Abstract

   Adenosine kinase (ADK) plays the major role in cardiac adenosine
   metabolism, so that inhibition of ADK increases myocardial adenosine
   levels. While the cardioprotective actions of extracellular adenosine
   against ischemia/reperfusion (I/R) are well-established, the role of
   cellular adenosine in protection against I/R remains unknown. Here we
   investigated the role of cellular adenosine in epigenetic regulation on
   cardiomyocyte gene expression, glucose metabolism and tolerance to I/R.
   Evans blue/TTC staining and echocardiography were used to assess the
   extent of I/R injury in mice. Glucose metabolism was evaluated by
   positron emission tomography and computed tomography (PET/CT).
   Methylated DNA immunoprecipitation (MeDIP) and bisulfite sequencing PCR
   (BSP) were used to evaluate DNA methylation. Lentiviral/adenovirus
   transduction was used to overexpress DNMT1, and the OSI-906 was
   administered to inhibit IGF-1. Cardiomyocyte-specific
   ADK/IGF-1-knockout mice were used for mechanistic
   experiments.Cardiomyocyte-specific ADK knockout enhanced glucose
   metabolism and ameliorated myocardial I/R injury in vivo.
   Mechanistically, ADK deletion caused cellular adenosine accumulation,
   decreased DNA methyltransferase 1 (DNMT1) expression and caused
   hypomethylation of multiple metabolic genes, including insulin growth
   factor 1 (IGF-1). DNMT1 overexpression abrogated these beneficial
   effects by enhancing apoptosis and decreasing IGF-1 expression.
   Inhibition of IGF-1 signaling with OSI-906 or genetic knocking down of
   IGF-1 also abrogated the cardioprotective effects of ADK knockout,
   revealing the therapeutic potential of increasing IGF-1 expression in
   attenuating myocardial I/R injury. In conclusion, the present study
   demonstrated that cardiomyocyte ADK deletion ameliorates myocardial I/R
   injury via epigenetic upregulation of IGF-1 expression via the
   cardiomyocyte adenosine/DNMT1/IGF-1 axis.

   Keywords: Adenosine, Ischemia-reperfusion, ADK, IGF-1, Epigenetics,
   Metabolism

1. Introduction

   Acute myocardial ischemia, resulting from arterial thrombosis, severely
   depresses cardiac contractile function. Paradoxically, while
   thrombolysis and reperfusion of the ischemic tissue is necessary to
   restore cardiomyocyte contractile function, a significant portion of
   the cell injury and death occur in response to reperfusion. The cardiac
   injury imposed by ischemia and reperfusion is defined as cardiac
   ischemia-reperfusion (I/R) injury, and contributes to the long-term
   poor prognosis in ischemic heart disease [[51]1]. Ischemic heart
   disease resulting from I/R injury is associated with high morbidity
   [[52]2]. Therefore, identifying cell and molecular mechanisms that
   contribute to cardiomyocyte death after reperfusion will be critical
   for development of adjunctive therapies for myocardial infarction.

   Adenosine is a purine nucleoside produced during cellular stress
   conditions such as inflammation, diabetes, and methionine cycle
   [[53][3], [54][4], [55][5]]. Adenosine activation of adenosine
   receptors or related kinase regulation exert important protective
   effects, including protection against ischemia/reperfusion injury
   [[56]6,[57]7], reduction of oxidative stress [[58]8], improved coronary
   flow/cardiac perfusion [[59]9], regulation of inflammatory response
   [[60]10,[61]11], and attenuation of hypertrophy and heart failure
   during pressure overload [[62]12]. Despite extensive evidence, however,
   that adenosine receptor activation can attenuate I/R injury in cell and
   animal models, the protective effects of adenosine receptor agonists,
   as well as other preconditioning or post-conditioning agents, have
   failed to translate into clinical practice [[63]13]. Thus, there is a
   clinical unmet need for identification of novel targets for attenuating
   reperfusion injury.

