Endometrial Metabolism of 17β-Estradiol during the Window of Implantation in Women with Recurrent Implantation Failure.

The Medical Ethics Committee of the Maastricht University Medical Centre+ (MUMC+) approved the study (registration number Dutch Trial Register NL7571, www.onderzoekmetmensen.nl ), and the study protocol was published previously [ 14 ]. Written informed consent was obtained from all participating subjects in accordance with the Helsinki Declaration. This is an observational study, part of the Mu ltidisciplinary R esearch on Repeated I mplantation Failure and Recurrent M iscarriages (MURIM) study. Two study cohorts were compared: (1) patients with RIF and (2) a control group consisting of female participants visiting the IVF clinic without infertility or any suspected endometrial contribution to infertility. RIF was defined as IVF patients aged 18–38 years with implantation failure after three transfers of high-quality embryos or after transferring ten or more embryos across multiple transfers. High-quality embryos were defined using Gardner criteria [ 15 ]. High quality was defined as (i) cleavage-stage embryos consisting of 4 cells (2 days post-fertilization) or ≥7 cells (3 days following fertilization) both with less than 20% fragmentation and the absence of multinucleation, (ii) 4 days post-fertilization: morula grade A or blastocyst grade B1 or B2, (iii) 5 or 6 days post-fertilization: blastocyst grade B3 or higher combined with inner cell mass quality A or B and trophectoderm epithelium quality A or B. Exclusion criteria for RIF were BMI >35 kg/m 2 ; clinically relevant intrauterine pathology; untreated endocrine abnormalities; or patients undergoing pre-implantation genetic testing (PGT) for monogenic disorders (PGT-M) or structural rearrangements (PGT-SR). PGT for aneuploidy (PGT-A) is not performed in the clinic because of national legislation and is therefore neither an inclusion nor exclusion criterion. The control group consisted of women aged 18 and 38 years who were referred to the Department of Obstetrics and Gynaecology of the MUMC+ for IVF or ICSI treatment because of a severe male factor (semen volume x concentration x good motility [VCM] < 1), bilateral tubal factor, or for PGT-M or PGT-SR without concurrent infertility. Exclusion criteria were a history of RIF or recurrent miscarriages, current or recent (35 kg/m 2 , and moderate or severe endometriosis (3rd–4th stage, according to the American Society of Reproductive Medicine) [ 14 ]. Patients with RIF ( n = 52) and controls ( n = 25) were included at the outpatient clinic of the Maastricht University Medical Centre+ between April 2019 and November 2022. The visit took place during the WOI, 5–8 days after a positive ovulation test. The ovulation tests, based on the luteinizing hormone (LH) surge in urine, were performed twice a day from cycle day 10 onward. Briefly, the visit included a questionnaire, BMI measurement, vaginal ultrasound including a measurement of the endometrial thickness (mm), endometrial biopsy, and collection of peripheral blood. The cause and type of infertility (primary [i.e., patient has never conceived before] or secondary) were determined. All included patients of whom enough endometrial tissue was available were included. RIF patients were referred from multiple centers in the Netherlands. The control group was recruited in the MUMC+ only, and subjects were asked to volunteer in this study, which included the collection of an endometrial biopsy. The sample size of this control group was lower than that of the RIF group. Participants of the MURIM study (both RIF patients and controls) had a 1-year follow-up as part of the study protocol, to determine long-term reproductive outcomes after treatment continuation. RIF patients with no ongoing pregnancy after 1-year of follow-up were classified as having “persistent” RIF ( n = 18). Patients from the control group who had an ongoing pregnancy during the follow-up period were considered as proven fertile controls ( n = 13). Endometrial tissue was obtained 5–8 days