Effects of Levosimendan on Cardiac Gene Expression Profile and Post-Infarct Cardiac Remodelling in Diabetic Goto-Kakizaki Rats


Author for correspondence: Eero Mervaala, Institute of Biomedicine, Pharmacology, University of Helsinki, PO Box 63, FI-00014 Helsinki, Finland (fax +358 9 191 25 364, e-mail eero.mervaala@helsinki.fi).


Abstract:  The calcium sensitizer levosimendan has shown beneficial effects on cardiac remodelling in spontaneously diabetic Goto-Kakizaki (GK) rats 12 weeks after experimental myocardial infarction (MI). However, the short-term effects and the cellular mechanisms remain partially unresolved. The aim was to study the effects of oral levosimendan treatment on the myocardial gene expression profile in diabetic GK rats 4 weeks after MI/sham operation. MI was induced to diabetic GK rats. Twenty-four hours after surgery, rats were randomized into four groups: MI, MI +levosimendan (1 mg/kg/day), sham-operated and sham-operated +levosimendan. Cardiac function and histology were examined 1, 4 and 12 weeks after MI. The effects of levosimendan on cardiac gene expression profile were investigated by microarray analysis. Levosimendan ameliorated post-infarct heart failure and cardiac remodelling. Levosimendan altered the expression of 264 of MI and sham rats, respectively; these changes were associated with alterations in two Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. Levosimendan up-regulated 3 genes in the renin-angiotensin system pathway [angiotensin receptor 1 (Agtr1), chymase 1 (Cma1) and thimet oligopeptidase 1 (Thop1)] and down-regulated 3 genes in the glycerolipid metabolism pathway [diacylglycerol kinase gamma (Dgkg), carboxyl ester lipase (Cel) and Diacylglycerol kinase iota]. Levosimendan induced opposite effects on the gene expression of pleckstrin homology (PH) domain?containing family f (Plekhf1), carboxymethylenebutenolidase homologue (Cmbl) (up-regulation) and hydroxyprostaglandin dehydrogenase 15 (Hpgd) (down-regulation) as compared with MI. MI versus sham affected 420 genes and was associated with alterations in 12 KEGG pathways. The beneficial effects of levosimendan on cardiac hypertrophy in sham-operated GK rats was associated with altered expression in 522 genes and associated with three KEGG pathways including purine metabolism, cell cycle pathway and pathways in cancer. Levosimendan protects against post-infarct heart failure and cardiac remodelling. Analysis of the cardiac transcriptome revealed several genes that are regulated by levosimendan. These genes may represent novel drug targets for heart failure and diabetic cardiomyopathy.

Recently, the term diabetic cardiomyopathy has been introduced to describe the independent adverse effects of diabetes on the heart [1]. Diabetes is presently recognized as a risk factor independent from other common comorbidities such as hypertension and/or dyslipidaemia in the development of cardiac complications including fatal MI [2]. Ventricular remodelling in diabetic cardiomyopathy includes left ventricular hypertrophy, cardiomyocyte apoptosis and increased interstitial collagen deposition [1]. Furthermore, microvascular changes in coronary arteries play a role in the development of atherosclerosis and increased risk for MI. The increased prevalence of type 2 diabetes requires the development of therapeutics related to its complications. To date, early therapeutic intervention has been the rule to prevent diabetes-related complications. The different treatment arms commonly used to reduce the risk for complications include lipid-lowering medication and antihypertensive drugs. However, for the failing heart, strikingly few drugs have been developed recently. Calcium-sensitizing treatment has been shown to improve cardiac function in patients with acutely decompensated heart failure [3], whereas long-term use of more common inotropic agents for heart failure, such as beta-adrenergic agonists and Phosphodiesterase (PDE) inhibitors, has been shown to increase mortality and morbidity. Levosimendan mediates its effect by sensitization of the contractile protein troponin C to calcium [4]. Levosimendan binds to the calcium-saturated N-terminal domain of troponin C in cardiac muscle and stabilizes the troponin molecule with subsequent prolongation of its effect on the contractile proteins [5]. Besides its effect as a positive inotrope, levosimendan exerts vasodilatory effects by opening of ATP-dependent K+ channels [6].

The spontaneously diabetic GK rats are salt sensitive and exhibit cardiac hypertrophy and increased interstitial collagen deposition [7]. These features are further exacerbated after experimental MI [8]. Our recent study showed the beneficial effects of a 12-week administration of oral levosimendan in diabetic GK rats with MI. Levosimendan treatment showed beneficial effects on post-infarct cardiac function, cardiomyocyte hypertrophy and inflammation in spontaneously diabetic GK rats [9]. In these rats, we have previously shown that the cardiovascular pathogenesis is mediated at least in part by increased activity of the renin-angiotensin system [10]. Our previous study with post-MI GK rats treated with OR-1896, a stable metabolite of levosimendan, showed a protective effect at 4 weeks after MI [11]. The mechanism behind the beneficial effects of chronic levosimendan treatment on left ventricle (LV) remodelling in diabetic rats was not elucidated in our previous studies. In addition to its acute effects on cardiac dysfunction, levosimendan has been suggested to exert long-term, cardioprotective effects through mechanisms other than calcium sensitization and vasodilatation [12]. A potential mechanism has been suggested to be the activation of mitochondrial ATP-dependent K+ channels and its negative effect on free radical production and apoptosis [12,13]. In the present study, we aimed to identify changes in myocardial gene expression after 4 weeks of levosimendan treatment, which might explain the beneficial effects on LV remodelling shown in previous studies. We used a microarray approach to analyse the effects of levosimendan on myocardial gene expression profile in post-MI and sham-operated spontaneously diabetic GK rats.

