Abstract
Introduction: During the COVID-19 pandemic, the effectiveness of vaccines against SARS-CoV-2 in immunodeficient patients not only did affect the individual risk of these vulnerable patients but endangered the selection of new variants of concern due to prolonged virus shedding by these patients. Methods: In a tertiary center for pulmonary diseases, we investigated the immune response of 11 patients with primary humoral immunodeficiency and 13 healthy controls on the humoral and cellular level after full vaccination with an mRNA or vector vaccine against SARS-CoV-2. Results: In the majority of patients (73%), we found antibodies against the spike protein above the threshold of positivity. Likewise, patients showed a promising cellular response: the upregulated production of INFγ, TNFα, and CXCL10 by T cells did not differ from the response of healthy controls. Conclusion: These results stress the importance to further discern an adequate immunological correlate of protection and the need to follow the effect of booster immunizations in this population at risk.
Introduction
SARS-CoV-2 is dreaded because of its capacity to trigger complex detrimental processes of the immune system leading to serious disease [1, 2]. The degree of the responsible cytokine storm seems to vary between individuals depending on specific characteristics of their immune system. Clinicians were especially concerned with immunocompromised populations at risk like patients with inborn errors of immunity (IEI).
In primary immunodeficiencies, the majority of patients lacks antibodies because of a dysfunction of B-cell lineage and consequently presents with recurrent infections of the upper or lower respiratory tract [3]. Among these, common variable immunodeficiency (CVID) is the largest, though very heterogeneous group, in which B-cell development and function might be impaired on different levels, implying T-cell dysfunction as well [4]. One-third of patients present with chronic lung disease at time of diagnosis [5] – a relevant risk factor for a serious COVID-19 disease as well [6]. In a large UK cohort of 310 patients with primary and secondary immunodeficiencies, the overall infection fatality rate was up to 17.7%, meaning a higher inpatient mortality rate at younger age than the general population [7]. In an international review based on 94 cases, an infection fatality rate of 9.6% is reported, while estimations for the general population vary significantly [8]. Patients with deficiencies of adaptive immunity such as CVID showed a milder disease which might be due to an attenuated immune response. On the other hand, case studies underline the importance to bear in mind the specific B- or T-cell deficiency of an individual patient. For example, an NF-κB2 loss of function might be associated with a serious disease by prompting an increase of IL-6 [9, 10].
Multiple studies underlie the importance of cellular immunity for effective protection in COVID-19, especially for its severity [11]. For infection control, a model of the timely orchestrated reaction of the three branches of adaptive immunity – CD4+ T cells, CD8+ T cells, and antibodies by B cells – is favored [11]. For mRNA vaccination, Liang et al. [12] observed a strong upregulation of INF-inducible genes such as CXCL10 in macaques. CD4+ memory T cells – along with memory B cells – seem to be robust over time [13]. The observation that patients with agammaglobulinemia, implying the absence of B cells and antibodies, might overcome the disease [9, 14] further stresses the importance of T-cell response [11]. Moreover, cross-reactive CD4+ memory T cells in immunocompetent individuals [15] and in CVID patients as well [16] are assumed to be pivotal. For these reasons, effects of cellular immunity were investigated and shown for both licensed mRNA vaccines, BNT162b2 und mRNA-1273 [17, 18], and the vector-based vaccine ChAdOx1 nCoV-19 [19].
In immunocompromised patients due to immunosuppressive treatments, reported antibody response rates vary between 10% in patients after heart (and/or lung) transplantation [20], 35.3% after organ transplantation (predominantly kidney) [21], and 50% under anti-CD20-therapy in multiple sclerosis [22]. The interpretation of these rates is complicated by differing strength of immunosuppressant regimes and the lack of clarity about the immunological correlate of clinical protection. For humoral response in IEI, there is a high variability of rates of seroconversion: Hagin et al. [23] found a strong production of anti-SARS-CoV-2-S-antibodies (81.8%) after full vaccination with BNT162b2 [23] in 26 patients; in a retrospective study, 90% of 11 patients with immune deficiency developed a positive titer [24] and Delmonte et al. [25] report a similar high percentage of humoral response in 63 of 74 (85.4%) IEI patients after the second dose. In an Italian longitudinal study, 21 patients with IEI developed an antibody response that was comparable to that of healthy controls (HC) 21 days after the first mRNA vaccine dose, but smaller 7 days after the second dose [26]. Ponsford et al. [27] reported a high seroconversion of IgG in their cohort of immunodeficient patients. Among these, 70% of 60 patients with CVID showed a positive, though differently high antibody titer after two doses of COVID-19 vaccination. By contrast, Salinas et al. [28] reported a response of only 20.6% in 34 CVID patients.
