Inhibition of the prolyl 4 hydroxylases (P4H) reduces infarct size and improves cardiac function after myocardial infarction [1,2] and reduces Left Ventricular (LV) hypertrophy and improves LV dysfunction in rats undergoing thoracic aortic banding [3]. Further it has been shown, that inhibition of the prolyl hydroxylase protects myocardium by activating NOS and generating mitochondrial ROS [4].

Two major isoforms of P4H, collagen and HIF P4H augmenting collagen maturation and facilitating HIF degradation, respectively, are currently known. Collagen P4H (C-P4H) is a key enzyme in collagen maturation catalyzing hydroxylation of proline residues, resulting in thermally stable triple helix collagen. Inhibition of C-P4H is therefore expected to favor production of unstable collagen within the endoplasmatic reticulum, which then is rapidly degraded. The C-P4H inhibitor (C-P4HI) FG 0041, which has been used here reduced the amount of collagen and collagen proline hydroxylation, prevented LV remodeling and improved survival following acute myocardial infarction when administered 48 h after induction of myocardial infarction for 4 weeks in rats [2]. We demonstrated that C-P4HI leads to an improvement of AoB-associated LV dysfunction and reduces imbalance of extracellular matrix turnover and hypertrophy-associated gene expression. C-P4H inhibition could therefore be of value in treatment of myocardial remodeling accompanying pressure overload hypertrophy [3].

In contrast we reported that stabilization of HIF rather than inhibition of collagen maturation by P4-HI may prevent cardiac remodeling after MI. Inhibition of HIF-P4H with the specific inhibitor FG 2216 increased myocardial HIF levels prior to induction of myocardial infarction and throughout the treatment period [1].

The aim of the present study was to determine the functional differences between two different orally available P4HIs, FG 2216 and FG 0041. We used the model of aortocaval shunt induced cardiac failure [5-7]. To identify the different properties of these substances we used the ACS model, in which fibrosis is negligible. Thus, we can distinguish the effect of both substances on left ventricular remodeling. This model seemed to be suitable comparing the beneficial effects of stabilizing HIF (FG 2216) compared to the production of unstable collagen within the endoplasmatic reticulum, which then is rapidly degraded (FG 0041).

Experimental heart failure model using infrarenal aortocaval shunt in the rat

The infrarenal Aortocaval Shunt (ACS) in male Wistar rats (weight on average 250 g) was used as a model of volume overload-induced heart failure as previously described [5-7]. Sham-operated rats used as controls were treated identically, except that no puncture of the vessels was performed. Four weeks after surgery rats were anaesthetized for the echocardiographic study. After 30 days hemodynamic measurements were performed. At the termination of the study the rats were anesthetized, the hearts removed, weighed and kept in liquid nitrogen until further analysis. For analysis of HIF by immunohistochemistry some hearts were cut into 1.5 mm transverse slices and fixed by immersion in 3% paraformaldehyde and embedded in paraffin (Figure 1).

Figure 1: Experimental protocol and ACS procedure for the animal groups. On day 0 the aortocaval shunt (ACS) was operated, on day 2 the P4H inhibitor or vehicle treatment started (Tx – treatment, control, vehicle, FG 2216 and FG 0041 BID), on day 28 echocardiographic measurements were done, on day 30 hemodynamic measurements (HD) were performed, the animals were sacrificed and the tissue was stored according to the necessary protocol. A schematic of the ACS procedure is indicated within the Figure. The animals received either P4-HI or vehicle, giving a total of four groups: sham (n = 14), control (n = 15), FG 2216 (n = 10), and FG 0041 (n = 10).

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) and was approved by the local committee.

Treatment with P4-HI

Prolyl 4 hydroxylase inhibitors (P4HI) FG 2216 (25 mg/kg twice daily) and FG 0041 (25 mg/kg twice daily) treatment was started 48 hours after surgery. The perioperative mortality rate was 15%. From days 2 to 30, ACS rats received the treatment by oral gavage. The animals received either P4-HI or vehicle, giving a total of four groups: sham (n=14), control (n=15), FG 2216 (n=10), and FG 0041 (n=10). No treatment related effect on mortality was observed.

