Introduction

Endometriosis is an estrogen-dependent disease, characterized by the appearance of ectopic endometrium. The progress of endometriosis is mostly chronic, and inflammatory mechanisms are pivotal. The most common symptoms are chronic pelvic pain and infertility [1].

Studies in different chronic inflammatory diseases, such as rheumatoid arthritis (RA), have elucidated interesting parallels between the diseases. Changes in the sensory and sympathetic innervation seem to be common phenomena. A noradrenergic hypoinnervation, for example, is found in peritoneal endometriosis, RA, psoriasis, Crohn’s disease, and pruritus and is involved in the progress and severity of the diseases [2–4]. In contrast, the peptidergic innervation is significantly increased in peritoneal endometriosis and in RA. The imbalance in the occurrence of neurotransmitters acting respectively anti- or pro-inflammatory, probably lead to the maintenance of the chronic inflammatory milieu.

Neuroimmunomodulatory processes might be responsible for the innervatory changes in endometriosis, since significantly increased amounts of macrophages are found in peritoneal fluid and tissue of EM patients [5, 6], and it is known that innervation can be modulated by a variety of immune cells [7–10]. Furthermore, the high concentration of cytokines and prostaglandins found in peritoneal fluid of women with endometriosis, possibly affects the neuromodulation and nociceptive stimuli on the nerve fibers during endometriosis [11]. However, this possibility has not been further elucidated. A specific group of semaphorins is largely known for their function as axonal guidance cues; the expression of semaphorins at specific regions prevents axons from entering inappropriate tissues [12–14]. Recently, different studies have suggested a role of semaphorins influencing innervation during the progress of diseases such as RA and Morbus Crohn [15, 16]. During tissue injury and consequent inflammation, the macrophages and fibroblasts supposed to antagonize the inflammation, secrete semaphorins [16–18].

A potential factor involved in the degeneration of sympathetic nerve fibers is semaphorin 3F. A group of investigators analyzed the influence of semaphorin 3F and estrogen on the sympathetic innervation in the rat uterus. Estrogen upregulation led to semaphorin 3F increased expression and also to a sympathetic denervation in the rat uterus [19–21]. This insight supports the thesis of changes in the innervation during endometriosis, due to its estrogen dependence. Semaphorin 3C and 3F are upregulated in synovial tissue and fluid accompanied by a sympathetic hypoinnervation in patients with RA [22, 23]. Semaphorins are secretory proteins, which require specific receptor binding to trigger their signal cascade. The necessary receptors consist of neuropilins and plexins; depending on the constellation, a repellence of specific nerve fibers occurs [24, 25]. To induce sympathetic nerve fiber collapse, semaphorin 3C requires Nrp1 and Nrp2 [12, 13, 24, 26], semaphorin 3F has a higher affinity to Nrp2 [25] and additionally binds plexinA3 and A4 [27–29]. Particularly, the receptor Nrp2 is proposed to aggravate sympathetic nerve fiber repulsion in RA and it is known that Nrp2 can be located on the surface of sympathetic nerve fibers and is responsible for their repulsion by Sema 3C/3F [22, 25, 30].

Neuroimmunomodulatory actions can only occur if immune cells express factors capable of modulating innervation and are able to reach the site of inflammation. In this study, we investigated the noradrenergic innervation in peritoneal endometriosis. Consequently, the expression of Sema 3C/3F and their main receptors was analyzed, in order to elucidate their possible role during neuroimmunomodulation and chronic inflammation.

Materials and Methods

Patient Samples

Peritoneal Tissue Samples

Endometriotic peritoneal tissue samples of 38 patients with clinical- and histological-proven endometriosis were collected during laparoscopy. Only women in reproductive age were included in this study (range: 19–53 years, mean: 32.8). The stage of endometriosis was provided during the surgery according to the revised American Society of Reproductive Medicine (rASRM). The menstrual cycle, endometriosis stage (rASRM), and the occurrence of pelvic pain of the patient samples utilized in this study are stated in Table 1 (endometriotic peritoneum).

Table 1 Endometriotic peritoneal tissue samples from endometriosis patients
Full size table

Furthermore, we collected unaffected peritoneal tissue samples of 24 patients with endometriosis of regions with no macroscopic or histological proof of endometriosis lesions. This tissue was collected from the opposite site of the affected peritoneum. The menstrual cycle, endometriosis stage (rASRM), and the occurrence of pelvic pain of the patient samples utilized in this study are stated in Table 2 (unaffected peritoneum of women with endometriosis).

