In vivo Remote Delivery of DNA Encoding for Hypoxia‐inducible Factor 1 Alpha Reduces Myocardial Infarct Size

Abstract

We tested if remote gene delivery of hypoxia‐inducible factor 1 alpha (HIF‐1α) protected hearts against induced ischemia, hypothesizing that gene delivery into skeletal muscle may lead to secretion of proteins with actions elsewhere.

Murine quadriceps muscles were transfected with DNA encoding for human HIF‐1α, which resulted in a local, but lasting expression (mRNA and protein, where the latter had nuclear localization). Subjection of isolated hearts to global ischemia and reperfusion 1, 4, and 8 weeks after gene delivery resulted in infarct size reduction (p < 0.05). Supporting that this was due to paracrine effects, HL‐1 cells treated with conditioned media from cells transfected with HIF‐1α or serum from HIF‐1α‐treated mice were protected against H₂O₂‐induced cell death (p < 0.05, respectively). The latter protection was reduced when a heme oxygenase activity blocker was used. Taqman low‐density array of 47 HIF‐1α‐regulated genes at the treatment site showed nine specific upregulations (p < 0.05). Of the corresponding proteins, PDGF‐B and adrenomedullin were upregulated in the heart. HIF‐1α treatment induced an increased vascularization of the heart and skeletal muscle.

In conclusion, remote delivery of DNAfor HIF‐1α was cardioprotective, represented by consistent infarct size reduction, which may be due to release of paracrine factors from the transfected muscle.

Introduction

Ischemic heart disease is the major cause of morbidity and mortality in the Western world. Revascularization of ischemic myocardium is important for its survival, and is mainly achieved through thrombolysis, percutanous coronary interventions, and coronary artery bypass grafting. However, in many patients, anatomical variations and the extent of coronary artery disease precludes revascularization. Thus, new therapeutic options are warranted for those patient populations.

Gene delivery of pro‐angiogenic, pro‐survival factors may provide an opportunity to rescue ischemic cardiomyocytes. Although several growth factors such as vascular endothelial growth factor A (VEGF‐A) and fibroblast growth factor 2 (FGF‐2) injected directly into ischemic tissue have presented positive results in preclinical studies, the first clinical studies showed little or no clinical gain in patients. ¹ , ² One reason maybe that growth factors such as VEGF‐A need cofactors to induce vessel formation.

Hypoxia‐inducible factor 1 alpha (HIF‐1α) is the oxygen‐sensitive subunit of the transcription factor HIF‐1 which conveys adaptive responses to hypoxia/ischemia. ³ In normoxia, HIF‐1α protein is degraded, while upon hypoxia, HIF‐1α is stabilized, heterodimerizes with HIF‐1β, and transactivates genes. HIF‐1α regulates genes involved in angiogenesis, vascular tone, oxygen transport, glycolysis, iron metabolism, cell survival, and proliferation. ³

HIF‐1α may potentially be therapeutic in the treatment of myocardial ischemia. ⁴ , ⁵ In a rabbit model of hind limb ischemia, locally injected DNA encoding for HIF‐1α significantly improved perfusion and capillary density. ⁶ Intramyocardial injection of naked DNA encoding for HIF‐1α enhanced angiogenesis and reduced infarct size in a rat model of in vivo myocardial infarction with permanent occlusion. ⁷ The aim of our study was to deliver DNA encoding for human HIF‐1α ⁸ to skeletal muscle of mice to test if it can cause an increased cardiac tolerance to ischemia‐reperfusion injury. We chose to deliver HIF‐1α to the skeletal muscle because 1) skeletal muscle can produce and secrete proteins efficiently ⁹ , ¹⁰ and 2) it provides an easy access for delivery which is clinically attractive. We hypothesized that secretory protein products coming from genes regulated by HIF‐1α may enter the circulation, reach the heart, and increase cardiac tolerance against ischemia. The possibility that these downstream factors could be heme oxygenase‐1 (HMOX‐1), adrenomedullin, and/or PDGF‐B was explored. To address the potential danger of a whole body angiogenic effect, angiogenesis was studied in the transfected and contralateral skeletal muscle in addition to the heart.

