Enhanced BNT162b2 vaccine-induced cellular immunity in anti-CD19 CAR T cell–treated patients

TO THE EDITOR:

Although severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections have spread worldwide, vaccines are highly effective in preventing severe COVID-19 and mortality.¹ The emergence of the Omicron variant with more than 30 mutations in the spike protein drastically reduced the neutralizing ability of vaccine-induced antibodies,² leading to frequent breakthrough infections.³ Nevertheless, vaccine-induced immunity remains highly effective in preventing severe COVID-19.⁴ This is probably mediated by Spike-specific T cells, whose inductions have been shown to correlate with onset of vaccine efficacy⁵ and are largely unaffected by Omicron mutations.^6-9 

CD19-targeting chimeric antigen receptor (CAR) T-cell therapy is a major breakthrough in the treatment of relapsed/refractory B-lineage malignancies, whose bystander effect is long-term B-cell aplasia.¹⁰ Consequently, these patients do not mount a humoral response after COVID-19 vaccination.^11-13 Initial data demonstrated an induction of Spike-specific CD4 T cells in adult patients¹⁴; however, an in-depth characterization of vaccine-induced T-cell responses and their resistance to emerging viral variants is lacking.

Here, we longitudinally studied anti-CD19 4-1BB-CD3z-CAR T cell–treated patients, mostly adolescents and young adults, before and after the first and second dose of vaccination with BNT162b2. We analyzed the magnitude of the T-cell response, the phenotypes of this response, their multispecificity, and their ability to respond to SARS-CoV-2 variants B.1.617.2 (Delta) and B.1.1.529 (Omicron).

We recruited 8 patients who had received single infusions of anti-CD19 CAR T cells at least 6 months before 2-dose vaccination with BNT162b2 3 weeks apart (supplemental Table 1 available on the Blood Web site) and 26 healthy controls (supplemental Table 2). None of the individuals in the study were infected with SARS-CoV-2 before vaccination or during the study period; none of the patients were receiving any immunosuppressive ,therapies and all were clinically well. Other baseline and disease specific characteristics are described in supplemental Table 1.

All 7 adolescents and young adults, except the 1 elderly patient in the study, experienced mild adverse events (local and systemic) especially after the second dose (Figure 1A). Seven of 8 patients mounted none to only minimal levels of Spike-specific antibodies following vaccination (Figure 1B). Only patient 9, who lost persistence of CAR T cells at the beginning of the study period and subsequently had detectable B cells, demonstrated an antibody response after the second dose.

The Spike-specific T-cell response induced by vaccination was longitudinally analyzed at 6 different time points: before vaccination, 10 and 21 days after the first and second dose, and then at 90 days after vaccination. Whole blood was stimulated with a Spike-peptide pool overnight, and secreted interferon γ (IFN-γ) and interleukin-2 (IL-2) were quantified.¹⁵ 

Before vaccination, supernatants of whole blood cultures from CAR T cell–treated patients and healthy controls contained low median IFN-γ (<1.7 pg/mL) and IL-2 (<1.4 pg/mL) levels (Figure 1C-D). Ten days after the first dose, peptide-triggered IFN-γ and IL-2 clearly increased in both groups similarly, in line with the previously demonstrated ability of BNT162b2 to rapidly induce Spike-specific T cells.^5,16 In contrast, after the second dose, we observed that the patients with CAR T cells produced quantities of IFN-γ (Figure 1C) and IL-2 (Figure 1D) after Spike-peptide stimulation that were 1 log higher than that detected in healthy controls. We followed 3 of the patients for 180 days after vaccination, and their peptide-induced cytokine release remained high (supplemental Figure 1).

To better define whether the enhanced response to Spike peptides observed after the second dose was derived from Spike-specific CD4 or CD8 T cells, we stimulated peripheral blood mononuclear cells (PBMCs) collected on days 0 and 42 with the Spike-peptide pool for expression of activation markers on gated nonnaïve CD4⁺ and CD8⁺ T cells (supplemental Figure 2). We visualized higher frequencies of Spike-specific CD4 T cells in all patients with CAR T cells (median = 4.4%) compared with healthy controls (median = 2.0%; Figure 1E). Similarly, higher median levels of Spike-specific CD8 T cells were also detected in patients with CAR T-cells (healthy median = 0.12%; CAR T cells median = 0.26%), with 3 of 8 showing frequencies that were above that of healthy controls (Figure 1F).

The emergence of highly mutated variants of concern (VOCs) has highlighted that vaccine-induced humoral immunity cannot prevent infection.^2,3 Yet, vaccines are still able to provide effective protection from severe COVID-19.⁴ This can be explained by the ability of cellular immunity to target numerous epitopes along the whole Spike protein and therefore remains largely intact against emerging VOCs.^6-9 We thus tested whether vaccine-induced T cells in patients with CAR T cells were also directed against multiple Spike regions.

PBMCs were stimulated with 7 Spike-peptide pools, each covering about 180 amino acids of the Spike protein and analyzed by IFN-γ-ELISpot assay. Figure 2A-C shows the ability of vaccine-induced T cells present in patients with CAR T cells to recognize multiple Spike regions, similar to those in healthy controls. Subsequently, we tested their capacity to respond to SARS-CoV-2 variants B.1.617.2 (Delta) and B.1.1.529 (Omicron). PBMCs were stimulated with 5 peptide pools covering the entire Spike (253 peptides) of the ancestral SARS-CoV-2 and the mutated regions in the Delta (30 peptides) and Omicron (68 peptides) variants, with and without the amino acid substitutions/deletions characteristic for these 2 VOCs. The calculated frequencies of IFN-γ-spot forming cells in response to Spike peptides of the VOCs was compared with the ancestral strain (Figure 2D). Vaccine-induced T-cell immunity was almost completely preserved against Delta in all patients, except in 1 where we detected a reduction of nearly 50% (mean = 9.7% inhibition; Figure 2Ei). Unsurprisingly, the higher mutated Omicron variant did impact the T-cell response more (mean = 20.8% inhibition), yet only in 1 patient did we observe a strong reduction in the T-cell response of nearly 80% (Figure 2Eii). Investigations are underway to identify which T-cell epitopes are affected by Omicron mutations in this individual, who is characterized by the HLA-type A*11:01, A*33:03, B*40:01, B*58:01, C*03:02, C*07:02, DRB1*03:01, and DRB1*11:01.

Although our study is limited by its small sample size and the use of a single vaccine platform (BNT162b2), our results are reassuring that a stronger induction of Spike-specific T cells after vaccination in patients with anti-CD19 CAR T-cells is seen compared with healthy individuals. Interestingly, this augmented response was only evident after the second dose, similar to observations in vaccinated anti-CD20–treated patients.¹⁷ In the absence of vaccine-induced antibody, mRNA-derived Spike antigens might persist longer and were more efficiently presented, resulting in a more robust T-cell response. This hypothesis is consistent with recent data in healthy individuals where low antibody titers were correlated with an enhanced boosting effect from the third vaccine dose.¹⁸ This observed enhanced cellular immunity, comprised of both CD4 and CD8 T cells and further characterized by broad specificity against various regions of Spike and its ability to tolerate mutations present in VOCs, might offer some protective efficacy.¹⁹ Our results support the clinical utility and importance of COVID-19 vaccination in patients after anti-CD19 CAR T-cell therapy despite B-cell aplasia.