SARS-CoV-2 spike-directed T cells and anti-receptor-binding-domain antibodies in hematological patients receiving three mRNA COVID-19 vaccine doses

Abstract

Introduction

Studies assessing SARS-CoV-2 antibodies and functional T-cell responses in hematological patients following receipt of a three-dose mRNA vaccine schedule have yielded conflicting results.

Methods

We conducted a retrospective observational study in which we measured SARS-CoV-2-S-directed cytokine-producing CD8⁺ and CD4⁺ T cells by flow cytometry and anti-receptor-binding domain (RBD) total antibodies in 32 hematological patients, including 10 allogeneic hematopoietic stem cell transplant (allo-HCT) recipients and 22 non-transplanted individuals. Immunological testing was performed after the second COVID-19 mRNA vaccine dose (Post-2D), prior to the third dose (Pre-3D), and after the third dose (Post-3D).

Results

At Post-2D, 72.4%, 44.8%, and 37.9% of participants had detectable IFN-γ-, TNF-α-, and IFN-γ/TNF-α-producing CD8⁺ T cells, respectively. A decrease in the rate of detectable and frequencies of monofunctional and bifunctional SARS-CoV-2-S CD8⁺ T-cell responses was observed at Pre-3D. By contrast, the number of participants displaying detectable cytokine-producing CD4⁺ T-cell responses and their frequencies were rather conserved at all testing times. Receiving a 3D had no significant impact on the rate of detection and frequencies of monofunctional and bifunctional SARS-CoV-2-S-directed T-cell subsets. Anti-RBD total antibody responses were detected in approximately half of participants at all time points. Overall, anti-RBD antibody levels decreased at Pre-3D and then increased following the 3D.

Conclusion

Administration of the 3D may have a limited impact on both cellular and humoral immune responses in a substantial proportion of hematological patients. These findings highlight the heterogeneity of vaccine-induced immunity and support the role of combined immune monitoring and alternative protective strategies for selected patients.

Resumen

Introducción

Los estudios que han evaluado las respuestas de anticuerpos y células T funcionales del SARS-CoV-2 en pacientes hematológicos tras recibir tres dosis de la vacuna ARNm han mostrado resultados contradictorios.

Métodos

Se realizó un estudio retrospectivo y observacional en el que se cuantificaron los linfocitos T CD8+ y CD4+ productores de citocinas dirigidos frente a SARS-CoV-2-S y los anticuerpos totales contra el dominio de unión al receptor (RBD) en 32 pacientes hematológicos vacunados contra la COVID-19. Las pruebas inmunológicas se realizaron tras la segunda dosis vacunal (Post-2D), antes de la tercera (Pre-3D) y después de la tercera dosis (Post-3D).

Resultados

En Post-2D, el 72,4%, el 44,8% y el 37,9% de los participantes presentaron células T CD8+ productoras de IFN-γ, TNF-α e IFN-γ/TNF-α detectables. Se observó una disminución en las tasas y frecuencias de células T CD8+ en Pre-3D. Al contrario, las respuestas de células T CD4+ productoras de citocinas y sus frecuencias se conservaron ampliamente en todos los momentos de la prueba. La administración de la 3D no modificó significativamente la detección y frecuencias de células T dirigidas a SARS-CoV-2-S. Se detectaron respuestas de anticuerpos anti-RBD en aproximadamente la mitad de los participantes en todos los puntos temporales. En general, los niveles de anticuerpos anti-RBD disminuyeron en Pre-3D y aumentaron en Post-3D.

Conclusión

La 3D parece tener un impacto limitado en las respuestas inmunitarias celular y humoral en pacientes hematológicos. Estos hallazgos resaltan la importancia de la monitorización inmunitaria combinada y de estrategias de protección alternativas para determinados pacientes.

Introduction

Hematological patients, including those with myeloid or lymphoid neoplasms—subjected or not to active treatment—or undergoing allogeneic hematopoietic stem cell transplantation (allo-HCT) or chimeric antigen receptor T (CAR T)-cell therapy are at higher risk of severe SARS-CoV-2 infection-associated complications, including hospitalization in the intensive care unit and death.^1–4 Due to the impairment of adaptive immunity mechanisms associated with these clinical conditions, COVID-19 vaccination might be less efficacious in these patient populations, compared with healthy individuals, who remain at risk of primary SARS-CoV-2 infection or reinfection by emerging (sub)variants.^5,6 In effect, a number of studies^7–19 have shown suboptimal adaptive immune responses in hematological patients following completion of a two-dose COVID-19 vaccination schedule using Spike(S) mRNA-based platforms, with variable rates of SARS-CoV-2-S-directed antibody and functional T-cell responses depending upon the nature of the underlying immunosuppressive condition. As a result, contemporary consensus guidelines recommend that hematological patients should receive a three-dose program of mRNA vaccines or a two-dose program with protein subunit vaccine, starting preferably before treatment of the underlying disease or allo-HCT, or during the maintenance or off-therapy phase.²⁰ Studies assessing SARS-CoV-2 antibody and functional T-cell responses in parallel following receipt of a three-dose mRNA vaccine schedule yielded rather heterogeneous findings as to their immunogenicity, with some showing robust and coordinated humoral and cellular immunity, while others observed weaker or discordant responses, both from qualitative and quantitative standpoints.^21–37 These inconsistencies highlight the need to better define the determinants of vaccine-induced immunogenicity. We hypothesized that a third COVID-19 vaccine dose would enhance SARS-CoV-2-specific immunity, leading to increased and more sustained humoral and cellular immune responses compared with a two-dose regimen. To further shed light on this issue, we aimed to characterize the kinetics of SARS-CoV-2-S cytokine-producing T cells and anti-receptor-binding domain (RBD) total antibodies following receipt of two and three COVID-19 vaccine doses. Results reported herein may be informative for establishing future COVID-19 vaccination strategies in this particularly vulnerable population.

