Original Article

Bone Marrow Transplantation (2015) 50, 1563–1571; doi:10.1038/bmt.2015.191; published online 14 September 2015

There is a Corrigendum (1 June 2016) associated with this article.

Graft-Versus-Host Disease

Identification and validation of biomarkers associated with acute and chronic graft versus host disease

S S Ahmed1,2, X N Wang1,2, J Norden1, K Pearce1, E El-Gezawy1, S Atarod1, I Hromadnikova3, M Collin1, E Holler4 and A M Dickinson1,2

  1. 1Haematological Sciences, Institute of Cellular Medicine, Newcastle University, Newcastle-upon-Tyne, UK
  2. 2Alcyomics Ltd, Newcastle-upon-Tyne, UK
  3. 3Third Faculty of Medicine, Charles University, Prague, Czech Republic
  4. 4Regensburg University Hospital, Department of Hematology and Oncology, Regensburg, Germany

Correspondence: Professor AM Dickinson, Haematological Sciences, Institute of Cellular Medicine, Medical School, University of Newcastle, William Leech Building, Newcastle upon Tyne NE2 4HH, UK. E-mail: anne.dickinson@ncl.ac.uk

Received 6 May 2015; Revised 14 July 2015; Accepted 15 July 2015
Advance online publication 14 September 2015

Top

Abstract

Graft versus host disease (GVHD) is a major complication of haematopoietic SCT (HSCT). A number of inflammatory cytokines/chemokines are implicated in GVHD and have been identified in numerous single centre studies as potential biomarkers for acute and/or chronic GVHD. In this study, we analysed candidate inflammatory biomarkers (B-cell activating factor (BAFF), interleukin 33 (IL-33), CXCL10 and CXCL11) in a two-centre study. Biomarkers were evaluated pre-transplant and at serial time points post transplant in acute and chronic GVHD patient sera with time-matched control samples from patients without GVHD. Further validation was performed using the human skin explant assay, clinical GVHD biopsies and mRNA expression analysis. BAFF was significantly increased pre-transplant. BAFF, IL-33, CXCL10 and CXCL11 showed increased levels in acute GVHD patient sera and high protein expression in grades II–III of the in vitro skin explant graft versus host reaction (GVHR) group. BAFF, CXCL10 and CXCL11 also showed increased mRNA expression levels in clinical biopsies compared with the no/low-grade GVHD group. BAFF, CXCL10 and CXCL11 levels were increased in chronic GVHD patient sera. The results identify BAFF and CXCL10 as predictive biomarkers for acute GVHD and BAFF, CXCL10 and CXCL11 as useful diagnostic biomarkers for acute GVHD and chronic GVHD.

Top

Introduction

Haematopoietic SCT (HSCT) is increasingly used to treat patients suffering from malignant and non-malignant haematopoietic disorders. A frequent and potentially life threatening complication is GVHD whereby donor T cells attack host target organs causing substantial damage and apoptosis of cells giving rise to clinical symptoms of acute GVHD (aGVHD) followed by chronic GVHD (cGVHD). Severity of aGVHD is graded according to the number of organs involved and the severity of organ dysfunction1 and occurs shortly after transplant whereas cGVHD either overlaps with aGVHD or is more delayed (after 100 days). The aGVHD process largely involves apoptosis and necrosis2 with cGVHD having features resembling autoimmune diseases.3

Allogeneic HSCT patients are given a conditioning regimen to maximise cytoreductive therapy and decrease graft rejection, however, this can cause the cytokine storm.4 Cytokines, chemokines and their receptors have important roles in all phases of GVHD and their differential expression may be a determining factor of GVHD development and provide insight into its pathology.5, 6 To date several studies have investigated predictive and prognostic biomarkers for GVHD including cytokines and chemokines.7, 8, 9, 10, 11, 12, 13, 14 However, information regarding diagnostic, prognostic or predictive significance of these molecules in GVHD is limited. We selected candidate molecules from the literature which had previously been associated with either aGVHD or cGVHD15, 16, 17, 18 for further validation as predictive, diagnostic or prognostic biomarkers. Donor-derived T cells are involved in initiation of aGVHD but also have a role in cGVHD, which has been classified as an independent disease.1 The ability of patients to develop cGVHD after T-cell depletion suggests that other cell populations may have a role in its pathophysiology.19 B cells have an important role in the immune response associated with cGVHD and are involved in GVHD biology.3 cGVHD is associated with perturbed B-cell homeostasis and B-cell activating factor (BAFF) is a key regulator of normal B-cell homeostasis in both mice and humans.20

Elevated levels of soluble BAFF have been reported in patients with autoimmune disease21 and cGVHD.22 Interleukin 33 (IL-33) is associated with inflammasome activation and danger signalling.23 Recently, Toll-like receptors and Nucleotide Binding Oligomerization domain-like receptors have been investigated in GVHD.24, 25 Both Toll-like receptors and Nucleotide Binding Oligomerization domain-like receptors have a role in sensing danger signals26 and with caspase-1 are involved in the formation of the multi-protein complex, inflammasome, which upon recognition of pathogens leads to the release of IL-1β, IL-18 and IL-33.27 Chemokines are small proteins that mediate migration of leukocyte trafficking in vivo.28 The chemokines CXCL9, CXCL10 and CXCL11 are induced by interferon-γ, which is produced by Th1 cells. The receptor for these ligands, CXCR3, is predominantly expressed on the surface of Th1 cells and recent studies have demonstrated the involvement of the CXCR3 ligands in GVHD.17

