A pivotal role of Krüppel-like factor 5 in regulation of cancer stem-like cells in hepatocellular carcinoma

Abstract

In hepatocellular carcinoma (HCC), there exists a highly tumorigenic subset of cells defined by high expression of CD44 and CD133 that has been reported to contain cancer stem-like cells (CSCs). Krüppel-like factor 5 (KLF5) regulates many factors involved in cell cycle, migration, inflammation, angiogenesis and stemness and has cancer-promoting effects in some cancers. While some reports have indicated that KLF5 may have important roles in regulation of CSCs, what role, if any, KLF5 plays in regulation of CSCs in HCC remains to be elucidated. Flow cytometric analysis of CD44 and CD133 in HCC cell lines revealed subpopulations of CD44^High/CD133^High and CD44^Low/CD133^Low cells. We subsequently sorted these subpopulations and identified KLF5 as a gene that is significantly upregulated in CD44^High/CD44^High cells via RNA sequencing using next generation sequencing technology. Moreover, KLF5 overexpression enriched the CD44^High/CD133^High subpopulation and, consistent with the up-regulation of CD44^High/CD133^High cells, KLF5 overexpressing cells were more resistant to anti-cancer drugs and displayed enhanced colony-formation capacity. By contrast, knock-down of KLF5 by siRNA diminished the CD44^High/CD133^High subpopulation. When KLF5 was acetylated by TGF-β1, the KLF5-mediated CD44^High/CD133^High subpopulation enrichment was abrogated. Oppositely, ectopic expression of an acetylation-deficient KLF5 mutant further increased CD44^High/CD133^High subpopulations as compared to cell expressing wild-type KLF5. These findings provide novel mechanistic insight into a pivotal role for KLF5 in the regulation of CSCs in HCC.

Keywords:

• acetylation
• cancer stem cell
• CD44
• CD133
• hepatocellular carcinoma
• KLF5

Abbreviations

5FU	=	5-fluorouracil
CDDP	=	cisplatin
CSCs	=	cancer stem-like cells
FACS	=	fluorescence-activated cell sorting
HCC	=	hepatocellular carcinoma
KLF5	=	Krüppel-like factor 5
NGS	=	next generation sequencing
TGF-β	=	transforming growth factor-β.

Introduction

Hepatocellular carcinoma (HCC) is the sixth most common cancer and the third most common cause of cancer death worldwide.¹ With the emergence of effective curative therapeutic options in HCC, including surgical resection, liver transplantation, radiofrequency ablation and trans-arterial chemoembolization, patient outcomes have dramatically improved. Unfortunately, patients with advanced HCC are often not eligible for such therapeutic interventions, resulting in maintained poor prognosis. This is particularly true of patients with extrahepatic metastases, a condition in which median survival time is estimated at 4.1 to 8.1 months.^2-4 Sorafenib is the only effective therapy for advanced HCC, however, the prognosis of advanced HCC patients treated with sorafenib is still not satisfactory.^5,6 Thus, further analysis of molecular mechanism of HCC is of critical importance to uncover novel therapeutic avenues against advanced HCC.

Cancer stem-like cells (CSCs) have been identified in many cancers including HCC.^7,8 CSCs are highly tumorigenic, metastatic and resistant to chemotherapeutic agents, resulting in dampened clinical response to therapy, tumor metastasis and recurrence, as well as deteriorated prognosis. In HCC, CD44, CD133, EpCAM, CD90 and CD24 are used to detect CSCs ^9-13 and CSCs have been shown to significantly correlate with malignant biological behavior and worse prognosis.^14-16 Although novel CSC-targeting therapeutic strategies are highly sought after, pivotal molecules and pathways regulating CSCs still remain to be elucidated and further studies on CSCs in HCC are necessary.

Kruppel-like factor 5 (KLF5) is a transcription factor containing 3 zinc finger domains and one transactivation domain. The DNA binding site of KLF5 has an acetylation site and acetylation status of KLF5 is important for its transcriptional activity. KLF5 regulates many factors related to cell cycle, angiogenesis, migration, inflammation and stemness.¹⁷ In cancers, KLF5 can act as either a tumor suppressor or an oncogene, depending on cancer cell type and context. KLF5 has been reported to have tumor-promoting effects in pancreatic cancer, breast cancer, lymphoblastic leukemia, bladder cancer and gastric cancer.^17-19 Furthermore, some reports have indicated critical roles of KLF5 in generation and maintenance of CSCs.^19-21 However, the function of KLF5 in HCC and CSCs in HCC has not been established.

