Chemokine expression is dysregulated in the endometrium of women using progestin-only contraceptives and correlates to elevated recruitment of distinct leukocyte populations

Rebecca L. Jones1, Naomi B. Morison1,4, Natalie J. Hannan1,2, Hilary O.D. Critchley3 and Lois A. Salamonsen1

1 Prince Henry’s Institute of Medical Research, PO Box 5152, Clayton, Victoria 3168, 2 Department of Obstetrics & Gynaecology, Monash University, Clayton, Victoria 3168, Australia and 3 Department of Reproductive and Developmental Sciences, University of Edinburgh, Centre for Reproductive Biology, Edinburgh, UK

4 To whom correspondence should be addressed. E-mail: naomi.morison{at}phimr.monash.edu.au
R.L.Jones and N.B.Morison contributed equally to this work


    Abstract
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Acknowledgements
 References
 
BACKGROUND: Breakthrough bleeding (BTB) is the most common reason for discontinuation of progestin-only (p-only) contraceptives, yet the causes are unknown. Use of p-only contraceptives is associated with elevated influx of endometrial leukocytes, similar to that observed perimenstrually or within decidualized endometrium. We hypothesized that chemokine expression is altered in women using p-only contraceptives, leading to abnormal leukocyte recruitment and BTB. METHODS: Expression of eight highly abundant endometrial chemokines was examined using immunohistochemistry and real-time PCR, in endometria from women using subdermal and intrauterine levonorgestrel and correlated to leukocyte subpopulations. RESULTS: Macrophage-derived chemokine (MDC), hemofiltrate CC chemokine-1 (HCC-1), monocyte chemoattractant protein-3 (MCP-3), interleukin-8 (IL-8) and eotaxin were strongly produced by epithelial glands, comparable to levels in premenstrual phase endometrium. Stromal cells were negative for chemokines in atrophic/shedding endometria, but intensely positive in highly decidualized tissues for MDC, MCP-3, HCC-1 and 6Ckine. Macrophage inflammatory protein-1{beta} (MIP-1b) and HCC-4 were suppressed in all p-exposed endometria. CONCLUSIONS: These data demonstrate that chemokine expression is dysregulated in p-exposed endometria, consistent with the morphological appearance of the endometrium and the leukocyte subsets present. This reinforces a potential role for chemokines in the elevated leukocyte recruitment that contributes to endometrial fragility and BTB.

Key words: chemokines/contraceptives/endometrium/leukocytes/progestin


    Introduction
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Acknowledgements
 References
 
Long-acting progestin-only (p-only) contraceptives are highly effective contraceptive methods, making them an appealing option for women in both developed and developing countries (Brache et al., 2003Go). However, the side-effect of breakthrough bleeding (BTB) remains the single most common reason for their discontinuation (Belsey et al., 1988Go). The causes of the endometrial fragility and breakdown are not fully understood; the main aims of current research are to develop new, improved contraceptive methods or therapeutic agents to prevent or treat BTB.

P-only contraceptives have different effects on the endometrium depending on the progestin used, the dose and the route of administration. Intrauterine delivery of progestin (levonorgestrel-releasing intrauterine system; LNG-IUS; Mirena®) results in high local concentrations, causing rapid and extensive decidualization of the endometrium, epithelial atrophy and loss of cyclicity (Nilsson et al., 1978Go; Silverberg et al., 1986Go). In contrast, subdermal implants (Norplant® and Implanon®) produce lower intrauterine concentrations of progestin and have a less dramatic and more variable effect on endometrial morphology (Marsh et al., 1995Go; Mascarenhas et al., 1998Go; Vincent and Salamonsen, 2000Go). Norplant endometrial samples have been subclassified based on morphology as atrophic (narrow proliferative-type epithelial glands, but with low mitotic activity), shedding (areas of distinct breakdown apparent) and progestin-modified (p-modified; decidualized) (Marsh et al., 1995Go; Vincent et al., 1999Go).

A common feature of all p-only contraceptives is an abnormal endometrial leukocyte infiltrate. The normal endometrium possesses an active and tightly regulated immune environment, with significant immune cell populations present associated with specific endometrial events (Bulmer et al., 1991Go). Few leukocytes are present in the proliferative phase, but there is a dramatic accumulation of leukocytes following ovulation, associated with implantation and decidualization. The majority are a unique endometrial-specific subset of natural killer (NK) cells (uNKs), which appear to be important for endometrial remodelling associated with the establishment of pregnancy when they are highly abundant in decidua (Moffett-King, 2002Go). Macrophages are present throughout the menstrual cycle but their numbers increase from the mid-secretory phase. Menstruation is associated with a dramatic influx of inflammatory leukocytes: neutrophils and eosinophils, and increased activation of resident mast cells, which provide proteases including matrix metalloproteinases and inflammatory mediators capable of initiating the tissue breakdown (Salamonsen and Lathbury, 2000Go).

Leukocyte trafficking appears to be dysregulated in p-exposed endometria, with excessive numbers of inflammatory leukocytes reported (Song et al., 1996Go; Vincent et al., 1999Go). Importantly, differences have been reported in the immune cell subpopulations in the endometria of users of different p-only contraceptives and between the different groups of Norplant-exposed endometrium; there are predominantly neutrophils and eosinophils in shedding endometria, whilst very few of these subtypes are in atrophic or p-modified endometria (Vincent et al., 1999Go). In the highly decidualized endometria of women using LNG-IUS, uNKs and macrophages are abundant (Critchley et al., 1998aGo).

