GTN057, a komaroviquinone derivative, induced myeloma cells' death in vivo and inhibited c‐MET tyrosine kinase

Objective: Despite the development of newly developed drugs, most multiple myeloma (MM) patients with high‐risk cytogenetic abnormalities such as t(4;14) or del17p relapse at anin early stage of their clinical course. We previously reported that a natural product,komaroviquinone (KQN), isolated from the perennial semi‐shrub Dracocephalum komarovi, i.e., komaroviquinone (KQN) and its derivative GTN024 induced the apoptosis of MM cells by producing reactive oxygen species (ROS), but both exhibited significant hematological toxicity. Aim of this study is to clarify anti‐tumor activity, safety and pharmacokinetics of GTN057, an optimization compound of KQN in vivo. Methods: ICR/SCID xenograft model of KMS11, a t(4;14) translocation‐positive MM cell line, was used for in vivo study. Mice pharmacokinetics of GTN057 and the degradation products were analyzed by LC‐MS/MS. Results: Herein, our in vitro experiments revealed that GTN057 is much less toxic to normal hematopoietic cells, induced the apoptosis of both MM cell lines andpatient samples, including those with high‐risk cytogenetic changes. A xenograft model of a high‐risk MM cell line demonstrated that GTN057 significantly delayed the tumor growth with no apparent hematological or systemic toxicities in vivo. The pathological examination of GTN057‐treated tumors in vivoshowed revealed apoptosis of MM cells and anti‐angiogenesis. In addition to the production of ROS, GTN057 inhibited the downstream signaling of c‐MET, a receptor tyrosine kinase a receptor forand hepatocyte growth factor (HGF) receptor. Thus, GTN057 is less toxic and is able tomay be a candidate drug for treating MM patients, via multifunctional mechanisms. We have also extensively studied the pharmacologyical analysis of GTN057. The metabolites of GTN057, (e.g.,such as GTN054), may also have anti‐tumorantitumor activity. Conclusion: Natural products or and their derivatives can could be good sources of antineoplastic drugs even for high‐risk cancer.

Keywords: hepatocyte growth factor; komaroviquinone; multiple myeloma; natural product; tyrosine kinase inhibition

GTN057, a komaroviquinone derivative, induced myeloma cell death in vivo. No significant hematological or systemic toxicity was observed in vivo. The pharmacokinetic and pharmacological actions of GTN057 are also studied.

As a hematopoietic tumor that is most common among elderly individuals,[1] multiple myeloma (MM) causes characteristic clinical symptoms: progressive anemia, hypercalcemia, lytic bone lesion, renal insufficiency, and occasionally amyloid light‐chain (AL) amyloidosis. The prognoses of MM patients began to improve significantly in the early 2000s with the use of newly developed drugs[2] including proteasome inhibitors,[[3]] immunomodulatory drugs (IMiDs),[5] histone‐deacetylase inhibitor,[6] and anti‐SLAMF7 (signaling lymphocytic activation molecule family member 7), and anti‐CD38 antibodies.[[7]] Proteasome inhibitors and IMiDs in particular have played a fundamental role in the treatment of MM and have significantly improved MM patients' prognoses.[9]

A number of MM patients whose MM cells have shown chromosomal abnormalities, for example, t(4;14), t(14;16), del17p, or 1q21 amplification, achieved significantly shorter survival regardless of the use of the above‐mentioned newly developed drugs and were thus categorized as having "high‐risk MM."[[10]] The Revised International staging system for multiple myeloma describes high‐risk cytogenetic abnormalities as independent risk factors for poor prognosis.[12] Since MM occurs mostly in elderly individuals, the side effects of new drugs occasionally impede the continuation of treatment. The further development of new drugs that are less toxic and more effective for high‐risk MM in clinical practice is greatly desired.

