| |

Received: 13 September 2017    Revised: 31 October 2017    Accepted: 17 November 2017

DOI: 10.1111/ajt.14600

MINIREVIEW

Review: The transcripts associated with organ allograft

rejection

Philip F. Halloran1,2 | Jeffery M. Venner1 | Katelynn S. Madill-Thomsen1,2 | 

Gunilla Einecke3 | Michael D. Parkes1 | Luis G. Hidalgo4 | Konrad S. Famulski1,4

1
Alberta Transplant Applied Genomics Centre,
Edmonton, AB, Canada The molecular mechanisms operating in human organ transplant rejection are best in-
2
Department of Medicine, Division ferred from the mRNAs expressed in biopsies because the corresponding proteins
of Nephrology and Transplant often have low expression and short half-­lives, while small non-­coding RNAs lack
Immunology, University of Alberta, Edmonton,
AB, Canada specificity. Associations should be characterized in a population that rigorously identi-
3
Medizinische Hochschule Hannover, fies T cell-­mediated (TCMR) and antibody-­mediated rejection (ABMR). This is best
Hannover, Germany
achieved in kidney transplant biopsies, but the results are generalizable to heart, lung,
4
Department of Laboratory Medicine and
Pathology, University of Alberta, Edmonton,
or liver transplants. Associations can be universal (all rejection), TCMR-­selective, or
AB, Canada ABMR-­selective, with universal being strongest and ABMR-­selective weakest. Top

Correspondence
universal transcripts are IFNG-­inducible (eg, CXCL11 IDO1, WARS) or shared by ef-
Philip F. Halloran fector T cells (ETCs) and NK cells (eg, KLRD1, CCL4). TCMR-­selective transcripts are
Email: phallora@ualberta.ca
expressed in activated ETCs (eg, CTLA4, IFNG), activated (eg, ADAMDEC1), or IFNG-­
induced macrophages (eg, ANKRD22). ABMR-­selective transcripts are expressed in
NK cells (eg, FGFBP2, GNLY) and endothelial cells (eg, ROBO4, DARC). Transcript as-
sociations are highly reproducible between biopsy sets when the same rejection defi-
nitions, case mix, algorithm, and technology are applied, but exact ranks will vary.
Previously published rejection-­associated transcripts resemble universal and TCMR-­
selective transcripts due to incomplete representation of ABMR. Rejection-­associated
transcripts are never completely rejection-­specific because they are shared with the
stereotyped response-­to-­injury and innate immunity

KEYWORDS
basic (laboratory) research/science, biopsy, kidney transplantation/nephrology, organ
transplantation in general, rejection

1 |  INTRODUCTION
This minireview discusses the protein-­coding transcripts (mRNAs) as-
sociated with rejection in transplant biopsies and the insight they pro-
vide into mechanisms and disease classification. We will not discuss
Abbreviations: ABMR, antibody-mediated rejection; AKI, acute kidney injury; ARTs, acute
rejection transcripts; AvEET, ABMR vs. everything else plus TCMR; cAR, clinical acute rejec-
tion; CRM, common rejection module; ETCs, effector T cells; ICR, immunological constant of
rejection; RTW, response to wounding; RvEEABMR, TCMR and mixed ABMR/TCMR rejec-
tion vs. everything else; TCMR, T cell-mediated rejection; TvEEA, TCMR vs. everything else
plus ABMR.
how molecular measurements are used for diagnosis, which is covered
elsewhere.1 The main focus is on kidney transplant biopsies because
the data sets are large and the rejection states are better character-
ized (by histology and by mRNA expression) than for other transplants
(as indicated by reproducibility of TCMR between pathologists: kidney
50%2; heart 28%3; and lung 18%4). In this review, we also compare
our findings in the Molecular Microscope® project1 to other published
literature, taking into account that many published biopsy studies used
different or outdated histology classifications and did not include or
recognize ABMR. Details are provided in recent reviews1 and in our
recent analysis of universal and selective rejection associations.5

Am J Transplant. 2018;18:785–795. © 2017 The American Society of Transplantation |  785

