DOCK8 Drives Src-Dependent NK Cell Effector Function Conor J

DOCK8 Drives Src-Dependent NK Cell Effector Function Conor J. Kearney, Stephin J. Vervoort, Kelly M. Ramsbottom, Andrew J. Freeman, Jessica Michie, Jane This information is current as Peake, Jean-Laurent Casanova, Capucine Picard, Stuart G. of September 29, 2021. Tangye, Cindy S. Ma, Ricky W. Johnstone, Katrina L. Randall and Jane Oliaro J Immunol published online 9 August 2017

http://www.jimmunol.org/content/early/2017/08/09/jimmun Downloaded from ol.1700751

Supplementary http://www.jimmunol.org/content/suppl/2017/08/09/jimmunol.170075 Material 1.DCSupplemental http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

by guest on September 29, 2021 • Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2017 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published August 9, 2017, doi:10.4049/jimmunol.1700751 The Journal of Immunology

DOCK8 Drives Src-Dependent NK Cell Effector Function

Conor J. Kearney,*,† Stephin J. Vervoort,‡ Kelly M. Ramsbottom,* Andrew J. Freeman,* Jessica Michie,* Jane Peake,x Jean-Laurent Casanova,{,‖,#,** Capucine Picard,{,#,†† Stuart G. Tangye,‡‡,xx Cindy S. Ma,‡‡,xx Ricky W. Johnstone,†,‡ Katrina L. Randall,{{,‖‖ and Jane Oliaro*,†

Mutations in the dedicator of cytokinesis 8 (DOCK8) gene cause an autosomal recessive form of hyper-IgE syndrome, characterized by chronic immunodeﬁciency with persistent microbial infection and increased incidence of malignancy. These manifestations suggest a defect in cytotoxic lymphocyte function and immune surveillance. However, how DOCK8 regulates NK cell–driven immune responses remains unclear. In this article, we demonstrate that DOCK8 regulates NK cell cytotoxicity and cytokine production in response to target cell engagement or receptor ligation. Genetic ablation of DOCK8 in human NK cells attenuated

cytokine transcription and secretion through inhibition of Src family kinase activation, particularly Lck, downstream of target Downloaded from cell engagement or NKp30 ligation. PMA/Ionomycin treatment of DOCK8-deﬁcient NK cells rescued cytokine production, indicating a defect proximal to receptor ligation. Importantly, NK cells from DOCK8-deﬁcient patients had attenuated production of IFN-g and TNF-a upon NKp30 stimulation. Taken together, we reveal a novel molecular mechanism by which DOCK8 regulates NK cell–driven immunity. The Journal of Immunology, 2017, 199: 000–000.

utosomal recessive hyper-IgE syndrome (AR-HIES) is a bacterial, and fungal infections; and recurrent pneumonia. Other http://www.jimmunol.org/ rare primary immunodeﬁciency disease characterized by clinical symptoms include eczema and an increased incidence of A elevated serum IgE levels; repeated cutaneous viral, severe allergies and asthma (1, 2). A major advance in the man- agement of patients with this rare disease was made in 2009 with *Immune Defence Laboratory, Cancer Immunology Division, Peter MacCallum Can- the discovery that mutations in the dedicator of cytokinesis 8 cer Centre, Melbourne, Victoria 3000, Australia; †Sir Peter MacCallum Department (DOCK8) gene, located on chromosome 9p, are responsible for of Oncology, The University of Melbourne, Parkville, Victoria 3052, Australia; most cases of AR-HIES (2, 3). ‡Gene Regulation Laboratory, Cancer Therapeutics Division, Peter MacCallum Can- cer Centre, Melbourne, Victoria 3000, Australia; xUniversity of Queensland and Lady DOCK8 is a member of the DOCK180-related family of guanine Cilento Children’s Hospital, Brisbane, Queensland 4006, Australia; {Laboratory of nucleotide exchange factors (GEFs), which promote the activity of

Human Genetics of Infectious Diseases, Necker Branch, INSERM U1163, Imagine Rho GTPases such as Rac and Cdc42 (4), and are involved in by guest on September 29, 2021 Institute, Necker Medical School, University Paris Descartes, 75015 Paris, France; ‖Pediatric Hematology and Immunology Unit, Necker Hospital for Sick Children, diverse cellular processes including cell migration, differentiation, AP-HP, Paris, France; #St. Giles Laboratory of Human Genetics and Infectious Dis- and cell–cell interactions (4). Although DOCK8 is ubiquitously eases, Rockefeller Branch, The Rockefeller University, New York, NY 10065; **Howard Hughes Medical Institute, New York, NY 10065; ††Study Center for expressed, it is particularly enriched in lymphocytes, suggesting it Primary Immunodeﬁciencies, Necker Hospital for Sick Children, AP-HP, 75015 functions to promote immunity (5). Indeed, recent studies have Paris, France; ‡‡Immunology Division, Garvan Institute of Medical Research, xx found that DOCK8 is required for the survival and persistence of Darlinghurst, New South Wales 2010, Australia; St. Vincent’s Clinical School, Uni- + versity of New South Wales, New South Wales 2052, Australia; {{Department of CD8 T cells (6) and NKT cells (7), and for long-term Ab pro- Immunology and Infectious Disease, The John Curtin School of Medical Research, duction by B cells (5). DOCK8 also appears to be essential for Australian National University, Acton, Australian Capital Territory 2601, Australia; and ‖‖Australian National University Medical School, Australian National University, maintaining the structural integrity of T cells, serving to repress Acton, Australian Capital Territory 2601, Australia cell death during migration into skin (8), which is essential for the ORCIDs: 0000-0001-9003-5861 (J.P.); 0000-0001-8788-5056 (C.P.); 0000-0002- control of herpes virus skin infection (8, 9). 5360-5180 (S.G.T.); 0000-0001-9708-6020 (K.L.R.). Consistent with the clinical characteristics of AR-HIES, DOCK8 Received for publication May 24, 2017. Accepted for publication July 12, 2017. is also implicated in the regulation of the NK cell and CD8+ T cell This work was supported by the National Health and Medical Research Council of immune synapse. The immune synapse is a specialized structure Australia (Grant 1079318). that forms between the plasma membrane of two cells, facilitating J.O., K.L.R., and C.J.K. designed the study; C.J.K., S.J.V., K.M.R., A.J.F., and J.M. signaling or the triggering of killer cell effector function (10, 11). performed experiments, analysis, and interpretation of the data; C.S.M., S.G.T., J.P., C.P., J.-L.C., and R.W.J. provided clinical samples and contributed to writing the Formation of the immune synapse results in rapid signal trans- manuscript; C.J.K. and J.O. wrote the manuscript. duction events that promote optimal cell activation, but also The sequencing data presented in this article have been submitted to the Gene Expression triggers reorganization of the cell cytoskeleton. These changes Omnibus (http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE101467. facilitate the spatial rearrangements of molecules that contribute Address correspondence and reprint requests to Dr. Jane Oliaro, Immune Defence Lab- to the success and quality of the immune synapse (11–13). Upon oratory, Cancer Immunology Division, Peter MacCallum Cancer Centre, 305 Grattan immune synapse formation in T cells, protein kinase C u and Street, Melbourne, VIC 3000, Australia. E-mail address: [email protected] leukocyte-speciﬁc protein tyrosine kinase (Lck) rapidly polarize The online version of this article contains supplemental material. into a region proximal to the synapse contact site called the central Abbreviations used in this article: AR-HIES, autosomal recessive hyper-IgE syndrome; DOCK8, dedicator of cytokinesis 8; GEF, guanine nucleotide exchange fac- SMAC region (11). These events trigger the TCR-mediated sig- tor; GSEA, gene set enrichment analysis; MTOC, microtubule-organizing center; naling cascades that culminate in T cell activation and acquisition NCR, natural cytotoxicity receptor; RNA-Seq, RNA sequencing; siRNA, small in- of effector functions. Furthermore, adhesion, costimulatory, and terfering RNA; WASp, Wiskott–Aldrich syndrome protein. conjugation-promoting molecules, such as LFA-1 and talin, mi- Copyright Ó 2017 by The American Association of Immunologists, Inc. 0022-1767/17/$35.00 grate to a ring that surrounds the central SMAC region (called the

