Transl. Stroke Res. DOI 10.1007/s12975-014-0377-3
ORIGINAL ARTICLE
Warfarin Pretreatment Reduces Cell Death and MMP-9 Activity in Experimental Intracerebral Hemorrhage Frieder Schlunk & Elena Schulz & Arne Lauer & Kazim Yigitkanli & Waltraud Pfeilschifter & Helmuth Steinmetz & Eng H. Lo & Christian Foerch
Received: 15 September 2014 / Revised: 6 November 2014 / Accepted: 11 November 2014 # Springer Science+Business Media New York 2014
Abstract Little is known about the pathophysiology of oral anticoagulation-associated intracerebral hemorrhage (OACICH). We compared hematoma volume, number of terminal deoxynucleotidyl dUTP nick-end labeling (TUNEL)-positive cells (indicating cell death), MMP-9 levels, and perilesional edema formation between warfarin-treated mice and controls. Intracerebral hemorrhage was induced by an injection of collagenase into the right striatum. Twenty-four hours later, hematoma volume was measured using a photometric hemoglobin assay. Cell death was quantified using TUNEL staining. MMP-9 levels were determined by zymography, and edema formation was assessed via the wet–dry method. Warfarin increased hematoma volume by 2.6-fold. The absolute number of TUNEL-positive cells in the perihematomal zone was lower in warfarin-treated animals (300.5±39.8 cells/mm2) than in controls (430.5±38.9 cells/mm2; p=0.034), despite the larger bleeding volume. MMP-9 levels were reduced in anticoagulated mice as compared to controls (p=0.018). Perilesional edema formation was absent in warfarin mice and modestly present in controls. Our results suggest differences in the pathophysiology of OAC-ICH compared to intracerebral hemorrhage occurring under normal coagulation. A likely explanation is that thrombin, a strong inductor of apoptotic cell death and blood–brain barrier disruption, is produced to a lesser extent in OAC-ICH. In humans, however, we assume that
F. Schlunk (*) : E. Schulz : A. Lauer : W. Pfeilschifter : H. Steinmetz : C. Foerch Department of Neurology, Johann Wolfgang Goethe University, Frankfurt am Main, Germany e-mail: [email protected] F. Schlunk : E. Schulz : A. Lauer : K. Yigitkanli : E. H. Lo Neuroprotection Research Laboratory, Massachusetts General Hospital, Harvard Medical School, Charlestown, USA
the detrimental effects of a larger hematoma volume in OAC-ICH by far outweigh potential protective effects of thrombin deficiency. Keywords Intracerebral hemorrhage . Warfarin . Animal model . Cell death . Edema . Blood–brain barrier
Introduction Oral anticoagulation-associated intracerebral hemorrhage (OAC-ICH) is a particularly deadly type of stroke with a short-term mortality rate above 50 % [1]. The prevalence of OAC-ICH is likely to rise in the future along with a more widespread use of long-term anticoagulation for the prevention of thrombotic and thromboembolic events. Despite its clinical importance, the pathophysiology of the disease is not well understood. Clinical data from observational series suggest that OAC-ICH is associated with a larger hematoma volume and a worse functional outcome as compared to brain bleedings occurring during a normal coagulation status [2–5]. Furthermore, hematoma expansion after the very first hours was reported to occur more frequently in OAC-ICH [6]. It was speculated that perihematomal edema formation in case of OAC-ICH is less pronounced as compared to bleedings in non-anticoagulated controls, despite a larger bleeding volume [7]. In fact, the coagulation factor II (thrombin) is known to be an important inductor of blood– brain barrier (BBB) disruption [8–10]. Thus, in turn, the reduced thrombin formation within the hematoma in case of OAC-ICH may prevent from disintegration of the BBB and edema formation [7]. Furthermore, as thrombin is also an inductor of apoptotic cascades, types and quantity of cell death may also differ between OAC-ICH and intracerebral hemorrhage occurring during a normal coagulation status.
Transl. Stroke Res.
