nerve fibres in peritoneal endometriosis

Human Reproduction Vol.21, No.11 pp. 3001–3007, 2006 doi:10.1093/humrep/del260

Advance Access publication September 1, 2006.

© The Author 2006. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. 3001For Permissions, please email: [email protected]

Nerve fibres in peritoneal endometriosis

Natsuko Tokushige1,3, Robert Markham1, Peter Russell2 and Ian S.Fraser1

1Department of Obstetrics and Gynaecology, Queen Elizabeth II Research Institute for Mothers and Infants and 2Department of Pathology, University of Sydney, Sydney, Australia3To whom correspondence should be addressed at: Department of Obstetrics and Gynaecology, Queen Elizabeth II Research Institute for Mothers and Infants, University of Sydney, Sydney, NSW 2006, Australia. E-mail: [email protected]

BACKGROUND: Endometriosis is a gynaecological disease that can be associated with severe pelvic pain; however,the mechanisms by which pain is generated remain unknown. METHODS: Peritoneal endometriotic lesions and nor-mal peritoneum were prepared from women with and without endometriosis (n = 40 and 36, respectively). Specimenswere also prepared from endosalpingiosis lesions (n = 9). These sections were stained immunohistochemically withantibodies against protein gene product 9.5, neurofilament (NF), nerve growth factor (NGF), NGF receptor p75(NGFRp75), substance P (SP), calcitonin gene-related peptide (CGRP), acetylcholine (ACh) and tyrosine hydroxylase(TH) to demonstrate myelinated, unmyelinated, sensory, cholinergic and adrenergic nerve fibres. RESULTS: Therewere significantly more nerve fibres identified in peritoneal endometriotic lesions than in normal peritoneum(P < 0.001) or endosalpingiosis lesions (P < 0.001). These nerve fibres were SP, CGRP, ACh or TH immunoreac-tive. Many of these markers were co-localized. There was an intense NGF immunoreactivity near endometrioticglands, and NGFRp75 immunoreactive nerve fibres were present near endometriotic glands and blood vessels inthe peritoneal endometriotic lesions. CONCLUSIONS: Peritoneal endometriotic lesions were innervated by sen-sory Ad, sensory C, cholinergic and adrenergic nerve fibres. These nerve fibres may play an important role in themechanisms of pain generation in this condition.

Key words: endometriosis/immunohistochemistry/nerve fibres/pain

Introduction

Endometriosis is a gynaecological disease defined by the histo-logic presence of endometrial glands and stroma outside theuterine cavity, most commonly implanted over visceral andperitoneal surfaces within the female pelvis. This disease isa benign disorder; however, it exhibits cellular proliferation,cellular invasion and neoangiogenesis (Giudice et al., 1998).Although the exact prevalence of endometriosis in the generalpopulation is not clear, the prevalence in women of reproduc-tive age is estimated to range between 10 and 15% (Lebovicet al., 2001). Endometriosis occurs throughout the pelvic cav-ity including pelvic organs such as the ovaries, pelvic perito-neum, uterus, Fallopian tube, vagina, rectovaginal septum,intestinal tract and the ureter.

Endometriosis is usually associated with pelvic pain such aschronic dysmenorrhoea, intermenstrual abdominal and pelvicpain, back pain, dysuria, dyschezia and dyspareunia. However,the relationship between pain and endometriosis is not wellunderstood.

Several researchers have investigated the presence of nervefibres in endometriotic lesions. Tamburro et al. (2003) demon-strated nerve fibres that expressed transforming growth factor-β1(TGF-β1) in endometriotic lesions from women with dysmen-orrhoea associated with endometriosis. Tulandi et al. (2001)

demonstrated neurofilament (NF) immunoreactive nerve fibresin endometriotic lesions in women with endometriosis, andthey reported that the distance between endometrial glands andnerve fibres in endometriotic lesions from women with painwas closer than in women with no pain. Berkley et al. (2004)demonstrated nerve fibres stained with calcitonin gene-relatedpeptide (CGRP), substance P (SP) and vesicular monoaminetransporter in endometriotic lesions in a rat model. We haverecently demonstrated the unexpected finding of multiple smallunmyelinated nerve fibres in the eutopic endometrium ofwomen with laparoscopically confirmed endometriosis butnot in the endometrium of women without endometriosis(Tokushige et al., 2006).

We have studied innervation with different types of spe-cific immunohistochemical neuronal markers in humanperitoneal endometriotic lesions from women with visuallyand biopsy proven endometriosis. Markers included poly-clonal rabbit anti-protein gene product 9.5 (PGP9.5) (Lundberget al., 1988), monoclonal mouse anti-human NF (Schlaepfer,1987), polyclonal rabbit anti-SP (Berkley et al., 2004), rabbitanti-CGRP (Berkley et al., 2004), polyclonal rabbit anti-acetylcholine (anti-ACh) (Janossy et al., 1998) and mono-clonal anti-tyrosine hydroxylase (anti-TH) (Calder et al.,1998) for myelinated, unmyelinated, sensory Aδ, sensory C,

by guest on March 16, 2016

http://humrep.oxfordjournals.org/

Dow

nloaded from


N.Tokushige et al.

