Predictive Values of Early Suppression of Tumorigenicity 2 for Acute GVHD and Transplant-related Complications after Allogeneic Stem Cell Transplantation : Prospective Observational Study

Address for Correspondence/Yazışma Adresi: Takuya Miyazaki, M.D., Yokohama City University Graduate School of Medicine, Department of Stem Cell and Immune Regulation, Kanagawa, Japan Phone : +81 45 787 2800 E-mail : takuya_m@yokohama-cu.ac.jp ORCID: orcid.org/0000-0003-1884-4757 Received/Geliş tarihi: April 2, 2019 Accepted/Kabul tarihi: August 28, 2019 ©Copyright 2020 by Turkish Society of Hematology Turkish Journal of Hematology, Published by Galenos Publishing House DOI: 10.4274/tjh.galenos.2019.2019.0139 Turk J Hematol 2020;37:20-29

ST2 is a membrane receptor expressed on several immune cell types that belong to the IL-1 receptor family.ST2 induces T helper type 2 (Th2) and Treg immune responses, which play important roles in GVHD [17].A soluble form of ST2 (sST2) is a decoy receptor for IL-33, which blocks the IL-33/ST2 pathway and drives Th2 cells toward a Th1 cell phenotype, and it is thought to be important in the pathophysiology of GVHD [18,19].In murine models, the ST2/IL-33 axis has been reported as a potential therapeutic target for GVHD [20,21].
Despite the accumulating evidence that higher sST2 levels are associated with increased GVHD risk and subsequent mortality, a clinical role of sST2 during the early phase of transplantation has not been fully elucidated.We performed a multicenter, prospective, observational study monitoring the serial changes in sST2 levels before and early after HSCT with the aim of determining the diagnostic and prognostic values for acute GVHD, other transplant-related complications, and mortality.

Patients and Transplant Procedure
Patients who received first allogeneic HSCT for hematological diseases were enrolled consecutively between February 2014 and July 2015 at Yokohama City University Hospital or Yokohama City University Medical Center.The selection of donor source and conditioning regimen was based on patients' hematological diagnosis, donor availability, and patients' clinical status.
Conditioning regimens were classified into myeloablative conditioning (MAC) and reduced-intensity conditioning (RIC) according to the definitions previously reported [22].The pre-transplant risk category was defined as standard or high according to the diagnosis and the disease stage at the time of transplantation, as described previously [23].
Written informed consent was obtained from all the patients enrolled in this study before the start of the trial.This study was approved by the institutional review boards of our university hospital and medical center and was conducted in accordance with the Declaration of Helsinki.All authors vouched for the accuracy and completeness of the reported data, analyses, and adherence to the study protocol.

Peripheral Blood Samples
Samples were collected prospectively before conditioning, on the day of HSCT (day 0), and on days 14, 21, and 28 after HSCT.Serum sST2 and IL-33 levels were measured by ELISA (Human ST2/IL-1 R4 ELISA Kit and Human IL-33 ELISA Kit; R&D Systems, Minneapolis, MN, USA).Absorbance was measured using a microplate reader (Powerscan HT, DS Pharma Biomedical, Osaka, Japan).

Transplant-related Complications
Acute GVHD was diagnosed clinically with histological confirmation when available.The classification of acute GVHD was based on the diagnostic criteria of the 1994 Consensus Conference on Acute GVHD Grading [24].GVHD prophylaxis consisted of cyclosporine or tacrolimus with short-term methotrexate.Anti-thymocyte globulin was administered for HLA-serological mismatched transplant based on each institution's criteria.
Other transplant-related complications were diagnosed as follows: TMA was diagnosed according to the Blood and Marrow Clinical Trials Network and European Group for Blood and Marrow Transplantation guidelines [25,26].VOD was diagnosed based on the Baltimore and Seattle criteria [27,28].

Statistical Analysis
Fisher's exact test and the Mann-Whitney U test were used to assess the categorical and continuous variables, respectively.The receiver operating characteristics (ROC) curve from logistic regression models with the area under the curve (AUC) was used to present the correlation between sST2 and other biomarkers.The Pearson test was used to determine the correlation between sST2 and other biomarkers.The Kaplan-Meier method was used to assess overall survival (OS) using the log-rank test.For multivariate analysis, the Cox proportional hazards method was used to assess the OS.The Gray test and Fine-Gray test were used to assess the cumulative incidence of GVHD and NRM.The competing risks were GVHD and death prior to GVHD.Values of p<0.05 were considered statistically significant.All analyses were performed using EZR version 1.36 Matsumura A, et al: ST2 for Transplant-related Complications Turk J Hematol 2020;37:20-29 statistical software, which is a graphical user interface for R version 3.4.1 [29].

