BCAT1 expression associates with ovarian cancer progression: possible implications in altered disease metabolism.

Previously, we have identified the branched chain amino-acid transaminase 1 (BCAT1) gene as notably hypomethylated in low-malignant potential (LMP) and high-grade (HG) serous epithelial ovarian tumors, compared to normal ovarian tissues. Here we show that BCAT1 is strongly overexpressed in both LMP and HG serous epithelial ovarian tumors, which probably correlates with its hypomethylated status. Knockdown of the BCAT1 expression in epithelial ovarian cancer (EOC) cells led to sharp decrease of cell proliferation, migration and invasion and inhibited cell cycle progression. BCAT1 silencing was associated with the suppression of numerous genes and pathways known previously to be implicated in ovarian tumorigenesis, and the induction of some tumor suppressor genes (TSGs). Moreover, BCAT1 suppression resulted in downregulation of numerous genes implicated in lipid production and protein synthesis, suggesting its important role in controlling EOC metabolism. Further metabolomic analyses were indicative for significant depletion of most amino acids and different phospho- and sphingolipids following BCAT1 knockdown. Finally, BCAT1 suppression led to significantly prolonged survival time in xenograft model of advanced peritoneal EOC. Taken together, our findings provide new insights about the functional role of BCAT1 in ovarian carcinogenesis and identify this transaminase as a novel EOC biomarker and putative EOC therapeutic target.

INTRODUCTION

Epithelial ovarian cancer (EOC) accounts for 4% of all cancers in women and is the leading cause of death from gynecologic malignancies [1]. Despite treatment improvements, long-term survival rates for patients with advanced disease remain disappointing [2]. The molecular basis of EOC initiation and progression is still poorly understood. To establish novel therapeutic and diagnostic strategies against this deadly disease, it is essential to understand its molecular pathology.

Recently, the importance of epigenetic perturbation of gene regulation in cancer [3], including EOC [4], has begun to be more fully appreciated. The most studied epigenetic alteration is DNA methylation, the addition of a methyl moiety to the cytosine-5 position within the context of a CpG dinucleotide, mediated by DNA methyltransferases [3]. Similar to other malignancies, aberrant DNA methylation occurs in EOC and contributes to ovarian tumorigenesis and mechanisms of chemoresistance [4]. Applying a more global array-based technology, several studies have demonstrated that DNA methylation changes in ovarian cancer are cumulative with disease progression and chemotherapy (CT) resistance [5-7]. Using a similar approach (methylated DNA immunoprecipitation coupled to CpG island tiling arrays) we have recently shown that DNA hypermethylation occurs in all (including less invasive/early) stages of ovarian tumorigenesis. Interestingly, advanced EOC was exclusively associated with DNA hypomethylation of a number of oncogenes, implicated in cancer progression, invasion/metastasis and probably chemoresistance [8]. The cytosolic form of the branched chain amino-acid transaminase 1 (BCAT1) was among the genes identified to be notably hypomethylated in low-malignant potential (LMP) and high grade (HG) serous EOC tumors [8]. BCAT enzymes comprise two isoforms: BCAT1 (cytosolic) and BCAT2 (mitochondrial). Both isoforms regulate the first step in degradation of branched-chain amino acids (BCAAs) including leucine, isoleucine and valine, which are essential for cellular metabolism and growth [9]. While BCAT2 is expressed in most tissues, BCAT1 expression is rather restricted to some highly specialized tissues, including brain, ovary and placenta [9]. Several groups have confirmed that BCAT1 is involved in cell proliferation, cell cycle progression, differentiation and apoptosis [10, 11]. Recent studies were indicative for the important role of BCAT1 in the progression of several malignancies, including medulloblastomas, nonseminomas, colorectal and nasopharyngeal carcinomas and glioblastomas [12-16]. Moreover, the role of BCAAs metabolism in cancer pathogenesis has been also a topic of interest [17].

This prompted us to further investigate if BCAT1 displays elevated expression levels in serous EOC tumors with different malignant potential, and whether this gene is functionally implicated in EOC dissemination. Here we present experimental data indicative for strong BCAT1 overexpression in LMP and HG serous EOC tumors, which probably correlates with its hypomethylated status. We also show that blocking BCAT1 expression inhibits proliferation, migration, invasion and peritoneal dissemination, possibly through altering EOC metabolism.

RESULTS

Overexpression of BCAT1 in both LMP and HG serous EOC tumors

Previously, we have identified the BCAT1 gene as hypomethylated in LMP and HG EOC tumors, when compared to normal tissues [8]. Here, we further evaluated BCAT1 protein expression by immunohistochemistry (IHC) in serous EOC tumors and ovarian normal tissue samples, using tissue microarrays (TMAs). Our TMAs included triplicate cores of 117 serous EOC tumors, including 13 LMP tumors and 104 HG ovarian tumors. Thirteen normal ovarian tissue samples were also included as controls. Table 1 shows the major clinical characteristics of these patients for whom extensive follow-up clinical data (up to 5-years) were available. The age ranged from 41 to 83 years (median: 66 years). High-grade tumors were all grade 3 (100%) including stage III (69%) and stage IV (31%) tumors. The majority of patients (93%) received a combination of platinum and paclitaxel. The median baseline CA125 was around 800. Forty percent of the patients had a progression or a recurrence within the first 6 months of follow-up; for 39% of the patients the progression-free survival (PFS) interval was in the range of 7 to 24 months, and 21% of the patients displayed PFS values higher than 25 months (Table 1).

Table 1

Detailed patients’ clinicopathological characteristics

Variable	Range	n/total	%
Age (years)
	≥ 65	64/130	49
	< 65	66/130	51
Median	66
Tissue/tumor type
	Normal	13/130	10
	LMP	13/130	10
	High-grade	52/130	40
	OM	52/130	40
Grade
	3	104/104	100
Stage
	III	72/104	69
	IV	32/104	31
Chemotherapy1
	platinum+taxol	97/104	93
	Other	13/104	7
CA125
	≥ 800	47/104	45
	< 800	53/104	55
PFS (months)2
	0–6	41/103	40
	7–24	40/103	39
	> 25	22/103	21

All patients were subjected to adjuvant therapy.

Extended follow-up, including PFS values, were available for 103 patients.

OM - omental metastasis

As seen from Figure 1, BCAT1 displayed significantly higher expression in LMP tumors and HG serous EOC tumors, when compared to normal tissues (p = 0.0003 and p = 0.0014, respectively), which correlates with BCAT1 hypomethylation status in these tumor types. However, we did not observe any significant differences between the levels of BCAT1 expression and patients’ PFS values (p = 0.0901; see Supplemental Figure 1), which suggests that staining intensity for BCAT1 in pre-treatment surgical specimens is not predictive of PFS.

Figure 1

Analysis of BCAT1 expression in serous EOC tumors by IHC

A. Representative IHC images of BCAT1 protein expression in normal ovarian tissues, low-malignant potential (LMP) tumors and high-grade (HG) tumors. B. Box-plot presentation of BCAT1 protein expression levels in normal ovarian tissues, LMP tumors and HG tumors.

