University of Groningen
Genetic susceptibility for inflammatory bowel disease across ethnicities and diseases
van Sommeren, Suzanne
DOI:
10.33612/diss.100597247
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from
it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2019
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
van Sommeren, S. (2019). Genetic susceptibility for inflammatory bowel disease across ethnicities and
diseases. Rijksuniversiteit Groningen. https://doi.org/10.33612/diss.100597247
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
CHAPTER
7
Clinical and genetic characterization
of patients with gastro-intestinal
graft-vs.-host disease (GI-GVHD) after
allogeneic hematopoietic cell
transplantation suggests a role for JAK2,
IL2RA and IL2.
S. van Sommeren, H.J.M. de Jonge, J. Kuball, L. te Boome, E. Vellenga, R.K. Weersma and G. Huls
ABSTRACT
Allogeneic hematopoietic cell transplantation (HCT) is the most effective immunotherapy available for a variety of hematological malignancies. Besides inducing graft-vs.-tumor effects, alloreac-tivity mediated by donor T-cells affects normal host tissues (partic-ularly the skin, liver and gastrointestinal (GI) tract), manifesting as graft-vs.-host disease (GVHD). GVHD, and especially GI-GVHD, is one of the major causes of mortality and morbidity after allogeneic HCT. Clinically and endoscopically GI-GVHD resembles inflammatory bowel disease (IBD). Genetic association studies have implicated several genes to be associated with GVHD and IBD, including NOD2, a well-known IBD gene.
In the current study, we studied a clinically well character-ized Dutch cohort of 185 patients, who all received a non- myeloablative conditioning regimen (Fludarabine + 2 Gy TBI), and donors. Of this cohort 35.7% developed substantial GI-GVHD (stage 2, 3 or 4). We hypothesized that genetic risk variants of IBD also might play a role in GI-GVHD. Therefore, both recipients and donors were genotyped using the Immunochip, a custom- made array including ~200,000 genetic variants with dense coverage of immune related genes.
We could not replicate the previously reported genetic associa-tions at NOD2 for GVHD in GI-GVHD or in overall GVHD. Inter-estingly, we identified suggestive signals (p-value < 1 × 10–4) for
8 loci. Loci harboring JAK2, IL2RA, and IL2 contained multiple genetic variants with signals for genetic association (JAK2 p-value 5.8 × 10–5, IL2RA p-value 7.3 × 10–6, IL2 p-value 7.45 × 10–5).
This implies that genetic variation in the immune system is likely to partly underlie the pathogenesis of (GI-) GVHD. Our results are in line with the recent reports which demonstrate that both JAK2 and IL2RA are promising therapeutic targets.
97
INTRODUCTION
Allogeneic hematopoietic cell transplantation (HCT) is a potentially curative procedure for a variety of hematologic malignancies and is nowadays the most effective tumor immuno-therapy available.1 Through the so called
graft-vs.-tumor (GVT) effect, donor T cells (graft) can produce durable immunologic control or eradication of residual malignancy in the recipi-ent. Unfortunately, the alloreactivity of donor T cells is not solely directed against tumor cells but can produce deleterious effects against normal host tissue in the recipient, manifesting as graft-vs.-host disease (GVHD). In HCT the activity of T cells against normal host tissue is strength-ened because the recipient’s tissue is damaged by the underlying disease and the conditioning regimen preparing the recipient for HCT. Dam-aged tissue conducts a pro-inflammatory state by producing pro-inflammatory cytokines and activating antigen presenting cells. Particularly the conditioning regime results in destruction of the barrier function of the gastrointestinal tract (GI), enabling microbial molecules to in-teract with antigen presenting cells, resulting in a stronger immune response by the donor T cells.2 The GI-tract is also one of the main sites of
acute GVHD, mostly leading to severe (bloody) diarrhea. Acute GVHD can also manifest in the skin and liver.3 The pathogenesis of chronic
GVHD is less clarified when compared to acute GVHD, possible mechanisms include persistent alloreactive T cells and loss of control by regu-latory T cells. The main sites of chronic GVHD include skin and lungs.4
Consistently reported factors associated with an increased risk of acute GVHD are recipient human leukocyte antigen (HLA) mismatching with the donor,5–7 alloimmunization of the
donor,8–11 the use of a female donor for male
recipients,8,10–12 and older patient age.10,12,13
Less consistently reported risk factors have in-cluded prior cytomegalovirus infection in the recipient,13–14 higher intensity of the
condition-ing regimen (irradiation),11–13 donor age,15 and
grafting with growth factor–mobilized blood cells.13,16 Despite optimal matching for these
factors, GVHD remains one of the major causes of mortality and morbidity after allogeneic HCT.
