Donor-Directed MHC Class I Antibody Is Preferentially Cleared from Sensitized Recipients of Combined Liver/Kidney Transplants

Introduction

Renal transplantation is a well-established treatment for end-stage renal disease. There are an increasing number of liver failure patients with coexisting renal disease either due to potentially reversible causes of acute renal failure such as hepatorenal syndrome, or to chronic renal disease. In the acute setting, renal function usually recovers after liver transplantation. For those patients with chronic renal diseases simultaneous liver and kidney transplantation (SLKT) is the best therapeutic option. Since its landmark discovery, a positive cytotoxic cross-match due to preformed donor-specific HLA antibodies (DSA) has been considered a relative contraindication to kidney transplantation (1). Although strategies to eliminate antidonor antibody levels in sensitized patients are now being successfully applied to bring them to transplant, such patients remain at risk for antibody-mediated graft injury (2).

Numerous reports suggest that the presence of alloantibody at the time of liver transplantation have no deleterious impact on immediate patient or liver allograft survival (3,4). However, recent studies suggest that a positive cross-match may adversely influence outcome (5). While the impact of alloantibody on liver lacks the same degree of systematic analysis applied to kidney transplantation, it appears that, compared to the kidney, the liver allograft is more resistant to hyperacute rejection. In fact, immediately following liver transplantation, kidney transplantation has been safely performed despite the presumed preoperative presence of alloantibody based on preoperative positive cross-match. Previous studies reported that these patients converted to a negative crossmatch after liver transplantation suggesting decreases in alloantibody (6). Based on these clinical observations, it has been proposed that a liver allograft protects the kidney from hyper acute rejection by either absorbing or neutralizing DSA.

It is still unclear whether there is any long-term impact of a pretransplant positive cross-match on organ survival with SLKT. In fact, a recent case report described acute humoral rejection in the renal allograft of a highly sensitized SLKT recipient. That study did not discriminate whether it was the HLA class I and/or class II DSA that caused the rejection (7). The purpose of this report is to describe our experience of the impact of pre-existing HLA class II-specific DSA on the outcomes of SLKT.

Material and Methods

An IRB-approved retrospective review was performed from 7/1/2008 to 10/1/2009 analyzing the results of SLKT performed at Emory University. All patients had at least 1 month of follow-up. Patients with chronic (stage IV or V) or end-stage renal disease along with end-stage liver disease were considered for SLKT. These patients underwent HLA typing and the characterization of their anti-HLA antibodies.

The technique used for measuring DSA was a solid phase HLA antibody assay. For identification of class I and class II specificities we utilized a Luminex platform. Our samples were not treated with DTT and our probe was IGG specific. Each MFI refers to an individual bead's value. If there was more than a single bead with the same antigen (meaning different alleles), we averaged the value of the beads. In measuring MFI, background values were subtracted.

Our flow cross-match results were reported as delta mean channel fluorescence (dMCF) as well as molecules of equivalent soluble fluorescence (MESF). MESF and dMCF values were calculated by generating a standard curve using a set of beads with a known amount of fluorescence. Values of >60 for T cells and >100 for B cells were positive when reported by dMCF. Values of >2000 for T cells and >5000 for B cells were positive when reported by MESF. A value of 2000 MFI was established for our assays when our assays were initially validated wherein we compared the Luminex results to data generated by flow cytometry. Values greater than 2000 MFI typically result in a positive cross-match or a positive DSA that correlates clinically with the finding of antibody-mediated rejection (AMR).

At the time of organ availability, a virtual cross-match was performed and patients with a potential positive cross-match were identified. Although HLA antibodies were identified, the corresponding antigens were not listed as unacceptable antigens with UNOS. As standard clinical practice, a prospective flow cytometric cross-match was performed and confirmed the virtual cross-match prediction in every patient. At our institution, a desensitization protocol was applied to those with a positive T or/and B-cell positive cross-match. These patients received a total of 2 g/kg of sucrose free intravenous immunoglobulin (IVIG) in two stages. The first dose of 1 g/kg was administered preoperatively as soon as cross-match results were confirmed. The second infusion was initiated postliver reperfusion with targeted completion prior to renal reperfusion.

