- Email: [email protected]
A pig allograft model of antibody-mediated rejection
Transplant Immunology 19 (2008) 167–172
Contents lists available at ScienceDirect
Transplant Immunology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t r i m
Brief communication
A pig allograft model of antibody-mediated rejection Paul E. Herbert a,c, M. Finn Morgan a,c, Ivan Berton a, Robert I. Lechler a, Anthony Dorling a, Gordon Williams c, Anthony N. Warrens a,b,⁎ a b c
Department of Immunology, Imperial College London, Hammersmith Campus, Ducane Road, London W12 0HS, UK Department of Renal Medicine, Imperial College London, Hammersmith Campus, Ducane Road, London W12 0HS, UK Department of Surgery, Hammersmith Hospital, Ducane Road, London W12 0NN, UK
A R T I C L E
I N F O
Article history: Received 24 March 2008 Accepted 7 May 2008 Keywords: Antibody-mediated rejection Anti-HLA antibodies Sensitisation Transplantation
A B S T R A C T Allograft rejection caused by antibodies in sensitised individuals remains a real problem in human allotransplantation. There would be value in a simple model of this process to evaluate the mechanisms involved in antibody-mediated damage and the development of accommodation, as well as the impact of potential interventions. We have thus developed a novel large animal model of this process using an allosensitisation system. Two inbred lines of miniature pigs that carry different major histocompatibility antigen haplotypes were used. Pigs of one line were sensitised by the sequential subcutaneous injection of major histocompatibility antigen-mismatched allogeneic peripheral blood mononuclear cells derived from the other inbred line. We demonstrated that this generated high titres of allospecific antibodies. We then transplanted carotid arteries from donors syngeneic to the priming cells into the sensitised animals. After 48 h these vessels showed a profound mononuclear cell inflammatory infiltrate in both intima and media, fibrin deposition, and luminal compromise with thrombus and antibody deposition. The mean endothelial surface affected by this process was 59.2%. No such pathology was seen in any of the controls. This model is technically simple to perform with few post-operative complications. It provides proof-of-principle of a model of antibody-mediated rejection which will be of potential value in elucidating the mechanisms underlying the process and the efficacy of interventions to prevent or treat it. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Pre-formed antibodies are a significant issue in allotransplantation, when the transplantation of sensitised recipients with pre-formed HLAspecific antibodies is being considered. In such situations, attempts to remove antibodies in advance of and following transplantation, combined with strategies to suppress further production of anti-HLA antibodies have met with some success [1–3]. However, the outcome in these situations is still inferior to transplantation in the non-sensitised recipient and the parameters that determine success have not been well defined. The situation is further complicated by the observation that, in some situations, the return of antibodies following transplantation does not cause the expected damage, because the graft has in some way adapted itself to be less susceptible to antibody-mediated damage, a process called “accommodation” [4,5]. In addition, new, highly sensitive techniques are now available that can identify the presence of low, previously undetected levels of antibodies [6,7]. Hence, there are several
variables that need to be teased out to understand the mechanisms underlying these phenomena. It is likely that they will only be elucidated using an appropriate animal model. Rodent models of vascular transplantation have been described, sometimes demonstrating the role of antibodies in transplant vasculopathy, but are technically difficult [8–10]. That is why we have developed a large animal model. 2. Objective In this study, we sought to prove the principle that it would be possible to study antibody-mediated allograft damage using a technically easy, large artery model in the pig. To do this we have studied a system with high levels of pre-formed allospecific antibodies. Having proved this principle, we believe this model will be useful to study the importance of different levels of antibodies, different timing of induction (before and after transplantation) and the mechanisms underlying different outcomes (rejection and accommodation). 3. Materials and methods
⁎ Corresponding author. Department of Immunology, Imperial College London, Hammersmith Campus, Du Cane Road, London, W12 0HS, UK. Tel.: +44 20 8383 2307; fax: +44 20 8383 2788. E-mail address: [email protected] (A.N. Warrens). 0966-3274/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.trim.2008.05.003
3.1. Animals Inbred lines of minipigs (c/c and d/d SLA (swine leucocyte antigen) phenotypes) [11] (kind gift of Dr David Sachs) were used. c/c
168
P.E. Herbert et al. / Transplant Immunology 19 (2008) 167–172
Fig. 1. Porcine carotid artery transplantation. The carotid artery was exposed (A) and a section excised (B). Another section of artery was transplanted with two end-to-end anastomoses (C) and the animal was allowed to wake up. Forty-eight hours later the vessel was exposed (D) and excised. Further details of the surgical methods involved are provided in Materials and methods.
animals were used as recipients of vessels (or untransplanted controls) and both c/c and d/d animals as vessel donors. All were greater than 20 kg at use. All procedures were performed in
accordance with the Animals (Scientific Procedures) Act 1986, after satisfying the internal ethical review process of Imperial College London.
Fig. 2. Generation of SLAd-allospecific antibodies in SLAc pigs. SLAc pigs were sensitised with weekly injections of SLAd cells (days 0, 7, 14 and 21) or not (controls). On regular occasions, blood was drawn and neat serum mixed with either SLAd (dd) (solid line) or control SLAc (cc) (broken line) cells and FITC-conjugated goat anti-pig IgG added. (A) Allospecific antibody is detectable by day 16. (B) High titres are observed after four weeks. In these flow cytograms, signal from SLAd cells are shown in outline and from control SLAc cells in solid graphs. This represents a representative example of four pigs.