In search of additional factors influencing the risk of developing acute GVHD and to elucidate the pathogenesis of acute GVHD researchers have turned to genetic studies. Recent devel-opments in molecular genetics, which use DNA arrays for genotyping a large number of single nucleotide polymorphisms (SNPs) in every in-dividual, have fueled research to identify can-didate genetic associations with acute GVHD. Indeed, various SNPs have been associated with acute GVHD.17–22 Several groups working on
immune mediated diseases aimed to maximize the yield of clinical significant loci in immune mediated diseases by developing the Immu-nochip, a custom made genotyping chip that holds nearly 200,000 SNPs concentrated in loci containing genes with known immune related functions. The Immunochip has been widely used to identify genetic associations with multi-ple immune mediated diseases.23–25 A trait that
has been very successfully studied with the Im-munochip are the inflammatory bowel diseases (IBD) such as Crohns disease (CD) and ulcerative colitis (UC), with currently 200 associated loci.26
IBD and GI-GVHD have several resemblances: there are shared symptoms such as diarrhea and weight loss and endoscopical findings of GI ulceration and pathological characteristics can be similar. Furthermore, the basis of both diseases is immune mediated and a few IBD loci have been reported in GVHD, examples are the CD associated genes NOD2 and CTLA4.17
Therefore, we hypothesized that genetic risk variants of IBD might also play a role in GI-GVHD and explored the custom made Immunochip in a cohort of well-defined patients with GI-GVHD who received non-myeloablative conditioning (Fludarabine + 2 Gy TBI).
MATERIAL AND METHODS
Patients
The source population included randomly se-lected consecutive patients with varying hemato-logical malignancies who all received allogeneic
98
HCT after the same non- myeloablative condi-tioning regimen (fludarabine + 2 Gy TBI (total body irradiation) at the University Medical Cen-ter Groningen and University Medical CenCen-ter Utrecht, the Netherlands between 2001 and 2011. The cohort included only recipients with HLA-matched related donors. Patient-donor HLA matching was determined on the basis of genotyping for HLA-A, B, C, HLA-DRB1 and DQB1 (so 10/10 matched). All patients re-ceived T cell–replete bone marrow or growth factor-mobilized blood cell grafts, and cyclospo-rine plus mycophenylate as post-engraftment immune suppression. All clinical data were prospectively collected and retrospectively re-viewed. Acute GVHD was graded according to Glucksberg’s grading system2, shown in table 1.
The endpoints were defined as development of acute GVHD, recurrent disease, death or lost to follow-up. Additional information regarding time to the endpoints and cause of death was documented. All recipients and donors pro-vided written informed consent in accordance with the Declaration of Helsinki, and the study was approved by all participating institutional review boards.
Genotyping and quality control
DNA was extracted from blood samples of both recipients undergoing HCT and their
donors. Genotyping was performed using the Immunochip, an Illumina Infinium mi-croarray comprising 196,524 SNPs and small indel markers selected based on results from genome-wide association studies of 12 different immune- mediated diseases. Genotype calling was performed using Genomestudio software (Illumina).
Sample and marker quality control were per-formed using PLINK.27 Samples with low callrate
(<98% of markers) were removed. There were no samples with outlying heterozygosity rate after removing samples with low callrate. Gender was predicted from genotype data using Genome-studio software and HLA haplotypes were im-puted using HLA-IMP.28 Predicted gender and
HLA haplotypes were compared to phenotype data and classical HLA typing and samples with mismatches were removed. In total 26 out of 369 samples were removed during sample QC.