Patients underwent orthotopic liver transplantation via piggy-back technique. Subsequently, kidney transplantation was performed with the typical retroperitoneal approach if possible through a separate lower quadrant incision. If clinically indicated, a midline extension incision was made and the renal allograft was placed intra-abdominally. Basiliximab (20 mg) was administered after liver reperfusion and on the 4th postoperative day. Patients received methylprednisolone (500 mg i.v.) intraoperatively. Maintenance immunosuppression consisted of standard triple therapy with prednisone, tacrolimus (target daily trough level of 8–12 ng/mL) and mycophenolate mofetil. Corticosteroids were avoided in patients who were seropositive for hepatitis C virus. Antiviral prophylaxis consisted of intravenous ganciclovir (5 mg/kg) daily until converted to oral valganciclovir (dosage adjusted for renal clearance) for 12 weeks. All SKLT were placed on oral fluconazole for 12 weeks. Finally, all patients received trimethoprim/sulfamethoxazole (or atovaquone) for Pneumocystis jiroveci prophylaxis. When possible, patients also underwent 6 month renal allograft protocol biopsies.

HLA antibody monitoring for both class I and II DSA as performed in those with a positive cross-match. Immediate perioperative monitoring was performed with serum samples collected between the liver and kidney transplant procedures and within 12 h of renal transplantation. Assessment and monitoring of the DSA was performed using single antigen bead assay (Luminex based). These DSA were serially monitored. The ‘strength’ of the DSA was measured and reported as normalized mean fluorescence intensity (MFI) on HLA-coated beads. A class I or class II DSA was defined when an antibody directed to a donor antigen was >2000 MFI.

All SKLT recipients with an unexplained increase in serum creatinine over 25% of baseline underwent kidney biopsy. Biopsies underwent routine analysis including hematoxylin-eosin, Masson's trichrome, periodic acid shift (PAS), SV40 and C4d and were interpreted/scored for rejection according to the Banff 97 criteria. Acute AMR required all the following criteria: allograft dysfunction, histological evidence of allograft injury (acute tubular necrosis or injury, peritubular capillaritis, diffuse peritubular capillary C4d staining) and DSA. In the setting of acute cellular rejection, the treatment involved a corticosteroid bolus regimen with increases in the trough tacrolimus levels if less than 10 ng/mL. Patients with AMR underwent combined plasmapheresis with subsequent IVIG (100 mg/kg) administered every other treatment until DSA was below threshold. No therapy was initiated in the setting of circulating DSA in the absence of renal dysfunction.

All SLKT patients with evidence of cholestasis underwent Doppler ultrasound to confirm vascular patency. Subsequent evaluation included an MRI and MRCP to further assess vasculature and the biliary system. Finally, a core needle biopsy was obtained and interpreted using the Banff criteria for rejection. In the setting of acute cellular rejection, treatment involved a corticosteroid bolus regimen with increases in the trough tacrolimus levels.

Patients were divided into cohorts based on (1) the absence of DSA, (2) the presence of class I DSA only, (3) class II DSA only or (4) both class I and class II DSAs. This report focuses exclusively on those patients with both DSA and positive cross-match. Recipient and donor characteristics, renal and liver allograft function, and patient outcomes were compared when possible. Continuous variables were summarized and reported as mean and standard deviations. Categorical variables were compared using Fisher's exact test. Actuarial patient and graft survival and freedom from AMR were estimated by Kaplan–Meier methods and compared using log-rank test. p-values <0.05 were considered significant. All analyses were performed using SPSS statistical software version 5.0.

Results

Immunological monitoring

Class I antibodies were rapidly cleared by the liver based on our monitoring. In contrast, class II antibodies persisted after liver transplantation. Of the five patients with class II DSA, only one patient (pt 4) demonstrated rapid clearance of class II DSA. Similarly, this patient did not manifest acute humoral rejection and has maintained excellent renal function. There was no significant difference in the pretransplant MFI of the class II DSA nor in the MHC class II antigens between this one patient and the other class II sensitized patients (Table 3). Strikingly, all patients who had humoral rejection of their kidney had persistent anticlass II antibodies (p = 0.25) while diminution of the class I DSA (p < 0.01) (Figure 1).

