PDSS2 Deficiency Induces Hepatocarcinogenesis by Decreasing Mitochondrial Respiration and Reprogramming Glucose Metabolism

Published OnlineFirst July 2, 2018; DOI: 10.1158/0008-5472.CAN-17-2172

Cancer Metabolism and Chemical Biology Research

PDSS2 Deﬁciency Induces Hepatocarcinogenesis by Decreasing Mitochondrial Respiration and Reprogramming Glucose Metabolism Yan Li1, Shuhai Lin2, Lei Li1, Zhi Tang2, Yumin Hu1, Xiaojiao Ban1, Tingting Zeng1, Ying Zhou1, Yinghui Zhu1, Song Gao1, Wen Deng3, Xiaoshi Zhang1, Dan Xie1, Yunfei Yuan1, Peng Huang1, Jinjun Li4, Zongwei Cai2, and Xin-Yuan Guan1,5

Abstract

Glucose metabolic reprogramming from oxidative phos- Reintroduction of full-length PDSS2 into HCC cells increased phorylation to glycolysis is one of the hallmarks of cancer CoQ10 and mitochondrial electron transport complex I activ- development. Coenzyme Q10 (CoQ10) is essential for elec- ity and subsequently induced a metabolic shift from aerobic tron transport in the mitochondrial respiratory chain and for glycolysis to mitochondrial respiration in cells. Reintroduc- antioxidant defense. Here, we investigated the role of a key tion of PDSS2 also inhibited foci formation, colony formation factor in CoQ10 synthesis, prenyldiphosphate synthase sub- in soft agar, and tumor formation in nude mice. Knockdown of unit 2 (PDSS2), in hepatocellular carcinoma (HCC) tumori- PDSS2 induced chromosomal instability in the MIHA immor- genesis. PDSS2 was frequently downregulated in HCC tissues talized human liver cell line. Furthermore, knockdown of and was significantly associated with poorer HCC prognosis PDSS2 in MIHA induced malignant transformation. Overall, (P ¼ 0.027). PDSS2 downregulation was a prognostic factor our findings indicate that PDSS2 deficiency might be a novel independent of T status and stage (P ¼ 0.028). Downregula- driving factor in HCC development. tion of CoQ10 was significantly correlated with downregulation of PDSS2 in HCC tumor tissues (R ¼ 0.414; P < 0.001). Of Significance: Downregulation of PDSS2 is a driving factor the six different splicing isoforms of PDSS2, the five variants in hepatocellular carcinoma tumorigenesis. Cancer Res; 78(16); other than full-length PDSS2 showed loss of function in HCC. 4471–81. 2018 AACR.

Introduction correlated with increased metastasis, tumor recurrence and poor outcome (9–11). Cancer cells reprogram their metabolism, which is character- Initially, metabolic reprogramming was thought to be a con- ized by a decrease in mitochondrial respiration and oxidative sequence of rapid cell proliferation, but recent data have dem- phosphorylation, together with strong enhancement of glycolysis, onstrated that metabolic reprogramming can actually drive tumor even in the presence of abundant oxygen (known as the Warburg development (12, 13). Inhibiting glucose ﬂux by suppressing key effect; refs. 1–3). Rapid consumption of glucose in tumors has enzymes such as PKM2 (14), LDH (15), and 6-phosphofructo-2- been associated with a poor prognosis in patients with oral kinase (16) could effectively reduce tumorigenicity, implying that squamous cell carcinoma (4), gastric cancer (5), esophageal metabolic reprogramming from oxidative phosphorylation to cancer (6), pancreatic cancer (7), and other tumors (8). Lactate, glycolysis plays an important role in cancer development and once considered a waste product of glycolysis, has now been progression (13). Recent studies showed that the oncogenes c-myc, Akt and NF-kB could upregulate many genes involved in glycolysis and glutaminolysis (1, 17). Similarly, loss of p53 function could lead to the Warburg effect through complex 1 State Key Laboratory of Oncology in South China, Collaborative Innovation regulation of glycolysis, ROS levels and apoptosis (18–21). Center for Cancer Medicine, Sun Yat-sen University Cancer Center, Guangzhou, ﬁ China. 2State Key Laboratory of Environmental and Biological Analysis, Hong In a previous study, we identi ed and characterized a tumor- PDSS2 Kong Baptist University, Hong Kong, China. 3Department of Anatomy, The suppressor (prenyldiphosphate synthase subunit 2), University of Hong Kong, Hong Kong, China. 4State Key Laboratory of Onco- which was interrupted by the 6q21 breakpoint in the recipro- genes and Related Genes, Shanghai Cancer Institute, Shanghai Jiaotong Uni- cal-like chromosomal translocation t(6;17)(6q21:17p13) in the 5 versity, Shanghai, China. Department of Clinical Oncology, The University of UACC930 melanoma cell line (22, 23). PDSS2 is important in Hong Kong, Hong Kong, China. determining the length of the side chain of ubiquinone in mam- Note: Supplementary data for this article are available at Cancer Research mals (24). PDSS2 catalyzes the addition of isopentenyl diphos- Online (http://cancerres.aacrjournals.org/). phate molecules to farnesyl diphosphate in multiple steps to Corresponding Authors: Xin-Yuan Guan, Sun Yat-sen University Cancer Center, produce the isoprenoid side chain of coenzyme Q10 (CoQ10). 651 Dongfeng East Road, Guangzhou 510060, China. Phone: 8620-87343166; A decrease in the plasma CoQ10 level has been detected in E-mail: [email protected]; and Yan Li, Sun Yat-sen University Cancer Center. patients with breast cancer (25), melanoma (26), and cervical Phone: 8620-87343166; E-mail: [email protected] cancer (27). 6q21 deletion is also one of the most frequently doi: 10.1158/0008-5472.CAN-17-2172 deleted regions in hepatocellular carcinoma (HCC; ref. 28). 2018 American Association for Cancer Research. Although the introduction of PDSS2 into melanoma and gastric

www.aacrjournals.org 4471

Li et al.

