Affymetrix gene expression data from IST Online database (ist.medisapiens.​com; Medisapiens Ltd.) were analyzed to characterize RTK expression in samples representing healthy heart (n = 62), acute myocardial infarction (n = 12) or ischemic cardiomyopathy (n = 63). Out of the 55 RTKs listed by HUGO Gene Nomenclature Committee, data were available for 49 genes in acute myocardial infarction samples and for 52 genes in ischemic cardiomyopathy samples (Additional file 1). Data were normalized by array-generation-based gene centering method [18] and log2-transformed. Expression levels of RTKs demonstrating statistically significant differences in two-group comparisons were visualized as box plots and heatmaps using Pretty Heatmaps package (pheatmap) [19] in RStudio [20].

Animal experiments were approved by the Laboratory Animal Care and Use Committee of the State Provincial Office of Southern Finland (license number: ESAVI/1167/04.10.03/2011). The landrace pig myocardial samples (n = 7) were a kind gift from Drs. Christoffer Stark and Timo Savunen. Experimental procedure has been described in detail earlier [21]. Pigs weighed 29–43 kg. Myocardial ischemia-reperfusion injury was produced by exposing anesthetized pigs (n = 4) to cardiopulmonary bypass with aortic cross-clamping and cardioplegic arrest for 60 min, causing global myocardial ischemia. A pediatric membrane oxygenator (Dideco 905 Eos, Dideco) was used for the bypass. Procedures were performed in the laboratory of Research Center of Applied and Preventive Cardiovascular Medicine, University of Turku, Turku, Finland. For pre-anesthesia, an intramuscular injection of 100 mg xylazin (Rompun vet, Bayer Animal Health GmbH) and 25 mg midazolam (Midazolam Hameln, Hameln pharmaceuticals GmbH) was used. For anesthesia, 20 mg boluses of propofol (PropofolLipuro, B. Braun Melsungen AG) and 150 μg phentanyl (Fentanyl-Hameln, Hameln pharmaceuticals GmbH) were administered via a cannulated ear vein, and pigs were intubated and connected to a respirator (Dräger Oxylog 3000, Drägerwerk AG), the respiratory rate set to 18–22 times/min with a tidal volume of 8–10 ml/kg using 40% oxygen. Continuous infusion of propofol 15–30 mg/kg/h, phentanyl 1.5 μg/kg/h and midazolam 100 μg/kg/h was used to maintain anesthesia. A right sided thoracotomy was performed and the ascending aorta and right atrium cannulated for the bypass. 500 ml of cold (10 °C) Modified St Thomas Hospital No II cardioplegia was used to protect the hearts during the bypass, administered via a cannula to the aortic root at the time of cross-clamping and 30 min later. Antibiotic prophylaxis (Cefuroxime 750 mg, Orion Pharma) was given preoperatively and then every 8 h. 10,000 IU of heparin (Heparin, LEO Pharma) was administered as a bolus before cannulation of the heart and this was repeated every 30 min during extracorporeal circulation. 14,000 IU of protamine sulphate (Protamin, LEO Pharma) was used to neutralize the heparin. For thrombosis prophylaxis, 20 mg of enoxaparin (Klexane, Sanofi) was administered 1 and 12 h after the surgery. 100–150 mg of lidocaine (Lidocain, Orion Pharmaceuticals) and 150–225 mg of amiodarone (Cordarone, Sanofi) were used for rhythm disorders, and 5 mg boluses of ephedrine (Efedrin, Stragen Nordic) and noradrenaline infusion (80–160 μg/h) (Noradrenalin Hospira, Hospira) were used for post-operative hemodynamic support, when needed. For post-operative analgesia, 50 mg of bupivacaine (Bicain, Orion Pharmaceuticals) was infiltrated to the wound. For monitoring of adequate ventilation and perfusion, blood gases (i-STAT, Abbott Laboratories), invasive central venous pressure, ECG and oxygen saturation were followed throughout the procedure.

Pig myocardial samples (280 to 460 mg) were homogenized and analyzed for phosphorylation status of 49 RTKs using the Proteome Profiler Human Phospho-RTK Array Kit (R&D Systems). Five hundred μg of protein was analyzed per sample. Receptors included in the analysis are listed in Additional file 2 B.

