For personal use only.
MECHANISMS OF EPIGENETIC SILENCING OF THE c21 GENE IN Y1 ADRENOCORTICAL TUMOR CELLS
Moshe Szyf* and Andrew D. Slack Department of Pharmacology and Therapeutics McGill University, 3655 Sir William Osler Promenade Montreal, Quebec H3G 1 Y6, Canada
ABSTRACT
We utilized Y1 adrenocortical carcinoma cell line as a model system to dissect the events regulating epigenomic gene silencing in tumor cells. We show here that the chromatin structure of c21 gene is inactive in Y1 cells and that it could be reconfigured to an active form by either expressing antisense mRNA to DNA methyltransferase 1 (dnmt1) or an attenuator of Ras protooncogenic signaling hGAP. Surprisingly however, the known inducer of active chromatin structure the histone deacetylase inhibitor trichostatin A TSA fails to induce expression of c21. These results suggest that the primary cause of c21 gene silencing is independent of histone deacetylation. We present a model to explain the possible roles of the different components of the epigenome and the DNA methylation and demethylation machineries in silencing c21 gene expression.
INTRODUCTION
Cellular transformation involves suppression of genes that block cellular growth such as tumor suppressors, activation of genes that promote cellular growth such as oncogenes and silencing of genes that encode specialized functions in differentiated tissues. We utilized the adrenocortical carcinoma cell line Y1 to study the common mechanisms that control both the suppression of cell growth functions and specialized gene expression during cellular transformation.
For personal use only.
Yl cells have lost the capacity to express the c21 gene (steroid-21-hydroxylase: Cyp21) (2). c21 is silenced in Y1 cells by an epigenetic mechanism (3) and is modified by DNA methylation.
DNA methylation of the cytosines found at the CG dinucleotide sequence is an epigenetic modification of DNA that marks the silenced genes in our genome in a tissue specific manner (4,5). DNA methylation silences genes by either preventing the binding of some transcription factors that interact with a sequence that bears a CG dinucleotide (6) or by attracting methylated DNA binding proteins that can precipitate an inactive chromatin structure by attracting corepressors and histone deacetylases HDAC to the promoter area (7).
We had previously studied the regulation of the gene encoding the enzyme responsible for replicating the DNA methylation pattern the DNA methyltransferase 1 dnmt1 in Y1 cells (8). We found that the dnmtl gene is upregulated in Y1 cells, that dnmtl is responsive to the Ras-Ap-1 signaling pathway that is upregulated in Y1 cells and that dnmtl plays a causal role in the transformation program in Y1 cells (9). Inhibiting dnmtl by dnmtl antisense oligonucleotides in mice bearing Y1 tumors inhibits tumor growth (10). A simple and attractive model that can explain both the silencing of differentiated genes like c21 in Y1 cells and tumor transformation is that ectopic induction of dnmlt (11,12) causes aberrant methylation and epigenetic modification of tumor suppressors (12) and other genes leading to the transformed phenotype (11,12).
Surprisingly however, our recent data suggests that dnmtl can control the p21 tumor suppressor gene expression by a mechanism that does not involve DNA methylation (13). We therefore proposed that dnmtl controls expression of some tumor suppressors and perhaps movement through the cell cycle by protein protein interactions (14). Two critical protein-protein interactions of dnmtl were recently described, its interaction with the replication fork protein PCNA (15) and its interaction with HDAC1 (16). We have also shown that inhibition of dnmtl results in inhibition of DNA replication (17) and proposed that induction of p21 results in inhibition of DNA replication (14).
For personal use only.
In this study we test the hypothesis that histone deacetylation triggered by DNA methylation and chromatin structure are the primary causes of inactivation of c21 in Y1 cells.
MATERIALS and METHODS
Cell lines and cell culture-Y1 cells culture conditions and the transfectants used in this study were previously described. Y1 were transfected with an antisense dnmt1 mRNA (9). Y1 pZEM were transfected with the control vector (9) and Y1 GAP were transfected with a cDNA expressing the human GAP cDNA (8).
