Increase in peripheral blood mononuclear cell (PBMC)- and CD56+ cell-mediated killing of endometrial stromal cells by mycobacteria; a possible role in endometriosis immunotherapy?

R.D. Clayton1,2,5, S.R. Duffy2, N. Wilkinson3, R. Garry2,4 and A.M. Jackson1

1 Applied Immunology Laboratory, Cancer Research UK Clinical Centre, 2 Department of Obstetrics and Gynaecology and 3 Department of Histopathology, St James's University Hospital, Leeds, LS9 7TF, UK and 4 Department of Gynaecology, University of Western Australia, King Edward Memorial Hospital, Perth, WA 6008, Australia

5 To whom correspondence should be addressed at: Christie Hospital, Wilmslow Road, Manchester M20 4BX, UK. Email: claytonrd{at}hotmail.com


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
BACKGROUND: Immunological therapies have shown promising results in the treatment of endometriosis. Mycobacteria are one of the most common immune therapies used in other diseases. We have assessed the effects of mycobacteria in altering the ability of peripheral blood mononuclear cells (PBMCs) and natural killer (NK) cells to kill endometrial stromal cells using an in vitro model. This may have implications in the immunotherapy of endometriosis. METHODS: Primary cultures of endometrial stromal cells were grown from female patients and PBMCs were extracted from healthy female volunteers. Effector cells (PBMCs or NK cells) were exposed to varying concentrations of mycobacteria before their ability to kill cultured endometrial cells was tested using a 51Cr-release assay. RESULTS: Treatment of effector cells with the Connaught Substrain Bacillus of Calmette and Guérin (BCG) led to increased killing of target cells by PBMCs and NK cells. The optimal concentration for treatment of effector cells with Connaught BCG was ~0.1 multiplicities of infection (m.o.i.). There was a trend towards increased killing after treatment with Pasteur BCG. CD56+ (NK) cells treated with BCG at 0.1 m.o.i. showed increased killing of target cells compared with untreated effector cells. CONCLUSIONS:Endometrial stromal cells are susceptible to killer cells activated by mycobacteria. This in vitro work suggests a possible role for mycobacteria in the immunotherapy of endometriosis.

Key words: BCG/endometriosis/endometrium


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
It is thought likely that endometriosis has an immunological basis, yet none of the current medical therapies for endometriosis target potential defects in immunological defences (Gleicher, 1995Go). Intraperitoneal (I.p.) administration of immunomodulatory agents such as interferon (IFN)-{alpha}2b on a trial basis in animals and humans has been reported recently for the treatment of endometriosis, and has been found to be successful in reducing implant size (Ingelmo et al., 1999Go; Keenan et al., 1999Go; Ali et al., 2000Go) although this was not found in a recent trial by Acien et al. (2002)Go.

One of the most effective immunotherapeutic agents currently in use is the Bacillus of Calmette and Guérin (BCG), particularly for the treatment of bladder cancer. It is given by intravesical instillation and its mechanism of action seems to depend largely on close contact with superficial bladder tumour cells (Bohle, 2000Go). Might BCG or other mycobacteria which are already established immunotherapeutic agents be of use in the treatment of endometriosis? Our group has reported previously on the decrease in proliferation in cultured endometrial stromal cells in response to BCG (Clayton et al., 2004Go). A recent report on the use of BCG vaccination to prevent implantation of endometriosis in rats has shown an inhibitory effect on endometrial transplantation (Gul et al., 2001Go). These findings may suggest a role for BCG or other mycobacteria in the treatment of endometriosis.

The response to the intravesical administration of BCG for treatment of bladder neoplasia is complex (Alexandroff et al., 1999Go), involving recruitment and activation of immune cells, direct effects on reduction of proliferation, production of large numbers of inflammatory mediators, including a variety of cytokines, and induction of adhesion molecules.

Analogous to the intravesical instillation of BCG for bladder tumours, i.p. administration of mycobacteria would bring the target cells into direct contact with mycobacteria.

