Journal of
Clinical Endocrinology and Metabolism
Volume 85 • Number 7 • July 2000
Copyright © 2000 The Endocrine Society
DANIELA HORNUNG 1
KATHRIN DOHRN 1
KARL SOTLAR 1
ROBERT R. GREB 1
DIETHELM WALLWIENER 1
LUDWIG KIESEL 1
ROBERT N. TAYLOR 2
1 Department of Obstetrics
and Gynecology (D.H., K.D., R.R.G., D.W., L.K.) and Department of Pathology
(K.S.), University of Tubingen, Germany
2 and
Center for Reproductive Sciences (D.H., R.N.T.), Department of Obstetrics,
Gynecology, and Reproductive Sciences, University of California, San
Francisco, California 94143-0556
Received November 29, 1999.
Revision received March 4, 2000.
Accepted March 22, 2000.
Address correspondence and requests for reprints to: Robert N. Taylor, M.D., Ph.D., Center for Reproductive Sciences, HSE 1689, Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, San Francisco, California 94143-0556.
Copyright © 2000 by The Endocrine Society
Our laboratories
have focused recently on the production and localization of eotaxin, a
C-C-chemokine of 8.4 kDa, whose major biological activity is the
chemoattraction of eosinophils. Given evidence of autoimmune activity in the endometriosis syndrome, we hypothesized that
eosinophil chemoattractants might be expressed in endometriosis.
In histological sections, we observed eotaxin protein localized mainly in
epithelial cells, with only very faint immunostaining in the surrounding
stromal cells. Prominent eotaxin accumulation was noted in the luminal
epithelium of secretory endometrium. Eotaxin distribution in endometriosis was similar to that seen in eutopic
endometrium but with higher levels of eotaxin staining in the glandular
epithelium. Peritoneal fluid concentrations of eotaxin were significantly
higher in women with moderate or severe endometriosis
than in women with minimal or mild endometriosis
or no disease. The treatment of isolated human endometriosis
epithelial cells with estradiol, medroxyprogesterone acetate, tumor necrosis
factor-alpha, and interferon-gamma stimulated measurable eotaxin secretion into
the conditioned media. The results indicate that eotaxin is produced in
epithelial cells of normal endometrium and endometriosis
tissues, varies across the menstrual cycle, and is elevated in women with endometriosis. We postulate that eotaxin, interacting
with other known cytokines and immune cells, contributes to an inflammatory
reproductive tract environment, leading to endometrial or blastocyst dysfunction.
( J Clin Endocrinol Metab 85: 2604-2608, 2000)
ENDOMETRIOSIS is a common, but complex,
gynecological syndrome of unknown pathogenesis that affects 5-15% of
reproductive-aged women. It commonly is accompanied by debilitating symptoms of
dysmenorrhea, pelvic pain, and reduced fecundity. Although multiple theories of
the histogenesis of endometriosis exist, the
implantation hypothesis of Sampson [1] is the most widely
accepted. Retrograde menstruation [2] and intraperitoneal spillage
of viable endometrial cells [3] occur frequently in cycling
women, and mullerian tract outflow obstruction is associated with an increased
prevalence of endometriosis [4] .
The
pathophysiology of endometriosis-associated pain and infertility remains
enigmatic, but current evidence suggests that these symptoms result from local
inflammation at the implant sites [5] . Pelvic implants are
associated with gross and microscopic evidence of local inflammatory changes, viz.
neovascularization, fibrous scarring, and accumulation of activated
inflammatory cells [6] . It has been proposed that
the recruitment of pelvic immune cells initiates an intraperitoneal cascade of
cytokines that mediate the pain and infertility that accompany endometriosis [7] .
Autoinflammatory
phenomena, including autoantibody production and atopy, have been associated
with endometriosis. Mathur and her colleagues [8] [9] were the first to describe,
in this disease, autoantibodies that recognize endometrial proteins ranging in
size from 34-140 kDa. Similar results have been confirmed by other groups [10] [11] . Gleicher and colleagues [12] noted that a significant
proportion (40-60%) of women with endometriosis
has elevated autoantibody titers when tested against a panel of common, generic
autoantigens ( e.g. phospholipid, ribonucleoprotein, and double-stranded
DNA). Thus, in addition to the development of specific antiendometrial
antibodies, generalized polyclonal B-cell activation is associated with some
cases of endometriosis [13] .