Phenotype analysis of BCAT1 suppression in EOC cells: possible implications in EOC cell proliferation, cell cycle control, migration and invasion

Next, we decided to verify if short-hairpin RNA (shRNA)-mediated BCAT1 gene knockdown could produce any cancer-related phenotypic changes in EOC cells. We tested several EOC cell lines for endogenous BCAT1 protein expression by Western blot analysis (see Supplemental Figure 2). Among these, the SKOV3 and the A2780s cell lines displayed strong BCAT1 expression and were further used to generate stably transfected shRNA-BCAT1 clones. Clone selection for further analyses was based on semi-quantitative (sq) RT-PCR and Western blot validation of the BCAT1 mRNA/protein expression in selected clones, compared to empty vector-transfected clones. Among the clones analyzed, the SKOV3 shRNA-BCAT1 knockdown clones sh-B1 and sh-B2 and the A2780s clones sh-A1 and sh-A2 displayed significant decrease of BCAT1 expression levels compared to the mock-transfected control (Supplemental Figure 2) and were selected for further analyses.

We investigated the impact of the BCAT1 gene suppression on SKOV3 cell proliferation, cell cycle control, migration, invasion, and sensitivity to cisplatin and paclitaxel (drugs, conventionally used for first-line EOC therapy). The BCAT1 gene knockdown led to sharp decrease of the number of viable adherent cells (represented by cell index), compared to control cells (Figure 2A). This observation was further supported by the reduced number of colony formation upon BCAT1 suppression (Figure 2B and 2C). Moreover, BCAT1 depletion induced S cell cycle arrest which could explain the drastic reduction in the proliferation rates of these EOC cells observed earlier (Figure 3). Additionally, BCAT1 suppression significantly inhibited both migration and invasion of SKOV3 cells. Indeed and as shown in Figures 4A and 4B (migration) and Figures 4C and 4D (invasion), the numbers of SKOV3 cells that passed through the filter using the sh-B1 and sh-B2 clones were remarkably less than in the negative control (Ctrl) clone.

Figure 2

ShRNA-directed knockdown of the BCAT1 expression in SKOV3 cells

A. effect of BCAT1 knockdown clones 1 and 2 (sh-B1 and sh-B2) on cell proliferation, compared to the control clone (Ctrl); B. Representative images of colony formation assays following BCAT1 knockdown; C. Quantitative determinations (graph-bars) of data obtained: results are expressed as numbers of sh-B1- and sh-B2-induced colony formation compared to the Ctrl-induced colony formation numbers. Differences were determined using the Student's t-test. Error bars denote mean ± SEM; *indicates statistical significance (p < 0.05)

Figure 3

Effect of BCAT1 suppression on cell cycle control in SKOV3 cells

Cell-cycle profile was examined by flow cytometry and percentages of cells in G0/G1, S, and G2/M phase in the shRNA-BCAT1 clone 1 (sh-B1) were compared to the mock-transfected control (Ctrl) clone. Propidium iodide staining shows an increased fraction of cells in the S-phase and corresponding decrease of cells in both G1- and G2/M-phase at 9 and 12 h after removing hydroxyurea in the shRNA-BCAT1 clone B1 (sh-B1), when compared with the control clone (Ctrl).

Figure 4

Effect of BCAT1 knockdown on SKOV3 cell migration and invasion

A. Migration was assessed using Boyden-chamber assay. Cells from the shRNA-BCAT1 clones 1 and 2 (sh-B1 and sh-B2) and the control (Ctrl) clone were seeded into the upper chambers in 0.1% FBS containing medium at a density of 2.5 × 104 per well, and 600 μl of 1% FBS containing medium was placed in the lower chamber as a chemoattractant. After 24 h at 37°C in 5% CO2, the cells were fixed with cold methanol and stained with blue trypan solution. Migrated cells on the underside of the filter were photographed and counted by phase contrast microscopy. C. Cell invasion was assayed in a similar way, as the upper chambers were coated with Matrigel. Here, NIH3T3 conditioned medium was added in the lower chamber as a chemoattractant (see Materials and Methods for details). All experiments were performed in triplicate. For each experiment, cell number was calculated as the total count from 10 random fields per filter (at magnification × 40). The bar graphs in panels B. and D. represent quantitative determinations of data obtained by selecting 10 random fields per filter under phase contrast microscopy and results are expressed as % change of the sh-G1 and sh-G2 clones over the Ctrl clone. Differences between shRNA-BCAT1-transfected and vehicle-transfected SKOV3 cells were determined by a Student's t-test; error bars denote mean ± SEM; *indicates statistical significance (p < 0.05).

Quite similar results were obtained when performing functional analyses in A2780s cells following BCAT1 knockdown (see Supplemental Figure 3).

Finally, BCAT1 knockdown had no significant impact on SKOV3 cisplatin and paclitaxel sensitivity (see Supplemental Figure 4).

Molecular mechanisms of BCAT1 action in EOC cells

To better understand the molecular mechanisms of BCAT1 action in EOC cells, we compared the gene expression of the previously selected shRNA-BCAT1 SKOV3 clones (sh-B1 and sh-B2) against the corresponding control clone (Ctrl). All microarray experiments were performed in duplicates, as two hybridizations were carried out for each of the two BCAT1 knockdown cell clones against the corresponding control, using a fluorescent dye reversal (dye-swap) technique. For both comparisons, a subset of common differentially expressed genes was selected by initial filtering on confidence at p-value ≤0.05, followed by filtering on expression level (≥2 fold). Using these stringent selection criteria, we found 878 genes were upregulated and 979 genes were downregulated in SKOV3 cells upon BCAT1 knockdown; among these, the BCAT1 gene displayed significant medium suppression value (−8.34 fold; see Supplemental Table 1B). Table 2 shows a list of selected functionally related groups of genes that were differentially expressed (≥2.0-fold at p value ≤0.05) in SKOV3 cells following BCAT1 suppression. As seen from Table 2, genes with previously shown implication in mechanisms of metabolism, signal transduction, regulation of transcription, transport, cell growth & proliferation and cell cycle, were predominantly or exclusively suppressed, while BCAT1 knockdown was related with the induction of some genes related with immune & inflammatory response and cell adhesion. Supplemental Table 1 shows the complete list of the differentially expressed genes (≥2.0-fold at p value ≤0.05) following BCAT1 knockdown in SKOV3 cells.

