Breast Cancer

The incidence rates for the three most common gynecological malignancies— breast cancer, endometrial adenocarcinomas, and ovarian adenocarcino-mas—increase sharply at menarche and decline abruptly at menopause (Pike et al. 2004). This was a clear indication that risk factors associated with the ovarian function are involved in the etiology of these tumors. Menopause is the most effective protective factor against each of these cancers (Pike et al. 2004). Early age at menarche, late age at menopause, and late age at first full-term pregnancy increase the risk for breast cancer, while removal of the ovaries at a younger age has a protective effect against breast cancer (Kelsey and Bernstein 1996). Ultimately, these risk factors modulate the status of the two ovarian hormones, estrogen and progesterone, that are known to control the normal development of the mammary gland and also induce breast tu-morigenesis. Estrogen is thought to serve either as a preinitiator or initiator of breast tumorigenesis, or as a growth promoter of existing breast malignancies (Hilakivi-Clarke 2000). The estrogen antagonist synthetic compounds such as tamoxifen, which block the action of estrogen, are very efficient in preventing the breast tumorigenesis and the recurrence of the disease (Gelber et al. 1996).

All sexual hormones exert their effects on their target tissues through steroid receptors. The mechanism by which sexual hormones can influence the pathogenesis of various hormonal-driven cancers may be related to the function of these receptors. The female sex steroid receptors, estrogen receptor a (ERa), estrogen receptor p (ERp), and progesterone receptor (PR), are DNA binding molecules that act as transcription factors. Following their activation by binding to their specific ligands, estrogen and progesterone steroid receptors recognize and bind specific hormone-responsive DNA elements situated in the promoter regions of the hormonally regulated genes. The steroid receptors are among the few receptors that interact directly with components of the transcriptional machinery and chromatin structure to regulate gene expression (Kinyamu and Archer 2004). The steroid receptors are expressed in a tissue- and cell-specific manner, and their expression can be affected by methylation during aging and tumorigenesis.

ESR1 and ESR2

Most of the biological effects of estrogen and its therapeutic synthetic antagonists are mediated via two distinct estrogen receptors called ERa and ERp, which are encoded by the ESR1 and ESR2 genes, respectively. Although they recognize the same estrogen-responsive elements, they have been shown to have opposing activities at activating protein-1 (AP1) sites (Paech et al. 1997) and to differ in the use of their ligand-independent activation function (AF1) domains for transactivation (Cowley and Parker 1999). In vitro studies have also shown that these two receptors have different responses to tamoxifen and other synthetic antagonists of estrogen (Nilsson et al. 2001). While estrogen-activated ERa stimulates cell proliferation (Nilsson et al. 2001), ERp has been shown to inhibit the proliferation and invasion of breast cancer cells (Lazen-nec et al. 2001). Consequently, the differential expression of these receptors in a tissue-specific manner may also explain some of the tissue-specific ef fects of estrogen. While both ERs are expressed in ductal and lobular breast epithelium (Flototto et al. 2001), ERp is more abundant than ERa in normal breast epithelium (Widschwendter and Jones 2002).

Both of these receptors have promoter-associated CpG islands that have the potential to become abnormally methylated in cancer. The subsequent loss of expression of these receptors can disrupt the normal estrogen-signaling pathway and result in inactivation of downstream targets of this pathway. Multiple promoters have been described for both ESRs corresponding to various isoforms of these receptors. However, most of the methylation analyses have been performed on the A promoter for the ESR1 and on the promoter associated with exon 1 for the ESR2.

