Elevated Bone Marrow Eosinophil Count Is Associated with High Incidence of Severe Acute Gvhd After Allogeneic Hematopoietic Stem Cell Transplantation

Home , Hematopoietic stem cell

ORIGINAL ARTICLE Elevated bone marrow eosinophil count is associated with high incidence of severe acute GvHD after allogeneic hematopoietic stem cell transplantation

Y Shimomura1, M Hara2, H Hashimoto3 and T Ishikawa1

Predicting severe acute GvHD (aGvHD) after allogeneic hematopoietic stem cell transplantation (HSCT) is challenging but critical. Mild aGvHD may have a favorable impact on relapse, whereas severe aGvHD is associated with poor outcomes. The aim of this study was to evaluate whether elevated eosinophil count in the bone marrow (BM) at 1 month after HSCT is associated with a high incidence of new and severe aGvHD. We enrolled 101 consecutive patients; median age was 50 years, and 50.5% patients were male. The median eosinophil concentration in BM at 1 month after HSCT was 1.1% (quartile 0.4–2.2%). The adjusted hazards ratio at 95% conﬁdence interval for severe aGvHD was 1.26 (1.12–1.42, Po0.001), per 1% increase in eosinophil concentration, and 3.76 (1.41–10.05, P = 0.008) for the high-risk group at a cutoff value of 4.0%. In addition, the predictive accuracy described by area under the curve of receiver operating characteristics, increased from 0.784 to 0.866 (P = 0.033) with the increasing concentration of eosinophils. In conclusion, elevated concentration of eosinophils in BM was associated with high incidence and predictive accuracy of severe aGvHD. BM eosinophil concentration can be one of the key markers to predict aGvHD.

Bone Marrow Transplantation (2017) 52, 1311–1316; doi:10.1038/bmt.2017.98; published online 19 June 2017

INTRODUCTION association between elevated eosinophils in BM and grades 3–4 Acute GvHD (aGvHD) is a common and significant complication aGvHD after HSCT, in their preliminary correspondence. The study of allogeneic hematopoietic stem cell transplantation (HSCT). included patients presenting with aGvHD on BM examination; It occurs when T lymphocytes of donor origin recognize the however, no multivariate analysis was performed despite the recipient cells as foreign and initiate immune reactions. The skin, presence of many confounding factors such as HLA mismatch. gastrointestinal tract and liver are principally involved, and the Hence, we evaluated whether eosinophil concentration in BM at severity of aGvHD is determined using the Glucksberg grade 1-month after HSCT is associated with subsequent development system,1,2 which calculates a score based on the degree of of severe aGvHD in patients with HSCT, who had no aGvHD at the involvement of these organs. Although it is known that severe time of BM examination. aGvHD defined as grades 2–4, is one of the leading causes of transplant-related mortality, recent evidence also suggests that PATIENTS AND METHODS aGvHD can have a twofold impact. It may have a favorable impact on relapse after HSCT, which is known as the graft-versus- Study patients leukemia effect.3–5 Hence, predicting severe aGvHD is clinically Of the 169 consecutive patients who underwent allogeneic HSCT at Kobe essential. The aGvHD prophylaxis regimen can be tailored for City Hospital Organization, Kobe City Medical Center General Hospital, or improved outcomes, based on the predicted severity of aGvHD.6–8 Foundation for Biomedical Research and Innovation between October Prediction of severe aGvHD using serum biomarkers is under 2010 and December 2015, 101 patients who underwent microscopic examination of BM at 1-month after HSCT, and were free of aGvHD, were rigorous investigation by researchers. Suppression of tumor- enrolled in the present study (Figure 1). We excluded 14 patients with no igenicity 2 (ST2) cell counts using flow cytometry or microRNA engraftment, 5 patients without microscopic examination of BM at are reported to be significant parameters for the prediction of 1-month and 49 patients who developed aGvHD before microscopic – severe aGvHD.9 11 examinations of BM, to avoid the impact of aGvHD and its treatment on On the other hand, several studies evaluating histological the predictability. The study protocol complied with the Helsinki findings of aGvHD, demonstrated infiltration of eosinophils in the Declaration standards, and was approved by the Ethics Committee of involved organs, indicating an association between eosinophils our institutions (Approval No. zn160706 at Kobe City Hospital Organization, and aGvHD.12–18 Eosinophil infiltration in the target tissue after Kobe City Medical Center General Hospital, and No. 16-08 at Foundation for Biomedical Research and Innovation). Written informed consent was HSCT is acquired from donor cells, and their production in the waivered, as the study used retrospective data obtained from hospital bone marrow (BM) is stimulated by several physiologically active records, and there were no patient interventions during the study. substances including interleukin-5, which are produced due to the The authors had complete access to the data and take responsibility for its allo-reactive immune response. Basara et al.19 reported the integrity. All authors have read, and agree to the manuscript as written.