Recently, an experimental model of OAC-ICH was developed, revealing an approximately twofold increase in hematoma volume after 24 h in mice anticoagulated to an international normalized ratio (INR) of 3.5±0.9 as compared to controls [11]. We used this model to further explore the pathophysiology of the disease, with a particular focus on cell death, MMP-9 activity, and perilesional edema formation.
Methods
skull 0 mm anterior and 2.0 mm lateral from the bregma and a 32-gauge 0.5 μl microinjection needle (Hamilton 7000 series) was lowered 3.5 mm in depth into the right striatum. Over a period of 5 min, 0.5 μl of saline containing 0.2 U of collagenase VII-s was injected. The needle was left in situ for 10 min and was slowly removed afterwards over 5 min. The borehole was sealed with bone-wax. After suturing, the mice were allowed to recover and returned to their cages (regular tap water was supplied at that time). During the surgical procedure (lasting approx. 30 min per mouse), a heat lamp was used to maintain body temperature.
Animals Outcome Assessment All experiments were performed under protocols approved by the Massachusetts General Hospital Institutional Animal Care and Use Committees (IACUC, protocol N-00087) following the NIH Guide for Use and Care of Laboratory Animals. For the entire study, male CD-1 mice aged 12 to 16 weeks were used. Mice were randomly assigned to warfarin or sham treatment, respectively, and were evaluated within the following four study collectives (group sizes defined according to prior studies [11–14]): (I) hematoma volume measurement (warfarin: n=5, controls: n=5), (II) cell death quantification (warfarin: n=10, controls: n=10), (III) MMP-9 activity determination (warfarin: n=6, controls: n=4), and (IV) perilesional edema measurement (warfarin: n=10, controls: n=10; due to an unexpectedly high mortality rate in the warfarin group, additional mice were randomized in a 2:1 ratio [warfarin: n=13, controls: n=6]). All mice were used for functional outcome determination 24 h after hemorrhage induction. For the entire study, researchers were fully blinded regarding the treatment status of the animals. Warfarin Administration Warfarin (2 mg/kg) was administered via drinking water for 24 h following a previously established protocol. Briefly, a 5mg coumadin tablet (crystalline warfarin sodium, Bristol Myers Squibb, New York City, NY, USA) was dissolved in 375 ml of tap water. Mice had free access to this solution in the cage during the feeding period. No other water was supplied. After the feeding period, mice revealed a mean INR of 3.5± 0.9 which is very close to the therapeutic range (INR 2–3.5) in humans [11]. A detailed analysis of coagulation factors in this model demonstrated a vast reduction of the activity of factors II, VII, IX, and X, mimicking full warfarin anticoagulation [11, 15]. Control mice received regular tap water during the entire feeding period (INR 0.8±0.1). Intracerebral Hemorrhage Induction Mice were anesthetized with isoflurane (1.5–2 %) under spontaneous respiration. A small borehole was drilled in the
Twenty-four hours after hemorrhage induction, functional outcome was measured using a well-characterized 5-point ordinal [11]. For doing so, mice were placed in the center of a steel table and movements were observed for 1 min. Neurological outcome was categorized in 0=no apparent deficit; 1=slight deficit as instability during walking, but no circling; 2=circling to the right with at least some straight movements and covering of distance; 3=heavy circling to the right, no gain of distance or no movements at all; and 4=death. Intracerebral Hemorrhage Volume Determination Twenty-four hours after intracerebral hemorrhage induction, mice were deeply anesthetized with 5 % isoflurane and transcardially perfused with 30 ml of phosphate-buffered saline (PBS). Brains were removed, divided into left and right hemisphere, and immersed in 3 ml of PBS. The cerebellum was discarded. Afterwards, the hemispheres were homogenized for 30 s, and ultrasound was applied to the homogenate for 1 min in order to lyse erythrocyte membranes. Following centrifugation (30 min, 13,000 rpm at 4 °C), 250 μl of supernatant was mixed with 1000 μl of Drabkin’s reagent. A photometer was used to determine absorption rates at 540 nm. Hematoma volumes were calculated for the entire brain using a standard curve. Mice which died within 24 h were not eligible for transcardial perfusion. In such cases, we subtracted 2.1 μl from the blood volumes for these brains. This value was found to be the mean difference in blood volumes between three perfused and three unperfused brains in our previous study in the same mouse strain [11]. TUNEL Staining Terminal deoxynucleotidyl dUTP nick-end labeling (TUNEL) staining was used to determine cell death 24 h after hemorrhage induction. Mice were deeply anesthetized and perfused transcardially with 30 ml of PBS. Brains were removed and frozen on dry ice. At the site of the needle insertion, where the
Transl. Stroke Res.
