+ All Categories
Home > Documents > Globular adiponectin and its downstream target genes are up-regulated locally in human colorectal...

Globular adiponectin and its downstream target genes are up-regulated locally in human colorectal...

Date post: 30-Dec-2016
Category:
Upload: ida
View: 222 times
Download: 0 times
Share this document with a friend
10
Globular adiponectin and its downstream target genes are up-regulated locally in human colorectal tumors: ex vivo and in vitro studies Katja Kannisto Vetvik a, b, , Tonje Sonerud a , Mona Lindeberg d, f , Torben Lüders a , Ragnhild H. Størkson a, e , Kristin Jonsdottir a , Eirik Frengen d, f , Kirsi H. Pietiläinen g, h , Ida Bukholm a, b, c a Department of Clinical Molecular Biology and Laboratory Sciences (EpiGen), Institute for Clinical Medicine, University of Oslo, Norway b Surgical Department, Akershus University Hospital, Lørenskog, Norway c Institute of Health Promotion, Akershus University Hospital, Lørenskog, Norway d Department of Medical Genetics, Oslo University Hospital, Oslo, Norway e Surgical Department, Østfold Hospital Trust, Fredrikstad, Norway f Department of Medical Genetics, University of Oslo, Norway g Obesity Research Unit, Department of Medicine, Division of Endocrinology, Helsinki University Central Hospital and University of Helsinki, Finland h Institute of Molecular Medicine Finland, FIMM, University of Helsinki, Helsinki, Finland ARTICLE INFO ABSTRACT Article history: Received 18 April 2013 Accepted 3 February 2014 Objective. Low plasma adiponectin levels are linked to obesity, insulin resistance, and the risk of several types of malignancy. Despite the decline in circulating adiponectin concentrations, the increase in the expression of adiponectin receptors AdipoR1 and AdipoR2 is greater in cancerous than in normal colonic tissue. The purpose of this study was to obtain new information regarding local adiponectin signaling in the pathogenesis of colorectal cancer (CRC). Methods. We characterized the expressions of adiponectin and several of its downstream targets in paired samples of tumor tissue and adjacent noncancerous mucosa in 60 surgical patients with colorectal adenocarcinomas. Keywords: Full-length adiponectin Colon cancer Insulin resistance AMPK AdipoR METABOLISM CLINICAL AND EXPERIMENTAL 63 (2014) 672 681 Abbreviations: AGRP, agouti-related peptide; ADIPOQ, collagen domain containing adiponectin; AdipoR1, adiponectin receptor type 1; AdipoR2, adiponectin receptor type 2; ACC, acetyl-CoA carboxylase; AMPK, 5AMP-activated protein kinase; CAB39, Calcium binding protein 39; CPT, carnitine palmitoyltransferase; fAd, full-length adiponectin; FABP, fatty-acid-binding protein; FATP, fatty acid transporter; FDR, false discovery rate; gAd, globular adiponectin; GLUT4, glucose transporter type 4; G6PC, glucose-6-phosphatase; HMW, high-molecular-weight polymer; JAK, janus kinase; LEP, leptin; LEPR, leptin receptor; LYK5, Protein kinase LYK5; MAPK, mitogen- activated protein kinase; MSH, α-melanocyte stimulating hormone; mTOR, mammalian target of rapamycin; NASH, nonalcoholic steatohepatitis; NPY, neuropeptide Y; pAMPK, phosphorylated AMPK; PGC, PPAR γ coactivator; PEPCK, phosphoenolpyruvate; POMC/ CART, proopiomelanocortin/cocaine- and amphetamine-regulated transcript; PRKAB1, protein kinase, AMP-activated, beta 1 noncatalytic subunit; qPCR, quantitative real-time PCR; RXR, retinoid X receptor; SAM, significant analysis of microarray; SHP2, tyrosine-protein phosphatase nonreceptor type 1; siRNA, silencing RNA; SLC2A4, solute carrier family 2 member 4; SLC27A1, long-chain fatty acid transport protein 1; STAT, signal transducers and activators of transcription; STK11, serine/threonine kinase 11; TPB, TATA-binding protein; TFIIB, transcription factor IIB. MICROARRAY DATA: The microarray data have been deposited in ArrayExpress (http://www.ebi.ac.uk/arrayexpress/) under accession no. E-MTAB-833. Corresponding author at: Akershus University Hospital, 1478 Lørenskog, Norway. Tel.: + 47 95796638. E-mail address: [email protected] (K.K. Vetvik). 0026-0495/$ see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.metabol.2014.02.001 Available online at www.sciencedirect.com Metabolism www.metabolismjournal.com
Transcript
Page 1: Globular adiponectin and its downstream target genes are up-regulated locally in human colorectal tumors: ex vivo and in vitro studies

M E T A B O L I S M C L I N I C A L A N D E X P E R I M E N T A L 6 3 ( 2 0 1 4 ) 6 7 2 – 6 8 1

Ava i l ab l e on l i ne a t www.sc i enced i r ec t . com

Metabolismwww.metabo l i sm jou rna l . com

Globular adiponectin and its downstream target genes areup-regulated locally in human colorectal tumors:ex vivo and in vitro studies☆

Katja Kannisto Vetvika, b,⁎, Tonje Soneruda, Mona Lindebergd, f, Torben Lüdersa,Ragnhild H. Størksona, e, Kristin Jonsdottir a, Eirik Frengend, f,Kirsi H. Pietiläineng, h, Ida Bukholma, b, c

a Department of Clinical Molecular Biology and Laboratory Sciences (EpiGen), Institute for Clinical Medicine, University of Oslo, Norwayb Surgical Department, Akershus University Hospital, Lørenskog, Norwayc Institute of Health Promotion, Akershus University Hospital, Lørenskog, Norwayd Department of Medical Genetics, Oslo University Hospital, Oslo, Norwaye Surgical Department, Østfold Hospital Trust, Fredrikstad, Norwayf Department of Medical Genetics, University of Oslo, Norwayg Obesity Research Unit, Department of Medicine, Division of Endocrinology, Helsinki University Central Hospital andUniversity of Helsinki, Finlandh Institute of Molecular Medicine Finland, FIMM, University of Helsinki, Helsinki, Finland

A R T I C L E I N F O

Abbreviations: AGRP, agouti-related peptidAdipoR2, adiponectin receptor type 2; ACC,protein 39; CPT, carnitine palmitoyltransftransporter; FDR, false discovery rate; gAd,HMW, high-molecular-weight polymer; JAK,activated protein kinase; MSH, α-melanocysteatohepatitis; NPY, neuropeptide Y; pAMPCART, proopiomelanocortin/cocaine- and amsubunit; qPCR, quantitative real-time PCR;phosphatase nonreceptor type 1; siRNA, sitransport protein 1; STAT, signal transduceprotein; TFIIB, transcription factor IIB.☆ MICROARRAY DATA: Themicroarray datano. E-MTAB-833.⁎ Corresponding author at: Akershus Universi

E-mail address: [email protected] (K.K. Vet

0026-0495/$ – see front matter © 2014 Elsevihttp://dx.doi.org/10.1016/j.metabol.2014.02.00

A B S T R A C T

Article history:Received 18 April 2013Accepted 3 February 2014

Objective. Low plasma adiponectin levels are linked to obesity, insulin resistance, and therisk of several types of malignancy. Despite the decline in circulating adiponectinconcentrations, the increase in the expression of adiponectin receptors AdipoR1 andAdipoR2 is greater in cancerous than in normal colonic tissue. The purpose of this study wasto obtain new information regarding local adiponectin signaling in the pathogenesis ofcolorectal cancer (CRC).