Array blot images were quantified by densitometry with NIH ImageJ v1.50i software. Intensity values were normalized by dividing each dot’s intensity with the sum of intensities of the whole array, allowing comparison between different samples. Data were scaled to interval 0–1 by dividing all values with the highest value. Normalized values were visualized as a bargraph (mean (SD)) and as a heatmap with Pretty Heatmaps package [19]. Receptors with at least two quantifiable results in both sample groups were included. Receptors were clustered using maximum distance method.

HL-1 mouse atrial cardiomyocytes were a kind gift from Dr. Pasi Tavi (University of Eastern Finland). HL-1 cells were maintained in Claycomb medium (Sigma) supplemented with 10% FBS, 0.1 mM norepinephrine, 50 U/ml penicillin, 50 U/ml streptomycin, and 2 mM UltraGlutamine (Lonza). Culture plates were coated at 37 °C with a solution containing 0.02% gelatin and 10 μg/ml fibronectin. Seeding densities were 140,000 cells/6-well plate well, 100,000 cells/12-well plate well, and 5,000 cells/96-well plate well. H9c2 rat cardiomyoblasts were purchased from ATCC and maintained in DMEM with 1.5 g/l NaHCO₃, supplemented with 10% FBS, 50 U/ml penicillin, 50 U/ml streptomycin, and 2 mM UltraGlutamine. Seeding density was 100,000 cells/6-well plate well. Cells were routinely checked for mycoplasma infection using MycoAlert Mycoplasma Detection Kit (Lonza). For hypoxia-reoxygenation experiments, the cells were cultured in 1% O₂ in a hypoxic work station (InVivo₂, Ruskinn Technology Ltd.) for the indicated periods of time, and returned to normal cell incubator (21% O₂) for reoxygenation. In ROR1 ligand activation experiments, 200–400 ng/ml of recombinant human/mouse Wnt-5a (R&D Systems) was added to medium at the time of plating (for MTT assays) or 24 h after plating for 30–60 min (for Western analyses).

One day after plating, HL-1 cells were transfected with siRNAs (Qiagen) targeting ROR1 (siRNA #1, SI01404655; siRNA #2, SI01404662), or ROR2 (siRNA #1, SI01404683; siRNA #2, SI01404690) or with AllStars Negative Control siRNA at a concentration of 100 nM, using Lipofectamine 2000 (Invitrogen). Immediately prior to transfection, medium was changed to antibiotic- and norepinephrine-free Claycomb medium. Four to six hours after transfection, medium was replaced with antibiotic-free, norepinephrine-supplemented Claycomb medium.

RNA was extracted from pig myocardium samples using TRIsure reagent (Bioline). Samples were treated with 10 units of DNAse I (Roche). cDNA was synthesized with SensiFAST cDNA Synthesis Kit (Bioline), using 1 μg of total RNA/sample. Real-time RT-PCR was carried out using QuantStudio 12 K Flex Real-Time PCR System thermal cycler (Thermo Fisher Scientific). For PCR reactions, 5 μl of TaqMan Universal Master Mix II (Thermo Fisher Scientific) and 10 ng of template cDNA were used in a reaction volume of 10 μl. Primer concentrations were 0.3 μM and probe concentration 0.1 μM. Primers were acquired from Eurofins Genomics and probes from Universal Probe Library (Roche). GAPDH was used as the reference gene [8]. ROR1 was analyzed using the primers 5’-GCGGCTCGCAATATTCTC-3′ and 5’-GAAAGCCCAAGGTCTGAAATC-3′, and the probe #108. GAPDH was analyzed using the primers 5’-ACAGACAGCCGTGTGTTCC-3′ and 5’-ACCTTCACCATCGTGTCTCA-3′, and the probe #28.

