ANGPTL8 reverses established adriamycin cardiomyopathy by stimulating adult cardiac progenitor cells

Established adriamycin cardiomyopathy is a lethal disease. When congestive heart failure develops, mortality is approximately 50% in a year. It has been known that ANGPTLs has various functions in lipid metabolism, inflammation, cancer cell invasion, hematopoietic stem activity and diabetes. We hypothesized that ANGPTL8 is capable of maintaining heart function by stimulating adult cardiac progenitor cells to initiate myocardial regeneration. We employed UTMD to deliver piggybac transposon plasmids with the human ANGPTL8 gene to the liver of rats with adriamycin cardiomyopathy. After ANGPTL8 gene liver delivery, overexpression of transgenic human ANGPTL8 was found in rat liver cells and blood. UTMD- ANGPTL8 gene therapy restored LV mass, fractional shortening index, and LV posterior wall diameter to nearly normal. Our results also showed that ANGPTL8 reversed established ADM cardiomyopathy. This was associated with activation of ISL-1 positive cardiac progenitor cells in the epicardium. A time-course experiment shown that ISL-1 cardiac progenitor cells proliferated and formed a niche in the epicardial layer and then migrated into sub-epicardium. The observed myocardial regeneration accompanying reversal of adriamycin cardiomyopathy was associated with upregulation of PirB expression on the cell membrane of cardiac muscle cells or progenitor cells stimulated by ANGPTL8.


INTRODUCTION
Adriamycin (ADM) is one of the most widely prescribed and effective cytotoxic drugs used in oncology. However, the utility of adriamycin is limited by cumulative, dose-related, progressive myocardial damage that may lead to congestive heart failure (CHF). Thus, patients who may benefit from continued administration of the drug must withdraw from adriamycin therapy and switch to an alternative agent, which may be less effective [1]. An estimated cumulative percentage of 5% of adriamycin induced CHF patients at a cumulative dose of 400 mg/m 2 , 26% of patients at 550 mg/m 2 , and 48% of patients at 700mg/m 2 . Although adriamycin cardiomyopathy accounts for only a small number of heart failure cases compared to ischemic cardiomyopathy or idiopathic dilated cardiomyopathy, it is a lethal disease. When congestive heart failure develops, mortality is approximately 50% in a year. Moreover, it often occurs in relatively young patients with curable malignancies such as breast cancer. Adriamycin is well known for its cardiac toxicity during chemotherapy for cancer. The cardiotoxicity is due to free radical formation and lipid peroxidation, which results in changes in lysosomes, sarcolemma's, mitochondria, and sarcoplasmic reticulum. These changes induce calcium overload, activation of