Discussion

SLKT is becoming increasingly accepted as the standard of care for patients with both end-stage liver and kidney disease including patients highly sensitized to HLA antigens (8–10). Nonetheless, the role of HLA antibodies, specifically DSA, remains controversial. Current published data indicate that kidneys are relatively protected from both antibody and cell-mediated rejection when cotransplanted with the liver in sensitized patients (3–4,8–10).

We present data in this report that suggest that in patients who undergo SLKT with preformed DSA and a positive cross-match, there may be a differential reduction of class I DSA versus class II DSA. The mechanisms for this finding are not entirely clear at this time. Due to the lack of mechanistic information in our study population we can only speculate as to the cause of our findings.

One avenue for speculation with regard to our findings would be that the liver allograft in our patients had a role in the clearance or neutralization of circulating HLA antibodies. It has been consistently established that liver transplantation can be performed safely and successfully in the setting of HLA antibodies and even in the setting of ABO incompatibility (11). Although the immunological mechanisms through which the liver could protect the kidney have not been fully elucidated, some reports indicate that protection of the kidney allograft from antibody is attributable to immunoabsorption of circulating antibodies by the liver allograft either by cell bound or soluble HLA antigens (12) or through phagocytosis via Kupffer cells (13). Multiple reports continue to describe the phenomenon of the liver protecting the kidney from antibody. In these cases, liver transplantation results in a decrease in circulating HLA antibodies sufficient to convert a positive cross-match to negative, thus avoiding hyperacute rejection (6,14,15). This model of the liver acting as a continuous absorptive source was the basis of a clinical trial in Sweden wherein partial liver transplantation was performed in sensitized patients who only required a renal transplant in order to achieve a negative cross-match and allowing the renal transplant to be performed (16).

In this regard, our finding of differential clearance of class I versus class II DSA in SLKT patients with a positive cross-match could be explained by differential expression of class I versus class II in the transplant allografts. There is differential expression of class II HLA molecules in the liver parenchyma and vasculature compared to the renal allograft (11). Prior studies have also indicated that liver allografts predominantly express MHC class I on hepatocytes, as compared to class II (17).

However, a number of caveats must be extended to this line of reasoning. First, we must question whether antibody clearance by the liver allograft through immunoabsorption is the only or most effective mechanism by which we can explain our findings. Reports of hyperacute (antibody mediated) rejection in ABO-compatible liver transplant recipients would indicate that despite a relative resistance of the liver to hyperacute and humoral rejection, that clearance of or protection from alloantibody is not complete even in the setting of liver transplantation alone (18–21). Further investigations also suggest that the immunoabsorption model is too simplistic as expression of MHC in liver allografts appears to be a far more complex with expression of class I or class II being dependent on factors such as cell type as well as states of liver inflammation or organ maturation (22–25). And beyond immunoabsorption, the liver allograft may produce factors which promote global immune tolerance. To that effect, recently, nonparenchymal liver cells were identified as critical to the liver tolerance effect through the production of anti-inflammatory cytokines such as TGF-beta and increased expression of inhibitory costimulation molecules such as programmed cell death ligand one (PD-L1) (26).

The next caveat is that simply the presence of simultaneous allografts may be a critical factor in terms of the changes in the level of DSA we observed and subsequent graft outcomes. This idea that the simultaneous nature of combined liver–kidney transplant is important to graft outcomes was given credence by the studies of Simpson et al. They performed an analysis of the UNOS database in which patients who received kidney transplants subsequent to liver transplantation had inferior 1–3 year rejection-free graft survival compared to SLKT (27). This report would suggest that a liver allograft alone is not sufficient to protect kidney allografts. This may be in part due to the donor MHC antigens shared by the liver and kidney in the case of SLKT or reading these results more broadly, it may be that the global immune response at the time of SLKT differs from that of liver transplantation alone, or of kidney transplantation after liver transplantation.

The final caveat is that kidney transplantation itself changes the level of DSA. Arias and colleagues monitored DSA levels prior to and after kidney transplantation and found that patients without pretransplant DSA developed de novo DSA post-transplant and those with pretransplant DSA developed changes in class I and class II DSA levels post-transplant (28). Other investigators, namely Kerman and colleagues and Rebellato et al., have reported fluctuations in the level of DSA postrenal transplant and in some cases the disappearance of DSA entirely following renal transplantation (29,30). These findings would indicate that DSA production and levels are altered simply by the presence of a kidney allograft and that our novel findings of differential clearance of class I versus class II DSA, though tempting to ascribe to the presence of the liver allograft, require a more complex explanation.