P.E. Herbert et al. / Transplant Immunology 19 (2008) 167–172
169
Fig. 3. Generation of antibody-mediated damage following allosensitisation. Arteries were excised and (in panels B–D and F–G) transplanted into recipients as indicated. They were then removed 48 h later. Tissues were studied for gross histology using H&E staining (A–D) and for deposition of fibrin using Martius scarlet (E–H). The figure illustrates an untransplanted artery (A, E), a transplanted syngeneic artery (c/c into c/c) (B, F), an allogeneic artery transplanted (d/d into c/c) into an unsensitised animal (C, G) and an allogeneic artery transplanted (d/d into c/c) into an animal sensitised to d/d cells. In each panel, a low power view is represented in the upper quadrant. Within that, the area represented in high power is outlined by a back box and the high power view occupies the larger part of the panel.
170
P.E. Herbert et al. / Transplant Immunology 19 (2008) 167–172
FACScalibur (Becton Dickinson, San Jose, CA, USA) flow cytometer using CellQuest software (Becton Dickinson, San Jose, CA, USA).
Table 1 Extent of thrombosis in transplanted arteries Recipient
Artery donor
c/c c/c c/c c/c c/c c/c c/c c/c Untransplanted c/c
d/d d/d d/d d/d d/d d/d d/d c/c −
Sensitised + + + − − − − − −
No. of levels cut 10 10 5 5 5 5 5 5 5
% of intima covered by thrombus 39.1 +/− 12.6 52.5 +/− 27 86 +/− 12 0 0 0 0 0 0
Results of staining showing percentage of intima covered by thrombus 48 h after transplantation. Intimal thrombin deposition occurred only in the sensitised animals and in each animal a significant proportion of the vessel intima was in contact with this thrombus. There is a significant statistical difference between the sensitised animals and the controls (p b 0.05).
3.2. Sensitisation Whole blood from d/d animals was heparinised and diluted 1:1 in RPMI 1640 medium (GIBCO, Invitrogen Corporation, Renfrew, UK). Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll gradient centrifugation and then frozen at −80 °C in RPMI 1640 medium containing 10% foetal calf serum (Biowest, Oxford, UK) until ready for use. Each c/c pig to be sensitised received approximately 1×108 d/d PBMCs, suspended in 2 mL phosphate buffered saline (PBS) (Oxoid Ltd., Basingstoke, UK), subcutaneously into the scruff of the neck weekly for four weeks. Unsensitised controls received no injections. Serum was sampled at various time points to test for the development of SLAd-specific alloantibodies. 3.3. Flow cytometry 2 × 105 PBMCs were resuspended in 200 μL neat pig serum and left at 4 °C for 30 min. The cells were washed and FITC-conjugated goat anti-pig IgM or IgG (Bethyl Laboratories, Montgomery, Texas, USA) added. After 30 min at 4 °C, the cells were washed and analysed in a
3.4. Carotid transplantation Pigs were anaesthetised with halothane (1%) (Zeneca, Macclesfield, UK) in nitrous oxide and oxygen after pre-medication with 10 mg/kg ketamine hydrochloride intramuscularly (Pharmacia Upjohn/Pfizer, New York, USA). The right common carotid artery was exposed by blunt dissection (Fig. 1A). A section of native artery was removed and replaced with a 3 cm section of transplanted artery. The donor vessel was placed by end-to-end anastomosis using 6/0 prolene (Ethicon Ltd., Edinburgh, U.K.) (Fig. 1B–C). After haemostasis had been achieved, 2 mL of papaverine was gently squirted by syringe over the vessel through a 23 G needle (Tyco, Gosport, UK) and patency confirmed both by digital palpation and by a hand-held Doppler ultrasound probe. The wound was then closed with 4-0 vicryl (Ethicon Ltd., Edinburgh, UK), by re-opposing the previously cut fibro-muscular layers and finally the skin, and the animal allowed to wake. Forty-eight hours after the operation, each animal was anaesthetised again and the transplanted vessel and contralateral carotid artery identified (Fig. 1D) and harvested. Sections were divided and placed in 10% paraformaldehyde or kept dry, prior to rapid cryopreservation. After the vessels had been taken, the animals were sacrificed. 3.5. Histological analysis A total of five segments were studied from each transplanted vessel. Sections were taken from segments at 1 mm intervals. Paraformaldehyde-fixed tissue was embedded in paraffin and cross-sectioned at 5 µm intervals. They were stained with haemotoxylin and eosin (H&E), or Martius scarlet blue, the latter to stain fibrin. The percentage of intima coated with thrombus in the H&E staining was recorded blindly to the nearest 5% using a grid to subdivide the circumference and, by counting the number of squares of the grid involved, calculate the extent of involvement. To detect the deposition of pre-formed antibodies, 7 µm transverse sections were made through cryo-embedded tissue. These were then
Fig. 4. Deposition of immunoglobulin in allogeneic artery transplanted into sensitised pig. Arteries were excised and (in panels A, B and D) transplanted into recipients as indicated. They were then removed 48 h later. Tissues were studied for the deposition of antibodies using goat anti-pig immunoglobulin. Nuclei were counterstained with DAPI.
