Dynamic expression of endometrial adhesion G protein-coupled receptors during the menstrual cycle and early mouse pregnancy: modulation by ovarian stimulation.

To quantify endometrial ADGR gene expression across the natural menstrual cycle, an in-silico analysis was performed on an existing GEO dataset (GDS2052). GDS2052 is derived from the microarray analysis of human endometrial gene expression throughout the menstrual cycle. The heatmap shows 32 known ADGR coding genes, representing 9 ADGR subfamilies that were expressed across the menstrual cycle ( Figure 5A ). Of the genes identified, 17 ADGRs were noted to have multiple probes per gene ( Supplemental Table 2 ), and overall trends of directionality in gene expression changes were conserved across the various probes (data not shown). ADGR gene expression (based on normalized hit counts) across the proliferative (PROLIF) and early, mid, and late secretory phases (ESE, MS and LSE, respectively) of the menstrual cycle was visualized on a heatmap ( Figure 5A ) and differential gene expression was identified by conducting pairwise comparisons of normalized hit counts ( Figure 5B – Q ). When comparing the PROLIF versus ESE, MS and LSE phases, 2, 3 and 11 differentially expressed ADGRs were identified, respectively and these represented a total of 13 unique ADGR genes ( Figure 5 ; Supplemental Table 7 ). Among the 2 differentially expressed ADGR genes between the PROLIF and ESE, ADGRC1 had significantly higher expression in the ESE phase while ADGRE5 had significantly lower expression. In the MS phase ADGRF1 had significantly higher expression across multiple probes, while ADGRB2 and ADGRL1 (also across multiple probes) had significantly lower expression when compared to the PROLIF phase. Lastly, when comparing the PROLIF to the LSE phase, 6 of the 11 differentially expressed ADGR genes demonstrated an increase in expression in the LSE across all probes ( ADGRE2, ADGRF1, ADGRF5, ADGRG3, ADGRG7, and ADGRL4), while another 5 ADGR genes demonstrated a decrease in expression in the LSE phase ( ADGRA2, ADGRB2, ADGRB3, ADGRL1 and ADGRL3). Next, a comparison of gene expression was made within the secretory phase. In the secretory phase, 4 ADGRs were found to be differentially expressed between the ESE and MS phases ( ADGRC1, ADGRF1, ADGRL2 , and ADGRL4 ). Three of the genes ( ADGRF1, ADGRL2 , and ADGRL4 ) increased in expression in the MS, while ADGRC1 was the only gene to decrease in expression ( Figure 5 ; Supplemental Table 8 ). Between the ESE and LSE phases, 13 ADGR genes were differentially expressed, with 9 genes exhibiting increased expression ( ADGRE2, ADGRE5, ADGRF1, ADGRF4, ADGRF5, ADGRG1, ADGRG3, ADGRL2 and ADGRL4) and four genes demonstrated decreased expression ( ADGRA2, ADGRC1, ADGRD1, and ADGRL1, Figure 5 ; Supplemental Table 8 ). Together, these data show that ADGR gene expression varies significantly with menstrual cycle stage, suggesting that ADGR expression is regulated by E2 and P4. The data also suggest that ADGRB2 , ADGRF1 , and ADGRL1 , which were differentially expressed in the MS phase relative to the PROLIF phase may regulate pre-decidualization during the window of implantation. The in-silico analysis of GEO dataset GDS2052 highlighted the differential expression of ADGRB2 , ADGRF1 , and ADGRL1 between the PROLIF and MS phases. We sought to validate the expression of these three genes using RT-qPCR in prospectively collected endometrial biopsies in the PROLIF (n=4) and MS (n=5) phases ( Figure 2 ) of ovulatory natural menstrual cycles. Histologic assessment confirmed menstrual cycle stage for all samples included in the analysis. The age and reproductive hormone profiles of the study participants are presented in Table 1 and Supplemental Table 3 . Since ADGRG2 was reported to enhance human endometrial stromal cell decidualization in vitro ( Yoo et al. 2017 ), we sought to also confirm its expression. RT-qPCR showed upward trends in expression for ADGRF1 and ADGRG2 , but overall, no significant change in expression for all four genes between the PROLIF and MS phases ( Figure 6A – D ). One PROLIF phase EMB sample did not express ADGRF1 and was thereby excluded from the calculation of normalized fold change for this gene. These findings differ from the in-silico analysis which demonstrated that ADGRF1 had significantly higher expression, while ADGRB2 and ADGRL1 had significantly lower expression in the MS when compared to the PROLIF phase ( Figure 5 , Supplemental Table 7 ). No significant difference was seen for ADGRG2 expression across menstrual cycle stages in the in-silico analysis. ADGR genes in multiple subfamilies ( ADGRA , ADGRB, ADGRC, ADGRE, ADGRF, and ADGRL) demonstrated differential expression between the PROLIF and secretory phases in the in-silico analysis ( Figure 5 , Supplemental table 7 ). Given our interest in these ADGR subfamilies and the inability to investigate human endometrial remodeling and stromal cell decidualization during early pregnancy, a mouse model of early pregnancy was studied to understand the potential role of ADGR genes in these processes. The time points chosen included embryonic day (E) E0.5 to represent a pre-implantation stage; E5.5, a post-implantation stage containing the newly