An important component of the immune response to mycobacteria during immunotherapy of bladder cancer involves the generation of BCG-activated killer cells (Bohle, 2000Go). These cells demonstrate enhanced killing of bladder tumour target cells. Exposure of immune cells in the peritoneal cavity to BCG might also be expected to lead to the generation of these killer cells. It is therefore important to assess whether deposits of endometriosis in the peritoneal cavity might be susceptible to this cell killing. In the experiments presented here, primary cultures of endometrial stromal cells were used as the targets to assess whether exposure of peripheral blood mononuclear cells (PBMCs) to mycobacteria would lead to any alteration in target cell killing. It is thought that in endometriosis patients, there is a defect in the ability of natural killer (NK) cells to clear refluxed endometrial cells from the peritoneal cavity (Witz, 1999Go), therefore an increase in cell killing by CD56+ cells might be of relevance to clearance of endometriosis.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Endometrial tissue samples were obtained at the time of laparoscopy, either by dilation and curettage or by pipette biopsy of the endometrium. Endometrium was obtained from 10 patients, with a total of 21 cultures used in experiments. Six patients were undergoing laparoscopic sterilization, three were undergoing diagnostic laparoscopy for pelvic pain, of whom two had endometriosis, and one laparoscopy and dye for investigation of infertility. Eleven secretory cultures and 10 proliferative cultures were used. The median age was 33 years (range 26–42). The local ethics committee approved this study and written consent was obtained from all study subjects. The method for processing endometrial tissue samples has been described previously (Clayton et al., 2004Go) and is based on the method described by Vigano et al. (1993)Go.

Tissue samples were processed within 60 min of being obtained. The tissue was minced using a crossed scalpel technique into small fragments <1 mm3 and placed in complete medium [50:50 Dulbecco's modified Eagle's medium (DMEM):RPMI-1640 (Gibco, Uxbridge, UK) supplemented with 10% heat-inactivated fetal calf serum (FCS; Sera Lab, Crawley Down, UK), 2 mmol/l L-glutamine (Gibco), 50 µg/ml gentamicin (Sigma, Poole, UK) and 5 µg/ml fungizone (Sigma)] with the addition of 0.1% collagenase 1A (Sigma), prior to incubation in a shaking water bath at 37°C for 2 h. The tissue was then repeatedly aspirated and pipetted using a narrow pastette (Alpha, Hants, UK) until the isolate appeared well dispersed. Single stromal cells were separated from clumps of epithelial cells by differential sedimentation at unity gravity for 10 min. The upper fraction, composed mainly of stromal cells, was removed to another tube, and washed twice. The pellet was resuspended in 5 ml of complete medium and allowed to adhere to a 25 cm2 tissue culture flask (Falcon, Fahrenheit, Leeds, UK) for 15 min at 37°C in 5% CO2. The non-adherent epithelial cells still present were then removed by careful washing of the culture flask with complete medium.

Part of the sample was placed in formalin for histopathological examination in order to confirm the stage of the menstrual cycle according to the criteria of Noyes et al. (1975).

Adherent endometrial stromal cells were cultured in a 5%CO2/95% air incubator at 37°C and were grown to confluence before passage. Experiments were performed in the second or third passage of cultured cells

The cell cultures were characterized morphologically by immunohistochemical staining with anti-vimentin and anti-cytokeratin 19 monoclonal antibodies (Dako, Ely, UK). Secondary antibody [goat anti-mouse IgG–fluorescein isthiocyanate (FITC) conjugate (Harlan Sera lab, Loughborough, UK)] was applied to each well including controls.

The degree of contamination of cell cultures by monocytes was assessed by labelling two separate cell lines after two passages, with anti-CD14 monoclonal antibody, and assessing the percentage of cells showing positive staining using flow cytometry

Cultures of BCG Pasteur and Mycobacterium smegmatis were grown from stocks frozen at –80°C in 15% glycerol (BDH Merck, Lutterworth, UK) and kindly donated by Dr A.Jackson. BCG Connaught was obtained from the manufacturer (Aventis Pasteur, Canada). BCG substrain Connaught is a freeze-dried preparation made from the Connaught strain of BCG, which is an attenuated strain of Mycobacterium bovis. BCG Pasteur is another of the many therapeutic strains that have been used in the treatment of bladder cancer. Taxonomically, mycobacteria are part of the genus Mycobacterium, which is the single genus in the family of Mycobacteriaceae, in the order Actinomycetales.

Heparinized venous blood (1.25 IU of heparin/ml of blood) was obtained from healthy female donors who were not taking any hormonal medication, and diluted 2-fold with phosphate-buffered saline (PBS). PBMCs were obtained by centrifugation through ficoll; 25 ml of diluted blood was layered onto 10 ml of lymphoprep solution (Nycomed, Birmingham, UK) in a sterile 50 ml tube (Falcon). Blood was then subjected to centrifugation at 350 g for 30 min. Cells at the interface were removed and washed twice in Hanks balanced salt solution (HBSS) (Sigma). Cells were resuspended in RPMI containing 2 mmol/l glutamine and 10% FCS and counted in a haemocytometer.

