Administration of Human Protein-C concentrate prevents apoptotic brain cell death after experimental sepsis

BR A I N R ES E A RC H 1 2 6 4 ( 2 00 9 ) 1 1 9 – 1 26

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

w w w. e l s e v i e r. c o m / l o c a t e / b r a i n r e s

Research Report

Administration of Human Protein-C concentrate prevents apoptotic brain cell death after experimental sepsis Nikolaos Memos a , Alex Betrosian b , Evangelos Messaris c , Maria Boutsikou a , Agapi Kataki a , Emmy Chatzigianni a , Marilena Nikolopoulou a , Emmanuel Leandros a , Manousos Konstadoulakis a,⁎ a

Laboratory of Surgical Research, First Department of Propaedeutic Surgery, Athens University Medical School, Hippocration General Hospital, Athens, Greece b Intensive Care Unit, Evgenidion Hospital, Athens, Greece c Department of Surgery, Brown University, Providence, RI, USA

A R T I C LE I N FO

AB S T R A C T

Article history:

Activated Protein C renders anti-apoptotic properties in neurons and endothelial cells. The

Accepted 24 January 2009

aim of the present study was to evaluate the in vivo cytoprotective role of Protein C zymogen

Available online 5 February 2009

(PC) administration in septic rat brain. Male Wistar rats (n = 60) were subjected to sepsis via Cecal Ligation and Puncture (CLP). Animals were randomly divided either to receive 100 IU/

Keywords:

kg human PC concentrate at 1, 7 and 13 h post CLP (CLP + PC group) or placebo treatment (CLP

Sepsis

group). At pre-specified time points (6, 12, 24, 36, 48 and 60 h post CLP) five animals from

Brain

either group were euthanized and the brain tissue was removed. Apoptosis in both neurons

Apoptosis

(Neu-N+) and astroglia (GFAP+) was assessed by flow cytometry using 7-aminoactinomycin

Protein C zymogen

D (7AAD). Immunohistochemical detection of cleaved caspase 3, bax, bcl-2, cytochrome c and caspase 8 was also performed. PC treated animals had significantly reduced apoptosis in neurons at 6 and 24 h post CLP (p = 0.04 and p = 0.016 respectively) and necrosis at 6, 12 and 60 h post CLP (p = 0.008, p = 0.012 and p = 0.032 respectively). Astrocyte necrosis was also decreased in septic rats receiving PC (6, 12 and 60 h post CLP p = 0.008, p = 0.016 and p = 0.008 respectively). In addition, active caspase 3, bax, cytochrome c and caspase 8 expression was significantly decreased during early sepsis (6–36 h) while bcl-2 expression was increased (24 h p = 0.001 and 60 h p = 0.001) in the PC treated animals compared to placebo. PC concentrate administration in experimental sepsis produced a time dependent inhibition of apoptosis in rat neurons and astrocytes. The inhibition of sepsis related apoptosis concerned both the mitochondrial and caspase 8 dependent pathways. © 2009 Elsevier B.V. All rights reserved.

⁎ Corresponding author. 1st Department of Propaedeutic Surgery, Athens Medical School, Kalvou 24 Str, Palaio Psychico, 15452 Athens, Greece. Fax: +302107707574. E-mail addresses: [email protected], [email protected] (M. Konstadoulakis). 0006-8993/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2009.01.053

120

1.