Marker quality control consisted of several phases. First, markers that are not present on autosomes were removed. Because the content of Immunochip is partly based on individual resequencing efforts and could be false posi-tives, we removed markers that did not repli-cate in 1000 Genomes Phase 2.29 Furthermore,
markers with a low callrate (<98% of samples), a minor allele frequency (MAF) below 0.01, fail Hardy Weinberg equilibrium (p < 1 × 10–5) and
Table 1. Glucksberg’s grading system.
Stage Skin Liver (bilirubin) Gut (stool output)
0 no rash <2 mg/dl <500 mL/day
1 Rash < 25% BSA 2–3 mg/dl 500–1000 mL/day
2 Rash 25–50% BSA 3–6 mg/dl 1000–1500 mL/day
3 Rash > 50% BSA 6–15 mg/dl >1500 mL/day
4 Bullae >15 mg/dl >2000 ml/day or pain or ileus
Grade Skin (stage) Liver (stage) Gut (stage)
I 1–2 0 0
II 1–3 1 1
III 2–3 2–3 2–4
IV 2–4 2–4 2–4
99
markers with a significant different callrate between cases and controls (p < 0.05) were re-moved. In total 62,653 markers were removed, 133,871 markers were left for analyses. Associ-ated markers were manually inspected by con-trolling genotype cluster plots for correct geno-typing. Because the Immunochip was designed to densely cover 180 genetic loci, most SNPs have neighboring SNPs in linkage disequilib-rium (LD) with them represented on the Immu-nochip. If a SNP is significantly associated with GVHD, one would expect to find an association signal for correlated SNPs. Therefore, we only report a SNP to be associated with GVHD when additional SNPs in LD within a 500 kb window give additional signals of association.
Association and statistical analyses
To assess the correlation between genetic mark-ers and the development of GI-GVHD we se-lected patients who developed severe GI-GVHD (stage 2, 3 and 4 – figure 1)) and compared ge-netic markers with patients that did not develop any form of GVHD (controls). In this way we ended up removing multiple cases, but we hy-pothesized that by selecting more extreme well defined cases and controls without any form of GVHD we maintain a clean case-control dataset
and optimize the power to detect any genetic signals. In order to incorporate genotypes of both the recipients and donors we combined their risk alleles in an additive model (0,1,2,3 or 4 risk alleles) and tested this risk profile for asso-ciation using two statistical methods: a simple logistic regression and in order to incorporate time-to-event as a variable we also performed a Cox regression analysis.
45 SNPs in 25 loci have previously been re-ported in GVHD.17–22 We tested these, or the
best proxy (LD r2 > 0.8) for association with
GI-GVHD in our cohort. One of the previously associated loci, containing the gene NOD2, contains three risk SNPs.17 In order to
maxi-mize power for this gene we combined the risk alleles of all three SNPs. Finally, we tested all remaining SNPs for genome-wide association. Genome-wide significance was defined as a p-value below 5 × 10–8, suggestive significance
when p-value was below 1 × 10–4. In order to
assess the relationship of associated SNPs with IBD risk variants we retrieved p-values of the SNP under investigation with Crohn’s disease, ulcerative colitis and the combined inflamma-tory bowel disease analysis out of publically available data via Ricopili (www.broadinstitute. org/mpg/ricopili).
Figure 1. Development of GVHD and GI-GVHD.
Number of patients developing GVHD and gastrointestinal (GI) GVHD. GI GVHD is divided according to the Gluckberg’s grading system.
100
Table 2. Clinical parameters.
