The role of myeloid cells as regulators of tumor progression that significantly impact the efficacy of cancer immunotherapies makes them an attractive target for inhibition. Here we explore the effect of a novel, potent, and selective inhibitor of serine/threonine protein kinase casein kinase 2 (CK2) on modulating myeloid cells in the tumor microenvironment. Although inhibition of CK2 caused only a modest effect on dendritic cells in tumor-bearing mice, it substantially reduced the amount of polymorphonuclear myeloid-derived suppressor cells and tumor-associated macrophages. This effect was not caused by the induction of apoptosis, but rather by a block of differentiation. Our results implicated downregulation of CCAAT-enhancer binding protein-α in this effect. Although CK2 inhibition did not directly affect tumor cells, it dramatically enhanced the antitumor activity of immune checkpoint receptor blockade using anti-CTLA-4 antibody. These results suggest a potential role of CK2 inhibitors in combination therapies against cancer.
Significance: These findings demonstrate the modulatory effects of casein kinase 2 inhibitors on myeloid cell differentiation in the tumor microenvironment, which subsequently synergize with the antitumor effects of checkpoint inhibitor CTLA4. Cancer Res; 78(19); 5644–55. ©2018 AACR.
It is now evident that the ultimate success of cancer immunotherapy depends on its ability to overcome limitations imposed by tumors. Moreover, the same is probably true for other types of oncologic therapies. Myeloid cells are critical components of the tumor microenvironment widely implicated in suppressing antitumor immunity (1–3). These cells are composed of tumor-associated macrophages (TAM), dendritic cells (DC), neutrophils, and pathologically activated immature myeloid cells with the potent ability to suppress immune responses, termed myeloid-derived suppressor cells (MDSC). In recent years, it has become clear that MDSC are not only an important negative regulator of immune responses, but also contribute to other aspects of tumor growth, namely tumor angiogenesis, tumor cell invasion, and formation of pre-metastatic niches (4). In tumor-bearing (TB) mice, the total population of MDSC consists of two large groups of cells; the most abundant (>75%) population are immature, pathologically activated neutrophils, termed polymorphonuclear MDSC (PMN-MDSC) and the less abundant (<20%) population are pathologically activated monocytes, termed monocytic MDSC (M-MDSC).
We have previously demonstrated that defective DC differentiation in cancer and the accumulation of PMN-MDSC was mediated by downregulation of Notch signaling (5). These results were later confirmed by others (6). Inhibition of Notch signaling was partially caused by phosphorylation of Notch by casein kinase 2 (CK2; ref. 5). This suggested that targeting CK2 could be potentially beneficial for improving immune responses in cancer.
CK2 is a serine/threonine protein kinase with numerous intracellular protein substrates. CK2 phosphorylates its substrates by utilizing either ATP or GTP as a phosphate donor. CK2 expression and activity are elevated in many types of cancer. Functionally, it is has been observed that activation of CK2 is associated with suppression of apoptosis of cancer cells (7–9). Conversely, downregulation of CK2 expression or inhibition of its activity facilitates cell death triggered upon drug exposure, ligation of death receptors, and ionizing radiation (8, 10). Inhibition of CK2 sensitized T-lymphoblastic cells to drug-induced apoptosis by increasing cellular uptake of the drug (11). CK2 inhibition also increased the sensitivity of rhabdomyosarcoma and colon carcinoma cells to TRAIL-induced apoptosis (12), and induced ROS-mediated apoptosis of leukemia cells (8). CK2 regulates a number of signaling pathways involved in tumor progression and also is critical for cell differentiation. Specifically, the phosphatase, PTEN, regulates cell survival by blocking the PI3K–AKT axis and has been shown to be phosphorylated by CK2 (13), which facilitates its proteasomal degradation (14). As a result of the inhibition of PTEN by CK2, the AKT survival signal is prolonged (15). In addition, CK2 can directly phosphorylate AKT at Ser129 (16). CK2 has also been implicated in direct targeting of the NF-κB and Wnt signaling pathways (16–18).
In this study we tested the effects of novel potent and selective CK2 inhibitors and found that in TB, but not in tumor-free mice, they caused a substantial decrease in PMN-MDSC and macrophage differentiation, and dramatically enhanced the effect of anti-CTLA-4 immune-based therapy.
Materials and Methods
Cord blood samples were collected at Helen F. Graham Cancer Center. All patients signed written informed consent form. The studies were conducted in accordance with recognized ethical guidelines and were approved by Institutional Review Board of the Christiana Care Health System at the Helen F. Graham Cancer Center and The Wistar Institute.
