The effect of hormone replacement therapy on the immunoreactive ...

Human Reproduction vol.15 no.1 pp.36–42, 2000

The effect of hormone replacement therapy on the immunoreactive concentrations in the endometrium of oestrogen and progesterone receptor, heat shock protein 27, and human β-lactoglobulin M.A.Habiba1, S.C.Bell1 and F.Al-Azzawi1,2,3 1Department of Obstetrics and Gynaecology, and 2Gynaecology Research Group, Faculty of Medicine and Biological Sciences, Leicester University, Leicester LE2 7LX, UK 3To

whom correspondence should be addressed at: Gynaecology Research Group, Faculty of Medicine and Biological Sciences, Leicester University, Leicester LE2 7LX, UK

We determined the expression of oestrogen receptor (ER), progesterone receptor (PR), heat shock protein 27 (HSP27) and human β-lactoglobulin in the endometrium under hormone replacement therapy (HRT). The immunohistochemical expression during the late progestogenic phase of sequential HRT was compared semi-quantitatively and using image analysis, to the early, mid-, and late luteal phase of the physiological cycle. Under sequential HRT, smaller glands were positive for the ER but larger glands with more advanced secretory features were negative. ER expression was lower in the stroma under HRT, and the difference was statistically significant compared with the early luteal phase (P < 0.05). Expression of HSP27 under HRT was lower in the epithelium but higher in the stroma compared with the physiological luteal phase. Epithelial PR expression was lower under HRT compared with the early, but not the mid- or the late luteal phase. The number of PR-positive stromal cells under HRT was lower compared with the physiological cycle, and the difference was statistically significant in comparison with the early luteal phase (P < 0.05). The glandular area expressing human β-lactoglobulin during the late progestogenic phase was statistically significantly higher compared with the early, but lower in comparison with the mid- or the late luteal phase (P < 0.05). The study demonstrates a subphysiological progestogenic response superimposed on evidence of a hypo-oestrogenism, and a differential response in the epithelium and stroma. Key words: endometrium/HSP27/hormone replacement therapy/ human β-lactoglobulin/steroid receptors

Introduction Cyclical oestrogen and progestogen administration to postmenopausal women aims to simulate the pre-menopausal hormone environment. We have previously demonstrated that the histological features of the endometrium of women using a sequential hormone replacement therapy (HRT) regimen differ significantly from those of the physiological cycle (Habiba et al., 1998). The exact cause of these differences 36

remains unknown but may be attributable to relative hypooestrogenism (Habiba et al., 1996), the type of progestogen used, or both (Habiba, 1998). Many markers have been employed to assess the hormonal influence in the endometrium; these include the level of oestrogen and progesterone receptors (ER and PR respectively), and the expression of heat shock protein 27 (HSP27) and human β-lactoglobulin [pregnancy-associated endometrial α2globulin (α2-PEG), placental protein 14 (PP14)]. However, no similar work has addressed this in relation to HRT. In the endometrium, both the concentration and the distribution of ER change during the normal menstrual cycle. These changes are predominantly under the influence of oestrogen and progesterone. There remains some disagreement on the exact pattern of expression, this particularly relates to whether the decline noted in the luteal phase commences at the beginning (Lessey et al., 1988; Amso et al., 1994) or the end (Garcia et al., 1988) of the early luteal phase. Stromal staining may either plateau (Amso et al., 1994) or decline (Lessey et al., 1988) from the time of ovulation into the late luteal phase. HSP27 (synonymous with p24 and p29), is a 27 kDa oestrogen-regulated phosphoprotein that may be involved in thermotolerance, protein degradation or as a molecular chaperone for other proteins including ER (Ciocca et al., 1993). The concentration of HSP27 is regulated by steroid hormones and cytokines. In the endometrial glandular epithelium, HSP27 appears during the late follicular phase and decreases after ovulation, and slightly rises premenstrually (Ciocca et al., 1983). This contrasts to the stronger expression in the luminal epithelium during the luteal phase (maximum around day 21). Expression in the glands is higher in conditions of excessive oestrogen stimulation as in hyperplasia (Ciocca et al., 1985). The stroma remains negative till the late luteal phase when the predecidual cells both around the spiral arterioles and beneath the superficial epithelium become positive (Ciocca et al., 1983). Stromal but not glandular epithelium expresses HSP27 under the influence of prolonged progestogen stimulation (Padwick et al., 1994). Oestrogen, through an intragenic oestrogen-responsive element (ERE), is the main stimulator of PR gene expression. Progesterone down-regulates PR through a protein–protein interaction between PR, ER and the same ERE (Savouret et al., 1991). In the mid-follicular phase (days 7–8), 25% of stromal and glandular cells are PR-positive. During the late follicular phase and the early luteal phase (days 9–19), the majority (75%) of glandular cells and half the stromal cells are positive. During the late luteal phase (days 21–27), PR disappears from the glands but remains faintly positive in the stroma (Garcia et al., 1988). © European Society of Human Reproduction and Embryology

