Eplet mismatch analysis in kidney transplantation: from concept to clinical practice

Abstract

Eplet mismatch analysis offers a refined approach to assessing donor-recipient compatibility in kidney transplantation, surpassing conventional antigen-level human leukocyte antigen (HLA) matching in predicting immunologic outcomes. By identifying polymorphic amino acid residues on HLA molecules recognized by B cell receptors, this method quantifies immunologic risk. Clinical studies demonstrate that high eplet mismatch loads, particularly at HLA-DQ, are strongly associated with de novo donor-specific antibody development, antibody-mediated rejection, and reduced graft survival. Single-molecule mismatch analysis further enables risk stratification when integrated with T cell epitope prediction tools such as PIRCHE-II (Predicted Indirectly Recognizable HLA Epitopes Presented by HLA Class II Molecules). This review outlines the immunologic basis, methodologies, and clinical evidence supporting eplet mismatch analysis. It also addresses current limitations in validation and clinical implementation, and proposes future directions for its use in personalized immunosuppression and organ allocation.

HIGHLIGHTS

Eplets are polymorphic residues on human leukocyte antigen (HLA) recognized by B cell receptors. High HLA-DQ eplet loads predict de novo donor-specific antibodies, antibody-mediated rejection, and poor graft survival. Eplet mismatch helps guide tailored immunosuppression and organ allocation. Combining eplet mismatch with PIRCHE-II improves alloimmune event prediction. Future research should focus on antibody verification, cohort validation, and standardizing analysis tools.

INTRODUCTION

The progress of kidney transplantation has been marked by significant advances in immunological knowledge, yet achieving optimal donor-recipient compatibility remains a major challenge. Despite improvements in immunosuppressive therapy, long-term graft survival remains limited, primarily due to antibody-mediated rejection (AMR) driven by donor-specific human leukocyte antigen (HLA) antibodies (DSAs) [1,2]. The immunogenicity of HLA is largely determined by differences in amino acid residues between donor and recipient HLA molecules [3].

Traditional HLA matching strategies, based on serological antigen-level compatibility, have long been the cornerstone of transplant immunology. However, they are increasingly recognized as inadequate for predicting long-term graft outcomes. The development of molecular-level compatibility assessment through eplet mismatch analysis represents a transformative shift toward precision medicine in transplantation [4].

With the introduction of high-resolution HLA typing and advances in structural immunology, a more refined approach to HLA incompatibility has emerged [5]. The concept of eplet mismatch, first described by Duquesnoy through the HLAMatchmaker algorithm, enables analysis of discrete polymorphic amino acid patches, termed eplets, on the surface of HLA molecules. These eplets are likely recognized as nonself by B cell receptors [6]. These structural motifs provide a more accurate measure of immunogenic potential than conventional antigen-level mismatches [7].

Multiple cohort studies have demonstrated that higher eplet mismatch loads, particularly at HLA-DQ loci, are strongly associated with the development of de novo DSAs (dnDSA), AMR, and reduced graft survival [5,8–10]. Moreover, several studies have underscored the importance of antibody-verified eplets [11,12]. Antibody-verified eplets are those for which direct evidence exists that they can be recognized by human alloantibodies, established through experimental validation [13]. Consequently, there is increasing interest in using eplet-based matching for immunologic risk stratification, tailored immunosuppression, and organ allocation algorithms [14,15].

This review summarizes current knowledge on eplet mismatch in kidney transplantation, highlights clinical evidence linking eplet load to transplant outcomes, and explores future directions for integrating molecular-level HLA compatibility into precision transplant medicine.

EPLET DEFINITION AND MOLECULAR MISMATCH ANALYSIS

Molecular Basis of Epitopes and Eplets

The concept of the epitope was first introduced by Niels Kaj Jerne in 1960 to describe the specific structural region on an antigen recognized by an antibody [16]. Epitopes are now a central concept in immunology, with applications in vaccine development and T cell immunity [14]. The functional epitope refers to the precise region that interacts with the complementarity-determining region 3 (CDR3) of the antibody’s heavy chain, thereby dictating antibody specificity [3]. In contrast, the structural epitope encompasses all amino acids of the HLA molecule that participate in binding to the antibody paratope, typically spanning a region with a radius of approximately 15 Å [3].

The difficulty of finding compatible donors for highly sensitized kidney transplant recipients prompted Rene Duquesnoy to develop the HLAMatchmaker software [17]. This tool enables detailed comparison of donor and recipient HLA amino acid sequences, identifying mismatched amino acid configurations that may act as immunogenic epitopes. Research has shown that mismatches involving these configurations are associated with alloantibody development in both kidney transplant recipients and women sensitized during pregnancy [18].

As HLAMatchmaker advanced, the concept of the “eplet” was introduced [6]. Unlike linear amino acid triplets, eplets are defined as polymorphic amino acid residues within a 3.0–3.5 Å radius. These residues may be discontinuous, reflecting the three-dimensional structure of epitopes recognized by B cell receptors (Fig. 1).

Fig. 1.

Schematic representation of the concepts of epitope and eplet. An epitope is a three-dimensional antigenic region on the human leukocyte antigen (HLA) molecule recognized by an antibody. In contrast, an eplet is a smaller, structurally defined configuration of polymorphic amino acids within the epitope, typically located within a 3–3.5 Å radius, and is considered the minimal functional unit for antibody binding. Reproduced from Mattoo et al. [21] according to the Creative Commons License.

