Parturition in dogs involves very complex hormonal regulation. However, the underlying prepartal regulatory mechanisms leading to the initiation of canine labor have been poorly understood so far. Canine pregnancy is maintained by progesterone (P4) of exclusively luteal origin, as the placenta lacks steroidogenic activity (Hoffmann et al., 1994; Nishiyama et al., 1999). The onset of parturition is preceded by a dramatic drop in P4 12–24 h before the first clinical signs of labor become visible (Concannon et al., 1977; Concannon et al., 1989; Veronesi et al., 2002). Simultaneously, a rapid increase in circulating prostaglandin F2a (PGF2α), measurable as its metabolite, 13,14-dihydro-15-keto-prostaglandin F2-alpha (PGFM), occurs (Concannon et al., 1977; Concannon et al., 1988; Nohr et al., 1993; Baan et al., 2005; Kowalewski et al., 2010; Gram et al., 2013; Kowalewski, 2014). Thus, different from the slow luteal regression in non-pregnant bitches, which is apparently biologically preprogrammed and associated with luteal aging and degeneration (Zatta et al., 2017), prepartum luteolysis is an actively regulated process in the dog (Kowalewski, 2014; Kowalewski et al., 2015a). The prepartum increase in PGF2α in bitches is suggested to originate from the placental trophoblast cells, as upregulation of prostaglandin-endoperoxide synthase 2 (PTGS2) in the fetal trophoblast was observed in response to altered feto-maternal communication (Kowalewski et al., 2010). Increased PTGS2 indicates the activation of the prostaglandin system. The reduced activity of the specific PGF2α-synthase (PGFS/AKR1C3), concomitant with increased expression of the PGE2-synthase (PTGES), demonstrates that an alternative pathway with PGE2 as substrate is crucial for local prepartum PGF2α synthesis (Kowalewski et al., 2010; Gram et al., 2013; Gram et al., 2014). Additionally, in vitro, it was shown that the endometrium of pregnant bitches produces PGF2α immediately prepartum (Luz et al., 2006). Following luteolysis, the P4-mediated inhibition of myometrial contractions in the pregnant canine uterus ceases, and the uterus becomes sensitive to contractile stimuli (Concannon et al., 1989; Lyse and Challis, 1989; Verstegen-Onclin and Verstegen, 2008). Myometrial contractile activity at parturition is generated by increased levels of contraction-associated proteins, including actin, myosin, and gap junction proteins, especially connexin 43 and the oxytocin receptor, but also PTGS2 and the receptors for PGF2α and PGE2 (PTGFR and PTGER1-4) (Cook et al., 2000; Patel and Challis, 2001; Challis et al., 2002; Olson, 2003; Kowalewski et al., 2010; Gram et al., 2013; Gram et al., 2014). Uterine contractions can then be evoked by contractile stimulators such as oxytocin and prostaglandins (Challis et al., 2002). Nevertheless, detailed knowledge of these processes in dogs is missing. A good understanding of myometrial (patho)physiology is crucial, especially in relation to the most common birth-associated pathology in bitches, uterine inertia (UI). UI is responsible for up to 70% of maternally caused dystocia cases, and in the absence of successful medical treatment options, a cesarean section is often performed immediately after diagnosis (Darvelid and Linde-Forsberg, 1994; Bergström et al., 2006; Prashantkumar and Walikar, 2018). However, in secondary uterine inertia (SUI), exhaustion of the myometrium occurs, usually because of prolonged labor due to a specific cause of dystocia, such as obstruction. The reasons for myometrial contractile failure in primary uterine inertia (PUI) are not fully elucidated (Feldman and Nelson, 2004; Linde Forsberg and Eneroth, 2005). Predisposing factors such as obesity, extremely small or large litter size, increasing age, nutritional and hormonal imbalances, stress, systemic diseases, and hypocalcemia or hypoglycemia of the bitch have been postulated (Johnston, 1986; Darvelid and Linde-Forsberg, 1994; Linde Forsberg and Persson, 2007; Münnich and Küchenmeister, 2009; Davidson, 2011). At the mRNA level, smooth muscle γ-actin, smooth muscle myosin heavy chain 11, and the Rho-associated coiled-coil-containing protein kinases 1 and 2 (involved in calcium sensitization and muscle contractions) are differentially expressed in the uterus between PUI and obstructive dystocia (OD) (Egloff et al., 2020; Frehner et al., 2021). In contrast, the expression of leptin as well as serum-ionized calcium and glucose concentrations did not differ between PUI and OD (Frehner et al., 2018; Frehner et al., 2021). Recently, compared to bitches with a fetal cause for dystocia, downregulated relative gene expression of myosin light chain kinase 4 was found in complete and partial PUI (Magnus et al., 2023). Additionally, in partial PUI, smooth muscle myosin heavy chain 2 was downregulated and protein kinase C expression increased, suggesting the possibility of a defect at the myometrial level (Magnus et al., 2023).