Patients and methods

Study population and specimen collection

The current single-center, retrospective, longitudinal and observational study was conducted at Hospital Clínico Universitario de Valencia (HCUV) and included 32 non-consecutive adult (≥18 years) hematological patients (Table 1). Patients were selected based on the availability of at least two longitudinal blood samples obtained at predefined time points as specified below. These samples were obtained from hematological patients routinely followed at our center for SARS-CoV-2 immune monitoring. Given the retrospective nature of the study and the limited availability of paired longitudinal samples, this study should be considered exploratory. All patients had completed a regular homologous vaccination schedule (two doses) with either SARS-CoV-2 Wuhan-Hu-1-based mRNA vaccines (Comirnaty®, n = 19; Spikevax®, n = 12) or recombinant ChAdOx1-S (Vaxzebria®, n = 1). Most participants (n = 28) received a homologous third vaccine dose. Third doses were administered between October 2021 and December 2022. The median of time between the first and the second dose of the vaccine was 23 days, and between the second and third vaccine doses was 159 days (range 87–351). The median of time for collecting samples after the second vaccine dose was 21 days (post-2D), and the median of time for collecting samples after the third dose was 31 days (post-3D). In the middle of the period between the second and the third dose new samples were taken (pre-3D). The median of time between pre-3D and the third dose of the vaccine was 89 days. A total of 80 whole blood specimens were tested. Data on previous SARS-CoV-2 infection were retrieved from electronic clinical records and laboratory databases at our institution. The study was approved (2022/351) by the Institutional (INCLIVA) Ethics Committee. All subjects agreed to voluntarily participate in the study and gave written consent.

Table 1.

Demographic and clinical characteristics of participants.

Male sex, no. (%)	19 (59.4)
Median age in years (IQR)	64 (56–75)
Underlying hematological disease, no. (%)
Multiple myeloma	7 (21.9)
Chronic lymphocytic leukemia	5 (15.6)
Acute myeloid leukemia	5 (15.6)
Follicular lymphoma	5 (15.6)
Hodgkin's lymphoma	2 (6.3)
Diffuse large B-cell lymphoma	2 (6.3)
Non-Hodgkin's lymphoma	2 (6.3)
Myelofibrosis	1 (3.1)
Myelodysplastic syndrome	1 (3.1)
Sezary syndrome	1 (3.1)
Mantle cell lymphoma	1 (3.1)
Treatment status, no. (%)
No current treatment	4 (12.5)
Active treatment	28 (87.5)
Treatment categories, no. (%)
Allogenic hematopoietic stem cell transplant	10 (31.2)
Autologous stem cell transplant	3 (9.4)
CD20 antibody therapy (rituximab, obinutuzumab)	5 (15.6)
CD38 antibody therapy (daratumumab)	4 (12.5)
Bruton tyrosine kinase (BTK) inhibitors (ibrutinib)	4 (12.5)
Antimetabolites therapy (azacitidine, methotrexate)	3 (9.4)
Immune modulatory therapy (lenalidomide, pomalidomide)	3 (9.4)
Vinca alkaloid therapy (vincristine, vindesine)	2 (6.3)
Alkylating agent therapy (cyclophosphamide)	1 (3.1)
Median time since allogeneic hematopoietic stem cell transplantation, no. of days (IQR)	545 (180–2369)
Type of COVID-19 vaccine (1st dose)
Comirnaty®	19 (59.4)
Spikevax®	12 (37.5)
Vaxzebria®	1 (3.1)
Type of vaccine (2nd dose)
Comirnaty®	19 (59.4)
Spikevax®	12 (37.5)
Vaxzebria®	1 (3.1)
Type of vaccine (booster dose)
Comirnaty®	19 (59.4)
Spikevax®	13 (40.6)
Vaccine homology
Until 2nd dose	32 (100)
Until booster dose	28 (87.5)
Heterologous in booster dose	4 (12.5)
SARS-CoV-2 infection status
No documented infection	26 (81.3)
Before the 1st dose	4 (12.5)
After the 2nd dose	1 (3.1)
After the 3rd dose	1 (3.1)

