Induction of annexin-1 at transcriptional and post-transcriptional level in rat brain by methylprednisolone and the 21-aminosteroid U74389F

Brain tissue of rats pretreated with methylprednisolone or with the 21-aminosteroid U74389F, and that of untreated control rats, was assessed for the expression of annexin-1 (Anx-1) and the transcription of its mRNA. For this purpose Anx-1 cDNA was amplified and simultaneously a T7-RNA-polymerase promoter was incorporated into the cDNA using a polymerase chain reaction (PCR). Then digoxigenin-11-UTP was incorporated into the transcribed cRNA with T7-RNA-polymerase. With this probe in situ hybridization was carried out on sections of the brain. The probe was visualized by an immunoassay using an antidigoxigenin antibody conjugate. Anx-1 protein was assessed by means of immunohistochemistry using a polyclonal antibody. The various brain areas of the control animals showed an appreciable amount of Anx-1 at mRNA or protein level; on the other hand, the animals which had been pretreated with either steroid, showed a more intense Anx-1 mRNA signal than the controls in many areas. In the pretreated animals Anx-1 immunostaining was unchanged in cortex, basal ganglia, amygdala and septum, but more intense in hippocampus, hypothalamus and thalamus. In ependyma, choroid plexus, meninges, and vascular walls there was no Anx-1 mRNA transcription detectable. An opposite profile was shown by the Anx-1 immunoreactivity, the protein was present in control animals as well as the steroid-pretreated animals, suggesting that here the protein was either from systemic origin, or has diffused from adjacent structures. The results indicated that Anx-1 mRNA transcription is upregulated by either steroid, and that in the untreated animals there is a resting level of Anx-1 mRNA transcription, presumably reflecting physiological influences on Anx-1 expression.


Introduction
Since the 1960s glucocorticoidsteroids have been successfully employed in the clinical treatment of vasogenic brain oedema. Oedema of this type originates from disruption of the blood-brain barrier, resulting in exudation of blood plasma into the brain parenchyma. The vasogenic oedema and the inherent disruption of the blood-brain barrier tends to accompany most focal lesions of the brain, such as tumours, inflammatory lesions, infarctions, trauma and haemorrhage. The disruption of the barrier which is to be localized in the endothelium of the cerebral capillaries, has been ascribed to changes of endothelial permeability, allegedly caused by oedema mediators, although the 370 Mediators of Inflammation Vol 5 1996 barrier defect in brain tumours is usually due to the occurrence of fenestrations in the endothelial cytoplasm.10edema mediators include free unsaturated fatty acids, especially arachidonic acid, and oxygen derived free radicals, which have been shown to be produced in brain tissue during injury. The unsaturated fatty acids are released from membrane phospholipids by phospholipase A2 attacking the glycerophospholipid molecule at the sn-2 position. Moreover, arachidonic acid, and its derivatives the eicosanoids, are allegedly involved in inflammatory processes in other tissues of the body.
Annexins, or lipocortins, have been detected as mediators of the anti-inflammatory action of glucocorticosteroids by inhibiting phospholipase A2 .2'3 This inhibition is both calcium-and (C) 1996 Rapid Science Publishers lipid-dependent. [4][5][6] Upon intravenous injection, Anx-1, 2 and 5 proved to exert a potent antiinflammatory effect in rat and mouse. 7 '8 The annexins are a family of proteins which all possess a conserved 70 amino acid repeating unit. In most annexins, there is a four-fold repeat, except for Anx-6 which has eight such repeats. Moreover, all annexins possess Ca 2/ and phospholipid binding sites. The differences between the annexin proteins are mainly located in the amino-terminus. 9 Anx-1 has been shown to be abundantly present in extracts of human placenta, and in the spleen, and kidneys of the rat. Readily detectable tissue levels in the rat are found in the lung, the thymus, the liver, the heart, the brain, the bone marrow and the ileum. In a previous immunocytochemical study on rat brain, annexin-immunoreactivity only appeared in sporadic microglial cells and in the choroid plexus, being absent in the greater part of the brain in animals not pretreated with steroids.
However, after administration of methylprednisolone or the 21-aminosteroid U74389F the neurones, ependyma, oligodendroglia and capillary endothelium of the brain showed an induction of annexin expression. 1 It has been the aim of the present investigation to assess in the central nervous system the expression of Anx-1 by methylprednisolone and the 21-aminosteroid U74389E mediated by the transcription of Anx-1 mRNA, thereby excluding the presence of Anx-1 which has originated from peripheral sources.