Although the failure of luteolysis was postulated to be causative for UI (van der Weyden et al., 1989; Irons et al., 1997; Bergström et al., 2010; McLean, 2012), it could not be confirmed in a recent study of our group (Jungmann et al., 2022). However, in vivo and in vitro studies showed a contractile response of the myometrium to exogenous PGF2α in humans (Senior et al., 1992; Senior et al., 1993; Patel and Challis, 2001; Janicek et al., 2007; Fischer et al., 2008), cattle (Takagi et al., 2002), and rats (Balki et al., 2010; Balki et al., 2012). The importance of prostaglandins as uterotonic agents has been postulated in dogs (Nohr et al., 1993; Hoffmann et al., 1999), but investigations of PGF2α′s actions in the canine myometrium are scarce. Gogny et al. (2010) assessed the uterokinetic properties of PGF2α on uterine tissue from cyclic bitches in vitro, and Ibuki et al. (1997) implanted force transducers in the uteri of non-pregnant bitches, measuring PGF2α-induced contractions. Regarding parturition, functional studies investigating the effect of PGF2α on canine pregnant myometrium are completely lacking. Similarly, local PGF2α concentrations were determined in canine corpora lutea (Kowalewski et al., 2015b), but the actual PGF2α concentration in the uterus itself at term is unknown. Previous studies investigating the prostaglandin system in relation to UI in dogs failed to identify significant correlations in uterine full-thickness biopsies (Rempel et al., 2021a; Rempel et al., 2021b). However, since oxytocin receptor expression and response to oxytocin stimulation in vitro differ between the two myometrial layers (Jungmann et al., 2022; Jungmann et al., 2023), slightly different functions of the circular and longitudinal myometrial layers can be assumed. To test for the hypothesized uterokinetic effect of PGF2α, we used exogenous PGF2α and investigated its ability to initiate canine myometrial contractions in vitro. Further aims of our study were to determine local PGF2α concentrations in interplacental tissue samples from bitches around parturition and investigate the mRNA and protein expression of PTGS2 and PTGFR in parturient canine uterine tissues in the context of PUI, considering the expression in the different myometrial layers.