amjtransplant.com  
and the American Society of Transplant Surgeons
 |
786       HALLORAN et al.
Changes in mRNA expression that are associated with rejection
present opportunities to infer mechanisms, identify druggable targets,
develop new diagnostic tests, and correct errors in the empirical his-
tology system.6,7 TCMR itself is also a model system for understand-
ing human T cell-­mediated inflammatory diseases such as ulcerative
colitis and multiple sclerosis, and the mechanisms of tumor rejection
activated by checkpoint inhibitors.8 Associations are only one pathway
in developing evidence but they are an important first step in under-
standing diseases, just as associated findings are important clues in a
crime scene investigation.
Transcripts increased or decreased in biopsies with rejection may
reflect transcription changes in resident cells (eg, IFNG-­inducible
changes in the donor endothelium) or changes in cell populations, eg,
infiltration and activation of ETCs and macrophages in TCMR or mar-
gination and activation of intracapillary NK cells in ABMR. Therefore
we never use the terms “upregulated” and “downregulated,” as “in-
creased” or “decreased” is both more accurate and more descriptive.
Interpretation of transcript changes can be assisted by studying rele-
vant mouse models, cell lines, and the literature for each gene (https://
www.ncbi.nlm.nih.gov/gene).
The identification of the mRNAs associated with a conventionally
defined disease state depends on the accuracy of the designation of
the disease state. The true disease state always remains a “latent vari-
able”—one that is not directly observed but rather inferred by rules or
algorithms from other variables that are directly measured.
1.1 | Why study mRNA?
Transcripts encoding proteins can be amplified whereas their protein
products are often undetectable. Small non-­coding RNAs (eg, miRNAs
and siRNA) provide less biological insight than mRNAs because they
have less diversity by an order of magnitude and lack specificity for
individual genes due to their promiscuous regulation of immune and
non-­immune processes. To date, small non-­coding RNAs have shown
9-12
limited ability to define rejection classes.
2 |  DISCOVERING REJECTION-­
ASSOCIATED TRANSCRIPTS
2.1 | Genome-­wide discovery of mRNAs associated
with rejection
This review focuses on microarrays as the discovery platform. Another
genome-­wide platform is RNA sequencing (RNAseq),13 which has
great potential for discovery, but has not been assessed for system-
atic examination of large, well-­characterized transplant biopsy sets.
One report found that microarrays and RNAseq have the ability to
develop similar signatures of rejection in kidney transplant patients,
although the sample size was small and the depth of sequencing was
low.14 We estimate that RNAseq needs >60 million “mappable” reads
to capture the relatively low expression mRNAs that are most specific
for rejection (eg, CTLA4, IFNG), increasing costs and time compared to
microarrays. RNAseq is not suited for routine diagnostics but is ideal
for research and discovery as the alternative mRNA splicing and al-
ternative promoters add intriguing levels of complexity and potential
specificity that have not yet been explored. An alternative to RNAseq
is the new Clariom D microarray, which can identify >540 000 coding
and non-­coding transcripts.
The identity of the mRNAs associated with rejection is inferred
from the microarray probe set with which it hybridizes. This distinc-
tion is important when considering mRNA that may have been chem-
ically modified by formalin-­fixation and paraffin-­embedding (FFPE).
Formalin chemically modifies mRNA and produces both false positive
and false negative hybridization.15,16 The use of FFPE tissue for di-
agnostic assessment is being examined in some centers but it is not
the preferred method for studies discovering transcript associations,
which typically require biopsies stabilized in RNAlater or snap-­frozen
to avoid chemical modification.
The raw data from microarrays requires normalization of the data-
set, and in many cases the data is variance filtered (ie, probe sets with
low variance are removed) to reduce the “noise,” the potential contam-
ination by accidental differences. The risk in variance filtering is that
important low-­expression signals may be discarded. For example, in
studying TCMR, variance filtering initially missed CTLA, a transcript
with low signal strength but one of the most highly associated with
TCMR.2,17,18
We use large data sets that include many well-­defined TCMR
cases, both early and late-­stage ABMR cases (including C4d neg-
ative ABMR); and many other diseases and injuries. Identifying
TCMR-­ or ABMR-­selective transcripts requires inclusion of the op-
posite phenotype in the comparator.19 We generally study biopsies
for clinical indications (eg, dysfunction, proteinuria, investigation
of DSA) in consented patients in prospective trials (Clinicaltrials.
gov NCT01299168) because rejection phenotypes are much more
frequent and severe in indication biopsies than protocol biopsies
of stable patients, though conclusions are applicable to protocol
biopsies.20
2.2 | Analysis strategies for discovering
rejection-­associated transcripts
Transcripts associated with rejection are usually discovered in binary
class comparisons of a positive class (TCMR or ABMR, or all rejection)
versus a negative class, using t-­tests. The classes are usually assigned
by histology but they can also be assigned by molecular classifiers.
The classes do not need to be perfect to permit associations to be
discovered.
In comparing disease states, the biopsy classification of both
ABMR and TCMR becomes important because it affects both the
positive class and the negative class (see below). For example, if C4d-­
negative ABMR in kidney transplants is misclassified as non-­rejection,
its undetected presence in the negative class limits the definition of
transcripts selective for ABMR.
Gene lists can also be derived using correlations with molecular
rejection states or scores such as those assigned in unsupervised ar-
chetypal analysis.21
HALLORAN et al.
The top ranked rejection-­associated transcripts in one study pop-
ulation will not necessarily predict the top-­ranked transcripts in a new
population because rank is influenced by random error (“noise”) as well
as the prevalence of disease states (“signal”). The problem should be
considered whenever any “top gene” list is presented: the goal is to
discover strong associations with different disease states, but the top-­
ranked transcripts are only a sketch. Generally, the strong associations
of transcripts with a disease state are preserved when exact rank is
not because they reflect specific biologic mechanisms operating in all
similar study cohorts.
2.3 | Distinguishing rejection associations from
response-­to-­wounding (RTW)
Tissue injury sustained from donation and implantation, rejection, and
diseases affects all transplanted organs, even live-­donor organs, and
it evokes a RTW. New diseases or recurrence of primary disease also
induce the RTW. Thus the RTW (recently reviewed22) impacts expres-
sion of thousands of transcripts that are shared across various injury
states and disease (including rejection), and involves parenchymal
cells, stroma, and microcirculation endothelium, as well as recruitment
and activation of inflammatory cells. RTW is underestimated by his-
tology and evolves over months to years.23,24 Distinguishing rejection
from response-­to-­wounding (RTW) requires comparison of the posi-
tive class to a population that includes a wide variety of non-­rejection
disease states, as well as relatively normal tissue (“everything else”).
3 | REJECTION-­A SSOCIATED
TRANSCRIPTS: UNIVERSAL, TCMR-­
SELECTIVE, AND ABMR-­S ELECTIVE
We used a discovery set—validation set approach of prospectively
collected biopsies to document the rejection-­associated transcript
changes that are universal,5,25 TCMR-­selective,2,26 and ABMR-­
selective.17,27 The initial discovery sets had 403 biopsies and valida-
tion sets contained 300 independent biopsies.
Associations were visualized as “landscapes” and top associated
transcripts (ie, on the x axis) were used for pathway analyses.5,18,28
(We use the term landscape to refer to volcano plots that visualize
association strength on the x axis and fold change in expression on
the y axis.) Altered expression of thousands of transcripts—fold change
and association strength—was highly conserved between the discov-
ery and validation sets.
When we annotated the top transcripts associated with kidney
transplant rejection by t-­tests in 703 biopsies, the adjusted P-­values
show a hierarchy from universal (eg, 10−52 to 10−40) to TCMR-­selective
(10−41 to 10−31) to ABMR-­selective (10−20 to 10−9). Thus the associ-
ations are strongest for universal changes and weakest for ABMR-­
selective changes.
The published landscapes of the transcript changes in the univer-
sal, ABMR-­selective, and TCMR-­selective binary algorithms are repro-
duced in the volcano plots in Figure 1. It is important to recognize that
|
      787
the actual phenotype of each disease manifests the selective plus the
universal features. Thus, TCMR or ABMR can be annotated by their
selective features and rejection transcripts (Table 1).
3.1 | Universal rejection-­associated transcripts
This list is dominated by transcripts that are IFNG-­inducible in pa-
renchymal cells, endothelial cells, and/or in macrophages eg, CXCL9,
WARS, IDO, and HLA-­E. In addition, about 10% are shared by acti-
vated ETCs and NK cells, eg, KLRD1, CCL4 and PRF1. IFNG inducibil-
ity is the principal feature of the universal transcripts and IFNG itself
ranks at the top of the transcripts induced in activated ETCs and NK
cells.29 The ranks of top 100 universal rejection transcripts were high
in both ABMR-­selective and TCMR-­selective transcript lists. Pathway
analyses confirmed the prominence of IFNG-­STAT mechanisms.5
3.2 | TCMR-­selective transcripts
The top TCMR-­selective transcripts reflect effector T cell and mac-
rophage infiltration and activation. Effector T cell transcripts in-
clude IFNG; checkpoint molecules CTLA4, ICOS, and BTLA; CD96,
LAG3, SIRPG, LCP2, DUSP2, and CD8A. ANKRD22 (IFNG-­inducible
in macrophages) and non-­IFNG-­inducible macrophage molecule
ADAMDEC1 are highly TCMR-­selective. TCMR-­selective transcripts
rank high among the universal rejection transcripts (mean 596) but
low in ABMR-­selective transcripts (mean 33573).2,18,30 In pathway
analysis of TCMR-­selective transcripts, the top pathways included
T cell receptor signaling and CTLA4 costimulation. Other features of
myeloid cells include CD80 and CD86, CD274/PD-­1 ligand 1; CD84,
IL21R, LAIR1, PHEX, SLA, SLAMF8, TNFSF8/CD30.
3.3 | ABMR-­selective transcripts
The ABMR-­selective transcripts (eg, CCL4, CXCL11, CXCL10, and
PLA1A) rank high among the universal rejection transcripts, but low
in TCMR-­selective transcripts.17,27,28 We believe that the ABMR-­
selective transcripts have weaker associations (P-­values) than the
TCMR-­selective transcripts because the ABMR process is largely con-
fined to the capillary compartment and affects fewer cells, as opposed
to the TCMR process in the expanded interstitium. ABMR-­selective
transcripts are expressed in NK cells (SH2D1B, GNLY, FGFBP2,
CD160), endothelial cells (DARC, ROBO4, CDH5, CDH13), or are
IFNG-­inducible in endothelium (eg, PLA1A and CXCL11).
Some ABMR-­selective transcripts are decreased (eg, scle-
rostin [SOST] and corticotropin releasing hormone binding
27
protein [CRHBP] ). We postulate that the endothelial injury and de-­
differentiation that is not visible histologically or by electron micros-
copy. These decreases are strongly associated with late-­stage ABMR,
but it is unclear how they relate to the histologic changes of late-­stage
ABMR (eg, capillary basement membrane multi-­layering or glomerular
double contours).
Pathway analysis of ABMR-­selective transcripts identified
angiogenesis-­like changes with roles for angiopoietin and vascular
 |
788       HALLORAN et al.