Another alternative explanation for these observations could include simply a dilutional effect from intraoperative fluid and blood product resuscitation associated with operative bleeding. However, this would not likely account for the unique finding of consistently high MFIs for the class II DSAs after liver transplantation and the persistence of class II DSA in the face of stable liver allograft function even at 18 months of follow-up. Furthermore, blood transfusions associated with the SLKT in this series were not excessive.

The high incidence of early acute humoral rejection in the cohort of patients with class II DSA is striking. In these patients, decline in renal allograft function occurred within the first 2 weeks post-transplant. A unique immunological aspect of this study is the relative absence of class I DSA with a persistence of the class II DSA at the time of AMR. The data would suggest that the class II DSA is the pathogenic agent. The deleterious impact of persistent DSA is becoming increasing established in the renal transplant literature (15,16). Among 2278 renal allograft recipients, 1 year graft failure was worse in those patients who developed HLA antibodies than those without (8.6% vs. 3.0%; p<0.01). Renal allograft recipients with a positive pretransplantation cross-match have a higher incidence of transplant glomerulopathy seen in 1 year protocol biopsies (31). In addition, acute humoral rejection itself has been shown to be a risk factor for transplant glomerulopathy (OR: 17.5) (31). Gloor et al. reported that pretransplant DSA, especially those that were against class II antigens, were associated with transplant glomerulopathy (32). Finally, transplant glomerulopathy is a morbid predictor of graft failure even in the subclinical setting with graft survival of only 50% 3 years after detection (31). At 10 months post-transplant, one patient in this study has significant transplant glomerulopathy with chronic humoral rejection. In a recent case report, one SLKT patient demonstrated acute humoral rejection in the renal allograft with demonstrated class I and II DSA (7). It is puzzling why one patient (pt 4) had significant and persistent elimination of both class I and II DSA. It is unclear if the desensitization effects of IVIG could account for the clinical success. Ongoing immunological mechanistic studies involving allospecific memory B-cell populations and isotype analysis of the DSA in these sensitized patients may shed light on this. In summary, one would anticipate that the SLKT recipients who demonstrated persistent antidonor HLA antibodies may have deteriorating renal function and earlier graft loss than those without alloantibodies.

This preliminary report has several deficiencies that include: single institution, limited sample size, early follow-up and limited mechanistic data. Nonetheless, these exciting early findings are consistent enough to warrant modifying clinical practice and potentially organ allocation strategies along with offering new opportunities for further investigation. First, a liver allograft may not be fully protective of the renal allograft, especially in the setting of a highly sensitized individual with pre-existing MHC class II DSA. Despite attempts to acutely desensitize the recipients post-transplant with high doses of IVIG, there was an 80% incidence of acute humoral rejection with renal dysfunction within the first 2 weeks post-transplant. We have modified our own clinical pathway such that future SLKT recipients with established class II DSA will be screened for unacceptable class II donor antigens and thus avoid transplantation with a donor to which they have preformed class II antibody. We also intend to use IVIG preoperatively and postoperatively with sequential monitoring of these anti-HLA antibodies. In addition, monthly IVIG therapy in highly sensitized patients while on the waiting list will be administered in an attempt to desensitize these patients prior to transplantation. Simply prolonging the interval between the liver and the renal allograft in the setting of class II antibodies would likely be insufficient given antibody persistence in the perioperative period. In the setting of highly sensitized patients who will require SLKT, careful monitoring of anti-HLA antibodies and aggressive interventions are critical in order to ensure successful long-term renal allograft function.

Finally, though our findings are preliminary, they do represent novel information about the role of antibody in the context of liver and kidney transplantation and have already significantly altered the clinical practice at our institution. Going forward, prospective comparison of these patients to those undergoing renal transplant alone with known DSA will help in determining the role of the liver in explaining our findings. Furthermore, our aims will be to prospectively study similar patients as reported in this study and obtain sufficient samples of allograft tissue (both kidney and liver) and cells at the time of transplantation to not only study MHC expression but to perform mechanistic studies that would provide better insight into our findings and help clarify the nature of the immune response in SLKT patients with positive cross-matches and thus hopefully leading to improved graft outcomes and patient survival.