INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

IGF-I AND IGF-II are 7- to 8-kDa structural homologs of insulin with multifunctional activities that affect many anabolic processes including promotion of cell growth, inhibition of apoptosis, and induction of cell differentiation (40). The IGFs in the circulation are primarily produced in the liver, but many cell types can synthesize IGFs for local use (9). IGF expression is stimulated in pathological situations in association with trauma including ischemia and wound healing, suggesting that these factors have important repair functions (9, 23).

There is considerable evidence to support a role for the IGFs in attenuating myocardial dysfunction due to ischemia. In experimental models, IGFs are overexpressed by cardiac myocytes in response to coronary occlusion (9, 25). Furthermore, both in vitro and in vivo studies support a role for the IGFs in promoting myocyte survival and improving myocardial function (6, 8, 21, 26, 40). A transgenic model that permits specific IGF-I overexpression in the postnatal heart conclusively demonstrated an important role for this factor in protecting myocytes from death after ischemic injury (29, 30, 46). Data obtained from this transgenic model and other experimental results also suggest that myocytes are able to proliferate postnatally in the presence of elevated IGF-I levels (3, 39). These results have led to increased interest in the therapeutic potential of the IGFs in the clinic (11, 12, 28).

Another area of considerable interest in cardiovascular biology is the use of angiogenesis therapy to rescue ischemic tissues (18, 31). In addition to their cardioprotective activity, some studies have suggested that the IGFs can stimulate angiogenesis (27, 38, 41). However, the proangiogenic activity appears less potent than that of other angiogenic factors (15, 38, 41) and in some cases is due to upregulated expression of VEGF in other cell types that subsequently activate endothelial cells (1, 7, 37, 45).

In the current study, we demonstrate that IGF-I and IGF-II are direct, potent chemotactic agents for human endothelial cells. Furthermore, we determined that laminin is a critical permissive extracellular matrix (ECM) protein for this chemotactic activity. Consistent with these findings we determined that, in the presence of laminin, these factors are also capable of stimulating vascular outgrowth in an ex vivo aortic ring model and an in vivo angiogenesis model. Adenoviral gene therapy vectors expressing IGF-I and IGF-II were able to efficiently transduce cardiomyocytes, with consequent IGF expression and secretion. The gene therapy vector-mediated expression of IGF-I effectively protected cardiomyocytes from apoptosis induced by ischemia-reoxygenation, ceramide, and heat shock and enhanced angiogenesis in vivo. These results provide additional support for local overexpression of the IGFs in the treatment of ischemic heart diseases.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell and reagents. The AE1-2a (S8) cell line expressing adenoviral E1 and E2a genes (derived from A549 cells) was generated at GTI (14) and routinely cultured in Richter's culture media (Invitrogen, Grand Island, NY) with 5% FBS (BioWhittaker, Walkersville, MD). Human umbilical vein endothelial cells (HUVECs, passage 3; Clonetics, San Diego, CA) were grown in 75-cm² tissue culture flasks (Falcon Primaria) in endothelial basal medium-2 (EBM-2; Clonetics) supplemented with 5% FBS, 0.5 µg/ml human recombinant epidermal growth factor, 5 µg/ml insulin, 1 µg/ml human recombinant FGF, 50 mg/ml gentamicin, and 50 µg/ml amphotericin B at 37°C in a 95% air-5% CO₂ humidified atmosphere. Cells were subcultured by aspiration of the growth medium followed by a 30-s rinse with a solution of 0.5 mM EDTA-0.25 mg/ml trypsin. Matrigel, laminin 1, type IV collagen, and fibronectin were obtained from Becton Dickinson (Franklin Lakes, NJ).

RT-PCR amplification of human IGF cDNAs. Adult and fetal human liver poly(A)+ RNA was purchased from ClonTech (Palo Alto, CA). RT-PCR was performed with the GeneAmp RNA PCR kit (Perkin-Elmer, Foster City, CA), and 1 µg of either human or fetal poly(A)+ RNA was used per reaction. Two primer sets for IGF-I and IGF-II were designed to amplify a fragment of cDNA ranging in size from 663 bp (IGF-I) and 562 bp (IGF-II). Each sense primer had a EcoR1 restriction site incorporated, and each antisense primer had an XbaI restriction site incorporated for subsequent cloning into the EcoR1 and XbaI sites of the plasmid, pcDNA3.1 (Invitrogen, Carlsbad, CA). In addition, the antisense primer also contained a SpeI site upstream of the XbaI site for easier orientation analysis in a subsequent cloning step into the adenoviral shuttle plasmid. A stop codon was also incorporated into the antisense primer for IGF-II upstream of the SpeI site. The human IGF cDNAs were assembled and were then cloned into the adenoviral shuttle plasmid pAvS6alx, which contains the left side of the adenoviral vector genome and a lox site (10) to create pAvlxIGF-I and pAvlxIGF-II. The IGF cDNAs were sequenced to ensure fidelity during PCR amplification.

