Received: 9 May 2018 | Revised: 6 July 2018 | Accepted: 29 July 2018

DOI: 10.1111/ajt.15055

ORIGINAL ARTICLE

Value of monitoring circulating donor‐reactive memory B

cells to characterize antibody‐mediated rejection after kidney
transplantation

Sergi Luque1 | Marc Lúcia1 | Edoardo Melilli2 | Carmen Lefaucheur3 | Marta Crespo4 |
3 1,5 6 1 1
Alex Loupy | David Bernal-Casas | Montse Gomà | Marta Jarque | Elena Crespo |
2 2 1,2 2
Núria Montero | Anna Manonelles | Josep M. Cruzado | Salvador Gil‐Vernet |
1,2 1,2
Josep M. Grinyó | Oriol Bestard

1
Experimental Nephrology
Laboratory, IDIBELL, Barcelona, Spain Antibody‐mediated rejection (ABMR) is defined by specific histopathological lesions
2
Kidney Transplant Unit, Nephrology and evidence of circulating donor‐specific antibodies (DSA). Although DSA are not
Department, Bellvitge University Hospital, always detectable, monitoring donor‐reactive memory B cells (mBC) could identify
Barcelona, Spain
3 patients at risk of developing ABMR. Peripheral donor‐reactive mBC using a novel
Paris Translational Research Center for
Organ Transplantation, Institut National de HLA B cell ELISpot assay, serum DSA, and numbers of different B cell subsets were
la Santé et de la Recherche Médicale, UMR‐
assessed in 175 consecutive kidney transplants undergoing either for‐cause or 6‐ and
S970, Paris, France
4
Kidney Transplant Unit, Nephrology
24‐month surveillance biopsies for their association with main histological lesions of
Department, Hospital del Mar, Barcelona, ABMR and impact on allograft outcome. In 85 incident for‐cause biopsies, high fre‐
Spain
5
quencies of donor‐reactive mBC were detected in all 16 (100%) acute ABMR/DSA+
Department of Genetics, Microbiology and
Statistics, Section of Statistics, University of and most chronic ABMR, with or without DSA (24/30[80%] and 21/29[72.4%], re‐
Barcelona, Barcelona, Spain spectively). In a longitudinal cohort of 90 nonsensitized patients, a progressively
6
Pathology Department, Bellvitge University
higher expansion of donor‐reactive mBC than de novo DSA was observed at 6 and 24
Hospital, Barcelona, Spain
months (8.8% vs 7.7% and 15.5% vs 11.1%, respectively) and accurately identified
Correspondence
patients with ongoing subclinical ABMR (area under the curve = 0.917 and area
Oriol Bestard, Kidney Transplant Unit,
Nephrology Department, Bellvitge under the curve = 0.809, respectively). An unsupervised hierarchical cluster analysis
University Hospital, Barcelona University,
revealed a strong association between donor‐reactive mBC with main fundamental
IDIBELL, Barcelona, Spain.
Email: obestard@bellvitgehospital.cat allograft lesions associated with ABMR and conferred a significant deleterious im‐
pact on graft outcome. Monitoring donor‐reactive mBC may be useful to further
Funding information
ISCiii, Grant/Award Number: PI16/01321, characterize humoral rejection after kidney transplantation.
PI13/01263, INT15/00112, PI16/00617,
PI13, /00598 and RD16/0009/0013; KEYWORDS
European Commission; Biomarker‐Driven
Immunosuppression Minimization
alloantibody, biomarker, clinical research/practice, histocompatibility, kidney transplantation/
Consortium (BIODRIM), Grant/Award nephrology, major histocompatibility complex (MHC), monitoring: immune, pathology/
Number: FP7; Catalan Society of histopathology, rejection: antibody‐mediated (ABMR), translational research/science
Transplantation

Abbreviations: ABMR, antibody‐mediated rejection; ASC, antibody‐secreting cells; AUC, area under the curve; CNI, calcineurin‐inhibitor; DSA, donor‐specific antibodies; IFTA, interstitial
fibrosis and tubular atrophy; IVIG, intravenous immunoglobulins; KT, kidney transplantation; mBC, memory B cells; MFI, mean fluorescence intensity; PBMC, peripheral blood mononu‐
clear cells; PRA, panel‐reactive antibodies; TCMR, T cell–mediated rejection; TFH, T‐follicular helper.