formed decidua; and E6.5, a stage of early pregnancy with angiogenesis and increased decidual size. E5.5 and E6.5 represent advancing decidual changes in mouse pregnancy. To begin, a RT-PCR screen was conducted to identify which Adgr genes, representing subfamilies, Adgra, Adgrb, Adgrc, Adgre, Adgrf, Adgrg , and Adgrl , were expressed in the non-pregnant (NP) uterus and E7.5 decidua. Pregnancy at E7.5 was chosen for this study since on this day, decidua formation is complete. Based on the RT-PCR screen, it was observed that the following Adgr genes were expressed either in the NP and/or E7.5 samples: Adgra1 , Adgra2 , Adgra3 ( Figure 7A ), Adgrc1 , Adgrc2 Adgrc3 ( Figure 8A ), Adgre1, Adgre5 ( Figure 9A ), Adgrf1 , Adgrf2, Adgrf4, Adgrf5 ( Figure 10A ) Adgrg1 , Adgrg2 , Adgrg3 , Adgrg5 , Adgrg6 , Adgrg7 ( Figure 1A ), Adgrl1 , Adgrl2 , Adgrl3 and Adgrl4 ( Figure 12A ). Following the RT-PCR screen, those genes that were expressed in the E7.5 decidua were chosen for qPCR characterization of expression in early pregnancy at E0.5, 5.5 and 6.5. These genes were Adgra1 , Adgra2 , Adgra3 ( Figure 7 ), Adgrc1 , Adgrc2, Adgrc3 ( Figure 8 ), Adgre1 , Adgre5 ( Figure 9 ), Adgrf2, Adgrf4, Adgrf5 ( Figure 10 ) Adgrg1 , Adgrg2 , Adgrg3 , Adgrg6 , Adgrg7 ( Figure 1 ) Adgrl1 , Adgrl2 and Adgrl4 ( Figure 12 ). None of the members within the Adgrb family were expressed in either the NP uterus or E7.5 decidua, and thus were not included in the RT-qPCR analysis. In the Adgra family, both Adgra1 and Adgra2 demonstrated significantly higher expression on E6.5 compared to E0.5 (5-fold and 4-fold, respectively). No significant changes in expression were seen in Adgra3 across all time points ( Figure 7B – D ). Only Adgrc3 demonstrated significant expression change in early mouse pregnancy within the Adgrc family, with a 20-fold increase from E0.5 to E6.5 ( Figure 8D ). In the Adgre family, Adgre1 demonstrated a 7-fold decrease in expression from E0.5 to E5.5 ( Figure 9B ), whereas Adgre5 had a 2.5-fold increase in expression from E5.5 to E6.5 ( Figure 9C ). Adgf4 demonstrated a 4-fold increase in expression from E0.5 to E6.5 ( Figure 10C ), but no changes in expression were seen for the other Adgrf genes ( Figure 10B , D ). In the Adgrg family, Adgrg1 , Adgrg6, and Adgrg7 all had significantly lower expression at E5.5 when compared to E0.5 with a 4-fold, 30-fold and 1000-fold decrease in expression respectively ( Figure 1B , E and F ). Whereas Adgrg2 and Adgrg3 had a 10- and 25-fold increase in expression respectively from E5.5 to E6.5 ( Figure 1C , D ). Lastly, in the Adgrl family, Adgrl1 and Adgrl4 had a 25-fold and 17-fold increase in expression from E0.5 to E6.5 ( Figure 12B , D ) and Adgrl2 had a 4-fold increase in expression from E5.5 to E6.5 ( Figure 12C ). Together these data demonstrate that Adgr gene expression across multiple families is altered in early mouse pregnancy. Similar to changes in gene expression seen between the proliferative and secretory phases of the menstrual cycle, ADGR gene expression also changes with decidualization of the pregnant mouse uterus. We performed mRNA sequencing (mRNA-Seq) on prospectively collected endometrial tissue samples to further characterize ADGR mRNA expression in ovulatory menstrual cycles. Endometrial samples from the PO and MS phases of natural cycles were analyzed to determine whether ADGR genes were among the differentially expressed genes (DEGs) ( Figure 13A ). Expression of the four DEGs ( ADGRC1, ADGRF1, ADGRL2, and ADGRL4 ) identified in the in-silico analysis when comparing the ESE to the MS phase ( Figure 5 , Supplemental Table 8 ), was also examined using RT-PCR in prospectively collected biopsies from the PO and MS periods of natural cycles. On the RT-PCR, ADGRF1 was not expressed in PO samples from a NC ( Figure 13B ). However, ADGRC1, ADGRL2 , and ADGRL4 were expressed in both the PO and MS phases of NC samples ( Figure 13B ). These genes were further analyzed in the mRNA seq dataset ( Figure 13C – F ; Supplemental Table 9 ). In the in-silico analysis, ADGRC1 expression was significantly decreased in the MS phase compared to the ESE phase ( Figure 5E , Supplemental Table 8 ). Whereas ADGRF1, ADGRL2 and ADGRL4 had significantly higher expression in the MS phase ( Figure 5I , O and Q , Supplemental table 8 ). When comparing the MS to the PO phase in our prospective NC cohort , ADGRF1 and ADGRL4 were found to be significantly upregulated, similar to what was seen in within the secretory phase of the in silico analysis; while ADGRC1 and ADGRL2 were unchanged in expression ( Figure 13C – F ; Supplemental Table 9 ). Based on the dynamic the ADGR expression observed within a natural menstrual cycle and early mouse pregnancy, we hypothesized that ADGR expression is hormone mediated. To determine if altered E2 and P4 levels would disrupt endometrial ADGR gene expression, EMB samples were prospectively collected during the secretory phase of ovarian stimulation cycles. Analysis of serum hormone profiles demonstrated significantly higher E2 and P4 levels during the PO phase with ovarian stimulation. Whereas in the MS phase, only E2 was significantly increased with ovarian stimulation ( Figure 4 , Table 2 ). RT-PCR findings in ovarian stimulation were similar to