P.E. Herbert et al. / Transplant Immunology 19 (2008) 167–172
fixed in acetone at −20 °C for 20 min, and dried at room temperature. Each section was sequentially blocked with avidin-blocking agent, biotin-blocking agent and goat serum (1:20 diluted in PBS), being washed in PBS after each reagent. Goat anti-pig IgG (Bethyl Laboratories, USA) was then added to each section in a 1:1000 dilution and the sections left covered in the dark at 4 °C for 12 h. Each slide was then washed with PBS, and goat anti-pig immunoglobulin (1:1000) placed on each section and left at room temperature for 1 h. In the experience of our laboratory, this gave optimal staining. After washing, flourescein–avidin DCS (Vector Labs Inc., Peterborough, UK) was added at a 1:100 dilution to each section. This was washed off with PBS after 1 h and the slides mounted and stored at 4 °C. Nuclei were counterstained with 4¢,6¢-diamidino-2-phenylindole (DAPI) (Sigma, MO, USA). 4. Results We initially attempted to perform this study using jugular venous transplantation since we were concerned that high arterial pressures would decrease the chances of identifying a phenotype. However, the transplantation of jugular veins was universally technically unsuccessful, always resulting in thrombosis before the end of the operation. By contrast, carotid artery transplantation was always successful in our hands (Fig. 1). In all cases the inoculation of allogeneic PBMCs led to the development of donor-specific alloantibodies (Fig. 2). Similarly, all arteries allotransplanted into sensitised animals showed changes consistent with antibody-mediated rejection. These involved a profound mononuclear inflammatory infiltrate in both intima and media, a histological pattern not seen in any of the unsensitised or untransplanted controls (Fig. 3 and Table 1) Fibrin deposition was seen as a layer deposited in the intima of the allografts into sensitised animals (Fig. 3 panels E–H). This was absent from all of the controls. A one-way ANOVA test, comparing extent of coverage of luminal surface with fibrin, showed this difference to be significant (p b 0.05). No mural necrosis was seen in these vessels. There was relatively little variation in the extent of the pathology. To confirm the deposition of antibody, we stained for deposited immunoglobulin and demonstrated its presence in allogeneic arteries transplanted into sensitised animals, but none in controls (Fig. 4). 5. Discussion This novel vessel transplant model of acute antibody-mediated rejection is simple and easy to perform. We believe it has the potential to be of value in determining the mechanisms underlying a range of alloantibody-mediated pathologies, accommodation and the efficacy of various interventions. As described, this model is superficially similar to hyperacute rejection (HAR) in that it is an early response to pre-formed antibody. However, it is unlike HAR in that it is not associated with complete occlusion of the vessel lumen with subsequent ischaemia, oedema, heamorrhage and infarction. Precisely because we are using a large vessel, this model allows us to isolate the effects of sensitisation and antibody-binding to the allograft (both of which we have shown and which are the events currently of particular interest in all forms of antibody-mediated allograft pathology) without the disadvantage of the effects of immediate and acute (or hyperacute) vascular occlusion. The next step will be to dissect the mechanisms underlying antibodymediated graft damage (and accommodation), for example by studying the effects of different titres of antibody, the speed with which they are induced (i.e. primary versus secondary antibody responses), and the differential effects of the presence of pre-formed antibodies at the time of transplantation (with all its acute ischaemic and inflammatory perturbations) versus those of alloreactive antibodies induced after the initial acute insult has begun to settle. One concern is the relevance of large vessel antibody-mediated damage to the small vessel pathology on which we focus in clinical
171
practice. In fact, we (fortunately) rarely see large vessels now in renal biopsies. However, the early literature [12–16] makes it clear that “vessels of any size may be involved” in antibody-mediated damage [12]. In his seminal paper, Porter describes the involvement of large vessels in rejection, in one case reporting the very dramatic abrupt commencement of histological abnormalities at the line of the anastomosis of the transplant renal artery [16]. Mihatsch et al. et al. reported their findings on 30 nephrectomy specimens and concluded that the pathology “involves the entire arterial vascular hierarchy” [15]. Indeed, the veins were also involved, although less commonly in their study. In another study, of 58 grafts, Busch et al. et al. described the vascular lesions in allograft injury and found that “cortical arteries were most severely affected, but similar changes were found in vessels as large as main renal artery branches” [13]. For these reasons, we suggest that this large vessel allograft model is a useful surrogate for alloantibody-mediated damage of clinical relevance. This is based on the assumption that some features of the response of endothelium to alloantibody (both to develop rejection and accommodation) are generic. We believe this is a reasonable and