Increased expression of HMGB1 in the implantation phase endometrium is related to recurrent implantation failure

Impaired endometrial receptivity was the main cause of recurrent implantation failure (RIF); however, its underlying mechanisms had not been elucidated. This study aimed to determine the expression level of high-mobility group box protein 1 (HMGB1) in the endometrium with RIF and its effect on endometrial receptivity. Genome-wide expression profiling, real-time reverse transcription PCR, immunohistochemical staining, western blot, and in vitro assays were performed in this study. We found that HMGB1 expression was significantly decreased in the implantation phase endometrium in the control group (patients with tubal infertility and successfully achieve conception after the first embryo transfer) (P = 0.006). However, the expression levels of HMGB1 mRNA and protein were significantly upregulated during the implantation phase in endometrial tissues obtained from patients with RIF compared to that in the control group (P = 0.001), consistent with the results of the genome-wide expression profiling. Moreover, in vitro assays showed that increased expression of HMGB1 in human endometrial epithelial cells dramatically displayed a marked deficiency in supporting blastocysts and human embryonic JAR cells adhesion, which mimic the process of embryo adhesion. These findings strongly indicated that increased HMGB1 levels suppressed the epithelial cell adhesion capability, therefore contributing to impaired endometrial receptivity in patients with recurrent implantation failure, which can be used as a target for the recognition and treatment of recurrent implantation failure in clinical practice.


Introduction
Recurrent implantation failure (RIF), defined as the absence of clinical pregnancy following more than three in vitro fertilization-embryo transfer (IVF-ET) or intracytoplasmic sperm injection (ICSI) treatments with a total of at least four good-quality embryos (score ≥ 7 or grade of 3BB or better) [1], accounts for approximately 10% of cases of sterility in women during IVF-ET treatment [2]. RIF is not only a challenge for clinicians but also increases the economic burden and psychological trauma for patients [3]. RIF etiologies are complex and include maternal endocrine abnormality, hydrosalpinx, intrauterine adhesions, etc. [4][5][6]. However, the precise mechanisms of RIF are not well-understood. Endometrial receptivity and embryo quality are the main factors influencing the embryo implantation rate; however, it was reported that impaired endometrial receptivity is the major cause of RIF [7].
Endometrial receptivity refers to the ability of the endometrium to allow embryos to adhere, invade, and induce a series of changes in endometrial cells, and finally accept embryo implantation [8]. Factors that may affect endometrial receptivity include uterine fibroids, intrauterine adhesions, endometriosis, and hydrosalpinx [9], among others. Establishment of stable adhesion between the embryo and Mi Han and Yi Cao have contributed equally to this work.
* Bufang Xu bufangxu@163.com * Aijun Zhang zhaj1268@163.com 1 3 maternal tissue is crucial for endometrial receptivity acquisition [10]. Many adhesion molecules have been identified, such as the endometrial receptivity marker ITGB3, which interacts with osteopontin to participate in adhesion of the embryo to the epithelium [11]. Nevertheless, the mechanism of impaired endometrial receptivity is not well-understood. Moreover, few effective treatments have been developed for patients with infertility showing low endometrial receptivity in clinical practice [12]. High-mobility group box protein 1 (HMGB1) is a non-histone nucleoprotein abundantly expressed in most mammalian cells [13] and participates in several cellular processes. For example, increased HMGB1 expression in astrocytes reduced tumor cell adhesion [14]. Furthermore, some researchers showed that HMGB1 expression was significantly increased in unexplained recurrent spontaneous abortion [15]. In rats, exogenous administration of recombinant human HMGB1 protein led to pregnancy failure, indicating a crucial role for HMGB1 in embryo implantation [16]. Our previous genome-wide expression profiling analysis showed that HMGB1 expression was significantly increased in the implantation phase endometrium in patients with RIF compared to in the control group [17]. However, the role of HMGB1 in RIF remains unclear. In this study, we examined the effect of high HMGB1 expression on endometrial receptivity in RIF.

