If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
There has been a recent resurgence of interest in the importance of donor-specific human leukocyte antigen antibodies (DSAs) after orthotopic liver transplantation (OLT). In 1999, Kasahara et al.1 showed that all 12 patients with DSAs detected by flow cytometry crossmatching experienced acute rejection during the first month after OLT, whereas only 17% of patients without DSAs did. Similarly, Kozlowski et al.2 recently found that in patients with preformed DSAs, 75% of patients with persistent DSAs after OLT developed acute rejection and had positive complement component 4d (C4d) staining, whereas 0% of patients who cleared their DSAs after OLT did. Fontana et al.3 demonstrated that among 95 consecutive OLT patients, 24% developed human leukocyte antigen (HLA) antibodies, and 4% developed DSAs. Among the patients with DSAs, 75% had biliary complications, and 66% had biopsy-proven chronic rejection. More recently, Musat et al.4 found that 54% of patients with biopsy-proven acute cellular rejection (ACR) had diffuse C4d staining and circulating DSAs, and this suggested a concomitant humoral component. Additionally, our group has demonstrated that almost all patients with biopsy-proven chronic rejection have circulating DSAs.5
Immunoglobulins are mediators of the immune response and are usually produced after the recognition of non-self molecules. In renal transplantation, antibodies of immunoglobulin G (IgG) isotypes directed at recipient-mismatched donor HLAs have been associated with organ dysfunction and graft failure.6, 7 The 4 IgG subclasses have different characteristics related to their immunological effects; for example, IgG3 is typically the strongest complement fixing subclass (followed by IgG1), whereas IgG2 weakly fixes complement, and IgG4 is believed not to fix complement. Similarly, IgG3 and IgG1 are the subclasses with the strongest Fc receptor–binding properties, whereas IgG2 and IgG4 bind weakly. The combination of these 2 effects is thought to result in the pathogenicity of each individual IgG subclass. However, debate exists about the importance of IgG subclasses in transplantation. Studies investigating the importance of IgG subclasses in pretransplant sera of kidney transplant recipients have shown conflicting results. One group observed that IgG subclasses were heterogeneously detected in patients with early graft loss and acute rejection, and although IgG1 was the most commonly found subclass, they could not identify a single IgG subclass associated with worse graft outcomes.8 However, another group demonstrated that the presence of IgG1 alone was associated with the worst graft survival.9 A recent study using more modern technology found that the determination of IgG subclasses in pretransplant serum did not add power for predicting rejection.10 However, when posttransplant sera of kidney transplant recipients were studied, the data consistently showed that the presence of IgG3 carried an increased risk for graft loss.11-13 Because of these findings, we analyzed IgG subclasses after transplantation.
Previously, we found that almost all patients with chronic rejection had DSAs.5 Much to our surprise, many patients with normal graft function and no history of rejection also had DSAs. Although the mean fluorescence intensity (MFI) played a role, we hypothesized that there was an intrinsic difference—specifically a difference in IgG subclasses–between the antibodies in patients with chronic rejection and the antibodies in patients with normal graft function. We used a modification of the Luminex-based assay to detect HLA antibodies of specific IgG subclasses.
ACR, acute cellular rejection; C4d, complement component 4d; CI, confidence interval; DSA, donor-specific human leukocyte antigen antibody; HLA, human leukocyte antigen; HR, hazard ratio; IgG, immunoglobulin G; IQR, interquartile range; MELD, Model for End-Stage Liver Disease; MFI, mean fluorescence intensity; OLT, orthotopic liver transplantation; SRR, steroid-resistant rejection.
