Prolyl-hydroxylase inhibitor activating hypoxia-inducible transcription factors reduce levels of transplant arteriosclerosis in a murine aortic allograft model

Abstract

OBJECTIVES

The development of transplant arteriosclerosis, the hallmark feature of heart transplant rejection, is associated with a chronic immune response and also influenced by an initial injury to the graft through ischaemia and reperfusion. Hypoxia-inducible transcription factor (HIF)-1 pathway signalling has a protective effect against ischaemia–reperfusion injury and has already been demonstrated to ameliorate allograft nephropathy in previous animal studies. Therefore, the aim of this study was to investigate the effect of stabilization of hypoxia-inducible transcription factors with a prolyl-hydroxylase domain (PHD) inhibitor on transplant arteriosclerosis in an experimental aortic allograft model.

METHODS

MHC-class I mismatched C.B10-H2(b)/LilMcdJ donor thoracic aortas were heterotopically transplanted into the abdominal aorta of BALB/c mice. Donor animals received a single dose of the PHD inhibitor 2-(1-chloro-4-hydroxyisoquinoline-3-carboxamido) acetate (ICA) (40 mg/kg) or vehicle i.p. 4 h before transplantation. Intragraft HIF accumulation after ICA treatment was detected by immunohistochemistry before and after cold ischaemia (n = 5). Grafts were harvested 30 days after transplantation and analysed by histology (n = 7) and immunofluorescence (n = 7). In addition, intragraft mRNA expression for cytokines, adhesion molecules and growth factors was determined on Day 14 (n = 7).

RESULTS

Donor preconditioning with ICA resulted in HIF accumulation in the aorta and induction of the HIF target genes vascular endothelial growth factor and transforming growth factor-beta. Vascular lesions were present in both experimental groups. However, there was significantly reduced intimal proliferation in preconditioned grafts when compared with vehicle controls [intimal proliferation 31.3 ± 8% (ICA) vs 55.3 ± 20% (control), P < 0.01]. In addition, experimental groups revealed a down-regulation of E-selectin (−57%) and MCP1 (−33%) expression after ICA pretreatment compared with controls, going along with decreased T-cell [1.4% CD4+ T-cell infiltration vs 8.4% (control) and 4.9% CD8+ T-cell infiltration vs 10.7% (control)], dendritic cell (0.6% dendritic cells infiltration vs 1.9% infiltration(control)] and macrophage infiltration [4.8% macrophages (ICA) vs 10.9% (control)] within vascular grafts.

CONCLUSIONS

These data of an animal transplant model show that the pharmaceutical activation of HIF with endogenous up-regulation of protective target genes leads to adaptation of the graft to low oxygen-saturation and hereby attenuates the development of transplant arteriosclerosis and allograft injury. Pharmaceutical inhibition of PHDs appears to be a very attractive strategy for organ preservation that deserves further clinical evaluation.

INTRODUCTION

Cardiac transplantation has evolved as a standard therapy for end-stage heart failure over the last decades. Transplant arteriosclerosis, being the hallmark feature of chronic cardiac transplant rejection, is still limiting long-term success after heart transplantation, despite advances in modern immunosuppressive regimens [1]. Recent studies have underscored the fact that both innate and adaptive immune response play a role in several aspects of the pathogenesis of transplant arteriosclerosis, including immune-mediated vascular injury, inflammation of vascular endothelium and ischaemia–reperfusion injury.