In this study, the role of BAFF, IL-33, CXCL10 and CXCL11 was investigated in aGVHD and cGVHD, pre and post allo-HSCT patient sera in a training and validation set as biomarker candidates for prediction or prognosis of acute or cGVHD. Results were further validated in the clinical GVHD biopsies and the human in vitro skin explant model, which can be used to induce a graft versus host reaction (GVHR) using donor and recipient HLA-matched skin and blood sample, as an in vitro model of GVHD.29, 30, 31

Top

Materials and methods

Patients

Patient samples were obtained after informed consent. This research was approved by Newcastle and North Tyneside Research Ethics Committee. Independent patient cohort samples were collected from two centres; Royal Victoria Infirmary, Newcastle (training cohort) and University Hospital, Regensburg (validation cohort). Patient clinical data were collected from the European Group of Blood and Marrow Transplantation (EBMT) database (https://www.ebmt.org.) Patient clinical characteristics and sample/cohort size are summarised in Table 1. Sample/cohort size was based on previous studies in this research group and has been shown to be of adequate power. Patients at both centres received either myeloablative or reduced intensity conditioning regimens. Healthy blood donors (from Newcastle) were used as normal controls to determine baseline levels/expression of biomarkers and compared to samples from GVHD patients taken post HSCT. Diagnosis of aGVHD was according to clinical and histopathological criteria32 and cGVHD by clinical assessment based on NIH criteria.33


Patient samples

In both centres peripheral blood and serum were collected from HSCT recipient patients before conditioning (7 days pre- transplant) and on days 0, 7, 14 and 28 and months 3, 6 and 12 following HSCT. Recipient skin biopsies for skin explant experiments were obtained pre-transplant and diagnostic skin/gut biopsies (4mm) were obtained at suspected clinical aGVHD onset post transplant.

ELISA

Soluble BAFF (R&D Systems, Minneapolis, MN, USA) and soluble IL-33 (ApoTECH Corporation, Epalinges, Switzerland) levels were measured in patient sera in duplicate by ELISA kits following the manufacturer’s recommended procedure.

FACS analysis

CXCL10 and CXCL11 levels were measured with a Cytometric Bead Array flex set (BD Biosciences, San Jose, CA, USA) in duplicate. Data were acquired on a FACS Calibur and analysed using FCAP software (BD Biosciences).

Skin explant model

The skin explant assay was performed as previously described29, 30 (for brief explanation, see Supplementary data).

Immunohistochemistry

Immunohistochemistry was performed on 3μM sections from in vitro skin explants, skin or gut biopsies from GVHD patients. BAFF and IL-33 (ApoTECH), CXCL10 (Peprotech, London, UK) and CXCL11 expression (Santa Cruz, Dallas, TX, USA) staining was performed using an automated immunostainer (Ventana, Benchmark, Tucson, AZ, USA). Staining was performed in duplicate. Negative control sections were stained without the primary antibody. Expression analysis was performed by counting the percentage of positive cells in each section in proportion to the total number of cells.

RNA extraction, cDNA production and real-time PCR

RNA was extracted from healthy volunteer skin samples, clinical GVHD skin biopsies or skin explant assay samples, using the mirVana miRNA Isolation Kit (Ambion, Paisley, UK). cDNA was generated by random hexamer priming. Real-time PCR was performed using TaqMan assays for CXCL10 (Hs00171042_m1), CXCL11 (Hs00171138_m1) and BAFF (Hs00198106_m1) and the control gene GAPDH (Life Technologies, Paisley, UK). Reactions were set up in triplicate. A 7900 qRT-PCR system (Life Technologies) was used to run reactions and analysis (SDS 2.3 software, Applied Biosystems, Paisley, UK). The relative changes in RNA expression were calculated using the comparative ΔΔCt method.34

Statistical analysis

Non-parametric Mann–Whitney U test, two sample proportion test and chi-square tests were used to compare changes in cytokine/chemokine levels in skin explant assays, biopsies and patient sera. A comparison was made at each time point between patients with or without aGVHD or cGVHD. mRNA expression of biomarkers in positive (grade II>) and negative (less than grade II) skin explant assays were compared with medium alone controls. Skin explant GVHR grade was compared with clinical aGVHD outcome. The diagnostic performance of each biomarker was evaluated using Receiver Operating Characteristic (ROC) curve analysis (method explained in Supplementary data). Statistical analysis was carried out using Prism V5 (GraphPad, San Diego, CA, USA). A probability (P)<0.05 was considered to be statistically significant. P-value index *P<0.01, **P<0.001 and ***P<0.0001. Results are displayed in graphs as distribution of values measured with centre values as mean with s.e.m.