Materials and Methods

Animals, myocardial infarction and levosimendan.

The investigation conforms to the Guide for the care and use of laboratory animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). Experimental MI was induced by ligating the left anterior descending coronary artery in 8-week-old spontaneously diabetic GK rats (M&B, Skensved, Denmark) rats as described previously [9] under ketamine (50 mg/kg, i.p.) and medetomidine (10 mg/kg, i.p.) anaesthesia. Long-acting insulin (1 IU/rat) was given 2 hr before anaesthesia to prevent hyperglycaemia. The rats received post-operative pain reliever (buprenorphine 0.01–0.05 mg/kg s.c.) twice a day for two consecutive days. Twenty-four hours after surgery, GK rats were randomized into the following groups: MI group without treatment (GK MI 1, 4 and 12 weeks, n = 9 each), levosimendan-treated MI group (GK MI + Levo, 1 mg/kg/day 1, 4 and 12 weeks, n = 10 each), sham-operated group (GK sham 1, 4 and 12 weeks, n = 10 each) and levosimendan-treated sham group (GK sham + Levo, 1 mg/kg/day, 4 weeks, n = 10). Levosimendan was given orally via drinking fluid (10 mg/l). Levosimendan dosage was chosen according to our previous dose–response experiment in Dahl salt-sensitive rats [14] and survival study in Wistar rats with MI [15]. Levosimendan given according to the protocol in the above-mentioned studies produced a significant plasma level concentration and exerted a clear therapeutic effect. The protocols were approved by the Animal Experimentation Committee of the University of Helsinki, Finland, and the Provincial State Office of Southern Finland (approval number STU 1187 A).

Echocardiography, blood pressure recordings and sample preparations.

Systolic blood pressure was measured using a tail cuff blood pressure analyzer (Apollo-2AB Blood Pressure Analyzer, Model 179-2AB; IITC Life Science, Woodland Hills, CA, USA) at week 4. Echocardiography was performed at 1, 4 and 12 weeks as described previously [8]. In short, transthoracic echocardiography (Toshiba Ultrasound, Tokyo, Japan) was performed under isoflurane anaesthesia (AGA, Riihimäki, Finland) in a blinded fashion by the same technician during the last study week. At weeks 1, 4 and 12 after start of treatment, the rats were anaesthetized with CO2/O2 (AGA) and decapitated. Blood samples were collected for biochemical measurements using EDTA as an anticoagulant. The hearts were excised, washed with ice-cold saline, blotted dry, weighed and snap-frozen in liquid nitrogen or isopentane −35°C). All samples were stored at −80°C until assayed.

Histology and immunohistochemistry.

Infarct size was determined planimetrically from picrosirius red–stained whole heart tissue slides as described previously [8]. Conventional light microscopy at ×40 magnification was used to determine cardiomyocyte cross-sectional area from tissue samples taken at 1, 4 and 12 weeks after MI, as described previously. Paraffin-embedded myocardial tissue sections were probed with angiotensin II type 1 receptor (Agt1r) mouse monoclonal antibody (anti-AT1, sc-57036, Santa Cruz biotech, Santa Cruz, CA, USA) according to the instructions of the manufacturer.

Biochemical measurements.

Blood glucose was determined with a handheld test meter (Glucocard II®, Arkray, Japan), and plasma BNP (BNP-45; Peninsula Laboratories, Belmont, CA, USA), plasma renin activity (Angiotensin I RIA kit; Diasorin, Saluggia, Italy), serum insulin (rat insulin RIA kit; Linco, St. Charles, MO, USA) and serum aldosterone (Coat-a-count Aldosterone RIA kit; DPC Biermann, Bad Neuheim, Germany) were determined by radioimmunoassay according to the instructions of the manufacturer. Plasma samples were analysed for levosimendan using liquid chromatography-tandem mass spectrometry (LC-MS/MS) [16].


Total RNA from the rat hearts was extracted with Trizol® (Gibco, Invitrogen, Carlsbad, CA, USA), treated with DNAse 1 (Deoxyribonuclease 1, Sigma Chemicals Co., St Louis, MO, USA) and reverse transcribed to cDNA by reverse transcription enzyme (Enhanced avian HS RT-PCR kit, Sigma Chemicals Co.). The expression of atrial natriuretic peptide (ANP) mRNA was measured using the Lightcycler® instrument (Roche diagnostics, Neuilly sur Seine, France) and FastStart DNA Master SYBR Green 1 master mix (Roche diagnostics) according to the protocol of the manufacturer. Ribosomal 18S was used as a control gene for ANP.

