Peripheral blood mononuclear cells (PBMCs) were isolated from leukocyte concentrates of healthy adult blood donors obtained from the Institute of Transfusion Medicine (Institute for Transfusion Medicine, University Hospital Schleswig-Holstein Campus Kiel) using Ficoll density gradient centrifugation (Merck, Darmstadt, Germany). Ethical approval was granted by the institutional ethics review board of the University Medical Faculty Kiel (approval number D405/10, D479/18, D546/16). The research was conducted in accordance with the Declaration of Helsinki. CD4⁺, CD8⁺, and γδ T cells or when required entire CD3⁺ T cells were purified from PBMCs by magnetic-activated cell sorting (MACS) using respective negative selection kits according to the manufacturer’s instructions (Miltenyi Biotec, Bergisch Gladbach, Germany).

To establish αβ T-cell lines, PBMCs were stimulated with Phytohemagglutinin A (PHA; 0.5 µg/ml; Thermo Fisher Scientific, Waltham, MA, USA) and recombinant interleukin-2 (rIL-2; 100 U/ml; Novartis, Basel, Switzerland) in culture medium which consists RPMI 1640 medium (Thermo Fisher Scientific) supplemented with penicillin (100 U/ml), streptomycin (100 µg/ml), and 10% heat-inactivated fetal bovine serum (FBS; Thermo Fisher Scientific), hereafter referred to as culture medium, following previously established protocols (37, 38). Vδ2 T cells were expanded using Zoledronate (Zole, 2.5 µM; Novartis) and rIL-2 (50 IU/ml) to establish short-term activated Vδ2 T-cell lines as previously reported (39). IL-2 was added every 2-3 days and cultures were split after day 6 or 7. After 14 days, non-viable cells were removed via Ficoll density centrifugation when necessary. The purity of expanded cell lines was assessed at day 14 by flow cytometry.

The human B-cell precursor ALL cell lines NALM-6 and HAL-01 obtained from DSMZ and maintained in required culture medium according to manufacturer’s instructions and utilized as target cells in co-culture assays.

Freshly isolated CD4⁺, CD8⁺, and γδ T cells were co-cultured with HAL-01 cells at an effector−to−target (E:T) ratio of 5:1, with or without BLN (20 ng/ml) for up to 7 days (40). T-cell numbers were assessed via manual counting at day 3 and day 7 and analyzed for their composition and phenotypes by flow cytometry at day 3. Supernatants were collected at specified day 3 and day 7, stored at −20°C, and analyzed via ELISA as described below. In vitro expanded PHA-activated αβ and Zole-activated γδ T cells were co-cultured with NALM-6 or HAL-01 target cells at various E:T ratios in the presence or absence of BLN (20 ng/ml or 0.5 ng/ml) for 24 hours or 3 days. PBMCs from 6 healthy donors were co-cultured for 7 days with the CD19⁺ BCP-ALL cell line HAL-01 at different effector-to-target (E:T) ratios: 1:1 (high CD19 load), 5:1 (medium CD19 load), or without addition of exogenous target cells (low CD19 load, autologous B cells only), in the presence of in the absence of BLN (20 ng/ml). Cells were harvested and analyzed on days 0, 3, and 7 by multiparametric flow cytometry. Effector cells at fixed numbers were cultured with varying amounts of target cells through all co-culture experiments. Recombinant human IL-2 (50 IU/ml) was added to all co-cultures at baseline and replenished every 48 h (no other exogenous cytokines were used). Media volume lost to sampling was replaced with fresh complete medium containing the same IL-2 concentration.

Purity of PHA- and Zole- expanded αβ and γδ T-cell cultures and MACS-isolated T-cell populations were assessed by flow cytometry using anti-CD3, CD4, CD8, γδ TCR, and αβ TCR monoclonal antibodies. Only populations with >80-90% purity were used in co-culture assays. For all analyses, cells were gated on singlets and live cells defined based on scatter properties, followed by CD3⁺ lymphocytes and subsequent αβ/γδ and memory−subset gates (Supplementary Table 1, gating strategy depicted in Supplementary Figures 1, 2).

