Effect of polyethylene glycol-based preservation solutions on graft injury in experimental kidney transplantation

Abstract

Background

New preservation solutions are emerging, of various ionic compositions and with hydroxyethyl starch replaced by polymers such as polyethylene glycols (PEGs), offering the potential for ‘immunocamouflage’. This experimental study investigated which of three clinically available preservation protocols offered the best graft protection, based on epithelial-to-mesenchymal transition (EMT) and fibrosis.

Methods

Kidneys were preserved for 24 h at 4 °C with University of Wisconsin solution (UW) as standard, compared with solutions containing either 1 g/l PEG 35 kDa (Institute Georges Lopez solution, IGL) or 30g/l PEG 20 kDa (solution de conservation des organes et des tissus, SCOT). Animals were followed for up to 3 months and development of EMT, tubular atrophy and fibrosis was evaluated in comparison with sham-operated animals.

Results

Functional recovery was better in the SCOT group compared with the other groups. Chronic fibrosis, EMT and inflammation were observed in the UW and IGL groups, but limited in the SCOT group. Levels of profibrosis markers such as transforming growth factor β1, plasminogen activator inhibitor 1 and connective tissue growth factor were increased in IGL and UW groups compared with the SCOT group. Hypoxia-inducible factor (HIF) 1α and 2α expression was increased at 3 months in grafts preserved in UW and IGL, but detected transiently on day 14 when SCOT was used. Expression of HIF-regulated genes vascular endothelial growth factor and erythropoietin was increased in UW and IGL groups.

Conclusion

The choice of colloid and ionic content is paramount in providing long-term protection against chronic graft injury after renal transplantation. Preservation solutions based on PEGs may optimize graft quality.

Introduction

Kidney transplantation is the preferred therapy for end-stage renal disease. Despite dramatic improvements in immunosuppression, the rate of chronic allograft injury remains high. The observed lesions of chronic allograft nephropathy, namely interstitial fibrosis (IF) and tubular atrophy (TA), are the end result of both immune and non-immune injury to the graft. Ischaemia–reperfusion injury (IRI) is inevitable in transplantation and is strongly correlated with early non-function or delayed graft function^1,2, as well as chronic graft failure and late graft loss.

Hypothermia per se causes damage, and is associated with energy depletion, acidosis, cell swelling and production of reactive oxygen species at the time of reperfusion. IRI also induces an innate immune response independently of allogenicity³, activating the endothelium and resulting in intense immune cell invasion to the kidney^4,5.

Increasing awareness that IRI partly determines outcome has stimulated research in the field of preservation and the development of new preservation solutions⁶. Although University of Wisconsin solution (UW) for static cold storage is still the standard preservation solution for kidney, liver, pancreas and small bowel^6,7, some studies have suggested that certain additives in UW may do more harm than good^8,9. UW is a high potassium ion (K⁺)/low sodium ion (Na⁺) solution¹⁰, originally thought to maintain intracellular ionic balance. However, high K⁺ induces cellular depolarization, decreases cellular adenosine 5′-triphosphate content and activates voltage-dependent channels, such as calcium channels^1,11. Recently low K⁺/high Na⁺ solutions have shown equal or improved results⁶.

Oedema represents one of the pivotal factors in organ preservation, characterized by mitochondrial and cell swelling. To mitigate this phenomenon, preservation solutions are supplemented with colloids. Hydroxyethyl starch (HES) is used in UW, but has been found to increase tubular damage¹² and red blood cell aggregation¹. Other colloids are emerging, such as polyethylene glycol (PEG), which have interesting properties in the context of organ preservation^13,14. PEG is a neutral, water-soluble, non-toxic polymer; in addition to its role as a colloid, adsorption of the PEG molecule to the cell membrane creates layers of ‘structured water’, offering an ‘exclusion volume’ and preventing the approach of other cells and protein fixation, as in the case of the immunological synapse^15–17. By interfering with identification of the graft as foreign by cells of the immune system, PEG prevents the activation of IRI-induced inflammation, which can have a long-term effect on graft outcome, particularly through interference with danger signal production¹⁸.

The aim of the present study was to compare the performance of two recently developed PEG-based solutions with the standard UW in an experimental kidney allotransplant pig model, which allowed measurement of functional recovery of the graft.