A highly
significant correlation has been noted between surgically documented endometriosis and the incidence of atopic allergic
symptoms [14] . This association was
confirmed in a recent survey, conducted by the Endometriosis
Association, of 4,000 North American women [15] . In that study, 41% and
17% of surveyed endometriosis patients reported
a history of pollen allergy and eczema, respectively, compared with 13% and 6%
of women in the general population. In a cohort of 40 endometriosis
patients with a predominant complaint of fatigue, 65% had allergy symptoms and
positive serum IgE and IgG radioallergosorbent tests (D. Metzger, personal
communication). On the basis of these clinical findings, which suggest an
autoinflammatory or allergic component in endometriosis,
we hypothesized that the biochemical mediators of atopic reactivity might be
increased in women with this syndrome. As a prototype for such molecules, we
selected the 8.4-kDa C-C eosinophil chemokine,
* These studies were
supported by Fortune-Project F1241133 and a grant from the NIH/NICHD, through
cooperative agreement U54-HD-37321, as part of the Specialized Cooperative Centers
Program in Reproduction Research.
eotaxin,
and first investigated its peritoneal fluid concentrations in a case-control
study of subjects with endometriosis and
matched, normal women. These findings were then extended, using an established endometriosis cell model [7] , to determine whether
eotaxin protein secretion could be induced in vitro via endocrine and/or
paracrine effectors.
Healthy
ovulatory women, who had not received hormones or GnRH agonist therapy for at
least 6 months before surgery, were recruited, after they had provided written
informed consent, under a study protocol approved by the UCSF Committee on
Human Research at the University of California, San Francisco. Women with endometriosis (n = 15) were staged intraoperatively,
according to a modification of the revised American Fertility Society system,
in which the extent of active endometriosis
lesions were scored [5] . Control subjects (n = 7)
were women with subserosal leiomyomata or uterine descensus or requesting tubal
ligation without evidence of pelvic pathology.
Tissue specimens
and peritoneal fluid samples were obtained from patients undergoing laparoscopy or laparotomy. Endometrial and endometriosis biopsies were collected under sterile
conditions for cell culture and immunohistochemical analyses. All samples and
cycle stages were estimated histologically according to the criteria of Noyes et
al. [16] . All normal endometrial
biopsies were in phase and consistent with the patients' menstrual dating.
Peritoneal fluid
was aspirated immediately upon entering the peritoneal cavity, cells were
removed by centrifugation, and aprotinin (15 nmol/L) was added to the
supernatant before freezing at -70 C. Pelvic fluid, endometriosis
specimens, and all biopsies for cell culture were taken in the midproliferative
phase of the cycle, as described previously [7] .
Endometrial and endometriosis tissues were either frozen or fixed for
24 h in 2% paraformaldehyde and 0.5% glutaraldehyde, paraffin-embedded, cut in
serial sections of 5 mum, and stained using the Vectastain Elite ABC kit
(Vector Laboratories, Inc. Burlingame, CA). Immunoperoxidase staining was
performed overnight at 4 C using mouse monoclonal IgG antibodies against human
cytokeratin 18 (1:2000 dilution, Sigma, Munchen, Germany) and eotaxin (1:50
dilution, R&D Systems, Wiesbaden, Germany). Controls for the immunostaining
specificity included sections stained with antieotaxin antibodies (1:50
dilution) immunoabsorbed with 1.2 mumol/L recombinant eotaxin protein (R&D
Systems). Diaminobenzidine (Zymed Laboratories, Inc., South San Francisco, CA)
was used as the chromagen. All sections also were lightly counterstained with
hematoxylin. Eight blinded observers scored the immunostaining intensity, from
0 (no staining) to 4 (most intense staining), relative to a cytokeratin
positive control. The mean ± SD score of the eight observers is reported.
