Molecular and Cellular Endocrinology, 0 1991 Elsevier Scientific Publishers
MOLCEL
41
82 (1991) 41-50 Ireland, Ltd. 0303-7207/91/$03.50
02628
Crotoxin, a phospholipase A, neurotoxin from snake venom, interacts with epithelial mammary cells, is internalized and induces secretion Michkle
Ollivier-Bousquet
‘, FranGois Radvanyi
* and Cassian Bon *
’ Laboratoire de Biologie Cellulaire et Molhdaire,
INRA, 78350 Jouy-en-Josas, France, and ’ Laboratoire des Venins, Unite Associee Pasteur/ INSERh4 285, Institut Pasteur, 75724 Paris, France (Received
Key words: Phospholipase
A,;
Neurotoxin;
27 March
Mammary
1991; accepted
cell; Casein
secretion;
12 July 1991)
Arachidonic
acid; Prolactin
Summary Prolactin (PRL) induces liberation of arachidonic acid (AA) from phospholipids of lactating mammary epithelial cells and stimulates casein secretion. In order to investigate the possible involvement of phospholipase A, (PLA,) activity in the hormonal control of casein secretion by PRL, we examined the effects of crotoxin, a PLA, neurotoxin from snake venom, on mammary epithelial cells. Crotoxin is made of two subunits: a basic PLA, with low toxicity (component B, CB) and an acidic, non-toxic and enzymatically inactive component A (CA) which enhances the pharmacological action of CB. While CA is inactive, the PLA, subunit (CB) induces an accumulation of secretory products in the lumen of mammary acini, an extensive development of the Golgi apparatus. The secretion of newly synthesized casein is increased in the presence of CB and this effect is inhibited by nordihydroguaiaretic acid (NDGA) and caffeic acid, two inhibitors of the lipoxygenase pathway which also prevent stimulation of secretion by PRL. Further, CB transiently induces the release of radiolabelled AA from mammary tissues previously labelled with [‘4C]AA, the highest release being observed between 15 s and 5 min of contact with CB and CA. Immunofluorescence labelling by anti-CB antibodies of epithelial mammary tissues previously incubated with CA, CB or a combination of CA and CB indicates that CB binds to epithelial cells and is internalized, at least in part, and that CA enhances both CB binding and its internalization. These observations emphasize the involvement of PLA, in the control of casein secretion and suggest that PLA, acts intracellularly.
Introduction Address for correspondence: Dr. Michele OllivierBousquet, Laboratoire de Biologie Cellulaire et Moleculaire, INRA, 78350 Jouy-en-Josas, France. Tel. (33-l) 34.65.25.49. Abbreviations: AA, arachidonic acid; BSA, bovine serum albumin; CA, component A of crotoxin; CB, component B of crotoxin; NDGA, nordihydroguaiaretic acid; PBS, phosphatebuffered saline; PLA,, phospholipase A,; PLP, periodate lysine paraformaldehyde; PRL, prolactin.
Prolactin (PRL) is a hormonal regulator of casein secretion by rabbit mammary epithelial cells (Ollivier-Bousquet, 1978). One of the early steps in the events following the binding of PRL on its receptor may be the activation of a phos-
42
pholipase A, (PLA,) and the liberation of arachidonic acid (AA) from the membrane phospholipids. Products of the oxidative metabolism of AA may be involved in basal and PRL-stimulated casein secretion (Blachier et al., 1988). Early observations, concerning the effects on mammary cells of exogenous AA, PLA,and inhibitors of AA metabolism are indeed in agreement with this hypothesis (Ollivier-Bousquet, 1984). Production of AA and AA metabolites from membrane phospholipids, as a consequence of activation of cell surface receptors, has been implied in the mechanism of action of many secretagogues in other cell types (Naor and Catt, 1981; Metz, 1988; Ross et al., 1988; Wang and Leung, 1988). AA release might occur by the direct action of exogenous PLA z on membrane phospholipids: PLA z of different origins have indeed been shown to stimulate secretion in various systems (Chi et al., 1982; Canonico et al., 1983; Yamamoto et al., 1983; Grandison, 1984; Abou-Samra et al., 1986; Kojima et al., 1986; Hirst et al., 1988). PLA, may, however, exert different effects on distinct physiological responses of mammary cells. It has been shown that, in cultured mammary tissues of midpregnant female rat and rabbit, PLA, is essential for PRL action on ornithine decarboxylase activity (Rillema et al., 1983) and lipid biosynthesis (Rillema et al., 1986) while in cultured pseudopregnant rabbit mammary tissues, PLA, did not affect the stimulation of casein synthesis by PRL (Dusanter-Fourt et al., 1983). In view of the different effects of PLA, on the lactogenic process in female rat and pseudopregnant rabbit, it is of interest to investigate the possible involvement of a PLAz activity on casein secretion by lactating mammary epithelial cells and to compare its action with that of PRL. We used crotoxin, a snake venom PLA, neurotoxin, from the venom of Crotalus durissus terrificus, which exerts its pathophysiological action primarily by altering neurotransmitter release from nerve endings. Crotoxin is made of two components: a basic PLA, with low toxicity (component B; CB) and an acidic subunit (component A; CA), inactive when applied alone, which enhances the lethal potency of CB by preventing its non-specific adsorption (Breithaupt et al., 1971; Hendon and Fraenkel-Conrat, 1971; Bon et al., 1979). We
used CB to investigate the action of a highly purified PLA, on the secretory process. The specificity conferred by CA to CB appeared as an additional useful tool. Materials
and methods
Animals New Zealand female rabbits and female tar rats were used on day 15 of lactation.
Wis-
Preparation of crotoxin and of its isolated subunits CA and CB Crotoxin was purified from Crotalus durissus terrificus venom (obtained from the stock of the Pasteur Institute) by gel filtration on Sephadex G-75 followed by ion exchange chromatography on diethylaminoethyl (DEAE) cellulose. Isolated CA and CB were prepared by cationic and anionic chromatography, both in 6 M urea, as described by Hendon and Fraenkel-Conrat (1971). Preparation of anti-CB antibodies CB was emulsified with 50% Freund’s adjuvant and administered to rabbits S.C. (750 pg per animal) at 3-week intervals. Boosts, which were performed when the serum titers were decreasing, were achieved by injecting i.m. the same dose of crotoxin component B in the presence of incomplete Freund’s adjuvant. Specific anti-CB antibodies were purified by a double ammonium sulfate precipitation at 30% saturation and chromatography on immunosorbent prepared with purified CB. After concentration, the purified antibodies were kept at - 18°C in 50% glycerol. Anti-CB reacted with isolated CB, crotoxin, crotoxin-like toxins and a few other snake venom PLA, neurotoxins, but not with non-toxic enzymes from snake venoms and mammalian pancreas (Choumet et al., 1989). Incubation Each experiment was performed with mammary tissue fragments from one animal (weight of each fragment: 0.1-0.2 mg; total weight per assay: 50-70 mg). Fragments were incubated at 37°C under an atmosphere containing 95% 0,
43
and 5% CO, in Hanks’ medium (pH 7.41 containing 2.2 g/l bicarbonate.
Light and electron microscopy
After 15 min of preincubation in Hanks’ medium, tissue fragments were incubated in the same medium in the presence of l-4 FM CA, 0.25-l PM CB or a combination of both components. After 1 h, fragments were fixed with 4% glutaraldehyde-4% paraformaldehyde (v/v) in 0.1 M cacodylate buffer, followed by 1% osmium tetroxide in the same buffer and finally embedded in Epon. Semi-thin sections (0.5-l pm> were stained with azure II-methylene blue and thin sections with uranyl acetate and lead citrate. In order to compare the volume occupied by subcellular structures in mammary epithelial cells in the presence of the different agents studied, we used the method of ‘spot numbering’, as described by Ollivier-B(~usquet (1978). Twenty micrographs were counted in each experiment.