Table 2 Unaffected peritoneal tissue samples from endometriosis patients.
Full size table

As control group, 10 healthy peritonea from women with macroscopic and histological proven exclusion of endometriosis were collected during hysterectomy (uterine fibroids). The tissue probes comprised peritoneal tissue and the underlying subperitoneal fat. The patients had a regular cycle. The mean age of the women in the control group was 41.3 (range 30–50 years).

Peritoneal Fluid Samples

Peritoneal fluid of women with and without endometriosis was collected for analysis of protein expression of Sema 3C, Sema 3F, Nrp1, and Nrp2, respectively (ELISA analysis). The menstrual cycle, endometriosis stage (rASRM), and the occurrence of pelvic pain of the patients probes utilized are stated in Table 3.

Table 3 Peritoneal fluid samples from endometriosis patients and control patients without endometriosis used for ELISA analysis
Full size table

RNA in Situ Hybridization

Peritoneal biopsies were directly incubated in 4 % paraformaldehyde at 4 °C overnight and were subsequently prepared for embedding in paraffin. Subsequently, the specimens were sectioned at 7-μm thickness and the slides were treated for RNA in situ hybridization. RNA in situ hybridization was essentially performed as previously described [31]. Paraffin sections were deparaffinized, rehydrated, treated with Proteinase K (Roche, Penzberg, Germany), and then acetylated. After washing in H2O, sections were dehydrated in different concentrations of ethanol (70, 80, 95, and 100 %) and incubated in chloroform to de-fat the sections. After hybridization with the cRNA probes (approximately 1 μg of the Sema 3C, Sema 3F, Nrp1, and Nrp2) and immunodetection of digoxigenin with alkaline phosphatase-conjugated antibody (Roche, Penzberg, Germany), sections were incubated with the BM purple (Roche, Penzberg, Germany). The color reaction was stopped in Tris–EDTA buffer and appeared purple. No background staining was used, due to the similar color of BM purple (RNA positive staining) and the Hematoxylin & Eosin (normally used as background staining). Peritoneal endometriotic sections of 10 patients and 5 control peritoneal tissues of women without endometriosis were studied. As a positive control, the mouse brain was used and negative controls were achieved by using sense probes of the RNA. The plasmids were a generous gift from Professor Joelle Roche (Sema 3F) and Alain Chedotal (Sema 3C, Nrp1, and 2).

Immunohistochemical Analysis

Samples collected during laparoscopy were immediately fixed in formalin (4 %) and consequently embedded in paraffin. The paraffin-embedded tissue was sliced in 2-μm thick serial sections. Thereafter, immunohistochemical stainings were performed on these sections using following antibodies: Sema 3C (polyclonal rabbit, dilution: 1:150, Sigma-Aldrich, Germany), Sema 3F (polyclonal goat, dilution:1:300, Santa Cruz Biotechnology, USA), Nrp1 (monoclonal mouse, dilution 1:100, Santa Cruz Biotechnology, USA), Nrp2 (monoclonal mouse, dilution 1:250, Santa Cruz Biotechnology, USA), PlxnA3 (monoclonal mouse, dilution 1:200, Millipore, USA), PlxnA4 (polyclonal rabbit, dilution 1:200, Sigma-Aldrich, Germany), and tyrosine hydroxylase (TH, monoclonal mouse, dilution, 1:250, Abcam, UK). Biotinylated secondary antibodies were used for detection of the first antibodies, rabbit anti-mouse IgG (dilution 1:400, Jackson Immuno Research Laboratories, Germany), mouse anti-rabbit IgG (dilution: 1:400, Jackson Immuno Research Laboratories, Germany), rabbit anti-rat IgG (dilution 1:400, Dako, Germany), and rabbit anti-goat IgG (dilution 1:400, Dako, Germany). To detect the biotin signal of biotinylated secondary antibodies, streptavidin-alkaline phosphatase (dilution 1:1000, Roche, Germany) was added, and as chromogen FastRed (Dako, Germany) or streptavidin-horse radish peroxidase (dilution 1:400, Dako, Germany) was added and then the chromogen DAB (Invitrogen, Germany). All antibodies and streptavidin were applied and incubated for 45 min on the tissue sample. Fast Red (red staining) was incubated for 10–15 min and DAB (brown staining) for 5 min. The stainings were then analyzed close to the endometriotic lesion. The numbers of Sema-positive cells, CD68-positive macrophages, and TH-positive nerve fibers per mm2 were determined by averaging the number of stained cells or nerves in three randomly selected high-power fields of view (magnification ×200) using semi-quantitative analyses. Power fields in the tissue were always chosen in the same pattern for all samples (L-shape: top left, bottom left, and bottom right) and cells counted by one individual.

