FBXW7 is a cell cycle regulatory gene that ubiquitinates positive cell cycle regulators such as c-Myc and cyclin E, allowing for cell cycle exit. Defects in the FBXW7 gene that lead to cell cycle re-entry and expedite the G1-S transition is thought to be one of the causes of cancer development. However, its clinical importance for breast cancer patients remains undetermined. This prompted us to investigate its expression level in breast cancer patients to establish its clinical significance. The expression level of FBXW7 mRNA was assessed in 186 cases of primary invasive breast cancer. Correlations between FBXW7 mRNA expression and clinicopathological factors, prognoses and immunohistochemical expression levels of Ki-67, FBXW7, c-Myc and cyclin E were analyzed. In vitro investigation of FBXW7 gene silencing in a breast cancer cell line was conducted. FBXW7 mRNA was expressed at significantly lower levels in patients with high histological grade and hormone receptor-negative tumors. Patients with lower FBXW7 mRNA expression had a poorer prognosis for breast cancer-specific survival than those with higher expression. A high Ki-67 labeling index and positive cyclin E protein expression were significantly correlated with lower FBXW7 mRNA expression. In vitro, silencing FBXW7 enhanced expression of c-Myc and cyclin E proteins and upregulated both cell proliferation and G1-S transition. In breast cancer, reduced FBXW7 mRNA expression may have independent prognostic potential through the enhanced function of cell cycle regulatory proteins. (Cancer Sci 2011; 102: 439–445)
Recent studies have revealed that the ubiquitin–proteasome system plays many important roles in human carcinogenesis, such as in cell cycle progression, apoptosis, signal transmission and DNA repair. Frequent deregulation of E3 ligase components in breast cancer has led to their investigation as oncogenes or tumor suppressor genes. Major examples are the amplification and overexpression of Mdm2 and Skp2, mutation of BRCA1 in familial breast cancer and, more recently discovered, mutation and loss of expression of FBXW7 (F-box and WD repeat domain-containing 7).(1,2)
FBXW7 is a component of SCF (complex of SKP1, CUL1 and F-box protein)-type ubiquitin ligases and regulates positive cell cycle regulators such as c-Myc, cyclin E, c-JUN, Aurora A and Notch, molecules commonly implicated in many cancers(3) including breast cancer.(4,5) c-Myc is a transcription factor and has a crucial role in deciding whether or not mammalian cells divide. Cyclin E coordinates S-phase entry from either G1 or quiescence (G0), and constitutive expression of cyclin E leads to genomic instability. One notable function of FBXW7 is induction of cell cycle exit (to G0-phase) by c-Myc degradation, whereas cell cycle re-entry (G0 to G1) is regulated by p27 and its F-box regulator Skp2. Therefore, altered expression of FBXW7 is hypothesized to cause carcinogenesis and cancer development.(6,7)
FBXW7 was first linked to carcinogenesis when mutations in FBXW7 were found in the breast cancer cell line SUM149PT, which has elevated levels of cyclin E.(1,8) In animal models, mice lacking FBXW7 die at approximately day 10 with multiple abnormalities, especially in vascular development,(9) and conditional knockout mice in which FBXW7 is disrupted in T cells show a relationship to carcinogenesis in solid organs, specifically, thymic hyperplasia and thymic lymphoma.(10)
The clinical significance of FBXW7 in human malignancy is underscored by the large number of recent reports.(7) Mutation analysis of FBXW7 in many primary human tumors revealed that approximately 6% of tumors have mutations in FBXW7. However, in breast cancer, FBXW7 mutation is rarely observed (<1%).(11,12) Lower expression of FBXW7 transcript levels in cancer tissues correlates with poor prognosis in glioblastoma,(13) gastric cancer(14) and colon cancer.(15) Nevertheless, few studies have reported on the relationship between FBXW7 expression level and prognoses in breast cancer.
In the present study, we investigated FBXW7 mRNA and protein expression and report a correlation between expression levels and several clinicopathological factors including prognoses. This experimental data was validated using RNA interference in a breast cancer cell line, specifically demonstrating the effects of lower expression of FBXW7 on proliferation and cell cycle transition.