All procedures were performed accordance with the NIH Guide for the Care and Use of Laboratory Animal guidelines and approved by the Institutional Animal Care and Use Committee at The Wistar Institute. Female 5- to 6-week-old C57BL/6 mice and Balb/c mice (Charles River Labs and Envigo) were maintained in a temperature controlled room with a 12/12-hour light/dark schedule and food provided ad libitum.
BMS-211 (FW: 825.24) is a prodrug of the parent pan-CK2 inhibitor, BMS-699 (Details to be published). Suspension formulations for BMS-211 were prepared in 0.5% methocel-A4m, 0.1% TWEEN-80. Particle size of the suspension was typically ∼10 μm (D50). Lower concentration formulations were prepared by appropriate dilution (v/v) of the highest concentration formulation. BMS-595 (FW: 532.44) is also pan-CK2 inhibitor and was solved with vehicle (Kolliphor TPGS:EtOH:PEG300 = 1:1:8) for in vivo experiments.
Lewis lung carcinoma (LLC), CT26 colon carcinoma, 4T1 breast carcinoma, and EL4 lymphoma were obtained from ATCC and cultured in DMEM (Corning Incorporated) supplemented with 10% FBS (Atlanta Biologicals, Inc.) and 1% antibiotics (Thermo Fisher Scientific Inc.). Cells were incubated in a 37°C and 5% CO2. 70% to 80% confluent cells were harvested using 0.25% Trypsin (Thermo Fisher Scientific Inc.) and passaged or used for experiments. Cells were tested on mycoplasma contamination every 3 months.
Balb/c mice were subcutaneously implanted with 1 × 106 4T1 or CT26 cells in 0.2 mL PBS on day 0 and C57BL/6 mice with 1 × 106 MC38 or 5 × 105 LLC cells in 0.2 mL PBS on day 0. Six or seven days post-implantation, TB mice were randomly divided into four groups (n = 7–10). Mice were treated with BMS-211 or BMS-595 solution orally daily for 21 days at 20 mg/kg or 60, respectively. In the 4T1 model, anti-CTLA-4 (clone 9D9) mIgG2a antibody or the isotype control was intraperitoneally injected into the mice at 20 μg/mouse on days 6, 13, and 20. In MC38 model, anti-CTLA-4 mIgG2a antibody or the isotype control was injected intraperitoneally into the mice at 200 μg/mouse on days 6, 13, and 20. In CT26 model, anti-CTLA-4 mIgG2a antibody was intraperitoneally injected at 5 μg/mouse on days 7, 14, and 21, and anti-CTLA-4 mIgG2b antibody was intraperitoneally injected at 200 μg/mouse on days 7, 14, and 21. In LLC model, anti-CTLA-4 mIgG2a antibody was intraperitoneally injected at 200 μg/mouse on day 7 and 14. For the CD8 T cell depletion study, LLC tumor bearing mice were intraperitoneally injected with 100 μg of anti-CD8 alpha antibody (BioXcell, clone 53–6.7) on day −3, 0, 3, 7, 10, and 14.
In vitro cytotoxic activity
The tumor cell lines were plated at 2 × 103 cells/well into 96-well plates and cultured 16 hours. BMS-595 was dissolved with DMSO and diluted with media, and added into the wells and incubated for 3 days. The cell viability was measured using MTS assay (Promega). IC50 values were calculated using linear regression analysis.
Single-cell suspensions were prepared from spleens and bone marrow. Tumor tissues were cut into small pieces, and digested with the Mouse Tumor Dissociation Kit (Miltenyi Biotec). Red blood cells in the cell suspensions were lysed using ammonium chloride lysis buffer. The list of antibodies is provided in Supplemental Table S1. Flow cytometry data were acquired using a BD LSR II flow cytometer and analyzed using FlowJo software (Tree Star).
Mouse hematopoietic progenitor cells culture
Lineage negative cells were purified from C57BL/6 naïve bone marrow using the Lineage Cell Depletion Kit (Miltenyi). Lineage negative cells were cultured in RPMI (Corning Incorporated) supplemented with 10% FBS, 1% antibiotics, and 50 μmol/L 2-mercaptoethanol (Thermo Fisher Scientific Inc.) with 20 ng/mL of recombinant GM-CSF (Invitrogen), at 50,000 cells/well in 24-well plates. Tumor explant supernatants (TES) were obtained by culturing small pieces of EL4 tumors with complete RPMI media for 24 hours. On day 1, TES or media with DMSO or BMS-595 solution was added into the wells at 10% to get the final concentration of 10 or 100 nmol/L BMS-595. On day 3, half of the culture supernatant was exchanged with fresh media supplemented with 20 ng/mL of GM-CSF, with or without 10% TES. On day 6, the cells were collected and analyzed by flow cytometry. To assess proliferation, hematopoietic progenitor cells (HPC) were cultured with 10 μmol/L of BrdUrd for 24 hours, and then analyzed by flow cytometry using the BrdUrd Detection Kit (BD Pharmingen).
