Early progression of chronic histologic lesions in kidney transplant biopsies is not associated with HLA histocompatibility

ABSTRACT

Background

Early progression of chronic histologic lesions in kidney allografts represents the main finding in graft attrition. The objective of this retrospective cohort study was to elucidate whether HLA histocompatibility is associated with progression of chronic histologic lesions in the first year post-transplant. Established associations of de novo donor-specific antibody (dnDSA) formation with HLA mismatch and microvascular inflammation (MVI) were calculated to allow for comparability with other study cohorts.

Methods

We included 117 adult kidney transplant recipients, transplanted between 2016 and 2020 from predominantly deceased donors, who had surveillance biopsies at 3 and 12 months. Histologic lesion scores were assessed according to the Banff classification. HLA mismatch scores [i.e. eplet, predicted indirectly recognizable HLA-epitopes algorithm (PIRCHE-II), HLA epitope mismatch algorithm (HLA-EMMA), HLA whole antigen A/B/DR] were calculated for all transplant pairs. Formation of dnDSAs was quantified by single antigen beads.

Results

More than one-third of patients exhibited a progression of chronic lesion scores by at least one Banff grade in tubular atrophy (ct), interstitial fibrosis (ci), arteriolar hyalinosis (ah) and inflammation in the area of interstitial fibrosis and tubular atrophy (i-IFTA) from the 3- to the 12-month biopsy. Multivariable proportional odds logistic regression models revealed no association of HLA mismatch scores with progression of histologic lesions, except for ah and especially HLA-EMMA DRB1 [odds ratio (OR) = 1.10, 95% confidence interval (CI) 1.03–1.18]. Furthermore, the established associations of dnDSA formation with HLA mismatch and MVI (OR = 5.31, 95% CI 1.19–22.57) could be confirmed in our cohort.

Conclusions

These data support the association of HLA mismatch and alloimmune response, while suggesting that other factors contribute to early progression of chronic histologic lesions.

Graphical Abstract

Video

Watch the video of this contribution at https://academic.oup.com/ndt/pages/author_videos

alloimmunity, chronic histologic lesion scores, HLA mismatch, kidney biopsy, kidney transplant

KEY LEARNING POINTS

What was known:

• •

  Early progression of chronic histologic lesions in kidney allografts represents the main risk factor of graft attrition.

• •

  HLA epitope mismatch scores are associated with development of de novo donor-specific antibodies (dnDSA) and reduced allograft survival.

• •

  However, the association of HLA mismatch scores such as eplet, predicted indirectly recognizable HLA-epitopes algorithm (PIRCHE-II), HLA epitope mismatch algorithm (HLA-EMMA) and the progression of chronic histologic lesions in the first year post-transplant remains unknown.

This study adds:

• •

  Our study included 117 adult kidney transplant recipients, who underwent surveillance biopsies at 3 and 12 months, dnDSA measurement and calculation of four different HLA mismatch scores, obtained through high-resolution HLA genotyping of donor and recipients by next-generation sequencing.

• •

  More than one-third of patients exhibited a progression of chronic lesion scores by at least one Banff grade in tubular atrophy, interstitial fibrosis, arteriolar hyalinosis and inflammation in the area of interstitial fibrosis and tubular atrophy from the 3- to the 12-month biopsy.

• •

  The main finding of our study was that the degree of HLA mismatch does not contribute to early progression of chronic histologic lesions in the transplant kidney.

Potential impact:

• •

  Our data support the association of HLA mismatch and alloimmune response, while suggesting that other factors contribute to early progression of chronic histologic lesions.

• •

  This is especially important to avoid overimmunosuppression and elucidate other independent factors for the progression of chronic lesions, such as traditional cardiovascular risk factors, like diabetes mellitus and arterial hypertension.

INTRODUCTION

Kidney transplantation is a successful procedure that restores quality of life in people with irreversible kidney failure relying on dialysis. The unsolved problem is the continuous transplant attrition at roughly four % per year [1]. Causes of this ongoing process are multifactorial but certainly include allo- and autoimmune reactions and classical cardiovascular risk factors. All these factors can potentially aggravate chronic histologic lesions in the vascular, interstitial, tubular and glomerular compartments, as detailed in the Banff classification [2].