Currently, the effect of prostaglandins, especially PGF2α, in parturition and particularly in the initiation and maintenance of uterine contractions in dogs is not fully understood. In non-pregnant bitches with open-cervix pyometra, PGF2α (natural/synthetic) is described as a conservative treatment option to promote the expulsion of uterine contents (Nelson et al., 1982; Meyers-Wallen et al., 1986; Günzel-Apel and Bostedt, 2016), assuming ecbolic/uterotonic activities. Moreover, PGF2α, or a combination of PGF2α and aglepristone, can be used to terminate pregnancies (Wichtel et al., 1990; Hubler et al., 1991; Romagnoli et al., 1991; Hoffmann et al., 1999). PGF2α, due to its action on smooth muscle cells, is known to cause dose-dependent severe side effects such as tachycardia, tachypnea, salivation, vomiting, diarrhea, and occasionally seizures or even collapse and death (Plumb, 2002; Kustritz and Hess, 2007). The aim of our current study was to gain deeper insights into the role of PGF2α on the canine parturient myometrium and assess its ecbolic capacities to activate uterine contractions in vitro. Experiments in the organ bath identified an effect of high dosages of PGF2α (50 µM); however, this effect was observed only on the circular myometrial layer. Our findings from the organ bath experiments contrast with those of Gogny et al. (2010), who described a concentration-dependent (10 nM–10 µM) contractile effect of PGF2α in circular and longitudinal myometrial fibers. Moreover, Gogny et al. identified a significantly higher responsiveness to PGF2α of longitudinal fibers from bitches in anestrus or metestrus receiving aglepristone treatment three to five days before ovariohysterectomy. However, the bitches used by Gogny et al. were non-pregnant and in metestrus or anestrus, and they lacked endogenous PGF2α production by placental trophoblast cells as well as other endocrine and physiological changes that distinguish the late-term and parturient uterus from other reproductive stages. Different responses to PGF2α depending on the cycle stage and/or pregnancy status have been described earlier for other species; for example, and similar to our results, a higher responsiveness to PGF2α was described in the circular myometrial layer compared to the longitudinal layer in rats at term (Tuross et al., 1987). Differences in response between the two myometrial layers are not surprising and were previously shown in the same experimental setup following oxytocin stimulation (Jungmann et al., 2023). During physiological parturition, interaction between both myometrial layers is essential, whereby both layers have slightly different functions and thus seem to have a differential responsiveness to stimulating agents. PGF2α interacts with its specific receptor, PTGFR. As a member of the G protein-coupled receptor family, PTGFR can be desensitized and downregulated by its own ligand (Watanabe et al., 1994; Ferguson et al., 1998; Xu et al., 2013). The reduction in PTGFR availability by increasing PGF2α concentrations around parturition could explain the low responsiveness of the myometrium to PGF2α stimulation observed in the organ bath. In IHC, we found a significantly higher PTGFR expression in the circular layer of the samples compared to the longitudinal layer, independent of the underlying reason for C-section (PUI or OD). This indicates that a higher PTGFR expression in the circular myometrial layer is a general parturition-associated phenomenon. In addition, this might explain, at least to some extent, why in the organ bath the response to PGF2α stimulation was restricted to the circular myometrial layer. Relative overexpression of PTGFR in the circular layer allows for the observed contractile response to at least high concentrations of PGF2α. Nevertheless, a larger number of bitches and a more homogeneous population sampled in the periparturient period, namely, prepartal, parturient, and post-parturient, would have been beneficial and would have allowed for even deeper insights into the role of PGF2α on the canine myometrium during parturition. However, neither age nor P4 concentration affected the investigated organ bath parameters, thus supporting our results in general. Different from this, the variable storage time of samples before processing affected the frequency of contractions. That is, the increase in the frequency of contractions after 50 µM PGF2α application in the circular layer was more pronounced after short-term storage. This seems plausible, as tissue quality cannot remain unaffected by increasing storage time. However, storage was necessary from an organizational point of view (with many C-sections being performed at night and laboratory availability restricted to the daytime). In contrast, in our previous study, no effect of storage time on the response to oxytocin stimulation was determined (Jungmann et al., 2023), and neither were the parameters A, MF, and AUC influenced by storage time in the present study.

