Search

Dopamine-containing neurones of the ventral tegmental area express neurotensin receptors which are involved in regulating cell firing and dopamine release. Although indirect evidence suggests that some neurotensin receptors may be localised on the nerve terminals of dopaminergic neurones in the striatum and thus locally regulate dopamine release, a clear demonstration of such a mechanism is lacking and a number of indirect sites of action are possible. We have taken advantage of a simplified preparation in which cultured rat ventral tegmental area dopaminergic neurones establish nerve terminals that co-release glutamate to determine whether neurotensin can act at presynaptic sites. We recorded glutamate-mediated synaptic currents that were generated by dopaminergic nerve terminals as an index of presynaptic function. The neurotensin receptor agonist NT(8-13) caused an inward current and an enhancement of the firing rate of dopaminergic neurones together with an increase in the frequency of spontaneous glutamate receptor-mediated excitatory postsynaptic currents (EPSCs). Incompatible with a direct excitatory action on nerve terminals, NT(8-13) failed to change the amplitude of individual action potential-evoked EPSCs or the frequency of miniature EPSCs recorded in the presence of tetrodotoxin. However, NT(8-13) reduced the ability of terminal D2 dopamine receptors to inhibit action potential-evoked EPSCs in isolated dopaminergic neurones. Taken together, our results suggest that in addition to its well-known somatodendritic excitatory effect leading to an increase in firing rate, neurotensin also acts on nerve terminals. The main effect of neurotensin on nerve terminals is not to produce a direct excitation, but rather to decrease the effectiveness of D2 receptor-mediated presynaptic inhibition.

Basic transcription element binding (BTEB, also designated BTEB1) protein is a member of the Sp-family of GC-box binding transcription factors that exhibit distinct patterns of expression in many cell types and tissues. A role for BTEB1 in the regulation of cell growth and gene transcription has been invoked, but little is known about the molecular mechanisms underlying these activities. The present study examined the functional consequences of high and low BTEB1 expression in the human endometrial carcinoma cell line Hec-1-A, by deriving stable clonal lines that expressed sense (S) and anti-sense (As) rat BTEB1 constructs. Clonal S lines, with BTEB1 mRNA and protein levels higher than in corresponding parent (N) and As lines, displayed enhanced DNA synthesis upon 3[H]-thymidine incorporation, in serum-containing but not in serum-free medium, and increased cell cycle kinetics, concomitant with the induction in expression of the genes for the cell cycle-associated components cyclin D1, PCNA, cyclin-dependent kinase (Cdk) inhibitor p21, and Cdk2. Compared to N and As lines, S lines also had diminished ability to grow in multi-layers and exhibited increased mRNA levels for plasminogen activator inhibitor-1 (PAI-1), secretory leukocyte protease inhibitor (SLPI), and tissue inhibitor of metalloproteinases (TIMP)-2. In serum-free medium, S, but not N nor As lines, had enhanced DNA synthesis with transforming growth factor (TGF)-beta1, albeit all lines demonstrated similar responses to insulin-like growth factor-I and to epidermal growth factor, respectively. The higher DNA synthesis in S relative to N and As, lines upon exogenous TGF-beta1 addition, was observed in concert with increased expression of cyclins D1 and E and p21, genes. Moreover, S and As lines had increased mRNA levels for TIMP-1, TIMP-2, PAI-1, and beta-catenin, and diminished SLPI, and to a lesser extent, Cdk4 mRNA levels, with TGF-beta1 treatment. These results suggest that BTEB1 may mediate cell growth, in part, by modulating gene expression levels of distinct cell cycle and growth-associated proteins. The correlation between serum- and TGF-beta1 induction of DNA synthesis with increased BTEB1 expression further suggests that BTEB1 may constitute an important downstream regulatory component of various signaling pathways utilized by serum-associated and other growth factors in endometrial epithelial cells.