cancer cells could inhibit tumor growth (23, 29), the mechanism were added, and analysis was performed after 24 hours or at underlying the role of PDSS2 in tumor genesis is not well the indicated time. The glucose and lactate levels in the culture understood. media were measured using a glucose assay kit (glucose oxidase– In the present study, we found that decreased expression of peroxidase method) and a lactate assay kit (LDH catalysis meth- PDSS2 was frequently detected in HCC tumor tissues and was od; Nanjing Jiancheng Technology), respectively. The cells were significantly associated with a poor outcome of HCC (P ¼ 0.027). collected, and the protein content was quantified. The glucose and PDSS2 downregulation was an independent prognostic factor, lactate levels were normalized by the protein content. Triplicate independent of T status and stage (P ¼ 0.028). Further study independent assays were performed. demonstrated that PDSS2 was able to increase mitochondrial electron transport complexes and then induce a metabolic shift Oxygen consumption rate analysis from aerobic glycolysis to mitochondrial respiration in hepato- Equal numbers of cells (5 106) were suspended in 1ml cellular carcinoma tumor cells. Silencing PDSS2 in the MIHA fresh medium pre-equilibrated with fresh air (21% oxygen immortalized human liver cell line promoted glycolytic flux and and 5% CO2)at37C and transferred to a chamber equipped tumor cell growth. We also found PDSS2-Del2, a previously with a thermostat controller and a microstirring device uncharacterized PDSS2 isoform, which was incapable of synthe- (Oxytherm, Hansatech Instrument). The oxygen consumption sizing CoQ10. The molecular mechanism of PDSS2 in HCC rate is expressed as nanomoles of oxygen per milliliter per development was also evaluated. 5 106 cells.

SDH, IDH, and Complex I activity assays Materials and Methods Cells were plated in 6-well plates, and thiazolyl blue tetrazo- Tissue microarray and staining for immunohistochemistry lium bromide (MTT, Sigma-Aldrich) was added to the cells. The Paraffin blocks from 295 patients with HCC and paired non- blue crystals were solubilized with DMSO and measured at tumorous tissues were collected from two hospitals (147 cases 570 nm. The results were normalized to the cell protein. An from the First Affiliated Hospital, Sun Yat-sen University, and 148 isocitrate dehydrogenase activity assay was performed using an cases from Sun Yat-sen University Cancer Center). Written IDH activity assay kit (Biovision) according to the manufacturer's informed consent was obtained from the patients and the study protocol. Complex I activity was determined by monitoring the þ was conducted in accordance with the Declaration of Helsinki. oxidation of NADH to NAD and the simultaneous reduction of a The study was approved by the Committees for Ethical Review of dye, which led to increased absorbance (OD450 nm; Complex I Research Involving Human Subjects in the Cancer Center and The activity kit, Abcam). Triplicate experiments were performed First Affiliated Hospital, SunYat-sen University. A tissue micro- independently. array (TMA) was constructed according to the method previously described (30). The age of patients ranged from 9 to 83 years at the Measurement of cellular ATP time of surgery (median age: 49.5 years) and the male/female Cellular ATP was measured using an ATP assay kit (Beyotime) ratio was 7.5:1. IHC was performed using the standard strepta- according to the manufacturer's protocol. Cells were seeded in a vidin–biotin–peroxidase complex method. After deparaffiniza- 6-well plate and lysed. The cellular ATP was normalized to the tion and endogenous peroxidase activity blockage, the section was cell protein content. Triplicate experiments were performed incubated with antibodies against PDSS2 (1:300, Wolwo) or independently. CoQ10B (1:800, Abcam). A staining score was obtained as the intensity of positive staining (negative ¼ 0, weak ¼ 1, moderate ¼ Detection of micronuclei 2, strong ¼ 3) multiplied by the proportion of immunopositive The cells were fixed and stained with DAPI. For the analysis, cells of interest (<25% ¼ 1, 26–50% ¼ 2, 51–75% ¼ 3, >75% ¼ 4). 3 103 cells were examined for each sample, and the cells with fi In the study, downregulation of PDSS2 was de ned as Nscore- micronuclei were counted. Data presented are the mean of three > fi Tscore 2, and downregulation of CoQ10 was de ned as Nscore- independent experiments. Tscore>3. Spectral karyotyping Tissue collection and CoQ10 analysis Spectral karyotyping (SKY) analysis was performed on 1 month Fresh HCC tumor tissues and the corresponding nontumorous consecutive cultured cells. Metaphase spreads were prepared on tissues were collected at Sun Yat-sen University Cancer Center. The glass slides, denatured and hybridized with denatured SKY probes clinical specimens used in this study were approved by the (SkyPaint Probes, Applied Spectral Imaging) and then analyzed Committees for Ethical Review of Research Involving Human with SkyView according to the manufacturer's protocol (32). Fifty Subjects at the Sun Yat-sen University Cancer Center. The tissue metaphase spreads from each cell group were analyzed. samples were frozen in liquid nitrogen, and then CoQ10 was analyzed. Ultra high-performance liquid chromatography Animal studies tandem mass spectrometric analysis was carried out on a Waters All procedures involving mice were approved by the Institu- ACQUITY UPLC/TQD system, which combines ultra high- tional Animal Care and Use Committee of Sun Yat-sen University performance liquid chromatography with triple quadrupole mass Cancer Center. Cells (enforced expression, knockdown and spectrometric (MS/MS) detection (31). vector controls) derived from HCC and MIHA cell lines were inoculated subcutaneously into the flanks of 4- to 6-week-old Measurement of glucose and lactate BALB/c nude mice. Tumor formation in nude mice was monitored The same number of cells from the experimental group and during the experimental period. Cells derived from the MIHA line from the control group were plated in a 6-well plate. Fresh media (PDSS2-KD1 and MIHA-C) were mixed with Matrigel (10 vs.

