Macrophage-derived Netrin-1 participates in endometriosis-associated pain

Background: Endometriosis is a common disease in reproductive-age women and usually causes pelvic pain. Endometriosis pain is considered as a kind of neuropathic pain and inltrating nerve ber in endometriotic lesions may play an important role. Netrin-1 is widely reported as an axon guidance cue that regulates axonal attraction or rejection in neural injury and regeneration. In this study, we aim to determine the role of Netrin-1 in endometriosis-related pain. Methods: Peripheral blood, peritoneal uid, and endometrial tissues were sampled from women with (n=37) and without (n=23) endometriosis. Serum Netrin-1 concentrations, endometrial expression levels of Netrin-1 and its receptors including DCC, A2BAR, UNC5B, UNC5C and DSCAM were assessed. The polarization phenotypes of the peritoneal macrophages were identied by detecting the marker expression of M1 (CD86+) /M2 (CD163+) macrophages via ow cytometry. Lipopolysaccharide (LPS) and interferon gamma (IFN-γ) stimulated human monocytic cell line (THP-1) and rat alveolar macrophage-derived cell line (NR8383) cells to induce M1 phenotype macrophages. The expression levels of M1 markers and Netrin-1 in THP-1/NR8383 cells were determined. Results: The expression levels of Netrin-1 in serum and endometriotic lesions were signicantly higher in women with endometriosis when compared with those in women without endometriosis (P<.05), and both were correlated with pain symptoms (P<.05). Netrin-1 was co-expressed with CD 68 (a macrophage marker) in endometriotic lesions, and was synthesized and secreted by THP-1 and NR8383 cells in process of M1 polarization. In women with endometriosis, peritoneal macrophages were polarized towards M1 phenotype. In addition, increased expression of DCC and A2BAR, and decreased expression of UNC5B, and DSCAM in found. Conclusions: These


Background
Endometriosis is a common gynecological disease among women of childbearing age, characterized by pain and infertility [1]. Pain symptoms in patients with endometriosis include dysmenorrhea, dyspareunia, dysuria, defecation pain and chronic pelvic pain, which have a signi cant impact on women's quality of life [2]. However, the exact mechanisms of endometriosis-associated pain remain unclear up to date despite extensive research efforts [3,4].
In 2000, Anaf et al. rstly reported S100-labeled nerves in ltrating in endometriotic lesions, and the percentages of nerves were signi cantly higher in the lesions of women suffered from severe endometriosis pain [5]. In subsequent studies, elevated speci c makers for sensory, sympathetic, and parasympathetic nerves [6][7][8] and growth factors such as nerve growth factor (NGF), brain derived neurotrophic factor (BDNF) and neurite growth factor 2 (NEGF2) [9][10][11] were identi ed in different types of endometriotic lesions, which were correlated with endometriosis pain. Obviously, endometriosis pain may be considered as a kind of neuropathic pain. Moreover, in a rat model of endometriosis, autotransplanted endometriotic lesions developed autonomic and sensory innervation [12]. Growthassociated protein 43 (GAP-43), a marker for neurite outgrowth and regeneration, was expressed in nerve bers in ltrating in the endometriotic lesions of peritoneal and deep in ltrating endometriosis [13,14].
However, it is still unclear how the nerve bers in endometriotic lesions sprout abnormally.
Netrins is a member of the laminin superfamily and has a similar amino acid terminal sequence [15]. The name Netrin is derived from the Sanskrit Netr, meaning 'guide' as its role in axon guidance [16]. Until now, three secreted Netrins (Netrins 1, 3 and 4), and two glycosylphosphatidylinositol-anchored membrane proteins, Netrins G1 and G2, have been identi ed in mammals. Netrin Gs regulate axon guidance by forming synaptic interactions between neurons with the transmembrane Netrin G ligands NGL1 and 2.
It has been demonstrated that Netrin-1, the most studied guidance cue, triggers attraction effect through DCC and neogenin, or repulsion effect via UNC5 A-D [15]. In the central nerve system, Netrin 1 is secreted by ventricular zone neural progenitors and oor-plate cells in the ventral embryonic spinal cord [23,24]. In the process of peripheral neural injury and regeneration, increased levels of Netrin-1 are mainly produced by Schwann cells [25]. Netrin-1 expression was increased in hypoxia conditions [22], in ammation and various diseases such as obesity [26], type 2 diabetes [27], acute lung injury [28], atherosclerosis [29] and abdominal aortic aneurysm (AAA) [30]. However, it is not clear whether Netrin-1 is involved in the pathogenesis of endometriosis.
Co-localization of Netrin-1 and CD68, a marker of macrophage, is identi ed in atherosclerotic plaques [29], adipose tissues [26] and in amed aortic vessel wall [30], suggesting that macrophages may participate in in ammation by secreting Netrin-1. In fact, the in ammatory microenvironment of endometriosis can promote macrophage in ltration [31,32], which play a crucial role in the etiology and pathogenesis of this disease including in ammatory response [33], angiogenesis [32], proliferation [34] and neurogenesis [35]. The interaction between macrophages and nerve bers contributes to neuroin ammation and pain generation in endometriosis [35]. A large number of up-regulated molecules released by nerve bers in endometriotic lesions such as monocyte chemotactic protein-1 (MCP-1), colony-stimulating factor 1 (CSF-1) and leukemia inhibitory factor (LIF), are responsible for the recruitment of macrophages from vessels within the lesions [36,37]. On the other hand, in ltrated macrophages in the lesions in turn mediate neurogenesis by secreting neurotrophins, semaphorins, and vascular epithelial growth factor (VEGF) [35,37,38]. Although macrophage activation is closely related to endometriosis, yet, macrophages differentiate into classically activated macrophages (pro-in ammatory, M1) or alternatively activated macrophages (anti-in ammatory, M2) [39], and the polarization of M1/M2 macrophages in endometriosis remains highly debated [31,32,40,41], Therefore, it is necessary to explore the role of macrophage-mediated Netrin-1 in the pathogenesis of endometriosis.
In the present study, we aimed to investigate the role of Netrin-1 mediated by macrophages in endometriosis-associated pain. To do so, we rstly detected expressions of Netrin-1 and its receptors in endometriotic lesions. Secondly, we localized Netrin-1 and macrophages in endometriotic lesions. Thirdly, we determined polarization phenotypes of peritoneal macrophages. Finally, we observed the expression and secretion of Netrin-1 after macrophage M1 polarization was induced in vitro.