The uterus during early pregnancy synthesizes a complex array of signaling molecules with specific spatial and temporal modes of expression and which are critical for embryo implantation and subsequent development. The mechanism(s) underlying the differential pattern of synthesis of these pregnancy-associated proteins is not understood very well. The present study evaluated the expression and trans-activation potential of the transcription factor Sp1 in the early pregnancy porcine endometrium to determine its temporal and functional association with the endometrial epithelial-specific genes encoding the transplacental iron-transport protein uteroferrin (UF) and an Sp-family member, basic transcription element-binding (BTEB) protein. Two identical Sp1 clones (717 bp) were isolated from a porcine endometrial cDNA library by polymerase chain reaction (PCR). The nucleotide sequence of these clones encodes a partial protein sequence of 238 amino acids encompassing the Zn-finger region and had significant identities with the corresponding regions in the rat and human proteins. By using a specific antibody raised against human Sp1, porcine endometrial Sp1 was found to exhibit a molecular weight of 110 kDa, was localized predominantly in the nuclei of glandular and luminal epithelial cells, and appeared to exist as a phosphorylated protein. Northern blot analysis demonstrated three distinct size transcripts of approximately 3.5, 5, and 8 kb for endometrial Sp1. The expression of Sp1 mRNA and protein, determined by RT-PCR and by its ability to bind Sp1 consensus motif in gel mobility shift assays, respectively, overlapped with, but did not parallel that of UF mRNA during early pregnancy. The effect of increased Sp1 expression on UF gene promoter activity was examined using a human Sp1 expression vector that was transiently transfected into primary cultures of pig endometrial glandular epithelial cells. Sp1 increased (P < 0.05) the promoter activities of various UF promoter-Luciferase reporter constructs by 2 to 4-fold, over those transfected with empty expression vector. Co-transfection of a BTEB expression vector with the Sp1 expression vector modified the effect of Sp1 on UF promoter activity in the shortest construct. These results suggest that Sp1 mediates the regulation of endometrial epithelial gene expression during pregnancy, and that this function is likely altered in vivo by co-expression of other family members, including BTEB.

Protease inhibitors are major secretory components of the mammalian uterus that are thought to mediate pregnancy-associated events primarily by regulating the activity of proteolytic enzymes. In the present study, we examined the mitogenic potentials of two serine protease inhibitors, namely secretory leukocyte protease inhibitor (SLPI) and uterine plasmin/trypsin inhibitor (UPTI) in primary cultures of glandular epithelial (GE) cells isolated from early pregnant (Day 12) pig endometrium, using the [(3)H]thymidine incorporation assay. Purified porcine SLPI (pSPLI), porcine UPTI (pUPTI), or recombinant human SLPI (rhSLPI), all of which exhibited anti-trypsin activity, increased (p < 0.05) labeled thymidine incorporation into DNA of serum-deprived GE cells when tested at a range of 10-1000-ng/ml concentrations. Polyclonal antibodies directed against either hSLPI or pSLPI abrogated the effect of SLPI. Co-addition of pSLPI and pUPTI increased DNA synthesis in these cells to a level higher (p < 0.05) than that observed with either protease inhibitor. The glycosaminoglycan heparin, which has been previously shown to increase the anti-protease activity of SLPI, exhibited a tendency (p = 0.08) to enhance SLPI and UPTI induction of cellular DNA synthesis. Reverse transcription-polymerase chain reaction indicated that the messenger RNAs for both protease inhibitors were present in the endometrium throughout pregnancy and, within this tissue, in GE cells to a greater extent (p < 0.05) than in stromal fibroblastic cells. Results demonstrate that, in addition to their well-documented anti-protease activities, SLPI and UPTI may constitute autocrine growth promotants for the uterine epithelium. These data suggest a novel mechanism whereby locally produced protease inhibitors may modulate periimplantation events and embryo-maternal communication.

