Porcine CMV Infection Is Associated with Early Rejection of Kidney Grafts in a Pig to Baboon Xenotransplantation Model

Abstract

Background

Recent survivals of our pig-to-baboon kidney xenotransplants have been markedly shorter than the graft survivals we previously reported. The discovery of high levels of porcine CMV (pCMV) in one of the rejected xenografts led us to evaluate whether this reduction in graft survival might be due to the inadvertent introduction of pCMV into our GalT-KO swine herd.

Methods

Archived frozen sections of xeno-kidney grafts over the past 10 years were analyzed for the presence of pCMV, using real-time PCR. Three prospective pig-to-baboon renal transplants using kidneys from swine delivered by caesarian section (C-section) and raised in isolation were likewise analyzed.

Results

Kidney grafts from which 8 of the 18 archived samples were derived were found to be pCMV-negative, had a mean graft survival of 48.3 days and were from transplants performed before 2008. None had shown signs of DIC and were lost due to either proteinuria or infectious complications. In contrast, 10 of the archived samples were pCMV positive, were from kidney transplants with a mean graft survival of 14.1 days, had been performed after 2008 and had demonstrated early vascular changes and decreased platelet counts. Three prospective xenografts from swine delivered by C-section were pCMV negative and survived an average of 53.0 days.

Conclusions

Decreased survivals of GalT-KO renal xenografts in this laboratory correlate temporally with latent pCMV in the donor animals and pCMV in the rejected xeno-kidneys. Transmission of pCMV to swine offspring may be avoided by C-section delivery and scrupulous isolation of donor animals.

INTRODUCTION

The shortage of human organ donors for transplantation has limited the care of patients with end–stage organ failure. The pig has attracted interest as the most likely xenograft donor for humans because of its size, physiologic compatibility, breeding characteristics and potential for genetic manipulation(1). With respect to genetic modifications, the elimination of the gene for α1,3–galactosyltransferase has provided α1,3-galactosyltransferase gene knockout (GalT-KO) pigs(2–4) which have overcome hyperacute rejection induced by anti-Gal antibody, considered the first major immunologic hurdle of xenotransplantation.

Initial trials using vascularized thymic tissue (as either a thymokidney or a vascularized thymic lobe plus kidney) in an attempt to induce tolerance of pig renal xenografts, showed marked improvement using GalT-KO donors, compared to results of previous studies in which hDAF or standard miniature swine donors were used (83 vs. 30 days maximal survival)(5,6). Furthermore, with a modified treatment regimen, in which we eliminated both steroids and whole body irradiation, we achieved a significant decrease in complication rates, extending the mean survival to over 50 days, and showed in vitro evidence of donor-specific non-responsiveness (7). However, since the end of 2008, renal xenograft survivals in this laboratory have decreased markedly, with most recipients losing their renal xenografts within 3 weeks, even with thymic co-transplantation. In one recent recipient, the finding of a high level of porcine CMV (pCMV) in the rejected kidney led us to hypothesize that pCMV, inadvertently introduced into our swine colony, may have resulted in early loss of porcine xenografts.

We have now carried out a retrospective analysis of the pig donors and baboon recipients of GalT-KO kidney xenotransplants this laboratory, to determine whether there was a correlation between the presence of porcine CMV (pCMV) and early graft loss. We have further tested our hypothesis by performing pCMV negative xenografts, from donor swine delivered through Caesarian section and raised in isolation.

RESULTS

Historical results (2003–2012) of life-supporting renal xenograft survivals with protocols directed toward tolerance induction

Figure 1 shows the survival of GalT-KO kidneys in recipient baboons treated with tolerance-induction protocols in this laboratory from 2003 to the year 2012. As seen in this figure, until middle of 2008, the average recipient survival was 53.2 days (n=18). All of the kidney xenografts maintained function, except for two cases in which immunosuppression was terminated in the third post-operative week due to infection. However, since late 2008, recipients survival declined significantly to 13.7 days (n=22 p<0.05), and all recipients died before day 30. Most of kidneys showed hemorrhagic changes.

Figure 1.

GalT-KO thymokidney xenograft survival in baboons treated with ant-CD40L based regimens at our center in the years 2003–2012. Bars indicate organ survival withmarkedly decreased graft survival has been observed since mid-2008. A difference of potential importance that was identified was the relocation of the breeding colony to a new facility in late 2007. Animals born in the new facility were used as donors of thymokidneys since mid 2008.

