Cytokine Response Following SARS-CoV-2 Antigen Stimulation in Patients with Predominantly Antibody Deficiencies

Predominantly antibody deficiencies (PADs) are inborn disorders characterized by immune dysregulation and increased susceptibility to infections. Response to vaccination, including severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), may be impaired in these patients, and studies on responsiveness correlates, including cytokine signatures to antigen stimulation, are sparse. In this study, we aimed to describe the spike-specific cytokine response following whole-blood stimulation with SARS-CoV-2 spike peptides in patients with PAD (n = 16 with common variable immunodeficiency and n = 15 with selective IgA deficiency) and its relationship with the occurrence of coronavirus disease 2019 (COVID-19) during up to 10-month follow-up period. Spike-induced antibody and cytokine production was measured using ELISA (anti-spike IgG, IFN-γ) and xMAP technology (interleukin-1β (IL-1β), IL-4, IL-6, IL-10, IL-15, IL-17A, IL-21, TNF-α, TGF-β1). No difference was found in the production of cytokines between patients with PAD and controls. Anti-spike IgG and cytokine levels did not predict contraction of COVID-19. The only cytokine that distinguished between vaccinated and naturally infected unvaccinated PAD patients was IFN-γ (median 0.64 (IQR = 1.08) in vaccinated vs. 0.10 (IQR = 0.28) in unvaccinated). This study describes the spike-specific cytokine response to SARS-CoV-2 antigens, which is not predictive of contracting COVID-19 during the follow-up.


Introduction
Predominantly antibody deficiencies (PADs) are a group of disorders characterized by impaired B cell differentiation and decreased synthesis of immunoglobulins, resulting in recurrent sinopulmonary infections and, in some cases, poor vaccine response [1]. The most common diagnoses include selective IgA deficiency (SIgAD) and common variable immunodeficiency (CVID). Patients with SIgAD have reduced serum immunoglobulin A (IgA) levels below 0.07 g/L, with normal levels of other immunoglobulins. They are often asymptomatic and do not usually have an impaired response to vaccination. Symptomatic patients may experience recurrent respiratory and gastrointestinal infections and have a higher risk of atopic or autoimmune diseases compared to the general population [2]. CVID is the most common symptomatic antibody deficiency. The European Society for Immunodeficiencies Registry working definition for CVID is decreased levels of immunoglobulins (IgG, IgA, and/or IgM) and low switched memory B cells or poor vaccine response in Peripheral venous blood samples were collected from all the participants. A peripheral blood sample for SARS-CoV-2 anti-spike IgG was collected in a serum collection tube, centrifuged after 30 min, and frozen at −20 • C until the day of measurement. A sample for peripheral blood mononuclear cell (PBMC) isolation and phenotyping of B and T lymphocyte subsets was collected in a lithium heparin-coated tube, and PBMCs were isolated as described previously [25]. Samples for the assessment of CD4+ and CD8+ cell cytokine responses to SARS-CoV-2 peptide stimulation and cytokine determination before/after SARS-CoV-2 peptide stimulation were collected in heparinized whole-blood QuantiFERON SARS-CoV-2 blood collection tubes (Qiagen, Hilden, Germany), incubated at 37 • C for 20 h, centrifuged according to the manufacturer's protocol, and frozen at −20 • C for subsequent analysis. QuantiFERON SARS-CoV-2 blood collection starter kit contains four types of blood collection tubes: negative control, Ag1 tube (contains epitopes from the Viruses 2023, 15, 1146 4 of 23 S1 fragment of the SARS-CoV-2 spike protein, measures response mediated by CD4+ cell response), Ag2 tube (contains epitopes from the S1 and S2 fragments of the SARS-CoV-2 spike protein, measures response mediated by both CD4+ and CD8+ cells), and positive control (mitogen control) [34,69]. Negative control was used to measure cytokine baseline levels and non-specific cytokine release during the incubation, while the Ag2 tube was used to measure cytokine levels after antigen stimulation.