PATIENTS AND METHODS
Patients and samples were selected from the biorepository and the liver transplant clinical research database of the Annette C. and Harold C. Simmons Transplant Institute. This biorepository was initiated in 1985 and collects serum and lymphocyte samples from all donors as well as protocol serial serum and lymphocyte samples from all recipients of OLT preoperatively and 1, 2, 5, 10, 15, and 20 years after transplantation. Selected patients had event-driven samples collected. The patients consented to their participation in this prospective collection at the time of OLT with the understanding that the samples would be used for research purposes. In all, 368 serum samples from 39 patients with biopsy-proven chronic rejection and 66 comparator patients were tested. Chronic rejection patients had a median of 3 samples [interquartile range (IQR) = 1-6] collected during the study period; in this, they were similar to comparator patients, who had a median of 4 samples (IQR = 3-4) collected. The median number of allograft biopsies during the study was 11 (IQR = 8-13) for chronic rejection patients and 5 (IQR = 3-6) for comparator patients (our program performs protocol liver biopsy). All but 2 patients received a liver transplant for the first time, and all liver allografts came from ABO-compatible donors. Institutional review board approval was granted before the initiation of this evaluation.
Chronic rejection patients were chosen for analysis if they were >18 years old, had undergone OLT without another organ transplant, had biopsy-proven (or in 2 cases explant-proven) chronic ductopenic rejection (defined as >50% ductopenia14 in the absence of recurrent primary sclerosing cholangitis with or without foamy arteriopathy) diagnosed by a hepatopathologist, did not have hepatitis C viremia or primary biliary cirrhosis, and had never experienced a vascular problem after OLT. Patients with primary biliary cirrhosis were excluded for fear that recurrent disease might have been mistaken for chronic rejection because both can cause ductopenia. One patient had recurrent autoimmune hepatitis and chronic rejection. Biopsy-proven chronic rejection did not always occur at the same time that serum samples were available for the analysis of DSAs. The median time between transplantation and the diagnosis of chronic rejection was 382 days (IQR = 166-1053 days). The median time between the diagnosis of chronic rejection and an antibody test was 31 days (IQR = 3-87 days).
All patients with abnormal liver function tests underwent a Doppler sonogram evaluation of the hepatic artery. If any abnormality was found on the sonogram, the patients underwent angiography. All yearly follow-up patients with normal or abnormal liver function tests also underwent Doppler sonogram screening. As part of our pre-OLT evaluation, the patients underwent a rigorous evaluation for causes of liver disease, including viral, genetic, and autoimmune causes. All explants with patients' clinical information were reviewed by the division director of hepatology, who certified the etiology of liver disease; this was then locked in the database.
For each chronic rejection patient, 1 comparator patient with no history of rejection (acute or chronic) was chosen, and they were matched by sex, the year of transplantation (within 5 years), age (within 5 years), and the calcineurin inhibitor used at 1 month (as previously published). However, before this, we analyzed 27 other comparator patients. Because this analysis focused on differences in IgG subclasses between chronic rejection patients and comparator patients, we wanted to maximize our ability to understand DSA-positive patients with normal graft function and no rejection. Therefore, all 66 comparator patients were used for this analysis. The comparator patients were not allowed to have had acute rejection in the past because others have reported a risk of rejection in the presence of DSAs.1, 4 The term comparator was used instead of control because significant unintended differences were found between these 2 groups.
HLA Tissue Typing
All patients and donors were typed for HLA-A, HLA-B, HLA-DRB1, HLA-DRB345, and HLA-DQ with commercially available serological typing trays (Terasaki HLA tissue typing trays) or molecular methods (Micro SSP or LabType SSO, One Lambda, Inc., Canoga Park, CA). Before 1998, serology was used exclusively. Since 1998, all donor class I and II HLA typing and patient class II typing have been performed with molecular methods, whereas patient class I typing has been performed with serology.
HLA IgG Antibody Determination
All sera were blinded and sent to the Terasaki Foundation Laboratory for evaluation. The detection of anti-HLA IgG antibodies was performed with LABScreen single-antigen class I (lot 6) and II (lot 8) beads (One Lambda). The assay was performed according to the manufacturer's protocol. We adopted an MFI cutoff of 1000 for the class I and II antibody interpretation because it has been used by many investigators of different organ transplants with Luminex-based assays.