Enumeration of SARS-CoV-2 spike-targeted IFN-γ- and TNF-α-producing T cells

The frequency of SARS-CoV-2-S-reactive cytokine-producing CD4⁺ or CD8⁺ T cells was determined using a commercially available flow cytometry for intracellular staining assay (FC-ICS), the SARS-CoV-2 T Cell Analysis Kit for peripheral blood mononuclear cells (PBMCs) (Miltenyi Biotec) as previously detailed.³⁸ Specifically, we enumerated monofunctional (IFN-γ- or TNF-α-producing) CD4⁺ and CD8⁺ T cells as well as bifunctional (IFN-γ- and TNF-α-producing) CD4⁺ and CD8⁺ T cells. Briefly, PBMCs were transferred into two tubes, antigen (Ag) and negative control (NC). Stimulation (Ag) was conducted with 50 μL of Peptivator SARS-CoV-2 S protein, a pool of lyophilized peptides (at 1 μg/mL) that covers the entire protein sequence of the Wuhan-Hu-1 S protein. The NC tube was mock-stimulated with dimethyl sulfoxide. A volume of 20 μL of Brefeldin A (2 μg/mL) was then added to each tube, and samples were incubated at 37 °C 5% CO₂ for 12 h. After incubation, cells were washed with autoMACS Running Buffer (PEB buffer; PBS, pH 7.4, 0.5% bovine serum albumin, and 2 mM EDTA), and then fixed with Fixation Mix. After two additional washing steps with autoMACS Running Buffer, 1 mL of StemMACS Cryo-Brew was added, and samples were kept at −80 °C until analysis by flow cytometry. On the day of analysis, samples were thawed and washed twice with PEB buffer, then 1 mL of Inside Perm was added. After centrifugation, 100 μL of antibody staining cocktail (CD3-APC, CD4-FITC, CD8-BV510, IFN-γ PE, TNF-α PE-Vio 770) was added to each sample and incubated for 10 min in the dark. Monoclonal antibody clones were the following: CD3+ (REA613), CD4⁺ (REA623), CD8⁺ (REA734), IFN-γ (REA600), and TNF-α (REA656). After incubation, an additional 1 mL of Inside Perm was added and samples were centrifuged. The resultant cell pellet was resuspended in 250 μL of PEB buffer and samples were acquired in an LSRFortessa™ flow cytometer (BD Biosciences Immunocytometry Systems) using FlowJo v-10 software (BD Biosciences). CD3⁺/CD8⁺ and CD3⁺/CD4⁺ events were gated and then analyzed for IFN-γ and TNF-α production. The gating strategy and representative plots are shown in Supplementary Fig. 1. All data were corrected for background cytokine production (NC tube) and expressed as the number of SARS-CoV-2-reactive monofunctional (IFN-γ-producing or TNF-α-producing) CD4⁺ or CD8⁺ T cells and bifunctional (IFN-γ and TNF-α-producing T cells) relative to the absolute number of CD4⁺ or CD8⁺ T cells, respectively, ×100 (%). Any value for SARS-CoV-2-reactive cytokine-producing CD4⁺ or CD8⁺ T cells after background subtraction was considered a positive (detectable) result and used for analytical purposes. Background IFN-γ-producing CD4⁺ or CD8⁺ T-cell frequencies were typically ≤ 0.10%. For some analyses, the integrated mean fluorescence intensity (iMFI) for cytokine-producing CD4⁺ or CD8⁺ T cells was considered, which was calculated using the formula iMFI = MFI × P, where P is the percentage of cells expressing the respective cytokine.

Antibody assays

Quantitation of total antibodies targeting the receptor-binding domain (RBD) of the SARS-CoV-2 Wuhan-Hu-1 S protein was carried out using the Roche Elecsys® Anti-SARS-CoV-2 S assay (Roche Diagnostics, Pleasanton, USA). Anti-SARS-CoV-2 nucleocapsid IgGs, a qualitative test for evaluating previous exposure to SARS-CoV-2, were assessed by the Roche Elecsys® Anti-SARS-CoV-2 Nucleocapsid IgG assay (Roche Diagnostics).

Statistical analyses

Frequency comparisons for categorical variables were carried out using Fisher's exact test. Comparisons of continuous variables between independent groups (allo-HCT vs non-allo-HCT patients) were performed using the Mann–Whitney U test, whereas comparisons between paired samples across time points (Post-2D, Pre-3D, and Post-3D) were performed using the Wilcoxon test. Correlations between continuous variables were evaluated using the Spearman rank correlation coefficient. When indicated, results were summarized as median and interquartile range (IQR). Integrated median fluorescence intensity (iMFI) values were analyzed among participants with detectable responses to estimate cytokine production per responding cell. The main outcome measures were the frequency and functionality of SARS-CoV-2-specific CD4⁺ and CD8+ T-cell responses, the levels of anti-RBD total antibodies, and the degree of concordance and correlation between humoral and cellular responses at each testing time point. Two-sided exact P-values are reported; a P-value < 0.05 was considered statistically significant. Statistical analyses and graphical representations were performed using GraphPad Prism version 10.0.0 for Windows.

