Licochalcone A Inhibits BDNF and TrkB Gene Expression and Hypoxic Growth of Human Tumor Cell Lines

Hypoxic cellular proliferation is a common feature of tumor cells and is associated with tumor progression. Therefore, the inhibition of hypoxic cellular proliferation is expected to regulate malignancy processes. Licochalcone A (LicA) is known to show inhibitory effects on cell growth in normoxia, but it is unclear whether LicA exerts similar effects in hypoxia. Here, we studied the inhibitory activity of LicA in the hypoxic cellular proliferation of tumor cells and its molecular mechanism using human cell lines derived from various tumors including neuroblastoma and colorectal cancer. LicA inhibited cell growth at a 50% inhibitory concentration between 7.0 and 31.1 µM in hypoxia. LicA significantly suppressed hypoxic induction of tropomyosin receptor kinase B (TrkB) gene expression at the transcription level. LicA also downregulated mRNA levels of the TrkB high-affinity ligand brain-derived neurotrophic factor, but not neurotrophin-4, another TrkB ligand, or glyceraldehyde-3-phosphate dehydrogenase, indicating that the inhibitory activity of LicA is selective. Since both LicA-treatment and TrkB-knockdown decreased activation of protein kinase B in hypoxia, LicA exerts its inhibitory effect against hypoxic cell growth through inhibition of the TrkB-AKT axis. These results suggest that LicA has therapeutic potential for malignant tumors including neuroblastoma and colorectal cancer.


Introduction
Hypoxic adaptation, which allows cellular proliferation and survival even at low oxygen levels, is a common feature of tumor cells and promotes tumor malignancy. To adapt to a hypoxic microenvironment, cells change the expression of the genes responsible for extracellular neo-vascularization and the intracellular metabolic shift to aerobic glycolysis [1,2]. The metabolic shift in hypoxia causes acidosis and induces cell invasion and metastasis via epithelial-to-mesenchymal transition (EMT) [3,4]. A key regulator of these processes is the phosphatidylinositol 3-kinase/protein kinase B/mammalian target of rapamycin (PI3K/AKT/mTOR) signaling pathway [1,5,6].
The PI3K/AKT/mTOR pathway is initiated by binding of ligands to receptor tyrosine kinases (RTKs) [1,6]. Tropomyosin receptor kinase B (TrkB), an essential signaling initiator for neurodevelopment, is one of the tumor-associated RTKs [7,8]. TrkB and its high-affinity ligand brain-derived neurotrophic factor (BDNF) had originally been recognized as notable markers for poor prognosis in neuroblastoma and as key players for its progression [9][10][11]. Recently, TrkB and its ligand, BDNF, have been shown to be overexpressed in various types of cancer, such as cervical, breast, and colorectal cancer (CRC), and to be associated with their poor prognosis [7,12,13]. TrkB has also been shown to be a key player in metastasis in studies using a rat intestinal epithelial cell line [14], a mouse xenograft model [15], and human lung cancer cell lines [16]. Furthermore, TrkB is upregulated in hypoxia [16,17]. Thus, hypoxic induction of BDNF/TrkB is expected to be an event that may serve as a therapeutic target for preventing the progression of advanced cancer, especially neuroblastoma and CRC.
Licochalcone A (LicA), a Glycyrrhiza flavonoid, has been shown to have anti-tumor activity in various systems. Potential therapeutic effects of LicA have been demonstrated in gastric, lung, and cervical cancer cells [18][19][20][21]. Although LicA has been reported to inhibit the activation of TrkB-downstream molecules, such as PI3K, AKT, and mTOR in previous studies, the effects of LicA on TrkB and BDNF as well as on hypoxic cellular proliferation are not yet known.
In this study, we examined the inhibitory effect of LicA on hypoxic growth in human cell lines derived from various tumors, including neuroblastoma and CRC. Furthermore, we examined the effect of LicA on TrkB/BDNF, an initial point of hypoxic cellular survival and proliferative signaling pathways, to explore the mechanism underlying LicA-induced growth inhibition.