ER, PR, HSP27, human β-lactoglobulin in HRT

Quantitatively, human glycosylated β-lactoglobulin (α2-PEG, PP14) is the major secretory protein product of the decidua and the major soluble product of the endometrium during the late luteal phase (Bell, 1990). Its function remains unknown, but it may have a role in maternal immune tolerance possibly through suppressing thymus-derived (T) and natural killer (NK) cells, and inhibiting interleukin-2 (IL-2)-mediated T cell proliferation (Seppala et al., 1994). Human β-lactoglobulin production is initiated on days 6–7 after the luteinizing hormone surge (LH⫹6 to LH⫹7) and rises thereafter to a maximum in the functionalis zone during the late luteal phase (Waites et al., 1988). Its concentration is related to the type of progesterone (levonorgestrel⬎medroxyprogesterone acetate), and is dependent on oestrogen pre-priming (Li et al., 1992). Despite a similar bleeding pattern, we have demonstrated significant histological differences between the late progestogenic phase endometrium from women on HRT and the endometrium of the late luteal phase, and whilst some of the observed features were shared with the early or mid-luteal phase others were not noted during the physiological luteal phase (Habiba et al., 1998). In this study, we aimed to investigate the discrepancy by measuring the level of ER, PR, HSP27, and human β-lactoglobulin as markers of hormone balance.

Immunohistochemistry (IHC) IHC was performed using the avidin–biotin complex immunolabelling method (Jackson and Blythe, 1993) with modifications. Briefly, sections were dewaxed in xylene, and rehydrated in grades of ethyl alcohol and distilled water. Sections were pretreated using enzyme digestion or microwave according to the antibody to be used (Table I). Endogenous peroxidase was blocked by freshly prepared 6% v/v hydrogen peroxide for 10 min. Sections were washed in tap water and Tris-buffered saline (TBS)–bovine serum albumin (BSA) buffer, and then covered in 100 µl of normal rabbit serum (NRS; Dako, Ely, Cambridgeshire, UK) at 1:10 in TBS (for anti-CD3 antibody, this was replaced with swine serum), and incubated for 20 min, these were then incubated with the primary antibody at the required dilution and time (Table I). Slides were washed in TBS–BSA and incubated for 30 min with the secondary antibody (100 µl of biotinylated antibody from a heterologous species) diluted at 1:150 in TBS–BSA. They were then washed in TBS–BSA and incubated with a freshly prepared solution Vectastain ABC peroxidase® (Vector Laboratories, Bretton, Peterborough, UK) for 30 min. Sections were washed and incubated with peroxidase substrate (DAB substrate, Vector Laboratories, UK) for 10 min. Slides were washed in running tap water, dried, dehydrated, cleared and mounted using XAM neutral medium (BDH, Poole, Dorset, UK). IHC for each of HSP27 and β-lactoglobulin were carried out in one batch, IHC for ER and PR were each divided into two batches with two sections duplicated for quality control. The inter-assay variation was 6%. Negative controls were used where the primary or the secondary antibody were missed.

Materials and methods

Image analysis This was performed using an image analysis system comprising a single chip colour video camera: Sony DXC-151P connected to the Sony CMA-151P camera adapter. This transmits the image to an Apple Macintosh® computer (Centris 650) via a RasterOps 24STV graphics display board (Rasterops Corporation, Santa Clara, CA, USA). This was analysed using the Colour Vision 1.7.4a program (Improvision, University of Warwick, Science Park, Coventry, UK). HSP27 or human β-lactoglobulin staining in the glands was measured by calculating the percentage of glands that stained positive in 17 low power fields (⫻100). Staining was evaluated by determining the proportion of tissue stained (P), multiplied by the intensity of staining (i), which was assigned scores, 0 ⫽ none, 1 ⫽ weak, 2 ⫽ distinct, and 3 ⫽ strong. The total score (T/score) was calculated using the equation: T/score ⫽ ΣPi (i ⫹ 1) (Lessey et al., 1988; Budwit-Novotny, 1986). This equation reduces the influence of a change in intensity on the overall score. Scoring was carried out by one of the investigators (M.A.H.) who had been blinded to the exact cycle phase. Stromal staining for HSP27 was expressed as the number of cases that were either negative, weak or positive. In order to measure the total secretory gland activity within the glandular unit, including luminal secretions, the total glandular area expressing β-lactoglobulin was measured in 17 high power fields (⫻200) per section (equivalent area of 1.59 mm2) and expressed in mm2 and as a percentage. The percentage of glands expressing ER and PR was counted in 17 random low power fields (⫻100)/section and the percentage of luminal epithelium expressing the receptor was also calculated in the same area. Receptor expression was measured assuming two levels of intensity above the preset threshold. The proportions of weakly stained cells were almost identical at 41, 38, 37 and 40% in the early, mid-, and late luteal phases and the late pseudoluteal phase respectively. Comparison thus yielded results that are almost identical to those calculated based on a single scale. We therefore present the