The expression of neutral endopeptidase (NEP), angiotensin II type 1 a receptor (AT1a), Mas receptor (MAS), angiotensin-converting enzyme (ACE), ACE2, prorenin receptor (ProRenR), angiotensin 4 receptor (AT4) and haem oxygenase-2 (HO-2) mRNAs was studied using ABI PRISM 7000 sequence detection system (Applied Biosystems, Foster City, CA, USA). SYBR Green Master mix (Applied Biosystems) was used for AT1a, MAS, ACE, ProRenR, AT4 and HO-2 PCRs and TaqMan Master mix (Applied Biosystems) for NEP and ACE2 PCRs. Data for NEP, AT1a, MAS, ACE, ACE2, ProRenR and AT4 were analysed using the absolute standard curve method as described earlier [17]. The amplification of HO-2 mRNA was used for normalizing the results of NEP, AT1a, MAS, ACE, ACE2, ProRenR and AT4 mRNAs. The un-normalized expression of HO-2 did not differ significantly in the experimental groups (data not shown). Primer and probe sequences are shown in table S2.


Four weeks after experimental MI or sham operation, the hearts were excised and total RNA was labelled and hybridized to rat whole-genome arrays (Rat Expression Array 230 2.0; Affymetrix, Santa Clara, CA, USA). Microarray was performed using four MI samples, four sham samples, three sham + levosimendan samples and three MI + levosimendan samples, each sample on separate chip. Sample labelling, hybridization to chips and image scanning (GeneChip Scanner 3000; Affymetrix) were performed according to the manufacturer’s instructions. The data were analysed by Genespring 7.2 (Agilent, Santa Clara, CA, USA). The Affymetrix data CEL files were pre-processed by Robust Multichip Average (RMA) method. All data from chips passed the quality control. The data were then normalized per chip to the median as normally distributed data.

The obtained expression values were used to calculate the differentially expressed probe sets, selected based on filtering by the parametric statistical analysis not assuming equal variances (Welch-type t-test) with p < 0.05 as a threshold for significance, followed by filtering for fold change (±1.2-fold) between the compared groups. The list of the obtained up- and down-regulated probe sets were inspected for the enriched canonical pathways of the KEGG among the genes using the ‘David 6.7’ programme [NIAD, (NIH), MD, USA] [18]. Our analysis identified canonical pathways that were most significantly enriched among the genes in the lists of up- and down-regulated genes.

Statistical analyses.

Data are presented as means ± S.E.M. Statistically significant differences in mean values were tested by analysis of variance (anova) and the Neuman–Keul’s post hoc test for comparisons of multiple groups.


Effects of levosimendan on systolic function and left ventricular hypertrophy 1, 4 and 12 weeks after myocardial infarction in spontaneously diabetic Goto-Kakizaki.

We studied the time-dependent effects of levosimendan on systolic function and LV hypertrophy. As shown by a decrease in ejection fraction (EF) in M-mode echocardiography, levosimendan exerted its beneficial effects in post-MI GK rats at week 1 after start of treatment (fig. 1A). However, at week 1 after MI, no changes in cardiomyocyte cross-sectional area were seen (fig. 1B). At 4 weeks after MI, systolic function was unchanged; however, left ventricular hypertrophy (LVH), as evaluated by cross-sectional area, was reduced significantly compared with GK + MI (fig. 1B). At 12 weeks after MI, levosimendan exerted a significant increase in ejection fraction together with a further amelioration in LVH (fig. 1).

Figure 1.

 Effects of levosimendan on systolic function and left ventricular hypertrophy 1, 4 and 12 weeks after myocardial infarction (MI) in spontaneously diabetic Goto-Kakizaki (GK) rats. (A) shows ejection fraction change in post-MI rats from baseline (sham) as evaluated by echocardiography. (B) shows the effects of levosimendan on LV hypertrophy in post-MI rats as evaluated by cardiomyocyte cross-sectional area. Data are presented as means ± S.E.M. * indicates < 0.05 versus GK MI.

Effects of levosimendan on systolic function, LV remodelling and neurohumoural activation 4 weeks after myocardial infarction or sham operation.

Levosimendan partly restored systolic function as indicated by a rise in cardiac output, increased ejection fraction and fractional shortening in post-infarct GK rats. Levosimendan did not affect ejection fraction or fractional shortening in sham-operated rats (table 1). Levosimendan partially prevented LV dilation as demonstrated by its effect on left ventricle diameter after MI. Levosimendan did not affect systolic blood pressure or heart rate. Results from echocardiography and blood pressure recordings at 4 weeks after MI or sham operation are given in table 1.

Table 1. 
Echocardiography, blood pressure and heart rate in spontaneously diabetic Goto-Kakizaki (GK) rats 4 weeks after myocardial infarction (MI) or sham operation. Oral levosimendan was given daily for 4 weeks. The data are presented as means ± S.E.M.
 GK sham (N = 10)GK sham + Levo (N = 10)GK MI (N = 9)GK MI + Levo (N = 9)
  1. (s) denotes systole and (d) denotes diastole. *p < 0.05 versus GK sham, **p < 0.05 versus GK sham + Levo and ***p < 0.05 versus GK MI.

