Background
Intracerebral hemorrhage (ICH) is a subtype of stroke that carries high morbidity and mortality. ICH causes brain injury through primary physical disruption of adjacent tissue and the mass effect and secondary injury such as brain edema and inflammation [
1,
2]. Evidence from preclinical and clinical studies suggests that inflammatory mechanisms contribute to the progression of secondary brain injury after ICH [
1,
3]. Inflammatory cells that have the potential to promote hemorrhagic brain damage include blood-derived leukocytes and macrophages, resident microglia, astrocytes, and mast cells. Molecular components of the inflammatory response include prostaglandins, chemokines, cytokines, extracellular proteases, and reactive oxygen species [
3‐
5].
Prostaglandin E
2 (PGE
2) is a key component in the initiation and propagation phases of the inflammatory process induced after various types of brain injury [
6‐
8]. It is synthesized from arachidonic acid in two enzymatic steps. Cyclooxygenase (COX) catalyzes the conversion of arachidonic acid to PGH
2, which is then converted to PGE
2 by PGE synthase (PGES). COX is present in two isoforms, constitutive COX-1 and inducible COX-2. Under normal conditions, COX-1, which is expressed in numerous tissues, has been classically considered to be the isoform principally responsible for homeostatic prostaglandin synthesis [
9]. In contrast, COX-2 is primarily expressed in glutamatergic neurons, wherein it is thought to participate in synaptic signaling [
10]. Three major isozymes of PGES have been isolated: microsomal PGES (mPGES)-1, mPGES-2, and cytosolic PGES (cPGES). Of these, mPGES-2 and cPGES are constitutively expressed in various cells and tissues; only mPGES-1 can be induced by pro-inflammatory stimuli and in ischemic stroke models [
7,
11‐
13]. A recent study showed that COX-2 and mPGES-1 are co-induced by excess glutamate in ischemic brain and that they act together to exacerbate stroke injury by amplifying PGE
2 production [
14].
Although COX/prostaglandin signaling has been implicated in the pathologic progression of ischemic stroke [
15‐
17], research on their role in ICH is limited. To our knowledge, no reports have been published regarding expression of COX-1 and PGES isoforms in ICH. Two studies have been published on COX-2 expression in blood models of ICH in rats, but the results are conflicting. One showed COX-2 immunoreactivity to be increased from 6 h to 3 days after ICH [
18], whereas the second showed a transient increase in COX-2 from 1 to 3 h after ICH followed by a significant down-regulation [
19]. To lay a foundation for understanding the roles of COX and PGES isoforms in ICH pathology, here we characterized the expression and cellular localization of COX-1, COX-2, mPGES-1, mPGES-2, and cPGES from 5 h to 3 days after collagenase-induced ICH in mice.
Materials and methods
Animals
This study was conducted in accordance with the National Institutes of Health guidelines for the use of experimental animals. Experimental protocols were approved by the Johns Hopkins University Animal Care and Use Committee. A total of 56 C57BL/6 male mice (25-35 g) were obtained from Charles River Laboratories (Wilmington, MA). All efforts were made to minimize the numbers of animals used and ensure minimal suffering.
ICH model
The procedure for modeling ICH by intrastriatal injection of collagenase was adapted to mice from an established rat protocol [
20] and has been described previously [
21,
22]. Briefly, C57BL/6 mice weighing 22-35 g were anaesthetized by isoflurane (3.0% for induction, 1.0% for maintenance) and ventilated with oxygen-enriched air (20%:80%) via a nose cone. We then injected mice in the left striatum with collagenase VII-S (0.075 U in 500 nL saline, C2399, Sigma, St. Louis, MO) at the following stereotactic coordinates: 1.0 mm anterior and 2.2 mm lateral of the bregma, 2.7 mm in depth. Collagenase was delivered over 5 min. The needle was held in place for an additional 5 min to prevent reflux. Sham-operated mice were injected with saline only. Rectal temperature was monitored and maintained at 37.0 ± 0.5°C with a heating pad throughout the experimental and recovery periods. This procedure resulted in reproducible lesions that were mostly restricted to the striatum.
Preparation of brain slices
At 5, 24, or 72 h after ICH, 3 mice per group were anesthetized deeply with phenobarbital and perfused transcardially with phosphate-buffered saline (PBS), pH 7.4. Sham-operated control mice (n = 2 at each time point) were perfused similarly. The brains were harvested and fixed with 4% paraformaldehyde overnight, cryoprotected in serial phosphate-buffered sucrose solutions (20%, 30%, and 40%) at 4°C, and then cut into 20-μm sections with a cryostat.