As normal menstruation is an inflammatory-type event, with leukocyte infiltrate thought to play a major role in the focal endometrial breakdown (Salamonsen and Woolley, 1999Go), the abnormal leukocyte infiltrate in women using p-only contraceptives has been postulated to contribute to BTB. Leukocyte trafficking into normal endometrium is tightly regulated (Hornung et al., 1997Go; Jones et al., 1997Go; Akiyama et al., 1999Go; Zhang et al., 2000Go). Our previous studies have identified the most abundant chemokines in the endometrium associated with selective leukocyte recruitment (Jones et al., 2004Go): monocyte chemoattractant protein (MCP)-3, 6Ckine, macrophage derived chemokine (MDC), hemofiltrate CC chemokine (HCC)-1, HCC-4, eotaxin, interleukin (IL)-8, and macrophage inflammatory protein (MIP)-1{beta}. Distinct cohorts of chemokines are up-regulated coincidental with the accumulation of decidual-associated leukocytes and with the influx of menstruation-associated leukocytes. We hypothesize that the normal regulation of leukocyte entry into the endometrium is disrupted by the use of progestins, resulting in the abnormal leukocyte infiltrate that is associated with BTB. We therefore examined the production of endometrial chemokines in women using two different methods of p-only contraceptive (LNG-IUS and Norplant), and correlated these with the presence of distinct leukocyte subpopulations.


    Patients and methods
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Acknowledgements
 References
 
Patient details and tissue collection
Endometrial tissue was obtained from women using p-only contraception. A single endometrial biopsy was taken from women in Indonesia (n = 18) who had been using Norplant for between 3 weeks and 12 months, by Pipelle suction curette. In Edinburgh, UK, serial biopsies were collected with a Pipelle suction curette from women seeking to use a LNG-IUS and attending for contraceptive care at a specialist gynaecology/family planning clinic. Biopsies from these women were taken prior to insertion of the LNG-IUS [proliferative (days 5–12; n = 2) or secretory (days 15–24; n = 3) phase of the menstrual cycle] and at 1, 3, 6 and 12 months after insertion from those same women (n = 5). Other women with a LNG-IUS in situ (n = 5) consented to endometrial biopsy collection at the time of hysterectomy. A further cohort of normal control endometrial samples (menstrual days 1–4, proliferative days 5–12, late secretory days 25–28; n = 6) were collected from Australian women of mixed ethnicity undergoing minor gynaecological surgical procedures, who had not been using any form of hormonal contraceptive method for the past 3 months. Decidual tissue was also collected from women (n = 6) undergoing elective termination of pregnancy (amenorrhoea 6–8 weeks).

Written informed consent was obtained from all participants and ethical approval was obtained from the appropriate institutional ethics committees (96014B, LREC/2003/6/27 and LREC/2003/6/24). Stage of cycle of normal endometrial biopsies was calculated from last menstrual period and confirmed by histological assessment by a gynaecological pathologist, according to the criteria of Noyes et al. (1975)Go. Norplant endometria were histologically stained with haematoxylin and eosin and subclassified into atrophic, shedding or p-modified groups based on morphological appearance (n = 4–7 per morphological group), as described previously (Marsh et al., 1995Go; Vincent et al., 1999Go).

Endometrial samples were fixed in 10% buffered formalin overnight prior to washing in Tris-buffered saline (TBS) and routine histological processing to paraffin blocks. A portion of each tissue from the LNG-IUS study and from the normal control tissues was immersed in RNALater solution (Ambion, Austin, TX, USA) prior to snap freezing and storage at –80°C for RNA extraction.

Leukocyte subtype immunohistochemistry
Immunohistochemistry was performed on Norplant and control endometrial biopsies using antibodies specific for leukocyte common antigen (LCA CD45; Dako, Glostrup, Denmark), macrophages (CD68; Dako) and uNK cells (CD56; Zymed; CA,USA) as described previously (Critchley et al., 1998aGo). Briefly, 5-µm sections of formalin-fixed, paraffin-embedded tissues were dewaxed and rehydrated. Antigen retrieval by microwaving was necessary for CD56 immunohistochemistry, whilst trypsin digestion was performed prior to CD68 immunolocalization. Endogenous hydrogen peroxidase activity was quenched using 3% H2O2 in dH2O for 5 min at room temperature. Non-specific binding was prevented by preincubation of tissue sections with a non-immune block containing 10% non-immune horse serum (Sigma-Aldrich, Sydney, Australia), 2% normal human serum (in house) in TBS-T (Tween-20; 0.1%). Primary antibodies were applied overnight (17 ± 1 h) at 4°C diluted to 1 µg/ml in non-immune block. Negative controls were included for each tissue by the substitution of the primary antibody for a matching concentration of non-immunized mouse IgG. Following stringent washing with TBS-T (0.6%), detection of positive binding was performed by the sequential application of biotinylated horse anti-mouse IgG (1:200 in non-immune block; Vector Laboratories, Burlingame, CA, USA) and avidin–biotin–peroxidase conjugate (ABC-HRP; Dako), followed by the substrate diaminobenzidine (Dako) for 2–10 min. All samples for comparison were included in the same run to exclude inter-assay variability. Sections were counterstained with Harris’ haematoxylin (Sigma), dehydrated and mounted from Histosol with DPX.

Chemokine immunohistochemistry
Immunohistochemistry was performed to localize eight chemokines (MCP-3, 6Ckine, MDC, HCC-1, HCC-4, eotaxin, IL-8 and MIP-1{beta}) in formalin-fixed endometrial biopsies from users of Norplant and LNG-IUS, compared with representative normal endometrial biopsies from the times of known specific chemokine production: menstrual, proliferative and late secretory phases. Immunolocalization was performed using a monoclonal antibody specific for eotaxin (LeukoSite Inc., Cambridge, MA, USA) and polyclonal antibodies raised against human chemokine peptides (MIP-1{beta}: R&D Systems Inc., Minneapolis, MN, USA; all other antibodies were from Santa Cruz Biotechnology, Santa Cruz, CA, USA, and were as described previously: Zhang et al., 2000Go; Jones et al., 2004Go). Antibodies were applied at 0.5–4 µg/ml overnight at 4°C following microwave antigen retrieval. A similar protocol to that described above was used for eotaxin, but with visualization of positive localization with ABC-alkaline phosphatase (Dako), with new fushin substrate (Dako). All other antibodies were detected with sequential application of biotinylated horse anti-goat IgG and ABC-HRP. Antibodies were preabsorbed for 48 h at 4°C with a five-fold excess of specific chemokine peptide (Santa Cruz) to verify specificity. Again, all samples to be analysed were included in the same run to allow comparison of staining intensity.

