The factor ETS1 is an important regulator of human NK cell development and terminal differentiation Sylvie Taveirne, Sigrid Wahlen, Wouter van Loocke, Laura Kiekens, Eva Persyn, Els van Ammel, Katrien de Mulder, Juliette Roels, Laurentijn Tilleman, Marc Aumercier, et al.

To cite this version:

Sylvie Taveirne, Sigrid Wahlen, Wouter van Loocke, Laura Kiekens, Eva Persyn, et al.. The transcrip- tion factor ETS1 is an important regulator of human NK cell development and terminal differentiation. Blood, American Society of Hematology, 2020, 136 (3), pp.288-298. ￿10.1182/blood.2020005204￿. ￿hal- 03015322￿

HAL Id: hal-03015322 https://hal.archives-ouvertes.fr/hal-03015322 Submitted on 19 Nov 2020

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés.

The ETS1 is an important regulator of human NK cell development and terminal differentiation

Short title: The role of ETS1 in human NK biology

S. Taveirne,1,2,* S. Wahlen,1,2,* W. Van Loocke,2,3 L. Kiekens,1,2 E. Persyn,1,2 E. Van Ammel,1 K. De Mulder,1 J. Roels,2,3 L. Tilleman,4 M. Aumercier,5,6 P. Matthys,7 F. Van Nieuwerburgh,4 T. Kerre,1,2 T. Taghon,1,2 P. Van Vlierberghe,2,3 B. Vandekerckhove1,2 and G. Leclercq1,2 1Department of Diagnostic Sciences, Laboratory of Experimental Immunology, Ghent University, Ghent, Belgium 2Cancer Research Institute Ghent (CRIG), Ghent University, Ghent, Belgium 3Department of Biomolecular Medicine, Ghent University and Cancer Research Institute Ghent, 9000 Ghent, Belgium; Cancer Research Institute Ghent (CRIG), Ghent, Belgium 4Department of Pharmaceutics, Ghent University, Ghent, Belgium 5Univ. Lille, Inserm, CHU Lille, Institut Pasteur de Lille, U1167 -RID-AGE- Risk Factors and Molecular Determinants of Aging-Related Diseases, F-59000 Lille, France 6CNRS ERL9002, Integrative Structural Biology, F-59000 Lille, France 7Laboratory of Immunobiology, Rega Institute for Medical Research, Department of Microbiology, Immunology and transplantation, KU Leuven, Leuven, Belgium *S.T. and S.W. contributed equally to this study.

Corresponding author: Georges Leclercq C. Heymanslaan 10, MRB2 entrance 38, 9000 Ghent, Belgium e-mail: [email protected] phone: +32 9 332 37 34 Fax: +32 9 332 36 59

Word count for text: 4104 Word count for abstract: 157 Figure count: 6 Reference count: 56

Primary scientific category: Hematopoiesis and stem cells Secondary scientific category: Immunobiology and Immunotherapy

Key Points • Human ETS1 deficiency inhibits expression of several NK cell-linked key transcription factors

and inhibits NK cell differentiation

• ETS1 is required for tumor cell-induced NK cell cytotoxicity and IFN-γ production

2

Abstract

Natural killer (NK) cells are important in the immune defense against tumor cells and pathogens, and regulate other immune cells by cytokine secretion. Whereas murine NK cell biology has been extensively studied, knowledge about transcriptional circuitries controlling human NK cell development and maturation is limited. By generating ETS1-deficient human embryonic stem cells (hESC) and by expressing the dominant-negative ETS1 p27 isoform in cord blood (CB) hematopoietic progenitor cells (HPCs), we show that the transcription factor ETS1 is critically required for human NK cell differentiation. Genome-wide transcriptome analysis determined by

RNA-sequencing combined with chromatin immunoprecipitation-sequencing (ChIP-seq) analysis reveals that human ETS1 directly induces expression of key transcription factors that control NK cell differentiation, i.e. E4BP4, TXNIP, TBET, GATA3, HOBIT and BLIMP1. In addition, ETS1 regulates expression of involved in apoptosis and NK cell activation. Our study provides important molecular insights into the role of ETS1 as an important regulator of human NK cell development and terminal differentiation.

3

Introduction

NK cells play a crucial role in the defense against viral infections and carcinogenesis through lysis of infected or transformed cells. Via secretion of cytokines, they also regulate T and B cells.1 In acute myeloid leukemia patients, transplanted donor NK cells expressing mismatched inhibitory receptors provide graft-versus-leukemia effect in absence of graft-versus-host disease.2 There are several ongoing studies and clinical trials for NK cell immunotherapy as treatment of leukemia and solid cancers.3

Differentiation and lineage specification of hematopoietic progenitor cells (HPCs) into myeloid and lymphoid cells is regulated by transcription factors. NK cells are derived from common lymphoid progenitors (CLPs). NK cell progenitors (NKP) express CD122, become IL-15 responsive and are NK cell-restricted. During further differentiation, inhibitory and activating NK receptors are expressed and functional capabilities, such as cytotoxicity and cytokine secretion, are eventually acquired. In contrast to mouse, detailed knowledge of the regulation of human NK cell differentiation is lacking. Due to increased interest in NK cells as therapeutic agents, a more complete understanding of the molecular basis of human NK cell differentiation is desirable.

However, as there are important differences between mouse and human NK differentiation, human models remain essential for translational purposes.

The transcription factor ETS1 is predominantly expressed in lymphoid cells. Ets1-/- mice have reduced NK cell numbers that display lower cytolytic activity.4,5 The role of ETS1 in human NK cell biology remains unknown. There are 3 human ETS1 isoforms: ETS1 p51 encoded by exons I to VII,

ETS1 p42 lacking the sequences encoded in exon VII, and ETS1 p27 lacking the sequences encoded in exons III-VI.6-8 ETS1 p42 acts as a gain-of-function as the auto-inhibitory motif

4 that suppresses DNA binding is absent. ETS1 p27 possesses the DNA-binding domain flanked by inhibitory domains, but lacks the Pointed and the transactivation domains which are required for the transactivation of ETS1 target genes. It has been shown that ETS1 p27 binds DNA in the same way as ETS1 p51 does, but it has a dual-acting dominant-negative effect as it not only acts at the transcriptional level by blocking ETS1 p51-mediated transactivation of target genes, but also affects the subcellular localization by inducing the translocation of ETS1 p51 from the nucleus to the cytoplasm.6,8

Here, we reveal the role of ETS1 in human NK cell differentiation and terminal differentiation by performing knockout, loss-of-function and overexpression experiments. NK cell differentiation is inhibited as a consequence of absent or reduced ETS1 function. Transcriptome and chromatin immunoprecipitation-sequencing (ChIP-seq) analysis reveals that ETS1 directly induces expression of 6 key transcription factors that have been previously reported to control NK cell differentiation and/or maturation. Finally, residual NK cells in ETS1 loss-of-function display defective cytotoxicity and cytokine production upon tumor cell contact, presumably due to reduced expression of activating NK receptors. Our data provide molecular insights in the role of

ETS1 as a critical regulator of human NK cell biology.

5

Methods

ETS1-deficient human embryonic stem cells (hESCs) and hematopoietic differentiation

WA01 hESCs were kept in undifferentiated state using the hESC medium (KOSR)/MEF platform/WiCell Feeder Dependent Protocol (http://www.wicell.org/). Single cell-adapted hESCs were generated as described9 and used to generate ETS1-/- hESCs via the CRISPR/Cas9 technology10,11 targeting exon A/1. Detailed information can be found in Supplemental Methods.

Umbilical CB HPC differentiation

CD34+ HPCs were in vitro differentiated into NK cells as described by Cichocki and Miller12 with modifications (see Supplemental Methods).

Viral constructs cDNA encoding human ETS1 p51 or ETS1 p27 was subcloned from the pLEXhETS1p51HAtag

(kindly provided by L. A. Garrett-Sinha, State University of New York, Buffalo, NY, U.S.A.) and pCDNA3hETS1p27 vector8 into the LZRS-internal ribosome entry site (IRES)-eGFP retroviral vector.13 The empty LZRS-IRES-eGFP vector was used as control. Virus was generated as described.13

Flow cytometry

Cultures were evaluated using an LSRII flow cytometer and data were analyzed with the FACSDiva

Version 6.1.2 (both from BD Biosciences, San Jose, CA, U.S.A.) and/or FlowJo_V10 software

(Ashland, OR, U.S.A.). Used antibodies are listed in Supplemental Methods.

Functional assays

6

IFN-γ assays

For intracellular IFN-γ flow cytometric detection, cells from day 21 (d21) co-cultures were stimulated with K562 cells at a 1:1 ratio for 6 h, or with IL-12/IL-18 (both 10 ng/ml) or IL-12/IL-15

(4 ng/ml)/IL-18 for 24 h. Brefeldin A (BD GolgiPlug, BD Biosciences) was added during the last 4 hours. NK cell surface marker staining was performed, hereafter cells were fixed and permeabilised (Cytofix/Cytoperm Kit (BD Biosciences)) and stained for IFN-γ.

For detection of secreted IFN-γ, eGFP+CD45+CD11a+CD56+CD94+ NK cells were sorted from d21 cultures using a BD FACSAria III cell sorter (BD Biosciences) and stimulated with IL-12/IL-18 or IL-

12/IL-15/IL-18 for 24 h. Secreted IFN-γ was quantified using the human IFN-γ ELISA PeliKine-Tool set (Sanquin, Amsterdam, The Netherlands).

CD107a detection assay

Cells from d18-21 co-cultures were stimulated by co-culture with K562 cells (ATCC, Manassas,

VA, U.S.A.) during 2 hours, as described by Bryceson et al.14

51Chromium release assay

51 K562 target cells were labeled with Na2 CrO4 (Perkin Elmer, Waltham, MA, U.S.A.). eGFP+CD45+CD11a+CD94+ NK cells were sorted from d18-21 co-cultures and added at various ratios to 51Cr-labeled K562 cells. After 4 hours, supernatant was harvested and measured in a

1450 LSC&Luminescence Counter (Wallac Microbeta TriLux, Perkin Elmer). The mean percentage of specific lysis of triplicate wells was calculated.

Library prep, sequencing, qPCR and identification of putative ETS1 binding sites

7

Viable day 18 eGFP+CD45+CD11a+CD56+CD94+ NK cells were sorted to perform transcriptome analysis. Samples were sequenced with the high throughput Illumina NextSeq 500 flow cell generating 75 bp single reads. Differential expression analysis was performed using DESeq2. Data are accessible on GEO (accession number GSE124104; token: adkxagywplqxhkp). Details can be found in supplemental data.

For ChIP-seq, sorted viable peripheral blood NK cells were crosslinked in 1% formaldehyde.

Protein-DNA complexes were immunoprecipitated with rabbit polyclonal anti-ETS1 Ab (Active

Motif, Carlsbad, CA, U.S.A). Immune complexes were isolated with protein A agarose beads

(Thermo Fisher Scientific, Waltham, MA, U.S.A.). Crosslinks were reversed at 65°C. After DNA purification, Illumina sequencing libraries were prepared from the ChIP and Input DNAs by the standard consecutive enzymatic steps of end-polishing, dA-addition, and adaptor ligation. After a final PCR amplification step, the resulting DNA libraries were quantified and sequenced on

Illumina’s NextSeq 500. Data are accessible on the GEO repository with accession number

GSE124104. See supplemental data for detailed information. Putative ETS1 binding sites in promotors and enhancers were detected by analysis using MACS2 and HOMER software.15,16

Statistical analysis

Statistical analysis was performed using SPSS Statistics 24 software (Armonk, NY, U.S.A.). A p- value generated by a t-test < 0.05 was considered statistically significant.

8

Results

Strongly reduced capacity of ETS1 knockout hESCs to differentiate into NK cells

An ETS1-specific guideRNA was selected that efficiently generated ETS1 knockout clones using the CRISPR/Cas9 technology in Jurkat cells (Figure S1). Using the same guideRNA in hESC, we selected three hESC ETS1-/- clones (15.3, 17.3, 21.1) that carried nucleotide deletions resulting in a bi-allelic frameshift. The wild type clone (16.1) underwent similar selection processes (Figure

1A). hESC clones were differentiated into HPCs (CD34+CD43+). Homozygous ETS1 deficiency had no significant effect on absolute HPC numbers (Figure 1B). hESC-derived HPCs were cultured in

NK differentiation conditions. Whereas WT hESCs efficiently generated NK cells, confirming earlier studies17, significantly less NK cells were generated from each ETS1-/- clone (Figure 1C-D).