1Department of Hematology, Kobe City Hospital Organization, Kobe City Medical Center General Hospital, Kobe, Japan; 2Department of Cardiovascular Medicine, Osaka City University Graduate School of Medicine, Osaka, Japan and 3Department of Cell Therapy, Foundation for Biomedical Research and Innovation, Kobe, Japan. Correspondence: Dr Y Shimomura, Department of Hematology, Kobe City Hospital Organization, Kobe City Medical Center General Hospital, Minamimati 2-1-1, Minatojima, Chuo-Ku, Kobe 650-0047, Japan. E-mail: [email protected] Received 16 February 2017; revised 15 April 2017; accepted 19 April 2017; published online 19 June 2017 Eosinophil and acute GvHD Y Shimomura et al 1312 Our basic protocol of HSCT conditioning was the others.20–22 All patients received a calcineurin All patients were provided with high-efficiency particulate air-filtered inhibitor-based aGvHD prophylaxis. A combined regimen of cyclosporine A rooms. Conditioning regimen varied in intensity according to the and short-term methotrexate was administered to patients who received patient’s age, disease severity and comorbidities, at the attending BM stem cell transplantation from a related donor. A regimen of tacrolimus physician’s discretion. Reduced intensity conditioning regimen was (Tac) and mycophenolate or Tac and short-term methotrexate was used for older patients (460 years) and/or patients with serious administered to patients who received cord blood stem cell transplantation, and a regimen of Tac and short-term methotrexate was prescribed complications. Myeloablative conditioning regimen included TBI ⩾ 8 Gy, for other patients. Post-transplant cyclophosphamide was administrated busulfan ⩾ 7.2 mg/kg or melphalan 4140 mg/m2 and reduced intensity to patients who received HLA-haploidentical stem cell transplantation. Anti-thymoglobulin was administered to patients with aplastic anemia and to some patients who received HLA-mismatch stem cell transplantation. Allogeneic HSCT during the study period Conditioning regimens and aGvHD prophylaxis tended to follow the 169 patients principal, but was finally determined by the respective attending physician. BM stem cell units and peripheral blood (PB) stem cell units were sourced Excluded 14 patients without engraftment from the Japan Marrow Donor Program or related donors, and cord blood stem cell units were sourced from the Japanese Cord Blood Bank Network. Successful engraftment 155 patients Microscopic examinations of BM Excluded 5 patients without microscopic examinations of BM at 1-month after HSCT Microscopic examinations of BM were performed to determine the concentration of eosinophils, and PB counts were performed for use as Eosinophil data of BM at 1-month was available reference. PB smear was obtained on the same day as examination of BM 150 patients (±1day). BM aspirate and PB smears were stained with May–Grünwald Giemsa stain and were each reviewed by a faculty hematologist or medical Excluded 49 patients who developed aGVHD before fi microscopic examinations of BM at 1-month after HSCT technologist, at a magni cation of × 60. A differential count of 500 cells was carried out for each BM smear and eosinophil concentrations were Final study population recorded, and a differential count of 200 cells was done for cases with low 101 patients cellularity. A 100 cell differential count was also carried out for each PB smear, and eosinophil concentrations were recorded. To minimize intra- Figure 1. Patient selection flow. and inter-observer variations, 30 randomly selected cases were reviewed