The present study examined the trans-activation potential of basic transcription element-binding protein (BTEB), a recently identified member of the Sp family of GC box-binding transcription factors, on the expression of the gene encoding the pregnancy-associated, epithelial-specific, and progesterone (P)-induced porcine uterine endometrial secretory protein, uteroferrin (UF). Endometrial expression of BTEB, P receptor (PR), and UF genes was analyzed by RT-PCR as a function of pregnancy stage and cell type and was correlated with the levels of endometrial BTEB that were quantified by Western blot and/or electrophoretic mobility shift assay. PR, BTEB, and UF messenger RNAs (mRNAs) were present in early (day 12) and mid(day 60) pregnancy pig endometrium, although expression levels varied for each mRNA (UF, day 12 << day 60; PR and BTEB, day 12 = day 60). Within the endometrium, glandular epithelial (GE) cells manifested higher amounts of UF mRNA than stromal fibroblastic cells, whereas both cell types had comparable amounts of BTEB and PR mRNAs. Expression of BTEB, however, was limited to endometrial GE cells. A BTEB expression vector (pcDNA-3BTEB) was used to examine the effect of increased BTEB protein on UF gene expression and promoter activity in primary cultures of pig endometrial GE cells. Cells transiently transfected with pcDNA-3BTEB had 2-fold higher UF mRNA levels than those transfected with the empty expression vector (pcDNA-3). Further, cells cotransfected with a UF promoter-luciferase (-1935UF-Luc) reporter gene and the BTEB expression vector had 2-fold higher Luc activity than those cotransfected with reporter gene and pcDNA-3. This effect of BTEB was not observed in transfected endometrial stromal fibroblastic cells, but was apparent in the human endometrial epithelial carcinoma cell lines ECC-1 and Hec-1-A, which exhibit low levels of BTEB protein and low or undetectable PR mRNA levels, respectively. The respective contributions of BTEB and PR to the modulation of UF promoter activity were examined by cotransfection of Hec-1-A and ECC-1 cells with expression plasmids for BTEB and PR and one of two UF promoter constructs (-831UF-Luc or -1935UF-Luc) in the absence or presence of P. The increase in UF promoter activity with BTEB was mimicked by PR in a P-dependent manner in both cell lines. The combined effect of PR/P and BTEB appeared additive in Hec-1-A cells and was synergistic in ECC-1 cells. These results highlight the cell context dependence of the trans-activation potential of BTEB and suggest its unique role, in concert with PR, in directing the temporal expression of endometrial epithelial genes of pregnancy.

Basic transcription element binding (BTEB) protein is a newly identified member of the C2H2 zinc finger family that also includes the transcription factors, Sp1, Sp2, Sp3, and Sp4. This family of proteins binds GC-rich motifs widely distributed in gene promoters, resulting in distinct activation or repression of transcriptional activities. Whereas Sp proteins are ubiquitously expressed, expression of BTEB appears more limited and has not been documented in the female reproductive tract of any mammalian species. This study was designed to identify and characterize the cellular distribution of BTEB in the porcine endometrium and placenta at known stages of pregnancy. Northern analysis of uterine endometrium detected BTEB mRNA that corresponds in size (5 kilobases) to that of the major BTEB transcript in rat brain. The steady-state levels of BTEB mRNA were higher (p < 0.05) in endometrium than placenta at corresponding days of pregnancy, although for each tissue, the levels did not change with pregnancy stage (p > 0.05). Luminal epithelial (LE), glandular epithelial (GE), and stromal (ST) cells isolated from pregnancy endometrium expressed the BTEB gene, but mRNA abundance varied with cell type (LE, GE > ST). Western blot analysis using an antiserum generated against the N-terminal region of a porcine BTEB fusion protein produced in Escherichia coli revealed the presence of BTEB protein only in endometrium, not in placenta. Immunohistochemical studies localized BTEB predominantly to the nuclei of endometrial GE and LE cells. Consistent with the presence of functional BTEB protein, binding to a double-stranded oligonucleotide containing multiple GC motifs was demonstrated in nuclear extracts prepared from endometrium and from endometrial LE and GE, but not ST, cells by electrophoretic mobility shift assay. These results demonstrate the preferential endometrial cell-type expression of BTEB and suggest its regulatory role in pregnancy-associated endometrial epithelial gene expression.

Cathepsin L is a major lysosomal cysteine protease produced by mouse placenta and fibroblasts. This study characterizes a novel cathepsin L-related mRNA expressed in rat placenta. Immunological and nucleotide screening of a rat placenta library identified six positive clones, the largest, pCLRP-9, being 924 base pairs in length. The combined sequences of all the clones contain an open reading frame of 711 nucleotides, a termination codon, a polyadenylation site, and 197 nucleotides of 3' untranslated region, but lack the 5' translation initiation codon. The pCLRP nucleotide sequence showed 60-64% identity to those of mouse, rat, and human cathepsin L. The deduced amino acid sequence of pCLRP codes for 237 amino acids, which align with the carboxy-terminal sequence of cathepsin L and has the active site residues characteristic of the cysteine protease family. Northern blot analysis showed hybridization of pCLRP with a major mRNA transcript of 1.3 kilobases expressed in placenta, but not kidney or liver. In contrast, a cDNA for mouse pro-cathepsin L hybridized with a transcript of 1.7 kilobases expressed in rat kidney, as well as placenta. During late gestation, steady-state levels of rat placental pCLRP mRNA were highest on day 18, whereas those of mouse procathepsin L were greatest on day 20 of gestation. Antiserum to mouse cathepsin L cross-reacted with four proteins of molecular weights 36,000 to 42,000 in rat placental culture medium, of which two were absent in the kidney. These data indicate that rat placenta expresses several species of cathepsin L-type proteins, which may be involved in placental function and nutrient supply.