Generation of recombinant adenoviral vectors. The IGF cDNAs were incorporated into an adenoviral vector with the lox recombination three-plasmid transfection system. AE1-2a cells (A549 cells stably expressing the E1/E2a regions under a dexamethasone inducible promoter; Ref. 14) were cultured in Richter's medium containing 5% heat-inactivated FBS. Transient transfections of the AE1-2a cells were performed to generate the recombinant adenoviral vector encoding the IGF transgenes (10, 14). Either Not1-digested pAvlxIGF-I or pAvlxIGF-II (0.5 µg), Cre-encoding plasmid (0.5 µg) and ClaI-digested pSQ3 DNA (1 µg; contains the right end of the adenoviral vector genome and a lox site) were transfected into AE1-2a cells with the Lipofectamine-Plus reagent system (Life Technologies, Rockville, MD). The AE1-2a cells were incubated with the Lipofectamine reagent-DNA precipitate at 37°C for 16 h. The precipitate was removed, and the monolayers were washed with PBS. Richter's medium containing 5% FBS and 0.3 µM dexamethazone was added to the cells. The cells were incubated at 37°C for ~5-7 days. The conditioned medium and cells were then collected, frozen, and thawed three times, and the cell debris was pelleted. The conditioned medium was then used to infect a fresh plate of dexamethazone-induced AE1-2a cells. A cytopathic effect was observed in the cells ~12-15 days after transfection. The virus was amplified in 15-cm dishes of dexamethazone-induced AE1-2a cells. The recombinant Av3IGF vectors were purified, and large-scale preparation seed lots were prepared.

Northern blot analysis. AE1-2a cells were transduced with Av3null, Av3IGFI, or Av3IGFII at the indicated doses. After 48 h, the cell pellets were obtained and total RNA was isolated with the RNAzol B (Tel-Test) extraction method. Ten-microgram aliquots of total RNA were applied per lane to two 1.2% agarose-formaldehyde gels. RNA was transferred to nylon membranes and prehybridized in 0.5 M NaPO₄, 1 mM EDTA, 0.5% BSA, 7% SDS at 65°C for 2 h. The membranes were then hybridized with either a 686-bp ³²P-labeled IGF-I or a 585-bp ³²P-labeled IGF II cDNA probe at 65°C and washed in SSC- and SDS-containing buffers at 65°C after standard protocols. Membranes were exposed to film for 1 h at room temperature.

Quantitation of IGFs by ELISA. AE1-2a cells were transduced with Av3null, Av3IGFI, or Av3IGFII at the indicated doses. After 48 h, the conditioned medium was collected, aliquoted, and stored frozen at 80°C. A quantitative measurement of IGF-I and IGF-II in the conditioned medium was determined with nonextraction IGF-I and IGF-II ELISA kits (Diagnostic Systems Laboratories, Webster, TX) using the suggested kit protocols. The level of IGF-I and IGF-II is reported as the mean ± SD concentration in nanograms per milliliter.

Cell proliferation assay. AE1-2a cells were transduced with Av3null, Av3IGFI, or Av3IGFII at the indicated doses. After 48 h, the conditioned medium was collected, aliquoted, and stored frozen at 80°C. On day 1, mouse myoblast C₂C₁₂ cells were seeded at 2,000 cells/well in a 96-well dish in DMEM containing 10% FBS. On day 2, 75 µl of conditioned medium was added to each of four wells in a total volume of 100 µl in DMEM containing 1% FBS. The cells were incubated at 37°C for 48 h. The number of viable cells per well was determined with the CellTiter 96 aqueous nonradioactive cell proliferation assay (Promega, Madison, WI). In this assay, 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfphenyl)-2H-tetrazolium salt is bioreduced by dehydrogenase enzymes found in metabolically active cells into formazan that is soluble in tissue culture medium and can be measured at OD₄₉₀. The level of formazan in the tissue culture medium was determined according to the suggested protocol, and the data are reported as means ± SD absorbance at 490 nm from four independent wells.