Aberrant methylation of the ESR1 gene promoter A has been documented in various normal epithelial tissues as an age-dependent modification, as well as in many types of cancers, including colon and breast cancers (Kondo and Issa 2004; Table 1). Due to the increase in methylation of the ESR1 gene with age, it has been hypothesized that hypermethylation of ESR1 in cancers may simply reflect the stochastic predisposition of the ESR1 gene to become methylated with progressive rounds of DNA replication (Velicescu et al. 2002), and thus may not be of consequence for the tumorigenic process (Kondo and Issa 2004). Indeed, breast tumors from older individuals are more likely to have ESR1 gene promoter methylation, whereas ESR1 methylation is less frequently methylated in women that develop breast cancer at younger age (M. Campan, D.J. Weisenberger, Q. Feng, S.E. Hawes, N.B. Kiviat, P.W. Laird, manuscript in preparation). Interestingly, the accumulation of ESR1 methylation does not appear to continue after cells become malignantly transformed. The majority of breast cancers do not have high levels of ESR1 methylation, despite the high rate of proliferation, characteristic for tumor cells. Loss of ESR1 expression has been recently shown to induce changes in the chromatin structure of the PGR gene and of many other downstream targets of the estrogen-signaling pathway, with accompanying promoter hypermethylation and transcriptional silencing (Leu et al. 2004). These results suggest that epigenetic inactivation of ESR1, even due to age-related stochastic events, can have important biological consequences that can result in tumorigenesis, by disrupting important growth regulatory pathways. These results also suggest that DNA methylation changes can be pathway specific, rather than as a consequence of stochastic processes, and this may help to explain the existence of tumor type-specific DNA methylation profiles. Based on these findings, it would be interesting to determine if the observed age-related methylation of the ESR1 gene is caused by similar mechanisms, as a consequence of reduction in the estrogen levels, as occurs during menopause in women. Almost all breast cancers show some degree of DNA methylation at the ESR1 gene promoter (Wid-

Table 1 DNA methylation frequencies of select genes in breast, endometrial, ovarian, colon, and proximal colon cancers. The methylation frequency for each gene represents a weighted average of methylation frequencies when multiple reports were available

HUGO

Chromo

Frequency (%)

References3

gene symbol

somal

Breast

Endo-

Ovarian

Colon

Proximal

location

cancer

metrial cancer

cancer

cancer

colon cancer

APC

5q21

28

37

14

21

52

[9, 33, 11, 16, 25, 53, 22, 8, 21, 35, 54-56]

ARF

9p21

20

16

13

30

29

[9, 11, 19, 7,41,42, 14,22,4,21,40,54-56]

BRCA1

17q21

16

ND

17

0

ND

[9, 11, 49, 33, 2, 5, 43, 36, 14, 3,47, 54, 56]

CDH1

16q22

41

26

26

49

64

[49, 9, 11, 2, 33, 35, 26, 30, 36, 22, 21, 13, 54-56]

CDKN2A

9p21

17

16

8

30

27

[9, 11, 19,7,41,42, 14,22,4,21,40,9,11,33, 19,42, 48, 51,43, 14, 23, 29,45, 22, 8, 39, 31,4, 21, 50, 56]

ESR1 promoter A

6q25.1

49

1

29

81

ND

[49, 20, 33, 2, 28, 38, 32, 8, 15, 54-56, 54-56b]

ESR1 promoter C

6q.25.1

ND

94

ND

ND

ND

[38]

ESR2

14q21

52

0

22

22

ND

[38, 49, 54-56b]

GSTP1

11q13

29

0

3

4

9

[2, 9, 11,33, 19, 10,36,21,54-56]

MGMT

10q26

8

0

4

38

29

[9, 11,49,21,22,54-56]

MLH1

3p21

29

41

10

20

40

[9, 11,49, 34, 27, 12,48,44,43, 17,45, 22, 8,24,21, 55, 56]

PGR promoter A

11q22

ND

0

ND

37

ND

[37,56]

PGR promoter B

11q22

66

75

0

80

ND

[37, 49,54,55]

HUGO Chromo- Frequency (%)

gene symbol somal Breast Endo- Ovarian Colon Proximal location cancer metrial cancer cancer colon cancer cancer

References3

PTGS2 RASSF1

3p21 77 ND 31 19 ND [2,18,34,49,6,1,19,36,14,52,21,54-56]