Thus, whereas ETS1 is not required for differentiation of hESCs into HPCs, it is crucial for NK cell development thereof.

ETS1 expression is essential for human NK cell differentiation from cord blood HPCs

HPCs generated from hESCs resemble yolk sac hematopoietic progenitors with limited stem cell activity.18 We therefore also studied the role of ETS1 in NK differentiation from CB HPCs, that are considered bona fide HPCs. We transduced dominant-negative ETS1 p27 and full-length ETS1 p51 into CB HPCs, next to a vector control, and stable overexpression was observed (Figure S2A-D).

ETS1 p27 transduction resulted in strong reduction of NK cell numbers (gated as shown in figure

S3A), whereas ETS1 p51 overexpression had no effect (Figure 2A). Since we observed less mature

NK cells in ETS1 loss-of-function conditions, we investigated whether ETS1 affects earlier NK developmental stages. Whereas stage 1 cells (CD34+CD45RA+CD117-) (gating shown in figure S3B)

9 were not affected, stage 2 cells (CD34+CD45RA+CD117+) displayed a trend in lower cell numbers in ETS1 p27 cultures, which became significant in stage 3 cells (CD34-CD117+CD94-NKp44-) (Figure

2B). These results show that ETS1 is an essential factor in human NK cell generation from CB HPCs starting at stage 3.

ETS1 regulates expression of key transcription factors linked to NK cell differentiation

To achieve mechanistic insight, the genome-wide transcriptome of sorted ETS1 p27 versus control NK cells was determined by RNA-sequencing. We identified 1178 differentially expressed genes (Figure 3A). ETS1 p27 NK cells displayed downregulated expression of molecules related to lymphocyte differentiation (Figure S4). Strikingly, several key transcription factors of murine and/or human NK cell differentiation were repressed, i.e. GATA3, HOBIT (ZNF683), TBET (TBX21),

TXNIP, E4BP4 (NFIL3) and BLIMP1 (PRDM1). Inversely, AIOLOS (IKZF3) and EOMES

() were upregulated (Figure 3A). This was confirmed by qPCR and/or flow cytometry (Figure 3B-D).

Expression of these 8 transcription factors was also tested in the precursor stages. None was differentially expressed at stage 1. In ETS1 p27 stage 2 cells, GATA3 expression was significantly reduced, whereas it was increased at stage 3. At stage 3, HOBIT, TBET and AIOLOS expression was reduced in ETS1 p27 cultures (Figure 3B-C).

Thus, ETS1 p27 first affects GATA3 (from stage 2), then HOBIT, TBET and AIOLOS (from stage 3), whereas at the mature NK cell stage the expression of all 8 transcription factors is compromised.

10

Multiple key NK cell-linked transcription factors are direct ETS1 targets

Whereas ETS1 ChIP data are available for human non-NK immune subpopulations and cell lines, it was not available for murine or human NK cells. As it is well known that target genes of transcription factors are cell type specific, we performed ETS1 ChIP-seq on unmanipulated human NK cells and detected 7738 ETS1 peaks. The primary motif identified was ETS1, validating the ChIP-seq experiment, followed by ETS1:RUNX and RUNX motifs (Figure 4A). This is in line with previous findings that ETS1 and RUNX1 bind in a cooperative manner to large cohorts of enhancers controlling T lineage commitment.19 The TBET motif was also enriched, albeit with lower significance (Figure 4A). This suggests that these transcription factors share regulatory regions. The ETS1 binding motif was mainly found within promoter regions (Figure 4B), whereas

RUNX and TBET motifs were enriched in enhancers (data not shown). This indicates that differential complex formation provides a mechanism by which ETS1 can be directed to other genomic locations.

Importantly, combined analysis of the ETS1 ChIP-seq data with available ATAC-seq as well as

H3K4me3 and H3K27ac ChIP-seq data from human NK cells20 revealed that ETS1 ChIP peaks were present in promotors or enhancers of the above indicated ETS1-regulated NK cell-linked key transcription factors (Figure 4C and Table S1).

There was a 44% and 48% overlap between ETS1 ChIP-seq peaks and ETS1 p27 down- and upregulated genes, respectively (Figure 4D). Non-biased analysis of enriched transcription factor motifs21 in ETS1 p27 down- and upregulated genes shows that motifs for BLIMP1, E4BP4, GATA3 and HOBIT were significantly enriched (Figure 4E).

11

Thus, several transcription factors identified by others to be important in NK cell biology are differentially expressed in ETS1 loss-of-function conditions, and ChIP data indicate that these are direct ETS1 targets. In addition, whereas part of the differentially expressed genes overlap with the ETS1 ChIP peaks, also motifs of several of the differentially expressed NK cell-linked transcription factors are enriched in down- and upregulated genes of the ETS1 p27 cultures.

Increased apoptosis in NK cells from ETS1 p27 cultures

We investigated whether the lower NK cell number in the ETS1 p27 cultures (Figure 2) resulted from increased apoptosis and/or decreased proliferation. The programmed cell death pathway was enriched in the upregulated genes of ETS1 p27 NK cells (Figure S5). qPCR confirmed higher expression of the pro-apoptotic PMAIP-1 and lower expression of several anti-apoptotic genes, including TSC22D3, BIRC3 and PIM2, in ETS1 p27 cultures (Figure 5A). This corresponded to increased apoptosis in ETS1 p27 NK cells (Figure 5B). Additionally, ChIP-seq analysis revealed that anti-apoptotic genes are direct ETS1 targets (Table S1).

The mitosis pathway was enriched in ETS1 p27 upregulated genes (Figure 5C), suggesting that proliferation would be increased. However, further analysis showed that the majority of these upregulated genes negatively regulate mitosis (Figure 5D and Table S2). analysis showed that the percentage of non-proliferating (G0) cells was not different in ETS1 p27 versus control NK cells. In proliferating cells, there was a small significant decrease in G1 cells, and an increase at the G2/M phase (Figure 5E). We observed similar proliferation rates in cell trace experiments starting at d14 of culture and analyzed after 7 days (data not shown).

12

In conclusion, increased apoptosis results in reduced NK cell numbers in ETS1 p27 cultures.

ETS1 is required for NK cell cytotoxicity and IFN-γ production upon tumor cell recognition

NK-mediated cytotoxicity is one of the significant pathways of downregulated genes in ETS1 p27

NK cells (Figure 5C). As activating receptors trigger NK cell cytotoxicity, we first analyzed NKG2D,

NKp30, NKp44 and NKp46 expression. Whereas NKG2D expression was unaffected, NKp30,

NKp44 and NKp46 expression was decreased in ETS1 p27 NK cells (Figure 6A). KIR expression was not different from control NK cells (Figure 6B). Functionally, ETS1 p27 NK cells displayed less effective degranulation upon K562 coculture (Figure 6C) as well as reduced killing

(Figure 6D). However, perforin and granzyme B protein levels tended to be higher,(Figure 6E).

The lower cytotoxicity of ETS1 p27 NK cells is not due to other killing pathways, as Fas ligand expression was not different in control versus ETS1 p27 cultures, neither on unstimulated nor on

K562 target-stimulated NK cells (Figure 6F) and TRAIL expression was undetectable (data not shown).

To analyze the role of ETS1 in IFN-γ production, cells were stimulated with K562 target cells. This resulted in lower IFN-γ production in ETS1 p27 NK cells as compared to the control (Figure 6G).

Upon stimulation with IL-12/IL-18 or IL-12/IL-15/IL-18, there was a not significant trend in increased IFN-γ production by ETS1 p27 NK cells, as evidenced by intracellular IFN-γ analysis

(Figure 6G), and IFN-γ secretion was significantly higher upon IL-12/IL-15/IL-18 stimulation as measured by ELISA (Figure 6H).

13

Thus, ETS1 loss-of-function inhibits expression of several activating NK receptors and reduces granule-mediated cytotoxicity and tumor target-induced IFN-γ production. Cytokine-induced

IFN-γ production is not affected.

To study whether ETS1 is required to maintain NK cell functionality, mature NK cells were analyzed upon transduction with control versus ETS1 p27 retrovirus. Degranulation, cytotoxicity, as well as IFN-γ production and secretion were not affected (Figure S6A-D). This indicates that

ETS1 mainly exerts its effects during NK cell differentiation and is not essential to maintain the functionality of mature NK cells.

14

Discussion

Insight into the regulatory role of transcription factors in human NK biology is still in its infancy.

We show that ETS1 regulates expression of several key transcription factors known to act at different levels of NK cell development, including E4BP4, TBET, TXNIP, GATA3, BLIMP1, HOBIT,

AIOLOS and EOMES. These are direct ETS1 targets as shown by ChIP-seq analysis. We therefore show that ETS1 is an important regulator of human NK cell development and terminal differentiation.

Endogenous ETS1 expression is detectable at stages 1 and 2 of human NK cell differentiation and rises at stage 3 to reach levels equivalent to mature NK cells.22 This is consistent with our findings, showing that ETS1 deficiency in hESCs does not influence the generation of HPCs and ETS1 p27 in CB HPCs does not affect development of stages 1 and 2, whereas stage 3 and mature NK cells are significantly decreased. These results are similar to what has been observed in Ets1-/- mice, where CLPs are not affected, and pre-NK cell progenitors (pre-NKP) and rNKP are decreased.5

We show that human ETS1 orchestrates expression of transcription factors that play key roles in

NK cell biology. Murine E4bp4 controls NK commitment. E4bp4-/- mice display reduced development of CLPs into NK progenitors, having less immature NK cells (NK1.1+DX5-) and mature

NK cells (NK1.1+DX5+) are almost absent.23-25 The requirement for E4bp4 in NK cell differentiation is restricted to NK cell progenitors as ablation of E4bp4 in NK cells does not affect NK cell numbers.26 Txnip is a stress-response gene expressed from NK progenitors. NK progenitors as well as mature NK cell numbers are reduced in Txnip-/- mice, correlating with decreased CD122 expression.27 Eomes-/- mice have reduced transition from early CD27+CD11b- into intermediate

CD27+CD11b+ NK cell stages, whereas T-bet-/- mice have reduced NK cell numbers in the final

15

CD27-CD11b+ stage.28,29 Mice deficient for both T-bet and Eomes have a complete loss of NK cells in all organs.29 T-bet and Eomes thus have both redundant and specific activities.28-34 Murine Ets1 induces T-bet expression5, whereas Eomes expression depends on E4bp4.24 It is thus remarkable that human EOMES expression is upregulated in response to ETS1 loss-of-function that, in contrast to observations in the mouse, results in E4BP4 downregulation. However, limited numbers of murine Eomes-positive tissue-resident NK cells are still able to develop in absence of

E4bp4, which suggests alternate molecular mechanisms for the induction of Eomes expression.35-

37 Also, Eomes and T-bet levels correlate inversely in differentiating murine NK cells, suggesting that negative feedback mechanisms regulate their balance30,38, which possibly explains enhanced

EOMES and decreased TBET expression in human ETS1 p27 NK cells. Murine Gata3-/- NK cells display reduced CD27+CD11b- to CD27+CD11b+ transition and decreased NK cell numbers in peripheral organs.39,40 Gata3 deletion ablates generation of murine IL-7Rα+ thymic NK cells.41

Conventional murine Gata3-/- NK cells have decreased T-bet expression.40 One of the most affected transcription factors in human ETS1 p27 NK cells is HOBIT. In mice, Hobit is highly expressed in liver tissue-resident NK cells, whereas it is virtually absent in liver and spleen conventional NK cells. In line with this, only murine liver tissue-resident NK cells are Hobit dependent.42 In contrast to murine findings, HOBIT is the most upregulated transcription factor during NK cell differentiation from CB HPCs, and HOBIT knockdown results in reduction of NK progenitors and largely abrogates NK cell generation. Remaining NK cells show unaltered degranulation, whereas IFN-γ production is increased.43 Blimp1, a homologue of Hobit, is required for terminal maturation of murine spleen NK cells. Residual NK cells have an unaltered capacity of IFN-γ production and cytotoxicity in vitro, although RMA-S tumor growth in vivo is

16 increased.44 Whereas most NK-linked transcription factors were repressed in ETS1 p27 NK cells, expression of AIOLOS (IKZF3) and EOMES was significantly increased. Murine Aiolos is constitutively expressed in the NK cell lineage and Ikzf3-/- mice have a block in terminal differentiation.45 This differentiation block resembles that found in mice lacking T-bet and

Blimp1, however, mice lacking either T‐bet, Blimp1 or both have normal Aiolos expression, suggesting that Aiolos functions at a different level in NK cell differentiation.45

Taken together, we have shown a remarkable influence of ETS1 on expression of several transcription factors that have key roles in NK cell development as demonstrated before by the generation of KO mice for these individual factors. Whereas the dominant-negative ETS1 p27 isoform significantly reduces expression of each of these 6 key transcription factors in human, it does not completely abolish expression. The severely compromised human NK cell differentiation in absence of functional ETS1 is therefore presumably due to the synergistic effect of reduction of expression of these different key NK cell-linked transcription factors. NK cells in Ets1-/- mice also have decreased Tbet expression, but, in contrast to ETS1 p27 NK cells, have decreased Eomes and increased E4bp4 and Helios expression. Also in contrast to findings in Ets1-/- mice5, we did not detect a difference in the other IKAROS family member, HELIOS (IKZF2), in ETS1 p27 NK cells.

