SEARCH

SEARCH BY CITATION

Keywords:

  • Biliary atresia;
  • living-related liver transplantation;
  • microchimerism

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

The presence of maternal cells in offspring may promote tolerance to noninherited maternal antigens (NIMAs). Children with biliary atresia (BA) have increased maternal cells in their livers, which may impact tolerance. We hypothesized that patients with BA would have improved outcomes when receiving a maternal liver. We reviewed all pediatric liver transplants recorded in the SRTR database from 1996 to 2010 and compared BA and non-BA recipients of maternal livers with recipients of paternal livers for the incidences of graft failure and retransplantation. Rejection episodes after parental liver transplantation were examined for patients transplanted at our institution. BA patients receiving a maternal graft had lower rates of graft failure compared to those receiving a paternal graft (3.7% vs. 10.5%, p = 0.02) and, consequently, fewer episodes of retransplantation (2.7% vs. 7.5%, p = 0.04). These differences were not seen among non-BA patients or among BA patients who received female deceased donor grafts. In patients transplanted at our institution, paternal liver transplantation was associated with an increased incidence of refractory rejection compared to maternal liver transplantation only in BA. Our data support the concept that maternal cells in BA recipients promote tolerance to NIMAs and may be important in counseling BA patients who require liver transplantation.


Abbreviations: 
ALT

alanine aminotransferase

ACR

acute cellular rejection

BA

biliary atresia

DD

deceased donor

IMA

inherited maternal antigen

IPA

inherited paternal antigen

IQR

interquartile range

LR

living related

NIMA

noninherited maternal antigen

NIPA

noninherited paternal antigen

RR

refractory rejection

Tregs

regulatory T cells

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

Biliary atresia (BA) is the most common cause of neonatal liver failure requiring liver transplantation. This disease is characterized by inflammation of the biliary tree of unclear etiology (1) and is almost uniformly fatal if untreated (2). A Kasai portoenterostomy is performed after diagnosis but liver transplantation is required in patients in whom long-term biliary drainage is not achieved.

Several groups have demonstrated higher numbers of maternal cells in the livers of patients with BA and hypothesized that these cells may contribute to disease pathogenesis by initiating a graft-versus-host reaction in the liver (3–6). The function of trafficking maternal cells and the consequences of persistent maternal microchimerism remain fascinating unanswered questions in the field. Increased numbers of maternal cells have been found in other patients with autoimmune diseases (7–9) and it is not clear whether they contribute to disease or simply proliferate in response to injury.

However, maternal microchimerism may also be tolerogenic. A recent study suggested that maternal cells in fetuses promote the generation of fetal regulatory T cells (Tregs) that suppress the fetal immune response to noninherited maternal antigens (NIMA) (10). It is possible that such tolerance is long lived and that the ‘NIMA effect’ may also play a role in postnatal transplantation when the mother serves as the donor. Indeed, a beneficial effect has been reported for bone marrow transplantation (11,12), but the results for kidney transplantation have been mixed (13–15). In contrast to renal transplantation, liver transplantation is less immunogenic, as indicated by less chronic immunosuppression requirements and the ability to withdraw immunosuppression completely in some patients (16). We therefore reasoned that if the NIMA effect leads to a subtle improvement in solid organ transplantation, its effect might best be discerned in such a setting, especially in a disease with elevated levels of maternal microchimerism.

To test the hypothesis that the increased maternal microchimerism observed in patients with BA may lead to different outcomes when receiving a maternal versus a paternal liver allograft, we examined the national SRTR database as well as the clinical records of patients transplanted at our institution. We report a significant beneficial effect of maternal liver transplantation compared to paternal on graft failure and retransplantation specifically in patients with BA.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

