Auersperg, Nelly M.D., Ph.D.
Ovarian cancer remains the most lethal gynecologic malignancy in developed countries, with a dismal 5-yr survival rate of only about 40%. One of the main reasons for this poor prognosis is the lack of means of early detection which stems, among other factors, from our ignorance regarding the precise site of origin of these neoplasms. In addition, ovarian cancers are heterogeneous, with widely differing phenotypes, genomics, and clinical characteristics, which raises the question whether or not all or some of these have a common cellular origin. Among ovarian cancers, the high-grade serous ovarian carcinomas (HGSOCs) are the most common and most lethal of all ovarian malignancies. Current hypotheses for the origin of the HGSOCs propose that these cancers originate in the epithelium of the oviductal fimbriae (FE) with metastasis to the ovary, in the ovarian surface epithelium (OSE), or in both 1–3. To further investigate the origin of the HGSOCs, we defined the expression of 5 stem-cell markers that may play a role in the oncogenic potential of these 2 epithelia.
In recent years there has been intense interest in cancer stem cells (CSCs) which have been defined as small subpopulations of cells within tumors that have the capacity to self-renew but also to give rise to heterogenous cancer cell lineages that comprise the tumor 4. Being stationary or slow dividing, they tend to be resistant to therapy and are thought to be a major factor in the development of drug resistance. CSCs are commonly identified by their resembance to stem cells of their normal precursors, most commonly by surface markers 4. The purpose of the present study was to compare stem-cell characteristics between OSE and fimbrial epithelium by immunohistochemistry. Currently, there seems to be little or no information about the stem-cell profile of normal fimbrial epithelial cells. In contrast, Bowen et al. 5 reported the presence of the stem-cell markers secreted frizzled-resembling protein 1 (SFRP1), ALDH1A2, and LHX9 in normal OSE and concluded that OSE has the potential to serve as CSCs and to undergo differentiation along different lineages.
In the present study, we compared the expression by OSE, FE, and oviductal ampullae of the stem-cell markers NANOG, LHX9, SFRPI, ALDH1A1, and ALDH1A2. NANOG is a transcription factor belonging to the homeobox gene family and is essential for maintaining pleuripotency and suppressing differentiation in embryonic stem cells 6. It also functions in normal and neoplastic adult cells, and its subcellular location and functions in adult cells are cell-type specific 7. LHX9 is a member of the Lim-homeodomain transcription factors, homeobox genes that have been implicated in maintaining the capacity for self-renewal in embryogenesis and in several adult stem-cell lineages 8,9, and one of these, LHX9 is strongly expressed in OSE 6. SFRP1 is a Wnt inhibitor which is secreted into extracellular space where it competes for Wnt with cell-surface Frizzled receptors. Wnt signaling is a highly context-dependent signaling pathway and functions in stem-cell maintenance, proliferation, differentiation, and neoplastic progression, depending on the tissues and conditions involved 10,11. Importantly, Wnt is implicated in promoting ovarian carcinogenesis 12,13. ALDH1A1 and ALDH1A2 (aldehyde dehydrogenases 1A1 and 1A2) belong to a family of 17 ALDH1 isoforms, intracellular enzymes that convert retinol to retinoic acid and oxidize aldehydes as a detoxifying function. In embryonic development and in adults, retinoic acid is an important inducer and regulator of differentiation. High ALDH1A1 expression has been associated with poor prognosis and development of drug resistance in a variety of neoplasms, including ovarian cancers, and has been detected in putative ovarian CSCs 14,15. In contrast, although ALDH1A2 in expressed by OSE 5, its mode of action of is similar to ALDH1A1, and it plays an important role in embryonic development, there is no evidence that it has a negative effect on outcomes in cancer.
Together, the 5 stem-cell components used in the present study vary widely in their roles and modes of action. The aim of the study was to determine whether they are expressed in adult OSE and/or FE, which would suggest that they may be the source of ovarian CSCs. In the course of the study, we unexpectedly discovered that stem-cell markers are preferentially located in the distal FE. This is, to our knowledge, the first explanation for the observation that Serous tubal intraepithelial carcinomas (STICs), the putative precursors of HGSOCs, are preferentially located at these sites 16.
