Transketolase Regulates the Metabolic Switch to Control Breast

Published OnlineFirst March 29, 2018; DOI: 10.1158/0008-5472.CAN-17-2906

Cancer Metabolism and Chemical Biology Research

Transketolase Regulates the Metabolic Switch to Control Breast Cancer Cell Metastasis via the a-Ketoglutarate Signaling Pathway Chien-Wei Tseng1,2, Wen-Hung Kuo3, Shih-Hsuan Chan1,2,4, Hong-Lin Chan5, King-Jen Chang6, and Lu-Hai Wang1,2

Abstract

Although metabolic reprogramming is recognized as a hall- mately suppressing breast cancer metastasis. Reduced TKT or mark of tumorigenesis and progression, little is known about addition of aKG mediated a dynamic switch of glucose metab- metabolic enzymes and oncometabolites that regulate breast olism from glycolysis to oxidative phosphorylation. Various cancer metastasis, and very few metabolic molecules have been combinations of the TKT inhibitor oxythiamine, docetaxel, and identified as potential therapeutic targets. In this study, the doxorubicin enhanced cell death in triple-negative breast cancer transketolase (TKT) expression correlated with tumor size in the (TNBC) cells. Furthermore, oxythiamine treatment led to 4T1/BALB/c syngeneic model. In addition, TKT expression was increased levels of aKG in TNBC cells. Together, our study has higher in lymph node metastases compared with primary tumor identified a novel TKT-mediated aKG signaling pathway or normal tissues of patients, and high TKT levels were associated that regulates breast cancer oncogenesis and can be exploited as with poor survival. Depletion of TKT or addition of alpha-keto- a modality for improving therapy. glutarate (aKG) enhanced the levels of tumor suppressors succi- Significance: These findings uncover the clinical significance nate dehydrogenase and fumarate hydratase (FH), decreasing of TKT in breast cancer progression and metastasis and demon- oncometabolites succinate and fumarate, and further stabilizing strate effective therapy by inhibiting TKT or by adding aKG. HIF prolyl hydroxylase 2 (PHD2) and decreasing HIF1a, ulti- Cancer Res; 78(11); 2799–812. 2018 AACR.

Introduction elevates the expression of glycolytic enzymes, including aldolase A, phosphoglycerate kinase 1, and pyruvate kinase (3). In addi- Patients with breast cancer have a 5-year survival rate over 90%; tion, a number of studies revealed that genetic defects in TCA however, for patients with distant metastasis, their survival rate cycle enzymes, such as succinate dehydrogenase (SDH) and decreases to only about 25% because of the lack of effective fumarate hydratase (FH), were also associated with tumor pro- strategies against breast cancer metastasis and recurrence (1). gression (4, 5). Tumor cells with altered metabolic program have high require- In this study, we used a proteomic approach to identify ments of glucose metabolism for rapid proliferation. Despite certain differentially expressed metabolic enzymes involved in some studies aiming at elucidating the correlation between tumor progression such as aldolase A (ALDOA), triose phos- aberrant metabolic behavior and tumor progression, how meta- phate isomerase (TPIS), a-enolase (ENOA), transketolase bolic processes regulate breast cancer cells growth and metastasis (TKT), and pyruvate dehydrogenase E1 (ODPB). Among them, is not fully understood. TKT is a metabolic enzyme involved in the nonoxidative branch A number of studies show that oncogenic signaling in cancers of the pentose phosphate pathway (PPP) and connects PPP drives metabolic reprogramming to generate large amounts of with glycolysis. Previous studies revealed that TKT was associ- biomass during rapid tumor growth (2). For example, HIF1a ated with metastasis of ovarian (6) and esophageal (7) cancers, as well as poor patient survival (6, 7). To date, no study has reported the effect of TKT-regulated metabolic signaling on 1Graduate Institute of Integrated Medicine, China Medical University, Taichung, tumor metastasis in breast cancer. Taiwan. 2Institute of Molecular and Genomic Medicine, National Health Research In this study, we reveal clinical signiﬁcance and regulatory Institutes, Zhunan, Miaoli County, Taiwan. 3Department of Surgery, National mechanism of TKT in progression and metastasis of breast cancer 4 Taiwan University Hospital, Taipei, Taiwan. Institute of Molecular Medicine, via alpha-ketoglutarate (aKG) signaling. TKT plays important National Tsing Hua University, Hsinchu, Taiwan. 5Institute of Bioinformatics and 6 roles in regulating dynamic switch of glucose metabolism. The Structural Biology, National Tsing Hua University, Hsinchu, Taiwan. Department a of Surgery, Taiwan Adventist Hospital, Taipei, Taiwan. combined therapy based on the new targets TKT or KG could be developed as an improved therapeutic approach for triple- Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). negative breast cancer (TNBC). Corresponding Author: Lu-Hai Wang, China Medical University, No. 91, Hsueh- Shih Road, Taichung 40402, Taiwan. Phone: 8864-2205-7153; Fax: 8864-2206- Materials and Methods 0248; E-mail: [email protected] Cell culture and transfection doi: 10.1158/0008-5472.CAN-17-2906 The human breast cancer MDA-MB-231, Hs578T and MCF-7 2018 American Association for Cancer Research. cells, and mouse breast cancer 4T1 cells were obtained from

www.aacrjournals.org 2799

Tseng et al.

