Biol Blood Marrow Transplant 22 (2016) 1552e1564

Biology of Blood and

Marrow Transplantation
journal homepage: www.bbmt.org

The Role of Biomarkers in the Diagnosis and Risk Stratification

of Acute Graft-versus-Host Disease: A Systematic Review
Alaa M. Ali, John F. DiPersio, Mark A. Schroeder*
Washington University School of Medicine, Department of Medicine, St. Louis, Missouri

Article history: a b s t r a c t
Received 21 December 2015 Allogeneic hematopoietic cell transplantation (HCT) is an increasingly used curative modality for hematologic
Accepted 26 April 2016 malignancies and other benign conditions. Attempts to reduce morbidity and mortality and improve survival
in patients undergoing HCT are crucial. The ability to diagnose acute graft-versus-host disease (aGVHD) in a
Key Words: timely manner, or to even predict aGVHD before clinical manifestations, along with the accurate stratification
Graft-versus-host disease of these patients, are critical steps to improve the treatment and outcomes of these patients. Many novel
Hematopoietic stem cell biomarkers that may help achieve these goals have been studied recently. This overview is intended to assist
transplantation
clinicians and investigators by providing a comprehensive review and analytical interpretation of the current
Biomarker
knowledge concerning aGVHD and biomarkers likely to prove useful in diagnosis and risk stratification of this
Proteomics
condition, along with the difficulties that hamper this approach.
Ó 2016 American Society for Blood and Marrow Transplantation.

INTRODUCTION made by combining the clinical impression, high pretest

Acute graft-versus-host disease (aGVHD) is a major likelihood of aGVHD, exclusion of other competing disorders,
complication of allogeneic hematopoietic cell trans- along with histological examination of the target tissue. Bi-
plantation (allo-HCT), occurring in up to 60% of transplant opsy alone is not the gold standard for diagnosis, owing to the
recipients, with well-described risk factors, including HLA false-negative results (patchiness of the disease, absence of
mismatch, age, stem cell source, donorerecipient sex the typical changes at early stages) and false-positive results
disparity, and conditioning regimen [1]. aGVHD is a systemic (residual conditioning regimen toxicity, infections) [3-5].
disorder driven by donor T cells with a pleomorphic clinical Furthermore, in many patients other causes of symptoms may
presentation involving multiple target organs, including the be found, confounding the diagnosis of aGVHD [8-12].
skin, liver, and gastrointestinal (GI) tract [2]. The diagnosis of Once diagnosed, the severity of aGvHD has historically
aGVHD with clinical and pathological confirmation is helpful been graded using the Keystone Consensus criteria [13] or
but lacks positive predictive value (PPV) and negative pre- CIBMTR criteria [14]. Grading was primarily developed as an
dictive value (NPV) [3-5]. important tool to determine the appropriate management of
Immunosuppressive therapy with steroids is the first-line aGvHD and to assess the response to therapy. Grading is also
therapy for clinically significant graft-versus-host disease important due to its impact upon survival and association
(GVHD). Mortality and morbidity from high-dose systemic with graft-versus-leukemia effect. It has been well recog-
immunosuppression is significant, and no other treatment nized that current grading systems overgrade some patients
added to upfront steroids has proven beneficial to date [6]. with a high likelihood of responding to immunosuppressive
Although the clinical diagnosis can be readily made in therapy and can not predict who will respond to steroids
patients with a classical presentation, some cases prove with certainty [15]. There are several limitations to current
challenging [7]. Currently, the diagnosis of aGVHD can be prognostic grading systems: despite the general relationship
to outcomes, inter-observer errors can occur due to subjec-
tive biases, initial grade may not reflect peak grade and the
time to response after therapy initiation is not accounted for.
Financial disclosure: See Acknowledgments on page 1562. Therefore, many patients who are classified as standard-risk
* Correspondence and reprint requests: Mark A. Schroeder, MD, Division
of Oncology, Washington University School of Medicine, 660 S Euclid Ave,
have their treatment fail while others classified as high- risk
Campus Box 8007, St. Louis, MO 63110. are over-treated. Recently, a refined clinical grading system
E-mail address: [email protected] (M.A. Schroeder). was introduced as a better tool not only to stratify patient’s

http://dx.doi.org/10.1016/j.bbmt.2016.04.022
1083-8791/Ó 2016 American Society for Blood and Marrow Transplantation.
 A.M. Ali et al. / Biol Blood Marrow Transplant 22 (2016) 1552e1564 1553

non-relapse mortality risk but also predict response to more accurate diagnosis and risk stratification of patients
therapy with high risk patients less likely to respond to with aGVHD. Recent research has applied proteomics tech-
steroids compared to standard risk patients [15]. However, nologies to identify aGVHD biomarkers. This has led, in a short
this refined scoring system could still be plagued by inter- period, to the identification of novel biological pathways and
observer biases with significant variability across BMT cen- biomarkers predictive of and associated with aGVHD [23].
ters [16]. It is hoped that the use of biomarkers can add to Nevertheless, no single biomarker or panel of biomarkers has
prediction accuracy and eliminate some of these problems. been validated for clinical use via large multicenter trials. In
High-dose systemic steroids are the standard first-line this article, we summarize the current knowledge of prom-
therapy for patients with grade II or higher aGVHD [6,17]; ising diagnostic and prognostic aGVHD biomarkers and
however, practices differ among institutions with respect to analyze the supporting data available in the literature.
initial steroid dose, the use of additional immunosuppressive
agents, and the approach to steroid tapering after initial REVIEW DESIGN
response. In general, therapy of clinically suspected aGVHD is We searched PubMed and MEDLINE up to December 31,
started before diagnosis is confirmed and potentially before 2015, to identify studies evaluating biomarkers in the setting
peak grade is achieved. Such an approach may expose the of aGVHD. Each biomarker (ie, micro RNA [miRNA], ST2, TNF
already immunosuppressed patients to unnecessary sys- receptor 1 [TNFR1], IL-7, sBAFF, REG3a, S100, TIM-3, CK-18,
temic steroids, with significant infectious and noninfectious hepatocyte growth factor [HGF], and elafin) was searched
complications. In fact, a study conducted to reduce the dose separately as well. Only full-text articles published in English
of initial steroid therapy for grade I-II acute GVHD found no were considered. The primary search was conducted using the
difference in outcomes between standard dose (2 mg/kg) terms “graft-versus-host disease” and “biomarker,” excluding
and reduced dose (1 mg/kg) arms [18]. Approximately one- reviews. Relevant references in the publications identified
half of patients have a complete response to steroids by were reviewed as well. Eligible studies included clinical
day 28 after therapy initiation [19]. studies with more than 5 patients. Studies investigating the
Steroid-refractory aGVHD is associated with high diagnostic and prognostic value of transcriptomic and prote-
transplantation-related mortality (TRM) and low overall omic biomarkers were reviewed. Here we discuss biomarkers
survival, even without relapse of the underlying disease that that were evaluated in at least 2 independent studies. The
necessitated HCT [20]. There are no standard second-line primary statistical outcomes were sensitivity, specificity, PPV,
therapies for steroid-refractory aGVHD, and responses vary, NPV, area under the curve (AUC), and hazard ratio (HR). The
but based on small retrospective and phase I/II studies, main outcomes of the remaining preclinical and clinical
responses are around 30% to 50% with a 6-month NRM of studies were reported in a table but not discussed in the text.
approximately 50% [6]. A biomarker-based grading system
may have the potential to more accurately stratify patients DIAGNOSTIC AND PROGNOSTIC BIOMARKERS
based on risk of failure to respond to steroids or alternative The biomarkers of aGVHD that have been discovered so
treatment approaches, and may be used to help identify far can be classified in various ways, including target (sys-
additional lines of therapy in those that fail upfront therapy. temic or particular organ/tissue) and biophysical properties.
Although mortality related to GVHD has been reduced in Systemic biomarkers lack organ specificity for skin or GI
recent years [21,22], aGVHD remains a major cause of TRM. aGVHD. These biomarkers rise in response to systemic injury
aGVHD represents the primary limitation to more wide- rather than to specific tissue damage. On the other hand,
spread use of allogeneic HCT as a potentially curative organ-specific biomarkers are expressed by target organs
modality for patients with malignant and nonmalignant rather than the effector cells that are damaging all tissues.
diseases. The field of biomarker research may provide more Identifying markers that are target tissue specific has been
accurate grading/risk stratification and identification of pa- technically challenging, owing to the cellular heterogeneity
tients at greater risk for refractoriness to therapy or GVHD of tissues and the difficulty of amplifying the amount of
progression. Furthermore, the treatment of aGVHD has protein required.
recently evolved from a one-size-fits-all approach to a more Another way of classifying these novel biomarkers is
refined strategy based on predicted outcomes. Patients who based on their biophysical properties. Transcriptomic bio-
are predicted to have low-risk aGVHD may benefit from lower markers are discovered by RNA expression profiling (mRNA,
doses and shorter courses of immune suppression. In addi- rRNA, tRNA, and other noncoding RNAs), whereas proteomic
tion, because not all cases of aGVHD progress in the same way biomarkers are discovered by the methodical study of the
or have the same outcome, the therapy should be tailored not protein profile of a biologic specimen. Finally, cellular bio-
only to the severity of the disease, but also to the predicted markers are discovered by the studying of the altered
rate of progression. As a result, numerous researchers have numbers and functions of several different immune cell
examined whether adding novel plasma biomarkers at subsets [24-26].
different time points before and after transplantation can add Below we review the systemic biomarkers (miRNA, ST2,
to the accuracy of prediction compared with other prognostic and markers of immune activation) and organ-specific bio-
tools. Timely recognition of patients who are at high risk for markers (Reg3a, S100, TIM-3, CK-18, HGF, and elafin). Our
aGVHD or who would likely demonstrate resistance to ste- findings are summarized in Table 1.
roids early in the course of transplantation may lead to more
stringent monitoring, better preventive care, and introduc- SYSTEMIC BIOMARKERS
tion of alternative and more effective immunosuppressive miRNAs
treatments earlier in the course of treatment. miRNAs are a class of small noncoding RNAs (21 to 25
It is reasonable to assume that plasma proteins involved in nucleotides) that negatively and positively regulate gene
the complex pathophysiology of aGVHD might be altered in expression by translational repression or induction of alter-
these patients. For the past 20 years, various groups have been ations in messenger RNA stability. miRNAs regulate gene
investigating potential biomarkers to enhance the early and function in various ways and at multiple levels, particularly
Table 1