In the present study, we performed RNA sequencing using next generation sequencing (NGS) technology to identify key molecules and pathways involved in CSC biology in HCC. We identify KLF5 as a pivotal factor regulating CSCs in HCC.

Results

CD44+/CD133+ cells have cancer stem-like characteristics in hepatocellular carcinoma

Yang et al. examined a relatively large number of patients with HCC and showed that CD44 and CD133 are statistically significant prognostic factors among CSC biomarkers.¹⁴ Thus, we utilized CD44 and CD133 to detect HCC CSCs in the present study. Indeed, flow cytometric analysis revealed heterogeneous populations in Huh7 and HepG2 upon staining with anti-CD44 and anti-CD133 (Fig. 1A). To characterize subpopulations of HCC cells, we then sorted cells with high expression of CD44 and CD133 (CD44+/CD133+ cells) and those with low expression of CD44 and CD133 (CD44−/CD133− cells) as shown in Fig. 1A. As expected, CD44+/CD133+ cells exhibited increased anchorage-independent growth capability and formed significantly more colonies in soft-agar than CD44−/CD133− cells (Fig. 1B). Furthermore, in both Huh7 and HepG2 cells, CD44+/CD133+ subpopulations displayed enhanced resistance to the anti-cancer drugs, CDDP and 5FU than their CD44−/CD133− counterparts (Fig. 1C). These data indicate that CD44+/CD133+ cells possess traits consistent with cancer stem-like cells in hepatocellular carcinoma.

Figure 1. CD44^High/CD133^High cells in hepatoma cell lines. (A) Hepatoma cell lines, Huh7 and HepG2, were stained with anti-CD44 and anti-CD133 and analyzed by FACS. Cells expressing high CD44 and CD133 (CD44+/CD133+ cells) and those expressing low CD44 and CD133 (CD44−/CD133− cells) were sorted by FACS Aria III as gated. (B) Anchorage independent growth was examined by soft-agar colony formation assay with the sorted CD44+/CD133+ cells and CD44−/CD133− cells. (*P < 0.05 vs. CD44−/CD133− cells) (C) Sensitivity to anti-cancer drugs, CDDP and 5FU, was analyzed by MTS assay (* P < 0.05 vs. CD44−/CD133− cells).

KLF5 is up-regulated in CD44+/CD133+ cells

To explore key molecules and pathways that may regulate CD44+/CD133+ subpopulations, we carried out RNA sequencing using NGS. The cut-off value of RPKM was set as 1.0. Interestingly, CD44+/CD133+ and CD44−/CD133− cells (Fig. 2A) displayed distinct gene expression profiles. 595 mRNAs were significantly upregulated over 1.5 fold and 465 mRNAs were down-regulated less than −1.5 fold in Huh7 cells (Fig. 2B). Among those mRNAs, we focused on KLF5. To validate the result of RNA sequencing, we performed real-time PCR. Indeed, expression of CD44 and CD133 were increased in CD44+/CD133+ cells, and consistent with the result of RNA sequence, KLF5 was significantly up-regulated in CD44+/CD133+ cells in both Huh7 and HepG2 cells (Fig. 2C).

Figure 2. KLF5 is significantly upregulated in CD44+/CD133+ cells. RNA sequence was performed using sorted Huh7 cells. Results of principal component analysis (A) and volcano plot (B). (C) Expression levels of CD44, CD133 and KLF5 were validated by real-time RT-PCR in Huh7 and HepG2 cells (* P < 0.05 vs. CD44−/CD133− cells).

