Hum. Reprod.

Advance Access published April 26, 2012

Human Reproduction, Vol.0, No.0 pp. 1 9, 2012 doi:10.1093/humrep/des134

ORIGINAL ARTICLE Early pregnancy

Expression of leukaemia inhibitory factor and interleukin 15 in endometrium of women with recurrent implantation failure after IVF; correlation with the number of endometrial natural killer cells
Downloaded from http://humrep.oxfordjournals.org/ by guest on April 27, 2012

N. Mariee 1, T.C. Li 2, and S.M. Laird 3,*

1 Academic Unit of Reproductive and Developmental Medicine, University of Shefeld, Jessop Wing, Shefeld Teaching Hospitals, Tree Root Walk, Shefeld S10 2SF, UK 2Jessop Wing, Shefeld Teaching Hospitals, Tree Root Walk, Shefeld S10 2SF, UK 3Biomedical Research Centre, Shefeld Hallam University, City Campus, Shefeld S1 1WB, UK

*Correspondence address. E-mail: [email protected]

Submitted on September 2, 2011; resubmitted on March 15, 2012; accepted on March 21, 2012

background: Several studies have suggested that endometrial interleukin 15 (IL-15) and the leukaemia inhibitory factor (LIF) may be
important in embryo implantation. IL-15 is postulated to play a role in the control of uterine natural killer (uNK) cell proliferation and function, and uNK cells are also known to play a role in implantation. The aims of this study was to (1) compare endometrial levels of IL-15 and the LIF in women with recurrent implantation failure (RIF) after IVF with those in fertile women (controls) and (2) examine the relation of IL15 and LIF levels to the uNK cell number.

methods: We investigated IL-15 and LIF in precisely timed endometrial biopsies (days LH + 7-LH + 9, where the day of the LH surge is
LH + 0) obtained from control women (n 15) and women with RIF (n 45) by immunohistochemistry. A semi-quantitative analysis was performed by the H-score analysis of staining intensity in the stroma, glandular epithelium and luminal epithelium, separately. We also correlated expression of LIF and IL15 with uNK cell numbers (obtained in an earlier study of the same samples).

results: The quantity of the LIF protein in endometrial glandular epithelium in women with RIF [median and range; 179 (70 365)] was lower (P 0.01) than in control women [median and range; 247 (120287)]. In contrast, the level of the IL-15 protein in the stroma in women with RIF [median and range; 90 (0 175)] was higher (P 0.009) than in control women [median and range; 60 (15 150)]. There was a signicant correlation between the uNK cell number and stromal expression of IL-15 (r 0.427, P 0.001). No correlation between the LIF expression in any compartment and the uNK cell number was seen. conclusions: The results show an altered expression of LIF and IL-15 in the endometrium of women with RIF. Despite the limitation of not identifying uNK cells by phenotypic markers, the correlation between the uNK cell number and the stromal cell IL-15 suggests that IL-15 may play a role in the control of endometrial uNK cell function or proliferation.
Key words: endometrium / recurrent implantation failure / leukaemia inhibitory factor / interleukin 15 / natural killer cells

Introduction
Successful implantation of the blastocyst into the endometrium is essential for reproduction. Implantation occurs during a very short period in the secretory phase of the menstrual cycle called the window of implantation. During this time, the embryo invades the

endometrium and gains access to the maternal blood supply. It is a complex process which requires synchronization of events in the developing embryo and receptive endometrium, and involves factors such as immune cells, cytokines, chemokines, growth factors and adhesion molecules (van Mourik et al., 2009). Various different immune cells and cytokines are postulated to participate

& The Author 2012. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.
For Permissions, please email: [email protected]