Human Leukocyte Antigen-DQ in Transplant Alloimmunity

Among HLA antigens, HLA-DQ has drawn increasing attention due to its strong association with adverse transplant outcomes, including graft dysfunction and loss [19]. Understanding HLA-DQ mismatches is complicated by the structural diversity of HLA-DQ molecules, which results from the ability of the HLA-DQB1-encoded β chain to pair with multiple α chains encoded by HLA-DQA1. These heterodimers can form within the same chromosome (cis) or between different chromosomes (trans) [20]. Depending on whether the donor is homozygous or heterozygous at the relevant loci, up to four distinct HLA-DQ heterodimers may be expressed [19]. HLA-DQ DSAs have been independently linked to an increased risk of acute rejection and inferior graft outcomes in kidney transplant recipients, beyond the predictive capacity of traditional HLA-A, -B, and -DR mismatches [22–25]. Among DSAs, those directed against HLA-DQ are more frequent than antibodies targeting other HLA loci, and they often appear at high titers, contributing substantially to allograft injury. In a prospective cohort, DeVos et al. [23] reported that 77% of patients who developed dnDSA had antibodies directed against HLA-DQ, and these patients exhibited worse graft outcomes. Similarly, Willicombe et al. [24] observed that 54% of dnDSA cases involved HLA-DQ antibodies, which were strongly associated with AMR, transplant glomerulopathy, and graft failure. A Korean cohort study also demonstrated that approximately 64.6% of patients with dnDSA after kidney transplantation had DQ-DSA [25]. The central role of HLA-DQ eplet mismatch in predicting dnDSA, rejection, and graft failure across diverse populations highlights its utility as a precision tool for risk prediction, personalized immunosuppression, and equitable donor-recipient matching in kidney transplantation [10,11,26,27].

Molecular Mismatch Analysis Tools

The advent of high-resolution HLA molecular typing has enabled the development of sophisticated computational tools for quantifying molecular mismatch load, thereby refining alloimmune risk assessment in transplantation. The clinical utility of these tools has been recognized by the Sensitization in Transplantation: Assessment of Risk (STAR) working group, which recommends integrating molecular mismatch analysis into pre- and posttransplant risk assessment protocols [28].

Table 1 summarizes the molecular mismatch analysis methods used in transplantation research. HLAMatchmaker, one of the earliest and most widely applied platforms, models eplet mismatches by comparing donor and recipient HLA amino acid sequences to identify immunologically relevant differences [29]. As of July 2020, version 3.1 of HLAMatchmaker was released for analyzing HLA-ABC and HLA-DRDQDP antibodies. The HLA Fusion software from OneLambda incorporates HLAMatchmaker within its antibody analysis platform, using the same database and calculation process while adding additional eplet data and offering a user-friendly eplet calculation tool [30]. The HLA Eplet Registry website also provides an accessible online calculator for eplet mismatch analysis, including single-allele mismatch load estimation (http://www.epregistry.com.br/calculator).

Table 1.

HLA molecular mismatch analysis tools

Program	Platform	Key features	Primary applications
HLAMatchMaker	Microsoft Excel	• Eplet-based structural modeling • Manual data entry • Version-specific reference tables • Includes antibody-verified eplets	• Detailed eplet-level compatibility assessment
One Lambda HLA Fusion	Clinical software	• Automated LABScreen/LABType data import • Software for HLA antibody screening & molecular typing • Graphical result review of antibody reactivities with adjustable cutoffs • Integrated HLAMatchmaker algorithms	• Standardized donor-recipient compatibility screening
HLA Eplet Registry Calculator	Web database	• Single-allele mismatch load calculator • Eplet frequency analysis • Publicly accessible via http://www.epregistry.com.br/calculator	• Quick, user-friendly eplet load calculation • Epidemiologic analyses
HLA-EMMA	Web-based	• Identifies solvent-accessible amino acid mismatches • Predicts B cell epitope recognition • Batch analysis capability	• Mapping of mismatch loci and epitope mapping
Cambridge EMS-3D (Electrostatic Mismatch)	MATLAB/Python	• Electrostatic mismatch score 3D (EMS3D) • Prioritizes charged residue differences • 3D protein structure modeling	• Physiochemical mismatch quantification • Immunogenicity prediction
PIRCHE-II	Web service/application programming interface	• T cell epitope prediction • NetMHCpan binding affinity scoring • Indirect allorecognition prioritization	• CD4+ T cell help risk stratification for donor-specific antibody production

HLA, human leukocyte antigen; HLA-EMMA, HLA Epitope Mismatch Algorithm; PIRCHE-II, Predicted Indirectly Recognizable HLA Epitopes Presented by HLA Class II Molecules.

The interaction between B cell epitopes and B cell receptors is mediated by surface-accessible amino acids forming noncovalent bonds, with binding affinity influenced by the electrostatic properties of polar and charged residues [31,32]. Physicochemical mismatches—particularly differences in electrostatic charge and hydrophobicity at HLA-A, -B, -DR, and -DQ loci—have been associated with alloantibody formation in sensitized kidney transplant recipients [33,34]. The Electrostatic Mismatch Score (EMS), developed by the Cambridge group, quantifies differences in electrostatic potential between donor and recipient HLA molecules. EMS scores for HLA-DR and -DQ have been shown to predict alloantibody formation in both graft failure [35] and in recipients with low immunologic risk in kidney transplantation [5]. The EMS program was later refined into EMS-3D by incorporating the tertiary structure of HLA molecules. This enhanced score has been linked to alloantibody formation in women sensitized by donor lymphocyte injections from their partners and in kidney transplant recipients who experienced graft failure [36]. HLA-EMMA is another tool that evaluates HLA class I and II compatibility at the amino acid level, with a focus on polymorphic, solvent-accessible residues likely to be recognized by B cell receptors [37]. HLA-EMMA enables batch analysis of amino acid mismatches between donor and recipient while identifying solvent-accessible residues that may interact with B cell receptors and antibodies. Because the catalog of eplets is continuously evolving, mismatch loads and thresholds vary across studies and remain difficult to compare. In contrast, amino acid mismatch analysis provides greater consistency, as HLA allele sequences are fixed [3].