Variables examined

In an attempt to determine the cause of our observed decrease in xenograft survival we examined several variables including the surgical procedure, pre- and post-operative care, immunosuppression, levels of preformed anti-non-Gal antibodies, immunosuppression and donor source. All kidney xenografts were transplanted by the same surgeon (K. Yamada) who also trained and supervised the team providing pre- and post-operative care. Donor kidney harvests and recipient transplants were performed in the same operating rooms at our center, and total vascular anastomosis time in each case was within 30 minutes. All recipients in this study had been tested for the levels of cytotoxic, preformed non-Gal Nab and none showed complement-mediated cytotoxicity on GalT-KO PBMC higher than 40%. Serum creatinine levels on POD 4 was 0.83 ± 0.90mg/dl in transplants performed prior to the middle of 2008 and 0.77 ± 0.36 mg/dl thereafter, indicating no change in surgery-dependent ATN or accelerated humoral rejection prior to POD 7. Immunosuppression was similar in all cases (see Methods), although small modifications in the immunosuppressive regimen did occur in 2005 (6, 7). None of these changes were deemed likely to have had a major negative effect on graft survival, since they were small andwere introduced 2 years before the reduction of xenograft survival was observed. However, a difference of potential importance that was identified was relocation of the swine breeding colony to a new facility in late 2007.

First clue to a potential role of porcine CMV in xenograft survival

A recipient baboon subjected to the injection of GalT-KO bone marrow (BM) directly into the tibial bone marrow space as part of a new protocol directed toward tolerance induction through mixed chimerism, provided the first indication that porcine CMV may impair xenograft survival. The details and results of this model will be reported separately (Tasaki M. et al., manuscript in preparation). Although serum creatinine (S-Cre) levels in the recipient were stable for the first 7 days, the kidney graft then unexpectedly lost function, with hemorrhagic changes (Fig 2A and B) and the animal developed severe thrombocytopenia. In order to evaluate the recipient immune response, we excised the primary kidney graft and untied the left native ureter on post-operative day (POD) 13 (83 days after BM transplantation) and stopped all immunosuppression. The platelet count returned to normal levels in 3 days and the DIC-like condition resolved. Unexpectedly, the excised kidney was found to be infected with pCMV (8.3×10^6 copy per 300ng DNA as assessed by qPCR). In order to determine whether pCMV had caused the early loss of the primary renal graft, a second kidney from a pCMV-negative donor SLA matched to the primary kidney was transplanted into the baboon on day 204 (121 days following the removal of the primary renal graft). This second xenograft maintained function for 60 days thereafter. Although the animal developed proteinuria, kidney graft showed microangiopathy but no hemorrhagic changes (Fig 2C and D). Tests for pCMV remained negative in the second graft and neither pCMV nor baboon CMV was detected in the baboon tissues collected at necropsy, as assessed by qPCR (less than 4 copies). These data suggested that active pCMV infection might have been the cause of early loss of the primary graft, with coagulation abnormalities associated with the pCMV infected kidney graft.

Figure 2.

A and B show histology and gross findings of the first GalT-KO kidney transplanted to B336 that was found to be pCMV positive. A): HE staining showed glomeruli with segmental necrosis and segmental tubular necrosis with luminal granular debris and red blood cells. B): On gross examination, the graft appeared hemorrhagic. C and D show histology and gross finding of a second GalT-KO kidney in B336 that found to be pCMV negative. C) HE staining showed chronic glomerulopathy and interstitial fibrosis. D) Gross finding of the graft revealed no hemorrhagic lesions.

Retrospective correlation of the presence of pCMV with early graft loss

We next investigated the potential role of pCMV in decreased xenograft survival by retrospective analysis of frozen tissues archived over the last 10 years at our center. All animals for which samples of thymokidneys or lymph nodes were available (n=18) were tested for the presence of pCMV using species-specific primers and quantitative PCR.

Strikingly, all samples collected before 2008 were found to be negative for pCMV, while most of the samples collected after 2008 were positive for pCMV (Fig 3-A). Only one animal (B313 in Fig. 3-A) in our most recent series had received a pCMV negative graft. This animal died of pneumonia on POD 11, without evidence of early rejection or vascular changes. The timing of the change in survivals, therefore, appeared to correlate with the relocation of our breeding facility. Although it remains unclear exactly how and when the latent pCMV infection was acquired within the herd, it is conceivable that inadvertent contamination occurred during the relocation process and subsequently reached a high prevalence in the herd since 2008. As shown in Fig. 3B, retrospective grouping of recipients according to pCMV status of the kidneys they received (both with the same treatment regimen), demonstrated a marked difference in average survival, with the pCMV negative kidney recipients averaging 48.3 days and the pCMV positive kidney recipients averaging only 14.1 days (p<0.05).