With regard to cellular immunity, Schmidt et al. [21] reported a significantly lower percentage of INFγ-producing T cells in organ transplant patients than in the control group and Schramm et al. [20] observed a lower concentration of INFγ released from T cells of transplant patients. Patients with multiple sclerosis under anti-CD20-therapy generated robust antigen-specific CD4 and CD8 T-cell responses [22]. In IEI, Hagin et al. [23] reported a positive response in an ELISpot for IL-2 and INFγ after stimulation with the spike protein in 73.1% of patients, whereby they did not differ from vaccinated HC. Amodio et al. [26] report a significant expansion of spike-specific CD4 T cells in patients and controls, while absolute levels at baseline and after the second vaccination seem to be higher in patients. Awuah et al. [29] found a significantly lower T-cell proliferation after vaccination in CVID patients than in HC. Remarkably, patients with X-linked agammaglobulinemia (XLA) had a robust T-cell response while producing no antibodies. In line, Salinas et al. [28] reported that spike-specific INFγ production by T cells did not increase in CVID patients but in HC and XLA patients. To further study the vaccine efficacy, we characterized the humoral and cellular immunity after SARS-CoV-2 vaccination in patients with primary humoral immunodeficiency from the perspective of a tertiary center for pulmonary diseases.
Methods
Study Design and Patient Population
Participants – patients with humoral immunodeficiency (IDP) and HC – were recruited via the social media platform Facebook and flyer and included after giving their informed consent from May until August 2021 (Table 2 for descriptive data). They were vaccinated according to the prioritization by the federal state of Germany independently of the enrollment of this study. Either they received a homologous or a heterogeneous sequence of vaccination type (vector-based vaccine ChAdOx1 nCoV-19 and/or mRNA vaccine BNT162b2/mRNA-1273) along their individual classification and choice. Except from longitudinal data in four participants, we measured the immune response 2 weeks (12–16 days) after the second dose of vaccination. All participants answered a questionnaire about their tendency to infections and reactions to the two vaccine doses. No subject was excluded from the study after giving their informed consent.
Routine Laboratory Testing
As an indicator of immune impairment, we measured levels of immunoglobulin A, M, G (plus subclasses), and E in all participants. For the four subjects longitudinally studied, we additionally performed flow cytometry (FACS) 1 week after the first dose. Antigen tests, which all remained negative, were done at the clinic entrance as a routine procedure from May up to end of June 2021 and terminated for administrative reasons.
Humoral Response
We looked for anti-nucleocapsid-peptide antibodies, indicating immunity after infection, by the ELECSYS anti-SARS-CoV-2 electrochemiluminescence-immunoassay by Roche Diagnostics (Mannheim, Germany) that offers a dichotomous result (positive speaking of infection). For detection of immunity after vaccination by anti-spike-peptide antibodies, more specifically against the S1 domain that contains the receptor-binding domain, we employed the anti-SARS-CoV-2 QuantiVac enzyme-linked immunosorbent assay (ELISA) by Euroimmun (Lübeck, Germany). The result is indicated in BAU/mL (binding antibody units) and defines a cut-off ≥35.2 BAU/mL as positive. Antibody titers <25.6 BAU/mL are defined as negative, those between 25.6 and 35.2 BAU/mL as borderline. In our laboratory, distinct values are obtained by dilution up to 384 BAU/mL. According to the manufacturer’s information, qualitative correspondence to neutralization tests was >95%.