Graft failure analysis

We reviewed all pediatric (age 0–6 years) primary isolated split liver transplants performed in the United States from 1996 to 2010 as recorded in the SRTR database. We restricted our analysis to recipients who were 0–6 years old since >80% of all recipients and >95% of BA recipients were within this age range at the time of transplantation (Figure S1). Recipient, donor and transplant characteristics were recorded among living related (LR; maternal or paternal) and split deceased donor (DD; female or male) transplants. We compared BA and non-BA recipients of maternal livers to recipients of paternal livers for graft failure at 6 months after transplantation. The 6-month time point was chosen since the majority of patients who developed early graft failure did so within 6 months after transplantation (Figure S2). Graft failure occurred if patients underwent retransplantation, if “primary graft failure” was listed as the cause of death, or if the graft was no longer functioning at the time of death. Patients who died from causes that were not attributed to graft failure were excluded from the analysis. Causes of graft failure are listed in Table S1. We also examined serum alanine aminotransferase (ALT) levels at discharge. As a comparison group, we analyzed BA and non-BA recipients of female and male split DD livers.

Analysis of rejection episodes

We retrospectively reviewed the clinical and pathologic records of all pediatric (age 0–6 years) primary liver transplants performed at our institution between 1993 and 2010 to identify biopsy-proven rejection episodes (17). All patients in this group received a prednisone taper along with either tacrolimus or cyclosporine (Neoral, Novartis, East Hanover, NJ, USA). The majority of patients (79%) received additional immunosuppression in the form of mycophenolate mofetil. While most cases of acute cellular rejection (ACR) resolve with a single steroid pulse, some patients are refractory to the initial therapy (refractory rejection, RR) (18) and require additional treatment. Patients were categorized as having ACR if they had a single episode of rejection which resolved with a course of steroids or as having RR if they required antibody therapy or an additional steroid bolus for a single rejection episode, if a subsequent biopsy demonstrated persistent rejection, or if they developed more than one rejection episode during the 6 month follow up period. This categorization was performed while blinded to the disease subtype and donor gender. Patients who did not have at least 6 months of follow-up or who lost their primary graft without biopsy-proven evidence of rejection were excluded from the analysis.

HLA matching analysis:  We compared BA and non-BA patients for compatibility with their parents at the HLA-A, HLA-B and HLA-DR loci. Complete data at each locus were available for >22% of patients. The following nomenclature was used: inherited maternal antigen (IMA) = the allele shared by the mother and the recipient; noninherited maternal antigen (NIMA) = the maternal allele not shared by the recipient; inherited paternal antigen (IPA): recipient allele not inherited from the mother; noninherited paternal antigen (NIPA) = the paternal allele not shared by the recipient. Compatibility with a maternal donor was then categorized as NIMA = IPA or NIMA/IPA = IMA. Compatibility with a paternal donor was defined as NIPA = IMA or NIPA/IMA = IPA. If a recipient/donor pair did not meet these criteria, they were categorized as incompatible.

Statistical analysis

We performed a univariate analysis to identify donor, recipient and transplant factors that were predictors of graft failure using a statistical significance level of 0.15. Variables found to be significant predictors of graft failure in the univariate analysis were entered into a forward stepwise logistic regression analysis, yielding adjusted odds ratios and 95% confidence intervals, using a statistical significance level of 0.05. Continuous and nonparametric variables were compared using the Wilcoxon rank-sum test while categorical variables were compared using the chi-square test. All statistical analyses were performed using STATA version 10 (Stata Corp. LP, College Station, TX, USA). Approval for this study was obtained from the UCSF Committee on Human Research (Approval # 10-01781).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

Graft failure after maternal or paternal liver transplantation:  We identified 611 patients who received LR parental liver allografts (363 maternal, 248 paternal) and 525 patients who received split DD allografts (199 female, 326 male). We separated patients with BA from those who received pediatric liver transplants for other reasons (‘non-BA’). We first compared BA (n = 321) and non-BA (n = 290) recipients of maternal and paternal livers for differences in recipient, donor and transplant factors (Table 1). There were no differences in recipient characteristics between patients who received a maternal liver and those who received a paternal liver for either disease. Maternal donors were slightly younger, shorter and weighed less than paternal donors in both BA and non-BA patients. Transplant characteristics were the same among recipients of maternal and paternal donors.