Ejection fraction (%)83.2 ± 2.4088.7 ± 1.9425.5 ± 4.94***40.7 ± 5.68******
Fractional shortening (%)46.8 ± 3.0256.5 ± 3.10*10.7 ± 2.30***18.5 ± 3.07***
Cardiac output (ml/min)197 ± 10.7231 ± 14.3157 ± 26.4257 ± 32.9***
LV inner diameter (d)6.70 ± 0.146.73 ± 0.1310.3 ± 0.30***9.43 ± 0.26******
LV inner diameter (s)3.60 ± 0.242.96 ± 0.259.23 ± 0.43***7.72 ± 0.44******
Anterior wall thickness, cm(d)1.93 ± 0.051.94 ± 0.071.78 ± 0.191.56 ± 0.17
Posterior wall thickness, cm(d)2.29 ± 0.082.19 ± 0.092.06 ± 0.182.26 ± 0.14
Systolic blood pressure (mmHg)127 ± 4.1116 ± 2.0124 ± 2.1126 ± 2.2
Heart rate (bpm)381 ± 16.5387 ± 9.0400 ± 9.7373 ± 15.7

MI induced a significant increase in LVH at 4 weeks compared with sham operation (fig. 2). Levosimendan decreased LVH in GK MI rats as evaluated by heart-weight-to-body-weight ratio and cardiomyocyte cross-sectional area (fig. 2). In sham-operated GK rats, levosimendan decreased LVH as evaluated by cardiomyocyte cross-sectional area (fig. 2). The plasma concentration of B-type natriuretic peptide (P-BNP), a marker of volume overload, was decreased in levosimendan-treated rats with MI (fig. 2). ANP mRNA expression was increased in GK MI rats compared with sham. Levosimendan tended to decrease ANP mRNA (fig. 2).

Figure 2.

 Left ventricular hypertrophy and markers of volume overload in spontaneously diabetic Goto-Kakizaki (GK) rats 4 weeks after experimental myocardial infarction (MI). In (A), the average cardiomyocyte cross-sectional area (μm2) calculated from an average of 350 cardiomyocytes in each group (N = 6 per group) is shown. Panel (B) shows the heart-weight-to-body-weight ratios (g/kg). Panel (C) shows plasma BNP concentration, and (D) shows ANP mRNA expression as determined by qRT-PCR form the LV:s of GK and Wistar rats (N = 6). Data are presented as means ± S.E.M; * indicates < 0.05 versus GK SHAM, #< 0.05 versus GK MI and < 0.05 versus GK MI + LEVO.

Planimetry of picrosirius red–stained collagen showed similar sized scars in the anterior wall of the left ventricle and a non-motile anterior wall (fig. 3). Levosimendan tended to reduce collagen volume fraction in post-MI and sham (fig. 3). Levosimendan did not influence serum insulin or blood glucose, serum aldosterone or plasma renin activity. Myocardial infarction did not change blood glucose or serum insulin concentration. Serum aldosterone tended to increase after MI. Basic biochemical parameters are given in table 2.

Figure 3.

 Interstitial fibrosis and M-mode echocardiography 4 weeks after myocardial infarction (MI) in spontaneously diabetic Goto-Kakizaki (GK) rats. The upper panel in (A) shows infarction area measured by planimetry from picrosirius red–stained slides, and the lower part shows picrosirius red–stained whole heart section images and 600ms from representative M-mode echocardiography. (B) shows average collagen volume fraction measured from picrosirius red–stained photomicrograph images and a picrosirius red–stained photomicrograph image (×100 original magnitude) from the remote area of the LV in GK + MI rat. Data are presented as means ± S.E.M. # indicates < 0.05 versus GK MI.

Table 2. 
Basic biochemical parameters in Goto-Kakizaki (GK) rats 4 weeks after myocardial infarction (MI) or sham operation. The data are presented as means ± S.E.M.
 GK sham (N = 10)GK sham + Levo (N = 10)GK MI (N = 9)GK MI + Levo (N = 9)
  1. *p < 0.05 versus GK MI.

fB-glucose (mmol/l)8.47 ± 0.859.88 ± 0.448.17 ± 0.677.37 ± 0.41
S-insulin (ng/ml)2.90 ± 0.284.09 ± 0.293.32 ± 0.582.90 ± 0.31
P-renin activity (ng/ml/h)2.11 ± 0.431.90 ± 0.451.85 ± 0.241.89 ± 0.30
P-BNP (mg/ml)3.73 ± 0.233.65 ± 0.194.54 ± 0.243.62 ± 0.14*
S-aldosterone (pg/ml)194 ± 27.3161 ± 16.9394 ± 95.4389 ± 53.5

Microarray analysis.