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1700751 2 DOCK8 PROMOTES CYTOKINE PRODUCTION FROM NK CELLS peripheral SMAC), which serves to improve the quality of the containing 5% BSA. After washing, proteins were detected using West immune synapse (10, 11). Coast SuperSignal. DOCK8 was initially identified as a regulator of the B cell im- Conjugation assay munological synapse, where it was found to be required for long-lived Ab production (5). Later, DOCK8 was found to also regulate the NK K562 and KHYG1 cells were labeled with CellTrace Violet and CFSE (Molecular Probes), respectively, for 20 min at 37˚C. Cells were washed twice, cell cytotoxic synapse (14, 15). In this study, it was demonstrated then added to Eppendorf tubes (K562 1 3 105, KHYG1 2 3 105)ina200ml that DOCK8 was essential to promote the polarization of cytotoxic final volume. At the indicated time points, cells were vortexed for 2 s, fixed granules to the synaptic cleft via an interaction with Talin and with 4% PFA, and then analyzed on a flow cytometer (BD Fortessa). Wiskott–Aldrich syndrome protein (WASp) (14). Similarly, NK Gene knockout/knockdown cells derived from patients with DOCK8 deficiency were characterized by poor actin accumulation at the synapse contact site For CRISPR, a guide RNA targeting DOCK8 (59-CAAAGTCCACTGG- (15). However, it remains unclear whether DOCK8 regulates other CTCCACA-39) was cloned into lentiviral FgH1t-UTG vector with inducible expression of guide RNA and a fluorescent GFP reporter. KHYG1 cells important NK cell functions, such as the signal transduction events stably expressing Cas9 were transduced and then sorted for GFP positivity. required to promote inflammatory cytokine production. Doxycycline was added for 5 d to induce guide expression. Single cells were Mechanistically, DOCK8 links the TCR to the actin cytoskeleton then seeded in 96-well plates and expanded. Clones were screened for loss of via a bridging interaction between WASp and WASp-interacting DOCK8 expression by Western blot, and a single clone was selected for all experiments using the knockout cell line. For RNA interference, KHYG1 or protein (16). DOCK8 also functions as an adaptor protein in primary human NK cells (cultured in IL-2 for 48 h) were electroporated with signaling pathways downstream of TLR9 receptor engagement on 100 nM small interfering RNA (siRNA; Amaxa program X001) to knock-

B cells essential for B cell proliferation and differentiation via down DOCK8 or Lck. After 48–72 h, cells were treated as indicated. The Downloaded from + siRNA sequences were as follows: scramble control: 59-AUGUUA- STAT3 (17). Similarly, DOCK8 regulates CD4 T cell TH17 polarization through a constitutive association with STAT3 (18, 19). GUAGCGAUUGUAU-39; DOCK8: 59-GGAGAUUAUUUGUGAACUU- 39; and Lck: 59-UAACCAGGUUGUCUUGCAGUG-39. Thus, evidence is emerging that suggests that DOCK8 not only promotes proper synapse formation, but also adopts important RNA sequencing and analysis roles in the signal transduction events that are required for optimal KHYG1 cells were cultured as indicated before stimulation with anti-Nkp30 immunological signaling cascades in diverse cell types. (duplicate samples). After 90 min, cell pellets were collected and total RNA http://www.jimmunol.org/ NK cell function is governed through the relative strength of was extracted using the NucleoSpin RNA extraction kit (Macherey-Nagel). activatory and inhibitory signals, delivered to the NK cell by li- RNA quality was checked on the Agilent 4200 TapeStation, and RNA with RIN values .9 were used for the subsequent analysis. Sequencing libraries gands expressed on target cells (12, 20). ITAM-bearing receptors, were prepared using the QuantSEquation 39 mRNA-Seq Library Prep Kit such as natural cytotoxicity receptors (NCRs) and NKG2D, pro- (Lexogen) according to the manufacturer’s instructions using 200 ng of mote NK cell function, including cytotoxic granule delivery and total RNA input. Single-end 75-bp RNA sequencing (RNA-Seq) was cytokine production, through activation of signaling cascades performed on the NextSEquation 500 (Illumina). Demultiplexing of the , initiated by Src kinase family members such as Lck and Fyn (20– reads was performed using CASAVAv1.8.2, and low-quality reads Q 30 were removed. Cutadapt (v1.9) was used to trim polyA-derived sequences 22). Src kinases phosphorylate the cytoplasmic tails of receptor and biased 39 reads resulting from random hexamer priming. HISAT2 was ITAMS to recruit the “second-line” proteins, such as ZAP70 and used to map the resulting reads to the human reference genome. Read by guest on September 29, 2021 SLP-76, which propagate signals leading to activation of MAPKs counting was performed using featureCounts, which is part of the subread and transcription factors that promote granule polarization and package (25). Voom-LIMMA workflow was used to normalize data for differential gene expression (26). Gene set enrichment analysis (GSEA) cytokine transcription (22, 23). Given that the clinical character- was performed using GSEA2-2.2.2 for identification of enriched signatures istics of DOCK8 deficiency suggest poor NK cell–mediated im- obtained from the MSigDB Hallmarks datasets (27). munity, we investigated whether DOCK8 regulates NK cell function by promoting the signal transduction events that drive Accession code both cytotoxicity and inflammatory cytokine production. Sequencing data have been deposited into the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE101467. Materials and Methods Statistical analyses Cells and assays Statistical significance was determined using an unpaired Student t test. NK cells were cultured in RPMI 1640 supplemented with 10% FCS plus Differences were considered significant when p , 0.05. IL-2 (450 U/ml for KHYG1 cells, 100 IU/ml for primary human NK cells). Cytotoxicity was measured using a chromium release assay (24). Cytokine secretion was measured using human Th1/Th2 Cytometric Bead Array Results (BD Biosciences) and analyzed on a FACSVerse. DOCK8 is required for NK cell cytotoxicity, but not Antibodies target conjugation Western blot analysis was performed using the following Abs: anti-human To investigate the proposed (14, 15) role of DOCK8 in human NK DOCK8 (Santa Cruz), phospho-Src, Src, LCK, phospho-LAT, LAT, cell effector function, we first genetically deleted DOCK8 from the phospho-ERK, ERK, phospho-P38, P38, phospho-P65, and P65 (Cell human NK cell line, KHYG1, using CRISPR/Cas9 technology. Loss Tyr505 Signaling). Phospho-LCK (Gentex) and NKp30-PE were used for of DOCK8 significantly attenuated cytotoxicity against both K562 FACS analysis (BioLegend). Agonistic NKp30 Ab was from R&D Sys- and HeLa target cells (Fig. 1A). To confirm the specificity of this tems. Agonistic NKp44 Ab was from Thermo Fisher. approach, we also demonstrated that siRNA knockdown of DOCK8 Western blotting reduced NK cell cytotoxicity against K562 and HeLa target cells (Fig. 1B). Thus, DOCK8 appears to be critical for the cytotoxic To analyze proteins, we used NaDodSO4 PAGE (SDS-PAGE). Samples were prepared in sample 2% SDS, 50 mM Tris-HCl (pH 6.8), 10% glyc- function of human NK cells. To test whether loss of DOCK8 was erol, 2.5% 2-ME, then boiled for 7 min. Lysates were then loaded into 8– impairing cytotoxicity through decreased conjugation to target cells, 12% polyacrylamide gels and electrophoresed at 75 V. Resolved proteins we performed a FACS-based conjugation assay using K562 cells as were then transferred onto nitrocellulose membranes at 40 mA overnight. targets. DOCK8 knockout NK cells conjugated to targets as effi- Membranes were blocked for 1 h (5% BSA, 0.05% NaN3 in PBS, Tween 20), then incubated with primary Ab overnight. Proteins were detected ciently as control NK cells (Fig. 1C); however, confocal analyses of using HRP-conjugated secondary Ab in 0.05% NaN3 in PBS and Tween 20 these conjugates confirmed previous reports that DOCK8-deficient The Journal of Immunology 3 Downloaded from http://www.jimmunol.org/

