Altered acylcarnitine metabolism and inflexible mitochondrial fuel utilization characterize the loss of neonatal myocardial regeneration capacity

Myocardial regeneration capacity declines during the first week after birth, and this decline is linked to adaptation to oxidative metabolism. Utilizing this regenerative window, we characterized the metabolic changes in myocardial injury in 1-day-old regeneration-competent and 7-day-old regeneration-compromised mice. The mice were either sham-operated or received left anterior descending coronary artery ligation to induce myocardial infarction (MI) and acute ischemic heart failure. Myocardial samples were collected 21 days after operations for metabolomic, transcriptomic and proteomic analyses. Phenotypic characterizations were carried out using echocardiography, histology and mitochondrial structural and functional assessments. In both groups, MI induced an early decline in cardiac function that persisted in the regeneration-compromised mice over time. By integrating the findings from metabolomic, transcriptomic and proteomic examinations, we linked regeneration failure to the accumulation of long-chain acylcarnitines and insufficient metabolic capacity for fatty acid beta-oxidation. Decreased expression of the redox-sensitive mitochondrial Slc25a20 carnitine-acylcarnitine translocase together with a decreased reduced:oxidized glutathione ratio in the myocardium in the regeneration-compromised mice pointed to a defect in the redox-sensitive acylcarnitine transport to the mitochondrial matrix. Rather than a forced shift from the preferred adult myocardial oxidative fuel source, our results suggest the facilitation of mitochondrial fatty acid transport and improvement of the beta-oxidation pathway as a means to overcome the metabolic barrier for repair and regeneration in adult mammals after MI and heart failure.


Animals
One-(P1) and seven-day (P7) old male Hsd:ICR(CD-1) mice were divided in four groups as follows: 1) 1-day-old sham operated (1d-SHAM+21d), 2) MI-induced 1-day old (1d-MI+21d), 3) 7-day-old sham-operated (7d-SHAM+21d), and MI-induced 7-day-old (7d-MI+21d). These animals were subjected to a follow up period of 21 days. A complementary study was made where the follow-up period was shortened to three days (1d-SHAM+3, 1d-MI+3, 7d-SHAM+3, 7d-MI+3). To carry out the necessary measurements from the small-sized hearts, a total of 193 animals were used. All animal experiments were carried out according to the European Community guidelines for use of experimental animals and approved by the Finnish National Animal Experiment Board (permission numbers ESAVI/8054/04. 10.07/2016, ESAVI/31851/2019). The animal model of the study was chosen based on the pioneering work of Porrello et al. 1 , who showed the different regenerative capacities between the P1 and P7 mice. Moreover, the chosen animal strain has been characterized as having large litters, excellent maternal characteristics and infrequent neonatal cannibalism (https://www.envigo.com/model/hsd-icr-cd-1).

Operations
The myocardial infarction model has been previously reported in detail by Mahmoud et al. 2 and Blom et al. 3 . In brief, one-and seven-day old male mice pups were removed from their maternal cage and were placed on a piece of sterile gauze lying on wet ice until slowing of breathing reflex and unresponsiveness for tail pinch was evident (from 2 to 3 min). After skin incision, thoracotomy was made through the 4th intercostal space with the aid of small scissors and forceps. Next, the left anterior descending coronary artery (LAD) was ligated with 10-0 prolene monofilament. Ischemia was visually confirmed as the myocardium distal from the suture became paler. The thoracal incision was sutured with 8-0 prolene, and the pups were carefully wiped clean and placed on a +37 °C heating pad until they regained consciousness. The pups were then moved into a heated chamber (+37 °C) holding their littermates with abundant bedding from their mother's cage. After the operated littermates were adequately awake, they were returned to their mother for the duration of the follow-up period (21 days or 3 days).

Echocardiography
Transthoracic echocardiography was performed on post-operative days 1, 3, 7 and 21 by using a Vevo 2100 ultrasound with MS700 (bandwidth 30-70 MHz) linear array transducer (VisualSonics Inc., Toronto, Canada). The images were stored and analyzed using Vevo LAB 1.7.1 software (VisualSonics Inc., Toronto, Canada). The chamber diameter was measured from the 2D-guided parasternal short axis M-mode images from which ejection fraction (EF) and fractional shortening (FS) were calculated. Of all animals used in this study, a total of 109 were subjected to echocardiography (n=8-14).