B-cell depletion in BLN-activated PBMCs or in co-cultures was analyzed on BD FACS Canto using antibodies directed against CD19 and CD3 or at Cytek Northern Lights cytometer using a 22-color antibody panel also incorporating anti-CD3 and anti-CD19 with the following antibodies directed against CD4, CD8, CD25, CD27, CD28, CD38, CD45, CD45RA, CD56, CD95, Annexin V, CCR7, DNAM-1, HLA-DR, PD-1, γδ TCR, TIGIT, TIM-3, Vδ2 γδ TCR and 7-AAD.

Selected samples undergoing single-cell RNA sequencing (BLN-treated and untreated samples of single healthy donor T cells) were analyzed using antibodies against CD3, CD4, CD8, CD16, CD19, CD22, CD24, CD45, CD45RA, CD56, CCR7, DNAM-1, γδ TCR and Vδ2 γδ TCR. Samples were analyzed on a BD FACSLyric. Detailed information of used antibody panels and used clones or formats is provided in Supplementary Table 2.

Cell culture supernatants were analyzed for Interferon-γ (IFNγ; DIF50C), Granzyme B (GrzB; DGZB00), Perforin (QK8011), Granulysin (Gnly; DY3138), Perforin (Prf; QK8011) and soluble Fas-Ligand (sFAS-L; DFL00B) using sandwich ELISA kits from R&D Systems, Minneapolis, USA according to the manufacturer’s instructions. Absorbance was measured at 450 nm using a microplate reader (Tecan Group Ltd, Männedorf, Switzerland) in the Institute of Immunology Kiel.

To quantify CD19 antigen density on the cell surface of BCP ALL cell lines (NALM6 and HAL-01), the specific antibody binding capacity (SABC) was determined using the purified anti-human CD19 antibody (clone HIB19) and the QIFIKIT^® (DAKO, Glostrup, DK) by flow cytometry according to the manufacturer’s instructions in the Division of Stem Cell Transplantation and Cellular Immunotherapies.

BD Rhapsody Single-Cell Analysis System (BD, Biosciences) were utilized using a targeted approach with human T-cell Expression Panel commercially available from BD Biosciences and covering 259 genes, commonly expressed in human T cells (Cat ID 633751, Supplementary Table 3). For this, PBMCs from one healthy donor were cultured for 72 hours with rIL-2 (50 IU/ml) + BLN (20 ng/ml) or with rIL-2 alone as a control. Afterwards, CD3⁺ cells were negatively isolated (human MACS Pan T-cell isolation Kit, Miltenyi Biotec), labeled using BD Single-Cell Multiplexing Kit and AbSeq (antibody-labeled with oligos) reagents targeting CD4, CD8, CD25, CD45RA, CD127, CCR7, and TCRγδ (Supplementary Table 2). Cells from BLN-treated and control samples were labeled with Sample Tags following the manufacturer’s instructions (Single Cell Labelling with the BD™ Single-Cell Multiplexing Kit and BD™ AbSeq Ab-Oligos (ID: 214419 Rev. 2.0)). 10, 000 single cells were load on a cartridge with preloaded beads to capture RNA of single cell, followed by cDNA synthesis using the BD Rhapsody Express Instrument and Scanner following BD protocols (Single Cell Capture and cDNA Synthesis with the BD Rhapsody™ Single-Cell Analysis System (210966 Rev. 1.0)). cDNA libraries of target transcripts, Samples Tags and Abseqs were prepared using the Manufactures instructions, following mRNA Targeted, Sample Tag, and BD™ AbSeq Library Preparation with the BD Rhapsody™ Targeted mRNA and AbSeq Amplification Kit (214508 Rev. 3). The final libraries were quantified using a Qubit Fluorometer with the Qubit dsDNA HS Kit (ThermoFisher) and the size-distribution was measured using the Agilent high sensitivity D5000 assay on a TapeStation 4200 system (Agilent Technologies). Sequencing was performed in paired-end mode (2x75 cycles) on NextSeq 500 System (Illumina) with NextSeq 500/550 Mid Output Kit, generating approximately 1.3 billion reads. Data processing utilized the Seven bridges pipeline BD Rhapsody™ Targeted Analysis Pipeline - Revision: 0. Generated demultiplexed matrices of ScRNA-seq UMI count were imported to R 4.3.3 and gene expression data analysis was performed using the R/Seurat package 5.3.0 (41). To remove doublets and cell fragments, we further fitted a linear model of detected genes versus transcript count per sample and excluded deviating outliers (0.5 - 5% of cells, depending on sample quality). Additionally, we removed genes that were expressed in less than ten cells. Before downstream analysis, LogNormalization (Seurat function) was applied separately to the RNA and AbSeq assays. The data were then scaled regressing for total UMI counts and principal component analysis (PCA) was performed. RNA and AbSeq data were integrated using weighted nearest neighbors (WNN) based on the first 8 RNA and first 6 AbSeq principal components. UMAP and clustering were performed on the resulting WNN graph. For two-dimensional data visualization we performed UMAP based on the first 20 dimensions of the resulting WNN graph. The cells were clustered using the Louvain algorithm based on the first 20 dimensions with a resolution of 0.8. Cluster annotation was based on gene expression signatures (top 5 characteristic genes per cluster) and AbSeq expression (Supplementary Table 4).

