Stimulation of an anti-tumor immune response with “chromatin-damaging” therapy

Abstract

Curaxins are small molecules that bind genomic DNA and interfere with DNA-histone interactions leading to the loss of histones and decondensation of chromatin. We named this phenomenon ‘chromatin damage’. Curaxins demonstrated anti-cancer activity in multiple pre-clinical tumor models. Here, we present data which reveals, for the first time, a role for the immune system in the anti-cancer effects of curaxins. Using the lead curaxin, CBL0137, we observed elevated expression of several group of genes in CBL0137-treated tumor cells including interferon sensitive genes, MHC molecules, some embryo-specific antigens suggesting that CBL0137 increases tumor cell immunogenicity and improves recognition of tumor cells by the immune system. In support of this, we found that the anti-tumor activity of CBL0137 was reduced in immune deficient SCID mice when compared to immune competent mice. Anti-tumor activity of CBL0137 was abrogated in CD8⁺ T cell depleted mice but only partially lost when natural killer or CD4⁺ T cells were depleted. Further support for a key role for the immune system in the anti-tumor activity of CBL0137 is evidenced by an increased antigen-specific effector CD8⁺ T cell and NK cell response, and an increased ratio of effector T cells to Tregs in the tumor and spleen. CBL0137 also elevated the number of CXCR3-expressing CTLs in the tumor and the level of interferon-γ-inducible protein 10 (IP-10) in serum, suggesting IP-10/CXCR3 controls CBL0137-elicited recruitment of effector CTLs to tumors. Our collective data underscores a previously unrecognized role for both innate and adaptive immunity in the anti-tumor activity of curaxins.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00262-020-02846-8.

Introduction

Immune checkpoint inhibitors have revolutionized the treatment of cancer by reinvigorating the ability of the immune system to recognize and kill tumor cells, leading to improve survival and even durable responses in a subset of patients. However, most tumors manage to escape anti-tumor immune attack using different mechanisms including active suppression of immune cells and a reduction in the immunogenicity of tumor cells.

Several DNA damaging agents increase the immunogenicity of tumor cells [1], and several mechanisms to explain this intriguing phenomenon were proposed [2]. A central role was given to the accumulation of cytoplasmic DNA due to release of damaged genomic DNA [3] from the nucleus. In particular, the role of cytoplasmic DNA-recognizing cyclic GMP-AMP synthase/stimulator of interferon (IFN) genes (cGAS/STING) pathway was established in several studies [4–6] since it is also known that induction of interferons through cGAS/STING pathway is critical for recognition of tumor cells by the immune system [7–9]. We have identified another way of causing activation of IFN signaling in tumor cells without induction of DNA damage. Curaxins are carbazole based DNA intercalators with demonstrated anti-cancer activity in a broad panel of preclinical models [10–15]. They were discovered in a phenotypic screening for simultaneous activation of the tumor suppressor p53 and inhibition of the pro-cancerous transcription factor nuclear factor kappa B (NF-κB) [10]. Analysis of changes in gene expression in cells treated with curaxins revealed that one of the most reactive pathways upregulated in response to the clinical lead compound, CBL0137, is the IFN pathway [16]. This unexpected combination of properties, inhibition of NF-κB and activation of IFN signaling, compelled us to look more closely into the mechanism(s) associated with CBL0137’s effects on tumor cells.

Binding of curaxins to DNA does not cause DNA damage but instead, interferes with DNA-histone interactions leading to a genome-wide destabilization of nucleosomes and decondensation of chromatin [17]. This makes many chromatin regions, including constitutive heterochromatin, accessible for transcription machinery. Divergent transcription of centromeric and pericentromeric repeats leads to the accumulation of double-stranded RNA (dsRNA) and mimics viral invasion. dsRNA is recognized by cytoplasmic nucleic acid sensitive receptors and this is what appears to activate the IFN response [16]. These findings suggested that curaxins may also increase immunogenicity of tumor cells through a pathway complementary to the one induced by DNA damaging drugs. This is an important possibility because the use of curaxins to increase tumor cell immunogenicity could be far less toxic due to the absence of permanent genotoxicity which persists even when DNA damaging drugs are no longer being given.