Gut samples served as positive control for TH, Sema 3C, Sema 3F, and P4HB stainings. Placenta samples served as positive control for Nrp1 and Nrp2 stainings. Tonsil samples served as positive control for CD68. For the negative control, the primary antibody was excluded during the staining.

Immunofluorescence Double Staining

For the immunofluorescence double stainings, we used the following antibodies: CD68 (monoclonal mouse, dilution 1:25,000, Dianova, Germany), Nrp1 (monoclonal mouse, dilution 1:100, Santa Cruz Biotechnology, USA), Nrp2 (monoclonal mouse, dilution 1:250, Santa Cruz Biotechnology, USA), prolyl 4-hydroxylase beta (marker for activated fibroblasts) (P4HB, polyclonal rabbit, dilution 1:1000, Sigma-Aldrich, Germany), Sema 3C (polyclonal rabbit, dilution: 1:150, Sigma-Aldrich, Germany), Sema 3F (polyclonal goat, dilution:1:300, Santa Cruz Biotechnology, USA), TH (polyclonal goat, dilution, 1:100, Abcam, UK), PlxnA3 (monoclonal mouse, dilution 1:200, Millipore, Temecula CA), and PlxnA4 (polyclonal rabbit, dilution 1:200, Sigma-Aldrich, Germany). Secondary antibodies are the following: Alexa 488 (donkey anti-mouse or donkey anti-rabbit, dilution 1:100, Abcam, UK) and Dylight 555 (donkey anti-mouse, donkey anti-goat or donkey anti-rabbit, dilution 1:100, Biomol, Germany). We analyzed the co-expression of Nrp1, Nrp2, PlxnA3, and PlxnA4 with TH-positive nerve fibers and evaluated the expression of Sema 3C/3F in P4HB-positive fibroblasts and CD68-postive macrophages. Nuclear counterstain was performed with Dapi (4′,6-diamidino-2-phenylindole). Positive and negative controls are described for IHC staining.

Enzyme Linked Immunosorbent Assay (ELISA)

ELISA was performed with peritoneal fluid of women with and without endometriosis. Peritoneal fluid was collected undiluted directly during laparoscopy and centrifuged for 5 min at 3000×g. Supernatant was collected, and pellet was discarded. ELISA procedure was as given by the company. ELISA Kits: Semaphorin 3C, Semaphorin 3F, Neuropilin 1, and Neuropilin 2 (USCN Business Co., Ltd., USA).

Statistical Analysis

Statistical analysis of the data proceeded with the program Prism 4 for Windows (GraphPad Software, 2003, San Diego, USA). Different tests were performed: non-parametric (Mann-Whitney test), two way analysis of variance, or non-parametric one-way analysis of variance (Kruskal Wallis test). Post hoc tests were Dunn’s multiple comparison and Bonferroni. Statistical significance was defined as p < 0.05.

Results

Quantification of Noradrenergic Nerve Fibers in Peritoneal Endometriosis

Noradrenergic nerve fibers (TH positive) were quantified in peritoneal endometriotic tissue and peritoneal tissue of women without endometriosis and compared. The number of noradrenergic nerve fibers in peritoneal endometriotic tissue and in unaffected peritoneal tissue of patients with endometriosis was significantly lower (mean TH-positive nerve fibers/mm2: 0.32 ± 0.09 and 1.25 ± 0.53) than in peritoneal tissue of women without endometriosis (mean TH-positive nerve fibers/mm2: 0.83 ± 0.19) (Fig. 1a–g). The noradrenergic nerve fibers found in healthy and endometriotic peritoneum were associated with vessels in the most cases. No significant differences in the amount of noradrenergic nerve fibers in the peritoneal tissue of women with and without endometriosis depending of menstrual cycle phase, endometriosis stage, or pelvic pain could be found (data not shown).

