Background
In the domestic dog (Canis familiaris), the establishment and maintenance of pregnancy depend on the luteal provision of progesterone (P4) as the only source of this hormone both during gestation and in the non-pregnant cycle [
1,
2]. Luteal function is similar in both pregnant and non-pregnant bitches (see reviews in [
3‐
5]). The circulating P4 levels start to deviate strongly only shortly prior to term on days 62/63, when a steep decline marks onset of prepartum luteolysis 12-24 h before the first clinical signs of labour become visible [
6,
7]. In non-pregnant animals, in contrast, luteal regression continues beyond this point of time and circulating P4 concentrations decrease slowly until levels below 1 ng/ml are reached about 10 to 20 days later and the cycle enters the phase of obligatory sexual quiescence, i.e., anoestrus [
6]. Consequently, in contrast to what is observed in non-pregnant dogs, prepartum luteolysis appears to be a process governed by active regulatory mechanisms. In respect to luteolytic factors, normal cyclic ovarian steroidogenic function has been observed in hysterectomized bitches, as in cats and primates, rendering luteal function as independent of a uterine luteolysin (prostaglandin (PG) F2α; PGF2α) in these species [
8,
9]. This is in contrast to livestock, where ovarian cyclicity is maintained due to periodic production and secretion of uterine PGF2α.
In dogs, prostaglandins synthesized locally within the corpus luteum appear to be involved in luteal maintenance rather than in luteolysis/regression, as indicated by increased expression of cyclooxygenase 2 (COX2, PTGS2) and PGE2-synthase (PGES) at the beginning of the canine luteal phase [
10‐
12]. This hypothesis has been substantiated recently by showing a luteotrophic role of PGE2 in canine lutein cells
in vitro. Acting mostly at the level of substrate provision facilitated by steroidogenic acute regulatory (STAR) protein expression and function, and not at the level of activity of the two major steroidogenic enzymes, 3βHSD and P450scc enzyme, PGE2 significantly increased progesterone output in lutein cells [
13].
In pregnant animals the prepartum luteolysis is concomitant with a rapid increase of PGF2α in maternal peripheral blood, implying its involvement during prepartum luteolysis and/or fetal expulsion [
14‐
16]. With regard to the endocrine mechanisms regulating the process of parturition in the dog, most of the currently available knowledge relates to the endocrine capabilities of the utero-placental compartments, comprising the placenta with its adjacent uterine tissues [
15,
17‐
20]. Recent investigations into the expression of factors involved in prostaglandin synthesis and function, the so-called PG-system, have indicated an upregulated expression of COX2 in fetal trophoblast cells as the possible origin of prepartum PGF2α release within the utero-placental compartment [
15]. This finding corroborates the observation from the same study that interfering with P4 receptor (PGR) function by application of an antigestagen in mid-pregnant dogs leads to upregulation of utero-placental COX2 expression, mostly in the fetal trophoblast, and results in significantly elevated peripheral PGF2α concentrations. The only known canine-specific PGF2α-synthase (PGFS), classified as aldo-keto reductase family 1, member C3 (AKR1C3) which is responsible for the direct conversion of PGH2 to PGF2α, was localized in epithelial cells of the superficial uterine glands, the so-called glandular chambers, and in fetal trophoblasts [
17]. Interestingly, however, expression of this enzyme was decreased during prepartum luteolysis. This was paralleled by the reduced expression of 15-hydroxyprostaglandin dehydrogenase (HPGD), an enzyme catabolising PGF2α and PGE2 to their inactive metabolites [
17]. Together with the concomitantly increased expression of
PGE2-synthase (PGES) [
15], a possible functional interrelationship between these entities has been suggested. The potential involvement of alternative pathways for PGF2α synthesis has not, however, yet been investigated. It is noteworthy that the results to date clearly show that there is no parturition-related increase in estrogens mediating the prepartum PGF2α release, and the sporadically observed elevated levels of cortisol measured in maternal blood peripartum seem not to be mandatory for normal canine parturition [
1,
21].
Less attention has been paid so far to uterine endocrine capabilities in the dog during pregnancy and parturition, especially those uterine sites not attached to the placenta. In one study, basal and 12,13-phorbol dibutyrate (PDBu; protein kinase C stimulator) - mediated capabilities of canine endometrial and placental explants to produce PGF2α were determined
in vitro, revealing that the endometrium of pregnant dogs seems to acquire an increased capacity to produce PGF2α immediately prepartum, which - on a tissue weight basis of the explants used - exceeded even that of the placenta [
20]. However, when taking into account the larger mass of placental tissue
in vivo, the authors inferred that also the placenta might substantially contribute to the prepartum PGF2α release. The latter conclusion was confirmed in own studies [
15] which showed an upregulation of
COX2 on the mRNA and protein level in the trophoblast prior to parturition, rather indicating that the placenta and not the endometrium is the main source of the prepartum PGF2α release. Thus, the question regarding the involvement of prostaglandins synthesized in the interplacental sites in regulating the processes of pregnancy and parturition in the bitch is still open. Such a functional interrelationship has previously been established for other species, e.g., cattle, where increased uterine, intercaruncular expression of COX2 and PGF2α-receptor (PTGFR, FP) was observed in animals with induced parturition, indicating the likely contribution of this increase to labour [
22].
Therefore, with the aim of improving our understanding of uterine endocrine function during canine gestation, here, the expression and localization of factors involved in prostaglandin synthesis and its biological activity, as well as the expression of progesterone receptor (PGR), were investigated in the interplacental uterine sites. Based on the tissue material available, gestational periods from the early pre-implantation stage until fully established pregnancy (mid-gestation) were included. Additionally, in order to investigate the possible functional pathways and aiming to gain further information on the underlying endocrine mechanisms, expression of the PG-system was assessed in mid-pregnant bitches in which PGR function was blocked with an antigestagen in order to induce preterm luteolysis/abortion.
