Conceivable difference in the anti-inflammatory mechanisms of lipocortins 1 and 5

Human recombinant lipocortins (LCT) 1 and 5 have been expressed in a yeast secretion vector and purified by ion exchange chromatography. The action of the proteins has been investigated in two models of experimental acute inflammation in the rat: carrageenin induced paw oedema and zymosan induced pleurisy. The effects of the proteins on PGE2 release in vitro by rat macrophages stimulated with zymosan and on rat neutrophil chemotaxis induced by FMLP have also been assessed. LCT-1 significantly inhibited both paw swelling in carrageenin oedema and leukocyte migration in zymosan pleurisy. Moreover it showed a dose dependent, inhibitory effect on PGE2 release. Neutrophil chemotaxis was only weakly affected by LCT-1. Conversely LCT-5 did not reduce carrageenin oedema and slightly inhibited PGE2 release, but showed profound, dose dependent inhibitory activity on leukocyte migration in zymosan pleurisy and on neutrophil chemotaxis. These data suggest that LCT-1 acts mainly by interfering with arachidonic acid metabolism via the inhibition of phospholipase A2. The anti-inflammatory activity of LCT-5, at variance with LCT-1, may be due to a direct effect on cell motility in addition to the interference with arachidonic acid metabolism.


Introduction
Glucocorticoids are drugs endowed with a potent anti-inflammatory effect, although the mechanism of action is not completely understood. It has been proposed that they inhibit phospholipase a 2 (PLA2) activity and thus prevent eicosanoid formation. The inhibitory effect of glucocorticoids on PLA2 appears to be mediated by inducible inhibitory proteins called lipocortins (tCTs). After the cloning and identification of the gene of lipocortin 1 (LCT-1) it has been shown that they belong to a larger family of proteins (annexins) with calcium and phospholipid binding properties, whose precise biological role is not yet clear. 2 In vivo anti-inflammatory activity has been reported in experimental inflammation for recombinant LCT-13 and for a purified 36 kDa protein immunologically related to lipocortin 5 (LCT-5). 4 Recently, by aligning the sequences of LCT-5 and uteroglobin (another anti-PLA2 protein) a nine aminoacid conservative region has been identified. The peptide deriving from this region of LCT-5 (amino acids 204-212) has shown anti-inflammatory activity both in vitro and in vivo suggesting that it could represent the active site of the anti-inflammatory effect of  In this paper the anti-inflammatory mechanism of LCT-1 and LCT-5 has been investigated both in vitro and in vivo. The results confirm that the two (C) 1993 Rapid Communications of Oxford Ltd proteins possess anti-inflammatory activity and suggest that they act differently in the control of experimental inflammation.