Human HPCs culture
CD34+ HPCs were purified from human cord blood by Ficoll-based density gradient centrifugation followed by CD34 magnetic beads selection (Miltenyi Biotec). OP9 mouse bone marrow stromal cells were cultured in the six-well plate at 104 cells/well the day before CD34+ purification. On day 0, CD34+ cells were cocultured with OP9 cells at 105 cells/well in Iscove's modified Dulbecco's medium supplemented with 20% FBS, 1% antibiotics, 20 ng/mL human GM-CSF, and 100 ng/mL human G-CSF (PeproTech). After overnight culture, tumor-conditioned media (TCM) from RPMI8226 human myeloma cell lines was added at 10% with DMSO or 100 nmol/L BMS-595. On day 3 and 6, fresh media containing 20 ng/mL GM-CSF, 100 ng/mL G-CSF, 10% TCM, and DMSO or 100 nmol/L BMS-595 was added in the wells. On day 9, the cells were collected and analyzed by flow cytometry.
Ly6G+ cells were purified from spleen cells or tumor cells by positive selection using biotinylated Ly6G antibody and streptavidin microbeads (Miltenyi). The purity of the cell populations was >95%. CD11b+Ly6ChighLy6G− cells were isolated from spleen cells by cell sorting on a FACSAria cell sorter (BD Biosciences). CD8+ T cells from PMEL mice that recognize the gp100-derived peptide were used as responders. Splenocytes from PMEL mice were mixed with splenocytes from naïve mice at 1:4 ratio in complete RPMI media and plated into 96-well U-bottom plates at 105 cells/well. Ly6G+ or CD11b+Ly6ChighLy6G− cells were added to the wells at 1:16–1:1 ratios. Murine gp100 peptide (25–33) EGSRNQDWL (AnaSpec, Inc.) was added into the wells at the final concentration of 0.1 μg/mL. After 48 hours, cells were pulsed with 3H-thymidine (1 μCi/well; GE healthcare) for 16 hours. 3H-thymidine uptake was counted using a liquid scintillation counter as counts per minute (cpm) and calculated the percentage of proliferation to the positive control (the wells with responder cells and peptide but without suppressive cells).
Ly6G+ cells and CD11b+Ly6G−F4/80+Ly6Clow macrophages were treated with BMS-595 for 24 to 72 hours. The cells were stained with FITC-Annexin V (BD Pharmingen) in Annexin V binding buffer (BD Pharmingen) for 30 minutes at room temperature, and then suspended with DAPI containing MACS buffer and analyzed by flow cytometry. HPCs were cultured under the same condition as previously described with BMS-595 and 12.5 and 25 μmol/L of Z-VAD-FMK (Calbiochem), and then analyzed using flow cytometry.
RNA was extracted with a total RNA Extraction Kit (Omega Bio-tek). cDNA was synthesized (cDNA Reverse Transcriptase Kit; Applied Biosystems), and PCR was performed in triplicate for each sample. To detect expression of Notch-, IRF8-, CCAAT-enhancer binding protein-α (C/EBPα-), or C/EBPβ-regulated genes, qRT-PCR was performed with 10 μL SYBR Master Mixture (Applied Biosystems) and the list of the primers is provided in Supplemental Table S2. Expressions of the different genes were normalized to β-actin. Relative expression was calculated using the 2−ΔΔCt method.
Western blot analysis
Nuclear and cytoplasm extracts were obtained from HPC treated with DMSO or 100 nmol/L of BMS-595 for 24 hours. Nuclear protein was extracted using CelLytic NuCLEAR Extraction Kit (Sigma) with phosphatase inhibitor cocktail (Sigma). To retrieve the cytoplasm lysate, the cells were lysed with RIPA buffer (Sigma) supplemented with a protease inhibitor cocktail (Sigma) and a phosphatase inhibitor cocktail. Denatured protein samples were separated by electrophoresis with 25 μg/lane and transferred onto PVDF membrane. The list of antibodies provided in Supplemental Table S1. The detected proteins were visualized by the ECL system (GE Healthcare Life Sciences).
