www.nature.

com/bmt

ARTICLE
The effect of danger-associated molecular patterns on survival
in acute graft versus host disease
1✉ 2
Serhat Çelik , Leylagül Kaynar , Zeynep Tuğba Güven3, Kübra Atasever Duran4, Olgun Kontaş5, Muzaffer Keklik 3
and Ali Ünal3

© The Author(s), under exclusive licence to Springer Nature Limited 2023

Danger-associated molecular patterns (DAMPs) are molecules that can initiate and maintain robust inflammatory responses and
were investigated in the pathogenesis of graft versus host disease (GvHD). Uric acid (UA) and fibrinogen (Fib) are DAMPs released
from damaged tissue during allogeneic hematopoietic stem cell transplantation (allo-HCT) and GvHD. We aimed to evaluate the
effects of UA and Fib levels on survival in GvHD. One hundred seventy-four patients with grade 2-4 acute GvHD were included. UA
and Fib levels were evaluated on allo-HCT day 0 and GvHD on days 0, 7, 14, and 28. Fib GvHD day 0 was the independent predictor
for overall survival (OS), non-relapse mortality (NRM), and progression-free survival in multivariable models (HR 0.98, p < 0.001; HR
0.98, p = 0.001, HR 0.98, p = 0.006, respectively). Also UA GvHD day 28 was the independent predictor for OS and NRM (HR 0.77,
p = 0.004; HR 0.76, p = 0.011, respectively). Our results indicated that hypouricemia and hypofibrinogenemia were associated with a
significantly shorter OS and higher NRM. UA and Fib are remarkable molecules in GvHD because they are routinely utilized, readily
1234567890();,:

available, can be therapeutic targets, and have DAMPs and antioxidant features.

Bone Marrow Transplantation; https://doi.org/10.1038/s41409-023-02145-7

INTRODUCTION METHODS
Graft versus host disease (GvHD) is one of the major complica- Patients and study design
tions in allogeneic hematopoietic stem cell transplantation (allo- Patients who underwent allo-HCT between 2010 and 2020 in Erciyes
HCT). GvHD has a prevalence of 20–60% and has risk factors University Şahinur Dedeman Hospital Bone Marrow Transplantation Unit
such as HLA incompatibility, the intensity of conditioning were included in this retrospective study. Data analysis was completed in
March 2021. Bone marrow and cord blood as stem cell source,
and prophylactic regimen, donor and recipient sex mismatch,
haploidentical transplantation, missing value, and diagnosis with
and graft source [1, 2]. Also, in the last few years, it has transplant-associated microangiopathy (TAM) were excluded. Those with
been observed that danger-associated molecular patterns HLA full-matched related or unrelated donors were included. Finally, 174
(DAMP)s have effects on GvHD [3]. DAMPs are molecules that patients with acute GvHD were included (Supplementary Fig. 1).
can initiate and maintain robust inflammatory responses First, baseline characteristics such as age, gender, diagnoses, blood
by activating the innate immune system to release cellular groups, and transplant characteristics were determined, with Fib and UA
stress or tissue damage and are considered endogenous danger levels examined afterward. Since the effect of DAMPs on survival in GvHD
signals [4]. was investigated, and it is still unclear which day is more effective, Fib and
Uric acid (UA) and fibrinogen (Fib) are DAMPs [5] and are UA levels were evaluated on the day of GvHD (day 0) and the following
released from damaged tissue during the conditioning regimen 7th, 14th, and 28th days. In addition, the day of allo-HCT (day 0) was also
evaluated. Fib levels (Fib-preconditioning) were also evaluated before
for allo-HCT and GvHD. UA stimulates T cells to release interleukin- starting the conditioning regimen. However, there was no definite day for
1β through activation of the NOD-like receptor protein (NLRP) 3, Fib-preconditioning. Fib values measured ranged from 23 to 8 days before
and Fib stimulates macrophages to produce inflammatory allo-HCT. Toxicity, and adverse events were classified using the Common
cytokines via toll-like receptor (TLR) 2 or 4 [5, 6]. Terminology Criteria for Adverse Events v5.0 [10].
No studies have examined the relationship between Fib and The criteria determined by the International Working Group were used
GvHD, but few studies with UA that contain controversial results to diagnose of TAM [11]. Apart from TAM, subclinical endothelial damage is
[7–9]. These studies focused more on incidence than survival. In also observed at a high rate in GvHD patients [12]. Therefore, we evaluated
addition, in these studies, DAMPs were evaluated during the the endothelial activation and stress index (EASIX), which includes lactate
period of allo-HCT administration and were not evaluated during dehydrogenase (LDH), creatinine, and platelet [13]. And ursodeoxycholic
acid 3 × 250 mg/day is routinely given to all patients as an endothelial
the development of GvHD. We aimed to examine the effect of protector until the 180th day after allo-HCT.
DAMPs released from damaged tissue during GvHD development All patients signed written informed consent. All ethical procedures and
and the subsequent impact on survival in GvHD. Among the standards were carried out in accordance with the 1975 Helsinki
DAMPs, Fib and UA, which are already routinely evaluated and Declaration. The study was approved by the Erciyes University Clinical
ready treatment targets, were preferred. Research Ethics Committee (Approval number: 2021/345).

