Transcriptional regulation of Munc13-4 expression in cytotoxic lymphocytes is disrupted by an intronic mutation associated with a primary immunodeficiency

Understanding how Munc13-4 expression is controlled in distinct lymphocyte subsets is of interest considering the pivotal role of this protein for lymphocyte cytotoxicity (Feldmann et al., 2003). A ubiquitous Munc13-4 expression pattern was initially reported by Koch and colleagues (Koch et al., 2000), but a more detailed analysis has not since been performed. To compare Munc13-4 expression levels between cytotoxic and noncytotoxic lymphocytes, we isolated lymphocyte subsets from the peripheral blood of healthy donors and performed Western blots for Munc13-4. Relative to CD19⁺ B cells, CD3⁻CD56⁺ NK cells expressed very high levels of Munc13-4. Moreover, CD4⁺ T cells expressed low levels of Munc13-4, whereas total CD8⁺ T cells expressed intermediate levels. To determine whether Munc13-4 is concomitantly induced along with the acquisition of lytic granules, we compared expression in CD56^bright and CD56^dim NK cells as well as in naive and effector CD8⁺ T cells. In both NK cells and CD8⁺ T cells, maturation and acquisition of lytic granules was associated with a marked increase in Munc13-4 expression, indicating that Munc13-4 is up-regulated during differentiation of cytotoxic lymphocyte subsets (Fig. 1, a and b). High levels of Munc13-4 expression thus correlate with a strong propensity to degranulate in response to engagement of activating receptors on cytotoxic T cells or NK cells (Chiang, et al., 2013). Notably, expression levels of syntaxin-11 and Munc18-2, also encoded by genes associated with FHL, were less correlated with cytotoxicity in lymphocyte subsets (Fig. 1, a and b). Thus, the data suggest an important role for induction of Munc13-4 expression in regulating the cytotoxic capacity of lymphocytes.

The marked increase in Munc13-4 expression upon differentiation indicated that lineage-specific mechanisms promote high levels of Munc13-4 expression. Recently, we identified a c.118-308C>T mutation in an evolutionarily conserved region deep within intron 1 of UNC13D (Fig. 2 a) in patients with FHL3 that leads to a lymphocyte-specific loss of Munc13-4 expression (Meeths et al., 2011). NK cells from these individuals exhibit defective degranulation (Fig. 2 b). These results led us to hypothesize that a regulatory region critical for Munc13-4 expression in cytotoxic lymphocytes is located within intron 1 of UNC13D. We used transcription factor binding prediction software (Heinemeyer et al., 1998) to identify potential binding sites within intron 1 of UNC13D that might be disrupted by the mutation and found that a consensus ETS site overlaps with the location of the c.118-308C>T mutation (Fig. 2 c). To determine whether an ETS family member binds to this site, we synthesized wild-type and mutant oligonucleotide probes corresponding to the sequences shown in Fig. 2 a for use in EMSAs with nuclear lysates from primary NK cells and the NK92 cell line. A clearly discernable protein complex bound to the wild-type UNC13D probe that was absent from the probe with the c.118-308C>T mutation. This protein complex migrated at the same size as a probe containing the consensus sequence for the ETS family member ELF1, and addition of an antibody against ELF1 resulted in a super-shift (Fig. 2 d). No super-shifts were observed with antibodies against other ETS family members, including ETS1 and ELF4 (unpublished data).

While examining the UNC13D intron 1 region, we identified potential binding sites for several other transcription factors that are associated with effector functions of cytotoxic lymphocytes, including AML/RUNX, NF-κB, and STAT, as well as an evolutionarily conserved transcriptional start site downstream of the ETS site (Fig. 2 c). This led us to hypothesize that the UNC13D intron 1 region could represent a lymphocyte-specific alternative promoter. To test this, we performed 5′-RACE on cDNA from primary peripheral blood NK cells and CD8⁺ T cells using an antisense primer specific for exon 6 of UNC13D. This yielded both the conventional UNC13D transcript and an additional transcript originating within intron 1 that contained an alternative first exon that is spliced in frame with the downstream coding exons (Fig. 2 e). Corroborating our findings, UNC13D transcripts initiating from the intron 1 promoter are present within the NCBI EST dataset (EST CR983520) for lymphocytes, and ELF1 binding to intron 1 is present in ENCODE chromatin immunoprecipitation (ChIP) data for K562 cells (ENCODE Project Consortium, 2011).