Methods

Plasmid DNA

The pCEP4/HIF‐1α, deriving from human HIF‐1α cDNA sequence downstream of a cytomegalovirus promoter was purchased from ATCC, Johns Hopkins Special Collection (Baltimore, MD, USA). ⁸

Animals

Male C57BL/6 mice (25–30 g) were used in the experiments. Animals were handled according to the Guide for the Care and Use of Laboratory Animals published by US National Institutes of Health and the study was approved by the lo cal ethics committee for animal research.

Gene delivery

Animals were anesthetized with Equithesin (35 mg pentobarbital and 153 mg chloral hydrate per kilogram of the animal) before gene delivery. The right hind limb was shaved, and 15 μg of naked DNA encoding for human HIF‐1α was injected into the right quadriceps muscle in a total volume of 50 μL saline solution. Uptake of the delivered DNA was enhanced through electroporation of the injected muscle with 10 trains of 1,000 Hz bipolar pulses at ±100 V amplitude, a length of 200 μs per phase and a current of 50 mA. In sham animals, an equivalent volume of saline was injected followed by the same protocol of electroporation. The electroporator and software were developed by Iacob Mathisen and Inovio. ⁹ , ¹⁰ LacZ‐staining was used to determine the localization of DNA after delivery of the reporter gene β‐gal as previously described. ¹¹

Isolated heart perfusion

One, 4, or 8 weeks later (n= 7–9 in sham‐ and HIF‐1α‐treated groups at the respective time‐points), animals were anesthetized with pentobarbital (60 mg/kg), and hearts were isolated and Langendorff‐perfused as described in detail elsewhere. ¹² Constant pressure perfusion (55 cm H₂o) with Krebs‐Henseleit buffer was used. A balloon was inserted into the left ventricle for determination of systolic (LVSP) and end‐diastolic (LVEDP) pressures and heart rate, while developed pressure (LVDP = LVSP – LVEDP) was calculated. Coronary flow was measured with an ultrasound probe (Transonic Systems, Ithaca, NY, USA). Data were continuously collected into the computer program PharmLab (AstraZeneca, Gothenburg,Sweden). After 20 minutes of stabilization, 40 minutes global ischemia was induced by clamping the inflow tubing followed by 60 minutes of reperfusion. At the end of reperfusion, hearts were harvested, sectioned, and stained with 1% TTC solution at 37°C for 20 minutes, then fixed in 4% paraformaldehyde for 1 hour. Digital images were taken of both sides of the sections and infarct size was measured by computerized planimetry using Photoshop 7.0 software (Adobe, San Jose, CA, USA). The average of one heart was used for statistics. Samples of treated and contralateral muscle, spleen, and plasma were harvested for immunoblot, immunohistochemistry, and RNA extractions. An additional series of tissues (including serum) were sampled 1 week after gene delivery, where hearts were not subjected to ischemia and reperfusion (n = 9 in each group).

Culture and transfection of HL‐1 cardiomyocytes

HL‐1 immortalized cardiomyocytes were a gift from Dr. William Claycomb (Louisiana State University, New Orleans, LA, USA). Cells were seeded on gelatine/fibronectin‐coated 6‐well plates at a density of 5 × 10⁵ cells/well and cultured in Claycomb medium with the supplementation described elsewhere. ¹³ Cells were transiently transfected with 1.6 μg of DNA for HIF‐1α or empty vector using Lipofectamine 2,000 (Invitrogen). To see if protection with HIF‐1α was due to paracrine effects, either conditioned media from transfected cells (9–10 experiments from empty vector/HIF‐transfected cells) or serum from HIF‐1α‐ or sham‐treated animals (six experiments each) were collected and applied on naïve cells. Forty‐four hours after transfection or 2 hours of serum incubation, cells were subjected to 300 μM H₂o₂ for 4 hours. To test if HMOX‐1 was a downstream candidate for HIF‐induced protection, the HMOX blocker zinc deuteroporphyrin 2,4‐bis glycol (ZnBG, Porphyrin Products, 30 mg/kg i.p. ¹⁴ ) was used in connection with HIF treatment, the serum collected 3 days later, and applied on HL‐1 cells with H₂O₂‐stimulation, Cell viability was determined using trypan blue exclusion assay and a total of 400 cells were evaluated under the microscope by a blinded observer. Due to variations in cell death induced by H₂O₂ (13–62%), results were related to H₂O₂‐treated, empty vector‐transfected cells in every single experiment, thus setting H₂O₂‐treated cells to 100% and the rest of the samples are relative to it.

