Cancer Research Cancer Epigenetics Telomeres
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
Cancer Research 66, 10308, November 1, 2006. doi: 10.1158/0008-5472.CAN-06-1052
© 2006 American Association for Cancer Research
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bistulfi, G.
Right arrow Articles by Sacchi, N.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bistulfi, G.
Right arrow Articles by Sacchi, N.

Molecular Biology, Pathobiology, and Genetics

A Repressive Epigenetic Domino Effect Confers Susceptibility to Breast Epithelial Cell Transformation: Implications for Predicting Breast Cancer Risk

Gaia Bistulfi, Silvia Pozzi, MingQiang Ren, Stefano Rossetti and Nicoletta Sacchi

Department of Cancer Genetics, Roswell Park Cancer Institute, Elm & Carlton Streets, Buffalo, New York

Requests for reprints: Nicoletta Sacchi, Roswell Park Cancer Institute, Elm and Carlton Streets, Buffalo, NY 14263. Phone: 716-845-1053; Fax: 716-845-1741; E-mail: nicoletta.sacchi{at}roswellpark.org.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Retinoic acid (RA) is a master epigenetic regulator that plays a pivotal role in both breast morphogenesis and development. Here, we show for the first time that RA, via the RA receptor {alpha} (RAR{alpha}), epigenetically regulates in a concerted fashion the transcription of two RA-responsive genes, the RA receptor ß2 (RARß2) and the cellular retinol-binding protein 1 (CRBP1). Specifically, an impaired RA signal through RAR{alpha} in human breast epithelial cells triggers a repressive epigenetic domino effect, involving first RARß2 and second CRBP1. The phenotype acquired by breast epithelial cells clearly implies that the resistance to RA-mediated growth inhibition precedes the acquisition of morphological epithelial transformation, thus supporting the occurrence of sequential transcriptional silencing of first RARß2 and second CRBP1. The identification of this epigenetic network mechanistically linking RARß2 and CRBP1 transcription provides the basis for devising more accurate epigenetic tests for the prediction of breast cancer risk.(Cancer Res 2006; 66(21): 10308-14)


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Retinoic acid (RA), the bioactive derivative of vitamin A or retinol, is a master epigenetic regulator of gene transcription (1), which plays a pivotal role in postpuberty mammary gland morphogenesis and development (2, 3). Once integrated through the RA receptor {alpha} (RAR{alpha}), RA signal elicits chromatin modifications that enable transcription of the RA receptor ß2 (RARß2; refs. 1, 4). RARß2 would then sustain its own transcription (5–7) and the transcription of a few downstream RA-responsive target genes.

By stably impairing the integration of RA signal through RAR{alpha} in RA-sensitive breast cancer cells, we found recently that RARß2 falls into an aberrant transcriptional inactive status, which is marked by a critical level of repressive chromatin modifications, including, but not limited to, DNA hypermethylation (8). Concomitant with the conversion of RARß2 alleles from a permissive transcriptional status into a nonpermissive status, cells are converted to RA resistance (8). These findings led us to originally speculate that the propensity to RARß2 epigenetic silencing and RA resistance could be consequent to an altered retinol/RA metabolism, capable of creating an aberrant RA signal through RAR{alpha} (8, 9). Indeed, several proteins involved in either retinol or RA metabolism/transport can be found deranged or down-regulated in breast cancer cells (10–13). One of these proteins is the cellular retinol-binding protein 1 (CRBP1). In this study, we set out to test whether silencing CRBP1 in RA-sensitive cells could mechanistically lead to RARß2 silencing and RA resistance. Because CRBP1, in addition of being a retinol transport protein, is also involved in the maintenance of apicobasal-differentiated morphology of human breast epithelial cells (14), we chose to knock down CRBP1 by RNA interference (RNAi) in nontransformed human breast epithelial cells (HME1). However, contrary to our hypothesis, we found that CRBP1 knockdown does not induce RARß2 silencing and RA resistance. We found instead that CRBP1 is a downstream RAR-regulated gene, consistent with reports pointing at CRBP1 like one of the few RAR targets (15–19). First, we identified in the human CRBP1 a RA-responsive element (RARE), which is evolutionary conserved (15, 16). Second, we found that both RAR{alpha} and RARß2 can bind the human CRBP1-RARE region. Finally, when we impaired the integration of RA through either RARß2 or RAR{alpha} in HME1 cells, we observed the conversion of CRBP1 alleles permissive for transcription into alleles nonpermissive for transcription. A fraction of nonpermissive alleles showed aberrant DNA hypermethylation. Apparently, an impaired integration of RA signal through RAR{alpha} leads to a repressive epigenetic "domino effect," involving first RARß2 and second CRBP1. The phenotypic analysis of HME1 clones, showing that the resistance to RA-mediated growth inhibition precedes the acquisition of morphologic phenotypes of epithelial transformation, further supported the occurrence of sequential transcriptional silencing of first RARß2 and second CRBP1.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Cells and cell cultures. The human telomerase-immortalized nontransformed breast epithelial cell strain HME1 (Clontech, Mountain View, CA) was grown in mammary epithelial growth medium (MEGM) plus bovine pituitary extract (Cambrex, Walkersville, MD) as per manufacturer's instructions. The monkey kidney COS cell line and the human breast cancer MDA-MB-231 cell line (American Type Culture Collection, Manassas, VA) were grown in DMEM plus 5% fetal bovine serum (Invitrogen, Carlsbad, CA).

