0013-7227/90/1264-2125$02.00/0 Endocrinology Copyright © 1990 by The Endocrine Society
Vol. 126, No. 4 Printed in U.S.A.
Vectorial Secretion of Prostaglandins by Polarized Rodent Uterine Epithelial Cells* ANDREW L. JACOBS, GLENN L. DECKER, STANLEY R. GLASSER, JOANNE JULIAN, AND DANIEL D. CARSON Department of Biochemistry and Molecular Biology, University of Texas M.D. Anderson Cancer Center (A.L.J., G.L.D., D.D.C.), Houston, Texas 77030; and the Department of Cell Biology, Baylor College of Medicine (S.R.G., J.J.), Houston, Texas 77030
ABSTRACT. Uterine epithelial cells (UEC) isolated from mature mice as well as immature mice and rats were cultured on EHS matrix-coated nitrocellulose filters in order to determine their ability to secrete prostaglandin (PG) F2a and PGE2 in a polarized manner. Ultrastructural analyses were performed to validate the polar nature of mouse UEC and demonstrate the presence of separate apical and basolateral plasma membrane domains. These properties included the presence of tightly juxtaposed lateral membranes, apical microvilli, and a relatively flat basal surface. Biochemical indices of polarity included the preferential (~5:1) basal uptake of [35S]methionine as well as a preferential (~9:1) apical secretion of protein. UEC isolated from mice during the estrous and diestrous stages of the estrous cycle did not differ in their degree of polarity, as measured by these morphological and biochemical indices. UEC of estrous and diestrous mice as well as immature mice and rats preferentially secreted PGF2n to the basal medium to an approximately 4-fold greater extent than to the apical medium. PGE2 was secreted at least 10-fold less than PGF2n) and a preferential basal secretion could not be demonstrated. Polarized UEC accumulated relatively large cellular pools of PGF2n, while nonpolarized cells
grown on matrix-coated plastic did not. This difference was reflected by the inability of an inhibitor of PG biosynthesis, indomethacin, to inhibit PGF2o secretion by polarized cells during short (4-h) incubations. In contrast, this drug effectively inhibited secretion in nonpolarized cells or polarized cells incubated with indomethacin for longer (24-h) intervals. Therefore, cellular PGF2a pools apparently support continued secretion of this lipid even when de novo synthesis is transiently inhibited. Preferential basal secretion of PGF2owas due to the polar nature of UEC, since disruption of tight junctions with EGTA modified the basal to apical ratio of PGF2a secretion to near unity. Sodium azide inhibited the secretion of PGF290%), and the extracts were redissolved in acetonitrile-acetic acid (100:0.05) and applied to the liquid chromatograph system (Beckman Instruments, Inc., Palo Alto, CA), which consisted of two model 100A pumps controlled by a model 421 controller. Chromatography was performed on a 0.46 X 25-cm column of octadecylsilane, equilibrated with H2O-acetic acid (100:0.05; vol/vol; buffer A) and developed with a linear gradient of acetonitrile-acetic acid (100:0.05; vol/vol; buffer B). The ratio of buffers A:B at time zero was 75:25, and buffer B was increased linearly to 50% by 60 min. At this time the concentration of buffer B was increased to 100% for the remaining 10 min. The buffer was pumped at room temperature at a flow rate of 1 ml/min, with a back pressure of approximately 1600 psi. Fractions were collected every 1 min. The elution profile of radioactive samples was determined by liquid scintillation counting. Recovery of radioactivity was more than 90%. The elution profile of nonradioactive standards was monitored at 195 nm using a Hitachi flow-through, variable wavelength monitor (Hitachi, Tokyo, Japan) interfaced with a CRIA integration. PG analysis UEC isolated from estrous and diestrous stage mice as well as immature mice and rats were cultured as described above. On the first day of confluency (day 2, mice; day 4, rats) medium in the apical and basal secretory compartments was removed, and both compartments were washed gently several times before fresh medium was added. After the indicated time of incubation, medium was collected from each compartment and frozen for subsequent analysis for PGF2«, 13,14-dihydro-15keto-PGF2a (PGFM), and PGE2 by RIA. Cells on filters were rinsed with PBS containing indomethacin (20 Mg/ml) and frozen before analysis for PGs.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 August 2016. at 17:56 For personal use only. No other uses without permission. . All rights reserved.