Here we explored, for the first time, possible immunogenic effects of curaxin in a mouse model of cancer and identified several likely mechanisms by which this effect is occurring. We used in this study the clinical lead compound, CBL0137, which is currently being tested in Phase I trials in patients with advanced metastatic cancers.

Materials and methods

Mice

Six to eight weeks old female BALB/c mice (Charles River) and SCID mice (the Laboratory Animal Resource at Roswell Park Comprehensive Cancer Center) were maintained in specific pathogen-free facilities and the protocols of all experiments with mice in this study were approved by the Institutional Animal Care and Use Committee at Roswell Park Comprehensive Cancer Center.

Cell culture

MM1.S cells (CRL-2974™), H522 cells (CRL-5810™), CT26.WT (CRL-2638™) and CT26.CL25 tumor cells (CRL-2639™) were purchased from and authenticated by American Type Culture Collection (ATCC). MM1.S, H522 and CT26.WT cells were cultured with Roswell Park Memorial Institute (RPMI) 1640 medium including 10% fetal bovine serum while CT26.CL25 cells were cultured with RPMI 1640 medium including 10% fetal bovine serum, 2 mM l-glutamine, 4.5 g/L glucose, 1.0 mM sodium pyruvate, 1.5 g/L sodium bicarbonate, 0.1 mM non-essential amino-acids, 10 mM HEPES, and 0.4 mg/mL G418. Once thawed, cells were cultured at 37 °C in an incubator with 5% carbon dioxide and 95% air and passed twice prior to tumor implantation.

Tumor implantation

5 × 10⁵ CT26.WT or CT26.CL25 cells in 50 μL PBS were subcutaneously implanted into hindlimbs of the BALB/c or SCID mice. Tumor growth was monitored by measuring perpendicular diameters (width/length) every other day and tumor volume was calculated by using the formula ((width² × length)/2). Five BALB/c mice per group were euthanized on Day 1 post the last dose of CBL0137 to collect tumors and spleens for immune analysis. Tumor growth of the remaining BALB/c mice was monitored until the width or length of tumors reached 2 cm at which point the mice were euthanized. Similar tumor growth criteria were used for tumor growth studies in SCID mice.

Treatment of CBL0137

For CT26.WT, BALB/c or SCID mice were injected via tail vein with 90 mg/kg CBL0137 [16] (provided by Cleveland Biolabs Inc.) in 5% dextrose while control mice received 5% dextrose on Day 8 post tumor implantation followed by another dose of CBL0137 on Day 7 post the first dose, respectively.

For CT26.CL25, BALB/c or SCID mice were injected via tail vein with 50 mg/kg CBL0137 [18, 19] (provided by Cleveland Biolabs Inc.) in 5% dextrose while control mice received 5% dextrose on Day 7 post tumor implantation followed by another two doses of CBL0137 on Day 4 and Day 8 post the first dose, respectively.

Depletion of immune cells

BALB/c mice were treated weekly with 400 μg anti-mouse CD8α (53-6.72, BioXCell), or anti-mouse CD4 (GK1.5, BioXCell) or corresponding isotype controls (2A3, LTF-2, BioXCell), or 30 μL anti asialo antibody (986-10001, Wako Chemicals) or normal serum (140-06571, Wako Chemicals) by i.p. injection 1 day prior to CT26.CL25 tumor implantation. The mice were treated with CBL0137 as mentioned before. CD8⁺, CD4⁺ T cell or NK cell depletion was confirmed by flow cytometry.

