Immunogenicity and Safety of SARS-CoV-2 mRNA Vaccines in a Cohort of Patients With Type 1 Diabetes

Patients affected by both type 1 diabetes (T1D) and type 2 diabetes may develop more severe outcomes during severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection (1,2), and glycometabolic control is altered during coronavirus disease 2019 (COVID-19) (3,4). The recent availability of novel vaccines against SARS-CoV-2 (5) suggests that primary prevention based on vaccination may be a key strategy to dampen the risks associated with COVID-19 in patients with T1D. Because an impaired humoral/cellular immune response to vaccination was hypothesized in patients with diabetes (6–8), careful consideration of age and timing of disease onset has been proposed when considering vaccination for patients with T1D. While benefits are known for common vaccinations, the use of novel vaccines, such as that for SARS-CoV-2, may be challenging (8–10). Given the more severe outcomes during SARS-CoV-2 infection (11), patients with T1D should be a priority group to receive the COVID-19 vaccine (12), although studies in this population are lacking (13). In the current study, we assessed the safety, immunogenicity, and glycometabolic effects of the SARS-CoV-2 mRNA vaccines in patients with T1D and provide new evidence for an impaired SARS-CoV-2–specific cellular cytotoxic immune response in patients with T1D.

This monocentric observational study was conducted at the Endocrinology and Diabetology Unit, ASST Fatebenefratelli Sacco in Milan, Italy, according to the principles of the Declaration of Helsinki, and the protocol was approved by the Comitato Etico Milano Area 1 (Effect of SARS-CoV-2 [COVID-19] Vaccination in Type 1 Diabetes [CoVaxT1D]). A total of 375 subjects (326 patients with T1D and 49 subjects without diabetes [frail patients without diabetes as defined on the basis of guidelines issued by the Italian Ministry of Health and caregivers/family members of patients with T1D) (Table 1) received two doses of SARS-CoV-2 mRNA vaccines (Supplementary Table 1) between 26 March and 18 April 2021 and signed informed consent (Supplementary Fig. 1). Patients with active COVID-19 or a diagnosis of COVID-19 within the previous 3 months were excluded. The primary outcomes of the current study were the safety of the mRNA vaccines in patients with T1D, as measured by report of local/systemic adverse events; humoral immunogenicity, as measured by anti-SARS-CoV-2 spike antibody levels; and cell immunogenicity, as measured by an increase in SARS-CoV-2 cytotoxic cellular response in vitro. The secondary outcome was the effect of the mRNA vaccines on glycemic control in patients with T1D, as measured by continuous glucose monitoring (CGM) (3). Samples and data were analyzed before vaccination and after doses 1 and 2.

The humoral response was analyzed using Elecsys Anti-SARS-CoV-2 serology test (Roche Diagnostics, Basel, Switzerland), which captured IgG, IgM, and, to a lesser extent, IgA against SARS-CoV-2 spike antigen (receptor-binding domain), with signal-to-cutoff values of ≥1.0 reported as positive (14). An in vitro neutralization assay was used as previously described (15) to validate anti-SARS-CoV-2 antibody titers detected in seronegative vaccinated patients with T1D by performing a Spearman correlation analysis between antibody titers detected by the neutralization and Elecsys assays.

Peripheral blood mononuclear cells (PBMCs) (1 × 10⁶) (16) obtained from frail subjects or patients with T1D before and after vaccination were cultured with and without SARS-CoV-2 spike and nucleocapsid recombinant proteins (PR-nCoV-3/PR-nCoV-1 500 ng/mL; Novatein Biosciences, Inc., Woburn, MA) for 48 h (17). Vaxigrip (quadrivalent split-virion influenza vaccine peptides 0.25 μg/mL; Sanofi-Pasteur) and glutamate transporter-1 50 μg/mL (GenScript, Piscataway, NJ) were included as positive and negative controls. Supernatants were tested for perforin/granzyme A (HCD8MAG-15K; Merck Millipore, Burlington, MA) and T-cell–related cytokines (M500KCAF0Y; Bio-Rad Laboratories, Hercules, CA) by a Bio-Plex 200 system. An increase in cytotoxic response was assessed as a fold increase >0.1 for perforin and >0.01 for granzyme A, respectively. Cut points of 0.1 and 0.01 were defined on the basis of the difference detected for each analyte, measured at each time point, and normalized to the baseline value. To test the specific response to SARS-CoV-2 antigens, the number of interferon-γ (IFN-γ)–producing cells (BD Biosciences, Franklin Lakes, NJ) was assessed by ELISpot assay as previously described (17).