Patients and Methods
Breast cancer tissues. Breast tumor specimens from 186 female patients with invasive breast carcinoma, who were treated at Kumamoto University Hospital between 2001 and 2008, were included in the present study. The patients were from a consecutive series, and no exclusion criteria were applied. Samples were snap frozen in liquid nitrogen at pretherapeutic biopsy or surgical treatment and stored at −80°C until RNA extraction. Informed consent was obtained from all patients. The ethics committee of Kumamoto University Graduate School of Medical Sciences approved the study protocol. The median age of the patients was 59 years (range, 27–93 years). Adjuvant and neoadjuvant treatment was done in accordance with the recommendations of the St Gallen international expert consensus on the primary therapy of early breast cancer.(16–19) On recurrence, patients with hormone receptor-negative tumors were treated with chemotherapy, such as anthracycline-contained regimens, taxanes, trastuzumab (for human epidermal growth factor receptor 2 [Her2] overexpression patients), capecitabine and vinorelbine. Patients with hormone receptor-positive tumors and nonvisceral metastases were treated with endocrine therapy, such as antiestrogens, aromatase inhibitors and medroxyprogesterone acetate. Patients were followed postoperatively every 3 months. The median follow-up period was 33 months (range, 4–90 months).
Immunohistochemical analysis. Histological sections (4 μm) were deparaffinized and incubated for 10 min in methanol containing 0.3% hydrogen peroxide. We used mouse monoclonal antibody against estrogen receptor α (ER α) (1D5, 1:50 dilution; Dako Japan, Tokyo, Japan) and progesteron receptor (PgR) (PgR636, 1:800; Dako Japan), and rabbit polyclonal antibody (1:200; Dako Japan) for detection of Her2, Ki-67 (MIB-1, 1:50; Dako Japan), FBXW7 (3D1, 1:200; Abnoova, Taipei, Taiwan), c-Myc (9E10, 1:800; Santa Cruz Biotechnology, Fremont, CA, USA) and cyclic E (13A3, 1:30; Neomarkers Laboratories, CA, USA). Expression using these antibodies was determined by the Histofine Simple Stain MAX-PO (Nichirei, Tokyo, Japan) method as described previously.(20) ERα and PgR status was considered positive when nuclear staining was ≥10%. Her2 immunostaining was evaluated using the same method as the HercepTest (Dako); the membrane staining was scored on a scale of 0 to 3+. Tumors with scores ≥3 or with a ≥2.2-fold increase in HER2 gene amplification as determined by fluorescence in situ hybridization were considered to be positive for Her2 overexpression. Ki-67 was scored for the percentage of nuclear staining cells out of all cancer cells in the invasive front of the tumor at ×400 high-power field (Ki-67 labeling index). For FBXW7, c-Myc and cyclin E, the complete H score was semiquantitatively calculated by summing the products of the percentage of cells stained at a given staining intensity (0–100) and the staining intensity score (0, none; 1, weak; 2, moderate; and 3, intense).
RNA isolation and real-time quantitative reverse transcription–polymerase chain reaction (RT-PCR). Total RNA was isolated from tissue specimens and treated cells using the RNeasy mini kit (Qiagen, Valencia, CA, USA). Total RNA (0.5 μg) was reverse transcribed to cDNA using ExScript RT reagent (Takara Bio Inc., Otsu, Japan), according to the manufacturer’s protocol. Each PCR was performed with 2 μL of the cDNA and 0.2 μmol/L of each probe in a LightCycler System with SYBR Premix Dimer Eraser (Takara Bio Inc.). The PCR primer sequences were as follows: for FBXW7 common isoform, forward 5′-CCACTGGGCTTGTACCATGTT-3′ and reverse 5′-CAGATGTAATTCGGCGTCGTT-3′; for glyceraldehyde-3-phosphate dehydrogenase receptor (GAPDH), forward 5′-GCACCGTCAAGGCTGAGAAC-3′ and reverse 5′- ATGGTGGTGAAGACGC-CAGT-3′. Each reaction (20 μL samples) was performed under the following conditions: initialization for 10 s at 95°C, and then 45 cycles of amplification, with 5 s at 95°C for denaturation and 20 s at 60°C for annealing and elongation. For each PCR run, a standard curve was constructed from cDNA from the T47D cell line. The level of expression of FBXW7 mRNA is given as relative copy numbers normalized against GAPDH mRNA.
