Expression of HOXA11 in the mid-luteal endometrium from women with endometriosis-associated infertility

Since advanced endometriosis may lead to anatomical distortions and adhesions resulting in infertility, we focused our study on infertility associated with minimal endometriosis. Data for women with infertility-associated endometriosis, fertile women and infertile patients with fallopian tubal occlusion were randomly collected in the Gynecologic and Obstetrical University Hospital, Division of Reproduction in Poznan, Poland (Table 1). Women with endometriosis and fertile women were examined for the cause of infertility, suspected pelvic endometriosis, or chronic pelvic pain. Then they were divided into eighteen infertile women with minimal endometriosis and sixteen fertile women. Minimal endometriosis in infertile women was diagnosed based on visualization of endometriotic lesions and histopathologic criteria (Table 1). The stage of endometriosis was evaluated according to the revised classification of the American Society for Reproductive Medicine [24]. The studied women with endometriosis displayed no anatomical changes in the reproductive tract. The women with endometriosis and with fallopian tubal occlusion exhibited regular menses and a minimum 1 year of infertility with a current desire for conception, and no contribution of male factor infertility. The fertile women assigned to the control group exhibited chronic pelvic pain without any pelvic abnormalities determined by laparoscopy. The fertile women were diagnosed as having varicose veins in the pelvic floor but no signs of past or present inflammation. These fertile women had at least one child born no later than 2 years before laparoscopy, regular menses, and no anatomical changes in the reproductive tract (Table 1). The second control group included the women with fallopian tubal occlusion diagnosed based on hysterosalpingography and subsequently verified by methyl blue administration to fallopian tubes during laparoscopy [25]. Furthermore, hysteroscopy and pipelle biopsy from women with endometriosis, and fertile women and infertile women with tubal occlusion were respectively employed for histopathologic evaluation to exclude individuals with pathological endometrium. All participating individuals had not used oral contraception, hormonal therapy, or an intrauterine device for half a year prior to the endometrial biopsy. Fertile women and infertile women with tubal occlusion were matched by age to the patients with endometriosis and all individuals were Caucasian from the same region of Poland (Table 1). Written informed agreement was obtained from all participating individuals. The procedures of the study were approved by the Local Ethical Committee of Poznań University of Medical Sciences. All biopsy specimens were collected during the middle secretory phase based on the endometrial dating criteria of Noyes et al. [26]. Sample of eutopic endometrial tissue from patients and controls was respectively obtained by pipelle or hysteroscopic biopsy during the implantation window, i.e. 7-9 days after ultrasound-confirmed ovulation. The eutopic endometrium samples were then used for total RNA, protein, and DNA isolation.

Goat polyclonal (Gp) anti-HOXA11 antibodies (Ab) (C-16, 200 μg/1.0 ml), donkey anti-goat horseradish peroxidase (HRP)-conjugated Ab (200 μg/0.5 ml), and anti-actin HRP-conjugated Ab (clone I-19, 200 μg/1.0 ml were provided by Santa Cruz Biotechnology (Santa Cruz, CA).

Total RNA was isolated by RNeasy Protect Mini Kit Qiagen (Hilden, Germany). RNA samples were treated with DNase I from DNase Set Qiagen (Hilden, Germany) and quantified. A volume of 0.5 μg RNA was reverse-transcribed into cDNA using QuantiTect Reverse Transcription Kit and oligo-dT primers from Qiagen (Hilden, Germany). The efficiency of reverse transcription was controlled by a series of RNA dilution-reverse-transcribed and comparison of RQ-PCR Ct differences for cDNA samples [27].

RQ-PCR was conducted in a Corbett Research Rotor-Gene 3000 thermocycler (Mortlake, Australia). Target cDNA was quantified using relative quantification method employing a calibrator. The calibrator was prepared as a cDNA mix from all patients and control samples and consecutive dilutions were used to create a standard curve as provided in Relative Quantification Manual Roche Diagnostics GmbH (Mannheim, Germany). The quantity of HOXA11, DNMT1, DNMT3A and DNMT3B transcript in each sample was standardized by human glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and β-actin (ACTB) transcript levels. For amplification, 60 ng cDNA solution was added to 18 μl (total 20 μl) of DyNAmo HS SYBR^® Green qPCR Kit from Finnzymes (Espoo, Finland) and primers (Additional file 1, Table S1). One RNA sample of each preparation was processed without the reverse transcription (RT)-reaction to provide a negative control in subsequent PCR. The quantity of HOXA11 transcripts in each sample was standardized by GAPDH and ACTB cDNA. Each sample was determined in triplicate and the HOXA11, DNMT1, DNMT3A and DNMT3B mRNA levels were expressed as the multiplicity of these cDNA concentrations in the calibrator.