Immunofluorescence
After 15 min of preincubation in Hanks’ medium, tissue fragments were incubated in the same medium in the presence of 4 FM CA, 1 FM CB or a combination of both components. After 5, 30 or 60 min, fragments were washed, then fixed with periodate lysine-paraformaldehyde (PLP) according to McLean and Nakane (19741, infused into 10% sucrose, frozen in liquid N, and sectioned in 2 E;Lrnthick sections at - 35°C with a Reichert cryocut. The sections were collected on poly-L-lysine-coated glass coverslips and sequentiaIly incubated with 50 mM NH,Cl, 0.2% gelatin and horse preimmune serum, l/3 in 0.01 M phosphate-buffered saline (PBS) containing 1% bovine serum albumin (BSA), then incubated 15 h with anti-CB antibody (7.5 pg/ml), washed with 0.01 M PBS containing 1% BSA, incubated 1 h with a biotinylated F(ab’1, anti-rabbit immunoglobulin (Amersham, U.K.) l/200 in 0.01 M PBS containing 1% BSA, washed in the same buffer and finalIy incubated 1 h with streptavidin-Texas Red (Amersham, UK) l/200 in the same buffer. After washes with 0.01 M PBS pH 8, the sections were mounted on a drop of glycerol
and observed under a Polyvar Reichert microscope equipped with a filter set for rhodamine. Label&g
with /‘4CJAA
and thin-layer chromatog-
ra~hy
After 30 min of preincubation in Hanks’ medium, the tissue fragments (50-70 mg/2 ml) were labelled for 45 min, in the same medium, with 1 pCi/ml of [13-‘4C]AA, specific activity 1.44 GBq/mmol (39 mCi/mmol! (CEA, Saclay, France). After labelling, the tissues were washed extensively and re-incubated in 2 ml of the same medium with or without 1 FM CB or with the combination 4 PM CA-t 1 &M CB. After 15 s, 1 min, 5 min and 15 min, the tissues were immediately homogenized in chloroform/ methanol/ water (8 : 4 : 3, v/v) and the lipids were extracted. Thin-layer chromatography was carried out on silica gel (Merck 601 and developed in hexane/ether/formic acid (85 : 20 : 2, v/v), then in hexane/ether/methanol/formic acid (40 : 60 : 0.8: 0.08, v/v) and revealed with iodide. The different spots corresponding to the different lipid classes (compared to standards) were scraped off and then counted by liquid scintillation in the presence of thixotropic gel powder (CAB-OSTL, Packard; 40 mg/ml) and scintillating toluene. The radioactivity of fatty acid was evaluated as a percentage of total radioactivity. Labelling with
L-I 3,4,.5-“Hlleucine After 30 min of preincubation in Hanks’ medium with or without inhibitor (see figure legends), the tissue fragments were pulse-labelled for 3 min with 40 pCi/ml of L-[3,4,5-~Hlleucine (CEA, Saclay, France, specific activity 2.22 TBq/mmol; 60 Ci/mmoll, then rinsed extensively with the medium and re-incubated in the presence or not of 10 pg/ml of PRL or CA and CB alone or in combination at different concentrations with or without 10 PM nordihydroguaiaretic acid (NDGA) or 100 PM caffeic acid (Sigma, St. Louis, MO, U.S.A.).
Labelled casein assay
The incubations were stopped 60 min after initial labelling and the tissues were weighed and homogenized in 300 ~1 of 10 mM phosphate
44
buffered saline (pH 7.2), 1% Triton X-100 and 0.5% sodium deoxycholate, then tissue proteins were precipitated with 300 ~1 of 20% trichloroacetic acid in the presence of 50 ~1 of 1% BSA and washed with 10% trichloroacetic acid. The labelied caseins secreted in the medium were precipitated at their isoelectric point with 0.5 M sodium acetate buffer (pH 4.6) in the presence of non-labelled caseins (final dilution 0.5 mg/ml). The radioactivity of secreted caseins and of tissue proteins was counted by liquid scintillation in a Packard spectrometer. The results were expressed as the percentage of radioactivity secreted versus the total radioactivity incorporated into the cells: Secreted
caseins
(76)
= 100 x [(radioactivity
of secreted
~(radioactivity
of tissue protein)
+ (radioactivity
of secreted
caseins)
caseins)]
_’
Statistical analysis Each experiment was made with mammary gland fragments from a single animal. Control and treated assays differed only by the addition of hormone, component A and/or component B. Student’s f-test on paired differences was used to statistically evaluate the difference between the treated and the control.
1). Epithelial cells were tall with extensive development of vesicles in the supranuclear region which could be attributed to the development of the Golgi complex. The aspect of the tissue treated with CB alone or in combination with CA was essentially identical to that seen in the case of tissue treated with PRL, indicating that the accumulation of secretory products in the lumen was not caused by a degenerative or a lytic effect of CB but was rather due to stimulation of secretion. The stimulation of casein secretion by CB alone or in the presence of CA was dose dependent and maximal for micromolar concentrations. Further, similar obse~ations were obtained with rabbit mammary tissue treated with similar concentrations of CB alone or in combination with CA. Examination by electron microscopy of the lactating rabbit-doe mammary epithelial cells incubated with CB or a combination of CB and CA reveaied an extensive development of one constituent of the Golgi complex: microvesicles. The modifications induced by crotoxin and its isolated components A and B were further evaluated in a quantitative manner by measuring the relative volumes of Golgi vesicles cisternae, microvesicles and secretory vesicles (Fig. 2). Neither PRL nor CA, CB and their combination modified the relative volumes of Golgi vesicles cisternae. In contrast, CB (1 FM) alone and a combination of CB (1 PM) and CA (1 ,uM) increased the relative volume of microvesicles as efficiently as 10 pg/ml PRL.
Results
Light and electron microscopy The effects of crotoxin and of its isolated components A and B were first investigated by examining, by light microscopy, rat mammary tissue incubated in Hanks’ medium with CA or CB or a combination of CA and CB. Fig. 1 shows the results of a typical experiment. Control and tissue treated with CA alone had similar aspects with well developed acini and cuboida1 epithelial cells containing ergastoplasm and Golgi apparatus. Tissue incubated for 1 h with 0.25 FM CB and with the combination of 0.25 PM CB and 1 PM CA showed an accumulation of secretory products in the lumen, particularly lipid droplets (Fig.
Immunoflrorescence Since polyclonal antibodies directed to CA and CB had been raised in rabbits, immuno~uores” cence experiments were performed with rat mammary tissue. Morphological and biochemical studies showed that CA, CB and their combination behave in the same manner in rat and rabbit. Rat mammary tissues were first incubated with CA, CB or their combination for various lengths of time, then treated for immunoffuorescence with anti-CB antibody, with a biotinyiated antirabbit immunoglobulin and finally with Texas Red-streptavidin. No fluorescence labelling was detectable in mammary tissues incubated with 4 PM CA (Fig. 3e). On the other hand, fluores-
Fig. 1. Micrographs of mammary tissue incubated with PRL, crotoxin and its components. Mammary tissues from female rat were incubated for 1 h in Hanks’ medium: (a) control, (b) 1 PM CA, (c) 10 pg/ml PRL, (d) 0.25 PM CB, (e) 1 PM CA-to.25 PM CB. Large white spheres inside the lumen are lipid droplets (big arrows). Smaller light areas adjacent to intracellular droplets or close to the nucleus are secretory vesicles (small arrows). Bar, 10 pm.