In contrast to ordinary arteriosclerosis that usually involves some parts of the artery more severely than other parts and produces prominent lesions, transplant arteriosclerosis is characterized by a diffuse, concentric and progressive thickening of the arterial intima that affects minor as well as major coronary arteries of transplanted cardiac allografts. This progressive narrowing of the vascular lumen of coronary arteries results in sudden or chronic progressive ischaemic damage to the graft with subsequent organ failure [2]. The development of transplant arteriosclerosis is closely associated with chronic immune responses and also influenced by the initial injury to the graft as a consequence of ischaemia and reperfusion. There is now accumulating experimental evidence that ischaemic or hypoxic preconditioning may be able to protect organs against ischaemia [3]. The hypoxia-inducible transcription factor (HIF), an α/β-heterodimer, plays a central role in the organism's adaption to low oxygen concentrations. The β-subunit (HIF-β) is constitutively expressed whereas the regulation of the two alternative HIF α-isoforms, HIF-1α and HIF-2α, is oxygen-dependent [4, 5]. Under normoxic conditions, HIF-α is rapidly degraded and virtually undetectable in most cells, whereas under hypoxic conditions HIF-α accumulates within the cell and forms heterodimers with HIF-β. This complex binds to specific hypoxia response elements as DNA regulatory sequence and subsequently leads to transcription of HIF target genes (Fig. 1). To date, more than 100 HIF target genes have been identified, including genes involved in haematopoiesis [such as erythropoietin (EPO)], in energy metabolism [e.g. glucose transporter 1 (GLUT-1)], in angiogenesis and the vascular tone [e.g. vascular endothelial growth factor (VEGF) and nitric oxide synthases (NOS)], in cell proliferation [such as transforming growth factor-beta (TGFβ)] and in vascular sclerosis [such as macrophage migration inhibitory factor (MIF)]. Normoxic inactivation of HIFs is regulated by oxygen-dependent hydroxylation of specific prolyl-residues by prolyl-hydroxylation domain (PHD) proteins. These enzymes require molecular oxygen as a substrate and 2-oxogluterate as a co-substrate. Therefore, inhibitors of the HIF prolyl-hydroxylases, which stabilize HIF and subsequently activate HIF targets, can be used to mimic the hypoxic response: Pharmaceutical inhibition of PHDs by the application of 2-oxoglutarate (2-OG) analogues such as 2-(1-chloro-4-hydroxyisoquinoline-3-carboxamido) acetate (ICA) can result in HIF accumulation with subsequent activation of tissue-protective genes [6].

Figure 1:

Rapid degradation of HIF-1α by the ubiquitin-proteasome pathway, during normoxia. Under hypoxic conditions, prolyl-hydroxylase activity is attenuated, which decreases the proteasomal degradation of intracellular HIF-1α. 2-OG analogues such as ICA can also result in HIF accumulation. The accumulated HIF-1α heterodimerizes with HIF-1β [aryl hydrocarbon receptor nuclear translocator protein (ARNT)] and translocates into the nucleus. The HIF-1 complex binds to DNA regulatory sequences (HRE), leading to subsequent transcription of target genes including VEGF, EPO and NOS (modified after Xia et al. Toxicol Sci. 2009 [26]). HIF-1α: hypoxia-inducible transcription factor-1α; ICA: 2-(1-chloro-4-hydroxyisoquinoline-3-carboxamido) acetate; HIF-1β: hypoxia-inducible transcription factor-1β; EPO: erythropoietin; HO-1: heme oxygenase-1; PHD: prolyl-hydroxylase domain; HRE: hypoxia response element; CXCL1: chemokine C-X-C ligand 1.

The hypothesis of this study was that HIF activation with a PHD inhibitor has an impact on the development of transplant arteriosclerosis, the hallmark feature of chronic heart rejection. Mouse abdominal aortic allografts were used as the experimental model because they have been shown to represent vascular lesions similar to those observed in human coronary arteries that are affected by transplant arteriosclerosis and allow a precise analysis of the composition of the vascular lesions. We were particularly interested in morphological changes of the allograft and in cytokine and adhesion molecule expression. The results of this study strengthen the concept that donor preconditioning is a very attractive strategy for organ protection that potentially can also improve long-term survival of cardiac transplantation.

MATERIALS AND METHODS

Animals

BALB/cJ mice were originally purchased from Janvier (Le Genest Saint Isle, France). C.B10-H2^b/LilMcdJ mice, a congenic strain harbouring a single MHC-class I mismatch compared with the BALB/cJ strain, were originally purchased from The Jackson Laboratory (Bar Harbor, ME, USA). C.B10-H2^b/LilMcdJ mice were used as donors and BALB/cJ as recipients of the aortic allografts. As the effect of the pretreatment protocol was unclear, we decided to use this particular mouse strain combination (C.B10-H2(b)/LilMcdJ > Balb/c) with a moderate rejection profile in order to detect even subtle differences in rejection kinetics. All 184 mice used in this study were aged between 6 and 8 weeks at the time of experimental use and were bred and maintained at the animal facility of the Preclinical Experimental Animal Center at the University of Erlangen-Nuremberg under specific pathogen-free conditions and treated in accordance with institutional and state guidelines.