For the generation of activated killer cells, isolated PBMCs were cultured for a period of 7 days in 6-well plates with different multiplicities of infection (m.o.i.) of BCG for the generation of BAK (BCG activated killer) cells; for example, if there were 1 x 107 PBMCs, then 5 x 106 BCG would be added to give an m.o.i. of 0.5. Control wells containing cells with complete medium alone were also cultured for 7 days.

For the isolation of NK cells, PBMCs were suspended in MACSTM buffer [PBS pH 7.2 + 0.5% bovine serum albumin (Sigma)+2 mmol/l EDTA (Gibco)] at a concentration of 1 x 107 cells in 80 µl of MACS buffer. A 20 µl aliquot of a suspension of CD56-labelled MACS beads (Miltenyi Biotec, Germany) was added per 1 x 107 PBMCs present. The solution was mixed thoroughly and placed at 6°C for 15 min. The cells were washed in MACS buffer and then resuspended in 0.5 ml of MACS buffer before being placed in a MACS RS+ selection column held in a VarioMACS magnetic separator (Miltenyi Biotec, Germany). The column was washed four times with MACS buffer and then removed from the separator. The column was placed over a 6 ml round-bottomed tube (Falcon) and 1 ml of MACS buffer added. A plunger was used to detach the CD56-MACS bead-labelled cells from the column. The cells were spun at 200 g for 5 min, resuspended in 10% RPMI and counted in a haemocytometer.

The ability to radioactively label cells and determine the amount of radioactivity released as a measure of cell lysis forms the basis of the 51Cr-release assay. All the experiments were performed in allogeneic conditions. Endometrial stromal cells were used as target cells. Target cells were grown almost to confluence in 75 cm flasks. Depending on the experiment to be performed, the cells were then exposed to fresh medium (untreated) or mycobacteria at concentrations of 1 x 105, 1 x 106 or 1 x 107 c.f.u./ml for 2 days (M. smegmatis only) or 7 days. Culture with M. smegmatis generally led to overgrowth of the culture with mycobacteria when prolonged exposure was used, hence the shorter time course in these experiments. The cells were harvested and counted. A total of 1 x 106 cells were pelleted in a 25 ml universal tube. A large number of initial experiments were performed to assess the optimum method of labelling target cells in order to reduce the spontaneous release rate (SRR). Repeated centrifugation to wash after labelling resulted in lysis of the fragile target cells; therefore, a method whereby target cells were allowed to stand and occasionally shaken gently was developed. This resulted in more viable target cells and acceptable SRR levels. All the supernatant was removed and the cells were resuspended in 100 µCi of sodium 51chromate (Amersham, Bucks, UK) and incubated for 1 h at 37°C. Cells were then washed in 20 ml of complete medium and resuspended in 20 ml of medium before allowing to stand for 90 min at 37°C. The cells were again washed and resuspended before standing for 120 min. After this time, the cells were counted and resuspended at a concentration of 5 x 104/ml after a further wash.

PBMCs or NK effector cells were counted and adjusted to a density of 2.5 x 106/ml. Using a U-bottomed 96-well plate (Nunc, Life Technologies, Glasgow, UK), a 200 µl volume was placed in the first three wells of a row and 100 µl of medium in the remaining wells of the row. Two-fold serial dilutions were prepared in triplicate to the end of the row and, depending on the minimum effector:target ratio, into the next row. Twelve wells were used to assess maximum and spontaneous release of radioactivity. To six of these wells, 100 µl of 0.1 mol/l HCl (BDH, Lutterworth, UK) was added and, to the remainder, medium alone was added. A 100 µl aliquot of target cell suspension was added to each well. Plates were incubated for 4 h at 37°C. Cells were pelleted by centrifugation of plates at 200 g for 5 min. Following this, 100 µl of supernatant was removed and placed into microtubes (Alpha, Hants, UK). Radioactivity was then measured using a gamma counter and percentage specific release was calculated as follows:


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Immunohistochemical staining with anti-vimentin and anti-cytokeratin antibodies demonstrated positivity for vimentin, with staining around the nucleus radiating into the cytoplasm. There was no staining of cytokeratin except in occasional epithelial cells (<1%) found amongst stromal cells. Less than 2% of cells labelled with FITC conjugated monoclonal antibody to CD14 in each cell line, indicating very low levels of contamination of cultures with macrophages.

