Abstract 33: Platelet 12-lipoxygenase is a Key Regulator of Platelet Reactivity and Thrombus Formation in vivo

2016 ◽  
Vol 36 (suppl_1) ◽  
Author(s):  
Reheman Adili ◽  
Katherine Mast ◽  
Theodore R Holman ◽  
Michael Holinstat

Background: Platelet reactivity is required to maintain hemostasis, however high platelet reactivity leads to thrombus formation, myocardial infarction, and stroke. Platelet 12-lipoxygenase (12-LOX) has been demonstrated by our lab and others to regulate agonist-mediated platelet reactivity suggesting a role for 12-LOX in regulation of in vivo thrombosis. The ability to target 12-LOX in vivo has not been established to date. Therefore, we sought to determine if 12-LOX regulates platelet reactivity and thrombus formation in vivo using the selective 12-LOX inhibitor ML355 to determine whether platelet 12-LOX is an effective target for anti-platelet therapeutics. Methods: ML355 effects on human platelet function was assessed in vitro by platelet aggregometry, ex vivo by perfusion chamber, and in vivo by thrombus formation and vessel occlusion in small and large vessels in 12-LOX -/- , WT mice, and mice treated with ML355 via intravital microscopy using the FeCl 3 and laser injury models. Results: In in vitro platelet aggregation, ML355 dose-dependently inhibited agonist-induced aggregation. In ex vivo flow chamber assays, platelet adhesion and thrombus formation on collagen-coated surfaces at high shear was attenuated in both mouse and human whole blood after incubation with ML355. Further, platelet aggregation and thrombus growth in 12-LOX -/- mice were impaired in both laser and FeCl 3 -induced mesenteric, carotid artery and cremaster arteriole thrombosis models. Thrombi in 12-LOX -/- mice were unstable and frequently formed emboli, which resulted in impaired vessel occlusion or reopening. Additionally, thrombus formation and vessel occlusion was impaired in ML355 treated WT mice. Conclusions: The 12-LOX inhibitor ML355 inhibits platelet aggregation induced by a number of platelet agonists. Ex vivo high shear conditions in both mice and human was attenuated in the presence of ML355. Thrombus formation and vessel occlusion were impaired in mice deficient in 12-LOX. Finally, ML355 attenuates thrombus formation and prevents vessel occlusion in vivo . Our data strongly indicates 12-LOX is an important determinant of platelet reactivity and inhibition of platelet 12-LOX may represent a new target for anti-platelet therapeutics.

Blood ◽  
2015 ◽  
Vol 126 (23) ◽  
pp. 3442-3442 ◽  
Author(s):  
Reheman Adili ◽  
Theodore R Holman ◽  
Michael Holinstat

Abstract Background: Adequate platelet reactivity is required for platelet adhesion and aggregation at the site of vascular injury to maintain hemostasis. However, excessive platelet reactivity can also lead to the formation of occlusive thrombi, the predominate underlying cause of myocardial infarction and stroke. While current anti-platelet treatments limit platelet function, they often result in an increased risk of bleeding. 12-lipoxygenase (12-LOX), an oxygenase highly expressed in the platelet, has been demonstrated by our lab and others to regulate PAR4 and GPVI-mediated platelet reactivity suggesting a role of 12-LOX in regulation of vivo thrombosis. However, the ability to pharmacologically target 12-LOX in vivo has not been established to date. Aims: To determine how 12-LOX regulates thrombus formation in vivo and whether platelet 12-LOX is an effective target for anti-platelet therapeutics, wild-type (WT) or 12-LOX deficient (12-LOX-/-) mice were treated with or without the 12-LOX inhibitor, ML355, and were assessed for inhibitory effects on platelet activation in vitro, ex-vivo and in vivo. Methods: The effect of the novel 12-LOX inhibitor ML355 on human platelet function was assessed in vitro by platelet aggregometry, ex vivo by perfusion chamber. In vivo thrombus formation and vessel occlusion in small and large vessels were studied in 12-LOX-/-, WT mice and mice treated with ML355 using intravital microscopy using the FeCl3 injury models. Results: Using in vitro platelet aggregation assays, ML355 dose dependently inhibited thrombin, PAR1-AP, and PAR4-AP-induced aggregation in washed human platelets. Interestingly, the negative regulatory effects of ML355 inhibition of 12-LOX can be overcome by high concentration of thrombin. Additionally, ML355 was able to attenuate ADP-induced platelet aggregation both in platelet-rich-plasma and whole blood. In ex vivo flow chamber assays, platelet adhesion and thrombus formation on collagen-coated surfaces at high shear was attenuated in both mouse and human whole blood after incubation with ML355. Further, platelet aggregation and thrombus growth in 12-LOX-/- mice was impaired in FeCl3-induced mesenteric or carotid artery thrombosis models. Thrombi in 12-LOX-/- mice were unstable and frequently form emboli, which resulted in impaired vessel occlusion or reopening. Additionally, thrombus formation and vessel occlusion was impaired in ML355 treated WT mice. Conclusions: The highly selective 12-LOX inhibitor ML355 inhibits platelets aggregation induced by various platelet agonists and ML355 inhibition of platelet function is not agonist specific. Platelet function at high shear in ex vivo conditions in both mice and human was attenuated in the presence of ML355. Thrombus growth, stability, and vessel occlusion was impaired in mice deficient for 12-LOX. Finally, the highly selective 12-LOX inhibitor ML355 attenuates thrombus formation and prevents vessel occlusion in vivo. Our data strongly indicates 12- LOX is an important determinant of platelet reactivity and inhibition of platelet 12-LOX may represent a new target for anti-platelet therapeutics. Disclosures No relevant conflicts of interest to declare.