Cellular Response: Gene Expression Analysis
We measured the production of RNA for INFγ, TNFα, and CXCL10 after stimulation with the spike protein and compared it to the respective production of RNA for these cytokines/chemokines after stimulation with a positive and a negative control. A 0.5-mL of heparinized whole blood was incubated in each of the tubes (Blank, Stim, and Cov-2) of the SARS-CoV-2 IGRA set (Euroimmun, Lübeck Germany) for 18 h at 37°C and 7% CO2. Red cells were lysed with ACK buffer 2 times. Afterward, the leukocytes were lysed with 400 μL of MagNA Pure lysis buffer (Roche) containing 1% 1,4-dithiothreitol (Roche Cat. #: 10 708 984 001) and the samples were frozen at −70°C. After thawing, the lysates were well mixed and transferred into the MagNA Pure sample cartridge and total RNA was isolated with the MagNA Pure LC device using the RNA standard protocol for cells. The elution volume was set to 50 µL. An aliquot of 8.2 μL mRNA was reversely transcribed using AMV-RT and oligo-(dT) as primer (First Strand cDNA Synthesis Kit, Roche) according to the manufacture protocol in a thermocycler. After termination of the cDNA synthesis, the reaction mix was diluted to a final volume of 500 μL and stored at −20°C until PCR analysis.
Primer sets optimized for the LightCycler® (RAS, Mannheim, Germany) were developed and purchased from SEARCH-LC GmbH (www.Search-LC.com). The PCR was performed with the LightCycler® FastStart DNA Sybr Green kit (RAS) according to the protocol provided in the parameter-specific kits. To control for specificity of the amplification products, a melting curve analysis was performed. The copy number was calculated from a standard curve, obtained by plotting known input concentrations of four different plasmids at log dilutions to the PCR cycle number (CP) at which the detected fluorescence intensity reaches a fixed value. To correct for differences in the content of RNA, the calculated transcript numbers were normalized according to the expression of the housekeeping gene peptidylprolyl isomerase B (PPIB). Values were thus given as transcripts per 1,000 transcripts of PPIB.
Statistical Analysis
For cellular responses, we performed a one-way MANOVA comparing IDP and HC on the three dimensions INFγ, TNFα, and CXCL10 as they would, already from a theoretical stance, correlate with each other (dependent measurements). Subjective impression of reactions to vaccine was analyzed by Mann-Whitney U test. A p value <0.05 was considered statistically significant. Statistical analyses were carried out by SPSS Statistics (Version 27).
Results
Patient Characteristics
Eleven IDP were included in this study (Table 1). Two patients could be monitored during their whole vaccination process, while the others were only seen 2 weeks after their second dose. Diagnosis was based on patients’ clinical history and laboratory parameters. In comparison with 13 HC, IDP showed more frequent infections per year leading to use of antibiotics (Table 2). Thanks to established immunoglobulin therapy, global levels of IgG and all but one subclass were within the normal range. One-third of IDP and 15% of HC received a heterologous vaccination scheme.