An increased relative PTGFR mRNA expression was reported using uteroplacental tissue samples, not only during preimplantation but also during prepartum luteolysis compared to mid-gestation (Kowalewski et al., 2010). During prepartum luteolysis, positive signals for PTGFR mRNA expression were localized by situ hybridization in all trophoblast cells and endometrial glandular chambers (Kowalewski et al., 2010), so increased expression of PTGFR in these compartments is likely. Nevertheless, the situation in the parturient canine myometrium remains unclear. Up until date, the myometrial distribution of PTGFR has never been investigated in detail by distinguishing between myometrial layers. In IHC, we confirmed the protein expression of PTGFR in both myometrial layers with a stronger staining intensity in the circular myometrial layer; thus, quantitative analysis of receptor expression at the mRNA level should be performed. Rempel et al. (2021a) investigated PTGFR expression in PUI and OD but failed to identify significant differences. These authors, however, used full-thickness uterine samples from interplacental and uteroplacental sites, with the latter including placental tissue. In our study, the myometrial layers were carefully separated from adherent connective tissue, endometrium, and placenta. Unfortunately, the determination of PTGFR expression in FFPE-derived mRNA was not successful. The use of FFPE tissue is known to be challenging, as the process of fixation and paraffin embedding often leads to RNA degradation and chemical alterations (Guo et al., 2016). Nonetheless, gene expression of reference genes and PTGS2 was measurable in all samples, indicating that the mRNA integrity was sufficient for more abundantly expressed genes. An additional RT-PCR using cDNA from fresh, frozen uteroplacental uterine tissue and the analysis of PTGFR mRNA expression carried out by Rempel et al. (2021a) prove the existence of PTGFR in the tissue. As quantitative analysis of myometrial PTGFR expression in dogs in the two layers is of great interest in the context of myometrial contractile activity, further studies should use uterine tissue that is immediately dissected into layers and frozen as soon as possible.

PTGS2 mRNA expression (ratio), the rate-limiting factor for prostaglandin synthesis, was also investigated in the canine pregnant uterus (Rempel et al., 2021a). PTGS2 expression did not differ between PUI and OD independent of tissue localization (interplacental versus uteroplacental tissue), but no conclusions could be drawn regarding layer-specific mRNA expression due to the full-thickness biopsies used. With respect to the previous IHC results for PTGS2, where the longitudinal myometrial layer stained significantly stronger than the circular layer, independently of the reason for the C-section (Rempel et al., 2021a), we anticipated differing mRNA expressions between both individual myometrial layers. The observed lack of differences in PTGS2 mRNA expression when comparing layers combined with the results of Rempel et al. (2021a) might indicate regulation on a post-transcriptional level and supports that, although crucial for prostaglandin supply and initiation of parturition in dogs (Kowalewski et al., 2010; Kowalewski et al., 2014), alterations in PTGS2 expression are not likely to be involved in the pathophysiology of canine UI. Interestingly, in human myometrial cells, a PGF2α-mediated positive feedback mechanism for PTGFR-induced PTGS2 expression has been shown (Xu et al., 2013). Thus, PGF2α is able to overcome the rate-limiting effect of PTGS2 in prostaglandin synthesis.