HLA IgG Subclass Determination
All sera positive for DSAs were further evaluated for IgG subclasses. We used a modified assay in which the conventional phycoerythrin-conjugated, anti-human IgG was replaced with a phycoerythrin-conjugated, IgG subclass–specific, anti-human IgG (IgG1 clone 4E3, IgG2 clone HP6002, IgG3 clone HP6050, and IgG4 clone HP6025, Southern Biotech, Birmingham, AL). The specificity of the anti-IgG subclass secondary antibodies was verified through the testing of these antibodies against a panel of 4 validation beads made in house (each coated with a single human IgG subclass protein from EMD Chemicals, Gibbstown, NJ) and the subsequent comparison of the corresponding MFIs. Anti-IgG2, anti-IgG3, and anti-IgG4 antibodies showed very specific patterns of binding to the corresponding IgG-coated beads with <5% cross-reactivity against other IgG-coated validation beads. Anti-IgG1 antibodies showed minimal cross-reactivity (12%) with IgG2-coated beads. The IgG subclass assay was performed with the same lot of LABScreen single-antigen beads and with the same protocol used for the detection of anti-HLA IgG antibodies. A normalized trimmed MFI higher than 500 was defined as positive on the basis of binding patterns after validation and dilution experiments.
Patient characteristics were compared with χ2 and Wilcoxon rank-sum tests, and the median values are reported. Categorical variables were summarized and compared with the χ2 test and Fisher's exact test. A Cox stepwise multivariate proportional hazards analysis for predicting patient and graft survival included all variables with a P value < 0.2 in the univariate analysis. The IgG3 subclass and multiple IgG subclasses were both forced into the final model to determine their individual predictive powers. These variables were built as time-dependent covariates, so patients were at risk for graft loss because of them only after their first positive test. Graft survival was examined with Kaplan-Meier analysis, and the log-rank test was used to detect differences. Statistical significance was defined as a P value < 0.05. All statistical analyses were performed with Stata/MP 10.1 (StataCorp LP, College Station, TX).
The characteristics of the 2 groups are summarized in Table 1. Patients in the chronic rejection group were younger and included fewer Hispanics and more African Americans than the comparator group. The donor ages, the Model for End-Stage Liver Disease (MELD) scores at the time of OLT, and the cold ischemia times were similar for the 2 groups. The median year for OLT was 1991 for the chronic rejection patients and 1997 for the comparator patients. The median number of HLA-A, HLA-B, HLA-DRB1, HLA-DRB345, and HLA-DQ mismatches was 7 for both groups. The tissue typing information was not available at the time of transplantation, so the allocation and selection of donor-recipient pairs were not based on this information. The use of calcineurin inhibitors, steroids, and antimetabolites at 1 month was similar in the 2 groups. Cytomegalovirus infections were more frequently observed in chronic rejection patients. More patients in the chronic rejection group underwent transplantation for primary sclerosing cholangitis or acute liver failure, whereas patients in the comparator group more often underwent transplantation for alcoholic liver disease or cryptogenic cirrhosis. The chronic rejection patients had a 5-year graft survival rate of 23%, whereas the comparator patients had a survival rate of 100%. All patients in the comparator group were alive 5 years after transplantation, whereas only 49% of the chronic rejection patients were. Most patients in the chronic rejection group had at least 1 episode of acute rejection, and 44% had at least 1 episode of steroid-resistant rejection (SRR). Comparator patients were chosen because they survived more than 5 years without evidence of rejection.
Table 1. Patient Demographics
Chronic Rejection Group (n = 39)
Comparator Group (n = 66)
NOTE: Bolded values are significant.
The data are presented as medians.
Wilcoxon rank-sum test.
Patients were included only if this occurred before OLT.