Patient characteristics
Patient No. . | Diagnosis . | Sex . | Age . | Manifestation/other relevant lung disease . | Frequency of respiratory infections per year . | Frequency of consequent use of antibiotics . | Medication (appl.) . | Dose, g . | Interval, weeks . |
---|---|---|---|---|---|---|---|---|---|
1 | Humoral B-cell defect, whole IgG, IgG 1, and 4 low | f | 41 | Recurrent infections, bronchiectasis | 4.5 | 2.5 | Privigen (i.v.) | 15 | 4 |
2 | Humoral immunodeficiency of unknown cause | f | 59 | Recurrent pneumonia | 0 | 0 | Octagam (i.v.) | 30 | 4 |
3 | CVID, suspected memory B-cell defect | f | 23 | Not known | 2 | 0.5 | Hizentra (s.c.) | 10 | 1 |
4 | CVID, suspected memory B-cell defect | m | 23 | Discrete bronchiectasis | 0 | 0 | Hizentra (s.c.) | 8 | 1 |
5 | Humoral immunodeficiency, probable CVID | m | 71 | Recurrent infections, discrete splenomegalie | 0 | 0 | Octagam (i.v.) | 20 | 4 |
6 | Hypogammaglobulinemia with deficiency of IgG, whole and subclasses IgG 2 and 4 | m | 38 | Granulomatous lymphocytic interstitial lung disease | 1 | 0 | Privigen (i.v.) | 20 | 4 |
7 | CVID | m | 59 | Bronchiectasis | 2 | 1 | Gamunex (i.v.) | 35 | 4 |
8 | IgG subclass dysregulation with reduced IgG 2 | f | 72 | Bronchiectasis | 0 | 0 | Gamunex (i.v.) | 15 | 4 |
9 | Humoral immunodeficiency, IgG 2 subclass deficiency | f | 65 | Pleural effusion, tree-in-bud sign | 10 | 3 | Octagam (i.v.) | 20 | 4 |
10 | Humoral immunodeficiency, IgG deficiency and low IgG 1 and IgG 3 | f | 69 | Recurrent infections, history of tuberculosis | 0 | 0 | Cutaquig (s.c.) | 4.95 | 1 |
11 | Good syndrome with hypogammaglobulinemia and thymoma | m | 49 | Recurrent infections | 2.5 | 1.5 | Octagam (i.v.) | 30 | 2 |
Patient No. . | Diagnosis . | Sex . | Age . | Manifestation/other relevant lung disease . | Frequency of respiratory infections per year . | Frequency of consequent use of antibiotics . | Medication (appl.) . | Dose, g . | Interval, weeks . |
---|---|---|---|---|---|---|---|---|---|
1 | Humoral B-cell defect, whole IgG, IgG 1, and 4 low | f | 41 | Recurrent infections, bronchiectasis | 4.5 | 2.5 | Privigen (i.v.) | 15 | 4 |
2 | Humoral immunodeficiency of unknown cause | f | 59 | Recurrent pneumonia | 0 | 0 | Octagam (i.v.) | 30 | 4 |
3 | CVID, suspected memory B-cell defect | f | 23 | Not known | 2 | 0.5 | Hizentra (s.c.) | 10 | 1 |
4 | CVID, suspected memory B-cell defect | m | 23 | Discrete bronchiectasis | 0 | 0 | Hizentra (s.c.) | 8 | 1 |
5 | Humoral immunodeficiency, probable CVID | m | 71 | Recurrent infections, discrete splenomegalie | 0 | 0 | Octagam (i.v.) | 20 | 4 |
6 | Hypogammaglobulinemia with deficiency of IgG, whole and subclasses IgG 2 and 4 | m | 38 | Granulomatous lymphocytic interstitial lung disease | 1 | 0 | Privigen (i.v.) | 20 | 4 |
7 | CVID | m | 59 | Bronchiectasis | 2 | 1 | Gamunex (i.v.) | 35 | 4 |
8 | IgG subclass dysregulation with reduced IgG 2 | f | 72 | Bronchiectasis | 0 | 0 | Gamunex (i.v.) | 15 | 4 |
9 | Humoral immunodeficiency, IgG 2 subclass deficiency | f | 65 | Pleural effusion, tree-in-bud sign | 10 | 3 | Octagam (i.v.) | 20 | 4 |
10 | Humoral immunodeficiency, IgG deficiency and low IgG 1 and IgG 3 | f | 69 | Recurrent infections, history of tuberculosis | 0 | 0 | Cutaquig (s.c.) | 4.95 | 1 |
11 | Good syndrome with hypogammaglobulinemia and thymoma | m | 49 | Recurrent infections | 2.5 | 1.5 | Octagam (i.v.) | 30 | 2 |
