Collection of human endometrium, decidua, and serum was carried out with the approvals of the Medical Ethics Committee of Nanfang Hospital, Southern Medical University, the Scientific Research Ethics Committee of the Drum Tower Hospital, Nanjing University Medical School, and the Ethical Committee of the Women’s Hospital, Zhejiang University School of Medicine. Serums (n = 15) and endometrial biopsies (n = 9) of the non-pregnant group were collected during a regular medical examination. Serums of the pregnant patients (n = 13) were collected during the first visit of antenatal care at 5.93 ± 0.71 weeks of gestation. 6 of them were sampled the 2nd time at about 2 weeks after elective abortion. Decidual tissues (n = 9) were collected from women undergoing an elective abortion in the first trimester of pregnancy. There are no significant differences in gravidity, parity, and BMI between the two groups. Serum levels of UA were measured by Blood Uric Acid Analyzer (Roche Applied Science, Basel, Switzerland). Tissues were fixed in 4% PFA and embedded in paraffin for further use. All the patients were written noticed before sampling. Demographic information on the patients who provided samples for this study is included in Table 1.

All mice (ICR strain) were housed in the SPF Experiment Animal Facility with the approval of the Institutional Animal Care and Use Committee of South China Agricultural University. Mice were maintained in a temperature- and light-controlled environment (12 h light and 12 h dark cycle). Mature female mice (8–10 weeks) were mated with fertile males of the same strain to induce pregnancy; the day of the vaginal plug was marked as D1. To artificially induce decidualization, mice were ovariectomized and treated with E2 (100 ng, Sigma) for 3 days and P4 (1 mg, Sigma, United Sates) plus E2 (6.7 ng) for another 3 days, with 2 days of rest in between to mimic the hormonal environment of normal early pregnancy. 6 h after the last injection, 10 μL sesame oil (Sigma, United Sates) or 2 mg/mL MSU (Sigma, United Sates) was infused intraluminally into one or two uterine horns of mice to induce decidualization (S). The uninjected horn served as a hormonal control (Non-S). Daily injections of P4 plus E2 were continued for 4 days to maximize the decidual response, the day of injection was marked as AD1. A diagram of the artificial decidualization protocol was shown in Figure 2. For hormone treatment, mice were ovariectomized and treated with E2 (100 ng, Sigma, United Sates) for 3 days and P4 (1 mg, Sigma, United Sates) for 6 or 24 h. Mice were then sacrificed, serum UA was measured by Mouse UA ELISA Kit (Huiyan, China), and uterine tissues were collected and used for UA measurement, the rest of uterus were fixed in 4% PFA or snap-frozen for further use.

mESCs were isolated as follow: Uteri were washed with Hanks’ balanced salt solution (HBSS, Sigma) and then digested by digest solution [1% (w/v) trypsin (Amresco, United Sates) and 6 mg/ml dispase (Roche Applied Science, Switzerland) in 3.5 ml HBSS] for 1 h at 4°C, followed by 1 h at room temperature and 10 min at 37°C. After the luminal epithelial cells were removed by HBSS washing, the remaining tissues were digested in 6 ml of HBSS containing 0.15 mg/ml collagenase I (Invitrogen, United Sates, 17100-017) at 37°C for 35 min. The stromal cells were collected and cultured in DMEM/F12 (Sigma, United Sates, # D2906) containing 2% heat-inactivated fetal bovine serum (FBS, Biological Industries, United Sates). In vitro decidualization of mESCs was induced with 10 nmol/L estrogen (E2) and 1 μmol/L progesterone (P4) in DMEM/F12 containing 2% charcoal-treated FBS (cFBS, Biological Industries, United Sates) in vitro. HESC was obtained from ATCC and cultured in DMEM/F12 supplemented with 10% FBS, and in vitro decidualization of HESC was induced by treatment of the cells with 1 mmol/L medroxyprogesterone (MPA, Sigma, United Sates), and 500 mmol/L cyclic adenosine monophosphate (cAMP, Sigma, United Sates) in DMEF/F12 supplemented with 2% cFBS. Cells and culture medium were collected, and the UA levels were measured by Mouse UA ELISA Kit (Huiyan, China).

Total RNAs were extracted from the whole uteri or cultured cells using TRIZOL (TaKaRa, Japan) method, and cDNAs were reverse-transcribed by PrimeScript reverse transcriptase reagent kit (Vazyme, China) according to the manufacturer’s introduction. qPCR was performed using an SYBR Premix Ex Taq kit (Vazyme, China) on the CFX96 Touch^TM Real-Time System (BioRad, United Sates). Data were analyzed and normalized to RPL7 for human data and Rpl7 for mouse data. The corresponding primer sequences of each gene used in this study were listed in Table 2.