1554
Summary of Biomarkers of aGVHD Classified According to Organ Specificity

Biomarker Studies Type (Number) Time Assay Conditioning GVHD Prophylaxis Outcomes
Regimen
Systemic
MicroRNA Ranganathan Murine/human (5) Day 21 post-HCT, 2-5 wk post-GVHD RT-PCR N/A N/A MiR-155 is up-regulated in T cells from aGVHD.
(eg, MiR- et al. [27] Human (64) Days 7, 14, 21, 30, 60, and 90 and at onset of RT-PCR MAC CSA, MMF, þ MTX Blocking MiR-155 may prevent aGVHD.
155, MiR- Xie et al. [28] Human (98) GVHD RT-PCR MAC Not specified MiR-155 is up-regulated in aGVHD following
586) Wang Human (196) 16 days before GVHD diagnosis qRT-PCR MAC and RIC Not specified allo-PBSCT, with a correlation with severity.
et al. [29] MiR-586 >2200 copies/uL at day 7 may predict
Xiao impending aGVHD (sensitivity, 87.5%;
et al. [30] specificity, 55.0%; AUC, 0.739).
A 4-miRNA panel predicted the probability of
aGVHD (sensitivity, 92%; specificity, 62%; AUC,
0.8). The panel was an independent predictor
for the development of aGVHD (HR, 1.478) at a

A.M. Ali et al. / Biol Blood Marrow Transplant 22 (2016) 1552e1564

median of 16 d before diagnosis.
miRNA level was significantly associated with
severity.
The panel was an independent unfavorable
prognostic factor for aGVHD OS (HR, 2.11).
The AUC of miRNA signature was higher than
that for sIL2Ra at 2 wk after HCT (0.86 vs. 0.76).
ST2 Vander Lugt Human (673), Day 14 after HCT Mass MAC and RIC CNI/MMF High ST2 level on day 14 was associated with an
et al. [31] 2 centers Days 16 and 28 after therapy for GVHD spectrometry/ MAC and RIC CNI/MTX increased NRM within 6 months of HCT (58%
Ponce Human (113) Day 28 after CBT ELISA CNI/MTX/Enbrel overall NRI).
et al. [32] ELISA CNI/MMF Patients with high ST2 levels at initiation of
therapy for aGVHD were 2.3-fold more likely to
have resistant aGVHD and 3.7-fold more likely
to die within 6 months compared with patients
with low levels.
High D28 ST2 levels (33.9 ng/mL) were
associated with increased risk of grade III-IV
aGVHD (HR, 2.62; AUC, 0.59) and increased TRM
(HR, 4.2; AUC, 0.75).
Ceruloplasmin Lv et al. [33] Human (98) Days -9, -1, 7, 14, 21, 28, 55-60, and 90-100 ELISA MAC CSA þ MMF þ MTX Ceruloplasmin level> 670 mg/mL on days 7, 14,
and 21 were predictive of aGVHD (sensitivity,
79.7%; specificity, 78.3%; AUC, 0.861), but not of
resistance to first-line therapy.
Immune activation
TNFR1 Choi Human (438) Before and at day 7 after HCT ELISA MAC FK506 þ mini-MTX The day 7 TNFR1 ratio was correlated with the
et al. [34] Human (82), Before and at day 7 after HCT ELISA MAC FK506 þ MMF eventual development of grade II-IV GVHD (HR,
Kitko pediatric Before and at day 7 after HCT ELISA RIC FK506 þ mini-MTX 2.44) and TRM (HR, 2.40).
et al. [35] Human (106) Before and at days 0, 5, 10, 15 after HCT ELISA RIC and MAC CNI þ MMF The day 7 TNFR1 ratio was correlated with
Willems Human (62), Not specified severity and 1-year OS in the pediatric
et al. [36] pediatric and population.
August adult The day 7 TNFR1 ratio was correlated with
et al. [37] grade II-IV (HR, 2.2) and grade III-IV (HR, 2.9)
aGVHD, but not with OS.
Day 15 TNFR1 levels (among other biomarkers:
sCD8, sIL-2R, sCD40, and sCD28) were
correlated with severe aGVHD (AUC, 0.77; PPV,
0.45; NPV, 0.87).
 IL-7 Dean Human (31) Days 7 and 14, and months 1, 2, 3, 6, 9, and 12 ELISA RIC CSA þ MTX IL-7 level at day 14 (cutoff value, 13 pg/mL)
et al. [38] Human (45) Days 0, 7, 14, 18, 25, 30, 60, and 90 ELISA RIC CSA þ MTX predicted the subsequent development of
Thiant Human (40) Days 0, 7, 14, 18, 25, 30, 60, and 90 ELISA MAC CSA þ MTX aGVHD (sensitivity, 85.7%; specificity, 88.2%;
et al. [39] PPV, 85.7%; NPV, 88.2%). Higher levels were
Thiant strongly associated with more severe grades of
et al. [40] aGVHD.
IL-7 and IL-15 have similar kinetics, peaking on
day 14 at 4- to 5-fold over preconditioning
values (median, 15.8 and 38.7 pg/mL,
respectively). The occurrence of grade II-IV
aGVHD is associated with peak IL-7 level (HR,
5.38).
sBAFF Cho Human (45) Days 0, 7, and 14 ELISA MAC CSA þ MTX sBAFF level >43 pg/mL at each time point was
et al. [41] FK506 þ MTX associated with a significantly lower cumulative
incidence of aGVHD (sensitivity, 75%;
specificity, 73%-82%; AUC, 0.7-0.8).
TIRC7 Zhu Human (39) Before and after GVHD therapy qPCR and ELISA MAC and RIC CSA þ MTX Higher levels of TIRC7 in aGVHD. Levels were

A.M. Ali et al. / Biol Blood Marrow Transplant 22 (2016) 1552e1564

et al. [42] correlated with severity and declined markedly
after therapy.
Organ-specific
Lower GI-specific
REG3a Harris Human (954), At onset of GVHD ELISA High and CNI/MMF REG3a (>151 ng/mL) distinguished lower GI
et al. [43] multicenter At onset of lower GI GVHD ELISA moderate CNI/MTX GVHD from non-GVHD diarrhea (AUC, 0.79;
Ferrara Human (1014), intensity At least 2 agents PPV, 95%; NPV, 34%).
et al. [44] multicenter High and including CNI REG3a at onset of lower GI GVHD was predictive
moderate of nonresponse to therapy at day 28 (PPV, 51%;
intensity NPV, 76%).
Higher REG3a concentration at the onset of
lower GI GVHD (above the median of 135 ng/
mL) was significantly correlated with higher
1-year NRM (52% versus 33%).
REG3a level distinguished LGI GVHD from non-
GVHD diarrhea (AUC, 0.80; PPV, 95%; NPV, 32%).
REG3a level >151 ng/mL at GVHD onset was
predictive of nonresponse to therapy at 4 weeks
(3-fold higher), 1-year NRM (59% versus 34%),
and 1-year OS (27% versus 48%).
REG3a level further risk-stratified patients with
either advanced clinical stage or histological
severity (34% for 1 risk factor versus 66% for 2
risk factors).
S100 Reinhardt Human (52) At onset or progression of GVHD Calprotectin MAC and RIC CNI, MMF, MTX, ATG S100A8/S100A9 and S100A12 levels were
et al. [45] Human (72) At onset of GVHD assay and MAC and RIC CSA þ MMF significantly increased in stool and serum of
Rodriguez- Human (59) At onset of GVHD sandwich ELISA Not specified CSA þ MTX patients with GI aGVHD and extensive cGVHD
Otero Sandwich ELISA Not specified compared with patients without GVHD. No
et al. [46] ELISA correlation with grade could be detected.
Chiusolo Fecal S100 concentrations were increased in
et al. [47] grade II-III GI aGVHD but not in grade I, causing
low sensitivity of the marker 31% (90%
specificity).
A high fecal S100 level (100 mg/g) was strongly
associated with SR-GVHD (95% CI of SR-GVHD,
93% versus 33%). S100 100 mg/g was
independently correlated with a lower

