You are here

  • Home
  • Archive
  • Volume 12, Issue 5
  • Therapeutic Inducers of Natural Killer cell Killing (ThINKK): preclinical assessment of safety and efficacy in allogeneic hematopoietic stem cell transplant settings

Article Text

Article menu
Immune cell therapies and immune cell engineering
Original research
Therapeutic Inducers of Natural Killer cell Killing (ThINKK): preclinical assessment of safety and efficacy in allogeneic hematopoietic stem cell transplant settings
  1. Nicolas Poirier1,2,
  2. Valérie Paquin1,2,
  3. Séverine Leclerc2,
  4. Véronique Lisi2,
  5. Carolina Marmolejo2,
  6. Hicham Affia2,
  7. Paulo Cordeiro2,
  8. Yves Théorêt3,4,
  9. Elie Haddad2,5,
  10. Gregor Andelfinger2,5,
  11. Vincent Philippe Lavallée2,5,
  12. Michel Duval2,5 and
  13. Sabine Herblot2,5
  1. 1Department of Microbiology, University of Montreal, Montreal, Quebec, Canada
  2. 2Centre de recherche, Centre Hospitalier Universitaire Sainte-Justine, Montreal, Quebec, Canada
  3. 3Centre Hospitalier Universitaire Sainte-Justine, Montreal, Quebec, Canada
  4. 4Department of Pharmacology, University of Montreal, Montreal, Quebec, Canada
  5. 5Department of Pediatrics, University of Montreal, Montreal, Quebec, Canada
  1. Correspondence to Dr Sabine Herblot; sabine.herblot{at}umontreal.ca

Abstract

Background Allogeneic hematopoietic stem cell transplantation (HSCT) remains the standard of care for chemotherapy-refractory leukemia patients, but cure rates are still dismal. To prevent leukemia relapse following HSCT, we aim to improve the early graft-versus-leukemia effect mediated by natural killer (NK) cells. Our approach is based on the adoptive transfer of Therapeutic Inducers of Natural Killer cell Killing (ThINKK). ThINKK are expanded and differentiated from HSC, and exhibit blood plasmacytoid dendritic cell (pDC) features. We previously demonstrated that ThINKK stimulate NK cells and control acute lymphoblastic leukemia (ALL) development in a preclinical mouse model of HSCT for ALL. Here, we assessed the cellular identity of ThINKK and investigated their potential to activate allogeneic T cells. We finally evaluated the effect of immunosuppressive drugs on ThINKK-NK cell interaction.

Methods ThINKK cellular identity was explored using single-cell RNA sequencing and flow cytometry. Their T-cell activating potential was investigated by coculture of allogeneic T cells and antigen-presenting cells in the presence or the absence of ThINKK. A preclinical human-to-mouse xenograft model was used to evaluate the impact of ThINKK injections on graft-versus-host disease (GvHD). Finally, the effect of immunosuppressive drugs on ThINKK-induced NK cell cytotoxicity against ALL cells was tested.

Results The large majority of ThINKK shared the key characteristics of canonical blood pDC, including potent type-I interferon (IFN) production following Toll-like receptor stimulation. A minor subset expressed some, although not all, markers of other dendritic cell populations. Importantly, while ThINKK were not killed by allogeneic T or NK cells, they did not increase T cell proliferation induced by antigen-presenting cells nor worsened GvHD in vivo. Finally, tacrolimus, sirolimus or mycophenolate did not decrease ThINKK-induced NK cell activation and cytotoxicity.

Conclusion Our results indicate that ThINKK are type I IFN producing cells with low T cell activation capacity. Therefore, ThINKK adoptive immunotherapy is not expected to increase the risk of GvHD after allogeneic HSCT. Furthermore, our data predict that the use of tacrolimus, sirolimus or mycophenolate as anti-GvHD prophylaxis regimen will not decrease ThINKK therapeutic efficacy. Collectively, these preclinical data support the testing of ThINKK immunotherapy in a phase I clinical trial.

  • Dendritic Cells
  • Hematologic Neoplasms
  • Immunity, Innate
  • Immunotherapy, Adoptive
  • Pediatrics

Data availability statement

Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/jitc-2023-008435

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • Hematopoietic stem cell transplantation (HSCT) is the standard of care for patients with high-risk hematological malignancies, although cure rates are still dismal. Harnessing the graft-versus-leukemia (GvL) effect of HSCT without increasing allo-immune reactions against normal cells has been a topic of intense research to reduce the risk of cancer relapse without worsening the graft-versus-host disease (GvHD).

    WHAT THIS STUDY ADDS

    • Our study, combined with our previous work, demonstrates that the adoptive transfer of Therapeutic Inducers of Natural Killer cell Killing (ThINKK) may improve the early GvL effect mediated by innate immune effectors without increasing the risk of severe GvHD mediated by donor derived T cells. Furthermore, ThINKK therapy is compatible with most GvHD prophylaxis regimen given to patients receiving allo-HSCT.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • These results open the way to a new class of post-transplant natural killer cell-based immunotherapy for patients with incurable leukemia or even solid cancers such as neuroblastoma.

    Introduction

    Acute lymphoblastic leukemia (ALL) is the most common cancer in children under the age of 15 and still one of the leading causes of death by childhood cancer, since 15%–20% of ALL patients relapse after chemotherapy and about 40% of them die from their disease.1 2 For patients with chemotherapy-refractory ALL, immunotherapy is the only hope of cure, as exemplified by the development of chimeric antigen receptor (CAR) T-cell therapy. Nonetheless, CAR T-cell therapy is available only for CD19+ leukemia, and cures only half of the patients.3 Allogeneic hematopoietic stem cell transplantation (HSCT) was the first established immunotherapy for leukemia.4 The eradication of the residual leukemia cells is achieved through their recognition and killing by donor-derived immune effectors, a process known as the graft-versus-leukemia (GvL) effect. Reinforcing the GvL effect is therefore a promising avenue to reduce the risk of relapse and improve the outcome of patients transplanted for high-risk leukemia.5

    Natural killer (NK) cells are the first immune cells to reconstitute after HSCT. Therefore, they play a major role in early GvL effect, when the leukemia burden is at its lowest and hence more amenable to immunotherapy.6–10 Novel strategies to increase NK cell cytotoxicity after allogeneic HSCT are though urgently needed. Activating cytokines such as interleukin (IL)-2 or IL-15 increase NK cell anticancer functions, but their use in transplanted patients is prohibited due to severe side effects or leukemia stimulation.11 12 On the opposite, type-I interferons (IFN) treatment has been shown to be safe after allogeneic HSCT but has demonstrated limited efficacy to prevent relapse.13

    Plasmacytoid dendritic cells (pDC) are the natural stimulators of NK cells. These specialized immune cells have been defined as type-I IFN producing cells with reduced capacity to trigger T cell activation, in sharp contrast with both subsets of classical DCs (cDCs), -DC1 and -DC2.14 15 pDC functional features are appealing for post-HSCT immunotherapy, as they would lead to increase the NK cell-mediated GvL effect without increasing the risk of T-cell mediated life-threatening graft-versus-host disease (GvHD).16 Indeed, whereas T cells induce both GvL and GvHD, harnessing NK cell cytolytic activity against cancer cells has not been associated with an increased risk of GvHD.17 18 In some clinical settings, activating NK cells even decreases GvHD incidence and severity by eliminating the antigen-presenting cells responsible for T cell activation.19–21 We showed, however, that pDC reconstitution and functions are delayed after allogeneic HSCT, precluding their activation with chemical compounds.22 Furthermore, pDC are so rare in blood that one cannot expect to harvest sufficient amounts from one donor to treat a patient. We thus designed a method to produce large numbers of pDC surrogates from CD34+ cord blood progenitors.22 23 We named these cells Therapeutic Inducers of Natural Killer cell Killing (ThINKK). We showed that ThINKK-stimulated NK cells efficiently kill childhood ALL cells both in vitro and in vivo. Importantly, weekly adoptive transfers of ThINKK prevented ALL development in a humanized preclinical model of HSCT for ALL, whereas 90% of control mice died from human leukemia.23

    Recent data underscore the heterogeneity and plasticity of pDC, therefore, the absence of T-cell activating properties remains to be documented for ThINKK. Indeed, recent unbiased single-cell RNA sequencing and cytometry by time-of-flight (CyTOF) experiments have demonstrated that the cell population previously identified as pDC encompasses two different subsets: canonical type-I IFN-producing AXL− pDC with low T cell activation capacity, and non-canonical AXL+SIGLEC6+ DC, able to give rise to classical DC and to activate T cells.24 Both subsets share the expression of E2-2 (TCF-4) but differ by their expression of genes such as BDCA-2, JCHAIN, IRF7, GZMB, MZB1 for canonical pDC, and AXL, ID2, CX3CR1 and SIGLEC6 for non-canonical AXL+DC.