F I G U R E   1   Molecular landscapes of TCMR (A), ABMR (B), and all rejection (C). Color keys (right lower quadrant) were assigned to the TCMR-­
selective, ABMR-­selective, and rejection-­associated transcripts (probe sets) as well as to the transcripts representing IFNG effects, B7 family
ligands and receptors, inflammasome, response to wounding, and parenchymal function

endothelial growth factors and leukocyte-­endothelial interactions, and
with evidence for NK CD16a Fc receptor signaling.1,29,31
ABMR-­associated transcripts in heart transplant endomyocar-
dial biopsies are very similar to a mix of the universal transcripts and
ABMR-­selective transcripts (eg, CXCL11, ROBO4) in kidney trans-
plants with ABMR.32,33
These results suggest models for TCMR and ABMR,1,34 summa-
rized here to encourage experimental validation. In TCMR, the cognate
effector T cells engage antigen on endothelium or antigen present-
ing cells (APC) prior to transendothelial migration35,36 and cross the
microcirculation without seriously damaging it. ETCs then engage the
APCs in the parenchyma and create an inflammatory compartment
dominated by the cognate engagement of T cell receptor and costim-
ulators with APCs (macrophages), regulated by inhibitors like CTLA4,
PD-­1, ICOS, and BTLA The inflammatory reaction in TCMR always
triggers a RTW in the parenchyma, analogous to acute kidney injury
(AKI). In ABMR, the NK cell CD16a Fc receptors (FCGR3A) engage
the Fc portions of IgG bound to HLA on the endothelium, activating
release of IFNG and other cytokines, potentially releasing cytotoxic
enzymes, and evoking an injury response in the endothelium. The mo-
lecular events in the CD16a-­activated NK cell inside the capillary are
probably similar to those in the T cell receptor-­activated effector T cell
in the interstitium, eg, IFNG production and CD160 expression.31 In
contrast to TCMR, ABMR initially induces more endothelial injury but
little parenchymal injury (which is why ABMR is often relatively silent
clinical problem).
3.4 | Comparing algorithms (Figure 2)
We compared the shared top transcripts for eight class comparisons
to interrogate microarray results from 703 kidney biopsies, 205 with
rejection. The positive comparators were all rejection, TCMR, or
ABMR. The negative comparators included all non-­rejecting biop-
sies (which include normal biopsies and other transplant diseases).
Selectivity for ABMR or TCMR (ie, generation of the ABMR-­ and
TCMR-­selective transcripts) required the other rejection class as well
as non-­rejection biopsies in the comparator to avoid selecting univer-
sal rejection transcripts.
When we generate subsets from our biopsy collection, the strong
associations are always conserved in the class comparisons but the
exact rank shows random variation. In one half of our 703 biopsy set,
the top 100 universal rejection transcripts in each algorithm will be
HALLORAN et al. |
      789

T A B L E   1   Most robust associations: Comparing universal (Ref. 5), ABMR-­selective (Ref. 27), and TCMR-­selective (Ref.17) top 30 transcripts*

Universal rejection-­associated TCMR-­selective (TvEEA)

Annotation of transcriptsa ABMR-­selective (AvEET) (N = 24) (RvEE) (N = 26) (N = 13)

