Potential Role of Semaphorin 3A and Its Receptors in Regulating Aberrant Sympathetic Innervation in Peritoneal and Deep Infiltrating Endometriosis

Endometriosis is a chronic, inflammatory, estrogen-dependent benign gynecological disease characterized by the presence of endometrial glands and stromal cells outside the uterine cavity. It is often associated with infertility and pelvic pain, affecting 10–15% of women of childbearing age worldwide [ 1 ]. The relationship between endometriosis and pain is still poorly understood. And the mechanisms underlying endometriosis-associated-pain remain to be further investigated. Recent studies have shown endometriotic lesions develop newly formed sensory and autonomic nerve fibers in rat models and in women with endometriosis [ 2 – 4 ], often in close contact with endometriotic implants and accompany with the immature blood vessels that vascularize the ectopic growth [ 5 ]. Tokushige et al. demonstrated that protein gene product (PGP) 9.5-immunoactive nerve fibers were present in peritoneal or deep infiltrating endometriotic lesions [ 1 ]. The density of endometriosis-associated nerve fibers were correlative with the severity of pain in women with endometriosis [ 6 ], suggesting a direct association between pain and nerve fiber density. An imbalance in sympathetic/sensory innervation in the inflamed area in women with peritoneal and intestinal endometriosis was demonstrated recently. They found significant loss of sympathetic innervation in the area near the peritoneal and intestinal endometriotic lesions, where the sensory nerve density remained unchanged or increased [ 7 – 8 ]. This aberrant sympathetic and sensory innervation is proposed to be an adaptive program in order to maintain the balance of pro-inflammatory and anti-inflammatory effects. This is in consistent with the observations in other autoimmune and chronic inflammatory disease, such as rheumatoid arthritis and Crohn disease [ 9 – 10 ]. However, the underlying pathophysiology of this interesting phenomenon is not well understood. Studies have shown that several nerve repellent factors (Semaphorin 3A, Semaphorin 3C, Semaphorin 3F, neuropilin 2, etc.) may play an important role in regulating the imbalanced innervation of different inflammatory diseases [ 11 – 13 ]. Accordingly, we speculate that semaphorin 3A (Sema 3A), a member of these nerve repellent factors, may also be involved in modulating the aberrant innervation within endometriotic lesions. Semaphorins are a group of evolutionarily highly conserved surface or locally secreted nerve repellent factors that not only steer and fasciculate axons in the developing nervous system [ 14 ] but also regulate both developmental [ 14 ] and tumor angiogenesis [ 15 – 17 ]. Sema 3A, a member of secreted class 3 semaphorins, is well reported as a potent chemorepellent that restricts axonal elongation and causes growth cone collapse [ 18 ] through binding to its specific receptor, neuropilin-1 (NRP-1), which is located on the surface of the target neurons. It is suggested that NRP-1 must bind with Plexin A1 [ 19 ], another receptor to Sema 3A, in order to form a multimeric holoreceptor signaling complex that triggers the down-stream signaling cascade [ 20 – 22 ]. Tolofari et al. demonstrated that with the reduced expression of Sema 3A in degenerated intervertebral disc, there was a disinhibition of neural ingrowth, thus resulting in increased density of PGP9.5 positively stained nerve fibers in the degenerative disc [ 23 ]. In addition, the increased uterine Sema 3A expression during pregnancy is associated with reduced sympathetic innervation [ 24 ]. Sema 3A has also been strongly implicated as an inhibitory factor determining density of sympathetic innervation of blood vessels [ 25 ]. It was reported that pro-inflammatory factors initiating up-regulation of Sema 3A was associated with decreased density of sympathetic and sensory nerve fibers in benign colorectal adenomatous polyps [ 13 ]. As is known to all, endometriosis is a chronic pelvic inflammatory disease. Many studies have shown that different expression of various cytokines or chemokines as well as infiltration of inflammatory cells in peritoneal microenvironment. And endometriotic lesions contribute to the inflammatory condition of endometriosis. We hypothesize that Sema 3A, a nerve repellent factor, may play a potential role in regulating the aberrant sympathetic innervation in peritoneal and deep infiltrating endometriosis. The present study tried to examine the expression of the chemorepulsive Sema 3A and its receptors (NRP-1 and Plexin A1) in peritoneal and deep infiltrating endometriosis first. And then we further investigated the distribution of PGP9.5-immunoactive or Tyrosine Hydroxylase-immunoactive sympathetic nerve fibers in/near the endometriotic lesions and analyze the association between expression of Sema 3A and the density of these nerve fibers.