Gene expression analysis

MM1.S cells were treated with three doses of CBL0137 (0.3, 1, 3 μM) for 6 h, respectively for microarray hybridization. Total RNA from MM1.S cells was isolated using TRIzol reagent (Invitrogen). Two biological replicates of each condition were used for microarray hybridization RNA processing, labeling, hybridization, generation of libraries, and sequencing were performed by the Genomics Shared Resource (Roswell Park). Human Expression BeadChip array (Illumina, San Diego, CA) was used for hybridization. Differential gene expression analysis was done using lumi (PMID: 18467348).

NY-ESO-1-specific T-cell activation assay

HLA-A*0201⁺ H522 cells were treated with or without 0.5 μM CBL0137 for 24 h. CBL0137 was washed out completely and adherent cells were harvested using trypsin. The H522 cells (2.5 × 10⁴) and HLA-A*0201-restricted NY-ESO-1-specific human CD8⁺ T cells [20] (5 × 10³) were seeded in 0.1 mL of RPMI-1640 medium per well of 96 well MultiSreen-HA filter plates (MAHAS4510, Millipore^®, Billerica, MA, USA) that had been coated with an antibody against IFN-γ (1-D1K, Mabtech). After 21–22 h of culture, the plates were washed and developed with biotinylated detection anti-IFN-γ antibody (7-B6-1, Mabtech), streptavidin-conjugated alkaline phosphatase (Mabtech^®, Nacka Strand, Sweden) and 5-bromo-4-chloro-3-indolyl-phosphate/nitro blue tetrazolium (Sigma^®) to count the number of IFN-γ-secreting cells on the plates using a CTL ImmunoSpot™ Analyzer (Cellular Technologies^®, Cleveland, OH, USA). Average counts from triplicate wells were used for analyses. The maximum count for wells with only T cells was 1 in all experiments.

Flow cytometry

Single-cell suspensions were created by excising and cutting mouse tumors into 2–3 mm pieces. CT26.CL25 tumors were dissociated with collagenase/hyaluronidase (STEMCELL Technologies, 07912) following the manufacturers’ protocols. Spleens were mechanically disrupted and red blood cells were lysed using ammonium–chloride–potassium (ACK) buffer (Gibco). All segregated cells from tumors, spleens were filtered using a 70 μM nylon cell strainer (Corning).

Cells were then washed with flow running buffer (0.1% bovine serum albumin in PBS) and incubated with anti-CD16/32 (Fc receptors blocker, 1:200) at 4 °C for 15 min. CD8⁺ T cell response to CT26 tumor antigen AH1 was assessed by the antigen-specific dextramer staining with 5 μL dextramer APC (H-2L^d SPSYVYHQF AH-1, JG3294, Immudex) per sample for 10 min at room temperature. Cells were then stained with the following extracellular antibodies: anti-CD8α BUV395 (53-6.7, BD), anti-CD4 PerCP (GK1.5, BD), anti-CD3 APC-Cy7 (145-2C11, BD), anti-CD45 FITC (30-F11, BD), anti-CD45 BUV395 (30-F11, BD), anti-CD44 BV711 or PECy7(J43, BioLegend), anti-CXCR3 BV605 (CXCR3-173, BioLegend), anti-CD11b BV605 (M1/70, BioLegend), anti-CD11b APC (M1/70, BioLegend), anti-CD27 BV786 (LG.3A10, BD), anti-CD27 BV421 (LG.3A10, BioLegend), anti-Ly6C PerCP (HK1.4, BioLegend), anti-Ly6G BV421 (1A8, BioLegend), anti-F4/80 PE-Cy7 (BM8, BioLegend), anti-CD206 PE (C068C2, BioLegend), anti-CD86 BV605 (GL-1, BioLegend) and anti-NK1.1 APC-Cy7 (PK136, BD). Live/dead aqua dye (Thermo Fisher) was used to gate out dead cells. For intracellular staining, cells were stimulated with cell activation cocktail containing phorbol myristate acetate (PMA)/ionomycin and brefeldin A at 37 °C for 4 h and then were surface-stained as above, then fixed and permeabilized using the forkhead box P3 (FoxP3)/Transcription Factor Staining Buffer Set (eBiosciences) according to the manufacturer’s protocol. Cells were then stained with anti-FoxP3 PE (MF23), anti-T-bet PE/Cy7 (4B10) from BD biosciences, anti-IFN-γ APC (XMG1.2) from eBiosciences. All data was collected on a LSR Fortessa flow cytometer (BD biosciences) and analyzed with FCS Express 7 software (De Novo Software).