Patients
A total of 60 women with (case group, n=37) and without endometriosis (other benign gynecologic diseases, control group, n=23) undergoing laparoscopic surgery at the Women's Hospital between January 2018 and June 2018 were recruited in this study. In endometriosis group, 17 women were at stage I-II while 20 of them were at stage III-IV, according to the Revised American Fertility Society Scoring system. Pain symptoms were observed in 48.6% (n = 18) and 0% (n = 0) of the endometriosis and nonendometriosis group, while infertility was 24.3% (n = 9) and 17.4% (n = 4), respectively. None of participants have received hormone therapy 6 month before surgery.

Samples Collection
Peripheral blood samples were obtained 1 day before the surgery while endometriotic lesions and endometrial tissues were collected during the surgical procedure. The bloods or cell supernatants were centrifuged at 1000 g for 10 minutes, and the supernatants was transferred into 1.5 mL tubes and stored at -80°C until processing. Half of the endometrial tissue was frozen in -80°C for mRNA detection, and the other half was immersed in formalin (Solarbio) for further immunohistochemical and immuno uorescence staining.

ELISA
The concentrations of Netrin-1 in serum or culture supernatants were quanti ed by ELISA kits (Cusabio) according to the manufacturer's instructions.
qRT-PCR and Western blot analyses Total RNA was extracted from endometrial samples or cells with TRIzol reagent (Takara) and reversed by PrimeScript Reverse Transcription (RT) reagent kit (Takara) according to the manufacturer's recommendations. SYBR Premix Ex Taq kit (Takara) was used to quantitative polymerase chain reaction (PCR) and the fold change was determined through 2 -△△ Ct method. The primers for Netrin-1, receptors and in ammatory factors were synthesized from Generay and the sequences were listed in Table 1 and  Table 2. Western blot analysis was used to test Netrin-1 protein expression levels with Netrin-1 antibody (Abcam) in THP-1 and NR8383 cells in M1 polarization process.

Immunohistochemical Staining
Speci c antibody of Netrin-1 (Abcam), DCC (Bioss), UNC5B (Bioss) and A2BAR (Invitrogen) were used to access the expression levels and localization of Netrin-1 and its receptors in endometrial tissues by immunohistochemistry (IHC). The sum of the percentage (0-3) and intensity scores (0-3) were represented as IHC scores to show the expression levels of molecule. Detailed immunohistochemical process and scoring were described before [10].

Double Immuno uorescence Staining
Slides of endometriotic lesions were incubated with primary antibody of Netrin-1 and CD68 (Abcam), and then were rinsed in PBS before mounted with corresponding uorescent secondary antibody (Abcam).

Statistics
Data were analyzed using SPSS Version 24.0 (IBM). The continuous data variables were quanti ed as mean ± SEM. Differences in variables between two groups and multiple groups were analyzed using unpaired Student t test and one-way ANOVA, respectively. Nonparametric testing was used where sample sizes were insu cient to con rm normality of data distribution. Statistical tests (χ2 and Mann-Whitney U test) were performed to compare the frequency and median among groups. P values of <.05 were considered statistically signi cant.

