{"paper_id":"2ea7830a-3232-4e2f-ab25-fdc2254ce6d5","body_text":"R E S E A R C H Open Access\nLocalization of TrkB and p75 receptors in\nperitoneal and deep infiltrating\nendometriosis: an immunohistochemical\nstudy\nAgung Dewanto 1,2, Jozsef Dudas 3, Rudolf Glueckert 3, Sylvia Mechsner 4, Anneliese Schrott-Fischer 3,\nLudwig Wildt 1 and Beata Seeber 1*\nAbstract\nBackground: The roles of the neurotrophins NGF (Neurotrophic growth factor) and BDNF (brain-derived\nneurotrophic factor) in neuronal growth and development are already known. Meanwhile, the neurotrophin\nreceptors TrkA (tropomyosin related kinase A), TrkB, and p75 are important for determining the fate of cells. In\nendometriosis, this complex system has not been fully elucidated yet. The aim of this study was to evaluate the\nexpression and location of these neurotrophins and their receptors in peritoneal (PE) and deep infiltrating\nendometriotic (DIE) tissues and to measure and compare the density of nerve fibers in the disease subtypes.\nMethods: PE lesions ( n = 20) and DIE lesions ( n = 22) were immunostained and analyzed on serial slides with\nanti-BDNF, −NGF, −TrkA, −TrkB, −p75,-protein gene product 9.5 (PGP9.5, intact nerve fibers) and -tyrosine\nhydroxylase (TH, sympathetic nerve fibers) antibodies.\nResult: There was an equally high percentage (greater than 75 %) of BDNF-positive immunostaining cells in both\nPE and DIE. TrkB (major BDNF receptor) and p75 showed a higher percentage of immunostaining cells in DIE\ncompared to in PE in stroma only ( p < 0.014, p < 0.027, respectively). Both gland and stroma of DIE lesions had a\nlower percentage of NGF-positive immunostaining cells compared to those in PE lesions ( p < 0.01 and p < 0.01,\nrespectively), but there was no significant reduction in immunostaining of TrkA in DIE lesions. There was no\ndifference in the mean density of nerve fibers stained with PGP9.5 between PE (26.27 ± 17.32) and DIE (28.19 ± 33.\n15, p = 0.8). When we performed sub-group analysis, the density of nerves was significantly higher in the bowel DIE\n(mean 57.33 ± 43.9) than in PE (mean 26.27 ± 17.32, p < 0.01) and non-bowel DIE (mean 14.6. ± 8.6 p < 0.002).\nConclusions: While the neurotrophin BDNF is equally present in PE and DIE, its receptors TrkB and p75 are more\nhighly expressed in DIE and may have a potential role in the pathophysiology of DIE, especially in promotion of cell\ngrowth. BDNF has a stronger binding affinity than NGF to the p75 receptor, likely inducing sympathetic nerve\naxonal pruning in DIE, resulting in the lower nerve fiber density seen.\nKeywords: Neurotrophins, BDNF, Endometriosis, Immunohistochemistry\n* Correspondence: beata.seeber@i-med.ac.at\n1Department of Gynecological Endocrinology and Reproductive Medicine,\nMedical University of Innsbruck, Anichstrasse 35, Innsbruck 6020, Austria\nFull list of author information is available at the end of the article\n© 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0\nInternational License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and\nreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to\nthe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver\n(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.\nDewanto et al. Reproductive Biology and Endocrinology  (2016) 14:43 \nDOI 10.1186/s12958-016-0178-5\n\nBackground\nEndometriosis is diagnosed by the presence of endomet-\nrial glands and stroma outside of the uterine cavity.\nThese endometriotic lesions are predominantly found in\nthe pelvis, namely on the peritoneum, ovaries, myome-\ntrium of the uterus (adenomyosis) and the bowel. The\ngold standard for the diagnosis of endometriosis is surgi-\ncal visualization, which is commonly accomplished with\nlaparoscopy, followed by histological confirmation of\noperatively excised lesions [1]. Endometriosis can be\ndivided into two subtypes: superficial peritoneal endo-\nmetriosis (PE) and deep infiltrating (DIE) disease, the\nlatter defined by lesions located at least 5 mm be-\nneath the peritoneal surface. The pathophysiology of\nDIE, including its development, innervation, and asso-\nciation with inflammation and vascularization, has\nbeen the subject of numerous studies. Most, but not\nall, conclude that these aspects differ between DIE\nand PE endometriosis [2 –4].\nIn addition, multiple investigators have aimed at\nelucidating the mechanisms of pain generation in\nendometriosis to improve the understanding of endo-\nmetriosis-associated pelvic pain [5, 6]. Previous studies\nhave evaluated for the presence of nerve fibers in endome-\ntriotic lesions, comparing the innervation density between\nPE, DIE and normal peritoneum using specific stains for\nall intact-nerve fibers (Protein Gene Product 9.5,PGP9.5),\nsensory nerve fibers-(Substance-P), and sympathetic-\nnerve fibers (tyrosine hydroxylase) [7, 8]. The studies have\nfound that: (1) there is little difference in overall nerve\nfiber density between PE and healthy peritoneum and (2)\nPE lesions have a higher density of sensory nerve fibers\nand a lower density of sympathetic nerves than healthy\nperitoneum [7]. In DIE, the density of nerve fibers stained\nwith PGP9.5 was the highest in endometrial lesions taken\nfrom rectum compared to the density seen in peritoneal\nlesions and in DIE lesions taken from other locations such\nas the uterosacral ligament or the retro-uterine cul de sac.\nSurprisingly, the nerve fiber density in non-bowel DIE\n(cul-de-sac and uterosacral ligament lesions) was not dif-\nferent than that observed in peritoneal endometriosis [8].\nThe presence of neurotrophins, a family of proteins\ncritical to supporting the growth and differentiation of\ndeveloping neurons and to maintaining neuronal sur-\nvival, has been studied in endometriosis [9]. The key\ncomponents of the neurotrophin system in humans are\nnerve growth factor (NGF), brain-derived neurotrophic\nfactor (BDNF), and the neurotrophins-3 (NT-3), −4/5\n(NT-4/5) [10] as well as their respective receptors, the\ntropomyosin kinase receptors (TrkA, TrkB, TrkC) and\nthe neurotrophin receptor p75, a member of the tumor\nnecrosis factor (TNF) receptor superfamily. Specifically,\nthe Trk receptors consist of: TrkA, the high affinity re-\nceptor for NGF and TrkB, the main receptor of BDNF\nand NT4/5. The binding of NGF and BDNF to the P75\nreceptor induces survival of the cell while the binding of the\npremature type of neurotrophin, pro-NGF and pro-BDNF,\nto the p75 receptor induces apoptosis in the cells [10, 11].