1555
probability of CR (HR, 0.47).
(continued on next page)
Table 1

1556
(continued )

Biomarker Studies Type (Number) Time Assay Conditioning GVHD Prophylaxis Outcomes
Regimen
Fecal S100 was higher in patients with aGVHD
than in those with non-GI aGVHD. Levels
increased with disease severity (with an
arbitrary cutoff of 160 mg/kg: sensitivity, 100%;
specificity, 81.8%; PPV, 86%; NPV, 100%; AUC,
0.942).
TIM-3 Oikawa Murine N/A Flow cytometry N/A N/A TIM-3 is up-regulated in aGVHD. AntieTIM-3
et al. [48] Murine N/A Flow cytometry N/A N/A antibody increases the severity of aGVHD.
Veenstra Human (127) At onset of aGVHD Sandwich ELISA/ MAC and RIC CNI þ MTX TIM-3/galectin-9 pathway acted as a suppressor
et al. [49] flow cytometry CNI þ MMF of aGVHD.
Hansen TIM-3 levels were significantly higher in
et al. [50] patients with more severe mid-gut GVHD

A.M. Ali et al. / Biol Blood Marrow Transplant 22 (2016) 1552e1564

compared with those with upper gut GVHD,
without GVHD, and normal controls (median,
11,550 versus 4670 versus 2710 versus 2285
pg/mL, respectively). The levels were higher in
samples collected closest to GVHD onset
compared with earlier samples. The levels were
correlated with severity.
Samples from patients with common diseases
did not yield increased levels of TIM-3.
CK-18 Luft Human (55) At the time of maximum GVHD ELISA MAC and RIC Not specified Both intestinal and hepatic GVHD were
et al. [51] Human (48) Longitudinally: before HCT, at escalation of ELISA MAC and RIC CNI/MMF consistently associated with significant
Luft immunosuppression, and late CNI/MTX elevations of CK18F levels over baseline
et al. [52] (5.6-fold change).
CK18F levels decreased in responsive GVHD to
immunosuppressive therapy and persisted in
resistant GVHD.
Conditions that might represent relevant
differential diagnoses (toxic mucositis,
noncomplicated infection-related diarrhea, and
VOD) were not associated with CK18F elevation.
Significantly higher CK18F levels were detected
in steroid-refractory GVHD. indicating the
ongoing epithelial death of target organs.
Liver-specific
HGF Okamoto Human (38) Serially from days 0-120 post-HCT ELISA Not specified Not specified HGF levels were increased in patients with
et al. [53] Human (424) Days 0, 7, 14, 21, 28, 56, and 100 and at onset of ELISA MAC and RIC FK506 and MTX aGVHD. Level was correlated with disease
Paczesny Human (954), GVHD ELISA High and CNI/MMF grade.
et al. [54] multicenter At onset of GVHD moderate CNI/MTX HGF as part of a panel (along with IL-2Ra,
Harris intensity MF þ MTX TNFR1, and IL-8) discriminated patients with
et al. [43] and without aGVHD (AUC, 0.86).
The panel predicted survival independent of
GVHD severity.
HGF was higher in visceral GVHD compared
with skin-only GVHD.
HGF concentrations were elevated in lower GI
GVHD compared with isolated skin GVHD.
 HGF level is a poor diagnostic biomarker for
distinguishing lower GI GVHD from non-GVHD
diarrhea (AUC, 0.60) and liver GVHD from other
causes of hyperbilirubinemia (AUC, 0.59).
High HGF concentration was correlated with
significantly higher 1-year NRM.
Skin specific
Elafin Paczesny [55] Human (492) At onset of skin GVHD ELISA MAC and RIC CNI and another Plasma elafin levels were twice as high in
Chacon [56] N/A N/A N/A N/A agent patients with skin GVHD compared with other
N/A groups (no GVHD, GI GVHD, and non-GVHD
rash).
Elafin levels at the time of diagnosis of skin
GVHD were correlated with maximum overall
GVHD grade.
Among the available biomarkers, Elafin is the
best single discriminator for the diagnosis of
GVHD in BMT recipients with a rash (AUC, 0.77).

A.M. Ali et al. / Biol Blood Marrow Transplant 22 (2016) 1552e1564

The 1-year NRM was more than double in the
high-elafin group (6000 pg/mL) compared
with low-elafin group (28% versus 11%).
High elafin level at the time of GVHD diagnosis
was significantly associated with a greater risk
of death (HR, 1.78).
There is immunohistochemical overexpression
of elafin in both skin aGVHD and engraftment
syndrome; however, an elevated serum elafin
level has not been reported in engraftment
syndrome.

ATG indicates antithymocyte globulin; BMT, bone marrow transplantation; CBT, cord blood transplantation; CI, cumulative incidence; CNI, calcineurin inhibitor; CR, complete response; CSA, cyclosporine; ELISA, enzyme-linked
immunosorbent assay; FK506, tacrolimus; N/A, not available; MMF, myecophenolate mofetil; MTX, methotrexate; MAC, myeloablative conditioning; NRI, net reclassification index; RIC, reduced-intensity conditioning; RT-PCR,
reverse-transcription polymerase chain reaction; SR-GVHD, steroid-resistant GVHD; VOD, veno-occlusive disease.