RESULTS
Inhibition of either dnmtl or the Ras signaling pathway induces changes in the chromatin structure the c21 gene
Methylated DNA binding proteins such as Mecp2 recruit histone deacetylases to methylated genes (7). Histone deacetylases modify histones by removing the acetyl groups on lysines residing in the N-terminal tails (18). The deacetylated histones are believed to form tight contact with DNA resulting in a closed and inactive chromatin structure (18). The state of openness of the chromatin at a specific gene locus is determined by measuring the sensitivity of the chromatin bound gene to endonucleases such as DNAsel (19) or restriction enzymes such as MspI (20). We therefore determined the sensitivity of chromatin bound c21 in either Y1 cells, Y1 cells transfected with a control vector neo, Y1 cells transfected with a dnmtl antisense (9) or Y1 cells transfected with hGAp cDNA (8) to increasing concentrations of the endonuclease Mspl. We have previously shown that expression of a dnmtl antisense mRNA and ectopic expression of hGAp which attenuates Ras signaling results in demethylation of c21 (9,10). As observed in Fig.1, the chromatin bound c21 gene in Y1 cells or Y1 cells transfected with the pZEM control vector is highly resistant to MspI digestion and
Y1
Y1 pZEM
Y1 GAP 7
Y1 ANT 9
Ex.C21 Ex.C21 Meth Non meth
Msp l:
21 -
…
Probe: 5’ C21
5.1-
3.5-
1.6-
kb
Probe:
21
…
3.8 C21
5.1- 3.5-
1.6
kb
Nuclei were prepared from Y1 cells, Y1 transfected with pZem control vector (YlpZEM), Y1 transfected with an antisense dnmtl mRNA (YIANT 9) Y1 transfected with the Ras attenuator hGAP (Y1GAP 7), Y1 transfected with an unmethylated cloned c21 gene (Ex. C21 Non Meth) or Y1 transfected with an in vitro methylated c21 gene (Ex. C21 Meth). The nuclei were incubated with increasing concentrations of MspI restriction enzyme (0,5,10,100 and 500 u/ml) for 1 h at 37℃. DNA was isolated from the treated nuclei and subjected to a Southern blot analysis using either a 5’ or a 3’ region of the c21 gene.
is cleaved by MspI only at high enzyme concentrations. However in Y1 cell expressing an antisense to dnmtl or hGAP, the chromatin bound c21 gene is sensitive to MspI at lower concentrations. Similarly an exogenously transfected unmethylated c21 gene is more sensitive to MspI digestion than a methylated exogenous c21. These results indicate that the c21 gene is found in an inactive chromatin structure in Y1 cells and that this inactive chromatin structure is dependent on DNA methylation as expected (5,7,18). Inhibition of dnmtl or its upstream regulators results in activation of chromatin structure.
A.
Adrenal
Y1 3UM TSA
B.
Y1 3UM TSA
Y1
Y1
500
400
45
300
+ C21
35
200
18
+ Acetyl-lysine H3
bp
8
+ Acetyl-lysine H4
Kd
FIGURE 2. Inhibition of histone deacetylase does not activate the c21 gene in Y1 cells. A Western blot analysis of whole cell extracts prepared from Y1 cells and Y1 cells treated with 3uM of TSA. The membrane was reacted with a monoclonal antibody against acetyl-lysine (Upstate Biotechnology 06-933). As observed, histones are hyperacetylated following TSA treatment. B. RNA was prepared from adrenal tissue, Y1 cells and Y1 cells treated with 3uM of TSA. The RNA was reverse transcribed and c21 mRNA was amplified using c21 specific amplimers as previously described.
Does chromatin structure play a causal role in the silencing of c21 in Y1 cells?
To determine whether the primary cause of repression of c21 in Y1 cells is its inactive chromatin structure, we took advantage of the specific histone deacetylase inhibitor trichostatin A (TSA) (21) which has been shown to increase histone acetylation and reconfigure the chromatin into an active structure and activate multiple genes (21). The results presented in fig. 2 show however that inhibition of histone deacetylation by TSA results in histone hyperacetylation (Fig. 2A) as determined by a Western blot analysis but does not result in induction of c21 expression as determined by a sensitive RT-PCR assay (3). These results suggest that the primary cause of c21 repression is not the
inactivation of chromatin structure brought about by either DNA methylation or a repressor complex.