Hemorrhagic injury volume
Mice (n = 5) were euthanized at 24 h after ICH. The entire brain of each mouse was cut into 50-μm sections with a cryostat. Sections were stained with Luxol fast blue (for myelin) and Cresyl Violet (for neurons) before being quantified for grey and white matter injury with SigmaScan Pro software (version 5.0.0 for Windows; Systat, San Jose, CA). The injury volume in cubic millimeters was calculated by multiplying the thickness by the damaged areas of each section [
21].
Immunofluorescence
Immunofluorescence was carried out as described previously [
23,
24]. Briefly, free-floating sections were washed in PBS for 20 min, blocked in 5% normal goat serum, and incubated overnight at 4°C with primary antibodies: rabbit anti-COX-1 (1:250; Catalog No. 160109, Cayman Chemical, Ann Arbor, MI); rabbit anti-COX-2 (1:250; Catalog No. 160106, Cayman Chemical); rabbit anti-mPGES-1 (1:250; Catalog No. 160140, Cayman Chemical); rabbit anti-mPGES-2 (1:250; Catalog No. 160145, Cayman Chemical); rabbit anti-cPGES (1:250; Catalog No. 10209, Cayman Chemical); mouse anti-NeuN (1:500; Catalog No. MAB377, Chemicon, Temecula, CA), specific for neurons; rat anti-GFAP (1:250; Catalog No. 13-0300, Invitrogen, Carlsbad, CA), specific for astrocytes; and rat anti-CD11b (1:1000; Catalog No. MCA711G, Serotec, Raleigh, NC), specific for microglia/macrophages. The sections were then washed with PBS and incubated with Alexa-488 (1:1000; Molecular Probes, Eugene, OR)- and/or Cy3 (1:1000; Jackson Labs, West Grove, PA)-conjugated secondary antibodies for 60 min. Stained sections (n = 3/mouse) were examined with a fluorescence microscope; the images were captured from the frontoparietal cortex and striatum and analyzed by SPOT advanced image software (Diagnostic Instruments Inc., Sterling Heights, MI). Control sections were processed identically, except that primary antibodies were omitted. Control sections lacked specific staining. The specificity of the antibodies against COX-1, COX-2, mPGES-1, mPGES-2, and cPGES was further confirmed by using the corresponding blocking peptides (Cayman Chemical).
To quantify the number of immunoreactive cells labeled with COX-1, COX-2, mPGES-1, mPGES-2, and cPGES, three sections per mouse (from the injection site and 360 μm on each side; n = 3 mice/group) were analyzed in the perihematoma region of the striatum. This region was defined within one 20× field that corresponded to ~460 μm from the edge of the hematoma. Positively stained cells were counted in three comparable, randomly selected 30× microscopic fields. The numbers of immunoreactive cells from nine locations per mouse (3 fields per section × 3 sections per mouse) were averaged and expressed as positive cells per field (30×).
Statistics
All data are expressed as means ± SD. The statistical comparisons among multiple groups were made using one-way ANOVA followed by Bonferroni correction. Statistical significance was set at P < 0.05.
Discussion
To our knowledge, this is the first systematic study performed to characterize the expression and cellular localization of COX and PGES isozymes in the hemorrhagic brain of mice. Using immunofluorescence staining, we observed constitutive expression of COX-1, mPGES-2, and cPGES in neurons; COX-1 was also constitutively expressed in microglia. In contrast, COX-2 and mPGES-1 immunoreactivity, which was minimal in the normal brain, underwent distinct time-dependent changes in neurons and astrocytes of the perihematomal region during the first 3 days post-ICH. Our data support the premise that the COX/PGES signaling pathway contributes to ICH pathology.
COX-1 is constitutively expressed in most tissues [
9]. We demonstrated for the first time that COX-1 is constitutively expressed in neurons and microglia in the hemorrhagic brain. Although involvement of COX-1 in ICH pathology has not been studied, the evidence that COX-1 is produced in microglia of the perihematomal region implies a toxic role of COX-1 in the pathophysiology of the disease. This hypothesis is based on the fact that activated microglia/macrophages are the major sources of proinflammatory mediators [
1,
3] and that inhibition of microglial activation before or early after ICH decreases neuronal death and improves neurologic function [
21,
25]. We therefore propose that microglial COX-1 might immediately initiate synthesis of prostaglandins in response to microglial activation and could be considered one of the major players in mediating neuroinflammation after ICH.