HME1 cells and derived clones were grown on reconstituted basement membrane in three-dimensional cultures to induce breast epithelial differentiation into acini-like structures essentially as described (20). Briefly, single cells were induced to form acini on chamber slides coated with Matrigel (BD Biosciences, San Jose, CA) in medium plus 2% Matrigel for 10 to 15 days. After fixation with 4% paraformaldehyde, the Golgi apparatus was stained with anti-GM 130 antibody (Ab) (1:400; BD Biosciences) followed by goat anti-mouse Alexa Fluor 546 Ab (1:500; Molecular Probes, Eugene, OR). Integrin was stained with anti-CD49f Ab (1:200; Chemicon, Temecula, CA) followed by anti-rat Alexa Fluor 488 Ab (1:400; Molecular Probes). Nuclei were counterstained with 300 nM 4',6-diamidino-2-phenylindole (Sigma, St. Louis, MO). At least 30 acini per each clone were analyzed by confocal microscopy (SP2 spectral confocal microscope, Leica, Microsystem, Wetzlar, DE) to inspect for the presence of a hollow lumen and apicobasal polarization. The phenotype observed in 70% or more of the acini was considered to be the prevalent phenotype.

For the colony formation assay, exponentially growing cells were seeded at 3 x 102 per well in six-well plates and allowed to attach for 48 hours. After treatment with either 0.1 µM RA or vehicle (ethanol) for 24 hours, the medium was replaced with drug-free medium and cells were allowed to grow for 10 to 14 days, until the appearance of colonies was observed. Colonies were then fixed with methanol, stained with Giemsa, and scored to establish the colony formation index.

For the anchorage-independent colony formation assay, 1 x 104 cells per well resuspended in semisolid medium containing 0.3% agarose were layered on 0.5% agarose in six-well plates and cultured for 4 weeks. Foci were counted in 10 random fields per each triplicate sample under a Nikon (Melville, NY) Eclipse E600 microscope. Each experiment was repeated three times.

Drugs. RA and 5-aza-2'-deoxycytidine (5-Aza; Sigma) were dissolved in 100% ethanol and 0.45% NaCl containing 10 nM sodium phosphate (pH 6.8), respectively, and stored in aliquots at –80°C in the dark.