Sample collection
At the end of the follow up period (3 days or 21 days), the animals were weighed, anesthetized with isoflurane and decapitated. The hearts were exposed, collected and weighed. The samples for conventional histology were collected as cross-sections of the whole heart distal from the LAD ligature. The samples for TEM, respirometry, RNASeq, miRNASeq, proteomics and metabolomics, in turn, were collected from the non-infarcted area of the left ventricle (border and remote zones). As the neonatal mouse heart is small, separate series for echocardiography and histology (n=6-10), electron microscopy (n=3-4), cardiac volumetry (n=4), mitochondrial highresolution respirometry and RNA sequencing (n=4), proteomic analysis (n=4) and metabolomic analysis (n=6) were carried out to ensure that enough representative sample was available for each analysis. Heart samples for histology were fixed in formalin, embedded in paraffin and processed to hematoxylin and eosin (HE)-stained sections with routine protocols. Samples (~1 mm 3 ) for transmission electron microscopy were collected from the non-infarcted area (remote zone) of the left ventricular wall, fixed in 2.5% glutaraldehyde in 100 mM sodium-phosphate buffer and subsequently embedded in resin by using the normal protocols. Samples for transcriptomics, proteomics and metabolomics were frozen in liquid nitrogen and stored at −80 °C until analysis. Following the samples collection, the gender of the pups was verified by gonadal examination.

Isolation of mitochondria
Neonatal hearts were pooled (from 3-4 mice) to obtain approximately 20 mg tissue. The heart tissue was minced with scissors in trypsin-EDTA phosphate-buffered saline solution for 15 min on ice. The trypsin digestion was stopped by addition of fetal bovine serum (10%). The tissue pieces were collected by centrifugation (200g, 1 min) and the buffer replaced to 225 mM mannitol, 75 sucrose, 1 mM EGTA, 0.1 mg/ml fatty-acid free BSA, and 10 mM Tris-Cl pH 7.4 (at 0 o C). Homogenization was performed with Wheaton 5 ml glass-teflon Potter-Elvehjem. Nuclei and unhomogenized material were pellet by a low-speed centrifugation (800g, 10 min at 4 o C). The supernatant was subjected to 7800g centrifugation (5 min at 4 o C) to obtain the mitochondrial fraction. The mitochondria were washed once with homogenization buffer lacking BSA and repelleted (7800g, 5 min).

Measurement of reactive oxygen species (ROS) production
Amplex UltraRed-peroxidase assay was employed to measure ROS production from isolated heart mitochondria. The measurements were performed at 37 o C in 96-well microplate format with a reaction volume of 0.3 ml and 12.5 µg mitochondrial protein. Mir05 buffer (Mir06 lacking catalase) served as the assay media. The Amplex UltraRed and horseradish peroxidase concentrations were 10 µM and 1 U/ml, respectively. Superoxide dismutase (5 U/ml) were included to ensure immediate dismutation of superoxide to H2O2. Catalase and thioredoxin system in samples were inhibited by 100 µM 3-amino-1,2,4-triazole and 0.5 µM auranofin. Background was determined in the presence of sample but in the absence of exogenous substrates required for the mitochondrial respiratory activity. Parallel reactions were spiked with a 50 pmol bolus of H2O2 (ε240 nm=43.6 M -1 cm -1 ) to calibrate and convert the fluorescence signal (530 nm excitation, 590 nm emission) to the amount of H2O2 produced. The incubations were kept short (< 6 min) to avoid fluctuation in oxygen concentration.