RNA Extraction

Total RNA was isolated using the RNeasy Mini Kit (QIAGEN Inc.) with an additional phenol‐chloroform (Sigma, St. Louis, MO, USA) extraction step and in‐column DNase treatment (QIAGEN, Hilden, Germany). The quantity of RNA was determined with spectrophotometer. RNA integrity was assessed with Bioanalyzer 2100.

cDNA Synthesis

One microgram of RNA was reverse transcribed using random hexamers for priming (3 minutes at 70°C) followed by a modified First Strand cDNA Synthesis Protocol with Superscript II (Invitrogen) and RNasin (Promega) enzymes (10 minutes at 25°C, 50 minutes at 42°C, and 4 minutes at 94°C).

Real‐time PCR

Oligos were designed with Primer Express 2.0 software and custom made (Applied Biosystems) on the basis of published cDNA sequence (Genebank #U22431). Taqman probe for human HIF‐1α were designed to span exon‐exon junctions and to recognize only human HIF‐1α, which was dry‐tested with BLAST against the RefSeq database. The primers were 5′TATGTGGATAGTGATATGGTCAATGAA and 3′ATTGGGATATAGGGAGCTAACATCTC and the probe was 5′AAGAACCCATTTTCTACTCAGGACACAGATTTAGACTT. 18S rRNA was used as endogenous control.

PCR‐reactions took place in 96‐well plates using 2 μL of cDNA, 12.5 μL of Universal Master Mix (Applied Biosystems, Foster City, CA, USA), primers, and probe at a final concentration of 900 nM and 200 nM, respectively, in a total volume of 25 μL. The PCR reaction had the standard amplification scheme: one cycle of 2 minutes at 50°C (AmpErase UNG activation), one cycle of 10 minutes at 95°C (Gold AmpliTaq activation, AmpErase UNG inactivation), followed by 40 cycles of denaturation for 15 seconds at 95°C and annealing/extension for 1 minute at 60°C in a ABI 7900 HT Sequence Detection System (Applied Biosystems). Human left atrial samples collected at the end of cardioplegia during open heart surgery were used as positive controls after informed consent was gained from patients (Eur J Cardio Thor Surg 2008; in press).

Taqman low‐density array

We selected 47 genes (Table S2), around 30 of which were known to be HIF‐1α‐regulated and the rest possibly influenced to determine their expression 1 week after HIF‐1α‐ or sham delivery (n= 7–9). Twenty nanograms cDNA was loaded in each port along with 50 μL Universal Master Mix (Applied Biosystems) and RNase‐free water to a total volume of 100 μL. For both realtime PCR and Taqman low‐density array samples were run in duplicates. Gene expression relative to 18S rRNA was calculated following the comparative C_T method (User Bulletin #2; Applied Biosystems).

Immunoblot

Whole cell protein extracts were prepared as previously. ¹² Thirty micrograms of protein extracts/lane were separated on polyacrylamide gel followed by transfer to a nitrocellulose membrane (Amersham, Chalfont St. Giles, UK). Membranes were blocked with 3% bovine serum albumin and incubated with goat adrenomedullin (ADM), the ADM receptors goat adrenomedullin receptor (ADMR), and rabbit calcitonin receptor‐like receptor (CRLR) (all Santa Cruz, Santa Cruz, CA, USA), rabbit platelet‐derived growth factor B (PDGF‐B, Abeam) antibodies in 1:1,000 overnight. The day after they were incubated with species‐specific horseradish peroxidase‐conjugated secondary antibodies and developed using the ECL‐kit (Amersham). Optical density of bands was measured with ImageJ software (NIH, Bethesda, MD, USA) and related to the optical density of Ponceau‐staining used as loading control. ¹²

Immunohistochemistry

Muscles and hearts were embedded in OCT. Twelve‐micrometer sections were fixed with 4% paraformaldehyde, and preincubated with 0.1% Triton X‐100 in PBS. Incubation with primary antibody (HIF‐1α, 1:50, goat polyclonal, Santa Cruz Biotechnology; CD31 1:50, rat monoclonal, BD Pharmingen; a‐actin, 1:50, rabbit polyclonal, Abeam, Cambridge, UK) was performed overnight at 4°C and secondary antibody (AlexaFluor488 and AlexaFluor568; Molecular Probes) at room temperature. Sections were treated with Hoechst 33342 for nuclear staining.