1557
1558 A.M. Ali et al. / Biol Blood Marrow Transplant 22 (2016) 1552e1564

transcription, translation, and protein degradation [57]. level (defined as >50% above the responder’s median value)
Circulating miRNAs have been studied as novel, noninvasive were less likely to respond to treatment, whereas those with
biomarkers for many diseases, such as cancer, sepsis, car- low levels were more likely to respond to treatment even if
diovascular disease, liver injury, organ transplant rejection, they had high-grade aGVHD using clinical criteria. The
and diabetes [58-61]. addition of ST2 to the clinical risk factors has reclassified
Various researchers have conducted animal and human almost 60% of patients receiving all types of conditioning
studies to investigate the role of miRNAs in the pathogenesis regimens. Furthermore, a strong additive effect was noted
of aGVHD and to screen for promising biomarkers and when both ST2 and REG3a levels (see below) were added to
therapeutic targets [27-30]. Studies have found some the clinical characteristics to improve the accuracy of the risk
lymphocyte-related serum miRNAs, such as MiR-155, to be stratification. Similar results were obtained when testing the
significantly up-regulated in aGVHD, with a correlation be- efficacy of ST2 in risk-stratifying cord blood transplantation
tween serum MiR-155 level and disease severity [28]; how- recipients [32].
ever, MiR-155 is known to play a key role in inflammation Despite the promising results of the foregoing studies in
regulation and immune response [62] and is unlikely to be demonstrating a particular significance of ST2 measure-
specific for aGVHD. Furthermore, the expression of MiR-155 ments in post-transplantation mortality risk stratification,
was the lowest among the miRNAs measured in the serum many questions remain. ST2 concentrations seem to vary
of patients with aGVHD [30]. Allo-HCT recipients with high significantly across conditioning intensities, necessitating
levels of another MiRNA, MiR-586, at day 7 were found to be separate thresholds to identify each conditioning intensity.
at high risk for developing aGVHD [29]; however, MiR-586 Furthermore, the timing for measuring ST2 was chosen
levels were readily affected by infections that commonly arbitrarily in these studies, and a prospective evaluation of
occur by day 7 after allo-HCT. The lack of specificity of these serial ST2 measurements is needed to refine the predictive
studied miRNAs has led to the simultaneous use of several value of this biomarker.
miRNAs to increase the specificity and diagnostic perfor-
mance of these biomarkers. Biomarkers of immune activation
Plasma miRNA signature has been studied as a biomarker The persistent activation of the immune system that oc-
for aGVHD. Xiao et al. [30] reported that evaluation of a panel curs early after donor graft infusion leads to excessive cyto-
of 4 miRNAs (miR-423, miR-199a-3p, miR-93, and miR-377) kine production (ie, cytokine storm). Cytokines are
was able to distinguish patients who developed aGVHD important to the pathogenesis of aGVHD [69]. Numerous
from those who did not (AUC, 0.80), and also to predict the studies have examined cytokines and their receptors as po-
severity of the disease [30]. High expression of the miRNA tential aGVHD biomarkers. Here we review the studies
panel was associated with poor overall survival. More examining the diagnostic and prognostic utility of serum
importantly, elevated miRNAs can be detected at a median of levels of TNFR1, IL-7, and sBAFF.
16 days before the diagnosis of aGVHD and were found to be
more predictive for the development of aGVHD compared TNFR1
with another biomarker, sIL2Ra. Furthermore, this miRNA An inflammatory cytokine frequently implicated in the
signature for aGVHD, unlike other studies of other miRNAs or pathogenesis of aGVHD, TNF-a is the most frequently re-
polypeptides, was not seen in the plasma of lung transplant ported immune activation marker elevated before the onset
recipients undergoing rejection, patients with non- of GVHD [70]. Researchers have measured TNFR1 as a sur-
transplanted sepsis, or patients with veno-occlusive disease rogate marker for TNF-a. Increased levels of TNFR1 at day 7
[30]. This suggests a higher specificity of this miRNA signa- to 2.5 times baseline (before HCT) were significantly
ture for aGVHD. correlated with the eventual development of severe aGVHD
Serum or plasma miRNAs are highly stable in human and with TRM in several studies [34-36]. Interestingly, day 7
peripheral blood [63] and are tissue-specific [64], making serum-soluble TNFR1 levels were not significantly associated
these biomarkers excellent targets for studying. There are with grade II-IV aGVHD in these studies, suggesting that
other advantages of an miRNA signature, including ease of TNFR1 day7/baseline ratio has better predictive value.
measurement, simplicity, and low cost. Organ-specific asso- Another important observation of these studies is the low
ciation studies of miRNAs in the setting of aGVHD are lacking, sensitivity of TNFR1 ratio (2.5 times baseline) in predicting
and could add to our improved stratification of patients to severe aGVHD (sensitivity <40%), given that the majority of
organ-specific treatment strategies. patients who develop grade II-IV aGVHD have a TNFR1 day 7/
baseline ratio <2.5. Nevertheless, in patients with a TNFR1
ST2 ratio 2.5 times baseline, the likelihood of developing sig-
ST2 (suppression of tumorigenicity 2) is a member of the nificant aGVHD is sufficiently high (w60%) to justify the
IL-1 receptor family that specifically binds to IL-33. ST2 is future use of TNFR1 ratioebased preemptive treatment
present in 2 isoforms: a membrane form expressed on he- strategies. The screening performance of the TNFR1 ratio
matopoietic cells, specifically T helper 2 (Th2) cells, and a may be strengthened by combining it with other biomarkers
soluble form secreted by endothelial and epithelial cells in as a composite panel (AUC, 0.77; NPV, 0.87) (Table 2) [37].
response to inflammatory injury [65,66]. Soluble ST2 appears
to act as a receptor for IL-33, limiting its access to Th2 cells IL-7
and promoting the Th1 phenotype, which has been associ- IL-7 is a growth factor that represents the principal ho-
ated in aGVHD pathophysiology [67,68]. meostatic cytokine for T cells and is important for B cell and
In a large 2-center study of 673 recipients of myeloa- T cell maturation. IL-7 promotes immune reconstitution after
blative or nonmyeloablative HCT, patients with a high ST2 allogeneic HCT and is required for the development of
level measured as early as 14 days after transplantation had aGVHD in murine models [75,76]. High levels of IL-7 along
an increased risk of NRM [31]. Moreover, ST2 level predicted with other cytokines leads to homeostatic peripheral
the response of aGVHD to treatment. Patients with a high ST2 expansion of donor T cells within 30 days post-HCT. Given
Table 2
Biomarker Panels That Combine Several Markers of aGVHD to Increase Specificity or Predictive/Diagnostic Power

Panel Biomarkers Study Type (Number) Time Assay Conditioning Regimen GVHD Prophylaxis Outcome
TIM3, IL-6, sTNFR1, ST2 McDonald et al. [71] Single center (317) Days 7-70 after HCT in cohorts 1 ELISA MAC CNI þ MTX; CNI þ MMF The panel was predictive of
and 2 (with GI GVHD); mean development of grade III-IV
day 14  3 after HCT in patients GVHD at a median of 4 days
without GVHD before initiation of therapy
(AUC, 0.88). TIM3 predicted
subsequent grade III-IV GVHD
(AUC, 0.76). Plasma ST2 and
sTNFR1 predicted 1-year NRM
(AUC, 0.90).
IL-2Ra, TNFR1, HGF, IL-8, Levine et al. [72] Multicenter (112) Onset of GVHD and days 0, 14, ELISA MAC/RIC Not specified The day 0 and 14 biomarker
elafin, REG3a and 28 after therapy initiation panels independently predicted
nonresponse at day 28 (OR, 2.98
for day 0; OR, 6.32 for day 14).

A.M. Ali et al. / Biol Blood Marrow Transplant 22 (2016) 1552e1564

The day 0 biomarker panel
independently predicted
mortality by day 180 (OR, 4.61;
AUC, 0.7210). The day 28
biomarker panel predicted day
180 NRM (OR, 7.43).
Several soluble and Te Boome et al. [73] Single center (48) Before and at predetermined ELISA MAC/RIC Not specified The panel was predictive for
cellular biomarkers time points after first MSC mortality (HR, 2.924) when
infusion measured before MSC
administration. ST2 was
predictive for mortality only at
2 weeks after, but not before,
MSC administration (HR,
2.389).
TNFR1, ST2, REG3a Levine et al. [74] Multicenter (492) At 48 h before or after therapy ELISA MAC/RIC CNI þ MTX; CNI þ MMF; The CI of 12-month NRM
initiation CNI þ sirolimus; post-HCT significantly increased as GVHD
cyclophosphamide score increased (8% vs 27% vs
46% for score 1, 2, and 3,
respectively). The response
rates to primary GVHD therapy
decreased as the GVHD score
increased (86% vs 67% vs 46%
for score 1, 2, and 3,
respectively). The development
of GI GVHD in patients who
presented with skin GVHD
increased only as the score
increased (19% vs 29% vs 34%
for score 1, 2, and 3,
respectively). This is the first
biomarker-based score that can
be used to guide risk-adapted
therapy.

MSC indicates mesenchymal stromal cell; OR, odds ratio.