Oligonucleotide primers and promoters
The sequences of the oligonucleotides used in this study and the predicted lengths of the PCR amplification products are listed in Table 1 Polymerase chain reaction (PCR) Circa 10ng annexin-1 (Anx-1) human cDNA was used as PCR template. The use of human cDNA is justified by the 87% homology with the murine cDNA. 11 The PCR mixture consisted of 10 1 PCR buffer (500 mM KC1, 15 mM MgC12 and 100 mM Tris-HC1 pH 9.0), 1 nmol start and stop primers, 10  Transcription of digoxigeninconjugated cRNA All reactions involving the use of RNA were performed with solutions treated with 0.1% diethylpyrocarbonate (DEPC). Labelled cRNA was produced with a T7-RNA-polymerase (Pharmacia Biotech, Uppsala, Sweden) reaction to which digoxigenin-ll-UTP (Boehringer Mannheim, Mannheim, Germany) was added. The following reaction mixture was incubated at 37C for 2 h: 1/g PCR-product, 2/1 transcription buffer (400 mM Tris-HC1 pH 8.0, 60 mM MgC12, 20 mM spermidine and 100 mM NaCD, 2/.tl NTP-mixture (10 mM ATE CTP, GTP, 6.5 mM UTP and 3.5 mM digoxigenin-11-UTP), 63 U T7-RNA-polymerase and DEPC treated H20 to a final volume of 20 bl. The reaction was arrested by the addition of 2 1 of 0.2 M EDTA pH 8.0, followed by RNA extraction and pre-  cipitation. The transcripts were subjected to agarose gel electrophoresis and Northern blotting and identified by immunoassay using an anti-digoxigenin antibody conjugate (Boehringer Mannheim, Mannheim, Germany) following the manufacturer's instructions.

Steroids and dosage
Adult Wistar rats with a mean body weight of 350-400 g, which had free access to food and water were used in the studies. They were divided into three experimental groups: untreated control animals (n 7), a group treated with 2 mg/kg methylprednisolone (Solumedrol, UpJohn Co., Kalamazoo, USA) dissolved in benzyl alcohol (n 8), and a group treated with the 21-aminosteroid U74389F (UpJohn Co., Kalamazoo, USA) dissolved in saline solution (n 4). All drugs were administered 24 and 2 h before the animals were killed. The animals were anaesthetized with ether and perfused with 0.9% NaC1 and 4% paraformaldehyde in 0.1 M phosphate buffered saline (PBS) (pH 7.4). The brains were removed and cryoprotected overnight in a 10% sucrose solution.
The brains were quickly frozen in CO2 and sections of 40 m thickness were cut on a cryostat microtome. The sections were stored in antifreeze (50% PBS, 30% ethylene glycol, 20% glycerol) at -20C. The sections were mounted on poly-L-lysine coated slides, air dried overnight and then used for in situ hybridization (ISH).

In situ hybridization
The slides were incubated in a Proteinase K solution (10 btg/ml in 10 mM Tris, 1 mM EDTA pH 7.5) for 10 min at 37C, rinsed twice in PBS and dehydrated in ethanol. On the slides 100 !,1 of hybridization mixture was applied. The hybridization mixture consisted of 4*SSC (I*SSC is 0.15M NaC1 and 0.015M sodium citrate), l*Denhardt's (0.02% Ficoll, 0.02% polyvinylpyrodilone, 10mg/ml BSA (Section V)), 50% deionized formamide, 5% dextran sulphate, 0.5 mg/ml tRNA and 1 bg/ml digoxigenin-labelled cRNA probe. The slides were covered with coverslips and then sealed in a humid plastic bag. Hybridization was conducted over- The sections used in the ISH and immunocytochemistry studies were analysed by two observers, who were unaware of the treatment of the rats. The relative intensities of the hybridization signal in the different brain regions were judged as weak (/), strong (++), or very strong (+++) ( Tables 2 and 3). No attempt was made to quantify the differences between the treated animals and controls.

Results
Production of labelled cRNAs By the inclusion of the T7-RNA-polymerase promoter sequence in the 5'-termini of the PCR primers ( Table 1) the promoter was incorporated into the terminus of the Anx-1 cDNA amplification products. Separate cDNA templates from which antisense and sense cDNAs could be transcribed, were generated for the Annexin-1 gene; and DNA fragments of the predicted size could be demonstrated by restriction and subsequent electrophoretic analysis of the amplification products. Moreover, Northern blot analysis showed discrete digoxigenin-conjugated sense and antisense cRNAs, which were transcribed from the amplification products by the T7-RNA-polymerase. The identities of the cRNAs were confirmed using the template cDNA and unlabelled cRNA as positive controls to which the labelled Anx-1 cRNA specifically hybridized.
In situ hybridization (ISH) Using the digoxigenin-labelled antisense cRNA probe for Anx-1 mRNA, in situ hybridization (ISH) on paraformaldehyde-fixed frozen sections of rat brains resulted in an adequate colour reaction with minimal background staining.
