Inhibition of Importin β1 Augments the Anticancer Effect of Agonistic Anti-Death Receptor 5 Antibody in TRAIL-resistant Tumor Cells

TNF-related apoptosis-inducing ligand (TRAIL) is a type II transmembrane protein belonging to the proapoptotic TNF family of molecules. In humans, four transmembrane TRAIL-specific receptors have been identified: two agonistic death receptors, DR4 and DR5, which possess a cytoplasmic death domain (DD), and two antagonistic decoy receptors, DcR1 and DcR2, which do not contain a DD (1–4). The TRAIL homotrimer binds to DR4 and DR5 on the cell surface and induces death receptor trimerization, which results in recruitment of Fas-associated death domain protein (FADD) and initiation of signal transduction through intracellular signaling machinery with various regulatory factors resulting in the activation of effector caspases (1–4). Because it can induce cell death in various types of tumor and transformed cells without apparent damage to normal cells, TRAIL has a crucial role in cancer elimination. For this reason, recombinant TRAIL (rTRAIL) and agonistic antibodies to DR4 and DR5 have been developed and tested in animal models (5, 6) and clinical trials are currently underway (7–11). However, some tumor cells have low sensitivity to TRAIL and as a result, clinical trials of combination therapy with other chemotherapeutic agents are being conducted (8, 9, 11, 12). Several reagents have been reported to upregulate cell surface expression of DR4 and DR5 cell and sensitize resistant tumor cells to cell death (13–17). Moreover, the correlation between the mislocalization of death receptors and TRAIL resistance or clinical prognosis was recently examined (12, 13, 15, 16, 18). Diminished surface expression of DR4 and DR5 appears to render tumors resistant to the targeted therapies, regardless of the expression levels of relevant death signaling molecules (11). Thus, upregulation of DR4 and DR5 on the cell surface of tumor cells is a straightforward strategy to improve the efficacy of TRAIL and anti-DR4/DR5 therapies.

Previously, we found that human DR5 (hDR5) is localized in the nucleus after transport by importin β1 in TRAIL-resistant human tumor cell lines. Furthermore, knockdown of importin β1 resulted in translocation of DR5, but not DR4, from the nucleus to the cell surface membrane and increased TRAIL/DR5-mediated cell death (19). Here, we investigated the effect of inhibition of importin β1 on the therapeutic efficacy of agonistic anti-hDR5 mAb.

Rabbit anti-human importin β1 was purchased from Santa Cruz Biotechnology, control rabbit IgG was from Cell Signaling Technology, mouse anti-human tubulin was from Sigma-Aldrich, horseradish peroxidase (HRP)-conjugated sheep anti-mouse IgG from GE Healthcare, and HRP-conjugated goat anti-rabbit IgG and biotin-conjugated goat anti-mouse IgG were from Dako Cytomation. Mouse anti-human DR5 mAb (clone HS201) and recombinant TRAIL (rTRAIL) were from Alexis Biochemicals, control mouse IgG1 (mIgG1) was from eBiosciences, and phycoerythrin (PE)-conjugated goat anti-mouse IgG F(ab′)₂, Alexa 488–conjugated wheat germ agglutinin (WGA) and Alexa 594–conjugated streptavidin were from Invitrogen. Avidin biotin blocking kit and VECTASHIELD Mounting Medium with DAPI were obtained from Vector Laboratories. CS-1008, a humanized agonistic anti-hDR5 mAb was kindly provided by the Daiichi Sankyo Company Ltd. (20). Human IgG (hIgG), puromycin, importazole, and protease inhibitor cocktail were obtained from Sigma-Aldrich, and doxycycline was from Clontech.

HeLa and HepG2 cells were obtained from ATCC and maintained in complete DMEM (Sigma-Aldrich), as described previously (19). HEK293T cells were obtained from ATCC and cultured in high glucose (hGlc)-DMEM, Hyclone (Takara) containing 10% tetracycline-free (Tet-free) \FBS (Clontech) and 100 U/mL of penicillin and 100 μg/mL of streptomycin (antibiotics: AB) (Thermo Fisher Scientific) for transfection, and were cultured in hGlc-DMEM containing 6 mmol/L glutamine, 5% Tet-free FBS, and AB for lentiviral particle production.