46 Secretory
vesicles ‘1
-r-
C
Micro
PRL
2 -
CB
CA+CB
vesicfes s-l
C Golgi
CA
vesicles 2 1
PRL
CA
CB
CA+CB
cisternae
T
C
PRL
CA
CB
CA+CB
Fig. 2. Effect of component A and component B on the relative volume of constituents of the Golgi complex. The relative volumes of secretory vesicles, microvesicles and Golgi vesicles cisternae were measured after incubation from female rabbit mammal tissues for 60 min in the presence of 10 &g/ml PRL. 1 JLM CA, I FM CB and 1 ,uM CA+ 1 &M CB. Means + SEM performed with three animals. * Significantly superior to control ( p < 0.05).
cence was observed at the edge of the tissue after 5 or 30 min of incubation with 1 PM CB. This fluorescence was associated with the connective tissue surrounding the fragments and with the basal membrane of epithelial cells of the acini located at the periphery of the fragments. Only little labelling was detectable inside the cells (Fig.
3u and cl. Finally, a bright fluorescence Iabelling was detectable on many acini and on the content of the lumen, in tissue fragments incubated during 5 or 30 min with a combination of CA 4 PM and I FM CB (Fig. 36 and cl). The labelling appeared as spots in a supra-nuclear position and close to the apex of the cells (Fig. 3f). The same labelling was observed in the case of tissue fragments incubated for longer periods of time (60 min) with a combination of CA and CB (results not shown). Release offree A.4 from rnarnrn~~ ep~i~e~~fffcells Rabbit mammal tissues were prelabelled with [ “C]AA for 45 min, then incubated without (control) or with 1 PM CB, alone or in combination with 4 PM CA. Tissue samples were removed after 15 s, 1 min, 5 min and 15 min and the radioactivity of free fatty acids, separated by thin layer chromatography, was measured. Fig. 4 shows that CB alone induced a transient increase of the radioactivity in free fatty acids (percent of total lipid radioactivity), which reached its maximum at 5 min of incubation, later decreasing to the control level after 15 min. A similar transient liberation of radiolabelled AA was observed when CB was combined with CA. However, the phenomenon was faster and started after 1.5 s of incubation. Stimulation of casein secretion CA, at concentrations below 1 wM, was unable to stimulate the secretion of newly synthesized caseins in rabbit mammary cells (Fig. 5). CB alone, at 0.1 ,uM or higher concentrations increased the secretion of newly synthesized caseins. This effect reached a plateau at 1 FM CB, a concentration of CB which stimulated the secretion as efficiently as 10 pg/ml PRL. Fig. 5 also shows that CA (at a molar ration of 4 CA/CB) increased the secretory response elicited by CB. Although a maximal effect was obtained with the combination of 8 PM CA and 2 FM CB (not shown), intermediate concentrations of 4 PM CA and 1 PM CB have been used for most of the present experiments. We previously reported that two inhibitors of the lipoxygenase pathway, NDGA and caffeic acid, inhibit PRL stimulated casein secretion
Fig. 3. Immunolocaiization of component B. Mammary tissues from female rat were incubated in Hanks’ medium in the presence of 4 I.IM CA (e), of I NM CB (a, c) or 4 FM CA+ 1 ,uM CB (b, d, fh fixed after 5 min (a, b) or 30 min (c, d, f) of incubation, then immunostained with anti-CB antibody, a second conjugate (biotinylated anti-rabbit immunoglobulin) and Texas Red-streptavidin. In the presence of CB (a, c), fluorescent label is detectable at the periphery of the fragments (big arrows) and slightly inside the cells (small arrows). In the presence of CA+CB (b. d) fluorescent label is detectable on the basal membrane (big arrows), inside the cells (small arrows) and in the content of the lumen (arrowheads). Panel (f) details experiment (d) at higher magnification, the fluorescent label is detectable on the basal membrane (wide arrow) and inside the cells (small arrows). Bar, 10 pm.
15 set
1 mim
5 min
15 min
Time
Fig. 4. Release of free AA by CB alone or in combination with CA. Mammary tissues from female rabbit were incubated for 45 min with 1 kCi/ml of [“‘CIAA. The tissues were washed extensively and incubated again without or with I FM CB, or 4 FM CA+ 1 FM CB. The different lipid constituents were separated by thin layer chromatography and the proportion of radioactivity in fatty acids was expressed as a percentage of control experiments. Means from four animals.
(Blachier et al., 1988). Although NDGA has been shown to be an inhibitor of both cyclooxygenase and lipoxygenase in several preparations (van Wayne and Goosens, 1983) in the lactating rabbit
*O1
f
,“/
l*;r(il)
Fig. 5. Effects of component A and component B alone or in combination on casein secretion. Mammary tissues from female rabbit were labelled for 3 min with L-[sH]leucine, washed, then incubated in the presence or not of 10 pg/ml PRL or of different concentrations of CA and CB. 60 min after the beginning of the pulse, radioactive secreted caseins were evaluated as described in Materials and methods. Mean iSEM from the number of animals indicated in brackets. * Significantly superior to control (p < 0.05). *** Significantly superior to control t p < 0.001).
Fig. 6. Effect of NDGA on casein secretion stimulated by PRL and crotoxin (CA+CB). Mammary tissue fragments from female rabbit were preincubated with or without IO PM NDGA, labelled for 3 min with t.-[‘Hlleucine, washed and finally incubated in the presence or not of 10 pg/ml PRL or 4 PM CA+ I PM CA, with or without 10 FM NDGA or 100 PM caffeic acid. 60 min after the beginning of the pulse. radioactive secreted caseins were evaluated as described in Materials and methods, Means+ SEM from four animals. ** Significantly superior to control (p < 0.01). *** Significantly superior to control (p < 0.001).