\nIn endometriosis, the expression of NGF was con-\nfirmed using immunohistochemistry in the gland,\nstroma, [12 –14] and in nerve fibers of endometriosis\nlesions [4, 8]. In addition, the concentration of the\nneurotrophin NGF has been reported by several groups\nto correlate with the density of nerve fibers, both in PE\n[4, 13, 14] as well as in DIE, including bowel endometri-\nosis [8, 12, 15]. Furthermore, Anaf et al. reported that\nNGF expression was higher in the glands and stroma of\nadenomyosis lesions compared to PE [13], but interest-\ningly did not differ from endometrium of disease-free\ncontrols [16].\nThe spatial relationship between the main neurotro-\nphins (NGF and BDNF) and their receptors (TrkA,\nTrkB, p75) in histological sections from endometriosis\ntissue has not been extensively studied so far. Tarjanne\net al. recently reported that NGF and its receptor TrkA\nwere strongly expressed in rectovaginal endometriosis\nbut they did not make comparisons to PE or DIE from\nother locations [12, 14]. No studies have evaluated the\npresence and location of BDNF in endometriosis lesions,\nalthough BDNF is present in eutopic endometrium from\nwomen with endometriosis.\nThe main objective of this immunohistochemistry\nstudy was to assess for the presence and localization of\nthe neurotrophins NGF and BDNF and their receptors,\nTrkA and, TrkB and p75, in endometriosis lesions and\nto compare these between PE and DIE by using quanti-\ntative methods. Secondarily, we compared the density of\nnerve fibers in PE versus non-bowel DIE and bowel DIE.\nMethods\nSubjects\nArchived endometriotic tissue samples were collected\nfrom 44 patients who had undergone surgery for pain,\ninfertility suspected uterus malformation or other gyne-\ncologic indication. Disease was confirmed histologically\nby the presence of both endometrial glands and stroma\nin an ectopic location. We analyzed PE lesions from 20\nwomen and DIE lesions from 22 women, 7 of whom had\ndeep endometriosis lesions located on the bowel ( n = 7),\nthe remaining in non-bowel pelvic locations ( n = 15).\nEndometriosis was staged surgically according to the\nAmerican Society for Reproductive Medicine (ASRM)\nrevised guidelines. The age of patients ranged from 27 to\n45 years (mean age ± Standard Deviation (SD): 33.15 ±\n5.79 in PE group) and 24 to 49 years (mean age ± SD:\n33.54 ± 6.72 in DIE group). Medical records were\nreviewed to collect relevant clinical information. All sub-\njects were confirmed to be premenopausal with regular\nDewanto et al. Reproductive Biology and Endocrinology  (2016) 14:43 Page 2 of 13\n\nmenstrual cycles. For the women in the PE group, 8\nwere in the menstrual or proliferative phase and 9 in the\nsecretory phase of the cycle, and for 3 women this infor-\nmation was not available. For the most of the women in\nthe DIE, menstrual phase information was not available,\nwhile 5 were confirmed to be in the menstrual or prolif-\nerative phase. Table 1 summarizes the relevant clinical\ndata of subjects.\nNone of the subjects received medical therapy for\nendometriosis nor hormonal contraception for at least\n3 months prior to laparotomy or laparoscopy for the exci-\nsion of endometriosis. This study was approved by the\nEthics Committee of the Medical University of Innsbruck\n(No. UN5130, July 2nd, 2013) and all patients gave their\ninformed consent for research participation.\nHistological specimens\nAll the specimens were immediately fixed in 4 % para-\nformaldehyde for 12 h, and processed and embedded in\nparaffin according to standard protocols. Each section\nwas cut at 3 μm thickness for 15 serial sections. Serial\nsections enabled us to study corresponding spatial/ana-\ntomical sites and evaluate the location of neurotrophins\nas well as their receptors. Routine hematoxylin and eosin\n(H&E) staining was performed for tissue overviews.\nImmunohistochemistry of sections with antibodies di-\nrected against NGF, BDNF, TrkA, TrkB, p75, PGP9.5,\nand TH were performed. Immunohistochemistry was\nperformed with an automated staining system Ventana®\nRoche® Discovery that ensured precise and equal treat-\nment of each slide. This was a prerequisite for quantifi-\ncation of immunostaining intensities. Sections were\nimmunohistochemically stained with the polyclonal\nrabbit anti-NGF (dilution: 1:400, sc-548, Santa Cruz\nBiotechnology, California), polyclonal rabbit anti-BDNF\n(dilution: 1:1200, ab101752, Abcam, England), polyclonal\ngoat anti-TrkA (dilution: 1:240, sc-20,537 Santa Cruz\nbiotechnology, California) polyclonal rabbit anti-TrkB\n(dilution: 1:60, sc-8316, Santa Cruz Biotechnology,\nCalifornia), polyclonal rabbit anti p75 (dilution: 1:3200,\na gift from Prof. Reichardt, University California San\nFrancisco), polyclonal rabbit anti PGP.9.5 (dilution:\n1:2400, ab10404, Abcam, England) monoclonal mouse\nanti-tyrosine hydroxylase (dilution 1:2000, T1299,\nSigma, USA). Healthy human colon and healthy hu-\nman endometrium were used as positive controls for\nneurotrophins and receptors. Healthy human skin and\nhealthy human colon were used as positive controls for\nnerve fibers. Negative controls were treated identically\nexcept that the primary antibody was replaced with\nIgG rabbit or goat isotype for polyclonal antibodies\nand IgG1 mouse for monoclonal antibody.\nCell counting\nFor each study participant, one representative gland and\narea of stroma were identified. Once the immunohisto-\nchemistry analysis for that gland and stroma was carried\nout with a single antibody, it was then carried out on the\nsame gland and stroma in the adjacent section for an-\nother antibody. With this method, we could determine\nthe precise localization of the neurotrophins and recep-\ntors on the same gland and stroma.