Primary
endometrial and endometriosis cell cultures were
prepared from biopsies, as we have described previously [7] . Glandular epithelial
cells were separated from stromal cells and debris by filtration through
narrow-gauge sieves. Stromal cells were subcultured to eliminate contamination
by macrophages or other leukocytes, and experiments were performed at passage
2. Extensive characterization of cell cultures, prepared using this protocol,
confirmed that they were more than 95% pure and retained functional markers of
their endometrial and endometriosis origin in
vivo [17] . At the end of each
experiment, cells were counted using the acid phosphatase colorimetric assay
described by Ueda et al. [18] .
When the primary
cell cultures approached confluence, the complete medium was removed and
replaced with fresh alpha-MEM containing 2.5% FCS and antibiotics, and the
cells were cultured for an additional 48 h with tumor necrosis factor-alpha (TNF-alpha;
5.9 nmol/L, Sigma), interferon-gamma (IFN-gamma; 4.2 nmol/L, Sigma), 10 nmol/L
estradiol (E2 , Merck & Co., Inc., Darmstadt, Germany), and/or 100 nmol/L
medroxyprogesterone acetate (MPA, Sigma). MPA was used in place of natural
progesterone because it is more slowly metabolized in tissue culture. This
combination of cytokines and steroids showed maximal stimulatory effects in
prior experiments [5] [7] [17] and was selected to model
the secretory phase of the endometrial cycle in vitro.
Eotaxin
concentrations were measured using a sandwich ELISA (R&D Systems). This
assay uses murine monoclonal antibodies and goat polyclonal antibodies against
eotaxin. The assay does not cross-react with several cytokines that are closely
related to eotaxin, including MCP-1, MCP-2, MCP-3, MIP-1alpha, MIP-1beta (macrophage
inflammatory protein-1alpha and beta), and RANTES (regulated upon activation,
normal T cell-expressed and secreted). The sensitivity limit of the assay was
0.6 pmol/L, with intraassay coefficients of variation less than 6% and
interassay coefficients of variation less than 12%.
All experiments
were repeated a minimum of three times, and the results are presented as the
mean ± SD. Kolmogorov-Smirnov
analyses demonstrated that the distribution of the results was Gaussian and did
not differ between normal and endometriosis
cases ( P = 0.31). The data were analyzed by ANOVA with Fisher's post
hoc tests for multiple comparisons. Linear regression analysis was
performed to determine the correlation between disease severity and pelvic
fluid eotaxin concentration. Significant differences were accepted when
two-tailed analyses yielded P < 0.05.
Immunohistochemistry
was used to localize eotaxin protein in fixed and frozen tissue. A section of
normal proliferative endometrium, stained with hematoxylin and eosin, is shown
in Fig.
1A . In an adjacent section, monoclonal antibodies against cytokeratin
specifically stained the glandular and luminal epithelium ( Fig.
1B ). Monoclonal mouse IgG antibodies against eotaxin showed this antigen
to be localized primarily in the epithelium, with the stromal compartment
appearing relatively free of the antigen ( Fig.
1C ). Control experiments were performed on serial sections of endometrium
using eotaxin antibodies immunoabsorbed with excess eotaxin protein ( Fig.
1D ). Experiments with frozen sections revealed the same staining pattern (data
not shown). During the secretory phase of the cycle, we observed increased
amounts of eotaxin immunoreactivity, and the epithelial cells were more
distinctively highlighted than in proliferative phase endometrium ( Fig.
1E ). The specificity of the staining pattern is shown again, after
immunoabsorption with an excess of pure recombinant human eotaxin ( Fig.
1F ). Normal secretory endometrium demonstrated more luminal than glandular
eotaxin staining ( Fig.
1G ). Endometriosis tissues also
demonstrated eotaxin protein. A section of an ovarian endometrioma stained with
eotaxin showed pat
Figure 1.
Immunohistochemistry of eotaxin in human endometrial and endometriotic tissues.
Paraffin-embedded sections of normal, proliferative-phase endometrium were stained
with hematoxylin and eosin (A), anticytokeratin antibodies to label epithelial
cells (B), and antieotaxin antibodies (C). The predominantly epithelial
staining of eotaxin was obliterated when the latter antibodies were preabsorbed
with excess eotaxin protein (D). Sections of secretory endometrium showed
strong, epithelial eotaxin staining (E), which also was eliminated by
preabsorption with the antigen (F). Several sections of secretory endometrium
revealed that eotaxin staining intensity was greater in the luminal epithelium
than in glandular epithelium (G). Note that the dark appearance of the
deep glands at the right side of the figure is attributable to densely
packed, hematoxylin-positive nuclei, rather than brown immunoperoxidase
product. Endometriotic epithelial cells stained intensely for eotaxin (H).