IFNG-­ IFNG-­inducible CXCL10, CXCL11 CXCL9, CXCL10, CXCL11, —

inducible chemokines
IFNG-­inducible PLA1A APOL3, BTN3A3, FAM26F,GBP1, ANKRD22, APOL2, GBP5,
enzymes, small GBP2, GBP4, GBP5, IDO1, IRF1, RARRES3, LAP3
molecules PLA1A, WARS
IFNG-­inducible — HLA-DMA, HLA-DOA, HLA-DPA1, —
MHC-­related HLA-DRA, HLA-DRB1, HLA-E,
HLA-F, NLRC5, PSMB8, PSMB9,
TAP1, UBD
NK cell NK CD16a inducedb CCL4c, CD160, YME1L1 CCL4 —
expressed NK/T cytotoxic FGFBP2, GNLY PRF1 —
molecules
NK/T receptors and CX3CR1, KLRD1, KLRF1, KLRD1, NKG7 —
receptor signaling SH2D1B, TRDV3
Endothelial cells CDH5, CDH13, COL13A1, DARC, — —
ECSCR, GNG11, ICAM2, MALL,
PECAM1, PGM5, RAMP3,
RAPGEF5, ROBO4, TM4SF18,
VWF
Effector T T cell checkpoints — — CTLA4, ICOS
cell-­ Others — — CD96, IFNG, LAG3, SIRPG, BTLA,
expressed LCP2, DUSP2, MYB, CD8A,
IL12RB1
Myeloid Monocytes KLF4, PPM1F — AOAH, TNFSF8
cell-­ Macrophages — — CD86, CD274ADAMDEC1,CD84,
expressed FGD2, IL21R, LAIR1, PHEX, SLA,
SLAMF8
B cells, macrophages — — CD72

RvEE, rejection vs. everything else; AvEET, ABMR vs. everything else including TCMR; TvEEA, TCMR vs. everything else including ABMR; ABMR, antibody-­
mediated rejection; TCMR, T cell-­mediated rejection.
*Published table of top 30 transcripts in 703 biopsies assessed with U133 microarrays (Ref. 5).
a
Annotation by highest expression in a cell panel, supplemented by literature. Genes that overlap between conditions have underlined symbols. Transcripts
expressed by T cells and NK cells are assigned as NK transcripts in ABMR and T cell transcripts in TCMR. Transcripts expressed in T cells and myeloid cells
are considered T cell transcripts for this summary.
b
Induced by anti CD16a in cultured NK cells.
c
As noted in text, CCL4 is expressed in various cell types and inducible by IFNG treatment. CCL4 is expressed in T and NK cells and inducible in macro-
phages but did not qualify as GRIT by our definition.

only 50-­60% shared with the top 100 in the other half, yet all retain
strong statistical associations and represent the same biological pro-
cess. Again, this illustrates the role of noise in determining exact rank.
3.5 | Rejection-­associated transcripts are expressed
during injury-­induced RTW
Transcripts associated with rejection are never truly specific for re-
jection. Many top transcripts in ABMR, TCMR, and all rejection are
higher in abnormal non-­rejecting than normal transplants (Figure 3).
ABMR-­selective and endothelial transcripts are least specific since
they are high in TCMR and non-­rejection injuries. This is not surpris-
ing: the canonical histologic lesions of ABMR (ptc-­, g-­, cg-­ lesions) and
TCMR (t-­, i-­, and v-­lesions) are also not specific, occurring in other
diseases and injury.37
After four months post-­transplantation, kidney transplants often
manifest immunoglobulin transcripts (representing plasma cells)38 and
certain mast cell transcripts.39 These emerge in kidneys developing
atrophy-­fibrosis as part of the RTW, and should not be mistaken for
adaptive immune mechanisms.24
4 | COMPARISON TO PUBLISHED
LISTS OF ACUTE REJECTION-­
ASSOCIATED TRANSCRIPTS
Other published studies of rejection-­associated transcripts40–46 differ
in platforms, histologic classifications, case mix, and organs studied,
which complicates direct comparisons. We have related our rejection
transcripts to lists of transcripts associated with “acute rejection”:
 790       | HALLORAN et al.

ABMR = RvEE plus AvEET RvEE 100 TCMR = RvEE plus TvEEA

AvN 76
AvEE 70

TvN 45
AvEET 39 TvEE 37

TvEEA 9

TvA 0
Relatively selective for ABMR Relatively selective for TCMR

Degree of sharing vs. selectivity: based on sharing transcripts associated with all rejection

F I G U R E   2   Illustration of the effect of changing the case mix in the positive and negative comparators on transcript sharing by ten alternative
rejection-­associated algorithms. Numbers indicate the non-­redundant and annotated transcripts shared with rejection vs. everything else (RvEE)
algorithm, based on the top 100 rejection-­associated transcripts. TvN -­ TCMR vs. Nephrectomies, TvEE -­ TCMR vs. everything else (excluding
ABMR), TvEET -­ TCMR vs. everything else including ABMR, TvA -­ TCMR vs. ABMR, AvEET -­ ABMR vs. everything else including TCMR, AvEE -­
ABMR vs. everything else (excluding TCMR), AvN -­ ABMR vs. Nephrectomies

ANOVA p<0.001
3.0
2.5
Mean expression of ABMR genes
(fold change vs. normal kidneys)

2.0

F I G U R E   3   The mean expression of the

1.5

top ABMR genes in rejecting and non-­

rejecting biopsies (previously published
1.0

in Ref.5). The Y scale values are the fold

changes vs. normal kidneys. The mean
0.5

expression of the top 30 ABMR transcripts

is highest in ABMR and very low in normal
kidneys, but is also increased in TCMR,
0.0

atrophy-­fibrosis (IFTA), some kidneys with

AKI, and even some normal transplants.
−0.5

Symbols indicate individual biopsies, boxes

ABMR AKI IFTA normal transplant TCMR normal kidney
and whiskers show the interquartile range
Histologic diagnosis (25-­75%) and 1.5 × Standard Deviation

liver acute rejection47; Common Rejection Module (CRM)43; Acute
48
Rejection Transcripts in kidney (ARTs) ; and clinical Acute Rejection
(cAR) in kidney.40 The CRM and the Immunologic Constant of
49,50
Rejection (ICR) were derived from multiple published lists and
across organs. All reflect class comparisons of transcript changes in
biopsies with acute rejection versus non-­rejection. The histologic
classification of rejection in these studies is outdated, and the case
mix is not clear: most biopsies were early when ABMR is infrequent
(in patients transplanted with no DSA), and were classified without
current criteria for ABMR. From the above consideration, these lists
should primarily detect the transcripts characterized in our universal
rejection transcripts and TCMR-­selective transcripts algorithms.5 Our
approach was to relate these acute rejection transcript lists (rank and
P-­values) to our universal rejection transcripts, TCMR-­selective, and
ABMR-­selective transcripts.5
The acute rejection transcripts in liver, CRM, ARTs, and cAR all
were highly enriched in universal rejection transcripts, eg, CXCL9,
CXCL10, PSMB9, TAP, UBD, and HLA. For example, 46 of 67 ARTs rank
in top 300 universal rejection transcripts, with 26 annotated as IFNG-­
inducible (Table 2). The median rank was lower in TCMR (Table 2)
because TCMR selectivity requires ABMR to be annotated and well
represented in the comparator.
T A B L E   2   Comparison of published Acute Rejection transcripts to the RvEE transcripts ordered by their rank in RvEE comparisona

Liver acute rejection (Ref. 47) Common rejection module across

n = 19 organs (CRM, Ref. 43) n = 12 Acute rejection transcripts in kidney (ARTS) and rank in the RvEE landscape by P-­value (Ref. 48) n = 67
HALLORAN et al.