Methods

Large White male pigs (Institut National de la Recherche Agronomique, Génétique et Expérimentation en Productions Animales, Surgères, France) weighing 30–35 kg were prepared as described previously^19–21 in accordance with the French guidelines of the Ethical Committee for Human and Animal Studies. Briefly, the kidney was harvested, cold flushed with approximately 300 ml solution, and preserved for 24 h at 4 °C immersed in the solution. Heparin (5000 units) was injected 10 min before constructing the anastomosis. Recipients' right kidneys were removed and the allograft was implanted by end-to-side anastomosis on the aorta and inferior vena cava. The left kidney was removed to mimic the nephron mass in the transplanted situation. Surgical teams were blinded to the preservation protocols. Reperfusion was started at time 0. For allograft matching, analysis of preoperative blood samples ensured compatibility for swine lymphocyte alloantigen (SLA) class I by microlymphotoxicity; the microsatellite technique was used to ensure less than 10 per cent recombination for SLA class II, permitting graft survival with low-grade acute rejection without immunosuppression²².

Preservation solutions

Three groups of 18 animals each were used to study the effect of three preservation solution protocols currently in clinical use. UW (ViaSpan®; Bristol-Myers Squibb, Garden City, New York, USA), containing high K⁺/low Na⁺, was used as the standard. This was compared with Institute Georges Lopez solution (IGL-1®; IGL Group, Saint-Didier au Mont d'Or, France), which contains low K⁺/high Na⁺ and has HES replaced with 1 g/l PEG 35 kDa, and solution de conservation des organes et des tissus (SCOT®; MacoPharma, Tourcoing, France), which contains low K⁺/high Na⁺ and 30 g/l PEG 20 kDa. The composition of each solution is shown in Table 1. Experimental groups were compared with a group of control, sham-operated, age- and weight-matched animals. Six animals from each group were killed on day 14, and at 1 and 3 months after surgery. Samples were collected and either frozen in liquid nitrogen or fixed in paraformaldehyde.

Table 1

Composition of the preservation solutions

	Blood	UW	IGL-1®	SCOT®
Sodium ion (mmol/l)	140	30	125	118
Potassium ion (mmol/l)	5	125	30	5
Magnesium ion (mmol/l)	0·8	5	5	1·20
Calcium ion (mmol/l)	2·5			1·75
Chloride ion (mmol/l)	104
Sulphate (mmol/l)	1·4	5	5
Phosphate (mmol/l)	3·2	25	25
Bicarbonate (mmol/l)	25			25
Glucose (mmol/l)	7			11
Raffinose (mmol/l)		30	30
Lactobionate (mmol/l)		100	100
Adenosine (mmol/l)		5	5
Glutathione (mmol/l)		4	3
Allopurinol (mmol/l)		1	1
HES (g/l)		50
PEG 20 kDa (g/l)				30
PEG 35 kDa (g/l)			1
pH	7·4	7·3	7·3	7·3
Viscosity (centistokes)	1·6	2·4	0·7	1·6
Osmolarity (mOsm)	308	320	320	320

UW, University of Wisconsin solution. IGL, Institute Georges Lopez solution; SCOT, solution de conservation des organes et des tissus; HES, hydroxyethyl starch; PEG, polyethylene glycol.

Functional parameters

Pigs were placed in a metabolic cage for measurement of creatinaemia and proteinuria as described previously^19–21.

Evaluation of tubular injury and fibrosis

Conventional stains were applied (haematoxylin and eosin, periodic acid–Schiff) to formalin-fixed sections (3 µm) of kidney. The magnitude of tubular atrophy was scored as described previously²⁰: 0, no abnormality; 1, less than 10 per cent of area affected; 2, 10–25 per cent; 3, 26–50 per cent; 4, 51–75 per cent; and 5, more than 75 per cent. All sections were examined under blinded conditions by a pathologist and a nephrologist. Tubulointerstitial fibrosis was determined using picrosirius red staining²³. Renal tissue was also used for a hydroxyproline assay^24,25. Immunostaining was carried out using the following primary antibodies: collagen I (1:100 dilution; Cosmo Bio, Tokyo, Japan), vimentin (1:100 dilution; Dako, Stockholm, Sweden), α-smooth muscle actin (1:100 dilution; Sigma, St Louis, Missouri, USA) and hypoxia-inducible factor (HIF) 1α (1:200 dilution; Novus Biologicals, Littleton, Colorado, USA). The percentage staining was determined by computerized image analysis in ten randomly selected fields of each slide (at × 200 magnification).

