Mentha pulegium L. (Pennyroyal, Lamiaceae) Extracts Impose Abortion or Fetal-Mediated Toxicity in Pregnant Rats; Evidenced by the Modulation of Pregnancy Hormones, MiR-520, MiR-146a, TIMP-1 and MMP-9 Protein Expressions, Inflammatory State, Certain Related Signaling Pathways, and Metabolite Profiling via UPLC-ESI-TOF-MS

Pregnant women usually turn to natural products to relieve pregnancy-related ailments which might pose health risks. Mentha pulegium L. (MP, Lamiaceae) is a common insect repellent, and the present work validates its abortifacient capacity, targeting morphological anomalies, biological, and behavioral consequences, compared to misoprostol. The study also includes untargeted metabolite profiling of MP extract and fractions thereof viz. methylene chloride (MecH), ethyl acetate (EtOAc), butanol (But), and the remaining liquor (Rem. Aq.) by UPLC-ESI-MS-TOF, to unravel the constituents provoking abortion. Administration of MP extract/fractions, for three days starting from day 15th of gestation, affected fetal development by disrupting the uterine and placental tissues, or even caused pregnancy termination. These effects also entailed biochemical changes where they decreased progesterone and increased estradiol serum levels, modulated placental gene expressions of both MiR-(146a and 520), decreased uterine MMP-9, and up-regulated TIMP-1 protein expression, and empathized inflammatory responses (TNF-α, IL-1β). In addition, these alterations affected the brain's GFAP, BDNF, and 5-HT content and some of the behavioral parameters escorted by the open field test. All these incidences were also perceived in the misoprostol-treated group. A total of 128 metabolites were identified in the alcoholic extract of MP, including hydroxycinnamates, flavonoid conjugates, quinones, iridoids, and terpenes. MP extract was successful in terminating the pregnancy with minimal behavioral abnormalities and low toxicity margins.


Introduction
Abortifacients are compounds that induce miscarriage or abortion, and the herbs used for such purposes, either intentionally or by misconception may pose toxicity risks to the administering mother. The use of herbs and plant extracts to intervene with pregnancy and enforce menses is one ancient folk practice that entails complex social, legal, ethical, and health aspects [1]. Values are presented as mean (n = 6) ± SD and statistical analysis was carried out using one-way ANOVA followed by Tukey's post hoc multiple comparison test. As compared with control (a), misoprostol (b); p < 0.05.

Abortifacient and Biochemical Effects of Different Doses of Mentha pulegium L. Body Weight-Related
Administration of Mentha pulegium L. extract in a dose of 250 and 500 mg/kg resulted in complete abortion of two of the six pregnant rats relative to the effect of the standard drug misoprostol (three of the six rats), whereas 125 mg/kg revealed less effectiveness as it did not show a complete abortion among its group. However, the 125 mg/kg dose resulted in a reduction in progesterone levels and showed signs of toxicity viz. asymmetric distribution of fetuses in the two uterine horns, absorption points, hematoma, and absence of boundaries between fetal balls; presented in Table 1, Figure 1 (panel I) as red brackets, green arrows, star, and yellow arc, respectively.  The paraffin sections as imaged in Figure 1(IIB) pregnant uterine tissue revealed more thickening of the uterine wall with mild vacuolization of lining epithelium (black arrow), as well as mild hyperplasia of uterine gland (yellow arrow) and many dilated and congested blood vessels of intermuscular tunica vascular (star) in pregnant rats as compared tox Figure 1(IIA) non-pregnant rats that demonstrated the apparent intact uterine wall with almost intact lining epithelium with scattered apoptotic cell bodies (black arrow) moderate submucosal mononuclear cells infiltrates (red arrow) alternated with normal connective tissue and intact myometrium, as well as normal intermuscular blood vessels (star). Furthermore, treatment with different doses of the MP extract presented an order in efficacy/power, that was augmented by increasing the extract dose, where Figure 1(IID) 125 mg/kg of MP showed mild endometritis with mild neutrophilic infiltrates (red arrow), as well as mild congested uterine vasculatures (star). The Figure 1(IIE) MP 250 mg/kg dose demonstrated more apparent endometritis with moderate vacuolar changes of lining uterine epithelium (black arrow). Eventually, Figure 1(IIF) MP 500 mg/kg dose showed the worst histo-findings where significant hyperplasia of lining epithelium (black arrow) with many intra epithelial inflammatory cells infiltrates with milder infiltrates (red arrow) in submucosal layers and mild congested uterine BVs (star) and abundant neutrophilic infiltrates (red arrow) with mild congested uterine BVs (star) were observed as compared to medically aborted rats using Figure 1(IIC) misoprostol that manifested significant atrophy of lining epithelium (black arrow), cystic dilatation of many endometrial glands with abundant intraluminal desquamated lining epithelium (yellow arrow) accompanied with sever endometritis and lymphocytic infiltrates (red arrow).

