Cancer/testis (CT) antigens are named after their expression pattern as they are typically present in various types of tumors and in the germ cells of normal adult testis. Adult ovarian tissue is usually reported to be CT antigen negative. Based on the differences in female versus male gonadal development, the ovarian counterpart of the most predominant CT antigen positive testicular germ cells are not prevalent in the adult ovary. Hence, we analyzed the protein expression of several CT antigens in fetal ovary by immunohistochemistry with various monoclonal antibodies (mAbs) previously generated by our group. The mAbs used were: MA454 (MAGE-A1), M3H67 (MAGE-A3), 57B (MAGE-A4), CT7-33 (CT7/MAGE-C1), and ES121 (NY-ESO-1). All mAbs showed some immunopositivity in fetal ovarian germ cells. The most intense staining was seen with mAbs M3H67, 57B, and CT7-33 during weeks 16 – 23 of gestation. The most prevalent cells stained were oogonia, with only focal staining of oocytes of the primordial follicle. We conclude that CT antigens are regularly expressed in fetal ovarian germ cells and might play an important role in male and female germ cell biology.
This article was published in Cancer Immunity, a Cancer Research Institute journal that ceased publication in 2013 and is now provided online in association with Cancer Immunology Research.
Cancer/testis (CT) antigens are named after their typical pattern of expression since they are found in various histological types of tumors but are restricted in normal tissues to germ cells of the testis (1). Since the isolation of its prototype MAGE-1 (now referred to as MAGE-A1), more than 40 genes or gene families of CT antigens have been discovered (2, 3, 4). Based on their tumor-associated expression pattern, CT antigens are regarded as potential target antigens for the immunotherapy of cancer and have become a major focus for the development of vaccine-based clinical trials in recent years (5, 6, 7). Knowledge about the restriction of CT antigen expression in normal tissues, i.e. their presence in testicular germ cells, is based on various RT-PCR and immunohistochemical analyses. At the protein level, CT antigens are present in spermatogonia and, to a variable degree, in later stages of sperm cell maturation (8, 9, 10, 11, 12, 13). Human ovarian tissue has been occasionally analyzed in the past and mostly found to be negative for most CT antigens, although single antigens were shown to be expressed in the ovary at the mRNA level (14, 15, 16, 17). However, these RT-PCR analyses were explicitly or presumably performed on adult ovarian tissue. Given their expression pattern in testicular germ cells, we assumed that any potential CT antigen expression in ovary would comprise germ cells of a similar maturation stage as in testis (18). Based on the biology of testicular germ cells, the analogous female germ cell maturation steps with presumably the most prominent CT antigen expression take place in the developing ovary (9, 10, 11, 12, 13, 18, 19, 20, 21). Consequently, in the present study we analyzed fetal ovary at various stages of maturation for the presence of several CT antigens by immunohistochemistry and demonstrate that CT antigens are regularly expressed in fetal ovarian tissue.
Based on availability, our analysis was restricted to a small collection of specimens derived from autopsy cases. Consequently, some of the tissues showed considerable autolysis. Furthermore, the age distribution of the available tissues was not even. However, ovarian tissue from various gestational ages starting week 13 was available to gain an adequate picture of CT antigen expression at different stages of development. The result of our immunohistochemical analysis is listed in Table 1, while Figures 1 to 3illustrate fetal ovary morphology and the expression of CT antigens in three representative specimens at gestational ages of 13, 22, and 40 weeks.
Conventional hematoxylin-eosin (HE) staining revealed details of the fetal ovarian tissue at different gestational ages (Figures 1-3; panels A, G, and M). At week 13, illustrated in Figure 1, mostly undifferentiated ovarian tissue was visible. Most cells resembled germ cells with large nuclei that were surrounded by clear cytoplasm (Figure 1; panels A, G, and M). Under the surface epithelium, cells with small and chromatin-dense rounded nuclei were present. Interspersed were longitudinal cells with elongated chromatin-dense nuclei. Plenty of dividing cells could be seen. At week 22 (Figure 2; panels A, G, and M), the periphery of the ovary consisted of cords and clusters of oogonia while in the central areas differentiation of primordial follicles was apparent. This was evidenced by the presence of germ cells with large nuclei and broad eosinophilic cytoplasm surrounded by much smaller cells with round cytoplasm, resembling early follicle epithelial cells (Figure 2, panel M). At week 40 (Figure 3; panels A, G, and M), the ovarian tissue consisted mostly of primordial follicles consisting of large centrally located germ cells surrounded by a layer of granulosa cells. The follicles were separated by a wide, cellular stroma (Figure 3, panel M).
Immunohistochemistry revealed that all five anti-CT mAbs stained at least some of the ovarian samples, albeit to a variable degree. In all cases, staining was observed in germ cells, though these were difficult to delineate in ovaries at early maturation stages. All five mAbs showed some staining at week 13, and more intense staining in ovaries at later stages of maturation.
