Long-term efficacy and safety of ixekizumab in Japanese patients with erythrodermic or generalized pustular psoriasis: subgroup analyses of an open-label, phase 3 study (UNCOVER-J).

BACKGROUND: Erythrodermic and generalized pustular psoriasis are rare, difficult to treat forms of psoriasis. In previous reports, we documented 24- and 52-week findings of an open-label, phase 3 trial (UNCOVER-J) of ixekizumab in Japanese patients with erythrodermic or generalized pustular psoriasis; most patients responded to treatment and maintained response through 52 weeks. OBJECTIVE: To assess the long-term (>3 years) efficacy and safety of ixekizumab in Japanese patients with erythrodermic or generalized pustular psoriasis. METHODS: These subgroup analyses were of a partial population of patients from UNCOVER-J (NCT01624233; Sponsored by Eli Lilly and Company), specifically those with erythrodermic psoriasis (N = 8) or generalized pustular psoriasis (N = 5). These patients received 160 mg ixekizumab at Week 0, ixekizumab 80 mg every 2 weeks through Week 12, and ixekizumab 80 mg every 4 weeks thereafter up to Week 244. This regimen is consistent with the regimen approved in Japan for plaque, erythrodermic, and generalized pustular psoriasis and psoriatic arthritis. Efficacy assessments included Global Improvement Score (GIS), Psoriasis Area and Severity Index (PASI), dermal symptoms (for patients with generalized pustular psoriasis), Dermatology Life Quality Index (DLQI) and Itch Numeric Rating Scale (NRS). Safety assessments included treatment-emergent adverse events and adverse events of special interest. RESULTS: Most patients had a GIS of resolved or improved from Week 12 onwards, and all patients had early and sustained improvement in PASI and dermal symptom (generalized pustular psoriasis only) scores. Mean improvements in DLQI and Itch NRS at Week 12 were sustained through Week 244. Ixekizumab was well tolerated over 3 years of treatment in patients with erythrodermic psoriasis or generalized pustular psoriasis, and no new safety concerns were identified. CONCLUSION: These findings suggest that ixekizumab can be an effective long-term treatment option for erythrodermic or generalized pustular psoriasis.

A new objective histological scale for studying human photoaged skin.

BACKGROUND: A quantitative understanding of the histological alteration of the skin is important for assessing the severity of photoaging. METHODS: We performed Elastica-van Gieson staining and immunohistochemistry for decorin on 34 facial skin sections. We evaluated the alteration of collagen fibers and decorin (a modulator for collagen fibrillogenesis), according to the 5 grades of morphological change in elastic fibers that was established by Kligman (1969). The objectivity of a stage (Stages I-VI), which was established in this study, was evaluated using weighted kappa statistical analysis based on the degree of agreement in stage determination by 11 observers using a blind procedure. Correlation between the crow's-feet-area wrinkles grades of another 26 women and stages was also analyzed. RESULTS: The initial alteration of elastic fibers was observed in the deep dermis. Decorin was not detected in very severely altered skin. Based on the combination of changes in the elastic fibers, collagenic fibers, and decorin, skin tissues were categorized into 6 stages according to severity. The statistical analysis showed almost perfect agreement between observers. Significant positive correlation between stages and wrinkle scores was found. CONCLUSIONS: We propose a new objective histological scale that is useful for assessing the severity of photoaging.

Involvement of stem cell factor and its receptor tyrosine kinase c-kit in pain regulation.

The c-kit receptor tyrosine kinase is expressed in a subpopulation of small- and medium-sized neurons of the dorsal root ganglia (DRG) and in the superficial layer of the spinal cord. Stem cell factor (SCF), a ligand of the c-kit receptor, induces neurite outgrowth from DRG and supports the survival of c-kit-expressing neurons. To clarify the possible function of the SCF/c-kit receptor system in the adult animal, we investigated the expression of c-kit receptor in the spinal cord and DRG in relation to pain by using H2C7, a newly developed anti-c-kit monoclonal antibody. S.c. and intrathecal injection of SCF markedly reduced the paw withdrawal threshold to mechanical stimuli and intrathecal SCF at 10 pg maximally induced mechanical allodynia in conscious mice. Intrathecal SCF also reduced the paw withdrawal latency to heat stimuli significantly but transiently. The c-kit receptor was co-expressed in 58.4% of calcitonin gene-related peptide (CGRP) -positive, but only 5.1% of isolectin B4-positive, DRG neurons. In the spinal cord, the c-kit receptor was detected in the superficial layer of the dorsal horn and co-localized there with CGRP in central terminals of DRG neurons. Selective elimination of unmyelinated C-fibers by neonatal capsaicin treatment resulted in marked reduction of the c-kit receptor and CGRP expression in the superficial layer of the spinal cord. Cell-size profiles showed that c-kit receptor expression was significantly up-regulated and down-regulated in medium-sized DRG neurons after neonatal capsaicin treatment and nerve injury, respectively. These results suggest that the c-kit receptor is mainly expressed in peptidergic small-sized DRG neurons and may be involved in pain regulation both peripherally and centrally.