Wiebe and Nickerson recently reviewed the evidence for HLA compatibility determined by the eplet score and its association with outcomes such as de novo donor-specific antibody (dnDSA) formation, antibody-mediated rejection (ABMR) and graft loss [3]. They showed a significant association of HLA-mismatch and immune-driven events, such as dnDSA appearance or ABMR. However, as stated in this review, estimation of the explained variability in graft loss by HLA mismatch remains elusive and no studies have elucidated the association of HLA-mismatch with progression of chronic histologic lesions as a main contributor to graft loss in sequential biopsies.

Moreover, Zahran et al. recently reported that the mismatch score of high risk eplets was associated with death-censored graft loss only in the higher mismatch range [4]. Accordingly, Senev et al. showed that eplet mismatches in HLA-DQ confer a risk of dnDSA formation, biopsy-proven acute rejection (BPAR) and graft failure [5].

There is evidence that the number of chronic lesions in donor kidney biopsies is among the best predictors of estimated glomerular filtration rate (eGFR) and graft patency [6]. Progression of these chronic lesions in the first year is ubiquitously observed, even in donor kidneys of excellent quality and low HLA mismatch. A progressive increase of chronic lesions is already present in calcineurin inhibitor (CNI)-treated patients in the first year after simultaneous kidney and pancreas transplantation [7, 8]. Causes of progression of chronic lesions as predictors of graft survival remain unclear, but CNI toxicity has been identified as contributor. However, lesions attributed to the use of CNI immunosuppression are unspecific and could also potentially be driven by other processes, such as diabetes mellitus, arterial hypertension, medication or HLA-incompatibility.

Therefore, there is a clear need to determine the contribution of HLA-mismatches and thus -incompatibilities between kidney donor and recipient to the early progression of chronic histologic lesions in stable patients with standard immunologic risk. We thus sought to elucidate this specific question in a well-defined cohort of patients with sequential surveillance biopsies and high-resolution HLA-typing of these recipients and their donors.

MATERIALS AND METHODS

Patients

In this retrospective cohort study, we studied 117 stable kidney allograft recipients transplanted at the Medical University of Vienna between 2016 and 2020, with CNI-based triple immunosuppression, who underwent kidney biopsies at 3 months (t3) and 12 months (t12) post-transplant, in accordance with post-transplant clinical protocols of our department [Strengthening the Reporting of Observational studies in Epidemiology (STROBE) chart Fig. 1]. Target tacrolimus trough levels in the first year post-transplant were 7–10 ng/mL in case of absence of DSA and 10–15 ng/mL in case of preformed DSAs. Deceased donor-to-recipient allocation was performed by the Eurotransplant Kidney Advisory Committee (ETKAC) algorithm in Eurotransplant. High-resolution HLA genotyping of donor and recipients was performed as part of the clinical routine by next-generation sequencing (NGS) for the genes of HLA-A, B, C, DRB1, DRB3/4/5, DQA1 and DQB1 [9]. eGFR was calculated according to the 2021 Chronic Kidney Disease Epidemiology Collaboration equation without race [10]. The study was approved by the Ethics Committee of the Medical University of Vienna (EK-Nr. 267/2011 renewed annually) and was conducted in accordance with the Declaration of Helsinki.

Figure 1:

STROBE flow chart of the study population. Adult kidney transplant recipients from the prospective Vienna transplant cohort study, who had a surveillance biopsy at 3 and 12 months between 2016 and 2020 were included.

Calculation of the HLA-mismatch scores and DSA measurements

We calculated four different HLA mismatch scores for each donor and recipient pair, i.e. HLA epitope mismatch algorithm (HLA-EMMA), predicted indirectly recognizable HLA-epitopes algorithm (PIRCHE-II) and eplet scores, as well as six antigen HLA mismatches (HLA whole-antigen mismatch). PIRCHE-II and HLA-EMMA scores were calculated using the PIRCHE-II prediction server (application version: v3.3.42; database version: 3.44) and HLA-EMMA program (version 1.05), respectively. eplet mismatch scores were computed using HLA-Matchmaker (ABC version 4; DRDQDP version 3.1) provided by Duquesnoy and Askar [11]. The sum score of the six HLA antigen mismatches (A/B/DR) was calculated from the Eurotransplant database.