Am J Transplant. 2018;1–13. amjtransplant.com

© 2018 The American Society of Transplantation | 1
and the American Society of Transplant Surgeons
2 |
1 | I NTRO D U C TI O N
The better identification of circulating donor‐specific HLA antibod‐
ies (DSA) and the better characterization of specific histopatholog‐
ical lesions have determined that the humoral alloimmune response
is a main cause of kidney graft loss. 1-4
For a precise diagnosis of antibody‐mediated rejection (ABMR),
specific histological lesions within different kidney graft compart‐
ments such as acute microvascular inflammation (glomerulitis and
peritubular capillaritis), 5,6
chronic transplant glomerulopathy, 7,8
and/or C4d positivity at the peritubular capillaries9-11 should be
accompanied by the concomitant serological evidence of DSA to
demonstrate the effector immune mechanism responsible for such
lesions. 12
Therefore, in the absence of detectable DSA despite the
presence of these histological features, the certainty of the diag‐
nosis of ABMR is incomplete, as no histology‐independent biopsy
diagnostic system has been fully validated yet. Indeed, an important
proportion of cases showing characteristic ABMR lesions assessed
at the light microscope, as well as using microarray‐derived molecu‐
lar ABMR scores as a histology‐independent estimate of ABMR, do
not always show circulating anti‐HLA DSA.13-16
In fact, the exclusive assessment of circulating DSA may signifi‐
cantly underestimate the magnitude of the entire humoral immune
response as it does not necessarily include the assessment of the
memory B cell (mBC) pool. Importantly, mBCs provide enhanced im‐
munity during secondary antigen exposure and rapidly differentiate
F I G U R E 1 Flowchart of the study. Tx, transplant; Bx, biopsy; DSA, donor‐specific antibodies; mBC, Memory B cells; HUB, Bellvitge
University Hospital; PTC, Paris Transplant Center; HdM, Hospital del Mar
LUQUE et al.
into antibody‐secreting cells (ASC) and release high‐affinity anti‐
bodies.17-20 Furthermore, because the mBC repertoire seems to be
broader than that of plasma cells, they may be present in the ab‐
sence of serum antibodies. 21,22 Indeed, emerging evidence in both
experimental animal models and humans strongly suggests that
posttransplant DSA predominantly derive from the generation and
reactivation of donor‐reactive mBCs leading to an increased risk of
acute ABMR. 23-27
Here, we hypothesized that the functional assessment of circu‐
lating donor‐reactive mBCs after kidney transplantation (KT) using
a novel HLA‐specific B cell ELISpot assay could refine the antidonor
humoral immune response of patients displaying histological lesions
compatible with ABMR, even in the absence of detectable serum
DSA. Further, the prospective monitoring after KT of both DSA and
donor‐reactive mBC would provide new insight regarding the gener‐
ation of the antidonor humoral immune response.
2 | M ATE R I A L A N D M E TH O DS
2.1 | Patients in the study and study design
Between January 2014 and January 2016, 250 KT patients from three
different transplant programs (Hospital Universitari de Bellvitge, Paris
Transplant Center for Organ Transplantation, and Hospital del Mar)
were enrolled (Figure 1). These patients were stratified in two groups: a
LUQUE et al.
first cross‐sectional group of 100 incident KT undergoing for‐cause allo‐
graft biopsies due to graft dysfunction and a second longitudinal cohort
of 150 consecutive KT patients transplanted undergoing surveillance
allograft biopsies at 6 and 24 months. As depicted in the flowchart of
the study, a total of 85 KT were assessed in the cross‐sectional cohort,
whereas a total of 90 KT patients were evaluated in the longitudinal
group; thus, a total of 175 patients were assessed and followed for at
least 2 years after last histological assessment.
As per protocol in all centers, patients with ABMR were treated with
plasmapheresis and intravenous immunoglobulins (IVIG). Circulating
DSA and donor‐reactive mBCs were monitored at the time of biopsy
assessment in the cross‐sectional cohort as well as at 6 and 24 months
after transplantation in the longitudinal cohort. Additionally, patients
from the cross‐sectional cohort were assessed for different B and T‐cell
subset phenotypes in peripheral blood at the time of biopsy.
All patients included in this project gave a written informed con‐
sent and the study was approved by the institutional review board at
Bellvitge University Hospital.
2.2 | Alloantibody detection and characterization
Alloantibody specificities against both class I and II HLA antigens
were done in banked serum samples and in supernatants of mBC
cultures using single‐antigen flow beads assays on a Luminex plat‐
form (Lifecodes, division of Immucor, Inc, Stamford, CT). All beads
showing a normalized mean fluorescence intensity (MFI) > 500 MFI
were considered positive if [MFI/(MFI lowest bead)] > 5.
2.3 | Renal allograft histology
Renal allograft biopsies used to characterize different histological
phenotypes were all performed as either for cause due to clinical
allograft dysfunction or as per protocol at 6 and 24 months after
transplantation. All biopsies were blindly graded from all clinical and
immunological data by an experienced transplant pathologist and
scored, following the last Banff classification. 28
2.4 | Assessment of circulating HLA‐specific
memory B cell frequencies
2.4.1 | Memory B cell stimulation assay
As thoroughly explained in the supplemental material, peripheral
blood mononuclear cells (PBMC) were first polyclonaly stimulated
in order to obtain an ASC‐like population to be subsequently chal‐
lenged against different HLA multimerized monomers in an ELISpot
platform.
2.4.2 | HLA‐specific and polyclonal IgG B cell
ELISpot assay
HLA‐specific IgG B cell ELISpot assays were performed following re‐
26
cently described standard operating procedures. Briefly, ELISpot
| 3
plates were coated with a mAb against human IgG and blocked to
avoid unspecific interactions. Previously stimulated PBMC were
seeded to ELISpot plates and incubated to release IgG. Polyclonal
IgG‐ASC were detected with antihuman IgG conjugated with alka‐
line phosphatase (Figure S1), and HLA‐sp IgG‐ASC were detected
using fluorescent dye labeled class I and II HLA dextramers.
Because of the limited number of ASC, we developed a multi‐
color HLA‐specific B cell ELISpot assay in order to increase the
number of HLA‐specific determinations in a single ELISpot well si‐
multaneously (Figure S2). A detailed description of the methodology
used is provided as supplemental material.
2.5 | Flow cytometry analysis of main B cell
subsets and T‐follicular helper cells
Different circulating B cell subset populations and T‐follicular helper
(TFH) cells were analyzed in peripheral blood using DuraClone IM B
cell Tube (Beckman Coulter Inc, Brea, CA). A detailed description of
the methodology used is provided as supplemental material.
2.6 | Statistical analysis
All data are presented as mean ± standard deviation (SD). Groups
were compared using the χ2 test for categorical variables, the one‐
way analysis of variance (ANOVA) or t test for normally distributed
data, and the nonparametric Kruskal‐Wallis or Mann‐Whitney U test
for nonnormally distributed variables. Bivariate correlation analyses
were done using Pearson or Spearman tests for nonparametric vari‐
ables. Binary logistic regression analyses were performed to deter‐
mine the independent correlation of variables that were statistically
significant in the univariate analyses and those potentially associ‐
ated with the presence of ABMR. Kaplan‐Meier probabilities of graft
survival were plotted and compared by different histopathological
phenotypes and presence or absence of DSA and/or donor‐­reactive
mBCs using log‐rank tests. Cox’s regression model was used to es‐
timate hazard ratios for univariate analyses for graft survival and to
compare clinical and immunological variables. Receiver operating
characteristic curve analysis was performed to assess the predictive
value of both DSA and donor‐reactive mBCs for the development of
ABMR. Analyses of graft loss were censored for patient’s death. The
two‐tailed statistical significance level used was P < .05. All statisti‐
cal analyses were performed with IBM® SPSS Statistics (version 23)
and GraphPad Prism (version 6.0; GraphPad Software, San Diego,
CA) and the integrated unsupervised heatmap analysis was done
using Matlab®.
3 | R E S U LT S
3.1 | HLA‐specific B cell ELISpot identifies HLA‐
specific mBCs in KT patients
As previously reported, 26 because mBCs do not secrete anti‐
bodies, the HLA‐specific B cell ELISpot assay was developed to
4 | LUQUE et al.

detect circulating HLA‐sp alloreactive mBCs capable of produc‐
ing anticlass I and II HLA antibodies after polyclonal stimulation.
In order to confirm the viability of the stimulated mBCs after
a 6‐day polyclonal stimulation, total IgG‐secreting cells were
counted in the ELISpot platform. As shown in Figure S1, all pa‐
tients of the study showed a broad range of detectable polyclonal
IgG‐ASC. Further, stimulated mBCs of each patient were tested
against medium alone and against a self HLA‐typed molecule to
rule out any unspecific mBC response. Unspecific spots detected
in negative control wells, if any, were subtracted from the result‐
ing HLA‐specific spot count. Because not all ASC are capable of
producing IgG, the ratio between donor‐reactive mBC over the
total polyclonal IgG mBC in each patient was used as the most
reliable approach to characterize the proportion or enhance‐
ment of a given HLA‐specific IgG‐ASC within the global IgG‐ASC
population. To further avoid unspecific HLA‐specific IgG spots, a
minimum of five spots/well, is considered as an acceptable IgG‐
specific mBC response.

TA B L E 1 Baseline clinical and demographic characteristics of the patient population

Main clinical and demographic Cross‐sectional cohort (For Longitudinal cohort

variables All (n = 175) cause Bx) (n = 85) (Surveillance Bx) (n = 90) P

Gender (female, %) 68 (39) 39 (46) 29 (32) .058

Age (y, mean ± SD) 51.2 ± 13.3 49.1 ± 13.7 52.8 ± 12.8 .084
Type of KT (deceased, %) 130 (74) 73 (86) 57 (63) .001
More than 1 previous KT (yes, %) 38 (22) 26 (31) 12 (13) .027
Time on dialysis prior to KT (months) 27.5 ± 37.1 33.8 ± 45.9 22.7 ± 26.9 .081
Type of ESRD (n, %) .229
Polycystic 21 (12) 11 (12) 10 (11)
Vascular 13 (7) 3 (3) 10 (11)
Interstitial 6 (3) 2 (2) 4 (5)
Glomerular 26 (15) 14 (17) 12 (13)
Diabetes 8 (5) 2 (2) 6 (7)
Unknown 101 (58) 53 (61) 48 (53)
DGF (yes, %) 34 (19) 21 (25) 13 (15) .108
Previous AR (yes, %) 20 (11) 13 (15) 7 (8) .042
Type of AR (cellular, %) 12 (60) 7(54) 5 (71) .444
Induction IS (%)
rATG/Basiliximab (%) 63(36)/94(53) 39(46)/28(33) 24(27)/66(73) <.001
Plasmapheresis/IVIG (%) 2(1)/6(3) 2(2)/3(3) 0(0)/3(3) .083
Maintenance IS: CNI‐based (TAC/ 140(80)/24(14) 61(72)/17(20) 79(89)/7(8) .010
CsA) (%)
HLA mismatch
Class I (A + B) (mean ± SD) (range) 2.81 ± 0.8 (1‐4) 2.75 ± 0.9 (1‐4) 2.89 ± 0.8 (1‐4) .379
Class II (DR + DQ) (mean ± SD) 1.94 ± 1.1 (0‐4) 2.02 ± 1.2 (0‐4) 1.85 ± 1.2 (0‐4) .291
(range)
Pretransplant cPRA (n, %); 51(29); 10.3 ± 23.2 44(51); 19.9 ± 29.6 7(8); 2.9 ± 12.3 <.001
(mean% ± SD)
Class I (A + B) (n, %) 26 (16) 23 (26) 3 (4) .290
Class II (DR+DQ) (n, %) 27 (16) 23 (26) 4 (5) .796
Pretransplant DSA (yes, %) 15 (9) 15 (17) 0 (0) <.001
Class I (A + B) (n, %) 8 (53) 8 (53) 0 (0) .302
Class II (DR+DQ) (n, %) 9 (60) 9 (60) 0 (0) .424
Mean time of Bx assessment (mo) 36.7 ± 31.5 (4‐96) 61.5 ± 32.6 (4‐96) 25.8 ± 4.6 (21‐31) <.001
(range)