LicA Inhibits Cell Growth in Hypoxia
We studied the effects of LicA on hypoxic and normoxic cell growth using the neuroblastoma SK-N-SH cell line, which showed somewhat lower proliferative ability in hypoxia than in normoxia based on cell counts ( Figure 1A) and mitochondrial activity ( Figure 1B). LicA at concentrations of 5-40 µM showed growth inhibitory activity in hypoxia-exposed cells at 72 h ( Figure 1C). LicA (20 µM) exerted an inhibitory activity in a time-dependent manner ( Figure 1D). Significant growth inhibition against pre-treatment (0 h) was detected at 24 h (P value for normoxia and hypoxia are 0.0385 and 0.0048, respectively) and continued for 72 h. The 50% inhibitory concentration (IC 50 ) for LicA was 8.4 µM (Table 1). LicA inhibition seems to be similar in hypoxic and normoxic cells.  Since the growth inhibitory activity of LicA was observed in neuroblastoma SK-N-SH cells, we examined the effect on six other cell lines derived from various origins including neuroblastoma, CRC, and cervix carcinoma, as listed in Table 1. All of the cell lines showed a proliferative ability in hypoxia as observed for SK-N-SH (Table 1). In addition to the effects seen in SK-N-SH cells, LicA showed potent inhibitory activity in the colorectal SW480 and SW620 cell lines and the cervix HeLa S3 cell line, with IC 50 values between 7.0 and 10.7 µM ( Table 1). The IC 50 values of the other cell lines examined were approximately 3-fold (24.9-31.1 µM). LicA exhibited inhibitory activity against hypoxic cellular proliferation regardless of tumor origin, even though IC 50 values varied. Since SK-N-SH is one of the most sensitive cell lines to LicA, we used this cell line in the subsequent investigations.