   Adenosine kinase (ADK) plays a major role in cardiomyocyte adenosine
   metabolism, so that its inhibition results in cellular adenosine
   accumulation and release of adenosine into the interstitial space
   [[64]14,[65]15]. While the cardioprotective roles of extracellular
   adenosine are established, the impact of cellular adenosine in
   cardioprotection is not clear. As a byproduct of s-adenosyl
   homocysteine (SAH) hydrolysis, adenosine can inhibit
   s-adenosylhomocysteine hydrolase (SAHH) activity, causing SAH to
   accumulate [[66]16]. Because SAH is a potent inhibitor of
   methyltransferase activity, inhibition of SAHH by cellular adenosine
   results in increased SAH to s-adenosylmethionine (SAM) ratio and can
   diminish DNA methylation [[67]17,[68]18]. While epigenetic effects of
   ADK disruption have been observed in several cells and tissues, the
   role of ADK and cellular adenosine in cardiomyocyte DNA methylation and
   cardioprotection are unknown.

   Recently, cross-talk between metabolism and epigenetics has been
   identified as a potential target for the treatment of cardiovascular
   diseases [[69]19,[70]20], and several studies have indicated a dynamic
   relationship between metabolic processes and gene expression
   [[71]21,[72]22]. Furthermore, cellular metabolites, such methyl and
   flavin adenine dinucleotide (FAD), have been shown to induce permanent
   alterations in cellular morphology and genetic structure in multiple
   disease models [[73]23,[74]24], likely dependent upon epigenetic
   modifications. This phenomenon is referred to as metaboloepigenetics.

   In this study, we investigated the role of ADK in cardiomyocyte DNA
   methylation, gene expression and cardioprotection against I/R in
   cardiomyocytes. Our findings indicate that inhibition of ADK protected
   heart against I/R injury by modulating epigenetic-metabolic crosstalk
   in cardiomyocytes.

2. Methods

   The data, analytical methods, and study materials that support the
   findings of this study are available from the corresponding author upon
   reasonable request.

2.1. Bioinformatic analysis to identify potential metabolic genes

   Raw data or series matrix files for microarray datasets were downloaded
   from the public Gene Expression Omnibus (GEO) database. The microarray
   sequencing results were derived from I/R samples of each group to
   reduce heterogeneity within the analysis. Analysis was performed for
   the corresponding annotation documents, pathway/process enrichment, and
   protein interactions identified by Search Tool for the Retrieval of
   Interacting Genes/Proteins (STRING) to transform the probes into gene
   symbols, while BioGPS tools were used to detect ADK expression in
   different tissues or organs as previously described [[75]22].

2.2. Transgenic mice and animal experiments

   Cardiomyocyte-specific ADK-knockout mice (ADK-Cre) were generated as
   previously described [[76]12] ([77]Fig. S12), and
   cardiomyocyte-specific insulin-like growth factor 1 (IGF-1)-knockout
   mice were purchased from Shanghai Model Organisms Center, Inc. These
   two mouse strains were crossed to establish cardiomyocyte-specific
   ADK/IGF-1 double-knockout mice. Myocardial I/R surgery was performed in
   the mice as previously described [[78]25]. Briefly, mice were
   anesthetized by a facemask connected to ventilator (2% isoflurane mixed
   with 100% O2, volume: 2 L/min, frequency: 60/minute). For each mouse, a
   left thoracic incision was made to expose the heart. Myocardial
   ischemia was induced by tying a slipknot around the left anterior
   descending coronary artery with a 6.0 silk suture (Surgical Specialties
   Corporation, England). After 30 min, the slipknot was released to allow
   reperfusion. Reperfusion times varied based on the experiment being
   performed. Reperfusion times ranged between 3 h and 24 h for
   investigation of adenosine-associated methylation levels. IGF-1R
   inhibitor (OSI-906, 10 mg/kg) was administered to mice via intragastric
   delivery, and DNMT1 adenovirus was administered via tail vein
   injection. To explore detailed changes in vivo, the specific cell
   sub-population were harvested and prepared immediately for further
   experiments, the details of protocols were described in supplementary
   files.

2.3. Echocardiography

   Mice were anesthetized with 1% pentobarbital after myocardial I/R
   surgery, and two-dimensional transthoracic echocardiography was
   conducted in a standard setting using a 30-MHz high-frequency scan head
   (VisualSonics Vevo770; VisualSonics Inc., Toronto, ON, Canada) as
   previously described [[79]22]. The left ventricular internal dimension
   at diastole (LVIDd), left ventricular internal dimension at systole
   (LVIDs), left ventricular end-diastolic volume (LVEDV), and left
   ventricular end-systolic volume (LVESV) were measured for each mouse,
   and the left ventricular fractional shortening (FS) and ejection
   fraction (EF) were calculated using Simpson's rule ([80]Fig. S13E). All
   parameters were measured by an experienced echocardiographer who was
   blinded to the treatment protocol ([81]Supplementary Table 4).