Table 1.  Demographics
 BA: living related donorsNon-BA: living related donorsBA: deceased donorsNon-BA: deceased donors
Maternal (n = 188)Paternal (n = 133)p-Value1Maternal (n = 175)Paternal(n = 115)p-Value1Female (n = 90)Male (n = 137)p-Value1Female (n = 109)Male (n = 189)p-Value1
  1. 1Nominal variables calculated by chi-square statistic, continuous and nonparametric with Wilcoxon rank sum.

  2. 2Median with range.

  3. 3Median with interquartile range (IQR).

Recipient characteristics
 Age, months  0.31  0.64  0.63  0.01
   Median29 (3–82)8 (3–60) 17 (0–80)21 (0–83) 11 (3–75)10 (4–83) 17 (0–82)26 (1–81) 
 Gender  0.57  0.24  0.94  0.76
   Male36.739.9 63.456.5 35.635.0 64.262.4 
 Race  0.35  0.77  0.31  0.38
   Caucasian56.464.7 60.664.4 46.740.9 50.546.0 
   African American13.814.3 12.614.8 24.418.3 13.816.4 
   Hispanic/Latino20.712.8 19.416.5 22.232.9 32.133.3 
   Asian5.96.0 5.12.6 5.67.3 1.84.2 
   Other3.22.2 2.31.7 1.10.7 1.80.0 
 Etiology     0.80     0.45
   Noncholestatic cirrhosis   5.16.1    2.86.9 
   Fulminant hepatic failure   23.419.1    39.535.5 
   Cholestatic liver disease   9.712.2    6.45.3 
   Metabolic disorder   14.316.5    17.415.3 
   Malignant neoplasm   13.116.5    9.214.8 
   Other/unknown   34.329.6    24.822.2 
 Medical condition  0.41  0.41  0.70  0.32
   ICU8.09.0 33.732.2 27.823.4 60.657.1 
   Hospitalized38.831.6 22.917.4 20.023.4 15.611.6 
   Not hospitalized53.259.4 43.350.4 52.253.3 23.931.2 
 Pretransplant ALT  0.50  0.47  0.88  0.71
   Median3103 (61–170)104 (66–181) 120 (56–356)124 (46–288) 107 (70–170)109 (70–177) 152 (70–633)135 (56–708) 
Donor characteristics
 Age, years  0.01  0.01  0.01  <0.001
   Median330 (18–45)32 (19–50) 29 (18–46)31 (18–49) 27 (18–54)24 (18–55) 34 (18–55)24 (18–53) 
 Race  0.70  0.73  0.01  0.13
   Caucasian59.065.4 63.461.7 80.056.9 71.662.4 
   African American12.214.3 12.614.8 6.715.3 6.416.4 
   Hispanic/Latino19.713.5 18.916.5 10.023.4 14.716.9 
   Asian6.95.3 4.64.4 3.31.5 4.62.7 
   Other2.11.5 0.62.6 0.02.9 2.81.6 
 Height, cm  <0.001  <0.001  <0.001  <0.001
   Median3163 (158–168)178 (170–183) 163 (157–168)178 (173–181) 164 (158–168)178 (170–183) 163 (157–168)178 (173–183) 
 Weight, kg  <0.001  <0.001  <0.001  <0.001
   Median364 (56–74)79 (68–87) 64 (57–71)80 (71–89) 60 (56–68)73 (64–81) 60 (55–68)75 (69–83) 
Transplant characteristics
 ABO type  0.14  0.26  0.66  0.56
   Identical85.679.7 81.783.5 80.078.8 74.377.3 
   Compatible12.219.6 17.713.9 18.918.3 22.918.5 
   Incompatible2.10.8 0.62.6 1.12.9 2.84.2 
 Warm ischemia time, min  0.78  0.93  0.19  0.02
   Median339 (30–50)40 (30–55) 40 (31–51)41 (32–48) 43 (35–58)47 (34–63) 44 (33–55)39 (30–47) 
 Cold ischemia time, hours 0.36  0.83  0.25  0.07 
   Median32.5 (1–4)2.0 (1–4) 2.0 (1–4)2.0 (1–4) 7.0 (6–10)7.0 (5–9) 8.0 (6–10)7.0 (5–9) 

In the control group of DD transplants, there were no significant differences in recipient characteristics in the BA group (n = 227). Among non-BA recipients (n = 298), male recipients were older than female recipients. Male deceased donors were younger, taller and weighed more, as has been previously reported (19), and there were differences in their ethnicity compared to female donors. Transplant characteristics showed a slightly increased warm ischemia time in non-BA patients receiving female deceased livers.