Patient Characteristics
The clinical characteristics of 32 patients are summarized in Table 1.The transplant procedure was deemed heterogeneous based on the donor source, conditioning regimen, and GVHD prophylaxis.

Expression Patterns of sST2 Following Transplantation
First, we evaluated the expression patterns of sST2 by serial sampling from individuals at a fixed time point before and early after HSCT.Different sST2 expression patterns were observed in each individual after the conditioning therapy (Figure 1A).Compared with median sST2 levels before conditioning (median=25.9 ng/mL; range=0-42.7 ng/mL), sST2 levels on day 0 were remarkably elevated in most of the patients (median=51.4 ng/mL; range=0-227.9ng/mL).The sST2 levels reached the maximum value on day 21 after HSCT (median=92.7 ng/mL; range=0-419.7 ng/mL) (Figure 1B).

Effects of Conditioning Damages and Inflammatory Conditions on sST2 Levels
As sST2 expression levels may depend on conditioning intensity [4], the sST2 levels in patients who underwent MAC and RIC were compared.The sST2 levels in patients who underwent MAC were higher than those in patients who underwent RIC on days 0, 14, 21, and 28 without statistical difference between the two groups (Figure 2).To determine whether sST2 was affected by various inflammatory conditions after HSCT, the correlations between sST2 and representative inflammatory markers, serum ferritin and C-reactive protein (CRP), were estimated (Table 2).sST2 levels were strongly correlated with ferritin and CRP levels in all samples (r=0.456 and 0.615, respectively).Specifically, sST2 was well correlated with CRP at days 0, 14, and 21 (r=0.717,0.630, and 0.628, respectively).Furthermore, levels of serum IL-33, the ligand of ST2, were under the detection limits in most of the patients, resulting in no correlation of IL-33 with sST2 levels (data not shown).

Association of sST2 Levels with Onset of Acute GVHD
The clinical courses of all patients are summarized in Figure 3.With a median follow-up of 21.5 months (range=0.9-35.4)after HSCT, 14 patients (43%) developed some grade of acute GVHD [median days to onset=39 days (range=9-84); median days to maximum grade=44 (range=15-94)].The maximum grade of GVHD was grade I in five patients, grade II in three, and grade III in six; none developed grade IV.Nine patients had involvement of the gastrointestinal tract (stage 1 in two patients, stage 2 in two, stage 3 in four, and stage 4 in one), and ten patients had skin involvement (stage 1 in one patient, stage 2 in five, and stage 3 in four).None developed liver GVHD.
When comparing the sST2 levels in patients with and without GVHD, median sST2 levels on day 14 after HSCT were relatively higher in patients with GVHD, but with no significant difference (Figure 4).Based on ROC curve analyses for predicting the onset of acute GVHD, the sST2 levels on day 14 showed the best AUC (0.66), with 66.7% sensitivity and 73.9% specificity (Figure S1, A).The optimal cut-off point of sST2 was 100 ng/ mL based on the ROC analysis.We focused on day 14 sST2 levels as the earliest time point for predicting acute GVHD.There was no significant difference in patient characteristics between the high-sST2 group (>100 ng/mL) and the low-sST2 group (≤100 ng/mL) (Table 1).The cumulative incidence of grades II-IV GVHD was significantly higher in the high-sST2 group (56.7%) than that in the low-sST2 group (16.5%) (p<0.01)(Figure 5A).Multivariate analyses showed that high sST2 was associated with higher incidence of acute GVHD (hazard ratio=9.35,95% confidence interval=2.92-30.0,p<0.01) (Table 3).sST2 levels at other time points of days 0, 21, and 28 were not associated with acute GVHD (data not shown).Although post-transplant CRP and ferritin levels were well correlated with sST2 (Table 2), these markers did not influence the cumulative incidence of acute GVHD (cut-off values were calculated by ROC analysis=CRP 6.0 mg/dL and ferritin 1700 mg/dL, respectively) (Figure 5B and Figure S1, B).These data suggest that the day 14 sST2 levels are supportive findings for predicting the onset of acute GVHD.
Among the nine patients who had grade II-IV GVHD, six patients were part of the high-sST2 group (Figure 6).The patients in the high-sST2 group developed acute GVHD relatively early after HSCT (median onset of GVHD=22 days; range=9-50), and sST2 levels were elevated several days before the onset of acute GVHD.Four patients (cases 6, 7, 11, and 12) developed acute GVHD before day 28, and the sST2 levels were the highest just before the development of GVHD.On the other hand, low-sST2 patients did not develop GVHD in the early phase of HSCT (median onset of GVHD=48 days; range=42-55).