    To prevent GvHD, transplanted patients receive a prophylactic regimen of immunosuppressive drugs targeting T cell proliferation.25 Calcineurin inhibitors such as cyclosporine A and tacrolimus, mTOR inhibitors such as sirolimus (rapamycine) and the inosine monophosphate dehydrogenase blocker mycophenolic acid are the most frequently used.26 Corticosteroids such as methylprednisolone can be used preventively or more often to treat overt GvHD. These drugs have low impact on NK cell activation and function. However, their influence on the ThINKK/NK cell axis remains to be established.27 28

    The safety of ThINKK immunotherapy depends on the risk of exacerbating GvHD, whereas its efficacy depends on the potential downregulation of its activity by the anti-GvHD prophylaxis regimen. Here we show that ThINKK share the key phenotypic and functional profile of naturally occurring canonical blood pDC. ThINKK do not induce T cell activation and neither increase allogeneic T cell proliferation in vitro nor in a mouse model of GvHD. We further define the GvHD prophylaxis regimen that could be used in transplanted patients receiving ThINKK immunotherapy without impairing its efficacy.

    Materials and methods

    Peripheral blood samples

    Peripheral blood samples were obtained from healthy volunteers after written informed consent was obtained in accordance with the Declaration of Helsinki and CHU Sainte Justine IRB approval. Peripheral blood mononuclear cells (PBMC) were isolated by density gradient centrifugation using Ficoll-Paque Plus (GE Healthcare Bio-Science AB, Uppsala, Sweden). Human PBMC were used for T, NK and CD14+ cell selections as well as for the induction of xeno-GvHD in NSG mice.

    ThINKK expansion and differentiation from cord blood CD34+ cells

    ThINKK were expanded and differentiated from purified cord blood CD34+ as previously described.29 Briefly, cord blood units were obtained from the CHU Sainte-Justine Research Center cord blood bank with the approval from the Institutional Review Board. Mononuclear cells were isolated by gradient centrifugation on Ficoll Paque and CD34+ cells were purified using positive selection kit from Miltenyi according to manufacturer’s instructions (Miltenyi Biotec, San Diego, CA). CD34+ were seeded in culture for 2 weeks in Stem Span serum free medium (StemCell Technologies, Vancouver, BC) containing cytokines (Miltenyi Biotec), and StemRegenin1 (Sigma-Aldrich, St-Louis, MO). ThINKK were then sorted by flow cytometry according to their HLA-DR and CD123 surface expression. Sorting according to AXL expression was used in some experiments. Purity was checked after each sort and was greater than 98%.

    Single-cell RNA sequencing and data analysis

    ThINKK cells were sorted by flow cytometry according to their expression of HLA-DR and CD123, viability and purity was assessed by flow cytometry and were ≥98% (online supplemental figure S1). We used the 10X Genomics platform to sequence 10,000-isolated ThINKK from two independent cultures of cord blood progenitors according to manufacturer’s instructions. cDNA libraries and RNA sequencing were performed according to manufacturer’s protocol (Chromium Single Cell 3′ reagent kit v2). The libraries were quantified using Agilent Bioanalyser 2100. Next generation sequencing was performed on Illumina HiSeq 2500 system (Illumina, San Diego, CA) at CHU Ste-Justine research center aiming at 500M reads per library.

    Supplemental material

    Data demultiplexing was performed using the software CellRanger V.3.0.1. The CellRanger filtered expression matrix was analyzed using the Seurat R package V.3. Cells expressing less than 200 genes and genes expressed in less than five cells were filtered out as likely empty droplets. Droplets containing more than one cell were identified as those were fulfilling any of the following: the percentage of reads originating from the mitochondrial genome was above 2.5 SD above the mean; the number of expressed genes was above 2 SD above the mean; the number of transcripts detected was above 2 SD above the mean. Cell cycle phase was assessed using Seurat CellcycleScoring function using the Nestorawa gene lists. After filtering, data normalization using a log transformed normalization was performed, the top 2000 variable features identified and data scaling performed without regressing any variables. Principal component analysis (PCA) dimensionality reduction was performed using Seurat RunPCA function. Dimension selection was decided after visual inspection of the scree plot. Further t-Distributed Stochastic Neighbor Embedding/Uniform Manifold Approximation and Projection (UMAP) dimensionality reduction was performed using Seurat RunTSNE/RunUMAP function for 2D visualization purposes. A Shared Nearest Neighbor graph using the Seurat FindNeighbors function and cell clustering was performed using Seurat FindClusters function.

    The individually filtered samples were then integrated together using the reciprocal PCA approach from Seurat. Integrated data was scaled, and embedded in two dimensions as described above. Visual inspection confirmed that cell clusters found in the original individual dataset were conserved in the integrated dataset.

    In vitro differentiation of monocyte-derived dendritic cells

    CD14+ cells were isolated from human PBMC using magnetic beads (EasySep Human Monocyte Isolation Kit, StemCell Technologies). Cells were then cultured at a concentration of 0.5×106 cell/mL for 7 days in RPMI 10% FBS containing 500 UI/mL Granulocyte-Macrophage Colony-Stimulating Factor, 100 UI/mL IL-4 and streptamycin, penicillin and fungizone (Fisher Scientific, CA, USA). Cultures were refreshed every 2 days and, on day 5, 1 µg/mL LPS was added. Purity and differentiation state were assessed by flow cytometry at day 6 using CD14 and CD1a markers.

    Proliferation assay

    T cells were negatively selected from human PBMC using magnetic bead selection kit (EasySep Human T Cell Enrichment Kit, StemCell Technologies). Purified T cells were labeled using 5 µM of carboxyfluorescein succinimidyl ester (CFSE; Life Technologies by Thermo Fisher Scientific). Labeled T cells were cocultured with in vitro differentiated monocyte-derived dendritic cell (moDC) and/or ThINKK. On day 6, T cells were stained for CD4 and CD8. CFSE dye dilution was assessed by flow cytometry using Canto or Fortessa cytometers (BD Biosciences, San Jose, CA). Cytometry data were further analyzed using the FlowJo proliferation platform (BD Biosciences). Starting from the CFSE dilution histograms, this tool quantifies cell divisions overtime and generates statistics. The percentages of dividing cells represent the proportion of initial population that underwent at least one division (also called the precursor frequency).30

    Xeno-GvHD in NSG mice

    NOD-scid IL2Rgammanull (NSG) mice were obtained from the Jackson Laboratory. Mice were bred in the CHU Ste-Justine research center animal care facility according to our in-house animal care ethic committee. A moderate GvHD was induced by injecting 106 human PBMC in 7–9 weeks old irradiated (1 Gy) NSG mice as previously described.31 Control mice were intravenously injected weekly with saline solution, while ThINKK treated mice received 105 ThINKK weekly for 5 weeks. A positive control group received daily subcutaneous injections of human 10,000 UI IL-2 (Miltenyi Biotec) for 2 weeks. Four independent experiments were performed with 12, 12, 14 and 16 animals, respectively for a total of 19 animals in the ThINKK-treated group, 26 animals in the control group and seven animals in the IL-2-treated group. Two mice were excluded during the study due to irradiation side effects unrelated to GvHD. Randomization of animals was used to mix male and female in each group and mix group in each cage. Sample size calculation was performed using G Power software assuming a size effect of 20 days and an SD of 30 days (primary outcome was survival). Clinical signs of GvHD were blindly evaluated three times a week using a GvHD assessment scale including scores for weight, posture, activity, fur texture and skin integrity.32 Mice were sacrificed when the clinical endpoints were reached (GvHD score of 5 or higher). Blood from each mouse was sampled weekly. After centrifugation plasma samples were stored at −80°C for further IFN-γ dosage. Blood cells were resuspended in red blood cell lysis buffer (BD Pharm Lyse, BD Biosciences) and washed in PBS containing 2 mM EDTA and 0.5% bovine serum albumin. Nucleated cells were stained using conjugated antibodies against mouse CD45, human CD45 and human CD3. Dead cells were excluded using 7AAD vital dye and 5000 counting beads (CountBright, Thermo Fisher Scientific) were added to each sample. Flow cytometry acquisition was performed on FACS Canto cytometer with a stopping gate of 2500 Count Bright beads. FlowJo software was used for data analysis (BD Biosciences). CD3 counts per µL of blood were calculated using the following formula: [(CD45+CD3+ absolute number/5000)×2500]/blood volume (70 µL)].