Cell culture and transfections. The T47D cell line (American Type Culture Collection) was grown in RPMI 1640 medium (Gibco, Grand Island, NY, USA) supplemented with 10% heat-inactivated fetal bovine serum (Gibco) in 5% CO2 at 37°C. The cell lysate was processed for western blot analysis as described below. FBXW7-specific siRNA (Silencer Predesigned siRNA) was purchased from Applied Biosystems Japan (Tokyo, Japan), sense 5′-GGAGUAUGGUCAU-CACAAAtt-3′ and antisense 5′-UUUGUGAUGACCAUACUCCac-3′. FBXW7 siRNA was diluted in Opti-MEM I Medium without serum (Life Technologies Japan, Tokyo, Japan). The diluted siRNA was mixed with the diluted LipofectAMINE 2000 (Life Technologies Japan) and incubated for 15 min at room temperature. Diluted logarithmic growth-phase T47D cells (without antibiotics) were seeded at 1 × 104 cells/mL in a final volume of 1 mL or 100 μL in 12- or 96-well flat bottom microtiter plates, respectively. The cells were incubated for 48 or 72 h before harvesting total RNA or protein, respectively. T47D cells transfected with control siRNA (Silencer Negative Contro1 siRNA; Applied Biosystems Japan) were used as the control.
Protein extraction and western blot. The T47D cells were washed in ice-cold PBS and lysed by CelLytic M Cell Lysis/Extraction Reagent (Sigma-Aldrich Japan, Tokyo, Japan). Equal amounts of protein were fractionated via 10 concentrated READY GELS J (Bio-Rad Laboratories, Hercules, CA, USA) and transferred to nitrocellulose membranes (Bio-Rad Laboratories). The membranes were blocked and incubated overnight at 4°C with anti-c-Myc (N-262, 1:500; Santa Cruz Biotechnology) and anti-cyclin E (13A3, 1:500; Neomarkers Laboratories). Proteins were normalized to the level of β-actin protein diluted 1:2000. After blots were washed, they were incubated for 1 h using the standardized biotin–streptavidin method. Specific protein bands were detected using the enhanced chemiluminescence system (Amersham Pharmacia Biotech, Buckinghamshire, UK).
Cell growth assay. Proliferation was determined using a Cell Counting Kit-8 (Dojindo Laboratories, Kumamoto, Japan). WST-8 reagent solution (10 μL) was added to each well, after which the microplate was incubated for 90 min at 37°C. Absorbance at 450 nm was then measured using an EMax Precision Microplate Reader (Molecular Devices Japan, Tokyo, Japan).
Flow cytometric analysis of asynchronous cell cultures. Cells were harvested with tripLE (Gibco), fixed with 70% ethanol, treated with RNase A (100 U/mL) and stained with propidium iodide (50 mg/mL). Flow cytometric analysis was done with FACScalibur (Becton Dickinson, San Jose, CA, USA) and Cell Quest software (Becton Dickinson).
Statistical analysis. The nonparametric Mann–Whitney U-test or Kruskal–Wallis test was adopted for statistical analysis of associations between FBXW7 mRNA expression and clinicopathological factors and Ki-67 labeling index, c-Myc and cyclin E. Disease-free survival and overall survival curves were generated using the Kaplan–Meier method and verified by the log-rank (Mantel–Cox) test. Cox’s proportional hazards model was used for the univariate and multivariate analyses of prognostic values. Student’s t test was used to assess differences in cell proliferation of the cell lines with two groups. Statistical significance was defined as P < 0.05. JMP software version 7.0.1 for Windows (SAS Institute Japan, Tokyo, Japan) was used for statistical analysis.
Lower FBXW7 mRNA expression correlates with higher nuclear grade, ER-negative and PgR-negative tumors. The clinical characteristics for the 186 cases analyzed in the present study are summarized in Table 1. Relative expression of FBXW7 mRNA ranged from 0.043 to 588.9 with a median of 118.
Table 1. FBXW7 gene expression and clinicopathological factors
FBXW7 mRNA median (25th, 75th percentile)
*The tumor subtype was grouped according to a combination of ER, PgR and Her2 status. Luminal: (ER[+] or PgR[+], any Her2); HER2: (ER[−]/PgR[−]/Her2[+]); and Triple Negative: (ER[−]/PgR[−]/Her2[−]).
ER, estrogen receptor; Her2, human epidermal growth factor receptor 2; PgR, progesteron receptor.