Association of sST2 Levels with GVHD Severity and Target Organ Involvement
The association of day 14 sST2 levels with GVHD severity was examined, but the sST2 levels were not associated with the grade of GVHD (data not shown).The cumulative incidence of gastrointestinal GVHD was significantly increased in the high-sST2 group (50% vs. 15%, p=0.03).However, there was no significant association between sST2 levels and skin GVHD.

Discussion
Previous studies have shown that ST2 may predict the onset of acute GVHD, but no detailed analysis regarding the serial monitoring of sST2 following HSCT has been reported thus far.
In this study, we performed a prospective evaluation of sST2 expression patterns by serial monitoring before and early after HSCT and identified that early assessment of sST2 after HSCT can yield predictive values for the onset of acute GVHD in addition to other severe transplant-related complications, such as TMA and VOD.
Although high sST2 levels on day 28 are associated with an increased risk of acute GVHD in cord blood transplantation [2], the optimal timing to measure sST2 remains unclear.We conducted serial sST2 measurements at fixed time points during HSCT and found that high sST2 levels around day 14 had the most significant association for predicting GVHD development.
The expression patterns of sST2 in individual patients showed that sST2 is relatively higher during exacerbation of acute GVHD.As shown in Figure 6, the sST2 levels in patients with early-phase GVHD were higher than those with late-phase GVHD, suggesting that sST2 is clinically useful in predicting the early phase of GVHD.As acute GVHD often develops in the early   phase of HSCT, it is advantageous to establish biomarkers that can predict the onset of GVHD at earlier time points.
A soluble form of ST2 is released from endothelial cells, epithelial cells, and fibroblasts in response to inflammatory stimuli [30,31].A previous study has shown that sST2 levels are associated with conditioning intensity; sST2 levels were two to four times higher after the MAC regimen than the RIC regimen [4].The present study showed that sST2 levels after MAC were relatively higher than those after RIC.To reduce the effect of conditioning intensity on sST2 levels, we also examined the predictive values of the day 14/day 0 sST2 ratio for acute GVHD, but we did not detect any relationship with the development of GVHD (data not shown).Furthermore, we examined the association between sST2 and targets of GVHD involvement (skin, gastrointestinal tract, and liver) and found that most of the patients in the high-sST2 group developed gastrointestinal GVHD, which is thought to occur due to endothelial damage.These findings imply that the release of sST2 during GVHD exacerbation partly depends on the degree of endothelial injury occurring after conditioning therapy.
On the other hand, examining the association of sST2 with other complications and mortality is also important.Vander Lugt et al. [4] showed that high sST2 levels were associated with NRM within 6 months after HSCT.Moreover, other studies showed that high sST2 levels had a similar association with increased risk of NRM [3,12,13].In this study, although it was difficult to evaluate this issue since only four patients developed NRM, they all showed high sST2 levels on day 14.Notably, they died of sepsis, graft failure, TMA, and pneumonia, without any GVHD, and all of them developed VOD.Recent studies demonstrated that high sST2 levels were significantly associated with TMA [15] and VOD [16], both of which are characterized by endothelial cell injury.
In accordance with these observations, two patients who developed TMA and five who developed VOD showed high sST2 levels on day 14, even in our small cohort.In addition, the correlation of sST2 with CRP and ferritin suggests that sST2 is released under a variety of inflammatory conditions.Several studies have shown that pro-inflammatory cytokines and their receptors are potential GVHD biomarkers, but some factors other than GVHD contribute to these cytokines' release (e.g., TNF-α, TNFR1, IL-6) [5,6,32].Therefore, caution should be taken when diagnosing patients who have high sST2 levels as sST2 is not a specific biomarker for acute GVHD.