LicA Inhibits Hypoxic Activation of TrkB and BDNF Expression
Proliferative signals in neuroblastoma cells are known to be initiated by ligand-binding to TrkB [22,23] (reviewed in [24,25]). To determine whether TrkB and its high-affinity ligand BDNF increase in hypoxia and whether LicA causes a decrease in either TrkB or BDNF, we focused on investigating the effects of LicA on the expression of the TrkB and BDNF genes. As expected, TrkB mRNA levels in SK-N-SH cells were increased by hypoxia in a time-dependent manner, and they were increased 2.9-fold of those seen in normoxic cells at 48 h ( Figure 2A). Hypoxic induction of TrkB mRNA levels was significantly inhibited to a basal level by 20 µM LicA detected at 24 h or 48 h (Figure 2A). A similar inhibitory activity of LicA was confirmed at the TrkB protein level, which was increased approximately 2-fold by hypoxia ( Figure 2D). Inhibition by LicA was detected at 12 h, and LicA-treated cells showed 51.6% and 60.7% of the band intensity of untreated cells at 12 h and 24 h, respectively. BDNF mRNA levels were also increased 3.7-fold at 48 h by hypoxia in a time-dependent manner ( Figure 2B). BDNF mRNA levels were also significantly inhibited by LicA, but this effect was detected only after 48 h and appeared to be weaker than that on TrkB. We also monitored the mRNA level of neurotrophin-4 (NT-4), a low affinity ligand for TrkB. While hypoxia transiently increased NT-4 mRNA levels at 24 h (1.9-fold) and decreased NT-4 mRNA levels at 48 h (1.4-fold), LicA continuously increased hypoxic NT-4 mRNA levels significantly at 48 h (2.1-fold; Figure 2C) without any inhibition. Taken together, LicA does not suppress universal mechanisms such as bindings of general transcription factors or RNA polymerization but specific mechanisms for TrkB and BDNF gene expression. TrkB as well as BDNF have been reported to be overexpressed in various types of cancer, such as colorectal and cervical cancers, and BDNF/TrkB overexpression has been shown to confer a migratory phenotype and chemotherapy resistance to tumor cells [12]. To determine whether the inhibitory activity of LicA on TrkB and BDNF expression in hypoxia is specific to the SK-N-SH cell line or if it is also exerted in cell lines from other neuroblastoma and/or tumors, we tested the effects of LicA on the amounts of TrkB and BDNF mRNA in the cell lines listed in Table 1. Various levels of TrkB mRNA were detected in the cell lines we tested compared to TrkB mRNA levels in SK-N-SH when cells were cultured in hypoxia for 48 h ( Table 2). LicA at a concentration of 20 µM inhibited hypoxic TrkB mRNA expression in all cell lines tested with a wide range of inhibitory activity (47.4% to 94.8%; Table 2). The effects of LicA on BDNF expression were examined in three other cell lines: neuroblastoma TGW, colorectal SW480, and cervical HeLa S3. Hypoxic BDNF expression in these cell lines was approximately 3.8-to 10-fold that of SK-N-SH. LicA inhibited BDNF expression by 13.7% to 96.6% (Table 2). Thus, the inhibitory activity of LicA may be seen generally in tumor cells which overexpress TrkB and BDNF. mRNA levels with or without 20 µM LicA were determined at 48 h in hypoxia. mRNA levels are expressed as values relative to the mRNA levels of untreated SK-N-SH for each gene. The relative values are shown as mean ± SD derived from 3 independent experiments. P values for differences between mRNA levels with and without 20 µM LicA in hypoxia were calculated using unpaired Student's t-test (two-tail). ** p < 0.01. N.D., not determined.
TrkB and BDNF were upregulated in hypoxia, and LicA inhibited the upregulation, as shown in Figure 2A. This result suggests that LicA suppresses transactivators or induces repressors. To ascertain whether LicA suppresses hypoxia-inducible transcription factor-1 (HIF-1), a key transactivator in hypoxia, we monitored the protein levels of the HIF-1 α-subunit (HIF-1α), an oxygen-sensitive subunit of HIF-1, in the cytosol and nucleus. We found that nuclear levels of HIF-1α were enhanced by LicA, whereas cytosolic levels were decreased ( Figure 2E), indicating that LicA likely activates HIF-1 by increasing the nuclear transport of HIF-1α. Consistently, hypoxic expression of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene, known to also be upregulated by HIF-1, was increased by LicA ( Figure 2F). These results indicate that LicA exhibits inhibitory activity on the hypoxic mRNA levels of TrkB and BDNF through a mechanism other than HIF-1 suppression. Further studies are needed to determine the target molecules of LicA.
The inhibitory effect of LicA on mRNA levels was detected earlier in the TrkB gene than in the BDNF gene. To determine whether the hypoxic growth inhibition of SK-N-SH by LicA is started by downregulation of the TrkB mRNA level, we performed knockdown experiments. TrkB was downregulated by siRNA for the TrkB gene at the mRNA and protein levels ( Figure 2G,H). Under these conditions, hypoxic cell growth of SK-N-SH was significantly inhibited ( Figure 2I). These results suggest that LicA exerts its inhibitory effect against hypoxic cell growth of SK-N-SH through the suppression of TrkB gene expression at the mRNA level.

LicA Inhibits TrkB-Downstream Cell Growth Signaling
We showed that LicA inhibited hypoxic cell growth of SK-N-SH by downregulation of the mRNA level of TrkB, which was known to transduce via downstream pathways including the PI3K/AKT cascade [22,23] (reviewed in [24,25]). To determine whether LicA exhibits its inhibitory effect on SK-N-SH cell growth via inhibition of the TrkB/PI3K/AKT axis, we investigated the effect of LicA on AKT activation regulated by phosphorylation at Thr308 (T308) and Ser473 (S473). LicA inhibited phosphorylation of AKT at both T308 and S473 ( Figure 3A). The inhibition was detected at 12 h, which was consistent with the effects of LicA on cell growth ( Figure 1D) and on TrkB protein level ( Figure 2D). To more directly evaluate whether the inhibition of AKT activation is due to downregulation of TrkB expression, we performed knockdown experiments. The knockdown of TrkB using siRNA also downregulated AKT activation at both phosphorylation sites by 38.8% at T308 and 33.6% at S473 against control siRNA ( Figure 3B). These results suggest that LicA exhibits growth inhibition through interception of the TrkB/PI3K/AKT cell growth signaling cascade.