BR A I N R ES E A RC H 1 2 6 4 ( 2 00 9 ) 1 1 9 –12 6

Introduction

During disease apoptosis is implicated in almost each form of brain injury such as stroke, traumatic brain injury, neurodegenerative disorders as well as brain infection (Muller et al. 2004; Zhang et al. 2003; Raghupathi et al., 2000). The role of apoptosis during sepsis in vital organs is well documented, but the role of apoptosis in septic brain remains unclear (Hotchkiss et al., 1997). Blood brain barrier breakdown as well as amino acid derangements, abnormal neurotransmitters and apoptosis have been documented to occur during the septic syndrome and seem to affect brain function (Papadopoulos et al., 2000). Protein C is a vitamin K dependent plasma protein which circulates in a precursor form (Esmon, 2000). In the presence of thrombin–thrombomodulin complex, Protein C is activated to Activated Protein C (APC). APC has anticoagulant actions via inhibition of factors Va and VIIa, as well as anti-inflammatory properties since in vitro studies showed that APC inhibits cellular activation, including nuclear translocation of NF-B and production of proinflammatory cytokines (White et al., 2000). Neutrophils and monocytes have also receptors that interact with protein C, APC, and rhAPC resulting in inhibition of neutrophil and monocyte chemotaxis (Sturn et al., 2003). In particular, incubation of neutrophils or monocytes with protein C, APC, or rhAPC decreases directional migration to chemotactins such as interleukin 8 (Abraham, 2005). Recent in vitro studies showed that APC renders antiapoptotic effects in brain endothelium as well as in neurons (Cheng et al., 2003, Guo et al., 2004). In addition APC inhibits apoptosis in hippocampal cells and in white matter of neonatal rats following hypoxia (Yesilirmak et al., 2007), in spinal cord injury and in ischemic brain (Cheng et al., 2003; Zlokovic et al., 2005). Septic patients experience acquired Protein C deficiency and PC depletion has been associated with a poor outcome (Shorr et al., 2006). Furthermore, human recombinant Acti-

vated Protein C (hrAPC) is an effective treatment for severe sepsis while Protein C zymogen supplementation has been used so far to treat meningococcal sepsis (O Brien et al., 2006; de Kleijn et al., 2003). In the present study we hypothesized that Human ProteinC administration in sepsis provides an anti-apoptotic effect. To test this hypothesis we investigated the anti-apoptotic effect of in vivo PC administration on brain following sepsis induced by cecal ligation and puncture.

2.

Results

2.1.

Effect of PC concentrate in neuronal cell death

Protein C concentrate administration resulted in a 3 fold reduction in neuronal apoptosis compared to untreated septic animals at 6 (p = 0.04) and a 2-fold decrease at 24 h post CLP (p = 0.016, Figs. 1 and 2). Neuronal necrosis was significantly decreased as well in the treatment group during the early phase of sepsis at 6 (p = 0.008) and 12 h (p = 0.012) and at 60 h post CLP (p = 0.032) compared to placebo (Fig. 3). Time distribution of neuronal apoptosis within treated group yielded a “U” shaped expression with increased levels of apoptosis at 6 h followed by a decrease at 24 h and raised again up to 60 h post CLP (p < 0.001). On the other hand, neuronal necrosis followed an S curve beginning with low levels during the first 12 h post CLP and significantly increased after 24 h (p < 0.001). At the placebo group apoptosis and necrosis in neurons did not significantly altered during the course of sepsis.

2.2.

Effect of PC concentrate in astrocyte cell death

Astrocyte necrosis was significantly reduced in treatment group at 6 (p = 0.008) and 12 h (p = 0.016 for astrocytes) compared to placebo as well as 60 h (p = 0.008) post procedure (Fig. 3). In the

Fig. 1 – Flow cytometry chart showing Neu-N(+) apoptotic cells from PC concentrate treated (A) and a relative untreated (B) animal. The presence of apoptotic cells (median bar) in treated animal is far less evident compared to cells from an untreated animal.

BR A I N R ES E A RC H 1 2 6 4 ( 2 00 9 ) 1 1 9 – 1 26

121

untreated animals (p = 0.016, Fig. 7B). Accordingly cytochrome c immunopositivity significantly declined in treated animals both at 6 and 24 h post CLP compared to CLP animals (p = 0.02, p = 0.016 respectively, Fig. 7C). Finally caspase 8 protein expression was significantly decreased in CLP + PC group at 12 h (p = 0.045, Fig. 7D). In the active treatment group, the time distribution of bcl-2 expression showed a rising potential until 24 h when a significant decrease was noticed (p = 0.014). On the contrary bax immunodetection in CLP + PC group was significantly low 6 h post CLP (p = 0.044). Cytochrome c immunodetection in the treatment group remained constant while being significantly decreased in placebo 48 h post CLP (p = 0.001).

3.