Uterine expression of the mRNA encoding antileukoproteinase (ALP) is highest in pig uterus during mid- to late pregnancy, suggesting a stage of pregnancy-dependent role for this elastase/cathepsin G protease inhibitor in feto-maternal interactions. To examine a potential relationship between uterine synthesis of ALP and the type of placentation in mammalian species, the expression of ALP mRNA and/or protein in pregnant mares, cows, rats, and mice was evaluated. Genomic DNA and mRNA hybridization analyses were performed using a porcine ALP cDNA as probe. The concentration of ALP protein in reproductive tissues was determined by RIA using a polyclonal antibody raised against a synthetic peptide (ALP 16P) corresponding to amino acid residues 21-36 of the porcine ALP protein. A single ALP mRNA transcript of approximately 0.8 kb in length was detected in equine and bovine uterine tissues. The relative abundance of ALP mRNA in equine endometrium increased between days 125-170 (mid-pregnancy), and then decreased by day 215 of pregnancy. Similarly, the steady state levels of ALP mRNA in bovine endometrium and myometrium were higher during mid- to late than during early pregnancy. The levels of ALP mRNA in bovine fetal cotyledon were low and did not change significantly with stage of pregnancy. No hybridization was detected to pregnant rat endometrial tissues, although high stringency Southern blot analysis of porcine, bovine, and rat genomic DNAs using porcine ALP cDNA as probe predicted a high degree of nucleotide sequence homology in their respective ALP genes.(ABSTRACT TRUNCATED AT 250 WORDS)

DNA-protein interactions within two putative regulatory regions distal from the transcription initiation site of the porcine uteroferrin (UF) gene were characterized. These regions, termed XB (-1,600 to -1,129 bp) and AB (-1,128 to -893 bp) exhibited transcriptional enhancer activities within the context of the heterologous SV40 promoter, that were specific to endometrial cells. DNase I and gel-shift assays demonstrated that both fragments contain a heptamer motif TGCTAGA that binds a nuclear protein present in crude and DEAE-fractionated nuclear extracts from porcine endometrium of pregnancy. This heptad sequence, designated as endometrial-associated sequence (EAS), is different from previously described nuclear protein-binding consensus sequences. Mutations in the heptamer motif abolished binding to the nuclear factor, as detected by gel-shift assays. The endometrial nuclear protein that interacts with the heptamer was characterized by Southwestern and UV cross-linking analysis. The protein has an approximate M(r) of 80 kD, is basic (pI 7.7-8.6) and is present in pig endometrium throughout pregnancy. The functional relevance of this DNA-binding protein in the control of UF gene transcription in the endometrium is discussed.

In the pig, iron transport to the developing fetus during pregnancy involves, in part, uteroferrin (UF), a secreted progesterone-induced protein of the uterus. Neonatal pigs suffer from anemia, and the decrease in the synthesis of UF protein in late pregnancy was suggested to be partly responsible for this condition. To examine whether diminished capacity for UF uptake by pig fetuses may also contribute to neonatal anemia, binding sites for 125I-UF were examined in plasma membrane-enriched fractions of fetal liver and spleen, which are sites of fetal hematopoiesis. In addition, changes in the number of these binding sites as a function of fetal development were evaluated. Binding of 125I-UF to liver membrane fractions was displaced by intact UF greater than deglycosylated (aglyco) UF greater than ovalbumin, but not by yeast mannan. Scatchard analysis of radioligand binding showed the presence of a single class of binding sites with a dissociation constant of 10(-7) M. During fetal development and at postpartum (day 5), liver binding sites for UF remained invariant and displayed the same affinity. In contrast, the number of binding sites for UF in fetal spleen increased from midpregnancy to parturition and remained elevated in day 5 neonatal spleen. Affinity cross-linking of 125I-UF to liver membrane-associated binding sites and subsequent analysis by gel electrophoresis and autoradiography demonstrated a single labeled protein complex of Mr 58,000 and 87,000 under denaturing and nondenaturing conditions, respectively. The appearance of these bands was inhibited by intact UF, but not ovalbumin. The characteristics of the membrane-associated binding sites for UF differed from those of the mannose-related receptor previously described in reticuloendothelial cells of fetal liver. The invariant presence of UF binding components in sites of hematopoiesis during fetal development suggests that mechanism(s) unrelated to specific uptake of UF are responsible for neonatal anemia.