what was seen in the natural cycle, ADGRF1 was not expressed in the PO phase, while ADGRC1, ADGRL2 , and ADGRL4 were expressed in both the PO and MS phases ( Figure 13B ). mRNA-Seq of the prospectively collected PO phase endometrium, demonstrated significantly lower expression of ADGRF1 and ADGRL2 with ovarian stimulation compared to a natural cycle ( Figure 13D – E , Supplemental table 9 ). Whereas ADGRC1 and ADGRL4 were unchanged in expression ( Figure 13C , F ; Supplemental Table 9 ). In the MS phase, none of the four genes exhibited a significant change in expression between natural and ovarian stimulation conditions ( Figure 13C – F ; Supplemental Table 9 ). However, when comparing the MS phase to the PO period within an ovarian stimulation cycle, three genes (ADGRF1, ADGRL2 and ADGRL4), had significantly higher expression in the MS phase ( Figure 13C – F ; Supplemental Table 9 ). Whereas, ADGRC1 had significantly lower expression in the MS phase ( Figure 13C , Supplemental Table 9 ) Together, these data show that ADGR gene expression is dynamic within the secretory phase and that ovarian stimulation alters ADGR expression. The changes in gene expression seen with ovarian stimulation are presumably hormone mediated given the significant changes in serum E2 and P4 which are also seen. The most significant alteration in both gene expression and hormone profiles seen with ovarian stimulation, appears to be in the periovulatory phase.

An in-silico analysis was performed using GEO dataset GDS2052, which consists of a microarray analysis of human endometrial gene expression across the menstrual cycle ( Talbi et al. 2006 ). GDS2052 includes 27 normo-ovulatory women ranging from 23 to 50 years old, with regular menstrual cycles lasting 24 to 35 days ( Figure 1 ). Recruited subjects were volunteers who had not used steroid medication for at least three months prior to the biopsy. Women with laparoscopy-proven endometriosis, or inflammation within the endometrium at the time of enrollment, were excluded from the study. Endometrial tissue was collected from either hysterectomy samples, performed for benign indications (n=20), or endometrial biopsies (EMBs) obtained via aspiration with a Pipelle catheter or endometrial curetting during a hysteroscopic procedure (n=7). Pathological diagnoses at the time of hysterectomy included uterine fibroids (n=13), adenomyosis (n=3), ovarian cyst (n=1) and pelvic organ prolapse (n=3). Endometrial tissue was collected at different phases of the menstrual cycle and staged in a blinded fashion by two or more pathologists using Noyes criteria ( Noyes, Hertig, and Rock 1975 ) for determination of menstrual cycle phase. For the in silico analysis reported in this study, 21 subjects were included, while six subjects with an ambiguous cycle stage were excluded ( Figure 1 ; Supplemental Table 1 ). Included subjects ranged from 23–49 years old. Histologically staged endometrial samples included proliferative (n=4), early secretory (n=3), mid secretory (n=8) and late secretory (n=6) phases. Endometrial ADGR gene expression in these 21 samples was determined using the high-density human genome (HG) U133 Plus 2.0 Arrays (Affymetrix), containing 42,203 genes and 12,397 expressed sequence tags (ESTs). This array contains probes against all 32 ADGR genes known to encode protein products ( Hamann et al. 2015 ) ( Supplemental Table 2 ). Probe level intensities were processed and hit, or raw counts were normalized as described ( Talbi et al. 2006 ). Differentially expressed genes were identified using limma/voom (v3.52.2) and the number of significant, differentially expressed genes was determined using 5% false detection rate (FDR) (q-value <0.05) as the threshold. Endometrial biopsy (EMB) samples were prospectively collected from women in natural, ovulatory menstrual cycles. Subjects were recruited under an Institutional Review Board approved protocol (Pro2018002041) at Rutgers New Jersey Medical School. Inclusion criteria included (1) age between 21–45; (2) regular menstrual cycle (21–35 days); (3) prior pregnancy. Exclusion criteria included: (1) endocrine or autoimmune disorders; (2) anatomic disorders of the reproductive tract, including hydrosalpinges, submucosal fibroids and endometrial polyps; (3) history of recurrent pregnancy loss; (4) use of hormonal medication within 3 months prior to enrollment; (5) history of pregnancy complications such as preeclampsia or intrauterine growth restriction; (6) pregnant or trying to conceive. A total of 15 women, ranging from 21 to 33 years of age, were recruited in natural cycles ( Figure 2 , Supplemental Table 3 ). EMB samples were collected in the proliferative (PROLIF) phase (cycle days 10–13; n=4), peri-ovulatory (PO) (n=6) period and mid-secretory (MS) (n=5) phase in natural cycles using a disposable endometrial Pipelle device (MedGyn Pipette, Westmont, IL, USA). Women recruited for secretory phase biopsies in a natural cycle (NC) used commercial ovulation predictor tests to detect the urinary LH surge prior to ovulation. The day of LH surge detection was deemed LH+0. Biopsies were either performed in the PO phase, 12–24 hours after detection of the urinary LH surge (PO-NC), or in the MS phase, 8–9 days after detection of the