2
in successful implantation, including uterine natural killer (uNK) cells, leukaemia inhibitory factor (LIF) and interleukin 15 (IL-15). Unregulated expression of these factors has been associated with several endometrial disorders including defective endometrial receptivity, endometriosis and recurrent miscarriage (Laird et al., 2003; Dimitriadis et al., 2005). LIF is a multifunction glycoprotein that has effects on many different tissues including the reproductive system (Kimber, 2005). Studies in mice have shown that LIF plays an important role in implantation. LIF knock-out female mice are infertile and their embryos do not implant. However, the same embryos when transferred to the wildtype mice do implant, suggesting that the defect is in the endometrium and not the embryo (Stewart et al., 1992). The role of LIF in human implantation is less clear. LIF is expressed in the human endometrium and the Fallopian tube, with increased expression in epithelial cells compared with stromal cells in both tissues. It is expressed in a menstrual cycle-dependent manner with maximum expression on Days 19 25 of the menstrual cycle, which is the time of blastocyst implantation (Keltz et al., 1996; Arici et al., 1999). IL-15 is expressed in the human endometrium, decidua and placenta (Kitaya et al., 2000; Verma et al., 2000) and is expressed in both stromal and epithelial cells of the endometrium. Its expression in epithelial cells is higher than in stromal cells during most of the cycle but increased expression is seen in stromal cells after decidualization (Kitaya et al., 2000; Chegini et al., 2003). uNK cells are the major leucocytes present in the endometrium in the secretory phase of the cycle and their numbers further increase in the rst trimester decidua (Bulmer et al., 1991). Several studies have shown increased numbers of endometrial uNK cells in women with recurrent miscarriage (Clifford et al., 1999; Quenby et al., 1999; Tuckerman et al., 2007). Other studies have shown high numbers of endometrial uNK cells in women with recurrent implantation failure (RIF) (Ledee-Bataille et al., 2004; Tuckerman et al., 2010). In vitro studies have shown that IL-15 is a strong inducer of proliferation of NK cells and is required for the activation and promotion of cytotoxity and cytokine production in NK cells (Carson et al., 1994). It is also postulated that IL-15 may be important in the recruitment of peripheral blood NK cells to the endometrium and their differentiation into uNK cells (Croy et al., 2003; Manaster et al., 2008). One of the limiting steps in IVF is the failure of embryo implantation. This may be caused by embryonic or endometrial factors or a combination of the two. For women who experience RIF with good quality embryos, it is likely that an endometrial factor may be the cause. Changes in expression of endometrial factors in women with RIF may, therefore, indicate their importance in the implantation process. To the best of our knowledge, no studies have examined the expression of LIF, IL-15 and uNK cell numbers in the same sample of precisely timed endometrial tissue from women with RIF after IVF. Therefore, the aims of this study were to compare the levels of LIF and IL-15 protein in precisely timed endometrial biopsies taken from women with RIF and from control women with proven fertility, and to correlate the levels of LIF and IL-15 with the number of uNK cells in the same tissue sample. The data for the numbers of uNK cells used in this study for correlations with the LIF and IL15 are the same as those reported in a previous study (Tuckerman et al., 2010).

Mariee et al.

Materials and Methods

Patients and endometrial samples
Endometrial biopsies for this study were obtained from women attending the Jessop Wing, Shefeld Teaching Hospitals, UK. All samples were collected with the informed consent of the patients and approval from the ethics committee was obtained for the study. Endometrial biopsies from control women (n 15) were obtained from women with proven fertility attending the gynaecological outpatient clinic. All these women had regular menstrual cycles of between 25 and 35 days, had not used any hormonal treatment for at least 2 months before the biopsy, had given birth at least once and had no history of infertility. Endometrial biopsies from women with RIF after IVF (n 45) were obtained from women attending the Assisted Conception Unit, Jessop Wing, Shefeld, UK. RIF was dened as the failure of good quality embryos to implant after either at least three cycles of fresh IVF, or two cycles of fresh IVF plus two cycles in which frozen embryos were replaced. All women had a good hormonal reserve (FSH on Day 5 of the cycle ,10 mIU/ml) and all embryos were at least of the four-cell stage by Day 2. The median age of the control group was 37 years (range 29 40 years), whereas the median age of the RIF group was 35 years (range 26 40 years). Timed biopsies were obtained from both groups of women. A daily measurement of LH in urine from Day 9 of the menstrual cycle onwards was used to identify the LH surge (designated LH + 0). All endometrial biopsies were collected with a Pipelle sampler between days LH + 7 and LH + 9. Biopsies were immediately xed overnight in formalin and then embedded in wax using an automated process.

Downloaded from http://humrep.oxfordjournals.org/ by guest on April 27, 2012

Hormone assays
Peripheral blood samples were obtained from RIF and control women during the early follicular phase and at the time of biopsy collection. Serum was immediately separated from cells by centrifugation and stored at 2208C until assayed. LH, FSH and estradiol were measured during the follicular phase (Day 5), and progesterone was measured in the mid-luteal phase. Levels of estradiol and progesterone were determined by a chemiluminescent assay (Abbott Axstm analyser: Abbott diagnostics). Levels of LH and FSH were measured using the Advia Centaur assay system (Bayer-Siemens plc). The intra- and inter-assay coefcients of variation were: estradiol 7.9 and 11.9%; progesterone 6.8 and 12.4%; LH 4.5 and 10.4%; FSH 5.9 and 8.0%, respectively.

