BET Bromodomain Inhibition Suppresses Human T Cell Function

Bromodomains are conserved structural motifs within diverse proteins that recognize and bind to acetylated lysine residues. The BET (bromodomain and extraterminal domain) family of adaptor proteins is characterized by two tandem N-terminal bromodomains (BD1 and BD2) and consists of four members: BRD2, BRD3, BRD4, and BRDT. Acetylated lysine residues found on histone tails, often histones 3 and 4, anchor BET proteins via BD1 and BD2 to discrete regions within the genome, where they can affect local gene expression. Broadly speaking, BET proteins act as epigenetic “readers” and regulate gene expression through a wide range of activities, including the control of RNA polymerase II (Pol II) transcriptional activity as well as through recruitment of the positive transcription elongation factor (P-TEFb) (reviewed in Ref. 1). BET proteins have been described as both transcriptional repressors and activators in many diverse biological processes, including proliferation and mitosis, neuronal function and memory, metabolism, cellular differentiation, and inflammation (reviewed in Refs. 2, 3).

Mitsubishi Tanabe Pharma first described the activity of thienotriazolodiazepine compounds to inhibit the binding of acetylated lysine residues to bromodomain-containing proteins in 2008 (patent PCT/JP2008/073864). This was the basis for discovery of Y-803 (subsequently called OTX015, then MK-8628) as well as the closely related compound, JQ1 (4). As exemplified by JQ1, these compounds exhibit high specificity for the BET family of bromodomains without detectable binding to bromodomains outside of the BET family (4, 5). Small molecule inhibitors of BET proteins have been widely used to demonstrate potent anti-inflammatory activity (6–8) and broad preclinical antitumor activity ranging from hematological models to nuclear protein in testis (NUT) midline carcinoma (NMC) samples carrying the NUT-BRD translocation (reviewed in Ref. 9). These data have led to development of several different BET inhibitors, many of which are or have been under evaluation in early stage clinical trials (reviewed in Ref. 1). In first-in-human clinical trials in oncology, the BET inhibitor MK-8628 has shown signs of activity in two hematologic studies (10, 11) as well as in a phase 1b clinical trial in patients with advanced solid malignancies in which partial responses were observed in cohorts of NMC patients (reviewed in Ref. 9). Mathematical modeling has been employed to predict the optimal dosing regimen for BET inhibitors in combination with immune therapies (12).

More recently, supportive evidence for various roles of BET inhibitors in antitumor immunity have been described. BET inhibitors were identified in a phenotypic screen for compounds that prolong functional properties of stem cell–like memory CD8⁺ T cells, suggesting that this epigenetic disruption in vitro could lead to enhanced T-effector cell function in adoptive therapy in vivo (13). JQ1-treated chimeric Ag receptor T cells displayed greater persistence and enhanced antitumor effects of chimeric Ag receptor-transduced T cells in a murine leukemia model (13). A different line of investigation demonstrated that BET inhibition in vivo showed maximal therapeutic effects in immunocompetent mice compared with immunocompromised mice (14). Among other things, the authors suggested that BET inhibition resulted in diminished constitutive and IFN-γ–induced PD-L1 expression (14, 15). Despite the recent literature describing activity and potential of BET inhibitors to promote T cell–driven antitumor immune responses, it remains largely unknown how pharmacologic disruption of bromodomain activity controls T cell functional responses in vitro and in vivo.

In this study, we investigate the immune modulatory properties of the BET inhibitor MK-8628 using various functional approaches in human T cells. We uncover intrinsic antiproliferative activity of MK-8628 in CD4⁺ and CD8⁺ T lymphocytes pursuant to an anti-inflammatory phenotype characterized by a reduction in proinflammatory cytokine production and impaired T lymphocyte activation.

