0013-7227/99/$03.00/0 Endocrinology Copyright © 1999 by The Endocrine Society
Vol. 140, No. 6 Printed in U.S.A.
Uterine Natural Killer Cells Are Targets for a Trophoblast Cell-Specific Cytokine, Prolactin-Like Protein A* ¨ LLER†, BING LIU, B. ANNE CROY, JUDITH R. HEAD, HEINER MU JOAN S. HUNT, GUOLI DAI, AND MICHAEL J. SOARES Departments of Molecular and Integrative Physiology (H.M., B.L., G.D., M.J.S.) and Anatomy and Cell Biology (J.S.H.), University of Kansas Medical Center, Kansas City, Kansas 66160; the Department of Biomedical Sciences, Ontario Veterinary College, University of Guelph (B.A.C.), Guelph, Ontario, Canada N1G 2W1; and the Green Center for Reproductive Biology Sciences, University of Texas Southwestern Medical Center (J.R.H.), Dallas, Texas 75235-9051 ABSTRACT PRL-like protein A (PLP-A) is a member of the PRL family expressed in trophoblast cells coincident with establishment of the chorioallantoic placenta. The purpose of this investigation was to identify targets for PLP-A. Using an alkaline phosphatase-tagging strategy, we show that PLP-A specifically interacts with a population of natural killer (NK) lymphocytes within the mesometrial compartment of decidua from pregnant and pseudopregnant rats. These observations are supported by the codistribution of PLP-A targets with cells ex-
T
ROPHOBLAST cells are essential for modification of the maternal uterine environment into a hospitable site for embryonic and fetal development. This important task is accomplished at least in part via the elaboration of cytokines/hormones belonging to the PRL/GH family (1). In the rat and mouse, this cytokine family contains at least 15 different members (1). At this juncture, information is available on target cells and biological actions for only six members of the family, including those with luteotropic and lactogenic actions (1) and those regulating angiogenesis (2). This includes PRL family members activating the PRL receptor [PRL and placental lactogen (PL) I, Iv, and II] (1) and those interacting with endothelial cells (proliferin and proliferinrelated protein) (2). Targets and actions for the remaining members of the PRL family are presently unknown. Based on their homology, abundance, and cell- and temporal-specific patterns of expression during pregnancy, we hypothesize that each member of the uteroplacental PRL family participates in the paracrine/endocrine signaling required for ensuring viviparity. Received December 8, 1998. Address all correspondence and requests for reprints to: Michael J. Soares, Ph.D., Department of Molecular and Integrative Physiology, University of Kansas Medical Center, 39th and Rainbow Boulevard, Kansas City, Kansas 66160. E-mail:
[email protected]. * This work was supported by grants from the NICHHD (HD-02528, HD-20676, HD-29036, HD-29797, HD-32552, and HD-33994,) and from the Natural Sciences and Engineering Council, Canada. † Supported by a fellowship from the Deutsche Forschungsgemeinschaft of Germany (Mu 1183/1–1). Present address: Department of Obstetrics and Gynecology, University of Rostock, Doberaner Strasse 142, 18057 Rostock, Germany.
pressing the rat NK cell surface marker, gp42, the absence of PLP-A binding in conceptuses from NK cell-deficient tge26 mice, and the specific interaction of PLP-A with a rat NK cell line, RNK-16. We have further demonstrated that PLP-A effectively suppresses RNK-16 cell cytolytic activities. Our results provide evidence for a new paradigm of embryonic-maternal communication involving a PLP-A signaling pathway between trophoblast cells and uterine NK lymphocytes. (Endocrinology 140: 2711–2720, 1999)
We have implemented a strategy of tagging members of the uteroplacental PRL family with alkaline phosphatase (AP). These AP-cytokine fusion proteins have facilitated the identification of target tissues for members of the uteroplacental PRL family (3). In the present report we begin to explore the biology of one member of the PRL family referred to as PRL-like protein A (PLP-A). PLP-A is expressed by trophoblast cells at the ontogeny of chorioallantoic placenta formation (4 – 8). PLP-A is considered a nonclassical PRL family member, in that it does not use the PRL receptor (1, 9). At present, the target cells and actions of PLP-A are unknown. Using the AP-tagging strategy, we show that PLP-A specifically interacts with a population of natural killer (NK) lymphocytes within the mesometrial compartment of uteri from pregnant rats and mice. NK cells participate in immune surveillance, where they effectively recognize and kill virally infected and transformed cells and secrete an array of bioeffector molecules (10 –12). NK cells are conspicuous residents of the uterus of rodents and humans during pregnancy (11, 13–15). The distribution and phenotype of uterine NK cells change from implantation to midgestation (11, 13). At midgestation in the mouse and rat, the mesometrial decidual region is characterized by the presence of a prominent population of NK cells (11, 13, 16). These uterine NK cells are juxtaposed to trophoblast cells of the developing chorioallantoic placenta and are distinct in their cytokine and bioeffector secretory profiles and their relative absence of cytolytic activities (17–21). Our data indicate that PLP-A is a candidate regulator of the pregnancy-associated uterine NK cell phenotype.