Cytokines/chemokines analysis

For detecting the levels of cytokines/chemokines, serum samples from the experimental mice were analyzed using Millipore MCYTOMAG-70K Panel according to manufacturer’s instructions. The plates were run on a Luminex 200 machine and the data were analyzed with Upstate BeadView software.

Statistical analysis

Student’s t test was used to compare data between two groups and tumor growth statistics were calculated using two-way ANOVA with Tukey analysis in GraphPad Prism. One-way ANOVA with Tukey posthoc tests was used to compare data between multiple groups. All data are presented as mean ± SEM unless otherwise specified. Differences were considered as statistical significance at P < 0.05.

Results

CBL0137 increases expression of genes associated with immunogenicity in tumor cells

To determine the ability of CBL0137 to augment tumor cell immunogenicity, we performed analyses of gene expression changes in tumor cells treated with CBL0137 using previously published data of micro-array hybridization and RNA sequencing [21–24]. In addition to expression of IFN-sensitive genes post-CBL0137 treatment [16], we found elevated expression of genes associated with MHC (HLA-A, HLA-B, HLA-G), immune activating factors (STAT1, CD226, PTGS2, IDO1) and one of the genes coding testes specific antigens NY-ESO-1, CTAG1A, (Fig. 1a; Supplementary Table 1). To test the functional importance of CBL0137-induced immunogenicity in tumor cells, we used NY-ESO-1 specific T cell clone which produces IFN-γ upon antigen activation measured by ELISpot assay. We observed that incubation of HLA compatible tumor cells with CBL0137 increased the ability of T cells to recognize tumor cells (Fig. 1b). Collectively, these data show that CBL0137 increases expression and markers of immunogenicity of tumor cells and their recognition by T cells.

Fig. 1.

CBL0137-induced immunogenicity in tumor cells. a Effect of CBL0137 on the expression of markers of immunogenicity in tumor cells. Heat plot of dose-dependent changes in mRNAs of different genes associated with cell immunogenicity in MM1.S cells treated with three doses of CBL0137 (0.3, 1, 3 μM) for 6 h assessed using microarray hybridization (see also supplementary Table 1). b Effect of CBL0137 on NY-ESO-1 antigen recognition by specific T cells. Co-culture of NY-ESO-1-specific T cells and the H522 cells treated with 0.5 μM CBL0137 for 24 h. IFNγ ELISpot count was shown (no drug: black; CBL0137: green, data were presented as mean ± SD, n = 3, **P < 0.01)

Anti-tumor effects of CBL0137 are blunted in mice lacking adaptive immunity

Anti-tumor activity of CBL0137 in several pre-clinical mouse tumor models was more pronounced in immune competent animals compared with immune deficient, suggesting an important role for the adaptive immune response in the mechanism by which curaxin exerts its anti-tumor activity [10–12]. To determine the impact of the adaptive immune response on the anti-tumor activity of CBL0137, we compared tumor growth kinetics of CT26.WT (parental) and a derivative called CT26.CL25 post-systemic CBL0137 treatment in syngeneic immunocompetent BALB/c mice or immunodeficient SCID mice, which genetically lack B and T cells. The CT26.CL25 mouse colon cancer model (hereafter referred to as CT26) is a preclinical model with a defined tumor associated antigen for immunotherapy [25, 26], which was generated from CT26.WT by retroviral transduction with the lacZ gene [27]. Mice were treated with three i.v. doses of CBL0137 (50 mg/kg), each separated by 4 days (Fig. 2a). Both tumor types grew faster in SCID mice than in BALB/c, a finding consistent with the concept of immunosurveillance. However, administration of CBL0137 caused stronger growth suppression of both tumor models in BALB/c compared with SCID mice (Fig. 2b; Supplementary Fig. 1). These data demonstrated that the presence of both innate and adaptive immunity contribute to the anti-cancer activity of CBL0137.