3002

cholinergic and adrenergic nerve fibres. PGP9.5 is a highlyspecific pan-neuronal marker and is also expressed in neuroen-docrine cells. NF is a highly specific marker for myelinatednerve fibres. SP and CGRP are sensory nerve fibre markers,and they can be present in both Aδ and C nerve fibres. ACh ispresent in sympathetic preganglionic neurons and parasympa-thetic preganglionic and postganglionic neurons, and it is aspecific marker for cholinergic nerve fibres. TH is present insympathetic postganglionic neurons, and it is a specific markerfor adrenergic nerve fibres. There is no cross-reactivity in theseantibodies. We also used polyclonal rabbit anti-nerve growthfactor (anti-NGF) and monoclonal mouse anti-human NGFreceptor p75 (NGFRp75), a low-affinity NGFR, to find outwhether these substances were associated with the developmentand survival of sensory, cholinergic and adrenergic nerve fibresin peritoneal endometriotic lesions in women with endometrio-sis. To demonstrate the relationship between nerve fibres andblood vessels, we performed double staining using PGP9.5 andCD34, a marker for endothelial cells.

We also studied nerve fibres in peritoneal lesions of endos-alpingiosis to see whether this related and commonly co-exist-ent pathological process increased nerve density in peritonealendometriotic lesions. Endosalpingiosis is defined as the pres-ence of epithelial inclusions, typified by tubal-like epithelium,with no endometrial cytogenic stroma, and the common pres-ence of microcalcification (psammoma bodies). Such lesionsmost commonly involve the peritoneal surface of the uterus,Fallopian tubes, ovaries or Pouch of Douglas.

Materials and methods

Collection of tissue

This study was approved by the Human Ethics Committees of theCentral Sydney Area Health Service and the University of Sydney,and all women gave their informed consent for participation.

Peritoneal endometriosis tissue was collected from 40 women withendometriosis (mean age 33.2 years; range 16–46 years) and normalperitoneum was also collected from 36 women without endometriosis(mean age 34.9 years; range 17–57 years) who underwent laparoscopycombined with hysteroscopy. The endometriosis patients all com-plained of dysmenorrhoea and a range of related pain symptoms. The36 normal peritoneum tissue samples collected from women withoutlaparoscopic evidence of endometriosis were from patients undergo-ing tubal sterilization, assessment before tubal reanastomosis or inves-tigation of infertility (without any recognizable pathology).

Peritoneal endosalpingiosis lesions were identified in biopsy mate-rial from nine women (mean age 27.6 years; range 14–44 years) whohad undergone laparoscopy combined with hysteroscopy for provenor suspected endometriosis.

The presence and nature of the endometriosis and endosalpingiosislesions in biopsies were confirmed by an experienced gynaecologicalpathologist (P.R.).

The severity of pain was not assessed systematically and prospec-tively in this preliminary observational study, but detailed clinicalinformation was recorded in a standard format. All endometriosispatients were staged according to the revised American Fertility Soci-ety (AFS) score (American Society for Reproductive Medicine,1996), and ranged from stages I to IV. None of the women hadreceived medical therapy for endometriosis before endoscopy and tis-sue sampling.

Immunohistochemistry

After resection, the specimens were immediately fixed in 10% neutralbuffered formalin for ∼18–24 h, processed and embedded in paraffinwax according to a standard protocol. Each section was cut at 4 μmand routinely stained with haematoxylin and eosin. Antigen retrievaltechniques for PGP9.5, NF and NGFRp75 were used. Serial sections,cut at 4 μm, were immunostained using antibodies for polyclonal rab-bit anti–PGP9.5 (dilution 1 : 700; Dako Australia, Sydney, Australia)(a highly specific pan-neuronal marker) (Lundberg et al., 1988), mon-oclonal mouse anti-human NF protein (dilution 1 : 200; Dako Australia)(a highly specific marker for myelinated nerve fibres) (Schlaepfer,1987), polyclonal rabbit anti-human NGF (dilution 1 : 1200; Santa CruzBiotechnology, CA, USA), monoclonal mouse anti-human NGFRp75(dilution 1 : 700; Dako Australia) (a low-affinity NGFR) and polyclo-nal rabbit anti-ACh (dilution 1 : 2500; Chemicon, CA, USA) for 30min at room temperature. Serial sections were also immunostainedwith rabbit anti-SP (dilution 1 : 4000; Serotec, NC, USA), polyclonalrabbit anti-CGRP (dilution 1 : 20 000; Sigma, MO, USA) and TH(dilution 1 : 1200; Sigma) overnight at 4°C. Sections stained withPGP9.5, NF, NGFRp75, TH and ACh were washed in Tris Buffer,incubated with Envision-labelled polymer-AP mouse/rabbit (DakoAustralia) for 30 min and stained with permanent fast red chromogen(Dako Australia) for 10 min. Sections stained with NGF, SP and CGRPwere washed in Tris Buffer, incubated with Envision+ dual link systemperoxidase (Dako Australia) for 30 min and stained with Liquid 3,3′-diaminobenzidine (DAB) + Substrate (Dako Australia) for 10 min.