Although the opposing behavior of circulating levels of serum P4 and PGF2α at term is well-described in periparturient bitches (Concannon et al., 1988; Nohr et al., 1993; Hoffmann et al., 1994; Luz et al., 2006; Kowalewski et al., 2010), little is known about the actual PGF2α concentrations in uterine tissue as the target organ during labor. Currently, based on in vitro and gene expression studies, the placenta, namely, the fetal trophoblast, is considered to be the major source of PGF2α (Luz et al., 2006; Kowalewski et al., 2010; Gram et al., 2013), despite evidence that the periparturient canine endometrium is also able to produce significant amounts of PGF2α in vitro (Luz et al., 2006). To the best of our knowledge, our study is the first to investigate PGF2α concentrations in uterine interplacental tissue obtained from bitches at term. As expected, we identified a negative correlation between local uterine PGF2α and peripheral P4 concentrations in serum. PGF2α concentrations were significantly higher in tissue obtained from bitches already in labor (IB, P4 < 2.5 ng/mL, indicating active luteolysis) than in bitches with still elevated P4 serum levels (NIB). Interplacental concentrations identified in the IB group (mean ± SD: 435.24 ± 281.51 pg/mg) were lower than levels obtained from endometrial explants at days 63 and 64 of pregnancy (mean ± SEM: 846 ± 118 pg/mL/mg), although P4 concentrations were not described in the latter study (Luz et al., 2006). Despite the lower concentrations, possibly related to the different tissue composition (full thickness versus endometrium only) and the method, our results clearly confirm the uterus, not only the placenta, as a relevant production site of PGF2α in the immediate prepartum period. Additionally, lower PGF2α concentrations were determined in PUI compared to IB, similar to NIB. Unfortunately, a correlation analysis was not possible, as the respective P4 levels were only available from two of the three PUI bitches and the generally low number of PUI animals. Whether this decrease in PGF2α levels is time-dependent, as bitches diagnosed with PUI are in labor longer than bitches scheduled for an elective C-section, or whether it is related to the occurrence of PUI requires further investigation with a larger number of animals. Additionally, the determination of uterine PGF2α concentrations from bitches within undisturbed labor that already gave birth to at least one puppy would be highly interesting to understand the course of PGF2α concentrations during parturition. Unfortunately, it is not possible to create a control group that would be of great interest in all the experiments conducted in this study because it is not feasible to take samples from clinically healthy periparturient dogs. Even if it is not completely transferable, a comparison of the average measured PGF2α concentration in the IB group with the PGF2α concentrations applied to the tissue in the organ bath revealed that the actual concentration corresponded to the dose of 50 pM. However, in the organ bath, 50 pM had no obvious effect on myometrial contractions. The dose able to achieve at least an effect in the circular layer is thus almost 98,000 times higher than the naturally occurring PGF2α concentrations measured in uterine interplacental tissue during prepartum luteolysis. Assuming that PGF2α promotes uterine contractions during labor, the corresponding dosage is supposed to show an effect in vitro. All results of this study question the direct uterotonic effect of PGF2 on canine myometrium at term. Nevertheless, future studies should investigate local PGF2α concentrations in interplacental and uteroplacental sites, as well as in the placenta, and perform ex vivo organ bath experiments with PGF2α stimulation in uteroplacental tissues as well as in interplacental tissues obtained from bitches with high P4. Studies on mice (Sugimoto et al., 1997) and humans (Xu et al., 2013) showed that PGF2α stimulates prepartal OXTR expression, presenting a trigger for active labor. PGF2α also stimulates the formation of gap junctions in the myometrium, improving cell–cell communication and, thus, enabling coordinated, strong myometrial contractions (Puri and Garfield, 1982; Xu et al., 2013; Xu et al., 2015). In addition, PGF2α is known to enhance the effects of oxytocin and corresponding agonists, presumably by providing Ca²⁺ from a pool inaccessible to oxytocin (Fuchs, 1995). Our data support these assumptions that the contractile effect of PGF2α might be predominantly indirect.

Our study could not prove an overall stimulatory effect of PGF2α on pregnant canine myometrium at term in vitro. High doses of this compound in the organ bath increased all contraction parameters except the F in the circular layer but not in the longitudinal layer, corresponding with higher PTGFR protein expression in the circular layer. However, the requirement for an effective uterotonic substance is likely to activate both layers, possibly ensuring functional labor. Downregulation of PTGFR following increased PGF2α release during prepartum luteolysis might be the reason for the observed low responsiveness of the myometrium to PGF2α treatment in vitro. Concentrations of PGF2α in the uterus were negatively correlated with serum P4 and were the highest during prepartum luteolysis, while contractile myometrial activity was shown to steadily increase until expulsion of the first puppy (van der Weyden et al., 1989). In conclusion, the tested PGF2α concentrations did not directly activate uterine contractions of both myometrial layers of the parturient bitch in an organ bath approach. Consequently, we suggest that PGF2α might affect uterine contractility mainly indirectly through the activation of different contraction-associated mechanisms. Further research is required to identify the underlying mechanisms.