Research Paper
Oncotarget 80392 www.impactjournals.com/oncotarget hydrolytic enzyme, and reduction in energy production. Reduction of cardiac function was a result of the loss of structural integrity. Because adriamycin cardiotoxicity is dose dependent, it has been used to predictably induce heart failure in different animal species considered as a decent animal model with heart failure [2][3][4][5][6]. It will be very crucial to study how to prevent or cure adriamycin cardiomyopathy for these cancer patients. Traditional medical therapy for CHF is directed toward relief of symptoms and blockade of neurohormonal activation. In patients with advanced stages of CHF, heart transplant or left ventricular assist devices may be appropriate, but are expensive and have limited availability. The ideal goal for CHF therapy is myocardial regeneration. It is currently known that cardiac myocytes are capable of regeneration, via various mechanisms, including selfreplication of pre-existing adult cardiac muscle cells [7][8], differentiation of resident cardiac progenitor cells [9][10], dedifferentiation and proliferation of adult cardiac muscle cells [11][12][13] and transdifferentiation of fibroblast cells into cardiac muscle cells [14][15]. However, the magnitude of myocardial regeneration triggered by such mechanisms is small and it remains unclear whether myocardial regeneration in heart failure is sufficient to reverse established cardiomyopathy. Multiple clinical trials of stem cell therapy have been attempted but without clear improvement in cardiac function [16]. Part of this problem may be the fact that patients with advanced HF often have lack of blood supply to the heart and extensive scar tissue due to ischemic cardiomyopathy or longstanding non-ischemic cardiomyopathy with extensive fibrosis. The LIM homeodomain transcription factor (ISL-1 or Islet1) is a biomarker of cardiac progenitors or neural stem cells that give rise to the right ventricle, atria and outflow tract during cardiac development [17]. These cells can differentiate to the different cell types that form an adult heart (cardiomyocytes, smooth muscle cells, and endothelial cells) [18][19], However ISL-1 positive progenitor cells are usually quiescent during adult heart, being present in parasympathetic neurons, smooth muscle cells, a few cardiomyocytes within the proximal aorta and pulmonary artery and more abundantly in the sinoatrial node. The epicardium contains epicardium-derived cells (EPDCs), which undergo epithelial-to-mesenchymal transition (EMT) and differentiate into all main cell lineages of the heart [20]. However some evidence shows that adult EMCs primarily undergo fibrogenic EMT upon cardiac stress, such as hypertension or infarction, to generate myofibroblast-like cells [21]. These cells can contribute to the development of cardiac disease, such as fibrosis, potentially leading to impaired cardiac performance and arrhythmias, including sudden death. It is very crucial to find out the molecular and cellular aspects of these EMCs as well as associated factors, and their implication in EMT and cardiac fibrosis in order to prevent or treat certain heart diseases.
Angiopoietin-like proteins are a family of 8 member proteins that are structurally similar to the angiopoietins, all containing an amino-terminal coiled-coil domain, a linker region, and a carboxy-terminal fibrinogenlike domain except ANGPTL8. ANGPTLs have various functions in lipid metabolism, inflammation, cancer cell invasion, hematopoietic stem activity. ANGPTL8 (also known as C19orf80) is a 596 bp cDNA with a short sequence in the N-terminus that is highly enriched in the liver and adipocytes [22][23][24][25][26][27]. Yi et al [28][29] found that a hepatic overexpression of ANGPTL8 after a hydrodynamic injection of non-viral sleeping beauty transposon plasmids into the tail vein of adult mice had several effects: it stimulated the proliferation of pancreatic beta cells, increased insulin secretion, improved glucose tolerance and expanded the mass of the beta cell islets. We have confirmed the pancreatic beta cell proliferation induced by ANGPTL8, but it was not able to reverse diabetes in rats with STZ-induced diabetes [30] or obese rhesus monkey with naturally occurring diabetes [data not shown]. To our knowledge, the effects of ANGPTL8 on myocardial regeneration in heart failure have not been reported. Our group developed a novel approach to gene therapy by targeting the delivery of non-viral DNA to the organs using ultrasound-targeted microbubble destruction (UTMD) [31][32][33][34][35][36][37]. In this study, we employed UTMD to deliver the ANGPTL8 gene to the liver of rats with established cardiomyopathy induced by adriamycin. We found that hepatic overexpression of ANGPTL8 can reverse established cardiomyopathy.

RESULTS
UTMD targeting of human ANGPTL8 gene to liver of rats Figure 1 shows that human ANGPTL8 signal was detected in liver cells after UTMD-pXL-BASII-CI-ANGPTL8/pCI-hyPB delivered to the liver ( Figure 1D) but was not detected in normal rat liver, nor in the controls treated with ADM only and ADM plus UTMD-DsRed reporter gene ( Figure 1A-1C). Human ANGPTL8 signal was detected in the cytoplasm of liver cells of rats from 3 days to 4 weeks after delivery of the ANGPTL8 gene to liver by UTMD. Figure 1E shows fasting plasma levels of human ANGPTL8 in rats. In the normal and DsRed reporter gene control rat groups, no human ANGPTL8 was detected in fasting plasma. However, in the treatment groups, human ANGPTL8 was detected in fasting plasma from day 1 to day 28 after UTMD. We further evaluated human ANGPTL8 mRNA levels using qRT-PCR. The results ( Figure 1F) showed that human ANGPTL8 mRNA levels after UTMD-ANGPTL8 gene delivery were 25± 8 fold greater, respectively, than for the normal, ADM www.impactjournals.com/oncotarget only and ADM plus UTMD-DsRed control groups (p<0.001). Western blotting was employed to detect human ANGPTL8 from liver protein extracts, and the results showed that the human ANGPTL8 signal existed in the liver protein extracts after UTMD-ANGPTL8 gene delivery (data not shown).

Reversal of established adriamycin cardiomyopathy after UTMD-ANGPTL8 gene therapy
We decided to evaluate if there is any possible pharmacological effects of ANGPTL8 on rat hearts with established adriamycin cardiomyopathy. We employed echocardiography to evaluate heart structure and function in all groups. Figure 1H-K demonstrated M-mode images derived from 2D parasternal short axis views of the left ventricle showing decreased LV fractional shortening and LV mass in adriamycin cardiomyopathy with restoration to normal values by UTMD-ANGPTL8 gene therapy ( Figure  1L and M).