Table 1. Patient characteristics

Parameters Total (n = 101) Low risk (n = 87) High risk (n = 14) P-value

Eosinophil concentration in BM, % 1.1 (0.4–2.2) 0.8 (0.4–1.6) 6.1 (4.7–8.6) o0.001 Eosinophil concentration in PB, % 0.5 (0.0–2.0) 0.0 (0.0–1.0) 4.0 (1.0–9.0) o0.001 Age, years old 50 (39–59) 50 (39–60) 51 (45–58) 0.633 Male 51 (50.5) 45 (51.7) 6 (42.9) 0.743 Diagnosis 0.566 AML 33 (32.7) 29 (33.3) 4 (28.6) ALL 23 (22.8) 19 (21.8) 4 (28.6) Myelodysplastic syndrome 12 (11.9) 9 (10.3) 3 (21.4) Malignant lymphoma 19 (18.8) 16 (18.4) 3 (21.4) Adult T-cell leukemia 8 (7.9) 8 (9.2) 0 (0) Aplastic anemia 6 (5.9) 6 (6.9) 0 (0) Disease states 0.72 Complete and PR 65 (64.4) 55 (63.2) 10 (71.4) Chemo naïve 22 (21.8) 19 (21.8) 3 (21.4) Active disease 14 (13.9) 13 (14.9) 1 (7.1) MAC 60 (59.4) 51 (58.6) 9 (64.3) 0.914 Stem cell source 0.031 BM 52 (51.5) 47 (54.0) 5 (35.7) Peripheral blood 18 (17.8) 12 (13.8) 6 (42.9) Cord blood 31 (30.7) 28 (32.2) 3 (21.4) Unrelated donor 71 (70.3) 65 (74.7) 6 (42.9) 0.035 HLA mismatch, 48 (47.5) 44 (50.6) 4 (28.6) 0.214 GvHD prophylaxis 0.876 Cyclosporine A/short-term methotrexate 14 (13.9) 12 (13.8) 2 (14.3) Tacrolimus/short-term methotrexate 70 (69.3) 61 (70.1) 9 (64.3) Tacrolimus/mycophenolate 17 (16.8) 14 (16.1) 3 (21.4) Additional GvHD prophylaxis 0.834 None 93 (92.1) 80 (92.0) 13 (92.9) Post-transplant cyclophosphamide 2 (2.0) 2 (2.3) 0 (0.0) Antithymocyte globulin 6 (5.9) 5 (5.7) 1 (7.1) Corticosteroid use at the moment of BM examination 7 (6.9) 7 (8.0) 0 (0.0) 0.589 Corticosteroid doses in patients who received corticosteroid 20 (15–27) 20 (15–27) NA NA Days from HSCT to BM examination 28 (27–28) 28 (27–28) 28 (28–29) 0.536 Abbreviations: BM = bone marrow; HSCT = hematopoietic stem cell transplantation; MAC = myeloablative conditioning regimen; NA = not applicable; PB = peripheral blood. Low risk group was deﬁned as patients with eosinophil concentration in BM ⩽ 4%, and high-risk group had concentration 44%. Continuous variables were summarized as medians and interquartile range (quartiles 1–3), and categorical variables were summarized as counts and percentages. Corticosteroid doses are expressed as the converted amount of methylprednisolone as calculated by multiplying the dose of prednisolone by 0.8.