Vessel formation

To quantify if HIF‐1α induced vessel growth, a scoring system for counting CD31 signals was made based on the printouts of 200× magnifications. Small dots or patches (probably corresponding to capillaries) were counted separately from longer and thicker signals in the longitudinal view (probably corresponding to arterioles, arteries). Capillaries were counted in numbers, and longer vessels in centimeters. For α‐actin, a different scoring system was used based on different signal density and pattern. Signals between 1–2 cm in length and less than 0.5 cm in diameter or smaller were counted numerically. Larger signals were measured and grouped: (1) length over 2 cm with (2) either less, or (3) more than 0.5 cm diameter, where for the latter the diameter was measured in addition. Random, nonoverlapping view‐fields were processed by a blinded person. For CD31,10–14 random view‐fields were counted in 3–4 samples of right muscle, left muscle, and heart per group 8 weeks after DNA or sham‐injections. For α‐actin, 10 random view‐fields were evaluated for hearts only (n= 4 for each group).

Additional mice were HIF‐1α‐ or sham‐pretreated (n= 3 per group), and 8 weeks later reanesthetized with pentobarbital (50 mg/kg) with concomitant injection of heparin (500IU). Hearts were excised and the aorta cannulated and retrogradely perfused with Krebs‐Henseleit buffer to wash out and dilate vessels. From a sidearm in the aortic cannula, monomer‐mix of a plastic replica and corrosion kit (Batson's No. 17, Polysciences Inc., Warrington, PA, USA) was injected to fill up the coronary circulation. Vascular casts were solidified in PBS at 4°C for 3–4 hours. Maceration solution was added at 50°C for 3–4 hours. Casts were imaged with an Olympus AX70 microscope at 40× original magnification, and signal quantified in Adobe Photoshop CS2 (total area of casts) and ImageJ (total number of pixels). Pixel density/cast area was plotted for statistical analysis.

Statistics

One‐way ANOVA was used to compare infarct sizes and vascular casts between HIF‐1α‐ and sham‐treated groups. ANOVA for repeated measurements was applied to follow hemodynamic changes in Langendorff experiments. The nonparametric Mann‐Whitney U‐test was used to evaluate cell death and gene expression data from Taqman low‐density array, where a non‐Gaussian distribution was assumed. A statistician calculated statistics on the CD31 and α‐actin dataset. The capillary and vessel data were log‐transformed and then analyzed by the same linear mixed model, using the lme function in R. ¹⁵ Fixed effects were included for sites and for combinations of treatment group and site, while random effects were assumed for animals and for sites inside animals. We assumed a model with variances that differ between sites (Appendix S1 and S2). Values are presented either as individual data + mean (infarct size, cell death) or mean ± SD (expression data, casts). Differences were considered significant when p < 0.05.

Results

Evaluation of gene delivery method and localization of delivered plasmid over time

There was no mortality due to gene delivery (injections, electroporation).

Gene delivery was evaluated by LacZ staining. This showed a very local and superficial distribution in the injected muscle ( Figure 1A ). These findings were verified by real‐time PCR. There was no expression of human HIF‐1α in hearts or untreated muscles 1 week after gene delivery, nor in any of these organs in the sham‐treated group ( Figure 1B ). In addition, we could not detect any human HIF‐1α mRNA in the kidney, lung, spleen, or liver (data not shown). Similarly, we found human HIF‐1α expression 4 and 8 weeks after gene delivery solely in treated muscles of the HIF‐1α‐treated animals ( Figures 1C and D ). The gene expression was reduced with time.

Figure 1.