ANGPTL8 significantly upregulated PirB expression on membrane of cardiac muscle cells
Zheng, et al [38] reported that Paired immunoglobulin-like receptor B (PirB) in rodents or LILRB2 in human is a specific receptor of ANGPTLs in membrane of hematopoietic stem cells. ANGPTLs stimulate expansion of hematopoietic stem cells ex vivo by activating PirB. It also is an inhibitory receptor in neurons that inhibited regeneration of adult neuron cells [39]. However, there are no reports of PirB in adult cardiac muscle cells. Figure 2 shows robust presence of PirB on the surface of cTNT positive cardiac muscle cells ( Figure 2A). However, under established cardiomyopathy induced by ADM, PirB was rarely seen ( Figure 2B-C), after reversal of ADM cardiomyopathy by ANGPTL8 gene liver delivery, PirB was again seen on the surface of membrane of cardiac muscle cells ( Figure 2D). We quantified PirB mRNA levels using qRT-PCR. The results ( Figure 2E) showed that PirB mRNA levels are 89±16, 12±4, 16±6, and 82±24 folds respectively in normal rat heart, ADM only control heart, ADM plus UTMD-DsRed Oncotarget 80394 www.impactjournals.com/oncotarget heart, and ADM plus UTMD-ANGPTL8 heart groups.
The results of western blotting shows that PirB existed in all the heart protein extracts, was significantly deceased in ADM treated heart samples, and recovered after UTMD-ANGPTL8 gene delivery ( Figure 2F). We decided to use cultured mouse HL-1 atrial cardiomyocyte cells to confirm what we found in vivo experiment, Figure 3A shown that PirB clearly localized on the membrane of HL-1 atrial cardiomyocyte cells as a membrane receptor, and exogenously ANGPTL8 peptide are binding with PirB on the membrane of HL-1 atrial cardiomyocyte cells ( Figure  3B-D). Figure 4D shown that ISL-1 positive signals were seen in nucleus of cardiac muscle cells 28 days after ADM plus UTMD-ANGPTL8 but not in other groups ( Figure  4A-C). The percentage of ISL-1 positive cardiac muscle cells was 3.8±1.2% in ADM plus UTMD-ANGPTL8 group (p<0.001 vs controlling groups) ( Figure 4F). qRT-PCR shown that mRNA level of ISL-1 is 21-fold higher in ADM plus UTMD-ANGPTL8 compared to controls Oncotarget 80395 www.impactjournals.com/oncotarget (p<0.001) ( Figure 4E). The results of western blotting shown that ISL-1 signal was scant in nuclear protein extracts of normal heart, ADM only and ADM plus UTMD-DsRed heart samples, but significantly increased in after UTMD-ANGPTL8 gene delivery( Figure 4K). Finally, the location of ISL-1 positive cells was confined to the epicardium and sub-epicardium ( Figure 4I) and not seen in endocardium and myocardium ( Figure 4G-H and J) (P<0.001).

ANGPTL8 significantly activated ISL-1 positive cardiac progenitor cells in epicardial layer
We decided further to investigate the effect of ANGPTL8 on ISL-1 positive cardiac progenitor cells with a time-course experiment. Figure 5 shows that no ISL-1 signal was seen in epicardial layer cells before UTMD-ANGPTL8 gene delivery ( Figure 5A), but was observed 3 days post UTMD-ANGPTL8 treatment ( Figure 5B). ISL-1 positive cells appeared to cluster into epicardial niches of cardiac progenitor cells 7 days post UTMD-ANGPTL8 gene delivery ( Figure 5C), further proliferating and migrating into sub-epicardial myocardium ( Figure 5D-E) and finally differentiating into new cardiac muscle cells ( Figure 5F).

Regenerating cardiac muscle cells are in proliferation
We used a mitotic marker (anti-phospho-histone H3 (Ser10) (PHH3)) to demonstrate if regenerating cardiac muscle cells were proliferating. We calculated the percentage of PHH3 positive cardiomyocytes by counting stained nuclei (pink color) from 1000 cTnT Oncotarget 80396 www.impactjournals.com/oncotarget positive cardiomyocytes cells in the sub-epicardium using serial sections through each rat heart (n=6 each group). Figure 6 shows that PHH3 signal was observed within the nuclear positive cells by confocal microscopy in ADM plus UTMD-ANGPTL8 groups. The percentage of PHH3 positive sub-epicardial muscle cells in the rats treated with UTMD-ANGPTL8 gene therapy 28 days post-UTMD ( Figure 6F) was 3.8±0.60% (p<0.001 vs controls). Figure 6D-E shows evidence of cell proliferation (PHH3 expression) confined to the epicardial layer. The western blotting results showed that PHH3 signal was not seen in the adult heart nuclear protein extracts except in the ADM plus UTMD-ANGPTL8 gene delivery hearts ( Figure 6H).