Bone Marrow Transplantation (2017) 1311 – 1316 © 2017 Macmillan Publishers Limited, part of Springer Nature. Eosinophil and acute GvHD Y Shimomura et al 1313 by 2 independent hematologists or medical technologists, who were variables between the 2 groups. Cumulative incidence of severe aGvHD blinded to patient information. was also estimated using the competing risk model with 95% confidence interval (CI). The prediction accuracy of eosinophil count for the Statistical analysis development of severe aGvHD was assessed comparing the area under the curve of receiver operating characteristics (ROC) using the logistic fi – The primary endpoint was severe aGvHD, de ned as grades 2 4 aGvHD regression model. We first created a standard model for prediction of within 70 days of the microscopic examinations of BM (100 days after the severe aGvHD based on patient background and previously reported risk – fi HSCT). Grades 2 4 aGvHD was de ned as per the Glucksberg grade factors for aGvHD. The standard model was adjusted for all clinical 1,2 fi system. Competing risks for severe aGvHD were de ned as relapse and parameters other than eosinophil concentration, as seen in Table 1. We death from all cause. All data were retrospectively obtained from patient then evaluated whether the incorporation of BM eosinophil concentration records. Continuous variables were summarized as medians and inter- – in the standard model would improve the predictive accuracy of severe quartile ranges (quartiles 1 3), and categorical variables were summarized aGvHD. Statistical significance was set as Po0.05. All statistical analyses as counts and percentages. Corticosteroid doses were summarized only for were performed using R software packages (version 3.1.1; R Development patients who received corticosteroids at the time of BM examination; they Core Team). are expressed as the converted amount of methylprednisolone as calculated by multiplying the dose of prednisolone by 0.8. Intra- and inter-observer variations in eosinophil counts of BM were assessed using RESULTS intraclass correlation coefficient on a randomly selected data of 30 cases. To evaluate the impact of BM and PB eosinophil concentration on the Patients characteristics primary endpoint, we employed multivariable Fine and Gray competing The characteristics of 101 patients enrolled and analyzed in the risks regression model, using eosinophil concentration as a continuous present study (Figure 1) are described in Table 1. There were no variable. Adjusted covariates were as follows; age, sex, donor source, missing data in this study. Median age was 50 years, and 51 unrelated donor, HLA mismatch and corticosteroid doses at the time of BM examination. To arrive at a simple risk stratification, we divided the (50.5%) patients were male. 52 (51.5%) patients received BM patients into two risk groups (high versus low risk groups) using the transplantation, 18 (17.8%) received PB transplantation and 31 survival classification and regression tree analysis. The classification and (30.7%) received cord blood transplantation. In the study regression tree method is an empirical, statistical technique based on population, 71 (70.3%) patients received the transplant from recursive partitioning analysis and is useful to arrive at clinical decisions unrelated donors and 48 (47.5%) patients received it from an 23 and risk stratification models. Both patient risk groups were treated as a HLA-mismatched donor. At the time of BM examination, 7 patients categorical variable, and the impact on primary endpoint was evaluated received corticosteroid therapy at a dose of 20 mg/day.15–27 using the multivariable Fine and Gray competing risks regression The median eosinophil concentration in BM was 1.1% (quartile model. Patients and disease characteristics were compared using the – – Mann–Whitney U test for continuous variables and χ2 test for categorical 0.4 2.2%) and that in PB was 0.5% (0.0 2.0%). The intraclass correlation coefficient of eosinophil concentration in BM was 0.908 (95% CI 0.818–0.955, Po0.001) and 0.911 (95% CI 0.816–0.957, Table 2. Total incidence of patient outcomes at 70 days after Po0.001) for intra- and inter-observer variations, respectively. The microscopic examination of BM survival classification and regression tree analysis revealed that the cutoff eosinophil concentration to predict the incidence of Parameters Total Low High P-value (n = 101) (n = 87) (n = 14) severe aGvHD was 4.0% in BM and no cutoff point could be determined in PB. Thus, we defined patients with eosinophil aGvHD 0.034 concentration in BM ⩽ 4.0% as low risk group, and patients Grade 1 27 (26.7) 22 (25.3) 5 (35.7) with eosinophil concentration in BM 44.0% as high-risk group. Grade 2 8 (7.9) 6 (6.9) 2 (14.3) There were no significant differences in patient characteristics Grade 3 7 (6.9) 6 (6.9) 1 (7.1) between the 2 groups, except for the source of stem cells and the Grade 4 3 (3.0) 1 (1.1) 2 (14.3) percentage of unrelated donors (Table 1). Relapse 12 (11.9) 8 (9.2) 4 (28.6) 0.102 Treatment related 5 (5.0) 5 (5.7) 0 (0.0) 0.798 death Development of aGvHD – Abbreviations: aGvHD = acute graft versus host disease; BM = bone There were 18 cases of grades 2 4 severe aGvHD, 12 cases of marrow. relapse and 5 treatment related deaths in this study (Table 2). The cumulative incidence of severe aGvHD at 70 days after BM

a Total population b Subgroup analysis 1.0 1.0

0.8 0.8 Adjusted P =0.008 0.6 0.6

High risk group 0.4 0.4 Cumulative incidence Cumulative incidence Low risk group of grade 2 to 4 aGVHD 0.2 of grade 2 to 4 aGVHD 0.2

0.0 0.0 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 Days after BM examination Days after BM examination Figure 2. Cumulative incidence of grades 2–4 aGvHD after microscopic examination of BM. Cumulative incidence of aGvHD in the total population (a) and in each risk group (b) with a competing risk model. Low risk group was deﬁned as patients with eosinophil concentration in BM ⩽ 4%, and high-risk group with concentration 44%.