ATCC. The 4T1 is a highly tumorigenic and invasive cell line University Hospital (Taipei, Taiwan). Other samples were from capable of metastasizing from the primary mammary gland commercial tissue arrays (US Biomax; SuperBioChips), including tumor to liver, lung, lymph nodes, and brain. The highly meta- 19 normal, 90 tumors, and 50 lymph node metastatic tissues. The static cell line MDA-MB-231-IV2-3 was previously established and slides were stained with mouse monoclonal anti-TKT antibody described (8). All cell lines were cultured in DMEM (Invitrogen) (clone 7H1AA1, ab112997; Abcam) using an automatic slide supplemented with 10% FBS (Biological Industries) at 37C with stainer BenchMark XT (Ventana Medical Systems). The staining 5% CO2. Cell lines were clear of Mycoplasma as determined by the intensities were evaluated and quantified by one pathologist Venor GeM kit (Minerva Biolabs) and were further authenticated (Pathology Core Lab, National Health Research Institutes) and in 2017 by Taiwan Bioresource Collection and Research Centre two independent investigators. The IHC scores of TKT for each (BCRC) using a short tandem repeat method. For transfection specimen were graded as follows: no expression, weak (þ); assay, cells were transfected with 20 mmol/L siTKT or 20 mmol/L moderate (þþ); and strong (þþþ).The expression levels of TKT siRNA control or TKT/pCMV plasmid (1 mg/mL) using Lipofecta- in tumor cells were quantified as a percentage. Paraffin-embedded mine RNAiMAX transfection reagent (Thermo Fisher Scientific). sections of tumor cells with TKTL1 overexpression (Origene, RG205218) were stained with mouse monoclonal anti-TKT anti- Protein extraction body (1 mg/mL, 1:75 dilution; clone 7H1AA1, ab112997; Cell samples were lysed in lysis buffer containing 7 mol/L urea, Abcam) or rabbit polyclonal anti-TKTL1 antibody (1 mg/mL, 2 mol/L thiourea, 4% w/v CHAPS, 10 mmol/L Tris-HCl pH 8.3, 1:75 dilution; clone N1C1, GTX109459; Genetex). We first used and 1 mmol/L EDTA. Protein lysates were extracted, sonicated, the D'Agostino and Pearson omnibus normality test to reveal that and centrifuged and the protein concentration was determined the quantitative results of IHC TKT expression were not Gaussian using Coomassie Protein Assay Reagent (Bio-Rad). distribution (P ¼ 0.0015). Thus, we used nonparametric Mann– Whitney test to analyze the quantitative results. 2-D DIGE gel image analysis and protein identification by MALDI-TOF-MS Proliferation assay The protein profiles of tumor tissues with 0.5, 1, and 2 cm in Cell proliferation was detected using CellTiter 96 Aqueous One size were analyzed using 2D differential gel electrophoresis Solution Cell Proliferation Assay (Promega). Assay was per- (DIGE). Protein samples were labeled with cyanine dyes Cy2, formed according to manufacturer's protocol. A total of 1.4 Cy3, and Cy5, and all procedures have been described previously 104 cells were cultured in a 24-well plate and incubated for (9, 10). The Cy-Dye–labeled 2-DE gels were visualized according different times. CellTiter 96 Aqueous One Solution reagent was to the previous report (10). For protein identification, the peptide added and incubated for 1 hour at 37C. The quantity of formazan mixture was loaded onto a MALDI plate and samples were product, proportional to living cell numbers, was measured at 490 analyzed using an Autoflex III mass spectrometer (Bruker Dal- nm using 96-well plate reader. Each experiment was performed in tonics) and parameters were described according to the previous triplicate and the shown data were mean SD. report (10). Cell invasion and migration assays Western blotting MDA-MB-231 and Hs578T cells were treated with 20 mmol/L Cells were lysed in the lysis buffer containing 7 mol/L urea, siTKT or 1 mmol/L aKG, or TKT/pCMV plasmid (1 mg/mL). After 2 mol/L thiourea, 4% w/v CHAPS, 10 mmol/L Tris-HCl (pH 8.3), 48 hours, these cells (1 105 cells) were seeded on Boyden 1 mmol/L EDTA, and phosphatase and protease inhibitors chamber, incubated for 8 hours, and then stained with 0.5% (Roche). Protein lysates were sonicated and centrifuged and the crystal violet dye. Cell invasion and migration were assayed in protein concentration was determined using protein assay kit 8-mm Falcon Cell Culture Inserts with or without Matrigel (BD (Thermo Fisher Scientific). The defined amount of final lysates Biosciences), respectively. All experiments were performed in was resolved in 8%–12% SDS-polyacrylamide gels, transferred triplicate. onto polyvinylidene difluoride membrane and probed with appropriate antibodies. Antibodies include rabbit polyclonal Soft agar colony formation assay anti-LDHA (GTX101416, Genetex), rabbit polyclonal anti–aKG MDA-MB-231 or MCF-7 cells at densities of 1 105 cells were dehydrogenase (clone C2C3, GTX105124, Genetex), rabbit poly- seeded in 6-well plate containing top layer of 0.4% agarose and clonal anti-SDH (GTX113833, Genetex), rabbit polyclonal anti- bottom layer of 0.6% agarose medium. The treatment group was FH (clone N2C2, GTX110128, Genetex), rabbit polyclonal anti- transfected with 20 mmol/L of siTKT for 48 hours. After one MDM2 (GTX100531, Genetex), mouse monoclonal anti-TKT month, colonies were stained with p-Iodonitrotetrazolium vio- (clone 7H1AA1, ab112997, Abcam), and mouse monoclonal let (1 mg/mL) for 48 hours and then counted. Data represent anti-PHD2 (clone 366G/76/3, Thermo Fisher Scientific). Mouse mean SD and the experiment was performed in triplicate. monoclonal anti-b-actin (clone SPM161, Santa Cruz Biotechnol- ogy) was used as the internal control, and protein expression Tail vein injection and orthotopic metastasis assays in levels were visualized with the Enhanced Chemiluminescence mouse models Detection Kit (Pierce Boston Technology) and exposed to X-ray To study the effects of aKG on tumor progression, MDA-MB-231 film. All experiments were repeated three times. cells (1 106) resuspended in 100-mLPBSwereimplanted orthotopically in 4th mammary fat pads of 8-week-old female IHC CB17-SCID mice (8). After implantation of MDA-MB-231 cells Paraffin-embedded matched normal, primary tumor, and for 24 hours, aKG reagent was intraperitoneally injected 3 times lymph node metastatic tissue sections of breast cancer specimens aweekuntil3months.aKG dissolved in PBS was used for (n ¼ 11) were provided by Dr. Wen-Hung Kuo, National Taiwan injection of 10 mg/kg each time. The tumor volume was

2800 Cancer Res; 78(11) June 1, 2018 Cancer Research

TKT Regulates Breast Cancer Metastasis via aKG Signaling

calculated by the formula: tumor volume (cm3) ¼ [length respiration, oligomycin treatment inhibited ATP synthase in mito- (cm) width (cm)2 0.5]. To study the effects of TKT chondria. FCCP, a proton ionophore in mitochondria, transports knockdown on tumor metastasis, MDA-MB-231-IV2-3 cells were protons across cell membranes to disrupt ATP synthesis. Finally, treated with 20 mmol/L siTKT. After 48 hours, MDA-MB-231-IV2-3 rotenone and antimycin are inhibitors for electric transport chain cells (1 106) resuspended in 100-mL PBS were injected per in mitochondria. mouse intravenously via tail veins into 6- to 8-week-old female CB17-SCID mice (BioLASCO). Tumor growth and metastasis Statistical analysis to individual organs were observed using live animal biolumi- Kaplan–Meier method (log-rank test) was used to analyze nescence imaging (BLI; Caliper IVIS system, PerkinElmer). Tumor survival data. Data were presented as mean SD. Student t test volume and weight were also measured at the end point. Cell was used to compare the differences between two experimental metastases were quantified by BLI signals of each mouse at the end groups and one-way ANOVA was used to compare the differences point. Animal experiments were approved by the Institutional among multiple groups using Tukey test in GraphPad. x2 test was Animal Care and Use Committee (IACUC). used to analyze the correlation between TKT levels and clinical factors; , P < 0.05; , P < 0.01; , P < 0.001. OCR and ECAR Orthotopic injection of stable TKT knockdown cells in mouse data were calculated by paired t test. model MDA-MB-231 cells were, respectively, transfected with two independent GFP-TKT/pCMV plasmids (Origene, NM001064). Results After 48 hours, these cells were selected by flow cytometry and Identification of metabolic proteins potentially involved in transfection efficiency was confirmed by Western blot analysis. breast cancer progression using proteomic analysis Stable shTKT cells were orthotopically injected at 1 106 cells per Using a proteomic approach and examining tumors of vary- mouse into 4th mammary fat pads of CB17-SCID mice (n ¼ 7) ing sizes, we attempted to identify differentially expressed and tumor volumes were recorded once a week during the 70 days proteins associated with breast cancer progression. We used period. Tumor volume ¼ 4/3pR3, R ¼ [length (cm) þ width syngeneicorthotopicimplantationof4T1cellsinBALB/cmice, (cm)]/2. Animal experiments were approved by IACUC. n ¼ 7. and tumors with 0.5, 1, and 2 cm in size were collected for further proteomic analyses. The protein profiles from the tumor TKT activity with0.5cminsizewerecomparedwiththosetumorswith1 MDA-MB-231 cells were transfected with 20 mmol/L siTKT. and 2 cm in size by two-dimensional protein gel analysis After 48 hours, these cells were lysed with 0.1 mol/L Tris- (Supplementary Fig. S1A–S1C). After spot detection and quan- HC1 buffer (pH 7.6), centrifuged, and the supernatant was tification from the two-dimensional gel images, a total of 21 collected (11). Supernatant (50 mL) was mixed with 200 mL differentially expressed proteins (P < 0.05) with 1.5-fold reaction mixture including 14.4 mmol/L ribose-5-phosphate, changes were chosen for further identification (Supplementary 190 mmol/L NADH, 380 mmol/L TP, >250 U/L glycerol-3- Fig. S1A–S1C) by using MALDI-TOF-MS and MASCOT data- phosphate dehydrogenase, and >6,500 U/L triose phosphate base (Supplementary Table S1). Three proteins related to isomerase (12). glycolysis were upregulated in the bigger tumors; they included Enzyme activity was detected at 340 nm. One unit of enzyme ALDOA, TPIS, and ENOA. Other metabolic enzymes included activity indicates the amount of enzyme catalyzing the oxidation the upregulated TKT involved in PPP and the downregulated of 1 mmol of NADH per minute. ODPB involved in pyruvate oxidation (Supplementary Table S1; Supplementary Fig. S1D). Tumors with 1 and 2 cm in size Metabolic assay had 1.5- and 2-fold, respectively, increased expression of TKT Oxygen consumption rate (OCR) is an indicator of mito- when compared with the 0.5-cm tumor (Supplementary Table chondrial oxidation and extracellular acidification rate (ECAR) S1; Supplementary Fig. S1D). is an indicator of lactate production that is equated to the glycolytic rate. OCR and ECAR were detected by XFe24 extra- TKT displays higher expression in metastatic lymph node cellular flux analyzer (Seahorse Bioscience). MDA-MB-231 cells tissues and patients with breast cancer with high TKT (7 104 cells) were cultured in X24 culture plate (Seahorse expression have poor overall survival Bioscience). OCR and ECAR were measured in XF base medium We analyzed TKT expression in normal and tumor tissues (Seahorse Bioscience). OCR was analyzed over time following according to gene expression arrays from Oncomine database injection of 1 mmol/L oligomycin, 2 mmol/L carbonyl cyanide- (Bild data). As compared with normal tissues, TKT displayed p-trifluoromethoxyphenylhydrazone (FCCP), and 0.5 mmol/L significantly higher expression in tumor tissues (Fig. 1A, nonpara- rotenone/antimycin. ECAR was measured over time follow- metric Mann–Whitney test, P ¼ 0.03). We also found that ing injection of 10 mmol/L glucose, 1 mmol/L oligomycin, and the levels of TKT in TNBC patients were significantly higher 50 mmol/L 2-deoxyglucose (2-DG). For ECAR, glucose than those in non-TNBC patients (Fig. 1B, P < 0.001). Kaplan– (10 mmol/L), oligomycin (1 mmol/L), and 2-DG (50 mmol/L) Meier survival curve (log-rank test) from Curtis 5-year overall were used to estimate glycolytic metabolism. Glucose treat- survival data showed that patients with higher TKT levels had ment could increase glycolytic metabolism in cells. 2-DG, a poorer 5-year survival than those with lower TKT levels (n ¼ synthetic glucose analogue, acted as a competitor for glucose 637, Fig. 1C; P ¼ 0.019, x2¼ 5.502, HR ¼ 1.3298). The similar and interfered with glucose metabolism. For OCR, oligomycin result is also observed in different clinical database (n ¼ 158; (1 mmol/L), carbonyl cyanide-4-(trifluoromethoxy)phenylhydra- Fig. 1D; P ¼ 0.003, c2 ¼8.7476, and HR ¼ 2.3131), suggesting that zone (FCCP; 1 mmol/L), and rotenone/antimycin (0.5 mmol/L) TKT has a prognostic potential. TNBC is the breast cancer subtype were used to estimate oxidative respiration. For mitochondrial with the poorest outcome; however, very few metabolic enzymes