1
Department of Hematology, Faculty of Medicine, Kırıkkale University, Kırıkkale, Türkiye. 2Department of Hematology, Faculty of Medicine, Medipol Mega University, İstanbul,
Türkiye. 3Department of Hematology, Faculty of Medicine, Erciyes University, Kayseri, Türkiye. 4Department of Internal Medicine, Faculty of Medicine, Erciyes University, Kayseri,
Türkiye. 5Department of Pathology, Faculty of Medicine, Erciyes University, Kayseri, Türkiye. ✉email: serhatcelikmd@gmail.com

Received: 24 August 2021 Revised: 2 October 2023 Accepted: 30 October 2023

 S. Çelik et al.
2
Table 1. Baseline characteristics. Table 2. Endpoints and treatments findings.
Variables All population n = 174 Variables All population n = 174
Age, years 37 (18–69) GvHD, n(%)
Gender, n(%) Skin 116 (66.7)
Female 68 (39.1) Grade 2 48 (41)
Male 106 (60.9) Grade 3 57 (48.7)
Diagnosis, n(%) Grade 4 12 (10.3)
Acute myeloid leukemia 95 (54.6) Gastrointestinal tract 63 (36.2)
Acute lymphoblastic leukemia 33 (19.0) Grade 2 26 (41.3)
Myelodysplastic syndrome 15 (8.6) Grade 3 29 (46.0)
Hodgkin lymphoma 7 (4) Grade 4 8 (12.7)
Non-Hodgkin lymphoma 7 (4) Liver 69 (39.7)
Aplastic anemia 4 (2.3) Grade 2 30 (41.7)
Chronic myeloid leukemia 5 (2.9) Grade 3 41 (56.9)
Miyelofibrozis 3 (1.7) Grade 4 1 (1.4)
Others 5 (2.9) Lung 2 (1.1)
Complete remission, n(%) Grade 2 2 (100)
1 61 (35.1) Prophylaxis and treatment, n(%)
2 66 (37.9) Cyclosporine 174 (100)
>2 47 (27) Corticosteroid 174 (100)
Conditioning regimen, n(%) Methotrexate 173 (99.4)
Anti-thymocyte globulin 20 (11.5) Cyclophosphamide 2 (1.1)
Myeloablative conditioning 141 (81) Mycophenolate mofetil 75 (43.1)
Reduced intensity conditioning 33 (19) Photopheresis 59 (33.9)
Recipient blood group, n(%) Mesenchymal stem cell 47 (27)
A 91 (52.3) Imatinib 18 (10.3)
B 24 (13.8) Ruxolitinib 17 (9.8)
AB 12 (6.9) Tacrolimus 8 (4.6)
0 47 (27) Alfa-beta T cell selection 4 (2.3)
Recipient Rh, n(%) Steroid refractory 99 (56.9)
Negative 21 (12.1) Acute GVHD latest status, n(%)
Positive 153 (87.9) fixed 42 (24.1)
Donor gender, n(%) has become chronic 112 (64.4)
Female 73 (42) did not improve 20 (11.5)
Male 101 (58) Median GVHD timea, days 63 (10–89)
Gender high risk, n(%) 44 (25.3) Relapse, n(%) 37 (21.3)
Donor, n(%) Median PFSa, months 15.7 (1–129.3)
Matched related 166 (95.4) Overall mortality, n(%) 84 (48.3)
Matched unrelated 8 (4.6) Median OSa, months 25.2 (1–129.3)
Donor age, years 37 (16–69) Non-relaps mortality, n(%) 56 (32.2)
Donor blood group, n(%) CMV, n(%) 141 (81.0)
A 79 (45.4) BK virus, n(%) 52 (29.9)
B 34 (19.5) CMV Cytomegalovirus, GvHD graft versus host disease.
a
AB 7 (4) Median (95% CI) from HSC to date of event. Categorical variables were
shown as number(%).
0 54 (31)
Donor Rh, n(%)
Negative 18 (10.3) Transplantation procedures and GvHD evaluation
Positive 156 (89.7) Acute GvHD was graded according to the modified Seattle Glucksberg
criteria [14], and those with grade 1 were excluded. Since the nature of the
Baseline LFT abnormalities, n(%)
study was retrospective, patients whose diagnosis of GvHD was confirmed
Grade 1 24 (13.8) by biopsy taken from the relevant tissue were included. A female donor
HSC, ×106/kg 7.18 (4.83–10.77) with a male recipient was defined as gender-high risk [15]. Myeloablative
Neutrophil engraftment day 16 (9–33) conditioning (MAC) was defined as cyclophosphamide 60 mg/kg ×2 days,
and intravenous busulfan 12.8 mg/kg total dose or total body irradiation 12
Platelet engraftment day 13 (7–47) Grey or fludarabine 120–180 mg/m2 ×4 days [16, 17]. Fib levels were
EASIX 1.5 (0.3–48.1) evaluated using the Sysmex CS-5100 Hemostasis System coagulation
EASIX The endothelial activation and stress index, HSC hematopoietic stem analyzer (Siemens), and UA levels were evaluated using a Cobas 8000 c702
cell, LFT liver function test. clinical chemistry analyzer (Roche-Hitachi). Allopurinol or rasburicase,
Categorical variables were shown as number(%), and numerical variables which causes a depletion of UA, was not used in the conditioning
as median (min-max). regimens. Although it varies according to the allo-HCT protocols, patients
were using cyclosporine, mycophenolate mofetil or methotrexate at the