Myocardial gene expression was examined in levosimendan-treated rats with MI and sham-operated rats by microarray-based pathway analysis. The results from hierarchal clustering showed altered gene expression in 1206 genes when comparing MI versus sham and the effect of levosimendan in MI and sham (fig. 4). Levosimendan in GK MI rats altered the expression of 264 genes (105 down-regulated and 159 up-regulated). Levosimendan-treated sham compared with sham alone resulted in altered expression in 522 genes (216 down-regulated and 306 up-regulated). MI operation compared with sham resulted in altered expression in 420 genes (223 down-regulated and 197 up-regulated) (fig. 4). Levosimendan targeted common genes in MI and sham rats; levosimendan up-regulated 10 and down-regulated 4 genes in both MI and sham rats (fig. 4). Further, when analysed which genes were down-regulated in MI/sham and up-regulated by levosimendan in MI, we found two genes, namely the Cmbl and the Plekhf1 that fulfilled these criteria. Conversely, one gene was shown to be down-regulated by levosimendan that was up-regulated in MI/sham rats, namely the Hpgd. These genes and their fold changes are given in table 3.

Figure 4.

 Microarray gene expression analysis showing the significantly up- and down-regulated genes in the myocardium of Goto-Kakizaki rats 4 weeks after myocardial infarction (MI) or sham operation. Levosimendan up-regulated 2 genes in post-MI rats that were down-regulated in MI/sham: Cmbl and Plekhf1; levosimendan down-regulated one gene in the MI + levo/MI comparison that was up-regulated in MI/sham: Hpgd. For a list of genes up/down-regulated by levosimendan in MI and sham, refer to table 3.

Table 3. 
(A) Significantly up-regulated (10) and down-regulated (4) genes in levosimendan-treated Goto-Kakizaki (GK) rats 4 weeks after myocardial infarction (MI) or sham operation. (B) shows up- and down-regulated genes in levosimendan-treated GK MI rats where MI/sham produced an opposite effect on gene expression.
Gene IDEntrez 2010Fold change
MI + Levo/MISham + Levo/shamMI/sham

According to the comparisons shown in fig. 4, pathway analysis was performed to gain further insight into the roles of these genes. We identified significant changes in 21 pathways (for a full list of these pathways and the genes, see table S1). For levosimendan + MI versus MI comparison, two significantly enriched pathways were identified (table 4). The most significantly (= 0.025) enriched pathway was the renin-angiotensin system pathway represented by three up-regulated genes, namely Agtr1a, Cma1 and Thop1. The second most significantly (= 0.036) enriched pathway was the glycerolipid pathway with three down-regulated genes, namely Dgkg, Cel and diacylglycerol kinase iota (Dgki) (table 4). The cell adhesion molecule pathway was enriched non-significantly (p = 0.11) with four up-regulated genes, including platelet and lymphocyte selectins (Selp and Sell), vascular cell adhesion molecule 1 (Vcam1) and intercellular adhesion molecule 1 (Icam1), and one down-regulated gene, neural cell adhesion molecule 1 (Ncam1). The other non-significantly enriched pathways included focal adhesion, cytokine–cytokine receptor interaction, purine metabolism pathways in cancer and Fc-gamma R-mediated phagocytosis.

Table 4. 
Significantly (< 0.05) enriched canonical pathways according to the Kyoto Encyclopedia of Genes and Genomes (KEGG) in the myocardium of levosimendan-treated GK rats.
KEGG pathwayEntrez 2010GeneFold change
GK MI + Levo versus GK MI
 Renin-angiotensin system24180Angiotensin II receptor, type 1aAgtr1a1.32
25627chymase 1, mast cellCma11.71
64517Thimet oligopeptidase 1Thop11.22
 Glycerolipid metabolism25666diacylglycerol kinase, gammaDgkg0.81
24254Carboxyl ester lipaseCel0.71
688705Diacylglycerol kinase, iotaDgki0.82
GK sham + Levo versus GK sham
 Cell cycle54237Cell division cycle 2, G1 to S and G2 to MCdc20.55
316273Similar to DNA replication licensing factor MCM3Mcm30.70
288532Minichromosome maintenance deficient 7 (S. cerevisiae)Mcm70.75
399489E2F transcription factor 1E2f10.77
360748Protein kinase, DNA-activated, catalytic polypeptidePrkdc0.80
311412Anaphase-promoting complex subunit 1Anapc10.83
83571Cyclin-dependent kinase inhibitor 1BCdkn1b1.20
313138Origin recognition complex, subunit 3-like (yeast)Orc3l1.24
 Purine metabolism25368Adenosine kinaseAdk1.26
295088Guanine monophosphate synthetaseGmps1.25
58964Non-metastatic cells 6, protein expressed in (nucleoside-diphosphate kinase)Nme61.20
50678Phosphodiesterase 3A, cGMP inhibitedPde3a1.26
25638Phosphodiesterase 4A, cAMP specific (phosphodiesterase E2 dunce homologue, Drosophila)Pde4a1.23
24626Phosphodiesterase 4B, cAMP specificPde4b1.21
117544Phosphoribosyl pyrophosphate amidotransferasePpat1.37
498109Polymerase (RNA) II (DNA directed) polypeptide H;Polr2h1.21
685679RAB11 family interacting protein 3 (class II); non-metastatic cells 4Nme41.29
684425Similar to adenylosuccinate synthetase isozyme 1LOC6844251.22
290029Nucleoside phosphorylaseNp0.81
304573Polymerase (DNA directed), epsilonPole0.69
 Pathways in cancer307403Colony-stimulating factor 1 receptorCsf1r0.76
140589GLI-Kruppel family member GLI1Gli10.76
290905Collagen type IV, alpha 1Col4a10.71
306628Collagen type IV, alpha 2Col4a20.74
309816Laminin, alpha 4Lama40.81
399489E2F transcription factor 1E2f10.77
499505MutS homologue 3 (E. coli)Msh30.78
56718Mechanistic target of rapamycin (serine/threonine kinase)Mtor0.80
84488Fibroblast growth factor 13Fgf131.30
140635Fms-related tyrosine kinase 3Flt31.34
83571Cyclin-dependent kinase inhibitor 1BCdkn1b1.20
25094Microphthalmia-associated transcription factorMitf1.36