VECTORIAL SECRETION OF PGs
2128 PGF2a assay
Secreted PGF2« was measured in heat-dried ethyl ether extracts of 75-^1 aliquots of acidified (20 jul 1 N HC1) medium using a PGF2« assay kit purchased from Amersham Corp. Cellassociated PGF2a was measured in heat-dried ethyl ether extracts from filters by adding 20 i*\ 1 N HC1 to each filter and vortexing filters in 2 ml ethyl ether. Extracts were redissolved in 200 /A assay buffer [0.05 M phosphate buffer, pH 7.3, plus 0.1% (wt/vol) gelatin and 0.01% (wt/vol) merthiolate] before [3H]PGF2« (60 nh -12,000 dpm) and rabbit anti-PGF2« (60 /x\; diluted 1:5,000) were added. Tubes were incubated at 4 C overnight before bound and unbound [3H]PGF2« were separated by the addition of 500 n\ dextran-coated charcoal [1% (wt/vol) Norit-A charcoal plus 0.5% (wt/vol) Dextran T-70 in 0.01 M phosphate buffer, pH 7.3] and centrifugation at 2,000 X g for 20 min, after which the supernatant was removed for quantification of radioactivity. Standards of PGF2« ranged from 6.1200 pg in a 200-jul volume of assay buffer. These tubes were processed identically to unknown samples. The cross-reactivity of the rabbit anti-PGF2o, was insignificant with other major arachidonic acid metabolites. Validation of this assay included the demonstration of accurate recovery of PGF2« added to culture medium and an inhibition of binding by serial dilutions of culture medium containing PGF2a in a parallel manner to the standard curve. Assays of ethyl ether extracts of culture medium, filter inserts, or Matrigel demonstrated no detectable PGF2« in any of these components. The intra- and interassay coefficients of variation for this assay were both 10% or less. The sensitivity of this assay was 3 pg/tube.
Endo • 1990 Vol 126 «No 4
variation were both 10% or less. The sensitivity of this assay was 8.2 pg/tube.
Results Homogeneity of epithelial cell cultures Indirect immunofluorescence staining with antibodies specific for cytokeratin-18 (ENDO B) and uvomorulin was performed after 72 h of culture on matrix-coated glass coverslips (Fig. 1). The characteristic appearance of confluent UEC is apparent before staining (Fig. 1A). It was difficult to identify any cells that were not positively stained for cytokeratin-18 (Fig. 1C) or uvomorulin (Fig. ID) in these cell cultures. Similar results were obtained when staining was performed on epithelial cells cultured on matrix-coated filters, which were cleared with xylene before viewing (data not shown). Indirect immunofluorescence staining using normal rabbit serum instead of primary antibody was negative (Fig. IB). Based on immunohistochemical staining, cells isolated and cultured in the manner described (see Materials and Methods) were determined to be more than 95% epithelial in nature. Indices of morphological and functional polarity: ultrastructure of UEC cultured on matrix-coated filters Transmission electron microscopy was used to determine the extent to which filter-cultured cells displayed
PGFM assay Concentrations of PGFM present in media from the apical and basal secretory compartments as well as cell-associated PGFM were measured by a RIA, described in detail previously (28). Before the assay, media and filters were processed as described for PGF2«. PGE2 assay
Secreted PGE2 was measured in ethyl ether extracts as described for PGF2a using a PGE2 assay kit purchased from Advanced Magnetics, Inc. (Cambridge, MA). Extracts were dissolved in 200 n\ assay buffer [0.01 M phosphate buffer, pH 7.0, plus 0.1% (wt/vol) bovine 7-globulin and 0.1% (wt/vol) sodium azide] before [3H]PGE2 (100 /xl; -15,000 dpm) and rabbit antiPGE2 (100 MI; 1:6,000 dilution) were added. Tubes were incubated overnight at 4 C, and unbound PGE2 was removed by the addition of dextran-coated charcoal and centrifugation, as described for PGF2a. Standards ranged from 8.2-2000 pg in a volume of 200 n\ and were processed in the same manner as unknown samples. The relative cross-reactivity with PGEi at 50% displacement of ligand was 60%. Therefore, concentrations of PGE2 are expressed as PGE. Cross-reactivity was insignificant with other arachidonic acid metabolites. Validation included the recovery of PGE2 added to culture medium and the measurement of PGE2 in serial dilutions of culture medium, as described for PGF2« The intra- and interassay coefficients of
FIG. 1. Indirect immune-fluorescent staining for cytokeratin-18 and uvomorulin in cultured mouse UEC. Random cycling mouse UEC cells cultured on glass coverslips for 72 h were processed for indirect immunofluorescent staining as described in Materials and Methods. A, Characteristic appearance of confluent mouse UEC; B, immunofluorescent staining using normal rabbit serum instead of primary antibody; C, staining with anti-Endo-B directed against cytokeratin-18; D, staining with antiuvomorulin. The general reactivity of these cells with antisera to cytokeratin-18 and uvomorulin demonstrated the homogeneity of the epithelial cell cultures used in these studies.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 August 2016. at 17:56 For personal use only. No other uses without permission. . All rights reserved.