KLF5 is a pivotal transcription factor in maintenance of CD44+/CD133+ cells

To analyze the functional role of KLF5 in regulation of CD44+/CD133+ subpopulations, we knocked down KLF5 using siRNA in Huh7 and HepG2 cells. Indeed, 2 independent siRNA sequences against KLF5 significantly suppressed the expression of KLF5 and the expression of CD44 was also significantly decreased by KLF5 knockdown (Fig. 3A). Consistent with the results of real-time PCR, the CD44+/CD133+ subpopulation was significantly suppressed by siRNA against KLF5 in both Huh7 and HepG2 cells (Fig. 3B), and the cells knocked down KLF5 by siRNA were significantly more sensitive to 5FU than the control cells (Fig. 3C). We next generated Huh7 cells ectopically expressing KLF5 to explore the consequences of KLF5 overexpression upon CD44+/CD133+ CSCs in HCC. Expression of KLF5 was confirmed by real-time RT-PCR (Fig. 4A) and western blot (Fig. 4B). Overexpression of KLF5 enriched the CD44+/CD133+ subpopulation (Fig. 4C). Furthermore, KLF5-overexpressing Huh7 cells displayed enhanced soft agar colony formation and resistance to CDDP and 5FU than the control cells (Fig. 4D). These results suggest that KLF5 may play critical roles in maintenance of CD44+/CD133+ cancer stem-like cells in hepatocellular carcinoma.

Figure 3. Knockdown of KLF5 decreases CD44+/CD133+ population in hepatoma cell lines. (A) Two independent sequences of siRNA against KLF5 were used to knock-down KLF5 at the concentration of 5nM. Expression levels of KLF5 and CD44 were determined by real-time RT-PCR 48 hours after siRNA transfection. (* P <0.05 vs. scramble siRNA) (B) KLF5 was knocked-down by KLF5-A siRNA at a concentration of 5nM. 48 hours after siRNA transfection, the CD44+/CD133+ subpopulation was analyzed by FACS. A histogram shows relative fold change of CD44+/CD133+ cells in Huh7 and HepG2 cells. (* P<0.05 vs. scramble siRNA) (C) KLF5 was knocked-down by KLF5-A siRNA at a concentration of 5nM. 48 hours after siRNA transfection, sensitivity to 5FU was analyzed by MTS assay (* P<0.05 vs. siKLF5).

Figure 4. Overexpression of KLF5 increases the CD44+/CD133+ subpopulation and renders cells more resistant to anti-cancer drugs in hepatoma cell lines. KLF5 was transduced by retrovirus-mediated gene transfer into Huh7 cells. Expression level of KLF5 was examined by real-time PCR (A) and western blot (B). (* P < 0.05 vs. empty) (C) The CD44+/CD133+ subpopulation was determined by FACS. A histogram shows relative fold change of CD44+/CD133+ cells in Huh7 cells. (* P < 0.05 vs. empty) (D) KLF5 overexpressing cells and empty control cells were subjected to soft-agar colony formation assay. Chemotherapeutic reagents, CDDP (10 μM) and 5FU (5 μM), were added to the overlaid medium. Histogram shows relative colony numbers (* P < 0.05 vs. empty control).

Acetylation status of KLF5 is critical for regulation of CD44+/CD133+ cells

As acetylation of KLF5 has been shown to alter the function of KLF5 in several cell lines,^22,23 we next examined the effect of acetylation of KLF5 on CD44+/CD133+ cancer stem-like cells. To induce acetylation of the DNA binding site of KLF5, cells were treated with TGF- β1.²³ Although TGF-β1 decreased endogenous KLF5 expression, ectopically expressed KLF5 was unaffected by TGF-β1 (Fig. 5A). However, expression level of CD44 was significantly suppressed by TGF-β1 in both control and KLF5-overexpressing cells (Fig. 5B). Consistent with decreased CD44 expression, the CD44+/CD133+ subpopulation was significantly suppressed by TGF-β1 in KLF5-overexpressing cells in which KLF5 expression was not affected by TGF-β1 (Fig. 5C). These results suggest that acetylation status of the DNA binding site of KLF5 may be important for regulation of CD44+/CD133+ cells. To further explore the role of acetylation of KLF5 in regulation of CD44+/CD133+ cells, we generated Huh7 cells stably overexpressing an acetylation-deficient mutant of KLF5 (KLF5-K369R). Protein levels of the wild type and the mutant KLF5-K369R were similar as confirmed by protein gel blotting (Fig. 6A). Induction of CD44 was robustly enhanced by the KLF5-K369R mutant (Fig. 6B). The known target genes of KLF5,¹⁷ Oct4 and Pim1 were further increased and p15 was further suppressed by the KLF5-K369R mutant than the wild type (Fig. 6C). The CD44+/CD133+ subpopulation was also dramatically increased by the KLF5-K369R mutant (Fig. 6D). Thus, acetylation status of KLF5 might be critical in the transcriptional activity of KLF5 and the regulation of CD44+/CD133+ cancer stem-like cells in HCC.

