Donor Chimerism, Kir Ligand Mismatch and Cytokine Levels (Tnfα, Il17, Gmcsf, Il1β) and their Association with Graft Rejection in Liver Transplant Recipients

Methods: Seventy-four liver transplant recipients and their respective donors were studied, providing a total of 468 post-transplantation samples in the chimerism study. A total of 23 Liver Transplant patients and their respective donors took part in HLA typing and KIR genotyping, phenotyping, and cytotoxicity studies. 62 liver transplant recipients were analyzed in this cytokine study (62 pre-transplant and 109 post-transplant serum samples). Chimerism study was carried out by quantitative real time PCRs and 7 indels and 14 SNPs were detected with TM probes. A study of NK-cell subsets was performed on fresh samples by multiparametric flow cytometry (Becton Dickinson). KIR genotyping and KIR ligand were analyzed by PCR with a KIR typing Kit (Milteny Biotec). Cytotoxicity assays were monitored in a conventional 2-hour europium-TDA release assay (Perkin-Elmer). Biorad “17-plex Kit on a luminex” platform was used for cytokines measurement.


Background
Long-term acceptance of transplanted organs without requirement for indefinite immunosuppression remains the ultimate goal of transplant clinicians and scientists [1]. Currently, more than 90% of all liver transplant recipients are treated with the calcineurin inhibitors cyclosporine or tacrolimus [2]. Patients are faced with severe complications of immunosuppressant therapy such as infections, metabolic disturbances, malignancies [3], and drug-specific toxicity such as nephrotoxicity when treated with these inhibitors [4,5].
The incidence of chronic rejection is much lower in liver transplantation cases, compared to other organ grafts [6], and approximately 20% of stable recipients can be weaned off immunossuppresive medication without developing graft rejection.
New recommendations on this subject, included in a meta-analysis study, have shown that randomized trials are needed to evaluate calcineurin inhibitor reduction versus continuation of calcineurin inhibitor treatment; but these randomized trials should only be performed in carefully selected patients [2].
The proposal that microchimerism induces tolerance was first put forward by Starzl et al. [7]. Starzl and Zinkernagel hypothesized that organ engraftment is a form of chimerism-dependent partial tolerance [7][8][9][10][11][12][13] and this tolerance can be inferred from the amount of immunosuppression necessary to maintain the stable function and structure of the graft.
Previously we have reported that "Stable, high levels of donor chimerism (DC), in the absence of other major clinical events, may be a marker of transplantation tolerance" [14] in a cohort of fortytwo liver transplant recipients without lowering the amount of immunossuppression. Nonspecific genomic (NK, CD T cell genes), immunophenotypic (CD4+CD25highFOXP3+, T cells), and HLA (soluble HLA-G) signatures also appear to be ready for testing as tolerance predictors in prospective studies [15][16][17][18].
It has been demonstrated that donor NK cells may promote engraftment, prevent graft-versus host disease, control infections, and reduce the risk of leukemia relapse in hematopoietic stem cell transplantation (HSCT). It has also been established that NK cell alloreactivity is affected by many determinants, including donor versus recipient KIR ligand compatibility, KIR-HLA receptor ligand mismatch, and the milieu of cytokine such as IL-12 and IL-18 [19][20][21]. KIR genotype and KIR-HLA-C ligand compatibility play roles in recurrence and progression of hepatitis C [22] and graft rejection [23] in liver transplant recipients.
In the field of hematology, the killer immunoglobulin-like receptor /HLA ligand matching plays a part in T-cell complete DC in hematopoietic stem cell transplantation. And Natural Killer cells can exert a graft vs. tumor effect in haploidentical stem cell transplantation for pediatric solid tumors. We hypothesized that NK function and KIR/ HLA matching could modify DC levels in liver recipients. Taking into account the immunological aspects, we also studied cytokine levels as potential biomarkers to predict acute cellular rejection. We also examined these cytokines in relation to DC and the NK function.

Chimerism study
Characteristics of recipients and donors: A total of 80 Deceased Donor Liver Transplant (DDLT) recipients and their respective donors took part initially in this study, but eventually it was conducted with 74 LT recipients. Six LT cases were not studied, non-informative polymorphism between recipient and donor was found in two of them, two patients had died before the first post-transplantation samples were collected, and in two cases, donor samples were not available. All patients underwent transplantation at the General Surgery of the Alimentary Tract and Abdominal Organ Transplantation Department of the 12 de Octubre Hospitals (Madrid, Spain) between October 2008 and January 2011. The study was approved by the Ethical Institutional Review Board of 12 de Octubre Hospital and all patients gave their consent to participate in the study in compliance with the Declaration of Helsinki. The characteristics of patients and immunosuppressive regimen are summarized in Table 1.
Sample collection and DNA preparation: A total of 792 samples were studied. 156 were pre-transplantation from both donors and recipients; 468 were post-transplantation peripheral blood samples from recipients (from 74 recipients) and 168 were post-transplantation serum samples from recipients (from 25 recipients).