Table 2

Selected differentially expressed gene groups in SKOV3 cells upon BCAT1 knock-down

	A. Upregulated genes
metabolism	ZNF404, ZNF556, ZNF608, ZNF626, ZNF654, ZNF781, AGBL5, CTSV, CTSZ, MIPEP, ZDHHC3, ADAMTS1, ADAMTS5, ADAMTS6, MMP13, PLAU, PROC, PRSS23, SENP2, C3orf38, USP46, LMLN, MME, RPL14, RPL15, RPL21, RPL24, TRIM6, IFT172, SPINT2, TCEAL5, TCEAL8, INSIG1, INSIG2, RBM3, NAP1L1, NAP1L2, RPL22, RPL27A, LACTB, DZIP1L, FSTL1, USPL1, ZC3H13, HNRNPA3, HNRNPH3, JRK, KLF8, LSM3, MID1, CXXC5, PHF11, SACS
regulation of transcription	CDKN2D, FAM172A, LMCD1, LPXN, PEX14, SERTAD2, ANKRD10, ARID5B, CTNNB1, ELL2, ETS1, FOXP1, HDAC6, HMGB1, HSBP1, JUN, KDM3A, KLF6, KLF7, MED4, MEIS1, MNT, MPHOSPH8, MSL3, MYCBP, NFKBIZ, NFYA, PFDN5, POU5F1, PPP1R13L, RBPJ, RUNX1T1, RUNX2, SAP18, SAP30BP, SMARCA4, SNAI1, SNAI2, STAT1, TAF9B, TARDBP, TBX3, TCEB2, TCF4, TSHZ3, WWTR1, EIF4H, EIF5, EIF5A2, MRPL43, RPSA, PAIP2B
signal transduction	PTPRB, PTPRS, SIRPA, PDCL, PDE11A, PDE4A, PDE4DIP, PLCB4, ARL4C, CETN2, GNAI2, GNB4, MRAS, RAB9B, RAP2A, RAP2B, CSF2, CXCL12, EDN1, IL11, TNFSF12, WNT5A, GPR126, GPRC5A, S1PR1, AKT3, CAMKK1, COL4A3BP, DYRK2, IP6K1, IP6K2, JAK1, MAP2K3, MYLK, NRBP2, PIK3CB, PRKAB2, PRKCA, SIK1, TRIO, CDK6, CLK1, NDRG1, NUCKS1, ACVR2B, EFNA5, EPHA4, THRB, ARHGAP1, ARHGAP23, GFRA1, JMJD6, LRPAP1, TNFRSF19
transport	PLIN2, VAMP8, PLIN3, STXBP6, CYB5B, CYBRD1, KCTD10, ANO7, CATSPER1, GABRQ, TRPM2, AP2M1, COG6, COPZ2, GGA2, PCTP, S100A6, SEC13, SEC22A, SLC9A9, STEAP3, SYT15, TMED9, UCP2, UCP3, APOL1, ATP13A3, ECM1, RBP4, TINAGL1, KPNA4, NUTF2, APOL3/APOL4, FAM101A/ZNF664-FAM101A, MFSD1, MFSD9, S100A3, SNX12, ABCA1, ABCC4, ABCC5, ATP2B4, ATP8A2, ITGB1BP1, PDZK1, SLC16A3, SLC2A1, SLCO2B1, TM9SF2
immune & inflammatory response	PSG1, PSG11, PSG3, PSG4, PSG6, PSG7, PSG9, IL7R, SRGN, PSG8, IGFN1, ZC3HAV1
cell growth	ACTC1, LOXL1, LOXL2, FGF5, GAS6, GDF6, INHBA, PTN, TGFB2, NRG1, FN1, KRT7, KRT8
cell adhesion	MEGF8, CDH13, CDH15, CDH2, CDH6, NKTR, PCDH20, MCAM, ACTC1, ACVR2B, CTNNB1, LMO7, MRAS, MYH10, MYL6, PTEN, TCF4
	B. Downregulated genes
metabolism	A4GALT, AADAT, AARS, ACAA2, ACOT1, ACOT2, ACOX2, AGPAT9, AHCYL1, AKR1B10, AKR1C1/AKR1C2, AKR1C3, ALDH3A1, ALDH3A2, ALDH9A1, ALOX5, AMACR, ASNS, ASPH, ASRGL1, BCAT1, CA2, CARS, CBR4, CBS, CES1, CHST6, COX5A, CYP2J2, DGAT2, DHX58, DIMT1, EBP, ECI1, EGLN3, EPRS, FA2H, FUT8, GALNT12, GALNT14, GARS, GATM, GCLC, GDA, GMPR2, GNPNAT1, GOT1, GPD1L, GPT2, GPX1, GSR, GSTT2/GSTT2B, HGD, HIBADH, HS3ST1, IARS, IDH1, IDH2, ISYNA1, KYNU, LRAT, MAN2A2, MARS, MGAT3, MPST, MTHFD2, MTHFD2L, MTHFS, NAPRT1, NCF2, NDUFC1, NDUFS1, NEDD4, NMNAT2, NMNAT3, PAPSS2, PDE8B, PGD, PHGDH, PIGF, PLCD3, PLCG2, PNPLA4, POLRMT, PPCDC, PSAT1, PTGS1, PTGS2, PTRH2, PYGB, RAB27B, RASD1, RHOU, RNF135, RPL4, SARS, SEPSECS, SHMT2, SMOX, SOD2, SORD, SQRDL, SRD5A3, SRXN1, ST3GAL1, ST6GAL1, SULT1A1, SULT1A2, SULT1A3/SULT1A4, TDO2, TGM2, TMOD1, TMX3, TRIM2, TST, UBB, UBC, UGT2B10, UGT2B7, UQCRC2, WARS, YARS, ZNRF1, DDHD1, ENPP5, LOXL4, PLA2G7, SDF2, USP53, VNN3, ARL4D, DDX52, ELAC1, FBXO25, NEDD8, PNP, RBBP8, RDH10, TARS, TOP2A, TRIM69, UGDH, UHRF2, ABHD3, CTAGE5, DDAH2, DHRS13, DHRS7, DHRSX, FAAH2, GLT8D2, GNPDA2, GPX8, HECTD1, VNN2, DLL1, GBP5, GCNT3, HS6ST2, NLGN4X, PDE6A, PIGP, SULF2, BLMH, CLGN, CPE, LONP1, MEST, PSMA3, PSMA6, PSME2, SENP6, UCHL1, USP18, USP3, C1R, C1S, CPD, PCSK5, PCSK6, PRSS2, PSMC6, CELA3B, CPVL, ADAM23, ANPEP, CPM, ECEL1, SPPL2A, TMPRSS3, GRSF1, JAKMIP1, MRPL46, RPL19, RPL23A, RPS29, SLPI, AHRR, CDC37L1, MKKS, MRPS35, MRPS36, UGT2B28, FGG, LGALS3, SERPINA3, SERPINB5, VWF, HIST1H1A, KLF14, RBPMS2, TRIM31, GRPEL2, CPEB4, ZBED3, FBXO33, LARP6
signal transduction	IBSP, ICAM2, SPRY4, ADAP2, ARHGEF2, NGEF, CLEC2B, CNIH1, FLOT2, PDLIM4, CISH, EFHD1, ERRFI1, RALGAPA1, SH3BGR, SHC2, TMEFF2, TMEM33, TMTC2, AIM1, NMB, NMU, SMOC2, IFRD1, RADIL, RASEF, ANXA1, PECAM1, SNTB1, DAAM1, GKAP1, GRB10, JKAMP, RGS16, S100P, IGFBP2, IGFBP5, MNAT1, NKD2, CGRRF1, GAREM, SAMD13, SAMD5, CCL26, IL18, IL1A, IL7, SCG2, SPP1, TSLP, VAV3, ANXA3, BIRC3, RAB39B, CCR7, F2RL2, FZD1, FZD8, GPR135, GPR64, LGR6, LPHN3, OR7E24, RXFP1, TAPT1, CAMK2D, DAPK2, ITPK1, ITPKA, MAP2K1, MAP2K5, MAP2K6, MAP3K5, MAPK6, MARK1, NME3, PCK2, PFKM, PIK3R1, PRKCH, PXK, SGK2, ULK1, NEK1, SRP72, STK39, TRIB3, HKDC1, MAST4, NME5, STK32A, BMPR1B, DDR2, ERBB2, PDGFRL, RIPK2, AR, ESR1, NR0B1, NR1D1, PGR, PPARG, RARA, ARHGAP24, ARHGAP25, ARHGAP26, PEX12, TMC5, TMEM128, TMEM170A, TMEM199, TMEM87A, ACP6, DUSP6, EYA4, INPP4B, NT5C3A, PPM1A, PPP1R1A, PPP1R1B, PSPH, PTP4A2, SGPP2, PTPDC1, DUSP5, EYA2, PPM1E, UBLCP1, DUSP14, ALPP, KTN1, CHRNA3, CLDN3, EPHA10, HLA-B, HLA-E, HLA-F, IL18R1, IL1R2, IL6ST, ITGA6, ITGB2, MSR1, PLXNC1, SFRP1, THBD, TLR3, TNFRSF11B, UNC5B
regulation of transcription	ANKRA2, CRLF3, GAS7, AJUBA, ATF3, ATF4, BATF3, CEBPA, CEBPB, CEBPG, DDIT3, DTX1, E2F5, EHF, ETV4, FOS, FOXA1, FOXC1, FOXP2, FOXQ1, GATA5, GATA6, GLI1, GSC, GTF2A2, HDAC4, HEXIM1, HEY1, HEYL, HIF1A, HLF, HOXC9, HOXD1, HOXD8, ID1, ILF2, IRF2BP2, JDP2, JMY, KCTD1, KLF4, KLF5, MECOM, MEF2A, MESP1, MSX1, MXD1, NAB1, NCOA7, NEUROD2, NFE2L1, NUPR1, ONECUT2, OVOL2, PATZ1, PIR, POLR3E, PRRX2, RELB, RUNX3, SIRT1, SMAD5, SNAPC5, SOX17, STRN3, SUPT6H, TBX1, TBX15, TCEA1, TCF12, TFCP2L1, TOX2, TP63, VAV1, XBP1, ZFP36L2, ZFX, ZNF143, ZNF397, ZSCAN16, ZSCAN31, NEO1, EEF1E1, EIF2S2, EIF3C/EIF3CL, EIF4EBP1, EIF1
transport	VPS37A, ITSN2, SVOPL, RIMS4, CACNA2D2, CACNG6, CLIC6, JPH3, KCNG1, KCNK1, KCNK5, PKD2, TRPC6, ABCB6, AP2B1, AP3S1, AP3S2, AP4S1, ATP6V0D2, CCT6B, COX18, ETFA, GOSR1, PITPNC1, SCFD1, SLC27A2, STARD3, TAP1, TIMM9, AMBP, LAMB3, SLC39A8, TC2N, FAM63B, SLC22A15, ABCC2, ABCC8, ABCG1, ATP10B, CDH23, EXOC6, MAL2, PDPN, SLC12A7, SLC22A16, SLC22A4, SLC22A5, SLC24A3, SLC30A1, SLC38A1, SLC38A2, SLC3A2, SLC47A1, SLC7A11, SLC7A5, SLCO4A1, STEAP1, STEAP2
cell growth & proliferation	EMILIN2, MYO5A, MYO5B, GMFB, AGT, AREG/AREGB, BMP6, BMP7, BTC, EREG, FGF18, FGF2, GDF15, IGF2, INHBB, MDK, NOV, NRG2, NRTN, NUDT6, TYMP, NGRN, PRX, GAS2L3, GAS5, CKAP4, KRT10, KRT19, MAP6, MID1IP1, MTSS1, PDLIM3, PDLIM5, RAB11FIP2, RAB11FIP4, FBLN1, FBLN2, FBN1, LAMA1, LAMA3, CABLES1, EMP1, FNDC1, FMNL1, TPD52L1, ECM2, FRAS1, LTBP1, NEBL
cell cycle	CCNB2, G0S2, CCNA1, CCPG1, TP53I13