Monoclonal antibody MA454 (Figures 1-3; panels B, H, and N) to MAGE-A1 and mAb ES121 (Figures 1-3; panels F, L, and R) to NY-ESO-1 showed the least staining. Both mAbs showed a comparable immunoreactivity pattern, with focal staining only. The earliest immunopositive cells could be seen in the ovary at week 13 of gestation for both mAbs but there was also staining in later ovarian stages, until week 40 for MA454 and up to week 31 for mAb ES121. In all ES121 and MA454 positive cases, the staining comprised a few scattered germ cells and no homogenous staining was seen in any of the specimens.
The restricted staining observed with mAbs MA454 and ES121 contrasts with the immunoreactivity seen with the other three mAbs of our analysis. The most intense immunoreactivity in our analysis was seen with mAbs M3H67 and CT7-33, while mAb 57B stained to a lesser extent but showed a comparable staining pattern. Similar to the staining seen with mAbs MA454 and ES121, there was little staining of the ovary at week 13 of gestation with mAbs M3H67, CT7-33 and 57B. However, in the fetal ovary specimens between week 16 and 23, there was intense immunoreactivity observed with all three reagents. Though the staining was variable and not fully congruent, the most intense expression of protein was seen in this time period. This is best illustrated in Figure 2 with mAbs M3H67 (panels C, I, and O) and CT7-33 (panels E, K, and Q). Here cords of immunopositive oogonia reaching from the surface of the medulla were present, while the centrally located oocytes within the early primordial follicles were mostly immunonegative though several showed weak to moderate intense staining (Figure 2; panels I, K, O, and Q). After week 23 of gestation (Figure 3), the number and intensity of immunopositive germ cells decreased. In the ovaries of the last trimester, only occasional germ cells were positive with mAbs M3H67 (panels C, I, and O), 57B (panels D, J, and P), and CT7-33 (panels E, K, and Q). These were easily visible as single stained primary oocytes of the primordial follicle. Staining was cytoplasmic and/or nuclear. Occasionally, clusters of germ cells without surrounding follicle epithelia and resembling focal groups of remaining oogonia were immunopositive (Figure 3, panel Q).
Within a short time of the initial isolation of the CT prototype antigen MAGE-1 (now MAGE-A1), CT antigens have become the focus of research as vaccine targets for the immunotherapy of cancer. This is based on their ability to elicit T-cell and serological immune responses in cancer patients (1, 22). Consequently immunological methods such as T-cell epitope cloning and SEREX (serological analysis of recombinant expression libraries) have been used for their identification (23). The number of genes and gene families identified as CT antigens has steadily increased and now classical (mapping to chromosome X) and non-classical CT antigens are distinguished (24). The expression of classical CT antigens in normal adult tissues is restricted to testicular germ cells and their acronymic name partially reflects this fact.
Several studies showed that protein expression of CT antigens is most prevalent in the germ cells located at the periphery of the seminiferous tubule. These cells resemble cells in the early maturation stages in the adult testis, such as spermatogonia (8, 10, 12, 13). The immunopositivity of testicular germ cells decreases from the periphery to the center of the tubule, with variable reactivity of spermatocytes while spermatids are negative. In adult testis, germ cell maturation is a permanent process commencing around puberty and continuing until late age. In ovary, germ cell development takes a different path. As opposed to testis, ovarian germ cell maturation is not a continuous process in adult life, but rather a process starting in utero which is then arrested around birth in order to continue only after puberty (18, 20, 25). During embryonal life, ovaries are populated by immature germ cells, which in the course of fetal development pass through the early maturation stages until they form primary follicles and arrest in early meiosis (18, 20, 21). Based on these facts, the ovarian germ cell most closely corresponding to its testicular counterpart with highest expression of CT antigens must be sought in the fetal ovary. Our study addressed this issue by analyzing fetal ovary specimens for the presence of CT antigens. Not surprisingly, we found all CT antigens present in fetal ovarian germ cells, albeit to variable degrees. Most immunostaining was seen with the mAbs 57B, M3H67, and CT7-33 while comparably little staining was observed with mAbs E978 and MA454 to NY-ESO-1 and MAGE-A1, respectively.
The differential immunoreactivity of the present set of CT antigen markers is partly reflected in testis. As demonstrated previously, CT7-33, 57B, and M3H67 frequently reveal a strong immunoreactivity in testicular germ cells, comprising intensely stained spermatogonia and spermatocytes stained to variable degrees (9, 12). The testicular reactivity pattern for MA454 and ES121 is usually more restricted to spermatogonia, oftentimes expressed in a patchy pattern throughout the testis with some tubules showing less spermatogonial staining than others (10, 11). There is also usually less immunoreactivity observed in spermatocytes. At this point, no explanation for this difference in CT antigen expression in germ cells can be given.