Source population Genetic analysis cohort Cases (grade
2,3,4 GI GVHD) (no GVHD)Controls Total number of patients - number 185 24 89
Gender - number (%) Male 108 (58.4%) 14 (58.3%) 52 (58.4%) Female 77 (41.6%) 10 (41.7%) 37 (41.6%) Age - years Mean 51 51 50 Median 54 52 55 Data available (%) 100% 100% 100% CMV positive (%) 53.5% 41.7% 55.6% Data available 61.6% 50% 70.8% Indication - number
Acute myeloid leukemia 75 10 37
Multiple myeloma 36 2 19
Acute lymphatic leukemia 16 3 8
Chronic lymphatic leukemia 13 2 7
Non-Hodgkin lymphoma 18 2 5
Chronic myeloid leukemia 10 2 5
Myelodysplastic syndrome 7 1 4
Other 10 2 4
Total number of donors - number 184 24 89
Gender - number (%) Male 94 (51.4%) 9 (37.5%) 45 (50.6%) Female 83 (45.4%) 15 (62.5%) 44 (49.4%) unknown 6 (3.3%) 0 0 Age - years Mean 50 47 51 Median 50 50 51 Data available (%) 56.8% 58.3% 65.2% CMV positive (%) 46.6% 57.9% 50.7% Data available 70.8% 79.2% 79.8%
Clinical parameters, including gender, age, CMV and indication for the whole clinical cohort and subset of genetic analysis cohort are presented. CMV cytomegalovirus
101
Overall survival (OS) was defined as time in months between date of SCT and death resulting from any cause or time between date of SCT and last follow-up time. Actuarial probabili-ties of OS (with death resulting from any cause) were estimated according to the Kaplan-Meier method.
Functional annotation and gene prioritizing
For every associated SNP we determined if the SNP, or any proxies with LD r2 > 0.8 has
func-tional consequences in translated proteins, in order to indicate biologically relevant muta-tions. Next we performed a number of analyses to prioritize genes within the newly associated loci. We assessed if the SNP (or proxies with LD r2 > 0.8) significantly influences gene
sion in a publically available database of expres-sion quantitative trait locus (eQTL) in peripheral blood samples.30 Finally, we used two network
approaches to identify genes that are highly connected to genes in the other associated loci. GRAIL creates such a network based on textual relationships among genes and DAPPLE uses protein-protein interactions to identify genes with known connections.31,32
RESULTS
Clinical data
We included 185 patients undergoing HCT for a variety of hematological indications (table 2). During a mean follow-up of 2.4 years 82 of our 185 (44.3%) patients developed GVHD. In 37/82 GVHD patients the GI tract was affected, of whom 29/82 patients classified as severe GI-GVHD (stage 2, 3 or 4) (figure 1). 29 patients (35.7% of all GVHD patients) developed a severe form of GI-GVHD (stage 2, 3 or 4).
The mortality rate is high in patients undergo-ing HCT, in total 94 (50.8%) patients died durundergo-ing follow up. Figure 2 shows the clinical outcome during follow up, focused on the causes of death, a separation has been made between patients that suffered from GVHD and patients that did not show any sign of GVHD.
Patients with GI-GVHD showed a signifi-cantly reduced overall survival compared to patient with GVHD without GI involvement (P = 0.001, figure 3A). Furthermore, figure 3B shows Kaplan Meier overall survival curves for patients with GI-GVHD, GVHD without GI- involvement and patients without GVHD, showing that patients with GI-GVHD have
Figure 2. Clinical outcome.
The clinical outcome during follow up is shown, a separation is made between patients that did or did not developed GVHD. The cause of death is specified.
102
Figure 3A. Survival of patients with GI-GVHD compared to patients developing GVHD without GI involvement.
Figure 3B. Survival of patients with GI-GVHD compared to patients developing GVHD without GI in-volvement and patients without GVHD.
(A) OS is shown in months for patients with GHVD and a distinction is made between GI-GVHD and GVHD without GI-involvement.
(B) OS is shown in months for patients without GHVD, patients with GVHD without GI-involvement and patients with GI-GVHD. P values are given for the overall comparison across all 3 groups.
103
within these three groups the most poor prog-nosis (P<0.001).
Genetic analyses
For our genetic analyses we used the data of 113 recipients and their donors. We removed 11 recipients and 10 donors because we did not have reliable post QC data of their corre-sponding donor or recipient. Furthermore we removed 48 recipients and their donors because the recipients developed GVHD without GI in-volvement (40 patients) or stage 1 GI-GVHD (8 patients). Of the 113 recipients, 24 developed severe GI-GVHD (cases) and they were compared to 89 recipient-donor sets who did not develop GVHD (controls). Clinical parameters for our case- control population are shown in table 2.
Replicating previously reported associations
Previously, 45 SNPs in 25 loci were found to be associated with GVHD. We did not find any of these SNPs, or SNPs in high LD with the reported SNP (proxy) in the loci significantly associated in our dataset with GI-GVHD or overall GVHD (Supplementary table 1). Also, we did not find an association with the combined NOD2 risk SNPs.