Figure 3.

A) Retrospective real time PCR analysis for pCMV: None of the baboons who received thymokidneys from GalT-KO swine donors bred at our original facility contracted pCMV infection detectable by q-PCR. In contrast, all the recipients of thymokidney from donors bred at our current facility contracted the infection, with the exception of B311, a baboon who died of pneumonia on POD11 with a pristine graft and without alterations in the coagulation profile.

B) Survival of baboon recipients of GalT-KO thymokidneys stratified according to the pCMV status of the donor swine: The survival of baboons that received GalT-KO kidneys from pCMV negative donors (top) markedly prolonged when compared to the survival of recipients of pCMV positive GalT-KO kidneys (bottom).

Up-regulation of ICAM in pCMV positive renal xenografts

To determine whether there was a correlation between pCMV and other parameters of endothelial cell activation in xenograft kidneys that might explain the early rejection observed, we compared the expression of ICAM-1 and MHC class II antigen on the pig kidneys by immunofluorescense microscopy. Figure 4A–a, A–b and A–c (left column) showed naïve kidney and grafts from pCMV negative pig donors and figure 4A–d, A–e and A–f (right column) show naïve kidney and grafts from pCMV positive donors. ICAM-1 was not expressed in pCMV negative kidney grafts except for the glomerulus in the rejected kidney grafts (Fig. 4A–c). In contrast, ICAM-1 expression increased on the peritubular capillary in the kidney grafts from pCMV positive donors (Fig. 4A–e and A–f). MHC class II antigens were expressed on the peritubular capillary of all kidneys, however not on the glomerular capillary in the naïve kidneys (Fig. 4B–a and B–d) and non-rejected kidney grafts from pCMV negative donors (Fig. 4B–b). MHC class II antigens were up-regulated on the glomerular capillary in the rejected kidney grafts from pCMV negative donors (Fig. 4B–c) and all examined kidney grafts from pCMV positive donors (Figure 4B–e and -f).

Figure 4. Immunofluorescence staining with ICAM-1 (A) and anti-swine class II antibodies.

A) Comparison of ICAM1 expression in pCMV positive kidneys with negative kidneys following xenotransplantation. a–c were from pCMV negative donors and e-f were from pCMV positive donors: a and d) native swine kidneys; b and e) non-rejected kidney xenografts; c and f): rejected kidney xenografts. ICAM-1 expression (white arrows) is seen on the peritubular capillaries in pCMV positive kidneys (e and f) while pCMV negative rejected kidney had only a minimal expression of ICAM-1 (c).

B) The comparison of class II DR expression in pCMV positive kidneys with negative kidneys following xenotransplantation. a–c were from pCMV negative donors and e-f were from pCMV positive donors: a and d): naive kidneys; b and e): non-rejected kidney grafts, c and f): rejected kidney grafts. Up-regulation of class II on peritubular capillaries and glomeruli (white arrows) in a rejected pCMV negative kidneys (c) as well as pCMV positive rejected kidneys (e and f).

Prospective study of a potential strategy to avoid pCMV transmission and increase renal xenograft survival

We next performed a prospective study in an attempt to further assess the hypothesis that pCMV was involved in early xenograft loss. Classical transmission of pCMV in adults animals is via the nasal mucosa (8). Although in utero infection has been reported, the porcine placenta does not support extensive viral replication (9). We therefore performed Cesarean sections in an attempt to deliver pCMV-free donors. We then raised these animals in a HEPA-filtered cage isolated from the rest of the herd. Thymokidney grafts were prepared when the piglets reached 2 months of age and were transplanted into baboons 6–8 weeks thereafter (Fig. 5-A) (10).

Figure 5.

A) Schematic of the prospective experiment planned to obtain PCMV negative donors. Animal #20441, a sow with latent PCMV infection, underwent Cesarean section and delivered 2 piglets (animals #21398 and #21396) that were housed in HEPA filtered units and raised in isolation from the rest of the herd. Pig 21398 donated a kidney to B336 and a thymokidney to B344; pig 21396 donated a thymokidney to B341. Samples from kidney biopsies taken at the time of transplantation confirmed the absence of PCMV in the graft.