Top

Results

Cytokine/chemokine levels in HSCT patient sera

Serum BAFF (sBAFF) levels were measured in recipient HSCT patients in a training cohort (Newcastle), validation cohort (Regensburg) and healthy individuals. Interestingly, sBAFF levels were significantly elevated pre-transplant in the aGVHD group compared with the no-aGVHD group in the training and validation cohort (Table 2, Figure 1a). Following conditioning, sBAFF levels decreased on day 0 in the aGVHD group compared with pre-transplant levels in the same group in both cohorts (training, P=0.01) (validation, P=0.04) (Figure 1a). Post-transplant, sBAFF levels were significantly elevated on day 14 in the aGVHD group compared with the no-aGVHD group in the both cohorts (Table 2, Figure 1a). sBAFF levels in healthy individuals were significantly lower (P=0.0003) compared with HSCT patients pre-transplant (Figure 1a). Development of aGVHD usually precedes cGVHD,35 within the training cohort (n=54) 45% of no-aGVHD patients developed cGVHD, compared with 75% of patients with aGVHD (P=0.04). This confirms the risk of developing cGVHD is higher in the aGVHD group36 and therefore may confirm BAFF as a predictive biomarker in cGVHD as recently reported.16, 22 In cGVHD patients, sBAFF levels were significantly elevated at 12 months post transplant in the training cohort and 6 and 12 months post transplant in the validation cohort compared with the no-cGVHD group (Table 2, Figure 1b). The onset of cGVHD at 6 months post transplant correlated with increased sBAFF levels (P=0.02, data not shown) in the training cohort. Increased sBAFF levels at 12 months in both cohorts suggest that BAFF is indeed implicated in cGVHD and could serve as a diagnostic biomarker for cGVHD.

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

sBAFF and sIL-33 levels in HSCT patient sera measured pre-transplant and at selected time points post transplant. Results are given for the training cohort (black bars) and validation cohort (grey bars) in HSCT patients with and without aGVHD (left) and cGVHD (right). Graph shows mean sBAFF and sIL-33 levels (ng/mL) at each time point tested. Results are shown for (a) healthy individuals (Normal, in training cohort only n=8) and sBAFF levels in HSCT patients with and without aGVHD. (b) sBAFF levels in HSCT patients with and without cGVHD. (c) sIL-33 levels in HSCT patients with and without aGVHD. (d) sIL-33 levels in HSCT patients with and without cGVHD. P-values indicated as *P<0.01, **P<0.001 and ***P<0.0001.

Full figure and legend (136K)


Treatment of patients with cGVHD often comprises steroid administration, which can reduce BAFF levels.37, 38 Patients in the training cohort treated with steroids had lower sBAFF levels compared with patients without steroid treatment (data not shown). However, in the validation cohort, patients treated with steroids showed increased sBAFF levels at 6 months (P=0.002, data not shown). In both cohorts, the majority of patients were administered with low-dose steroids (1–2mg/kg), two patients from the training cohort and one patient from the validation cohort received high-dose steroids (>2mg/kg). Data analysis in relation to the different conditioning regimens showed no difference in sBAFF levels between patients receiving reduced intensity conditioning or myeloablative conditioning treatments (data not shown).

Next, we investigated levels of cytokine IL-33 in both cohorts. Soluble IL-33 levels were significantly higher in the aGVHD group compared with the no-aGVHD group on day 0 and day 28 post transplant in the training cohort and day 28 in the validation cohort (Table 2, Figure 1c) indicating a potential role of IL-33 in predicting GVHD. sIL-33 levels in patients with cGVHD showed a significant increase at 3 months post transplant (*P=0.03) in the cGVHD group, which could possibly be a continuation of aGVHD. However, in the validation cohort a significant decrease was observed in cGVHD patients compared with the no-cGVHD group (Table 2, Figure 1d).

We then analysed CXCL10 and CXCL11 levels. CXCL10 levels were significantly increased pre-transplant in the aGVHD group in the training cohort only. Post-transplant, CXCL10 was significantly decreased in the training cohort aGVHD group on day 0 but then significantly increased on days 14 and 28 post transplant compared with patients with no-aGVHD. Correspondingly, a significant increase was also observed in the validation cohort on days 14 and 28 (Table 2, Figure 2a) confirming an important role for CXCL10 in aGVHD. CXCL11 levels were also increased pre-transplant in aGVHD patients (*P=0.03) in the training cohort only. Post transplant a significant increase was observed in both cohorts on days 14 and 28 in aGVHD patients (Table 2, Figure 2c). In cGVHD patients, CXCL10 levels in the training cohort were significantly increased at 3, 6 and 12 months post transplant. This was confirmed at 6 and 12 months post transplant in the validation cohort (Table 2, Figure 2b). CXCL11 levels in cGVHD patients were significantly increased at 6 and 12 months post transplant in the training cohort and significantly increased at 3, 6 and 12 months in the validation cohort (Table 2, Figure 2d). The corresponding results from both cohorts confirm the suitability of CXCL10 and CXCL11 as diagnostic biomarkers for acute and cGVHD. The expression of pro-inflammatory chemokines in target tissues has been shown to be dependent on the conditioning regimen39 and reported to be significantly increased after conditioning in comparison with normal controls. We also demonstrated increased levels of both CXCL10 and CXCL11 levels after either myeloablative or reduced intensity conditioning (see Supplementary data).