FIGURE 1. DOCK8 is required for NK cell cytotoxicity, but not target conjugation. (A) Control or DOCK8 CRISPR knockout KHYG1 cells were used at the indicated E:T ratio in a chromium release assay with K562 or HeLa cells as targets. Killing kinetics are represented as a Michaelis–Menten plot and relative killing on the right. (B) Control or DOCK8 knockdown KHYG1 cells were used at the indicated E:T ratio in a chromium release assay with K562 or HeLa cells as targets. Killing kinetics are represented as a Michaelis–Menten plot and relative killing on the right. (C) Control or DOCK8 CRISPR knockout KHYG1 cells were subject to a conjugation assay with K562 cells. At the indicated time points, cell conjugates were analyzed by ﬂow cytometry, by guest on September 29, 2021 as described in Materials and Methods. Representative FACS plots are displayed. Data are representative of at least three independent experiments. Error bars on the Michaelis–Menten plots represent the mean 6 SEM of triplicate determinations from a representative experiment. Error bars on graphs (relative killing) are mean 6 SEM pooled from three independent experiments. *p , 0.05 by unpaired Student t test.

NK cells do not efficiently polarize the microtubule-organizing center DOCK8 control and knockout NK cells showed that the magni- (MTOC) and perforin-containing granules to the immune synapse (14, tude of the response elicited by NKp30 was impaired in DOCK8 15) (Supplemental Fig. 1A). These data suggest that DOCK8 regulates knockout cells. DOCK8 loss resulted in reduced expression across NK cell cytotoxicity downstream of NK cell–target conjugation. the majority of NKp30 target genes, including key inflammatory factors. In accordance with this observation, GSEA on the RNA- DOCK8 regulates inflammatory transcriptional events upon seq data revealed that genes involved in inflammation, K-RAS activation through NKp30 engagement signaling, and the IFN-g response were significantly reduced in Recognition of targets by NK cells results in receptor-mediated NKp30-stimulated NK cells upon DOCK8 deletion compared activation of signaling pathways that drive granule polarization with their control counterpart. Importantly, gene signatures asso- to the cytotoxic synapse and gene transcription leading to rapid ciated with distinct signaling pathways such as TGF-b signaling, cytokine synthesis and secretion (23). ITAM-bearing receptors, PI3K/AKT signaling, and protein secretion pathways were unaf- such as NCRs, deliver activating signals through signaling motifs fected in DOCK8 knockout NK cells (Fig. 2C). Interestingly, within their cytoplasmic tails. The NCR, NKp30, is known to upregulation of genes encoding the cytokines IFN-g,TNF-a, trigger NK cell activity against K562 cells upon binding to its IL-10, IL-8, and IL-2 was attenuated in response to NKp30 ligand B7-H6 (28). To determine the effects of DOCK8 deletion stimulation in DOCK8 knockout NK cells (Fig. 2D, 2E), whereas on NK cell activation, we performed genome-wide gene expres- expression of DOCK8 or key signaling mediators such as Lck and sion analysis using 39 mRNA-Seq on DOCK8 control and knockout NKp30 was unaffected by NKp30 stimulation (Fig. 2E). These NK cells, untreated or after NKp30 stimulation (Fig. 2). data demonstrate that loss of DOCK8 leads to a reduction in cy- As expected, DOCK8 expression was reduced in DOCK8 tokine gene transcription, suggesting that DOCK8 deficiency may CRISPR knockout NK cells, regardless of treatment (Fig. 2A). impair signaling pathways leading to transcriptional regulation Differential gene expression analysis of transcripts regulated by and secretion of key NK cell cytokines such as IFN-g and TNF-a. NKp30 in control NK cells demonstrated that stimulation caused an increase in multiple transcripts associated with immune cell DOCK8 is required for cytokine secretion after target cell activation. The immune activation signature was present in both conjugation or NKp30 engagement DOCK8 control and knockout NK cells (Fig. 2B, Supplemental To confirm that DOCK8 is involved in receptor-induced signal- Tables I, II). However, a direct comparison of NKp30-stimulated ing cascades leading to cytokine transcription and production, we 4 DOCK8 PROMOTES CYTOKINE PRODUCTION FROM NK CELLS Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 2. DOCK8 regulates inflammatory transcriptional events upon activation. (A) Control or DOCK8 CRISPR knockout KHYG1 cells (n = 2) were either left untreated or activated with anti-NKp30 (1 mg/ml) for 90 min followed by 39 mRNA-Seq. Normalized Log2CPM (count per million) DOCK8 expression is depicted. (B) Heat map of the RNA-Seq data from (A) depicting the top 100 regulated genes upon Nkp30 stimulation. (C) GSEA enrichment score plots from RNA-Seq data showing significant correlation between DOCK8 expression, or not, and genes regulating the indicated cellular processes. (D) Heat map of the RNA-Seq data from (A), showing a selection of inflammatory cytokines. (E) Fold gene expression from the RNA-Seq data in (A), for a selection of inflammatory cytokine genes and control genes. The Journal of Immunology 5 exposed DOCK8 control or knockout NK cells to K562 target cells attenuated in the absence of DOCK8, leading to reduced phos- and measured the release of cytokines. Consistent with the RNA- phorylation of ERK and p38 MAPK downstream (Fig. 4B, seq data, DOCK8-deficient NK cells secreted significantly less Supplemental Fig. 2A). DOCK8 knockdown NK cells also failed IFN-g, IL-2, TNF-a, and IL-10 compared with control NK cells to activate Src kinases as efficiently as control cells (Fig. 4C, (Fig. 3A), demonstrating that DOCK8 drives cytokine secretion Supplemental Fig. 2B). PMA/Ionomycin treatment activated ERK upon target cell engagement. Furthermore, when stimulated via and p38 MAPK equivalently in both DOCK8 control and the NCR, NKp30, DOCK8 knockout NK cells exhibited reduced knockout NK cells (Fig. 4D), and we could rescue the defect in cytokine release compared with control cells (Fig. 3B). To dem- cytokine production using nonspecific PMA/Ionomycin treatment onstrate the specificity of this approach, we repeated the experi- (Fig. 4E), indicating that loss of DOCK8 attenuates NKp30- ment using NK cells with siRNA-mediated knockdown of DOCK8 induced signaling pathways, but does not invoke a global inabil- and found a similar reduction in the secretion of IFN-g, IL-2, ity to trigger cytokine synthesis. There was no significant differ- TNF-a, and IL-10 (Fig. 3C). ence in cytokine production when DOCK8 control or knockout NK cells were stimulated through engagement of another NK cell DOCK8 drives cytokine production through activation of Src NCR, NKp44 (Fig. 4F), even though DOCK8 control and kinases, independent of the cytoskeleton knockout NK cells expressed equivalent levels of both NKp30 and Because DOCK8 acts as a signaling adaptor to promote Src kinase NKp44 (Supplemental Fig. 1B). These data suggest that DOCK8 activation upon TLR9 stimulation in B cells (17), we hypothesized may be a specific downstream effector of NKp30 engagement. that DOCK8 may drive NK cell function by linking ITAM-based DOCK8 has previously been shown to play an important role in signaling to Src kinase activation, which is known to be crucial for establishing the cytoskeletal networks required for efficient syn- Downloaded from NK cell activation (23). We exposed DOCK8 control or knockout apse formation and granule delivery leading to cytotoxic activity in NK cells to K562 targets and monitored activation of proximal NK cells (14, 15). To determine whether the attenuation in cyto- kinases. Unlike control cells, DOCK8-deficient NK cells largely kine production in the absence of DOCK8 was dependent on the failed to activate Src kinases, including Lck phosphorylation on cytoskeleton, we treated the cells with nocodazole, which disrupts the activation motif, Tyr505 (Fig. 4A). We next asked whether the tubule network and impairs translocation of the microtubule-