Clinical sample collection
Uterine tissues were collected via pipe suction curettage (Wallace) from patients who underwent IVF-ET/ICSI treatment at the Reproductive Medical Center of Ruijin Hospital, School of Medicine, Shanghai Jiao Tong University between January 2016 and December 2020. All patients signed an informed consent form before they were recruited in the study. Patients with a history of tubal infertility who successfully achieved conception after the first embryo transfer were recruited as the control group (n = 38) and divided to three phases of the menstrual cycle: early proliferative phase (days 4-5, n = 10); late proliferative phase (days 11-13, n = 10); implantation phase (days 20-23, n = 18). Patients with no pregnancy after at least three embryo transfers including a total of ≥ 4 good-quality embryos were recruited as the RIF group (n = 18). The inclusion criteria for all participants included: 25-35 years old, regular ovulatory cycles with every 28-32 days, normal serum levels including follicle stimulating hormone (FSH < 10 mIU/mL), luteinizing hormone (LH < 10 mIU/mL) and estradiol (E 2 < 50 mIU/ mL) on day 3 of the menstrual cycle. The exclusion criteria for all patients included: presence of intrauterine pathology, hydrosalpinx, salpingitis, polycystic ovary syndrome, endometriosis, adenomyosis, chromosome abnormalities, or autoimmune disease, and any intrauterine device or contraceptive drug within the last 6 months.
For endometrial biopsy, specimens collected from the control group in the three phases were used to analyze differences in the HMGB1 phase-specific expression in the normal natural menstrual cycle (Supplementary Table 1). The implantation phase endometrium from 18 patients in the control group and 18 patients in the RIF group was collected on day LH + 7 to examine the expression of HMGB1 and HOXA10 (Supplementary Table 2). As we described previously [18], patients were monitored for follicle development by transvaginal ultrasound (Hitachi EUB-2000, Tokyo, Japan) until the diameter of the dominant follicle was ≥ 15 mm, and then serum LH and E2 levels were quantified daily. The day on which the LH level peaked (≥ 20 mIU/mL) was considered as day LH 0, the window/implantation phase was defined as day LH + 7. To isolate primary endometrial cells, the endometrium was collected from the control group (late proliferative phase, n = 12).

Microarray assay
Whole-genome gene expression profiling was conducted using uterine tissues from three patients with RIF and three tissues from the control group in the implantation phase as described previously [18]. Data were analyzed using the Gene Expression Omnibus database (GSE103465), and genes were considered as differentially expressed when they showed a fold-change ≥ 2 and adjusted P-value ≤ 0.05.

Real-time reverse transcription PCR (qRT-PCR)
Total RNA was extracted from cultured cells and endometrial biopsy tissues using the TaKaRa9767 kit (Takara, Shiga, Japan). Next, 1 μg RNA was reverse-transcribed into cDNA using Primescript Reverse transcriptase (Takara). The synthesized template cDNA, gene-specific primers (Supplementary Table 3), and fluorescent probes (Takara) were used for qRT-PCR. PCR was conducted on a QuantStudio™ 6 Flex Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA, USA). The mRNA expression level was normalized to that of GAPDH. Relative mRNA expression levels were calculated using the 2 −ΔΔCt method.

Immunohistochemical staining
Immunohistochemical (IHC) staining was performed on 5-µm sections from 4% paraformaldehyde-fixed paraffinembedded endometrial tissue samples. The slides were incubated with 5% bovine serum albumin (Amresco, Solon, OH, USA), and then incubated with anti-HMGB1 (dilution 1:400, ab79823, Abcam) and normal rabbit IgG antibodies overnight at 4 °C. Images were acquired using an Olympus BX53M fluorescence microscope (Tokyo, Japan). The H-score was calculated based on the percentage of cells stained at each intensity level, ranging from 3 to 0 (equaling to strong, moderate, weak, and negative staining, respectively).