Methods.We characterized the expressions of adiponectin and several of its downstreamtargets in paired samples of tumor tissue and adjacent noncancerous mucosa in 60 surgicalpatients with colorectal adenocarcinomas.

Keywords:Full-length adiponectinColon cancerInsulin resistanceAMPKAdipoR

e; ADIPOQ, collagen domain containing adiponectin; AdipoR1, adiponectin receptor type 1;acetyl-CoA carboxylase; AMPK, 5′ AMP-activated protein kinase; CAB39, Calcium bindingerase; fAd, full-length adiponectin; FABP, fatty-acid-binding protein; FATP, fatty acidglobular adiponectin; GLUT4, glucose transporter type 4; G6PC, glucose-6-phosphatase;janus kinase; LEP, leptin; LEPR, leptin receptor; LYK5, Protein kinase LYK5; MAPK, mitogen-te stimulating hormone; mTOR, mammalian target of rapamycin; NASH, nonalcoholicK, phosphorylated AMPK; PGC, PPAR γ coactivator; PEPCK, phosphoenolpyruvate; POMC/phetamine-regulated transcript; PRKAB1, protein kinase, AMP-activated, beta 1 noncatalyticRXR, retinoid X receptor; SAM, significant analysis of microarray; SHP2, tyrosine-proteinlencing RNA; SLC2A4, solute carrier family 2 member 4; SLC27A1, long-chain fatty acidrs and activators of transcription; STK11, serine/threonine kinase 11; TPB, TATA-binding

have been deposited in ArrayExpress (http://www.ebi.ac.uk/arrayexpress/) under accession

ty Hospital, 1478 Lørenskog, Norway. Tel.: +47 95796638.vik).

er Inc. All rights reserved.1

Page 2: Globular adiponectin and its downstream target genes are up-regulated locally in human colorectal tumors: ex vivo and in vitro studies

673M E T A B O L I S M C L I N I C A L A N D E X P E R I M E N T A L 6 3 ( 2 0 1 4 ) 6 7 2 – 6 8 1

Results. Adiponectin was expressed in both colorectal tumors and the adjacent mucosa.The expressions of adiponectin mRNA and its globular protein variant (gAd), adiponectinreceptor type 1 and 5′AMP-activated protein kinase (AMPK)mRNAwere significantly higherin colorectal tumors than in the adjacent mucosa. This finding was accompanied byincreased mRNA expression of genes encoding proteins involved in fatty-acid traffickingand oxidation. The potential interference between adiponectin stimulation and AMPKactivation through AMPK1 was examined in an in vitro model with the aid of silencing-RNAexperiments. Furthermore, AMPK mRNA expression on tumors was positively correlatedwith a more advanced tumor stage in the patients.

Conclusion. We propose that the globular adiponectin-AMPK pathway functions in anautocrinemanner in colorectal tumors, explaining some of the beneficial changes in cellularoxidative capacity in tumors in favor of tumorigenesis.

© 2014 Elsevier Inc. All rights reserved.

1. Background

Adipose tissue is involved in the metabolic and inflammatoryresponses of the body by secreting adipokines, which aresystemic mediators [1]. These cytokines play multiple roles incellular processes such as cell proliferation, differentiation,and apoptosis [2]. Adiponectin is a hormone secreted byadipocytes that has diverse biological functions, includingstimulation of glucose utilization, fatty-acid oxidation, andinhibition of gluconeogenesis [3]. Low levels of circulatingadiponectin are associated with obesity and insulin resis-tance, as well as increased risk for the development andprogression of several obesity-related malignancies, such asbreast, endometrial, prostate, and colon cancers [4].

The plasma levels of full-length adiponectin (fAd) (alsoknown as 30-kDa adipocyte complement-related protein)exceed 10 μg/ml and constitute ≈0.01% of the total amountof plasma protein [3,5]. Proteolytic cleavage of fAd at aminoacid 110 produces globular adiponectin (gAd) [6]. fAd candimerize and most of the adiponectin present in plasma isoligomerized to form a high-molecular-weight polymer(HMW) [7,8], whereas the abundance of the low-molecular-weight form, or gAd, is normally low [9,10]. It has beendemonstrated that oligomerization and posttranslationalmodifications of adiponectin are critical for binding to itsreceptors and to determine its biological activity. Thus,different oligomeric forms of adiponectin, or gAd versusfAd, exhibit distinct biological effects through the differen-tial activation of downstream signaling pathways [5] whichmay explain their eventual different roles in tumordevelopment [11].

Two adiponectin receptors have been described: AdipoR1and AdipoR2 [12]. These receptors mediate the signaling ofboth fAd and gAd. AdipoR1 has been shown to bind gAd withhigher affinity than fAd, whereas AdipoR2 seems to be anintermediate-affinity receptor for both forms [7,12]. AdipoR1 isubiquitously expressed, but is present most abundantly inskeletal muscle, whereas AdipoR2 is predominantlyexpressed in the liver [12]. Activation of the two receptorvariants reveals clearly opposing metabolic functions in vivo[13], as observed in studies on knockout animal models.AdipoR1 signaling occurs mainly through 5′ AMP-activatedprotein kinase (AMPK), involving increased energy consump-

tion in the form of increased glucose and fatty-acid oxidationand energy expenditure in vivo, whereas AdipoR2 acts throughperoxisome proliferator-activated receptor α-related path-ways and seems to be involved in increased storage of energy,glucose intolerance, and reduced energy expenditure. Theexpressions of both types of adiponectin receptor have beenidentified in several cancer types, whereas AdipoR1 wasrecently shown to be more ubiquitously expressed in obesity-associated cancers [14,15]. Previous studies have demonstratedincreased adiponectin receptor expression in colon cancer butnot in prostate cancer [16,17]. Investigation of the role of thesetwo receptors and the pathways that are involved wouldprovide additional information regarding the role of adiponec-tin signaling during tumor development.

It has been hypothesized that the generally beneficialmetabolic and antioxidative actions of adiponectin can beexplained in part through its activation of the AMPK pathway[18]. There is mounting evidence for both this observation andthe finding that activation of AMPK constitutes a generalintracellular response to adiponectin mainly by activationthrough AdipoR1: it has been shown that AMPK is activated inresponse to adiponectin treatment in adipocytes [19], myo-cytes and hepatocytes [20], pancreatic β-cells [21], cardiomyo-cytes [22], and endothelial cells [23], as well as in colon (HT-29)and prostate (PC-3) cancer cells [18]. AMPK activation sup-presses cell proliferation through multiple mechanisms,including suppression of the mammalian target of rapamycin(mTOR) pathway and inhibition of the enzymes implicated inthe regulation of protein and fatty-acid and triacylglycerolsynthesis [acetyl-CoA carboxylase (ACC2) and fatty-acidsynthase]. In addition, adiponectin, through activation ofAMPK, negatively regulates the downstream members in themTOR pathway, such as the p70 S6 kinase and S6 protein[20,24,25]. It is speculated that activation of the mTORpathway directly promotes colonic epithelial cell prolifera-tion, and thereby colorectal carcinogenesis [26].