Cells were lysed with lysis buffer [22] supplemented with Pierce Protease Inhibitor Mini Tablets (Thermo Fisher Scientific). Lysates were centrifuged for 15 min at 16,000 g and the supernatants were collected. Snap-frozen pig heart tissue samples were dissolved in ice-cold Lysis Buffer 17 supplied with the Proteome Profiler Human Phospho-RTK Array Kit. Samples were separated on 8–10% polyacrylamide gels. Protein amounts loaded on the gel were 20–35 μg for cell samples and 100 μg for pig tissue samples. Separated samples were transferred to nitrocellulose membrane which was blocked with 5% non-fat milk or bovine serum albumin in 10 mM Tris-HCl (pH 7.4), 150 mM NaCl and 0.05% Tween-20 (blocking solution) for 1 h at room temperature. Membranes were incubated with primary antibodies overnight at 4 °C in the blocking solution. The following primary antibodies and dilutions were used: anti-ROR1 (sc-83033 and sc-130386, Santa Cruz Biotechnology; 1:250 and 1:125, respectively), anti-actin (sc-1616, Santa Cruz Biotechnology; 1:1,000), anti-α-tubulin (sc-5546, Santa Cruz Biotechnology; 1:1,000), anti-phospho-Akt (#4060, Cell Signaling Technology; 1:1,000), anti-Akt (#9272, Cell Signaling Technology; 1:1,000), anti-phospho-p38 (#9211, Cell Signaling Technology; 1:500), anti-p38 (#9212, Cell Signaling Technology; 1:500), anti-phospho-tyrosine (4 g10, Upstate; 1:500), and anti-GAPDH (G8795, Sigma-Aldrich; 1:1,000). Incubation with secondary HRP-conjugated antibodies (Santa Cruz Biotechnology; 1:5,000) or IRDye secondary antibodies (LI-COR, 1:10,000) was carried out for 1 h at room temperature in the blocking solution. The immunosignals were visualized with WesternBright ECL HRP substrate reagent (Advansta) and imaged with ImageQuant LAS 4000 (GE Healthcare Life Sciences) or with Odyssey imaging system (LI-COR). Densitometric analysis of the Western signals was carried out with ImageJ.

For immunoprecipitation analyses, approximately 2 mg of total protein from HL-1 cell lysates was incubated overnight at 4 °C with 4 μg of mouse monoclonal anti-ROR1 antibody (sc-130386, Santa Cruz Biotechnology). Immunoprecipitated lysates were incubated with Protein G PLUS-Agarose beads (sc-2002, Santa Cruz Biotechnology) for one hour at 4 °C. After washing the beads three times with lysis buffer without protease inhibitors, samples from the supernatants were loaded onto 10% polyacrylamide gels for subsequent Western analysis.

CellTiter 96 Non-Radioactive Cell Proliferation Assay (MTT) (Promega) was used to measure viability of HL-1 cells. The assay was performed on 96-well plates with plating density of 5,000 cells/well. Before addition of the MTT dye solution, the culture medium was replaced by norepinephrine-free Claycomb medium (100 μl/well). Wells only containing the medium were used for background subtraction. Absorbances at 570 nm were detected with EnSight Multimode Plate Reader (PerkinElmer) and results were normalized to untreated control sample level.

Human and pig RTK protein sequences from UniProt database were aligned using EMBOSS Needle tool for global alignments and EMBOSS Water tool for local alignments by the European Bioinformatics Institute (http://​www.​ebi.​ac.​uk/​Tools/​psa/​) [23]. The longest pig sequences available were used.

For in silico transcriptomics analyses, Student’s t-test was used to assess the significance of differences between the expression levels of an RTK between sample groups. Correction for multiple testing was performed using false discovery rate (FDR) method [24]. Genes with significant differences in expression were selected for visualizations. Two-tailed Student’s t-test was used for testing the significance of differences in phosphorylation array data (for RTKs with at least three quantifiable results in both treatments) and in expression levels in Western analyses. For in vitro experiments, data are represented as box plots depicting median (black horizontal line), first and third quartile (box) and the range of the data (whiskers). Multiple group comparisons were performed with Kruskal-Wallis test with Dunn’s post-hoc test using FDR method for P-value adjustments. P-values < 0.05 were considered significant. Analyses were performed with RStudio.