10.3 (4.2, 38.7)
19.3 (2.2, 38.4)
17.9 (5.2, 42.0)
10.3 (4.0, 38.5)
Tumor size (mm)
10.9 (4.9, 36.5)
13.7 (3.9, 38.9)
12.0 (4.1, 40.7)
9.1 (4.4, 33.9)
11.5 (4.3, 32.4)
13.3 (4.3, 40.7)
13.3 (2.8, 41.6)
14.0 (5.2, 43.6)
7.4 (2.3, 30.3)
Lymphatic vessel invasion
10.9 (3.9, 77.0)
19.7 (6.1, 41.2)
Blood vessel invasion
11.3 (4.1, 37.0)
20.4 (9.1, 79.1)
7.1 (2.3, 31.2)
13.7 (5.2, 42.4)
8.8 (3.1, 32.9)
13.7 (5.2, 45.1)
13.1 (4.5, 42.7)
7.1 (3.4, 31.4)
13.5 (5.1, 42.7)
7.4 (2.6, 32.1)
6.9 (2.0, 31.1)
The level of FBXW7 mRNA expression was observed to be significantly lower in the group of patients classified as higher nuclear grade (grade 3) than in those classified as lower nuclear grade (grade 1 or 2; P = 0.024). The level of FBXW7 mRNA expression was lower in the ER-negative group than in the ER-positive group (P = 0.009). Also, FBXW7 mRNA was expressed at lower levels in the PgR-negative group than in the PgR-positive group (P = 0.027). There was no significant difference between Her2-positive and Her2-negative groups (P = 0.21; Table 1). Although there is a tendency for positive correlation between the expression levels of FBXW7 mRNA and protein for each patients (P = 0.031, R2 = 0.025; Fig. S1a), the FBXW7 protein expression levels itself had less remarkable characteristics in relation to clinicopathological factors (Table S1).
Low FBXW7 mRNA is an independent prognostic factor for breast cancer-specific survival. In the analysis of relapse-free survival, local recurrences and distant metastases were considered an event. Among 23 recurrent cases, there were 20 cases of distant metastases and three cases of local recurrences. Fourteen cases died as a result of breast cancer, which were regarded as events when analyzing breast cancer-specific survival.
To identify a clinically meaningful cutoff point for FBXW7 mRNA expression in prognostic analysis, various levels of FBXW7 mRNA expression were tested by the Kaplan–Meier method and verified by the log-rank (Mantel–Cox) test. When the cut-off point for FBXW7 mRNA level was set at 9.3, patients with low expression levels (median, 3.6; 25th percentile, 2.0; 75th percentile, 5.7; n = 80) had significantly poor breast cancer-specific survival than those with high expression levels (median, 33.0; 25th percentile, 14.9; 75th percentile, 72.0; n = 106) (P = 0.011; Fig. 1a). However, when the same cut-off point was adopted there was no significant association in relapse-free survival (P =0.34; Fig. 1b). Among 23 recurrent cases, the interval from recurrence to death was significantly shorter in low FBXW7 mRNA expression cases (n = 12; median, 0.87 months; 25th percentile, 0.54; 75th percentile, 15.3) than in high FBXW7 mRNA expression cases (n = 11; median, 3.96 months; 25th percentile, 3.05; 75th percentile, 5.42) (P = 0.024, Log-rank [Mantel–Cox] test; data not shown). For FBXW7 protein expression, effect on prognoses was weakly observed in the log-rank (Mantel–Cox) test of the Kaplan–Meier method by the best cut-off point of H-score 70 (Fig. S2).
Univariate and multivariate prognostic analysis showed FBXW7 mRNA expression to be an independent prognostic factor for breast cancer-specific survival (P = 0.033; Table 2).
Table 2. Univariate and multivariate analysis for breast cancer-specific survival (Cox proportional regression model)
Relative risk (95% CI)
Relative risk (95% CI)
CI, confidence interval; ER, estrogen receptor; Her2, human epidermal growth factor receptor 2; PgR, progesteron receptor.