4472 Cancer Res; 78(16) August 15, 2018 Cancer Research

PDSS2 Deﬁciency Induces Hepatocellular Carcinoma

100 mL cell suspension) before inoculation since the MIHA cell nontumorous information, CoQ10 downregulation was detected line is not tumorigenic. Tumor growth was examined by measur- in 86/144 (59.72%) of tumor tissues compared with matched ing the tumor size with calipers. The tumor volume was calculated nontumor tissues (Fig. 1E). The correlation study showed that (V ¼ 0.5 L W2). After sacrifice, the tumors that formed in the decreased expression of CoQ10 in HCCs was correlated with T tested mice were excised, fixed, and embedded in a paraffin block status marginally (P ¼ 0.053, Supplementary Table S2). In addi- for hematoxylin and eosin staining or IHC study. tion, CoQ10 was measured by Ultra high-performance liquid chromatography tandem mass spectrometric analysis in 13 pairs Statistical analysis of freshly prepared HCC tumor tissues and corresponding non- Results are presented as the average SEM. The correlation tumor tissues. The results showed that CoQ10 was decreased in between PDSS2 or CoQ10 downregulation and the clinicopath- tumor tissues compared with that in the corresponding nontumor ologic features of patients with HCC was assessed by the Pearson's tissues (P < 0.01, Fig. 1F). A CoQ10 decrease was also observed in x2 test. Survival curves were generated according to the Kaplan– HCC cells compared with that in MIHA cells (Supplementary Meier method and statistical analysis was performed by the log- Fig. S1A). Correlation analysis was also performed in 114 cases rank test. Multivariate survival analysis was performed on all that had both PDSS2 and CoQ10 information. The results parameters that were found to be significant in univariate analysis demonstrated that downregulation of CoQ10 was positively using a Cox regression model. All statistical analysis was carried correlated with downregulation of PDSS2 in tumor tissues out using statistical software (SPSS 16.0 for Windows; SPSS, Inc.). (R ¼ 0.414, P < 0.001; Pearson correlation; Supplementary Differences were considered significant if the P value was less Table S3). IDH1 is an important enzyme in the krebs cycle. To than 0.05. confirm whether PDSS2 downregulation affects the enzymes in Supplementary Information about the Materials and Methods the krebs cycle, IDH1 expression was evaluated in HCC tumor is provided in the Supplementary Information. tissues, and downregulation of IDH1 was positively correlated with downregulation of PDSS2 (R ¼ 0.260, P ¼ 0.017; Pearson correlation; Supplementary Tables S4–S5; Supplementary Results Fig. S1B). PDSS2 was frequently downregulated in HCC and significantly associated with poorer prognosis PDSS2 introduction increased mitochondrial Complex I The expression of PDSS2 was analyzed by immunohistochem- activity istry on two sets of HCC tissue microarrays containing a total of The results demonstrated higher CoQ10 in PDSS2-transfected 295 pairs of HCC cases (tumor vs. corresponding nontumor cells than in empty vector-transfected cells (Fig. 2A; Supplemen- tissues). Informative data were obtained from 226 tumor tissues tary Fig. S1C). Because CoQ10 plays a key role in the electron and 197 nontumor tissues. In 164 cases with both tumor and transfer chain, and because CoQ10 addition has been reported to nontumor information, PDSS2 downregulation was observed in be associated with an increase in the Complex I-linked respiratory 57.32% (94/164) of the tumor tissues compared with that in the rate (35), we next examined the expression of the mitochondrial matched nontumor tissues (Fig. 1A). Although the correlation respiratory chain components by Western blot analysis. The study demonstrated that PDSS2 downregulation was not signif- results showed that the level of Complex I increased as expected, icantly correlated with gender, age, T status, and stage (Supple- whereas the Complex II level increased slightly in PDSS2-trans- mentary Table S1), Kaplan–Meier survival analysis revealed fected cells, compared with those in empty vector-transfected cells that PDSS2 downregulation was significantly associated with a (Fig. 2B). For Complexes III and IV and ATP synthase (Complex poorer overall survival (OS) rate in patients with HCC (log-rank V), no obvious change was observed (Fig. 2B). Complex I activity test, P ¼ 0.027; Fig. 1B). The mean OS time of the PDSS2 groups was then determined by measuring the oxidation of NADH to þ with and without downregulation was 41.366 months [95% NAD using a Complex I enzyme activity assay kit. The result confidence interval (CI), 33.997–48.736) and 52.197 months demonstrated that the activity of Complex I increased in PDSS2- (95% CI, 43.956–60.438), respectively. The multivariate analysis transfected cells compared with that in empty vector-transfected result demonstrated that PDSS2 downregulation was an inde- cells (Fig. 2C). pendent prognostic factor for a poorer OS (P ¼ 0.028), independent of T status and stage (Table 1). Further analysis showed that PDSS2 induced metabolic reprogramming from glycolysis to downregulation of PDSS2 could be detected in early stages, mitochondrial respiration suggesting that PDSS2 might play an important role in HCC As the respiratory chain activity was upregulated in PDSS2- tumor initiation (Fig. 1C). The expression of PDSS2 was also transfected cells, we proposed that PDSS2 could promote mito- determined in four HCC cell lines (SMMC7721, QGY7703, chondrial respiration. Western blot results demonstrated that Huh7, and PLC8024) and an immortalized human liver cell line, the levels of pyruvate dehydrogenase (PDH), a key mitochondrial MIHA. The results demonstrated that PDSS2 expression was enzyme that converts pyruvate to acetyl-CoA for further metab- significantly decreased in HCC cell lines compared with that in olism through the Krebs cycle, and three molecules involved in the the MIHA cells (Fig. 1D; Supplementary Fig. S1A). Krebs cycle (isocitrate dehydrogenase, IDH; malate dehydrogenase, MDH2; and citrate synthase, CS) increased in PDSS2-trans- CoQ10 level decreased in HCC tumor tissues fected cells compared with those in vector-transfected cells, except As PDSS2 is important in CoQ10 synthesis and lower CoQ10 CS decreased in Huh7-PDSS2 cells (Fig. 2D–E and Supplementary plasma level has been reported in many cancers (26, 27, 33, 34), Fig. S2A). The SDH and IDH activity also increased in PDSS2- we next measured the CoQ10 level in HCC tumor tissues by IHC. overexpressing cells (Fig. 2F and G). Informative IHC data were obtained from 211 tumor tissues and After reintroduction of PDSS2, both the glucose uptake and 165 nontumor tissues. In 144 cases with both tumorous and lactate level decreased (Fig. 3A). The oxygen consumption rate is

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4473

Li et al.

Figure 1. Downregulation of PDSS2 and CoQ10 was frequently detected in HCCs. A, Representative IHC images of PDSS2 staining in HCC tumor tissues and paired nontumor tissues (original magniﬁcation, 200). B, Kaplan–Meier curve showing the overall survival rates in patients with HCC with or without PDSS2 downregulation. C, The percentage of PDSS2 downregulation in HCCs at different clinical stages determined by IHC. The numbers above the bars indicate the number of PDSS2 downregulated cases/total cases in each stage. D, PDSS2 expression was evaluated by IHC in HCC cell lines (SMMC7721, QGY7703, Huh7, and PLC8024) and an immortalized human liver cell line MIHA (original magniﬁcation, 200). E, Representative IHC images of CoQ10 staining in HCC tumor and paired nontumor tissues. F, CoQ10 was measured in 13 pairs of HCC tumor tissues and paired nontumor tissues by ultra high-performance liquid chromatography tandem mass spectrometric analysis.

another major biochemical parameter indicating glycolytic activ- maximal respiration also increased in 7721-PDSS2 cells com- ity. As shown in Fig. 3B, reintroduction of PDSS2 in SMMC7721, pared with that in vector control cells (P < 0.01; Supplementary QGY7703, and Huh7 cells led to a nearly 25% increase in the Fig. S2B). As glucose uptake, lactate production and oxygen oxygen consumption rate compared with that in control cells. The consumption are three major biochemical parameters for

Table 1. Univariate and multivariate analyses of different prognostic variables in patients with HCC Univariate analysisa Multivariate analysisa Variable HR (95% CI) P HR (95% CI) P Gender 1.208 (0.584–2.495) 0.611 Age 0.778 (0.488–1.240) 0.291 T status T1-T2 1 T3-T4 2.796 (1.842–4.243) <0.001 2.104 (1.332–3.321) 0.001 Stage I–II 1 III–IV 3.098 (2.014–4.766) <0.001 2.327 (1.458–3.714) <0.001 PDSS2 downregulation in T 1.601 (1.045–2.454) 0.031 1.622 (1.054–2.495) 0.028 Abbreviations: CI, conﬁdence interval; HR, hazard ratio. aCox regression model.