Komaroviquinone (KQN) is a natural product extracted from Dracocephalum komarovii, a perennial semi‐shrub native to central Asia. KQN shows anti‐protozoal activities against the organism that causes Chagas disease, that is, Trypanosoma cruzi.[13] KQN killed T. cruzi via its reduction by TcOYE (T. cruzi old yellow enzyme) and its production of reactive oxygen species (ROS).[14] Several new anti‐neoplastic drugs were isolated from natural products, especially those with anti‐protozoal activity,[15] and we thus constructed a chemical library for KQN derivatives and screened the derivatives for the inhibition of the growth of MM cells that have high‐risk cytogenetic alterations.[[16]] By conducting a structure–activity relationship study, we found that the hydroquinone moiety is important for the potent antitumor effects of KQN derivatives.[17] We observed that GTN024, a benzoquinone derivative of KQN, had high antitumor activities in MM cells. GTN024 also induced apoptosis in MM cells in vivo by producing ROS.[18] However, GTN024 exhibited significant hematological toxicity in vitro, which hampers the clinical applications of this compound.[17] To identify less‐toxic compounds, we continued our screening of the KQN‐derivatives library and observed that the KQN derivative GTN057 has an anti‐myeloma effect with negligible hematological toxicity in a colony assay (compound 18 in ref. ([17])). Thus, GTN057 is a compound that merits further pharmacological study in vivo.

We speculated that GTN057 would have antitumor activities in MM cells in vivo. We conducted the present investigation to test this hypothesis, and we also investigated the anti‐MM mechanisms, safety, and pharmacokinetics of GTN057 in mice.

Prof. Takemi Otsuki (Kawasaki Medical School, Kurashiki, Japan) kindly provided the cell lines KMM1, KMS11, KMS21, KMS26, KMS27, KMS28, and KMS34 from Japanese patients with MM.[19] The MUM24 cells were obtained from a thalidomide‐resistant MM patient, and the authentication was confirmed by an STR (short tandem repeat)‐PCR (polymerase chain reaction) analysis (Japan National Institute of Biomedical Innovation, Osaka).[20] All of the cell lines were cultured in RPMI1640 medium (Sigma‐Aldrich) containing 10% FBS (fetal bovine serum) (Gibco, Life Technologies) and 1% penicillin–streptomycin (Pen‐Strep, Gibco). Chromosomal abnormalities such as del(17p13.1) and t(4;14) were examined by a FISH (fluorescence in situ hybridization) analysis (LSI Medience, Tokyo). We have confirmed that del(17p13.1) is positive in the cell lines MUM24, KMS26, KMS28, KMS34, KMM1, and KMS11.[20] In addition, t(4;14) translocation was observed in the MUM24, KMS26, KMS28, KMS34, and KMS11 cells[[19]] (see also Intini et al., Br J Haematol. 2004;126:437–9). The cell line A549 was obtained from the JCRB (Japanese Cancer Research Resources Bank) in 2002 and cultured in Eagle's MEM medium with 10% FBS and 1% penicillin–streptomycin (Gibco).

GTN057 was synthesized as described.[[13], [16]] Briefly, GTN054 was synthesized from komaroviquinone by palladium catalyzed hydrogenation (Figure 1A). The acetylation of GTN054 by acetic anhydride in the presence of triethylamine gave GTN057. The structure of GTN057 was revealed by 1H NMR (nuclear magnetic resonance), 13C NMR, IR (infrared) spectrometry, and mass spectrometry (see ref. [17] for details). For the in vitro investigation, phosphate‐buffered saline (PBS) containing 1% Tween® 80 (Otsuka Pharmaceuticals, Tokyo) and 10% dimethylsulfoxide (DMSO) was used as a solvent for GT057 dilution. For the in vivo administration, GTN057 was diluted in 1% Tween® 80% + 10% sodium dodecyl sulfate (SDS) in saline.

Mononuclear cells were obtained from the bone marrow aspirates of MM patients treated at Tokyo Saiseikai Central Hospital, by centrifugation with Lymphoprep™ (Axis‐Shield). The mononuclear cells were then labeled with CD138 MicroBeads from Miltenyi Biotec (Bergisch Gladbach). The labeled CD138‐positive cells were magnetically purified using MACS Columns (Miltenyi Biotec).

For the trypan blue exclusion assay, 2 × 105 cells/mL of each MM cell line were incubated with various concentrations of GTN057 (0–20 μM). The cells were cultured in triplicate plates, and viable cells were evaluated by a trypan blue exclusion assay (Trypan Blue Stain 0.4%, Gibco) using an automatic cell counter (TC20™, Bio‐Rad).