Analysis of immunostaining
Abundance of leukocyte subtypes was analysed semi-quantitatively using a scoring system between 0 (no leukocytes) and 4 (highly abundant), as described previously (Vincent et al., 1999Go). Chemokine immunostaining was analysed semi-quantitatively by two independent observers blind to the identity of the tissue as previously described (Hannan et al., 2004Go; Jones et al., 2004Go). Staining intensity and heterogeneity in each of the endometrial compartments [epithelium, stroma (including decidualized stromal cells) and vasculature] was assessed and allocated a score between 0 and 3, where 0 = no stain and 3 = very intense staining. Data were statistically analysed using ANOVA with Tukey’s post-hoc test. A P-value of <0.05 was deemed significantly different.

RNA extraction and purification
Total RNA was extracted from LNG-IUS (n = 6) and control endometrial samples from the menstrual, proliferative and late secretory phases of the menstrual cycle and from early pregnancy (n = 3/stage) by homogenization in Trizol reagent (Qiagen, MD, USA), according to the manufacturer’s instructions, with the exception of an additional chloroform extraction step to minimize carryover of phenol into the precipitated RNA. All samples were treated with RNase-free DNase (Ambion) to remove contaminating genomic DNA. RNA samples were then analysed by spectrophotometry to determine RNA concentration, yield and purity. Any samples with ratios of A260/280 <1.7 or A230/280 >1 were purified through RNeasy spin columns (Qiagen, MD, USA) according to manufacturer’s instructions and thereafter reanalysed by spectrophotometry. Purified RNA (0.5 µg) was then run on a 1% agarose (Roche, Castle Hill, NSW, Australia) gel to ensure integrity of rRNA subunits, or analysed using the Agilent 2100 Bioanalyzer.

RNA concentrations were also analysed by Ribogreen fluorescence RNA assay (Molecular Probes, Eugene, OR, USA). RNA samples were diluted to ~40 ng/ml based on spectrophotometric readings and analysed in a 96-well plate by addition of Ribogreen fluorescent dye at a dilution of 1:500. A standard curve of serially diluted rRNA (Molecular Probes) between 0 and 80 ng/well was run on the same plate. This assay gave an intra-assay coefficient of variation (CV), as defined by the repeated analysis of a single RNA sample, of 3.3% (n = 16) and an inter-assay variation of 10.3% (n = 14).

Real time reverse transcription (RT)–PCR
mRNA expression levels of MCP-3, 6Ckine, MDC, HCC-1, HCC-4, IL-8 and MIP-1{beta} were analysed by real time PCR in LNG-IUS exposed endometrium compared with normal endometrium. Standards for real time PCR were generated by conventional PCR as described previously (Jones et al., 2004Go). RNA samples were reverse transcribed (RT) using AMV-RTase (Promega, Annandale, NSW, Aust.) and 100 ng random hexanucleotide primers (Amersham Biosciences, Piscataway, NJ, USA). The efficiency and reproducibility of the reverse transcription step was then tested by real time PCR analysis of 18S expression in the triplicate RTs. The %CV was calculated and samples outside 15% variability were excluded as outliers prior to pooling of triplicates, which were thereafter used for analysis of chemokine expression. This method overcomes the inherent high variability of this technique, ensuring the RTs for subsequent analysis are representative of the original mRNA sample.

Real time RT–PCR for chemokines was performed using a Roche Light Cycler (Roche) as described previously. Expression levels were quantified by comparison with standards of known concentration, using SYBR green fluorescence detection. Amplification conditions were carefully optimized for each chemokine as described previously (Jones et al., 2004Go). Fluorescence was monitored during cycling at the end of each elongation phase and analysis of expression was performed when amplified products were in the log-linear phase and parallel to the standards. At the end of each program, melting curve analysis was carried out to ensure specificity of the reaction products.

All samples to be compared were included within the same run (14.9% intra-assay variability). Mean expression levels were calculated from samples within each group and statistical analysis was performed using ANOVA with Tukey’s post-hoc test. A P-value of <0.05 was considered statistically significant. Data were not normalized for the expression of a housekeeping gene owing to the regulated expression of all known housekeeping genes examined in the endometrium (18S, cyclophilin, GAP-DH,{beta}-actin; data not shown). Instead RNA samples were meticulously purified and quantitated by three distinct methods as described above to ensure confidence in equal RNA loading and reverse transcription efficiency and reproducibility.


    Results
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Acknowledgements
 References
 
Leukocyte subpopulations in Norplant endometrium
Semi-quantitative analysis of numbers of CD45+ leukocytes demonstrated that leukocytes were highly abundant in shedding and p-modified groups, and significantly fewer were present in atrophic endometrium (P < 0.05 versus shedding and 0.01 versus p-modified) (Figure 1A). There were also differences in the distribution of leukocytes, with positively stained cells scattered throughout the stromal compartment in atrophic and p-modified endometria, whereas in shedding endometrium, distinct clusters of leukocytes were evident (Figure 2B). Intra-epithelial leukocytes were present in all Norplant groups, with increased density of leukocytes frequently observed in the immediate subepithelial region (not shown).



View larger version (16K):
[in this window]
[in a new window]
 
Figure 1. Immunohistochemical identification of leukocyte subpopulations in endometrium of Norplant users. (A) Semi-quantitative analysis of total leukocyte numbers using leukocyte common antigen (LCA, CD45), in the three morphological sub-groups of Norplant users: atrophic (Atr), shedding (Shed) and p-modified (P-mod). Each tissue section was allocated an immunostaining score between 0 and 4 based on the abundance of immunopositive cells. Significantly higher numbers of leukocytes are present in shedding (*P < 0.05) and p-modified endometria (**P < 0.01) compared with atrophic endometrium. (B) Macrophages were identified by surface expression of CD68. Significantly higher numbers of macrophages were present in p-modified endometrium compared witht other groups (*P < 0.05). (C) Uterine NK cells were immunostained with an antibody against CD56. Significantly higher numbers were present in p-modified endometria (**P < 0.01) compared with shedding and atrophic groups. All data are mean ± SEM (n = 4–7/group).