Doxycycline-inducible lentiviral vectors TRIPZ carrying importin β1 shRNA clones V3THS_353236 (designated B) and V3THS_353240 (designated C), nonsilencing-verified negative TRIPZ clone RHS4743 (designated control), and transduction kit for lentiviral shRNA were purchased from Thermo Fisher Scientific. Vector plasmid DNA was transfected into HEK293T cells, and viral particles were prepared using the Trans-Lentiviral packaging kit (Thermo Fisher Scientific) according to manufacturer's instructions. One day after transfection, medium was changed to hGlc-DMEM containing 5% Tet-free FBS, 6 mmol/L l-glutamine, 1 mmol/L sodium pyruvate and antibiotics, and harvested at 37°C for 48 hours. Lentiviral particle-containing supernatants were collected by centrifugation, filtered with a 0.22 μm Steriflip-GP filter (Millipore), and viral particles were concentrated using a Lenti-X Concentrator (Clontech). Particles were the suspended in DMEM without FBS and antibiotics, and transduced into HeLa cells, which were then incubated in DMEM in the absence of FBS and antibiotics at 37°C for 8 hours. The transduction mixture was then replaced with DMEM containing 10% Tet-free FBS, and transduced cells were incubated for an additional 48 hours. Infected cells were selected by 3 μg/mL of puromycin, and TurboRFP⁺ cells were sorted by using a FACS Aria (BD Biosciences) after incubation with 1 μg/mL of doxycycline. Colonies were then seeded separately in a 96-well plate to establish independent clones, which were maintained with puromycin. One clone was selected from cells transduced with importin β1 shRNA designated B or C (importin β1 shRNA-#B1 and importin β1 shRNA-#C14) and used for subsequent experiments. These clones were subjected to Western blot analysis for importin β1 and tubulin.

Transduced HeLa cells (8 × 10⁴) were seeded into 6-well plate (Corning) preincubated in the presence or absence of 1 μg/mL of doxycycline for 72 hours at 37°C, and lysed with RIPA buffer [150 mmol/L NaCl, 50 mmol/L Tris-Cl (pH 8.0), 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS] containing protease inhibitor cocktail (Sigma-Aldrich). Protein concentration was measured using the BCA Protein Assay Kit (Thermo Fisher Scientific). The lysates (1 μg) were separated in 7.5% SDS-PAGE under reducing conditions and transferred onto polyvinylidene difluoride membranes (Millipore). The membranes were then probed using rabbit anti-human importin β1 and HRP-conjugated goat anti-rabbit immunoglobulins for the detection of importin β1, and mouse anti-α-tubulin and HRP-conjugated sheep anti-mouse IgG for α-tubulin detection, and subsequently analyzed as described previously (19).

Transduced HeLa cells (8 × 10⁴) were seeded into 6-well plates and preincubated in the presence or absence of 1 μg/mL of doxycycline containing puromycin for 72 hours at 37°C. These cells were collected and stained with mouse anti-human DR5 mAb or isotype-matched mIgG1, followed by FITC-conjugated secondary antibody. TurboRFP⁺ living cells and FITC⁺ cells were analyzed on a FACScan (BD Biosciences). For cellular (cell surface and intracellular) staining, cells were fixed and permeabilized with a Foxp3/Transcription Factor Staining Buffer Set (eBioscience). In some experiments, HeLa cells (5 × 10⁴) or HepG2 cells (3 × 10⁵) in 6-well plate were added by 100 mmol/L importazole in DMSO to final concentration of 10, 20 or 30 μmol/L, or vehicle, and were incubated for 24 hours, then cells were permeabilized with 70% ethanol in PBS. The net mean fluorescent intensity (net MFI) was calculated as described previously (21).

A total of 2 × 10³ of shRNA-transduced HeLa cells were seeded into flat-bottomed 96-well plate (Corning), cultured in the presence or absence of 1 μg/mL doxycycline in tetracycline-free DMEM for 48 hours at 37°C, and then further incubated with 125 ng/mL of rTRAIL for 24 hours at 37°C. Culture supernatants were collected and cell viability was determined by WST assay as described previously (19).