Total RNA was isolated using the MagMAX mirVana Total RNA isolation kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. DNase-treated total RNA was reverse-transcribed using QuantiTect Reverse Transcription (Qiagen) according to manufacturer’s instructions. Primer assays (20×) were obtained commercially from Thermo Fisher Scientific. Gene specific preamplification was done on 10 ng cDNA per Fluidigm Biomark manufacturer’s instructions (Fluidigm). Real-time quantitative PCR was then done on the Fluidigm Biomark using 20× TaqMan primer assays (Thermo Fisher Scientific) with TaqMan Fast Universal PCR Master Mix with no AmpErase UNG‎. Samples and primers were run on the 96.96 Dynamic Array per manufacturer’s instructions (Fluidigm). Ubiquitin levels were measured in a separate reaction and used to normalize the data by the δ cycle threshold (Ct) method. Using the mean cycle threshold value for ubiquitin and the gene of interest for each sample, the equation 1.8^{Ct ubiquitin − Ct gene of interest} × 10,000 was used to obtain the normalized values. Normalized values were averaged across all experiments and Log2 transformed prior to statistical testing.

Annexin V and the following Abs were used: anti-CD4 (RPTA-4), anti-CD8 (RPTA-8), anti-Ki-67 (20raj1), anti–cyclin-B1 (GNS-1), anti–cyclin-D3 (G107-565), anti-CD71 (CYIG4), anti-CD25 (M-A251), anti–IL-2 (MQI-17HR), anti–IFN-γ (4S.B3), and anti–TNF-α (Mab11). Viability assessments were conducted with the use of fixable viability stain 510 or 780 (BD Biosciences). Cell suspensions were incubated for 15 min with anti-CD16/CD32, and then surface markers were stained by incubation for 30 min with Abs in cell staining buffer (BioLegend). Staining of cyclin B1 and cyclin D3 was performed using the FOXP3 Fix/Perm Buffer Set (BioLegend) following the manufacturer’s instructions. For cytokine detection, cell suspensions were preincubated with Dynabeads Human T-Activator CD3/CD28 (Invitrogen) for 24 h in complete RPMI 1640 media (cRPMI) and treated with protein transport inhibitor mixture (Invitrogen) for 4 h before blockade of CD16/CD32 followed by cell surface staining, permeabilization, and intracellular staining of IL-2, IFN-γ, and TNF-α. For analysis of 2-[N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino]-2-deoxyglucose (2-NBDG), cell suspensions were preincubated with Dynabeads Human T-Activator CD3/CD28 (Invitrogen) for 24 h in complete medium followed by incubation with 2-NBDG (100 μM) for 2 h at 37°C. After incubation, cells were washed followed by blockade of CD16/CD32 and cell surface staining. For apoptosis studies, cells were stimulated with Dynabeads Human T-Activator CD3/CD28 (Invitrogen) for 96 h, and FITC Annexin V Apoptosis Detection Kit I (BD Biosciences) was used following the manufacturer’s instructions. For proliferation studies, T lymphocytes were loaded with CFSE using the CellTrace CFSE Proliferation Kit (Invitrogen) and stimulated with plate-bound anti-CD3/CD28 mAbs (Invitrogen) at 1 μg/ml in cRPMI for 96 h followed by blockade of CD16/CD32 and cell surface staining. For Ki-67 studies cells were stimulated with plate bound anti-CD3/CD28 mAbs (Invitrogen) at 1 μg/ml in cRPMI for 96 h, followed by blockade of CD16/CD32, cell surface staining, and intracellular Ki-67 staining using the FOXP3 Fix/Perm Buffer Set (BioLegend) following the manufacturer’s instructions. All acquisitions for flow cytometry were performed on an LSR Fortessa cytometer with DIVA software (BD Biosystems) and were analyzed using FCS Express Software (De Novo Software). In all flow cytometry studies, gates were drawn on T lymphocytes by forward and side scatter comparison (side scatter area versus forward scatter area), followed by doublet exclusion (forward scatter height versus forward scatter area) and dead cell exclusion using fixable viability dyes (FVS510 or FVS780). Fluorescence minus one, unstained, and isotype controls were included for assessment of surface and intracellular proteins. Gates were then drawn on CD4 and CD8 populations, and expression of Ki-67, cyclin-D3, cyclin-B1, VAD-FMK, annexin V, propidium iodide (PI), CD25, CD69, CD71, CD154, IL-2, TNF-α, IFN-γ, granzyme B, and perforin was assessed by drawing the respective gates based on fluorescence minus one and unstained controls.