Immunohistochemistry
Expression of LIF and IL-15 in all endometrial samples was determined by immunohistochemistry. The numbers of uNK cells in each endometrial sample was determined by staining with an anti-CD56 antibody. Sections (5 mm) were dewaxed in xylene, rehydrated through descending alcohols to Tris-buffered saline (TBS) or phosphate-buffered saline (PBS), (pH 7.6), and quenched in 0.3% hydrogen peroxide in methanol for 20 min. A negative control in which the specic antibody was replaced by an equivalent concentration of mouse or goat immunoglobulin G was included for each marker. After washing, unmasking was performed in an 800 W microwave oven in 10 mmol/l citrate buffer (pH 6.0). Buffer was heated in the microwave oven until boiling. The slides were added to the buffer, and left covered at high heat (800 W) for 3 min. Slides were further incubated for 12 min on medium heat and allowed to cool for 20 min. ABC kits (Vector Laboratories, UK) were used according to the manufacturers instructions, except for the following adaptations. Slides were washed in TBS or PBS and blocked in blocking buffer containing 250 ml avidin/ml (Vector Laboratories) for 1 h at room temperature,

Natural killer cells and cytokines in human implantation

Table I Details of the antibodies used in immunohistochemistry studies of the endometrium from women with RIF after IVF.
Antigen CD56 LIF IL-15 Primary antibody Mouse monoclonal anti-human antibody (NCAM) (NovaCastra NCL-CD56 504) Polyclonal goat anti-human LIF antibody (R&D Systems AF-250-NA) Mouse monoclonal anti-human antibody (AbCam ab62834) Secondary antibody system Universal Elite ABC kit (Vector PK-6200) Elite ABC kit (Vector PK-6200) ImmPRESS universal kit (Vector MP-7500) Dilution of primary antibody 1:50 1:25 1:75 Buffer TBS PBS TBS

.............................................................................................................................................................................................

LIF, leukaemia inhibitory factor; IL-15, interleukin 15; TBS, Tris-buffered saline; PBS, phosphate-buffered saline.

and incubated overnight at +48C with the primary antibody containing 250 ml/ml biotin (see Table I for details of the different antibodies used). The addition of avidin in the blocking buffer and biotin with the antibody blocks endogenous activity. The slides were washed in TBS or PBS throughout, and after application of a secondary antibody and Vectorstain, positive cells were visualized by incubation with a peroxidase substrate (3.3 diaminobenzidene terahydrochloride; Vector Laboratories). The slides were washed in distilled water and counterstained with 20% haematoxylin for 30 min, and dehydrated through ascending alcohols, cleared in xylene and mounted in Vectormount (Vector Laboratories). Immunostaining was carried out on sections of tissue from control women and women with RIF at the same time. For LIF and IL-15, the immunostaining and analysis was repeated on sections from the same biopsy on three separate occasions.

Statistical analysis
Data were analysed using the Statistical Package for the Social Sciences version 14. Signicance of differences in the LIF and IL-15 levels between control women and women with RIF was assessed using a nonparametric Mann Whitney test. The Spearman correlation was used to assess correlations between the levels of LIF and IL-15 in endometrial stroma, glandular epithelium and luminal epithelium and number of uNK cells. P , 0.05 was considered signicant.

Downloaded from http://humrep.oxfordjournals.org/ by guest on April 27, 2012

Results
Successful immunostaining for IL-15 was carried out in 36 of 45 biopsies from women with RIF and all 15 biopsies from control women. Successful immunostaining for LIF was carried out in 43 biopsies from women with RIF and 14 biopsies from control women. There was no signicant difference in the mean age, FSH, LH, estradiol and progesterone levels between the control women and the women with RIF.

Analysis of immunostaining
The numbers of CD56-positive stromal cells and the total number of negative stromal cells were counted by eye in 10 microscopic elds (at 400 magnication) for each biopsy. Cell counting was carried out by one individual (N.M.), who had undergone a period of training and a quality control of her counting had been undertaken and was shown to differ by ,5% compared with others in the laboratory. The number of CD56+ cells was expressed as a percentage of all stromal cells. Cell counting was carried out independent of the LIF and IL-15 analysis, and the observer was blinded both with respect to these results and whether the sample was from a control woman or woman with RIF.

