Amphotrophic rhEPO (pBabe-EPO) retrovirus-containing packaging cell line PT67 was used for retroviral vector production 15 . Twenty-four hour filtered (0.45 μm) culture virus-containing supernants were stored at -80°C for future experiments. The presence and titer of rhEPO-retroviral vector were assayed by infecting BALB/3T3 (ATCC number CCL-163) mouse fibroblast cell line and a CFU (colony-forming unit) number determination.

Cardiac fibroblasts and cardiomyocytes were isolated from 1–2 day old neonatal Sprague-Dawley rats (Harlan Lab, Jerusalem, Israel) as previously described 16 . We used a 30 minute pre-plating procedure to obtain cardiac fibroblasts and to reduce the number of non-myocyte cells in cardiomyocyte culture. The purity of obtained cardiac fibroblast culture was confirmed microscopically by characteristic cell morphology. Neonatal rat cardiomyocytes were used for oxidative stress studies and cardiac fibroblasts were used for transduction and transplantation studies.

BALB/3T3 (ATCC number CCL-163) mouse fibroblast cell line and isolated cardiac fibroblasts were used for transduction experiments. RhEPO retrovirus-containing supernants were first incubated at room temperature with a Superfect cationic transfection reagent (QIAGEN, Westburg Ltd., Be'er-Sheva, Israel) to a final concentration of 10 μg/ml for 15 minutes 17 . Finally, the viral stock was added to 70–80% confluent BALB/3T3 or cardiac fibroblast culture. After 24 h incubation at 37°C, puromycin selection was applied (2.5 μg/ml). Transduction efficiency was approximately 80%, as evaluated qualitatively by relative confluence determination (before and after puromycin selection). Twenty-four hour culture supernants were collected for EPO assessment. Supernants were stored at -20°C for future experiments.

After 6 days in culture, isolated rat neonatal cardiomyocytes were pretreated with supernants collected from rhEPO-transduced or non-transduced (for control) cardiac fibroblasts for 24 h, after which H₂O₂ was directly added to culture medium to a 150 μM final concentration for 60 minutes, and the degree of apoptosis was tested.

Male Sprague-Dawley rats (approximately 250 g; Harlan Lab, Jerusalem, Israel) were subjected to acute myocardial infarction (AMI) induction by permanent coronary artery occlusion, as previously described 19 .

RhEPO-transduced and non-transduced fibroblast transplantations were performed 7 days after AMI. After 10 days in culture (7 days after transduction, 3 passages), cardiac fibroblasts were prepared for transplantation by trypsinization with 0.25% Trypsin-EDTA solution for 1 minute at 37°C. The rhEPO expression level in prepared cells was 7.0 ± 0.23 mIU/24 h/10⁶ cells. Following centrifugation (1,500 rpm, 10 minutes at room temperature) the cells were re-suspended in saline. The final cell density for implantation was 1 × 10⁶ cells/100 μl. Rats were anesthetized and the chest was opened under sterile conditions. The infarcted area was identified visually by the surface scar and wall motion akinesis. Rats were randomized and selected either for injection of 1 × 10⁶ rhEPO-transduced cells, 1 × 10⁶ non-transduced cells, or saline (n = 5 in each group). All injections were made into the center of the scar. After the injection, the air was expelled from the chest, and the surgical incision was sutured closed 19 20 .

Transthoracic echocardiography was performed on all animals before transplantation (baseline echocardiogram) and 5 and 9 weeks later (4 and 8 weeks post-transplantation, respectively). Echocardiograms were performed with a commercially available echocardiography system (Sonos 7500; Philips) equipped with 12.5 MHz phased-array transducer (Hewlett-Packard, Andover, MA, USA) as previously described 21 22 . All measurements were averaged over three consecutive cardiac cycles and were performed by an experienced technician who was blinded to the treatment group.

For evaluation of possible systemic effects of rhEPO produced from transplanted cells, hematocrit measurements were performed before or 1 and 4 weeks after transplantation.