T lymphocytes were isolated using a CD4⁺ and CD8⁺ T Cell Isolation Kit (Miltenyi Biotec) following the manufacturer’s instructions. Cells were then stimulated with plate-bound anti-CD3/CD28 mAbs (Invitrogen) at a concentration of 1 μg/ml in cRPMI for 72 h in the presence of various doses of MK-8628 at 37°C. Ten microcoulomb of thymidine (Methyl-³H) (PerkinElmer) were added to cell cultures and plates were further incubated at 37°C for an additional 12 h. Plates were harvested onto a 96-well UniFilter-96 GF/B plate (PerkinElmer) followed by addition of 50 μl of ultima gold liquid scintillation mixture (PerkinElmer) and incubation at room temperature (RT) for 1 h. Plates were read using a Microbeta² microplate counter (PerkinElmer).

Naive (CD3⁺CD45RA⁺CCR7⁺) and memory (CD3⁺CD45RO⁺ CD45RA⁻CCR7⁻) T cell subsets were sorted using a BD FACS Aria II instrument, whereas bulk T lymphocytes were isolated using a CD4⁺ and CD8⁺ T cell isolation kit (Miltenyi Biotec). Cells were treated with various doses of MK-8628 and stimulated with Dynabeads Human T-Activator CD3/CD28 (Invitrogen) for 24 h in cRPMI, and protein transport inhibitor mixture (Invitrogen) was added for the last 4 h of stimulation. For secreted cytokines in the supernatant, V-PLEX Proinflammatory Panel 1 Human kit (Meso Scale Discovery), which measures IFN-γ, IL-1β, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12p70, IL-13, and TNF-α, was used. MSD plates were analyzed on the MS2400 imager (Meso Scale Discovery). All assays were performed according to the manufacturer’s instructions.

Human T lymphocytes were isolated using a CD4⁺ and CD8⁺ T cell isolation kit (Miltenyi Biotec) following the manufacturer’s instructions. T lymphocytes were then stimulated with Dynabeads Human T-Activator CD3/CD28 (Invitrogen) for 48 h in cRPMI and various concentrations of MK-8628. Following treatment, cells were washed and incubated with 20 μM 5-ethynyl-2′-deoxyuridine (EdU) for 2.5 h followed by staining, according to manufacturer’s protocol. Briefly, cells were washed with 3 ml of 1% BSA PBS solution and resuspended in 100 μl of Click-iT fixative, followed by incubation at RT for 15 min in the dark. Cells were then washed with 3 ml of a 1% BSA PBS solution, pelleted, and resuspended in 100 μl of Click-iT saponin-based permeabilization and wash reagent for 15 min at RT, followed by addition of 500 μl of Click-iT reaction mixture. Cells were then incubated at RT for 30 min in the dark, followed by a wash using 3 ml of 1× Click-iT saponin-based reagent. Cell pellets were then resuspended in 500 μl of FxCycle PI/RNase solution and incubated at RT for 30 min in the dark prior to filtration through polystyrene tubes with strainer caps and flow cytometry analysis.

Data were analyzed with the Student t test and one-way ANOVA with posttest analysis. Differences in means were considered significant at a p value <0.05.