Expression of human RTKs in healthy heart, heart with ischemic cardiomyopathy and heart with acute myocardial infarction was analyzed in silico using the IST Online database. Comparison of healthy samples to ischemic cardiopathy samples revealed a set of 14 receptors that were significantly differently expressed at the mRNA level (Fig. 1a). The most significantly up- and downregulated RTKs were ROR1 and EPHA2, respectively, when ischemic cardiomyopathy was compared to normal heart (Fig. 1b). Interestingly, another ROR family member, ROR2, was downregulated in the ischemic cardiomyopathy samples (Fig. 1a). When comparing healthy hearts with hearts representing acute myocardial infarction, significant downregulation of EGFR, ERBB2, ERBB3, as well as EPHA2, was discovered in the infarcted heart, ERBB2 demonstrating the greatest difference in expression (Fig. 1c and d). Statistical test values and differences in expression are shown in Additional file 1 and heatmaps generated from the data in Additional file 3.

To experimentally address RTK activation in the ischemic heart, an ischemia-reperfusion model in pig, a relevant model animal close to human as a large mammal, was used. Pigs were subjected to ischemia-reperfusion injury by cardiopulmonary bypass and reperfusion, followed by phospho-RTK array analysis of myocardial samples. Increased amount of troponin T in the plasma (Additional file 4 A) and thinned cardiomyocytes with condensed nuclei as demonstrated by histological analysis (Additional file 4 B) indicated ischemic injury. The phosphoarray analysis was carried out after approximately 30 h of reperfusion. While the observed phosphorylation-specific signal was relatively weak (possibly due to loss of phosphosignal during sample preparation and/or low cross-reactivity of the array antibodies between human and pig epitopes) (Additional file 2 A), the data indicated the presence of 23 activated RTKs in the heart. As expected, based on publications about RTK activity in the heart [8‐12], the array identified phosphorylated EGFR, INSR, VEGFR2, and ERBB2 in both ischemic (n = 4) and control samples (n = 3) (Fig. 2a). Phosphorylation of a group of receptors, that to our knowledge have not previously been shown to be active or regulated in ischemic heart, was also detected. This group included both ROR receptors ROR1 and ROR2; the EPH receptors EPHB2, EPHB3, EPHB6, and EPHA10; TYRO3; RYK; FGFR3 and ALK (Fig. 2a).

Densitometric quantitation of the data indicated highest activity for EGFR, ROR2, INSR, EPHB3, EPHB2, TEK, TYRO3, RYK, ROR1 and EPHB6 (phosphosignal exceeding the 75th percentile of mean phosphorylation levels of all RTKs in all samples). Phosphorylation of ROR1 was significantly downregulated (P = 0.022), and phosphorylation of VEGFR2 significantly upregulated (P = 0.043), in ischemia-reperfusion-injured hearts. Unsupervised clustering of the RTKs on a heatmap grouped highly active EGFR, INSR and ROR2 together. ROR1 and RYK, demonstrating downregulation of phosphorylation by ischemia-reperfusion, were grouped together in the other end of the heatmap (Fig. 2b).

As ROR1 was one of the most significantly regulated RTKs in the human heart expression analysis (Fig. 1b), and significantly downregulated in the ischemic pig heart (Fig. 2), it was selected for further validation. To confirm the presence of ROR1 in the heart in vivo, lysates from pig myocardium were analyzed by Western blotting with a polyclonal anti-ROR1 antibody. A band of the predicted size for full-length ROR1 protein (130 kDa) was detected in both control and ischemia-reperfusion-injured samples (Fig. 3a and b). The expression of ROR1 mRNA was confirmed by RT-PCR (Fig. 3c). In both analyses, ROR1 expression did not demonstrate statistically significant difference between control and ischemia-reperfusion samples.

Analyses of the samples available from the IST Online database (Fig. 1) and the pig in vivo model (Figs. 2 and 3) indicated the presence of ROR1 in the heart tissue after ischemia and reperfusion, but did not allow for temporal analyses of ROR1 expression or activity at different time points after hypoxia and reoxygenation. To address the ROR1 regulation in vitro in cardiomyocytes under conditions simulating ischemia-reperfusion, HL-1 mouse atrial cardiomyocytes and H9c2 rat cardiomyoblasts were exposed to hypoxia (1% O₂) in a hypoxic workstation for 1, 3 or 24 h. Cells were subsequently either returned to normoxia for 3 or 24 h (hypoxia-reoxygenation), or directly lysed for expression analysis (hypoxia alone). ROR1 protein expression was analyzed by Western blotting.