4474 Cancer Res; 78(16) August 15, 2018 Cancer Research

PDSS2 Deﬁciency Induces Hepatocellular Carcinoma

Figure 2. Reintroduction of PDSS2 enhanced mitochondrial respiration. A, Representative IHC images of CoQ10 in PDSS2- and vector-transfected cells (original magniﬁcation, 200). B, The protein levels of mitochondrial complexes IV were determined by Western blotting in HCC cell line derivatives (PDSS2-and vector control). C, Complex I activity was compared between PDSS2- and vector-transfected cells (left) and the data are summarized (right). V, vector control; P, PDSS2. , P < 0.01. D, Schematic diagram of key factors involved in glycolysis and the Krebs cycle. E, The molecules involved in glycolysis (TPI, GAPDH, and LDH), PDH, and molecules associated with the Krebs cycle (IDH, MDH2, and CS) were determined by Western blotting. F, SDH activity was compared between PDSS2- and vector-transfected cells. G, IDH activity was compared between PDSS2- and vector-transfected cells.

estimating glycolytic activity (36), the ratio of glycolytic indexes upregulate mitochondrial respiration and down-regulate glycol- upon reintroduction of PDSS2 was calculated. The ratio of ysis. To examine whether the ATP level changed in the PDSS2- glycolytic indexes in PDSS2-transfected cells relative to that in transfected cells after the glycolytic activity decreased in the cells, empty vector-transfected cells was 0.47:1 (SMMC7721), 0.45:1 we compared the cellular ATP levels in cells between PDSS2- and (QGY7703), and 0.38:1 (Huh7), indicating that the glycolytic vector-transfected cells. The reintroduction of PDSS2 increased activity decreased in PDSS2-transfected cells. Taken together, the production of cellular ATP compared with that in vector- these data suggest that the reintroduction of PDSS2 could transfected cells (Fig. 3C). As the glycolytic index decreased to

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4475

Li et al.

Figure 3. PDSS2 inhibited aerobic glycolysis and tumorigenicity. Reintroduction of PDSS2 could decrease glucose uptake and lactate secretion (A) and increase oxygen consumption (B) and the ATP level (C) compared with that in control cells. D, IHC staining confirmed the expression of ectopic PDSS2 in 7,721 cells (original magnification, 200). E, The representative pictures and summary of the foci formation assay (left) and the colony formation assay in soft agar (right) demonstrated that the anchorage-dependent and anchorage-independent growth ability was inhibited in 7721-PDSS2 cells compared with those in control cells (7721-Vec). F–H, An in vivo tumor formation assay demonstrated that PDSS2 could inhibit tumor growth in nude mice. Compared with tumors induced by 7721-Vec cells (n ¼ 6), tumors induced by 7721-PDSS2 cells (n ¼ 6) showed a smaller size (F), slower growth rate and lower tumor weight (G) 5 weeks after subcutaneous injection. H, Representative images of xenograft sections stained with PDSS2 antibody (original magnification, 200).

30% to 50% in PDSS2-transfected cells, our data suggest that ATP The results showed that the reintroduction of PDSS2 (Fig. 3D; was generated through mitochondrial oxidative phosphorylation Supplementary Fig. S3A–S3B) could significantly inhibit foci rather than through glycolysis. We also detected an increased formation (P < 0.01) and colony formation in soft agar (P < abundance of metabolic intermediates specific to the tricarboxylic 0.01) in three tested cells (SMMC7721, QGY7703 and Huh7) acid (TCA) cycle and decreased pyruvate in PDSS2-transfected compared with that in their control cells (Fig. 3E; Supplementary SMMC7721 cells (Supplementary Fig. S2C). This finding is con- Fig. S3C–S3F). The cell-cycle distribution results indicated that sistent with the mitochondrial respiration activation results in compared with vector-transfected cells, 7721-PDSS2 cells were 7721-PDSS2 cells. arrested at the G1–S checkpoint, manifested as an accumulation in G1-phase and a decrease in S-phase cells. The G2 distribution was PDSS2 suppressed tumor growth in vitro and in vivo also decreased significantly in 7721-PDSS2 cells (Supplementary Metabolic reprogramming was thought to be a consequence of Fig. S4A). The apoptotic index did not differ between 7721-PDSS2 rapid cell proliferation; however, recent findings have suggested cells and 7721-Vec cells (Supplementary Fig. S4B). The in vivo that it might be a driving factor in tumorigenesis (12, 13). tumorigenic assay was performed by subcutaneously injecting Therefore, both in vitro and in vivo functional assays were used PDSS2- and vector-transfected cells into the right and left flanks of to investigate the potential role of PDSS2 in HCC pathogenesis. nude mice (n ¼ 6), respectively. Tumor formation in nude mice

4476 Cancer Res; 78(16) August 15, 2018 Cancer Research

PDSS2 Deﬁciency Induces Hepatocellular Carcinoma

was monitored for 5 weeks, the mice were sacrificed, and the compared with MIHA-C (Fig. 4A, Supplementary Fig. S5D). tumor volume and tumor weight were examined. The results Compared with vector control cells, MIHA-KD cells displayed showed that the reintroduction of PDSS2 could significantly decreased PDH, IDH, and MDH2 and increased TPI1 protein inhibit tumor growth (P < 0.01) compared with that in control levels (Fig. 4B; Supplementary Fig. S5E). SDH activity also mice (Fig. 3F–G). IHC staining showed that the expression of decreased in the MIHA-KD cells compared with that in the PDSS2 was detected only in the tumors induced by PDSS2- vector control cells (P < 0.05; Fig. 4C). IDH activity decreased transfected cells (Fig. 3H). The tumor formation assay was repeat- slightly in MIHA-KD cells compared with that in vector control ed in BEL7402 cells, and the results were consistent with the above cells (Fig. 4C). referenced results (Supplementary Fig. S4C–S4D). MIHA-KD1 showed increased glucose uptake and lactate production and decreased oxygen consumption (Fig. 4D). The Silencing PDSS2 enhanced aerobic glycolysis and MIHA-KD1:MIHA-C ratio of glycolytic index was 4.75:1, indicat- tumorigenicity in immortalized liver cells ing that PDSS2 knockdown cells had approximately 4-fold higher To directly test whether PDSS2 is functionally important for glycolytic activity than the vector control cells. We also noticed HCC tumorigenesis, we used short-hairpin RNA (shRNA) to that the ATP level increased in MIHA-KD1 cells (Fig. 4D), which stably silence its expression in the MIHA immortalized liver cell might be the result of increased glycolysis to produce more ATP to line (Supplementary Fig. S5A–S5B). The knockdown efficiency facilitate cell proliferation. Taken together, the results indicate that was also confirmed in Huh7 cells (Supplementary Fig. S5C). PDSS2 deficiency promotes a switch to a state of glucose addic- Mitochondrial complex I decreased slightly in MIHA-KD cells tion, a hallmark of cancers undergoing aerobic glycolysis.

