A Novel Peptidylarginine Deiminase 4 (PAD4) Inhibitor BMS-P5 Blocks Formation of Neutrophil Extracellular Traps and Delays Progression of Multiple Myeloma

Multiple myeloma is a plasma cell malignancy that grows preferentially in the bone marrow (BM; ref. 1). Crosstalk of multiple myeloma cells with the BM microenvironment plays a critical role in promoting multiple myeloma cell survival and growth (2, 3). Although several mechanisms within the BM niche have been implicated in disease progression, our understanding of the contribution of interactions between tumor cells and their microenvironment remains scarce.

Neutrophils represent one of the predominant cell types in the BM and are essential effector cells of the innate immune system. Neutrophils serve as the first line of defense against a wide range of pathogens including bacteria, fungi, and protozoa, by utilizing two major mechanisms: phagocytosis and degranulation (4). Recently, formation of neutrophil extracellular traps (NET) has been described as another host defense mechanism of neutrophils against pathogens (5, 6). NETs are extracellular web-like structures composed of chromatin backbone with peptides and proteins assembled on it. Proteins and antimicrobial peptides that are normally found in neutrophil granules, cytoplasm, or nucleus, including neutrophil elastase and myeloperoxidase, are found in NETs (5, 7). NETs are formed via a novel cell death pathway called NETosis. During NETosis, disintegration of the nuclear envelope allows mixing of chromatin with cytoplasmic and granular proteins followed by their release outside the cell (7). A number of stimuli have been described to prime neutrophils to undergo NETosis including lipopolysaccharide (LPS), IL8, G-CSF, TGFβ among others, as well as artificial stimuli such as phorbol myristate acetate and calcium ionophore (8).

The occurrence of NETosis is not limited to infection; the presence of NETs has recently been documented in other pathologies including solid and blood cancers (8, 9) and was associated with promoting metastasis and poor prognosis (10–12). NETs were shown to trap circulating lung carcinoma cells (10), to stimulate cancer cell adhesion, migration, and invasion in vitro (13, 14). In addition, NETs promoted cancer-associated thrombosis (15, 16) and had pro-coagulant activity in patients with gastric cancer (17).

The mechanism of NETosis is not entirely understood. However, it has been postulated that one of the steps in NET formation involves decondensation of chromatin and that citrullination of histones, specifically histone H3, is necessary for chromatin decondensation (18). Citrullination is a process of converting arginine residues into citrulline and is mediated by a family of enzymes called peptidylarginine deiminases (PAD; ref. 19). PAD4, a member of this family, is responsible for the citrullination of histone H3 at arginine residues 2, 8, and 17, which leads to loss of the positive charge of chromatin and consequently weakens histone-DNA binding facilitating chromatin decondensation during NET formation (20). PAD4 knockout (KO) mice have shown reduced bacterial killing by NETs in a mouse infectious disease model of necrotizing fasciitis (21), reduced NET-dependent inflammatory and procoagulant host response to endotoxin in a polymicrobial sepsis model (22), and protection from NET-dependent cardiac fibrosis (23).

To date, several compounds inhibiting PAD activity have been reported. Those include Cl-amidine and related compounds that are covalent nonselective irreversible PAD inhibitors (24, 25) and GSK-484, a reversible selective PAD4 inhibitor (26). GSK-484 blocked calcium ionophore-induced NET formation by mouse neutrophils in vitro (26) and through blocking LPS-induced NET generation prevented awakening of dormant cancer cells in an in vivo breast cancer model (27). Although GSK-484 provided a proof-of-concept that pharmacologic targeting of PAD4 blocks NET formation, this compound does not possess a strong potency as it required micromolar concentrations to be effective and thus, it is not suitable for clinical development. Therefore, a need for development of more potent PAD4 inhibitors exists.