For paired comparisons of matched donors, we used two−tailed paired t−tests. For comparisons between non−overlapping donor sets, we used two-tailed unpaired t-tests. For multiparameter immunophenotyping, p-values were adjusted for multiple testing using the Benjamini-Hochberg false discovery rate (FDR) method across the set of markers/subsets tested within each comparison. In all figures, n denotes the number of independent donors (biological replicates). No multiple-comparison adjustment was applied to ELISA readouts because each analyte was analyzed as an independent, pre-specified endpoint (single assay per analyte), whereas FDR correction was applied to high-dimensional flow-cytometry marker panels. Statistical significance was set at p<0.05 (*), p<0.01 (**), p<0.001 (***). Analyses were performed using GraphPad Prism 8.4.3 and RStudio 4.3.3.

We first evaluated BLN-induced killing of CD19⁺ target cells by freshly isolated γδ, CD4⁺ and CD8⁺ αβ T-cell populations. Isolated T-cell populations (γδ, CD4⁺ and CD8⁺ αβ) were co-cultured with the CD19⁺ malignant BCP ALL cell line HAL-01 at an effector-to-target ratio of 5:1 in the presence of BLN at concentration of 20 ng/ml, previously demonstrated to effectively induce immunological synapse formation (40). At day three, target CD19⁺ cells were consistently reduced in all co-cultures in the presence of BLN, as measured by assessing residual CD19⁺ cells using flow cytometry (Figure 1A, top panel shown for CD8⁺ αβ T cells). However, γδ T cells displayed lower cytolytic activity compared to CD8⁺ and CD4⁺ αβ T cells, demonstrated by persistence of CD19⁺ target cells in γδ and not in αβ T-cell cultures in the presence of BLN (Figure 1A, bottom panel). To account for donor−dependent background killing, we additionally quantified BLN−specific target elimination by normalizing each donor to its no BLN control (Supplementary Figure 3A): residual CD19⁺ cells in the no BLN condition were set to 100%, and the corresponding +BLN condition was expressed relative to this baseline (i.e., a donor-normalized fold-reduction metric). This normalization did not change the overall conclusion that CD8⁺ and CD4⁺ αβ T cells mediated stronger BLN-dependent killing than γδ T cells.

ELISA assays revealed significantly increased secretion of Interferon-γ (IFNγ) and cytotoxic effector molecule Granzyme B (GrzB), Perforin (Prf) and late mediator of cytotoxicity Granulysin (Gnly) (42). At day 3 in BLN-treated cultures, with αβ T cells releasing consistently higher levels than γδ T cells. Prf levels were highest in CD8⁺ T-cell cultures at day 3, whereas Gnly levels, assessed after seven days, were significantly elevated in both CD8⁺ and γδ T-cell cultures treated with BLN (Figure 1B).

Overall, BLN enhanced cytotoxic potential of freshly isolated γδ and αβ T-cell populations. However αβ T-cell populations demonstrated superior BLN-mediated cytolytic activity compared to γδ T cells, with CD8⁺ T cells having the potentially highest cytotoxic potential, as indirectly indicated by perforin secretion (43).

