In animal species, where infertility factors of selected breeders can be discarded, it is well documented that changes induced by exposure of embryos to suboptimal conditions can derive in phenotypic defects [17, 23, 24]. Indeed, in cattle and sheep it has been noted a higher incidence of embryo loss, higher pregnancy length, large body size and birth weight in ART-derived gestations [25]. In addition, the large offspring syndrome described in cattle is an imprinting disorder with similar features to Beckwith-Wiedemann Syndrome in humans, such as macrosomia, macroglossia or visceromegaly [26, 27]. Other anomalies observed in IVP-derived cattle due to an overgrowth in late gestation are a higher incidence of dystocia, hydrallantois, and neonatal mortality [23, 24].
Even though numerous studies in cattle have enabled insights into the impact derived from ART, the information is almost inexistent in porcine species. Thus, it becomes relevant to describe the phenotypical consequences in the offspring and, if possible, to understand the mechanisms involved in potential anomalies derived from ART-derived piglets.
Results from the present study showed that, when the reproductive fluids were introduced in the culture media at the different steps of the in vitro production system, the percentages of cleavage and blastocysts were similar to those obtained in absence of fluids. As the use of reproductive fluids is not yet a common practice in the field, the only work previously published to compare with is that from our own group [11]. In that study, the results showed a 5% higher cleavage rate in the control group than in the group using reproductive fluids, although both values were below the values in the present study (lower than 50% in Cánovas et al. for both groups vs. higher than 65% in the present study for both groups). As in the present study, the percentage of blastocysts in Cánovas et al. study was not different regarding the presence or absence of reproductive fluids. These results confirm what was previously proposed about the low impact that using reproductive fluids has on the final number of embryos obtained. However, at this point, the higher quality of the embryos described by Canóvas et al. cannot be confirmed because no other parameters where analysed in the present study. Instead, we transferred most of the produced embryos to investigate their ability to implant and develop to term.
First data sets that called to our attention regarding the pregnancy rate after transfers was the higher percentage (> 35% in both, C-IVP and RF-IVP groups) of positive pregnancies with the in vitro produced embryos compared to those in our previous study using in vivo produced embryos (27%, Paris-Oller et al., unpublished). Such apparent paradox can be explained because the recommended range of asynchrony between donor and receptors for pig ET must be between 0 and (-) 48 h [18, 28] and this was the rule we followed in the present study, but not in our previous one. Thus, we assume this was the reason for the higher pregnancy and parturition outcomes obtained here. Also the use of a minimally invasive procedure based on a LESS approach might have influenced positively the pregnancy rate because, as demonstrated in previous works [29, 30], the trauma to the uterus was minimal and the full recovery of the animal after procedure significantly reduced compared to laparotomy. Despite this, the percentage of non-pregnant animals after embryo transfer in our study was still higher than 64% in both groups. As it has been recently proposed, this could be associated to a dysregulation of pro and anti-inflammatory cytokine levels in recipient sows that, in turn, induce embryonic mortality [31]. However, many other factors related to either the quality of the embryos or the recipient’s status and age could have been affecting these rates of unsuccessful transfers [32]. On the one hand, and as for the own ET procedure, the reported levels of embryonic mortality by using non-surgical procedures are approximately of 70% [33, 34], while using surgical laparoscopic procedures, Wieczorek et al. [35] reported 50% of successful pregnancies but those were after transferring in vivo produced embryos. This is, on the other hand, other crucial factor to be considered because all the above referred rates derive from embryos produced in vivo and ours were produced in vitro. In fact, for most of the researchers, the transfer of embryos produced in vivo (but not in vitro) is the only one with “possible short-term use in pig production” [32] although, from our results, this statement should probably be reconsidered since our pregnancy rates are comparable to those obtained transferring in vivo collected embryos.
Actually, the high farrowing rates in our experiment, with only one miscarriage happening in the RF-IVP group at day 24 post-transfer (which is regarded as the time frame for implantation), can be considered good indexes of the quality of the embryos transferred, although more studies with higher sample sizes are needed to confirm this statement.
As for the gestation period, it is an index depending on the litter size, but it is well known that some other factors such as farm, parity, number of inseminations or genetic line can affect it [36]. While short gestation lengths are associated with higher stillborn, long pregnancies are not desired by the farmers and the advantages of inducing and attending farrowing compared to letting the sows go on their own are a matter of current debate [36]. In our experience, delivery inductions with oxytocin were necessary in 1, 2 and 4 animals from AI, C-IVP and RF-IVP groups respectively due to delays between the birth of the first and following piglets, but not because the parturition did not start spontaneously. Our presence during delivery, in order to take the individualized umbilical cord and placenta samples, could have acted as an additional stressor contributing to the delays and, consequently, we cannot affirm at this point if such problems were related to the embryo transfer procedure, the embryo source, or our own presence.
