Dr H. Maia Jr, CEPARH, Rua Caetano Moura, 35, 40210-341 Salvador, Bahia, Brazil.
Objective To determine the presence of proteins related to proliferation (Ki-67) and apoptosis (Bcl-2, p53) in endometrial polyps and normal endometrium during the menstrual cycle.
Design Retrospective study using paraffin embedded tissue.
Setting Hospital affiliated to the university.
Population Premenopausal patients with endometrial polyps.
Methods Seventy-eight premenopausal patients in different phases of the menstrual cycle were submitted to polypectomy using the Bettocchi hysteroscope. Immunohistochemistry was used to detect the expression of these proteins in endometrial polyps. One hundred and eighteen normal endometrial biopsies were used as controls.
Main outcome measures Detection of Bcl-2 and Ki-67 expression by immunohistochemistry.
Results In endometrial polyps, Ki-67, p53 and Bcl-2 expression was detected with more frequency during the proliferative than during the luteal phase of the cycle. Similar findings were observed in the normal endometrium.
Conclusion Endometrial polyps undergo cyclic changes in the expression of their proteins related to proliferation and apoptosis during the menstrual cycle, similar to those of the cycling endometrium.
The prevalence of endometrial polyps reaches almost 15% during the premenopausal years and their presence is related to infertility and to alterations in the menstrual cycle.1,2 Although the gold standard for the treatment of endometrial polyps is polypectomy by hysteroscopy,3 there have been reports advocating other forms of treatment including expectant management as small endometrial polyps can regress spontaneously.4 This indicates that the natural history of these benign tumours may depend on the intricate balance between proliferation and apoptosis, which may be under the control of ovarian steroid hormones. In the cycling endometrium, proliferation and apoptosis are tightly regulated by the interplay between oestrogens and progesterone, with the latter down-regulating the expression of apoptosis-blocking proteins such as Bcl-2.5 Apoptosis is thought to play an important role in the process of implantation in the endometrium, a process that is orchestrated by progesterone.6 In endometrial polyps, little is known on the balance between apoptosis, proliferation rates and how ovarian hormones can affect this process. However, the observation that endometrial polyps can regress spontaneously, suggests that the post-ovulatory rise in progesterone could affect polyp growth by preventing proliferation while stimulating apoptosis in a similar way to that which occurs in the cycling endometrium.5,7
The present study was designed to investigate whether there are changes in the expression of the proteins related to proliferation, Ki-67, or apoptosis, Bcl-2 and p53, in endometrial polyps removed by hysteroscopy during different phases of the menstrual cycle, and to compare them with changes observed in the normal cycling endometrium.
Seventy-eight premenopausal patients with endometrial polyps diagnosed either by transvaginal sonography or sonohysterography were submitted to polypectomy using the Bettocchi hysteroscope (Karl Storz, Tuttligen, Germany) from March 2001 to November 2003. The procedure was carried out in different phases of the menstrual cycle and, in most cases, cervical dilatation was not required. All polypectomies were carried out under intravenous propofol, administered continuously with an automatic pump, and a paracervical block using 5% lidocaine. The endometrial polyp was sliced under visual control using the Versapoint (Gynecare, Somerville, New Jersey, USA) electrode (tweezer) with a cutting current of 100 W, VC1 mode, and the fragments removed with a specially designed alligator forceps (Karl Storz). At the end of the procedure, a biopsy of the surrounding endometrium was carried out using a 4-mm Karman curette attached to a 10-cc syringe to produce vacuum. Both the endometrium and the polyp were immediately fixed in formalin 4% and sent to pathology. Standard haematoxylin–eosin staining was carried out in tissue samples to confirm the diagnosis of polyp and to date the endometrium using standard histologic criteria. The histologic criteria for the diagnosis of polyp were the presence of non-functional glands, fibrous stroma and thick-walled blood vessels. These criteria were used in this study in order to prevent the inclusion of polypoid overgrowth of the luteal phase endometrium that might be confused with functional polyps. The first day of menstruation was noted and compared with endometrial dating. The patients were arbitrarily divided into four groups according to the histology of the endometrium and the day of the menstrual cycle: menstruation/early proliferative phase (days 1–8), late proliferative phase (days 9–16), early luteal phase (first four days after ovulation) and late luteal phase (>4 days following ovulation).
The presence of p53, Bcl-2 and Ki-67 expression was detected by immunohistochemistry in the glandular epithelium of endometrial polyps using monoclonal antibodies obtained through commercially available sources (Dako, Carpinteria, California, USA). To enhance antigen retrieval, the tissues were boiled in 10 nM citrate buffer hydrogen peroxide. The immunohistochemical detection of p53, Bcl-2 and Ki-67 was carried out using the streptavidin–biotin method. Tissue slides were examined by certified pathologists (AM and ES). The polyp was considered positive for p53 if the chromogen was detected in any cell nuclei. The negative polyps were those in which no immunostaining reaction was detected. The percentage of cell nuclei positive for Ki-67 was determined in the glands in 10 random microscopic fields (×25). The polyps were considered positive for Bcl-2 only when a strong reaction was detected in the glandular epithelium. Tissues with attenuated or null expression were analysed together and considered for statistical purposes to be negative.