It can be argued that the dominant-negative ETS1 isoform also impacts binding and transcriptional regulation by other ETS family proteins, as members of the ETS family are characterized by a well-conserved DNA-binding-domain. Approximately half of the ETS members are significantly expressed in mature NK cells (Table S3). However, in this context it is important that the full length ETS1 (p51) is auto-inhibited for its binding to DNA due to the presence of 2 inhibitory domains surrounding and interacting with its DNA-binding-domain.46 To counteract

17 autoinhibition and improve DNA binding, ETS1 interacts with partners, enabling their cooperative binding to adjacent DNA elements.47 This binding auto-inhibition brings an additional degree of complexity and specificity in the transcriptional function of ETS1 p51 and enhances the specificity of function of ETS1 in comparison to other ETS proteins that are not auto-inhibited. This compelling need for partners is essential for ETS1 function and results in a particular space time function. Both inhibitory domains surrounding the DNA-binding domain are conserved in ETS1 p27, suggesting that ETS1 p27, like the full-length ETS1 p51 isoform, is autoinhibited for DNA binding. In this way, Ets-1 p27 may not be a dominant-negative protein for other members of the

ETS family.

To study the underlying cause of the lower NK cell number in ETS1 p27 cultures, we tested both apoptosis as well as proliferation. Apoptosis was significantly increased and correlated with decreased expression of anti-apoptotic genes that were also direct ETS1 targets as shown by

ChIP-seq, including TSC22D3, BIRC3 and PIM2, as well as with increased expression of pro- apoptotic genes such as PMAIP1. Both pro- and anti-apoptotic features have been ascribed to

ETS1, depending on the biological context.48-51 Increased apoptosis also accounts for the markedly decreased murine mature thymocyte and peripheral numbers generated from

Ets1-/- precursors.52,53 Transcriptome analysis showed that, although the mitosis pathway was enriched in ETS1 p27 upregulated genes, the majority of these genes negatively modulate mitosis. Cell cycle analysis further showed that there was no major difference in ETS1 p27 versus control NK cells. Overall, this indicates that the lower ETS1 p27 NK cell number is due to increased apoptosis rather than to a difference in proliferation.

18

ETS1 has an important role in human NK cell terminal differentiation and function as ETS1 p27

NK cells show reduced cytotoxicity and target cell-induced IFN-γ production. The activating receptors NKG2D, NKp30, NKp44 and NKp46 are important to activate both NK cell cytotoxicity and cytokine production upon tumor target recognition.54,55 Whereas NKG2D expression was unaffected, expression of NKp30, NKp44 and NKp46 was decreased in ETS1 p27 NK cells. Also NK cells in Ets1-/- mice have unaffected expression of NKG2D and decreased expression of NKp46.5

Human ETS1 ChIP peaks were present in promotors or enhancers of NCR1, NCR2 and NCR3, that encode NKp46, NKp44 and NKp30, respectively (Table S1). This suggests that ETS1 directly regulates expression of these activating NK receptors. Additionally, the reduced ETS1 p27 NK cell functionality can be enforced by the altered expression of several transcription factors proven to be involved in functional maturation of NK cells, i.e. Aiolos, E4BP4, TXNIP, GATA3 and Tbet.

Murine Aiolos-/- NK cells display a hyperreactive response towards tumor cells in vivo, suggesting a negative regulatory role of Aiolos in NK activation.45 The enhanced AIOLOS expression in human

ETS1 p27 NK cells might therefore contribute to decreased cytotoxicity. Murine E4bp4-/- and

Txnip-/- NK cells show impaired cytotoxicity and Gata3 and Txnip are required for efficient IFN-γ production.23,27,39 Also, T-bet has been described to be essential for functional maturation of murine NK cells.28 Although TBET was repressed in ETS1 p27 NK cells, the decreased cytotoxicity of murine T-bet-/- NK cells has been attributed to the decreased production of perforin and granzyme B29, whereas both perforin and granzyme B were increased in human ETS1 p27 NK cells. Similarly, NK cells from Ets1-/- mice contain higher amounts of perforin and granzyme B.5

This study provides mechanistic insight in the role of ETS1 in human NK cell biology by showing that ETS1 loss-of-function results in altered expression of 8 key transcription factors that have

19 been linked to NK cell development and terminal differentiation in mouse and/or human.

Additionally, ETS1 loss-of-function leads to altered linked to increased apoptosis and decreased expression of activating NK receptors. Finally, cytotoxicity and IFN-γ production upon tumor cell contact are decreased.

The crucial role of ETS1 in development and function of not only B- or T- cells but also of NK cells, suggests that genetic aberrations within the gene give rise to immune-related diseases. Genome- wide association studies have identified single nucleotide polymorphisms in or near the ETS1 gene or in ETS1-specific binding motifs of target genes. Accordingly, ETS1 is regarded as a of susceptibility for a wide range of autoimmune and inflammatory disorders, including systemic lupus erythematosus, rheumatoid arthritis, psoriasis, multiple sclerosis, ulcerative colitis, type 1 diabetes and allergy.20,56 However, the causal relationship between ETS1 and these disorders is still a subject of future research.

20

Acknowledgments

This work was supported by grants from the Research Foundation - Flanders (FWO) (grants

G.0444.17N, 12N4515N, 1S45317N, 1S29317N, 1831317N) (G.L., S.T., S.W., L.K. and T.K., respectively) and Kinderkankerfonds (a non-profit childhood cancer foundation under Belgian law to P.V.V. and W.V.L.). The computational resources (Stevin Supercomputer Infrastructure) and services used in this work were provided by the VSC (Flemish Supercomputer Center), funded by Ghent University, FWO and the Flemish Government – department EWI.

We thank Imke Velghe for help with hESC maintenance and culture.

Authorship Contributions: S.T., S.W. and G.L. designed research, analyzed and interpreted data, and wrote the manuscript; S.T., S.W., L.K., E.P., E.V.A. and K.D.M. performed research; W.V.L.,

J.R. and L.T. performed bioinformatic analysis; M.A., P.M., F.V.N., T.K., P.V.V., T.T. and B.V. contributed vital new reagents or analytical tools; and S.T. and S.W. collected data and performed statistical analysis.

Disclosure of Conflicts of interests

The authors declare no competing financial interests.

The current affiliation for S.T. is Orionis Biosciences, Ghent, Belgium, and for K.D.M. is AZ Sint-

Lucas, Ghent, Belgium.

Correspondence: Georges Leclercq, C. Heymanslaan 10, MRB2 entrance 38, 9000 Ghent, Belgium e-mail: [email protected]

21

References

1. Vivier E, Raulet DH, Moretta A, et al. Innate or adaptive immunity? The example of natural killer cells. Science. 2011;331(6013):44-49. 2. Ruggeri L, Capanni M, Urbani E, et al. Effectiveness of donor natural killer cell alloreactivity in mismatched hematopoietic transplants. Science. 2002;295(5562):2097-2100. 3. Barrow AD, Colonna M. Exploiting NK cell surveillance pathways for cancer therapy. Cancers (Basel). 2019;11(1):E55. 4. Barton K, Muthusamy N, Fischer C, et al. The Ets-1 transcription factor is required for the development of natural killer cells in mice. Immunity. 1998;9(4):555-563. 5. Ramirez K, Chandler KJ, Spaulding C, et al. Gene deregulation and chronic activation in natural killer cells deficient in the transcription factor ETS1. Immunity. 2012;36(6):921-932. 6. Russell L, Garrett-Sinha LA. Transcription factor Ets-1 in cytokine and chemokine gene regulation. Cytokine. 2010;51(3):217-226. 7. Koizumi S, Fisher RJ, Fujiwara S, et al. Isoforms of the human ets-1 protein: generation by alternative splicing and differential phosphorylation. Oncogene. 1990;5(5):675-681. 8. Laitem C, Leprivier G, Choul-Li S, et al. Ets-1 p27: a novel Ets-1 isoform with dominant-negative effects on the transcriptional properties and the subcellular localization of Ets-1 p51. Oncogene. 2009;28(20):2087-2099. 9. Costa M, Dottori M, Sourris K, et al. A method for genetic modification of human embryonic stem cells using electroporation. Nat Protoc. 2007;2(4):792-796. 10. Mali P, Yang L, Esvelt KM, et al. RNA-guided engineering via Cas9. Science. 2013;339(6121):823-826. 11. Cho SW, Kim S, Kim JM, Kim JS. Targeted genome engineering in human cells with the Cas9 RNA-guided endonuclease. Nat Biotechnol. 2013;31(3):230-232. 12. Cichocki F, Miller JS. In vitro development of human killer-immunoglobulin receptor- positive NK cells. Methods Mol Biol. 2010;612:15-26. 13. Taghon T, De Smedt M, Stolz F, Cnockaert M, Plum J, Leclercq G. Enforced expression of GATA-3 severely reduces human thymic cellularity. J Immunol. 2001;167(8):4468-4475. 14. Bryceson YT, Fauriat C, Nunes JM, et al. Functional analysis of human NK cells by flow cytometry. In: Campbell KS, ed. Natural Killer Cell Protocols: Cellular and Molecular Methods. Totowa, NJ: Humana Press; 2010:335-352. 15. Zhang Y, Liu T, Meyer CA, et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol. 2008;9(9):R137. 16. Heinz S, Benner C, Spann N, et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol Cell. 2010;38(4):576-589. 17. Woll PS, Martin CH, Miller JS, Kaufman DS. Human embryonic stem cell-derived NK cells acquire functional receptors and cytolytic activity. J Immunol. 2005;175(8):5095-5103. 18. Vanhee S, De Mulder K, Van Caeneghem Y, et al. In vitro human embryonic stem cell hematopoiesis mimics MYB-independent yolk sac hematopoiesis. Haematologica. 2015;100(2):157-166. 19. Zhao J-y, Osipovich O, Koues OI, Majumder K, Oltz EM. Activation of Mouse Tcrb: Uncoupling RUNX1 Function from Its Cooperative Binding with ETS1. J Immunol. 2017;199(3):1131-1141.