\nStained tissue sections were analyzed at × 20 magnifi-\ncation using a Tissuefax Plus system based on a Zeiss®\nAxioImagerZ2 Microscope (Jena, Germany). Images\nwere acquired with the TissueFaxs (Tissue-Gnostics®,\nVienna, Austria) software. The percentage of NGF-,\nBDNF-, TrkA-, TrkB-, and p75 –positive immunostain-\ning cells in each endometriosis tissue specimen was\nquantified using HistoQuest® (Tissue-Gnostics) software.\nThis software has been used by previous researchers\n[17–19] and has the advantage of being more objective\nthan the subjective assessment by an investigator.\nHistoquest® is an analytical tool used to quantify im-\nmunostaining based on single cells using the cell specific\nnucleus structure as the primary identification marker\n(hematoxylin), followed by an automatic segmentation of\nTable 1 Clinical characteristics of subjects\nPatient Information PE DIE\nn 20 22\nAge (mean ± SD) 33.15 ± 5.79 33.54 ± 6.72\nrASRM\nI7 0\nII 10 6\nIII 3 11\nIV 0 5\nMenstrual Phase\nProliferative/menstruation 8 5\nSecretory 9 –\nUnknown/missing 3 17\nPain Type/location\nLower abdominal pain 4 7\nDysmenorrhea only 8 8\nMixed Pain 1 2\nNo pain 7 5\nLocation of DIE lesion\nBowel (rectum, appendix, colon/sigmoid) – 7\nNon-bowel: – 15\nParaureteral – 2\nBladder – 2\nPararectal/rectovagina – 10\nvagina – 1\nDewanto et al. Reproductive Biology and Endocrinology  (2016) 14:43 Page 3 of 13\n\nthe immunostaining confined to the corresponding nu-\ncleus. A ring mask around this nucleus is interactively\ndefined and set as parameter for all sections stained with\na certain marker-specific channel named single reference\nshade. The brown staining caused by chromogen (3,3 ’-\ndiaminobenzidine, DAB) is automatically separated from\nthe blue hematoxylin staining into their optical density\ncounterparts. The mean optical density per cell is quan-\ntified by the segmentation method.\nRegions of interest (ROIs) were defined separately for\nglandular epithelial tissue and stromal tissue. Identifica-\ntion of cell types was accomplished through morphomet-\nric parameters such as the nuclear size, shape and staining\nintensity. A background threshold for hematoxylin stain-\ning was determined interactively. Immunostaining cut-offs\nwere determined as well (this tool differentiates between\npositive and negative cells; these were set in the dot blots).\nAll images were acquired with the same setting parame-\nters. The representative brown color (DAB chromogen)\nwas picked by the color picker tool. Positive staining cells\nwere shown in the scatter gram of forward gating tool.\nThe raw data of the analysis were imported into SPSS 21.0\n(IBM, Armonk, NY, USA) for further statistical analysis.\nThe number of NGF-, BDNF-, TrkA-, TrkB-, and p75 –\npositive immunostaining cells was divided by the total\nnumber of cells in each gland or stroma of endometriosis\ntissue (hematoxylin counterstain), yielding a percentage of\nstaining. More detailed methodology and setting parame-\nters of Histoquest® can be found in the Supplemental\nS e c t i o n( A d d i t i o n a lf i l e s1 ,2 ,3 ,4a n d5 :F i g s .S 1 ,S 2 ,S 3 ,\nS4 and S5).\nNerve fiber density\nEach stained endometriosis section was imaged at × 40\nmagnification using the above described TissueFaxs Plus\nsystems. The evaluation area was randomly set to 1 mm 2\nsurrounding the nerve fiber located nearest to the lesion.\nThe area of interest was manually highlighted with a tool\nprovided by HistoQuest® software to comprise an area of\n1m m 2 as shown in Additional file 6: Fig. S6. Thus, the\ntotal nerve fiber density was calculated by averaging the\namount of nerve fibers defined by PGP9.5 positive stain-\ning in 3 different areas of endometriosis of 1 mm 2 each\n(Additional file 6: Fig. S6). The same method was used\nto calculate the density TH-positive staining sympathetic\nnerve fibers.\nStatistical analysis\nThe Mann-Whitney U test was used to evaluate the dif-\nference between the percentage of positive immuno-\nstaining cells in PE and DIE. The student ’s t-test was\nused to compare mean nerve fiber density between\ngroups. Pearson correlation was used to analyze the\ncorrelation between neurotro phins and the nerve fibers.\nP < 0.05 was considered statistically significant at 95 %\nof confidence interval.\nResults\nStaining for BDNF and TrkB\nBDNF exhibited strong staining intensity in both gland\nand stroma cells of PE and DIE, with no differences seen\n(Fig. 1e). Nonetheless, in stroma, there was a signifi-\ncantly higher percentage of cells expressing TrkB posi-\ntive immunostaining in DIE compared to PE. No such\ndifference in TrkB staining was seen in glands of PE and\nDIE (Fig. 1f ).\nStaining for NGF and TrkA\nThe expression of NGF positive immunostaining cells\nwas higher in both the gland and stroma of PE com-\npared to the gland and stroma of DIE, as shown in Fig. 2.\nNevertheless, staining for TrkA, the main receptor of\nNGF, showed no differences in glands or stroma be-\ntween PE and DIE (Fig. 2f ).\nStaining for p75\nThere was a high percentage of p75-positive immuno-\nstaining seen in the endometrial glands of both PE and\nDIE. In the stroma, there was a significantly higher per-\ncentage of p75 staining in DIE compared to PE (Fig. 3c).\nStaining nerve fibers for PGP9.5 in PE, bowel DIE and\nnon-bowel DIE\nNerve density stained with PGP9.5 was similar between\nPE (26.27 ± 17.32) and DIE (28.19 ± 33.15), p = 0.81.\nWhen we performed sub-group analysis and separately\nevaluated the bowel and non-bowel DIE, the density of\nnerves of was significantly higher in the bowel DIE\n(57.33 ± 43.9) than in the PE ( p < 0.01) and non-bowel\nDIE (14.6. ± 8.6 p < 0.002) (Fig. 4g).\nStaining nerve fibers using TH compared to PGP9.5 in\nnon-bowel DIE\nMean density of sympathetic nerve fibers stained with\nTH was not different between PE and DIE. When the\nDIE group was split to bowel and non-bowel endometri-\nosis, the pattern of mean density of nerves for each\ngroup stained with TH was similar to that seen with\nPGP9.5, with markedly higher concentration of sympa-\nthetic nerves in bowel DIE, and similarly lower sympa-\nthetic innervation of non-bowel DIE and PE (Fig. 4h).