© 2017 Macmillan Publishers Limited, part of Springer Nature. Bone Marrow Transplantation (2017) 1311 – 1316 Eosinophil and acute GvHD Y Shimomura et al 1314 examination was 17.8% (95% CI 11.1–25.9%) in the entire cohort (HR) for eosinophil concentration was 1.26 (95% CI 1.12–1.42, (Figure 2a), 35.7% (12.2–60.5%) in the high-risk group and 14.9% Po0.001) per 1% increase in eosinophil concentration, and 3.76 (8.4–23.3%) in the low risk group (adjusted P = 0.008; Figure 2b). (1.41–10.05, P = 0.008) for the high-risk group compared to low The correlation between eosinophil concentration in BM and the risk group. On the other hand, no significant correlation between primary endpoint is shown in Table 3. The adjusted hazard ratio eosinophil concentration in PB and the primary endpoint was observed with adjusted HR of 0.98 (95% CI 0.87–1.10, P = 0.750). In addition, ROC analysis demonstrated that inclusion of eosinophil concentration in BM in the prediction model, Table 3. Impact of eosinophil concentration in BM on grades 2–4 significantly improved the predictive accuracy of severe aGvHD acute GvHD from area under the curve of 0.784 (95% CI 0.668–0.901) with the – Adjusted HR 95% CI P-value standard model, to 0.866 (95% CI 0.779 0.954) with eosinophil concentration incorporated model (Delong P = 0.033; Figure 3). High (44%) vs Low (⩽4%) 3.76 1.41–10.05 0.008 A representative case is shown in Figure 4. Eosinophil in BM per 1% 1.26 1.12–1.42 o0.001 Abbreviations: aGvHD = acute graft versus host disease; BM = bone DISCUSSION marrow; CI = confidence interval; HR = hazard ratio. Multivariate analysis was adjusted for age, sex, donor source, unrelated donor, HLA mismatch In the present study, we demonstrated that elevated concentra- and corticosteroid dose at the time of BM examination. tion of eosinophils in the BM is associated with high incidence of severe aGvHD after HSCT and determining eosinophil concentration during BM microscopic examination significantly improved the predictive accuracy of future severe aGvHD. 1.0 Delong P =0.033 Since little evidence is available on the association between eosinophil concentration in BM and incidence of aGvHD, our 0.8 results provide physicians with additional information on the subject. We hope eosinophil counts in BM after HSCT would be 0.6 useful for predicting severe aGvHD, and for designing risk-adapted tailored immunosuppression intensity in the future. The possible association between eosinophils and aGvHD has 0.4 Sensitivity been previously reported. Infiltration of eosinophils in the skin, AUC 0.784 gastrointestinal tract and liver in patients with aGvHD is often (95% CI 0.668-0.901) 12–14,18 19 0.2 without eosinophil count observed and may be harmful. Basara et al. reported that AUC 0.866 an elevated concentration of eosinophils in BM was associated (95% CI 0.779-0.954) – fi 0.0 without eosinophil count with grades 3 4 aGvHD after HSCT in their preliminary ndings. On the other hand, there are some reports about an association 0.0 0.2 0.4 0.6 0.8 1.0 between elevated concentration of eosinophils in PB and aGvHD; 1- Specificity however, the association is still controversial, possibly due to the 15–17,24,25 Figure 3. ROC curve for predicting severe aGvHD. Black line influence of corticosteroid therapy. As mentioned in indicates ROC curve with standard model, and blue line indicates previous report from Basara et al., our results also showed that ROC curve with eosinophil concentration incorporated model. increased concentration of eosinophils in BM was associated with

x 60 x 60 Figure 4. Representative case presentation. A 36-year-old female underwent allogenic HSCT due to ALL. May–Grünwald Giemsa staining with air-dry and methyl alcohol fixation of BM smear at 28 days after the HSCT demonstrated an increase in eosinophil concentration to 4.6% (a, arrow). The patient developed class 3 aGvHD at 43 days after HSCT with erythroderma 450% of the whole body. She had no elevation of total bilirubin or gastrointestinal symptoms. May–Grünwald Giemsa staining with formalin fixation of skin revealed eosinophil infiltration in the dermis (b, arrows). Magnification is shown at the right bottom corner.