MK-8628 (birabresib, OTX015, Y-803) is a triazole inhibitor of BRD2, BRD3, BRD4, and BRDT (Supplemental Fig. 1A, 1B). Consistent with the highly selective profile of its structurally related derivative JQ1, MK-8628 binds to BET family bromodomains but not to other bromodomain-containing proteins (4, 5). This compound demonstrates sub-micromolar potency for disrupting the binding between BD1/BD2 and acetylated lysine as well as cellular toxicity effects (Supplemental Fig. 1C). To elucidate the impact of MK-8628 on transcriptional profiles of T lymphocytes, CD4⁺ and CD8⁺ T cells were isolated from healthy human volunteers and stimulated with anti-CD3/CD28 mAbs in the presence or absence of MK-8628. A clinically relevant dose corresponding to MK-8628 plasma trough concentrations previously reported in two phase 1 clinical trials was selected for these studies (10, 11). Hierarchal clustering analysis of genes whose expression was significantly altered by MK-8628 revealed a subgroup of genes that were either upregulated (e.g., Ccr7, Cd45ra, Casp7, Casp9) or downregulated (e.g., Lag3, Glut1, Ifng, Grzb, Hif1) following T cell activation in an MK-8628–dependent manner (Fig. 1A–C). The latter included genes critically involved in T cell proliferation and cell cycle progression such as Mki67, Cdk2, Cdk4, Cdk6, Ccnd3, Ccne1, and Ccne2 (Fig. 1A–C). A decrease in genes directing the late G1/S phase cell cycle transition, including the Cdk-activating tyrosine phosphatase (Cdc25a) and E2F transcription factor (E2f1), was observed. In addition to this antiproliferative signature, we observed downregulation of genes associated with canonical T cell activation (e.g., Il7r, Tnfsf5, Icos, Ctla4, and Tfrc) and TCR signaling components including Lck, Mapk3, Lat, Cd28, Cd3e, Cd3d, and Cd3g (Fig. 1A–C).

MK-8628 treatment reduced expression of proinflammatory cytokine genes, including Il6, IL1b, and Csf2, and anti-inflammatory cytokine genes such as Il10, Il4, and Il13 (Fig. 1A–C) in activated T cells. In addition, MK-8628 treatment resulted in significant reduction of Il17a and Il23a transcripts in activated CD4⁺ but not CD8⁺ T cells treated with MK-8628 (Fig. 1A–C). These findings aligned with previous reports demonstrating attenuation of inflammatory gene expression following BET inhibition (6) and a requisite role for BET proteins in T_H17 cell differentiation and inflammation (7, 16, 17). Previous studies in ALL and AML leukemic cell lines have found that mRNA expression levels of BRD2, BRD3, and BRD4 did not correlate with MK-8628–induced loss of cell viability (6). Consistent with its restricted pattern of expression, we did not detect transcripts of the testis associated gene Brdt in human T lymphocytes (Fig. 1A–C). However, MK-8628 increased expression of Brd2, Brd3, and Brd4 transcripts in activated CD4⁺ T cells, although this increase was not statistically significant in the CD8⁺ T cell compartment (Fig. 1A–C). Taken together, these results reveal profound effects of MK-8628 on transcripts whose protein products play a critical role in TCR signaling pathways essential for the activation, proliferation, and effector molecule production of human T lymphocytes.

To functionally validate transcriptional alterations indicative of antiproliferative activity by MK-8628 (Fig. 2A), we first evaluated Ki-67 protein expression in MK-8628–treated T lymphocytes stimulated with anti-CD3/CD28 mAbs. T lymphocyte viability was assessed in these cultures by multiparameter flow cytometry using a fixable viability probe (FVS510). Treatment with MK-8628 resulted in a modest but statistically significant dose-dependent decrease in the frequency of viable (FVS510^negative) T lymphocytes (Supplemental Fig. 1D, 1E). Assessment of Ki-67 expression in FVS510^negative T lymphocytes revealed a striking dose-dependent reduction in the frequency of viable Ki-67^positive cells, demonstrating that this antiproliferative phenotype did not result from MK-8628 toxicity (Fig. 2B–F). In thymidine incorporation (Fig. 2G, 2H) and cell trace (CFSE) proliferation assays (Supplemental Fig. 1F, 1G), treatment with MK-8628 similarly resulted in robust inhibition of CD4⁺ and CD8⁺ T cell proliferation at clinically relevant doses of MK-8628. In additional proliferation assays using lethally irradiated HLA-DR⁺ cells as APC, stimulation of CD4⁺ and CD8⁺ T with an anti-CD3 mAb (OKT3) impaired T cell proliferative responses in the presence of MK-8628, but not in MK-8628 untreated controls, or MK-8628 pretreated APCs, further suggesting these effects were T cell intrinsic (data not shown).