LIF in the endometrium

Figure 1a and b shows immunostaining for LIF in the endometrium from control women and women with RIF. Positive staining for the LIF was seen in both stromal and glandular compartments, with much more intense staining in the glandular cells than stromal cells. Staining in the luminal epithelium was less intense than staining in the glandular epithelium. H-score analysis agreed with this observation and showed that expression in stromal cells was very low [median and range; 30 (10 161)]. There was signicantly higher expression of LIF in the glandular epithelium [median and range; 247 (120287)] than in the luminal epithelium [median and range; 188 (20235)] (P , 0.001). This pattern of staining was similar in tissues obtained from both control women and women with RIF. Analysis of H-scores showed that there was no signicant difference (P 0.167) in the intensity of the LIF staining in the stroma of endometrial biopsies from control women versus women with RIF (Fig. 2). In contrast, analysis of H-scores showed that the LIF staining in glandular epithelium in biopsies from women with RIF [median and range; 179 (70 365)] was signicantly less intense than in control women [median and range; 247 (120287)] (P 0.01). The LIF staining in the luminal epithelium of the endometrium from women with RIF was also lower than in control women; however, the difference was not statistically signicant (P 0.159).

H-score
The staining intensity of the LIF and IL-15 in endometrial sections was graded and calculated according to an H-score equation: H-score = Pi (i + 1),

where i staining intensity (1 weak, 2 moderate, 3 strong) and Pi percentage of cells staining at each intensity (0100%). The intensity of the LIF and IL-15 staining was assessed separately in stroma, glandular epithelium and luminal epithelium. Sections were scored independently by two observers, both of whom were blinded to the source of the biopsy. In the event of differences in the score obtained, slides were re-examined and the H-score for that section agreed by both observers. The nal score for LIF and IL-15 staining in each compartment was obtained by taking the mean of the agreed LIF and IL-15 scores of three runs for each compartment (stromal, glandular epithelium and luminal epithelium).

Mariee et al.

Downloaded from http://humrep.oxfordjournals.org/ by guest on April 27, 2012

Figure 1 Immunostaining for LIF and IL15 in endometrium. A photomicrograph of immunostaining for (a) LIF in the endometrium from a control woman; (b) LIF in the endometrium from a woman with RIF after IVF; (c) IL-15 in the endometrium from a normal fertile woman; (d) IL-15 in the endometrium from a woman with RIF; (e and f) negative controls. Scale bar 50 mm.

IL-15 in the endometrium

Figures 1c and d show immunostaining for IL-15 in the endometrium from control women and women with RIF. Positive staining for IL-15 was seen in all endometrial compartments (stromal and glandular and luminal epithelium), with more intense staining seen in the glandular epithelial cells than stromal cells and the luminal epithelium. H-score analysis agreed with this observation and showed that in the samples from control women, there was a signicantly higher intensity of IL-15 staining in the glandular epithelium of mid-luteal phase endometrium [median and range; 235 (200 357)] than in the luminal epithelium [median and range; 208

(150 280)] (P 0.003) and stromal cells [median and range; 60 (15 150)] (P , 0.001). An analysis of H-score values showed that there was signicantly higher intensity (P 0.009) of IL-15 staining in the stroma of endometrial biopsies from women with RIF compared with control women (Fig. 3). The median and the range of H-scores were 90 (0175) and 60 (15 150) for women with RIF and control women, respectively. In contrast, there was no signicant difference (P 0.185) in H-score values for IL-15 in the glandular epithelium of biopsies from control women and women with RIF. H-score analysis showed that IL-15 expression in the luminal epithelium of endometrium from women

Natural killer cells and cytokines in human implantation

Downloaded from http://humrep.oxfordjournals.org/ by guest on April 27, 2012

Figure 2 LIF in the endometrium. Boxplot comparing LIF staining in

(a) stroma, (b) glandular epithelium and (c) luminal epithelium in tissue from the control women (n 14) and the women with RIF (n 44). (An open circle indicates outlier and an asterisk indicates far outlier.) LIF staining in glandular epithelium was signicantly lower in the women with RIF (P 0.01).

Figure 3 IL15 in the endometrium. Boxplot comparing IL-15 staining in (a) stroma, (b) glandular epithelium and (c) luminal epithelium in tissue from control women (n 15) and women with RIF (n 36). (An open circle indicates an outlier and an asterisk indicates a far outlier.) IL-15 staining in stroma was signicantly higher in women with RIF (P 0.009).

with RIF was no different from IL-15 expression in the luminal epithelium of endometrium from control women.