Gene symbol Rank in RvEEb Gene symbol Rank in RvEE Gene symbol Rank in RvEE Gene symbol Rank in RvEE Gene symbol Rank in RvEE

CXCL10c 7 CXCL10 7 WARS 4 FCER1G 109 T3JAM (TRAF3IP3) 388

CXCL9 12 PSMB9 10 CXCL10 7 SLA 111 CD53 457
TAP1 18 CXCL9 12 GBP1 9 ISG20 112 LYZ 466
HLA-F 20 TAP1 18 PSMB9 10 GZMA 119 CASP4 476
HLA-DMA 31 NKG7 42 CXCL9 12 HLA-DPA1 142 LCK 335
HLA-DRA 41 ISG20 112 TAP1 18 CD8A 142 NMI 498
UBD 43 RUNX3 191 HLA-F 20 CCL5d 145 IFITM1 707
HLA-DQB1 51 INPP5D 325 HLA-DRB3 23 GMFG 146 ARPC2 709
CD2 274 LCK 335 HLA-DMA 31 HLA-DQB1 158 CD163 814
RFX5 334 CD7 1002 GBP2 35 PLEK 182 TCIRG1 838
SLC1A3 994 CD6 1277 PSMB8 36 HCLS1 202 TNFRSF7 (CD27) 852
IL32 1352 BASP1 2283 HLA-DRA 41 STAT1 204 SERPING1 880
ITM2A 1611 UBD 43 CD48 216 TIMP1 1429
ANXA2 1664 HLA-E 46 LCP1 226 CD44 1469
PLA2G7 2409 HLA-B 61 ARHGDIB 233 FER1L3 (MYOF) 1619
FABP5 7001 HLA-C 62 RAC2 275 HLA-DQB2e 1724
IL8 12074 HLA-G 64 RUNX3 280 IGHM 2539
GPNMB 25510 PSMB10 77 PRKCB1 (PRKCB) 295 PLSCR1 2547
CD83 54237 CASP1 81 ITGB2 299 PRG1 (TPRG1L) 3251
UBE2L6 87 IFI30 328 CSPG2 (VCAN) 3797
IL10RA 101 LAPTM5 353 TNC 5311
HLA-DMB 105 CD52 368 LTF 6750
HCK 108 LCP2 378 MMP7 9006
WFDC2 28888
Median rank in RvEE 334 Median rank in RvEE 151 Median rank in RvEE 210
Median rank in TvEEA 578 Median rank in TvEEA 336 Median rank in TvEEA 390
Median rank in AvEET 1148 Median rank in AvEET 584 Median rank in AvEET 1564

Underlined are transcripts overlapping with the union of rejection transcripts in kidneys (Ref. 5).
RvEE, rejection vs. everything else, TvEEA, TCMR vs. everything else including ABMR, AvEET, ABMR vs. everything else including TCMR.
a
Genes in the Acute Rejection (AR) data sets were selected by comparing biopsies with AR to biopsies from well-­functioning transplants.
b
Order by their rank in the RvEE landscape by corrected P-­value.
c
Bolded gene symbols indicate IFNG-­inducible (http://atagc.med.ualberta.ca/Research/GeneLists/Pages/default.aspx).
|

d
CCL5 expression is inducible by IFNG but is also highly expressed in effector T cells.
e
      791

HLA-DQB2 expression was very low across the cell cultures.

 |
792       HALLORAN et al.

T A B L E   3   Ranking of the top 50 increased transcripts (out of 2653) from the cAR (internal) transcript list (Ref. 40) in the RvEE and TvEEA
comparisons ordered by their rank in the RvEE algorithm

cAR internal

Gene symbol Rank in cAR Rank in RvEEb Rank in TvEEAc Gene symbol Rank in cAR Rank in RvEE Rank in TvEEA

CCL4 18 2 794 BCL11B 48 530 401

KLRD1 9 6 584 IL7R 26 564 974
GBP1a 7 9 91 RGS18 33 567 1009
CXCL9 13 12 480 FPR1 3 643 82
IRF1 10 30 169 EMB 49 755 995
CTSS 21 57 392 LILRB1 24 787 210
CST7 19 110 524 EGR2 45 818 771
GZMA 35 119 351 BOD1L 5 841 2007
FCGR1B 44 130 60 RAB27A 47 898 428
SLAMF8 12 137 1 SERPINE1 41 1209 2195
CD8A 40 141 29 RARRES1 22 1418 1717
ITGAL 50 147 272 CXCL13 15 1930 201
HLA-DQB1 38 158 1213 CYAT1(IGLV1-44) 46 2314 18770
PTPRC 34 252 397 C3 2 2435 2100
MS4A6A 8 253 86 LOC100287723 32 2775 38707
NCF2 27 261 419 GALNT6 29 3324 5291
RNASE6 23 302 597 TSPO 43 4490 2243
GZMK 20 304 288 TNFRSF17 6 5208 19499
MYO1G 42 333 743 CCL18 25 6457 70
LAIR1 36 356 38 MOXD1 28 10697 25876
ZAP70 11 370 170 FGB 17 16246 30784
EVI2A 4 384 761 THBS2 16 16450 7420
FLJ32255 1 448 567 SFN 31 16528 12122
IGSF6 37 487 93 RRAD 30 30896 29018
MNDA 14 515 804 IGLL1/IGLL3 39 48382 34564

Median ranking in AvEET (the ABMR-­selective transcripts) was 5214.