Cell culture and transfection
Isolation of primary human endometrial stromal cells (HESCs) and primary human endometrial epithelial cells (HEECs) from endometrial samples was performed as described previously [17]. Briefly, fresh endometrial biopsy samples were rinsed twice with PBS, and then minced and digested with 1 mg/mL collagenase type IV (Thermo Fisher Scientific) at 37 °C for 30-45 min. After enzymatic digestion, the mixture was passed through a 100-mm sieve followed by a 40-mm sieve, with HESCs collected by passing the mixture through a 40-μm sieve and HEECs collected on the opposite side of the 40-μm sieve. All cells were cultured in Dulbecco's modified Eagle medium (DMEM) (Thermo Fisher Scientific) supplemented with 10% fetal bovine serum at 37 °C.
Negative control and HMGB1 over-expressing plasmids were purchased from GeneCopoeia (Rockville, MD, USA) and transfected into cells using transfection reagent (Roche) following the manufacturer's instructions.

Proliferation assay
The capacity for cell proliferation was examined using Cell Counting Kit-8 (CCK-8, Dojindo, Kumamoto, Japan) in endometrial stromal cells (ESCs) including T-HESCs and HESCs. After 2 days of plasmid transfection, the cells were seeded into 96-well plates at 2000 cells/well and 10 µL of CCK-8 reagent was added to each well. After incubating at 37 °C for 2 h, the optical densities were determined at 450 nm excitation using a microplate reader.

Cell migration assay
Ishikawa cells and HEECs were transfected with negative control and HMGB1 over-expressing plasmids and plated into 6-well plates. After transduction for 3 days when the cell confluence reached 80-90%, the cells were cultured in DMEM containing 0.1% fetal bovine serum for 24 h to achieve deprivation conditions. A vertical linear scratch was created in the Ishikawa monolayer cells using a 200-µL micropipette tip, with the floating cells removed by washing with sterile PBS. The cell migration distance was measured under a light microscope via Image J 1.46r (Nunes and Dias, NIH, Bethesda, MD, USA) at 0, 24, and 48 h after continuous culturing of the scratched cells in serum deprivation media at 37 °C.

Embryo adhesion assay
The co-culture model of "embryo" adhesion was constructed as described by Jemma Evans et al. [19]. Briefly, HEECs were isolated from four late proliferative-phase endometrium samples from controls as described above. Ishikawa cells, possessing apical adhesiveness and similar characteristics of endometrial luminal and glandular epithelium, were commonly used to study endometrial receptivity and embryo adhesion [20,21]. Ishikawa cells and HEECs transfected with HMGB1 over-expressing or negative control plasmids were grown to confluence in a 48-well plate. Endometrial cells were then treated with a hormonal paradigm to mimic hormonal exposure throughout the menstrual cycle: cells were primed with 10 -8 M β-estradiol (E2758, Thermo Fisher Scientific, referred to as E) for 24 h, and then treated with combined E and 10 -7 progesterone (V900699, Thermo Fisher Scientific, referred to as P) for 24 h, followed by treatment with E/P and 10 IU human chorionic gonadotropin for 24 h. An adhesion assay was then performed. A total of 44 hatched blastocysts from C57BL/6 mice with a normal morphology was cocultured with confluent monolayers of acceptable HEECs for 6 h in DMEM/F12 complete medium. After washing the cells three times with PBS, images of the adherent blastocysts were captured for morphological examination with a TS2 microscope (Nikon, Tokyo, Japan). JAR cells (human trophoblastic cell line) labeled with PKH26 (PKH26GL-1KT, Abcam) for 10 min at 37 °C were then added to fully confluent endometrial epithelial cells (EECs) including Ishikawa cells and HEECs monolayers for 1 h at 37 °C, followed by vigorous agitation three times at 140 rpm for 5 min before refreshing the medium. The attached JAR cells were then imaged with a fluorescence microscope (Leica V3.8, Wetzlar, Germany).

Statistical analysis
All experiments were independently performed at least three times. The results are shown as the mean ± standard error of the mean (SEM) and analyzed using SPSS software (version 23.0, SPSS, Inc, Chicago, IL, USA). Data were analyzed using Student's t-test between two groups and one-way analysis of variance among more than two groups. Data were considered as statistically significant when P < 0.05.

Baseline characteristics
The characteristics of the control and RIF groups are shown in Supplementary Table 2. There were no significant differences between groups in average age, body mass index (BMI), follicle-stimulating hormone (FSH), luteinizing hormone (LH), or estradiol (E 2 ), although the number of embryo transfers differed significantly (P < 0.001). Similarly, there were no significant differences among the different phases of the menstrual cycles, as indicated by demographic characteristics (Supplementary Table 1).