1559
1560 A.M. Ali et al. / Biol Blood Marrow Transplant 22 (2016) 1552e1564

this key role in initial T cell recovery, many groups have estimate the maximal severity would be helpful. The most
measured IL-7 levels over sequential time points before and widely studied LGI-specific biomarkers are: REG3a, CK-18,
after allo-HCT. In a small study, IL-7 level at day þ14 was S100, and TIM-3.
predictive of the subsequent development of aGVHD (PPV,
85%; NPV, 88%), and higher levels were strongly associated REG3a
with more severe grades of aGVHD [38]. In a multivariate REG3a (regenerating islet-derived 3-a) is an antimicrobial
model, high serum IL-7 level by day 14 after transplantation protein expressed in Paneth cells and secreted into the crypt
was the factor most strongly associated with the probability microenvironment. REG3a may have a protective effect for
of developing grade II-IV aGVHD after both myeolablative intestinal stem cells, which are important cellular targets of
[40] and reduced- intensity conditioning [39]. It is important GVHD in the GI tract [81].
to note, however, that some other studies have reported less REG3a has been shown to have very good diagnostic and
TRM in patients who recover T cells more rapidly [77], sug- prognostic performance in multicenter studies with large
gesting that the ratio among the subsets of the lymphocytes, numbers of patients. REG3a distinguishes LGI GVHD from
rather than the size, determines the aGVHD developments non-GVHD diarrhea with excellent PPV and AUC (95% and
and the outcomes. The predictive value of IL-7 level still re- 0.8, respectively). REG3a was the best diagnostic biomarker
quires support from larger, prospective trials. for LGI GVHD, and additional biomarkers (HGF and CK18) as a
composite panel provided minimal increased sensitivity or
sBAFF specificity. Furthermore, the levels of REG3a at the onset of
sBAFF (B cell activating factor, also known as B cell sur- LGI GVHD predicts the response to first-line therapy and 1-
vival and activation factor), acts as a potent B-cell activator year NRM [43,44].
and it has been shown to play a crucial role in the prolifer- REG3a, along with clinical severity and histological
ation and differentiation of these cells in mice and humans. It severity, provides important prognostic information before
has been linked to a variety of autoimmune diseases. sBAFF is the initiation of therapy rather than at the time of maximum
required for the reconstitution of B cells after myeloablation grade of GVHD. The 3 factors (REG3a level, clinical and his-
in animal models. Cho et al. [41] found that elevated sBAFF tological severity) independently predict the lack of response
level at any time point in the early days after HCT was to therapy and can be integrated into a single grading system
associated with decreased risk of developing aGVHD. These that will permit better risk stratification of patients with
findings suggest not only that there may be a predictive role severe LGI aGVHD.
of sBAFF during the peritransplantation period, but also that
elevated levels may even confer protection against the S100
development of aGVHD. This is in contrast with notable S100s are proinflammatory protein that play important
recent reports demonstrating an association between high roles in many inflammatory disorders, including inflamma-
sBAFF level and the occurrence of chronic GVHD (cGVHD) tory bowel disease (IBD) and rheumatoid arthritis (RA)
[78]. This highlights the difference in pathogenesis between [82,83]. S100 proteins are released by activated or damaged
aGVHD, which is mediated mainly by donor T cells, and phagocytes under conditions of cell stress during infections
cGVHD, where donor B cells play an important role in the and autoimmune disease and represent promising novel
pathophysiology of cGVHD. Despite the recent reports that therapeutic targets. Previous studies have attempted to
have linked B cells to aGVHD [79], the role of B cells in the evaluate the diagnostic and prognostic value of serum and
pathogenesis of aGVHD remains uncertain. Furthermore, fecal levels of some of S100 proteins (S100A12 and calpro-
T cells have been shown to express BAFF receptors [80]; thus, tectin [S100A8/S100A9]) in the setting of LGI aGVHD [45-47].
there could be a role for sBAFF in the regulation of The major advantage of these markers lies not in their
T celledependent immunity. Further experimental and ability to diagnose GI GVHD, but rather in their ability to
clinical studies are needed to gain more insight into the roles predict the response to steroids, and thus in risk-stratifying
of BAFF and B cells in the pathogenesis of aGVHD. these patients. The sensitivity of fecal S100 proteins for the
The use of sBAFF as a biomarker for aGVHD prediction diagnosis of grade I GI aGVHD is quite low (with no sensitivity
may be complicated by various factors. First, sBAFF levels are improvement with a lower cutoff) and is unable to discrimi-
increased in the setting of B lymphopenia. Furthermore, the nate aGVHD from other causes of diarrhea; however, S100
levels are also affected by medications, particularly high- level is predictive for response to treatment. Furthermore,
dose steroids that are commonly used to treat aGVHD. fecal S100 (along with other fecal markers, a1-antitrypsin and
Finally, current ELISA assays underestimate sBAFF levels, and elastase) is particularly useful for restratifying patients with
precise quantification of the biomarker is lacking. grade II GI GVHD, allowing identification of patients with the
worst prognosis. Interestingly, the probability of steroid-
ORGAN-SPECIFIC BIOMARKERS resistant GI aGVHD was as high as 100% if 2 fecal markers
Lower GIeSpecific biomarkers were high and as low as 0% if the 2 markers were low.
Lower GI (LGI) GVHD occurs in up to 60% of patients after Fecal proteins represent attractive noninvasive bio-
allo-HCT. The involvement of the lower GI tract with aGVHD markers for evaluating LGI aGVHD. These biomarkers are
is often severe and is characterized by voluminous secretory immediately available and can be easily collected. Moreover,
diarrhea, with or without hematochezia, and abdominal if validated, they may represent a better tool for grading, risk-
cramps. The etiology of diarrhea following allo-HCT is a stratifying, and monitoring the response to treatment in
common diagnostic dilemma often indistinguishable from patients with LGI aGVHD compared with histologically var-
other causes of diarrhea (conditioning chemotherapy iable GI biopsies and measurement of stool volumes.
related, administration of antibiotics, Clostridium difficilee
associated diarrhea, or cytomegalovirus infection). The GI TIM-3
involvement of aGVHD remains a major cause of morbidity, The TIM (T-cell immunoglobulin domain and mucin
and the use of biomarkers to predict the occurrence or domain) family of genes was first described in 2001 [84].
 A.M. Ali et al. / Biol Blood Marrow Transplant 22 (2016) 1552e1564 1561

The critical role of these gene products (3 proteins: TIM-1, -3, hyperbilirubinemia caused by non-GVHD transplantation-
and -4) in regulating immunity is beginning to emerge [85]. related complications, such as veno-occlusive disease, and
TIM-3, the first product described, is the most widely studied these complications can be discriminated clinically from
member of the family [86]. To date, the in vivo functions of GVHD. Despite this, however, a biomarker that can differ-
TIM-3 and its role in immunoregulation remain largely un- entiate liver aGVHD from these complications would be
known, with discrepant study results. Interactions between clinically meaningful. HGF seems to perform significantly
TIM-3 and its ligand galectin-9 play a role in autoimmune better as a diagnostic biomarker when integrated in a com-
disorders, chronic infection, tumor immunity, and trans- posite biomarker panel [54] (Table 1).
plantation [87,88]. In murine models of aGVHD, TIM-3 has Despite HGF’s poor diagnostic capabilities, this biomarker
been shown to be up-regulated, with a possible inhibitory has significantly better prognostic performance. Biomarker
role of the interaction between TIM-3 and its ligand in the panels that involve HGF have been shown to predict 1-year
pathogenesis of aGVHD [48,49]. NRM and long-term survival independently of GVHD
The predictive value of serum TIM-3 level in the setting of severity, and HGF was the primary contributor to the prog-
GVHD has been studied by Hansen et al. [50]. Although TIM-3 nostic significance of these panels [43,54].
levels correlated with the occurrence and severity of the
disease, the AUC was significant only for severe midgut Skin-Specific Biomarkers
GVHD compared with all other types of aGVHD, including Elafin
upper GI aGVHD (0.79 versus 0.59). These findings are Skin-derived antileukoproteinase (SKALP), otherwise
inconsistent with the recent study from McDonald et al. [71], known as elafin, is an elastase-specific protease inhibitor
where the measurement of TIM-3 was the most useful ana- expressed mainly in epithelial cells. Elafin is not present in
lyte among those studied for predicting grade III-IV versus normal keratinocytes, but is overexpressed in various in-
grade 0-II aGVHD (AUC, 0.76) (Table 2). flammatory skin disorders, including psoriasis [97]. In vivo
studies have shown that inflammatory cytokines secreted in
CK-18 patients with aGVHD can stimulate the expression of elafin
CK-18 (cytokeratin-18) is an intermediate filament typi- in epidermal cells [98]. Immunohistochemistry studies have
cally expressed in the epithelia of the digestive, respiratory, revealed an overexpression of elafin in GVHD skin biopsy
and urogenital tracts [89]. Initiation of the apoptotic process specimens. Elafin is expressed by the target epidermal cells
results in activation of certain caspases that cleave various rather than by the effector cells that injure all 3 target organs
cellular substrates, including CK18. Cleavage at a particular (GI, liver, and skin), which explains the greater specificity of
site (DALD-S) gives rise to a neo-epitope, cytokeratin-18 elafin as a biomarker for skin aGVHD. Although produced
fragments (CK18Fs), which are released into the serum [90]. locally as an antiprotease secreted in response to cytokines,
Intracryptal apoptosis is the histopathological hallmark in elafin is readily detected in the systemic circulation as well.
GVHD and is used in the grading of GVHD [91]. Based on the A comparison of the diagnostic features of serum elafin as
hypothesis that apoptosis at the level of target organs in a biomarker of skin aGVHD with other biomarkers of aGVHD
aGVHD leads to an increase in CK18F, and that CK18F mirrors (TNFR1, IL2Ra, IL8, and HGF) identified elafin as the best
the pathogenetic endpoint of GVHD, researchers have single marker for differentiating skin GVHD from other eti-
investigated whether CK18F provides a tool for the sensitive ologies of rash (ie, drug rash, engraftment syndrome, leu-
assessment of GVHD-associated apoptotic activity and dis- kemia cutis, and viral or fungal rash) (Table 1). The AUC of
ease grading [51,52]. Although CK18F level correlated with GI elafin was the highest (0.77) of all the biomarkers studied.
and hepatic GVHD, as well as with the response to immu- Furthermore, elafin was correlated with the maximum
nosuppressive therapy, these studies were not designed to overall GVHD grade and 1-year NRM, making its prognostic
systematically determine the diagnostic features of this value appealing as well [55].
biomarker (sensitivity, specificity, PPV, and NPV). Apoptosis
is not GVHD-specific, and more studies are needed to vali- BIOMARKER PANEL
date the diagnostic performance of its associated biomarkers When several clinically useful biomarkers are present,
in the setting of aGVHD. typically no particular one is satisfactory alone in terms of
sensitivity or specificity for the diagnosis or prediction of a
Liver-Specific Biomarkers specific disease. Researchers have identified potential
HGF biomarker panels for aGVHD either by combining individual
HGF is a cellular growth factor that targets epithelial and biomarkers that have been studied previously or by per-
endothelial cells, including hepatocytes, by activating a forming discovery studies that compare samples from pa-
tyrosine kinase signaling cascade after binding to proto- tients with severe aGVHD with samples from patients
oncogene c-Met receptors [92,93]. Serum HGF concentra- without aGVHD. Analyzing these samples identifies bio-
tions have been reported to be elevated in liver disease, markers that provide the best discrimination between the 2
chronic renal failure, systemic inflammatory response syn- groups. Subsequently, the panel is developed, validated, and
drome, and such malignant diseases as leukemia, breast tested for operating characteristics. Table 2 presents a sum-
cancer, and gastric cancer [94-96]. mary of biomarker panels.
Although HGF levels have been found to be elevated in Recently, numerous groups have used biomarker panels
patients with visceral GVHD, HGF performs poorly as a to identify patients at the extremes of outcomes. These
diagnostic biomarker. HGF fails to distinguish LGI aGVHD panels have been shown to provide prognostic and predic-
from non-GVHD diarrhea (AUC, 0.6) [43]. Likewise, HGF tive information on aGVHD outcomes, including maximal
concentrations are elevated in patients with liver GVHD severity, response to therapy, and mortality [71,72]. The
compared with asymptomatic patients, but are comparable predictive values of these panels are independent and ad-
to those in patients with non-GVHD hyperbilirubinemia ditive of the clinical status of the patient (onset grade, un-
(AUC, 0.59) [43]. A small number of patients develop related versus related donor, and peripheral versus bone
1562 A.M. Ali et al. / Biol Blood Marrow Transplant 22 (2016) 1552e1564