2.4. Evans blue and 2,3,5-triphenyltetrazolium chloride (TTC) staining

   Mice were sacrificed at 3- and 24-h after I/R surgery, and the hearts
   were quickly sectioned for Evans blue and TTC staining as previously
   described [[82]22]. Evans blue staining was used to assess the risk
   zones, and TTC staining was used to measure the areas of the infarcted
   tissue and entire left ventricle. Analysis was conducted using ImageJ
   software as previously described [[83]22].

   Assessment of ADK expressions, adenosine levels, SAM/SAH-associated
   metabolites, global DNA methylation, DNMT1 activity and specific
   histone methylation.

   Quantitative measurements of cellular adenosine, SAM/SAH-associated
   metabolites were assessed by reversed-phase HPLC on freshly isolated
   cells obtained through the sub-population isolation method, and
   isolated cardiomyocytes were cultured and verified by commercial kits.
   Relative amounts were calculated with the average OD of wild type as
   respective reference in our study for ELISA experiment.The ADK
   expressions, ADK activity, global DNA methylation, DNMT1 activity and
   specific histone methylation were determined using commercial kits
   according to the manufacturer's instructions. The detailed information
   of kits was presented in [84]Supplementary Table 1. The HPLC assay
   method is detailed in the Appendix Supplementary Method.

2.5. Detection of adenine nucleotide related metabolites

   The cAMP concentrations were determined using a Cyclic AMP XP® assay
   kit (Cell Signaling Technology, Danvers, MA, USA) according to the
   manufacturer's protocol. AMP, ADP, ATP and adenine were measured by
   liquid chromatography (LC)-mass spectrometry (MS) analysis, and the
   details were provided in Appendix Supplementary Method.

2.6. Terminal transferase-mediated dUTP nick end-labeling (TUNEL) assay

   Myocardial tissue sections after I/R surgery were stained using a
   fluorescein-conjugated TUNEL in situ cell death detection kit (Roche.
   Catalog No. 11684795910). Procedures were performed according to
   manufacturer's instructions. The number of TUNEL-positive cells was
   determined in 10 independent fields for each sample.

2.7. Hypoxia/reoxygenation (H/R) injury cell model

   The H/R injury cell model was established as previously described
   [[85]25]. Briefly, healthy adult mouse ventricular cardiomyocytes were
   isolated (Details in supplementary appendix files), or HCM-a human
   cardiomyocyte were cultured for ADK overexpression experiment. Cells
   for in vivo experiments were further under isolation and prepared for
   test immediately, while cells for in vitro experiments cultured in
   laminin-coated dishes in a standard tissue culture incubator (37 °C, 5%
   CO[2]) for 24 h. Then, the cardiomyocytes were placed in hypoxic buffer
   (118 mmol/L NaCl, 24 mmol/L NaHCO[3], 1 mmol/L NaH[2]PO[4], 2.5 mmol/L
   CaCl[2]·2H[2]O, 1.2 mmol/L MgCl[2], 20 mmol/L sodium lactate, 16 mmol/L
   KCl, and 10 mmol/L 2-deoxyglucose [pH 6.2]) and placed in a hypoxia
   incubator (1% O[2], 94% N[2], 5% CO[2]). Reoxygenation was achieved by
   replacing the hypoxic buffer with fresh culture medium and incubating
   the cells in a standard tissue culture incubator for the indicated time
   [[86]25]. The cells and their supernatants were harvested separately
   after reoxygenation. Further adenosine supplementation is 10 μmol/L,
   and lentivirus transduction was performed for DNMT overexpression in
   vitro.

2.8. Flow cytometry

   Isolated cardiomyocytes (10^7 cells) were trypsinized for 5 min and
   then resuspended and stained using an apoptosis detection kit (BD,
   Catalog No. 559763). The samples were incubated for 60 min at 4 °C in
   the dark, and washed three times with a flow cytometry staining buffer.
   Apoptotic cells were detected using a BD LSRFortessa Cell Analyzer, and
   data were analyzed with FlowJo software.