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

sCXCL10 and sCXCL11 levels in HSCT patient sera measured pre-transplant and at different time points post transplant. Results are given for the training cohort (black bars) and validation cohort (grey bars) in HSCT patients with and without aGVHD (left) and cGVHD (right). Graph shows mean sCXCL10 and CXCL11 levels (pg/mL) at each time-point tested. (a) Healthy individuals (Norm, in training cohort only) and sCXCL11 levels in HSCT patients with and without aGVHD. (b) sCXCL10 levels in HSCT patients with and without cGVHD. (c) sCXCL11 levels in HSCT patients with and without aGVHD. (d) sCXCL11 levels in HSCT patients with and without cGVHD. P-values indicated as *P<0.01, **P<0.001 and ***P<0.0001.

Full figure and legend (131K)

ROC curves

Training and validation cohort results were combined to determine diagnostic performance of each biomarker by ROC curve analysis pre-/post-transplant for aGVHD and cGVHD. The 3-month post-transplant time point was excluded as no significant changes were observed. An area under the curve (AUC) between 0.9 and 1.0 is considered excellent, between 0.8 and 0.9 good, 0.7–0.8 fair and 0.6–0.7 poor. Biomarkers that gave an AUC of 75> are listed in Figure 3. For aGVHD, the AUC for IL-33 at day 0 was 0.78 and for CXCL10 at day 14 was 0.79. For cGVHD, the AUC was good for all biomarkers at 12 months post transplant (BAFF 0.85, IL-33 0.80, CXCL10 0.79 and CXCL11 0.75) at 6 months post transplant for BAFF (0.75) and pre-transplant for CXCL10 (0.78).

Figure 3.
Figure 3 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

ROC curve data for predicting aGVHD and cGVHD. (a) The AUC, s.e., P-value and optimal cutoff point is given for biomarkers, which showed an AUC of 0.75 or above for aGVHD or cGVHD patients. (b) A representative image of ROC analysis of HSCT patients with cGVHD 12 months post transplant.

Full figure and legend (122K)

Confirmatory studies on skin explants and clinical biopsies

Cytokine/chemokine expression in the human in vitro skin explant assay
 

We investigated BAFF, IL-33, CXCL10 and CXCL11 protein expression in skin explant assays. BAFF expression was analysed in 23 allogeneic skin explants. The GVHR results were correlated with clinical aGVHD outcome for each patient and an 83% correlation (data not shown) was observed as previously reported.31 Controls were skin incubated in medium alone or with autologous cells. BAFF expression in controls was similar to GVHR grade 0–I explant assays (54 and 56%, respectively), but significantly increased in grade II–III GVHR explants (mean positive cells 66%, P=0.04) (Figure 4a) demonstrating BAFF expression correlates with severity of GVHR. BAFF expression in normal healthy skin was significantly lower (P=0.01) compared with grades II–III GVHR (data not shown).

Figure 4.
Figure 4 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Cytokine/chemokine expression in in vitro skin explants of HSCT patients. Box plots show the overall percentage of cells, which stained positive for each cytokine/chemokine. (a) BAFF expression was significantly (P=0.04) increased in GVHR-positive skin explant assays (n=17 grades II–III GVHR) compared with GVHR-negative skin explants (n=6, grades 0–I GVHR). (b) IL-33 expression was significantly (P=0.004) increased in GVHR-positive skin explant assays (n=19 grades II–III GVHR) compared with GVHR-negative skin explants (n=7, grades 0–I GVHR). (c) CXCL10 and (d) CXCL11 expression was significantly increased (P=0.007 and P=0.01, respectively) in GVHR-positive skin explant assays compared with GVHR-negative skin explants. Representative immunohistochemistry images are given below each box plot showing expression in a GVHR-negative skin explant and a GVHR-positive skin explant for each cytokine/chemokine. All images were acquired using a light microscope (Leitz Wetzer) at × 50 magnification (1.32 OPL) and a digital camera (Canon, DS126071; Canon Utilities Browser EX software).

Full figure and legend (236K)

IL-33 expression was analysed in 26 skin explant assays and was significantly increased (P=0.004, Figure 4b) in grade II–III GVHR skin explants (mean positive cells 41%) compared with grade 0–I GVHR (15%, data not shown). Normal healthy skin showed low IL-33 expression in comparison with skin explant assays (data not shown).

CXCL10 and CXCL11 expression was analysed in 20 skin explant assays. CXCL10 expression was significantly increased in the II–III GVHR group in comparison with the control group (medium only) (P=0.007). No difference was observed between the control group and the grade 0–I GVHR group (Figure 4c). CXCL11 expression was significantly increased in grade II–III GVHR skin explants compared with grade 0–I GVHR group (P=0.01) (Figure 4d). Moreover, only basal levels of CXCL10 and CXCL11 were observed in normal healthy skin (data not shown).