DISCUSSION
Our studies of the regulation of DNA methylation in Y1 cells in the last decade have shown that aberrant regulation of dnmtl and methylation of specific genes such as c21 are tightly associated with the transformation process (3,8,9,10,22). We have shown that dnmtl is regulated by the nodal Ras signaling pathway indicating that aberrant dnmtl expression is a downstream effector of a transformation program (8,9,10,11,23). These data lend support to the simple and attractive hypothesis that the mechanism through which increased dnmtl activity leads to transformation is mediated by hypermethylation of DNA which is responsible for silencing of tumor suppressors as well as other specific genes such as c21 (11,23). However our recent experiments in human cancer lines suggest that ectopic expression of dnmtl might be involved in transformation by a mechanism that does not involve DNA methylation (11, 17, 14, 23). I previously proposed that dnmtl transform cells through its protein-protein interactions in the replication fork (14, 23).
Our results show that although the maintenance of the inactive chromatin structure at the c21 gene locus is dependent on the state of its methylation (fig. 1), as predicted by the current understanding of the relationship between DNA methylation and chromatin structure, inhibition of histone deacetylation (fig. 2) does not result in activation of c21. These results indicate that an additional mechanism, perhaps an active repressor suppress the c21 gene and that this repressor can act independent of the chromatin structure.
Do these result imply that DNA methylation does not play a role in the silencing of c21 and similar genes in transformed cells? The fact that complete methylation of c21 plays a causal role in the chromatin structure of c21 (fig. 1) and that c21 is specifically methylated in Y1 cell even when exogenously introduced into the cells (23) is consistent with a role for DNA methylation in the
HDAC inactive chromatin
Mbd binding methylation
HDAC Inactive chromatin
HDAC Inactive chromatin
Mbd binding ☒ methylation
press
r
c21 5
c21 5
Methylation and silencing of c21, a model.
We tested two alternative models of how c21 might be repressed by epigenetic factors in Y1 cells. According to the first model (left panel) methylation of c21 in Y1 cells attracts the binding of methylated DNA binding proteins (Mbd) which bind HDACs resulting in deacetylation of chromatin and silencing of the gene. This model is inconsistent with the observation that inhibition of HDAC by TSA (fig. 2) does not activate c21. An alternative model that is consistent with our data is presented on the left panel. A trans- acting repressor binds the c21 gene and represses it. The repressed state of the gene attracts HDACs to the gene. The inactive chromatin is inaccessible to demethylases resulting in progressive methylation of the c21 gene. This in turn amplifies the inactive state by attracting Mbds as discussed above.
silencing program. A possible role for DNA methylation is that it fixes by a stable covalent modification the inactive state induced by a trans-acting repressor. DNA methylation and resulting changes in chromatin structure are proposed to stabilize and enhance the inactive state of the gene. Demethylation and histone deacetylation do not result in activation of c21 as long as the primary cause, an active repressor is still present in the cell (fig. 3). Our previous data shows that methylation of c21 follows its inactivation in Yl cells supporting the model proposed here (3).
Multiple paradigms of regulation of genes by the epigenome are emerging from recently published data. A proposed unifying hypothesis is that genes are regulated by a dynamic equilibrium between competing histone deacetylases and
For personal use only.
histone acetylases as well as DNMT and demethylase. These modifications of histones and DNA are interdependent on each other and are also influenced by the interaction of transacting factors with the gene. The relative contributions of the histone and DNA modifications versus the role of trans-acting factors vary between the different paradigms.
One obvious question is why does the inactive c21 gene become hypermethylated? Our data derived from experiments in other cancer cell lines might provide an answer to this question. Traditionally methylation of DNA was looked upon from the point of view that only DNMTs are responsible for maintenance of DNA methylation in somatic cells. We however have recently discovered that mammalian cells bear a bona fide demethylase activity (27,28). We have also shown that Y1 cells express a demethylase activity and bear a cDNA encoding demethylase activity (unpublished data). Thus, there is a tone of demethylase activity in Y1 cells in addition to DNMT activity. The methylation pattern reflects the balance of the methylation and demethylation tones in a cell. We have recently shown that the access of demethylase to genes depends on the chromatin structure. Inactive genes are relatively inaccessible to demethylase. It is possible that the progressive increase in methylation of c21 gene reflects the increased resistance of the gene to the demethylating tone. While additional experiments are required to further test these hypotheses, the silencing of c21 in Y1 cells provides an interesting model system to address these questions.