www.aacrjournals.org Cancer Res; 78(11) June 1, 2018 2801

Tseng et al.

Figure 1. Clinical signiﬁcance of TKT in patients with TNBC. A, The expression levels of TKT in tumor (n ¼ 40) and normal (n ¼ 7) tissues were analyzed according to gene expression arrays in Oncomine database (nonparametric Mann–Whitney test, P ¼ 0.03). B, The levels of TKT in non- TNBC (n ¼ 1725) and patients with TNBC (n ¼ 250) from Curtis data were compared (, P < 0.001). C, Kaplan– Meier curve for TKT expression in association with 5-year survival of 637 patients with breast cancer. Patients were divided into high (blue line) and low (red line) TKT expression groups based on the mean þ SD levels among the patients analyzed (log-rank test, P ¼ 0.019). D, Kaplan–Meier curve for TKT expression in association with overall survival (n ¼ 158). Patients were divided into high (blue line) and low (red line) TKT expression groups based on the median levels among the patients analyzed (log-rank test, P ¼ 0.003). E, Kaplan–Meier curve for TKT expression in association with 5-year survival of 106 patients with TNBC among the 637 patients with breast cancer. Patients were divided into high (blue line) and low (red line) TKT expression groups based on the mean þ SD levels among the patients analyzed (log-rank test, P ¼ 0.0006). F, Representative pictures of TKT IHC from normal, primary tumor and lymph node metastatic tissues (scale bar, 1 mm; Supplementary Fig. S2 shows quantitative results).

as prognostic indicators for patients with TNBC are known. The tumor size (Supplementary Table S2). We also analyzed TKT role of TKT in patients with TNBC has not been reported; thus, expression in normal, primary tumor, and lymph node metastatic we further analyzed the correlation between TKT expression levels tissues by using IHC. First, we checked whether TKT antibody used and patients with TNBC 5-year overall survival. Among the 637 in the IHC staining cross-reacted with TKTL-1. To address this, we cases, there were a total of 106 patients with TNBC. Our analysis used TKTL1/pCMV plasmid to overexpress TKTL1 in MDA-MB- showed that patients with TNBC with higher TKT levels had poorer 231 cells. The overexpression efficiency was verified (Supplemen- 5-year overall survival than those with lower TKT levels (n ¼ tary Fig. S2A). The paraffin-embedded sections of tumor cells with 106; Fig. 1E; P ¼ 0.0006, x2 ¼ 11.7166, HR ¼ 2.3758), showing TKTL1 overexpression were stained with anti-TKT or anti-TKTL1 that TKT might have a prognostic potential in patients with TNBC, antibody. Our results displayed high staining intensity of TKTL1 in and it could play a role in TNBC progression. The clinicopatho- tumor cells overexpressing TKTL1 using anti-TKTL1, whereas logic features of TKT in patients with breast cancer from Curtis staining intensity using TKT antibody was insignificant (Supple- data showed that TKT levels were significantly associated with mentary Fig. S2B). These results suggest that TKT antibody used in some clinical factors, including stage, age, grade, type, TNBC, and the IHC staining does not cross-react with TKTL1.