Results

Relevant features of the study population

The median age of participants was 64 years (IQR, 56–75), and 19 (59.4%) were male. Relevant features of the cohort are as follows: (i) of the 32 patients included, 22 (68.7%) were non-allo-HCT recipients, whereas 10 (31.3%) had undergone allo-HCT before receiving any COVID-19 vaccine dose, a median of 545 days before recruitment; (ii) 28 patients (87.5%) were under active treatment at baseline (Table 1); (iii) all but one patient received mRNA COVID-19 vaccines; (iv) homologous COVID-19 vaccination schedules were employed in most patients (100% for the second vaccine dose and 87.5% for the third); and (v) based upon registry records and/or the presence or absence of anti-SARS-CoV-2-N-specific antibodies, 26/32 patients (81.3%) were categorized as naïve for SARS-CoV-2 infection throughout the study period. The remaining six patients contracted SARS-CoV-2 infection either prior to receipt of the first vaccine dose (n = 4), after the second dose (n = 1), or after the third dose (n = 1).

Kinetics of SARS-CoV-2 Spike-directed T-cells We used a commercially available FC-ICS assay for measuring SARS-CoV-2-S-T-cell frequencies in whole blood. The number of participants exhibiting detectable SARS-CoV-2 T-cell responses at the different testing times is shown in Table 2. The most relevant observations of these analyses can be summarized as follows: (i) at Post-2D, around 72.4%, 44.8%, and 37.9% of participants had detectable IFN-γ-, TNF-α-, and IFN-γ/TNF-α-producing CD8⁺ T cells, respectively; (ii) a drop in the rate of detectable monofunctional and bifunctional SARS-CoV-2-S CD8⁺ T-cell responses was seen at Pre-3D, which was more marked for IFN-γ-producing CD8⁺ T cells. By contrast, the number of participants displaying detectable CD4⁺ T-cell responses was rather conserved or showed a modest increase; and (iii) receiving a third vaccine dose did not substantially modify the rate of patients with detectable responses, regardless of the functional T-cell subset considered for the analyses. No major differences in the detection rate of detectable cytokine-producing T-cell responses were seen between allo-HCT and non-allo-HCT patients. From a quantitative standpoint, the following observations were made. First, SARS-CoV-2-S–specific CD4+ T-cell frequencies remained largely unchanged over time, whereas CD8⁺ T-cell frequencies tended to decline between Post-2D and Pre-3D. This decrease reached statistical significance only for bifunctional IFN-γ/TNF-α–producing CD8⁺ T cells (P = 0.02), while all other comparisons did not achieve statistical significance (Fig. 1, panels A–F). Secondly, a third vaccine dose had no significant impact on the frequencies of any of the above functional SARS-CoV-2-S-directed T-cell subsets (Fig. 1, panels A–F). Frequency of SARS-CoV-2–reactive CD4⁺ and CD8⁺ T cells producing TNF, IFN-γ or both cytokines is consistently low across the cohort. Maximum values reached 2.1%, 3.7%, and 2.6% for CD4⁺ T cells and 4.3%, 13.8%, and 5% for CD8⁺ T cells, respectively, while most samples showed frequencies below 0.5% or even undetectable responses. Despite the above trends, as shown in Fig. 2 (panels A–F), individual variations were observed across participants. To assess the magnitude of cytokine production by stimulated T cells, we next examined the iMFI of IFN-γ- and TNF-α-producing CD8⁺ and CD4⁺ T cells among participants displaying detectable responses. As shown in Fig. 3 (panels A–D), no significant variations in iMFI were observed across testing times.

Table 2.

Detectable SARS-CoV-2 spike cytokine-producing T-cell responses in the participants.

Study group and testing time point	Detectable IFN-γ CD8^+ T-cell response		P-value	Detectable TNF-α CD8^+ T-cell response		P-value	Detectable IFN-γ and TNF-α CD8^+ T-cell response		P-value
	Yes	No		Yes	No		Yes	No
All participantsPost-2D/Pre-3D	21/17	9/12	0.12	13/12	16/20	0.56	19/10	10/22	0.001
All participantsPre-3D/Post-3D	17/8	12/11	0.45	12/7	20/12	0.96	10/6	22/13	0.98
Allo-HCT/Non-Allo-HCT Post-2D	6/15	4/4	0.28	4/9	6/10	0.70	6/13	4/6	0.65
Allo-HCT/Non-Allo-HCT Pre-3D	2/13	8/6	0.01	2/9	8/10	0.15	2/9	8/10	0.15
Allo-HCT/Non-Allo-HCT Post-3D	2/5	1/9	0.32	1/5	2/9	0.94	1/5	2/9	0.69

Study group and testing time point	Detectable IFN-γ CD4^+ T-cell response		P-value	Detectable TNF-α CD4^+ T-cell response		P-value	Detectable IFN-γ and TNF-α CD4^+ T-cell response		P-value
	Yes	No		Yes	No		Yes	No
All participantsPost-2D/Pre-3D	16/22	13/10	0.27	14/13	15/19	0.55	12/15	17/17	0.67
All participantsPre-3D/Post-3D	22/10	10/9	0.25	13/10	19/9	0.40	15/8	17/11	0.74
Allo-HCT/Non-Allo-HCT Post-2D	6/10	4/9	0.70	4/10	6/9	0.52	3/9	7/10	0.37
Allo-HCT/Non-Allo-HCT Pre-3D/	5/10	5/9	0.89	3/10	7/9	0.24	5/10	5/9	0.89
Allo-HCT/Non-Allo-HCT Post-3D	1/6	2/8	0.09	1/6	2/8	0.76	2/5	1/9	0.32

Allo-HCT, allogeneic hematopoietic stem cell transplant recipient; D, vaccine dose.