LH surge (MS-NC). To determine if altered serum E2 and P4 levels would disrupt endometrial ADGR gene expression, EMB samples were also prospectively collected from ovarian stimulation cycles. Women undergoing ovarian stimulation were recruited from a university-affiliated fertility clinic, under the same Institutional Review Board approved protocol stated above (Pro2018002041). Inclusion criteria were (1) age between 21–45; (2) IVF without fresh embryo transfer; (3) diagnosis of male factor, tubal factor, or unexplained infertility; (4) undergoing ovarian stimulation for oocyte cryopreservation; (5) plan to freeze all oocytes or embryos obtained during the treatment. The same exclusion criteria applied to women in natural cycles were applied to women undergoing ovarian stimulation. Nine women were recruited, ages ranging from 31–42 years old ( Figure 2 , Supplemental Table 4 ). OS was performed using a GnRH-antagonist protocol with Cetrotide ® (ASTA Medica AG, Frankfurt, HE, Germany), Ganirelix ® (Organon, Jersey City, NJ, USA) or Fyremadel ® (Sun Pharmaceutical Industries Ltd., Halol, GJ, India). Gonadotropins used for stimulation included Menopur ® (Ferring Pharmaceuticals INC, Parsippany, NJ, USA) in combination with Follistim ® (Organon, Jersey City, NJ, USA) or Gonal-F ® (Merck, Madrid, Spain). Starting gonadotropin dose was determined based on the patient’s age, body mass index (BMI) and serum anti-Mullerian hormone (AMH) level. During stimulation, patients were monitored regularly using ultrasound and serum hormone levels. Once follicles reached 17–19 mm, final oocyte maturation was triggered with human chorionic gonadotropin (hCG) (10,000 units), followed by oocyte retrieval 34–36 hours after hCG administration. The EMB samples were obtained following the hCG trigger, in either the peri-ovulatory phase (OS-PO/hCG+2; n=5) or the mid-secretory phase (OS-MS/hCG+9; n=4), using a disposable endometrial Pipelle device (MedGyn Pipette) under sterile conditions ( Figure 2 ). The ovarian stimulation protocols were managed by the treating physician while the EMB samples were collected by study physicians. Following collection, 100mg of endometrial tissue was preserved in RNAprotect (Qiagen, Frederick, MD, USA) for RNA extraction, and 100mg was fixed in formalin at room temperature. The formalin-fixed tissue samples were paraffin-embedded, sectioned and stained with hematoxylin and eosin ( Chemerinski et al. 2024 ). Histologic review was performed by two gynecological pathologists who were blinded to the study participant group. Endometrial dating was performed according to the histopathological criteria set forth by Noyes et al ( Noyes, Hertig, and Rock 1975 ). Subjects were excluded if samples had inconclusive endometrial dating or if histologic dating between the two pathologists differed by more than three days. Natural cycle subjects were excluded if histologic dating and phase of the menstrual cycle (based on the last menstrual period) were inconsistent ( Chemerinski et al. 2024 ). For subjects in a natural cycle, blood samples were collected at the first study visit and time of EMB. During an OS cycle, blood was collected throughout the cycle per treatment protocol as well as at the time of EMB. The blood (5mL) was collected in nonheparinized, serum separator tubes, allowed to clot, and centrifuged at 4°C. An electrochemiluminescence immunoassay (ECLIA) was then utilized to measure serum estradiol and progesterone (E2: Labcorp test: 004515 and P4: Labcorp test: 004317, Raritan, NJ, USA). Demographic data and reproductive hormone profiles of the study participants are presented in Table 1 and Supplemental Table 3 . There were no significant differences in median age, BMI, gravidity or parity between subjects in the proliferative, periovulatory and mid-secretory phases ( Table 1 ). Median serum E2 levels were similar in the proliferative, periovulatory and mid-secretory phases (PROLIF: 212.5 pg/mL, NC-PO: 175.0 pg/mL, and NC-MS: 156.0 pg/mL, p=0.9381, Table 1 , Figure 3 . Median serum P4 levels were significantly higher in the mid-secretory phase as compared to proliferative or periovulatory phases (NC-MS: 6.5 ng/mL vs PROLIF: 0.3 ng/mL, p=0.0173 and NC-MS: 6.5 ng/mL vs NC-PO: 0.5 ng/mL, p=0.0284 Table 1 , Figure 3 ). Demographic and clinical data for participants in the OS cohort are presented in Table 1 and Supplemental Table 4 . No differences were seen in age, BMI, gravidity, parity, and number of oocytes retrieved between subjects in the periovulatory (OS-PO) and mid-secretory (OS-MS) groups ( Table 1 ). Similarly, median serum E2 (OS-PO: 541.0 pg/mL vs OS-MS 800.0 pg/mL, p=0.4127) and P4 levels (OS-PO: 13.5 ng/mL vs OS-MS: 18.6 ng/mL, p=0.7302) were not significantly different ( Table 1 , Figure 3 ). Subjects undergoing ovarian stimulation were significantly older than those recruited in a natural cycle (35.1 vs 28.5 years, p=0.002). Within the subjects in the periovulatory cohort, there was no difference in median age between those in a NC and those undergoing OS (NC-PO 29.5 vs OS-PO 33.0 years, p=0.0561, Table 2 ). However, within the midsecretory phase, the median age was significantly lower in those in a NC compared to those undergoing OS (NC-MS 28.0 vs OS-MS 37.5 years, p=0.0221) ( Table 