Women on sequential HRT received 2 mg of oestradiol valerate daily with the addition of 1 mg of norethisterone on days 17–28 of each treatment cycle (Solvay Pharma GmBH, Hannover, Germany). The inclusion and exclusion criteria, patient characteristics, and schedule have been reported previously (Habiba et al., 1998). Endometrial pipelle biopsies (n ⫽ 10) were obtained between days 27–29 and before the onset of bleeding in the sixth treatment cycle. These will be referred to as the late pseudoluteal phase endometrium. The LH-dated natural cycle endometrial biopsies were obtained by dilatation and curettage from healthy women with regular cycles at the time of scheduled tubal sterilization. They were all given a urine ovulation detection kit (Clearplan; Unipath, Basingstoke, Hampshire, UK) and instructed in its use. Sterilization and dilatation and curettage were performed in the early luteal phase (days 2–6 after the LH peak, n ⫽ 10), the mid-luteal phase (days 7–11 after the LH peak, n ⫽ 10), or the late luteal phase (days 12–14 after the LH peak, n ⫽ 10), with the day of LH peak being day 0. All endometrial samples were fixed in formalin and routinely processed and stained with haematoxylin and eosin to confirm normality and the LH dating. Quantifying ER and PR using image analysis was shown to have a high degree of agreement when compared to the dextran charcoalcoated (DCC) method (86.7 and 91.1% respectively) (Esteban et al., 1993). Percentage of positive area (PA ⫽ positive nuclei/total nuclei) measurement was used in this study as it was demonstrated to have a high degree of correlation with the more complicated percentage of positive stain (PS ⫽ summation of optical density of the positive nuclear area divided by the summation of the optical density of all the nuclei studied). Furthermore, the latter was shown not to improve the correlation between PA and DCC (Esteban et al., 1993). Thus image analysis, although semi-quantitative, combines an acceptable correlation with receptor content (correlation between the PA and DCC for ER, r ⫽ 0.71 (Esteban et al., 1993), with the added advantage of enabling antigen localization.

37

M.A.Habiba, S.C.Bell and F.Al-Azzawi

Table I. The source of antibodies used in this work and the method used in immunohistochemistry Primary antibody antigen

Supplier

Antibody type/clone

Ig isotype

Pre-treatment, time

Dilution, time, temp

Oestrogen receptor (ER)a Heat shock protein 27 (HSP27) Progesterone receptor (PR) β-lactoglobulin

Neo-markers Neo-markers Neo-markers Dr S.C.Bell

mAb mAb mAb mAb

IgG1/lambda IgG1 IgG1 None

Mv, 30 min Pep, 20 min, 37°C Mv, 30 min 1:2000, 2 h, 37°C

1:50, o/night, 4°C 1:100, 2 h, 37°C 1:50, o/night, 4°C

mouse mouse mouse mouse

anti-hum, AER314 anti-HSP27, G3.1 anti-hum, hPRa 3 anti-hum, Code 2CH11

aNot

specific to either type of receptor (ERα or ERβ). IgG ⫽ immunoglobulin G; CB ⫽ citrate buffer; hum ⫽ human; mAb ⫽ monoclonal antibody, Mv ⫽ microwave, o/night ⫽ overnight; pAb ⫽ polyclonal antibody; Pep ⫽ pepsin; RT ⫽ room temperature.

Table II. The percentage of the luminal epithelium that was oestrogen receptor (ER)-positive, and the percentage of glands that express ER in 17 low power fields (⫻100), and the number and percentage of stromal cells that express ER in 17 high power fields (⫻400). Values are given as numbers with (SD) and [range] Phase of the cycle

% Positive Luminal epithelium

Early-LP Mid-LP Late-LP Late-PLP

30 (40) [0–10] 22 (34) [0–80] 20 (34) [0–80] 42 (45) [0–100]

Positive stromal cells Glands 33 (40) [0–10] 20 (27) [0–60] 20 (30) [0–80] 40 (36) [0–80]

Stromal cells

Total number

No. corrected for densityb

47 (19) [38–56] 27 (26)a [0–60] 25 (29)a [0–76] 22 (25)a [0–67]