Humoral Response to SARS-CoV-2
The humoral response to SARS-CoV-2 (IgG class antibodies to SARS-CoV-2 S1 domain of the spike protein) was assessed using the semi-quantitative enzyme-linked immunosorbent assay (ELISA) from Euroimmun (Anti-SARS-CoV-2 IgG assay, Euroimmun, Lübeck, Germany), following the manufacturer's recommendations and as described previously. An IgGAM ratio (optical density compared with the calibrator) > 1.1 was considered a positive response. SARS-CoV-2 anti-spike response was considered low if the levels were <1.1, moderate if the anti-spike IgG levels were between 1.1 and 10, and high if levels were >10.

Cytokine Detection
Cytokine levels before and after SARS-CoV-2 S1 and S2 pool peptide stimulation were determined by Luminex xMAP technology using Luminex 200 equipment (A DiaSorin Company, Austin, TX, USA) and a multiplex assay (MILLIPLEX MAP Human TH17 Magnetic Bead Panel, Cat#: HTH17MAG-14 K and MILLIPLEX MAP TGF-ß1 Magnetic Bead Single Plex Kit, Cat#: TGFBMAG-64K-01, both from Merck Millipore, Darmstadt, Germany), following the manufacturer's protocol. Levels of the following cytokines were measured: transforming growth factor beta 1 (TGF-ß1), IL-1β, IL-4, IL-6, IL-10, IL-15, IL-17A, IL-21, and TNF-α. Antigen-specific responses were quantified as the increase or decrease in cytokine levels in SARS-CoV-2 S1 and S2 pool peptide-stimulated samples compared with paired unstimulated negative controls. For the analysis, cytokines falling below the lowest detection limit were adjusted to 0.5, and absent values were removed.

T and B Cell Subset Phenotyping
T and B cell subpopulations from freshly isolated PBMCs were determined by flow cytometry as described previously [25].

Statistical Analysis
The Shapiro-Wilk test was used to determine whether continuous variables were normally distributed. The results indicated that the data were not normally distributed; therefore, medians and interquartile ranges (IQRs) were used in data presentation, and nonparametric statistical methods were used in subsequent analysis. The differences in categorical variables were examined using the chi-square and Fisher exact tests. The Mann-Whitney U or Kruskal-Wallis test was used to compare continuous variables between two or more groups, respectively. The Wilcoxon signed-rank test was used to compare two dependent samples. Spearman's rank test was used to assess the correlation between continuous variables. A binominal regression analysis was used to predict the occurrence of COVID-19. In binominal regression analysis, the dependent variable was whether COVID-19 was present during the follow-up period, while the independent variables were measured at the beginning of this period as follows: levels of anti-spike IgG, changes in cytokine levels (IFN-γ, TGF-β, IL-1β, IL-4, IL-6, IL-10, IL-15, IL-17A, IL-21, TNF-α) following whole-blood stimulation with SARS-CoV-2 S1 and S2 pool antigens, as well as levels of total IgG, IgA, and IgM, and different lymphocyte subpopulations as previously noted. Statistical significance was set at p value < 0.05. Statistical analysis was performed using IBM SPSS Statistics version 23 (IBM, New York, NY, USA). Graphs were generated using GraphPad Prism version 8 (GraphPad Software, San Diego, CA, USA).

Study Population
Overall, 38 individuals were enrolled in the study: 31 patients with predominantly antibody deficiency (29.0% male, median age 40 years, IQR = 22) and seven control subjects (28.5% male, median age 50 years, IQR = 21). In the patient group, 16 and 15 patients were diagnosed with CVID and SIgAD, respectively. The clinical characteristics of the participants are presented in Table 1. All patients and control group subjects had been previously exposed to the SARS-CoV-2 antigen at the time of blood sample collection via natural infection (4 patients, 0 controls), vaccination (12 patients, 3 controls), or both (15 patients, 4 controls). Of all vaccinated individuals (27 patients and 7 controls), the median time to completion of the vaccination regimen was 164 (IQR = 114) days: median 153 (IQR = 90) days in patients and 203 (IQR = 169) days in the healthy vaccine recipient group. The median duration between positive SARS-CoV-2 PCR test results and the date of evaluation for unvaccinated patients was 106 days (IQR = 144). Seven patients and two controls developed COVID-19 during the follow-up period, all of whom had mild disease according to the WHO classification (up to a score of 3) and were treated in an out-patient setting.