Acute Toxicity Test
No deaths were observed. However, at the dose level of 500 mg/kg, signs of toxicity on uterine tissue were observed as mentioned previously. Rats treated with MP extract or misoprostol behaved similarly in terms of the number of completely aborted rats, the change in weight of pregnant rats/mothers and the number and weights of remaining fetuses relative to positive control mothers and healthy fetuses (Table 2, Figure 2I). On the other hand, the different fractions did not show complete abortion, but a significant reduction in fetal weight was observed, and was more pronounced with the but. and remaining aqueous fractions showing the maximum weight reduction in mothers on 18th day compared to the 15th day of gestation and the highest intrauterine growth retardation (IUGR) detected by fetal weight. Additionally, the MP and different fractions generated a different pattern of toxicity on uterine tissue, as shown in Figure 2(II), such as hematoma, intensive bleeding, asymmetrical distribution of fetuses in the two horns of uterus, complete absence of fetus on one uterine horn, resorption sites, absence of boundaries between fetal balls, and small fetal balls. Panel III presents IUGR associated with Figure 2(IIIb) misoprostol Figure 2(IIIc) MP extract and its fractions administration, as implied by their weights and lengths, morphologic malformations, hematoma, transparency of skin, and intensive bleeding as compared to those of fetuses in Figure 2(IIIa) control group. We omitted the fetal lengths measurements although there was significant difference between groups as high elasticity and flexibility of fetus may be a misleading during measurements and increase error. in Figure 2(II), such as hematoma, intensive bleeding, asymmetrical distribution of fetuses in the two horns of uterus, complete absence of fetus on one uterine horn, resorption sites, absence of boundaries between fetal balls, and small fetal balls. Panel III presents IUGR associated with Figure 2(IIIb) misoprostol Figure 2(IIIc) MP extract and its fractions administration, as implied by their weights and lengths, morphologic malformations, hematoma, transparency of skin, and intensive bleeding as compared to those of fetuses in Figure 2(IIIa) control group. We omitted the fetal lengths measurements although there was significant difference between groups as high elasticity and flexibility of fetus may be a misleading during measurements and increase error.  day) was utilized to analyze the change in rat weight and (B) the non-parametric data of fetal weight were expressed as the median (min-max) and were analyzed using the non-parametric Kruskal-Wallis followed by Dunn's as the post hoc test and student t test (Panel I). As shown in Figure 3I, pregnant female rats treated with misoprostol (100 ug/kg) and 250 mg/kg of MP extract both, had a marked decline in serum Figure 3(IA) progesterone levels and this effect reached Figure 3(IB) estradiol levels which was increased as compared to untreated pregnant female rats. It is worth mentioning that both MP extract and the standard drug, misoprostol produced similar values, where the progesterone levels were 11.17 ± 1.6 and 9.70 ± 2.0 ng/mL and the estradiol levels were 76.30 ± 3.0 and 72.55 ± 7.0 pg/mL, respectively. Meanwhile, the different fractions of MP extract also exhibited an abortifacient effect by the observed decline in serum progesterone levels. Additionally, the butanol and the remaining aqueous fractions effectively increased the levels of estradiol as compared to healthy pregnant one and those treated with total extract. balls, (E,F); asymmetrical distribution of fetuses in the two horns, absence of boundaries between fetal balls, reduced sized fetal balls, notable resorption sites, and intensive bleeding. The asymmetric distribution of fetuses in the two uterine horns: red brackets; resorption points: green arrows; hematoma: star; and absence of boundaries between fetal balls: yellow arc; reduced fetal balls: black circle (Panel II). MP (extract or fractions) and Misoprostol-mediated structural abnormalities in fetus including height, weight, and physical appearance; (a,a*) pregnant, (b,b*) Misoprostol, and (c,c*) MP; extract or fractions groups respectively (Panel III).