Characteristics of the Patient
There were no signi cant differences between endometriosis group and non-endometriosis group with respect to their age, gravidity, parity, abortion, infertility, or cycle stage, except for pain symptoms (Table  3).

Netrin-1 concentrations in serum is associated with endometriosis pain in the patients
The serum concentrations of Netrin-1 in endometriosis women were signi cantly higher than those from women without endometriosis (P<.05, Fig. 1A), and were correlated with endometriosis pain (P<.05, Fig.  1B).
Endometriosis polarizes peritoneal macrophages towards the M1 phenotype CD68+ cells were recognized as peritoneal macrophage in women with endometriosis ( Fig. 2 A, C) or without endometriosis (Fig. 2 B, D). Flow cytometry plots revealed a signi cantly higher proportion of M2 macrophages (CD86-CD163+) in non-endometriosis women compared to endometriosis women (Fig. 3C). The proportion of M1 macrophages (CD68+CD163-) was high in endometriosis women, but it did not reach statistical signi cance (Fig. 3A). In short, the proportion of CD86+ macrophages signi cantly increased in endometriosis patients compared with those without endometriosis, while the proportion of CD163+ macrophages did not differ (Fig. 3E, F). Netrin-1 receptors was more expressed in endometriotic tissues
Immunohistochemistry results showed that DCC was mainly expressed in epithelial and stromal cells and interstitium in endometriotic lesions (Fig. 6A-C), and showed a higher IHC score in endometriotic lesions than those in eutopic endometrium form women with or without endometriosis (P<.0001). UNC5B was mainly expressed in glandular epithelial cells (Fig. 6D-F) and was less expressed in endometriotic lesions than those in eutopic endometrium from women with (P<.01) or without (P<.05) endometriosis. A2BAR was widely expressed in endometriotic lesions (Fig. 6G-I), and the IHC scores were higher than those in eutopic endometrium from endometriosis (P<.05) and non-endometriosis (P<.05) women.