Recombination activating gene 2- (RAG-2-) deficient (RAG-2^−/−) BALB/c mice were derived as described previously (22), and were maintained under specific pathogen–free conditions. Mice were used in accordance with the institutional guidelines and approval of the Juntendo University Animal Experimental Ethics Committee. In experiments using atelocollagen or importazole, RAG^−/− BALB/c mice were inoculated subcutaneously (s.c.) with HeLa cells (1 × 10⁶) or HepG2 cells (5 × 10⁶). Tumor size was periodically monitored using a digital caliper and calculated according to the following equation (23). Tumor size (Tumor surface area) (mm²) = (length) × (width). When tumor size reached approximately 16 mm² (∅: 4 mm), the mice were randomized into four groups and treatment started with drugs and antibodies on the indicated days. Tumor growth is shown as a percentage of original tumor size, calculated as described previously (23). The grafted tumors were dissected from some mice during the experiments, and treated with 1 mg/mL collagenase and 0.1 mg/mL DNase I (Wako) in PBS for 1 hour at 37°C. After washing with PBS, tumor cells were collected and analyzed by flow cytometry. At the end of the experiment, mice were sacrificed, tumors were excised and hematoxylin and eosin (H&E) staining was conducted for histologic examinations.

AteloGene Local Use Quick Gelation (atelocollagen; Koken Co., LTD) was used for in vivo transfection and knockdown induction (24, 25) according to the manufacturer's instructions. Briefly, AteloGene Local Use Quick Gelation was mixed with 4 mg/mL of pTRIPZ vector plasmid DNA containing control shRNA, importin β1 shRNA-B, or importin β1 shRNA-C, respectively. The mixture was rotated slowly at 4°C for 10 minutes, centrifuged for 1 minute to degas, and 100 μL of the mixture was injected subcutaneously per mice to cover the tumor lump.

HeLa cells (4 × 10⁴) or HepG2 cells (1 × 10⁵) on 24-well plates were preincubated with importazole or vehicle at 37°C for 24 hours. Then, cells were incubated with HS201 or mIgG1 at 37°C for 30 minutes followed by rTRAIL treatment (100 ng/mL) at 37°C for 2 hours. Cells were lysed with RIPA buffer containing protease inhibitors, as described in the preparation for Western blot samples, and the supernatants were subjected to protein assay. Cell lysates (0.4 μg) were diluted with 200 μL of caspase-3 assay buffer (100 mmol/L HEPES-KOH (pH 7.4), 220 mmol/L mannitol, 68 mmol/L sucrose, 2 mmol/L NaCl, 2.5 mmol/L KH₂PO₄, 0.2 mmol/L EGTA, 2 mmol/L MgCl₂, 5 mmol/L sodium pyruvate), and incubated at 37°C for 60 minutes with 100 μmol/L fluorescence substrate, DEVO-MCA (Peptide Institute). Fluorescence intensity was measured using Flex Station III (Molecular Devices) at an excitation wavelength of 380 nm and emission wavelength of 460 nm.

Tumor cells were incubated with or without of importazole at 37°C for 24 hours on a poly-l-lysine–coated 4-well chamber slide (Nalgene Nunc), rinsed with PBS, and fixed with 8% paraformaldehyde in 100 mmol/L phosphate buffer for 30 minutes at 4°C. After permeabilization with permeabilization buffer (Takara Bio Inc.), the cells were stained with Alexa 488-conjugated WGA for Golgi apparatus and anti-hDR5 mAb, followed by biotin-conjugated secondary antibody and Alexa 594–conjugated streptavidin. Cell nuclei were counterstained with DAPI, and cells were viewed using a confocal microscope, LSM510 (Zeiss), as described previously (19).

We used a two-sample Student t test for statistical analysis of flowcytometric net MFI, WST assay, and caspase-3 activity in vitro. Statistical analysis for xenograft tumor growth was performed by one-way ANOVA test using Prism software (GraphPad).

We first established stable importin β1 knockdown clone cells (importin β1 shRNA-#B1 and importin β1 shRNA-#C14) from TRAIL-resistant HeLa cells by transducing doxycycline-inducible TRIPZ lentiviral vectors containing two distinct importin β1 shRNAs (shRNA-B or shRNA-C). Expression of importin β1 in these cells was clearly inhibited after incubation with doxycycline (Fig. 1A). The doxycycline-inducible TRIPZ lentiviral vector is arranged with the sequences of importin β1 shRNA and TurboRFP in tandem, thus doxycycline-induced gene knockdown can be evaluated by TurboRFP expression. Approximately 80%–90% of transduced cells were found to express TurboRFP after doxycycline treatment (Fig. 1B). Superficial and cellular DR5 expression was compared among control shRNA-transduced cells, importin β1 shRNA-#B1 cells, importin β1 shRNA-#C14 cells, and nontransduced HeLa cells (Fig. 1C). Doxycycline pretreatment augmented DR5 expression on the cell surface of both importin β1 shRNA-#B1 and importin β1 shRNA-#C14 cells, but not on nontransduced or control shRNA–transduced cells, whereas cellular DR5 expression level was similar among all cells (Fig. 1C and D). These results suggest that shRNA-induced importin β1 knockdown augments DR5 expression on the cell surface.