Given the known cytotoxic potential of Zole-activated Vδ2 T cells and their potential clinical application (19, 44), we next assessed BLN-dependent killing using Zole−expanded Vγ9Vδ2 γδ T-cell lines. Using the same flow-based readout (residual CD19⁺ cells), we compared their activity with PHA-expanded αβ T cells against HAL−01 and NALM−6, which differ in CD19 surface density (HAL−01 ca. 2.5−fold higher CD19 antigen density than NALM−6; Supplementary Figure 4). We studied release of cytokine mediators/effector molecules at 24 hours anticipating stronger effector functions (Figure 2B) in the presence of BLN by Zole-expanded γδ T cells (Figure 2A, green) and compared it with effector functions mediated by PHA-expanded αβ T cells (Figure 2A, grey) against two BCP-ALL cell lines, HAL-01 and NALM-6, with high and low expression CD19-patterns respectively (Supplementary Figure 4). At an effector-to-target ratio of 5:1 and BLN concentration of 20 ng/ml, both γδ and αβ activated cultures resulted in significant CD19⁺ reduction. However, PHA-expanded αβ T cells induced a slightly more pronounced B-cell killing with complete clearance of CD19⁺ cells. Target cells still persisted in Zole-expanded γδ T cells after 24 hours at low levels, particularly in a case of NALM6 (residual CD19⁺ cells at median value of 4.4% of co-culture), expressing lower levels of CD19 compared to HAL-01 (Supplementary Figure 4). Interestingly, we observed CD3 downregulation, that selectively occurred in PHA-expanded αβ T cells in the presence of BLN, whereas Zole-expanded γδ T cells maintained stable CD3 expression (Figure 2A, bottom panel with exemplary histograms). BLN-independent T-cell responses were also observed, as indicated by partial B-cell elimination in the absence of BLN occurring in a donor-dependent manner, with PHA-expanded αβ T cells (dark grey) outperforming Zole-expanded γδ T cells (dark green). BLN-dependent T-cell responses were also calculated by normalized residual CD19-cells to their counts in the absence of BLN and resulted in the same outcome (Supplementary Figure 3B).

Both expanded γδ and αβ T-cell lines exhibited increased secretion of IFNγ, GrzB, soluble FAS ligand (sFAS-L), and Prf following BLN stimulation in the presence of tumor cells. IFNγ and GrzB were lower in Zole-expanded cultures compared to PHA-expanded cultures (Figure 2B). Interestingly, high levels of sFAS-L, known to occur during activation induced cell death (AICD) expected in case of excess of T-cell activation (45, 46), was higher in the presence of PHA-expanded αβ T cells compared to Zole-expanded γδ T cells. Thus, when comparing Zole and PHA-activated T-cell cultures, αβ T-cells exhibit superior cytotoxicity; however, release higher levels of soluble FAS-L and potentially undergo CD3 downregulation in contrast to γδ T cells.

In clinical setting, response to BLN is largely dependent on ALL blast counts (8). Blinatumomab has higher affinity to CD19 and lower affinity to CD3 (47). BLN binds to CD19 and acts as activation matrix for T cells, that can detach and kill up to 5‐10 target cells within 9 hours (Serial killing) (48, 49). In the absence of CD19⁺, no T-cell activation takes places (47). Thus, CD19 target expression is relevant for BLN mediated T-cell activation and target antigen density/target cell load, as well as the concentration of BLN might impact T-cell signaling and activation (8). Thus, we next explored how varying the effector-to-target ratio and BLN concentration influenced γδ T-cell cytotoxicity. Using the HAL-01 cell line, which exhibits higher CD19 antigen density than NALM-6 (ca. 2.5−fold; Supplementary Figure 4), we cultured Zole-expanded γδ T-cell lines under low target burden (E:T 5:1; MRD-like) and high target burden (E:T 1:5; diagnosis/relapse−like) conditions (Figure 3). BLN was tested at 20 ng/ml and at concentration of 0.5 ng/ml, close to steady-state serum level reported in patients (50). Residual target (CD19⁺) cells and T-cell composition/phenotypes linked with T-cell functionality were evaluated after 24 hours and following three days by 22-color spectral flow cytometry (Supplementary Table 2).

Under low target burden (E:T 5:1; low CD19:CD3), both γδ (Zole-expanded) and αβ (PHA-expanded) cultures efficiently eliminated CD19⁺ target cells (Figure 3A, right panel). Under high target burden (E:T 1:5; high CD19:CD3, Figure 3A, left panel), αβ T-cell cytotoxicity appeared near-maximal already at 24 h and did not increase further with higher BLN, whereas γδ T-cell killing improved with 20 ng/ml BLN, particularly by day 3 (3/5 donors; Figure 3A, left panel, green bars). BLN-specific target cell elimination based on donor-normalized residual target cells yielded the same pattern (Supplementary Figure 3C).