Description and administration of the prolyl-hydroxylase domain inhibitor ICA

The inhibitor of the PHDs ICA was locally synthesized at the Department of Chemistry and Pharmacy (Nicolai Burzlaff) according to the structural formula of the substance ‘bicyclic isoquinolinyl inhibitor’ as published by Stubbs et al. [7]. ICA predominantly inhibits the PHDs and has only marginal inhibitory effects on factor inhibiting HIF. ICA was dissolved in 10% dimethyl sulphoxide and 90% phosphate buffered saline. A total volume of 1 ml was administered intraperitoneally as single dose 4 h prior to transplantation containing the weight-adapted dosage of 40 mg/kg body weight.

Hypoxia-inducible transcription factor immunohistochemistry

A prerequisite for testing the potential protective effect of ICA in an experimental mouse aortic allograft model was to test if ICA was able to activate intracellular HIF in aortic grafts. For the experimental set-up of the transplantation model, a procurement timepoint of 4 h prior to transplantation appeared suitable as this timeframe is consistent with a realistic clinical application of donor pretreatment. In addition, experimental data have shown that a 4-h procurement is sufficient to up-regulate HIF target genes in a kidney transplant model [8]. Four hours after donor treatment with ICA, aortic grafts were removed and subjected ex vivo to additional cold ischaemia periods of 0 h (immediate analysis), 2 h and 6 h (n = 5). Paraffin sections (2–4 µm) were dewaxed in xylene and rehydrated in a series of ethanol washes. SuperfrostPlus (Menzel) slides were used. For detection of HIF-1α, monoclonal mouse anti-human HIF-1α antibody (α67; Novus Biologicals) was used as described [8]. Biotinylated secondary anti-mouse antibody (Dako) was used. For signal amplification and visualization, a catalysed signal amplification system (catalyzed signal amplification system-Kit; Dako) was used according to the manufacturer's instructions. Antigen retrieval was performed for 6 min in preheated target retrieval solution (Dako), using a pressure cooker. Between incubations, specimens were washed two times in buffer (50 mM Tris-HCl, 300 mM NaCl, 0.1% Tween-20, pH 7.6). Diaminobenzidine was used as a chromogen for the peroxidase reaction. Control samples were from vehicle-pretreated donor mice. Signals were analysed with a Leica DMRB microscope, using differential interference contrast at a magnification of ×400. A total of 30 animals were included in this analysis (n = 5).

Quantitative reverse transcription polymerase chain reaction of hypoxia-inducible transcription factor target genes

To assess the expression of the HIF target genes mouse heme oxygenase-1 (mHO-I), erythropoietin receptor (EpoR), inducible nitric oxide synthase (iNOS), endothelial nitric oxide synthase (eNOS), Glut-1, MIF, TGFβ3 and VEGF in the donor aorta removed before the onset of ischaemia (n = 7 per group), expression profiles were conducted in triplets by applying the StepOne reverse transcription polymerase chain reaction System and the TaqMan Gene Expression Master Mix (Applied Biosystems, Forster City, CA, USA). Isolation of the RNA and the cDNA synthesis were performed according to standard protocols. Oligonucleotide sequences were generated with the free software Primer3 as previously described [9]. Primer sets and probes were commercially synthesized by Eurofins MWG GmbH (Ebersberg, Germany). To generate PCR standards, the respective PCR product was cloned into the cloning vector PCR 2.1® TOPO (Invitrogen, Germany) and sequenced. Standard curves with known concentrations of template copy numbers were used to determine the expression of the amplified target. The samples were normalized against the expression of the gene 18S rRNA, which is described as appropriate housekeeping gene for hypoxia experiments [10]. The results are given as relative expression units, which were calculated as a quotient of the respective gene expression relative to the housekeeping gene expression. A total of 42 animals were included in target genes analysis performed in triplicates (n = 7).

Abdominal aortic transplantation

The procedure was performed using a modified technique initially described by Koulack et al. [11]. In brief, the donor thoracic aorta was isolated, resected and stored in NaCl at 4°C (cold ischaemia). After laparotomy and bowel/tissue retraction, the recipient abdominal aorta was clamped and then transected with sharp microvascular scissors. The donor aorta of ∼5 mm was transplanted into the abdominal aorta by first a proximal end-to-end anastomosis. Then, the aortic graft was repositioned and the anastomosis continued with single interrupted sutures. Strain combinations BALB/cJ to BALB/cJ and C.B10-H2b/LilMcdJ to BALB/cJ were used for syngeneic and allogeneic transplantations, respectively [12].