Correlation of uNK cell number and LIF and IL-15 staining

We have previously published the data for uNK cell number in these endometrial biopsies and shown that numbers in women with RIF

were signicantly higher (P , 0.005) than those in control women (Tuckerman et al., 2010). Figures 4 and 5 show the correlation between IL-15 and uNK cell numbers and LIF and uNK cell numbers in biopsies from both control women and women with RIF. A signicant positive correlation was found between uNK cell number and IL-15 in the stroma (P 0.001, r 0.427). There was no correlation between uNK cell number and IL-15 in the glandular epithelium (P 0.370, r 0.164)

Mariee et al.

Downloaded from http://humrep.oxfordjournals.org/ by guest on April 27, 2012

Figure 4 The numbers of uNK cells and IL15 staining. Correlation between the number of uNK cells and IL-15 staining in (a) stroma (P 0.001, r 0.427), (b) glandular epithelium (P 0.370, r 0.164) and (c) luminal epithelium (P 0.078, r 0.292) in biopsies from both control women (n 15) and women with RIF (n 36).

Figure 5 The numbers of uNK cells and LIF staining. Correlation of the number of uNK cells and LIF staining in (a) stroma (r 0.224, P 0.052), (b) glandular epithelium (r 0.089, P 0.596) and (c) luminal epithelium (r 0.044, P 0.662) in biopsies from both control women (n 15) and women with RIF (n 36).

Natural killer cells and cytokines in human implantation

7
KIR, ILT-2 and CD94. Flow cytometry or co-labelling using uorescent immunohistochemistry could have been used to determine expression of these markers, and omission of this analysis is another limitation of the study. In this study, our control group consisted of women of proven fertility, so that we did not include any women with unknown fertility problems. However, it is possible that these previous pregnancies may have long-lasting effects on the immunoregulatory cells of the endometrium that could explain the difference compared with the never-pregnant group. Optimally, this type of study should also include a control group of women who had never been pregnant. The expression of the LIF by normal human endometrium has been previously reported. In agreement with others and with our previous results (Arici et al., 1999; Tuckerman et al., 2004), we observed increased staining for LIF in endometrial epithelial cells compared with stromal cells. The results of the current study showed a signicantly reduced expression of LIF in the glandular epithelium of endometrium from women with RIF compared with control women. To the best of our knowledge, this is the rst study that reports on LIF expression in precisely timed endometrium from women with RIF. Several studies have shown that reduced expression of LIF, or components of its receptor (LIFR and gp130), in the endometrium may be associated with a higher risk of recurrent miscarriage and unexplained infertility (Laird et al., 1997; Hambartsoumian, 1998; Sharkey, 1998; Dimitriadis et al., 2005; Aghajanova et al., 2009). However, studies of LIF expression in subgroups of infertility, including RIF, are limited. So far, only one in vitro study has reported on the expression of the LIF in cultured endometrial cells from women with RIF (Hambartsoumian, 1998). Our results agree with this in vitro study, which found that LIF secretion in endometrial explants cultures from fertile women obtained on Days 1821 of the menstrual cycle was 3.5 times higher than from women with RIF and 2.2 times higher than from infertile women (Hambartsoumian, 1998). A previous study from our laboratory has also shown that, on day LH + 10, women with unexplained infertility had a lower concentration of the LIF in uterine ushings than normal fertile women (Laird et al., 1997). Studies have suggested that LIF expression in the endometrium during the luteal phase could be used to predict the success of IVF treatment in subsequent cycles. One study reported that infertile women who had a strong mid-luteal LIF expression had a 6.4 times higher chance of achieving a pregnancy than those who had a weaker mid-luteal LIF expression (Serani et al., 2008). In contrast, another study reported that decreased endometrial LIF levels are associated with a successful pregnancy outcome in women undergoing IVF (Ledee-Bataille et al., 2002); however, in this study, the LIF was measured in endometrial ushings obtained on Day 26 of the menstrual cycle, which is at a time when LIF levels are normally decreasing. Therefore, a high LIF at this time of the cycle may indicate endometrial dysfunction, and this may be the cause of an unsuccessful pregnancy outcome. In this study, we also investigated IL-15 in the endometrium of the control women and women with RIF. In the control endometrial samples, we found maximal staining for IL-15 in the glandular epithelium and the lowest in the stroma. Our ndings are in agreement with previous reports that also showed increased expression of IL-15 in epithelial cells compared with stromal cells (Chegini et al., 2002),

or in luminal epithelium (P 0.078, r 0.292) (Fig. 4). In contrast, there was no correlation between uNK cell number and LIF in the stroma (P 0.052, r 0.224), the glandular epithelium (P 0.596, r 0.089) or luminal epithelium (P 0.662, r 0.044) (Fig. 5).