(A) LacZ staining after gene transfer of β‐galactosidase followed by electroporation, with longitudinal‐ and cross‐sections of skeletal muscle. The pictures show a very local distribution of β‐galactosidase (blue) in treated quadriceps muscles (3 left) not present in shams (right). (B), (C), and (D) show expression of mRNA (in log2) for human HIF‐1α in right and left quadriceps muscles and hearts from HIF‐1α‐ and sham‐treated mice 1 (C), 4 (C), or 8 (D) weeks after injection of HIF‐1α DNA or saline followed by electroporation (n= 6–8 organs in each bar graph; real‐time PCR). Note that human HIF‐1α mRNA was detected only in HIF‐1α‐treated right quadriceps muscles and in human heart samples (positive control). H = heart; huH = human heart; ML = left, untreated muscle; MR = right, treated muscle. Data are presented as mean ± SD.

Evaluation of HIF‐1α protein expression by immunohistochemistry in heart, right and left muscles of both sham, and HIF‐1α‐treated animals 1 week after gene delivery showed a similar pattern: at the site of injection HIF‐1α‐treated animals expressed HIF‐1α protein abundantly, while other tissues did not express any ( Figure 2A ). A similar expression pattern was found 4 (data not shown) and 8 weeks after gene delivery. The expression of HIF‐1α in the treated muscle was nuclear, which was verified by Hoechst 33342 co‐localization ( Figure 2B ) and remained nuclear up to 8 weeks after gene delivery ( Figure 2C ).

Figure 2.

(A) HIF‐1α protein expression by immunofluorescence in treated and untreated quadriceps muscles and hearts after gene delivery of HIF‐1α or sham. One week after injections, HIF‐1α was expressed in treated right quadriceps muscles, but not in other organs or in shams. (B) High power magnifications show that the HIF‐1α protein colocalizes with the nuclear stain Hoechst 33342 in cross‐ and longitudinal sections. (C) Eight weeks after injections, HIF‐1α protein expression was restricted to a small area of the electroporated muscle, but clearly present and colocalized with the nuclear marker Hoechst 33342. Original magnifications: 400× for panel A, 1000× for panel B, and 200× for panel C

Effects of HIF‐1α gene delivery on cardiomyocyte survival ex vivo and in vitro

Hearts were isolated, retrogradely perfused, and subjected to induced global ischemia and reperfusion serially after gene‐ or sham delivery. Infarct size 1 week after electroporation was 54%± 13% (mean ± SD) in shams, and reduced to 34%± 6% in HIF‐1α‐treated animals (p < 0.005; Figure 3A ). Infarct size 4 weeks later was 57%± 7% in shams and 38%± 11% in treated animals (p < 0.005), while after 8 weeks it was 48%± 8% in shams and 36%± 8% in HIF‐1α‐treated animals (p < 0.05; Figures 3B and C ).

Figure 3.

Myocardial infarct size in mice (A) 1, (B) 4, and (C) 8 weeks after delivery of HIF‐1α to the skeletal muscle or sham‐treatment (n= 7–9 per group). Hearts were isolated and Langendorff‐perfused with global ischemia and reperfusion. Representative images of TTC stained hearts are shown where infarctions are unstained (white) and viable myocardium red. (D) Cell death (trypan blue) in naïve HL‐1 cells after treatment with conditioned media from HL‐1 cells transfected with HIF‐1α or empty vector and injured with H₂O₂, (E) Cell death after treatment of na'ive HL‐1 cells with serum from sham or HIF‐1α‐transfected mice with or without the addition of the HMOX‐1 blocker ZnBG before H₂O₂‐induced injury. Data are presented as individual data + mean. * denotes p < 0.05, + denotes p < 0.005.

Heart function was evaluated through an intraventricular balloon. There was no difference in left ventricular function evaluated by left ventricular developed pressure, left ventricular end‐diastolic pressure, heart rate, or coronary flow at any time‐point after gene delivery (Table S2).

To confirm the hypothesis that the infarct‐sparing effect of HIF‐1α was due to paracrine factors, HL‐1 cells were transfected with HIF‐1α or empty vector, and conditioned media was applied on naïve cells before they were subjected to H₂O₂‐induced injury. Conditioned media from HIF‐1α‐transfected cells reduced cell death (p < 0.05; Figure 3D ). Similarly, serum from HIF‐1α‐transfected mice applied on HL‐1 cells reduced cell death (p < 0.005; Figure 3E ). This reduction was diminished when mice received the HMOX blocker ZnBG, though it remained reduced compared to the effect of serum from sham‐treated animals (p < 0.05, Figure 3E ).