Insulin-like growth factor-I (IGF-I), synthesized by the uterine endometrium of cyclic and early pregnant gilts, accumulates in the uterine luminal fluid, where it comes in contact with the developing conceptus and the rapidly growing uterus. The uterus and the conceptus thus represent potential target sites for the biological effects of IGF-I, provided high-affinity Type I receptors are present. This study was undertaken to evaluate the expression of functional IGF-I receptors in the endometrium and myometrium of pregnant (Day 10, 12, and 15) gilts and in the endometrium of cyclic (Day 15) and pseudopregnant (Day 15) gilts and to correlate levels of these receptors with temporally regulated uterine production of IGF-I. Specific binding of 125I-IGF-I to endometrial membranes pretreated with MgCl2 (4 M) at 4 degrees C for 16 h, was saturable and membrane concentration-dependent. Competition of 125I-IGF-I binding to endometrial membranes was highest with unlabeled IGF-I greater than IGF-II much greater than insulin, whereas porcine relaxin was noncompetitive. Affinity cross-linking of endometrial membranes with 125I-IGF-I followed by SDS-PAGE and autoradiography revealed two labeled bands of Mr greater than 200,000 and Mr 135,000, with the major band being the Mr 135,000 species. Scatchard analysis of 125I-IGF-I binding to endometrial membranes from Day 12 pregnant gilts revealed a single class of binding sites with a dissociation constant (Kd) = 4.08 +/- 0.09 nM. Membranes prepared from endometrium of Day 10, 12, and 15 pregnant gilts exhibited comparable 125I-IGF-I binding (p greater than 0.05) that was higher (p less than 0.001) than that for the corresponding myometrial membranes.(ABSTRACT TRUNCATED AT 250 WORDS)

Expression of the gene for the porcine transplacental iron transport protein uteroferrin (UF) is largely restricted to the uterus, where it is differentially regulated by estrogen (E) and progesterone (P). To study the regulatory mechanisms subserving these effects, a 2-kilobase genomic fragment corresponding to -2005 to 48 nucleotides of the UF gene was ligated up-stream to the reporter gene chloramphenicol acetyltransferase (CAT). This construct (UF-CAT) was transiently transfected into rabbit endometrial (HRE-H9), mouse fibroblastic (AKR-2B), and human choriocarcinoma (JEG-3) cells. The basal gene promoter activity of UF-CAT was exhibited in H9 cells, but not in AKR-2B or JEG-3 cells. In contrast, a simian virus-40 early promoter (SV2) was functional in all three cell lines. The H9 cells were used to examine steroid regulation of the UF gene promoter. The CAT expression in H9 cells primed with E and PRL, but not with E or PRL alone, was stimulated by P. In contrast, basal activity of SV2 in these cells was unaffected by hormones, singly or in combination. To examine the basis for the E/PRL-dependent response to P, levels of P and E receptors in H9 cells were quantified. PRL and E plus PRL increased the number of high affinity sites for P, but had little effect on levels of high affinity sites for E in treated vs. untreated H9 cells. In vivo administration of PRL to cyclic gilts had no effect on levels of endometrial UF mRNA and secreted UF protein; however, E- plus PRL-treated gilts had higher (P less than 0.05) levels of endometrial UF mRNA and luminal UF than PRL-treated gilts. These results demonstrate in vitro functional activity of the UF gene promoter and associated 5' flanking region and suggest that sequences within this region may mediate tissue-specific and steroid hormone-regulated expression of the UF gene. Moreover, interactions among E, PRL, and P modulate UF gene expression in vivo and in vitro.