Determination of inflammatory cells

CD3-positive cells, macrophages and monocyte invasion were measured on frozen sections from the graft at day 14, 1 and 3 months, with antibodies against CD3 (1:100 dilution; SouthernBiotech, Birmingham, Alabama, USA) and ED1 (1:100 dilution; AbD Serotec, Oxford, UK). Ten high-power fields (at × 400 magnification) were selected randomly and the number of positive cells determined in a blinded fashion. Quantitative determination of tubulitis was adapted from the Banff classification²⁶: 0, no mononuclear inflammatory cell in tubules; 1, foci with one to four mononuclear cells per tubular cross-section or ten tubular cells; 2, foci with five to ten mononuclear cells per tubular cross-section; and 3, foci with more than ten mononuclear cells per tubular cross-section.

Western blotting

A standard western blotting protocol was used, as described previously^21,27. Antibodies used were against transforming growth factor (TGF) β, HIF-1α and vimentin (1:600, 1:200 and 1:200 dilution respectively; Santa Cruz Biotechnology, Santa Cruz, California, USA), E-cadherin and zonula occludens (ZO) 1 (both 1:200 dilution; Invitrogen, Carlsbad, California, USA), connective tissue growth factor (CTGF) (1:500 dilution; BioVision, San Francisco, California, USA), plasminogen activator inhibitor (PAI) 1 (1:200 dilution; BD Biosciences, San Diego, California, USA), matrix metalloproteinase (MMP) 2 (1:10 dilution; Chemicon International, Temecula, California, USA), bone morphogenetic protein (BMP) 7 (1:5000 dilution; AbD Serotec, Raleigh, North Carolina, USA), HIF-2α (1:500 dilution; Abcam, Cambridge, Massachusetts, USA) and loading control β-actin (1:500 dilution; Sigma). Protein band intensities were quantified using ImageJ (National Institutes of Health, Bethesda, Maryland, USA).

Determination of mRNA expression by quantitative real-time polymerase chain reaction

Tissues snap-frozen in liquid nitrogen were processed for RNA extraction using TRIzol® (Invitrogen; supplied by Fisher Scientific, Illkirch, France) according to the manufacturer's recommendations. Genomic DNA was removed using DNA-free kit (Applied Biosystems, Foster City, California, USA). Each cDNA template for reverse transcriptase–polymerase chain reaction (PCR) was prepared by first-strand reverse transcription (Applied Biosystems). Real-time PCR assays were performed with 10 ng total cDNA templates on an ABI Prism 7300 Sequence Detection System (Applied Biosystems) following the manufacturer's recommendations. Porcine primers were designed using OligoPerfect™ (Invitrogen), with sequences as detailed in Table 2. Finally, the mRNA expression level was obtained by the 2^(−ΔΔCt) method²⁸ using 18S as internal control.

Table 2

Primer sequences for real-time quantitative polymerase chain reaction analysis in pig kidneys

Gene	Forward	Reverse
18S	AGCCTGCGGCTTAATT	AACCAGACAAATCGCT
	TGAC	CCAC
HIF-1α	TGGCAGCAATGACACAG	GAGGCAGGCAATGGAG
	AAAC	ACAT
EPO	TGGGCTTGCCGAAA	TCCATGGCCTGCTGCT
VEGF	GCCCACTGAGGAGTTCAA	GGCCTTGGTGAGGTTT
	CATC	GATC
BMP-7	TCGGACCTGTTCCTGCTC	TGCTGGTGGCCGTGA

HIF, hypoxia-inducible factor; EPO, erythropoietin; VEGF, vascular endothelial growth factor; BMP, bone morphogenetic protein.