Fig. 1.

Frequencies of cytokine-producing SARS-CoV-2 spike-directed CD8⁺ and CD4⁺ T cells after receipt of the second COVID-19 vaccine dose (Post-2D), before receipt of the third dose (Pre-3D), and after receipt of the third vaccine dose (Post-3D). Panels A and B: IFN-γ-producing T cells; Panels C and D, TNF-α-producing T cells. Panels E and F, bifunctional IFN-γ-/TNF-α-producing T cells. Bars indicate medians and interquartile ranges. P-values for comparisons are shown.

Fig. 2.

Individual kinetics of cytokine-producing SARS-CoV-2 spike-directed CD8⁺ and CD4⁺ T cells. Panels A and B: IFN-γ-producing T cells; Panels C and D: TNF-α-producing T cells. Panels E and F: bifunctional IFN-γ-/TNF-α-producing T cells. Immunological testing after receipt of the second COVID-19 vaccine dose (Post-2D), before receipt of the third dose (Pre-3D), and after receipt of the third vaccine dose (Post-3D).

Fig. 3.

Integrated mean fluorescence intensity (iMFI) for cytokine-producing CD4⁺ or CD8⁺ T cells. Panels A and C: IFN-γ-producing T cells. Panels B and D: TNF-α-producing T cells. Immunological testing after receipt of the second COVID-19 vaccine dose (Post-2D), before receipt of the third dose (Pre-3D), and after receipt of the third vaccine dose (Post3D). Bars indicate medians and interquartile ranges. P-values for comparisons are shown.

The sample size was limited; however, we next conducted a subanalysis comparing monofunctional and bifunctional CD8⁺ and CD4⁺ T-cell frequencies between allo-HCT and non-allo-HCT patients. The data are shown in Supplementary Fig. 2. No significant differences were noted for any functional T-cell subset.

Overall, as shown in Fig. 4, the correlation between cytokine-producing SARS-CoV-2 CD8⁺ and CD4⁺ T-cell frequencies was moderate for TNF-α-producing T cells (Rho, 0.50) and weak for IFN-γ-producing (Rho = 0.23) and IFN-γ/TNF-α-producing T cells (Rho = 0.16).

Fig. 4.

Overall correlation between frequencies of SARS-CoV2-S cytokine-producing CD8⁺ and CD4⁺ T cells. Panel A: IFN-γ-producing T cells as assessed by the Spearman Rank test. Panel B: TNF-α-γ-producing T cells. Panel C: bifunctional IFN-γ-/TNF-α-producing T cells. Rho and P-values are shown.

Kinetics of SARS-CoV-2-receptor-binding-domain antibodies

It was not possible to determine the proportion of patients lacking B lymphocytes in our study, as B-cell counts (e.g., CD19⁺ cells) were not available due to the characteristics of the assay used (Miltenyi kit). Overall, anti-RBD total antibody responses were detected in 51.7%, 46.8% and 52% participants at Post-2D, Pre-3D, and Post-3D testing time points. Undetectable responses were more frequently noted among non-allo-HCT patients than allo-HCT patients (Table 3). Overall, anti-RBD antibody levels decreased at Pre-3D (P = 0.85) and then increased following receipt of the 3D (P = 0.12). At all testing times, anti-RBD antibody levels were significantly higher in allo-HCT than in non-allo-HCT patients (Table 4).

Table 3.

Detectable SARS-CoV-2 spike-receptor-binding-domain total antibodies in the participants.

Study group and testing time point	Detectable anti-SARS-CoV-2-RBD total antibodies		P-value
	Yes	No
All participants Post-2D/Pre-3D	15/15	14/17	0.71
All participants Pre-3D/Post-3D	15/10	17/9	0.69
Allo-HCT/Non-Allo-HCT Post-2D	9/5	1/11	<0.01
Allo-HCT/Non-Allo-HCT Pre-3D/	9/5	1/14	<0.01
Allo-HCT/Non-Allo-HCT Post-3D	4/6	0/9	0.09

Allo-HCT, allogeneic hematopoietic stem cell transplant recipient; D, vaccine dose.

Table 4.

SARS-CoV-2 spike receptor-binding domain total antibody levels in the participants.