2.9. Western blot

   Protein lysates were prepared from isolated cardiomyocytes or from the
   infarction border zone tissues isolated from mice after I/R surgery
   according to previously established protocols. Samples containing an
   equal amount of protein were separated by SDS-polyacrylamide gel
   electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes
   (Millipore). The membranes were blocked with 5% bovine serum albumin
   (BSA) in Tris-buffered saline for 1 h prior to overnight incubation
   with the respective primary antibodies at 4 °C. Levels of target
   proteins were analyzed with β-actin (1:5000) as the loading control.
   The detailed information of antibodies are described in
   [87]Supplementary Table 2.

2.10. Real-time quantitative polymerase chain reaction (qPCR)

   Total RNA was isolated from myocardial tissues or cardiomyocytes, and
   any contaminating genomic DNA was removed by DNase digestion.
   Complementary DNA (cDNA) was synthesized from 1 μg total RNA at 37 °C
   for 60 min in a 20-μL reaction system (Superscript III). The data were
   analyzed for the target genes. The primers used for qPCR are provided
   in the [88]Supplemental Table 3.

2.11. Methylated DNA immunoprecipitation (MeDIP)–qRT–PCR and bisulfite PCR
for IGF-1 promoter

   MeDIP–qRT–PCR was carried out with a Magnetic Methylated DNA
   Immunoprecipitation kit (Diagenode, Denville, NJ, USA). The MeDIP DNA
   was used for qPCR. Bisulfite PCR was performed to detect IGF-1 promoter
   methylation. Details were performed as described in the Appendix
   Supplementary Method. The primers used for BSP are provided in the
   [89]Supplemental Table 3.

2.12. Lentivirus/adenovirus production and transduction

   The full-length sequences of the DNMT1 gene were produced by PCR. The
   set of primers were as follows (forward primer:
   GCGAATTCGAAGTATACCTCGAGGCCACCATGGCTGCCAAACGGAGACC; reverse primer:
   CATGGTCTTTGTAGTCCATGGATCCGTCCTTGGTAGCAGCCTCCTCTTT). The products of the
   DNMT1 gene were amplified, purified, digested and ligated into the
   respective sites in the PGMLV-6395 vector (Genomeditech, China). 10 μg
   DNMT1 over-expression vectors were mixed with 10 μl Lenti-HG mix and
   60 μl HG transgene reagent (Genomeditech, China) for the lentivirus
   package, and transfection of 293 T cells was performed with
   Lipofectamine 2000 reagent (Invitrogen, Thermo Fisher Scientific, USA).
   100X Enhancing buffer (Genomeditech, China) was used for better
   transfection efficiency after 12 h incubation. The virus-containing
   supernatants were collected after 48 h transfection, and then
   transformed with 0.22 μm cellulose acetate filters (Merck Millipore,
   USA) for ultracentrifugation. For transductions, the respective
   Lentivirus (1 × 10^6 PFU/mL in culture medium) was incubated with
   cardiomyocytes (10^5 cells/mL) for 48 h, and the transduction
   efficiency was analyzed by Western blot.

   As for adenovirus, the same over-expression sequences were used as the
   Lentivirus production, and adenovirus was established with routine
   protocols. In short, the products of the DNMT1 gene were cloned into
   the vector. The vectors were then packaged into a recombinant
   adenovirus using U293 cells (Invitrogen, Carlsbad, USA) according to
   the manufacturer's instructions. In control group, the mice were
   injected with ADV-NC (1 × 10^10 infectious units/mouse) by tail vein.
   In DNMT1 overexpression group, the mice were infected with
   ADV-DNMT1(1 × 10^9 infectious units/mouse) by tail vein.

2.13. Immunostaining and immunofluorescence analyses

   For histological analyses, myocardial tissues were serially
   cryosectioned at a 5-μm thickness and mounted onto slides. The
   cryosections were fixed in 4% paraformaldehyde for 30 min and then
   rinsed in PBS. Plasma membranes were permeabilized with Triton X-100
   (Sigma, St. Louis, USA). Nonspecific binding sites were blocked by
   incubation with 5% BSA (Sigma, St. Louis, USA) in PBS for 1 h. The
   slides were then incubated with primary antibodies overnight at 4 °C.
   Then, the slides were rinsed and incubated with secondary antibodies at
   room temperature for 2 h. The stained slides were photographed with an
   Olympus IX71 or Zeiss Pascal confocal microscope.

2.14. Detection of glucose metabolism

   Quantitative measurements of glucose metabolism were performed with
   commercial kits according to the manufacturer's instructions. The kit
   information is described in [90]Supplementary Table 1.