For double staining, serial sections were immunostained with CD34(dilution 1 : 700; Dako Australia) for 30 min at room temperature, washedin Tris Buffer, incubated with Envision+ dual link system peroxidase for30 min and stained with Liquid DAB + Substrate for 10 min. The sectionswere then treated with acid block (Dako Australia) for 3 min and stainedwith polyclonal rabbit anti-PGP9.5 (dilution 1 : 500) for 30 min, washedin Tris Buffer, incubated with Envision-labelled polymer-AP mouse/rab-bit for 30 min and stained with permanent fast red chromogen for 10 min.

All immunostaining was carried out on a Dako Autostainer ModelS3400 (Dako USA, Carpinteria, CA, USA). Images of the sectionswere captured using Olympus microscope BX51 and digital cameraDP70 (Olympus, Tokyo, Japan). We used normal skin as a positivecontrol as it reliably contains myelinated and unmyelinated nervefibres and endothelial cells expressing PGP9.5, NF, NGF, NGFRp75,SP, CGRP, ACh, TH and CD34. The functional layer of eutopicendometrium from women without endometriosis was used as a nega-tive control as it does not contain any nerve fibres expressing PGP9.5,NF, NGF, NGFRp75, SP, CGRP, ACh and TH.

Statistical analysis

The images were captured using Olympus microscope BX51 and dig-ital camera DP70, and an assessment of nerve fibre density was per-formed by Image Pro Plus Discovery (MediaCybernetics, MD, USA).Once the images’ features were acquired, an orthogonal grid maskwas sketched above the original images. The sections of the grid were250 μm per side. Once the grid was in position, nerve fibres in theperitoneal endometriotic lesions, peritoneal endosalpingiosis lesionsand normal peritoneum, and the total number of squares covering thesections of peritoneal endometriotic lesions, peritoneal endosalpingio-sis lesions and normal peritoneum were counted. The total number ofnerve fibres was divided by the total number of squares covering theperitoneal endometriotic lesions, peritoneal endosalpingiosis lesionsor the normal peritoneum to obtain an average of nerve fibres persquare (each square of 250 × 250 μm). The results were expressed asthe mean (±SD) number of nerve fibres per mm2 in each specimenfrom all peritoneal endometriotic lesions, peritoneal endosalpingiosislesions and normal peritoneum samples, stained with the polyclonal



Dow

nloaded from



3003

antibody against PGP9.5 and the monoclonal antibody against NF.The counting procedure was carried out twice by two independentobservers (each blinded to the other) without any knowledge of theclinical parameters or other prognostic factors. The concordance ratewas >95% between the observers. Nerve fibre density between perito-neal endometriotic lesions from endometriosis patients, peritonealendosalpingiosis lesions from patients with endosalpingiosis alone orpatients with both endometriosis and endosalpingiosis and normalperitoneum from women without endometriosis was compared usingthe Kruskal–Wallis test and the Mann–Whitney test. Differences wereconsidered to be significant at P < 0.05.

Nerve fibres stained with SP, CGRP, TH, ACh, NGF andNGFRp75 were assessed semiquantitatively due to the partial cross-reactivity of some of these markers with some blood vesselsand connective tissues. Their staining pattern was compared with thatof PGP9.5 to obtain a semiquantitative representation of these nervefibres. Grading of density of these nerve fibres was represented asthe percentage of total nerve fibres demonstrated by PGP9.5 immuno-reactivity. + few (10–20%), ++ moderate (20–70%), +++ many (70–90%),++++ abundant (90–100%).

In one series of experiments, serial sections were sequentiallystained with all the markers listed above to assess the extent of co-localization of the markers around small nerve tracts within peritonealendometriotic lesions and normal peritoneum.

Results

There were more nerve fibres stained with PGP9.5 in theperitoneal endometriotic lesions (n = 40) from women withendometriosis (mean 16.3 ± 10.0/mm2; range 6.8–53.9/mm2)than in normal peritoneum (n = 36) from women withoutendometriosis (mean 2.5 ± 1.3/mm2; range 0.5–5.5/mm2;P < 0.001; Table I). These nerve fibres were more frequentlyseen near endometriotic glands and blood vessels than else-where in the stroma (Figure 1A).