2802 Cancer Res; 78(11) June 1, 2018 Cancer Research

TKT Regulates Breast Cancer Metastasis via aKG Signaling

The staining intensity of TKT was evaluated and quantified To further assess whether decreased lung metastasis by the ranging from no expression to the highest expression by a pathol- depletion of TKT resulted from decreased targeting of the tumor ogist and two independent investigators of our team. As summa- cells to lung, the cells transfected with the control or TKT siRNA rized in Supplementary Fig. S2C, a high percentage of normal were injected into CB17-SCID mice through tail vein (n ¼ 8) and tissues displayed insignificant TKT intensities (60%) or low inten- after 24 hours, the lung was perfused with PBS to flush out sities of TKT (30%) when compared with those of tumor tissues intravascular tumor cells and subsequently the expression (P < 0.001). Moreover, metastatic lymph node tissues displayed levels of human GAPDH reflecting the injected cells in lung a higher percentage of high intensities of TKT (56%) when tissues were measured. BLI analysis exhibited about equivalent compared with primary tumor (25%, P < 0.001). The represen- signals in the lungs of siCon or siTKT-transfected cells 30 minutes tative staining photographs are shown in Fig. 1F. The percentage after injection (Fig. 3D, P ¼ 0.294). The qPCR data confirmed of TKT expression in tumor cells, not including stroma cells, from the result (Fig. 3E, P ¼ 0.222). To confirm the inhibitory effects of primary tumor and lymph node metastatic tissue sections was transient TKT knockdown on tumor growth, two different knock- further quantified. Metastatic lymph node tissues displayed a down stable lines, MDA-MB-231-shTKT1 and MDA-MB-231- higher percentage of TKT expression in tumor cells when com- shTKT2, as well as MDA-MB-231-shNC line, were established, pared with the primary tumor (Supplementary Fig. S2D; P < and each (1 106 cells) were implanted orthotopically into 0.001). These results showed that TKT expression levels were the 4th mammary fat pad of CB17-SCID mouse (n ¼ 7). The the highest in lymph node metastases, suggesting that a possible knockdown efficiency of shTKT was verified (Fig. 3F) and the correlation of TKT levels with progression of metastasis in result showed that tumor sizes in both TKT knockdown groups breast cancer. were significantly smaller than those in the control group (Fig. 3G and H; Supplementary Fig. S3G–S3I). These findings indicated Downregulation of TKT suppresses metastatic functions that knockdown of TKT did not inhibit lung targeting (Supple- and affects cell-cycle distribution mentary Fig. S4A–S4D), but inhibited the subsequent lung col- To further elucidate the functional role of TKT, we manip- onization ability of the tumor cells. ulated TKT expression by siRNA depletion of TKT in MDA-MB- 231 and Hs578T TNBC cells (Fig. 2A). The downregulation of Identification of TKT-regulated metabolites in breast TKT in MDA-MB-231 cells resulted in significantly decreased cancer cells cell proliferation (Fig. 2B–D). This phenomenon was also Recent reports indicated the involvement of Warburg effect in observed in Hs578T cells (Fig. 2E–G). The inhibition by tumor metastasis and suggested that the molecules participating TKT knockdown in both cell lines was significantly rescued by in metabolic modulation were potential targets for antimetas- TKT/pCMV overexpression (Fig. 2B–G). Cell migration and tasis therapy (13). To address TKT-regulated metabolic path- invasion were carried out by transwell Boyden chamber assays. ways in breast cancer cells, we manipulated TKT expression by Downregulation of TKT led to a significant inhibition of inva- siRNA treatment of MDA-MB-231 cells for 48 hours and then sion (Fig. 2H) and migration (Fig. 2I) of MDA-MB-231 and cell lysates were harvested for identifying altered metabolites by Hs578T cells, whereas the inhibitory effects were almost LC-MS/MS (Waters Corporation). The differentially expressed completely rescued by TKT overexpression. MDA-MB-231 cells metabolites in siTKT-treated cells were identified when compar- with the inhibited TKT expression displayed reduced ability of ing with the siRNA control cells. Knockdown of TKT increased colony formation (Fig. 2J). TKT knockdown increased the some TCA-cycle intermediates including aKG (Fig. 4A) and P < percentage of cells in the G2–MphaseinMDA-MB-231and malate, while decreased succinate and fumurate ( 0.05). Hs578T cells (Fig. 2K). Taken together, these data suggested Reports indicated that the alternation of metabolites in the TCA that the depletion of TKT impaired tumor cell growth and cycle was associated with tumor formation (4). For example, metastasis-related abilities. succinate and fumarate accumulated in the mitochondria leaked out to the cytosol because of inactivation of the tumor sup- Knockdown of TKT suppresses lung metastasis of breast pressors SDH and FH, resulting in promoting cancer formation cancer cells (14). Currently, the potential role of aKG and the relationship To evaluate whether the depletion of TKT suppressed between TKT and aKG in TNBC are still unclear. Our findings cancer cell metastasis in vivo, we used tail vein injection of that TKT might play an important role in metastasis and its the highly invasive MDA-MB-231-IV2-3 cells (1 106 cells) knockdown led to increased aKG prompted us to further in CB17-SCID mice (n ¼ 8). The highly metastatic MDA-MB- investigate the potential effect of aKG in oncogenic behavior 231-IV2-3 sublines derived from the MDA-MB-231 parental of cancer cells. line were established and described previously (8). The MDA-MB-231-IV2-3 cells exhibited dramatically higher inva- aKG suppresses tumor cell growth, migration, and invasion siveness than the MDA-MB-231 parental cells in vitro and We further found that TKT overexpression attenuated aKG they also exhibited more aggressive lung and lymph node levels (Fig. 4B, P < 0.001), which was consistent with the result metastasis in vivo (8). The data from tail vein injection model from TKT knockdown. The physiologic concentration of aKG in showed that knockdown of TKT resulted in greatly decreased healthy brain tissues ranges from 1 to 3 mmol/L, whereas its lung metastasis of the MDA-MB-231-IV2-3 cells [Fig. 3A concentration is decreased to 100 to 300 mmol/L in gliomas (P ¼ 0.005) and B (P ¼ 0.002)] by BLI as also reflected in (15). IDH1-mutated tumor cells exhibited decreased aKG, hematoxylin and eosin staining (Fig. 3C). These findings leading to increased HIF1a levels (16). The similar results were indicated that knockdown of TKT inhibited lung metastasis observed in aKG derivatives treatments in IDH1-mutated gli- of the highly invasive breast cancer cells (Supplementary omas (17) or SDH-deficient tumor cells (18). Despite its tumor Fig. S3A–S3F). suppressor role of artificial aKG derivative in cancers, many

www.aacrjournals.org Cancer Res; 78(11) June 1, 2018 2803

Tseng et al.

Figure 2. Downregulation of TKT suppresses growth, invasion/migration and colony formation, and affects cell-cycle distribution of breast cancer cells. A, Twenty mmol/L siTKT reduced TKT expression in MDA-MB-231 and Hs578T cells, whereas its inhibitory effects were rescued by TKT/pCMV overexpression (1 mg/mL). The effects of TKT expression on cell proliferation in MDA-MB-231 (B–D) and Hs578T cells (E–G) were measured after siTKT or siTKT, and TKT/pCMV cotreatment for 24, 48, and 72 hours (, P < 0.05; , P < 0.01; and , P < 0.001). For invasion (H)andmigration(I) assays, MDA-MB-231 and Hs578T cells were treated with siTKT or siTKT, and TKT/pCMV cotreatment for 48 hours (, P < 0.001) and then incubated on Boyden chamber for 8 hours. J, For colony assay, 1 105 cells MDA-MB-231 or MCF-7 (Supplementary Fig. S4E) cells were transfected with siTKT. K, MDA-MB-231 and Hs578T cells were transfected with siTKT. Forty-eight hours later, tumor cells were harvested for analysis of cell-cycle distribution after propidium iodide staining. The percentage of cells was quantiﬁed by FlowJo 7.6 (, P < 0.05; , P < 0.001).

2804 Cancer Res; 78(11) June 1, 2018 Cancer Research

TKT Regulates Breast Cancer Metastasis via aKG Signaling

Figure 3. Knockdown of TKT does not inhibit early targeting to lung but suppresses lung metastasis of breast cancer cells. A, MDA-MB-231-IV2-3 cells (1 106 cells) were transfected with 20 mmol/L of siCon or siTKT. After 48 hours, 1 106 cells per mouse were injected intravenously into CB17-SCID mice via tail veins (n ¼ 8). Lung metastases as reflected by amount of cancer cells in lung in vivo (A) and ex vivo (B) were quantified using BLI signal (n ¼ 8). C, Images show hematoxylin and eosin staining of lung metastases. More detailed data are shown in Supplementary Fig. S3A–S3F. Scale bar, 1 mm. T, tumor cells in the lung. D, MDA-MB-231-IV2-3 cells (1 106 cells) transiently transfected with siCon or siTKT were injected at 1 106 cells per mouse into CB17-SCID mice via tail veins (n ¼ 8). BLI images showed lung metastasis of tumor cells in siCon and siTKT-treated mice 30 mins after injection (P ¼ 0.294). E, Twenty-four hours after injection, the mice were perfused with PBS to rid of blood and lung tissues were harvested. Specific qPCR primers for human GAPDH were used to detect injected cells in lung tissues and mouse actin mRNA was used as the internal control (P ¼ 0.222). F, Knockdown efficiency of shTKT in the two independent stable lines, MDA-MB-231-shTKT-1 and MDA-MB-231-shTKT-2 were confirmed when compared with the control group (stable MDA-MB-231-shNC cells). G and H, Stable shTKT cells were orthotopically injected at 1 106 cells per mouse into 4th mammary fat pads of CB17-SCID mice (n ¼ 7; G) and tumor volumes (H) were recorded once a week during the 70 days period (, P < 0.001).

studies revealed that non-aKG derivative could attenuate cell with the control (Fig. 4C). Furthermore, treatment of aKG led proliferation of colon cancer (19) and reduces the levels of to a significant inhibition of cell invasion (Fig. 4D) and migra- VEGF and erythropoietin through decreasing HIF1a,thereby, tion (Fig. 4E). MCF-7 cells with the inhibited TKT expression inhibiting angiogenesis ability of the Hep3B hepatoma cells displayed reduced ability of colony formation (Supplementary (20). These findings suggested the potential tumor suppressing Fig. S4E). TKT overexpression promoted cell proliferation in role of aKG. Furthermore, G-protein–coupled receptor GPR99 MDA-MB-231 (Fig. 4F–H) and Hs578T (Supplementary Fig. was reported to function as a receptor for the TCA cycle S4F–S4H) cells, whereas its effect was substantially reversed by intermediate aKG (21). Although, previous studies indicated aKG treatment. These findings indicate that aKG can impair that aKG–dependent dioxygenases signaling pathways func- metastatic-related abilities of breast cancer cells. We further tioned as tumor suppressors (22), the regulatory role of aKG in verified that the promotion of TKT on invasion (Fig. 4I) and breast cancer is unclear. Treatment of aKG resulted in signif- migration (Fig. 4J) in MDA-MB-231 and Hs578T cells was icantly decreased MDA-MB-231 cell growth when compared substantially reversed by aKG treatment, suggesting that TKT

www.aacrjournals.org Cancer Res; 78(11) June 1, 2018 2805

Tseng et al.