Nine experiments were performed using Connaught BCG at varying concentrations for treatment of effector cells and, in some cases, treatment of target cells. In all cases, treatment of target and effector cells was for 7 days. Combining individual experiments for untreated target cells (Figure 1a), it can be seen that treatment of PBMCs with Connaught BCG increases 51Cr release at an m.o.i. of 0.01, 0.05 and 0.1, differences being more marked at the higher effector:target ratios. At an effector:target ratio of 50:1, the data were normally distributed. There was found to be a significant difference in 51Cr release between the different treatment groups (P=0.008), using one-way analysis of variance (ANOVA). There was poor homogeneity of variance; however, post hoc comparison using Tamhane's test (homogeneity of variances not assumed) showed a significant difference between control and an m.o.i. of 0.1 (P=0.03)



View larger version (17K):
[in this window]
[in a new window]
 
Figure 1. 51Cr release from endometrial stromal cells. (a) With no BCG exposure using PBMCs as effectors exposed to varying Connaught BCG treatment concentrations for 7 days. Overall spontaneous release rate (SRR)=31%. In the series of nine experiments, all experiments used control PBMCs and PBMCs treated at an m.o.i. of 0.1, whereas seven experiments used an m.o.i. of 0.5 and four used an m.o.i. of 0.01 and 0.05. (b) With no BCG exposure. PBMCs treated with Pasteur BCG for 7 days. Overall SRR = 37%. There were seven experiments using control and an m.o.i. of 0.05 and 0.1, and five using an m.o.i. of 0.01 and 0.5

 
Seven experiments were performed using Pasteur BCG at varying concentrations for treatment of effector cells and, in some cases, treatment of target cells. In all cases, treatment of target and effector cells was for 7 days (Figure 1d). When untreated target cells are used, there appears to be a non-significant, marginal increase in 51Cr release at effector:target ratios of 50:1 and 25:1 when effector cells are treated at an m.o.i. of 0.01 and 0.05 (Figure 1b)

The overall effects of the series of experiments in which the target cells were exposed to Connaught BCG at a concentration of 1 x 106 or 1 x 107 c.f.u./ml are shown in Figure 2a and b. At a concentration of 1 x 106, no significant difference was noted between the groups at any effector:target ratio. For target cells at 1 x 107 c.f.u./ml (Figure 2b), at an effector:target ratio of 50:1, the data were normally distributed. One-way ANOVA showed a significant difference between the groups; however, there was poor homogeneity of variance. Logarithmic transformation of the data resulted in acceptable homogeneity of variance. The ANOVA after transformation showed a significant difference (P<0.001) and post hoc analysis using Gabriel's test showed significant differences between control and an m.o.i. of 0.1 (P<0.001) and between control and an m.o.i. of 0.5 (P<0.001). At an effector:target ratio of 25:1, ANOVA showed a significant difference between the groups after transformation of the data had resulted in acceptable homogeneity of variance (P=0.026). Post hoc analysis showed a difference between control and an m.o.i. of 0.5 (P=0.032), but not between control and an m.o.i. of 0.1 (P=0.105). At an effector target ratio of 12.5:1, no significant difference was noted between groups (P=0.059).



View larger version (18K):
[in this window]
[in a new window]
 
Figure 2. 51Cr release from endometrial stromal cells. (a) With Connaught BCG exposure at a concentration of 1 x 106 c.f.u./ml for 7 days. PBMC treated with Connaught BCG. Overall SRR = 25%. There were five experiments with untreated PBMCs and PBMCs treated at an m.o.i. of 0.1, and four using cells treated at an m.o.i. of 0.5. (b) With Connaught BCG exposure at a concentration of 1 x 107 c.f.u./ml for 7 days. PBMCs treated with Connaught BCG. Overall SRR = 29%. There were five experiments with untreated PBMCs and PBMCs treated at an m.o.i. of 0.1, and three using cells treated at an m.o.i. of 0.5. PBMCs at an m.o.i. of 0.05 and 0.01 were used only once. (c) With Pasteur BCG exposure at a concentration of 1 x 107 c.f.u./ml for 7 days. PBMCs treated with Pasteur BCG. Overall SRR = 36%. There were seven experiments, of which all used untreated PBMCs and PBMCs at an m.o.i. of 0.1, six used an m.o.i. of 0.01 and 0.05, and two used an m.o.i. of 0.5.