Bone Marrow Transplantation (2017) 1311 – 1316 © 2017 Macmillan Publishers Limited, part of Springer Nature. Eosinophil and acute GvHD Y Shimomura et al 1315 development of severe aGvHD in the future, even after adjust- Finally, the optimal time to perform microscopic examinations of ments of covariates including corticosteroid doses; however, no BM is yet to be determined. Further studies are necessary to find association was found between eosinophil concentration in PB these answers. and aGvHD. The results suggest that an elevated concentration of In conclusion, an elevated concentration of eosinophils in BM eosinophils in BM might represent an initial immune reaction was associated with high incidence of severe aGvHD and leading to severe aGvHD. As the corticosteroid is used for improved the predictive accuracy of severe aGvHD after HSCT. treatment or prophylaxis of aGvHD, the impact of steroid therapy Eosinophil counts in BM are an important marker to predict on eosinophil count in the BM is another area of interest. aGvHD, and the optimal timing of microscopic examinations of The role of eosinophils in aGvHD is still unclear. The release of BM and phenotype of eosinophils are subjects of further study. eosinophil peroxidase into the extracellular space has cytopatho- genic effects on the intestinal epithelium in celiac and Crohn’s disease.26 In organ transplant, the correlation between infiltration CONFLICT OF INTEREST and activation of eosinophils and allograft rejection has been The authors declare no conflict of interest. previously reported.27,28 Similar outcomes were observed for aGvHD.13 In addition, some reports revealed the correlation between aGvHD and interleukin-5, which suggests a correlation between ACKNOWLEDGEMENTS eosinophils and aGvHD, because interleukin-5 stimulates production We thank the medical and nursing staff working at our institution, and would like to – of eosinophils in BM and inhibits apoptosis.29 31 Eosinophils may express our gratitude to the Japan Society of Clinical Research (http://www.japanscr. proliferate on activation by interleukin-5 and other cytokines org/) for their dedicated support. produced by T cells, following which they infiltrate into the target tissue and cause injury to these tissues by releasing peroxidase. REFERENCES Prediction of severe aGvHD is challenging and critical after HSCT. One of the most important markers for the prediction of 1 Thomas ED, Storb R, Clift RA, Fefer A, Johnson L, Neiman PE et al. Bone-marrow 292 – aGvHD is ST2, which acts as a decoy receptor for interleukin-33, transplantation (second of two parts). N Engl J Med 1975; :895 902. limiting its access to type 2T helper cell and promoting the type 1T 2 Rowlings PA, Przepiorka D, Klein JP, Gale RP, Passweg JR, Henslee-Downey PJ et al. 9 IBMTR severity Index for grading acute graft-versus-host disease: retrospective helper cell phenotype. Ponce et al. reported that the ST2 levels at comparison with Glucksberg grade. Br J Haematol 1997; 97:855–864. 28 days after cord blood stem cell transplantation appeared to 3 Bacigalupo A, Van Lint MT, Occhini D, Gualandi F, Lamparelli T, Sogno G et al. predict the development of grades 3–4 aGvHD. The adjusted Increased risk of leukemia relapse with high-dose cyclosporine A after allogeneic HR (95% CI) of ST2 high group compared to low group was 2.62 marrow transplantation for acute leukemia. Blood 1991; 77:1423–1428. (1.06–6.45, P = 0.036). Another study observed that the 4 plasma 4 Locatelli F, Zecca M, Rondelli R, Bonetti F, Dini G, Prete A et al. Graft versus host microRNA biomarker panel, which includes inflammation, tissue disease prophylaxis with low-dose cyclosporine-A reduces the risk of relapse damage, regulation of cell proliferation and in tissue repair, could in children with acute leukemia given HLA-identical sibling bone marrow predict further progress of grades 2–4 aGvHD.11 The area under transplantation: results of a randomized trial. Blood 2000; 95: 1572–1579. the curve (95% CI) for the combined microRNA biomarker panel 5 Baron F, Labopin M, Niederwieser D, Vigouroux S, Cornelissen JJ, Malm C et al. Impact of graft-versus-host disease after reduced-intensity conditioning was 0.85 (0.76–0.93, Po0.001).11 In the present study, we showed allogeneic stem cell transplantation for acute myeloid leukemia: a report from the that elevated