Evidence for an Important Role of CIDEA in Human Cancer Cachexia

Research Article

Jurga Laurencikiene,1 Britta M. Stenson,1 Elisabet Arvidsson Nordstro¨m,1 Thorhallur Agustsson,2 Dominique Langin,3,4,5 Bengt Isaksson,2 Johan Permert,2 Mikael Ryde´n,1 and Peter Arner1

Departments of 1Medicine and 2Surgery, Karolinska Institutet, Karolinska University Hospital, Huddinge, Stockholm, Sweden; 3Institut National de la Sante´et de la Recherche Me´dicale,U858, Obesity Research Laboratory, 4Louis Bugnard Institute, IFR31, Paul Sabatier University;and 5CHU de Toulouse, Biochemistry Laboratory, Biology Institute of Purpan, Toulouse, France

Abstract from subjects with cancer cachexia. This augments the lipolytic Loss of fat mass in cancer cachexia is linked to increased effects of the major stimulatory hormones in humans, which are adipocyte lipolysis; however, the fate of the excess fatty acids catecholamines and natriuretic peptides (11). (FA) generated by lipolysis is not known. We investigated if FAs are the most energy-rich molecules in the body. It is unclear the adipocyte-specific gene cell death–inducing DNA frag- how they are metabolized after lipolysis in cancer cachexia. mentation factor-A–like effector A (CIDEA) could be involved. Increased FA oxidation has been shown in cancer patients, which CIDEA mRNA expression was assessed in s.c. white adipose was more pronounced among those losing weight (9, 12). Skeletal tissue from 23 cancer cachexia patients, 17 weight-stable muscle is the major site for FA oxidation in humans, whereas in cancer patients, and 8 noncancer patients. CIDEA was also rodents, brown fat cells contribute as well. Recent rodent studies overexpressed in adipocytes in vitro. CIDEA expression was suggest that FA oxidation in white adipose tissue is an additional increased in cancer cachexia (P < 0.05) and correlated with important regulator of FA oxidation (13). In weight-stable humans, elevated levels of FAs and reported weight loss (P < 0.001). adipocytes contribute only little to total FA combustion (14, 15). CIDEA overexpression in vitro increased FA oxidation 2-to However, total energy expenditure in human fat cells (glucose and 4-fold (P < 0.01), decreased glucose oxidation by 40% FA oxidation) is significant and constitutes up to 14% of whole (P < 0.01), increased the expression of pyruvate dehydroge- body heat production (16). This percentage is markedly increased nase kinase (PDK) 1 and PDK4 (P < 0.01), and enhanced the following body weight reduction (17). phosphorylation (inactivation) of the pyruvate dehydrogenase The capacity of human white adipose tissue to oxidize FAs may complex (PDC). Inactivation of PDC facilitates FA oxidation by increase markedly if the oxidative machinery is altered (14). A key favoring the metabolism of FAs over glucose to acetyl-CoA. In factor in regulating the switch from glucose to FA oxidation is the accordance with the in vitro data, PDK1 and PDK4 expression pyruvate dehydrogenase complex (PDC;ref. 18). PDC plays a major correlated strongly with CIDEA expression in white adipose role in glucose metabolism by catalyzing the oxidative decarbox- tissue (P < 0.001). We conclude that CIDEA is involved in ylation of pyruvate to acetyl-CoA and NADH, which can then be adipose tissue loss in cancer cachexia and this may, at least in used for the tricarboxylic acid cycle and various biosynthetic part, be due to its ability to inactivate PDC, thereby switching processes. PDC is a highly organized multienzyme complex (19) substrate oxidation in human fat cells from glucose to FAs. and is inactivated via phosphorylation of the E1 (PDC-E1) subunit [Cancer Res 2008;68(22):9247–54] of the complex (20). Phosphorylation of PDC-E1 is regulated by four pyruvate dehydrogenase kinases (PDKs), which are hetero- Introduction dimeric enzymes with tissue-specific distribution (18, 19). The expression level of PDK4 is regulated by the nuclear receptor Cancer cachexia is associated with poor survival rate and peroxisome proliferator–activated receptor a,whichisalso impaired response to chemotherapy (1, 2). It is characterized by activated by FAs (21). PDC needs to be active when the complete loss of both adipose tissue and skeletal muscle (1, 2). The former oxidation of glucose is required for the generation of energy, often precedes and is more rapid than the latter (3). Several factors whereas its activity has to be suppressed when glucose is in short may induce loss of adipose tissue in cancer cachexia. However, increased lipid mobilization due to enhanced adipocyte lipolysis supply. Inactivation of PDC favors the preferential oxidation of seems to be a major course (1, 2, 4, 5). Lipolysis is the hydrolysis of long-chain FAs over glucose. triglycerides to glycerol and fatty acids (FA) in adipocytes, where We hypothesized that in response to the increased lipolysis in triglycerides constitute >95% of adipose mass. Increased lipolysis white fat cells of cancer cachexia patients, adipocytes may activate in cancer cachexia has been shown in vivo (6–10) and results in compensatory FA oxidative pathways and suppress glucose decreased fat cell volume and size. In agreement with this, fat cells oxidation to rid themselves of excess FAs. A putative candidate in cancer cachexia patients are smaller (6). The mechanism behind for such an adaptation is a novel gene termed cell death–inducing CIDEA CIDEA increased adipocyte lipolysis has recently been elucidated (6). It DNA fragmentation factor-a–like effector A ( ). was shown that the expression of the adipocyte hormone-sensitive belongs to a family of apoptosis genes with unclear function (22). lipase (the rate-limiting enzyme of lipolysis) is increased in fat cells However, recent studies suggest a role in lipid metabolism and body weight regulation. In rodents, CIDEA deficiency leads to increased FA oxidation in brown fat cells, acceleration of lipolysis, and resistance to diet-induced obesity (23, 24). In humans, unlike Requests for reprints: Peter Arner, Karolinska Institutet, Karolinska University rodents, the gene is predominantly expressed in white fat cells Hospital, Huddinge, CME-M61, SE-141 86 Stockholm, Sweden. Phone: 46-8-58582342; and inhibits adipocyte lipolysis (25, 26). In transcriptomic Fax: 46-8-58582347. E-mail: [email protected]. CIDEA I2008 American Association for Cancer Research. profiling study, was the most up-regulated gene in human doi:10.1158/0008-5472.CAN-08-1343 white adipose tissue following dietary weight reduction (27). www.aacrjournals.org 9247 Cancer Res 2008; 68: (22). November 15, 2008