Immunochip wide analyses
Neither our logistic regression analysis nor Cox regression analysis identified hits with a p-value below a genome-wide significant sta-tistical threshold of 5 × 10–8. Including time as a
covariant in a Cox regression analysis we found eight suggestive hits, with p-values < 5 × 10–4.
The association results are presented in table 3. Indeed, 3 out of 8 suggestive hit SNPs cover loci that have been firmly established in IBD: rs61268988 in the IL2 locus, rs57143124 in the
JAK2 locus and rs942200 in the IL2RA locus.
figure 4 represents locus plots of these 3 loci, including the top SNP, surrounding SNPs and the position of genes relative to the hit SNPs. Only rs57143124 shows some association in IBD cases (p-value of 0.008 in Crohn’s disease, 0.03 in ulcerative colitis and 0.003 in combined analyses of the inflammatory bowel diseases, (supplementary table 2), Locus plots of the other suggestive hits harboring the genes PDE4B,
STK39 and PVT1 are presented in
supplemen-tary figure 1.
Functional annotation and gene prioritizing
None of the suggestive hit SNPs, or any of their proxies are protein altering or have functional
Table 3. Suggestive hits after genetic analysis. SNP Chr Position (bp) Candidate
genes Risk allele Reference allele Risk allele frequency P-value Hazard ratio
rs7367706 1 66.707.856 PDE48 A C 0.04 8.47 × 10–7 3.90 rs6714707 2 169.010.297 STK39 A G 0.27 1.05 × 10–5 2.00 rs61268988 4 122.977.466 IL2,IL21 A C 0.02 7.45 × 10–5 3.37 rs1016579 8 129.199.961 PVT1, BC009730 A G 0.02 9.26 × 10–7 10.63 rs57143124 9 4.958.009 JAK2 A G 0.03 5.80 × 10–5 3.41 rs942200 10 6.063.674 IL2RA, IL15RA A G 0.02 7.30 × 10–6 8.31 rs7128743 11 41.925.699 No genes A C 0.1 8.86 × 10–5 2.17 rs17728625 18 37.969.946 No genes G A 0.04 1.79 × 10–5 2.64
Suggestive hits after genetic analysis, with a p-value below 1 × 10–4 are presented. The table includes SNP name,
chro-mosome and basepair postion on the chrochro-mosome (build hg19), candidate genes in the locus, risk and reference allele with risk allele frequency and p-value and hazard ratio of the Cox regression analysis of development of severe GI GVHD. Chr. Chromosome. Bp basepair
104
Figure 4A. Locusplot IL2A.
consequences in translated proteins (Supple-mentary table 3). Only SNP rs6714707 influ-ences gene expression of a nearby gene (cis eQTL), STK39, a gene that encodes a serine/ threonine kinase that is thought to function in the cellular stress response pathway (Supple-mentary table 4). A proxy of SNP rs6714707, SNP rs 6749447 influences gene expression of a gene on a different chromosome (trans eQTL),
RPAP2 on chromosome 1. This gene encodes
a polymerase II associated protein. How these eQTL effects would influence disease pathogen-esis remains uncertain. The network approach DAPPLE highlights the protein-protein interac-tion between JAK2 and proteins encoded by two genes from the same locus: IL2RA and IL15RA (Supplementary figure 2).
DISCUSSION
We report the first study that investigates the ge-netic basis of GI-GVHD using the Immunochip, a custom made genotyping chip focusing on genomic loci involved in immune-mediated dis-eases. Further, we used a cohort that received the same non-myeloablative conditioning reg-imen prior to allogeneic hematopoietic cell transplantation (i.e. Flu + 2 Gy TBI). In the ex-tremes of our cohort, i.e. 24 SCT recipients that developed severe GI-GVHD and 89 recipients
that did not develop GVHD, we found 8 ge-nomic loci with a suggestive association with GI-GVHD.