For the 3‐[4,5]‐2,5‐diphenyltetrazolium bromide (MTT) assay (Roche Diagnostics), 6 × 104 cells/mL of CD138‐positive bone marrow cells were cultured in 96‐well plates in triplicate with various concentrations of GTN057 for 48 h. The cell viability was examined by MTT dye absorbance in accord with the manufacturer's manual. We obtained the IC50 values by drawing approximate straight lines in semilogarithmic growth curves, using Microsoft Excel.

First, 2 × 105 cells/mL of MUM24 cells were incubated with or without 6 mM NAC (N‐acetyl cysteine; Sigma‐Aldrich) or 3 mM GSH (glutathione; Sigma‐Aldrich) for 2 h. Next, 20 μM GTN057 was added and cultured for an additional 72 h. Cells were incubated with 1 μM CM‐H2DCFDA (i.e., 5‐(and‐6)‐chloromethyl‐2′, 7′‐dichlorodihydrofluorescein diacetate) for 30 min. For the evaluation of ROS production, the cells were stained with CM‐H2DCFDA and analyzed using a FACS (fluorescence‐activated cell sorting) system (LSR II, BD Biosciences). For the assessment of growth recovery by NAC or GSH, 2 × 105 cells/mL of MUM24 cells were also treated with 20 μM GTN057 in the presence or absence of 6 mM NAC or 3 mM GSH as described above. We conducted a trypan blue exclusion test to determine the numbers of viable cells.

MM cells (4 × 105 cells/mL) were incubated with 0–20 μM GTN057 for 12 h. The cells were then co‐cultured with 0–50 ng/mL of hepatocyte growth factor (HGF) (PeproTech, London) or 0–20 ng/mL of fibroblast growth factor (FGF)‐2 (PeproTech) with 1 unit/mL of heparin in RPMI containing 4% dialyzed FCS (Gibco) for the KMS11 and KMS34 cells and 10% dialyzed FCS for the KMS21 cells for 15 min at RT (room temperature). A549 cells were co‐cultured with 0–50 ng/mL of epidermal growth factor (EGF) (PeproTech) in Eagle's MEM containing 4% dialyzed FCS for 15 min at RT. The cells were lysed in 1% NP‐40 buffer containing 50 mM Tris/HCl, pH 8.0, 1% NP‐40, 150 mM NaCl, 1 mM PMSF, 1 mM Na3VO4, 20 mM NaF, 2 mM Na4P2O7, and protease inhibitors (Complete Protease Inhibitor Mixture, Roche Diagnostics, Mannheim, Germany). After centrifugation of the lysates (15,000 rpm for 10 min), the supernatants were saved and stored at −80°C until use.

The protein concentrations were measured using a BCA Protein Assay Kit (Thermo Fisher Scientific). The lysates were diluted with 3 × Laemmli's sample buffer (0.25 M Tris/HCl, pH 6.8, 4% SDS, 0.006% bromophenol blue, and 6% 2‐mercaptoethanol) and boiled for 5 min. The lysates were then subjected to 10% SDS‐PAGE (sodium dodecyl sulfate‐polyacrylamide gel electrophoresis) and transferred to PVDF (polyvinylidene fluoride) membranes. Five‐percent skim milk was used to block the membranes overnight at 4°C, and the membranes were then subjected to Western blotting.

Antibodies against phospho‐STAT3 (Tyr705), phospho‐Akt (Ser473), p44/42 MAPK (Erk 1/2), phosphor‐p44/42 MAPK (Thr202/Tyr 204), and phospho‐c‐Met (Tyr1234/1235) (all from Cell Signaling Technology, Danvers, MA) or STAT3, Akt, c‐Met, and β‐actin (all from Santa Cruz Biotechnology, Santa Cruz, CA) or phospho‐c‐Met (Tyr1230/1234/1235) (Biosource, Camarillo, CA for KMS34 cells in Figure 3A) were each used at 1:1000 dilution. The second antigen‐antibodies, that is, HRP (horseradish peroxidase)‐conjugated anti‐rabbit or anti‐mouse Ig antibody, were each used at 1:5000 dilution. Details regarding the antibodies used for Western blotting are provided in Table S1. The Western blotting signals were detected by enhanced ECL chemiluminescence (Amersham). The signal intensity of each band was quantified densitometrically with ImageJ software (ver. 1.48, U.S. National Institutes of Health).