Furthermore, a polymorphism in the CIDEA gene reduces the risk (Invitrogen) used for amplification were 5¶-GGGAATTCGCGCCATGGAGG- for obesity (28). 3¶ and 5¶-CTCTCGAGCAGCGTGACCAG-3¶, respectively, with restriction Eco Xho In the present study, the possible role of CIDEA in white adipose sites for RI and I underlined. Constructs were verified by DNA in vitro tissue for FA oxidation in vitro and in vivo was investigated. Cancer sequencing (MWG Biotech). Protein expression was verified by translation using a Transcend Non-Radioactive Translation Detection patients with or without cachexia and control (noncancer) patients System (Promega). were compared. The direct effect of CIDEA on FA and glucose Transient transfections. 3T3-L1 cells were grown to confluence and oxidation was investigated by overexpressing the gene in fat cells. differentiated for 3 d according to standard procedures. Cells were then seeded into 12-well plates (Corning, Inc.) at a density of f10,000/cm2. After 24 h in postdifferentiation medium, the 3T3-L1 adipocytes were transfected Materials and Methods using Lipofectamine and Plus transfection reagents (Invitrogen) according Patients. All patients scheduled for gastrointestinal cancer operation to the manufacturer’s instructions. Transfection efficiency in 3T3-L1 cells between January 2004 and December 2006 (about 450 patients) were was optimized using a green fluorescence protein–containing plasmid. evaluated for the study and patients who (a) were fit in spite of their cancer, Transfection efficiency was 30% to 40%. One microgram of CIDEA (b) had not received prior anticancer treatment, and (c) were willing to expression vector, 6 AL of Plus reagent, and 4 AL of Lipofectamine were participate were included (n = 48). The remaining ones were excluded due used per well for transfection. Cells were analyzed after 48 h. After to severe illness, disinclination, communication problems, and logistic overexpression, levels of hCIDEA mRNA reached those detected in human reasons. The study was approved by the Regional Ethics committee. The mature adipocytes. Human adipocytes differentiated from s.c. precursor investigation was explained in detail to each patient and written informed cells were transfected as described previously (32). Cell Death Detection consent was obtained. None of the patients had jaundice. They were divided ELISA Plus (Roche Diagnostics GmbH) was used to determine apoptosis into three groups based on diagnosis after surgery. One group had cancer and necrosis. cachexia (n = 23), which was defined as gastrointestinal cancer with Western blot and immunoprecipitation. 3T3-L1 cells were transfected unintentional weight loss of >5% of the habitual weight during the previous with CIDEA expression vector or pcDNA3 as described above. After 48 h, the 3 mo or >10% unintentional weight loss during the previous 6 mo (6). The medium was removed and cells were lysed in RIPA buffer [50 mmol/L Tris- primary location of cancer was pancreas (n = 10), esophagus (n = 10), HCl (pH 7.4), 150 mmol/L NaCl, 1 mmol/L EDTA 1 mmol/L PMSF, 1% stomach (n = 2), and colon (n = 1). One group (n = 17) consisted of patients Triton X-100, 1% sodium deoxycholate, 0.1% SDS] containing phosphatase with gastrointestinal cancer who reported no important weight change inhibitors (1 mmol/L sodium orthovanadate and 1 mmol/L sodium during the last year. The localization of malignancy was pancreas (n = 5), fluoride) and proteinase inhibitor cocktail (Roche Diagnostics GmbH). Cell esophagus (n = 1), stomach (n = 3), colon (n = 4), gall bladder (n = 2), and lysates were centrifuged at 14,000 rpm for 15 min at +4jC;the supernatant liver (n = 2). The control group (n = 8) contained subjects with was removed and the protein concentration was measured using the prediagnosed gastrointestinal cancer but who did not have a malignancy bicinchoninic acid protein assay kit (Pierce). according to final histologic evaluations. The noncancer control patients Four hundred micrograms of protein lysate were used for PDC had chronic pancreatitis (n = 5) or cholecystitis (n = 3). The study was immunoprecipitation according to the manufacturer’s instructions (PDC designed as ‘‘intention to compare’’;therefore, all included subjects were pull-down kit, Mitosciences). PDC was eluted with 1% SDS. Gel kept in the analyses despite a few missing values of some of the electrophoresis and blotting were done as described previously (33). Anti- measurements. phosphoserine antibody (Zymed Laboratories) was used at the concentra- Clinical examination. The patients came to the laboratory after an tion of 2 Ag/mL to detect phosphorylation of the PDC-E1 subunit. overnight fast. Height, weight, body composition by bioimpedance using Antibodies that were used for CIDEA detection were anti-CIDEA polyclonal Quad Scan 4000 (Bodystat Ltd.), and indirect calorimetry using Deltatrac antibodies C-16 and N-19 from Santa Cruz Biotechnology, ab5692 from (Datex-Engstroms) were determined. A venous blood sample was obtained Abcam, and AB16508 and AB16922 from Chemicon International. for the measurement of lipids, glycerol, FA, albumin, and transferrin. Palmitate oxidation. 3T3-L1 cells were transfected with CIDEA Nutritional status was assessed by using a standardized questionnaire for expression vector or control pcDNA3 vector. After 48 h, the medium oncology termed Subjective Global Assessment (SGA;ref. 29). Tumor stage was removed and FA oxidation determined as described previously (34). was classified postoperationally as described previously (6). Indirect As a control, FA oxidation was stimulated for 3 h with 10 Amol/L calorimetry could not be done on one weight-stable cancer patient due m-chlorocarbonylcyanide phenylhydrazone (Sigma). CIDEA overexpression to technical reasons. To indirectly assess lipolytic activity in vivo, glycerol was controlled in each experiment, and palmitate oxidation was normalized and FA concentrations were divided by body fat weight (6). to total protein amount in each experimental group. Fat biopsies. After the clinical examination, an abdominal s.c. fat biopsy Glucose oxidation. Oxidation of glucose to CO2 in 3T3-L1 cells (f0.5 g) was obtained by needle biopsy as described (30). Tissue pieces differentiated to adipocytes was assayed as previously described (35) with were rapidly rinsed in saline, frozen in liquid nitrogen, and kept at À70jC some modifications. 3T3-L1 cells were transfected with CIDEA expression for later gene expression studies. Tissue pieces removed and frozen in this vector or empty pcDNA3 vector as described above. After 48 h, the cell way are free from damaged cells and blood (31). In three cachexia patients, medium was removed and the cells were rinsed with PBS. Krebs-Ringer- the quality of the RNA was not sufficient for mRNA analysis. HEPES (0.5 mL) containing 2% BSA and 0.55 mmol glucose was added. Cell Gene expression. Total RNA was extracted from adipose tissue or 3T3-L1 suspensions from three wells (1.5 mL) were transferred to Carbosorb E cells using the RNeasy mini kit (Qiagen GmbH). RNA quality was analyzed flasks containing 1.5 mL of Krebs-Ringer-HEPES with 2% BSA, 0.55 mmol 14 using the Agilent 2100 Bioanalyzer (Agilent Technologies) before cDNA glucose, and 0.4 ACi/mL D-[U- C]glucose (Amersham Biosciences). Flasks j 14 synthesis. cDNA synthesis was done on 1 Ag of total RNA using the were closed and incubated at 37 C with shaking for 1.5 h. CO2 was Omniscript RT kit (Qiagen) and random hexamer primers (Invitrogen). liberated by acidification with 10 N H2SO4, collected for 1 h on Carbosorb, Quantitative real-time PCR was done in an iCycler IQ (Bio-Rad Laboratories, and measured by scintillation counting. Inc.) using TaqMan probes (Assays-on-demand, Applied Biosystems). mRNA Statistical analysis. Values are mean F SD or range in text, tables, and levels in patient material were determined using a comparative Ct method, figures. Results were compared using ANOVA and appropriate post hoc 2 as described (6). In CIDEA overexpression experiments, cDNA from empty tests, m test, unpaired or paired t test, and linear or multiple regression vector–transfected cells was used as the calibrator and acidic ribosomal analysis. Gene expression values and values for body fat CIDEA were log10 protein 1 (Arp1) was used as the reference gene. transformed to obtain normal distribution. Kruskal-Wallis and Mann- CIDEA expression vector. CIDEA expression vector was obtained by Whitney tests were used to compare SGA and tumor score, respectively. A insertion of a PCR product corresponding to the human CIDEA cDNA into power calculation was made before starting to recruit subjects and was the pcDNA3.1(+) plasmid (Invitrogen). The forward and reverse primers based on our previously known distribution of CIDEA mRNA expression in