Bone Marrow Transplantation

 S. Çelik et al.
3
Table 3. Optimal thresholds for uric acid in different periods of time for predicting of endpoints.

Fibrinogen HSCT GvHD

Day 0 Day 0 Days 7 Days 14 Days 28
Median (min-max) 290 (117–776) 217 (66–665) 289 (81–665) 310.5 (88–808) 327.5 (105–768)
Relapse
Thresholds ≤342 ≤189* ≤223* ≤269 ≤213
Sensitivity - Specificity (%) 86.5–32.9 67.6–66.4 48.7–73.5 46.0–70.4 35.1–87.2
AUC(Confidence Interval) 0.56 (0.48–0.63) 0.64 (0.56–0.71) 0.59 (0.52–0.67) 0.52 (0.44–0.59) 0.55 (0.47–0.62)
NRM
Thresholds ≤203 ≤225* ≤278* ≤218 ≤347
Sensitivity - Specificity (%) 28.6–85.6 80.4–60.2 56.4–60.2 35.2–85.6 63.5–47.5
AUC(Confidence Interval) 0.53 (0.45–0.61) 0.72 (0.64–0.78) 0.60 (0.53–0.67) 0.59 (0.51–0.66) 0.53 (0.46–0.61)
OS
Thresholds ≤192 ≤194* ≤193* ≤232 ≤233
Sensitivity - Specificity (%) 22.6–91.1 73.8–85.6 37.4–91.1 39.0–86.7 31.3–87.8
AUC(Confidence Interval) 0.56 (0.48–0.63) 0.84 (0.77–0.89) 0.65 (0.57–0.72) 0.59 (0.50–0.67) 0.56 (0.48–0.63)
Uric acid HSCT GvHD
Day 0 Day 0 Days 7 Days 14 Days 28
Median (min-max) 3.6 (1.0–10.0) 3.0 (0.8–10.4) 3.1 (0.8–10.5) 3.4 (0.8–10.2) 3.8 (1.0–9.8)
Relapse
Thresholds ≤2.4 ≤1.7 ≤2.2 ≤2.3 <4.3
Sensitivity - Specificity (%) 37.8–78.8 35.1–91.8 37.8–77.2 32.4–80.7 45.9–68.4
AUC(Confidence Interval) 0.53 (0.45–0.61) 0.58 (0.49–0.67) 0.54 (0.47–0.62) 0.55 (0.47–0.62) 0.52 (0.44–0.59)
NRM
Thresholds ≤6.7 ≤3.9 ≤2.7* ≤2.9* ≤4.3*
Sensitivity - Specificity (%) 98.2–11.9 78.6–32.2 50.9–70.3 55.6–71.2 86.5–44.1
AUC(Confidence Interval) 0.50 (0.43–0.58) 0.53 (0.45–0.61) 0.61 (0.53–0.68) 0.62 (0.55–0.70) 0.68 (0.60–0.75)
OS
Thresholds ≤3.7 ≤3.2* ≤2.8* ≤2.9* ≤3.4*
Sensitivity - Specificity (%) 57.1–54.4 69.1–51.1 53.0–76.7 53.7–77.8 55.0–70.0
AUC(Confidence Interval) 0.55 (0.47–0.62) 0.62 (0.55–0.70) 0.65 (0.58–0.72) 0.67 (0.59–0.74) 0.66 (0.58–0.73)
The measure of all fibrinogen and uric acid levels is mg/dL.
GvHD graft versus host disease, HSCT hematopoietic stem cell transplantation.
*p < 0.05 statistical significant.

onset of GvHD. However, when GvHD was suspected in all patients, Cox proportional hazards model was used for multivariable regression. All
corticosteroids were started after blood tests were taken. Therefore, all of variables found to be significant in the univariable regression analysis were
our patients were steroid free at the time of GvHD onset. included in the stepwise multivariable regression model. Results were
expressed as the hazard ratio (HR) with a 95% confidence interval (95% CI).
P < 0.05 was taken as statistical significance.
Statistical analysis The threshold values for serum Fib and UA levels in five different
Statistical analyses of collected data were conducted using R 3.4.2 (R Core
periods were determined using the receiver operating characteristic
Team (2017). R: A language and environment for statistical computing. R
(ROC) curves and median. ROC curves analysis also assessed the
Foundation for Statistical Computing, Vienna, Austria. URL https://www.R- sensitivity, specificity, and the area under curve (AUCs). The data that
project.org/.). Determination of the normally distributed data was support the findings of this study are available from the corresponding
conducted using the Kolmogorov–Smirnov test. Numerical variables with author upon request.
normal distribution were expressed as the mean ± standard deviation,
while those with non-normal distribution were expressed as the median
(min-max). The categorical variables were expressed as numbers and
percentages. RESULTS
Endpoints were relapse incidence, non-relapse mortality (NRM), overall Patients and transplant characteristics
survival (OS), and progression-free survival (PFS). Relapse incidence was The median age of the patients was 37 (18–69) years and 60.9%
defined as the probability of having had a relapse during follow-up time. (n = 106) were male. The most common diagnosis was acute
Death without experiencing a relapse was a competing event. NRM was myeloid leukemia constituting 54.6% (n = 95), followed by acute
defined as death without evidence of relapse or progression. OS was lymphoblastic leukemia with 19% (n = 33). In disease states,
defined as the time from allo-HCT to death, regardless of the cause. PFS
complete remission (CR) 2 was observed most frequently in 37.9%
was defined as survival with no evidence of relapse or progression.
Cumulative incidence was used to estimate the endpoints of relapse (n = 66) of patients, followed by CR1 in 35.1% (n = 61) and CR > 2
incidence and NRM. Probabilities of OS and PFS were calculated using the in 27% (n = 47) of patients. Gender high risk was present in 44
Kaplan–Meier method. Univariate analyses were done using the Gray test (25.3%) patients and the median EASIX score was 1.5 (0.3–48.1).
for cumulative incidence functions and the log-rank test for OS and PFS. A The essential characteristics are described in Table 1.