Consistent with a decrease in cell cycle progression–related transcripts (Fig. 2A), expression of cell cyclin B1 and D3 was significantly decreased by MK-8628 in CD4⁺ and CD8⁺ T lymphocytes following activation, as assessed by intracellular flow cytometry analysis (Fig. 3A–H). To obtain further mechanistic insight into these observations we conducted DNA replication analysis studies using the BrdU fluorescent analogue EdU–AF488 (Click-iT Plus EdU AF488) and FxCycle Far Red dye. Flow cytometry analysis revealed an impairment in the G1 to S phase cell cycle transition with a concomitant decrease in T cells undergoing mitotic division observed at higher doses of MK-8628 (Fig. 3I–L). These observations are consistent with reduced expression of cyclin D3 (Fig. 3E–H), which is critically involved in regulating passage through the G1 phase (18). Collectively, these studies demonstrate that MK-8628 induces T cell–intrinsic antiproliferative activity in human T lymphocytes, as evidenced by inhibition of cell cyclin expression and induction of G1 cell cycle arrest.

Consistent with a reduction in T cell viability, MK-8628 treatment resulted in the upregulation of a subgroup of genes (e.g., Casp7, Casp9) that play a central role in the execution phase of apoptosis (Fig. 1A–C). To determine whether MK-8628 augments T lymphocyte apoptosis, annexin V expression was assessed in tandem with a PI probe following TCR activation in the presence or absence of MK-8628. Treatment with MK-8628 resulted in a significant dose-dependent increase in the frequency of early apoptotic annexin V⁺/PI⁻ as well as late stage apoptotic annexin V⁺/PI⁺ CD4⁺ and CD8⁺ T cells (Fig. 4A, 4B). Caspase activity in T lymphocytes was further assessed using a fluorescent analogue of the pan-caspase inhibitor Z-VAD-FMK. Consistent with augmentation of annexin V expression and increased levels of caspase-related transcripts, treatment with MK-8628 lead to a significant dose-dependent increase in the frequency of FITC-VAD-FMK⁺ CD4⁺ and CD8⁺ T cells (Fig. 4C, 4D), indicative of increased caspase activity. To further validate these observations, we examined mitochondrial membrane potential changes associated with apoptosis driven by the mitochondrial death pathway. There was a significant reduction of mitochondrial membrane potential in MK-8628–treated CD4⁺ and CD8⁺ T cells as assessed by the membrane potential sensitive cyanine dye DiOC₂ (data not shown). Collectively, these results demonstrate that MK-8628 exacerbates T cell apoptosis.

To validate transcriptional signatures (Fig. 5A) indicative of attenuated TCR signaling by MK-8628, we next assessed activation marker expression and cytokine production in bulk CD4⁺ and CD8⁺ T lymphocytes. Expression of canonical activation markers such as CD25, CD71, CD69, and CD154 were significantly decreased in MK-8628–treated CD4⁺ and CD8⁺ T lymphocytes (Supplemental Fig. 2A–D and data not shown). Levels of effector cytokines such as IFN-γ, TNF-α, and IL-2 were significantly decreased in the supernatants of CD4⁺ and CD8⁺ T cells treated with MK-8628 (Fig. 5B, 5C). Consistent with this observation, there was a significant reduction in the frequency of CD4⁺ and CD8⁺ T cells producing IFN-γ, TNF-α, and IL-2 as assessed by intracellular flow cytometry (Fig. 5D, 5E). Decreased expression of effector molecules that play a critical role in T lymphocyte–mediated target cell killing, such as granzyme B and perforin (data not shown), were similarly observed in MK-8628–treated CD4⁺ and CD8⁺ T cells. Importantly these results were not an artifact of MK-8628 cytotoxicity, as the viability of TCR-stimulated T lymphocytes treated with MK-8628 did not differ from that of untreated controls (data not shown). Furthermore, these results are consistent with MK-8628–driven transcriptional repression of TCR signaling components (Fig. 5A) and their downstream gene products (Fig. 1A–C).