Although the majority of patients with multiple myeloma respond well to the first line of treatment, they inevitably develop therapy-resistance. Novel agents that are effective in refractory multiple myeloma as a monotherapy or in combination with commonly used treatments are needed. Here, we investigated whether multiple myeloma cells can stimulate NET formation and if so, whether interruption of this process would delay disease progression. We utilized the novel PAD4-specific inhibitor BMS-P5 developed by Bristol-Myers Squibb (BMS). Our data provide evidence that pharmacologically targeting PAD4 with BMS-P5 blocks multiple myeloma-induced NET formation and delays progression of multiple myeloma in a syngeneic mouse model.

Human multiple myeloma RPMI8226 and MM1.S cell lines were obtained from ATCC. Mouse MM DP42 cell line was a gift from Dr. Van Ness (University of Minnesota); 5TGM1 cell line was provided by Dr. Oyajobi (UT Health San Antonio). All multiple myeloma cells were cultured in RPMI1640 medium supplemented with L-glutamine, 10% FBS, and 1% antibiotic-antimycotic. DP42 cells were also supplemented with 0.5 ng/mL murine IL6 (catalog no. 406-ML; R&D Systems). Cell lines were mycoplasma negative as detected by PCR. Cells were maintained in culture for a maximum of 6 weeks.

BMS-P5, [(cis)-5-Amino-2-methylpiperidin-1-yl)(2-(1-(cyclopropylmethyl)-1H-pyrrolo[2,3-b]pyridine-2-yl)-7-methoxy-1-methyl-1H-benzo[d]imidazole-5-yl)methanone, hydrochloride; Example 24, patent US9127003B2; ref. 28] was provided by BMS. Cl-amidine (Cl-A, catalog no. 10599) and GSK-484 (catalog no. 17488) were purchased from Cayman Chemical and calcium ionophore (CaIo, catalog no. A23187) was obtained from Sigma-Aldrich.

Measurements of the inhibitory potency of BMS-P5 against PAD enzymes (PAD4, PAD1, PAD2, or PAD3) employed recombinant human proteins with recombinant histone H3 as a substrate in 100 mmol/L Tris, pH 7.5, containing 2 mmol/L DTT and 0.65 mmol/L CaCl₂. The citrullinated histone H3 was detected by either immunoblot with LiC or quantitation or an ELISA-based assay using anti-citrullinated histone H3 rabbit polyclonal antibody (Abcam ab5103) with either IRDye 800CW-conjugated or HRP-conjugated donkey anti-rabbit IgG secondary detection antibodies (Licor 926-32213 or Invitrogen A16029).

All animal experiments were carried out in accordance with institutional guidelines and were approved by the Institutional Animal Care and Use Committee of the Wistar Institute. C57BL/6 and FVB/N mice were purchased from Charles River Laboratories, and were crossed to obtain F1 progeny of mixed C57BL/6 × FVB/N background. In these mice, multiple myeloma tumors were established by intravenous inoculation into the tail vein of DP42 cells (5 × 10³). Six- to 10-week-old male and female mice were used for experiments. For in vivo treatment studies, DP42-bearing mice were split into two groups and treated with BMS-P5 or vehicle control (VC, 0.5% Methocel A4M and 0.1% polysorbate 80 in 100 mmol/L sodium acetate pH 4.6). BMS-P5 was given at a dose of 50 mg/kg and administered by oral gavage twice a day beginning on day 3 after tumor cell injection. Onset of symptoms (paralysis and hunched posture) were monitored. Survival of mice was determined as they were euthanized at the humane endpoint.

PAD4 flox mice [PAD4^fl/fl, B6(Cg)-Padi4^tm1.2Kmow/J, stock no.026708] and Mrp8Cre^Tg mice [B6.Cg-Tg (S100A8-cre,-EGFP)1Ilw/J, stock no.021614] were purchased from the Jackson Laboratory, and crossed to obtain PAD4^fl/flMrp8Cre⁺ mice (further designated as PAD4 KO) and their PAD4^fl/flMrp8Cre⁻ littermates control.