Materials and Methods
Expression and purification of the proteins: Human LCT-1 and LCT-5 have been obtained as recombinant proteins by using a yeast expression/ secretion.system. The published cDNAs 6'7 for both proteins have been subcloned into the Saccharomyces cerevisiae secretion vector YEpsec 1 previously described. Yeast cells transformed with the recombinant plasmids secrete the heterologous proteins into the culture medium containing galactose as inducer of expression. The proteins secreted in the yeast culture medium were purified by ion exchange chromatography. The medium containing LCT-1 was first applied to a preequilibrated 20 2.5 cm column of DEAE cellulose (DE-52 Whatman, Maidstone, UK) in 25 mM Tris buffer pH7.7 and thoroughly washed at 1 ml/min with the same buffer. In these conditions the LCT-1 was recovered in the flow-through fractions that were pooled, dialysed and lyophilized. These samples were resuspended in 25 mM diethanolamine buffer pH 8.8 and applied to a 8 x 1 cm column of Q-Sepharose (Pharmacia, Uppsala, Sweden) and run at 1.1 ml/min in the same buffer. The fractions containing the protein (flow-through) were collected, dialysed against 1 x 100 vol of 20 mM ammonium carbonate buffer pH 8.0 and lyophilized. LCT-5 was purified by applying the yeast medium to a QAE-Zprep 60 disk (Cuno, Meriden, CT, USA) previously equilibrated in 250 mM Tris buffer pH 7.0. After washing with 25 mM Tris pH 7.0, the elutio.n of the protein was performed by applying a gradient of NaC1 (0.1-0.5 M) in the same buffer (100 ml). LCT-5 eluted in 0.2M NaC1 and was dialysed and lyophilized as above. To use as negative control in inflammation experiments (see below) a sham protein (SHAM) was prepared using a yeast strain transformed with a plasmid lacking the LCT cDNA and purified as described for LCT-1. Western blotting analysis: Western blots were performed on the proteins separated in 10% polyacrylamide gels and electroblotted to nitrocellulose membranes as described previously. Immunodetection was carried out using specific polyclonal antibodies raised against the amino terminus of the proteins. The amino termini (amino acids 15-31, ENEEQEYVQTVKSSKGG, for LCT-1 and amino acids 1-11, MAQVLRGTVTD, for  were synthesized with an Applied Biosystems 430 A peptide synthesizer. The peptides (1 mg/ml in PBS) were cross-linked to 400#g keyhole limpet haemocyanin (Calbiochem. San Diego, CA, USA) by incubation in 0.5 ml of 20 mM glutaraldehyde at room temperature for 30 min. Antibodies were raised in New Zealand male rabbits by subcutaneous injection of the crosslinked peptides (1 mg in complete Freund's adjuvant). Booster injections of the cross-linked peptides in incomplete Freund's adjuvant were given 7, 14, 24 and 60 days later and rabbits bled 2 weeks after the last injection. Additional Western blots were also performed with polyclonal antibodies raised against the entire proteins (kindly supplied by Dr J. L. Browning, Biogen, Cambridge, MA, USA). Animals: Male Wistar rats (200-250 g body weight) were purchased from Charles River, Calco, Italy and housed for a week before experiments.
PGEe release by rat peritoneal macrophages: Macrophages were collected as described previously, Briefly, rat peritoneal cavities were washed with 20 ml PBS + 5 U/ml heparin (Roche, Basel, Switzerland) and cells plated at I x 106 per well in cluster 24-well plates (Costar, Cambridge, MA, USA), in RPMI 1640 medium (Gibco, Paisley, UK) with 20% foetal calf serum. After 3 h incubation at 37C in 5% COg, non-adherent cells were removed by washing with serum-free medium and the adherent macrophages incubated with the proteins for 30min. After stimulation with 200 g/ml opsonized zymosan (Sigma, St Louis, MO, USA) for 60 min, in a final volume of 0.5 ml of serum free medium, supernatants were collected and PGE2 concentration assessed by a specific radioimmunoassay (NEN-DuPont, Dreieich, Germany). Cell viability was always greater than 98% as measured by the trypan blue exclusion test.
Polymorphonuclear leukocyte chemotaxis: The chemotaxis assay was performed according to Harvarth et  for 45 min at 37C in 5% CO2. The migrating cells were stained with Diff-Quik reagent (Merz-Dade AG, Dudingen, Switzerland) and counted with a light microscope. Ten separated fields of each well were counted at 100 .
Carrageenin oedema: The oedema was induced by the subplantar injection of 0.1 ml of 1% carrageenin (Sigma, St Louis, MO, USA) in 0.9% NaC1 into the right hindpaw of rats, as described previously, The proteins were administered i.v. (1 ml/kg) immediately before carrageenin injection (time 0). Control animals received the vehicle (saline). The paw volume was measured in a double blind manner with a water plethysmometer (Basile, Comerio, Italy) at time 0 and every 60 min up to 5 h. Zymosan pleurisy: The pleurisy was induced as described by Perretti et al. 11 Briefly, rats were injected intrapleurally with 0.2 ml of 2 mg/ml zymosan in 0.9% NaC1 and killed by CO2 inhalation 4 h later. The pleural cavities were washed with 1 ml of PBS + 5 U/ml heparin, the fluids collected and total and differential count of cells performed by using Turk's staining and a Neubauer haematocytometer. The proteins were administered i.v. (1 ml/kg) 30 min before zymosan injection. The doses of the proteins used in the oedema and pleurisy experiments were established on the basis of preliminary experiments. Control animals received the vehicle (saline).

Results
Purity and identification of the proteins: Figure 1 shows representative SDS-PAGE gels of LCT-5 (lane 1) and LCT-1 (lane 2) after purification. Purity was > 90% for LCT-1 and > 80% for LCT-5. Figure 1 also shows that the proteins were recognized in Western blots by polyclonal antibodies raised against the entire molecules (lane 3 for LCT-5 and lane 4 for LCT-1) as well as by antibodies raised against the amino termini (lane 5 for LCT-5 and lane 6 for LCT-1). Efect of lipocortins on PCEe release" LCT-1 reduced in a dose dependent manner the release of PGE2 from activated macrophages with a maximum inhibition of about 70% at 1 ng/ml. LCT-5 showed a weaker effect with a peak inhibition of about 40% at 1 ng/ml. The sham protein preparation had no significant inhibitory effect on PGE2 release ( Table   1).
Effect of lipocortins on PMN chemotaxis: The chemotaxis of rat PMN was dose dependently reduced by LCT-5 with significant inhibition at 10 and 100 ng/ml (21.8% and 41.7% respectively). LCT-1 caused a 30% inhibition at 1 ng/ml which did not increase at higher concentrations. PMN chemotaxis was not inhibited by the sham protein ( Table 2).
Eect of lipocortins on carrageenin oedema: The i.v. administration of LCT-1 (1 mg/kg) resulted in a profound and long-lasting reduction of the paw swelling induced by the phlogogen compound. The inhibition was significant at 2, 3 and 4 h after carrageenin injection. In contrast, the same i.v. dose of LCT-5 was completely ineffective in reducing carrageenin oedema. As a rhatter of fact, the first hour oedema was substantially increased by LCT-5 (Fig. 2). Leukocyte migration was not significantly modified by i.v. injection of the sham protein (Fig. 3).