ANGPTL8 protein therapy also reversed established ADM cardiomyopathy
We wonder if exogenous ANGPTL8 protein delivery has the same effects as its gene liver delivery. We set three different dosage of ANGPTL8 protein for intramuscular injection daily for 14 days; Supplemental Figure 1A shows fasting plasma levels of human ANGPTL8 protein in rats. No human ANGPTL8 was detected in fasting plasma in ADM plus saline injection, in the three different dosages groups. However, human ANGPTL8 was detected in fasting plasma from day 3, 7, 10 and 14 during daily injection of ANGPTL8 protein for consistently 14 days. Supplementary Figure 1B-C demonstrated M-mode images derived from 2D parasternal short-axis views of the left ventricle, compared with normal rat heart, ADM plus saline injection control rat heart, ADM plus ANGPTL8-5µg/kg rat heart, ADM plus ANGPTL8-20µg/ kg rat heart and ADM plus ANGPTL8-40µg/kg rat heart. There was decreased LV fractional shortening and LV mass in adriamycin cardiomyopathy with some degree of restoration by ANGPTL8 protein therapy. Supplementary  Figure 2 shows that ISL-1 positive cells in epicardial layer were activated by ANGPTL8 protein injection with a dose dependent manner after 14 days treatment. Supplementary  Figure 3 displayed that some epicardial cells and subepicardial cardiac muscle cells are in proliferation after ANGPTL8 protein therapy.

DISCUSSION
Balmer, et al [40] reported that cardiac progenitor cells in epicardial layer of adult heart are a resource for Oncotarget 80397 www.impactjournals.com/oncotarget myocardial regeneration and they are able to differentiate into cardiac muscle cells, coronary artery muscle cells and vascular endothelial cells. They appear to originate from migration of hematopoietic stem cells from bone marrow. Under some pathological states, these cells may differentiate into fibroblasts, forming scar tissue in heart. It will be a crucial to study how to drive these cardiac progenitor cells to differentiate into muscle cells but not fibroblasts. There is some evidence that ANGPTL1-7 expands hematopoietic stem cells ex vivo and stimulates the formation of niches of hematopoietic cells in bone marrow. Our study demonstrates that ANGPTL8 can stimulate adult myocardial regeneration by proliferation of cardiac progenitor cells located at epicardium. Timecourse experiments suggest that ANGPTL8 drove the differentiation of cardiac progenitor/stem cells into ISL-1 positive cardiac progenitor cells which formed niches and migrated into sub-epicardial myocardium. It still is not clear what the relationship is between WT1 positive progenitor cells and ISL-1 cardiac progenitor cells in epicardium. The exact molecular mechanism by which ANGPTL8 activated these quiescent adult cardiac  Oncotarget 80399 www.impactjournals.com/oncotarget progenitor cells in adult hearts with established ADM cardiomyopathy is also not known.
Zheng, et al [38,[41][42][43] showed that the human immune inhibitory receptor leukocyte immunoglobulinlike receptor B2 (LILRB2) and its mouse ortholog paired immunoglobulin-like receptor (PirB) are receptors for several angiopoietin-like proteins. LILRB2 and PirB are expressed on human and mouse hematopoietic stem cells, respectively, and the binding of ANGPTLs to these receptors supported ex vivo expansion of hematopoietic stem cells. In mouse transplantation acute myeloid leukemia models, a deficiency in intracellular signaling of PirB resulted in increased differentiation of leukemia cells, revealing that PirB supports leukemia development. Their study indicated an unexpected functional significance of classical immune inhibitory receptors in maintenance of normal adult stem cells and in support of cancer development. Supplemental Figure  4 show that PirB exists on the surface membrane of ISL-1 positive cardiac progenitor cells. Our study suggests that ANGPTL8 expands cardiac progenitor cells in epicardium by the activation of PirB signal pathway. Although some data indicates that this inhibitory receptor signal is related to the Notch1 control pathway [44][45][46], further studies are needed to decipher the role of PirB in adult cardiac muscle cells under physiological or pathological states. Nevertheless, our finding that ANGPTL8 reverses established ADM induced cardiomyopathy offers hope that pathways exist for myocardial regeneration in adult animals. In conclusion, ANGPTL8 has previously unrecognized effects that result reversal of established cardiomyopathy by activation of PirB receptor on the cell membrane of resident adult cardiac progenitor cells. Further evaluation of the mechanism of the effect of ANGPTL8 on cardiac PirB receptors is needed.

