Cancer-testis antigens and the MAGE-A family are sometimes described as expressed exclusively in cancer (Sang et al. 2011) . However, proteins expressed in cancer can have functions in normal cells during development or at times of rapid cell growth or differentiation. The full developmental spectrum of MAGE-A11 expression in humans remains to be examined rigorously. The MAGE-A11 gene promoter contains a CpG island subject to extensive DNA methylation that limits expression in normal tissues. However, there are examples of MAGE-A11 expression in normal tissues of the human reproductive tract.

MAGE-A11 is expressed in normal human endometrium during the menstrual cycle. MAGE-A11 mRNA increases 20-50-fold in mid-secretory endometrial biopsies obtained from normal cycling women (Bai et al. 2008) . MAGE-A11 immunostaining increases in mid-secretory endometrial epithelial cells. Highest levels of MAGE-A11 mRNA and protein are in the mid-secretory window of implantation. MAGE-A11 expression in human endometrium increases after the surge in luteinizing hormone (LH) that follows ovulation. Up-regulation of MAGE-A11 mRNA by cyclic AMP in endometrial and prostate cancer cell lines (Bai et al. 2008; Karpf et al. 2009) suggests that the LH-induced increase in cyclic AMP increases MAGE-A11 expression in the mid-secretory endometrium and may be required to establish receptivity for embryo implantation in humans. The hormone-dependent cycling of human endometrium does not occur in less evolved mammals. The hormone-regulated expression of MAGE-A11 suggests that MAGE-A11 evolved in primates for successful reproduction.

Progesterone receptor (NR3C3)-A (PR-A) and PR-B are isoforms transcribed from the same gene using different promoters (Richer et al. 2002) . PR-B is identical to PR-A except for 164 extra NH₂-terminus amino acids. PR-A is the predominant isoform in mouse uterus, whereas the more transcriptionally active PR-B is predominant in human endometrium (Igarashi et al. 2005) . MAGE-A11 increases the transcriptional activity of PR-B in an isotype-specific manner by interacting with a unique PR-B NH₂-terminal ¹¹⁰LLXXVLXXLL¹¹⁹ motif (Fig. 2) (Su et al. 2012) . The transcriptional enhancing effects of MAGE-A11 for PR-B were not seen with PR-A or the PR-A/B heterodimer. MAGE-A11 interaction with human PR-B is regulated by phosphorylation and ubiquitination, functions synergistically with p300, and amplifies the coregulator effects of p160 nuclear receptor coactivators. Interactions between MAGE-A11 and PR-B may explain the greater transcriptional activity of PR-B relative to PR-A. A ligand-dependent N/C interaction reported for PR-B (Tetel et al. 1999) may be modulated by MAGE-A11.

The coregulator function of MAGE-A11 therefore includes human AR and human PR-B. MAGE-A11 may function primarily as a PR-B coregulator in human endometrium, although effects on AR in uterus have not been excluded. The hormone-regulated cyclic function of human endometrium and evolution of MAGE-A11 is consistent with its function as an amplifying signaling molecule for steroid hormone receptors.

Cancer cells interfere with the retinoblastoma (Rb) family of proteins to achieve uncontrolled growth. The Rb family consists of the Rb tumor suppressor, p107 and p130. The major function of the Rb-related proteins is to regulate cell growth by inhibiting the activity of E2F transcription factors that drive the cell cycle. Inactivation of Rb family members in cancer occurs through several mechanisms. Rb gene mutations are common in late stage cancer and block the ability of Rb to inhibit E2F transcription factors resulting in cell cycle progression (Chellappan et al. 1991) . DNA tumor viruses, such as adenovirus early protein E1A, inactivate Rb family members to release transcriptionally active E2Fs and promote transition through the G1/S phase of the cell cycle (Fattaey et al. 1993) . MAGE-A11 interacts selectively with p107 of the Rb family, stabilizes p107 by inhibition of ubiquitination, and increases the transcriptional activity of E2F1 (Su et al. 2013) . MAGE-A11 interacts with hypophosphorylated E2F1 and does not interact with hyperphosphorylated E2F1, suggesting it is directly involved in E2F1 activation.

The ability of MAGE-A11 to increase E2F1 transcriptional activity resembles the activity of adenovirus early protein E1A, a DNA tumor virus protein that promotes cell transformation by disrupting Rb-related protein interactions with E2F transcription factors (Liu and Marmorstein 2007) . The specificity of MAGE-A11 interaction with p107, and its ability to increase E2F transcription factor activity, suggest that MAGE-A11 promotes cell cycle progression independent of its function as a steroid hormone receptor coregulator. MAGE-A11 may also provide an important link between steroid receptor transcriptional activity and cell cycle regulation.

Castration-recurrent prostate cancer growth during androgen deprivation therapy is a major clinical challenge. AR continues to drive the growth of castration-recurrent prostate cancer based on the expression of androgen-regulated genes, and inhibition of prostate cancer cell growth by silencing AR (Gregory et al. 1998; Zegarra-Moro et al. 2002; Ponguta et al. 2008) . AR coregulators such as p160 nuclear receptor coactivators are often overexpressed in castration-recurrent prostate cancer (Gregory et al. 2001a). However, it remains controversial whether AR splice variants exist and are at sufficiently high levels to impact castration-recurrent prostate cancer (Dehm et al. 2008; Sun et al. 2010) . AR splice variants may be largely a phenomenon of prostate cancer cell lines that are subject to extensive gene rearrangements. The level of reported AR variant mRNA in tumor tissue is very low compared to full-length AR (Watson et al. 2010) and could represent incompletely processed RNAs detected using the sensitive technique of quantitative RT-PCR. AR is also highly susceptible to partial proteolysis during isolation, which generates forms that migrate as reported AR variants (Wilson and French 1979; Gregory et al. 2001b).