22

20. Koues OI, Collins PL, Cella M, et al. Distinct gene regulatory pathways for human innate versus adaptive lymphoid cells. Cell. 2016;165(5):1134-1146. 21. Janky R, Verfaillie A, Imrichova H, et al. iRegulon: from a gene list to a gene regulatory network using large motif and track collections. PLoS Comput Biol. 2014;10(7):e1003731. 22. Freud AG, Yokohama A, Becknell B, et al. Evidence for discrete stages of human natural killer cell differentiation in vivo. J Exp Med. 2006;203(4):1033-1043. 23. Gascoyne DM, Long E, Veiga-Fernandes H, et al. The basic transcription factor E4BP4 is essential for natural killer cell development. Nat Immunol. 2009;10(10):1118- 1124. 24. Male V, Nisoli I, Kostrzewski T, et al. The transcription factor E4bp4/Nfil3 controls commitment to the NK lineage and directly regulates Eomes and Id2 expression. J Exp Med. 2014;211(4):635-642. 25. Kamizono S, Duncan GS, Seidel MG, et al. Nfil3/E4bp4 is required for the development and maturation of NK cells in vivo. J Exp Med. 2009;206(13):2977-2986. 26. Firth MA, Madera S, Beaulieu AM, et al. Nfil3-independent lineage maintenance and antiviral response of natural killer cells. J Exp Med. 2013;210(13):2981-2990. 27. Lee KN, Kang HS, Jeon JH, et al. VDUP1 is required for the development of natural killer cells. Immunity. 2005;22(2):195-208. 28. Townsend MJ, Weinmann AS, Matsuda JL, et al. T-bet Regulates the Terminal Maturation and Homeostasis of NK and Vα14i NKT Cells. Immunity. 2004;20(4):477-494. 29. Gordon SM, Chaix J, Rupp LJ, et al. The transcription factors T-bet and Eomes control key checkpoints of natural killer cell maturation. Immunity. 2012;36(1):55-67. 30. Daussy C, Faure F, Mayol K, et al. T-bet and Eomes instruct the development of two distinct natural killer cell lineages in the liver and in the bone marrow. J Exp Med. 2014;211(3):563-577. 31. Madera S, Geary CD, Lau CM, Pikovskaya O, Reiner SL, Sun JC. Cutting Edge: Divergent requirement of T-Box transcription factors in effector and memory NK Cells. J Immunol. 2018;200(6):1977-1981. 32. Harmon C, Robinson MW, Fahey R, et al. Tissue-resident Eomes(hi) T-bet(lo) CD56(bright) NK cells with reduced proinflammatory potential are enriched in the adult human liver. Eur J Immunol. 2016;46(9):2111-2120. 33. Stegmann KA, Robertson F, Hansi N, et al. CXCR6 marks a novel subset of T- bet(lo)Eomes(hi) natural killer cells residing in human liver. Sci Rep. 2016;6:26157. 34. Collins A, Rothman N, Liu K, Reiner SL. Eomesodermin and T-bet mark developmentally distinct human natural killer cells. JCI Insight. 2017;2(5):e90063. 35. Cortez VS, Fuchs A, Cella M, Gilfillan S, Colonna M. Cutting Edge: Salivary gland NK cells develop independently of Nfil3 in steady-state. J Immunol. 2014:4487-4491. 36. Boulenouar S, Doisne JM, Sferruzzi-Perri A, et al. The residual innate lymphoid cells in NFIL3-deficient mice support suboptimal maternal adaptations to pregnancy. Front Immunol. 2016;7:43. 37. Seillet C, Huntington ND, Gangatirkar P, et al. Differential requirement for Nfil3 during NK cell development. J Immunol. 2014;192(6):2667-2676. 38. Simonetta F, Pradier A, Roosnek E. T-bet and Eomesodermin in NK Cell Development, Maturation, and Function. Front Immunol. 2016;7:241.

23

39. Ali AK, Oh JS, Vivier E, Busslinger M, Lee SH. NK cell-specific ablation identifies the maturation program required for bone marrow exit and control of proliferation. J Immunol. 2016;196(4):1753-1767. 40. Samson SI, Richard O, Tavian M, et al. GATA-3 promotes maturation, IFN-gamma production, and liver-specific homing of NK cells. Immunity. 2003;19(5):701-711. 41. Vosshenrich CA, Garcia-Ojeda ME, Samson-Villeger SI, et al. A thymic pathway of mouse natural killer cell development characterized by expression of GATA-3 and CD127. Nat Immunol. 2006;7(11):1217-1224. 42. Mackay LK, Minnich M, Kragten NA, et al. Hobit and Blimp1 instruct a universal transcriptional program of tissue residency in lymphocytes. Science. 2016;352(6284):459-463. 43. Post M, Cuapio A, Osl M, et al. The transcription factor ZNF683/HOBIT regulates human NK-cell development. Front Immunol. 2017;8:535. 44. Kallies A, Carotta S, Huntington ND, et al. A role for Blimp1 in the transcriptional network controlling natural killer cell maturation. Blood. 2011;117(6):1869-1879. 45. Holmes ML, Huntington ND, Thong RP, et al. Peripheral natural killer cell maturation depends on the transcription factor Aiolos. EMBO J. 2014;33(22):2721-2734. 46. Lee GM, Donaldson LW, Pufall MA, et al. The structural and dynamic basis of Ets-1 DNA binding autoinhibition. J Biol Chem. 2005;280(8):7088-7099. 47. Pufall MA, Graves BJ. Autoinhibitory domains: modular effectors of cellular regulation. Annu Rev Cell Dev Biol. 2002;18:421-462. 48. Xu D, Wilson TJ, Chan D, et al. Ets1 is required for transcriptional activity in UV- induced apoptosis in embryonic stem cells. Embo J. 2002;21:4081-4093. 49. Teruyama K, Abe M, Nakano T, et al. Role of transcription factor Ets-1 in the apoptosis of human vascular endothelial cells. J Cell Physiol. 2001;188:243-252. 50. Zhang C, Kavurma MM, Lai A, Khachigian LM. Ets-1 protects vascular smooth muscle cells from undergoing apoptosis by activating p21WAF1/Cip1: ETS-1 regulates basal and inducible p21WAF1/Cip1 transcription via distinct CIS-acting elements in the p21WAF1/Cip1 promoter. J Biol Chem. 2003;278:27903-27909. 51. Dittmer J. The role of the transcription factor Ets1 in carcinoma. Semin Cancer Biol. 2015;35(Supplement C):20-38. 52. Muthusamy N, Barton K, Leiden JM. Defective activation and survival of T cells lacking the Ets-1 transcription factor. Nature. 1995;377:639-642. 53. Bories JC, Willerford DM, Grevin D, et al. Increased T-cell apoptosis and terminal B-cell differentiation induced by inactivation of the Ets-1 proto-oncogene. Nature. 1995;377:635-638. 54. Bauer S, Groh V, Wu J, et al. Activation of NK cells and T cells by NKG2D, a receptor for stress-inducible MICA. Science 1999;285(5428):727-729. 55. Pende D, Parolini S, Pessino A, et al. Identification and molecular characterization of NKp30, a novel triggering receptor involved in natural cytotoxicity mediated by human natural killer cells. J Exp Med. 1999;190(10):1505-1516. 56. Garrett-Sinha LA, Kearly A, Satterthwaite AB. The Role of the Transcription Factor Ets1 in Lupus and Other Autoimmune Diseases. Crit Rev Immunol. 2016;36(6):485-510.

24

Figure legends

Figure 1. ETS1 deletion in human embryonic stem cells severely impairs NK cell development.

(A) Three homozygous ETS1 KO (ETS1-/-) hESC clones were generated using the CRISPR/Cas9 technology to introduce mutations in exon A/1 by non-homologous end joining. The wild type

(WT) clone underwent the same selection processes. The sequence of the targeted region of exon

A/1 is indicated for both alleles, showing bi-allelic frameshifts in the KO clones. (B) WT and mutated hESCs were differentiated into hematopoietic CD34+CD43+ progenitor cells. Cell numbers are indicated (mean ± SEM; n=4-6). (C-D) CD34+CD43+ cells were subsequently cultured in NK cell differentiating conditions. (C) The ratio of the cell numbers of generated NK cells (gated as CD45+CD56+CD94+) over the input precursor stem cell number was determined (mean ± SEM; n=4-6) on d14 and d21. (D) Representative dot plots of d21 cultures are shown, with the indicated

NK cell percentage in the CD45+ gated population. * indicates a significant difference compared to the WT clone, p<0.05 (student’s t-test).

Figure 2. ETS1 loss-of-function in cord blood HPC inhibits NK cell differentiation.

(A) Upon transduction with control, ETS1 p27 or ETS1 p51 vectors, sorted CB CD34+Lin-eGFP+ precursor cells were cultured in NK differentiating conditions. Absolute cell numbers of NK cells were determined at the indicated time points. (B) Absolute cell numbers of differentiation stage

1 to stage 3 were determined at d14. (A, B) Data are presented as mean ± SEM, and the number of experiments is indicated. * indicates a significant difference compared to control, p<0.05

(student’s t-test).

25

Figure 3. ETS1 directly regulates expression of several key transcription factors linked to NK cell differentiation.

(A) NK cells were sorted from d18 control and ETS1 p27 cultures, and RNA-sequencing was performed. Volcano plot showing down- (blue) and upregulated (red) genes in NK cells from ETS1 p27 versus control (n=4). Differentially expressed NK-related transcription factors are indicated.

(B-C) NK-linked transcription factors that were differentially expressed in (A) were evaluated by qPCR analysis in all stages. Data are presented as mean ± SEM, the number of experiments is indicated. (D) Representative dot plots of flow cytometry analysis of TBET and EOMES expression in mature NK cells. The mean MFI (mean fluorescence intensity) ± SEM (n=8) is shown. (B,C,D) * and *** indicates a significant difference compared to control of p<0.05 and p<0.001, respectively (student’s t-test).

Figure 4. ETS1 ChIP analysis of human ex vivo peripheral NK cells.

(A-D) ETS1 ChIP-seq analysis was performed on sorted human peripheral blood NK cells. (A)

Known motifs were identified with the HOMER package and the 4 most significant motifs are shown. (B) Locations of ETS1 ChIP peaks relative to genomic annotations. (C) Genome browser tracks of NK cell-linked transcription factor genes are shown for human NK cell ETS1 ChIP-seq versus input DNA, versus H3K4me3 and H3K27ac ChIP-seq and ATAC-seq. (D) Identification of the proportion of genes that are potentially directly regulated by ETS1 (E) Non-biased analysis of enriched transcription factor motifs in the down- and upregulated genes of ETS1 p27 NK cells by using the computational method iRegulon.

26

Figure 5. ETS1 p27 expression increases apoptosis in NK cells.

(A) NK cells were sorted from d14 control and ETS1 p27 cultures. qPCR results are shown for the relative expression of indicated pro- and anti-apoptotic genes (mean ± SEM, n=3-7). (B) NK cells from control and ETS1 p27 cultures were analyzed at d14 for apoptosis using annexin V and propidium iodide (PI) staining. The percentage annexin V and/or PI positive cells in the NK cell gate is shown (mean ± SEM; n=3-4). (C) DAVID analysis of up- and downregulated genes of the

RNA-seq analysis of ETS1 p27 vs control NK cells. The 10 most significant pathways are shown.

(D) The number of up- and downregulated genes of ETS1 p27 vs control NK cells in the Gene

Ontology positive and negative regulation of the mitotic cell cycle pathways. (E) Cell cycle analysis of d18 NK cells by 7-AAD/Ki67 staining. The percentage of cells at the indicated cell cycle phases is shown (n=4). (A, B, F) * indicates a significant difference compared to control, p<0.05 (student’s t-test).

Figure 6. ETS1 influences NK cell terminal differentiation and function.

(A, B) Flow cytometric expression of the indicated receptors on gated NK cells from d21 control and ETS1 p27 cultures (mean ± SEM, n=5-15). (C) Day 18-21 NK cells were stimulated with an equal number of K562 target cells and CD107a expression was determined (mean ± SEM, n=8).

(D) The cytotoxic capacity of sorted d18-21 NK cells against K562 targets was determined in a 51Cr release assay. Data are expressed as the percentage specific lysis as a function of the effector:target ratio (mean ± SEM, n=6). (E) The MFI of perforin and granzyme B expression in control and ETS1 p27 NK cells (n=6). (F) Fas ligand expression of NK cells that were unstimulated

(US) or stimulated for 2 h with K562 cells (n=5). (G) Cells from d21 control and ETS1 p27 cultures were stimulated with K562 cells for 6 h, or with IL-12/IL-18 or IL-12/IL-15/IL-18 for 24 h and IFN-

27

γ production was determined by flow cytometry in gated NK cells (mean ± SEM, n=4-12). (H)

Sorted d21 NK cells were stimulated with IL-12/IL-18 or IL-12/IL-15/IL-18 for 24 h and secreted

IFN-γ was measured by ELISA. The relative IFN-γ production compared to control is shown (mean

± SEM, n=6-7). *, ** and *** indicate a significant difference compared to control of p<0.05, p<0.01 and 0.001, respectively (student’s t-test).