Discussion
Implantation of the embryo in a receptive endometrium occurs during a well-dened period of the menstrual cycle which is known as the window of implantation. Both the embryo and the endometrium produce factors that may contribute to successful implantation (Giudice, 1999; Kao et al., 2002). LIF and IL-15 are among several factors that have been previously studied and results have suggested that they may play a role in the implantation process. The exact role of the LIF in implantation is unknown. IL-15 is postulated to play a role in the control of uNK cell proliferation and function, and this may be the mechanism by which it affects embryo implantation (Dimitriadis et al., 2005). In this study, we examined the endometrial LIF and IL-15 expression the mid-luteal phase of the cycle in control women and women with RIF after IVF. This is the time of embryo implantation and also the time in the cycle of maximum endometrial expression of LIF and IL-15 (Keltz et al., 1996; Arici et al., 1999; Kitaya et al., 2000). Dating the endometrium using the timing of the LH surge is more accurate than dating based on the onset of the menstrual cycle (Li et al., 1987).Therefore, all endometrial biopsies, in this study, were precisely timed according to the LH surge and obtained in a well-dened period between days LH + 7 and LH + 9. We assessed expression of LIF and IL-15 by immunohistochemistry and used an H-score analysis to determine the amount and intensity of the staining. Although this method is only semi-quantitative, it has been used extensively to assess the expression of various factors in the endometrium (Lessey et al., 1996; Gomes et al., 2009). More recent studies have used image analysis to provide a more quantitative measurement of the staining intensity (Quenby et al., 2009). However, there are still limitations with these methods: in particular, there is often interference from the haematoxylin counterstain and therefore we chose to use H-score analysis in this study. The study could potentially have been strengthened by the measurement of IL-15 and LIF mRNA by quantitative RTPCR. The advantage of using immunohistochemistry rather than real-time PCR is that it allows a separate assessment of expression in the endometrial stroma, glandular epithelium and luminal epithelium. Different results were obtained when IL-15 and LIF staining were assessed in the different compartments and this may have been missed if real-time PCR analysis had been used. Analysis of uNK cells can be carried out either by ow cytometry or immunostaining. The advantage of using ow cytometry is that you can distinguish between the CD56dim and CD56bright populations of uNK cells. The CD56dim and CD56bright populations may have different functions in early pregnancy (Bulmer and Lash, 2005) and it would have been interesting to correlate IL-15 and LIF expression with each of these populations separately. The fact that we have only used immunostaining and therefore could not carry out this type of analysis is a major limitation of the usefulness and novelty of this study. The study would also be strengthened by investigation of additional uNK cell markers, particularly the HLA receptors, such as

Downloaded from http://humrep.oxfordjournals.org/ by guest on April 27, 2012

8
although an intense expression was seen in perivascular stomal cells invading the spiral arteries (Kitaya et al., 2000). There are few reports on the expression of IL-15 in women with reproductive failure. One study has suggested that women with recurrent miscarriage have elevated levels of IL-15 expression in the stroma and glandular epithelium of the endometrium compared with normal fertile women (Chegini et al., 2002). So far, only one study has examined expression of the IL-15 mRNA in women with RIF and showed that women with unexplained IVF-embryo transfer failure have signicantly lower expression of the IL-15 mRNA compared with the control women (Ledee-Bataille et al., 2004). Our study is the rst to examine the expression of IL-15 in precisely timed endometrium from women with RIF. We found signicantly higher levels of IL-15 in the endometrial stroma of women with RIF compared with the control women. In contrast to the previous study by Ledee-Bataille et al. (2004), we were able to assess IL-15 in the stromal, luminal and glandular epithelium separately and found that the increase was only seen in the stromal compartment. In this study, we also found a signicant positive correlation between the numbers of endometrial uNK cells and IL-15 staining in the stroma. In contrast to IL-15, the levels of the LIF in none of the three endometrial cell types correlated with the number of uNK cells. This suggests that the correlation of IL-15 with the number of uNK cells is not related to a general endometrial defect but is specic for IL-15. Our ndings are in agreement with the ndings by LedeeBataille et al. (2004) who also found a signicant positive correlation between the IL-15 mRNA and the number of uNK cells in women with IVF-embryo transfer failure. A correlation between IL-15 and the number of uNK cells was only seen for stromal IL-15 and not for IL-15 in the luminal and glandular epithelium. uNK cells are present in the endometrial stroma and these results, therefore, suggest that IL-15 produced by the stroma may play a role in controlling uNK cell function. Other studies have suggested that stromal IL-15 is increased during decidualization (Kitaya et al., 2000) at the start of pregnancy, when the number of uNK cells in the endometrium increases dramatically, further suggesting a role for IL-15 in uNK cell function. In summary, this study has shown that women with RIF have signicantly lower LIF protein levels in the endometrial glandular epithelium compared with controls. It has also shown that women with RIF had more IL-15 in the stroma compared with control women. Moreover, this was the rst study to show a positive correlation between the number of endometrial uNK cells and IL-15 in the stroma. The increase in both the number of uNK cells and staining for IL-15 support the theory that IL-15 is involved in the recruitment and proliferation of uNK cells in the endometrium.