2 ). During the periovulatory period, there was no difference in BMI between those undergoing a NC and OS (NC-PO 29.4 vs OS-PO 24.4 kg/m 2 , p=0.0823). However, median gravidity and parity were significantly lower among the OS subjects compared to those in a NC (gravidity: NC-PO 2.5 vs OS-PO 0.0, p=0.0087 and parity: NC-PO 2.0 vs OS-PO 0.0, p=0.009) ( Table 2 ). Serum hormone levels were also compared. In the periovulatory phase, E2 and P4 levels were significantly higher with ovarian stimulation (E2: OS-PO: 800.0 vs NC-PO: 175.0 pg/mL, p=0.0043; P4: OS-PO: 13.5 vs NC-PO: 0.5 ng/mL, p=0.0043, Table 1 , Figure 4 ). In the mid-secretory phase, only E2 was significantly increased with ovarian stimulation (OS-MS: 541.0 vs NC-MS: 156.0 pg/mL, p=0.0317, Table 2 , Figure 4 ). Mid-secretory P4 levels, although slightly higher with ovarian stimulation, were not significantly different (OS-MS: 18.6 vs NC-MS: 6.5 ng/mL, p=0.5556, Table 2 , Figure 4 ). RNA was extracted from EMB tissues using RNeasy Plus Mini Kit (Qiagen, Frederick, MD, USA). Following extraction, the RNA concentration was determined using the NanoDrop Spectrophotometer (ThermoFisher Scientific, Waltham, MA, USA). 300 ng total RNA per sample was then sent to the Computational Genomics Core at Albert Einstein College of Medicine, Bronx, NYC, for next-generation mRNA sequencing. The raw transcriptome reads were assessed for quality control (FASTQC v 0.11.8) and trimmed for quality/adapter contamination (cutadapt v2.5). Trimmed reads were aligned to the human genome (GRCh37) using STAR (v2.6.1), followed by transcript abundance calculations and hit count extraction with StringTie (v2.0) and featureCounts (v1.6.4) respectively. Hit count normalization and differential gene expression group cross comparisons were performed using DESeq2 (v1.26.0). Significant differentially expressed gene thresholds were set to FDR adjusted p<0.05 ( Chemerinski et al. 2024 ). Total RNA (0.5 μg) was extracted using the RNeasy Mini Kit (Qiagen 74136, Frederick, MD, USA)) from EMB tissues and reverse transcribed using qScript cDNA Supermix (Quantabio, Beverly, MA, USA). Gene expression was evaluated using human gene-specific primers ( Supplemental Table 5 ). qPCR was conducted using the QuantiNova SYBR Green PCR kit (Qiagen, Frederick, MD, USA). All measurements were performed in triplicate and the gene expression level was determined as a mean of the triplicate. Relative gene expression was determined using the 2 −∆∆Ct method ( Livak and Schmittgen 2001 ) and expressed as fold change normalized to the human RNA18SN5 housekeeping gene. Animal studies were approved by the Institutional Animal Care and Use Committee at Rutgers New Jersey Medical School (PROTO2018000086) and were conducted in compliance with relevant regulations, as well as the ARRIVE guidelines 2.0 ( Percie du Sert et al. 2020 ). All data presented in this study were generated using C57BL/6J wild type mice (Jackson Laboratories). For studies involving non-pregnant mice, females were euthanized at 10 weeks of age. The uterus was removed and preserved in RNAlater (Qiagen, Frederick, MD, USA). For studies involving pregnant mice, females (8–10 weeks old) were mated with stud males, and the day of mating was designated as day 0, with day 0.5 being noon on the day a mating plug was observed. Mice were then euthanized on E0.5, E5.5, E6.5 and E7.5. On E0.5, the entire uterus was isolated and used for subsequent analysis, while on E5.5, E6.5 and E7.5, the myometrium was removed to isolate the underlying decidua. Whole uteri from E0.5, as well as decidua from E5.5, E6.5 and E7.5 were preserved in RNAlater (Qiagen, Frederick, MD USA). RT-PCR was performed on whole uteri from non-pregnant mice and decidua from pregnant dams at E7.5. RT-qPCR was performed on whole uteri from pregnant dams at E0.5 and decidua on E5.5 and 6.5. For both RT-PCR and qPCR, total RNA (0.5 μg) was extracted and reverse transcribed using qScript cDNA Supermix (Quantabio, Beverly, USA). Gene expression was evaluated using mouse gene-specific primers ( Supplemental Table 6 ). For RT-PCR, PCR products were visualized on an agarose gel after 30 cycles of amplification. qPCR was conducted using the QuantiNova SYBR Green PCR kit (Qiagen, Frederick, MD, USA) in the same manner described above for human ADGR expression. Relative gene expression was expressed as fold change normalized to the mouse Rn18s housekeeping gene. For all analyses a Shapiro-Wilk test was used to determine normality of the data set. Normally distributed data were described using means +/− standard deviations (SD). Parametric testing including an analysis of variance (ANOVA) and a Tukey’s post hoc analysis were used to compare three or more samples, while an unpaired t-test was used to compare two samples. For non-normally distributed data, values were expressed as medians with interquartile ranges (IQR) to assess the spread of the data. Nonparametric testing was performed using a Kruskal-Wallis test and Dunn’s post hoc analysis for three or more samples with data expressed in mean rank, while a Mann-Whitney U test was used for two samples. Statistical analyses were performed using Prism Version 9.0 (GraphPad). Statistical significance was defined as p<0.05.