Different roles of nitric oxide synthase-1 and -2 between herpetic and postherpetic allodynia in mice.

We investigated using the mice role of nitric oxide synthase (NOS) in the spinal dorsal horn in herpetic and postherpetic pain, especially allodynia, which was induced by transdermal inoculation of the hind paw with herpes simplex virus type-1 (HSV-1). The virus inoculation induced NOS2 expression in the lumbar dorsal horn of mice with herpetic allodynia, but not postherpetic allodynia. There were no substantial alternations in the expression level of NOS1 at the herpetic and postherpetic stages. Herpetic allodynia was significantly inhibited by i.p. administration of the selective NOS2 inhibitor S-methylisothiourea, but not the selective NOS1 inhibitor 7-nitroindazole. NOS2 expression was observed around HSV-1 antigen-immunoreactive cells. On the other hand, postherpetic allodynia was significantly inhibited by i.p. administration of 7-nitroindazole, but not S-methylisothiourea. The activity of reduced nicotinamide adenine dinucleotide phosphate diaphorase, an index of NOS1 activity, significantly increased in the laminae I and II of the lumbar dorsal horn of mice with postherpetic allodynia, but not mice without postherpetic allodynia. The expression level of NOS1 mRNA in the dorsal root ganglia was similar between mice with and without postherpetic allodynia. The results suggest that herpetic and postherpetic allodynia is mediated by nitric oxide in the dorsal horn and that NOS2 and NOS1 are responsible for herpetic and postherpetic allodynia, respectively. It may be worth testing the effects of NOS2 and NOS1 inhibitors on herpetic pain and postherpetic neuralgia in human subjects, respectively.

Phosphorylation of cAMP response element-binding protein in hippocampal neurons as a protective response after exposure to glutamate in vitro and ischemia in vivo.

Although accumulating evidence indicates that cAMP response element-binding protein (CREB) phosphorylation mediates not only synaptic plasticity but also survival of certain neurons, it remains uncertain whether CREB phosphorylation induced after metabolic insult leads to CRE-mediated gene transcription and is involved in cell survival or not. In the present study, we clarified that (1) CREB phosphorylation and ischemic tolerance induced after preconditioning ischemia in the hippocampal neurons was abolished by MK801 administration in gerbil global ischemia model, (2) CREB phosphorylation induced after exposure to glutamate in cultured neurons was inhibited by removal of extracellular calcium, by MK801 and by an inhibitor of calcium-calmodulin-dependent protein kinase (CaMK) II and IV, (3) inhibitor of CaMK II-IV or CRE-decoy oligonucleotide suppressed upregulation of BCL-2 expression and accelerated neuronal damage after exposure to glutamate, and (4) CREB phosphorylation induced in the hippocampal neurons after ischemia and in cultured neurons after exposure to glutamate was followed by CRE-mediated gene transcription in transgenic mice with a CRE-LacZ reporter. Our results suggest that CREB phosphorylation in neurons after ischemia and exposure to glutamate is induced by NMDA receptor-gated calcium influx and subsequent activation of CaMK II-IV and that CREB phosphorylation after metabolic stress might show a neuroprotective response through CRE-mediated gene induction.

Involvement of Rho-kinase in inflammatory and neuropathic pain through phosphorylation of myristoylated alanine-rich C-kinase substrate (MARCKS).

