The MHC class I MICA gene is a histocompatibility antigen in kidney transplantation

The identity of histocompatibility loci, besides human leukocyte antigen (HLA), remains elusive. The major histocompatibility complex (MHC) class I MICA gene is a candidate histocompatibility locus. Here, we investigate its role in a French multicenter cohort of 1,356 kidney transplants. MICA mismatches were associated with decreased graft survival (hazard ratio (HR), 2.12; 95% confidence interval (CI): 1.45–3.11; P < 0.001). Both before and after transplantation anti-MICA donor-specific antibodies (DSA) were strongly associated with increased antibody-mediated rejection (ABMR) (HR, 3.79; 95% CI: 1.94–7.39; P < 0.001; HR, 9.92; 95% CI: 7.43–13.20; P < 0.001, respectively). This effect was synergetic with that of anti-HLA DSA before and after transplantation (HR, 25.68; 95% CI: 3.31–199.41; P = 0.002; HR, 82.67; 95% CI: 33.67–202.97; P < 0.001, respectively). De novo-developed anti-MICA DSA were the most harmful because they were also associated with reduced graft survival (HR, 1.29; 95% CI: 1.05–1.58; P = 0.014). Finally, the damaging effect of anti-MICA DSA on graft survival was confirmed in an independent cohort of 168 patients with ABMR (HR, 1.71; 95% CI: 1.02–2.86; P = 0.041). In conclusion, assessment of MICA matching and immunization for the identification of patients at high risk for transplant rejection and loss is warranted.

Here, we evaluate the role of MICA matching and donor-specific MICA immunization in a retrospective multicenter French cohort of 1,356 patients who had undergone kidney transplantation. All known covariates relevant to graft failure and acute rejection were considered in the analysis. The results highlight the relevance of both MICA matching and donor-specific immunization for kidney transplantation outcomes.

Results
Baseline characteristics of kidney transplant recipients. The main analysis involved 1,356 patients who underwent kidney transplantation in six French medical centers between 2002 and 2011: 104 in Montpellier, 107 in Paris-Saint-Louis, 188 in Toulouse, 262 in Paris-Necker, 304 in Nancy and 391 in Nantes. The demographics of this study population are listed in Table 1. Most patients were recipients of their first transplant (95%). One hundred and two patients received organs from living donors and 9% of patients received simultaneous kidney-pancreas transplantations. All but two of the relevant covariates for the clinical outcomes analyzed were equally distributed in the MICA-matched and -mismatched patients. There were more retransplantations in the MICA-matched than in the MICA-mismatched groups (10% versus 5%, P = 0.04), and MICA-mismatched transplantations had more HLA mismatches (P < 0.001, P < 0.001 and P = 0.01 for HLA-A, -B and -DRB1 mismatches, respectively; Table 1); both observations are probably due to linkage disequilibrium between MICA and HLA-B.
MICA matching and graft survival. The median follow-up after transplantation was 6.3 years, with a maximum of 12.9 years. The median follow-up was 6.5 and 6.3 years for the MICA-matched and -mismatched patients, respectively. A total of 192 patients (14.2%) had graft failure during follow-up; 1,208 patients (89.1%) survived. Compared with MICA-mismatched patients, MICA-matched patients had a significantly improved graft survival rate (P log-rank = 0.017), which was the primary endpoint of the study (Fig. 1a). At 5 years after transplantation, graft survival was 96% and 88% for MICA-matched and -mismatched patients, respectively, and this difference in survival rate was also observed when comparing the different mismatching possibilities at the MICA locus (0 versus 1 versus 2 mismatches, P log-rank = 0.008) (Fig. 1b). The most important impact on graft survival was observed for the case of two mismatches, with rates of 87% and 76% at 5 and 10 years after transplantation, respectively. Based on multivariate Cox regression, MICA mismatching was an independent factor associated with graft loss (HR, 2.12; 95% CI: 1.45-3.11; P < 0.001). Other independent risk factors in the model included age of the donor and recipient, dialysis duration, initial nephropathy, older transplantations, delayed graft function and absence of induction treatment ( Table  2). HLA-A, -B and -DRB1 mismatching at a low level of resolution had no impact on graft failure (Extended Data Table 1).
To exclude potential bias due to the difference in the resolution of MICA and HLA genotypes, we analyzed a subset of 862 transplants in which both donor and recipient were retrospectively HLA-typed at second-field resolution, which corresponds to allele-level resolution of MICA typing. Multivariate analysis confirmed the HLA-independent association of MICA mismatches with a higher incidence of graft loss (HR, 1.53; 95% CI: 1.07-2.19; P = 0.018; Extended Data Table 2). Other risk factors for graft loss in the model included age of the donor and recipient, dialysis duration, initial nephropathy, pre-transplantation anti-HLA DSA, number of transplantations, absence of induction treatment, depleting induction treatment and HLA-DQB1 mismatches (Extended Data Table  2). We also confirmed the HLA-B-independent effect of MICA by analyzing HLA-B-matched transplantations in this subset of transplants (n = 33), in which MICA mismatches were still associated with lower graft survival (P log-rank = 0.015, Extended Data Fig. 1).
Finally, MICA eplet mismatches had a similar association with graft loss, but did not reach statistical significance (P log-rank = 0.11, Supplementary Fig. 1).