To further elucidate the nature of the UNC13D intron 1 regulatory element, we designed primers and probes specific for the conventional and intron 1 UNC13D transcripts and performed qRT-PCR on sorted CD19⁺ B cells, CD4⁺ T cells, CD8⁺ T cells, CD3⁻CD56⁺ NK cells, and monocytes from healthy donors and an individual homozygous for the UNC13D c.118-308C>T mutation. Conventional UNC13D transcripts were expressed at comparable levels between subsets in healthy donors and were significantly lower, but still readily detectible, in all analyzed cellular subsets from the individual with the UNC13D c.118-308C>T mutation (Fig. 3 a). UNC13D transcripts originating from intron 1 were expressed in a lymphocyte–specific pattern in healthy donors, with the highest levels observed in CD8⁺ T cells and CD3⁻CD56⁺ NK cells. Intron 1 transcripts were observed at the limit of detection in B cells and were not detectable in CD4⁺ T cells, CD8⁺ T cells CD3⁻CD56⁺ NK cells, or monocytes from the homozygous UNC13D c.118-308C>T patient (Fig. 3 b).

To test the transcriptional activity of the UNC13D intron 1 promoter, we created a luciferase reporter vector and tested its activity in the readily transfectable YT-Indy NK cell line. Luciferase expression levels from the intron 1 promoter were comparable to those of the full-length UNC13D promoter. We also created a version of the intron 1 promoter reporter vector with the UNC13D c.118-308C>T mutation. Remarkably, luciferase levels were not statistically different between the wild-type and mutated versions (Fig. 4 a), suggesting that the ELF1-binding site is not essential for promoter transactivation. Moreover, CD19⁺ B cells, CD4⁺ T cells, CD8⁺ T cells, and CD3^–CD56⁺ NK cells all expressed similar levels of ELF1 (Fig. 4, b and c). Together, the results suggest that the UNC13D intron 1 regulatory element serves as both an enhancer for the conventional promoter and a lymphocyte-specific alternative promoter, and the intronic ELF1 site alone may not explain the observed increase in Munc13-4 expression upon cytotoxic lymphocyte differentiation.

Because the c.118-308C>T mutation did not significantly affect promoter transactivation in luciferase reporter assays and ELF1 expression levels do not correlate with lymphocyte cytotoxicity, we hypothesized that an intact ELF1-binding site is necessary for the recruitment of additional factors that facilitate the establishment of open chromatin within the UNC13D regulatory region, ultimately leading to enhanced Munc13-4 expression. Notably, two potential STAT binding sites are located in close proximity to the ETS consensus site in intron 1 of UNC13D (Fig. 2 c). Furthermore, the expression pattern of STAT4 in lymphocyte subsets (Fig. 5, a and b) mirrors that of Munc13-4 (Fig. 1, a and b). To determine whether these putative STAT sites were active, we synthesized probes containing either the STAT site 5′ to the ELF1 site or 3′ to the ELF1 site and tested them in an EMSA with NK92 nuclear protein lysate. A protein complex was clearly visible with the 5′ STAT probe, but not the 3′ STAT probe (Fig. 5 c). To test whether there is interdependence between the ELF1 and 5′ STAT sites, we synthesized 87 bp probes containing either the wild-type or c.118-308C>T mutation along with either the 5′ STAT or 3′ STAT site. In EMSA assays using nuclear lysate from primary NK cells, the presence of a protein complex was evident on the probe with the wild-type ELF1 site and 5′ STAT site, and the c.118-308C>T mutation led to a reduction in binding. No discernable protein complex was observed with probes containing the 3′ STAT site (Fig. 5 d, left). Using nuclear protein from NK92 cells, we also observed the presence of a protein complex on the probe with the wild-type ELF1 site and the 5′ STAT site that was diminished in the presence of the c.118-308C>T mutation. Importantly, binding of the protein complex could be almost entirely competed away with cold probes containing either the consensus ELF1 site or the consensus STAT4 site (Fig. 5 d, right). Binding of the protein complex to the probe could also be inhibited by the addition of an antibody against STAT4 and to a somewhat lesser extent with the addition of an antibody against ELF1 (Fig. 5 d right). The weak effect of the ELF1 antibody in this assay is likely caused by epitope masking by bound STAT4. STAT4 binding to intron 1 of UNC13D was confirmed in NK cells by ChIP analysis (Fig. 5 e).