Morphometric analysis of the aortic graft

Aortic grafts were removed under anaesthesia on Day 30 after transplantation, the timepoint when transplant arteriosclerosis has fully developed in MHC-class I-mismatched grafts [13]. Grafts were perfused with saline and were flash frozen in OCT medium (Tissue-Tek®, Sakura, Netherlands) in liquid nitrogen for morphometric analysis of 7-µm cryostat sections. A minimum of 10 transverse sections were analysed from each graft. Five sections from each graft were stained with Elastin/Van Gieson and haematoxylin/eosin and analysed by two independent examiners blinded to the experimental conditions at an original magnification of ×100 using a conventional light microscope. A digitized image of each section was captured and the areas within the lumen and the internal and external elastic lamina were circumscribed manually and measured as previously described [12, 13]. All image analyses were carried out on a colour display monitor using the cellSens Dimension® software (Olympus, Hamburg, Germany). A total of 28 recipient mice were included in this histomorphological analysis (n = 7).

Antibodies

The following antibodies and conjugates were used for this study: anti-CD4 and anti-CD8 were purchased from BD Bioscience (Heidelberg, Germany). Anti-macrophage F4/80 was obtained from AbD Serotec MorphoSys (Duesseldorf, Germany). The antibody to CD205 was grown from a hybridoma (HB-290/Clone NLCD145) obtained from American Type Culture Collection (Manassas, VA, USA) and purified from tissue culture supernatant for in vitro work. These antigens were detected by indirect immunofluorescence with the secondary antibody mouse anti-rat IgG-Cy3 (Dianova, Hamburg, Germany). Endothelial cells were detected with CD31-FITC (BD Biosciences, Heidelberg, Germany). The slides were covered with Vectashield Hard Set Mounting medium (Vector Laboratories, Burlingame, CA, USA) for 4’,6-Diamidin-2-phenylindol (DAPI) detection [12].

Immunofluorescence

Aortic grafts were removed and frozen as described in the morphometric analysis above [12]. Cryostat sections of 7 µm on gelatine-coated slides were air-dried overnight and then fixed in acetone for 10 min. Slides were then rehydrated and preincubated in staining buffer (0.1 M Tris, pH 8.0 and 0.1% Tween 20%), containing 5% heat-inactivated mouse serum (Invitrogen, San Diego, CA, USA) for 15 min. Afterwards, sections were incubated with purified primary antibodies in a humidified chamber at room temperature for 1 h. After three washes with staining buffer, detection of antigen was revealed with secondary antibody mouse anti-rat IgG-Cy3 (1 : 100). Unspecific Fc-receptors were blocked with staining buffer containing 5% heat-inactivated rat serum (Invitrogen) for 15 min followed by incubating the slides with the CD31-FITC for 1 h. After three final washes, the slides were mounted with Vectashield Hard Set Mounting medium and analysed by epifluorescence microscopy (Olympus, Germany). The quantification of the intragraft cellular infiltrate on Day 30 after transplantation was performed with computerized image analysis using the cellSens software. The positive stained area in relation to the total area of intima and media of each section was analysed using an original magnification of ×100 [12]. The 28 recipient animals from the histomorphological group were used for this analysis (n = 7).

Analysis of intragraft mRNA expression

Grafts were harvested at Day 14, which was previously established as the time point of highest cytokine expression [13], flushed with sterile saline and stored in RNAlater (Qiagen, Hilden, Germany). The expression profile for E-selectin, tumor necrosis factor-alpha (TNFα), interleukin 4 (IL-4), interleukin 6 (IL-6), interleukin 12 (IL-12), TGFβ, interferon gamma (IFNγ), intercellular adhesion molecule 1 (ICAM-1), platelet-derived growth factor-beta (PDGF-β), monocyte chemotactic protein 1 (MCP-1) and vascular cell adhesion molecule 1 (VCAM-1) was determined (see above). Oligonucleotide sequences were previously published [9]. For calculating the expression of E-selectin, we used the TaqMan® Gene Expression Assay (Life Technologies GmbH, Darmstadt, Germany). A total of 28 grafts were included in this analysis (n = 7).