Cell migration. Transwell (Corning Scientific Products, Acton, MA) inserts were coated with Matrigel (1:50 dilution; Becton Dickinson, Bedford, MA) and incubated overnight at 4°C . HUVECs were seeded onto the top well of the inserts (10⁴-10⁵ cells) in a total volume of 0.4 ml of serum-free DMEM containing 0.5% BSA. Recombinant FGF-2, IGF-I, and IGF-II (R&D Systems, Minneapolis, MN) at the concentrations indicated were added to the bottom wells of the inserts, and the plates were incubated for 6 h at 37°C and 5% CO₂. After the incubation period, the cells on the top well were removed with a cotton swab and washed extensively with Tris-buffered saline. HUVECs that had transversed to the bottom part of the Transwell membrane were quantitated by measuring intracellular acid phosphatase activity. Para-nitrophenyl phosphate (PNPP) was dissolved in 0.1 M sodium acetate, pH 5.5 containing 0.1% Triton X-100. This assay buffer (0.5 ml) was applied to the Transwells and incubated for 1 h at 37°C. The cleaved PNPP is colorless at pH 5.5. One hundred microliters of the reaction mixture was removed from the Transwell insert and mixed with fifty microliters of 1 M Tris pH 8.0 in a 96-well plate. The absorbance of the cleaved PNPP was determined at 405 nm.

Cell survival assay. The ability of IGF-I, IGF-II, and FGF-2 to increase the survival of HUVECs in serum-depleted medium was determined by microscopy and quantitated by protein determination. HUVECs (10⁶ cells) were plated onto 12-well tissue culture dishes and allowed to adhere for 18 h in medium containing 10% serum. The medium was replaced by serum-free medium containing 0.5% BSA alone or with increasing concentrations of IGF-I, IGF-II, or FGF-2. Cells were incubated at 37°C for 2 days, after which plates were washed three times with serum-free medium to remove all nonadherent or loosely adherent cells or cell fragments remaining on the plate. The status of the cells was then checked under a microscope. The remaining cellular protein was determined by bicinchoninic acid protein determination.

Apoptosis assay. Ischemia in cardiac myocyte cultures was induced by adding serum- and glucose-free DMEM (Life Technologies) and incubating the cells in an incubator perfused with 95% N₂-5% CO₂. After the indicated time periods, the cells were removed from the hypoxic incubator, reoxygenated with complete growth medium containing glucose and serum, and placed at 37°C in a 95% air-5% CO₂ humidified atmosphere. Cardiac myocytes were heat shocked by immersion into a temperature-controlled water bath at 42°C for 30 min. After exposure to heat shock, the cells were incubated at 37°C in a CO₂ incubator for 6 h and then analyzed for apoptosis induction. Cells were treated with 10 µM C6-ceramide (N-hexanoylsphingosine) overnight (Biomol Research Laboratories, Plymouth Meeting, PA) in serum-free medium before assessment for apoptosis. Cardiac myocytes were examined for morphological features of apoptosis (chromatin condensation and fragmentation) by fluorescence microscopy with acridine orange and ethidium bromide as described previously (17). Briefly, apoptosis was assessed by the addition of 50 µl of a 1:1 stock solution of ethidium bromide-acridine orange (Sigma, St. Louis, MO) to 1 ml of culture medium on the cells growing in 35-cm culture dishes. A coverslip was attached, and the morphological features of apoptosis were monitored by fluorescence microscopy with a microscope equipped with a FITC filter at ×600. At least 200 cells from randomly selected fields were counted and quantitated according to the following formula: %apoptotic cells = no. of apoptotic cells/total no. of cells counted × 100. A similar profile of ischemia-reoxygenation-induced apoptosis was also observed with either a Annexin-V-FLUOS staining kit or a DNA fragmentation photometric ELISA kit (both from Roche Molecular Biochemicals, Mannheim, Germany) according to the manufacturer's instructions.

Aortic ring explant assay. This assay was described in detail elsewhere (38). Briefly, Matrigel was thawed overnight on ice in the refrigerator. Matrigel (200 µl/well) was placed in each well of a 48-well plate. The dish was placed in an incubator at 37°C to allow the Matrigel to solidify. Aorta from Sprague-Dawley rats was harvested and cleaned in Hanks' solution under sterile conditions. The aorta was sliced into 48 rings and placed into individual wells of a 48-well plate containing the Matrigel. Another 50 µl of Matrigel was added on top, the plate was incubated for 30 min at 37°C, and serum-free growth medium (200 µl) containing the test factors was added to the wells. Six rings were used for each treatment. After 7 days the rings were stained with 1 mM calcein AM (Molecular Probes, Eugene, OR) and examined under confocal microscope. Ten optical sections were taken for each sample and collected for computer image analysis.

Adenovirus transduction of cardiac myocytes. Transduction of cardiac myocytes with adenoviral vectors was optimized with the adenoviral vector encoding the marker gene nuclear -galactosidase (Av3nBg). Briefly, rat cardiac myocytes, grown in 35-mm Primaria tissue culture dishes, were transduced with adenoviral vectors at 10-500 multiplicity of infection (MOI) in 0.5 ml of serum-free medium for 6 h at 37°C, and then 1 ml of complete growth medium was added for overnight incubation. Human cardiac myocytes were transduced as described above with the exception that cells were transduced with Av3nBg in the presence of a 1:1,000 µl dilution of Fugene 6.

