This research belongs to Ajinomoto Pharmaceutical Company. The authors declare that they have applied for a patent related to BCAA used in this study (Patent name: Agent for enhancing the antitumor activity of chemotherapeutic drug, Patent application No: WO2012/111790). SN, MH and SI are employed by Ajinomoto Pharmaceuticals, Co, Ltd. This study is conducted on animal models and has no direct implications on human subjects. However, the work of this study will be translated to human subjects in future studies and hence this publication may add some benefit to Ajinomoto Pharmaceuticals, Co, Ltd. There are no further patents, products in development to declare. Compounding ratio of BCAA used in this study is same as a compounding ratio of BCAA granules which are sold by Ajinomoto Pharmaceuticals Co. Ltd. in Japan by the name of LIVACT ® Granules. However, the company doesn’t draw a direct commercial benefit from this study as it is not conducted in human subjects. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.
Conceived and designed the experiments: SN HY. Performed the experiments: SN MH. Analyzed the data: SN SI. Contributed reagents/materials/analysis tools: HY. Wrote the manuscript: SN SI.
Differentiation of cancer stem cells (CSCs) into cancer cells causes increased sensitivity to chemotherapeutic agents. Although inhibition of mammalian target of rapamycin (mTOR) leads to CSC survival, the effect of branched chain amino acids (BCAAs), an mTOR complex 1 (mTORC1) activator remains unknown. In this study, we examined the effects of BCAA on hepatocellular carcinoma (HCC) cells expressing a hepatic CSC marker, EpCAM. We examined the effects of BCAA and/or 5-fluorouracil (FU) on expression of EpCAM and other CSC-related markers, as well as cell proliferation in HCC cells and in a xenograft mouse model. We also characterized CSC-related and mTOR signal-related molecule expression and tumorigenicity in HCC cells with knockdown of Rictor or Raptor, or overexpression of constitutively active rheb (caRheb). mTOR signal-related molecule expression was also examined in BCAA-treated HCC cells.
The term “cancer stem cell” (CSC) refers to a cancer cell with the characteristics of a stem cell. Stem cells carry the potential to self-renew and differentiate into other cell types, and can therefore restore cells undergoing apoptosis. It is hypothesized that carcinomas developing from stem cells undergo a process of asymmetric division. CSCs were initially identified in acute myeloid leukemia [
Two approaches to cancer therapy have been suggested: Wake up therapy and Sleep therapy. Wake up therapy enhances sensitivity to chemotherapeutic agents by inducing differentiation from CSCs to cancer cells. Oncostatin M, a cytokine of IL-6 family, is a well-known inducer of differentiation. Combination of this cytokine with a chemotherapeutic agent improved antitumor efficacy in an HCC xenograft model [
All animal work was conducted according to relevant national and international guidelines.
Two HCC cell lines, HAK1-B [
Flag-caRheb plasmid was provided by Dr. Tomohiko Maehama (The Tokyo Metropolitan Institute of Medical Science, Tokyo, Japan).
Small hairpin RNA (shRNA) plasmids were obtained from iGENE therapeutics (USA).
The shRNA lipofectamine system was designed according to the manufacturer’s instructions. The target sequences of shRNA were as follows: sh-Raptor:
Q-PCR was used to detect mRNA levels of CYP3A4, Bmi1, EpCAM, FOXO3a, Raptor, and Rictor. A total of 1 × 105 HAK-1B cells were seeded in RPMI1640 medium containing 10% FBS for 24 h prior to experiments. BCAA (4 mM) was added to the medium and maintained for 72 h. RNA was isolated using TriPure isolation reagent (Roche Applied Science) and complementary DNA (cDNA) was synthesized using the High Capacity cDNA reverse transcription kit (Applied Biosystems, Carlsbad, CA). Real-time polymerase chain reaction (PCR) was performed using the 7500 Real-Time PCR System (Applied Biosystems) and Power SYBR Green PCR Master Mix (Applied Biosystems) containing specific primers, according to the manufacturer’s instructions. Each sample was normalized to GAPDH expression.