    IFN-α intracellular staining

    ThINKK were cultured for 4 hours in the presence or the absence of a TLR9 agonist, CpG oligonucleotide 2216 (ODN 2216, 10 µg/mL, Invivogen). Golgi Stop (BD Biosciences) was then added and cells were further incubated at 37°C, 5% CO2 for 10 hours. Unstimulated and stimulated cells were harvested, washed once in PBS and fixed/permeabilized using the Cytofix/Cytoperm Plus Fixation/Permeabilization Kit from BD Biosiences. Intracellular IFN-α was stained using a FITC-conjugated anti-human IFN-α specific antibody (Miltenyi Biotec). Samples were analyzed by flow cytometry on Canto cytometer and FlowJo software was used for data analysis (BD Biosciences).

    NK cell cytotoxicity assays

    NK cells were then negatively selected from human PBMC samples using the EasySep Human NK Cell Enrichment Kit (Stemcell Technologies). Isolated NK cells were cocultured with CpG-activated ThINKK at a ratio of 10:1 in RPMI medium supplemented with 10% FBS for 18–20 hours (overnight). Sub-therapeutic, therapeutic and supra-therapeutic concentrations of immunosuppressive drugs were added in the culture medium during the overnight NK:ThINKK coculture. We tested cyclosporine A (166.6, 500.0 and 1500.0 ng/mL), tacrolimus (6.6, 20.0 and 60.0 ng/mL), sirolimus (5.0, 15.0 and 45.0 ng/mL) and mycophenolic acid (1000.0, 3000.0 and 9000.0 ng/mL). Methylprednisolone was also used as a positive control at therapeutic concentration (500 ng/mL). Drug concentrations in culture supernatant were verified by MS/HPLC (cyclosporine A, tacrolimus and sirolimus) or MS/MS.26 Unstimulated NK cells and NK:ThINKK cocultures were then harvested and incubated for 2 hours with target cells, that is, GFP expressing REH cells, at 10:1, 5:1 and 1:1 effector to target (E:T) ratios in triplicates. Specific lysis of target cells was calculated by quantification of live GFP+ cells excluding dead cells using propidium iodide dye. Samples were analyzed by flow cytometry on a Fortessa cytometer (BD Biosciences). FlowJo software was used for data analysis.

    Luminex assay

    NK cell/ThINKK coculture supernatants were collected and stored at −80°C until use. The concentrations of IFN-α, IFN-γ, IL-6 and IL-8 were assessed using MILLIPLEX Human/Chemokine/Growth Factor Panel A Luminex assay from Millipore Sigma-Aldrich according to the supplier’s instructions.

    Statistics

    Two-way analysis of variance tests or Kruskal-Wallis non-parametric test were used for multiple group comparisons of paired data, and paired t-tests were used for single data comparisons. The log-rank test was used to compare survival curves. A value of p<0.05 (*) was considered significant with a CI of 99% (GraphPad Software, San Diego, CA, USA).

    Results

    ThINKK display features of canonical blood pDC

    ThINKK are identified as CD123 and HLA-DR expressing cells following in vitro expansion and differentiation of cord blood CD34+ progenitors. To establish the cellular identity of ThINKK at the single-cell level, we used the 10X Genomics platform. Two independent cord blood units were used to expand and select ThINKK (online supplemental figure S1).23 10,000 cells per sample were sequenced and quality control revealed optimal library assembly and sequencing with median gene counts per cell over 1800 in each sample (online supplemental table S1). We integrated the two samples using a reciprocal PCA approach and confirmed, using visual inspection, that the integration recapitulated the initial individual clustering. Clustering the data with a small number of clusters identified four clusters (figure 1A,B). Based on the differential gene expression analysis and similarity with previously published DC datasets we annotated each cluster according to their transcriptional profiles (online supplemental table S2).14 15 Cluster 0 represented 68% of total cells and expressed the gene signature of canonical pDC: BDCA2, CD123, GZMB, JCHAIN, IRF-7, MZB1, E2-2 (TCF4) (figure 1C). Cluster 1 (20% of the cells) expressed canonical pDC genes such as E2-2, IRF7, JCHAIN, but lower levels of GZMB and BDCA2 and higher levels of ID2 genes. They also expressed genes associated with cell cycle (figure 1A, online supplemental table S2). Therefore, cells in cluster 1 have the profile of cycling canonical pDC. Cells in cluster 2 (11%) expressed genes associated with AXL+ non canonical DC, such as AXL, CX3CR1, CD141 and CD33 but, as opposed to natural AXL+ DC, they expressed low levels of SIGLEC6. Furthermore, cells in cluster 2 expressed some genes associated with the cDC2 subtype such as CD1c and CLEC10A (figure 1C, online supplemental table S2).14 Finally, cells in cluster 3 (1.2%) expressed genes associated with lymphoid (CD79+) and myeloid (MPO+) lineages. Cluster 3 was thus marginal and not expected to play a role in ThINKK function. ThINKK are therefore a relatively homogenous population with strong homology with canonical pDC (cluster 0) and probably their immediate precursors (cluster 1). The AXL+ subset exhibits a mixed expression profile that differs from blood DC subsets, including transitional or non-canonical DC described by others.14 33 34

    Supplemental material

    Figure 1
    Figure 1

    sc-RNA profiling of HLA-DR+ CD123+ ThINKK expanded from CD34+ cord blood progenitors. (A) sc-RNA datasets obtained from two ThINKK samples were integrated and unbiased clustering was performed. RNA clusters (left) and cell cycle phase scores (right) are represented as t-SNE plots embedding 18,915 cells in which each cell is a dot. (B) Relative proportions of clusters within each sample and in the integrated dataset are presented. (C) The expression of indicated genes are presented as t-SNE plots. Colors indicate the level of expression in each cell. (D) t-SNE after cycling cells and contaminants removal. The first principal component of the PCA reduction orders the cells on the canonical versus non-canonical pDC axis. Genes abundantly expressed in canonical pDC are identified as those with high correlation with PC1 (in red in the waterfall graph) whereas genes expressed in non-canonical pDC have a low correlation with PC1 (in green in the waterfall graph). PCA, principal component analysis; pDC, plasmacytoid dendritic cell; ThINKK, Therapeutic Inducers of Natural Killer cell Killing; t-SNE, t-Distributed Stochastic Neighbor Embedding.

    To better characterize the differences between clusters 0 and 2, we removed clusters 1 and 3. On this reduced dataset, we recomputed the principal components and observed that the first principal component (PC1) captured the differences between cluster 0 (high PC1 values) and cluster 2 (low PC1 values—figure 1D). To identify genes whose expression varies in this same gradient, we computed the correlation of each gene with the PC1 value. All genes associated with canonical pDC displayed a strong positive correlation with PC1, whereas genes associated with other natural DC subsets negatively correlate with PC1.

    We then confirmed gene expression results at the protein level using flow cytometry. As expected, most ThINKK expressed the specific marker for canonical pDC BDCA2 (figure 2A). AXL surface expression varied from one ThINKK sample to another (quartile 3.03%–15.1%, median 6.03%, figure 2B,C). Although mutually exclusive in normal cells, DC1 specific marker CD141 and specific DC2 marker CLEC10A were expressed together in about half of AXL+ cells, in good agreement with scRNA sequencing data (figure 2D).