The PIRCHE-II (Predicted Indirectly Recognizable HLA Epitopes Presented by HLA Class II Molecules) algorithm predicts the repertoire of donor-derived peptides that can be presented by recipient HLA class II molecules to CD4+ T cells, thereby quantifying the risk of indirect allorecognition and subsequent DSA generation [38]. More recently, the Snow algorithm, developed by the PIRCHE group, has been introduced as an advanced B cell epitope matching approach [39,40]. The Snow algorithm evaluates antibody epitope mismatches by identifying surface-exposed amino acid differences between donor and recipient HLA molecules. The resulting Snow score quantifies the number of surface-accessible mismatched amino acids and serves as a surrogate measure of antibody epitope burden. The Snow algorithm remains under clinical validation and is currently restricted to research settings. Fig. 2 illustrates representative platforms commonly used for eplet and epitope mismatch analysis.

Fig. 2.

Representative tools for eplet and epitope mismatch analysis in transplantation. (A) HLAMatchmaker: a downloadable program that calculates B cell eplet mismatches using antibody-verified and theoretical eplet definitions. It enables detailed analysis of human leukocyte antigen (HLA) class I and II mismatches in Microsoft Excel format. (B) HLA Epitope Registry Mismatch Calculator: an online tool that identifies nonself eplets from donor/recipient HLA typing, referencing curated eplet data from the international HLA Epitope Registry. (C) HLA-EMMA (HLA Epitope Mismatch Algorithm): a web-based platform that performs automated B cell epitope mismatch analysis using public HLA sequence databases. It considers surface accessibility, physicochemical properties, and mismatch load at the epitope level. (D) PIRCHE-Tx Predictor: a web interface that calculates PIRCHE-II (Predicted Indirectly Recognizable HLA Epitopes Presented by HLA Class II Molecules) for CD4+ T cell responses. It estimates the number of mismatched peptides potentially presented by recipient HLA class II molecules, providing insights into T cell-mediated alloreactivity.

CLINICAL EVIDENCE OF EPLET MISMATCH IN KIDNEY TRANSPLANTATION

Eplet Mismatch and De Novo Donor-Specific Antibody Formation

The association between HLA eplet mismatch and dnDSA development has been consistently demonstrated across diverse transplant populations (Table 2) [1,4,8,10–12,18,26,27,39–53]. Dankers et al. [18] first reported that HLA triplet mismatches predicted dnDSA, forming the foundation for subsequent eplet-based models. In 2013, Wiebe et al. [41] showed that class II eplet thresholds (DR<10, DQ<17) outperformed antigen-level matching in predicting dnDSA. In a larger cohort, Wiebe et al. [42] further demonstrated that class II eplet mismatch influenced both dnDSA risk and the tacrolimus trough level required to prevent dnDSA development.

Table 2.