DSA measurements were performed as part of standard patient care by Luminex One Lambda [12] in the first year post-transplant, at the time of surveillance biopsies, i.e. at t3 and t12. According to the local laboratory's cut-off values, a mean fluorescence intensity >1000 was regarded as positive DSA. Furthermore, DSAs appearing within 30 days post-transplant were not considered to be dnDSAs, but rather preformed DSAs that were already present pre-transplant, as suggested by previous studies [13]. Moreover, in patients with prior immunization, DSA appearance against another locus than pre-transplant DSAs was mandatory for dnDSA diagnosis.

Histopathology and Banff scores

Progression of histologic lesions was assessed in the paired 117 kidney biopsies between three months (t3) and twelve months (t12) post-transplant. Percutaneous renal allograft biopsies were obtained under real-time ultrasound guidance with a 16-gauge needle. All biopsies were analyzed by three experienced renal transplant pathologists (H.R., N.K., J.K.).

Three months was chosen as baseline to avoid confounding by early post-transplant events. Perioperative donor biopsies at t0 were thus not included in our analyses of chronic lesions, as early histologic changes between t0 and t3 are rather due to acute rejection or medical and surgical problems. Furthermore, we refrained from including them into our analyses given that no standardized Banff scoring has been validated for perioperative donor biopsies.

Histopathologic workup of kidney biopsies was performed applying standard methodology on several consecutive slides stained with hematoxylin–eosin, periodic acid–Schiff, Methenamin-Silver and Acid Fuchsin-Orange, according to the recommendation of the Banff classification meeting report [14]. Surveillance biopsies were scored at t3 and t12 according to the Banff classification of renal allograft pathology valid at the time point of biopsy. However, due to several changes to this classification during the study period, inflammation in the area of interstitial fibrosis and tubular atrophy (i-IFTA) lesions were re-evaluated in cases with interstitial fibrosis (ci > 0) and cases without interstitial fibrosis (ci = 0) were set as i-IFTA = 0. The Banff lesion scores range from 0 to 3 in all compartments with the exception of microvascular inflammation [MVI, i.e. glomerulitis (g) + peritubular capillaritis (ptc)], which ranges from 0 to 6. Scoring for i-IFTA also ranged from 0 to 3, with 0 in case of absence or <10%, 1 in case of 10%–25%, 2 in case of 26%–50% and 3 in case of >50%. Glomerular basement membrane double contours (cg) were further assessed in light microscopy. Both de novo occurrence and aggravation of a histologic score were considered as progression of the respective histologic lesion in this study. Borderline readings were defined as interstitial inflammation (i) = 1 and tubulitis (t) ≥ 1 or i ≥ 1 and t = 1 without intimal arteritis (v) according to Banff 2019 [15]. In case of rejection on allograft biopsy with preceding decline of kidney function (i.e. creatinine increase), treatment consisted of a 3-day intravenous glucocorticoid regimen, followed by oral prednisone taper. Additionally, in Banff IIb rejections (i.e. n = 1), rabbit antithymocyte globulin was also administered.

Statistical analysis

Continuous variables were expressed as means ± standard deviations if normally distributed and medians and interquartile ranges (Q1–Q3) otherwise. Violation of normality was assessed through visual inspection of histograms and boxplots. Categorical variables were presented as numbers and percentages. Fractions of missing data were examined and reported for each variable. A complete case analysis was carried out for each of the models.