HMGB1 expression in the endometrium of the menstrual cycles
In the control group endometrium, the mRNA levels of HMGB1 were high during the proliferative phases and significantly decreased during the implantation phase (Fig. 1a). However, HMGB1 protein levels increased gradually from the early to late proliferative phase, and then decreased during the implantation phase (Fig. 1b, c). IHC staining showed that HMGB1 was localized in the stroma and epithelium of LP late proliferative phase. Data are presented as the mean ± SEM. *P < 0.05; ***P < 0.001 the human endometrium, and mostly located in the epithelium (Fig. 1d-f).

HMGB1 expression significantly increased in the RIF group during the implantation phase
According to our previous microarray results [17], the HMGB1 expression level was increased in the RIF group (fold-change = 2.00, P < 0.05, Fig. 2a). Consistently, qRT-PCR showed that the mRNA levels of HMGB1 were significantly increased in the RIF group during the implantation phase (P < 0.01) (Fig. 2b). Similarly, as shown in Fig. 2c and d, HMGB1 protein levels were considerably increased in the endometrium of patients with RIF during the implantation phase (P < 0.001), and HOXA10 protein expression was significantly lower in the RIF group. IHC staining showed that HMGB1 expression in the implantation phase endometrium of the RIF group was increased compared to in the control group (Fig. 2e, f), in line with the above-mentioned findings. Through receiver operating characteristic curve analysis, we verified the potential functional role of HMGB1 in endometrial receptivity, as the area under the curve was 0.8878 [0.803-0.972] for tissue HMGB1 measurements (Fig. 2g).

HMGB1 overexpression did not influence cell proliferation and decidualization of T-HESCs and HESCs
Based on the higher expression of HMGB1 in the endometrial stroma of RIF compared to in controls, HMGB1overexpressing T-HESCs and HESCs were successfully constructed to examine the impact of HMGB1 on biological function, including stromal cell proliferation and in vitro decidualization. Compared with the negative control group, the mRNA expression of HMGB1 was 869 times higher in the HMGB1-overexpressing T-HESCs group and 66 times higher in the HMGB1-overexpressing HESCs group (Fig. 3b, d). As shown in Fig. 3, HMGB1 overexpression did not affect cell proliferation or decidualization in T-HESCs and HESCs.

HMGB1 overexpression in EECs inhibited embryo adhesion in vitro
HMGB1-overexpressing HEECs and Ishikawa cells were successfully constructed, as confirmed by qRT-PCR and western blotting. Elevated HMGB1 levels had no effect on the migration ability of EECs (Fig. 4). However, overexpression of HMGB1 suppressed cell adhesion of embryonic JAR cells onto Ishikawa cells (Fig. 5a) and HEECs (Fig. 5c). The fluorescence intensity of fluorescence-labeled JAR cells adherent on over-expressing HMGB1 EECs cells was decreased compared to that on control EECs (Fig. 5b, d). We further examined a heterologous co-culture model of "embryo" adhesion. As shown in Fig. 5e, 44 mouse embryos were randomly assigned to the two groups in five parallel experiments, after co-culture for 6 h with the acceptable HEECs, in the negative control group, out of 22 blastocysts, 13 attached; however, in the HMGB1-overexpression group, of 22 blastocysts, five attached. HMGB1-overexpression in EECs significantly inhibited embryo attachment to the endometrial epithelium (Fig. 5f, g).