Figure 4. aKG inhibits growth, lymph node, and lung metastases of breast cancer cells in CB17-SCID mice. MDA-MB-231 cells were treated with siTKT (A) or TKT/pCMV (B). After 48 hours, their effects on aKG levels were measured by LC-MS. C, MDA-MB-231 cells were treated with or without 100 or 1000 mmol/L aKG for 18, 24, and 48 hours and cell growth was measured using MTS assay. , P < 0.05; , P < 0.01; , P < 0.001. For invasion (D)andmigration(E)assays, MDA-MB-231 cells were treated with 1 mmol/L aKG (treatment) for 48 hours and then incubated on Boyden chamber for 8 hours. Each experiment was repeated three times. TKT overexpression promoted cell proliferation of MDA-MB-231 (F–H) and Hs578T (Supplementary Fig. S4F–S4H). A total of 1 mmol/L aKG treatment decreased the phenomenon. TKT overexpression promoted cell invasion (I)andmigration(J), whereas its effects were decreased by aKG treatment. , P < 0.05; , P < 0.01; , P < 0.001. MDA-MB-231 cells were orthotopically injected with 1 106 cells per mouse into 4th mammary fat pads of CB17-SCID mice. The next day, the mice were intraperitoneally injected with aKG (10 mg/kg) or PBS control three times per week. The images of tumor cells in tumors (K) and various organs, including spleen, lung, liver, and lymph node (N) from individual mice (n ¼ 10) were monitored by BLI signal. Representative BL1 images are shown after 3 months of continuous treatment with PBS or aKG. Tumor weight (L) and tumor volume (M) quantiﬁcations in aKG or PBS control were measured. , P < 0.05; , P < 0.01; , P < 0.001.

2806 Cancer Res; 78(11) June 1, 2018 Cancer Research

TKT Regulates Breast Cancer Metastasis via aKG Signaling

regulated invasion and migration of tumor cells via aKG TKT regulates tumor suppressors SDH and FH signaling signaling. In this study, we observed the cellular levels of pathways aKG were increased after the treatment of aKG (Supplementary Our data showed that knockdown of TKT decreased the Fig. S5A and S5B). expression levels of metabolites succinate and fumarate (Fig. 5C). Previous studies indicated that the inactivation aKG suppresses lung metastasis of breast cancer cells mutations in SDH and FH led to abnormal accumulation of We next assessed the effect of this metabolic pathway on metabolites succinate and fumarate in TCA cycle, which in turn tumor growth and metastasis using a mouse model. A total of inhibited PHD and increased HIF1a in tumors (4, 5). The 1 106 MDA-MB-213 cells were implanted orthotopically correlation between TKT, SDH, and FH in breast cancer is still into mammary fat pads of CB17-SCID mice (n ¼ 10). One unclear; thus, we investigated the effects of TKT knockdown on day after implantation, intraperitoneal aKG (10 mg/kg) the expression levels of SDH and FH. We found that knock- administration was started three times a week for 3 months. down of TKT increased the levels of SDH and FH (Fig. 5A), BLI data revealed that aKG treatment led to a significant leading to decreased levels of succinate and fumarate and thus reduction of primary tumor growth (Fig. 4K, P < 0.001). There stabilizing the PHD2-regulated signaling pathway. were significant differences in the weights (Fig. 4L, P ¼ 0.024) Previous studies reported aKG–dependent dioxygenases sig- and sizes (Fig. 4M, P ¼ 0.004) of primary tumors between naling pathways functioning as tumor suppressors (22). In control and the aKG–treated groups after 3 months. Individual addition, SDH and FH have been reported to be targets of organ metastases were also examined, and we found that aKG aKG–dependent dioxygenases, including JmjC domain- treatment significantly diminished lung and lymph node containing histone demethylase (KDMs) and DNA demethy- metastases (Fig. 4N, P < 0.05). Overall, our data for the first lases (27). These studies suggest that TKT may control tran- time demonstrated that TKT-mediated aKG signaling sup- scriptional regulation of SDH and FH via aKG–dependent pressed growth and metastases of breast cancer. dioxygenases. To elucidate the potential underlying mechanism, we detected the effects of TKT depletion or aKG treat- TKT regulates breast cancer metastasis via the aKG ment on RNA levels of SDH and FH. Our results showed that signaling pathway TKT depletion (Fig. 5D, P < 0.001) or aKG treatment (Fig. 5E, To further explore TKT-regulated downstream pathways in P < 0.01) indeed increased RNA levels of SDH and FH, suggest- breast cancer metastasis, the effects of TKT on the aKG and ing regulation at the transcriptional level. Overall, the likely TCA-cycle enzymes were examined. Previous studies indicated regulatory mechanism of TKT via aKG signaling in breast cancer that accumulation of aKG enhanced the activity of PHD and metastasis is depicted in Fig. 5F. subsequent destabilization of its downstream target HIF1a (23). To assess the relationship between TKT and HIF1a in Reduced TKT or aKG treatment regulates glucose metabolism MDA-MB-231 cells, the impact of TKT on PHD2 was investi- and mitochondrial oxygen consumption gated. Results revealed that downregulation of TKT enhanced Tumor cells predominantly metabolize glucose through gly- PHD2 expression (Fig. 5A) and this phenomenon was also colysis instead of oxidative phosphorylation in TCA cycle to observed in the aKG–treated cells (Fig. 5B). Moreover, knock- rapidly produce ATPs and nucleic acid building stones for down of TKT reduced HIF1a expression (Fig. 5A), suggesting supporting their high rate of growth (28). The effect of TKT that TKT affected HIF1a expression via the PHD2 signaling on metabolic activities in cancers was unclear; thus, we exam- pathway. HIF1a has been reported to be associated with tumor ined the relationship among glycolysis, mitochondrial meta- metastasis (24) and is known to be a transcription factor bolism, and TKT signaling. The knockdown efficiency of TKT in regulating the expression of LDHA (25). Other studies revealed MDA-MB-231 cells was initially estimated (Fig. 6A). TKT that knockdown of LDHA inhibited breast cancer metastasis knockdown (Fig. 6B, P < 0.001) or aKG treatment (Fig. 6C, (26). Currently, the relationship between TKT and LDHA is not P < 0.001) exhibited decreased ECAR. TKT knockdown (Fig. 6D, known; thus, we further assess the effect of TKT on LDHA P < 0.001) or aKG treatment (Fig. 6E, P < 0.001) elevated OCR. expression. Our data showed that knockdown of TKT inhibited These results demonstrated that reduced TKT led to switch of LDHA expression (Fig. 5A) and this phenomenon was also glucose metabolism from glycolysis to mitochondrial respira- observed in aKG–treated cells (Fig. 5B). These results suggested tion via the aKG signaling pathway. that TKT decreased LDHA expression and promoted HIF1a To further confirm whether knockdown of TKT drove the degradation through the aKG signaling pathway, leading to switch of glucose metabolism from glycolysis to TCA cycle, we the inhibition of breast cancer metastasis. Our data suggest that used mass spectrometry to measure expression levels of meta- a regulatory network of those metabolites and their corre- bolites in glycolysis and TCA cycle. Reduction of TKT dimin- sponding catalyzing enzymes are involved in the regulation of ished the levels of glycolytic metabolites including glucose-6- breast cancer metastasis. phosphate (G6P), pyruvate, and lactic acid, while increased the Previous study indicates that L-2HG dehydrogenase TCA-cycle metabolites including aKG and malate (Supplemen- (L2HGDH) and D-2HG dehydrogenase (D2HGDH) prevent tary Fig. S5A). We treated the cancer cells with aKG and oncometabolites L-2HG and D-2HG from accumulating in nor- observed a similar result like TKT knockdown (Supplementary mal cells, respectively, by converting them back to aKG (22). Fig. S5B), suggesting that reduction of TKT drove the switch of We have found that TKT depletion enhanced the levels of glucose metabolism from glycolysis to mitochondrial metab- L2HGDH and D2HGDH (Fig. 5A). Overall, these results indicate olism at least in part through the aKG signaling pathway. that TKT depletion enhances L2HGDH and D2HGDH levels, To further verify this, the effect of decreased TKT on the resulting in the increase of aKG and PHD2 levels and thereby expression levels of metabolic enzymes in TCA cycle was promoting HIF1a degradation. evaluated. We found that the depletion of TKT resulted in

www.aacrjournals.org Cancer Res; 78(11) June 1, 2018 2807

Tseng et al.

