Defibrotide inhibits donor leucocyte‐endothelial interactions and protects against acute graft‐versus‐host disease

Abstract Allogeneic hematopoietic stem cell transplantation (allo‐HCT) is an effective therapy for the treatment of high‐risk haematological malignant disorders and other life‐threatening haematological and genetic diseases. Acute graft‐versus‐host disease (aGvHD) remains the most frequent cause of non‐relapse mortality following allo‐HCT and limits its extensive clinical application. Current pharmacologic agents used for prophylaxis and treatment of aGvHD are not uniformly successful and have serious secondary side effects. Therefore, more effective and safe prophylaxis and therapy for aGvHD are an unmet clinical need. Defibrotide is a multi‐target drug successfully employed for prophylaxis and treatment of veno‐occlusive disease/sinusoidal obstruction syndrome. Recent preliminary clinical data have suggested some efficacy of defibrotide in the prevention of aGvHD after allo‐HCT. Using a fully MHC‐mismatched murine model of allo‐HCT, we report here that defibrotide, either in prophylaxis or treatment, is effective in preventing T cell and neutrophil infiltration and aGvHD‐associated tissue injury, thus reducing aGvHD incidence and severity, with significantly improved survival after allo‐HCT. Moreover, we performed in vitro mechanistic studies using human cells revealing that defibrotide inhibits leucocyte‐endothelial interactions by down‐regulating expression of key endothelial adhesion molecules involved in leucocyte trafficking. Together, these findings provide evidence that defibrotide may represent an effective and safe clinical alternative for both prophylaxis and treatment of aGvHD after allo‐HCT, paving the way for new therapeutic approaches.


| INTRODUC TI ON
Acute graft-versus-host disease (aGvHD) is the most frequent life-threatening complication after allogeneic hematopoietic stem cell transplantation (allo-HCT). The pathobiology of aGVHD results from immunocompetent allogeneic T cells contained in the marrow graft that recognize the recipient as foreign and migrate and attack specific host tissues (typically the skin, gastrointestinal tract and liver), resulting in significant immune/cytokine-mediated tissue injury, endothelial dysfunction, and high morbidity and mortality. [1][2][3][4][5] Different prophylactic and treatment regimens are currently employed to reduce the incidence and severity of aGvHD, most commonly based on corticosteroids, calcineurin inhibitors or other immunosuppressants. However, responses are unsatisfactory and aGvHD remains a significant problem. Thus, there is an unmet need for the development of newer strategies to prevent and treat this disease. 6 Defibrotide is a polydisperse mixture of 90% single-stranded and 10% double-stranded phosphodiester oligonucleotides derived from controlled depolymerization of porcine intestinal mucosal DNA that has proved to be effective for prophylaxis and treatment of hepatic veno-occlusive disease/sinusoidal obstruction syndrome (VOD/ SOS) following allo-HCT. 7,8 It has been proposed to have a protective effect on activated endothelial cells, [9][10][11] and among others, pro-fibrinolytic, anti-thrombotic and anti-inflammatory activities have been reported. 12 Defibrotide has been found to suppress heparanase gene expression, whose high levels have been postulated as a risk factor for aGvHD development, 13,14 and it was initially identified as an adenosine receptor agonist, whose activation inhibits T cell function. 10,15,16 Importantly, preliminary evidence in recent clinical trials employing defibrotide for VOD suggests that it could also have a prophylactic effect on the occurrence of aGvHD after allo-HCT: patients who received VOD prophylaxis with defibrotide had also a lower incidence and severity of aGvHD than untreated control group of patients, and it also did not seem to interfere with a graft-versus-leukaemia effect. 17,18 Based on these indirect data, defibrotide received an orphan designation for the prevention of aGvHD. However, its efficacy for the prevention and treatment of aGvHD as the only drug without the administration of other immunosuppressant/s, as well as the molecular mechanisms mediating its aGvHD-protective effect, has not been previously addressed.
Given such prior findings, we sought the mechanistic basis of defibrotide effects on evolving aGvHD. Here, using a murine model of fully major histocompatibility complex (MHC)-mismatched allo-HCT to induce fulminant aGvHD, we report that administration of defibrotide with same total daily dose employed for VOD in humans is also effective for both prophylaxis and treatment of murine aGvHD. This effect is mediated via a significant early inhibition of donor alloimmune effector T cell and neutrophil tissue infiltration, a substantial decrease in aGvHD-associated tissue injury, and a reversal of plasma pro-inflammatory/anti-inflammatory cytokines ratios.
These effects yield a significant increased survival. Mechanistic in vitro studies using human samples reveal that defibrotide inhibits peripheral blood leucocyte transendothelial migration by down-regulating expression of key endothelial adhesion molecules involved in leucocyte trafficking such as E-selectin, P-selectin, ICAM-1 and VCAM-1. Collectively, these findings provide novel perspectives on the clinical applicability of this multi-target drug in aGvHD therapy following allo-HCT.
Housing was maintained under specific pathogen-free conditions in the animal facilities of the University of Murcia. All animal protocols were approved by the Institutional Animal Care and Use Committee at University of Murcia (protocol A13150201).

