Research on genesis of adipocytic metaplasia in uterine fibroids

Introduction

Uterine fibroids with adipocytic metaplasia are rare in women. According to our previous studies they appeared in 1.33% of all performed hysterectomies [1].
The histogenesis of these changes arouses controversy [2, 3, 4]. Immunohistochemically, the examined lipocytes always yielded a positive reaction with S-100 antibody [5, 6], and a negative reaction with muscle antibodies [5, 6, 7]. According to Aung et al. [4], the majority of the lipomatous component in the uterine fibroids yielded a negative reaction with SMA antibodies, but in some parts lipocytes showed SMA positivity. There were fewer estrogen and progesterone receptors in lipoleiomyomas than in surrounding uterine corpus muscle [8, 9], which may indicate the role of hormonal disorders in their histogenesis.
In current research we decided to follow the development of adipocytic metaplasia process in uterine fibroids depending on: 1) the age of patients, 2) the type of the primary tumor stroma in which the process of metaplasia took place, 3) the participation of estrogen and progesterone receptors in the primary tumor.

Material and methods

The archive material originating from 17 uteri with previously diagnosed foci of adipocytic metaplasia was used in the examination. Age of patients ranged between 43 and 82 years, with the average age of 51.0 ±9 years.
Paraffin sections were cut at 4 µm. In addition to hematoxylin and eosin staining, sections were stained with the Mallory method and with toluidine blue. Moreover, in selected cases, they were stained with periodic acid – Schiff (PAS), with alcian blue and with Grocott-Gomori methenamine silver staining. Immunohistochemical reactions were performed on all the specimens numbered 1-8 presented in Table I, and also 10 times with CD68 and NSE, 9 times with Ki-67, and 7 times with WT1, CK 5/6 and FVIII.
Reactions to thrombomodulin, calretinin and vimentin were conducted 5 times. HMB 45 staining was conducted four times.
Four types of myomas were classified from the material: 1) pure leiomyoma, 2) fibroleiomyoma, 3) hyalinizing leiomyoma, 4) strongly hyalinized myoma.
The degree of adipocytic metaplasia was determined as: 1) if the percentage of lipocytes in the histological specimen was 25%, 2) if the percentage of lipocytes was between 25% and 50%, 3) if the percentage of lipocytes was over 50%.
Immunohistochemical reactions were conducted according to standards of the local laboratory. Paraffin sections were placed on silanized slides (DAKO cat. S 3003), deparaffinized in xylene and then passed through a descending alcohol series into water.
Antibodies of tissues fixed in formalin were discovered in Target Retrieval Solution (DAKO cat. 3307) by heating at 96 degrees for 20 minutes. Endogenous peroxidase was blocked in 3% H₂O2 solution for 10 minutes. Primary antibodies were placed on paraffin sections. Sections were rinsed in TBS. The visualization system En Vision + HRP (DAKO cat. 4007) was put on sections, which were later incubated for 30 minutes at room temperature. The immunohistochemical reaction was carried out with a solution of 3.3’-diaminobenzidine tetrachloride (DAB, liquid K 3486 DAKO). Next, the sections were rinsed in running water and graded alcohol dehydration was performed. The material was then exposed in xylene and encased in balm. The evaluation of staining intensity was half-quantitative according to the following gradation: 0 – lack of positive reaction, 1 – weak reaction, 2 – strong reaction. The final evaluation of the separate reactions was based on the average taken from the three readings carried out at monthly intervals. In this way, the results (in figures) were subject to statistical analysis with the Statistica 9 program. Differences at the p ≤ 0.05 level were considered significant.