Placental lactogen (PL) was isolated from goat cotyledonary tissue by a combination of mild alkaline extraction, anion and cation exchange chromatography, chromatofocussing and molecular filtration. The product, enriched 15,000-fold from the initial extract, was homogeneous when examined by SDS-gel electrophoresis (Mr 22,500) and isoelectricfocussing indicated a pI of 8.35 with a trace contaminant of pI 8.0. When assessed by relative binding activity in radioreceptor assays (RRA), goat PL exhibited somatotropic activity equivalent to 2.2 units/mg dry weight and lactogenic activity equivalent to 28.5 units/mg. A radioimmunoassay (RIA) for goat PL is described that is highly sensitive (190 pg/tube) and has acceptable repeatability within and between assays (6 and 13%, respectively). The assay is not affected by goat pituitary extracts or partly purified goat growth hormone and prolactin. Despite the marked increase in sensitivity of the RIA over that previously available when goat PL was measured by RRA, the hormone was not detected in jugular plasma of goats before Day 44 of pregnancy; concentrations increased thereafter and highest levels were measured during the last third of pregnancy in animals bearing triplets. Measurements by RIA are in general agreement with those obtained earlier in several studies in which RRAs were used. The hormone was detected in amniotic fluid. Maternal concentrations of goat PL declined before parturition and were undetectable by 18 h post partum.

Jugular plasma concentrations of oestradiol-17 beta, prolactin, progesterone and 13,14-dihydro-15-keto-prostaglandin F-2 alpha (PGFM) were measured at 2-h intervals during the last 4 days of pregnancy in 6 goats. During advanced labour and delivery, samples were obtained more frequently and assayed for oxytocin. The animals were housed in a barn with continuous dim lighting. A distinct pattern of oscillation in prolactin concentrations, with peaks during the late afternoon, was apparent during the last 3 days. Geometric means of peak concentrations doubled each day and became of longer duration; night-time nadir values remained low except during the last night before parturition. A progressive increase in oestradiol-17 beta, with mean levels doubling every 36 h, was apparent during the last 3 days. There was no sharp pre-partum increase in oestradiol-17 beta. Correlated (r = 0.83) with the increase in oestradiol-17 beta was a gradual increase in PGFM and when the latter reached approximately 1000 pg/ml, the non-reversible decline in progesterone reflecting pre-partum luteolysis occurred. Subsequent changes in PGFM related closely to an approximately 20-fold increase in the ratio of oestradiol-17 beta to progesterone until maximal PGFM levels of 26.5 +/- 4.2 ng/ml were reached at delivery. Basal concentrations of oxytocin (8-15 microU/ml) were measured before the last 60 min and markedly higher, though erratic, concentrations were detected at various times before appearance of the allantochorion. Maximal oxytocin values (range 180-1570 microU/ml) occurred within minutes before or after delivery of the first fetus. The results suggest that increased pre-partum production of oestradiol-17 beta, in addition to provoking sufficient release of prostaglandins to cause luteolysis, may modulate either the sensitivity or set-points for an endogenous rhythm in prolactin secretion at the end of pregnancy. The nature of the oxytocin changes suggest that, after labour has evolved sufficiently, delivery is precipitated by an abrupt increase in oxytocin secretion.

Concentrations of progesterone and of 13,14-dihydro-15-keto-prostaglandin F-2 alpha (PGFM) were measured in plasma collected from 6 bitches every 3 h starting 2.8-4.6 days before parturition (birth of first pup) and continuing until 0.4-0.8 days post partum, and in additional samples collected less frequently. Progesterone concentrations at 48, 24, 12 and 3 h pre partum averaged 2.8 +/- 0.3, 2.2 +/- 0.4, 1.0 +/- 0.3 and 0.7 +/- 0.2 ng/ml. At those times PGFM values averaged 380 +/- 80, 800 +/- 220, 1450 +/- 450 and 1930 +/- 580 pg/ml, respectively. Mean concentrations of PGFM increased about 2.5-fold between 48 and 15 h pre partum in association with the onset of luteolysis, and then increased another 2.5 times before parturition as progesterone fell to nadir values. Peak levels of PGFM ranged from 1060 to 7150 pg/ml (2100 +/- 600 pg/ml) and occurred within 1-9 h after the birth of the first pup and before the birth of the last pup. These results suggest that prepartum luteolysis in dogs is initiated by increases in maternal concentrations of PGF, and that progesterone withdrawal causes a further increase in PGF which completes luteolysis and provides a major portion of the uterotonic activity causing expulsion of pups.