marrow versus cord blood as a stem cell source), with a very composite panels, at this time the use of these data in clinical
good AUC (0.7 to 0.9) (Table 2). However, the PPVs of the practice should be reserved for use in well-designed pro-
panels remained in the range of 40% to 50%, indicating that spective clinical trials.
the false-positive results would still equal or outnumber the Although the data derived from biomarker-based scoring
true-positive results. A recent reiteration of a biomarker systems predicts outcomes better than clinical scoring sys-
panel in a well-powered study by the University of Michigan tems, the PPV of these biomarkers is still low, and these
group (Ann Arbor GVHD score) examined the utility of systems are most useful for patients who score at either end
measuring 3 biomarkers (TNFR1, REG3a, and ST2) at the time to justify preemptive therapy. Moreover, reconciling the re-
of aGVHD diagnosis to create a model algorithm that predicts sults of the various studies done at different centers to
NRM 6 months later [74]. The algorithm defined 3 distinct examine the value of different biomarkers in the prediction
scores whose risk of NRM increased with each increasing of aGVHD outcomes will be difficult owing to the differences
grade (independent of other risk factors, such as donor type, in timing, intensity of conditioning therapy, choice of GVHD
recipient age, conditioning regimen, and HLA match) and prophylaxis, hematopoietic cell source, and statistical
may prove useful for guiding aGVHD therapy. The value of methods. This suggests that predictive biomarkers might
biomarkers in predicting the risk of complications and require a coordinated multicenter approach with coordi-
mortality after HCT will be studied in Blood and Marrow nated times of sampling and centralized analysis of
Transplant Clinical Trials Network Protocol 1202. biomarker levels. Another attractive approach would be to
Although the integration of a biomarker panel in clinical incorporate the clinical characteristics and risk factors with
practice requires further investigation, these panels may also biomarker scores (such as the Ann Arbor score) to improve
provide a valuable alternative or addition to composite the predictive power of these scores and their treatment
endpoints in clinical trials evaluating novel therapies for algorithms. Biomarkers represent a compelling, yet still un-
aGVHD, and improve the success rate of these trials. proven, tool for reproducible prediction of aGVHD, severity
Biomarker panels may provide an opportunity for a better of aGVHD, and outcomes of therapy.
selection of patients who are likely to benefit from expensive
or potentially risky therapies tested in clinical trials, as well
ACKNOWLEDGMENTS
as close monitoring of responses. To this end, Te Boome et al.
Financial disclosure statement: The authors have nothing
[73] evaluated the utility of biomarkers for predicting either
to disclose.
resolution of aGVHD or survival in patients with steroid-
Conflict of interest statement: There are no conflicts of in-
refractory aGVHD treated with mesenchymal stromal cells
terest to report.
in a prospective phase II trial [73]. Surprisingly, biomarkers
that were previously reported to be predictive of outcomes
(ST2 [31] and the Levine panel [72]) failed to correlate with REFERENCES
clinical severity or predict resolution of aGVHD, likely owing 1. Lee SE, Cho BS, Kim JH, et al. Risk and prognostic factors for acute GVHD
to the relatively low number of patients. These studies based on NIH consensus criteria. Bone Marrow Transplant. 2013;48:
587-592.
represent the first demonstrations of how GVHD biomarker 2. Ferrara JL, Levine JE, Reddy P, Holler E. Graft-versus-host disease.
panels and biomarker-based scores potentially may be Lancet. 2009;373:1550-1561.
incorporated into clinical care. 3. Iqbal N, Salzman D, Lazenby AJ, Wilcox CM. Diagnosis of gastrointes-
tinal graft-versus-host disease. Am J Gastroenterol. 2000;95:3034-3038.
4. Roy J, Snover D, Weisdorf S, et al. Simultaneous upper and lower
SUMMARY endoscopic biopsy in the diagnosis of intestinal graft-versus-host dis-
Many flaws exist in the current clinically based a GVHD ease. Transplantation. 1991;51:642-646.
5. McDonald GB. How I treat acute graft-versus-host disease of the
grading systems. Maximal clinical severity, on which grading gastrointestinal tract and the liver. Blood. 2016;127:1544-1550.
tools rely, can be assigned only retrospectively once the 6. Martin PJ, Rizzo JD, Wingard JR, et al. First- and second-line systemic
response to treatment is known. Withholding treatment for treatment of acute graft-versus-host disease: recommendations of the
American Society of Blood and Marrow Transplantation. Biol Blood
GVHD or waiting for 14 to 28 days to evaluate for refracto-
Marrow Transplant. 2012;18:1150-1163.
riness to therapy may lead to extensive and irreversible 7. Dignan FL, Clark A, Amrolia P, et al. Diagnosis and management of acute
damage, especially in the GI tract, resulting in unacceptably graft-versus-host disease. Br J Haematol. 2012;158:30-45.
high rates of morbidity and mortality. If it were possible to 8. Bombí JA, Nadal A, Carreras E, et al. Assessment of histopathologic
changes in the colonic biopsy in acute graft-versus-host disease. Am J
predict the ultimate severity of aGVHD before or at the onset Clin Pathol. 1995;103:690-695.
of symptoms, preemptive therapy (eg, antieT cell therapy) 9. Ponec RJ, Hackman RC, McDonald GB. Endoscopic and histologic
could be provided to blunt the intensity of tissue damage. diagnosis of intestinal graft-versus-host disease after marrow trans-
plantation. Gastrointest Endosc. 1999;49:612-621.
Whether this approach, even when implemented early, will 10. Liakopoulou E, Mutton K, Carrington D, et al. Rotavirus as a significant
work to change the natural history of GVHD remains to be cause of prolonged diarrhoeal illness and morbidity following alloge-
seen and proven in prospective clinical trials. Even patients neic bone marrow transplantation. Bone Marrow Transplant. 2005;36:
691-694.
who respond to low doses of steroids can experience un- 11. Selbst MK, Ahrens WA, Robert ME, et al. Spectrum of histologic changes
necessary infectious complications as well as other morbid- in colonic biopsies in patients treated with mycophenolate mofetil.
ities, such as diabetes mellitus and avascular necrosis. Mod Pathol. 2009;22:737-743.
12. Alonso CD, Treadway SB, Hanna DB, et al. Epidemiology and outcomes
Many studies have demonstrated a correlation between of Clostridium difficile infections in hematopoietic stem cell transplant
individual as well as combined biomarkers and outcomes of recipients. Clin Infect Dis. 2012;54:1053-1063.
aGvHD. However, inconsistent results among different clin- 13. Przepiorka D, Weisdorf D, Martin P, et al. 1994 Consensus Conference
on Acute GVHD Grading. Bone Marrow Transplant. 1995;15:825-828.
ical centers and studies are noted and mentioned in this
14. Rowlings PA, Przepiorka D, Klein JP, et al. IBMTR Severity Index for
review. This inconsistency can be attributed to the high grading acute graft-versus-host disease: retrospective comparison
heterogeneity of patient groups receiving various treat- with Glucksberg grade. Br J Haematol. 1997;97:855-864.
ments, as well as to variations in study protocols and treat- 15. MacMillan ML, Robin M, Harris AC, et al. A refined risk score for acute
graft-versus-host disease that predicts response to initial therapy,
ments. Despite the statistically significant operating survival, and transplant-related mortality. Biol Blood Marrow Trans-
characteristics of many of the biomarkers discussed and the plant. 2015;21:761-767.
 A.M. Ali et al. / Biol Blood Marrow Transplant 22 (2016) 1552e1564 1563