Thirty-six of the 39 patients (92%) with biopsy-proven chronic rejection had detectable DSAs, whereas only 37 of the 66 patients (56%) who never experienced any type of rejection had detectable DSAs (P < 0.001). A summary of the frequencies and characteristics of the DSAs in each group is shown in Table 2. Twenty-one of the 35 patients (60%) in the chronic rejection group with available pre-OLT serum samples had preformed DSAs, whereas 27 of the 66 patients (41%) in the comparator group did (P = 0.07). Twenty-one of the 34 patients (62%) in the chronic rejection group with available post-OLT serum samples had de novo DSAs, whereas only 19 of the 66 patients (29%) in the comparator group did (P = 0.001). Twenty-seven of the 34 patients (79%) with chronic rejection had DSAs after OLT (preformed or de novo), whereas 32 of the 66 patients (48%) in the comparator group did (P = 0.003). DSAs against class II HLA were more commonly observed overall. In the chronic rejection group, 10% had DSAs exclusively against class I, 39% had DSAs exclusively against class II, and 44% had DSAs against classes I and II. In the comparator group, 8% had DSAs against class I, 27% had DSAs against class II, and 21% had DSAs against both classes.
Table 2. HLA Antibody Characteristics
Chronic Rejection Group (n = 39)
Comparator Group (n = 66)
NOTE: Bolded values are significant. - For preformed DSAs, chronic rejection (n = 35). - For de novo and post-OLT DSAs, chronic rejection (n = 34). - Post-OLT IgG subclass DSAs and DSA profile, chronic rejection (n = 32).
IgG DSAs (%)
Any DSAs ever
De novo DSAs
DSA class I only
DSA class II only
DSA classes I and II
Post-OLT IgG subclass DSAs (%)
Post-OLT IgG subclass DSA profile (%)
Single IgG subclass
Multiple IgG subclasses
When we analyzed the presence or absence of DSAs, we found that the median MFI for chronic rejection patients who were DSA-positive after OLT (79%) was 10,171 (IQR = 2987-22,804). The median MFI for comparator patients who were DSA-positive after OLT (48%) was 4250 (IQR = 2020-9152).
IgG Subclass DSAs
For the evaluation of IgG subclasses, we analyzed patients with DSAs after OLT (27 chronic rejection patients and 32 comparator patients). All DSA-positive sera were examined with the modified single-antigen beads assay for IgG subclasses. Two patients in the chronic rejection group did not have enough serum and were excluded from the analysis. DSAs of the IgG1 subclass were most common and were found in 45% of all patients, followed by IgG3 DSAs (21%), IgG4 DSAs (14%), and IgG2 DSAs (13%). In the chronic rejection group, 23 of the 32 patients had IgG1 DSAs (72%), whereas 21 of the 66 comparator patients (32%) did (P < 0.001). Six of the 32 chronic rejection patients (19%) and 7 of the 66 comparator patients (11%) had the IgG2 subclass (P = 0.27). Twelve of the 32 chronic rejection patients (38%) and 9 of the 66 comparator patients (14%) had the IgG3 subclass (P = 0.007). Eight of the 32 chronic rejection patients (25%) and 6 of the 66 comparator patients (9%) had the IgG4 subclass (P = 0.04). One chronic rejection patient and 7 comparator patients had no IgG subclass despite the presence of DSAs. Figure 1 shows the timing and pattern of IgG subclasses among DSA-positive patients in the chronic rejection group (Fig. 1A) and the comparator group (Fig. 1B) during the follow-up period.
The median IgG subclass MFI in chronic rejection patients with post-OLT DSA (71%) was 5596 (IQR = 898-18,453); the median IgG subclass MFI in comparator patients with post-OLT DSA (38%) was 4720 (IQR = 1734-12,405). For 5 chronic rejection patients and 1 comparator patient, the IgG subclass MFI was 500 to 1000.