Four weeks after transplantation, hearts were arrested with 15% KCl, perfused with formaldehyde 4% (15 mmHg) for 20 minutes, sectioned into 3–4 transverse slices, and parallel to the atrioventricular ring. Each slice was fixed with 10% buffered formalin, embedded in paraffin, and sectioned into 5 μm slices. Serial sections were stained with the following agents: anti-α-smooth muscle actin monoclonal antibodies (Sigma-Aldrich, Rehovot, Israel) to localize pericytes and arterioles for neovascularization evaluation; anti-active caspase-3 polyclonal antibodies (Biocare Medical, Walnut Creek, CA, USA) for apoptosis detection; anti-rhEPO monoclonal antibodies (R&D Systems, Minneapolis, MN, USA) to localize rhEPO production and distribution; lectin (Bandeirea simplicifolia agglutinin BS-1; Sigma-Aldrich, Rehovot, Israel) to localize vascular endothelium; and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) using Apoptag Peroxidase In situ Apoptosis Kit (S7101, Chemicon, Temecula, CA, USA) to assess apoptosis. The degree of neovascularization was evaluated by capillary blood vessel and arteriole counts in the infarct border zone (at least three different fields were counted). Apoptosis evaluation was made by active caspase-3 or TUNEL-positive cell counts in the infarct border zone (at least five different fields were counted). Co-staining was made with haematoxylin or methyl green for TUNEL.

After perfusion pressure fixation (15 mmHg), slides stained with Masson's trichrome (Dexmore, Israel) were used to assess left ventricular (LV) remodeling by morhpometric analysis, as previously described 23 . The following parameters were measured: LV maximal diameter (mm), average wall thickness (mm; averaged from three measurements of septum thickness), average scar thickness (mm; averaged from three measurements of scar thickness), relative scar thickness (average scar thickness/average wall thickness), LV muscle area (mm²; including the septum), LV cavity area (mm²), whole LV area (mm²), infarct expansion index ([LV cavity area/Whole LV area]/Relative scar thickness), epicardial scar length (mm), and endocardial scar length (mm).

All values are given as means ± standard error of the mean (SEM) from at least three independent experiments. RhEPO activity results were compared with Student's t-test for unpaired data. Cardiomyocyte or general cell apoptosis results and blood vessel counts were compared by one-way ANOVA with Bonferroni post test.

In the echocardiography study, because each animal was used as its own control, changes between baseline and 4 weeks in the control and treated groups were assessed with paired t-tests. Comparisons of the changes from baseline to 5 and 9 weeks in the control and treatment groups were made with repeated-measure two-way ANOVA. The ANOVA model included control versus treatment and baseline versus 5 and 9 weeks as factors, as well as the interaction between the two factors. GraphPad Prism version 4.00 for Windows (GraphPad Software, San Diego, California, USA) was used for analysis.

Comparison of morphometry parameters and hematocrit measurements among treatment groups was analyzed by one-way ANOVA with Bonferroni post test. All tests were performed using GraphPad Prism Version 4.0. P < 0.05 was considered statistically significant.

To test retroviral ability for stable expression of biologically active rhEPO, the initial transduction experiments were carried out using BALB/3T3 mouse fibroblast cell line. Table 1 describes the activity of EPO produced from transduced cells. ELISA was used to evaluate EPO concentrations, and its biological activity was determined by proliferation of the EPO-sensitive human erythroleukemia cell line TF-1. The activity induced by recombinant human EPO (Eprex) was used as a standard in this assay. Produced EPO protein was found to be 35 times more biologically active than recombinant protein (Eprex; Table 1, p = 0.0006). The rhEPO production level in transduced cultures was 11.5 ± 0.8 mIU/24 h/10⁶ cells.

To generate viable cardiac cell culture producing rhEPO for subsequent transplantation, we used rat neonatal cardiac fibroblasts. EPO activity parameters of transduced cardiac fibroblasts are presented in Table 1. The produced protein was found to be 12 times more active than recombinant protein (Eprex; p = 0.03). The rhEPO production level in transduced cells was 14.2 ± 3.1 mIU/24 h/10⁶ cells.

The rate of apoptosis was significantly lower in cultured cardiomyocytes treated with supernants derived from rhEPO-tranduced fibroblasts. To determine the protective effect of rhEPO produced from genetically modified cardiac fibroblasts on cardiomyocyte apoptosis, we used an H₂O₂-induced oxidative stress model 24 . Isolated rat neonatal cardiomyocytes were treated with 150 μM H₂O₂ for 60 minutes, with or without 24 h preconditioning with supernants containing rhEPO (12 mIU/ml) from transduced cardiac fibroblasts. The degree of apoptosis was assayed by Annexin V/PI staining and morphologically evaluated using a DAPI nuclear stain. FACS analysis of Annexin V/PI staining showed that the percentage of apoptotic cells was significantly lower in cardiomyocytes pretreated with rhEPO-containing supernants compared with control (Figure 1A): 17.8 ± 2.3% versus 32.1 ± 3.7% apoptotic cells (p < 0.05).