Results

The results of the study are depicted in Tables II-IV. According to Table II, 41% were pure myomas, 23.5% were fibroleiomyomas and 17% were hyalinizing leiomyomas and strongly hyalinized myomas.
58.8% of tumors exhibited adipocytic metaplasia in early phases of development: 35.3% in a semi- advanced phase, and 5.9% in a very advanced phase.
Table III and Figure 1 show that the dynamism of adipocytic metaplasia in myoma was positively correlated with the individual age of patients (r = 0.514, p = 0.035). This dynamism depended also on the activity of the estrogen receptor in myoma (r = 0.53, p = 0.030), but it was not correlated with the activity of the progesterone receptor, nor with the type of regressive changes in the primary tumor (Table IV, Figure 2). Adipocytes presented a close relation to blood vessels (Fig. 3). Some delicate connective tissue component around lipocytes was always reported in staining with the Mallory method and in the reaction with vimentin, even in “full” myomas. Cells of this type were always S-100 positive (Fig. 4). Inflammatory infiltrate composed of acidophilic granulocytes (Fig. 5) was sometimes observed around blood vessels and lipocytes. Mast cells were more visible after reaction with CD117 than after staining with toluidine blue. Mast cells were almost always present in uterine stroma and around lipocytes. Reaction with antibody CD68 was strongly positive in 8.3% of cases (Fig. 6). An identical proportion to this reaction was however weakly positive. In fibrous and hyaline masses of the primary tumor, we observed the spontaneous creation of lipocytes with SMA positive circumference of those cells (Fig. 7) and strongly positive reaction for S-100 antibody (Fig. 8). Positive reaction with Ki-67 antibodies (Fig. 9) was reported twice in primary tumor (11.8%) and once in lipocytes (5.9%). In the examined material we reported that in the case of two women (11.8%) young mucous tissue developed out of fibrotic myoma masses on which formed lipocytes. These changes took place in close connection with capillaries (Figs. 10, 11). In myomas in three women (17.65%) (women numbered 8, 15, 16) under the endothelium of large blood vessels the appearance of polygonal cells with a centrally located nucleus and a large amount of amphoteric granules (Figs. 12, 13) was reported. These cells were always SMA positive (Figs. 14-16). They were proliferating strongly and moving the endothelium to the vessel lumen, closing it more than once (Figs. 15, 16). Among other reactions, these cells were CD68 positive (Fig. 17), they reacted poorly with toluidine and S-100, they did not stain with desmin (Fig. 18), and they gave a negative reaction with CD117 and HMB 45. Reactions with calretinin were carried out 5 times. In 40% of the cases, the reaction was positive in isolated cells of the main tumor (Fig. 19). Once there was intensive staining of the lipocyte circumference in a 61-year-old woman numbered 17 (Fig. 20).
Reactions with vimentin were always positive, with antibody D2-40 always negative. Reactions with WT1, CK5/6, thrombomodulin and NSE were always negative.
Summing up, immunohistochemical staining in lipocytes varied depending on the tumors.
The reaction with actin antibody (group average of 1.08 points) was the strongest, while the reaction with S-100 (0.97 points) and desmin (0.68 points) was weaker.