ACKNOWLEDGEMENTS
We thank V. Bozovic and J. Theberge for excellent technical assistance. The studies reported here were funded by the MRC Canada. Address correspondence to - Dr. Moshe Szyf, Tel: 514-398-7107, FAX: 514-398-6690
E. mail: mszyf@pharma.mcgill.ca
REFERENCES
1. Yasumura Y, Buonassisi V, Sato G 1966 Cancer Res 26:529-535.
2. Schimmer BP 1981 In: Sato G (ed) Functionally differentiated Cell Lines, Lis, New York, pp 61-92
3. Szyf m, Milstone DS, Schimmer BP, Parker KL, Seidman JG 1990 Mol. Endocrin. 4:1144-1152
4. Razin A, Riggs AD 1980 Science 210:604-610
5. Razin A, Cedar H 1991 Microbiol. Rev. 55:451-458
6. Comb M, Goodman HM 1990 Nucleic Acids Res. 18:3975-3982.
7. Nan X, Ng H-H, Jonson CA, Laherty CD, Turner. BM, Eisenman RN, Bird A 1998 Nature 393: 386-389
8. MacLeod AR, Rouleau J. Szyf M 1995 J. Biol. Chem. 270:11327-11337.
9. Macleod AR, Szyf M 1995 J. Biol. Chem. 270:8037-8043.
10. Ramchandani S, MacLeod AR, Pinard M, von Hofe E,, Szyf M 1997 Proc. Natl. Acad. Sci. USA 94: 684-688
11. 1. Szyf M 1994 Trends Pharmacol. Sci. 7: 233-237.
12. Merlo A, Herman JG, Mao L, Lee DJ, Gabrielson E, Burger P, Baylin SB, Sidransky D 1995 Nature Med. 1: 686-692
13. Milutinovic S, Knox DJ, Szyf M 2000 J. Biol. Chem. 275:6353-6359
14. Szyf M, Knox DJ, Milutinovic S, Slack AD, Araujo FD 2000 Ann. NY. Acad. Sci. 910:156-177
5. Chuang L,s-H, Ian H-I, Koh T-W, Ng H-H, Xu G, Li BFL 1997 Science 277:1996-2000
16. . Fuks F, Burgers WA, Brehm A, Hughes-Davies L, Kouzarides T 2000 Nature Genet. 24:88-91
17. Knox DJ, Araujo FD, Bigey P, Slack AD, Price GP, Zannis-Hadjopoulos M, Szyf M 2000 J. Biol. Chem. 2675: 17986-17990
18. Ng HH, Bird A 2000 Trends Biochem. Sci. 25:121-126
19. Garel A, Axel R 1976 Proc. Natl. Acad. Sci. USA 73:3966-3970
20. Tazi T, Bird A 1990 Cell 60:909-920
21. Bartsch J, Truss M, Bode J, Beato M 1996 Proc. Natl. Acad. Sci. USA 93:10741-10746
22. Szyf M, Schimmer BP, Seidman JG, 1989 Proc. Natl. Acad. Sci. USA 86:6853-6857
23. Szyf M 1998 Cancer Metast. Rev. 17: 219-231
24. Yang ZY, Perkins ND, Ohno T, Nabel EG, Nabel GJ 1995 Nat. Med. 1:1052-1056
25. Waga S, Hannon GJ, Beach D, Stilman B 1994 Nature 369:574-578
26. Szyf M, Bozovic V, Tanigawa G 1991 J. Biol. Chem. 266:10027-10030
27. Bhattacharya SK, Ramchandani S, Cervoni N, Szyf M 1999 Nature 397:579-583.
28. Ramchandani S, Bhattacharya SK, Cervoni N, Szyf M 1999 Proc. Natl. Acad. Sci. USA 96:6107-6112
Endocr Res Downloaded from informahealthcare.com by CDL-UC Davis on 10/29/14 For personal use only.