IgG Subclass DSA Profile
After we observed that patients displayed different IgG subclass DSA profiles, we regrouped the patients according to whether they had a single IgG subclass or a combination of multiple IgG subclasses. Similar proportions of patients with a single IgG subclass were observed in the 2 groups: 8 of the 32 chronic rejection patients (25%) and 16 of the 66 comparator patients (24%, P = 0.94). Conversely, 16 of the 32 patients with chronic rejection showed a combination of multiple IgG subclasses, whereas 9 of the 66 comparator patients did (50% versus 14%, P < 0.001). The use of different immunosuppressive agents was studied to investigate any association with the production of certain IgG subclasses or profiles. Only steroid use 1 month after OLT was statistically significantly associated with a lower likelihood of developing multiple IgG subclasses; 83% of the patients not receiving steroids had multiple IgG subclasses, whereas only 21% of the patients receiving steroids did [P = 0.005 (not shown)].
Risk Factors Associated With Graft and Patient Survival
In the univariate analysis, we found 12 variables associated with graft survival with a P value < 0.2, and we entered these variables into a Cox stepwise multivariate regression analysis (Table 3). The multivariate analysis showed that the development of IgG3 DSAs, the presence of ACR, and the use of cyclosporine versus tacrolimus at 1 month increased the likelihood of graft loss. Specifically, patients with IgG3 DSAs had a hazard ratio (HR) for graft loss of 3.35 [95% confidence interval (CI) = 1.39-8.05, P = 0.007]. Patients with a history of ACR at any time after OLT had an HR for graft loss of 10.4 (95% CI = 5.42-19.8, P < 0.001). Finally, patients who received cyclosporine instead of tacrolimus 1 month after OLT had an HR for graft loss of 2.32 (95% CI = 1.07-5.03, P = 0.03).
Table 3. Cox Regression Analysis for Predicting Graft Survival
HR (95% CI)
HR (95% CI)
All bolded values are not necessarily significant. Bolded values in the univariate analysis (left column) indicate p-values less than 0.2, which were introduced into the multivariate model. In the multivariate model (right column), the bolded p-values DO indicate that are significant. NOTE: Bolded values are significant.
Included in the Cox stepwise multivariate regression model.
Mycophenolate mofetil versus azathioprine or nothing at 1 month
Similarly, 13 variables in the univariate analysis were associated with patient survival and were, therefore, included in the multivariate analysis (Supporting Table 1). Similarly to the analysis of graft survival, the development of IgG3 DSAs and the presence of ACR were independent predictors of patient death. Patients with IgG3 DSAs had an HR for death of 2.88 (95% CI = 1.08-7.76, P = 0.04), and patients with a history of ACR had an HR for death of 6.42 (95% CI = 3.52-11.7, P < 0.001).
We also evaluated the effects of particular IgG subclasses and profiles on graft survival. Patients with a single IgG subclass had a significantly higher graft survival rate than patients with a combination of IgG subclasses (Supporting Fig. 1A). Moreover, a combination of IgG subclasses that included IgG3 led to the worst graft survival. The 1-year graft survival rate was 50% for patients with a combination of IgG subclasses that included IgG3, 71% for patients with a combination of IgG subclasses without IgG3, and 79% for patients with a single IgG subclass (Supporting Fig. 1B).