LicA inhibits TrkB and BDNF Promoter Activity
We observed that LicA selectively downregulated the mRNA levels of TrkB and BDNF genes. mRNA levels depend on a balance between transcription and mRNA degradation. To determine if LicA inhibits transcription of TrkB and BDNF mRNA, we constructed reporter plasmids covering the promoter region of each gene and performed reporter gene assays in two neuroblastoma cell lines, SK-N-SH and TGW, and a CRC cell line, SW480. The promoter activities of TrkB and BDNF were higher in hypoxia than in normoxia. LicA (20 µM) reduced the promoter activity of TrkB and BDNF in all cell lines (Figure 4). The inhibition of TrkB was approximately 34.8% in SK-N-SH, 47.1% in TGW, and 10.9% in SW480. The inhibition of BDNF was approximately 46% in SK-N-SH, 57.3% in TGW, and 27.4% in SW480. These results suggest that LicA inhibits TrkB and BDNF gene expression at the transcriptional level.

Discussion
LicA blocks cell proliferation in hypoxia as well as in normoxia. LicA also selectively suppresses the hypoxic expression of TrkB and BDNF at the transcription level. The selectivity of LicA for TrkB and BDNF might be based on these gene-specific regulatory factors involved in the transcription of the genes such as activators or repressors. HIF-1 is a key transactivator that accumulates in hypoxia and has been reported to upregulate TrkB transcription in human neuroblastoma and lung cancer cell lines in hypoxia [16,17]. However, the inhibitory activity of LicA might not be due to suppression of HIF-1, since LicA enhanced nuclear levels of HIF-1α in hypoxia ( Figure 2E). We suggest that, for the TrkB and BDNF genes, LicA may inhibit other transactivating factors, or that it may enhance suppressive factors. LicA also inhibits normoxic TrkB expression (our unpublished results), indicating that the factors might play in both normoxia and hypoxia. Based on computational prediction using the JASPAR database [26], several binding sites for transcription factors, such as LMX1B, NOTO, MZF1, MSX2, NKX3-2, and AP-1, are commonly found in the regulatory regions used in the reporter gene assay. Further investigation is required to determine the details of the transcriptional inhibition observed with LicA.
TrkB has been reported to regulate cellular proliferation in various cancer cells (reviewed in [12,27]). In accordance with this notion, TrkB knockdown in the SK-N-SH cell line in hypoxia caused inhibition of cellular proliferation in this study ( Figure 2I). However, the extent of the growth inhibition achieved by TrkB knockdown was less extensive than that seen with LicA. This suggests that LicA is most likely to suppress not only BDNF and TrkB but also other genes responsible for hypoxic proliferation.
Potential therapeutic effects of LicA have been demonstrated using various human cancer cell lines [18][19][20][21]. In these reports, LicA was shown to exert the effects by inhibiting the activating phosphorylation of signal transducers, such as PI3K, AKT, mTOR, and mitogen-activated protein kinases (MAPKs). In the present study, we show that LicA suppresses expression levels of the BDNF and TrkB genes and also decreases AKT activation (Figure 3).
In this study, IC 50 values of LicA varied widely, ranging from 7.0 to 35.5 µM (Table 1). Less sensitive cell lines were shown to have abnormal activation downstream of growth signaling pathways. TGW and GOTO are known to have several hundred MYCN gene copies and produce an enormous amount of the MYCN protein [28] downstream of TrkB and mTOR [29]. Activated downstream signaling events may not be enough to detect the inhibitory capability of LicA, indicating that a molecular-targeted compound in combination with LicA might be useful for the treatment of solid tumors.
It has been reported that TrkB activation initiates three downstream signaling cascades, namely PI3K/AKT/mTOR, Ras/extracellular signal-regulated kinase (ERK), and phospholipase C-γ/protein kinase C (PLCγ/PKC) [12,27]. The activating mutations of the K-ras gene constitutively turn on the downstream K-ras-transduced signaling pathways, Ras/ERK and PI3K/AKT. However, two CRC cell lines harboring the G12V K-ras mutation, SW480 and SW620 [30], have been shown to be relatively sensitive to LicA (Table 1). This finding may indicate that the K-ras-transduced pathways are less essential for SW480 and SW620 than the PLCγ/PKC pathway, which is not transduced by K-ras, and that BDNF/TrkB seems to be important for the proliferation of these CRC cell lines. Taken together with previous studies reporting that TrkB overexpression has been observed in advanced CRC [25,31], our results indicate that curbing of hypoxic TrkB and BDNF expression may be a potent therapeutic strategy for certain types of ras-mutated CRCs.
In summary, we demonstrated here that LicA is a novel BDNF/TrkB inhibitor that suppresses gene transcription selectively; therefore, LicA may potentially be a key therapeutic compound for the treatment of malignant tumors including neuroblastoma and CRC.