\nStaining for nerve fibers using PGP9.5 in DIE of bowel\n(colon)\nThe plexus myentericus is located in between two\nmuscle layers, circular and longitudinal fibers. In normal\ncolon (Fig. 5a) the plexus was seen as a regular straight\nline interrupted by ganglia. In colon infiltrated by\nDewanto et al. Reproductive Biology and Endocrinology  (2016) 14:43 Page 4 of 13\n\nendometriosis, the plexus myentericus lost its longitu-\ndinal configuration, as depicted in Fig. 5b and c. Further-\nmore, the regularity for the muscle layers was lost and\ntheir structure became aberrant.\nThe correlation between neurotrophin expression and\nnerve fibers density\nWe performed correlation analyses between NGF and\nBDNF immunostaining and nerve fiber density (PGP9.5\nFig. 1 Formalin-fixed paraffin-embedded serial section of PE ( a, c) and DIE ( b, d) stained for BDNF ( a, b) and TrkB ( c, d). Original magnification ×\n100. The percentage TrkB positive immune staining cells in stroma showed a significant difference between stroma groups* P < 0.014 (f), but not\nsignificant difference in other groups ( P > 0.05) (e). Scale bar, 100 μm\nDewanto et al. Reproductive Biology and Endocrinology  (2016) 14:43 Page 5 of 13\n\nFig. 2 Formalin-fixed paraffin-embedded serial section of PE ( a, c) and DIE ( b, d) stained against NGF ( a, b) and TrkA ( c, d). Original magnifica-\ntion × 100. The percentage NGF positive immune staining cells in stromal cells showed significant difference (* P < 0.00, **P < 0.00) in glandular\nepithelial cells and stromal cells between PE and DIE ( e). No differences were seen for TrkA-positive immunostaining ( f). High degree of positive\nstaining was observed in the lumen of the endometriotic gland reflecting immune cells ( b, d). Scale bar, 100 μm\nDewanto et al. Reproductive Biology and Endocrinology  (2016) 14:43 Page 6 of 13\n\nstaining) for both gland and stroma, and for the sub-\ntypes of endometriosis (PE and non-bowel DIE). We\nchose to exclude bowel DIE in these analyses because of\nthe inability to differentiate between the intrinsic and ex-\ntrinsic innervation, thus making endometriosis-specific\nnerve growth impossible to study.\nWhen both subtypes of endometriosis tissues were\ntaken together (PE and DIE), there was no significant\ncorrelation between BDNF-positive staining and nerve\ndensity assessed by PGP9.5, neither for gland nor stroma\n(data not shown). When only the gland was assessed,\nthere was a negative correlation seen only in non-bowel\nDIE ( P < 0.007, r = −0.618). We then did analogous cor-\nrelations for NGF and PGP9.5, BDNF and TH and also\nNGF and TH.\nThere were no significant correlations between NGF-\npositive staining and nerve fiber density assessed by\nPGP9.5 as can be seen in Table 2 (NGF vs PGP9.5) and\nnerve fiber density stained with TH, except in gland PE\nthat showed a significant negative correlation ( P = 0.038,\nr = −0.406). We found a negative correlation between\nBDNF-positive staining and the density of nerve fi-\nbers stained with TH in the gland of non-bowel DIE\n(P < 0.036, r = −0.478) (Table 2).\nStaining for neurons (PGP9.5), p75 and trks receptors in\nendometriosis in submucosa of colon\nThe submucosal plexus in the bowel strongly stained\nwith p75 and TrkB antibodies, and faintly stained with\nTrkA antibody. The submucosal plexus has an import-\nant role in innervating mucous glands of the bowel. As\ndepicted in Additional file 7: Fig. S7, endometrial glands\nwere surrounded by small nerve fibers.\nIn glandular epithelial cells surrounded by p75 immu-\nnostaining cells typical formations of mitosis telophase\nwith two separated cell nuclei were identified (Fig. 6).\nFig. 3 Formalin-fixed paraffin-embedded serial section of PE ( a) and DIE ( b) stained for p75 ( a, b). Original magnification × 100. The percentage\np75 positive immune staining cells in stromal cells showed a significant difference * p < 0.027 between PE and DIE ( c). Scale bar, 100 μm\nDewanto et al. Reproductive Biology and Endocrinology  (2016) 14:43 Page 7 of 13\n\nDiscussion\nCell growth and development\nBased on the high degree of staining for BDNF in both\nglands and stroma of PE and DIE, this neurotrophin ap-\npears to play a role in the innervation of endometriosis.\nTrkB, the main receptor of BDNF is expressed more\nhighly in glands than in stroma, and highest in the\nstroma of DIE, but with high variability. The other re-\nceptor for BDNF, p75, showed similar expression,\nnamely high expression in glands of PE and DIE, moder-\nate expression in stroma of DIE and relatively low ex-\npression in stroma of PE. BDNF mRNA expression has\nFig. 4 Formalin-fixed paraffin-embedded serial section of PE ( a, b, c, d) and DIE non bowel ( e, f) stained for PGP9.5 ( a, c, e) and . Original\nmagnification × 100 (a, b) and × 200 (c, d, e, f). Picture c and d are an inset from ( a)t o( b) respectively. Red arrow showed nerve fibers. Nerve fibers\nstained with PGP9.5 were significantly different between PE and non-bowel DIE endometriosis *p < 0.041, between PE and bowel **p < 0.01 and\nbetween non-bowel and bowel *** p <0 . 0 0 2g. Nerve fibers stained with TH were not significantly different between PE and non-bowel DIE\nendometriosis P = 0.21, between PE and bowel P = 0.06 and significantly different between non-bowel and bowel ****p <0 . 0 4(h). Scale bar, 100 μm\nDewanto et al. Reproductive Biology and Endocrinology  (2016) 14:43 Page 8 of 13\n\npreviously been shown in eutopic endometrium of\nwomen with endometriosis [20] and also in normal hu-\nman and mammalian uterus as well as endometrium\n(glandular epithelium and stroma) [21]. We confirm, for\nthe first time in histological sections, that BDNF is\npresent in endometriosis lesions, both DIE and in peri-\ntoneal lesions.\nNGF and BDNF are both important for axonal growth\nof sensory neurons, but each induces growth of a differ-\nent type of sensory neuron. In vitro studies in chick dor-\nsal root ganglion showed that neuronal growth cones\nturned and migrated under NGF-coated beads through\nthe expression of TrkA receptors [22]. Another study\nshowed that visceral afferent neurons in the nodose/pe-\ntrosal sensory ganglion complex innervated vascular\nafferents that express high levels of BDNF in the devel-\nopment of arterial baroreceptors. The survival of these\nneurons was reduced by TrkB-Fc blocking [23]. Analo-\ngous mechanisms may be responsible for inducing\nsensory neuron growth in endometriosis lesions via both\nTrkA and TrkB.