**3. Discussion**

Summarizing our findings, we demonstrate a su fficient induction of intrauterine growth restriction in both rodent models, as determined by fetal weight reduction, when compared to the respective controls. Rarres1 and 2 were pre-dominantly located in the spongiotrophoblast and giant cells of both rat and mouse placenta. In the rat, no consistent regulation of Rarres1/2 was detected under maternal protein restriction. Similarly, we did not find changes in *Rarres1* and *2* mRNA or protein expressions in eNOS−/− mice. In the mouse, we observed a temporal increase in placental *Rarres1* mRNA expression from E15 to E18.5 independent of their genotype. Moreover, we found an IL-11 induction at E15 in the eNOS−/− mice, suggestive of increased placental inflammation, a common feature of human PE [24].

#### *3.1. Expression of Rarres1 in Rodent Placenta*

Based on its localization in the human placenta, we have previously hypothesized that RARRES1, as a tumor-suppressor gene, might slow down invasion and migration of EVTs in terminal placentas and that it might regulate proliferation, syncytialization, and apoptosis of villous trophoblasts [1,2]. Thus, an increased RARRES1 expression in human PE might represent a state of reduced trophoblast proliferation and syncytialization with increased apoptosis [1,38,39].

In our current study, we determined that rodent Rarres1 was predominantly located in the placental junctional zone (giant cells and the spongiotrophoblast) with only minor expression in the labyrinth layer (resembling the human syncytiotrophoblast). Based on this observation, it could be assumed that Rarres1 plays a minor role in rodent placental syncytial physiology at the examined gestational stages. Other than a common hemochorial nature, murine placental anatomy shares limited features with the human placenta [40,41]. Nevertheless, the rodent placenta junctional zone (JZ) shares similarities with the human extravillous compartment [42–45], as it is positioned between the labyrinth and the maternal decidua [46]. From ~E12.5 onwards, trophoblast cells of the JZ invade into the decidua, where they become associated with maternal blood spaces. As a counterpart to the human placental syncytium, the JZ also constitutes the main placental endocrine compartment a ffecting both maternal and fetal physiology [46]. We have previously demonstrated that the placental distribution pattern of RARRES1 changes throughout human gestation [1]. Since we were able to detect temporal changes in the gestational expression of Rarres1 in mice, an involvement of Rarres1 in placental development and growth seems feasible, potentially regulating invasiveness of JZ trophoblast.

In contrast to our previous results from human PE placentas, *Rarres1* expression was not induced in placentas of eNOS−/− mice. In fact, late gestational expression of placental *Rarres1* was rather reduced by dietary-induced IUGR in the rat. However, our *Rarres1* mRNA data were not supported by Western blot analysis, potentially owing to alterations in translation rate/protein degradation or transcription/mRNA stability (reviewed by [47]).

Our findings argue for a species-specific role of RARRES1 in the human syncytium and its involvement in PE [48,49]. While in the murine placenta temporal changes of placental *Rarres1* expression were noted, gestational changes in Rarres1 expression in the rat were not studied. No di fferences in Rarres1 expression were detected in placentas of male or female fetuses.

#### *3.2. Expression of Rarres2 in Rodent Placenta*

We found that our Wistar rats expressed Rarres2 in the same placental compartments as Rarres1 (see above). The pronounced localization of Rarres2 in rat trophospongium supports our previous findings in human placenta [1], where RARRES2 was specifically expressed in extravillous trophoblasts (see above).