Whole-mount immunostaining and volumetric microscopy
For whole-mount immunostaining, the animals were perfused with saline to remove excess blood and then perfusion-fixed with 10% paraformaldehyde. Hearts were removed and kept in 10% paraformaldehyde solution for 24 hours until stored in methanol. Immunolabeling and tissue clearing process was carried out with the iDISCO method out as described previously in detail by Renier et al. 8 . Cardiac Troponin C (ab30807, dilution 1:500, Abcam, Cambridge, United Kingdom) antibody was used for immunolabeling of cardiomyocytes. Images were captured by using a Bioptonics OPT Scanner 3001 M (LifeArc, London, United Kingdom) and 3D reconstructions from the data were made using NRecon software (LifeArc). Segmentation and analysis of MI size was made with Imaris software (Bitplane AG, Zürich, Switzerland). Briefly, MI size was calculated as a proportion of viable myocardium recognized with Troponin C antibody compared to whole heart quantified with visible light channel.

General histology and mitochondrial ultrastructure
Histology of the myocardial samples was evaluated in a blinded fashion with conventional light microscopy. In brief, histopathology of the arteries and cardiac muscle tissue was examined. The severity of the observed lesions in the myocardium, including tissue necrosis, fibrosis, and inflammatory cell infiltration, was graded with numerical values denoting to the degree of damage at the whole tissue level. The following system of severity scoring was used: 0, no abnormalities detected; 1, minimal; 2, mild; 3, moderate; 4, marked; or 5, severe 9,10 . For the transmission electron microscopy, thin-sectioned samples were viewed and photographed with a Jeol JEM-1400 transmission electron microscope (Jeol Ltd., Tokyo, Japan). Mitochondrial number, volume fraction and trans-sectional area were analyzed with ImageJ software 11 (v1.47). For the number of cristae per μm, the ImageJ line tool was used to draw a line across the stack of cristae, perpendicular to their orientation. The number of cristae crossing the line was counted and expressed in relation to the line length 12 .

Global metabolomics analysis
Metabolomics analysis was carried out as previously described 4 . Hearts from the SHAM and MI-treated groups collected after 21-day follow-up period were shipped to Metabolon, Inc (Durham, NC, USA) where they were stored at −80 °C until analysis. Sample preparation, quality control, UPLC-MS/MS and data analysis were all carried out by Metabolon, Inc. Samples were prepared using the automated MicroLab STAR system with standards for QC (Hamilton, Reno, NV, USA). Proteins were precipitated with methanol followed by shaking and centrifugation. The extract was divided into following fractions: two for analysis by two separate reverse-phase (RP)/UPLC-MS/MS methods with positive ion mode electrospray ionization (ESI), one for analysis by RP/UPLC-MS/MS with negative ion mode ESI, and one for analysis by HILIC/UPLC-MS/MS with negative ion mode ESI. All methods utilized a Waters ACQUITY ultra-performance liquid chromatography (UPLC) with Waters (UPLC BEH C18-2.1×100 mm, 1.7 μm and BEH Amide 2.1x150 mm, 1.7 µm) columns and a Thermo Scientific Q-Exactive mass spectrometer interfaced with a heated electrospray ionization (HESI-II) source and Orbitrap mass analyzer operated at 35,000 mass resolution 13 . Metabolites were identified by comparison to library entries of purified standards or recurrent unknown entities based on authenticated standards that contains the retention time/index (RI), mass to charge ratio (m/z), and chromatographic data (including MS/MS spectral data) on all molecules present in the library 14 . Peaks were quantified using areaunder-the-curve. The data underwent random forest analysis, and two-way Anova tests followed by Welch's two-sample t-test according to the statistics workflow of Metabolon Inc. were used to identify molecules that differed significantly between experimental groups (p≤0.05).

Transcriptomics analysis
The cardiac samples for whole transcriptome and miRNA analysis were removed from −80 °C and processed with the Precellys 24 homogenizer with ceramic beads (Bertin Technologies, Montigny-le-Bretonneux, France). RNA was isolated with TRIzol reagent, followed by DNAse I treatment. Agilent Bioanalyzer RNA pico chip (Agilent Technologies Inc., Santa Clara, CA, USA) was used to evaluate the integrity of RNA and Qubit RNA-kit (Thermo Fisher Scientific Inc., Waltham, MA, USA) to quantitate RNA.