Matrigel plug assay and analysis. Liquid Matrigel mixed with test substances was prepared beforehand and preloaded into 1.0-ml tuberulin syringes with G27 needles. Animals (female nude mice, 6-8 wk old) were anesthetized under isofluorane, and 0.5 ml of the undiluted Matrigel containing 250 ng/ml of FGF-2 and cells that were transduced with adenoviral vectors was injected subcutaneously into the caudal portion of the midline (1 injection/animal) with a G27 needle. The Matrigel will rapidly form a solid gel that persists for over 7 days. On day 7, all animals were euthanized with CO₂ and Matrigel plugs were harvested for histological analysis and hemoglobin measurement. Matrigel plugs were dissected and fixed in 10% neutral buffered formalin for 2 h at 4°C. The tissue was then processed and embedded in paraffin. Five-micrometer-thick paraffin sections were cut. Three sections, at least 50 µm apart, were stained with anti-smooth muscle -actin monoclonal antibody (Pharmingen, San Diego, CA) followed by standard immunoperoxidase detection (Vector Laboratories, Burlingame, CA). Positively stained cells were visualized with 3,3'-diaminobenzidene (Sigma), and slides were counterstained with hematoxylin. For quantitative analysis, Matrigel plugs were snapped frozen in a dry ice-alcohol bath and dried overnight. The dry weight was recorded, and the plug was rehydrated with 0.5 ml of 0.5% Tween 20. After homogenization, the plug was centrifuged at 14,000 rpm for 30 min (Eppendorf 5415C; Brinkmann Instruments, Westbury, NY). Supernatant fluid was collected, and the absorbance was read at 405 nm and converted to micrograms of hemoglobin per milligram of Matrigel with a standard curve generated with hemoglobin standards (Sigma).

Data analysis. ANOVA was performed, and comparison of two group means was subsequently made with the Fisher's protected least-square difference test. Student's t-test was used to determine statistical significance when only two groups were compared. Data analysis and graph generation were performed with Prism (GraphPad Software, San Diego, CA), and statistical analysis was performed with InStat (GraphPad Software). A P value of <0.05 was considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

IGFs are potent stimulants of human endothelial cell migration. Although studies have demonstrated that IGF-I stimulates angiogenesis in complex assays (38, 41), the basis for this activity remains unclear. Endothelial cell migration is an important component of the proangiogenic response, and we tested IGF-I and IGF-II for chemotactic activity. As shown in Fig. 1A, we demonstrated that both of the IGFs are potent stimulants of human endothelial cell migration. The chemotactic activity of the IGFs was comparable to FGF-2-induced activity (Fig. 1A). To better understand this finding, we next compared IGF-I-elicited endothelial cell migration on Matrigel versus fibronectin. As shown in Fig. 1, B and C, IGF-I only stimulated migration of human endothelial cells when the cells were seeded on Matrigel. In contrast, FGF-2 was chemotactic for endothelial cells seeded on either fibronectin or Matrigel (Fig. 1, B and C). These results indicate that IGFs are potent stimulants of endothelial cell migration but this activity is highly dependent on the substratum.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Human endothelial cell migration in response to IGF-I, IGF-II, or FGF-2. Transwell inserts were coated either with Matrigel (1:50 dilution) or fibronectin (10 µg/ml) and incubated overnight at 4°C. Human umbilical vein endothelial cells (HUVECs) were seeded onto the top well of the inserts (104-105 cells) in a total volume of 0.4 ml of serum-free DMEM containing 0.5% BSA. FGF-2 or IGF-I at the concentrations indicated was added to the bottom wells of the inserts, and the plates were incubated for 6 h at 37°C. After the incubation period, the cells on the top of the insert membrane were ablated with a cotton swab and washed extensively with Tris-buffered saline. HUVECs on the bottom part of the Transwell membrane were quantitated by measuring acid phosphatase activity. A: plate coated with Matrigel (1:50 dilution) and treated with FGF-2 (50 ng/ml), IGF-I (50 ng/ml), or IGF-II (50 ng/ml). B: plate coated with Matrigel (1:50 dilution) and treated with IGF-I (10 and 20 ng/ml) or FGF-2 (10 and 20 ng/ml). C: plate coated with fibronectin (10 µg/ml) and treated with IGF-I (10 and 20 ng/ml) or FGF-2 (10 and 20 ng/ml). All experiments were done at least twice. * Significantly different from nontreated control cells (P < 0.05).