GPCRs are major regulators of the hypothalamic-pituitary-ovarian axis and have crucial functions in female reproductive physiology and fertility ( Babwah 2015 ). Studies in mice have revealed critical roles for GPCRs in the regulation of early pregnancy at the level of endometrial receptivity, embryo implantation and decidualization ( Yoo et al. 2017 ). Less is known about the roles of GPCRs in regulating these events in human reproduction, but one in vitro study showed a potential role for ADGRG2 in regulating human endometrial stromal cell decidualization ( Yoo et al. 2017 ). ADGRs constitute the second largest GPCR subfamily in humans and are represented by at least 32 coding ADGR genes. Herein, we sought to characterize ADGR expression in decidualization and endometrial remodeling in women and mice. We performed an in-silico analysis of ADGR microarray expression using a preexisting database of endometrial samples from normo-ovulatory women across the menstrual cycle ( Talbi et al. 2006 ), and found that multiple ADGRs are differentially expressed between the proliferative and secretory phase, as well as within the secretory phase. We then sought to confirm the differential gene expression seen in the in-silico analysis using prospectively collected human secretory endometrium within natural cycles, and characterize ADGR expression with respect to decidualization in early mouse pregnancy. Given our finding that members of the ADGR gene family were dynamically expressed across the natural menstrual cycle and early mouse pregnancy, we then tested the hypothesis that endometrial ADGR gene expression could be perturbed by exposure to altered levels of E2 and P4 associated with ovarian stimulation ( Kalakota et al. 2022 ). The differential expression of ADGR observed between the proliferative and secretory phases of the natural menstrual cycle, along with changes in expression seen with OS, suggest that ADGR expression in the endometrium and decidua is regulated by sex hormones. Both E2 and P4 are also responsible for remodeling of the endometrium into a secretory phenotype following ovulation, a step which is essential for endometrial receptivity and embryo implantation. Thus, changes in ADGR expression suggest roles in endometrial development and receptivity necessary for embryo implantation. Our study identified novel ADGRs which have not been previously implicated in endometrial decidualization or female fertility. In our analysis of the endometrium, Both ADGRF1 and ADGRL4 were found to be significantly upregulated in the mid-secretory phase compared to the periovulatory period of ovulatory natural cycles, a trend seen both in the in-silico and prospective mRNA-Seq analyses. While no difference in expression was seen for ADGRL4 when comparing the proliferative to early or mid-secretory phases in the in-silico analysis, a significant increase in ADGRL4 expression was seen between the proliferative and late secretory phases. Ovarian stimulation does not appear to disrupt the expression of ADGRL4 as we found similar levels of expression in the PO and MS phases between OS and NC groups. However, ovarian stimulation does appear to alter the expression of ADGRF1, as a significantly lower level of expression was seen in the periovulatory period of the OS cycle compared to a NC. The expression of ADGRG2 has been investigated in both human and mouse decidualization, and ADGRG2 is one of 238 genes used in a commercially available endometrial receptivity array (ERA) that was developed with the intention of identifying a women’s “personal” window of implantation ( Diaz-Gimeno et al. 2011 ). In the ERA, ADGRG2 expression decreased 5.5-fold between the early to mid-secretory phase ( Diaz-Gimeno et al. 2011 ). In our study of ADGRG2 expression in the endometrium, no significant changes were seen across the menstrual cycle or with ovarian stimulation in both the in-silico analysis and prospectively collected endometrial samples. Discrepancies in gene expression seen between our investigation and previously published data by Diaz-Gimeno et al. 2011 , may be explained by differences in sample size and methods of evaluating gene expression. Our sample size of 21 women included in the in-silico analysis and 15 women recruited for prospective collection of endometrium in a natural cycle, was substantially smaller than the 95 endometrial samples utilized by Diaz-Gimeno et al. 2011 . Additionally, the previously mentioned study evaluated gene expression using microarray hybridization similar to the in-silico analysis data set, rather than RNA-sequencing which was used to analyze the prospectively collected endometrial samples. Nevertheless, previous in