Gene transcription induced by HIF‐1α

Transcriptional activity at the site of HIF‐1α gene delivery was explored by Taqman low‐density array for 47 selected genes (see list in Table S1). As HIF‐1α mRNA expression was highest 1 week after gene delivery, we used this time‐point for Taqman low‐density array studies. We found that 9 out of 47 genes were significantly upregulated in the HIF‐1α‐treated group versus the sham‐treated group, and they are shown in Figure 4 .

Figure 4.

Taqman low density array of RNA extracted from quadriceps muscles of HIF‐1α‐ or sham‐treated animals 1 week after gene delivery (n= 7–9 per group). Of a total of 47 genes (Table S1), the genes above were all significantly upregulated (p < 0.05). TGF‐β1 = transforming growth factor‐β1; MCP‐1 = monocyte chemoattractant protein‐1; Car9 = carbonic anhydrase‐9; HMOX‐1 = heme oxygenase‐1; IGF‐2 = insulin‐like growth factor 2; p21 = cyclin‐dependent kinase inhibitor 1A; and PDGF‐B = platelet derived growth factor B (all p < 0.05 compared to sham).

To see if this gene regulation was specific to the site of HIF‐1α‐delivery, we screened hearts and untreated, contralateral muscles. Downregulation of angiopoietin‐1 and phosphofruktokinase‐L (PFK‐L) were found in the untreated, contralateral muscle while monocyte chemoattractant protein‐1 (MCP‐1), plasminogen activator inhibitor‐1, and transforming growth factor‐β3 were downregulated in the heart ( Figures 5A and B ).

Figure 5.

Taqman low‐density array was performed on RNA from (A) contralateral muscles and (B) hearts of the mice in Figure 4 (n= 6 in each group). The genes depicted in the bar graph were all significantly downregulated (p < 0.05). PFK‐L = phosphofruktokinase‐L; MCP‐1 = monocyte chemoattractant protein‐1; PAI‐1 = plasminogen activator inhibitor‐1; and TGFβ3 = transforming growth factory.

Upregulated proteins after HIF‐1α gene delivery

Protein expression of PDGF‐B, ADM and its receptors ADMR and CRLR were studied by immunoblotting 1 week after gene delivery and electroporation. As platelets are degraded in the spleen, spleen samples were included to study PDGF‐B.

PDGF‐B was upregulated in the treated skeletal muscle, spleen and the heart of HIF‐1α‐treated animals (p < 0.05). We could not evaluate plasma samples for PDGF‐B because of heavy background, probably due to platelet‐derived contamination. ADM expression was not influenced in the treated skeletal muscle or plasma after HIF‐1α gene delivery (data not shown), but was upregulated in the heart (p < 0.05; Figure 6 ). This was not accompanied by upregulation of its two receptors ADMR and CRLR (data not shown).

Figure 6.

Immunoblots with antibodies to platelet‐derived growth factor B (PDGF‐B) and adrenomedullin. Representative blots with primary antibody are shown on the upper panel, with Ponceau‐staining for protein loading below. PDGF‐B expression in proteins extracted from sham‐ and HIF‐1α‐treated (A) muscles, (B) spleens, and (C) hearts 1 week after HIF‐1α gene delivery to the skeletal muscle or sham‐treatment. (D) adrenomedullin expression in heart. (E) optical density of the blots (n= 4–7 in each group, mean ± SD). * denotes p < 0.05.

Vascular growth

The endothelial marker CD31 was used to detect possible endothelization by HIF‐1α pretreatment. One week after gene delivery we detected an increased expression of CD31 in the HIF‐1α‐treated right muscles, hearts, and left muscles compared to the same organs in the sham‐treated group (qualitative analysis, not quantified; Figure S1). Eight weeks after gene delivery we quantified the formation of capillaries and vessels. There was a significant increase of vessels, but not capillaries according to our scoring system in the treated as well as untreated, contralateral muscles of HIF‐1α‐treated compared to shams (Appendix S1, Figures 7A and B ). There was no significant difference in either capillary or vessel pattern of CD31 expression in hearts.

Figure 7.