To confirm whether suppression of importin β1 augments TRAIL sensitivity, we investigated TRAIL sensitivity in established importin β1 knockdown clones. As shown by both WST assay (Fig. 2A) and phase-contrast microscopy (Fig. 2B), preincubation with doxycycline significantly augmented TRAIL-induced cell death in importin β1 shRNA-#B1 and importin β1 shRNA-#C14 cells, but not in the nontransduced HeLa cells or control shRNA-transduced cells. On the other hand, incubation with rTRAIL alone induced a comparable level of cell death among these cells. Taken together, these results suggest that knockdown of importin β1 results in increased DR5 cell surface expression and augments susceptibility to TRAIL-induced cell death. These findings are consistent with our previous report (19).

To examine the consequence of importin β1 knockdown on the therapeutic effect of agonistic anti-hDR5 mAb, we established a mouse xenograft model by subcutaneous inoculation of importin β1 shRNA-#B1, importin β1 shRNA-#C14, or control shRNA HeLa cells into RAG-2^−/− mice. Intraperitoneal (i.p.) treatment with the humanized agonistic anti-hDR5 mAb, CS-1008, which was developed for clinical usage in cancer therapies (20), did not demonstrate a significant effect on the growth of control shRNA HeLa cells, even when treatment was combined with doxycycline administration (Fig. 3A and B). In contrast, importin β1 shRNA-#B1 cells exhibited significant regression in response to combination therapy with CS-1008 and doxycycline, but not the administration of CS-1008, control human IgG (hIgG), or doxycycline and control hIgG (Fig. 3A). Moreover, combination therapy with CS-1008 and doxycycline exerted a dramatic antitumor effect against importin β1 shRNA-#C14 cells compared with the other treatments, which resulted in complete rejection (Fig. 3B). Consistently, we observed a significant increase in cleaved caspase-8 and cleaved caspase-3–positive apoptotic cells in the importin β1 shRNA-#B1 and importin β1 shRNA-#C14 cell tumors treated with both doxycycline and CS-1008 compared with the tumors in the other treatment groups (Supplementary Fig. S1A and S1B). Furthermore, more than 75% of isolated tumor cells expressed TurboRFP after doxycycline administration, suggesting that doxycycline treatment induced shRNA expression in the majority of tumor cells in vivo (Fig. 3C and D). Regression of the importin β1 shRNA-#B1 tumor and complete rejection of the importin β1 shRNA-#C14 tumor (5/5) was confirmed by histologic examination (Fig. 3E).

We next tested the effect of in vivo importin β1 knockdown via the atelocollagen shRNA delivery system. To confirm in vivo transduction of shRNA, we examined live TurboRFP⁺ cells in the engrafted tumors by flow cytometry. A significant, but small, population of HeLa cells and HepG2 cells isolated from engrafted tumors treated with control shRNA/pTRIPZ or importin β1 shRNA/pTRIPZ expressed TurboRFP (Fig. 4A and B). These live TurboRFP⁺ cells transduced with importin β1 shRNA are expected to eventually undergo cell death due to the anti-DR5 mAb treatment.

When combined with the knockdown of importin β1 by shRNA-B/pTRIPZ or shRNA-C/pTRIPZ with the atelocollagen delivery system, CS-1008 administration significantly inhibited the growth of HeLa cells (Fig. 4C and D; Supplementary Fig. S2A and S2B). In contrast, tumor growth was not inhibited by delivery of control shRNA/pTRIPZ regardless of CS-1008 administration (Fig. 4C and D; Supplementary Fig. S2A and S2B). When we investigated the therapeutic effect in a xenograft model using HepG2 cells, both shRNA-B/pTRIPZ and shRNA-C/pTRIPZ–induced importin β1 knockdown followed by CS-1008 administration completely eradicated a majority of established tumors (2/3 and 3/3, respectively; Fig. 4E and F; Supplementary Fig. S2C and S2D).