2711
2712
PLP-A AND NATURAL KILLER CELLS Materials and Methods
Reagents Reagents for PAGE were purchased from Bio-Rad Laboratories, Inc. (Hercules, CA). All restriction enzymes, polymerases, and DNA ligase were purchased from New England Biolabs, Inc. (Beverly, MA). The 293 human fetal kidney, the Chinese hamster ovary (CHO), and the YAC-1 cell lines were obtained from American Type Culture Collection (Manassas, VA). RNK-16 cells were a gift from Dr. Craig W. Reynolds, NCI (Frederick, MD). Dispase was purchased from Boehringer Mannheim (Indianapolis, IN). Superscript preamplification kits, Taq polymerase, and Lipofectamine reagent for transfection were obtained from Life Technologies (Gaithersburg, MD). Transformation-competent Sure bacterial cells, Pfu polymerase, and random primer labeling kits were acquired from Stratagene (La Jolla, CA). DNA extraction kits were purchased from Qiagen (Chatsworth, CA). Nitrocellulose was obtained from Schleicher & Schuell, Inc. (Keene, NH). Ovine PRL was purchased from Nobl Laboratories, Inc. (Sioux City, IA). Bovine GH was obtained from the USDA hormone distribution service. Monoclonal antibody to gp42 (3G7-E hybridoma) was a gift from Dr. William E. Seaman, V.A. Medical Center (San Francisco, CA). T7 DNA sequencing kits were acquired from U.S. Biochemical Corp. (Cleveland, OH). The pCMV/ SEAP vector was acquired from Tropix, Inc. (Bedford, MA). The rat PLP-A complementary DNA (cDNA) was a gift from Dr. Mary Lynn Duckworth of the University of Manitoba (Winnipeg, Canada). Radiolabeled nucleotides were purchased from DuPont NEN (Boston, MA). Avidin-biotin immunoperoxidase kits were purchased from Zymed Laboratories (South San Francisco, CA). Antimouse IgG MicroBeads were obtained from Miltenyi Biotech (Auburn, CA). Reagents for the detection of immune complexes by enhanced chemiluminescence were acquired from Amersham (Arlington Heights, IL). Unless otherwise noted, all other chemicals and reagents were purchased from Sigma Chemical Co. (St. Louis, MO).
Animals and tissue preparation Holtzman rats were obtained from Harlan Sprague Dawley, Inc. (Indianapolis, IN). A breeding colony of transgenic tge26 mice (22) was maintained in the barrier-sustained environment of the isolation unit at the University of Guelph. Timed pregnancies and pseudopregnancies and tissue dissections were performed as previously described (23–26). The presence of a copulatory plug or sperm in the vaginal smear was designated day 0 of pregnancy. The first day of pseudopregnancy was defined as the first day of leukocytic vaginal smears after stimulation. Deciduomal responses were induced in pseudopregnant rats by the injection of 50 –100 ml sesame oil/uterine horn on day 4 of pseudopregnancy (25). Protocols for the care and use of animals were approved by the University of Kansas animal care and use committee.
Cell culture Human fetal kidney 293 cells were used as hosts for the expression of AP fusion proteins. 293 cells were routinely maintained in MEM supplemented with 20 mm HEPES, 100 U/ml penicillin, 100 mg/ml streptomycin, and 10% FBS. Some aspects of NK cell biology were evaluated in the rat NK cell line, RNK-16 (27, 28). RNK-16 cells were maintained in RPMI 1640 culture medium supplemented with HEPES (10 mm), l-glutamine (2 mm), sodium pyruvate (1 mm), nonessential amino acids, 50 mm 2-mercaptoethanol, 100 U/ml penicillin, 100 mg/ml streptomycin, and 10% FBS. YAC-1 cells were used as targets for cytotoxicity assays and were maintained in RPMI 1640 culture medium supplemented with 100 U/ml penicillin, 100 mg/ml streptomycin, and 10% FBS. CHO cells were used as hosts for the PRL receptor expression plasmid and were routinely maintained in DMEM/MCDB-302 culture medium containing 1 mm proline, 100 U/ml penicillin, 100 mg/ml streptomycin, and 10% FBS.
Generation and characterization of the AP-PLP-A fusion protein A fusion protein consisting of a modified human placental AP (PLAP) and rat PLP-A was generated and used to monitor PLP-A target cell interactions. Procedures for generating the fusion protein were similar
Endo • 1999 Vol 140 • No 6
to those previously reported for the generation of an AP-PL-I fusion protein (3). A vector containing ampicillin and neomycin resistance genes and a secreted version of PLAP (SEAP) (29) situated downstream of a cytomegalovirus (CMV) promoter (pCMV/SEAP) was commercially obtained (Tropix). A nucleotide region representing the mature rat PLP-A protein was then amplified and ligated into the pCMV/SEAP vector. Ligation with the PLP-A insert resulted in a CMV promoterdriven vector containing the ligated cDNAs encoding a SEAP-PLP-A fusion protein (AP-PLP-A; see Fig. 1). DNA sequencing of the insert was performed to verify the accuracy of the PCR amplification. After linearization with BglII, the AP-PLP-A construct was electroporated into 293 human fetal kidney cells. After a 2-week selection with 500 mg/ml G418, single clones were isolated by limiting dilution and were screened for AP expression. An unmodified pCMV-SEAP vector (AP) was similarly transfected, selected, and served as a negative control. Transfected 293 cells were cultured in MEM supplemented with 20 mm HEPES, 100 U/ml penicillin, 100 mg/ml streptomycin, and 10% FBS in an atmosphere of 5% CO2-95% air at 37 C in a humidified incubator. After the cells reached confluence, the medium was changed to serum-free MEM and HEPES, further conditioned for 72 h, collected and clarified by centrifugation, sterile filtered (0.22 mm), and stored at 220 C until used. AP activity was measured from conditioned medium via a colorimetric assay (3). Western blot analysis for PLP-A was performed as previously described (9). AP-PLP-A preparations were isolated from conditioned medium using immunoprecipitation with monoclonal antibodies to AP conjugated to agarose (Sigma Chemical Co.). Native PLP-A isolated from junctional zone conditioned medium and recombinant PLP-A isolated from 293 cells were used as controls for the Western blotting (8, 9). PRL-like biological activities were assessed through the use of the rat Nb2 lymphoma cell proliferation assay (30, 31) and through incubation with CHO cells transiently transfected with the long form of the rat PRL receptor (32).