mammary tissues, it strongly inhibits the LTC, production (Blachier et al., 1988) but does not significantly decrease the prostaglandin production (Ollivier-Bousquet and Lacroix, 1986). Further, caffeic acid has been described as a relatively specific inhibitor of the transformation of LTA, into LTC, (Leung, 1986). In order to examine if stimulation of casein secretion by CB requires the integrity of the lipoxygenase pathway, tissue fragments were incubated with CB in the presence of these inhibitors. Fig. 6 shows that the stimulation induced by 1 PM CB in combination with 4 mM CA was completely inhibited by 10 PM NDGA or 100 PM caffeic acid, like in the case of PRL, therefore confirming the hypothesis. Also, as previously observed (Blachier et al., 1988>, the two inhibitors increased the basal level of casein secretion. Discussion This investigation shows that the PLA, subunit of crotoxin, component B (CB), used alone or in combination with the non-catalytic component A (CA), is able to bind to epithelial mammary cells and to efficiently stimulate casein se-
49
cretion, whereas non-enzymatic component A alone is totally inactive. Since it is the first time that such effects are described, it was of importance to examine if this increase in cell secretion does not result from a lytic action of the phospholipase on the membranes. In fact, cytomorphological observation of the mammary tissue shows that acini exhibit an active secretory aspect after treatment with crotoxin. Moreover, in vitro incorporation of [3H]leucine and neosynthesis of caseins confirmed that the metabolic activity of the tissue was unaffected. Immunofluorescence investigations performed with anti-CB antibodies indicated that CB alone strongly adsorbs on membranes, as shown by the non-selective labelling of the peripheral structures of mammary tissue fragments, while in the presence of CA it preferentially interacts with epithelial cell membranes. The binding of CB, on mammary epithelial cells, in the presence of the non-catalytic subunit CA, is in agreement with the properties of these components on neuromuscular transmission (Bon et al., 1979; Hawgood and Santana de Sa, 1979). It has been shown: (i) that the two subunits of crotoxin complex (CACB) separate upon interaction with membranes: the phospholipase subunit CB binds, whereas the non-catalytic subunit is released in solution; and (ii> that CA prevents non-specific absorption of CB on membrane phospholipids, this favoring its binding to high affinity sites (Bon et al., 1979). Immunofluorescence investigations carried out with crotoxin components A and B indicated that a similar mechanism of binding occurs in the case of mammary epithelial cells. Once it is bound to the plasma membrane of mammary epithelial cells, CB is rapidly internalized since after several minutes it appears as immunofluorescent spots localized in a supranuclear position near the apex of the cells. This localization was confirmed with immuno-electron microscopy on tissues embedded in Lowicryl. In the presence of CA, CB was mainly located on the Golgi apparatus and in secretory vesicles (not shown). The internalization of CB in mammary epithelial cells is an interesting new observation. Indeed, an exogenous PLA, which, in a cellular system, is expected to interact primarily with the external side of the plasma membrane, will not
necessarily act in the same manner as an intracellular enzyme. In fact, the importance of an intracellular localization of PLA, in the regulatory process has been recently emphasized. For example, fibroblasts transformed by a rus oncogene in which the PLA, activity is increased, accumulate the intracellular enzyme on membrane ruffles and on vesicles (Bar-Sagi et al., 1988). Neutrophil activation correlates with an increase of PLA, activity in combination with the translocation of the enzyme from plasma membrane to intracellular structures (Balsinde et al., 1988). Similarly, a translocation of PLA, from cytosol to membranes has been reported in the case of macrophage activation (Schontardt and Ferber, 1987). As in the case of PRL (Blachier et al., 19881, the action of crotoxin on mammary epithelial cells led to a rapid and transient release of AA and required the integrity of the lipoxygenase metabolic pathway. This confirms previous observations according to which the secretagogue action of PRL was thought to be mediated by the activation of an endogenous PLA, and observations showing that exogenous enzymes can mimic the hormone action on mammary epithelial cells (Ollivier-Bousquet et al., 1984). Indeed, exogenous PLA, have been shown to stimulate numerous secretory responses (Burgoyne et al., 1987). The fact that the extent of CB internalization in the presence of CA was correlated with an increased stimulation of casein secretion suggests a specific action of crotoxin on the secretory system of mammary epithelial cells. It emphasizes the importance of the subcellular localization of PLA, for the production of AA. All eukaryotic tissues possess intracellular PLA, located on plasma membranes, lysosomes, mitochondria, etc. (Van den Bosch, 1980) and it is likely that the activity of one or several of these enzymes is regulated through hormonal receptors. Since PRL is very rapidly internalized in the lactating mammary epithelial cell and entrapped by the Golgi apparatus, secretory vesicles and lysosomes (Kane et al., 1983), it would be interesting to determine whether the same intracellular compartments are involved in the action of hormone and of PLA,. Comparison with the action of the toxin in the blockade of synaptic transmission at the neuromuscular junction (Bon et al., 1979) showed that
so the step of binding of CB in the presence of CA seemed similar in the two systems. However, differences appear in the mode of action of crotoxin since blockade of neuromuscular transmission results from its effects on voltage-sensitive ions channels (Hawgood and Santana de Sa, 1979) whereas stimulation of casein probably involves the eicosanoids pathway. Furthermore, internalization of crotoxin does not seem to be implied at the neuromuscular junction (Trivedi et al., 1989) while it is quite evident in lactating mammary epithelial cells. In conclusion, crotoxin is able to mimic the secretagogue effect of PRL on mammary epithelial cells and appears to be a useful tool to study the regulation of secretory processes in these cells. Acknowledgments The authors thank Dr. A. Tixer-Vidal and Dr. J. Massoulie for their helpful comments upon the manuscript and M. Ahmed Ali and S. Delpal for excellent technical assistance. References Abou-Samra, A.B., Catt, J. and Aguilera, G. (1986) Endocrinology 119, 1427-1431. Balsinde, J., Die,, E., Schuller, A. and Mollinedo, F. (1988) J. Biol. Chem. 263. 1929-1936. Bar-Sagi, D., Suhan, J.P., McCormick, F. and Feramisco, J.R. (1988) J. Cell Biol. 106. 1649-1658. Blachier, F., Lacroix, M.C., Ahmed-Ali, M., Leger, C. and Ollivier-Bousquet, M (1988) Prostaglandins 35, 259-276. Bon, C., Changeux, J.P.. Jeng, T.W. and Fraenkel-Conrat, H. (1979) Eur. J. Biochem. YY, 471-481. Breithaupt, H.. Riibsamen, K., Walsch, P. and Haberman, E. (1971) Naunyn-Schmied. Arch. Pharmacol. 269, 403-404. Burgoyne. R.D., Cheek, T.R. and O’Sullivan, A.J. (1987) Trends Biochem. Sci. 12, 3322333. Canonico, L., Schettini, G., Valdenegro, CA. and McLeod, R.M. (19X3) Neuroendocrinology 37, 212-217.
Chi, E.Y., Henderson, W.R. and Lebanoff. S.J. (1982) Lab. Invest. 47, 579-585. Choumet, V., Jiang, M.S., Radvanyi, F., Ownby, C. and Bon, C. (1989) FEBS Lett. 244. 367-173. Dusanter-Fourt, I., Djiane, J. and Houdebine, L.M. (19X3) Mol. Cell. Endocrinol. 31, 2X7-299. Grandison. L. (1984) Endocrinology 114, 1-7. Hawgood, B.J. and Santana de Sa. S. (1979) Neurosciences 4, 293-303. Hendon. R.A. and Fraenkel-Conrat, H. (1971) Proc. Nat]. Acad. Sci. U.S.A. 6X. 1560-1563. Hirst. J.J., Rice, G.E., Jenkins, G. and Thornburn, G.D. (1988) Endocrinology 122, 774-781. Kane, S., Raymond, M.N., Dusanter-Fourt, I., Houdebine. L.M., Djiane, J. and Ollivier-Bousquet, M. (1983) Eur. J. Cell. Biol. 30, 24442.53. Kojima. I., Kojima, K. and Rasmussen. H. (1986) Endocrinology 117, 1057-1066. Leung, K.H. (1986) Biochem. Biophys. Res. Commun. 137, 195-200. McLean, I.W. and Nakane. P.K. (1974) J. Histochem. Cytochem. 22. 1077-1083. Metz, S.A. (1988) Prostaglandins Leukotrienes Essential Fatty Acids 32, 187-202. Naor, Z. and Catt. K.J. (1981) J. Biol. Chem. 256, 2226-222’). Ollivier-Bousquet, M. (1978) Cell. Tissue Res. 187, 25-43. Ollivier-Bousquet, M. (1984) Biol. Cell 51, 327-334. Ollivier-Bousquet, M. and Lacroix, M.C. (1986) Reprod. Nutr. Dev. 26. 575-582. Rillema, J.A.. Wing, L.Y.C. and Foley, K.A. (1983) Endocrinology 113, 202442028. Rillema, J.A., Etindi, R.N. and Cameron, M. (1986) Horm. Metab. Res. 18, 672-674. Ross, P.C., Judd. A.M. and McLeod, R.M. (1088) Endocrinology 123, 2445-2453. Schontardt, T. and Ferber. E. (1987) B&hem. Biophys. Res. Commun. 149. 7699775. Trivedi. S.. Kaiser, I.]., Tanaka, M. and Simpson, L. (1989) J. Pharmacol. Exp. Ther. 25 1, 490-496. Van den Bosch, H. (1980) Biochim. Biophys. Acta 604. 19l246. Van Wayne, J. and Goosens, J. (1983) Prostaglandins 26, 725-730. Wang, J. and Leung. P.C.K. (1988) Endocrinology 122, 9066 911. Yamamoto, S., Nakai, T., Nakade. T. and Kato, R. (1983) Eur. J. Pharmacol. 86, 121-124.