The role of adiponectin in malignancies may be explainedboth by its effect at the systemic level, whereby it regulateswhole-body insulin sensitivity or inflammatory processes,and at the cellular level, whereby it acts directly on cancercells. In line with epidemiological data, there is accumulatingevidence from several previous in vitro studies that adiponec-tin is an important regulator of cell proliferation [27]. A

Page 3: Globular adiponectin and its downstream target genes are up-regulated locally in human colorectal tumors: ex vivo and in vitro studies

674 M E T A B O L I S M C L I N I C A L A N D E X P E R I M E N T A L 6 3 ( 2 0 1 4 ) 6 7 2 – 6 8 1

previous study found that both specific inhibitors of AMPKand a specific silencing RNA (siRNA) to AdipoR1 blocked theantiproliferative effects of adiponectin [28]. Plasma adiponec-tin administration suppressed the growth of intestinaladenomas in the Apc(Min)(/+) mice [29], and reducedimplanted tumor growth and angiogenesis in tumors obtainedfrom mice fed a high-fat diet and from adiponectin-deficientC57BL/6 mice [30]. Consistent results were obtained whenapplying a high-fat diet to adiponectin-receptor-knockoutmice [26]. Despite an increasing amount of experimental data,the mechanisms underlying the antiproliferative and tumor-suppressing effects of adiponectin are not fully understood,and some of the results show opposite actions. For example,Ogunwobi and Beales demonstrated that adiponectin stimu-lates the proliferation of colon cancer HT-29 cells [31]. Thediscrepancies between the aforementioned and other previ-ous results could be partly explained by the findings of Habeebet al., which suggest that the bimodal action of adiponectindepends upon the glucose level in the culture medium, suchthat adiponectin supports cell survival in a glucose-deprivedmedium but not in a glucose-enriched medium [32]. However,most of the previous studies focused on the role of fAd ratherthan gAd stimulation in cell proliferation.

For the present study it was proposed that autocrineactivation of the genes in the gAd-AdipoR1-AMPK pathway bytumor cells could be one of the mechanisms underlying theimprovement in tumor-cell metabolism, providing a linkbetween colorectal tumor disease, obesity and the body’smetabolic state. This study is the first to examine adiponectinsecretion in tumors in vivo, and the first to reveal increasedgAd expression in tumors.

2. Methods

Details regarding the materials and methods are provided inthe Supplementary Materials and Methods section. Thematerials and methods are summarized below.

2.1. Study subjects

The study comprised samples of tumor and the adjacentnormal mucosa, which were removed surgically from 60patients with colorectal adenocarcinoma at the AkershusUniversity Hospital, Norway. All patients had read, under-stood, and signed an informed consent document before theircolorectal surgery and the collection of either samples or anypatient data. Patients eventually using any drugs that maypossess anticancer effects, such as nonsteroidal anti-inflam-matory drugs and metformin, were not excluded. Samplesfrom colorectal tumors and adjacent noncancerous mucosawere collected from the patients during open colorectalsurgery. The tumor tissue was collected from solid tumorsthat were macroscopically visible, and subsequently investi-gated microscopically and characterized histologically. Onlypatients with malignant tumors were included. Radical tumorresection was ensured by employing a resection margin ofseveral centimeters. Adjacent mucosal tissue was collectedwith a margin of several centimeters from the solid tumor,

from an area that was confirmed to be tumor free throughboth macroscopic and microscopic investigation. The prepa-ration of tissues is reported elsewhere [33]. The study wasperformed in accordance with the principles of the HelsinkiDeclaration and was approved by the local ethics committee(REK Sør-Øst, Blindern, Oslo) before any study-related activ-ities took place.

2.2. RNA isolation and quantitative real-time PCR (qPCR)

Total RNA was isolated using a Trizol/RNeasy hybrid protocol[34]. RNA was reverse transcribed to cDNA using the Super-Script First Strand Synthesis System (Invitrogen, Carlsbad, CA,USA). TaqMan gene-expression assays were purchased fromApplied Biosystems (Bedford, MA, USA) and the qPCR wasperformed with the aid of the ABI PRISM 7900 HT Fast qPCRsystem (Applied Biosystems), as reported previously [33]. Theprimers are listed in the Supplementary Materials andMethods. General transcription factor IIB [TATA-bindingprotein (TPB)/transcription factor IIB (TFIIB)] expression wasused as a reference gene (TPB/TFIIB) for normalization. All datawere calculated relative to TPB, and the ratios between theexpression in the tumor samples and normal colon samplesare presented.

2.3. Oligo microarray analysis

Oligo microarray analyses were available for 23 of the 60sample pairs. Agilent’s Human Whole Genome Oligo Micro-arrays (44 K), which contain ~41,000 probes of genes andtranscripts, were used. Data collection and quality assessmentwere performed using Agilent G2567AA Feature Extractionsoftware v8.5 (with the default parameters). Data analysis wasperformed using J-express Pro v2.7 (Molmine, Bergen, Nor-way). The LSimpute function calculated missing values forgenes with less than 30% missing values. A two-classSignificant Analysis of Microarray (SAM) [35] was used toidentify genes that were significantly differentially expressedin the tumor samples compared to the samples from adjacentnormal mucosa, implemented in J-express with 1,000 permu-tations and a false discovery rate (FDR) of <10%. Pathwayanalyses were carried out in the Database for Annotation,Visualization, and Integrated Discovery (http://david.abcc.ncifcrf.gov/), which applies the gene list found in the KyotoEncyclopedia of Genes and Genomes pathway. The microar-ray data have been deposited in ArrayExpress (http://www.ebi.ac.uk/arrayexpress/) under accession no. E-MTAB-833.

2.4. Cell culture and transient transfection of siRNA forthe knockdown of AdipoR1

The colorectal adenocarcinoma cell line WiDr (ATCC NumberCCL-218) was purchased from LGC Promochem (Boras, Sweden).ON-TARGETplus siRNA AdipoR1 (Catalog nr J-007800-10) waspurchased from Thermo Fisher Scientific (Lafayette, CO, USA).Four different siRNAs for AdipoR1, as well as a mix of all four,were tested on WiDr colon cancer cells, and the best siRNA wasselected. Transient transfections were carried out with Dharma-fect4 transfection reagent (Thermo Fisher Scientific). Since

Page 4: Globular adiponectin and its downstream target genes are up-regulated locally in human colorectal tumors: ex vivo and in vitro studies

675M E T A B O L I S M C L I N I C A L A N D E X P E R I M E N T A L 6 3 ( 2 0 1 4 ) 6 7 2 – 6 8 1

AdipoR1 activation is dependent upon the presence of its ligand[5], gAd was added to the cell culture in order to reveal thedownstream effect of the AdipoR1 knockdown on AMPKphosphorylation. The transfected cells were incubated with orwithout 2.5 μg/ml gAd (AXXORA, San Diego, CA, USA) for 15 min,30 min, or 24 h before harvesting them.

2.5. Protein extraction and Western blot analysis

Detailed procedures for protein extraction from the colorectaltissue and Western blot analysis, as well as a list of theantibodies used, are provided in the Supplementary Materialsand Methods.

2.6. Statistical analyses

The mRNA expression of each gene was compared betweencolorectal tumor and normal tissue samples using pairedWilcoxon’s rank-sum tests for nonparametric data. Anunpaired Mann–Whitney U test was used for comparisons ofgene expressions between males and females. Relativechanges (tumor divided by normal tissue) in gene expressionswere used when evaluating the role of clinical pathologicalstages on changes in gene expression. The Mann–Whitney Utest was also used for comparisons between medians in geneexpressions in Dukes’ B and C stages, and Spearman’scorrelation coefficient was calculated between gene expres-sion and categorical or continuous tumor characteristics(localization, size, differentiation, and lymph nodes affected).Spearman’s correlation analyses between selected geneexpressions involved either the relative changes (tumordivided by normal tissue) or tumor values. Statistical analyseswere performed using Stata statistical software (release 9.0,Stata Corporation, College Station, TX, USA).