mRNA analysis
Library preparation and NGS of samples collected after 21 days of follow-up were made by the Institute for Molecular Medicine Finland (FIMM), University of Helsinki, Finland. rRNA depletion was done according to RiboZero magnetic Gold protocol (Illumina Inc.) and the library preparation was done by using ScriptSeqV2 library preparation kit Library Prep Kit (Illumina Inc.). The libraries were quantified and pooled for sequencing based on Agilent Bioanalyzer HiSens kit. Sequencing was performed on Illumina HiSeq 2000 instrument (2x100 cycles), according to the manufacturer instructions. Library preparation and next generation sequencing (NGS) of samples collected at 3 days follow-up were made with TruSeq stranded total RNA Library Prep Kit and NextSeq550 instrument (Illumina Inc., San Diego, CA, USA) by QIAGEN Sciences Inc. (Germantown, MD, USA, http://www.qiagen.com). The starting material (500 ng) of total RNA was rRNA depleted using biotinylated, target-specific oligos combined with Ribo-Zero rRNA removal beads. The isolated RNA was subsequently fragmented using enzymatic fragmentation. Then first and second strand synthesis were performed, and the double stranded cDNA was purified (AMPure XP, Beckman Coulter Inc., Brea, CA, USA). The cDNA was end-repaired, 3'adenylated and Illumina sequencing adaptors ligated onto the fragments ends, and the library was purified (AMPure XP). The RNA stranded libraries were pre-amplified with PCR and purified (AMPure XP). The libraries size distribution was validated, and quality inspected on a Bioanalyzer 2100 or BioAnalyzer 4200 TapeStation (Agilent Technologies). High quality libraries were pooled in equimolar concentrations based on the Bioanalyzer Smear Analysis tool (Agilent Technologies). The library pool(s) were quantified using qPCR and optimal concentration of the library pool used to generate the clusters on the surface of a flow cell before sequencing on a NextSeq550 instrument (2x75 cycles), according to the manufacturer instructions (Illumina Inc.).
The bioinformatic analysis of NGS data was carried out with "Tuxedo" software consisting of Bowtie2 (v. 2.2.2), Tophat (v2.0.11), Cufflinks (v2.2.1; Qiagen) including alignment of reads to the mouse reference genome (GRCm38; Ensembl_81), feature summation, and normalization of expression estimates. To allow for the comparison of differential expression estimates between samples, the normalized expression values for each transcript were calculated as Fragments per Kilobase per Million Mapped Reads (FPKM) with Cuffdiff2. The expression profiles of different treatment groups were compared and the transcript ratios between the average FPKMs were calculated. Transcripts showing a LogFC ≥1 and a False Discovery Rate (FDR) <0.05 were classified as differentially expressed.

miRNA analysis
Library preparation and NGS of samples collected at the 21-day follow-up were made by the Institute for Molecular Medicine Finland (FIMM), University of Helsinki, Finland. The library preparation was done using the TruSeq Small RNA Sample Prep kit (Illumina Inc., San Diego, CA, USA). The small RNA fractions of the libraries were enriched by using Caliper XT system (PerkinElmer, Waltham, MA, USA) Caliper XT 300 DNA chip was used and 147bp fragments were selected for the enrichment. The size selected libraries were quantified for sequencing based on Agilent Bioanalyzer HiSens kit. Sequencing was performed on Illumina HiSeq 2000 instrument (1x100 cycles), according to the manufacturer instructions (Illumina Inc., San Diego, CA, USA) Annotation of the obtained sequences was performed by using the reference annotation (GRCm38; miRbase_20). Library preparation and NGS of samples collected at 3-days follow-up were made by QIAGEN Sciences Inc. using QIAseq miRNA Library Kit (QIAGEN Sciences Inc.) and NextSeq500 instrument (Illumina Inc.) by QIAGEN Sciences Inc. The library preparation was done using the QIAseq miRNA Library Kit (Qiagen). A total of 100ng total RNA was converted into miRNA NGS libraries. Adapters containing UMIs were ligated to the RNA. Then RNA was converted to cDNA. The cDNA was amplified using PCR (16 cycles) and during the PCR, indices were added. After PCR, the samples were purified. Library preparation QC was performed using either Bioanalyzer 2100 (Agilent) or TapeStation 4200 (Agilent). Based on quality of the inserts and the concentration measurements, the libraries were pooled in equimolar ratios. The library pool(s) were quantified using qPCR. The library pool(s) were then sequenced on a NextSeq500 sequencing instrument according to the manufacturer instructions.
The data was UMI corrected and mapped (Bowtie2 v.2.2.2.). Annotation of the obtained sequences was performed by using the reference annotation (GRCm38; miRbase_20). Differential expression analysis was performed using the EdgeR statistical software package (Bioconductor, http://bioconductor.org/). For normalization, the trimmed mean of M-values method based on LogF and absolute gene-wise changes in expression levels between samples (TMM normalization) were used. Transcripts showing a LogFC ≥1 and a False Discovery Rate (FDR) <0.05 were classified as differentially expressed.