Changes in Cytokine Production Following SARS-CoV-2 Antigen Stimulation
To examine the changes in cytokine levels in response to SARS-CoV-2 antigen stimulation, we initially assessed the overall differences in cytokine production by comparing the baseline levels with those after stimulation in all study participants. Subsequently, we performed a similar analysis after excluding individuals who had not been vaccinated. Furthermore, we compared cytokine levels between the patient and control groups to determine whether significant differences existed in both baseline levels and changes in cytokine levels following antigen stimulation. Additionally, we examined the potential differences in cytokine levels and changes between patients diagnosed with common variable immunodeficiency (CVID) and selective IgA deficiency (SIgAD). All relevant statistical data are presented in Supplementary Table S1, which provides a detailed display of all pertinent statistical analyses.
Overall, we found a statistically significant increase in the levels of IFN-γ, IL-10, IL-15, IL-17A, IL-1β, and TNF-α, but not IL-21, IL-4, and IL-6, and a decrease in TGF-β1 following SARS-CoV-2 antigen stimulation (see Figure 1 and Supplementary Table S1a). After excluding patients who had not been vaccinated, changes were observed in the levels of the cytokines mentioned previously, except for changes in IL-15, which were not statistically significant (see Supplementary Table S1a). However, in the Mann-Whitney U test, we did not observe a significant difference between the patient and control groups regarding the increase in cytokine levels after SARS-CoV-2 antigen stimulation (see Supplementary Table  S1b). In contrast, such differences were observed in the baseline cytokine levels; patients were characterized by significantly elevated baseline levels of IL-10 and IL-4 compared with the control group (see Supplementary Table S1b).
In addition, when categorizing patients based on diagnosis, we observed significant changes in IFN-γ, IL-10, IL-15, and IL-4 levels in the SIgAD subgroup, whereas only changes in the levels of IFN-γ and TGF-β1 were observed in the CVID subgroup (see Supplementary Table S1a). We also found a statistically significant difference in the increase in IL-4 synthesis following SARS-CoV-2 antigen stimulation between different diagnosis groups: IL-4 did not increase in the CVID group (median 0 ng/mL), but we found a median 1.2 ng/mL increase in the SIgAD group (see Supplementary Table S1b).
Viruses 2023, 15, 1146 9 of 23 changes in IFN-γ, IL-10, IL-15, and IL-4 levels in the SIgAD subgroup, whereas only changes in the levels of IFN-γ and TGF-β1 were observed in the CVID subgroup (see Supplementary Table S1a). We also found a statistically significant difference in the increase in IL-4 synthesis following SARS-CoV-2 antigen stimulation between different diagnosis groups: IL-4 did not increase in the CVID group (median 0 ng/mL), but we found a median 1.2 ng/mL increase in the SIgAD group (see Supplementary Table S1b).

Predictors of COVID-19 during the Follow-Up Period in PAD Patients
To determine potential clinical parameters that may be associated with the occurrence of COVID-19 during the follow-up period, we conducted a binomial logistic regression analysis and employed the Mann-Whitney U test. All the relevant statistical data are presented in Supplementary Table S2. Binomial logistic regression analysis indicated that variables such as anti-spike IgG levels, cytokine levels following SARS-CoV-2 antigen stimulation, total IgG, IgM, and IgA levels, and lymphocyte subpopulations cannot be used to predict the likelihood of contracting COVID-19 (see Supplementary Table S2a). However, the Mann-Whitney U test demonstrated significant differences in cytotoxic T and NK cell subpopulations between patients who contracted COVID-19 during the follow-up period and those who did not (see Figure 2 and Supplementary Table S2b). However, the effect size of this test was moderate. After excluding patients who had not been vaccinated, significant differences in cytotoxic T cells, but not NK cells, were observed (see Supplementary Table S2b).
When categorizing patients based on their diagnosis, a consistent observation of differences in cytotoxic T cells, but not NK cells, was found in the CVID group, while in the SIgAD group, we did not find a statistically significant difference in lymphocyte subpopulations between the patients who contracted COVID-19 and those who did not.
We did not find a statistically significant relationship between the occurrence of COVID-19 during the follow-up period and anti-spike IgG or cytokine levels after the SARS-CoV-2 antigen stimulation (see Supplementary Table S2b).