Analysis of AP-PLP-A binding to tissues and cells Tissues were frozen in heptane and stored at 270 C until tissue sections were prepared with the aid of a cryostat. Sections were mounted onto glass slides, washed in a modified HBSS (containing 20 mm HEPES, 0.5 mg/ml BSA, and 0.1% NaN3), and incubated with AP, AP-PLP-A, or AP-PLP-A plus excess recombinant rat PLP-A, ovine PRL, or bovine GH for 75 min (5 mg/ml each). Recombinant rat PLP-A was generated in 293 cells and purified as previously described (8, 9). After incubation, the sections were washed with a modified HBSS supplemented with 0.1% Tween-20, heated at 65 C for 30 min to inactivate endogenous tissue AP activity, and processed for detection of the heat-stable AP activity associated with the fusion proteins, and coverslips were mounted in aqueous mounting medium. The specificity of binding was assessed by incubation with the AP control or addition of various peptide hormones. Cytospins of cultured RNK-16 cells (27, 28), enzymatically dissociated uterine cells (see below), and transfected CHO cells directly adherent to culture plates were similarly processed. CHO cells were transiently transfected with an expression vector containing the long form of the rat PRL receptor (pECE/long) (32) using Lipofectamine according to the manufacturer’s instructions. The AP-PL-I fusion protein (3) was used as a positive control for binding to the rat PRL receptor.
Identification of PLP-A in the midgestation rat conceptus PLP-A messenger RNA was detected in frozen tissue sections from the rat day 9 conceptus using previously described procedures (33). The full-length rat PLP-A cDNA (4) was linearized and used as a template for the synthesis of 35S-labeled sense and antisense RNA probes. PLP-A protein was identified by immunocytochemistry in rat day 9 conceptus tissue sections as previously described (8).
Immunocytochemical localization of gp42 The protein, gp42, is expressed on the surface of interleukin-2-activated rat NK cells and can be specifically recognized with the monoclonal antibody, 3G7-E (34, 35). The gp42 protein was localized in frozen tissue sections of uteri from pregnant and pseudopregnant/decidualized rats using the 3G7-E monoclonal antibody and a streptavidin-biotin
PLP-A AND NATURAL KILLER CELLS
2713
FIG. 1. Construction of the AP-PLP-A expression vector and characterization of the AP-PLP-A fusion protein by Western blotting. A, AP-PLP-A expression vector construction. A fusion protein consisting of a modified human placental AP and rat PLP-A was generated and used to monitor PLP-A target cell interactions. Human placental AP in its native form is a heat-stable membrane-associated AP. The carboxyl-terminus of human placental AP was truncated, resulting in a secreted AP referred to as SEAP and situated downstream of the CMV promoter in a vector containing ampicillin and neomycin resistance genes (pCMV/SEAP; Tropix). A nucleotide region representing the mature rat PLP-A was amplified and ligated into the pCMV/SEAP vector. Ligation with the PLP-A insert resulted in a CMV promoter-driven vector containing the ligated cDNAs encoding a SEAP-PLP-A fusion protein (AP-PLP-A). B, Immunocharacterization of the AP-PLP-A fusion protein. Samples were separated by PAGE in 7.5% gels under reducing conditions and were electrophoretically transferred to nitrocellulose. Polyclonal antibodies generated against PLP-A were used as probes. Immune complexes were detected using the enhanced chemiluminescence system. Lane A, Native PLP-A isolated from day 19 rat junctional zone conditioned medium; lane B, recombinant rat PLP-A; lane C, AP-PLP-A preparations from 293 cells stably transfected with the pCMV AP-PLP-A expression vector; lane D, 293 cells stably transfected with the pCMV AP control expression vector; lane E, nontransfected 293 cells. Note that the AP-PLP-A fusion protein migrated at a Mr approximating 110 kDa, whereas native PLP-A migrated at Mr of 29 and 33 kDa. immunoperoxidase kit for mouse IgG (33, 34). The specificity of the immunoreactions was demonstrated using nonimmune mouse IgG.
Immunomagnetic isolation of rat uterine NK cells The 3G7-E monoclonal antibody to gp42 and magnetic cell sorting with antimouse IgG MicroBeads (Miltenyi Biotech, Auburn, CA) were used to isolate NK cells from uteri of pregnant and pseudopregnant/ decidualized rats. Decidual cells were dissociated from mesometrial decidua isolated on day 10 of gestation or from whole deciduoma isolated on day 10 of pseudopregnancy. Decidual tissue was minced with fine scissors and dispersed in Dispase (2.4 U/ml) containing deoxyribonuclease I (80 U/ml) for 1 h at 37 C. Dispersed cells were recovered by centrifugation and washed with HBSS to remove residual Dispase. Red blood cells were lysed via incubation of the cell suspensions with 10 vol lysis buffer (155 mm NH4Cl, 10 mm NaHCO3, and 0.1 mm EDTA) for 5 min at room temperature. Cells were recovered by centrifugation and washed twice in 10 vol PBS containing 2 mm EDTA and 0.5% BSA (isolation buffer). The cell preparation was adjusted to 107 cells/100 ml and incubated with anti-gp42 (34) for 10 min at 5 C with continual mixing. Cells were then washed twice with 10 vol isolation buffer, recovered by centrifugation, resuspended at a final concentration of 107 cells/100 ml, and incubated with antimouse IgG MicroBeads (Miltenyi Biotech) for 15 min at 5 C. After incubation, the cells were washed, resuspended in a final volume of 500 ml isolation buffer, and loaded onto a Magnetic Separation Plus column (Miltenyi Biotech) affixed to a magnet. The column was washed three times with isolation buffer (500 ml each), and the eluting cells were collected (gp42-immunonegative fraction). The column was then removed from the magnet, and cells were eluted in 1 ml isolation buffer (gp42-immunopositive fraction). Aliquots of cells from both fractions were adhered to glass slides and processed for PLP-A binding.