Sema 3A, NRP-1 and Plexin A1 were expressed in both the endometrial glandular epithelial cells and stromal cells ( Fig 1A , S4 , S5 , S6 and S7 Figs). For qualitative immunostaining, Sema 3A, NRP-1 and Plexin A1 were stained positively in all of the detected sections. Higher immunostaining of Sema 3A and Plexin A1 were found in the eutopic endometrial glandular epithelial cells from patients with endometriosis (p = 0.041 and p = 0.025, respectively) than those without endometriosis. But the expression of Sema 3A as well as Plexin A1 between the stromal cells from patients with endometriosis and those without endometriosis were not significantly different (p = 0.067 and p = 0.319, respectively). There was no significant difference between the HSCORE of NRP-1 of eutopic endometrial glandular epithelial cells from patients with endometriosis and that from patients without endometriosis (p = 0.120), but the HSCORE of NRP-1 of stromal cells from patients with endometriosis was obviously higher than that from patients without endometriosis (p = 0.002). The semi-quantitative immunostaining of Sema 3A, NRP-1 and Plexin A1 in the endometrial samples using HSCORE analysis is shown in S1 Table ( Fig 1B and 1C ). A: Immunohistochemical staining of three proteins in eutopic endometrium. EuE-EM: Eutopic Endometrium from patients with Endometriosis; EuE-NEM: Eutopic Endometrium from patients without Endometriosis (immunohistochemical stain, 200 × magnification). B and C: Semi-quantitative expression of Sema 3A, Plexin A1 and NRP-1 of glandular epithelial cells (B) as well as stromal cells (C) from eutopic endometrium from patients with endometriosis (EM) and without endometriosis (NEM). * P <0.05 (Mann–Whitney). Sema 3A, NRP-1 and Plexin A1 immunostaining were found 100% of peritoneal and deep infiltrating endometriosis samples ( Fig 2C ). Sema 3A immunostaining was higher in glandular epithelial cells of peritoneal endometriosis (HSCORE = 356.77±74.3) compared with glandular epithelial cells of the endometrium from women with endometriosis (HSCORE = 309.09±66.61, P <0.001) or glandular epithelial cells of the endometrium from women without endometriosis (HSCORE = 271.15±49.34, P <0.001, S12 Fig ). The expression of both NRP-1 and Plexin A1 in glandular epithelial cells of peritoneal endometriosis were also higher than that from eutopic endometrium of women with endometriosis as well as women without endometriosis (all the p values were less than 0.001) ( Fig 2A ). A: Comparison of semi-quantitative expression of Sema 3A, Plexin A1 and NRP-1 of glandular epithelial cells between peritoneal endometriosis, deep infiltrating endometriosis of uterosacral ligament and eutopic endometrium from patients with endometriosis (EM) as well as without endometriosis (NEM). B: Comparison of semi-quantitative expression of Sema 3A, Plexin A1 and NRP-1 of stromal cells between peritoneal endometriosis, deep infiltrating endometriosis of uterosacral ligament and eutopic endometrium from patients with endometriosis (EM) as well as without endometriosis (NEM). C: Immunohistochemical staining of Sema 3A, Plexin A1 and NRP-1 in PEM and USL-EM. PEM: peritoneal endometriosis; USL-EM: deep infiltrating endometriosis of uterosacral ligament. (immunohistochemical stain, 200 × magnification) Mann-Whitney test: * P <0.05 (versus NEM); ** P <0.001 (versus EM); § P <0.05 (versus NEM); §§ P <0.001(versus EM); NS, not significant. On the other hand, Sema 3A immunostaining in stromal cells of peritoneal endometriosis ( S9 Fig ) (HSCORE = 158.33±94.60) was significantly lower than that from eutopic endometrium of women with endometriosis (HSCORE = 236.36±49.24, P = 0.002) as well as women without endometriosis (HSCORE = 242.31±48.36, P <0.001) ( Fig 2B ). This lower expression was also observed for Plexin A1 ( S11 Fig ) (p = 0.005, p = 0.005, respectively). The expression of NRP-1 in stromal cells of peritoneal endometriosis was lower than the stromal cells of endometrium of women with endometriosis (p = 0.001), while there was no difference between stromal cells of peritoneal endometriosis and eutopic endometrium of women without endometriosis (p = 0.082). Relative to glandular epithelial cells of deep infiltrating endometriotic lesions of uterosacral ligament, glandular epithelial cells of eutopic endometrium from women with or without endometriosis showed lower expression of Sema 3A (p = 0.028, p<0.001, respectively), NRP-1 (p<0.001, p<0.001, respectively) and Plexin A1 (p<0.001, p<0.001, respectively) ( Fig 2A ). Similarly, Sema 3A immunostaining in stromal cells of deep infiltrating endometriosis ( S8 Fig ) (HSCORE = 143.75±109.99; p<0.001, p<0.001, respectively) was lower than that from eutopic endometrium