Mariee et al.

the paper. S.M.L., the principal investigator, supervised and provided advice on the work; wrote the paper.

Funding
The work was funded by a studentship to N.M. from the Libyan government.

Conict of interest
None declared.

References
Aghajanova L, Altmae S, Bjuresten K, Hovatta O, Landgren BM, Stavreus-Evers A. Disturbances in the LIF pathway in the endometrium among women with unexplained infertility. Fertil Steril 2009;91:2602 2610. Arici A, Engin O, Attar E, Olive DL. Modulation of leukaemia inhibitory factor gene expression and protein biosynthesis in human endometrium. J Clin Endocrinol Metab 1999;80:1908 1914. Bulmer JN, Lash GE. Uterine natural killer cells; a reappraisal. Mol Immunol 2005;42:511 521. Bulmer JN, Morrison L, Longfellow M, Ritson A, Pace D. Granulated lymphocytes in human endometrium: histochemical and immunohistochemical studies. Hum Reprod 1991;6:791 798. Carson WE, Giri JG, Lindemann MJ, Linett ML, Ahdieh M, Paxton R, Anderson D, Eisenmann J, Grabstein K, Caligiuri MA. IL-15 is a novel cytokine that activates human natural killer cells via components of the IL-2 receptor. J Exp Med 1994;180:1395 1403. Chegini N, Ma C, Roberts M, William RS, Ripps BA. Differential expression of IL-13 and IL-15 throughout the menstrual cycle in endometrium of normal fertile women and women with recurrent abortion. J Reprod Immunol 2002;56:93 110. Chegini N, Roberts M, Ripps B. Differential expression of IL-13 and IL-15 in ectopic and eutopic endometrium of women with endometriosis and normal fertile women. Am J Reprod Immunol 2003;49:75 83. Clifford K, Flanagan AM, Regan L. Endometrial CD56+ natural killer cells in women with recurrent miscarriage. Hum Reprod 1999; 14:2727 2730. Croy BA, Esadeg S, Chantakru S, van de Heuval M, Paffaro VA, He H, Black GP, Ashkar AA, Kiso Y, Zhang J. Update on pathways regulating the activation of uterine natural killer cells, their interactions with decidual spiral arteries and homing precursors to the uterus. J Reprod Immunol 2003;59:175 191. Dimitriadis E, White CA, Jones RL, Salamonsen LA. Cytokines, chemokines and growth factors in endometrium related to implantation. Hum Reprod Update 2005;11:613 630. Giudice LC. Potential biochemical markers of uterine receptivity. Hum Reprod 1999;14(Suppl. 2):3 16. Gomes MKO, Roas-e-Silva JC, Garcia SB, Rosa-e-Silva ACJD, Turatti A, Viera CS, Ferriani RA. Effects of levonorgestrel-releasing intrauterine system on cell proliferation, Fas expression and steroid receptors in endometriosis lesions and normal endometrium. Hum Reprod 2009; 24:2736 2745. Hambartsoumian E. Endometrial LIF as a possible cause of unexplained infertility and multiple failures of implantation. Am J Reprod Immunol 1998;39:137 143. Kao LC, Tulac S, Lobo S, Imani B, Yang JP, Germeyer A, Osteen K, Taylor RN, Lessey BA, Giudice LC. Global gene proling in human

Downloaded from http://humrep.oxfordjournals.org/ by guest on April 27, 2012

Acknowledgements
We would like to thank the staff at the assisted conception unit at the Jessop Wing, Shefeld Teaching Hospitals for their help in the collection of endometrial biopsy samples.