These data suggest that atelocollagen effectively transduces shRNA into the tumor cells in vivo. Adverse systemic effects were not observed for any of the treatments, as assessed by the gross appearance and behavior of mice, over the duration of the experiment. Taken together, knockdown of importin β1 using the atelocollagen delivery system exerts therapeutic antitumor effects against TRAIL/DR5–resistant tumors when combined with agonistic anti-hDR5 antibody treatment.

Next, we assessed whether a small-molecule inhibitor of importin β1, importazole, augments surface DR5 expression and TRAIL/DR5–mediated tumor cell death in HeLa and HepG2 cells. Confocal microscopic analysis demonstrated that DR5 is located in the nucleus in intact cells, and incubation with importazole results in localization of DR5 to the cell surface and cytoplasm, including the Golgi apparatus, significantly (Fig. 5A; Supplementary Fig. S3A). In addition, the shape of the Golgi apparatus appeared to change, which was possibly due to the inhibition of the nuclear import of proteins (26). Furthermore, as revealed by flow cytometric analysis, preincubation of HeLa or HepG2 cells with importazole increased cell surface expression of DR5, although cellular DR5 expression did not significantly change (Fig. 5B and C). Phase-contrast microscopic analysis revealed that cell death was induced in HeLa and HepG2 cells by 100 ng/mL of rTRAIL when cells were preincubated with the indicated concentrations of importazole. Moreover, this induction of cell death was diminished by the addition of antagonistic anti-hDR5 mAb (Supplementary Fig. S3B and S3C). Consistently, caspase-3 activation by rTRAIL was significantly increased in HeLa and HepG2 cells by preincubation with importazole, and was almost completely blocked by antagonistic anti-hDR5 mAb (Fig. 5D). These results suggest the incubation with importazole results in translocation of DR5 from the nucleus to the cytoplasm and cell surface, resulting in augmentation of TRAIL/DR5–mediated apoptosis of HeLa and HepG2 cells.

Finally, we investigated the therapeutic anticancer effect of importazole, importin β inhibitor, combined with agonistic anti-hDR5 mAb treatment. The growth of HeLa cells was significantly inhibited only by combination therapy with CS-1008 and importazole (Fig. 6A and B). Consistent with the results seen following combination treatment with importin β1 shRNA and CS-1008 (Fig. 4E and F), the combination of importazole with CS-1008 drastically inhibited the growth of HepG2 cells (Fig. 6A), and complete tumor eradication was confirmed in some HepG2 tumors (2/3) by histologic analysis (Fig. 6C). Moreover, adverse systemic effects were not observed through changes in body weight (Supplementary Fig. S4), gross appearance, or behavior in any mice over the duration of the experiment. Taken together, these results suggest that importazole increases superficial DR5 expression and DR5-induced apoptosis in TRAIL/DR5–resistant tumor cells, and combination therapy of anti-hDR5 mAb and importazole exerts a therapeutic effect against TRAIL/DR5–resistant tumor cells.

DR5, death-inducing receptor for TRAIL, is selectively expressed in tumor cells making it an attractive target molecule for cancer therapy. We have previously reported that the importin β1 transport pathway critically regulates nuclear and cell surface expression of human DR5 and sensitivity to TRAIL/DR5–induced cell death (19). In this study, we demonstrate that inhibition of importin β1 by shRNA or a small-molecule inhibitor, importazole, exerts therapeutic antitumor activity against TRAIL/DR5–resistant human tumor cells when combined with agonistic anti-hDR5 mAb treatment.

Localization of DR5 to the nucleus is observed in various tumor cells (27), including human breast cancer cells (18), colorectal carcinoma cells, and pancreatic cancer cells (28). It has also been reported that nuclear localization of DR5 correlates with resistance to TRAIL-induced cell death (18, 19, 27, 28), and increased nuclear DR5 expression correlate with poor outcome in pancreatic tumors (28). Although DR5 expressed on the cell surface acts as a functional death-inducing receptor for TRAIL, nuclear DR5 has been identified to inhibit microRNA let-7 maturation, thereby promoting tumor cell proliferation in pancreatic cancer cells, breast cancer cells, and colorectal carcinoma cells (28). Thus, targeting the transport of DR5 from the nucleus could be an interesting approach to treat some cancers.