    Figure 2
    Figure 2

    Flow cytometry analysis of ThINKK phenotype. The proportions of ThINKK expressing (A) BDCA2 pDC specific marker, (B–C) AXL tyrosine kinase receptor and (D) DC markers (CD11c, CD141, CLEC10A) were determined by flow cytometry. Representative histograms are displayed. The doted lines represent the negative control (unstained cells) while the dark lines represent the stained sample. (C) The mean±SD of positive ThINKK for indicated markers are presented (n=5 independent experiments). (E) Unstimulated (unst. doted line) and CpG-stimulated (CpG st. black line) ThINKK were stained for intracellular IFN-α. A representative histogram is presented. (F) The phenotype of unstimulated (unst.) and CpG-stimulated (CpG) ThINKK was assessed using PDL-1 and CD80 markers. Representative contour plots and percentage of P1 (PDL1+CD80−), P2 (PDL1+CD80+) and P3 (PDL1−CD80+) are displayed (n=5 independent experiments). IFN, interferon; ThINKK, Therapeutic Inducers of Natural Killer cell Killing.

    As type-I IFN production on TLR stimulation is the hallmark of blood pDC, we assessed IFN-α production by intracellular staining and showed that more than 80% of ThINKK secrete IFN-α in response to TLR-9 agonist (figure 2E). Alculumbre et al have demonstrated that TLR stimulation triggers blood pDC diversification in three stable subpopulations: PD-L1+CD80− IFN-producing cells, PD-L1−CD80+ T-cell activating cells and intermediate PD-L1+CD80+ cells with both functions.35 We thus stained ThINKK surface with anti-PDL1 and anti-CD80 antibodies after TLR-9 activation. Most activated ThINKK expressed PD-L1 but not CD80, a minority expressed an intermediate PDL1+CD80+ phenotype, whereas none expressed the activating T cell PDL1−CD80+ phenotype (figure 2F).

    Collectively, single-cell RNA sequencing and flow cytometry analysis indicate that ThINKK display a high phenotypic and functional homology with canonical blood pDC, while the AXL+CD141+CLEC10A+ minor subset exhibit a mixed phenotype.

    ThINKK do not increase allogeneic T cell proliferation

    Although phenotypic studies at the RNA and protein levels showed that most ThINKK were analogs to canonical blood pDC, this does not completely exclude that ThINKK might activate T cells or increase allogeneic T cell activation. We therefore assessed ThINKK capacity to induce T cell proliferation or to exacerbate allogeneic T cell proliferation. We cocultured CFSE-labeled T cells with bonafide allogeneic antigen-presenting cells, that is, moDC, in the presence or absence of ThINKK. Peripheral blood T cells were isolated from healthy volunteers, moDC were in vitro differentiated from CD14+ peripheral blood monocytes and ThINKK were differentiated from cord blood CD34+ cells.23 36 CD3-CD28 T cell stimulation was used a positive control. We first assessed whether allogeneic T cells are able to kill ThINKK when cocultured. We observed that the percentage of recovery following overnight ThINKK:T cocultures was 90±8% and the percentage of Annexin V+ThINKK was 3.8±1.2% (mean±SD). This result was similar for ThINKK:NK coculture (91±13% recovery and 8.4±13% Annexin V+ThINKK) (online supplemental figure S2). As anticipated, moDC induced the proliferation of 6.8±3.1% (mean±SD) and 3.1±0.88% of CD4 and CD8 allogeneic T cell respectively, corresponding to the expected frequency of allogeneic T cells against another individual. On the opposite, ThINKK induced the proliferation of only 1.3±0.3% and 1.2±0.47% of allogeneic CD4 and CD8 T cells, respectively (figure 3A,B). Although not statistically significant, the percentage of T cells proliferating in response to moDC was slightly decreased by the addition of ThINKK (4.7±2.45% vs 6.8±3.1% for CD4 T cells and 2.5±0.62% vs 3.1±0.88% for CD8 T cells) (figure 3A,B).

    Supplemental material

    Figure 3
    Figure 3

    ThINKK do not increase allogeneic T cell proliferation. Mixed lymphocyte reactions were performed by CFSE dilution assays. CFSE-stained purified T cells were cultured with monocyte-derived DC (moDC) in the presence or the absence of ThINKK, AXL+ ThINKK or AXL− ThINKK. (A) Representative histograms of CFSE dilution assays are displayed for CD4 and CD8 T cells with proliferation modeling profiles (FlowJo Software). The dark lines represent the raw data of CFSE dilution assays, the colored histograms represent the model of cell division according to a peak ratio of 0.5 (purple peaks represent undivided cells, orange peaks with darker overlaps represent cells that underwent 1–7 divisions). (B–C) Means of percentage of dividing CD4 and CD8 cells are showed with SD for the indicated culture conditions (n=5 independent experiments). The percentages of dividing cells represent the proportions of the initial population that underwent at least one division (precursor frequency) and are not equal to the percentages of cells in the final culture that divided at least once (% CFSElow as presented in panel A). Tukey’s multiple comparison test was used to compare matched culture conditions, *p≤0.05, **p≤0.01, ***p≤0.0001. CFSE, carboxyfluorescein succinimidyl ester; ThINKK, Therapeutic Inducers of Natural Killer cell Killing.

    We further investigated the potential of the minor AXL+CD141+CLEC10A+ subset to induce in vitro T cell proliferation. We sorted AXL+ and AXL− ThINKK subsets using flow cytometry and cocultured them with allogeneic T cells in the presence or the absence of moDC. We observed that AXL− ThINKK did not induce allogeneic T cell proliferation, while AXL+ ThINKK induced the proliferation of 2.3±0.8% and 1±1% of allogeneic CD4 and CD8 T cells, respectively. Therefore, ThINKK T-cell activating capacity is low and resides mostly in the AXL+CD141+CLEC10A+ minor subset. Accordingly, AXL+ ThINKK express slightly higher levels if HLA class I molecules as compared with AXL− ThINKK (online supplemental figure S3A,B). In response to TLR stimulation, AXL+ are more prone to differentiate in P2 subset (PDL1+CD80+ intermediate population) than AXL− subset, although this phenotype was not observed in all samples (online supplemental figure S3B). Most importantly, neither AXL− nor AXL+ ThINKK were able to increase CD4 or CD8 T cell proliferation induced by moDC (figure 3C).

    Supplemental material

    Collectively, these results indicate that although ThINKK had low T-cell activation capacities confined to their minor AXL+CD141+CLEC10A+ subset, they do not increase T cell proliferation induced by allogeneic antigen-presenting cells. ThINKK adoptive therapy is thus not expected to increase the risk of GvHD after allogeneic HSCT.

    ThINKK injections do not exacerbate in vivo xeno-GvHD

    To confirm that ThINKK do not increase the risk of GvHD, we used an in vivo human-to-mouse model of xeno-GvHD. As described previously, the injection of human blood mononuclear cells in immune-deficient NSG mice induces a xeno-GvHD characterized by the proliferation of human CD3+ T cells recognizing mouse antigens in the context of murine major histocompatibility complex, the increase of blood human IFN-γ, and the clinical signs of GvHD (skin alteration, decreased activity, loss of weight).31 We injected 106 human PBMC in NSG mice followed by 5-weekly injections of 105 ThINKK, saline solution or human IL-2 as a positive control. ThINKK were in vitro stimulated with a TLR-9 ligand for 4 hours and then washed before the injection. As we showed that neither AXL− nor AXL+ subsets increased moDC-induced T-cell activation, ThINKK were not selected based on AXL expression for these in vivo experiments. Mice were evaluated in a blinded fashion every other day using a GvHD assessment scale and weekly blood analysis.32 Mice were sacrificed when GvHD clinical scores reached 5–7 out of 10. We did not observe significant difference in blood CD3+ T cell counts and IFN-γ serum concentrations between the ThINKK-treated and saline-treated mice (figure 4A,B). On the opposite, human CD3 counts were significantly higher in IL-2 injected animals at week 2 and 3 (online supplemental figure S4A). GvHD developed with the same kinetic in ThINKK treated mice (n=19) and in controls (n=19) with median survival of 59 and 66 days, respectively and no significant difference in survival curves (p=0.09, figure 4C), while the median survival of IL-2-treated mice was significantly lower (16 vs 25 days, online supplemental figure S4B). These results corroborate our in vitro data and demonstrate that ThINKK are not expected to exacerbate GvHD in transplanted patients.