16. Harris AC, Young R, Devine S, et al. International, multicenter stan- 42. Zhu F, Qiao J, Chen W, et al. Increased expression of T cell immune
dardization of acute graft-versus-host disease clinical data collection: a response cDNA 7 in patients with acute graft-versus-host disease. Ann
report from the Mount Sinai Acute GVHD International Consortium. Hematol. 2015;94:1025-1032.
Biol Blood Marrow Transplant. 2016;22:4-10. 43. Harris AC, Ferrara JL, Braun TM, et al. Plasma biomarkers of lower
17. Ruutu T, Gratwohl A, de Witte T, et al. Prophylaxis and treatment of gastrointestinal and liver acute GVHD. Blood. 2012;119:2960-2963.
GVHD: EBMT-ELN working group recommendations for a standardized 44. Ferrara JL, Harris AC, Greenson JK, et al. Regenerating islet-derived 3-
practice. Bone Marrow Transplant. 2014;49:168-173. alpha is a biomarker of gastrointestinal graft-versus-host disease.
18. Mielcarek M, Furlong T, Storer BE, et al. Effectiveness and safety of lower Blood. 2011;118:6702-6708.
dose prednisone for initial treatment of acute graft-versus-host disease: 45. Reinhardt K, Foell D, Vogl T, et al. Monocyte-induced development of
a randomized controlled trial. Haematologica. 2015;100:842-848. Th17 cells and the release of S100 proteins are involved in the
19. MacMillan ML, Weisdorf DJ, Wagner JE, et al. Response of 443 patients pathogenesis of graft-versus-host disease. J Immunol. 2014;193:
to steroids as primary therapy for acute graft-versus-host disease: 3355-3365.
comparison of grading systems. Biol Blood Marrow Transplant. 2002;8: 46. Rodriguez-Otero P, Porcher R, Peffault de Latour R, et al. Fecal calpro-
387-394. tectin and alpha-1 antitrypsin predict severity and response to corti-
20. Levine JE, Logan B, Wu J, et al. Graft-versus-host disease treatment: costeroids in gastrointestinal graft-versus-host disease. Blood. 2012;
predictors of survival. Biol Blood Marrow Transplant. 2010;16: 119:5909-5917.
1693-1699. 47. Chiusolo P, Metafuni E, Giammarco S, et al. Role of fecal calprotectin as
21. Gooley TA, Chien JW, Pergam SA, et al. Reduced mortality after allo- biomarker of gastrointestinal GVHD after allogeneic stem cell trans-
geneic hematopoietic-cell transplantation. N Engl J Med. 2010;363: plantation. Blood. 2012;120:4443-4444.
2091-2101. 48. Oikawa T, Kamimura Y, Akiba H, et al. Preferential involvement of Tim-
22. Jenq RR, van den Brink MR. Allogeneic haematopoietic stem cell 3 in the regulation of hepatic CD8þ T cells in murine acute graft-versus-
transplantation: individualized stem cell and immune therapy of host disease. J Immunol. 2006;177:4281-4287.
cancer. Nat Rev Cancer. 2010;10:213-221. 49. Veenstra RG, Taylor PA, Zhou Q, et al. Contrasting acute graft-versus-
23. Lochmanová G, Jedli cková L, Pote sil D, et al. Potential biomarkers for host disease effects of Tim-3/galectin-9 pathway blockade dependent
early detection of acute graft-versus-host disease. Proteomics Clin Appl. upon the presence of donor regulatory T cells. Blood. 2012;120:
2012;6:351-363. 682-690.
24. Lau J, Sartor M, Bradstock KF, et al. Activated circulating dendritic cells 50. Hansen JA, Hanash SM, Tabellini L, et al. A novel soluble form of Tim-3
after hematopoietic stem cell transplantation predict acute graft- associated with severe graft-versus-host disease. Biol Blood Marrow
versus-host disease. Transplantation. 2007;83:839-846. Transplant. 2013;19:1323-1330.
25. Magenau JM, Qin X, Tawara I, et al. Frequency of CD4(þ)CD25(hi) 51. Luft T, Conzelmann M, Benner A, et al. Serum cytokeratin-18 fragments
FOXP3(þ) regulatory T cells has diagnostic and prognostic value as a as quantitative markers of epithelial apoptosis in liver and intestinal
biomarker for acute graft-versus-host-disease. Biol Blood Marrow graft-versus-host disease. Blood. 2007;110:4535-4542.
Transplant. 2010;16:907-914. 52. Luft T, Dietrich S, Falk C, et al. Steroid-refractory GVHD: T-cell attack
26. Teshima T, Maeda Y, Ozaki K. Regulatory T cells and IL-17eproducing within a vulnerable endothelial system. Blood. 2011;118:1685-1692.
cells in graft-versus-host disease. Immunotherapy. 2011;3:833-852. 53. Okamoto T, Takatsuka H, Fujimori Y, et al. Increased hepatocyte growth
27. Ranganathan P, Heaphy CE, Costinean S, et al. Regulation of acute graft- factor in serum in acute graft-versus-host disease. Bone Marrow
versus-host disease by microRNA-155. Blood. 2012;119:4786-4797. Transplant. 2001;28:197-200.
28. Xie LN, Zhou F, Liu XM, et al. Serum microRNA155 is increased in pa- 54. Paczesny S, Krijanovski OI, Braun TM, et al. A biomarker panel for acute
tients with acute graft-versus-host disease. Clin Transplant. 2014;28: graft-versus-host disease. Blood. 2009;113:273-278.
314-323. 55. Paczesny S, Braun TM, Levine JE, et al. Elafin is a biomarker of graft-
29. Wang Y, Zhao X, Ye X, et al. Plasma microRNA-586 is a new biomarker versus-host disease of the skin. Sci Transl Med. 2010;2:13ra2.
for acute graft-versus-host disease. Ann Hematol. 2015;94:1505-1514. 56. Chacon AH, Farooq U, Shiman MI, Elgart GW. Elafin: a possible new
30. Xiao B, Wang Y, Li W, et al. Plasma microRNA signature as a nonin- biomarker and immunohistochemical stain for pre-engraftment syn-
vasive biomarker for acute graft-versus-host disease. Blood. 2013;122: drome. J Am Acad Dermatol. 2013;69:e102-103.
3365-3375. 57. Bartel DP. MicroRNAs: target recognition and regulatory functions. Cell.
31. Vander Lugt MT, Braun TM, Hanash S, et al. ST2 as a marker for risk of 2009;136:215-233.
therapy-resistant graft-versus-host disease and death. N Engl J Med. 58. Ng EK, Chong WW, Jin H, et al. Differential expression of microRNAs in
2013;369:529-539. plasma of patients with colorectal cancer: a potential marker for
32. Ponce DM, Hilden P, Mumaw C, et al. High day 28 ST2 levels predict for colorectal cancer screening. Gut. 2009;58:1375-1381.
acute graft-versus-host disease and transplant-related mortality after 59. Li LM, Hu ZB, Zhou ZX, et al. Serum microRNA profiles serve as novel
cord blood transplantation. Blood. 2015;125:199-205. biomarkers for HBV infection and diagnosis of HBV-positive hep-
33. Lv M, Ye HG, Zhao XS, et al. Ceruloplasmin is a potential biomarker for atocarcinoma. Cancer Res. 2010;70:9798-9807.
aGVHD following allogeneic hematopoietic stem cell transplantation. 60. Zampetaki A, Kiechl S, Drozdov I, et al. Plasma microRNA profiling
PLoS One. 2013;8:e58735. reveals loss of endothelial miR-126 and other microRNAs in type 2
34. Choi SW, Kitko CL, Braun T, et al. Change in plasma tumor necrosis diabetes. Circ Res. 2010;107:810-817.
factor receptor 1 levels in the first week after myeloablative allogeneic 61. Mas VR, Dumur CI, Scian MJ, et al. MicroRNAs as biomarkers in solid
transplantation correlates with severity and incidence of GVHD and organ transplantation. Am J Transplant. 2013;13:11-19.
survival. Blood. 2008;112:1539-1542. 62. Tili E, Croce CM, Michaille JJ. miR-155: on the crosstalk between
35. Kitko CL, Paczesny S, Yanik G, et al. Plasma elevations of tumor necrosis inflammation and cancer. Int Rev Immunol. 2009;28:264-284.
factor receptor 1 at day 7 postallogeneic transplant correlate with 63. Mitchell PS, Parkin RK, Kroh EM, et al. Circulating microRNAs as stable
graft-versus-host disease severity and overall survival in pediatric blood-based markers for cancer detection. Proc Natl Acad Sci USA. 2008;
patients. Biol Blood Marrow Transplant. 2008;14:759-765. 105:10513-10518.
36. Willems E, Humblet-Baron S, Dengis O, et al. Elevations of tumor ne- 64. Chang J, Guo JT, Jiang D, et al. Liver-specific microRNA miR-122 en-
crosis factor receptor 1 at day 7 and acute graft-versus-host disease after hances the replication of hepatitis C virus in nonhepatic cells. J Virol.
allogeneic hematopoietic cell transplantation with nonmyeloablative 2008;82:8215-8223.
conditioning. Bone Marrow Transplant. 2010;45:1442-1448. 65. Dinarello CA. Interleukin-1 in the pathogenesis and treatment of in-
37. August KJ, Chiang KY, Bostick RM, et al. Biomarkers of immune acti- flammatory diseases. Blood. 2011;117:3720-3732.
vation to screen for severe, acute GVHD. Bone Marrow Transplant. 66. Ito S, Barrett AJ. ST2: the biomarker at the heart of GVHD severity.
2011;46:601-604. Blood. 2015;125:10-11.
38. Dean RM, Fry T, Mackall C, et al. Association of serum interleukin-7 67. Reichenbach DK, Schwarze V, Matta BM, et al. The IL-33/ST2 axis
levels with the development of acute graft-versus-host disease. J Clin augments effector T-cell responses during acute GVHD. Blood. 2015;
Oncol. 2008;26:5735-5741. 125:3183-3192.
39. Thiant S, Labalette M, Trauet J, et al. Plasma levels of IL-7 and IL-15 68. Schmitz J, Owyang A, Oldham E, et al. IL-33, an interleukin-1elike
after reduced-intensity conditioned allo-SCT and relationship to cytokine that signals via the IL-1 receptorerelated protein ST2 and
acute GVHD. Bone Marrow Transplant. 2011;46:1374-1381. induces T helper type 2eassociated cytokines. Immunity. 2005;23:
40. Thiant S, Yakoub-Agha I, Magro L, et al. Plasma levels of IL-7 and IL-15 479-490.
in the first month after myeloablative BMT are predictive biomarkers 69. Henden AS, Hill GR. Cytokines in graft-versus-host disease. J Immunol.
of both acute GVHD and relapse. Bone Marrow Transplant. 2010;45: 2015;194:4604-4612.
1546-1552. 70. Remberger M, Ringden O, Markling L. TNF alpha levels are increased
41. Cho BS, Min CK, Kim HJ, et al. High levels of B cell activating factor during bone marrow transplantation conditioning in patients who
during the peritransplantation period are associated with a reduced develop acute GVHD. Bone Marrow Transplant. 1995;15:99-104.
incidence of acute graft-versus-host disease following myeloablative 71. McDonald GB, Tabellini L, Storer BE, et al. Plasma biomarkers of acute
allogeneic stem cell transplantation. Biol Blood Marrow Transplant. GVHD and nonrelapse mortality: predictive value of measurements
2010;16:629-638. before GVHD onset and treatment. Blood. 2015;126:113-120.
1564 A.M. Ali et al. / Biol Blood Marrow Transplant 22 (2016) 1552e1564

