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

Abstract

Autosomal recessive mutations in UNC13D, the gene that encodes Munc13-4, are associated with familial hemophagocytic lymphohistiocytosis type 3 (FHL3). Munc13-4 expression is obligatory for exocytosis of lytic granules, facilitating cytotoxicity by T cells and natural killer (NK) cells. The mechanisms regulating Munc13-4 expression are unknown. Here, we report that Munc13-4 is highly expressed in differentiated human NK cells and effector CD8(+) T lymphocytes. A UNC13D c.118-308C>T mutation, causative of FHL3, disrupted binding of the ETS family member ELF1 to a conserved intronic sequence. This mutation impairs UNC13D intron 1 recruitment of STAT4 and the chromatin remodeling complex component BRG1, diminishing active histone modifications at the locus. The intronic sequence acted as an overall enhancer of Munc13-4 expression in cytotoxic lymphocytes in addition to representing an alternative promoter encoding a novel Munc13-4 isoform. Mechanistically, T cell receptor engagement facilitated STAT4-dependent Munc13-4 expression in naive CD8(+) T lymphocytes. Collectively, our data demonstrates how chromatin remodeling within an evolutionarily conserved regulatory element in intron 1 of UNC13D regulates the induction of Munc13-4 expression in cytotoxic lymphocytes and suggests that an alternative Munc13-4 isoform is required for lymphocyte cytotoxicity. Thus, mutations associated with primary immunodeficiencies may cause disease by disrupting transcription factor binding.

Figure 1.

Munc13-4 expression levels correlate with lymphocyte cytotoxicity. (a) Western blot analysis of Munc13-4, Stx11, and Munc18-2 expression in eight different isolated subsets of human peripheral blood lymphocytes from a healthy peripheral blood donor. All cell types were isolated by negative magnetic bead selection except for CD56^bright and CD56^dim NK cells, which were FACS sorted. Munc13-4 was run on a separate Western blot gel from Stx11 and Munc18-2. Munc13-4 and β-actin for CD8⁺ T cell and NK cell subsets (rows 1 and 2 in columns 2 and 3) were run on the same Western blot gel as those shown for STAT4 in Figure 5 a (columns 2 and 3) and BRG1 in Figure 5 j (columns 2 and 3), accounting for the shared β-actin image. (b) Cumulative Munc13-4, Stx11, and Munc18-2 fold expression values relative to B cells, naive CD8⁺ T cells, or CD56^bright NK cells from four healthy donors in three independent experiments. All values are normalized to β-actin. Fold expression differences with P ≤ 0.05 are marked with an asterisk.

Figure 2.

A UNC13D c.118-308C>T mutation causative of FHL3 lies within an intronic regulatory element of UNC13D and abrogates binding of the transcription factor ELF1. (a) Evolutionary conservation of nucleotides (gray bars) as predicted by the AlaMut algorithm in intronic (black line) and exonic (black boxes) regions of UNC13D. The c.118-308C>T mutation is highlighted in red. (b) Freshly isolated and IL-2–stimulated NK cells from 36 healthy adults, 17 healthy children, 10 FHL2 patients, 24 FHL3 patients with bialleic UNC13D mutations, and 5 FHL3 patients with bialleic UNC13D mutations, of which at least one allele harbored an UNC13D intron 1 mutation were stimulated with P815 cells, P815 cells with anti-CD16 antibody, or K562 targets. Degranulation was measured by FACS analysis of CD107a expression on the cell surface. Results represent cumulative data from multiple independent experiments. (c) Nucleotide sequence of intron 1 of UNC13D. Predicted transcription factor binding sites are labeled, and the location of the c.118-308C>T mutation is marked in red. Putative alternative transcriptional start site (TSS) and alternative first exon (blue text) are also labeled. (d) EMSA with nuclear lysates from primary NK cells and NK92 cells showing that a protein complex containing ELF1 binds to intron 1 of UNC13D, and the c.118-308C>T mutation abrogates binding. Data are representative of three independent experiments with NK cell lysates from two healthy donors and NK92 cells. (e) Location of transcriptional starts relative to the conventional translational start site of UNC13D in primary CD56⁺ NK cells and CD8⁺ T cells as identified by 5′ RACE in four healthy individuals in four independent experiments.

Figure 3.

The UNC13D intron 1 regulatory element acts as an enhancer and lymphocyte-specific alternative promoter. Quantitative RT-PCR using primers and probes specific for the (a) conventional and (b) intron 1 UNC13D transcripts in sorted B cells, CD4⁺ T cells, CD8⁺ T cells, NK cells, and monocytes from a fresh, healthy buffy coat donor, healthy donor cells that were previously frozen, healthy donor cells that were frozen and transported, and a homozygous c.118-308C>T patient that were frozen and transported. All PCR data are normalized to 18S RNA levels, and fold expression values were calculated for each donor relative to B cells. All PCRs were performed in triplicate and averaged. All data were generated together in a single experiment with sorted cells from the homozygous donor.

Figure 4.

An intact ELF1-binding site is not necessary for active transcription from the regulatory element within intron 1 of UNC13D. (a) Luciferase assays showing the relative transcriptional activity of the full UNC13D promoter region, the intron 1 promoter region and the intron 1 promoter region with the c.118-308C>T mutation in the YT-Indy NK cell line. Cumulative results from three independent experiments are shown. (b) Western blot analysis of ELF1 expression in isolated subsets of human peripheral blood lymphocytes from a healthy donor. (c) Cumulative ELF1 fold expression values relative to B cells from three donors in two independent experiments. All values are normalized to β-actin.

Figure 5.