Primer sequences for PCR were as follows: CYP3A4: forward
FOXO3a: forward
Rictor: forward
Cells were cultured in the presence or absence of BCAA for 72 h, then fixed using 4% paraformaldehyde and stained with Hoechst 33342 (10 mg/mL, Dojindo, Tokyo, Japan) for 1 min. Cell counts were performed using the target activation protocol and an array scan system. EpCAM-positive cells were detected by staining the fixed cells with EpCAM-conjugated FITC antibody for 1 h, and the number of EpCAM-positive cells per 5000 cells were detected by the scan system
Cells were cultured with 5-FU (0, 1, 2 μg/mL) in the presence or absence of 2 mM BCAA for 72 h in 96-well plates (BD Biosciences). Annexin V binding to cell membranes was visualized with an Annexin V-FITC detection kit (TAKARA BIO INC, Tokyo, Japan) and Hoechst 33342 solution by array scan.
A total of 1 × 105 Huh7 cells were seeded in DMEM medium 24 h prior to study experiments. The medium was exchanged with LC medium for 3 days or DMEM medium with various treatments: knockdown for 5 days, overexpression for 1 day, and BCAA treatment for 30 min or 3 days. Western blotting was performed in the usual manner. Cells were washed in phosphate-buffered saline (PBS) and lysed in RIPA buffer containing complete protease and phosphatase inhibitor cocktail (Roche Applied Science, Indianapolis, INC). The membranes were blocked in Blocking One-P (Nacalai Tesque, Japan). Antibodies included rabbit anti-Rictor (Cell Signaling Technology, Beverly, MA), rabbit anti-raptor (Bethyl Laboratories, Montgomery, TX), rabbit anti-p-70S6 kinase, anti-total-p-70S6 kinase, rabbit anti-p-4EBP1 (T37/46), rabbit anti-p-Akt (T308 or S473), rabbit anti-p-GSK3β (S9), rabbit anti-β-catenin (Cell Signaling Technology, Beverly, MA), and mouse anti-α-tubulin (Sigma-Aldrich, St Louis, MO). Densitometry was conducted directly on the blotted membrane using a charge-coupled device camera system (LAS-4000 Mini, Fujifilm, Tokyo, Japan).
Huh-7 cells were transfected with negative control (NC) or target (mTORC1 or mTOR2 regulatory associated protein of Raptor and Rictor) small hairpin RNA (shRNA) using Lipofectamine™ LTX Reagent (Invitrogen Technology, USA) according to the manufacturer’s instructions. Stable transfectants were selected, as previously described [
The following experimental protocol was reviewed and approved by the Animal Care Committee of Ajinomoto Co., Inc. Female BALB/c nude mice or NOD/SCID mice at age 6 weeks were obtained from Charles River Japan (Yokohama, Japan). They were maintained in individual cages in a clean, air-conditioned room (24 ± 1°C) with a 12 h-12 h light-dark cycle (lights on from 0700 to 1900). The animals were fed a stock sterile powder diet (CRF-1, Oriental Yeast, Tokyo, Japan).
One million HAK-1B cells were suspended in 100 μL of RPMI1640 with FBS, and a subcutaneous injection was performed. The incidence and size of subcutaneous tumors were recorded when the average volume had reached 100 mm3 as previously described [
One million Huh7 cells transfected with lipofectal particles containing NC, Raptor, or Rictor shRNA, control plasmid cDNA (pcDNA), or Flag caRheb plasmid were harvested and injected subcutaneously into NOD/SCID mice (n = 5 in each group). Tumor volumes were evaluated as before. Mice bearing the NC, knockdown (KD), or over expression xenografts were sacrificed after 4 weeks.
Results were expressed as mean ± SE. Significance was determined in EXSUS version 7.7.1 (CAC Corporation, Tokyo, Japan) by performing Dunnett’s test, Tukey’s test and Student’s
HAK-1B cells were approximately 10% EpCAM-positive (
Dunnett's test, *p < 0.05, **p < 0.01 n = 6, mean ± SE.