Importin β1, on the other hand, is known to be overexpressed in cervical cancer (29), advanced prostate cancer (30), ovarian cancer, breast cancer, and several transformed cells compared with normal cells (31). Suppression of importin β1 has been reported to result in mitotic arrest and activation of the intrinsic apoptotic pathway in cervical cancer cells (32), the inhibition of proliferation in hepatocellular carcinoma (33), and the suppression of prostate tumor growth in vivo (30). Thus, the inhibition of importin β1 has been previously considered as a possible anticancer therapeutic strategy (31, 34). In this study, doxycycline and hIgG treatment inhibited tumor growth of importin β1 shRNA-#C14 cells (Fig. 3B, right), but not importin β1 shRNA-#B1 cells, compared with hIgG treatment when examined by Student t test. As presented in Supplementary Fig. S5, importin β1 expression in #C14 tumors was more substantially downregulated by doxycycline treatment when compared with that in #B1 tumors in vivo. Thus, we do not exclude the possibility that inhibition of importin β1 exerts antitumor effects by direct inhibition of tumor proliferation in some tumor cells. On the other hands, the antitumor effects of inhibition of importin β1 was dramatically augmented when combined with agonistic anti-DR5 mAb, (Fig. 3B, right). Thus, while inhibition of importin β1 can reduce tumor growth by itself, its antitumor efficacy is increased when combined with TRAIL/DR5–induced cell death therapies. It was recently reported that inhibition of importin β1 induces various positive and negative regulatory effects on death signals in glioblastoma cells (35). Inhibition of nuclear transport of molecules other than DR5 by importin β1 inhibition might also modulate the efficacy of anticancer drugs. Thus, it will be interesting to further examine the antitumor effect of inhibition of importin β1 in combination with other anticancer drugs.

Small-molecule inhibitors of importin β may be more attractive for clinical use than shRNA delivery. Importazole was identified through a FRET-based, high-throughput small-molecule screening for compounds that interfere with the interaction between RanGTP and importin β (36). A cell-permeable and reversible inhibitor, importazole, binds preferentially to importin β and specifically inhibits importin β–mediated protein import to nucleus (36). Although importazole was subcutaneously administered to mice in this study, it was recently reported that intravenous injection of nanoparticles containing importazole can inhibit tumor growth (30). Thus, identification of the most effective method to administer importazole will be required for favorable clinical application. Recently, inhibitor of nuclear import-43 (INI-43) was identified as another importin β1 inhibitor and was found to exert a cytotoxic effect on cervical and esophageal cancer cell lines, as well as to inhibit the growth of xenograft tumors (37). Moreover, ivermectin, a broad-spectrum antiparasite medication, is reportedly a specific inhibitor of importin α/β1–mediated nuclear import (38). In fact, intraperitoneal administration of ivermectin was found to suppress tumor growth in a xenograft model of glioblastoma (39). While nuclear hDR5 is known to be transported through the importin β1–mediated pathway (19), the involvement of importin α in nuclear transport of hDR5 is unclear. Thus, it would also be interesting to explore the antitumor effects of INI-43 or ivermectin in combination with agonistic anti-hDR5 mAb against TRAIL-resistant cancer. For clinical application, careful selection of appropriate strategies and/or reagents to inhibit importin β1 will be needed to achieve the most effective antitumor effects without unfavorable side-effects caused by modification of its molecular translocation.

In this study, we show the beneficial effect of inhibition of importin β1 on agonistic anti-hDR5 antibody therapy against TRAIL-resistant tumor cells. Although further studies are needed to fully elucidate the mechanism of TRAIL/DR5 resistance in various tumor cells, the suppression of importin β1 shows promise as a cotreatment strategy for TRAIL-resistant cancers.

No potential conflicts of interest were disclosed.

Conception and design: Y. Kojima, K. Takeda

Development of methodology: Y. Kojima, H. Nakano, K. Takeda

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Kojima, K. Takeda

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Kojima, K. Takeda

Writing, review, and/or revision of the manuscript: Y. Kojima, K. Takeda

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): T. Nishina, H. Nakano, K. Takeda

Study supervision: K. Okumura, K. Takeda

We deeply thank Dr. Toshiaki Ohtsuka (Daiichi Sankyo Company Ltd.) for CS-1008, Dr. Akemi Koyanagi and Dr. Tamami Sakanishi (Laboratory of Cell Biology, Research Support Center, Juntendo University Graduate School of Medicine) for technical advice and cell sorting. We also thank members of the Laboratory of proteomics and Biomolecular Science, Research Support Center, Juntendo University Graduate School of Medicine for technical assistance. This work was supported by the Grants-in-Aid for Scientific Research from Ministry of Education, Culture, Sports, Science and Technology, Japan (19K07648, 24590092, and 26460397).

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.