Discussion

In 1955 Brandfass and Everts-Suarez [10] quoted from academic literature 7 theories concerning the origin of uterine lipoleiomyomas. They also wondered whether pure forms of lipomas had the same genesis as mixed forms. According to the authors, pure forms may originate in stray mesenchymal cells of stroma, and mixed forms appear as a result of fat degeneration of myocytes. Similar suggestions were made by Meinhof and Bersh [11]. However, they were contradicted by Salm [12]. According to this author, lipoleiomyomas were never observed in children and adolescents. On the other hand, mixed forms appear usually before menopause, and pure forms after menopause. Similar observations were made in our previous work [1]. It may be concluded from the above data that mixed forms, constantly growing, eventually lead to the appearance of “pure” forms.
In our current research we observed the development of lipocytes in the connective tissues situated close to the blood vessels. Negative results with D2-40, WT1 and with other antibodies [13, 14] ruled out the possibility that lymphatic vessels and mesothelium of peritoneum took part in their development.
According to other authors [2, 15], immunohistochemical reactions carried out in lipoleiomyomas gave various results, staining uterus myocytes SMA positive and adipocytes S-100 positive. No transitory forms have been observed between myocytes and lipocytes [3], which provided grounds to reject the hypothesis that the developed lipocytes in uterine fibroids were of a degenerative nature and instead supported the hypothesis of them being cancerous in nature [12]. It is assumed that there are mesenchymal totipotential pericapillary cells [2, 16, 17], with numerous possibilities of diversification, including those necessary to create lipocytes [2, 3, 9], which eventually lead to the occurrence of adipocytic neometaplasia [9, 15] and eventually to the development of lipoma in the uterus [12]. Lipocytes developed this way are characterized by a high proliferative index Ki-67, which supports their cancerous, not degenerative nature [4, 15]. Cytogenetic examinations of lipoleiomyomas in the uterus indicated the possibility of different components of the primary tumor existing [18, 19]. That is why the type of components found in uterine fibroids depends only on the accuracy of research and on the number of sampled tissues of the main tumor [20].
Ultrastructural examination of myxomas revealed that myofilaments are present in myxoma cells [17]. They were particularly exposed in cells close to the vessel wall. The results of our research were compatible with the data cited above from the academic literature. Spontaneous development of lipocytes in the stroma of fibromatous, or even hyalinized myoma, is a surprising phenomenon. These cells were both SMA and S-100 positive. The creation of mucous tissue around the vessels and based on it formation of lipocytes was particularly interesting. Positive reaction with Ki-67 antibodies in myoma cells as well as in lipocytes of our material confirms the suggestions of the academic literature about the cancerous character of formed lipocytes. Granulated subendothelial cells of large blood vessels, containing both muscle antigens and antigens which characterize phagocytic cells, are the second important element we observed in our examination. There is a question whether they represent the inflammatory reaction to the creation of regressive changes in the wall of blood vessels, or they are primal changes, reverting from myocytes to lipocytes as mentioned by Mazur and Kraus [17]. These are not mast cells or cells from the PEComa group, which is confirmed by negative staining results for HMB 45 and CD117. However, they were CD68 positive. According to Resta et al. [15], no positive reaction to CD68 and MAC 387 was ever reported in the examined lipoleiomyomas. We carefully analyzed the available literature and we did not find similar descriptions.
The presence of mastocytes was reported in almost all analyzed specimens in the whole visual field. They were localized not only in the tumor pulp, but also around lipocytes. Their role in carcinogenesis is controversial [5, 21, 22, 23, 24].
According to Özdemir et al. [23], it should be asked whether they are an active player in the process of adipocytic “neometaplasia” of uterine fibroids, or they are just a harmless “passerby”. In our current material they appeared most frequently in lipoleiomyoma with the most advanced metaplasia in the 61-year-old woman (case 17).
Kinins are released as a result of changes in the environment of the main tumor, such as hyperemia, exudate and other regressive changes, and mastocytes play a significant role in these processes [24]. It is possible that interactions between these components could have stimulated the pluripotential cells already existing in mesenchymal stroma to create lipocytes. In the current research, the formation of such lipocytes was noted even in a strongly hyalinized stroma. However, this phenomenon was not statistically significant.
The hormonal state of a woman with adipocytic metaplasia in uterine fibroids requires further research. Sieinski et al. [9, 25] recorded a higher concentration of estrogen and progesterone receptors in uterine fibroids than in the surrounding muscle of the uterine corpus. Akpolat et al. [8] reported in a 54-year-old woman fewer estrogen receptors in lipoleiomyoma than in the surrounding muscle of the uterus. In our own material, we observed a positive correlation between the quantity of estrogen receptors in the primary tumor and the rate of development of adipocytic metaplasia. This correlation was additionally positively related with the individual age of operated patients. However, in comparison with the quantity of progesterone receptors, as well as in comparison with the degree of regressive changes in the main tumor, as mentioned above, such a correlation was not demonstrated. Similar research carried out on a larger scale and with more precise biochemical methods needs to be performed.

Conclusions

The rate of development of adipocytic metaplasia in uterine fibroids is positively correlated with the age of patients.
Increased activity of estrogen receptors in primary tumor supports rapid development of adipocytic metaplasia in the uterine fibroid.
Immunohistochemical examination results suggest that adipocytes in uterine fibroids come from multipotential stroma cells.
Type of regressive changes in uterine fibroids was not significantly correlated with the rate of adipocytic metaplasia in primary tumor.
We are greatly indebted to mgr Grażyna Erbel and mgr Teresa Szkudlarek for their excellent technical assistance.

The authors declare no conflict of interest.