This study was performed in strict compliance with Keio University's Institutional Guidelines on Animal Experimentation to minimize the animals' suffering. All animal experiments (including the toxicity assay, the in vivo tumor growth assay, and the pharmacokinetic study) were prospectively approved by Keio University's Institutional Animal Care and Use Committee (no. 12067‐(5)). All mice were kept in a specific pathogen‐free (SPF) space in which pathogen monitoring was carried out every 6 months. As humane endpoints in in vivo tumor growth assay, if the inoculated tumor size reached 3000 mm3, the mouse was sacrificed by isoflurane treatment or cervical dislocation. In our experiments including the toxicity assessment, in vivo tumor growth assay, or pharmacokinetic study, all mice were euthanized by isoflurane treatment or cervical dislocation. No mouse died of any other reason. Nine male 5‐week‐old ICR mice were used for the toxicity assessment. Twelve male 5‐week‐old ICR/SCID mice were used for the in vivo xenograft model and histopathologic investigation. The pharmacokinetic study was performed using three male 5‐week‐old ICR mice.

For the toxicity assessment of GTN057 in vivo, we intraperitoneally (i.p.)‐injected male 5‐week‐old ICR mice (CLEA, Tokyo) with GTN057 (0, 50, or 100 mg/kg) dissolved in 1% Tween® 80 + 10% DMSO in saline on 2 consecutive days of every 3‐day period for 2 weeks. The mice were weighed every week. Peripheral blood was collected with a heparinized hematocrit tube (Terumo, Tokyo) from the tail vein of each mouse every week and was stained with Türk's solution (Merck). The numbers of leukocytes and neutrophils were counted using light microscopy.

For the in vivo tumor growth assay, we injected 3 × 107 cells of KMS11 cells subcutaneously into the flank of male 5‐week‐old ICR/SCID mice. When the tumors reached 100 mm3, we injected (i.p.) 20 or 100 mg/kg of GTN057 solubilized in 1% Tween® 80 + 10% DMSO in saline twice every 3 days for 2 weeks. We measured the tumor volumes every day for 2 weeks as the length × width2 × 0.52.[21]

Xenografts were excised and fixed with 10% formalin and embedded in paraffin. Five‐μm sliced sections were used for hematoxylin and eosin (H&E) staining. For the immunohistochemical evaluation, we used anti‐human cleaved poly(ADP‐ribose), polymerase (PARP), polyclonal antibody (pAb) (Cell Signaling Technology Japan, Tokyo), anti‐human Ki‐67 monoclonal antibody (mAb) (clone MIB‐1) (Dako Japan, Tokyo), anti‐human cleaved caspase‐3 (Asp175), pAb (Cell Signaling Technology Japan), and anti‐human Factor VIII mAb (Atlas Antibodies). Detailed information about these antibodies is provided in Table S1.

For the pharmacokinetic study, 0 or 100 mg/kg of GTN057 (1% Tween® 80 + 10% DMSO in saline) was injected (i.p.) into male 5‐week‐old ICR mice. Peripheral blood samples were taken using heparinized hematocrit tubes from the tail vein of each mouse at 0, 15, and 30 min and 1, 2, 4, and 24 h. The blood samples were applied to centrifugation (1100 × g, 15 min at 4°C) and we then collected the supernatant plasma. CH2CN (Wako) and MeOH (Kanto, Tokyo) for LC–MS/MS (liquid chromatography–tandem mass spectrometry) measurements were mixed in a 1:1 ratio, and verapamil hydrochloride (an internal standard substance) was dissolved to the final concentration of 120 ng/mL to prepare a deproteinizing agent. We diluted the blood samples four times with this deproteinizing agent. After 15 min on ice, each deproteinized plasma sample was centrifuged (9100 × g, 10 min, 4°C). The supernatants were collected and used for the LC–MS/MS analyses.

All compounds including GTN057 were analyzed with a mass spectrometer (AB SCIEX) coupled to an HPLC (high‐performance liquid chromatography) system (Shimadzu). The peptides were separated with the use of an Atlantis® dC18, 3‐μm, 100 Å column (Waters). The solvent composition was 15% for solution A (100 mM CH3COONH4) and 85% for solution B (CH3OH). The HPLC measurement settings were as follows. Constant flow rate: 0.2 mL/min, column temperature: 40°C. The MS settings were: spray voltage, 5500 V; turbo heater temperature, 300°C.