Predictors of COVID-19 during the Follow-Up Period in PAD Patients
To determine potential clinical parameters that may be associated with the occurrence of COVID-19 during the follow-up period, we conducted a binomial logistic regression analysis and employed the Mann-Whitney U test. All the relevant statistical data are presented in Supplementary Table S2.
Binomial logistic regression analysis indicated that variables such as anti-spike IgG levels, cytokine levels following SARS-CoV-2 antigen stimulation, total IgG, IgM, and IgA levels, and lymphocyte subpopulations cannot be used to predict the likelihood of contracting COVID-19 (see Supplementary Table S2a). However, the Mann-Whitney U test demonstrated significant differences in cytotoxic T and NK cell subpopulations between patients who contracted COVID-19 during the follow-up period and those who did not (see Figure 2 and Supplementary Table S2b). However, the effect size of this test was moderate. After excluding patients who had not been vaccinated, significant differences in cytotoxic T cells, but not NK cells, were observed (see Supplementary Table S2b).
When categorizing patients based on their diagnosis, a consistent observation of differences in cytotoxic T cells, but not NK cells, was found in the CVID group, while in the SIgAD group, we did not find a statistically significant difference in lymphocyte subpopulations between the patients who contracted COVID-19 and those who did not.
We did not find a statistically significant relationship between the occurrence of COVID-19 during the follow-up period and anti-spike IgG or cytokine levels after the SARS-CoV-2 antigen stimulation (see Supplementary Table S2b).

Associations between the Cytokine Levels and Previous SARS-CoV-2 Vaccination
To ascertain potential differences in cytokine levels between vaccinated and unvaccinated patients and controls, the Kruskal-Wallis test was conducted. Furthermore, we assessed whether the cytokine response was correlated with the type of vaccination received within the patient group. All the relevant statistical data are presented in Supplementary Tables S3 and S6. The Kruskal-Wallis test indicated a significant difference in the median increase in IFN-γ levels in the vaccinated, unvaccinated, and control groups (see Figure 3a and Supplementary Table S3a). When categorizing patients based on their diagnosis, a consistent observation of a difference between vaccinated and unvaccinated individuals was found in the CVID group, but not in the SIgAD group (see Supplementary Table S3b). No other significant differences were found in the increase in cytokine levels after SARS-CoV-2 antigen stimulation between the vaccinated and unvaccinated patients or vaccinated controls.

Associations between the Cytokine Levels and Previous SARS-CoV-2 Vaccination
To ascertain potential differences in cytokine levels between vaccinated and unvaccinated patients and controls, the Kruskal-Wallis test was conducted. Furthermore, we assessed whether the cytokine response was correlated with the type of vaccination received within the patient group. All the relevant statistical data are presented in Supplementary  Tables S3 and S6. The Kruskal-Wallis test indicated a significant difference in the median increase in IFN-γ levels in the vaccinated, unvaccinated, and control groups (see Figure 3a and Supplementary Table S3a). When categorizing patients based on their diagnosis, a consistent observation of a difference between vaccinated and unvaccinated individuals was found in the CVID group, but not in the SIgAD group (see Supplementary Table S3b). No other significant differences were found in the increase in cytokine levels after SARS-CoV-2 antigen stimulation between the vaccinated and unvaccinated patients or vaccinated controls. Furthermore, there was no significant correlation between the type of vaccine administered and changes in cytokine levels (see Supplementary Table S3c).
Regarding the cytokine response correlation with time after the last vaccine dose, the only cytokine whose level changed after antigen stimulation correlated with time after vaccination was TGF-β1 (in the patient group) (Figure 3b and Supplementary Table S6c).  Furthermore, there was no significant correlation between the type of vaccine administered and changes in cytokine levels (see Supplementary Table S3c).

Associations between the Cytokine Levels and SARS-CoV-2 Humoral Response
Regarding the cytokine response correlation with time after the last vaccine dose, the only cytokine whose level changed after antigen stimulation correlated with time after vaccination was TGF-β1 (in the patient group) (Figure 3b and Supplementary Table S6c).

Associations between the Cytokine Levels and SARS-CoV-2 Humoral Response
We also assessed whether the cytokine response to SARS-CoV-2 antigen stimulation was related to the anti-spike IgG antibody levels. All the relevant statistical data are presented in Supplementary Tables S4 and S6.
No significant correlation was observed between the levels of anti-spike IgG and the increase in cytokine levels following SARS-CoV-2 antigen stimulation in patients with PAD (see Supplementary Table S6c). However, when categorizing patients based on low, moderate, or high anti-spike antibody response, patients with low antibody response exhibited a higher median decrease in TGF-β1 levels after stimulation with SARS-CoV-2 antigen and compared to patients with a moderate or high humoral response or control group individuals (see Figure 4 and Supplementary Table S4a).