MATERIALS AND METHODS
Twenty-one normal ovaries and 21 normal salpingo-oophorectomy specimens from reproductive age women were formalin-fixed and paraffin embedded, and 5-µm sections of these specimens were stained immunohistochemically for the following stem-cell markers: NANOG, SFRP1, LHX9, ALDH1A1, and ALDH1A2. Sources and dilutions used of antibodies to these markers are listed in Table 1.
Antigen retrieval was carried out in a steamer using 10 mM sodium citrate buffer for 20 min at 95°C. Endogenous peroxidase was blocked with 3% hydrogen peroxide in methanol for 15 min and unspecific protein binding was blocked with protein block for 30 min. The sections were incubated with the following primary antibodies or the corresponding negative isotype controls at the following dilutions: SFRP1 (1:400), LHX9 (1:200), NANOG (1:50), ALDH1A1 (1:650), and ALDH1A2 (1:400). The polymer detection method was used for SFRP1, LHX9, and NANOG and the avidin-biotin peroxidase complex method for ALDH1A1 and ALDH1A2. The sections were developed with 3,3′-diaminobenzidine and counterstained with Mayer hematoxylin. Goat and rabbit immunoglobulin G were used as negative controls, and human testis was the positive control.
The OSE and ovarian epithelial inclusion cysts were analyzed for the approximate percentage of cells expressing the stem-cell markers and for variations in cell morphology (flat, cuboidal, and columnar). OSE is loosely attached to the ovary and is therefore rarely preserved intact, as parts of it are detached during handling before fixation. In our series, epithelium was found to be retained on the ovarian surfaces over a range of 0% to 90%, with an average of about 30%, and most values ranging from 10% to 60% (Table 2). It was necessary to correct for this variation when reporting on the average proportion of stained cells in a group of ovaries, by weighing staining data by the amount of OSE that they were based on (see Statistical analysis section).
To analyze the oviductal FE, each oviductal slide was screened in its entirety; the approximate proportion of FE staining for each stem-cell marker was estimated and the predominant subcellular location of the staining was determined. During this process it was noted that distal parts of the FE seemed to stain preferentially. To obtain semiquantitative information about this potentially important observation, the following procedure was carried out for each of the 5 markers, on 300 FE per marker: Sections through FE were selected randomly from slides derived from at least 5 oviducts. To be included in this study each fimbria had to be continous with its underlying stroma, had to be covered by intact epithelium throughout, and had to be taller than wide by at least a factor of 2, although the majority were considerably taller.
The project was approved by the Ethics Committee of the University of British Columbia.
To correct for the variation in amounts of OSE present per ovary when reporting on the average proportion of stained cells in a group of ovaries, staining data were weighed by the amount of OSE that they were based on. This was carried out by using the following formula, where % ep indicates the % of OSE retained by an ovary, and % +mk signifies the proportion of that OSE staining for a stem-cell marker. The subscripts identify individual ovaries:
The statistical significance of the difference between groups in Table 5 was determined using the Student t test.
In the ovaries, all 5 stem-cell markers were detected in the OSE, although there were differences in the proportion of cells staining, the location (ovarian surface or cysts), and cell type (flat, cuboidal, or columnar) (Fig. 1, Tables 2 and 3). On the ovarian surface, columnar cells were rare. Among flat and cuboidal cells, NANOG, LHX9, and SFRP1 were observed in ≥75% cells; ALDH1A2 was essentially ubiquitous, whereas only a low proportion was positive for ALDH1A1. NANOG and LHX9 were localized predominantly in the nuclei, whereas the location of SFRP1, ALDH1A1, and ALDH1A2 was cytoplasmic.
The number of epithelial inclusion cysts examined for each marker ranged from 35 to 52, some of which were likely serial sections from the same cysts (Table 3). The cysts were characterized by an increase in columnar cells, which lined them either uniformly or in combination with cuboidal cells. There was a striking increase in ALDH1A1 but a reduction of SFRP1 expression in both flat/cuboidal and in columnar cells in cysts compared with surface OSE.