239 (148) [194–285] 116 (113) [0–252] 160 (185) [0–480] 86 (109)1 [0–245]

188 (116) [152–224] 109 (106) [0–238] 100 (116) [0–301] 88 (112)a [0–277]

LP ⫽ luteal phase; PLP ⫽ pseudoluteal phase. aStatistically significantly lower compared with the early-LP (P ⬍ 0.05) (Mann–Whitney U-test). bAs described in Habiba et al. (1998).

results using the percentage and the numerical scores (Tables II and IV). The number of stromal cells expressing ER or PR was counted using image analysis in 17 random high power fields (⫻400), and the number of positive cells was calculated, taking into account the difference in stromal cell density noted previously (Habiba et al., 1998). Statistical analysis was performed as described previously, using the Mann–Whitney U-test (Habiba et al., 1998).

Results Expression of ER ER expression in the physiological cycle was uniform within each section, and was stronger in the epithelium and the stroma of the early luteal phase compared with the mid- or late luteal phase. No staining was seen in the vascular endothelium. ER expression did not exhibit any site-specific preferential expression or any predilection depending on gland size or shape. The percentage of positive stromal cells in the early luteal phase was significantly higher compared with the midluteal phase and the late luteal phase (P ⬍ 0.05). In the late pseudoluteal phase, smaller glands were predominantly ER positive whilst the larger glands, which exhibited more advanced secretory features, were negative (Figure 1a). Some glands contained positive and negative cells, although the majority were uniformly either positive or negative. Staining in the surrounding stroma followed a similar pattern, with the stroma surrounding positive glands being predominantly positive, and vice versa. The higher expression 38

of ER in the glandular epithelium of the late pseudoluteal phase endometrium is attributable to higher expression in the more prevalent smaller glands. There was, however, a wide patient-to-patient variability. The only statistically significant differences between the late pseudoluteal phase and the early luteal phase were in the number and the proportion of positive stromal cells (P ⬍ 0.05) (Table II). Expression of HSP27 HSP27 expression was strong in the luminal epithelium throughout the luteal phase, but the predominantly HSP27-positive epithelium was interrupted by HSP27-negative segments. At some gland openings, positive luminal and glandular epithelium appeared continuous. The overall expression in the glandular epithelium was very weak, and positive and negative glands were scattered in the stroma and only rarely appearing in clusters. There was no discernible preferential expression in relation to gland size or location. Semi-quantitative assessment demonstrated higher HSP27 in the early-luteal phase compared with the late-luteal phase in both the luminal and the glandular epithelium, but the differences were not statistically significant. The stroma was predominantly negative during the early luteal phase and the mid-luteal phase, but HSP27 expression was higher in the late luteal phase (Figure 1b). The pattern of expression in the luminal epithelium in the late pseudoluteal phase was similar, but the HSP27-negative segments were larger and this accounted for the weaker overall expression. These differences were statistically significant


Figure 1. The expression during the late pseudoluteal phase of (a) the oestrogen receptor (ER). The large gland (bottom right quadrant) shows negative staining, whereas two smaller glands (above and left) are positive; (b) heat shock protein 27 (HSP27); (c) human β-lactoglobulin; and (d) the expression of human β-lactoglobulin during the late luteal phase. Scale bar ⫽ 0.05 mm.

compared with the physiological cycle (P ⬍ 0.05). Glandular staining followed the same pattern as in the physiological cycle, although staining was very weak. Expression of HSP27 was lower in glands in the late pseudoluteal phase compared to all stages of the luteal phase, but the difference was statistically significant only in comparison to the early luteal phase (P ⬍ 0.05). There was no preferential expression of HSP27 in relation to glandular size or distribution. The late pseudoluteal phase exhibited the highest stromal HSP27 expression (Table III). Expression of PR PR expression in the glandular and luminal epithelium was stronger during the early luteal phase and gradually regressed and almost disappeared during the mid- and late luteal phase. The stroma remained weakly positive during the mid- and late luteal phase. The vascular endothelium was negative. The late pseudoluteal phase exhibited little PR in the glandular and luminal epithelium but stronger staining in the stroma. PR expression was statistically significantly higher in the luminal and the glandular epithelium during early luteal phase compared with the mid- and the late luteal phases (P ⬍ 0.05) (Table IV). Expression of PR in the glandular and the luminal

epithelium during the late pseudoluteal phase was lower compared with the early luteal phase (P ⬍ 0.05), and although it was higher compared to the mid- or the late luteal phases, the difference was not statistically significant. During the late pseudoluteal phase, PR expression was higher in some of the smaller size glands, but the smallest glands that were ER-positive were PR-negative. The number of PR-positive stromal cells during the late pseudoluteal phase was lower compared with all stages of the physiological cycle, but the difference was statistically significant only when compared with the early luteal phase (P ⬍ 0.05). Expression of human β-lactoglobulin Glandular and luminal human β-lactoglobulin was only rarely detected during the early luteal phase. This peaked during the mid- and late-luteal phase where most glands were positive. There was no staining in the stroma. The difference between the early luteal phase and both the mid- and late luteal phase was statistically significant (P ⬍ 0.05). During the late pseudoluteal phase the majority of the glands expressed human β-lactoglobulin (Figure 1c), but a higher proportion (compared with the mid- and late luteal phases), especially of the smaller 39


