Article by Natalie Banet and Robert J. Kurman: Two types of ovarian cortical inclusion cysts: proposed origin and possible role in ovarian serous carcinogenesis; Int. J. Gynecol. Pathol. 2015;34:3-8.

In Reply:

The above publication is very interesting and deals with an important topic, that is, the source of high-grade serous ovarian carcinomas. The authors propose that all of these highly malignant tumors originate from oviductal fimbriae, or from ovarian cortical inclusion cysts (CICs) that are derived from fimbriae, rather than from CICs derived from ovarian surface epithelium (OSE). Several points in the publication require clarification before this conclusion can be accepted:

The authors used calretinin and PAX8 as the specific histochemical markers for OSE-derived and fimbriae-derived CICs, respectively. However, while calretinin is a consistently positive marker for mesothelial cells including OSE, and is absent in fimbriae, PAX8 is found in both cell types. In the ovary, PAX8 is mainly, although not completely absent in OSE on the ovarian surface, but it is expressed by most OSE-lined CICs 1–3. This shift in gene expression parallels the general tendency of OSE to acquire epithelial characteristics when translocated from the surface to the ovarian stroma 4. Panels J–O in figure 2, meant to illustrate the specificity of calretinin and PAX8, are of such poor quality as to be inconclusive.

The authors assume that all “mixed” CICs (those being lined partially by flat, OSE-like epithelium and partially by columnar, ciliated epithelium) are derived from fimbriae, and speculate that “expansion” (not further defined) causes flattening of some of the ciliated cells to resemble OSE. However, the flat component of mixed CICs does not only resemble OSE morphologically but also by differentiation: it expresses calretinin and lacks several epithelial markers (eg, EPCA, EMA, cilia) present in the ciliated cells within the same CICs 2,3. Therefore, mixed CICs do not result simply through distortion of cell shapes but, rather, through altered gene expression, that is, metaplasia from an OSE-like to a fimbria-like phenotype. The only alternative interpretation of the coexistence of OSE-like and fimbriae-like cells within the same CICs would be metaplasia from a fimbrial to a mesothelial phenotype, which is unlikely. Metaplasia of OSE to fimbriae-like epithelium is more likely to occur because (1) OSE cells are pleuripotential stem cells with the capacity to differentiate along more than one pathway 5–7 and (2) metaplasia tends to lead to developmentally related cell types, and OSE originates from the same embryonic field as the fimbrial epithelium 8,9. It is misleading that, in figure 1, the authors pooled the mixed CICs with ciliated CICs. As mixed CICs include OSE-derived CICs undergoing metaplasia, it is impossible to determine the true proportion of ciliated CICs from figure 1. In the Results section the authors report that 60% of CICs are ciliated, which is closer to results by others 4 and is significantly lower than suggested by figure 1.

The article is based on the assumption that fragments of fimbrial epithelium enter the ovarian stroma through ovulatory ruptures of the ovarian surface. This idea has been suggested in many publications but there seems to be no evidence supporting it. In contrast, there are indications contrary to this hypothesis: (1) Immediately after ovulation, the site of follicular rupture is filled with a mass of cells, coagulated fluids and blood, which must impede displacement of any fimbrial epithelial fragments. It would be expected, therefore, that such fragments would occasionally be found within or near freshly ovulated follicles and that ciliated CICs would be located predominantly in or near ovulated follicles or c. lutea. No such spatial relationships have been reported. (2) The authors conclude, based on table 1, that the number of CICs increases with age. Interestingly though, in this table, the highest number of total ciliated CICs occurs in the 71+ age group, that is, in women long past menopause and ovulations. (3) CICs are significantly more numerous in women with polycystic disease who ovulate rarely or not at all, and the number of CICs is proportional to parity, that is, it is inversely proportional to the number of ovulations 10,11. These reports support the hypothesis that CICs arise from invaginations of surface OSE 11.

In conclusion, the hypothesis that all high-grade serous ovarian carcinomas arise from fimbrial cells requires further studies.