Figure 5. TKT and aKG reversely regulate glucose metabolic enzymes. Knockdown of TKT (A)oraKG treatment (B)signiﬁcantly altered the expression of TCA-cycle enzymes. C, LC-MS data showed reduction of TKT decreased the levels of succinate and fumarate. The effects of TKT knockdown (D)oraKG treatment (E) on RNA levels of SDH and FH were measured by qPCR. GAPDH served as the internal control. , P < 0.01; , P < 0.001. F, Model of breast cancer cell metastasis suppressed by downregulation of TKT via aKG and SDH and FH commonly mediated signaling pathways.

2808 Cancer Res; 78(11) June 1, 2018 Cancer Research

TKT Regulates Breast Cancer Metastasis via aKG Signaling

Figure 6. Knockdown of TKT or aKG addition affects glucose metabolism and mitochondrial oxygen consumption. A, Knockdown efﬁciency of siTKT in MDA-MB-231 cells was conﬁrmed. Reduced TKT or aKG addition decreased glycolytic metabolism (ECAR; P < 0.001; B and C) while it increased OCR (D and E; P < 0.001). The ECAR and OCR values were normalized with 7 104 MDA-MB-231 cells per well.

increased expression levels of metabolic enzymes in TCA cycle are unclear. In this study, we first assessed the effect of including aconitase, aKG dehydrogenase, SDH, FH, and malate oxythiamine on TKT activity according to previous study dehydrogenase (Supplementary Fig. S6A). The similar result (12). Tumor cells were treated with 5 mmol/L oxythiamine was obtained in aKG–treated cells (Supplementary Fig. S6B). In for 48 hours. Our results revealed that TKT activity was contrast, the depletion of TKT resulted in decreased levels of significantly reduced by oxythiamine treatment (Fig. 7A, P < glycolytic enzymes including PKM2, HK, and PFK (Supplemen- 0.01). In addition, we found oxythiamine treatment elevated tary Fig. S6C), and the similar results were also observed in the levels of aKG in MDA-MB-231 (Fig. 7B, P < 0.001) and aKG–treated cells (Supplementary Fig. S6C). Taken together, Hs578T (Fig. 7C, P < 0.001) cells as expected, suggesting that these results indicate that reduced TKT leads to the alteration of oxythiamine suppressed tumor growth could in part through glucose metabolism by switching it from glycolytic to mito- the aKG signaling pathway. Then, we analyzed whether oxy- chondrial metabolism via the elevation of metabolic enzymes thiamine treatment affected growth of breast normal cells. The in TCA cycle through the aKG signaling pathway. As tumor cells results showed that cell viabilities of nontumorigenic human depend on glycolysis for their rapid growth, inhibition of TKT breast epithelial cell line H184 for 24 (P ¼ 0.16), 48 (P ¼ or addition of aKG could be used as a modality for developing 0.08), and 72 hours (P ¼ 0.07) were not significantly cancer therapeutics not only for breast cancer including triple- decreased by 5 mmol/L oxythiamine treatment when com- negative breast cancer as shown in this study, but for other pared with those without oxythiamine treatment (Fig. 7D), typesofcanceraswell. indicating there was no significant side effects of oxythiamine in human breast normal cells. We observed that docetaxel or Oxythiamine in combination with docetaxel and/or doxorubicin treatment increased aKG levels (Fig. 7E, P < doxorubicin enhances inhibitory effects of TNBC cells 0.001). Moreover, previous studies reported that docetaxel or Docetaxel and doxorubicin are commonly used drugs for doxorubicin treatment attenuated HIF1a levels (31, 32), TNBC, but their efficiencies are limited as a result of the further supporting our findings that TKT affects HIF1a expres- development of drug resistance. Oxythiamine inhibits TKT sion via aKG signaling. Thus, we tested the inhibitory effects of and thus could lead to downregulation of glycolysis, not oxythiamine in combination with docetaxel and/or doxorubi- targeted by the above two drugs. Thus combinatory treatment cin on cell proliferation. We treated TNBC cell lines MDA-MB- of oxythiamine together with the two drugs may enhance the 231 (Fig. 7F–H) and Hs578T (Fig. 7I–K) with 5 mmol/L killing effect of cancer cells. Oxythiamine, an antimetabolite oxythiamine, 1 mmol/L docetaxel, 1 mmol/L doxorubicin and thiamine analogue, induces cell apoptosis and suppresses oxythiamine in combination with docetaxel and/or doxorubi- tumor cell growth in cancers by targeting TKT (29, 30). cin for 24, 48, and 72 hours. Treatment of oxythiamine had Although some studies indicate that oxythiamine can suppress significant inhibitory effects for 24 (Fig. 7F and I) and 48 hours tumor progression, the effects of oxythiamine in breast cancer (Fig. 7G and J) in both cell lines. Although treatment of

www.aacrjournals.org Cancer Res; 78(11) June 1, 2018 2809

Tseng et al.

Figure 7. Oxythiamine in combination with docetaxel and/or doxorubicin enhances inhibitory effects on TNBC cell viability. A, MDA-MB-231 cells were treated with 5 mmol/L oxythiamine (OT). After 48 hours, the effect of oxythiamine on TKT activity was measured. The effects of 5 mmol/L oxythiamine treatment on the levels of aKG for 24 hours in MDA-MB-231 (B)andHs578T(C) cells. D, The effects of OT treatment on viabilities of nontumorigenic human normal breast cell line H184 for 24 hours (P ¼ 0.16), 48 (P ¼ 0.08,) and 72 hours (P ¼ 0.07) were assessed. E, The effects of docetaxel (Doc) or doxorubicin (Dox) on the levels of aKG were measured by LC-MS (, P < 0.001). The effects of oxythiamine in combinationwithdocetaxeland/ordoxorubicinoncellviabilitiesofMDA-MB-231(F–H)andHs578T(I–K) were assessed. Cell viabilities for 24 (F and I), 48 (G and J), and 72 hours (H and K)weremeasured.

docetaxel or doxorubicin had inhibitory effects of TNBC cells, Discussion the killing effects of oxythiamine combining with docetaxel or Increasing evidence suggests that some pivotal genes, includ- doxorubicin could be strengthened in TNBC cells, In addition, ing HIF1a, are able to regulate certain enzymes to induce combining of the three drugs had maximum killing effects metabolic reprogramming in cancers. HIF1 has been reported (>90% decrease) for 72 hours in both TNBC cell lines (Fig. 7H to induce glycolytic enzymes, including aldolase A, phospho- and K). These ﬁndings indicate that oxythiamine could glycerate kinase 1, and pyruvate kinase (3). HIF1a regulates enhance the two-drug sensitivities of TNBC cells.