28

Figure 1

A B CD34+CD43+ C NK / input HSC

150 * ) 3 WT 16.1 100 ETS1-/-15.3 WT 16.1

ETS1-/-15.3 ratio ETS1-/-17.3 50 ETS1-/-17.3 * -/- ETS1-/-21.1 no. (x10 Absolute ETS1 21.1 0 d 0 d 14 d 21

D WT 16.1 ETS1-/-15.3 ETS1-/-17.3 ETS1-/-21.1

20 + 3.7 % 2.1 *+ 0.9 % 0.5 *+ 0.3 % 4.0 *+ 0.2 % CD56

CD94 Figure 2 A NK cells d14 (n=8) d21 (n=11-15) 8000 * 500000 6000 400000 300000 4000 * 200000 2000 100000 Absolute cell number Absolute 0 cell number Absolute 0 ctrl p27 p51 ctrl p27 p51

B Stage 1 Stage 2 Stage 3 (n=11) (n=11) (n=11) 1000 15000 * 800 10000 600

400 5000 200 Absolute cell number Absolute cell number Absolute 0 0 ctrlp27 ctrl p27 ctrl p27 Figure 3 A B HOBIT TBET 0.08 GATA3 * 0.05 0.08 * 0.04 (n=2-5) 0.04 (n=3-4) 0.04 (n=3-4) 0.03 * 0.005 0.005 30 DOWN: 690 UP: 488 0.02 0.004 0.01 * 0.004 0.003 0.003 0.0006 0.002 0.002 * 0.0004 Relative expression 0.001 * * 0.0002 0.001 20 0.000 0.0000 0.000 stage 1 stage 2 stage 3 mNK stage 1 stage 2 stage 3 mNK stage 1 stage 2 stage 3 mNK

HOBIT 0.8 TXNIP 0.04 E4BP4 BLIMP1 * 0.04 (n=3-7) (n=3-11) * (n=3-4) * 0.03 -log10(padj) 0.6 0.03 0.02 10 TXNIP AIOLOS TBET 0.4 0.02 0.01 E4BP4 0.0020 GATA3 0.0015 EOMES 0.2 0.01 0.0010 Relative expression 0.0005 0 BLIMP1 0.0 0.00 0.0000 stage 1 stage 2 stage 3 mNK stage 1 stage 2 stage 3 mNK stage 1 stage 2 stage 3 mNK -5 0 +5 ctrl p27 log2FoldChange C D ctrl p27 AIOLOS EOMES ctrl 0.25 (n=3-7) * 0.15 (n=2-3) p27 4000 *** 0.20 0.15 relative expression3000 0.10 0.10 *** 2000 0.005 MFI 0.004 * 0.05 Counts Counts 0.003 1000 0.002 Relative expression 0.001 0.000 0.00 0 stage 1 stage 2 stage 3 mNK stage 1 stage 2 stage 3 mNK TBET EOMES TBET EOMES

NFIL3

GATA3-AS GATA3 Figure 4

A B promoter ETS1 (p = 10-220)

ETS1:RUNX (p = 10-148) intergenic

RUNX (p = 10-52) intron

TBET(p = 10-6) exon

C 31 kb 35 kb

ETS1 [0-41] ETS1 [0-17] ATAC [0-40] ATAC [0-24] H3K27ac [0-33] H3K27ac [0-42] H3K4me3 [0-51] H3K4me3 [0-59] input [0-41] input [0-17] AIM1L HOBIT E4BP4

31 kb 68 kb

[0-14] [0-22] ETS1 ETS1 relative expression ATAC [0-42 ATAC [0-32] H3K27ac [0-54] H3K27ac [0-43] H3K4me3 [0-108] H3K4me3 [0-73] input [0-14] input [0-22] TXNIP GATA3-AS GATA3

42 kb 43 kb

ETS1 [0-37] ETS1 [0-16] ATAC [0-28] ATAC [0-35] H3K27ac [0-56] H3K27ac [0-54] H3K4me3 [0-89] H3K4me3 [0-82] input [0-37] input [0-16] TBET BLIMP1

21 kb 142 kb

ETS1 [0-25] ETS1 [0-40] NFIL3 ATAC [0-42] ATAC [0-42] H3K27ac [0-45] H3K27ac [0-60] H3K4me3 [0-88] H3K4me3 [0-70] input [0-25] input [0-40] EOMES AIOLOS ZPBP2

D E Motif clusters in DOWNregulated genes Motif clusters in UPregulatedGATA3-AS genesGATA3 Figure 5

A Pro-apoptotic Anti-apoptotic B ctrl p27 0.03 0.20 25 * * 20 0.15 0.02 * 15 0.10 * * 10 0.01 Percentage 0.05 5 Relative expression 0.00 0.00 0 annexin V + + PIM2

BIRC3 PI - + PMAIP1 TSC22D3 C RNA-seq: UP RNA-seq: DOWN

Mitosis Ribosome DNA repair Transmembrane Proteasome Immunoglobin Cyclin SH3 domain Cell cycle Ig family Lectin CH domain NK mediated cytotoxicity ATP binding RGS domain Tubulin Kinase DNA binding TNF signaling 0 2 4 6 0 5 10 15 Enrichment score Enrichment score

D Mitotic gene regulation E 30 80 ctrl 28 * 20 p27 60 10 9

UP 40 0

2 Percentage 10 12 20 * Differential genes DOWN 20 0 G0 G1 S G2/M regulation regulation Negative Positive Figure 6 A B ctrl p27

80 100 100 20

** 80 80 * 60 60 60 40 40 15 20 20 *** Counts 0 0 -103 0 103 104 105 -103 0 103 104 105 40 NKG2D NKp30 10

100 100

80 80 % expression % expression 20 60 60 5

40 40

20 20 Counts

0 0 0 0 103 104 105 0 103 104 105 0 NKG2D NKp30 NKp44 NKp46 KIR2DL1 KIR2DL2 KIR3DL1 KIR2DS4 NKp44 NKp46 KIR2DS1 KIR2DL3 KIR3DS1 C D E 80 100 * 15000 * 80 60 *

* Counts 10000 60 40 * MFI Perforin 40 5000 % expression 20 % specific lysis ctrl 20 p27 0 0 0 Counts CD107a 0.3:1 1:1 3:1 10:1 Perforin GzmB effector:target GzmB

0 F G H US K562 6h 24h 100 15 100 8 ctrl 80 p27 * 80 6 10 60 NK

+ 60 4 * 40 40 5 % expression 2

20 % IFN- γ 20 0 0 0

0 Relative IFN- γ secretion FasL K562 IL-12 IL-12 IL-12 IL-12 IL-18IL-18 IL-18 IL-18 IL-15 IL-15

Supplemental Methods

ETS1-deficient human embryonic stem cells and hematopoietic differentiation of hESCs hESC usage in this study was approved by the Ethics Committee of the Faculty of Medicine and

Health Sciences. Single cell suspensions were made using TrypLE select (Gibco, Life Technologies,

Grand Island, NY, U.S.A) and cultured on mouse embryonic feeders in the presence of 10 µM

Rock inhibitor (Y-27632, Selleckhem, Munich, Germany; the first 24 h after split). Cultures were split every 3 days and single cell adapted hESCs were used for nucleofection at single cell passage

4 to 10.

To generate ETS1-deficient hESCs, hETS1 guideRNA (gRNA) was designed using the software available at http://crispr.mit.edu from the Zhang lab (Massachusetts Institute of Technology,

Cambridge, MA, U.S.A.). The sequence of ETS1 gRNA was: AAGACGGAAAAAGTCGATC. This gRNA was used to target exon A/1. The gRNA was inserted in the gRNA_Cloning vector (#41824,

Addgene, Teddington, U.K.) using the Gibson Assembly Master Mix (New England Biolabs,

Ipswich, MA, U.S.A.) and nucleofected together with the pCas9_GFP vector (#44719, Addgene).

Nucleofection in the Jurkat cell line was performed using the Neon Transfection System and Neon kit 100 µl (Invitrogen, U.K.), after which eGFP positive cells were single cell sorted and cultured in complete Iscove's Modified Dulbecco's Medium (IMDM) containing 10% FCS (Life

Technologies, Leiden, The Netherlands). Single cell adapted ESCs were nucleofected using the

Nucleofector 2b Device (Lonza, Basel, Switzerland) and the Human Stem Cell Nucleofector kit 2

(Lonza). Nucleofected hESCs were seeded at low densitiy on mouse embryonic feeders in 6-well plates in the presence of 10 µM Rock inhibitor (first 24 h only) and 10 ng/ml basic Fibroblast

Growth Factor (bFGF, Peprotech). 7 to 14 days after nucleofection, individual clones were

1 transferred onto mouse embryonic feeders in 24-well plates and gDNA was isolated for mutation analysis using the Surveyor Mutation Detection Kit (Transgenomic, Omaha, NE, U.S.A.).

Subsequent TA-cloning (pGEMT Vector System, Promega, Madison, WI, U.S.A.) and Sanger sequencing (BigDye Terminator v1.1 Cycle sequencing kit, Thermo Fisher Scientific, Waltham,

MA, U.S.A.) was used to select clones that carried a frameshift mutation on both alleles. Selected clones were single cell sorted after a second round of single cell adaptation to ensure clonality and expanded in the presence of 10 µM Rock inhibitor (first 24 h) and 40 ng/ml bFGF until stable colonies were established. gDNA was isolated and the target region was amplified with the

Expand High Fidelity PCR System (Roche, Bazel, Switzerland) followed by TA-cloning and Sanger sequencing.

To differentiate hESCs into hematopoietic stem cells, 5x103 single cell adapted hESCs in APEL medium containing 10 µM Rock inhibitor, 40 ng/ml SCF, 2 ng/ml BMP4 (R&D Systems) and 20 ng/ml VEGF165 (Peprotech) were transferred to each well of a 96-well low attachment plate, spun at 480 g for 7 min and cultured at 37°C, 5% CO2. After 4 days, spin embryoid bodies (EB) were transferred onto OP9 feeder cells (kindly provided by J. C. Zúñiga-Plfücker, University of

Toronto, Toronto, Canada) on 0.1% gelatin and further cultured for an additional week, for a total of 11 days of differentiation culture. Half the medium was refreshed on day 7. Spin EBs were dissociated at day 11 by forceful pipetting and subsequent incubation with 0.1% collagenase type

IV (Gibco, Life Technologies) and trypsin-EDTA (17.000 U/L, Lonza). Cells obtained from one 96- well plate were seeded in twelve 24-well plates that contained mitomycin C-inactivated (Sigma-

Aldrich, St. Gallen, Switzerland) EL08-1D2 cells (kindly provided by E. Dzierzak, Erasmus University

MC, Rotterdam, the Netherlands), a mouse embryonic liver cell line. At this point, the amount of

2 hematopoietic precursor cells was determined by CD34/CD43/CD45 staining. Cells were co- cultured in Dulbecco’s modified Eagle medium plus Ham’s F-12 medium (2:1 ratio), supplemented with 100 U/mL penicillin, 100 μg/mL streptomycin, 2 mM glutamine, 10 mM sodium pyruvate (all from Life Technologies), 20% of heat-inactivated human AB serum (Merck,

Darmstadt, Germany), 24 μM β-mercaptoethanol, 20 μg/mL ascorbic acid and 50 ng/mL sodium selenite (all from Sigma-Aldrich). The following cytokines were added: IL-3 (5 ng/mL, first week only), IL-7 (20 ng/mL), IL-15 (10 ng/mL) (all from R&D Systems), SCF (20 ng/mL), and Flt3-L (10 ng/mL). Culture medium was refreshed on day 7 by addition of the same volume of fresh medium with cytokines, and split and transferred to new inactivated EL08-1D2 feeder cells on day 14 and day 21.

EL08-1D2 cells were maintained in 50% Myelocult M5300 medium (Stem Cell

Technologies, Grenoble, France), 35% α-MEM, 15% FCS, supplemented with 100 U/mL penicillin,

100 μg/mL streptomycin, 2 mM glutamine and 10 μM β-mercaptoethanol on 0.1% gelatin-coated plates at 32°C. Cell proliferation was blocked by adding 10 µg/ml mitomycin C to the culture medium during 2-3 h. Thereafter, cells were thoroughly rinsed before harvesting using trypsin-

EDTA. Cells were plated at a density of 50.000 cells per well of a 0.1% gelatin-coated tissue culture-treated 24-well plate at least 24 h before adding HPCs or differentiating NK.

Umbilical cord blood HPC differentiation co-culture

CD34+ HPCs were isolated from umbilical CB (Cord Blood Bank, University Hospital Ghent). Cord blood usage in this study was approved by the Ethics Committee of the Faculty of Medicine and

Health Sciences (Ghent University, Ghent, Belgium) and informed consent was obtained in accordance with the Declaration of Helsinki. Mononuclear cells were isolated by Lymphoprep

3 density gradient centrifugation, followed by CD34+ MACS microbead (Miltenyi Biotec, Leiden, The

Netherlands) enrichment according to the manufacturer’s guidelines. For viral transduction,

CD34+ precursors were cultured in complete Iscove's Modified Dulbecco's Medium (IMDM) containing 10% FCS (all from Life Technologies) and supplemented with thrombopoietin (TPO)

(20 ng/ml), stem cell factor (SCF) (100 ng/ml) (all from Peprotech) and FMS-like tyrosine kinase 3 ligand (FLT3-L) (100 ng/ml, R&D Systems) during 48 h. Subsequently, pre-cultured CD34+ cells were transduced using RetroNectin (Takara Bio, Saint-Germain-en-Laye, France)-coated plates in case of retroviral infection, and using RetroNectin-coated plates combined with the addition of polybrene (8 µg/ml, Sigma-Aldrich) in case of lentiviral infection. In both cases, a 950 g at 32°C spin during 90 min was performed. Additional cytokines were added to keep the concentration constant after virus addition. 24 h after transduction, the medium in polybrene-containing wells was replaced with medium without polybrene. 48 h after transduction, lineage-

+ + (CD3/CD14/CD19/CD56) CD34 eGFP cells were sorted using a FACS ARIA III cell sorter (BD

Biosciences, San Jose, CA, U.S.A.). Sorted HPCs were co-cultured with mitomycin C-inactivated

EL08-1D2 as described in hematopoietic differentiation of hESCs.