after a positive urine ovulation test. Biopsies were sampled using a Pipelle catheter (CCD International, France) and stored at −80°C. At the same day of the endometrial biopsy, serum was isolated from peripheral blood (10 mL), which was left to clot for 30–60 min and centrifuged (2,000 g for 10 min) and stored at −80°C. Enzyme activity of reductive and oxidative 17β-HSDs was determined in endometrial tissue extracts using HPLC as previously described [ 13 , 16 ]. Briefly, endometrial tissue was homogenized in RIPA buffer (50 m m Tris-HCl, pH 7.4, 1% NP-40, 0.25% sodiumdeoxycholate, 150 m m NaCl, 1 m m EDTA). Protein concentration was determined by a BC Assay Protein Quantitation Kit (Uptima-Interchim, Montlucon, France). The conversion of estrone to 17β-estradiol was carried out as follows: a 500 µL reaction mixture containing 3.2 m m NADP+, 9 m m glucose-6-phosphate, 1 U glucose-6-phosphate dehydrogenase, 3.3 m m magnesium chloride, and 10 nmol estrone in 50 m m potassium phosphate buffer (pH 7.4) was incubated with 24–232 µg of the protein lysate at 37°C for 16 h. The conversion of 17β-estradiol into estrone was measured in a 500 µL reaction mixture containing 5 m m NADP + , 10 nmol 17β-estradiol in 50 m m potassium phosphate buffer (pH 7.4), and 24–232 µg of the protein lysate incubated at 37°C for 16 h. HPLC (Shimadzu LC-10AD setup; Kyoto, Japan) was used to determine estrogen formation. After adding the internal standard butyl-4-hydroxybenzoate (300 ng), 2 mL of H 2 O, and 2.5 mL of chloroform to the reaction mixture, the estrogens were extracted in the organic phase. The chloroform phase was evaporated under nitrogen at 45°C and estrogens were derivatized with 250 μL 2-(4-carboxy-phenyl)-5,6-dimethylbenzimidazole (synthesized as part of previous work [ 13 ]) and 250 μL 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride solution at 50°C for 20 min. The reaction mix was diluted in acetonitrile and 10 μL was injected in the HPLC. LabSolutions HPLC (Shimadzu, Kyoto, Japan) was used for instrument control, data acquisition, and calculation of peak areas. The derivatized steroids were injected into a LiChroCART 250-4 RP 18 column (Merck, Darmstadt, Germany) and eluted with a gradient of methanol/H 2 O (80/20; v/v) at time 0 to methanol/water (90/10; v/v) at time 20 min. Reactions were carried out with or without the presence of specific 17β-HSD inhibitors. First, the 3-[15b-estronyl]-N-(5-methyl-thiazol-2-yl)-propionamide, also described as compound 21 [ 17 ], is a 17β-HSD1 inhibitor. The inhibitor has an estrone substitute at the 15th carbon position and has previously being characterized as a potent and specific inhibitor of 17β-HSD1 with little affinity toward other 17β-HSDs and estrogen receptor α (ERα) [ 7 , 18 ]. Second, the 1-{1-([5-methoxy-1H-indol-2-yl] carbonyl) piperidin-4-yl}c-2-methylpropan-2-ol, also known as ASP9521, inhibits 17β-HSD5 (AKR1C3) [ 19 ]. Compounds have been previously evaluated for their ability to inhibit the reductive transformation of [ 14 C]-4-androstene-3,17-dione into [ 14 C]-testosterone by 17β-HSD5 transfected in HEK-293 cells in culture, as previously reported [ 20 ]. Third, the inhibitor 17β-HSD7, a 4-methyl-4-aza-5R-androstane derivative, inhibits reductive transformation E1 into E2 by 17β-HSD7 [ 20 , 21 ] and lastly INH-12, also described as compound 97, is an inhibitor of 17β-HSD12 [ 22 ]. The inhibition of the 17β-HSD1 enzyme was computed as percentage of inhibition (or inhibition of estrone to 17β-estradiol conversion) by the formula: 100% − ( formation of 17β-estradiol with inhibitor/formation of 17β-estradiol without inhibitor × 100% ). Inhibitor of 17β-HSD5, 7, and 12 were analyzed separately and in combination with the inhibitor of 17β-HSD1 (5 μm). Tissue biopsies were fixed in 3.7% formaldehyde, embedded in paraffin, and processed for histological examination. For immunostaining of 17β-HSD7, paraffin sections of 5 μm were deparaffinized, 2 × 5 min in xylene and 2 × 5 min in 100% ethanol. The endogenous peroxidases were blocked by incubating 20 min in 0.3% H 2 O 2 in