Cancer Res 2008; 68: (22). November 15, 2008 9248 www.aacrjournals.org

Table 1. Characteristics of study groups

Measure Cancer cachexia Weight-stable cancer Control P

Gender, M/F 20/3 12/5 4/4 0.11 Age, y 65 F 764F 10 59 F 7 0.24 BMI, kg/m2 22.6 F 4.2 25.7 F 3.5 28.0 F 7.6 0.016 Body fat, % 18 F 923F 737F 23 0.001 Body fat, kg 13.3 F 9.9 21.2 F 6.1 32.9 F 23.7 0.001 Lean body mass, kg 57.9 F 10.2 58.1 F 13.3 46.7 F 10.4 0.064 Weight loss, % of habitual weight 12.4 F 7.6 3.9 F 7.4 3.7 F 7.3 0.002 P-glucose, mmol/L 6.1 F 1.3 6.2 F 1.5 6.6 F 2.2 0.76 P-triglycerides, mmol/L 1.0 F 0.3 1.3 F 0.4 1.4 F 0.4 0.027 P-cholesterol, mmol/L 4.3 F 1.1 4.8 F 1.1 4.8 F 1.2 0.25 S-albumin, g/L 34.9 F 3.6 37.6 F 3.5 38.9 F 3.3 0.016 S-transferrin 2.1 F 0.3 2.4 F 0.4 2.5 F 0.4 0.031 P-glycerol, Amol/L/kg body fat 12.8 F 12.8 4.4 F 3.7 4.5 F 4.2 0.016 P-FA, Amol/L/kg fat 97 F 75 41 F 26 45 F 45 0.008 SGA score, points 10 (1–19) 3 (1–13) 4 (1–10) <0.001 Tumor stage, points 3 (1–4) 4 (0–4) — 0.036 Respiratory quotient 0.798 F 0.033 0.838 F 0.066 0.831 F 0.038 0.031 Resting energy expenditure, kcal/d 1,681 F 273 1,698 F 316 1,584 F 191 0.65