P values correspond to intergroup comparisons for X 2 and one way analysis of variance.
AR, acute rejection; IS, immunosuppression; rATG, rabbit anti‐thymocyte globulin; IVIG, intravenous immunoglobulin; CNI, calcineurin inhibitor; TAC,
tacrolimus; CsA, cyclosporine; cPRA, calculated panel‐reactive antibodies; DSA, Donor Specific Antibodies; Bx, Biopsy; ESRD, End‐stage renal
disease.
LUQUE et al. | 5

longitudinal cohort, patients evaluated in for‐cause biopsies showed

3.2 | Clinical, histological, and immunological
characteristics of the study population
Main baseline clinical and demographic characteristics of the en‐
tire study population are depicted in Table 1. As compared to the
higher pretransplant sensitization, by means of both pretrans‐
plant calculated panel‐reactive antibodies (PRA) and preformed
DSA, received previous transplants and T‐cell depletion as induc‐
tion therapy, although the majority received a calcineurin‐inhibitor

TA B L E 2 Main clinical, histological, and immunological characteristics at the time of biopsy assessment

Longitudinal cohort
(Surveillance Bx)
Cross‐sectional cohort (n = 90)
Main clinical and demo‐ (For cause Bx)
graphic variables (n = 85) 6‐mo Bx 24‐mo Bx P

Allograft function at the time of biopsy

eGFR (mL/min) 32.4 ± 19.3 52.1 ± 17.4 51.2 ± 16.8 <.001
sCreat (µmol/L) 233.7 ± 166.9 132.1 ± 39.4 131.4 ± 59.1 <.001
Proteinuria g/day 2.17 ± 2.47 0.27 ± 0.66 0.34 ± 0.66 <.001
Anti‐HLA Ab at the time of biopsy
Class I (A + B) (yes, %) 44 (52) 6 (6) 7 (8) <.001
Class II (DR + DQ) (yes, %) 55 (65) 10 (11) 15 (16) <.001
DSA at the time of biopsy
Class I (A + B) (yes, %) 32 (37) 2 (2) 2 (2) <.001
Class II (DR + DQ) (yes, %) 23 (27) 4 (4) 6 (6) <.001
DSA MFI at the time of 3330 ± 5350 466 ± 1878 1197 ± 3802 <.001
biopsy (mean ± SD) (range) (0‐20953) (0‐13607) (0‐18127)
Histological lesions in renal compartments:
Acute lesions(mean ± SD;
range)
Interstitial inflammation (ai) 0.85 ± 0.9 (0‐3) 0.54 ± 0.8 (0‐3) 0.51 ± 0.7 (0‐3) .011
Tubulitis (at) 0.59 ± 0.8 (0‐3) 0.37 ± 0.8 (0‐3) 0.36 ± 0.7 (0‐3) .062
Glomerulitis (ag) 1.38 ± 1.0 (0‐3) 0.20 ± 0.5 (0‐2) 0.41 ± 0.8 (0‐3) <.001
Peritubular capillaritis (ptc) 1.18 ± 0.9 (0‐3) 0.14 ± 0.4 (0‐2) 0.35 ± 0.7 (0‐3) <.001
MIV (ptc+ag/2) 1.20 ± 0.9 (0‐3) 0.21 ± 0.4 (0‐1) 0.37 ± 0.7 (0‐3) <.001
C4d+ 0.93 ± 1.2 (0‐3) 0.12 ± 0.5 (0‐3) 0.22 ± 0.7 (0‐3) <.001
Endothelialitis (av) 0.18 ± 0.6 (0‐3) 0.02 ± 0.1 (0‐1) 0.05 ± 0.3 (0‐2) .094
Chronic lesions(mean ± SD; range)
Transplant glomerulopathy 1.11 ± 0.9 (0‐3) 0.02 ± 0.1 (0‐1) 0.12 ± 0.5 (0‐3) <.001
(cg)
Interstitial fibrosis (ci) 1.41 ± 0.8 (0‐3) 0.42 ± 0.5 (0‐2) 0.48 ± 0.6 (0‐2) <.001
Tubular atrophy (ct) 1.38 ± 0.8 (0‐3) 0.54 ± 0.5 (0‐1) 0.52 ± 0.5 (0‐2) <.001
Chronic Endothelialitis (cv) 0.70 ± 0.9 (0‐3) 0.10 ± 0.3 (0‐1) 0.10 ± 0.3 (0‐2) <.001
Banff phenotypes (n, %)
aABMR/DSA+/cABMR/ 16(19)/30(35) 2(2)/0(0) 5(5)/0(0) <.001
DSA+
ABMR/DSA− 29 (34) 3 (3.5) 7 (8)
TCMR/DSA− 0 (0) 5 (6) 4 (4)
IFTA/DSA− 10 (12) 5 (6) 14 (15)
STA/DSA− 0 (0) 75 (82) 60 (66)