Underlined are transcripts overlapping with the union of rejection genes in kidneys (Ref. 5).
RvEE, rejection vs. everything else; TvEEA, TCMR vs. everything else (including ABMR).
a
Bolded gene symbols indicate IFNG-­inducible (http://atagc.med.ualberta.ca/Research/GeneLists/Pages/default.aspx).
b
Rank in the RvEE comparison by corrected P-­value, median rank 522.
c
Rank in the TvEEA comparison by corrected P-­value, median rank 590.

The top 50 cAR transcripts (by P-­value) in the original supple-
mentary material40 included our universal rejection transcripts CCL4,
KLRD1, GBP1, and CXCL9 (ranks 2-­12) (Table 3). Their median rank in
our all rejection data set was 522, and in TCMR was 590, indicating
universal/selective mix with essentially no ABMR-­selectivity.
For the ICR,50 of 263 transcripts that mapped to HG 133 2.0 Plus
Affymetrix arrays, 81 overlapped our top 300 universal rejection
transcripts. The ICR universal rejection transcripts included CXCL11,
CCL4, WARS, CXCL10, and CXCL9 (ranks 1-­12) and 49 IFNG-­induced
transcripts.
Overall, these previously published gene lists reflect universal
and TCMR-­selective transcripts, (Table 4). The liver acute rejection,
CRM, and ARTs lists had stronger associations with universal rejection
(median P-­values) than with TCMR (nine to six orders of magnitude).
Interestingly, the cAR and ICR lists had similar median P-­values for
universal rejection transcripts and TCMR-­selective transcripts, al-
though median ICR P-­values were weaker. The median P-­values of
ABMR-­selective transcripts were poor for all of the acute rejection
transcript lists. Thus published lists of rejection transcripts resemble
our universal rejection transcripts and TCMR-­selective transcripts, yet
lack selectivity for ABMR.
5 | RELEVANCE TO
OTHER ORGAN TRANSPLANTS
The mRNAs associated TCMR and ABMR in kidney transplants are
highly conserved in heart transplants,32,33 giving us confidence to use
HALLORAN et al.       793|
T A B L E   4   Median rank in the ATAGC
Acute rejection data RvEE median P TvEEA median P AvEET median P
703 dataset of transcripts annotated in
set value* (range) value* (range) value* (range)
other published acute rejection data sets
Liver acute rejection 1.8 × 10−27 2.5 × 10−21 0.03 (3.0 × 10−09-­0.98)
(n = 19) (1.2 × 10−47-­0.98) (3.2 × 10−31-­0.30)
Common rejection 2.2 × 10−32 1.2 × 10−23 3.8 × 10−03
module (CRM) (1.2 × 10−47-­ (3.2 × 10−31-­ (3.0 × 10−09-­0.72)
(n = 12) 9.2 × 10−10) 2.6 × 10−12)
ARTs study (n = 67) 2.9 × 10−32 3.0 × 10−23 9.0 × 10−3
(3.4 × 10−36-­0.04) (3.1 × 10−25-­0.03) (3.0 × 10−9-­0.53)
CAR study (internal) 1.4 × 10−23 3.5 × 10−21 0.04 (6.1 × 10−12-­0.91)
(n = 50)** (1.3 × 10−36-­0.64) (3.09 × 10−41-­0.35)
Immunological 5.1 × 10−16 5.3 × 10−15 0.07 (8.0 × 10−19-­0.99)
constant of rejection (3.6 × 10−52-­0.54) (9.6 × 10−41-­1.0)
(n = 263)***

RvEE, Rejection versus everything else (non rejecting); TvEEA, TCMR vs. everything else (including
ABMR); AvEET, ABMR vs. everything else (including TCMR).
*Association with rejection (P values with Benjamini-­Hochberg correction) in RvEE comparison.
**The top 50 transcripts selected by P values from the original publication were used.
***number of transcripts that mapped to HG133 2.0 Plus Affymetrix arrays.

T A B L E   5   Lessons about the transcripts associated with rejection*

Algorithms Type Description and their characteristics

Three main “universal” (RvEE) most significant P values; IFNG effects and shared effector T/NK cell molecules
algorithms “TCMR-­selective” (TvEEA) activated effector T cell and macrophage transcripts; selected IFNG effects
“ABMR-­selective” (AvEET) least significant P values; IFNG effects, endothelial injury response; NK cells
Associations and
selectivity Description Comments
Four factors related algorithm, case mix, histology Strong associations are maintained if the four factors are maintained. Exact rank will
to associations: classification system, and differ in otherwise similar populations due to random variation between popula-
measurement platform tions (noise).
Other associations associations with response to Rejection associations are shared with response to wounding: not specific.
wounding and across organs Considerable conservation of rejection-­associated transcripts across organs.
Selectivity Selectivity for ABMR or TCMR If ABMR is not included and well characterized, class comparisons regress toward
requires inclusion of the opposite RvEE. Need selectivity for stages in ABMR: early-­stage, fully-­developed, late-­stage.
class and of EE in the comparator.

ABMR, antibody-mediated rejection; TCMR, T cell-mediated rejection; EE, everything else; RvEE, TCMR and ABMR vs. everything else (non rejecting);
TvEEA, TCMR vs. everything else plus ABMR; AvEET, ABMR vs. everything else plus TCMR.
*
Usually annotated by P value in binary class comparison: positive class vs. negative class. An alternative is to use correlations with molecular scores for each
biopsy.