    Supplemental material

    Figure 4
    Figure 4

    ThINKK injections do not exacerbate in vivo GvHD. 106 human peripheral blood mononuclear cells were injected in NSG mice to induce a moderate xeno-GvHD triggered by functional human T cells. Control mice received weekly injections of saline solution (saline-treated group) while treated mice received five weekly injections of 105 ThINKK (ThINKK-treated group). (A) Human CD3+ cell counts were monitored by flow cytometry on weekly blood samples. (B) Serum human IFN-γ was measured by ELISA on blood samples. (C) Kaplan-Maier survival curves were established according to GvHD clinical scores assessed three times a week. Mice were sacrificed when critical endpoints were reached (GvHD score ≥5–7). Log-rank test was used to compare the two groups (p=0.0945). IFN, interferon; GvHD, graft-versus-host disease; ThINKK, Therapeutic Inducers of Natural Killer cell Killing.

    Tacrolimus, sirolimus and mycophenolate do not decrease ThINKK-induced NK cell cytotoxicity against childhood ALL cells

    In allogeneic transplantation settings, patients routinely receive a GvHD prophylactic regimen of immunosuppressive drugs to suppress allogeneic T cell activation and proliferation. We tested the functions of the ThINKK/NK cell axis in the presence of the immunosuppressive drugs most used after HSCT. ThINKK were activated for 4 hours with a TLR-9 ligand. Activated ThINKK were then cocultured with peripheral blood NK cells in the presence or the absence of cyclosporine A, tacrolimus, sirolimus, mycophenolate or methylprednisolone. We tested sub-therapeutic (C1), therapeutic (C2) and supra-therapeutic (C3) doses for each drug, corresponding to usual residual, intermediate and peak plasmatic concentrations in patients. We confirmed the in vitro dosages of each drug using MS/HPLC or MS/MS dosages in culture supernatants. As a positive control, we verified that these immunosuppressive drugs inhibited in vitro CD3-CD28-induced polyclonal T cell proliferation, even at sub-therapeutic dosage (online supplemental figure S5). We first assessed ThINKK viability by flow cytometry using vital 7AAD dye. Only methylprednisolone decreased ThINKK viability (figure 5A). In addition, cyclosporine, tacrolimus, sirolimus and mycophenolate did not decrease the production of IFN-α, IL-6 and IL-8, as assessed by multiplex Luminex dosages of these cytokines in the supernatants of ThINKK/NK cell cocultures (figure 5B). Thus, ThINKK remain viable and functional in an immunosuppressive environment.

    Supplemental material

    Figure 5
    Figure 5

    Tacrolimus, sirolimus and mycophenolate do not impair ThINKK-induced NK cell cytotoxicity against ALL. ThINKK were cultured in the presence of vehicle (Veh), methylprednisolone (mPred), cyclosporine A (CsA), tacrolimus (Tacro), sirolimus (Siro) or mycophenolic acid (MPA) at three concentrations corresponding to residual (C1), therapeutic (C2) and peak plasmatic (C3) concentrations in patients. (A) After overnight incubation, live ThINKK cells were numerated by flow cytometry using 7ADD vital dye. Percentages of viable cells were calculated as compared with ThINKK in vehicle. Averages of three experiments are presented with SD. (B) ThINKK-secreted cytokines (IFN-α, IL-6 and IL-8) were measured by multiplex Luminex in ThINKK-NK cell coculture supernatants following overnight stimulation with CpG oligonucleotide (CpG st.) and in controls (unst.). The Kruskal-Wallis non-parametric multiple comparison test was used to compare stimulated conditions in the presence or absence of drugs. (C) NK cell cytotoxic assays were performed against REH ALL cells and specific lysis was measured by flow cytometry. Sorted blood NK cells were cocultured overnight in the presence (right column, ThINKK) or in the absence of ThINKK (left column, Unst.). Methylprednisolone, cyclosporine, tacrolimus, sirolimus or mycophenolate or vehicle were added in overnight NK-ThINKK cocultures at three indicated concentrations. The graphs represent the average specific lysis for three independent experiments with SD for CsA, Tacro, Siro and MPA and 13 experiments for mPred used as positive control in each experiment. Two-way analysis of variance test was used to compare the results in the cytotoxic activity in the presence or the absence of drugs. (D) IFN-γ was measured by multiplex Luminex in NK cell culture supernatants following overnight culture in the absence (unst.) or in the presence of ThINKK and CpG oligonucleotide. *p<0.05; ****p<0.0001. ALL, acute lymphoblastic leukemia; IFN, interferon; IL, interleukin; NK, natural killer; ThINKK, Therapeutic Inducers of Natural Killer cell Killing.

    We then assessed the NK cell response to ThINKK stimulation in the presence of immunosuppressive drugs. Since ThINKK stimulation do not induce NK cell proliferation, we could not use CSFE dilution assays as for T cells, so we assessed the impact of immunosuppressive drugs on NK cell functions. We first monitored TRAIL expression by NK cells, as we previously showed that very high TRAIL expression is not only the hallmark of their activation by ThINKK but also plays a major role in ALL killing.29 TRAIL upregulation in response to ThINKK stimulation was decreased by half on NK cell surface in the presence of cyclosporine and methylprednisolone but was not affected by the other drugs (online supplemental figure S5). We performed cytotoxic assays against REH childhood ALL cell line using NK cells stimulated by ThINKK in the presence or absence of immunosuppressive drugs. We showed that tacrolimus, sirolimus and mycophenolate did not decrease NK cell cytotoxicity against ALL cells, while cyclosporine and methylprednisolone decreased ALL lysis by ThINKK-stimulated NK cells (figure 5C). Finally, methylprednisolone but none of the other immunosuppressive drugs decreased the production of IFN-γ by ThINKK-stimulated NK cells (figure 5D). Collectively, these data indicate that the ThINKK/NK cell axis is functional in the presence of tacrolimus, sirolimus and mycophenolate. However, the prophylactic anti-GvHD regimen should not include cyclosporine or corticosteroids in transplanted patients receiving ThINKK therapy.

    Discussion

    Our results indicate that ThINKK immunotherapy is expected to be safe and effective in transplanted patients, even in allogeneic settings. Indeed, our data showed that ThINKK display the phenotypic and functional properties of canonical blood pDC, including type-I IFN production on TLR stimulation, and do not induce T cell proliferation in vitro or worsen GvHD in vivo. We nonetheless identified an AXL expressing subset displaying a mixed phenotype with partial similarity to blood AXL+ DC described by others. Following TLR9 stimulation, these AXL+ ThINKK modestly activated T cells but did not increase allogeneic T cell proliferation induced by antigen-presenting cells and neither worsened GvHD in a preclinical mouse model. We further demonstrate that ThINKK immunotherapy is compatible with anti-GvHD prophylaxis since we showed that tacrolimus, sirolimus and mycophenolate do not decrease the anti-leukemic effect of ThINKK-stimulated NK cells.