Figure 4. Silencing PDSS2 increased aerobic glycolysis and tumorigenicity. A, The protein levels of mitochondrial complexes IV were determined by Western blotting in MIHA-KD1 and MIHA-C cells. Actin was used as a loading control. B, The levels of PDH and the molecules associated with glycolysis and the Krebs cycle were determined by Western blotting. Actin was used as a loading control. C, SDH activity and IDH activity were compared between MIHA-KD1 and MIHA-C cells. D, Silencing PDSS2 expression in MIHA-KD1 cells significantly increased the glucose uptake and lactate secretion, decreased the oxygen consumption, and increased the ATP level compared with those in control cells. E and F, Knockdown of PDSS2 was able to significantly increase the cell growth rate (E)and foci formation frequency (F). G, MIHA-KD1 and control cells were subcutaneously injected into the left and right flanks of nude mice (n ¼ 6), respectively. Tumor weight was compared 3 weeks after inoculation. H, Expression of PDSS2 was detected by IHC in xenograft sections (original magnification, 400).

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4477

Li et al.

Figure 5. Silencing PDSS2 increased chromosomal instability. A, Representative images and summary of the frequency of micronuclei in MIHA-KD1 and MIHA-C cells. White arrows, micronuclei formed in the cells. B, Representative images of spectral karyotyping of one-month consecutive cultured cells (MIHA and MIHA-KD1). More chromosomal aberrations were observed in metaphase spreads of MIHA-KD1 cells than in those of MIHA cells. Green arrows, whole chromosome losses or gains; red arrows, novel structural aberrations. C, Summary of changes in chromosomal aberrations in 50 metaphase spreads of cells.

Both in vitro and in vivo functional assays were performed to cells (Fig. 6A). We sequenced the 50 and 30 flanking intron determine whether the metabolic shift toward aerobic glycolysis sequences of exon 2 in 8 HCC cases, and no SNV (single nucleoid by PDSS2 knockdown could lead to tumorigenesis. Silencing variation) or mutation was observed. The antibody used to detect PDSS2 in MIHA cells led to increased proliferation and greater PDSS2 could only detect the full-length PDSS2 but not the PDSS2- colony formation in the foci formation assay (Fig. 4E–F). Because Del2 (Fig. 6B). We synthesized an anti–PDSS2-Del2 antibody that MIHA cells cannot form xenografts in vivo, we mixed the cells with could specifically stain PDSS2-Del2 (Fig. 6B). The polyprenyl Matrigel and injected into nude mice. Our results demonstrated synthesis domain of PDSS2 protein was disrupted in PDSS2-Del2 that silencing PDSS2 in MIHA cells could promote tumor forma- (Fig. 6A and C; Supplementary Fig. S7B). Consequently, tion in nude mice compared with the tumor formation induced PDSS2-Del2 was incapable of changing CoQ10 levels in cells by control cells (Fig. 4G–H; Supplementary Fig. S6A). When with ectopic expression (Supplementary Fig. S7C). In addition, CoQ10 was added to PDSS2-deficient MIHA (MIHA-KD1) PDSS2-Del2 could not change the protein levels of the mitochon- cells, the glucose uptake ratio and lactate secretion ratio drial electron transport complexes in PDSS2-Del2 cells (Fig. 6D). þ (CoQ10 /CoQ10 ) decreased in MIHA-KD1 cells compared Furthermore, Complex I activity decreased slightly in PDSS2-Del2 with that in control cells (Supplementary Fig. S6B). Cell growth cells compared with that in vector control cells (Fig. 6D). decreased more significantly in MIHA-KD1 cells than in control Considering that many proteins involved in glycolysis and cells (Supplementary Fig. S6C). When cells were treated with a the mitochondrial respiratory chain were changed in the glycolysis inhibitor, 2-deoxy-D-glucose (2-DG), the cell growth PDSS2-overexpressing cells, we next studied whether proteins was inhibited more significantly in MIHA-KD1 cells than in were affected by PDSS2-Del2. Similar to PDSS2-overexpressing MIHA-C, indicating that MIHA-KD1 relied more on glycolysis cells, cells overexpressing PDSS2-Del2 had increased PDH and than did control cells (Supplementary Fig. S6D). Taken together, MDH2 protein levels but no significant changes in the IDH and CS these results indicate that the metabolic shift induced by PDSS2 protein levels (Fig. 6E). Unlike that in PDSS2-overexpressing cells, knockdown results in a more malignant phenotype that favored LDH increased significantly in PDSS2-Del2–overexpressing cells tumor cell growth. (Fig. 6E). We next evaluated whether PDSS2-Del2 could reduce glycolysis Silencing PDSS2 increased chromosomal instability and upregulate the mitochondrial respiratory chain as PDSS2 did. We also found that PDSS2 knockdown led to chromosomal As shown in Fig. 6F, glucose uptake showed no difference between instability. The frequency of the formation of micronuclei, which PDSS2-Del2 cells and control cells. However, lactate production contain unstable extranuclear chromosomal fragments, was used was higher in PDSS2-Del2 cells than in control cells. Oxygen to evaluate the chromosomal instability in MIHA-KD1 and consumption decreased in PDSS2-Del2 cells compared with vector control cells after stable cell lines were established. The that in PDSS2-transfected cells (Fig. 6F). To determine whether frequency of micronuclei formation was significantly higher PDSS2-Del2 had similar effects on tumor cell growth to those of (10.56% 0.26%) in MIHA-KD1 cells than in vector cells full-length PDSS2, SMMC7721 cells transfected with PDSS2-Del2 (MIHA-C; 8.41% 0.78%, P < 0.01; Fig. 5A). SKY analysis was and PDSS2 were inoculated into the left and right flanks of nude performed on MIHA-KD1, MIHA-C, and MIHA cells after 1 mice, respectively. The results showed that the tumors induced by month consecutive culture. The results demonstrated that 7721-PDSS2-Del2 grew much faster than the tumors induced the chromosomal aberrations in the MIHA-KD1 cells were by 7721-PDSS2 cells (Fig. 6G–H), suggesting that the tumor- significantly higher than that in the vector and parental cells suppressive ability was absent in 7721-PDSS2-Del2 cells. (P < 0.01; Fig. 5B and C). Discussion PDSS2-Del2 is a new variant of PDSS2 We next investigated whether there were other variants of Metabolic dysregulation is one of the hallmarks of tumor cells PDSS2. The mRNA products of PDSS2 were cloned into the pMD to support rapid cell proliferation (37). In proliferative tumor 19-T vector, 30 clones were sequenced, and several different cells, changes occur in the metabolism of carbohydrates, lipids splicing isoforms were found, including PDSS2-Del2, in which and peptides to meet increased energy demands and support alternative splicing deleted exon 2 (Supplementary Fig. S7A). This nucleic acid and protein biosynthesis and membrane biogenesis isoform is exclusively expressed in HCC cells, but not in MIHA (38, 39). Here, we demonstrated that the downregulation of