Table III. The total score (T/score) of heat shock protein 27 (HSP27) expressed in the luminal and the glandular epithelium in 17 low power fields (⫻100), and the number and staining intensity of the stroma in 17 high power fields (⫻400) Phase of the cycle


Luminal T/score

174 170 151 103

(47) (43) (76) (38)a

Glandular T/score

91 60 58 43

(75) (57) (67) (43)b

Stroma Negative

Weak

Positive

14/17 14/17 9/17 0/17

3/17 3/17 3/17 0/17

0/17 0/17 5/17 17/17

significantly lower compared with the early-, mid-, and late-luteal phase (LP) (P ⬍ 0.05; Mann–Whitney U-test). significantly lower compared to the early-LP (P ⬍ 0.05; Mann–Whitney U-test).

aStatistically

bStatistically

Table IV. The percentage of the luminal epithelium that was PR positive, and the percentage of glands that express PR in 17 low power fields (⫻100), and the number and percentage of stromal cells that express PR in 17 high power fields (⫻400). Values are given as numbers with (SD) and [range] Phase of the cycle


% Positive

Positive stromal cells

Luminal epithelium

Glands

Stromal cells

Total number

No. corrected for densityb

47 (48)a [20–100] 6.7 (9.8) [0–20] 8.2 (13) [0–40] 20 (25) [0–70]

55 (48)a [20–100] 12 (19) [0–55] 15 (21) [0–60] 18 (29) [0–75]

47.7% (16) [35–75] 48.5% (18) [27–75] 34% (9.7) [22–52.5] 42.5% (16) [22.5–80]

242 (81) [176–383] 205 (87) [135–365] 217 (62) [108–329] 167 (65)b [91–313]

190 (64) [139–302] 194 (82) [127–343] 136 (38)2 [68–207] 171 (66) [93–320]

aStatistically

significantly higher than the mid-luteal phase (LP), late-LP, and the late pseudoluteal phase (PLP) (P ⬍ 0.05) (Mann–Whitney U-test). bStatistically significantly lower than the early-LP (P ⬍ 0.05) (Mann–Whitney U-test).

Table V. The total glandular area expressing β-lactoglobulin during the different stages of the luteal phase (LP) and the late-pseudoluteal phase (PLP), expressed as the percentage of area stained per section, and in mm2/17 high power fields (⫻200), equivalent area of 1.59 mm2. Values are given as numbers with (SD) and [range]. The total score (T/score) for β-lactoglobulin expression in the luminal and the glandular epithelium in 17 low power fields (⫻100) Phase of the cycle

Percentage of total area

mm2/1.59mm2

Luminal T/score

Glandular T/score

Early-LP

0.06 (0.82)a [0.0–1.89] 8.57 (4.6) [0.077–14.7] 7.68 (3.34) [2.02–12.8] 3.02 (1.2)c [0.52–4.8]

0.01 (0.013)a [0.02–0.03] 0.136 (0.073) [0.001–0.23] 0.122 (0.053) [0.032–0.204] 0.048 (0.019)c [0.008–0.076]

14 (13)

33 (53)a

49 (63)

140 (60)

68 (48)b

120 (61)

22 (11)

60 (40)a

Mid-LP Late-LP Late-PLP aStatistically aStatistically

U-test). cStatistically

significantly lower compared with the mid- and late-LP (P ⬍ 0.05; Mann–Whitney U-test). significantly higher compared with the early-LP and the late-PLP (P ⬍ 0.05; Mann–Whitney significant difference from the early-, mid-, and late-LP (P ⬍ 0.05; Mann–Whitney U-test).

glands were negative. The glandular area expressing human β-lactoglobulin during the late pseudoluteal phase was significantly higher compared with the early luteal phase, but lower compared with the mid-luteal phase or the late luteal phase (P ⬍ 0.05) (Figure 1c,d; Table V). 40

Discussion In this study, the pattern of ER expression observed in the endometrium during the physiological cycle was similar to that described in the literature. In agreement with previous observations, endothelial cells were found to be ER negative