We used an unpaired Student's t‐test with a two‐tailed distribution to determine the significance of differences in the growth inhibition of MM cells in vitro (Figure 1B–D, Figure S1B), in the in vivo tumor growth assay (Figure 2A,B), Western blots (Figure 3), and pharmacokinetics (Figure 2B, Figure S4B) with Microsoft Excel. Probability (p)‐values <0.05 were accepted as significant.

We investigated whether GTN057 would inhibit the growth of the five MM cell lines in vitro by conducting a trypan blue exclusion assay. In each of the cell lines examined, GTN057 inhibited the cells' proliferation in a dose‐dependent manner. The IC50 (half‐maximal inhibitory concentration) values calculated for the MUM24, KMS11, KMS21, KMS26, and KMS27 cells were 5.29, 11.2, 10.1, 9.91, and 7.93 μM, respectively (Figure 1B). In the KMM1, KMS28, and KMS34 cells, 50% growth inhibition (IC50) was observed at 20.5, 24.0, and 20.1 μM, respectively. Chromosome 17 was deleted in MUM24, KMM1, KMS11, KMS26, KMS28, and KMS34 cells, and t(4;14) translocation was detected in the MUM24, KMS11, KMS26, KMS28, and KMS34 cells (see the Materials and Methods section). Thus, GTN057 was able to kill multiple lines of myeloma cells, including the MM cell lines with high‐risk cytogenetic changes.

We conducted an MTT assay to examine whether GTN057 can inhibit the growth of CD138+ cells obtained from four MM patients. As shown in Figure 1C, GTN057 inhibited the growth of the cells from each of the patients. The IC50 values were 12.1 μM in Patient 1, 16.3 μM in Patient 2, 26.3 μM in Patient 3, and 18.7 μM in Patient 4. Four patients showed high‐risk cytogenetic changes: Patient 1 had 1q21 amplification; Patient 2 had 1q21 amplification, t(4;14) and t(14;16); Patient 3 had 1q21 amplification and del 17p; and Patient 4 showed 1q21 amplification and t(11;14) and a complex karyotype change.

We also examined the combination treatment of GTN057 with bortezomib and lenalidomide, two major standard myeloma drugs. As shown in Figure 1D, GTN057 exerted an additive anti‐myeloma effect with these two drugs.

For the assessment of the toxicity of GTN057 in the mouse model, 40 or 100 mg/kg of GTN057 was injected (i.p.) to ICR mice on 2 consecutive days of each 3‐day period for 2 weeks (e.g., on days 1, 2, 4, 5, 7, 8, 10, 11, 13, and 14). When the mice were treated with either dose of GTN057, we observed neither a loss of body weight nor leukocytopenia (Figure 2A). We thus considered GTN057 to be a highly safe compound, with 100 mg/kg or higher as the maximal tolerated dose in the animal experiments at the injection schedule of 2 days of every 3‐day period.

For the determination of the anti‐multiple myeloma activities of GTN057 in vivo, we injected (i.p.) 100 mg/kg per day of GTN057 into KMS11‐xenografted SCID mice. As illustrated in Figure 2B, GTN057 significantly inhibited the MM cells' growth in the KMS11‐xenografted mice compared to the controls. After 2 weeks of treatment, the average tumor volume of the GTN057‐treated mice was significantly smaller than that of the control mice (1.76 ± 0.36 cm3 vs. 0.71 ± 0.17 cm3, p = 0.00016). The average tumor weight was also significantly lower in the GTN057‐treated mice (1.70 ± 0.34 g vs. 0.98 ± 0.16 g, p = 0.0015). We performed LC–MS/MS to determine the plasma concentrations of GTN057 after a single i.p. injection (100 mg/kg) in ICR mice. The Cmax was 790.1 ng/mL (1.95 μM), the tmax was 30 min, and the plasma disappearance half‐time was 110 min (Figure 2B). At 24 h post‐injection, the injected GTN057 had almost completely disappeared from the blood.