The expression and regulation of Rarres2 during rat pregnancy has been previously studied by Garces et al. [6] in Sprague Dawley rats at multiple gestational timepoints. In contrast to our findings, Garces et al. [6] found relevant Rarres2 staining in the labyrinthine trophoblast, besides the trophospongium in rats, which might have been due to the di fference in employed rat species (Wistar vs. Sprague Dawley). Similarly, they found syncytial RARRES2 expression besides its extravillous localization in the human placenta, which was not supported by our previous studies [1]. This difference requires further investigation. It might be due to divergent IHC techniques (fixation: methyl Carnoy's solution (our study) versus paraformaldehyde [6], rabbit antibody vs. recombinant full length human RARRES2 (our study) versus goa<sup>t</sup> antibody vs. N-terminal human RARRES2 [6]). The placental *Rarres2* expression increased until E16 in rats and then decreased until term, while rat maternal serum levels (ELISA) steadily decreased over the course of pregnancy in their animals [6]. This finding is in contrast to analyses of the same research group in humans, where RARRES2 levels were shown to rise significantly over the course of pregnancy [50]. This might argue for species-specific differences in the regulation of gestational Rarres2 expression and/or different functions of Rarres2 in murine and human placenta. The expression of *Rarres2* in rat mesenteric adipose tissue remained mainly constant, despite a singular increase at E19 [6]. IUGR (30% of total ad libitum maternal diet) resulted in a ~50% reduction of placental *Rarres2* expression in Sprague Dawley rats. The gestational expression pattern of *Rarres2* (i.e., maximum of placental expression around E16), however, remained unchanged [6]. Rarres2 expression in rodents was higher at the end of the second trimester, than at the end of the third trimester. This is in line with our findings of a more prominent Rarres2 expression in mouse placenta at gestational E15 compared to gestational day E18.5.

In contrast to Garces et al. [6], we did not observe a reduction of *Rarres2* in our isocaloric rat model of maternal protein restriction. This might be due to the divergent use of maternal diets (i.e., isocaloric protein restriction vs. total caloric reduction). We have just recently shown that our diet does not resemble a stress-model, unlike other models of total intake restriction [36]. Furthermore, models with total calorie restriction [51,52] seem more prone to develop insulin resistance after IUGR. Thus, the observed placental reduction of rat *Rarres2* expression under the condition of profound maternal food restriction [6] might underscore its proposed role as placental adipocytokine [6,53], with putative involvement in the development of maternal insulin resistance [54,55] and in feto-maternal metabolic homeostasis during pregnancy [56]. However, this hypothesis and the association of markers of insulin sensitivity with circulating RARRES2 is controversially discussed [50,57]. Unfortunately, there was no description of the influence of IUGR on *Rarres2* levels in rat maternal adipose tissue or serum by Garces et al. [6].

In vitro findings of Wang et al. [3] indicated that RARRES2 induces NO production in HUVECs. Our PE mouse model did not show local induction of *Rarres2* despite the knockout of eNOS, which suggests a lack of local negative feedback. At this point, systemic feedback-signaling cannot be ruled out, as circulating Rarres2 remained undetermined in our study.

#### *3.3. Expression of CmklR1 in Rodent Placenta*

Our finding of stable placental*CmklR1* expression levels in our Wistar rats resembled the gestational findings of Sanchez-Rebordelo et al. in Sprague Dawley rats [58]. Similarly, we could not detect significant differences or gestational changes of *CmklR1*-expression in our eNOS−/− mice. This finding is of interest, as CMKLR1 activation has been shown to induce vasoconstriction of peripheral vessels [59] and a reduced uterine blood flow is characteristic in eNOS−/− mice [60]. Thus, CmklR1 might play a minor role in the dysregulation of vascular tone in these mice. However, as knockout of CmklR1 seems to induce higher abortion rates in mice [61], different mechanisms of action need to be taken into consideration.

#### *3.4. IL-11 as a Novel Regulatory Cytokine in eNOS*−/− *Mice*

The level of IUGR in our rats, as determined by fetal weight, was comparable to our previous experience with this model [36,62]. Additionally, the observed level of fetal and placental weight reduction in eNOS−/− mice (18% and 22%, respectively; E18.5) was similar to findings from the literature (11% and 10%, respectively; E17 [28]), when compared to C57BL/6 control mice.

Interestingly, we found an induction of placental *IL-11* in our eNOS−/− mice at midgestation, which has not been shown previously. As IL-11 has been demonstrated to contribute to the development of inflammation in PE and placental vascular changes in mice [24,25], our finding might indicate respective changes in our rodent placentas, which are found in a similar manner in human PE.