(Lessey et al., 1988; Perrot-Applanat et al., 1988), but the two-fold increase in ER expression in the glandular epithelium during the late pseudoluteal phase, compared with the late luteal phase although it was not statistically significant, was mostly accounted for by the higher incidence of smaller glands lacking secretory features. The expression of ER amidst apparent retardation of glandular development suggests inadequate oestrogen priming, with glands failing to reach a stage where they can respond optimally to progestogens. This hypothesis is supported by the finding of a low HSP27 expression in the luminal glandular epithelium during the late pseudoluteal phase, which (as discussed below) reflects poor oestrogenic response. On the other hand, stromal cells express low levels of ER, but exhibit high HSP27 expression, which may indicate adequate progestogenic response. This discrepancy may be explained by the selective differential effect of oestrogens and progestogens on the glands and stroma, and by the observation that a dose or preparation of progestogen that is adequate for one compartment may not be so for the other (King et al., 1981). For example, 19-nortestosterone derivatives used in oral contraceptives, have been shown to induce relatively more stromal decidualization and gland atrophy compared with progesterone derivatives (DallenbachHellweg, 1987). The observation that some of the smaller diameter glands strongly expressed ER during the late pseudoluteal phase whilst adjacent glands were ER-negative, may be related to a threshold effect of steroids during gland recruitment into proliferation and differentiation. The expression of HSP27 in the endometrial glandular epithelium was found to be lower during the mid-progestogenic phase in women treated with sequential HRT in comparison with the endometrium during the physiological proliferative phase, and to be similar to that observed during the mid-luteal phase of the physiological cycle (Padwick et al., 1994). The expression of HSP27 may be inhibited by progestogens, and this inhibition may be type- or dose-related (Padwick et al., 1994). Thus, the low values of HSP27 in the glandular epithelium during the late pseudoluteal phase observed in our study suggest, in the absence of other evidence of progestogenic influence, low induction rather than adequate inhibition. It was argued that the late expression of human β-lactoglobulin (luteal day 5) suggests that it is not directly dependent on either oestrogen or progesterone (Manners, 1990). Factors known to affect serum levels of human β-lactoglobulin include the anti-oestrogen tamoxifen, the antiprogestin RU486, clomiphene citrate, HCG and the GnRH analogue buserelin (Seppala et al., 1994). Glucocorticoid/progesterone regulatory elements (PREs) have been found in the human β-lactoglobulin gene, and these elements bind purified PR in vitro (Vaisse et al., 1990). It is therefore possible that factors affecting human β-lactoglobulin secretion (other than progestogens) may act by modifying PRE response, with human β-lactoglobulin secretion reflecting the balance of these factors. In sequential HRT specimens human β-lactoglobulin production was reduced, despite evidence of stromal decidualization (Habiba et al., 1998), thus emphasizing the dissociation between stromal and gland development, and between gland development and menstrual function.

There appears to be an inverse relationship between the expression of PR and human β-lactoglobulin in both the physiological cycle and the late pseudoluteal phase. Semiquantitative expression of human β-lactoglobulin demonstrated that the late pseudoluteal phase cannot be compared with either stage of the luteal phase, expression being higher than in the early luteal phase, but lower than in the mid- or lateluteal phases. This may indicate a suboptimal progestogenic response, which may be secondary to a suboptimal oestrogenic induction of PR, or a function of the type or dose of progestogen used, or both. The higher ER expression in the glands and tendency to lower expression in the stroma suggest adequate progestogenic effect in the latter. This is supported by the finding of normal levels of PR during the late pseudoluteal phase. This, however, does not rule out the possibility of low initial levels of PR. These findings suggest adequate progestogenic response on the stroma. Although the observation that some glands are ER-positive, but PR-negative, suggests that suboptimal PR expression in the glands may have contributed to the low human β-lactoglobulin expression. A suboptimal progestogenic response may be due to the overall low level of receptor. This is supported by the finding of a difference in the level of PR but not in the PRA/PRB ratio in relation to the differential effect of progesterone and its antagonists on fibroids (Viville et al., 1997). However, an alteration of the PRA/PRB balance cannot be ruled out. This study demonstrates a suboptimal progestogenic response in this particular sequential HRT preparation that incorporates a progestogen with relatively strong androgenic properties and a high affinity to PR, and it is possible that the hypo-oestrogenism is secondary to relatively higher anti-oestrogenic properties. The differences observed in the oestrogenic and progestogenic effects on the endometrium may reflect the non-phasic administration mode of these steroids in this regimen, as is the case with all currently available HRT regimens. Alternatively, it may reflect the serum concentration of the steroids. This study did not address this question and further research is needed. The value of urinary LH home detection kits in relation to infertility treatment is subject to debate (Robinson et al., 1992; Anderson et al., 1996), and its accuracy may be reduced in infertile women with a high endogenous LH (Anderson et al., 1996). It has, however, been shown to be highly sensitive and specific in women with normal cycles and no history of infertility (Schmutzler et al., 1995; Miller and Soules, 1996). Furthermore, we used histological assessment in order to confirm LH-dating in our study group. The difference in the level of expression of functional markers between the physiological cycle and women on sequential HRT, despite similar bleeding pattern (Habiba et al., 1996), demonstrates dissociation between endometrial behaviour and functional markers. This parallels the observed dissociation between ER and PR and the amount of blood loss (Critchley et al., 1994), and may be a function of the role of second messengers such as cAMP, protein kinase A, protein kinase C, activator protein-1 (AP-1), insulin-like growth factor-I (IGF-I), epidermal growth factor (EGF), and the proto-oncogene erb-2 ligand, glycoprotein 30 (gp30), in controlling or modifying the action of 41