methanol. For heat-induced epitope retrieval, sections were boiled in citrate buffer (pH 6.0) in a microwave for 20 min. After blocking (1% BSA/PBST) for 1 h at room temperature, staining with a rabbit polyclonal antibody 17β-HSD7 (1:50; Invitrogen ™ ) was performed overnight at 4°C. To visualize the antibody, Ultra-Sensitive ABC Peroxidase Rabbit IgG Staining Kit (Thermo Scientific ™ ) and 3,3-diaminobenzidine solution (Dako, Glostrup, Denmark) were used. The AllRed Scoring System was used for 17β-HSD7 evaluation [ 23 ]. This score combines the percentage of positive cells (no staining = 0, <1% = 1, 1–10% = 2, 11–33% = 3, 34–66% = 4, 67–100% = 5) and the intensity of the reaction product (no staining = 0, weak = 1, moderate = 2, strong = 3) to determine a final score (the staining index [SI]). The SI was determined independently by two researchers and the average SI between the RIF ( n = 19) and control ( n = 18) group was compared using the Mann-Whitney U test. Endometrial tissue from a study participant and a paraffin section of colon tissue were used, respectively, as negative and positive control for 17β-HSD7 staining. The Shapiro-Wilk test was used to test for normality. Data that were not normally distributed were log10F transformed. A Student’s t test was used for continuous data if normally distributed or Wilcoxon-Mann-Whitney U test if non-normally distributed or in case of n < 30. Categorical data were compared using the Fisher’s exact test. Multivariate logistic regression analysis was used to assess the influence of age at inclusion and confounder-adjusted p values are reported. Statistical significance was defined as a p value <0.05. Missing data are reported and not imputed. The Statistical Package for the Social Sciences (SPSS v 28 for Windows, Chicago, IL, USA) was used for data analysis. Graphs were created with R (v4.2.2; R Foundation for Statistical Computing, Vienna, Austria) using ggplot2 v. 3.5.1.

RIF patients ( n = 52) were 33.7 ± 2.8 years old with an infertility duration of 36.0 ± 19.9 months. On average, 7.0 ± 2.6 embryos had been transferred upon inclusion (range: 3–14). Controls ( n = 25) were 30.4 ± 3.3 years old, which were significantly younger ( p < 0.001). In concordance with the study design, control women had significantly shorter durations of infertility ( p < 0.001) and had no previous embryo transfers. There were no differences in BMI, number of smokers, number of patients with primary infertility, and endometrial thickness between groups ( Table 1 ). Characteristics of RIF ( n = 52) and control ( n = 25) participants Data are presented as mean ± standard deviation or number (%). a By Mann-Whitney U test. b By chi-square test. c Endometrial thickness was measured during the study visit, 5–8 days after the LH surge. Median reductive 17β-HSD activity did not differ between RIF and control women (1,074.5 [IQR = 789.8–1,519.8] vs. 1,259 [749.5–1,854.5] pmol 17β-estradiol/hr/mg protein, p = 0.31) (shown in Fig. 2 a). Oxidative 17β-HSD activity was log10 transformed to allow for parametric testing. Oxidative 17β-HSD activity resulted non-significantly different between RIF patients and controls (786.5 [509–1,193.5] vs. 1,012 [647–1,532] pmol estrone/hr/mg protein, p = 0.21) (shown in Fig. 2 b). The ratio of reductive 17β-HSD to oxidative 17β-HSD enzyme activity was similar between RIF patients and controls (1.27 vs. 1.20, p = 0.22) (shown in Fig. 2 c). Subgroup analysis of women with primary infertility and unknown infertility cause gave similar results. Activity of reductive ( a ) and oxidative ( b ) 17β-HSD enzymes and activity ratio ( c ) of the reductive and oxidative 17β-HSD enzymes. The conversion of estrone to 17β-estradiol ( a ), conversion of 17β-estradiol to estrone ( b ), and the activity ratio of reductive and oxidative 17β-HSDs ( c ) were similar in the RIF ( n = 52) and control ( n = 25) group. Box plots show the 25th and 75th percentile within the median, and whiskers indicate 1.5 times the interquartile range. As part of the MURIM study, participants had a 1-year follow-up