Western blot analysis

Cells were lysed in Laemmli buffer (1 M Tris pH 6.8, 2% (w/v) SDS, 10% (v/v) glycerol) and protein concentration was determined using the DC protein assay (Biorad, Temse, Belgium).

Bromophenol blue (0.1% (w/v)) and Bolt Reducing agent (10% (v/v)) (Thermo Fisher Scientific) were added to 20 µg of protein, and the samples and a protein ladder (WesternSure Pre-stained chemiluminescent protein ladder, Licor Biotechnology, Bad Homburg, Germany) were separated on a Bolt 4-12% Bis-Tris Plus gel (Thermo Fisher Scientific) and transferred to a PVDF membrane

4

(Invitrogen). After blocking, the membrane was incubated with a primary antibody. Figure S1 was constructed with two separate blots for ETS1 and Vinculin. The membranes were treated with either monoclonal mouse anti-human ETS1 (Santa Cruz Biotechnology, Heidelberg, Germany, clone C-4) or with mouse anti-human Vinculin (Sigma-Aldrich). Thereafter, the membranes were incubated with horse radish peroxidase linked anti-mouse IgG (Sigma-Aldrich). For figure S2, the expression of ETS1 p27 and β-actin was shown on the same blot. Sequential incubation was performed with polyclonal rabbit anti-human ETS1 (Santa Cruz Biotechnology, clone C-20) and the secondary antibody horse radish peroxidase linked anti-rabbit IgG (Cell Signaling Technology,

Leiden, The Netherlands). To detect β-actin, the membrane was stripped using the Restore

Western Blot Stripping Buffer (Thermo Fisher Scientific), followed by incubation with β-actin loading control monoclonal antibody (Thermo Fisher Scientific, clone BA3R) and horse radish peroxidase linked anti-mouse IgG (Sigma-Aldrich).

Flow cytometry analysis and sorting

The following antibodies were used in the flow cytometric analysis and/or sorting: anti-CD3 allophycocyanin (APC)-conjugated, clone SK7; anti-CD11a phycoerythrin (PE)-conjugated, clone

Hl111; anti-CD16 PE, clone Leu11c; anti-CD43 PE, clone 1G10; anti-CD45RA APC, clone TD611; anti-CD56 V450, clone B159; anti-CD94 peridinin chlorophyll protein/Cy5.5-conjugated, clone

HP-3D9; anti-CD107a PE, clone H4A3; anti-CD335 (anti-NKp46) PE/Cy7, clone 9E2 (all from BD

Biosciences/BD Pharmingen); anti-CD11a APC, clone Hl111; anti-CD45 APC/Fire750, clone 2D1

(all from Biolegend, Koblenz, Germany); anti-CD14 APC, clone TUK4; anti-CD34 Fluorescein isothiocyanate (FITC), PE or VioBlue, clone AC136 (all from Miltenyi Biotec); anti-CD19 APC, clone

SJ25C1; anti-CD117 PE/Cy7, clone 104D2; anti-CD336 (anti-NKp44) APC, clone P44.189 or eFluor

5

450, clone 48.189; anti-eomes PE or eFluor 660, clone WD1928; anti-GATA3 PE, clone TWAJ; anti- granzymeB PE, clone BG11; anti-IFN-γ eFluor 660, clone 4SB3; anti-Ki-67 PE, clone SolA15; anti- perforin PE, clone dG9; anti-T-bet eFluor 660, clone ebio4B10; Annexin V apoptosis detection kit

APC (all from eBioscience/Thermo Fisher Scientific); anti-CD158b1/b2 PE (KIR2DL2/KIR2DL3), clone GL183; anti-CD158a/h PE (KIR2DL1/KIR2DS1), clone EB6.B; anti-CD158e1/e2 PE

(KIR3DL1/KIR3DS1), clone Z27.3.7; anti-CD158i PE (KIR2DS4), clone FES172; anti-CD314 (anti-

NKG2D), clone ON72 (all from Beckman Coulter, Brea, CA, U.S.A); anti-ETS1, clone 8A8 (the epitope recognized by the antibody is absent in the ETS1 p27 isoform) (Novus Biologicals,

Littleton, CO, U.S.A), in house labeled with Dylight 650 (Thermo Fisher Scientific).

Before staining, cells were blocked with anti-mouse FcγRII/III (unconjugated, clone 2.4G2, kindly provided by Dr J. Unkeless, Mount Sinai School of Medicine, New York, NY) and FcR blocking reagent (Miltenyi Biotec). Propidium iodide or the LIVE/DEAD Fixable Aqua Dead Cell Stain Kit

(Life Technologies) was used to discriminate live and dead cells. Transcription factor staining and cell cycle analysis was performed using the FoxP3/Transcription Factor Staining buffer set

(eBioscience).

RNA-seq and ChIP-seq analysis

After RNA extraction (RNeasy micro kit, Qiagen, Hilden, Germany), concentration and quality of the total extracted RNA was checked by using the 'Quant-it ribogreen RNA assay' (Life

Technologies, Grand Island, NY, U.S.A) and the RNA 6000 nano chip (Agilent Technologies, Santa

Clara, CA, U.S.A), respectively. Subsequently, 50 ng of RNA was used to perform an Illumina sequencing library preparation using the QuantSeq 3' mRNA-Seq Library Prep Kits (Lexogen,

Vienna, Austria) according to manufacturer's protocol. Libraries were quantified by qPCR,

6 according to Illumina's protocol 'Sequencing Library qPCR Quantification protocol guide', version

February 2011. A high sensitivity DNA chip (Agilent Technologies, Santa Clara, CA, U.S.A.) was used to control the library's size distribution and quality. Sequencing was performed on a high throughput Illumina NextSeq 500 flow cell generating 75 bp single reads.

Per sample, on average 1.0 x 107 ± 1.3 x 106 reads were generated. Quality Control on fastq files was performed with FastQCv0.11.7 66. Fastq files were aligned to hg38 using STAR2.4.2a and genes (Gencode v25) were quantified on the fly. Differential expression analysis was done in R with DESeq2 with design formula replicates +condition 67. Genes with an FDR <0.1 were considered significantly differential.

To test whether differential genes are enriched in genes controlled by a common transcription factor, we used iRegulon. iRegulon is a computational method designed to reverse-engineer the transcriptional regulatory network underlying a co-expressed gene set using cis-regulatory sequence analysis. iRegulon implements a genome-wide ranking-and-recovery approach to detect enriched transcription factor motifs and their optimal sets of direct targets. In this study, we provided iRegulon with 2 gene sets, the significantly up and downregulated genes after ETS1 perturbation. The regulatory regions of these genes were scanned for transcription factor binding sites (motifs) and were subsequently scored and compared.

For ChIP-seq experiments mature NK cells were isolated from peripheral blood (Belgian

Red Cross, Mechelen, Belgium; usage was approved by the Ethics Committee of the Faculty of

Medicine and Health Sciences). Mononuclear cells from buffy coats were isolated by Lymphoprep density gradient centrifugation, followed by CD3 and CD19 cell removal by anti-CD3/19-biotine labeling and subsequent anti-biotin MACS depletion. Enriched NK cells were subsequently sorted

7 as CD56+CD3- (residual CD3+ and CD19+ cells were identified via streptavidin staining and propidium iodide was added to discriminate live and dead cells) using a BD FACSAria III cell sorter.

NK cells were fixed with 1% formaldehyde for 15 min and quenched with 0.125 M glycine.

Chromatin was isolated by the addition of lysis buffer, followed by disruption with a Dounce homogenizer. Lysates were sonicated on a microtip sonicator and the DNA sheared to an average length of 300-500 bp. Control genomic DNA (further referred to as input DNA) was prepared by treating aliquots of chromatin with RNase, proteinase K and heat for de-crosslinking, followed by ethanol precipitation. Pellets were resuspended and the resulting DNA was quantified on a

NanoDrop spectrophotometer. Extrapolation to the original chromatin volume allowed quantitation of the total chromatin yield.

An aliquot of chromatin (30 µg) was precleared with protein A agarose beads. Genomic

DNA regions of interest were isolated using 10 µl of antibody serum against ETS1. Complexes were washed, eluted from the beads with SDS buffer, and subjected to RNase and proteinase K treatment. Crosslinks were reversed by incubation overnight at 65° C, and ChIP DNA was purified by phenol-chloroform extraction and ethanol precipitation.

Quantitative PCR reactions were carried out in triplicate on specific genomic regions using

SYBR Green Supermix (Bio-Rad, Hercules, CA, U.S.A). The resulting signals were normalized for primer efficiency by carrying out QPCR for each primer pair using Input DNA.

Illumina sequencing libraries were prepared from the ChIP and Input DNAs by the standard consecutive enzymatic steps of end-polishing, dA-addition, and adaptor ligation. After a final PCR amplification step, the resulting DNA libraries were quantified and sequenced on Illumina’s

NextSeq 500 (75 nt reads, single end).

8

QC on fastq files was performed with FastQCv0.11.7. Fastq files were aligned to hg38 using

STAR2.4.2a with parameters --outFilterMultimapNmax 1 --alignIntronMax 1. Peak calling was done with MACS2 v2.1.0.

The ETS1 ChIP-seq peaks were aligned to H3K4Me3 and H3K27Ac ChIP-seq and ATAC-seq results for human peripheral NK cells 19 that were downloaded from GEO with accession number

GSE77299. For H3K4Me3 and H3K27Ac, peaks were called from bedGraph files with the MACS2 peak calling software 20. First, MACS2 bdgcmp was used to compare ChIP and input files.

Thereafter, peaks were called using MACS2 bdgpeakcall. Options for MACS2 bdgpeakcall were set to -c 2, -g 100 and -l 100 for H3K4Me3 or 150 for H3K27Ac. qPCR analysis

RNA was converted into cDNA using Superscript RT III (Life Technologies). Real-time PCR reactions were performed using the LightCycler 480 SYBR Green I Master mix (Roche) on a

LightCycler 480 real-time PCR system (Roche) using the following primers:

- BIRC3 FW: 5’- CCGTCAAGTTCAAGCCAGTT-3’

- BIRC3 RV: 5’- ATCTCCTGGGCTGTCTGATG-3’

- ETS1 FW: 5’-AGATGGCTGGGAATTCAAAC-3’

- ETS1 RV: 5’-TTCCTCTTTCCCCATCTCCT-3’

- IRF8 FW: 5’-ATGTGTGACCGGAATGGTGG-3’

- IRF8 RV: 5’-AGTCCTGGATACATGCTACTGTC-3’

- IKZF3 FW: 5’-GCTCATACAGACCCGCATGAT-3’

- IKZF3 RV: 5’- AACTGGAACCATCTCCGAGGT-3’

- NFIL3 FW: 5’-AAAATGCAGACCGTCAAAAAGGA-3’

9

- NFIL3 RV: 5’-TGACACTTCCGTTAAAGCAGAAT-3’

- PIM2 FW: 5’-TCACAGATCGACTCCAGGTG-3’

- PIM2 RV: 5’-ACCTGCACCCACTTTCCATA-3’

- PRDM1 FW: 5’- GTGTCAGAACGGGATGAACA-3’

- PRDM1 RV: 5’- GCTCGGTTGCTTTAGACTGC-3’

- TSC22D3 FW: 5’-CAAGATCGAACAGGCCATGG-3’

- TSC22D3 RV: 5’-TCTCCACCTCCTCTCTCACA-3’

- TXNIP FW: 5’-ATATGGGTGTGTAGACTACTGGG-3’

- TXNIP RV: 5’-GACATCCACCAGATCCACTACT-3’

- ZNF683 FW: 5’-GTCCCATTGAGTTCCCAGA-3’

- ZNF683 RV: 5’- GTGGGCAAGTTGAGTGAAG-3’

- GAPDH FW: 5’-TCCTCTGACTTCAACAGCGACA-3’

- GAPDH RV: 5’-GTGGTCGTTGAGGGCAATG-3’

- YWHAZ FW: 5’-ACTTTTGGTACATTGTGGCTTCAA-3’

- YWHAZ RV: 5’-CCGCCAGGACAAACCAGTAT-3’

Relative expression levels were calculated for each gene using YWHAZ and GAPDH for normalization.

