- Top of page
The detrimental effect of alloantibodies directed against donor HLA has been widely demonstrated and accepted in all solid organ transplantation except the liver [1-4]. Studies have shown that preformed donor-specific HLA antibodies (DSA) increase the risk of acute rejection and lower graft survival [5-7], leading this to be commonly considered a contraindication to transplant. In addition, many studies have shown that de novo DSA after transplantation is associated with higher rates of rejection and lower rates of survival [3, 4, 8, 9].
In sharp contrast, livers are transplanted regardless of the crossmatch result. This is because posttransplant outcomes were thought to be similar regardless of the pre-transplant DSA status [10-12]. These perceived clinically equivalent outcomes in crossmatch positive and negative patients were supported by experiments in rats from the 1980s where DSA was thought to be an integral part of tolerance development [13, 14].
Despite the early clinical evidence showing no difference in patient or graft survival, later studies demonstrated that patients transplanted with a positive crossmatch had an increased risk of early graft loss [15, 16]. This has been further supported by later evidence in rats that hyperacute rejection can be induced by adoptively transferred DSA , and the group previously claiming DSA was an integral part of tolerance [13, 14] later found another protein (other than DSA) in serum to explain their tolerance induction . However, since consistent results are lacking, practice has not changed .
Although pre-transplant DSA's impact on outcome has been evaluated numerous times, the importance of DSA after liver transplantation has more limited data. Only in recent years have studies suggested that post-liver transplant DSA may play a role in acute and chronic rejection [20-25]. One of the first studies to link the role of alloantibodies and chronic rejection was published by Demetris et al.  where they showed 7 of 22 patients with chronic rejection had high PRA (panel reactive antibodies) peri-transplant and the majority showed deposition of Ig or complement components in the rejected liver tissue. Later, Piazza et al.  showed that 65% of liver transplant recipients had DSA after transplant, and this was associated with rejection. Kasahara et al.  demonstrated that if patients had DSA within the first month after transplant, 100% experienced rejection, compared to only 17% if no DSA was found. Similarly, Kozlowski et al.  found that preformed DSA that persists after transplant was associated with severe early rejection. Recently, Musat et al.  demonstrated that DSA is present in up to 75% of patients experiencing rejection, and both DSA and C4d staining was present in 54% of patients diagnosed with rejection, showing a previously unrecognized humoral component to these rejections. Likewise, 70% of patients with ductopenia had DSA and 60% had both DSA and C4d staining in their biopsy. We similarly showed an association between DSA in serum and biopsy-proven chronic rejection . Despite these reports on the association between DSA in serum and rejection, to date no group has investigated the frequency of de novo DSA formation after liver transplantation in a large cohort. Therefore, we sought to define the frequency and consequences of de novo DSA formation after liver transplantation.
- Top of page
With the improvement in patient and allograft survival that has occurred over the last several decades in liver transplantation, the re-evaluation of DSA's importance in liver transplantation is essential and has demonstrated a higher risk of rejection and decreased allograft survival in most, but not all studies [12, 19, 26, 27]. However, most prior studies focused on the impact of preformed DSA, treated class I and II DSA equally, and used a low MFI cutoff adopted from renal transplantation. As a result, the impact of DSA on patient and liver allograft survival may have been under appreciated. Specifically, one area not investigated is the impact of de novo DSA formation on patient and liver allograft survival. We found that 8.1% of liver transplant recipients develop de novo DSA with a MFI >5000. These findings are similar to the only other report using modern technology to screen for DSA after liver transplantation by Fontana et al.  where they observed DSA in 4.2% of the 95 patients analyzed. Our findings are also similar to previous observations in kidneys, where the prevalence of de novo DSA ranges from 6% to 27% [29-32]. This supports the hypothesis that the liver allograft is able to induce a humoral response in a similar fashion and frequency to other solid organ allografts.
We observed that DSA against HLA class II antigens was more common than class I and represented 95% of all de novo DSA. The higher prevalence of DSA against class II over class I observed in this study has also been shown in other solid organs but the explanation is still not clear [32-34]. It is known that there is a differential expression of class I and II antigens on the surface of liver cells, including endothelial cells, hepatocytes and biliary epithelial cells; and that this expression can be altered by different events, such as rejection episodes, infections, and inflammation related to transplantation [35-37]. Knowing that there is a higher constitutive expression of some HLA molecules (especially class I) in the liver, one may expect to find more alloantibodies against such molecules, as higher expression of alloantigen could translate into a stronger humoral response. Conversely, the higher expression of HLA antigens may facilitate more complete absorption thereby inhibiting one's ability to detect DSA in serum. Since DQ has the least expression in liver of all the HLA molecules, we cannot rule out the possibility that the higher prevalence of HLA-DQ antibodies is secondary to their inability to be absorbed by the liver [35, 36, 38]. However, if this were the case one would not necessarily find them to be detrimental, and we have shown that they are in some cases.
Another finding of our study is the discovery of four factors associated with the development of de novo DSA. First, the use of cyclosporine as opposed to tacrolimus carries more than twice the risk of de novo DSA formation. Since DSA and cyclosporine use have independently been linked to higher rates of acute and chronic rejection , it is no surprise that cyclosporine is now associated with an increased rate of DSA formation. Second, low levels of calcineurin inhibitors increase the risk of de novo DSA formation. It is not surprising that under immunosuppression increased the likelihood of mounting an alloimmune response . Third, a MELD score >15 at the time of transplant reduce the risk of de novo DSA formation in half. Patients transplanted with higher MELD scores are known to be sicker and at higher risk for infection pre-transplant. However, it is intriguing that they have a lower risk of de novo DSA formation by 1-year posttransplant. Lastly, recipients over 60 years of age displayed a lower likelihood of de novo DSA formation, which likely results from a less reactive immune system found in older patients .
Finally, we show that de novo DSA is an independent predictor of patient death and liver allograft loss, as shown in the Cox proportional hazards multivariable analysis. Although significant, the small difference in survival rates between groups observed in the Kaplan–Meier analysis (only 7% difference at 5 years for patient survival and 6% for graft survival) shows that the impact of DSA alone is not as pronounced as in other organs during the ≤5 years of follow up for this study, and opens the possibility for future research to find other cofactors or diagnostic tools to better identify patients at the highest risk for losing their allografts from DSA. Complement-fixing DSA, IgG subclasses of DSA , IgM DSA, and antibodies against non-HLA targets are some potential markers to examine in the near future. However, recent data has shown us that longer-term studies may be needed to fully understand the implications of post-liver transplant DSA, as slowly progressive idiopathic fibrosis may be a previously unrecognized consequence . Therefore, while we found an association between DSA with MFI >5000 and adverse outcomes, this does not rule out the possibility that DSA at lower MFI may be deleterious.
Even though this is a large cohort study of 749 patients, there are limitations. First, in an effort to achieve a homogeneous cohort, we did exclude patients without samples available for analysis at the predetermined time points and those with preformed DSA. Second, sample acquisition did not occur frequently enough for us to determine a temporal-relationship between the formation of de novo DSA and graft loss or patient death. Third, the lack of complete donor and recipient HLA-C and HLA-DP typing information did not allow for a complete DSA determination of all potential HLA loci, since these two loci can result in graft rejection [44, 45].
In conclusion, de novo DSA at 1-year post-liver transplant occurred in 8.1% of patients and was found to be an independent predictor of patient death and allograft loss. Future research will need to focus on improved mechanisms to identify pathologic DSA, outline monitoring and diagnostic algorithms, and design prospective treatment trials.