Animal protocols
Animal studies were performed according to National Institutes of Health (NIH) recommendations and approved by our institutional animal research committee. Adult male Sprague-Dawley rats were purchased from Harlan Laboratories (Indianapolis, IN, USA).
The protocol was planned to test the hypothesis that delivery of ANGPTL8 by UTMD could reverse established adriamycin (ADM) cardiomyopathy defined as a fractional shortening < 30% by echocardiography after injection of ADM [47] at total dose of 15 mg/kg/ip, 2.5mg/ kg/ip 6 times over 2 weeks. Roughly 70% of animals developed an established cardiomyopathy after 6 doses of ADM. Only the animals with a fractional shortening <30% were selected into the protocols.
Male Sprague-Dawley rats (230-270g) were anesthetized with intraperitoneal ketamine (60 mg/ kg) and xylazine (5 mg/kg), and a polyethylene tube (PE 50, Becton Dickinson, Franklin Lakes, TN, USA) was inserted into the right internal jugular vein by cutdown. Piggybac transposon donor plasmids and helper plasmids ratio (pXL-BASII-CI-ANGPTL8/pCI-hyPB) was 5:1. Microbubble or control solutions (0.5 ml diluted with 0.5 ml phosphate-buffered solution (PBS)) were infused over 1 min 30 seconds via pump (Genie, Kent Scientific, Torrington, CT). During the infusion, ultrasound was directed to the liver using a commercially available ultrasound transducer (S3, Sonos 5500, Philips Ultrasound, Bothell, WA). Ultrasound was then applied in ultraharmonic mode (transmit 1.3 MHz/receive 3.6 MHz) at a mechanical index of 1.2. Four bursts of ultrasound were triggered to every fourth end-systole by electrocardiogram using a delay of 45-70 ms after the peak of the R wave. These settings have shown to be optimal for plasmid delivery by UTMD using this instrument [48]. Bubble destruction was visually apparent in all rats. After UTMD, the jugular vein was tied off, the skin closed, and the animals allowed to recover. All of rats were euthanized using an overdose of sodium pentobarbital (120 mg/kg).

Manufacture of plasmid-containing lipidstabilized microbubbles
Lipid-stabilized microbubbles were made as previously described [48]. Briefly, 250 mg of 1,2-dipalmitoyl-sn-glycero-3-phosphocholine, 50 mg of 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine and 10% glucose were mixed with PBS and boiled in water until fully dissolved. Next 2 mg of plasmid DNA was mixed with 0.5 ml of ethyl alcohol and centrifuged at 10,000 g for 5 min. The supernatant fraction was removed, and the DNA pellet was placed in an incubator at 37°C for 5 min to remove any remaining ethyl alcohol. The DNA was then added to 50 µl of Lipofectamine 2000 (1 mg/ml; Invitrogen, Carlsbad, CA, USA) and mixed for 20 min. This mixture was added to 250 μl of liposome solution, 5 μl of 10% albumin and 50 μl of glycerol in 1.5 ml vials, and placed on ice. The headspace of the vials was filled with perfluoropropane gas, and the vials were then shaken for 30 seconds at 4°C. The mean diameter and concentration of the microbubbles were 1.9 ±0.2 μm and 5.2 ± 0.3 × 10 9 bubbles per ml, respectively. The concentration of plasmid carried by the microbubbles was 250±10 μg/ml.

Plasmid constructs
Human ANGPTL8 cDNA with Sal1/Not1 cutting sites (Integrated DNA Technologies, Coralville, IA, USA) was subcloned into PiggyBac transposon plasmids (pXL-BSII donor plasmid) provided by Dr. Fraser at the University of Notre Dame (Notre Dame, IN, USA) [49], and hyperactive piggyBac transposase helper plasmid was provided by Dr. Bradley at Wellcome Trust Sanger Institute (Cambridge, UK) [50]. Cloning, isolation and purification of the plasmids were performed by standard procedures, and the PCR products were sequenced to confirm that no artefactual mutations were present.