P values correspond to intergroup comparisons for X2 and one way analysis of variance. Statistical significant differences were observed between for
cause and both time points of Surveillance biopsy groups.
eGFR, estimated glomerular filtration rate; sCreat, serum creatinine; DSA, Donor Specific Antibodies.
6 |
F I G U R E 2 Presence of donor‐reactive mBC frequencies
between different histological phenotypes in cross‐sectional cohort.
(A) Donor‐reactive mBC frequencies between different histological
phenotypes in for‐cause biopsies. Donor‐reactive mBC frequencies in
aABMR/DSA+, cABMR/DSA+, cABMR/DSA‐, and IFTA/DSA− were:
0.38 ± 0.20; 0.25 ± 0.28; 0.18 ± 0.21, and 0.08 ± 0.08, respectively.
Groups were compared using the one‐way analysis of variance
(ANOVA). (B) Percentage of patients with detectable donor‐reactive
mBC in peripheral blood in different clinical phenotypes. Percentage
of patients with detectable donor‐reactive mBC in peripheral blood
in aABMR/DSA+, cABMR/DSA+, cABMR/DSA−, and IFTA/DSA−
were 16/16(100%); 24/30(80%); 21/29(72.4%), and 2/10(20%),
respectively. (C) Percentage of patients showing more than one
donor‐antigen specificity according to DSA or donor‐reactive mBC.
The percentage of patients showing more than one antidonor antigen
mismatch positivity was higher when evaluating donor‐reactive
mBC as compared to circulating DSA; 9/46(19.5%) patients with DSA
showed more than one antidonor positivity, whereas 27/63(42.8%)
patients with detectable donor‐reactive mBC displayed more than
one detectable mBC donor‐antigen specificities (P < .001)
(CNI)‐based immunosuppressive regimen. Mean time of biopsy as‐
sessment from transplant surgery in the cross‐sectional group was
longer (mean 61.5 ± 32.6 months, range 4‐96 months) than the
LUQUE et al.
longitudinal cohort that were all assessed at 6 and 24 months after
transplantation.
At the time of biopsy assessment (Table 2), patients from the cross‐
sectional cohort showed significantly worse kidney allograft function
and higher levels of anti‐HLA antibodies and DSA than the longitudi‐
nal cohort. As per purposes of the study, patients of the cross‐sec‐
tional cohort had significantly higher acute and chronic fundamental
Banff score allograft lesions than patients evaluated at 6‐ and 24‐
month surveillance biopsies. Main histopathological phenotypes
within the cross‐sectional cohort were ABMR (75/85, 88%) either on
its acute (aABMR) or chronic (cABMR) forms, with or without detect‐
able DSA in sera (46/85[54%] ABMR/DSA+ and 29/85[34%] ABMR/
DSA−). Also, 10/85(12%) of patients showed unspecific interstitial fi‐
brosis and tubular atrophy without inflammation and DSA (interstitial
fibrosis and tubular atrophy [IFTA]/DSA−). Conversely, whereas most
patients in the longitudinal cohort showed a preserved, stable (STA)
allograft histology both at 6‐ and 24‐month surveillance biopsies (75
[83.5%] and 60 [67%], respectively), 5 (5.5%) patients displayed his‐
tological lesions compatible with ABMR, 5 T cell‐mediated rejection
(TCMR; 5.5%), and 5 IFTA (5.5%) at 6 months, whereas 12 (13.5%)
showed ABMR, 4 (4%) TCMR, and 14 (15.5%) IFTA at 24 months.
3.3 | Patients with ABMR in for‐cause allograft
biopsies show high frequencies of circulating donor‐
reactive mBCs
As shown in Figure 2A, high donor‐reactive mBC frequencies against
class I and class II donor‐HLA mismatch antigens were observed in all
aABMR/DSA+, being significantly higher than those detected among
cABMR/DSA+ and cABMR/DSA− patients (0.38 ± 0.20; 0.25 ± 0.28;
0.18 ± 0.21 donor‐reactive mBCs, respectively, P = .036 and
P = .003, respectively). Most IFTA/DSA‐ patients (8/10, 80%) did not
show donor‐reactive mBC frequencies. As shown in Figure 2B, all 16
(100%) aABMR/DSA+ patients and most cABMR/DSA+ (24/30;80%)
displayed detectable donor‐reactive mBCs. Interestingly, 21/29
(72.4%) cABMR/DSA− showed donor‐reactive mBCs despite no evi‐
dence of serum DSA, whereas only 2/10 (20%) IFTA/DSA− patients
did display donor‐reactive mBCs. Of note, although all antidonor
mBC specificities fitted with the same donor mismatch antigen of
the respective DSA found in serum, some patients showed addi‐
tional donor‐reactive mBCs against other donor‐mismatch antigens
(Figure 2C). The number of DSA and donor‐reactive mBC positivities
per patient is illustrated in Figure S3.
3.4 | Numbers of different circulating B cells and
T‐follicular helper cells do not differ between distinct
histological phenotypes
Numbers and percentages of different B and T‐cell subsets at dif‐
ferent activation stages were also evaluated in patients undergo‐
ing for‐cause biopsies as well as in a group of STA/DSA− patients
of the longitudinal cohort at the time of 24‐month protocol biop‐
sies. As illustrated in Figure S4, neither numbers nor percentages of
LUQUE et al.
F I G U R E 3 Evolution of histological allograft lesions and anti‐
donor humoral immunity between 6 and 24 months after kidney
transplantation. (A) Progression of histological allograft lesions
between 6 and 24 months after kidney transplantation. At 6 months,
75/90 (83.5%) showed an STA allograft, 5/90 (5.5% ABMR, 5/90
(5.5%) TCMR, and 5/90 (5.5%) IFTA allograft, whereas at 24‐ months,
60/90 (67%) displayed an STA graft, 12/90(13%) ABMR, 4/90 (4.5%)
TCMR, and 14/90(15.5%). (B) Progression of antidonor humoral
immunity between 6 and 24 months after kidney transplantation.
Whereas 7/90 (7.7%) developed de novo DSA and 8/90(9%)
donor‐reactive mBC at 6 months, 10/90 (11%) showed DSA and
14/90(15.5%) donor‐reactive mBC 2 years after transplantation
different B cell subtypes such as naive (CD19+CD27−IgD+), transi‐
tional (CD19+IgM+CD27−CD24+CD38+), unswitched memory (CD19
+ (66.6%), of the biopsy‐proven subclinical ABMR at 24‐month
CD27+CD38−IgD−IgM+), switched memory (CD19+CD27+CD38−IgD
−
IgM−), plasmablast (CD19+CD27++CD38++IgD−IgM−), and TFH cells
(CD3+CD4+CXCR5+) were found to be different between the differ‐
ent histological phenotypes.
3.5 | Progression of antidonor humoral immunity and
allograft lesions between 6 and 24 months after KT
Ninety nonsensitized KT were prospectively assessed for histologi‐
cal changes and development of de novo DSA and donor‐reactive
mBC at the time of 6‐ and 24‐month surveillance biopsies. As de‐
picted in Figure 3A, whereas at 6‐month 75/90 (73.5%) patients
displayed a STA allograft, 5/90 (5.5%) ABMR, 5/90 (5.5%) TCMR,
and 5/90 (5.5%) IFTA lesions, at 24 months, 60/90 (67%) remained
with a STA graft, 12/90 (13%) showed ABMR, 4/90 (4.5%) TCMR,
and 14/90 (15.5%) IFTA. Out of the 12 ABMR at 24 months, 5 had
| 7
F I G U R E 4 Presence of DSA or donor‐reactive mBC in
longitudinal cohort patients. (A) Presence of DSA or donor‐reactive
mBC at 6‐month protocol biopsies. At 6‐month surveillance biopsies
DSA were detected in 4/5 (80%) ABMR, 0/5 (0%) TCMR, 0/5 (0%)
IFTA, and 3/75 (4%) STA patients, whereas donor‐reactive mBC
frequencies were observed in 5/5 (100%) ABMR, 1/5 (20%) TCMR,
0/5 (0%) IFTA, and 2/75 (2.6%) STA patients. (B) Presence of DSA
or donor‐reactive mBC at 24‐month protocol biopsies. At 24‐month
surveillance biopsies, DSA were detected in 8/12 (66.6%) ABMR,
0/4 (0%) TCMR, 0/6 (0%) IFTA, and 2/60 (3.3%) STA patients,
whereas donor‐reactive mBC frequencies were observed in 10/12
(83.3%) ABMR, 0/4 (0%) TCMR, 2/14 (14.2%) IFTA, and 2/60 (3.3%)
STA patients. (C) Progression of antidonor humoral biomarkers
in patients undergoing ABMR at 24‐month surveillance biopsies.
The assessment of donor‐reactive mBC prior to or at the time of
24‐month surveillance biopsies predicted 10/12 (83.3%) biopsy‐
proven subclinical ABMR at 24‐month, whereas DSA predicted 8/12
ABMR, 1 TCMR, and 6 a STA graft at 6 months. All four patients with
TCMR at 24 months had the same lesions at 6 months. All five IFTA
patients at 6 months displayed the same lesions at 24 months and
the additional nine had a STA graft at 6 months. Whereas 7/90 (7.7%)
developed de novo DSA and 8/90 (8.8%) donor‐reactive mBCs at 6
months, 10/90 (11.1%) showed DSA, and 14/90 (15.5%) donor‐reac‐
tive mBCs 2 years after transplantation (Figure 3B).
3.6 | Donor‐reactive mBC associates with
fundamental histological lesions of ABMR in stable
KT patients
We next analyzed the association between de novo DSA and donor‐
reactive mBCs in relation to the histological lesions found at 6‐ and
8 |
A 24‐month surveillance biopsies. As illustrated in Figure S5, patients
B
F I G U R E 5 Discrimination capacity of circulating DSA and
donor‐reactive mBC for histological lesions compatible of ABMR.
(A) Receiver operating characteristic curves of DSA (MFI) and
donor‐reactive mBC (frequencies) predicting ABMR. Receiver
Operating Characteristic curves were plotted for the highest mean
fluorescence intensity (MFI) DSA in each patient (gray) and for
the highest frequency of donor‐reactive mBC (red) (AUC = 0.809
95%CI 0.7042‐0.876, P < .001 for DSA [MFI] and AUC = 0.917 95%
CI 0.879‐0.956, P < .001 for donor‐reactive mBC frequencies). (B)
Unsupervised hierarchical clustering analysis of main histological
lesions in all kidney compartments, diagnosis of ABMR and DSA and
donor‐reactive mBC. Segregation of allograft rejection phenotypes
according to circulating DSA and donor‐reactive mBC. Variables
considered in this analysis were all histologic lesions at all kidney
allograft compartments (glomerulitis [ag], peritubular capillaritis
[ptc], microvascular inflammation [miv], endarteritis [av], interstitial
inflammation [ai], tubulitis [at], C4d complement fraction deposition
in ptc, transplant glomerulopathy [cg], interstitial fibrosis [ci], and
tubular atrophy [ct]) and presence or absence of DSA and donor‐
reactive mBC. Data are on the basis of 175 kidney allograft biopsies
performed either for cause or for protocol. Each histological lesion
found in each compartment in an individual patient is colored
according to the threshold for each parameter (zero to three, with
higher score translating more severe injury). Cluster ABMR was
enriched with patients with a kidney graft histology highly suggestive
of ABMR, and cluster No‐ABMR was enriched with patients without
such histological lesions. Immediately below these gray color clusters,
blue/green/red/white color bars indicate the combination of DSA and
donor‐specific mBC frequencies as depicted in the legend
LUQUE et al.
with DSA and donor‐reactive mBC showed significantly higher mi‐
crovascular inflammation scores (ag, ptc, and miv) and transplant
glomerulopathy (cg) than patients without DSA or donor‐reactive
mBCs. Presence of donor‐reactive mBC was also significantly associ‐
ated with higher chronic scores at the interstitium (ci) and tubuli (ct),
whereas patients with DSA showed significantly higher C4d stain‐
ing at the ptc. Notably, ai and at lesions were not associated with
either DSA or donor‐reactive mBC. No association was observed in
tacrolimus trough levels at all times of assessment between patients
developing donor‐reactive mBC and/or DSA and patients that did
not (data not shown).
3.7 | Donor‐reactive mBC progressively expands
over time and predicts subsequent ABMR
The progression of both DSA and donor‐reactive mBC over time in re‐
lation with graft lesions in 6‐ and 24‐month surveillance biopsies was
next investigated. As illustrated in Figure 4A, at 6 months, all 5 (100%)
ABMR patients displayed donor‐reactive mBC and 4/5 (80%) were
DSA+. Also, 3/75 (4%) and 2/75 (2.6%) STA patients showed DSA and
donor‐reactive mBC, respectively. Conversely, only 1/5 (20%) patient
with TCMR showed donor‐reactive mBC but none displayed DSA. All
IFTA patients were DSA and donor‐reactive mBC negatives. At 24
months (Figure 4B), whereas 10/12 (83%) and 8/12 (66.6%) patients
with ABMR displayed donor‐reactive mBC and DSA, respectively,
2/60 (3%) STA patients did also show DSA and donor‐reactive mBC.
Also, two IFTA patients displayed donor‐reactive mBC but no DSA
and none of TCMR patients displayed DSA or donor‐reactive mBC.
Importantly, although out of the eight patients who had detectable
donor‐reactive mBC at 6 months, seven showed subclinical ABMR at
24‐month surveillance biopsies, three out of the five additional pa‐
tients with ABMR at 24 months had developed donor‐reactive mBC
between 6 and 24 months. Thus, the presence of donor‐reactive mBC
predicted 10/12 (83.3%) of the ­biopsy‐proven subclinical ABMR at
24‐month, whereas DSA predicted 8/12 (66.6%) of the biopsy‐proven
subclinical ABMR at 24‐month (Figure 4C). Figure S6 summarizes the
association between de novo DSA and donor‐reactive mBC with the
advent of 6‐ and 24‐month subclinical ABMR.
3.8 | Donor‐reactive mBC independently
predicts ABMR
We merged all patients of the study and assessed the predictive
capacity of DSA and donor‐reactive mBC for identifying ABMR. As
shown in Figure 5A, Receiver Operating Characteristic curve analyses
showed high AUC for both mean fluorescence intensity (MFI)‐DSA
and donor‐reactive mBC frequencies predicting ABMR (AUC = 0.809
95%CI 0.7042‐0.876, P < .001 for MFI‐DSA and AUC = 0.917 95%CI
0.879‐0.956, P < .001 for donor‐reactive mBC frequencies). Whereas
DSA showed high specificity (97.1%) and positive predictive value
(PPV) (92.1%) but lower sensitivity (63.1%) and negative predictive
value (NPV) (83.1%), donor‐reactive mBC displayed higher sensitivity
LUQUE et al. | 9