Consistent with previous findings, marked CD3 downregulation was observed predominantly in αβ T cells upon BLN treatment, especially at high tumor burden conditions (Figures 3B, C; Supplementary Figure 5). Because CD3 down-modulation occurred only at high target burden, and spared γδ T cells while primarily affecting αβ T cells under identical BLN and staining conditions, this pattern is consistent with CD3 internalization after strong T-cell stimulation rather than epitope masking (51), supported by the condition-matched unstained controls and representative histograms (Supplementary Figure 5).

The differential CD3 modulation observed in γδ and αβ T cells suggested distinct BLN-mediated signaling through CD3, which we hypothesized would result in phenotypic and functional differences, affecting their fitness and potentially susceptibility to AICD, as supported by the high levels of sFAS-L detected in αβ T-cell cultures. Of note, sFas-L reflects ADAM10/17−mediated shedding of the pro-apoptotic membrane FasL and elevated sFasL in CD8⁺ cultures are here interpreted as a biomarker of strong re-stimulation and shedding, not as a direct driver of AICD (45, 46, 52).

To investigate this, we compared the composition and phenotype of Zole- and PHA-expanded T cells from 5 healthy donors at baseline (prior to co-culture) and analyzed their numbers and phenotypic changes at day 1 and day 3 using spectral flow cytometry (Figure 4, Supplementary Figure 6). At baseline prior co-culture (d0), after 14 days of expansion, T-cell composition differed markedly: Zole-expanded cultures consisted expectedly of Vδ2 γδ T cells (median 90%, range 80-97%), while PHA cultures comprised both CD4⁺ and CD8⁺ αβ T cells with median proportions of 14% (range 6%-31%) and 66% (range 63%-84%), respectively (Supplementary Table 1).

Zole-expanded Vδ2 γδ T cells displayed a highly homogeneous effector memory (EM) phenotype (CD27^-CD45RA^-), consistent with previous observations (39, 44). Their absolute cell counts did not significantly change during three days of co-culture across different BLN concentrations at high ALL load (Figure 4C, left top panel, black line). We observed slight decrease in absolute counts in low CD19:CD3 conditions (Figure 4C, right top panel, black line). Day 14 Zole-expanded γδ T cells expressed high levels of activation markers (CD38, HLA-DR, DNAM-1, CD56) while showing lower or absent levels of exhaustion markers such as TIGIT (baseline, d0). During co-culture there was not significant change of their phenotype compared to baseline (Figure 4A). TIGIT negativity on day 14 cultures goes in line with a reported transient peak around day 4 and declining by day 7 in Zole-expanded T-cell cultures (35).

In contrast, PHA-expanded CD8⁺ αβ T cells were more heterogeneous with naïve and memory subsets at baseline. Upon BLN stimulation, especially in the context of high tumor load, naïve cells declined and EM and TEMRA (CD27^-CD45RA⁺) subsets increased after 24 hours, with EM cells becoming dominant by day 3 (Figure 4B). This was reflected by an increase in total EM counts in the presence of high tumor load at day 3 (unadjusted p=0.019; FDR−adjusted p=0.09, trend), as single dominant population. In cultures with lower CD19⁺ burden (Figure 4C, left bottom panel), in line with less activation provided by low target cell density, naïve CD8⁺ T cells transitioned toward both EM and central memory (CM) phenotypes. Interestingly absolute counts in the presence of low target cells and lower BLN amount was higher compared to high target density and low BLN concentration. Phenotypic shifts observed in PHA-expanded CD4⁺ αβ T cells following BLN are shown as Supplementary Figure 6.

Phenotypically, PHA-expanded CD8⁺ αβ T cells showed strong activation in response to BLN, characterized by increased expression of CD25 and HLA-DR in line to observed differentiation from naïve to memory phenotypes. TIGIT expression was markedly high in αβ T cells at baseline. We also attempted to study differences in AICD at 24h or at day 3, however we did not observe differences in Annexin-V expression on T cells following BLN stimulation (readout for AICD), possibly because AICD often peaks within 4-12 h of stimulation.