vitro studies demonstrate increased ADGRG2 expression with decidualization of human endometrial stromal cells (HESCs), and identified a functional role, given decidualization in HESCs was impaired after knockdown of ADGRG2 ( Yoo et al. 2017 ). In mice, Adgrg2 has also been identified as a key regulator of decidualization and previous studies have demonstrated significant increases in mRNA and protein expression from E0.5 to E7.5 ( Yoo et al. 2017 ). Our data also demonstrated expression of Adgrg2 in the mouse decidua at E7.5. A significant increase in expression was observed from E5.5 to E6.5, a time period of increasing decidualization. In addition to Adgrg2, Adgrd1 has also been implicated in mouse reproduction. However, Adgrd1 has not been identified as being involved with decidualization, but rather plays a role in embryo transit from the oviduct ( Bianchi et al. 2021 ). In our investigation utilizing mouse models, no significant changes in expression were identified, suggesting little involvement in early mouse pregnancy or decidualization. When investigating differentially expressed ADGRs across the menstrual cycle and the context of early mouse pregnancy, Adgrl4 demonstrated a 17-fold increase in expression from E0.5 to E6.5. The higher expression seen with increased decidualization reflects the higher expression seen in the MS phase compared to the ESE/PO phase in both the in-silico and mRNA-seq analyses. Conversely, ADGRL1 was found to be significantly decreased in expression in the MS phase when compared to the PROLIF phase in the in-silico analysis. However, in mouse studies, Adgrl1 demonstrated a 25-fold increase in expression from E0.5 to E6.5. No significant changes in ADGRL1 expression were identified in the mRN-seq analysis. Other differentially expressed genes identified included ADGRB2 and ADGRF1 , however neither were expressed in the pregnant mouse uterus. Whereas the involvement of ADGRs in non-gynecologic cancers has been well described (Ping Lei 2022 ), members of the ADGR family have more recently been implicated in the pathogenesis of endometrial cancer ( Ahn et al. 2019 ). Genes including ADGRD1, ADGRF1 and ADGRL4, have been associated with cancers including glioblastoma, osteosarcoma, as well as retinoblastoma, respectively (Ping Lei 2022 ). Emerging data examining ADGR expression in endometrial cancer tissues demonstrate significantly elevated expression of several ADGR families, with ADGRC3 and ADGRF1 found to also have a significant correlation with pathological staging (Ping Lei 2022 ). ADGRG2 has also been implicated in endometrial cancer with findings suggestive of its role as a tumor suppressor ( Ahn et al. 2019 ). In vitro studies demonstrated siRNA mediated downregulation of ADGRG2 allows for increased cell proliferation, migration, and invasion of human endometrial cancer cells ( Ahn et al. 2019 ). Although G protein coupled receptors are ubiquitous proteins involved in almost every biological process in humans, there remains much to be discovered regarding their role in the endometrium. Emerging data, including our own, suggests key roles for ADGRs in both physiologic and pathological processes in this tissue. For this reason, it is imperative to better understand their role, especially in endometrial receptivity and embryo implantation. Strengths of our study include multiple methods of evaluating gene expression including microarray data from the in-silico analysis, mRNA sequencing, RT-PCR and RT-qPCR. Additionally, our investigation was able to characterize ADGR expression changes that were conserved across decidualization in the menstrual cycle and in early mouse pregnancy. Limitations of our study include the small sample size and a lack of functional studies to investigate the role of ADGRs in decidualization and endometrial remodelling. Our investigation reveals novel ADGRs that demonstrate dynamic expression in both human and mouse decidualization. Both ADGRL1 and ADGL4 were differentially expressed in the natural menstrual cycle and early mouse pregnancy. Future investigations will be targeted towards in vitro studies and mouse models to determine the impact on decidualization when expression of ADGRL1 and ADGL4 is altered. These ADGRs may serve as novel targets for functional studies in in-vitro and mouse models, to assess how impaired expression may impact decidualization. Investigating the role of ADGRs in female reproduction could advance the field of assisted reproduction by possibly uncovering new genes that are involved in endometrial receptivity, thereby offering targets for novel prophylactic or corrective interventions to improve impaired endometrial receptivity following IVF.