RNAi. Short hairpin RNA (shRNA) sequences targeting CRBP1, RARß2, and RAR{alpha} mRNAs were cloned into the pSUPER-retro vector (Oligoengine, Seattle, WA). The sequences targeted by RNAi were as follows: CRBP1-A, 5'-GTGCATGACAACAGTGAGC-3'; CRBP1-B, 5'-GGTGTGGTCTGCAAGCAAG-3' (Genbank NM_002899, nucleotides 371-389 and 495-513, respectively); RARß2-A, 5'-GCTGGCTTGTCTGTCATAA-3'; RARß2-B, 5'-GGGGCAGAGTTTGATGGAG-3' (Genbank NM_000965.2, nucleotides 303-321 and 371-389, respectively); and RAR{alpha}-B (8), 5'-AGCGCACCAGGAAACCTTC-3' (Genbank NM_000964, nucleotides 681-699). The control mock sequence 5'-ACGTACGTACGTAGTGGGG-3', which should not recognize any human mRNA, was cloned in the pSUPER-retro vector and used as a control. The silencing effect of each of these sequences was tested by transient cotransfection with the cognate exogenous cDNAs in COS cells. The cDNA sequences included the following: (a) the RAR{alpha}1 cDNA cloned into pSG5 plasmid (kindly provided by Dr. Fausto Andreola, NCI, Bethesda, MD); (b) the CRBP1 cDNA cloned from MDA-MB-468 cell line in frame with the TAG sequence present into the pCMV-TAG vector (Stratagene, La Jolla, CA) after PCR amplification using the primers 5'-TATGGAATTCCTGGCTCCAGTCACTCCCGAA-3' (sense) and 5'-TATCTCGAGCTGATTGGTTGGGACAAGGTTGTCT-3' (antisense) containing restriction sites for EcoRI and XhoI, respectively; and (c) the RARß2 cDNA cloned from T47D into pCDNA3.1+ (Invitrogen) after PCR amplification using the primers 5'-TATGGATCCGCAAGGGAGATCATGTTT-3' (sense) and 5'-TATAAGCTTTTATTGCACGAGTGGTGACTG-3' (antisense) containing restriction sites for BamHI and HindIII, respectively. Stable transfections in HME1 cells were carried out with Lipofectamine Plus (Invitrogen). Single stable clones were selected in puromycin 1 µg/mL and four clones per each RNAi sequence were selected for further analysis.

Retroviral infection. Supernatants containing either the RAR{alpha} dominant-negative LXRAR{alpha}403SN or the empty LX SN (mock) retroviral particles (kindly provided by Dr. Fausto Andreola) were used to infect HME1 cells as described (21). Four independent clones were used for further analysis.

Real-time reverse transcription-PCR. Total RNA obtained with the single-step method using Trizol (Invitrogen) was treated with DNase I (Ambion, Austin, TX). For each sample, 1 µg total RNA was retrotranscribed with SuperScript First-Strand Synthesis System (Invitrogen) in a 20 µL reaction. One microliter of the so obtained cDNA was then used for each triplicate in real-time reverse transcription-PCR, which was done on an iCycler (Bio-Rad, Hercules, CA) using the iQ SYBR Green Supermix (Bio-Rad) with primers specific for CRBP1 [5'-GGTACTGGAAGATGTTGGTC-3' (sense) and 5'-CATCTCTAGGTGCAGCTCAT-3' (antisense)], RARß2 [5'-GACTGTATGGATGTTCTGTCAG-3' (sense) and 5'-ATTTGTCCTGGCAGACGAAGCA-3' (antisense)], and RAR{alpha} [5'-TGTGGACTTCGCCAAGCA-3' (sense) and 5'-CGTGTACCGCGTGCAGA-3' (antisense)]. The mRNA levels were normalized to the mRNA levels of the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase [GAPDH; 5'-GAAGGTGAAGGTCGGAGTC-3' (sense) and 5'-GAAGATGGTGATGGGATTTC-3' (antisense)].

Western blot. Western blot analysis of RARß2, FLAG-CRBP1, and GAPDH protein expression was done with standard protocols using an anti-RARß2 antibody (Santa Cruz Biotechnology, Santa Cruz, CA), an anti-FLAG M2 monoclonal antibody (Sigma) recognizing the TAG encoded by the CMV-TAG vector, and an anti-GAPDH antibody (Santa Cruz Biotechnology). Appropriate horseradish peroxidase–conjugated secondary antibodies (Amersham, Piscataway, NJ) were used for detection.