Next, we asked whether BLN treatment in a mixed PBMC context could induce γδ T-cell proliferation alongside αβ T cells, despite unfavorable abundances of γδ T cells. This particular point is important in the clinical setting. To this end we stimulated PBMCs (n = 6) together with CD19⁺ HAL-01, at an E:T ratio of 1:1 (corresponding more to high CD19⁺ load when taking CD3 T-cell proportions of PBMCs into account), at an E:T ratio of 5:1 (medium CD19⁺ load) and without addition of CD19-expressing cell lines (low CD19⁺ load, autologous B-cells) for 7 days with (red) and without (grey) 20 ng/ml BLN (Figure 5A). We assessed absolute cell counts of CD4⁺, CD8⁺ and γδ T cells at day 0, day 3 and day 7 of the co-culture (Supplementary Figure 7) and assessed immune profiles of T cells. We noticed a significant expansion of CD4⁺ and CD8⁺ T cells in all three culture conditions with BLN treatment with low, medium and high CD19⁺ load. In contrast to CD4⁺ and CD8⁺ T cells, we did not see a noteworthy expansion of γδ T cells under BLN treatment from PBMCs up to day 7. As expected there was no significant increase in T-cell counts in the absence of BLN (Figure 5, grey lines). Interestingly we saw higher expansion rates of both CD4⁺ and CD8⁺ T cells in the presence of high target cell density compared to other two conditions with medium and low target cell density at day 3. At day 7, there were no differences among these conditions in CD4⁺ T cells. In CD8⁺ T cells however, the cultures with the highest CD19⁺ cells at baseline had the lowest CD8⁺ T-cell counts at day 7 compared to the other two cultures with medium and low frequencies of CD19⁺ cells (red filled mark). This was confirmed by strongest activation profile (CD25⁺, HLA-DR⁺) in the presence of high target cells (not shown), compared to other cultures, where EM CD8⁺ T cells predominated the BLN-containing cocultures on day 7, whereas the “naïve” population almost disappeared. In summary BLN application did not lead to γδ T-cell expansion in the presence of other T-cell populations, CD8⁺ T cells were the most expanded subset, as previously reported (53) and target load was inversely correlated with absolute CD8⁺ T-cell counts, suggestive of activation-associated contraction following strong activation.

To compare transcriptional profiles of αβ and γδ T cells following BLN stimulation at single-cell resolution, we performed targeted single-cell RNA sequencing of T cells from one healthy donor using the BD Rhapsody™ platform with a human T-cell panel focusing on key genes involved in T-cell responses. PBMCs from a healthy donor (containing autologous B cells) were cultured for 3 days with or without BLN (20 ng/ml). Subsequently, CD3⁺ T cells were negatively isolated for sequencing. In parallel, immunophenotyping was performed using AbSeq reagents and conventional flow cytometry to support cluster identification and validate phenotypic states (Supplementary Figures 8A-D). After quality control and removing non-viable cells, 6872 high-quality T cells were included in the analysis. Dimensionality reduction and Seurat-based clustering on UMAP space identified 14 initial clusters (Supplementary Figure 8C), which were manually curated and consolidated into eight main functional clusters (groups) based on top 5 characteristic genes (Supplementary Table 4, Figure 6C, Supplementary Figure 8). UMAP projection showed a clear separation between BLN-treated (red, 3, 576 cells) and untreated (grey 3, 296 cells) conditions, indicating major shifts in transcriptomic profiles upon BLN stimulation. UMAP projection shows minimal transcriptomic shifts in γδ T cells (predominantly Seurat clusters 6 and 13; Figures 6A, B, blue) following BLN exposure, with substantial overlap between -BLN and +BLN cells. In contrast, αβ T cells, especially CD8⁺ T cells, redistributed into BLN-enriched regions and upregulated proliferation/effector and exhaustion-associated signatures (Figures 6B, C). BLN stimulation led to increased expression of genes associated with proliferation and effector function, as well as exhaustion markers including TIGIT, TIM-3 (HAVCR2), LAG3, and CD160 (Supplementary Figure 8B). In contrast, γδ T-cell frequencies remained largely stable, consistent with the phenotypic data and aligned with in vitro, with lack of γδ T cell expansion from PBMCs at day 3 as shown in Figure 5. Of note again, also on transcript level, exhaustion marker TIGIT was not expressed on γδ T cells following BLN exposure (Supplementary Figure 8B) (54).