To determine whether MK-8628 might have differential effects within the CD4⁺ and CD8⁺ T cell compartment, we sorted CD3⁺CD45RO⁺CD45RA⁻CCR7⁻ memory and CD3⁺CD45RA⁺CCR7⁺ naive T cell subsets (purity >98%, data not shown) and stimulated them in the presence or absence of MK-8628. In line with its inhibitory effects in bulk CD4⁺ and CD8⁺ T lymphocytes (Fig. 5, Supplemental Fig. 2A–D), MK-8628 significantly reduced the frequency of naive and memory T cells expressing the activation markers CD25 and CD71 and significantly reduced IFN-γ, TNF-α, and IL-2 levels in supernatants of MK-8628–treated cells (Supplemental Fig. 3A–D). There was a parallel decrease in the frequency of cells expressing IFN-γ, TNF-α, and IL-2 (Supplemental Fig. 4A–F), consistent with observations in bulk CD4⁺ and CD8⁺ T lymphocytes, although we did not observe a significant change in IFN-γ intracellular expression in memory CD8⁺ T cells treated with MK-8628. Collectively, these results show that MK-8628 globally suppresses human T cell activation and effector cytokine production.

The capacity of naive T lymphocytes to undergo effector differentiation entails changes in their bioenergetics pathways resulting in the preferential use of aerobic glycolysis as well as augmentation of oxidative phosphorylation (19). To examine potential metabolic perturbations in naive T cells treated with MK-8628, we measured glucose uptake in naive T cells following activation using fluorescent glucose (2-NBDG). Naive T cells activated in the presence of MK-8628 had a reduced competency for glucose uptake, as the frequency of 2-NBDG⁺ cells was significantly lower in MK-8628–treated cells compared with untreated controls (Supplemental Fig. 4G–L). Consistent with this observation, a reduced expression of Glut1 in naive T cells activated in the presence of MK-8628 was observed when compared with untreated controls (data not shown). In line with these observations, transcripts requisite for induction of a glycolytic switch (19, 20) including Glut1 and Hif1a, were significantly decreased in MK-8628–treated CD4⁺ and CD8⁺ T lymphocytes (Supplemental Fig. 4L, 4M). Taken together, these results illustrate that MK-8628 inhibits glycolysis in the naive T cell compartment.

In this study, we demonstrate negative regulation of human T cell function by the BET bromodomain inhibitor MK-8628. Exposure of TCR-activated lymphocytes to MK-8628 reduces expression of genes critically involved in cell cycle progression, including key cell cyclins and cyclin dependent kinases. Assessment of transcripts essential for proliferation revealed defects in Mki67 and E2f1 expression, the latter playing a critical role in the activation of promoters of cell cycle regulatory genes such as cyclin A and cyclin E (21, 22). Flow cytometry studies demonstrated defects in intracellular Ki-67 protein expression coupled to reduced expression levels of cyclin B1 and cyclin D3 proteins in MK-8628–treated T lymphocytes. These results are consistent with MK-8628–mediated transcriptional repression of Cdk4 and Cdk6 genes, whose protein products form complexes with cyclin D3 and are necessary for the G1/S transition (23, 24). Furthermore, BRD2 forms complexes with RNA Pol II and can escort E2F transcription factors to regulatory sequences, thereby coupling histone acetylation to transcription (25–27). BRD4 also directs the assembly of chromatin complexes, where it plays a critical role in transcriptional initiation and elongation, the latter driven by its capacity to interact with P-TEFb (28, 29). Furthermore, BRD4 acts as a mitotic bookmark and is indispensable for postmitotic transcription, including G1 gene expression (30–32). In line with these observations, we demonstrate in this study for the first time, to our knowledge, precipitation of G1 cell cycle arrest in TCR-activated lymphocytes exposed to a small molecule inhibitor targeting BET proteins. This antiproliferative phenotype is consistent with observations of cell cycle arrest in several models of hematologic and solid malignancies, including human MLL–fusion leukemic cell lines as well as BRD4/NUT-dependent cell lines and patient-derived xenograft models (4, 33–35). Furthermore, we uncover robust inhibition of T cell activation and effector cytokine production in the naive and memory T cell compartments following treatment with MK-8628. Collectively, our results emphasize how pharmacological BET inhibition represses inflammatory cytokine production and impairs polyclonal T cell proliferation.