Morphological changes in cells that were exposed to oxidative stress were evaluated after DAPI staining. Apoptotic nuclei counts showed a significant reduction in cell death in cells treated with rhEPO-containing supernants, compared to untreated cells: 13.5 ± 0.6% versus 19.0 ± 0.7%, p < 0.001 (Figure 1B).

There was no difference among the groups in the levels of hematocrit at all time points (Table 2). Four weeks after cell transplantation, heart tissue sections were stained with anti-rhEPO monoclonal antibodies to detect rhEPO expression from implanted cells. Human kidney tissue was used as positive control (Figure 2A). RhEPO was detected only in hearts treated with transduced cardiac fibroblasts 5 weeks post MI (Figure 2B–D). The staining was identified in all animals treated with transduced cells (n = 5), towards the center of the infarct zone.

The degree of vascularization was evaluated by counting capillary blood vessels in the infarct border zones using α-smooth muscle actin and lectin stainings for vascular endothelium localization. Compared with non-transduced fibroblasts and saline, implanted rhEPO-producing fibroblasts enhanced angiogenesis, as shown in both α-smooth muscle actin (Figure 3A; 15.27 ± 0.37 versus 9.93 ± 0.97 versus 9.57 ± 0.72 vessels/mm², respectively, p < 0.05), and lectin stained sections (Figure 3B; 70.1 ± 8.5 versus 47.5 ± 5.9 versus 20.7 ± 3.6 vessels/mm², respectively, p < 0.05).

Apoptosis in infarcted hearts was detected by active caspase-3 staining with polyclonal antibodies detecting only the active (effector) form of caspase-3 (Figure 4A) 25 26 . Human tonsil tissue was used as a positive control (Figure 4A, a). RhEPO-transduced cell transplantation resulted in a significant decrease in the number of active-caspase-3 positive cells (24 ± 6 versus 176 ± 20 versus 110 ± 27 cells per mm² in saline and non-transduced cell injections, p < 0.05; Figure 4C, a). The staining revealed mainly nuclear localization of active caspase-3 (Figure 4A) 26 . Apoptosis was also evaluated by commercially available TUNEL kit (Figure 4B). Rat mammary gland tissue served as a positive control (Figure 4B, a). Similar to active caspase-3 staining, TUNEL staining showed that rhEPO-transduced cell transplantation resulted in a significant decrease in the number of apoptotic cells (118 ± 6 versus 182 ± 12 versus 157 ± 10 cells per mm² in saline and non-transduced cell injections, p < 0.05; Figure 4C, b). These results suggest that local rhEPO expression prevented cardiac cell apoptosis in vivo.

To examine potential cardioprotective effects of rhEPO-producing fibroblasts in vivo, we used a rat model of AMI, which resulted in progressive LV remodeling and dysfunction 19 . Overall, 45 rats were included in the final transplantation study. Rats were randomized into three treatment groups (n = 15 in each group): transduced rhEPO-producing cells, non-transduced cells, and saline injection. Ten animals from each group (five additional animals in each group were sacrificed one month after transplantation for immunohistochemical examination) were assessed with echocardiography.

Serial echocardiography tests showed that EPO-transduced and non-transduced cell injection did not reduce LV remodeling and dysfunction 5 and 9 weeks post MI (Table 3). At 5 weeks, there was a trend towards more scar thinning in the EPO-treated group (anterior wall diastolic thickness, p = 0.07; LV end diastolic dimension, p = 0.08).

Morhpometric analysis of the infarcted hearts showed that EPO-transduced cell injection also fails to prevent infarct expansion and scar thinning (Table 4). For major parameters (average scar thickness, LV cavity area, whole LV area and expansion index), EPO-transduced cells showed significant adverse remodeling compared to non-transduced cells with a magnitude similar to saline-treated animals. Furthermore, non-transduced cell injection attenuated infarct expansion and preserved LV area, compared to saline-treated animals.

Cell engraftment and survival was not quantified. Bromodeoxyuridine (BrdU) or other cellular markers were not used to avoid interference with cell capacity to colonize and proliferate. Such interference could cause a decrease in the proposed EPO effect.

Retroviral transduction itself can potentially alter the phenotype of the transduced cells. However, the expression profile of transduced cells did not exhibit any negative effects when the transduced cell culture supernants were added to cardiomyocyte culture in vitro.

In order to avoid possible systemic interference, we did not apply an immunosuppression protocol. However, EPO detection 1 month after cell transplantation suggests that immune response, if evoked, is limited in magnitude. The EPO tissue level could not be quantified due to technical constraints for detection of predicted low levels of the protein.