after the study visit. A subgroup analysis was performed to compare RIF patients without pregnancy at the end of the follow-up period (“persistent” RIF, n = 18) with control participants who had an ongoing pregnancy (proven fertile controls, n = 13). No differences were found in oxidative 17β-HSD activity, reductive 17β-HSD activity, and ratio (shown in online suppl. Table 2). Mean percentage of inhibition of 17β-HSD1 was not significantly different in RIF patients ( n = 52) versus controls ( n = 25) (30.6% ± 8.5% vs. 27.7 ± 12.4%, p = 0.30) (shown in Fig. 3 ). The combined inhibition of 17β-HSD5, 17β-HSD7, and 17β-HSD12 was determined by subtracting inhibition of 17β-HSD1 from the maximum inhibition of all reductive 17β-HSD enzymes in the two groups (shown in online suppl. Table 3). Inhibition of 17β-HSD5, 7, and 12 was significantly lower in the RIF group ( n = 52) compared to controls ( n = 25) (20.4% ± 9.5% vs. 27.1% ± 14.1%, p = 0.04) ( Fig. 3 ). After adjusting for age as a possible confounder, this difference remained significant (adjusted p = 0.01). Inhibition of the 17β-HSD1 and 17β-HSD5, 7, and 12 enzyme activity (%) in the control ( n = 25) and RIF ( n = 52) group. The difference in inhibition of 17β-HSD5, 7, and 12 between the control ( n = 25) and RIF ( n = 52) group was significant. * p < 0.05. To understand which enzyme might contribute to this observation, we performed explorative pilot experiments testing the contribution of each one of the enzymes 17β-HSD5, 17β-HSD7 (RIF, n = 7, controls, n = 7), and 17β-HSD12 (RIF, n = 2, controls, n = 2) in combination with 17β-HSD1. Inhibition of 17β-HSD1+5 and 17β-HSD1+12 was not different between groups, but inhibition of 17β-HSD1+7 was lower in the RIF group versus controls (26.9 ± 15.7 vs. 47.4% ± 15.1, p = 0.03) (online suppl. Table 4). Therefore, we performed the analyses of the 17β-HSD1+7 combination in a larger group (RIF, n = 31 versus controls, n = 24) and confirmed lower inhibition of 17β-HSD1+7 in the RIF group versus controls (inhibition of 17β-HSD1+7 = 26.4% ± 10.8 vs. 42.6% ± 21.7, p = 0.002 [by Mann-Whitney U test]). This difference remained significant after age correction (adjusted p = 0.001) ( Fig. 4 ). Inhibition of the 17β-HSD1 and 7 in the RIF ( n = 31) versus control ( n = 24) group. The difference in inhibition was significant. * p < 0.05. When inhibition of 17β-HSD1, 5, and 7 was analyzed separately, no statistically significant differences were found between RIF ( n = 5) and controls ( n = 5), although 17β-HSD7 inhibition was nearly twofold lower in the RIF’s group endometrium compared to control subjects (15.3% ± 15.3 vs. 27.0% ± 16.4, p = 0.55) (online suppl. Table 5), indicating that in this group this enzyme might be less involved in the reductive reaction since adding this inhibitor did not decrease 17β-estradiol formation as substantially as in the control group. Next, we tested the presence of the 17β-HSD7 in the human endometrium. Immunohistochemistry on paraffin sections showed that 17β-HSD7 had both a cytoplasmic and nuclear localization, and across samples a higher staining intensity was observed in the cytoplasm compared to the staining intensity in the nucleus. Furthermore, staining was generally stronger in the glandular epithelium and luminal epithelium compared to the stromal compartment, although measurements were semi-qualitative ( Fig. 5 ). The positive and negative control are shown in online supplementary Figure 1. The AllRed SI was used to compare 17β-HSD7 staining in the endometrium (secretory phase) of RIF ( n = 19) and control ( n = 18) samples. The SI did not differ between RIF and controls ( p = 0.58). Representative example of microscopic immunohistochemical 17β-HSD7 staining in RIF patients and control subjects. Upper left: control subject, ×40 magnification. Upper right: RIF patient, ×40 magnification. Lower left: control subject, ×100 magnification. Lower right: RIF patient, ×100 magnification. Scale bar is given for reference. Glands, stroma, lumen, luminal epithelium, and glandular epithelium are indicated.