To more definitively demonstrate specificity and determine whether the c.118-308C>T mutation abrogates STAT4 binding to intron 1 of UNC13D in the context of native chromatin, we took advantage of linked polymorphisms surrounding the mutation to design PCR primers that specifically amplify the wild-type or mutant UNC13D alleles (Fig. 5 f). These primers were used in a ChIP experiment with NK cells from an individual heterozygous for the c.118-308C>T mutation. Using this approach, we could detect STAT4 binding to the wild-type allele, but not to the mutant allele (Fig. 5 g).

Zhang and Boothby (2006) demonstrated that STAT4 recruits the Switch (swi)-sucrose nonfermenter (SNF) component Brahma-related gene 1 (BRG1) to the IFN-γ gene in CD4⁺ T cells undergoing Th1 differentiation. The Swi–SNF remodeling complex uses ATP hydrolysis to slide histone octamers and enhance nucleosomal DNA accessibility, which is a fundamental requirement for transcription (Narlikar et al., 2002). BRG1 recruitment is associated with high levels of H3K27 acetylation on surrounding chromatin (Rada-Iglesias et al., 2011). To determine whether BRG1 is recruited to the intron 1 promoter of UNC13D and is associated with high H3K27ac levels, we performed BRG1 and H3K27ac ChIP analysis with cells from the UNC13D heterozygous donor. Using allele-specific primers, we were able to detect BRG1 binding to the wild-type allele, whereas binding was absent on the mutant allele (Fig. 5 h). Similarly, we observed strong enrichment of H3K27ac on the wild-type allele, whereas the mutant allele exhibited reduced levels of H3K27ac (Fig. 5 i). Therefore, an intact ELF1-binding site is necessary for the recruitment of STAT4 and BRG1 and high levels of active chromatin marks within intron 1 of UNC13D. Additionally, we analyzed BRG1 expression in lymphocyte subsets and found a strong positive correlation between BRG1 levels and differentiation of cytotoxic lymphocyte subsets (Fig. 5, j and k).

To analyze UNC13D with high resolution, we designed 24 primer pairs covering a region extending from –2,433 to +12,689 nt relative to the translational start site (Fig. 6 a). These primers were used to identify nucleosome-depleted, open chromatin using the formaldehyde-assisted isolation of regulatory elements (FAIRE) approach (Simon et al., 2012). We observed an enrichment of open chromatin in the conventional and intron 1 promoters in NK cells relative to B cells (Fig. 6 b), which express very low levels of STAT4 (Fig. 5, a and b). Similarly, effector CD8⁺ T cells displayed an enrichment of open chromatin in the promoter and intron 1 regions relative to naive CD8⁺ T cells. High-resolution ChIP analysis with an antibody against STAT4 showed that STAT4 (Fig. 6 c) and BRG1 (Fig. 6 d) binding are enriched within intron 1 in NK cells and effector CD8⁺ T cells. The active histone marks (Karlić et al., 2010) H3K4me3 (Fig. 6 e) and H3K27ac (Fig. 6 f) were enriched throughout the conventional and intron 1 promoters in NK cells and effector CD8⁺ T cells. Collectively, the ChIP data indicate a poised environment surrounding the first exon of UNC13D in all lymphocyte subsets, with an increase in STAT4 and BRG1 binding in NK cells and effector T cells concomitant with increases in active histone marks.