Continued AR signaling in castration-recurrent prostate cancer could represent to some extent ligand-independent AR activation. However, early studies raised the possibility of intratumoral DHT production from adrenal androgen precursors (Geller et al. 1978) . Improved techniques of radioimmunoassay and mass spectrometry demonstrated that testosterone and DHT are at sufficient concentrations to activate AR in castration-recurrent prostate cancer tissue in patients undergoing androgen deprivation therapy (Mohler et al. 2004; Titus et al. 2005; Mostaghel et al. 2007; Montgomery et al. 2008; Locke et al. 2008) . Adrenal androgen precursors of DHT synthesis circulate in the blood of patients undergoing androgen deprivation therapy by medical castration. Adrenal androgen supplementation increased prostate DHT in rodent models (Labrie et al. 1988) . Normal prostate and prostate cancer cells have enzymes that convert adrenal androgens to DHT (Mohler et al. 2011a, 2011b). Continued reliance on ligand-activated AR in recurrent prostate cancer growth suggests that adrenal androgens are important precursors for intratumoral DHT synthesis when circulating testosterone and DHT are low in patients undergoing androgen deprivation therapy. It has also been suggested that cholesterol may be a source of intratumoral androgen synthesis in castration-recurrent prostate cancer (Leon et al. 2010; Twiddy et al. 2011; Mostaghel et al. 2012) .

MAGE-A11 levels increase in prostate cancer during progression to castration-recurrent growth (Karpf et al. 2009) . Immunohistochemistry shows an increase in MAGE-A11 protein in the CWR22 human prostate cancer xenograft after castration of tumor-bearing mice and in clinical specimens of castration-recurrent prostate cancer. The increase in MAGE-A11 protein correlates with a 20–100 fold increase in MAGE-A11 mRNA in the CWR22 xenograft with time after castration. MAGE-A11 mRNA can increase up to 1000 fold in clinical specimens of castration-recurrent prostate cancer. An increase in MAGE-A11 mRNA was seen in ~ 30 % of castration-recurrent prostate cancer specimens analyzed, suggesting its overexpression represents one mechanism that contributes to castration-recurrent growth. MAGE-A11 is also overexpressed in epithelial ovarian cancers (James et al. 2013) .

The major cause of the increase in MAGE-A11 in prostate and ovarian cancer is hypomethylation of a CpG island at the transcription start site of the X-linked MAGE-A11 gene promoter (Karpf et al. 2009; James et al. 2013). Progressive hypomethylation of the promoter occurs with time after castration of mice with the CWR22 human prostate cancer xenograft. In a limited number of castration-recurrent prostate cancers analyzed, there was an inverse relationship between AR and MAGE-A11 mRNA levels. One castration-recurrent prostate cancer sample with undetectable AR mRNA determined using quantitative RT-PCR had 1000-fold higher levels of MAGE-A11 mRNA. This suggests a compensatory relationship between AR and MAGE-A11 contributes to prostate cancer growth.

MAGE-A11 mRNA also increases in response to cyclic AMP in human prostate cancer and endometrial cell lines (Bai et al. 2008; Karpf et al. 2009) . Cyclic AMP up-regulation of MAGE-A11 is another mechanism that increases MAGE-A11 in castration-recurrent prostate cancer because cyclic AMP levels increase in prostate cancer cells during androgen deprivation (Burchardt et al. 1999) . The increase in AR transcriptional activity in response to cyclic AMP (Merkle and Hoffmann 2011) may reflect an increase in MAGE-A11. MAGE-A11 also interacts with and stabilizes the ligand-free AR, which may account in some cases for the increase in AR in castration-recurrent prostate cancer. Higher levels of MAGE-A11 enhance androgen-stimulated AR transcriptional activity through interactions with p300 and p160 coactivators, and drive prostate cancer cell growth through its ability to sequester p107, increase E2F transcriptional activity that drives the cell cycle.

MAGE-A11 is a multifunctional protein expressed at low levels in normal human cells and at higher levels in cancer. Extensive DNA methylation of the promoter that limits MAGE-A11 expression in normal cells is reversed in cancer cells by DNA hypomethylation. The ability of MAGE-A11 to increase the transcriptional activity of AR and PR-B by interacting with unique NH₂-terminal motifs also depends on its interaction with p300 and p160 nuclear receptor coactivators (Fig. 3). MAGE-A11 interacts with p107, a member of the Rb family that regulates cell growth by sequestering E2F transcription factors. MAGE-A11 stabilizes p107 in association with increased E2F1 transcriptional activity. Inhibition of prostate cancer cell growth by suppressing MAGE-A11 expression suggests a critical role in cell cycle regulation. The primate-specific expression of MAGE-A11 suggests it introduces a gain-in-function that enhances steroid hormone receptor transcriptional activity and provides a direct link to cell cycle progression. The multiple interacting partners suggest that MAGE-A11 functions in a transcriptional hub to increase steroid receptor activity and cell cycle progression in rapidly dividing normal cells and in the uncontrolled growth of cancer cells. Based on crystal structure of MAGE-G1 (Doyle et al. 2010) , MAGE-A11 may be a hub protein with ordered and disordered regions separated by flexible hinges similar to steroid receptors (Dunker et al. 2005) . (Fig. 3)