TA B L E 3 Predictive values of DSA and

Biomarkers predicting
donor‐reactive mBC identifying ABMR
ABMR Sensitivity (%) Specificity (%) PPV (%) NPV (%)

Donor‐reactive mBC 82.6 94.7 89.4 91.1

(frequencies)
DSA (MFI) 63.1 97.1 92.1 83.1

PPV, positive predictive value; NPV, negative predictive value; MFI, mean fluorescent intensity.

TA B L E 4 Univariate and multivariate binary logistic regression analysis of main clinical, demographic variables predicting ABMR

Univariate analysis Multivariate analysis

Variables predicting ABMR OR CI 95% P value OR CI 95% P value

Recipient age (at 0.983 0.960‐1.006 .153

transplantation)
Recipient gender (female) 2.125 1.141‐3.959 .018 1.277 0.344‐4.744 .715
Induction therapy (rATG) 2.453 1.270‐4.738 .008 0.521 0.112‐2.430 .406
Previous transplants 3.106 1.537‐6.279 .002 0.318 0.056‐1.807 .196
Previous AR (yes) 6.109 1.940‐19.239 .002 6.739 0.846‐53.699 .072
HLA mismatch 1.125 0.915‐1.385 .264
HLA mismatch harbored in 2.556 1.006‐6.496 .049 2.907 09195‐9.200 .069
previous Tx
Pre‐TX DSA (POS) 18.400 6.795‐49.825 <.001 19.697 2.770‐140.085 .003
DSA at time of biopsy (POS) 123.250 16.402‐926.167 <.001 37.681 2.847‐498.642 .006
Donor‐reactive mBC at time 60.646 22.521‐163.310 <.001 47.157 10.048‐221.311 <.001
of biopsy (POS)

P values are shown for binary logistic regression. Statistically significant values are represented in bold characters.
rATG, rabbit antithymocyte globulin; AR, acute rejection; Tx, transplants; cPRA, calculated panel‐reactive antibodies; POS, positive; DSA, donor‐spe‐
cific antibodies; mBC, memory B cells.