2014 ◽  
Vol 111 (02) ◽  
pp. 279-289 ◽  
Author(s):  
Yan Hou ◽  
Hui Zhou ◽  
Yiming Wang ◽  
Alexandra Marshall ◽  
Chaofan Liang ◽  
...  

SummaryPlatelet adhesion and aggregation at the sites of vascular injury are key events for thrombosis and haemostasis. It has been well demonstrated that interaction between glycoprotein (GP) Ib and von Willebrand factor (VWF) initiates platelet adhesion and contributes to platelet aggregation, particularly at high shear. GPIb has long been suggested as a desirable antithrombotic target, but anti-GPIb therapy has never been successfully developed. Here, we evaluated the antithrombotic potential of Anfibatide, a novel snake venom-derived GPIb antagonist. We found Anfibatide inhibited washed murine platelet aggregation induced by ristocetin and recombinant murine VWF. It also blocked botrocetin-induced binding of murine plasma VWF to recombinant human GPIb . Interestingly, Anfibatide did not inhibit botrocetin- induced aggregation of platelet-rich plasma, indicating that its binding site may differ from other snake venom-derived GPIb antagonists. Anfibatide strongly inhibited platelet adhesion, aggregation, and thrombus formation in perfusion chambers at high shear conditions and efficiently dissolved preformed thrombi. Anfibatide also inhibited thrombus growth at low shear conditions, though less than at high shear. Using intravital microscopy, we found that Anfibatide markedly inhibited thrombosis in laser-injured cremaster vessels and prevented vessel occlusion in FeCl3-injured mesenteric vessels. Importantly, Anfibatide further inhibited residual thrombosis in VWF-deficient mice, suggesting that Anfibatide has additional antithrombotic effect beyond its inhibitory role in GPIb-VWF interaction. Anfibatide did not significantly cause platelet activation, prolong tail bleeding time, or cause bleeding diathesis in mice. Thus, consistent with the data from an ongoing clinical trial, the data from this study suggests that Anfibatide is a potent and safe antithrombotic agent.


2017 ◽  
Vol 37 (suppl_1) ◽  
Author(s):  
Reheman Adili ◽  
Katherine Mast ◽  
Michael Holinstat