\nBDNF and NGF play different roles in sensory nerve\ndevelopment, with BDNF influencing axonal branching\nand the growth of lathellipodia and NGF influencing\naxonal elongation of sensory neurons from the dorsal\nroot ganglion [24]. Furthermore, endometriosis is an\nestrogen-dependent disease in which lesions stimulate\ntheir own growth by producing estrogen via aromatase\nactivity [25] and also via the mechanism of tissue injury\nand repair [26]. Wessels et al. showed that estrogen ex-\nposure may activate BDNF-TrkB pathways in a mouse\nmodel, exerting wide ranging effects such as neural de-\nvelopment, cell differentiation, growth and maintenance,\nangiogenesis, proliferation, and resistance to apoptosis\n[17]. In clinical studies in humans, Wessels et al. re-\nported that that plasma BDNF concentrations were sig-\nnificantly higher in women with endometriosis than in\ncontrols, whereas other neurotrophins, NGF and NT4/5,\nwere not different [18].\nIt is known that estrogen may have local proliferative\nactions as well as neuromodulatory effects on the in-\nnervation of endometeriosis [19]. In this retrospective\nstudy, we were not able to fully account for the men-\nstrual cycle phase at which the histological sample was\nobtained. However, in a recent report, there was no ef-\nfect of menstrual cycle phase on circulating BDNF levels\nFig. 5 Ganglion in plexus myentericus of healthy human colon ( a)\nand colon endometriosis lesion ( b, c) stained with PGP9.5. Original\nmagnification 40 ×. Black arrows showed ganglions, red arrows\nshowed nerve fibers, blue arrows showed endometriosis. Irregular\nform of plexus myentericus depicted in picture b and c. Scale\nbar, 200 μm\nTable 2 The correlation between BDNF-positive immunostain-\ning and the density of nerve fibers stained with PGP9.5\nGland Stroma\nBDNF vs PGP9.5 P value r P value r\nPE 0.196 −0.202 0.324 0.109\nDIE non-bowel 0.007 −0.618 0.426 −0.053\nNGF vs PGP9.5\nPE 0.218 −0.185 0.248 0.149\nDIE non-bowel 0.187 0.247 0.404 −0.069\nBDNF vs TH\nPE 0.160 −0.235 0.299 0.125\nDIE non-bowel 0.036 −0.478 0.368 −0.095\nNGF vs TH\nPE 0.038 −0.406 0.483 0.010\nDIE non-bowel 0.322 0.130 0.289 −0.156\nStatistically significant differences are shown in bold\nDewanto et al. Reproductive Biology and Endocrinology  (2016) 14:43 Page 9 of 13\n\nin women with endometriosis [18]. In mice, estrogen ex-\nposure after ovariectomy significantly increases uterine\nBDNF, but the hormonal fluctuations of the murine es-\ntrous cycle do not [17]. Importantly, none of the women\nin this study used hormonal medication for at least\n3 months prior to sample collection, as hormonal treat-\nment has been shown to decrease the nerve fiber density\nin peritoneal endometriosis lesions [27].\nWe found that NGF was higher in PE than in DIE, in\nboth glands and stroma. Previous studies looked only at\nendometrioma, adenomyosis, peritoneal lesions [13, 28,\n29] and eutopic endometrium from endometriosis pa-\ntients [20]. The main receptor of NGF , TrkA was not dif-\nferent between PE and DIE in both locations. These\nresults differ from those of Anaf et al. who found that\ndeep adenomyotic lesions had higher expression of NGF\nimmunohistochemically than peritoneal endometriosis\n[13] and imply that DIE and adenomyosis are not compar-\nable entities when it comes to neurotrophin expression.\nWe found a low NGF expression as well as low nerve\nfiber density in DIE, especially in non-bowel DIE.\nPerhaps, the loss of NGF-TrkA signaling transduction\ncauses failure of NGF-dependent neuron to survive [30].\nFurthermore, in sensory developmental studies, NGF is\nimportant for nociceptor development and BDNF is im-\nportant for mechanoreceptor development [31]. Our re-\nsults are consistent with those of Arellano et al. who\nimplicated the peritoneum as an important location for\npathogenesis and pain generation in endometriosis [32].\nThus, our results support the concept that NGF is in-\nvolved in neuronal development and likely pain gener-\nation from peritoneum in endometriosis [12, 13, 32].\nMany studies have evaluated the role of Trk receptors in\nthe invasiveness or degree of progressiveness in cancer.\nTrkA receptors appear to promote the growth and metas-\ntasis in breast cancer [33, 34], while TrkB receptors\npromote invasion in choriocarcinoma cells [35]. We found\nthat TrkB expression was high in DIE, especially in stro-\nmal cells. This pattern is similar to that seen for p75 ex-\npression, but not for TrkA expression. This finding could\nbe secondary to an effect of estrogen [17]. Therefore, we\nspeculate that p75 receptor is involved in survival rather\nthan apoptosis in endometriosis.\nBowel innervation\nOn first analysis, it appears that there is no difference in\nmean nerve density between PE and DIE. However, crit-\nical differences appear when DIE from bowel and non-\nbowel locations was evaluated separately, as also shown\nby Wang [8]. We found that the sub-group of non-\nbowel DIE endometriosis actually had a lower density of\ninnervation than the PE.\nThe DIE from bowel showed markedly higher innerv-\nation, both total nerve density as well as the density of\nsympathetic nerves stained with TH. The innervation of\nbowel is highly complex, stemming from both intrinsic\n(from the enteric nervous system) as well as extrinsic\nnerves (from autonomic nervous system —parasympa-\nthetic and sympathetic nervous system). In addition, the\nintestine has sensory afferents originating from the vagus\nnerve (nodose ganglion) [36, 37] and sensory afferents ori-\nginating from dorsal root ganglion [38]. A previous study\nlikewise demonstrated that normal bowel has a rich in-\nnervation base on nerve fiber density area stained with\nPGP9.5 and other markers [39, 40]. The plexus of\nAuerbach and Meissner (submucosal plexus) likewise\nexpress PGP9.5, and may be damaged by the invasion\nof an endometriosis lesion, as we also showed [41 –43].