Effect of Mentha pulegium L. on the Serum Levels of Progesterone and Estradiol
As shown in Figure 3I, pregnant female rats treated with misoprostol (100 ug/kg) and 250 mg/kg of MP extract both, had a marked decline in serum Figure 3(IA) progesterone levels and this effect reached Figure 3(IB) estradiol levels which was increased as compared to untreated pregnant female rats. It is worth mentioning that both MP extract and the standard drug, misoprostol produced similar values, where the progesterone levels were 11.17 ± 1.6 and 9.70 ± 2.0 ng/mL and the estradiol levels were 76.30 ± 3.0 and 72.55 ± 7.0 pg/mL, respectively. Meanwhile, the different fractions of MP extract also exhibited an abortifacient effect by the observed decline in serum progesterone levels. Additionally, the butanol and the remaining aqueous fractions effectively increased the levels of estradiol as compared to healthy pregnant one and those treated with total extract. Values are presented as mean (n = 6) ± SD and statistical analysis was carried out using one-way ANOVA followed by Tukey's post hoc multiple comparison test. As compared with control (a), misoprostol (b), MP (c); p < 0.05. But.: butanol; EtOAc: ethyl acetate; MecH: methylene chloride; MiR-520: micro ribonucleic acid-520; MiR-146a: micro ribonucleic acid-146a; MMP-9: matrix metalloproteinase-9; MP: Mentha pulegium L.; TIMP-1: tissue inhibitor matrix metalloproteinase-1; Rem. aq.: remaining aqueous.  Figure 3II illustrates the disturbance in micro-RNA (MiR) expression in all the pregnant rats after misoprostol or MP (extract; 250 mg/kg) treatment where both increased Figure 3(IIA) MiR-520 and decreased Figure 3(IIB) MiR-146a protein expression as compared to the pregnant ones. Furthermore, these perturbations in both MiRs' expression were also figured in groups of the different fractions, knowing that Rem.aq. showed the highest increase in MiR-520 among fractions as compared to MP values, however all the fractions showed almost the same inhibitory effect as the total extract on MiR-146a. As revealed in Figure 3III, a marked elevation in Figure 3(IIIA) TIMP-1 that in sequence decreased Figure 3(IIIB) MMP-9 protein expression was detected on treatment with misoprostol or MP extract as compared to untreated pregnant female. Furthermore, medically aborted rats by our chosen standard showed a higher effect on the uterine TIMP-1 and MMP-9 expression as compared to values of MP extract. Regarding the fractions, they showed similar activity to total MP extract on TIMP-1 and MMP-9 expression with the MecH fraction exhibiting the highest effect on TIMP-1 and MMP-9 expressions, whereas the Rem. aq. fraction was more dominant on TIMP-1 expression. Figure 4I depicts the effect of MP extract administration on the levels of inflammatory mediators, Figure 4(IA) tumor necrosis factor alfa (TNF-α) and Figure 4(IB) interlukin-1 beta (IL-1β) which were emphatically spiked after its administration for three days and even exceeding them misoprostol effects on TNF-α levels (29.80% and 56.17% of control pregnant rats, respectively). Furthermore, the inflammatory activity of all the fractions under investigation were confirmed by the elevated inflammatory markers as compared to MP extract. Nevertheless, the Rem. Aq. Fraction recorded the highest increase in TNF-α among all the animals to reach 76.67% of normal pregnant ones. Concerning IL-1β, the But. And Rem. Aq. Fractions gave no significant difference compared to MP extract which marks a more significant effect in accordance with their employed doses.  Values are presented as mean (n = 6) ± SD and statistical analysis was carried out using one-way ANOVA followed by Tukey's post hoc multiple comparison test. As compared with control (a), misoprostol (b), MP (c); p < 0.05. BDNF: brain derived neurotrophic factor; But.: butanol; EtOAc: ethyl acetate; GFAP: Glial fibrillary acidic protein; 5-HT: 5-Hydroxytryptamine; IL-1β: interlukin- In the open field test ( Figure 4II), female rats behaved differently after misoprostol, MP, and different fraction administration, indicating a variety of behavioral abnormalities associated with medically induced abortion. The ones treated with misoprostol and EtOAc fraction showed immobilization for Misoprostol treatment affected the brain where a decreased in cortical contents of Figure 4(IIIA) GFAP was detected. Surprisingly, MP treatment on day 15th of gestation and for three days saved GFAP levels in cortical brain tissue and showed comparable levels of GFAP to the positive control. Additionally, all the fractions exerted the same effect of MP extract except for the Rem.aq. fraction which demonstrated a marked decline in GFAP levels as compared to MP extract results. However, all the rats when compared to the positive control demonstrated a significant reduction in cortical Figure 4(IIIB) BDNF levels after misoprostol and MP extract and/or fractions treatments; where all the fractions showed similar patterns with no statistical difference when compared to their corresponding extract except for But. The fraction did not affect BDNF levels showing the same values of control pregnant rat. It is worth mentioning that the decline in BDNF content by misoprostol treated group reached almost half of contents of the control group. Ultimately, Figure 4(IIIC) 5HT cortical contents were massively increased in misoprostol and MP extract treated rats as compared to the positive control. However, all the fractions did not affect serotonin levels except EtOAc fraction which caused a significant increase in its levels as compared to the ones with normal gestation. These results were recapitulated in Figure 4III. (MP, pennyroyal) and fractions thereof in a holistic, untargeted manner and correlate its phytochemical configuration to the anticipated abortifacient effect. The analytical procedure adopted herein warranted an adequate elution of the analytes within 30 min in an order of decreased polarity ( Figure 5). A total of 128 metabolites were detected in the 70% ethanol extract of pennyroyal, classified according to their phytochemical structures into hydroxycinnamates, several flavonoid conjugates, quinones, iridoids, and terpenes. A Tandem mass experiment was useful in distinguishing between O/C attached flavonoid and phenolic conjugates. Generally, the O-glycosyl attachment can be readily identified in MS 2 spectra by the neutral losses of 162, 146, and 132 amu indicative of hexose, deoxyhexose or pentose moieties, respectively, whereas the fragmentation pattern of the C-glycosyl conjugates shows major fragment ions resulting from neutral losses of 90 and 120 amu of pentose and hexose sugars, respectively [17].  Table 3.

Effect of Mentha pulegium L. on Serum Inflammatory Markers
Details for detection and assigning major metabolites, listed in Table 3 and depicted in Figures 5 and 6, are explained in the following section. Table 3. Metabolites tentatively identified in the 70% ethanol extract of (Mentha pulegium L.) and its fractions methylene chloride (MecH), ethyl acetate (EtOAc), butanol (But) and remaining liquor (Rem. aq.) via UPLC-ESI-TOF-MS in both; negative/positive ESI ionization modes.   Table 3.

Flavones and Derivatives
The   Figures S15 and S16). A major peak (39), eluded at t R = 21.51 min in the positive ionization mode, produced an intense molecular ion at m/z 611.2864 that exhibited a predicted chemical formula [C 28 H 35 O 15 ] + . The metabolite was annotated hesperetin 7-O-rutinoside (hesperidin) as reported previously in pennyroyal [21] and confirmed by the fragment ion at m/z [465.2265] + implying the loss of a deoxyhexose portion (146 amu) ( Figure S17).

Prenyl Flavanones
Mass spectrometry-based detection is a powerful tool for discovering minor secondary metabolites of potential medicinal merit in plant extracts [22]. The analytical procedure adopted in this study succeeded in tentative identification of seven prenylated flavones in MP extract for the first time.  Figure S19).