Based on our hypothesis that BA patients would have improved outcomes when receiving a graft from their mother, we first compared graft failure at 6 months after transplantation between BA patients who received a maternal or paternal liver. We observed a significantly lower incidence of graft failure in BA patients who received a maternal liver compared to those who received a paternal liver (Figure 1A; maternal 3.7%, paternal 10.5%, chi-square test 0.02). BA patients receiving a maternal liver also had a lower incidence of retransplantation (Figure 1B; maternal 2.7%, paternal 7.5%, chi-square test 0.04). Importantly, the finding of improved graft survival with maternal organs was specific to those recipients with BA: among non-BA patients, there were no differences in the rates of graft failure or retransplantation between recipients of maternal and paternal organs (Figure 1A, B). Median discharge ALT levels (an indirect measure of ongoing hepatic injury (18)) were significantly lower among BA patients receiving a maternal liver compared to those receiving a paternal liver (maternal 52.0, [IQR 31.0–89.0] vs. paternal 67.0, [IQR 44.0–108.0]; p = 0.007, by Wilcoxon rank-sum), although pretransplant ALT levels were similar in both groups.

image

Figure 1. Incidence of (A) graft failure and (B) retransplantation at 6 months among BA and non-BA recipients of LR (maternal and paternal) or DD (female and male) livers (LR BA: maternal n = 188, paternal n = 133; LR non-BA: maternal n = 175, paternal n = 115; DD BA: female n = 90, male n = 137; DD non-BA: female n = 109, male n = 189). The percentage of patients who experienced graft failure or retransplantation was compared between recipients of maternal/female and paternal/male livers using chi-square analysis. (C) Biopsy-confirmed rejection episodes among BA and non-BA recipients of maternal and paternal livers from a single institution (LR BA: maternal n = 21, paternal n = 12; LR non-BA: maternal n = 8, paternal n = 12). The percentage of patients who experienced refractory rejection (RR) was compared between recipients of maternal and paternal livers using chi-square analysis. *p-Value <0.05.

Download figure to PowerPoint

To determine if the salutary effect of maternal livers in BA was secondary to the transplantation of a female liver, we analyzed a control group of patients receiving split DD organs. We found no differences in the rates of graft failure or retransplantation when comparing male and female DD organs in BA or non-BA recipients (Figure 1A, B). These results indicate that the beneficial effect of maternal livers is limited to those patients whose underlying disease is BA and is due to the maternal origin of the organ and not simply to donor gender.

BA patients are younger than non-BA patients at the time of liver transplantation and it is possible that recipient age influences the levels of maternal microchimerism and tolerance. We therefore analyzed graft failure in the 0–1 age group and found that there is still decreased graft failure in BA patients receiving maternal livers compared to paternal livers, while no significant difference is observed in non-BA patients (0–1 years: BA, maternal 3.5%, paternal 11.1%, chi-square test 0.02; non-BA, maternal 7.9%, paternal 11.6%, chi-square test 0.52). To further rule out the possibility that younger patients fare better with maternal livers compared to older ones, we compared graft failure in BA patients who were 0–1 or 1–6 years old at the time of transplantation. We found similarly low rates of graft failure in both groups (<1 year 3.5%; 1–6 years 4.6%, chi-square test 0.74). Taken together, these results indicate that the benefit of maternal livers in BA is not simply because BA patients may be younger than non-BA patients at the time of transplantation.