4478 Cancer Res; 78(16) August 15, 2018 Cancer Research

PDSS2 Deﬁciency Induces Hepatocellular Carcinoma

Figure 6. PDSS2-Del2 was a new PDSS2 isoform. A, Different variants of PDSS2 were detected in three HCC cell lines (SMMC7721, QGY7703, and Huh7) and the MIHA immortalized liver cell line. Schematic showing the amino acid sequence of PDSS2 FL (full length) and PDSS2-Del2 (dark green, polyprenyl synthesis domain). B, The cell derivatives of 7721 (-Vec, -PDSS2 and -PDSS2-Del2) were detected by IHC with anti-PDSS2 and anti-Del2. IHC staining showed that Del2 protein could be detected by the anti-Del2 antibody but not by the anti-PDSS2 antibody (original magniﬁcation, 200). C, 3-D structure of PDSS2 (exon 2, purple). D, Del2 overexpression in 7,721 cells did not affect the protein levels of mitochondrial complexes IV. The Complex I activity decreased slightly in PDSS2-Del2 cells compared with that in vector control cells. E, The levels of PDH and molecules associated with glycolysis and the Krebs cycle were determined by Western blotting in Del2-transfected 7721 cells and vector control cells. F, Glucose uptake, lactate secretion, and oxygen consumption were compared between Del2-transfected and vector-transfected cells or PDSS2-transfected cells. G, The tumor volume and tumor weight demonstrated that xenograft inhibition ability was absent in Del2-transfected cells, in contrast to PDSS2-transfected cells. H, Representative images of xenograft sections stained with anti-PDSS2 and anti-Del2 antibodies (original magniﬁcation, 200).

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4479

Li et al.

PDSS2 might be a driving factor in HCC development by repro- Genomic instability is one of the most common hallmarks of gramming glucose metabolism from aerobic oxidation to aerobic cancer and can be responsible for the accumulation of mutations glycolysis and increasing chromosomal instability. Sequencing affecting tumor cell malignant properties and response to ther- analysis found that PDSS2 has at least 6 different splicing apies (42). Chromosomal alterations are frequently detected in isoforms. No coding proteins could be predicted for four variants HCCs, including loss of 1p, 4q, 6p, 16q, and 17p, and gain of 1q, due to loss of the start codon. Only two transcripts, the full-length 8q, and 20q (28, 43–45). Another interesting finding is that PDSS2 and a variant with exon 2 deleted, could encode proteins: knockdown of PDSS2 in MIHA cells could increase chromosomal PDSS2 and PDSS2-Del2, respectively. Sequencing and IHC data instability. The formation of micronuclei and frequency of chro- showed that the downregulation of PDSS2 mainly affected the mosomal rearrangements both increased significantly in PDSS2 full-length isoform. knockdown cells. In addition, we found that PDSS2 deficiency not As PDSS2 is one of the key enzymes in CoQ10 synthesis, we only increased genomic instability but also transformed immor- next studied the effect of PDSS2 on CoQ10 synthesis. Our data talized hepatocytes (MIHA) into malignant cells, which has been demonstrated that CoQ10 downregulation was significantly cor- demonstrated by both in vitro and in vivo functional assays. In related with PDSS2 downregulation in clinical HCC samples, summary, we found that PDSS2 deficiency might be a driving indicating that the downregulation of PDSS2 in HCC could factor in hepatocarcinogenesis through reprograming energy negatively affect CoQ10 biosynthesis. CoQ10 deficiency has been metabolism from mitochondrial respiration to aerobic glycolysis, reported in many solid tumors, and a recent long-term clinical increasing chromosomal instability, and finally transforming a study demonstrated that patients with breast cancer who received normal cell into a malignant cell. adjuvant CoQ10 therapy showed a better outcome (40). Because the PDSS2 knockout mouse reveals mitochondrial defects and Disclosure of Potential Conflicts of Interest CoQ10 is essential for electron transport in the mitochondrial No potential conflicts of interest were disclosed. respiratory chain and antioxidant defense (24, 41), the effect of PDSS2 reintroduction on the expression level and bioactivity of Authors' Contributions mitochondrial respiratory chain components was investigated. Conception and design: Y. Li, Y. Hu, X.-Y. Guan Our results suggested that Complex I might have a greater con- Development of methodology: S. Lin, Z. Cai Acquisition of data (provided animals, acquired and managed patients, tribution to mitochondrial respiration changes in cells with PDSS2 provided facilities, etc.): Y. Li, L. Li, Z. Tang, X. Ban, T. Zeng, Y. Zhou, Y. Zhu, ectopic expression. W. Deng, X. Zhang, D. Xie, Y. Yuan Our next question was what will happen when Complex I Analysis and interpretation of data (e.g., statistical analysis, biostatistics, activity decreases? Our results showed that reintroduction of computational analysis): Y. Li, Z. Tang, S. Gao, P. Huang PDSS2 in HCC cell lines could increase the expression of the Writing, review, and/or revision of the manuscript: Y. Li, X.-Y. Guan Krebs cycle-related proteins PDH, IDH, MDH2 and CS, as well as Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J. Li SDH and IDH activity. This finding suggests that the PDSS2 Study supervision: Y. Li, X.-Y. Guan deficiency might inhibit TCA activity because the decreased activity of Complex I blocked the electron transport in the mitochon- Acknowledgments drial respiratory chain, which subsequently inhibited TCA cycle We thank Dr. Tao Wang (Guangzhou Institutes of Biomedicine and Health, activity. Therefore, we hypothesized that the PDSS2 deficiency Chinese Academy of Sciences) for his technical support. This work was might switch glucose metabolism from mitochondrial respiration supported by National Key R&D Program of China [(2017YFC1309000 to to glycolysis. A further study demonstrated that PDSS2 was able to X.Y. Guan and Y. Li), NSFC (81472255 to Y. Li), 81472250 (to X.Y. Guan)], decrease glucose uptake and lactate secretion and increase the National Key Sci-Tech Special Project of Infectious Diseases (2018ZX10723204- PDSS2 fi 006-005 to X.Y. Guan), GDNSF (2014A030313071 to Y. Li), GECI (M201511 oxygen consumption rate, indicating that the de ciency to Y. Li), GDSTP (2016A020214008 to Y. Li), SYSUIP (16ykjc34 to Y. Li), Hong could indeed increase glycolysis, which is beneficial for biomass Kong RGC Collaborative Research Funds (C7038-14G, HKBU5/CRF/10 to X.Y. production and tumor cell growth. All these findings support the Guan), RGC GRF Funds (767313 and 17143716 to X.Y. Guan), Theme-based possibility that PDSS2 deficiency is a driving factor in HCC Research Scheme Fund (T12-704/16-R to X.Y. Guan). X.-Y. Guan is a Sophie YM carcinogenesis through reprogramming glucose metabolism from Chan Professor in Cancer Research. mitochondrial respiration and oxidative phosphorylation to aerobic glycolysis. If this hypothesis is true, PDSS2 deficiency should The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in also affect cell behavior besides metabolic reprogram of glucose. accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Indeed, our functional studies demonstrated that reintroduction fi of PDSS2 into HCC cell lines could signi cantly suppress tumor Received July 18, 2017; revised December 12, 2017; accepted June 19, 2018; cell growth in vitro and in vivo. published first July 2, 2018.