Cytokine/chemokine expression in clinical biopsies

Clinical biopsies from HSCT patients were analysed for BAFF, IL-33 CXCL10 and CXCL11 protein expression by immunohistochemistry using clinical gastrointestinal (GI) tract biopsies (n=17, 8 patients with clinical aGVHD and 9 patients with no-aGVHD). High BAFF expression was observed in 6/8 aGVHD biopsies (Figure 5a), the no-aGVHD biopsies tested negative for BAFF staining (P=0.001). The aGVHD GI biopsies all showed positive IL-33 staining (66.7% positive cells P=0.01) (Figure 5b) whereas 7/9 biopsies from no-aGVHD patients were negative. To investigate further, mRNA expression was investigated in clinical skin biopsies taken from allo-HSCT patient’s pre-/post transplant. BAFF mRNA expression in pre-transplant biopsies was used as a calibrator for expression in post-transplant biopsies and the histopathological aGVHD grades were used for analysis. BAFF mRNA expression was 23-fold higher in aGVHD (grade II–III) patients (n=8) in comparison with no-aGVHD (grade 0–1) patients (n=15) (P=0.019) (Figure 6e) and significantly lower in normal healthy control skin biopsies compared with no-aGVHD patients (P=0.03) and aGVHD (P=0.007) (data not shown). CXCL10 and CXCL11 mRNA expression showed 10/17 (CXCL10) and 12/17 (CXCL11) aGVHD post-HSCT clinical biopsies tested positive. However, CXCL10 (Figure 5c) and CXCL11 (Figure 5d) expression was more strongly distinctive in biopsies from patients with aGVHD. CXCL10 and CXCL11 mRNA expression in normal and clinical aGVHD skin showed low levels in normal skin, and significantly increased in aGVHD skin (P=0.001 and P=0.0003, respectively) (Figures 6a and b). Moreover, gene expression of CXCL10 and CXCL11 was also significantly increased in GVHR (grades II–III) positive skin explants (Figures 6c and d) in comparison with the medium control (P=0.02, P=0.02).

Figure 5.
Figure 5 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Cytokine/chemokine expression in clinical biopsies of HSCT patients. Representative immunohistochemistry images of GI tract biopsies of aGVHD patients (left) and HSCT patients with no aGVHD (right). (a) Membrane-bound staining of BAFF in a gut biopsy of an aGVHD patient (left) and duodenum biopsy of HSCT patient with no aGVHD (right). (b) IL-33, (c) CXCL10 and (d) CXCL11 staining in a gut biopsy of an aGVHD patient (left) and of HSCT patient with no aGVHD right. All images were acquired using a light microscope (Leitz Wetzer) at × 50 magnification (1.32 OPL) and a digital camera (Canon, DS126071; Canon Utilities Browser EX software).

Full figure and legend (487K)

Figure 6.
Figure 6 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

CXCL10, CXCL11 and BAFF mRNA expression analysis in aGVHD clinical skin biopsies and in skin explant assays by RT-PCR. Graphs show mRNA expression in either clinical skin biopsies from HSCT aGVHD patients in comparison with expression in normal healthy individuals or mRNA expression in skin explants compared with medium control (skin incubated in medium alone). The point 0 (dotted line) defines the expression of mRNA above or below this point. (a) CXCL10 and (b) CXCL11 expression in acute GVHD clinical biopsies was significantly increased in comparison with normal skin (control) (**P=0.001, ***P=0.0003). (c) CXCL10 and (d) CXCL11 expression in GVHR grade III skin explants was significantly increased when compared with medium control (**P=0.02, **P=0.02) (Mann–Whitney U test). (e) Quantitative RT-qPCR analysis of BAFF mRNA expression in aGVHD clinical skin biopsies. Results are given relative to GAPDH and BAFF expression in pre-HSCT biopsies. BAFF expression was significantly higher in grades II-III acute GVHD patient biopsies in comparison with grades 0–I aGVHD. Mann–Whitney U test was used to determine significance.

Full figure and legend (98K)

Top

Discussion

Elevated BAFF levels were previously reported in plasma of cGVHD patients.16, 22 Our results show that sBAFF levels were already elevated pre-transplant in patients subsequently developing aGVHD indicating graft outcome may be influenced by transplant factors such as BAFF. Decreased sBAFF levels at day 0 suggested an effect of the conditioning treatment on BAFF producing cells. Following conditioning, a defect in cell-mediated immunity occurs resulting in disruption of B-cell homeostasis causing slow B-cell reconstitution and impaired functionality.40 The sustained increase in BAFF levels observed in aGVHD patients following HSCT suggested delayed B-cell recovery, which may have contributed to loss of B-cell tolerance and development of cGVHD.41 Elevated BAFF mRNA and protein expression confirmed its association with increased risk of developing aGVHD. Moreover, increased sBAFF levels in cGVHD patients provided strong support for BAFF involvement in cGVHD with pre-transplant levels predicting aGVHD outcome. Therefore, BAFF can be important for diagnosis or monitoring GVHD in response to acute and cGVHD therapy. Similarly, sIL-33 data suggest an important role for IL-33 in development of aGVHD as recently reported.42

Chemokines have been implicated in allogeneic transplant rejection.43, 44 The CXCR3 receptor and its ligands CXCL9/10/11 constitute an important pathway for effector cell recruitment post-transplant.18, 45 Increased CXCL10 levels in sera and aGVHD skin biopsies17, 46 have been reported but no involvement in cGVHD. The increase in CXCL10/CXCL11 levels in acute and cGVHD patient sera suggests a potential biomarker role for both chemokines and suggests the perseverance of elevated CXCL10/CXCL11 levels during aGVHD may lead to cGVHD development. Both CXCL10/CXCL11 were strongly influenced by conditioning treatment indicating conditioning induces early chemokine expression promoting T-cell migration into GVHD target organs and GVHD. Activation of chemokine expression using skin explant assays has recently been shown to induce T-cell migration and interferon-γ secretion exacerbating the GVHR.47