We reasoned that the lack of a significant decrease in graft failure with maternal organs in non-BA recipients could be explained by one of the two following possibilities: BA recipients of paternal livers could have more graft failure compared to non-BA recipients of paternal livers or, BA recipients of maternal livers could have less graft failure compared to non-BA recipients of maternal livers. We observed near identical rates of graft failure in BA and non-BA recipients of paternal livers (BA 10.5% vs. non-BA 9.6%, chi-square test 0.80). In contrast, there was a lower rate of graft failure with maternal organs in BA patients compared to non-BA patients, although this difference was not statistically significant (graft failure: BA 3.7%, non-BA 6.3%, p = 0.26). These results suggest that the finding of less graft failure with maternal organs only in BA may be due to the differential tolerance to NIMAs between BA and non-BA recipients.

Analysis of rejection episodes:  Although the improvement in graft survival with maternal liver transplantation is striking, the causes of graft failure are not clearly immune-mediated and include etiologies such as vascular thrombosis (Table S1). To obtain detailed information on biopsy-proven rejection, a parameter not well-recorded in the SRTR database, we next examined the clinical and histological records for the subset of LR recipients transplanted at our own institution (n = 60) and categorized each patient as having either a single steroid responsive episode of ACR or RR within the first 6 months after transplantation. BA patients receiving paternal livers had more episodes of RR compared to those receiving a maternal liver (Figure 1C, maternal 11.1% vs. paternal 36.4%, chi-square test 0.10), while this trend was not seen in non-BA patients (Figure 1C, maternal 12.5% vs. paternal 18.2%, chi-square test 0.74). To address the possibility that BA recipients may uniquely react against paternal antigens, we compared the rate of RR between BA and non-BA recipients of paternal grafts. We found a trend towards a higher rate of RR among BA patients (36.4%) compared to non-BA patients (18.2%, p = 0.34), although this difference was not significant. Thus, analysis of rejection episodes in this smaller cohort supports our overall conclusion that there is a discernible difference in the ability to tolerate a maternal versus paternal graft only when the underlying disease is BA.

Risk factors for graft failure:  Since graft failure differed between BA patients receiving maternal versus paternal organs, we performed a univariate analysis for graft failure to determine if other factors contribute to this difference (Table 2A). Our analysis confirmed that paternal organs have a higher risk of graft failure specifically in BA patients (OR 3.04, p = 0.02) but not in non-BA patients (OR 1.58, p = 0.31). African American recipients with BA also had a higher risk of graft failure (OR 2.68, p = 0.06). ABO identical grafts had a significantly lower risk of graft failure (OR 0.19, p < 0.001) in BA. Surprisingly, BA patients who received ABO compatible grafts had a much higher risk of graft failure (OR 6.08, p < 0.001), whereas this risk was not seen if the graft was ABO incompatible (OR 1.25, p = 0.88). This increased risk was seen only if BA patients received LR organs and was not seen in ABO-compatible DD organs (data not shown) or among non-BA patients. Detailed analysis of the impact of ABO type on graft failure (Table S2) demonstrated that this result is secondary to a high risk of graft failure in BA patients with blood group AB (OR 6.98, p = 0.01). Factors such as recipient age and donor/recipient CMV mismatch did not contribute to graft failure in BA patients.