Magnification, ×400.
terns
very similar to those observed in normal endometrium ( Fig.
1H ).
The scoring of
eotaxin immunoreactivity in normal endometrial glands, by blinded observers,
demonstrated greater intensity in the secretory phase (2.9 ± 0.6) than in the
proliferative phase (1.8 ± 0.6). The immunoabsorption controls
had
significantly lower intensity scores (0.1 ± 0.3). Eotaxin staining intensity
was greater in glandular epithelium of endometriosis
implants biopsied in the proliferative phase (3.0 ± 0.6) than in endometrium
biopsies from the same phase of the cycle (1.8 ± 0.6). Each of the above
comparisons differed significantly [ P < 0.05 (ANOVA with Fisher's
tests)].
Peritoneal fluid
from women with endometriosis and from women
without the disease was collected and assayed for secreted eotaxin, by ELISA.
All peritoneal fluids were obtained in the midproliferative phase of the cycle.
There was a slight increase, but not statistically significant difference,
between controls (11.0 ± 3.7 pmol/L) and patients with minimal or mild endometriosis (13.8 ± 6.3 pmol/L; modified American
Fertility Society (AFS) stages I and II). But women with moderate-to-severe endometriosis, based on active implant scoring (see
Ref. [24] ; modified AFS stages III
and IV), had significantly higher concentrations of peritoneal fluid eotaxin
(20.6 ± 6.5 pmol/L) than women without disease or those with minimal or mild
disease ( P < 0.05; Fig.
2 ). There was a positive correlation (r = 0.56) between the severity of endometriosis and the concentration of eotaxin in
peritoneal fluid ( P < 0.01).
To determine
whether isolated human endometrial cells preserved the capacity to secrete
eotaxin, these were grown to confluence and treated for 48 h with E2 (10 nmol/L), MPA (100
nmol/L), TNF-alpha (5.9 nmol/L), and IFN-gamma (4.2 nmol/L) or not stimulated.
Conditioned media were collected after 48 h and assayed for secreted eotaxin,
by ELISA. Neither normal endometrial nor endometriosis
stromal cells secreted
Figure 2.
Peritoneal fluid concentrations of eotaxin were determined using a sensitive
ELISA. Pelvic fluid was collected during the midproliferative phase of the
menstrual cycle. Subjects were classified as having no evidence of endometriosis (controls, n
= 7), minimal-moderate endometriosis (modified AFS stages I-II, n = 8), or
moderate-severe endometriosis (modified AFS stages III-IV, n = 7).
The latter group had significantly higher peritoneal fluid eotaxin
concentrations (ANOVA with Fisher's post hoc tests; * , P
< 0.05).
measurable
eotaxin protein under any conditions (<1.2 fmol eotaxin/100,000 cells).
Likewise, normal endometrial epithelial cells failed to secrete eotaxin under
any of the above conditions. However, conditioned media from isolated endometriosis epithelial cells contained 856 ± 181
fmol eotaxin/100,000 cells after 48 h stimulation with E2 (10 nmol/L), MPA (100
nmol/L), TNF-alpha (5.9 nmol/L), and IFN-gamma (4.2 nmol/L). Unstimulated endometriosis epithelial cells or those stimulated
with either cytokines or steroids alone had no measurable eotaxin protein. Only
the combination of these factors stimulated the endometriosis
epithelial cells to secrete detectable eotaxin (ANOVA, P < 0.01).
Eotaxin was
initially identified as a specific chemoattractant protein for eosinophils;
however, recent studies indicate that it may dictate a broader scope of
activities on myeloid cells during development and in pathological states [19] . The detection of
immunoreactive eotaxin protein in normal endometrium ( Fig.
1 ; C, E, and G) and more intense staining in endometriosis
biopsies ( Fig.