Interestingly, the biomarkers investigated work collectively downstream of each another in inflammatory response initiation. Inflammasome assembly at tissue damage sites results in caspase-1 release48 and subsequent IL-33 activation49 and binding to its receptor ST2 (suppressor of tumorigenicity-2). This elicits danger signals recognised by immune cells, which initiate an inflammatory response and express the CXCR3 receptor on the cell surface, resulting in CXCL10/CXCL11 binding and infiltration of effector T cells. ST2 itself has recently implicated as a biomarker for aGVHD.50

In conclusion, the training and validation cohort results highlight association of the investigated biomarkers to GVHD. Independent evaluation of the biomarkers in clinical biopsies and the skin explants strongly confirms association of each biomarker in predictive or diagnostic roles in aGVHD or cGVHD. Our results show a combination of different markers can be used to predict both acute and cGVHD; however, the use of a biomarker panel may be more effective to predict GVHD occurrence.11

Top

Conflict of interest

The authors declare no conflict of interest.

Top

References

  1. Flowers ME, Inamoto Y, Carpenter PA, Lee SJ, Kiem HP, Petersdorf EW et al. Comparative analysis of risk factors for acute graft-versus-host disease and for chronic graft-versus-host disease according to National Institutes of Health consensus criteria. Blood 2011; 117: 3214–3219. | Article | PubMed | ISI | CAS |
  2. Reddy P, Teshima T, Kukuruga M, Ordemann R, Liu C, Lowler K et al. Interleukin-18 regulates acute graft-versus-host disease by enhancing Fas-mediated donor T cell apoptosis. J Exp Med 2001; 194: 1433–1440. | Article | PubMed | ISI | CAS |
  3. Zhang C, Todorov I, Zhang Z, Liu Y, Kandeel F, Forman S et al. Donor CD4+ T and B cells in transplants induce chronic graft-versus-host disease with autoimmune manifestations. Blood 2006; 107: 2993–3001. | Article | PubMed | ISI | CAS |
  4. Ferrara JL, Deeg HJ. Graft-versus-host disease. N Engl J Med 1991; 324: 667–674. | Article | PubMed | ISI | CAS |
  5. Miura Y, Thoburn CJ, Bright EC, Chen W, Nakao S, Hess AD. Cytokine and chemokine profiles in autologous graft-versus-host disease (GVHD): interleukin 10 and interferon gamma may be critical mediators for the development of autologous GVHD. Blood 2002; 100: 2650–2658. | Article | PubMed | ISI | CAS |
  6. Jaksch M, Remberger M, Mattsson J. Increased gene expression of chemokine receptors is correlated with acute graft-versus-host disease after allogeneic stem cell transplantation. Biol Blood Marrow Transplant 2005; 11: 280–287. | Article | PubMed | CAS |
  7. Paczesny S, Krijanovski OI, Braun TM, Choi SW, Clouthier SG, Kuick R et al. A biomarker panel for acute graft-versus-host disease. Blood 2009; 113: 273–278. | Article | PubMed | ISI | CAS |
  8. Miyamoto T, Akashi K, Hayashi S, Gondo H, Murakawa M, Tanimoto K et al. Serum concentration of the soluble interleukin-2 receptor for monitoring acute graft-versus-host disease. Bone Marrow Transplant 1996; 17: 185–190. | PubMed | ISI | CAS |
  9. Or R, Kalinkovich A, Nagler A, Weisman Z, Naparstek E, Weiss L et al. Soluble tumor necrosis factor (sTNF) receptors: a possible prognostic marker for bone marrow transplantation-related complications. Cytokines Mol Ther 1996; 2: 243–250. | PubMed | ISI | CAS |
  10. Imamura M, Hashino S, Kobayashi H, Kubayashi S, Hirano S, Minagawa T et al. Serum cytokine levels in bone marrow transplantation: synergistic interaction of interleukin-6, interferon-gamma, and tumor necrosis factor-alpha in graft-versus-host disease. Bone Marrow Transplant 1994; 13: 745–751. | PubMed | ISI | CAS |
  11. Levine JE, Logan BR, Wu J, Alousi AM, Bolanos-Meade J, Ferrara JL et al. Acute graft-versus-host disease biomarkers measured during therapy can predict treatment outcomes: a Blood and Marrow Transplant Clinical Trials Network study. Blood 2012; 119: 3854–3860. | Article | PubMed | ISI | CAS |
  12. Levine JE, Braun TM, Harris AC, Holler E, Taylor A, Miller H et al. A prognostic score for acute graft-versus-host disease based on biomarkers: a multicenter study. Lancet Haematol 2015; 2: e21–e29. | Article | PubMed |
  13. MacMillan ML, Robin M, Harris AC, DeFor TE, Martin PJ, Alousi A et al. A refined risk score for acute graft-versus-host disease that predicts response to initial therapy, survival, and transplant-related mortality. Biol Blood Marrow Transplant 2015; 21: 761–767. | Article | PubMed |