\nThus, it is very difficult and likely error-prone to com-\npare the endometriosis-influenced innervation of PE to\nbowel endometriosis because the intrinsic and extrin-\nsic innervations of bowel cannot be differentiated with\nFig. 6 Mitotic activity in glandular epithelial cells surrounded by p75 immunostaining cells. Original magnification was × 400. Scale bar, 100 μm\nDewanto et al. Reproductive Biology and Endocrinology  (2016) 14:43 Page 10 of 13\n\nimmunohistochemistry. In future studies, investigators\nshould focus on DIE from non-bowel sources when\nmaking comparisons to PE.\nTable 3 summarizes the relative abundance of NGF,\nBDNF, TrkA, TrkB, p75, and nerve fibers densities for\nPE, non-bowel and bowel DIE.\nCorrelation between cell expressing neurotrophins and\nthe nerve fibers density\nBDNF showed a negative correlation to total nerve fibers\nstained with PGP9.5 in glands of DIE non-bowel only,\nbut not in other sites and not in the stroma. These re-\nsults may reflect the phenomenon of axonal pruning,\nthrough which BDNF may induce pruning in neurons\nand thus lead to a lower nerve fiber density in the tissue.\nOur results are consistent with those of Singh et al. who\nshowed that BDNF via p75 receptor influenced the de-\nvelopment of sympathetic axon pruning, despite the\nsimultaneous presence of NGF at the same site [44].\nWe found no correlation between NGF and nerve fiber\ndensity, as also reported by others [45, 46]. NGF com-\npetes for the same axons as BDNF to determine the fate\nof nerve fibers/neurons. In in vitro studies, the pruning/\napoptosis effects of BDNF override the growth effects of\nNGF. Induction of pruning is stronger than the induction\nof growth [44]. This would explain why NGF-positive im-\nmunostaining did not correlate to a higher density of\nnerve fibers but instead there was a negative correlation\nbetween BDNF and nerve fiber density seen in DIE.\nWe did find a negative correlation between BDNF-\npositive staining and the density of nerve fibers stained\nwith TH in the gland of non-bowel DIE. These results\nare consistent with the study by Krizsan-Agbas et al.\nwhich showed in a rat model that BDNF suppresses\nsympathetic neurite growth and that this effect is medi-\nated by estrogen [45].\nSemi-automated counting\nTo our knowledge, ours is the first study to employ a\nsemi-quantitative method to analyzing immunohisto-\nchemical staining in endometriosis [47]. While the soft-\nware used does not replace the skills and expertise of\nthe experienced human observer (pathologist), it is gain-\ning popularity in the research and clinical setting [48].\nThe emergence of computerized image analysis systems\nfor accurate analysis of immunohistochemistry specimen\nis increasingly needed. In breast cancer, a disease where\nthe degree of estrogen receptor and progesterone receptor\nexpression predicts outcome,Walker et al. argue that com-\nputerized image analysis systems present more accurate\nmeans of quantification. It is well accepted that manual\ncounting is time consuming [47] and automated imaging\nmethods are immune to fatigue and subjectivity [49].\nConclusions\nThe neurotrophins and their receptors are part of a\ncomplex signaling system that are present in endometri-\nosis. We showed that endometriotic lesions, especially\nepithelial glandular cells and stromal cells, express neu-\nrotrophins BDNF and NGF and their receptors, TrkA,\nTrkB and p75. The spatial arrangement of these agonists\nand receptors suggests an autocrine function in endo-\nmetriosis, though a clear causative picture does not\nemerge given the redundancy and complexity of the sig-\nnaling system. BDNF has a stronger binding affinity than\nNGF to the p75 receptor, likely inducing sympathetic\nTable 3 Relative abundance of the neurotophins NGF and BDNF, their receptors and nerve fibers in PE and DIE from non-bowel\nsources\nRed arrows and * highlight statistically significant differences for each neurotrophin/receptor/nerve fiber in gland or stroma\nDewanto et al. Reproductive Biology and Endocrinology  (2016) 14:43 Page 11 of 13\n\nnerve axonal pruning in DIE, resulting in the lower\nnerve fiber density seen in DIE. The differences in neu-\nrotrophin expression between PE and DIE may reflect\nthe differencing innervations and cell fates, namely\ngrowth and infiltration and have been shown to be\nhormone-mediated. Thus, in future studies, differences\nin local estrogen action as measured by the distribution\nof estrogen receptors in co-localization with the BDNF\nneurotrophin signaling system and the density of nerves\nshould be evaluated to more closely delineate this com-\nplex system and to further explain differences between\nPE and DIE.\nAdditional files\nAdditional file 1: Figure S1. The ROIs were determined manually,\nseparating glandular epithelial tissue and stromal tissue ( a). The nuclear\nmorphometric and staining parameter enable the identification in\nepithelial cells ( b). Gray levels of separated “blue staining channel ” with\nsegmented structures as an overlay ( c). DAB staining mask as color\nlabeled areas ( d). Backward visualization positive stained cells ( e),\nBackward visualization negative stained cells ( f). (JPG 944 kb)\nAdditional file 2: Figure S2. The selected endometriosis tissue\nrequired the presence of glandular epithelial cells and stromal cells. ROI\nwere developed by selecting ‘custom’ mode applied to separate lumen,\nepithelial gland tissue and stromal tissue. The lumen ROI was excluded\nfrom analysis. Black arrow showing custom mode ( a), ROI has been\ndeveloped, blue arrow showing lumen, red arrow showing epithelial\ntissue, and black arrow showing stromal tissue. (JPG 282 kb)\nAdditional file 3: Figure S3. Markers determination. Nucleus\nidentification was determined by selecting blue color as marker. First, the\nmarkers button was chosen. Hematoxillin staining button was chosen.