KIR genotyping and KIR ligand (HLA-I typing):
Fifteen human KIR genes and two pseudo genes were analyzed by PCR with a KIR typing kit (Miltenyi Biotec, Bergisch Gladbach, Germany). KIR A and B haplotype were performed as described previously by A. Pérez-Martínez et al (BMT 2012) KIR ligand HLA-C allotypes (C1 and C2) and HLA-B allotypes (Bw4) were determined using high-resolution PCR-sequence-based typing. The equations used to calculate spontaneous and specific cytotoxicity were also previously defined by Perez-Martinez et al. [26].

Statistical analyses
Statistical analyses were performed with SPSS statistical software, version 15.0 (SPSS, Chicago, IL). To analyze associations between discrete and categorical numerical variables, the χ 2 test and Fisher test were used. A comparison of the mean values of different groups was carried out with the Student unpaired t test, the Wilcoxon test,   and the analysis of variance test. A forward stepwise regression logistic analysis was performed to find variables associated with graft rejection. Survival curves were calculated according to the Kaplan-Meier method, and differences between curves were evaluated with the log-rank test. A P value less than 0.05 was considered to be significant. We defined "graft rejection" as liver damage diagnosed by biopsy or that had good evolution with an increased immunosuppressant therapy. We only considered acute rejection episodes in this work.
The diagnosis of liver toxicity was made by excluding other potential causes, including infections, congestive heart failure, liver disease relapse or graft rejection, and was considered highly probable if the patients showed elevated liver enzymes (>10 fold increase in γ-glutamyl transferase ± aspartate aminotransferase ± alanine aminotransferase) without associated infections or without response to antirejection therapy.

High levels of donor hematopoietic chimerism were associated with non-rejection liver recipients
An informative polymorphism to differentiate recipient and donor was found in 74 of 76 cases (97.4%). DC was detected in 241 of the 468 samples (51.5%) with a median of 0.02% (range 0.0045-175%). DC was expressed as a percentage of donor DNA in relation to total DNA in the post-transplant recipient samples. No difference in mean DC in peripheral blood was detected between the rejection and non-rejection group, when we compared mean DC of all patient samples (in log) ( Figure 1A). Nevertheless, high levels of DC in serum samples were associated with non-rejection liver recipients ( Figure 1B). Donor/recipient chimerism was studied along 30 months in peripheral blood and serum samples ( Figures 1C and 1D). The analysis of posttransplant evolution by time intervals showed high DC during the first three months in the group without graft rejection in peripheral blood and serum samples (p=0.005 and p=0.016, respectively) ( Figure 2). Higher DC levels were detected in serum samples than in peripheral blood samples (7.23% ± 24.86 vs a 1.77% ± 13.59; p=0.006) (Figures  2A and 2B).
Chimerism levels were correlated with platelet count (r0.234, p<0.001); or inversely with leucocyte count (r-0.139; p-0.005) and lymphocyte count (r-0.188; p<0.001). We observed higher DC in female recipients and male donors than other combinations (p=0.003), and higher DC in RH+ recipients and RH-donors than other combinations (p<0.001). There were no differences in distribution depending on gender or blood group between recipients and donors. In the same way, significant differences were not detected in relation to split graft or nonheart-beating donor. Similar to bone marrow transplantation, DC level was correlated inversely with Mycofenolate level (r-0.173; p=0.022), but no correlations were found for TAC or CYA levels.  post-transplant CD56 bright NK cells were 4.61% vs 15.70% (p=0.0029) (Figure 3). No differences were observed at other times or NK subsets. Only data from one pair of donor/recipient samples corresponding to a liver rejection case was available. CD56 dim NK percentage for this recipient was similar to non-rejection median (98.60% vs 87.89% vs) and CD56 bright NK was 1.51% (median 11.49%).
Disparities in both NK cell subsets were found when comparing liver recipient and donor percentages. Median CD56dim NK cells in recipients was 89.034% vs 56.86% in donors (p=0.009). Median CD56bright NK cells in recipients was 4.61% vs 2.94% in donors (p=0.015); Median NK bright/NK dim ratio was 0.0512 vs 0.17 in recipients and donors respectively (Figure 4).