NOTE: Values are mean F SD or mean (range). They were compared by ANOVA, m2 test (for gender), Kruskal-Wallis test (for SGA), or Mann-Whitney test (for tumor stage). Abbreviations: P, fasting plasma;S, fasting serum.

white adipose tissue of healthy human subjects (27). Based on our previous Results study of cancer cachexia (6), it was decided to recruit control patients, Clinical findings. The clinical results are shown in Table 1. The weight-stable cancer patients, and cancer cachexia patients in the ratio of 1:2:3. This ratio was based on the assumption of a 10% experimental failure subjects in the three groups were of similar age. Body mass index rate, and for the mean and SD of CIDEA mRNA expression we had to recruit (BMI) was markedly decreased in subjects with cancer cachexia in at least 8 controls, 16 weight-stable, and 24 cachexia patients to detect a comparison with the other two groups;the weight-stable cancer 2-fold difference between cachexia and the other two groups at <0.05 group had an intermediate value. The same was true for percentage (ANOVA) with 80% power using Sample Power (SPSS, Inc.). of body fat and total amount of body fat. In contrast, lean body mass

Figure 1. CIDEA mRNA expression in white adipose tissue in cancer cachexia, weight-stable cancer,and control state (top left) and the relation between CIDEA mRNA and BMI (top right),body fat (bottom left),and body weight loss ( bottom right). Filled circles, cancer cachexia. Open circles, weight-stable cancer. Crosses, control. ANOVA is used in the top left graph; otherwise linear regression is used.

www.aacrjournals.org 9249 Cancer Res 2008; 68: (22). November 15, 2008

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 2008 American Association for Cancer Research. Cancer Research was similar in the two cancer groups. The cancer cachexia group parameters (Fig. 2A). Therefore, low RQ (indicative of high fat reported a marked decrease in their habitual weight. Markers of oxidation) was associated with a high gene expression of CIDEA anabolism (i.e., plasma triglycerides, serum albumin, and serum and could explain as much as 25% of the variation seen in RQ (i.e., transferrin values) were significantly lower and SGA scores were adjusted r2). Furthermore, reported weight loss was inversely significantly higher among cancer cachexia patients than in the correlated with RQ (Fig. 2B). RQ tended to correlate with BMI other two groups. There was no difference between the weight- (P = 0.052) and body fat (P = 0.077), but not with lean body mass stable cancer patients and control subjects with respect to these (P = 0.23;graphs not shown). Because CIDEA expression was parameters. Somewhat unexpectedly, tumor stage was more markedly influenced by BMI and total body fat, a multiple advanced in the weight-stable than in the weight-losing cancer regression analysis was also done. When either BMI or body fat group. Lipid mobilization was analyzed by measuring either plasma was used together with CIDEA as an independent regressor and RQ glycerol or plasma FA concentrations and corrected for total body was the dependent variable, only CIDEA was a significant regressor fat. It was markedly increased in the cancer cachexia group when for RQ (partial r = 0.54, P = 0.002 together with BMI;partial compared with the weight-stable cancer and control groups. r = 0.59, P = 0.002 together with log10 body fat). There was no effect The latter two groups did not differ between each other in lipid of age when this parameter was included in the multiple regression mobilization or respiratory quotient (RQ). In contrast, RQ was analysis. The reported decrease in body weight correlated strongly significantly decreased in the cancer cachexia group, which is and negatively with RQ (Fig. 2B). indicative of increased FA oxidation. Total resting energy expendi- Transient CIDEA overexpression increases FA oxidation and ture was similar in all three groups. decreases glucose oxidation but does not cause cell death in CIDEA is highly expressed in white adipose tissue of cancer adipocytes. To investigate the role of CIDEA in regulating cachexia patients. The average mRNA expression of CIDEA was substrate oxidation, a human CIDEA expression vector was much higher in cancer cachexia patients than in patients from the transfected into differentiated 3T3-L1 cells and human adipocytes. other two groups, which, in turn, did not differ between each other Forty-eight hours posttransfection, FA oxidation was determined (Fig. 1). Figure 1 also shows a strong negative correlation between using 14C palmitate. Overexpression of CIDEA enhanced palmitate BMI or total body fat and CIDEA mRNA levels in the entire oxidation two to four times in 3T3-L1 cells (Fig. 2C) and up to two material put together. Furthermore, CIDEA mRNA correlated times in differentiated human adipocytes (P < 0.01;graph not strongly and positively with the reported decrease in body weight shown). A low RQ indicates the use of FAs as the preferred in the whole material. However, there was no correlation between substrate for energy production. We hypothesized that in CIDEA- lean body mass and CIDEA expression (r = 0.18;graph not shown). transfected cells the oxidation of glucose to CO2 could be CIDEA expression levels correlate negatively with RQ. The decreased. Therefore, a glucose oxidation assay was done in 3T3- relationship between in vivo fat oxidation and CIDEA expression L1 cells transfected with either a CIDEA expression vector or a was examined by comparing RQ with CIDEA mRNA in all subjects. control vector (Fig. 2D). We observed that glucose oxidation in A strong inverse relationship was observed between these two CIDEA-overexpressing cells was 40% lower (P < 0.01) than in

Figure 2. Relationship between CIDEA, RQ,FA,and glucose oxidation. Relationships between CIDEA mRNA expression and respiratory quotient (A)and between body weight loss and respiratory quotient (B) were analyzed by linear regression. See legend to Fig. 1 for further details. Effect of CIDEA overexpression on FA oxidation (C) and glucose oxidation (D). Differentiated 3T3-L1 cells were transfected with CIDEA or empty vector (control). FA oxidation and glucose oxidation were measured as described in Materials and Methods. The amount of palmitate or glucose oxidation is adjusted for total protein amount and expressed as fold change over the control (**, P < 0.01; n = 5–7).