For levosimendan-treated sham rats, three significantly enriched pathways were identified (table 4). One enriched (= 0.02) pathway included the purine metabolism pathway with 1.21- to 1.37-fold up-regulations of adenosine kinase (Adk), guanine monophosphate synthetase (Gmps), phosphodiesterase subtypes 3A and 4A and 4B (PDE3A/4A/4B), phosphoribosyl pyrophosphate amidotransferase (Ppat), polymerase II polypeptide H (Pol2rh), RAB11 family interacting protein 3 non-metastatic cells 4 (Nme4), similar to adenylosuccinate synthetase isoenzyme 1 (LOC684425) and 0.81- to 0.69-fold down-regulation of nucleoside phosphorylase (Np) and polymerase epsilon (Pole). The second significantly (= 0.014) enriched cell cycle pathway included 0.83- to 0.55-fold down-regulation in the following genes: cell division cycle 2, G1 to S and G2 to M (Cdc2 or Cdk1), minichromosome maintenance complex component 3 (Mcm3), minichromosome maintenance deficient 7 (Mcm7), E2F transcription factor 1 (E2f1), protein kinase, DNA-activated, catalytic polypeptide (Prkdc), anaphase-promoting kinase inhibitor 1B (Anapc1) and 1.20- to 1.24-fold up-regulation of cyclin-dependent kinase inhibitor 1B (Cdkn1b) and origin recognition complex, subunit 3-like (Orc31). The third pathway included eight significantly down-regulated genes in the pathways in cancer with 0.71- to 0.80-fold changes in colony-stimulating factor 1 receptor (Csf1r), GLI-Kruppel family member GLI1 (Gli1), collagen type IV, alpha 1 (Col4a1), collagen type IV alpha 2 (Col4a2), laminin, alpha 4 (Lama4), E2F transcription factor (E2F1), MutS homologue 3 (E.coli) (Msh3) and mechanistic target of rapamycin (Mtor) (table 4).

In GK MI/sham comparison, the significantly enriched pathways included extracellular membrane –receptor interaction (13 up-regulated genes and one down-regulated gene), valine, leucine and isoleucine degradation pathway (11 down-regulated), focal adhesion (17 up-regulated and one down-regulated), fatty acid metabolism (nine down-regulated), propanoate metabolism (eight down-regulated), cytokine–cytokine receptor interaction (seven up-regulated, three down-regulated), asthma (three up-regulated), pyruvate metabolism (four down-regulated), intestinal immune network for IgA production (four up-regulated), Fc-gamma R-mediated phagocytosis (five up-regulated), systemic lupus erythomatosus (five up-regulated), cell adhesion molecules (six up-regulated), butanoate metabolism (four down-regulated) and fatty acid elongation in mitochondria (three down-regulated) (for a full list, see table S1).

The effects of levosimendan on components of the tissue renin-angiotensin system.

The gene expression of several components of the renin-angiotensin system (RAS) was confirmed by quantitative RT-PCR. Our results showed that levosimendan did not increase the mRNA expression of Agtr1 (fig. 5). In the sham-GK rat myocardium, levosimendan increased the mRNA expression of MAS and NEP. MI increased the mRNA expression of Agtr1, ACE and ACE2 compared with sham.

Figure 5.

 Effects of levosimendan on components of intracardiac renin-angiotensin system mRNA expression in spontaneously diabetic Goto-Kakizaki (GK) rats with myocardial infarction (MI). (A) angiotensin II type 1 receptor (AT1R), (B) AT1R-–antibody-stained myocardial tissue samples (×400) upper panel: GK + MI rat heart, lower panel: GK + sham, (C) Mas receptor mRNA expression, (D) neutral endopeptidase, (E) angiotensin-converting enzyme (ACE), (F) ACE2, (G) prorenin receptor and (H) angiotensin receptor type 4. Data are presented as means ± S.E.M; * indicates < 0.05 versus GK SHAM, #< 0.05 versus GK MI and < 0.05 versus GK MI + Levo.


Our earlier studies have shown the beneficial effects of calcium sensitizer therapy in ameliorating both structural and functional myocardial damage in spontaneously diabetic GK rats 12 weeks after MI [9,11]. The present study showed that 4 weeks of oral levosimendan treatment exerts beneficial effects on LV remodelling in spontaneously diabetic GK rats with or without experimental myocardial infarction. The results were associated with changes in the expression of genes that have not before been characterized as targets of levosimendan. Pathway analysis showed that levosimendan down-regulated the expression of genes associated with glycerolipid metabolism and up-regulated genes in the renin-angiotensin system pathway in GK rats with post-MI ventricular remodelling. Further, in sham-operated GK rats, levosimendan treatment was associated with down-regulation of cell cycle genes, up-regulation of purine metabolism genes and down-regulation of genes in the pathways in cancer.