Statistical analysis

Results are presented as mean(s.e.m.). Within-group comparisons were performed by use of paired Student's t test, and comparisons among groups by two-way ANOVA with the Bonferroni correction for multiple group comparisons followed by Student–Newman–Keuls test²⁹. For histological data, the Kruskal–Wallis test for multiple comparisons was used followed by the Connover test²⁹. P < 0·050 was considered to be significant.

Results

The time taken to complete the anastomosis ranged from 25 to 35 min. Haemodynamics were comparable between groups and there were no postoperative complications.

Grafts preserved with SCOT showed the best renal recovery by day 14, and at 1 and 3 months' follow-up (Fig. 1a). Proteinuria remained low in the SCOT group (0·30(0·10) g per 24 h), whereas it was critically raised at 3 months in the UW (3·40(0·25) g per 24 h; P = 0·046) and IGL (3·20(0·29) g per 24 h; P = 0·037) groups (Fig. 1b)

Fig. 1

Kidney function. a Serum creatinine and b proteinuria at 14 days, 1 month and 3 months after transplantation in control, University of Wisconsin solution (UW), Institute Georges Lopez solution (IGL) and solution de conservation des organes et des tissus (SCOT) groups. Values are mean(s.e.m.). *P < 0·050 versus control and SCOT groups at each time point (n = 6) (two-way ANOVA with Bonferroni correction for multiple group comparisons followed by Student–Newman–Keuls test)

Chronic fibrosis

The development of fibrosis in kidney grafts was evaluated by picrosirius red staining (Fig. 2a–d). Quantification showed that grafts preserved with SCOT demonstrated little fibrosis, comparable to controls. However, grafts preserved with UW and IGL had a significantly higher level of fibrosis (Fig. 2b). The hydroxyproline level and collagen I staining were reduced in the SCOT group (Fig. 2c,d).

Fig. 2

Fibrosis development. a Representative images (original magnification × 200) of typical picrosirius red staining of kidneys from control, University of Wisconsin solution (UW), Institute Georges Lopez solution (IGL) and solution de conservation des organes et des tissus (SCOT) groups at 3 months after surgery. b–d Graphical representation of b amount of fibrosis, c collagen I score and d hydroxyproline level in each group at 14 days, 1 month and 3 months. Values are mean(s.e.m.). *P < 0·050 versus control and SCOT groups at each time point (n = 6) (Kruskal–Wallis test for multiple comparisons followed by Connover test)

TGF-β is a key element in chronic kidney fibrosis. Western blot analysis of TGF-β at day 14, 1 and 3 months revealed increased expression in grafts preserved in UW and IGL compared with control, whereas expression remained low when SCOT was used (Fig. 3a). The MMP-2 level was increased in UW and IGL-1 groups, and unchanged in the SCOT group (Fig. 3b). PAI-1 (Fig. 3c) and CTGF (Fig. 3d) expression was increased in UW and IGL groups, and stable in the SCOT group.

Fig. 3

Epithelial-to-mesenchymal transition development at 14 days, 1 month and 3 months after transplantation in control, University of Wisconsin solution (UW), Institute Georges Lopez solution (IGL) and solution de conservation des organes et des tissus (SCOT) groups: protein levels of a transforming growth factor (TGF) β, b matrix metalloproteinase (MMP) 2, c plasminogen activator inhibitor (PAI) 1, d connective tissue growth factor (CTGF), e vimentin, f E-cadherin, g zonula occludens (ZO) 1 and h bone morphogenetic protein (BMP) 7. Values are mean(s.e.m.). *P < 0·050 versus control and SCOT groups at each time point (n = 6) (two-way ANOVA with Bonferroni correction for multiple group comparisons followed by Student–Newman–Keuls test)

Epithelial-to-mesenchymal transition development

To refine the analysis of epithelial-to-mesenchymal transition (EMT) development within the grafts, several markers were measured at day 14, and 1 and 3 months. Vimentin was detected on day 14 and expressed at control level 1 and 3 months after transplantation in the SCOT group, whereas its expression was increased in the UW and IGL groups at 1 and 3 months (Fig. 3e). These findings were confirmed by immunohistochemistry for vimentin and α-smooth muscle actin (data not shown). Protein levels of E-cadherin (Fig. 3f), ZO-1 (Fig. 3g) and BMP-7 (Fig. 3h) were attenuated in IGL and particularly UW groups.