DOCK8 was required for ITAM-triggered signaling upon NKp30 organizing network (29). Although treatment of the NK cells with http://www.jimmunol.org/ ligation. In control NK cells, NKp30 stimulation potently trig- nocodazole signiﬁcantly impaired cytotoxic activity, we found no gered activation of Src family kinases, including Lck (Fig. 4B). effect on the secretion of either IFN-g or TNF-a (Fig. 4G, 4H). However, activation of Src family kinases, including Lck, was These data demonstrate that, under these conditions, NK cell by guest on September 29, 2021

FIGURE 3. DOCK8 is required for NK cell cytokine production. (A) Control or DOCK8 CRISPR knockout KHYG1 cells were exposed to K562 cells at the indicated E:T ratios. After 4 h, cytokines in the supernatants were analyzed by cytometric bead array (CBA). (B) Control or DOCK8 CRISPR knockout KHYG1 were stimulated with anti-NKp30 (1 mg/ml). At the indicated time points, cytokines in the supernatants were analyzed by CBA. (C) Control or DOCK8 knockdown KHYG1 cells were stimulated with anti-NKp30 (1 mg/ml). At the indicated time points, cytokines in the supernatants were analyzed by CBA. Knockdown efﬁciency is shown on the right. Data are representative of at least three independent experiments. Error bars represent the mean 6 SEM of triplicate determinations from a representative experiment. *p , 0.05 by unpaired Student t test. 6 DOCK8 PROMOTES CYTOKINE PRODUCTION FROM NK CELLS Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 4. DOCK8 is required for optimal Src kinase activation. (A) Control or DOCK8 knockout KHYG1 cells were exposed to K562 cells (10:1 E:T ratio). At the indicated time points, cell lysates were analyzed for the indicated proteins by Western blotting. (B) Control or DOCK8 knockout KHYG1 cells were stimulated with anti-NKp30 (10 mg/ml). At the indicated time points, cell lysates were analyzed for the indicated proteins by Western blotting. (C) KHYG1 cells were electroporated with scramble or DOCK8 siRNA. After 72 h, cells were stimulated with anti-NKp30 (10 mg/ml). Cell lysates were then probed for Src kinase activation. (D) Control or DOCK8 knockout KHYG1 cells were treated with PMA (10 ng/ml) and Ionomycin (1 mg/ml). At the indicated time points, cell lysates were analyzed by Western blotting. (E) Control or DOCK8 knockout KHYG1 cells were treated with PMA (10 ng/ml) and Ionomycin (1 mg/ml); then cytokines in the supernatants were analyzed by cytometric bead array (CBA) at the indicated time points. (F) Control or DOCK8 knockout KHYG1 cells were treated with anti-NKp30 (1 mg/ml) or anti-NKp44 (1 mg/ml). After 4 h, cytokines in the supernatants were analyzed by CBA. (G) KHYG1 cells were pretreated with DMSO or the indicated concentration of nocodazole for 30 min, then subjected to a killing assay using K562 cells as targets. Nocodazole concentration was maintained throughout the assay. (H) Supernatants from (G) were analyzed by CBA. Data are representative of at least three independent experiments. Error bars represent the mean 6 SEM of triplicate determinations from a representative experiment. Error bars on graph of relative killing (G) are mean 6 SEM pooled from three independent experiments. *p , 0.05 by unpaired Student t test. cytotoxicity and cytokine secretion are uncoupled, and suggest duction through the failure to optimally activate Src, we used the that the role of DOCK8 is more complex than simply acting as a speciﬁc Src kinase inhibitor, PP2. As expected, PP2 potently scaffolding protein to facilitate cytoskeletal changes required for inhibited NKp30-induced activation of Src kinases, including Lck cytotoxicity. (Fig. 5A), which translated into attenuated ERK activation downstream of Src. Importantly, inhibition of Src-mediated sig- The Src kinase, Lck, and ERK are required for NK cell naling entirely abolished the cytotoxicity of KHYG1 cells against cytotoxicity and cytokine production K562 targets (Fig. 5B). Src inhibition also inhibited NKp30- Src kinases, including Lck and Fyn, play a crucial role in the induced secretion of IFN-g, TNF-a, and IL-2 (Fig. 5B), thus activation of NK cells (23). To determine whether DOCK8 deﬁ- suggesting that DOCK8 is required for NK cell function via ac- ciency might attenuate NK cell cytotoxicity and cytokine pro- tivation of Src kinases and downstream MAPKs. The Journal of Immunology 7 Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 5. Src kinase Lck and ERK are required for NK cell cytokine production. (A) KHYG1 cells were treated with DMSO or PP2 (5 mM) for 15 min, then stimulated with anti-NKp30 (10 mg/ml). At the indicated time points, cell lysates were analyzed by Western blotting. (B) KHYG1 cells were treated with DMSO or PP2 (5 mM) for 15 min, then subjected to a chromium release assay using K562 cells as targets. KHYG1 cells were treated with DMSO or PP2 (5 mM) for 15 min, then stimulated with anti-NKp30 (1 mg/ml). At the indicated time points, cytokines in the supernatants were analyzed by cytometric bead array (CBA). (C and D) Control or Lck knockdown KHYG1 cells were used at the indicated E:T ratio in a chromium release assay with K562 or HeLa cells as targets. Killing kinetics are represented as a Michaelis–Menten plot and relative killing on the right. (E and F) Control or Lck knockdown KHYG1 cells were stimulated with anti-NKp30 (1 mg/ml). After 6 h, cytokines in the supernatants were analyzed by CBA. (G) KHYG1 (Figure legend continues) 8 DOCK8 PROMOTES CYTOKINE PRODUCTION FROM NK CELLS Downloaded from

FIGURE 6. NK cells from DOCK8-deﬁcient patients fail to respond to NKp30 stimulation. (A) NK cells were isolated from PBMCs from a healthy donor using negative selection, then incubated overnight in IL-2 (100 IU/ml). NK cells were then electroporated with the indicated siRNAs. Seventy-two hours later, cells were collected for Western blot analysis and treated with anti-NKp30 (1 mg/ml) for 4 h. Cytokines in the supernatants were analyzed by cytometric bead array (CBA). Data are representative of results obtained from four individual donors. (B) NK cells were isolated from PBMCs from control or DOCK8-deﬁcient patients (n = 2) using negative selection, then incubated overnight in IL-2 (100 IU/ml). The cells were then analyzed for NKp30 expression by FACS or stimulated with anti-NKp30 (1 mg/ml). After 4 h, cytokines in the supernatants were analyzed by CBA. Error bars represent the http://www.jimmunol.org/ mean 6 SEM of triplicate determinations from a representative experiment; *p , 0.05 by unpaired Student t test.

To determine whether the Src kinase, Lck, was the key signaling NKp30 ligation, whereas control NK cells produced abundant molecule affected by DOCK8 deficiency, as suggested by the amounts of these cytokines (Fig. 6B). Thus, DOCK8 regulates Western blots (Fig. 4A, 4B), we used siRNA to specifically knock cytokine production in human NK cells, suggesting that DOCK8 down Lck in KHYG1 cells. Knockdown of Lck reduced cyto- deficiency in humans likely results in poor NK cell–driven im- toxicity of NK cells against both K562 and HeLa target cells munity through suboptimal cytokine production.