Impact of preformed anti-MICA DSA on graft outcome.
Although there is no functional analogy between HLA and MICA molecules, however, to establish whether the observed lower graft survival associated with donor-recipient MICA mismatches might be explained by immunization against MICA (similarly to the situation between HLA mismatches and anti-HLA DSA), we analyzed the pre-transplant sera of 524 patients for the presence of anti-MICA DSA. In this subset of patients, the median follow-up was 5.80 years (with a maximum at 9.58 years) in those with anti-MICA DSA, and 6.04 years (with a maximum at 10.09 years) in those without anti-MICA DSA (Supplementary Table 1). Given that acute rejection is a major cause of kidney transplantation failure (HR, 2.64; 95% CI: 2.15-3.25; P < 0.001, Extended Data Table 3), we assessed whether donor-specific immunization against MICA had a role in this clinical event, which was the secondary endpoint of the study. Acute clinical rejection developed in 77 patients: TCMR in 52 (9.9%) and ABMR in 35 (6.7%), and of those 10 were mixed-type rejections (1.9%). The presence of anti-MICA DSA was found to be an independent risk factor for acute rejection, with a borderline but significant effect on TCMR (HR, 2.11; 95% CI: 1.01-4.42; P = 0.047) and a more important effect on ABMR (HR, 3.79; 95% CI: 1.94-7.39; P < 0.001; Fig. 2a and Table 3). Preformed anti-MICA DSA were not associated with graft loss (HR, 1.32; 95% CI: 0.82-2.10; P = 0.25; Table 3). The association of eplet-specific anti-MICA DSA with ABMR was similar to that of all anti-MICA DSA (Supplementary Fig. 2 and Extended Data Table 4).
One year post-transplant anti-MICA DSA and graft outcome. Immunization against MICA was analyzed using 225 serum samples collected 1 year after transplantation. In this subset of patients the median follow-up was 7.37 years (with a maximum at 9.58 years) and 7.34 years (with a maximum at 9.65 years) in those with and without anti-MICA DSA, respectively (Supplementary Table 2).
Although the presence of anti-MICA DSA at 1 year after transplantation was not associated with a higher incidence of graft failure, it was a risk factor for both TCMR (HR, 1.60; 95% CI: 1.01-2.53; P = 0.043) and ABMR (HR, 9.92; 95% CI: 7.43-13.20; P < 0.001; Fig. 2b and Table 3). Moreover, these associations were maintained when considering only the de novo fraction of these antibodies. Interestingly, the presence of de novo anti-MICA DSA was also a risk factor for graft survival (HR, 1.29; 95% CI: 1.05-1.58; P = 0.014; Table 3). Finally, the presence of anti-MICA DSA after transplantation was associated with a higher frequency of MICA mismatches whether considering all DSA present at 1 year after transplantation (0% versus 24.6% in matched versus mismatched patients, P = 0.0017) or only the de novo fraction of these antibodies (0% versus 13.5% in matched versus mismatched patients, P = 0.05).
We also tested whether specific MICA alleles were more prone to elicit DSA than others. For this purpose, we conducted a chi-squared test for equality of proportions on the proportion of individuals developing de novo anti-MICA DSA conditional on the presence of a specific MICA allele in the donor. There was no specific MICA allele that was associated with a higher rate of de novo anti-MICA DSA (Extended Data Table 5). Finally, when considering only eplet-specific anti-MICA DSA, the association with ABMR was similar to that of all anti-MICA DSA (Supplementary Fig. 3 and Extended Data Table 4).