Figure 5. TGF-β1 decreases the CD44+/CD133+ subpopulation independent of altered KLF5 expression. Control (Empty) and KLF5 overexpressing (KLF5) Huh7 cells were treated with TGF-β1 (5ng/ml) for 48 hours. (A) Expression of KLF5 was determined by WB. (B) Expression of CD44 was determined by real-time PCR (* P < 0.05 vs. Empty/TGF-β1 (−), # P < 0.05 vs. Empty/TGF-β1 (−), ** P < 0.05 vs. KLF5/TGF-β1 (−)). (C) Expression of CD44/CD133 and percentage of CD44+/CD133+ cells were determined by FACS.

Figure 6. Acetylation status of KLF5 is critical for regulation of the CD44+/CD133+ subpopulation. Wild type (KLF5-WT) and acetylation-deficient (KLF5-K369R) KLF5 were transduced by retrovirus-mediated gene transfer into Huh7 cells. (A) Expression of KLF5 was determined by WB. (B) Expression of CD44 was determined by real-time PCR (* P < 0.05 vs. Empty, # P < 0.05 vs. KLF5-WT). (C) Expression levels of the known KLF5 target genes, Oct4, Pim1 and p15 were determined by real-time PCR (* P < 0.05 vs. Empty, # P < 0.05 vs. KLF5-WT). (D) Expression of CD44/CD133 and percentage of CD44+/CD133+ cells were determined by FACS.

Discussion

In the present study, we have shown that KLF5 is a pivotal transcription factor involved in the regulation of CSCs in HCC. Recent genome-wide sequencing studies have revealed that 10 to 30% of HCC cases have mutations in exon 3 of CTNNB1, a key mediator of Wnt/β-catenin signaling.²⁴ CD44 and Myc are targets of Wnt/β-catenin signaling and aberrant activation of Wnt/β-catenin signaling leads to expansion of CSCs in HCC.^11,25,26 Activation of Wnt/β-catenin has been reported to increase expression of EpCAM, one of cancer stem cell markers,²⁷ and Wnt/β-catenin and EpCAM form multiprotein complexes to regulate shared target genes.²⁸ It is obvious that Wnt/β-catenin signaling is important for generation and maintenance of CSCs in HCC. Since it has been reported that deletion of KLF5 from the intestinal mucosa results in attenuation of Wnt/β-catenin signaling,²⁹ KLF5 and Wnt/β-catenin signaling may intersect in the regulation of CSCs in HCC. Inflammation also plays important roles in regulation of CSCs in HCC. IL-6 secreted by tumor-associated macrophages promotes expansion of CD44 positive CSCs via activation of STAT3,³⁰ and IL-8 has been shown to enhance self-renewal and tumor initiation of CSCs through activation of the mitogen-activated protein kinase (MAPK) pathway in HCC.³¹ Most HCCs arise from chronic inflammation such as viral hepatitis, non-alcoholic steatohepatitis and alcoholic hepatitis. KLF5 might have some roles in hepatitis and further study is necessary to clarify the role of KLF5 in inflammation in regulation of CSCs in HCC.

A role of KLF5 in the regulation of CSCs has not been well established so far. In Helicobacter pylori infection, KLF5 is induced by CDX1 and KLF5 promotes trans-differentiation of gastric epithelial cells into intestinal epithelial cells, leading to carcinogenesis and enrichment of CSCs in gastric cancer.^19,20 Nakaya et al. have reported that KLF5 acts as a key regulator of intestinal carcinogenesis at the stem-cell level in their mouse model.³² These reports indicate that KLF5 might play important roles in the regulation of CSCs; however, detailed mechanisms through which KLF5 regulates CSCs have yet to be elucidated. In the present study, we have shown that KLF5 regulates expression of CD44, one of the most important stem cell markers. Although Pagliuca et al. have shown that both CD44 and KLF5 are regulated by miR-143 and miR-145,³³ regulation of CD44 by KLF5 has not been reported thus far. Our findings indicate that expression of CD44 is controlled by expression level of KLF5 in overexpression and knockdown experiments (Figs. 3 and 4). CD44 interacts with xCT, a light-chain subunit of the cystine-glutamate transporter xc(−), and reduces reactive oxygen species (ROS). CD44 protects CSCs from excessive ROS under stressful conditions, such as rapid tumor growth and metastasis, and maintain CSCs.^34,35 Thus, KLF5 may enrich and maintain CSCs via induction of CD44, one of important functional molecules in CSC biology. In the present study, we have not determined if KLF5 directly upregulates expression of CD44. Smith et al. have shown that NF-κB directly up-regulates CD44 expression.³⁶ KLF5 may indirectly regulate CD44 expression through NF-κB, a known KLF5 target gene.³⁷ Further analysis is required to identify the detailed regulatory mechanism of CD44 by KLF5.