of women with endometriosis as well as women without endometriosis. This lower expression was also observed for Plexin A1 ( S10 Fig ) (p<0.001, p<0.001, respectively). But no differences were observed for NRP-1 ( Fig 2B ). PGP 9.5 is a highly specific pan-neuronal marker for both myelinated and unmyelinated nerve fibers, including Aα, Aβ, Aγ, Aδ, B and C fibers [ 29 ], while Tyrosine Hydroxylase (TH) is a specific marker for sympathetic nerve fibers. Anti-PGP 9.5 and Anti-TH positively stained nerve fibers were observed in peritoneal endometriotic lesions, para-endometriotic tissue and healthy peritoneum ( S13 Fig ). The nerve fiber density (NFD) of Anti-PGP 9.5 (+) endometriosis-associated nerve of peritoneal endometriosis (EAN-PEM) was significantly higher than NFD of para-endometriotic nerve of peritoneal endometriosis (PEN-PEM) and NFD of nerve of peritoneum of control (N-PC) (p<0.001, p = 0.002, respectively) ( Fig 3A , S3 Table ). NFD of Anti-TH (+) endometriosis-associated sympathetic nerve of peritoneal endometriosis (EASN-PEM) was significantly lower than NFD of para-endometriotic sympathetic nerve of peritoneal endometriosis (PESN-PEM) (p<0.001). And NFD of PESN-PEM was also lower than that of sympathetic nerve of peritoneum of control (SN-PC) (p = 0.028) ( Fig 3B , S4 Table ). (A) Total nerve fibers were positively stained by Anti-PGP 9.5 antibody. Comparison of the total nerve fiber density (NFD, NF/mm 2 ) of endometriosis-associated nerve of peritoneal endometriosis (EAN-PEM), NFD of para-endometriotic nerve of peritoneal endometriosis (PEN-PEM) and NFD of nerve of peritoneum of control (N-PC). (B) Sympathetic nerve fibers were positively stained by Anti-TH antibody. Comparison of endometriosis-associated sympathetic NFD of peritoneal endometriosis (ESAN-PEM), para-endometriotic sympathetic NFD of peritoneal endometriosis (PESN-PEM) and sympathetic NFD of peritoneum of control (SN-PC). * P <0.05. Similarly, Anti-PGP 9.5 and Anti-TH positively stained nerve fibers were observed in uterosacral endometriotic lesions, para-endometriotic tissue and healthy uterosacral ligament ( S13 Fig ). The NFD of Anti-PGP 9.5 (+) endometriosis-associated nerve of uterosacral ligament endometriosis (EAN-USL-EM) was significantly higher than NFD of para-endometriotic nerve of uterosacral ligament endometriosis (PEN-USL-EM) as well as NFD of nerve of uterosacral ligament of control (N-USL-C) (p<0.001, p<0.001, respectively) ( Fig 4A , S5 Table ). NFD of Anti-TH (+) endometriosis-associated sympathetic nerve of uterosacral ligament endometriosis (EASN-USL-EM) was also lower than NFD of para-endometriotic sympathetic nerve of uterosacral ligament endometriosis (PESN-USL-EM) (p<0.001). And NFD of PESN-USL-EM was also lower than that of sympathetic nerve of uterosacral ligament of control (SN-USL-C) (p = 0.012) ( Fig 4B , S6 Table ) ( Fig 5 ). (A) Total nerve fibers were positively stained by Anti-PGP 9.5 antibody. Comparison of the total nerve fiber density (NFD, NF/mm 2 ) of endometriosis-associated nerve of deep infiltrating endometriosis of uterosacral ligament (EAN-USL-EM), NFD of para-endometriotic nerve of deep infiltrating endometriosis of uterosacral ligament (PEN-USL-EM) and NFD of nerve of uterosacral ligament of control (N-USL-C). (B) Sympathetic nerve fibers were positively stained by Anti-TH antibody. Comparison of endometriosis-associated sympathetic NFD of USL-EM (ESAN-USL-EM), para-endometriotic sympathetic NFD of USL-EM (PESN-USL-EM) and sympathetic NFD of uterosacral ligament of control (SN-USL-C). * P <0.05. Total nerve fibers were orange-yellow stained by Anti-PGP 9.5 antibody (A: EAN-USL-EM; B: PEN-USL-EM; C: N-USL-C). Sympathetic nerve fibers were also orange-yellow stained by Anti-TH antibody (D: EASN-USL-EM; E: PESN-USL-EM; F: SN-USL-C). Ectopic endometrial glandular epithelial cells was stained in green by anti-Sema 3A antibody staining (A was the merge image of double staining of both Sema 3A and PGP 9.5, D was the merge image of double staining of both Sema 3A and TH), nuclei were stained in blue by DAPI staining. White triangle: orange-yellow stained nerve fibers. (Original magnification 200×) Patients with peritoneal endometriosis or deep infiltrating endometriosis of uterosacral ligament were divided into three groups: mild pain group, moderate pain group and severe pain group. We found out that as the aggravation of the grade of dysmenorrhea, the expression of Sema 3A in stromal cells of endometriotic lesion decreased (PEM p = 0.004; USL-EM p = 0.046); while the Anti-PGP 9.5 (+) endometriosis-associated nerve fiber density increased (PEM: p = 0.004; USL-EM p = 0.017) ( Fig 6 ). A and B: Patients with higher pain VAS score presents with lower Sema 3A HSCORE of stromal cells of peritoneal endometriosis (SC-PEM), but with higher PGP 9.5 (+) total nerve density. C and D: Patients with higher pain VAS score presents with lower Sema 3A HSCORE of stromal cells of deep infiltrating endometriosis of uterosacral ligament (SC-USL-EM), but with higher PGP 9.5 (+) total nerve fiber density. * P <0.05.