Study group and time of immunological testing	Number of patients	Anti-SARS-Cov-2-RBD total antibodies (log10 BAU/mL), median (IQR)	P-value
Post-2D
Allo-HCT	10	3.23 (0.71–3.39)	<0.01
Non-allo-HCT	16	0 (0–0)
All participants	29	0 (0–2.19)
Pre-3D
Allo-HSCT	10	2.9 (0.42–3.47)	<0.01
Non-allo-HSCT	19	0 (0–0.19)
All participants	32	0 (0–1.70)
Post-3D
Allo-HSCT	4	3.80 (2.10–4.84)	0.01
Non-allo-HSCT	15	0 (0–2.64)
All participants	21	1.34 (0–3.06)

Allo-HCT, allogeneic hematopoietic stem cell transplant recipient; D, vaccine dose.

Concordance between SARS-CoV-2 antibody and T-cell immune responses

Overall, at all testing times, around 50% (33–66%) exhibited qualitative discordant adaptive immune responses, in most cases monofunctional or bifunctional cytokine-producing T-cell detectable responses and undetectable anti-RBD antibody levels (Supplementary Table 1). The percentage of patients without detectable T-cell and anti-RBD antibody responses varied depending upon the cytokine-producing T-cell response considered; nonetheless, overall, it was <20% of patients. Finally, the degree of correlation between T-cell and anti-RBD antibody responses was poor for all cytokine-producing T-cell subsets (Fig. 5).

Fig. 5.

Overall correlation between anti-receptor-binding domain (RBD) total antibodies and frequencies of SARS-CoV2-S cytokine-producing CD8⁺ and CD4⁺ T cells, as assessed by the Spearman Rank test. Panels A and B: IFN-γ-producing T cells; Panels C and D: TNF-α-producing T cells. Panels E and F: bifunctional IFN-γ-/TNF-α-producing T cells as assessed by the Spearman Rank test. Panel B: TNF-α-γ-producing T cells. Panel C: bifunctional IFN-γ/TNF-α-producing T cells. Rho and P-values are shown.

Studies evaluating both SARS-CoV-2 S protein directed antibody and T-cell responses in hematological patients following receipt of a third mRNA COVID-19 dose are relatively limited and yield variable results regarding the magnitude, kinetics and correlation of humoral and cellular responses.^21–36 Likely, discrepancies across published studies are related to several factors, including the variable timing of immunological screening after vaccination, the wide variety of underlying hematological conditions in patients across cohorts, and, importantly, the methodology used for antibody and T-cell immunological evaluations. Here, we assessed SARS-CoV-2-S-directed cytokine-producing monofunctional (IFN-γ or TNF-α) and bifunctional (IFN-γ/TNF-α) CD8⁺ and CD4⁺ T-cell responses by FC-ICS and anti-RBD total antibodies using a commercially-available chemiluminescent assay in a cohort mainly comprising SARS-CoV-2 naïve non-allo-HCT hematological patients that underwent B-cell-depleting therapies, received Bruton's tyrosine kinase, Janus kinase inhibitors, or conventional chemotherapy. The cohort also included allo-HCT patients vaccinated a median of 545 days after transplantation. All but one participant received three mRNA COVID-19 vaccine doses, with the third dose being homologous in most cases. Our main interest was to gauge the impact of receiving a third vaccine dose (Post-3D), whose administration is recommended by consensus guidelines²⁰ on SARS-CoV-2-S T-cell and antibody responses. However, we also assessed the kinetics of these adaptive immune responses following receipt of the second vaccine dose until the administration of the third vaccine dose (Post-2D/Pre-3D). Immunological evaluations following the 3D were conducted a median of 31 days after vaccinations, when de novo elicited or boosted responses were likely captured. From a qualitative standpoint, our data revealed a conserved rate of detectable SARS-CoV-2-S CD4⁺ T-cell responses Post-2D compared with that Pre-3D; importantly, around half of the participants, irrespective of their baseline clinical characteristics, failed to mount detectable responses. In contrast, the percentage of participants with detectable SARS-CoV-2-S CD8⁺ T-cell responses at Post-2D was higher than that of CD4⁺ T-cell responses and tended to decrease over time. Strikingly, receipt of the third vaccine dose failed to increase the percentage of patients with detectable T-cell responses, regardless of the functional T-cell subset considered and whether or not the patients had undergone allo-HCT. Quantitative analyses showed conserved SARS-CoV-2-S cytokine-producing CD4⁺ T-cell responses and a non-statistically significant drop in the frequencies of SARS-CoV-2-S cytokine-producing CD8⁺ T cells between Post-2D and Pre-3D and a lack of a boosting effect of 3D, either in terms of T-cell frequencies or iMFI values, although individual variations were observed across participants. Importantly, T-cell response kinetics were rather similar in allo-HCT and non-allo-HCT patients.