Proteomics analyses
Hearts from the 1d-MI+21 and 7d-MI+21 and their corresponding controls were homogenized at 4 °C in 7M urea, 2M thiourea, 4% CHAPS using a Precellys® 24 homogenizer (Bertin Technologies) with ceramic beads. Samples were centrifuged for 1 minute at 16.000 g. 10 µg of total protein lysates were processed using modified FASP protocol as described 15 . The samples were analyzed by DIA-HDMS E (data independent acquisition high-definition ionmobility enabled tandem mass spectrometry), as described 4 .
Relative quantification between samples using precursor ion intensities was achieved with Progenesis QI for ProteomicsTM Informatics for Proteomics software (Nonlinear Dynamics/ Waters Corporation, Milford, MA, USA) and ProteinLynx Global Server (PLGS v3.0) 16 . Labelfree protein quantitation method was applied for post processing data analysis using precursor ion intensity data and standardized expression profiles. Database searches were carried out against Mus musculus (release 2018_16993 entries) UniProtKB/SwissProt, reviewed database (with primary digest reagent set for trypsin, 2 missed cleavages allowed, cysteine residue carbamidomethylation as fixed modification and as variable modifications, deamidation of asparagine and glutamine residues and oxidation of methionine). Differentially expressed proteins (DEP) were identified based on the number of unique peptides used for label-free quantitation (≥2), at the FDR<0.01 and the fold change (FC) ratio from averaged, normalized protein intensities |FC≥1.5|, and p≤ 0.05 by ANOVA in all comparisons.
The lists of up/down regulated protein changes with their corresponding, unique UniProtKB-database identifiers served as inputs into Ingenuity Pathways (IPA, Ingenuity systems, Redwood City, CA, USA; www.ingenuity.com), with a focus on Canonical pathways, and Disease and Function annotation.

Ingenuity Pathway Analysis (IPA)
Pathway analyses were carried out as described earlier [17][18][19] . Briefly, pre-filtered lists of differentially expressed genes were imported into and analyzed using the IPA software (Qiagen). Analysis of miRNA utilized the IPA miRNA filter and only miRNAs' and their target genes' expressions at opposite directions were considered meaningful as miRNA-dependent regulation. Exploratory analysis and hypothesis-generating network building included genes with differential |log2| expression fold changes greater than 1 between the 1d-MI+21 and 7d-MI+21 groups and one-sided t-test p values less than 0.05. The top-scoring, differentially expressed genes identified were then evaluated using R Studio (2021.09.1+372 "Ghost Orchid" http://www.rstudio.com/) Benjamini-Hochberg FDR-controlling method between groups. For visualization, the String network was imported into Cytoscape 3.9.0 20 . Visualization used Nested network style and yFiles Organic layout under academic licensing 21 (http://www.yworks.com). Final graphical layout design of the networks was carried out in Inkscape (1.0.1; https://inkscape.org).

Quantitative RT-PCR (qRT-PCR)
The hearts from 1-day-old and 7-day-old mice were snap-frozen in liquid nitrogen and processed with the Precellys 24 homogenizer with ceramic beads (Bertin Technologies, Montignyle-Bretonneux, France). RNA from the homogenized samples was isolated with TRIzol as described 22 , followed by DNAse I treatment. Agilent Bioanalyzer RNA pico chip (Agilent Technologies, Santa Clara, CA, USA) was used to evaluate the integrity of RNA and Qubit RNAkit (Life Technologies, Carlsbad, CA, USA) to quantitate RNA.