| Induction and assessment of murine aGvHD and histopathology analysis
A model of fully MHC-mismatched allo-HCT was carried out by transplanting bone marrow mononuclear cells from donor BALB/c mice into recipient C57BL/6J mice that were previously irradiated by with a lethal dose of 10 Gy split into two equal doses of 5 Gy and 24 hours apart (days −1 and +0). To induce aGvHD, on day +0 recipient C57BL/6J mice were injected intravenously with 1 × 10 7 bone marrow mononuclear cells from donor BALB/c mice and 1.5 × 10 7 BALB/c donor splenocytes. Mice survival post-transplant was monitored daily, and clinical aGvHD was evaluated using a scoring evidence that defibrotide may represent an effective and safe clinical alternative for both prophylaxis and treatment of aGvHD after allo-HCT, paving the way for new therapeutic approaches.

K E Y W O R D S
acute GvHD, defibrotide, hematopoietic stem cell transplantation system that generates a composite aGvHD score composed of individual scores for weight loss, posture, activity, skin integrity and fur texture. 19 Histopathology analysis of aGvHD was performed in different sections of liver, intestine (colon), skin and oral mucosa (tongue) collected from the recipients 10 days after allo-HCT and analysed as described previously 19,20 by a single pathologist blinded to the treatment groups. Samples from the different organs were fixed in formalin and embedded in paraffin. After, three-micrometre sections were stained with haematoxylin and eosin stain for histopathologic examination. In brief, the skin histopathologic lesions were grade 0 (normal), grade I (slight vacuolar degeneration of epidermal basal cells), grade II (scattered individual apoptotic epidermal basal cells, spongiosis), grade III (separation of dermoepidermal junction) and grade IV (diffuse and severe ulceration, with extensive destruction of epidermis). Colon aGvHD scores were either grade 0 (normal), grade I (scattered individual apoptotic cells, with inflammatory cell infiltrate), grade II (crypt epithelial cell apoptosis, with villous blunting and/or exploding crypts), grade III (focal mucosal ulceration, with moderate villous atrophy) and grade IV (diffuse and severe mucosal ulceration). Liver aGvHD was scored according to degree of bile ducts affected: 0 (normal), grade I (≤25% bile ducts affected), grade II (25%-49% bile ducts affected), grade III (50%-74% bile ducts affected) and grade IV (≥75% bile ducts affected). The scoring system used for histopathologic evaluation of the tongue samples was based on intraepithelial inflammatory infiltrate as follows: grade 0 (absence of inflammatory infiltrate), grade 1 (weak), grade 2 (moderate), grade 3 (intense) and grade 4 (severe). The degree of T-lymphocyte infiltrate or induction of apoptosis was evaluated by immunohistochemical labelling by using a commercial kit (EnVision Flex, Dako-Agilent Technologies, Barcelona, Spain). Briefly, tissue sections were submitted to deparaffinization and rehydration, followed by antigen demasking procedure and endogenous peroxidase inhibition. After, sections were incubated with primary antibodies (anti-T-CD3, Dako) for 1 hour at 37ºC, following by secondary anti-rabbit labelled polymer incubation for 20 minutes at 37ºC. Immunoreaction was finally revealed by incubation with 3-3´diaminobenzidine (Sigma-Aldrich, St. Louis, MO). Positive immunoreaction was identified as a darkbrown pericellular staining. In addition, polymorphonuclear (PMN) neutrophils were identified on H&E-stained sections according to morphologically evident segmented nuclei. All analyses were performed by using a standard direct-light microscope (Axio A10, Carl Zeiss, Barcelona, Spain).
Untreated controls were injected daily with same volume of saline solution without defibrotide.