Bone Marrow Transplantation

 S. Çelik et al.
4
a 100
b 100
Fibrinogen > 225 mg/dL
90 90 Fibrinogen > 225 mg/dL

80 80
Progression free survival (%)

Survival probability (%)

70 70

60 60
Fibrinogen d 225 mg/dL
50 50
Fibrinogen d 225 mg/dL
40 40
HR = 4.74
30 95% CI = 2.76-8.14 30 HR = 6.13
955 CI = 3.92-9.58
20 20

10 10
Logrank P < 0.001 Logrank P < 0.001
0 0
0 12 24 36 48 60 72 84 96 108 120 132 0 12 24 36 48 60 72 84 96 108 120 132
Follow-up time (months) Follow-up time (months)

c d

Cumulative incidence of non-relaps mortality

1.0
Gray’s test
1.0 P < 0.001
Cumulative incidence of relaps

Gray’s test
0.8
0.8 P < 0.001
Fibrinogen d 225 mg/dL
0.6 0.6 Fibrinogen d 225 mg/dL

0.4 0.4
Fibrinogen > 225 mg/dL
0.2 Fibrinogen > 225 mg/dL
0.2

0
0
0 12 24 36 48 60 72 84 96 108 120 132 0 12 24 36 48 60 72 84 96 108 120 132
Follow-up time (months) Follow-up time (months)

Fig. 1 Endpoints results by fibrinogen level. A The progression-free survival was lower in those with fibrinogen ≤225 mg/dL. B The risk of
mortality was higher in those with fibrinogen ≤225 mg/dL. C The cumulative incidence of relapse was higher in those with fibrinogen
≤225 mg/dL. D The cumulative incidence of non-relapse mortality was higher in those with fibrinogen ≤225 mg/dL.

Acute GvHD characteristics and survival endpoints have PFS than patients with >225 mg/dL (95% CI = 2.76–8.14;
In GvHD organ involvement, the skin was observed most p < 0.001) (Fig. 1A), and the overall mortality risk was 6.13 times
frequently with 66.7% (n = 116), followed by the liver in 41% more (95% CI = 3.92–9.58; p < 0.001) (Fig. 1B). The cumulative risk
(n = 69) of patients, the gastrointestinal (GI) tract in 39.7% of relapse was 3.60 times higher (95% CI = 31.76–7.34; Gray’s test
(n = 63), and 1.1% (n = 2) had lung involvement. All patients p < 0.001) (Fig. 1C), and the cumulative risk of NRM was 4.74 times
received cyclosporine and steroid. Other treatments are detailed higher (95% CI = 2.76–8.14; Gray’s test p < 0.001) (Fig. 1D). The
in Table 2. The median follow-up time after allo-HCT was positive predictive value (PPV) of fibrinogen for overall mortality
25.2 (1–129.3) months. Relapse occurred in 21.3% of the patients, was 57.5%, while the negative predictive value (NPV) was 81.1%.
the NRM rate was 32.2%, and the OS rate was 51.7% (Table 2). The The PPV of Fib for relapse was 23.5%, while the NPV was 79.1%.
cumulative incidence of 100-day, 1-year, and 2-year relapses was The PPV of Fib for NRM was 23.5%, while the NPV was 79.1%.
3%, 19%, and 28%, respectively. The cumulative incidence of 100- Grade 1 baseline LFT abnormalities were observed in 6 (15%)
day, 1-year, and 2-year NRM was 8%, 28%, and 33%, respectively. patients with Fib levels ≤225 mg/dL at GvHD day 0, and grade 1
The cumulative probability of 100-day, 1-year, and 2-year survival baseline LFT abnormalities were observed in 18 (13.4%) patients
was 77%, 61%, and 54%, respectively. with Fib levels above 225 mg/dL (p = 0.8). C-reactive protein (CRP)
levels were examined on GvHD day 0 because Fib has an acute
Threshold values of Fib and UA phase reactant feature. In those with Fib level ≤225 mg/dL at
Fib and UA levels at different follow-up points from the day of GvHD day 0, the median CRP level was 60.5 (2.9–221) mg/dL, while
allo-HCT are shown in Table 3. For relapse, NRM, and OS, the AUC those above 225 mg/dL were 45.5 (1–259) mg/dL (p = 0.14).
of Fib levels on the day of development of GvHD was higher than Disseminated intravascular coagulation or sepsis was not present
at other follow-ups. The cut-off value of Fib level at the time of in any of our patients at GvHD day 0.
GvHD development was 189 mg/dL for relapse, 225 mg/dL for UA levels were not statistically significant in predicting relapse
NRM, and 194 mg/dL for overall mortality. On the other hand, to in all follow-ups. In predicting NRM and OS, the AUC values of UA
determine a common cut-off value for all endpoints of Fib level, levels at day 28 of GvHD were higher than in other follow-ups. The
225 mg/dL, which is the cut-off value in predicting NRM, was cut-off value of UA level on day 28 of GvHD was 4.3 mg/dL for
chosen because it includes other cut-off values. This selection NRM and 3.4 mg/dL for overall mortality. To determine a common
increased sensitivity for relapse (70.3%) but decreased specificity cut-off value of UA level for all endpoints, 4.3 mg/dL, which is the
(51.8%), likewise increased sensitivity for overall mortality (83.3%) cut-off value for predicting NRM, was chosen because it includes
but decreased specificity (75.6%). According to this, patients with the overall mortality cut-off value. Thus, it increased sensitivity for
Fib level ≤225 mg/dL at GvHD day 0 were 4.74 times less likely to overall mortality (75.0%) but reduced specificity (43.3%). Patients