Approximately 10% of reproductive-aged women in the United States suffer from infertility and a large number of them will utilize in vitro fertilization (IVF) to conceive ( Centers for Disease Control and Prevention 2018 ). The use of IVF has become increasingly popular, with the number of annual cycles doubling in the last decade, from 150,000 cycles per year in 2010 to more than 300,000 cycles per year in 2019 ( Centers for Disease Control and Prevention 2021 ). Despite this increase, the live birth rate (LBR) after IVF has only increased by approximately 5%, from 32% in 2010, to 37.2% in 2019 ( Centers for Disease Control and Prevention 2021 ). Successful embryo implantation depends on the coordinated development of the embryo and endometrium, so that when the implantation-competent embryo arrives at the endometrium, the endometrium is receptive to implantation. The period of opportunity to implant (called the window of implantation) lasts only 30–36 hours and occurs 6 to 9 days following the surge of luteinizing hormone (LH) in a natural cycle, or 4 to 7 days following progesterone administration in a hormone replacement cycle ( Simon et al. 2020 ). Under the tight control of estradiol (E2) and progesterone (P4), the endometrium develops a secretory phenotype shortly after ovulation. Through the actions of P4 and cAMP in the mid-secretory phase, stromal cells decidualize and remodeling of the endometrium gives rise to the “predecidua” that regulates embryo implantation ( Gellersen and Brosens 2014 ). Following implantation, the decidua that is associated with pregnancy regulates trophoblast invasion and development of the placenta. Consequently, dysregulated decidualization compromises placenta formation and/or function leading to late sporadic miscarriage, normotensive fetal growth restriction, preeclampsia, placental abruption and preterm labor ( Conrad 2020 ). Ovarian stimulation, utilized for the development of multiple ovarian follicles for IVF, induces supraphysiologic levels of E2 and an early rise in P4 ( Kalakota et al. 2022 ). These hormonal changes induce morphological, biochemical, and functional genomic modifications that can disrupt endometrial receptivity ( Kalakota et al. 2022 ). Thus, with fresh embryo transfers, when the embryo is returned to the uterine cavity following ovarian stimulation, it may encounter a non-receptive endometrium, resulting in implantation failure. Alternatively, the embryo may encounter an endometrium that is receptive with impaired decidualization, resulting in successful embryo implantation, but affected by the “ripple effect” leading to placental insufficiency and pregnancy complications ( Rabaglino and Conrad 2019 ; Conrad 2020 ). To improve pregnancy rates and reduce the risk of pregnancy complications associated with IVF, it is crucial to advance our molecular understanding of endometrial receptivity and decidualization. Previous studies, including those from our laboratory, have shown that in the mouse, uterine G protein-coupled receptors (GPCRs) are critical regulators of endometrial receptivity and decidualization ( Schaefer et al. 2021 ; de Oliveira et al. 2019 ; Radovick and Babwah 2019 ; Babwah 2015 ; Fayazi et al. 2015 ; Parobchak et al. 2020 ; Calder et al. 2014 ; Mohri et al. 2010 ; Sone et al. 2013 ; Kida et al. 2014 ; Diao et al. 2015 ; Verma and Arora 1990 ; Leon et al. 2016 ). GPCRs are transmembrane proteins that regulate almost every cellular and physiologic process, including responses to gonadotropin releasing hormone (GnRH), follicle stimulating hormone (FSH) and LH ( Babwah 2015 ). Given their ubiquitous nature, these receptors are targets of approximately 35% of all drugs approved by the United States Food and Drug Administration ( Babwah 2015 ). In women, GPCRs might also regulate the acquisition of endometrial receptivity; this is based on their differential expression between the pre-receptive and receptive period in the secretory endometrium ( Diaz-Gimeno et al. 2011 ). Members of the adhesion G protein-coupled receptor (ADGR) family, the second largest GPCR subfamily with at least 32 members distributed over nine subfamilies ( Hamann et al. 2015 ), are involved in both mouse and human reproduction. In mice, ADGRD1 is needed for embryo transit from the oviduct, while ADGRG2 (GPR64) is involved in fluid reabsorption and sperm maturation in the mouse epididymis ( Bianchi et al. 2021 ). In women, the development of a receptive endometrium is associated with the reduced expression of endometrial ADGRG2 ( Diaz-Gimeno et al. 2011 ; Yoo et al. 2017 ). Conversely, in isolated human endometrial stromal cells (HESCs), ADGRG2 gene expression is reported to be required for HESC decidualization in vitro ( Yoo et al. 2017 ). In mice, the ADGRG2 protein is expressed in the developing decidua, and development of the receptive endometrium is associated with the increased expression of both ADGRG2 mRNA and protein ( Yoo et al. 2017 ). These findings strongly suggest that in the human endometrium, ADGRG2 is an E2- and/or P4-regulated gene, and this is further supported by mouse data that demonstrated Adgrg2 mRNA expression is P4-dependent ( Yoo et al. 2017 ). Given the potential importance of ADGRs in decidualization, endometrial remodeling, and embryo implantation, we sought to characterize their expression in the human endometrium and in early mouse pregnancy. We determined ADGR expression across the menstrual cycle via in silico analysis of a publicly available endometrial gene expression dataset, generated from a retrospective cohort of ovulatory women ( Talbi et al. 2006 ). We found that many ADGRs are differentially expressed between the proliferative and secretory phases. We sought to confirm these findings, and thus analyzed ADGR expression in prospectively collected endometrial biopsy samples within natural ovulatory menstrual cycles, as well as in animal studies using non-pregnant and pregnant mice. Based on our observations from the in-silico analysis, as well as prospective studies in humans and mice, we hypothesized that ADGR expression is likely hormone mediated. To determine if altered E2 and P4 expression would disrupt endometrial ADGR gene expression, we then prospectively collected endometrial biopsy samples from women undergoing ovarian stimulation. Our findings demonstrate that endometrial ADGRs are dynamically expressed within the menstrual cycle and early mouse pregnancy, and their gene expression is altered by ovarian stimulation.