Tumor size (> 20 mm)
Nuclear grade (1 or 2/3)
Relationship between FBXW7 mRNA expression and the Ki-67 labeling index, c-Myc protein expression and cyclin E protein expression in breast cancer tissues. We examined the association between FBXW7 mRNA expression and the Ki-67 labeling index, c-Myc protein expression and cyclin E protein expression in serial sections of 183 patients whose FBXW7 mRNA level had been determined. The low FBXW7 mRNA expression group showed a higher Ki-67 labeling index than the high FBXW7 mRNA expression group. In addition, there was a significant correlation between low FBXW7 mRNA expression and positive cyclin E staining (Table 3). No significant correlation was observed for c-Myc protein expression. A representative case for the low FBXW7 mRNA group showed a high Ki-67 labeling index (91.3%) and lower staining of FBXW7, positive staining for both c-Myc and cyclin E (Fig. 2a–d). This typical combination was observed in 10.2% (8 out of 78) of low FBXW7 mRNA expression cases. Meanwhile, a representative case for the high FBXW7 mRNA group showed a low Ki-67 labeling index (11.1%) and higher expression of FBXW7, negative staining for both c-Myc and cyclin E (Fig. 2e–h). This type was observed in 57.1% (60 out of 105) of high FBXW7 mRNA expression cases.
Table 3. Relationships between FBXW7 mRNA level and Ki-67 labeling index, c-Myc protein expression and cyclin E protein expression in tumors (n =183)
Low expression group (n = 78) (42.6%)
High expression group (n = 105) (57.4%)
Ki-67 index (%), median (25th, 75th percentile)
50.6 (37.3, 67.5)
30.7 (15.9, 46.2)
c-Myc H-score (mean ± SD)
17.0 ± 37.8
26.23 ± 43.9
Cyclin E H-score (mean ± SD)
24 ± 41.13
8.3 ± 27.03
Downregulation of FBXW7 mRNA expression promotes cell cycle transition and proliferation in vitro. Because FBXW7 mRNA suppression in cancer tissues indicated a poor prognosis, the protein levels of c-Myc and cyclin E (degradation targets of FBXW7) were examined to evaluate FBXW7 function in breast cancer cell lines. FBXW7 suppression analysis was performed with siRNA targeting FBXW7 (FBXW7 siRNA) using the breast cancer cell line T47D. Quantitative RT-PCR analysis confirmed that FBXW7 siRNA transfection reduced FBXW7 mRNA expression by 71% compared with control siRNA transfection (Fig. 3a).
As a result of FBXW7 knockdown, the expression levels of both c-Myc and cyclin E proteins were significantly enhanced (Fig. 3b). In addition, in a cell cycle assay, the fraction of cells in G1-phase decreased in the FBXW7 siRNA group (62.4%) compared with the control siRNA group (77.9%). The S-phase fraction also increased in the FBXW7 siRNA group (20.3%) compared with the control siRNA group (14.8%). In turn, the percentage of cells in G2-phase increased in the FBXW7 siRNA group (17.3%) compared with the control siRNA group (8.2%; Fig. 3c). In agreement with these data, the WST-8 assay showed that proliferation rates were significantly enhanced by FBXW7 siRNA transfection in a time-dependent manner (P = 0.016; Fig. 3d).
In the present study, we investigated the level of FBXW7 gene expression and its protein expression in 186 corresponding clinical cases of breast cancer. We found that FBXW7 mRNA was suppressed significantly in patients with high histological grade (P = 0.024), ER-negative (P = 0.009) and PgR-negative (P = 0.027; Table 1) tumors. This evidence supports the idea that low FBXW7 mRNA expression affects the cell cycle of tumor cells because histological grading reflects DNA synthesis and mitosis by morphological assessment of tumor cell nuclei.(21) The complementary in vitro study also showed that knockdown of FBXW7 accelerated the cell cycle and promoted proliferation (Fig. 3c,d). Recent DNA microarray profiling studies of breast cancer identified that hormone receptor negative subtypes are characterized by high expression of a cluster of proliferation genes.(22) In our study, low FBXW7 mRNA expression was observed in hormone receptor-negative subtypes: HER2 (median 74) and Triple Negative (median 69) compared with the Luminal subtype (median 135; Table 1). Furthermore, there was a significant relationship between low FBXW7 mRNA expression levels and a high Ki-67 labeling index (Table 3). Yu et al.(23) suggested that homozygous single nucleotide polymorphism in the intron 2 block of FBXW7 are associated with ER-negative tumors and higher clinical stages (stage III and metastatic phase) in breast cancer. Although there are no other reports about the detailed relationship between FBXW7 and hormone receptors, deregulation of this gene can potentially result in some of the same aggressive characteristics observed in highly proliferative breast cancer, including hormone receptor-negative subtypes. Although FBXW7 protein expression showed the tendency of positive correlation with mRNA expression levels and a similar pattern of prognostic outcomes (Figs S1,S2), its clinical significance remained limited. Regarding the regulation of FBXW7 expression, we speculate there may be epigenetic transcriptional regulation,(24) translational regulation by non-coding RNA (including micro RNA(25)) because FBXW7 mutation is reported to be rarely observed(11,12) in breast cancer.