 


View larger version (127K):
[in this window]
[in a new window]
 
Figure 2. Representative photomicrographs showing immunohistochemical localization of leukocytes in endometrium from women using p-only contraceptives. (A–C) Immunohistochemical identification of leukocytes using leukocyte common antigen (LCA, CD45) in atrophic (Atr), shedding (Shed) and p-modified (P-mod) Norplant endometrium. Leukocytes were highly abundant in the latter two groups, with clusters of leukocytes evident in shedding endometrium. (D–F) Immunolocalization of CD68+ macrophages (Mac) in Atr, Shed and P-mod Norplant endometria. Few macrophages were present in Atr or Shed endometrium, but were more abundant in P-mod endometrium. Generally, macrophages were located in the subluminal zone and not in deeper regions. (G–I) uNKs were highly abundant in p-mod endometrium, whilst Atr and Shed endometrium contained fewer numbers of uNKs. Scale bars = 50 µm.

 

Analysis of leukocyte subtypes revealed that generally very few CD68-positive macrophages were present in shedding and atrophic endometrium, but there were significantly higher numbers in p-modified samples (Figures 1B and 2D–F). These tended to be in the subepithelial zone (Figure 2D–F), with few positive cells in deeper tissue. CD56 identifies uNK cells, and these were highly abundant in p-modified tissues (P < 0.01 versus atrophic and shedding groups) (Figures 1C and 2G–I), comparable to numbers seen in early pregnancy decidua (not shown). In shedding and atrophic groups there were significantly fewer uNK cells (Figure 2G and H). A summary of this data, together with existing information in the literature, is provided in Table I, clearly illustrating the distinct differences in leukocyte subpopulations between p-only groups.


View this table:
[in this window]
[in a new window]
 
Table I. Summary of leukocyte subpopulations present in endometrium from different groups of p-only contraceptive users

 

Chemokines in Norplant endometrium
All chemokines examined (Table II) were detected in endometrium from women using Norplant, although significant differences were seen in chemokine production between different Norplant groups and in comparison with normal endometrium.


View this table:
[in this window]
[in a new window]
 
Table II. Chemokine ligand-receptor pairings, demonstrating potential actions of the selected chemokines in recruitment of leukocyte subtypes to human endometrium

 

In non-pregnant endometrium, the luminal and glandular epithelia are the predominant sites of chemokine synthesis. MDC, HCC-1, MCP-3 and eotaxin were highly expressed in epithelial cells in endometria from all women using Norplant, with levels equivalent to maximal immunostaining levels seen during the normal menstrual cycle (Figures 3 and Figures 4A). Very intense staining was identified for MDC and HCC-1 in epithelial cells of p-modified endometrium (Figure 3). In contrast, 6Ckine, MIP-1{beta} and HCC-4 were present in lower levels in epithelial cells from Norplant tissues than normal control endometrial samples, with a similar pattern in all three groups (Figure 3). IL-8 followed a distinct pattern, with elevated epithelial immunostaining in the shedding Norplant group comparable to levels seen in normal endometrium perimenstrually (Figure 3). Immunostaining was low or absent in atrophic and p-modified endometria. Representative photomicrographs of chemokines are shown in Figure 4, illustrating the faint immunostaining for most chemokines (eoxtain is a notable exception) in atrophic endometrium (Figure 4B); intense staining for MCP-3 and IL-8 in areas of breakdown in shedding endometrium (Figure 4C); and reduced staining for HCC-4 and MIP-1{beta} compared with proliferative controls in p-modified endometrium, whilst 6Ckine, MCP-3 and MDC are more highly expressed, consistent with secretory levels (Figure 4D).



View larger version (29K):
[in this window]
[in a new window]
 
Figure 3. Immunohistochemical localization of chemokines in endometria of Norplant users. Semi-quantitative assessment of immunostaining was performed based on the intensity of staining in epithelial and stromal (including decidualized stroma) cells. Each chemokine was assessed in the three morphological groups of Norplant endometria (Atr = atrophic; Shed = shedding; P-mod = p-modified), compared with normal endometrium from the menstrual (ME), proliferative (PR) and late secretory (LS) phases of the menstrual cycle. In each case chemokine immunostaining varied in cycling endometria consistent with previous observations (Jones et al., 2004Go) (significant differences in immunostaining not marked for clarity). Significant differences in epithelial immunostaining levels between linked groups are represented by asterisks (*P < 0.05 and **P < 0.01). In Atr and Shed groups all chemokines, except for eotaxin, were absent from stromal cells, whilst immunostaining for MDC, HCC-1, MCP-3, 6Ckine and MIP-1{beta} in decidualized stroma in P-mod endometrium was strongly positive (**P < 0.01) compared with other Norplant groups and normal menstrual and proliferative endometria. *P < 0.05; **P < 0.01. Data are mean ± SEM of n = 5–8.

 


View larger version (103K):
[in this window]
[in a new window]
 
Figure 4. Immunohistochemical localization of chemokines in endometrium of women using p-only contraceptives. (A) Representative photomicrographs of chemokine immunostaining in control endometrium. HCC-4 and MIP-1{beta} are maximal during the proliferative phase; 6Ckine, MCP-3 and MDC are maximal in the secretory phase. (B) In atrophic endometrium from women using Norplant, chemokine immunostaining is faint and predominantly in epithelial cells. (C) Similarly, chemokine immunostaining was generally faint in shedding endometrium from women using Norplant. However, IL-8 was markedly up-regulated in epithelial glands and infiltrating leukocytes (arrowheads). Other chemokines, e.g. MCP-3 was strongly up-regulated in regions exhibiting signs of breakdown. Eotaxin was present in endometrial vasculature in all Norplant groups. (D) Proliferative associated chemokines (e.g. HCC-4 and MIP-1{beta}) and menstrual-associated chemokines (IL-8) were lowly abundant in p-modified Norplant endometrium. Other chemokines (6Ckine, MCP-3, MDC and HCC-1) were up-regulated in epithelial and decidual cells, and in infiltrating leukocytes (arrowheads). (E) Similar chemokine immunostaining patterns were observed in the endometrium of women using LNG-IUS, with increased immunostaining intensity in the highly decidualized endometrium. Representative photomicrographs shown for HCC-1, 6Ckine, MCP-3 and MDC. Scale bars = 50 µm.