Fig. 1
figure 1

a Quantification of noradrenergic nerve fibers in peritoneum of women without endometriosis and peritoneum of women with endometriosis, lesional, and non-lesional (mean and standard deviation; Kruskal Wallis test with Dunn’s multiple comparison test) *p < 0.01. HP: healthy peritoneum of women without endometriosis, UP: unaffected peritoneum of women with endometriosis, EMP: peritoneum with endometriosis lesion. TH-positive noradrenergic nerve fibers (red) in (b, c) peritoneum of women without endometriosis and (d–g) peritoneum of women with endometriosis (d, e) lesional peritoneum and (f, g) unaffected, non-lesional peritoneum (background staining with hemalaun). Arrowhead points at nerve fiber and pointed square corresponds to picture in higher magnification. Arrow points at endometriosis lesion

Full size image

Sema 3C and Sema 3F are Expressed Surrounding Endometriosis Lesions

The in situ hybridization analysis revealed that Sema 3C and Sema 3F are expressed at the RNA level in the endometriosis affected peritoneum (Fig. 2a, b). The Sema expression in the peritoneal endometriotic tissue was confirmed by immunochemical analysis. Sema 3C- and 3F-positive cells were detected close to the endometriotic lesion (Fig. 2c, f). Due to morphologic evaluation, endothelial cells of blood vessels, stromal-like cells, and immune cells were defined as Sema 3C and 3F positive. In the healthy peritoneum of women without endometriosis and unaffected peritoneum of women with endometriosis Sema 3C and Sema 3F were only expressed in scattered cells in the connective tissue, presumably fibroblasts (Fig. 2d, e, g, h). The one-way ANOVA test revealed that the expression of Sema 3C is significantly increased in the affected peritoneum of women with endometriosis compared to the unaffected peritoneum of women with endometriosis and the control group of women without endometriosis (mean cells/mm2: 27.76 ± 2.64, 2.92 ± 1.09, and 2.33 ± 1.01, respectively) (Fig. 2i). Sema 3F expression in the affected peritoneum of women with endometriosis is significantly increased compared to the unaffected peritoneum of women with endometriosis and control group (mean cells/mm2: 23.61 ± 2.50, 4.08 ± 1.19, and 3.56 ± 2.48, respectively) (Fig. 2j).

Fig. 2
figure 2

Expressions of Sema 3C and Sema 3F in peritoneal endometriotic lesions. RNA in situ hybridization of peritoneal endometriotic tissue (positive staining appears blue, no background staining was used) a Sema 3C and b Sema 3F (magnification 200×). Sema RNA positive staining is surrounded with a blue dashed line; the endometrial gland is marked with a surrounding black dashed line. Immunohistochemical staining against Sema 3C c in peritoneal endometriotic peritoneum (twofold magnification of stained cells in marked area), d in healthy peritoneum of women without endometriosis, and in e unaffected peritoneum of women with endometriosis. Immunohistochemical staining against Sema 3F in f peritoneal endometriotic lesions (twofold magnification of stained cells in marked area), in g healthy peritoneum of women without endometriosis (magnification 200×), and in h non-lesional peritoneum of women with endometriosis. Arrows show endometriotic lesion and arrowheads show Sema-positive cells in brown (background staining with hemalaun). Quantification of the i Sema 3C and j Sema 3F expressions in healthy peritoneum of women without endometriosis (HP), unaffected peritoneum of women with endometriosis (UP), and affected peritoneum of women with endometriosis (EMP) (mean and standard deviation; Kruskal Wallis test with Dunn’s multiple comparison test) ***p < 0.001

Full size image

No significant difference of Sema 3C or Sema 3F expression in different EM stages could be shown (data not shown). Furthermore, the Sema 3C and 3F expressions were independent from the menstrual cycle phase. When comparing the expression of Sema 3C/3F in women with endometriosis, experiencing pelvic pain and without painful symptoms, also no significant differences could be found (Sema 3C pain vs. no pain: mean cells/mm2: 26.03 ± 2.99 and 32.57 ± 5.74; Sema 3F pain vs. no pain: mean cells/mm2: 23.10 ± 2.88 and 24.17 ± 6.75).