VECTORIAL SECRETION OF PGs
ultrastructural features characteristic of UEC in situ (11). UEC processed for electron microscopy on day 2 of culture displayed distinct apical and basal cell surface domains as well as several other features indicative of differentiation to a polarized state (Fig. 2). Specific structures associated with the attainment of polarity in these cells included the presence of apical microvilli (Fig. 2, A and C), a relatively flattened basal surface associated with a thin extracellular lamina (Fig. 2A), tight lateral cellular junctions with the frequent appearance of desmosomes (Fig. 2, A-C), and the presence of an apical glycocalyx (Fig. 2C). When the ultrastructure of cells isolated from estrous and diestrous stage mice were com-
2129
pared, these cells displayed the same morphological characteristics as those described above, with no obvious differences between stages of the estrous cycle (data not shown). Basal vs. apical uptake of P5S]methionine The ability of UEC to incorporate [35S]methionine presented to either the apical or basal surface was examined (Fig. 3). In addition, it was of interest to determine if the hormonal state at the time uteri were removed influenced this index of polarity. UEC from estrous, diestrous, and immature mice that had not yet reached confluency (day 1) exhibited a similar rate of uptake of [35S]methionine from the apical and basal surfaces. Once confluency had been achieved, the uptake of [35S]methionine in all three cell stages reached a basal/apical ratio of approximately 5 (~200,000:40,000 dpm), which was maintained throughout culture. Cells from immature mice exhibited a similar pattern, although uptake was transiently higher on day 4 of culture. Thus, UEC from mice cultured on matrix-coated filters exhibited a preferential basal uptake of [35S]methionine associated with the attainment of polarity. Protein secretion to apical and basal compartments As an additional biochemical index of polarity, [35S] methionine incorporation into protein by UEC and the relative proportion of protein secreted into the apical and basal secretory compartments were determined. When cells from randomly cycling mice were labeled
Vol. 126, No. 4 Printed in U.S.A.
Vectorial Secretion of Prostaglandins by Polarized Rodent Uterine Epithelial Cells* ANDREW L. JACOBS, GLENN L. DECKER, STANLEY R. GLASSER, JOANNE JULIAN, AND DANIEL D. CARSON Department of Biochemistry and Molecular Biology, University of Texas M.D. Anderson Cancer Center (A.L.J., G.L.D., D.D.C.), Houston, Texas 77030; and the Department of Cell Biology, Baylor College of Medicine (S.R.G., J.J.), Houston, Texas 77030
ABSTRACT. Uterine epithelial cells (UEC) isolated from mature mice as well as immature mice and rats were cultured on EHS matrix-coated nitrocellulose filters in order to determine their ability to secrete prostaglandin (PG) F2a and PGE2 in a polarized manner. Ultrastructural analyses were performed to validate the polar nature of mouse UEC and demonstrate the presence of separate apical and basolateral plasma membrane domains. These properties included the presence of tightly juxtaposed lateral membranes, apical microvilli, and a relatively flat basal surface. Biochemical indices of polarity included the preferential (~5:1) basal uptake of [35S]methionine as well as a preferential (~9:1) apical secretion of protein. UEC isolated from mice during the estrous and diestrous stages of the estrous cycle did not differ in their degree of polarity, as measured by these morphological and biochemical indices. UEC of estrous and diestrous mice as well as immature mice and rats preferentially secreted PGF2n to the basal medium to an approximately 4-fold greater extent than to the apical medium. PGE2 was secreted at least 10-fold less than PGF2n) and a preferential basal secretion could not be demonstrated. Polarized UEC accumulated relatively large cellular pools of PGF2n, while nonpolarized cells