Multivariable proportional odds logistic regression (POLR) was used in order to assess the association between progression of histologic lesion scores between t3 and t12 post-transplant [Δ(t12, t3)] as the dependent variable and different HLA molecular mismatch scores, i.e. eplet, PIRCHE-II, HLA-EMMA, HLA whole-antigen mismatch, as the independent variable. We used baseline adjustment for the respective histologic lesion score present at t3 to avoid regression to the mean. The underlying proportional odds assumption of the model implies that the effect of the independent variable remains constant at all levels of [Δ(t12, t3)] and this was checked by the Brant–Wald test.

Multivariable logistic regression (LR) models were performed to determine the association between dnDSA formation or rejection as the dependent variables and each molecular mismatch score, adjusted for prior immunization (i.e. pre-transplant DSAs) as the independent variables. Results of the models were reported as adjusted odds ratio (OR), 95% confidence interval (CI), P-value and McFadden's Pseudo R2. Furthermore, multivariable Cox proportional hazard models (CPH) were used to estimate unadjusted hazard ratios and 95% CIs of time to dnDSA formation and rejection and each HLA mismatch score as the independent variables. POLR was further performed to determine the latter association. Finally, Spearman correlation coefficients were computed in order to quantify non-linear associations between pairs of histocompatibility scores.

All statistical analyses were performed using the software R Project for Statistical Computing version 4.2.0. Two-sided P-values <.05 were considered statistically significant, but were only of exploratory character. We therefore refrained from multiplicity adjustments.

RESULTS

Baseline characteristics of the study population (i.e. N = 117) were similar to those of the remaining, excluded kidney transplant recipients between 2016 and 2020 (i.e. N = 401) (Supplementary data, Table S1). Characteristics of the study cohort are depicted in Table 1. Donors and recipients were predominantly Caucasian and most kidney allografts came from deceased donors (86%). Forty-nine percent of donors and 33% of recipients were females and both donor and recipient groups were of similar age (54 ± 5 vs 54 ± 12 years). eGFR at baseline (t3) was 42 ± 16 mL/min/1.73 m² and remained stable throughout the study to t12 at 42 ± 17 mL/min/1.73 m². Patients had CNI-based triple immunosuppression, with a mean tacrolimus trough level during the study period of 8 ± 2 ng/mL. Fifteen percent of patients showed dnDSA formation, with a median time to dnDSA detection of 242 (103, 364) days. Most patients developed HLA class II dnDSAs (10%), especially against DQB1 (5%). dnDSA formation against specific loci is depicted in Supplementary data, Table S2. Due to high variability in patient compliance, appointment availability and the occasional necessity of indication biopsy, t3 was on average 102 ± 31 and t12 391 ± 52 days post-transplant. Detailed histologic lesion scores at 3 and 12 months post-transplant, with corresponding tacrolimus levels at time of biopsy, are shown in Supplementary data, Table S3. Eighteen percent exhibited a BPAR, which included 10% of T-cell mediated rejection (TCMR) and 8% of ABMR in surveillance biopsies. One patient presented with mixed rejection (i.e. TCMR and ABMR) and was thus counted in both categories. Moreover, recurrence of disease was seen in two patients with membranous glomerulonephritis and pauci-immune crescentic glomerulonephritis respectively, within 3 months post-transplant. One patient also developed de novo glomerulonephritis within 3 months. Furthermore, two patients presented with polyomavirus nephropathy within 3 months post-transplant and allograft pyelonephritis was seen in one patient.

Table 1:

Baseline demographic and clinical characteristics of the study population.