Discussion

Apoptosis in brain following sepsis has been documented in autonomic centres of patients with septic shock and in

Fig. 2 – Time distribution of Neu-N(+) apoptosis from treated and untreated septic rats at 6, 12, 24, 36, 48 and 60 h post CLP (n = 5 at each time point). Apoptosis significantly differed in the treated group at 6 and 24 h compared to placebo. (Error bars represent 1SEM, *: denotes statistical significance)

same group, necrosis of GFAP(+) cells was consistent with neuronal necrosis showing a significant lower pattern at 6 and 12 h and a significant increase thereafter until 60 h when a significant decline was noticed (p < 0.001). GFAP(+) apoptotic cells followed the same time pattern as neurons showing a “U” shaped curve (p=0.005). On the contrary, the time trend of astroglial apoptosis in placebo group was not significantly altered.

2.3. Effect of PC concentrate in caspase 3 immunoreactivity Cleaved caspase 3 mainly occurred in Purkinje cells of the cerebellum, pyramidal cells of Hippocampus, astroglial cells and ependymal cells of choroid plexus (Fig. 4). In PC treated animals active caspase 3 expression was significantly decreased at 6 h (p = 0.012), 12 h (p = 0.026), 24 (p = 0.018) and 36 h (p = 0.02) post CLP compared to untreated animals (Fig. 5). Although the time distribution of active caspase 3 immunoreactivity within placebo septic animals was not found statistically significant (Kruskal Wallis, p = 0.22), the protein expression within PC group was significantly lower in early sepsis and gradually increased at late stages (Kruskal Wallis, p = 0.014)

2.4. Effect of PC concentrate on apoptosis-related protein expression Immunodetection of apoptosis-related proteins occurred at same brain regions as active caspase 3 (Fig. 6). Protein C concentrate administration increased bcl-2 protein immunodetection at 24 h and at 60 h post CLP compared to placebo (p = 0.001 in both time points, Fig. 7A). On the contrary bax protein expression was significantly decreased in neuronal cells at 6 h post CLP in PC treated animals compared to

Fig. 3 – Time distribution of necrosis in Neu-N (A) and GFAP (B) labeled cells from treated and untreated septic rats at 6, 12, 24, 36, 48 and 60 h post CLP (n = 5 at each time point). PC zymogen administration significantly reduced necrosis at the time points of administration compared to placebo. (Error bars represent 1SEM, *: denotes statistical significance)

122

BR A I N R ES E A RC H 1 2 6 4 ( 2 00 9 ) 1 1 9 –12 6

Fig. 4 – Immunohistochemical detection of active caspase 3 in brain sections from septic rats. (A) Cleaved caspase 3 immunoreactive hippocampal neurons (×100). (B) Nuclear localization of cleaved caspase 3 in neurons (arrows) from septic animal without treatment (× 200).

porcine cortex in experimental sepsis (Sharshar et al., 2003; Papadopoulos et al., 1999). Recent reports have shown that during hypoxia, APC has anti-apoptotic properties independent from its anticoagulant and anti-inflammatory properties in endothelial cells and in neurons both in vitro and in vivo via protease activated receptors 1 and 3 (Cheng et al., 2003; Guo et al., 2004). The present study provides data supporting that PC precursor administration reduces the apoptotic process in brain cells during sepsis. Furthermore, suggests that the antiapoptotic effect of PC is evident in both the caspase-8 and mitochondrial pathway of apoptosis as documented by the loss of caspase 8 and cytochrome c immunopositivity and the increase of bcl-2 immunoreactivity in brain cells from animals receiving PC.

Fig. 5 – Time distribution of positive cells for cleaved caspase 3. PC administration reduced the active caspase 3 positive cells in brain during the early sepsis. (Error bars represent 1SEM, *: denotes statistical significance)