References

1. Sośnik H, Sośnik K. Investigations on uterine fatty lesions. Pol J Pathol 2013; 64: 136-143.
2. El Amine El Hadj O, Bonraoui S, Be Fadhel CG, et al. Lipoma of the uterus: clinical and ethiopathological approach of 7 cases with immunohistohemical study of histogenesis. Tunis Med 2010; 88: 916-919.
3. Honoré LH. Uterine fibrolipoleiomyoma; Report of a case with discussion of histogenesis. Am J Obstet Gynecol 1978; 15: 635-636.
4. Aung T, Goto M, Nomoto M, et al. Uterine lipoleiomyoma: a histopathological review of 17 cases. Pathol Int 2004; 54: 751-758.
5. Özdemir O. Mast cells and the tumor-associated neoangiogenesis. Med Sci Monit 2006; 12: LE 9-11.
6. Takeuchi K, Kitazawa S, Tsujino T, et al. Uterine bizarre epithelioid lipoleiomyoma with a myxoid stroma. Eur J Gynecol Oncol 2006; 27: 273-274.
7. Akbulut M, Soysal ME, Duzcan SE. Giant lipoleiomyoma of the uterine corpus. Arch Gynecol Obstet 2008; 278: 291-293.
8. Akpolat I, Sertcelik A, Cӧmert S, et al. ERRP-29 and ER staining in uterine lipoma and lipoleiomyoma. Acta Oncol 1996; 35: 108.
9. Sieiński W. Lipomatous neometaplasia of the uterus. Report of 11 cases with discussion of histogenesis and pathogenesis. Int J Gynecol Pathol 1989; 8: 357-363.
10. Brandfass RT, Everts-Suarez EA. Lipomatous tumors of the uterus. A review of the world’s literature with report of a case of true lipoma. Am J Obstet Gynec 1955; 70: 359-367.
11. Meinhof U, Bersch W. Beitrag zur Kenntnis der fettgewebshaltigen Uterustumoren. Zentralbl Allg Pathol 1975; 119: 369-363.
12. Salm R. The histogenesis of uterine lipomas. Histogenese der Uteruslipome. Beitr Pathol 1973; 149: 284-292.
13. Kaiserling E. Immunohistochemical identification of lymph vessels with D2-40 in diagnostic pathology. Pathologe 2004; 25: 362-374.
14. Kahn HJ, Bailey D, Marks A. Monoclonal antibody D2-40, a new marker of lymphatic endothelium, reacts with Kaposi’s sarcoma and a subset of angiosarcomas. Mod Pathol 2002; 15: 434-440.
15. Resta L, Maiorano E, Piscitelli D, Botticella MA. Lipomatous tumor of the uterus. Clinico-pathological features of 10 cases with immunocytochemical study of histogenesis. Path Res Pract 1994; 190: 378-383.
16. Scully R. Smooth-muscle differentiation in genital tract disorders (Editorial). Arch Pathol Lab Med 1981; 105: 505-507.
17. Mazur MT, Kraus FT. Histogenesis of morphologic variations in tumor of the uterine wall. Am J Surg Pathol 1980; 4: 59-74.
18. Havel G, Wedell B, Dahlenfors R, Mark J. Cytogenetic relationship between uterine lipoleiomyomas and typical leiomyomas. Virch Arch B 1989; 57: 77-79.
19. Hu J, Surti U, Tobon H. Cytogenetic analysis of a uterine lipoleiomyoma. Cancer Genet Cytogenet 1992; 62: 200-202.
20. Morelli L, Pusiol T, Parolari AM, Piscioli I. Plexiform lipoleiomyoma of the uterus: first case report. Arch Gynecol Obstet 2006; 274: 117-118.
21. Goksu Erol AY, Tokyol C, Özdemir O, et al. The role of mast cells and angiogenesis in benign and malignant neoplasms of the uterus. Pathol Res Pract 2011; 207: 618-622.
22. Dyduch G, Okoń K, Pescarini E. Mast cells in melanocytic skin lesions. An immunohistochemical and quantitative study. Pol J Pathol 2011; 62: 139-144.
23. Özdemir O, Dağoğlu R, Goksu Erol A. Antitumor actions of human mast cells. Pol J Pathol 2012; 63: 292.
24. Dyduch G, Kaczmarczyk K, Okoń K. Mast cells and cancer: enemies or allies? Pol J Pathol 2012; 63: 1-7.
25. Sieinski W, Chrapusta S, Konopka B, et al. Struktura morfologiczna mięśniaków macicy a zawartość receptorów estrogenów i progestagenów. Nowotwory 1988; 38: 29-34.

Address for correspondence

Prof. Michał Jeleń
Department of Pathomorphology
and Oncological Cytology
Borowska 213, 51-213 Wrocław
tel. +48 71 734 39 60
e-mail: michal.jelen@umed.wroc.pl