Abstract
Obesity is a rapidly growing public health issue among women of reproductive age associated with decreased reproductive function including implantation failure. This can result from a myriad of factors including impaired gametes and endometrial dysfunction. The mechanisms of how obesity-related hyperinsulinaemia disrupts endometrial function are poorly understood. We investigated potential mechanisms by which insulin alters endometrial transcript expression. Ishikawa cells were seeded into a microfluidics device attached to a syringe pump to deliver a constant flow rate of 1 µL/min of the following: (i) control (ii) vehicle control (acidified PBS), or (iii) insulin (10 ng/mL) for 24 h (n = 3 biological replicates). Insulin-induced transcriptomic response of endometrial epithelial cells was determined via RNA sequencing, and DAVID and Webgestalt to identify Gene Ontology (GO) terms and signalling pathways. A total of 29 transcripts showed differential expression levels across two comparison groups (control vs vehicle control; vehicle control vs insulin). Nine transcripts were differentially expressed in vehicle control vs insulin comparison (P < 0.05). Functional annotation analysis of transcripts altered by insulin (n = 9) identified three significantly enriched GO terms: SRP-dependent co-translational protein targeting to membrane, poly(A) binding, and RNA binding (P < 0.05). The overrepresentation analysis found three significantly enriched signalling pathways relating to insulin-induced transcriptomic response: protein export, glutathione metabolism, and ribosome pathways (P < 0.05). Transfection of siRNA for RAPSN successfully knocked down expression (P < 0.05), but this did not have any effect on cellular morphology. Insulin-induced dysregulation of biological functions and pathways highlights potential mechanisms by which high insulin concentrations within maternal circulation may perturb endometrial receptivity.
Lay summary
Changes in components of blood associated with obesity in women of reproductive age can have consequences for pregnancy success. These changes to circulating molecules associated with obesity can alter the ability of the endometrium (the innermost lining of the womb/uterus) to be receptive to an embryo to implant – a key stage of successful pregnancy. Understanding which molecules contribute to this is difficult and one in particular, insulin, can change the role of the endometrium. Studying this is limited to static culture, that is, the cells are not exposed to sustained and high concentrations of Insulin that could occur in the mother. In this study, we use a new laboratory-based approach (microfluidics) that allows us to mimic maternal circulation. We have determined that exposure of these endometrial cells to insulin changes the expression of specific genes that may lead to the inability of the endometrium to support implantation and early pregnancy.
Introduction
Obesity is a complex disease with multifactorial aetiologies and in the UK, almost half of women within the childbearing age range are overweight or obese. Due to complex biochemical changes in the endocrinological and metabolic profile of those suffering from obesity, there are resulting complex health issues including insulin resistance and compensatory hyperinsulinaemia (Talmor & Dunphy 2015). This has been linked to poor reproductive outcomes in women of childbearing age such as miscarriage, anovulation, irregular menstrual cycles, and decreased conception rates following the use of assisted reproductive technologies (Pasquali et al. 2003, Silvestris et al. 2018).
The impact of obesity on ovarian function dysregulation is well recognised although complex. The higher concentrations of insulin in circulation associated with obesity can inhibit sex hormone binding globulin (SHBG) production by the liver (Daka et al. 2013) along with aromatisation of androgens to oestrogens associated with increased adipose tissue. As SHBG binds to circulating sex hormones, the combined effects of high peripheral aromatisation of androgens and decreased SHBG leads to an overall increase in bioavailable oestradiol and testosterone, therefore contributing to hyperandrogenism in the theca cells of the ovarian follicles (Pasquali et al. 2003, Talmor & Dunphy 2015). Hyperinsulinaemia and subsequent hyperandrogenism coupled with altered hormonal milieu can lead to premature follicular atresia and anovulation (Pasquali et al. 2003). Furthermore, increased insulin in circulation associated with obesity can lead to inhibition of hepatic and ovarian IGF binding protein 1 (IGFBP1) expression, an important regulator of ovarian and endometrial function. Overall, the combined effects of systemic insulin resistance and hyperinsulinaemia in obesity contribute to biochemical and sex hormone dysregulation within the female reproductive system. Such symptoms are common features of polycystic ovarian syndrome (PCOS), and obesity-related hyperinsulinaemia seems to exacerbate PCOS symptoms (Broughton & Moley 2017).
The effects that this dysregulation in metabolic and biochemical profiles associated with obesity has on endometrial function and uterine receptivity remain unclear. Conflicting results and discrepancies in methodologies between the studies may explain in part why this is the case. A series of studies from ovum donation cycles identified significantly higher spontaneous abortion rates in obese women (38.1%) compared to women with normal body mass index (BMI: 13.3%) (Bellver et al. 2003), with a follow-up study of 2656 first ovum donation cycles concluded that high BMI altered endometrial environment and function, although these effects are quite small (Bellver et al. 2007). A further investigation into 9587 first-cycle ovum donations showed that implantation, clinical pregnancy, and live-birth rates were significantly reduced in obese women, although miscarriage rates showed no difference with BMI (Bellver et al. 2013) similar to what was observed by Cano et al., (Cano et al. 2001). In contrast, no significant differences in receptivity impairment between obese and non-obese women were observed (Wattanakumtornkul et al. 2003) or implantation rate or miscarriage rate in high BMI groups compared to lower BMI (Styne-Gross et al. 2005) although in this study miscarriage rates were disproportionately high.
Using the ovum donation model to isolate the impact of endometrial factors from the embryonic influence on implantation capability is controversial as it limits study participants only to those participating in oocyte donation. Given there are likely fundamental differences in those that require ovum donation in obese and non-obese scenarios, these are likely to compound results (Howards & Cooney 2008, Brewer & Balen 2010). Overall, these clinical studies highlight the gap in knowledge in regard to the mechanism by which the metabolic and biochemical changes in circulation contribute to dysfunction in endometrial function and receptivity.
Glucose metabolism is important during the peri-implantation period of pregnancy with expression of the facilitative glucose transporters (GLUT) in endometrial epithelia modulated by oestradiol and progesterone (Kim & Moley 2009). Furthermore, the upregulation of glucose and its transporter (GLUT1) is critical for endometrial stromal cell decidualisation in the mouse (Schulte et al. 2015). Obesity-related insulin resistance and hyperinsulinaemia lead to disturbances in glucose metabolism, which may have downstream effects on endometrial function and receptivity (Shanik et al. 2008). Dissecting the complex interactions between insulin-specific alterations on endometrial function without the confounding influence of additional metabolic stressors in circulation is difficult. Moreover, static culture systems in vitro do not recapitulate exposure in vivo. Evidence in sows looking at the proteomic analysis on endometria from those with normal reproductive performance and low reproductive performance to pinpoint the differentially expressed genes (DEGs) and pathways (i.e. insulin-signalling pathway and lipid metabolism) associated with uterine dysfunction. Chen et al. subsequently used insulin-resistant animal models driven by a high-fat diet and Ishikawa cells (human endometrial adenocarcinoma cells) to create in vitro implantation models and reported that insulin resistance reduces receptivity through mitochondrial dysfunction and consequent oxidative stress (Chen et al. 2021). Additionally, the use of a microfluidics endometrium-on-a-chip model in bovine demonstrated cellular transcriptional and secretome differences if the endometrium following exposure to physiological extremes of metabolic factors glucose and insulin in microfluidics (De Bem et al. 2021). However, these studies do not recapitulate the complexity of the in vivo scenario in humans and extrapolating findings in animals can be limited given differences in endometrial morphologies. There is an overall scarcity of literature highlighting hyperinsulinaemia in obesity as an aetiology of poor uterine function and implantation. We therefore propose to investigate the impact of obesity-related metabolic stressors of insulin on endometrial function. To achieve this, we used a microfluidics approach to better mimic in vitro, exposure of the endometrial epithelium to Insulin that would occur in vivo. This allowed us to identify the transcriptional response of the endometrial epithelium revealing the potential mechanism by which hyperinsulinaemia in obesity may contribute to endometrial dysfunction.
Materials and methods
Unless otherwise stated all consumables were sourced from Sigma Aldrich.
Culture of Ishikawa cells and device preparation
Passage 20–24 Ishikawa cells (Immortalised human endometrial cells; ECACC #99040201) cultured in DMEM/F-12 (Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12), containing 10% fetal bovine serum (FBS), 1% glutamine, streptomycin, penicillin (GSP)) were trypsinised from a T75 flask when 90% confluent and counted using a haemocytometer. Media (described earlier) was added to cells until they reached a concentration of 1,000,000/mL. The microfluidics device (Ibidi u-slide V0.4) was placed in the incubator at 37°C. One hundred microlitres of the 1,000,000/mL Ishikawa cell suspension was subsequently seeded into each channel of V0.4 slides, and left to attach for 12 h. Media (DMEM/F12, 10% FBS, 1% GSP) was added to each channel until almost full to prevent the channel from drying. Ishikawa cells were left for 24 h in a 37°C incubator with 5% CO2 before undergoing treatment in the microfluidics device.
Treatment of Ishikawa cells with insulin in a microfluidic device
Each channel of the microfluidics device was flushed five times with 37°C PBS and the inlet syringes were connected with sterile tubing. Elbow connectors were then used to attach the sterile tubing to the microfluidics device inlet. Sterile tubing was filled with medium from the syringes and attached to the device. A droplet-to-droplet method was used to prevent the introduction of bubbles into the device. Syringes were pushed to fill both the chamber and outlet. The outlet tubing was filled with sterile PBS and connected to the outlets on microfluidics channels using the droplet-to-droplet method. The other end of each outlet tube was connected to a 7 mL bijou container to collect and save the conditioned medium. After setting up the microfluidics device, the medium was pushed through the entire device system to check for bubbles and to get a droplet at the end of each outlet tube inside the bijou container. The microfluidics device set-up can be pictured in Fig. 1.
Schematic diagram of experimental setup and workflow of experiment. Diagrammatic representation of experimental approach using a microfluidics approach to mimic exposure to endometrial epithelial cells (n = 3 biological replicates) to media control, vehicle control, or Insulin (10 ng/mL) for 24 h. Cells were analysed for a transcriptional response via RNA sequencing. Images produced with Biorender.