PC administration suppressed active caspase 3 detection in areas known to be vulnerable in hypoxic and inflammatory insult in brain such as the choroid plexus, the Purkinje cells of the cerebellum and the CA1 hippocampal cells (Messaris et al., 2004; Semmler et al., 2005). The effect of Protein C administration was more pronounced at 6 h post CLP since a single infusion 90 min post procedure resulted in caspase 3, bax and cytochrome c suppression and reduced neuronal apoptosis and necrosis and astrocyte necrosis. Furthermore, a second neuroprotective effect regarding the mitochondrial pathway was also seen at 24 h with caspase 3 and cytochrome c reduction and bcl-2 elevation which was synchronously related to a significant reduction in neuronal apoptosis and astrocyte necrosis. The activation of the alternative caspase 8 apoptotic pathway was prevented at 12 h post CLP since at the specific time point a major decrease in caspase 8 and caspase 3 immunodetection was noticed related to a significant reduction in neuronal apoptosis. Finally a significant decrease in neuronal and astrocyte necrosis and a significant elevation of bcl-2 was observed 60 h post CLP. This is consistent with in vitro studies showing activated Protein C to inhibit both the caspase 8 dependent as well as the bax/bcl-2 mitochondrial biochemical pathway in cultured murine neurons following N-methyl-Daspartate (NMDA) and staurosporine induced apoptotic process (Cheng et al., 2003, Guo et al., 2004). Both pathways result in caspase 3 activation and execution of programmed cell death. The CLP model mimics the biphasic nature of septic syndrome in rats, showing a hyperactive phase at first (up to 16–24 h) followed by hypodynamic phase (Wichterman et al., 1980). PC administration altered the time distribution of cell death in neuronal cells resulting in a “U” shape reaching the lowest apoptotic value at 24 h. The half life of PC zymogen in humans is approximately 6 h which may explain that the cytoprotective effect lasted 12 h following last infusion. Accordingly, neuronal necrosis was significantly reduced in PC group at 6 and 12 h post CLP enhancing the neurocytoprotective effect of zymogen infusion. In addition PC protected astroglial cells from necrosis but not from apoptosis at 6 and 12 h, suggesting that the anti-apoptotic effect of Protein C is predominant on the neuronal cell population and the astrocytic cytoprotection may be secondary to neuronal cytoprotection.

BR A I N R ES E A RC H 1 2 6 4 ( 2 00 9 ) 1 1 9 – 1 26

123

Fig. 6 – Immunohistochemical detection of apoptosis-related proteins in brain sections from septic rats. (A) Hippocampal cells immunopositive for bax in a septic rat at 6 h post CLP (× 200). (B) Hippocampal cells from septic rat positive for bcl-2 treated with Protein C (× 400). (C) Choroid plexus with cytoplasmic immunolocalization of caspase 8 protein in a septic rat without treatment (×200). (D) Purkinje cells in the cerebellum of a septic rat without treatment positive for cytochrome c (×100).

In our study we have used the dosage of 100 IU/kg as used in humans since no available data has been performed in rats. The half life of Protein C zymogen in humans is approximately 6 h and that time interval was used for the three consecutive infusions. It seems from our results that the anti-apoptotic effect of Protein C ceased at 24 h which is 12 h following the last PC infusion. After that time point neuronal apoptosis was the same as the placebo. Thus the cytoprotective effect of Protein C in neuronal cells was sufficient in the early sepsis rather than late sepsis probably due to elimination of Protein C by rat metabolism. In clinical trials the use of activated Protein C is the standard treatment in septic patients with high risk of death but provides a significant risk of bleeding as a side effect (Dellinger et al., 2008). An alternative could be the use of modified APC with non-anticoagulant properties (Kerschen et al., 2007). On the other hand, the theoretical advantage of PC zymogen is its on site and on demand activation at regions of increased thrombin generation and APC generation ceases once thrombin generation is under control of anticoagulant mechanisms. To date the use of PC was evaluated in pediatric patients with meningococcal sepsis without major side effects (de Kleijn et al., 2003) and in septic adult patients with contraindication to APC restoring the coagulation imbalance (Baratto et al., 2008). In addition, recent studies by Feistritzer et al. (2006) showed that the protective signaling in endothelial cells is linked to PC activation. They demonstrated that an

efficient barrier enhancement is coupled with the endogenous PC activation pathway which gives the rationale for the use of Protein C zymogen in acquired Protein C deficiency. Recombinant human activated Protein C administration acts anti-apoptotically in human umbilical vein endothelial cells (O'Brien et al., 2007), in diabetic nephropathy by inhibiting glomerular apoptosis (Isemann et al., 2007) and in myocardial cells following ischemia–reperfusion (Pirat et al., 2007). Regarding sepsis APC's anti-apoptotic properties have been documented in circulating mononuclear cells (Bilbault et al., 2007), traumatized skeletal muscle during endotoxemia (Gierer et al., 2007), in lung after polymicrobial sepsis (Shires et al., 2007) and in developing rat brain following endotoxinemia (Yesilirmak et al., 2007). From the present study it cannot be distinguished whether Protein C acted via its activated form or the Protein C concentrate exerts the same anti-apoptotic effects as Activated Protein C. However published data does not support the latter hypothesis since Protein C zymogen per se failed to protect cortical neuronal and brain endothelial cells in vitro from NMDA and staurosporine induced apoptosis (Cheng et al., 2003). Thus the cytoprotective effect seen might be due to direct effect of the active form of PC in neuronal cells or secondary to overall cytoprotective properties of the active form of PC in other vital organs. In conclusion, our data revealed a strong relation between PC zymogen administration and inhibition of apoptosis in