Fisher exact test, one-way and two-way ANOVA, Kruskal-Wallis test adjusted for multiple comparisons, and Mann-Whitney U test were used in the analyses of baseline characteristics, side effects, humoral and cellular response, and CGM parameters. Multivariable logistic regression was used to model the relationships between cofactors and cellular response (Stata 12 software; StataCorp). Statistical significance was determined at α <0.05. GraphPad Prism 7 (GraphPad Software, San Diego, CA) was used to generate graphs.

The data sets and resources generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.

We included 375 patients consecutively administered a SARS-CoV-2 mRNA vaccine, of whom 326 had long-term T1D (22.9 ± 14.2 years) and 49 had no diabetes but were considered to be medically frail or were caregivers/family members of patients with T1D (Supplementary Fig. 1 and Table 1). The mean age of patients with T1D was 45.9 ± 9.0 years, and 180 (55%) were male. The mean age of subjects without diabetes was 61.0 ± 12.9 years, and 21 (43%) were male (Table 1). No differences were observed with regard to type of vaccine administered (Supplementary Table 1), BMI, concomitant therapy (except for insulin treatment), and major clinical comorbidities, with a higher frequency of autoimmune thyroiditis observed in patients with T1D (Table 1).

No differences were observed between patients with T1D and subjects without diabetes with regard to local and systemic side effects after vaccination (Supplementary Fig. 2A and B). The most common side effect was local pain reported equally after dose 1 and 2 of the vaccine (72.6% vs. 73.7%, P = 0.78) (Supplementary Fig. 2A and B). Systemic adverse events, consisting mainly of weakness and fever, were more frequent after dose 2 (Supplementary Fig. 2A and B).

With regard to SARS-CoV-2 humoral response, patients with T1D who were seronegative for COVID-19 showed an increase in median IgG levels after both dose 1 and dose 2 (60.4 and 2058.0 units/mL, respectively) (Fig. 1A–C). This observation was paralleled in subjects without diabetes at both time points (46.3 and 1,965.0 units/mL, respectively) (Fig. 1A–C), with a response rate comparable between the two groups. These data indicate that the administration of two doses of the SARS-CoV-2 mRNA vaccine elicited a humoral response in patients with T1D. In a subgroup analysis performed on 15% (49 of 326) of patients with T1D and 14.3% (7 of 49, P = 0.99) of subjects without diabetes who underwent full vaccination but were seropositive for COVID-19 (i.e., experienced asymptomatic SARS-CoV-2 infection, baseline signal-to-cutoff IgG value of ≥1.0), the analysis of the humoral response revealed no differences between the two groups (Supplementary Fig. 3). Finally, our assay captured neutralizing activity as shown by a positive correlation between anti-SARS-CoV-2 titers determined by our assay with titers of neutralizing antibodies as measured by an in vitro neutralization assay (Fig. 1D).

To address whether an immune cytotoxic response against SARS-CoV-2 was elicited with vaccination, we rechallenged isolated PBMCs with SARS-CoV-2 peptides at baseline and after doses 1 and 2 in a MILLIPLEX MAP Human assay. Patients with T1D showed no increase in granzyme A and perforin SARS-CoV-2–specific T-cell release in the supernatant after vaccination compared with baseline, while a 2–3-fold increase in both granzyme A and perforin production was evident in subjects without diabetes (Fig. 2A and B), which was also confirmed by testing negative (<1-fold increase) and positive (10- and 7-fold increase for granzyme A and perforin, respectively) controls. Interestingly, only 30% of patients with T1D showed an increased SARS-CoV-2–specific T-cell cytotoxic response after dose 1 compared with 80% of subjects without diabetes (Fig. 2C). This observation was confirmed after dose 2 in patients with T1D, with 22% developing/maintaining an increased SARS-CoV-2–specific T-cell cytotoxic response, although the rate of subjects without diabetes developing a cytotoxic response was lower (57%) (Fig. 2D). Furthermore, the reduced cytotoxic response detected in T1D was confirmed after controlling for other factors potentially involved (age, sex, and concomitant diseases) (Supplementary Table 2). These findings suggest that SARS-CoV-2 mRNA vaccines fail to elicit a robust cytotoxic response against SARS-CoV-2 in the majority of patients with T1D. Finally, a multiplex T-cell–related cytokine analysis revealed lower levels of interleukin-2 and tumor necrosis factor-α measured in the supernatant of PBMCs cultured with SARS-CoV-2 peptides in patients with T1D compared with subjects without diabetes after vaccination (Fig. 3A), which was associated with a low response in an IFN-γ ELISpot assay (Fig. 3B), regardless of the type of mRNA vaccination administered, thus confirming the existence of an altered immune cellular response to SARS-CoV-2 mRNA two-dose vaccination in T1D.