10

Supplemental table 1: ETS1 ChIP-seq results of a selected gene set, Gene Gene Interval Interval Dists to Gene Name Gene Start Gene End Interval Pos Peak Value Altsymbols Length Strand Count Start ZNF683 1 26361632 26374533 12901 - 1 -3899 upstream 41 ZNF683, Hobit, MGC33414 TBX21 17 47733244 47746122 12878 + 3 132, 676, 8004 in gene, in gene, in gene 37 TBX21, TBET, T-bet, T-PET, TBLYM -8004, 4300, 5004, upstream, in gene, in gene, in HDR, MGC2346, MGC5445, MGC5199, GATA3, GATA3 10 8045428 8075201 29773 + 6 9356, 10492, 22 gene, in gene, in gene HDRS 16492 MGC118922, MGC118923, MGC118924, PRDM1 6 106086320 106109939 23619 + 1 11568 in gene 16 PRDM1, PRDI-BF1, BLIMP1, MGC118925 IKZF3 17 39757715 39864188 106473 - 2 109660, 44732 downstream, in gene 41 RP11-94L15,2, IKZF3, ZNFN1A3, AIO, AIOLOS EOMES 3 27715949 27722715 6766 - 3 1547, 411, -325 in gene, in gene, upstream 25 TBR2, EOMES PIM2 X 48913182 48919136 5954 - 2 7520, 181 downstream, in gene 35 PIM2 DIP, DKFZp313A1123, TSC22D3, DSIPI, TSC- TSC22D3 X 107713221 107777329 64108 - 2 60497, 34993 in gene, in gene 17 22R, GILZ, hDIP CIAP2, HIAP1, AIP1, RNF49, c-IAP2, API2, BIRC3 11 102317373 102339403 22030 + 2 3, 29795 in gene, downstream 31 HAIP1, MIHC, MALT2, BIRC3 KIAA0075, , MGC104598, E2F3 6 20401906 20493714 91808 + 1 -2 upstream 18 DKFZp686C18211, E2F-3 BRF1, RNF162B, TIS11B, ZFP36L1, ERF1, ZFP36L1 14 68787655 68796243 8588 - 3 6046, 2195, 211 in gene, in gene, in gene 41 cMG1, Berg36, ERF-1 7054, 2286, 1358, - downstream, in gene, in gene, ERF-2, RNF162C, ERF2, ZFP36L2, BRF2, ZFP36L2 2 43222402 43226606 4204 - 4 58 994 upstream TIS11D NCR1 19 54905992 54938208 32216 + 1 24 in gene 26 NCR1, FLJ99094, NK-p46, NKP46, LY94, CD335 dJ149M18,1, NKP44, NK-p44, CD336, NCR2, NCR2 6 41335790 41350887 15097 + 1 -1518 upstream 25 LY95 NCR3 6 31588883 31593019 4136 - 1 -37 upstream 28 CD337, MALS, NCR3, 1C7, LY117, NKp30 PFN1, MGC65093, PFP, FLH2, P1, PRF1, PRF1 10 70597348 70602775 5427 - 2 -73, -7913 upstream, upstream 63 HPLH2 CTLA1, CTSGL1, CGL-1, CSP-B, CCPI, CSPB, GZMB 14 24630948 24634226 3278 - 1 -142 upstream 27 HLP, SECT, GZMB, CGL1

SGA72M, FLJ45597, EXO4, MGC102768, SYTL2 11 85694221 85811159 116938 - 1 58103 in gene 37 PPP1R151, FLJ34802, KIAA1597, SLP2, SLP2A, FLJ42764, CHR11SYT, SYTL2, FLJ21219

MGC118883, MGC118885, SYTL3, MGC118884, SYTL3 6 158644899 158764876 119977 + 3 -339, 8797, 42141 upstream, in gene, in gene 54 MGC105130, SLP3 FLJ23544, DXS648E, AUTSX5, L10, NOV, RPL10 X 154398065 154402339 4274 + 1 335 in gene 23 FLJ27072, DXS648, DKFZp686J1851, RPL10, QM RPL10A 6 35468401 35470781 2380 + 1 143 in gene 28 Csa-19, RPL10A, NEDD6, L10A, CSA19 RPL12 9 127447674 127451432 3758 - 1 -56 upstream 29 RPL12, L12 FLJ27454, BBC1, FLJ38635, FLJ27453, RPL13, RPL13 16 89560657 89566829 6172 + 3 -177, 399, 943 upstream, in gene, in gene 28 MGC117342, D16S44E, L13, MGC71373, D16S444E upstream, downstream, RPL13A 19 49487554 49492308 4754 + 3 -114, 8462, 13262 26 L13A, RPL13A, TSTA1 downstream RPLY10, MGC88603, RPYL10, RPL10, L15, RPL15 3 23916804 23923696 6892 + 1 348 in gene 37 DBA12, EC45, RPL15, FLJ26304 rpL23, FLJ92089, RPL17, MGC117162, RPL23, RPL17 18 49488481 49492565 4084 - 1 5333 downstream 49 FLJ18762, L17 RPL17-C18orf32 18 49481178 49492565 11387 - 1 5333 in gene 49 RPL17-C18orf32 RPL18A 19 17859878 17863324 3446 + 1 122 in gene 31 RPL18A, L18A RPL19, FLJ27452, L19, MGC71997, RPL19 17 39200283 39204727 4444 + 1 133 in gene 22 DKFZp779D216 FLJ27458, L21, DKFZp686C06101, MGC71252, RPL21 13 27251555 27256568 5013 + 1 -67 upstream 35 MGC104274, RPL21, HYPT12, MGC104275 RPL22L1 3 170864875 170870256 5381 - 1 -48 upstream 35 MGC104449, RPL22L1 MGC72008, RPL23, MGC111167, MGC117346, RPL23 17 38850068 38853800 3732 - 1 232 in gene 17 L23, rpL17 in gene, downstream, RPL23A 17 28719982 28724356 4374 + 3 178, 6855, 8930 23 FLJ27455, MDA20, RPL23A, L23A downstream RPL23AP97 13 114338798 114347209 8411 + 1 1522 in gene 24 RPL23AP97 RPL24 3 101681090 101686719 5629 - 2 9695, 95 downstream, in gene 41 RPL24, L24, HEL-S-310 RPL26 17 8377516 8383247 5731 - 1 143 in gene 16 L26, DBA11, RPL26 RPL27 17 42998429 43002959 4530 + 1 131 in gene 35 RPL27, L27 MGC87238, FLJ43464, RPL27A, MGC10850, RPL27A 11 8682448 8689872 7424 + 2 208, 6480 in gene, in gene 34 MGC17878, L27A RPL28 19 55385101 55403245 18144 + 1 23027 downstream 25 FLJ43307, RPL28, L28 HUMRPL29, HIP, RPL29_3_370, RPL29, RPL29 3 51993628 51995942 2314 - 1 38 in gene 42 MGC88589, L29, RPL29P10 RPL3 22 39312882 39319665 6783 - 1 113 in gene 40 RPL3, TARBP-B, L3, ASC-1, MGC104284 RPL30 8 98041710 98045590 3880 - 1 102 in gene 20 RPL30, L30 RPL31 2 101002229 101019693 17464 + 1 75 in gene 25 RPL31, MGC88191, L31 RPL32 3 12834945 12841582 6637 - 1 -82 upstream 181 PP9932, RPL32, L32 8585, 73, -6999, - downstream, in gene, RPL35 9 124857879 124861961 4082 - 4 52 RPL35, L35 7831 upstream, upstream RPL35A 3 197950181 197955851 5670 + 2 -144, 10059 upstream, downstream 17 L35A, DBA5, RPL35A RPL36 19 5690261 5691667 1406 + 3 -9765, -3365, 43 upstream, upstream, in gene 47 DKFZp566B023, RPL36, L36 RPL36AL 14 49618688 49620685 1997 - 1 -115 upstream 52 MGC111574, RPL36A, RPL36AL RPL37 5 40831328 40835285 3957 - 1 21 in gene 35 DKFZp686G1699, RPL37, MGC99572, L37 RPL37A 2 216498797 216501467 2670 + 1 179 in gene 19 L37A, RPL37A, MGC74786, FLJ13357 RPL3L 16 1944579 1954678 10099 - 1 -10026 upstream 21 RPL3L RPL4 15 66499315 66504855 5540 - 1 7015 downstream 27 L4, RPL4 RPL41 12 56116590 56117832 1242 + 1 1682 downstream 30 RPL41, L41 L5, MGC117339, DBA6, MSTP030, RPL5, RPL5 1 92832037 92841924 9887 + 1 123 in gene 33 PPP1R135 RPL6 12 112405190 112418838 13648 - 2 9366, -74 in gene, upstream 39 SHUJUN-2, TXREB1, TAXREB107, RPL6, L6 RPL7 8 73290639 73294494 3855 - 2 382, -162 in gene, upstream 31 L7, humL7-1, MGC117326, RPL7 RPL7A 9 133348214 133351425 3211 + 2 202, 8362 in gene, downstream 26 SURF3, TRUP, RPL7A, L7A RPL7L1 6 42879917 42889889 9972 + 1 -237 upstream 25 RPL7L1, MGC62004, dJ475N16,4 RPL8 8 144789767 144792420 2653 - 2 5348, 84 downstream, in gene 25 L8, RPL8 RPL9P9, RPL9P7, RPL9P8, RPL9, MGC15545, RPL9 4 39454124 39458948 4824 - 1 228 in gene 33 FLJ27456, DKFZp313J1510, NPC-A-16, L9 MGC111226, L10E, PRLP0, RPLP0, LP0, P0, RPLP0 12 120196700 120201211 4511 - 2 6283, 139 downstream, in gene 32 RPP0, MGC88175 RPLP1 15 69452820 69455545 2725 + 2 236, 812 in gene, in gene 25 RPLP1, RPP1, P1, FLJ27448, MGC5215, LP1 -4432, -976, -464, upstream, upstream, RPLP2 11 809936 812876 2940 + 4 66 MGC71408, RPLP2, RPP2, D11S2243E, LP2, P2 10096 upstream, downstream RPS10 6 34417454 34426125 8671 - 1 45 in gene 18 DBA9, RPS10, MGC88819, S10 RPS10-NUDT3 6 34287194 34426125 138931 - 3 34178, 33373, 45 in gene, in gene, in gene 41 RPS10-NUDT3 upstream, upstream, RPS11 19 49496365 49499712 3347 + 3 -8925, -349, 4451 26 RPS11, S11 downstream RPS14 5 150444229 150449756 5527 - 1 -132 upstream 51 S14, RPS14, EMTB RPS15 19 1438364 1440494 2130 + 1 68 in gene 47 RPS15, S15, MGC111130, RIG RPS15A 16 18782955 18790334 7379 - 1 254 in gene 25 S15a, FLJ27457, MGC111208, RPS15A RPS16 19 39433207 39436020 2813 - 1 -12 upstream 74 RPS16, S16 DBA4, RPS17L2, MGC72007, S17, RPS17L1, RPS17 15 82536750 82540544 3794 - 1 64 in gene 42 RPS17, RPS17L KE3, D6S218E, MGC126837, KE-3, HKE3, RPS18 6 33272075 33276504 4429 + 2 -427, 5333 upstream, downstream 55 MGC126835, MGC117351, RPS18, S18 RPS19 19 41859918 41871416 11498 + 2 -30, 11890 upstream, downstream 42 DBA1, S19, DBA, RPS19 MGC52010, dJ1104E15,4, AROS, FLJ21770, RPS19BP1 22 39529093 39532855 3762 - 2 12279, -9 downstream, upstream 30 RPS19BP1, S19BP MGC117344, MGC117345, MGC102851, S2, RPS2 16 1962061 1964826 2765 - 3 122, -326, -7110 in gene, upstream, upstream 90 RPS2, LLREP3 RPS20 8 56068180 56074581 6401 - 1 53 in gene 80 MGC102930, RPS20, FLJ27451, S20 RPS21 20 62387065 62388520 1455 + 1 -89 upstream 46 S21, RPS21, HLDF RPS23 5 82273320 82278416 5096 - 1 48 in gene 28 S23, FLJ35016, RPS23 RPS24 10 78033760 78056813 23053 + 1 -4256 upstream 31 DBA3, RPS24, DKFZp686N1586, S24 RPS25 11 119015712 119018347 2635 - 1 -149 upstream 30 S25, RPS25 RPS26 12 56041902 56044223 2321 + 1 114 in gene 31 DBA10, MGC104292, S26, RPS26 RPS27 1 153990763 153992155 1392 + 1 69 in gene 22 MPS-1, RPS27, MPS1, S27 RPS27A, CEP80, UBC, UBCEP1, UBA80, RPS27A 2 55231903 55235853 3950 + 1 -8255 upstream 22 UBCEP80, S27A, HEL112 RPS28 19 8321500 8322396 896 + 1 -220 upstream 22 DBA15, RPS28, S28 RPS29 14 49576672 49586416 9744 - 2 336, -144 in gene, upstream 18 S29, RPS29, DBA13 RPS3 11 75399491 75422302 22811 + 1 157 in gene 37 RPS3, FLJ26283, MGC87870, S3, FLJ27450 RPS3A 4 151099573 151104652 5079 + 1 -149 upstream 40 FTE1, MGC23240, S3A, RPS3A, MFTL RPS5 19 58387269 58394804 7535 + 2 -6485, -197 upstream, upstream 50 S5, RPS5 RPS6 9 19376256 19380237 3981 - 1 184 in gene 37 RPS6, S6 12290, 12610, MAPKAPK1A, RSK, RSK1, HU-1, RPS6KA1, RPS6KA1 1 26529758 26575030 45272 + 3 in gene, in gene, in gene 37 16034 p90Rsk S6K-alpha, HU-2, RSK, MAPKAPK1C, S6K- RPS6KA2 6 166409366 166862550 453184 - 2 234998, 234358 in gene, in gene 24 alpha2, RPS6KA2, RSK3, p90-RSK3, pp90RSK3