References 1. Levine AJ, Puzio-Kuter AM. The control of the metabolic switch in are associated with a poor prognosis in patients with oral squamous cell cancers by oncogenes and tumor suppressor genes. Science 2010;330: carcinoma. Cancer 2003;97:1015–24. 1340–4. 5. Mochiki E, Kuwano H, Katoh H, Asao T, Oriuchi N, Endo K. Evaluation of 2. Warburg O. On the origin of cancer cells. Science 1956;123:309–14. 18F-2-deoxy-2-ﬂuoro-D-glucose positron emission tomography for gastric 3. Gasparre G, Porcelli AM, Lenaz G, Romeo G. Relevance of mitochondrial cancer. World J Surg 2004;28:247–53. genetics and metabolism in cancer development. Cold Spring Harbor 6. Schreurs LM, Smit JK, Pavlov K, Pultrum BB, Pruim J, Groen H, et al. Perspect Biol 2013;5. pii: a011411. Prognostic impact of clinicopathological features and expression of bio- 4. Kunkel M, Reichert TE, Benz P, Lehr HA, Jeong JH, Wieand S, et al. markers related to (18)F-FDG uptake in esophageal cancer. Ann Surg Oncol Overexpression of Glut-1 and increased glucose metabolism in tumors 2014;21:3751–7.