Comparison of IDP and HC for demographics and immunological features
. | IDP . | HC . | Normal range . |
---|---|---|---|
N (after second vaccination) | 11 | 13 | |
Age: M (SD) | 51.73 (18.23) | 48.46 (14.76) | |
Sex: f/m | 6/5 | 6/7 | |
Respiratory infections per year: M (SD), n | 2 (3.02) | 0.67 (0.69)a | |
Consequent use of antibiotics: M (SD), n | 0.77 (1.1) | 0.25 (0.58)a | |
Type of received vaccine (B, M, A, AB, AM), n | 8 (B), 2 (AM), 1 (AB) | 8 (B), 3 (A), 2 (AB) | |
Global IgG 2 weeks after second dose: M (SD) | 8.2 (2.42) g/L | 11 (2.44) g/L | 7–16 g/L |
Subclass IgG 1: M (SD) | 5.52 (1.8) g/L | 6.68 (2.18) g/L | 2.8–8 g/L |
Subclass IgG 2: M (SD) | 2.54 (0.81) g/L | 3.68 (1.66) g/L | 1.15–5.7 g/L |
Subclass IgG 3: M (SD) | 0.21 (0.18) g/L | 0.34 (0.1) g/L | 0.24–1.25 g/L |
Subclass IgG 4: M (SD) | 0.14 (0.09) g/L | 0.87 (0.79) g/L | 0.05–1.25 g/L |
IgA 2 weeks after second dose: M (SD) | 1 (1.06) g/L | 2.3 (0.81) g/L | 0.7–4 g/L |
IgM 2 weeks after second dose: M (SD) | 1.03 (1.71) g/L | 1.07 (0.75) g/L | 0.4–2.3 g/L |
IgE 2 weeks after second dose: M (SD) | 13.91 (21.73) U/mL | 94.69 (139.27) U/mL | <100 U/mL |
. | IDP . | HC . | Normal range . |
---|---|---|---|
N (after second vaccination) | 11 | 13 | |
Age: M (SD) | 51.73 (18.23) | 48.46 (14.76) | |
Sex: f/m | 6/5 | 6/7 | |
Respiratory infections per year: M (SD), n | 2 (3.02) | 0.67 (0.69)a | |
Consequent use of antibiotics: M (SD), n | 0.77 (1.1) | 0.25 (0.58)a | |
Type of received vaccine (B, M, A, AB, AM), n | 8 (B), 2 (AM), 1 (AB) | 8 (B), 3 (A), 2 (AB) | |
Global IgG 2 weeks after second dose: M (SD) | 8.2 (2.42) g/L | 11 (2.44) g/L | 7–16 g/L |
Subclass IgG 1: M (SD) | 5.52 (1.8) g/L | 6.68 (2.18) g/L | 2.8–8 g/L |
Subclass IgG 2: M (SD) | 2.54 (0.81) g/L | 3.68 (1.66) g/L | 1.15–5.7 g/L |
Subclass IgG 3: M (SD) | 0.21 (0.18) g/L | 0.34 (0.1) g/L | 0.24–1.25 g/L |
Subclass IgG 4: M (SD) | 0.14 (0.09) g/L | 0.87 (0.79) g/L | 0.05–1.25 g/L |
IgA 2 weeks after second dose: M (SD) | 1 (1.06) g/L | 2.3 (0.81) g/L | 0.7–4 g/L |
IgM 2 weeks after second dose: M (SD) | 1.03 (1.71) g/L | 1.07 (0.75) g/L | 0.4–2.3 g/L |
IgE 2 weeks after second dose: M (SD) | 13.91 (21.73) U/mL | 94.69 (139.27) U/mL | <100 U/mL |
B, BioNTech; M, Moderna; A, AstraZeneca; AB, heterologous (1. AstraZeneca, 2. BioNTech); AM, heterologous (1. AstraZeneca, 2. Moderna).
aOne missing value.
Safety
IDP and HC reported similar frequency of combined local and systemic adverse events (Fig. 1). Concerning subjective severity, both groups rated the reactions to the first dose as “mild” (seven missing values). Whereas IDP perceived those after the second dose as “moderate” (three missing values), HC (five missing values) described still “mild” reactions. However, impressions did not differ significantly between groups (U = 41, Z = −0.841, p = 0.40).
Relative frequencies of self-reported type of reaction after first and second vaccine dose in IDP and HC, local: pain/swelling at injection site, lymphadenopathy, skin reaction; systemic: fatigue, muscle pain, headache, fever; IDP with one missing value.
Relative frequencies of self-reported type of reaction after first and second vaccine dose in IDP and HC, local: pain/swelling at injection site, lymphadenopathy, skin reaction; systemic: fatigue, muscle pain, headache, fever; IDP with one missing value.