Discussion
The present results indicate that LCT-1 and LCT-5 possess anti-inflammatory activity with some differences. Both proteins inhibited leukocyte migration in vivo during zymosan induced pleurisy whereas only LCT-5 caused a dose dependent inhibition of PMN chemotaxis in vitro. On the other hand, LCT-5 was less potent than LCT-1 in reducing PGE2 release from activated macrophages and was ineffective in reducing carrageenin induced paw oedema where LCT-1 gave a significant and long-lasting inhibition.
The anti-inflammatory profile of LCT-1 is consistent with inhibition of PLA2 activity. In fact the preferential effect on leukocyte migration in vivo is likely due to the down-regulation of the synthesis of chemotactic mediators like LTB 4 and PAFacether. LCT-1 is indeed capable, like glucocorticoids, of impairing the formation of all lipid metabolites originating from PLA 2 activation. Moreover, LCT-1 and dexamethasone have been shown to reduce PMN migration induced in vivo by interleukin-1 (IL-1), 12 which is known to activate PLA2 .13 These data support a regulatory feedback mechanism by LCT-1 and IL-1 on PLA2 activity, as suggested previously. 14  purified from human placenta. v This discrepancy is likely due to the different biological activity of the proteins, given the extreme sensitivity of LCT-1 to denaturation and the structural heterogeneity of the various preparations. 18 The anti-inflammatory profile of LCT-5 suggests that this protein directly affects cell motility in addition to the interference with arachidonic acid metabolism. The carrageenin induced oedema was not inhibited by LCT-5, rather the paw swelling at the first hour was increased by the protein, an effect possibly due to contaminants in the preparation. The lack of inhibition of carrageenin oedema by LCT-5 can be ascribed to the relatively weak effect of the protein on PGE2 release, since prostaglandins are important mediators in this model of inflammation. 9 Since LCT-5 was administered intravenously, it is possible that, in the paw, it did not reach the concentration necessary to inhibit eicosanoid release. This view is supported by the fact that a peptide from LCT-5, injected directly in the paw, was able to inhibit the carrageenin oedema. It is of interest that a 36 kDa protein, related to LCT-5, was able to inhibit PMN migration and eicosanoid formation in a model of murine inflammation. This protein also inhibited PMN chemotaxis in vitro. 4 The decreased eicosanoid formation in vivo could have been a consequence of the reduced cell migration. The anti-inflammatory 11;2 Mediators of Inflammation. Vol 2.1993 effect of LCT-5 could be related to the inhibition of protein kinase C (PKC) activity. 2 It has been suggested that PKC activation in leukocytes leads to cellular responses like lysosomal enzyme release, arachidonate release and superoxide formation. 21 In agreement with this hypothesis it has been shown recently that PKC inhibition results in antiinflammatory activity. 22 Alternatively or additionally, the inhibitory effect of LCT-5 on cell motility could be due to the property of the protein to tightly bind to negatively charged phospholipids present on activated cell surfaces. It has been shown that LCT-5 binds strongly to membranes of activated macrophages, but not of resting cells. 23 This is supported by the recent observation that peripheral blood leukocytes have very few binding sites for LCT-5 (N. Goulding, personal communication).
Whatever the mechanism of action, which deserves further study, LCT-1 and LCT-5 possess anti-inflammatory activity with some conceivable differences. This observation is very interesting in the light of recent data from the authors' laboratory on the expression of lipocortins brought about by glucocorticoids. The authors have shown that dexamethasone induces the release of LCT-1 and LCT-5, but not of LCT-2, from differentiated U-937 cells. Therefore it can be suggested that glucocorticolds could control the different facets of the inflammatory process through the release of lipocortins endowed with different actions.