Impact of KIR-ligand mismatch on allograft outcome in liver transplantation
In the study of alloreactivity, a higher NK activity was shown in pre-transplant recipients than in post-transplant recipients or in donor samples ( Figure 5). KIR-ligand mismatch/match was defined according to van Bergen et al. [27]. KIR ligand mismatch was detected in 12/22 cases (54.5%) and a tendency to association was observed between acute rejection and KIR ligand mismatch: patients with alloreactivity, 5 cases in the group with no rejection and 7 in the rejection group; patients without alloreactivity, 8 cases in the group with no rejection and 2 in the rejection group (p=0.069). We investigated the impact of the number of active or inhibitory KIR genes in recipients. The presence of KIR haplotype B (containing multiple activating KIR genes) has been reported to affect outcomes in hematopoietic stem cell transplantation [28,29]. But we did not observe influence of recipient and donor KIR genotypes on liver rejection.

Donors lacking C1 ligands (C2/C2) exhibited increased risk of acute rejection:
We observed no impact of HLA-C2 (lacking C1 ligands) of liver recipients on graft rejection rates (p=0.831), but the existence of lacking C1 ligands (C2/C2) in the donors was associated with graft rejection (17 cases with donor HLA-C2, 7 of them in the non-rejection group and 10 of them in the rejection group). 6 cases without donor HLA-C2 were in the non-rejection group (p=0.012). We observed no association between activating recipient KIR genes and graft rejection. We found 9 cases with activating KIR genes in recipients (7 cases nonrejection and 2 cases with graft rejection); and 14 cases without activating KIR genes (6 cases non-rejection and 8 cases with rejection, p=0.099).

Pre and post-transplant IL6 levels showed an impact on recipient overall survival. Lower TNF and IL17 levels and higher GMCSF and
IL1β levels were associated with low rejection free survival: Only pre and post-transplant IL6 levels showed an impact on recipient overall survival (OS). Lower OS with high IL6 levels was observed in both determinations (p=0.022 and p=0.022) (Figures 8A and 8B).
We investigated the relevance of cytokine levels on rejection free survival (RFS) and observed an impact on rejection related to IL17, TNFα, IL1β and GMCSF levels. Lower IL17 (p=0.008) and TNF levels (p=0.009) and higher IL1β (p=0.010) and GMCSF levels (p=0.008) were associated with low RFS (Figures 8C-8F  This cytokine study was made in 62 pre-transplant recipient samples: Graph A: Low cytokine levels in pretransplant recipient samples were associated with a tendency to graft rejection: mean IL10 was 1.14 vs 7.91 pg/ml (p=0.076). Graph B: Low MIP1b levels in pretransplant recipient samples were associated with a tendency to graft toxicity: mean MIP1b was 1.33 vs 7.09 pg/ml (p=0.093). Graph C: Low TNF alpha levels in pretransplant recipient samples were associated with liver disease relapse: mean TNF alpha was 5.015 vs 62.102pg/ml, p=0.020. But this result could be related to a particular liver disease such as HCV. This cytokine study was made in 109 post-transplant recipient samples: Graph D: Low GMCSF levels in post-transplant recipient samples were associated with graft rejection: mean GMCSF was 28.9 vs 81.7 pg/ml (p=0.006).
p=0.089   Starzl [7] was the first to make a strong argument about allogeneic microchimerism, in which the donor passenger leukocytes that migrate widely into the recipient's lymphoid tissues are essential for the maintenance of clonal exhaustion-deletion. This is induced by the initial flood of passenger leukocytes during the first few weeks after transplantation, and the survival of passenger leukocytes is associated with long-term acceptance of the graft.