As shown in Table 4, all 5 markers were detected in the oviductal FE, with variations in the proportion of FE stained. In the majority of oviducts, only ≤25% FE expressed NANOG and SFRP1, whereas in most cases the majority of FE stained positively for ALDH1A1 and ALDH1A2. Subcellular staining varied among the markers (Fig. 2). Cytoplasmic staining for NANOG was low in the secretory cells and somewhat more intense in ciliated cells. Sporadic nuclei were intensely stained. Unexpectedly, these intensely NANOG-positive nuclei belonged predominantly to ciliated cells, which are supposed to be postmitotic 17. In contrast, most nuclei of secretory cells appeared to contain little or no NANOG.
ALDH1A1 and ALDH1A2 showed similar intracellular distributions: nuclei were negative, and the cytoplasmic staining was limited to secretory cells, whereas ciliated cells tended to be negative. In both secretory and ciliated cells, SFRP1 was localized mainly in the apical cytoplasm, and LHX9 in the nuclei. Importantly, the distribution of the stem cells within the fimbrial epithelium was not random. Analysis of 300 FE per marker (see Materials and methods section) showed that FE could be divided into 2 groups: 1 those that were uniformly marker positive or negative throughout the epithelium; 2 FE where marker distribution was uneven, with a gradient between proximal and distal parts of the FE. The results were striking: there was a highly significant (P≤0.00001) trend for most marker-positive cells to be located in the distal parts of the FE. This was true for all 5 markers (Table 5, Figs. 3–5).
Sections through oviductal ampullae were found in several of the fimbrial slides. Among these, 9 sections that were stained for NANOG and 7 sections stained for SFRP1 were all negative. In contrast, 5 of 9 LHX9 samples contained positive cells, and most cells were positive for ALDH1A1 and ALDH1A2 in the majority of samples. In ampullae that were otherwise negative, rare positive cells were localized at their periphery, suggesting a possible location for stem cells (Fig. 6).
This study demonstrates that both OSE and FE express several stem-cell markers, including the pleuripotency marker NANOG. This observation indicates that OSE and FE are both pleuripotent, with the capacity to give rise to CSCs, which are the source of different ovarian cancer subtypes. The data confirm and expand evidence that OSE is a multipotential epithelium with stem-cell characteristics 3,5,18,19. They show that OSE that lines cortical inclusion cysts exhibits changes in its stem-cell profile from surface OSE which may be indicative of premalignant progression. The results show further that FE, despite its high degree of differentiation, also expresses all 5 stem-cell markers that are present in OSE. Importantly, the markers occur preferentially in the distal parts of FE, which are also the sites of STICs, the putative precursors of HGSOCs 16. This discovery is a plausible explanation for the as yet unexplained predilection of STICs to occur at these sites.
Among the stem-cell markers that characterize OSE and FE, one of the most significant is NANOG, an important factor that maintains pleuripotency in embryonic development 6 and likely also in CSCs 20. NANOG also controls gene expression in normal adult tissues and can be found in either the nucleus or cytoplasm in a stage-dependent and cell type-dependent manner 7. An unexpected observation in the present study was the subcellular location of NANOG in the FE. NANOG was present in low amounts in both the cytoplasm and nuclei of most cells, but intensely stained nuclei were observed almost exclusively in ciliated cells (Fig. 2). This was unexpected, as the presence of NANOG is generally associated with pleuripotentiality and limited differentiation; yet, there is considerable evidence that in fimbrial epithelium the interspersed secretory cells are the source of proliferation, whereas ciliated cells are terminally differentiated and postmitotic 17. Furthermore, the nuclear staining was so intense that it seems unlikely to only reflect the distribution of a normally functioning transcription factor. Although this observation requires further study, it is tempting to speculate that in analogy with p53, which is normally undetectable by immunohistochemistry but is detectable when it is overexpressed in response to inactivation, NANOG might be similarly inactivated in the nonreplicative ciliated cells. Like NANOG, the transcription factor LHX9 was expressed in the nuclei of OSE and FE. LHX9 is a member of the Lim-homeodomain transcription factors that specify regional and cellular identities in early embryogenesis 21, and also have the capacity to maintain the stem-cell