Fig. 2.

CBL0137 treatment enhances the tumor control in a CT26.CL25 colon carcinoma model. a Experimental design. b Comparison of tumor growth in BALB/c mice and SCID mice which received vehicle, or CBL0137. c Comparison of tumor growth in BALB/c mice received vehicle or CBL0137 which were depleted of CD8⁺ T cells, CD4⁺ T cells, NK cells or corresponding isotype, respectively. n = 5–12, data are presented as mean ± SEM. *P < 0.05; **P < 0.01; ****P < 0.0001

To investigate which immune cells are involved in tumor control of CBL0137, BALB/c mice were treated with CD8⁺, CD4⁺ T cells or NK cells depleting antibody and compared anti-cancer activity of CBL0137 in these mice comparing with animals treated with corresponding isotype control antibodies. We found that depletion of any of cells reduced anti-cancer effect of CBL0137. Depletion of CD8⁺ caused the strongest reduction in anti-tumor activity of CBL0137, while depletion of CD4⁺ T cells or NK cells had milder effect (Fig. 2c). These data suggest that CD8⁺ are critical component on CBL0137 induced anti-tumor immunity with additional contributing effect of CD4⁺ T cells and NK cells.

CBL0137 stimulates anti-tumor immunity while reducing tumor-induced immunosuppression

To clarify the role of immune system in the anti-tumor activity of CBL0137, we assessed the number and activation status of immune cells in tumors and spleens isolated from BALB/c mice on day 1 after three doses of CBL0137 (same scheme as illustrated in Fig. 2a). We first measured the CD8⁺ T cell response, since CD8⁺ T cells are the main effectors of adaptive immunity that control tumor growth [28] and depletion of CD8⁺ T cell caused the greatest reduction of anti-tumor activity of CBL0137 (Fig. 2c). We found an increased presence of total CD8⁺ T cells in tumors of mice treated with CBL0137, compared to vehicle treated mice (Fig. 3a top left). In addition, the number of effector CD8⁺ T cells marked by the CD44 receptor was significantly higher in CBL0137 treated mice than in control vehicle treated mice (Fig. 3a top right), suggesting that CBL0137 increases the number of antigen specific effector CD8⁺ T cells.

Fig. 3.

The expression of important effector molecules in CD8⁺ T cells in tumor (a), and spleen (b) including IFNγ⁺, CD44 and CXCR3 in CT26.CL25 tumor-bearing BALB/c mice which received vehicle (black), or CBL0137 (yellow). n = 5, data are presented as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001

To test this hypothesis we used the fact that the tumor-specific CD8⁺ T cell response against CT26 is directed against a single immunodominant peptide antigen, AH1 (SPSYVYHQF), which is derived from gp70 and corresponds to amino acids^423-431 [29]. We used a dextramer complexed to the AH1 peptide sequence to track the tumor antigen specific CD8⁺ T cell response post-CBL0137 treatment. We found that AH1 dextramer reactive CD8⁺ T cells (Fig. 3a middle left) and those expressing the CD44 receptor (Fig. 3a middle right) were significantly increased in the tumor after CBL0137 treatment, indicating that CBL0137 triggers a tumor antigen specific effector CD8⁺ T cell response.