Bone Marrow Transplantation

 S. Çelik et al.
5
a 100
b 100
Uric acid > 4.3 mg/dL
90 90

80 80
Progression free survival (%)

70 70 Uric acid > 4.3 mg/dL

Survival probability (%)

60 Uric acid d 4.3 mg/dL 60

50 50 Uric acid d 4.3 mg/dL

40 HR = 3.22 40
95% CI = 1.87-5.56
HR = 2.07
30 30 95% CI = 1.33-3.22

20 20

10 10
Logrank P < 0.001 Logrank P = 0.001
0 0
0 12 24 36 48 60 72 84 96 108 120 132 0 12 24 36 48 60 72 84 96 108 120 132
Follow-up time (months) Follow-up time (months)

c d

Cumulative incidence of non-relaps mortality

1.0 1.0
Gray’s test
P < 0.001
Gray’s test
Cumulative incidence of relaps

0.8 P = 0.383 0.8

Uric acid d 4.3 mg/dL

0.6 0.6

Uric acid d 4.3 mg/dL

0.4 0.4

0.2 0.2 Uric acid > 4.3 mg/dL

Uric acid > 4.3 mg/dL

0 0
0 12 24 36 48 60 72 84 96 108 120 132 0 12 24 36 48 60 72 84 96 108 120 132
Follow-up time (months) Follow-up time (months)

Fig. 2 Endpoints results by uric acid level. A The progression free survival was lower in those with uric acid ≤ 4.3 mg/dL. B The risk of
mortality was higher in those with uric acid ≤4.3 mg/dL. C The cumulative incidence of relapse did not differ significantly in those with uric
acid ≤4.3 mg/dL. D The cumulative incidence of non-relapse mortality was higher in those with uric acid ≤4.3 mg/dL.

with UA level ≤ 4.3 mg/dL at GvHD day 28 had lower PFS and DISCUSSION
higher overall mortality compared to patients with elevated UA In the development of GVHD, cells are damaged; as a result,
levels (Fig. 2A, HR 3.22, 95% CI: 1.87–5.56; p < 0.001, Fig. 2B, HR molecules called DAMPs are released from these damaged cells [4,
2.07, 95% CI: 1.33–3.22; p = 0.001 respectively). There was no 18]. Adenosine-5′-triphosphate, a DAMP, was demonstrated to
significant difference in cumulative incidence of relapse risk in play a role as an endogenous danger signal in GvHD [19, 20]. The
those with low UA levels (Gray’s Test P = 0.383) (Fig. 2C). However, relationship between UA, a DAMP acting on NLRP3, and GvHD was
NRM was significantly increased with low UA levels (HR 3.22, 95% investigated in a preclinical and phase 1 clinical study [20, 21]. In
CI: 1.87–5.56; Gray’s test p < 0.001) (Fig. 2D). The PPVs of UA were both studies, depletion of UA with urate oxidase and allopurinol
54.1%, 40.5%, and 18% for overall mortality, NRM, and relapse, reduced the frequency and severity of GvHD. Two later retro-
respectively. The NPVs of UA were 66.1%, 88.1%, and 71.2% for spective studies showed that those with low UA levels had a
overall mortality, NRM, and relapse, respectively. higher incidence of GvHD. A prospective study found no
correlation between UA levels and GvHD incidence [7–9]. The
Independent predictors of survival endpoints preclinical and phase 1 study assessed the effects of urate oxidase
Potential risk factors associated with relapse, NRM, and overall and allopurinol on GvHD rather than UA levels.
mortality by univariate analyzes are shown in Supplementary Ostendorf et al. reported no relationship between UA levels and
Tables 1–3. In the multivariate cox regression model in which OS and relapse [7]. Penack et al. observed that OS and PFS were
potential risk factors were included, Fib level in GvHD day 0 was significantly shorter in patients with UA levels above the median
determined as an independent risk factor for all endpoints (HR measured before the onset of the conditioning regimen [9].
0.98, p = 0.006 for relapse, HR 0.98, p = 0.001 for NRM and HR 0.98, Conversely, Ghasemi et al. demonstrated that those with low UA
p < 0.001 for overall mortality). UA level on GvHD day 28 was levels had an increased risk of death and inferior OS, and those
determined as an independent risk factor for NRM and overall with above median UA levels had a 35% less risk of death [8].
mortality (HR 0.76, p = 0.011 and HR 0.77, p = 0.004, respectively). Although there are controversial results, it was observed in our
The most effective independent variable in the multivariate study that those with low UA and Fib levels had significantly
analysis for relapse was CR > 2 (HR 6.09, p < 0.001). Unimproved shorter OS and higher NRM. UA and Fib increase and maintain
acute GvHD is the most effective independent variable for both non-infectious inflammatory responses. In addition, Haen et al.
NRM and overall mortality (HR 10.22, p < 0.001; HR 8.87, p < 0.001, suggested that UA may indicate incipient or remaining immuno-
respectively). All independent predictors of survival endpoints are logical activity after induction chemotherapy or SCT and may be
shown in Table 4. effective in immune remodeling [22]. Immune remodeling and