and/or progenitor status in the adult 9. All markers except for ALDH1A1 were widespread in OSE on the ovarian surface. In contrast, in OSE-lined epithelial inclusion cysts, SFRP1 was rare, whereas ALDH1A1 was detected in over 80% of cysts. These changes may indicate acquisition of properties related to neoplastic transformation: SFRP1, a frizzled-related protein, inhibits Wnt signaling by competing for Wnt in extracellular spaces with the cell-surface receptor Frizzled. The reduced expression of SFRP1 therefore permits increased Wnt activity in OSE-lined inclusion cysts. Importantly, aberrant activation of Wnt has been associated with the initiation and progression of malignant tumors, including ovarian cancers 13,22. The other striking change in cysts versus surface OSE was the increased presence of ALDH1A1 in the cysts. ALDH1A1 is a member of 17 ALDH1 isoforms which oxidize aldehydes and thereby play a detoxifying role, and also convert retinol to retinoic acid, an important mediator of differentiation pathways in the embryo, and of differentiation, apoptosis, and growth regulation in the adult. Among the ALDH1 isoforms, ALDH1A1 seems to be the only member that has been associated with ovarian cancer progression 15. High ALDH1A1 content has been shown to be a possible cause of chemoresistance in recurrent ovarian carcinoma, perhaps through it detoxifying function 15, but it is also associated with increased tumorigenicity 14. Thus, the increase in ALDH1A1 expression in inclusion cysts, compared with the ovarian surface, may contribute to the malignant potential of OSE cells. In contrast to ALDH1A1, ALDH1A2, another ALDH1 isozyme and stem-cell component 5 had a very different immunohistologic expression: it was the most widespread of all the markers tested and it, with ALDH1A1 were the only markers expressed throughout the oviductal ampullae. These results suggest that ALDH1A2 has different functions from ALDH1A1, which might include regulation by retinoic acid of the secretory activities of the fimbriae and ampulla independently of its stem-cell characteristics.
Importantly, the stem-cell markers of FE were found predominantly in the distal parts of the FE (Figs. 3–5, Table 5). It is well known that STICs, the putative fimbrial precursors of HGSOCs 16, are characteristically located distally on FE, but the reason for this location has been unknown. The present study provides a reasonable explanation for this location—the presence of stem cells that are considered likely precursors of cancers stem cells and their derivatives, the tumor cells 23. An interesting question that arises is why fimbrial stem cells would be at this location. Stem cells, in general, must be located in a niche that provides the necessary environment which will allow them to maintain their stem-cell status. This status includes a state of proliferative quiescence and incomplete differentiation 24. The location of OSE on the ovarian surface is a perfect example of a stem-cell niche. OSE is isolated basally from stromal and circulatory growth-promoting and differentiation-promoting factors and metabolites by the collagenous tunica albuginea 18. It thus has little access to agents that would foster growth and/or differentiation. In analogy, cells at the tip of FE are the ones that are furthest removed from sources of nutrients, hormones, and other factors that are required for growth and differentiation that would interfere with the maintenance of the stem-cell status.
The location of stem cells in the distal parts of the FE also reflects the close developmental relationship between the oviduct and OSE. Throughout differentiation of the mullerian ducts, the distal FE remain their most lateral components, closest to their tissue of origin which is the highly pleuripotential coelomic epithelium, and to the developing OSE. The discovery of stem cells in both of these locations supports the concept that OSE and the distal FE remain incompletely differentiated in the adult, which may contribute to their propensity to undergo neoplastic transformation 25.
The author thanks Dr Blake C. Gilks for his invaluable advice and discussions, and for providing the tissue specimens.
1. Pothuri B, Leietao MM, Levine DA, et al .Genetic analysis of the early natural history of epithelial ovarian carcinoma.PLoS One. 2010; 5:e10358
2. Kurman RJ, Shih IeM .Molecular pathogenesis and extraovarian origin of epithelial ovarian cancer—shifting the paradigm.Hum Pathol. 2011; 42:918–31
3. Auersperg N .The origin of ovarian carcinomas: a unifying hypothesis.Int J Gynecol Pathol. 2011; 30:12–21.
4. Conic I, Dimov I, Tasic-Dimov D, et al .Ovarian epithelial cancer stem cells.ScientificWorldJournal. 2011; 11:1243–69.