Endometriosis-Associated Macrophages are Sema 3C and Sema 3F Positive

Peritoneal macrophages are significantly increased in endometriotic peritoneum when compared to the unaffected peritoneum of women with endometriosis and the healthy peritoneum of women without endometriosis (mean macrophages/mm2: 7.46 ± 0.90, 0.50 ± 0.13, and 0.22 ± 0.17, respectively; Kruskal Wallis test with Dunn’s comparison: p < 0.001) (Fig. 3a–d, i). Macrophage conglomeration seems not to be influenced by the endometriosis stage (data not shown). Peritoneal CD68-positive macrophages expressed Sema 3C in 86.62 % of the cases (double labeling) in the endometriosis group, while in the unaffected peritoneum of women with endometriosis only 50 % of the macrophages were Sema 3C positive. In the control group without endometriosis, no peritoneal macrophages were identified. Sema 3F was expressed in 96.64 % of the peritoneal CD68-positive macrophages in the endometriosis group whereas in the unaffected peritoneum of women with endometriosis only 33.33 % of the macrophages were Sema 3F positive. In two samples of the control group without endometriosis, four macrophages could be identified; three of them were Sema 3F positive.

Fig. 3
figure 3

Immunofluorescence double staining of Sema (green) and CD68 or P4HB (red) (colocalization of markers appears yellow, background staining with dapi in blue). a, b Sema 3C/CD68 staining in a peritoneal endometriotic tissue and b healthy peritoneum of women without endometriosis; c, d Sema 3F/CD68 staining in c peritoneal endometriotic tissue and d healthy peritoneum of women without endometriosis. Sema 3C/P4HB staining in e peritoneal endometriotic tissue and f healthy peritoneum of women without endometriosis; Sema 3F/P4HB staining in g peritoneal endometriotic tissue and h healthy peritoneum of women without endometriosis (Arrowheads show colocalization, and arrows point at EM lesion) (magnification 400×). i Quantification of the amount of macrophages healthy peritoneum of women without endometriosis (HP), unaffected peritoneum of women with endometriosis (UP), and affected peritoneum of women with endometriosis (EMP) (mean and standard deviation; Kruskal Wallis test with Dunn’s multiple comparison test) ***p < 0.001

Full size image

Activated Fibroblasts in Peritoneal Endometriosis are Sema 3C and Sema 3F Positive

Endometriosis-associated activated fibroblasts, characterized by the marker P4HB, could be identified in the affected peritoneum of women with endometriosis and were always Sema 3C and Sema 3F positive(Fig. 3e, g). In the unaffected peritoneum of women with endometriosis, just few activated fibroblasts could be found and these were also Sema 3C and Sema 3F positive. In healthy peritoneum of women without endometriosis, just very few activated fibroblasts could be found and these did not show Sema expression (Fig. 3f, h). Furthermore, activated fibroblasts could be found between stromal cells around an endometriosis lesion and in the fibro connective tissue, which surrounds the fat tissue of the peritoneum of women with endometriosis. In contrast, in women without endometriosis this could not be shown.

Nrp1 and Nrp2 are Expressed in the Affected Peritoneum of Women with Endometriosis

The Sema receptors Nrp1 and Nrp2 expression at RNA level was identified through RNA in situ hybridization in affected peritoneum of women with endometriosis. Mostly, the tissue surrounding the lesion and the lesion were positive for the staining (Fig. 4a, b). The expression of these receptors was further analyzed at protein level through immunohistochemical staining. The staining revealed Nrp1 and Nrp2 expressions in the affected peritoneum of women with endometriosis (Fig. 4c, d). Morphologically, we defined the expression of the receptors in stromal cells, immune cells, and endothelial cells and in blood vessels in the area close to the endometriotic lesion, but also in the surrounding fibro connective tissue even so to a lesser extent, similarly to the expression in the unaffected peritoneum of women with endometriosis (Fig. 4c, d). In healthy peritoneum, Nrp1 and Nrp2 are only expressed in blood vessels (Fig. 4g, h). Expression of these receptors in blood vessels during endometriosis could also be identified (Fig. 4e, f).