Nerve fibres were present in greater density in peritonealendometriotic lesions (mean 16.3 ± 10.0/mm2; range 6.8–53.9/mm2) than in peritoneal endosalpingiosis lesions (n = 9)

Table I. A quantitative and semiquantitative assessment of the density of different types of nerve fibres in peritoneal endometriotic lesions

PGP9.5, protein gene product 9.5; NF, neurofilament; SP, substance P; CGRP, calcitonin gene-related peptide; TH, tyrosine hydroxylase; ACh, acetylcho-line; NGFRp75, nerve growth factor receptor p75; NGF, nerve growth factor.Grading of density of these nerve fibres was represented as the percentage of total nerve fibres demonstrated by PGP9.5 immunoreactivity. +, few (10–20%); ++, moderate (20–70%); +++, many (70–90%); ++++, abundant (90–100%).

Normal peritoneum

Peritoneal endometriotic lesion

Number of specimens 36 40Total nerve fibre density (mean ± SD/mm2) (stained with PGP9.5)

2.5 ± 1.3 16.3 ± 10.0

Total nerve fibre density (mean ± SD/mm2) stained with NF

1.0 ± 0.8 6.7 ± 3.7

Presence of SP immunoreactive nerve fibres

++ ++

Presence of CGRP immunoreactive nerve fibres

++ ++

Presence of TH immunoreactive nerve fibres

++ ++

Presence of ACh immunoreactive nerve fibres

+ +

Presence of NGFRp75immunoreactive nerve fibres

++++ ++++

Presence of NGF immunoreactivenerve fibres

++++ ++++

Figure 1. Nerve fibres in peritoneal endometriotic lesions and a peri-toneal endosalpingiosis lesion. (A) Peritoneal endometriotic lesionstained with protein gene product 9.5 (PGP9.5) and permanent fastred chromogen (magnification ×200). Arrows denote PGP9.5-positivenerve fibres near an endometriotic gland. (B) Peritoneal endosalping-iosis lesion stained with PGP9.5 and permanent fast red chromogen(magnification ×200). Arrows denote tiny PGP9.5-positive nervefibres. (C) Peritoneal endometriotic lesion stained with neurofilament(NF) and permanent fast red chromogen (magnification ×200).Arrows denote NF-positive myelinated nerve fibres near an endome-triotic gland. (D) Peritoneal endometriotic lesion stained with sub-stance P (SP) and 3,3′-diaminobenzidine (DAB) (magnification×400). Arrows denote tiny SP-positive nerve fibres near an endome-triotic gland. (E) Peritoneal endometriotic lesion stained with calci-tonin gene-related peptide (CGRP) and DAB (magnification ×200).Arrows denote CGRP-positive nerve fibres near an endometrioticgland. (F) Peritoneal endometriotic lesion stained with acetylcholine(ACh) and DAB (magnification ×200). Arrow denotes ACh-positivenerve fibres near an endometriotic gland. (G) Peritoneal endometri-otic lesion stained with tyrosine hydrolase (TH) and DAB (magnifi-cation ×400). Arrows denote TH-positive nerve fibres near anendometriotic gland. (H) Peritoneal endometriotic lesion stained withnerve growth factor (NGF) and DAB (magnification ×200). (I) Peri-toneal endometriotic lesion stained with nerve growth factor receptorp75 (NGFRp75) and permanent fast red chromogen (magnification×200). Black arrows denote NGFRp75-positive nerve fibres andnerve fibre trunks near an endometriotic gland (red arrow). Magnifi-cation ×400 was used due to small size of nerve fibres. Scale bars rep-resent 50 μm in D, G and H; and 100 μm in A–C, E, F and I.



Dow

nloaded from


N.Tokushige et al.

3004

(mean 3.8 ± 0.9/mm2; range 2.7–5.8/mm2; Figure 1B; P < 0.001).The density of nerve fibres in endosalpingiosis lesions was alsosignificantly greater than in normal peritoneum (P < 0.002).

There were also more NF immunoreactive nerve fibres inperitoneal endometriotic lesions (6.7 ± 3.7/mm2; range 2.0–20.0/mm2) than in normal peritoneum (1.0 ± 0.8/mm2; range0–2.7/mm2) and peritoneal endosalpingiosis (1.8 ± 0.4/mm2;range 1.3–2.8/mm2; P < 0.001; Figure 1C; Table I). However,only 4 out of 40 (10%) peritoneal endometriotic lesionsshowed NF immunoreactive nerve fibres near to endometrioticglands.

With confirmation by the specific markers for SP, CGRP,ACh and TH, nerve fibres in the peritoneal endometrioticlesions were demonstrated to be sensory Aδ, sensory C, cholin-ergic and adrenergic nerve fibres (Figure 1D–G, Table I).