Discussion
Pregnant women usually turn to self-medication, relying on herbal and natural products to relieve pregnancy related ailments, or even for pregnancy termination, regardless of the lack of evidence concerning the safety profile of these products. This study targets the biological, behavioral, pathological, and toxicological consequences prior to administering the 70% ethanol extract of Mentha pulegium L. (MP) in pregnant female rats. This aim was executed over two parts; the first part includes probing the abortifacient potential using different doses of MP 70% ethanol extract to specify the effective abortifacient dose. Designating the abortifacient effect will rely on assessing some parameters, including percentage of abortion per group, number of resorbed fetuses, and further correlation to serum progesterone levels for confirmation of the results. On the 18th day of gestation, animals were sacrificed, and the uteri were examined. After specifying the dose of 250 mg/kg of MP extract as an abortifacient dose comparable to the standard misoprostol, the study attempts, in the second part, to unravel the mechanistic pathway underlying the abortifacient action. Additionally, the fractions of the 70% ethanol extract of MP, prepared as described in Section 5.1, were also similarly examined to trace the phytochemical constituents responsible for this effect.
Maternal exposure to MP (extract and/or fractions thereof) and misoprostol resulted in teratogenic anomalies represented by the asymmetrical distribution of fetuses in both uterine horns, resorption sites, intensive bleeding, complete absence of fetus on one uterine horn, absence of boundaries between fetal balls, and small fetal balls. Additionally, their fetuses suffered from IUGR, morphologic malformations, hematoma, transparency of skin, and intensive bleeding. In line with our findings, a previous documentation of teratogenicity and reproductive toxicity signs induced in pregnant rats exposed to pesticides on uterus and fetus investigation [31]. As a reinforcement to our data, unsuccessful use of misoprostol as an abortifacient was associated with preterm delivery, uterine rupture, and congenital malformations [15,16].
Estrogen and progesterone are cornerstones of healthy successful gestation with normal fetal growth [32]. Both hormones increase throughout pregnancy compared to unpregnant states, followed by a sharp decline in progesterone levels towards the end of gestation for labor induction [33]. Affecting these steroidal hormone trajectories, such as their production, transport, and receptors, may result in abortion (intended) or miscarriage (unintended), and IUGR [34,35]. In the current work, treatments with MP extract and misoprostol significantly affected serum sex hormones, which may be a reason for the complete abortion or fetal resorption observed in treated rats. In agreement with a previous study that demonstrated misoprostol power to modulate serum level of estrogen and progesterone to induce abortion, and misoprostol-mediated structural abnormalities in survival fetus [14]. Furthermore, all the fractions showed an anti-progesterone effect, whereas But. and the Rem.aq. exhibited additional definite estrogenic activities. This fluctuation in sex hormone levels were manifested by the obvious growth resistance and the increased fetal resorption that was documented in their groups. The findings reported herein imply that a possible abortifacient effect of MP is to decrease progesterone levels and increase estradiol levels.
MicroRNAs (MiRs) are non-coding small molecules known as "fine-tuners" rather than "switches" as they balance gene expression in the posttranscriptional state [36,37]. Nowadays, MiRs are gaining significant attention and became a field of study by researchers targeting its implication in the disease progression and maintaining tissue homeostasis. On the other side, MiRs titers during normal pregnancy or disorder states, such as abortion, preterm labor, pre-eclampsia, IUGR, and congenital deformity did not receive the same recognition. Hence, investigating them will aid in establishing new therapeutic and diagnostic targets in pregnancy related complications. A single change in miRNA expression, at any stage of placental or fetal development, can lead to serious complications which was observed herein. All the treated rats exhibited disrupted MiRs expression that may be one of the contributing factors to IUGR, fetal resorption, and even complete abortion. The administration of Rem.aq was associated with the highest up regulation in MiR-520 among fractions that sequentially resulted in additional toxicity observed in its group. Previous studies documented that MiR-520 upregulation in both placenta and maternal circulation is highly linked to placental pathologies, such as IUGR [38] and our records of declined fetal weight supports this fact. Concerning MiR-146a, its up-regulation is highly associated with balanced micro-environment of pregnancy, and its downregulation was involved in unexplained spontaneous abortion [39] in agreement with our data.
Substances that mimic or stimulate different MiR-520 isoforms regulate and increase estradiol [40] and that is concurrent with our findings where misoprostol, MP extract, and its fractions upregulated MiR-520 that, in turn, increased estradiol ( Figure 3II), resulting in abortion or IUGR. Concerning miRNAs-146, formerly it was documented that mature miRNAs-146 in endometrial epithelial cells are upregulated by progesterone that, in turn, will antagonize estrogen in endometrial epithelium [41], all these facts were reverted by MP (extract and fraction) and misoprostol administration to induce IUGR, fetal morphological deformity, or even complete abortion, where MP; total extract and or its fractions, and misoprostol decreased progesterone that resulted in MiR-146a down-regulation resulting in inhibition of the posttranscriptional regulation of estradiol and increasing its levels.
Matrix metalloproteinases (MMPs) are the most important among the classes of proteolytic enzyme in matrix degradation with similar structures and variety of functions [42]. MMPs show elevated levels during normal pregnancy, as a part of essential regulation of vascular and uterine remodeling, since significant changes in the hemodynamic and uterus is required to allow adequate uteroplacental blood flow and uterine expansion for appropriate fetal growth [43]. MMP-9 upregulation has been implicated in placentation, vasodilation, and uterine expansion during normal pregnancy. Hence, our investigation probed MMPs expression and its regulator, the tissue inhibitors of metalloproteinases (TIMPs), where MMPs are inactivated through TIMPs [43], that is produced by the same cells that produce MMP-9. In this study, all the pregnant rats treated with misoprostol, MP (250 mg/kg) extract, and fractions thereof at a dose of 125 mg/kg, demonstrated a marked inhibition in MMP-9 protein expression because of upregulation of its inhibitor the TIMP-1 protein expression. Previous studies reported the decreased MMP-9 levels may be involved in the IUGR pathogenesis during pregnancy [44], and that was evident in this work where all the fetuses showed retarded growth expressed in reduced fetal weights comparable to fetuses of untreated mothers. Moreover, MMP-9 levels are affected by pregnancy hormones (progesterone and estradiol) [45] as it was depicted formerly that progesterone regulates MMP expression/activity, which helps in management of the initiation, maintenance, and progression of endometrium [46]. Hence, the MP associated disturbance in both hormones may be the reason behind the elevated TIMP-1 and the decrease in MMP-9 uterine expression.
A certain amount of inflammatory response is paramount to normal healthy gestation, however, its exaggeration by the imbalance between pro-inflammatory (TNF-α, IL-6, and IL-8) and anti-inflammatory (IL-10, and IL-4) cytokines/chemokines, and the stimulation of different signaling molecules may sequel in many complications, such as preterm labor, IUGR, and spontaneous abortion, which may even increase the risk of morbidity and mortality to both the mother and their offspring [47,48]. Furthermore, increasing maternal inflammation through periods of fetal development is considered as a significant risk factor for birth deficits, morphological deformity, congenital diseases, and even still birth [49,50]. In our study, a marked inflammatory state in both, misoprostol and MP treated rats was manifested by elevation in the levels of TNF-α and IL-1β which participates in the observed fetal deformity in different groups, in agreement with previously published data [51]. Notably, the cytokine IL-1β was the first parameter to be linked to infection-induced premature labor. In addition, preterm labor and delivery were induced by IL-1β administration in experimental animals [52].
Phytochemical constituents in MP possess potential modulation of circulatory maternal hormones may trigger the inflammatory response as progesterone hinders the inflammatory related markers and molecules and keep the gestation environment protected [53]. Furthermore, the exaggerated inflammatory response observed herein after MP administration rationaled the disturbance in MMP-9/TIMP-1 expressions in their relevant treated groups (extract and fractions). Supporting this hypothesis, a previous study documented that overexpression of TNF-α, reduce MMP-2 and MMP-9, while an aggravation of the severity of abortion in rats further prompted TNF-α expression that sequentially decreased MMP-2 and MMP-9 levels [54]. In addition, using anti-inflammatory medication such as lactoferrin was able to prevent miscarriage/abortion by modulating MMP-9/TIMP-1 ratio and promoting repair [55]. Collectively, MP shifted the inflammatory state toward stimulation that increased risk of abortion or fetal resorption, IUGR, and morphological deformity.