Pathway and network analyses, generated through the use of the IPA software confirmed the major functionally related gene groups, found to be differentially expressed in the BCAT1 knockdown clones sh-B1 and sh-B2. As seen from Figure 5, pathways linked to cell death and survival, cellular movement, cell cycle regulation, cellular growth and proliferation, cellular function and maintenance and cell-to-cell signaling and interaction were both induced and suppressed. Notably, pathways implicated in gene expression, molecular transport, as well as carbohydrate, lipid, amino acid and nucleic acid metabolism, energy production and protein biosynthesis, were exclusively suppressed upon BCAT1 depletion (Figure 5B). Common networks obtained upon merging the top-scoring networks recognized some important gene nodes and genes that are specifically up- or downregulated upon BCAT1 suppression in SKOV3 cells (Figure 6). Thus, most of gene nodes that were induced upon BCAT1 knockdown are presented in Figure 6A; they are predominantly implicated in transcription regulation (EIF5A2, EIF5, RPSA, ETS1, HDAC6, RUNX1T1, TAF9B, TARDBP, TCEB2, TCF4), metabolism (CUL4B, HSPA8, PLAU, RBM3, MME, MMP13, ADAMTS1, EIF5A2, PFDN5), cell growth and maintenance (TUBB, MPP1, THBS1, FN1, Ecm, PLEC), signal transduction (EDN1, WNT5A, CXCL12), immune response (SRGN, IL7R) and transport (importin alpha). The majority of gene nodes that were suppressed upon BCAT1 knockdown in SKOV3 cells are presented in Figure 6B; these genes were mostly involved in metabolism (PLC, IDH1, IDH2, AKR1B10, GOT1, ESR1, PTGS1, PTGS2, AKR1C3, TGM2, CPE, SULT1A3/A4, sulfotransferase, CDH1, RPL4, RPL19, RPL23A, NEDD4, TP63, EPRS, WARS, MARS, GARS, IARS, SARS), signal transduction (ANXA1, IGF2, AGT, ULK1, UNC119, NROB1, FGF2, PIK3R1, CLDN3, MSR1, Pde, PPARG, NR0B1, PGR, CCNA1, GREB1), transcription regulation (LMO2, GLI1, GATA6, TCF12, KLF5, FOXA1) and cell growth and proliferation (NRG, EREG, MYO5A, MYO5B, MYO5C).

Figure 5

Functional analysis for a dataset of differentially expressed genes (≥ 2-fold) following BCAT1 suppression in SKOV3 cells

A. Functional analysis of upregulated genes; B. Functional analysis of downregulated genes. Top functions that meet a p-value cutoff of 0.05 are displayed.