In the normal brain, COX-2 is constitutively expressed in neurons of the cortex, hippocampus, and striatum [
8]. COX-2 is mainly induced in response to inflammatory stimuli; deletion of the COX-2 gene or selective COX-2 inhibition reduces infarction volume and neuronal death after cerebral ischemia [
26‐
28]. To our knowledge, only two studies have investigated post-ICH COX-2 expression, and the results were conflicting [
18,
19]. Zhao et al. [
19] demonstrated that COX-2 mRNA and protein were increased within 3 h after ICH and that COX-2 immunoreactivity was increased in blood vessels and neurons in the perihematomal region at 4 h; the increase in immunoreactivity was transient, followed by a significant down-regulation at days 1 and 3. In the present study, we found that the immunoreactivity of COX-2 in astrocytes increased gradually in the perihematomal region from 5 h to 3 days, whereas neuronal COX-2 increased only transiently at 5 h after ICH. Most of our results are consistent with the findings by Gong et al. [
18], except that we found COX-2 to be increasingly induced in astrocytes from 1 to 3 days, whereas Gong et al. reported that COX-2 was induced in endothelial cells, perivascular cells, and infiltrating leukocytes at 1 day after blood infusion. The reason for the discrepancy between the previous studies and our own is not clear, but it may be a result of differences in ICH models and species used and the size of the intrastriatal hematoma formed. In the two previous studies, investigators modeled ICH by injecting rats intrastriatally with differing amounts of autologous blood. In contrast, we used a collagenase-induced ICH model in mice that may cause gradual hematoma growth over the first few hours with subsequent inflammation. Our data suggest that astrocytic COX-2 in concert with microglial COX-1 might contribute to collagenase-induced post-hemorrhagic neuroinflammation. More studies are warranted to understand whether the functions of neuronal and astrocytic COX-2 differ after ICH.
Induction of mPGES-1 expression has been observed in various conditions, such as inflammation, fever, pain, tissue repair, and cancer, in which COX-2-derived PGE
2 plays a critical role [
29]. In the ischemic brain, mPGES-1 and COX-2 are both induced in neurons, microglia, and endothelial cells in the ipsilateral cerebral cortex and striatum [
7]. It has been confirmed that mPGES-1 and COX-2 are co-localized and co-induced in the infarct region of the cortex, and it has been suggested that they act together to exacerbate stroke injury [
14]. At 1 day post-ICH in our model, mPGES-1 expression in the ipsilateral cortex was elevated primarily in neurons whereas in the perihematomal region, it was elevated in astrocytes. Astrocytic expression continued to increase for at least 3 days. However, we observed no apparent changes in the expression of neuronal mPGES-2 or cPGES. The different cellular expression profiles for mPGES-1 between cortex and striatum suggests that mPGES-1 may be involved in different signaling pathways within the cortical neurons and striatal astrocytes after ICH. In line with the results from a lipopolysaccharide-induced inflammation model [
30], we found that the induction of COX-2 protein expression was more rapid than that of mPGES-1 in the hemorrhagic brain. These results suggest that the sequential up-regulation and co-induction of COX-2 and mPGES-1 in astrocytes in the perihematomal region might contribute to inflammation-mediated secondary brain injury after ICH, possibly through excessive PGE
2 production.
In conclusion, our data provide novel evidence that COX-1, mPGES-2, and cPGES are constitutively expressed in the hemorrhagic brain; COX-2 is induced early in neurons and later in astrocytes. Although neuronal COX-2 is induced earlier than astrocytic COX-2, the latter is induced in parallel with astrocytic mPGES-1. Together, our data suggest that microglial COX-1, neuronal COX-2, astrocytic COX-2, and astrocytic mPGES-1 may work sequentially to affect ICH outcomes. Based on our previous observations that neuroinflammation affects the normal function of the entire brain in patients with lethal ICH [
31], these findings have implications for efforts to develop anti-inflammatory strategies that target the COX/PGES pathway to reduce ICH-induced secondary brain damage. Indeed, a recent study showed that inhibition of COX-2 attenuated inflammation, neuronal death, and gliosis and promoted long-term recovery in motor function and myelination in rabbit pups with intraventricular hemorrhage [
32].
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
TW, HW, and Jessica W carried out the ICH model and immunofluorescence staining and helped draft the manuscript. Jian W conceived of the study, participated in its design and conduct, and helped draft the manuscript. All authors read and approved the final manuscript.