Fig. 4
figure 4

Expressions of Nrp1 and Nrp2 in peritoneal endometriotic lesions. RNA in situ hybridization of peritoneal endometriotic tissue (positive staining appears blue, no background staining was used) a Nrp1 and b Nrp2 (magnification 200×). Nrp RNA positive staining is surrounded with a blue dashed line; the endometrial gland is marked with a black dashed line. Immunohistochemical staining against Nrp1 c in peritoneal endometriotic lesions (twofold magnification of stained cells in marked area) and against Nrp2 d in peritoneal endometriotic lesions (twofold magnification of stained cells in marked area). Arrows indicate endometrial gland; arrowheads show Nrp expression in brown, background staining with hemalaun. Immunofluorescence staining of Nrp1 or Nrp2. e, g Nrp1 expression in blood vessels in e peritoneal endometriotic tissue and g healthy peritoneum of women without endometriosis; f, h Nrp2 expression in blood vessels in f peritoneal endometriotic tissue and h healthy peritoneum of women without endometriosis. Intramural green staining in Fig. e, f are erythrocytes, which are very often positive because of the high autofluorescence of hemoglobins. Arrowheads show Nrp1 or Nrp2 expression in green and background staining with dapi in blue (magnification 600×)

Full size image

Nrp1 and Nrp2 must be expressed on the membrane of nerve fibers to be able to respond to a Sema signaling. Therefore, characterization of the precise localization of the Nrp’s expression was necessary.

Sema Receptors are Expressed on the Membrane of Peritoneal Noradrenergic Nerve Fibers in Endometriotic but not in Healthy Peritoneum

Nrp1 and Nrp2 expressions on the membrane of nerves in the peritoneal cavity of women with endometriosis could be identified as noradrenergic nerve fibers by the use of Immunofluorescence double staining. The Nrp1- and Nrp2-positive nerves were detectable in endometriotic lesions as well as in unaffected peritoneum of women with endometriosis (Fig. 5 a, d, c, f). The noradrenergic nerve fibers in the peritoneal cavity of women without endometriosis did neither express Nrp1 nor Nrp2 (Fig. 5b, e). Interestingly, the tyrosine hydroxylase expression was stronger in the nerves of the healthy patients than of the patients with peritoneal endometriotic lesions (Fig. 5a, d, c, f).

Fig. 5
figure 5

Immunofluorescence double staining of Nrp and TH or plexin and TH. Nrp1/TH staining in a peritoneal endometriotic tissue, b healthy peritoneum of women without endometriosis, and c non-lesional peritoneum of women with endometriosis. d–f Nrp2/TH staining in d peritoneal endometriotic tissue, e healthy peritoneum of women without endometriosis, and f unaffected peritoneum of women with endometriosis. g–i Plexin A3/TH staining in g peritoneal endometriotic tissue, h healthy peritoneum of women without endometriosis, and i non-lesional peritoneum of women with endometriosis. d–f Plexin A4/TH staining in d peritoneal endometriotic tissue, e healthy peritoneum of women without endometriosis, and f unaffected peritoneum of women with endometriosis. Arrowheads show Nrp1, Nrp2, Plexin A3, and plexin A4 expressions in green, TH-positive nerve fibers appear red; Plexin and Nrp stainings appear green, background staining with dapi in blue (magnification 600×)

Full size image

PlxnA3 and PlxnA4 were expressed on the membrane of noradrenergic nerve fibers in the peritoneal cavity of women with peritoneal endometriosis, lesional and non-lesional, but not on noradrenergic nerve fibers in the peritoneal cavity of women without endometriosis (Fig. 5g–l).

Sema 3C Concentration is Increased in Peritoneal Fluid of Women with Endometriosis

The concentration of Sema 3C in peritoneal fluid of endometriosis patients was significantly higher than in women without endometriosis (mean Sema3C ng/ml: 23.56 ± 3.22 and 17.73 ± 4.02, respectively, (Mann-Whitney test, p = 0.0432) (Fig. 6a). Nevertheless, Sema 3F expression remained unmodified with or without endometriosis (mean Sema3F ng/ml: 137.7 ± 26.02 and 201.7 ± 24.82 in EM and healthy PF, respectively. Mann-Whitney test. p = 0.09) (Fig. 6b). Soluble forms of Sema receptors are known, and these can act to promote signal transduction. Therefore, Nrp1 and Nrp2 expressions in peritoneal fluid were also analyzed.