3. Results

3.1. Clinical characteristics

This study analyzed paired colorectal tissues (tumor andadjacent normal mucosa) from 60 patients for pathwaysinvolved in the regulation of tumor metabolism. The ratioof males/females was 40/20, and the age at surgery of theentire study population was 61.8 ± 1.5 years (mean ± SEM).The distribution of tumor location was as follows: 34% in therectum, 25% in the sigmoid colon, and the remainder in themore proximal portions of colon. The tumors were 5.0 ±0.3 cm in diameter. Histologically, most (80%) of the patientshad moderately differentiated tumors. The tumors werestaged as either Dukes’ B (51.7%) or C (48.3%), and the numberof lymph nodes affected ranged from 0 to 22, with half of thepatients having no affected nodes.

3.2. Increased mRNA expression of AMPK-related genes intumor tissue, as revealed by microarray analysis

Microarray expression arrays were first used to measuremRNA transcript profiles from 23 paired samples of colorectal

tumor and adjacent normal tissue from the same patients.The sampleswere randomly selected frommaterial consistingof paired samples. SAM with 1,000 permutations and an FDRof 0% yielded 3,738 significant genes. The regulators of theAMPK pathway were studied in order to identify the micro-matrix mRNA expression of the leptin and adiponectinpathways in tumors, which involve two adipocyte-specifichormones previously shown to activate AMPK (Fig. 1, Table 1)[19,36]. The genes encoding glucose transporter type 4 (GLUT4;FDR = 0.1), carnitine palmitoyltransferase I (CPT1; FDR = 0.0X),and carnitine palmitoyltransferase 2 (CPT2; FDR = 0.0X) weresignificantly elevated in tumors relative to the adjacentnormal colorectal tissue from the same patients. The expres-sion of AMPKwas also increased, but with a lower significance(FDR = 13.1). The expressions of the genes encoding tyrosine-protein phosphatase nonreceptor type 11 (SHP2; FDR = 0.0–0.6)and those encoding long-chain fatty acid transport protein 1(SLC27A1; FDR = 3.0) were significantly lower in the tumorsthan in the normal tissue in the paired samples.

3.3. Increased mRNA expression of AMPK-related genes intumors, as revealed by qPCR

To verify the results from the global gene-array analysis andto further identify the regulators of the AMPK pathway, qPCRanalysis was performed for the following genes, which areknown to be involved in the AMPK pathway: adiponectin,ADIPOR1, ADIPOR2, AMPK, ACC2, plasma membrane fatty-acid-binding protein (FABP-pm), fatty acid transporter (FATP),CPT1, and fatty-acid-binding protein 4 (FABP-4). qPCR analysisconfirmed the results from the gene-array analysis regardingleptin, whose expressionwas the same in normalmucosa andtumor tissue, and hence it was not therefore a candidate foran intrinsic activator of the AMPK pathway in the colorectalmucosa cells. In contrast, adiponectin mRNA expression wasdetected in colorectal epithelium, andwas significantly higherin tumor tissue than in adjacent normal mucosa from thesame patient (2.7-fold, p < 0.0001). Adiponectin exerts itseffects through its receptors AdipoR1 and AdipoR2. AdipoR1(1.6-fold, p = 0.0037) but not AdipoR2 (p = 0.42) mRNA expres-sion was significantly higher in tumor tissue than in thesample from adjacent normal mucosa from the same patient.The ratio of the expressions of AdipoR1 to AdipoR2 was higherin the tumors (1.50 ± 0.12) than in the normal samples (1.19 ±0.08, p = 0.012). It has been demonstrated that activation ofAdipoR1 leads to an increase in AMPK activation in several celltypes [18–23]. In our material, AMPK mRNA expression was1.5-fold (p = 0.034) higher in the tumor tissue than in theadjacent normal mucosa, and was significantly correlatedwith AdipoR1 mRNA expression (r = 0.92, p < 0.0001). Therewas also a significant positive correlation (r = 0.85, p < 0.0001)between AdipoR1 and AMPKmRNA expressions in the normalmucosal tissue, but with much lower expression levels.

A downstream target of AMPK, ACC2 mRNA, was 0.5-folddecreased (p < 0.0001) in the tumor tissue; this phenomenonis known to switch on fatty-acid oxidation [37]. The mRNAexpression of a downstream target of AMPK and ACC,carnitine palmitoyltransferase (CPT)1, was positively corre-lated with the AMPK expression in both tumors (r = 0.90,p < 0.0001) and normal mucosal tissue (r = 0.83, p < 0.0001).

Page 5: Globular adiponectin and its downstream target genes are up-regulated locally in human colorectal tumors: ex vivo and in vitro studies

Fig. 1 – mRNA expressions in gene arrays (n = 23) were used to measure the transcript profiles of the adiponectin pathway intumor and normal tissue samples obtained from the same patients. The genes in the adiponectin-AMPK pathway that showedsignificantly increased or decreased expression in tumors in the global gene array are marked with red stars.

676 M E T A B O L I S M C L I N I C A L A N D E X P E R I M E N T A L 6 3 ( 2 0 1 4 ) 6 7 2 – 6 8 1

Several genes involved in the transport of fatty acids into thecells (i.e., FATP, FABP-pm, and FABP4) were simultaneouslyup-regulated in the tumor relative to the adjacent normaltissue (mRNA increase of 1.8 for FATP and FABP-pm, p =0.0229 and 0.056, respectively; mRNA increase of 1.6 fold forFABP, p < 0.001). Differential expressions of GLUT4, CPT2,SHP2, and SLC27A1 were also detected in the microarraydata. These results were not validated using qPCR.

3.4. Increased gAd versus fAd protein expression intumors, as revealed by Western blot analysis

Western blot analyses were conducted to confirm andfurther study the adiponectin protein secretion in thecolorectal samples (Fig. 2). The expressions of both the15-kDa gAd and 30-kDa fAd were quantified, since theseproteins exhibit distinct biological activities. Fifty pairedtumor and normal tissue samples were available for theseanalyses. The gAd protein was increased by 1.7-fold incolorectal tumor tissue compared pairwise with the adjacentnormal mucosa from the same patient (p = 0.0088, Fig. 3a).fAd expression in tumor tissue was decreased to 60 % of thatin the adjacent normal mucosa (p = 0.0007, Fig. 3b). The ratiobetween gAd and fAd was significantly higher in the tumorsamples (Fig. 3c). This increased relative amount of gAd wasassociated with increased AMPK mRNA expression (r = 0.34,p = 0.018) and tended to be correlated with AdipoR1 mRNAexpression (r = 0.25, p = 0.089) but not with AdipoR2 expres-sion (r = 0.01, p = 0.95). This suggests that activation ofAMPK is can be facilitated by increased sensitivity to gAdin colorectal tumors through increased AdipoR1 expression,but not AdipoR2 expression.

3.5. Correlation between the AMPK activation and gAdlevels, as revealed by RNA interference (RNAi)

The existence of a causal relationship between the expres-sion of gAd and AMPK activation in colon cancer cells wasinvestigated using an in vitro model by down-regulating theexpression of AdipoR1 using RNAi in cell-culture experi-ments. WiDr colon cancer cells were first incubated withgAd for 15 min, which resulted in a more than twofoldincrease in phosphorylation of the α-subunit of AMPK(15-min gAd), an effect that was suppressed by simulta-neous treatment of the cells with siRNA targeting the geneencoding AdipoR1 (siRNA ADIPOR1, 15 min) (Fig. 4). Thedifference in AMPK phosphorylation evident betweenuntreated and siRNA-treated WiDr colon cancer cells in thebaseline situation is thought to depend upon the alteredcircumstances in cells due to siRNA treatment. These resultsindicate that the correlation between increased gAd levelsand increased AMPK activation detected in tumor samplesmay be mediated at least in part via the activation ofADIPOR1 in colon cancer cells.