 
Comparison of the experiments using untreated targets and targets treated with Connaught BCG at 1 x 107 (Figures 1a and 2b) showed a significant increase in killing of treated targets by effector cells treated with Connaught BCG at an m.o.i. of 0.5 using effector:target ratios of 50:1 (P=0.03) and 25:1 (P=0.003)

When endometrial stromal cells were pre-treated with BCG Pasteur strain at a concentration of 1 x 107 c.f.u./ml (Figure 2c), there was a trend towards a greater difference between 51Cr release by untreated effectors and effectors at an m.o.i. of 0.01 and 0.05 when compared with the untreated target cells; however, this difference did not reach statistical significance (P=0.138). Insufficient experiments were performed using target cells treated at 1 x 106 c.f.u./ml to show any discernible trends (data not shown).

Three experiments were performed using M. smegmatis at effector and target cell treatment concentrations identical to those used in the experiments detailed above for Pasteur and Connaught BCG, except that, in addition, target cells treated at a concentration of 1 x 105 c.f.u./ml were used. Treatment of target and effector cells was in all cases for 48 h. There was no discernible difference in 51Cr release between target cells and effector cells at the various concentrations (data not shown).

Target cells treated at 1 x 105, 1 x 106 or 1 x 107 c.f.u./ml with M. smegmatis for 48 h and exposed to untreated PBMCs did show a trend towards greater 51Cr release than untreated target cells (Figure 3), a trend not noted with the experiments using BCG. This trend did not reach statistical significance.



View larger version (10K):
[in this window]
[in a new window]
 
Figure 3. 51Cr release from endometrial stromal cells with M. smegmatis exposure at a concentration of 0, 1 x 105, 1 x 106 or 1 x 107 c.f.u./ml for 48 h using untreated PBMCs as effectors.

 
Effect of CD56+ cells
Cells isolated using CD56-MACS beads were found to be >90% CD56+ by fluorescence-activated cell sorting (FACS) analysis. 51Cr release assays were performed three times using untreated endometrial stromal cells as targets or cells treated with Connaught BCG at 1 x 107 c.f.u./ml for 7 days. Isolated NK cells, untreated or exposed to an m.o.i. of 0.1 of Connaught BCG for 7 days before isolation, were compared with untreated PBMCs and PBMCs treated with Connaught BCG at an m.o.i. of 0.1. In addition, 51Cr release caused by untreated PBMCs and PBMCs treated with Connaught BCG at an m.o.i. of 0.1 after depletion of CD56+ cells was assessed. The results are shown in Figures 4a and b. It can be seen that untreated PBMCs and PBMCs exposed to an m.o.i. of 0.1 of Connaught BCG led to 51Cr release similar to that found in Figures 1a and 2b, as would be expected. When untreated PBMCs are CD56 depleted, there is almost no change in 51Cr release, although, particularly against untreated targets, there appears to be a slight reduction in release. Isolated CD56+ cells were only available in sufficient quantities to allow serial dilutions from an effector:target ratio of 25:1. Untreated CD56+ cells elicited increased 51Cr release when compared with PBMCs, particularly against untreated targets, and were found to lead to 51Cr release similar to treated PBMCs, although against treated targets, the treated PBMCs showed a trend towards greater release.



View larger version (16K):
[in this window]
[in a new window]
 
Figure 4. (a) 51Cr release from endometrial stromal cells with no BCG exposure using PBMCs and NK cells as effectors. Overall SRR = 37%. (b) 51Cr release from endometrial stromal cells with BCG exposure at a concentration of 1 x 107 c.f.u./ml for 7 days using PBMCs and NK cells as effectors. Overall SRR = 36%.

 
CD56+, treated cells led to the greatest 51Cr release by both treated and untreated target cells; however, the level of release was greater in the case of treated target cells. Using treated target cells, there was a statistically significant difference between CD56+ treated cells and controls at effector:target ratios of 25:1 (P=0.002) and 12.5:1 (P=0.003), but no significant differences were found for untreated target cells.