4480 Cancer Res; 78(16) August 15, 2018 Cancer Research

PDSS2 Deﬁciency Induces Hepatocellular Carcinoma

7. Xi Y, Guo R, Hu J, Zhang M, Zhang X, Li B. 18F-fluoro-2-deoxy-D-glucose 27. Palan PR, Mikhail MS, Shaban DW, Romney SL. Plasma concentrations of retention index as a prognostic parameter in patients with pancreatic coenzyme Q10 and tocopherols in cervical intraepithelial neoplasia and cancer. Nucl Med Commun 2014;35:1112–8. cervical cancer. Eur J Cancer Prev 2003;12:321–6. 8. Podoloff DA, Advani RH, Allred C, Benson AB 3rd, Brown E, Burstein HJ, 28. Huang SF, Hsu HC, Fletcher JA. Investigation of chromosomal aberrations et al. NCCN task force report: positron emission tomography (PET)/ in hepatocellular carcinoma by fluorescence in situ hybridization. Cancer computed tomography (CT) scanning in cancer. J Nat Compr Cancer Genet Cytogenet 1999;111:21–7. Network 2007;5:S1–22. 29. Chen P, Zhao SH, Chu YL, Xu K, Zhu L, Wu Y, et al. Anticancer activity of 9. Doherty JR, Cleveland JL. Targeting lactate metabolism for cancer thera- PDSS2, prenyl diphosphate synthase, subunit 2, in gastric cancer tissue and peutics. J Clin Invest 2013;123:3685–92. the SGC7901 cell line. Anticancer Drugs 2009;20:141–8. 10. Walenta S, Wetterling M, Lehrke M, Schwickert G, Sundfor K, Rofstad EK, 30. Wang Y, Wu MC, Sham JS, Zhang W, Wu WQ, Guan XY. Prognostic et al. High lactate levels predict likelihood of metastases, tumor recurrence, significance of c-myc and AIB1 amplification in hepatocellular carcinoma. and restricted patient survival in human cervical cancers. Cancer Res A broad survey using high-throughput tissue microarray. Cancer 2002;95: 2000;60:916–21. 2346–52. 11. Brizel DM, Schroeder T, Scher RL, Walenta S, Clough RW, Dewhirst MW, 31. Tang Z, Li S, Guan X, Schmitt-Kopplin P, Lin S, Cai Z. Rapid assessment of et al. Elevated tumor lactate concentrations predict for an increased risk of the coenzyme Q10 redox state using ultrahigh performance liquid chro- metastases in head-and-neck cancer. Int J Radiat Oncol Biol Phys 2001; matography tandem mass spectrometry. Analyst 2014;139:5600–4. 51:349–53. 32. Deng W, Tsao SW, Guan XY, Lucas JN, Si HX, Leung CS, et al. Distinct 12. Sebastian C. Tracking down the origin of cancer: metabolic reprogram- profiles of critically short telomeres are a key determinant of different ming as a driver of stemness and tumorigenesis. Crit Rev Oncog chromosome aberrations in immortalized human cells: whole-genome 2014;19:363–82. evidence from multiple cell lines. Oncogene 2004;23:9090–101. 13. De Miguel MP, Alcaina Y, de la Maza DS, Lopez-Iglesias P. Cell metabolism 33. Cobanoglu U, Demir H, Cebi A, Sayir F, Alp HH, Akan Z, et al. under microenvironmental low oxygen tension levels in stemness, prolif- Lipid peroxidation, DNA damage and coenzyme Q10 in lung cancer eration and pluripotency. Curr Mol Med 2015;15:343–59. patients–markers for risk assessment? Asian Pacific J Cancer Prev 14. Goldberg MS, Sharp PA. Pyruvate kinase M2-specific siRNA induces 2011;12:1399–403. apoptosis and tumor regression. J Exp Med 2012;209:217–24. 34. Cooney RV, Dai Q, Gao YT, Chow WH, Franke AA, Shu XO, et al. Low 15. Fantin VR, St-Pierre J, Leder P. Attenuation of LDH-A expression uncovers a plasma coenzyme Q(10) levels and breast cancer risk in Chinese women. link between glycolysis, mitochondrial physiology, and tumor mainte- Cancer Epidemiol Biomark Prev 2011;20:1124–30. nance. Cancer Cell 2006;9:425–34. 35. Fisar Z, Hroudova J, Singh N, Koprivova A, Maceckova D. Effect of 16. Telang S, Yalcin A, Clem AL, Bucala R, Lane AN, Eaton JW, et al. simvastatin, coenzyme Q10, resveratrol, acetylcysteine and acetylcarnitine Ras transformation requires metabolic control by 6-phosphofructo-2- on mitochondrial respiration. Folia Biol 2016;62:53–66. kinase. Oncogene 2006;25:7225–34. 36. Xu RH, Pelicano H, Zhou Y, Carew JS, Feng L, Bhalla KN, et al. Inhibition of 17. Miller DM, Thomas SD, Islam A, Muench D, Sedoris K. c-Myc and cancer glycolysis in cancer cells: a novel strategy to overcome drug resistance metabolism. Clin Cancer Res 2012;18:5546–53. associated with mitochondrial respiratory defect and hypoxia. Cancer Res 18. Bensaad K, Tsuruta A, Selak MA, Vidal MN, Nakano K, Bartrons R, et al. 2005;65:613–21. TIGAR, a p53-inducible regulator of glycolysis and apoptosis. Cell 37. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 2006;126:107–20. 2011;144:646–74. 19. Feng Z, Hu W, de Stanchina E, Teresky AK, Jin S, Lowe S, et al. The regulation 38. Robey RB, Weisz J, Kuemmerle NB, Salzberg AC, Berg A, Brown DG, et al. of AMPK beta1, TSC2, and PTEN expression by p53: stress, cell and tissue Metabolic reprogramming and dysregulated metabolism: cause, conse- specificity, and the role of these gene products in modulating the IGF-1- quence and/or enabler of environmental carcinogenesis? Carcinogenesis AKT-mTOR pathways. Cancer Res 2007;67:3043–53. 2015;36:S203–31. 20. Hu W, Zhang C, Wu R, Sun Y, Levine A, Feng Z. Glutaminase 2, a novel p53 39. DeBerardinis RJ, Lum JJ, Hatzivassiliou G, Thompson CB. The biology of target gene regulating energy metabolism and antioxidant function. PNAS cancer: metabolic reprogramming fuels cell growth and proliferation. Cell 2010;107:7455–60. Metab 2008;7:11–20. 21. Matoba S, Kang JG, Patino WD, Wragg A, Boehm M, Gavrilova O, et al. p53 40. Bjorklund G. The adjuvant nutritional intervention in cancer (ANICA) regulates mitochondrial respiration. Science 2006;312:1650–3. Trial. Nutr Cancer 2015:1–4. 22. Guan XY, Zhang HE, Zhou H, Sham JS, Fung JM, Trent JM. Characterization 41. Lu S, Lu LY, Liu MF, Yuan QJ, Sham MH, Guan XY, et al. Cerebellar defects of a complex chromosome rearrangement involving 6q in a melanoma cell in Pdss2 conditional knockout mice during embryonic development and line by chromosome microdissection. Cancer Genet Cytogenet 2002;134: in adulthood. Neurobiol Dis 2012;45:219–33. 65–70. 42. Nieborowska-Skorska M, Kopinski PK, Ray R, Hoser G, Ngaba D, Flis S, 23. Fung JM, Smith R, Brown MA, Lau SH, Xie D, Lau GK, et al. Identification et al. Rac2-MRC-cIII-generated ROS cause genomic instability in chronic and characterization of a novel melanoma tumor suppressor gene on myeloid leukemia stem cells and primitive progenitors. Blood 2012; human chromosome 6q21. Clin Cancer Res 2009;15:797–803. 119:4253–63. 24. Quinzii CM, Lopez LC, Von-Moltke J, Naini A, Krishna S, Schuelke M, et al. 43. Hertz S, Rothamel T, Skawran B, Giere C, Steinemann D, Flemming P, et al. Respiratory chain dysfunction and oxidative stress correlate with severity of Losses of chromosome arms 4q, 8p, 13q and gain of 8q are correlated with primary CoQ10 deficiency. FASEB J 2008;22:1874–85. increasing chromosomal instability in hepatocellular carcinoma. Patho- 25. Jolliet P, Simon N, Barre J, Pons JY, Boukef M, Paniel BJ, et al. Plasma biology 2008;75:312–22. coenzyme Q10 concentrations in breast cancer: prognosis and therapeutic 44. Nishida N, Nishimura T, Ito T, Komeda T, Fukuda Y, Nakao K. Chromo- consequences. Int J Clin Pharmacol Ther 1998;36:506–9. somal instability and human hepatocarcinogenesis. Histol Histopathol 26. Rusciani L, Proietti I, Rusciani A, Paradisi A, Sbordoni G, Alfano C, 2003;18:897–909. et al. Low plasma coenzyme Q10 levels as an independent prognostic 45. Guan XY, Fang Y, Sham JS, Kwong DL, Zhang Y, Liang Q, et al. Recurrent factor for melanoma progression. J Am Acad Dermatol 2006;54: chromosome alterations in hepatocellular carcinoma detected by compar- 234–41. ative genomic hybridization. Gen Chromosom Cancer 2000;29:110–6.

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4481

PDSS2 Deficiency Induces Hepatocarcinogenesis by Decreasing Mitochondrial Respiration and Reprogramming Glucose Metabolism

Yan Li, Shuhai Lin, Lei Li, et al.

Cancer Res 2018;78:4471-4481. Published OnlineFirst July 2, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-17-2172

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2018/06/30/0008-5472.CAN-17-2172.DC1

Cited articles This article cites 43 articles, 12 of which you can access for free at: http://cancerres.aacrjournals.org/content/78/16/4471.full#ref-list-1

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/78/16/4471. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.