MOLCEL
41
82 (1991) 41-50 Ireland, Ltd. 0303-7207/91/$03.50
02628
Crotoxin, a phospholipase A, neurotoxin from snake venom, interacts with epithelial mammary cells, is internalized and induces secretion Michkle
Ollivier-Bousquet
‘, FranGois Radvanyi
* and Cassian Bon *
’ Laboratoire de Biologie Cellulaire et Molhdaire,
INRA, 78350 Jouy-en-Josas, France, and ’ Laboratoire des Venins, Unite Associee Pasteur/ INSERh4 285, Institut Pasteur, 75724 Paris, France (Received
Key words: Phospholipase
A,;
Neurotoxin;
27 March
Mammary
1991; accepted
cell; Casein
secretion;
12 July 1991)
Arachidonic
acid; Prolactin
Summary Prolactin (PRL) induces liberation of arachidonic acid (AA) from phospholipids of lactating mammary epithelial cells and stimulates casein secretion. In order to investigate the possible involvement of phospholipase A, (PLA,) activity in the hormonal control of casein secretion by PRL, we examined the effects of crotoxin, a PLA, neurotoxin from snake venom, on mammary epithelial cells. Crotoxin is made of two subunits: a basic PLA, with low toxicity (component B, CB) and an acidic, non-toxic and enzymatically inactive component A (CA) which enhances the pharmacological action of CB. While CA is inactive, the PLA, subunit (CB) induces an accumulation of secretory products in the lumen of mammary acini, an extensive development of the Golgi apparatus. The secretion of newly synthesized casein is increased in the presence of CB and this effect is inhibited by nordihydroguaiaretic acid (NDGA) and caffeic acid, two inhibitors of the lipoxygenase pathway which also prevent stimulation of secretion by PRL. Further, CB transiently induces the release of radiolabelled AA from mammary tissues previously labelled with [‘4C]AA, the highest release being observed between 15 s and 5 min of contact with CB and CA. Immunofluorescence labelling by anti-CB antibodies of epithelial mammary tissues previously incubated with CA, CB or a combination of CA and CB indicates that CB binds to epithelial cells and is internalized, at least in part, and that CA enhances both CB binding and its internalization. These observations emphasize the involvement of PLA, in the control of casein secretion and suggest that PLA, acts intracellularly.
Introduction Address for correspondence: Dr. Michele OllivierBousquet, Laboratoire de Biologie Cellulaire et Moleculaire, INRA, 78350 Jouy-en-Josas, France. Tel. (33-l) 34.65.25.49. Abbreviations: AA, arachidonic acid; BSA, bovine serum albumin; CA, component A of crotoxin; CB, component B of crotoxin; NDGA, nordihydroguaiaretic acid; PBS, phosphatebuffered saline; PLA,, phospholipase A,; PLP, periodate lysine paraformaldehyde; PRL, prolactin.
Prolactin (PRL) is a hormonal regulator of casein secretion by rabbit mammary epithelial cells (Ollivier-Bousquet, 1978). One of the early steps in the events following the binding of PRL on its receptor may be the activation of a phos-
42
pholipase A, (PLA,) and the liberation of arachidonic acid (AA) from the membrane phospholipids. Products of the oxidative metabolism of AA may be involved in basal and PRL-stimulated casein secretion (Blachier et al., 1988). Early observations, concerning the effects on mammary cells of exogenous AA, PLA,and inhibitors of AA metabolism are indeed in agreement with this hypothesis (Ollivier-Bousquet, 1984). Production of AA and AA metabolites from membrane phospholipids, as a consequence of activation of cell surface receptors, has been implied in the mechanism of action of many secretagogues in other cell types (Naor and Catt, 1981; Metz, 1988; Ross et al., 1988; Wang and Leung, 1988). AA release might occur by the direct action of exogenous PLA z on membrane phospholipids: PLA z of different origins have indeed been shown to stimulate secretion in various systems (Chi et al., 1982; Canonico et al., 1983; Yamamoto et al., 1983; Grandison, 1984; Abou-Samra et al., 1986; Kojima et al., 1986; Hirst et al., 1988). PLA, may, however, exert different effects on distinct physiological responses of mammary cells. It has been shown that, in cultured mammary tissues of midpregnant female rat and rabbit, PLA, is essential for PRL action on ornithine decarboxylase activity (Rillema et al., 1983) and lipid biosynthesis (Rillema et al., 1986) while in cultured pseudopregnant rabbit mammary tissues, PLA, did not affect the stimulation of casein synthesis by PRL (Dusanter-Fourt et al., 1983). In view of the different effects of PLA, on the lactogenic process in female rat and pseudopregnant rabbit, it is of interest to investigate the possible involvement of a PLAz activity on casein secretion by lactating mammary epithelial cells and to compare its action with that of PRL. We used crotoxin, a snake venom PLA, neurotoxin, from the venom of Crotalus durissus terrificus, which exerts its pathophysiological action primarily by altering neurotransmitter release from nerve endings. Crotoxin is made of two components: a basic PLA, with low toxicity (component B; CB) and an acidic subunit (component A; CA), inactive when applied alone, which enhances the lethal potency of CB by preventing its non-specific adsorption (Breithaupt et al., 1971; Hendon and Fraenkel-Conrat, 1971; Bon et al., 1979). We
used CB to investigate the action of a highly purified PLA, on the secretory process. The specificity conferred by CA to CB appeared as an additional useful tool. Materials
and methods
Animals New Zealand female rabbits and female tar rats were used on day 15 of lactation.
Wis-
Preparation of crotoxin and of its isolated subunits CA and CB Crotoxin was purified from Crotalus durissus terrificus venom (obtained from the stock of the Pasteur Institute) by gel filtration on Sephadex G-75 followed by ion exchange chromatography on diethylaminoethyl (DEAE) cellulose. Isolated CA and CB were prepared by cationic and anionic chromatography, both in 6 M urea, as described by Hendon and Fraenkel-Conrat (1971). Preparation of anti-CB antibodies CB was emulsified with 50% Freund’s adjuvant and administered to rabbits S.C. (750 pg per animal) at 3-week intervals. Boosts, which were performed when the serum titers were decreasing, were achieved by injecting i.m. the same dose of crotoxin component B in the presence of incomplete Freund’s adjuvant. Specific anti-CB antibodies were purified by a double ammonium sulfate precipitation at 30% saturation and chromatography on immunosorbent prepared with purified CB. After concentration, the purified antibodies were kept at - 18°C in 50% glycerol. Anti-CB reacted with isolated CB, crotoxin, crotoxin-like toxins and a few other snake venom PLA, neurotoxins, but not with non-toxic enzymes from snake venoms and mammalian pancreas (Choumet et al., 1989). Incubation Each experiment was performed with mammary tissue fragments from one animal (weight of each fragment: 0.1-0.2 mg; total weight per assay: 50-70 mg). Fragments were incubated at 37°C under an atmosphere containing 95% 0,
43
and 5% CO, in Hanks’ medium (pH 7.41 containing 2.2 g/l bicarbonate.
Light and electron microscopy
After 15 min of preincubation in Hanks’ medium, tissue fragments were incubated in the same medium in the presence of l-4 FM CA, 0.25-l PM CB or a combination of both components. After 1 h, fragments were fixed with 4% glutaraldehyde-4% paraformaldehyde (v/v) in 0.1 M cacodylate buffer, followed by 1% osmium tetroxide in the same buffer and finally embedded in Epon. Semi-thin sections (0.5-l pm> were stained with azure II-methylene blue and thin sections with uranyl acetate and lead citrate. In order to compare the volume occupied by subcellular structures in mammary epithelial cells in the presence of the different agents studied, we used the method of ‘spot numbering’, as described by Ollivier-B(~usquet (1978). Twenty micrographs were counted in each experiment.