72. Levine JE, Logan BR, Wu J, et al. Acute graft-versus-host disease bio- 85. Geng H, Zhang GM, Li D, et al. Soluble form of T cell Ig mucin 3 is an
markers measured during therapy can predict treatment outcomes: a inhibitory molecule in T cellemediated immune response. J Immunol.
Blood and Marrow Transplant Clinical Trials Network study. Blood. 2006;176:1411-1420.
2012;119:3854-3860. 86. Monney L, Sabatos CA, Gaglia JL, et al. Th1-specific cell surface protein
73. Te Boome LC, Mansilla C, van der Wagen LE, et al. Biomarker profiling Tim-3 regulates macrophage activation and severity of an autoimmune
of steroid-resistant acute GVHD in patients after infusion of mesen- disease. Nature. 2002;415:536-541.
chymal stromal cells. Leukemia. 2015;29:1839-1846. 87. Sehrawat S, Reddy PB, Rajasagi N, et al. Galectin-9/TIM-3 interaction
74. Levine JE, Braun TM, Harris AC, et al. A prognostic score for acute graft- regulates virus-specific primary and memory CD8 T cell response. PLoS
versus-host disease based on biomarkers: a multicenter study. Lancet Pathog. 2010;6:e1000882.
Haematol. 2015;2:e21-e29. 88. Hastings WD, Anderson DE, Kassam N, et al. TIM-3 is expressed on
75. Chung B, Dudl E, Toyama A, et al. Importance of interleukin-7 in the activated human CD4þ T cells and regulates Th1 and Th17 cytokines.
development of experimental graft-versus-host disease. Biol Blood Eur J Immunol. 2009;39:2492-2501.
Marrow Transplant. 2008;14:16-27. 89. Chu PG, Weiss LM. Keratin expression in human tissues and neoplasms.
76. Chung B, Dudl EP, Min D, et al. Prevention of graft-versus-host disease Histopathology. 2002;40:403-439.
by antieIL-7Ra antibody. Blood. 2007;110:2803-2810. 90. Bivén K, Erdal H, Hägg M, et al. A novel assay for discovery and char-
77. Savani BN, Mielke S, Rezvani K, et al. Absolute lymphocyte count on acterization of pro-apoptotic drugs and for monitoring apoptosis in
day 30 is a surrogate for robust hematopoietic recovery and patient sera. Apoptosis. 2003;8:263-268.
strongly predicts outcome after T celledepleted allogeneic stem 91. Shulman HM, Kleiner D, Lee SJ, et al. Histopathologic diagnosis of
cell transplantation. Biol Blood Marrow Transplant. 2007;13: chronic graft-versus-host disease. National Institutes of Health
1216-1223. Consensus Development Project on Criteria for Clinical Trials in
78. Sarantopoulos S, Stevenson KE, Kim HT, et al. Altered B-cell homeo- Chronic Graft-versus-Host Disease, II: Pathology Working Group
stasis and excess BAFF in human chronic graft-versus-host disease. Report. Biol Blood Marrow Transplant. 2006;12:31-47.
Blood. 2009;113:3865-3874. 92. Matsumoto K, Nakamura T. Emerging multipotent aspects of hepato-
79. Ratanatharathorn V, Logan B, Wang D, et al. Prior rituximab correlates cyte growth factor. J Biochem. 1996;119:591-600.
with less acute graft-versus-host disease and better survival in B-cell 93. Nakamura T, Nishizawa T, Hagiya M, et al. Molecular cloning and
lymphoma patients who received allogeneic peripheral blood stem cell expression of human hepatocyte growth factor. Nature. 1989;342:
transplantation. Br J Haematol. 2009;145:816-824. 440-443.
80. Ye Q, Wang L, Wells AD, et al. BAFF binding to T celleexpressed BAFF-R 94. Shiota G, Okano J, Kawasaki H, et al. Serum hepatocyte growth factor
costimulates T cell proliferation and alloresponses. Eur J Immunol. levels in liver diseases: clinical implications. Hepatology. 1995;21:
2004;34:2750-2759. 106-112.
81. Takashima S, Kadowaki M, Aoyama K, et al. The Wnt agonist R-spon- 95. Hino M, Inaba M, Goto H, et al. Hepatocyte growth factor levels in bone
din1 regulates systemic graft-versus-host disease by protecting intes- marrow plasma of patients with leukaemia and its gene expression in
tinal stem cells. J Exp Med. 2011;208:285-294. leukaemic blast cells. Br J Cancer. 1996;73:119-123.
82. Foell D, Roth J. Proinflammatory S100 proteins in arthritis and auto- 96. Taniguchi T, Toi M, Inada K, et al. Serum concentrations of hepatocyte
immune disease. Arthritis Rheum. 2004;50:3762-3771. growth factor in breast cancer patients. Clin Cancer Res. 1995;1:
83. Foell D, Wittkowski H, Ren Z, et al. Phagocyte-specific S100 1031-1034.
proteins are released from affected mucosa and promote immune 97. Nonomura K, Yamanishi K, Yasuno H, et al. Up-regulation of elafin/
responses during inflammatory bowel disease. J Pathol. 2008;216: SKALP gene expression in psoriatic epidermis. J Invest Dermatol. 1994;
183-192. 103:88-91.
84. Kuchroo VK, Umetsu DT, DeKruyff RH, Freeman GJ. The TIM gene 98. Pfundt R, Wingens M, Bergers M, et al. TNF-alpha and serum induce
family: emerging roles in immunity and disease. Nat Rev Immunol. SKALP/elafin gene expression in human keratinocytes by a p38 MAP
2003;3:454-462. kinase-dependent pathway. Arch Dermatol Res. 2000;292:180-187.