Next, we sought to address the question of how signaling may induce Munc13-4 expression. We chose to focus on CD8⁺ T cells for these experiments due to low basal levels of STAT4 in naive CD8⁺ T cells and the fact that STAT4 is dramatically up-regulated in effector CD8⁺ T cells (Fig. 5, a and b), which exhibit strong degranulation, relative to naive CD8⁺ T cells (Chiang, et al., 2013). To this end, we stimulated naive CD8⁺ T cells for 16 h with IL-2 and IL-12, IFN-γ, IFN-α, or anti-CD3/CD28 beads and determined Munc13-4, STAT4, phosphorylated STAT4 (pY693; pSTAT4), and BRG1 protein expression levels by Western blot. Cytokine stimulation alone did not induce expression of any of these proteins. However, CD3/CD28 stimulation robustly increased the expression of all three proteins along with robust pSTAT4 (Fig. 7, a and b). Phosphorylation on tyrosine 693 is important for STAT4 transcriptional activity (Visconti et al., 2000). Of note, we observed an approximately threefold induction of both conventional and alternative intron 1 transcripts in CD3/CD28-stimulated cells (unpublished data).

To more accurately delineate relative temporal expression patterns, we stimulated naive CD8⁺ T cells with CD3/CD28 beads for 2, 4, 8, or 16 h and performed Western blots. We observed rapid induction of BRG1 between 2 and 4 h, followed by STAT4 induction between 4 and 8 h. Significant Munc13-4 induction was observed between 8 and 16 h (Fig. 7, c and d). Thus, antigen receptor stimulation is sufficient to induce expression of all three proteins, with more rapid kinetics observed for BRG1 and STAT4. To determine whether STAT4 is activated before Munc13-4 induction, we stimulated naive CD8⁺ T cells with CD3/CD28 beads for 0.5, 1, 2, 4, 6, 8, 12, or 16 h and performed Western blots for total and phosphorylated STAT4. Induction of pSTAT4 was consistently observed between 2 and 4 h and increased over time, with a peak at 12 h after stimulation (Fig. 7, e and f). Therefore, STAT4 phosphorylation precedes Munc13-4 induction in naive CD8⁺ T cells upon engagement of the TCR signaling pathway and could thus promote Munc13-4 expression upon T cell activation.

To establish a direct link between STAT4 and the induction of Munc13-4 expression after TCR stimulation, naive CD8⁺ T cells were isolated from peripheral blood and transfected with either control siRNA or siRNA-targeting STAT4 (siSTAT4). Upon CD3/CD28 stimulation, STAT4 was highly induced in cells transfected with the control siRNA. No STAT4 induction was observed in cells transfected with siSTAT4, confirming the efficacy of the knockdown. As expected, Munc13-4 was induced upon stimulation in cells transfected with control siRNA. Notably, Munc13-4 protein levels were 76 ± 12% lower in siSTAT4-transfected cells compared with siRNA controls, supporting the hypothesis that STAT4 is important for Munc13-4 induction. BRG1 was comparably induced in both control and siSTAT4 transfected cells, suggesting that TCR signaling induces BRG1 expression independent of STAT4 (Fig. 8, a and b). These results provide conclusive evidence demonstrating a significant role for STAT4 in the regulation of Munc13-4 expression.

By studying a mutation that causes specific defects in lymphocyte cytotoxicity, we have delineated a critical role for STAT4 in regulating Munc13-4 expression upon cytotoxic lymphocyte differentiation and identified a novel, evolutionarily conserved promoter within intron 1 of UNC13D. The intronic c.118-308C>T mutation causative of FHL3 abolished an ELF1-binding site that is necessary for the recruitment of STAT4 and BRG1 as well as the acquisition of histone marks permissive for transcription. TCR stimulation of naive CD8⁺ T cells led to robust induction of Munc13-4, STAT4, and BRG1, and knockdown of STAT4 abrogated induction of Munc13-4 expression. Our data show that the UNC13D intronic sequence can act both as an alternative promoter and as a general enhancer, promoting Munc13-4 expression upon cytotoxic lymphocyte differentiation.