Endothelial cell adhesion to laminin is essential for chemotactic activity of IGF-I. Matrigel is a complex basement membrane mixture that is >95% laminin and type IV collagen. We therefore tested whether purified laminin or type IV collagen is the ECM component permissive for the robust endothelial cell migration response to the IGFs. As shown in Fig. 2B, IGF-I was a potent stimulant of cell migration only when HUVECs were seeded on laminin. Indeed, we observed a consistently better response to IGF-I when the cells were seeded on laminin versus Matrigel, and the chemotactic activity of IGF-I was better than that of FGF-2 under these conditions. In contrast, IGF-I had no effect on endothelial cell migration when seeded on type IV collagen (Fig. 2C). These results indicate that laminin is a critical ECM component in Matrigel that is permissive for the stimulation of cell migration by IGF-I.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Migration of HUVECs on laminin and type IV collagen in response to IGF-I and FGF-2. Transwell inserts were coated either with laminin or type IV collagen (10 µg/ml) and incubated overnight at 4°C. HUVECs were seeded onto the top well of the inserts, and the migration assay was performed as described in Fig. 1. A: plate coated with Matrigel (1:50 dilution) and treated with IGF-I (10 ng/ml) or FGF-2 (10 ng/ml). B: plate coated with laminin (10 µg/ml) and treated with IGF-I (10 ng/ml) or FGF-2 (10 ng/ml). C: plate coated with type IV collagen (10 µg/ml) and treated with IGF-I (10 ng/ml) or FGF-2 (10 ng/ml). All experiments were done at least twice. * Significantly different from nontreated control cells (P < 0.05).

IGF-I and IGF-II stimulate capillary outgrowth in organ ring culture assay. The functional importance of IGF-mediated endothelial cell migration in the presence of laminin-containing substratum was further evaluated in an explant model of aortic sprouting. IGF-I, IGF-II, FGF-2, and VEGF were compared for their ability to elicit vascular outgrowth from explanted rat aortic rings. As shown in Fig. 3A, 7 days after explant, a significant angiogenic response was observed in the presence of growth factors. Under confocal microscopy, 10 optical sections from each ring were taken and collated for quantitative analysis. The results, shown in Fig. 3B, clearly demonstrated that IGF-I and IGF-II elicited the most robust sprouting from the aortic explants. Interestingly, the increased angiogenic response to IGF-I was statistically better than VEGF- or FGF-2-treated aortic vessels. These observations are consistent with the cell migration results and indicate that in the presence of an appropriate substratum, IGFs can elicit a potent angiogenic response.

View larger version (86K): [in this window] [in a new window]   Fig. 3.   IGF-I and IGF-II are potent inducers of capillary formation in aortic ring explants. A: photomicrograph of aortic ring cultures. Two hundred microliters of serum-free growth medium was added with the test factors indicated. Organ rings were incubated for 7 days and stained with calcein AM, and pictures were taken with a phase contrast microscope. B: quantitation of aortic ring assay. Six rings were used for each treatment. Ten optical sections were taken 50 µm apart and collected for computer image analysis. Pixel intensity of the calcein AM-stained cells was quantitated with a Bio-Rad image analysis system. Data are represented as means ± SD; n = 6. * Significantly different from untreated control (P < 0.01); ** significantly different from FGF-2 (P < 0.05).

Generation of recombinant adenoviral vectors encoding human IGF-I and IGF-II. Adenoviral vectors (Av3IGFI and Av3IGFII) containing the respective human IGF cDNAs were constructed, and Southern blot analysis demonstrated the expected hybridization pattern for both vectors (data not shown). Expression of human IGF RNA and protein were analyzed by Northern and ELISA analysis. Northern blot analysis of total RNA isolated from transduced AE1-2a cells demonstrated expression of a 1-kb IGF-I and a 0.9-kb IGF-II RNA (Fig. 4A). IGF RNAs were not detected in nontransduced or Av3null-transduced AE1-2a cells. Conditioned medium from transduced cells was also analyzed for human IGF-I and IGF-II protein by ELISA with anti-human IGF-I and IGF-II monoclonal antibodies. The human IGF proteins were detected in the conditioned medium of IGF vector-transduced cells and not in mock- or Av3null-transduced cells (Fig. 4B). The functional activity of the expressed human IGF-I protein was further verified by demonstration of mitogenic activity on C₂C₁₂ myocytes (results not shown). We conclude that these adenoviral vectors can generate robust expression and secretion of IGFs.