Variables	Study population (N = 117)	Not available
Donor and recipient characteristics
Donor age (years)	54.1 ± 14.6	0
Donor sex (female)	57 (48.7)	0
Donor type (living)	17 (14.5)	0
KDRI	1.2 ± 0.4	2
Recipient age (years)	54.0 ± 12.0	0
Recipient sex (female)	39 (33.3)	0
Diabetes in recipients	17 (14.5)	0
First kidney transplant recipients	101 (86.3)	0
Recurrent kidney transplant recipientsb	16 (13.7)	0
Latest PRAs (%)	0 (0, 0)a	0
Donor cause of ESKD
Immune-mediated nephropathy	44 (37.6)	0
ADPKD	20 (17.1)	0
Diabetic nephropathyc	11 (9.4)	0
Vascular nephropathyc	7 (6.0)	0
Else	36 (30.8)	0
Laboratory measurements
eGFR at t3 (mL/min/1.73 m2)	42 ± 16	0
eGFR at t12 (mL/min/1.73 mc)	42 ± 17	2
Mean tacrolimus trough levels 1–6 months post-Tx (ng/mL)d	8.2 ± 1.5	3
Tacrolimus trough levels 1 month post-Tx (ng/mL)	9.0 ± 1.9	6
Tacrolimus trough levels 2 months post-Tx (ng/mL)	9.2 ± 2.4	5
Tacrolimus trough levels 3 months post-Tx (ng/mL)	8.6 ± 2.7	8
Tacrolimus trough levels 4 months post-Tx (ng/mL)	8.0 ± 2.8	5
Tacrolimus trough levels 5 months post-Tx (ng/mL)	7.5 ± 2.1	9
Tacrolimus trough levels 6 months post-Tx (ng/mL)	7.2 ± 1.8	16
Pre- and post-transplant characteristics
Mean follow-up time post-Tx (days)	391 ± 52	0
eplet total	32.7 ± 16.8	0
eplet class I	14.5 ± 7.3	0
eplet class II	18.2 ± 12.5	0
eplet DRB1	9.9 ± 7.6	0
eplet DQB1	6.0 ± 5.4	0
EMMA total	51.9 ± 27.6	0
EMMA class I	21.8 ± 10.5	0
EMMA class II	30.0 ± 21.7	0
EMMA DRB1	7.0 ± 5.8	0
EMMA DQB1	4.6 ± 7.0	0
PIRCHE-II	72.3 ± 46.5	0
PIRCHE DRB1	11.5 ± 9.2	0
PIRCHE DQB1	18.7 ± 17.2	0
HLA-A/-B/-DR mismatch	3.0 ± 1.4	0
HLA-A/-B mismatch	2.0 ± 1.0	0
HLA-DR mismatch	1.0 ± 0.6	0
Preformed DSA	22 (18.8)	0
dnDSA formation post-TX	18 (15.4)	0
dnDSA class II	12 (10.3)	0
dnDSA DRB1	0 (0.0)	0
dnDSA DQB1	6 (5.1)	0
Time to dnDSA detection (days)	242 (103, 364)a	99
Rejection in any biopsy (surveillance or indication)
Borderline	24 (20.5)	0
TCMRe	12 (10.3)	0
ABMRe	9 (7.7)	0
Time to rejection (days)	106 (18, 358)a	73

Values are presented as means ± standard deviations, numbers and percentages (%) or

^amedians and interquartile ranges.

b

Two to four kidney transplant recipients.

c

One patient had both a diabetic and vascular nephropathy and was thus counted in both categories.

d

Patients had CNI-based triple immunosuppression (tacrolimus, mycophenolate mofetil, prednisolone). In two cases, tacrolimus had to be switched to ciclosporine due to tremor and two patients received belatacept, sirolimus and prednisolone.

e

One patient presented with mixed rejection (i.e. ABMR and TCMR) and was thus counted in both categories.

KDRI: kidney donor risk index; ESKD: end-stage kidney disease; ADPKD: autosomal dominant polycystic kidney disease; Tx: kidney transplant.

More than one-third of patients exhibited a progression of chronic lesion scores by at least one Banff grade in tubular atrophy (ct), ci, arteriolar hyalinosis (ah) and i-IFTA from the 3- to the 12-month biopsy. Highest progression was observed for ct in 46.2% of cases, followed by ci with 45.3%, i-IFTA with 37.6% and ah with 33.3%. Total inflammation (ti) was observed in 29.9% and vascular fibrous intimal thickening (cv) in 27.4% (Table 2, Fig. 2). Progression of acute lesions, i.e. g, ptc, i and v, as well as of mesangial matrix expansion (mm), was rare and is not discussed further. The chronic lesion score cg was also excluded from further analyses due to a lack of progression events (Fig. 2).