Discussion
By showing strong functional interrelationships, and to some extent opposite effects, prostaglandins (PGs) unequivocally play an important role in regulating reproductive functions. Thus, whereas PGE2 was shown to be involved, e.g., in softening the cervix during parturition [
27,
28], PGF2α is the best-known luteolytic factor, while both PGs are involved in coordinating myometrial contractile activity [
29‐
31]. Accordingly, uterine expression of the respective genes encoding for factors involved in PGs synthesis and metabolism, the so-called prostaglandin system, was shown in several species, such as cattle [
22], sheep [
32,
33] and rats [
34].
In the present study, the expression and localization of the PG-system was investigated in canine interplacental uterine tissues at selected time points of pregnancy, i.e., pre-implantation, post-implantation and mid-gestation (there was no tissue material available for experiments from the interplacental sites from dogs during normal perpartum luteolysis).
Our data clearly indicate the basic capability of these tissues to synthesize and respond to prostaglandins, because the expression of all major components of the PG-system was detectable in all tissue samples investigated, both at the mRNA and protein level.
In general, the gene expression patterns observed resembled their expression in the corresponding utero-placental compartments [
15,
17]. The expression of
COX2 remained low until mid gestation, and the post-implantation period was characterized by increased
PGFS/AKR1C3 expression. This, together with the gradually rising
PGES expression levels and the concomitant presence of the respective prostaglandin receptors, imply a local role of these hormones during canine decidualization and implantation, as indicated previously [
15,
17]. Also, in agreement with our previous conclusion [
17], a locally increased
HPGD expression during earlier stages of pregnancy, which - in the present study - shows up as a tendency, could additionally restrict and coordinate the endocrine/paracrine effects of PGs.
The antigestagen-mediated blocking of the uterine PGR receptor, whose expression was significantly downregulated after implantation, resulted in upregulated uterine
COX2 and
PGES expression, thereby resembling their expression patterns in utero-placental compartments during normal and preterm parturition/abortion [
15]. Interestingly, however, their localization patterns did not change significantly in response to the antigestagen treatment and, especially for COX2, the strongest signals were still observed in the myometrium. This was compared with their distribution pattern in the corresponding utero-placental compartments, confirming their presence in the fetal part of the placenta, i.e., in the trophoblast cells, as described in our previous research [
15,
17]. Less dramatic changes, but also with an overall expression pattern resembling that observed in utero-placental samples, pertain to the uterine expression of
PGFS/AKR1C3,
FP,
EP4 and
HPGD in antigestagen-treated bitches.
Consequently, based on our observations, we infer that the upregulated uterine expression of COX2 and PGES may originate mostly in the myometrial compartment, because their localization pattern did not differ significantly following treatment with Aglepristone®. This is in contrast to observations made in cattle where the strongly upregulated COX2 expression was predominantly localized in surface epithelial cells, both in animals undergoing spontaneous labour and those in which it was induced prematurely [
22,
35]. However, as in cattle, in dogs PGs derived from the interplacental uterine sites seem to contribute primarily to myometrial contractility in conjunction with fetal expulsion, with the utero-placental compartment, however, remaining an important source of PGs around term [
36,
37].
The alterations in placental feto-maternal communication occurring in response to the local withdrawal of progesterone appear to be important for subsequently increasing output of the luteolytic hormone PGF2α, which with respect to the placenta must originate in the fetal component, i.e., in the trophoblast cells, where PGs synthesizing enzymes are expressed and localized. This conclusion seems to relate to both the bovine and canine species [
15,
36,
37].
Furthermore, also in accordance with our previous studies [
15], the upregulated expression of COX2 during induced parturition points towards the substantial role of COX2 (PTGS2) as a rate-limiting factor in provision of prepartum prostaglandins in the dog. Such a role of COX2 as a rate-limiting factor has also been observed in the horse [
38], where blocking of endometrial COX2 expression by the conceptus at day 15 of early pregnancy prevented PGF2α-induced luteal regression and resulted in continuation of pregnancy.
At the functional level, our descriptive findings concerning the basal capability of canine uterine and placental tissues to produce and respond to PGs, are not conclusive and will require further studies including, e.g., measurements of the local PGs content within those tissues during different stages of canine gestation.
Conclusions
Our study presents the expression and localization patterns of factors involved in the synthesis of prostaglandins, the so-called prostaglandin-system, in canine interplacental uterine sites at selected times of pregnancy (pre-implantation, post-implantation and mid-gestation). Additionally, changes in the expression and availability of these factors and, hence, their potential contribution to the process of parturition, have been investigated in dogs in which progesterone receptor function was blocked by an antigestagen.
Based on the observations presented herein, in addition to prostaglandins of placental origin, the canine pregnant uterus also appears to be an organ actively involved in prostaglandin synthesis during canine parturition, most probably contributing predominantly to myometrial contractility. From a practical point of view, our findings can help us to better understand the mechanisms responsible for parturition in the dog, especially because decreased synthesis of prostaglandins and/or reduced uterine sensitivity to them could result in serious clinical conditions such as uterine inertia.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MPK: Idea of the study, collecting tissue samples, Real Time PCR experiments, statistical evaluation, interpretation of data, manuscript writing. EK: RNA isolation, Real Time PCR experiments, immunohistochemical procedures. EH: Tissue preparation and processing, immunohistochemical procedures. BH: coordination of animal experiments and collection of tissue samples, knowledge transfer, critical discussion of the data, editing of the manuscript. AB: knowledge transfer, critical discussion of the data, editing of the manuscript. All authors read and approved the final manuscript.