Each probe’s cDNA template was amplified with the specific primers (listed in Table 2) and cloned into a pGEM-T plasmid (Promega, United Sates). Digoxigenin-labeled anti-sense or sense cRNA probes were transcribed in vitro using a digoxigenin RNA labeling kit (Roche Applied Science, Switzerland). In situ hybridization was performed as previously described (Hu et al., 2008). In brief, uteri were frozen sectioned into 10 mm, and then fixed in 4% (wt/vol) paraformaldehyde, permeabilized, and hybridized with each anti-sense cRNA probe (1:100) at 55°C overnight. A digoxigenin-labeled sense probe was used as the negative control. Following hybridization and post-hybridization washes, sections were incubated in an anti-digoxigenin antibody conjugated with alkaline phosphatase at 4°C overnight (Roche Applied Science, Switzerland). The signal was visualized as dark brown by incubating within a substrate solution containing 5-bromo-4-chloro-3-indolyl phosphate (Amresco, United Sates) and nitro blue tetrazolium (Amresco, United Sates). The activity of endogenous alkaline phosphatase was blocked by levamisole (2 mM, Sigma, United Sates). The sections were counterstained with 1% methyl green.

The 10 μm frozen sections were fixed in 4% paraformaldehyde for immunofluorescence and incubated in a permeabilizing solution containing 0.01% Triton X-100 (Sigma, United Sates). Sections were then blocked and incubated overnight at 4°C in an anti-precipitated UA antibody (1: 100, AB-T168, Advanced Targeting Systems, United Sates). Next, sections were incubated in respective species-specific fluorochrome-conjugated secondary antibody and then mounted with Vectashield Antifade Mounting Medium with DAPI. For immunohistochemistry, 4% paraformaldehyde-fixed, paraffin-embedded uterine tissues were sectioned into 6 μm. Following deparaffinization and hydration, sections were subjected to antigen retrieval in citrate buffer and hydrogen peroxide treatment. They were then incubated with an anti-XDH primary antibody (1: 100, Abcam, United Sates) overnight. After incubated with secondary antibody, HRP and diaminobenzidine (Vector Laboratories, United Sates) were used to visualize antigens. Sections were counterstained with hematoxylin. Normal IgG was used as a negative control to validate the specificity of antibodies.

Western blot was performed as described previously (Hu et al., 2008). Briefly, proteins were extracted from cultured cells by RIPA lysis buffer, and the concentrations were calculated using the bicinchoninic acid (BCA) method (Thermo Fischer Scientific, United Sates). 10 mg protein was separated on SDS/PAGE gels and transferred onto a PVDF membrane (Millipore, United Sates). The membranes were then blocked and incubated overnight with anti-COX2 (1: 1,000, Cell Signaling Technology, United Sates) or anti-α-tubulin (1: 2,000, Cell Signaling Technology, United Sates) antibodies overnight at 4°C. After incubation with respective secondary antibodies labeled with HRP, immunocomplexes were visualized by enhanced chemiluminescence (ECL) on X-ray films.

Frozen uterine sections were fixed with 4% paraformaldehyde in PBS for 10 min at 4°C and washed with 1 × PBS three times for 5 min. The uterine sections were then washed in PBS, followed by incubation at room temperature in an AP substrate solution containing 5-bromo-4-chloro- 3-indolyl phosphate (Amresco, United Sates) and nitro blue tetrazolium (Amresco, United Sates).

Data are expressed as means ± SEMs. Mann-Whitney test (unpaired) or Wilcoxon test (paired) was used for human data with two groups. Mouse data were analyzed using the Student’s t-test or one-way ANOVA, followed by Tukey’s post hoc multiple-range test. p < 0.05 was considered significant. All statistical analyses were performed by GraphPad Prism 8.0 (GraphPad Software). Power analysis was performed by using G × Power. The powers of all the human data in this study were more than 0.85.

We first explored the expression pattern of Xdh, the key synthetase of uric acid, in the mouse uterus from early pregnancy. There was no detectable signal in the uterus during natural pregnancy in mice, from D1 to D4 (Figure 1A). However, from D5, the time point that blastocyst implantation and decidualization occur, the signal of Xdh mRNA is observed to be localized in the stromal cells surrounding the implanting embryo and further diffused to the decidua zone through D6 to D8 (Figure 1A). On the other hand, mRNA expression of Uox, the gene that encodes urate oxidase, the metabolic enzyme of uric acid, was highly expressed on D1 of early pregnancy, but not detected in mouse endometrium from D2 through D8 (Figure 1B).