Previous reports have shown that post-translational modification of KLF5 dramatically alters transcriptional activity and function of KLF5.¹⁷ Phosphorylation of KLF5 increases its transactivation activities,³⁸ ubiquitination of KLF5 by WWP1 or FBW7 regulates its stability,^39,40 and sumoylation regulates the nuclear localization of KLF5 by inactivating nuclear export signal.⁴¹ Interestingly, acetylation of KLF5 by TGF-β signaling converts the function of KLF5 from tumor suppressive to oncogenic by changing target gene expression profile of KLF5.^22,23 Therefore, we hypothesized that acetylation status of KLF5 in addition to expression level might play an essential role in maintaining the malignant properties of CSCs in HCC. The present study indicates that TGF-β1 decreased expression of CD44 at both the mRNA and protein level although the protein expression level of KLF5 was not changed (Fig. 5). In addition, expression of a point mutation in KLF5 at amino acid 369 (KLF5-K369R), resulting in an acetylation-deficient mutant of KLF5, significantly increased CD44 expression compared with wild-type KLF5, where protein expression levels of KLF5-K369R and wild-type KLF5 were equal (Fig. 6). This suggests that KLF5, especially deacetylated KLF5, is important for upregulation of CD44 in HCC. We have shown for the first time that acetylation status of KLF5 is critical for regulation of CSCs in HCC.

In conclusion, we have identified KLF5 as one of the pivotal transcription factors regulating CSCs in HCC. KLF5 upregulates CD44 expression and expands CSCs. Furthermore, acetylation status of KLF5 is critical for regulation of CSCs and deacetylated KLF5 robustly increases subpopulation of CSCs in HCC. Thus, modulation of KLF5 expression and acetylation status may be investigated as potential novel therapeutic targets against CSCs in HCC.

Materials and methods

Cell lines and reagents

Hepatoma cell lines, Huh7 and HepG2 cells were grown in DMEM supplemented with 10% fetal bovine serum, 100 units/ml of penicillin G and 100 µg/ml of streptomycin. Anticancer drugs, cisplatin (CDDP) and 5-fluorouracil (5FU) were purchased from Sigma Aldrich (Saint Louis, MI). Recombinant human TGF-β1 (PHG9214) was purchased from Life Technologies (Carlsbad, CA).

DNA constructs

KLF5 was amplified by PCR using ADV-KLF5 ⁴² as a template with primers 5′- ACCGGATCCGCCACCATGGCTACAAGGGTGCTGAGCATGAG -33′ and 5′- ACCGTCGACTCAGTTCTGGTGCCTCTTCAT-33′. Amplified KLF5 was ligated at the BamHI and SalI sites of pBABE puro and designated as pBABE-KLF5. A point mutation was introduced into KLF5 by PCR using pBABE-KLF5 as a template with primers 5′-AGACGACGCATCCACTACTGCGAT-33′ and 5′-CTCCAAATCGGGGTTACTCCTTCT-33′. Mutated KLF5 was ligated at the BamHI and SalI sites of pBABE puro and designated as pBABE-KLF5-K369R. The DNA binding site of KLF5-K369R is acetylation deficient.

Retrovirus-mediated gene transfer

Replication-incompetent retrovirus expressing KLF5 and KLF5-K369R were produced and transduced as described previously.^43,44 Stable cell lines were established by drug selection with 2 μg/ml puromycin for 7 days.

Real-time PCR

RNA was extracted by RNeasy mini kit (QIAGEN, Valencia, CA USA) and cDNA was synthesized using Super Script II (Life Technologies) according to the manufacturer3′s instruction. Real-time PCR was performed with SYBR® Green (QIAGEN) and ABI 7300 real time PCR system (Life Technologies). β-actin was used as an internal control. Primers used for real-time PCR are listed in Table 1.