(Fig. 5C, 5D, respectively). Knockdown of Lck also attenuated the by guest on September 29, 2021 secretion of both IFN-g and TNF-a upon NKp30 stimulation Discussion (Fig. 5E, 5F, respectively), demonstrating that Lck is essential for DOCK8 has recently emerged as a key regulator of lymphocyte NK cell cytotoxicity and cytokine production. Furthermore, inhi- function, both in the context of proper immune synapse forma- bition of both ERK and p38 significantly reduced IFN-g and TNF tion and through contributing to signal transduction events that secretion (Fig. 5G), and ERK inhibition reduced intracellular promote immune cell function (30). In this article, we have levels of TNF-a in both control and DOCK8 knockout NK cells demonstrated that DOCK8 regulates the signal transduction events (Fig. 5H), demonstrating the importance of these downstream that are required for NK cell effector function acquisition, by signaling molecules in the regulation of these cytokines. promoting the activation of the Src kinase Lck, downstream of target cell recognition or NKp30 stimulation. NK cells from DOCK8-deficient patients fail to respond to NK cells are innate immune effector cells, which are often the NKp30 stimulation first responders to microbial infection or malignant transformation. To investigate the role of DOCK8 in the function of primary NK cells, Upon recognition of infected or transformed cells, NK cells rapidly we isolated NK cells from the peripheral blood of healthy individ- acquire cytotoxic effector function and secrete inflammatory cy- uals and used siRNA to knock down DOCK8 (Fig. 6A). Impor- tokines such as IFN-g and TNF. TNF is a highly proinflammatory tantly, DOCK8 knockdown NK cells failed to produce IFN-g apical cytokine, which acts upon diverse tissue types to amplify and had attenuated secretion of TNF-a in response to NKp30 inflammation. Importantly, TNF triggers the secretion of numer- ligation (Fig. 6A), supporting our previous data with KHYG1 cells. ous downstream cytokine and chemokines that promote immune We next investigated NK cells isolated from the peripheral blood cell infiltration into the localized site of infection (31). Further- of two patients suffering from AR-HIES as a result of genetic loss more, TNF acts directly on immune cells, enhancing their acti- of DOCK8 (Table I). DOCK8-deficient patient NK cells expressed vation status, and facilitates Ag cross-presentation by promoting NKp30 to comparable levels as healthy donor-derived NK cells macrophage and dendritic cell maturation (32). Indeed, neutrali- (Fig. 6B), but failed to produce IFN-g and TNF-a in response to zation of TNF is often used to treat inflammatory disorders such as

cells were pretreated with ERK inhibitor (cobimetinib, 125 nM) or p38 inhibitor (SB203580, 125 nM) for 15 min, followed by stimulation with anti-NKp30. After 6 h, cytokines in the supernatants were analyzed by CBA. (H) KHYG1 cells were treated with DMSO or cobimetinib (125 nM) for 15 min in the presence of GolgiPlug, then stimulated with anti-NKp30 (1 mg/ml). After 2 h, intracellular accumulation of TNF was measured by FACS staining. Data are representative of at least three independent experiments. Error bars represent the mean 6 SEM of triplicate determinations from a representative experiment. Error bars on graphs of relative killing (C and D) are mean 6 SEM from pooled from three independent experiments. *p , 0.05 by unpaired Student t test. The Journal of Immunology 9

Table I. Characteristics of DOCK8-deﬁcient patients (AR-HIES)

DOCK8-Deficient Patients Mutation Sex Age (y) IgE (IU/ml) Infections Allergies Patient 1 Unidentified: lack Male 5 17,300 HSV, Streptococcus Food allergies (milk, egg, protein expression13 pyogenes, Haemophilus cashew, pistachio, beef, influenzae, Candida lamb) albicans, adenovirus, Eczema norovirus, HHV6, EBV, Asthma CMV, VZV, Aspergillus Bronchiectasis niger, Cladosporium Patient 2 C.5748insC(p.T1917H Male 13 1,716 Molluscum contagiosum, Severe eczema fsX30) recurrent viral and Food, environmental, and bacterial respiratory drug allergies infection Asthma Bronchiectasis Clinical analysis of AR-HIES patients from Fig. 6. HHV6, human herpesvirus 6; VZV, varicella zoster virus.

Crohn’s disease and arthritis. However, these patients become transcriptional-dependent manner. Interestingly, we did not observe vulnerable to opportunistic bacterial infections, highlighting the a similar decrease in cytokine production after activation through Downloaded from importance of TNF in driving both innate and adaptive immu- the NKp44 receptor, suggesting that the DOCK8 may be a speciﬁc nity (33). Similarly, IFN-g is rapidly produced upon NK cell adaptor for the NKp30 receptor. However, we cannot rule out a role activation, where it serves to promote the effector status of for DOCK8 in mediating signaling through other NK cell–activat- cytotoxic lymphocytes and prime APCs to drive initiation of ing receptors without further investigation. adaptive immunity (34). IFN-g is also critical for promoting Another outstanding question in the ﬁeld is the relative con-

antitumor immunity (35, 36), by elevating MHC class I ex- tribution of DOCK8 in promoting the cytoskeletal rearrangements http://www.jimmunol.org/ pression on tumor cells, invoking tumor cell cytostasis and in- that facilitate signal transduction events, versus a more direct role creasing effector cell cytotoxic activity. Our data suggest that in signal initiation/transduction. Indeed, evidence suggests that DOCK8 deficiency in patients is likely to impair NK cell–driven DOCK8 is involved in the physical reorientation of the MTOC immune responses because of a defect in cytotoxicity and in the through its activity as a GEF-dependent activator of Cdc42 (14, 16). secretion of inflammatory cytokines, particularly TNF and IFN-g, We have addressed this question by destabilizing microtubules, which are required for efficient mobilization of the immune sys- which are required for MTOC polarization. As expected, nocodazole tem and subsequent clearance of infections. Reduced NK cell attenuated lytic activity; however, under these conditions, target cell function may also lead to increased susceptibility to malignancies synapse-mediated cytokine production was unaffected. These data in DOCK8-deficient patients because of inefficient immune sur- suggest that rapid, DOCK8-dependent signal transduction events by guest on September 29, 2021 veillance. proceed efficiently and independently of a reorganized cytoskeleton. NK cell activation results in a rapid cascade of signaling events The rapid activation kinetics of Src kinases we have monitored by that culminate in translocation of lytic granules via MTOC po- Western blotting (within 5 min) strongly support this concept, larization in the cytotoxic synapse (12). Early studies into the where signaling to cytokine transcription/translation precedes role of DOCK8 in NK cells identified a defect in cytotoxicity granule delivery to the synaptic cleft. It is also worth noting that this against tumor cell targets (14, 15). This was found to occur would be consistent with many other receptors that promote in- through a failure to polarize key molecules, accumulate actin, flammatory cytokine production. For example, ligation of the TNFR and translocate the MTOC to the immune synapse. However, it results in rapid formation of a signaling complex composed of was unclear whether DOCK8 was directly involved in these nonenzymatic adaptors, such as TRADD, that serve as platforms for cytoskeletal changes, or whether DOCK8 was involved in activation of kinases such as RIPK1, triggering chemokine tran- transducing the signals that promote these events, or indeed both. scription within minutes (38). Further studies will be required to Since these discoveries, several lines of evidence have pointed to determine the exact kinetic relationship between gene transcription, a role for DOCK8 in regulating signal transduction events (30). cytoskeletal rearrangements caused by cytotoxic synapse formation, Indeed, recent studies have identified DOCK8 as a critical cell- and the molecular pathways involved in these processes. intrinsic regulator of CD4+ T cell polarization. In this study, DOCK8 was found to constitutively associate with STAT3 in a GEF-independent manner, but promote STAT3 activation using Acknowledgments GEF activity upon cytokine stimulation (18, 19). Similarly, We thank Dr. Ilia Voskoboinik for critical reading of the manuscript. DOCK8 also regulates the production of IL-31 in CD4+ Tcells by controlling the transcription factor EPAS1 (37). DOCK8 also Disclosures acts in a nonenzymatic role upon TLR9 stimulation in B cells. In The authors have no financial conflicts of interest. this study, DOCK8 bridges the interaction of the Src kinase, Lyn, with Syk, facilitating STAT3 translocation to the nucleus to drive References transcription (17). Importantly, this study demonstrated that 1. Su, H. C. 2010. Dedicator of cytokinesis 8 (DOCK8) deficiency. Curr. Opin. DOCK8 constitutively associates with Src kinases, which is Allergy Clin. Immunol. 10: 515–520. enhanced upon receptor triggering. These observations are con- 2. Zhang, Q., J. C. Davis, I. T. Lamborn, A. F. Freeman, H. Jing, A. J. Favreau, sistent with our data that demonstrate that DOCK8 plays a crucial H. F. Matthews, J. Davis, M. L. Turner, G. Uzel, et al. 2009. Combined im- role in linking the NKp30 receptor to rapid and efficient activation munodeficiency associated with DOCK8 mutations. N. Engl. J. Med. 361: 2046– 2055. of the Src kinase, Lck, which is required for downstream sig- 3. Su, H. C. 2010. Combined immunodeficiency associated with DOCK8 mutations naling that triggers lytic activity and cytokine production in a and related immunodeficiencies. Dis. Markers 29: 121–122. 10 DOCK8 PROMOTES CYTOKINE PRODUCTION FROM NK CELLS