pragmatic assumption, as it has to be offset against the facts that (a) performing whole organ grafts (to allow assessment of microvasculature) is a much more demanding and costly exercise than the vascular transplants outlined here and (b) if we are to use this model to dissect mechanisms underlying antibody-mediated graft damage and accommodation and to develop interventional strategies (for example by introducing putatively protective genes), this will be much easier to assess first in a simple arterial graft than in a whole organ. Indeed, we have already developed a system for introducing genes into pig carotid artery endothelium [17]. We next plan to use this to study interventions in antibody-mediated rejection using the model described in this paper. One such strategy involves the use of anticoagulant molecules to circumvent this process. The potential value of this process has already been demonstrated in a mouse model within our group [18]. It will also allow us to study the parameters that determine the development of accommodation with a view to being able to reproduce them clinically. Several models of vascular transplantation have been described, but most have involved rodents which have very small vessels which are difficult to manipulate [8–10]. The technique described in this paper is not especially demanding surgically and thus has a very high technical success rate. This makes it significantly better than solid organ models which are much more difficult to perform, have a high failure rate and are therefore expensive. However, it still makes it possible to focus on the endothelium and vessel wall, the primary area of interest in antibody-mediated rejection. Finally, the model is highly discriminating and can be very easily quantified. Acknowledgments This work was supported by a grant from the Hammersmith Hospitals Trustees’ Trustees' Research Committee. PEH was a Medical Research Council/Royal College of Surgeons of England Clinical Training Fellow. We would like to thank Professor Terry Cook for assistance with the histological analysis of the specimens and helpful discussion. References [1] Hippen B. The sensitized recipient: what is to be done? Am J Transplant 2006;6 (10):2230–1. [2] Bray RA, Nolen JD, Larsen C, Pearson T, Newell KA, Kokko K, et al. Transplanting the highly sensitized patient: the Emory algorithm. Am J Transplant 2006;6(10): 2307–15. [3] Stegall MD, Gloor J, Winters JL, Moore SB, Degoey S. A comparison of plasmapheresis versus high-dose IvIg desensitization in renal allograft recipients with high levels of donor specific alloantibody. Am J Transplant 2006;6(2):346–51. [4] Delikouras A, Dorling A. Transplant accommodation. Am J Transplant 2003;3 (8):917–8.
172
P.E. Herbert et al. / Transplant Immunology 19 (2008) 167–172
[5] Salama AD, Delikouras A, Pusey CD, Cook HT, Bhangal G, Lechler RI, et al. Transplant accommodation in highly sensitized patients: a potential role for Bcl-Xl and alloantibody. Am J Transplant 2001;1(3):260–9. [6] Gibney EM, Cagle LR, Freed B, Warnell SE, Chan L, Wiseman AC. Detection of donorspecific antibodies using HLA-coated microspheres: another tool for kidney transplant risk stratification. Nephrol Dial Transplant 2006;21(9):2625–9. [7] Vasilescu ER, Ho EK, Colovai AI, Vlad G, Foca-Rodi A, Markowitz GS, et al. Alloantibodies and the outcome of cadaver kidney allografts. Hum Immunol 2006;67(8):597–604. [8] Barcelos AS, Rodrigues AC, Silva MD, Padovani CR. Inside-out vein graft and insideout artery graft in rat sciatic nerve repair. Microsurgery 2003;23(1):66–71. [9] Isik FF, McDonald TO, Ferguson M, Yamanaka E, Gordon D. Transplant arteriosclerosis in a rat aortic model. Am J Pathol 1992;141(5):1139–49. [10] Alkhatib B, Freguin-Bouilland C, Litzler PY, Jacquot S, Lallemand F, Henry JP, et al. Antidonor humoral transfer induces transplant arteriosclerosis in aortic and cardiac graft models in rats. J Thorac Cardiovasc Surg 2007;133(3):791–7. [11] Pennington LR, Lunney JK, Sachs DH. Transplantation in miniature swine. VIII. Recombination within the major histocompatibility complex of miniature swine. Transplantation 1981;31(1):66–71.
[12] Porter KA. Renal transplantation. In: Heptinstall RH, editor. Pathology of the kidney. Fourth ed. Boston: Little Brown; 1992. p. 1865. [13] Busch GJ, Reynolds ES, Galvanek EG, Braun WE, Dammin GJ. Human renal allografts. The role of vascular injury in early graft failure. Medicine (Baltimore) 1971;50(1):29–83. [14] Kincaid-Smith P. Histological diagnosis of rejection of renal homografts in man. Lancet 1967;2(7521):849–52. [15] Mihatsch MJ, Zollinger HU, Gudat F, Schuppler J, Riede UN, Thiel G, et al. Transplantation arteriopathy. Pathol Microbiol (Basel) 1975;43(2-O):219–23. [16] Porter KA, Thomson WB, Owen K, Kenyon JR, Mowbray JF, Peart WS. Obliterative vascular changes in four human kidney homotransplants. Br Med J 1963;2(5358): 639–45. [17] Chen D, Morgan F, Berton I, Herbert PE, Lechler RI, Dorling A, et al. Developing a porcine transplantation model: efficient gene transfer into porcine vascular cells. Transplantation 2004;77(9):1443–51. [18] Chen D, Giannopoulos K, Shiels PG, Webster Z, McVey JH, Kemball-Cook G, et al. Inhibition of intravascular thrombosis in murine endotoxemia by targeted expression of hirudin and tissue factor pathway inhibitor analogs to activated endothelium. Blood 2004;104(5):1344–9.