Tissue samples were minced in liquid nitrogen followed by lysis in RIPA buffer. Next, 30 μg of protein were resuspended in sample buffer and separated on 12% Tris-glycine gel using the SDS-PAGE system. Gel proteins were transferred to PVDF membrane, which was blocked with 5% milk in Tris-buffered saline/Tween. Immunodetection was performed with Gp anti-HOXA11 Ab (C-16, 1:500 dilution, 4.0 μg) followed by incubation with donkey anti-goat horseradish peroxidase (HRP)-conjugated Ab (1:3000 dilution, 1.3 μg). The membranes were then stripped and incubated with anti-β-actin HRP-conjugated Ab (clone I-19, 1:3000 dilution, 0.66 μg) to ensure equal protein loading of the lanes. Bands were revealed using ECL kit and Hyperfilm ECL Amersham (Piscataway, NJ). The quantities of western blot-detected HOXA11 and β-actin proteins were determined based on the band optical density. The band densitometry readings were normalized to β-actin loading control to calculate the HOXA11-to-β-actin optical density ratio.

Genomic DNA was isolated by the salting out method [28], and DNA cytosine bases were converted to uracil using the EZ DNA Methylation Kit™ procedure from Zymo Research Corporation (Orange, CA). The locations of CpG island in regions I, II, and III of the HOXA11 gene (Additional file 2, Figure S1) was determined based on two online programs [29, 30].

The HOXA11 regions I, II, and III were amplified from the bisulfite-modified DNA by the three pairs of primers complementary to the bisulfite-DNA modified sequence (Additional files 1, Table S1; Additional file 2, Figure S1). PCR amplification was conducted by FastStart Taq DNA Polymerase from Roche Diagnostic GmbH (Mannheim, Germany). The PCR products were purified using Agarose Gel DNA Extraction Kit Roche (Mannheim, Germany) with subsequent cloning into pGEM-T Easy Vector System I Promega (Madison, WI) and transformation into TOPO10 E. coli strain cells. Plasmid DNA isolated from ten positive bacterial clones was used for commercial sequencing of the cloned fragment of DNA. The results of bisulphite sequencing were assessed and presented using BiQ analyzer software [31] and BDPC web server [32].

Statistical analysis was conducted by Systat Software Inc (2006). SIGMASTAT (data analysis software system) version 3.5 [33]. Data groups were analyzed by Mann-Whitney Rank Test to evaluate if there was significance (P < 0.05) between the groups.

We used RQ-PCR and western blotting analysis to evaluate HOXA11 transcript and protein levels, respectively, in eutopic mid-secretory endometrium from infertile women with endometriosis, fertile women and women with tubal occlusion. We observed significantly lower levels of HOXA11 transcript in women with endometriosis as compared to fertile women (p = 0.003) and women with tubal occlusion (p = 0.041) (Table 2, Figure 1A). We also found significantly reduced HOXA11 protein levels in eutopic endometrium from infertile women with endometriosis than in fertile women (p = 0.004) and women with tubal occlusion (p = 0.001) (Table 2, Figures 1B and 1C).

RQ-PCR analysis showed significantly increased levels of DNMT3A transcript in eutopic mid-secretory endometrium from women with endometriosis compared to fertile women (p = 0.044) and women with tubal occlusion (p = 0.047) (Table 3 and Figure 2A). However, we did not observe significant differences in DNMT1 (p = 0.621, p = 0.470) and DNMT3B (p = 0.717, p = 0.100) transcript levels between the investigated groups (Table 3 and Figures 2B and 2C).

We performed sodium bisulfite DNA sequencing of HOXA11 regions I, II, and III (Additional file 2, Figure S1). There was no DNA methylation in HOXA11 regions I and III in all infertile women with endometriosis and all fertile women and women with tubal occlusion. However, we found significantly higher methylation levels of HOXA11 region in II the in eutopic mid-secretory endometrium obtained from infertile women with endometriosis as compared to material obtained from the fertile women (p < 0.001) and women with tubal occlusion (p < 0.001) (Figure 3). In the group of eighteen infertile women with endometriosis, we found fifteen (83.3%) individuals with methylation in HOXA11 region II (Figure 3). By contrast, in both fertile women and women with tubal occlusion we found one (6.2%) subject with DNA methylation in HOXA11 region II (Figure 3).