The inhibition of reductive 17β-HSD enzymes (reflected by the formation of 17β-estradiol in the presence of specific 17β-HSD enzyme inhibitors) was altered between two well-defined groups of RIF patients and control subjects. Immunohistochemistry confirmed the presence of 17β-HSD7 in the endometrium, but the SI did not differ between the study groups. A possible role for 17β-HSD7 in human embryo implantation has not been previously described. The observed differences warrant future research on these 17β-HSD enzymes as modulators of the WOI, to improve our understanding of endometrial 17β-estradiol metabolism in achieving endometrial receptivity.

This study aimed to assess the role of the local estrogen metabolism in endometrial receptivity. The activity and ratio of reductive and oxidative 17β-HSD enzymes were not significantly different between RIF and control subjects. Combined inhibition of the three reductive 17β-HSD enzymes 17β-HSD5, 7, and 12 was lower in the RIF group compared with controls. A lower combined inhibition of 17β-HSD1+7 in the RIF group might suggest a novel, context-dependent, or modulatory role for the lesser-known 17β-HSD7 in endometrial steroid metabolism. However, no differences were found between groups after separate analysis of 17β-HSD7 inhibition. A potential explanation is that 17β-HSD7 activity varies due to differential cofactor availability or local interactions within the endometrial steroidogenic network. Additionally, the absence of a significant effect when targeting 17β-HSD7 individually may reflect limited statistical power in the absence of simultaneous 17β-HSD1 inhibition, particularly if 17β-HSD1 plays a more dominant role in this context. Also, immunostaining revealed the presence of 17β-HSD7 in the endometrium of all samples, but the SI of 17β-HSD7, although semiquantitative of nature, was not different between groups. Taken together, these findings suggest that 17β-HSD7 may still contribute to the fine-tuning of endometrial steroid homeostasis, but further investigation with a larger sample size and the addition of additional experiments at RNA level are necessary. Expression of 17β-HSD1 and 7 is detected in the endometrium at the mRNA and/or protein level. Although their function is important for the final androgenic/estrogenic balance in the tissue, little is known about the activity of these enzymes in relation to endometrial receptivity. In human ectopic pregnancies, 17β-HSD1 showed the highest immunoreactivity at the fetal-maternal interface, suggestive for a role in implantation [ 24 ]. Additionally, SNPs of 17β-HSD1 are associated with infertility [ 25 ]. A recent study found that inhibition of 17β-HSD1 was lower in a primary infertile subgroup of pregnant IVF patients compared to non-pregnant IVF patients. Included patients received an embryo transfer after a previous failed first IVF attempt. Although other reductive 17β-HSD enzymes were not determined, this outcome led to the speculation that enzymes other than 17β-HSD1 might be important in achieving endometrial receptivity [ 13 ]. In our study, we could not reproduce the lower inhibition of 17β-HSD1 in the presumed receptive endometria of the control group. However, in line with the previous study, the presumed receptive endometria of the control group showed higher inhibition of 17β-HSD5, 7, and 12, indicating that these enzymes might play a larger role in this population. The enzyme 17β-HSD7 has been associated with implantation in rodent models. In mice, the expression of HSD17B7 (encoding 17β-HSD7) varies spatiotemporally, starting around the implantation site and later appearing in the decidua and spongiotrophoblasts of the placenta [ 26 ]. Heterozygous HSD17B7 +/− mice were expected to show impaired fertility but cycled normally and were fertile. As HSD17B7 −/− offspring was absent, prevention of implantation by absence of 17β-HSD7 in these knock-out mice was hypothesized. However, in pregnant mice, HSD17B7 −/− embryos were recovered, indicating that the absence of 17β-HSD7 activity affects fetal survival