Cancer Res 2008; 68: (22). November 15, 2008 9250 www.aacrjournals.org

extracts or different amounts of PDC in specific tissues. We failed to visualize CIDEA as a single protein band from extracts of human adipose tissue despite trying four different commercial antibodies (figure not shown). CIDEA expression correlates positively with expression levels of PDK1 and PDK4 in human white adipose tissue. PDK mRNA expression was examined in human white adipose tissue. There was a slight increase in the expression of PDK1 (P = 0.03) and PDK4 (P = 0.06) in cancer cachexia when compared with weight-stable cancer (graph not shown). In the entire examined white adipose tissue material, CIDEA mRNA expression correlated strongly and positively with mRNA expression levels of PDK1 (r = 0.7, P < 0.0001) and PDK4 (r = 0.8, P < 0.0001;Fig. 4). As much as 46% and 58% of the variation seen in PDK1 and PDK4, respectively, could be explained by the variation in CIDEA (i.e., adjusted r2). These correlations are specific for CIDEA because the mRNA expression of other genes involved in FA oxidation, uncoupling protein (UCP)-1, and AMPK did not correlate with CIDEA expression (Fig. 4). One subject had a very high relative CIDEA expression (>10 log 2);however, the correlation between Figure 3. Effect of CIDEA overexpression on the PDH complex regulation in mRNA for CIDEA and PDK1 or PDK4 remained significant 3T3-L1 cells. Top, relative expression of mRNA for PDK1 and PDK4 in (P < 0.01, r = 0.5–0.6) even when this value was excluded from CIDEA-transfected cells (**, P < 0.01; n = 6). Bottom, detection of phosphorylated PDH-E1 component in CIDEA-transfected cells. The two lanes the analysis. to the left show protein extracts used for immunoprecipitation (input) and the three lanes to the right show immunoprecipitated PDH-E1 detected by anti-phosphoserine antibodies. Three independent experiments were run with similar results. Discussion Results from this study suggest that CIDEA is involved in the altered lipid metabolism observed in cancer cachexia. We also control cells. Transient overexpression of CIDEA did not induce confirm that lipolysis, FA mobilization, and lipid oxidation are cell death in 3T3-L1 fat cells, and if anything, it tended to decrease increased in cancer cachexia when compared with weight-stable apoptosis (P = 0.10) and necrosis (P = 0.06;graph not shown). cancer patients and noncancer patients. The cancer cachexia CIDEA increases expression of PDK1 and PDK4 in 3T3-L1 patients showed clear signs of malnutrition. However, they adipocytes. The expression levels of several genes involved in displayed no evidence of lean body mass loss. Because loss of fat h-oxidation and FA turnover were examined. These included is faster than loss of skeletal muscle in cancer cachexia (3), it is carnitine palmitoyl-CoA transferase A (Cpt-1A), Cpt-1B, carnitine- likely that we investigated an early phase of the condition when acylcarnitine translocase (CACT), and several mitochondrial matrix the patients had only lost adipose tissue. The two noncachectic enzymes [acyl-CoA long-chain synthetase 1 (ACSL1) and enoyl-CoA groups were comparable in all investigated parameters except for hydratase/3-hydroxyacyl-CoA dehydrogenase (EHHADH)]. The body composition, with weight-stable cancer patients having less mRNA levels of all of these genes remained unchanged in body fat than noncancer patients. CIDEA-transfected 3T3-L1 adipocytes (values not shown). In Human CIDEA is predominantly expressed in white adipocytes contrast, mRNA expression levels of PDK1 and PDK4 were and there is no or very low expression of CIDEA in human skeletal increased in CIDEA-transfected 3T3-L1 cells compared with muscle (26). In this study, mRNA expression of CIDEA in white control vector–transfected cells (Fig. 3). Expression levels of adipose tissue was markedly increased in patients with cancer PDK2, PDK3, and AMP-activated kinase (AMPK) remained cachexia. This is most likely explained by their low BMI and unchanged (values not shown). PDK1 showed the highest reduced adipose mass because these two parameters correlated expression level in 3T3-L1 cells when compared with other PDKs strongly and negatively with CIDEA expression. In contrast, lean (values not shown). body mass was not related to CIDEA expression. It seems that total CIDEA overexpression increases the phosphorylation of fat mass and adipocyte CIDEA expression are tightly coregulated. PDC. To assess whether increased expression of PDKs has any This is supported by previous findings in white adipose tissue of functional consequences, the phosphorylation of the PDC in obese healthy subjects that a decrease in BMI is followed by an CIDEA-transfected 3T3-L1 cells was examined. A pull-down assay increase in CIDEA mRNA expression (25–27). In the present study, using anti-PDH antibodies coupled to Sepharose beads was done reported weight loss was strongly associated with CIDEA using protein lysates from CIDEA or control vector–transfected expression in white adipose tissue. Whether this is due to the cells. Heart mitochondrial lysate was used as a positive control. loss of body weight per se or secondary to the catabolic state Anti-phosphoserine antibodies were used for the detection of cannot be elucidated because prospective investigations before immunoprecipitated proteins. We observed a 2-fold stronger band, the development of cancer and associated cachexia are impossible representing the phosphorylated PDC-E1 component in lysates to perform. However, studies of obese subjects before and after from CIDEA-transfected cells when compared with control cells weight loss suggest that catabolism is more important than weight (Fig. 3). The mitochondrial extracts (positive control) gave the loss for CIDEA up-regulation (26). strongest band of phosphorylated PDC-E1, and this may be RQ was decreased in cancer cachexia, and in the whole material explained by a higher content of mitochondrial proteins in these we also observed a strong negative relationship between RQ and www.aacrjournals.org 9251 Cancer Res 2008; 68: (22). November 15, 2008