LLC-TB mice were treated with vehicle or 60 mg/kg of BMS-595 for 2 weeks. HPC was isolated from the bone marrow of the vehicle- or BMS-595-treated mice. Total RNA was extracted from the HPC using the Qiagen's Rneasy Mini Kit and RNA quality was assessed using the Agilent RNA ScreenTape Assay on Agilent's 2200 TapeStation. RNA-seq library was prepared using TruSeq Stranded Total RNA Kit (Illumina) and ran in a 75bp paired end run, 40 million reads on Illumina NextSeq instrument.
RNA-seq data was aligned using bowtie2 (19) against mm10 version of the human genome and RSEM v1.2.31 software (20) was used to estimate raw read counts and FPKM values using transcriptome information from Ensemble v84 set. DESeq2 (21) was used to estimate significance of differential expression between five replicates of untreated and BMS-595–treated samples. Gene expression changes were considered significant if passed FDR <5% threshold and had expression for all replicates from one group higher or lower than any replicates of another group. Gene set enrichment analysis of gene sets was done using QIAGEN's Ingenuity Pathway Analysis software (IPA; QIAGEN Redwood City, www.qiagen.com/ingenuity) using “Canonical Pathways,” “Diseases & Functions,” and “Upstream Analysis” options. Only significantly enriched at P < 0.05 functions and regulators with significantly predicted activation state (Z-score |Z| > 2) were considered. Pathways significantly affected at FDR <5% with predicted activation stated Z-score of at least 1 were reported. The data were submitted to the Gene Expression Omnibus (GEO) database and can be accessed using accession number: GSE117712.
Statistical analyses were performed using two-tailed Student t test or Mann–Whitney U test and GraphPad Prism 5 software (GraphPad Software Inc.). Paired t test was used since data were normally distributed. All the data are presented as mean ± SD and P value less than 0.05 was considered significant.
CK2 inhibitor has potent antitumor effects in combination with anti-CTLA4 antibody
Potent and selective, ATP-competitive CK2 inhibitors (BMS-595, BMS-699, and BMS-211) were developed by Bristol-Myers Squibb (Supplementary Fig. S1A). Inhibition of proliferation of various mouse cell lines (LLC, CT26 colon carcinoma, EL4 lymphoma) were at relatively high IC50 values between 250 nmol/L and 1 μmol/L, and the 4T1 breast adenocarcinoma cells were even less sensitive (Supplementary Fig. S1B). The treatment of immunocompetent C57BL/6 mice bearing LLC tumors (Fig. 1A) or immunodeficient NSG mice bearing 4T1 or MC38 tumors with CK2 inhibitors demonstrated modest direct antitumor effects of the compounds (Supplementary Fig. S1C and S1D). BMS-595 inhibited the in vitro proliferation of human colorectal and lung cancer cell lines with an IC50 ranging from 10 nmol/L to greater than 1 μmol/L (Supplementary Fig. S2). In sensitive cell lines, antiproliferative effects of BMS-595 and its structurally related analog strongly correlated with CK2 kinase inhibition.
However, although as a single-agent neither the CK2 inhibitor nor the immune check-point inhibitor, anti-CTLA-4-mIgG2a antibody at low dose, potently impacted antitumor activity, in combination, they showed remarkable antitumor effects in all tested tumor models (LLC, 4T1, MC38, and CT26; Fig. 1A). Complete rejection (CR) was observed in 60% to 90% of mice treated with the combination therapy (Fig. 1A). Because anti-CTLA-4-mIgG2a has the potential to deplete Treg cells in tumors, in a separate set of experiments, we tested anti-CTLA-4-mIgG2b antibody, which was reported not to deplete intratumoral Treg cells in the CT26 model (22). In 4T1 model, anti-CTLA-4-mIgG2b did not deplete Treg cells in the tumors (Fig. 1B). Combination of the CK2 inhibitor with anti-CTLA-4-mIgG2b resulted in complete tumor rejection in all 10 mice used in the experiment (Fig. 1C), suggesting that Treg depletion is not involved in the effects mediated by CK2 inhibition. Mice that completely rejected tumor in the combination treatment groups were rechallenged with the same tumor cells after 90 days. All of these mice rejected a secondary tumor challenge, indicating that immune memory was generated in the initial treatment (Fig. 1D). Anti-CD8 antibody injection of LLC TB mice completely cancelled the antitumor efficacy of CK2 inhibitor, which suggested that CD8+ T cells were major effector cells in CK2 inhibitor-induced antitumor effect (Fig. 1E). Thus, although the CK2 inhibitor had limited antitumor effects, it has potent antitumor activity in combination with CTLA-4 immune check-point blockade and this effect was mediated by an immunological mechanism.