    Harnessing innate immune responses is an attractive approach to reinforce the early GvL effect of HSCT and prevent leukemia relapse, without inducing adverse GvHD. Indeed, reinforcing the anticancer properties of the innate immune lymphocytes, that is, the NK cells, does not induce nor exacerbate the GvHD mediated by allogeneic donor T cells. In contrast, activated NK cells have been shown to protect against GvHD in some clinical settings.19 37 We previously demonstrated that adoptive transfers of ThINKK stimulate NK cell lytic activity and control leukemia development in a preclinical humanized model of HSCT for ALL.23 29 However, pDC defective reconstitution following HSCT forbids any attempt to activate directly these cells with a TLR ligand. In addition, their rarity in blood precludes their harvest from healthy donors for adoptive transfers.22 We therefore developed a method for the in vitro expansion and differentiation of pDC analogs from cord blood progenitors. These cells are called ThINKK for Therapeutic Inducers of NK cell Killing. We have previously showed that blood TLR9-stimulated pDC and ThINKK share similar NK cell activation properties.23 29 However, ThINKK cellular identity and safety profile needed to be established before the first-in-human clinical trial. Our data demonstrated that ThINKK are homogeneous and mainly composed of canonical blood pDC analogs. In addition, we show that, like activated blood pDC, ThINKK are not killed by either allogeneic T or NK cells.38 80% of ThINKK produce IFN-α and express PDL1+ but not CD80− on TLR stimulation, a phenotype characteristic of type I IFN-producing cells, but not antigen-presenting cells. A subset of ThINKK expresses the AXL tyrosine kinase receptor but differs from the blood AXL+ DC subset identified by CyTOF analysis and scRNA sequencing of blood and tissue DC.14 15 34 These blood AXL+ DC display a continuum from pDC transcriptional state (high E2-2 and BDCA2 expression) to a CD1c+ or CD141 DC transcriptional state (low E2-2 expression and CD1c or CD141 expression, specific markers of DC2 or DC1 cells respectively). Following in vitro differentiation in the presence of IL-3 and CD40L, these cells downregulate E2-2 gene expression and exhibit cDC phenotype and functions, such as T cell activation.14 34 Here we observed that, following TLR9 stimulation, AXL+CD141+CLEC10A+ ThINKK display low T cell activating properties, and importantly do not exacerbate T cell proliferation induced by allogeneic antigen-presenting cells. In addition, our in vivo study showed that ThINKK do not worsen GvHD in a mouse model of xeno-GvHD, despite the presence of AXL+ ThINKK. Although the xeno-GvHD model does not assess the capacity of human ThINKK to present murine antigens to grafted T cells, our results reveal that ThINKK-derived cytokines and chemokines do not exacerbate T cell responses against recipient’s antigens. Collectively, our results suggest that ThINKK adoptive transfers in patients will not worsen allogeneic T cell activation that can lead to life-threatening GvHD. The therapeutic efficacy of ThINKK adoptive therapy will therefore rely only on the stimulation of innate immunity and in particular, NK cells via the production of type-I IFN and other cytokines and chemokines.

    Our data and previous reports showed that ThINKK and human blood pDC express high levels of Granzyme B and, furthermore, that TLR-stimulation increased its expression.14 15 39 Direct tumor cytotoxicity of pDC against tumor cells has also been described, although the ratio of effector:target ratio used in these reports was not compatible with the rarity of natural pDC.39 Importantly, we never observed direct cytotoxic activity of ThINKK or blood pDC against ALL cells.29 In the absence of Perforin expression, the role of pDC-derived Granzyme B may differ from its function in cytotoxic effectors. Jahrsdörfer et al have suggested that pDC-derived Granzyme B might regulate T cell proliferation in a Treg reminiscent manner.40 Further investigations are needed to determine whether this is the case with ThINKK-derived Granzyme B.

    Two other groups have proposed the adoptive transfers of pDC to fight cancer.41–43 Both approaches aim to increase T-cell-based adaptive immunity against specific cancer antigens using pDC alone or in combination with cDC. ThINKK immunotherapy distinguishes itself from these approaches by harnessing only the innate immune system to fight cancer in a non-antigen specific manner. Since NK cell-mediated killing involves multiple lytic pathways including death receptor-induced apoptosis and cytotoxic granule release via the activation of multiple NK receptors, the risk of immune evasion by the down regulation of targeted antigen is lowered. As NK cell activation has been shown to increase the efficacy of allogeneic HSCT, we plan to use our novel approach in that setting.5 6 19 44 45 In a preclinical humanized HSCT model, the elimination of residual leukemia by ThINKK-stimulated NK cells prevented leukemia development.23 Patients with high-risk leukemia could thus receive ThINKK adoptive therapy within the first months following HSCT, when leukemia burden has been reduced by the conditioning regimen and when NK cells are the only GvL effectors.

    We showed here that ThINKK does not induce T cell proliferation, and therefore are not expected to induce GvHD, a life-threatening side effect of HSCT. In mice, Tian et al recently reported that adoptive transfers of pDC repressed GvHD through (1) T-cell repression by a type-I IFN-signaling dependant mechanism and (2) the restoration of functional pDC that further repressed GvHD.46 Importantly, this allogeneic T-cell repression did not decrease the GvL effect. In our preclinical model of human-to-mice xeno-GvHD, ThINKK therapy did not decrease the GvHD rate because of the low persistence of human NK cells in our model. However, several lines of evidence indicate that ThINKK may reduce the risk of GvHD in transplanted patients. Indeed, type I IFN has been showed to reduce CD8 T cell proliferation47 and clinical studies showed that higher numbers of NK cells as well as the presence of NK cell alloreactivity reduce the rate of GvHD in allogeneic transplanted patients.48–50 Mice studies indicate that the involved mechanisms include the secretion of transforming growth factor-β and the ablation of recipient antigen-presenting cells or donor alloreactive T cells.19–21

    The pharmacological strategy to prevent GvHD is the inhibition of T cell activation using immunosuppressive drugs targeting the cytoplasmic enzyme calcineurin (cyclosporine, tacrolimus), the mTOR pathway (sirolimus) or purine synthesis (mycophenolate). Prophylaxis regimen composed of these drugs alone or in combination is routinely administrated to HSCT patients, while overt GvHD is treated with corticosteroids such as methylprednisolone. The effect of these immunosuppressive drugs on NK cell activation remains controversial and may vary on NK cell stimulation status.27 28 51 The drugs with the most potent effect on NK cell function are corticosteroids. We indeed observed that a therapeutic dose of methylprednisolone decreases ThINKK viability and ThINKK-induced NK cell cytotoxicity against ALL. Importantly, our results demonstrated that ThINKK-induced NK cell cytotoxicity was preserved in the presence of therapeutic and supra-therapeutic doses of tacrolimus, sirolimus and mycophenolate, while cyclosporine decreased by half the NK cytotoxic activity against ALL independently of the dose. This difference between the two calcineurin inhibitors remains unexplained: cyclosporine and tacrolimus inhibit calcineurin by binding to two different immunophilins, cyclophilin A for the former and FK-binding proteins for the latter.52 The impact of calcineurin inhibition on type-I IFN production by pDC has been shown recently to depend on calcium concentration.53 However, the presence and role of immunophilins in pDC remain to be explored. Collectively, our results indicate that any GvHD prophylaxis regimen, excluding cyclosporine and corticosteroids, will be compatible with ThINKK adoptive immunotherapy.

    As a conclusion, our preclinical data on ThINKK indicate that post-transplant ThINKK immunotherapy is expected to improve the GvL effect without worsening GvHD in allogeneic HSCT settings. This hypothesis will be tested in an upcoming phase I clinical trial in children transplanted for high-risk leukemia.

    Supplemental material

    Data availability statement

    Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.

    Ethics statements

    Patient consent for publication

    Ethics approval

    This study involves human participants and the human study was approved by Ethics Review Board of CHU Ste-Justine (CER - Comité d’éthique à la Recherche) (ID# MP-21 2019-2028). Animal study was approved by Comité Institutionnel des Bonnes Pratiques Animales en Recherche (CIBPAR) du CHU Ste-Justine (ID: 2020-2687). Participants gave informed consent to participate in the study before taking part.

    Acknowledgments

    The authors thank Ines Boufaied for cell sorting and Thomas Sontag for the management of cord blood units. The authors also thank all participating volunteers for their valuable help.