CIDEA mRNA expression, which was independent of BMI or fat are lost. This could indicate insufficient energy supply and result mass. Patients with a high ability to oxidize lipids, including FAs, in protective mechanisms being switched on. Further proof had high CIDEA expression and vice versa. Furthermore, there was about the involvement of CIDEA in the regulation of alternative a strong negative association between weight loss and RQ. Because substrate oxidation in human white adipose tissue would require CIDEA is absent or expressed at very low levels in skeletal muscle direct measurements in vivo, which, unfortunately, are not possible (26), these data suggest that CIDEA in white adipose tissue may to perform due to methodologic reasons. participate in the oxidation of FAs in weight-losing subjects. UCP-1 is essential for FA oxidation in brown fat cells and can Moreover, because FA mobilization is elevated in cancer cachexia be activated in human white fat cells (34). As opposed to CIDEA, due to increased adipocyte lipolysis (6), it could be speculated that we found no correlation between UCP-1 and cachexia or fat CIDEA may regulate oxidation of the excess FAs. As mentioned oxidation. In addition, mRNA for CIDEA and UCP-1 did not earlier, FA oxidation in white adipose tissue can be increased by correlate, which is in contrast to earlier findings in healthy changes in the regulation of oxidative pathways. For example, subjects (26). A possible explanation for these divergent findings is overexpression of the peroxisome proliferator–activated receptor-g that we have compared a cohort of patients with and a cohort coactivator 1a markedly increased the capacity of human white fat without cancer cachexia. cells to oxidize FAs (34). We provide evidence that CIDEA can up- Activation of AMPK in adipocytes up-regulates FA oxidation and regulate FA oxidation in white fat cells via an alternative inhibits catecholamine-induced lipolysis (37). A recent study mechanism (i.e., inhibition of the PDH complex). Our findings in showed CIDEA-mediated degradation of AMPK in murine brown 3T3-L1 fat cells show that CIDEA does not regulate genes of the adipocytes (38). High AMPK expression in brown adipocytes of h-oxidation cascade, but rather attenuates the activity of the PDH CIDEA-null mice results in increased FA oxidation. It was proposed complex. CIDEA overexpression up-regulates mRNA levels of that high activity of AMPK is a mechanism causing resistance to À À PDKs, which results in increased phosphorylation of PDH-E1. diet-induced obesity in CIDEA / mice. In contrast, we did not These data are supported by clinical findings showing a strong, observe altered levels of AMPK mRNA following CIDEA over- positive, and selective correlation between CIDEA and PDK1 and expression in white adipocytes (P = 0.17;graph not shown). PDK4 mRNAs in human adipocytes. Taken together, our data Moreover, in the clinical material, CIDEA expression did not suggest that CIDEA may function as a regulatory protein, which correlate with AMPK mRNA (Fig. 4D). Cancer cachexia patients facilitates the preferential oxidation of FAs instead of glucose in display increased catecholamine-induced lipolysis (6–10), suggest- white adipocytes. Under conditions of high FA mobilization such as ing that elevated activity of AMPK in these patients is unlikely. cancer cachexia, CIDEA stimulates oxidation of excess FAs so that Although we cannot exclude AMPK involvement in CIDEA- glucose recycling could be optimized in these cells. It is well known mediated FA oxidation at the posttranscriptional level, our present that the activity of PDC is inhibited in muscle tissue during and previous data do not support this mechanism in human starvation and other catabolic states such as diabetes (36);thus, if adipocytes. glucose supply is scarce or FA supply and oxidation is sufficient to The findings on CIDEA and FA oxidation in white adipose tissue meet cellular energy requirements, mammalian PDC activity is are different from those observed in brown adipose tissue of mice, suppressed. In cancer cachexia, large quantities of adipose tissue where CIDEA inhibits FA oxidation (23). It was previously reported

Figure 4. Relationship between CIDEA mRNA expression in white adipose tissue and mRNA expression of PDK1 (A),PDK4 (B),UCP-1 ( C),and AMPK ( D). See legend to Fig. 1 for further details.