Quantitative chromatin immunoprecipitation. Chromatin immunoprecipitation (ChIP) was done using reagents purchased from Upstate (Lake Placid, NY) according to the manufacturer's protocol. Chromatin was immunoprecipitated with antibodies against acetyl histone H4 (Upstate), RARß [Santa Cruz Biotechnology and Active Motif (Carlsbad, CA)], and RAR{alpha} (Santa Cruz Biotechnology). The immunoprecipitated DNA was amplified by real-time PCR with specific primers encompassing the CRBP1-RARE [5'-AGCCTGCACTGTGAGAACACAT-3' (sense) and 5'-CCACCAAGTAGATGACATAATCA-3' (antisense)], a control region 8.7 kb upstream of the CRBP1-RARE [5'-GGACCTTGGTGAGTGGAGGATA-3' (sense) and 5'-CCCTGCAGGTCCTCCACTAT-3' (antisense)], and the GAPDH promoter region [5'-GGTGCGTGCCCAGTTGAACCA-3' (sense) and 5'-AAAGAAGATGCGGCTGACTGTCGAA-3' (antisense)]. The relative enrichment of immunoprecipitated DNA was calculated by normalizing the PCR signals of the samples to both the input and the no antibody controls. In ChIP experiments with anti-acetyl histone H4 antibody, we used as an internal control the amplification of the GAPDH promoter. In control ChIPs with anti-RARß and anti-RAR{alpha} antibodies, we amplified a region 8.7 kb upstream of the CRBP1-RARE as a negative control.

DNA methylation analysis. Genomic DNA was extracted with DNAzol (Invitrogen) and modified with sodium bisulfite as described previously (22). We amplified by seminested PCR a 449-bp region of the CRBP1 CpG island encompassing 45 CpG sites using specific primer sets [5'-AGGTTTTAGATAAAGGTTTGTAAGTG-3' (sense-1), 5'-GTGGTTGTTGAGTGTGAGAAG-3' (sense-2), and 5'-CACCAAACCACAACTCACCAAA-3' (antisense)] and a region encompassing a CpG stretch, which we defined as the RARß2 methylation epicenter (RME; ref. 8). To sequence single alleles, the PCR products were cloned into the pCR4-TOPO plasmid vector (Invitrogen).

In silico analysis of human CRBP1. To analyze the DNA region 5' to hCRBP1, the hCRBP1 transcript sequence from the Ensemble database (ENST00000232219) and the National Center for Biotechnology Information (NCBI) database (NM_002899.2) were compared with both the sequence encoding human DNA for CRBP1 and exons 1 and 2 (and joined coding sequence) from the NCBI database (X07437), and the complete sequence of the human 3 BAC RP11-553K23 clone (Roswell Park Cancer Institute Human Bacterial Artificial Chromosome Library, Buffalo, NY) deposited at the NCBI database (AC046134).

Statistical analysis. Statistical significance was calculated based on three independent experiments using the Student's t test.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
CRBP1 knockdown in HME1 cells does not lead to RARß2 silencing and RA resistance. CRBP1 is a protein critically involved in the transport and mobilization of retinol, the precursor of RA (23, 24). Previously, we speculated that lack/derangement of proteins involved in retinol/RA metabolism could alter the intracellular concentration of RA, thus leading to RARß2 epigenetic silencing (8, 9). The findings presented here discount that this might be the case for CRBP1. When we simulated the occurrence of CRBP1 epigenetic down-regulation in HME1 cells by knocking down CRBP1 transcription by RNAi with two sequences, CRBP1-A and CRBP1-B (Fig. 1A, top ), selected to efficiently silence an exogenously expressed human CRBP1 protein in COS cells (Fig. 1A, bottom), we did not induce RARß2 silencing, and consistently, we did not observe the conversion to RA resistance (Fig. 1C). Four independent HME1 clones, two per each sequence, showing significant CRBP1 transcriptional down-regulation relative to a prototypic control clone, mock-7, carrying a scrambled sequence, were analyzed. Here, we show that for two representative clones, Si-Crbp1-A1 and Si-Crbp1-B4, which were stably transfected with the sequences CRBP1-A and CRBP1-B, respectively (Fig. 1B, top, left), the induction of CRBP1 knockdown confers the two phenotypes expected for CRBP1 down-regulation (14, 25), including anchorage-independent growth in soft agar (Fig. 1B, top, right) and impaired apicobasal polarization when grown on basement membrane in three-dimensional cultures. Both clones formed mainly acini with a filled lumen (Fig. 1B, bottom); however, they maintained a transcriptionally active RARß2 (Fig. 1C, left) and remained sensitive to the growth-inhibitory action of RA (Fig. 1C, right).