Statistical analysis

Results are given as the mean per group ± standard deviation, which was derived from the mean per graft. Statistical analysis was done using a one-way ANOVA followed by a Bonferroni correction. Differences between groups are considered as significant when P < 0.05.

RESULTS

ICA stabilizes hypoxia-inducible transcription factor-α and induces hypoxia-inducible transcription factor target genes

In a first series of experiments, 40 mg/kg b.w. of ICA was administered i.p. to the respective animals, and we could demonstrate that this stabilized HIFα in vivo (Fig. 2). ICA was able to activate intracellular HIF in aortic grafts and to consecutively up-regulate different HIF target genes (Fig. 3), which is a prerequisite for testing the potential protective effect of ICA in an experimental mouse aortic allograft model. Four hours after donor treatment, aortic grafts were removed and subjected ex vivo to additional cold ischaemia periods of 0 h (immediate analysis), 2 h and 6 h (n = 5). Intracellular HIF expression was analysed using immunohistochemistry. There was no apparent difference in the distribution or intensity of HIF signalling (Fig. 2A–C) under these experimental conditions for up to 6 h of cold ischaemia time, a realistic timeframe for a clinical setting in human cardiac transplantation. Vehicle treatment did not lead to any relevant detectable HIF-α accumulation without (Fig. 2D) or with (Fig. 2E and F) additional cold ischaemia. Preliminary experiments showed that additional cold ischaemia did not result in significant HIF-dependent effects within 6 h [8]. Simulating a clinical set-up for thoracic organ transplantation, the following experiments were performed with a procurement timepoint of 6 h prior to the surgical procedure, followed by a short warm ischaemia time during the surgical procedure of an aortic allograft usually taking between 30 and 90 min. In order to evaluate the transcriptional response at the time of transplantation, 6 h after ICA treatment, mRNA levels of different HIF target genes were quantified in the aortic graft. After pretreatment with ICA, MIF, VEGF and TGFβ were strongly up-regulated, when compared with vehicle control-treated animals (Fig. 3), whereas the expression level iNOS, eNOS, EpoR and Glut-1 was not different after ICA treatment.

Figure 2:

HIF accumulation after ICA treatment in the aortic grafts of C.B10-H2^b/LilMcdJ mice before and after cold ischaemia. At 4 h after i.p. injection of ICA, HIF-1α was immunohistochemically detectable in endothelial cells of aortic grafts independent from time of additional cold ischaemia (B and C). Vehicle treatment with (E and F) or without (D) additional cold ischaemia did not lead to detectable HIF-1α accumulation. HIF-1α: hypoxia-inducible transcription factor-1α; ICA: 2-(1-chloro-4-hydroxyisoquinoline-3-carboxamido) acetate.

Figure 3:

Effect of ICA on the expression of HIF target genes. Potentially tissue-protective HIF target genes were quantified by real-time PCR 6 h after ICA treatment. The tissue-protective genes MIF, VEGF and TGFβ were significantly up-regulated after ICA treatment (P ≤ 0.05). HIF-1α: hypoxia-inducible transcription factor-1α; ICA: 2-(1-chloro-4-hydroxyisoquinoline-3-carboxamido) acetate.

ICA procurement results in significantly lower amounts of transplant arteriosclerosis

To detect differences in the extent of vascular lesions within aortic allografts of the respective experimental groups, C.B10-H2^b/LilMcdJ aortic allografts from BALB/c-recipients were analysed 30 days after transplantation. When donor mice were pretreated with ICA (Fig. 4D), transplant arteriosclerosis was significantly reduced in recipient mice when compared with vehicle control (Fig. 4C)-treated animals [intimal proliferation 31.3 ± 8% (ICA) vs 55.3 ± 20% (control) n = 7/P ≤ 0.01] (Fig. 4E). Syngeneic control grafts (BALB/c), which were studied to investigate the isolated role of ischaemia–reperfusion injury in the absence of a relevant allogeneic immune response, from recipients (BALB/c) treated with either ICA or vehicle control, did not show any relevant vascular lesions 30 days after transplantation, indicating that non-immunological mechanisms alone (cold ischaemia time up to 6 h) were not sufficient to initiate the development of transplant arteriosclerosis in this experimental model (Fig. 4A, B and E). None of the recipient mice showed any kind of thrombosis or major postoperative bleeding.