Associations between the Cytokine Levels and Patient's Demographic and Clinical Characteristics
We also assessed whether there were any correlations between age and cytokine responses to SARS-CoV-2 antigen stimulation as well as any differences in cytokine responses in relation to various clinical manifestations. All relevant statistical data are presented in Supplementary Tables S5 and S6. In the patient group, age was significantly correlated with changes in the levels of TGF-β1 and IL-1β following SARS-CoV-2 antigen stimulation (see Supplementary Table  S6a).
Distinct variations were observed in the cytokine profiles when comparing various complications and comorbidities. The Kruskal-Wallis test indicated that there was a significant difference in the median increase in the level of IL-4 and the frequency of detected autoantibodies in personal medical history, while in the control group or patients without autoantibodies, no increase was observed; in patients who had detected autoantibodies in their personal medical history, we found a median 1.26 ng/mL increase in the levels of IL-4 following SARS-CoV-2 antigen stimulation (see Figure 5a and Supplementary Table  S5a). Additionally, baseline IL-4 and TNF-α levels were higher in patients with autoanti-

Associations between the Cytokine Levels and Patient's Demographic and Clinical Characteristics
We also assessed whether there were any correlations between age and cytokine responses to SARS-CoV-2 antigen stimulation as well as any differences in cytokine responses in relation to various clinical manifestations. All relevant statistical data are presented in Supplementary Tables S5 and S6. In the patient group, age was significantly correlated with changes in the levels of TGF-β1 and IL-1β following SARS-CoV-2 antigen stimulation (see Supplementary Table S6a).
Distinct variations were observed in the cytokine profiles when comparing various complications and comorbidities. The Kruskal-Wallis test indicated that there was a significant difference in the median increase in the level of IL-4 and the frequency of detected autoantibodies in personal medical history, while in the control group or patients without autoantibodies, no increase was observed; in patients who had detected autoantibodies in their personal medical history, we found a median 1.26 ng/mL increase in the levels of IL-4 following SARS-CoV-2 antigen stimulation (see Figure 5a and Supplementary Table S5a). Additionally, baseline IL-4 and TNF-α levels were higher in patients with autoantibodies. We found no association between increased cytokine levels, including IL-4, after SARS-CoV-2 antigen stimulation and clinically detectable autoimmune diseases (see Supplementary  Table S5a). However, when categorizing patients according to diagnosis, in the CVID group, TGF-β1 levels were related to the presence of autoimmune disease (see Supplementary  Table S5c). An increase in the level of IL-4 following SARS-CoV-2 antigen stimulation, but not the baseline IL-4 level, was also correlated with the levels of total IgG and total IgM (see Figure 6 and Supplementary Table S6c). . (e) Mann-Whitney U test results for associations between the changes in TNF-α following SARS-CoV-2 antigen stimulation in patients with B+SmB-(n = 4), B+SmB+ phenotype (n = 9), and control group individuals (n = 7). (f) Spearman's correlation between the CVID patients' severity score and increase in levels of IL-21 after whole-blood stimulation with SARS-CoV-2 antigen in CVID patients (n = 9). ns-non-significant.
Changes in IL-10 levels after SARS-CoV-2 antigen stimulation were found to be significantly associated with lymphadenopathy, splenomegaly, and hepatomegaly. IL-10 levels decreased after SARS-CoV-2 antigen stimulation in patients with lymphadenopathy, whereas an increase was observed in those without the condition and in the control group. A similar trend was observed for splenomegaly and hepatomegaly, as well as for changes in IL-15 levels and splenomegaly (see Figure 5b-d and Supplementary Table S5a). When categorizing patients based on their diagnosis, a consistent observation of changes in IL-10 levels was found in the CVID group, but not in the SIgAD group, since only one patient with each of these conditions was present in the SIgAD subgroup (see Supplementary  Table S5c,d). Changes in IL-10 levels after SARS-CoV-2 antigen stimulation were found to be significantly associated with lymphadenopathy, splenomegaly, and hepatomegaly. IL-10 levels decreased after SARS-CoV-2 antigen stimulation in patients with lymphadenopathy, whereas an increase was observed in those without the condition and in the control group. A similar trend was observed for splenomegaly and hepatomegaly, as well as for changes in IL-15 levels and splenomegaly (see Figure 5b-d and Supplementary Table S5a). When categorizing patients based on their diagnosis, a consistent observation of changes in IL-10 levels was found in the CVID group, but not in the SIgAD group, since only one patient with each of these conditions was present in the SIgAD subgroup (see Supplementary Table S5c,d).
In the CVID patient subgroup, we also assessed the association with the EUROclass classification groups and found a statistically significant difference between the increase In the CVID patient subgroup, we also assessed the association with the EUROclass classification groups and found a statistically significant difference between the increase in levels of TNF-α after SARS-CoV-2 stimulation; patients with low switched memory B cell percentages (EUROclass group B+SmB-) had a higher increase in TNF-α levels compared to patients with normal switched memory B cell percentages or the control group subjects (see Figure 5e and Supplementary Table S5b). There was only one patient with B-and one patient with CD21 low high phenotype and two with transitional cell high phenotype; therefore, the median values for these patients have not been reported.
In the CVID patient subgroup, we assessed a correlation with the CVID severity score, and only spike-induced IL-21 response correlated with the severity of CVID (rs = 0.809, p = 0.009, n = 9).