| Determination of plasmatic cytokines by enzyme-linked immunosorbent assays (ELISA)
Murine IFNγ, TNFα, IL-6, IL-12, TGFβ and IL-10 were measured in blood plasma by ELISA (Diaclone, bioNova Cientifica, Madrid, Spain; and Elabscience, Bethesda, MD). All experimental samples and standards were assayed in triplicate according to the manufacturer's instructions. Human peripheral blood mononuclear cells (MNCs) were isolated using Ficoll-Paque density gradient centrifugation from heparinized blood samples obtained from healthy volunteers after informed consent. In some experiments, pooled sera samples obtained from three different patients with aGvHD on the day of diagnosis, or from same number of healthy donors, were used.

Hospital Virgen de la Arrixaca and Hospital Clinic de Barcelona
Ethics Committees approved the protocols used to obtain and process all human samples.

| Parallel-plate flow chamber adhesion assays
The effects of defibrotide on leucocyte adhesion on endothelial cells were analysed using a parallel-plate perfusion chamber. SK-Hep1 monolayers were treated previously with 100 µg/mL defibrotide for 24 hours or left untreated and then grown for 48 hours in medium supplemented with 20% pooled sera of aGvHD patients (n = 15, Table 1) or of healthy donors (n = 15). Then, SK-Hep1 cells were perfused with control citrated blood from a healthy donor at a shear rate of 300 s −1 for 10 minutes. After perfusion, coverslips were washed with PBS and fixed in 3% paraformaldehyde in PBS.

| Transwell migration assays and flow cytometry analysis
HUVEC in culture were exposed to medium containing 20% pooled sera of aGvHD patients (n = 4, Table 1

| Statistical analysis
Results are expressed as mean ± standard deviation. Student's paired t test or one-way ANOVA followed by Bonferroni's post hoc comparison tests was used for statistical comparisons between groups.
Survival curves were plotted using Kaplan-Meier estimates and statistically analysed using the Mantel-Cox log-rank test. P values < .05 were considered statistically significant.

| Defibrotide aGVHD treatment reduces aGvHD clinical manifestations and reduces level of donor T cell and PMN neutrophil infiltrates in aGvHD-target organs
To test the hypothesis that defibrotide has direct effects in aGvHD onset and/or evolution, mice with ongoing aGvHD (from day +7 after allo-HCT) were treated with a daily i/p infusion of 25 mg/ kg of defibrotide for 20 days, following the same total daily dose used in human VOD/SOS patients (schematic diagram shown in Figure 1A) showed a rapid severe aGvHD occurrence, which caused the early death in all mice ( Figure 1C). However, daily administration of defibrotide showed a significantly lower aGvHD-related mortality than untreated mice group (P < .001) (60% survival in defibrotide group versus 0% survival in untreated group at day +60 post-transplantation, P < .001) ( Figure 1C). Also, defibrotide-treated mice significantly improved their clinical aGvHD score compared to untreated mice group, being significantly lower from the onset of the disease (P < .001) ( Figure 1D). To evaluate whether defibrotide had effects in the prophylaxis of aGVHD, mice were treated with a daily intraperitoneal infusion of 25 mg/kg of defibrotide from 2 days before allo-HCT to 17 days after (schematic diagram shown in Figure 1B).
Remarkably, mice receiving daily prophylaxis with defibrotide dis- In addition, the tongue was also analysed as a representative tissue of oral mucosa, which is also affected during aGvHD progression (Figure 2). Tongue dermis of untreated animals displayed a weak or moderate inflammatory infiltrate (grade 0-I), which was absent in defibrotide-treated mice.
These observations suggest that the improved survival observed in the group of animals receiving defibrotide was correlated with a substantially reduced degree of tissue-specific destruction.
Tissue damage in aGvHD-target organs is largely related with the ability of alloreactive donor T cells to traffic and infiltrate specifically these tissues. [21][22][23] Thus, we next investigated levels of T-CD3 + cell infiltrates in skin, liver, colon and tongue of untreated animals as compared to animals receiving prophylactic defibrotide at day +10 after allo-HCT. Significantly, we observed that livers, colon and tongues of untreated allo-HCT recipients displayed significant higher levels of T-CD3+ infiltrates than BM transplanted or syn-HCT animals. However allo-HCT mice that received defibrotide showed a substantial decrease in T-CD3 + infiltrates compared to their untreated counterparts, mainly in the colon (P < .001) ( Figure 3A,B). On the other hand, skin sections of mice among the two groups of animals only displayed scattered infiltrates of donor T cells, which correlated with no observable lesions in this tissue at this time post-transplantation.
Apart from tissue trafficking of T cells, it has been previously described that neutrophils can also infiltrate the aGvHD-target organs, being considered as a bad prognostic marker of early allo-HCT-associated mortality. 24 As such, we also examined the