Contents lists available at ScienceDirect
Transplant Immunology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t r i m
Brief communication
A pig allograft model of antibody-mediated rejection Paul E. Herbert a,c, M. Finn Morgan a,c, Ivan Berton a, Robert I. Lechler a, Anthony Dorling a, Gordon Williams c, Anthony N. Warrens a,b,⁎ a b c
Department of Immunology, Imperial College London, Hammersmith Campus, Ducane Road, London W12 0HS, UK Department of Renal Medicine, Imperial College London, Hammersmith Campus, Ducane Road, London W12 0HS, UK Department of Surgery, Hammersmith Hospital, Ducane Road, London W12 0NN, UK
A R T I C L E
I N F O
Article history: Received 24 March 2008 Accepted 7 May 2008 Keywords: Antibody-mediated rejection Anti-HLA antibodies Sensitisation Transplantation
A B S T R A C T Allograft rejection caused by antibodies in sensitised individuals remains a real problem in human allotransplantation. There would be value in a simple model of this process to evaluate the mechanisms involved in antibody-mediated damage and the development of accommodation, as well as the impact of potential interventions. We have thus developed a novel large animal model of this process using an allosensitisation system. Two inbred lines of miniature pigs that carry different major histocompatibility antigen haplotypes were used. Pigs of one line were sensitised by the sequential subcutaneous injection of major histocompatibility antigen-mismatched allogeneic peripheral blood mononuclear cells derived from the other inbred line. We demonstrated that this generated high titres of allospecific antibodies. We then transplanted carotid arteries from donors syngeneic to the priming cells into the sensitised animals. After 48 h these vessels showed a profound mononuclear cell inflammatory infiltrate in both intima and media, fibrin deposition, and luminal compromise with thrombus and antibody deposition. The mean endothelial surface affected by this process was 59.2%. No such pathology was seen in any of the controls. This model is technically simple to perform with few post-operative complications. It provides proof-of-principle of a model of antibody-mediated rejection which will be of potential value in elucidating the mechanisms underlying the process and the efficacy of interventions to prevent or treat it. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Pre-formed antibodies are a significant issue in allotransplantation, when the transplantation of sensitised recipients with pre-formed HLAspecific antibodies is being considered. In such situations, attempts to remove antibodies in advance of and following transplantation, combined with strategies to suppress further production of anti-HLA antibodies have met with some success [1–3]. However, the outcome in these situations is still inferior to transplantation in the non-sensitised recipient and the parameters that determine success have not been well defined. The situation is further complicated by the observation that, in some situations, the return of antibodies following transplantation does not cause the expected damage, because the graft has in some way adapted itself to be less susceptible to antibody-mediated damage, a process called “accommodation” [4,5]. In addition, new, highly sensitive techniques are now available that can identify the presence of low, previously undetected levels of antibodies [6,7]. Hence, there are several
variables that need to be teased out to understand the mechanisms underlying these phenomena. It is likely that they will only be elucidated using an appropriate animal model. Rodent models of vascular transplantation have been described, sometimes demonstrating the role of antibodies in transplant vasculopathy, but are technically difficult [8–10]. That is why we have developed a large animal model. 2. Objective In this study, we sought to prove the principle that it would be possible to study antibody-mediated allograft damage using a technically easy, large artery model in the pig. To do this we have studied a system with high levels of pre-formed allospecific antibodies. Having proved this principle, we believe this model will be useful to study the importance of different levels of antibodies, different timing of induction (before and after transplantation) and the mechanisms underlying different outcomes (rejection and accommodation). 3. Materials and methods
⁎ Corresponding author. Department of Immunology, Imperial College London, Hammersmith Campus, Du Cane Road, London, W12 0HS, UK. Tel.: +44 20 8383 2307; fax: +44 20 8383 2788. E-mail address: [email protected] (A.N. Warrens). 0966-3274/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.trim.2008.05.003
3.1. Animals Inbred lines of minipigs (c/c and d/d SLA (swine leucocyte antigen) phenotypes) [11] (kind gift of Dr David Sachs) were used. c/c
168
P.E. Herbert et al. / Transplant Immunology 19 (2008) 167–172
Fig. 1. Porcine carotid artery transplantation. The carotid artery was exposed (A) and a section excised (B). Another section of artery was transplanted with two end-to-end anastomoses (C) and the animal was allowed to wake up. Forty-eight hours later the vessel was exposed (D) and excised. Further details of the surgical methods involved are provided in Materials and methods.
animals were used as recipients of vessels (or untransplanted controls) and both c/c and d/d animals as vessel donors. All were greater than 20 kg at use. All procedures were performed in
accordance with the Animals (Scientific Procedures) Act 1986, after satisfying the internal ethical review process of Imperial College London.
Fig. 2. Generation of SLAd-allospecific antibodies in SLAc pigs. SLAc pigs were sensitised with weekly injections of SLAd cells (days 0, 7, 14 and 21) or not (controls). On regular occasions, blood was drawn and neat serum mixed with either SLAd (dd) (solid line) or control SLAc (cc) (broken line) cells and FITC-conjugated goat anti-pig IgG added. (A) Allospecific antibody is detectable by day 16. (B) High titres are observed after four weeks. In these flow cytograms, signal from SLAd cells are shown in outline and from control SLAc cells in solid graphs. This represents a representative example of four pigs.