Cells and Cell Culture
The human neuroblastoma cell line SK-N-SH [32], the human CRC cell lines SW480 and SW620, and the human cervix carcinoma cell line HeLa S3 were obtained from American Type Cell Culture (Rockville, MD, USA). The human neuroblastoma cell lines TGW and GOTO were described previously [33]. All cell lines were maintained in MEM supplemented with 10% fetal bovine serum. The cells were cultured at 37 • C in a humidified atmosphere with 5% CO 2 . Cell passaging was performed weekly. In hypoxia, cells were cultured in 5% CO 2 with 0.1% O 2 atmospheric conditions, and the O 2 concentration was controlled using a ProOx model 110 controller (BioSpherix, Parish, NY, USA) as previously described [34]. The cells were continuously subjected to hypoxia until they were harvested for the subsequent experiments.

Cell Growth Monitoring
Cell growth was monitored using a 2-(2-methoxy-4-nitrophenyl)-3-4-nitrophenyl)-5-(2,4disulfophenyl)-2H-tetrazolium, monosodium salt (WST-8) kit (Kishida Chemical, Osaka, Japan). Cells were plated in 96-well plates at a density of 3-7 × 10 3 cells/well in a total volume of 100 µL. After culturing the cells for 24 h in normoxia, various dosages of compounds were added in triplicate, and culturing was continued under either normoxic or hypoxic conditions. After cultivation for the appropriate periods, 10 µL of WST-8 solution was added and the cells were incubated for 4 h at 37 • C. The absorbance at 450 nm was measured using a microplate reader (Model 680; Bio-Rad Laboratories, Hercules, CA, USA) and analyzed.

Quantitative Reverse-Transcription Polymerase Chain Reaction
Gene expression at the mRNA level was quantified by quantitative reverse-transcription polymerase chain reaction (qRT-PCR). Exponentially growing cells were seeded in a 24-well plate at an initial density of 2-4.5 × 10 4 /well and cultured at 37 • C. After 24 h, the plates were placed in a normoxic or hypoxic (0.1% O 2 ) chamber. Total RNA was extracted using Isogen-II (Nippon Gene, Tokyo, Japan). A One Step SYBR PrimeScript PLUS RT-PCR kit (Takara Bio, Shiga, Japan) was used for the subsequent cDNA synthesis and RT-PCR. Each sample was amplified in triplicate and analyzed using an Applied Biosystems 7500 Fast Realtime PCR system (Life Technologies, Carlsbad, CA, USA). Ribosomal 18S RNA was used for normalization. The primer sequences used in this study are listed in Table 3.