(82.6%) and NPV (91.1%) although lower specificity (94.7%) and PPV
(89.4%) for characterizing ABMR (Table 3).
Using integrated unsupervised hierarchical clustering analysis
(Figure 5B), we confirmed that high histological score parameters as‐
sociated with ABMR such as ag, ptc, miv, cg, and c4d could identify
patients with DSA and in a greater proportion, patients with donor‐
reactive mBC.
The multivariate binary logistic regression analysis taking into
account well‐known clinical, demographic, and immunological
characteristics associated with ABMR, revealed that together with
preformed and posttransplant DSA, donor‐reactive mBC were inde‐
pendent predictors of ABMR (Table 4).
As illustrated in Figure 7A, patients with ABMR and IFTA lesions
showed significantly lower survival rates than STA and TCMR patients
(Log‐rank P < .001) (HR = 31.848, 95% CI 4.372‐231.998, P = .001
for ABMR; HR = 14.901, 95% CI 1.792‐123.907, P = .012 for IFTA;
HR = 0.702, 95% CI 0.096‐5.100, P = .726 for TCMR, and HR = 0.025,
95%CI 0.002‐0.309, P = .004 for STA, respectively). When the impact
on graft survival was analyzed according to DSA and donor‐reactive
mBC, patients with donor‐reactive mBC, regardless of the presence of
DSA, showed the highest allograft failure rates (Log‐rank, P < .001) (HR
= 8.162, 95% CI 3.699‐18.010, P < .001 for DSA+/mBC+; HR = 5.750,
95% CI 2.296‐14.402, P < .001for DSA−/mBC+; HR = 0.745, 95%
CI 0.180‐3.081, P = .684 for DSA+/mBC−; HR = 0.127, 95% CI
0.054‐0.299, P < .001 for DSA−/mBC−) (Figure 7B).

3.9 | Donor‐reactive mBC confer a strong

deleterious effect on kidney allograft outcome
Finally, we analyzed the impact of donor‐reactive mBC and DSA on
kidney graft function and survival. As shown in Figure 6A, presence
of mBC, regardless of DSA, was associated with worse graft function
at 6, 12, and 24 months as compared to other groups. When only pa‐
tients of the longitudinal cohort were assessed, patients with donor‐
reactive mBC did also display the worse graft function progression
as compared to all the other groups (Figure 6B).
4 | D I S CU S S I O N
In this study, we aimed at evaluating the value of tracking donor‐re‐
active mBC circulating in peripheral blood after KT to characterize
histopathological lesions associated with ABMR and also, investigate
the kinetics of appearance of donor‐reactive mBC as compared to
serum DSA and their impact on kidney graft outcome, in order to
gain further insight on the pathophysiology of humoral transplant
10 | LUQUE et al.

F I G U R E 6 Impact of donor‐reactive mBC and DSA on kidney graft function progression. (A) Graft function progression (serum creatinine)
over time according to the presence or absence of DSA and/or donor‐reactive mBC after last biomarker assessment in all patients of the
study. Serum creatinine (mmol/L) progression after 6 months of last biomarker determination was (135.9 ± 45.4; 113.4 ± 31.0; 183.4 ± 66.5;
243.1 ± 187.8, for DSA−/mBC−, DSA+/mBC−, DSA−/mBC+, and DSA+/mBC+, respectively, P < .001), after 12‐ months (154.2 ± 78.8;
122.1 ± 56.7; 215.7 ± 107.1; 184.6 ± 101.9, for DSA−/mBC−, DSA+/mBC−, DSA−/mBC+, and DSA+/mBC+, respectively, P = .017) and
24 months after (147.9 ± 64.8; 106.3 ± 25.7; 208.6 ± 104.6; 193.1 ± 98.7, for DSA−/mBC−, DSA+/mBC−, DSA−/mBC+, and DSA+/mBC+,
respectively, P = .012). (B) Graft function progression over time after 24‐month protocol biopsy according to DSA or donor‐reactive mBC
in the longitudinal cohort. Serum creatinine (mmol/L) after 6 months was (127.9 ± 39.7; 100.6 ± 11.9; 157.4 ± 37.4; 181.7 ± 102.3, for
DSA−/mBC−, DSA+/mBC−, DSA−/mBC+, and DSA+/mBC+, respectively, P = .019), after 12 months (141.5 ± 50.5; 98.3 ± 11.1; 163.5 ± 37.9;
210.3 ± 155.9, for DSA−/mBC−, DSA+/mBC−, DSA−/mBC+, and DSA+/mBC+, respectively, P = .052) and after 24 months (133.5 ± 41.2;
109.0 ± 29.7; 182.2 ± 108.7; 172.3 ± 53.9, for DSA−/mBC−, DSA+/mBC−, DSA−/mBC+, and DSA+/mBC+, respectively, P = .109)