P.E. Herbert et al. / Transplant Immunology 19 (2008) 167–172
169
Fig. 3. Generation of antibody-mediated damage following allosensitisation. Arteries were excised and (in panels B–D and F–G) transplanted into recipients as indicated. They were then removed 48 h later. Tissues were studied for gross histology using H&E staining (A–D) and for deposition of fibrin using Martius scarlet (E–H). The figure illustrates an untransplanted artery (A, E), a transplanted syngeneic artery (c/c into c/c) (B, F), an allogeneic artery transplanted (d/d into c/c) into an unsensitised animal (C, G) and an allogeneic artery transplanted (d/d into c/c) into an animal sensitised to d/d cells. In each panel, a low power view is represented in the upper quadrant. Within that, the area represented in high power is outlined by a back box and the high power view occupies the larger part of the panel.
170
P.E. Herbert et al. / Transplant Immunology 19 (2008) 167–172
FACScalibur (Becton Dickinson, San Jose, CA, USA) flow cytometer using CellQuest software (Becton Dickinson, San Jose, CA, USA).
Table 1 Extent of thrombosis in transplanted arteries Recipient
Artery donor
c/c c/c c/c c/c c/c c/c c/c c/c Untransplanted c/c
d/d d/d d/d d/d d/d d/d d/d c/c −
Sensitised + + + − − − − − −
No. of levels cut 10 10 5 5 5 5 5 5 5
% of intima covered by thrombus 39.1 +/− 12.6 52.5 +/− 27 86 +/− 12 0 0 0 0 0 0
Results of staining showing percentage of intima covered by thrombus 48 h after transplantation. Intimal thrombin deposition occurred only in the sensitised animals and in each animal a significant proportion of the vessel intima was in contact with this thrombus. There is a significant statistical difference between the sensitised animals and the controls (p b 0.05).
3.2. Sensitisation Whole blood from d/d animals was heparinised and diluted 1:1 in RPMI 1640 medium (GIBCO, Invitrogen Corporation, Renfrew, UK). Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll gradient centrifugation and then frozen at −80 °C in RPMI 1640 medium containing 10% foetal calf serum (Biowest, Oxford, UK) until ready for use. Each c/c pig to be sensitised received approximately 1×108 d/d PBMCs, suspended in 2 mL phosphate buffered saline (PBS) (Oxoid Ltd., Basingstoke, UK), subcutaneously into the scruff of the neck weekly for four weeks. Unsensitised controls received no injections. Serum was sampled at various time points to test for the development of SLAd-specific alloantibodies. 3.3. Flow cytometry 2 × 105 PBMCs were resuspended in 200 μL neat pig serum and left at 4 °C for 30 min. The cells were washed and FITC-conjugated goat anti-pig IgM or IgG (Bethyl Laboratories, Montgomery, Texas, USA) added. After 30 min at 4 °C, the cells were washed and analysed in a
3.4. Carotid transplantation Pigs were anaesthetised with halothane (1%) (Zeneca, Macclesfield, UK) in nitrous oxide and oxygen after pre-medication with 10 mg/kg ketamine hydrochloride intramuscularly (Pharmacia Upjohn/Pfizer, New York, USA). The right common carotid artery was exposed by blunt dissection (Fig. 1A). A section of native artery was removed and replaced with a 3 cm section of transplanted artery. The donor vessel was placed by end-to-end anastomosis using 6/0 prolene (Ethicon Ltd., Edinburgh, U.K.) (Fig. 1B–C). After haemostasis had been achieved, 2 mL of papaverine was gently squirted by syringe over the vessel through a 23 G needle (Tyco, Gosport, UK) and patency confirmed both by digital palpation and by a hand-held Doppler ultrasound probe. The wound was then closed with 4-0 vicryl (Ethicon Ltd., Edinburgh, UK), by re-opposing the previously cut fibro-muscular layers and finally the skin, and the animal allowed to wake. Forty-eight hours after the operation, each animal was anaesthetised again and the transplanted vessel and contralateral carotid artery identified (Fig. 1D) and harvested. Sections were divided and placed in 10% paraformaldehyde or kept dry, prior to rapid cryopreservation. After the vessels had been taken, the animals were sacrificed. 3.5. Histological analysis A total of five segments were studied from each transplanted vessel. Sections were taken from segments at 1 mm intervals. Paraformaldehyde-fixed tissue was embedded in paraffin and cross-sectioned at 5 µm intervals. They were stained with haemotoxylin and eosin (H&E), or Martius scarlet blue, the latter to stain fibrin. The percentage of intima coated with thrombus in the H&E staining was recorded blindly to the nearest 5% using a grid to subdivide the circumference and, by counting the number of squares of the grid involved, calculate the extent of involvement. To detect the deposition of pre-formed antibodies, 7 µm transverse sections were made through cryo-embedded tissue. These were then
Fig. 4. Deposition of immunoglobulin in allogeneic artery transplanted into sensitised pig. Arteries were excised and (in panels A, B and D) transplanted into recipients as indicated. They were then removed 48 h later. Tissues were studied for the deposition of antibodies using goat anti-pig immunoglobulin. Nuclei were counterstained with DAPI.