The ELF1-binding site within the UNC13D intron 1 regulatory element is important for the recruitment of STAT4, the chromatin remodeling complex BRG1 and high levels of H3K27ac. (a) Western blot analysis of STAT4 expression in isolated subsets of human peripheral blood lymphocytes from a healthy donor. STAT4 and β-actin for CD8⁺ T cell and NK cell subsets (columns 2 and 3) were run on the same Western blot gel as those shown for Munc13-4 in Fig. 1 a (rows 1 and 2 in columns 2 and 3), accounting for the shared β-actin image. (b) Cumulative STAT4 fold expression values relative to B cells, naive CD8⁺ T cells or CD56^bright NK cells from four donors in three independent experiments. All values are normalized to β-actin. Fold expression differences with P ≤ 0.05 are marked with a single asterisk, and with P ≤ 0.01 are marked with a double asterisk. (c) EMSA with probes for the UNC13D 5′ and 3′ STAT sites with nuclear lysates from NK92 cells. Data are representative of three independent experiments. (d) EMSA with longer probes spanning the wild-type or mutant ELF1 and 5′ or 3′ STAT sites within intron 1 of UNC13D with lysates from primary NK cells (left). EMSA with longer probes spanning the wild-type or mutant ELF1 and 5′ STAT sites with and without cold consensus ELF1- and STAT4-binding sequences or anti-ELF1 and anti-STAT4 antibodies with lysates from NK92 cells (right). All three sections in this figure are from one gel. Data are representative of two independent experiments. (e) ChIP analysis of STAT4 and RNA Pol II binding to intron 1 of UNC13D. Cumulative quantitative RT-PCR fold enrichment values relative to input for three separate experiments are shown. (f) Control PCRs showing the specificity of primers used to distinguish between the wild-type and mutant intron 1 regions of UNC13D in an individual heterozygous for the c.118-308C>T intron 1 mutation. ChIP analysis of allele-specific (g) STAT4 and (h) BRG1 binding, as well as (i) H3K27ac, in primary NK cells from an individual heterozygous for the c.118-308C>T intron 1 mutation. All data were generated in a single experiment with the heterozygous donor. (j) Western blot analysis of BRG1 expression in isolated subsets of human peripheral blood lymphocytes from a healthy donor. BRG1 and β-actin for CD8⁺ T cell and NK cell subsets (columns 2 and 3) were run on the same Western blot gel as those shown for Munc13-4 in Fig. 1 a (rows 1 and 2 in columns 2 and 3), accounting for the shared β-actin image. (k) Cumulative BRG1 fold expression values relative to B cells, naive CD8⁺ T cells, or CD56^bright NK cells from four healthy donors in three independent experiments. All values are normalized to β-actin. Fold expression differences with P ≤ 0.01 are marked with a double asterisk.

Figure 6.

High-resolution comparison of the UNC13D chromatin state in lymphocyte subsets. (a) Locations of primer pairs that were used for high-resolution analysis of UNC13D. The primers map a region spanning from –2,433 to +12,689 (relative to the conventional translational start site), and each primer pair was designed to amplify a 150–200-bp product. The location of the ELF1-binding site is indicated with an open circle. The alternative first exon is labeled in blue. (b) FAIRE analysis of open chromatin in primary B, NK, CD8⁺ naïve, and CD8⁺ effector T cells. Quantitative RT-PCR values averaged from two donors in two independent experiments are shown. (c–f) ChIP assays with antibodies against STAT4, BRG1, H3K4me3, and H3K27ac. Expression values normalized to an IgG-negative control and to input are shown. Quantitative RT-PCR values averaged from two donors in two independent experiments are shown.

Figure 7.

CD3–CD28 signaling induces Munc13-4, STAT4, and BRG1 expression in naive CD8⁺ T cells. (a) Representative Western blots for Munc13-4, STAT4, pSTAT4, and BRG1 in naive CD8⁺ T cells from a healthy donor stimulated with IL-2 and IL-12, IFN-γ, type I IFN, or anti-CD3/CD28 microbeads. (b) Cumulative Munc13-4, STAT4, pSTAT4, and BRG1 fold expression values from each stimulation condition relative to unstimulated cells from five healthy donors in three independent experiments. All values are normalized to β-actin. (c) Representative Western blots for Munc13-4, STAT4 and BRG1 in naive CD8⁺ T cells from a healthy individual stimulated with anti-CD3/CD28 microbeads for 0, 2, 4, 8, and 16 h. (d) Cumulative Munc13-4, STAT4, and BRG1 fold expression values from each stimulation time point relative to unstimulated cells from three healthy individuals in two independent experiments. All values are normalized to β-actin. (e) Representative Western blots for total and phosphorylated (pY693) STAT4 expression in naive CD8⁺ T cells from a healthy donor at the indicated time points after stimulation with anti-CD3/CD28 microbeads. (f) Cumulative STAT4 and pSTAT4 fold expression values from each stimulation time point relative to unstimulated cells from three healthy donors in two independent experiments. All values are normalized to β-actin. Fold expression differences with P ≤ 0.05 are marked with a single asterisk, and with P ≤ 0.01 are marked with a double asterisk.

Figure 8.

STAT4 is necessary for Munc13-4 induction in response to CD3/CD28 stimulation. (a) Naive CD8⁺ T cells from a healthy donor were transfected with control siRNA or siRNA targeting STAT4 and stimulated overnight with anti-CD3/CD28 microbeads or left unstimulated. Western blots were performed with antibodies against Munc13-4, STAT4, and BRG1. (b) Cumulative Munc13-4, STAT4, and BRG1 fold expression values normalized to β-actin from control siRNA and siSTAT4 transfection and stimulation experiments from four healthy donors in two independent experiments. Fold expression differences between control and STAT4 knockdown conditions with P ≤ 0.05 are marked with an asterisk.