4. Laurin, M., and J. F. Coˆte´. 2014. Insights into the biological functions of dock 21. Oykhman, P., M. Timm-McCann, R. F. Xiang, A. Islam, S. S. Li, D. Stack, family guanine nucleotide exchange factors. Genes Dev. 28: 533–547. S. M. Huston, L. L. Ma, and C. H. Mody. 2013. Requirement and redundancy of 5. Randall, K. L., T. Lambe, A. L. Johnson, B. Treanor, E. Kucharska, the Src family kinases Fyn and Lyn in perforin-dependent killing of Crypto- H. Domaschenz, B. Whittle, L. E. Tze, A. Enders, T. L. Crockford, et al. 2009. coccus neoformans by NK cells. Infect. Immun. 81: 3912–3922. Dock8 mutations cripple B cell immunological synapses, germinal centers and 22. Chini, C. C., M. D. Boos, C. J. Dick, R. A. Schoon, and P. J. Leibson. 2000. long-lived antibody production. [Published erratum appears in 2010 Nat. Regulation of p38 mitogen-activated protein kinase during NK cell activation. Immunol. 11: 644.] Nat. Immunol. 10: 1283–1291. Eur. J. Immunol. 30: 2791–2798. 6. Randall, K. L., S. S. Chan, C. S. Ma, I. Fung, Y. Mei, M. Yabas, A. Tan, 23. Vivier, E., J. A. Nune`s, and F. Ve´ly. 2004. Natural killer cell signaling pathways. P. D. Arkwright, W. Al Suwairi, S. O. Lugo Reyes, et al. 2011. DOCK8 defi- Science 306: 1517–1519. ciency impairs CD8 T cell survival and function in humans and mice. J. Exp. 24. House, I. G., K. Thia, A. J. Brennan, R. Tothill, A. Dobrovic, W. Z. Yeh, Med. 208: 2305–2320. R. Saffery, Z. Chatterton, J. A. Trapani, and I. Voskoboinik. 2015. Heterozy- 7. Crawford, G., A. Enders, U. Gileadi, S. Stankovic, Q. Zhang, T. Lambe, gosity for the common perforin mutation, p.A91V, impairs the cytotoxicity of T. L. Crockford, H. E. Lockstone, A. Freeman, P. D. Arkwright, et al. 2013. primary natural killer cells from healthy individuals. Immunol. Cell Biol. 93: DOCK8 is critical for the survival and function of NKT cells. Blood 122: 2052– 575–580. 2061. 25. Liao, Y., G. K. Smyth, and W. Shi. 2014. featureCounts: an efficient general 8. Zhang, Q., C. G. Dove, J. L. Hor, H. M. Murdock, D. M. Strauss-Albee, purpose program for assigning sequence reads to genomic features. Bio- J. A. Garcia, J. N. Mandl, R. A. Grodick, H. Jing, D. B. Chandler-Brown, et al. informatics 30: 923–930. 2014. DOCK8 regulates lymphocyte shape integrity for skin antiviral immunity. 26. Law, C. W., Y. Chen, W. Shi, and G. K. Smyth. 2014. voom: precision weights J. Exp. Med. 211: 2549–2566. unlock linear model analysis tools for RNA-seq read counts. Genome Biol. 15: 9. Flesch, I. E., K. L. Randall, N. A. Hollett, H. Di Law, L. A. Miosge, Y. Sontani, R29. C. C. Goodnow, and D. C. Tscharke. 2015. Delayed control of herpes simplex 27. Liberzon, A., C. Birger, H. Thorvaldsdo´ttir, M. Ghandi, J. P. Mesirov, and virus infection and impaired CD4(+) T-cell migration to the skin in mouse P. Tamayo. 2015. The molecular signatures database (MSigDB) hallmark gene models of DOCK8 deficiency. Immunol. Cell Biol. 93: 517–521. set collection. Cell Syst. 1: 417–425. 10. de la Roche, M., Y. Asano, and G. M. Griffiths. 2016. Origins of the cytolytic 28. Brandt, C. S., M. Baratin, E. C. Yi, J. Kennedy, Z. Gao, B. Fox, B. Haldeman, C. D. Ostrander, T. Kaifu, C. Chabannon, et al. 2009. The B7 family member

synapse. Nat. Rev. Immunol. 16: 421–432. Downloaded from 11. Dustin, M. L. 2014. The immunological synapse. Cancer Immunol. Res. 2: B7-H6 is a tumor cell ligand for the activating natural killer cell receptor NKp30 in humans. J. Exp. Med. 206: 1495–1503. 1023–1033. 29. Mentlik, A. N., K. B. Sanborn, E. L. Holzbaur, and J. S. Orange. 2010. Rapid 12. Kearney, C. J., A. J. Brennan, P. K. Darcy, and J. Oliaro. 2015. The role of the lytic granule convergence to the MTOC in natural killer cells is dependent on immunological synapse formed by cytotoxic lymphocytes in immunodeﬁciency dynein but not cytolytic commitment. Mol. Biol. Cell 21: 2241–2256. and anti-tumor immunity. Crit. Rev. Immunol. 35: 325–347. 30. Kearney, C. J., K. L. Randall, and J. Oliaro. 2017. DOCK8 regulates signal 13. Orange, J. S. 2007. The lytic NK cell immunological synapse and sequential transduction events to control immunity. Cell. Mol. Immunol. 14: 406–411. steps in its formation. Adv. Exp. Med. Biol. 601: 225–233. 31. Sedger, L. M., and M. F. McDermott. 2014. TNF and TNF-receptors: from 14. Ham, H., S. Guerrier, J. Kim, R. A. Schoon, E. L. Anderson, M. J. Hamann,