Inflammatory response

To evaluate the graft inflammatory response, immuno-staining for monocytes and macrophages (Fig. 4a) and CD3, a lymphocyte marker, was carried out. Quantification of these cells revealed reduced numbers in kidneys preserved in SCOT (Fig. 4c,e). This was further shown by measurement of tubulitis (Fig. 4b,d).

Fig. 4

Inflammation and hypoxia. a,b Representative images (original magnification × 200) of typical staining for a monocytes and macrophages (ED1 immunostain) and b tubulitis (haematoxylin and eosin stain) in kidneys from control, University of Wisconsin solution (UW), Institute Georges Lopez solution (IGL) and solution de conservation des organes et des tissus (SCOT) groups at 3 months after surgery. Arrows represent invading cells. c–h Quantification of c monocytes and macrophages (ED1 expression), d tubulitis, e CD3-positive cells, and mRNA expression of f hypoxia-inducible factor (HIF) 1α, g vascular endothelial growth factor (VEGF) and h erythropoietin in each group at 14 days, 1 month and 3 months. Values are mean(s.e.m.). *P < 0·050 versus control and SCOT groups at each time point (n = 6) (Kruskal–Wallis test for multiple comparisons followed by Connover test for histological data; two-way ANOVA with Bonferroni correction for multiple group comparisons followed by Student–Newman–Keuls test for other data)

Long-term hypoxic injury

Chronic fibrosis reduces the blood supply to the tubules, inducing the activation of oxygen-sensitive pathways. To evaluate hypoxia in the kidney, HIF-1α and HIF-2α expression was determined. Levels of these markers were increased at 1 and particularly 3 months after transplantation in UW and IGL groups (Fig. 4f). Expression was increased at day 14 and then declined by 1 and 3 months after transplantation in the SCOT group, suggesting that hypoxia was reduced in this group. Expression of HIF-activated genes encoding vascular endothelial growth factor (VEGF) and erythropoietin was also increased in UW and IGL groups at 1 and 3 months after transplantation compared with the SCOT group, in which expression of these genes was increased transiently at day 14 (Fig. 4g,h).

Discussion

This experimental study has demonstrated the importance of ionic composition and choice of colloid in allograft preservation. The plasma-like content associated with PEG 20 kDa at 30 g/l in SCOT provided a significantly better outcome.

The mechanisms of chronic allograft failure reflect a complex interplay between immune and non-immune mechanisms³⁰. Acute kidney injury immediately after transplantation not only impairs early graft function but also is associated with reduced long-term graft survival^31,32. EMT, a process by which differentiated epithelial cells undergo a phenotypic conversion that gives rise to matrix-producing fibroblasts and myofibroblasts, is increasingly recognized as an integral part of tissue fibrogenesis after injury^33,34. The role of PEG during conservation and the importance of polymer size for immunocamouflage have been discussed in recent reports^16,35,36. The present study brings new insights into the role of PEG and ionic composition of the preservation solution using a kidney transplantation model.

EMT is thought to be a repair mechanism that can be deregulated during injury and promote IF in four key steps: loss of epithelial proteins such as E-cadherin and ZO-1; de novo expression of vimentin and α-smooth muscle actin; disruption of the tubular basement membrane and degradation of internal architecture; and, finally, enhanced cell migration and invasion with production of profibrotic molecules such as type I collagen^37,38. The role of preservation conditions and PEG in the development of IF/TA has been demonstrated previously^{19,20,39–41}. In addition, early phenotypic changes indicative of EMT were shown to be predictive of the progression toward IF in human renal allografts⁴².

In the present study, kidney preservation with SCOT was associated with an improvement in renal functional recovery and a reduction in urinary protein output. Proteinuria is one of several mechanisms by which the disease process is transmitted to the interstitial space⁴³, as a good correlation between proteinuria and the occurrence of EMT has been shown in rats⁴⁴. Proteins per se increase secretion of tubular cytokines such as TGF-β, the most potent inducer of EMT⁴⁵.