12-lipoxygenase (12-LOX) has been demonstrated to regulate platelet function, hemostasis, and thrombosis ex vivo , supporting a key role for 12-LOX in regulation of in vivo thrombosis. While pharmacologically targeting 12-LOX in vivo has been a challenge to date, the recent development of the 12-LOX selective inhibitor, ML355, as an effective antiplatelet therapeutic in vivo was assessed. ML355 potently inhibited thrombin and other agonist-induced platelet aggregation ex vivo in washed human platelets and inhibited downstream oxylipin production of platelet 12-LOX as confirmed by Mass spectrometry analysis. Ex vivo flow chamber assays confirmed that human platelet adhesion and thrombus formation at arterial shear over collagen was attenuated in human whole blood treated with ML355 to a greater extent compared to aspirin. In vivo , PK assessment of ML355 showed reasonable 12-LOX plasma levels 12 hours following administration of ML355. FeCl 3 -induced injury of the mesenteric arterioles resulted in less stable thrombi in 12-LOX -/- mice and ML355-treated WT mice resulting in impairment of vessel occlusion. Additionally, ML355 dose-dependently inhibited laser-induced thrombus formation in the cremaster arteriole thrombosis model in WT, but not in 12-LOX -/- mice. Importantly, hemostatic plug formation and bleeding following treatment with ML355 were not affected in response to laser ablation on the saphenous vein or in a cremaster microvasculature laser-induced rupture model. Our data strongly supports 12-LOX as a key determinant of platelet reactivity in vivo and inhibition of platelet 12-LOX with ML355 may represent a new class of antiplatelet therapeutics.


Blood ◽  
2011 ◽  
Vol 118 (21) ◽  
pp. 3361-3361
Author(s):  
Riitta Lassila ◽  
Annukka Jouppila ◽  
Ulla M Marzec ◽  
Stephen R Hanson

Abstract Abstract 3361 We have developed a semi-synthetic antithrombotic heparin complex, APL001, to mimic mast cell-derived natural heparin proteoglycans (HepPG). HepPG attenuate platelet-collagen interactions under blood flow by inhibiting VWF- and GPIIb/IIIa -mediated platelet aggregation. In addition, rat-derived HepPG arrest platelet thrombus growth on collagen surfaces or at vascular injury sites, both in vitro and in vivo (Lassila et al.ATVB 1997, Kauhanen et al. ATVB 2000, Olsson et al. Thromb Haemost 2002). Our objective was to study the inhibitory capacity of APL001 for preventing human platelet aggregation in vitro and acute thrombosis in a baboon model in vivo. The effects of unfractionated heparin (UFH) and APL001 were compared in relevant coagulation assays (APTT, PT, thrombin time, anti-FXa activity, fibrinogen, FVIII:C and VWF activity (VWF:RCo) and antigen). Additionally, agonist-induced (collagen, ristocetin and ADP) platelet aggregation in citrate or hirudin-anticoagulated whole blood (Multiplate®) (n=10 healthy subjects), and platelet function analysis (PFA100®) in citrated platelet rich plasma (PRP) were assessed. In a well-established baboon thrombosis model a collagen-coated PTFE graft (length 2 cm, lumen 4 mm) was placed in an arterio-venous shunt. Prior to blood contact the thrombogenic surface was treated for 10 min with UFH or APL001 (both at 4 mg/mL). Thrombus formation was initiated by exposing the surface to blood flow (100 mL/min, shear rate 265−1), and the deposition of 111-In-labeled platelets and of fibrin was quantified continuously over 1h. Fibrin thrombus accumulation was assessed from the incorporation of circulating 125-I-fibrinogen. In the heparin-relevant coagulation tests APL001 was comparable or 20–30% more potent than UFH while FVIII, fibrinogen and VWF variables remained unaltered. In contrast to UFH, APL001 (300 μg/mL) consistently inhibited collagen- and ristocetin-induced platelet aggregation, whereas UFH had only a modest effect in comparison with PBS control (Table). ADP-induced aggregation was unaffected. Comparable results were observed in the PRP aggregation assay. PFA100 testing also demonstrated inhibitory effects. In the in vivo thrombosis model (n=4) APL001 reduced platelet deposition on collagen (vs. the results with UFH) by 34% (p=0.01), while platelet accumulation in distal propagated thrombus was reduced by 61% (p=0.16). APL001-treated surfaces accumulated 45% less fibrin than the UFH-treated surfaces (p=0.008). In conclusion, when compared with UFH APL001 inhibited both collagen- and ristocetin-induced platelet aggregation in human blood, while anticoagulant properties were comparable. In the absence of systemic antithrombotic drugs, exposure of APL001 to a highly thrombogenic collagen surface arrested thrombus formation in an in vivo baboon model. This finding suggests that locally administered APL001 alone, due to its dual antiplatelet and anticoagulant effects, may limit the growth and size of thrombus and thereby prevent subsequent thrombo-occlusion.TableAnticoagulantInhibition-% of platelet aggregation ± SDConc. 300 μg/mLnColl (3.2 μg/mL)Ristocetin (0.77 mg/mL)ADP (6.4 μM)CitrateAPL0011033 ± 1543 ± 166 ± 24UFH1011 ± 1323 ± 153 ± 7p value0.0030.0100.700HirudinAPL0011032 ± 1043 ± 178 ± 10UFH108 ± 1116 ± 166 ± 9p value0.0000.0020.600 Disclosures: Lassila: Aplagon: Chief Scientific Advisor.