MI in GK rats produced a marked decrease in haemodynamic parameters including a significant drop in ejection fraction and systolic blood pressure and an increase in left ventricular hypertrophy and collagen volume fraction. Microarray showed that MI significantly altered gene expression within pathways implicated in energy metabolism including a decrease in expression of 9 genes in the fatty acid metabolism pathway. Diabetes is associated with an increase in FA oxidation in the heart [19]; antioxidants and Peroxisome proliferator activated receptor (PPAR)-alpha agonists that reduce fatty acid oxidation are known to have therapeutic potential in diabetes-induced LV remodelling [20]. MI further impairs myocardial energy metabolism, and it has been proposed that the underlying mismatch in energy metabolism predisposes the diabetic heart to complications after MI [21]. The data from microarray provide new insight into the potential genes that may be involved in the pathogenesis of post-MI diabetic cardiomyopathy.

In the present study, we showed that levosimendan attenuated left ventricular hypertrophy and interstitial collagen deposition in GK rats after MI. The beneficial effect on hypertrophy was present in both sham and MI rats. Interestingly, our results showed that the expression of mTOR was decreased in levosimendan-treated sham rats. As the mTOR protein is known to promote cell growth via PI3K/Akt signalling [22], the potential effect of levosimendan on the gene product should be further investigated. Further, microarray suggested that decreased interstitial fibrosis may be at least in part because of the decrease in mRNA expression of collagen type IV.

We showed here that levosimendan down-regulated the Hpgd gene in GK MI rats. The Hpgd gene encodes the NAD+-dependent 15-dehydroprostaglandin dehydrogenase enzyme [15-prostaglandin dehydrogenase (PGDH)], one of the main inactivators of prostaglandins [23]. Interestingly, PPAR-gamma agonists used in the management of diabetes, including ciglitazone and pioglitazone, have been shown to inhibit 15-PGDH [24]. These results are interesting and may in part explain the beneficial effects of levosimendan in the spontaneously diabetic GK rats. However, these results should be confirmed in subsequent studies. The Plekhf1 gene encodes the protein PH domain–containing family F member 1. Plekhf1 was up-regulated in levosimendan-treated GK MI rats and down-regulated in GK MI versus sham. Proteins containing the PH domain are implicated in induction of apoptosis via caspase-independent pathways [25]. The implications here remained unclear because levosimendan treatment is known to be associated with decreased cardiomyocyte apoptosis [9]. The other gene that levosimendan up-regulated in the opposite direction to MI was the Cmbl gene that encodes the human homologue of Pseudomonas carboxymethylenebutenolidase enzyme. Its effect in human beings has not been elucidated, but recently a study showed that Cmbl is involved in the conversion of the angiotensin II type I receptor antagonist (AIIRA) olmesartan medoxomil to its bioactive metabolite olmesartan [26]. In both MI and sham, levosimendan up-regulated the Arl6ip2 gene (or Atlastin GTPase2) that encodes the adenosine diphosphate-ribosylation factor-like six interacting protein 2, implicated in Golgi apparatus and endoplasmic reticulum organization. Furthermore, levosimendan up-regulated the Atp5s gene that encodes a subunit for the ATP synthase protein. ATP synthase is localized in the mitochondria and produces energy in the form of ATP coupled to the transport of hydrogen atoms. An increase in ATP production may help the energy-starved and infracted heart and may at least in part explain the beneficial effects seen here. The Cxl11 gene was up-regulated in levosimendan-treated rats; this gene encodes the C-X-C-motif chemokine implicated in homing of CD4+ cells of the immune system. Further, levosimendan up-regulated the Mlx gene encoding the MAX-like protein X transcription factor. The Mlx is a mediator of glucose-induced gene expression in the liver and a key target in the signal transduction of glucose and polyunsaturated fatty acids [27]. The mitochondrial ribosomal protein L41 (Mrpl41) was up-regulated by levosimendan; this protein is an integral constituent of the ribosome. Mrpl41 is a tumour suppressor, implicated in stabilization of p53 (apoptosis) and p27(Kip1) (G1 cell cycle arrest) [28]. Levosimendan further up-regulated Vps28, a gene encoding the yeast homologue of the vacuolar protein sorting 28, and no conclusive role for this protein in mammals has been established. Levosimendan down-regulated the expression of Etv5, a gene encoding the ETS translation variant five protein; studies have implicated Etv5 in the modulation of oxidative stress in endometrial carcinoma [29]. However, no studies concerning Etv5 in the myocardium were identified. Lphn1 or latrophilin 1 was down-regulated in levosimendan-treated MI and sham rats. This gene encodes the protein latrophilin or calcium-independent alpha-latrotoxin receptor 1 implicated in neurodegeneration [30]. The Paxip1 gene encodes the PAX interacting (with transcription activation domain) protein 1. The gene was down-regulated by levosimendan here; its protein controls DNA stability by histone methylation, acetylation and interacting with p53 [31]. Levosimendan further down-regulated the Tln1 gene encoding the talin 1 protein implicated in focal adhesion and actin cytoskeleton organization. Taken together, these results yielded interesting novel gene targets of levosimendan and may partly explain the beneficial effects on LV remodelling.