Mouse neutrophils were isolated from the BM of tumor-free mice using magnetic cell separation technique. Briefly, BM cells were labeled with biotin-Ly6G antibody followed by incubation with streptavidin conjugated microbeads and positive selection using LS columns (all from Miltenyi Biotech). The purity of isolated Ly6G⁺ cells was more than 98% as detected by flow cytometry.

Collection of BM samples from patients with multiple myeloma at the Abramson Cancer Center was approved by the Institutional Review Board (IRB) of the University of Pennsylvania. All patients signed IRB-approved consent forms. Collection of peripheral blood samples from healthy donors was approved by the Wistar IRB. Neutrophils were isolated using Percoll gradient from cells pelleted following Ficoll-Paque gradient centrifugation. Briefly, pelleted cells were resuspended in PBS and were layered over 63%/72% Percoll and centrifuged at 700 × g for 30 minutes. Neutrophils were collected from the interface between 72% (bottom) and 63% (top) layers. Primary multiple myeloma cells were isolated from the BM mononuclear fraction by positive selection of CD138 cells using CD138 MicroBeads and MACS columns (all from Miltenyi Biotec).

DP42 or 5TGM1 cells were plated at 2 × 10⁶ cells/mL in phenol-free RPMI1640 supplemented with L-glutamine and 2% FBS. Cells were collected after 24 hours and centrifuged for 5 minutes at 300 × g. Supernatants were collected and frozen.

Neutrophils were placed in 24-well flat bottom plates and cultured for 8 hours with or without multiple myeloma cells separated by Transwell insert or DP42 or 5TGM1 CM followed by fixation of neutrophils with 500 μL of 4% paraformaldehyde and staining with Sytox green nucleic acid stain (catalog no. S7020; Thermo Fisher Scientific, final concentration of 0.25 μmol/L). NETs were visualized with a TE300 inverted microscope (Nikon) equipped with a motorized XY stage. An average of 25 z-stacked images with random location (consistent pattern set-up) were acquired using a 20x objective and analyzed with the digital image analysis NIS-Elements software (Nikon). Data presented as NET area per cell (total NET area measured across 25 images divided by the number of neutrophils).

Neutrophils were plated in phenol-free RPMI1640 medium containing 2% FBS at a density of 2 × 10⁶ cells/mL and cultured alone or in the presence of DP42 or 5TGM1 cells separated by Transwell insert, DP42 or 5TGM1 CM, or 5 μmol/L calcium ionophore for the indicated time periods. In some experiments, neutrophils were pretreated for 0.5 hour with PAD inhibitors Cl-Amidine, GSK-484, or BMS-P5.

Neutrophils were lysed using RIPA buffer supplemented with 1 mmol/L EDTA and 1× Halt protease and phosphatase inhibitor (catalog no. 1861281; Thermo Fisher Scientific). Membranes were blocked with 5% nonfat dry milk in TBS-T and incubated with anti-histone H3 (citrulline R2+R8+R17; catalog no. 5103; Abcam) or PAD4 antibodies (catalog no. ABIN2856939; Antibodies Online) followed by incubation with secondary HRP-conjugated antibodies. Membranes were reprobed with antihistone H3 antibody (catalog no. 9715s; Cell Signaling Technology), and then reprobed with antibody against β-actin (catalog no. sc-47778; Santa Cruz Biotechnology).

Femurs and tibias were collected from mice and flushed with 500 μL of cold PBS. Cold acetone (30 μL) was added to 120 μL of sample, vortexed, and incubated for 30 minutes on ice. Tubes were then centrifuged at 14,000 × g at 4°C for 10 minutes. The supernatant was discarded, and the protein pellet was then rinsed with 1 mL 100% ethanol prior to centrifugation at 14,000 × g at 4°C for 10 minutes. This step was repeated followed by drying the pellet. The protein pellet was dissolved in 100 μL of 15 mmol/L Tris-HCl pH 8.8 and 6× loading buffer (catalog no. BP-111R; Boston BioProducts) was added per tube and heated for 5 minutes at 95°C. TCA precipitated samples were subjected to Western blotting.