Table 2.  Univariate and multiple regression analysis of graft failure
(A) Univariate analysis of graft failure
 BA living related unadjusted ORNon-BA living related unadjusted OR
OR95% CIp-ValueOR95% CIp-Value
Donor predictors      
 Paternal donor3.04(1.19–7.76)0.021.58(0.66–3.77)0.31
 Age1.02(0.95–1.10)0.621.06(0.99–1.13)0.10
 Race
   Caucasian1.01(0.41–2.51)0.981.04(0.42–2.57)0.93
   African American1.62(0.52–5.08)0.411.02(0.29–3.61)0.98
   Hispanic/Latino0.79(0.23–2.80)0.720.44(0.10–1.93)0.27
   Asian0.32(0.02–5.44)0.434.07(1.03–16.06)0.05
   Other3.93(0.61–25.25)0.151.69(0.08–33.67)0.73
 Height (cm)1.03(0.97–1.10)0.371.03(0.97–1.09)0.30
 Weight (kg)1.01(0.97–1.06)0.541.01(0.97–1.06)0.55
Recipient predictors
 Male gender0.63(0.24–1.68)0.361.15(0.46–2.82)0.77
 Age1.31(0.90–1.91)0.161.13(0.91–1.41)0.27
 Race
   Caucasian1.10(0.44–2.73)0.840.87(0.36–2.12)0.77
   African American2.68(0.98–7.31)0.062.02(0.70–5.84)0.19
   Hispanic/Latino0.48(0.11–2.12)0.330.43(0.10–1.88)0.26
   Asian0.34(0.02–5.75)0.452.58(0.53–12.59)0.24
   Other0.71(0.04–12.68)0.820.90(0.05–16.45)0.94
 Medical condition
   ICU1.92(0.53–6.97)0.322.15(0.90–5.16)0.09
   Hospitalized0.89(0.35–2.27)0.810.36(0.08–1.59)0.18
   Not hospitalized0.86(0.36–2.10)0.750.79(0.33–1.92)0.61
Transplant predictors
 ABO compatibility
   Identical0.19(0.08, 0.47)<0.0010.70(0.25–2.01)0.51
   Compatible6.08(2.42–15.27)<0.0011.58(0.55–4.52)0.39
   Incompatible1.25(0.07–23.35)0.881.31(0.07–25.04)0.86
 Warm ischemia time1.01(1.00–1.03)0.171.01(0.99–1.03)0.37
 Cold ischemia time1.01(0.91–1.12)0.921.03(0.97–1.10)0.32
 Year of transplant0.96(0.85–1.08)0.530.88(0.77–1.00)0.05
 CMV IgG+ donor/IgG- receipt1.00(0.24–4.22)1.000.56(0.10–3.01)0.50
 Zero HLA-A mismatches0.56(0.06–4.94)0.600.68(0.07–6.26)0.74
 Zero HLA-B mismatches0.48(0.03–9.10)0.635.80(0.84–39.86)0.07
 Zero HLA-DR mismatches0.83(0.09–7.54)0.870.31(0.02–5.94)0.44
(B) Multiple regression analysis of graft failure
PredictorBA living related adjusted OR Non-BA living relate adjusted OR
OR95% CIp-Value OR95% CIp-Value
Paternal donor2.72(1.03–7.17)0.04Asian donor7.00(0.52–94.81)0.14
African American recipient3.18(1.09–9.34)0.04Zero HLA-B mismatches7.00(0.97–50.57)0.05
ABO compatible6.00(2.31–15.58)<0.001    

Interestingly, risk factors for graft failure among non-BA patients were different from those seen in BA patients: in non-BA patients, older recipient age (OR 1.06, p = 0.10) and year of transplantation (OR 0.88, p = 0.05) correlated with a higher risk of graft failure. There was also a significantly increased risk of graft failure with Asian donors in non-BA (OR 4.07, p = 0.05), although the distribution of diseases in this subgroup was not different from the group as a whole (data not shown).

We next performed a multiple regression analysis incorporating all significant variables found in the univariate analysis (Table 2B). We found that the risk of graft failure in BA patients was still significantly higher with paternal donors (adjusted OR 2.72, p = 0.04), African American recipients (adjusted OR 3.18, p = 0.04), and with ABO compatible organs (adjusted OR 6.00, p < 0.001).

Gender matching:  Several groups have examined whether gender matching can impact liver transplant outcomes (20–22). Male-to-female mismatch has been identified as a risk factor for chronic rejection in adult patients with primary biliary cirrhosis (20) but there has not been a recent disease-specific analysis of gender matching in pediatric liver transplantation. To determine whether gender matching impacted our results, we analyzed rates of graft failure and retransplantation among female and male recipients of maternal or paternal livers separately (Table 3). Among BA patients, we found that the highest rate of graft failure is for father-to-daughter transplants (12.5%), which is significantly higher than that seen in mother-to-daughter transplants (4.2%, chi-square test 0.03). However, there were no differences in the rates of graft failure or retransplantation among BA recipients in DD male-to-female and female-to-female transplants (Table 3), indicating this effect is likely specific to the presence of paternal antigens in the graft and not due to the presence of H-Y antigen. Furthermore, paternal livers also fared worse in sons with BA compared to maternal livers (graft failure rates: F->S 7.6% vs. M->S 2.9%, chi-square test 0.24), suggesting that the higher risk of graft failure with paternal organs is not simply due to gender mismatch. This analysis indicates that the difference in graft survival observed with maternal versus paternal organs is due to the specific donor/recipient relationship and that the H-Y effect does not contribute significantly to graft failure for either BA or non-BA patients.