Myristoylated alanine-rich C-kinase substrate (MARCKS) is a major in vivo substrate for protein kinase C in the brain and has been implicated in cellular processes associated with cytoskeletal restructuring such as synaptic trafficking and neurotransmitter release. A phosphorylation-site specific antibody against Ser159-phospho-MARCKS (pS159-Mar-Ab) revealed that MARCKS is phosphorylated at Ser159 by Rho-kinase and that its phosphorylation is inhibited by the Rho-kinase specific inhibitor H-1152. Since the function of MARCKS is regulated by phosphorylation at multiple sites, here we examined the involvement of Rho-kinase in relation to phosphorylation of MARCKS at Ser159 in inflammatory and neuropathic pain by H-1152. When intrathecally administered 10 min before s.c. injection of formalin, H-1152 at 10 and 100 ng attenuated the second-phase, but not the first-phase, pain-like behaviors in the formalin test. Neuropathic pain induced by selective L5 spinal nerve transection was also relieved by intrathecal injection of H-1152. Nitric oxide synthase activity visualized by NADPH diaphorase histochemistry increased in the superficial layer of the spinal cord 30 min after formalin injection and 7 days after nerve transection, which were blocked by H-1152. Phosphorylation of MARCKS at Ser159 was detected in the spinal cord by pS159-Mar-Ab and the level of phosphorylation increased in the superficial layer after nerve transection. In contrast, immunoreactivities of neuronal nitric oxide synthase and MARCKS did not change significantly in the spinal cord before and after nerve transection. Taken together, the present study demonstrates that Rho-kinase is involved in inflammatory pain and the maintenance of neuropathic pain through phosphorylation of MARCKS at Ser159.

PRKAR1A gene mutation in patients with cardiac myxoma.

BACKGROUND: PRKAR1A gene encodes the type 1A regulatory subunit of protein kinase A. The mutation of this gene causes Carney complex which is an autosomal dominant multiple neoplasia syndrome characterized by spotty pigmentations, endocrine overactivity and cardiac myxoma. We hypothesized that cardiac myxoma may be associated with PRKAR1A gene mutation and determined whether mutation in the PRKAR1A gene is the cause of familial and sporadic cardiac myxoma. METHODS: We studied seven patients (three males and four females) with cardiac myxoma. Two of them had familial cardiac myxoma complicated with Carney complex. The other five patients were characterized as sporadic cardiac myxomas. We analyzed the PRKAR1A gene of all patients by the polymerase chain reaction (PCR)-single-strand conformation method, followed with direct sequence analysis. RESULTS: We identified a novel mutation (494delTG) in exon 4A of the PRKAR1A gene in the patients with Carney complex. A 16-year-old proband had a left atrial myxoma, pituitary adenoma and skin pigmentation. His father also had left atrial myxoma and skin pigmentation. In contrast, no mutations in the PRKAR1A gene were identified in the other five patients with sporadic cardiac myxomas. CONCLUSIONS: These results suggest that mutation of the PRKAR1A gene may be associated with familial cardiac myxoma in Carney complex but may not be associated with sporadic cardiac myxoma.

Cerebral ischemia after bilateral carotid artery occlusion and intraluminal suture occlusion in mice: evaluation of the patency of the posterior communicating artery.

Cerebral ischemia models using mice have drawn increasing attention, particularly because of the availability of transgenic animals. However, the variability of intracranial vasculature at the circle of Willis in mice can influence the degree of ischemia in both the bilateral common carotid artery (CCA) occlusion and intraluminal suture occlusion models. We have developed a method to predict the extent of the anastomosis between carotid and vertebrobasilar circulation in three mouse strains (C57BL/6, CBA, and DBA/2) by measuring cortical microperfusion with laser Doppler flowmetry during a 1-minute occlusion of both CCA. When animals showed residual cortical microperfusion of less than 12% during bilateral CCA occlusion, the mice showed absence of functional anastomosis, developed ATP depletion in the frontal cortex during occlusion, and had ischemic neuronal death in the hippocampus and caudoputamen after occlusion for 15 minutes and recirculation for 7 days. Furthermore, those mice exhibited decreased local cerebral blood flow and associated ischemic neuronal death in the hippocampus, within the territory supplied by the posterior cerebral artery, with the intraluminal suture occlusion model. The current study demonstrates the need for assessment of intracranial vasculature in each animal by measuring cortical microperfusion during temporary occlusion of both CCA, no matter whether cerebral ischemia is produced by bilateral CCA occlusion or intraluminal suture occlusion in transgenic mice.