Figure 2:

Distribution of histologic lesion scores according to time of kidney biopsy according to the Banff classification [2]. Bar charts represent the progression of the different histologic lesion scores between months 3 and 12 after transplantation. It is of note that MVI (g + ptc) ranges from 0 to 5, all others from 0 to 3. The bar NA indicates the number of non-available values.

Table 2:

Distribution of chronic histologic lesion score progression and regression in kidney biopsies from Month 3 to 12 after transplantation according to the Banff classification^a.

		Regression levels				Progression levels
Histologic lesion scores	N	–3	–2	–1	0	1	2	3	NA	Progression events	Regression events
Chronic histologic lesion scores
Δct	117	0	0	11	51	35	14	5	1	54	11
Δci	117	0	3	13	48	25	21	7	0	53	16
Δah	117	1	4	13	59	32	7	0	1	39	18
Δti	117	0	3	8	69	25	6	4	2	35	11
Δcv	117	2	10	16	45	20	10	2	12b	32	28
Δcg	117	0	0	0	112	2	0	1	2	3	0
Δi-IFTA	117	2	1	7	60	20	10	14	3	44	10
Δmm	117	0	0	0	110	4	0	0	3	4	0
Acute histologic lesion scores
Δg	117	1	0	3	105	7	0	0	1	7	4
Δptc	117	1	0	6	101	3	2	0	4	5	7
Δ(g + ptc)	117	2	0	8	93	8	1	0	5	9	10
Δi	117	1	3	6	94	11	1	0	1	12	10
Δt	117	3	7	13	61	16	11	6	0	33	23
Δv	117	0	0	0	108	0	1	0	8	1	0

a

Acute histologic lesion scores are also shown for completeness.

b

Missing values of cv are due to the absence of arteries in the examined biopsy.

NA: not available.

The HLA-mismatch scores eplet, PIRCHE-II, HLA-EMMA and HLA whole-antigen mismatch are detailed in Table 1. Mean mismatch loads were 33 ± 17 for total eplet, 72 ± 47 for PIRCHE-II, 52 ± 28 for HLA-EMMA and 3 ± 1 for HLA-A/-B/-DR. Positive pairwise correlations were observed between the different mismatch scores (Supplementary data, Fig. S1), the strongest being between class II HLA-EMMA and total HLA-EMMA mismatch (ρ = 0.94, P < .001) and the weakest being between PIRCHE-II DRB1 and class I HLA-EMMA mismatch (ρ = 0.20, P = .04). In addition, eplet and HLA-EMMA mismatch also showed high positive correlations for all of their subscores.

Multivariable POLR with progression of histologic lesion scores between t3 and t12 as the dependent variable and different HLA molecular mismatch scores, i.e. eplet, PIRCHE-II, HLA-EMMA, HLA whole-antigen mismatch, as the independent variable, revealed no association of progression of chronic histologic lesions with any HLA mismatch score, except for ah, which showed a weak relationship with HLA-EMMA and eplet (Fig. 3). In more detail, ah was associated with HLA-EMMA DRB1 (OR = 1.10, 95% CI 1.03–1.18, P = .01), eplet DRB1 (OR = 1.07, 95% CI 1.01–1.12, P = .02), class I HLA-EMMA (OR = 1.06, 95% CI 1.01–1.10, P = .01) and total HLA-EMMA (OR = 1.02, 95% CI 1.00–1.04, P = .02).

Figure 3:

Forest plots of the ORs and 95% CIs of multivariable proportional odds logistic regression models with progression of histologic lesion scores between t3 and t12 as outcome variables and mismatch scores as independent variables, adjusted for the respective histologic lesion score at t3. A triangle denotes a statistically significant adjusted association between progression of the histologic lesion score between t3 and t12 and the respective mismatch score. The vertical dashed line indicates an OR of 1.