2810 Cancer Res; 78(11) June 1, 2018 Cancer Research

TKT Regulates Breast Cancer Metastasis via aKG Signaling

dynamic switch from oxidative to glycolytic metabolism by MDM2 levels had better RFS than those with lower MDM2 activating glucose transporters and glycolytic enzymes (33). (N ¼ 3951, P ¼ 0.0019, Supplementary Fig. S6K). As both Certain metabolic enzymes involved in glucose transport, gly- PKM2 and MDM2 could regulate HIF1a stability, our results colysis and lipid metabolism are targets of HIF1a (34). In our suggest that aside from the TKT/aKG–mediated regulation of study, we found that TKT depletion promoted HIF1a degra- PHD2 and HIF1a degradation, PKM2 and MDM2 could also dation via aKG signaling. These results suggest that TKT-medi- play a role in TKT-mediated control of HIF1a stability. ated signaling pathways may collaborate to regulate dynamic aKG functions as a cosubstrate for Fe (II)/aKG–dependent switch of glucose metabolism. Xu and colleagues (11) reported dioxygenases, including KDMs and the TET (ten-eleven trans- that TKT reduced oxidative stress and played important roles in location) family of DNA hydroxylases (27). They catalyze glycolysis and glutathione synthesis in hepatocellular carcino- hydroxylation in diverse substrates including proteins, alky- ma (HCC) cells. TKT knockdown attenuated NADPH produc- lated DNA/RNA and 5-methylcytosine (5mC) of genomic DNA tion and led to the increase of reactive oxygen species (ROS; (27). TET family of DNA hydroxylases catalyzes a three-step ref. 11). TKT knockdown decreased glucose flux, and purine oxidation reaction to convert 5mC to 5-carboxylcytosine metabolites including AMP, ADP, ATP, and GTP (11). Together, (5caC) and subsequent decarboxylation of 5caC, leading to these results provide evidence that TKT may play an important DNA demethylation (27). SDH and FH have been reported to role in metabolic reprogramming in tumors. be the targets of aKG–dependent dioxygenases, including The emerging evidence demonstrates that several TCA cycle KDMs and DNA demethylases (27). Our results showed that enzymes are tumor suppressors, such as SDH and FH, and their TKT depletion or aKG treatment increased RNA levels of SDH genetic defects are associated with tumorigenesis. The inacti- and FH. Together, these studies suggest that TKT may control vation mutations in SDH and FH leads to abnormal accumu- transcription of SDH and FH via aKG–dependent dioxygenases lation of metabolites succinate and fumarate in TCA cycle, and signaling. the subsequent inhibition of PHD and enhancement of HIF1a TKT inhibitor oxythiamine had been reported to have antican- pathways in tumors (4, 5). Here, we have demonstrated that cer activity (29, 30). For example, oxythiamine in combination reduction of TKT augments levels of SDH, FH, and PHD2, but with sorafenib had enhanced effects on HCC cell growth by in vivo decreased levels of HIF1a. In addition, levels of oncometabo- assay (11). Despite its potential therapeutic development, at lites succinate and fumarate are significantly reduced by TKT present, the targeted therapy of TKT against TNBC cells has not knockdown, which is likely due to increased levels of SDH and been reported. Our results showed that the combinations of FH, which in turn affects PHD2 stabilization and HIF1a oxythiamine with docetaxel and doxorubicin had maximum degradation. HIF1a is a transcription factor regulating the inhibitory effects in TNBC cells, suggesting combinatory drug expression of LDHA (25) and its knockdown inhibits breast treatment as a novel therapy against TNBC. Our study for the first cancer metastasis (26). We have also noticed that knockdown time revealed that oxythiamine treatment elevated the levels of of TKT decreases levels of LDHA, suggesting that reduction of aKG in TNBC cells, suggesting that oxythiamine suppressed TKT resulted in decreased HIF1a and LDHA via elevated levels tumor cell growth via aKG signaling pathway. Together, it is of SDH and FH, leading to the inhibition of tumor metastasis. feasible to develop a combinatory drug treatment with the con- Previous reports indicate that a glycolytic enzyme pyruvate ventional therapeutic drugs to improve treatment benefits for kinase M2 (PKM2) is a transcriptional coactivator for HIF1, TNBC. amplifying HIF1 activity via a positive feedback regulation, and thereby promoting cancer progression (35). To date, the under- Disclosure of Potential Conflicts of Interest lying mechanism of TKT-mediated regulation of PKM2 via aKG No potential conflicts of interest were disclosed. signaling is unclear. We found that TKT depletion or aKG treatment reduced PKM2 levels (Supplementary Fig. S6C) and Authors' Contributions promoted HIF1a degradation. A significant positive correlation Conception and design: L.-H.Wang,C.-W.Tseng,H.-L.Chan,and existed between TKT and PKM2 (r ¼ 0.4635, P < 0.0001, K.-J. Chang Supplementary Fig. S6D). Patients with breast cancer (N ¼ Development of methodology: L.-H. Wang, C.-W. Tseng, S.-H. Chan, and P < H.-L. Chan 3951, 0.001, Supplementary Fig. S6E) including patients Acquisition of data (provided animals, acquired and managed pati- N ¼ P ¼ with TNBC ( 255, 0.045, Supplementary Fig. S6F) with ents, provided facilities, etc.): C.-W. Tseng, W.-H. Kuo, S.-H. Chan, and higher TKT and PKM2 levels had poorer recurrence-free survival H.-L. Chan (RFS) than those with lower TKT and PKM2. We also observed Analysis and interpretation of data (e.g., statistical analysis, biostatistics, that patients with breast cancer (N ¼ 3951, P < 0.001, Sup- computational analysis): L.-H. Wang, C.-W. Tseng, W.-H. Kuo, S.-H. Chan, plementary Fig. S6G) including patients with TNBC (N ¼ 255, and H.-L. Chan P ¼ Writing, review, and/or revision of the manuscript: L.-H. Wang and 0.0049, Supplementary Fig. S6H) with higher TKT, PKM2, C.-W. Tseng a and HIF1 levels had poorer RFS than those lower TKT, PKM2 Administrative, technical, or material support (i.e., reporting or organiz- and HIF1a. On the other hand, a study indicated that p53 ing data, constructing databases): L.-H.Wang,C.-W.Tseng,S.-H.Chan, induced tumor suppressor MDM2 E3-ubiquitin-mediated deg- and H.-L. Chan radation of HIF1a (36). To date, the underlying mechanism of Study supervision: L.-H. Wang, C.-W. Tseng, and K.-J. Chang TKT-mediated regulation of MDM2 via aKG signaling is not known. We found that TKT depletion or aKG treatment Acknowledgments enhanced MDM2 levels (Supplementary Fig. S6I) and promot- We thank the Protein Chemistry Core Lab, Pathology Core Lab, and Cell a fi Sorter Core Lab of the National Health Research Institutes for mass ed HIF1 degradation. A signi cant negative correlation existed spectrometric analysis, H&E and IHC staining, and technical assistance of between TKT and MDM2 (r ¼0.2618, P < 0.0001, Supple- cell cycle, respectively. This study has been supported by Ministry of Science mentary Fig. S6J). Patients with breast cancer with higher and Technology (MOST), Taiwan (MOST 104-2320-B-039-055-MY3, MOST

www.aacrjournals.org Cancer Res; 78(11) June 1, 2018 2811

Tseng et al.

104-2320-B-039-054-MY3, and MOST 106-2811-B-039-004), and National advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate Health Research Institutes (NHRI06A1-MGPP09-014) grants. this fact.

The costs of publication of this article were defrayed in part by the Received September 27, 2017; revised January 31, 2018; accepted March 8, payment of page charges. This article must therefore be hereby marked 2018; published ﬁrst March 29, 2018.