3.6. mRNA expression in tumors relative to pathologicalstaging, tumor size, lymph node involvement, andcolorectal location

No significant association between the Dukes’ stages andadiponectin expression was found. In contrast, there was agreater increase in AMPK expression in later-stage tumors.The relative change in AMPK mRNA expression (tumor/normal tissue) was significantly higher in Dukes’ class C(median, 1.8 ± 0.4) than in class B tumors (1.0 ± 0.2; p = 0.022).

Page 6: Globular adiponectin and its downstream target genes are up-regulated locally in human colorectal tumors: ex vivo and in vitro studies

Table 1 –mRNA expression of the adiponectin-AMPK pathway in tumor and normal tissues, as assessed using gene-arrayanalyses [50].

Array ID Genesymbol

Synonyms Gen Bank Gene name FDR

A_23_P369237 ADIPOQ NM_004797 Adiponectin, C1Q, and collagen domain containing (ADIPOQ), mRNA 26.291A_23_P46627 ADIPOR1 NM_015999 AdipoR1, mRNA 14.925A_24_P13376 ADIPOR2 NM_024551 AdipoR2, mRNA 23.762A_23_P48121 ADIPOR2 NM_024551 AdipoR2, mRNA 25.674A_23_P204503 PRKAB1 AMPK NM_006253 Protein kinase, AMP-activated, beta 1 noncatalytic subunit (PRKAB1), mRNA 13.089A_23_P16483 STK11 LKB1 NM_000455 Serine/threonine kinase 11 (Peutz-Jeghers syndrome) (STK11), mRNA; part

of the AMPKK complex8.264

A_24_P159866 STK11 LKB1 NM_000455 Serine/threonine kinase 11 (Peutz-Jeghers syndrome) (STK11), mRNA; partof the AMPKK complex

8.87

A_24_P147407 STRADA STRAD,LYK5

NM_153335 Protein kinase LYK5 (LYK5), transcript variant 3, mRNA; part of the AMPKKcomplex

27.094

A_23_P79426 CAB39 MO25 NM_016289 Calcium binding protein 39 (CAB39), mRNA; part of the AMPKK complex 0.0A_23_P104563 CPT1A CPT1 NM_001876 Carnitine palmitoyltransferase 1A (liver) (CPT1A), nuclear gene encoding

mitochondrial protein, mRNA0.0

A_23_P148919 CPT2 NM_000098 Carnitine palmitoyltransferase II (CPT2), nuclear gene encodingmitochondrial protein, mRNA

0.0

A_23_P8820cdø-ø…

FABP-4 NM_001442 Fatty-acid-binding protein 4, adipocyte (FABP-4), mRNA 19.722

A_23_P107350 SLC2A4 GLUT4 NM_001042 Solute carrier family 2 (facilitated glucose transporter), member 4 (SLC2A4),mRNA

0.098

A_24_P410678 JAK1 NM_002227 Janus kinase 1 (a protein tyrosine kinase; JAK1), mRNA 9.577A_23_P97005 JAK1 NM_002227 Janus kinase 1 (a protein tyrosine kinase; JAK1), mRNA 25.528A_23_P123608 JAK2 NM_004972 Janus kinase 2 (a protein tyrosine kinase) (JAK2), mRNA 35.921A_24_P59667 JAK3 BC028068 Janus kinase 3 (a protein tyrosine kinase, leukocyte), mRNA (cDNA clone

MGC:39993 IMAGE:5212575), complete cds0.0

A_24_P308096 JAK3 NM_000215 Janus kinase 3 (a protein tyrosine kinase, leukocyte; JAK3), mRNA 2.029A_23_P329112 JAK3 NM_000215 Janus kinase 3 (a protein tyrosine kinase, leukocyte; JAK3), mRNA 12.941A_23_P99027 PTPN11 SHP2 NM_002834 Protein tyrosine phosphatase, nonreceptor type 11 (Noonan syndrome 1;

PTPN11), mRNA0.021

A_32_P61757 PTPN11 SHP2 NM_002834 Protein tyrosine phosphatase, nonreceptor type 11 (Noonan syndrome 1;PTPN11), mRNA

0.557

A_23_P204090 PTPN11 SHP2 NM_002834 Protein tyrosine phosphatase, nonreceptor type 11 (Noonan syndrome 1;PTPN11), mRNA

17.255

A_24_P382489 SLC27A1 FATP NM_198580 Solute carrier family 27 (fatty acid transporter), member 1 (SLC27A1), mRNA 2.969A_32_P40744 THC2302865 FABP THC2302865 FABE_HUMAN (Q01469) Fatty-acid-binding protein, epidermal (E-FABP)

(Psoriasis-associated fatty-acid-binding protein homolog; PA-FABP), partial(53%)

8.366

Fig. 2 – Adiponectin protein expression was confirmed in both normal colorectal tissue and tumor samples from the samepatient byWestern blot analyses. The sampleswith different numbers represent different patients. N = normalmucosal tissue,T = tumor from a patient with the same patient number. The quantified expressions of both gAd (15 kDa; 60-min exposure ofthe membrane) and fAd (30 kDa; 3-min exposure of the membrane) are also shown. The red oblong illustrates the proteinexpression of gAd versus fAd in normal versus tumor tissue from a single patient.

677M E T A B O L I S M C L I N I C A L A N D E X P E R I M E N T A L 6 3 ( 2 0 1 4 ) 6 7 2 – 6 8 1

Page 7: Globular adiponectin and its downstream target genes are up-regulated locally in human colorectal tumors: ex vivo and in vitro studies

Fig. 3 – Protein expression of gAd was increased in colorectal tumors relative to normal colorectal tissue samples (a). Theexpression of fAd protein was lower in the tumor than in normal tissue samples (b). The ratio between gAd and fAd proteinexpressions (gAd/fAd ratio) was increased in colorectal tumors relative to normal colorectal tissue samples (c). The p valueswere obtained using pairwise Wilcoxon’s tests.

678 M E T A B O L I S M C L I N I C A L A N D E X P E R I M E N T A L 6 3 ( 2 0 1 4 ) 6 7 2 – 6 8 1

In addition, neither tumor size nor differentiation stage wasassociated with adiponectin expression. However, the num-ber of samples was small and more than 80% of the tumorshad a moderate differentiation grade. Interestingly, theexpression of FATP changed significantly more betweennormal tissue and tumor tissue in class C tumors (1.7 ± 2.3)than in class B tumors (1.0 ± 0.3; p = 0.018). The geneexpression was correlated with the size of the tumor for

Fig. 4 –WiDr colon cancer cells were first incubated with gAdfor 15 min, which increased the phosphorylation of theα-subunit of AMPK by more than twofold (15-min gAd; thetwo first bars from left to right). This effectwas suppressed bysimultaneous treatment of the cells with siRNA ADIPOR1(15-min; the third and fourth bars from the left). Thedifference inAMPKphosphorylation seen betweenuntreatedand siRNA-treated WiDr colon cancer cells in the baselinesituation (baseline versus siRNA ADIPOR1 15-min) is thoughtto depend upon the altered circumstances in cells due tosiRNA treatment. This experiment confirms the connectionbetween AdipoR1 stimulation and AMPK phosphorylationunder in vitro conditions.

both FATP (r = 0.28, p = 0.035) and FABP-pm (r = 0.24, p = 0.066).There was no statistically significant association betweenFATP (r = 0.25, p = 0.053) and AMPK expressions (r = 0.24, p =0.065), and the number of lymph nodes affected. However, thenumber of samples here are also small, since only 50% of thepatients had lymph-node metastases. Furthermore, geneexpression patterns were not associated with tumor locationin the colon, sigmoid colon, or rectum.