We next characterized the nature of the anti-tumor CD8⁺ T cell response post-CBL0137 treatment. Cytotoxic CD8⁺ T cells (CTLs) produce copious amounts of IFN-γ [30] and express the homing receptor CXCR3 [31], which controls T cell migration into inflamed sites such as tumors. We observed an increase in the number of IFN-γ expressing CD8⁺ T cells in the tumor of mice treated with CBL0137 (Fig. 3a bottom left). In addition, CBL0137 increased the number of CXCR3-expressing CD8⁺ T cells in the tumor (Fig. 3a bottom right), indicating that CBL0137 may improve the homing capacity of CD8⁺ T cells into tumors.

The spleen is a secondary lymphoid organ indispensable for supporting the generation of an anti-tumor immune response, providing the rationale to measure whether CBL0137 also elicits an anti-tumor immune response in the spleen. Similar to the tumor, CBL0137 augmented the total number of tumor antigen specific CD8⁺ T cells in the spleen (Fig. 3b left) and those expressing CD44 (Fig. 3b right). These data indicate that systemic CBL0137 treatment stimulates a tumor specific cytotoxic CD8⁺ T cell response in both the tumor and spleen.

The CD8⁺ T cell infiltrate is considered beneficial for patient survival and a high ratio of T effector/Tregs in the tumor microenvironment correlates with the anti-tumor effects of targeted therapy for different types of cancer [32–34]. CBL0137 treatment increased ratios of CD44⁺ CD8⁺ T cells to Tregs, AH1⁺ CD8⁺ T cells to Tregs as well as IFN-γ⁺ CD8⁺ T cells to Tregs in the tumor (Fig. 4a). The same changes as well as decreased number of Tregs were also observed in spleen (Fig. 4b). In addition, CBL0137 decreased the number of immunosuppressive M2 tumor-associated macrophages (TAM) and myeloid-derived suppressor cells (MDSC), including both M-MDSC and PMN-MDSC subsets while having no effect on anti-tumor M1 TAM (Supplementary Fig. 2). These data demonstrate that CBL0137 treatment relieves tumor-induced immunosuppression while also eliciting a tumor-specific effector CD8⁺ T cell response with homing potential for the tumor.

Fig. 4.

The ratios of effector CD8⁺ T cells to Tregs in tumor (a), and spleen (b) of CT26.CL25 tumor-bearing BALB/c mice which received vehicle (black), or CBL0137 (yellow). n = 5, data are presented as mean ± SEM. *P < 0.05; **P < 0.01

CD4⁺ and CD8⁺ T cells interplay in the control of tumor growth [32] and CD4⁺ T cells release help signals to CD8⁺ T cells to optimize the magnitude and quality of the cytolytic response against tumor antigen [35]. Total CD4⁺ T cells and those expressing the CD44 receptor as a marker of effector/memory similar to CD8⁺ T cells increased in the tumor after CBL0137 treatment (Fig. 5a), indicating the potential role for CD4⁺ T cell help in the generation of an anti-tumor CD8⁺ T cell response post-CBL0137. Similar to the CD8⁺ T cell response, CBL0137 also increased the number of IFN-γ-expressing CD4⁺ T cells in the spleen (Fig. 5b), further supporting CBL0137 improving the cytotoxic function of T cells towards the tumor.

Fig. 5.

The expression of important effector molecules in CD4⁺ T cells in tumor (a), and spleen (b) as well as NK cells in tumor (c) and spleen (d) including CD44⁺, IFNγ⁺ and T-bet⁺ of CT26.CL25 tumor-bearing BALB/c mice which received vehicle (black), or CBL0137 (yellow). n = 5, data are presented as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001

We also looked whether CBL0137 treatment affects innate anti-tumor immunity which plays a critical role in triggering the adaptive immune response. NK cells are part of the innate immune system [36]. Thus, we measured the number and activation status of NK cells in the tumor and spleen post-CBLB0137 therapy. The quantity of NK cells in the tumor and spleen of CBL0137 treated mice was significantly higher than in vehicle treated mice (Fig. 5c, d left panels). NK cells constitutively express IFN-γ mRNA, which allows for rapid synthesis and secretion of IFN-γ protein [37]. In addition, the anti-tumor activity of NK cells in metastasized cancer strongly depends on T-bet expression in NK cells [38–40]. Treatment with CBL0137 stimulated an accumulation of NK cell with both of these functional markers in the tumor (Fig. 5c middle and right panels) and spleen (Fig. 5d middle and right panels).