rejection. In a first group of incident patients undergoing for‐cause
biopsy, we demonstrate that although KT developing ABMR lesions
with DSA also show circulating donor‐reactive mBC, the majority
of patients (72.4%) with such lesions but undetectable serum DSA
(cABMR/DSA−) do also display high frequencies of circulating donor‐
reactive mBC thus, suggesting that these lesions may likely be driven
by a similar antidonor HLA B cell effector immune mechanism, al‐
though we cannot exclude other roles of circulating mBC spreading
out the alloimmune response. Further, the prospective evaluation of
both DSA and donor‐reactive mBC after transplantation, revealed
a progressive expansion of donor‐reactive mBC that discriminated
KT with ongoing subclinical ABMR and importantly, those at higher
risk of subsequently developing such lesions over time. Of note, the
presence of circulating donor‐reactive mBC conferred a significant
deleterious impact on KT outcome.
As recently reported in experimental transplantation23,29 and
in humans, 26 KT developing acute ABMR short after transplant
surgery might show high frequencies of preformed donor‐reactive
mBC circulating in peripheral blood, even in the absence of detect‐
able DSA. Interestingly, at the time of biopsy assessment, patients
with acute ABMR showed significantly higher donor‐reactive mBC
responses than those displaying chronic ABMR lesions, suggesting
a stronger alloimmune response driven by circulating alloreactive
mBC, which may not be sufficiently controlled by conventional
chronic immunosuppression. These higher donor‐reactive mBC fre‐
quencies may also suggest an already preformed sensitization sta‐
tus in early acute ABMR patients; in fact, pretransplant DSA within
the acute ABMR was higher than among the late ABMR group thus,
reinforcing this hypothesis. Circulating donor‐reactive mBC also
identified transplant patients with subclinical ABMR in surveillance
biopsies. Notably, an unsupervised analysis of fundamental graft
lesions taking into account the presence of donor‐reactive mBC or
DSA, revealed a strong association between the two biomarkers
and microvascular inflammation (miv), glomerulitis, and peritubular
capillaritis as well as transplant glomerulopathy (cg), highlighting
their direct contribution to the development of the same lesions
LUQUE et al.
associated to ABMR. Interestingly, patients with donor‐reactive
mBC did more likely show higher IFTA scores in 24‐month surveil‐
lance biopsies, a finding in line with recent reports suggesting a di‐
rect contribution of humoral immunity on premature development
of allograft fibrosis. 30
A main new finding in our work relies on the kinetics of the
antidonor humoral immune activation after KT through the assess‐
ment of the global antidonor immune repertoire by means of either
DSA or donor‐reactive mBC. Here, we confirm on the one hand,
previous data showing that among non‐sensitized individuals the
yearly rates of de novo DSA formation ranges between 5%‐7%31,32
and on the other, that donor‐reactive mBC may appear in periph‐
eral blood at early stages, progressively leading to smoldering but
constant allograft damage, even prior to DSA detection in a number
of patients.
| 11
F I G U R E 7 Kaplan‐Meier graft survival after biopsy assessment
according to Banff histological lesions and presence or absence of
DSA and/or donor‐reactive mBC. (A) Kaplan‐Meier graft survival after
biopsy assessment according to different Banff histological lesions.
Patients with ABMR and IFTA showed significantly lower survival
rates than STA and TCMR patients (Log‐rank P < .001) (HR = 31.848,
95%CI 4.372‐231.998, P = .001 for ABMR; HR = 14.901, 95% CI
1.792‐123.907, P = .012 for IFTA; HR = 0.702, 95% CI 0.096‐5.100,
P = .726 for TCMR, and HR = 0.025, 95%CI 0.002‐0.309, P = .004
for STA, respectively). (B) Kaplan‐Meier graft survival after biopsy
assessment according to the presence or absence of DSA and/or
donor‐reactive mBC. According to the combination of biomarkers,
patients with donor‐reactive mBC, regardless of the presence of DSA,
showed significantly worse allograft survival than all the other groups
(Log‐rank, P < .001) (HR = 8.162, 95% CI 3.699‐18.010, P < .001
for DSA+/mBC+; HR = 5.750, 95% CI 2.296‐14.402, P < .001 for
DSA−/mBC+; HR = 0.745, 95% CI 0.180‐3.081, P = .684 for DSA+/
mBC−; HR = 0.127, 95% CI 0.054‐0.299, P < .001 for DSA−/mBC−,
respectively)
The predictive value of both MFI‐DSA and donor‐reactive mBC
frequencies for the detection of ABMR revealed that both biomark‐
ers can discriminate the presence of ongoing ABMR lesions with
high accuracy (AUC > 0.85). When main clinical, demographic, and
immunological variables were analyzed together in a multivariate bi‐
nary logistic regression for the prediction of ABMR, the presence of
donor‐reactive mBC was an additional independent variable predict‐
ing ABMR. Notably, although the presence of DSA has a high spec‐
ificity and PPV identifying ABMR, circulating donor‐reactive mBC
showed higher sensitivity and NPV, thus illustrating their potentially
important role to rule out the risk of ABMR. Further, patients with
donor‐reactive mBC were at particularly higher risk of both worse
graft function progression and poorer graft survival thus, highlight‐
ing that ABMR driven by circulating mBC seems to present a more
severe clinical course, most likely due to an earlier occurrence of
graft damage.
Interestingly, the higher antidonor immune specificities when
measuring donor‐reactive mBC as compared to circulating DSA,
illustrate a broader B cell repertoire of mBC than serum DSA.
Importantly, the exclusive evaluation of different circulating B cell
subset numbers at different maturation stages as well as TFH cells
using flow cytometry, was not able to discriminate between differ‐
ent histological lesions. This is most likely due to the lack of antigen‐
specific functional insight provided by circulating cell phenotypes
as compared to assessing donor‐reactive mBC. Importantly, a novel
achievement in this study has been the development of a new mul‐
ticolor HLA‐specific B cell ELISpot assay, which allows the detec‐
tion of multiple donor‐reactive mBC specificities within the same
ELISpot well. This technical improvement significantly increases the
efficiency of the assay as it enables testing for distinct HLA antigens
with less numbers of polyclonal activated mBC.
This study has some limitations. We did not measure non‐HLA
antibodies, which could have eventually lead to similar histolog‐
ical ABMR lesions in the absence of DSA. 33 However, the clear
detection of donor‐reactive mBC specifically in patients with
graft injury and not in patients with preserved grafts, significantly
12 | LUQUE et al.
counterbalances this limitation. Surprisingly, in a small group of
patients with ABMR and DSA (16%), no donor‐reactive mBC were
detected. A potential explanation could rely on the very low an‐
tidonor mBC frequencies that could not be detected even using
an ELISpot assay or conversely, that such DSA were directly
produced by bone marrow‐residing plasma cells. 34,35 We unfor‐
tunately could not evaluate donor‐reactive mBC at baseline in
the longitudinal cohort because of lack of biological specimens.
Nonetheless, and while we cannot rule out the possibility that
some of the 6‐month donor‐reactive mBC were already present
prior to transplantation, the similar proportion of donor‐reactive
mBC detected at 24‐months, highly suggests a constant activation
of donor‐reactive mBC over time.
In conclusion, our study shows for the first time that circu‐
lating donor‐reactive mBC might be monitored using a sensitive
HLA B cell ELISpot assay after transplantation and their detec‐
tion is highly associated with the presence of ABMR lesions, also
among those with quiescent allograft function and undetectable
serum DSA. Notably, donor‐reactive mBC seems to progressively
expand in peripheral blood over the course of the transplant and
negatively impact on allograft outcome. These data provides new
insight into the effector mechanism triggering humoral rejection
and highlights a central role of alloreactive mBC. Prospective, in‐
terventional biomarker‐driven trials targeting alloreactive mBC
are highly warranted, most preferable within collaborative inter‐
national networks.
AC K N OW L E D G M E N T S
We thank our biobank unit at IDIBELL for careful management of all
biological samples. We thank all the patients who voluntarily pro‐
vided their informed consent to participate in the study. OB received
two competitive Spanish research grants (ISCiii PI16/01321, ISCiii
PI13/01263) a Feder funding a way to build Europe and a European