    References

    1. ↵
      1. Hunger SP,
      2. Mullighan CG
      . Acute lymphoblastic leukemia in children. N Engl J Med 2015;373:1541–52. doi:10.1056/NEJMra1400972
      OpenUrlCrossRefPubMed
    2. ↵
      1. Malard F,
      2. Mohty M
      . Acute lymphoblastic leukaemia. The Lancet 2020;395:1146–62. doi:10.1016/S0140-6736(19)33018-1
      OpenUrl
    3. ↵
      1. Maude SL,
      2. Laetsch TW,
      3. Buechner J, et al
      . Tisagenlecleucel in children and young adults with B-cell lymphoblastic leukemia. N Engl J Med 2018;378:439–48. doi:10.1056/NEJMoa1709866
      OpenUrlCrossRefPubMed
    4. ↵
      1. Mathan T,
      2. Figdor CG,
      3. Buschow SI
      . Human plasmacytoid dendritic cells: from molecules to intercellular communication network. Front Immunol 2013;4:372. doi:10.3389/fimmu.2013.00372
    5. ↵
      1. Cichocki F,
      2. Verneris MR,
      3. Cooley S, et al
      . The past, present, and future of NK cells in hematopoietic cell transplantation and adoptive transfer. Springer International Publishing, 2015:225–43.
    6. ↵
      1. Locatelli F,
      2. Pende D,
      3. Falco M, et al
      . NK cells mediate a crucial graft-versus-leukemia effect in haploidentical-HSCT to cure high-risk acute leukemia. Trends Immunol 2018;39:577–90. doi:10.1016/j.it.2018.04.009
      OpenUrl
    7. ↵
      1. Beziat V,
      2. Nguyen S,
      3. Lapusan S, et al
      . Fully functional NK cells after unrelated cord blood transplantation. Leukemia 2009;23:721–8. doi:10.1038/leu.2008.343
      OpenUrlCrossRefPubMed
    8. ↵
      1. Foley B,
      2. Cooley S,
      3. Verneris MR, et al
      . NK cell education after allogeneic transplantation: dissociation between recovery of cytokine-producing and cytotoxic functions. Blood 2011;118:2784–92. doi:10.1182/blood-2011-04-347070
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Szabolcs P,
      2. Niedzwiecki D
      . Immune reconstitution in children after unrelated cord blood transplantation. Biol Blood Marrow Transplant 2008;14:66–72. doi:10.1016/j.bbmt.2007.10.016
      OpenUrlCrossRefPubMed
    10. ↵
      1. Charrier E,
      2. Cordeiro P,
      3. Brito R-M, et al
      . Reconstitution of maturating and regulatory lymphocyte Subsets after cord blood and BMT in children. Bone Marrow Transplant 2013;48:376–82. doi:10.1038/bmt.2012.176
      OpenUrlCrossRefPubMed
    11. ↵
      1. Bayer AL,
      2. Chirinos J,
      3. Cabello C, et al
      . Expansion of a restricted residual host T reg-cell repertoire is dependent on IL-2 following experimental autologous hematopoietic stem transplantation. Eur J Immunol 2011;41:3467–78. doi:10.1002/eji.201141611
      OpenUrlPubMed
    12. ↵
      1. Williams MTS,
      2. Yousafzai Y,
      3. Cox C, et al
      . Interleukin-15 enhances cellular proliferation and upregulates CNS homing molecules in pre-B acute lymphoblastic leukemia. Blood 2014;123:3116–27. doi:10.1182/blood-2013-05-499970
      OpenUrlAbstract/FREE Full Text
    13. ↵
      1. Liu S,
      2. Luo X,
      3. Zhang X, et al
      . Preemptive interferon-alpha treatment could protect against relapse and improve long-term survival of ALL patients after Allo-HSCT. Sci Rep 2020;10:20148. doi:10.1038/s41598-020-77186-9
      OpenUrl
    14. ↵
      1. Villani A-C,
      2. Satija R,
      3. Reynolds G, et al
      . Single-cell RNA-Seq reveals new types of human blood dendritic cells, monocytes, and progenitors. Science 2017;356:eaah4573. doi:10.1126/science.aah4573
    15. ↵
      1. See P,
      2. Dutertre C-A,
      3. Chen J, et al
      . Mapping the human DC lineage through the integration of high-dimensional techniques. Science 2017;356:eaag3009. doi:10.1126/science.aag3009
    16. ↵
      1. Socié G,
      2. Blazar BR
      . Acute graft-versus-host disease: from the bench to the bedside. Blood 2009;114:4327–36. doi:10.1182/blood-2009-06-204669
      OpenUrlAbstract/FREE Full Text
    17. ↵
      1. Gao F,
      2. Ye Y,
      3. Gao Y, et al
      . Influence of KIR and NK cell reconstitution in the outcomes of hematopoietic stem cell transplantation. Front Immunol 2020;11:2022. doi:10.3389/fimmu.2020.02022
    18. ↵
      1. Maurer K,
      2. Soiffer RJ
      . The delicate balance of graft versus leukemia and graft versus host disease after allogeneic hematopoietic stem cell transplantation. Expert Rev Hematol 2023;16:943–62. doi:10.1080/17474086.2023.2273847
      OpenUrl
    19. ↵
      1. Ruggeri L,
      2. Capanni M,
      3. Urbani E, et al
      . Effectiveness of donor natural killer cell alloreactivity in mismatched hematopoietic transplants. Science 2002;295:2097–100. doi:10.1126/science.1068440
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Olson JA,
      2. Leveson-Gower DB,
      3. Gill S, et al
      . NK cells mediate reduction of GVHD by inhibiting activated, alloreactive T cells while retaining GVT effects. Blood 2010;115:4293–301. doi:10.1182/blood-2009-05-222190
      OpenUrlAbstract/FREE Full Text
    21. ↵
      1. Asai O,
      2. Longo DL,
      3. Tian ZG, et al
      . Suppression of graft-versus-host disease and amplification of graft-versus-tumor effects by activated natural killer cells after allogeneic bone marrow transplantation. J Clin Invest 1998;101:1835–42. doi:10.1172/JCI1268
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Charrier E,
      2. Cordeiro P,
      3. Brito R-M, et al
      . Impaired interferon-alpha production by plasmacytoid dendritic cells after cord blood transplantation in children: implication for post-transplantation toll-like receptor ligand-based Immunotherapy. Biol Blood Marrow Transplant 2014;20:1501–7. doi:10.1016/j.bbmt.2014.06.007
      OpenUrl
    23. ↵
      1. Díaz-Rodríguez Y,
      2. Cordeiro P,
      3. Belounis A, et al
      . In vitro differentiated plasmacytoid dendritic cells as a tool to induce anti-leukemia activity of natural killer cells. Cancer Immunol Immunother 2017;66:1307–20. doi:10.1007/s00262-017-2022-y
      OpenUrl
    24. ↵
      1. Reizis B
      . Plasmacytoid dendritic cells: development, regulation, and function. Immunity 2019;50:37–50. doi:10.1016/j.immuni.2018.12.027
      OpenUrlCrossRefPubMed
    25. ↵
      1. Cangemi M,
      2. Montico B,
      3. Faè DA, et al
      . Dissecting the multiplicity of immune effects of immunosuppressive drugs to better predict the risk of de novo malignancies in solid organ transplant patients. Front Oncol 2019;9:160. doi:10.3389/fonc.2019.00160
    26. ↵
      1. Leung W,
      2. Campana D,
      3. Yang J, et al
      . High success rate of hematopoietic cell transplantation regardless of donor source in children with very high-risk leukemia. Blood 2011;118:223–30. doi:10.1182/blood-2011-01-333070
      OpenUrlAbstract/FREE Full Text
    27. ↵
      1. Pradier A,
      2. Papaserafeim M,
      3. Li N, et al
      . Small-molecule immunosuppressive drugs and therapeutic immunoglobulins differentially inhibit NK cell Effector functions in vitro Front Immunol 2019;10:556. doi:10.3389/fimmu.2019.00556
    28. ↵
      1. Meehan AC,
      2. Mifsud NA,
      3. Nguyen THO, et al
      . Impact of commonly used transplant immunosuppressive drugs on human NK cell function is dependent upon stimulation condition. PLoS One 2013;8:e60144. doi:10.1371/journal.pone.0060144
    29. ↵
      1. Lelaidier M,
      2. Dìaz-Rodriguez Y,
      3. Cordeau M, et al
      . TRAIL-mediated killing of acute lymphoblastic leukemia by plasmacytoid dendritic cell-activated natural killer cells. Oncotarget 2015;6:29440–55. doi:10.18632/oncotarget.4984