| Prophylactic administration of defibrotide reverses ratios of plasma pro-inflammatory and antiinflammatory cytokines
To further evaluate the effects of defibrotide administration on aGvHD development, we analysed the release of several proinflammatory cytokines described to have key roles during the effector phase of aGVHD. 25 Mice with ongoing aGvHD after prophylactic daily administration induced a marked decrease in these pro-inflammatory mediators (P < .01, P < .001) and significant increased levels of the anti-inflammatory cytokines TGFβ and IL-10 (P < .01) (Figure 4). Absolute T-CD3 + cell or PMN neutrophil counts are presented as mean ± SD per high-power field from counts relative to 10 high-power fields (magnification ×200 or ×400, respectively). T-CD3 + cell or PMN neutrophil infiltrates were significantly increased compared to syn-HCT group, **P < .01, ***P < .001, or decreased compared to untreated allo-HCT group, ∆∆∆ P < .001, respectively

| Defibrotide impairs human leucocyte trafficking and down-regulates endothelial adhesion molecules expression: mechanism of action involved in the protective effect of defibrotide.
Acute GvHD is associated with the up-regulation of several endothelial cell adhesion molecules such as E-selectin, P-selectin, VCAM-1 and ICAM-1 in the affected organs in response to the effect of proinflammatory cytokines, favouring donor alloreactive leucocytes to engage the endothelium and extravasate to the target tissue(s). [26][27][28] In order to extrapolate the findings obtained in the aGvHD murine model to the human disease, we first analysed whether defibrotide interferes with peripheral blood MNC or T cell adhesion and migration on endothelial HUVEC and SK-Hep1 cell lines. As previously reported, 11 defibrotide does not interact with peripheral blood MNCs, and when leucocytes were directly incubated with defibrotide, leucocyte-endothelial rolling adhesive interactions under hemodynamic flow conditions were not affected (data not shown). However, subsequent Transwell migration assays showed that when HUVEC were exposed to aGvHD sera and treated with defibrotide afterwards ( Figure 5A) or before ( Figure 5B