4. Discussion

Adiponectin is generally believed to exert a protective roleagainst carcinogenesis [5,38]. Low plasma levels of adiponec-tin have been linked to increased colorectal cancer (CRC) riskand tumor grade, especially in male patients [38,39]. It hasbeen shown that the rs266729 single-nucleotide polymor-phism, which tags the 5′ flanking region of the adiponectingene, is associated with decreased CRC risk [40]. In 2007,Körner et al. showed that women with the highest adiponec-tin levels had a 65% reduced risk of breast cancer [41]. Thesefeatures are based on the measurement of plasma levels offAd and HMW, but neither the local expression of adiponectinin the tumor tissue nor expression of its globular form havebeen either discussed or demonstrated previously in CRC. Thepresent study is the first to show that adiponectin is expressedlocally by both colorectal tumors and the adjacent normalmucosa. In line with previous results, levels of the fAd proteinwere lower in the tumors than in the adjacent healthy colonmucosa. In contrast, the total expression of adiponectinmRNAwas increased in tumors. fAd can be cleaved to producea fragment containing the globular domain (gAd), whichexerts potent metabolic effects [6]. One especially interestingfinding of the present study is that the relative ratio betweengAd and fAd was significantly higher in the tumor samplesthan in the adjacent healthy mucosa from the same patients.These results suggest that colorectal tumors produce adipo-nectin locally and then cleave it to produce gAd to a greaterextent than in the adjacent healthy colorectal mucosa.

Page 8: Globular adiponectin and its downstream target genes are up-regulated locally in human colorectal tumors: ex vivo and in vitro studies

679M E T A B O L I S M C L I N I C A L A N D E X P E R I M E N T A L 6 3 ( 2 0 1 4 ) 6 7 2 – 6 8 1

While it is currently generally accepted that adiponectin isa circulating hormone that is secreted mainly by adiposetissue [3], there are a few other indications supporting thepresent finding that it is also synthesized in cell types otherthan adipocytes, and suggesting both endocrine and auto-/paracrine effects. The autocrine production of adiponectin inendothelial cells has been correlated with the antiatherogenicproperties of the hormone [42], and the expressions of bothadiponectin and its receptor AdipoR1 have been reported inairway epithelial cells [43]. In addition, proinflammatorycytokines in muscle elicit an autocrine production of adipo-nectin [44]. Furthermore, myogenesis has been shown toinduce an autocrine production of adiponectin, which canagain be increased by treatment with proinflammatorycytokines or exposure to oxidative stress, thereby suggestingautocrine-loop-sustaining myogenesis [42]. Therefore, inaddition to the results reported here, there are reasons tobelieve that the autocrine production of gAd in response tooxidative stress or as a response to proinflammatory cyto-kines could also occur in colorectal tumor cells and act as atumor-specific mechanism for achieving metabolic modifica-tions that are oxidatively beneficial for the tumor cells.

In addition to its eventual autocrine functions, there isevidence that the adiponectin response can vary according tothe local circumstances in the tissue. In the study of Habeeb etal. showing bimodal effects of adiponectin on cell survival, theeffect of gAd appeared to be more potent for supporting cellsurvival than fAd in glucose-deprived conditions [32]. Accord-ing to that study, the effect of adiponectin would changeduring the carcinogenic process by suppressing cell prolifer-ation in favorable conditions with abundant nutrients pre-sent, while supporting cell survival during stress conditionssuch as glucose deficiency [32]. The bimodal effects ofadiponectin have also been seen in nonalcoholic steatohepa-titis (NASH). Reduced serum adiponectin levels have beenimplicated in the pathogenesis of NASH; in contrast, adipo-nectin levels were shown to increase while hepatic fatdecreased during the development of liver cirrhosis [45].

The physiological effect of adiponectin is mediated by itsreceptors, whose various isoforms can be regulated byhyperinsulinemia and hyperglycemia, with the consequenceof increased sensitivity or resistance to specific forms ofadiponectin [7,12]. The different forms of adiponectin, as wellas differential expression or activation of its receptors, mayalso partly explain some previous conflicting results regardingthe downstream effects of adiponectin stimulation in tumors[5]. AdipoR1 and AdipoR2 expressions have been reported inboth CRC and normal mucosa, with the surface expressionsbeing greater in cancerous tissue than in either normalmucosa or gastrointestinal stromal tumors. This findingsupports the hypothesis implicating adiponectin in thepathogenesis of CRC [17]. A hypoadiponectinemia-inducedup-regulation of adiponectin receptors in CRC tissues maycompensate and maintain the adiponectin signaling path-ways. Again, the same mechanism has also been discussedwith respect to NASH. The specific roles of AdipoR1 andAdipoR2, and their relative amounts in advanced NASH maybe important for estimating the net response due to therelative increase in AdipoR1 expression leading to increasedfatty acid oxidation [46]. Habeeb et al. found that the

expressions of both AdipoR1 and AdipoR2 were significantlyincreased in cell culture after glucose deficiency, mimickingthe situation in nutrient-deprived tumor cells; however, theincrease in the expression of AdipoR1 was more marked [32].The expression of AdipoR1 (but not AdipoR2) mRNA was alsosignificantly increased in the tumors in the present material.Giamalas et al. revealed the same expression pattern, exceptthat in their study the expression of AdipoR2 was alsoincreased [39]. The expected beneficial effect of adiponectinin cancer disease has prompted ongoing work in designingand developing an adiponectin receptor agonist for cancertreatment [47]. However, the findings of the present studyconfirm that a fundamental understanding of the reportedbimodal effects of adiponectin with respect to the pathophys-iology of tumor development and progression is necessarywhen evaluating possible treatment options with the adipo-nectin analogs.

It has been demonstrated that gAd is a strong activator ofAMPK, mainly by inducing its phosphorylation [42]. Knock-out of AdipoR1 by siRNA in human WiDr colon cancer cellsin the present study prevented the phosphorylation ofAMPK, thus confirming the connection between AdipoR1stimulation and AMPK phosphorylation under in vitroconditions. AMPK is a central sensor of cellular energyrequirements and can also be activated through a high AMP/ATP ratio [37]. There is evidence that gradual alterations ingene expression promote metabolic changes during tumor-igenesis [48,49]. Early premalignant lesions benefit from ageneral inhibition of AMPK activity [48], which occurs instates such as obesity and insulin resistance. In contrast,oncogene-mediated changes at the beginning of the diseaseare subsequently followed by hypoxia-induced-factor-mediated gene expression. The AMPK activation benefitsadvanced neoplasms by increasing their intracellular sub-strate availability for energy breakdown processes [48]. Thisassumption is supported by several previous studies linkinghigh AMPK expression with advanced tumor disease [48].The increased gAd expression in the tumor tissue relative tothe adjacent healthy mucosa in the present study wasassociated with increased AMPK mRNA expression, andtended to be correlated with AdipoR1 but not AdipoR2 mRNAexpression. These results should be interpreted with cau-tion, since neither the protein levels of AMPK nor itsphosphorylation were measured in the patient material.gAd-induced activation of AMPK is involved in the inhibitionof ACC2 and increased fatty-acid oxidation, glucose uptake,and lactate production [20]. However, the expressions ofseveral genes involved in the transport of fatty acids intocells (i.e., FATP, FABP-pm, and FABP-4) were simultaneouslyup-regulated in the tumor tissues compared with theadjacent healthy tissue in the present study, supportingthe biological relevance of our findings. Furthermore,Habeeb et al. showed recently that during glucose depriva-tion, the AMPK–mTOR cascade leads to the activation ofmicrotubule-associated protein 1 light chain 3, resulting incell survival via activation of the autophagic machinery ofthe malignant DLD-1 cell line [32]. In the present study,AMPK expression was activated mostly in later-stagetumors; these findings are in line with previously reporteddata on AMPK activation in tumors [48].