The developing embryo and the decidualized endometrial stromal cells are two major contributors to the physiological changes during early pregnancy. To test if the increase of Xdh expression was caused by either decidualized stromal cells or implanting embryos, we used a mouse model named artificial decidualization. In this model, mice were ovariectomized, and serial E2 and P4 were given exogenously to mimic hormone levels of early pregnancy. Then, one horn of the endometrium was then stimulated to decidualize by injection of sesame oil without the presence of an embryo (Dec S), while another horn received no stimulation and served as control (Dec Non-S) (Su et al., 2016). Meanwhile, we have experimented with an additional group of mice named Non-Dec: the hormone primed mouse without oil stimulation in either of the two horns. A diagram of this experimental design was shown in Figures 2A,B. We then examined the expression pattern of Xdh in uterine horns of Non-Dec mice and both the stimulated (S) and non-stimulated (Non-S) horns of Dec mice. The result showed higher levels of Xdh mRNA and protein in artificial decidualized endometrium (Dec S) than those in non-stimulated control horn of the same mouse (Dec Non-S) or horns from Non-Dec mice (Figures 2C,D).

Decidualization of the endometrial stromal cell was controlled by both E2 and P4 (Griffith et al., 2017). To test whether E2 or P4 induced higher expression of Xdh, we treated ovariectomized mice with either E2 or P4. The results showed that E2 but not P4 induced Xdh expression in only 6 h, indicating a regulation of the estrogen pathway on Xdh (Figures 1C,D).

The spatiotemporal expression pattern of Xdh and Uox mRNAs suggested that in situ synthesis and accumulation of uric acid very likely occurred in the endometrial stromal cells during the decidualization process.

To test whether the higher expressed Xdh in decidua tissue could in situ synthesize UA and further increased the serum UA level, we measured serum UA level (sUA) and endometrial tissue UA level (eUA) in both Dec and Non-Dec groups of mice. Compared to Non-Dec mice, which received no stimulation, the sUA levels of decidualized mice were approximately twice higher (500.37 ± 18.97 vs. 273.78 ± 34.74 μmol/L, p < 0.01, Figure 3A). As described above, the artificial stimulation by oil injection in one uterine horn of Dec mice was the only variable between the two groups of mice. Therefore, the significant difference in serum UA levels of the two groups of mice resulted from decidualization induction, indicating the higher expression of Xdh in decidua resulted in in situ synthesis of UA. Furthermore, in Dec mice, endometrial tissue UA levels (eUA) were significantly higher in the stimulated uterine horn (Dec S) than the unstimulated horn (Dec Non-S) of the same mouse (4.81 ± 0.14 vs. 3.84 ± 1.09 μmol/g, p < 0.05, Figure 3B), further confirmed the in situ synthesis of UA by decidualization induced higher level of Xdh. In addition, we tested the Xdh expression and the levels of UA in in vitro decidualized mESCs. After 4 days of induction, we detected a significantly increased level of Xdh mRNA in the E + P treated decidualization group compares to the vehicle-treated group (Figure 3C), associated with increased UA levels in these cells (Figure 3D), as well as in the supernatant of the culture medium (Figure 3E).

A generally agreed saturation point of MSU crystallization is sUA > 410 μmol/L (6.8 mg/dL) (Dalbeth et al., 2019). To explore the possibility that the in situ synthesized UA results in the crystallization of MSU in decidualized endometrium, we further performed immunofluorescence using an antibody that stains explicitly precipitated but not free uric acid (van de Hoef et al., 2013). Our result showed that MSU crystal existed in the endometrium 24 h after artificial decidualization (S), but no MSU signal was detected in the endometrium without decidualization (NS) (Figure 3F). During natural pregnancy, the MSU signal is presented in the stroma surrounding the embryo, where decidualization is initiated, on D5 and D6 of pregnancy (Figure 3G). In contrast, non-decidualized endometrium from D1 through D4 showed no detectable staining of MSU antibody (D3 was shown representatively in Figure 3G). These results suggested that the high level of in situ synthesized UA in the decidualized endometrium led to MSU crystallization.