Both the corresponding sense probe hybridization experiment and the ISH, which had been performed with the digoxigenin-labelled antisense neo-RNA were negative, confirming the specificity of the Anx-1 cRNA hybridization reaction.  The areas of the foreand midbrain which showed a hybridization signal of Anx-1 mRNA are summarized in Table 2. They include the following areas of the telencephalon: layers 2-6 of the cortex, the olfactory bulb, the nuclei of the amygdala, in the hippocampus the pyramidal and molecular layers of the CA1, CA2 and CA3 area and the granular layer of the dentate gyrus, furthermore in the septum and the striatum. Scattered cells were labelled in the hypothalamus, the thalamus and the epithalamus. In the cortex, especially layer 2 of the pyriform and cingulate area showed a high signal intensity (Fig. 2B). Along the fibre bundles of the corpus callosum the oligodendoglia were labelled (Fig. 2A).
No labelling was observed in layer 1 of the cortex and the fimbria of the hippocampus, and notably in the cells constituting the vascular wall, the meninges, the ventricular ependyma, and the choroid plexus ( Fig. 2A). 374 Mediators of Inflammation Vol 5 1996 The Anx-1 mRNA in situ hybridization signal in the various groups The control animals showed labelling of structures shown in Table 2. The methylprednisolone-treated animals showed Anx-1 mRNA hybridization signals in the same areas as the controls, but far more intense, although the sections had been treated in the same session. Especially the various structures of the hippocampus and the habenular nucleus exhibited a strong increase of signal intenstiy when compared with the controls (Fig. 1). The areas in which Anx-1 mRNA labelling was absent in the controls, such as the cells of the meninges, the ependyma, the vascular wall, and the choroid plexus also lacked Anx-1 hybridization signals in the steroid-treated animals.
Pretreatment with the 21-aminosteroid U74389F gave similar results as pretreatment with methylprednisolone. The only difference was found in the CA3 area of the hippocampus in which Anx-1 mRNA expression was less than in the methylprednisolone-treated rats ( Fig. 1C  and 1E). Again no signal was observed in the cells of the meninges, the ependyma, the vascular wall, and the choroid plexus.
Immunocytochemistry of Anx-1 protein In the brains of the control rats, Anx-1 immunoreactivity was observed in layers 2 and 5 of the cerebral cortex, and to a lesser extent in the other cortical layers; furthermore in the basal ganglia, the amygdala, hippocampus, hypothalamus and thalamus ( Table 3). The distribution of Anx-1 immunoreactivity was only slightly different from that of Anx-1 mRNA on the ISH: Anx-1 immunostaining of the neuronal cells was most pronounced both in layers 2 and 5, whereas only layer 2 showed the most abundant Anx-1 mRNA expression. In addition, immunoreactive Anx-1 was found in small varicose nerve fibres of neuronal cells, the ventricular ependyma, the choroid plexus, the meningeal cells and sporadically in cells of the vascular wall (Figs 1B, 2C, 2D and 2E). Contrary to the ISH, immunostaining was found in the fibres but not in the oligodendroglia of the corpus callosum.
Methylprednisolone or the 21-aminosteroid pretreatment increased Anx-1 immunoreactivity in other structures, notably the hippocampus, the hypothalamus and thalamus (Fig. 1), but did not change the Anx-1 immunostaining of cerebral cortex, basal ganglia, amydala, septum, meninges, vascular wall, choroid plexus, and ependyma. was carried out for a shorter period on tissue Discussion sections which had been attached to glass Non-radioactive ISH with digoxigenin-labelled slides. Moreover, the presence of Anx-1 protein probes as applied in the present study, is a rapid in the group of control animals is consistent technique with reasonably high spatial resoluwith the finding of Anx-1 mRNA in the controls, tion. The use of a PCR to incorporate a RNAalbeit in low, presumably resting levels of polymerase promoter-region in the eDNA contranscription. This resting state of Anx-1 mRNA stitutes a simple way to produce a template for transcription may reflect hormonal influences the RNA-transcription reaction and requires no under physiological conditions, such as the cloning. 2 Labelling of the probe with dig-normal glucocorticoid secretion by the adrenals oxigenin-ll-UTP also proved a simple and as governedbythe hypothalamo-pituitaryaxis. effective procedure. The ISH protocol used in The only type of glial cells which could be this study has evolved from a number of differrecognized with reasonable certainty in the ent protocols. [13][14][15][16] present study, without requiring cell-type speci-In the present study Anx-1 mRNA has pretic immunocytochemical markers were the olidominantly been shown in neuronal cells withgodendroglia in the white matter. Astrocytes in various structures of the brain. Notably, no have been shown to express Anx-1 mRNA 21 and Anx-1 mRNA has been found in the capillary to possess glucocorticoid receptors, and may endothelium and the choroid plexus. Our pretherefore be assumed to be among the labelled vious immunocytochemical studies have demoncells in the cortex. In the hippocampus, on the strated annexins in the choroid plexus of other hand, no cells with the morphological control animals. 1 Also, Anx-1 protein has been features of astrocytes have been found to be demonstrated in choroid plexus and in cellular labelled.