Echocardiography

Echocardiography was performed on day 28. A two-dimensional short-axis and long-axis view of the left ventricle was obtained with a 15-MHz transducer (Acuson Sequoia, Erlangen, Germany) in anesthetized rats. M-mode tracings were recorded and used to determine the diameter of the left ventricle at the end of the diastole (LVEDD) and systole (LVESD). Ejection fraction was calculated using short and long-axis views. Measurements were done online by an observer blinded to treatment according to the American Society for Echocardiography leading-edge method.

Hemodynamic measurements

Hemodynamic measurements were performed on day 30 before sacrifice. Rats were intubated and artificially ventilated under chloral hydrate anesthesia. A PE 50-catheter was inserted through the right jugular vein into the superior vena cava to measure central venous pressure. Arterial blood pressure was measured directly via the left carotid artery with a Millar tip high fidelity catheter. Left ventricular hemodynamics were measured with a Millar catheter and registered with a Statham transducer (P23XL) and a Gould amplifier (AMP 4600). Left ventricular contractility (dP/dtmax) was obtained from the ventricular pressure curves, which were converted with a Gould differentiator (G4615). Heart rate was measured with a 3-channel ECG.

Determination of HIF by immunehistochemistry

Immunohistochemistry was done on 4 μm sections of paraffinembedded formaldehyde-fixed tissue as previously [8,9] described. HIF-1α was detected by a mouse monoclonal antibody, which also cross reacts with HIF-1α rat (α67, Novus Biologicals, Littleton) at a dilution of 1:6,000. HIF-2α was detected by polyclonal rabbit antimouse antibody (PM9, obtained from a rabbit immunized against a fusion-protein containing a 337-439 of mouse HIF-2α), at a dilution of 1:3,000. All incubations were performed in a humidified chamber. Between incubations, specimens were washed 2-4 times in TBST (50 mM Tris-HCl, 300 mM NaCl, 0.1% Tween-20, pH 7.6). Biotinylated secondary anti-mouse or anti-rabbit antibodies and a catalyzed signal amplification system (CSA-DAKO, Hamburg, Germany), based on a streptavidin-biotin-peroxidase reaction, were used. ‘Backto- back’ sections were processed for co-localization of HIF-1α and HIF-2α, of which the first was mounted up side down on the slide. For double-labeling immunohistochemistry, the anti-HIF antibodyantigen- complexes were first revealed using streptavidin-texas-red in combination with the CSA-system. Subsequently, the sections were incubated with the anti-ED-antibody followed by a CY2-labeled anti-mouse antibody. Controls included sham-operated animals, the omission of primary antibodies and the use of preimmune serum of animals immunized against HIF-2α. Sections were counterstained with aqueous hematoxylin before application of gelatin mounting medium. Signals were analyzed with a Leica DMRB microscope (Leica, Bensheim, Germany) in differential interference contrast. Photographs were digitally recorded by means of a Visitron system (Visitron, Puchheim, Germany) and mirrored in case of adjacent “back to back” processing.

Quantitation of mRNA by real time PCR

Total RNA preparation, Deoxyribonuclease (DNase) digestion and reverse transcription were performed as described previously [3,10]. Briefly, RNA was extracted from LV samples using the RNAzol B™ reagent (Lorei & Pasel, Germany). RNA was digested with RNase-free DNase I (Boehringer, Germany). From each sample, 250 ng DNase digested RNA was reverse transcribed with random hexamers and using Superscript™ RNaseH-Reverse Transcriptase according to the manufacturers protocol (Gibco BRL, Germany). A “hot start” real time PCR procedure with SYBR Green that was validated in respect of reproducibility and linearity within the measuring range was performed in duplicates with the TaqMan™ 7700 instrument (ABI). Primers for all target genes were designed using Dnasis version 2.1 and the Primer Express Software (PE Applied Biosystems). The primer sequences are shown in Supplemental Table 1. All PCR reactions had efficiencies of about 1.9. A calibration curve containing 50 ng, 25 ng, 12.5 ng, 6.25 ng and 3.125 ng cDNA pooled from all samples analysed was used to estimate relative changes of mRNA expression within each sample and was run in each PCR reaction with the specific primers used. To correct for potential variances between samples in mRNA extraction or in RTefficiency, the mRNA content of the target genes was normalized to the expression of the stably expressed reference gene GAPDH in the same sample.

Protein extraction and immunoblotting

LV myocardial samples were homogenized (30 sec, 2000 rpm) in ice-cold extraction buffer (1:3 wt/vol) containing (10 mM Tris HCL, pH 7.5, 140 mM NaCl, 1 mM EDTA 25% Glycerol, 0.5% SDS, 0.5% Nonident P-40, 0.1 mM DTT, 0.5 mM PMSF, 100 ng/ml Protease inhibitor cocktail) and then cleared by centrifugation (4°C, 10 min, 14.000 rpm). The supernatant was assayed for protein concentration using Bio-Rad Protein Assay and flash frozen in liquid nitrogen at –80°C. Ten to twenty μg of protein was separated by 10% SDSPAGE and blotted onto nitrocellulose membranes (Fa. Amersham). Membranes were incubated with specific monoclonal antibodies. As secondary antibody, specific HRP-conjugated antibodies were used and the signal was visualized with the ECL detection kit (Amersham Pharmacia Biotech). The immunoblots were digitized on a transluminate scanner and the specific protein band was analyzed with AlphaEaseFC (Software, version 3.1.2, Alpha Innotech Corporation). To normalize for different protein content, we stripped the membrane from the first antibody complex with buffer A (Buffer A: 200 mM Glycin pH2.2, 0.1% SDS, 1% TWEEN 20) for 12 hours at 4°C and rehybridized for GAPDH (primary antibody: Chemicon, MAB- 374, 1:5.000; secondary antibody: Donkey anti-mouse, Dianova, 1:50.000) as described above.

Statistical methods

Values are given as mean ± SEM except where indicated. Differences in echocardiographic and hemodynamic parameters as well as organ weights between MI-P4-HI, MI-control, and sham were analyzed by one-way ANOVA (Tukey’s range test), followed by a modified Student’s t-test. RT-PCR Statistics were calculated with the Excel 2000, Sigma Plot 8.0 and SPSS 11.0 software. The t-test was used to calculate differences between the examined groups and ANOVA corrections for multiple testing were used. A p-value <0.05 was considered to be statistically significant.

Morphology and functional data

Aortocaval Shunt (ACS) caused an increase in heart (right and left ventricular weight) and lung weight compared to sham operated rats. Treatment with FG 0041 did not affect heart weight but significantly reduced lung weight. Treatment with FG 2216 significantly reduced heart (both right and left ventricular) weight and lung weight (Figure 2). There was no difference in the relative heart and lung weight.

Figure 2: Body and organ weights after aortocaval shunt. Weights of sham- rats (n = 14), vehicle-treated rats (n = 15) and rats receiving either FG 0041 (n = 10) or FG 2216 (n = 10) are shown. BW = bodyweight, HW = heart weight, LV = left ventricular weight, RV= right ventricular weight, lung = lung weight. *p<0.05, **p<0.01 vs. Sham, ^#p<0.05, ^##p<0.01 vs. Vehicle, ^§p<0.05 vs. FG 0041. Data are expressed as the mean value ± SEM.

ACS reduced systolic blood pressure and mildly increased central venous pressure. Left ventricular contractility was reduced in rats with ACS. Treatment with both FG 2216 and FG 0041 normalized left ventricular contractility and increased systolic blood pressure (Figure 3).

Figure 3: Echocardiographic data 28 days after aortocaval shunt. Echocardiographic variables of sham-operated rats, vehicle-treated rats and rats receiving either FG 0041 or FG 2216. P4-HI.
a. LVEDD = left ventricular end-diastolic diameter, LVESD = left ventricular end-systolic diameter, LVH = left ventricular hypertrophy (thickness of interventricular septum and left ventricular posterior wall / 2). *p<0.05, **p<0.01 vs. sham, ^##p<0.01 vs. Vehicle, ^§p<0.05 vs. FG 0041. Data are expressed as the mean value ± standard error of the mean.
b. EF = left ventricular ejection fraction, measured in 4 chamber view. ***p<0.001 vs. sham, ^##p<0.01 vs. Vehicle, ^§§p<0.01 vs. FG 0041. Data are expressed as the mean value ± standard error of the mean.

ACS caused a dilatation of the left ventricle with an increase in left ventricular end-diastolic (LVEDD) and end-systolic diameter (LVESD), increase in left ventricular hypertrophy (LVH) and a reduction of the ejection fraction (EF). Whereas treatment with FG 0041 had no affect on any of these parameters, FG 2216 reduced LVEDD, LVESD and LVH and normalized EF (Figure 4).

Figure 4: Hemodynamic parameters after aortocaval shunt. Hemodynamic parameters of sham-rats and ACS operated rats receiving either vehicle, FG 0041 or FG 2216 are shown. Sham (n = 14), control (n = 15), FG 2216 (n = 10), and FG 0041 (n = 10). SBP = systolic blood pressure; CVP = central venous pressure; LVEDP = left ventricular end-diastolic pressure, dP/dt max = peak positive value of rate of change in left ventricular pressure. *p<0.05 vs. sham. Data are expressed as the mean value ± SEM.

Induction of HIF by inhibition of prolyl 4-hydroxylase in rats with ACS in vivo

It has been shown that P4HI can be used to stabilize HIF α and to enhance HIF activity and its downstream effects, including the induction of angiogenesis [11,12]. In this study we used two specific inhibitors of the P4H–FG 2216 and FG 0041[13]. Previously we demonstrated that FG 2216 induced both HIF-1α and HIF-2α in the heart, kidney and liver after 24 hours of treatment [1]. The regional distribution was ubiquitous and persisted in all organs over the treatment period. We now show that treatment with FG 2216 resulted in a clear increase of HIF-2α protein within the left ventricle, whereas FG 0041, the vehicle-treated, and the control rats did not (Figure 5).

Figure 5: Expression of HIF after P4-HI treatment. Immunohistochemistry of HIF-1α and HIF-2α after treatment with the prolyl 4-hydroxylase inhibitor (P4-HI). The nuclear accumulation of the transcript factor is detectable in cardiomyocytes, interstitial cells and endothelial cells.

Effects of different P4-HI on collagen Collagen expression in ACS

Collagen type I and III mRNA expression was increased under ACS (p<0.01 for all). FG 0041 significantly reduced ACS-induced collagen I and III mRNA expression (Figure 6). Treatment with FG 2216 did not affect the collagen I or III mRNA expression.

Figure 6: Collagen I and III, CTGF, ITG ß1, TGF ß1 MMP 2 and TIMP 2 mRNA expression. mRNA expressions of specific genes are normalized to GAPDH. Mean of sham rats was set as 100% and ACS and ACS +FG 0041 and ACS +FG 2216 are expressed in % of sham. * vs. sham, ^# vs. Vehicle, $vs. FG 2216. Data are presented as mean value ± SEM.

Effect of ACS and P4HI treatment on the matrix remodelling MMP/TIMP system

MMP-2 mRNA (p<0.01) and TIMP 2 mRNA (p<0.05) were induced by ACS and FG 0041 almost normalized both. FG 2216 had no affect on ACS induced MMP/TIMP mRNA expression (Figure 6).

Regulation of growth factors

CTGF, TGFß1 and ITGß1 mRNA were increased in ACS (p<0.01 for all) (Figure 6). ACS-induced CTGF, TGFß1 and ITGß1 mRNA expression were significantly reduced by FG 0041 but not by FG 2216.Thus, FG 2216 had no effect on any of the measured cardiac growth factors (Figure 6).

We reported for the first time that inhibition of the prolyl 4-hydroxylase counteracts changes in cardiac function rather by stabilization of HIF than by modulation of the collagen matrix.

We used in this study two P4H inhibitors with different properties. The collagen P4HI FG 0041 reduces the amount of collagen and collagen proline hydroxylation and therefore prevents collagen maturation [3]. However, FG 0041 treatment had no or minor effect on ACS induced LV hypertrophy, dilatation, or cardiac dysfunction. Interestingly, the cardiac growth factors ITG ß1, CTGF and TGF ß1 have all been significantly reduced by FG 0041. This might suggest that the beneficial effects of FG 0041 might be detectable at a later point, maybe during severe heart failure. Thou in previous studies the beneficial effect was detectable as early as 30 days after MI or AoB [1,3].

Compared to FG 0041 the HIF-P4HI FG 2216 is capable of stabilizing HIF [1]. As previously shown, FG 2216 has no effect on collagen metabolism [1]. In our model of ACS treatment with FG 2216 inhibitor prevented changes in cardiac function independent of a reduction of collagen maturation or altering growth factors expression.Inhibition of HIF-P4H led to an improvement of cardiac function by increasing LV contractility and reduction of LVEDP. It prevented LV dilatation and normalizes LV ejection fraction.

Since the HIF system is stabilized by inhibition of P4H, this protective system might be involved in the beneficial effects observed in our study. The transcription factor HIF is a primary regulator of the hypoxic response, controlling genes involved in diverse processes that balance metabolic supply and demand within tissues [14,15]. Stabilization of HIF α - polypeptides by pharmacological inhibitors of HIF-P4Hs is believed to be therapeutically beneficial in diseases characterized by acute or chronic ischemia, such as myocardial infarction, stroke, peripheral vascular disease and diabetes [16-18]. Under normoxia HIF is continuously generated and then rapidly degraded, mRNA changes are normally not seen. Neither hypoxia nor inhibition leads to an increase of HIF-RNA but rather alters the P4HmRNA [19]. We and others demonstrated that the P4-HI is capable of stabilizing and therefore inducing HIF protein [1,20] in vivo thus demonstrating the potential of a new non-peptide protein hydroxylase inhibitor as a pharmacological activator of the HIF pathway and as a therapeutic tool in treating ischemic diseases. Apparently the same system is involved in the progression of volume-overload induced left ventricular dilatation and reduction of cardiac function.

The beneficial effects of FG 2216 are independent of extracellular collagen metabolism or growth factors, although other effects, particular on cellular metabolism was not looked at and therefore, cannot be ruled out. Recently we demonstrated that inhibition of HIF-P4H protected the myocardium by induction of NOS and generation of free oxygen radicals [4]. Furthermore it has been discovered, that inhibition of the prolyl hydroxylase stimulates erythropoietin production [21,22]. Hence erythropoietin [23,24] is known to be supportive in chronic heart failure possibly by induction of VGEF and NOS this might be another possible mechanism behind the beneficial effects of FG 2216.

Different P4-HIs seem to target different subunits of the enzymes [16] resulting in stabilizing HIF and thereby inducing NOS, VEGF and EPO or in modulation of the collagen matrix.

In summary, inhibition of P4H with FG 2216 improves cardiac function and reduces cardiac hypertrophy with negligible effect on collagen production or extracellular matrix composition. This mechanism might lead to a new therapeutic approach in heart failure.

The authors thank Jeannette Mothes, Jutta Meisel and Rita Günzel for excellent technical and secretarial assistance. This study has been in part financially supported by FibroGen, Inc. Jens Fielitz received the Pfizer fellowship of the German Society of Cardiology. The study was supported by the german institute for high blood pressure.