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Difficulty can be encountered with high SRRs when primary cell cultures are used as targets compared with cell lines such as K562, as has been noted by other workers (Coligan et al., 1994Go; D'Hooghe and Hill, 1995Go). 51Cr seems to be more toxic to primary cell cultures. An SRR of ≤35–40% was aimed for, a level used by other workers with endometrial stromal cells as targets (P.Vigano, personal communication). Experiments in which SRR levels were significantly higher were rejected.

Thanhauser et al. (1993)Go and Bohle et al. (1993)Go found that in relation to killing of bladder carcinoma cell lines, BAK cells were optimally generated with a 7 day incubation at an m.o.i. of 0.04. The series of experiments conducted for the present study therefore used an m.o.i. at around this level for generation of effectors. Using Connaught BCG, it is difficult to generate a precise m.o.i., as the manufacturer gives a range of possible c.f.u. for each vial of 10.5±8.7 x 108 c.f.u. per vial. The mid point of this range was taken for our experiments. Initial experiments with Connaught BCG suggested that an m.o.i. of 0.1 or 0.5 may be most effective; therefore, experiments tended to use these levels of m.o.i. Ideally, all possible m.o.is would have been used for each experiment; however, this was not possible in many cases because of a shortage of isolated PBMCs. More precise estimates of the m.o.i. were possible with Pasteur BCG as this mycobacterium had been stored at known concentrations. The greatest trend towards an increase in killing with Pasteur BCG occurred at an m.o.i. of 0.01. A further dilution was not used in these experiments, and it is therefore not possible to assess whether this would have led to further improvement in killing.

An interesting finding of the experiments with BCG is increased cell killing of target cells which have themselves been treated with BCG. It is possible that exposure to BCG, which upregulates expression of adhesion molecules such as intracellular adhesion molecule-1 (ICAM-1), may predispose to enhanced binding of BAK cells to targets and hence increased cell killing. Another possible explanation would be that these cells are more fragile after BCG exposure, and are therefore more prone to spontaneous lysis. This, however, was not found to be the case, as the SRRs of target cells treated with mycobacteria; other than M smegmatis were found to be no greater than those of the untreated targets.

The optimal time for generation of M. smegmatis activated killer cells is unknown. In the initial experiments presented here, a time of 48 h for exposure was selected, as some increase in killing had been noted at this time point using bladder cancer cell lines (A.M.Jackson, unpublished observations). Culture with M. smegmatis generally leads to overgrowth of the culture with mycobacteria when prolonged exposure is used; hence the shorter time course in these experiments. It is possible that if more prolonged exposure of PBMCs were possible, then significant killing effects may be found. Of interest in the experiments presented here was the increased susceptibility of target cells treated with M. smegmatis to killing by unstimulated PBMCs.

In the series of experiments using isolated CD56+ cells, the finding of increased killing by these effectors suggests that this is the population of cells which leads to most of the target cell killing. This evidence is strengthened by a reduction in killing below the level of control PBMCs when these cells are depleted from the population. Most of these will be NK cells, killing in a non-MHC-restricted fashion. The target cells are allogenic rather than autologous; therefore, killing by T cells in an MHC-restricted fashion is unlikely.

Recent evidence suggests that as endometriosis has an immunological basis, immunotherapy may lead to disease regression. Ingelmo et al. (1999)Go used a murine model of experimentally induced endometriosis to evaluate the effect of IFN-{alpha}2b on implant size. They found that a single i.p. dose of IFN-{alpha}2b led to an initial reduction (40% on day 6) in implant size compared with control animals, and this effect plateaued at 25% in the longer term (120 days). Three s.c. doses led to a smaller initial reduction (13%) which plateaued at 23% in the longer term. Ali et al. (2000)Go used i.p. injection of IFN-{alpha}2b at the time of diagnosis of endometriosis by laparoscopy in women undergoing investigation of infertility, and noted a reduction in symptoms including dyspareunia, pain with defaecation and dysmenorrhoea by 3 months. Second-look laparoscopy at 3 months suggested a reduction in the stage of the disease. They concluded that a controlled, randomized study would be worthwhile to investigate this effect further. These findings contrast with those of Acien et al. (2002)Go whose group found an increased late recurrence rate in patients treated with IFN-{alpha}2b. It was interesting to find that GnRH analogues made no difference to recurrence given that other studies of post-surgical treatment with GnRH have shown a reduction in recurrence of symptoms (Hornstein et al., 1997Go; Vercellini et al., 1999Go).

A novel study by Gul et al. (2001)Go, using a rat model, showed that BCG vaccination significantly reduced the incidence of endometrial implantation when endometrium was injected into the eye. This suggests generalized enhancement of the immunological ability to reject implantation of ectopic endometrial cells.