There was an intense NGF immunoreactivity near endome-triotic glands in peritoneal endometriotic lesions (Figure 1H).More than 90% of the total PGP9.5 immunoreactive nervefibres were stained with NGF and NGFRp75 in peritonealendometriotic lesions (Figure 1I).

Serial sections stained with PGP9.5, NF, NGF, NGFRp75,TH, ACh, SP, CGRP and CD34 showed that several nervefibres in peritoneal endometriotic lesions showed co-localizationof SP, CGRP, TH, NGF and NGFRp75. Those nerve fibreswere a mixture of sensory C, sensory Aδ and adrenergic nervefibres (Figure 2A–H) and associated with blood vessels (Figure2I). Nerve fibres stained with ACh were not co-localized withSP, CGRP, TH, NGF and NGFRp75.

Discussion

This study has demonstrated multiple, small unmyelinated nervefibres in peritoneal endometriotic lesions in 40 women with con-firmed endometriosis. The density of nerve fibres in peritonealendometriotic lesions was much greater than in normal perito-neum in women with no endometriosis and much greater than inlesions of endosalpingiosis. It was somewhat surprising that thedensity of nerve fibres in endosalpingiosis lesions was no differ-ent from that in normal peritoneum, suggesting that increasednerve fibre density in endometriosis lesions and in the eutopicendometrium of endometriosis patients (Tokushige et al., 2006)is a specific finding for that condition rather than a generalresponse of peritoneum to development of a pathological lesion.Nerve fibres were more frequently seen near endometrioticglands in peritoneal endometriotic lesions. Only 4 out of 40(10%) peritoneal endometriotic lesions showed NF immunore-active (myelinated) nerve fibres near endometriotic glands, andNF immunoreactive nerve fibres demonstrated in the other 36peritoneal endometriotic lesions were distant from endometrioticglands. Therefore, most of the nerve fibres demonstrated in closeproximity to endometriotic glands were unmyelinated. Thosenerve fibres stained with NF in four peritoneal endometrioticlesions may be Aδ fibres that transmit sharp, pricking localizedpain to the central nervous system.

Confirmation by the specific markers for sensory, cholinergicand adrenergic nerve fibres demonstrated that single trunks andnerve fibre trunks in peritoneal endometriotic lesions werefound to contain sensory Aδ, sensory C, cholinergic and adrenergicnerve fibres. Moderate numbers of sensory C and adrenergicnerve fibres were present near endometriotic glands. The othermarkers for sensory Aδ and cholinergic nerve fibres were notobviously located in close relation to glands. Several singletrunks and nerve fibre trunks in peritoneal endometrioticlesions showed co-localization of SP, CGRP, TH and NF, andthose nerve fibre trunks were a mixture of sensory C, sensoryAδ and adrenergic nerve fibres. Nerve fibres stained with AChwere not co-localized with SP, CGRP and TH.

Tulandi et al. (2001) demonstrated NF immunoreactivenerve fibres in peritoneal endometriotic lesions and reported

Figure 1. Continued



Dow

nloaded from



3005

that the distance between endometriotic glands and nerve fibresin endometriotic lesions from women with pain was closer thanin women with no pain. We also demonstrated SP, CGRP, TH,NF and ACh immunoreactive nerve fibres close to endometri-otic glands in women with pain associated with endometriosis.

It is not yet clear what neural stimuli may initiate the severepain sensations that some women experience with endometriosis.Endometriosis is an inflammatory disease and many inflammatory

Figure 2. Different types of nerve fibres in a peritoneal endometri-otic lesion (serial sections). (A) Peritoneal endometriotic lesionshowing a large nerve trunk stained with protein gene product 9.5(PGP9.5) (magnification ×400). Nerve fibres were stained red withpermanent fast red chromogen. (B) Peritoneal endometriotic lesionfrom the same patient stained with neurofilament (NF) (magnifica-tion ×400) as nerve fibres were stained red with permanent fast redchromogen. (C) Peritoneal endometriotic lesion from the samepatient stained with nerve growth factor (NGF) and 3,3′-diami-nobenzidine (DAB) (magnification ×400). Nerve fibres and somesurrounding tissues were weakly stained brown. (D) Peritonealendometriotic lesion from the same patient stained with nervegrowth factor receptor p75 (NGFRp75) and permanent fast redchromogen (magnification ×400). Nerve fibres and some closelylocalized and related surrounding tissues were stained red. (E) Peri-toneal endometriotic lesion from the same patient stained with tyro-sine hydrolase (TH) and permanent fast red chromogen(magnification ×400). Nerve fibres were weakly stained red. (F)Peritoneal endometriotic lesion from the same patient stained withacetylcholine (ACh) and DAB (magnification ×400). No nervefibres were stained. (G) Peritoneal endometriotic lesion from thesame patient stained with substance P (SP) and DAB (magnification×400). Nerve fibres were weakly stained brown. (H) Peritonealendometriotic lesion from the same patient stained with calcitoningene-related peptide (CGRP) and DAB (magnification ×400).Nerve fibres were weakly stained brown. (I) Peritoneal endometri-otic lesion from the same patient double-stained with PGP9.5 andCD34 (magnification ×400). The nerve trunk was intensely stainedred (black arrow) and blood vessels (blue and red arrows) werestained brown. Red arrow denotes high density of capillaries andblue arrow denotes thin walled blood vessel. Scale bars represent 50 μm.