A surprising finding from our analysis of lymphocyte subsets was that Munc13-4 expression levels are selectively up-regulated upon cytotoxic lymphocyte differentiation. Feldmann et al. (2003) reported similar levels of UNC13D mRNA transcripts by conventional RT-PCR analysis in all hematopoietic cells analyzed including CD4⁺ and CD8⁺ T cells and CD19⁺ B cells, but did not explore expression in well-defined naive and cytotoxic subsets. Given that Munc13-4 protein levels closely correlate with degranulative and cytotoxic potential (Chiang et al., 2013), it is likely that Munc13-4 expression represents a critical rate-limiting step for cytotoxic lymphocyte degranulation, as previously shown for platelets (Shirakawa et al., 2004).

The ability of a cell to respond to environmental stimuli and appropriately alter gene expression is a fundamental component of evolutionary adaptation. The activation or repression of cell-type-specific gene expression programs takes place at the level of chromatin and is influenced by regulatory elements such as enhancers. The c.118-308C>T mutation causative of FHL3 abolished an ELF1-binding site. ELF1 was expressed at uniform high levels in disparate lymphocyte subsets and did not correlate with expression of Munc13-4, suggesting that other transcription factors might be implicated in induction of Munc13-4 expression upon differentiation. A series of elegant experiments by Vahedi et al. (2012) identified STATs, particularly STAT1 and STAT4, as predominant players in shaping the active global enhancer landscape during CD4⁺ T cell fate specification. Interestingly, phenotype-defining master regulators such as T-bet had only a modest impact upon the active enhancer landscape.

Our data demonstrate that STAT4 can bind upstream of the ELF1-binding site and that the c.118-308C>T mutation interferes with STAT4 recruitment to intron 1 of UNC13D. A strong case can be made for STAT4 as a central transcriptional regulator of effector function in cytotoxic lymphocytes, as it also regulates the expression of perforin (Yamamoto et al., 2002), IFN-γ (Zhang and Boothby, 2006), IL-12R (Letimier et al., 2007), T-bet (Yang et al., 2007), and type 1 IFN (Miyagi et al., 2007). During mouse T_H1 cell differentiation, STAT4 and BRG1 act together in regulation of the Ifng promoter (Zhang and Boothby, 2006) and Il12rb2 enhancer (Letimier et al., 2007). On a genome-wide scale, STAT4 appears to facilitate T_H1 cell differentiation by promoting the acquisition of active epigenetic marks, particularly tri-methylation of lysine 4 on histone 3, at intergenic and promoter-associated binding sites (Wei et al., 2010). We show that STAT4 plays a similar role in altering the epigenetic state of UNC13D to allow for transcriptional activation in NK cells and CD8⁺ T cells. The requirement for STAT4 in UNC13D transcriptional regulation may provide an explanation for defective cytotoxicity observed in STAT4-deficient mice (Kaplan et al., 1996). Therefore, STAT4 also represents a driver of epigenetic programming required for degranulation by cytotoxic lymphocytes.

Whether STAT4 directly interacts with ELF1 is also an open question. In a study by Serdobova et al. (1997), ELF1 was shown to bind cooperatively, but not directly, with STAT5 to an IL-2–responsive enhancer in mouse Il2ra. Our data demonstrates that a similar cooperative interaction between ELF1 and STAT4 occurs within intron 1 of UNC13D. Furthermore, two mechanisms have been identified for the recruitment of BRG1 and the associated Swi–SNF complex to regions of regulatory DNA. In a model proposed by Hassan et al. (2001), retention of SWI–SNF is mediated by histone acetyltransferase (HAT) complexes and acetylated histones. The second mechanism for BRG1 recruitment is through interactions with transcription factors. STAT2, which displays 42% homology amino acid to STAT4, has been shown to directly interact with BRG1 in cell lines (Huang et al., 2002). However, we were not able to detect a direct interaction between STAT4 and BRG1 in primary human NK cells (unpublished data).

