Anti‐HLA alloantibodies of the IgA isotype in re‐transplant candidates part II: Correlation with graft survival

We reported previously on the widespread occurrence of anti‐HLA alloantibodies of the IgA isotype (anti‐HLA IgA) in the sera of solid‐organ re‐transplantation (re‐tx) candidates (Arnold et al., ). Specifically focussing on kidney re‐tx patients, we now extended our earlier findings by examining the impact of the presence and donor specificity of anti‐HLA IgA on graft survival. We observed frequent concurrence of anti‐HLA IgA and anti‐HLA IgG in 27% of our multicenter collective of 694 kidney re‐tx patients. This subgroup displayed significantly reduced graft survival as evidenced by the median time to first dialysis after transplantation (TTD 77 months) compared to patients carrying either anti‐HLA IgG or IgA (TTD 102 and 94 months, respectively). In addition, donor specificity of anti‐HLA IgA had a significant negative impact on graft survival (TTD 74 months) in our study. Taken together, our data strongly indicate that presence of anti‐HLA IgA, in particular in conjunction with anti‐HLA‐IgG, in sera of kidney re‐tx patients is associated with negative transplantation outcome.


Summary
We reported previously on the widespread occurrence of anti-HLA alloantibodies of the IgA isotype (anti-HLA IgA) in the sera of solid-organ re-transplantation (re-tx) candidates (Arnold et al., 2013). Specifically focussing on kidney re-tx patients, we now extended our earlier findings by examining the impact of the presence and donor specificity of anti-HLA IgA on graft survival. We observed frequent concurrence of anti-HLA IgA and anti-HLA IgG in 27% of our multicenter collective of 694 kidney re-tx patients. This subgroup displayed significantly reduced graft survival as evidenced by the median time to first dialysis after transplantation (TTD 77 months)

| INTRODUC TI ON
Chronic antibody-mediated graft rejection (AMR) is a major cause for late loss of transplant function (Einecke et al., 2009;Gaston et al., 2010;Lee et al., 2002;Sellares et al., 2012). Accordingly, the presence of antibodies specific to allogenic human leukocyte antigen (HLA) in the serum of recipients is strongly associated with an increased risk of graft failure. Therefore, detection of donor-specific alloantibodies of the immunoglobulin G (IgG) subclasses is an integral component of most current AMR risk stratification strategies after kidney transplantation. Recently, however, the impact of other complement-fixing and also noncomplement-fixing antibody isotypes on graft survival has gained attention (Chen, Sequeira, & Tyan, 2011). As part of the 16th International HLA and Immunogenetics Workshop, we reported a high prevalence of anti-HLA IgA in a large cohort of re-transplantation candidates (Arnold et al., 2013).
Expanding these findings, we now assessed the impact of the presence and specificity of anti-HLA IgA alloantibodies on graft survival in a subset of 694 kidney re-transplantation patients.

| Patients and study design
Sera from 694 patients collected for our previous report (Arnold et al., 2013)  were typed for HLA-A, -B and -DR, 571 for HLA-DQ, and 439 for HLA-C (Table 1). A total of five donors had no available typing data and were therefore excluded from the analysis of donor specificity.
The number of mismatched HLA-loci for each transplantation was determined by comparing the split level HLA type of recipient and donor for HLA-A, -B, -DR and -DQ.

| Serum IgA levels
Quantification of human IgA in patients' sera was carried out by IMMAGE ® Immunochemistry Systems Kit #446460 (Beckman Coulter Inc., Fullerton, CA, USA) according to the manufacturer's instructions. Results for serum IgA levels are reported as mg/dl.

| Antibody screening and specification
All sera were tested for the presence and specificity of anti-HLA IgG and IgA as previously described (Arnold et al., 2013). For the purpose of this study, IgA1 and IgA2 subclasses are summarized as IgA.
DSA status was determined by matching anti-HLA alloantibodies compared to patients carrying either anti-HLA IgG or IgA (TTD 102 and 94 months, respectively). In addition, donor specificity of anti-HLA IgA had a significant negative impact on graft survival (TTD 74 months) in our study. Taken together, our data strongly indicate that presence of anti-HLA IgA, in particular in conjunction with anti-HLA-IgG, in sera of kidney re-tx patients is associated with negative transplantation outcome.  Bonferroni-corrected p-value cut-off (of p = .017 for 3 and p = .0083

| Statistical analyses
for four subgroups) was used to assert statistical significance.

| Anti-HLA IgA antibodies are frequent in renal re-transplantation candidates
For this follow-up study, 694 kidney re-transplantation patients were selected from a cohort of 803 patients collected for the 16th IHIW workshop based on serum and TTD availability (Arnold et al., 2013). Despite exclusion of all nonkidney re-transplantation candidates from the original cohort, demographics, HLA typing coverage as well as antibody prevalence in this sub-cohort are comparable to our earlier report (Table 2). Specifically, anti-HLA antibodies of the IgG and/or IgA isotype were detectable in 547 (79%) patients.
Stratifying by isotype revealed that 333 (48%) patients carried anti HLA-IgG exclusively. Moreover, 29 (4%) and 185 (27%) patients carried either anti-HLA IgA or anti-HLA antibodies of both isotypes respectively (Table 2, Figure S1). Therefore, a total of 31% of patients had detectable anti-HLA IgA antibodies in the serum, raising the question of the relevance of IgA in general and anti-HLA IgA in particular for graft survival.