Bone Marrow Transplantation

 S. Çelik et al.
6
However, Luft et al. developed a parameter called EASIX, which
Table 4. Independent predictors of endpoints.
contains LDH, creatinine, and platelet [13]. Unlike other para-
Variables Multivariable Cox Regression meters, EASIX is easy to use, standardized, and cost-effective.
EASIX at GvHD day 0 (EASIX-GvHD) was shown to be strongly
HR 95% CI p predictive in NRM and OS (HR 1.24, p < 0.0001; HR 1.23, p < 0.0001,
Relapse respectively) but no difference for relapse. In a large cohort that
Complete remission EASIX was evaluated before conditioning (EASIX-pre), it was found
1 ref
to be predictive for OS and NRM in the training group (HR 1.09,
p = 0.0033; HR 1.23, p = 0.0002, respectively) and predictive for
2 1.28 0.49–3.34 0.62 OS, NRM, and relapse in the validation group (HR 1.23, p < 0.001;
>2 6.09 2.48–14.96 <0.001* HR 1.31, p < 0.0001; HR 1.11, p = 0.003, respectively) [30]. Also, in
Fibrinogen GvHD Day 0 0.98 0.97–0.99 0.006* our study, EASIX-GvHD was an independent predictor of OS and
−2 Log Likelihood = 313.1; p < 0.001
NRM. In our study, Fib was the independent variable predicting
OS, NRM, and relapse, and its relationship with endothelial
Non-relapse mortality functions is known [31]. Therefore, modification of EASIX with
EASIX 1.04 1.01–1.07 0.011* Fib may be considered in future studies.
Age 1.03 1.01–1.05 0.021* There are many predictive markers that assess NRM, overall
Acute GvHD latest status, n(%)
mortality and prognosis in GvHD. In the study evaluating
angiopoietin 2 and endothelial cell-derived microparticles, it was
fixed ref observed that they were associated with the severity of GvHD [32].
has become chronic 0.8 0.36–1.82 0.6 It was stated that interleukin-2 receptor alpha, interleukin-8,
did not improve 10.22 3.80–27.50 <0.001* hepatocyte growth factor, and tumor necrosis factor receptor-1
Fibrinogen GvHD Day 0 0.98 0.97–0.99 0.001*
and T cell Ig and mucin domain 3 (Tim-3) levels, which show T cell
exhaustion, were associated with GvHD severity and NRM [33, 34].
Uric acid GvHD Day 28 0.76 0.62–0.94 0.011* The scoring system with suppression of tumorigenicity 2 (ST2) and
−2 Log Likelihood = 411.9; p < 0.001 regenerating islet-derived 3 alpha (REG3 α) was shown to predict
Overall mortality six-month NRM [35]. Although there are many predictive markers
EASIX 1.04 1.01–1.06 0.003*
for GvHD in the literature, these are not routine parameters in
every clinic. However, the UA and Fib used in our study are
Age 1.02 1.01–1.04 0.008* biomolecules that are easily accessible, can be found in every
Acute GvHD latest status, n(%) center and have easily targeted for therapy.
fixed ref There are some limitations in our study. First is the time interval
has become chronic 0.77 0.43–1.41 0.406
in which UA and Fib levels are evaluated. Ostendorf et al.
evaluated UA levels at allo-HCT day 0. This was because while peri-
did not improve 8.87 3.75–21.02 <0.001* transplant UA depletion decreased the incidence of GvHD in the
Fibrinogen GvHD Day 0 0.98 0.97–0.99 <0.001* preclinical study, depletion five days after allo-HCT did not change
Uric acid GvHD Day 28 0.77 0.65–0.92 0.004* the severity of GvHD [7, 21]. Penack et al. evaluated the time
Relapse 1.7 1.02–2.84 0.041*
before starting the conditioning regimen, which ranged from day
22 to day 0 before allo-HCT [9]. In another study, UA levels were
CMV 2.68 1.25–5.76 0.011* measured on day 7 before allo-HCT, day 0, and day 7 and 14 after
−2 Log Likelihood = 528.6; p < 0.001 allo-HCT [8]. However, these studies focused more on incidence
CI Confidence interval, CMV Cytomegalovirus, EASIX The endothelial than survival. Since we examined the effect of UA and Fib on
activation and stress index, GvHD graft versus host disease, HR Hazard ratio. survival in GvHD, we evaluated the day when DAMPs are released
*p < 0.05. from the damaged tissue - that is, the day GvHD develops - and
the first, second, and fourth weeks after that. Another limitation is
that haploidentical allo-HCTs were not included. Haploidentical
allo-HCTs are being performed more and more, but were excluded
increased inflammatory responses in allo-HCT recipients are because of the well-known risk factor for GvHD, affecting survival
crucial for anti-leukemic and anti-infectious efficacy. Therefore, and the homogeneity of our study group. Yet another limitation is
we hypothesize that inadequate anti-leukemic and anti-infectious that our study’s retrospective nature and structure did not answer
inflammatory responses why allo-HCT recipients with low UA and questions about the effects of UA and especially Fib at the
Fib levels during the GvHD period have less OS and more NRM molecular level in GvHD. The last limitation is that our study may
and even more relapses for low Fib. not be generalizable due to the single-center results.
UA and Fib have antioxidant activity [23, 24]. UA is one of the In conclusion, this study supports that UA and Fib are
most essential antioxidant molecules in the periphery, and a biomolecules that show OS and NRM in GvHD. UA and Fib are
decrease in UA levels reduces antioxidant capacity [7]. The remarkable molecules in GvHD because they are routinely utilized,
association of decreased antioxidative capacity with GvHD has readily available, easily targeted for therapy, and have DAMPs and
been demonstrated [25]. Nagler et al. reported decreased UA antioxidant features. Although our study is promising given it is
levels and decreased antioxidative capacity in the saliva of GvHD the first report on this subject, comprehensive studies still need to
patients [26]. In addition, the association of hypouricemia with support our results.
some neurodegenerative and inflammatory diseases was also
observed [7, 27, 28]. We think that UA and Fib were the most
influential factors in our study demonstrating survival in GvHD
patients because they are both DAMP and antioxidant molecules. DATA AVAILABILITY
Endothelial dysfunction has a crucial role in the pathogenesis of The datasets generated during and/or analyzed during the current study are available
GvHD [12]. Some parameters indicating endothelial dysfunction in from the corresponding author on reasonable request. The Figshare DOI is https://
GvHD are associated with prognosis and survival [12, 29]. doi.org/10.6084/m9.figshare.22302763.