Effect of CK2 inhibition on host cells in the tumor microenvironment
To better understand the mechanism of action of the CK2 inhibitor, we investigated the presence of myeloid cells and lymphocytes in LLC TB mice treated for 2 weeks with BMS-595. The treatment did not affect the presence of T, B, or NK cells in spleens or tumors of TB mice. It also did not affect populations of CD11b+Ly6G−Ly6Chi M-MDSC and CD11c+I-Abhi DC. However, substantial decreases in the presence of CD11b+Ly6G+Ly6Clo PMN-MDSC and CD11b+Ly6G−F4/80hi macrophages (MΦ) were observed in the spleen (Fig. 2A). In tumors, the frequency of PMN-MDSC was not significantly changed, whereas populations of TAMs and DC were decreased (Fig. 2A). CK2 inhibition caused significant, albeit a modest decrease in granulocytic cells in the bone marrow (BM) (Fig. 2A). The discrepancy between a decreased presence of granulocytic cells in the BM and spleen after CK2 inhibitor treatment and the lack of changes in the tumor site could be explained by the fact that the decrease of granulocytic cells in the periphery may not be sufficient to prevent the migration of substantial numbers of PMN-MDSC to the tumor site. To verify this possibility, we depleted MDSC from spleens and blood using agonistic DR5 (TRAIL-R) antibody previously shown to selectively deplete a substantial portion of PMN-MDSC in TB mice (23). Consistent with previous observations, treatment with DR5 antibody alone had minor antitumor activity (Fig. 2B), while causing a significant reduction in PMN-MDSC in the spleens of mice (Fig. 2C). However, treatment with DR5 antibody did not change the presence of PMN-MDSC in the tumor site (Fig. 2C).
Because the CK2 inhibitor affected populations of granulocytic cells and MΦ in TB mice, we tested its effect on these cells in naïve, tumor-free mice. Two-week treatment with BMS-595 did not change the proportion of neutrophils in the BM or the absolute number of these cells in the spleen. Similarly, no changes were seen in the presence of splenic MΦ (Fig. 2D).
We verified that the populations of granulocytic and monocytes cells in the spleen and tumor of TB mice were immune suppressive and can be indeed identified as PMN-MDSC and M-MDSC, respectively (Supplementary Fig. S3). Treatment with the CK2 inhibitor did not abrogate suppressive activity of these cells (Supplementary Fig. S3). Thus, the CK2 inhibitor caused a significant reduction in the presence of PMN-MDSC in the BM and spleens of TB mice and decreased the population of splenic MΦ and TAMs. It did not, however, affect these cells in tumor-free mice.
Mechanisms regulating the effect of CK2 inhibition on PMN-MDSC and macrophages
We then evaluated the effect of CK2 inhibition on the generation of myeloid cells. Treatment of LLC TB mice with BMS-595 did not affect common myeloid (CMP) and megakaryocyte–erythroid (MEP) progenitor cells in the BM, but slightly increased the proportion of granulocyte-macrophage progenitors (GMP) (Fig. 3A). This increase in GMP likely reflects compensatory changes resulting from the decreased presence of granulocytes and macrophages in tissues. In the naïve mice, BMS-595 treatment increased the proportion of GMP but decreased that of MEP (Fig. 3A). We assessed the differentiation of myeloid cells in vitro from enriched BM-derived HPCs with and without tumor explant supernatant (TES). BMS-595 had a modest effect on the total number of cells generated from HPC during 6-day culture with GM-CSF (Fig. 3B). However, it caused a dramatic (almost three-fold) decrease in the proportion of granulocytic cells (Fig. 3C). We evaluated the effect of BMS-595 on the myeloid cell differentiation from CD34+ progenitor cells from human cord blood. After 9 days of culture, BMS-595 decreased the proportion and the absolute number of granulocytic cells (Fig. 3D). In contrast, the proportion and the number of monocytic cells were increased by BMS-595.
Such a dramatic decrease in granulocytic cells could be the result of BMS-595 induced apoptosis, a block of proliferation of the precursors, or inhibition of differentiation. Ly6G+ cells were isolated after 3-day culture of HPC with GM-CSF and were treated, in the presence of GM-CSF, with BMS-595 for 24 hours. This time frame was selected because a 24-hour incubation of granulocytes with GM-CSF sustained viability of granulocytes above 70%. After 48 hours, the viability dropped below 40% and any effect of CK2 inhibition on apoptosis is difficult to establish. The CK2 inhibitor caused a very small decrease in the total number of cells recovered after incubation (Fig. 4A), which was associated with a modest (5%) upregulation of apoptosis (Fig. 4B). No difference in cell proliferation was observed (Fig. 4C).