In our pathological examinations of the GTN057‐treated xenografts, the H&E staining revealed a chromatin aggregation in the GTN057‐treated tumor cells' nuclei (Figure 2C). In addition, the immunohistochemical examinations revealed significantly increased numbers of cleaved‐caspase‐3‐positive and PARP‐positive cells in the GTN057‐treated tumors. By contrast, GTN057 treatment weakened the staining for the cell proliferation marker MIB‐1/Ki67 (Figure 2C). These pathology results suggested that GTN057 induced apoptosis and growth inhibition in MM cells in vivo. We also observed that the number of Factor VIII‐positive epithelial vessel cells was decreased in the GTN057‐treated xenografts, suggesting that GTN057 could inhibit tumor angiogenesis.

KQN was reported to show anti‐protozoal activity via the production of ROS, mediated by protozoal old yellow enzyme.[[13]] In our previous study, KQN and its derivative GTN024 produced ROS in MM cells, which was suspected to be the antitumor mechanism of this derivative.[[17]] We thus investigated whether GTN057 can also produce ROS and induce ROS‐dependent MM cell death in the present study, and we observed that GTN057 also produced ROS in MM cells (Figure S1). However, the MM cells' growth inhibition and apoptosis induced by GTN057 were not significantly abrogated by co‐culturing with anti‐oxidants (Figure S1). We thus speculated that an additional anti‐myeloma mechanism of GTN057 exists.

To clarify the mechanisms that underlie the functions of GTN057, we attempted to elucidate GTN057's effect on intracellular growth signaling. We focused on the phosphorylations of ERK1/2, STAT3, and Akt1, which are known to be activated by various growth factors and cytokines. Since we previously reported the growth‐promoting effect of HGF in MM cells, in the present study we investigated whether GTN057 regulated the phosphorylations of ERK1/2, STAT3, and Akt1 in MM cells when they were stimulated by HGF or other growth factors.[21] As shown in Figure 3, the phosphorylations of ERK1/2, STAT3, and Akt1 were enhanced when KMS11, KMS21, and KMS34 cells were treated with HGF. The addition of GTN057 diminished the degrees of the phosphorylation of ERK1/2, STAT3, and Akt1.

In addition, anti‐phospho c‐MET antibody, which reacts with ERK1/2 and STAT3‐binding site y1234 y1235, successfully detected the phosphorylation of c‐MET/HGF receptor induced by HGF treatment. Co‐culture with GTN057 and HGF abolished the phosphorylation at y1234 y1235 of the c‐MET protein (Figure 3A). The basal phosphorylation level of c‐MET and the downstream signal transducers varied in myeloma cells. One of the reasons for this variation is the autocrine activation of c‐MET in KMS11 and KMS34 cells, as our previous observation showed that these two cell lines produce HGF.[21] In the present study we also performed a Western blot analysis using various concentrations of GTN057 in KMS34 cells to examine the inhibition of c‐MET and the downstream signaling (Figure 3B). The 50% inhibition of the phosphorylations of c‐MET and ERK1/2 by GTN057 was observed at 35.9 μM and 21.3 μM, respectively (Figure 3B), and these results are comparable with the IC50 value for growth inhibition, 20.1 μM (Figure 1B). However, GTN057 did not clearly change the ERK1/2, STAT3, or AKT1 phosphorylation levels in KMS11 MM cells, in which FGFR1 and FGFR3 were expressed, when the cells were stimulated by FGF‐2 (Figure S2). In the A549 lung cancer cell line (in which EGF receptors are overexpressed), the phosphorylation levels of ERK1/2, STAT3, and AKT1 were not altered by GTN057 when these cell lines were activated by epidermal growth factor (EGF) stimulation (Figure S2). GTN057 thus preferentially inhibited HGF signaling.