P.E. Herbert et al. / Transplant Immunology 19 (2008) 167–172
fixed in acetone at −20 °C for 20 min, and dried at room temperature. Each section was sequentially blocked with avidin-blocking agent, biotin-blocking agent and goat serum (1:20 diluted in PBS), being washed in PBS after each reagent. Goat anti-pig IgG (Bethyl Laboratories, USA) was then added to each section in a 1:1000 dilution and the sections left covered in the dark at 4 °C for 12 h. Each slide was then washed with PBS, and goat anti-pig immunoglobulin (1:1000) placed on each section and left at room temperature for 1 h. In the experience of our laboratory, this gave optimal staining. After washing, flourescein–avidin DCS (Vector Labs Inc., Peterborough, UK) was added at a 1:100 dilution to each section. This was washed off with PBS after 1 h and the slides mounted and stored at 4 °C. Nuclei were counterstained with 4¢,6¢-diamidino-2-phenylindole (DAPI) (Sigma, MO, USA). 4. Results We initially attempted to perform this study using jugular venous transplantation since we were concerned that high arterial pressures would decrease the chances of identifying a phenotype. However, the transplantation of jugular veins was universally technically unsuccessful, always resulting in thrombosis before the end of the operation. By contrast, carotid artery transplantation was always successful in our hands (Fig. 1). In all cases the inoculation of allogeneic PBMCs led to the development of donor-specific alloantibodies (Fig. 2). Similarly, all arteries allotransplanted into sensitised animals showed changes consistent with antibody-mediated rejection. These involved a profound mononuclear inflammatory infiltrate in both intima and media, a histological pattern not seen in any of the unsensitised or untransplanted controls (Fig. 3 and Table 1) Fibrin deposition was seen as a layer deposited in the intima of the allografts into sensitised animals (Fig. 3 panels E–H). This was absent from all of the controls. A one-way ANOVA test, comparing extent of coverage of luminal surface with fibrin, showed this difference to be significant (p b 0.05). No mural necrosis was seen in these vessels. There was relatively little variation in the extent of the pathology. To confirm the deposition of antibody, we stained for deposited immunoglobulin and demonstrated its presence in allogeneic arteries transplanted into sensitised animals, but none in controls (Fig. 4). 5. Discussion This novel vessel transplant model of acute antibody-mediated rejection is simple and easy to perform. We believe it has the potential to be of value in determining the mechanisms underlying a range of alloantibody-mediated pathologies, accommodation and the efficacy of various interventions. As described, this model is superficially similar to hyperacute rejection (HAR) in that it is an early response to pre-formed antibody. However, it is unlike HAR in that it is not associated with complete occlusion of the vessel lumen with subsequent ischaemia, oedema, heamorrhage and infarction. Precisely because we are using a large vessel, this model allows us to isolate the effects of sensitisation and antibody-binding to the allograft (both of which we have shown and which are the events currently of particular interest in all forms of antibody-mediated allograft pathology) without the disadvantage of the effects of immediate and acute (or hyperacute) vascular occlusion. The next step will be to dissect the mechanisms underlying antibodymediated graft damage (and accommodation), for example by studying the effects of different titres of antibody, the speed with which they are induced (i.e. primary versus secondary antibody responses), and the differential effects of the presence of pre-formed antibodies at the time of transplantation (with all its acute ischaemic and inflammatory perturbations) versus those of alloreactive antibodies induced after the initial acute insult has begun to settle. One concern is the relevance of large vessel antibody-mediated damage to the small vessel pathology on which we focus in clinical
171
practice. In fact, we (fortunately) rarely see large vessels now in renal biopsies. However, the early literature [12–16] makes it clear that “vessels of any size may be involved” in antibody-mediated damage [12]. In his seminal paper, Porter describes the involvement of large vessels in rejection, in one case reporting the very dramatic abrupt commencement of histological abnormalities at the line of the anastomosis of the transplant renal artery [16]. Mihatsch et al. et al. reported their findings on 30 nephrectomy specimens and concluded that the pathology “involves the entire arterial vascular hierarchy” [15]. Indeed, the veins were also involved, although less commonly in their study. In another study, of 58 grafts, Busch et al. et al. described the vascular lesions in allograft injury and found that “cortical arteries were most severely affected, but similar changes were found in vessels as large as main renal artery branches” [13]. For these reasons, we suggest that this large vessel allograft model is a useful surrogate for alloantibody-mediated damage of clinical relevance. This is based on the assumption that some features of the response of endothelium to alloantibody (both to develop rejection and accommodation) are generic. We believe this is a reasonable and