Figure 2. Continued



Dow

nloaded from


N.Tokushige et al.

3006

mediators are released from damaged cells and tissues in thevicinity of endometriotic lesions. These substances include his-tamine, serotonin (5-HT), bradykinin, prostaglandins (PGs),leukotrienes, interleukin (IL)-1β, IL-6, IL-8, ACh, vascularendothelial growth factor (VEGF), tumour necrosis factor-α(TNF-α), endothelins, epidermal growth factor (EGF), TGF-β,platelet-derived growth factor (PDGF) and NGF (Julius et al.,2001; Mantyh, 2002; Richardson and Vasko, 2002). Moleculessecreted directly from endometriotic lesions include PGs(Ylikorkala and Viinikka, 1983; De Leon et al., 1988), TNF-α(Bergqvist et al., 2001) and NGF (Anaf et al., 2002). These sub-stances can all sensitize and/or activate sensory nerve endingsby interacting with cell-surface receptors (Mantyh, 2002; Schaibleet al., 2002). It is clear that many peritoneal endometrioticlesions are locally tender and this may be a manifestation of aninflammatory reaction. Palter et al. (1996) first described activeendometriotic lesions as a source of pain using ‘conscious painmapping’. Almeida and Val-Gallas (1997) found 48 positivefindings in 50 women (96%) with conscious pain mapping, andDemco (1998) also found that most women localized or mappedtheir pain to their endometriotic lesions.

Tamburro et al. (2003) demonstrated TGF-α1 immunoreac-tive nerve fibres in peritoneal endometriotic lesions fromwomen with dysmenorrhoea. Because TGF-α1 can increasecyclooxygenase-2 activity (Fong et al., 2000), TGF-α1 mayincrease PG production in sensory and sympathetic nervefibres in women with endometriosis, constituting one furtherpossible mechanism of pain generation.

When sensory nerve fibres are stimulated by inflammatorysubstances listed above, neurotransmitters such as SP, CGRP,endothelin, histamine, glutamate, PGs and vaso-intestinalpeptide can be secreted from sensory nerve endings (Mantyh,2002; Liddle and Nathan, 2004). SP and CGRP can contributeto the inflammatory response by causing vasodilation, plasmaextravasation (leakage of proteins and fluid from postcapillaryvenules) and cellular infiltration by interacting with endothelialcells, arterioles, mast cells, neutrophils and immune cells(Maggi, 1995). SP can also act on mast cells in the vicinity ofsensory nerve endings to evoke degranulation and the releaseof TNF-α, histamine, PGD2, IL-8 and leukotriene B4, as wellas acting on platelets to release 5-HT, providing a positivefeedback (Steinhoff et al., 2003). CGRP has a wide range ofbiological activities, including sensory transmission (Gibsonet al., 1984), regulation of glandular secretion (Onuoha andAlpar, 2001), proliferative effect on human endothelial cells(Onuoha and Alpar, 2001) and inhibiting SP degradation by aspecific endopeptidase (Le Greves et al., 1985) to enhance SPrelease, thereby amplifying the effects (Steinhoff et al., 2003).

C nerve fibres can develop a greater sensitivity to nore-pinephrine (NE) during inflammation (Banik et al., 2001).Sympathetic postganglionic neurons (adrenergic nerve termi-nals) can synthesize PGE2 and PGI2 (Gonzales et al., 1989).Taiwo et al. (1990) reported that the hyperalgesia induced byNE is mediated by PGs produced by sympathetic postgangli-onic neuron terminals, as NE is considered to stimulate theproduction of PGs through α2-adrenoceptors (Levine et al.,1986). ACh can be released from sensory nerve fibres (Tata et al.,1994), damaged endothelial cells, platelets and parasympathetic

postganglionic neurons. ACh can produce burning pain whenapplied to human skin (Rukwied and Heyer, 1999) by activat-ing sensory C nerve fibres (Steen and Reeh, 1993) throughnicotinic receptors (Jinks and Carstens, 1999; Bernardiniet al., 2001).