To address the potential role of apoptosis in the substantial loss of granulocytic cells during HPC differentiation, experiments were repeated in the presence of the apoptosis inhibitor z-VAD. z-VAD at optimal doses (12–25 μmol/L) did not prevent inhibition of granulocytic cell differentiation by BMS-595 (Fig. 4D).
To clarify the effect of the CK2 inhibitor on differentiated PMN-MDSC, these cells were isolated from BM and spleens of LLC TB mice. BMS-595 caused a small decrease in the number of cells recovered after a 24-hour culture of BM PMN-MDSC, and no increase in apoptosis was observed (Fig. 5A). Similar results were observed with treatment of splenic PMN-MDSC with the CK2 inhibitor (Fig. 5B). Altogether, these results indicate that cell death plays a relatively minor role in the decrease of PMN-MDSC by CK2 inhibition in TB mice. It raised the possibility that the CK2 inhibitor may interfere with granulocytic differentiation.
In parallel, we evaluated the effect of the CK2 inhibitor on MΦ. First, we tested the hypothesis that the CK2 inhibitor causes TAM cell death. CD11b+Ly6CloLy6G−F4/80hi MΦ were sorted from the spleens and tumors of LLC TB mice and then were treated with the CK2 inhibitor. BMS-595 did not affect the total number of MΦ (Fig. 5C) or cause apoptosis in these cells (Fig. 5D). Next, we evaluated the effect of the CK2 inhibitor on MΦ differentiation. We tested the possibility that the CK2 inhibitor can block differentiation of M-MDSC to MΦ. Monocytic cells were isolated from bone marrow of naive mice and were cultured for 3 days with BMS-595 in the presence of TES. The CK2 inhibitor significantly reduced the proportion of macrophages generated from monocytic cells. The proportion of DC generated from monocytic cells was not changed (Fig. 5E). These results indicate that BMS-595 inhibited differentiation of PMN-MDSC and macrophages from their precursors without causing apoptosis of these cells.
Molecular mechanisms regulating the effect of the CK2 inhibitor on myeloid cells
Although the pro-apoptotic effects of CK2 inhibition on tumor cells is well established (7, 8, 24), the role of CK2 in the inhibition of granulocyte and macrophage differentiation is a novel finding. To elucidate this mechanism, LLC TB mice were treated for 2 weeks with BMS-595. HPCs were isolated from the BM and transcriptome was evaluated by RNA-seq. CK2 inhibitors caused changes in 2633 genes with FDR <5%. Most significantly changed genes are presented in Fig. 6A. Although no changes in apoptosis were determined in these cells, genes associated with apoptosis were upregulated in HPC in mice treated with the CK2 inhibitor whereas downregulated pathways included oxidation of lipids (Fig. 6B). The CK2 inhibitor caused increased expression of genes regulated by proinflammatory mediators: LPS, IL1β, TNF, NF-κB, and IL6. It also included E1A-binding protein P400, which is involved in transcriptional activation of E2F1, which is regulated by the retinoblastoma (Rb) protein. Not surprisingly, Rb1 was one of the regulators that were prominently inhibited by BMS-595 in HPC (Fig. 6). Consistent with previous results, the CK2 inhibitor caused increased expression of genes regulated by Notch signaling (Supplementary Fig. S4). However, it did not prevent the substantial decrease in PMN-MDSC. Therefore, we further studied genes regulated by transcription factors that were previously directly implicated in PMN-MDSC differentiation: IRF8, C/EBPα, and C/EBPβ. BM-derived HPC from TB mice were treated with GM-CSF and BMS-595 for 24 and 48 hours. CK2 inhibition caused a significant increase in the expression of only few genes regulated by IRF8 (Fig. 7A). To test whether this upregulation was sufficient to block granulocytic differentiation, we used HPC isolated from the BM of IRF8 knockout (KO) mice. CK2 inhibitor caused similar reduced differentiation of granulocytes from WT and IRF8 KO HPC (Fig. 7B), which exclude possible involvement of IRF8 in CK2 inhibitor-mediated effects. Treatment of HPC with BMS-595 did not affect genes regulated by C/EBPβ, however, it caused downregulation of many genes controlled by C/EBPα (Fig. 7A). C/EBPα is a critical regulator of myeloid differentiation. Active form of C/EBPα is C/EBPα-p42 protein. C/EBPα-p30 protein is produced by alternative transcription initiation. It retains the DNA-binding domain but lacks the N terminal transactivation domain of the longer form of C/EBPα-p42. C/EBPα-p30 exhibits a dominant negative function over C/EBPα-p42 (25, 26). Treatment of HPC with the CK2 inhibitor reduced the presence of the functionally active p42 subunit of C/EBPα in the cytoplasm and increased the dominant negative p30 subunit of C/EBPα in both the cytoplasm and nucleus (Fig. 7C). No differences were found in phosphorylation of three major sites of C/EBPα (S21, T222/226, S193; Fig. 7D). These results suggest possible mechanism regulating function of C/EBPα by CK2 inhibition.