grown on matrix-coated plastic did not. This difference was reflected by the inability of an inhibitor of PG biosynthesis, indomethacin, to inhibit PGF2o secretion by polarized cells during short (4-h) incubations. In contrast, this drug effectively inhibited secretion in nonpolarized cells or polarized cells incubated with indomethacin for longer (24-h) intervals. Therefore, cellular PGF2a pools apparently support continued secretion of this lipid even when de novo synthesis is transiently inhibited. Preferential basal secretion of PGF2owas due to the polar nature of UEC, since disruption of tight junctions with EGTA modified the basal to apical ratio of PGF2a secretion to near unity. Sodium azide inhibited the secretion of PGF290%), and the extracts were redissolved in acetonitrile-acetic acid (100:0.05) and applied to the liquid chromatograph system (Beckman Instruments, Inc., Palo Alto, CA), which consisted of two model 100A pumps controlled by a model 421 controller. Chromatography was performed on a 0.46 X 25-cm column of octadecylsilane, equilibrated with H2O-acetic acid (100:0.05; vol/vol; buffer A) and developed with a linear gradient of acetonitrile-acetic acid (100:0.05; vol/vol; buffer B). The ratio of buffers A:B at time zero was 75:25, and buffer B was increased linearly to 50% by 60 min. At this time the concentration of buffer B was increased to 100% for the remaining 10 min. The buffer was pumped at room temperature at a flow rate of 1 ml/min, with a back pressure of approximately 1600 psi. Fractions were collected every 1 min. The elution profile of radioactive samples was determined by liquid scintillation counting. Recovery of radioactivity was more than 90%. The elution profile of nonradioactive standards was monitored at 195 nm using a Hitachi flow-through, variable wavelength monitor (Hitachi, Tokyo, Japan) interfaced with a CRIA integration. PG analysis UEC isolated from estrous and diestrous stage mice as well as immature mice and rats were cultured as described above. On the first day of confluency (day 2, mice; day 4, rats) medium in the apical and basal secretory compartments was removed, and both compartments were washed gently several times before fresh medium was added. After the indicated time of incubation, medium was collected from each compartment and frozen for subsequent analysis for PGF2«, 13,14-dihydro-15keto-PGF2a (PGFM), and PGE2 by RIA. Cells on filters were rinsed with PBS containing indomethacin (20 Mg/ml) and frozen before analysis for PGs.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 August 2016. at 17:56 For personal use only. No other uses without permission. . All rights reserved.
VECTORIAL SECRETION OF PGs
2128 PGF2a assay
Secreted PGF2« was measured in heat-dried ethyl ether extracts of 75-^1 aliquots of acidified (20 jul 1 N HC1) medium using a PGF2« assay kit purchased from Amersham Corp. Cellassociated PGF2a was measured in heat-dried ethyl ether extracts from filters by adding 20 i*\ 1 N HC1 to each filter and vortexing filters in 2 ml ethyl ether. Extracts were redissolved in 200 /A assay buffer [0.05 M phosphate buffer, pH 7.3, plus 0.1% (wt/vol) gelatin and 0.01% (wt/vol) merthiolate] before [3H]PGF2« (60 nh -12,000 dpm) and rabbit anti-PGF2« (60 /x\; diluted 1:5,000) were added. Tubes were incubated at 4 C overnight before bound and unbound [3H]PGF2« were separated by the addition of 500 n\ dextran-coated charcoal [1% (wt/vol) Norit-A charcoal plus 0.5% (wt/vol) Dextran T-70 in 0.01 M phosphate buffer, pH 7.3] and centrifugation at 2,000 X g for 20 min, after which the supernatant was removed for quantification of radioactivity. Standards of PGF2« ranged from 6.1200 pg in a 200-jul volume of assay buffer. These tubes were processed identically to unknown samples. The cross-reactivity of the rabbit anti-PGF2o, was insignificant with other major arachidonic acid metabolites. Validation of this assay included the demonstration of accurate recovery of PGF2« added to culture medium and an inhibition of binding by serial dilutions of culture medium containing PGF2a in a parallel manner to the standard curve. Assays of ethyl ether extracts of culture medium, filter inserts, or Matrigel demonstrated no detectable PGF2« in any of these components. The intra- and interassay coefficients of variation for this assay were both 10% or less. The sensitivity of this assay was 3 pg/tube.
Endo • 1990 Vol 126 «No 4
variation were both 10% or less. The sensitivity of this assay was 8.2 pg/tube.