 
With the exception of eotaxin, chemokines were not produced by stromal cells in atrophic and shedding groups. However, the highly decidualized stromal cells in p-modified endometria were strongly positive for MDC, MCP-3, HCC-1 and 6Ckine (Figures 3 and Figures 4D). Faint staining was detected for HCC-4, MIP-1{beta}, eotaxin and IL-8 in decidual cells.

Individual stromal cells, exhibiting histological characteristics of leukocytes, stained intensely for a number of the chemokines: MDC, HCC-4, eotaxin, MIP-1{beta}, 6Ckine and IL-8. MDC- and MCP-3-positive cells were particularly abundant in p-modified endometrium (Figure 4D), whilst those producing IL-8 were abundant in breakdown areas in shedding tissues (Figure 4C, arrowheads). Occasional immunoreactivity for chemokines was detected in vascular endothelial cells of Norplant users, as illustrated for eoxtain (Figure 4C).

Chemokines in LNG-IUS endometrium
Chemokine immunostaining after insertion of LNG-IUS was consistent with the highly decidualized nature of the tissue and closely resembled that seen in the p-modified Norplant endometrium. Most marked was the strong up-regulation of many chemokines (HCC-1, 6Ckine, MCP-3 and MDC) in the decidualized stroma (Figures 4E and 5). However, HCC-4 and eotaxin exhibited faint immunostaining (Figure 5), whilst MIP-1{beta} was markedly absent (data not shown). Epithelial staining remained high post-insertion for MDC, HCC-1 and MCP-3, whilst again 6Ckine and HCC-4 production by epithelial glands was reduced with continued exposure to intrauterine levonorgestrel, although this failed to reach significance (Figure 5). No significant changes were seen with continued exposure to levonorgestrel, except that stromal chemokine immunostaining was limited to those cells that had undergone decidualization, and thus the staining became more homogeneous as the degree of decidualization increased. It is important to note that while some samples were collected during times of bleeding there was no correlation with bleeding and chemokine expression compared with those samples collected during non-bleeding episodes.



View larger version (34K):
[in this window]
[in a new window]
 
Figure 5. Immunohistochemical localization of chemokines in endometrium from LNG-IUS users. Semi-quantitative assessment of immunostaining was performed based on the intensity of staining in epithelial and stromal (including decidualized stroma) cells. Each chemokine was assessed before insertion of the LNG-IUS, in the proliferative (PR) or secretory (SEC) phase, and at 1, 3 and 6 months postinsertion. For every chemokine, stromal staining was significantly elevated with respect to preinsertion proliferative levels (P < 0.05). Data are mean ± SEM of n = 3–5.

 

Real time RT–PCR
mRNA expression levels of seven of the chemokines were analysed by real time RT-PCR in endometrium from LNG-IUS users, and compared with normal non-pregnant and pregnant endometrial levels. Most of the chemokines examined were expressed at elevated levels after insertion of LNG-IUS compared with normal cycling endometrium (Figure 6) and were generally consistent with those seen in the decidua of early pregnancy, although HCC-1 was slightly reduced (in keeping with the lower staining intensity in glandular epithelium seen in these tissues; Figure 5). MCP-3 was highly elevated in the LNG-IUS users, consistent with the intense immunostaining seen after levonorgestrel exposure (Figure 5). IL-8 mRNA was only detected in menstrual phase tissue, with a low degree of expression in endometrium from LNG-IUS users. These expression patterns are consistent with and support the protein expression data from immunohistochemical studies.



View larger version (25K):
[in this window]
[in a new window]
 
Figure 6. Real time RT–PCR quantitative analysis of chemokine mRNA expression in endometrium from women before and after insertion of LNG-IUS. Chemokine expression levels were analysed in normal endometrium during the menstrual (ME), late proliferative (LP) and late secretory (LS) phases, and also in decidua (DEC) from early pregnancy (n = 3/group); and compared with expression levels following insertion of LNG-IUS (n = 5). Chemokine mRNA levels were consistent with expression levels in decidua from early pregnancy, except for HCC-1, where it was significantly down-regulated, and MCP-3 and IL-8, where levels in levonorgestrel-treated endometrium were significantly higher (*P < 0.05). Significant changes between normal women are not shown for clarity, but are consistent with previously published data (Jones et al., 2004Go).

 


    Discussion
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Acknowledgements
 References
 
In this study we present data on the expression of chemokines in endometrium of women using two methods of p-only contraception and correlate this to the subsets of leukocytes present. Whilst we observe specific expression patterns for the individual chemokines in the different endometrial types, we see an overall pattern of strong expression of chemokines selective for the leukocytes present.

P-only contraceptive methods induce differential effects on endometrial morphology and the different morphologies contain different populations of immune cells. Norplant-treated endometrium with an atrophic appearance contains very few leukocytes. However, there is a large leukocyte infiltrate evident in the shedding group, which previous studies show to be predominantly neutrophils and eosinophils (Vincent et al., 1999), and also in the p-modified tissues with an increase in macrophages and uNK cells, as we have shown here. In LNG-IUS users where morphology is highly decidualized (p-modified), macrophages and uNK cells also predominate (Critchley et al., 1998a), as seen normally in early pregnancy decidual samples, where uNK cells comprise 40% of all cells in the stromal compartment (Moffett-King, 2002Go). Consistent with these leukocyte patterns, we present evidence for different repertoires of chemokines in each of the three different groups of Norplant users, with LNG-IUS-exposed endometria closely resembling that of the p-modified Norplant endometrium.