Page 9: Globular adiponectin and its downstream target genes are up-regulated locally in human colorectal tumors: ex vivo and in vitro studies

680 M E T A B O L I S M C L I N I C A L A N D E X P E R I M E N T A L 6 3 ( 2 0 1 4 ) 6 7 2 – 6 8 1

The present study shows for the first time that theglobular form of adiponectin is increasingly produced bycolorectal tumors ex vivo in combination with up-regulationof AdipoR1 on the cell surface. This might be a compensa-tory mechanism that occurs in response to low plasmaadiponectin levels, as reported previously in CRC, or to othercircumstances in the tumor microenvironment, such asreduced oxygen or glucose levels. Whatever the reason forthe response, this finding implicates a beneficial metabolicadvantage in the tumor cells as compared with those of thenormal mucosa. The biological relevance of these results issupported by the findings regarding several gAd targetedgenes, showing the expected increased or decreased expres-sion ex vivo. However, the major weakness of this is study isthat the results do not highlight the impact of local versussystemic changes in adiponectin signaling during tumori-genesis. The eventual bimodal actions of adiponectin shouldbe investigated during tumorigenesis, including both plasmalevels and local expression of the hormone during differentphases of the disease. The results of this study do notresolve the issue of how adiponectin expression is regulatedduring the tumorigenic process, and it is unclear whetherthese results simply represent an epiphenomenon orwhether local adiponectin activation in tumors is a signif-icant driving force for the tumorigenic process. The clinicalor translational potential of these results include a potentialtarget for pharmaceutical or lifestyle interventions, al-though activation of genes in the gAd-AdipoR1-AMPKpathway in advanced cancer disease needs to be confirmedand studied further.

Author contributions

K.K.V. designed the study, supervised the experiments, andwrote the paper. T.S. conceived and carried out experiments.K.K.V. and M.L. conceived and carried out some of theexperiments. I.B. revised the paper and coordinated thecollection of the clinical samples, and R.H.S. contributed tothe collection of clinical samples. K.J. carried out parts of theRNA isolation and the microarray experiments. K.H.P. andK.K.V. analyzed the data. T.L. analyzed the microarray data.E.F. conceived and supervised some of the experiments, andrevised the paper. All authors were involved in writing andrevising the paper, and provided their final approval of thesubmitted version.

Funding

This study was supported by grants from Akershus Univer-sity Hospital, Helsinki University Central Hospital, Jalmari jaRauha Ahokas, and Novo Nordisk Foundation (K.H.P.). E.F.was supported by The Research Council of Norway throughgrants from the Functional Genomics Program (FUGE, grantnumber 151882), Helse Sør-Øst (grant number 2007060),“Ullevål University Hospital Research Fund (VIRUUS),” and“University of Oslo Research Fund (UNIFOR).”

Acknowledgments

We thank English Science Editing for improving the use ofEnglish in this manuscript.

Conflict of interest

While working on this research article, Katja Kannisto Vetvikalso worked in parallel as Medical Director at Merck SeronoNorway, an affiliate of Merck Serono S.A. There is no conflictof interest or any linkage between this position and thecontent or results of this work. Katja Kannisto Vetvik has notreceived any commercial research grants.

Appendix A. Supplementary data

Supplementary data to this article can be found online athttp://dx.doi.org/10.1016/j.metabol.2014.02.001.

R E F E R E N C E S

[1] Vázquez-Vela ME, Torres N, Tovar AR. White adipose tissueas endocrine organ and its role in obesity. Arch Med Res2008;39:715–28.

[2] Renehan AG, Roberts DL, Dive C. Obesity and cancer:pathophysiological and biological mechanisms. Arch PhysiolBiochem 2008;114:71–83.

[3] Goldfine AB, Kahn CR. Adiponectin: linking the fat cell toinsulin sensitivity. Lancet 2003;362:1431–2.

[4] Kelesidis I, Kelesidis T, Mantzoros CS. Adiponectin andcancer: a systematic review. Br J Cancer 2006;94:1221–5.

[5] Barb D, Williams CJ, Neuwirth AK, et al. Adiponectin inrelation to malignancies: a review of existing basic researchand clinical evidence. Am J Clin Nutr 2007;86:858–66.

[6] Waki H, Yamauchi T, Kamon J, et al. Generation of globularfragment of adiponectin by leukocyte elastase secreted bymonocytic cell line THP-1. Endocrinology 2005;146:790–6.

[7] Kadowaki T, Yamauchi T. Adiponectin and adiponectinreceptors. Endocr Rev 2005;26:439–51.

[8] Scherer PE, Williams S, Fogliano M, et al. A novel serumprotein similar to C1q, produced exclusively in adipocytes.J Biol Chem 1995;270:26746–9.

[9] Fruebis J, Tsao TS, Javorschi S, et al. Proteolytic cleavageproduct of 30-kDa adipocyte complement-related proteinincreases fatty acid oxidation in muscle and causes weightloss in mice. Proc Natl Acad Sci U S A 2001;98:2005–10.

[10] Peake PW, Kriketos AD, Campbell LV, et al. The metabolismof isoforms of human adiponectin: studies in human subjectsand in experimental animals. Eur J Endocrinol 2005;153:409–17.

[11] Pajvani UB, Hawkins M, Combs TP, et al. Complexdistribution, not absolute amount of adiponectin, correlateswith thiazolidinedione-mediated improvement in insulinsensitivity. J Biol Chem 2004;279:12152–62.

[12] Yamauchi T, Kamon J, Ito Y, et al. Cloning of adiponectinreceptors that mediate antidiabetic metabolic effects. Nature2003;423:762–9.

[13] Bjursell M, Ahnmark A, Bohlooly YM, et al. Opposing effects ofadiponectin receptors 1 and 2 on energy metabolism.Diabetes 2007;56:583–93.

Page 10: Globular adiponectin and its downstream target genes are up-regulated locally in human colorectal tumors: ex vivo and in vitro studies

681M E T A B O L I S M C L I N I C A L A N D E X P E R I M E N T A L 6 3 ( 2 0 1 4 ) 6 7 2 – 6 8 1

[14] Yamauchi T, Nio Y, Maki T, et al. Targeted disruption ofAdipoR1 and AdipoR2 causes abrogation of adiponectinbinding and metabolic actions. Nat Med 2007;13:332–9.

[15] Chou SH, Tseleni-Balafouta S, Moon HS, et al. Adiponectinreceptor expression in human malignant tissues. HormCancer 2010;1:136–45.

[16] Michalakis K, Williams CJ, Mitsiades N, et al. Serumadiponectin concentrations and tissue expression ofadiponectin receptors are reduced in patients with prostatecancer: a case control study. Cancer Epidemiol BiomarkersPrev 2007;16(2):308–13.