In conclusion, the results obtained in the present study suggest that Sema 3A and its receptors play an important potential role of regulating the aberrant sympathetic innervation of peritoneal and deep infiltrating endometriosis. The identification of sympathetic axonal repulsive effect induced by Sema 3A provides a new insight of nerve repellent factors in modulating the abnormal innervation of endometriotic lesions.

The study was conducted in the Department of Obstetrics and Gynecology of the First Affiliated Hospital of Sun Yat-sen Universigy from December 2012 to January 2014. Tissue samples consisted of 24 peritoneal endometriotic tissues and 20 deep infiltrating endometriotic tissues (all of which are from uterosacral ligaments) from 44 premenopausal women requiring surgical treatment. Endometrial tissue samples were collected from the same patient in 22 cases (women with endometriosis, secretory phase n = 8 and proliferative phase n = 14). The mean ages of patients according to the type of tissue collected were: endometriotic lesions 32 years (range 23–43 years), eutopic endometrium from women with endometriosis 34 years (range 26–40 years). And all of the patients had a regular menstrual cycle (secretory phase n = 18 and proliferative phase n = 26), without malignant or inflammatory diseases. The presence of endometrial glands and surrounding stromal cells was regarded as histological proof of endometriosis. In addition, normal endometrial samples were obtained from 26 women (mean age: 37 years; range: 33–47 years; secretory phase n = 10 and proliferative phase n = 16) undergoing hysterectomy or endometrial biopsy for benign diseases, either uterine fibroids or infertility. Healthy peritoneum (n = 13) from the lateral pelvic wall and healthy uterosacral ligament (n = 13) were obtained from patients (mean age: 39 years; range: 34–49 years) with a regular menstrual cycle (secretory phase n = 12 and proliferative phase n = 14). Patients who had no surgical and histological proof of endometriosis during hysterectomy (laparoscopy or laparotomy) for uterine fibroids were included in these groups. The lower basis of all peritoneal biopsies was the sub-peritoneal fat. The uterosacral ligaments were sampled 1 cm from the uterus. None of the patients had received hormonal treatment for at least 3 months prior to surgery. All of the patients had no malignant or inflammatory disease. All the tissue samples were obtained with written, full and informed consent from all the patients. The research protocol was approved by the Research Ethics Committee of the First Affiliated Hospital of Sun Yat-sen University. Tissues were fixed immediately in formalin (10%) and then processed as paraffin blocks. Four micrometer-thick sections of formalin-fixed tissues were deparaffinated in xylene and rehydrated through a graded series of ethanol solutions. The initial section was stained with hematoxylin-eosin for tissue diagnosis. Immunohistochemical staining was performed on paraffin sections with rabbit polyclonal antibodiy directed against Sema 3A (dilution 1:50, Sigma), rabbit polyclonal antibodiy against NRP-1 (dilution 1:20, Sigma), rabbit polyclonal antibodiy against Plexin A1 (dilution 1:50, Sigma). The antigen retrieval was performed in a pressure cooker in a sodium citrate buffer (10 mM, pH 5.5) for 2 minutes (start the time as soon as the cooker has reached full pressure) using SP Rabbit HRP kit (CWBIO Tech, Beijing, China). The sections were incubated with endogenous peroxidase for 10 minutes, washed three times with phosphate buffered saline (PBS, pH7.4) and then blocked with 5% normal goat serum for 30 minutes. After Washing with PBS, the sections were incubated with primary antibodies against Sema 3A, NRP-1 and Plexin A1 overnight at 4°C. After washing with PBS, the sections were incubated with secondary antibodies for 1 hour at room temperature and then washed with PBS three times. Detection of bound antibody was with 1- to 5-minute incubation at room temperature with 3, 3’-diaminobenzidine (DAB) substrate. Slides that were used as specificity controls underwent the same procedure without primary antibodies. These procedures resulted in negative staining. Positive controls are presented in S1 , S2 and S3 Figs. The histological slides were scanned using an Olympus BX51 microscope (Tokyo, Japan) and microscopic images containing