Cytotoxicity assays NK cell cytolytic activities were monitored by 51Cr release assay (36). YAC-1 target cells were incubated for 1 h at 37 C with 200 mCi 51Cr. Radiolabeled YAC-1 target cells were washed three times and then incubated with RNK-16 cells to evaluate killing activity. The two populations of cells were coincubated at various ratios at 37 C, and after 4 h the amount of 51Cr released into the culture medium was measured with a g-counter. Cells were exposed to medium conditioned by 293 cells mock transfected or transfected with a PLP-A expression vector, or they were exposed to an enriched preparation of recombinant PLP-A (purity, .90% by PAGE). Target cells in medium alone served as spontaneous release controls, and target cells in Triton X-100 served as maximal release controls. Spontaneous release values were subtracted from the test and maximal release values to calculate specific release. Data are presented as the mean 6 sem percentage of maximal killing.
Statistical analysis Data were analyzed by ANOVA. The source of variation from significant F ratios was determined with Student’s two-tailed t test (37).
Results Generation and characterization of the AP-PLP-A fusion protein
As a first step toward evaluating the biology of PLP-A, we generated and characterized an AP-PLP-A fusion protein. The construction of the AP-PLP-A vector was achieved by in-frame insertion of the cDNA sequence of mature rat
2714
PLP-A AND NATURAL KILLER CELLS
PLP-A downstream of the SEAP-coding sequence within the pCMV-SEAP vector (Fig. 1A). The AP-PLP-A construct and an unmodified AP control vector were transfected via electroporation into 293 cells. The presence of AP activity was evaluated in conditioned medium from transfected cells and nontransfected cells using a colorimetric assay. Anti-PLP-A antibodies specifically recognized the AP-PLP-A fusion protein. AP and the AP-PLP-A fusion protein were initially enriched by immunoprecipitation with a monoclonal antibody to human PLAP conjugated to agarose. PLP-A antibodies recognized an AP-PLP-A protein species approximating 110 kDa and native PLP-A protein species of 29 and 33 kDa (Fig. 1B). The 110-kDa Mr of the AP-PLP-A fusion protein was consistent with the predicted Mr of its AP and PLP-A components. The AP control preparation was not recognized by the PLP-A antibodies. The AP-PLP-A fusion protein was also evaluated for its ability to activate the PRL receptor signaling system. Both AP-PLP-A and AP failed to stimulate the proliferation of rat Nb2 lymphoma cells (data not shown) and failed to bind to CHO cells transiently transfected with the long form of the rat PRL receptor (data not shown). In contrast, AP-PL-I controls effectively stimulated rat Nb2 lymphoma cell proliferation and interacted with the PRL receptor (data not shown).
Endo • 1999 Vol 140 • No 6
These observations are consistent with earlier reports that PLP-A does not activate the PRL receptor signaling pathway (9). Identification of PLP-A targets with the AP-PLP-A fusion protein
We next examined the binding of AP-PLP-A to uteroplacental tissues during various stages of gestation. APPLP-A specifically bound to a population of cells situated in the midgestation mesometrial compartment of the deciduum immediately overlying the developing chorioallantoic placenta (Fig. 2, A–C). This distinct binding pattern was first observed on day 9 of gestation in the rat, earlier than reported for PLP-A expression (4 – 6). Consequently, we examined the expression of PLP-A messenger RNA by the rat day 9 conceptus and observed strong hybridization in developing trophoblast (Fig. 2, D and E). Similar localization results were obtained with PLP-A immunocytochemistry. Thus, the expression of the PLP-A ligand coincided with the appearance of cells capable of binding PLP-A. AP-PLP-A binding to cells in the day 9 mesometrial deciduum was specific for PLP-A. Incubation of compa-
FIG. 2. PLP-A binding and PLP-A expression in the midgestation rat conceptus. A, Binding of AP control probe to day 9 rat conceptus. Note the absence of binding. B, Binding of AP-tagged PLP-A (AP-PLP-A) probe to day 9 rat conceptus. Note the presence of discrete binding in the mesometrial decidua. C, Binding of AP-PLP-A and excess PLP-A. Note the successful competition for binding. D, Hematoxylinstained section from day 9 rat conceptus. E, Darkfield microscopy of antisense PLP-A hybridization to trophoblast cells of the day 9 rat conceptus. F, Schematic diagram showing the distribution of NK cells in the midgestation conceptus (26 –29). Analyses were performed on semiserial sections. Magnification: A–C, 310; D and E, 350. M, Myometrium; M.D., mesometrial decidua; AMD, antimesometrial decidua; T, trophoblast; E, embryo.
PLP-A AND NATURAL KILLER CELLS
2715
FIG. 3. Comparison of PLP-A binding with the intrauterine distribution of NK cells. gp42 immunocytochemical localization of rat NK cells (A–C) and PLP-A binding (D–I) within mesometrial decidua from day 10 or 11 of pseudopregnancy (A, D, and G), decidua from day 10 or 11 of pregnancy (B, E, and H), and metrial gland tissue from day 14 of pregnancy (C, F, and I). D and E represent mesometrial decidua from day 10 of pseudopregnancy and pregnancy, respectively. Magnification for A–F, 350. G–I are high magnification micrographs of PLP-A-binding cells surrounding vasculature from day 11 of pseudopregnancy (G), day 11 of pregnancy (H), or day 14 of pregnancy. Magnification for G–I, 3250. Note the similar patterns of distribution of gp42-positive NK cells and PLP-A-binding cells.