Citation: Reproduction and Fertility 4, 2; 10.1530/RAF-21-0120
Cells within the devices were exposed to one of the following three treatments for 24 h (n = 3 biological replicates): (i) media control, (ii) vehicle control (VC; phosphate-buffered saline (PBS) acidified to pH 3 with acetic acid (A6283)), and (iii) Insulin (10 ng/mL) in PBS acidified to pH 3 with acetic acid). In addition, we also exposed cells to an additional vehicle of 70% ethanol to be able to compare how treatment under flow differs from that under static conditions that we have previously published (Hume et al. 2023). Each treatment was loaded into a 5 mL syringe and put in a 37°C incubator overnight and loaded onto the syringe pump on the day of the microfluidics run. The syringe pump and microfluidics system were placed into a 37°C incubator and the syringe pump was set to flow at the rate of 1 µL/min in order to mimic secretion in vivo (as previously described; (De Bem et al. 2021)).
Cell recovery and RNA extraction and RNA sequencing
After disconnecting the microfluidics device from the system, every channel from the device was flushed with PBS. Ishikawa cells were lifted with trypsin, neutralised with media, and centrifuged at 500 g for 5 min to pellet the cells. The resulting cell pellet was then snap-frozen in liquid nitrogen until RNA extraction. RNA extraction was carried out by using the Mini RNeasy Kit (Qiagen), according to instructions provided by the manufacturer. Extracted RNA was sent to Novogene (Cambridge, UK) for subsequent library preparation. Ribosomal RNA was first removed and a directional sequencing library was constructed using NEBNext® UltraTM Directional RNA Library Prep Kit for Illumina® (NEB, Ipswich, MA, USA) following the manufacturer’s protocol. Indices were included to multiplex multiple samples. Briefly, the first strand cDNA was synthesized using random hexamer primers followed by the second strand cDNA synthesis. The strand-specific library was ready after end repair, A-tailing, adapter ligation, size selection, and USER enzyme digestion. After amplification and purification, the insert size of the library was validated on an Agilent 2100 and quantified using quantitative PCR (Q-PCR). Libraries were then sequenced on Illumina NovaSeq 6000 S4 flowcell with PE150 according to results from library quality control and expected data volume.
Bioinformatic analysis
The raw and processed FASTQ files were scanned for the fundamental quality control metrics using FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/). The adapter sequences in the reads were trimmed off using Cutadapt (Martin 2011) and poor-quality bases or reads were trimmed or discarded using PRINSEQ (Schmieder & Edwards 2011). The human reference genome file (GRCh38) and corresponding GTF file were downloaded from GENCODE Human Release 31 (Frankish et al. 2019). The processed paired-end reads were aligned onto the human reference genome using STAR (Dobin et al. 2013). The resultant alignment files were converted, sorted, and indexed with SAMtools (Li et al. 2009), and only uniquely mapped reads were selected for the downstream analysis. The reads were counted for each gene using featureCounts function of Rsubread package and a matrix of raw read counts for all genes was produced (Liao et al. 2019a ). The differential expression testing was performed using DESeq2 with apeglm method for the log2 fold change shrinkage > 0.1 (or <−0.1) and P value of <0.05 (Love et al. 2014).
Differentially expressed protein-coding genes found in three treatment groups were subjected to Gene Ontology (GO) functional annotations and overrepresentation enrichment analysis using DAVID (The Database for Annotation, Visualization and Integrated Discovery; Bioinformatics Resourced 6.8 (Huang et al. 2009)) and WebGestalt (WEB-based Gene SeT AnaLysis Toolkit (Liao et al. 2019b )), respectively. Gene identifier for DAVID was chosen as ‘OFFICIAL_GENE_SYMBOL,’ with list type selected as ‘Gene List.’ When using WebGestalt, the functional database was chosen as ‘KEGG Pathway,’ and gene ID type was ‘Gene symbol.’ The gene reference set when using WebGestalt was ‘genome protein-coding.’ P-value cutoff was P < 0.05. P-values for both DAVID and WebGestalt were adjusted as FDR using the Benjamini–Hochberg Method, with the statistical significance threshold set as FDR < 0.05. GO functional annotation terms were presented by their Enrichment scores, calculated by −log10(Pvalue), and overrepresentation analysis on WebGestalt was displayed by presenting each signalling pathways’ enrichment ratios and FDR.
We also sought to compare differences in Ishikawa cells cultured under static and flow conditions using previously published data sets (Hume et al. 2023, GEO number GSE211151).
Knockdown of RAPSN in vitro
We chose to knockdown RAPSN as it was the only transcript that was upregulated following Insulin treatment. Confluent Ishikawa cells were treated with one of the following (n = 5 biological replicates per treatment; n = 2 technical duplicate): (i) Control (NT; no treatment), (ii) V (vehicle) (iii) I (insulin: 10 ng/mL), (iv) sC V (siRNA control, vehicle), (v) sC I (siRNA control, insulin), (vi) sR V (siRNA RAPSN, vehicle), or (vii) sR I (siRNA RAPSN, insulin). RAPSN siRNA SMARTpools and scrambled control SMARTpools (Dharmacon, Lafayette, CO, USA) were used at 20 nM to transfect cells using DharmaFECT1 (Dharmacon), according to the manufacturer's instructions. siRNA treatment began 24 h after seeding in the antibiotic-free growth medium and vehicle/insulin treatment began 48 h after siRNA treatment. Cells were fixed in 4% formaldehyde for immunofluorescence staining or lysed for RNA extraction after 24 h of treatment. RNA was analysed on a Nanodrop 1000, diluted to 200 ng/μL, and reverse transcribed into cDNA using the high-capacity reverse transcription kit (Thermo Fisher, 4368814) as per the manufacturer’s protocol. cDNA was diluted to 10 ng/μL and 2 μL from each sample plated for qRT-PCR in technical triplicate. RAPSN cDNA was analysed on a Lightcycler 480, using 5 μL SYBR Green (Roche). Twenty micrometres of forward and reverse primers (Integrated DNA Technologies) for RAPSN (NM_005055.5; Forward – GGAGGTGGGGAACAAGCTGA; Reverse – CGATGGCATCCAGAGCCTTG) or ACTB (NM_001101.4; Forward – AGAAAATCTGGCACCACACC: Reverse – TAGCACAGCCTGGATAGCAA) and 2.5 μL water per well were used. ACTB expression was used as a normaliser gene. Programme was run as per the Roche SYBR Lightcycler 480 recommended protocol for 35 cycles. A Ct value of 35 was assigned to wells with no signal by the Lightcycler programme. Technical triplicate Ct values for each sample were averaged and processed as per the 2−DDCt method (Livak & Schmittgen 2001). One-way ANOVA analysis was conducted in GraphPad Prism with Tukey’s multiple comparison correction on the DDCt values. Percentage knockdown calculated by (1 − DDCt) × 100, DDCt being ‘DCt RAPSN siRNA - DCt non-targeting siRNA’.
Staining and fluorescence microscopy
Fixed Ishikawa cells on glass coverslips were quenched with 50 mM ammonium chloride solution and then permeabilization in 0.5% Triton-X 100 PBS before incubation with the primary antibody in PBS for 2 h (mouse anti-ZO-1, BD Biosciences, Wokingham, Berkshire, UK). After washes in PBS, coverslips were incubated for 1 h in PBS containing donkey anti-mouse secondary antibody conjugated with Alexa-488 (Life Technologies), phalloidin conjugated with Alex-568 (Life Technologies), and 4′,6-diamidino-2-phenylindole (DAPI) (Sigma). Washed coverslips were mounted on glass slides with mowiol (Sigma). Optical sectioning fluorescence microscopy was performed using an Apotome-equipped Zeiss Axiophot microscope and Zeiss Zen software for image capture. Maximum intensity projections were formed from 18 to 21 optical sections.
Results
Differences in transcriptional response of Ishikawa cells to static vs microfluidic culture conditions
In order to understand how Ishikawa cells differed following treatment under static and control conditions, we compared control and EtOH-exposed cells under static conditions (Hume et al. 2023) and under flow (this experiment). EtOH treatment under flow altered the expression of 100 transcripts (Fig. 2). In comparison, 1155 transcripts were differentially expressed between control and ethanol-treated cells under a static system (Fig. 2). Only a small cohort of transcripts are different all in both static and flow conditions (MT-CYB, SRRM2, FTH1,MYL9, MT-ND5, MT-ND4, SPEN, NFIC, PRR12, MT-ATP6, BCL9L, MBD6, MT-ND1, RPL14, and ZFHX3).
Venn diagram analysis of differentially expressed transcripts in Ishikawa cells cultured under static or flow conditions. Comparison of differentially expressed transcripts from Ishikawa cells cultured in a Microfluidic device and Static culture (Hume et al. 2023). Ishikawa cells were cultured (n = 3 biological replicates) as control or with vehicle control (70% EtOH routinely used for steroid hormone treatment). Statistical analysis for differential gene expression was conducted via DESeq2 with cut-offs of log2 fold change > 0 and P < 0.05. P adj is also shown – P-value adjusted for multiple testing. All transcripts in the Venn diagram are comparisons between control and vehicle control. Transcripts with the highest log2 fold change for each list are shown in the boxes.
Citation: Reproduction and Fertility 4, 2; 10.1530/RAF-21-0120
Of the 85 DEGs found exclusively in cells treated under flow, these were overrepresented in pathways associated with systemic lupus erythematosus, alcoholism, ribosome, necroptosis, basal transcription factors, endocrine and other factor-related calcium reabsorption, central carbon metabolism in cancer, taste transduction, and MAPK signalling pathway (Fig. 3; Table 1). Overrepresented pathways in the static culture only were proteasome, propanoate metabolism, oxidative phosphorylation, carbon metabolism, Parkinson’s disease, Alzheimer’s disease, Huntington’s disease, ribosome pathway, non-alcoholic fatty liver disease, and thermogenesis pathways (Table 2). Of the DEGs identified in both microfluidic and static culture, overrepresented pathways were involved in ferroptosis, mineral absorption, leukocyte transendothelial migration, vascular smooth muscle contraction, ribosome, signalling pathways regulating pluripotency of stem cells, oxytocin signalling pathway, necroptosis, cGMP-PKG signalling, and tight junction pathways (Table 3).