Luo et al. (1999, 2001)Go, Luo et al. (2001)Go) have found that IFN-{alpha}2b acts synergistically with BCG in the treatment of superficial bladder cancer by enhancing Th1 cytokine responses. This effect may have implications for the possible benefits of mycobacteria in the treatment of endometriosis. Additionally, IFN-{alpha}2b has been used as adjuvant treatment for high risk melanoma, leading to improved survival (Agarwala and Kirkwood, 2000Go).

We have shown using an in vitro model that exposure to BCG can increase the ability of PBMCs and NK cells to kill target endometrial cells. In addition, previous work from this group has shown a reduction in proliferation of endometrial stromal cells in response to BCG (Clayton et al., 2004Go)

In vivo work, perhaps in an animal model, may therefore be justified to test this effect. Success in animal models would mean that a trial of therapy in humans could be justified; however, a number of potentially serious problems exist if mycobacteria were to be introduced into the peritoneal cavity. The only mycobacterium currently in widespread use in bladder cancer, BCG, carries a small but definite mortality risk from BCG sepsis (Lamm, 1992Go). Whilst this risk may be acceptable in bladder neoplasia, it would not be acceptable for a condition such as endometriosis which is not fatal. Likewise, a risk of serious morbidity would be unacceptable for most women. BCG would lead to an inflammatory response, and to what extent this would be symptomatic in terms of acute pain and adhesion formation is uncertain. Information on the likelihood of adhesion formation would be available from the preliminary animal work discussed above.

It is therefore likely that BCG as currently used in bladder cancer would be unacceptable in endometriosis. In contrast, M. smegmatis is non-pathogenic in humans and may therefore be more acceptable. A variety of strategies have been proposed to increase the future efficacy of mycobacteria whilst at the same time reducing the chance of unwanted pathogenic side effects. Recombinant mycobacteria have been developed which can secrete a variety of cytokines, enhancing the immunotherapeutic effect (Jackson and Murphy, 1998Go). Recombinant M. smegmatis engineered to secrete tumour necrosis factor-{alpha} combines the advantages of increased therapeutic effect with the non-pathogenic nature of the mycobacteria (Haley et al., 1999Go). Suicide genes incorporated into pathogenic mycobacteria, or genes that induce sensitivity to certain antibiotics in case of possible sepsis might also be useful (O'Donnell, 1997Go).

Future developments in bladder cancer treatment may therefore translate into benefits for endometriosis sufferers if animal work were to show efficacy.


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Acien P, Quereda F, Campos A, Gomez-Torres MJ, Velasco I and Gutierrez M (2002) Use of intraperitoneal interferon alpha-2b therapy after conservative surgery for endometriosis and postoperative medical treatment with depot gonadotropin-releasing hormone analog: a randomized clinical trial. Fertil Steril 78, 705–711.[CrossRef][Medline]

Agarwala SS and Kirkwood JM (2000) Update on the role of adjuvant interferon for high risk melanoma. Forum (Genova) 10, 230–239.[Medline]

Alexandroff AB, Jackson AM, O'Donnell MA and James K (1999) BCG immunotherapy of bladder cancer: 20 years on. Lancet 353, 1689–1694.[CrossRef][Medline]

Ali AF, Fateen B, Ezzet A, Badawy H, Ramadan A and El-tobge A (2000) Laparoscopic intraperitoneal injection of human interferon-alpha2b in the treatment of pelvic endometriosis: a new modality. Obstet Gynecol 95, S47–S48.

Bohle A (2000) BCG's mechanism of action—increasing our understanding. For the EBIN Group. Eur Urol 37 Suppl 1, 1–8.

Bohle A, Thanhauser A, Ulmer AJ, Ernst M, Flad HD and Jocham D (1993) Dissecting the immunobiological effects of Bacillus Calmette-Guerin (BCG) in vitro: evidence of a distinct BCG-activated killer (BAK) cell phenomenon. J Urol 150, 1932–1937.[Medline]

Clayton RD, Duffy SR, Wilkinson N, Garry R and Jackson AM (2004) Anti-proliferative effect of mycobacteria, IFN-gamma and TNF-alpha on primary cultures derived from endometrial stroma: possible relevance to endometriosis? Am J Reprod Immunol 51, 63–70.[CrossRef][Medline]

Coligan JE, Kruisbeek AM, Margulies DH, Shevach EM and Strober W (1994) Assays for T cell function. In Coico R (ed.) Current Protocols in Immunology. John Wiley, New York, pp. 3.11.4–3.11.15.