Discussion
Reproductive failure is a common social and economic issue delaying parenthood [22]. Although remarkable advances have been made in IVF-ET in clinical pregnancy, the technique fails repeatedly for several couples who are infertile [23]. Currently, RIF, largely defined as implantation failure due to impaired endometrial receptivity, is a challenging clinical problem in assisted reproductive technology [7]. Therefore, it is imperative to determine the potential mechanism of RIF. In this study, we found that HMGB1 overexpression impaired endometrial receptivity by affecting cell adhesion between the embryo and endometrial epithelium.
We previously identified CDYL [24] and PIBF1 [17], which were consistent with HOXA10 [18], an accepted endometrial receptivity marker, expressed dynamically in the endometrium of the normal menstrual cycles in the control group. The dynamic expression changes of the mentioned above molecules in the different phases of normal menstrual cycles suggested that they play important roles in the acquisition of endometrial receptivity. Similarly, in T-HESCs and HESCs after HMGB1 overexpression during in vitro decidualization. Experiments were repeated three times independently. All data are presented as the mean ± SEM. *P < 0.05, **P < 0.01 HMGB1 expression showed dynamic changes during the menstrual cycles. During normal menstrual cycles in the control group, HMGB1 expression in the endometrium gradually increased from the early to late proliferative phase but decreased during the implantation phase, indicating that decreased expression of HMGB1 during the implantation phase is favorable for the acquisition of endometrial receptivity, in other words, increased expression of HMGB1 may inhibit embryo implantation.
In our previous study, genome-wide expression profiling suggested that HMGB1 expression was significantly increased in patients with RIF during the implantation phase, which was verified by qRT-PCR and western blotting in this study. Furthermore, the endometrial receptivity marker HOXA10 was significantly decreased in the RIF group, supporting impaired endometrial receptivity in patients with RIF. In addition, receiver operating characteristic analysis revealed the high sensitivity and specificity of HMGB1 in endometrial tissues for discriminating endometrial receptivity. Additionally, low expression of HMGB1 in the implantation phase may be necessary for establishing endometrial receptivity, and increased expression of HMGB1 may contribute to impaired endometrial receptivity in patients with RIF.
IHC staining was performed to identify HMGB1 localization in the endometrium. HMGB1 was localized in the stroma and epithelium in the human endometrium. The proliferation and decidualization capability of ESCs and the migration and adhesion capability of EECs are essential for embryo implantation and for building endometrial Experiments were repeated three times independently. All data are presented as the mean ± SEM receptivity [17,25]. To examine the mechanism by which HMGB1 affects endometrial receptivity, we constructed HMGB1-overexpressing ESCs and EECs to evaluate changes in cell biology processes compared to the negative control group. In vitro experiments with plasmid transfection to overexpress HMGB1 in T-HESCs and HESCs showed no effect on IGFBP1 and PRL (decidualization markers) levels, indicating that increased HMGB1 levels did not affect endometrial decidualization. The results depicted in Fig. 4 demonstrated that the effect of HMGB1 overexpression on endometrial receptivity was not dependent on endometrial epithelial cell migration capability. However, Fig. 5 showed that overexpression of HMGB1 significantly inhibited the adhesion of EECs to JAR cells. Additionally, in the heterologous co-culture model of "embryo" adhesion, HMGB1 overexpressing in EECs displayed a marked deficiency in supporting blastocyst adhesion, which mimic the process of embryo adhesion.
HMGB1 influences cell proliferation, invasion and motility, adhesion, and inflammation [26,27]. Although only a few studies on HMGB1's direct effect on cell adhesion have been performed, it has been shown to reduce cell adhesion in astrocytes. In a previous report, HMGB1 was found to bind to the CpG islands of the SASH1 gene, and its overexpression promoted SASH1 methylation and reduced the expression of adhesion molecule integrin β8, thereby reducing cell adhesion [14]. In our study, we found that HMGB1 overexpression caused a deficiency in embryo-epithelium adhesion. There exists a possibility that HMGB1 may affect cell adhesion through certain epigenetic modifications. Recently, an increasing number of studies on the role of HMGB1 in the maternal-fetal interface have suggested that HMGB1 plays an important role in embryo implantation. Aikawa found that HMGB1 sustains P4-PR signaling and uterine deficiency of HMGB1 protein causes implantation defects [28]. Furthermore, it was reported that HMGB1 overexpression at the maternal-fetal interface might account for pathophysiological responses during pregnancy, and even led to miscarriage [15]. Taken together, an adequate expression of HMGB1 is essential for embryo implantation, while deletion or overexpression is unfavorable to embryo implantation and pregnancy. Future research should focus on the underlying mechanisms by which increased HMGB1 expression inhibits embryo adhesion and decreases endometrial receptivity. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.