HUGO, The Human Genome Organisation; ND, not determined a References: 1Agathanggelou et al. 2001; 2Bae et al. 2004; 3Baldwin et al. 2000; 4Burri et al. 2001; 5Catteau et al. 1999; 6Dammann et al. 2001; 7Dominguez et al. 2003; 8Eads et al. 1999; 9Esteller et al. 2001a; 10Esteller et al. 1998a; 11Esteller et al. 2001b; 12Esteller et al. 1998b;13Garinis et al. 2002; 14Ibanez de Caceres et al. 2004; 15Issa et al. 1994; 16Jin et al. 2001; 17Kane et al. 1997; 18Kondo and Issa 2004; 19Krassenstein et al. 2004; 20Lapidus et al. 1996; 21Lee et al. 2004; 22Lind et al. 2004; 23McCluskey et al. 1999; 24Miyakura et al. 2001; 25Moreno-Bueno et al. 2002; 26Moreno-Bueno et al. 2003; 27Murata et al. 2002; 28Navari et al. 2000; 29Niederacher et al. 1999; 30Nishimura et al. 2003; 31Norrie et al. 2003; 32O'Doherty et al. 2002; 33Parrella et al. 2004; 34Paz et al. 2003; 35Pijnenborg et al. 2004; 36Rathi et al. 2002; 37Sasaki et al. 2001a;38 Sasaki et al. 2001; 39Schneider-Stock et al. 2003; 40Shen et al. 2003; 41Silva et al. 2001; 42Silva et al. 2003; 43Strathdee et al. 2001; 44Strathdee et al. 1999; 45Toyota et al. 1999; 46Toyota et al. 2000b; 47Wang et al. 2004; 48Whitcomb et al. 2003; 49Widschwendter et al. 2004; 50Wiencke et al. 1999; 51Wong et al. 1999; 52Yoon et al. 2001; 53Zysman et al. 2002; 54M. Campan, D.J. Weisenberger, Q. Feng, S.E. Hawes, N.B. Kiviat, P.W. Laird, manuscript in preparation; 55Ehrlich et al. 2006; 56D.J. Weisenberger, K. Siegmund, M. Campan, J. Young, T.I. Long, M.A. Faasse, G.H. Kang, M. Widschwendter, D. Weener, D. Buchanan, H. Koh, L. Simms, M. Barker, B. Leggett, J. Levine, A.J. French, S.N. Thibodeau, J. Jass, R. Haile, P.W. Laird, submitted b Methylation frequencies from cell lines have been included in the analysis schwendter et al. 2004). However, only 30% of these tumors may have high enough levels of DNA methylation at this locus (Bae et al. 2004; Lapidus et al. 1996; Parrella et al. 2004) that can result in loss of ESR1 expression. This is in agreement with the finding that two-thirds of breast cancers express ERa.

The hormonal receptor (HR) status of breast tumors, defined by the presence or absence of ER and PR, constitutes an important indicator of response to therapy and survival. For instance, patients with ER+ breast tumors have better survival rates, respond better to anti-estrogenic therapy, and are less likely to have tumor recurrence than those with ER- or ER-/PR- breast tumors (Li et al. 2003). We have recently shown an association between DNA methylation changes of ESR1 and PGR in breast tumors and the HR status and response to anti-estrogenic therapy (Widschwendter et al. 2004). A molecular profiling of breast tumors, using DNA methylation profiles, identified two distinct groups of tumors that differed with respect to their HR status. DNA methylation of neither of the two HR genes was the best predictor of the overall HR status, suggesting either a complex interplay between hormone receptor gene methylation and hormone receptor status, or that DNA methyl-ation markers may not always correlate with the gene expression status, due to threshold effects. Nevertheless, we found that ESR1 methylation was the best predictor of PR status and of response to tamoxifen treatment, whereas the methylation of PGR was the best predictor of ER status. Interestingly, the association between ESR1 methylation and PR status was not inversely correlated, even though estrogen signaling is known to activate PGR expression (Widschwendter et al. 2004). It is clear that a full understanding of the relationship between hormone receptor gene expression and DNA methyl-ation, and its role in breast carcinogenesis will require further investigation. ESR1 methylation was also shown to be associated with the methylation of CDH1, GSTP1, CCND2, and TRp1, suggesting that this association may represent a molecular signature of a specific subset of breast cancers with yet-unidentified common phenotypic characteristics (Nass et al. 2000; Parrella et al. 2004).

The methylation status of the ESR2 gene promoter in normal or various cancerous tissues has not yet been investigated to a large extent. As in the case of ESR1, low levels of ESR2 gene promoter methylation are detected in the majority (79%) of breast cancers (Widschwendter et al. 2004), although only about 10% may have high levels of methylation (M. Campan, D.J. Weisenberger, Q. Feng, S.E. Hawes, N.B. Kiviat, P.W. Laird, manuscript in preparation).

0 0

Post a comment