1989 ◽  
Vol 67 (9) ◽  
pp. 989-993 ◽  
Author(s):  
A. W. Ford-Hutchinson ◽  
Y. Girard ◽  
A. Lord ◽  
T. R. Jones ◽  
M. Cirino ◽  
...  

L-670,596 ((−)6,8-difluoro-9-p-methylsulfonyl benzyl-1,2,3,4-tetrahydrocarbazol-1-yl-acetic acid) has been shown to be a potent receptor antagonist as evidenced by the inhibition of the binding of 125I-labeled PTA-OH to human platelets (IC50, 5.5 × 10−9 M), inhibition of U-44069 induced aggregation of human platelet rich plasma (IC50, 1.1 × 10−7 M), and competitive inhibition of contractions of the guinea pig tracheal chain induced by U-44069 (pA2,9.0). The compound was also active in vivo as shown by inhibition of arachidonic acid and U-44069 induced bronchoconstriction in the guinea pig (ED50 values, 0.04 and 0.03 mg/kg i.v., respectively), U-44069 induced renal vasoconstriction in the pig (ED50, 0.02 mg/kg i.v.), and inhibition of ex vivo aggregation of rhesus monkey platelets to U-44069 (active 1–5 mg/kg p.o.). The selectivity of the compound was indicated by the failure to inhibit, first, ADP-induced human or primate platelet aggregation and, second, bronchoconstriction in the guinea pig in vivo and contraction of the guinea pig tracheal chain in vitro to a variety of agonists. It is concluded that L-670,596 is a potent, selective, orally active thromboxane A2/prostaglandin endoperoxide receptor antagonist.Key words: thromboxane A2, thromboxane antagonist, prostaglandin endoperoxides, platelet aggregation.


1995 ◽  
Vol 73 (02) ◽  
pp. 318-323 ◽  
Author(s):  
K Azzam ◽  
L I Garfinkel ◽  
C Bal dit Sollier ◽  
M Cisse Thiam ◽  
L Drouet

SummaryTo assess the antithrombotic effectiveness of blocking the platelet glycoprotein (GP) Ib/IX receptor for von Willebrand factor (vWF), the antiaggregating and antithrombotic effects were studied in guinea pigs using a recombinant fragment of vWF, Leu 504-Lys 728 with a single intrachain disulfide bond linking residues Cys 509-Cys 695. The inhibitory effect of this peptide, named VCL, was tested in vitro on ristocetin- and botrocetin-induced platelet aggregation and compared to the ADP-induced platelet aggregation. In vivo, the antithrombotic effect of VCL was tested in a model of laser-injured mesentery small arteries and correlated to the ex vivo ristocetin-induced platelet aggregation. In this model of laser-induced thrombus formation, five mesenteric arteries were studied in each animal, and the number of recurrent thrombi during 15 min, the time to visualization and time to formation of first thrombus were recorded.In vitro, VCL totally abolished ristocetin- and botrocetin-induced platelet aggregation, but had no effect on ADP-induced platelet aggregation. Ex vivo, VCL (0.5 to 2 mg/kg) administered as a bolus i. v. injection inhibits ristocetin-induced platelet aggregation with a duration of action exceeding 1 h. The maximum inhibition was observed 5 min after injection of VCL and was dose related. The same doses of VCL had no significant effect on platelet count and bleeding time. In vivo, VCL (0.5 to 2 mg/kg) had no effect on the appearance of the thrombi formed but produced dose-dependent inhibition of the mean number of recurrent thrombi (the maximal effect was obtained at 5 min following i. v. injection of the highest dose: 0.8 ± 0.2 thrombi versus 4 ± 0.4 thrombi in controls). The three doses of VCL increased the time in which the first thrombus in a concentration-dependent manner was formed. However, the time to visualize the first thrombus was only prolonged in the higher dose-treated group.These in-vivo studies confirm that VCL induces immediate, potent, and transient antithrombotic effects. Most importantly, this inhibition was achieved without inducing thrombocytopenia nor prolongation of the bleeding time.