B) Survival of recipients of organs from swine prospectively delivered via Cesarean section and raised in isolated units in order to avoid PCMV infection. Average baboon survival was 50.3 days, consistent to the survival of animals that received organs from PCMV negative donors in our historical cohort (48.3 days).

Including the intra-bone BM recipient that received a pCMV negative kidney as a second graft (see above), these baboons (2 thymokidneys and one kidney) showed an average survival of 50.3 days (Fig 5-B). These animals were sacrificed due to causes other than rejection (pleural effusion and nephrotic syndrome), showing neither the marked decrease in platelets nor the alterations of coagulation profiles observed for the pCMV positive kidney recipients. Their survival rates were thus similar to those observed in animals receiving pCMV-negative TKs in the historical cohort. Comparison of the outcomes of the 3 animals prospectively transplanted with grafts from pCMV-negative donors with those of the historical recipients of pCMV-positive kidneys, reached statistical significance (p<0.05).

DISCUSSION

This study demonstrates a strong association between presence of pCMV in GalT-KO porcine kidneys and markedly decreased graft survival in our pig-to-baboon xeno-kidney transplantation model. There is likewise a strong association between pCMV and endothelial cell activation, presumably leading to graft loss. The data in this study confirm our previous report on the activation of pCMV following pig-to-baboon kidney transplantation (11) and clarify that an anti-Gal response was likely not the only cause of the resulting DIC and ureteral necrosis observed in that study. Moreover, our prospective study suggests that pCMV transmission can be avoided by C-section delivery of offspring even in the presumed presence of latent pCMV infection in the pregnant sow. These data strongly support raising of xeno-kidney donor candidates in a pCMV-free environment for successful pig-to-primate kidney xenotransplantation.

The mechanism by which pCMV induces endothelial cell activation leading to early and dramatic loss of xenografts remains unclear. A previous in vitro study showed that human umbilical vein endothelial cells (HUVEC) infected with human CMV exhibited markedly enhanced cell surface expression of ICAM-1 (12). Moreover, the up-regulation of ICAM-1 appeared to be a direct virus-mediated effect that required the presence of infectious virus, while effects such as cytokine-stimulation (13,14) and up-regulation of MHC class II antigens appeared to be an indirect effect of CMV (15). In our model, all of grafts had stable renal function for the first 4 days, indicating no graft damage associated with preformed Nab or surgical ATN. Immunofluorescence microscopy of excised grafts as early as day 10 showed the up-regulation of the ICAM-1 and MHC class II antigens on the endothelial surface in pCMV positive xeno-kidney grafts, while ICAM-1 expression was less obvious in pCMV negative kidneys. The activation of pCMV up-regulated ICAM-1 expression on the endothelial cells in the grafts was followed by the adhesion of activated lymphocytes and platelets. It seems likely that once the kidney was removed from the donor, and the latent pCMV was no longer controlled by the recipient pig’s immune system, it became activated, with subsequent proliferation, especially in the face of systemic immunosuppression of the recipient baboon (16). The resultant profound activation of endothelium of tubular capillaries might then lead to interstitial hemorrhage as early as day 7 in this protocol. In addition to the direct effects, indirect effects might have been present, such as a cytokine release and tissue modifications induced by replication of pCMV. Although we did not observe the characteristic pCMV inclusion bodies often associated with CMV infection, clinical cases have been reported in which CMV-specific inclusion bodies were not detected in the transplanted tissues even when CMV DNA and proteins were found in the transplanted kidney by PCR and immunohistochemistry (17). The histologic appearance of these tissues may also have been modified by use of ganciclovir prophylaxis (18).

The observation that the DIC and gastrointestinal bleeding observed in one baboon in this study disappeared after graftectomy suggests that the pCMV infection was restricted only to the porcine kidney. Thus, the systemic manifestations, including consumptive coagulopathy, were presumably due to factors released by baboon cells reacting to infection-induced damage within the xenograft kidney. All of our baboon recipients have been given ganciclovir as prophylaxis against baboon CMV (bCMV) infection, but unfortunately, as we have previously reported, ganciclovir has reduced efficacy against pCMV compared with human strains (18). Therefore, these results support the need to exclude pCMV from xenograft donors in the absence of effective anti-pCMV therapy. Like the human virus, pCMV is usually acquired early in life and remains latent in most animals (8,19); the prevalence of pCMV in commercial swine herds has been reported to be high (20). It seems therefore conceivable that many experiments using GalT-KO or transgenic GalT-KO swine may have been influenced by the presence of pCMV in the donor animals. Several recent studies have reported the effectiveness of mTOR inhibitors in preventing CMV reactivation in transplant recipients (21–23). Modification of immunosuppressive regimens in favor of mTOR inhibitors could therefore be beneficial in those xenotransplantation models in which pCMV represent a major obstacle.