1H ) indicates that these tissues have the potential to accumulate this
chemokine. A caveat is that the latter observation is cross-sectional, because
the same women did not have simultaneous eutopic and ectopic biopsies compared.
The distribution of eotaxin staining essentially was confined to the epithelial
layer, similar to observations made in guinea pig lung [20] . Whereas this pattern is
the same as that reported for the related C-C monocyte chemokine MCP-1 [21] , it is quite distinct from
our localization of the C-C monocyte chemokine RANTES in the stromal
compartment of normal endometrium and endometriosis
implants [7] . These findings indicate
that the accumulation of immune cell activator proteins in human endometrium
and endometriosis is remarkably cell specific.
As reported for MCP-1 [22] and RANTES [7] , normal endometrial
eotaxin accumulation seems to be regulated during the menstrual cycle, with
enhanced protein in the secretory phase ( Fig.
1G ).
Our recent
detection of eotaxin mRNA transcripts in endometriosis
implants, by reverse transcription-PCR (data not shown), and the de novo
synthesis of eotaxin by isolated endometriosis
epithelial cells suggest that the eotaxin gene is expressed in this tissue.
Stromal cells isolated from normal endometrium and endometriosis
lesions failed to secrete eotaxin under basal, cytokine, and/or steroid
hormone-stimulated conditions (<1.2 fmol/100,000 cells). Epithelial cells
isolated from normal endometrium likewise did not secrete immunodetectable
eotaxin under basal, cytokine, and/or steroid hormone-stimulated conditions
(<1.2 fmol/100,000 cells). However, epithelial cells derived from
endometriomas were induced to secrete 856 ± 181 fmol eotaxin/100,000 cells
after treatment with cytokines and steroid hormones ( P < 0.01).
Pelvic fluid
concentrations of eotaxin were correlated positively with endometriosis stage (r = 0.56, P < 0.01) and
were statistically elevated in women with advanced stages (AFS stages III and
IV) of active disease. This observation is similar to previous studies of
TNF-alpha [23] , interleukin-8 [24] , and vascular endothelial
growth factor [25] and supports the proposal
that active endometriosis lesions secrete
eotaxin
and
other cytokines into the peritoneal environment. Our results indicate that endometriosis epithelial cells have a preferential
ability to synthesize and secrete eotaxin, relative to those derived from
normal endometrium. Production of this eosinophil chemokine seems to be under
both endocrine and paracrine control, with enhanced secretion under conditions
that mimic the secretory phase of the menstrual cycle. The precise role of
eotaxin in the pathogenesis of endometriosis
remains to be elucidated; however, our findings suggest that autoinflammatory
and allergic phenomena associated with this syndrome may be linked to the
recruitment of eosinophils and other myeloid cells into the peritoneal cavity.
This hypothesis is currently under investigation in our laboratories.
1. Sampson
J. 1927 Peritoneal endometriosis due to
menstrual dissemination of endometrial tissue into the peritoneal cavity. Am J
Obstet Gynecol. 14:422-429.
2. Halme
J, Hammond MG, Hulka JF, Raj SG, Talbert LM. 1984 Retrograde menstruation in
healthy women and in patients with endometriosis.
Obstet Gynecol. 64:151-154. abstract
3. Kruitwagen
RF, Poels LG, Willemsen WN, Jap PH, Thomas CM, Rolland R. 1991 Endometrial
epithelial cells in peritoneal fluid during the early follicular phase. Fertil
Steril. 55:297-303. abstract
4. Cramer
DW, Hornstein MD, Ng WG, Barbieri RL. 1996 Endometriosis
associated with the N314D mutation of galactose-1-phosphate uridyl transferase
(GALT). Mol Hum Reprod. 2:149-152. abstract
5. Taylor
RN, Ryan IP, Moore ES, Hornung D, Shifren JL, Tseng JF. 1997 Angiogenesis and
macrophage activation in endometriosis. Ann NY
Acad Sci. 828:194-207.
6. Halme
J, Becker S, Haskill S. 1987 Altered maturation and function of peritoneal
macrophages: possible role in the pathogenesis of endometriosis.