The reduction of FBXW7 expression is associated with the dysregulation of c-Myc and cyclin E. We showed that protein expression of c-Myc and cyclin E is enhanced when FBXW7 is suppressed in a breast cancer cell line (Fig. 3b). In addition, the G1-S transition in these cells was observed by a cell cycle assay (Fig. 3c), which facilitated the proliferation rates significantly (Fig. 3d). For clinical specimens, low expression of FBXW7 mRNA was significantly correlated with cyclin E protein expression (P = 0.0017), but not with c-Myc protein expression (P = 0.97; Table 3). The combination of low FBXW7 mRNA expression and both c-Myc and cyclin E-positive staining were observed in only 4.3% (8 out of 183) of subject cases. In breast cancer, amplification and overexpression of these genes and their clinical significance have been thoroughly investigated.(4,26) Cyclin E, the maintainer of the restriction point, is overexpressed in up to 30% of breast cancers and is associated with hormone receptor-negative, high histological grade tumors and a poor prognosis.(5) c-Myc is one of the key effectors of estrogen signaling, and MYC gene amplification is reported to be associated with a poor prognosis.(4,27,28) Otherwise, c-MYC mRNA and immunohistochemical studies report their conflicting clinical significance.(29–32) In another report, there are some controversies regarding c-Myc antibodies, which are very sensitive to fixation conditions of breast cancer tissues.(33) Our heterogeneous collection of immunohistochemistry data, especially the accumulation of c-Myc, might reflect these complex routes (estrogen signaling, gene amplification, methodology of detection, etc.) rather than only by FBXW7 dysfunction itself.
In the current study, patients with low levels of FBXW7 mRNA expression had poorer prognoses than those with high expression when analyzing breast cancer-specific survival (Fig. 1a). Multivariate analysis showed that FBXW7 mRNA expression is an independent prognostic factor in breast cancer-specific survival (Table 2), but not in relapse-free survival. A meta-analysis of cyclin E expression and prognosis in breast cancer patients similarly showed that a high level of cyclin E is an independent prognostic factor for breast cancer-specific survival, but not for relapse-free survival.(34) Because a major mechanism governing the effects of c-Myc on cell cycle progression in breast cancer cells is suggested to be activation of cyclin E/Cdk2 by suppression of the Cdk inhibitor p21WAF1/Cip1,(35,36) our results on prognoses might indirectly reflect the characteristics of FBXW7 targeting proteins.
Interestingly, among 23 recurrent cases the interval from recurrence to death was significantly shorter in low FBXW7 mRNA expression cases than in high FBXW7 mRNA expression cases. Our recurrent cases have been treated with as many available anticancer drugs as their general condition permits. Putting these observations together, it may be that downregulation of FBXW7 has some effect on the sensitivity or resistance to anticancer drugs. Thus, FBXW7 status may be a potentially useful predictive marker for drug treatment, or the protein could be a therapeutic target itself. Another study analyzing the relationship between FBXW7 expression levels and therapeutic responses to anticancer drugs is warranted.
In conclusion, downregulation of FBXW7 might contribute to the malignant behavior of breast tumors. FBXW7 mRNA expression has its prognostic potential and this molecule might eventually be a novel target for the treatment of breast cancer.
The authors thank Dr. K. Kai (Division of Gene Regulation, Institute for Advanced Medical Research, Keio University School of Medicine, Tokyo) and Dr. T. Ota (Department of Oral and Maxillofacial Surgery, Graduate School of Medical Sciences, Kumamoto University, Kumamoto) for helpful comments; Y. Azakami, Y. Sonoda and S. Sakata for excellent technical support; and A. Okabe for clinical data management.
This research was supported by the Ministry of Education, Culture, Sports, Science and Technology of Japan grant-in-aid for scientific research #21591671 and the Japan Society for the Promotion of Science (Y. Yamamoto; researcher number 20398217).