Figure 6

Network analysis of dynamic gene expression in SKOV3 cells based on the 2-fold common gene expression list obtained following BCAT1 knockdown

A. Upregulated networks; B. Downregulated networks. The four top-scoring networks (upregulated genes) and the five top-scoring networks (downregulated genes) were merged and are displayed graphically as nodes (genes/gene products) and edges (the biological relationships between the nodes). Intensity of the node color indicates the degree of up- (red) or downregulation (green). Nodes are displayed using various shapes that represent the functional class of the gene product (square, cytokine, vertical oval, transmembrane receptor, rectangle, nuclear receptor, diamond, enzyme, rhomboid, transporter, hexagon, translation factor, horizontal oval, transcription factor, circle, other). Edges are displayed with various labels that describe the nature of relationship between the nodes: __ binding only, → acts on. The length of an edge reflects the evidence supporting that node-to-node relationship, in that edges supported by article from literature are shorter. Dotted edges represent indirect interaction.

To validate microarray results, we arbitrarily selected 11 differentially expressed genes and quantified their expression by qPCR in SKOV3 cells following shRNA-BCAT1 knockdown compared to control (vehicle transfected) SKOV3 cells. We found that both methods (microarray analysis and qPCR) detected similar patterns for the up- and down-regulated genes selected for validation (Figure 7). Additionally, the BCAT1 knockdown-directed suppression of some important metabolism modulators, including IDH1, IDH2, AKR1C1 and PHGDH was confirmed by Western blot analysis (see Figure 8).

Figure 7

Quantitative PCR validation of microarray results

The figure shows bar graphs presentation of the differential expression of the selected genes in SKOV3 cells following BCAT1 knockdown compared to control (mock-transfected) SKOV3 cells. The relative copy number was calculated based on the target gene/18S ribosomal RNA ratio. Values more than or equal to 1 represent gene upregulation and less than 1 display gene downregulation. The analysis confirmed higher levels of expression for TGFBI, THBS1, ZP1 and ADAMTS5, and lower levels of expression for S100P, GREB1, RXFP1, GDF15, IL1A, SLP1 and IGFBP2 upon BCAT1 knockdown.

Figure 8

Western blot analysis of IDH1, IDH2, PHGDH and AKR1C1/2/3 protein expression in BCAT1 knockdown SKOV3 clones

These proteins displayed lower expression in BCAT1 clones (sh-B1 and sh-B2), when compared to the control (Ctrl).

Metabolomic validation of altered metabolites’ formation in EOC cells upon BCAT1 knockdown

By using a kit-based high-throughput flow injection mass spectroscopy (MS) approach, we analyzed the mean intracellular metabolic changes in both BCAT1 knockdown SKOV3 clones sh-B1 and sh-B2 when compared to the Ctrl SKOV3 clone. BCAT1 suppression resulted in lowering the concentrations of the major metabolite groups, as mean values for many amino acids, glycerophospholipids and sphingolipids were reduced to 69%, 72% and 80% respectively, when compared to the Ctrl clone (Table 3). BCAT1 inhibition was also associated with lower concentrations of two BCAAs – Leu (69.43%) and Ile (74.31%). A significant depletion of numerous glycerophospholipids and several sphingolipids with both normal (SM) and hydroxylated SM(OH) fatty acid chains was also observed, as well as variations in some polyanines (Table 3).

Table 3

Mean metabolites’ values in BCAT1 knockdown SKOV3 clones sh-B1 and sh-B2, compared to the Ctrl SKOV3 clone

Metabolites	Mean % (± SD) in BCAT1- KD clones sh-B1 and sh-B2 vs. the control clonea
Amino acids1
Ala	72.01 ± 0.64b
Arg	50.10 ± 0.64
Asn	70.20 ± 1.87
Asp	78.70 ± 1.27
Glu	44.01 ± 14.78
Gly	50.00 ± 0.34
Ile	74.31 ± 0.91
Leu	69.43 ± 1.34
Lys	65.74 ± 2.39
Phe	66.36 ± 0.68
Pro	39.51 ± 2.33
Tyr	70.20 ± 0.79
Amino acids pool	69.09%
Glycerophospholipids2
PC aa C30:0	47.62 ± 0.73
PC aa C32:1	54.05 ± 0.22
PC aa C32:2	42.41 ± 0.35
PC aa C32:3	32.83 ± 0.03
PC aa C34:1	54.13 ± 0.31
PC aa C34:2	48.13 ± 7.88
PC aa C34:3	39.05 ± 0.36
PC aa C34:4	33.62 ± 0.05
PC aa C36:2	69.68 ± 4.32
PC aa C36:3	68.89 ± 1.20
PC aa C36:4	57.70 ± 1.77
PC aa C36:5	53.51 ± 0.03
PC aa C36:6	44.47 ± 0.01
PC aa C38:0	71.81 ± 0.03
PC aa C40:2	64.65 ± 0.03
PC aa C42:5	68.69 ± 0.06
PC ae C30:0	52.62 ± 0.11
PC ae C30:1	60.79 ± 0.04
PC ae C32:1	57.79 ± 0.50
PC ae C32:2	56.82 ± 0.05
PC ae C34:1	48.39 ± 0.23
PC ae C34:2	53.81 ± 0.01
PC ae C34:3	51.57 ± 0.07
PC ae C36:1	62.10 ± 0.09
PC ae C36:2	56.09 ± 0.22
PC ae C36:3	48.63 ± 0.31
PC ae C36:4	47.37 ± 0.28
PC ae C36:5	53.38 ± 0.24
PC ae C38:2	64.17 ± 0.10
PC ae C38:3	62.70 ± 0.10
PC ae C38:4	61.80 ± 0.18
PC ae C38:5	51.21 ± 0.26
PC ae C38:6	60.48 ± 0.21
PC ae C40:2	64.49 ± 0.01
PC ae C40:3	60.70 ± 0.03
PC ae C40:4	67.78 ± 0.01
PC ae C40:5	66.57 ± 0.06
PC ae C40:6	62.58 ± 0.07
PC ae C42:4	63.32 ± 0.03
PC ae C44:6	45.19 ± 0.01
PC pool	72.00%
Sphingolipids3
SM (OH) C14:1	75.99 ± 0.08
SM (OH) C16:1	67.81 ± 0.01
SM (OH) C22:2	72.10 ± 0.09
SM C16:0	72.90 ± 1.21
SM C16:1	74.62 ± 0.59
SM C18:0	77.05 ± 0.04
SM C18:1	71.02 ± 0.23
SM C24:0	70.19 ± 0.10
SM C24:1	69.24 ± 0.41
SM C26:1	65.51 ± 0.03
SM pool	79.65%
Acylcarnitines4
C14:1	70.02 ± 0.00
Biogenic amines5
Ac-Orn	45.81 ± 0.01
Putrescine	527.34 ± 0.01
Sarcosine	69.77 ± 1.13

Percentage are derived from medium values for both BCAT1 knockdown clones (shB1 & sh-B2)

Standard Deviation (SD) values are displayed

Ala: Alanine; Arg: Arginine; Asn: Asparagine; Asp: Aspartic acid; Glu:Glutamic acid; Gly :Glycine; Ile: Isoleucine; Leu: leucine; Lys: Lysine; Phe: Phenylalanine; Pro: Proline; Ser: Serine; Thr: Threonine; Tyr: Tyrosine.