In normal endometrium chemokines are produced predominantly by epithelial cells, with cyclical variations in expression levels. We have previously identified groups of chemokines that are up-regulated in specific stages of the cycle: menstrual-associated (IL-8), proliferative phase-associated (MIP-1{beta} and HCC-4), decidual-associated (MDC and 6Ckine) and constantly present (MCP3, FKN and HCC-1) (Jones et al., 2004Go). Expression of most chemokines examined in this study (HCC-1, MCP-3, eotaxin and 6Ckine) was similar in the shedding and atrophic groups and was consistent in localization and expression levels with normal non-decidualized endometrium. IL-8 was strikingly elevated in the shedding Norplant group, consistent with the high levels seen perimenstrually, supporting a role for IL-8 in the recruitment of the large neutrophil infiltrate peculiar to this Norplant group and absent from the p-modified tissue of some Norplant and all LNG-IUS users (Jones and Critchley, 2000Go). The marked reduction of the proliferative-associated chemokines, HCC-4 and MIP-1{beta} (Jones et al., 2004Go), in the atrophic quiescent endometrium compared with proliferative endometrium reinforces a role for these chemokines in the regeneration and proliferation of the normal endometrium.

In the p-modified Norplant and LNG-IUS endometrium, epithelial immunostaining levels were consistent with those seen in the mid-late secretory phase of the cycle, with very strong staining for HCC-1, MDC, MCP-3 and eotaxin, but low expression of HCC-4 and MIP-1{beta}. The former group are all chemoattractants for the leukocyte subsets abundant in p-modified Norplant and LNG-IUS tissues (macrophages, uNKS) (Table I and II), and confirms an important role for the glands in positioning and activation of leukocytes in the decidualized endometrium. However, in marked contrast to normal decidualized endometrium, 6Ckine and IL-8 were selectively down-regulated in epithelial cells in p-modified endometria from women using both Norplant and LNG-IUS. Such down-regulation has been reported previously for fractalkine (Hannan et al., 2004Go) and indicates either differential regulation of these three chemokines in artificially induced or extensive decidualization, or differential effects of levonorgestrel and natural progesterone. One possible explanation is that levonorgestrel is a highly androgenic progestin (Kloosterboer et al., 1988Go) and may indirectly influence chemokine expression through interactions with stromal androgen receptor (Mertens et al., 2001Go).

The most marked finding in these studies relates to stromal production of chemokines, and again we confirm that stromal cells are not a source of chemokines until they undergo decidualization, when they become the predominant production site in the endometrium (Jones et al., 2004Go). This is entirely consistent with the abundance of uNKs and macrophages in the decidua and in the highly decidualized LNG-IUS and p-modified Norplant endometrium, and with the strong expression of their chemoattractants (MDC, HCC-1, MCP-3, MIP-1{beta}) (Bacon et al., 2002Go; Rabin, 2003Go). Decidual-derived chemokines have been implicated in the recruitment of uNKs to the implantation site (Jones et al., 2004Go), and also for specific targeted migration of a cytotrophoblast to decidual vessels (Sato et al., 2003Go; Drake et al., 2004Go). The high chemokine production and uNK cell abundance in these groups of p-only contraceptive users is likely to be unrelated to BTB, but rather, a consequence of extensive decidualization.

This work, and much other research in the literature, is driven by the hypothesis that immune cells (particularly inflammatory neutrophils) are key effector cells initiating menstruation and BTB. Their absence in the decidualized endometrial groups of p-only users therefore seems to vilify this hypothesis. However, most studies examining bleeding patterns in Norplant users have not distinguished between the different morphological groups. One study from our laboratory reported that women with a p-modified endometria experience almost half the number of bleeding days (24 days) than the shedding or atrophic groups (37 and 40, respectively), and more than double the number of non-bleeding days (16 versus 5.8 and 6.4) in the 90-day period prior to endometrial biopsy (Vincent et al., 1999Go). This finding, however, failed to reach significance when the statistical analysis was corrected for body mass index (BMI), as the women with a p-modified endometrium had an overall higher BMI (although still in the normal weight range). This trend towards less BTB in women with a decidualized endometrium than in those with an atrophic or shedding endometrium, is consistent with the leukocyte profile of this tissue, and raises the suggestion that achieving a decidualized endometrium, complete with non-inflammatory chemokines and leukocytes, is desirable to reduce BTB. Women using LNG-IUS do experience BTB, although there is a notable improvement in bleeding patterns after the first 6 months of use. There are likely to be different mechanisms involved in endometrial fragility between Norplant and LNG-IUS users owing to the considerable difference in endometrial exposure to progestin, leading to distinct imbalances in the endometrial microenvironment. Indeed, local delivery of progestin results in suppression of progesterone receptors in the first 6 months of use (Critchley et al., 1998aGo), whilst Norplant endometrium remains strongly responsive to steroids (Critchley et al., 1993Go), indicating major differences in the functional downstream effects of progestin in these groups of women.

Certain chemokines are known to be progesterone regulated (Kelly et al., 1994Go; 1997Go), and generally inflammatory chemokines are inhibited by glucocorticoids (Miyamasu et al., 1998Go). From the differential expression of chemokines between the three groups of Norplant endometrium it seems unlikely that levonorgestrel is exerting direct effects on chemokine expression. This is supported by the fact that in women using LNG-IUS there is an almost total down-regulation of nuclear progesterone receptor (PR) in epithelial cells (Critchley et al., 1998bGo), although this could be explained by the presence of membrane bound PR (Zhu et al., 2003Go). Alternatively, the progestin effect on epithelial cells could be mediated through paracrine interactions with underlying stromal cells, which also possess glucocorticoid and androgen receptors (Bamberger et al., 2001Go; Mertens et al., 2001Go), with which the high concentrations of levonorgestrel may interact. Epithelial and decidual chemokine production during the normal secretory phase and in women using p-only contraceptives is therefore more likely to be stimulated indirectly by progesterone-regulated locally produced factors.