Mood disorders viz depression and anxiety, addiction, and attempted suicide are mental health problems associated with abortion and placental pathologies induced by either stress or drugs [56]. Hence, brain investigations in this milieu have preoccupied researchers, and still there is no definite explanation available for its incidence nor prevention [57][58][59]. In this study, an open field test (OFT) was employed to examine the mental and behavioral status of the pregnant female rats under investigation. OFT is a widely used method of assessing rodent exploratory behavior, anxiety-like behaviors and determining whether a substance is sedative, poisonous, or stimulating. Consequently, the OFT assesses a variety of behavioral characteristics in addition to locomotion [60][61][62][63].
Taking into consideration all the analyzed parameters in OFT, only EtOAc and Rem.aq. treatment showed contrast behavioral effects, EtOAc treated rats showed depressive behaviors whereas Rem.aq. treated ones presented a state of anxiety and stress. Additionally, misoprostol treated group showed similar behavior to EtOAc treated group, even more pronounced. Several studies agree with the records of OFT variables and our interpretation [64][65][66]. In addition, this variance in behavior outcome, depression and/or anxiety, is common following early pregnancy loss or termination [67].
Glial fibrillary acidic protein (GFAP), is a major intermediate filament protein of differentiated astrocytes that plays a fundamental role in brain activity, not only in pregnancy but also physiologically as it modulates astrocytic motility and shape by supplying structural stability to astrocytic processes [68]. A previous study monitoring the changes in GFAP expression in the brain of rats during pregnancy and the beginning of lactation reported that GFAP expression is markedly increased on days 18 and 21 of gestation and is a sequence to changes in circulating sex steroid hormones. In the current study, administration of MP extract and its fraction on day 15th of gestation maintained GFAP cortical content on day 18th, except for Rem.aq., which showed a reduced GFAP content. Misoprostol treated rats presented the worst GFAP contents, implying a higher hazard/deleterious potential and this is the first documentation of this effect according to the author knowledge. Misoprostol and Rem.aq. changes in sex steroid hormones distinctly decreased GFAP cortical contents. Further, the decreased levels of GFAP may be in charge of the behavioral disturbance observed in rats treated with them [69].
Placental pathologies, mediated by altered MiRs levels, target gene expression of various signaling molecules that are important not only for fetal neurodevelopment but also for maintaining a normal maternal brain environment [70]. MiRs and other secreted compounds from the preeclamptic placenta showed direct effects on neurons and astrocytes [70], accordingly the changes of GFAP contents, may be related to misoprostol and Rem.aq. higher potential to modulate MicroRNAs.
Brain-derived neurotrophic factor (BDNF) is member of nerve growth factor family that has a structural role in neuronal development and function that is closely related to mood and behavioral disorders in many conditions [71]. All the treatments, except for But., decreased cortical contents of BDNF as compared to those of positive control reflecting neurogenesis inhibition. The reason behind this reduction may be the disturbance in hormonal status caused by both medical or herbal treatments for abortion induction [72,73]. Additionally, the decline in the cortical contents of BDNF contributes to the disruption in the behavior of pregnant rats as evidenced by OFT behavioral records after the different treatments. This was reinforced previously that the decline in BNDF levels have been linked to faster cognitive decline, poor memory performance, learning ability and even anxiety related behavior [62].
Serotonin, or 5-Hydroxytryptamine (5-HT) is the most manipulated monoamine for its role as a neurotransmitter in experimental animal models to establish its role in humans physiologically and pathologically. Misoprostol, MP extract and only EtOAc fraction increased cortical content of 5-HT, but the rest of the fractions did not affect brain 5-HT levels. Toxic placental effects perceived by these groups may also have a role in this monoamine increase. Our findings are supported by the theory which states that maternal mood and behavior are controlled by the placenta and that exposure to stressors or pathological conditions may result in an increase in 5-HT [74]. Additionally, altered serum progesterone and estradiol levels marked herein, participated in the increased 5-HT cortical content. As the regulation and modulation of neuronal excitability and neurotransmitter systems (5-HT is a one member of this system) is under the control of neurosteroids, steroid hormones with activity in the nervous system [75].
During the last century, the estrogenic properties of certain plants were recognized, initially in grazing animals, following the ingestion of clover revealed to contain high levels of the isoflavone compounds; formononełin and biochanin A [76]. Ever since, the term phytoestrogens have been established, referring to polyphenolic, non-steroidal plant constituents exerting estrogen-like biological action. Increasing epidemiological and chemical investigations pinpointed phytochemical classes viz. isoflavones, stilbenes, lignans, terpenes, and certain saponins as phytoestrogen candidates [77]. Phytoestrogens intake can modulate the endocrine system either by acting as agonists/antagonists for the estrogen receptors or by interfering with the estrogen biosynthesis process [78]. Apart from the estradiol-activating pathway, phytoestrogens can also exhibit their effects through non-genomic mechanisms, via interactions with cell surface receptors or by epigenetic mechanisms [41].
Like other members of the mint family, administration of Mentha pulegium L. extract and oil are known to pose fluctuations in estrogenic/progesterone levels and are used for terminating pregnancy [78,79]. The metabolic profile of MP extract reported herein in the results Section 2.3, encompasses a plethora of phytochemical constituents that were reported to possess effects on sex hormones.
Caffeic acid conjugates and depsides are the major water-soluble components detected in MP extract using UPLC-ESI-MS-TOF. A wide variety of pharmacological activities of these bioactive compounds have been evaluated including modulation of estrogen receptors [80,81]. Rosmarinic acid, a caffeic acid depside synthesized ubiquitously in the Lamiceae family and regarded as a marker, was reported to increase serum estradiol contents in diabetic and estrogen-deficient rats [82,83].
The metabolite profile of MP extract revealed the abundance of the flavone apigenin and several of its conjugates in addition to amentoflavone, which are reported in pennyroyal for the first time. Amentoflavone, a biflavonoid composed of two apigenin molecules, as well as its derivatives, were reported to be estrogen receptor agonists and inhibitors of aromatase enzymes in various experimental models [84,85]. A major metabolite of MP profile also was the flavone compound acacetin, which was proclaimed to be estrogen receptors agonists promoting uterine weight gain in mice [86]. Isoflavones are the most extensively studied compounds with established phytoestrogen activities [87], are represented in MP profile by a single constituent viz. daidzein-hexoside, enriched in the But. and Rem.aq. fractions. Daidzein was recorded to inhibit progesterone secretion, decrease the mRNA expression of α and βestrogen receptors in medium ovarian follicles of pigs [88]. Prenylated flavonoids are bioactive metabolites that possess a prenyl (isoprene) group attached to the flavane nucleus identified in about 37 genera of plant families viz. Leguminosae, Rutaceae, Guttiferae, and Apiaceae [89]. Prenylated flavonoids are mainly regarded as phytoalexins produced in response to pathogenic microorganism offense [90] and were identified in MP profile herein for the first time owing to the high sensitivity offered by the mass detection. These biometabolites are regarded as potent phytoestrogens with 8-Prenylnaringenin capacity to compete with 17 β-estradiol for estrogen receptors binding at a dose of 10 µM [89]. Although the analytical procedure adopted herein is not enough to assign the prenylation position, it is worth mentioning that the substitution at C8 is crucial for the activity, and 6-prenylnaringenin was proven to exert a much lower estrogenic activity than the 8-prenylated isomer [91].
Quinones are another group of metabolites reported previously in Salvia miltiorrhiza [92], and detected in MP extract and its fractions (Table 3) for the first time in this study. Tanshinone IIA, a reported potent phytoestrogen with active conformation analogous to 17 β-estradiol and high binding affinity to estrogen receptors, was detected in the 70% ethanol extract of MP, MecH and Rem. aq. fractions [93].
It is worth mentioning that all MP fractions imposed fetal morphologic malformations, if not complete abortion. In brief, all the fractions presented a significant reduction in fetal weight especially with the But. and rem. Aq. fractions in addition to the highest intrauterine growth retardation (IUGR). Both fractions also caused elevation in the levels of estradiol with a decline in serum progesterone levels as compared to healthy pregnant ones and those treated with MP extract. The rem. Aq. fraction showed the highest increase in MiR-520 and TNF-α values, compared to MP, whereas MecH fraction exhibited the highest effect on TIMP-1 and MMP-9 expressions. EtOAc treated rats showed depressive behaviors contrarily to Rem. aq. treated ones which demonstrated a state of anxiety and stress. Misoprostol treated groups showed similar, or even more pronounced behavioral state to EtOAc treated group. MP extract and its fractions caused a reduced GFAP content, and misoprostol resulted in the least GFAP contents. Misoprostol, MP extract, and EtOAc fraction increased cortical content of 5-HT. All the treatments, except for the But. fraction, decreased the cortical contents of BDNF as compared to those of positive control reflecting neurogenesis inhibition.