PC aa: phosphatidylcholine diacyl; PC ae: phosphatidylcholines acyl-alkyl

SM (OH): Hydroxysphingomyeline; SM:sphingomyeline

Tetradecenoylcarnitine

Ac-Orn: Acetyl-ornithine

BCAT1 suppression inhibits tumor expansion and metastasis in nude mice

In order to investigate the effect of BCAT1 knockdown on EOC dissemination in vivo, intact SKOV3 cells, as well as mock (Ctrl) and shRNA-BCAT1 (clone sh-B1) transduced SKOV3 cells were injected intraperitoneally (IP) in nude mice (n = 6–7 per experimental group). The BCAT1 suppression in tumors of mice injected with the BCAT1-knockdown cells was confirmed by Western blot and IHC staining (Figures 9A and 9B). Moreover, global gene expression comparison in tumor tissues extracted from two nude mice i.p. injected with sh-B1 cells, versus tumors from two mice injected with Ctrl cells, displayed evident similarity between the major functionally related gene groups that were differentially expressed upon BCAT1 silencing both in vitro and in vivo (see Figure 5 and Supplemental Figure 5).

Figure 9

A. Western-blot analysis of BCAT1 protein expression in tumor tissues extracted from nude mice, injected with the parental SKOV3 cells, the mock-transfected (Ctrl) cells, and the BCAT1 knockdown (sh-B1 cells

B. Representative IHC images of BCAT1 expression in tumor tissues extracted from nude mice, injected with the parental SKOV3 cells, Ctrl cells, and sh-B1 cells. C. Survival curves for mice injected with parental, Ctrl, and sh-B1 SKOV3 cells. The median survival of mice injected with the Ctrl cells (87 days, n = 6) is similar to the parental SKOV3 cells (85 days, n = 7). Survival of mice injected with the BCAT1 knockdown cells is significantly longer than the vector control (165 days; P = 0.0031, Log rank test), with 5 mice still alive more than 200 days after injection with the BCAT1 knockdown cells.

The total tumor burden and the ascites volume were not different among the three experimental groups (data not shown). However, BCAT1 knockdown had a major impact on survival. While the median survival of mice injected with the vector-control cells (87 days, n = 6) was similar to the parental SKOV3 cells (85 days, n = 7), the survival of mice injected with BCAT1 knockdown cells was significantly longer than the vector-control (165 days; p = 0.0031, Logrank test), with 5 mice still alive more than 200 days after injection with the BCAT1 knockdown cells (Figure 9C).

The transcription factor c-Myc regulates BCAT1 expression in EOC cells

c-Myc represents a major oncogene, frequently overexpressed in EOC [18]. BCAT1 has been previously identified as one of the c-Myc target genes in different cancers [15, 19, 20]; however, the c-Myc-BCAT1 link has been not studied in EOC. To investigate the regulation of BCAT1 by c-Myc in EOC cells, shRNA-related c-Myc knockdown was performed in SKOV3 cells. Clone selection for further analyses was based on Western blot validation of the c-Myc protein expression in selected clones, compared to empty vector-transfected clones. Among the analyzed clones, the shRNA-c-Myc knockdown clones sh-C1 and sh-C2 displayed a significant decrease of c-Myc protein expression levels compared to the mock-transfected control (Ctrl) (Figure 10); as expected, the c-Myc knockdown resulted in evident suppression of the BCAT1 protein expression (Figure 10).

Figure 10

Western blot analysis of BCAT1 protein expression in shRNA-mediated c-Myc knockdown clones sh-C1 and sh-C2, compared to the control (Ctrl-C) clone

DISCUSSION

We have recently identified a 425 nt CpG-rich region of the BCAT1 gene that is significantly hypomethylated in HG serous EOC tumors, compared to normal tissue [8]. This region is positioned within the first intron of BCAT1 gene isoforms I, II and III; however, it covers a potential control (promoter 2) region for BCAT1 isoforms IV and V. Indeed, the 425 nt hypomethylated region spans the transcription initiation site (−105 nt to +320 nt) of BCAT1 gene isoform IV, and is located upstream of the transcription initiation site (−370 nt to −795 nt) of BCAT1 gene isoform V (comprising CpG island 70 - see Supplemental Figure 6). Since all five BCAT1 isoforms share highly homologous amino acids sequences (ranging from 89% to 99%), our findings confirm the previously proposed role of DNA methylation in controlling alternative promoter usage, both during normal cell differentiation and neoplastic transformation [16, 21]. The implication of aberrant DNA methylation in the control of BCAT1 expression in cancer has been also demonstrated in gliomas and colorectal cancer [16, 21]. These findings suggest that epigenetic mechanisms could account for altered BCAT1 expression in different cancer types, including EOC.

Moreover, our functional analyses are strongly indicative for evident oncogenic capacity of BCAT1 in EOC, including its potential implication in EOC cell proliferation, cell cycle control and cell migration/invasion (see Figures 2, 3, 4), in concordance with similar findings about the role of BCAT1 in other cancers [12-16]. The consecutive gene expression experiments and related network and pathway analyses were quite confirmatory of the data obtained by the BCAT1 functional assays (see Table 2 and Figure 5).

IPA network analysis was indicative for some important gene nodes linked to BCAT1 suppression in EOC cells, as most of these substantiate and/or complement the functional data obtained. As shown on Figure 6A, BCAT1 knockdown resulted in upregulation of a number of gene nodes with previously suggested TSG function in EOC, including RBM3, THBS1, RUNX1T1 and WNT5A [22-25], as well as some TSGs, characterized in other cancer types (ADAMTS1, PFDN5/MM-1, MME, MPP1, TCEB2) [26-30]. Other induced gene nodes comprised genes, known in promoting cellular death/apoptosis, such as ETS1, SRGN, TCF4 and CUL4B [31-34].

BCAT1 suppression in SKOV3 cells led to the induction of several gene nodes (EIF5A2, EIF5 and EIF4H), as part of the eukaryotic initiation factors (eIFs) network, shown to display all elements of a complete oncogenic, as well as tumor suppressive cascade [35]. Similarly, some of the induced genes nodes upon BCAT1 knockdown (PLEC, HDAC6, CXCL12, DCUN1D1/SCCRO) have exhibited dual functions in tumorigenesis [36-43]. Some known oncogenes (MMP13, FN, EDN1 and PLAU) [44-47] also displayed increased expression upon BCAT1 knockdown, which could be indicative for some compensatory carcinogenic mechanisms in SKOV3 cells following BCAT1 suppression.

In parallel, upon BCAT1 knockdown we have observed substantial downregulation of major gene nodes known to be implicated in EOC tumorigenesis (PPARG, ANXA3, TGM2, Ap1/c-jun, NRG & NRG2, TP63, PGR, MSR1, PIK3R1, KLF5, TCF) [48-58], including EOC dissemination/metastasis (CLDN3, EPCAM, PTGS1/COX1 & PTGS2/COX2, ESR1, GLI1, GREB1, ERBB, IGF2, FGF2, CDH1) [59-69] (Figure 6B). The BCAT1 silencing in SKOV3 cells also resulted in suppression of numerous gene nodes with proven oncogenic potential in different cancers (including EREG, RPL23A, LMO2, TCF12, FMNL1, MYO5A, CCR7, NEDD4, ANXA1, TSLP, RPL4, AGT, FOXA1, CPE, RPL19) [70-84] (Figure 6B). Thus, our findings are strongly indicative for the oncogenic functionality of BCAT1 in EOC etiology.