Study Limitations
The limitations of this study include the small number of patients and the heterogeneous patient populations and transplant procedures, such as conditioning regimen, donor source, and disease status.Even though we conducted detailed analyses of sST2 expression patterns in individual patients, it was difficult to exclude the multiple factors that can cause an increase in sST2 (e.g., cardiac overload or infection).

Conclusion
We revealed that sST2 levels increased not only in patients with acute GVHD but also in those with other life-threatening complications, such as TMA, VOD, and severe infection, as identified during patient monitoring.Although these complications often overlap with each other in the clinical settings for HSCT, gastrointestinal GVHD, TMA, and VOD, all of which are linked to endothelial injury, may be key complications related to high sST2 release.The early assessment of sST2 after HSCT may be a predictive indicator for acute GVHD and other transplant-related complications.Further studies with larger sample sizes and serial monitoring are needed to clarify the clinical value of sST2.

29 Figure 1 .
Figure 1.Expression patterns of sST2 following transplantation.A) Expression patterns of sST2 before conditioning (Pre) and on days 0, 14, 21, and 28 after transplantation in individual patients.B) Expression patterns of median sST2 values before conditioning (Pre) and on days 0, 14, 21, and 28 after transplantation.Box plots indicate medians, interquartiles, and ranges of sST2 levels.sST2: Suppression of tumorigenicity 2.

Figure 2 .
Figure 2. Comparison of sST2 expression patterns based on conditioning intensity.Median sST2 levels in patients who received MAC and RIC following transplantation.Box plots indicate medians, interquartiles, and ranges of sST2 levels.RIC: Reduced-intensity conditioning, MAC: Myeloablative conditioning, sST2: Suppression of tumorigenicity 2.

Figure 4 .
Figure 4. Comparison of sST2 expression patterns in patients with and without acute GVHD.Mean sST2 levels in patients who developed grade II-IV acute GVHD or not following transplantation.Bars indicate the mean ± SEM of sST2 levels.aGVHD: Acute graft-versus-host disease, sST2: Suppression of tumorigenicity 2, GVHD: Graft-versus-host disease.

Matsumura A, et
al: ST2 for Transplant-related Complications pneumonia (n=1), which often overlapped.Four patients in the high-sST2 group died of transplant-related complications including sepsis, graft failure, TMA, and pneumonia (cases 1, 3, 4, and 5) without developing grade II-IV GVHD.The two patients who had TMA (cases 3 and 9) and five who had VOD (cases 1-5) were included in the high-ST2 group.All patients who died had severe overlapping complications including VOD.

TurkJ 29 Figure 5 .
Figure5.High sST2 on day 14 is correlated with subsequent GVHD development.A) Cumulative incidence of grade II-IV acute GVHD by day 100 in patients with high sST2 (>100 ng/mL) and low sST2 (≤100 ng/mL) levels on day 14 after transplantation.B) Cumulative incidence of grade II-IV acute GVHD by day 100 in patients with high-CRP (>6.0 mg/dL) and low-CRP (≤6.0 mg/dL) levels on day 14 after transplantation.C) Cumulative incidence of grade II-IV acute GVHD by day 100 in patients with high ferritin (>1700 mg/dL) and low ferritin (≤1700 mg/dL) levels on day 14 after transplantation.The CRP and ferritin cut-off values were calculated by ROC curve analysis for predicting the onset of acute GVHD (ROC curves are available in FigureS1).sST2: Suppression of tumorigenicity 2, CRP: C-reactive protein, GVHD: Graft-versus-host disease.

Figure 6 .
Figure 6.Association of early-phase sST2 levels with subsequent development of acute GVHD.Patient numbers correspond with those in Figure 3. Clinical course of 14 patients who developed grade I-IV acute GVHD.Dotted bars (grade I GVHD) or striped bars (grade II-IV GVHD) indicate the duration of acute GVHD clinical symptoms.Cut-off point of sST2 is 100 ng/mL.aGVHD: Acute graft-versus-host disease, Pt.No.: Patient number, GVHD: Graft-versus-host disease, sST2: Suppression of tumorigenicity 2.

Matsumura A, et
al: ST2 for Transplant-related Complications