2716
PLP-A AND NATURAL KILLER CELLS
Endo • 1999 Vol 140 • No 6
FIG. 4. PLP-A binding to immunoisolated uterine NK cells. PLP-A binding to gp42 negative (A and D) and positive cells (B, C, E, and F) isolated from decidua of day 10 pseudopregnant (A–C) or pregnant (D–F) rats. Magnification for A, B, D, and E, 3200. C and F are high magnification micrographs of B and E. Magnification for C and F, 3800. Note the significant enrichment of PLP-A-binding cells in the gp42-positive cell fractions.
rable tissue sections with AP failed to produce evidence of binding (Fig. 2A), and the binding of AP-PLP-A was specifically competed by the addition of PLP-A (Fig. 2C), but not by coincubation with ovine PRL or bovine GH (data not shown). The distribution of PLP-A binding in the midgestation rat conceptus was reminiscent of the distribution of NK cells in rat and mouse mesometrial decidua (Fig. 2F) (13, 14, 16, 38). To clarify this relationship, we compared PLP-A binding with the localization of rat NK cells using a monoclonal antibody to gp42, a protein expressed on the surface of activated rat NK cells (34, 35). gp42 has been previously shown to be expressed on differentiated rat uterine NK cells (13).
Similar distributions of PLP-A binding and gp42 immunoreactivity were observed in decidual tissues from days 9 –11 of pseudopregnancy or pregnancy and in the metrial gland from day 14 of pregnancy (Fig. 3). The number of gp42positive cells always outnumbered the number of AP-PLPA-binding cells, especially on days 9 –11 of pregnancy or pseudopregnancy. Of particular interest was the association of gp42-positive cells and PLP-A-binding cells with the vasculature of the mesometrium and the developing metrial gland (Fig. 3). In specimens from pregnant females, uterine NK cells radiated dramatically from developing trophoblast cells and were prominently associated with mesometrial vasculature, whereas NK cells developing in the absence of
PLP-A AND NATURAL KILLER CELLS
2717
FIG. 5. PLP-A interactions with normal and NK cell-deficient conceptuses and a rat NK cell line, RNK-16. A, Binding of AP-PLP-A to the mesometrial compartment of day 10 normal mouse conceptus; B, binding of AP-PLP-A to mesometrial compartment of day 10 NK cell-deficient tge26 mouse conceptus. Note the significant reduction in binding in the NK cell-deficient tge26 mouse conceptuses. C, Binding of AP to RNK-16 cells. D, Binding of AP-PLP-A to RNK-16 cells. Magnification: A and B, 3125; C and D, 3250. Note the specific and somewhat variable binding of AP-PLP-A to the rat NK cells.
trophoblast were more randomly dispersed throughout the mesometrial compartment. To further investigate the relationship between NK cells and PLP-A target cells, we used a monoclonal antibody to gp42 and an immunomagnetic cell separation strategy to isolate NK cells from decidual tissues of pregnant and pseudopregnant rats. The monoclonal antibody to gp42 facilitated the enrichment of PLP-A-binding cells (Fig. 4). Pseudopregnant/decidualized uteri proved to be a richer source of gp42-positive cells than mesometrial decidua from conceptuses. Additional evidence for uterine NK cells as targets for
PLP-A was generated from binding experiments with midgestation conceptuses (day 10) from NK cell-deficient tge26 mice (22) and a rat NK cell line, RNK-16 (27, 28). AP-PLP-A specifically bound to cells in the mesometrial compartment of controls, but not in the NK cell-deficient conceptuses (Fig. 5). Finally, AP-PLP-A specifically bound to cells of the rat NK cell line, RNK-16 (Fig. 5). The extent of binding to RNK-16 cells was variable and may relate to their differentiation state. In summary, the data suggest that PLP-A specifically binds to a population of NK cells located in the mesometrial compartment of the decidua.
2718
PLP-A AND NATURAL KILLER CELLS
Endo • 1999 Vol 140 • No 6
FIG. 6. PLP-A modulation of RNK-16 cell cytolytic activities. Effects of PLP-A on RNK-16 cell cytolytic activity against YAC-1 cells. Cells were exposed to medium conditioned by 293 cells (mock transfected or transfected with a PLP-A expression vector), or they were exposed to an enriched preparation of recombinant PLP-A (;1 mg/ml; purity, .90% by PAGE). Note that control cells killed to a level of 36%, whereas NK cells exposed to PLP-A showed diminished killing. NK cells exposed to 100% 293-PLP-A conditioned medium (293-PLP-A CM) significantly decreased killing (P , 0.05), whereas exposure to the recombinant PLP-A preparation diminished killing even more (P , 0.01). PLP-A did not significantly affect the viability or proliferation rates of RNK-16 or YAC-1 cells (data not shown). CM, Conditioned medium.
Effects of PLP-A on NK cell cytotoxicity
NK cells are components of natural/innate immunity and participate in immune surveillance through their targeting of virus-infected cells, tumor cells, and potentially other cells for destruction without prior sensitization (10 –12). The phenotype of uterine NK cells during pregnancy is characterized by their reduced capacity for killing (11, 17–19, 39). Thus, we examined the effects of PLP-A on killing activity of RNK-16 cells toward YAC-1 cells. PLP-A significantly inhibited RNK-16 cell killing in a dose-dependent manner (Fig. 6) without directly affecting the viability or proliferation of either RNK-16 or YAC-1 cells. The data are consistent with the participation of PLP-A in the regulation of the pregnancyspecific uterine NK cell phenotype. Discussion
The uteroplacental PRL family is emerging as a family of hormones/cytokines pivotal to the establishment of pregnancy (1, 2, 40). In this report we have shown that the trophoblast-specific cytokine, PLP-A, specifically interacts with a population of cells situated in the mesometrial decidual compartment of both pregnant and pseudopregnant uteri. The distribution of cells binding PLP-A mirrors the intrauterine distribution of NK cells. In fact, PLP-A specifically binds to NK cells isolated from the uterus and from a rat NK cell line. Uterine NK cells have been proposed to participate in immunological adjustments to pregnancy and in the establishment of the chorioallantoic placenta (13, 14, 38). During pregnancy, NK cells undergo considerable morphological and functional changes, creating, in effect, a NK cell with a unique phenotype. Midgestation uterine NK cells are distinct in their cytokine and bioeffector secretory profile and their relative absence of cytolytic activities (17–21). Further evidence from genetically mutant mice possessing defective cytolytic activities clearly indicate that NK cell cytolytic activities are not required for a successful pregnancy (41, 42).