1998 ◽  
Vol 79 (01) ◽  
pp. 222-227 ◽  
Author(s):  
F. Stockmans ◽  
W. Deberdt ◽  
Å. Nyström ◽  
E. Nyström ◽  
J. M. Stassen ◽  
...  

SummaryIntravenous administration of piracetam to hamsters reduced the formation of a platelet-rich venous thrombus induced by a standardised crush injury, in a dose-dependent fashion with an IC50 of 68 ± 8 mg/kg. 200 mg/kg piracetam also significantly reduced in vivo thrombus formation in rats. However, in vitro aggregation of rat platelets was only inhibited with piracetam-concentrations at least 10-fold higher than plasma concentrations (6.2 ± 1.1 mM) obtained in the treated animals. No effects were seen on clotting tests.In vitro human platelet aggregation, induced by a variety of agonists, was inhibited by piracetam, with IC50’s of 25-60 mM. The broad inhibition spectrum could be explained by the capacity of piracetam to prevent fibrinogen binding to activated human platelets. Ex vivo aggregations and bleeding times were only minimally affected after administration of 400 mg/kg piracetam i.v. to healthy male volunteers, resulting in peak plasma levels of 5.8 ± 0.3 mM.A possible antiplatelet effect of piracetam could be due to the documented beneficial effect on red blood cell deformability leading to a putative reduction of ADP release by damaged erythrocytes. However similarly high concentrations were needed to prevent stirring-induced “spontaneous” platelet aggregation in human whole blood.It is concluded that the observed antithrombotic action of piracetam cannot satisfactorily be explained by an isolated direct effect on platelets. An additional influence of piracetam on the rheology of the circulating blood and/or on the vessel wall itself must therefore be taken into consideration.


Blood ◽  
2009 ◽  
Vol 114 (22) ◽  
pp. 156-156
Author(s):  
Christopher M. Spring ◽  
Wuxun Jin ◽  
Hong Yang ◽  
Adili Reheman ◽  
Guangheng Zhu ◽  
...  

Abstract Abstract 156 Platelet adhesion and aggregation at sites of vascular injury are key events required for haemostasis and thrombosis. It has been documented that von Willebrand factor (VWF) and fibrinogen (Fg) are required for platelet adhesion and aggregation. However, we previously showed that occlusive thrombi still form in mice deficient for both Fg and VWF (Fg/VWF−/−) via a β3 integrin-dependent pathway. Here, we have investigated novel, non-classical ligands of β3 integrin that may regulate platelet adhesion and aggregation. To identify potential ligand(s) of β3 integrin, latex beads were coated with purified human platelet β3 integrin and incubated with human plasma. Protein(s) specifically associated with β3 integrin were electrophoresed and apolipoprotein AIV (ApoA-IV) was identified by mass spectrometry. We found that ApoA-IV binds to the surface of stimulated platelets, but not to quiescent platelets or β3−/− platelets, and ApoA-IV/platelet association was blocked by the addition of a specific anti-β3 integrin monoclonal antibody. It appears that ApoA-IV binds to, but is not internalized by platelet β3 integrins. ApoA-IV-deficient (ApoA-IV−/−) mice exhibited enhanced platelet aggregation induced by ADP, Collagen, and TRAP in plasma (but not PIPES buffer) compared to wild type (WT) littermates. This enhancement was diminished when ApoA-IV−/− plasma was replaced by WT plasma, indicating that the reduction was due to plasma ApoA-IV and not an unrelated platelet effect. When platelets were incubated with FITC-Fg, ApoA-IV was able to reduce platelet/Fg association, indicating that ApoA-IV may act to displace pro-thrombotic β3 integrin ligand(s). In support of this, ApoA-IV reduced the number of adherent platelets on immobilized Fg in perfusion chamber assays and enhanced thrombus formation was observed when ApoA-IV−/− mouse blood was perfused over collagen. We found that addition of recombinant ApoA-IV inhibited platelet aggregation and thrombus formation in vitro, while the control apolipoprotein ApoA-I did not. Using intravital microscopy, we further demonstrated that early platelet deposition was increased, and the time for thrombus formation and vessel occlusion were shorter in ApoA-IV−/− mice, which can be corrected by recombinant ApoA-IV transfusion. Furthermore, recombinant ApoA-IV inhibited WT platelet aggregation, thrombus formation and enhanced thrombus dissolution both in vitro and in vivo. Our data demonstrate for the first time that ApoA-IV is a novel ligand of platelet β3 integrin that negatively regulates thrombosis. These new data are consistent with the reported association between ApoA-IV and reduced cardiovascular diseases, and establish the first link between ApoA-IV and thrombosis. Disclosures: No relevant conflicts of interest to declare.