Expression of the endothelial marker CD31 in right and left quadriceps muscles as well as hearts of animals treated with HIF‐1α or sham 8 weeks earlier. (A) representative immunofluorescent staining (original magnification: 200×); (B) quantification corresponding to vessels and to capillaries as described in Methods. H = heart; MR = right, treated muscle; ML = left, untreated muscle (n= 3–4 organs/group, mean ± SD). * denotes p < 0.05.

The question of vessel growth was further pursued in the heart by using α‐actin immunostaining for smooth muscle cell detection. We found no difference for smaller signals or the length of bigger signals. However, the breadth of bigger signals, probably corresponding to larger vessel diameter, was significantly greater in hearts of HIF‐1α‐treated animals than in sham‐treated hearts (p < 0.05; Figure 8 , Appendix S2). When plastic replica casting of the coronary vasculature was performed, increased cardiac vascularization 8 weeks after HIF‐1α treatment was confirmed ( Figures 8C and D ).

Figure 8.

Expression of smooth muscle cell α‐actin in hearts of HIF‐1α‐versus sham‐treated animals 8 weeks after gene delivery. (A) representative immunofluorescent images (original magnification: 200×); (B) quantification of larger vessels as described in Methods (n= 4 per group); (C) representative cast of the coronary vasculature; and (D) quantification thereof (n= 3 per group, mean ± SD). * denotes p < 0.05.

Discussion

In our present study, we demonstrate that remote delivery of DNA encoding for human, constitutively active HIF‐1α protected murine hearts from ischemia and reperfusion. The novel finding with this approach is that the delivered factor was not detectable in the organ of pharmacological action, only at the site of delivery, while the desired cardioprotective effect was achieved. This suggests secondary mechanisms mediating cardioprotection. The potential therapeutic gain of this approach is a long‐lasting cardioprotection with a single injection at a clinically easily accessible site.

We hypothesize that the transfected skeletal muscle secretes factors that can protect the heart by activating downstream cardioprotective mechanisms (or inhibiting deleterious ones), which probably alter cardiac gene expression profile and structure. Secretion of paracrine factors is supported by our findings in cell culture, where either conditioned media from HIF‐1α‐transfected HL‐1 cells or serum from HIF‐1α‐treated animals reduced H₂O₂‐induced cell death. The fact that the HMOX blocker ZnBG diminished this protection indicates that it may be a mediator of remote, HIF‐induced cardioprotection. We found upregulated genes with angiogenic/cardioprotective potential locally at the site of HIF‐1α treatment. The protein products of two of these genes, PDGF‐B and adrenomedullin were upregulated in the heart, making them also potential candidates for underlying the infarct size‐limiting effect of HIF‐1α. The potential of ADM, PDGF‐B, and HMOX‐1 as candidates of HIF‐1α‐induced protection needs to be explored in vivo in the future.

PDGF‐B is a growth factor expressed in a variety of cell‐types with autocrine, paracrine and endocrine effects. ¹⁶ After forming a homodimer or heterodimer with PDGF‐A, the PDGF‐complex exerts a number of actions (mitogenic, chemotactic, and antiapoptotic) via activation of diverse pathways. PDGF‐B is known to phosphorylate and activate cardioprotective chaperons such as αβ‐crystallin and heat shock protein 27 in cardiomyocytes. ¹⁷ Furthermore, locally administered PDGF‐B decreased the extent of myocardial infarction in rats. ¹⁸ Similarly, adrenomedullin is expressed in most cell‐types, ¹⁹ is secretory and known to activate multiple second messenger systems. Adrenomedullin is diuretic, has vasodilatory, antiproliferative, and antiapoptotic effects in the myocardium, and protects against ischemia‐reperfusion injury ²⁰ HMOX‐1 is an inducible cytoprotective enzyme that catabolizes heme to free iron, carbon monoxide, and biliverdin, which is endogenously converted to bilirubin. ²¹ Bilirubin and carbon monoxide may exert cardioprotection through antioxidant, vasodilatory, antiapoptotic, and antiinflammatory actions, while HMOX‐1 itself has been used for cardioprotection. ²¹ , ²² In the present study, mRNA for HMOX‐1 was increased locally in the HIF‐1α‐treated muscle, and inhibition of its action diminished the protection of serum from treated mice on HL‐1 cells. In the present study, the production origin of HMOX‐1, ADM, or PDGF is unlikely to be the heart, as there was no transcriptional upregulation of these genes in the heart. PDGF‐B was also upregulated in the spleen. The spleen is the degradation place of the major PDGF‐B‐secreting platelets, and at the same time, another distant organ. The increased expression of the endothelial marker CD31 and smooth muscle cell α‐actin in the untreated muscle and heart, respectively, further suggests release of growth factors from the transfected muscle into the circulation, leading to vessel formation elsewhere.