However, there is some evidence suggesting that aberrant uterine NK cell cytolytic activity is associated with pregnancy termination (43– 47). Presumably, dysregulated maternal uterine NK cells target genetically foreign extraembryonic and embryonic cells for destruction. Specific cause and effect relationships between uterine NK cells and pregnancy termination have not been established. Although NK cell cytolytic activities are not required for normal progression of pregnancy, it is apparent from studies on NK celldeficient mice that noncytolytic actions of NK cells are essential for normal uterine vasculature and chorioallantoic placental development (48 –50). Thus, it is evident that uterine NK cell transformation into a pregnancy-dependent phenotype is vital for normal progression of the gestational state. Some of the maturational changes accompanying the pregnancy-dependent uterine NK cell phenotype occur in the absence of extraembryonic and embryonic tissues (51–53), whereas other changes are probably dependent upon modulators produced by extraembryonic sources. The nature of the modulators controlling the pregnancy-dependent uterine NK cell phenotype is yet to be resolved. Several years ago, Parr and co-workers postulated the idea of PRL or PRL-related proteins arising from decidual or placental tissues as regulators of uterine NK cells (54). They proposed that uteroplacental PRL family members participated in transforming uterine NK cells into their pregnancyspecific phenotype. Indeed, ligands for the PRL receptor are secreted throughout gestation via either decidual (decidual luteotrophin) or trophoblast (PL-I, PL-II, and PL-I variant) tissues (1, 55). Furthermore, PRL is known to interact with and influence an array of different immune cells, including NK cells (56, 57). The surfaces of NK cells possess PRL receptors (58), and PRL appears to synergize with interleukin-2 to promote NK cell activation, including killing activities (59). Such actions are not entirely consistent with the midgestation uterine NK cell phenotype, which exhibits limited cytolytic activities (17–19). Another PRL family member in-
PLP-A AND NATURAL KILLER CELLS
vestigated in the present report, PLP-A, is a new candidate regulator of uterine NK cell function during pregnancy. PLP-A specifically binds to uterine and some nonuterine NK cells and can inhibit the cytolytic activities of rat NK cells (present study). PLP-A binding appears to be best associated with NK cell activation, as demonstrated by rat uterine NK cells expressing the activation surface protein, gp42. In further support of a regulatory role for PLP-A during pregnancy, trophoblast cell expression of PLP-A coincides with the appearance of proximally located NK cells with diminished cytolytic activities (Refs. 4 – 8 and 19 and the present study). Whether PLP-A participates in the modulation of uterine NK cell killing or other aspects of the pregnancydependent uterine NK cell phenotype remains to be determined. The nature of PLP-A’s actions on NK cells is not known. PLP-A does not interact with the PRL receptor (Ref. 9 and the present study). Consequently, PLP-A actions are probably mediated by alternative signaling pathways. It is not entirely clear whether specific and unique receptor signaling systems have coevolved with each PRL family member, whether PRL family members use signaling pathways for other known ligands, or whether PRL family members act through receptor-independent mechanisms (e.g. transport proteins, binding proteins, etc.). PGs have been implicated in mediating pregnancy-associated changes in the uterine NK cell phenotype (47, 60, 61) and may represent an intracellular mediator of PLP-A action. In summary, using the AP-tagging strategy, PLP-A is shown to specifically interact with a population of NK cells within the mesometrial compartment of the pregnant uterus. NK cells comprise part of the maternal immune surveillance with potent cytolytic activities and represent a rich source of bioeffector molecules. Appropriate modulation of uterine NK cells is critical for the establishment of vascular connectivity between mother and embryo without immunological rejection of the embryo. We propose that PLP-A participates in the regulation of uterine NK cells during pregnancy. Our results provide evidence for a new paradigm of embryonicmaternal communication involving a PLP-A signaling pathway between trophoblast cells and uterine NK cells. Acknowledgments We thank Drs. Mary Lynn Duckworth, Craig W. Reynolds, William E. Seaman, and Paul Kelly for rat PLP-A cDNA, RNK-16 cell line, gp42 monoclonal antibody, and PRL receptor expression vector, respectively.