| D ISCUSS I ON
Allogeneic HCT provides a replacement of the diseased marrow with healthy hematopoietic progenitor cells derived from the donor.
Moreover, allo-HCT is an important type of adoptive cell immunotherapy in which donor alloreactive immune cells confer a potent graft-versus-malignancy effect. However, it can also induce immunopathological processes such as aGvHD, which is related to the immunologic recognition and destruction of specific host tissues that notably increase the mortality of the procedure. 29 In the present study, we provide evidence on the therapeutic efficacy of defibrotide in prevention and treatment of aGvHD after allo-HCT. Although two preliminary clinical studies have suggested indirect evidence that defibrotide prophylaxis might decrease the risk of aGvHD in paediatric and adult patients receiving allo-HCT, 17,18 this is the first comprehensive preclinical study demonstrating that administration of defibrotide as a single agent, without other concomitant treatment, is sufficient to effectively prevent and/or treat this immunopathological complication.
Defibrotide has been previously described as a multi-target drug, showing anti-thrombotic/thrombolytic, anti-ischaemic, pro-fibrinolytic and anti-angiogenic effects. [30][31][32][33][34] But defibrotide has also demonstrated important anti-inflammatory properties and a protective effect on endothelial cells from HCT conditioning. 12,35,36 These findings and the fact that aGvHD occurrence is associated with higher activation and dysfunction of endothelial cells 37 led us to hypothesize that defibrotide may be an effective agent for aGvHD prevention and/or treatment.
Our findings indicate that daily infusions of defibrotide employing the same total daily dose as in human VOD/SOS patients, given prior to allo-HCT (prophylaxis) or once aGvHD has already devel- that an abundance neutrophil infiltrates in host tissues contribute synergistically to increase aGvHD severity and related mortality by F I G U R E 5 Defibrotide inhibits human peripheral blood MNC and T cell migration and down-regulates endothelial adhesion molecules expression. A, HUVEC were exposed to culture medium containing pooled sera of aGvHD patients or from healthy donors (control) and treated with defibrotide, or (B) incubated with defibrotide before adding sera (prophylaxis). Then, peripheral blood MNC and T-CD3 cell transendothelial migration towards aGvHD or control sera were analysed in Transwell assays in the presence or absence ('untreated', ie UNT) of defibrotide. Data are expressed as percentages of migrated cells related to the total number of MNCs or T-CD3 cells added to the upper chamber. Migration was significantly increased in the aGvHD condition, *** P < .001, or significantly inhibited after defibrotide incubation, ∆∆∆ P < .001, respectively. C, Also, effect of defibrotide on the MNC or T-CD3 cell rolling and firm adhesion on SK-Hep1 cells was analysed using a parallel-plate perfusion chamber under dynamic flow conditions. Adhesion was significantly increased in the aGvHD condition, ** P < .01, *** P < .001, and significantly inhibited after defibrotide incubation, ∆ P < .05, ∆∆ P < .01, respectively. D, Representative micrographs showing adherent leucocytes (green staining with Alexa488-conjugated anti-human CD45, red staining with PE-conjugated anti-CD3 and blue staining with DAPI) obtained in the different experimental conditions are shown (magnification ×40). After, expression of endothelial adhesion molecules on HUVEC exposed to pooled aGvHD sera with DF or without prophylactic defibrotide (UNT) was analysed by flow cytometry. E, Representative histograms showing E-selectin, P-selectin, VCAM-1 and ICAM-1 expression on HUVEC in the different experimental conditions are shown. F, Relative expression of the different adhesion molecules was normalized to untreated HUVEC (ie w/o prophylactic defibrotide) incubated with healthy donor pooled sera (UNT control). Expression was up-regulated in the aGvHD condition compared to UNT control, *** P < .001, or down-regulated compared to UNT aGvHD, ∆∆ P < .01, ∆∆∆ P < .001, respectively production of reactive oxygen species that leads to T cell activation and subsequent potentiation of the tissue damage, mainly in the gastrointestinal tract. 24,42 In this study, we also found that similarly to what was observed with T cells, mice treated with prophylactic infusions of defibrotide showed markedly lower neutrophils infiltrates than allo-HCT untreated mice.
The pathogenesis of aGvHD is closely linked to cytokines produced by T cells and other immune cells that infiltrate aGvHD-target organs. 43,44 Here, we found that mice receiving prophylactic In summary, our findings indicate that defibrotide, used as a concomitant prophylactic or therapeutic drug to the usual standard immunosuppressants employed for aGvHD, is effective in ameliorating the inflammatory response and tissue damage associated with this immunopathological disease, reducing its incidence and severity and significantly improving survival after allo-HCT.
In addition, defibrotide is not associated with relevant secondary side effects associated with other established treatments such as with corticosteroids. Therefore, defibrotide might represent a promising option for the prophylaxis and treatment of aGvHD in patients receiving allo-HCT.

ACK N OWLED G M ENTS
This study has been funded by

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding authors upon reasonable request.