D'Hooghe TM and Hill JA (1995) Killer cell activity, statistics, and endometriosis [letter; comment]. Fertil Steril 64, 226–228.

Gleicher N (1995) Immune dysfunction—a potential target for treatment in endometriosis. Br J Obstet Gynaecol 102 Suppl 12, 4–7.[Medline]

Gul A, Yasar T and Ugras S (2001) BCG vaccination to prevent implantation of endometriosis: an experimental study in rats. Eur J Obstet Gynecol Reprod Biol 98, 209–212.[CrossRef][Medline]

Haley JL, Young DG, Alexandroff A, James K and Jackson AM (1999) Enhancing the immunotherapeutic potential of mycobacteria by transfection with tumour necrosis factor-alpha. Immunology 96, 114–121.[CrossRef][Medline]

Hornstein MD, Hemmings R, Yuzpe AA and Heinrichs WL (1997) Use of nafarelin versus placebo after reductive laparoscopic surgery for endometriosis. Fertil Steril 68, 860–864.[CrossRef][Medline]

Ingelmo JM, Quereda F and Acien P (1999) Intraperitoneal and subcutaneous treatment of experimental endometriosis with recombinant human interferon-alpha-2b in a murine model. Fertil Steril 71, 907–911.[CrossRef][Medline]

Jackson AM and Murphy M (1998) The use of mycobacteria for the delivery of cytokines in gene therapy. Curr Res Mol Ther 1, 256–261.

Keenan JA, Williams-Boyce PK, Massey PJ, Chen TT, Caudle MR and Bukovsky A (1999) Regression of endometrial explants in a rat model of endometriosis treated with the immune modulators loxoribine and levamisole. Fertil Steril 72, 135–141.[CrossRef][Medline]

Lamm DL (1992) Complications of bacillus Calmette–Guerin immunotherapy. Urol Clin North Am 19, 565–572.[Medline]

Luo Y, Chen X, Downs TM, DeWolf WC and O'Donnell MA (1999) IFN-alpha 2B enhances Th1 cytokine responses in bladder cancer patients receiving Mycobacterium bovis Bacillus Calmette–Guerin immunotherapy. J Immunol 162, 2399–2405.[Abstract/Free Full Text]

Luo Y, Chen X, Han R and O'Donnell MA (2001) Recombinant bacille Calmette–Guerin (BCG) expressing human interferon-alpha 2B demonstrates enhanced immunogenicity. Clin Exp Immunol 123, 264–270.[CrossRef][Medline]

Noyes RW, Hertig AT, Rock J (1975) Dating the endometrial biopsy. Am JOG 122 (2), 262–263.

O'Donnell MA (1997) The genetic reconstruction of BCG as a new immunotherapeutic tool. Trends Biotechnol 15, 512–517.[CrossRef][Medline]

Thanhauser A, Bohle A, Flad HD, Ernst M and Mattern T ander AJ (1993) Induction of bacillus-Calmette–Guerin-activated killer cells from human peripheral blood mononuclear cells against human bladder carcinoma cell lines in vitro. Cancer Immunol Immunother 37, 105–111.[Medline]

Vercellini P, Crosignani PG, Fadini R, Radici E, Belloni C and Sismondi P (1999) A gonadotrophin-releasing hormone agonist compared with expectant management after conservative surgery for symptomatic endometriosis. Br J Obstet Gynaecol 106, 672–677.[Medline]

Vigano P, Di Blasio AM, Dell'Antonio G and Vignali M (1993) Culture of human endometrial cells: a new simple technique to completely separate epithelial glands. Acta Obstet Gynecol Scand 72, 87–92.[Medline]

Witz CA (1999) Current concepts in the pathogenesis of endometriosis. Clin Obstet Gynecol 42, 566–585.[CrossRef][Medline]

Submitted on January 5, 2004; accepted on May 6, 2004.





This Article
Abstract
Full Text (PDF )
All Versions of this Article:
19/8/1886    most recent
deh340v1
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Request Permissions
Google Scholar
Articles by Clayton, R.D.
Articles by Jackson, A.M.
PubMed
PubMed Citation
Articles by Clayton, R.D.
Articles by Jackson, A.M.