Humoral Response
None of IDP or HC showed presence of anti-N antibodies (Roche ECLIA) during data collection. As for anti-S-antibodies (Euroimmun ELISA), all HC had a positive response with an exhaustive titer >384 BAU/mL. Among IDP, 8/11 (73%) produced anti-S-IgG and 5 patients had a titer >384 BAU/mL. Of note, one of them was tested for anti-S-antibodies 43 days after his second dose. Three IDP had a positive titer <384 BAU/mL. As well, due to organizational delays, we performed the Euroimmun ELISA for two of them 40 and 91 days after their respective second dose.
Cellular Response
We measured the expression of INFγ, TNFα, and CXCL10 genes in whole blood after stimulation with the spike protein and compared it to the respective expression of transcripts of these cytokine/chemokine genes after stimulation with a positive and a negative control. As the positive control is a polyclonal stimulation, the response to it might be stronger than that to the spike protein. Since individual absolute values diverged a lot, we calculated a quotient comparing normalized transcripts in response to the spike protein and transcripts in the positive control for each participant and cytokine/chemokine. For technical reasons, we did not get results for two IDP. Extreme values, defined as exceeding three times the interquartile range [30], are depicted in Figure 2. All calculated quotients were positive and above 1, meaning more transcripts were produced in response to spike than transcripts in response to positive control. Finally, one-way MANOVA showed no significant difference between IDP and HC on the combined variables CXCL10, TNFα, and INFγ (F(3, 18) = 0.94, p = 0.44, partial η2 = 0.14, Wilk’s Λ = 0.86). IDP seemed to have slightly higher responses to spike than HC (Table 3). For control, we repeated the MANOVA without the 4 cases with extreme values (see above), which did not change this result (F(3, 14) = 0.18, p = 0.91, partial η2 = 0.04, Wilk’s Λ = 0.96), though the direction of means changed. As well, there were no significant group differences when we built other quotients for CXCL10, TNFα, and INFγ with the following structure: (normalized transcripts to spike – normalized transcripts to negative control)/normalized transcripts to positive control (F(3, 17) = 0.95, p = 0.44, partial η2 = 0.14, Wilk’s Λ = 0.86).
Box plots for relative transcripts of CXCL10, TNFα, and INFγ in reaction to spike protein/positive control in IDP and HC, extreme values (>3× IQR) are named, outliers (>1.5× IQR) only marked by a point.
Box plots for relative transcripts of CXCL10, TNFα, and INFγ in reaction to spike protein/positive control in IDP and HC, extreme values (>3× IQR) are named, outliers (>1.5× IQR) only marked by a point.
Descriptive statistics for different comparisons of cellular response between IDP and HC
Relative response of cytokine/chemokine . | IDP: M (SD) . | HC: M (SD) . |
---|---|---|
CXCL10 (spike/positive) | 1.94 (3.49) | 1.05 (1.01) |
CXCL10 (spike/positive) without cases with extreme values | 0.72 (0.66) | 1.06 (1.02) |
CXCL10 (spike-negative/positive) | 1.94 (3.49) | 1.03 (1.05) |
TNFα (spike/positive) | 1.97 (3.91) | 0.81 (0.59) |
TNFα (spike/positive) without cases with extreme values | 0.64 (0.30) | 0.67 (0.35) |
TNFα (spike-negative/positive) | 1.66 (3.44) | 0.59 (0.61) |
INFγ (spike/positive) | 0.58 (0.64) | 0.51 (0.60) |
INFγ (spike/positive) without cases with extreme values | 0.29 (0.25) | 0.36 (0.40) |
INFγ (spike-negative/positive) | 0.53 (0.64) | 0.49 (0.63) |
Relative response of cytokine/chemokine . | IDP: M (SD) . | HC: M (SD) . |
---|---|---|
CXCL10 (spike/positive) | 1.94 (3.49) | 1.05 (1.01) |
CXCL10 (spike/positive) without cases with extreme values | 0.72 (0.66) | 1.06 (1.02) |
CXCL10 (spike-negative/positive) | 1.94 (3.49) | 1.03 (1.05) |
TNFα (spike/positive) | 1.97 (3.91) | 0.81 (0.59) |
TNFα (spike/positive) without cases with extreme values | 0.64 (0.30) | 0.67 (0.35) |
TNFα (spike-negative/positive) | 1.66 (3.44) | 0.59 (0.61) |
INFγ (spike/positive) | 0.58 (0.64) | 0.51 (0.60) |
INFγ (spike/positive) without cases with extreme values | 0.29 (0.25) | 0.36 (0.40) |
INFγ (spike-negative/positive) | 0.53 (0.64) | 0.49 (0.63) |
Longitudinal Data
Fortunately, IDP No. 5 and 7 as well as HC No. 9 and 12 were followed throughout the process of vaccination. Levels of IgG (global and subclasses 1–4), IgA, IgM, and IgE remained stable within subjects from first to last visit. In flow cytometry, we detected 3 plasmablasts/μL in both HC whereas IDP showed 0. As to cellular response, the relative response was recognizable only after the second dose for most subjects and parameters (Fig. 3a–c). It seemed to be more pronounced for TNFα and for INFγ than for CXCL10.