Although our experiments have focused on the role of STAT4 in regulation of Munc13-4 expression and shown that knockdown of STAT4 can reduce the TCR-induced up-regulation of Munc13-4 expression by 76% compared with siRNA controls (Fig. 8, a and b), other STAT family members may well contribute to transcriptional regulation of UNC13D. ENCODE ChIP-seq data show binding peaks for STAT1, STAT2, STAT3, and STAT5A within intron 1 and the proximal UNC13D promoter in IFN-α–stimulated K562 cells (ENCODE Project Consortium, 2011). ChIP-seq analysis for STAT4 binding is not available within this dataset. The binding of each individual STAT member to the UNC13D locus in primary cytotoxic lymphocytes and their functional significance will require a more extensive analysis.

Our results demonstrate a direct role for STAT4 in regulating genes required for lymphocyte cytotoxicity downstream of TCR engagement in naive CD8⁺ T cells. This finding is supported by in vitro TCR stimulation experiments with CD4⁺ T cells (Bacon et al., 1995) and a recent in vivo study of CD8⁺ T cell activation by lymphocytic choriomeningitis virus (LCMV) infection in the mouse where stimulation through the TCR induced elevated STAT4 expression and was important for antigen-specific proliferation and activation (Gil et al., 2012). The link between STAT4 and Munc13-4 may have implications in the study of diseases that have been linked to STAT4 polymorphisms. A genetic variant of the STAT4 locus that results in reduced levels of STAT4 mRNA expression is associated with an increased risk of hepatitis B virus-related hepatocellular carcinoma (Jiang et al., 2013). Association studies have also identified STAT4 as a genetic risk factor for rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), and Sjögren’s syndrome. The genetic associations with STAT4 involve SNP variants located within the gene’s third intron. Despite clear clinical associations between STAT4 polymorphisms and autoimmunity, a mechanistic explanation for the role of STAT4 is lacking (Remmers et al., 2007; Korman et al., 2008). The identification of STAT4 targets critical for cytotoxic lymphocyte function, such as UNC13D, may be relevant in the context of complex autoimmune syndromes.

Remarkably, our data also revealed that the UNC13D c.118-308C>T mutation specifically abrogates transcription of an alternative form of Munc13-4 in lymphocytes. Residual transcripts of the conventional Munc13-4 isoform appear not to be sufficient for supporting any cytotoxic lymphocyte degranulation. The predicted alternative Munc13-4 protein has a unique amino acid sequence at the amino terminus and is 19 amino acids shorter than the conventional Munc13-4 protein. To our knowledge, no disease causing UNC13D mutations have hitherto been reported in the conventional first coding exon of UNC13D. Thus, it is possible that the alternative, lymphocyte-specific isoform rather than the conventional isoform of Munc13-4 is critical for lymphocyte cytotoxicity. It is therefore of high interest to study the trafficking and function of this alternative Munc13-4 isoform in the setting of cytotoxic lymphocyte degranulation.

In conclusion, the results of this study provide novel insights into the regulatory networks that control lymphocyte cytotoxic function and explain different UNC13D intron 1 mutations that represent a frequent cause of FHL3 (Meeths et al., 2011; Entesarian et al., 2013; Seo et al., 2013). Specific mutations that interrupt immune transcription factor binding within cis-regulatory elements likely represent a mechanism causative of many known primary immunodeficiencies, as well as other immune diseases, warranting further investigation.

FHL patients were included in this study, as were individuals with monoallelic mutations and/or polymorphisms in UNC13D. The patients have been previously described (Meeths et al., 2011). The regional ethics committee at Karolinska Institutet approved this study, and written consent was obtained from all participants.