124

BR A I N R ES E A RC H 1 2 6 4 ( 2 00 9 ) 1 1 9 –12 6

Fig. 7 – Time distribution of apoptosis related proteins in septic brain from treated and untreated animals (A–D). (A) Bax immunodetection was significantly decreased at 6 h while antiapoptotic bcl-2 significantly increased in PC treated animals compared to placebo (B). Cytochrome c detection was constant during early sepsis in PC treated animals and gradually increased afterwards (C). Protein C delayed caspase 8 immunoexpression as compared to caspase 8 expression in placebo animals (D). (Error bars show 1 SEM, *: denotes statistical significance).

brain sections from septic rats providing evidence of an in vivo anti-apoptotic role during sepsis. This inhibition of apoptosis may be due to direct anti-apoptotic properties of the Protein C molecule recently recognized or as a secondary effect of its anticoagulant and anti-inflammatory properties.

4.

Experimental procedures

4.1.

Animals

Male pathogen-free Wistar rats (n = 60) weighting 250 g in average were used in the experiments. Animals were housed for at least 7 days before manipulations. Animal care and handling were performed in accordance with the guidelines of the National Institute of Health, the European Guidelines for

Animal Experimentation, under the supervision of the Greek Ministry of Health and were approved by the Animal Welfare Committee of Athens University Medical School.

4.1.1. Experimental sepsis induced by cecal ligation and puncture (CLP) Sepsis was induced by the well established CLP model (Wichterman et al., 1980). Anesthesia was performed by ketamine (60 mg/ kg body weight) and xylazine (3 mg/kg body weight) intramuscularly. A 3-cm ventral midline incision was performed, the cecum was exposed, ligated just distal to the ilieocecal valve to avoid intestinal obstruction and punctured twice with an 18gauge needle. Cecum was then gently compressed to extrude a small amount of cecal contents and returned to the abdomen. The abdominal incision was then closed in layers and animals received 3 ml/100 g body weight saline solution subcutaneously

BR A I N R ES E A RC H 1 2 6 4 ( 2 00 9 ) 1 1 9 – 1 26

in order to maintain normal circulation. The injected saline was completely absorbed in 10 min.

4.2.

Experimental design

After the induction of sepsis animals were randomly assigned to two major groups: the CLP + PC (n = 30) and the CLP group (n = 30). Within 60 min post CLP each animal of the assigned groups received intravenously either 100 IU/kg of Human Protein-C (Ceprotin, ATC Code: B01 AX, Baxter AG, Hellas) or 30 μl of normal saline. The injection of PC or placebo was repeated 7 and 13 h post CLP. Five sham operated animals were served as controls. Animals from both groups were euthanised at 6, 12, 24, 36, 48 and 60 h (n = 5 at each time point) post procedure. Immediately after each rat's death, brain was removed and separated in two symmetric halves through a midline sagittal incision. One half was stored at 80 °C and the remaining immersed in formalin buffer for paraffin fixation.

4.2.1.

Cell preparation for flow cytometry

Brain frozen tissue were cut into four to five pieces approximately 2.5 mm3 in size and placed into a Medicon (BD Biosciences, San Jose, CA) wetted with 1.0 ml of 1 × PBS. The Medicon is a polystyrene chamber containing an impeller and an immobile stainless steel screen with approximately 100 hexagonal holes, each surrounded by six microblades. The Medicon was inserted into the Medimachine (BD Biosciences) and was operated at 100 rpm for 15 s. The cell suspension was filtered using a Filcon (BD Biosciences), a disposable filter device, and washed twice with 5 ml of 1 × PBS. The cells were then counted with a hemocytometer and adjusted to a concentration of approximately 1 × 107 cells/ml.