Studies of eplet mismatch and clinical outcomes in kidney transplantation

Study	Cohort	Type	Outcome	Eplet mismatch cutoff (thresholds)	Main findings
dnDSA
Dankers et al. (2004) [18]	144 Sensitized kidney transplant recipients	Triplet	dnDSA	≥11 or 12 triplet mismatches	• Number of HLA triplet mismatches correlated strongly with antibody formation, predicting both graft rejection and pregnancy alloimmunization
Wiebe et al. (2013) [41]	286 Kidney transplant recipients	Eplet	dnDSA	10 For HLA-DR, 17 for HLA-DQ eplet mismatch	• Class II HLA eplet matching outperformed traditional antigen-based matching in predicting dnDSA, with optimal thresholds to minimize dnDSA development
Wiebe et al. (2017) [42]	596 Kidney transplant recipients	Eplet–class II	dnDSA	>11 HLA-DR or DQ eplet mismatch	• Highly mismatched eplets were more likely to develop dnDSA, and class II eplet mismatch modulated the tacrolimus trough levels required to prevent DSA
Wiebe et al. (2019) [26]	664 Kidney transplant recipients	Eplet–class II	dnDSA, AMR, TCMR	Low, DR<7 and DQ<9; intermediate, DR≥7 and DQ<14; high, DR 7–22 and DQ 15–31 for single molecular mismatch	• HLA‐DR/DQ single-molecule eplet mismatches were significantly correlated with primary alloimmune events, including T cell-mediated rejection, HLA‐DR/DQ dnDSA development, and AMR
Philogene et al. (2020) [43]	110 Pediatric kidney transplant recipients who received their first organ from a donor of the same race (SRT) versus a donor of a different race (DRT)	Eplet	dnDSA	Class I eplet load was≥70	• For the entire population, the risk of de novo HLA-DSA development was significantly increased with higher eplet loads • The risk of rejection increased significantly for DRT compared with SRT, only when class I eplet load was≥70
Charnaya et al. (2021) [44]	125 Pediatric kidney transplant recipients	Eplet–class II	dnDSA	Not available	• Patients who developed dnDSA had a significantly higher median eplet load compared to those who did not, 64 (IQR, 46–83) vs. 77 (IQR, 56–98) • The most frequent dnDSA targets were eplets found on HLA-A*11 and A2, and HLA-DQ6 and DQA5 • Frequent eplet mismatches did not necessarily correlate with an increased likelihood of dnDSA formation
Tafulo et al. (2021) [45]	96 Kidney transplant recipients	Total and AbVer eplet mismatch	dnDSA	Not available	• High class II antibody-verified eplet mismatch (hazard ratio, 1.105) was a superior predictor of dnDSA compared to antigen mismatch
Lee et al. (2022) [10]	347 Korean kidney transplant recipients	Eplet–class II	dnDSA	Single molecular eplet (DQ≥10), total eplet (DQ≥12), antibody-verified eplet (DQ≥4), and antibody-verified single molecular eplet (DQ≥4)	• A high level of mismatch significantly correlated with HLA class II dnDSA development • HLA class II epithelial mismatch may improve risk stratification for dnDSA development in combination with tacrolimus trough levels
Wiebe et al. (2023) [46]	949 Kidney transplant recipients	Eplet–class II	dnDSA	Low, DR<7 and DQ<9; intermediate, DR≥7 and DQ ≤14; high, DR 7–22 and DQ 15–31 for single molecular mismatch	• Younger age and high HLA-DR/DQ single molecular mismatch were independent predictors of dnDSA • Age + mismatch risk grouping improved risk stratification
Tran et al. (2024) [47]	21 Patients with failed kidney transplants	Eplet–class II	dnDSA, cPRA	DR≥12, DQ≥12	• High HLA-DQ eplet mismatches were associated with an increase in cPRA and de novo DQ-DSA in recipients with failed kidney • No significant impact from HLA-DR eplet mismatch
Wong et al. (2024) [27]	234 Southeast Asian kidney recipients	Eplet–class II	dnDSA	Low, DR<7 and DQ<9; intermediate, DR<7–11 and DQ≥9 or DR≥12 and DQ 9–14; high, DR≥12 and DQ≥15 for single molecular mismatch	• HLA-DR/DQ single molecular mismatch predicts dnDSA development in Southeast Asian recipients • Low-risk eplet mismatch → only 1% dnDSA in 5 years
Rejection and graft failure
Tafulo et al. (2019) [8]	151 Kidney transplant recipients	Eplet	AMR	HLA-II: T1, 0–5; T2, 6–12; T3, ≥13 HLA-DR: T1, 0–1; T2, 2–5; T3, ≥6 HLA-DQ: T1, 0–1; T2, 2–6; T3, ≥7	• High HLA-DR/DQ eplet mismatch was an independent predictor of AMR compared to traditional antigen mismatch
Senev et al. (2020) [11]	926 Kidney transplant recipients	Eplet	dnDSA, graft failure	Not available	• Antibody-verified eplet mismatch load, particularly at HLA-DQ, is independently associated with dnDSA development and graft failure
Sypek et al. (2020) [48]	196 Pediatric recipients	Eplet	dnDSA, graft survival	Not available	• HLA class I eplet mismatches predicted graft survival, and HLA class II eplet mismatches predicted retransplantation risk • Eplet mismatch was associated with dnDSA
Sapir-Pichhadze et al. (2020) [12]	118,382 U.S. first kidney transplant recipient cohort + 1,753 Canadian cohort for validation	Eplet–antibody-verified	Graft failure, glomerulopathy	Per single antibody-verified mismatch, per single overall mismatch, per 10 antibody-verified mismatches, per 10 overall mismatches	• U.S. cohort: antibody-verified eplet mismatches were found to be independent predictors of death-censored graft failure with hazard ratios of 1.231, 1.268 and 1.411 for class I (HLA-A, B, and C), -DRB1 and -DQB1 loci, respectively • Canadian cohort: antibody-verified eplet mismatches were independent predictors of transplant glomerulopathy with hazard ratios of 5.511 for HLA-DRB1 and 3.640 for -DRB1/3/4/5
Mohammadhassanzadeh et al. (2021) [49]	118,313 From the Scientific Registry of Transplant Recipients	Eplet	Graft failure	Not available	• Data were analyzed from U.S. kidney transplant recipients to determine whether specific HLA eplet mismatches are associated with death-censored graft failure, and 15 eplet mismatches were analyzed as significant predictors
Arches et al. (2024) [50]	45 Kidney retransplantation	Eplet	AMR, graft survival	DSA targeting repeated eplet mismatches (DREMM)	• The presence of DREMM was associated with a higher frequency of AMR and a lower death-censored graft survival • However, multivariate analysis did not show that DREMM were associated with the risk of AMR
Arana et al. (2025) [51]	122 High-risk kidney recipients	Eplet–DQA1	AMR	DQA1 ≥2	• Very early AMR linked to DQA1 eplet mismatches
Li et al. (2025) [52]	64 dnDSA+ kidney recipients	Eplet–class II	AMR, dnDSA	Identify dnDSA-specific and AMR-associated mismatched eplets	• AMR was significantly associated with a greater number of mismatched eplets, especially those containing polar and positively charged amino acids • dnDSA-specific and AMR-associated mismatched eplets overlapped • The dominant mismatched eplets: 41 T, 163R, 25Q, 78 V, 47QL, 55PP
Intergration with complementary assessment tools
Ashimine et al. (2024) [1]	691 Kidney transplant recipients (112 sensitized, 579 unsensitized)	Eplet+PIRCHE-II	dnDSA	8 (eplet)/ 230 (PIRCHE-II) in sensitized and 5 (eplet)/ 176 (PIRCHE‐II) in unsensitized groups	• PIRCHE-II score and a history of acute TCMR were significant predictors of dnDSA • Eplet mismatch was a significant predictor only in the unsensitized recipients
Jäger et al. (2024) [4]	439 Kidney recipients	Eplet mismatch count + the number of highly immunogenic eplets + PIRCHE-II score	AMR, dnDSA	Low risk: total eplet mismatch count <73 + top 10 immunogenic eplet mismatch count <4 + PIRCHE II score <93 High risk: total eplet mismatch count ≥73 + top 10 immunogenic eplet mismatch count ≥4 + PIRCHE II score ≥93	• Identify top 10 immunogenic eplets for each class/locus • If all three molecular mismatch scores exceeded the cutoffs, the patient was categorized as high-risk • Patients in the high-risk group had significantly higher rates of rejection and dnDSA formation
Chou-Wu et al. (2025) [40]	843 Kidney recipients	Verified eplet mismatch (vEplet) + Snow + PIRCHE-II	dnDSA	HLA-A+B+C: Snow (>9), verified eplet (>8), PIRCHE-II (>43) HLA-DRB1+DQB1: Snow (>7), verified eplet (>3), PIRCHE-II (>30)	• Higher Snow scores, eplet mismatches, and PIRCHE-II scores are strongly linked to dnDSA development—individually and combined
Niemann et al. (2025) [39]	400,935 Kidney transplants (SRTR dataset)	Eplet + Snow + PIRCHE-II + amino acid	Graft survival	Optimal PIRCHE-II interval: [0–23), [23–45), [45-69], [69–∞] Optimal Snow interval: [0-17], [17-35], [35-47], [47–∞] Optimal antibody-verified eplet interval: [0-6], [6-16], [16-22], [22–∞]	• Eplet, PIRCHE-II, amino acid, Snow (B cell epitope mismatch) were calculated • Combination of B cell and T cell mismatch scores yielded the most precise classification of immunological risk for dnDSA development and graft outcomes
González-López et al. (2023) [53]	42 Kidney transplant recipients	Eplet + dd-cfDNA	Rejection	dd-cfDNA>1.0%	• The only pretransplant variable associated with dd-cfDNA >1.0% was HLA-DQB1 eplet mismatch load • Increased dd-cfDNA levels were associated with rejection