In HL-1 cells, the total ROR1 protein level was significantly reduced by 1 h treatment in hypoxia (P = 0.014), remained low 3 h after reoxygenation, and was recovered with a high variation in expression after 24 h of reoxygenation (Fig. 4a and b). Treatment for 3 h in hypoxia also initially slightly reduced ROR1 protein expression, but the expression recovered already 3 h after reoxygenation (Fig. 4a and b). The longest time point of hypoxia analyzed, 24 h, did not result in changes in ROR1 protein levels (Fig. 4a and b). In H9c2 cells, downregulation of ROR1 was similarly observed in response to treatment for 1 h in hypoxia (P = 0.002), followed by a trend of expression returning back to the level of control samples after reoxygenation (Fig. 4c and d). In contrast to ROR1, ROR2 expression was not significantly regulated by hypoxia-reoxygenation (Additional file 5 A and B).

To address whether ROR1 phosphorylation was regulated in response to hypoxia, ROR1 was immunoprecipitated from HL-1 cell lysates followed by anti-phosphotyrosine Western analysis. The samples were collected at two time points: after one hour of hypoxia, a time point that demonstrated significant downregulation of total ROR1 protein level in both HL-1 and H9c2 cells (Fig. 4), and after one hour of hypoxia followed by 24 h of reoxygenation, a schedule chosen to reflect the ischemia-reperfusion treatment used for the in vivo analysis (Fig. 2). ROR1 phosphorylation was downregulated in response to both treatments (Additional file 6), suggesting that hypoxia alone is sufficient for the observed effect.

The regulation of ROR1 expression in both ischemia-reperfusion and hypoxia-reoxygenation models indicated functional relevance for ROR1 activity in the hypoxic heart. To address the contribution of ROR1 for the cardiomyocyte viability, expression of the receptor was knocked down in HL-1 cells by RNA interference, and the relative amount of viable cells was measured by MTT assay. Two ROR1-targeting siRNAs significantly enhanced the cellular viability both in normoxia (P = 0.004 and P < 0.001 for siRNAs #1 and #2, respectively; n = 18) and after 24 h of hypoxia followed by 24 h of reoxygenation (P = 0.003 and P = 0.001 for siRNAs #1 and #2, respectively; n = 18), when compared to negative control siRNAs (Fig. 5a and Additional file 5 E). Consistent with the expression data, knockdown of ROR2 expression did not significantly affect HL-1 cell viability (Additional file 5 C-E), suggesting that the effect was specific for ROR1 within the ROR subfamily of RTKs.

To study the downstream signaling effects of ROR1 knockdown, phosphorylation status of Akt and p38 was studied by Western blotting, as these pathways have been shown to be involved in ROR1 signaling [25, 26]. Phosphorylation of both Akt and p38 was decreased in HL-1 cells 24 h after siRNA transfection (Fig. 5b).

Wnt-5a is an activating ligand of ROR1 [27, 28] in addition to other receptor such as ROR2 [27‐29], RYK [30] and Frizzled receptors [31]. Stimulation of HL-1 cells for 96 h with 200 or 400 ng/ml Wnt-5a significantly reduced the viability of the cells in both normoxia (P = 0.012 and P = 0.015, for 200 and 400 ng/ml, respectively; n = 9) and after 48-h hypoxia followed by 24 h of reoxygenation (P = 0.006 and P < 0.001, for 200 and 400 ng/ml, respectively; n = 9), as compared to non-treated controls (Fig. 5c). In accordance with the decrease in the level of phosphorylated Akt and p38 by ROR1-targeted RNA interference, phosphorylation of both Akt and p38 was enhanced in response to treatment with 400 ng/ml of Wnt-5a for 30 min (Fig. 5d).

Taken together, these results indicate that ROR1 expression and activity suppress cardiomyocyte viability in vitro, and that the pathways involved include Akt and p38.