a–c Relative cellular response for IDP No. 5 and 7 (missing for time points 2 and 3) as for HC No. 9 and 12 throughout the vaccination process.
a–c Relative cellular response for IDP No. 5 and 7 (missing for time points 2 and 3) as for HC No. 9 and 12 throughout the vaccination process.
Discussion
We studied humoral and cellular vaccination responses after SARS-CoV-2 vaccination in patients with primary humoral immunodeficiency. Remarkably, the majority (73%) of IDP had a positive anti-S-antibody response 2 weeks after full vaccination. For cellular immunity, we found no difference between IDP and HC in the expression of INFγ, TNFα, and CXCL10 genes.
The percentage of IDP that had a positive humoral response in our cohort is clearly above that found for transplanted patients under immunosuppressant medication [20, 21] and in line with most of the reported rates in similar samples [23‒25, 27]. Considering the ongoing debate about a correlate of sufficient clinical protection, interpretation should be careful. For once, there is conflicting evidence as to the necessity of neutralizing antibodies [31, 32]. Furthermore, there is a growing agreement about the importance of memory B cells, whose frequencies tend to increase after infection while neutralizing antibodies decay [33]. This process might be impaired in CVID patients [34, 35]. Likewise, the decay kinetics and the immunological meaning of other binding antibodies are still to be investigated [36].
As for cellular immunity comparison to other studies, investigating cellular response is complicated by the fact that employed methods differ a lot. Nonetheless, our data support the promising results seen in the Israelian [23] and Italian [26] cohorts. Once again, the definition of a positive, clinically protective response is still a matter of debate. As Zonozi et al. [37] observed a remarkable risk reduction of severe disease by vaccination even without an antibody response, also the clinical relevance of cellular immunity is demonstrated. Furthermore, the fact that cellular responses seemed even more pronounced in IDP than in HC and the finding of extreme values fits well with results of preceding studies. Hagin et al. [23] reported stronger cellular responses in XLA, and Amodio et al. [26] found higher absolute levels of cellular response for patients than for controls. More recently, Zonozi et al. [37] found a more proliferative CD8+ T-cell response in CVID patients than in HC. Also in patients under anti-CD-20 therapy, there is evidence of a stronger CD8+ T-cell response [22, 37]. This might be interpreted as a compensation of B-cell deficiency but could, as well, be seen in the context of immunological dysregulation in CVID that is supposed to drive noninfectious morbidity such as autoimmunity [38].