dnDSA, de novo donor-specific antibody; HLA, human leukocyte antigen; AMR, antibody-mediated rejection; IQR, interquartile range; cPRA, calculated panel-reactive antibody; PIRCHE-II, Predicted Indirectly Recognizable HLA Epitopes Presented by HLA Class II Molecules; TCMR, T cell-mediated rejection; dd-cfDNA, donor-derived cell-free DNA.

The concept of single-molecule eplet mismatch, introduced by Wiebe et al. [26], allowed precise thresholding at the allele level. Using HLAMatchmaker and ROC curve analysis, they established mismatch thresholds for each individual HLA-DR and -DQ molecule, enabling stratification of recipients according to whether a molecule exceeded its threshold (low: DR<7 and DQ<9; intermediate: DR≥7 and DQ<14; high: DR≥7 and DQ≥15).

Pediatric studies also support the association between class II eplet mismatch and dnDSA formation. Philogene et al. [43], in a cohort of 110 pediatric kidney transplant recipients, found that higher eplet loads significantly increased the risk of dnDSA development. Importantly, the risk of rejection was higher in recipients of kidneys from donors of a different race compared with same-race donors, but only when the class I eplet load was ≥70. Charnaya et al. [44] reported that pediatric recipients who developed dnDSA had significantly higher median eplet loads, with the most frequent dnDSA targets being eplets on HLA-A*11, A2, DQ6, and DQA5. Tafulo et al. [45] studied 96 living donor kidney transplants and found that higher HLA class II eplet mismatch loads were significantly associated with dnDSA development. Notably, eplet mismatch analysis outperformed traditional antigen mismatch in predicting alloimmune risk and graft survival [45]. In a cohort of 347 Korean kidney transplants, Lee et al. [10] identified specific HLA-DQ eplet mismatch cutoffs—single-molecule ≥10, total eplet ≥12, and antibody-verified or antibody-verified single-molecule eplet ≥4—that were significantly associated with class II dnDSA development. Wiebe et al. [46] proposed a model combining recipient age with single-molecule eplet mismatch to identify patients with excellent long-term outcomes, supporting cost-effective DSA surveillance. Tran et al. [47] reported that high DQ mismatch was associated with elevated calculated panel-reactive antibody (cPRA) and de novo DQ DSAs in recipients with failed allografts. More recently, a Southeast Asian cohort study (n=234) validated single-molecule HLA-DR/DQ eplet mismatch categories as predictors of dnDSA development [27]. Using both published and cohort-specific thresholds, the authors showed that eplet mismatch-based stratification correlated with dnDSA-free survival, with improved predictive value when thresholds were tailored to the cohort’s immunosuppressive regimen.

Impact of Eplet Mismatch on Rejection and Graft Survival

Beyond dnDSA, eplet mismatch has been strongly associated with both acute and chronic rejection and with long-term graft outcomes. Tafulo et al. [8] examined 151 living donor kidney transplants and showed that a high HLA class II eplet mismatch load (≥13) independently predicted AMR, whereas traditional antigen mismatch did not. Senev et al. [11] analyzed 926 kidney transplant pairs and demonstrated that higher antibody-verified HLA-DQ eplet mismatch load was independently and linearly associated with dnDSA formation, rejection, and graft failure. In pediatric recipients, Sypek et al. [48] studied 196 cases and found that antibody-verified eplet mismatches were associated with dnDSA development. Although eplet mismatch significantly predicted graft failure and reduced the likelihood of retransplantation in univariate analysis, this association did not remain after adjustment [48]. Sapir-Pichhadze et al. [12] analyzed over 118,000 U.S. kidney transplant recipients and found that higher antibody-verified HLA eplet mismatches were independently associated with increased risk of graft failure, particularly at HLA-DQ and -DR loci. These findings were confirmed in a Canadian cohort, where eplet mismatches also predicted transplant glomerulopathy. Mohammadhassanzadeh et al. [49] analyzed 118,313 first kidney transplant recipients with 0% PRA and identified 15 specific HLA eplet mismatches that significantly increased the risk of death-censored graft failure. Using network and variable selection analyses, they demonstrated that a small subset of immunodominant eplet mismatches could predict long-term outcomes, supporting their role in personalized risk stratification and organ allocation. Arches et al. [50] reported that repeated mismatches at immunogenic eplets contributed to persistent DSAs, higher AMR incidence, and reduced graft survival after retransplantation. Arana et al. [51] showed that higher DQA1 eplet mismatches were significantly associated with early AMR in highly sensitized patients. Li et al. [52] analyzed 64 dnDSA-positive kidney transplant cases and found that HLA class II dnDSAs developed later, had higher mean fluorescence intensity (MFI) values, and were more frequently linked to AMR than class I dnDSAs. In these patients, the rejection group exhibited significantly more mismatched eplets, and dnDSA-specific eplets overlapped with those associated with AMR. Interestingly, these eplets were enriched in polar and positively charged amino acids, suggesting that specific biochemical properties of eplets may contribute to AMR development.