Overrepresented pathways associated with response to treated cells under static and flow conditions. Overrepresented pathways from differentially expressed genes are significantly expressed in (A) microfluidic culture only, (B) static culture only, and microfluidic and static culture, and (C) static culture only. Dark blue indicates significant FDR and P-value. Light blue indicates a significant P-value and grey indicates the pathway was not significantly overrepresented.
Citation: Reproduction and Fertility 4, 2; 10.1530/RAF-21-0120
Overrepresentation analysis of transcripts significantly expressed in microfluidic culture only.
| Gene Set | Description | Size of set | Overlap | Gene symbol | Expected value | Enrichment ratio | P-value† | FDR |
|---|---|---|---|---|---|---|---|---|
| Hsa05322 | Systemic lupus erythematosus | 133 | 14 (10.5) | ACTN4, HIST1H2AB, HIST1H2AC, HIST1H2BD, HIST1H2BF, HIST1H2BN, HIST1H3B, HIST1H4C, HIST1H4E, HIST12H2AB |
0.76826 | 18.223 | 7.9936e-15* | 2.6059e-12* |
| Hsa05034 | Alcoholism | 180 | 14 (7.78) | HIST1H2AB, HISTAH2AC, HIST1H2BD, HIST1H2BF, HIST1H2BN, HISTAH3B, HIST1H4C, HIST1H4E, HIST2H2AB, HIST2H2BE | 1.0397 | 13.465 | 5.6533e-13* | 9.2148e-11* |
| Hsa03010 | Ribosome | 134 | 7 (5.22) | MRPL24, MRPS2, RPL19, RPL36, RPL41, RPS15, RPS6 | 0.77403 | 9.0435 | 0.0000097243* | 0.00079253* |
| Hsa04217 | Necroptosis | 162 | 4 (2.47) | FTL, HIST1H2AB, HIST1H2AC, HISTA2H2AB | 0.93577 | 4.2745 | 0.013765* | 0.89751 |
| Hsa03022 | Basal transcription factors | 45 | 2 (4.44) | GTF2F1, TAF3 | 0.25994 | 7.6942 | 0.027570* | 1 |
| Hsa04961 | Endocrine and other factor-regulated calcium reabsorption | 47 | 2 (4.26) | AP2S1, PRKACA | 0.27149 | 5.3668 | 0.029887* | 1 |
| Hsa05230 | Central carbon metabolism in cancer | 65 | 2 (3.08) | G6PD, MYC | 0.37546 | 5.3267 | 0.053896 | 1 |
| Hsa04742 | Taste transduction | 83 | 2 (2.41) | PRKACA, TAS2R14 | 0.47944 | 4.1715 | 0.082680 | 1 |
| Hsa04010 | MAPK signalling pathway | 295 | 4 (1.36) | FGF18, MYC, PRKACA, STM1 | 1.7040 | 2.3474 | 0.089146 | 1 |
*Indicates significant following microfluidic exposure of Ishikawa cells (n = 3); † P < 0.05.
FDR, false discovery rate.
Overrepresentation analysis of transcripts significantly expressed in microfluidic culture only.
| Gene set | Description | Size of set | Overlap | Expected value | Enrichment ratio | P-value* | FDR† |
|---|---|---|---|---|---|---|---|
| Hsa03050 | Proteasome | 45 | 16 (35.56) | 2.8160 | 5.6819 | 5.0948e-9 | 2.3727e-7 |
| Hsa00640 | Propanoate metabolism | 32 | 10 (31.25) | 2.0025 | 4.9938 | 0.000015412 | 0.00050244 |
| Hsa00190 | Oxidative Phosphorylation | 133 | 38 (28.57) | 8.3228 | 4.5658 | 4.4409e-16 | 1.4477e-13 |
| Hsa01200 | Carbon metabolism | 116 | 28 (24.14) | 7.2590 | 3.8573 | 3.0705e-10 | 1.6683e-8 |
| Hsa05012 | Parkinson’s disease | 142 | 33 (23.24) | 8.8860 | 3.7137 | 2.2925e-11 | 1.8684e-9 |
| Hsa05010 | Alzheimer’s disease | 171 | 37 (21.64) | 10.701 | 3.4577 | 1.3167e-11 | 1.4308e-9 |
| Hsa05016 | Huntington’s disease | 193 | 41 (21.24) | 12.077 | 3.3948 | 1.8334e-12 | 2.9885e-10 |
| Hsa03010 | Ribosome | 134 | 28 (20.90) | 8.3854 | 3.3392 | 1.0115e-8 | 4.1220e-7 |
| Hsa04932 | NAFLD | 149 | 29 (19.46) | 9.3240 | 3.1101 | 2.9670e-8 | 0.0000010747 |
| Hsa04714 | Thermogenesis | 229 | 42 (18.34) | 14.330 | 2.9309 | 1.4802e-10 | 9.6510e-9 |
*P < 0.05; †FDR < 0.05.
FDR, false discovery rate; NAFLD, non-alcoholic fatty liver disease.
Overrepresentation analysis of transcripts significantly expressed in microfluidic and static culture.
| Gene set | Description | Gene set size | Overlap (%) | Gene symbols | Expected value | Enrichment ratio | P-value† | FDR‡ |
|---|---|---|---|---|---|---|---|---|
| Hsa04216 | Ferroptosis | 40 | 1 (2.5) | FTH1 | 0.022005 | 45.444 | 0.021829* | 1 |
| Hsa04978 | Mineral absorption | 51 | 1 (1.96) | FTH1 | 0.028057 | 35.642 | 0.027769* | 1 |
| Hsa04670 | Leukocyte transendothelial migration | 112 | 1 (0.89) | MYL9 | 0.061615 | 16.230 | 0.060218 | 1 |
| Hsa04270 | Vascular smooth muscle contraction | 121 | 1 (0.83) | MYL9 | 0.066566 | 15.023 | 0.064936 | 1 |
| Hsa03010 | Ribosome | 134 | 1 (0.75) | RPL14 | 0.073718 | 13.565 | 0.071719 | 1 |
| Hsa04550 | Signalling pathways regulating pluripotency of stem cells | 139 | 1 (0.72) | ZFHX3 | 0.076468 | 13.077 | 0.074318 | 1 |
| Hsa04921 | Oxytocin signalling pathway | 152 | 1 (0.66) | MYL9 | 0.083620 | 11.959 | 0.081050 | 1 |
| Hsa04217 | Necroptosis | 162 | 1 | FTH1 | 0.089121 | 11.221 | 0.086204 | 1 |
| Hsa04022 | cGMP-PKG signalling pathway | 162 | 1 (0.66) | MYL9 | 0.089671 | 11.152 | 0.086718 | 1 |
| Hsa04530 | Tight junction | 170 | 1 (0.59) | MYL9 | 0.093522 | 10.693 | 0.090311 | 1 |
*Indicates significance following static culture of Ishikawa cells (n = 3); †P < 0.05; ‡FDR < 0.05.
FDR, false discovery rate.
Acidified PBS used as a vehicle alters the transcriptome of endometrial epithelial cells exposed in a microfluidic device
DEGs from each of the three treatment groups (control, VC, and insulin) were compared with one another to minimise the confounder of VC added in with insulin. Comparison of control to VC altered the expression of 22 transcripts in total (Fig. 4: Table 4). Genes involved in the biological processes of mitochondrial electron transport, NADH to ubiquinone, and mitochondrial respiratory chain complex I assembly were overrepresented in the list of DEGs (Table 5). There were more transcripts associated with the cellular components of the respiratory chain, mitochondrial inner membrane, mitochondrial respiratory chain complex I, and mitochondrion as well as the molecular functions of NADH dehydrogenase (ubiquinone) activity and cytochrome-c oxidase activity than one would have expected by chance (for full details see Table 5). While there were a number of overrepresented pathways (including those associated with thiamine metabolism and folate biosynthesis), none had an FDR of <0.05 (Fig. 5: Table 6).
Venn diagram of DEGs identified in endometrial epithelial cells following exposure to control media, vehicle, or insulin in a microfluidic device for 24 h. Ishikawa cells (n = 3) were exposed to control medium, vehicle control (acetic acid: VC), or insulin (10 ng/ml) for 24 h in a microfluidics device (1 μL/min). Following RNA sequencing DEGs in treatment groups were identified using a cut-off of P < 0.05 and log2 fold change > 0 or log2 fold change < 0; fold change differences for transcripts were determined using paired t-tests. ‘Control vs VC’ demonstrates the transcripts altered when exposing Ishikawa cells to control compared to those altered in VC. ‘VC vs insulin’ shows the comparison of DEGs expressed when Ishikawa cells were exposed to VC and insulin treatment. Results were found to be significant when P < 0.05, and the results were derived from n = 3 biological replicates.
Citation: Reproduction and Fertility 4, 2; 10.1530/RAF-21-0120
Overrepresentation analysis (ORA) graphs depicting enriched KEGG pathways associated with DEGs found in control vs vehicle control groups. ORA was produced by Webgestalt with the gene identifier ‘Gene Symbol’ and functional database of ‘KEGG pathways.’ Benjamini–Hochberg method was used to adjust P-values and presented as FDR, FDR ≤ 0.05 being the threshold for statistical significance. (A) Bar chart illustrates the enrichment ratio plotted against the enriched pathway identified through ORA, with the blue-colour shade indicating statistical significance based on FDR (dark blue = FDR ≤ 0.05; light blue = FDR > 0.05). The chart is organised according to decreasing enrichment ratio, with the top pathway being the most enriched and the bottom being the least enriched. (B) Volcano plot plotting −log10 of FDR against log2 of enrichment ratio for each pathway. Darker purple shade of plotted dots indicates pathway of interest, as it demonstrates notable magnitude in both log2 of enrichment ratio and −log10 of FDR compared to other regulatory networks.