5. Bowen NJ, Walker LD, Matyunina LV, et al .Gene expression profiling supports the hypothesis that human ovarian surface epithelia are multipotent and capable of serving as ovarian cancer initiating cells.BMC Med Genomics. 2009; 2:71
6. Mullin NP, Yates A, Rowe AJ, et al .The pluripotency rheostat Nanog functions as a dimer.Biochem J. 2008; 411:227–31.
7. Ambady S, Malcuit C, Kashpur O, et al .Expression of NANOG and NANOGP8 in a variety of undifferentiated and differentiated human cells.Int J Dev Biol. 2010; 54:1743–54.
8. Dahl L, Richter K, Hagglund AC, et al .Lhx2 expression promotes self-renewal of a distinct multipotential hematopoietic progenitor cell in embryonic stem cell-derived embryoid bodies.PLoS One. 2008; 3:e2025
9. Rhee H, Polak L, Fuchs E .Lhx2 maintains stem cell character in hair follicles.Science. 2006; 312:1946–9.
10. De Sousa E, Melo F, Medema JP .Axing Wnt signals.Cell Res. 2012; 22:9–11.
11. Hanahan D, Weinberg RA .The hallmarks of cancer.Cell. 2000; 100:57–70.
12. Gatcliffe TA, Monk BJ, Planutis K, et al .Wnt signaling in ovarian tumorigenesis.Int J Gynecol Cancer. 2008; 18:954–62.
13. Barbolina MV, Burkhalter RJ, Stack MS .Diverse mechanisms for activation of Wnt signalling in the ovarian tumour microenvironment.Biochem J. 2011; 437:1–12
14. Landen CN Jr, Goodman B, Katre AA, et al .Targeting aldehyde dehydrogenase cancer stem cells in ovarian cancer.Mol Cancer Ther. 2010; 9:3186–99.
15. Steg AD, Bevis KS, Katre AA, et al .Stem cell pathways contribute to clinical chemoresistance in ovarian cancer.Clin Cancer Res. 2012; 18:869–81.
16. Crum CP, McKeon FD, Xian W .The oviduct and ovarian cancer: causality, clinical implications, and targeted prevention.Clin Obstet Gynecol. 2012; 55:24–35
17. Kuhn E, Kurman RJ, Sehdev AS, et al .Ki-67 labeling index as an adjunct in the diagnosis of serous tubal intraepithelial carcinoma.Int J Gynecol Pathol. 2012; 31:416–22.
18. Auersperg N, Wong AS, Choi KC, et al .Ovarian surface epithelium: biology, endocrinology, and pathology.Endocr Rev. 2001; 22:255–88
19. Strauss R, Li ZY, Liu Y, et al .Analysis of epithelial and mesenchymal markers in ovarian cancer reveals phenotypic heterogeneity and plasticity.PLoS One. 2011; 6:e16186
Erratum in: PLoS One. 2011;6(2)
20. Hu L, McArthur C, Jaffe RB .Ovarian cancer stem-like side-population cells are tumourigenic and chemoresistant.Br J Cancer. 2010; 102:1276–83.
21. Abellan A, Menuet A, Dehay C, et al .Differential expression of LIM-homeodomain factors in Cajal-Retzius cells of primates, rodents, and birds.Cereb Cortex. 2010; 20:1788–98.
22. Kinzler KW, Vogelstein B .Lessons from hereditary colorectal cancer.Cell. 1996; 87:159–70
23. Li Y, Laterra J .Cancer stem cells: distinct entities or dynamically regulated phenotypes? Cancer Res. 2012; 72:576–80
24. Li l, Neaves WB .Normal stem cells and cancer stem cells: the niche matters.Cancer Res. 2006; 66:4553–7.
25. Auersperg N .The origin of ovarian cancers—hypotheses and controversies.Front Biosci. 2013; S5:709–19.