Integration of Eplet Mismatch With Complementary Assessment Tools

Several recent studies have underscored the value of integrating B cell and T cell epitope prediction models [1,4,39,40,54]. Ashimine et al. [1] analyzed 691 living donor kidney transplants to examine the relationship between B cell (eplet mismatch) and T cell (PIRCHE-II score) epitope matching and dnDSA development, stratified by prior sensitization status. In unsensitized recipients, both eplet mismatch and PIRCHE-II scores were significant predictors of class II dnDSA. In contrast, among previously sensitized recipients, only PIRCHE-II and a history of T cell-mediated rejection remained significant. Jäger et al. [4] studied 439 standard-risk kidney transplant recipients and demonstrated that combining three molecular mismatch methods—eplet mismatch count, highly immunogenic eplets, and PIRCHE-II score—enabled effective risk stratification. Patients with low scores across all measures had a substantially reduced incidence of rejection (12%) compared with those with high scores (33%, P=0.003). This integrated model consistently predicted clinical rejection, antithymocyte globulin-treated rejection, and dnDSA development, with high-risk patients showing more than a threefold increase in risk.

Chou-Wu et al. [40] further demonstrated that combining Snow (B cell epitope) and PIRCHE-II (T cell epitope) algorithms more than doubled the accuracy of DSA risk prediction. Niemann et al. [39] systematically examined over 400,000 kidney transplants from the SRTR (Scientific Registry of Transplant Recipients) database and confirmed that integrating B cell (Snow, eplet) and T cell (PIRCHE-II) mismatch scores provided complementary, independent information. The numerical combination of these scores produced the most precise classification of dnDSA risk and graft outcomes. Collectively, these findings support the clinical utility of incorporating molecular mismatch analysis into routine immunologic risk assessment.

In addition to epitope-based algorithms, emerging molecular biomarkers such as donor-derived cell-free DNA (dd-cfDNA) provide complementary insights for early risk stratification and monitoring. González-López et al. [53] examined the role of dd-cfDNA alongside eplet mismatch in kidney transplantation. At 1 month posttransplant, 26.2% of 42 kidney transplant recipients had dd-cfDNA levels ≥1.0%. Among pretransplant variables, only HLA class II eplet mismatch load—particularly HLA-DQB1—was significantly associated with elevated dd-cfDNA. Moreover, dd-cfDNA effectively distinguished patients with AMR (area under the curve [AUC], 87.3%), acute rejection (AUC, 78.2%), and compromised grafts (AUC, 81.4%). These results suggest that integrating eplet mismatch analysis with complementary tools enhances immunologic risk assessment, facilitates earlier detection of allograft injury, and supports personalized management in kidney transplantation.

Eplet-Guided Immunosuppression Strategies

Recent studies have highlighted the clinical utility of eplet mismatch analysis in guiding immunosuppressive strategies for kidney transplant recipients (Table 3) [10,42,46,55–57]. Wiebe et al. [42] demonstrated that HLA-DR/DQ eplet mismatch and tacrolimus trough levels independently predicted dnDSA. They showed that recipients with high eplet mismatch should avoid maintaining tacrolimus levels below 5 ng/mL unless absolutely necessary and should undergo close monitoring for dnDSA development [42]. Bestard et al. [55] reported that minimization to tacrolimus monotherapy was successful in low-risk recipients, defined by low eplet load and absence of preformed T cell alloimmunity. Davis et al. [56] found that tacrolimus trough levels <6 ng/mL were associated with a 2.3-fold increased risk of dnDSA in intermediate- and high-risk eplet mismatch groups, whereas no significant association was observed in low-risk groups. Moreover, HLA eplet mismatch predicted dnDSA risk and supported the strategy of maintaining higher tacrolimus exposure (≥8 ng/mL) in high-risk patients while permitting dose reduction in low-risk populations. Lee et al. [10] further emphasized the compounded risk when both high HLA-DQ eplet mismatch and low tacrolimus trough levels were present, showing a synergistic increase in dnDSA development. In a more recent study, Wiebe et al. [46] proposed a model combining eplet mismatch and recipient age to define a low-risk group with a 97% 10-year dnDSA-free survival rate, while reducing DSA surveillance costs by more than 50%. Johnson et al. [57] evaluated 588 kidney transplant recipients and found that class II eplet mismatch levels correlated with immune event rates in both belatacept/tacrolimus (Bela/TacTL) and tacrolimus-only groups. Notably, Bela/TacTL was associated with significantly lower risks of DSA formation, AMR, and rejection, particularly in recipients with low eplet mismatch. These findings suggest that eplet mismatch analysis can inform immunosuppression strategies, helping to identify high-risk patients who require intensified therapy while supporting safe immunosuppression minimization in low-risk populations.

Table 3.

Studies using eplet mismatch to guide immunosuppressive strategies

Study	Design	Key findings	Clinical implications
Wiebe et al. (2017) [42]	Retrospective cohort, kidney transplant recipients (n=596)	• HLA-DR/DQ eplet mismatch and tacrolimus trough <5 ng/mL were independent predictors of dnDSA • Recipients with high eplet mismatch are less likely to tolerate low tacrolimus without dnDSA development	• In high eplet mismatch patients, avoid targeting tacrolimus <5 ng/mL unless essential; monitor for dnDSA
Bestard et al. (2021) [55]	Prospective RCT, kidney transplant recipients (n=167)	• In low-risk patients (low eplet mismatch, low preformed T cell immunity), tacrolimus monotherapy was associated with higher acute rejection but fewer viral infections compared to standard therapy	• Eplet mismatch and T cell immunity can help identify candidates for immunosuppression minimization; caution is required in high-risk groups
Davis et al. (2021) [56]	Retrospective, U.S. multicenter cohort (n=444)	• Intermediate/high-risk eplet groups with a mean tacrolimus blood concentration <6 ng/mL exhibited a 2.34-fold increased risk of dnDSA development compared to those with levels >8 ng/mL • In low-risk eplet mismatch, dnDSA risk was low regardless of tacrolimus level	• High-risk eplet mismatch requires adequate tacrolimus; low-risk may allow dose reduction
Lee et al. (2022) [10]	Retrospective, Korean kidney transplant recipients (n=347)	• High HLA-DQ eplet mismatch and tacrolimus trough <5 ng/mL each independently increased dnDSA risk • Tacrolimus trough variability (time-weighted coefficient of variability) was not a significant risk factor	• Risk stratification for dnDSA should focus on eplet mismatch and tacrolimus trough, not tacrolimus variability
Wiebe et al. (2023) [46]	Predictive model, large cohort	• HLA-DR/DQ eplet mismatch plus age identified a low-risk group with 97% 10-year dnDSA-free survival and enabled >50% reduction in surveillance	• Surveillance can be reduced in low-risk patients, focusing resources on high-risk groups
Johnson et al. (2024) [57]	Prospective, belatacept with tacrolimus induction and 294 tacrolimus regimen	• The single-molecule DR/DQ eplet risk score from the Wiebe study was used • Class II eplet mismatch levels were associated with immune event rates (DSA, AMR, rejection) in both Bela/TacTL and tacrolimus-only groups • Bela/TacTL was linked to significantly lower risks of DSA formation, AMR, and rejection, especially in low eplet mismatch recipients	• The integration of eplet with immunosuppression protocols, such as belatacept combined with time-limited tacrolimus, reduces dnDSA incidence and AMR, especially in high eplet mismatch groups, while improving graft survival in low-risk patients