Immunofluorescence
After 15 min of preincubation in Hanks’ medium, tissue fragments were incubated in the same medium in the presence of 4 FM CA, 1 FM CB or a combination of both components. After 5, 30 or 60 min, fragments were washed, then fixed with periodate lysine-paraformaldehyde (PLP) according to McLean and Nakane (19741, infused into 10% sucrose, frozen in liquid N, and sectioned in 2 E;Lrnthick sections at - 35°C with a Reichert cryocut. The sections were collected on poly-L-lysine-coated glass coverslips and sequentiaIly incubated with 50 mM NH,Cl, 0.2% gelatin and horse preimmune serum, l/3 in 0.01 M phosphate-buffered saline (PBS) containing 1% bovine serum albumin (BSA), then incubated 15 h with anti-CB antibody (7.5 pg/ml), washed with 0.01 M PBS containing 1% BSA, incubated 1 h with a biotinylated F(ab’1, anti-rabbit immunoglobulin (Amersham, U.K.) l/200 in 0.01 M PBS containing 1% BSA, washed in the same buffer and finalIy incubated 1 h with streptavidin-Texas Red (Amersham, UK) l/200 in the same buffer. After washes with 0.01 M PBS pH 8, the sections were mounted on a drop of glycerol
and observed under a Polyvar Reichert microscope equipped with a filter set for rhodamine. Label&g
with /‘4CJAA
and thin-layer chromatog-
ra~hy
After 30 min of preincubation in Hanks’ medium, the tissue fragments (50-70 mg/2 ml) were labelled for 45 min, in the same medium, with 1 pCi/ml of [13-‘4C]AA, specific activity 1.44 GBq/mmol (39 mCi/mmol! (CEA, Saclay, France). After labelling, the tissues were washed extensively and re-incubated in 2 ml of the same medium with or without 1 FM CB or with the combination 4 PM CA-t 1 &M CB. After 15 s, 1 min, 5 min and 15 min, the tissues were immediately homogenized in chloroform/ methanol/ water (8 : 4 : 3, v/v) and the lipids were extracted. Thin-layer chromatography was carried out on silica gel (Merck 601 and developed in hexane/ether/formic acid (85 : 20 : 2, v/v), then in hexane/ether/methanol/formic acid (40 : 60 : 0.8: 0.08, v/v) and revealed with iodide. The different spots corresponding to the different lipid classes (compared to standards) were scraped off and then counted by liquid scintillation in the presence of thixotropic gel powder (CAB-OSTL, Packard; 40 mg/ml) and scintillating toluene. The radioactivity of fatty acid was evaluated as a percentage of total radioactivity. Labelling with
L-I 3,4,.5-“Hlleucine After 30 min of preincubation in Hanks’ medium with or without inhibitor (see figure legends), the tissue fragments were pulse-labelled for 3 min with 40 pCi/ml of L-[3,4,5-~Hlleucine (CEA, Saclay, France, specific activity 2.22 TBq/mmol; 60 Ci/mmoll, then rinsed extensively with the medium and re-incubated in the presence or not of 10 pg/ml of PRL or CA and CB alone or in combination at different concentrations with or without 10 PM nordihydroguaiaretic acid (NDGA) or 100 PM caffeic acid (Sigma, St. Louis, MO, U.S.A.).
Labelled casein assay
The incubations were stopped 60 min after initial labelling and the tissues were weighed and homogenized in 300 ~1 of 10 mM phosphate
44
buffered saline (pH 7.2), 1% Triton X-100 and 0.5% sodium deoxycholate, then tissue proteins were precipitated with 300 ~1 of 20% trichloroacetic acid in the presence of 50 ~1 of 1% BSA and washed with 10% trichloroacetic acid. The labelied caseins secreted in the medium were precipitated at their isoelectric point with 0.5 M sodium acetate buffer (pH 4.6) in the presence of non-labelled caseins (final dilution 0.5 mg/ml). The radioactivity of secreted caseins and of tissue proteins was counted by liquid scintillation in a Packard spectrometer. The results were expressed as the percentage of radioactivity secreted versus the total radioactivity incorporated into the cells: Secreted
caseins
(76)
= 100 x [(radioactivity
of secreted
~(radioactivity
of tissue protein)
+ (radioactivity
of secreted
caseins)
caseins)]
_’
Statistical analysis Each experiment was made with mammary gland fragments from a single animal. Control and treated assays differed only by the addition of hormone, component A and/or component B. Student’s f-test on paired differences was used to statistically evaluate the difference between the treated and the control.
1). Epithelial cells were tall with extensive development of vesicles in the supranuclear region which could be attributed to the development of the Golgi complex. The aspect of the tissue treated with CB alone or in combination with CA was essentially identical to that seen in the case of tissue treated with PRL, indicating that the accumulation of secretory products in the lumen was not caused by a degenerative or a lytic effect of CB but was rather due to stimulation of secretion. The stimulation of casein secretion by CB alone or in the presence of CA was dose dependent and maximal for micromolar concentrations. Further, similar obse~ations were obtained with rabbit mammary tissue treated with similar concentrations of CB alone or in combination with CA. Examination by electron microscopy of the lactating rabbit-doe mammary epithelial cells incubated with CB or a combination of CB and CA reveaied an extensive development of one constituent of the Golgi complex: microvesicles. The modifications induced by crotoxin and its isolated components A and B were further evaluated in a quantitative manner by measuring the relative volumes of Golgi vesicles cisternae, microvesicles and secretory vesicles (Fig. 2). Neither PRL nor CA, CB and their combination modified the relative volumes of Golgi vesicles cisternae. In contrast, CB (1 FM) alone and a combination of CB (1 PM) and CA (1 ,uM) increased the relative volume of microvesicles as efficiently as 10 pg/ml PRL.
Results
Light and electron microscopy The effects of crotoxin and of its isolated components A and B were first investigated by examining, by light microscopy, rat mammary tissue incubated in Hanks’ medium with CA or CB or a combination of CA and CB. Fig. 1 shows the results of a typical experiment. Control and tissue treated with CA alone had similar aspects with well developed acini and cuboida1 epithelial cells containing ergastoplasm and Golgi apparatus. Tissue incubated for 1 h with 0.25 FM CB and with the combination of 0.25 PM CB and 1 PM CA showed an accumulation of secretory products in the lumen, particularly lipid droplets (Fig.
Immunoflrorescence Since polyclonal antibodies directed to CA and CB had been raised in rabbits, immuno~uores” cence experiments were performed with rat mammary tissue. Morphological and biochemical studies showed that CA, CB and their combination behave in the same manner in rat and rabbit. Rat mammary tissues were first incubated with CA, CB or their combination for various lengths of time, then treated for immunoffuorescence with anti-CB antibody, with a biotinyiated antirabbit immunoglobulin and finally with Texas Red-streptavidin. No fluorescence labelling was detectable in mammary tissues incubated with 4 PM CA (Fig. 3e). On the other hand, fluores-
Fig. 1. Micrographs of mammary tissue incubated with PRL, crotoxin and its components. Mammary tissues from female rat were incubated for 1 h in Hanks’ medium: (a) control, (b) 1 PM CA, (c) 10 pg/ml PRL, (d) 0.25 PM CB, (e) 1 PM CA-to.25 PM CB. Large white spheres inside the lumen are lipid droplets (big arrows). Smaller light areas adjacent to intracellular droplets or close to the nucleus are secretory vesicles (small arrows). Bar, 10 pm.