Conclusions
The data of the current study verifies and documents the traditional reputation of employing Mentha pulegium L. (Pennyroyal) as an abortifacient by modulating pregnancy hormone levels, gene/protein expression, and the inflammatory process. All these cues afflicted the fetal morphology and the mothers' brains and behavior. Indeed, MP was successful in terminating pregnancy with minimal behavioral abnormalities and low toxicity margin, yet contrarily to misoprostol, MP extract in the dose of 250 mg/kg maintained cortical contents of GFAP. UPLC-ESI-TOF-MS examination revealed MP and its fractions to encompass several metabolites with phytoestrogenic potential and mediating for the abortifacient effect. The findings reported herein encourage the future perspective of possible use of MP extract in contraception if administered at an earlier phase of the estrous cycle. Major phytochemical constituents identified in this study viz. rosmarinic acid, hesperidin and ursolic acid can be examined for their individual effects following the same presented model. The workflow reported in this study can be employed to investigate the safety profile of other edible or medicinal plants, particularly the herbs of family Lamiaceae, which are commonly consumed during pregnancy.

Plant Material
Aerial parts of Mentha pulegium L. (pennyroyal) were collected from El-Beheira Governorate in May, after blooming of the flowers. The taxonomic authentication was confirmed by Dr. Abdel Halem Abdel El Mogali, Professor in flora and phytotaxonomy research, Agricultural Research Centre, where a voucher specimen No. (M.P. 998) is kept. Pennyroyal was dried in the shade then pulverized, and 200 g of the powdered plant was prepared by maceration in 70% ethanol. The extract was concentrated on a rotary evaporator (Buchi R-210 evaporator, Flawil, Switzerland) until completely dry, yielding 45 g for biological investigations and UPLC-ESI-TOF-MS analyses. Portion of AE (25 g) was reconstituted in water and partitioned with increasing polarity organic solvents, as methylene chloride (MecH), ethyl acetate (EtOAc), and butanol (But). Each organic fraction was carefully evaporated until completely dry, and the remaining liquor (Rem. aq.) was also concentrated under vacuum and freeze dried for UPLC-ESI-TOF-MS examination. Dried fractions were freshly dissolved in distilled water by sonication and 2 drops of tween 80 if needed before oral administration to the rats.