steroids (Katzenellenbogen, 1996; Saceda, 1996). Nuclear receptor co-regulators and co-repressors modulate ligandinduced transcription (McKenna et al., 1999). In relation to ER, oestrogen and raloxifene have been shown to induce distinct conformational changes in the transactivation domain of the ligand-binding domain (Brzozowski et al., 1997; Grese et al., 1997), which may affect the type of co-activator(s) recruited. This provides a putative mechanism for the differential effect of agonists and antagonists binding to the steroid receptor. Similarly, a differential binding of coactivators is a putative mechanism for the discrepant effect of norethisterone, and it is possible that pregnane and 19norpregnane derivatives might have different molecular effects. Recently, it was suggested that the 19-nortestosterone derivative, levonorgestrel, may induce its progestational effects through non-receptor-mediated mechanisms (Salmi et al., 1998). These observations highlight the need for further studies to clarify the molecular pathways of progestogen action.

References Amso, N.N., Crow, J. and Shaw, R.W. (1994) Comparative immunohistochemical study of oestrogen and progesterone receptors in the Fallopian tube and uterus at different stages of the menstrual cycle and the menopause. Hum. Reprod., 9, 1027–1037. Anderson, R.A., Eccles, S.M. and Irvine, D.S. (1996) Home ovulation testing in a donor insemination service. Hum. Reprod., 11, 1674–1677. Bell, S.C. (1990) Assessment of endometrial differentiation and function. Br. Med. Bull., 46, 720–732. Brzozowski, A.M., Pike, A.C.W., Dauter, Z. et al. (1997) Molecular basis of agonism and antagonism in the oestrogen receptor. Nature, 389, 753–757. Budwit-Novotny, D.A., McCarty, S.K.S., Cox, E.B. et al. (1986) Immunohistochemical analyses of estrogen receptor in endometrial adenocarcinoma using a monoclonal antibody. Cancer Res., 46, 5419–5425. Ciocca, D.R., Asch, R.H., Adams, D.J. and McGuire, W.L. (1983) Evidence for modulation of a 24K protein in human endometrium during the menstrual cycle. J. Clin. Endocrinol. Metab., 57, 496–499. Ciocca, D.R., Puy, L.A., Edwards, D.P. et al. (1985) The presence of an estrogen-regulated protein detected by monoclonal antibody in abnormal human endometrium. J. Clin. Endocrinol. Metab., 60, 137–143. Ciocca, D.R., Oesterreich, S., Chamness, G.C. et al. (1993) Biological and clinical implications of heat shock protein 27000 (Hsp27): a review. J. Natl Cancer Inst., 85, 1558–1570. Critchley, H.O.D., Abberton, K.M., Taylor, N.H. et al. (1994) Endometrial sex steroid receptor expression in women with menorrhagia. Br. J. Obstet. Gynaecol., 101, 428–434. Dallenbach-Hellweg, G. (1987) Histopathology of the Endometrium, 4th edn. Springer-Verlag, Berlin, pp. 1–25. Esteban, J.M., Kandalaft, P.L., Mehta, P. et al. (1993) Improvement of the quantification of estrogen and progesterone receptors in paraffin-embedded tumors by image analysis. Am. J. Clin. Pathol., 99, 32–38. Garcia, E., Bouchard, P., De Brux, J. et al. (1988) Use of immunocytochemistry of progesterone and estrogen receptors for endometrial dating. J. Clin. Endocrinol. Metab., 67, 80–87. Grese, T.A., Sluka, J.P., Bryant, H.U. et al. (1997) Molecular determinants of tissue selectivity in estrogen receptor modulators. Proc. Natl. Acad. Sci. USA, 94, 14105–14110. Habiba, M. (1998) Endometrial responses to hormone replacement therapy. PhD thesis, University of Leicester. Habiba, M., Bell, S.C., Abrams, K. and Al-Azzawi, F. (1996) Endometrial responses to hormone replacement therapy: the bleeding pattern. Hum. Reprod., 11, 503–508. Habiba, M., Bell, S.C. and Al-Azzawi, F. (1998) Endometrial responses to hormone replacement therapy: histological features compared with those of late luteal phase endometrium. Hum. Reprod., 13, 1674–1682. Jackson, P., and Blythe, D. (1993) Immunolabelling techniques for light microscopy. In Beesley J.E. (ed.), Immunocytochemistry. A Practical Approach. IRL Press, Oxford, UK, pp. 15–42.