Authors roles
N.M. carried out all the laboratory work and contributed to writing the paper. T.C.L. provided human tissue and contributed to writing

Natural killer cells and cytokines in human implantation

9
Endometrial NK cells are special immature cells that await pregnancy. J Immunol 2008;181:1869 1876. Quenby S, Bates M, Doig T, Brewster J, Lewis-Jones DI, Johnson PM, Vince G. Pre-implantation endometrial leukocytes in women with recurrent miscarriage. Hum Reprod 1999;14:2386 2391. Quenby S, NIK H, Innes B, Lash G, Turner M, Drury J, Bulmer J. Uterine natural killer cells and angiogenesis in recurrent reproductive failure. Hum Reprod 2009;24:45 54. Serani P, Rocha AM, Osorio CT, da Silva I, Motto EL, Baracat EC. Endometrial leukemia inhibitory factor as a predictor of pregnancy after in vitro fertilisation. Int J Gynecol Obstet 2008;102:23 27. Sharkey AM. Cytokines and implantation. Rev Reprod 1998;3:52 61. Stewart CL, Kaspar P, Brunet LJ, Bhatt H, Gadi I, Kontgen F, Abbondanzo SJ. Blastocyst implantation depends on maternal expression of leukaemia inhibitory factor. Nature 1992;359:76 79. Tuckerman E, Laird SM, Stewart R, Wells M, Li TC. Markers of endometrial function in women with unexplained recurrent pregnancy loss: a comparison between morphologically normal and abnormal endometrium. Hum Reprod 2004;19:196 205. Tuckerman E, Laird SM, Prakash A, Li TC. Prognostic value of the measurement of uterine natural killer (uNK) cells in the endometrium of women with recurrent miscarriage. Hum Reprod 2007; 22:2208 2213. Tuckerman E, Mariee N, Prakash A, Laird SM, Li TC. Uterine natural killer cells in peri-implantation endometrium from women with repeated implantation failure after IVF. J Reprod Immunol 2010; 87:60 66. Van Mourik MSM, Macklon NS, Heijnen CJ. Embryonic implantation: cytokines, adhesion molecules and immune cells in establishing an implantation environment. J Leukoc Biol 2009;85:4 19. Verma S, Hiby SE, Loke YW, King A. Human decidual natural killer cells express the receptor for and respond to the cytokine IL-15. Biol Reprod 2000;62:959 968.

endometrium during the window of implantation. Endocrinology 2002; 143:21192138. Keltz MD, Attar E, Buradagunta S, Olive DL, Klima HJ, Arici A. Modulation of leukemia inhibitory factor gene expression and protein synthesis in the human fallopian tube. Am J Obstet Gynecol 1996;175:1611 1619. Kimber SJ. Leukaemia inhibitory factor in implantation and uterine biology. Reproduction 2005;130:131 145. Kitaya K, Yasuda J, Yagi I, Tada Y, Fushiki S, Honjo H. IL-15 expression in human endometrium and decidua. Biol Reprod 2000;63:683 687. Laird SM, Tuckerman EM, Dalton CF, Dunphy BC, Li TC, Zhang X. The production of leukaemia inhibitory factor (LIF) by human endometrium: presence in uterine ushings and production by cells in culture. Hum Reprod 1997;12:569574. Laird SM, Tuckerman EM, Cork BA, Linjawi S, Li TC. A review of immune cells and molecules in women with in recurrent miscarriage. Hum Reprod Update 2003;9:163174. Ledee-Bataille N, Lepree-Delage G, Taupin JL, Dubanchet S, Frydman R, Chaouat G. Concentration of LIF in uterine ushing uid is highly predictive of embryo implantation. Hum Reprod 2002;17:213 218. Ledee-Bataille N, Bonnet-Chea K, Hosny G, Dubanchet SY, Frydman R, Chaouat G. Role of endometrial tripod IL18 -15 and 12 in inadequate uterine receptivity in patients with a history of repeated in vitro fertilisation-embryo transfer. Fertil Steril 2004;8:598605. Lessey BA, Yeh IT, Castelbaum AJ, Fritz MA, Ilesanmi AO, Korzeniowski P, Sun JH, Chwalisz K. Endometrial progesterone receptors and markers of uterine receptivity in the window of implantation. Fertil Steril 1996;65:477 483. Li TC, Rogers AW, Lenton EA, Dockery P, Cooke I. A comparison between 2 methods of chronological dating of human endometrial biopsies during the luteal phase and their correlation with histologic dating. Fertil Steril 1987;48:928 932. Manaster I, Mizrahi S, Golman-Wohl D, Sela H, Stern-Ginossar N, Lankry D, Gruda R, Hurwitz A, Bdolah Y, Haimov-Kockman R et al.

Downloaded from http://humrep.oxfordjournals.org/ by guest on April 27, 2012