Chemicals and Reagents
Thermo-Fisher Scientific Co. (Waltham, MA, USA) supplied the solvents acetonitrile and methanol (HPLC grade), while Sigma-Aldrich Co. (St. Louis, MO, USA) supplied the mobile phase solvents ammonium hydroxide, formic acid 98 %, and misoprostol standard drug. Analytical grade solvents (methylene chloride, ethyl acetate, and butanol) used in preparing the fractions were also purchased from Sigma-Aldrich Company.

Part I: Dose Selection
A dose range-finding study was conducted to screen and select the optimal abortive dose of MP extract using 48 mature albino rats, comprised of40 females (180-200 g) and 8 males (230-250 g) [94] purchased from the National Research Centre, Giza, Egypt and housed in clean plastic cages open to the room environment temperature of 24 ± 1 • C, humidity (55 ± 5%), exposed to 12 h light/dark cycle, and allowed free access to food and water throughout the study. The study was approved by Misr University for Science and Technology's institutional ethics committee for the use of laboratory animals with approval No. PT13 on 6 March 2021 and was supervised by the Guide for the Care and Use of Laboratory Animals (NIH publication, 1996).
After acclimation for one week, mating was ordered in a 1:5 ratio, where one male rat was introduced into five caged fertile female rats overnight [95]. Mating was confirmed on the next day in the early morning by carefully inserting a cotton-tipped swap moistened with normal saline into the vaginal cavity, then withdrawing after gently rolling it against the vaginal wall [14,96]. Successful mating is marked by the presence of vaginal plugs, positive animals were considered pregnant, and this day was considered as their first day of gestation [97][98][99]. Pregnant females were also confirmed by non-invasive techniques as monitoring the physical appearance of the enlarged abdomen and eventually the noticeable body mass increase [95,100]. Hence, we used six additional un-mated female rats as a negative control to compare changes in weight through the experiment, and two of them were sacrificed at the end of the experiment, and their uteruses were preserved in formalin for histopathological examination.
On the 15th day of gestation, female rats with confirmed pregnancies were randomly divided into five groups as follows: Group I: pregnant rats serving as positive control group. Group II: rats were treated by misoprostol (100 ug/kg; once per day orally for three days) to induce abortion and served as standard control group. This dose was selected after a pilot study (data not included), guided by previous studies used misoprostol alone in abortion induction [14,101]. Group III, IV, and V received 125, 250, and 500 mg/kg of MP extract, respectively (once per day for three days starting on day 15th). The number of the completely aborted rats per group, abortion percent were calculated, the number of fetuses remaining were counted, and serum progesterone was leveled. These parameters along with the histopathological examination (n = 2/per group) were used for the assessment of the optimal abortion dose of different doses of MP extract.

Part II: Identification of the Abortifacient Mechanism of MP Extract and Fractions Thereof
In this part, the sample size of male rats was increased as a conclusion from the different performed trials in the dose selection part where mating ratio 1:5 was associated with higher pregnancy failure and time consuming. Consequently, 20 male rats were randomly used for the mating of 42 female rats with a ratio 1:1 with the same previously described procedures of pregnancy confirmation in the previous section (dose selection). After the verification, the positive female rats were randomly allocated as follows: Group I: pregnant female rats that is served as positive control; Group II: rats in this group received misoprostol (100 ug/kg, orally) on day 15th of gestation; Group III: received the optimal dose of MP extract selected according to the results perceived from the first part of the study; 250 mg/kg once per day for three consecutive days starting on day 15 till day 17; Groups IV, V, VI, and VII received different fractions of MP orally (125 mg/kg; MecH, EtOAc, But., and Rem. Aq., respectively).

Tissue and Serum Preparation
On the 18th day of gestation, all animals' blood samples were collected via the retroorbital sinus under anesthesia, followed by scarification by cervical dislocation, then decapitated. The uterus, placenta, and brain cortex were rapidly removed and washed with saline. The fetuses were isolated from the uterus of different groups to screen for the changes in weight and morphology to healthy ones correspondingly.

Behavioral Open Field Test (OFT)
The test was commenced 24 h after the last treatment. The apparatus matches the specifications described by previous studies [102,103]. Each rat was placed individually in the center of the apparatus to be tested, the apparatus was cleaned with 70% ethanol and allowed to dry prior to each individual trial. A video camera was placed in the center of the arena above the apparatus to record the experiment for the subsequent analysis of (i) latency time (time spent from placing the animal into the center of the device till its first move), (ii) ambulation frequency (number of squares crossed by the animal over 3 min.), (iii) rearing frequency (number of times the animal stood extended and extended on rear appendages with no forelimb support), and (iv) grooming frequency (number of times of rapid cleaning movements of the forelegs towards the face and/or the body (licking the fur, washing the face, or scratching behavior) as guided by a previous study [104].