rather than implantation [ 27 ]. This was later confirmed, indicating that 17β-HSD7 is essential for fetal development [ 28 ]. However, since endometrial receptivity and implantation as well as the substrate specificity of 17β-HSD enzymes differ greatly between species, results should be interpreted with great caution. In our study, all endometrial samples were taken in the secretory phase; therefore, we cannot determine if there are menstrual cycle-related changes in activity of 17β-HSD7. Gene expression analysis revealed that HSD17B7 was upregulated in patients with endometriosis compared to controls [ 29 ]. In IVF patients, while HSD17B1 and HSD17B7 were expressed in the endometrium, no differences in their expression levels were found between patients and controls, unlike the findings from our analysis. Likewise, the differences in inhibition observed in this study did not correspond with the immunohistochemistry results. It is possible that post-translational modifications of the enzymes, such as phosphorylation, may explain the differences at the mRNA, protein, and activity levels.

RIF patients included in this study were significantly older than the presumed fertile controls. Although it is known that aging influences the expression of genes involved in endometrial epithelial growth and endometrial receptivity [ 30 ], differences in inhibition of the 17β-HSD enzymes remained statistically significant after correction for age. Since pre-implantation genetic testing for aneuploidy (PGT-A) is not performed in our center, the study included RIF patients whose embryos were not tested for aneuploidy. Therefore, some RIF subjects may have had embryo-related factor rather than endometrial receptivity problems. This might have contributed to less pronounced differences between the two study groups. Although including RIF patients with failed transfers of euploid embryos could have been insightful in this case, we did not pursue this approach because (1) the Dutch legislative framework restricts PGT-A and (2) it is not recommended in the good clinical practice guidelines for RIF as it does not increase cumulative pregnancy rates per started IVF cycle [ 31 ]. Intra-individual, cycle-related variation in 17β-HSD enzyme activity cannot be ruled out, as the invasive endometrial samples were only taken once per participant, in a natural cycle during which the embryo transfer was not performed. Additionally, the study design describes a simplified cell-free situation (in vitro), and extrapolation to the in vivo situation should be done with caution. High concentrations of substrate are used to activate the enzymes and determine their metabolization capacity. Therefore, the biological significance of these enzymes may be overestimated, particularly given the uncertainty about whether the17β-HSD 5, 7, and 12 enzymes are already active at low estrone concentrations, as is the case for 17β-HSD1 [ 32 ]. While our study provides important insights at the enzyme activity and inhibition level, the absence of additional experiments at the RNA level limit further interpretation of gene regulatory dynamics and potential functional implications. In the future, comprehensive omics analysis (such as genomics, transcriptomics, and proteomics) of 17β-estradiol metabolism during the WOI is warranted to obtain a better understanding of the relevant processes for implantation to occur. To do so, three-dimensional in vitro endometrial structures (endometrial organoids) could be helpful as they recapitulate the original tissue while being expandable and remaining phenotypically and genetically stable [ 33 ]. These endometrial organoid cultures can be combined with a model of the human blastocyst (blastoids) to perform functional testing of implantation under different circumstances [ 34 ]. Mimicking the adhesion and implantation of blastoids in the peri-implantation endometrial organoid model can be a valuable tool to determine which phenotype of 17β-estradiol metabolism in RIF patients and controls contributes to implantation success or failure.