View larger version (44K): [in this window] [in a new window]   Fig. 4.   Characterization of adenoviral vector-mediated expression of human IGF-I and IGF-II. A: nontransduced AE1-2a cells (mock), Av3null-transduced cells at 1,000 multiplicity of infection (MOI) (null), or Av3IGFI (IGF-I)- or Av3IGFII (IGF-II)-transduced cells at 100 or 1,000 MOI were harvested 48 h after infection. Total RNA was isolated, transferred to a nylon membrane, and probed with IGF-I- and IGF-II-specific cDNA probes as described in METHODS. RNA size markers are indicated. B: conditioned medium from nontransduced AE1-2a cells, Av3null-transduced cells at 1,000 MOI, or Av3IGFI-transduced or Av3IGFII-transduced cells at 100 MOI was harvested 48 h after infection, aliquoted, and stored frozen. An aliquot of conditioned medium from each sample was analyzed for human IGF protein by ELISA as described in METHODS. Values represent means ± SD of triplicate samples. At least 3 independent experiments were performed for the ELISAs with similar results. * Significantly different from conditioned medium from nontransfected cells (P < 0.05).

Adenoviral vectors containing IGF transgenes can efficiently transduce cardiac myocytes and induce secretion of IGF-I and IGF-II. We optimized transduction of rat cardiac myocytes with a third-generation adenoviral vector encoding -galactosidase (Table 1). In these conditions, rat cardiac myocytes were transduced with Av3null, Av3IGFI, or Av3IGFII and the effects on IGF-I and IGF-II protein released into the cell medium or contained within the cell pellet were assessed. A negligible level of IGF-I protein was detected in the supernatant or cell pellets of nontransfected cardiac myocytes, and transduction with 500 MOI Av3null did not alter this profile (Fig. 5A). Transduction of cardiac myocytes with 50, 200, or 500 MOI Av3IGFI resulted in a dose-dependent increase in IGF-I protein expression after either 2 or 5 days of transduction (Fig. 5A). Five days after transduction of cells with 200 or 500 MOI, ELISA analysis revealed 1.4 ± 0.5 and 3.5 ± 0.3 mg/ml IGF-I, respectively. On both days 2 and 5, there was a greater proportion of IGF-I protein in the supernatants compared with the cell pellet in cells transduced with Av3IGFI. Similar experiments were performed with cardiac myocytes transduced with Av3IGFII. In contrast to IGF-I, nontransfected myocytes produced a higher basal amount of IGF-II protein (Fig. 5B). However, transduction of cardiac myocytes with Av3IGFII also resulted in a significant increase in IGF-II protein detected in the supernatant at both days 2 and 5 after infection. Treatment of cardiac myocytes with Av3null had no effect on IGF-II protein levels. Together, these data indicate that transduction of cardiac myocytes with Av3IGFI or Av3IGFII results in substantial transgene expression that is both time and dose dependent. Furthermore, the expressed IGF-I and IGF-II proteins are efficiently secreted.

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Adenoviral vectors can efficiently transduce cardiac myocytes and stimulate the expression of IGF-I and IGF-II. A: IGF-I protein expression is both time and dose dependent after transduction of rat cardiac myocytes with Av3IGFI. Rat cardiac myocytes, grown in 6-well plates, were transduced with the indicated concentrations of Av3IGFI (in MOI). Supernatant and cell pellet fractions were analyzed at 2 or 5 days after infection. B: IGF-II protein expression is both time and dose dependent after transduction of rat cardiac myocytes with Av3IGFII. Rat cardiac myocytes, grown in 6-well plates, were transduced with the indicated concentrations of Av3IGFII (in MOI). Supernatant and cell pellet fractions were analyzed at 2 or 5 days after infection. Values shown are means ± SE from 3-4 experiments. * Significantly different from nontransfected cells (P < 0.05).

IGF-I and IGF-II expressed via gene therapy vectors protect cardiac myocytes from apoptosis. Previous studies demonstrated that administration of IGF-I protein protects against apoptosis in several cell types including cardiac myocytes (36, 44). Pretreatment of cardiac myocytes with increasing concentrations of either IGF-I or IGF-II protein for 30 min before reoxygenation resulted in a significant attenuation in apoptosis (57 ± 1% and 67 ± 8% less apoptotic cells than untreated cells; n = 3; Fig. 6). With DNA fragmentation ELISA as another measure of apoptosis, 100 ng/ml IGF-I reduced ischemia-reoxygenation-induced apoptosis to a similar level (61 + 3%; n = 3) compared with nontreated cells. The IGFs were substantially more effective than VEGF in this activity, and similar results were obtained with human cardiac myocytes (data not shown).