mediators of cell death and inflammation to therapeutic giants - past, present and http://www.jimmunol.org/ Z. Lou, and D. D. Billadeau. 2013. Dedicator of cytokinesis 8 interacts with talin future. Cytokine Growth Factor Rev. 25: 453–472. and Wiskott-Aldrich syndrome protein to regulate NK cell cytotoxicity. J. 32. Trevejo, J. M., M. W. Marino, N. Philpott, R. Josien, E. C. Richards, Immunol. 190: 3661–3669. K. B. Elkon, and E. Falck-Pedersen. 2001. TNF-alpha-dependent maturation of 15. Mizesko, M. C., P. P. Banerjee, L. Monaco-Shawver, E. M. Mace, W. E. Bernal, local dendritic cells is critical for activating the adaptive immune response to J. Sawalle-Belohradsky, B. H. Belohradsky, V. Heinz, A. F. Freeman, K. E. Sullivan, virus infection. Proc. Natl. Acad. Sci. USA 98: 12162–12167. et al. 2013. Defective actin accumulation impairs human natural killer cell function in 33. Murdaca, G., F. Spano`, M. Contatore, A. Guastalla, E. Penza, O. Magnani, and patients with dedicator of cytokinesis 8 deficiency. J. Allergy Clin. Immunol. 131: F. Puppo. 2015. Infection risk associated with anti-TNF-a agents: a review. 840–848. Expert Opin. Drug Saf. 14: 571–582. 16. Janssen, E., M. Tohme, M. Hedayat, M. Leick, S. Kumari, N. Ramesh, 34. Schroder, K., P. J. Hertzog, T. Ravasi, and D. A. Hume. 2004. Interferon-gamma: M. J. Massaad, S. Ullas, V. Azcutia, C. C. Goodnow, et al. 2016. A DOCK8- an overview of signals, mechanisms and functions. J. Leukoc. Biol. 75: 163–189. WIP-WASp complex links T cell receptors to the actin cytoskeleton. J. Clin. 35. Zaretsky, J. M., A. Garcia-Diaz, D. S. Shin, H. Escuin-Ordinas, W. Hugo, Invest. 126: 3837–3851. S. Hu-Lieskovan, D. Y. Torrejon, G. Abril-Rodriguez, S. Sandoval, L. Barthly, by guest on September 29, 2021 17. Jabara, H. H., D. R. McDonald, E. Janssen, M. J. Massaad, N. Ramesh, et al. 2016. Mutations associated with acquired resistance to PD-1 blockade in A. Borzutzky, I. Rauter, H. Benson, L. Schneider, S. Baxi, et al. 2012. DOCK8 melanoma. N. Engl. J. Med. 375: 819–829. functions as an adaptor that links TLR-MyD88 signaling to B cell activation. 36. Gao, J., L. Z. Shi, H. Zhao, J. Chen, L. Xiong, Q. He, T. Chen, J. Roszik, Nat. Immunol. 13: 612–620. C. Bernatchez, S. E. Woodman, et al. 2016. Loss of IFN-gamma pathway genes 18. Keles, S., L. M. Charbonnier, V. Kabaleeswaran, I. Reisli, F. Genel, N. Gulez, in tumor cells as a mechanism of resistance to anti-CTLA-4 therapy. Cell 167: W. Al-Herz, N. Ramesh, A. Perez-Atayde, N. E. Karaca, et al. 2016. Dedicator 397–404.e399. of cytokinesis 8 regulates signal transducer and activator of transcription 3 ac- 37. Yamamura, K., T. Uruno, A. Shiraishi, Y. Tanaka, M. Ushijima, T. Nakahara, tivation and promotes TH17 cell differentiation. J. Allergy Clin. Immunol. 138: M. Watanabe, M. Kido-Nakahara, I. Tsuge, M. Furue, and Y. Fukui. 2017. The 1384–1394.e2. transcription factor EPAS1 links DOCK8 deficiency to atopic skin inflammation 19. Tangye, S. G., B. Pillay, K. L. Randall, D. T. Avery, T. G. Phan, P. Gray, via IL-31 induction. Nat. Commun. 8: 13946. J. B. Ziegler, J. M. Smart, J. Peake, P. D. Arkwright, et al. 2017. DOCK8- 38. Kearney, C. J., C. Sheridan, S. P. Cullen, G. A. Tynan, S. E. Logue, I. S. Afonina, deficient CD4+ T cells are biased to a Th2 effector fate at the expense of Th1 D. Vucic, E. C. Lavelle, and S. J. Martin. 2013. Inhibitor of apoptosis proteins and Th17 cells. J. Allergy Clin. Immunol. 139: 933–949. (IAPs) and their antagonists regulate spontaneous and tumor necrosis factor 20. Lanier, L. L. 2008. Up on the tightrope: natural killer cell activation and inhi- (TNF)-induced proinflammatory cytokine and chemokine production. J. Biol. bition. Nat. Immunol. 9: 495–502. Chem. 288: 4878–4890. A

WT - IgG WT - IgG KO - IgG KO - IgG WT - NKp30 WT - NKp44 KO - Nkp30 KO - Nkp44

NKp30 - PE NKp44 - PE Supplemental Figure 1: A. K562 cells were seeded in chamber slides using serum free media, then overlaid with control or DOCK8 knockout KHYG1 NK cells. After 30 minutes, conjugates were fixed, permeablilised and stained as indicated for confocal microscopy analysis. Blue = DAPI. Representitive images are shown (n=20 for each protein). B. Control or DOCK8 knockout KHYG1 cells were stained with anti-NKp30 or anti-NKp44 and analyzed by flow cytometry.

A Replicate 2 Control KO 30 0 15 30 α-NKp30 (mins:) 0 5 15 60 5 60 phospho ERK (60 min) 50 p ERKThr 202/Tyr 204 * 1.0 *

50 ERK

Replicate 3 0.5 Control KO α-NKp30 (mins:) 0 5 15 30 60 0 5 15 30 60

50 RelativePhosphorylation p ERKThr 202/Tyr 204

50 WT KO ERK

Relating to Figure 4C B Replicate 2 Scramble DOCK8 α-NKp30 (mins:) 0 5 15 30 60 0 5 15 30 60 DOCK8 25 0 p SrcTyr 416 75 phospho Src (15 min) * 75 Src 1.0 *

Replicate 3 Scramble DOCK8 0.5 NKp30 (mins) 0 15 30 0 15 30 25 DOCK8

0 RelativePhosphorylation p Src 75 WT KO 75 p Src (light)

75 Src

Supplemental Figure 2: A. Two aditional replicates for phospho ERK, from the experiment in Figure 4B. Relative densitometry at 60 minutes is displayed on the right. B. Two additional replicates for phospho Src, from the experiment in Figure 4C. Relative densitometry at 15 minutes is displayed on the right. Error bars represent mean +/- SEM of 3 pooled independent experiments. * p < 0.05 Supplemental Table I: Top 100 (1-50) NKp30 Regulated Genes