Discussion
This study demonstrated that Netrin-1 increased in serum and endometriotic lesions in endometriosis women. Since Netrin-1 is an axon guide molecule, we speculate that Netrin-1 mediates endometriosis pain by promoting nerve ber in ltration in endometriotic lesions. Actually, Netrin-1 has bi-functionality on axonal guidance. It has been reported that Netrin-1 leads to axon attraction binding to DCC receptor or repulsion binding to UNC5A-D receptors in the same cells [15]. Neogenin, DSCAM and CD146 also show promoting effect on axon extension [18]. The bi-functionality on axon guidance is based on the crystal structure of Netrin-1 and its receptors [42]. Besides, cross-link with different receptor types and the charge changes on Netrin-1 and receptors also determine promotion or inhibition of Netrin-1 in axon guidance [18]. Our study demonstrated that endometriotic lesions showed a signi cantly higher expressions levels of DCC, neogenin and lower expression levels of UNC5B and UNC5C compared with eutopic endometrium, which further supported that Netrin-1 is responsible for endometriosis pain by promoting nerve in ltration in endometriotic lesions. Dorsal root ganglion (DRG) neurons express DCC, neogenin and UNC5A-D receptors [43]. In a rat model of sciatic nerve transection, the expression of DCC receptor is up-regulated while the expression of UNC5B and UNC5C is down-regulated in sensory neurons [43]. Transplantation of Netrin-1 overexpression bone marrow mesenchymal stem cells promotes axon regeneration and functional recovery of the sciatic nerve after crush injury [44]. However, Netrin-1 treatment (500 ng/mL) inhibits neurite outgrowth of adult DRG neurons in explant and dissociated cultures, which may be mediated by UNC5A-C receptors on DRG [45]. Thus, the effect of Netrin-1 on Nerve outgrowth is complex and the mechanism of Netrin-1 involved in endometriosis pain remains to be further studied. M1 macrophages, activated by IFN-γ, LPS or TNF-α, participate in tissue injury, in ammatory and immune responses by producing pro-in ammatory cytokines and chemokines [46]. In contrast, M2 macrophages can be activated by IL-4, IL-10, IL-13, or transforming growth factor-β (TGF-β), thus participating in tissue repair tumor angiogenesis and vessel normalization [47]. In endometriosis, several studies have reported that the macrophages in women with endometriosis are predominantly of M2 phenotype (CD163+/CD206+), which play an important role in the development of endometriosis [31,41,[48][49][50]. As endometriosis is an estrogen-dependent disease, it has been reported that estrogen promotes M2 polarization through activation of the signal transducers and activators of transcription (STAT3) and P38mitogen activated protein kinases (MAPK) pathway [51,52]. However, another study shows opposed result that 17β-estradiol represses suppressor for M2 polarization via inhibiting the JAK1-STAT6 signaling pathway [53]. Takebayashi et al. have reported that macrophage population slants toward M1 in the endometrium of endometriosis patients due to signi cantly lower ratio of the number of CD163+ or CD206+ macrophages to CD68+ macrophages [40]. In this study, we found that the peritoneal macrophages of endometriosis were mainly of CD86+CD163+ type, while those of non-endometriosis women were mainly of CD86-CD163+ type (M2). The percentage of CD86+ macrophages in endometriosis women was signi cantly higher than that in control group, which displayed a unique M1/M2 polarization signature that was skewed towards the classical M1 activation phenotype. This was consistent with the subsequent results of experiments in vitro that M1 polarization induced up-regulation of Netrin-1 synthesis and secretion in human and rat cell line. A recent study also has reported that Netrn-1 enriched macrophages are those that highly expressed pro-in ammatory markers as Netrin-1 mRNA expression levels are increased in CD68+/CD206− pro-in ammatory phenotypes rather than CD68+/CD206+ samples [30]. Actually, polarization is a dynamic process as the signals are temporally and dynamic, and the use of terms M1 and M2 is confusing due to the lack of speci c phenotypic scoring criteria [54,55]. Many physiological or pathological macrophages did not show a clear M1 or M2 phenotype [56], or macrophages with combinations of M1 and M2 markers can be found during M1/M2 polarization [57,58]. New methods and technical advances are needed to reassess activation and classi cation of macrophage.
Netrin-1 gene is a direct transcriptional target of nuclear factor (NF)-κB, which up-regulates Netrin-1 in colorectal carcinoma and mammary epithelial cells in response to in ammation [59]. However, another study has reported that activation of NF-κB represses Netrin-1 expression levels in adenocarcinomic alveolar epithelial cells and dermal microvessel endothelial cells [28]. Actually, macrophages from endometriosis patients show a statistically signi cant higher proportion of NF-κB nuclear translocation, and release various cytokines, growth factors and angiogenic factors to participate in endometrial fragment adhesion, proliferation and neovascularization [33]. Since we only tested cell lines in vitro, primary macrophages from endometriosis women are more valuable depending on the amount of cytokine, time of exposure, and the competition for cytokine [54]. Thus, the regulatory mechanism of Netrin-1 in endometriosis macrophages needs to be further studied.
Netrin-1 also plays a role in angiogenesis, cell migration, cell proliferation and cell survival. The present study con rmed the higher expression of Netrin-1, constant expression of CD146 and lower expression of UNC5B in endometriotic lesions. Netrin-1 enriched macrophages also highly express pro-angiogenic markers [30] and treatment of Netrin-1 with low doses (50-200 ng/mL) on endothelial cells promotes proliferation, migration and tube formation via binds to high a nity receptor CD146 [19]. However, high concentrations of Netrin-1 (1000-2000 ng/mL) inhibit the above effects, possibly via UNC5B signaling pathway [19]. In endometriosis, nerve bers are accompanied by immature blood vessels within endometriotic lesions [13]. Netrin-1 also dose-dependently regulates cell migration of Schwan cell, endothelial cell via activating or inhibiting MAPK pathway by CD146 or UNC5B receptor [19,60,61]. In in ammatory, it has been reported that Netrin-1 in endothelial cells inhibits in ammatory cell migration, such as leukocyte and macrophage [30,62,63]. This may be an anti-in ammatory response in the body, but it can lead to the accumulation of in ammatory cells in the lesions, resulting in AAA or atherosclerosis [29,30]. In endometriosis, macrophage retention in endometriotic lesions increases the local concentration of Netrin-1, which may play a role in the in ltration of nerve bers. In tumorigenesis, Netrin-1 promotes cell survival, proliferation, invasion and migration in different types of cancer, such as prostate carcinoma, hepatocellular carcinoma, gastric cancer and breast cancer [64][65][66][67]. Although endometriosis is a benign gynecological disease, it has malignant behaviors in adhesion, proliferation, invasion, metastasis and recurrence [68][69][70][71]. Thus, elevated Netrin-1 may be also involved in the growth of endometriotic lesions.

Conclusions
In summary, the present study indicated that increased Netrin-1 in women with endometriosis may participate in endometriosis pain. Target therapy toward Netrin-1 and macrophages may not only interfere with the process of endometriosis pain, but also inhibit the progression of endometriosis.    Polarized phenotypes of peritoneal macrophages in women with or without endometriosis.
Representative dot plot and statistical chart of CD68+ cells in peritoneal uids from endometriosis (A, C, n=8) and non-endometriosis (B, D, n=6) women by ow cytometric analysis using CD86 and CD163 markers.