Cancer Res 2008; 68: (22). November 15, 2008 9252 www.aacrjournals.org

CIDEA is a multifunctional protein, although the exact intracellular mechanism of its function in human white adipocytes remains to be defined. Because CIDEA inhibits basal (i.e., nonhormonal) lipolysis (25, 39), we made a retrospective analysis of basal adipocyte lipolysis (i.e., glycerol release) from our previous study (6) comparing cancer cachexia patients with control patients. The values were 3 F 2 and 6 F 4 Amol of glycerol/2 h/107 fat cells, respectively (P = 0.039, Mann-Whitney test). Thus, in addition to its effect on FA oxidation, the elevated CIDEA expression in cancer cachexia may also inhibit basal FA release in fat cells due to the antilipolytic effect. However, under hormone-stimulated conditions (natriuretic peptides, catecholamines), the effect of CIDEA seems to be masked by the activity of hormone-sensitive lipase. Finally, the tumor stage score was higher in weight-stable than in weight-losing cancer patients. This excludes the possibility that increased FA oxidation in cancer cachexia is secondary to a more severe tumor stage. It should be pointed out that we have only investigated gastrointestinal cancer. Therefore, we do not know if FA oxidation is altered in other forms of tumor-induced cachexia. On the basis of previous (6) and current findings, we propose the Figure 5. Model for altered lipid metabolism in white fat cells of cancer cachexia following model of disturbed white adipose tissue lipid mobiliza- patients. Unknown factor(s) directly or indirectly mediated by the cancer tumor enhances the expression of hormone-sensitive lipase (HSL) and CIDEA. tion in cancer cachexia, which could, at least in part, explain the Hormone-sensitive lipase increases the rate of lipolysis stimulated by loss of fat mass in this condition (Fig. 5). The expression of the rate- catecholamines and natriuretic peptides so that more FAs are produced via limiting enzyme hormone-sensitive lipase is elevated in fat cells, hydrolysis of triglycerides (red arrows). Increased expression of CIDEA may up-regulate the expression of PDK1 and PDK4,which in turn phosphorylate the which enhances the lipolytic effect of stimulatory hormones and PDC. Phosphorylation results in the inactivation of the PDC,thereby inducing therefore increases the availability of FAs. Perhaps as a protective the fat cell to use FAs instead of glucose for acetyl-CoA oxidation via the citric mechanism, the PDC of white fat cells is inactivated by CIDEA, acid cycle. CIDEA has previously been shown to inhibit (green dotted line) basal (green arrow),but not hormone-stimulated ( red arrow),lipolysis. The favoring FA oxidation and decreasing the oxidization of glucose to importance of this inhibitory action remains to be determined. acetyl-CoA. that CIDEA is localized to the mitochondria of brown fat cells (23). Disclosure of Potential Conflicts of Interest Recent findings have identified CIDEA as a lipid droplet–coating No potential conflicts of interest were disclosed. protein and mitochondrial localization was not confirmed (39). Instead, some cytosolic and nuclear localization was observed. Acknowledgments In our laboratory, we have also detected CIDEA in the cytoplasm and the nucleus.6 Localization to the surface of lipid droplets does Received 4/15/2008;revised 8/13/2008;accepted 9/18/2008. Grant support: Swedish Research Council, AFA, and the European Union not explain how CIDEA regulates FA oxidation. We believe that (HEPADIP;LSHM-CT-2005-018734). 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 accordance with 18 U.S.C. Section 1734 solely to indicate this fact. We thank Britt-Marie Leijonhufvud, Katarina Hertel, Eva Sjo¨lin, Kerstin Wa˚hle´n, 6 Unpublished observations. and Gaby A˚stro¨m for their excellent technical assistance.

References 7. Hyltander A, Daneryd P, Sandstrom R, Korner U, 13. Wilson-Fritch L, Nicoloro S, Chouinard M, et al. Lundholm K. h-Adrenoceptor activity and resting Mitochondrial remodeling in adipose tissue associated 1. Esper DH, Harb WA. The cancer cachexia syndrome: a energy metabolism in weight losing cancer patients. with obesity and treatment with rosiglitazone. J Clin review of metabolic and clinical manifestations. Nutr Eur J Cancer 2000;36:330–4. Invest 2004;114:1281–9. Clin Pract 2005;20:369–76. 8. Klein S, Wolfe RR. Whole-body lipolysis and triglycer- 14. Frayn KN, Karpe F, Fielding BA, Macdonald IA, 2. Tijerina AJ. The biochemical basis of metabolism in ide-fatty acid cycling in cachectic patients with Coppack SW. Integrative physiology of human cancer cachexia. Dimens Crit Care Nurs 2004;23:237–43. esophageal cancer. J Clin Invest 1990;86:1403–8. adipose tissue. Int J Obes Relat Metab Disord 2003; 3. FouladiunM,KornerU,BosaeusI,DanerydP, 9. Legaspi A, Jeevanandam M, Starnes HF, Jr., Brennan 27:875–88. Hyltander A, Lundholm KG. Body composition and MF. Whole body lipid and energy metabolism in the 15. Wang T, Zang Y, Ling W, Corkey BE, Guo W. time course changes in regional distribution of fat and cancer patient. Metabolism 1987;36:958–63. Metabolic partitioning of endogenous fatty acid in lean tissue in unselected cancer patients on palliative 10. Shaw JH, Wolfe RR. Fatty acid and glycerol kinetics in adipocytes. Obes Res 2003;11:880–7. care-correlations with food intake, metabolism, exercise septic patients and in patients with gastrointestinal 16. Klaus S. Adipose tissue as a regulator of energy capacity, and hormones. Cancer 2005;103:2189–98. cancer. The response to glucose infusion and parenteral balance. Curr Drug Targets 2004;5:241–50. 4. Kalra PR, Tigas S. Regulation of lipolysis: natriuretic feeding. Ann Surg 1987;205:368–76. 17. Sorbris R, Monti M, Nilsson-Ehle P, Wadso I. Heat peptides and the development of cachexia. Int J Cardiol 11. Arner P. Human fat cell lipolysis: biochemistry, production by adipocytes from obese subjects before 2002;85:125–32. regulation and clinical role. Best Pract Res Clin and after weight reduction. Metabolism 1982;31:973–8. 5. Tisdale MJ. Molecular pathways leading to cancer Endocrinol Metab 2005;19:471–82. 18. Holness MJ, Sugden MC. Regulation of pyruvate cachexia. Physiology (Bethesda) 2005;20:340–8. 12. Korber J, Pricelius S, Heidrich M, Muller MJ. dehydrogenase complex activity by reversible phosphor- 6. Agustsson T, Ryden M, Hoffstedt J, et al. Mechanism of Increased lipid utilization in weight losing and weight ylation. Biochem Soc Trans 2003;31:1143–51. increased lipolysis in cancer cachexia. Cancer Res 2007; stable cancer patients with normal body weight. Eur J 19. Patel MS, Korotchkina LG. Regulation of mammalian 67:5531–7. Clin Nutr 1999;53:740–5. pyruvate dehydrogenase complex by phosphorylation: www.aacrjournals.org 9253 Cancer Res 2008; 68: (22). November 15, 2008