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Inhibition of reoxygenation-induced apoptosis in cardiac myocytes by recombinant IGF-I or IGF-II proteins. Rat cardiac myocytes were exposed to 4 h of ischemia followed by 4 h of reoxygenation (I/R). Cells were exposed to the indicated concentrations of IGF-I or IGF-II during the last 30 min of ischemia before reoxygenation. Apoptosis was monitored by DNA-binding dyes. Results are expressed relative to cells that had no exposure to growth factors. Values shown are means ± SE from 3-4 experiments. * Significantly different from cells exposed to I/R only (P < 0.05).

View larger version (29K): [in this window] [in a new window]   Fig. 7.   Inhibition of cardiac myocyte apoptosis by adenovirus encoding IGF-I and IGF-II. A: rat cardiac myocytes, grown in 30-mm culture dishes, were transduced with the indicated concentrations of Av3IGFI or Av3IGFII (in MOI) or 500 MOI Av3null. After 2 or 5 days, cells were exposed to 4 h of ischemia followed by 4 h of reoxygenation (I/R). B: rat cardiac myocytes were transduced with 500 MOI of Av3null, Av3IGFI, or Av3IGFII. After 5 days, cells were exposed to ischemia (4 h)-reoxygenation (4 h), ceramide (10 µM), or heat shock (42°C, 30 min). A and B: values shown are means ± SE from 3 experiments. * Significantly different from nontransfected cells exposed to the relevant apoptosis inducer (P < 0.05). C: human cardiac myocytes were transduced with 250 or 500 MOI Av3IGFI or 500 MOI Av3null. After 3 days, cells were exposed to 4 h of ischemia followed by 4 h of reoxygenation. Apoptotic cells were determined by fluorescent microscopy as described in METHODS and are shown as a % of the total cells counted. Results are expressed relative to cells that had no exposure to growth factors. Values shown are means ± SE from 3-4 experiments. * Significantly different from cells exposed to I/R only (P < 0.05).

IGFs are potent stimulants of angiogenesis in an in vivo model. We next tested whether gene therapy-mediated overexpression of IGF is also able to modulate angiogenesis in an in vivo Matrigel model. IGF-I and IGF-II vector-transduced cells in Matrigel plugs were implanted in the flank of nude mice, and neovascularization was assessed. As shown in Fig. 8A, IGF-I significantly enhanced neovascularization compared with cells transduced with null vector in the presence of 250 ng/ml FGF-2 as quantified by hemoglobin levels in the Matrigel plug. IGF-II also stimulated angiogenesis but not to as great an extent as IGF-I (Fig. 8A). Platelet-endothelial cell adhesion molecule-1 and CD31 staining were consistent with the quantitative analysis, revealing substantially more blood vessels in the Av3IGFI-treated Matrigel plugs than FGF-2- or Av3null-treated samples (Fig. 8B). We also observed the presence of more vessels larger than capillaries in the Av3IGFI-treated plugs. The size of these large vessels was in the range of 50-100 µm. These vessels also stained positive for -smooth muscle actin (results not shown). Our data indicate that overexpression of IGFs via an adenoviral vector can substantially enhance blood vessel formation in vivo.

View larger version (75K): [in this window] [in a new window]   Fig. 8.   IGF-I and IGF-II gene therapy vector can enhance in vivo angiogenesis. Matrigel plug model of in vivo angiogenesis was performed in nude mice with the indicated treatments. Angiogenesis was induced by the inclusion of FGF-2 (250 ng/ml) in 450 µl of Matrigel. A: plugs contained 5 × 105 S8 cells previously transduced (48 h) with Av3IGFI, Av3IGFII, Av3null, or mock infection. Matrigel plugs were harvested 7 days after implantation, and Hb concentration was quantified as described in METHODS. Data shown represent means ± SE; n = 6. Av3IGF1 treatment was significantly different from Av3null-treated controls (P < 0.01). B: histological analysis of angiogenesis induced by Av3IGFI. a-f: hematoxylin and eosin-stained cross-sections of Matrigel plug. a: Matrigel alone. b: Av3null. c: Av3IGFl. d: FGF-2. e: FGF-2 + Av3null. f: FGF-2 + Av3IGFI. g-l: CD31-stained cross-sections of Matrigel plug. g: Matrigel alone. h: Av3null. i: Av3IGFI. j: FGF-2. k: FGF-2 + Av3null. l: FGF-2 + Av3IGFI. Small arrows indicate capillary formation in the Matrigel plug; large arrows indicate large vessel formation. Scale bar = 50 µM.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Apoptosis of cardiomyocytes contributes to myocardial dysfunction in various cardiac diseases including ischemia-reperfusion injury and heart failure (16, 22, 26, 33). Previous studies demonstrated that IGF-I protein suppresses apoptosis in cultured cardiomyocytes (13, 36, 44). In the present study, adenoviral transduction of cardiomyocytes resulted in a dose- and time-dependent expression of IGF-I or IGF-II protein and a concomitant inhibition of apoptosis. Nearly complete protection against ischemia-reoxygenation stress can be achieved with either IGF-I or IGF-II in these cells. It was particularly interesting to note that at 5 days after transduction, effective protection against the apoptotic signal was observed, even at a MOI of 50 when only a small increase in secreted IGF was detected. The reason for this robust protective effect in the presence of a modest increase in the IGFs is unclear but suggests that even a small but sustained increase in IGF concentration can effectively protect the entire cardiomyocyte population against cell death. Our results are consistent with other reports in the literature indicating that systemic administration of IGF-I protein or overexpression of IGF-I in the heart increases cardiomyocyte proliferation, reduces apoptosis, and improves cardiac function in animal models of ischemic injury and heart failure (6, 8, 21, 26, 28, 30, 39, 40). Furthermore, recent transgenic studies indicate that IGF-I can normalize hemodynamic and cellular parameters in a model of dilated cardiomyopathy, suggesting that IGF-I can also improve cardiomyocyte performance and that congestive heart failure may benefit from early interventional IGF-I treatment (46).