Deficiency of intercellular adhesion molecule 1 attenuates microcirculatory disturbance and infarction size in focal cerebral ischemia.

Recent evidence has shown crucial roles for cell-adhesion molecules in inflammation-induced rolling, adhesion, and accumulation of neutrophils in tissue. Intercellular adhesion molecule-1 (ICAM-1) is one of these adhesion molecules. Previous studies have shown marked reduction in the size of infarction after focal cerebral ischemia by depletion of granulocytes and administration of the antibody against ICAM-1. In the present study we investigated the role of ICAM-1 in the size of ischemic lesions, accumulation of granulocytes, and microcirculatory compromise in focal cerebral ischemia by using ICAM-1-knockout mice. Ischemic lesions were significantly mitigated in knockout mice after permanent and transient focal ischemia, even though the number of granulocytes in the infarcted tissue was almost the same between knockout and wild-type mice. Depletion of granulocytes further decreased the size of ischemic lesions after transient focal ischemia in ICAM-1-knockout mice. Microcirculation was reduced after focal ischemia, but it was better preserved in the cerebral cortex of knockout mice than that of wild-type mice. The present study demonstrated that ICAM-1 played a role in microcirculatory failure and subsequent development and expansion of infarction after focal cerebral ischemia. However, it is highly unlikely that ICAM-1 played a key role in accumulation of granulocytes after focal cerebral ischemia.

Extended neuronal protection induced after sublethal ischemia adjacent to the area with delayed neuronal death.

In the present study, we investigated whether neurons adjacent to an ischemic lesion acquire tolerance against subsequent ischemia or not. We initially used unilateral hemispheric ischemia for 3 min in gerbils to produce an ischemic lesion confined to the unilateral CA1 sector, and the presence of tolerance was examined in the adjacent CA3 sector through transient global ischemia by occlusion of both common carotid arteries. Attenuation of neuronal damage was clearly observed in neurons in the CA3 sector adjacent to the ischemic lesion in the CA1 sector. The phenomenon lasted for up to two weeks after the initial hemispheric ischemia, but was no longer present two months later. Reactive astrocytes as identified by the presence of glial fibrillary acidic protein were visible in the CA3 hippocampus four days and two weeks after hemispheric ischemia, but they were scarce two months later. Expression of heat shock protein 72 in the CA3 neurons was observed four days after hemispheric ischemia, but the reaction returned to the control level two weeks later. In conclusion, the present study showed that tolerance in the neurons adjacent to an ischemic lesion could be sustained at least for two weeks, and raised the possibility that reactive astrocytes might contribute to the extended tolerance in neurons.

Ischemic tolerance in hippocampal CA1 neurons studied using contralateral controls.

We induced ischemic tolerance unilaterally in gerbil hippocampus using the contralateral hippocampus as control. Ischemia for 2 min of right common carotid occlusion was reversible but sufficient to cause heat-shock protein 70 production in CA1 neurons. This pretreatment given four days prior to occlusion of both common carotids for 5 min, but not at longer preceding intervals, induced tolerance in right CA1 neurons. Neuroprotection was still evident two months after the 5 min occlusion. Adenosine triphosphate content and immunoreactive microtubule associate protein 2 in the hippocampus showed that the 5 min ischemic insult was essentially equal in both hemispheres. Repetitive pretreatments at two day intervals caused almost complete protection of CA1 neurons against subsequent 5 min ischemia, while a single pretreatment showed 80% protection. However, the increase in heat-shock protein 70 with repeated pretreatments was not significantly more than with one pretreatment. We concluded that true ischemic tolerance was induced by ischemic stress itself, was long-lasting, was not due to mitigation of subsequent ischemia, and was augmented by repetition without further increase of heat-shock protein 70.

DNA cleavage and proteolysis of microtubule-associated protein 2 after cerebral ischemia of different severity.