[17] Williams CJ, Mitsiades N, Sozopoulos E, et al. Adiponectinreceptor expression is elevated in colorectal carcinomas butnot in gastrointestinal stromal tumors. Endocr Relat Cancer2008;15(1):289–99.

[18] Zakikhani M, Dowling RJ, Sonenberg N, et al. The effects ofadiponectin and metformin on prostate and colon neoplasiainvolve activation of AMP-activated protein kinase. CancerPrev Res (Phila) 2008;1:369–75.

[19] Wu X, Motoshima H, Mahadev K, et al. Involvement ofAMP-activated protein kinase in glucose uptake stimulatedby the globular domain of adiponectin in primary ratadipocytes. Diabetes 2003;52:1355–63.

[20] YamauchiT,Kamon J,MinokoshiY, etal.Adiponectinstimulatesglucose utilization and fatty-acid oxidation by activatingAMP-activated protein kinase. Nat Med 2002;8:1288–95.

[21] Huypens P, Moens K, Heimberg H, et al. Adiponectin-mediated stimulation of AMP-activated protein kinase(AMPK) in pancreatic beta cells. Life Sci 2005;77:1273–82.

[22] Li L, Wu L, Wang C, et al. Adiponectin modulates carnitinepalmitoyltransferase-1 through AMPK signaling cascade inrat cardiomyocytes. Regul Pept 2007;139:72–9.

[23] Chen H, Montagnani M, Funahashi T, et al. Adiponectinstimulates production of nitric oxide in vascular endothelialcells. J Biol Chem 2003;278:45021–6.

[24] Sugiyama M, Takahashi H, Hosono K, et al. Adiponectininhibits colorectal cancer cell growth through the AMPK/mTOR pathway. Int J Oncol 2009;34:339–44.

[25] Tomas E, Tsao TS, Saha AK, et al. Enhanced muscle fatoxidation and glucose transport by ACRP30 globular domain:acetyl-CoA carboxylase inhibition and AMP-activated proteinkinase activation. Proc Natl Acad Sci U S A 2002;99:16309–13.

[26] Fujisawa T, Endo H, Tomimoto A, et al. Adiponectinsuppresses colorectal carcinogenesis under the high-fat dietcondition. Gut 2008;57:1531–8.

[27] Dieudonne MN, Bussiere M, Dos Santos E, et al. Adiponectinmediates antiproliferative and apoptotic responses in humanMCF7 breast cancer cells. Biochem Biophys Res Commun2006;345:271–9.

[28] Ogunwobi OO, Beales IL. Globular adiponectin, acting viaadiponectin receptor-1, inhibits leptin-stimulatedoesophageal adenocarcinoma cell proliferation. Mol CellEndocrinol 2008;285:43–50.

[29] Otani K, Kitayama J, Yasuda K, et al. Adiponectin suppressestumorigenesis in Apc(Min)(/+)mice. Cancer Lett 2010;288:177–82.

[30] Moon HS, Liu X, Nagel JM, et al. Salutary effects of adiponectinon colon cancer: in vivo and in vitro studies in mice. Gut2013;62(4):561–70.

[31] Ogunwobi OO, Beales IL. Adiponectin stimulates proliferationand cytokine secretion in colonic epithelial cells. Regul Pept2006;134:105–13.

[32] Habeeb BS, Kitayama J, Nagawa H. Adiponectin supports cellsurvival in glucose deprivation through enhancement of

autophagic response in colorectal cancer cells. Cancer Sci2011;102:999–1006.

[33] Størkson RH, Aamodt R, Vetvik KK, et al. mRNA expressionof adipocytokines and glucocorticoid-related genes areassociated with downregulation of E-cadherin mRNA incolorectal adenocarcinomas. Int J Colorectal Dis 2012;27:1021–7.

[34] Wei JS. Purification of total RNA from mammalian cells andtissues. In: Bowtell D, Sambrook J, editors. DNAmicroarrays: amolecular cloning manual. New York: Cold Spring HarborLaboratory Press; 2003. p. 110–7.

[35] Tusher VG, Tibshirani R, Chu G. Significance analysis ofmicroarrays applied to the ionizing radiation response. ProcNatl Acad Sci U S A 2001;98:5116–21.

[36] Villanueva EC, Myers MG. Leptin receptor signaling and theregulation of mammalian physiology. Int J Obes (Lond)2008;32(Suppl 7):S8–S12.

[37] Kahn BB, Alquier T, Carling D, et al. AMP-activated proteinkinase: ancient energy gauge provides clues to modernunderstanding of metabolism. Cell Metab 2005;1:15–25.

[38] Guadagni F, Roselli M, Martini F, et al. Prognostic significanceof serum adipokine levels in colorectal cancer patients.Anticancer Res 2009;29:3321–7.

[39] Gialamas SP, Petridou ET, Tseleni-Balafouta S, et al. Serumadiponectin levels and tissue expression of adiponectinreceptors are associated with risk, stage, and grade ofcolorectal cancer. Metabolism 2011;60:1530–8.

[40] Kaklamani VG, Wisinski KB, Sadim M, et al. Variants of theadiponectin (ADIPOQ) and adiponectin receptor 1 (ADIPOR1)genes and colorectal cancer risk. JAMA 2008;300(13):1523–31.

[41] Korner A, Pazaitou-Panayiotou K, Kelesidis T, Kelesidis I,Williams CJ, Kaprara A, et al. Total and high-molecular-weight adiponectin in breast cancer: in vitro and in vivostudies. J Clin Endocrinol Metab Mar 2007;92(3):1041–8.

[42] Fiaschi T, Cirelli D, Comito G, et al. Globular adiponectininduces differentiation and fusion of skeletal muscle cells.Cell Res 2009;19:584–97.

[43] Miller M, Cho JY, Pham A, et al. Adiponectin and functionaladiponectin receptor 1 are expressed by airway epithelial cells inchronic obstructive pulmonary disease. J Immunol 2009;182:684–91.

[44] Delaigle AM, Jonas JC, Bauche IB, et al. Induction ofadiponectin in skeletal muscle by inflammatory cytokines: invivo and in vitro studies. Endocrinology 2004;145:5589–97.

[45] Van Der Poorten D, Samer CF, Ramezani-MoghadamM, et al.Hepatic fat loss in advanced nonalcoholic steatohepatitis: arealterations in serum adiponectin the cause? Hepatology 2013;57:2180–8.

[46] Claudel T, Trauner M. Adiponectin, bile acids, and burnt-outnonalcoholic steatohepatitis: new light on an old paradox.Hepatology 2013;57:2106–9.

[47] Otvos L, Haspinger E, La Russa F, et al. Design anddevelopment of a peptide-based adiponectin receptor agonistfor cancer treatment. BMC Biotechnol 2011;11:90–5.

[48] Ashrafian H. Cancer's sweet tooth: the Janus effect of glucosemetabolism in tumorigenesis. Lancet 2006;367:618–21.

[49] Smolková K, Plecitá-Hlavatá L, BellanceN, et al.Waves of generegulationsuppressand then restoreoxidativephosphorylationin cancer cells. Int J Biochem Cell Biol 2011;43:950–68.

[50] Hawley SA, Boudeau J, Reid JL, et al. Complexes between theLKB1 tumor suppressor, STRAD alpha/beta and MO25 alpha/beta are upstream kinases in the AMP-activated proteinkinase cascade. J Biol 2001;2:28.1–28.16.


Recommended