WT UT WT UT WT NKP30 WT NKP30 KO UT KO UT KO NKP30 KO NKP30 1 2 1 2 1 2 1 2 ARL5B 3.13 2.36 5.53 5.2 2.57 2.55 5.5 5.44 ATF3 1.77 2.3 5.54 5.09 2.86 2.71 4.88 4.81 BCL2L11 4.53 4.42 5.91 5.85 4.61 4.32 5.71 5.78 BIRC3 7.68 7.81 10.03 9.86 7.32 7.32 9.18 9.07 BTG2 2.5 2.72 7.7 7.31 3.44 3.4 7.36 7.15 CD83 4.86 4.91 7.15 6.93 4.82 4.43 6.79 6.79 CDKN1A 3.69 3.92 5.95 5.7 2.81 2.17 5.01 4.9 CISH 7.66 7.66 6.05 6.13 7.75 7.29 6.13 6.32 CREM 4.35 4.31 6 5.98 4.85 4.46 6.17 5.95 CRTAM 2.22 2.55 6.95 6.58 0.69 0 4.95 4.81 CSF1 3.44 3.35 5.6 5.38 1.86 2.88 4.79 4.99 CSRNP1 4.51 4.33 7.39 7.01 5.07 4.46 6.75 6.74 CYTOR 6.36 6.43 7.55 7.28 5.77 6.38 7.51 7.42 DUSP2 5.71 5.65 9.62 9.31 5.89 5.98 9.36 9.28 DUSP5 3.86 3.76 7.08 6.94 3.94 4.18 7.2 6.84 EGR1 5.19 5.35 9.17 8.87 6.32 5.8 8.53 8.73 EGR2 1.48 1.22 8.38 8.11 3.02 2.65 7.45 7.73 EGR3 -4.29 -2.03 4.19 4.06 -3.7 -4.29 4.03 3.52 EGR4 1.73 1.64 8.38 8.15 2.36 2.25 7.72 7.67 EVI2A 3.06 3.27 6.52 6.42 3.59 3.49 6.23 6 EVI2B 2.99 3.14 4.9 4.7 2.9 3.29 4.53 4.65 FASLG 5.65 5.55 9.64 9.39 6.59 5.64 9.04 8.83 FBXO30 3.38 3.09 4.84 4.54 2.81 2.68 4.88 4.57 FOSB 3.74 3.52 7.8 7.53 5.01 4.37 7.06 7.24 FOSL2 3.72 3.72 7.51 7.39 4.82 4.71 7.63 7.71 GADD45B 4.84 4.89 7.14 7.05 4.7 4.43 7.17 7.09 GBP2 3.72 3.75 6.38 6.27 2.9 3.22 5.02 5.03 GPR18 3.45 3.2 5.13 5.04 3.59 3.71 4.98 4.87 GZMB 5.8 5.95 8.8 8.66 4.92 4.87 8.08 7.88 ICAM1 3.78 3.5 6.16 6.01 3.29 3.26 5.71 5.46 IER3 8.98 9.22 11.67 11.34 7.6 8.87 11.31 10.94 IFNG 1.87 1.45 7.85 7.16 1.56 2.21 5.91 5.46 IL10 4.78 4.7 9.21 8.78 4.67 4.04 7.58 7.26 JUN 5.69 5.6 8.14 7.69 5.82 5.57 7.22 7.32 KDM6B 2.69 2.81 6.75 6.7 2.39 2.6 6.06 6.46 KLF9 2.89 2.92 5.66 5.51 3.33 3.28 5.57 5.5 LMO4 6.33 6.07 7.03 7.12 6.14 6.17 7.01 7.05 MAP2K3 7.19 7.08 8.46 8.46 7.34 7.12 8.56 8.77 MAP3K8 4.56 4.49 7 6.77 4.82 5.12 6.67 6.65 MCL1 8.16 8.04 9.44 9.38 8.01 7.64 9.03 9.1 MIR155HG 2.46 2.72 6.52 5.98 2.33 2.01 4.46 4.68 MVB12B 1.51 1.94 4.13 3.98 1.82 1.82 3.86 3.42 NAB1 3.6 3.68 4.85 5.04 3.26 3.2 5.18 4.89 NAB2 5.05 4.85 6.11 6.06 4.63 4.53 6.31 6.59 NAMPT 7.15 7.12 9.36 9.16 7.48 7.7 9.21 9.15 NF1 3.86 3.88 5.23 5.27 3.92 3.7 4.99 5 NFKBIA 8.85 8.79 10.84 10.64 8.67 8.9 10.18 10.19 NFKBID 1.18 1.05 5.8 5.65 1.22 1.45 5.02 4.42 NINJ1 1.56 1.64 3.72 3.55 0.33 0.74 3.65 3.46 NR4A2 -0.04 0.16 8.52 8.09 1.71 2.17 7.76 7.74 NR4A3 -0.62 0.57 8.55 8.33 1.43 1.5 8.2 8.25

Normalised Log2 expression data from wildtype or DOCK8 knockout KHYG1 cells, either left untreated or treated with anti-NKp30 (n=2). Data was used to generate the heat maps in Figure 2 of the manuscript.

Supplemental Table II: Top 100 (50-100) NKp30 Regulated Genes

WT UT WT UT WT NKP30 WT NKP30 KO UT KO UT KO NKP30 KO NKP30 1 2 1 2 1 2 1 2 PFKFB3 4.28 4.45 6.79 6.51 4.59 4.28 6.14 6.36 PHLDA1 5.58 5.83 9.46 9.12 5.49 5.94 8.81 8.68 PIM3 9.97 9.95 11.31 11.22 8.86 9.27 10.82 10.87 PNRC1 2.27 2.28 4.26 4.13 3.13 2.58 4.16 4.26 PPP1R15A 5.79 5.85 8.17 7.89 5.68 5.84 7.35 7.12 PRDM1 2.62 2.74 5.83 5.41 3.59 3.33 4.97 5.03 PRF1 5.56 5.45 6.22 6.36 4.87 5.08 5.96 5.9 PRNP 4.78 4.71 7.88 7.56 5.51 5.3 7.07 7.07 PTGER4 2.14 2.1 4.67 4.48 2.1 1.56 4.24 3.88 RAB11FIP1 6.96 7.09 8.54 8.52 7.46 7.52 8.58 8.55 RDH10 2.87 3.2 5.76 5.4 4.5 3.97 6.01 5.85 REL 3.22 3.83 6.4 6.5 4 3.74 6.29 6.36 RGS16 6.82 6.82 8.6 8.54 7.26 6.43 8.5 8.59 RILPL2 5.65 5.7 7.33 7.39 5.48 5.08 6.85 6.88 RNF19A 4.29 4.34 6.72 6.74 3.97 4.08 6.41 6.34 RPSAP12 6.94 6.65 6.3 6.09 7.5 7.54 6.6 6.5 SAT1 3.65 3.81 6.52 6.25 4.25 3.85 5.51 5.42 SDC4 3.69 3.88 6.36 6.28 2.17 3.15 6.12 6.03 SEMA7A 3.32 3.43 5.67 5.37 3.11 2.71 5.47 5.52 SEPHS2 6.61 6.62 5.66 5.73 6.88 7.07 5.91 6.04 SERPINE1 4.11 4.27 7.56 7.27 3.47 4.07 6.51 6.37 SERTAD1 3.99 4.08 6 5.97 4.08 4.15 6.03 5.91 SIAH2 5.02 4.83 7 6.85 4.93 5.13 6.68 6.59 SLA 5.46 5.41 7.35 7.23 4.36 4.64 5.73 5.63 SLC30A1 4.74 4.56 4 3.95 5.23 4.98 3.99 4.1 SOCS1 4.63 4.64 6.19 6.16 4.58 4.68 6.22 6.42 SPAG1 2.76 2.62 4.83 4.67 2.48 2.48 4.54 4.16 STAT6 5.05 5.16 5.8 5.7 4.49 4.85 5.97 5.84 TBX21 8.18 8.12 9.21 9.29 7.52 7.75 9.61 9.59 TMEM2 5.55 5.66 7.52 7.38 6.41 5.72 7.53 7.71 TNF 6.13 6.11 8.96 8.72 6.05 6.24 8.68 8.5 TNFAIP3 4.62 4.71 6.62 6.34 4.36 4.22 5.7 5.66 TNFSF14 5.21 5.68 9.12 8.89 5.06 5.13 8.53 8.11 TNFSF9 2.01 2.21 6.88 6.47 2.14 2.29 5.72 5.43 TRAF1 2.6 2.95 5.56 5.43 1.22 1.56 4.84 4.33 TRIB1 -2.34 0.57 7.76 7.53 0.79 1.01 7.09 6.93 TSC22D2 2.6 2.33 4.74 4.37 2.61 2.94 4.35 4.43 XCL1 7.02 6.9 10.89 10.52 6.51 6.57 10.1 10.04 XCL2 6.83 6.62 10.43 10.02 6.54 6.48 9.77 9.76 XIRP1 -2.03 -2.43 5.77 5.35 -3.7 -1.64 4.73 4.75 ZC3H12A 3.3 3.24 6.14 5.77 3.11 3.16 5.47 5.49 ZFP36L1 7.94 7.93 11.02 10.83 7.55 8.06 10.34 10.34