Several anti‐neoplastic drugs have been isolated and developed from products that are present in plants. For example, a derivative of camptothecin, that is, irinotecan, inhibits topoisomerase I and was isolated from Camptotheca acuminata, a deciduous tree.[22] Etoposide (VP16) also inhibits topoisomerase II and is a derivative of podophyllotoxin, which is an ingredient of the perennial plant Podophyllum peltatum.[23] Paclitaxel, which depolymerizes and consequently stabilizes microtubules, was isolated from the tree Taxus brevifolia.[24] Vincristine and paclitaxel inhibit tubulin polymerization and were isolated from the flowering plants Vinca rosea (Catharunthus roseus; niche niche sou) and Taxus brevifolia, respectively.[[25]] Doxorubicin and mitomycin, both of which target DNA, were derived from Streptomyces species.[[27]]

In recently developed cancer treatments, molecular targeted drugs are preferred due to their lower toxicity to normal cells. Some molecular targeted drugs have also been developed from products found in nature. For example, trichostatin A, an extract from Streptomyces hygroscopicus, contributed to the development of vorinostat, which inhibits histone deacetylase activity.[[29]] ATRA (all‐trans retinoic acid), which is used for the differentiation induction therapy of APL (acute promyelocytic leukemia), is a biologically active form of vitamin A.[31] Natural products are thus expected to continue to be a source of lead compounds for novel antineoplastic drugs. In the present study, we chose a novel natural compound, KQN, as a lead compound that could act against high‐risk MM.

KQN and KQN‐related compounds were successfully synthesized by Suto and colleagues.[[13], [17]] We reported that GTN024 effectively induced myeloma cell‐death; however, the toxicity of GTN024 to normal hematopoietic cells was of serious concern.[18] It was reported that GTN057 did not inhibit the colony formation of bone marrow cells from mice.[17] As described in the Results section, when mice were injected with 100 mg/kg of GTN057 on 2 days of each 3‐day period over a 2‐week period, no significant neutropenia or systemic toxicity was observed, indicating the low toxicity of this compound.

Our present findings demonstrated that GTN057 significantly inhibited the proliferation of both the five MM cell lines and the CD138‐positive cells obtained from the bone marrow of patients with MM. Our mouse xenograft model showed that significant anti‐MM effects occurred following the injection schedule of 2 days of each 3‐day period without significant toxicities. The histopathological examination demonstrated that GTN057 treatment resulted in the apoptosis of xenografted KMS11 cells, which have t(4;14) translocation and TP53 gene deletion (Figure 2). Thus, GTN057 is a promising candidate molecule for overcoming high‐risk MM.

It has also been shown that c‐MET gene knock‐down in myeloma cells increased the cells' susceptibility to bortezomib and doxorubicin.[[32]] Thus, sensitization to existing anti‐myeloma drugs might also be an anti‐myeloma effect of GTN057. Despite the significant anti‐myeloma effect of GTN057 in the present xenograft model, the Cmax value of GTN057 was lower than the IC50 values for the MM cells (Figure 2) when mice were injected with GTN057. One of the possible reasons for this result is that the degradation products of GTN057 also decrease myeloma cells' proliferation in vitro.[17] Our preliminary data demonstrated that GTN057 (m/z 405 in Figure S3) was immediately dehydrated (m/z 387) or deacetylated (m/z 363) in plasma into several compounds (Figure S3). A deacetylated product of GTN057, that is, GTN054 (m/z 363) (Figure 1A), compound 17 in ref. ([17]) showed potent antiproliferative activity in our previous study, in which the IC50 of MUM24 cells was 6.13 μM.[17] By conducting an ion scan analysis, we also detected and analyzed putative peaks of GTN057 and the metabolites in LC–MS/MS (Figure S3). The time course changes of the plasma concentrations of GTN054 in mice are illustrated in Figure S4. We speculate that further dehydrated compounds of GTN054 (e.g., m/z 345 in Figure S3) may also have antitumor activity. We recognized that GTN057 per se possesses anti‐myeloma activity and also works as a prodrug. Thus, GTN057 exerted significant antiproliferative activity with a low Cmax and a short half‐life in the mouse xenograft model.