During inflammation, NGF is markedly up-regulated innerve fibres associated with the inflamed area (Steinhoff et al.,2003) and estrogen also increases endometrial stromal NGFprotein expression by 43% (Krizsan-Agbas et al., 2003). NGFcan increase the synthesis of SP and CGRP and also directlystimulate degranulation of mast cells (Schaible et al., 2005).NGF can also sensitize (lower the threshold) or excite the ter-minals of sensory nerve fibres by interacting with cell-surfacereceptors expressed by these neurons (Julius et al., 2001).Hence, multiple potential molecular mechanisms exist withinthe cells of endometriotic lesions to cause pain sensation bysensitizing and/or activating sensory nerve fibres.

It will be important to see how future studies reveal varia-tions in these potential mechanisms in endometriosis womenwho present with pain compared with those who have no com-plaint of pain, or those who present solely with infertility.

The greatly increased expression of NGF and NGFRp75 byendometriosis lesions may also play a role in inducingingrowth of nerve fibres into endometriotic tissue in the firstplace. This expression may play a primary role in setting up themechanisms for generation of pain from endometriotic lesions.

In summary, human peritoneal endometriotic lesions wereinnervated by sensory, cholinergic and adrenergic nerve fibres,and sensory and adrenergic nerve fibres were often co-localized.These nerve fibres may interact with each other and with manyactive molecules released by endometriotic lesions to causepain and local tenderness. These mechanisms require muchmore detailed study.

AcknowledgementsThis study was supported by research funding from the Department ofObstetrics and Gynaecology, University of Sydney. The authors thankLawrence Young (Dako, Australia) for technical support.

References

Almeida OD Jr and Val-Gallas JM (1997) Conscious pain mapping. J AmAssoc Gynecol Laparosc 5,587–590.

American Society for Reproductive Medicine (1996) Revised American Soci-ety for Reproductive Medicine classification of endometriosis. Fertil Steril67,817–821.

Anaf V, Simon P, Nakadi IEI, Fayt I, Simonart T, Buxant F and Noel JC(2002) Hyperalgesia, nerve infiltration and nerve growth factor expressionin deep adenomyotic nodules, peritoneal and ovarian endometriosis. HumReprod 17,1895–1900.

Banik RK, Sato J, Yajima H and Mizumura K (2001) Differences betweenthe Lewis and Sprague-Dawley rats in chronic inflammation inducednorepinephrine sensitivity of cutaneous C-fiber nociceptors. NeurosciLett 299,21–24.

Bergqvist A, Bruse C, Carlberg M and Carlstrom K (2001) Interleukin 1beta,interleukin-6, and tumour necrosis factor-alpha in endometriotic tissue andin endometrium. Fertil Steril 75,489–495.

Berkley KJ, Dmitrieva N, Curtis KS and Papka RE (2004) Innervation ofectopic endometrium in a rat model of endometriosis. Proc Natl Acad SciUSA 101,11094–11098.

Bernardini N, Sauer SK, Haberberger R, Fischer MJ and Reeh PW (2001)Excitatory nicotinic and desensitizing muscarinic (M2) effects on C-nocice-ptors in isolated rat skin. J Neurosci 21,3295–3302.



Dow

nloaded from



3007

Calder JS, Holten I and McAllister RMR (1998) Evidence for immune systeminvolvement in reflex sympathetic dystrophy. J Hand Surg 23,147–150.

De Leon FD, Vijayakumar R, Rao CV and Yussman M (1988) ProstaglandinF2 alpha and E2 release by peritoneum with and without endometriosis. Int JFertil 33,48–51.

Demco L (1998) Mapping the source and character of pain due to endome-triosis by patient-assisted laparoscopy. J Am Assoc Gynecol Laparosc5,241–245.

Fong CY, Pang L, Holland E and Knox AJ (2000) TGF-α1 stimulates IL-8release, COX-2 expression, and PGE2 release in human airway smooth mus-cle cells. Am J Physiol Lung Cell Mol Physiol 279,201–207.

Gibson SJ, Polak JM, Bloom SR, Sabate IM, Mulderry PM, Ghatei MA,McGregor GP, Morrison JF, Kelly JS and Evans RM (1984) Calcitoningene-related peptide immunoreactivity in the spinal cord of man and of eightother species. J Neurosci 4,3101–3111.

Giudice LC, Tazuke SI and Swiersz L (1998) Status of current research onendometriosis. J Reprod Med 43,252–262.

Gonzales R, Goldyne ME, Taiwo YO and Levine JD (1989) Production ofhyperalgesic prostaglandins by sympathetic postganglionic neurons. J Neu-rochem 531,595–598.

Janossy A, Orso E, Szalay KS, Juranyi Z, Beck M, Vizi ES and Vinson GP(1998) Cholinergic regulation of the rat adrenal zona glomerulosa. J Endo-crinol 157,305–315.