2016 ◽  
Vol 116 (12) ◽  
pp. 1100-1110 ◽  
Author(s):  
Isabella Massimi ◽  
Lavinia Lotti ◽  
Flavia Temperilli ◽  
Massimo Mancone ◽  
Gennaro Sardella ◽  
...  

SummaryPlatelet multidrug resistance protein4 (MRP4)-overexpression has a role in reducing aspirin action. Aspirin in vivo treatment enhances platelet MRP4 expression and MRP4 mediated transport inhibition reduces platelet function and delays thrombus formation. The aim of our work was to verify whether MRP4 expression is enhanced in platelets obtained from patients under chronic aspirin treatment and whether it correlates with residual platelet reactivity. We evaluated changes on mRNA and protein-MRP4 expression and platelet aggregation in four populations: healthy volunteers (HV), aspirin-free control population (CTR), patients who started the treatment less than one month ago (ASA<1 month patients) and aspirinated patients who started the treatment more than two months ago (ASA>2 months patients). In platelets obtained from ASA>2 months patients, it was found a statistically significant MRP4 enhancement of both mRNA and protein expression compared to HV, CTR and ASA<1 month patients. Platelets obtained from ASA>2 months patients that present high levels of platelet MRP4, have higher serum TxB2 levels and collagen-induced platelet aggregation compared to patient with low levels of MRP4 in platelets. In addition collagen induced platelet aggregation is higher in in vitro aspirinated platelets obtained from patients with high levels of MRP4 patients compared to those obtained from patients with low MRP4 levels. We can assert that, in patients under chronic aspirin treatment, platelets that present high MRP4 levels have an increase of residual platelet reactivity, which is due in part to incomplete COX-1 inhibition, and in part to COX-1–independent mechanism.


1981 ◽  
Author(s):  
H D Lehmann ◽  
J Gries ◽  
D Lenke

6- [p-(2-(Chiorpropionylamino)phenyl] -4.5-dihydro-5-methyl-3(2H)-pyridazinone, LU 23051, is primarily characterized by its strong inhibition of platelet aggregation under in vitro and in vivo conditions. In vitro there is a concentration-dependent inhibition of ADP and collagen induced aggregation in platelet rich plasma of man, rat and dog. The inhibitory concentration EC 33 % is 0.0010-0.030 mg/1 (man: ADP-0.030, col 1.-0.013 mg/l) depending on species and type of aggregation. When administered orally in ex vivo experiments on rats and dogs the substance is found to have a dose-dependent antiaggregatory effect in the range from 0.1-3.16 mg/kg. The ED 33 % is 0.27-0.63 mg/kg.-In addition after oral administration the substance has a good inhibitory effect in models being based on intravascular platelet aggregation. Thus, a dose of 1 mg/kg inhibits laser-induced aggregation in mesenteric venules of rats. Mortality after i.v. injection of collagen in mice is reduced by 50 % after a dose of 0.02 mg/kg. A dose of 0.039 mg/kg prolongs the bleeding time of rats by 50 %. The aggregation-inhibiting action is of long duration (0.1 mg/kg p.o.∼24 h). The substance does not interfere with clotting.Besides its effect on platelet aggregation LU 23051 acts as vasodilatator as well. Dilatation of coronary vessels by 100 % is seen in isolated guinea-pig hearts at a concentration of 0.1 mg/l. In spontaneously hypertensive rats the substance has an anti hypertensive effect. The ED 20 % is 0.36 mg/kg p.o.The combination of antiaggregatory and vasodilatatory effects opens up interesting aspects with respect to the pharmacotherapeutic use of the new substance


Sign in / Sign up

Export Citation Format

Share Document