Pathway analysis indicated that components of the local RAS are influenced by levosimendan in post-MI rats. The local RAS promotes hypertrophy in the heart through the angiotensin II type 1 receptor (AT1R) [32]. Accordingly, previous studies have shown the efficacy of angiotensin II type 1 receptor blocker (ARB) and ACE inhibitors in alleviating cardiovascular complications including cardiac hypertrophy in diabetic GK rats [10]. We observed that three genes encoding molecules that participate in RAS peptide metabolism were up-regulated in levosimendan-treated GK MI rats compared with untreated GK MI rats; chymase 1 (CMA1), thimet oligopeptidase1 (THOP1 or EP24.15) and the angiotensin receptor subtype 1a (AT1R). Our results from microarray showed that the angiotensin receptor subtype 1 was increased in GK + MI rats treated with levosimendan. However, when we performed RT-PCR, the results showed a modest 1.25-fold, non-significant change in AT1R mRNA expression in levosimendan-treated versus untreated GK + MI rats. In the present study, cardiac hypertrophy and fibrosis were ameliorated by levosimendan, and thus, the potentially harmful effects of increased myocardial AT1R expression seen here could not be confirmed, possibly due to the other beneficial effects of the calcium sensitizer treatment. In contrast, the beneficial effects of the local RAS via angiotensin (1–7)–MAS axis activation include vasodilatory and antihypertrophic effects [33,34]. Ang (1-7) is produced from angiotensin I (ang I) by enzymes including THOP1, NEP (neprilysin or MME) and prolyl oligopeptidase [35]. Of these, levosimendan increased THOP1 expression in MI and NEP mRNA in sham rats. Levosimendan increased MAS mRNA expression in GK sham rats. These results give an indication that levosimendan may be able to shunt ang I from forming ang II to the ang (1-7)–MAS pathway. Our results further showed that CMA1 was up-regulated 1.7-fold in levosimendan-treated GK MI rats. The enzyme is a converter of ang I to ang II in the heart and blood vessels. This would suggest that local conversion to ang II may have been increased in levosimendan-treated rat hearts.

Levosimendan decreased the expression of three genes in the glycerolipid pathway, namely Dgkg, Dgki and Cel. The functions of Dgkg and Dgki are to metabolize diacylglycerol into phosphatidic acid and synthesize glycerolipids, whereas Cel is involved in steroid and glycerolipid synthesis. Diacylglycerol that is cleaved by Dgk:s is a second messenger that activates novel and classical Ca2+-dependent Protein kinase C (PKC), and hence, Dgk:s are negative regulators of PKC [36]. Therefore, Dgkg and Dgki down-regulation would, at least in theory, be associated with increased PKC activity. In the heart, PKC is inked with increased proliferation, cardiac contraction and hypertrophy. In sham-operated levosimendan-treated rats, we showed here that several genes in the cell cycle pathway were down-regulated, including Cdc2 and E2f1 transcription factor that regulate transition from the gap (G) to the DNA synthesis (S) phase. The implications of these results are not entirely clear.

One study limitation in the present study was that we did not investigate protein expression of the genes in microarray that were regulated by levosimendan. The present study aimed to profile the myocardial transcriptome in diabetic rats with MI after levosimendan treatment. Whether the changes seen in the myocardial mRNA expressions were caused by levosimendan directly or whether they were an effect of amelioration of cardiac function remains somewhat unclear. Further investigation is warranted to investigate the effects of levosimendan on protein expression and to elucidate the implications of these changes in more detail. In earlier studies, levosimendan has been shown to protect the heart from ischaemia/reperfusion injury and reduce cardiac remodelling in experimental models of post-MI heart failure [9,37]. Recent evidence suggests that one of the mechanisms may be opening of mitochondrial KATP channels [38]. Taken together, these results support the hypothesis that levosimendan reduces cardiac remodelling by mechanisms independent from its commonly known calcium-sensitizing and systemic vasodilatory actions. The consequences have been shown here and in earlier studies to include beneficial effects on cardiomyocyte apoptosis, LVH and fibrosis. Future studies should examine the effects of levosimendan on LV remodelling in, e.g., knockout animal models deficient of the targets identified here. The important finding in the present study was the characterization of new molecular targets of calcium sensitizer therapy. Our results indicate that the use of oral calcium sensitizer is an alternative to be reckoned with in the management of post-MI heart failure in diabetes.


This study was supported by grants from the Academy of Finland, Finnish Foundation of Cardiovascular Research, Päivikki and Sakari Sohlberg Foundation and the Sigrid Jusélius Foundation. We are grateful to Ms. Anneli von Behr, Ms. Sari Laakkonen and Ms. Essi Martonen for expert technical assistance.