concentration of eosinophils in BM at 28 days after – Acute Leukemia Working Party of the European group for blood and marrow HSCT predicted the progression of grades 2 4 aGvHD. The transplantation. Leukemia 2012; 26: 2462–2468. evaluation of eosinophils in BM is less expensive and more 6 Malard F, Szydlo RM, Brissot E, Chevallier P, Guillaume T, Delaunay J et al. Impact accessible, compared to previously reported indices such as ST2 or of cyclosporine-A concentration on the incidence of severe acute graft-versus- microRNA. However, BM aspiration is more invasive and cannot be host disease after allogeneic stem cell transplantation. Biol Blood Marrow Trans- performed frequently. However, the evaluation of eosinophils in plant 2010; 16:28–34. BM can be performed in patients with indices predictive of severe 7 Mori T, Kato J, Shimizu T, Aisa Y, Nakazato T, Yamane A et al. Effect of early aGvHD, and may be clinically useful because of its cost and posttransplantation tacrolimus concentration on the development of acute accessibility. graft-versus-host disease after allogeneic hematopoietic stem cell transplantation from unrelated donors. Biol Blood Marrow Transplant 2012; 18:229–234. We described that an elevated concentration of eosinophils in 8 Rogosheske JR, Fargen AD, DeFor TE, Warlick E, Arora M, Blazar BR et al. Higher BM may be a marker for prediction of severe aGvHD. As the therapeutic CsA levels early post transplantation reduce risk of acute GvHD and previously reported markers such as ST2 and microRNA are not improves survival. Bone Marrow Transplant 2014; 49: 122–125. known to be associated with eosinophils, a combined use of 9 Ponce DM, Hilden P, Mumaw C, Devlin SM, Lubin M, Giralt S et al. High day 28 ST2 concentration of eosinophils in BM and other predictive markers levels predict for acute graft-versus-host disease and transplant-related mortality may have the potential to improve predictive value. The optimal after cord blood transplantation. Blood 2015; 125:199–205. timing to perform microscopic examinations of BM is yet to be 10 Chen YB, McDonough S, Hasserjian R, Chen H, Coughlin E, Illiano C et al. determined. We performed microscopic examinations of BM at Expression of CD30 in patients with acute graft-versus-host disease. Blood 2012; 120 – 28 days after allogeneic HSCT. Considering that almost one-third :691 696. of patients (49 patients) had developed aGvHD before the 11 Xiao B, Wang Y, Li W, Baker M, Guo J, Corbet K et al. Plasma microRNA signature as a noninvasive biomarker for acute graft-versus-host disease. Blood 2013; 122: microscopic examinations of BM, determination of the optimal 3365–3375. timing of the BM evaluation could be a future objective. Early 12 Nonomura A, Kono N, Mizukami Y, Nakanuma Y. Histological changes of the liver prediction of the risk of mild or severe aGvHD can lead to in experimental graft-versus-host disease across minor histocompatibility barriers. preemptive interventions such as administering higher therapeu- VIII. Role of eosinophil infiltration. Liver 1996; 16:42–47. tic calcineurin inhibitor levels to prevent severe aGvHD or tapering 13 Daneshpouy M, Socie G, Lemann M, Rivet J, Gluckman E, Janin A. Activated immunosuppressive agents rapidly to produce stronger graft eosinophils in upper gastrointestinal tract of patients with graft-versus-host versus leukemia effect, which may improve the prognosis of disease. Blood 2002; 99: 3033–3040. patients undergoing HSCT. 14 Janin A, Ertault-Daneshpouy M, Socie G. Eosinophils and severe forms of 101 There are some limitations in this study. First, this is a graft-versus-host disease. Blood 2003; : 2073. 15 Sato T, Kobayashi R, Nakajima M, Iguchi A, Ariga T. Significance of eosinophilia retrospective single institution study. Second, we used morpho- after stem cell transplantation as a possible prognostic marker for favorable logic analysis to detect the presence of eosinophils. It is still outcome. Bone Marrow Transplant 2005; 36: 985–991. unknown which phenotype of eosinophils is associated with 16 Aisa Y, Mori T, Nakazato T, Shimizu T, Yamazaki R, Ikeda Y et al. Blood eosinophilia aGvHD, although activated phenotypes of eosinophils in the as a marker of favorable outcome after allogeneic stem cell transplantation. 13 tissues has been reported to be associated with severe aGvHD. Transpl Int 2007; 20: 761–770.