Figure 1
View larger version (29K):
[in this window]
[in a new window]
  		Figure 1. CRBP1 silencing does not induce RARß2 silencing and RA resistance. A, two sequences, CRBP1-A and CRBP1-B (top), cloned in the pSUPER-retro vector and able to efficiently abrogate the expression of exogenous CRBP1 in COS cells transfected with human CRBP1 cDNA (bottom), were stably transfected into HME1 cells. B, two HME1 clones, Si-Crbp1-A1 carrying CRBP1-A and Si-Crbp1-B4 carrying CRBP1-B, showing significant (P < 0.001) CRBP1 transcriptional down-regulation compared with the control clone, mock-7, carrying a scrambled sequence (top, left), displayed both anchorage-independent growth (top, right) and an impaired epithelial polarization in three-dimensional culture [blue, nuclei; red, Golgi apparatus; green, integrin (bottom)]. C, Si-Crbp1-A1 and Si-Crbp1-B4 with unaffected RARß2 transcriptional activity (left) are RA sensitive (right).

 
Evidence that hCRBP1 is a direct RAR target, epigenetically regulated in response to RA. Previous reports implied that CRBP1 might be a direct RAR target gene (15–19). We located by in silico analysis (see Materials and Methods) a human CRBP1-RARE 1,377 bp upstream of the hCRBP1 transcription start site (Fig. 2A ), which is highly homologous to both mouse and rat CRBP1-RARE sequences (15, 16). Second, we found that RA (1 µM, 72 hours) activated both RARß2 and CRBP1 transcription in HME1 cells (Fig. 2B, top) in association with a significant increase of histone H4 acetylation at both the RARß2-RARE and the CRBP1-RARE chromatin regions (Fig. 2B, middle), which is consistent with the scanty DNA methylation in just a minority of HME1 alleles (Fig. 2B, bottom). Conversely, both RARß2 and CRBP1 in MDA-MB-231 cells were transcriptionally silent (Fig. 2C, top), hypoacetylated at histone H4 (Fig. 2C, middle), and showed DNA hypermethylation (Fig. 2C, bottom). The transcription of both genes was concomitantly reactivated by RA (1 µM) after partial reversion of the DNA hypermethylated status by 5-Aza (0.8 µM, 72 hours; Fig. 2C, top). Altogether, these data implied that the transcription of both RARß2 and CRBP1 is epigenetically regulated in a concerted fashion in response to RA. This supposition was reinforced by ChIP analysis with anti-RARß2-specific antibodies, which detected RARß2 binding at the CRBP1-RARE but not at an adjacent 5' control region (Fig. 2D, left). In addition, the level of RARß2 binding increased in response to RA, concomitant with a decrease of RAR{alpha} binding at the CRBP1-RARE (Fig. 2D, right). These data prompted us to test whether CRBP1 transcription is under direct RAR regulation.


Figure 2
View larger version (50K):
[in this window]
[in a new window]
  		Figure 2. CRBP1 is epigenetically regulated in response to RA. A, in silico analysis of the human CRBP1 showing a RARE 1,377 bp upstream of the transcription start site. B, RA significantly induces both transcription (P < 0.001; top) and histone H4 acetylation (P < 0.05; middle) of both RARß2 and CRBP1 in HME1 cells where both genes are not hypermethylated (bottom). C, both RARß2 and CRBP1 remain transcriptionally silent in response to RA in MDA-MB-231 cells (top) consistent with lack of histone H4 acetylation (middle) and the presence of DNA methylation (bottom). RARß2 RA-induced and CRBP1 RA-induced transcription can be concomitantly reactivated after treatment with 5-Aza (top). D, ChIP analysis with either anti-RAR{alpha} or anti-RARß antibodies showing that both RAR{alpha} and RARß2 bind to the CRBP1-RARE in HME1 cells significantly (P < 0.05) more than to an adjacent [control (CTRL)] region 8.7 kb 5' to the RARE. RARß2 binding and RAR{alpha} binding are increased and decreased, respectively, in response to RA.