Fig. 6
figure 6

ELISA analysis of a Sema 3C, b 3F, c Nrp1, and d Nrp2 expressions in peritoneal fluid of patients with endometriosis (EM) and patients without endometriosis (control)

Full size image

Peritoneal Fluid of Endometriosis Patients has Elevated Nrp2 Levels but not Nrp1

Nrp1 concentration resulted to be extremely low in peritoneal fluid, independent of endometriosis occurrence. Therefore, no significant differences could be identified. ±0.23 and 0.67 ± 0.16 in endometriosis and healthy probes, respectively. Mann-Whitney test. p = 0.84) (Fig. 6c). However, Nrp2 expression was identified in peritoneal fluid. And in this case, Nrp2 concentration was significantly higher in women with endometriosis when compared to samples from women without endometriosis (mean Nrp2 ng/ml: 206.3 ± 54.60 and 80.12 ± 25.25, respectively. Mann-Whitney test. p = 0.04) (Fig. 6d).

Discussion

Depletion of sympathetic nerve fibers seems to be a common phenomenon in various chronic inflammatory diseases. The mentioned reduction of noradrenergic nerve fibers has also been shown in intestinal endometriosis [32] and could be shown in peritoneal endometriosis in this study. Further, we attempted to find out if the reduction of noradrenergic nerve fibers occurs locally in dependence of the endometriosis lesion (in direct vicinity of the lesion) or if the entire peritoneum is affected. Therefore, we analyzed non-lesional peritoneal tissue of women with endometriosis. Here, too, we found significantly reduced amounts of noradrenergic nerve fibers, suggesting an overall affected peritoneum. An overall affected peritoneum could contribute to a reduced anti-inflammatory response, thereby maintaining and promoting the chronic inflammatory status in peritoneal endometriosis. Until now, the factors and mechanisms, leading to the reduction of the noradrenergic nerve fibers in endometriosis or other chronic inflammatory diseases, remain unclear.

Different studies have proposed specific semaphorins as possible elicitors of the nerve fiber depletion [15, 16], since Sema 3C and Sema 3F act as repellents for sympathetic nerve fibers [13, 33, 34]. Semaphorins class 3 are highly conserved secretory proteins, which are long renowned as axonal guidance cues during neural development [12, 35, 36]. Recently, semaphorins are receiving more attention for their role during inflammation [37].