The percentage of Annexin V-positive cells (C) and relative viable cell number (D) by array scan in HAK-1B cells cultured in RPMI1640 containing 10% FBS with or without 2 mM BCAA in the presence (1 or 2 µg/mL) or absence of 5-FU for 72 h by using target activation protocol of array scan. Student
The same results were obtained with Huh7 cells and are showed in
To test the hypothesis that EpCAM positivity increases chemotherapeutic sensitivity in the presence of BCAA, we used combined BCAA and 5-FU treatment in a BALB/c nude mouse model transplanted with subcutaneous HAK-1B.
Antitumor efficacy was significant with 5-FU alone, and an even stronger antitumor effect was achieved with combination BCAA and 5-FU treatment. However, the antitumor effect of BCAA alone was not remarkable (
The relative expression of mRNA of various molecules of each tumor was associated with CSC properties (C, D). Control: 10% DMSO/saline/tumor injection + 3% casein containing diet, 5-FU: 250 µg/tumor injection + 3% casein containing diet, BCAA diet: 10% DMSO/saline/tumor injection + 3% BCAA containing, 5-FU+BCAA diet: 250 µg/tumor injection + 3% BCAA containing diet for 14 days. Tukey’s test: *p < 0.05, **p < 0.01, ***p < 0.001 vs. control, #p < 0.05 vs. 5-FU, $$p < 0.01 vs. BCAA, n = 6, mean ± SE.
Considering the experimental transfection efficiency of Huh7 cells, we used them to evaluate the mechanism of BCAA, assuming that 40% of Huh7 were EpCAM-positive as previously reported [
The rate change of EpCAM-positive cells in 5000 cells with overexpression of caRheb or control plasmid cDNA (pc DNA) in control medium (DMEM containing 10% FBS) with and without 4 mM BCAA stimulation for 24 h in Huh7 (C).
The detection of P70S6 kinase phosphorylation, a member of downstream mTOR signaling, in the presence of DMEM, BCAA treatment, pretreatment with rapamycin and BCAA treatment, or LC stimulation for 72 h in Huh7 (A,B).
Tukey’s test **p < 0.01 vs. control, $$$p < 0.001 vs. BCAA, n = 8, mean ± SE (A).
Student t-test *p < 0.05, ****p < 0.0001, n = 8, mean ± SE (B,C).
Finally, EpCAM-positive cells were significantly reduced by mTORC1 activation via overexpression of caRheb, an activating factor of mTORC1 without BCAA stimulation (
The mTOR signals include two subtypes: mTORC1 and mTORC2 [
Rapamycin inhibits mTORC1 and mTORC2, depending on the duration of stimulation [
Dunnett's test, *p < 0.05, ***p <0.001 vs. control, n = 6, mean ± SE.
The relative expressions of EpCAM, c-myc, and FOXO3a mRNA upon Raptor and Rictor knockdown (B-D).
Dunnett's test, **p < 0.01, ***p < 0.001 vs. control n = 8, mean ± SE.
These findings suggest the presence or absence of EpCAM-positive cells depends on mTORC1 or mTORC2, respectively. We then examined the expression of EpCAM mRNA in the presence of mTORC1 or mTORC2 loss of function by knockdown of Raptor or Rictor, respectively. Loss of mTORC1 function caused EpCAM mRNA expression to increase similarly to mTORC1 inhibition by pretreatment with rapamycin for 1 h. However, loss of mTORC2 function caused EpCAM mRNA expression to decrease significantly. Furthermore, c-myc and FOXO3a mRNA expression reacted similarly to EpCAM mRNA expression (
Rictor, Raptor, and β-catenin were also associated with mTOR and Wnt/β-catenin signals (
Rictor or Raptor Knockdown compared to negative control (NC), caRheb compared to control plasmid cDNA (pc DNA), BCAA treatment compared to DMEM (FBS 10%) only (Ctrl) (A). The average tumor volumes and tumorigenesis ratio at the 4th week in a xenograft model with transplanted cells with negative control, knockdown of Raptor, Rictor for 5 days, or overexpression of control plasmid DNA, caRheb for 1 day (C), and tumorigenesis rate (B).