pragmatic assumption, as it has to be offset against the facts that (a) performing whole organ grafts (to allow assessment of microvasculature) is a much more demanding and costly exercise than the vascular transplants outlined here and (b) if we are to use this model to dissect mechanisms underlying antibody-mediated graft damage and accommodation and to develop interventional strategies (for example by introducing putatively protective genes), this will be much easier to assess first in a simple arterial graft than in a whole organ. Indeed, we have already developed a system for introducing genes into pig carotid artery endothelium [17]. We next plan to use this to study interventions in antibody-mediated rejection using the model described in this paper. One such strategy involves the use of anticoagulant molecules to circumvent this process. The potential value of this process has already been demonstrated in a mouse model within our group [18]. It will also allow us to study the parameters that determine the development of accommodation with a view to being able to reproduce them clinically. Several models of vascular transplantation have been described, but most have involved rodents which have very small vessels which are difficult to manipulate [8–10]. The technique described in this paper is not especially demanding surgically and thus has a very high technical success rate. This makes it significantly better than solid organ models which are much more difficult to perform, have a high failure rate and are therefore expensive. However, it still makes it possible to focus on the endothelium and vessel wall, the primary area of interest in antibody-mediated rejection. Finally, the model is highly discriminating and can be very easily quantified. Acknowledgments This work was supported by a grant from the Hammersmith Hospitals Trustees’ Trustees' Research Committee. PEH was a Medical Research Council/Royal College of Surgeons of England Clinical Training Fellow. We would like to thank Professor Terry Cook for assistance with the histological analysis of the specimens and helpful discussion. References [1] Hippen B. The sensitized recipient: what is to be done? Am J Transplant 2006;6 (10):2230–1. [2] Bray RA, Nolen JD, Larsen C, Pearson T, Newell KA, Kokko K, et al. Transplanting the highly sensitized patient: the Emory algorithm. Am J Transplant 2006;6(10): 2307–15. [3] Stegall MD, Gloor J, Winters JL, Moore SB, Degoey S. A comparison of plasmapheresis versus high-dose IvIg desensitization in renal allograft recipients with high levels of donor specific alloantibody. Am J Transplant 2006;6(2):346–51. [4] Delikouras A, Dorling A. Transplant accommodation. Am J Transplant 2003;3 (8):917–8.
172
P.E. Herbert et al. / Transplant Immunology 19 (2008) 167–172
[5] Salama AD, Delikouras A, Pusey CD, Cook HT, Bhangal G, Lechler RI, et al. Transplant accommodation in highly sensitized patients: a potential role for Bcl-Xl and alloantibody. Am J Transplant 2001;1(3):260–9. [6] Gibney EM, Cagle LR, Freed B, Warnell SE, Chan L, Wiseman AC. Detection of donorspecific antibodies using HLA-coated microspheres: another tool for kidney transplant risk stratification. Nephrol Dial Transplant 2006;21(9):2625–9. [7] Vasilescu ER, Ho EK, Colovai AI, Vlad G, Foca-Rodi A, Markowitz GS, et al. Alloantibodies and the outcome of cadaver kidney allografts. Hum Immunol 2006;67(8):597–604. [8] Barcelos AS, Rodrigues AC, Silva MD, Padovani CR. Inside-out vein graft and insideout artery graft in rat sciatic nerve repair. Microsurgery 2003;23(1):66–71. [9] Isik FF, McDonald TO, Ferguson M, Yamanaka E, Gordon D. Transplant arteriosclerosis in a rat aortic model. Am J Pathol 1992;141(5):1139–49. [10] Alkhatib B, Freguin-Bouilland C, Litzler PY, Jacquot S, Lallemand F, Henry JP, et al. Antidonor humoral transfer induces transplant arteriosclerosis in aortic and cardiac graft models in rats. J Thorac Cardiovasc Surg 2007;133(3):791–7. [11] Pennington LR, Lunney JK, Sachs DH. Transplantation in miniature swine. VIII. Recombination within the major histocompatibility complex of miniature swine. Transplantation 1981;31(1):66–71.
[12] Porter KA. Renal transplantation. In: Heptinstall RH, editor. Pathology of the kidney. Fourth ed. Boston: Little Brown; 1992. p. 1865. [13] Busch GJ, Reynolds ES, Galvanek EG, Braun WE, Dammin GJ. Human renal allografts. The role of vascular injury in early graft failure. Medicine (Baltimore) 1971;50(1):29–83. [14] Kincaid-Smith P. Histological diagnosis of rejection of renal homografts in man. Lancet 1967;2(7521):849–52. [15] Mihatsch MJ, Zollinger HU, Gudat F, Schuppler J, Riede UN, Thiel G, et al. Transplantation arteriopathy. Pathol Microbiol (Basel) 1975;43(2-O):219–23. [16] Porter KA, Thomson WB, Owen K, Kenyon JR, Mowbray JF, Peart WS. Obliterative vascular changes in four human kidney homotransplants. Br Med J 1963;2(5358): 639–45. [17] Chen D, Morgan F, Berton I, Herbert PE, Lechler RI, Dorling A, et al. Developing a porcine transplantation model: efficient gene transfer into porcine vascular cells. Transplantation 2004;77(9):1443–51. [18] Chen D, Giannopoulos K, Shiels PG, Webster Z, McVey JH, Kemball-Cook G, et al. Inhibition of intravascular thrombosis in murine endotoxemia by targeted expression of hirudin and tissue factor pathway inhibitor analogs to activated endothelium. Blood 2004;104(5):1344–9.