Jinks SL and Carstens E (1999) Activation of spinal wide dynamic rangeneurons by intracutaneous microinjection of nicotine. J Neurophysiol82,3046–3055.

Krizsan-Agbas D, Pedchenko T, Hasan W, Smith PG (2003) Oestrogen regu-lates sympathetic neurite outgrowth by modulating brain derived neuro-trophic factor synthesis and release by the rodent uterus. Eur J Neurosci18,2760–2768.

Lebovic DI, Mueller MD and Taylor RN (2001) Immunobiology of endometri-osis. Fertil Steril 75,1–10.

Le Greves P, Nyberg F, Terenius L and Hokfelt T (1985) Calcitonin gene-related peptide is a potent inhibitor of substance P degradation. Eur J Phar-macol 115,309–311.

Levine JD, Taiwo YO, Collins SD and Tam JK (1986) Noradrenaline hyperal-gesia is mediated through interaction with sympathetic postganglionicneurone terminals rather than activation of primary afferent nociceptors.Nature 323,158–160.

Liddle RA and Nathan JD (2004) Neurogenic inflammation and pancreatitis.Pancreatology 4,551–559.

Lundberg LM, Alm P, Wharton J and Polak JM (1988) Protein gene product9.5 (PGP9.5): a new neuronal marker visualizing the whole uterine innerva-tion and pregnancy-induced and developmental changes in the guinea pig.Histochemistry 90,9–17.

Maggi CA (1995) Tachykinins and calcitonin gene-related peptide (CGRP) asco-transmitters released from peripheral endings of sensory nerves. ProgNeurobiol 45,1–98.

Mantyh PW (2002) Neurobiology of substance P and the NK1 receptor. J ClinPsychiatry 63,6–10.

Onuoha GN and Alpar EK (2001) Levels of vasodilators (SP, CGRP) and vaso-constrictor (NPY) peptides in early human burns. Eur J Clin Invest 31,253–257.

Palter SF, Duleba A, Zreik T, DiStasio C, Mandel T, Guarnaccia M, Polcz T,Amesse L and Olive DL (1996) Office microlaparoscopy under localanesthesia. J Am Assoc Gynecol Laparosc 3,S38.

Richardson JD and Vasko MR (2002) Cellular mechanisms of neurogenicinflammation. J Pharmacol Exp Ther 302,839–845.

Rukwied R and Heyer G (1999) Administration of acetylcholine and vasoac-tive intestinal polypeptide to atopic eczema patients. Exp Dermatol 8,39–45.

Schaible HG, Ebersberger A and Von Banchet GS (2002) Mechanisms of painin arthritis. Ann N Y Acad Sci 966,343–354.

Schaible HG, Del Rosso A and Matucci-Cerinic M (2005) Neurogenic aspectsof inflammation. Rheum Dis Clin North Am 31,77–101.

Schlaepfer WW (1987) Neurofilaments: structure, metabolism and implica-tions in disease. J Neuropathol Exp Neurol 46,117–129.

Steen KH and Reeh PW (1993) Actions of cholinergic agonists and antagonistson sensory nerve endings in rat skin, in vitro. J Neurophysiol 70,397–405.

Steinhoff M, Stander S, Seeliger S, Ansel JC, Schmelz M and Luger T (2003)Modern aspects of cutaneous neurogenic inflammation. Arch Dermatol139,1479–1488.

Taiwo YO, Heller PH and Levine JD (1990) Characterization of distinct phos-pholipases mediating bradykinin and noradrenaline hyperalgesia. Neuro-science 39,523–531.

Tamburro S, Canis M, Albuisson E, Dechelotte P, Darcha C and Mage G(2003) Expression of transforming growth factor beta1 in nerve fibers isrelated to dysmenorrhea and laparoscopic appearance of endometrioticimplants. Fertil Steril 80,1131–1136.

Tata AM, Plateroti M, Cibati M, Biagioni S and Augusti-Tocco G (1994)Cholinergic markers are expressed in developing and mature neurons ofchick dorsal root ganglia. J Neurosci Res 37,247–255.

Tokushige N, Markham R, Russell P and Fraser IS (2006) High density ofsmall nerve fibres in the functional layer of the endometrium in women withendometriosis. Hum Reprod 21,782–787.

Tulandi T, Felemban A and Chen MF (2001) Nerve fibers and histopathology ofendometriosis-harboring peritoneum. J Am Assoc Gynecol Laparosc 8,95–98.

Ylikorkala O and Viinikka L (1983) Prostaglandins and endometriosis. ActaObstet Gynecol Scand Suppl 113,105–107.

Submitted on March 22, 2006; resubmitted on May 26, 2006; accepted on June 5,2006



Dow

nloaded from


nerve fibres in peritoneal endometriosis

Documents