© 2017 Macmillan Publishers Limited, part of Springer Nature. Bone Marrow Transplantation (2017) 1311 – 1316 Eosinophil and acute GvHD Y Shimomura et al 1316 17 Imahashi N, Miyamura K, Seto A, Watanabe K, Yanagisawa M, Nishiwaki S et al. 25 Cromvik J, Johnsson M, Vaht K, Johansson JE, Wennerås C. Eosinophils in the Eosinophilia predicts better overall survival after acute graft-versus-host-disease. blood of hematopoietic stem cell transplanted patients are activated and have Bone Marrow Transplant 2010; 45:371–377. different molecular marker profiles in acute and chronic graft-versus-host disease. 18 Weaver J, Bergfeld WF. Quantitative analysis of eosinophils in acute graft-versus- Immun Inflamm Dis 2014; 2:99–113. host disease compared with drug hypersensitivity reactions. Am J Dermatopathol 26 Hallgren R, Colombel JF, Dahl R, Fredens K, Kruse A, Jacobsen NO et al. Neutrophil 2010; 32:31–34. and eosinophil involvement of the small bowel in patients with celiac disease 19 Basara N, Kiehl MG, Fauser AA. Eosinophilia indicates the evolution to acute and Crohn's disease: studies on the secretion rate and immunohistochemical graft-versus-host disease. Blood 2002; 100: 3055. localization of granulocyte granule constituents. Am J Med 1989; 86:56–64. 20 Giralt S, Ballen K, Rizzo D, Bacigalupo A, Horowitz M, Pasquini M et al. 27 Hällgren R, Bohman SO, Fredens K. Activated eosinophil infiltration and deposits Reduced-intensity conditioning regimen workshop: defining the dose spectrum. of eosinophil cationic protein in renal allograft rejection. Nephron 1991; 59: Report of a workshop convened by the center for international blood and marrow 266–270. transplant research. Biol Blood Marrow Transplant 2009; 15: 367–369. 28 de Groen PC, Kephart GM, Gleich GJ, Ludwig J. The eosinophil as an effector cell 21 Bacigalupo A, Ballen K, Rizzo D, Giralt S, Lazarus H, Ho V et al. Defining the of the immune response during hepatic allograft rejection. Hepatology 1994; 20: intensity of conditioning regimens: working definitions. Biol Blood Marrow 654–662. Transplant 2009; 15: 1628–1633. 29 Imoto S, Oomoto Y, Murata K, Das H, Murayama T, Kajimoto K et al. Kinetics 22 Hamadani M, Mohty M, Kharfan-Dabaja MA. Reduced-intensity conditioning of serum cytokines after allogeneic bone marrow transplantation: interleukin-5 allogeneic hematopoietic cell transplantation in adults with acute myeloid as a potential marker of acute graft-versus-host disease. Int J Hematol 2000; 72: leukemia. Cancer Control 2011; 18: 237–245. 92–97. 23 Fonarow GC, Adams KF Jr, Abraham WT, Yancy CW, Boscardin WJ et al. Adhere 30 Fujii N, Hiraki A, Aoe K, Murakami T, Ikeda K, Masuda K et al. Serum cytokine Scientific Advisory Committee SG Risk stratification for in-hospital mortality in concentrations and acute graft-versus-host disease after allogeneic peripheral acutely decompensated heart failure: classification and regression tree analysis. blood stem cell transplantation: concurrent measurement of ten cytokines and JAMA 2005; 293: 572–580. their respective ratios using cytometric bead array. Int J Mol Med 2006; 17: 24 Beier F, Arbter K, Kittan NA, Andreesen R, Krause SW, Holler E et al. Regression of 881–885. eosinophil counts after diagnosis of chronic graft-versus-host disease as a 31 Dent LA, Strath M, Mellor AL, Sanderson CJ. Eosinophilia in transgenic mice potential marker for improved clinical outcome. Mol Clin Oncol 2014; 2:81–86. expressing interleukin 5. J Exp Med 1990; 172: 1425–1431.