Moreover, after adjusting for presence of pre-transplant DSAs, multivariable LR with dnDSA formation post-transplant as dependent variable and mismatch scores as independent variable revealed an association of different HLA mismatch scores with dnDSA development. In fact, HLA-EMMA DQB1 was associated with class II dnDSA (OR = 1.12, 95% CI 1.05–1.22, P = .01) and dnDSA DQB1 formation (OR = 1.13, 95% CI 1.03–1.26, P = .02) and PIRCHE-II DQB1 with class II dnDSA (OR = 1.04, 95% CI 1.01–1.08, P = .01) and dnDSA DQB1 development (OR = 1.06, 95% CI 1.02–1.11, P = .01, Fig. 4). Finally, there was also an association between eplet DQB1 and dnDSA DQB1 formation (OR = 1.17, 95% CI 1.01–1.38, P = .05, Fig. 4). However, the different mismatch scores were not associated with time to dnDSA formation, except for HLA-EMMA DQB1 (OR = 1.06, 95% CI 1.01–1.12, P = .03, Supplementary data, Fig. S2). Besides, dnDSA development was associated with progression of non-chronic lesions suggestive of humoral alloimmunity, i.e. MVI (OR = 5.31, 95% CI 1.19–22.57, P = .004, Supplementary data, Fig. S3).

Figure 4:

Forest plots of the ORs and 95% CIs of multivariable logistic regression models with dnDSA formation post-transplant as outcome variable and mismatch scores as independent variables, adjusted for presence of pre-transplant DSAs. A triangle denotes a statistically significant adjusted association between dnDSA formation (i.e. dnDSA, class II dnDSA, dnDSA DQB1) and the respective mismatch score. The vertical dashed line indicates an OR of 1.

DISCUSSION

The present study showed in a series of sequential surveillance biopsies of a well-defined cohort of mainly deceased donor kidney transplant recipients that HLA mismatch did not contribute to early progression of chronic lesions. However, as previously reported in the literature, higher molecular mismatch load, especially in DQB1, was found to be associated with class II dnDSA formation, which in return was linked to progression of MVI.

In fact, unfavorable early histologic changes seem to have a negative impact on graft and long-term survival, which underlines their impact on long-term outcomes [16]. So far, it has been shown that HLA incompatible transplantations exhibit a higher risk of active ABMR and premature graft loss [17]. Accordingly, Lemieux et al. found an association of HLA-mismatch score and graft loss, but only in the higher mismatch range. Therefore, it is likely that the overall mismatch load, rather than the few highly immunogenic mismatches, drives the development of graft loss [18]. Moreover, higher eplet mismatch load has been associated with dnDSA formation, ABMR, TCMR, transplant glomerulopathy and allograft failure [5, 19, 20].

Research in recent years has revealed inflammation as the main cause driving progression of chronic allograft lesions. However, early rejection episodes occurring within 3 months post-transplant were found to have no or only a weak impact on the subsequent development of chronic lesions if treated adequately [21]. In fact, a group of patients exhibiting chronic stable pathology at 1 year post-transplant, but no subclinical active inflammation in the follow-up biopsies, demonstrated similar long-term graft function to those with normal histology. Inflammation thus only seems to have the greatest effect if it persists and therefore has enough time to cause tissue remodeling and fibrosis. The noted progression of chronic histologic lesions in this study may thus be explained by other factors, e.g. arterial hypertension, diabetes, atherosclerosis, CNI levels or ischemia–reperfusion injury, as suggested by Nankivell et al. [7, 8]. In fact, the observed association between ah and eplet and HLA-EMMA mismatch scores, respectively, may be a sign of potentially higher immunosuppression in patients with higher immunologic risk, as ah lesions are the result of CNI treatment [8, 22].

Moreover, the established associations of dnDSA development, especially of class II, with HLA mismatch and progression of MVI, the defining histopathologic feature of active ABMR, were also found in this study, thus allowing for comparability of results with other study cohorts [23].

As previously reported, HLA matching at the molecular level was superior to allelic antigen mismatch regarding prediction of dnDSA formation, given that only PIRCHE-II DQB1, HLA-EMMA DQB1 and eplet DQB1 were associated with dnDSA DQB1 development in this study [24]. In fact, the large heterogeneity of mismatched antigens cannot be reflected by whole-antigen mismatch. Class II eplet mismatch and PIRCHE scores have been associated with class II dnDSA formation in former studies [25]. A strong relationship between higher mismatch scores and dnDSA development has also been shown by Wiebe et al. [20], who showed that alloimmune risk group stratification (low, intermediate, high) according to eplet mismatch load correlated with dnDSA formation. Wiebe et al. further reported an OR of 1.04 per eplet HLA-DQ epitope mismatch for HLA-DQ dnDSA formation, which is in accordance with our results [26].