the kidney results as a basis for defining rejection in other organs. For
liver TCMR this is also true as we have seen in the liver acute rejec-
47
tion study above. We expect this will be the case in lung transplant
ABMR and TCMR and for liver transplant ABMR, but the histological
rejection phenotypes are so poorly defined in lung (and ABMR in liver)
that this will take time to develop.
6 | PRINCIPLES OF MOLECULAR
ASSOCIATION STUDIES
Some principles emerging from our analyses are listed in Table 5. Ranking
of probe sets varies with the positive class and negative class, ie, case
mix matters for label classes. Selectivity for TCMR or ABMR requires
having large numbers of the opposite rejection phenotype included in
the negative class and proper histologic annotation, so that TCMR and
ABMR do not contaminate each other. Otherwise, the selectivity for
ABMR or TCMR decreases when the negative class does not include
non-­rejection biopsies and the opposite rejection class. The biology
represented by the top transcripts changes with case mix in the nega-
tive class: for example, for ABMR-­selectivity, endothelial changes only
emerge as top transcripts when TCMR and EE are in the negative class.
When rejection transcript measurements are used in molecular di-
agnostics, they should be incorporated into machine-­learning derived
classifiers, which increase their accuracy compared to simple measure-
ments and cut-­offs. Machine learning can assign disease diagnoses
using existing classifications but can also be used to discovery new
classifications using clustering methods such as archetype analysis.21
 |
794       HALLORAN et al.
ACKNOWLE DG ME NTS
This research has been supported by funding and/or resources from
the Industrial Research Assistance Program (IRAP) of the National
Research Council of Canada;, Canada Foundation for Innovation,
the University of Alberta Hospital Foundation, Roche Molecular
Systems, Hoffmann-­La Roche Canada Ltd., Genome Canada, the
Alberta Ministry of Advanced Education and Technology, the Roche
Organ Transplant Research Foundation, and Astellas. Dr. Halloran
holds the Muttart Chair in Clinical Immunology.
D ISCLOSURE
The authors of this manuscript have conflicts of interest to disclose
as described by the American Journal of Transplantation. P.F. Halloran
holds shares in Transcriptome Sciences Inc., a company with an inter-
est in molecular diagnostics. The other authors have no conflicts of
interest to disclose.
REFERENCES
1. Halloran PF, Famulski KS, Reeve J. Molecular assessment of
disease states in kidney transplant biopsies. Nat Rev Nephrol.
2016;12(9):534‐548.
2. Reeve J, Sellares J, Mengel M, et al. Molecular diagnosis of T cell-­
mediated rejection in human kidney transplant biopsies. Am J
Transplant. 2013;13(3):645‐655.
3. Crespo-Leiro MG, Zuckermann A, Bara C, et al. Concordance among
pathologists in the second cardiac allograft rejection gene expression
observational study (CARGO II). Transplant. 2012;94(11):1172‐1177.
4. Arcasoy SM, Berry G, Marboe CC, et al. Pathologic interpretation of
transbronchial biopsy for acute rejection of lung allograft is highly
variable. Am J Transplant. 2011;11(2):320‐328.
5. Halloran PF, Venner JM, Famulski KS. Comprehensive analysis of tran-
script changes associated with allograft rejection: Combining univer-
sal and selective features. Am J Transplant. 2017;17(7):1754‐1769.
6. Halloran PF, Merino LM, Barreto PA. Identifying subphenotypes of
antibody-­mediated rejection in kidney transplants. Am J Transplant.
2016;16(3):908‐920.
7. Reeve J, Chang J, Salazar IDR, Lopez MM, Halloran PF. Using molecu-
lar phenotyping to guide improvements in the histologic diagnosis of
T cell-­mediated rejection. Am J Transplant. 2016;16(4):1183‐1192.
8. Sharma P, Allison JP. The future of immune checkpoint therapy.
Science. 2015;348(6230):56‐61.
9. Vitalone MJ, Sigdel TK, Salomonis N, Sarwal RD, Hsieh SC, Sarwal
MM. Transcriptional perturbations in graft rejection. Transplant.
2015;99(9):1882‐1893.
10. Wilflingseder J, Regele H, Perco P, et al. miRNA profiling discriminates
types of rejection and injury in human renal allografts. Transplant.
2013;95(6):835‐841.
11. Mas VR, Dumur CI, Scian MJ, Gehrau RC, Maluf DG. MicroRNAs
as biomarkers in solid organ transplantation. Am J Transplant.
2013;13(1):11‐19.
12. Scian MJ, Maluf DG, David KG, et al. MicroRNA profiles in allograft
tissues and paired urines associate with chronic allograft dysfunction
with IF/TA. Am J Transplant. 2011;11(10):2110‐2122.
13. Mimura I, Kanki Y, Kodama T, Nangaku M. Revolution of nephrology
research by deep sequencing: ChIP-­seq and RNA-­seq. Kidney Int.
2014;85(1):31‐38.
14. Kurian SM, Velazquez E, Thompson R, et al. Orthogonal comparison
of molecular signatures of kidney transplants with subclinical and
clinical acute rejection: equivalent performance is agnostic to both
technology and platform. Am J Transplant. 2017;17(8):2103‐2116.
15. Masuda N, Ohnishi T, Kawamoto S, Monden M, Okubo K. Analysis of
chemical modification of RNA from formalin-­fixed samples and opti-
mization of molecular biology applications for such samples. Nucleic
Acids Res. 1999;27(22):4436‐4443.
16. Farragher SM, Tanney A, Kennedy RD, Paul HD. RNA expression anal-
ysis from formalin fixed paraffin embedded tissues. Histochem Cell
Biol. 2008;130(3):435‐445.
17. Halloran PF, Pereira AB, Chang J, et al. Microarray diagnosis of
antibody-­mediated rejection in kidney transplant biopsies: an in-
ternational prospective study (INTERCOM). Am J Transplant.
2013;13(11):2865‐2874.
18. Venner JM, Famulski K, Badr D, Hidalgo LG, Chang J, Halloran PF.
Molecular landscape of T cell-­mediated rejection in human kidney
transplants: prominence of CTLA4 and PD ligands. Am J Transplant.
2014;14(11):2565‐2576.
19. Loupy A, Haas M, Solez K, et al. The Banff 2015 kidney meeting
report: current challenges in rejection classification and prospects
for adopting molecular pathology. Am J Transplant. 2017;17(1):
28‐41.
20. Mengel M, Chang J, Kayser D, et al. The molecular phenotype of
six-­week protocol biopsies from human renal allografts: reflections