Peripheral blood mononuclear cells were freshly isolated from buffy coats by density gradient centrifugation. Untouched lymphocyte subsets were further isolated by negative magnetic bead selection (Miltenyi Biotec or STEMCELL Technologies, Inc.). CD19⁺ B cells, CD4⁺ T cells, CD8⁺ T cells, CD3-CD56⁺ NK cells, monocytes, CD56^bright NK cells, and CD56^dim NK cells were isolated by cell sorting (FACSAria; BD) in some experiments.

Cells were stained with fluorochrome-conjugated antibodies and incubated for 30 min on ice. The following antibodies were used for labeling cells before sorting: CD56 PE-Cy7 (BD), CD3 PE-Cy5.5 (Invitrogen), CD19 APC-Cy7 (BD), CD8 QDot 705 (Invitrogen), CD45 QDot 655 (Invitrogen), and CD57 Pacific Blue (BioLegend). Cells were sorted with a FACSAria II (BD).

Cells were lysed on ice for 30 min in lysis buffer (20 mM Tris-HCl, 150 mM NaCl, 1% Triton-X 100, 2 mM EDTA, 0.2% SDS, and complete protease inhibitor (Roche). Protein lysates were quantified with Bradford analysis (Thermo Fisher Scientific), and 10 µg of each lysate was separated through SDS-PAGE gel electrophoresis (Invitrogen). Protein was then transferred onto Immobilon Transfer Membranes (EMD Millipore) through the fast semi-dry transfer method (Thermo Fisher Scientific). Membranes were stained with anti–Munc13-4, anti–Munc18-2, anti-syntaxin 11 (Proteintech), anti-STAT4, anti-pSTAT4 (Abcam), anti-BRG1 (Abcam), anti-ELF1 (Santa Cruz Biotechnology, Inc.), or anti–β-actin (Sigma-Aldrich) primary antibodies. Secondary staining was performed with anti–rabbit or anti–mouse antibodies conjugated to horseradish peroxidase (Invitrogen). Chemiluminescence was visualized with an ImageQuant LAS 4000 Mini (GE Healthcare), and densitometry values were quantified with ImageJ software (National Institutes of Health).

Nuclear extracts were prepared from NK92 cells using the CellLytic NuCLEAR extraction kit (Sigma-Aldrich). Protein concentration was measured with a Bio-Rad protein assay, and samples were stored at −70°C until use. The following DNA oligonucleotide probes were synthesized: UNC13D intron 1 WT/Mut ETS probes AGCCTCACTTC(C/T)GCTGCTGGGTGAGG; UNC13D intron 1 5′ STAT site probe GTGGGGTGAGCTGTTCGGGGAGCAGCAGAC; UNC13D intron 1 3′ STAT site probe GAGCTGTTTCCTCTGGGGAAGGGAGCG; UNC13D intron 1 WT/Mut ETS + 5′ STAT site long probes GTGAGCTGTTCGGGGAGCAGCAGACCCCCAGGCCGACGTGGCACCCCCCTCACCTCCGCTGCCGAGCCTCACTTC(C/T)GCTGCTGGCTC. Probes were labeled with α-[³²P]deoxycytidine triphosphate (3000 Ci/mmol; PerkinElmer Life and Analytical Sciences) by fill-in using the Klenow fragment of DNA polymerase I (Invitrogen). [³²P]-labeled double-stranded oligonucleotides were purified using mini Quick Spin Oligo Columns (Roche). DNA–protein binding reactions were performed in a 10-µl mixture containing 5 µg nuclear protein and 1 µg poly(dI-dC; Sigma-Aldrich) in 4% glycerol, 1 mM MgCl₂, 0.5 mM ethylenediaminetetraacetic acid, 0.5 mM dithiothreitol, 50 mM NaCl, 10 mM Tris-HCl (pH 7.5). Nuclear extracts were incubated with 1 µl ³²P-labeled oligonucleotide probe (10,000 cpm) at room temperature for 20 min and then loaded on a 5% polyacrylamide gel (37:5:1). Electrophoresis was performed in 0.5x TBE for 2 h at 130 V, and the gel was visualized by autoradiography.