Parameters Assessed by ELISA Technique
As mentioned in Appendix A.

Histopathological Examinations
As mentioned in Appendix B.

Quantitative Real-Time PCR for miR-520 and miR146a Placental Expression
As mentioned in Appendix C.

Western Blotting
As mentioned in Appendix D.

Statistical Analysis
GraphPad Prism software (version 5.0d; GraphPad Software, Inc., San Diego, CA, USA) was used for all statistical analyses. The data were articulated using analysis of variance (ANOVA), followed by Tukey's post hoc test. A Two-Way Repeated Measures (RM) ANOVA with one factor repetition (time: day) was employed to analyze the change in rat weight. The non-parametric data were expressed as the median (min-max) and were analyzed using the non-parametric Kruskal-Wallis followed by Dunn's as the post hoc test.
All the values were expressed as the mean of six rats ± S.D.A and the difference at p < 0.05 was considered statistically significant.

Metabolite Profiling of Pennyroyal (MP) Extract and Active Fractions by
UPLC-ESI-TOF-MS 5.11.1. Sample Preparation Stock solutions of pennyroyal (MP) extract and the active fractions (MecH, EtOAc, But., and Rem. Aq.), were prepared by separately dissolving 50 mg of each extract in 1 ml of the solvent mixture (Water: Methanol: Acetonitrile, 50:25:25 v/v). The samples were vortexed for 2 min, ultrasonicated for 10 min, and centrifuged at 10,000 rpm for 5 min, to obtain a final concentration of 2.5 µg/µL and 10 µL of the prepared solution was injected. The LC-MS analysis procedure was also used for blank samples, quality control samples, and internal standards used for experiment setup validation. Instrument setup, spectral acquisition, and MS-data processing were adjusted as mentioned in [105,106], and provided in the Supplementary Materials. Mass spectral data of identified metabolites in both ionization modes were compared to available online libraries and databases, viz. Human Metabolome Database and Pubchem. In addition, retention times and fragmentation patterns were compared to available reference standards.

Instrument and Spectral Acquisition
As mentioned in Appendix E.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/toxins14050347/s1, Figure S1: ESI-MS/MS Spectrum of peak (1) in the negative ion mode; Figure S2: ESI-MS/MS Spectrum of peak (5) in the positive ion mode; Figure S3: ESI-MS/MS Spectrum of peak (11) in the positive ion mode; Figure S4: ESI-MS/MS Spectrum of peak (8) in the positive ion mode; Figure S5: ESI-MS/MS Spectrum of peak (10) in the positive ion mode; Figure S6: ESI-MS/MS Spectrum of peak (9) in the negative ion mode; Figure  S7: ESI-MS/MS Spectrum of peak (15) in the negative ion mode; Figure S8: ESI-MS/MS Spectrum of peak (16) in the positive ion mode; Figure S9: ESI-MS/MS Spectrum of peak (21) in the positive ion mode; Figure S10: ESI-MS/MS Spectrum of peak (22) in the positive ion mode; Figure S11: ESI-MS/MS Spectrum of peak (20) in the positive ion mode; Figure S12: ESI-MS/MS Spectrum of peak (17) in the negative ion mode; Figure S13: ESI-MS/MS Spectrum of peak (29) in the positive ion mode; Figure S14: ESI-MS/MS Spectrum of peak (23) in the positive ion mode; Figure S15: ESI-MS/MS Spectrum of peak (33) in the positive ion mode; Figure S16: ESI-MS/MS Spectrum of peak (36) in the negative ion mode; Figure S17: ESI-MS/MS Spectrum of peak (39) in the positive ion mode; Figure S18: ESI-MS/MS Spectrum of peak (41) in the negative ion mode; Figure S19: ESI-MS/MS Spectrum of peak (45) in the negative ion mode; Figure S20: ESI-MS/MS Spectrum of peak (50) in the positive ion mode; Figure S21: ESI-MS/MS Spectrum of peak (51) in the positive ion mode; Figure S22: ESI-MS/MS Spectrum of peak (70) in the positive ion mode; Figure S23: ESI-MS/MS Spectrum of peak (75) in the negative ion mode; Figure S24: ESI-MS/MS Spectrum of peak (69) in the positive ion mode; Figure S25: ESI-MS/MS Spectrum of peak (71) in the positive ion mode; Figure S26: ESI-MS/MS Spectrum of peak (76) in the positive ion mode; Figure S27: ESI-MS/MS Spectrum of peak (83) in the positive ion mode; Figure S28: ESI-MS/MS Spectrum of peak (85) in the positive ion mode; Figure S29: ESI-MS/MS Spectrum of peak (92) in the positive ion mode; Table S1: The PCR primers used for quantitative PCR analysis  Acknowledgments: Authors would like to acknowledge Mohamed Rasheed for procuring the plant material and extend their appreciation to staff of "The Proteomics and Metabolomics Unit" (https:// www.57357.org/en/department/proteomics-unit-dept/about-department/), at the Children Cancer Hospital (CCHE-57357), for their tenacious efforts in empowering research in Egypt.

Conflicts of Interest:
The authors declare no conflict of interest. In a total volume of 25 µL per reaction volume, quantitative RT-PCR was performed using the miScript SYBR ® Green PCR kit and protocol for mature miRNAs quantitative assessment (Qiagen, Valencia, CA, USA) utilizing the specific MiR-(520 and 146a) primers. The PCR primers used are presented in Table S1. The fold change in MiRs expression were calculated by Equation 2−∆∆Ct.