Figure 4:

Histopathological evaluation of the morphology of vehicle-treated partial allogeneic C.B10-H2^b/LilMcdJ aortic grafts implanted into Balb/C recipients showed elevated levels of intimal proliferation (C) which was significantly decreased after pretreatment with ICA (D) showed significantly decreased intimal proliferation 30 days after transplantation. Syngeneic controls did not show any vascular lesions (A and B). The tissue was snap-frozen and sections were stained with Miller's Elastin/van Gieson stain, original magnification, ×100. For the morphometric analysis (E) of the degree of intimal thickening areas within the lumen and the internal and external elastic lamina were circumscribed manually and measured. From these measurements, a quotient for the thickness of the intima (Q_int) was calculated. Q_int indicates the relative thickness (%) of the intima. Five measurements from different areas of each aortic allograft were obtained for this analysis (n = 7 animals per group/P-values as indicated in the diagram). ICA: 2-(1-chloro-4-hydroxyisoquinoline-3-carboxamido) acetate.

E-selectin and MCP-1 mRNA expression in aortic allografts was significantly down-regulated 14 days after ICA procurement

Transendothelial migration of leucocytes is considered to be an important component of chronic graft arteriosclerosis. Thus, we were next interested if early HIF stabilization induced by ICA pretreatment has an impact on mRNA expression in allografts of signature cytokines and adhesion molecules involved in this cellular transendothelial process. For this purpose, intragraft cytokine mRNA expression was measured 14 days after transplantation, a time point when changes in cytokine expression pattern are most prominent [13]. Aortic allografts from recipients pretreated with ICA exhibited significantly decreased intragraft mRNA expression for E-selectin (−57%), MCP-1 (−33%) and TGFβ (−36%) compared with vehicle-treated controls (Fig. 5). Intragraft mRNA expression of IFN-γ (−22%), IL-4 (-55%), TNF-α (−13%) and VCAM-1 (−16%) showed reduced levels after ICA procurement but the difference did not reach significance. Analysis of baseline cytokine and adhesion molecule gene expression within syngeneic controls was characterized by low levels of ICAM-1, VCAM-1, IFN-γ, IL-12 and TNF-α mRNA (data not shown).

Figure 5:

Quantitative reverse transcription polymerase chain reaction analysis of intragraft cellular adhesion molecule, growth factors and cytokine production. C.B10-H2^b/LilMcdJ aortic grafts implanted into Balb/C recipients were analysed 14 days after transplantation. Analysis revealed significant decrease of E-selectin, MCP-1 and TGFβ in ICA-pretreated animals. Data are shown as the mean of seven animals from each group (n = 7 animals per group/P-values as indicated in the diagram). ICA: PHD inhibitor 2-(1-chloro-4-hydroxyisoquinoline-3-carboxamido) acetate.

Cellular infiltration of the aortic grafts showed significantly decreased CD4+ T cells, dendritic cells (CD205+) and macrophages in ICA-pretreated animals

Next we were interested if observed differences in mRNA expression after ICA pretreatment were associated with differences in cellular migration and infiltration patterns. Aortic allografts recovered on Day 30 after transplantation displayed significantly reduced numbers of CD4+ T cells, dendritic cells (CD205) and macrophages (F4/80) within the intima and media after ICA procurement. (Fig. 6A—A–B,E–F,G–H). However, there was no marked difference in CD8+ T-cell infiltration. Quantification of the intragraft cellular infiltrate on Day 30 after transplantation was performed for CD4+ T cells [cell infiltration 8.4 ± 0.7% (control) vs 1.4 ± 0.8% (ICA) n = 7/P ≤ 0.01], CD8+ T cells [cell infiltration 10.7 ± 3.1% (control) vs 4.9 ± 3.5% (ICA) n = 7/P = 0.08], dendritic cells (CD205+) [cell infiltration 1.9 ± 2.0% (control) vs 0.6 ± 0.3% (ICA) n = 7/P ≤ 0.05] and macrophages (F4/80) [cell infiltration 10.9 ± 3.1% (control) vs 4.8 ± 2.0% (ICA) n = 7/P ≤ 0.01 (Fig. 6B)].