46 Secretory
vesicles ‘1
-r-
C
Micro
PRL
2 -
CB
CA+CB
vesicfes s-l
C Golgi
CA
vesicles 2 1
PRL
CA
CB
CA+CB
cisternae
T
C
PRL
CA
CB
CA+CB
Fig. 2. Effect of component A and component B on the relative volume of constituents of the Golgi complex. The relative volumes of secretory vesicles, microvesicles and Golgi vesicles cisternae were measured after incubation from female rabbit mammal tissues for 60 min in the presence of 10 &g/ml PRL. 1 JLM CA, I FM CB and 1 ,uM CA+ 1 &M CB. Means + SEM performed with three animals. * Significantly superior to control ( p < 0.05).
cence was observed at the edge of the tissue after 5 or 30 min of incubation with 1 PM CB. This fluorescence was associated with the connective tissue surrounding the fragments and with the basal membrane of epithelial cells of the acini located at the periphery of the fragments. Only little labelling was detectable inside the cells (Fig.
3u and cl. Finally, a bright fluorescence Iabelling was detectable on many acini and on the content of the lumen, in tissue fragments incubated during 5 or 30 min with a combination of CA 4 PM and I FM CB (Fig. 36 and cl). The labelling appeared as spots in a supra-nuclear position and close to the apex of the cells (Fig. 3f). The same labelling was observed in the case of tissue fragments incubated for longer periods of time (60 min) with a combination of CA and CB (results not shown). Release offree A.4 from rnarnrn~~ ep~i~e~~fffcells Rabbit mammal tissues were prelabelled with [ “C]AA for 45 min, then incubated without (control) or with 1 PM CB, alone or in combination with 4 PM CA. Tissue samples were removed after 15 s, 1 min, 5 min and 15 min and the radioactivity of free fatty acids, separated by thin layer chromatography, was measured. Fig. 4 shows that CB alone induced a transient increase of the radioactivity in free fatty acids (percent of total lipid radioactivity), which reached its maximum at 5 min of incubation, later decreasing to the control level after 15 min. A similar transient liberation of radiolabelled AA was observed when CB was combined with CA. However, the phenomenon was faster and started after 1.5 s of incubation. Stimulation of casein secretion CA, at concentrations below 1 wM, was unable to stimulate the secretion of newly synthesized caseins in rabbit mammary cells (Fig. 5). CB alone, at 0.1 ,uM or higher concentrations increased the secretion of newly synthesized caseins. This effect reached a plateau at 1 FM CB, a concentration of CB which stimulated the secretion as efficiently as 10 pg/ml PRL. Fig. 5 also shows that CA (at a molar ration of 4 CA/CB) increased the secretory response elicited by CB. Although a maximal effect was obtained with the combination of 8 PM CA and 2 FM CB (not shown), intermediate concentrations of 4 PM CA and 1 PM CB have been used for most of the present experiments. We previously reported that two inhibitors of the lipoxygenase pathway, NDGA and caffeic acid, inhibit PRL stimulated casein secretion
Fig. 3. Immunolocaiization of component B. Mammary tissues from female rat were incubated in Hanks’ medium in the presence of 4 I.IM CA (e), of I NM CB (a, c) or 4 FM CA+ 1 ,uM CB (b, d, fh fixed after 5 min (a, b) or 30 min (c, d, f) of incubation, then immunostained with anti-CB antibody, a second conjugate (biotinylated anti-rabbit immunoglobulin) and Texas Red-streptavidin. In the presence of CB (a, c), fluorescent label is detectable at the periphery of the fragments (big arrows) and slightly inside the cells (small arrows). In the presence of CA+CB (b. d) fluorescent label is detectable on the basal membrane (big arrows), inside the cells (small arrows) and in the content of the lumen (arrowheads). Panel (f) details experiment (d) at higher magnification, the fluorescent label is detectable on the basal membrane (wide arrow) and inside the cells (small arrows). Bar, 10 pm.
15 set
1 mim
5 min
15 min
Time
Fig. 4. Release of free AA by CB alone or in combination with CA. Mammary tissues from female rabbit were incubated for 45 min with 1 kCi/ml of [“‘CIAA. The tissues were washed extensively and incubated again without or with I FM CB, or 4 FM CA+ 1 FM CB. The different lipid constituents were separated by thin layer chromatography and the proportion of radioactivity in fatty acids was expressed as a percentage of control experiments. Means from four animals.
(Blachier et al., 1988). Although NDGA has been shown to be an inhibitor of both cyclooxygenase and lipoxygenase in several preparations (van Wayne and Goosens, 1983) in the lactating rabbit
*O1
f
,“/
l*;r(il)
Fig. 5. Effects of component A and component B alone or in combination on casein secretion. Mammary tissues from female rabbit were labelled for 3 min with L-[sH]leucine, washed, then incubated in the presence or not of 10 pg/ml PRL or of different concentrations of CA and CB. 60 min after the beginning of the pulse, radioactive secreted caseins were evaluated as described in Materials and methods. Mean iSEM from the number of animals indicated in brackets. * Significantly superior to control (p < 0.05). *** Significantly superior to control t p < 0.001).
Fig. 6. Effect of NDGA on casein secretion stimulated by PRL and crotoxin (CA+CB). Mammary tissue fragments from female rabbit were preincubated with or without IO PM NDGA, labelled for 3 min with t.-[‘Hlleucine, washed and finally incubated in the presence or not of 10 pg/ml PRL or 4 PM CA+ I PM CA, with or without 10 FM NDGA or 100 PM caffeic acid. 60 min after the beginning of the pulse. radioactive secreted caseins were evaluated as described in Materials and methods, Means+ SEM from four animals. ** Significantly superior to control (p < 0.01). *** Significantly superior to control (p < 0.001).