Discussion

Our observations on the effect of the 3D on SARS-CoV-2-S-directed T-cell responses in hematological patients align with those of most studies addressing this issue. For example, Kimura et al.²¹ evaluated the immunogenicity of a three-dose mRNA SARS-CoV-2 vaccination schedule in a cohort of adult allo-HCT recipients at varying times after transplantation. The authors observed a trend toward an increase in the frequency of IFN-γ monofunctional CD4⁺ T-cells after 3D mRNA vaccination, as assessed by FC-ICS, however, this did not reach statistical significance. Likewise, in agreement with our data, the authors found no differences in the frequency of S-specific total IFN-γ expressing CD8⁺ T-cells measured between the 2D and 3D. Using a cytometric bead array, Body et al.²⁴ evaluated humoral and cellular immune responses to COVID-19 vaccination in a heterogeneous cohort of adults with hematological and solid malignancies, and also reported a minimal impact of the 3D on IFN-γ T-cell responses in patients with hematological cancer. Ko et al.³⁷ investigated cellular and humoral immune responses after a three-dose SARS-CoV-2 vaccination schedule in a mixed cohort comprising healthy individuals and immunocompromised patients, including a substantial subgroup of patients with hematological malignancies receiving diverse immunosuppressive treatments. They reported impaired T-cell responses as measured by an Interferon-gamma release assay (IGRA) following the 3D in the cohort of hematological patients. Likewise, Haidar et al.,²⁶ who prospectively evaluated humoral and cellular immune responses following administration of a 3D in a large heterogeneous cohort of immunocompromised individuals, including patients with hematological malignancies, found that the 3D did not improve SARS-CoV-2 T-cell responses among hematological patients. Gressens et al.²⁸ reported a limited impact of a booster dose (3D) of BNT162b2 in patients with B-cell lymphomas receiving anti-CD20 monoclonal antibody therapy. Azeem et al.²⁹ showed that three doses of SARS-CoV-2 mRNA vaccines failed to yield detectable SARS-CoV-2-S-directed CD8⁺ T cells in >80% of patients with multiple myeloma. In contrast, other studies^23,25,32,33 found that a 3D boosted SARS-CoV-2 cytokine-producing T-cell responses in allo-HCT recipients or patients with B-cell malignancies, as measured by various immunological procedures. Pinder et al.²³ observed that successive vaccination, including a 3D administered approximately 4–6 months after the second dose, improved cytokine-producing T-cell responses and antibody titers in patients with B-cell malignancies. Campanella et al.²⁵ reported that an additional booster (3D) given at a median of 3 months post-second dose elicited robust humoral and T-cell responses in chronic lymphocytic leukemia patients, comparable to those seen after natural infection, regardless of ongoing therapies. Re et al.³² found that a 3D of BNT162b2 vaccine enhanced both humoral responses and functional T-cell activity in patients with lymphoid malignancies, with a marked increase in IFN-γ production. Similarly, Albiol et al.³³ showed that in recently transplanted allo-HCT recipients, administration of a mRNA booster (3D) approximately 2–3 months after the initial vaccine regimen significantly increased the proportion of patients with detectable spike-specific T-cell responses and neutralizing antibody levels. The persistently low frequencies of SARS-CoV-2-reactive CD4⁺ and CD8⁺ T cells producing TNF, IFN-γ or both cytokine across our cohort suggests a limited magnitude of cytokine-producing T-cell responses to SARS-CoV-2 in these hematological patients. In contrast, studies in healthy, non-immunocompromised individuals generally describe higher frequencies of SARS-CoV-2-specific cytokine-producing T cells and the induction of robust and coordinated immune responses, supporting the idea that cellular immune responses may be quantitatively and/or functionally compromised in this immunodepressed population. In these individuals, most CD4⁺ and CD8⁺ T cells specific for SARS-CoV-2 produce cytokines such as IFN-γ and TNF-α, and nearly all individuals develop detectable levels of neutralizing antibodies.^40–42 Importantly, the magnitude of T-cell responses correlates with antibody production, highlighting the integrated nature of humoral and cellular immunity in non-immunodeficient hosts. These findings serve as a reference point for evaluating the lower or heterogeneous responses observed in immunocompromised patients. Regarding anti-RBD total antibody responses, our data revealed that a large percentage of patients (overall, around 50%), in particular non-allo-HCT participants, failed to mount detectable responses even after the 3D, in line with published data [discussed in 17–19]. Gagelmann et al.¹⁷ reported seroconversion rates below 60% after full vaccination in this population, with even lower responses observed in subgroups such as chronic lymphocytic leukemia and those receiving B-cell-depleting therapies. Teh et al.¹⁸ similarly found markedly reduced immunogenicity in hematologic malignancies, with substantial heterogeneity depending on disease type and treatment status. Additionally, prospective cohort data from Haggenburg et al.¹⁹ demonstrated that despite a 3D mRNA schedule, a significant proportion of patients with hematologic malignancies remained seronegative or exhibited low anti-spike antibody levels. Nevertheless, receipt of the 3D seemed to boost anti-RBD antibody responses in allo-HCT with detectable responses at the 2D, as previously reported in a cohort of recently transplanted allo-HCT patients.²⁴

Moreover, our data further reinforced the idea that SARS-CoV-2-S-antibody and T-cell responses may follow dissimilar kinetics in hematological patients following 2D and 3D as previously suggested.^23–26,32 In this context, Fig. 5 further illustrates this dissociation, showing that a substantial proportion of patients with high or very high anti–SARS-CoV-2 spike antibody levels exhibited very low, and in many cases undetectable, frequencies of IFN-γ- and/or TNF-producing CD4⁺ and CD8⁺ T cells. The results obtained suggest that, in this patient cohort, the magnitude of cytokine production by SARS-CoV-2-specific T lymphocytes is not proportional to the humoral response. This finding is consistent with impaired functional capacity of this T-cell compartment in immunocompromised hematological patients, such as those undergoing hematopoietic stem cell transplantation.