Commission grant from the Biomarker‐Driven Immunosuppression
Minimization Consortium (BIODRIM) (No. FP7/2007‐2017). We
thank CERCA programme/Generalitat de Catalunya for institutional
support. OB has also received two intensification research grants
from the ISCiii (INT15/00112). SL and ML received two research
fellowship grants from the Catalan Society of Transplantation. MC
was supported by three research grants (ISCiii PI16/00617, ISCiii
PI13/00598, RD16/0009/0013 (RedinRen).
C O N FL I C T S O F I N T E R E S T
The authors of this manuscript have no conflicts of interest to dis‐
close as described by the American Journal of Transplantation.
ORCID
Marta Crespo http://orcid.org/0000-0001-6992-6379
Oriol Bestard https://orcid.org/0000-0001-9468-7920
REFERENCES
1. Sellarés J, de Freitas DG, Mengel M, et al. Understanding the causes
of kidney transplant failure: the dominant role of antibody‐mediated
rejection and nonadherence. Am J Transplant. 2012;12(2):388‐399.
https://doi.org/10.1111/j.1600-6143.2011.03840.x.
2. Halloran PF, Schlaut J, Solez K, Srinivasa NS. The significance of
the anti‐class I response. II. Clinical and pathologic features of
renal transplants with anti‐class I‐like antibody. Transplantation.
1992;53(3):550–555.
3. Lefaucheur C, Loupy A, Hill GS, et al. Preexisting donor‐specific HLA
antibodies predict outcome in kidney transplantation. J Am Soc Nephrol.
2010;21(8):1398‐1406. https://doi.org/10.1681/ASN.2009101065.
4. Wiebe C, Gibson IW, Blydt‐Hansen TD, et al. Evolution and clini‐
cal pathologic correlations of de novo donor‐specific HLA antibody
post kidney transplant. Am J Transplant. 2012;12(5):1157‐1167.
https://doi.org/10.1111/j.1600-6143.2012.04013.x.
5. Porter KA, Dossetor JB, Marchioro TL, et al. Human renal trans‐
plants. I. Glomerular changes. Lab Invest. 1967;16(1):153‐181.
6. Regele H, Böhmig GA, Habicht A, et al. Capillary deposition of
complement split product C4d in renal allografts is associated with
basement membrane injury in peritubular and glomerular capillar‐
ies: a contribution of humoral immunity to chronic allograft rejec‐
tion. J Am Soc Nephrol. 2002;13(9):2371‐2380.
7. Gloor JM, Sethi S, Stegall MD, et al. Transplant glomeru‐
lopathy: subclinical incidence and association with alloan‐
tibody. Am J Transplant. 2007;7(9):2124‐2132. https://doi.
org/10.1111/j.1600-6143.2007.01895.x.
8. Maryniak RK, First MR, Weiss MA. Transplant glomerulopa‐
thy: evolution of morphologically distinct changes. Kidney Int.
1985;27(5):799‐806.
9. Feucht HE, Schneeberger H, Hillebrand G, et al. Capillary deposi‐
tion of C4d complement fragment and early renal graft loss. Kidney
Int. 1993;43(6):1333‐1338.
10. Feucht HE. Complement C4d in graft capillaries – the missing
link in the recognition of humoral alloreactivity. Am J Transplant.
2003;3(6):646‐652.
11. Mauiyyedi S, Pelle PD, Saidman S. Chronic humoral rejection:
identification of antibody‐mediated chronic renal allograft rejec‐
tion by C4d deposits in peritubular capillaries. J Am Soc Nephrol.
2001;12(3):574‐582.
12. Halloran PF, Wadgymar A, Ritchie S, Falk J. Solez K, Srinivasa
NS. The significance of the anti‐class I antibody response : I.
Clinical and pathologic features of anti‐class I‐mediated rejection.
Transplantation. 1990;49(1):85‐91.
13. Sis B, Jhangri GS, Riopel J, et al. A new diagnostic algorithm for
antibody‐mediated microcirculation inflammation in kidney
transplants. Am J Transplant. 2012;12(5):1168‐1179. https://doi.
org/10.1111/j.1600-6143.2011.03931.x.
14. Gebel HM, Bray RA. In search of perfection. Am J Transplant.
2013;13(8):1951‐1952. https://doi.org/10.1111/ajt.12335.
15. Halloran PF, Merino Lopez M, Barreto PA. Identifying subphe‐
notypes of antibody‐mediated rejection in kidney transplants.
Am J Transplant. 2016;16(3):908‐920. https://doi.org/10.1111/
ajt.13551.
16. Halloran PF, Famulski KS, Chang J. A probabilistic approach to his‐
tologic diagnosis of antibody‐mediated rejection in kidney trans‐
plant biopsies. Am J Transplant. 2017;17(1):129‐139. https://doi.
org/10.1111/ajt.13934.
17. AmannaIJ,SlifkaMK.Mechanismsthatdetermineplasmacelllifespanand
the duration of humoral immunity. Immunol Rev. 2010;236(1):125‐138.
https://doi.org/10.1111/j.1600-065X.2010.00912.x.
18. Tarlinton D, Good‐Jacobson K. Diversity among memory B cells: or‐
igin, consequences, and utility. Science. 2013;341(6151):1205‐1211.
https://doi.org/10.1126/science.1241146.
LUQUE et al.
19. McHeyzer‐Williams MG, Ahmed R. B cell memory and the long‐
lived plasma cell. Curr Opin Immunol. 1999;11(2):172‐179.
20. Chong AS, Sciammas R. Memory B cells in transplantation.
Transplantation. 2015;99(1):21‐28. https://doi.org/10.1097/
TP.0000000000000545.
21. Bauer T, Jilg W. Hepatitis B surface antigen‐specific T and B cell
memory in individuals who had lost protective antibodies after
hepatitis B vaccination. Vaccine. 2006;24(5):572‐577. https://doi.
org/10.1016/j.vaccine.2005.08.058.
22. Han M, Rogers JA, Lavingia B, Stastny P. Peripheral blood B cells
producing donor‐specific HLA antibodies in vitro. Hum Immunol.
2009;70(1):29‐34. https://doi.org/10.1016/j.humimm.2008.10.013.
23. Young JS, Chen J, Miller ML, et al. Delayed cytotoxic T lymphocyte‐
associated protein 4‐immunoglobulin treatment reverses ongoing
alloantibody responses and rescues allografts from acute rejection.
Am J Transplant. 2016;16(8):2312‐2323. https://doi.org/10.1111/
ajt.13761.
24. Yang J, Chen J, Young JS, et al. Tracing donor‐MHC class II reac‐
tive B cells in mouse cardiac transplantation: delayed CTLA4‐
Ig treatment prevents memory alloreactive B‐cell generation.
Transplantation. 2016;100(8):1683‐1691. https://doi.org/10.1097/
TP.0000000000001253.
25. Heidt S, Roelen DL, de Vaal Y, et al. A NOVel ELISPOT assay
to quantify HLA‐specific B cells in HLA‐immunized individ‐
uals. Am J Transplant. 2012;12(6):1469‐1478. https://doi.
org/10.1111/j.1600-6143.2011.03982.x.
26. Lúcia M, Luque S, Crespo E, et al. Preformed circulating HLA‐­
specific memory B cells predict high risk of humoral rejection
in kidney transplantation. Kidney Int. 2015;874‐887. https://doi.
org/10.1038/ki.2015.205.
27. Luque S, Lúcia M, Bestard O. Refinement of humoral immune mon‐
itoring in kidney transplantation: the role of “hidden” alloreactive
memory B cells. Transpl Int. 2017;30(10):955‐968. https://doi.
org/10.1111/tri.13014.
28. Haas M, Sis B, Racusen LC, et al. Banff 2013 meeting report: in‐
clusion of C4d‐negative antibody‐mediated rejection and antibody‐­
associated arterial lesions. Am J Transplant. 2014;14:272‐283.
https://doi.org/10.1111/ajt.12590.
29. Chen J, Wang Q, Yin D, Vu V, Sciammas R, Chong AS. Cutting edge:
CTLA‐4Ig inhibits memory B cell responses and promotes allograft
| 13
survival in sensitized recipients. J Immunol. 2015;195(9):4069‐4073.
https://doi.org/10.4049/jimmunol.1500940.
30. Gosset C, Viglietti D, Rabant M, et al. Circulating donor‐specific
anti‐HLA antibodies are a major factor in premature and acceler‐
ated allograft fibrosis. Kidney Int. 2017;92(3):729‐742. https://doi.
org/10.1016/j.kint.2017.03.033.
31. Wiebe C, Pochinco D, Blydt‐Hansen TD, et al. Class II HLA epi‐
tope matching ‐ a strategy to minimize de novo donor‐specific
antibody development and improve outcomes. Am J Transplant.
2013;13:3114‐3122. https://doi.org/10.1111/ajt.12478.
32. Crespo E, Cravedi P, Martorell J, et al. Posttransplant peripheral
blood donor–specific interferon‐γ enzyme‐linked immune spot
assay differentiates risk of subclinical rejection and de novo donor‐
specific alloantibodies in kidney transplant recipients. Kidney Int.
2017;92(1):201‐213. https://doi.org/10.1016/j.kint.2016.12.024.
33. Dragun D, Scornik J, Meier‐Kriesche H‐U. Kidney‐transplant rejec‐
tion and anti‐MICA antibodies. N Engl J Med. 2008;358(2):196.
34. Kwun J, Burghuber C, Manook M, et al. Humoral compensation after
bortezomib treatment of allosensitized recipients. J Am Soc Nephrol.
2017;28(7):1991‐1996. https://doi.org/10.1681/ASN.2016070727.
35. Perry DK, Pollinger HS, Burns JM, et al. Two novel assays of
­alloantibody‐secreting cells demonstrating resistance to desen‐
sitization with IVIG and rATG. Am J Transplant. 2008;8:133‐143.
https://doi.org/10.1111/j.1600-6143.2007.02039.x.
S U P P O R T I N G I N FO R M AT I O N
Additional supporting information may be found online in the
Supporting Information section at the end of the article.  
How to cite this article: Luque S, Lúcia M, Melilli E, et al.
Value of monitoring circulating donor‐reactive memory B
cells to characterize antibody‐mediated rejection after
kidney transplantation. Am J Transplant. 2018;00:1–13.
https://doi.org/10.1111/ajt.15055