HU-3, MAPKAPK1B, RSK, RPS6KA3, CLS, RPS6KA3 X 20149911 20267095 117184 - 1 87 in gene 20 pp90RSK2, S6K-alpha3, p90-RSK2, ISPK-1, RSK2, MRX19 RPS6KA5 14 90868122 91060649 192527 - 1 73 in gene 49 RLPK, MSPK1, RPS6KA5, MGC1911, MSK1 P70-beta-1, S6K-beta2, p70S6Kb, STK14B, KLS, 52, 6836, 7764, in gene, in gene, downstream, RPS6KB2 11 67428460 67435408 6948 + 4 28 P70-beta, SRK, S6K2, RPS6KB2, P70-beta-2, 8948 downstream p70(S6K)-beta RSKL1, RPS6KC1, humS6PKh1, S6PKh1, RPS6KC1 1 213051233 213273465 222232 + 1 -65 upstream 39 RPK118 RPS7 2 3575263 3580919 5656 + 1 -95 upstream 38 RPS7, DBA8, S7 RPS8 1 44775574 44778740 3166 + 1 106 in gene 30 RPS8, S8 RPS9 19 54200557 54207647 7090 + 2 -9517, -9181 upstream, upstream 27 RPS9, S9 Supplemental Table 2: Differentially expressed genes in ETS1 p27 versus control NK cells in positive and negative regulation of mitotic cell cycle (GO) pathways.

Differential gene GO_Positive regulation of GO_Negative regulation of expression in ETS1 p27 mitotic cell cycle mitotic cell cycle versus control NK cells

UP APP, UBE2C, DLGAP5, CDC25A, TTL, TOP2A, PSMA7, PSMB9, CDKN2A, CCND3, FGFR1, CCNB1, CDC25B GTSE1, AURKB, RPS27L, SPDL1, , TRIP13, ZWILCH, TFDP1,

PLK1, ZWINT, PSMD2, CLSPN, PSMC5, CENPF, PSMD3, PSMB5, PSME2, TRIAP1, CCNB1, HMGA2, E2F7, PSMD11, CDKN1A, TIPIN

DOWN RGCC; IGF1 BTG1, ZFP36L1, CDKN1B, ZFP36L2, RPL26, RPL24, BTG2, CTDSPL, RGCC, , PLAGL1, PSMA5

Supplemental Table 3: Expression of ETS family members in human NK cells. The expression level (Base Mean = mean normCounts) and differential expression of ETS1 p27 vs control NK cells is indicated. * n.s.: not significant; ** n.d.: not detectable.

ETS1 Family Base Subfamily p27 vs log2FoldChange padj members Mean ctrl ELF1 450 n.s.* ELF ELF2 208 n.s. ELF4 428 n.s.

ELG GABPA 133 n.s. ERG 1 n.s. ERG FLI1 456 down -0,95 0,06 FEV n.d.** ERF 74 n.s. ERF ETV3 51 n.s. ELF3 2 n.s. ESE ELF5 n.d. ESE3 n.d. ETS1 2504 up 0,71 0,01 ETS ETS2 82 n.s.

PDEF SPDEF n.d. ETV1 7 n.s. PEA3 ETV4 2 n.s. ETV5 24 n.s. ER71 ETV2 13 n.s. SPI1 7 n.s. SPI SPIB 7 up 5,70 0,03 SPIC n.d. ELK1 251 n.s. TCF ELK4 246 n.s. ELK3 236 n.s. ETV6 231 n.s. TEL ETV7 n.d.

Cas9 (ctrl) gRNA + Cas9 (KO) Clone nr. 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 ETS1

Clone nr. 1 2 3 4 5 7 6 8 9 1 2 3 4 5 6 7 8 9 Vinculin

Supplemental Figure 1. Validation of gRNA ETS1 in Jurkat cells using CRISPR/Cas9. Jurkat cells were nucleofected with either Cas9_eGFP vector, or in combination with a vector containing a human ETS1 specific gRNA. Two days after nucleofection, eGFP+ Jurkat cells were single cell sorted and cultured. When a sufficient number of cells was obtained, nine random control and nine random knockout Jurkat clones were harvested, lysed and analyzed by western blot for ETS1 and vinculin expression. ctrl, control; KO, knockout. A CB HPC B CB HPC C CB HPC D NK cells day 0 day 0 day 0 day 21 * 8 * 2.0 100 ctrl * ctrl 6 * p51 p27 80 1.5 4 60 ETS1 0.50 1.0 40 ActinB Counts 0.5 0.25 20

0.00 0 0.0 3 3 4 5 expression Relative ETS1 expression Relative ETS1 ctrl p27 p51 -10 0 10 10 10 ctrl p27 p51 ETS-1

Supplemental Figure 2. ETS1 overexpression analysis. (A-C) Cord blood HPCs were transduced with the indicated ETS1 isoforms at d-2 and analysed at day 0. (A) CD34+eGFP+ cells were sorted and expression of ETS1 was determined by qPCR. Data show the relative expression to GAPDH and YWHAZ (mean + SEM, n=3). (B) ETS1 expression was determined by flow cytometry on gated eGFP+CD34+Lin- cells. Data are representative for 3 experiments. (C) CD34+eGFP+ cells were sorted and expression of ETS1 was determined by Western blot analysis. (D) ETS1 qPCR was performed on sorted NK cells at day 21 of the cultures. Data show the relative expression to GAPDH and YWHAZ (mean + SEM, n=3). A count CD56 CD45

eGFP CD11a CD94

B stage 1 stage 2 CD34 CD34 eGFP

FSC CD45RA CD117 CD34 eGFP NKp44

stage 3

CD45 CD94 CD117 Supplemental Figure 3. Gating strategy (A) Mature NK cells were gated as CD45+eGFP+CD11a+CD56+CD94+ cells. Representative plots of day 21 analysis is shown. (B) Stage 1 cells were gated as eGPF+CD45RA+CD34+ CD117-, stage 2 cells as eGFP+CD45RA+CD34+CD117+ and stage 3 cells as eGFP+CD45+ CD94-NKp44-CD117+CD34+. Representative plots of day 14 analysis are shown. GATA3 ● ● TNFSF4 ● ● FUT7 CBFB INHBA ● CD3E ATP7A ● ● DNAJB9 ● NTRK1 MMP14 ● ● ● CD80 EGR3 IL15 ● PRDM1 ● ● ● SOS1 EOMES ●

● ● SASH3 LFNG ● ● ID2 ● Lymphocyte differentiation IKZF1 TBX21 JAG2 BCL11B ● ● ZBTB1 ● ● IKZF3 ZNF683● PIK3R6 ● TNFRSF18● ● SH3RF1 TESPA1 ● ● ● TGFBR2 IRF2BP2 ● CAMK4 ADGRG3 ● ● ● CD27 IL2RA

● ● CD2 ● ● PPP2R3C CD8A CD3D ● ● ● RBPJ KIT RORA ● ● PREX1 ITK ZFP36L1 ● ● ZFP36L2

Supplemental Figure 4. Network plot of lymphocyte differentiation Network plot of the GO Biological Process pathway 'lymphocyte differentiation' showing the up- (red) and down-regulated (bleu) genes determined in the RNA-seq results of ETS1 p27 versus control NK cells. A Peptide chain elongation SRP−dependent cotrans...targeting to membrane Immunoregulatory inte...d a non−Lymphoid cell Neurodegenerative Diseases Mitotic G1−G1/S phases Neutrophil degranulation Transcriptional regulation by RUNX3 M Phase Cellular responses to stress RHO GTPases Activate Formins Signaling by NOTCH Transcriptional Regulation by TP53 Programmed Cell Death Signaling by Rho GTPases Transcription of E2F ...trol by DREAM complex p53−Dependent G1 DNA Damage Response Lysosphingolipid and LPA receptors Glycolysis Degradation of beta−c...e destruction complex TNFR2 non−canonical NF−kB pathway L1CAM interactions Cross−presentation of... antigens (endosomes) Activation of Matrix Metalloproteinases Signaling by NTRKs MAPK6/MAPK4 signaling Repression of WNT target genes Interleukin−1 signaling Transcriptional regulation by RUNX1 Activation of E2F1 target genes at G1/S Platelet degranulation

0 5 10 15

B organic cyclic compound catabolic process ● aromatic compound catabolic process ● lymphocyte activation ● negative regulation ...velopmental process ● regulation of MAPK cascade ● cell−cell adhesion ● cardiovascular system development ● G protein−coupled re...r signaling pathway ● neutrophil activatio... in immune response ● Count taxis ● inflammatory response ● ● 20 developmental proces...ved in reproduction ● ● 40 lymphocyte differentiation ● 60 leukocyte migration ● ● response to hypoxia ● ● 80 translational initiation ● regulation of endopeptidase activity ● establishment of pro...doplasmic reticulum ● p.adjust nuclear−transcribed ...ense−mediated decay ● SRP−dependent cotran...rgeting to membrane ● 0.0025 extracellular structure organization ● epidermis development ● 0.0050 muscle cell proliferation ● 0.0075 axon guidance ● projection guidance ● humoral immune response ● cell killing ● negative regulation ...flammatory response ● killing of cells of other organism ● disruption of cells of other organism ● regulation of cardia... cell proliferation ● antimicrobial humora...timicrobial peptide ● cardiac epithelial t...enchymal transition ● 0.025 0.050 0.075 GeneRatio

Supplemental Figure 5. Additional pathway analysis of ETS1 p27 NK RNA-seq data (A) Reactome Pathway and (B) GO Biological Process enrichment analysis of the RNA-seq results of ETS1 p27 versus control NK cells. A B 40 ctrl 100 p27 30 80 60 20 40 ctrl % expression 10 % specific lysis 20 p27

0 0 CD107a 0.03:1 0.1:1 0.3:1 1:1 effector:target C D 80 1.5 ctrl p27 60

NK 1.0 + 40 0.5

% IFN- γ 20

0 Relative IFN- γ secretion 0.0 IL-12 IL-12 IL-12 IL-12 IL-18IL-18 IL-18IL-18 IL-15 IL-15

Supplemental Figure 6. ETS1 is not essential to maintain the functionality of mature NK cells Mature NK cells from dl8 control cultures were transduced with control or ETS1 retrovirus, and functionally analyzed 4 d later. (A) NK cells were stimulated with K562 target cells and CD107a expression was determined (mean ± SEM, n=5). (B) The cytotoxic capacity against K562 targets was determined in a 51Cr release assay. Data are expressed as the percentage specific lysis as a function of the effector:target ratio (mean ± SEM, n=5). (C) NK cells were stimulated with IL-12/IL-18 or IL-12/IL-15/IL-18 for 24 h and IFN-y production was determined by flow cytometry (mean ± SEM, n=5). (D) NK cells were stimulated with IL- 12/IL-18 or IL-12/IL-15/IL-18 for 24 h and secreted IFN-y was measured by ELISA. The relative IFN-y production compared to control is shown (mean ±SEM, n=5).