The production of ROS, which is a major anti‐myeloma activity of GTN024,[18] is not the only antiproliferative mechanism of GTN057; we observed herein that GTN057 inhibited HGF receptor/c‐Met in addition to the downstream signaling mediated by ERK1/2, STAT3, and AKT1. The exact molecular mechanisms underlying the inhibition of c‐Met by GTN057 are not yet known. To reveal the direct binding of GTN057 to c‐Met, biophysical experiments such as a three‐dimensional structure analysis are needed for future study. The activation of the HGF/c‐Met pathway has been elucidated in various types of cancer including gastric cancer, non‐small lung cancer, and hepatic cancer.[34] The HGF/c‐Met signal pathway is also known to have roles in angiogenesis and the invasion and metastasis of tumors, and it is thought to be a promising treatment target.[[35]] Investigations of multiple myeloma have shown that (i) both HGF and c‐Met signaling are involved in the growth of MM cells, and (ii) the plasma concentration of HGF is related to disease progression and/or prognosis.[[21], [37]] It was also suggested that an HGF/c‐Met signal is involved in bone marrow angiogenesis and disease activity in MM.[[21], [39]]

In our present histopathological observations using the mouse model, the number of Factor VIII‐positive vascular endothelial cells in the GTN057‐treated tumors decreased, suggesting that GTN057 could inhibit tumor angiogenesis. Our earlier study showed that an antagonist against HGF, that is, NK4, also inhibited HGF/c‐MET signaling and anti‐angiogenic activity in a mouse MM xenograft model, indicating the growth dependency of myeloma cells on HGF/c‐MET.[21] It was also reported that HGF inhibited bone morphogenetic protein (BMP), which is an important factor for the differentiation of osteoblasts.[40] This suggested that GTN057 might improve bone lesions in MM patients by inhibiting the HGF signaling pathway.

In conclusion, we screened a KQN‐derivative library considering the structure‐antitumor activity relationship, and we identified GTN057, which is less toxic to normal tissues and induced MM cells' apoptosis in vivo via multifunctional mechanisms including tyrosine kinase inhibition, ROS production, and anti‐angiogenesis.

Mikio Okayama: Investigation (equal). Kota Fujimori: Investigation (equal); methodology (equal); writing – original draft (supporting). Mariko Sato: Investigation (equal); methodology (equal). Koichi Samata: Investigation (equal). Koki Kurita: Investigation (equal); methodology (equal). Hiromu Sugiyama: Investigation (equal); methodology (equal). Yutaka Suto: Conceptualization (equal); data curation (equal); resources (lead); writing – original draft (equal). Genji Iwasaki: Resources (equal). Taketo Yamada: Investigation (equal); methodology (equal). Fumiyuki Kiuchi: Conceptualization (equal); data curation (equal); writing – original draft (equal). Daiju Ichikawa: Software (equal). Maiko Matsushita: Writing – original draft (equal). Maki Hirao: Resources (equal). Hisako Kunieda: Resources (equal). Kohei Yamazaki: Resources (equal). Yutaka Hattori: Conceptualization (lead); data curation (lead); formal analysis (lead); funding acquisition (lead); investigation (equal); methodology (lead); project administration (lead); resources (supporting); software (equal); supervision (lead); validation (lead); visualization (lead); writing – original draft (lead).

We appreciate Professor Takemi Otsuki for providing us with MM cell lines, Shotaro Kitabatake, Shogo Mori, and Koya Kakimoto for their contribution to the screening and the growth inhibition assay and Saori Nishiyama for performing pharmacokinetics. We are also grateful for the helpful discussions with and experimental suggestion by Mr. Shinobu Kudoh regarding the pharmacokinetic study of GTN057. This work was supported by Grants‐in‐Aid for Scientific Research (#20K08763 and #17K09940 to Y.H.) and by a grant from the Private University Strategic Research Base Development Program, MEXT (Y.H.). The study was also supported by grants‐in‐aid from the Translational Research Network Program, AMED, Japan (#15lm0103010j0002 and #16lm0103010j0003, to Y.H.) and a Japanese Society of Hematology Research Grant (Y.H.).

The following potential conflicts of interests exist: Y.H. has received research grants from Takeda Pharmaceutical Co., MSD, Astellas Pharma, Daiichi Sankyo Co., Pfizer, Ono Pharmaceutical Co., and AbbVie GK.

The data obtained in this study are available from the corresponding author upon reasonable request.

The Ethics Committee of Tokyo Saiseikai Central Hospital (No. 28–66) and the Faculty of Pharmacy, Keio University (No. 220518–5) approved the use of bone marrow samples from myeloma patients. The study conforms to the provisions of the Declaration of Helsinki. All of the patients provided written informed consent for their samples to be used.

GRAPH: Table S1.