maximum of the bleeding is assumed, cryosections (20 μm) were cut. To detect cells with fragmented DNA, TUNEL was performed using the TACS™ 2 TdT-Fluor in situ apoptosis detection kit (Trevigen). Sections were analyzed using an Olympus microscope (BX51; Olympus). With the magnification set to 200-fold, four images of randomly chosen areas representing the rim around the hematoma were obtained (Fig. 1a-c). The number of TUNEL-positive cells was quantified using ImageJ software (NIH, Bethesda, USA) and expressed as cell density (cells/mm2). MMP-9 Activity Determination MMP-9 activity was determined 24 h after hemorrhage induction. For doing so, mice were deeply anesthetized and transcardially perfused with PBS as described above. Using a mouse brain matrix, 1-mm slices were cut and divided into left and right hemisphere. On both sides, the striatum and the striatal hematoma, respectively, were removed leaving no relevant traces of blood behind. Afterwards, the hemispheres were frozen and stored at −80 °C. For MMP-9 measurements,
Fig. 1 TUNEL-positive cells in the area representing the rim around the hematoma 24 h after hemorrhage induction. a Image of a 20-μm thick cryosection. The boxes indicate the sites of the analyzed region in every slice. b Representative images of TUNEL-positive cells in control mice. Images were taken 24 h after hemorrhage induction from randomly chosen areas representing the rim around the hematoma with 200-fold
brain slices were thawed and homogenized in 600 μl cell lysis buffer containing protease inhibitors on ice. Following a centrifugation step (30 min, 13,000 rpm at 4 °C), the protein concentration of the supernatant was measured using the Bradford assay (Bio-Rad). Samples with equal protein concentration were then prepared, loaded into a 10 % Tris-glycine gel with 0.1 % gelatin as substrate, and submitted to gel electrophoresis, followed by a 24-h incubation at 37 °C. Thereafter, gels were stained with a Coomassie brilliant blue solution which positively stains gelatin and hence MMP-9 activity leads to colorless bands in the gel. To quantify the MMP-9 activity, pictures of the gels were taken, and the mean gray value was determined using the ImageJ software. Values from non-injured contralateral hemispheres were subtracted from the values obtained for the ipsilateral sides. Brain Edema Quantification To measure perilesional edema formation, mice were deeply anesthetized (isoflurane 5 %) and euthanized. After removing the brain, a 3-mm block was cut with the needle insertion point
magnification. c Representative images of TUNEL-positive cells in anticoagulated mice (see b). d Number of TUNEL-positive cells per mm2 in controls (n=10) and in animals pretreated with warfarin (n=8). Anticoagulated mice showed significantly less TUNEL-positive cells as compared to controls
Transl. Stroke Res.
in the middle using a mouse brain matrix and divided into left and right hemisphere. After quickly removing the striatum and the striatal hematoma, respectively, the cortex (adjacent to the hematoma on the ipsilateral side, but free of blood) was weighed on a preweighed aluminum foil using an electronic balance in order to obtain the wet weight. After drying the samples for 48 h at 80 °C, the dry weight was determined. Results were expressed in percent as follows: (wet weight − dry weight) / wet weight × 100. The cerebellum served as a control. Statistical Analysis SPSS version 12.0 (SPSS Inc, Chicago, III) was used for the statistical analysis. Results for TUNEL-positive cells, MMP9, and edema were compared using the parametric t test. Statistical analysis of ordinal data was performed with the χ2 test. For correlation analysis between outcome, cell death, MMP-9, and edema, the Spearman rank test was used.