In this study, we have not identified a single chemokine that is responsible for BTB, leading to the conclusion that BTB is most likely caused by very focal perturbations in the endometrial environment leading to loss of tissue integrity in discrete focal areas. This is supported by the hysteroscopic visualization of dilated bleeding vessels (Hickey et al., 2000Go), and histologically by the focal clusters of leukocyte infiltrate in breaking-down tissue, surrounded by unaffected atrophic endometrium (Clark et al., 1996Go; Song et al., 1996Go; Vincent et al., 1999Go). Selective leukocyte accumulation and activation in breakdown foci must be caused by focal up-regulation of specific chemokines. The chemokines examined in this study were selected on the basis of their abundant expression in normal endometrium, from global gene array studies (Jones et al., 2004Go), which would not identify chemokines that are lowly and focally expressed. Nevertheless, closer examination of normal endometrium using immunohistochemistry identified local up-regulation of neutrophil and macrophage chemokines (IL-8, FKN, MDC, 6Ckine) in vascular endothelium, only in the immediate premenstrual phase (Jones et al., 1997Go; 2004Go; Hannan et al., 2004Go). Identification of the critical effector cells and mediators involved in BTB is likely to be possible only by detailed analysis of actual bleeding sites, as described by Lockwood et al. (2000)Go, correlating leukocyte infiltrate, localized chemokine expression, blood vessel abnormalities and progesterone receptor content in the areas known to be breaking down.

To summarize, the current study verifies that leukocytes are abundant in the endometrium of women using p-only contraceptives, but that the leukocyte profile of the endometrium differs significantly between subgroups of contraceptive users. In keeping with the elevated numbers of leukocytes present, we show that chemokine production is overall elevated and comparable to levels seen during maximal leukocyte recruitment in normal endometrium. Furthermore, we see significant differences in chemokine expression patterns in the different groups of p-only users, consistent with endometrial morphology and leukocyte content. Importantly, these findings reinforce that these contraceptives not only differentially affect endometrial morphology, but also endometrial physiology. Overall, these data demonstrate chemokine expression is dysregulated in p-exposed endometria, reinforcing a critical role for chemokines in abnormal leukocyte recruitment in women using p-only contraceptives.


    Acknowledgements
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Acknowledgements
 References
 
We acknowledge Associate Professor Peter Rogers and Dr Biran Affandi for providing the Norplant endometrial biopsies, and Dr Lishi Udabage for technical assistance with real time PCR. This investigation received financial support from the UND/UNFPA/WHO/World Bank Special Programme of Research, Development, and Research Training in Human Reproduction, World Health Organization. LNG-IUS sample collection made possible from funding support to HODC from The Wellcome Trust (grant No. 044744) and the MRC Programme (grant G0000066). L.A.S. is supported by the NH&MRC of Australia (#143798).


    References
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Acknowledgements
 References
 
Akiyama M, Okabe H, Takakura K, Fujiyama Y and Noda Y (1999) Expression of macrophage inflammatory protein-1alpha (MIP-1alpha) in human endometrium throughout the menstrual cycle. Br J Obstet Gynaecol 106,725–730.[ISI][Medline]

Bacon K, Baggiolini M, Broxmeyer H, Horuk R, Lindley I, Mantovani A, Maysushima K, Murphy P, Nomiyama H, Oppenheim J et al. (2002) Chemokine/chemokine receptor nomenclature. J Interferon Cytokine Res 22,1067–1068.[CrossRef][ISI][Medline]

Bamberger AM, Milde-Langosch K, Loning T and Bamberger CM (2001) The glucocorticoid receptor is specifically expressed in the stromal compartment of the human endometrium. J Clin Endocrinol Metab 86,5071–5074.[Abstract/Free Full Text]

Belsey EM, d’Arcangues C and Carlson N (1988) Determinants of menstrual bleeding patterns among women using natural and hormonal methods of contraception. II. The influence of individual characteristics. Contraception 38,243–257.[CrossRef][ISI][Medline]

Brache V, Faundes A and Alvarez F (2003) Risk–benefit effects of implantable contraceptives in women. Expert Opin Drug Saf 2,321–332.[CrossRef][Medline]

Bulmer JN, Longfellow M and Ritson A (1991) Leukocytes and resident blood cells in endometrium. Ann N Y Acad Sci 622,57–68.[Medline]

Clark DA, Wang S, Rogers P, Vince G and Affandi B (1996) Endometrial lymphomyeloid cells in abnormal uterine bleeding due to levonorgestrel (Norplant). Hum Reprod 11,1438–1444.[Abstract/Free Full Text]

Critchley HO, Bailey DA, Au CL, Affandi B and Rogers PA (1993) Immunohistochemical sex steroid receptor distribution in endometrium from long-term subdermal levonorgestrel users and during the normal menstrual cycle. Hum Reprod 8,1632–1639.[Abstract]

Critchley HO, Wang H, Jones RL, Kelly RW, Drudy TA, Gebbie AE, Buckley CH, McNeilly AS and Glasier AF (1998a) Morphological and functional features of endometrial decidualization following long-term intrauterine levonorgestrel delivery. Hum Reprod, 13,1218–1224.[Abstract]

Critchley HO, Wang H, Kelly RW, Gebbie AE and Glasier AF (1998b) Progestin receptor isoforms and prostaglandin dehydrogenase in the endometrium of women using a levonorgestrel-releasing intrauterine system. Hum Reprod 13,1210–1217.[Abstract]

Drake PM, Red-Horse K and Fisher SJ (2004) Reciprocal chemokine receptor and ligand expression in the human placenta: implications for cytotrophoblast differentiation. Dev Dyn 229,877–885.[CrossRef][ISI][Medline]

Hannan NJ, Jones RL, Critchley HO, Kovacs GJ, Rogers PA, Affandi B and Salamonsen LA (2004) Coexpression of fractalkine and its receptor in normal human endometrium and in endometrium from users of progestin-only contraception supports a role for fractalkine in leukocyte recruitment and endometrial remodeling. J Clin Endocrinol Metab 89,6119–6129.[Abstract/Free Full Text]

Hickey, M, Dwarte D and Fraser IS (2000) Superficial endometrial vascular fragility in Norplant users and in women with ovulatory dysfunctional uterine bleeding. Hum Reprod 15,1509–1514.[Abstract/Free Full Text]