Figure 6:

Immunofluorescence (A) of frozen sections from aortic allografts 30 days after transplantation. A significant decrease in CD4+ T cell (A and B), dendritic cell (CD205) (E and F) and macrophage (F4/80) (G and H) infiltration was seen in aortic allografts after ICA pretreatment. (n = 7 animals per group). Quantification (B) of the intragraft cellular infiltrate on Day 30 after transplantation was performed with computerized image analysis using cellSens. The positive stained area in relation to the total area of intima and media of each section was analysed, using an original magnification of ×100. Quantification was performed for CD4+, CD8+, macrophages (F4/80) and dendritic cells (CD205+) (n = 7 animals per group/P-values as indicated in the diagram). ICA: 2-(1-chloro-4-hydroxyisoquinoline-3-carboxamido) acetate.

DISCUSSION

Previous studies about protective mechanisms of hypoxia-inducible gene expression in an animal model of allograft nephropathy [8] encouraged us to adapt this approach into a widely accepted animal model of transplant vasculopathy after cardiac transplantation. In the current study, we could show (i) that donor preconditioning using a single dose of ICA, a small molecule inhibitor of HIF PHDs resulted in HIF accumulation and subsequent induction of HIF target genes, such as VEGF and TGFβ, (ii) that vascular lesions were significantly reduced in preconditioned allografts when compared with vehicle controls, (iii) that this goes along with an intragraft down-regulation of E-selectin and MCP1 expression after ICA pretreatment and (iv) that it is accompanied with decreased T cell, dendritic cells and macrophage infiltration within the vascular grafts.

There is now accumulating evidence that short periods of ischaemia/reperfusion are able to protect against subsequent ischaemic injury [3, 14]. The heart was the first organ for which the principle of this so-called hypoxic preconditioning was described [15]. Hereby hypoxia prior to myocardial infarction reduced infarct area and improved left ventricular contractility [15]. As cardiac myocyte-specific knockout of HIF resulted in impaired vascularization and altered contractility of non-hypoxic hearts [16], it was concluded that functioning HIF is necessary for myocardial integrity. Myocardial infarction induces up-regulation of HIF, suggesting that the HIF system plays an important role in ischaemic myocardial damage and adaptive remodelling and that therefore preconditional HIF activation may protect the myocardium in ischaemic donor organs. The impact of myocardial protection by activating the HIF system has been described by several authors [17–19] but the impact of preconditioning donor organs for heart transplantation via induction of HIF and HIF target genes has been reported controversially in the literature [20, 21]. In addition, clinical application of ischaemic preconditioning in heart transplantation is obviously limited by difficulties to induce controlled ischaemia in donor organs in vivo. Therefore, the approach used in this study relies on mimicking the response to hypoxia by inhibiting prolyl-hydroxylases that mediate degradation of the transcription factor HIF. In our experiments, the 2-oxoglutarate analogue ICA as competitive inhibitor of HIF PHDs resulted in the accumulation of HIFα and induction of known HIF target genes in the aortic graft. In other context in rodents [20, 22] and in humans [23], 2-oxoglutarate analogues have previously been used.

There is a variety of HIF target genes which may play a role in the protection of vessels and neoangiogenesis. Our data show that genes involved in angiogenesis and regulation of the vascular tone, such as VEGF and TGFβ3, were up-regulated in ICA-pretreated grafts. Genes involved in cellular metabolism like Glut-1 were similar in pretreated and control groups, implying that these genes may not play a role in tissue protection of vascular grafts under the current conditions. Macrophage migration inhibitory factor, as an important parameter in vascular sclerosis, was significantly up-regulated and led to decreased macrophage infiltration within the vascular grafts. Interestingly, TGFβ expression was down-regulated after 14 days within the aortic wall, potentially as a consequence of the strong reduction of macrophage infiltration. EPO and heme oxygenase-1 (HO-1) demonstrated their potential to confer nephroprotection in a kidney allograft model [8], but, in our experimental model, expression levels were not significantly elevated.

Significantly reduced transplant arteriosclerosis of allografted vessels after ICA pretreatment correlated with an intragraft down-regulation of E-selectin, an important cell adhesion molecule expressed on endothelial cells and participating in recruiting leucocytes. In addition, MCP-1, was also down-regulated in preconditioned allografts, typically recruiting monocytes, T cells and dendritic cells to the sites of inflammation [24]. All these findings were supported by immunohistological analysis of infiltrating cells, revealing a reduced accumulation of the respective cell types within the pretreated grafts with significantly decreased infiltration of CD4+ T cells, dendritic cells (CD205+) and macrophages (F4/80). The lower CD4+ T cell infiltration may also be the reason for down-regulation of TGFβ1 expression potentially via IL-17 and thereby mediating the alloimmune response [25].