   Preparation of fluorine-18 (18 F)-labeled positron emission tomography
   (PET) probes and micro-PET/computed tomography (CT) imaging.

   Standard 18 F-labeled fluorodeoxyglucose (FDG) and 4 (R,
   S)-[18 F]fluoro-6-thia-heptadecanoic acid (FTHA) probes were prepared
   on the same day of examination by the Department of Nuclear Medicine,
   Fudan University Shanghai Cancer Center. Each probe was prepared
   according to a standard protocol, and quality control analysis was
   conducted to ensure that they met the radiopharmaceutical requirements
   as previously reported [[91]26]. Micro-PET/CT scanning and image
   analysis were performed with an Inveon micro-PET/CT system (Inveon
   Research Workplace, Siemens Medical Solution, California, USA). Each
   mouse was intraperitoneally injected with approximately 200 μCi of the
   18 F-labeled probe. Then, 1 h after injection, a 20-min static scan was
   acquired for each mouse while under isoflurane anesthesia. Mice were
   fasted overnight prior to injection and were kept under isoflurane
   anesthesia during the 1-h period prior to imaging. Images were
   reconstructed and standardized uptake values were quantified from
   regions of interest (ROIs) using Inveon Research Workplace after
   recalibration according to the percentage of the injected dose per gram
   of tissue (ID/g %).

2.15. S-adenosylhomocysteine hydrolase (SAHH) activity assay

   SAHH activity was assessed using the adenosylhomocysteinase activity
   fluorometric assay kit (Abcam, ab197002, USA) per the manufacturer's
   instructions. SAHH catalyze hydrolysis of SAH to homocysteine and
   adenosine.

2.16. Statistical analysis

   Data were presented as the mean ± standard error of the mean (SEM).
   Significance between comparisons was determined by one-way analysis of
   variance (ANOVA) followed by Tukey's post-hoc test or a two-tailed
   Student's t-test. A p-value<0.05 was considered statistically
   significant.

3. Results

3.1. I/R resulted in the upregulation of ADK in the acute period

   RNA sequencing was performed to detect the change of ADK in the setting
   of I/R injury. Compared with WT group, WT mice with I/R injury had
   multiple changed candidate genes in ADK related pathways, while ADK was
   up-regulated in the setting of I/R ([92]Fig. 1A and B, [93]Fig. S9E).
   Further analysis to ADK related pathways showed that ADK, DNMT1 and
   metabolism related genes were significantly changed in the setting of
   I/R ([94]Fig. 1C). Moreover, correlation analysis showed that ADK had
   strong negative association with IGF-1, and middle-level negative
   association with DNA methylation related enzymes ([95]Fig. 1D).
   Pathways and bioinformatic analysis using datasets from three databases
   was also performed to identify potential myocardial candidate genes
   with altered expression in response to I/R. I/R resulted in the altered
   expression of metabolic genes ([96]Fig. S9A). Further analysis revealed
   that ADK was significantly increased in the acute period of I/R (less
   than 24-h) (Supplementary S9C, D), while decreased in the chronic
   period of I/R ([97]Fig. S9B). Overall, above analysis indicated the
   association between ADK and other metabolic genes in myocardial I/R.

Fig. 1.

   [98]Fig. 1
   [99]Open in a new tab

   ADK was identified as a potential candidate gene involved in the
   metabolic changes during I/R. (A) Pathways with potential difference in
   the setting of I/R based on RNA-sequencing, the red column marks
   pathways involved with ADK. (n = 3 per group). (B) Potential candidate
   genes with changed expression in the setting of I/R based on
   RNA-sequencing, the bold black arrow marks the position of ADK. (n = 3
   per group). Log2 fold change in X axis legend means the expression
   changes of target genes. Log10 adjust P value means the difference of
   target genes between groups.(C) Comparison of target genes between WT
   and WT + I/R group in the setting of 3-h I/R injury, Group A, wild-type
   mice; Group B, wild-type mice with I/R surgery, other details were
   demonstrated in the [100]Supplementary Fig. 9. (n = 3 per group). (D)
   Logistics analysis of ADK to other genes. Red color represents positive
   association. Blue color represents negative association. Abbreviations
   in C and D represents the candidate genes with potential difference.
   (For interpretation of the references to color in this figure legend,