Immunohistochemistry
Tissue samples were fixed in 10% formalin for 24 hours and transferred into 70% alcohol for paraffin embedding and 4% paraformaldehyde and 20% sucrose overnight at 4 ºC for frozen sections. Cryostat sections 5-8 µm in thickness were further fixed with acetone (-20ºC) for 5 min and quenched for 5-20 min with 10 mM glycine in PBS. Sections were then rinsed in PBS 3 times, and permeabilized with 0.5% Triton X-100 in PBS for 15 min. The slides that needed further nuclear protein retrieval were subjected to boiling citrate buffer solution with tween 20 at pH 6.0 for 5 minutes. Sections were blocked with Aquablock solution (EastCoast Bio, North Berwick, ME) at room temperature for 1hr and washed with PBS 1 time. The primary antibodies rat anti-PirB, 1: 200 dilution ( BD biosciences, San Jose, CA ) , rabbit anti-Notch1, 1:200, rabbit anti-FoxO1, 1:250, and rabbit anti-ISL-1, 1: 500 dilution, and mouse anti-cardiac troponin T, 1: 250 dilution, rabbit anti-PHH3, 1:200 dilution, rabbit anti-human ANGPTL8 at 1:100 dilution, rabbit anti-albumin, 1:250 dilution (Abcam Inc, Cambridge, MA), mouse anti-human ANGPTL8 1:200 dilution (Phoenix Pharmaceuticals, Burlingame, CA, USA), were added and incubated for 2 hrs at RT or overnight at 4 ºC. After washing with PBS three times for 5 min, the secondary antibody (Sigma, St; Louis, MO) anti-mouse lgG conjugated with FITC; anti-rabbit IgG-conjugated with Texas Red, or anti-donkey lgG conjugated with Cy5) (1:250 dilution in block solution) were added and incubated for 1 hr at RT. Sections were rinsed with PBS for 10 min, 3 times, and incubated with Dapi (Invitrogen, Carlsbad, CA), 1: 5000 dilution for 5 min and washed 3 times with PBST, then mounted. A confocal microscope was used to take pictures.

Culture of HL-1 Atria Muscle Cells line.
The HL-1 cell line was a generous gift from Dr. William C. Claycomb [51] in Louisiana State University Medical Center, New Orleans, LA. The cells were maintained in Claycomb basal medium (Sigma) supplemented with 10% fetal bovine serum, 0.1 mM norepinephrine and 2 mM L-glutamine. HL-1 cells were treated with adriamycin at 0 µmol/L, 0.25 µmol/L, 0.50 µmol/L, or 1.00 µmol/L for 48 hrs in complete growth medium, or add ANGPTL8 gene plasmids transfection with lipofectamin2000, or 2 pmol/L ANGPTL8 peptide (Phoenix Pharmaceuticals, Burlingame, CA, USA) to a dish and were subjected to immunofluorescent staining.

ELISA for detecting plasma ANGPTL8
Fasting venous blood was collected in EDTA tubes from the tail vein, and plasma lipids were measured using standard enzymatic assays. Human ANGTPL8 was measured using ELISA (Phoenix Pharmaceuticals, Burlingame, CA, USA).

RNA isolation and quantitative RT-PCR analysis
Total RNA was isolated from 100 mg of heart using the RNeasy mini kit (Qiagen). Real-time quantitative RT-PCR (qRT-PCR) analysis was performed on an ABI 7700 Sequence Detector (Applied Biosystems, Grand Island, NY, USA) using SYBR Green (RT2 SYBR Green qPCR Kit; Qiagen, Boston, MA, USA). Data were normalized to the expression of housekeeping genes (as an endogenous control). Changes in gene expression were normalized to Oncotarget 80401 www.impactjournals.com/oncotarget control liver and cardiac muscle samples.

Data analysis
Data were analysed using Statview software (SAS, Cary, NC, USA). The values are presented as mean ± SEM. Differences were analyzed by repeated measures ANOVA with Fisher's post hoc test and were considered significant at p<0.05.

Supplementary material
Supplementary material is available at Oncotarget online.

ACKNOWLEDGMENTS
SC and PAG were responsible for the conception and design of the study, acquisition and analysis of the data and drafting of the article. SC and JC performed data acquisition and analysis and contributed to writing the manuscript. JS, XM, PH, JH, ZP, and NHM made substantial contributions to the conception of the study and critically revised the manuscript. PAG is the guarantor of this work and as such had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All listed authors approved the final version of the manuscript.

CONFLICTS OF INTEREST
The authors declare that there is no conflict(s) of interest associated with this manuscript.