The locus covered by SNP rs57143124 con-tains the gene JAK2 (janus-activated kinase 2), a tyrosine kinase that functions as a signaling molecule in, amongst others, both the myeloid and lymphoid lineage of hematopoiesis. The V617F mutation in JAK2 is a well known and strongly associated risk factor for several my-eloproliferative neoplasms like polycythemia vera, essential trombocythemia and primary myelofibrosis.33 SNPs correlated with this
mutation are not associated with GI-GVHD in our dataset (data not shown). Ruxolitinib is a therapeutic agent that diminishes proliferation of cells in myeloproliferative neoplasms by dis-abling signaling by inhibiting JAK1 and JAK2. In 2011 it was approved as a treatment for pri-mary myelofibrosis, where it diminishes spleno-megaly and improves quality of life,34 and since
mutations in all four JAK family members, in-cluding JAK2, are increasingly being described in lymphoid malignancies like ALL the possible therapeutic range of JAK inhibitors is broadened. Interestingly, ruxolitinib caught attention in the GVHD field when it was shown to reduce development of GVHD after non-HLA matched allogeneic hematopoietic cell transplantation without diminishing the graft versus tumor ef-fect in a murine model.35 A retrospective study
105
Figure 4B. Locusplot JAK2.
Figure 4C. Locusplot IL2RA.
Regional locus plots for the IL2-locus (a), the JAK2-locus (b) and the IL2RA-locus (c). The x-axis represents the position of the SNPs on the chromosome (build hg19). Genes are depicted in the lower box. The y-axis represents the –log p-value. The name of the top SNP in the locus is presented in the figure, linkage disequilibrium in r2 in relation to the top SNP is
106
showed response rates >80% after treatment with ruxolitinib in corticosteroid refractory GVHD patients.36 Our findings emphasize the
importance of JAK2 signaling from a genetic perspective and are in line with the intriguing clinical data with ruxolutinib in animal models and clinical practice.
Cytokine IL2 is considered the general acti-vator of proliferation of Tcells, the significance of IL2 for GI-GVHD is accentuated by a genetic association with SNP rs61268988 situated near the gene coding for IL2 and an association with rs942200 that covers the IL2RA locus, which encodes the alpha chain of the IL2 receptor. Interestingly, IL2RA has a proteprotein in-teraction with JAK2. Although the gene IL2RA has never been associated with GVHD, Dacli-zumab, a monoclonal antibody against IL2RA (CD25) is an established and effective treatment for steroid-refractory acute GVHD.37 IL2 has
previously been reported by Chien et al. to be associated with GVHD, but only based on the donor genome.17
The loci encompassing JAK2, IL2 and IL2RA are established IBD loci, and although the risk SNPs in GI-GVHD and IBD seem to reflect other underlying genetic variation in the loci and the small number of suggested loci makes it impos-sible to conclude that we see a true enrichment of IBD genes within the list of suggested associ-ated loci, it remains possible that similar disease pathways underlie GI-GVHD and IBD.
Previous studies established replication of genetic associations in GVHD. Although our cohort is similar in size compared to some of the cohorts published in literature, we did not rep-licate any of the previously reported GVHD loci, including NOD2. A possible explanation could be the conditioning regimen, other cohorts con-sist of patients generally treated with myeloabla-tive conditioning regimen, while all patients in our cohort received a non- myeloablative condi-tioning regimen. It could be hypothesized that genes involved in integrity of the epithelial lin-ing and inflammation are differently involved, depending on conditioning strategy. Further, it is possible that due to the relative small sample sizes the results from previously reported and
our cohorts are liable to sampling bias and are therefore less consistent in reported candidate loci. The relatively small sample size might also explain why we only find low frequency SNPs with a suggestive association, as we only have enough power to detect these SNPs. Further-more, although we tried to investigate a homo-geneous cohort of patients by including only patients with a severe form of GI GVHD, it might be possible that GVHD remains a more hetero-geneous disease in terms of underlying genetic make-up. This warrants a larger independent and uniformly genotyped cohort to replicate our and previously published findings, which would also give enough statistical power to reach ge-nome wide significance for suggestive hits. In the mean time, it seems reasonable to focus our interest to gene products in newly suggested loci, because they might be good therapeutic targets as was shown for JAK2 and IL2RA.
REFERENCES
1. Gooley TA, Chien JW, Pergam SA, et al. Reduced mortality after allogeneic hematopoietic-cell transplantation. N Engl J Med. 2010;363(22): 2091–2101.