Importantly, humoral immunity does not rely directly on IFN-γ- or TNF-producing CD4⁺ or CD8⁺ T cells per se, but rather on specialized CD4⁺ T-cell subsets, particularly T follicular helper (Tfh) cells, which play a central role in germinal center reactions and antibody affinity maturation. These T-cell subsets were not specifically evaluated in the present study, which may partially explain the lack of correlation observed between cellular cytokine responses and antibody levels.

In immunocompetent individuals, SARS-CoV-2 vaccination and infection have been shown to induce coordinated humoral and cellular immune responses, including higher frequencies of cytokine-producing CD4⁺ and CD8⁺ T cells, as well as robust Tfh responses that correlate with antibody titers. In contrast, patients with hematological malignancies or receiving B-cell-depleting therapies frequently display a dissociation between humoral and cellular immunity, as observed in our cohort. In this sense, monofunctional or bifunctional cytokine-producing T-cell detectable responses were observed concomitantly with undetectable anti-RBD antibody levels. Strikingly, a non-negligible percentage of hematological patients (20% in this series) fail to develop both detectable antibody and T-cell responses after a three-dose vaccination regimen. Interestingly, more than half of the patients in our series exhibited detectable SARS-CoV-2-specific T-cell responses despite undetectable anti-RBD antibody levels. This observation underscores the contribution of cellular immunity to vaccine-induced protection in hematological patients with defective humoral responses, particularly those under B-cell-depleting therapies or following allo-HCT.

The current study is primarily limited by its small size, missing specimens at some time points, and the rather heterogeneous baseline conditions of the participants. These precluded a more comprehensive analysis of post-vaccination immune responses according to the underlying hematological disease. In addition, the absence of a control group followed over a comparable period limits our ability to definitively assess whether the third dose enhances SARS-CoV-2-S-directed immune responses or primarily contributes to their maintenance by preventing a time-dependent decline. Likewise, the ability of vaccine-elicited antibodies and T cells to recognize currently circulating SARS-CoV-2 (sub)variants was not explored. In conclusion, our data strongly suggested that receipt of the 3D may have a minimal impact on SARS-CoV-2-S-directed T-cell responses both from qualitative and quantitative standpoints in hematological patients. This also appears to be the case after receipt of a fourth vaccine dose in allo-HCT.³⁶ Analogous results were obtained for anti-RBD total antibodies in non-allo-HCT participants. Despite the proven efficacy of mRNA-based COVID-19 vaccines against SARS-CoV-2 Omicron subvariants in patients with hematological malignancies,^38,39 our data warn of a possible suboptimal immunogenicity of repeat boosting with COVID-19 vaccines in certain hematological patients. Our findings suggest that repeated mRNA vaccine boosting may have a limited and heterogeneous immunological impact in hematological patients, particularly among non-allo-HCT recipients. In this sense, further studies are required to determine whether currently available antibody or T-cell assays could identify those patients who may benefit most from seasonal revaccination. Moreover, the identification of patients who fail to mount both antibody and T-cell responses after three vaccine doses highlights the need for alternative or complementary protective strategies in these high-risk subgroups.

Author contributions

Diego Carretero (Data curation, Formal Analysis, Writing – original draft, Writing – review & editing), Estela Giménez (Conceptualization, Methodology, Data curation, Formal Analysis), Ariadna Pérez (Conceptualization, Methodology), Rafael Hernani (Conceptualization, Methodology), Jose Luis Piñana (Conceptualization, Methodology), Carlos Solano (Conceptualization, Methodology), David Navarro (Conceptualization, Investigation, Writing – original draft, Writing – review & editing, Supervision), Eliseo Albert (Conceptualization, Data curation, Methodology, Supervision, Investigation). All authors have read and agreed to the published version of the manuscript.

Ethical considerations

The study was approved (2022/351) by the Institutional (INCLIVA) Ethics Committee.

Informed consent

All subjects agreed to voluntarily participate in the study and gave written consent.

Use of artificial intelligence

Not apply.

Funding

This study was supported by the Instituto de Salud Carlos III, Madrid, Spain (FIS, PI21/00563) granted to David Navarro.

Conflicts of interest

The authors declare no conflicts of interest.

Data availability statement

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

Acknowledgements

Eliseo Albert (Juan Rodés Contract; JR20/00011) holds a contract funded by the Carlos III Health Institute (co-financed by the European Regional Development Fund, ERDF/FEDER).