Many inflammatory factors and immune cells have been proved to play essential roles during embryo implantation and decidualization (Mor et al., 2011). We then explored if the increased local UA level or MSU crystal, a well-known sterile inflammatory mediator, can affect the process of decidualization. To test our hypothesis, we treated mouse endometrial stromal cells (mESC) with UA or MSU crystal in vitro, together with a well-established decidualization cocktail (E + P). As shown in Figures 4A,B, only MSU crystal but not soluble UA was able to enhance the decidualization response of mESC in only 6 h, evidenced by the highly expressed decidualization marker gene Prl8a2. Furthermore, MSU crystal induced higher expression of Prl8a2 even without combining with the E + P cocktail (Figure 4C). In addition, human endometrial stromal cells (HESC) were also used to test the effects of MSU crystal on decidualization. Similar to mESC, the response of HESC to decidualization treatment (MPA + cAMP) was significantly enhanced by MSU crystal (Figure 4D).

Next, we tested the effect of MSU crystal on decidualization in vivo by replacing sesame oil, the commonly used inducer in the artificial decidualization model, with MSU crystal inartificial decidualization mouse model. A total amount of 2 mg/mL of MSU crystal can perfectly mimic sesame oil in inducing artificial decidualization of hormone primed receptive endometrium. The decidual weight, alkaline phosphatase staining, and expression of decidualization marker genes Prl8a2, Bmp2, and Wnt4 were comparable in the two groups (Figures 4E–H), indicating that MSU crystal act as an inducer for decidualization similar to sesame oil.

To further confirm that MSU is necessary for the decidualization of mESCs, we used Allopurinol, an antagonist of xanthine and hypoxanthine, to inhibit the activity of XDH, therefore which in turn reduces the level of sUA and MSU (Dalbeth et al., 2019). In our study, administration of Allopurinol could inhibit mESCs decidualization in a dose-dependent manner. When the dose reached 3 mmol/L, allopurinol treatment significantly repressed expression of Prl8a2 in in vitro decidualized mESC (Figure 4I), further proved the importance of the in situ synthesis UA and MSU to the decidualization of mESC. These results demonstrated that higher expressed Xdh during decidualization in endometrium resulted in in situ synthesis UA and MSU, which in turn promoted decidualization itself.

Many inflammatory factors have been shown as essential mediators of decidualization, such as Cyclooxygenase 2 (COX2, encoded by Ptgs2) and Interleukin 11 (IL-11, encoded by Il11) (Lim et al., 1997; Dimitriadis et al., 2002). Previously, MSU crystal was also reported to significantly stimulate Cox2 and IL-11 expression in other cell types such as osteocyte and macrophage (Scanu et al., 2015; Chhana et al., 2018; Liu et al., 2018). In this study, we tested the effect of MSU crystal in the expression of these pro-inflammatory genes. The results showed that expressions of both Ptgs2 and Il11 were induced by MSU treatment in mESC (Figure 5A,C,F). Mpges1, the synthetase of Prostaglandin E, was also increased by MSU crystal treatment, but without a statistical significance (Figure 5B). Moreover, the expression pattern of Ptgs2 in mouse decidua induced by MSU was comparable to decidua induced by sesame oil or implanting blastocyst (Figure 5D). Furthermore, NS398, an inhibitor to COX2 activity, was able to abolish the up-regulation of Prl8a2 expression induced by MSU crystal (Figure 5E). We also confirmed the capacity of IL-11 to induce in vitro decidualization on mESC (Figure 5G). Our results taken together suggested that MSU may induce decidualization via these inflammatory pathways.

At last, we wondered if the in situ synthesis of UA in decidua was also the case in women. We first tested the expression level of XDH in decidual tissue obtained from elective abortion women, which showed dramatic higher XDH expression than non-decidualized endometrium from non-pregnant women, consistent with the observation in mice (Figures 6A,B). We then assessed serum UA (sUA) levels of women in their early stages of the first trimester of gestation (5.93 ± 0.71 weeks). As expected, the sUA levels in these pregnant women were significantly higher than those in women without pregnancy (308.15 ± 30.16 vs. 263.07 ± 29.21 μmol/L, p < 0.05, Figure 6C). Additionally, 6 women from the pregnant group then accepted elective abortion. About 2 weeks after the elective abortion, the sUA levels of the same set of patients went down significantly compared to the data before abortion (279.17 ± 10.69 vs. 323.67 ± 25.89 μmol/L, p < 0.05, Figure 6D), and were comparable to that of non-pregnant women. These data suggested a short-term increase of sUA level at the early stage of the first trimester. Taken together, our data from human biopsy suggested a similar in situ synthesis of UA in women decidua during early pregnancy to that in mice.