42

Katzenellenbogen, B.S. (1996) Estrogen receptors: bioactivities and interactions with cell signaling pathways. Biol. Reprod., 54, 287–293. King, R.J.B., Townsend, P.T., Whitehead, M.I. et al. (1981) Biochemical analyses of separated epithelium and stroma from endometria of premenopausal and postmenopausal women receiving estrogen and progestins. J. Steroid Biochem., 14, 979–987. Lessey, B.A., Killam, A.P., Metzger, D.A. et al. (1988) Immunohistochemical analysis of human uterine estrogen and progesterone receptors throughout the menstrual cycle. J. Clin. Endocrinol. Metab., 67, 334–340. Li, T.C., Dalton, C., Bolton, A.E. et al. (1992) An analysis of the variation of plasma concentrations of placental protein 14 in artificial cycles. Fertil. Steril., 57, 776–782. Manners, C.V. (1990) Endometrial assessment in a group of infertile women on stimulated cycles for IVF: immunohistochemical findings. Hum. Reprod., 5, 128–132. McKenna, N.J., Lanz, R.B. and O’Malley, B.W. (1999) Nuclear receptor coregulators: cellular and molecular biology. Endocr. Rev., 20, 321–344. Miller, P.B. and Soules, M.R. (1996) The usefulness of a urinary LH kit for ovulation prediction during menstrual cycles of normal women. Obstet. Gynecol., 87, 13–17. Padwick, M.L., Whitehead, M. and King, R.J. (1994) Hormonal regulation of HSP27 expression in human endometrial epithelial and stromal cells. Mol. Cell. Endocrinol., 102, 9–14. Perrot-Applanat, M., Groyer-Picard, M.T., Garcia, E. et al. (1988) Immunocytochemical demonstration of estrogen and progesterone receptors in muscle cells of uterine arteries in rabbits and humans. Endocrinology, 123, 1511–1519. Robinson, J.N., Lockwood, G.M., Dalton, J.D. et al. (1992) A randomized prospective study to assess the effect of the use of home urinary luteinizing hormone detection on the efficiency of donor insemination. Hum. Reprod., 7, 63–65. Saceda, M., Grunt, T.W., Colomer, R. et al. (1996) Regulation of estrogen receptor concentration and activity by an erbB/HER ligand in breast carcinoma cell lines. Endocrinology, 137, 4322–4330. Salmi, A., Pakarinen, P., Peltola, A.-M., and Rutanen, E.-M. (1998) The effect of intrauterine levonorgestrel use on the expression of c-JUN, oestrogen receptors, progesterone receptors and Ki-67 in human endometrium. Mol. Hum. Reprod., 4, 1110–1115. Savouret, J.F., Bailly, A., Misrahi, M. et al. (1991) Characterization of the hormone responsive element involved in the regulation of the progesterone receptor gene. EMBO J., 10, 1875–1883. Schmutzler, R.K., Diedrich, K., and Kerbs, D. (1995) Value of the Clearplan Ovulation Test in sterility treatment. [In German.] Geburtshilfe Frauenheilk., 55, 266–269. Seppala, M., Koistinen, R., and Rutanen, E.-M. (1994) Uterine endocrinology and paracrinology: insulin-like growth factor binding protein-1 and placental protein-14 revisited. Hum. Reprod., 9, 917–925. Vaisse, C., Atger, M., Potier, B. and Milgrom, E. (1990) Human placental protein 14 gene: sequence and characterization of a short duplication. DNA Cell Biol., 9, 401–413. Viville, B., Charnock-Jones, D.S., Sharkey, A.M. et al. (1997) Distribution of the A and B forms of the progesterone receptor messenger ribonucleic acid and protein in uterine leiomyomata and adjacent myometrium. Hum. Reprod., 12, 815–822. Waites, G.T., Wood, P.L., Walker, R.A. and Bell, S.C. (1988) Immunohistological localization of human endometrial secretory protein, ‘pregnancy-associated endometrial α2-globulin’ (α2-PEG), during the menstrual cycle. J. Reprod. Fertil., 82, 665–672. Received on June 1, 1999; accepted on October 12, 1999