mostly stained epithelial cells were captured and digitized using a digital camera connected to a computer. Five mostly stained microscopic fields at a magnification of ×200 were included in the measurement. For qualitative analysis, samples were considered negative when no labelled cells were observed on the tissue section and positive in all other cases. The immunohistochemical result was scored in a semiquantitative fashion incorporating both the intensity of specific staining and the percentage of stained epithelial cells. The intensity of the staining was evaluated as follows: 0 = no, 1 = weak, 2 = moderate and 3 = intense immunostaining. The percentage of stained epithelial cells was calculated by two independent experienced pathologists. Differences in opinion between the observers were resolved at a discussion microscope. For each observed slide, we used a calculated value known as HSCORE for further analysis. HSCORE = ∑ P i ( i + 1 ) , where i is the intensity of the staining varying from 0 to 3 and P i is the percentage of stained cells [ 26 ]. Each slide was examined at least twice. Immunofluorescence histochemical double staining technique was used to identify the target nerve fibers. Antibodies against protein gene product 9.5 (Anti-PGP 9.5, monoclonal mouse, dilution 1:500, Abcam, UK), tyrosine hydroxylase (Anti-TH, monoclonal mouse, dilution 1:500, Abcam, UK) were used to identify total nerve fibers and sympathetic nerve fibers, respectively. Antibody against Sema 3A (polyclonal rabbit, dilution 1:50, Abcam, UK) was used to identify the endometrial epithelial cells. The procedures before the incubation of primary antibody were the same as the immunohistochemical staining described above. Incubate the sections in the mixture of two primary antibodies (rabbit against Sema 3A and mouse against PGP 9.5 or TH) in 1% BSA in PBS in a humidified chamber overnight at 4°C. Wash the sections for three times with PBS and then incubate the sections with the mixture of two secondary antibodies which are raised in different species (DyLight 488 AffiniPure Goat Anti-Rabbit IgG and DyLight 549 AffiniPure Goat Anti- Mouse IgG, EarthOx, LLC, San Francisco, CA, USA) in 1% BSA for 1 hour at room temperature in dark. Decant the mixture of the secondary antibody solution and wash three times with PBS for 5 minutes each in dark. DAPI staining solution (Sigma) is used to identify the neuclei (incubate for 10 minutes). Rinse the slides with PBS (3×5 minutes). Drain excess buffer from the coverslip and mount with a mounting medium containing an antifade reagent (Product number: ab104136, Abcam, UK). The double immunofluorescence stained slides were scanned using an Olympus BX51 microscope (Tokyo, Japan) and microscopic images were captured and digitized using a digital camera connected to a computer. Five clearly stained microscopic fields at a magnification of ×200 were included in the measurement. Under microscope, the endometrial epithelial cells were markedly labelled by green fluorescence, while the nerve fibers were labelled by orange-yellow fluorescence. Any nerve found within 1.5 mm of an endometriotic lesion was defined as endometriosis-associated nerve (EAN) [ 27 ], while the nerve found in the area (with a normal histological tissue) at least 4-mm distance from the endometriotic lesion was defined as para-endometriotic nerve (PEN) [ 8 ]. The number of nerve fibers per square millimeter (nerve fiber density, NFD) was determined by averaging the number of counted nerve fibers, using a grid of 1 mm 2 , in 5 non-overlapping indiscriminately selected high-power fields (magnification ×200) [ 28 ]. All of the slides were examined and the NFD were calculated blindly by two independent investigators. Finally, the average NFDs of the results from two investigators were employed for further analysis. Using a standardized questionnaire with a visual analogue scale (VAS), the degree of dysmenorrhea was documented before surgery. In VAS evaluation, the degree of dysmenorrhea was quantified on a scale of 0–100 mm. “No pain” was indicated at the left side of the scale and “the maximum pain you could imagine” at the right side of the scale. The value of VAS was collected by the investigator responsible for preoperative data collection. The severity of dysmenorrhea was defined as the following: 1–3 points (mild pain), 4–6 points (moderated pain), and 7–10 points (severe pain). The Kruskal-Wallis and Mann-Whitney tests were used for comparisons of semi-quantitative immunostaining HSCORE and NFDs. Statistical significance was defined P <0.05. Statistical analysis was performed using GraphPad Prism 5 for Windows (GraphPad Software, 2003, USA). Means and standard deviations are shown.