Table 1. Primers used for real-time RT-PCR

	Forward primer	Reverse primer
CD44	5′-taaggacaccccaaattcca-3′	5′-actgcaatgcaaactgcaag-3′
CD133	5′-tggccaccgctctagatact-3′	5′-gcttttcctatgccaaacca-3′
KLF5	5′-aaggagtaaccccgatttgg-3′	5′-tggcttttcaccagtgtgag-3′
β-actin	5′-ttgccgacaggatgcagaa-3′	5′-gccgatccacacggagtact-3′

Western blotting

Western blotting was carried out as described previously ⁴³ using the following primary antibodies at the indicated titers; anti-KLF5 (ab24331, Abcam) (1:1000) and anti-β-actin (AC74; Sigma-Aldrich) (1:5000).

Drug sensitivity assay

Five thousands cells were seeded per well in a 96 well plate. Twenty-four hours after seeding, drugs were added at indicated concentrations. After 72 hours incubation, MTS assay was carried out using CellTiter 96® AQueous One Solution Cell Proliferation Assay (Promega, Fitchburg, Wisconsin) according to the manufacturer3′s instruction. Absorbance (A₄₉₀ - A₆₃₀) was measured by Sunrise™ plate reader (F039302, TECAN, Durham, NC) and survival rate was determined.

Soft-agar colony formation assay

Soft agar colony formation assays were performed as described previously.⁴⁵ In brief, 1 × 10³ cells were suspended in 0.67% agarose in growth media and then overlaid on top of 1% agarose in growth media per well. (24 well plate). 500 µl of medium with indicated reagents was added twice a week in each well and grown for 3 weeks. Colonies over 100 µm were counted. The experiments were performed in hexaduplicate.

Flow cytometry and Fluorescence Activated Cell Sorting (FACS)

FACSCantoII (BD Biosciences, San Jose, CA) and FlowJo (Tree Star, Ashland, OR) were used for flow cytometry. Cells were suspended in Hank's balanced salt solution (Invitrogen) containing 1% BSA (Sigma-Aldrich) and stained with APC-anti-CD44 (BD Biosciences) at 1:20 and PE-anti-CD133 (Miltenyi Biotec, Auburn, CA) at 1:20 on ice for 30 min. Dead cells were eliminated by 7-AAD (BD Biosciences) according to the manufacturer3′s instruction. To purify CD44^High / CD133^High CSCs and CD44^Low / CD133^Low non-CSCs, cells were sorted using a FACS Aria III (BD Bioscience).

siRNA transfection

Small interfering RNA (siRNA) sequences directed against KLF5 (S104161381 and S204281074) and a non-silencing control sequence (AllStars Negative Control siRNA) (QIAGEN) were purchased from QIAGEN. siRNAs were transfected transiently with Lipofectamine™ RNAi Max reagent (Life Technologies), following the manufacturer's instructions.

RNA sequence

RNA was extracted using RNeasy mini kit (QIAGEN) from sorted CD44⁺/CD133⁺ and CD44⁻/CD133⁻ Huh7 cells. Intact polyadenylated (poly A) mRNAs were enriched by poly A selection using Dynbeads^® mRNA DIRECT ™ Micro Kit (Life Technologies). Libraries for RNA sequence were constructed by Ion Total RNA-Seq Kit version 2 (Life Technologies) according to the manufacturer3′s instruction. Quality of the library was checked using Agilent Bioanalyzer 2100 (Agilent Technologies, Waldbronn, Germany). Emulsion PCR was carried out using Ion One Touch™ Duo and Ion One Touch™ ES (Life Technologies) and sequencing was performed using the Ion PGM™ system (Life Technologies), following the manufacturer3′s instruction. Acquired sequences were mapped to homo sapiens Hg19 and analyzed by CLC Genomics Workbench (CLC bio, Aarhus, Denmark). We set a cut-off value of RPKM (reads per kilobase of exon per million mapped sequence reads) ⁴⁶ as 1.0 in the present study.

Statistical analysis

Data from triplicate and hexaduplicate experiments are presented as mean ± SD and were analyzed by 2-tailed Student3′s t test. P < 0.05 was considered significant.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Funding

This study was supported in part by the Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT) KAKENHI Grant Numbers 25860512, 26460933, and 15K08943, and Japanese Health Labor Sciences Research Grant.