References 1. Soares MJ, Dai G, Orwig KE, Peters TJ, Mu¨ller H 1998 The uteroplacental prolactin family and pregnancy. Biol Reprod 58:273–284 2. Jackson D, Volpert OV, Bouck N, Linzer DIH 1994 Stimulation and inhibition of angiogenesis by placental proliferin and proliferin-related protein. Science 266:1581–1584 3. Mu¨ller H, Dai G, Soares MJ 1998 Placental lactogen-I (PL-I) target tissues identified with an alkaline phosphatase-PL-I fusion protein. J Histochem Cytochem 46:737–744 4. Duckworth ML, Peden LM, Friesen HG 1986 Isolation of a novel prolactin-like cDNA clone from developing rat placenta. J Biol Chem 261:10879 –10884 5. Campbell WJ, Deb S, Kwok SCM, Joslin JA, Soares MJ 1989 Differential expression of placental lactogen-II and prolactin-like protein-A in the rat chorioallantoic placenta. Endocrinology 125:1565–1574 6. Duckworth ML, Schroedter IC, Friesen HG 1990 Cellular localization of rat placental lactogen-II and rat prolactin-like proteins A and B by in situ hybridization. Placenta 11:143–155
2719
7. Lin J, Poole J, Linzer DIH 1997 Three new members of the mouse prolactin/ growth hormone family are homologous to proteins expressed in the rat. Endocrinology 138:5541–5549 8. Mu¨ller H, Ishimura R, Orwig KE, Liu B, Soares MJ 1998 Homologues for prolactin-like protein-A and B are present in the mouse. Biol Reprod 58:45–51 9. Deb S, Hamlin GP, Roby KF, Kwok SCM, Soares MJ 1993 Heterologous expression and characterization of prolactin-like protein-A. Identification of serum binding proteins. J Biol Chem 268:3298 –3305 10. Gumperz JE, Parham P 1995 The enigma of the natural killer cell. Nature 378:245–248 11. Whitelaw PF, Croy BA 1996 Granulated lymphocytes of pregnancy. Placenta 17:533–543 12. Yokoyama WM 1996 Natural killer cells and the NK gene complex. In: Herzenburg LA (ed) Weir’s Handbook of Experimental Immunology, ed 5. Blackwell, Cambridge, vol 2:65.1– 65.21 13. Head JR 1996 Uterine natural killer cells during pregnancy in rodents. Nat Immun 15:7–21 14. Liu C-C, Parr EL, Young JD-E 1994 Granulated lymphoid cells of the pregnant uterus: morphological and functional features. Int Rev Cytol 153:105–136 15. King A, Burrows T, Loke YW 1996 Human uterine natural killer cells. Nat Immun 15:41–52 16. Head JR, Kresge CK, Young JD, Hiserodt JC 1994 NKR-P11 cells in the rat uterus: granulated metrial gland cells are of the natural killer cell lineage. Biol Reprod 51:509 –523 17. Stewart IJ 1991 Granulated metrial gland cells: pregnancy specific leukocytes? J Leukocyte Biol 50:198 –207 18. Stewart IJ 1994 Granulated metrial gland cells–not part of the natural killer cell lineage? J Reprod Immunol 26:1–15 19. Croy BA, Gambel P, Rossant J, Wegmann TG 1985 Characterization of murine decidual natural killer (NK) cells and their relevance to the success of pregnancy. Cell Immunol 93:315–326 20. Croy BA, Guilbert LJ, Browne MA, Gough NM, Stinchcomb DT, Reed N, Wegmann TG 1991 Characterization of cytokine production by the metrial gland and granulated metrial gland cells. J Reprod Immunol 19:149 –166 21. Hunt JS, Miller L, Vassmer D, Croy BA 1997 Expression of the inducible nitric oxide synthase gene in mouse uterine leukocytes and potential relationships with uterine function during pregnancy. Biol Reprod 57:827– 836 22. Wang B, Biron C, She J, Higgins K, Sunshine M-J, Lacy E, Lonberg N, Terhorst C 1994 A block in both early T lymphocyte and natural killer cell development in transgenic mice with high-copy numbers of the human CD3E gene. Proc Natl Acad Sci USA 91:9402–9406 23. Soares MJ, Julian JA, Glasser SR 1985 Trophoblast giant cell release of placental lactogens: temporal and regional characteristics. Dev Biol 107:520 –526 24. Faria TN, Deb S, Kwok SCM, Talamantes F, Soares MJ 1990 Ontogeny of placental lactogen-I and placental lactogen-II expression in the developing rat placenta. Dev Biol 141:279 –291 25. DeFeo VJ 1963 Temporal aspects of uterine sensitivity in the pseudopregnant or pregnant rat. Endocrinology 73:488 – 497 26. Rasmussen CA, Hashizume K, Orwig KE, Xu L, Soares MJ 1996 Decidual prolactin-related protein: heterologous expression and characterization. Endocrinology 137:5558 –5566 27. Reynolds CW, Bere EW, Ward JM 1984 Natural killer activity in the rat. III. Characterization of transplantable large granular lymphocyte (LGL) leukemias in the F344 rat. J Immunol 132:534 –540 28. Axberg I, Nose M, Reynolds CW, Wigzell H 1988 Features of the in vitro established granular lymphocyte leukaemia RNK-16. Scand J Immunol 27:271–278 29. Berger J, Hauber J, Hauber R, Geiger R, Cullen BR 1988 Secreted placental alkaline phosphatase: a powerful new quantitative indicator of gene expression in eukaryotic cells. Gene 66:1–106 30. Tanaka T, Shiu RPC, Gout PW, Beer CT, Noble RL, Friesen HG 1980 A new sensitive and specific bioassay for lactogenic hormones: measurement of prolactin and growth hormone in human serum. J Clin Endocrinol Metab 51:1058 –1063 31. Cohick CB, Dai G, Xu L, Deb S, Kamei T, Levan G, Szpirer C, Szpirer J, Kwok SCM, Soares MJ 1996 Placental lactogen I variant utilizes the prolactin receptor signaling pathway. Mol Cell Endocrinol 116:49 –58 32. Ali S, Edery M, Pellegrini I, Lesueur L, Paly J, Djiane J, Kelly PA 1992 The Nb2 form of prolactin receptor is able to activate a milk protein gene promoter. Mol Endocrinol 6:1242–1248 33. Rasmussen CA, Orwig, KE, Vellucci S, Soares MJ 1997 Dual expression of prolactin-related protein in decidua and trophoblast tissues during pregnancy. Biol Reprod 55:647– 654 34. Imboden JB, Eriksson EC, McCutcheon M, Reynolds CW, Seaman WE 1989 Identification and characterization of a cell-surface molecule that is selectively induced on rat lymphokine-activated killer cells. J Immunol 143:3100 –3103 35. Seaman WE, Niemi EC, Stark MR, Goldfien RD, Pollock AS, Imboden JB 1991 Molecular cloning of gp42, a cell-surface molecule that is selectively induced on rat natural killer cells by interleukin 2: glycolipid membrane anchoring and capacity for transmembrane signaling. J Exp Med 173:251–260 36. Kiessling R, Klein E, Wigzell H 1975 “Natural” killer cells in the mouse. I.