Dunnett's test, *p<0.05, ***p < 0.001 vs. N.C. n = 5, mean ± SE.
When mTORC1 was activated by caRheb overexpression or 4 mM BCAA treatment, the phosphorylation of p70S6 kinase increased and phosphorylation of Akt decreased. Furthermore, β-catenin protein was decreased by inhibition of phosphorylation of GSK3β via mTORC1 activation.
In addition, GSK3β was phosphorylated by Akt (Ser473) in response to Raptor knockdown; however, phosphorylation of GSK3β and β-catenin protein was not affected by Rictor knockdown.
The activation of mTORC1 or inhibition of mTORC2 was hypothesized to repress EpCAM-positive cells. The tumorigenic ability of CSCs was examined using a xenograft model subcutaneously implanted with Raptor or Rictor knockdown, control plasmid cDNA, or caRheb overexpressing Huh7 cells in mice for 4 weeks (
BCAA activated mTORC1; however, BCAA has dual effects on mTORC1. Stimulation with high doses of BCAA decreased EpCAM-positive cells via activation of mTORC1, while stimulation with low doses increased EpCAM-positive cells via inhibition of mTORC1 (
Although tumor size was smaller in the group implanted with Raptor knockdown cells than in the negative control group (
One of the novel contributions of this study was the finding that cancer cells experienced increased sensitivity to chemotherapy agents after EpCAM-positive cells differentiated by BCAA treatment via mTORC1 activation (
We also found that mTORC1 activation by BCAA treatment suppressed EpCAM-positive cells and enhanced the sensitivity of chemotherapeutic agents in HCC tumor (
The mechanism of the effects of BCAA treatment on EpCAM-positive cells was evaluated by accelerated differentiation to cancer cells via activation of mTORC1. Phosphorylation of GSK3β was inhibited by activating mTORC1 and the downstream Wnt/β-catenin signal, while phosphorylation of β-catenin was maintained, resulting in β-catenin degradation. It was inferred that this inhibited the nuclear translocation of β-catenin, and subsequently inhibited transcription of, c-myc, EpCAM, and Bmi1. Furthermore, the downstream signals of EpCAM were determined to have a c-myc transcription signal [
Moreover, we found that mTORC1 activation decreased the protein expression of Rictor via overexpression of caRheb (
The inhibition of mTORC2 function by Rictor knockdown led to the activation of p70S6 kinase, a signal of mTORC1, decreased EpCAM, c-myc, and FOXO3a expression, and phosphorylation of Akt, but had no effect on GSK3β and protein expression of β-catenin. It was not clear whether p70S6 kinase was directly activated by Rictor knockdown, although it is possible that mTORC1 activation and inhibition of Akt signals led to decreased EpCAM expression. Several studies have suggested that inhibition of mTORC2, and not mTORC1, suppresses carcinogenesis [
In this study, our findings suggest that CSC potential is strongly associated with mTORC signals. Furthermore, mTORC2 has a role in maintaining the stem cell potential whereas mTORC1 suppresses this potential (
Malnutrition or low Fischer ratio suppressed mTORC1 activation which led to increased cancer stemness. By contrast, mTORC2 maintains cancer stemness. However, mTORC1 inhibits mTORC2. BCAA suppresses cancer stemness by activation of mTORC1, and enhances antitumor effects of chemotherapy agents.
It is possible that activation alone of mTORC1 by BCAA treatment led to decreased Rictor protein expression, which was similar to Rictor knockdown, thus inhibiting mTORC2 and activating mTORC1. These findings suggest that mTORC2 may be a true cancer therapeutic target, especially in carcinogenesis related to CSC and the activation of mTORC1 alone leading to decreases in EpCAM and c-myc expression. Thus, we recommend that new cancer therapy strategies aim to inhibit mTORC2 and activate mTORC1 concurrently.
(TIF)
(TIF)
(TIF)
(TIF)