Moreover, BCAT1 suppression was notably associated with downregulation of numerous metabolism-related gene nodes, predominantly implicated in regulation of lipid metabolism (PLC, NROB1, AKR1B10, SULT1A3/A4, GOT1, ESR1, PGR, MSR1, ANXA1, PTGS1, PTGS2, AGT, KLF5, PIK3R1, MYO5A, AKR1C3, FOXA1, PPARG) (Figure 6B). Among these, sulfotransferases (SULT1A1, SULT1A2, SULT1A3) and aldo-keto reductases (AKR1C1, AKR1C2, AKR1C3) were shown to contribute to disease progression and/or represent therapeutic targets in hormone-dependent forms of cancer [85, 86], including EOC [87].

A number of aminoacyl-tRNA synthetases (ARSs), including EPRS, MARS, SARS, WARS, IARS, GARS, AARS, were suppressed upon BCAT1 silencing in EOC cells. These enzymes play a critical role in protein biosynthesis by charging tRNAs with their cognate amino acids; moreover, recent findings are indicative the potential pathophysiological implications of ARSs in tumorigenesis [88].

The two isoforms of the isocitrate dehydrogenase enzyme (IDH1 and IDH2) were also among the major gene nodes that were suppressed upon BCAT1 silencing (see also Figure 6B). The IDH1 and IDH2 enzymes catalyze the oxidative decarboxylation of isocitrate to α-ketoglutarate, generating NADPH from NADP+. IDH enzymes are associated with the tricarboxylic acid (TCA) cycle for energy production and are thereby involved in multiple metabolic processes [89]. Mutations of the IDH1/2 genes represent significant driver mutations in gliomas and acute myeloid leukemia development, but are quite rare/almost undetectable in solid tumors [90]. Recent study has demonstrated a link between IDH1 function and BCAT1 expression in gliomas, as BCAT1 was shown to be involved in tumor growth and disease progression only in gliomas with wild-type (non-mutated) IDH1 [16]. Moreover, shRNA-directed knockdown of IDH1 led to strong downregulation of BCAT1 expression in glioblastoma cell lines [16]. Our data are indicative for the inverse relationship regarding the impact of the BCAT1 expression on the expression of the IDH genes. Additionally, the aldo-keto reductase AKR1C1 and the phosphoglycerate dehydrogenase (PHGDH) were among the most strongly suppressed genes (−15.20 and −14.58, respectively) in both (sh-B1 and sh-B2) BCAT1 knockdown clones (see Suppl. Table 1B). Both these genes are actively implicated in cellular metabolism and have been previously characterized as therapeutic cancer targets. Indeed, AKR1C1 was shown to induce resistance to cisplatin in human EOC cells [91] and was thus recognized as potential EOC therapeutic target [87]. PHGDH catalyzes the first step in the serine biosynthesis pathway by diverting glucose-derived carbon into serine and glycine metabolism. Recently, the essential role of the serine biosynthesis pathway in metabolic reprogramming in cancer has been recognized, and PHGDH was found to be overexpressed and to contribute to tumor cell growth in a significant subset of human cancers [92, 93]. The IDH1, IDH2, ACR1C1 and PHGDH downregulation following BCAT1 knockdown was also confirmed on the protein level (see Figure 8).

Therefore, BCAT1 silencing could exert multiple inhibitory effects on the TCA cycle-associated energy production, and could impact EOC metabolism by suppressing lipid production and protein synthesis, including the serine biosynthesis pathway. This was further supported by our initial and rather exploratory metabolomic investigation. Indeed, the BCAT1 knockdown phenotype was associated with significant depletion of most amino acids and numerous phospho- and sphingolipids, as well as with perturbations in metabolism of some polyamines.

EOC spreads by intraperitoneal (IP) sloughing, lymphatic invasion, and hematogenous dissemination [94]. IP dissemination is the most common; after malignant cells have evaded from the ovarian capsule, they are shed from the tumor surface into the peritoneal cavity where they follow normal routes of peritoneal fluid [95]. Hence, IP injection of cancer cells in animal models can accurately model advanced disease, as EOC metastases frequently appear disseminated throughout the peritoneum [96]. We used a similar in vivo approach to investigate the role of BCAT1 in EOC progression. We found that SKOV3 cells dissemination was strongly abrogated following BCAT1 knockdown, which led to significant increase of survival rate in animals treated with BCAT1 knockdown cells.

In conclusion, we have shown that BCAT1 is significantly overexpressed in LMP tumors and HG serous EOC tumors compared to normal ovarian tissues, as epigenetic mechanisms (DNA hypomethylation) could be involved in BCAT1 overexpression in this tumor type. We also confirmed that c-Myc regulates BCAT1 expression in EOC cells. Functional analyses pointed towards BCAT1 implication in the control of EOC cell proliferation (including cell cycle control), migration and invasion. Gene expression profiling and associated network and pathway analyses confirmed these findings, as numerous genes and pathways known previously to be implicated in ovarian tumorigenesis, including EOC tumor invasion and metastasis, were found to be suppressed upon BCAT1 knockdown, while some TSGs were induced. Importantly, our microarray data and consecutive metabolomics analyses are indicative for a major BCAT1 implication in controlling EOC metabolism. Finally, BCAT1 knockdown led to reduced dissemination and prolonged survival time in xenograft models of advanced peritoneal EOC. Our data provide both in vitro and in vivo proof-of-concept that BCAT1 represents a major EOC oncogene and suggest that BCAT1 should be further explored as a potential EOC therapeutic target.

MATERIALS AND METHODS

Patients and tissue specimens

Snap frozen and formalin-fixed paraffin-embedded (FFPE) tissues of 117 EOC tumors were obtained at the Hotel-Dieu de Quebec Hospital, Quebec, Canada. These included 13 borderline, or LMP tumors and 104 HG adenocarcinomas. None of the patients received CT before surgery (see Table 1 for detailed clinicopathological characteristics). All tumors were histologically classified according to the criteria defined by the World Health Organization [97]. The CT treatment was completed for all patients and the response to treatment was known. Disease progression was evaluated following the guidelines of the Gynecology Cancer Intergroup [97]. PFS was defined as the time from surgery to the first observation of disease progression, recurrence or death. Thirteen normal ovarian samples were derived from women subjected to hysterectomy with oophorectomy due to non-ovarian pathologies. The study was approved by the Clinical Research Ethics Committee of the Hotel-Dieu de Quebec Hospital and all patients signed an informed consent for voluntary participation.

Cell cultures

The EOC cell lines OVCAR3 and SKOV3 were purchased from American Tissue Type Collection (Manassas, VA); OV-90, OV2008, TOV-112 and TOV-21 cell lines were a kind gift from Dr. Anne-Marie Mes-Masson (Université de Montréal), while A2780s and A2780cp cell lines were a kind gift from Dr. Benjamin Tsang (University of Ottawa). The cell lines were passed in different culture media supplemented with 10% fetal bovine serum, as previously described [98].