\nColor picker button was chosen. Blue color taken from the cell that may\nrepresent all nucleus. With the same procedure, brown color was chosen\nto detect brown color as result of DAB/IHC staining in cells. Blue color\nwas chosen to detect hematoxillin staining in nucleus ( a), brown color\nwas chosen to detect result of IHC staining in cells ( b). (JPG 203 kb)\nAdditional file 4: Figure S4. Setting parameter for nuclei size. The\nnucleus size was determined as depicted in Fig. 4a. It was based on blue\ncolor detection (hematoxylin staining). Brown color expressed by cell was\nrestricted from interior radius −4,81 μm to exterior radius 25.3 μm( a) and\nbrown color resulted from IHC staining ( b). (JPG 191 kb)\nAdditional file 5: Figure S5. Scatter gram showing the result of\nhematoxylin staining ( a) and DAB staining ( b). The last step was\ndetermination of the immune positive or negative expression resulting\nfrom IHC staining. The ‘cut off ’ option was used to set new values for\none axis (y axis in DAB staining). The cut off menu was set at 18 and only\nfor DAB staining and applied to all ROI ( b). (JPG 150 kb)\nAdditional file 6: Figure S6. Manual counting of nerve fibers. Region\nof interests were selected randomly and marked with a border to obtain\nan area of 1 mm 2 each. A single nerve fiber was marked manually by\nusing a tool available in Histoquest® software. (JPG 565 kb)\nAdditional file 7: Figure S7 . Endometriosis lesion in submucosa of\ncolon stained with antibody anti PGP9.5 ( a), p75 ( b), and TrkB ( c) and anti\nTrkA ( d). Red arrow shows nerve fiber, black arrow shows ganglion-like\nform. Original magnification × 200. Scale bare, 100 μm. (JPG 713 kb)\nAbbreviations\nASRM: American Society of Reproductive Medicine; BDNF: brain-derived\nneurotrophic factor; DAB: 3,3 ’-diaminobenzidine; DIE: deep infiltrating\nendometriosis; H&E: hematoxyline and eosin; NGF: nerve growth factor; NT-3:\nneurotrophins-3; NT-4: neurotrophins-4; NT-5: neurotrophins-5; PE: Peritoneal\nendometriosis; PGP9.5: Protein Gene Product 9.5; ROI: region of interest; TH:\ntyrosine hydroxylase; TNF: tumor necrosis factor; Trk: tropomyosin kinase\nAcknowledgments\nThe antibody anti p75 was a kind gift from Professor Dr. Louis F. Reichardt\nfrom University of California, San Francisco, California.\nFunding\nThere were no sources of external funding for this study.\nDr. Dewanto was supported by scholarships from the Indonesian Directorate\nGeneral of Higher Education, Ministry of National Education, Indonesia; i-med\nForschungsstipendien (Nachwuchsförderung), Medical University of Innsbruck,\nand ASEA-UNINET.\nAvailability of data and material\nThe dataset analysed during the study is available upon request from the\ncorresponding author.\nAuthors’ contributions\nAD performed, analyzed and interpreted the histological examinations,\nperformed statistical analyses and contributed to writing the manuscript; JD\nanalyzed and interpreted the histological examinations, assisted with\nstatistical analyses, and contributed to conceptualization of the study; RG\nanalyzed and interpreted the histological examinations, assisted with\nstatistical analyses, and contributed to conceptualization of the study; SM\nanalyzed and interpreted the data, conceptualized the study and edited the\nmanuscript for key content; ASF analyzed and interpreted the data,\nconceptualized the study and edited the manuscript for key content; LW\nanalyzed and interpreted the data, conceptualized the study and edited the\nmanuscript for key content; BS analyzed and interpreted the data, assisted\nwith statistical analyses, conceptualized the study and was a major\ncontributor to writing, editing and revising the manuscript. All authors read\nand approved the final manuscript.\nAuthors’ information\nNot applicable.\nCompeting interests\nThe authors declare that they have no competing interests.\nConsent for publication\nSubjects gave their informed consent for research participation.\nEthics approval and consent to participate\nThe study was approved by the Ethics Committee of the Medical University\nof Innsbruck, UN5130, 02.08.2013.\nAuthor details\n1Department of Gynecological Endocrinology and Reproductive Medicine,\nMedical University of Innsbruck, Anichstrasse 35, Innsbruck 6020, Austria.\n2Department of Obstetrics and Gynecology, Gadjah Mada University,\nYogyakarta, Indonesia. 3Department of Otorhinolaryngology, Medical\nUniversity of Innsbruck, Innsbruck, Austria. 4Endometriosis Centre Charité,\nDepartment of Gynecology - Campus Benjamin Franklin, Charité\nUniversitätsmedizin Berlin, Berlin, Germany.\nReceived: 2 April 2016 Accepted: 29 July 2016\nReferences\n1. Sourial S, Tempest N, Hapangama DK. Theories on the pathogenesis of\nendometriosis. Int J Reprod Med. 2014;2014:179515.\n2. Koninckx PR, et al. Deeply infiltrating endometriosis is a disease whereas\nmild endometriosis could be considered a non-disease. Ann N Y Acad Sci.\n1994;734:333–41.\n3. Tosti C, et al. Pathogenetic Mechanisms of Deep Infiltrating Endometriosis.\nReprod Sci. 2015;22(9):1053 –9.\n4. Tokushige N, et al. Nerve fibres in peritoneal endometriosis. Hum Reprod.\n2006;21(11):3001–7.\n5. Mckinnon BD, et al. Inflammation and nerve fiber interaction in\nendometriotic pain. Trends Endocrinol Metab. 2015;26(1):1 –10.\n6. Mechsner S, et al. Oxytocin receptor expression in smooth muscle cells of\nperitoneal endometriotic lesions and ovarian endometriotic cysts. Fertil\nSteril. 2005;83 Suppl 1:1220 –31.\nDewanto et al. Reproductive Biology and Endocrinology  (2016) 14:43 Page 12 of 13\n\n7. Arnold J, et al. Neuroimmunomodulatory alterations in non-lesional\nperitoneum close to peritoneal endometriosis. Neuroimmunomodulation.\n2013;20(1):9–18.\n8. Wang G, et al. Rich innervation of deep infiltrating endometriosis. Hum\nReprod. 2009;24(4):827 –34.\n9. Barcena De Arellano ML, et al. Evidence of neurotrophic events due to\nperitoneal endometriotic lesions. Cytokine. 2013;62(2):253 –61.\n10. Caporali A, Emanueli C. Cardiovascular actions of neurotrophins. Physiol Rev.\n2009;89(1):279–308.\n11. Huang EJ, Reichardt LF. Neurotrophins: roles in neuronal development and\nfunction. Annu Rev Neurosci. 2001;24:677 –736.\n12. Tarjanne S, et al. Use of hormonal therapy is associated with reduced nerve\nfiber density in deep infiltrating, rectovaginal endometriosis. Acta Obstet\nGynecol Scand. 2015;94(7):693 –700.\n13. Anaf V, et al. Hyperalgesia, nerve infiltration and nerve growth factor\nexpression in deep adenomyotic nodules, peritoneal and ovarian\nendometriosis. Hum Reprod. 2002;17(7):1895 –900.