In the affected peritoneal endometriosis tissue, we could show significantly increased levels of Sema 3C and Sema 3F when compared to the peritoneal tissue of women without endometriosis. The levels of Sema 3C and 3F were not elevated in the unaffected peritoneum of women with endometriosis as we expected, since the noradrenergic nerve fibers are reduced in the unaffected peritoneum. Possibly, Sema 3C and Sema 3F expressions are not increased in the unaffected peritoneum due to the lack of activated fibroblasts and macrophages in the unaffected tissue. However, the noradrenergic nerve fibers in the unaffected peritoneum of women with endometriosis are Nrp1 and Nrp2 positive and, therefore, capable of being regulated by semaphorins. Possibly, the repellence of noradrenergic nerve fibers in the unaffected peritoneum is promoted by Semas expressed in the peritoneal fluid, which is constantly in contact with the whole peritoneum, since the peritoneal fluid is in constant movement. In this study, we were able to demonstrate a significant increase of Sema 3C and soluble neuropilin 2 in peritoneal fluid of women with endometriosis. Therefore, suggesting these factors might regulate sympathetic innervation in the peritoneal cavity, in endometriosis-affected regions but also in unaffected regions. Studies in rheumatoid arthritis demonstrated that expression of soluble neuropilin 2 in synovial fluid aggravates the repellence of noradrenergic nerve fibers in synovial tissue and thereby aggravates the course of the disease [22]. Furthermore, other factors known as key factors in endometriosis, such as estrogen, might influence the sympathetic innervation in the peritoneum. Richeri et al. analyzed the sympathetic innervation in the peritoneum of the rat after treatment with estrogen. Their studies revealed a loss of sympathetic nerve fibers and interestingly also a significant increase in the expression level of Sema 3F [20]. However, it remains unclear which cells are responsible for the increased expression of these Semas. According to different studies, Sema 3C and 3F are supposed to be secreted by macrophages and activated fibroblasts in inflamed tissue [16, 38]. In our study, we could identify the expression of Sema 3C and 3F in endometriosis-associated macrophages and activated fibroblasts. Thereby, the number of macrophages and activated fibroblasts was significantly higher in peritoneal endometriotic tissue, pointing at the inflammatory state in endometriosis. The elevation of macrophages in peritoneal endometriosis was already discovered in 2009 by Tran [5]. Since no activated fibroblasts could be identified in healthy peritoneum of women without endometriosis, factors only present during endometriosis must be responsible for the activation of the fibroblasts. In general, macrophages and activated fibroblasts seem to be highly dependent of the endometriosis lesion, since the number of macrophages and activated fibroblasts in unaffected peritoneum of women with endometriosis was similar to the healthy peritoneum of women without endometriosis. These results suggest that the Sema expression in endometriosis is partially regulated by macrophages and activated fibroblasts and might also explain the low levels of Semas in the unaffected peritoneum of women with endometriosis. During tissue injury or inflammation, fibroblasts are activated and modulate pro-inflammatory responses, promote angiogenesis, and attract leukocytes. Normally, the activated fibroblasts would go through this procedure and then undergo apoptosis. In the case of chronic inflammation, the fibroblasts remain active and additionally promote the activation of further fibroblasts, thereby aggravating the inflammation [39]. By now, several studies revealed that fibroblasts are able to synthesize and secrete Sema 3C and 3F [16, 40, 41], which we now demonstrated in peritoneal endometriosis, suggesting that fibroblasts might be involved in the depletion of noradrenergic nerve fibers and maintenance of the inflammation in endometriosis. However, the exact role or mechanisms by which macrophages and fibroblasts interact and affect the inflammatory status or innervation in endometriosis remains widely opaque and has to be further analyzed.

Nevertheless, the mechanisms by which Semas are able to repel noradrenergic nerve fibers are well understood. Nrp 1 and Nrp2, which form specific complexes with PlxnA3 and A4, are specific receptors necessary for specific Semas to repel nerve fibers [26, 28, 29, 42–44]. For this purpose, the receptors of the Semas have to be expressed on the neurons, which are supposed to be repelled [25, 43]. Interestingly, Nrp1, Nrp2, PlxnA3, and PlxnA4 expressions could be identified on the membrane of noradrenergic nerve fibers in peritoneum of women with endometriosis but not in the peritoneum of women without endometriosis. Expression of neuropilins on sympathetic nerve fibers is usually limited to the time during neural development; therefore, the expression of neuropilins after neural development events that has finished can only be based on tissue injury, neural regeneration, or other pathological events [41]. Therefore, expression of neuropilins on the surface of TH-positive nerve fibers during endometriosis reveals a clear pathological mechanism by which nerve fibers are highly affected. Our studies reveal that Sema 3C and 3F would be able to lead to the repellence of noradrenergic nerve fibers through the Nrp-Plx receptor complex in peritoneal endometriosis, in affected and unaffected tissues.

Our results show a significant reduction of noradrenergic nerve fibers in affected and unaffected peritoneum of women with endometriosis. Importantly, increased expressions of Sema 3C and 3F could be found in endometriotic peritoneum. Thereby, the numerous macrophages and activated fibroblasts were Sema 3C and 3F positive and noradrenergic nerve fibers were positive for the specific Sema receptors.

Interestingly, although unaffected peritoneum shows a reduced amount of Sema receptors positive sympathetic nerve fibers, Semas themselves are not increased in the unaffected peritoneum, probably due to the poor number of macrophages and activated fibroblasts found in this tissue. The reduction of noradrenergic nerve fibers in unaffected peritoneum might be regulated by semaphorins or other factors present in the peritoneal fluid of the endometriosis patients. The condition found in the unaffected peritoneum suggests pathological changes, which could be an early phase of the inflammatory alterations typical for endometriosis, consequently leading to the activation of fibroblasts and macrophages and prompting them to secrete Semas and further impair the local microenvironment.

These leads to the assumption that the immune reaction, responsible for the macrophages release and fibroblasts activation during the chronic inflammation in endometriosis, might partially modulate the noradrenergic innervation in peritoneal endometriosis.