The litter size, in our case, was not a factor that affected gestation length in IVP pregnancies because the sow with the longest period (121 days) delivered only 5 piglets while the sow with the highest litter size (10 piglets) delivered at day 115. Similarly, the fact that embryos were in vitro produced was not a factor affecting gestation length because the AI animals showed similar periods of pregnancy length. N
Umbilical and placental abnormalities are relatively common in clones but information about the impact of porcine in vitro-derived embryos on these defects is still limited. In the present study, expression level of key genes related with imprinting, angiogenesis or glucose transport were analysed in placenta and umbilical cord. Fetal weight was found to be proportional to placental weight in several studies [37–39] whereas reduced placental weight has been reported in somatic cell nuclear transfer derived piglets vs those produced by AI [40]. However, while this parameter has been studied in pig with productive purpose [41, 42], in other species such as mice, changes in placental weight have been related to exposure to stressors during in vitro production [43, 44]. Our results show significant differences in the placental weight between the experimental groups (C-IVP and RF-IVP) and AI. Nonetheless, placental area is considered a better marker for postpartum piglets’ viability. Placental area is also highly associated with birth weight [45] and lower placental area was found in piglets dead at weaning vs piglets alive at weaning [42]. In our study, both RF-IVP and C-IVP group displayed larger placental area than AI placentas. However, it should be noted that both IVP groups had higher birth weights (data not shown) while the litter size was smaller than AI. This is in accordance with the negative association reported between birth weight and placental weight with live litter size [42].
Complementary to placental area, placental efficiency (PE) could provide information about placental functionality, with high PE values associated with greater nutrient transport capacity. PE was significantly decreased in the RF-IVP vs AI. Nonetheless, this parameter shows natural variation in pigs and between breeds. Even within the same litter, PE can vary significantly, with piglets having similar birthweight but very different placental weight (up to 25%) [41, 46]. Moreover, the use of PE as a selection tool to increase litter size is debatable, because an increase in litter size could result in reduced birth weight and higher mortalities. It is controversial regarding welfare animal conditions, even in the hypothetic situation that global outcome would remain beneficial.
As for the molecular analyses, it is known that placental nutrient transport capacity is related with gene expression of transporter genes. In mice, Slc2a1 (glucose transporter) and Slc38a2 (amino acid transporter) were upregulated in the lightest placentas, confirming that placentas with high PE adapt and increase nutrient transport efficiency. Contrary, SLC7A1, a cationic amino acid transporter, was found negatively related to PE [46]. Placentas in the RF-IVP group, with the lowest PE, showed double SCL7A1 expression than C-IVP or AI, but these differences were not statistically significant, perhaps by the reduced number of samples. In the umbilical cord, SCL2A1 expression did not show significant differences, but there was a tendency (p = 0.0502) with RF-IVP and C-IVF showing expression values over AI.
LUM and VIM genes were selected as possible markers for placental functionality because it was previously described that children born by IVF showed higher levels of LUM and lower levels of VIM expression in their umbilical vein endothelia cells than children naturally conceived [47]. Altered expression of these proteins would explain, according with [47], the cardiovascular dysfunction and vascular remodelling occurring in IVF offspring. Our data, however, did not show differences in VIM or LUM expression between AI, RF-IVP and C-IVP in umbilical cord samples, suggesting that our IVP system could not be affecting in such way the cardiovascular function of the offspring. However, placentas from AI piglets showed significantly lower LUM expression than those from C-IVP piglets, whereas those from RF-IVP were not different from AI. This result could be suggesting a protective effect of reproductive fluids on the altered expression of LUM, whose consequences in the short-term development of the offspring should be further studied.
As for PEG3, it is well known that placenta from ART-derived animals exhibits higher probability of perturbations in genomic imprinting (mouse, [48]; pig [40]; bovine [26]; human [49]). Under our experimental conditions, PEG3 expression was upregulated in C-IVP embryos vs. AI in placenta but the same did not happened for RF-IVP group, that was similar to AI. Even though PEG3 expression level is sexually biased, with two-fold higher levels in males than females [50], this cannot be the explanation for our results because the proportions of males was 39% in C-IVP piglets. Again, this finding could be of high interest and could confirm the beneficial effect of reproductive fluids on the phenotype of the offspring, but additional studies with higher sample sizes are still necessary to confirm this hypothesis.
Finally, it is worth mentioning that all the animals derived from the present study were kept alive and studies of their growth, glucose metabolism, and haematological and biochemical profiles along their lives are being currently undergone.