HIF‐1α is important for vascular development and is exploited in therapeutic vascularization. ⁴ , ⁵ In a recent study, local HIF‐1α caused therapeutic arteriogenesis in rabbits with hind limb ischemia ²³ and phase I clinical trials have been conducted for critical limb ischemia. ²⁴ The present findings with increased vascularization, and increased CD31 and α‐actin expression concur with previous findings. ⁶ , ⁷ , ²³ PDGF‐B, locally upregulated in the heart, could contribute to the enlargement of vessels via its chemoattractant effect for smooth muscle cells and its being a stabilizer of blood vessel walls. ²⁵ Vessel enlargement was not followed by an increased coronary flow, but this could be due to the limitation of isolated heart perfusion. Whether the induction of vessel growth by HIF‐1α contributes to cardioprotection remains to be elucidated. The finding that H₂O₂‐induced cell death in HL‐1 cells was reduced by treatment with conditioned media from HIF‐1α‐transfected cells or serum from HIF‐1α‐treated animals suggests that vascular growth is not essential for cardiomyocyte protection.

We found an increased vessel growth in the untreated, contralateral quadriceps muscle. HIF‐1α is upregulated in vascular malignancies including von‐Hippel Lindau disease. ²⁶ It is likely that HIF‐1α has a role in the development of other malignant tumors as well, possibly by its growth‐promoting effect or by causing an angiogenic switch crucial for tumor progression. ³ Thus, if we want to use HIF‐1α as a therapeutic agent in the treatment of ischemic diseases, we need to know the systemic effects of HIF‐1α administration, including possible side effects. Most of the studies to date that have used HIF‐1α therapeutically to fight ischemic disorders have investigated only local effects at the site of delivery. ⁶ , ²³ , ²⁴ The increased expression of the endothelial marker CD31 in the untreated quadriceps muscle may be an unwanted side effect of HIF‐1α treatment, potentially reducing the therapeutic applicability of this approach. Thus, we believe that further investigations are necessary to identify downstream targets of the HIF‐1 complex that can be controlled easier, yet still have the desired cardioprotective effects.

The remote administration of HIF‐1α plasmid DNA reduced infarct size after ischemia and reperfusion 1, 4, and 8 weeks after gene delivery. However, the infarct size reduction was not accompanied by improvement of left ventricular function. One possibility for this discrepancy is that the net pharmacodynamic effect of HIF‐1α products on the heart is infarct size reduction without changing hemodynamics. This is supported by a recent study, where mice were pretreated with the HIF‐1α stabilizer cobalt chloride. ²⁷ When isolated hearts were subjected to ischemia and reperfusion, the authors found myocardial infarct size reduction, without influence on hemodynamics. ²⁷ This possibility is further supported by increased cardiac adrenomedullin protein expression in the current study, a factor that is known to have infarct‐sparing effect without a direct influence on inotropic function. ²⁸ The nature of this intriguing finding remains unknown, though.

Conclusion

In summary, our findings indicate that remote delivery of the transcription factor HIF‐1α provided a lasting cardioprotection against ischemia and reperfusion in a mouse Langendorff model, despite the fact that the delivered HIF‐1α was localized only at the site of delivery in the quadriceps muscle. At this site, an increased transcriptional activity hallmarked the functionality of the delivered transcription factor. The protection could be transferred by applying conditioned media from HIF‐1α‐transfected HL‐1 cells or serum from HIF‐1α‐treated animals to naïve HL‐1 cells. In the heart, PDGF‐B and adrenomedullin were upregulated at protein level, where increased vascularization was found following HIF‐1α treatment. The increased expression of the endothelial marker CD31 in the contralateral skeletal muscle indicates vascular growth, and also warns about a possible limitation of this treatment.

Supporting information

Supporting info item

Supporting info item

Supporting info item

Supporting info item

Supporting info item