  14. Harris AC, Ferrara JL, Braun TM, Holler E, Teshima T, Levine JE et al. Plasma biomarkers of lower gastrointestinal and liver acute GVHD. Blood 2012; 119: 2960–2963. | Article | PubMed | ISI |
  15. Shin OS, Harris JB. Innate immunity and transplantation tolerance: the potential role of TLRs/NLRs in GVHD. Korean J Hematol 2011; 46: 69–79. | Article | PubMed |
  16. Sarantopoulos S, Stevenson KE, Kim HT, Bhuiya NS, Cutler CS, Soiffer RJ et al. High levels of B-cell activating factor in patients with active chronic graft-versus-host disease. Clin Cancer Res 2007; 13: 6107–6114. | Article | PubMed | ISI | CAS |
  17. Piper KP, Horlock C, Curnow SJ, Arrazi J, Nicholls S, Mahendra P et al. CXCL10-CXCR3 interactions play an important role in the pathogenesis of acute graft-versus-host disease in the skin following allogeneic stem-cell transplantation. Blood 2007; 110: 3827–3832. | Article | PubMed | ISI | CAS |
  18. Croudace JE, Inman CF, Abbotts BE, Nagra S, Nunnick J, Mahendra P et al. Chemokine-mediated tissue recruitment of CXCR3+ CD4+ T cells plays a major role in the pathogenesis of chronic GVHD. Blood 2012; 120: 4246–4255. | Article | PubMed |
  19. Pavletic SZ, Carter SL, Kernan NA, Henslee-Downey J, Mendizabal AM, Papadopoulos E et al. Influence of T-cell depletion on chronic graft-versus-host disease: results of a multicenter randomized trial in unrelated marrow donor transplantation. Blood 2005; 106: 3308–3313. | Article | PubMed | CAS |
  20. Khan WN. B cell receptor and BAFF receptor signaling regulation of B cell homeostasis. J Immunol 2009; 183: 3561–3567. | Article | PubMed | ISI | CAS |
  21. Melchers F. Actions of BAFF in B cell maturation and its effects on the development of autoimmune disease. Ann Rheum Dis 2003; 62(Suppl 2): ii25–ii27. | Article | PubMed |
  22. Sarantopoulos S, Stevenson KE, Kim HT, Cutler CS, Bhuiya NS, Schowalter M et al. Altered B-cell homeostasis and excess BAFF in human chronic graft-versus-host disease. Blood 2009; 113: 3865–3874. | Article | PubMed | ISI |
  23. Nasti TH, Timares L. Inflammasome activation of IL-1 family mediators in response to cutaneous photodamage. Photochem Photobiol 2012; 88: 1111–1125. | Article | PubMed |
  24. Landfried K, Bataille F, Rogler G, Brenmoehl J, Kosovac K, Wolff D et al. Recipient NOD2/CARD15 status affects cellular infiltrates in human intestinal graft-versus-host disease. Clin Exp Immunol 2009; 159: 87–92. | Article | PubMed |
  25. Watson MJ, Ke B, Shen XD, Gao F, Busuttil RW, Kupiec-Weglinski JW et al. Intestinal ischemia/reperfusion injury triggers activation of innate toll-like receptor 4 and adaptive chemokine programs. Transplant Proc 2008; 40: 3339–3341. | Article | PubMed |
  26. Zeiser R, Penack O, Holler E, Idzko M. Danger signals activating innate immunity in graft-versus-host disease. J Mol Med (Berl) 2011; 89: 833–845. | Article | PubMed |
  27. Schroder K, Tschopp J. The inflammasomes. Cell 2010; 140: 821–832. | Article | PubMed | ISI | CAS |
  28. Zlotnik A, Yoshie O. Chemokines: a new classification system and their role in immunity. Immunity 2000; 12: 121–127. | Article | PubMed | ISI | CAS |
  29. Dickinson AM, Sviland L, Carey P, Reid MM, Hamilton PJ, Pearson AJ et al. Skin explant culture as a model for cutaneous graft-versus-host disease in humans. Bone Marrow Transplant 1988; 3: 323–329. | PubMed | ISI | CAS |
  30. Vogelsang GB, Hess AD, Berkman AW, Tutschka PJ, Farmer ER, Converse PJ et al. An in vitro predictive test for graft versus host disease in patients with genotypic HLA-identical bone marrow transplants. N Engl J Med 1985; 313: 645–650. | Article | PubMed | ISI | CAS |
  31. Wang XN, Collin M, Sviland L, Marshall S, Jackson G, Schulz U et al. Skin explant model of human graft-versus-host disease: prediction of clinical outcome and correlation with biological risk factors. Biol Blood Marrow Transplant 2006; 12: 152–159. | Article | PubMed |
  32. Dignan FL, Clark A, Amrolia P, Cornish J, Jackson G, Mahendra P et al. Diagnosis and management of acute graft-versus-host disease. Br J Haematol 2012; 158: 30–45. | Article | PubMed |
  33. Filipovich AH, Weisdorf D, Pavletic S, Socie G, Wingard JR, Lee SJ et al. National Institutes of Health consensus development project on criteria for clinical trials in chronic graft-versus-host disease: I. Diagnosis and staging working group report. Biol Blood Marrow Transplant 2005; 11: 945–956. | Article | PubMed | ISI |