Correlation between the Cytokine Levels and Lymphocyte Subsets
Correlations between changes in cytokine levels following SARS-CoV-2 antigen stimulation and immunological parameters are shown in Figure 6 and Supplementary Table S6.
Overall, we found a significant correlation between the increase in the levels of Th17 cytokines (IL-17A and IL-21) after SARS-CoV-2 antigen stimulation, as well as TGF-β with the pro-inflammatory cytokines IL-1β and TNF-α (see Supplementary Table S6b). We also observed a correlation between the baseline levels of pro-inflammatory cytokines: IL-1β correlated with the levels of IL-6 and TNF-α, and IL-6 with TNF-α and IL-10 correlated with IL-4 and IL-15 (see Supplementary Table S6a).
Regarding correlations with lymphocyte subsets, several B cell subsets were correlated with changes in cytokine levels after SARS-CoV-2 antigen stimulation: marginal zone-like B cells and IgM-only memory B cells correlated with changes in IFN-γ and IL-10 levels, while IFN-γ was also correlated with CD21low B cells. Changes in IL-4 levels also correlated with IgM-only memory B cells, CD21low B cells, and T helper cells. Changes in IL-1β levels were related to total and atypical B cells, whereas changes in IL-6 levels were related to naïve B cells. Natural killer T cells correlated with changes in IL-10, IL-17A, and IL-21 levels. Changes in IL-17A levels also correlated with total T cells and terminally differentiated T cells. Natural killer cells were correlated with changes in IL-17A and TNF-α levels (see Figure 6 and Supplementary Table S6c).

