-11.1 mmol/L or >11.1 mmol/L) as independent predictors of mortality in COVID-19. A point system was given to each predictive variable based on the beta coefficients allocated to each variable. The cut-off of the score to predict death was 5.5, showing a specificity of 86.3% and sensitivity of 75% (AUC= 0.901). The clinical risk score requires internal and external validation to assess the potential to predict the primary outcome. Two cohorts were used for internal validation, the initial N=417 Kuwaiti COVID-19 group used to build the score and a separate cohort of N=923 Kuwaiti COVID-19 patients admitted from May 4th to August 26th, 2020, both admitted within one COVID-19 center within Kuwait. External validation was performed using an N=178 CoViDiab Italian cohort. The score was calculated for each patient and then tested against the primary outcome (in-hospital mortality from COVID-19); the score was then plotted as a ROC curve with the AUC calculated. The AUC showed 0.901 ± 0.20 fit for the score for the 417 Kuwaiti cohort, 0.826 ± 0.91 fit for the score for the 923 Kuwaiti cohort, and a 0.687 ± 0.06 fit for the score for the CoViDiab cohort, with respective negative predictive values of 95.4%, 93.9%, and 94.1%. Conclusions Diabetic patients are characterized by hyperglycemia and chronic low-grade inflammation. Upon viral infection, these patients present with more exacerbated immune responses, that may lead to severe disease, ICU admission, and potentially death. The findings in this study suggest that dysregulated fructosamine levels are more strongly correlated with a decreased antibody response towards NT- and GLY-S protein of SARS-CoV-2. Additionally, when assessing the IgG and neutralization antibody response in patients with T1D in association with CGM data, there appeared to be a stronger association with more improved TIR and TAR glucose measurements and a stronger antibody response. This again suggests that better controlled glucose measurements aid in improving the protective immune response towards SARS-COV-2. Finally, the development of the clinical risk score demonstrated that elevated glucose was a stronger predictor of negative outcomes and mortality in COVID-19 related infections than diabetes. In fact, the addition of glucose measurements removed diabetic state as an independent predictor of in-hospital death. In conclusion, maintaining key glucose targets may aid in preventing detrimental outcomes towards not only COVID-19 and other viral infections.
22-mar-2023
Autoimmunity and Antibody Response in Diabetes: The Role of Hyperglycemia and Antigen Post Translational Modifications / Ghadeer Alhamar , 2023 Mar 22. 35. ciclo, Anno Accademico 2019/2020.
File in questo prodotto:
File Dimensione Formato  
PhD_Alhamar_Ghadeer_.pdf

accesso aperto

Descrizione: PhD_Alhamar_Ghadeer
Tipologia: Tesi di dottorato
Licenza: Creative commons
Dimensione 5.33 MB
Formato Adobe PDF
5.33 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.12610/71903
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact