Tissue factor (TF), the primary initiator of the extrinsic coagulation cascade, is a 263 amino acid transmembrane protein, which is the cellular receptor for factor VII (1- 3). Beyond its role in hemostasis, TF is abundantly expressed by cardiomyocytes in the human heart but not in skeletal muscle (4- 5). In the adult heart, TF is present in the transverse part of the intercalated disk and co-localizes with the cytoskeletal proteins desmin and vinculin (4). The cytoplasmic domain of TF interacts with actin-binding protein 280 (filamin), thereby influencing adhesion and migration of cells (6). The myocardial TF expression was found to correlate with the number of contact sites of cardiomyocytes, pointing to an important role of TF for maintaining the structural integrity and contractility of the myocardial muscle (4- 5). In patients with hypertension and ventricular hypertrophy, the TF level was found to be reduced in structurally altered ventricular myocardium (4). In addition, a complete TF knockout in the murine 129/Sv genetic background leads to TF null embryos that do not survive beyond mid-gestation (7). Defects in cytoskeletal and junctional proteins in dilated cardiomyopathy (DCM) are known to contribute to cardiac dilation and dysfunction (8). Whether alterations in the myocardial TF expression also occur in the cardiac muscle of patients with DCM has not been examined to date. However, mice expressing low levels of human TF showed a myocardial defect with myocardial fibrosis, hemosiderin deposition, and ventricular dysfunction (9). In addition, alternative splicing of the TF pre-messenger ribonucleic acid (mRNA) leads to the generation of a soluble TF isoform (asHTF), which has recently been detected in the circulating blood (10). Leukocytic cells have been suggested to be the source of TF and asHTF in blood (10). Stimulated endothelial cells also express TF mRNA isoforms (11) and therefore may contribute to asHTF in blood as well. Growth factors and cytokines are known to induce the TF expression in a variety of cardiovascular cells (11- 13). Moreover, cytokines may also contribute to the development of DCM (14). Whether asHTF is constitutively expressed in the human myocardium is unknown. Therefore, we examined the expression and localization of TF and its soluble isoform in the cardiac tissue of patients with a normal ejection fraction (EF) and of patients with DCM.

Myocardial biopsies of 79 patients with the presumptive clinical diagnosis of cardiomyopathy and ventricular tissue from 12 explanted DCM hearts were examined. Tissue specimens were taken from the right ventricular septum and immediately frozen in liquid nitrogen in order to preserve RNA and protein content. Coronary angiography was performed in all patients to rule out coronary artery disease. The presence of valvular heart disease, amyloidosis, or other accumulation disease was also excluded. According to the 1995 World Health Organization/International Society and Federation of Cardiology report on the definition and classification of cardiomyopathies, 79 biopsied patients were divided into two groups dependent on the EF measured during left ventricular catheterization: 37 patients had normal to mildly reduced EF (≥50%) and 42 patients moderate to severely reduced EF (<50%). The patients with normal EF had loco-regional wall motion disturbances in at least two different wall segments on echocardiography and/or had clinical symptoms such as unexplained chest pain or heart rhythm disturbances. Ventricular septum of explanted hearts from patients with DCM and severely impaired myocardial function was also examined in a separate group (n = 12). To keep the amount of myocardial biopsies as low as possible and to avoid unnecessary risks for the patients, either TF mRNA or protein levels were determined from a myocardial biopsy. To assess the impact of cardiomyopathy on TF expression, mRNA levels were quantified by real-time polymerase chain reaction (PCR) in 46 of 79 biopsied patients (n = 27 with an EF <50% and n = 19 with an EF ≥50%) and in the explanted hearts (n = 12). The content of TF protein in myocardial biopsies was measured by TF enzyme-linked immunosorbent assay (ELISA) in 33 of the 79 biopsied patients (n = 15 with an EF <50% and n = 18 with an EF ≥50%). Immunohistochemistry was performed to visualize TF protein and localization within the myocardial tissues. All procedures were approved by our institutional review board and performed in accordance with ethical standards and the Helsinki Declaration of 1975. All patients gave informed consent for the study.

Total RNA of human heart biopsies was isolated by the TRIzol method according to manufacturer's instruction (Invitrogen, Karlsruhe, Germany), reverse transcribed, and analyzed by PCR using primers listed in (Table 1)A. Products were separated by electrophoresis on 1% agarose, visualized with ethidium bromide under ultraviolet light, and bands were excised. To ensure that the products obtained by transcription represented TF isoform sequences, full-length TF as well as asHTF (Gene Bank accession number AF 487337) specific sequences were verified by direct sequencing of the PCR products. Conventional PCR was performed under the following conditions: 94°C, 2 min; 94°C, 15 s; 64°C, 30 s; 68°C, 1 min for 30 cycles.

Total RNA was isolated as described previously. First, 0.5 μg total RNA of each sample was transcribed into complementary deoxyribonucleic acid (DNA) by using avian mycoblastosis virus reverse transcriptase (AMV RT) and random hexamer primers according to the supplier's instructions (Roche Applied Sciences, Mannheim, Germany). Then, 3 μl of each complementary DNA preparation was diluted to a final PCR volume of 25 μl containing 12.5 μl TaqMan Universal Master Mix (Applied Biosystems, Foster City, California), and 0.25 μl (200 nM) primers (TIB Molbiol, Berlin, Germany) and 0.3 μl (100 nM) probes as listed in (Table 1). Each sample was tested for the following mRNAs: TF, asHTF, and GAPDH. In order to differentiate between TF and asHTF, primers and probe for detection of TF have been positioned in exon 5, which is missing in asHTF, whereas asHTF primers have been positioned in exon 4 and 6, respectively, with the probe spanning exon 4/6 boundary, which is not present in TF. Real-time PCR was performed using ABI Prism 7000 Sequence Detection System (Applied Biosystems) under the following conditions: 50°C, 2 min; 95°C, 10 min; 40 cycles 95°C, 15 s, 60°C, 1 min. Standards covering the complete coding sequence for each target were generated by RT-PCR. Serial dilutions of each standard were made.

For TF and asHTF staining, specimens were immediately snap-frozen and stored at −80°C. Cryostat sections were washed with phosphate-buffered saline, incubated in 4% H₂O₂, and immunostained either with polyclonal goat anti-human antibody against TF (American Diagnostica Inc., Stamford, Connecticut), diluted 1:200 in phosphate-buffered saline/fetal calf serum, or with polyclonal rabbit anti-human antibody against asHTF, respectively. Polyclonal asHTF antibodies were raised against the last 29 amino acids of the unique C-terminal asHTF domain, conjugated to keyhole limpet hemocyanin (Pineda Antikörper-Service, Berlin, Germany). A second polyclonal antibody specific for asHTF as previously described (10) was also used for staining. Myocardial staining patterns for asHTF obtained by these two antibodies were identical. After incubation with primary antibody, specimens were washed, incubated with an appropriate biotinylated secondary antibody (DakoCytomation, Hamburg, Germany), and counterstained with hematoxylin (Merck, Darmstadt, Germany). For detection, the Vectastain ABC kit was used according to manufacturer's instructions (Vector Laboratories, Burlingame, California).

Alternatively, immunohistochemistry was performed on paraffin-embedded tissue blocks from biopsies. Several slides made from every block were deparaffinated overnight by incubation with Roti-Histol (Roth, Karlsruhe, Germany). After rehydration they were incubated in a trypsin solution (1 mg/ml, Sigma, Munich, Germany) for 10 min at 37°C to enhance the penetration of the antibodies, washed twice in Tris-buffered saline (TBS) and blocked with Avidin/Biotin Blocking Kit according to manufacturer's protocol (Vector Laboratories). The washed slides were incubated either with a polyclonal anti-TF antibody (1,12) (pAb-sTF, 6 μg/ml) or a monoclonal anti-desmin antibody (clone D33, DakoCytomation, 2.5 μg/ml) overnight at 4°C; negative controls were performed without the use of primary antibodies. The detection of primary antibodies was performed as described earlier. To ensure that the counterstaining with hematoxylin had no influence on the TF staining, control staining was performed without the use of hematoxylin.

Samples from explanted heart tissues were homogenized in radio-immunoprecipitation assay-buffer supplemented with 1% protease-inhibitor cocktail (P-8340, Sigma). For immunoprecipitation extracts were incubated overnight at 4°C with a monoclonal antibody anti-TF (MabTFH clone TFE, Enzyme Research Laboratories, South Bend, Indiana) and afterwards precipitated with protein-G Plus Agarose (Santa Cruz Biotechnology, Santa Cruz, California). The precipitates were separated by electrophoresis on a 12% sodium dodecyl sulfate-polyacrylamide gel (SDS-PAGE). The gels were transferred to a nitrocellulose membrane (Protran BA 85, Schleicher and Schüell, Dassel, Germany), which was blocked in 5% nonfat dry milk diluted in Tris-buffered Saline. Blots were incubated for 3 h at 4°C with the primary polyclonal antibody anti-TF (pAb-sTF, 6 μg/ml). Washed blots were incubated for 90 min with a secondary horseradish peroxidase-conjugated antibody (DakoCytomation). Washed blots were subjected to the Immun-Star HRP Chemiluminescent Kit (BioRad, Hercules, California) for detection of the immunoreactive signal by chemiluminescence, and the membranes were then exposed to X-ray film (Kodak X-OMAT AR, Kodak, Stuttgart, Germany). The membranes were stripped in 100 mM glycine (pH 2.9) at room temperature for 60 min. The stripped blots were blocked again for 3 h and incubated with a primary polyclonal antibody anti-asHTF (dilution 1:100, Pineda Antikörper-Service). After the incubation, the blots were washed and treated as described. The membrane was exposed to X-ray film.

Myocardial biopsies were homogenized in a glass potter containing 100 μl ice-cold radio-immunoprecipitation assay buffer, supplemented with 1% protease-inhibitor cocktail, incubated on ice for 30 min and centrifuged at 10,000 g for 10 min at 4°C. Total protein content of the cell lysate was measured using a bicinchoninic acid protein assay according to the manufacturer's protocol (Pierce, Rockford, Illinois). The protein was precipitated with acetone and afterwards resuspended in TBS, pH 8.5. To quantify the total TF protein content (full-length TF and asHTF) in myocardial biopsies, TF ELISA was performed according to the instruction manual (Imubind Tissue Factor ELISA Kit, American Diagnostica Inc.).

SPSS statistical software version 11.0.1 was used for statistical analysis. Nominal data were tested using chi-square and Fisher exact test. Tissue factor mRNA ratios obtained by real-time PCR were presented as mean ± SD. If more than two groups were compared, metric normally distributed data were analyzed by one-way analysis of variance. If this test revealed significant differences, a Bonferroni-corrected (correction factor 3) nonpaired t test was performed. Other metric normally distributed data were expressed as mean ± SD and compared by nonpaired t test. A p value ≤0.05 was regarded as significant.

The patients, from whom myocardial biopsies were obtained, were divided into two groups according to the EF (EF ≥50% [n = 37] or EF <50% [n = 42]). These groups were then subdivided into another two groups depending on whether mRNA or protein levels were determined. The explanted heart tissues from DCM patients were assessed as a separate group (n = 12). The left ventricular end-diastolic diameter was significantly increased in the DCM patients compared to those patients with normal EF (Tables 2, 3). There was no significant difference in cardiovascular risk factor profile between the above groups (Tables 2, 3).

A statistical difference between the biopsied patients with regard to the medication with beta-blocker, diuretics, and digitalis was found as shown in (Tables 2, 3).

AsHTF mRNA was identified in myocardial biopsies and explanted hearts from DCM patients by conventional RT-PCR. The gel electrophoresis showed not only the PCR product for full length TF but also an amplification product for asHTF ((Table 1)A, Figure 1). The direct sequencing of the PCR products revealed specific sequences for full-length TF and asHTF. The asHTF sequence and protein were previously described to be present in human leukocytic cells but not in human myocardial tissue (10,13). This is the first report about the presence of asHTF in human adult myocardial tissues. The quantification of TF and asHTF mRNA levels by real-time PCR in myocardial biopsies taken from the right ventricular septum of patients with a normal EF revealed myocardial asHTF expression level to be significantly lower than myocardial TF mRNA level (13.91 × 10⁻⁵ ± 11.20 × 10⁻⁵ for asHTF/GAPDH ratio vs. 1.76 × 10⁻¹ ± 6.08 × 10⁻² for TF/GAPDH ratio; p ≤ 0.0001).

To find out whether asHTF protein was expressed in the myocardium, western blot analysis was performed. Full length TF and asHTF were both detected by the polyclonal anti-TF antibody (pAb-sTF, 6 μg/ml) directed against the extracellular domain of the TF protein (Figure 2, lane 1). Rehybridization of the membrane with the asHTF-specific antibody yielded a single protein band at 30 kDa, whereas full-length TF was not detected (Figure 2, lane 2).

Immunohistochemistry showed cardiomyocytes and endothelial cells to be positive for asHTF in myocardial biopsies of patients with normal EF and DCM. However, in contrast to full-length TF, which is abundantly expressed in the cardiomyocytes, asHTF-associated cardiomyocyte staining appeared to be faint ((Figure 3)A and Figure 3B). There was no difference in asHTF myocardial staining pattern comparing patients with normal EF and with DCM (data not shown).

Tissue factor isoform specific real-time PCR was performed to quantify the expression levels of TF and asHTF mRNA in patients with normal EF and with DCM. The expression levels of TF and asHTF were lower in DCM patients than in those with normal EF. Mean and standard deviation of TF/GAPDH and asHTF/GAPDH ratios are shown in (Figure 4). The myocardial TF expression in the biopsied DCM patients decreased to 60.2% (TF/GAPDH ratio 1.76 × 10⁻¹ ± 6.08 × 10⁻² in patients with normal EF vs. 1.06 × 10⁻¹ ± 5.26 × 10⁻² in biopsied DCM patients; p < 0.001). The myocardial asHTF expression in biopsied DCM patients decreased also to 51.5% (asHTF/GAPDH ratio 13.91 × 10⁻⁵ ± 11.2 × 10⁻⁵ in patients with normal EF vs. 7.17 × 10⁻⁵ ± 3.82 × 10⁻⁵ in biopsied DCM patients; p = 0.014). Compared with the level present in patients with normal EF, TF mRNA levels in explanted heart tissues from DCM patients were also decreased to 50.6% for TF (TF/GAPDH ratio 1.76 × 10⁻¹ ± 6.08 × 10⁻² vs. 8.91 × 10⁻² ± 2.95 × 10⁻²; p < 0.0001) and to 42.1% for asHTF (asHTF/GAPDH ratio 13.91 × 10⁻⁵ ± 11.20 × 10⁻⁵ vs. 5.86 × 10⁻⁵ ± 2.93 × 10⁻⁵; p = 0.014). The TF/GAPDH ratio and EF were positively correlated in the patients included (r = 0.504; p < 0.0001; Figure 5), whereas the TF/GAPDH ratio and left ventricular end-diastolic diameter (LVEDD) were negatively and only weakly correlated (r = −0.309; p = 0.018). Correlations between the asHTF/GAPDH ratio and EF (r = 0.382; p = 0.003; Figure 5) as well as between asHTF/GAPDH ratio and LVEDD (r = −0.278; p < 0.04) were also weakly significant.

In line with the mRNA results, protein quantification of total TF content (full-length TF and asHTF) by ELISA revealed a down-regulation of TF in the myocardium from patients with normal to mildly impaired EF compared with those having moderate to severely impaired EF (8.8 ± 3.0 ng/mg total protein (n = 18) vs. 6.5 ± 2.7 ng/mg (n = 15); p = 0.032; Figure 6). Thus, the total myocardial TF protein expression was reduced by approximately 26%. Myocardial TF protein correlated with the EF in biopsied patients (Figure 6B).

Immunohistochemical staining of adjacent specimens from myocardial biopsies with desmin and TF antibodies revealed desmin and TF to be present in the Z-bands of the myocardium in patients with normal EF ((Figure 3)C and Figure 3D). Tissue factor was also found in the sarcolemma and intercalated disks of the myocardial tissue. In contrast to desmin and TF in hearts with normal EF, TF staining of the Z-bands was faint or completely absent in DCM ((Figure 3)F and Figure 3G). However, the Z-bands stained positive for desmin in myocardial biopsies of the same patients with DCM (Figure 3E). In the myocardial biopsies from DCM patients, TF was localized in the perinuclear cytosol of the cardiomyocytes ((Figure 3)F and Figure 3G). Thus, the myocardial specimens from DCM hearts showed an altered cellular localization of TF, redistributing from the Z-bands to the perinuclear cytosol in DCM. To ensure that the counterstaining with hematoxylin had no influence on the TF staining, control staining was performed without the use of hematoxylin ((Figure 3)I and Figure 3J).

Immunohistochemistry with and without hematoxylin counterstaining showed both TF staining of Z-bands in patients with normal EF ((Figure 3)D and Figure 3I), whereas an altered staining pattern without TF-positive Z-bands was found in DCM ((Figure 3)E and Figure 3J).

This study shows TF expression to be down-regulated in the myocardium of patients with DCM. Tissue factor has been reported to be expressed by cardiomyocytes in the sarcolemma and in the transverse part of the intercalated disks, where it co-localizes with the cytoskeletal proteins desmin and vinculin (4- 5). The observation that TF expression correlated with the number of contact sites of cardiomyocytes led to the hypothesis that TF might play an important role in maintaining the structural integrity of the myocardial muscle. The cytoplasmic domain of TF was shown to interact with actin-binding protein 280 (filamin), thereby influencing the adhesion and migration of vascular cells (6). During postnatal maturation the increased degree of TF immunostaining present in the intercalated disks correlates positively with the formation of intercardiomyocyte desmosomal and junctional contacts and has been suggested to possibly influence the electromechanical function of the myocardial muscle (5,15). Lower TF levels have been reported in structurally altered ventricular myocardium of patients with hypertension and ventricular hypertrophy. Tissue factor protein content was decreased approximately 20% in tissue extracts from the myocardium of severely hypertrophic hearts (4). Here, we demonstrate the TF mRNA level in the ventricular septum of patients with DCM to be down-regulated to approximately 60% of that in patients with normal cardiac function. Tissue factor protein content in biopsied DCM patients was reduced by approximately 26%. In addition to the reduction in TF expression, cellular TF localization was altered, with a dislocation of TF from the Z-bands to the cytosol of the cardiomyocytes, suggesting that the reduction and redistribution of TF may contribute to myocardial dilation and contractile dysfunction. Whether down-regulation of TF expression and its cellular translocation contributes to ventricular dilation and cardiac dysfunction needs to be further elucidated. Tissue factor interacts with tissue factor pathway inhibitor, which is expressed on cardiomyocytes (16). Of note, TF pathway inhibitor is known to bind heparin-sulfated proteoglycans on various cell types, thereby contributing to cell stabilization. A reduction in myocardial TF expression is likely to produce an overall decrease in the number of TF molecules localized, via its modified intracellular cysteine residue (17), to lipid rafts—membrane formations involved in regulation of actin cytoskeleton (18).

It was previously reported (19) that the spatial expression of membrane-bound TF corresponded to that of force-transmitting cytoskeletal proteins in actin-rich membrane areas of in vitro cultured epithelial cells. In line with the co-localization of TF with other structural proteins is the finding that the myocardial TF expression correlated with the ejection fraction in our patients. Data from animal knockout experiments and clinical findings on patients with DCM suggest that defects in cytoskeletal and junctional proteins may cause cardiac dilation, contributing to the development of cardiac pump failure (8). A reduced TF protein content of the left ventricular myocardium has been described for patients with sepsis. It has been suspected that the reduced TF expression by cardiomyocytes may contribute to cardiac failure during sepsis (5). The reason for the reduction of TF expression during sepsis is thus far unknown (20).

Differences in medication with beta-blocker, diuretics, and digitalis were present between our patient groups. More patients with DCM were treated with these drugs than patients with normal or mildly reduced EF (Tables 2, 3). It is not known to date whether medication with beta-blocker, diuretics, or digitalis influences TF gene expression. However, clinical and experimental studies suggest that statins and angiotensin-converting enzyme (ACE) inhibitors reduce the TF expression in human monocytes (21- 22) and endothelial cells (23- 24). No differences in the frequency of statin or ACE inhibitor administration were observed between our patient groups. Thus, we can exclude that a difference in TF gene expression between our patients groups is related to differences in therapy with statins or ACE. Nevertheless, it should be emphasized that the mechanisms and regulation pathways of TF gene expression in cardiomyocytes are poorly understood.

Here we also report the presence of asHTF in cardiomyocytes and endothelial cells of the human myocardium. Alternatively spliced human tissue factor has been demonstrated to circulate in the blood of healthy human beings, being able to induce factor Xa generation (10). Acellular TF has also been reported to be present in blood of patients undergoing coronary artery bypass grafting (25). In these patients, the fluid-phase form of TF failed to support thrombin generation, whereas full-length TF present in microparticles in the blood promoted thrombin generation (25). Acellular TF may bind to TF pathway inhibitor, thereby modulating blood thrombogenicity. Leukocytes have been suggested as a source for asHTF in blood, as these cells are able to synthesize asHTF (10). Whether endothelial cells contribute to the presence of asHTF in the blood under pathologic conditions is unknown. In comparison with full-length TF, only a limited amount of asHTF protein was found to be present in the right ventricular septum. The quantification of myocardial TF expression showed that both full-length TF and asHTF were reduced in patients with DCM. In addition, the expression of full-length TF in myocardial tissues correlated well with myocardial contractility, pointing to an association of transmembrane TF expression and myocardial function. A physiological role for asHTF, which is a soluble protein, has thus far not been identified for the heart. In addition, the correlation between the soluble isoform of TF and EF was only weak. Whether asHTF is expressed to a larger extent within the myocard of the atrium and/or the reduction of asHTF in DCM contributes to dysfunction of the myocardial muscle remains a subject of further investigation.

In summary, the TF expression was down-regulated in the myocardium of DCM patients. The reduction in TF expression and its dislocation from the Z-bands into the cytosol of the cardiomyocytes may adversely affect the cell-to-cell contact stability, thereby contributing to ventricular dilation and cardiac dysfunction in DCM.

The authors acknowledge Prof. Dr. Körfer, Director, Clinic of Thoracic and Cardiovascular Surgery, Herzzentrum Nordrhein-Westfalen, Bad Oeynhausen, Germany, for providing some of the explanted DCM hearts. The authors thank Dr. S. Schwartz, Department of Haematology, Charité-Universitätsmedizin Berlin, Germany, for technical support on light microscopy and digital photography, and Michael Piorkowski for support in the manuscript preparation.