Discussion
In this study, we present the spike-specific cytokine response observed after wholeblood stimulation with the SARS-CoV-2 antigen in naturally infected unvaccinated and vaccinated PAD patients and explore the cytokine profile and other immunological parameters as markers for predicting the contraction of COVID-19 infection during the up to 10-month follow-up period.
Overall, we were not able to predict the occurrence of COVID-19 based on the antispike humoral response, baseline cytokine levels, changes in cytokine levels following whole-blood SARS-CoV-2 antigen stimulation, or other immunological parameters in patients with PAD. However, patients with higher percentages of cytotoxic T and NK cells showed a lower incidence of COVID-19 during the follow-up period. This finding is consistent with the current understanding that these specific cell types play a crucial role in the antiviral immune response, including that against SARS-CoV-2 [70]. Functional exhaustion of cytotoxic lymphocytes (such as CD8+ cytotoxic T cells and NK cells) has been associated with poor COVID-19 prognosis [33,44], whereas substantial CD8+ T cell responses have been associated with mild COVID-19 disease [71]. Cytokine levels after SARS-CoV-2 antigen stimulation did not prove to be predictive of COVID-19. To the best of our knowledge, there have been no prior studies examining the potential of cytokine levels as predictive markers of occurrence of COVID-19 in patients with PAD; however, previous research conducted on healthy individuals during an 8-month follow-up period following CoronaVac vaccination found that those who exhibited lower levels of IFN-γ in the IFN-γ release assay were at a higher risk of contracting COVID-19 [72]. Similarly, we were unable to predict the likelihood of being infected with SARS-CoV-2 through anti-spike antibody levels, which is consistent with the findings of the COV-AD study [72].
Cytokines that have previously exhibited alterations in their levels following vaccination in immunocompetent individuals include IFN-γ, IL-10, IL-15, IL-17A, IL-1 β, TNF-α, IP-10/CXCL10, IL-6, IFN-α2, IL-12p70, IL-18, IL-23, and IL-33 [73][74][75][76][77][78]. Interestingly, in addition to several of the mentioned cytokines, we also observed a decrease in the level of TGF-β1 following whole-blood stimulation with the SARS-CoV-2 antigen, and this decrease was more prominent in patients who showed a low anti-spike IgG response as well as in CVID patients with an autoimmune disease. The main source of TGF-β1 is epithelial cells, but it can be produced by most immune cells in response to infection, and it mainly acts as a regulator of multiple types of immune cells, including T regulatory (Treg) cells, NK cells, and macrophages [79]. The TGF-β1 role in antibody synthesis is mainly induced by the induction of T regulatory cells to suppress B cells [80,81] and the induction of an isotype switch to mainly IgA1 and IgA2 class antibodies [82]. In a study on the dynamics of adaptive immune response in severe COVID-19 examining the plasmablast transcriptome changes over the course of eight weeks, plasmablasts showed a continuous immune reaction; during the first week, plasmablasts showed an immune response directed against SARS-CoV-2, characterized by the synthesis of IgG antibodies against the spike and nucleocapsid proteins, but later response switched to IgA-expressing plasmablasts, which were are not specific to SARS-CoV-2 proteins and reflected continued instruction of the B lymphocytes by TGF-β1 [82]. In addition, in the upper airways of immunocompetent patients, TGF-β1 transcript level expression was lower in SARS-CoV-2-infected patients than in controls, and in asymptomatic individuals, TGF-β1 correlated negatively with IFN-γ, suggesting its role in the regulation of early antiviral inflammatory response [83]. In our cohort, changes in the levels of TGF-β1 were significantly correlated with the levels of IL-1β and TNF-α. Indeed, TGF-β1 can activate nuclear factor-kappa-light-chain-enhancer of activated B cells (NF-kB), which can further upregulate various cytokines, including IFN-γ, TNF-α, and IL-1 β [84,85]. Further research is warranted to determine the function of TGF-β1 in the adaptive immune reactions related to COVID-19 in patients with PAD.
This report confirms a higher IFN-γ response to SARS-CoV-2 antigen stimulation in vaccinated PAD patients than in unvaccinated patients who were exposed to the SARS-CoV-2 virus via natural COVID-19, similar to that previously reported in individuals in the general population [74,86], although conflicting evidence exists [87]. In our cohort, we found no difference in the changes in other cytokine levels after SARS-CoV-2 antigen stimulation when comparing vaccinated and unvaccinated patients, although previously in a study with healthy individuals, TNF-α levels in response to stimulation with peptide pools corresponding to the SARS-CoV-2 spike, nucleocapsid, or membrane protein were significantly higher in individuals who had completed a vaccination regimen than in unvaccinated individuals [74]. These differences could be because we only examined cytokine responses to S1 and S2 pools in this study.