BM cells were labeled with CD138-APC antibody (catalog no. 347207; BD Biosciences), washed with ice cold PBS, and resuspended in PBS containing DAPI. Apoptosis of multiple myeloma cells was evaluated using Annexin V binding assay. Briefly, cells were washed twice with ice cold PBS and once with binding buffer followed by staining with Annexin V-FITC (BioLegend) and DAPI (Life Technologies). At least 10,000 events were acquired using LSR II flow cytometer (BD Biosciences). Data were analyzed with FlowJo software (TreeStar).

The level of DNA present in the serum from multiple myeloma-bearing mice was determined using the Qubit dsDNA HS assay (catalog no. Q32851; Invitrogen) as per the manufacturer's protocol.

Data are presented as mean ± SD. Statistical analysis was performed using GraphPad Prism 5 software (GraphPad Software Inc). A two-tailed Student t test was used to detect significant differences (P < 0.05) between treatment groups. A log-rank test was used for the comparison of survival curves. A P value of less than 0.05 was considered statistically significant.

We evaluated the ability of multiple myeloma cells to induce citrullination of histone H3 and formation of NETs by mouse BM neutrophils. Coculturing of neutrophils with murine multiple myeloma DP42 or 5TGM1 cells separated by Transwell inserts resulted in increased histone H3 citrullination in neutrophils. Similar results were obtained when neutrophils were kept in the presence of DP42 or 5TGM1 cell CM (Fig. 1A). To evaluate NETosis, BM neutrophils were cultured with murine multiple myeloma cells or multiple myeloma cell CM followed by staining with SYTOX Green Nuclear Acid Stain, a cell-impermeable fluorescent dye, and assessment of NET formation by fluorescent microscopy. The presence of multiple myeloma cells or multiple myeloma cell CM equally and significantly up-regulated NET production by neutrophils (Fig. 1B and C).

PAD4 is required for citrullination of R2, R8, and R17 residues of histone H3. We utilized PAD4 KO mice (Fig. 2A) to investigate whether PAD4 is involved in multiple myeloma-induced NET formation. Citrullination of histone H3 was significantly increased in control neutrophils upon addition of multiple myeloma cells. In contrast, PAD4-deficient neutrophils were not able to upregulate citrullination of histone H3 in response to multiple myeloma cells (Fig. 2B). In line with these data, multiple myeloma cells were not able to induce NETosis in PAD4-deficient neutrophils in contrast to control neutrophils (Fig. 2C and D).

BMS-P5 (Fig. 3A) is a potent selective pharmacological inhibitor of PAD4 (Supplementary Table S1) and was used to evaluate the effect of PAD4 inhibition on multiple myeloma-induced NET formation. Our preliminary experiments demonstrated a lack of cytotoxicity of BMS-P5 on neutrophils (Fig. 3B and C) and myeloma cells (Fig. 3D).

Initially, the ability of BMS-P5 to inhibit NETosis was compared with that of known PAD inhibitors: nonselective inhibitor Cl-A and selective PAD4 inhibitor GSK-484. Both inhibitors were used at the previously reported doses of 100 and 10 μmol/L, respectively. All three compounds blocked calcium ionophore-induced citrullination of histone H3 (Fig. 4A). Next, we evaluated whether BMS-P5 could prevent multiple myeloma-induced NET formation. Neutrophils were pretreated with 1 μmol/L BMS-P5, 1 or 10 μmol/L GSK-484, or 100 μmol/L Cl-A for 30 minutes followed by addition of DP42 or 5TGM1 CM. As anticipated, both DP42 and 5TGM1 CM induced citrullination of histone H3 (Fig. 4B and C) and formation of NETs (Fig. 4D and E). Addition of all three PAD inhibitors significantly reduced multiple myeloma-induced citrullination of histone H3 (Fig. 4B and C) and NET formation (Fig. 4D and E). In the absence of PAD4, no induction of NET formation by multiple myeloma cells was observed, and BMS-P5 did not have any effect (Fig. 4F).