4.2.2. Two-step GFAP immunocytochemical reaction using FITC-conjugated goat anti-rabbit IgG (whole fragment) Expression of GFAP in the cell suspension was measured by an indirect labeling procedure using FITC-conjugated goat antirabbit IgG. After fixation and permeabilization using 1% paraformaldehyde–1% saponin in PBS for 20 min at 4 °C, cells were incubated for 60 min at room temperature with 50 μl per 106 cells anti-GFAP primary antibody (Dako, dilution 1:150). The cells were then washed with PBS and treated with 50 μl anti-rabbit IgG-FITC conjugate antibody (Sigma, dilution 1:100) as a secondary antibody. Following additional 90-min incubation in the dark at room temperature with gently agitation, cells were washed with PBS. Finally, cells were incubated with 10 μl of 7-aminoactinomycin D (7AAD) (BD) for 15 min at room temperature before analyzed by flow cytometry. 7-Aminoactinomycin D (7AAD) intercalates into double-stranded nucleic acids. It is excluded by viable cells but can penetrate cell membranes of dying or dead cells.

4.2.3. One-step Neu-N immunocytochemical reaction using Zenon™ R-phycoerythrin rabbit IgG, Labeling Kit (Invitrogen) For Neu-N immunocytochemical reaction, 1 μg anti-Neu-N antibody (Chemicon) per 106 cells was used. Five microliters of Zenon rabbit IgG labeling reagent, which contains a fluorophore-labeled Fab fragment, were mixed with 1 μg (5 μl) of the antibody solution. The mixture was further incubated for 20 min at room temperature and applied to the cell sample previously

125

fixed and permeabilized using 1% paraformaldehyde–1% saponin in PBS for 20 min at 4 °C and washed with PBS. The samples were incubated for 20 min at room temperature, 10 μl of 7AAD was then added (BD) and samples were further incubated for another 15 min before analyzed by flow cytometry. For all staining procedures, a negative control was processed in parallel with each sample.

4.2.4.

Flow cytometry analysis

Measurements of fluorescence intensity were performed on a COULTER® EPICS® XL-MCL™. Fluorochromes were excited with a single 488-nm Argon laser. FITC fluorescence was detected through a 530-nm 15-nm band pass filter and 7AAD fluorescence was detected through a 650-nm long pass filter. A broad gate was established for data collection using a Cartesian plot of FSC (forward scatter) versus SSC (side scatter) to exclude cellular fragments and debris from the cell population. For each sample, at least 10,000 events were counted.

4.3. Immunohistochemical determination of apoptosis regulating proteins Immunohistochemical staining was performed on 5 μm sections of the tissue paraffin embedded blocks. Paraffin sections were mounted on superfrost glass slides, deparaffinized and rehydrated in a graded ethanol series, and then subjected to microwave antigen retrieval (10 min in 0.1 M citrate acid buffer solution, pH 6). Endogenous peroxidase activity was blocked using 0.3% hydrogen peroxide. Sections were incubated at 4 °C overnight with the following primary antibodies at the dilutions specified: bcl-2 (Santa-Cruz, Biotechnologies) dilution 1:12, cleaved caspase-3 (Cell Signaling Technology) dilution 1:100, caspase-8 (Neomarkers, CA) dilution 1:150, cytochrome c (Neomarkers, CA) dilution 1:100 and polyclonal antibody for bax (Oncogene, Campridge, MA) dilution 1:30. Immunohistochemical staining was performed using the rabbit or mouse DAKO ChemMateTM EnVisionTM system. Sections were then counterstained with Mayer hematoxylin and then dehydrated, cleared and mounted. Appropriate positive controls were used according to the manufacturer's guide for each antibody. Two independent pathologists by means of a light microscope evaluated the slides. Those sections where less than 10% of cells were stained were considered negative (−).

4.4.

Statistical analysis

Data is presented as mean ± SEM values. Non parametric statistics was used for the analysis of quantitative variables. Differences between two groups were analyzed using Mann Whitney. The time distribution of a variable within group was analyzed using the Kruskal Wallis test. Differences were considered to be statistically significant when the two sided probability value was