Table 3.  Impact of gender matching on graft failure
 BA: living relatedNon-BA: living relatedBA: deceasedNon-BA: deceased
Maternal donorPaternal donorp-ValueMaternal donorPaternal donorp-ValueFemale donorMale donorp-ValueFemale donorMale donorp-Value
Graft failure (%)
 Female recipient4.212.50.036.38.00.7215.515.70.9712.89.90.63
 Male recipient2.97.60.246.310.80.2921.922.90.9118.615.30.55
 p-Value0.650.36 0.990.62 0.450.30 0.440.29 
Retransplantation (%)
 Female recipient3.410.00.054.72.00.7212.112.40.9610.35.60.37
 Male recipient1.53.80.414.510.80.1118.814.60.6212.913.60.89
 p-Value0.430.18 0.960.07 0.390.71 0.690.09 

HLA compatibility:  It has been reported that BA patients share a high degree of HLA compatibility with their mothers and that this phenomenon may contribute to the higher levels of maternal microchimerism in this disease (23). To determine if increased HLA compatibility between BA patients and their mothers could account for the improved outcomes seen with maternal liver transplantation, we compared BA and non-BA recipients and their parents for compatibility at the HLA A, B and DR loci. We found no differences in HLA compatibility with the mother in BA patients compared to non-BA patients, indicating that the improvements in graft survival were not secondary to a unique situation of increased HLA compatibility in the setting of BA (Table 4A). We also compared compatibility with the mother versus the father in BA patients and found no increase in incompatibility with paternal donors compared to maternal donors, indicating if there is specific rejection of paternal antigens in BA, minor antigens are more likely to be involved (Table 4B).

Table 4.  Parental HLA and maternal versus paternal HLA compatibility in BA versus non-BA (%)
(A) Parental HLA compatibility in BA versus non-BA (%)
 Maternal donorsPaternal donors
BA recipientsNon-BA recipientsp-ValueBA recipientsNon-BA recipientsp-Value
HLA-A compatibility
 Compatible23.828.90.5916.711.10.53
 Incompatible76.271.10.5983.388.90.53
HLA-B compatibility
 Compatible10.610.00.92 7.911.10.66
 Incompatible89.490.00.9292.188.90.66
HLA-DR compatibility
 Compatible20.017.40.7610.018.50.36
 Incompatible80.082.60.7690.081.50.36
(B) Maternal versus paternal HLA compatibility in BA and non-BA (%)
 BA recipientsNon-BA recipients
Maternal donorsPaternal donorsp-ValueMaternal donorsPaternal donorsp-Value
HLA-A compatibility
 Compatible23.816.70.4428.911.10.08
 Incompatible76.283.30.4471.188.90.08
HLA-B compatibility
 Compatible10.6 7.90.6710.011.10.88
 Incompatible89.492.10.6790.088.90.88
HLA-DR compatibility
 Compatible20.010.00.2617.418.50.90
 Incompatible80.090.00.2682.681.50.90

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

This is the first study to examine whether the NIMA effect has an impact on the success of LR liver transplantation in biliary atresia, a disease with increased levels of maternal microchimerism. Consistent with our hypothesis, we report that BA patients who receive a maternal liver have a significantly lower rate of graft failure and a lower incidence of refractory rejection compared to those receiving paternal livers. The beneficial effect of maternal livers is seen only when BA is the underlying disease and is not seen in deceased donor transplants from female deceased donors. These results indicate that BA patients have differential tolerance to maternal versus paternal antigens and may have implications for patient counseling.