Am J Obstet Gynecol. 156:783-785. abstract
7. Hornung
D, Ryan IP, Chao VA, Vigne J-L, Schriock ED, Taylor RN. 1997 Immunolocalization
and regulation of the chemokine RANTES in human endometrial and endometriosis tissues and cells. J Clin Endocrinol
Metab. 82:1621-1628. full text
8. Mathur
S, Peress MR, Williamson HO. 1982 Autoimmunity to endometrium and ovary in endometriosis. Clin Exp Immunol. 50:259-265. abstract
9. Mathur
S, Garza DE, Smith LF. 1990 Endometrial autoantigens eliciting immunoglobulin
(Ig)G, IgA, and IgM responses in endometriosis.
Fertil Steril. 54:56-63. abstract
10. Kennedy
SH, Starkey PM, Sargent IL, Hicks BR, Barlow DH. 1990 Antiendometrial
antibodies in endometriosis measured by an enzyme-linked
immunosorbent assay before and after treatment with danazol and nafarelin.
Obstet Gynecol. 75:914-918. abstract
11. Wild
RA, Shivers CA, Medders D. 1992 Detection of antiendometrial antibodies in
patients with endometriosis: methodological
issues. Fertil Steril. 58:518-521. abstract
12. Gleicher
N, El-Roeiy A, Confino E. 1987 Is endometriosis
an autoimmune disease? Obstet Gynecol. 70:115-121. abstract
13. Badawy
SZ, Cuenca V, Kaufmann L, Stitzel A, Thompson M. 1989 The regulation of
immunoglobulin production by B cells in patients with endometriosis.
Fertil Steril. 51:770-773. abstract
14. Nichols
TR, Lamb K, Arkins JA. 1987 The association of atopic diseases with endometriosis. Ann Allergy. 59:360-363. abstract
15. Ballweg
ML. 1998 New endometriosis association research
shows disease is starting younger and is more severe. Endomet Ass Newsletter.
19:6-10.
16. Noyes
RW, Hertig AT, Rock J. 1950 Dating the endometrial biopsy. Fertil Steril.
1:3-25.
17. Ryan
IP, Schriock ED, Taylor RN. 1994 Isolation, characterization, and comparison of
human endometrial and endometriosis cells in
vitro. J Clin Endocrinol Metab. 78:642-649. abstract
18. Ueda
Y, Walsh E, Nakanishi H, Yoshida K. 1994 A colorimetric assay method for the
evaluation of neurotrophic activity in vitro. Neurosci Lett.
165:203-207. abstract
19. Quackenbush
EJ, Wershil BK, Aguirre V, Gutierrez-Ramos JC. 1998 Eotaxin modulates
myelopoiesis and mast cell development from embryonic hematopoetic progenitors.
Blood. 92:1887-1897. abstract
20. Cook
EB, Stahl JL, Lilly CM, et al. 1998 Epithelial cells are a major source of the
chemokine eotaxin in the guinea pig lung. Allergy Asthma Proc. 19:15-22. abstract
21. Jolicoeur
C, Boutouil M, Drouin R, Paradis I, Lemay A, Akoum A. 1998 Increased expression
of monocyte chemotactic protein-1 in the endometrium of women with endometriosis. Am J Pathol. 152:125-133. abstract
22. Jones
RL, Kelly RW, Critchley HO. 1997 Chemokine and cyclooxygenase-2 expression in
human endometrium coincides with leukocyte accumulation. Hum Reprod.
12:1300-1306. abstract
23. Eisermann
J, Gast M, Pineda J. 1988 Tumor necrosis factor in peritoneal fluid of women
undergoing laparoscopic surgery. Fertil Steril. 50:573-579. abstract
24. Ryan
IP, Tseng JF, Schriock ED, Khorram O, Landers DV, Taylor RN. 1995 Interleukin-8
concentrations are elevated in peritoneal fluid of women with endometriosis. Fertil Steril. 63:929-93. abstract
25. Shifren
JL, Tseng JF, Zaloudek CJ, et al. 1996 Ovarian steroid regulation of vascular
endothelial growth factor in the human endometrium: implications for
angiogenesis during the menstrual cycle and in the pathogenesis of endometriosis. J Clin Endocrinol Metab. 81:3112-3118. full text
ana sayfaya dönmek için tikla http://www.geocities.com/sinandoganturk/kadindogum1.html