Control samples were obtained from 118 patients with regular menstrual cycles who underwent hysteroscopy with endometrial biopsy as part of their infertility diagnostic work-up prior to IVF. None of the control biopsies came from uterine cavities that had any detectable abnormality at hysteroscopy. The biopsies were obtained using a 4-mm Karman curette attached to a 10-cc syringe to produce vacuum. The phase of the cycle was determined by the menstrual history and histologic dating of the endometrium. The patients were similarly divided into four groups and immunohistochemical detection of Bcl-2, p53 and Ki-67 expression was carried out using the methods described above.
Statistical analysis was carried out using Student's paired t test for differences between means and the χ2 test for differences between percentages, with significance at P < 0.05.
The mean age of the patients in the polyp group was 39 (7) years and in the control group was 33 (8) years. The majority of the patients with endometrial polyps had symptoms related to increased menstrual bleeding or intermenstrual spotting. Only 10% of the patients were asymptomatic and in these cases the endometrial polyp was discovered during routine transvaginal sonography.
The number of cell nuclei positive for Ki-67 in endometrial polyps and in the normal endometrium varied according to the phase of the menstrual cycle. During menstruation and in the early phases of the cycle, the percentages of positive cells in the glandular epithelium were low and not significantly different in either the endometrium or the polyp. However, the number of Ki-67 positive cells increased significantly (P < 0.01) in both polyp and endometrium throughout the proliferative phase, reaching peak values during the late proliferative phase when there were significantly more (P < 0.01) Ki-67 positive cells in endometrial polyps than in the normal cycling endometrium. Following ovulation, on the other hand, there was a rapid fall in the number of Ki-67 positive cells in the glandular epithelium of both polyp and endometrium, which was already significantly lower (P < 0.01) in the early luteal phase. During the late luteal phase, there was a further decrease in the number of Ki-67 positive cells in both tissues, reaching values close to zero (Fig. 1). In contrast to the late proliferative phase, there were no significant differences between polyps and the endometrium in the number of Ki-67 positive cells during this period. These results are summarised in Table 1.
Table 1. Percentage of Ki-67 positive cells in endometrial polyps and in the normal endometrium during the menstrual cycle. Values are presented as mean [SEM].
Phase of the cycle
Significantly higher than the other phases, P < 0.01.
Routine histology using haematoxylin–eosin staining revealed that while in the underlying endometrium the glandular epithelium showed signs of secretory transformation, in the polyp the glands remained non-functional despite the reduction in their proliferation rates.
Strong expression of Bcl-2 was detected by immunohistochemistry in both endometrial polyps and in the normal endometrium, and this expression was similarly affected by the phase of the menstrual cycle. In the first half of the menstrual cycle, from menstruation to the late proliferative phase, there was a significant increase in the number of endometrial polyps showing strong Bcl-2 expression in the glandular epithelium (P < 0.05). In the normal endometrium, on the other hand, there were no significant differences between the early and the late proliferative phases in the percentage of Bcl-2 positive cases (P= 0.3854). Following ovulation, however, there was a marked reduction in the intensity of Bcl-2 expression in both tissues and only 11% of the polyps and 7% of the endometrium remained strongly positive for this protein during the late luteal phase (Figs 2 and 3). The decrease in the percentage of Bcl-2 positive cases was already significant (P < 0.01) in both tissues in the first days following ovulation when compared with values found during the late proliferative phase. These results are shown in Table 2.
Table 2. Percentage of endometrial polyps and normal endometrium showing strong Bcl-2 expression during the menstrual cycle. Values are presented as n/n (%).
Phase of the cycle
Significantly higher than menstruation, P < 0.05.
Significantly higher than early and late luteal phase, P < 0.01.
Not significantly different from menstruation, P= 0.3854.
Both endometrial polyps and the cycling endometrium showed focal positive expression for p53 protein in the glandular epithelium. However, the percentage of endometrial polyps showing p53 expression in the glands increased during the proliferative phase of the menstrual cycle. In the late proliferative phase, 50% of polyps were positive for p53, which was significantly higher (P < 0.05) than during menstruation and the early proliferative phase. During the luteal phase, there was a trend towards a decrease in the number of polyps showing p53 expression but this did not reach statistical significance (P= 0.059). In the normal endometrium, p53 expression was detected in the glandular epithelium during both the proliferative and luteal phases of the menstrual cycle. Although there were no significant changes in p53 expression during the proliferative phase, there were, however, significantly fewer p53 positive cases during the late luteal phase (P < 0.01). These results are summarised in Table 3. In p53-positive polyps and in the endometrium, the expression was focal and usually affected very few cell nuclei in the glands (<10%) (Fig. 4).
Table 3. Percentage of endometrial polyps and normal endometrium showing p53 expression during the menstrual cycle. Values are presented as n/n (%).
Phase of the cycle
Significantly higher than the late luteal phase, P < 0.01.