Medically assisted reproduction has advanced over the years as treatment protocols and laboratory techniques have improved. However, embryo implantation remains a bottleneck in in vitro fertilization (IVF) success, even when transferring morphologically high-quality embryos. In approximately 10–15% of IVF patients, viable embryos fail to implant repeatedly, a phenomenon termed recurrent implantation failure (RIF). The relative contribution of endometrial versus embryonic factors remains an important topic of debate in these patients. Chromosomal abnormalities are often found in embryos [ 1 ], and implantation rates improve drastically when only euploid embryos are transferred [ 2 ]. However, implantation failure can still occur after transfer of high-quality/euploid embryos, in which case endometrial receptivity might be impaired [ 3 ]. Sex steroids play a critical role prior and during implantation. In particular, 17β-estradiol and progesterone are essential for establishing the “window of implantation” (WOI), the period when the endometrium is receptive for an embryo. Deviations in the levels of these steroids may cause shifts or shortening of the WOI [ 4 , 5 ], potentially leading to implantation failure. Although estrogen and progesterone are primarily produced in the ovaries and secreted in the blood, the endometrium is not merely passively exposed to these circulating steroids. Through the local expression of multiple steroid-metabolizing enzymes, it can regulate the availability of active estrogens and other sex steroids, a process known as intracrinology [ 6 ]. Disruptions in the local estrogen metabolism can lead to an imbalance in endometrial 17β-estradiol levels, which has been linked to endometriosis, endometrial cancer, and infertility [ 7 – 9 ]. In the context of RIF, altered endometrial estrogen metabolism may impair the local hormonal milieu, potentially leading to an inadequate or mistimed WOI. Endometrial intracrinology and its relation to endometrial receptivity has been addressed in two recent investigations [ 10 , 11 ]. In one study, endometrial receptivity was determined based on the outcome of the endometrial receptivity array (ERA) test. This test, based on gene expression profile from an endometrial biopsy, predicts if the WOI is displaced. In the group of patients with high endometrial progesterone levels, a receptive endometrium was more frequently observed compared to patients whose progesterone levels were below the relative maximum progesterone cutoff level. Low 17OH-progesterone levels were associated with a lower percentage of a receptive endometrium [ 10 ]. No link was found between endometrial estrogen levels and receptivity. In another study, receptivity was determined by embryo implantation success in an IVF procedure that followed the spontaneous cycle, in which the biopsy was taken. It was found that lower serum estrone concentrations and lower serum estrone to androstenedione ratios were positively associated with receptivity. However, this association was only found in a subgroup of patients with primary infertility and not in the complete patient cohort [ 11 ]. Although these studies are pioneering, attempts to chart and decipher the complete intracrine network is particularly challenging due to its complexity and dynamicity where steroids levels fluctuate constantly. Hence, it is also important to dissect the role and contribution of individual steps in the steroid metabolism since differences can occur at multiple levels: in the tissue’s steroid hormone levels, in the activity of the enzymes that metabolize the steroid, and in enzyme expression (mRNA) as measured by RNA sequencing. Two important groups of steroidogenic enzymes are the reductive 17β-hydroxysteroid dehydrogenases (17β-HSDs), which catalyze the reduction of 17-ketosteroids (such as estrone), and the oxidative 17β-HSDs, which catalyze the oxidation of 17β-hydroxysteroids (such as 17β-estradiol), thereby regulating the final estrogen activity in the endometrium ( Fig. 1 ). While 17β-HSD1 is the major reductive 17β-HSD enzyme and 17β-HSD2 is the major oxidative 17β-HSD enzyme, 15 human 17β-HSDs exist and multiple enzymes can use