BIOLOGY OF REPRODUCTION (2013) 88(5):122, 1–8Published online before print 10 April 2013.DOI 10.1095/biolreprod.112.103598

Caldesmon: New Insights for Diagnosing Endometriosis1

Juliana Meola,2,3 Gabriela dos Santos Hidalgo,3 Julio Cesar Rosa e Silva,3 Lilian Eslaine Costa MendesSilva,4 Claudia Cristina Paro Paz,5 and Rui Alberto Ferriani3

3Department of Gynecology and Obstetrics, School of Medicine of Ribeirao Preto, University of Sao Paulo, RibeiraoPreto, Brazil4Department of Physiology, School of Medicine of Ribeirao Preto, University of Sao Paulo, Ribeirao Preto, Brazil5Department of Genetics, School of Medicine of Ribeirao Preto, University of Sao Paulo, Ribeirao Preto, Brazil

ABSTRACT

Considerable effort has been invested in searching for lessinvasive methods of diagnosing endometriosis. Previousstudies have indicated altered levels of the CALD1 gene(encoding the protein caldesmon) in endometriosis. The aimsof our study were to investigate whether average CALD1expression and caldesmon protein levels are differentiallyaltered in the endometrium and endometriotic lesions and toevaluate the performance of the CALD1 gene and caldesmonprotein as potential biomarkers for endometriosis. Pairedbiopsies of endometrial tissue (eutopic endometrium) andendometriotic lesions (ectopic endometrium) were obtainedfrom patients with endometriosis to evaluate CALD1 geneexpression and caldesmon protein levels by real-time PCR andWestern blot analysis, respectively. In addition, immunostain-ing for caldesmon to determine cellular localization was alsoperformed. Endometrium from women without endometriosiswas used as a control. Increased CALD1 expression andcaldesmon levels were detected in the endometriotic lesions.The electrophoretic profile of caldesmon by Western blotanalysis was clearly different between the control group(endometrium of women without endometriosis) and thegroup of women with endometriosis (eutopic endometriumand endometriotic lesions). Caldesmon expression as deter-mined by immunostaining showed no variation among the celltypes in endometriotic lesions and eutopic endometrium.Stromal cells marked positively in eutopic endometrium fromcontrol patients and in the endometriotic lesions. Thepresence of caldesmon in the endometrium of patients withand without endometriosis permitted diagnoses with 95%sensitivity (specificity 100%) and 100% sensitivity (specificity100%) for the disease and for minimal to mild endometriosisin the proliferative phase of the menstrual cycle, respectively.In the secretory phase, minimal to mild endometriosis wasdetected with 90% sensitivity and 93.3% specificity. Caldes-mon is a possible predictor of endometrial dysregulation inpatients with endometriosis. A potential limitation of ourstudy is the fact that other endometrial diseases were notexcluded, and therefore prospective studies are needed to

confirm the potential of caldesmon as a biomarker exclusivelyfor endometriosis.

biomarker, caldesmon, diagnosis, endometriosis, Western blotanalysis

INTRODUCTION

Endometriosis is a common benign gynecological diseasethat affects at least 10% of women of reproductive age [1]. Thiscondition is characterized by the growth of glands andendometrial stroma outside the uterine cavity, which areknown as ectopic tissues [2]. These lesions may be associatedwith infertility and pain symptoms, including cyclic pelvicpain, dysmenorrhea, dyspareunia, dysuria, and dyschezia.There is a low correlation between the type of lesion and thesymptoms of the disease [3].

The gold standard for diagnosing endometriosis is directinspection of the abdominal cavity by laparoscopy [4] plushistological confirmation of ectopic endometrial tissue [5].However, this procedure is costly; moreover, the diagnosticaccuracy of the method is affected by the level of experience ofthe surgeon [6]. This heterogeneity in clinical presentation andthe need for an invasive procedure for diagnostic confirmationjustify the search for a sensitive marker of low invasiveness [7].

Considerable effort has been invested in searching fornoninvasive methods for diagnosing endometriosis and inunderstanding the physiopathology of the disease [8]. Manyreports have suggested that various serum, peripheral blood,peritoneal fluid, urine, endometrial fluid, and biopsy markersare associated with endometriosis, such as tumor markers andpolypeptides (CA-125, CA 19-9, SICAM-1), immunologicalmarkers (IL-6, TNF, antiendometrial antibodies and autoanti-bodies, and markers of oxidative stress), genetic markers(EGR-1, P450 aromatase, and PP14), circulating cell-freeDNA, and tissue markers (P450 aromatase, cytokeratins,hormone receptors, PGP9.5-immunoactive nerve fibers) [7,9–19]. The most commonly used serum marker for the clinicaldiagnosis of endometriosis, CA-125, is inadequate for theroutine clinical assessment of endometriosis [20] because it haslimited diagnostic accuracy in identifying early-stage endome-triosis [21].

The expression of the CALD1 gene determined by screeninghas been reported to be increased in patients with endometri-osis [22–25]. The human CALD1 gene encodes caldesmon, aprotein found in the cytoskeleton [26]. Cytoskeletal proteinshave been shown to play an important role in many cellularfunctions, such as proliferation, apoptosis, cell motility,adhesion, receptor function, and second messenger pathways.Several studies have shown that genes and proteins of thecytoskeleton, such as transgelin, moesin, and caldesmon,among others, are related to endometriosis [10, 12, 19, 22,24, 27]. The objectives of the present study were to investigate

1Supported by grants from Fundacao de Amparo a Pesquisa do Estadode Sao Paulo (FAPESP) and Instituto Nacional de Ciencia e Tecnologia(INCT).2Correspondence: Juliana Meola, School of Medicine of Ribeirao Preto,University of Sao Paulo, Av Bandeirante 3900, HC/FMRP, laboratoriode GO, Ribeirao Preto, Sao Paulo 14049-900, Brazil.E-mail: jumeola@yahoo.com.br

Received: 25 July 2012.First decision: 14 September 2012.Accepted: 1 April 2013.� 2013 by the Society for the Study of Reproduction, Inc.eISSN: 1529-7268 http://www.biolreprod.orgISSN: 0006-3363

1 Article 122

Dow

nloaded from w

ww

.biolreprod.org.

whether average CALD1 expression and caldesmon proteinlevels are altered in the endometrium and in endometrioticlesions and to evaluate the performance of the CALD1 geneand caldesmon protein as potential biomarkers for endometri-osis.

MATERIALS AND METHODS

This was not a blind study. The study was approved by the Research EthicsCommittees of the Faculdade de Medicina de Ribeirao Preto. The participantswere recruited at the tertiary hospital of the School of Medicine of RibeiraoPreto (University of Sao Paulo), and all of the participants gave informedwritten consent.