mammary tissues, it strongly inhibits the LTC, production (Blachier et al., 1988) but does not significantly decrease the prostaglandin production (Ollivier-Bousquet and Lacroix, 1986). Further, caffeic acid has been described as a relatively specific inhibitor of the transformation of LTA, into LTC, (Leung, 1986). In order to examine if stimulation of casein secretion by CB requires the integrity of the lipoxygenase pathway, tissue fragments were incubated with CB in the presence of these inhibitors. Fig. 6 shows that the stimulation induced by 1 PM CB in combination with 4 mM CA was completely inhibited by 10 PM NDGA or 100 PM caffeic acid, like in the case of PRL, therefore confirming the hypothesis. Also, as previously observed (Blachier et al., 1988>, the two inhibitors increased the basal level of casein secretion. Discussion This investigation shows that the PLA, subunit of crotoxin, component B (CB), used alone or in combination with the non-catalytic component A (CA), is able to bind to epithelial mammary cells and to efficiently stimulate casein se-
49
cretion, whereas non-enzymatic component A alone is totally inactive. Since it is the first time that such effects are described, it was of importance to examine if this increase in cell secretion does not result from a lytic action of the phospholipase on the membranes. In fact, cytomorphological observation of the mammary tissue shows that acini exhibit an active secretory aspect after treatment with crotoxin. Moreover, in vitro incorporation of [3H]leucine and neosynthesis of caseins confirmed that the metabolic activity of the tissue was unaffected. Immunofluorescence investigations performed with anti-CB antibodies indicated that CB alone strongly adsorbs on membranes, as shown by the non-selective labelling of the peripheral structures of mammary tissue fragments, while in the presence of CA it preferentially interacts with epithelial cell membranes. The binding of CB, on mammary epithelial cells, in the presence of the non-catalytic subunit CA, is in agreement with the properties of these components on neuromuscular transmission (Bon et al., 1979; Hawgood and Santana de Sa, 1979). It has been shown: (i) that the two subunits of crotoxin complex (CACB) separate upon interaction with membranes: the phospholipase subunit CB binds, whereas the non-catalytic subunit is released in solution; and (ii> that CA prevents non-specific absorption of CB on membrane phospholipids, this favoring its binding to high affinity sites (Bon et al., 1979). Immunofluorescence investigations carried out with crotoxin components A and B indicated that a similar mechanism of binding occurs in the case of mammary epithelial cells. Once it is bound to the plasma membrane of mammary epithelial cells, CB is rapidly internalized since after several minutes it appears as immunofluorescent spots localized in a supranuclear position near the apex of the cells. This localization was confirmed with immuno-electron microscopy on tissues embedded in Lowicryl. In the presence of CA, CB was mainly located on the Golgi apparatus and in secretory vesicles (not shown). The internalization of CB in mammary epithelial cells is an interesting new observation. Indeed, an exogenous PLA, which, in a cellular system, is expected to interact primarily with the external side of the plasma membrane, will not
necessarily act in the same manner as an intracellular enzyme. In fact, the importance of an intracellular localization of PLA, in the regulatory process has been recently emphasized. For example, fibroblasts transformed by a rus oncogene in which the PLA, activity is increased, accumulate the intracellular enzyme on membrane ruffles and on vesicles (Bar-Sagi et al., 1988). Neutrophil activation correlates with an increase of PLA, activity in combination with the translocation of the enzyme from plasma membrane to intracellular structures (Balsinde et al., 1988). Similarly, a translocation of PLA, from cytosol to membranes has been reported in the case of macrophage activation (Schontardt and Ferber, 1987). As in the case of PRL (Blachier et al., 19881, the action of crotoxin on mammary epithelial cells led to a rapid and transient release of AA and required the integrity of the lipoxygenase metabolic pathway. This confirms previous observations according to which the secretagogue action of PRL was thought to be mediated by the activation of an endogenous PLA, and observations showing that exogenous enzymes can mimic the hormone action on mammary epithelial cells (Ollivier-Bousquet et al., 1984). Indeed, exogenous PLA, have been shown to stimulate numerous secretory responses (Burgoyne et al., 1987). The fact that the extent of CB internalization in the presence of CA was correlated with an increased stimulation of casein secretion suggests a specific action of crotoxin on the secretory system of mammary epithelial cells. It emphasizes the importance of the subcellular localization of PLA, for the production of AA. All eukaryotic tissues possess intracellular PLA, located on plasma membranes, lysosomes, mitochondria, etc. (Van den Bosch, 1980) and it is likely that the activity of one or several of these enzymes is regulated through hormonal receptors. Since PRL is very rapidly internalized in the lactating mammary epithelial cell and entrapped by the Golgi apparatus, secretory vesicles and lysosomes (Kane et al., 1983), it would be interesting to determine whether the same intracellular compartments are involved in the action of hormone and of PLA,. Comparison with the action of the toxin in the blockade of synaptic transmission at the neuromuscular junction (Bon et al., 1979) showed that
so the step of binding of CB in the presence of CA seemed similar in the two systems. However, differences appear in the mode of action of crotoxin since blockade of neuromuscular transmission results from its effects on voltage-sensitive ions channels (Hawgood and Santana de Sa, 1979) whereas stimulation of casein probably involves the eicosanoids pathway. Furthermore, internalization of crotoxin does not seem to be implied at the neuromuscular junction (Trivedi et al., 1989) while it is quite evident in lactating mammary epithelial cells. In conclusion, crotoxin is able to mimic the secretagogue effect of PRL on mammary epithelial cells and appears to be a useful tool to study the regulation of secretory processes in these cells. Acknowledgments The authors thank Dr. A. Tixer-Vidal and Dr. J. Massoulie for their helpful comments upon the manuscript and M. Ahmed Ali and S. Delpal for excellent technical assistance. References Abou-Samra, A.B., Catt, J. and Aguilera, G. (1986) Endocrinology 119, 1427-1431. Balsinde, J., Die,, E., Schuller, A. and Mollinedo, F. (1988) J. Biol. Chem. 263. 1929-1936. Bar-Sagi, D., Suhan, J.P., McCormick, F. and Feramisco, J.R. (1988) J. Cell Biol. 106. 1649-1658. Blachier, F., Lacroix, M.C., Ahmed-Ali, M., Leger, C. and Ollivier-Bousquet, M (1988) Prostaglandins 35, 259-276. Bon, C., Changeux, J.P.. Jeng, T.W. and Fraenkel-Conrat, H. (1979) Eur. J. Biochem. YY, 471-481. Breithaupt, H.. Riibsamen, K., Walsch, P. and Haberman, E. (1971) Naunyn-Schmied. Arch. Pharmacol. 269, 403-404. Burgoyne. R.D., Cheek, T.R. and O’Sullivan, A.J. (1987) Trends Biochem. Sci. 12, 3322333. Canonico, L., Schettini, G., Valdenegro, CA. and McLeod, R.M. (19X3) Neuroendocrinology 37, 212-217.
Chi, E.Y., Henderson, W.R. and Lebanoff. S.J. (1982) Lab. Invest. 47, 579-585. Choumet, V., Jiang, M.S., Radvanyi, F., Ownby, C. and Bon, C. (1989) FEBS Lett. 244. 367-173. Dusanter-Fourt, I., Djiane, J. and Houdebine, L.M. (19X3) Mol. Cell. Endocrinol. 31, 2X7-299. Grandison. L. (1984) Endocrinology 114, 1-7. Hawgood, B.J. and Santana de Sa. S. (1979) Neurosciences 4, 293-303. Hendon. R.A. and Fraenkel-Conrat, H. (1971) Proc. Nat]. Acad. Sci. U.S.A. 6X. 1560-1563. Hirst. J.J., Rice, G.E., Jenkins, G. and Thornburn, G.D. (1988) Endocrinology 122, 774-781. Kane, S., Raymond, M.N., Dusanter-Fourt, I., Houdebine. L.M., Djiane, J. and Ollivier-Bousquet, M. (1983) Eur. J. Cell. Biol. 30, 24442.53. Kojima. I., Kojima, K. and Rasmussen. H. (1986) Endocrinology 117, 1057-1066. Leung, K.H. (1986) Biochem. Biophys. Res. Commun. 137, 195-200. McLean, I.W. and Nakane. P.K. (1974) J. Histochem. Cytochem. 22. 1077-1083. Metz, S.A. (1988) Prostaglandins Leukotrienes Essential Fatty Acids 32, 187-202. Naor, Z. and Catt. K.J. (1981) J. Biol. Chem. 256, 2226-222’). Ollivier-Bousquet, M. (1978) Cell. Tissue Res. 187, 25-43. Ollivier-Bousquet, M. (1984) Biol. Cell 51, 327-334. Ollivier-Bousquet, M. and Lacroix, M.C. (1986) Reprod. Nutr. Dev. 26. 575-582. Rillema, J.A.. Wing, L.Y.C. and Foley, K.A. (1983) Endocrinology 113, 202442028. Rillema, J.A., Etindi, R.N. and Cameron, M. (1986) Horm. Metab. Res. 18, 672-674. Ross, P.C., Judd. A.M. and McLeod, R.M. (1088) Endocrinology 123, 2445-2453. Schontardt, T. and Ferber. E. (1987) B&hem. Biophys. Res. Commun. 149. 7699775. Trivedi. S.. Kaiser, I.]., Tanaka, M. and Simpson, L. (1989) J. Pharmacol. Exp. Ther. 25 1, 490-496. Van den Bosch, H. (1980) Biochim. Biophys. Acta 604. 19l246. Van Wayne, J. and Goosens, J. (1983) Prostaglandins 26, 725-730. Wang, J. and Leung. P.C.K. (1988) Endocrinology 122, 9066 911. Yamamoto, S., Nakai, T., Nakade. T. and Kato, R. (1983) Eur. J. Pharmacol. 86, 121-124.