This study was based on previous observations that CK2 may be involved in the regulation of MDSC and DC differentiation in cancer via inhibition of Notch signaling (5). We suggested that inhibition of CK2 activity with potent and selective inhibitors could improve differentiation of DC and potentially antitumor immune responses. In TB mice, the effect of the CK2 inhibitors on DC was modest. Unexpectedly, we observed a dramatic decrease in the presence PMN-MDSC and TAM. This effect was not observed in tumor-free mice. It is possible CK2 activity is more critical for abnormal myelopoiesis observed in TB mice. CK2 inhibitors dramatically enhanced the antitumor activity of anti-CTLA-4 antibody in several mouse tumor models. Since direct antitumor effects of CK2 inhibitor are marginal, it suggested that regulation of myeloid cells could be responsible for the observed synergistic antitumor effect. CK2 inhibition has a known pro-apoptotic effect in various solid and hematologic malignancies (7, 8, 24). However, its effect on myeloid cells was described in only several studies. CK2 was implicated in the regulation of phagocytosis by macrophages (27). It was also involved in the regulation of p47phox and p40phox components of NADPH oxidase in neutrophils and macrophages (28–30). No information existed on the effect of CK2 inhibition on myeloid cell differentiation.
Based on known mechanisms of action, we expected that the CK2 inhibitors would cause apoptosis of myeloid cells. However, this was not the case. Instead, its effect was associated with a block of differentiation from the precursors of granulocytes and macrophages. Because M-MDSC differentiate into TAM in tumors and MΦ in the spleens, this may explain the observed substantial decrease of MΦ after the treatment. PMN-MDSC do not differentiate from precursors in tumors and accumulate there after migration from the circulation. The decrease of PMN-MDSC caused by CK2 inhibition was observed in the BM and spleen but not in tumors. This can be explained by the fact that despite depletion, there was still enough PMN-MDSC in circulation to migrate to the tumor site. The biological effect of CK2 inhibition apparently involves a systemic decrease of PMN-MDSC that attenuates the suppression of T cells expanded in peripheral lymphoid organs and depletion of TAM in tumors that decrease local suppression.
In the search for a mechanism of how CK2 inhibition could affect myeloid cell differentiation, we found that CK2 inhibition significantly reduced the expression of genes regulated by the C/EBPα transcription factor, a leucine zipper transcription factor mainly involved in monopoiesis and granulopoiesis. It also can affect gene expression independent of DNA-binding via interaction with E2F1 (31, 32), HDAC1 or HDAC3 (33, 34). A low level of Cebpa RNA expression is detected in Lin−Sca-1+c-Kit+ (LSK) cells; Cebpa expression increases twofold in CMP and tenfold in GMP (35). Reduced levels of C/EBPα may contribute to monopoiesis by hetero-dimerizing with AP-1 proteins such as c-Jun and c-Fos (36, 37). This may explain the fact that we did not observe changes in monocytic cells generated from HPC in the presence of the CK2 inhibitor. C/EBPα induces transcription of several regulatory proteins required for subsequent lineage maturation. This includes the transcription factors C/EBPϵ, Gfi-1, KLF5 (38). C/EBPα cooperates with PU.1, c-Myb, and RUNX1 to activate genes such as myeloperoxidase, neutrophil elastase, lysozyme, lactoferrin, CSF1R, CSF2R, CSF3R in immature granulocytic or monocytic cells (39). Although the role of C/EBPα in granulocytic differentiation is well established, it is only in recent years that evidence has emerged suggesting it has a possible regulatory role in the differentiation of MΦ. Transient expression of C/EBPα and PU.1 in THP-1 cells synergistically promoted differentiation of monocytes to MΦ (40). C/EBPα enabled the induction of a monocytic cell differentiation program (41). Our results suggest that activation of CK2 in cancer can contribute to accumulation of PMN-MDSC and MΦ. The specific molecular mechanisms by which CK2 influences their accumulation warrants further investigation.