Citation: Reproduction and Fertility 4, 2; 10.1530/RAF-21-0120
Transcript ID and description, associated with differentially expressed transcripts between control, vehicle control, and insulin-treated groups. DEGs were identified following RNA sequencing of cells treated with control, vehicle control (PBS acidified with acetic acid), or insulin (10 ng/mL) exposed endometrial epithelial (Ishikawa) cells cultured in a microfluidics device for 24 h (n = 3 biological replicates).
| Transcript | Description |
|---|---|
| Control vs vehicle control | |
| MT-CYB | Mitochondrially encoded cytochrome B |
| MT-CO1 | Mitochondrially encoded cytochrome C oxidase 1 |
| MT-CO2 | Mitochondrially encoded cytochrome C oxidase 2 |
| MT-ND3 | NAD + hydrogen (NADH)-ubiquinone oxidoreductase chain 3 |
| MT-ND5 | NADH-Ubiquinone oxidoreductase chain 5 |
| COA6 | Cytochrome C oxidase assembly factor 6 |
| ALPG | Alkaline phosphatase, germ cell |
| THEGL | Theg spermatid protein-like |
| HIST1H2BN | Histone H2B type 1-N |
| HMGA1 | High mobility group AT-Hook 1 |
| SCX | Scleraxis BHLH transcription factor |
| MKI67 | Marker of proliferation Ki-67 |
| RPL36AL | Ribosomal protein L36A-like |
| RPL19 | Ribosomal protein L19 |
| PRKACA | Protein kinase cAMP-activated catalytic subunit alpha |
| CTSZ | Cathepsin Z |
| PPDPF | Pancreatic progenitor cell differentiation and proliferation factor |
| TCEAL4 | Transcription elongation factor A like 4 |
| MT-ND1 | Mitochondrially encoded NADH: ubiquinone oxidoreductase core subunit 1 |
| MT-ND4 | Mitochondrially encoded NADH: ubiquinone oxidoreductase core subunit 4 |
| TMSB10 | Thymosin beta 10 |
| RPS6 | Ribosomal protein S6 |
| Vehicle control vs insulin | |
| SSB | Small RNA binding exonuclease protection factor La |
| SPCS1 | Signal peptidase complex subunit 1 |
| BTF3 | Basic transcription factor 3 |
| GGCT | Gamma-glutamyl cyclotransferase |
| RPL7A | Ribosomal protein L7a |
| RAPSN | Receptor-associated protein of the synapse |
| HTATSF1 | HIV-1 Tat specific factor 1 |
| TMSB10 | Thymosin beta 10 |
| RPS6 | Ribosomal protein S6 |
Gene Ontology terms involved in biological processes, cellular components, and molecular functions associated with differentially expressed transcripts between control vs vehicle control groups. Overrepresented GO terms and their descriptions associated with DEGs identified between control and vehicle-exposed Ishikawa cells cultured in a microfluidics device for 24 h (n = 3 biological replicates). DEGs were identified following RNA sequencing of cells and overrepresentation analysis determined using Webgestalt.
| Gene ontology terms/GO ID | GO ID | P-v | FDR | Enrichment score† | Transcripts involved |
|---|---|---|---|---|---|
| Biological processes | |||||
| Mitochondrial electron transport, NADH to ubiquinone* | GO:0006120 | 0.000026 | 0.0042 | 4.585 | MT-ND1,MT-ND3, MT-ND4, MT-ND5 |
| Mitochondrial respiratory chain complex I assembly* | GO:0032981 | 0.000055 | 0.0045 | 4.260 | MT-ND1,MT-ND3, MT-ND4, MT-ND5 |
| ATP synthesis coupled electron transport* | GO:0042773 | 0.0059 | 0.33 | 2.229 | MT-ND4, MT-ND5 |
| Response to hypoxia* | GO:0001666 | 0.018 | 0.58 | 1.745 | MT-ND4, MT-ND5, MT-CYB |
| Response to copper ion* | GO:0046688 | 0.018 | 0.58 | 1.745 | MT-CYB, MT-CO1 |
| Mitochondrial electron transport, cytochrome-c to oxygen* | GO:0006123 | 0.024 | 0.65 | 1.620 | MT-CO1, MT-CO2 |
| Mesoderm formation* | GO:0001707 | 0.033 | 0.74 | 1.481 | PRKACA, SCX |
| Translation* | GO:0006412 | 0.036 | 0.74 | 1.444 | RPL19, RPL36AL, RPS6 |
| Cerebellum development* | GO:0021549 | 0.043 | 0.79 | 1.367 | MT-ND4, MT-CO1 |
| Organ regeneration | GO:0031100 | 0.055 | 0.85 | 1.260 | MT-CYB, MKI67 |
| Response to organic cyclic compound | GO:0014070 | 0.057 | 0.85 | 1.244 | MT-ND1, MKI67 |
| Cellular component | |||||
| Respiratory chain* | GO:0070469 | 4.0E-9 | 2.5E-7 | 8.398 | MT-ND3, MT-ND5, MT-CYB, MT-CO1, MT-CO2 |
| Mitochondrial inner membrane* | GO:0005743 | 0.0000040 | 0.00012 | 5.398 | MT-ND1, MT-ND3, MT-ND4, MT-ND5, MT-CYB, MT-CO1, MT-CO2 |
| Mitochondrial respiratory chain complex I* | GO:0005747 | 0.000017 | 0.00035 | 4.770 | MT-ND1, MT-ND3, MT-ND4, MT-ND5 |
| Mitochondrion* | GO:0005739 | 0.000029 | 0.00044 | 4.538 | MT-ND1, MT-ND3, MT-ND4, MT-ND5, MT-CYB, COA6, MT-CO1, MT-CO2, PRKACA |
| Respiratory chain complex IV* | GO:0045277 | 0.0052 | 0.063 | 2.284 | MT-CO1, MT-CO2 |
| Membrane* | GO:0016020 | 0.021 | 0.21 | 1.678 | MT-ND1, MT-CYB, MT-CO2, MKI67, PRKACA, RPL19, RPS6 |
| Cytosolic large ribosomal subunit | GO:0022625 | 0.069 | 0.60 | 1.161 | RPL19, RPL36AL |
| Mitochondrial membrane | GO:0031966 | 0.094 | 0.71 | 1.027 | MT-ND1, MT-ND3 |
| Molecular function | |||||
| NADH dehydrogenase (ubiquinone) activity* | GO:0008137 | 0.000020 | 0.00097 | 4.699 | MT-ND1, MT-ND3, MT-ND4, MT-ND5 |
| Cytochrome-c oxidase activity* | GO:0004129 | 0.00051 | 0.012 | 3.292 | COA6, MT-CO1, MT-CO2 |
| Structural constituent of ribosome* | GO:0003735 | 0.025 | 0.33 | 1.602 | RPL19, RPL36AL, RPS6 |
| Transcriptional activator activity, RNA polymerase II distal enhancer sequencer-specific binding* | GO:0003735 | 0.028 | 0.33 | 1.553 | HMGA1, SCX |
| Copper ion binding | GO:0005507 | 0.061 | 0.59 | 1.215 | COA6, MT-CO2 |
*Marks statistical significance: P ≤ 0.05, or FDR (false discovery rate) ≤ 0.05; †(−log10 (Pvalue)).
Overrepresented signalling pathways in DEGs between control vs vehicle control groups. Overrepresented signalling pathways and their descriptions associated with DEGs identified between control and vehicle-exposed Ishikawa cells cultured in a microfluidics device for 24 h (n = 3 biological replicates). DEGs were identified following RNA sequencing of cells and overrepresentation analysis determined using Webgestalt.
| Gene set | Pathway description | Gene set size | DEG overlap | P-value | FDR | Enrichment ratio | Differentially expressed transcripts |
|---|---|---|---|---|---|---|---|
| hsa00730 | Thiamine metabolism* | 16 | 1 | 0.0175 | 0.7774 | 56.81 | ALPG |
| hsa00790 | Folate biosynthesis* | 26 | 1 | 0.0283 | 0.7778 | 34.96 | ALPG |
| hsa05020 | Prion disease* | 35 | 1 | 0.0379 | 0.7778 | 25.97 | PRKACA |
| hsa03010 | Ribosome* | 134 | 3 | 0.0003 | 0.1044 | 20.35 | RPL19, RPL36AL, RPS6 |
| hsa04371 | Apelin signalling pathway* | 137 | 2 | 0.0092 | 0.7466 | 13.27 | PRKACA, RPS6 |
| hsa04910 | Insulin signalling pathway* | 137 | 2 | 0.0092 | 0.7466 | 13.27 | PRKACA, RPS6 |
| hsa04714 | Thermogenesis* | 229 | 3 | 0.0015 | 0.2503 | 11.91 | COA6, PRKACA, RPS6 |
| hsa05034 | Alcoholism* | 180 | 2 | 0.0155 | 0.7774 | 10.10 | HIST1H2BN, PRKACA |
| hsa05205 | Proteoglycans in cancer* | 198 | 2 | 0.0185 | 0.7774 | 9.181 | PRKACA, RPS6 |
| hsa05203 | Viral carcinogens* | 201 | 2 | 0.0191 | 0.7774 | 9.044 | HIST1H2BN, PRKACA |
*Marks statistical significance; P ≤ 0.05, or FDR (False Discovery Rate) ≤ 0.05.
Exposure of endometrial epithelial cells to insulin in a microfluidics device altered the transcriptional profile
Exposure of Ishikawa cells to Insulin altered the expression of nine genes (Fig. 4: Table 4). Two of these transcripts were altered in both comparisons; however, TMSB10 and RPS6 were increased in expression in control vs VC but were significantly decreased in VC compared to insulin-treated cells. The molecular function of Poly(A) RNA binding was significantly overrepresented in cells exposed to insulin in a microfluidic device (Fig. 6: Table 7). While genes associated with the pathways of protein export, glutathione metabolism, and ribosome (Table 8) were significantly overrepresented following insulin exposure. Comparisons of the overrepresented biological processes, cellular components, and molecular functions are provided in Fig. 7.