Patients and Samples

A case-control study was conducted on women with and withoutendometriosis in the proliferative and secretory phases of the menstrual cycle.Forty patients with a laparoscopic and histological diagnosis of endometriosiswere selected. The subjects were 18 to 40 years old (mean, 33.5 6 0.72 yr),were not menopausal, had not been taking any hormone therapy for at least 6mo before sample collection, and had no other reproductive disorders or anytumors. The indication for laparoscopy was infertility in 21 patients and pelvicpain in 19. The stage of endometriosis was determined according to theclassification of the American Society for Reproductive Medicine [28]. Pairedsamples of eutopic endometrium and endometriotic lesions were obtained from40 women. Only one lesion was collected per patient. Twenty peritoneal lesionswere studied in patients staged as grade I (n¼ 6, secretory phase), grade II (n¼11, 7 proliferative and 4 secretory phase), grade III (n¼ 2, proliferative phase),and grade IV (n ¼ 1, proliferative phase), and 20 lesions of ovarianendometrioma were studied in patients staged as grade III (n¼8, 4 proliferativeand 4 secretory phase) and grade IV (n ¼ 12, 6 proliferative and 6 secretoryphase). Eutopic endometrium (n ¼ 40) was collected with a Novak curetteduring laparoscopy.

The control group consisted of 15 women of reproductive age withoutendometriosis, fibrosis, pelvic adhesions, or infertility, aged 18 to 40 years old(mean, 33.87 6 0.69 yr). These women were subjected to laparoscopy for tubeligation, with laparoscopic confirmation of the absence of endometriotic lesionsand had not been taking hormone medication for at least 6 mo before biopsycollection. Two paired endometrial biopsies (control endometrium) werecollected from these patients according to the phase of the menstrual cycle.Fifteen biopsies were obtained during the proliferative phase and 15 during thesecretory phase of the menstrual cycle, as confirmed by histological dating. Thebiopsies were collected with a Novak curette, the first during the surgicalprocedure (during the proliferative phase of the menstrual cycle) and the secondscheduled and obtained during the secretory phase of the same menstrual cycle.

The samples were stored in a freezer at �808C in a cryotube containingRNAlater Solution (Ambion Life Technologies) for RNA and proteinpreservation. In addition, part of the fresh tissues were fixed in neutral 4%formalin and embedded in paraffin for tissue structure preservation.

Gene Expression of CALD1

The samples were washed in 13 PBS (8.50 g/L NaCl, 1.11 g/L Na2HPO

4,

2.81 g/L Na2HPO

4

.12H2O, 0.20 g/L KH

2PO

4, pH 7.0) to remove the RNAlater

Solution. Next, total RNA (50 mg tissue) was extracted with TRIzol reagent(Invitrogen Life Technologies) and treated with DNase I (Invitrogen LifeTechnologies) according to the manufacturer’s instructions. The interphase andthe phenol/chloroform organic phase were stored in a freezer at�808C for theisolation of total protein from the same samples. The RNA integrity wasconfirmed by the presence of 28S and 18S ribosome bands when analyzed by1% agarose gel electrophoresis with 13 3-(N-morpholino) propane sulfonicacid buffer. Total RNA concentration was determined by spectrophotometry(NanoDrop 2000c; Thermo Scientific) at 260 nm. The extracted total RNA wasstored in a freezer at�808C for subsequent use. One microgram of total RNAfrom each sample was reverse transcribed using the High Capacity cDNATranscription Kit (Applied Biosystems Life Technologies) according to themanufacturer’s instructions.

Relative quantification of the CALD1 gene in the 110 tissue samples thatwere collected were analyzed using an ABI PRISM 7500 FAST instrument(Applied Biosystems Life Technologies). The reactions were conducted usingthe TaqMan Gene Expression Assays (TaqMan minor groove binder probes, 6-carboxyfluorescein dye labeled) by Applied Biosystems Life Technologies.The assays were probed for CALD1 (Hs00189021_m1), GAPDH(Hs99999905_m1), and ACTB (Hs99999903_m1). Quantitative real-timePCR was performed in triplicate in a final volume of 20 ll using the following

conditions: 10 ll TaqMan Universal PCR Master Mix (23) (AppliedBiosystems Life Technologies), 1 ll TaqMan Gene Expression Assay (203)(Applied Biosystems Life Technologies), and 9 ll of cDNA diluted 1:50. Thereaction conditions were as follows: 508C for 2 min, 958C for 10 min, and 40cycles of 958C for 15 sec and 608C for 1 min.

The relative quantification of the CALD1 gene was calculated for eachsample according to the 2-delta delta Ct (2�DDCT) method. A pool of cDNAcontaining equal quantities of endometrium samples obtained from womenwithout endometriosis (control group) was used to calibrate the reactions. TheGAPDH and ACTB genes were used as reference genes to normalize thereactions.

Western Blot Analysis

Total protein was extracted from the interphase and from the phenol/chloroform organic phase obtained after total RNA isolation according to theTRIzol reagent protocol (Invitrogen Life Technologies). The pellet was dilutedwith 1% SDS, and protease and phosphatase inhibitors were added (Sigma-Aldrich). The protein concentration was determined in duplicate by thebicinchoninic assay method (Pierce). The proteins were denatured by boiling in13 Laemmli sample buffer. Next, 10 lg of total protein per well was separatedon a 1-mm thick 10% SDS-PAGE polyacrylamide gel (25 V for 16 h) andelectroblotted onto (0.45-lm pore size) nitrocellulose membranes (Hybond; GEHealthcare) using the Mini Trans-Blot Electrophoretic Transfer Cell tankmethod (Bio-Rad) (350 mA, 90 min, cold). The efficiency of the transfer to thenitrocellulose membranes was verified by a brief staining of the blots withPonceau red stain. These membranes were then blocked with Tris-bufferedsaline (TBS) containing 5% bovine serum albumin and 0.1% Tween-20 for 1 hat 48C. Next, the membranes were incubated for 16 h at 48C with an anti-caldesmon antibody [12B5] (mouse monoclonal ab8247, 0.5 lg/ml dilution;Abcam) and monoclonal anti-b-actin (clone AC-15, A5441, 0.1 lg/ml dilution;Sigma-Aldrich). After washing with TBS containing 0.1% Tween-20, all of themembranes were incubated for 1 h at 48C with a rabbit polyclonal secondaryantibody to mouse immunoglobulin G conjugated to horseradish peroxidase(ab6728, 1 lg/ml dilution; Abcam). The specific immunoreactive bands weredetected by chemiluminescence (ECL Plus Western Blotting Detection System;GE Healthcare). Images were acquired with the Molecular Image ChemidocXRS Imaging System (Bio-Rad) and quantified using Image Lab 3.0 (Bio-Rad). The caldesmon levels were normalized to a b-actin reference control. Toavoid bias in the quantification, comparisons were made before and afterstripping on the membrane. Thus, the membrane was cut below the 55-kDaposition of the molecular marker to hybridize the primary antibodies separately.