CBL0137 modulates expression of systemic cytokines and chemokines

Cytokines and chemokines play a crucial role in modulating the immune response by mediating effector function and supporting migration of immune cells into inflamed sites [41, 42]. We used a multiplex assay to measure the levels of systemic cytokines and chemokines associated with an anti-tumor immune response in the serum from BALB/c mice in the experiments presented above. We found that IP-10 (CXCL10), RANTES, Eotaxin, and MCP-1 strikingly increased in the serum of mice treated with CBL0137 (Fig. 6), suggesting that CBL0137 improves tumor control by releasing tumoricidal cytokines/chemokines to attract anti-tumor effector T cells into the tumor and enhance the presentation of tumor antigens.

Fig. 6.

The levels of cytokines and chemokines in serum from CT26.CL25 tumor-bearing BALB/c mice which received vehicle (black), or CBL0137 (yellow). n = 4–5, data are presented as mean ± SEM. *P < 0.05; **P < 0.01

Discussion

Curaxin CBL0137 is the first-in-class “chromatin damaging” agent, i.e. DNA binding compounds which disrupts DNA/histone interactions without causing direct DNA damage, i.e. alterations of chemical structure of DNA. Instead, CBL0137 alters biophysical properties of DNA (shape, charge, flexibility) and thus interferes with DNA packaging into chromatin [17]. There are other DNA binding compounds which destabilize chromatin in cells, but they also damage DNA [43, 44]. Interestingly, some of them (e.g. anthracyclines) were shown to be strong inducers of tumor cell immunogenicity via DNA damaging activity [45]. However, the input of chromatin damaging activity into the elevation of tumor cell immunogenicity were not evaluated. Although full mechanisms of how chromatin damage induces enhanced immunogenicity of tumor cells is not fully revealed yet, our data suggests that they all depend on the emergence of cryptic transcription in tumor cells due to chromatin decondensation [16]. Chromatin decondensation may also explain expression of epigenetically silenced MHC class of molecules and embryo-specific antigens etc., which we observed in this study. Although the consequence of these mechanisms on tumor cell immunogenicity still needs to be demonstrated, here we provide the first evidence that CBL0137 treatment significantly increases tumor recognition by immune cells.