elements of some other circumventricular or-In the steroid-pretreated animals, there was gans, such as the median eminence. 17a8 The increased, transcription of Anx-1 mRNA in all choroid plexus as well as the other circumven-structures of the brain, as in other tissue. 22'23 tricular organs are notably devoid of a blood-Evidence of an increase of Anx-1 immunosta'inbrain barrier. Therefore, the Anx-1 protein ohing, however, was only found in some strucserved in the choroid plexus may well be of tures, notably the hippocampus, the systemic origin, having entered because of the hypothalamus and thalamus. Another inconsislack of a blood-brain barrier. This contention tency was the absence of Anx-1 in the oligodenseems now to be confirmed by the absence of droglia in spite of ISH evidence of Anx-1 mRNA. Anx-1 mRNA in these structures as demon-It may be speculated that in these cases transstrated by ISH. lation of the Anx-1 protein has not followed the In addition, Anx-1 protein could be demon-mRNA transcription, presumably impeded by strated in the meninges, the ventricular ependyinfluences yet to be clarified. ma, and the vascular wall, whereas ISH showed Glucocorticoids can cross the blood-brain no Anx-1 mRNA transcription. Since these barrier, and after entering the parenchymal cells structures are located behind the blood-brain bind to the intracellular glucocorticosteroid barrier, the protein may have originated by receptor. The receptor and its mRNA have been diffusion possibly followed by preferential accu-demonstrated by immunohistochemistry to ocmulation from adjacent areas; diffusion of procur in nuclei of neuronal and glial cells of the teins is a well-known and aspecific process in central nervous system with a widespread, but brain tissue. 19 An alternative explanation would heterogeneous distribution. 2425 On comparing be the disparate turn-over rates of mRNA, 2 the distribution of Anx-1 mRNA with that of which may explain subsequent rapid disappear-glucocorticoid receptors (Table 2), a correlation ance of the mRNA. seems to be apparent. The recently introduced Although in our previous study evidence of 21-aminosteroids have been claimed to exert Anx-1 immunocytostaining has not been found their action by a direct membrane-protective in the larger part of the brain of control animals, effect a,ainst oxidation by oxygen-derived free in the present study Anx-1 immunostaining radicals, 6 lacking the glucocorticosteroid effect occurred in quite a number of cerebral struc-on glucose metabolism; however, our results tures. The differences are most probably attribu-indicate that they induce expression of annextable to the far greater sensitivity of the freeins as well.
floating technique, by which in the present Although in the early 1980s the annexins study the immunohistochemistry was carried have been discovered mainly in their role of out overnight on 40 tzm sections, whereas in mediators of glucocorticosteroid anti-inflammathe previous study the immunohistochemistry tory action by their alleged inhibition of phos-pholipase A2mwith a similar effect on brain oedema being the object of the present studym other properties of the annexins have been revealed in the meantime, including anticoagulant activity (placental Anx-3 and Anx-4), channel activity (Anx-7, Anx-5 and Anx-1), trafficking of vesicles like endocytosis and exocytosis and aggregation of synaptic vesicles (Anx-7, Anx-2, Anx-3 and Anx-4), and involvement in proliferation and differentiation processes (Anx-1 and Anx-2).27 The recent elucidation of the crystal structures of Anx-5 and Anx-12 has revived the discussion on their role of channels, particularly in view of the hexameric structure of Anx-12. 28,29 In conclusion the present study has demonstrated the presence of annexin-1 in the brain as a result of local transcription of the Anx-1 mRNA, exhibiting an upregulation following administration of a glucocorticosteroid as well as a 21-aminosteroid. In view of the previously demonstrated anti-inflammatory and anti-oedematous effects of these steroids, this upregulation of annexin expression is presumably involved in mediating these steroid actions on phospholipase A2 activity in the arachidonic acid cascade and eicosanoid formation. Besides, a resting level of annexin expression has been observed which presumably reflects physiological conditions, in which their alleged other functions such as those of channels or in vesicle formation are relevant. It has also been demonstrated by the absence of mRNA transcription, that their occurrence in the choroid plexus can be attributed to a lack of the blood-brain barrier, and that in this respect annexins do not differ from other proteins.