2720
37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47.
48. 49.
PLP-A AND NATURAL KILLER CELLS
Cytotoxic cells with specificity for mouse Moloney leukemia cells. Specificity and distribution according to genotype. Eur J Immunol 5:112–117 Keppel G 1973 Design and Analysis. Prentice Hall, Englewood Cliffs Croy BA, Luross JA, Guimond M-J, Hunt JS 1996 Uterine natural killer cells: insights into lineage relationships and functions from studies of pregnancies in mutant and transgenic mice. Nat Immun 15:22–33 King A, Loke YW 1990 Human trophoblast and JEG choriocarcinoma cells are sensitive to lysis by IL-2-stimulated decidual NK cells. Cell Immunol 129:435– 448 Lin J, Linzer DIH, Regulation of erythropoiesis by mouse placental prolactinlike proteins. 80th Annual Meeting of The Endocrine Society, New Orleans LA, 1998 (Abstract P3– 463) Roder J, Duwe AK 1979 The beige mutation in the mouse selectively impairs natural killer cell function. Nature 278:451– 453 Stallmach T, Ehrenstein T, Isenmann S, Muller C, Hengartner H, Kagi D 1995 The role of perforin-expression by granular metrial gland cells in pregnancy. Eur J Immunol 25:3342–3348 DeFougerolles AR, Baines MG 1987 Modulation of the natural killer cell activity in pregnant mice alters the spontaneous abortion rate. J Reprod Immunol 11:147–153 Gendron RL, Baines MG 1988 Infiltrating decidual natural killer cells are associated with spontaneous abortion in mice. Cell Immunol 113:261–267 Gendron RL, Baines MG 1989 Morphometric analysis of the histology of spontaneous fetal resorption in a murine pregnancy. Placenta 10:309 –318 Chaouat G, Menu E, Clark DA, Dy M, Minkowski M, Wegmann TG 1990 Control of fetal survival in CBA 3 DBA/2 mice by lymphokine therapy. J Reprod Fertil 89:447– 458 Lala PK, Scodras JM, Graham CH, Lysiak JJ, Parhar RS 1990 Activation of maternal killer cells in the pregnant uterus with chronic indomethacin therapy, IL-2 therapy, or a combination therapy is associated with embryonic demise. Cell Immunol 127:368 –381 Guimond M-J, Luross JA, Wang B, Terhorst C, Danial S, Croy BA 1997 Absence of natural killer cells during pregnancy is associated with reproductive compromise in TgE26 mice. Biol Reprod 56:169 –179 Croy BA, Ashkar AA, Foster RA, DiSanto JP, Magram J, Carson D, Gendler
50.
51. 52.
53. 54.
55.
56. 57. 58.
59. 60. 61.
Endo • 1999 Vol 140 • No 6
SJ, Grusby MJ, Wagner N, Muller W, Guimond M-J 1997 Histological studies of gene-ablated mice support important functional roles for natural killer cells in the uterus during pregnancy. J Reprod Immunol 35:111–133 Guimond M-J, Wang B, Croy BA 1998 Engraftment of bone marrow from severe combined immunodeficient (SCID) mice reverses the reproductive deficits in natural killer cell-deficient tge26 mice. J Exp Med 187:217–223 Peel S, Stewart I, Bulmer D 1979 Metrial gland cells in deciduomata of pseuodopregnancy. J Anat 129:21–30 Stewart I 1988 Granulated metrial gland cells in the nontraumatized regions of the uterus of ovariectomized mice with deciduomata maintained on progesterone. J Endocrinol 116:11–15 Zheng LM, Joag SV, Parr MB, Parr EL, Young JD-E 1991 Perforin-expressing granulated metrial gland cells in murine deciduoma. J Exp Med 174:1221–1226 Parr EL, Parr MB, Zheng LM, Young JD 1991 Mouse granulated metrial gland cells originate by local activation of uterine natural killer lymphocytes. Biol Reprod 44:834 – 841 Gibori G, Jayatilak PG, Khan I, Rigby B, Puryear T, Nelson S, Herz Z 1987 Decidual luteotropin secretion and action: its role in pregnancy maintenance in the rat. In: Mahesh VB, Dhindsa DS, Anderson E, Kalra SP (eds) Regulation of Ovarian and Testicular Function. Plenum Press, New York, pp 379 –397 Gala RR 1991 Prolactin and growth hormone in the regulation of the immune system. Proc Soc Exp Biol Med 198:513–527 Yu-Lee L-Y 1997 Molecular actions of prolactin in the immune system. Proc Soc Exp Biol Med 215:35–52 Chambers WH, Amoscato AA, Smith MS, Kenniston TW, Herberman RB, Appasamy PM 1995 Prolactin receptor expression by rat NK cells. Nat Immun 14:145–156 Matera L 1996 Endocrine, paracrine, and autocrine actions of prolactin on immune cells. Life Sci 59:599 – 614 Lala PK 1989 Similarities between immunoregulation in pregnancy and malignancy: the role of prostaglandin E2. Am J Reprod Immunol 19:141–149 Linnemeyer PA, Pollack SB 1993 Prostaglandin E2-induced changes in the phenotype, morphology, and lytic activity of IL-2 activated natural killer cells. J Immunol 150:3747–3754