\n14. Arnold J, et al. Imbalance between sympathetic and sensory innervation in\nperitoneal endometriosis. Brain Behav Immun. 2012;26(1):132 –41.\n15. Kuteken FS, et al. Expression of neurotrophic and inflammatory mediators in\nrectosigmoid endometriosis. Rev Bras Ginecol Obstet. 2012;34(12):568 –74.\n16. Barcena De Arellano ML, et al. Eutopic endometrium from women with\nendometriosis does not exhibit neurotrophic properties. J Neuroimmunol.\n2012;249(1–2):49–55.\n17. Wessels JM, et al. Estrogen induced changes in uterine brain-derived\nneurotrophic factor and its receptors. Hum Reprod. 2015;30(4):925 –36.\n18. Wessels JM, et al. Assessing brain-derived neurotrophic factor as a novel\nclinical marker of endometriosis. Fertil Steril. 2016;105(1):119 –28.\n19. Liang Y, Yao S. Potential role of estrogen in maintaining the imbalanced\nsympathetic and sensory innervation in endometriosis. Mol Cell Endocrinol.\n2016;424:42–9.\n20. Browne AS, et al. Proteomic identification of neurotrophins in the eutopic\nendometrium of women with endometriosis. Fertil Steril. 2012;98(3):713 –9.\n21. Wessels JM, et al. The brain-uterus connection: brain derived neurotrophic\nfactor (BDNF) and its receptor (Ntrk2) are conserved in the mammalian\nuterus. PLoS One. 2014;9(4), e94036.\n22. Gallo G, Lefcort FB, Letourneau PC. The trkA receptor mediates growth cone\nturning toward a localized source of nerve growth factor. J Neurosci.\n1997;17(14):5445–54.\n23. Brady R, et al. BDNF is a target-derived survival factor for arterial\nbaroreceptor and chemoafferent primary sensory neurons. J Neurosci.\n1999;19(6):2131–42.\n24. Lentz SI, et al. Neurotrophins support the development of diverse sensory\naxon morphologies. J Neurosci. 1999;19(3):1038 –48.\n25. Bulun SE, et al. Aromatase and endometriosis. Semin Reprod Med.\n2004;22(1):45–50.\n26. Leyendecker G, Wildt L. A new concept of endometriosis and adenomyosis:\ntissue injury and repair (TIAR). Horm Mol Biol Clin Investig. 2011;5(2):125 –42.\n27. Tokushige N, et al. Effect of progestogens and combined oral contraceptives\non nerve fibers in peritoneal endometriosis. Fertil Steril. 2009;92(4):1234–9.\n28. Barcena De Arellano ML, et al. Neurotrophin expression is not affected in\nuteri of women with adenomyosis. J Mol Neurosci. 2012;47(3):495 –504.\n29. Kajitani T, et al. Possible involvement of nerve growth factor in\ndysmenorrhea and dyspareunia associated with endometriosis. Endocr J.\n2013;60(10):1155–64.\n30. Indo Y. Neurobiology of pain, interoc eption and emotional response: lessons\nfrom nerve growth factor-dependent neurons. Eur J Neurosci. 2014;39(3):375–91.\n31. Farinas I, et al. Regulation of neurogenesis by neurotrophins in developing\nspinal sensory ganglia. Brain Res Bull. 2002;57(6):809 –16.\n32. Barcena De Arellano ML, Mechsner S. The peritoneum —an important factor\nfor pathogenesis and pain generation in endometriosis. J Mol Med (Berl).\n2014;92(6):595–602.\n33. Lagadec C, et al. TrkA overexpression enhances growth and metastasis of\nbreast cancer cells. Oncogene. 2009;28(18):1960 –70.\n34. Davidson B, et al. Altered expression and activation of the nerve growth factor\nreceptors TrkA and p75 provide the first evidence of tumor progression to\neffusion in breast carcinoma. Breast Cancer Res Treat. 2004;83(2):119–28.\n35. Kawamura K, et al. Suppression of choriocarcinoma invasion and metastasis\nfollowing blockade of BDNF/TrkB signaling. Cancer Med. 2013;2(6):849 –61.\n36. Straub RH, et al. The role of the sympathetic nervous system in intestinal\ninflammation. Gut. 2006;55(11):1640 –9.\n37. Furness JB. The enteric nervous system: normal functions and enteric\nneuropathies. Neurogastroenterol Motil. 2008;20 Suppl 1:32 –8.\n38. Su HC, et al. Dual intrinsic and extrinsic origins of CGRP- and\nNPY-immunoreactive nerves of rat gut and pancreas. J Neurosci.\n1987;7(9):2674 –87.\n39. Oh JT, et al. Morphometric evaluation of PGP9.5 and NCAM expressing\nnerve fibers in colonic muscle of patients with Hirschsprung ’s disease.\nYonsei Med J. 2002;43(1):31 –6.\n40. Southwell BR, et al. Decrease in nerve fibre density in human sigmoid colon\ncircular muscle occurs with growth but not aging. Neurogastroenterol\nMotil. 2010;22(4):439–45. e106.\n41. Tomita T. Localization of nerve fibers in colonic polyps, adenomas, and\nadenocarcinomas by immunocytochemical staining for PGP 9.5. Dig Dis Sci.\n2012;57(2):364–70.\n42. Sams VR, et al. Evaluation of PGP9.5 in the diagnosis of Hirschsprung ’s\ndisease. J Pathol. 1992;168(1):55 –8.\n43. Ferrero S, et al. Loss of sympathetic nerve fibers in intestinal endometriosis.\nFertil Steril. 2010;94(7):2817 –9.\n44. Singh KK, et al. Developmental axon pruning mediated by BDNF-p75NTR-\ndependent axon degeneration. Nat Neurosci. 2008;11(6):649 –58.\n45. Krizsan-Agbas D, et al. Oestrogen regulates sympathetic neurite outgrowth\nby modulating brain derived neurotrophic factor synthesis and release by\nthe rodent uterus. Eur J Neurosci. 2003;18(10):2760 –8.\n46. Brauer MM, et al. The role of NGF in pregnancy-induced degeneration and\nregeneration of sympathetic nerves in the guinea pig uterus. J Auton Nerv\nSyst. 2000;79(1):19 –27.\n47. Walker RA. Quantification of immunohistochemistry —issues concerning\nmethods, utility and semiquantitative assessment I. Histopathology.\n2006;49(4):406 –10.\n48. Schlederer M, et al. Reliable quantification of protein expression and cellular\nlocalization in histological sections. PLoS One. 2014;9(7), e100822.\n49. Taylor CR, Levenson RM. Quantification of immunohistochemistry —issues\nconcerning methods, utility and semiquantitative assessment II. Histopathology.\n2006;49(4):411–24.\n•  We accept pre-submission inquiries \n  Our selector tool helps you to find the most relevant journal\n  We provide round the clock customer support \n  Convenient online submission\n  Thorough peer review\n  Inclusion in PubMed and all major indexing services \n  Maximum visibility for your research\nSubmit your manuscript at\nwww.biomedcentral.com/submit\nSubmit your next manuscript to BioMed Central \nand we will help you at every step:\nDewanto et al. Reproductive Biology and Endocrinology  (2016) 14:43 Page 13 of 13","source_license":"CC0","license_restricted":false}