In the present study relative increases in TGF-β expression in the kidney were accompanied by increased IF/TA. TGF-β signalling is involved in cell proliferation, tissue growth and matrix remodelling by enhanced collagen synthesis⁴⁶. Proteases are factors in the peritubular microenvironment that induce or promote EMT^{33,34,37,38,45}. It has been reported that MMP-2 is sufficient to induce tubular EMT in vitro, and overexpression of MMP-2 in transgenic mice promotes renal fibrosis⁴⁷. PAI-1 and CTGF expression was increased in IGL and UW groups. PAI-1 controls plasmin formation, acts as an inhibitor of fibrinolysis and matrix degradation, and is involved in the development of tissue fibrosis and IF/TA⁴⁸. PAI-1 expression is increased among the EMT proteome⁴⁹. CTGF, a downstream effector of TGF-β, is involved in induction of cell proliferation, collagen synthesis and chemotaxis in a variety of cells⁵⁰, and has a critical role in graft fibrogenesis⁵¹. The present study also showed modulation of loss of epithelial proteins such as E-cadherin and ZO-1. Loss of tight and adherent junctions is followed by reorganization of actin stress fibres and de novo expression of vimentin, a mesenchymal marker^33,38. The present in vivo model, which was free from any effects due to immunosuppression⁵², allowed investigation of the role of ionic composition and choice of PEG size and concentration in the preservation solution on graft outcome and different proteins of the EMT proteome.

The expression of factors that negatively regulate the EMT was also evaluated. In the SCOT group, BMP-7 protein and mRNA was upregulated during follow-up. BMP-7 signals through the BMP receptors to phosphorylate different Smad proteins (1, 5 and 8), which are in balance with Smad 2 and 3 regulated by TGF-β^33,34,37,38. Consequently, BMP-7 can counterbalance the profibrotic effects of TGF-β^53,54. These results suggest that is possible to restore homeostatic balance after cold ischaemia and reperfusion.

Preservation conditions can also influence the peritubular environment during cold ischaemia and reperfusion. The role of PEG 20 kDa in the inflammatory response and its influence on the development of IF has been reported previously^19,39. Direct contact of tubular epithelial cells with T cells may induce EMT^33,34,38. Interestingly, in the present study modulation of the inflammatory response in SCOT was associated with modulation of the endo- genous TGF-β/BMP-7 balance, fibrosis development and preservation of the epithelial protein junction.

HIF-1α has been described as a powerful cue for inducing tubular EMT in vivo and in vitro in several studies^55,56. Stable expression of HIF-1α consistently promotes IF^57–63. However, the HIF pathway also induces protective responses against ischaemia, such as the VEGF pathway^59–61. In the present study, preservation with SCOT was associated with early expression of HIF and its related genes, which then declined to control level by 1 and 3 months after transplantation. Expression of HIF was delayed in UW and IGL groups. Thus, consistent and delayed HIF expression in grafts preserved in UW and IGL is probably a sign of chronic hypoxia and development of fibrosis, whereas early and regulated expression in kidneys preserved in SCOT may have favoured repair processes associated with improved outcome. This is in accordance with the recently developed notion that HIF-mediated cellular responses differ under conditions of acute or chronic oxygen deprivation^61,64.

The superiority of IGL and PEG 35 kDa over UW has been reported previously^65,66, in contrast to the present findings. The choice of model is the most likely explanation for this discrepancy, as moderate allogenicity was introduced into the model described here. IGL and SCOT have a similar ionic composition, suggesting that the superiority of SCOT was due to better immunocamouflage provided by PEG 20 kDa at 30 g/l, whereas PEG 35 kDa in IGL at low concentration (1 g/l) appears incapable of protecting against the immune response. The protection afforded by PEG appears highly dependent on the polymer size and dose, warranting further studies to determine optimal parameters for the use of PEG in organ preservation.

Acknowledgements

The authors are grateful to Professor G. Mauco for his guidance and support during this study.

This study was funded by Conseil Général de la Vienne, Région Poitou Charentes, the Banque Tarneaud, Poitiers, Centre Hospitalier Universitaire de Poitiers and Institut National de la Santé et de la Recherche Médicale, the Société Francophone de Transplantation and the French Foundation of Transplantation. The authors declare no conflict of interest.