Results Functional Outcome Determined over all subgroups and experiments, mice that were anticoagulated with warfarin had a higher mortality rate after 24 h (14/44) as compared to nonanticoagulated controls (0/35, p
ORIGINAL ARTICLE
Warfarin Pretreatment Reduces Cell Death and MMP-9 Activity in Experimental Intracerebral Hemorrhage Frieder Schlunk & Elena Schulz & Arne Lauer & Kazim Yigitkanli & Waltraud Pfeilschifter & Helmuth Steinmetz & Eng H. Lo & Christian Foerch
Received: 15 September 2014 / Revised: 6 November 2014 / Accepted: 11 November 2014 # Springer Science+Business Media New York 2014
Abstract Little is known about the pathophysiology of oral anticoagulation-associated intracerebral hemorrhage (OACICH). We compared hematoma volume, number of terminal deoxynucleotidyl dUTP nick-end labeling (TUNEL)-positive cells (indicating cell death), MMP-9 levels, and perilesional edema formation between warfarin-treated mice and controls. Intracerebral hemorrhage was induced by an injection of collagenase into the right striatum. Twenty-four hours later, hematoma volume was measured using a photometric hemoglobin assay. Cell death was quantified using TUNEL staining. MMP-9 levels were determined by zymography, and edema formation was assessed via the wet–dry method. Warfarin increased hematoma volume by 2.6-fold. The absolute number of TUNEL-positive cells in the perihematomal zone was lower in warfarin-treated animals (300.5±39.8 cells/mm2) than in controls (430.5±38.9 cells/mm2; p=0.034), despite the larger bleeding volume. MMP-9 levels were reduced in anticoagulated mice as compared to controls (p=0.018). Perilesional edema formation was absent in warfarin mice and modestly present in controls. Our results suggest differences in the pathophysiology of OAC-ICH compared to intracerebral hemorrhage occurring under normal coagulation. A likely explanation is that thrombin, a strong inductor of apoptotic cell death and blood–brain barrier disruption, is produced to a lesser extent in OAC-ICH. In humans, however, we assume that
F. Schlunk (*) : E. Schulz : A. Lauer : W. Pfeilschifter : H. Steinmetz : C. Foerch Department of Neurology, Johann Wolfgang Goethe University, Frankfurt am Main, Germany e-mail: [email protected] F. Schlunk : E. Schulz : A. Lauer : K. Yigitkanli : E. H. Lo Neuroprotection Research Laboratory, Massachusetts General Hospital, Harvard Medical School, Charlestown, USA
the detrimental effects of a larger hematoma volume in OAC-ICH by far outweigh potential protective effects of thrombin deficiency. Keywords Intracerebral hemorrhage . Warfarin . Animal model . Cell death . Edema . Blood–brain barrier
Introduction Oral anticoagulation-associated intracerebral hemorrhage (OAC-ICH) is a particularly deadly type of stroke with a short-term mortality rate above 50 % [1]. The prevalence of OAC-ICH is likely to rise in the future along with a more widespread use of long-term anticoagulation for the prevention of thrombotic and thromboembolic events. Despite its clinical importance, the pathophysiology of the disease is not well understood. Clinical data from observational series suggest that OAC-ICH is associated with a larger hematoma volume and a worse functional outcome as compared to brain bleedings occurring during a normal coagulation status [2–5]. Furthermore, hematoma expansion after the very first hours was reported to occur more frequently in OAC-ICH [6]. It was speculated that perihematomal edema formation in case of OAC-ICH is less pronounced as compared to bleedings in non-anticoagulated controls, despite a larger bleeding volume [7]. In fact, the coagulation factor II (thrombin) is known to be an important inductor of blood– brain barrier (BBB) disruption [8–10]. Thus, in turn, the reduced thrombin formation within the hematoma in case of OAC-ICH may prevent from disintegration of the BBB and edema formation [7]. Furthermore, as thrombin is also an inductor of apoptotic cascades, types and quantity of cell death may also differ between OAC-ICH and intracerebral hemorrhage occurring during a normal coagulation status.
Transl. Stroke Res.
Recently, an experimental model of OAC-ICH was developed, revealing an approximately twofold increase in hematoma volume after 24 h in mice anticoagulated to an international normalized ratio (INR) of 3.5±0.9 as compared to controls [11]. We used this model to further explore the pathophysiology of the disease, with a particular focus on cell death, MMP-9 activity, and perilesional edema formation.