of prior injury but not future course. Am J Transplant. 2011;11(4):
708‐718.
21. Reeve J, Bohmig GA, Eskandary F, et al. Assessing rejection-­related
disease in kidney transplant biopsies based on archetypal analysis of
molecular phenotypes. JCI Insight. 2017;2(12):pii 94197
22. Kreepala C, Famulski KS, Halloran PF. Fundamental Concepts
Regarding Graft Injury and Regeneration: Tissue Injury, Tissue Quality
and Recipient Factors. In: Kirk AD, Knechtle SJ, Larsen CP, Madsen JC,
Pearson TC, Webber SA, eds. Textbook of Organ Transplantation, 1st
edn. Hoboken, NJ: John Wiley & Sons; 2014:99‐118.
23. Famulski KS, de Freitas DG, Kreepala C, et al. Molecular pheno-
types of acute kidney injury in human kidney transplants. JASN.
2012;23(5):948‐958.
24. Venner JM, Famulski KS, Reeve J, Chang J, Halloran PF. Relationships
among injury, fibrosis, and time in human kidney transplants. JCI
Insight. 2016;1(1):e85323.
25. Reeve J, Einecke G, Mengel M, et al. Diagnosing rejection in renal
transplants: a comparison of molecular-­ and histopathology-­based
approaches. Am J Transplant. 2009;9(8):1802‐1810.
26. Halloran PF, Pereira AB, Chang J, et al. Potential impact of microar-
ray diagnosis of T cell-­mediated rejection in kidney transplants: the
INTERCOM study. Am J Transplant. 2013;13(9):2352‐2363.
27. Sellares J, Reeve J, Loupy A, et al. Molecular diagnosis of antibody-­
mediated rejection in human kidney transplants. Am J Transplant.
2013;13(4):971‐983.
28. Venner JM, Hidalgo LG, Famulski KS, Chang J, Halloran PF. The mo-
lecular landscape of antibody-­mediated kidney transplant rejection:
evidence for NK involvement through CD16a Fc receptors. Am J
Transplant. 2015;15(5):1336‐1348.
29. Parkes MD, Halloran PF, Hidalgo LG. Direct evidence for CD16a-­
mediated NK cell stimulation in antibody-­mediated kidney transplant
rejection. Transplant. 2017;101(4):e102‐e111.
30. Halloran PF, Chang J, Famulski K, et al. Disappearance of T cell-­
mediated rejection despite continued antibody-­mediated rejection in
late kidney transplant recipients. JASN. 2015;26(7):1711‐1720.
31. Parkes MD, Halloran PF, Hidalgo LG. Mechanistic sharing between
NK cells in ABMR and effector T cells in TCMR. Am J Transplant.
2017;00:1‐11. https://doi.org/10.1111/ajt.14410
32. Halloran PF, Potena L, Duong Van Huyen JP, et al. Building a tissue-­
based molecular diagnostic system in heart transplant rejection:
the heart molecular microscope MMDx. J Heart Lung Transplant.
2017;36(11):1192‐1200.
HALLORAN et al.
33. Loupy A, Duong Van Huyen JP, Hidalgo LG, et al. Gene expression
profiling for the identification and classification of antibody-­mediated
heart rejection. Circulation. 2017;135(10):917‐935.
34. Reeve J, Halloran PF, Kaplan B. Common errors in the imple-
mentation and interpretation of microarray studies. Transplant.
2015;99(3):470‐475.
35. Walch JM, Zeng Q, Li Q, et al. Cognate antigen directs CD8  +
T cell migration to vascularized transplants. J Clin Invest.
2013;123(6):2663‐2671.
36. Zhuang Q, Liu Q, Divito SJ, et al. Graft-­infiltrating host dendritic
cells play a key role in organ transplant rejection. Nat Commun.
2016;24(7):12623.
37. Salazar IDR, Lopez MM, Chang J, Halloran PF. Reassessing the sig-
nificance of v-­lesions in kidney transplant biopsies. J Am Soc Nephrol.
2015;26(12):3190‐3198.
38. Einecke G, Reeve J, Mengel M, et al. Expression of B cell and immuno-
globulin transcripts is a feature of inflammation in late allografts. Am J
Transplant. 2008;8(7):1434‐1443.
39. Mengel M, Reeve J, Bunnag S, et al. Molecular correlates of scarring
in kidney transplants: the emergence of mast cell transcripts. Am J
Transplant. 2009;9(1):169‐178.
40. Modena BD, Kurian SM, Gaber LW, et al. Gene expression in biop-
sies of acute rejection and interstitial fibrosis/tubular atrophy reveals
highly shared mechanisms that correlate with worse long-­term out-
comes. Am J Transplant. 2016;16(7):1982‐1998.
41. Hruba P, Brabcova I, Gueler F, et al. Molecular diagnostics identifies
risks for graft dysfunction despite borderline histologic changes.
Kidney Int. 2015;88(4):785‐795.
42. Broin O, Hayde N, Bao Y, et al. A pathogenesis-­based transcript sig-
nature in donor-­specific antibody-­positive kidney transplant patients
with normal biopsies. Genom Data. 2014;2:357‐360.
43. Khatri P, Roedder S, Kimura N, et al. A common rejection module
(CRM) for acute rejection across multiple organs identifies novel
|
      795
therapeutics for organ transplantation. J Exp Med. 2013;210(11):
2205‐2221.
44. Dean PG, Park WD, Cornell LD, Gloor JM, Stegall MD. Intragraft gene
expression in positive crossmatch kidney allografts: ongoing inflam-
mation mediates chronic antibody-­mediated injury. Am J Transplant.
2012;12(6):1551‐1563.
45. Rodder S, Scherer A, Korner M, Marti HP. A subset of metzincins and
related genes constitutes a marker of human solid organ fibrosis.
Virchows Arch. 2011;458(4):487‐496.
46. Rodder S, Scherer A, Korner M, et al. Meta-­analyses qualify metz-
incins and related genes as acute rejection markers in renal transplant
patients. Am J Transplant. 2010;10(2):286‐297.
47. Bonaccorsi-Riani E, Pennycuick A, Londono MC, et al. Molecular char-
acterization of acute cellular rejection occurring during intentional im-
munosuppression withdrawal in liver transplantation. Am J Transplant.
2016;16(2):484‐496.
48. Saint-Mezard P, Berthier CC, Zhang H, et al. Analysis of independent
microarray datasets of renal biopsies identifies a robust transcript sig-
nature of acute allograft rejection. Transpl Int. 2009;22(3):293‐302.
49. Wang E, Worschech A, Marincola FM. The immunologic constant of
rejection. Trends Immunol. 2008;29(6):256‐262.
50. Spivey TL, Uccellini L, Ascierto ML, et al. Gene expression profiling
in acute allograft rejection: challenging the immunologic constant of
rejection hypothesis. J Transl Med. 2011;9:174.
How to cite this article: Halloran PF, Venner JM, Madill-
Thomsen KS, et al. Review: The transcripts associated with
organ allograft rejection. Am J Transplant. 2018;18:785–795.
https://doi.org/10.1111/ajt.14600