| Total serum IgA levels do not impact graft survival in kidney re-transplantation candidates
Early studies hinted at a positive correlation of elevated serum IgA levels with enhanced kidney transplant survival Koka et al., 1993). In addition, a correlation between anti-HLA IgA and total serum IgA was also proposed .
Therefore, we first measured total serum IgA levels in our cohort  Figure 1), diverging from earlier observations in pretransplantation sera . Taken together, our data demonstrate that total serum IgA levels in renal re-transplantation candidates are globally similar to normal levels and not significantly associated with transplantation outcome.

| Donor specificity of anti-HLA IgA antibodies
Given   IgA alloantibodies can supplement routine IgG testing to improve risk stratification for kidney transplantations. However, effective treatment strategies of chronic AMR still prove to be controversial and understudied (Fehr & Gaspert, 2012). Accordingly, clinical impact of an improved stratification remains largely speculative. As our dataset did not include the cause for the loss of graft function, further studies are needed to examine whether IgA alloantibodies are specifically associated with any specific cause of graft loss like AMR.
For this study, we tested sera from patients with a previously failed graft who were relisted for transplantation. In addition, patients' sera were collected at or after the first dialysis after the subsequent transplantation (on average 48 months). Therefore, we cannot address the aspect of the initial time point of IgG and/  (Moroni et al., 2013).
Moreover, post-transplantation IgA deposition appears to be associated with reduced graft function in IgA nephropathy patients (Lefaucheur et al., 2016). Interestingly, another recent study including but not limited to IgA nephropathy patients found a similar association of post-transplantation IgA deposition and graft dysfunction (Sofue et al., 2015). Therefore, prolonged IgA deposition is a plausible mechanism for a reduction in graft function eventually resulting in graft failure. In contrast to these studies, however, our analysis specifically focuses on anti-HLA IgA. Consequently, F I G U R E 3 Graft survival after stratification by IgA DSA status several points argue against increased IgA deposition as a major contributor to graft dysfunction in anti-HLA IgA-positive patients.
First, anti-HLA IgA likely represent a minor fraction of total serum IgA. And second, we do not observe elevated IgA serum levels in IgA-positive compared to IgA-negative patients in our study.
Functional studies are therefore mandated to clarify if anti-HLA IgA contributes directly to graft failure and elucidate the mechanisms involved.
Several earlier studies established links between the presence of IgA autoantibodies against human immunoglobulin domains and good kidney graft survival (Susal, Kropelin et al., 1995;. Importantly, elevated total serum IgA was also proposed to be associated with the presence of anti-HLA IgA and enhanced 3-year graft survival at least in primary recipients . Taken together, these studies therefore eluded to a potential positive correlation of anti-HLA IgA with superior transplantation outcome. Our data, however, demonstrate clearly that the presence of anti-HLA IgA in kidney re-transplantation candidates is a hallmark of patients with significantly worse outcome, in particular in conjunction with anti-HLA IgG. This apparent inconsistency likely reflects differences in study design. First, our multicenter study exclusively analyzed highly allosensitized re-transplantation candidates. This patient collective allows for the assessment of the time frame between transplantation and first dialysis after transplantation, a very consistent estimate for the lifetime of a functioning kidney graft.
Second, our observation period extended significantly beyond the initial 3 years after transplantation and therefore included many late graft rejection events unobserved in earlier studies. Therefore, our study significantly expands previous findings and, to our knowledge, is the largest systematic assessment of a correlation between presence and specificity of anti-HLA IgA and kidney graft survival.
Notably, anti-HLA IgA-positive patients displayed an elevated average number of HLA mismatches compared to IgA-negative patients, while age and gender were not correlated with the presence of anti-HLA IgA. An increase in the number of HLA mismatches clearly represents a potential confounding effect for the observed shorter TTD (Opelz & Dohler, 2007). However, anti-HLA IgG-positive patients in our cohort were also characterized by an increased number of HLA mismatches compared to anti-HLA IgG-negative patients. Therefore, our data are in line with a nonisotype-specific positive correlation of the degree of immunization and the development of anti-HLA antibodies, rather than a specific correlation with IgA. Nevertheless, a larger study allowing for multivariate analyses will be required to investigate dependency of covariates, including anti-HLA IgA, contributing to graft loss. Alternatively, longitudinal studies can permit assessment of the timing of IgG and IgA antibody development after transplantation. In the context of such studies it would be particularly interesting to compare and contrast IgG and IgA antibody (donor-) specificities, which would also help clarify a possible interdependence of anti-HLA IgG and IgA antibodies.
In summary, our study demonstrates significantly reduced kidney graft survival for patients carrying anti-HLA IgA and IgG alloantibodies in a large multicenter cohort of patients with previously failed grafts. Therefore, inclusion of anti-HLA IgA screening in postand potentially pretransplantation testing can be advisable for improved risk stratification for kidney transplantations.

ACK N OWLED G EM ENTS
We thank Dr. Daniel Zecher for helpful discussion of results and conclusions and are indebted to Prof. Ilias Doxiadis for help with collecting data on the time to first dialysis after transplantation.

E TH I C S
Due to the study design, no explicit IRB approval was required.

CO N FLI C T S O F I NTE R E S T
The authors declare no conflicts of interest.