Discussion
We reported in a previous work [14] that stable levels of DC, in the absence of other major clinical events, may be a marker of transplantation tolerance, and may help to tailor immunosuppressive treatment in liver transplantation. We have confirmed, in this prospective work, the relevance of DC observed in the first ten post-transplantation months. However, other variables might influence the levels of DC, especially when they are determined in serum samples. High chimerism levels in serum samples in cases with liver disease relapse can be related to DNA released from the graft where the graft has an endothelial injury. In this sense, our DC results in peripheral blood would reflect more appropriately the positive effect on tolerance.
Natural killer (NK) cell function can be modulated by killer cell immunoglobulin-like recipients (KIR), which interact with human leukocyte antigen (HLA) class I molecules on target cells. KIRligand mismatching has been shown by van Bergen et al. [27] to be a significant risk factor for long-term graft loss in HLA-A, -B and -DR compatible kidney transplants. Nevertheless, Tran et al. [30] studied this finding, and performed genotyping on HLA-A, -B and -DR of 608 deceased-donor kidney graft recipients, and found no impact of KIR-ligand mismatch on allograft outcome in HLA-compatible kidney transplantation. Recently, Minguela et al. [31] studied 402 liver recipients and found that KIR gene-gene mismatch is associated with acute rejection and KIR2DS4+/C ligands also influence short-term graft survival. Our results support the findings that KIR mismatch has influence on acute rejection although the difference is not significant, probably because of the very small cohort in our study. Other authors [22] described that KIR genotype and KIR-HLA-C ligand compatibility play roles in the recurrence and progression of hepatitis C in liver transplant recipients but our results have not shown this.
The expansion of the cytokine-producing CD56 (bright) NK cell subset is a main feature of lymphocyte reconstitution after allogeneic hematopoietic stem cell transplantation (HSCT) [32]. We observed an increase of CD56(bright) NK cells during first post-transplant month compared to the pre-transplant period. Despite our cohort to study CD56 NK subtypes only included a graft rejection recipient, it should be noted that a clearly lower NKbright/NKdim ratio was obtained in this case.
We detected lower cytokine levels in pre-transplant recipient samples associated with liver toxicity (MIP1b) and liver disease relapse (TNFαlpha). These results have not been previously published. Papers with correlations between graft outcome and pre-transplant cytokine determinations are few in liver transplantation, but Friedman´s work [33] showed in multivariate analyses that patients experiencing early allograft dysfunction had lower pre-operative IL-6 and higher IL-2R levels.
Nader Najafian et al. published [34] a paper aimed at determining the utility of a noninvasive cytokine assay in screening of acute rejection and their analyses revealed a strong association between posttransplant IL-6 levels and acute rejection after multivariate adjustment for clinical characteristics (P<0.001) . Although six cytokines (IL-IB, IL6, TNF, IL4, GM-CSF and MCP-1) distinguished acute rejectors in the training cohort, logistic regression modeling identified one IL-6 as the best predictor. But we found no relation between post-transplant IL-6 levels and acute rejection cases although we observed lower OS with high IL6 levels. We detected lower post-transplant GMCSF levels to be associated with graft rejection. Nevertheless, lower IL17 (p=0.008) and TNF levels (p=0.009) and higher IL1β (p=0.010) and GMCSF levels (p=0.008) were associated with low RFS in the survival analysis. A possible origin for the high GMCSF secretion pointed at the increased CD56 Bright NK population [35]. Elevated GMCSF might in turn induce macrophage differentiation and as result an increased IL1β production. This hypothesis would imply an important role for CD56 Bright NK cells on graft rejection. IL17 is a cytokine which acts as a potent mediator in delayed-type reactions, with a pro-inflammatory function and acts synergistically with TNF and interleukin-1. IL17, secreted by Th 17 cells, is a highly inflammatory cytokine with a robust effect on stromal cells in many tissues [36][37][38] and its expression is increased in bronchoalveolar lavages in patients with lung transplantation rejection. Hsieh et al. [39] reported that IL-17 could serve as a predictive parameter for borderline subclinical renal allograft rejection in the future. TNF alpha is an adipokine involved in systemic inflammation and is a member of a group of cytokines which stimulate the acute phase reaction. The ratio of IL10/TNFα expression in regulatory B cells may be a better indicator of regulatory function than IL10 expression alone for the prediction of immunologic human allograft injury [40]. Our results agree with previously published papers and also detected lower TNFα and IL17 levels associated with low RFS.
On the other hand, cytokine levels could be influenced by the particular immunosuppressant therapy. Patients on calcineurin inhibitors had impaired inflammatory cytokine production in response to TLR2, TLR4, and TLR7/8 stimulation compared with healthy controls [41]. In our work, the influence of immunosuppressive regimen (2 or 3 drugs vs 1 drug) on cytokine levels showed a tendency to decrease the level of IL4.
Gender donor/recipient mismatch was independently associated with the probability of graft rejection even when clinical and biological tolerance markers were considered. A higher risk of graft loss in male recipients after liver transplantation from female donors has been reported before by other authors [42]. Also male-to-female liver transplantations seem to present in some cases complications during surgery intervention with a negative impact in the patient outcome [43]. As Rustgy et al. remarked immunologic factors are influenced by gender; anyhow, additional studies are required to confirm this finding and to clearly uncover its causes. Although only gender donor/ recipient mismatch was associated with graft rejection in multivariate analysis, DC in peripheral blood/serum samples, KIR ligand mismatch and cytokine levels (TNFα, IL17, GMCSF, IL1β and IL6) should be incorporated into the evaluation of liver recipients in the screening of operational tolerance programs.
In conclusion, we have confirmed that high chimerism levels in peripheral blood during the first months after transplantation are related with a lower probability of rejection. A CD56 bright NK cell subset dominates the early post-transplant period following liver transplantation. A tendency to high frequency of graft rejection was observed in cases with KIR-ligand mismatches. Donors lacking C1ligands exhibited increased risk of acute rejection. Cytokine levels predict events in liver transplant recipients: low TNFα levels in pretransplant recipient samples were associated with liver disease relapse and low IL17 and TNFα levels and high GMCSF and IL1β levels are associated with low rejection free survival. Finally IL6 levels show an adverse impact on recipient overall survival.