(DOCX) Click here for additional data file. (DOCX) Click here for additional data file. EAN-PEM: endometriosis-associated nerve of peritoneal endometriosis; PEN-PEM: para-endometriotic nerve of peritoneal endometriosis; N-PC: nerve of peritoneum of control. (DOCX) Click here for additional data file. ESAN-PEM: Endometriosis-associated sympathetic nerve of peritoneal endometriosis; PESN-PEM: Para-endometriotic sympathetic nerve of peritoneal endometriosis; SN-PC: sympathetic nerve of peritoneum of control. (DOCX) Click here for additional data file. EAN-USL-EM: endometriosis-associated nerve of deep infiltrating endometriosis of uterosacral ligament; PEN-USL-EM: para-endometriotic nerve of deep infiltrating endometriosis of uterosacral ligament; N-USL-C: nerve of uterosacral ligament of control. (DOCX) Click here for additional data file. ESAN-USL-EM: endometriosis-associated sympathetic nerve of USL-EM, PESN-USL-EM: para-endometriotic sympatheticnerve of USL-EM; SN-USL-C: sympathetic NFD of uterosacral ligament of control. (DOCX) Click here for additional data file. Sema 3A was positively stained in glands of intestinal mucosa. (immunohistochemical stain, 100 × magnification) (DOCX) Click here for additional data file. Plexin A1 was positively stained in glandular cells of rectal mucosa. (immunohistochemical stain, 100 × magnification) (DOCX) Click here for additional data file. NRP-1 was positively stained in nerve fibers. (immunohistochemical stain, 100 × magnification) (DOCX) Click here for additional data file. (immunohistochemical stain, 200 × magnification) (DOCX) Click here for additional data file. (immunohistochemical stain, 200 × magnification) (DOCX) Click here for additional data file. (immunohistochemical stain, 200 × magnification) (DOCX) Click here for additional data file. (immunohistochemical stain, 200 × magnification) (DOCX) Click here for additional data file. (immunohistochemical stain, 200 × magnification) (DOCX) Click here for additional data file. (immunohistochemical stain, 200 × magnification) (DOCX) Click here for additional data file. (immunohistochemical stain, 200 × magnification) (DOCX) Click here for additional data file. (immunohistochemical stain, 200 × magnification) (DOCX) Click here for additional data file. PEM: peritoneal endometriosis; USL-EM: deep infiltrating endometriosis of uterosacral ligament. (immunohistochemical stain, 200 × magnification) (DOCX) Click here for additional data file. A: Merged image of double staining of both Sema 3A and TH (white arrow: peritoneal endometriotic glands, Sema 3A positive stained in green; white triangle: endometriosis-associated sympathetic nerve, TH positive stained in orange yellow; nuclei were stained in blue by DAPI staining). B: Merged image of double staining of both Sema 3A and PGP 9.5 (white arrow: endometriotic glands of deep infiltrating endometriosis of uterosacral ligament, Sema 3A positive stained in green; white triangle: PGP 9.5 positive stained endometriosis-associated nerve, orange yellow; nuclei were stained in blue by DAPI staining). Original magnification: 200×. TH: Tyrosine hydroxylase. (DOCX) Click here for additional data file.