RNA was isolated from primary CD56⁺ NK cells and CD8⁺ T cells from the peripheral blood of healthy donors, and 5′ RACE was performed with the 5′/3′ RACE kit, second generation (Roche) using primers listed in Table S1.

pGL3 (Promega) luciferase reporter constructs were created by PCR using the USER cloning system (NEB) for the UNC13D intron 1 promoter region (+5 to +607). A version of the intron 1 promoter with the c.118-308C>T mutation and a full-length version of the UNC13D promoter (−1,374 to +607) were also created by USER cloning. Primer sequences are provided in Table S1. These constructs, along with a pGL3⁺ control, were tested for activity in the YT-INDY NK cell line using the dual luciferase kit (Promega). Analysis was performed on a VICTOR² 1420 multilabel counter (PerkinElmer). Cells were co-transfected with a renilla plasmid for normalization measurements. Nucleofections were performed using the Amaxa Human Monocyte Nucleofector kit (Lonza) using the Y-10 program.

Primary human peripheral blood NK cells or B cells were cross-linked with 1% formaldehyde and sheared with a Misonix sonicator. For each sample, 10⁶ cell equivalents of chromatin were immunoprecipitated with the EZ-ChIP chromatin immunoprecipitation kit (Millipore) using 1 µg each of the following ChIP-grade antibodies: mouse IgG (Millipore), RNA polymerase II (Millipore), STAT4 (Abcam), anti-histone H3 (acetyl K27; Abcam), anti-BRG1 (Abcam), and anti-histone H3 (tri-methyl K4; Abcam). Cross-links were reversed by incubation at 65°C with the addition of proteinase K overnight.

Primary human peripheral blood NK cells or B cells were cross-linked with 1% formaldehyde and sheared with a Misonix sonicator. Input and phenol chloroform-extracted DNA was isolated from 5 × 10⁶ cell equivalents of chromatin according to published methods (Simon et al., 2012), and open chromatin was assessed by qRT-PCR.

PCR was performed using Platinum Taq DNA Polymerase (Invitrogen) with 35 cycles of the following incubations: 94°C for 20 s, 60°C for 30 s, and 72°C for 1 min. Quantitative real-time PCR was performed using SYBR Green or TaqMan reagents (QIAGEN) on a 7500 real time PCR system (Applied Biosystems). All primer sequences are provided in Table S1.

Naive CD8⁺ T cells were isolated from the peripheral blood of healthy donors. 2 × 10⁶ cells were cultured overnight in the following conditions: no stimulation, 200 U/ml IL-2 (PeproTech) plus 10 ng/ml IL-12 (PeproTech), 1 ng/ml IFN-γ (R&D Systems), 1 U/ml IFN-α (PBL), or 75 µl/ml CD3/CD28 microbeads (Life Technologies).

3 × 10⁶ naive CD8⁺ T cells were transfected with 40 pmol control siRNA or siRNA targeting STAT4 (QIAGEN) using the Amaxa Human T Cell Nucleofector kit (Lonza). The U-014 Nucleofector program was used for high cell viability.

Statistical significance was determined by paired Student’s t tests using GraphPad software.

Table S1 shows custom UNC13D primers.

This work was supported by the European Research Council under the European Union’s Seventh Framework Program (FP/2007-2013)/ERC Grant Agreement n. 311335, Swedish Research Council, Swedish Foundation for Strategic Research, Swedish Cancer Foundation, Swedish Children’s Cancer Foundation, Histiocytosis Association, Jeansson’s Foundation, Åke Wiberg’s Foundation, the Karolinska Institute Research Foundation, the Frontiers in Biomedical Research Fellowship, and University of Minnesota T32 Haematology Training Grant. This project has been funded in whole or in part with Federal funds from the Frederick National Laboratory for Cancer Research, National Institutes of Health, under contract HHSN261200800001E. This research was supported in part by the Intramural Research Program of National Institutes of Health, Frederick National Laboratory, Center for Cancer Research. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government.

The authors have no conflicting financial interests.