For the studies presented in this article, the functional assessment of MK-8628 was evaluated predominantly in in vitro primary T cell assays. Although the evaluation of in vivo and in vitro correlations of MK-8628 treatment on T cells is beyond the scope of this work, these studies align and expand upon the reported roles of BET inhibitors on T cell biology. Our observations of transcriptional repression and attenuated production of inflammatory cytokines are consistent with several early studies demonstrating anti-inflammatory functions of small molecule inhibitors of BET proteins (6, 7). For example, the first reports establishing an anti-inflammatory role for BET inhibitors demonstrated selective repression of proinflammatory gene expression in LPS-treated bone marrow–derived macrophages (6, 7). In the former study, the BET inhibitor I-BET attenuated LPS-induced endotoxic shock and conferred protection against cecal ligation and puncture–induced septic lethality. Similarly, Belkina and colleagues demonstrated that JQ1 suppressed key LPS-inducible genes in bone marrow–derived dendritic cells (DCs) and blunted the cytokine storm in endotoxemic mice by reducing levels of the proinflammatory cytokines IL-6 and TNF-α (8). Observations of an anti-inflammatory phenotype were also extended to human monocytes, where BET inhibition attenuated TLR-4 and TNF-induced IFN-β transcriptional responses (36, 37). Furthermore, a study by Bandukwala and colleagues (7) demonstrated that the BET inhibitor I-BET-762 limited the capacity of Ag-specific T cells differentiated under T_H1 but not T_H17 conditions in vitro to induce pathological T cell responses in an adoptive transfer model of experimental autoimmune encephalomyelitis. In this study, I-BET-762 attenuated production of GM-CSF by CNS-infiltrating T cells, leading to decreased recruitment of CD11b⁺ macrophages in the CNS. Interestingly, our studies also uncovered a striking decrease in mRNA expression levels of GM-SCF in T lymphocytes activated in the presence of MK-8628, further underscoring the concept of leveraging BET inhibitors in autoimmune settings where GM-CSF promotes T_H17-mediated inflammation in murine models of autoimmunity. Consistent with this hypothesis, several studies have demonstrated a requisite role for BET proteins in directing T_H17 cell differentiation as well as T_H17-mediated pathology in murine models of autoimmunity (16, 17, 38, 39). Furthermore, the BET inhibitor JQ1 was able to impair LPS-induced maturation of DCs by inhibiting GM-CSF–mediated STAT5 activation and limiting the capacity of DCs to induce allogeneic T cell proliferation and cytokine production (40). Although BET inhibition did not interfere with regulatory T (Treg) cell function and expansion in vivo and suppressed graft-versus-host disease in a preclinical model, a recent study reported downregulation of Foxp3 and CTLA-4 as a mechanism of attenuated Treg suppression in lung adenocarcinomas treated with JQ1 (41, 42). This is particularly important given the strict dependency of Treg cells on STAT5-derived signals for their homeostatic proliferation, peripheral maintenance, and suppressor function (43). In line with this reasoning, combination of JQ1 with anti-PD1 in a KRAS-mutant nonsmall-cell lung carcinoma model was sufficient to delay tumor growth and improve survival through a JQ1-mediated reduction of suppressive KLRG1⁺ Treg cells (44).