Immunohistochemistry

Fresh tissue had been routinely fixed in 4% neutral formalin and embeddedin paraffin. Immunohistochemical staining was developed using the NovolinkMax Polymer Detection System (Novocastra). Briefly, 3-lm thick sectionswere cut from paraffin blocks containing representative samples. Paraffinsections were dewaxed in xylene, rehydrated through a graded alcohol series,placed in 10 mM citrate buffer, and submitted to heat retrieval using a vaporlock for 40 min. After heating, the slides were allowed to cool to roomtemperature and briefly washed with TBS. Endogenous peroxidase activity wasblocked with 0.3% hydrogen peroxide in methanol for 30 min, followed by asingle wash in ethylenediaminetetraacetic acid, pH 8.5. Normal serum(Novostain super-ABC kit; Novocastra) was used for 30 min in order to blockany nonspecific immunoassay. The primary antibody (caldesmon E 89, 1:40dilution; Cell Marque) was incubated overnight at 48C. The slides were thenincubated with the postprimary solution for 30 min and incubated with thepolymer for 30 min (both provided by Novolink Max Polymer DetectionSystem). The reaction was stained with diaminobenzidine followed byhematoxylin counterstaining. The slides were then dehydrated in an ethanolseries and mounted with Permount (Fischer). Myoma cases were used as thepositive control for caldesmon. Negative control for immunostaining wasprepared by omission of the primary antibody.

Statistical Analysis

The data are represented in figures on the original scale (box-and-whiskerand dots). However, the variables of gene expression and protein levels werelog

10transformed. The logarithmic transformation was necessary because one

of the assumptions (i.e., linearity) in the linear model analysis was not satisfied.These analyses specify that the conditional average E (yx ¼ x

0) from the

response variable y (CALD1 measures), given the x0

value of the predictorvector x (independent variable; endometriosis), is linear in x

0.The application

of the linear models can be extended assuming that an appropriate

MEOLA ET AL.

2 Article 122

Dow

nloaded from w

ww

.biolreprod.org.

transformation of the answer variable is given by t (y), where E (t [y] x) is linearin x and t (y)¼ b

0(overall average CALD1) þ bTx þ e, for unknown b

0e bT

(the nonlinear coefficient that estimates the effects of the endometriosis inCALD1 measures). The term e (random error) is independent of x and has anaverage of zero [29]. The analyses were performed using the SAS 2003software (SAS Institute Inc.). The analysis of variance that was performed inthe statistical model included the effects of phases of menstrual cycle(proliferative and secretory), the type of tissue (ovary and peritoneum lesionsand eutopic and control endometrium), as well as the interactions among them.A t-test for paired samples (eutopic endometrium versus ectopic lesions) andWelch ANOVA for unpaired samples (control endometrium versus eutopicendometrium of patients with endometriosis; control endometrium versuslesions of patients with endometriosis; peritoneal lesions versus ovarianlesions) were used to compare the means of the data obtained for the differenttissues. The study samples were sufficient for CALD1 analyses with a power (1� b) of at least 0.80 and a level of significance (a)¼ 0.05 and a power (1� b)of 0.99 for protein comparisons between control endometrium and eutopicendometrium of women with endometriosis.

The diagnostic performance of CALD1 expression and caldesmon proteinlevels were assessed using receiver operating characteristic (ROC) curves(MedCalc statistical software) [30] to plot the test sensitivity versus the false-positive rate and to determine the usefulness of a diagnostic test over a range ofpossible clinical results; 95% confidence intervals were obtained for the variousROC curves. The data used in this analysis were obtained from theendometrium of women with and without endometriosis during both theproliferative and the secretory phases of the menstrual cycle.

RESULTS

Comparisons between peritoneal lesions and endometriomaswere made both regarding expression of the CALD1 gene andits protein levels. No significant difference was observed inthese analyses. The remaining analyses were carried out bygrouping the peritoneal and ovarian lesions (see Figs. 1 and 3).

CALD1 Expression

Increased CALD1 expression was detected in endometrioticlesions compared with the eutopic endometrium of womenwith endometriosis and control women in both the proliferative(Fig. 1A) and secretory (Fig. 1B) phases of the menstrual cycle.The phase of the menstrual cycle did not change CALD1expression in any of the tissues analyzed (data not shown).There was a difference in expression between the eutopicendometrium of women with endometriosis and of controls inthe secretory phase (Fig. 1B) of the menstrual cycle.

Caldesmon Protein Levels

The SDS-PAGE profile of the caldesmon protein wasclearly different among the eutopic endometrium of bothgroups and the endometriotic lesion tissue. Two bands withstrong intensity were visualized, one of approximately 77 kDa(low molecular weight caldesmon, i.e., l-CaD) and the other of120 kDa (high molecular weight caldesmon, i.e., h-CaD). Thetwo protein isoforms were strongly detected in the endometri-otic lesions but were only weakly detected in the eutopicendometria of the women with endometriosis. In contrast, the l-CaD isoform predominated in control endometrium (Fig. 2).

The phase of the menstrual cycle did not influence theprotein levels in any of the tissues analyzed (data not shown).In addition, the caldesmon levels detected in the differenttissues showed the same pattern of protein expression in theproliferative and secretory phases of the menstrual cycle. Thechanges detected are presented in Figure 3, A–F.

The data for the h-CaD (120 kDa) and l-CaD (77 kDa)isoforms were analyzed separately and then as a whole(caldesmon protein ¼ h-CaD þ l-CaD). A significant increasein caldesmon protein was observed in the endometriotic lesionscompared with the eutopic endometrium of women withendometriosis (Fig. 3, A and B). However, a significantly

lower amount of protein was detected in the endometrium ofwomen with endometriosis compared with control endometri-um. In addition, there was no significant increase in caldesmonprotein in the lesions compared with the controls.

A separate analysis of the caldesmon isoforms also revealedtheir increased levels in the lesions compared with the pairedendometrium (Fig. 3, C–F), although the difference betweenthe endometria of the patients with endometriosis and thecontrols was significant only for l-CaD (Fig. 3, E and F). For h-CaD, the protein levels were increased in the lesions comparedwith the control endometrium and the eutopic endometrium inthe same patients (Fig. 3, C and D).