Synergetic effect of anti-MICA and anti-HLA DSA on ABMR.
To evaluate the additive or synergetic impact of anti-MICA and anti-HLA DSA on ABMR, we analyzed the cumulative incidence of ABMR as a function of the presence or the absence of these antibodies before and after transplantation, as determined by single-antigen Luminex assays. The presence of anti-MICA or anti-HLA DSA, before and after transplantation, was a risk factor for ABMR (Fig. 3). In addition, both anti-MICA and anti-HLA DSA had an independent effect on ABMR, before and after transplantation (Extended Data  Fig. 3 and Extended Data Table 6).    Fig. 2a). Of note, the graft survival was worst when both anti-MICA and anti-HLA DSA antibodies were present, confirming a synergetic effect of these antibodies on graft survival (Extended Data Fig. 2b).   P values were determined using the two-sided log-rank test without correction.

Discussion
Here, we report that kidney transplantation from MICA-mismatched donors carries a significantly higher risk of graft failure. The lower graft survival can be explained by an increased rate of ABMR, which is independently associated with anti-MICA DSA. The present data formally define MICA as a bona fide transplantation antigen in kidney organ transplants and provide the rationale for including MICA genotyping and immunization monitoring in the pre-and post-transplantation workup. These results could be contextualized within several key, convergent and divergent, aspects of HLA and MIC genetics and immunobiology. On the genetic side, one of the major challenges in any association study involving MHC genes is the high degree of linkage disequilibrium within the complex, here exemplified using that between MICA and HLA-B, which are separated by a 46 kb stretch of DNA (Extended Data Table 7 provides an update on linkage disequilibrium between MICA and all classical HLA genes). This could mean that some of the observed associations could indeed be due to linkage disequilibrium rather than being a primary association. However, the contribution of linkage disequilibrium to our results was ruled out by inclusion of all HLA mismatches as covariates in the multivariate Cox model, as well as by the observation of a still-significant association of graft survival with MICA mismatches in the subset of donors and recipients who were allele-matched for HLA-B (Table 2 and Extended Data Fig. 1). This is also in line with an independent assessment of the contribution of MICA mismatching to the outcome of hematopoietic cell transplants 14,15 .
Despite attention to long-term follow-up, it should also be noted that HLA-A, -B and -DRB1 mismatches had no impact on graft survival in this cohort (Extended Data Tables 1 and 2). This is probably due to the comparatively smaller size of our cohort with respect to large, (multi) continent-wide cohorts, which have been able to show HLA-dependent disease outcome in kidney transplant recipients; for example the Collaborative Transplant Study (CTS), UK Transplant and Eurotransplant, with more than 100,000 donorrecipient pairs 16,17 . The necessity of having large cohorts to show an HLA-mismatching effect is due to the following: there is only a 15% survival difference at 10 years after transplantation between fully matched kidneys and kidneys mismatched for both alleles at HLA-A, -B and -DRB1 loci 18 ; and the magnitude of this effect has decreased over the years as a positive effect from many allocation policies taking matching into account 19 . The absence of MICA from these allocation policies may indeed explain why fewer donorrecipient pairs are needed to highlight a significant impact of MICA mismatching on graft outcome and, in consequence, to further incentivize its inclusion in a pre-transplant workup. Interestingly, in the subset of transplants with high-resolution typing of six HLA loci, only HLA-DQB1 mismatches were associated with lower graft survival (HR, 1.71; 95% CI: 1.35-2.17; P < 0.001; Extended Data Table 2). This observation is in line with recent reports showing associations of HLA-DQB1 mismatches with acute rejection 20,21 and decreased graft survival 22 .