We report temporal profiles of cytoplasmic proteolysis and genomic DNA cleavage after cerebral ischemia of different severity in gerbils. Global forebrain ischemia by bilateral common carotid artery occlusion for 5 min with reperfusion, severe unilateral hemispheric ischemia by unilateral common carotid artery occlusion for 30 min with reperfusion, and complete ischemia by decapitation were used. The hippocampus was examined for proteolysis by using immunohistochemistry for microtubule-associated protein 2, DNA cleavage by using in situ nick-end labelling, and nuclear morphology by Hematoxylin staining. During evolution of delayed neuronal death after transient forebrain ischemia, loss of the immunoreaction for microtubule-associated protein 2 occurred almost in parallel with DNA cleavage in the CA1 region. In contrast, disappearance of the immunoreaction for microtubule-associated protein 2 was much faster than genomic DNA cleavage after unilateral hemispheric ischemia and reperfusion. The microtubule-associated protein 2 immunoreactivity was completely lost before development of changes in nuclear morphology or DNA cleavage after complete ischemia. The present study demonstrated the differences between necrosis and delayed neuronal death, but the nuclear morphology in the latter was not exactly the same as seen in apoptosis. Some elements of both necrotic and apoptotic machineries may work following transient ischemia, and the degree of ischemic insult may determine the character of cell death process.

Ischemic damage and subsequent proliferation of oligodendrocytes in focal cerebral ischemia.

In order to achieve a better understanding of the pathophysiology of ischemic white matter lesions, oligodendrocytic degeneration and subsequent proliferation were examined in the mouse model of middle cerebral artery occlusion. In situ hybridization histochemistry for proteolipid protein messenger RNA was employed as a sensitive and specific marker of oligodendrocytes, and immunohistochemistry for myelin basic protein was used as a compact myelin marker. Immunohistochemistry for microtubule-associated protein 2 and albumin was employed to monitor neuronal degeneration and the breakdown of the blood brain barrier, respectively. In the ischemic core of the caudoputamen, the immunoreactivity for microtubule-associated protein 2 disappeared and massive albumin extravasation occurred several hours after vessel occlusion, while proteolipid protein messenger RNA signals remained relatively strong at this time. The messenger RNA signals began to attenuate 12 h after ischemia and were hardly detectable 24 h after ischemia in the whole ischemic lesion. In situ end-labeling of fragmented DNA showed some cells with proteolipid protein messenger RNAs to have DNA fragmentation at this period. In contrast to proteolipid protein messenger RNA signals, the immunoreactivity for myelin basic protein was detected as long as five days after ischemia. An apparent increase in the cells possessing strong proteolipid protein messenger RNA signals was found five days after ischemia, mainly in the corpus callosum and the cortex bordering the infarcted areas. A double simultaneous procedure with in situ hybridization for proteolipid protein messenger RNA and immunohistochemistry for glial fibrillary acid protein or lectin histochemistry for macrophages/microglia showed proliferating oligodendrocytes to be co-localized with reactive astrocytes and macrophages/microglia. These findings show that oligodendrocytic damage occurred following ischemic neuronal damage and the breakdown of the blood brain barrier, but preceded the breakdown of myelin proteins in the ischemic lesion, that an apoptosis-like process was involved in ischemic oligodendrocytic death, and that surviving oligodendrocytes responded and proliferated in the outer border of the infarcted area.

Acute increases in plasma oxidized low-density lipoprotein immediately after percutaneous transluminal coronary angioplasty.

To investigate the acute changes in plasma oxidized low-density lipoprotein before and immediately after coronary angioplasty, we studied 132 consecutive patients who successfully underwent this procedure. Plasma oxidized low-density lipoprotein levels were significantly increased immediately after coronary angioplasty in patients with stable angina pectoris as well as in those with acute coronary syndromes.

Nucleotide sequence involved in the replication of cloned yeast mitochondrial DNA.

A fragment of yeast mitochondrial DNA, Alu B, has two subfragments, Alu B1 and Alu B2. They were each cloned and sequenced. The autonomously replicating function of the curtailed Alu B1 (342 bp) was defined within 186 bp. A GC-rich sequence identical to the oris sequence in the curtailed Alu B1 was unnecessary for its autonomously replicating function. The 186 bp sequence had an ATATAAAT sequence and the stem and loop structures. The base sequence of Alu B2 also contained the same octanucleotides, the stem and loop structures, one oris sequence and one unique GC cluster. Yeast transformants with cloned Alu B2 grew slowly. The cloned Alu B2 was enlarged in the yeast host concomitantly with compensation of the slow growth of the transformants.

Nucleotide sequence of an essential region for autonomous replication of cloned yeast mitochondrial DNA.