Miniature swine have many characteristics making them ideal potential donors for clinical xenotransplantation (1). Due to their similar size to humans, they require far less space than domestic swine, which is an advantage for housing in closed, Specific Pathogen-Free or Good Laboratory Practice facilities. Our demonstration that delivery by cesarean section may eliminate pCMV from the progeny of infected animals would support the construction of such facilities, starting with cesarean-derived breeders. Availability of pCMV-free animals could be an important step toward achieving long-term kidney function in pig-to-primate models.

Materials and Methods

Animals

All animals were cared for according to the guidelines of the Massachusetts General Hospital Institutional Animal Care and Use Committee. Male or female recipient baboons (Papioanubis) were purchased from Mannheimer Foundation, Homestead, FL. Xenogeneic organs were obtained from by GalT-KO miniature swine. The generation of these animals has been published previously (4,7).

Surgery

(a) Thymokidney preparation

Thymokidneys were prepared by the implantation of autologous thymic tissue under the kidney capsule, allowed to heal in for 6–8 weeks to permit thymic revascularization prior to transplantation (24).

(b) Recipient preparation

All operations were performed under general anesthesia. Baboon recipients of thymokidneys were thymectomized 3 weeks prior to thymokidney transplantation. Indwelling plastic catheters (Saint Gobain Performance Plastics, Reading, PA) were inserted into the carotid artery and the internal and external jugular veins and splenectomy was performed 7 days prior to thymokidney transplant. The thymokidney transplantation procedure is identical to that of a normal orthotopic renal transplantation (5). All anastomoses were performed by the same surgeon (K. Yamada) and no surgical ATN developed.

Immunosuppression

All recipients received our tolerance inducing induction regimen that included transient T cell and B cell depletion. Details of immunosuppressive regimen described in Supplemental Digital Content.

Pathologic examination

Excised GalT-KO kidney tissue samples were divided into three sections. Two were histologic examination and one was for CMV analysis. Formaldehyde fixed tissue samples were stained using either hematoxylin and eosin (H&E) and periodic acid-Schiff (PAS) stains. Coded samples were examined by light microscopy, and rejection was diagnosed according to a standardized grading system (25). Immunohistochemistry was also performed on frozen sections using polyclonal antibodies reactive to ICAM-1, porcine MHC Class II, and IgG (polyclonal rabbit anti-human IgG (Dako, Denmark).

Preparation of peripheral blood mononuclear cells (PBMC)

PBMCs were prepared from whole blood as previously described (5). Details of immunosuppressive regimen described in Supplemental Digital Content.

Quantitative Real-Time PCR

Target DNA sequences were quantified by real-time PCR using the ABI PRISM 7700 Sequence Detection System (Perkin-Elmer, Foster City, CA). (See Supplemental Digital Content).

Quantification of pCMV DNA

Primers and probes specific for pCMV DNA polymerase gene showed no cross-reactivity with PLHV-1: pCMV sense, 5′-GTTCTGGGATTCCGAGGTTG-3′; antisense, 5′-ACTTCGTCGCAGCTCATCTGA-3; probe, 5′-6FAM-CAGGGCGGCGGTCGAGCTC-TAMRA-3′. Results are expressed as copy number of pCMV per 300ng of total DNA (tissues), The quantitative detection limit for this assay was 7 copies.

Statistical analysis

Statistical analysis: Creatinine values and days of survival were compared between groups by the Mann-Whitney U test using GraphPad Prism, Version 5.04 (Graphpad Software Inc.).

Abbreviations

ATN

acute tubular necrosis

GalTKO

alpha-1,3-galactosyltransferase knockout

NHPs

nonhuman primates

MGH

Massachusetts General Hospital

pCMV

porcine cytomegalovirus

Nab

natural antibodies

S-Cre

serum creatinine

CBC

complete blood counts

H&E

hematoxylin and eosin

PAS

periodic acid-Schiff

DIC

disseminated intravascular coagulation

POD

post-operative day

PLT

platelet