Immunolocalization of Caldesmon

The caldesmon immunolocalization did not identify differ-ences between the marked cell types in the endometrioticlesions and in the eutopic endometrium. Stromal cells markedpositively in the eutopic endometrium from control patientsand in the endometriotic lesions (peritoneal and endometrioma)(Fig. 4, A, C, and D). However, no marked cells were observedin the eutopic endometrium of patients with endometriosis (Fig.4B). In addition, no glands were marked in any of the analyzedtissues.

ROC Curves

The diagnostic performance of CALD1 transcripts and ofcaldesmon protein (h-CaD þ l-CaD) for endometriosisaccording to disease stage is shown in Table 1. The geneexpression of CALD1 has limited value as a potentialdiagnostic biomarker for endometriosis. However, the protein

FIG. 1. Relative quantification data for the CALD1 gene by the 2�DDCt

method in the analysis of different tissues in the proliferative (A) andsecretory (B) phases of the menstrual cycle. Control, endometrium ofwomen without endometriosis; Eutopic, endometrium of women withendometriosis; Lesion, endometriotic lesions (peritoneal and ovarianendometrioma). Comparisons showing significant differences are indicat-ed by the P value in parentheses. The sample number for each groupstudied is also given. Comparisons between control and eutopic tissueswere not paired analyses; comparisons between eutopic tissue and lesionswere paired analyses.

CALDESMON IS A POTENTIAL MARKER FOR ENDOMETRIOSIS

3 Article 122

Dow

nloaded from w

ww

.biolreprod.org.

encoded by this gene, caldesmon, could be used to discriminatebetween patients with and without endometriosis with aplausible diagnostic potential both during the proliferativeand secretory phases of the menstrual cycle, although it isrecommended that this analysis be performed during theproliferative phase because of safety and performance (Fig. 5).

DISCUSSION

The increased expression of the CALD1 gene in endome-triotic lesions has been previously reported in studies of large-scale hybridization based on subtractive libraries and micro-arrays [22–25]. We used real-time PCR to quantify theexpression of this gene to confirm these results. Thismethodology has been described as the gold standard for theanalysis of gene expression because it has high sensitivity,good reproducibility, and a wide quantification range;moreover, this technique reveals potential false-positive resultswhen used for the validation of different screening methodol-ogies [31, 32]. The present results confirm the increasedexpression of CALD1 in endometriotic peritoneal and ovarianlesions regardless of the phase of the menstrual cycle. Inaddition, the levels of caldesmon protein were also increased in

FIG. 2. The results obtained by Western blot analysis for h-CaD (120kDa) and l-CaD (77 kDa) isoforms. The SDS-PAGE profile of h-CaD and l-CaD proteins obtained for the control, eutopic endometrium, andendometriotic lesion groups.

FIG. 3. Data for the relative quantification of caldesmon protein in the proliferative phase (A) and in the secretory phase (B) of the menstrual cycle;relative quantification of h-CaD in the proliferative (C) and secretory phases (D); and relative quantification of l-CaD in the proliferative phase (E) andsecretory phase (F). Control, endometrium of women without endometriosis; Eutopic, endometrium of women with endometriosis; Lesion, endometrioticlesions (peritoneal and ovarian endometrioma). Comparisons showing significant differences are indicated by the P value in parentheses. The samplenumber for each group studied is also given. Comparisons between control and eutopic tissues were not paired analyses; comparisons between eutopictissue and lesion were paired analyses.

MEOLA ET AL.

4 Article 122

Dow

nloaded from w

ww

.biolreprod.org.

endometriotic lesions, but immunolocalization of caldesmondid not identify differences between the marked cell types inthe endometriotic lesions and in the eutopic endometrium.

The human caldesmon protein has six isoforms http://www.uniprot.org/uniprot/Q05682). Because of the high content ofglutamine residues, the h-CaD and l-CaD isoforms migrateanomalously slowly during SDS-PAGE and have apparentmolecular masses of 120–130 kDa and 70–80 kDa, respec-tively [33, 34]. A differential profile of caldesmon was detectedin primary human endometrium stromal cells with induceddecidualization. Bands of 77, 95, and 120 kDa were observedby Western blot analyses [35, 36]. In the current study, 77-kDa(l-CaD) and 120-kDa (h-CaD) isoforms were observed inendometrial tissues and endometriotic lesions.

The human CALD1 gene, which encodes the caldesmon(CaD) protein, consists of 14 exons mapped to the 7q33-q34region [26]. A single gene is encoded by the alternative splicingof exons (exons 3a, 3b, and 4) and distinct promoters (starting

with exon 1a-1, 1a-2, 1a-3, or 1b) in the following isoforms:aorta h-CaD (793 amino acid [aa] residues, 93.2 kDa), WI-38 l-CaD I (564 aa, 65.7 kDa),WI-38 l-CaD II (538 aa, 62.7 kDa),HeLa l-CaD I (558 aa, 64.3 kDa), and HeLa l-CaD II (532 aa,61.2 kDa) [26, 37, 38]. This protein has binding sites for actin,myosin, tropomyosin, and Ca2-calmodulin [39]. The caldesmonfamily is composed of two subfamilies: h-CaDs, which areexpressed in differentiated smooth muscle cells, and l-CaDs,which are expressed in dedifferentiated smooth muscle cells andin nonmuscle cells [40]. The main N- and C-terminal functionaldomains of h-CaD and l-CaD are identical, except for the lack ofa central region in l-CaD [41].

The h-CaD and l-CaD isoforms have similar biochemicalproperties, and when bound to actin, they inhibit actomyosinATPase activity [42]. The actomyosin system converts thechemical energy of ATP to mechanical force and is themolecular basis for smooth muscle contraction and severalevents of contraction and motility in nonmuscle cells, such as

FIG. 4. Immunolocalization of caldesmon in control endometrium (A), eutopic endometrium of women with endometriosis (B), endometrioma (C), andperitoneal lesion (D). The positive stain is in the stromal compartment (marked by arrows). Original magnification 340.