Bone Marrow Transplantation

 S. Çelik et al.
7
REFERENCES 26. Nagler R, Barness-Hadar L, Lieba M, Nagler A. Salivary antioxidant capacity in graft
1. Chao NJ. Clinical manifestations, diagnosis and grading of acute graft-versus- versus host disease. Cancer Investig. 2006;24:269–77.
host disease. Upto Date. 2013. https://www.uptodate.com/contents/clinical- 27. Auinger P, Kieburtz K, Mcdermott MP. The relationship between uric acid levels
manifestations-diagnosis-and-grading-of-acute-graft-versus-host-disease. and Huntington’s disease progression. Mov Disord. 2010;25:224–8.
2. Flowers ME, Inamoto Y, Carpenter PA, Lee SJ, Kiem H-P, Petersdorf EW, et al. 28. Andreadou E, Nikolaou C, Gournaras F, Rentzos M, Boufidou F, Tsoutsou A, et al.
Comparative analysis of risk factors for acute graft-versus-host disease and for Serum uric acid levels in patients with Parkinson’s disease: their relationship to
chronic graft-versus-host disease according to National Institutes of Health treatment and disease duration. Clin Neurol Neurosurg. 2009;111:724–8.
consensus criteria. Blood. 2011;117:3214–9. 29. Ali AM, DiPersio JF, Schroeder MA. The role of biomarkers in the diagnosis and
3. Ramadan A, Paczesny S. Various forms of tissue damage and danger signals risk stratification of acute graft-versus-host disease: a systematic review. Biol
following hematopoietic stem-cell transplantation. Front Immunol. 2015;6:14. Blood Marrow Transplant. 2016;22:1552–64.
4. Zeiser R, Blazar BR. Acute graft-versus-host disease—biologic process, prevention, 30. Luft T, Benner A, Terzer T, Jodele S, Dandoy CE, Storb R, et al. EASIX and mortality
and therapy. N Engl J Med. 2017;377:2167–79. after allogeneic stem cell transplantation. Bone Marrow Transplant.
5. Roh JS, Sohn DH. Damage-associated molecular patterns in inflammatory dis- 2020;55:553–61.
eases. Immune Netw. 2018;18:e27. 31. Lip G, Blann AD, Zarifis J, Beevers M, Lip P-L, Beevers DG. Soluble adhesion molecule
6. Braga TT, Forni MF, Correa-Costa M, Ramos RN, Barbuto JA, Branco P, et al. P-selectin and endothelial dysfunction in essential hypertension: implications for
Soluble uric acid activates the NLRP3 inflammasome. Sci Rep. 2017;7:1–14. atherogenesis? A preliminary report. J Hypertens. 1995;13:1674–8.
7. Ostendorf BN, Blau O, Uharek L, Blau IW, Penack O. Association between low uric 32. Nomura S, Ishii K, Inami N, Kimura Y, Uoshima N, Ishida H, et al. Evaluation of
acid levels and acute graft-versus-host disease. Ann Hematol. 2015;94:139–44. angiopoietins and cell-derived microparticles after stem cell transplantation. Biol
8. Ghasemi K, Parkhideh S, Kazemi MH, Salimi M, Salari S, Nalini R, et al. The role of Blood Marrow Transplant. 2008;14:766–74.
serum uric acid in the prediction of graft‐versus‐host disease in allogeneic 33. Paczesny S, Krijanovski OI, Braun TM, Choi SW, Clouthier SG, Kuick R, et al. A
hematopoietic stem cell transplantation. J Clin Lab Anal. 2020;34:e23271. biomarker panel for acute graft-versus-host disease. Blood. 2009;113:273–8.
9. Penack O, Peczynski C, van der Werf S, Finke J, Ganser A, Schoemans H, et al. 34. Hansen JA, Hanash SM, Tabellini L, Baik C, Lawler RL, Grogan BM, et al. A novel
Association of uric acid levels before start of conditioning with mortality after soluble form of Tim-3 associated with severe graft-versus-host disease. Biol Blood
allogeneic hematopoietic stem cell transplantation–a prospective, non- Marrow Transplant. 2013;19:1323–30.
interventional study of the EBMT Transplant Complication Working Party. Hae- 35. Srinagesh HK, Özbek U, Kapoor U, Ayuk F, Aziz M, Ben-David K, et al. The MAGIC
matologica. 2020;105:1977. algorithm probability is a validated response biomarker of treatment of acute
10. Health UDo, Services H. Common Terminology Criteria for Adverse Events graft-versus-host disease. Blood Adv. 2019;3:4034–42.
(CTCAE) version 5.0. Published November 27, 2017. 2021. https://ctep.cancer.gov/
protocoldevelopment/electronic_applications/docs/ctcae_v5_quick_reference_
5x7.pdf. ACKNOWLEDGEMENTS
11. Ruutu T, Barosi G, Benjamin RJ, Clark RE, George JN, Gratwohl A, et al. Diagnostic The authors thank all the patients who participated in this study and all the doctors,