References 1. Ci Y, Qiao J, Han M. Molecular mechanisms and metabolomics of natural pseudohypoxia in succinate dehydrogenase-deficient cells. Mol Cell polyphenols interfering with breast cancer metastasis. Molecules 2016;21, Biol 2007;27:3282–9. pii:E1634. 19. Rzeski W, Walczak K, Juszczak M, Langner E, Pozarowski P, Kandefer- 2. Schulze A, Harris AL. How cancer metabolism is tuned for proliferation and Szerszen M, et al. Alpha-ketoglutarate (AKG) inhibits proliferation of colon vulnerable to disruption. Nature 2012;491:364–73. adenocarcinoma cells in normoxic conditions. Scand J Gastroenterol 3. Semenza GL, Roth PH, Fang HM, Wang GL. Transcriptional regulation of 2012;47:565–71. genes encoding glycolytic enzymes by hypoxia-inducible factor 1. J Biol 20. Matsumoto K, Imagawa S, Obara N, Suzuki N, Takahashi S, Nagasawa T, Chem 1994;269:23757–63. et al. 2-Oxoglutarate downregulates expression of vascular endothelial 4. Selak MA, Armour SM, MacKenzie ED, Boulahbel H, Watson DG, Mans- growth factor and erythropoietin through decreasing hypoxia-inducible field KD, et al. Succinate links TCA cycle dysfunction to oncogenesis by factor-1alpha and inhibits angiogenesis. J Cell Physiol 2006;209:333–40. inhibiting HIF-alpha prolyl hydroxylase. Cancer Cell 2005;7:77–85. 21. He W, Miao FJ, Lin DC, Schwandner RT, Wang Z, Gao J, et al. Citric acid 5. Yang M, Soga T, Pollard PJ. Oncometabolites: linking altered metabolism cycle intermediates as ligands for orphan G-protein-coupled receptors. with cancer. J Clin Invest 2013;123:3652–8. Nature 2004;429:188–93. 6. Ricciardelli C, Lokman NA, Cheruvu S, Tan IA, Ween MP, Pyragius CE, 22. Losman JA, Kaelin WG Jr. What a difference a hydroxyl makes: mutant IDH, et al. Transketolase is upregulated in metastatic peritoneal implants and (R)-2-hydroxyglutarate, and cancer. Genes Dev 2013;27:836–52. promotes ovarian cancer cell proliferation. Clin Exp Metastasis 2015; 23. Vatrinet R, Leone G, De Luise M, Girolimetti G, Vidone M, Gasparre G, et al. 32:441–55. The alpha-ketoglutarate dehydrogenase complex in cancer metabolic plas- 7.ChaoYK,PengTL,ChuangWY,YehCJ,LiYL,LuYC,etal.Transke- ticity. Cancer Metab 2017;5:3. tolase serves a poor prognosticator in esophageal cancer by promoting 24. Zhao T, Zhu Y, Morinibu A, Kobayashi M, Shinomiya K, Itasaka S, et al. HIF- cell invasion via epithelial–mesenchymal transition. J Cancer 2016;7: 1-mediated metabolic reprogramming reduces ROS levels and facilitates 1804–11. the metastatic colonization of cancers in lungs. Sci Rep 2014;4:3793. 8. Chan SH, Huang WC, Chang JW, Chang KJ, Kuo WH, Wang MY, et al. 25. Miao P, Sheng S, Sun X, Liu J, Huang G. Lactate dehydrogenase A in cancer: MicroRNA-149 targets GIT1 to suppress integrin signaling and breast a promising target for diagnosis and therapy. IUBMB Life 2013;65:904–10. cancer metastasis. Oncogene 2014;33:4496–507. 26. Rizwan A, Serganova I, Khanin R, Karabeber H, Ni X, Thakur S, et al. 9. Lai TC, Chou HC, Chen YW, Lee TR, Chan HT, Shen HH, et al. Secretomic Relationships between LDH-A, lactate, and metastases in 4T1 breast and proteomic analysis of potential breast cancer markers by two-dimen- tumors. Clin Cancer Res 2013;19:5158–69. sional differential gel electrophoresis. J Proteome Res 2010;9:1302–22. 27. Xiao M, Yang H, Xu W, Ma S, Lin H, Zhu H, et al. Inhibition of alpha-KG- 10. Chou HC, Lu YC, Cheng CS, Chen YW, Lyu PC, Lin CW, et al. Proteomic dependent histone and DNA demethylases by fumarate and succinate that and redox-proteomic analysis of berberine-induced cytotoxicity in breast are accumulated in mutations of FH and SDH tumor suppressors. Genes cancer cells. J Proteomics 2012;75:3158–76. Dev 2012;26:1326–38. 11. Xu IM, Lai RK, Lin SH, Tse AP, Chiu DK, Koh HY, et al. Transketolase 28. Soga T. Cancer metabolism: key players in metabolic reprogramming. counteracts oxidative stress to drive cancer development. Proc Natl Acad Sci Cancer Sci 2013;104:275–81. U S A 2016;113:E725–34. 29. Wang J, Zhang X, Ma D, Lee WP, Xiao J, Zhao Y, et al. Inhibition of 12. Zhao Y, Wu Y, Hu H, Cai J, Ning M, Ni X, et al. Downregulation of transketolase by oxythiamine altered dynamics of protein signals in transketolase activity is related to inhibition of hippocampal progenitor pancreatic cancer cells. Exp Hematol Oncol 2013;2:18. cell proliferation induced by thiamine deficiency. BioMed Res Int 2014; 30. Zhao F, Mancuso A, Bui TV, Tong X, Gruber JJ, Swider CR, et al. Imatinib 2014:572915. resistance associated with BCR-ABL upregulation is dependent on HIF- 13. Doppler H, Storz P. Differences in metabolic programming define the site 1alpha-induced metabolic reprograming. Oncogene 2010;29:2962–72. of breast cancer cell metastasis. Cell Metab 2015;22:536–7. 31. Oh ET, Kim CW, Kim SJ, Lee JS, Hong SS, Park HJ. Docetaxel induced-JNK2/ 14. King A, Selak MA, Gottlieb E. Succinate dehydrogenase and fumarate PHD1 signaling pathway increases degradation of HIF-1alpha and causes hydratase: linking mitochondrial dysfunction and cancer. Oncogene 2006; cancer cell death under hypoxia. Sci Rep 2016;6:27382. 25:4675–82. 32. Blagosklonny MV, An WG, Romanova LY, Trepel J, Fojo T, Neckers L. p53 15. Thirstrup K, Christensen S, Moller HA, Ritzen A, Bergstrom AL, Sager inhibits hypoxia-inducible factor-stimulated transcription. J Biol Chem TN, et al. Endogenous 2-oxoglutarate levels impact potencies of 1998;273:11995–8. competitive HIF prolyl hydroxylase inhibitors. Pharmacol Res 2011; 33. Semenza GL. HIF-1: upstream and downstream of cancer metabolism. Curr 64:268–73. Opin Genet Dev 2010;20:51–6. 16. XuW,YangH,LiuY,YangY,WangP,KimSH,etal.Oncometabolite2- 34. Keith B, Johnson RS, Simon MC. HIF1alpha and HIF2alpha: sibling rivalry hydroxyglutarate is a competitive inhibitor of alpha-ketoglutarate- in hypoxic tumour growth and progression. Nat Rev Cancer 2011;12:9–22. dependent dioxygenases. Cancer Cell 2011;19:17–30. 35. Luo W, Semenza GL. Pyruvate kinase M2 regulates glucose metabolism by 17. Zhao S, Lin Y, Xu W, Jiang W, Zha Z, Wang P, et al. Glioma-derived functioning as a coactivator for hypoxia-inducible factor 1 in cancer cells. mutations in IDH1 dominantly inhibit IDH1 catalytic activity and induce Oncotarget 2011;2:551–6. HIF-1alpha. Science 2009;324:261–5. 36. Ravi R, Mookerjee B, Bhujwalla ZM, Sutter CH, Artemov D, Zeng Q, et al. 18. MacKenzie ED, Selak MA, Tennant DA, Payne LJ, Crosby S, Frederiksen Regulation of tumor angiogenesis by p53-induced degradation of hypoxia- CM, et al. Cell-permeating alpha-ketoglutarate derivatives alleviate inducible factor 1alpha. Genes Dev 2000;14:34–44.

2812 Cancer Res; 78(11) June 1, 2018 Cancer Research

Transketolase Regulates the Metabolic Switch to Control Breast Cancer Cell Metastasis via the α-Ketoglutarate Signaling Pathway

Chien-Wei Tseng, Wen-Hung Kuo, Shih-Hsuan Chan, et al.

Cancer Res 2018;78:2799-2812. Published OnlineFirst March 29, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-17-2906

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2018/03/29/0008-5472.CAN-17-2906.DC1

Cited articles This article cites 36 articles, 9 of which you can access for free at: http://cancerres.aacrjournals.org/content/78/11/2799.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/78/11/2799.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/78/11/2799. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.