IRF8 is an essential transcription factor for the development of myeloid cells. It plays an important role in the switch between monocytic and granulocytic differentiation (42, 43) and its downregulation was implicated in the accumulation of PMN-MDSC (44–46). It was previously shown that IRF8 can physically interact with C/EBPα and prevents its binding to chromatin blocking the ability of C/EBPα to stimulate transcription and neutrophil differentiation. A partial inhibition of C/EBP activity in irf8−/− hematopoietic progenitors alleviates the neutrophil overproduction in vivo (47). In our experiments, CK2 inhibition did not reverse granulocyte hyperproduction from irf8−/− progenitors in vitro, suggesting that its effect was independent of Irf8. C/EBPα is phosphorylated at several sites, indicating a potential role for posttranslational modifications in mediating C/EBPα activity. Several kinases were implicated in phosphorylation of C/EBPα. S21 phosphorylation is mediated by Erk1/2 and interferes with granulocytic differentiation (48). Phosphorylation of C/EBPα at S248 is mediated by activated Ras, leading to the enhanced ability of C/EBPα to upregulate the expression of the granulocyte-colony stimulating factor receptor (49). Phosphorylation at S193 blocks proliferation of hepatocytes (50). Dephosphorylation of S193 promotes proliferation by preferentially binding the retinoblastoma protein (Rb) in the repressive Rb–E2F complex. A role for S193 in modulating granulopoiesis has not been reported. Finally, phosphorylation at T222 was associated with glycogen synthase kinase 3 (GSK3 kinase) activity, which regulates preadipocyte differentiation (51). We did not observe changes in phosphorylation of C/EBPα caused by the CK2 inhibitor. It is possible, however, that CK2 utilizes different sites for phosphorylation. In our study, we observed up-regulation of the 30 kDa dominant-negative isoform of C/EBPα-p30 by the CK2 inhibitor. It is known that p30 can compete for C/EBPα specific promoters with the active p42 isoform (52). Our data suggest that CK2 may affect the balance of these isoforms and thus inhibit C/EBPα activity. It is possible that the sensitivity of monocytic cells to CK2 inhibitors is different at different stages of maturation or CK2 inhibitors may affect other cell types (for instance, fibroblasts) that support myeloid cell differentiation. These possibilities require further elucidation.
Thus, our study demonstrates a novel effect of CK2 inhibition, resulting in a decrease of immune suppressive and tumor-promoting population of PMN-MDSC and TAM in TB mice. This resulted in a very substantial augmentation of antitumor activity of immune therapy. Our results suggest a novel mechanism by which CK2 may interfere with myeloid cell differentiation, leading to the inhibition of C/EBPα activity. Taken together, small molecule inhibitors targeting CK2 should be considered as a potentially valuable addition to immune-based combination therapies.
Disclosure of Potential Conflicts of Interest
A. Hashimoto is a senior researcher at Daiichi Sankyo Co., Ltd. Heshani Desilva has ownership interest (including stock, patents, etc.) in BMY. M. Jure-Kunkel is a director at Bristol-Myers Squibb and has ownership interest (including stock, patents, etc.) in Bristol-Myers Squibb. No potential conflicts of interest were disclosed by the other authors.
Conception and design: A. Hashimoto, C. Gao, A.V. Purandare, M. Jure-Kunkel, D.I. Gabrilovich
Development of methodology: A. Hashimoto, A.V. Purandare, D.I. Gabrilovich
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A. Hashimoto, C. Gao, J. Mastio, H. Desilva, J. Hunt
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A. Hashimoto, C. Gao, J. Mastio, A. Kossenkov, S.I. Abrams, S. Wee, M. Jure-Kunkel, D.I. Gabrilovich
Writing, review, and/or revision of the manuscript: A. Hashimoto, C. Gao, J. Mastio, A. Kossenkov, S.I. Abrams, A.V. Purandare, H. Desilva, S. Wee, M. Jure-Kunkel, D.I. Gabrilovich
Study supervision: D.I. Gabrilovich
Other (design and synthesis of novel CK2 inhibitors used in the publication): A.V. Purandare
Other (major role in designing the previously undisclosed CK2 inhibitor used in biological experiments): J. Hunt
This work was supported in part by grant from Bristol-Myers Squibb and by NIH grant CA 084488 to D. Gabrilovich.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.