TABLE 1. The performance of the potential biomarkers CALD1 and caldesmon in the diagnosis of endometriosis according to disease stage.a

Potential biomarker Groups

Proliferative phaseb Secretory phasec

Sensitivity (%) Specificity (%) AUCd Pe Sensitivity (%) Specificity (%) AUCd Pe

CALD1 expression Control versus endometriosis 60 86.7 0.717 0.0122 75 66.7 0.727 0.0078Control versus stages I–II 42.9 86.7 0.505 0.9719 48.4 93.3 0.713 0.0519Control versus stages III–IV 61.5 100 0.831 0.0001 80 66.7 0.74 0.024

Caldesmon protein Control versus endometriosis 95 100 0.95 0.0001 90 93.3 0.963 0.0001Control versus stages I–II 100 100 1.0 0 90 93.3 0.973 0.0001Control versus stages III–IV 92.3 100 0.923 0.0001 90 93.3 0.953 0.0001

a The data used in this analysis were obtained from the endometrium of women with and without endometriosis during both the proliferative and thesecretory phases of the menstrual cycle by Western blot analysis.b Control, n¼ 15; endometriosis, n ¼ 20; I–II, n ¼ 7; III–IV, n¼ 13.c Control, n¼ 15; endometriosis, n¼ 20; I–II, n ¼ 10; III–IV, n¼ 10.d AUC, area under the ROC curve.e Significant differences at P , 0.05.

CALDESMON IS A POTENTIAL MARKER FOR ENDOMETRIOSIS

5 Article 122

Dow

nloaded from w

ww

.biolreprod.org.

cytokinesis and phagocytosis [43]. However, this effect can bereversed by Ca2þ-calmodulin [44] and by phosphorylation,which reduces the affinity of caldesmon for actin [45]. Both h-CaD and l-CaD are phosphorylated in vivo, either by smoothmuscle stimulation for contraction or by the activation ofnonmuscle cells for proliferation or migration [46, 47]. The twoisoforms are heat stable; bind to calmodulin, tropomyosin, andactin; inhibit the actomyosin ATPase system (actin-activatedMg2þ-ATPase myosin activity); and undergo the so-called flip-flop and bind to both actin filaments and calmodulin,depending on the concentration of free calcium. Thus, whenthe Ca2þ-calmodulin-caldesmon complex is formed, it bindsweakly to actin. Regarding intracellular localization, h-CAD islocated in the fine filaments (part of the contractility apparatusof smooth muscle cells) and regulates their contractions byinhibiting the actin-myosin interaction, whereas l-CAD ispresent in actin filaments and stress fibers [40].

Moreover, l-CaD has alternative functions that includestabilizing the actin filaments of nonmuscle cells to regulatecell contraction, adhesion signaling, and cytoskeleton organi-zation. This isoform also influences granule movement,hormone secretion, and microfilament reorganization duringmitosis via specific phosphorylation by cdc2 protein kinase[39]. Several other functions have been directly or indirectlyassociated with l-CaD, such as the inhibition of cellularcontractility and signaling-dependent adhesion [48] and thestabilization of stress fibers, thereby affecting cell morphology,growth, and motility [49]. There is controversy about theregulatory role of this isoform in the formation of podosomes,highly dynamic structures found in mobile and invasive cells,and in cellular invasion [50] and cell migration [51].

The deregulated expression of CALD1 and its alternativesplicings [52, 53], as well as its loss of function, have beenassociated with oncogenesis [50]. Zheng et al. [53] have relatedthe differential expression of CALD1 to neovascularization and

angiogenesis in gliomas and in tumors of the breast, lung,kidney, stomach, ovary, uterus, and liver. This gene isconsidered to be a potential serologic marker for gliomas[54], and its overexpression is related to possible therapeuticfunctions for glaucoma [55]. Alternative splicing and increasedexpression of the gene have also been shown in cancer of thecolon, bladder and prostate [52]. A study of 60 cell linesderived from human tumors has revealed overexpression of thegene in glioblastomas and renal cell tumors and moderateexpression in other sets of tumor lines [56].

In the present study, higher protein levels of caldesmon weredetected in the lesions and in control endometrium comparedwith the endometrium of women with endometriosis. Eutopicendometrium in patients with endometriosis is an importantfactor for the understanding of the physiopathology of thedisease because this tissue appears to be the center of variousalterations [57]. Lower levels of caldesmon in the endometriumof women with endometriosis may increase the motility andinvasiveness of this endometrium because if this protein is notbinding to actin, then it cannot inhibit the activity ofactomyosin ATPase. However, endometriotic lesions areknown to have a greater invasive potential than the endome-trium [58]. The activity of caldesmon is regulated byphosphorylation [45] and by Ca2þ-calmodulin [44], and theexpression of the CALM2 gene (which encodes calmodulin) isincreased in endometriotic lesions [22]. Thus, regardless of itslevels, the function of caldesmon of inhibiting the invasivecellular properties when bound to actin is related to thequantities of Ca2þþ available in the medium, to phosphoryla-tion, and to the quantity of calmodulin expressed. Wespeculated that, even if caldesmon is present at the sameconcentrations in the control endometrium and in endometrioticlesions, the lesions would have a greater invasive potentialbecause of the greater expression of the CALM2 gene, which isfound to be increased in the lesions, a fact that would prevent

FIG. 5. Receiver operating characteristic curves for the caldesmon protein in the proliferative and secretory phases of the menstrual cycle obtained in thecomparison of the control group to the groups with endometriosis, minimal–mild endometriosis (stage I–II), and moderate–severe endometriosis (stage III–IV). For these analyses, we used data obtained in unpaired endometrium of women with and without endometriosis.

MEOLA ET AL.

6 Article 122

Dow

nloaded from w

ww

.biolreprod.org.

caldesmon from binding to actin and inhibiting cell invasionand motility. Functional studies are needed to clarify theparticipation of caldesmon in the development of endometri-osis.

Several noninvasive tests for endometriosis have beenproposed as follows: a combination of CCR1 mRNA, MCP1,and CA-125 measurements showed a sensitivity of 92.2% anda specificity of 81.6% [59]; IL-6 with a cutoff point of 1.9 pg/ml showed 71% sensitivity and 66% specificity [60]; the totalamount of ccf nuclear DNA provided 70% sensitivity and 87%specificity [18]; a serum anti-syntaxin 5 autoantibody showed53.6% sensitivity and 72.2% accuracy [61]; a combinedanalysis of the mRNA levels of MMP-3, MMP-9, VEGF andsurvivin in the peripheral blood and the serum levels of CA-125 and CA19-9 in women with and without endometriosisshowed 87% sensitivity [62]; serum IL-6 and peritoneal fluidTNF-a showed 90% sensitivity and 67% specificity and 90%specificity and 100% sensitivity, respectively [63]; thesimultaneous evaluation of serum activin A and follistatinconcentrations for the diagnosis of ovarian endometriomashowed 41% sensitivity and 90% specificity [64]; thesensitivity and specificity were 61% and 96% for a serumanti-TMOD3b-autoantibody, 78% and 93% for an anti-TMOD3c-autoantibody and 78% and 96% for an anti-TMOD3d-autoantibody (used in the detection of the earlystages of endometriosis) [13]; the combined performance of IL-6, IL-8, TNF-a, hsCRP, CA-125, and CA19-9 diagnosedminimal to mild endometriosis with a sensitivity of 94%(specificity of 61%) during the secretory phase and with asensitivity of 92% (specificity of 63%) during the menstrualphase [15]; and as a diagnostic marker, urinary vitamin D-binding protein exhibited 58% sensitivity and 76% specificity[65].