Methods
skull 0 mm anterior and 2.0 mm lateral from the bregma and a 32-gauge 0.5 μl microinjection needle (Hamilton 7000 series) was lowered 3.5 mm in depth into the right striatum. Over a period of 5 min, 0.5 μl of saline containing 0.2 U of collagenase VII-s was injected. The needle was left in situ for 10 min and was slowly removed afterwards over 5 min. The borehole was sealed with bone-wax. After suturing, the mice were allowed to recover and returned to their cages (regular tap water was supplied at that time). During the surgical procedure (lasting approx. 30 min per mouse), a heat lamp was used to maintain body temperature.
Animals Outcome Assessment All experiments were performed under protocols approved by the Massachusetts General Hospital Institutional Animal Care and Use Committees (IACUC, protocol N-00087) following the NIH Guide for Use and Care of Laboratory Animals. For the entire study, male CD-1 mice aged 12 to 16 weeks were used. Mice were randomly assigned to warfarin or sham treatment, respectively, and were evaluated within the following four study collectives (group sizes defined according to prior studies [11–14]): (I) hematoma volume measurement (warfarin: n=5, controls: n=5), (II) cell death quantification (warfarin: n=10, controls: n=10), (III) MMP-9 activity determination (warfarin: n=6, controls: n=4), and (IV) perilesional edema measurement (warfarin: n=10, controls: n=10; due to an unexpectedly high mortality rate in the warfarin group, additional mice were randomized in a 2:1 ratio [warfarin: n=13, controls: n=6]). All mice were used for functional outcome determination 24 h after hemorrhage induction. For the entire study, researchers were fully blinded regarding the treatment status of the animals. Warfarin Administration Warfarin (2 mg/kg) was administered via drinking water for 24 h following a previously established protocol. Briefly, a 5mg coumadin tablet (crystalline warfarin sodium, Bristol Myers Squibb, New York City, NY, USA) was dissolved in 375 ml of tap water. Mice had free access to this solution in the cage during the feeding period. No other water was supplied. After the feeding period, mice revealed a mean INR of 3.5± 0.9 which is very close to the therapeutic range (INR 2–3.5) in humans [11]. A detailed analysis of coagulation factors in this model demonstrated a vast reduction of the activity of factors II, VII, IX, and X, mimicking full warfarin anticoagulation [11, 15]. Control mice received regular tap water during the entire feeding period (INR 0.8±0.1). Intracerebral Hemorrhage Induction Mice were anesthetized with isoflurane (1.5–2 %) under spontaneous respiration. A small borehole was drilled in the
Twenty-four hours after hemorrhage induction, functional outcome was measured using a well-characterized 5-point ordinal [11]. For doing so, mice were placed in the center of a steel table and movements were observed for 1 min. Neurological outcome was categorized in 0=no apparent deficit; 1=slight deficit as instability during walking, but no circling; 2=circling to the right with at least some straight movements and covering of distance; 3=heavy circling to the right, no gain of distance or no movements at all; and 4=death. Intracerebral Hemorrhage Volume Determination Twenty-four hours after intracerebral hemorrhage induction, mice were deeply anesthetized with 5 % isoflurane and transcardially perfused with 30 ml of phosphate-buffered saline (PBS). Brains were removed, divided into left and right hemisphere, and immersed in 3 ml of PBS. The cerebellum was discarded. Afterwards, the hemispheres were homogenized for 30 s, and ultrasound was applied to the homogenate for 1 min in order to lyse erythrocyte membranes. Following centrifugation (30 min, 13,000 rpm at 4 °C), 250 μl of supernatant was mixed with 1000 μl of Drabkin’s reagent. A photometer was used to determine absorption rates at 540 nm. Hematoma volumes were calculated for the entire brain using a standard curve. Mice which died within 24 h were not eligible for transcardial perfusion. In such cases, we subtracted 2.1 μl from the blood volumes for these brains. This value was found to be the mean difference in blood volumes between three perfused and three unperfused brains in our previous study in the same mouse strain [11]. TUNEL Staining Terminal deoxynucleotidyl dUTP nick-end labeling (TUNEL) staining was used to determine cell death 24 h after hemorrhage induction. Mice were deeply anesthetized and perfused transcardially with 30 ml of PBS. Brains were removed and frozen on dry ice. At the site of the needle insertion, where the
Transl. Stroke Res.