criteria for hematopoietic stem cell transplant-associated microangiopathy: nurses, staff and patient relatives in the clinic for their hard work for the patients.
results of a consensus process by an International Working Group. Haematolo-
gica. 2007;92:95–100.
12. Luft T, Dreger P, Radujkovic A. Endothelial cell dysfunction: a key determinant for
the outcome of allogeneic stem cell transplantation. Bone Marrow Transplant. AUTHOR CONTRIBUTIONS
2021;56:2326–35. LK and SÇ designed the study, SÇ and ZTG performed the statistical analysis, LK and
13. Luft T, Benner A, Jodele S, Dandoy CE, Storb R, Gooley T, et al. EASIX in patients SÇ wrote the manuscript, KAD, SÇ, MK, and AÜ did the data managements and OK
with acute graft-versus-host disease: a retrospective cohort analysis. Lancet provided the pathological information. All authors reviewed and approved the final
Haematol. 2017;4:e414–23. version of the manuscript. Each Author has contributed substantially to the research,
14. Przepiorka D, Weisdorf D, Martin P, Klingemann H, Beatty P, Hows J, et al. 1994 preparation and production of the paper and approves of its submission to the Bone
Consensus conference on acute GVHD grading. Bone Marrow Transplant. Marrow Transplantation.
1995;15:825–8.
15. Nannya Y, Kataoka K, Hangaishi A, Imai Y, Takahashi T, Kurokawa M. The negative
impact of female donor/male recipient combination in allogeneic hematopoietic COMPETING INTERESTS
stem cell transplantation depends on disease risk. Transpl Int. 2011;24:469–76. The authors declare no competing interests.
16. Carreras E, Dufour C, Mohty M, Kröger N. The EBMT handbook: hematopoietic
stem cell transplantation and cellular therapies. Springer Nature, Cham, 2019.
17. Negrin RS. Preparative regimens for hematopoietic cell transplantation. UpTo-
Date. 2010;Topic:3557.
ETHICS APPROVAL AND CONSENT TO PARTICIPATE
All patients signed written informed consent. All ethical procedures and standards
18. Shi Y, Evans JE, Rock KL. Molecular identification of a danger signal that alerts the
were carried out following the 1975 Helsinki Declaration. The study was approved by
immune system to dying cells. Nature 2003;425:516–21.
Erciyes University Clinical Research Ethics Committee.
19. Wilhelm K, Ganesan J, Müller T, Dürr C, Grimm M, Beilhack A, et al. Graft-versus-
host disease is enhanced by extracellular ATP activating P2X 7 R. Nat Med.
2010;16:1434–8.
20. Zeiser R, Penack O, Holler E, Idzko M. Danger signals activating innate immunity ADDITIONAL INFORMATION
in graft-versus-host disease. J Mol Med. 2011;89:833–45. Supplementary information The online version contains supplementary material
21. Jankovic D, Ganesan J, Bscheider M, Stickel N, Weber FC, Guarda G, et al. The available at https://doi.org/10.1038/s41409-023-02145-7.
Nlrp3 inflammasome regulates acute graft-versus-host disease. J Exp Med.
2013;210:1899–910. Correspondence and requests for materials should be addressed to Serhat. Çelik.
22. Haen S, Eyb V, Mirza N, Naumann A, Peter A, Löffler M, et al. Uric acid as a novel
biomarker for bone-marrow function and incipient hematopoietic reconstitution Reprints and permission information is available at http://www.nature.com/
after aplasia in patients with hematologic malignancies. J Cancer Res Clin Oncol. reprints
2017;143:759–71.
23. Ames BN, Cathcart R, Schwiers E, Hochstein P. Uric acid provides an antioxidant Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims
defense in humans against oxidant-and radical-caused aging and cancer: a in published maps and institutional affiliations.
hypothesis. Proc Natl Acad Sci. 1981;78:6858–62.
24. Olinescu RM, Kummerow FA. Fibrinogen is an efficient antioxidant. J Nutr Bio-
chem. 2001;12:162–9. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to
25. Suh JH, Kanathezhath B, Shenvi S, Guo H, Zhou A, Tiwana A, et al. Thiol/redox this article under a publishing agreement with the author(s) or other rightsholder(s);
metabolomic profiling implicates GSH dysregulation in early experimental graft author self-archiving of the accepted manuscript version of this article is solely
versus host disease (GVHD). PloS One. 2014;9:e88868. governed by the terms of such publishing agreement and applicable law.

Bone Marrow Transplantation