Results Homogeneity of epithelial cell cultures Indirect immunofluorescence staining with antibodies specific for cytokeratin-18 (ENDO B) and uvomorulin was performed after 72 h of culture on matrix-coated glass coverslips (Fig. 1). The characteristic appearance of confluent UEC is apparent before staining (Fig. 1A). It was difficult to identify any cells that were not positively stained for cytokeratin-18 (Fig. 1C) or uvomorulin (Fig. ID) in these cell cultures. Similar results were obtained when staining was performed on epithelial cells cultured on matrix-coated filters, which were cleared with xylene before viewing (data not shown). Indirect immunofluorescence staining using normal rabbit serum instead of primary antibody was negative (Fig. IB). Based on immunohistochemical staining, cells isolated and cultured in the manner described (see Materials and Methods) were determined to be more than 95% epithelial in nature. Indices of morphological and functional polarity: ultrastructure of UEC cultured on matrix-coated filters Transmission electron microscopy was used to determine the extent to which filter-cultured cells displayed
PGFM assay Concentrations of PGFM present in media from the apical and basal secretory compartments as well as cell-associated PGFM were measured by a RIA, described in detail previously (28). Before the assay, media and filters were processed as described for PGF2«. PGE2 assay
Secreted PGE2 was measured in ethyl ether extracts as described for PGF2a using a PGE2 assay kit purchased from Advanced Magnetics, Inc. (Cambridge, MA). Extracts were dissolved in 200 n\ assay buffer [0.01 M phosphate buffer, pH 7.0, plus 0.1% (wt/vol) bovine 7-globulin and 0.1% (wt/vol) sodium azide] before [3H]PGE2 (100 /xl; -15,000 dpm) and rabbit antiPGE2 (100 MI; 1:6,000 dilution) were added. Tubes were incubated overnight at 4 C, and unbound PGE2 was removed by the addition of dextran-coated charcoal and centrifugation, as described for PGF2a. Standards ranged from 8.2-2000 pg in a volume of 200 n\ and were processed in the same manner as unknown samples. The relative cross-reactivity with PGEi at 50% displacement of ligand was 60%. Therefore, concentrations of PGE2 are expressed as PGE. Cross-reactivity was insignificant with other arachidonic acid metabolites. Validation included the recovery of PGE2 added to culture medium and the measurement of PGE2 in serial dilutions of culture medium, as described for PGF2« The intra- and interassay coefficients of
FIG. 1. Indirect immune-fluorescent staining for cytokeratin-18 and uvomorulin in cultured mouse UEC. Random cycling mouse UEC cells cultured on glass coverslips for 72 h were processed for indirect immunofluorescent staining as described in Materials and Methods. A, Characteristic appearance of confluent mouse UEC; B, immunofluorescent staining using normal rabbit serum instead of primary antibody; C, staining with anti-Endo-B directed against cytokeratin-18; D, staining with antiuvomorulin. The general reactivity of these cells with antisera to cytokeratin-18 and uvomorulin demonstrated the homogeneity of the epithelial cell cultures used in these studies.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 August 2016. at 17:56 For personal use only. No other uses without permission. . All rights reserved.
VECTORIAL SECRETION OF PGs
ultrastructural features characteristic of UEC in situ (11). UEC processed for electron microscopy on day 2 of culture displayed distinct apical and basal cell surface domains as well as several other features indicative of differentiation to a polarized state (Fig. 2). Specific structures associated with the attainment of polarity in these cells included the presence of apical microvilli (Fig. 2, A and C), a relatively flattened basal surface associated with a thin extracellular lamina (Fig. 2A), tight lateral cellular junctions with the frequent appearance of desmosomes (Fig. 2, A-C), and the presence of an apical glycocalyx (Fig. 2C). When the ultrastructure of cells isolated from estrous and diestrous stage mice were com-
2129
pared, these cells displayed the same morphological characteristics as those described above, with no obvious differences between stages of the estrous cycle (data not shown). Basal vs. apical uptake of P5S]methionine The ability of UEC to incorporate [35S]methionine presented to either the apical or basal surface was examined (Fig. 3). In addition, it was of interest to determine if the hormonal state at the time uteri were removed influenced this index of polarity. UEC from estrous, diestrous, and immature mice that had not yet reached confluency (day 1) exhibited a similar rate of uptake of [35S]methionine from the apical and basal surfaces. Once confluency had been achieved, the uptake of [35S]methionine in all three cell stages reached a basal/apical ratio of approximately 5 (~200,000:40,000 dpm), which was maintained throughout culture. Cells from immature mice exhibited a similar pattern, although uptake was transiently higher on day 4 of culture. Thus, UEC from mice cultured on matrix-coated filters exhibited a preferential basal uptake of [35S]methionine associated with the attainment of polarity. Protein secretion to apical and basal compartments As an additional biochemical index of polarity, [35S] methionine incorporation into protein by UEC and the relative proportion of protein secreted into the apical and basal secretory compartments were determined. When cells from randomly cycling mice were labeled