estrogens as substrate [ 6 , 12 ]. Aberrant oxidative and reductive activity in the endometrium could directly impair implantation as estrogen imbalances have been associated with reduced receptivity. In IVF patients, supra-physiological estrogen levels can induce advanced endometrial maturation such as the early appearances of pinopodes, leading to endometrial-blastocyst asynchrony [ 4 ]. Furthermore, inactivation of 17β-estradiol by oxidative 17β-HSD enzymes is one of the progesterone-dependent uterine anti-estrogenic effects in the luteal phase. Therefore, oxidative and reductive activity of 17β-HSDs can also indirectly reflect alterations in the intracrine network [ 6 ]. Schematic simplified representation of the conversion of estrone to 17β-estradiol and vice versa by 17β-HSD in the endometrium. The enzymes 17β-HSD1, 5, 7, and 12 have common catalytic reductive enzyme activity (reduction of estrone into 17β-estradiol). The oxidative enzyme 17β-HSD2 catalyzes the oxidation of 17β-estradiol to estrone. Subsequently, 17β-estradiol binds to the ER leading to estrogen-sensitive gene expression. HSD, hydroxysteroid dehydrogenase; ER, estrogen receptor. In a recent study, we determined the activity of reductive 17β-HSD enzymes in the context of implantation failure in IVF patients. We used endometrial biopsies and the oxidative and reductive activities were measured by high-performance liquid chromatography (HPLC). The contribution of 17β-HSD1 was deduced by using a specific inhibitor. We concluded first that the reductive activity in the human endometrium is not performed exclusively by 17β-HSD1 as its inhibition could not block completely reductive 17β-HSD enzyme activity; further, an increased contribution of 17β-HSD1 compared to the other reductive 17β-HSD enzymes impaired endometrial receptivity. These findings were observed specifically in IVF patients with a history of primary infertility, but not in the complete study group, that included participants after a first failed IVF attempt, among which were patients with both primary and secondary infertility status. We speculated that the primary infertile group is more likely to represent women with an underlying endometrial cause of their infertility [ 13 ]. In the present study, we attempted to confirm our previous results (described above) in a more homogeneous study group, including only patients with RIF, and compare them with (presumed) fertile controls. Reductive 17β-HSD activity was determined in endometrial biopsies obtained during the WOI, and inhibitors specific to distinct reductive 17β-HSD enzymes were used to measure the contribution of enzymes other than 17β-HSD1 to the total reductive 17β-HSD activity. We also determined, for the first time, the contribution of oxidative 17β-HSD enzymes in these groups. In online supplementary Table 1 (for all online suppl. material, see https://doi.org/10.1159/000546442 ), we provide an overview of these reductive and oxidative enzymes. We hypothesize that endometrial metabolism of estrogen is altered during the WOI in women with RIF compared to controls.

The authors have no conflicts of interest to declare.

The authors would like to thank Dr. Koskimies, Forendo Pharma Ltd., for gifting the compounds.

The current project was funded by the Executive Board of Maastricht University and MUMC+. The project is supported by the EVA (Erfelijkheid Voortplanting and Aanleg) specialty program (Grant No. KP111513) of the MUMC+. The funder had no role in the design, data collection, data analysis, and reporting of this study.

This study protocol was reviewed and approved by the medical Ethical Committee of the Maastricht University Medical Centre+ (MUMC+), registration number Dutch Trial Register NL7571, www.onderzoekmetmensen.nl . Written informed consent was obtained from all participants to participate in the study.

Study design: L.B.P.M.S.B., B.D., A.R., R.J.T.G., and J.E.H. Tissue collection: S.X., L.B.P.M.S.B., and B.D. Data collection: L.B.P.M.S.B., B.D., D.O., A.R., F.M., L.B.P.M.S.B, and S.X. Data analysis: L.B.P.M.S.B., B.D., D.O., and A.R. Manuscript writing: L.B.P.M.S.B., A.R., R.J.T.G., J.E.H., and S.X. Review and approval: all authors.