Given that COVID-19 is associated with the development of dysglycemia and diabetes, we finally assessed whether the administration of SARS-CoV-2 mRNA vaccine may affect glycometabolic control in patients with T1D. Data collected using CGM showed no difference in all the parameters examined at each time point (Supplementary Table 3). Moreover, a multivariable analysis showed no association between estimated HbA_1c levels and cytotoxic response to vaccination in patients with T1D after adjusting for age, sex, and concomitant therapies (P = 0.23 and P = 0.65 for response to doses 1 and 2, respectively). These findings suggest that the administration of the SARS-CoV-2 mRNA vaccine is not associated with altered glycometabolic control in patients with T1D.

Our study evaluated the safety, immunogenicity, and glycometabolic effects of two doses of SARS-CoV-2 mRNA vaccines in patients with T1D. No differences were observed with regard to the humoral response or the incidence of side effects in patients with T1D compared with subjects without diabetes, and no effects on glycometabolic control were observed using CGM analysis during follow-up. Interestingly, in the majority of patients with T1D, both dose 1 and dose 2 of SARS-CoV-2 mRNA vaccines were not associated with an increase in cytotoxic factors granzyme A and perforin or with an increase in the production of the T-cell–related cytokines interleukin-2 and tumor necrosis factor-α compared with subjects without diabetes, who showed an increased rate of both cytotoxic factors and cytokines after vaccination. The SARS-CoV-2 mRNA vaccines have been shown to elicit a humoral and T-cell–specific immune response in healthy subjects (18,19), but little is known about their effects in patients with T1D. An impairment in the immune response following vaccination in patients with T1D has been previously hypothesized for other immunization strategies, such as influenza, rotavirus, and Haemophilus influenzae type B (6,8), and attributed to an impaired cellular immune response. Therefore, boosting immunization strategies with multiple doses has been recommended (8,20). Indeed, immune dysregulation of both T- and B-cell compartments is a key feature of patients with T1D from the onset of the disease and may indicate abnormalities in the immune response (21,22), including a higher risk of developing infectious disease (23,24). This was particularly evident during the COVID-19 pandemic, with patients with T1D experiencing severe COVID-19 disease and worsened outcomes compared with the general population without diabetes (1,2,11). In our study, patients with T1D did not show an increase in T-cell–specific SARS-CoV-2 cytotoxic factor release (granzyme A and perforin), which may indicate a defective ability in inactivating the virus. Indeed, perforin and granzyme A levels, as observed for other vaccines and immunization strategies (25,26), detected after vaccination increased in subjects without diabetes in response to SARS-CoV-2 in vitro challenge, while no effect was observed in the majority of patients with T1D. This suggests a reduced cytotoxic effector function in T1D and a less immunogenic/efficient vaccination in these patients. Nevertheless, no significant differences were observed in SARS-CoV-2 spike–specific neutralizing antibody titers between patients with T1D and subjects without diabetes. To our knowledge, this study is the first to demonstrate the safety and immunogenicity of SARS-CoV-2 mRNA vaccines in a large cohort of patients with T1D without any significant perturbation in CGM parameters (13,27).

We acknowledge that this study has some limitations. The study design may have led to potential selection bias primarily due to the COVID-19 pandemic, such as the smaller and not fully age-matched frail control population and the limited number of samples available for some analyses (e.g., ELISpot), and the assay used captured only one aspect of the vaccine-specific cellular response.

In summary, an impaired cytotoxic SARS-CoV-2–specific cellular immune response has been observed in the majority of patients with T1D following two doses of SARS-CoV-2 mRNA vaccines, with the humoral response and glycemic control unaffected.

Acknowledgments. The authors thank the “Fondazione Romeo e Enrica Invernizzi” for extraordinary support.

Funding. F.D. is supported by a Società Italiana di Diabetologia Lombardia grant and European Foundation for the Study of Diabetes grant EFSD/JDRF/Lilly Programme on Type 1 Diabetes Research 2019. P.F. is supported by Italian Ministry of Health grant RF-2016-02362512 and Università di Milano Linea-2 2019 funding.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. F.D. and G.S. designed the study, analyzed data, and wrote the manuscript. V.U., E.A., A.A., A.M., A.J.S, M.B.N., C.L., D.M., and G.R. performed experiments and analyzed data. I.P., L.M., P.S.M., L.P., A.R., E.C., A.M.B., E.L., C.M., and M.M. collected human data and samples. S.A., S.R., M.Gallie., C.B., F.F., M.Galli, M.R.G., and G.V.Z. coordinated and designed the research and edited the manuscript. P.F. conceived the idea, designed the study, and wrote and edited the manuscript. P.F. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.