We also examined whether the spike-specific cytokine response was associated with any specific non-infectious complications. In this study, we found an association between the magnitude of the increase in IL-4 and the presence of autoantibodies in patients' serum, as well as an increase in IL-4 correlated with total IgG and IgM levels, but not with anti-spike IgG. This finding is in line with a study of COVID-19 patients where excessive plasmablast expansion was correlated with autoantibody production, and these plasmablasts developed according to IL-4− and BAFF-driven developmental trajectories. Although they were not enriched in autoreactive B cells, two distinct memory populations (CD80+/ISG15+ and CD11c+/SOX5+/T-bet+/−) with signs of autoreactivity were identified, which were considered to be the source of COVID-19 autoantibodies [88]. Indeed, the evidence of the development of autoimmune conditions following COVID-19 has accumulated during the past few years [44,[89][90][91][92][93][94][95]. In addition, we found a correlation between IL-21 and CVID severity score in patients with CVID. Within secondary lymphoid organs, T follicular helper (Tfh) cells are primarily engaged in ensuring B cell survival, proliferation, and differentiation by producing significant quantities of IL-21 and IL-4 [96], and the majority of inborn errors of immunity patients exhibit spike-specific circulating Tfh cells [28]. Regarding CVID and SIgAD patients, association studies suggest that defective IL-4 and IL-21 signaling has been linked to an increased prevalence of non-infectious complications, including autoimmunity [4,[97][98][99][100].
IL-10 was identified as another cytokine whose increase was associated with a specific patient phenotype. IL-10 is an anti-inflammatory cytokine that is capable of inhibiting the synthesis of pro-inflammatory cytokines and is produced mainly by T cells, especially Treg and Tfh cells, as well as monocytes and B regulatory cells [7]. In our study, the extent of the spike-specific IL-10 response was significantly lower in patients with benign polyclonal lymphoproliferation, manifested as hepatomegaly, splenomegaly, and lymphadenopathy. In the EUROclass trial, splenomegaly was associated with dysregulation of B cell homeosta-sis, including reduced switched-low memory B cells, which could indicate dysregulated germinal center development [67]. We also observed that the level of the pro-inflammatory cytokine TNF-α was significantly higher in patients with low switched memory B cell counts (EUROclass B+SmB−). Although the cytokine response to SARS-CoV-2 antigen has not been examined previously in the context of non-infectious complications of PAD, the response of circulating Tfh isolated from CVID patients to submitogenic PHA+IL-2 stimulation, as indicated by the intracellular expression of Tfh cytokines (IL-4, IL-10, IL-21), with non-infectious complications (autoimmunity and/or granulomatous disease) was lower than that in patients without these complications. In contrast, the IL-4 response was higher in patients with autoimmunity and/or granulomatous disease [101]. Studies of IL-10 levels without antigen stimulation suggest that low IL-10 production in patients with CVID-like disorders with homozygous deletion of the inducible T cell co-stimulator (ICOS), a key receptor of Tfh help to germinal center B cells, may contribute to the disturbed germinal center reaction in secondary lymphoid organs, and patients with this genetic defect can be associated with splenomegaly [102]. In addition, CVID patients have been shown to have a decreased frequency of naïve regulatory T cells, one of the major sources of IL-10, in CVID patients with splenomegaly [103].
Several major limitations should be considered. First, the sample size was limited due to the rarity of these conditions, which reduced the statistical power and limited the ability to detect significant associations; therefore, the utility of predicting factors should be studied in larger cohorts. Second, the timeframe within the measurements after completion of the immunization schedule was broad in this cohort. Third, whole-blood stimulation has limitations that should be considered: T cell responses are complex, and natural or vaccine-induced T cell stimulation can activate different cell subpopulations and molecular pathways, resulting in distinct functional cytokine expression profiles and clinical outcomes. In addition, we examined only the spike-induced cytokine response, and other SARS-CoV-2 proteins, such as the nucleocapsid or membrane proteins, were not included. The cytokine IL-2 was not included in the analysis due to technical reasons. Therefore, it is possible that predicting vaccine-induced or natural protection may be more accurate if we consider the activation of T cells that express different cytokines or activation markers in addition to those currently being examined. In addition, there is a limitation in comparing our results with those of other studies, because the methods of investigation vary significantly between studies.

Conclusions
No significant association was observed between spike-specific cytokine response, anti-spike IgG levels, or other relevant factors and the likelihood of contracting COVID-19. Similar to the healthy control groups, the most distinct difference between vaccinated and naturally infected unvaccinated patients with PAD was the higher increase in IFN-γ levels in the vaccinated patients.

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/v15051146/s1. Table S1. Changes in cytokine levels following SARS-CoV-2 stimulation. Table S2. Predictors of COVID-19 during the follow-up period in PAD patients. Table S3. Associations between the cytokine levels and previous SARS-CoV-2 vaccination. Table S4. Associations between the changes in cytokine levels following SARS-CoV-2 antigen stimulation and SARS-CoV-2-specific humoral response. Table S5. Associations between the cytokine levels and patient's demographic and clinical characteristics. Table S6. Spearman correlation analysis.