Significantly higher than the early proliferative phase, P < 0.05.
The present study shows that both endometrial polyps and the normal endometrium undergo changes in the expression of proteins related to proliferation and apoptosis during the menstrual cycle. This supports previous observations showing that Bcl-2 expression increases in the cycling endometrium in response to oestrogens and this response is attenuated by progesterone leading to an increase in apoptosis at the end of the menstrual cycle.5,7 In this respect, endometrial polyps behave similarly to the underlying endometrium in terms of proliferation and Bcl-2 expression. Following ovulation and the rise in progesterone secretion by the corpus luteum, the glandular epithelium in the endometrium undergoes secretory transformation while in polyps the glands remain non-functional. This apparent non-responsiveness of endometrial polyps to progesterone in terms of glandular differentiation has been reported previously.8 Because they derive from a basalis type of endometrium, endometrial polyps are able to respond to oestrogens with proliferation but are unable to undergo differentiation and secretory transformation in response to progesterone.9 However, despite this lack of secretory differentiation in the glandular epithelium, there was a marked decrease in the expression of both Ki-67 and Bcl-2 during the luteal phase, which might favour the occurrence of apoptosis. The decrease in proliferation associated with the increase in apoptosis could leads in some cases to the spontaneous regression of endometrial polyps as previously reported.4 These findings also suggest that progestins could be effective in preventing the development of endometrial polyps. This has indeed been observed in patients using tamoxifen for breast cancer treatment in whom the concomitant use of an intrauterine device that releases levonorgestrel was associated with a marked reduction in the incidence of endometrial polyps in this high risk population.10 The effects of levonorgestrel on the endometrium are in part mediated by a decrease in the expression of Bcl-2 and Ki-67, and it is possible that this may prevent polyp formation11 through similar mechanisms. This might suggest the possibility of preventing or even treating endometrial polyps with the continuous use of progestins. However, the presence of breakthrough bleeding caused by progestin-only therapy in the presence of endometrial polyps may interfere with these results.12 The possibility, although rare in premenopausal patients, of the presence of occult carcinoma in polyps may also favour the removal of these lesions. Although there were no cases of carcinoma in our small series of premenopausal patients treated by polypectomy, they have been found in endometrial polyps of postmenopausal patients.13
Recently, it was reported that endometrial polyps showed stronger Bcl-2 expression than the normal endometrium during the proliferative phase and that this may result in failure of the polyps to undergo apoptosis during the late luteal phase of the menstrual cycle.14 The presence of a strong Bcl-2 expression in endometrial polyps was confirmed in the present article but in contrast to the previous findings, there were no differences between the normal endometrium and polyps either in intensity of expression or the percentage of positive cases during the proliferative phase. The decrease in Bcl-2 expression in the luteal phase was also similar in both tissues. In endometrial polyps removed from menopausal patients, Bcl-2 expression was likewise decreased by the use of progestins present in HRT regimens.15
Oestrogens are known to increase p53 expression in the endometrium of ovariectomised monkeys.16 In this study, 9/20 monkeys (45%) treated with conjugated oestrogens showed focal p53 positivity, a number similar to that reported in the present article for the normal endometrium during the late proliferative phase of the menstrual cycle. In vitro studies using cancer cells that express the wild type of p53 have shown that oestrogens can increase the levels of this protein by affecting gene transcription.17 However, it is not possible to conclude from these data whether the increased number of p53 positive endometrial polyps during the proliferative phase resulted from a direct effect of oestradiol on gene transcription or whether it was the consequence of the accumulation of DNA errors caused by the high rates of cell division, which in turn indirectly increased p53 intracellular levels in order to correct them or to induce apoptosis.18 In this context, it may be worth mentioning that one particular oestradiol metabolite, 2-methoxyestradiol, can increase p53 levels in lung cancer cells and this effect can enhance the therapeutic effect of radiation on this tumour.19 However, it remains to be established whether oestradiol or its metabolites may have similar effects on the endometrium, increasing p53 levels to induce apoptosis or correct DNA defects, thereby preventing the accumulation of genomic errors in the endometrium before the diminution in both Bcl-2 expression and proliferation occurs in response to progesterone.5,7 In fact, both mechanisms can act synergistically and may be responsible for the spontaneous regression of endometrial polyps observed in some cases.4 Heterogeneous and weak expression of p53 in the non-cancerous endometrium has been reported before in cases of endometrial hyperplasia20 and metaplasia.21 In both cases, the presence of p53 expression was thought to be a consequence of elevations of wild type p53 to correct DNA damage and not the accumulation of the stable mutant form.18
Recently, it was reported that tamoxifen but not HRT is a risk factor for the development endometrial polyps, which indicates that the progestins present in HRT regimens may play a role in suppressing their appearance.22 Furthermore, endometrial polyps in HRT users not only occur with less frequency than in controls,23 but are also smaller than those found in patients taking tamoxifen.24 These findings are in concordance with the results reported here that there is an increase in both Bcl-2 and Ki-67 expression in endometrial polyps during the proliferative phase of the menstrual cycle and a sharp decrease during the luteal phase.