HLA, human leukocyte antigen; dnDSA, de novo donor-specific antibody; AMR, antibody-mediated rejection.

Clinical Implementation in Transplantation Programs

Eplet-based matching provides a sophisticated method for assessing donor-recipient compatibility and holds substantial promise for improving equity in national kidney transplant programs. Traditional HLA antigen matching often disadvantages candidates of non-European ancestry—such as Black, Asian, and Hispanic populations—because their HLA types are underrepresented in donor registries, limiting access to well-matched grafts [58,59]. Simulation analyses by Bekbolsynov et al. [59] have shown that race-informed molecular matching, using metrics such as hydrophobicity, electrostatic charge, and amino acid composition, can improve donor compatibility for underrepresented groups. Furthermore, eplet-based matching strategies validated in large-scale datasets, including the National Kidney Registry (n=5,193), demonstrated improved match quality for Black and Hispanic/Latino candidates in deceased donor allocation [58]. These findings support the incorporation of eplet-based mismatch assessment into national organ allocation policies to reduce racial disparities and promote equitable access to transplantation, while simultaneously optimizing long-term graft outcomes.

CRITICAL ISSUES IN EPLET MISMATCH ANALYSIS

Limitations in Eplet Immunogenicity Validation

While eplet-based HLA matching provides a more detailed approach to alloimmune risk assessment than conventional antigen-level matching, not all eplet mismatches carry equal immunologic risk [60,61]. The HLA Epitope Registry serves as a valuable resource for cataloging eplets and their antibody-verified status, but it also has notable limitations [13,14]. Many eplets were inferred from sequence data rather than experimentally validated, and some antibody-verified eplets were confirmed only with nonhuman antibodies. Bezstarosti et al. [13] reported that only 12% of eplets meet the highest verification standards (Level A1/A2), underscoring the need for standardized validation methods.

Devriese et al. [62] provided crucial evidence by systematically testing 29 previously unverified DQ eplets using patient sera and adsorption-elution assays. Among the 83 DQ eplets listed in the Registry, 66 remain unverified, highlighting the urgent need for further experimental validation. The authors confirmed the antigenicity of 24 DQ eplets, including both DQα and DQβ variants, thereby distinguishing truly immunogenic eplets from those inferred only by sequence alignment. In parallel, pregnancy has been used as a natural model of alloimmunization to investigate eplet immunogenicity [63–65]. Schawalder et al. [65] systematically mapped the immunogenicity of DQ eplets, identifying specific DQB1 eplets such as 55PP, 52PR, 52PQ, 85VG, and 45EV as highly immunogenic based on antibody response frequency in pregnancy cohorts. Recent technological advances, including the generation of recombinant human monoclonal antibodies, have enabled more precise functional validation of eplet immunogenicity [61,66,67]. Kramer et al. [61] evaluated HLA class I and class II eplets using a panel of human monoclonal HLA antibodies and single antigen bead assays, generating high-level evidence supporting or refuting the immunogenicity of several clinically relevant eplets. These findings have strong clinical implications. In kidney transplantation, Li et al. [52] confirmed that mismatches involving eplets such as 41T, 163R, 25Q, 78V, 47QL, and 55PP were dominant drivers of dnDSA formation and AMR. Similarly, Arches et al. [50] demonstrated that repeated mismatches at immunogenic eplets were associated with higher rates of AMR and reduced graft survival, with DSAs targeting repeated eplets being more persistent and exhibiting higher MFI values.

Human Leukocyte Antigen Nomenclature and Allelic Variability

A recent study introduced a novel framework for understanding HLA allele relationships by applying amino acid sequence analysis and advanced visualization tools. The epiArt project developed an R-based computational and graphical platform to visualize HLA allele relationships based on antibody-confirmed eplet disparities [68]. By translating amino acid sequences of confirmed eplets into a comprehensive atlas of HLA class I and II antigens, the researchers generated antigen-specific disparity graphs to display pairwise sequence differences among alleles. These visualizations revealed significant intragroup heterogeneity and highlighted the presence of shared eplets and epitopes across different antigen groups and loci. Amino acid motif plots further illustrated the prevalence and distribution of polymorphic residues within each antigen group. Collectively, these findings underscore the limitations of traditional HLA group nomenclature, as alleles within the same serologic group can differ substantially at the eplet level, while alleles from different groups may share critical epitopes.