Overrepresentation analysis (ORA) graphs depicting enriched KEGG pathways associated with DEGs found in vehicle control vs insulin-treated cells. ORA was produced by Webgestalt with the gene identifier ‘Gene Symbol’ and functional database of ‘KEGG pathways.’ Benjamini–Hochberg method was used to adjust P-values and presented as FDR, FDR ≤ 0.05 being the threshold for statistical significance. (A) Bar chart illustrates the enrichment ratio plotted against the enriched pathway identified through ORA, with the blue-colour shade indicating statistical significance based on FDR (dark blue = FDR ≤ 0.05; light blue = FDR > 0.05). The chart is organised according to decreasing enrichment ratio, with the top pathway being the most enriched and the bottom being the least enriched. (B) Volcano plot plotting −log10 of FDR against log2 of enrichment ratio for each pathway. The darker purple shade of plotted dots indicates a pathway of interest, as it demonstrates notable magnitude in both log2 of enrichment ratio and −log10 of FDR compared to other regulatory networks.
Citation: Reproduction and Fertility 4, 2; 10.1530/RAF-21-0120
Overrepresentation analysis (ORA) graphs depicting enriched KEGG pathways associated with DEGs found in vehicle control vs insulin groups. ORA was produced by Webgestalt with the gene identifier ‘Gene Symbol’ and functional database of ‘KEGG pathways.’ Benjamini–Hochberg method was used to adjust P-values and presented as FDR, FDR ≤ 0.05 being the threshold for statistical significance. (A) Bar chart illustrates the enrichment ratio plotted against the enriched pathway identified through ORA, with the blue-colour shade indicating statistical significance based on FDR (dark blue = FDR ≤ 0.05; light blue = FDR > 0.05). The chart is organised according to decreasing enrichment ratio, with the top pathway being the most enriched and the bottom being the least enriched. (B) Volcano plot plotting −log10 of FDR against log2 of enrichment ratio for each pathway. The darker purple shade of plotted dots indicates a pathway of interest, as it demonstrates notable magnitude in both log2 of enrichment ratio and −log10 of FDR compared to other regulatory networks.
Citation: Reproduction and Fertility 4, 2; 10.1530/RAF-21-0120
Bar charts indicating overrepresented GO terms for biological processes, cellular components, and molecular functions associated with DEGs. Charts are divided into two groups: (A) biological processes, (B) cellular component, and (C) molecular function for n = 22 DEGs involved in control vs VC group, and (D), (E), and (F) depicting gene ontology terms found in from n = 9 DEGs involved in VC vs insulin group. All DEGs involved in GO terms in (A), (B), and (C) were upregulated, thus shown in orange. All DEGs involved in the GO terms in (D), (E), and (F) were downregulated as RAPSN (the only upregulated DEG out of n = 9 DEG in VC vs insulin) was not associated with any of the functional annotations, thus shown in blue. GO terms and P-values were determined by overrepresented functional annotation analysis using DAVID. P-values were used to calculate enrichment scores (−log10( Pvalue)). The colour gradient depicts the size of the enrichment score, darker being more enriched (or having a higher P-value), and lighter being less enriched.
Citation: Reproduction and Fertility 4, 2; 10.1530/RAF-21-0120
Gene Ontology terms involved in biological processes, cellular components, and molecular functions associated with differentially expressed transcripts between vehicle control vs insulin-treated groups. Overrepresented GO terms and their descriptions associated with DEGs identified between vehicle and insulin (10 ng/mL)-exposed Ishikawa cells cultured in a microfluidics device for 24 h (n = 3 biological replicates). DEGs were identified following RNA sequencing of cells and overrepresentation analysis determined using Webgestalt.
| Gene Ontology terms | GO ID | P-value | FDR | Enrichment Score† | Transcripts involved |
|---|---|---|---|---|---|
| Biological function | RPL7A, RPS6 | ||||
| SRP-dependent cotranslational protein targeting to membrane* | GO:0006614 | 0.044 | 0.82 | 1.357 | |
| Viral transcription | GO:0019083 | 0.052 | 0.82 | 1.284 | |
| Nuclear-transcribed mRNA catabolic process, nonsense-mediated decay | GO:0000184 | 0.055 | 0.82 | 1.260 | |
| Translational initiation | GO:0006413 | 0.063 | 0.82 | 1.201 | |
| rRNA processing | GO:0006364 | 0.098 | 0.99 | 1.009 | |
| Cellular component | |||||
| Intracellular ribonucleoprotein complex | GO:1990904 | 0.58 | 1.0 | 0.2366 | SSB, RPS6 |
| Ribosome | GO:0005840 | 0.71 | 1.0 | 0.1487 | RPL7A, RPS6 |
| Molecular function | |||||
| Poly(A) binding* | GO:0008143 | 0.0011 | 0.021 | 2.959 | HTATSF1, SSB, BTF3, RPL7A, RPS6 |
| RNA binding* | GO:0003723 | 0.026 | 0.24 | 1.585 | HTATSF1, SSB, RPL7A |
*Indicates statistical significance: P ≤ 0.05, or FDR (false discovery rate) ≤ 0.05; †−log10 (Pvalue).
Overrepresented signalling pathways in DEGs between control vs vehicle control groups. Overrepresented signalling pathways and their descriptions associated with DEGs identified between vehicle and insulin (10 ng/mL)-exposed Ishikawa cells cultured in a microfluidics device for 24 h (n = 3 biological replicates). DEGs were identified following RNA sequencing of cells and overrepresentation analysis determined using Webgestalt.
| Gene set | Pathway description | Gene set size | DEG overlap | FDR | P-value | Enrichment ratio | Differentially expressed transcripts |
|---|---|---|---|---|---|---|---|
| Hsa03060 | Protein export* | 23 | 1 | 1 | 0.0157 | 63.23 | SPCS1 |
| Hsa00480 | Glutathione metabolism* | 56 | 1 | 1 | 0.0379 | 25.97 | GGCT |
| Hsa03010 | Ribosome* | 134 | 2 | 1 | 0.0033 | 21.70 | RPL7A, RPS6 |
| Hsa01521 | EGFR tyrosine kinase inhibitor resistance | 79 | 1 | 1 | 0.0532 | 18.41 | RPS6 |
| Hsa04066 | HIF-1 signalling pathway | 100 | 1 | 1 | 0.0669 | 14.54 | RPS6 |
| Hsa05322 | Systemic lupus erythematosus | 133 | 1 | 1 | 0.0882 | 10.93 | SSB |
| Hsa04371 | Apelin signalling pathway | 137 | 1 | 1 | 0.0908 | 10.62 | RPS6 |
| Hsa04910 | Insulin signalling pathway | 137 | 1 | 1 | 0.0908 | 10.62 | RPS6 |
| Hsa04150 | mTOR signalling pathway | 151 | 1 | 1 | 0.0996 | 9.631 | RPS6 |
| Hsa05205 | Proteoglycans in cancer | 198 | 1 | 1 | 0.1290 | 7.344 | RPS6 |
*Indicates statistical significance; P ≤ 0.05, or FDR (false discovery rate) ≤ 0.05.
Knockdown of RAPSN in endometrial epithelial cells in vitro
Transfection of Ishikawa cells with siRNA for RAPSN resulted in decreased expression of more than 45% (Fig. 9). Examination of epithelial morphology in insulin and RAPSN siRNA-treated Ishikawa cells revealed no differences in the apical tight junction (arrowheads) and microvilli formation (arrows) in cells treated with insulin or RAPSN siRNA (Fig. 9).
Knockdown of RAPSN in Ishikawa cells. Relative normalised expression of RAPSN (left hand panel) in Ishikawa cells treated with (i) no treatment (NT), (ii) vehicle only (V), (iii) insulin (I), (iv) siRNA control + vehicle (Sc V), (v) siRNA control + insulin (Sc I), (vi) siRNA RAPSN + vehicle (Sr V), and (vii) siRNA RAPSN + insulin (Sr I). Expression calculated using the 2−DDCt method with all treatment groups compared to the no treatment (NT) group and normalised to the expression of ACTB. One-way ANOVA analysis conducted in GraphPad Prism with Tukey’s multiple comparison correction, statistical significance (P < 0.05) between groups denoted above each group. Maximum intensity projection images of Ishikawa cells treated with the stated conditions and processed for immunofluorescence staining (right-hand panel). Arrowheads indicate ZO-1-positive tight junctions and arrows indicate actin-positive microvilli, both seen at the apical domain. Scale bars 20 μm.
Citation: Reproduction and Fertility 4, 2; 10.1530/RAF-21-0120
Discussion
To the best of our knowledge, our study is the first to use a microfluidics approach to mimic exposure of the endometrial epithelium to obesogenic concentrations of insulin associated with maternal circulation to enhance our understanding of the mechanism by which endometrial function may be affected by obesity. This was conducted by allowing the exposure of Ishikawa cells in a microfluidics device to insulin concentrations associated with an obese environment and identifying the associated pathways and functional annotations that were modified. Our results indicate that maternal metabolic stressors, such as insulin, may alter the uterine function and receptivity by altering specific transcripts in the endometrial epithelium and propose the mechanisms by which this may occur (Fig. 10).
Proposed mechanism by which insulin alters endometrial function. Created using Biorender.
Citation: Reproduction and Fertility 4, 2; 10.1530/RAF-21-0120
We first sought to determine how Ishikawa cells respond to the same treatment under static and flow conditions. To do this, we compared previously publish data (Hume et al. 2023) to control and EtOH-treated cells in the device (this study). Ethanol is a widely acceptable vehicle for human endometrial cells and Ishikawa cells and facilitates treatment with steroid hormones (Lee et al. 2016) and is thought to not alter the proliferation of Ishikawa cells compared (Duman et al. 2021). Our findings have shown that there are substantial differences in the transcriptional response of cells to static compared to flow conditions and these should be taken into consideration when planning experiments or considering discrepancies between the expression of the same molecule under different conditions. Moreover, proteins involved in the ribosome pathway (MRPL24, MRPS2, RPL19, RPL36, RPL41, RPS15, RPS6) may be examined in this flow system but other pathways such as the ferroptosis pathway (previously identified as altered in patients with RIF (Bielfeld et al. 2019)) are modified in both static and flow systems. These data demonstrate the dynamic a microfluidic device better mimics the dynamic flow the endometrium is exposed to in vivo but not all transcriptional responses are seen in this system (Fig. 2).