2. Ferrara JLM, Levine JE, Reddy P, Holler E. Graft-versus-host disease. Lancet. 2009;373(9674): 1550–1561.
3. Glucksberg H, Storb R, Fefer A, et al. Clinical man-ifestations of graft-versus-host disease in human recipients of marrow from HLA-matched sibling donors. Transplantation. 1974.18(4);295–304. 4. Socie G, Stone JV, Wingard JR et al.
Long-term survival and late deaths after allogeneic bone marrow transplantation. Late Effects Working Committee of the International Bone Marrow Transplant Registry. N Engl J Med. 1999;341(18):14–21.
5. Anasetti C, Beatty PG, Storb R, et al. Effect of HLA incompatibility on graft-versus-host disease, re-lapse, and survival after marrow transplantation for patients with leukemia or lymphoma. Hum
Immunol. 1990;29(2):79–91.
6. Beatty PG, Clift RA, Mickelson EM, et al. Mar-row transplantation from related donors other than HLA-identical siblings. N Engl J Med. 1985;313(13):765–771.
107
7. Lee SJ, Klein J, Haagenson M, et al.High-resolu-tion donor-recipient HLA matching contributes to the success of unrelated donor marrow trans-plantation. Blood. 2007;110(13):4576–4583. 8. Gale RP, Bortin MM, van Bekkum DW, et al. Risk
factors for acute graft-versus-host disease. Br J
Haematol. 1987;67(4):397–406.
9. Flowers ME, Pepe MS, Longton G, et al. Previ-ous donor pregnancy as a risk factor for acute graftversus-host disease in patients with aplastic anaemia treated by allogeneic marrow transplan-tation.Br J Haematol. 1990;74(4):492–496. 10. Weisdorf D, Hakke R, Blazar B, et al. Risk
fac-tors for acute graft-versus-host disease in histo-compatible donor bone marrow transplantation.
Transplantation. 1991;51(6):1197–1203.
11. Nash RA, Pepe MS, Storb R, et al. Acute graft versus-host disease: analysis of risk factors after allogeneic marrow transplantation and prophy-laxis with cyclosporine and methotrexate. Blood. 1992;80(7):1838–1845.
12. Eisner MD, August CS. Impact of donor and recip-ient characteristics on the development of acute and chronic graft-versus-host disease following pediatric bone marrow transplantation. Bone
Marrow Transplant. 1995;15(5):663–668.
13. Hahn T, McCarthy PL, Jr., Zhang MJ, et al. Risk factors for acute graft-versus-host disease af-ter human leukocyte antigen-identical sibling transplants for adults with leukemia. J Clin Oncol. 2008;26(35):5728–5734.
14. Jacobsen N, Badsberg JH, Lonnqvist B, et al. Graft-versus-leukaemia activity associated with CMV-seropositive donor, post-transplant CMV infection, young donor age and chronic graft-versus- host disease in bone marrow allograft re-cipients. The Nordic Bone Marrow Transplanta-tion Group. Bone Marrow Transplant. 1990;5(6): 413–418.
15. Kollman C, Howe CW, Anasetti C, et al. Do-nor characteristics as risk factors in recipients after transplantation of bone marrow from un-related donors: the effect of donor age. Blood. 2001;98(7):2043–2051.
16. Cutler C, Giri S, Jeyapalan S, Paniagua D, Viswanathan A, Antin JH. Acute and chronic graft versus-host disease after allogeneic pe-ripheral bloodstem-cell and bone marrow transplantation:a meta-analysis. J Clin Oncol. 2001;19(16):3685–3691.
17. Chien JW, Zhang XC, Fan W, et al. Eval-uation of published single nucleotide
polymorphisms associated with acute GVHD.
Blood. 2012;119(22):5311–5319.
18. Espinoza JL, Takami A, Onizuka M et al. Re-cipient PTPN22 - 1123 C/C Genotype predicts acute graft-versus-host disease after HLA fully matched unrelated bone marrow transplantation for hematologic malignancies. Biol Blood Marrow
Transplant. 2013;19(2):240–246.