maximum of the bleeding is assumed, cryosections (20 μm) were cut. To detect cells with fragmented DNA, TUNEL was performed using the TACS™ 2 TdT-Fluor in situ apoptosis detection kit (Trevigen). Sections were analyzed using an Olympus microscope (BX51; Olympus). With the magnification set to 200-fold, four images of randomly chosen areas representing the rim around the hematoma were obtained (Fig. 1a-c). The number of TUNEL-positive cells was quantified using ImageJ software (NIH, Bethesda, USA) and expressed as cell density (cells/mm2). MMP-9 Activity Determination MMP-9 activity was determined 24 h after hemorrhage induction. For doing so, mice were deeply anesthetized and transcardially perfused with PBS as described above. Using a mouse brain matrix, 1-mm slices were cut and divided into left and right hemisphere. On both sides, the striatum and the striatal hematoma, respectively, were removed leaving no relevant traces of blood behind. Afterwards, the hemispheres were frozen and stored at −80 °C. For MMP-9 measurements,
Fig. 1 TUNEL-positive cells in the area representing the rim around the hematoma 24 h after hemorrhage induction. a Image of a 20-μm thick cryosection. The boxes indicate the sites of the analyzed region in every slice. b Representative images of TUNEL-positive cells in control mice. Images were taken 24 h after hemorrhage induction from randomly chosen areas representing the rim around the hematoma with 200-fold
brain slices were thawed and homogenized in 600 μl cell lysis buffer containing protease inhibitors on ice. Following a centrifugation step (30 min, 13,000 rpm at 4 °C), the protein concentration of the supernatant was measured using the Bradford assay (Bio-Rad). Samples with equal protein concentration were then prepared, loaded into a 10 % Tris-glycine gel with 0.1 % gelatin as substrate, and submitted to gel electrophoresis, followed by a 24-h incubation at 37 °C. Thereafter, gels were stained with a Coomassie brilliant blue solution which positively stains gelatin and hence MMP-9 activity leads to colorless bands in the gel. To quantify the MMP-9 activity, pictures of the gels were taken, and the mean gray value was determined using the ImageJ software. Values from non-injured contralateral hemispheres were subtracted from the values obtained for the ipsilateral sides. Brain Edema Quantification To measure perilesional edema formation, mice were deeply anesthetized (isoflurane 5 %) and euthanized. After removing the brain, a 3-mm block was cut with the needle insertion point
magnification. c Representative images of TUNEL-positive cells in anticoagulated mice (see b). d Number of TUNEL-positive cells per mm2 in controls (n=10) and in animals pretreated with warfarin (n=8). Anticoagulated mice showed significantly less TUNEL-positive cells as compared to controls
Transl. Stroke Res.
in the middle using a mouse brain matrix and divided into left and right hemisphere. After quickly removing the striatum and the striatal hematoma, respectively, the cortex (adjacent to the hematoma on the ipsilateral side, but free of blood) was weighed on a preweighed aluminum foil using an electronic balance in order to obtain the wet weight. After drying the samples for 48 h at 80 °C, the dry weight was determined. Results were expressed in percent as follows: (wet weight − dry weight) / wet weight × 100. The cerebellum served as a control. Statistical Analysis SPSS version 12.0 (SPSS Inc, Chicago, III) was used for the statistical analysis. Results for TUNEL-positive cells, MMP9, and edema were compared using the parametric t test. Statistical analysis of ordinal data was performed with the χ2 test. For correlation analysis between outcome, cell death, MMP-9, and edema, the Spearman rank test was used.
Results Functional Outcome Determined over all subgroups and experiments, mice that were anticoagulated with warfarin had a higher mortality rate after 24 h (14/44) as compared to nonanticoagulated controls (0/35, p