Tissue microarrays (TMAs) construction and immunohistochemistry (IHC)

TMAs construction and IHC analyses were performed as previously described [98-100]. Briefly, one representative 4 μm tissue section was cut from the tissue array blocks. Sections were deparaffinized and rehydrated in graded alcohols, then incubated with blocking serum for 20 min. Sections were incubated overnight at room temperature with the BCAT1 rabbit polyclonal antibody (Proteintech) at dilution 1:50. Sections were then incubated with a biotinylated secondary antibody (Dako, Carpinteria, CA) and then exposed to a streptavidin complex (Dako, Carpinteria, CA). Complete reaction was revealed by 3–3 diaminobenzidine and slides were counterstained with hematoxylin.

BCAT1 protein expression was assessed by semi-quantitative scoring of the intensity of staining and recorded as absent (0), weak (1+), moderate (2+) or strong (3+). The IHC staining was analyzed independently by 2 pathologists blinded to clinical data and progression. The relationship between BCAT1 expression in serous ovarian carcinomas and normal ovarian tissues was evaluated by the Wilcoxon two-sample test. A significant association was considered when p-value was below 0.05. A Kaplan-Meier curve and the log-rank test were performed based on PFS values to test the effect of the intensity of BCAT1 (3, 2 versus 0, 1) on disease progression.

Short Hairpin RNA (shRNA) – mediated BCAT1 and c-Myc knockdowns in SKOV3 and A2780s cells

A shRNA, targeting the BCAT1 sequence 5′-GAAGAACTGGCAACTCCTC-3′, was designed using the siRNA Ambion Target Finder software (http://www.ambion.com/techlib/misc/siRNA_finder.html), and subcloned into the pSilencer 4.1 puro vector (Ambion). SKOV3 cells were stably transfected with the shRNA-BCAT1 plasmid using the ExGen 500 transfection reagent (Fermentas Canada Inc., Burlington ON), according to the manufacturer's instructions. Cells were consecutively grown for 2 weeks in selection medium containing 0.5 μg/ml puromycin (Wisent, Canada) to isolate stable clones. Cells were also mock-transfected with the pSilencer 4.1 puro vector, and the stably-transfected clones were isolated as controls. Stable clones with inhibited BCAT1 expression were evaluated and validated by semi-quantitative RT-PCR and Western blot, as previously described [101].

A c-Myc shRNA cloned into the pLKO.1-puro vector was retrieved from the Sigma Mission TRC human 1.5 shRNA library (clone number TRCN0000039642). Viral supernatants were generated by transfecting 293T cells with the shRNA construct and the packaging vectors psPAX2 and pMD2.G (Addgene, Cambridge, MA). The high-titer lentiviral supernatants in the presence of 5 μg/ml polybrene were used to infect SKOV3 cells. Two days later, infected cells were treated with puromycin (0,5 μg/ml) for the selection of stably-transduced clones. The pLKO.1-puro vector encoding a scramble sequence not matching any mammalian sequence was used for the generation of mock-transduced (Ctrl-C) clones. Stable clones with inhibited c-Myc expression were evaluated and validated by Western blot.

Functional assays

Different functional assays, including real time cell proliferation monitoring (using the xCELLigence Real-Time Cell Analyzer), colony formation, cell cycle analysis, cell migration and invasion assays and cell cytotoxity (MTT) assays were performed as previously shown [98, 101, 102].

Gene expression profiling and data analysis

Gene expression profiling and microarray analysis was carried out using the Agilent Whole Human Genome microarrays, containing 44,000 genes, as described before [98, 100, 102]. Network analysis of the microarray data was completed using the Ingenuity Pathway Analysis (IPA) software (see http://www.Ingenuity.com). The microarray data have been deposited to the GEO database (http://www.ncbi.nlm.nih.gov/geo/) with accession number GSE64424. Selected microarray data were confirmed by quantitative PCR (qPCR) as described previously [102]. The primers used for sqRT-PCR and qPCR validation are listed in Supplemental Table 2.

Western blotting

Western blot analysis was performed as previously described [98, 100, 102]. Antibodies used for monitoring protein expression were the anti-BCAT1 rabbit polyclonal Ab (1:500) (Proteintech), the anti-IDH1 goat polyclonal Ab (1:500) (Santa Cruz Biotechnology), the anti-IDH2 mouse monoclonal Ab (1:500) (Santa Cruz Biotechnology), the anti-PHGDH mouse polyclonal Ab (1:500) (Santa Cruz Biotechnology), the anti-AKR1C1, 2, 3 goat polyclonal Ab (1:500) (Santa Cruz Biotechnology), and the anti-β-actin Ab (1:5000) (Santa Cruz Biotechnology), the latter used as internal standard. The anti-c-Myc Ab was a gift from Dr. Gobeil's lab.

Metabolomics analyses

5 × 106 SKOV3 cells were plated in 75 cm flasks and grown for 48 h before counting and collecting the cells by trypsinization. Upon 2 × washing with cold PBS, cells were lysed by adding 10 mM phosphate buffer, pH 7.4 (15% v/v) and cold 100% EtOH (85% v/v). Lysed cell suspensions were subjected to three rounds of sonication-freeze-thaw to further ensure cell disruption and protein denaturation and centrifuged at 20000 g for 10 minutes at 4°C to remove any pellet. Supernatants were transferred into cryogenic vials, and stored at −160°C.

Metabolomics analysis was performed using a commercially available assay using the LC-MS/MS based AbsoluteIDQ p180 Kit (Biocrates Life Sciences AG, Austria), which allows the quantification of up to 188 endogenous metabolites from 5 different compound classes including acylcarnitines, amino acids, hexoses, phospho- and sphingolipids and biogenic amines [103]. MS analyses were done on a API 4000 LC/MS/MS System (AB Sciex, Concord, ON, Canada) equipped with a UFLC Prominence (Shimadzu Scientific Instrument Inc., Columbia, MD, USA) and controlled by the software Analyst version 1.6.2. Isotope-labeled internal standards and other internal standards are integrated into a kit plate filter for metabolite quantification. The Biocrates MetIDQ software was used to control the assay workflow, from sample registration to automated calculation of metabolite concentrations and the export of data into other data analysis programs. The metabolomics analyses were performed at the Pharmacogenomics laboratory, CHU Research Center, Québec, Canada (http://www.pharmacogenomics.pha.ulaval.ca/).

Peritoneal tumor formation in mice

Intact SKOV3 cells, as well as the SKOV3 sh-B1 clone and the SKOV3 Ctrl clone (1 × 107 cells in 500 μL of PBS), were IP injected into 8 × 8 week old CB17 SCID female mice (CB17/Icr-Prkdcscid/IcrIcoCrl strain code 236, Charles River) using a 25G5/8 needle. Mice were monitored daily by staff blinded to the cell type injected and euthanized when they reached a loss of wellness endpoint that was most often respiratory distress associated with ascites accumulation. The animals had free access to food and water and experiments were done in accordance with the Canadian Council on Animal Care's Guidelines for the Care and Use of Animals. Protocols were approved by the University of Ottawa Animal Care Committee.