The VC used to deliver insulin to the cells altered biological functions of mitochondrial electron transport (NADH to ubiquinone), mitochondrial respiratory chain complex I assembly, ATP synthesis; molecular functions of NADH dehydrogenase (ubiquinone activity), and cytochrome-c oxidase activity were altered. While NADH dehydrogenase promotes the smooth functioning of respiration and ATP synthesis by stimulating the assembly of respiratory chain complex I, they may have an additional role of producing apoptosis-inducing fragments to promote cell death (Saladi et al. 2020, Herrmann & Riemer 2021). This alternative role of NADH dehydrogenase in our findings with VC can be further substantiated by the fact acetic acid is reported to induce apoptosis in a gastric cell line (Okabe et al. 2014). Apoptosis in cells is initiated by acetic acid by mitochondrial dysfunction, cytochrome-c release, and along with reactive oxygen species accumulation, which seem to correlate with the GO terms found in control vs VC (Marques et al. 2013). Using acetic acid might have stimulated mitochondria-induced apoptosis in the Ishikawa cells and moreover may mask some of the insulin-specific effects on the endometrial transcriptome.
The processes and signalling pathways modified by insulin treatment may provide a major insight into the function of insulin endometrial receptivity, specifically those related to processes involving translation. Translation initiation is facilitated by the binding of ribosomes to the 5′cap of mRNAs and is mediated by the eukaryotic translation initiation factor 4F (eIF4F) complex. One of the major components of the eIF4F complex is the eIF4G, which aids in recruiting and binding of the ribosome. The 3′ poly(A) tail end of the mRNA binds to the poly(A) binding protein (PABP) with PABP and elF4G interaction synergistically initiates translation (Prévôt et al. 2003, Berlanga et al. 2006). The mTOR signalling pathway was also shown to be modified by insulin in our study and this signalling pathway is implicated in translational initiation by phosphorylating translation factors such as EIF4E binding protein (Saxton & Sabatini 2017). Furthermore, RPS6 phosphorylation (a DEG significantly downregulated by insulin in our findings) within the mTOR pathway is widely accepted to stimulate protein synthesis and translation (Mok et al. 2013); cellular protein synthesis capacity is widened by the mTOR pathway and is critical to cell survival (Johnson et al. 2013). Translation modulation and increased ribosome synthesis are implicated in endometrial epithelial cell function and receptivity given the dynamic nature of the human endometrium (Berlanga et al. 2006, Wang et al. 2015). A recent study identified the proteins involved in repeated implantation failure were involved in translation pathways, posttranslational modification and ribosomal pathways (Wang et al. 2021). It is thus clear that our findings show evidence that insulin treatment potentially modifies the translational profile of endometrial epithelial cells.
Insulin treatment induced differential expression of transcripts from Ishikawa cells, which may provide insight into how obesity-related insulin changes can alter endometrial function. RPS6 (Ribosomal protein S6) is a transcript that is associated with many of the enriched and overrepresented pathways in analysis. The gene is highly expressed in the endometrium, with 1752 transcripts per million expressed in the endometrial tissue (Howe et al. 2020). Aside from its role in translation regulation in the ribosomal and mTOR pathway, it is also implicated in immune response regulation (Howe et al. 2020). Phosphorylation of RPS6 is necessary for T-cell differentiation in the thymus (Salmond et al. 2015) and plays a key role in the mTOR pathway, which has been demonstrated to be involved in T-cell signalling (Salmond et al. 2009). The immune profile of the endometrium plays a critical role in implantation success via T-cell-mediated immune tolerisation (Robertson et al. 2018). In fact, research showed that patients experiencing recurrent implantation failure during IVF had a dysregulated endometrial uterine immune profile (Lédée et al. 2016). Taking into consideration RPS6 regulation of T cell expression during the immune response and that insulin downregulated RPS6 expression, it can be hypothesised that high circulating insulin in the endometrium may compromise the uterine immune profile and consequent receptivity.
GGCT encodes an enzymatic protein essential for glutathione metabolism. Dysregulation in the glutathione metabolism pathway is associated with insulin resistance, hyperinsulinaemia, and type 2 diabetes (Kobayashi et al. 2009, Hakki Kalkan & Suher 2013). It is thought that GGCT is important in cellular protective mechanisms by salvaging glutathione, thereby providing an antioxidant effect to cells (Kageyama et al. 2018). Glutathione is a critical modulator of normal cellular function such as gene expression, protein synthesis, and immune response but is primarily involved in antioxidative roles and nutrient metabolism (Wu et al. 2004). Glutathione deficiency leads to the accumulation of free radicals and oxidative stress (Pham-Huy et al. 2008). Glutathione metabolism also may play a major role in fertility and endometrial function, as Xu et al. (Xu et al. 2014) demonstrated that this pathway is crucial in reducing hydrogen peroxide (a type of reactive oxygen species that causes oxidative stress) during decidualisation. High concentrations of reactive oxygen species and oxidative stress adversely affect the implantation process, early embryo development, and can eventually cause implantation failure (Agarwal et al. 2005) as well as abnormal endometrial function, making the environment unsuitable to support the growth and development of the embryo (Adeoye et al. 2018). We propose that insulin can disturb glutathione homeostasis in the endometrium, thereby leading to oxidative stress and subsequent effects on uterine receptivity to implantation.
The downregulation of TMSB10 with insulin treatment encodes a protein that aids the processes of actin filament organization and actin binding (Howe et al. 2020). During the window of receptivity, the human endometrial surface goes through ultrastructural modifications which include the development of pinopodes, which are tightly managed by cytoskeletal actin filaments (Qiong et al. 2017, Quinn et al. 2020). Although previously controversial, recent literature has deemed the presence of pinopodes in the endometrial epithelia as a credible biomarker of uterine receptivity (Jin et al. 2017, Qiong et al. 2017). Filamentous actin expression coincides with pinopode formation during receptivity on the apical surface of luminal epithelial cells (D'Ippolito et al. 2020). In the endometria of those suffering recurrent pregnancy loss, both ezrin and thrombomodulins were downregulated which disrupted the organisation of cytoskeletal actin filaments and may ultimately impair pinopode formation. TMSB10 was differentially expressed and highly responsive to oestradiol and progesterone (Tamm-Rosenstein et al. 2013) and has been associated with a functional and receptive endometrium (Ullah et al. 2017). We conclude that insulin concentrations can downregulate TMSB10, which may contribute to impaired implantation capacity of the endometrium.
The fact that RAPSN was the only transcript increased in expression by Insulin in this system is interesting. While RAPSN is mostly associated with developmental disorders of the neuronal system, more recently it has been identified as a potential biomarker for epithelial cell tumours including lung (Qiao et al. 2020) and breast (Lei et al. 2021). This could suggest a potential role in cellular proliferation – particularly given its role in the regulation of actin filaments (Howe et al. 2020). While knockdown of RAPSN was successful, this, or treatment with insulin did not alter cellular morphology (Fig. 9). This could be due to different mechanisms of action in endometrial cells or indeed differences in how this transcript is expressed in a static scenario compared to under flow as described earlier (Fig. 2). In addition to this we did not see similar changes to the endometrial transcriptome in human or bovine epithelial cells exposed to similar concentrations and flow rates of Insulin (De Bem et al. 2021). While there some molecular mechanisms to establish receptivity to implantation which are conserved in placental mammals (for example, PGR signalling) there are species-specific signals such as that for pregnancy recognition (hCG and Interferon Tau in humans and bovine respectively) (Taylor et al. 2021). Therefore, the actions of insulin on these cell systems could be species-specific. While expression of a number of the glucose family members is detected in the Ishikawa cells in this study, they were not regulated by insulin in this system indicating potential species- or in vitro model-specific regulation of these glucose transporters. Moreover, the bovine system used a dual-chambered approach with underlying stromal cells and crosstalk between the different cells of the endometrium to alter the response to external cues. Future efforts will focus on developing human dual-cell systems with the possibility of using primary cells derived from patient samples. An additional limitation to this system is the fact the Ishikawa cells are tumour-derived and will not recapitulate in all scenarios the transcriptional landscape of the luminal epithelium of the non-tumourigenic endometrium. The ease with which human endometrial organoids (Turco et al. 2017) can be derived could add complexity to this system facilitating cross talk between cell types, however how difficult it would be to culture them underflow has not yet been established.
Despite numerous reports on the potential contribution of insulin as a metabolic stressor in obesity, the mechanism by which uterine receptivity may be compromised is still unknown. Existing literature studying this phenomenon and underlying mechanisms utilised hyperinsulinaemic animal models (Li et al. 2017, Chen et al. 2021), which may not accurately reflect the inner workings of human endometrium. Our findings aimed to provide insight into this gap in knowledge by using a microfluidics approach. Microfluidics can simulate shear stress and communications between cells, mimicking the constant flow of maternal circulation the endometrium is exposed to. Building up the different layers of the endometrium under flow conditions (similar to the approach used in the bovine model (De Bem et al. 2021)) will allow this to be modelled effectively using these types of devices (Campo et al. 2020). In conclusion, our results indicate that insulin alters the transcriptional profile of the endometrial epithelial cells when cultured in a microfluidics device. The biological processes and signalling pathways associated with differentially expressed transcripts found following insulin treatment were related to certain mechanisms of endometrial receptivity and implantation (Fig. 10). Specifically, genes and processes related to translation, immune response regulation, glutathione metabolism, and actin filament organisations seem to be dysregulated with insulin treatment, which may ultimately impair implantation success and reduce pregnancy success in those suffering from obesity.
Declaration of interest
SYB, HT, DW, DJA, and PTR have nothing to declare. NF is a current AE for the sister journal Reproduction and Council member for SRF.