Of note, our study has some considerable limitations. The small sample size impairs statistical power and substantially diminishes the generalizability of the results. The representativeness of our small cohort of patients is limited by their heterogeneous individual defects. Due to the small sample size, building subcategories for more detailed examination would have been desirable but was not possible in a statistically comprehensive manner. Of note, two participants only had a subclass deficiency implying a milder disease. In a larger sample, it would be interesting to further differentiate the immunological reaction depending on the underlying defect. In addition, the high IgE levels in HC point to a possible subclinical allergic predisposition that might have influenced their immune response. Conflicting comparability, three IDP were tested for antibodies with substantial delay which urges prudent interpretation. Concerning methods, we did not perform neutralization assays. But, there is convincing evidence that implies that the anti-SARS-CoV-2-QuantiVac ELISA by Euroimmun we employed predicts neutralization capacity and could therefore guide the vaccination process with lower costs [39, 40]. Of note, in large cohort studies, spike-binding – not neutralizing – IgG antibodies after infection were associated with protection from reinfection [41] and IgG concentrations higher than 500 BAU/mL were associated with clinical protection [42]. Unfortunately, at the time of our study our laboratory did not further dilute titers >384 BAU/mL. Data in the manufacturer’s information on correlation to neutralization are limited to a probably immunocompetent, partly convalescent sample and might not be transferable to immunocompromised patients. In solid transplant recipients, however, anti-S-antibodies measured by Euroimmun ELISA and neutralization activity were similar in magnitude [21]. Furthermore, it would have been interesting to compare different vaccination schemes which we could not provide because of small case numbers.
After the pandemic and in light of future risks, the long-term follow-up remains a relevant clinical question on an individual scale. In this context, it would be of special interest to measure humoral and cellular responses for a longer period, to correlate the data to clinical protection and incorporate the effect of further vaccinations. In this context, extending the dose interval is supposed to improve humoral immunity through better affinity maturation [43], which should be investigated in newly vaccinated immunodeficient patients. As well, as prolonged viral shedding and intrahost evolution in immunocompromised patients seem to play a pivotal role in emergence of variants, following the effect of further vaccine doses in primary immunodeficiencies is critical on a global health scale [44, 45]. Intriguingly, a Korean case study revealed 18 mutations in one B-cell-depleted individual over 68 days of infection [46]. Still for the omicron period, a recent multicentered, prospective analysis revealed B-cell dysfunction or depletion as a major risk factor for longer infection, which favorizes mutations of the spike protein and asks for monitoring for antiviral resistance in patients at risk [47]. To integrate studies on cellular immunity, a more standardized, clinically feasible measuring of T-cell response and its correlation to protection would be helpful.
Acknowledgments
The authors thank all study participants.
Statement of Ethics
This study protocol was approved by the Ethics Committee of the Medical Faculty of the University of Heidelberg (reference S-243/2021). Due to the fast vaccination process, an amendment of the original protocol was necessary and approved. In addition, it was notified to the Paul-Ehrlich-Institute as a non-interventional study (reference NIS 578) with protocol and report available online (https://www.pei.de/SharedDocs/awb/nis-0501-0600/0578.html). Written informed consent was obtained from all participants.
Conflict of Interest Statement
Sophie Rosendahl, Felix J.F. Herth, Michael Kreuter, and Thomas Giese have no conflicts of interest to declare. Franziska C. Trudzinski states that she has received payment or honoraria for lectures, presentations, speaker bureaus, manuscript writing, or educational events from GlaxoSmithKline, Chiesi, Boehringer Ingelheim GmbH, Grifols, CSL Behring, Streamed Up, RG Gesellschaft für Information and Organisation mbH, and AstraZeneca. She also states participation on a Data Safety Monitoring Board or Advisory Board of GlaxoSmithKline and Chiesi as well as support for attending meetings and/or travel by Grifols and CSL Behring. She receives grants from Knorr Stiftung. Markus Polke states that he received lecture fees from AstraZeneca, Boehringer Ingelheim, and Novartis.
Funding Sources
This study was not supported by any sponsor or funder.
Author Contributions
S. Rosendahl – investigation, formal analysis, data curation, and writing; T. Giese – investigation, methodology, formal analysis, data curation, and writing; F.C. Trudzinski – investigation, methodology, supervision, and editing; M. Kreuter – methodology, supervision, and editing; M. Polke, F.J.F. Herth – review and editing.
Additional Information
Notified to Paul-Ehrlich-Institute as a non-interventional study (reference NIS 578).
Data Availability Statement
Data are not publicly available due to ethical reasons. Further inquiries can be directed to the corresponding author.