  34. Schmittgen TD, Livak KJ. Analyzing real-time PCR data by the comparative C(T) method. Nat Protoc 2008; 3: 1101–1108. | Article | PubMed | ISI | CAS |
  35. Atkinson K, Horowitz MM, Gale RP, van Bekkum DW, Gluckman E, Good RA et al. Risk factors for chronic graft-versus-host disease after HLA-identical sibling bone marrow transplantation. Blood 1990; 75: 2459–2464. | PubMed | ISI | CAS |
  36. Arora M, Klein JP, Weisdorf DJ, Hassebroek A, Flowers ME, Cutler CS et al. Chronic GVHD risk score: a Center for International Blood and Marrow Transplant Research analysis. Blood 2011; 117: 6714–6720. | Article | PubMed | ISI |
  37. Stohl W, Metyas S, Tan SM, Cheema GS, Oamar B, Xu D et al. B lymphocyte stimulator overexpression in patients with systemic lupus erythematosus: longitudinal observations. Arthritis Rheum 2003; 48: 3475–3486. | Article | PubMed | ISI |
  38. Quartuccio L, Fabris M, Moretti M, Barone F, Bombardieri M, Rupolo M et al. Resistance to rituximab therapy and local BAFF overexpression in Sjogren's syndrome-related myoepithelial sialadenitis and low-grade parotid B-cell lymphoma. Open Rheumatol J 2008; 2: 38–43. | Article | PubMed |
  39. Mapara MY, Leng C, Kim YM, Bronson R, Lokshin A, Luster A et al. Expression of chemokines in GVHD target organs is influenced by conditioning and genetic factors and amplified by GVHR. Biol Blood Marrow Transplant 2006; 12: 623–634. | Article | PubMed | ISI | CAS |
  40. Storek J, Ferrara S, Ku N, Giorgi JV, Champlin RE, Saxon A. B cell reconstitution after human bone marrow transplantation: recapitulation of ontogeny? Bone Marrow Transplant 1993; 12: 387–398. | PubMed | CAS |
  41. Brink R. Regulation of B cell self-tolerance by BAFF. Semin Immunol 2006; 18: 276–283. | Article | PubMed | CAS |
  42. Reichenbach DK, Schwarze V, Matta BM, Tkachev V, Lieberknecht E, Liu Q et al. The IL-33/ST2 axis augments effector T cell responses during acute GVHD. Blood 2015; 125: 3183–3192. | Article | PubMed |
  43. Hancock WW, Lu B, Gao W, Csizmadia V, Faia K, King JA et al. Requirement of the chemokine receptor CXCR3 for acute allograft rejection. J Exp Med 2000; 192: 1515–1520. | Article | PubMed | ISI | CAS |
  44. Hancock WW, Gao W, Faia KL, Csizmadia V. Chemokines and their receptors in allograft rejection. Curr Opin Immunol 2000; 12: 511–516. | Article | PubMed | ISI | CAS |
  45. Uppaluri R, Sheehan KC, Wang L, Bui JD, Brotman JJ, Lu B et al. Prolongation of cardiac and islet allograft survival by a blocking hamster anti-mouse CXCR3 monoclonal antibody. Transplantation 2008; 86: 137–147. | Article | PubMed | ISI | CAS |
  46. Hancock WW, Gao W, Csizmadia V, Faia KL, Shemmeri N, Luster AD. Donor-derived IP-10 initiates development of acute allograft rejection. J Exp Med 2001; 193: 975–980. | Article | PubMed | ISI | CAS |
  47. Mavin E, Ahmed SS, O'Boyle G, Turner B, Douglass S, Norden J et al. Regulatory T cells inhibit CD8(+) T-cell tissue invasion in human skin graft-versus-host reactions. Transplantation 2012; 94: 456–464. | Article | PubMed |
  48. Grenier JM, Wang L, Manji GA, Huang WJ, Al-Garawi A, Kelly R et al. Functional screening of five PYPAF family members identifies PYPAF5 as a novel regulator of NF-kappaB and caspase-1. FEBS Lett 2002; 530: 73–78. | Article | PubMed | ISI | CAS |
  49. Cayrol C, Girard JP. The IL-1-like cytokine IL-33 is inactivated after maturation by caspase-1. Proc Natl Acad Sci USA 2009; 106: 9021–9026. | Article | PubMed |
  50. Ponce DM, Hilden P, Mumaw C, Devlin SM, Lubin M, Giralt S et al. High day 28 ST2 levels predict for acute graft-versus-host disease and transplant-related mortality after cord blood transplantation. Blood 2014; 125: 199–205. | Article | PubMed | ISI |
Top

Acknowledgements

We thank Elizabeth Douglas, Cindy Carr for technical support, Shelagh Lowerson for providing Clinical Data from Euro Bank and help with ROC analysis and Shazmeen Surtee for helping as part of her undergraduate project. This study was funded by FP6 EC Stemdiagnostics (Contract No. 037703), FP7 Marie Curie Initial Training Network Celleurope (Contract No. 315963), Leukemia and Lymphoma Research and Tyneside Leukaemia Research Association.

Author contributions

SSA, XNW, JN and EE-G performed experiments. SSA, KP and SA carried out the statistical analysis. SSA and AMD wrote the paper. EH, MC and IH contributed to collaborating on the research and contributed patient samples and clinical data. AMD and EH designed the research.

Supplementary Information accompanies this paper on Bone Marrow Transplantation website