Next, we investigated the effect of PAD4 inhibition with BMS-P5 on NET formation by human neutrophils. Neutrophils were isolated from peripheral blood of healthy donors and cultured in the presence or absence of human multiple myeloma cell lines with or without addition of 1 μmol/L BMS-P5. In parallel, cells were treated with 1 or 10 μmol/L GSK-484. As anticipated, both RPMI8226 and MM1.S multiple myeloma cell lines induced NET formation (Fig. 5A; Supplementary Fig. S1). Both PAD4 inhibitors reduced NETosis stimulated by multiple myeloma cells; however, the effect of BMS-P5 was more substantial (Fig. 5A; Supplementary Fig. S1). BMS-P5 was also able to inhibit formation of NETs induced by primary multiple myeloma cells isolated from the BM of patients with multiple myeloma (Fig. 5B)

We next investigated whether pharmacological targeting of PAD4 with BMS-P5 could have an antitumor effect in a syngeneic mouse model of multiple myeloma. DP42-bearing mice were treated with BMS-P5 or vehicle control and the onset of disease symptoms (paralysis and hunched posture) and mice survival were evaluated. Administration of BMS-P5 significantly delayed development of symptoms (Fig. 6A) and significantly prolonged survival of multiple myeloma-bearing mice (Fig. 6B). In parallel, groups of BMS-P5 and vehicle control-treated mice were euthanized 13 days after tumor cell inoculation for analysis. Selection of this time point was based on known kinetics of tumor growth when a substantial proportion of DP42 cells is already present in the BM microenvironment while other hematopoietic cells still represented the majority of the BM cells. The proportion and absolute number of multiple myeloma cells were significantly reduced in BMS-P5–treated mice as compared with vehicle control-treated mice (Fig. 6C). DNA represents one of the main NET components and the presence of citrullinated histone H3 has been associated with NETosis. Therefore, we measured levels of cell-free DNA and histone H3cit in the BM plasma. Administration of PAD4 inhibitor BMS-P5 resulted in significantly reduced plasma levels of cell-free DNA (Fig. 6D) and BM level of cell-free citrullinated histone H3 (Fig. 6E). Taken together, our data suggest that pharmacological targeting of PAD4 could be beneficial for the treatment of multiple myeloma.

The overall objective of this study was to evaluate whether multiple myeloma cells can stimulate neutrophils to form NETs and whether pharmacologic blocking of NET formation would delay multiple myeloma progression. Our results, for the first time, demonstrated that multiple myeloma cells induce NETosis and that PAD4 is required for multiple myeloma-induced NET formation. We further reported a novel small molecule PAD4-specific inhibitor BMS-P5 and demonstrated that this compound was able to block multiple myeloma-induced NET formation. Administration of BMS-P5 in vivo in multiple myeloma-bearing mice delayed appearance of symptoms and prolonged mice survival in a syngeneic model of this disease.

Although the presence of NETs has been reported for several cancer types, their involvement in the pathogenesis of multiple myeloma has not yet been demonstrated. We observed an increase in citrullination of histone H3 associated with an increase in NET formation in neutrophils stimulated by multiple myeloma cells. These effects were abrogated when multiple myeloma cells were cocultured with PAD4-deficient neutrophils indicating that PAD4 is indeed required for multiple myeloma-induced NET formation. Our data are consistent with previous studies in other tumor types demonstrating that tumor cells can induce PAD4-depenedent NETosis (12, 13, 15, 10, 11, 31–34).