These body fluid markers have limited success in practice[13, 15, 18, 59–65]. A double-blind study evaluated theefficacy of an endometrial biopsy to detect nerve fibers(PGP9.5) signifying endometriosis and had a detectionsensitivity of 98% (specificity of 83%) [9]. However, a rolefor PGP9.5-immunoactive nerve fibers in the functional layerof the endometrium has been suggested in pain generation indisorders such as endometriosis, adenomyosis, uterine fibroids,and endometriosis with adenomyosis [66], which could hinderthe clinical use of this feature for the diagnosis of endometri-osis. Another group using a combination of PGP9.5, vasoactiveintestinal peptide, and substance P was able to predict thepresence of minimal to mild endometriosis with 95%sensitivity (100% specificity) [67]. These tests require a carefulendometrial biopsy and meticulous attention to immunohisto-chemistry. Another study used a combined mRNA microarrayand proteomic analysis on the same eutopic endometriumsample obtained by biopsy from patients with and withoutendometriosis, which allowed the diagnosis of endometriosiswith 91% sensitivity (80% specificity) [68].

In the present study, we suggest a possible predictor ofendometrial dysregulation in patients with endometriosis thatinvolves a simple methodology. A potential limitation of ourstudy is the fact that others endometrial diseases were notexcluded, and therefore prospective studies are needed toconfirm the potential of caldesmon as a biomarker exclusivelyfor endometriosis.

ACKNOWLEDGMENT

We are grateful to the patients for their participation in this study, to theDepartment of Pathology, and to the multidisciplinary team of the HumanReproduction Division of the Department of Gynecology and Obstetrics

(Faculdade de Medicina de Ribeirao Preto, Universidade de Sao Paulo[FMRP–USP]) for sample collection and technical support.

REFERENCES

1. Yang WC, Chen HW, Au HK, Chang CW, Huang CT, Yen YH, TzengCR. Serum and endometrial markers. Best Pract Res Clin Obstet Gynaecol2004; 18(2):305–318.

2. Sampson JA. Peritoneal endometriosis due to menstrual dissemination ofendometrial tissue into the peritoneal cavity. Am J Obstet Gynecol 1927;14:422–429.

3. Berkley KJ, Stratton P. Mechanisms: lessons from translational studies ofendometriosis. In: Giamberardino MA (ed.), Visceral Pain: Clinical,Pathophysiological and Therapeutic Aspects, 1st ed. Oxford: OxfordUniversity Press: 2009:39–50.

4. Bulun S. Endometriosis. N Engl J Med 2009; 360:268–279.5. Ozkan S, Murk W, Arici A. Endometriosis and infertility: epidemiology

and evidence based treatments. Ann N Y Acad Sci 2008; 1127:92–100.6. Wanyonyi SZ, Sequeira E, Mukono SG. Correlation between laparoscopic

and histopathologic diagnosis of endometriosis. Int J Gynaecol Obstet2011; 115(3):273–276.

7. Bedaiwy MA, Falcone T. Laboratory testing for endometriosis. Clin ChimActa 2004; 340:41–56.

8. Meola J, Rosa e Silva JC, Hidalgo GS, Rocha MG, Ferriani RA. Diagnosisof endometriosis: laboratory and imaging diagnosis and future perspec-tives. In: Mitchell LA (ed.), Endometriosis: Symptoms, Diagnosis andTreatments, 1st ed. New York: Nova Science Publishers Press; 2010:73–92.

9. Al-Jefout M, Dezarnaulds G, Cooper M, Tokushige N, Luscombe GM,Markham R, Fraser IS. Diagnosis of endometriosis by detection of nervefibres in an endometrial biopsy: a double blind study. Hum Reprod 2009;24:3019–3024.

10. Ametzazurra A, Matorras R, Garcıa-Velasco JA, Prieto B, Simon L,Martınez A, Nagore D. Endometrial fluid is a specific and non-invasivebiological sample for protein biomarker identification in endometriosis.Hum Reprod 2009; 24:954–965.

11. Ferrero S, Gillott DJ, Remorgida V, Anserini P, Leung KY, Ragni N,Grudzinskas JG. Proteomic analysis of peritoneal fluid in women withendometriosis. J Proteome Res 2007; 6:3402–3411.

12. Fowler PA, Tattum J, Bhattacharya S, Klonisch T, Hombach-Klonisch S,Gazvani R, Lea RG, Miller I, Simpson WG, Cash P. An investigation ofthe effects of endometriosis on the proteome of human eutopicendometrium: a heterogeneous tissue with a complex disease. Proteomics2007; 7:130–142.

13. Gajbhiye R, Sonawani A, Khan S, Suryawanshi A, Kadam S, Warty N,Raut V, Khole V. Identification and validation of novel serum markers forearly diagnosis of endometriosis. Hum Reprod 2012; 27:408–417.

14. Liu H, Lang J, Zhou Q, Shan D, Li Q. Detection of endometriosis with theuse of plasma protein profiling by surface-enhanced laser desorption/ionization time-of-flight mass spectrometry. Fertil Steril 2007; 87:988–990.

15. Mihalyi A, Gevaert O, Kyama CM, Simsa P, Pochet N, De Smet F, DeMoor B, Meuleman C, Billen J, Blanckaert N, Vodolazkaia A, Fulop V, etal. Non-invasive diagnosis of endometriosis based on a combined analysisof six plasma biomarkers. Hum Reprod 2010; 25:654–664.

16. Seeber B, Sammel MD, Fan X, Gerton GL, Shaunik A, Chittams J,Barnhart KT. Proteomic analysis of serum yields six candidate proteinsthat are differentially regulated in a subset of women with endometriosis.Fertil Steril 2010; 93:2137–2144.

17. Tokushige N, Markham R, Crossett B, Ahn SB, Nelaturi VL, Khan A,Fraser IS. Discovery of a novel biomarker in the urine in women withendometriosis. Fertil Steril 2011; 95:46–49.