DSAs have been investigated in OLT recipients for many decades. It was originally believed that they did not play a role in liver rejection or allograft loss.15, 16 However, in the early days of OLT, graft loss rates were high, and the technology for detecting antibodies was limited to the crossmatch. After allograft survival rates increased and technology improved, new investigations were warranted. In fact, later studies showed an association between DSAs and graft loss.17-20 Although this correlation exists, many patients with DSAs have normal graft function. Therefore, there must be characteristics that distinguish pathological DSAs from those that are intrinsically less pathological or nonpathological. With the goal of identifying these characteristics, we evaluated the IgG subclasses of DSAs in patients with chronic rejection and compared them to the IgG subclasses of patients with normal graft function. Our rationale for focusing on patients with DSAs after transplantation was derived from previous publications showing that the determination of pretransplant IgG subclasses did not improve the predictive power of renal transplant rejection in comparison with a general DSA determination.8, 10
IgG1 was the most predominant subclass found, and this is consistent with previous studies in renal transplant recipients.1, 8-10, 12, 13, 21 However, we discovered that half of the nonrejection patients with DSAs had IgG1 alone, whereas patients with chronic rejection usually had a combination of IgG subclasses. It is unclear whether IgG1 alone or the production of a single IgG subclass has less pathologic potential or, more likely, the presence of multiple IgG subclasses serves as a surrogate marker for the immunogenicity of the antigen or the strength of the host immune response; a stronger immune response that incorporates the production of multiple IgG subclasses may lead to or be part of accelerated graft destruction.
We then separately analyzed the individual IgG subclasses because others have previously described an association between IgG3 and rejection in renal transplant recipients. Gao et al.11 studied 19 recipients of different organs and found 3 patients who experienced renal allograft rejection. DSAs of the IgG3 subclass were exclusively observed in the 3 cases of rejection, which included a case of hyperacute rejection. Taniguchi et al.12 investigated the role of IgG3 DSAs in 60 renal transplant recipients and observed that IgG3 was strongly associated with both allograft dysfunction and graft loss in patients with isolated class II DSAs. Similarly, Rerolle et al.13 found that IgG3 was statistically significantly more common in patients with graft loss versus patients with a functioning graft. In contrast, other IgG subclasses were found to have similar frequencies in the 2 groups. They also observed a trend toward lower graft survival in patients with IgG3 versus patients without IgG3. In our study, the Cox regression analysis found that the development of IgG3 DSAs carries decreased chances of patient and graft survival. According to these findings, complement-dependent or Fc receptor–dependent mechanisms of allograft injury (which are mechanisms associated with IgG3) may be important in the pathophysiology of chronic rejection in liver transplants, and this warrants further study.22
Because chronic rejection is relatively uncommon in the modern era of immunosuppression, our cohort of patients underwent transplantation many years ago. This greatly limited our access to optimal control patients and did not allow matching for every desirable characteristic. As a result, there were differences between our chronic rejection patients and control patients that led us to name the latter patients comparators. The chronic rejection patients more frequently received cyclosporine than the comparator patients, although this was not statistically significant. It has previously been shown that tacrolimus decreases the risk of both acute and chronic rejection; however, calcineurin inhibitor effects on DSAs after liver transplantation are unknown.23, 24 Another limitation is the lack of C4d staining of liver allograft biopsy samples. However, in our experience, C4d staining of formalin-fixed tissue has suboptimal sensitivity and specificity, and although this is improved in fresh tissue, we did not have access to fresh tissue. Finally, a potential technical limitation related to the use of biological samples was identified. It is possible that sera can be affected by both repeated thawing and freezing cycles and the storage time. Fortunately, all the serum samples used in this study were thawed for a second time, and even though some sera were kept in the freezer for more than 15 years (with potential alterations to the concentrations of certain proteins25), we found that sera from different time points displayed similar rates of positive DSAs (49% of samples before 1995 and 39% of samples after 1996); this leads us to believe that our results were not affected by the mechanism or time of storage.
In conclusion, DSAs are associated with chronic rejection after OLT. However, some patients have DSAs and do not develop chronic rejection. Many factors likely play a role in determining the difference. Although not all the factors are understood, patients with normal graft function in the presence of DSAs often have isolated IgG1, patients with chronic rejection often have a combination of IgG subclasses, and patients at the greatest risk of graft loss often have IgG3 DSAs.
The authors thank Tho Pham and Vadim Jucaud (Terasaki Foundation Laboratory) for their technical assistance with the antibody assays and the staff of the Baylor Transplantation Immunology Laboratory for maintenance of the biorepository.