However, our study has intrinsic limitations. We used established scores as a measure of HLA-mismatches, which all rely on amino acid or eplet mismatches. Nevertheless, it is of note that only 44 of 224 HLA class I eplets and 19 of 288 HLA class II eplets have been experimentally verified by monoclonal antibodies [27]. Furthermore, all eplet mismatches are considered equally immunogenic in these scores, which is certainly incorrect. Furthermore, sampling error, especially of focally distributed lesions such as i-IFTA, is an intrinsic limitation of all biopsy studies. Moreover, the categorical Banff scoring system itself represents a limitation, as it reflects an estimation based on percentages. However, interobserver variability of trained renal transplant pathologists, especially in regards to i-IFTA and t-IFTA, has been shown to exhibit kappa values of agreement from 0.5 to 0.6 and can thus be interpreted as moderate to good [28]. It is of note that we included 51 patients featuring one histologic lesion score of 3 at baseline, as those cannot show progression at t12 anymore. However, they were still included in this study, as their other histologic lesion scores were below 3 with consequently the possibility of progression. Our study exhibits limit external validity as our patient population mainly consists of Caucasian donors and recipients with a subsequently higher genetic compatibility. Furthermore, donor organs from Eurotransplant are preferably allocated on the basis of HLA-A, -B and -DR match between donor and recipient, thus avoiding severe mismatches.

Nevertheless, the narrower population ethnicity may be a strength, as people seldom relocate in Austria. There is a mandatory resident address registration, thus follow-up and biopsy performance is high. Besides, this study was conducted in recent eras (2016–20) with CNI-based triple immunosuppression and therefore provides representative up-to-date data that are comparable with other centers. As in most HLA laboratories of transplant centers and organ procurement organizations, donor and recipient HLA typing was performed at high-resolution using standardized NGS. Furthermore, the sample size of 117 timed sequential transplant biopsies, analyzed by well-established kidney pathologists, with complete adjacent data guarantees high data quality and validity of results. An additional strength is the comprehensive statistical analysis that includes baseline adjustment for the corresponding histologic lesion score at t3, preventing overestimation of effects and thereby avoiding biased or misleading results. However, given the limited sample size an undetectable small effect size cannot be ruled out. Thus, the present study does not represent a definitive answer to the question of HLA histocompatibility not being associated with progression of chronic histologic lesions in the first year post-transplant. Prospective studies with a larger sample size are therefore needed to address this issue and confirm our results.

ACKNOWLEDGEMENTS

We are indebted to the staff of our transplant center who supported the study with their expertise and were of tremendous administrative help in obtaining, labeling and biobanking biopsies and plasma samples from the many patients.

FUNDING

Funding was obtained through the Vienna Science and Technology Fund (WWTF Genomics based immunologic risk stratification in kidney transplantation #10.47379/LS20081).

AUTHORS’ CONTRIBUTIONS

R.O., A.H. and R.R.-S. were responsible for conception, design, financial support, critical revision and final approval of the manuscript. R.O. and R.J. were responsible for manuscript writing, design, critical revision and final approval of the manuscript. A.H. and M.G.G. were responsible for statistical data analysis and revision of the manuscript. H.R., N.K. and J.K. performed kidney biopsy analyses and critical revision of the manuscript. G.F., A.K., J.U.B. and C.W. were responsible for critical revision of the manuscript. All authors contributed to the article and approved the submitted version.

DATA AVAILABILITY STATEMENT

The raw data of this article are available in the online Supplementary data.

CONFLICT OF INTEREST STATEMENT

The authors of this manuscript have no conflicts of interest to disclose as described by Nephrology Dialysis Transplantation.