Surprisingly, MK-8628 downregulated the expression of multiple genes associated with a canonical T cell activation signature including Tfrc, Il7r, Icos, Tnfsf5, and Cd25 while reciprocally upregulating genes reflective of a naive T cell state such as Cd45ra and Ccr7. We validated these observations on the protein level in TCR-stimulated naive and memory lymphocyte subsets and observed parallel dysfunction in their capacity to produce inflammatory cytokines such as IFN-γ and TNF-α. In contrast to our studies, a recent report demonstrated that the BET inhibitor JQ1 inhibited expression of the BRD4-regulated transcription factor BATF and promoted the in vivo persistence of stem cell–like memory CD8⁺ T cells, including their capacity to produce polyfunctional cytokines such as IL-2 (13). Furthermore, in syngeneic ovarian cancer xenografts and a Myc-driven B cell lymphoma model, BET inhibition suppressed PD-L1 expression in tumor cells and promoted cytotoxic T cell activity (14, 15). Importantly, the study by Zhu and colleagues demonstrated that Cd274 is a direct target gene of BRD4, which is subjected to JQ1-mediated transcriptional repression, but did not reveal whether augmentation of cytotoxicity in T cells was a JQ1 CD8⁺ T cell–intrinsic effect (15). Notably, the capacity of BRD4 to preferentially localize to noncoding regions of DNA known as superenhancers might account for the selectivity of molecules like JQ1 to target genes with super enhancer signatures such as Myc (45). In addition, several reports have now shown activity of BET inhibitors in preclinical cancer models as a monotherapy or in combination with histone deacetylase (HDAC) and MAPK inhibitors (46–48). In contrast, promising clinical activity has been short-lived and restricted to small cohorts of patients with NMC: a disease directly driven by a BRD4 fusion oncoprotein. In other malignancies in which BET proteins do not seem to play direct roles, responses have been modest and associated with significant dose limiting toxicities (reviewed in Ref. 1). Our results therefore underscore the importance of demarcating the activity of BET inhibitors on healthy tissues and cells, including antitumor T lymphocytes.

One interesting question that remains to be elucidated is how transcriptional landscapes of T cell subsets, including distinct chromatin arrangement, may direct differential responses to BET inhibition. For example, memory CD8⁺ T cells display a preprogrammed chromatin accessibility landscape with unique sets of loci and memory-primed genes states that are absent in effector and naive CD8⁺ T cells (49, 50). It is reasonable to speculate therefore that chromatin accessibility to bromodomain-mediated remodeling may be dependent on the stage of T cell differentiation. Consistent with this notion, Mele et al. (16) demonstrated that JQ1 can abrogate T_H17, but not T_H1 or T_H2, cell differentiation, by preventing BRD2/BRD4 association at the Il17 locus. In contrast, I-BET-762 was able to limit both IFN-γ and IL-4 production under T_H1 and T_H2 polarizing conditions (6). Furthermore I-BET-762 did not impact the expression of Rorgc, although it was able to affect expression of Il17a (6). Taken together, these studies provide impetus for a better understanding of the contribution of individual BRD proteins in diverse T cell activities and phenotypes, including their distribution and regulation in the context of T cell differentiation and activation. In summary, our studies reveal a previously unappreciated T cell–intrinsic antiproliferative activity of the BET inhibitor MK-8628 and extend an anti-inflammatory role for BET inhibition on human T cell function. These studies provide a strong impetus for assessment of BET inhibitors in autoimmune T cell–driven inflammation.

We thank Dr. Peter S. Kim and Dr. Sheila Ranganath for careful reading of the manuscript. We would also like thank Dr. Andrew M. Haidle for providing the MK-8628 chemical structures.

All authors own stock or stock options in and are current employees of Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Kenilworth, NJ.