      OpenUrl
    30. ↵
      1. Roederer M
      . Interpretation of cellular proliferation data: avoid the panglossian. Cytometry A 2011;79:95–101. doi:10.1002/cyto.a.21010
      OpenUrlCrossRefPubMed
    31. ↵
      1. Gregoire-Gauthier J,
      2. Durrieu L,
      3. Duval A, et al
      . Use of Immunoglobulins in the prevention of GVHD in a xenogeneic NOD/SCID/Gammac- mouse model. Bone Marrow Transplant 2012;47:439–50. doi:10.1038/bmt.2011.93
      OpenUrlCrossRefPubMed
    32. ↵
      1. Cooke KR,
      2. Kobzik L,
      3. Martin TR, et al
      . An experimental model of idiopathic pneumonia syndrome after bone marrow transplantation: I. The roles of minor H antigens and Endotoxin. Blood 1996;88:3230–9.
      OpenUrlAbstract/FREE Full Text
    33. ↵
      1. Sulczewski FB,
      2. Maqueda-Alfaro RA,
      3. Alcántara-Hernández M, et al
      . Transitional dendritic cells are distinct from conventional Dc2 precursors and mediate proinflammatory antiviral responses. Nat Immunol 2023;24:1265–80. doi:10.1038/s41590-023-01545-7
      OpenUrl
    34. ↵
      1. Alcántara-Hernández M,
      2. Leylek R,
      3. Wagar LE, et al
      . High-dimensional phenotypic mapping of human dendritic cells reveals Interindividual variation and tissue specialization. Immunity 2017;47:1037–50. doi:10.1016/j.immuni.2017.11.001
      OpenUrlCrossRefPubMed
    35. ↵
      1. Alculumbre SG,
      2. Saint-André V,
      3. Di Domizio J, et al
      . Diversification of human plasmacytoid predendritic cells in response to a single stimulus. Nat Immunol 2018;19:63–75. doi:10.1038/s41590-017-0012-z
      OpenUrlCrossRefPubMed
    36. ↵
      1. Thurner B,
      2. Röder C,
      3. Dieckmann D, et al
      . Generation of large numbers of fully mature and stable dendritic cells from leukapheresis products for clinical application. J Immunol Methods 1999;223:1–15. doi:10.1016/s0022-1759(98)00208-7
      OpenUrlCrossRefPubMedWeb of Science
    37. ↵
      1. Ruggeri L,
      2. Mancusi A,
      3. Burchielli E, et al
      . NK cell alloreactivity and allogeneic hematopoietic stem cell transplantation. Blood Cells Mol Dis 2008;40:84–90. doi:10.1016/j.bcmd.2007.06.029
      OpenUrlCrossRefPubMedWeb of Science
    38. ↵
      1. Gerosa F,
      2. Gobbi A,
      3. Zorzi P, et al
      . The reciprocal interaction of NK cells with plasmacytoid or myeloid dendritic cells profoundly affects innate resistance functions. J Immunol 2005;174:727–34. doi:10.4049/jimmunol.174.2.727
      OpenUrlAbstract/FREE Full Text
    39. ↵
      1. Wu J,
      2. Li S,
      3. Yang Y, et al
      . TLR-activated plasmacytoid dendritic cells inhibit breast cancer cell growth in vitro and in vivo. Oncotarget 2017;8:11708–18. doi:10.18632/oncotarget.14315
      OpenUrl
    40. ↵
      1. Jahrsdörfer B,
      2. Vollmer A,
      3. Blackwell SE, et al
      . Granzyme B produced by human Plasmacytoid Dendritic cells suppresses T-cell expansion. Blood 2010;115:1156–65. doi:10.1182/blood-2009-07-235382
      OpenUrlAbstract/FREE Full Text
    41. ↵
      1. Bol KF,
      2. Schreibelt G,
      3. Rabold K, et al
      . The clinical application of cancer immunotherapy based on naturally circulating dendritic cells. J Immunother Cancer 2019;7:109. doi:10.1186/s40425-019-0580-6
    42. ↵
      1. Charles J,
      2. Chaperot L,
      3. Hannani D, et al
      . An innovative plasmacytoid dendritic cell line-based cancer vaccine primes and expands antitumor T-cells in Melanoma patients in a first-in-human trial. Oncoimmunology 2020;9:1738812. doi:10.1080/2162402X.2020.1738812
    43. ↵
      1. van Eck van der Sluijs J,
      2. van Ens D,
      3. Thordardottir S, et al
      . Clinically applicable Cd34(+)-derived blood dendritic cell subsets exhibit key subset-specific features and potently boost anti-tumor T and NK cell responses. Cancer Immunol Immunother 2021;70:3167–81. doi:10.1007/s00262-021-02899-3
      OpenUrl
    44. ↵
      1. Cooley S,
      2. Weisdorf DJ,
      3. Guethlein LA, et al
      . Donor selection for natural killer cell receptor genes leads to superior survival after unrelated transplantation for acute myelogenous leukemia. Blood 2010;116:2411–9. doi:10.1182/blood-2010-05-283051
      OpenUrlAbstract/FREE Full Text
    45. ↵
      1. Handgretinger R,
      2. Lang P,
      3. André MC
      . Exploitation of natural killer cells for the treatment of acute leukemia. Blood 2016;127:3341–9. doi:10.1182/blood-2015-12-629055
      OpenUrlAbstract/FREE Full Text
    46. ↵
      1. Tian Y,
      2. Meng L,
      3. Wang Y, et al
      . Graft-versus-host disease depletes plasmacytoid dendritic cell progenitors to impair tolerance induction. J Clin Invest 2021;131:e136774. doi:10.1172/JCI136774
    47. ↵
      1. Sköld AE,
      2. Mathan TSM,
      3. van Beek JJP, et al
      . Naturally produced type I IFNs enhance human myeloid dendritic cell maturation and IL-12P70 production and mediate elevated effector functions in innate and adaptive immune cells. Cancer Immunol Immunother 2018;67:1425–36. doi:10.1007/s00262-018-2204-2
      OpenUrl
    48. ↵
      1. Simonetta F,
      2. Alvarez M,
      3. Negrin RS
      . Natural killer cells in graft-versus-host-disease after allogeneic hematopoietic cell transplantation. Front Immunol 2017;8:465. doi:10.3389/fimmu.2017.00465
    49. ↵
      1. Ludajic K,
      2. Balavarca Y,
      3. Bickeböller H, et al
      . KIR genes and KIR ligands affect occurrence of acute GVHD after unrelated, 12/12 HLA matched, hematopoietic stem cell transplantation. Bone Marrow Transplant 2009;44:97–103. doi:10.1038/bmt.2008.432
      OpenUrlCrossRefPubMed
    50. ↵
      1. Sivori S,
      2. Carlomagno S,
      3. Falco M, et al
      . Natural killer cells expressing the KIR2DS1-activating receptor efficiently kill T-cell blasts and Dendritic cells: implications in Haploidentical HSCT. Blood 2011;117:4284–92. doi:10.1182/blood-2010-10-316125
      OpenUrlAbstract/FREE Full Text
    51. ↵
      1. Vitale C,
      2. Chiossone L,
      3. Cantoni C, et al
      . The corticosteroid-induced inhibitory effect on NK cell function reflects down-regulation and/or dysfunction of triggering receptors involved in natural cytotoxicity. Eur J Immunol 2004;34:3028–38. doi:10.1002/eji.200425418
      OpenUrlCrossRefPubMedWeb of Science
    52. ↵
      1. Barbarino JM,
      2. Staatz CE,
      3. Venkataramanan R, et al
      . Pharmgkb summary: cyclosporine and tacrolimus pathways. Pharmacogenet Genomics 2013;23:563–85. doi:10.1097/FPC.0b013e328364db84
      OpenUrlCrossRefPubMed
    53. ↵
      1. Raychaudhuri D,
      2. Duttagupta P,
      3. Liu CSC, et al
      . Role of Ca(2+) in toll-like receptor 9 activation in human plasmacytoid dendritic cells. Cytokine 2020;125:154822. doi:10.1016/j.cyto.2019.154822

    Supplementary materials

    Footnotes

    • Contributors NP performed the experiments and wrote the manuscript. VP, PC, CM and SL performed the experiments. VL and HA performed the bioinformatics analysis. YT, GA, VPL and EH provide their expertise in pharmacology (YT), scRNA sequencing (GA and VPL) and mouse models (EH). MD and SH designed the project, analysed the data and wrote the manuscript. SH is the guarantor.

    • Funding This study has been financed by private and governmental grants (Fondation Charles Bruneau, BioCanRx Network #CAT18 and Canadian Institute of Health Research grants #142373).

    • Competing interests None declared.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.