Immunoblotting
Cells at an initial density of 5 × 10 5 cells per 10 cm dish were cultured for 24 h under normoxic condition and transferred to a hypoxic chamber (0.1% O 2 ). After the indicated hours of hypoxia exposure, the cells were harvested, lysed, and immunoblotting analysis was performed as described previously [37]. For HIF-1α, nucleus and cytosol fractions were prepared as described previously [38]. Proteins in heat-denatured cell lysates were separated by SDS-PAGE and then transferred to PVDF membranes in Tris-Gly buffer. For TrkB, a Tris-lactic acid buffer system [39] was used. To detect proteins, the primary antibodies listed above were used. α-Tubulin was used as a loading control. The blots were visualized using the Amersham ECL-plus system (GE Healthcare UK, Buckinghamshire, UK) and imaged and quantified using a chemiluminescent image analyzer (EZ-Capture MG; Atto, Tokyo, Japan).

siRNAs and Transfection
Cells at an initial density of 2 × 10 5 cells per 10 cm dish were cultured for 24 h under normoxic conditions and transfected with siRNA by changing the culture medium to Dharmacon Accell siRNA delivery medium (Horizon Discovery, Cambridge, UK) containing 1 µM Accell SMARTpool siRNA (Horizon Discovery, E-003160-01 for TrkB, D-001910-10 for non-targeting control). After 24 h culture in normoxia, the Accell siRNA delivery medium was changed to culture medium and cells were further cultured in hypoxia. After 48 h, cells were subjected to WST-8 assay, qRT-PCR, and immunoblotting.

Luciferase Reporter Gene Analysis
The regulatory regions of the TrkB gene (−1880 to +42 relative to the transcription start site of exon 1a, NCBI accession no. AF410902.1) [40] and the BDNF gene (−629 to +281 relative to the transcription start site of exon 4, NCBI accession no. NM_170733.3; promoter IV described by Yasuda et.al) [41] were amplified using genomic DNA of SK-N-SH as a template and TrkB-F (5 -TGG GTA CCT GCC TTT ATG TAT GTG TTC ATT GGG-3 ) and R (5 -TGG GTA CCG GTG ACA AAC CGT AAC TAT TCT GC-3 ) and BDNF-F (5 -ATT ACG CTA GCC AGG TGA TTT TTA TGC TCC GAG G-3 ) and R (5 -ATT ACG CTA GCT CGG CCC CAA AAC TCC CAC-3 ) as primers, which were designed with the NCBI's Primer-BLAST online tool (https://www.ncbi.nlm.nih.gov/tools/primer-blast). Amplified fragments were digested by KpnI for TrkB and NheI for BDNF and cloned to a corresponding multiple cloning site of a promoter-less luciferase reporter vector, pGL4.10 (Promega, Madison, WI, USA). The sequence of the clones was confirmed by a capillary sequencer, 3130xl Genetic analyzer (ABI), to be identical to the wild-type sequence found in GenBank. Sequencing primers are listed in Table 4. All procedures for construction of reporter plasmids were approved by the Toho University Administration Panel for Recombinant DNA (19-54-307). Luciferase reporter gene analysis was done as described previously with some modification [34]. Prior to transfection with the reporter plasmids, 8 × 10 3 SK-N-SH, TGW, or SW480 cells were plated in a well of a 96-well plate. The reporter plasmids were transected in normoxia for 24 h before adding LicA with the lipofection reagent Invitrogen Lipofectamine 3000 (Thermo Fisher Scientific, Waltham, MA, USA). The cells were cultured in hypoxia with or without LicA for 6 h. Luciferase activity was determined using the One-Glo luciferase assay system (Promega, Madison, WI, USA) and the multimode microplate reader model TriStar LB 941 (Berthold Technologies, Bad Wildbad, Germany). Two identical series of the transfected cell cultures were prepared: one for the luciferase assay and the other for counting cell density. Cell density measured by WST-8 assay was used for normalization. Promoter activity was expressed relative to values obtained under normoxic LicA-untreated conditions.

Statistical Analysis
Statistical significance was determined by Student's t-test using the online software QuickCalcs (GraphPad, La Jolla, CA, USA). A P value of less than 0.05 indicated statistical significance.