19. Xiao HW, Luo Y, Lai XY et al. Donor TLR9 gene tagSNPs influence susceptibility to aGVHD and CMV reactivation in the allo-HSCT set-ting without polymorphisms in the TLR4 and NOD2 genes. Bone Marrow Transplantation. 2014;49(2):241–247.
20. Cao W, Xiao H, Lai X et al. Genetic variation in the mycophenolate mofetil target enzyme are associated with acute GVHD risk after related and unrelated hematopoietic cell transplanta-tion. Biol Blood Marrow Transplant. 2012;18(2): 273–279.
21. Zhang L, Mao L, Xu J. Transforming growth factor-B1 polymorphisms and graft-versus-host disease risk: a meta-analysis. Oncotarget. 2015;7(3):2455–2461.
22. Rayes A, Morrow AL, Payton LR, Lake KE, Lane A, Davies SM. A genetic modifier of the gut mi-crobiome nfluences the risk of graft-versus-host disease and bacteremia after hematopoietic stem cell transplantation. Biol Blood Marrow Transplant. 2016;22(3):418–422.
23. Trynka G, Hunt KA, Bocket NA, et al. Dense geno-typing identifies and localizes multiple common and rare variants associated signals in celiac dis-ease. Nat Genet. 2011;43(12):1193–1201. 24. Eyre S, Bowes J, Diogo D, et al. High-density
genetic mapping identifies new susceptibility loci for rheumatoid arthritis. Nat Genet. 2012; 44(12):1336–1340.
25. Onengut-Gumuscu S, Chen WM, Burren O, et al. Fine mapping of type 1 diabetes susceptibility loci and evidence for colocalization of causal vari-ants with lymphoid gene enhancers. Nat Genet. 2015;47(4):381–386.
26. Liu JZ, van Sommeren S, Huang H, et al. Asso-ciation analyses identify 38 susceptibility loci for inflammatory bowel disease and highlight shared genetic risk across populations. Nat Genet. 2015;47(9):979–986.
27. Purcell S, Neale B, Todd-Brown K, et al. PLINK: a tool set for whole-genome association and pop-ulation based linkage analyses. Am J Hum Genet. 2007;81(3):559–575.
108
28. Leslie S, Donnely P, McVean G. A statistical method for predicting classical HLA alleles from SNP data. Am J Hum Genet. 2008;82(1);48–56. 29. The 1000 Genomes Project Consortium. A global
reference for human genetic variation. Nature. 2015;526(7571):68–74.
30. Westra HJ, Peters MJ, Esko T, et al. Systematic identification of trans eQTLs as putative drivers of known disease associations. Nat Genet. 2013;45 (10):1238–1243
31. Rossin EJ, Lage K, Raychaudhuri S, et al. Pro-teins encoded in genomic regions associated with immune mediated disease physically inter-act and suggest underlying biology. PLos Genet. 2011;7:e1001273.
32. Raychaudhuri S, Plenge RM, Rossin EJ, et al. Identifying relationships among genomic dis-ease regions: predicting genes at pathogenic SNP associations and rare deletions. PLoS Genet. 2009;5:e1000534.
33. Jones AV, Kreil S, Zoi K, et al. Widespread occurrence of the JAK2 V617F mutation in chronic myeloproliferative disorders. Blood. 2005;106(6):2162–2168.
34. Verstovsek S, Mesa RA, Gotlib J, et al. A double-blind, placebo-controlled trial of Ruxolitinib for Myelofibrosis. N Engl J Med. 2012;366(9):799–807.
35. Choi J, Cooper ML, Alahmari B, Ritchey J, Collins L, Holt M, DiPersio JF. Pharmacologic blockade of JAK1/JAK2 reduces GvHD and preserves the graft-versus-leukemia effect. PLoS One. 2014;9:e109799.
36. Zeiser R, Burchert A, Lengerke C, et al. Ruxoli-tinib in corticosteroid-refractory graft-versus-host disease after allogeneic stem cell trans-plantation: a multicenter survey. Leukemia. 2015;29(10):2062–2068.
37. Bordigoni P, Dimicoli S, Clement L, et al. Dacli-zumab, an efficient treatment for steroid-refrac-tory acute graft-versus-host disease. Br J Haema-tol. 2006;135(3):382–385