Although the data supporting a positive role for IGF in the heart are now quite convincing, its therapeutic use is complicated by the finding that circulating IGF-I may also play a role in tumor growth (5, 32, 47) and proliferative retinopathy (42). It has also been reported that multiple intravenous administration of high-dose recombinant IGF-I to humans is associated with substantial side effects (19). However, tissue-targeted overexpression of IGF-I via gene transfer augments the local production of IGF-I but does not increase the serum levels of IGF-I (2). Systemic delivery of IGFs is further complicated by the presence of multiple IGF-binding proteins in the serum (9). Thus a gene therapy-mediated approach for local overexpression of IGFs in the heart appears to have substantial advantage over systemic IGF protein delivery.

A proangiogenic role for IGF-I in the ischemic heart was proposed previously (24). In the current study, we demonstrated that IGF-I and IGF-II are both able to stimulate endothelial cell migration when the cells are seeded on Matrigel or the Matrigel component laminin. Under these conditions, IGF-I is equivalent or more potent than FGF-2 in stimulating both cell migration and capillary outgrowth from explanted aortic rings. This contrasts with previous reports indicating that the IGFs are less potent than other angiogenic factors (15, 38, 41) or act secondarily by the upregulation of VEGF expression in tumor cells (1, 7, 45). We speculate that this apparent discrepancy is caused by the specific condition of the assays. For example, in the ex vivo angiogenesis assay, Nicosia et al. (38) embedded the aortic rings in interstitial collagen gel whereas we used the IGF-permissive ECM Matrigel. Unlike that observed for the protection of cardiomyocyte apoptosis, IGF-II is less active than IGF-I in all the angiogenesis assays. Assuming that IGF-II is working through the IGF-I receptor, one explanation for this is that only a small increase in IGF-I receptor activity is sufficient to protect cardiomyocytes but a more robust response is required for angiogenesis on endothelial cells. A recent study using a pig model of postinfarct repair reported that IGF-I was the most effective molecule in improving myocardial function and also increased neovascularization in the ischemic heart (26). We believe the data presented here represent the first demonstration of the intriguing possibility that the proangiogenic activity of IGF-I requires the ECM protein laminin. Previous studies implicated laminin and laminin fragments as important mediators of angiogenesis (35, 43). Given the ubiquitous nature of laminin as a component of the basement membrane, we speculate that laminin or laminin fragments are likely present in the ischemic heart and could synergize with exogenously provided IGFs to stimulate angiogenesis.

IGF-I has been shown to improve myocardial function in healthy humans as well as in patients with chronic heart failure (11, 12). Furthermore, patients with a higher serum level of IGF-I immediately after an acute myocardial infarction had better myocardial remodeling and ventricular function as well as a positive clinical outcome than patients with lower IGF-I serum (28). These and the preclinical model results discussed above indicate that sustained, local expression of IGFs may represent a viable treatment of ischemic heart disease and heart failure. A variety of gene therapy vectors including adenovirus-associated virus (4), DNA-liposomes (20), and adenoviral vectors (31) can be used depending on the desired level and duration of expression. Given the multiple cellular activities of the IGFs, it remains unclear as to the level and duration of expression that will be required for optimal therapy. Although continuous expression of IGFs may be ideal for its cardioprotective activity, sustained proangiogenesis signaling could eventually be detrimental and the less proangiogenic IGF-II may be a better choice with a sustained expression vector. Controlled gene expression of a long-duration vector such as adenovirus-associated virus, lentiviral vector, or adenoviral vectors deleted of all viral coding sequences (34) is a possible solution.