complexity of multiple phosphorylation sites and population-based and dietary intervention studies. 33. Ryde´nM, Dicker A, van Harmelen V, et al. Mapping of kinases. Exp Mol Med 2001;33:191–7. J Clin Endocrinol Metab 2007;92:4759–65. early signaling events in tumor necrosis factor-a- 20. Korotchkina LG, Patel MS. Probing the mechanism of 27. Dahlman I, Linder K, Arvidsson Nordstrom E, et al. mediated lipolysis in human fat cells. J Biol Chem inactivation of human pyruvate dehydrogenase by Changes in adipose tissue gene expression with energy- 2002;277:1085–91. phosphorylation of three sites. J Biol Chem 2001;276: restricted diets in obese women. Am J Clin Nutr 2005;81: 34. Tiraby C, Tavernier G, Lefort C, et al. Acquirement of 5731–8. 1275–85. brown fat cell features by human white adipocytes. 21. Kersten S, Desvergne B, Wahli W. Roles of PPARs in 28. Dahlman I, Kaaman M, Jiao H, Kere J, Laakso M, J Biol Chem 2003;278:33370–6. health and disease. Nature 2000;405:421–4. Arner P. The CIDEA gene V115F polymorphism is 35. Fawcett J, Tsui BT, Kruer MC, Duckworth WC. 22. Inohara N, Koseki T, Chen S, Wu X, Nunez G. CIDE, a associated with obesity in Swedish subjects. Diabetes Reduced action of insulin glargine on protein and lipid novel family of cell death activators with homology to 2005;54:3032–4. metabolism: possible relationship to cellular hormone the 45 kDa subunit of the DNA fragmentation factor. 29. Ottery FD. Definition of standardized nutritional metabolism. Metabolism 2004;53:1037–44. EMBO J 1998;17:2526–33. assessment and interventional pathways in oncology. 36. Wu P, Inskeep K, Bowker-Kinley MM, Popov KM, 23. Zhou Z, Yon Toh S, Chen Z, et al. Cidea-deficient Nutrition 1996;12:S15–9. Harris RA. Mechanism responsible for inactivation of mice have lean phenotype and are resistant to obesity. 30. Kolaczynski JW, Morales LM, Moore JH, Jr., et al. A skeletal muscle pyruvate dehydrogenase complex in Nat Genet 2003;35:49–56. new technique for biopsy of human abdominal fat under starvation and diabetes. Diabetes 1999;48:1593–9. 24. Lin SC, Li P. CIDE-A, a novel link between brown local anaesthesia with Lidocaine. Int J Obes Relat Metab 37. Daval M, Foufelle F, Ferre´ P. Functions of AMP- adipose tissue and obesity. Trends Mol Med 2004;10: Disord 1994;18:161–6. activated protein kinase in adipose tissue. J Physiol 434–9. 31. Ryden M, Elizalde M, van Harmelen V, et al. Increased 2006;574:55–62. 25. Nordstrom EA, Ryden M, Backlund EC, et al. A human- expression of eNOS protein in omental versus subcuta- 38. Qi J, Gong J, Zhao T, et al. Downregulation of AMP- specific role of cell death-inducing DFFA (DNA fragmen- neous adipose tissue in obese human subjects. Int J activated protein kinase by Cidea-mediated ubiquitina- tation factor-a)-like effector A (CIDEA) in adipocyte Obes Relat Metab Disord 2001;25:811–5. tion and degradation in brown adipose tissue. EMBO J lipolysis and obesity. Diabetes 2005;54:1726–34. 32. Laurencikiene J, van Harmelen V, Arvidsson Nord- 2008;27:1537–48. 26. Gummesson A, Jernas M, Svensson PA, et al. strom E, et al. NF-nB is important for TNF-a-induced 39. Puri V, Ranjit S, Konda S, et al. Cidea is associated Relations of adipose tissue CIDEA gene expression to lipolysis in human adipocytes. J Lipid Res 2007;48: with lipid droplets and insulin sensitivity in humans. basal metabolic rate, energy restriction, and obesity: 1069–77. Proc Natl Acad Sci U S A 2008;105:7833–8.

Cancer Res 2008; 68: (22). November 15, 2008 9254 www.aacrjournals.org

Jurga Laurencikiene, Britta M. Stenson, Elisabet Arvidsson Nordström, et al.

Cancer Res 2008;68:9247-9254.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/68/22/9247

Cited articles This article cites 39 articles, 11 of which you can access for free at: http://cancerres.aacrjournals.org/content/68/22/9247.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/68/22/9247.full#related-urls

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

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

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/68/22/9247. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.