A 341 bp sequence from yeast mtDNA was cloned, which consisted of an upstream 98 bp AT stretch and a downstream 206 bp AT stretch separated by a single 37 bp GC cluster. Cleavage of this GC cluster did not cause loss of the autonomously replicating function of this sequence. The recloned first 98 bp AT stretch was incapable of replication, while the recloned 206 bp AT stretch could replicate. We were able to confine an essential sequence for autonomous replication within a 186 bp AT stretch. Sequencing data revealed a sequence of ATATAAAT and stem and loop structures within the AT stretch.

Two autonomously replicating sequences near oli-1 gene of yeast mitochondrial DNA.

We cloned two autonomously replicating sequences from a short segment of mtDNA of an oligomycin-resistant petite yeast, O-111, into a vector pYleu 12 constructed from yeast LEU 2 gene and pBR 322. These plasmids, pYmit 4 and pYmit 1, had frequencies of transformation of yeast as high as that of YEp 13, having a replicator of 2 mu DNA. They were maintained as plasmids in yeast under selective conditions and shuttled from yeast to E. coli. No evidence was obtained that these plasmids were incompatible with the wild-type mitochondrial genome. These sequences were located in intergenic regions.

C57BL/6 strain is most susceptible to cerebral ischemia following bilateral common carotid occlusion among seven mouse strains: selective neuronal death in the murine transient forebrain ischemia.

Rats and gerbils have been used widely to investigate the molecular mechanism of selective neuronal death following transient global ischemia. Recently, the availability of transgenic mice has enabled us to examine the involvement of specific gene products in various pathophysiological conditions. However, there has been only limited information about the experimental model of cerebral ischemia in mice, particularly in regard to selective neuronal death. We examined whether bilateral carotid occlusion produced global forebrain ischemia in seven common mouse strains including C57BL/6, ICR, BALB/c, C3H, CBA, ddY and DBA/2, based on neurological signs, histological findings and cortical microcirculatory as well as India ink perfusion patterns. The C57BL/6 strain was found to be the most susceptible among seven strains. All C57BL/6 mice died within 6 h after permanent bilateral carotid occlusion. After transient bilateral carotid occlusion for 20 min, more than 90% of C57BL/6 mice showed typical neurological signs such as torsion of the neck and rolling fits, and developed selective neuronal death in the hippocampus and caudoputamen. Hypothermia prevented the neuronal death. Visualization of brain vasculature by India ink perfusion indicated that the susceptibility of the mice after bilateral carotid occlusion depended mainly on the degree of anastomosis between carotid and basilar arteries. Our results showed the feasibility of investigating selective neuronal death in transgenic mice with simple temporary occlusion of both common carotid arteries, when those from the C57BL/6 strain or inbred transgenic mice from other strains with the C57BL/6 strain in a back-cross manner are used.

Ischemic tolerance in moderately symptomatic gerbils after unilateral carotid occlusion.

Ischemic tolerance following transient global cerebral ischemia has drawn considerable attention because of the putative cell defense mechanism which may be inducible by ischemic stress. The purpose of this study is to investigate the inducibility of ischemic tolerance in moderately symptomatic gerbils after unilateral carotid occlusion. Adult Mongolian gerbils were used. Under ether inhalation, the right common carotid artery was occluded for up to 30 min with an aneurysmal clip. Immediately after occlusion, neurological signs and motor function were evaluated and gerbils with moderate signs were selected for investigation of ischemic tolerance. Ischemia for 30 min to gerbils with moderate signs constantly caused neuronal death in the caudoputamen, but it was prevented by pretreatment with 10 min ischemia which was reversible but strong enough to produce heat shock protein 70. The results show that ischemic tolerance can be induced after hemispheric cerebral ischemia as in the case of global cerebral ischemia and suggest that ischemic tolerance may be relevant in human stroke.

Induction of cyclooxygenase-2 mRNA in gerbil hippocampal neurons after transient forebrain ischemia.

We examined the effect of brain ischemia on neuronal expression of cyclooxygenase-2 gene in the hippocampus. Transient forebrain ischemia was produced by occluding bilateral carotid arteries for 5 min in Mongolian gerbil. Northern blotting and in situ hybridization demonstrated that expression of cyclooxygenase-2 mRNA was transiently induced in the hippocampal neurons. Although future studies will be needed to clarify if induced cyclooxygenase-2 following ischemia is involved in neuronal damage or neuronal protection, selective cyclooxygenase-2 inhibitors may be a new therapeutical approach for the treatment of stroke.