Histone H3 citrullination plays an important role in NET formation and is mediated by the PAD4 enzyme. PAD4 is receiving increased interest in the pharmaceutical field due to its implication in the biology of cancer and other pathologies. In this study, we aimed to understand the in vitro and in vivo effects of selective PAD4 inhibition in multiple myeloma. Several previously reported PAD inhibitors have been shown to prevent NETosis (12, 26, 27, 35); however, only one of them, GSK-484, possesses selectivity for PAD4 (26). Our findings demonstrated that, in vitro, BMS-P5 significantly reduced multiple myeloma-stimulated PAD4-dependent NET formation by mouse BM neutrophils and human neutrophils isolated from peripheral blood of healthy donors. Similar results were obtained with GSK-484 confirming robustness of these data. In vivo, administration of BMS-P5 significantly improved survival of multiple myeloma-bearing mice. Ex vivo data demonstrated that the proportion and absolute number of multiple myeloma cells in multiple myeloma-bearing mice treated with BMS-P5 was significantly reduced as compared with vehicle control-treated mice. In addition, the level of citrullinated histone H3 was reduced in the BM flushes from BMS-P5-treated mice as compared with control mice. There is a possibility that mechanisms other than blocking NETosis are responsible for the anti-multiple myeloma effect of BMS-P5. Identification of these mechanisms could represent a subject of future studies.

The protumorigenic effect of NETs may be due to many mechanisms. NETs contain a number of proteins including high mobility group box 1 (HMGB1), neutrophil elastase (NE), and matrix metalloproteinase 9 (MMP9) among others. HMGB1 has been implicated in the angiogenesis, invasion, progression, metastasis, and drug resistance of cancers (36–38). Li and colleagues showed that HMGB1 can cause immunosuppression by stimulating myeloid-derived suppressor cells and promote peritoneal spread of colon cancer cells after surgical resection (39). NE and MMP9 have also been implicated in tumor progression (40, 41). A study by Albrengues and colleagues demonstrated that NET-associated NE and MMP9 induced proteolytic cleavage of laminin which drove awakening of cancer cells and caused metastasis in mice (27). In the context of multiple myeloma, the accumulation of NET-derived DNA scaffold whereby a lot of proteases are concentrated could also alter the tumor microenvironment and provide a growth advantage to multiple myeloma cells. Therefore, by targeting the process of NET formation, BMS-P5 may attenuate the presence of protumorigenic proteins in the tumor microenvironment, and thus delay tumor progression.

Taken together, our data for the first time demonstrate that multiple myeloma cells may induce NETosis associated with PAD4 activation and that disruption of the NET formation process by pharmacologic inhibition of PAD4 has antitumor effects in a syngeneic model of multiple myeloma. These data suggest that targeting PAD4 could modify the tumor microenvironment and may be beneficial for the treatment of patients with multiple myeloma.

M. Li is a research scientist at Bristol-Myers Squibb. D.T. Vogl is a paid consultant at Karyopharm Therapeutics, Janssen, Takeda Oncology, Active Biotech, Celgene Corporation; and has report of receiving commercial research grant from Active Biotech. J.J. Burke is a Group Director at Bristol-Myers Squibb; and has ownership interest (including patents) in Bristol-Myers Squibb. No potential conflicts of interest were disclosed by the other authors.

Conception and design: C. Lin, D.T. Vogl, J.J. Burke, Y. Nefedova

Development of methodology: M. Li, C. Lin, J. Strnad, J.J. Burke

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Li, C. Lin, H. Deng, L. Bernabei, D.T. Vogl

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Li, C. Lin, D.T. Vogl, J.J. Burke, Y. Nefedova

Writing, review, and/or revision of the manuscript: M. Li, C. Lin, J. Strnad, D.T. Vogl, J.J. Burke, Y. Nefedova

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L. Bernabei

Study supervision: Y. Nefedova

This work was supported by NIH/NCI grant R01CA196788 (to Y. Nefedova). Support for the Shared Resources utilized in this study was provided by Cancer Center Support Grant P30CA010815 to The Wistar Institute.

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.