Hornung D, Ryan IP, Chao VA, Vigne JL, Schriock ED and 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.[Abstract/Free Full Text]

Jones RL and Critchley HO (2000) Morphological and functional changes in human endometrium following intrauterine levonorgestrel delivery. Hum Reprod 15 (Suppl 3), 162–172.[Medline]

Jones RL, Kelly RW and Critchley HO (1997) Chemokine and cyclooxygenase-2 expression in human endometrium coincides with leukocyte accumulation. Hum Reprod 12,1300–1306.[CrossRef][ISI][Medline]

Jones RL, Hannan NJ, Kaitu’u TJ, Zhang J and Salamonsen LA (2004) Identification of chemokines important for leukocyte recruitment to the human endometrium at the times of embryo implantation and menstruation. J Clin Endocrinol Metab 89,6155–6167.[Abstract/Free Full Text]

Kelly RW, Illingworth P, Baldie G, Leask R, Brouwer S and Calder AA (1994) Progesterone control of interleukin-8 production in endometrium and chorio-decidual cells underlines the role of the neutrophil in menstruation and parturition. Hum Reprod 9,253–258.[Abstract]

Kelly RW, Carr GG and Riley SC (1997) The inhibition of synthesis of a beta-chemokine, monocyte chemotactic protein-1 (MCP-1) by progesterone. Biochem Biophys Res Commun 239,557–561.[CrossRef][ISI][Medline]

Kloosterboer HJ, Vonk-Noordegraaf CA and Turpijn EW (1988) Selectivity in progesterone and androgen receptor binding of progestagens used in oral contraceptives. Contraception 38,325–332.[CrossRef][ISI][Medline]

Lockwood CJ, Runic R, Wan L, Krikun G, Demopolous R and Schatz F (2000) The role of tissue factor in regulating endometrial haemostasis: implications for progestin-only contraception. Hum Reprod 15 (Suppl 3), 144–151.[Medline]

Luster AD (1998) Chemokines—chemotactic cytokines that mediate inflammation. N Engl J Med 338,436–445.[Free Full Text]

Marsh MM, Butt AR, Riley SC, Rogers PA, Susil B, Affandi B, Findlay JK and Salamonsen LA (1995) Immunolocalization of endothelin and neutral endopeptidase in the endometrium of users of subdermally implanted levonorgestrel (Norplant). Hum Reprod 10,2584–2589.[Abstract]

Mascarenhas L, van Beek A, Bennink HC and Newton J (1998) A 2-year comparative study of endometrial histology and cervical cytology of contraceptive implant users in Birmingham, UK. Hum Reprod 13,3057–3060.[Abstract]

Mertens HJ, Heineman MJ, Theunissen PH, de Jong FH and Evers JL (2001) Androgen, estrogen and progesterone receptor expression in the human uterus during the menstrual cycle. Eur J Obstet Gynecol Reprod Biol 98,58–65.[CrossRef][ISI][Medline]

Miyamasu M, Misaki Y, Izumi S, Takaishi T, Morita Y, Nakamura H, Matsushima K, Kasahara T and Hirai K (1998) Glucocorticoids inhibit chemokine generation by human eosinophils. J Allergy Clin Immunol 101,75–83.[ISI][Medline]

Moffett-King A (2002) Natural killer cells and pregnancy. Nat Rev Immunol 2,656–663.[CrossRef][ISI][Medline]

Nilsson CG, Luukkainen T and Arko H (1978) Endometrial morphology of women using a D-norgestrel-releasing intrauterine device. Fertil Steril 29,397–401.[ISI][Medline]

Noyes RW, Hertig AT and Rock J (1975) Dating the endometrial biopsy. Am J Obstet Gynecol 122,262–263.[Medline]

Rabin R (2003) CC, C and CX3C chemokines. In Henry HL and Norman AW (eds) Encyclopedia of Hormones. Elsevier Science, Oxford, UK, pp. 255–263.

Salamonsen LA and Woolley DE (1999) Menstruation: induction by matrix metalloproteinases and inflammatory cells. J Reprod Immunol 44,1–27.[CrossRef][ISI][Medline]

Salamonsen LA and Lathbury LJ (2000) Endometrial leukocytes and menstruation. Hum Reprod Update 6,16–27.[Abstract/Free Full Text]

Sato Y, Higuchi T, Yoshioka S, Tatsumi K, Fujiwara H and Fujii S (2003) Trophoblasts acquire a chemokine receptor, CCR1, as they differentiate towards invasive phenotype. Development 130,5519–5532.[Abstract/Free Full Text]

Silverberg SG, Haukkamaa M, Arko H, Nilsson CG and Luukkainen T (1986) Endometrial morphology during long-term use of levonorgestrel-releasing intrauterine devices. Int J Gynecol Pathol 5,235–241.[ISI][Medline]

Song, JY, Russell P, Markham R, Manconi F and Fraser IS (1996) Effects of high dose progestogens on white cells and necrosis in human endometrium. Hum Reprod 11,1713–1718.[Abstract]

Vincent AJ and Salamonsen LA (2000) The role of matrix metalloproteinases and leukocytes in abnormal uterine bleeding associated with progestin-only contraceptives. Hum Reprod 15 (Suppl 3), 135–143.[Abstract/Free Full Text]

Vincent AJ, Malakooti N, Zhang J, Rogers PA, Affandi B and Salamonsen LA (1999) Endometrial breakdown in women using Norplant is associated with migratory cells expressing matrix metalloproteinase-9 (gelatinase B). Hum Reprod 14,807–815.[Abstract/Free Full Text]

Zhang J, Lathbury LJ and Salamonsen LA (2000) Expression of the chemokine eotaxin and its receptor, CCR3, in human endometrium. Biol Reprod 62,404–411.[Abstract/Free Full Text]

Zhu Y, Bond J and Thomas P (2003) Identification, classification, and partial characterization of genes in humans and other vertebrates homologous to a fish membrane progestin receptor. Proc Natl Acad Sci USA 100,2237–2242.[Abstract/Free Full Text]

Submitted on January 25, 2005; resubmitted on March 28, 2005; accepted on May 11, 2005.