18. Zachariah R, Schmid S, Radpour R, Buerki N, Fan AX, Hahn S,Holzgreve W, Zhong XY. Circulating cell-free DNA as a potentialbiomarker for minimal and mild endometriosis. Reprod Biomed Online2009; 18:407–411.

19. Zhang H, Niu Y, Feng J, Guo H, Ye X, Cui H. Use of proteomic analysisof endometriosis to identify different protein expression in patients withendometriosis versus normal controls. Fertil Steril 2006; 86:274–282.

20. Somigliana E, Vigano P, Tirelli AS, Felicetta I, Torresani E, Vignali M, DiBlasio AM. Use of the concomitant serum dosage of CA 125, CA 19-9and interleukin-6 to detect the presence of endometriosis. Results from aseries of reproductive age women undergoing laparoscopic surgery forbenign gynaecological conditions. Hum Reprod 2004; 19:1871–1876.

21. Rosa e Silva A, Rosa e Silva JC, Ferriani RA. Serum CA-125 in thediagnosis of endometriosis. Int J Gynaecol Obstet 2007; 96:206–207.

22. Meola J, Rosa e Silva JC, Dentillo DB, da Silva WA Jr, Veiga-Castelli LC,

CALDESMON IS A POTENTIAL MARKER FOR ENDOMETRIOSIS

7 Article 122

Dow

nloaded from w

ww

.biolreprod.org.

Bernardes LA, Ferriani RA, de Paz CC, Giuliatti S, Martelli L.Differentially expressed genes in eutopic and ectopic endometrium ofwomen with endometriosis. Fertil Steril 2010; 93:1750–1773.

23. Mettler L, Salmassi A, Schollmeyer T. Comparison of c-DNA microarrayanalysis of gene expression between eutopic endometrium and ectopicendometrium (endometriosis). J Assist Reprod Genet 2007; 24:249–258.

24. Vouk K, Smuc T, Guggenberger C, Ribic-Pucelj M, Sinkovec J, Husen B,Thole H, Houba P, Thaete C, Adamski J, Rizner TL. Novel estrogen-related genes and potential biomarkers of ovarian endometriosis identifiedby differential expression analysis. J Steroid Biochem Mol Biol 2011; 125:231–242.

25. Zafrakas M, Tarlatzis BC, Streichert T, Pournaropoulos F, Wolfle U,Smeets SJ, Wittek B, Grimbizis G, Brakenhoff RH, Pantel K, Bontis J,Gunes C. Genome-wide microarray gene expression, array-CGH analysis,and telomerase activity in advanced ovarian endometriosis: a high degreeof differentiation rather than malignant potential. Int J Mol Med 2008; 21:335–334.

26. Hayashi K, Yano H, Hashida T, Takeuchi R, Takeda O, Asada K,Takahashi E, Kato I, Sobue K. Genomic structure of the human caldesmongene. Proc Natl Acad Sci U S A 1992; 89:12122–12126.

27. Hidalgo GS, MeolaJ, Rosa e Silva JC, Paz CCP, Ferriani RA. TAGLNexpression is deregulated in endometriosis and may be involved in cellinvasion, migration, and differentiation. Fertil Steril 2011; 96:700–703.

28. American Society for Reproductive Medicine. Revised American Societyfor Reproductive Medicine classification of endometriosis: 1996. FertilSteril 1997; 67:817–821.

29. Cook RD, Weisberg S. Transforming a response variable for linearity.Biometrika 1994; 81:731–737.

30. Zweig MH, Campbell G. Receiver-operating characteristic (ROC) plots: afundamental evaluation tool in clinical medicine. Clin Chem 1993; 39:561–577.

31. Goodsaid FM, Smith RJ, Rosenblum IY. Quantitative PCR deconstructionof discrepancies between results reported by different hybridizationplatforms. Environ Health Perspect 2004; 112:456–460.

32. Pfaffl MW, Horgan GW, Dempfle L. Relative expression software tool(REST) for group-wise comparison and statistical analysis of relativeexpression results in real-time PCR. Nucleic Acids Res 2002; 30:e36.

33. Bretscher A, Lynch W. Identification and localization of immunoreactiveforms of caldesmon in smooth and nonmuscle cells: a comparison with thedistributions of tropomyosin and a-actinin. J Cell Biol 1985; 100:1656–1663.

34. Graceffa P, Wang CL, Stafford WF. Caldesmon. Molecular weight andsubunit composition by analytical ultracentrifugation.J Biol Chem 1988;263:14196–14202.

35. Kilpatrick LM, Stephens AN, Hardman BM, Salamonsen LA, Li Y, StantonPG, Nie G. Proteomic identification of caldesmon as a physiologicalsubstrate of proprotein convertase 6 in human uterine decidual cellsessential for pregnancy establishment. J Proteome Res 2009; 8:4983–4992.

36. Paule SG, Airey LM, Li Y, Stephens AN, Nie G. Proteomic approachidentifies alterations in cytoskeletal remodeling proteins during decidual-ization of human endometrial stromal cells. J Proteome Res 2010; 9:5739–5747.

37. Humphrey MB, Herrera-Sosa H, Gonzalez G, Lee R, Bryan J. Cloning ofcDNAs encoding human caldesmons. Gene 1992; 112:197–204.

38. Novy RE, Lin JLC, Lin JJC. Characterization of cDNA clones encoding ahuman fibroblast caldesmon isoform and analysis of caldesmon expressionin normal and transformed cells. J Biol Chem 1991; 266:16917–16924.

39. Huber PA. Caldesmon. Int J Biochem Cell Biol 1997; 29:1047–1051.40. Sobue K, Sellers JR. Caldesmon, a novel regulatory protein in smooth

muscle and nonmuscle actomyosin systems. J Biol Chem 1991; 266:12115–12118.

41. Ball EH, Kovala T. Mapping of caldesmon: relationship between the highand low molecular weight forms. Biochemistry 1988; 27:6093–6098.

42. Ngai PK, Walsh MP. Inhibition of smooth muscle actin-activated myosinMg2þ-ATPase activity by caldesmon. J Biol Chem 1984; 259:13656–13659.

43. Sellers JR. Regulation of cytoplasmic and smooth muscle myosin. CurrOpin Cell Biol 1991; 3:98–104.

44. Horiuchi KY, Miyata H, Chacko S. Modulation of smooth muscleactomyosin ATPase by thin filament associated proteins. BiochemBiophys Res Commun 1986; 136:962–968.