On the biological front, despite the fact that both MICA and HLA class I genes and molecules have a similar and unique tri-dimensional structure, major differences exist in their respective functions, for example HLA class I require both the β 2 -microglobulin and an endogenously derived peptide antigen for proper surface expression, and interact with the T cell receptor, whereas MICA does not require either β 2 -microglobulin or any peptide cargo for surface expression and interacts with a distinct receptor, NKG2D. Other differences include (and this is despite the fact that after HLA genes, MIC genes are the most polymorphic loci in the human genome) a substantially higher degree of diversity (for example, >8,000 HLA-B alleles versus >300 MICA alleles, vastly higher numbers of polymorphic positions for HLA molecules than MICA; see http://hla.alleles.org/alleles/index.html), and substantially stronger tissue expression for HLA class I than MICA (see comparative RNA sequencing data at https://gtexportal.org/home/ multiGeneQueryPage/MICA,HLA-B). Incidentally, the last two facts are probably the reason for the higher antigenicity of HLA compared with MICA molecules, as evidenced by the disparity in the level of mean fluorescence intensity for anti-MICA compared with anti-HLA antibodies.
Independently of the influence of MICA genetic incompatibility on graft outcome, our study equally showed that the presence of pre-and post-transplantation anti-MICA DSA was strongly associated with an increased incidence of ABMR ( Fig. 2 and Table 3), an effect that was independent of, and synergetic with, that of anti-HLA DSA (Fig. 3 and Extended Data Table 1). Indeed, because they were also associated with transplantation failure, de novo    anti-MICA DSA appeared to be more harmful than preformed antibodies ( Table 3). Given that these harmful antibodies are associated with MICA mismatches (0% versus 13.5% of patients with de novo antibodies in MICA-matched and -mismatched transplantations, respectively), they can be anticipated by performing pre-transplant MICA genotyping. Finally, anti-MICA DSA were confirmed to be harmful because they were associated with graft loss in an independent cohort of ABMR patients (Extended Data Fig. 2). Some of these observations were made in two subcohorts (pre-transplant and post-transplant) of the initial (master) cohort. Of note, patient inclusion in each subcohort depended solely on the availability of their sera (Supplementary Tables 1 and 2); and the incidence of the main endpoint analyzed in these subcohorts, ABMR, was not significantly different from that observed in the main cohort, that is: 6.3%    generally more immunized. The other unique covariate that was not equally distributed in patients with and without anti-MICA DSA was the proportion of potential recurrent nephropathies (11.7% versus 4.7%, P = 0.03), which was probably due to the fact that there were more retransplantations in these patients with potentially recurrent nephropathies than in those without (13.3% versus 6.9%). Based on structural accessibility, MICA polymorphic residues can be grouped in small patches of surface-exposed amino acids, called eplets, using HLAMmatchmaker 23 . According to work by Duquesnoy et al., first for classical HLA molecules 24 and later for MICA 25 , donor-specific eplets are thought to represent surface-accessible polymorphic amino acids prone to elicit DSA. Even though this theory has been verified for HLA (for example ref. 26 ), when considering MICA eplet mismatches instead of global MICA mismatches and eplet-specific anti-MICA DSA instead of all donor-specific anti-MICA DSA, similar results but no improvements in terms of associations with graft loss or ABMR could be evidenced in our dataset ( Supplementary Figs. 1-3 and Extended Data Table 4). This discrepancy with HLA might be explained by the fact that MICA-mismatched alleles considered as matched at the eplet level may have immunogenic characteristics that cannot be identified using the HLAMmatchmaker approach. The limited number of reported eplet validation sera for MICA and the less extensive knowledge of MICA structures and polymorphisms may also be reasons for the non-superiority of associations measured when restricting the analysis to eplets. To sum up, in contrast to the HLA setting, the global and eplet mismatching models performed equally well for MICA. Although immunologically more correct, the eplet model and the number of identified eplets for MICA might still need improvements to demonstrate its superiority over the global mismatching model. The outcomes of this study warrant further detailed investigations on the eplet model for MICA.
In conclusion, molecular typing of MICA in association with screening for anti-MICA antibodies has the potential to lower the incidence of kidney transplantation rejection and loss.

Online content
Any methods, additional references, Nature Research reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/ s41591-022-01725-2.