In contrast to our results, a recent study by Keränen et al. could not show a substantial benefit by pharmacological HIF-α preconditioning of cardiac allograft donors in a rat heart transplantation model [20]. The authors found that pharmacological HIF preconditioning of the recipient for 4 h reduced cardiac allograft inflammation, and mildly improved the long-term allograft survival, but surprisingly preconditioning of the donor even enhanced acute myocardial injury in their rat cardiac allografts 6 h after reperfusion [20]. Several aspects may be responsible for the differences between our results compared with those of the study by Keränen et al. [20]: Firstly, vascular lesions develop in mouse abdominal aortic transplants over the entire experimental timeframe and no concomitant application of immunosuppression is required that could potentially interact with HIF activation in the allogeneic setting. Both the current study and a previous study in an experimental kidney transplant model [8] were performed without the need of immunosuppression. In contrast, abdominal cardiac allografts as used in the study by Keränen et al. [20] require the application of cyclosporine to prevent acute rejection of the cardiac allograft which might interfere with HIF gene up-regulation. This assumption is also supported by the finding of Keränen et al. that NF-κB is down-regulated in their animal model [20], whereas we could not find an effect on NF-kB in our previous study [8]. Further differences include the substances that were used to inhibit HIF degradation and their mode of application. In addition, the effect of passenger leucocytes is different in between the experimental models and more substantial when cardiac allografts are performed as they harbour many more passenger leucocytes compared with aortic transplants. However, the specific interaction of HIF activation and passenger leucocytes is currently unclear and we would speculate at this time derived from our results that grafts characterized by only few passenger leucocytes such as the aortic allograft may be more suitable for organ preconditioning by HIF activation including the up-regulation of tissue-protective genes. Finally, the mouse abdominal aortic model has been shown to represent vascular lesions similar to those observed in human coronary arteries that are affected by transplant arteriosclerosis and allows a precise analysis of the composition of the vascular lesions [13].

In their recent study, Hegedus et al. described the concept of pharmaceutical donor preconditioning with a PHD inhibitor by stabilizing HIFs in a rodent model of brain death-associated donor heart dysfunction. They demonstrated the beneficial effects of pretreatment with the HIF stabilizing drug DMOG against I/R injury, another PHD inhibiting substance leading to HIF accumulation. This PHD inhibitor could ameliorate the brain-dead-associated mechanisms and hereby clearly improve graft function after heart transplantation [22]. In contrast to the findings by Keränen et al. [20], these results of cardioprotective HIF preconditioning in terms of improved graft function after rodent heart transplantation are in accordance with our experimental results of significantly decreased development of vasculopathy after heart transplantation following pharmaceutical HIF stabilization.

In summary, our study confirms the marked tissue-protective effect of oxoglutarate analogues in a murine experimental aortic allograft model of ischaemia–reperfusion injury. Our findings support the concept that this protective effect is mediated through concerted induction of HIF target genes and also indicate that HIF stabilization seems to be an attractive strategy for clinical organ preservation that deserves further investigation.

LIMITATIONS OF THIS STUDY

Certain limitations of this study are necessary to address. We used the murine aortic transplantation model to evaluate the effect of pretransplant HIF-1 stabilization on the development of transplant arteriosclerosis. All basic and preclinical science experiments evaluate specific points of interest by standardizing all variables. These clinical variables are to be taken into account when results of animal models are translated to human research. The murine aortic transplantation model was developed because the intima proliferation developing in an allograft situation is much likely to the development of neointimal formation in vessels of transplanted human hearts [13]. While this vascular transplantation model has several advantages (clear readout, no immunosuppression needed etc.), the effect of passenger leucocytes cannot be investigated in this set-up and needs further evaluation in other animal models to find out the specific interaction of HIF activation and passenger leucocytes.

Funding

Research support was obtained from the ELAN and IZKF-trust of the University of Erlangen-Nuernberg and the ADUMED-foundation.

Conflict of interest: none declared.