45. Mak AS, Watson MH, Litwin CM, Wang JH. Phosphorylation ofcaldesmon by cdc2 kinase. J Biol Chem 1991; 266:6678–6681.

46. Franklin MT, Wang CL, Adam LP. Stretch-dependent activation anddesensitization of mitogen-activated protein kinase in carotid arteries. AmJ Physiol 1997; 273:C1819–C1827.

47. Yamboliev IA, Gerthoffer WT. Modulatory role of ERK MAPK-

caldesmon pathway in PDGF-stimulated migration of cultured pulmonaryartery SMCs. Am J Physiol Cell Physiol 2001; 280:C1680–C1688.

48. Helfman DM, Levy ET, Berthier C, Shtutman M, Riveline D, Grosheva I,Lachish-Zalait A, Elbaum M, Bershadsky AD. Caldesmon inhibitsnonmuscle cell contractility and interferes with the formation of focaladhesions. Mol Biol Cell 1999; 10:3097–3112.

49. Li Y, Lin JLC, Reiter RS, Daniels K, Soll DR, Lin JJC. Caldesmon mutantdefective in Ca2þ-calmodulin binding interferes with assembly of stressfibers and affects cell morphology, growth and motility. J Cell Sci 2004;117:3593–3604.

50. Tanaka J, Watanabe T, Nakamura N, Sobue K. Morphological andbiochemical analyses of contractile proteins (actin, myosin, caldesmon andtropomyosin) in normal and transformed cells. J Cell Sci 1993; 104:595–606.

51. Mayanagi T, Morita T, Hayashi K, Fukumoto K, Sobue K. Glucocorticoidreceptor-mediated expression of caldesmon regulates cell migration via thereorganization of the actin cytoskeleton. J Biol Chem 2008; 283:31183–31196.

52. Thorsen K, Sorensen KD, Brems-Eskildsen AS, Modin C, Gaustadnest M,Hein AMK, Kruhoffer M, Laurberg S, Borre M, Wang K, Brunak S,Krainer AR, et al. Alternative splicing in colon, bladder, and prostate canceridentified by exon array analysis. Mol Cell Proteomics 2008; 7:1214–1224.

53. Zheng PP, Sieuwerts AM, Luider TM, Weiden M, Sillevis-Smitt PAE,Kros J. Differential expression of splicing variants of the humancaldesmon gene (CALD1) in glioma neovascularization versus normalbrain microvasculature. Am J Pathol 2004; 164:2217–2228.

54. Zheng PP, Hop WC, Smitt PAES, Bent MJ, Avezaat CJJ, Luider TM,Krod JM. Low-molecular weight caldesmon as a potential serum markerfor glioma. Clin Cancer Res 2005; 11:4388–4392.

55. Gabelt BT, Hu Y, Vittitow JL, Rasmussen CR, Grosheva I, Bershadsky AD,Geiger B, Borras T, Kaufman PL. Caldesmon transgene expression disruptsfocal adhesion in HTM cells and increases outflow facility in organ-culturedhuman and monkey anterior segments. Exp Eye Res 2006; 82:935–944.

56. Ross DT, Scherf U, Eisen MB, Perou CM, Rees C, Spellman P, Iyer V,Jeffrey SS, Rijn MV, Waltham M, Pergamenschikov A, Lee JC, et al.Systematic variation in gene expression patterns in human cancer celllines. Nat Gen 2000; 24:227–235.

57. Brosens I, Brosens JJ, Benagiano G. The eutopic endometrium inendometriosis: are the changes of clinical significance? Reprod BiomedOnline 2012; 24:496–502.

58. Jiang QY, Wu RJ. Growth mechanisms of endometriotic cells in implantedplaces: a review. Gynecol Endocrinol 2012; 28:562–567.

59. Agic A, Djalali S, Wolfler MM, Halis G, Diedrich K, Hornung D.Combination of CCR1 mRNA, MCP1, and CA125 measurements inperipheral blood as a diagnostic test for endometriosis. Reprod Sci 2008;15:906–911.

60. Othman Eel-D, Hornung D, Salem HT, Khalifa EA, El-Metwally TH, Al-Hendy A. Serum cytokines as biomarkers for nonsurgical prediction ofendometriosis. Eur J Obstet Gynecol Reprod Biol 2008; 137:240–246.

61. Nabeta M, Abea Y, Takaokaa Y, Kusanagi Y, Ito M. Identification of anti-syntaxin 5 autoantibody as a novel serum marker of endometriosis. JReprod Immunol 2011; 91:48–55.

62. Mabrouk M, Elmakky A, Caramelli E, Farina A, Mignemi G, Venturoli S,Villa G, Guerrini M, Manuzzi L, Montanari G, De Sanctis P, Valvassori L,et al. Performance of peripheral (serum and molecular) blood markers fordiagnosis of endometriosis. Arch Gynecol Obstet 2011; 285:1307–1312.

63. Bedaiwy MA, Falcone T, Sharma RK, Goldberg JM, Attaran M, NelsonDR, Agarwal A. Prediction of endometriosis with serum and peritoneal fluidmarkers: a prospective controlled trial. Hum Reprod 2002; 17:426–431.

64. Reis FM, Luisi S, Abrao MS, Rocha ALL, Vigano P, Rezende CP, FlorioP, Petraglia F. Diagnostic value of serum activin A and follistatin levels inwomen with peritoneal, ovarian and deep infiltrating endometriosis. HumReprod 2012; 27:1445–1450.

65. Cho S, Choi YS, Yim SY, Yang HI, Jeon YE, Lee KE, Kim HY, Seo SK,Lee BS. Urinary vitamin D-binding protein is elevated in patients withendometriosis. Hum Reprod 2012; 27:515–522.

66. Zhang X, Lu B, Huang X, Xu H, Zhou C, Lin J. Endometrial nerve fibersin women with endometriosis, adenomyosis, and uterine fibroids. FertilSteril 2009; 92:1799–1801.

67. Bokor A, Kyama CM, Vercruysse L, Fassbender A, Gevaert O,Vodolazkaia A, Moor BD, Fulop V, D’Hooghe T. Density of smalldiameter sensory nerve fibres in endometrium: a semi-invasive diagnostictest for minimal to mild endometriosis. Hum Reprod 2009; 24:3025–3032.

68. Fassbender A, Verbeeck N, Bornigen D, Kyama CM, Bokor A,Vodolazkaia A, Peeraer K, Tomassetti C, Meuleman C, Gevaert O, Vande Plas R, Ojeda F, et al. Combined mRNA microarray and proteomicanalysis of eutopic endometrium of women with and without endometri-osis. Hum Reprod 2012; 27(7):2020–2029.

MEOLA ET AL.

8 Article 122

Dow

nloaded from w

ww

.biolreprod.org.