Imputation Accuracy and Population-Specific Limitations

A key limitation of eplet mismatch analysis in clinical practice is its dependence on high-resolution HLA genotyping, which is often not feasible in deceased donor transplantation because of time constraints. To address this, allele-level genotypes are frequently imputed using population-specific haplotype frequency data. However, concerns remain about the accuracy of imputed genotypes, particularly in non-European populations where limited population-specific HLA data increases the likelihood of imputation errors [69,70]. Despite these concerns, recent studies support the clinical utility of genotype imputation. Cohen et al. [71] demonstrated that imputation preserved mismatch risk classification in over 90% of both racially concordant and discordant donor-recipient pairs, with 97.1% concordance in a real-world validation cohort. Similarly, Mangiola et al. [72] reported that common eplets are consistently expressed across six global population groups, reinforcing the feasibility of imputation in diverse clinical contexts.

Nevertheless, accurate risk stratification still requires expanding ethnicity-specific HLA databases and refining imputation algorithms. Senev et al. [73] showed that 23.3% of DSA assignments were inaccurate when based on imputed genotypes, highlighting the need for direct high-resolution typing whenever possible. Matern et al. [74] further demonstrated that ancestry misclassification during imputation increases mismatch uncertainty, particularly in populations such as Alaska Native/American Indian. Their findings suggest that multiple imputation combined with ancestry-informed models can improve accuracy but cannot fully substitute for high-resolution data [74]. Thus, while genotype imputation facilitates timely use of eplet analysis—especially in deceased donor transplantation—its accuracy remains population dependent, underscoring the importance of expanding high-resolution typing and improving imputation models.

Integration of Eplet Mismatch Into Clinical Practice and Allocation Systems

The potential of eplet mismatch analysis to improve long-term graft survival has fueled growing interest in its incorporation into kidney allocation policies [59,75,76]. Some living donor programs, including the National Kidney Registry in the United States and the Royal Children’s Hospital Melbourne transplant program, have already integrated eplet mismatch considerations into their allocation algorithms [77,78].

At their 2022 consensus meeting, the STAR Working Group emphasized the need to move beyond traditional antigen-level matching. They recommended incorporating molecular mismatch analysis—including eplet-based and physicochemical scoring—into pre- and posttransplant risk assessment strategies. The group highlighted that molecular mismatch provides greater granularity in predicting alloimmune responses, but also stressed the importance of standardized methodologies, validation across diverse populations, and careful consideration of equity implications before applying these tools in organ allocation systems [28]. These recommendations reflect a growing consensus that molecular mismatch–based tools, including eplet mismatch, can improve both immunologic precision and fairness in transplantation—provided their implementation is supported by robust evidence.

Standardization and Validation Requirements

Although clinical interest in eplet mismatch analysis is rapidly expanding, challenges remain in achieving consistent standardization and validation. Notably, eplet mismatch scores may vary depending on the computational platform used, as well as differences in allele databases and software versions. Tassone et al. [30] compared eplet assessments using HLAMatchmaker (ver. 2.1) and One Lambda Fusion MatchMaker (ver. 4.2) with identical donor-recipient pairs. They reported significant differences in total mismatch counts, the list of mismatched eplets, and antibody verification status. In some cases, incorrect assignment of eplets to specific HLA alleles was also observed. Such discrepancies complicate study comparability and reduce the reliability of results in clinical practice. Furthermore, because eplet databases continue to evolve as new alleles are characterized, consistent version tracking is critical to ensure reproducibility across studies and platforms. In this regard, recent efforts to evaluate version consistency in computational immunogenetics tools are encouraging. For example, an analysis of PIRCHE algorithm updates demonstrated that major version changes (from ver. 3 to ver. 4) produced consistent T cell epitope mismatch scores in both solid organ and stem cell transplantation modules. These findings suggest that, with appropriate validation and version control, stable scoring can be maintained even as algorithms evolve [79].

FUTURE DIRECTIONS AND RESEARCH PRIORITIES

Future research should aim to optimize the clinical utility of eplet mismatch analysis by addressing current limitations and identifying areas for refinement. Although the single-molecule mismatch cutoff proposed by Wiebe et al. [26] has been widely applied in subsequent studies, population-specific thresholds require validation across diverse ethnic groups. Large-scale, population-specific studies are needed to establish clinically relevant mismatch cutoffs, particularly for underrepresented populations disproportionately affected by disparities in HLA allele distribution. Emerging evidence highlights the contribution of non-HLA immunogenetic variants to graft outcomes [21]. Future compatibility algorithms should therefore incorporate polygenic risk scores that integrate both HLA eplet mismatches and non-HLA immunogenetic factors. Prospective clinical trials are also necessary to evaluate eplet-guided immunosuppression reduction strategies, particularly in retransplant candidates who remain at heightened risk. Widespread adoption of eplet mismatch analysis in clinical practice will depend on the availability of cost-effective and rapid second-field HLA typing, alongside continued investment in robust imputation algorithms and centralized eplet databases. These developments will be particularly important to enable real-time compatibility assessment in deceased donor transplantation, where time constraints often limit access to high-resolution typing [71].

CONCLUSION

Eplet mismatch analysis represents a significant advance in transplant medicine by providing a more precise assessment of donor-recipient compatibility than traditional antigen-level matching. Extensive evidence demonstrates that high HLA-DQ eplet mismatches are strongly associated with adverse outcomes, including dnDSA development, AMR, and graft failure. The risk increases with higher mismatch load; however, no universally safe threshold has yet been defined. Importantly, recent findings emphasize that not all eplet mismatches carry equal immunologic risk, as certain high-risk eplets and repeated mismatches exert a disproportionately greater effect on graft outcomes. This underscores the importance of individualized risk stratification. By integrating eplet mismatch scores with other clinical and immunologic parameters, clinicians can more effectively tailor immunosuppressive regimens, particularly in sensitized and retransplant candidates. Moreover, eplet analysis offers the potential to reduce racial disparities in transplantation by providing more equitable access to well-matched organs for patients from diverse ethnic backgrounds. As eplet databases continue to expand and computational algorithms become more robust, molecular-level HLA compatibility assessment will evolve into a cornerstone of precision transplant medicine, guiding both clinical management and equitable organ allocation.