Funding
NF’s lab is supported by funding from N8 agri‐food pump priming, QR GCRF, UN-CRP, as well as BBSRC grant numbers BB/R017522/1 and BB/X007367/1.
Author contribution statement
NF designed the study. SYB, HT, DJA, and PTR undertook lab work. DW, SYB, and PTR performed the analysis. SYB drafted the manuscript. All authors edited the manuscript.
Acknowledgements
RNA sequencing analysis was undertaken on ARC3, part of the High-Performance Computing facilities at the University of Leeds, UK. Figures 1 and 8 were created using Biorender. Research in the Forde group is supported by N8 agri‐food pump priming, QR GCRF, BBSRC grant number BB/R017522/1 as well as support from LTHT.
References
Adeoye O, Olawumi J, Opeyemi A & & Christiania O 2018 Review on the role of glutathione on oxidative stress and infertility. JBRA Assisted Reproduction 22 61–66. (https://doi.org/10.5935/1518-0557.20180003)
Agarwal A, Gupta S & & Sharma RK 2005 Role of oxidative stress in female reproduction. Reproductive Biology and Endocrinology 3 28–28. (https://doi.org/10.1186/1477-7827-3-28)
Bellver J, Melo MA, Bosch E, Serra V, Remohí J & & Pellicer A 2007 Obesity and poor reproductive outcome: the potential role of the endometrium. Fertility and Sterility 88 446–451. (https://doi.org/10.1016/j.fertnstert.2006.11.162)
Bellver J, Pellicer A, García-Velasco JA, Ballesteros A, Remohí J & & Meseguer M 2013 Obesity reduces uterine receptivity: clinical experience from 9,587 first cycles of ovum donation with normal weight donors. Fertility and Sterility 100 1050–1058. (https://doi.org/10.1016/j.fertnstert.2013.06.001)
Bellver J, Rossal LP, Bosch E, Zúñiga A, Corona JT, Meléndez F, Gómez E, Simón C, Remohí J & & Pellicer A 2003 Obesity and the risk of spontaneous abortion after oocyte donation. Fertility and Sterility 79 1136–1140. (https://doi.org/10.1016/s0015-0282(0300176-6)
Berlanga JJ, Baass A & & Sonenberg N 2006 Regulation of poly(A) binding protein function in translation: characterization of the Paip2 homolog, Paip2B. RNA 12 1556–1568. (https://doi.org/10.1261/rna.106506)
Bielfeld AP, Pour SJ, Poschmann G, Stühler K, Krüssel JS & & Baston-Büst DM 2019 A proteome approach reveals differences between fertile women and patients with repeated implantation failure on endometrial Level⁻Does hCG Render the Endometrium of RIF Patients? International Journal of Molecular Sciences 20 425. (https://doi.org/10.3390/ijms20020425)
Brewer CJ & & Balen AH 2010 The adverse effects of obesity on conception and implantation. Reproduction 140 347–364. (https://doi.org/10.1530/REP-09-0568)
Broughton DE & & Moley KH 2017 Obesity and female infertility: potential mediators of obesity's impact. Fertility and Sterility 107 840–847. (https://doi.org/10.1016/j.fertnstert.2017.01.017)
Campo H, Murphy A, Yildiz S, Woodruff T, Cervelló I & & Kim JJ 2020 Microphysiological modeling of the human endometrium. Tissue Engineering. Part A 26 759–768. (https://doi.org/10.1089/ten.tea.2020.0022)
Cano F, Landeras J, Molla M, Gomez E, Ballesteros A & & Remohí J 2001 The effect of extreme of body mass on embryo implantation at oocytes donation program. Fertility and Sterility 76 S160–S161. (https://doi.org/10.1016/S0015-0282(0102472-4)
Chen M, Li J, Zhang B, Zeng X, Zeng X, Cai S, Ye Q, Yang G, Ye C, Shang L, et al.2021 Uterine insulin sensitivity defects induced embryo implantation loss associated with mitochondrial dysfunction-triggered oxidative stress. Oxidative Medicine and Cellular Longevity 2021 6655685. (https://doi.org/10.1155/2021/6655685)
Daka B, Rosen T, Jansson PA, Råstam L, Larsson CA & & Lindblad U 2013 Inverse association between serum insulin and sex hormone-binding globulin in a population survey in Sweden. Endocrine Connections 2 18–22. (https://doi.org/10.1530/EC-12-0057)
De Bem THC, Tinning H, Vasconcelos EJR, Wang D & & Forde N 2021 Endometrium on-a-chip reveals insulin- and glucose-induced alterations in the transcriptome and proteomic secretome. Endocrinology 162. (https://doi.org/10.1210/endocr/bqab054)
D'ippolito S, Di Nicuolo F, Papi M, Castellani R, Palmieri V, Masciullo V, Arena V, Tersigni C, Bernabei M, Pontecorvi A, et al.2020 Expression of pinopodes in the endometrium from recurrent pregnancy loss women. Role of thrombomodulin and Ezrin. Journal of Clinical Medicine 9 2634. (https://doi.org/10.3390/jcm9082634)
Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M & & Gingeras TR 2013 STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29 15–21. (https://doi.org/10.1093/bioinformatics/bts635)
Duman İ, Tiftik RN & & Ün İ 2021 Effects of vitamin D analogs alfacalcidol and calcitriol on cell proliferation and migration of HEC1A endometrial adenocarcinoma cells. Nutrition and Cancer 73 273–281. (https://doi.org/10.1080/01635581.2020.1764066)
Frankish A, Diekhans M, Ferreira AM, Johnson R, Jungreis I, Loveland J, Mudge JM, Sisu C, Wright J, Armstrong J, et al.2019 GENCODE reference annotation for the human and mouse genomes. Nucleic Acids Research 47 D766–D773. (https://doi.org/10.1093/nar/gky955)
Hakki Kalkan I & & Suher M 2013 The relationship between the level of glutathione, impairment of glucose metabolism and complications of diabetes mellitus. Pakistan Journal of Medical Sciences 29 938–942. (https://doi.org/10.12669/pjms.294.2859)
Herrmann JM & & Riemer J 2021 Apoptosis inducing factor and mitochondrial NADH dehydrogenases: redox-controlled gear boxes to switch between mitochondrial biogenesis and cell death. Biological Chemistry 402 289–297. (https://doi.org/10.1515/hsz-2020-0254)
Howards PP & & Cooney MA 2008 Disentangling causal paths between obesity and in vitro fertilization outcomes: an intractable problem? Fertility and Sterility 89 1604–1605. (https://doi.org/10.1016/j.fertnstert.2007.03.093)
Howe KL, Achuthan P, Allen J, Allen J, Alvarez-Jarreta J, Amode MR, Armean IM, Azov AG, Bennett R, Bhai J, et al.2020 Ensembl 2021. Nucleic Acids Research 49 D884–D891.
Huang DW, Sherman BT & & Lempicki RA 2009 Systematic and integrative analysis of large gene lists using David bioinformatics resources. Nature Protocols 4 44–57. (https://doi.org/10.1038/nprot.2008.211)
Hume L, Edge JC, Tinning H, Wang D, Taylor AS, Ovchinnikov V, Geijer-Simpson AV, Vrljicak P, Brosens JJ, Lucas ES, et al.2023 MicroRNAs emerging coordinate with placental mammals alter pathways in endometrial epithelia important for endometrial function. iScience 26 106339. (https://doi.org/10.1016/j.isci.2023.106339)
Jin XY, Zhao LJ, Luo DH, Liu L, Dai YD, Hu XX, Wang YY, Lin X, Hong F, Li TC, et al.2017 Pinopode score around the time of implantation is predictive of successful implantation following frozen embryo transfer in hormone replacement cycles. Human Reproduction 32 2394–2403. (https://doi.org/10.1093/humrep/dex312)
Johnson SC, Rabinovitch PS & & Kaeberlein M 2013 mTOR is a key modulator of ageing and age-related disease. Nature 493 338–345. (https://doi.org/10.1038/nature11861)
Kageyama S II, Ii H, Taniguchi K, Kubota S, Yoshida T, Isono T, Chano T, Yoshiya T, Ito K, Yoshiki T, et al.2018 Mechanisms of tumor growth inhibition by depletion of γ-glutamylcyclotransferase (GGCT): A novel molecular target for anticancer therapy. International Journal of Molecular Sciences 19 2054. (https://doi.org/10.3390/ijms19072054)
Kim ST & & Moley KH 2009 Regulation of facilitative glucose transporters and AKT/MAPK/PRKAA signaling via estradiol and progesterone in the mouse uterine epithelium. Biology of Reproduction 81 188–198. (https://doi.org/10.1095/biolreprod.108.072629)
Kobayashi H, Matsuda M, Fukuhara A, Komuro R & & Shimomura I 2009 Dysregulated glutathione metabolism links to impaired insulin action in adipocytes. American Journal of Physiology. Endocrinology and Metabolism 296 E1326–E1334. (https://doi.org/10.1152/ajpendo.90921.2008)
Lédée N, Petitbarat M, Chevrier L, Vitoux D, Vezmar K, Rahmati M, Dubanchet S, Gahéry H, Bensussan A & & Chaouat G 2016 The uterine immune profile may help women with repeated unexplained embryo implantation failure after in vitro fertilization. American Journal of Reproductive Immunology: New York, N.Y. 75 388–401. (https://doi.org/10.1111/aji.12483)
Lee II, Maniar K, Lydon JP & & Kim JJ 2016 Akt regulates progesterone receptor B-dependent transcription and angiogenesis in endometrial cancer cells. Oncogene 35 5191–5201. (https://doi.org/10.1038/onc.2016.56)
Lei S, Li L, Yang X, Yin Q, Xu T, Zhou W, Gu W, Ma F & & Yang R 2021 The association between RAPSN methylation in peripheral blood and breast cancer in the Chinese population. Journal of Human Genetics 66 1069–1078. (https://doi.org/10.1038/s10038-021-00933-x)
Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, Durbin R & 1000 Genome Project Data Processing Subgroup 2009 The Sequence Alignment/Map format and SAMtools. Bioinformatics 25 2078–2079. (https://doi.org/10.1093/bioinformatics/btp352)
Li R, Wu J, He J, Wang Y, Liu X, Chen X, Tong C, Ding Y, Su Y, Chen W, et al.2017 Mice endometrium receptivity in early pregnancy is impaired by maternal hyperinsulinemia. Molecular Medicine Reports 15 2503–2510. (https://doi.org/10.3892/mmr.2017.6322)
Liao Y, Smyth GK & & Shi W 2019a The R package Rsubread is easier, faster, cheaper and better for alignment and quantification of RNA sequencing reads. Nucleic Acids Research 47 e47–e47. (https://doi.org/10.1093/nar/gkz114)
Liao Y, Wang J, Jaehnig EJ, Shi Z & & Zhang B 2019b WebGestalt 2019: gene set analysis toolkit with revamped UIs and APIs. Nucleic Acids Research 47 W199–W205. (https://doi.org/10.1093/nar/gkz401)
Livak KJ & & Schmittgen TD 2001 Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25 402–408. (https://doi.org/10.1006/meth.2001.1262)
Love MI, Huber W & & Anders S 2014 Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biology 15 550–550. (https://doi.org/10.1186/s13059-014-0550-8)
Marques C, Oliveira CSF, Alves S, Chaves SR, Coutinho OP, Côrte-Real M & & Preto A 2013 Acetate-induced apoptosis in colorectal carcinoma cells involves lysosomal membrane permeabilization and cathepsin D release. Cell Death and Disease 4 e507–e507. (https://doi.org/10.1038/cddis.2013.29)
Martin M 2011 Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet.Journal 17 3. (https://doi.org/10.14806/ej.17.1.200)
Mok KW, Mruk DD & & Cheng CY 2013 Regulation of blood-testis barrier (BTB) dynamics during spermatogenesis via the "Yin" and "Yang" effects of mammalian target of rapamycin complex 1 (mTORC1) and mTORC2. International Review of Cell and Molecular Biology 301 291–358. (https://doi.org/10.1016/B978-0-12-407704-1.00006-3)
Okabe S, Okamoto T, Zhao CM, Chen D & & Matsui H 2014 Acetic acid induces cell death: an in vitro study using normal rat gastric mucosal cell line and rat and human gastric cancer and mesothelioma cell lines. Journal of Gastroenterology and Hepatology 29(Supplement 4) 65–69. (https://doi.org/10.1111/jgh.12775)
Pasquali R, Pelusi C, Genghini S, Cacciari M & & Gambineri A 2003 Obesity and reproductive disorders in women. Human Reproduction Update 9 359–372. (https://doi.org/10.1093/humupd/dmg024)
Pham-Huy LA, He H & & Pham-Huy C 2008 Free radicals, antioxidants in disease and health. International Journal of Biomedical Science : IJBS 4 89–96.