A cautionary note as to the reactivity of anti-CT antibodies is warranted. Due to the high homology of several CT antigens, cross-reactivities of some reagents to CT antigens cannot be fully excluded. This is best illustrated by mAb 57B, which was generated to MAGE-A3 and subsequently considered to react with several MAGE-A antigens, but was eventually found to be predominantly reactive with MAGE-A4. However, no cross reactivity of these reagents has been found to any non-CT antigen (26, 27, 28). Nevertheless, the fine specificity of these reagents should be regarded with caution.
In our analysis, CT antigen expression appeared to be most intense in fetal ovarian tissue between weeks 16 to 23. This time period corresponds to the late proliferation and the entire maturation phase of germ cells (20). This is congruent with the presence of CT antigens in testis where they were found to be most prevalent in spermatogonia (9, 12, 13). Interestingly, immunostaining with all our reagents of fetal ovary specimens at week 13 was considerably less. This suggests that CT antigen expression in germ cells of earlier fetal stages, such as primordial germ cells, is less prevalent. The limited CT antigen expression in early stages of ovarian development is corroborated by an immunohistochemical analysis of NY-ESO-1 and MAGE-A4 in fetal testis (29, 30). NY-ESO-1 positive germ cells were only present in fetal testis older than week 16, and increased and peaked at week 40. Comparable data for MAGE-A4 have been reported (30). However, the overall number of cells was minute compared with the expression in adult testis and the corresponding developmental stages of the fetal ovary (29). Little knowledge exists as to the presence of CT antigens on a protein level in non-gonadal fetal tissue. We previously did an immunohistochemical analysis in a limited number of fetal specimens from week 8 to week 12 and did not find any staining in tissues outside the gonads (data not included).
Consequently, our study and the corresponding findings in testis suggest a role of CT antigens in germ cell development, mainly during the period around meiosis rather than in the early phase of primordial germ cells or later stages such as folliculogenesis. However, it is entirely possible that CT antigens can be up-regulated during folliculogenesis in adult life. Since CT antigen expression in the oocyte of the maturating follicle may be restricted to a short time period, detection of CT antigen expression in the adult ovary may be difficult. We found no immunoreactivity in adult ovarian tissue with any of the serological reagents used in this study (unpublished results). However, GAGE protein was recently found to be expressed in oocytes of the adult ovary (8). Since at present the function of CT antigens remains largely unknown, the interpretation of our findings in the biological context of germ cell maturation is difficult. Recent studies have suggested the involvement of CT antigens in gene expression by interacting with DNA binding proteins, affecting p53 via histone deacetylase 3 (HDAC3), as well as anti-apoptotic activity (4, 31, 32, 33). However, a specific role for their biological function in germ cells cannot be assigned as yet. Consequently it would be interesting to compare the presence of CT antigens in germ cells with other biological parameters, such as proliferation and apoptosis. However, in the present study we intended and succeeded to demonstrate that CT antigens are in fact regularly expressed in the female germ cells. Obviously future analyses of the characteristics of the CT-positive germ cell population in the fetal ovary are necessary.
Materials and methods
Ovarian specimens and antibodies
Standard formalin-fixed paraffin-embedded tissues were retrieved from the archives of the Department of Pathology of the University of Pennsylvania. Ovarian specimens from fifteen fetal autopsy cases, with a gestation age of week 13 to week 40, were available. The following monoclonal antibodies were used: mAb MA454 (10) to MAGE-A1, mAb M3H67 to MAGE-A3, mAb 57B to MAGE-A4, mAb CT7-33 to CT7/MAGE-C1, and mAb ES121 (11) to NY-ESO-1.
Immunohistochemistry was performed as previously described. Briefly, 5 µm paraffin sections were applied to slides (Menzel, Braunschweig, Germany) for immunohistochemistry (IHC) and heated in an oven at 60˚C overnight. Antigen retrieval (AGR) was performed by heating the slides for 30 minutes in a buffer solution. Citrate buffer (10 mM, pH 6.0) was used for mAb CT7-33, EDTA buffer (1 mM, pH 8.0) for mAbs MA454, M3H67, and 57B. For mAb ES121 DAKO hipH antigen retrieval solution (DAKO, Carpentaria, CA) was employed. Primary antibody incubation was done overnight at 4˚C. A biotinylated horse anti-mouse antibody (Vector Labs, Burlingame, CA) was used as a secondary Ab for mAbs MA454, CT7-33, M3H67, and 57B, followed by an avidin-biotin-complex detection system (ABC Elite, Vector). For the detection of mAb ES121, DAKO Envision plus (DAKO) was employed. Diaminobenzidine (DAB) was used as a chromogen.
Part of this study has been presented at the 2001 annual meeting of the United States and Canadian Academy of Pathology (USCAP) in Atlanta, GA.
- Received November 21, 2006.
- Accepted December 19, 2006.
- Copyright © 2007 by Achim A. Jungbluth