In stark contrast to DNA damaging therapies that kill immune cells, chromatin damaging therapies such as CBL0137 may in fact activate immune cells via decondensation of chromatin and reversal of epigenetic mechanisms of T cell exhaustion [46]. In support of this idea, we show that the anti-tumor activity of CBL0137 is blunted in SCID mice as well as BALB/c mice depleted of CD8⁺, and to a lesser extent, CD4⁺ T cells or NK cells, which is consistent with our findings showing that CBL0137 increases the effector T cell response, NK cell response, and ratio of effector T cells to Treg in the tumor and spleen. Effector T and NK cells produce IFN-γ upon CBL0137 treatment, which signifies tumor cytolytic activity and an IFN-γ orchestrated anti-tumor immune response via different mechanisms. For example, IFN-γ stimulates expression of MHC class I molecules on tumor cells [37], and CBL0137 independently of IFN-γ increases expression of these molecules, since we observed this in monoculture of tumor cells in the absence of immune cells. These two potentially synergistic effects should make tumor cells more readily recognized by CTLs. Second, IFN-γ regulates the expression of multiple chemokines including IP-10 [47], RANTES [48], and Eotaxin [49], all of which are elevated in the serum post-CBL0137 and support the homing and migration of effector and memory T cells into tumors. In particular, IP-10 binds to the CXCR3 receptor on activated CD8⁺ T cells and NK cells [50, 51]. Our findings that CBL0137 elevates the number of CXCR3-expressing CTLs in the tumor and the level of IP-10 in serum support the idea that IP-10/CXCR3-dependent migration of effector CTLs into tumors post-CBL0137 contributes to its anti-cancer activity. Third, CBL0137 increases the number of IFN-γ⁺ CD4⁺ T cells, suggesting that IFN-γ-producing Th1 cells rather than Th17 cells [52] mainly contribute to tumor control post-CBL0137. Fourth, IFN-γ rapidly induces T-bet expression in NK cells followed by acquisition of anti-tumor effector function [53]. The increased number of T-bet⁺ NK cells in both the spleen and tumor post-CBL0137 supports NK cells as an innate immune response that contributes to tumor control. Fifth, data from depletion of immune cells in CBL0137 treated tumor-bearing mice suggests that CBL0137 triggers antitumor immunity through a bidirectional NK and T cell cross-talk. Although the exact mechanism of cross-talk remains unclear and will be examined in further research, it is plausible that IFN-γ secreted by NK cells stimulates T cell activation and their proliferation into tumoricidal CTL and Th while NK cells receive an activation signal from tumor antigen-specific CD8⁺ T cells.

CBL0137 has the rather unique ability to simultaneously cause opposite effects on two closely related signaling pathways: it causes activation of IFN signaling [16] while it inhibits transcriptional activity of NF-κB [10]. Chronic inflammation in the tumor microenvironment stimulates recruitment of MDSCs. NF-κB plays a crucial role in the activation and function of MDSC [54]. Moreover, MDSC recruitment also promotes NF-κB-controlled Tregs to promote tumor angiogenesis and immune evasion [55], and NF-κB maintains the tumor-promoting M2 phenotype of TAM in cancer [56]. In the present study, a significant decrease in NF-κB-driven MDSC, Tregs and M2 in CBL0137 treated tumors is consistent with the ability of CBL0137 to act as an inhibitor of NF-κB [10].

In summary, the data shown in the present study provides evidence for a beneficial cell-mediated immune modulation by the lead curaxin CBL0137. Treatment with CBL0137 improves tumor control by mechanisms that involve the release of cytokines and chemokines that attract anti-tumor immune cells, including NK and CD8⁺ and CD4⁺ T cells, to the tumor and enhancing the effector function of these immune cells. These data support the use of CBL0137 as a safe and efficacious clinical agent that can control tumors in cancer patients by eliciting improved anti-tumor immunity.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

MC: performed most of experiments, wrote the manuscript, CMB: design of experiments, analyses of data, editing of manuscript, LB: performed part of mouse experiments, edited manuscript, AP: performed NY-ESO-1-specific ELISPOT assay, SKP: analyzed data, edited manuscript, JM: performed NY-ESO-1-specific ELISPOT assay, edited manuscript, AO: edited manuscript, AG: edited manuscript, AKS: design of experiments, analyses of data, editing of manuscript, ER: design of experiments, analyses of data, editing of manuscript, KG: design of experiments, analyses of data, editing of manuscript.

Funding

This study was funded in part by Roswell Park Alliance Foundation (KG), National Cancer Institute (NCI) grant R01CA197967 (KG), Incuron, LLC (KG), and NCI Cancer Center Support Grant P30CA16056 (FICSR).

Data availability

All materials are available upon request.

Compliance with ethical standards

Conflict of interest

K. Gurova is co-inventor of curaxins and recipient of research grants and consulting payments from Incuron, LLC.

Ethical approval

The protocols of all experiments with mice in this study were approved by the IACUC at Roswell Park Comprehensive Cancer Center.