Prévôt D, Darlix JL & & Ohlmann T 2003 Conducting the initiation of protein synthesis: the role of eIF4G. Biology of the Cell 95 141–156. (https://doi.org/10.1016/s0248-4900(0300031-5)
Qiao R, Di F, Wang J, Wei Y, Zhang Y, Xu T, Wang Y, Gu W, Han B & & Yang R 2020 The association between RAPSN methylation in peripheral blood and early stage lung cancer detected in case-control cohort. Cancer Management and Research 12 11063–11075. (https://doi.org/10.2147/CMAR.S275321)
Qiong Z, Jie H, Yonggang W, Bin X, Jing Z & & Yanping L 2017 Clinical validation of pinopode as a marker of endometrial receptivity: a randomized controlled trial. Fertility and Sterility 108 513–517.e2. (https://doi.org/10.1016/j.fertnstert.2017.07.006)
Quinn KE, Matson BC, Wetendorf M & & Caron KM 2020 Pinopodes: recent advancements, current perspectives, and future directions. Molecular and Cellular Endocrinology 501 110644–110644. (https://doi.org/10.1016/j.mce.2019.110644)
Robertson SA, Care AS & & Moldenhauer LM 2018 Regulatory T cells in embryo implantation and the immune response to pregnancy. Journal of Clinical Investigation 128 4224–4235. (https://doi.org/10.1172/JCI122182)
Saladi S, Boos F, Poglitsch M, Meyer H, Sommer F, Mühlhaus T, Schroda M, Schuldiner M, Madeo F & & Herrmann JM 2020 The NADH dehydrogenase Nde1 executes cell death after integrating signals from metabolism and proteostasis on the mitochondrial surface. Molecular Cell 77 189–202.e6. (https://doi.org/10.1016/j.molcel.2019.09.027)
Salmond RJ, Emery J, Okkenhaug K & & Zamoyska R 2009 MAPK, phosphatidylinositol 3-kinase, and mammalian target of rapamycin pathways converge at the level of ribosomal protein S6 phosphorylation to control metabolic signaling in CD8 T cells. Journal of Immunology 183 7388–7397. (https://doi.org/10.4049/jimmunol.0902294)
Salmond RJ, Brownlie RJ, Meyuhas O & & Zamoyska R 2015 Mechanistic target of rapamycin Complex 1/S6 kinase 1 signals influence T cell activation independently of ribosomal protein S6 phosphorylation. Journal of Immunology 195 4615–4622. (https://doi.org/10.4049/jimmunol.1501473)
Saxton RA & & Sabatini DM 2017 MTOR signaling in growth, metabolism, and disease. Cell 168 960–976. (https://doi.org/10.1016/j.cell.2017.02.004)
Schmieder R & & Edwards R 2011 Quality control and preprocessing of metagenomic datasets. Bioinformatics 27 863–864. (https://doi.org/10.1093/bioinformatics/btr026)
Schulte MMB, Tsai JH & & Moley KH 2015 Obesity and PCOS: the effect of metabolic derangements on endometrial receptivity at the time of implantation. Reproductive Sciences 22 6–14. (https://doi.org/10.1177/1933719114561552)
Shanik MH, Xu Y, Skrha J, Dankner R, Zick Y & & Roth J 2008 Insulin resistance and hyperinsulinemia: is hyperinsulinemia the cart or the horse? Diabetes Care 31(Supplement 2) S262–S268. (https://doi.org/10.2337/dc08-s264)
Silvestris E, De Pergola G, Rosania R & & Loverro G 2018 Obesity as disruptor of the female fertility. Reproductive Biology and Endocrinology: 16 22–22. (https://doi.org/10.1186/s12958-018-0336-z)
Styne-Gross A, Elkind-Hirsch K & & Scott RT 2005 Obesity does not impact implantation rates or pregnancy outcome in women attempting conception through oocyte donation. Fertility and Sterility 83 1629–1634. (https://doi.org/10.1016/j.fertnstert.2005.01.099)
Talmor A & & Dunphy B 2015 Female obesity and infertility. Best Practice and Research. Clinical Obstetrics and Gynaecology 29 498–506. (https://doi.org/10.1016/j.bpobgyn.2014.10.014)
Tamm-Rosenstein K, Simm J, Suhorutshenko M, Salumets A & & Metsis M 2013 Changes in the transcriptome of the human endometrial Ishikawa cancer cell line induced by estrogen, progesterone, tamoxifen, and mifepristone (RU486) as detected by RNA-sequencing. PLoS One 8 e68907. (https://doi.org/10.1371/journal.pone.0068907)
Taylor AS, Tinning H, Ovchinnikov V, Edge J, Smith W, Pullinger AL, Sutton RA, Constantinides B, Wang D,, Forbes K, Forde N & & O’Connell MJ 2021.07.22.453388 A burst of regulatory and protein innovation at the origin of placental mammals drove the emergence of placenta and underpins divergent early pregnancy strategies in modern mammals. Communications Biology 6 459. (https://doi.org/10.1038/s42003-023-04809-y)
Turco MY, Gardner L, Hughes J, Cindrova-Davies T, Gomez MJ, Farrell L, Hollinshead M, Marsh SGE, Brosens JJ, Critchley HO, et al.2017 Long-term, hormone-responsive organoid cultures of human endometrium in a chemically defined medium. Nature Cell Biology 19 568–577. (https://doi.org/10.1038/ncb3516)
Ullah K, Rahman TU, Pan HT, Guo MX, Dong XY, Liu J, Jin LY, Cheng Y, Ke ZH, Ren J, et al.2017 Serum estradiol levels in controlled ovarian stimulation directly affect the endometrium. Journal of Molecular Endocrinology 59 105–119. (https://doi.org/10.1530/JME-17-0036)
Wang C, Feng Y, Zhou WJ, Cheng ZJ, Jiang MY, Zhou Y & & Fei XY 2021 Screening and identification of endometrial proteins as novel potential biomarkers for repeated implantation failure. PeerJ 9 e11009. (https://doi.org/10.7717/peerj.11009)
Wang Y, Zhu L, Kuokkanen S & & Pollard JW 2015 Activation of protein synthesis in mouse uterine epithelial cells by estradiol-17β is mediated by a PKC-ERK1/2-mTOR signaling pathway. PNAS 112 E1382–E1391. (https://doi.org/10.1073/pnas.1418973112)
Wattanakumtornkul S, Damario MA, Stevens Hall SA, Thornhill AR & & Tummon IS 2003 Body mass index and uterine receptivity in the oocyte donation model. Fertility and Sterility 80 336–340. (https://doi.org/10.1016/s0015-0282(0300595-8)
Wu G, Fang YZ, Yang S, Lupton JR & & Turner ND 2004 Glutathione metabolism and its implications for health. Journal of Nutrition 134 489–492. (https://doi.org/10.1093/jn/134.3.489)
Xu X, Leng JY, Gao F, Zhao ZA, Deng WB, Liang XH, Zhang YJ, Zhang ZR, Li M, Sha AG, et al.2014 Differential expression and anti-oxidant function of glutathione peroxidase 3 in mouse uterus during decidualization. FEBS Letters 588 1580–1589. (https://doi.org/10.1016/j.febslet.2014.02.043)