The idea that maternal-fetal tolerance may have long-term effects on LR organ transplantation has vital clinical significance, but studies in other transplant settings have obtained mixed results. In bone marrow transplantation, one report has demonstrated improved patient survival after haploidentical stem cell transplantation with maternal donors (11). However, in kidney transplantation, two studies have reported no improvement in the survival of mother-to-child transplants (14,15), while one study of sibling recipients of kidney transplants did report higher survival when the donor kidney expressed NIMAs (13). One potential explanation for the lack of a beneficial “NIMA effect” among recipients of maternal kidneys is that the immunogenicity of a renal allograft may mask any potential contribution of maternal-fetal tolerance mechanisms.

An alternative explanation of our data is that instead of maternal organs faring better, paternal organs fare worse, either due to the presence of the H-Y antigen or due to specific rejection of paternal antigens in donor livers in the setting of BA. Since our analysis of non-BA patients and of patients receiving DD transplants did not indicate a gender effect, our findings are more likely to be secondary to the latter hypothesis. It will be interesting to determine whether paternal antigens may be specifically more immunogenic in patients with BA. For example, higher levels of maternal cells in these patients may lead to rejection of paternal antigens, especially if they have been sensitized from prior pregnancies. This theory is consistent with the idea that maternal cells contribute to pathogenesis of BA due to an immune response (5).

While it has been reported that BA patients and their mothers share a high degree of HLA compatibility (23), which could explain heightened levels of maternal microchimerism, we found that BA patients had a similar degree of HLA compatibility with their parents as non-BA patients. This discrepancy may be explained by the less diverse HLA expression among Japanese patients, which were examined in the previous study (23), compared to the United States (24). The differences we observed with graft failure in ABO compatible and incompatible grafts for BA were surprising and unlike what has been reported for chronic rejection (20) or graft survival (25). We think that our results reflect our separate analysis of BA and non-BA patients and are due to the observation that the recipient blood group AB has a worse outcome overall in BA patients.

Although we demonstrate a clear advantage for maternal liver transplantation in BA patients, we did not directly measure maternal microchimerism and therefore cannot conclude that it leads to improved tolerance. In addition, the causes of graft failure include entities that are not obviously immunologic, such as hepatic artery thrombosis, a limitation we addressed by reviewing biopsy records from our institution. We are currently designing a prospective analysis of maternal microchimerism, patient/parental alloreactivity and transplant outcomes to directly measure the influence on maternal microchimerism on transplant tolerance.

Beyond our original hypothesis, our analysis of various outcome measures in both LR and DD pediatric liver transplants have allowed us to draw other important conclusions. Among BA recipients, we found that recipient age and health status are not significant prognostic indicators, as has been reported previously (26,27), indicating it is possible to carefully monitor patients who drain intermittently after a Kasai procedure prior to making a decision regarding transplantation. Most importantly, we show that the risk factors for graft failure in non-BA patients are different from those in BA, suggesting these two categories of patients should be considered separately in analyses of transplant outcomes.

Biliary atresia is the most common indication for liver transplantation in children and the lack of availability of deceased donor organs has led to a growing interest in LR transplantation. Our results may have important implications when counseling parents of BA patients for LR organ donation.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

We thank Drs. Ryutaro Hirose, John Roberts and Peter Stock for helpful discussions, Drs. Mike McCune and Alan Flake for their critical review of the manuscript, and Ms. Elizabeth Gress for her assistance with manuscript preparation. This work was supported by funds from the Irene Perstein Award (TCM) and was made possible by a grant from the California Institute for Regenerative Medicine, grant number TG2-01153 (AN). The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of CIRM or any other agency of the State of California.

Disclosure

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

Table S1: Causes of graft failure

Table S2: Impact of ABO type on graft failure

Figure S1: Distribution of living-related and split deceased donor liver transplants in the SRTR database based on recipient age.

Figure S2: Time to graft failure after transplantation.

FilenameFormatSizeDescription
AJT_3895_sm_suppmat.doc89KSupporting info item
AJT_3895_sm_FigS1.pdf13KSupporting info item
AJT_3895_sm_FigS2.pdf294KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.