 
RARß2 knockdown in HME1 cells leads to CRBP1 epigenetic silencing. When we down-regulated RARß2 transcription in HME1 by RNAi, we observed down-regulation of CRBP1 transcription, which remained unresponsive to RA. We used two sequences, RARß2-A and RARß2-B (Fig. 3A, top ), which efficiently silenced RARß2 expression from an exogenous RARß2 cDNA transfected in COS cells (Fig. 3A, bottom), to stably transfect HME1 cells. Four independent clones, two per each sequence, showing significant RARß2 transcriptional down-regulation and RA resistance (Fig. 3B, top, left and right, respectively) yet unaffected levels of RAR{alpha} transcription (Fig. 3B, bottom) relative to a representative control clone, mock-8, were selected for further analysis. The clones tested included Si-ß-A8, carrying the RARß2-A sequence, and Si-ß-B4, carrying the RARß2-B sequence (Fig. 3B, left). In both Si-ß-A8 and Si-ß-B4 CRBP1 alleles were clearly nonpermissive for transcription (Fig. 3C). A fraction of the alleles nonpermissive for transcription (shown here are the alleles of the Si-ß-A8 clone) acquired de novo DNA methylation over the CRBP1 CpG island (Fig. 3D, top and bottom).


Figure 3
View larger version (28K):
[in this window]
[in a new window]
  		Figure 3. RARß2 knockdown in HME1 cells leads to CRBP1 epigenetic silencing. A, the RARß2-A and RARß2-B sequences (top) efficiently silenced exogenous RARß2 expression in COS cells cotransfected with human RARß2 cDNA (bottom). B, two HME1 clones, Si-ß-A8 carrying RARß2-A and Si-ß-B4 carrying RARß2-B, showing specific and significant down-regulation of RARß2 (P < 0.01; top, left) but not of RAR{alpha} (bottom) acquired RA resistance (top, right) compared with the control clone mock-8, carrying a scrambled sequence. C, both Si-ß-A8 and Si-ß-B4 show significant (P < 0.001) CRBP1 transcriptional down-regulation. D, distribution of CpG methylated sites (top) in de novo methylated, nonpermissive alleles of the Si-ß-A8 clone (bottom). E, Si-ß-A8 and Si-ß-B4 clones showing anchorage-independent growth (left) and impaired epithelial cell differentiation in three-dimensional cultures (right).

 
Both Si-ß-A8 and Si-ß-B4 failed to form polarized, hollow lumen acini when grown on reconstituted basement membrane (three-dimensional cultures; Fig. 3E, right), and both displayed anchorage-independent growth (Fig. 3E, left). Apparently, RARß2 transcriptional repression leads to both CRBP1 epigenetic repression and associated morphologic phenotypes. These data led us to test whether an impaired RA signal at RAR{alpha} can trigger a repressive epigenetic "domino effect" involving both RARß2 and its target CRBP1.

An impaired RA signal at RAR{alpha} leads to concomitant RARß2 and CRBP1 epigenetic repression. Interference with either RAR{alpha} transcription by RAR{alpha} RNAi or RAR{alpha} function with a dominant-negative RAR{alpha} mutant (RAR{alpha}403; ref. 21) in HME1 cells confirmed that indeed this was the case. Both the RAR{alpha} knockdown clone Si-{alpha}-1, carrying the RAR{alpha}-B silencing sequence (Fig. 4A, top, left ; ref. 8), and the DN {alpha}4 clone, carrying the dominant-negative RAR{alpha}403 mutant (Fig. 4A, top, right), displayed stable repression of endogenous RARß2 and CRBP1 transcription relative to the respective control clones, mock-1, carrying a scrambled sequence, and LX C3, carrying an empty vector (Fig. 4A, middle and bottom). The chromatin of both genes underwent histone H4 hypoacetylation. We conclude that this is the case because RA did not raise significantly histone H4 acetylation (Fig. 4B, left and right). In addition, we observed de novo DNA methylation in the RME (8) and CRBP1 CpG island (Fig. 4C, top and bottom) in clones impaired for RAR{alpha}, but not in the mock controls.


Figure 4
View larger version (31K):
[in this window]
[in a new window]
  		Figure 4. Functional inhibition of RAR{alpha} induces concomitant CRBP1 and RARß2 epigenetic silencing. A, HME1 stable clones, such as the prototypical Si-{alpha}-1, carrying the RAR{alpha}-B RNAi sequence (top, left), or the prototypical DN {alpha}4, carrying the dominant-negative construct RAR{alpha}403 (top, right), showing significant (P < 0.01) down-regulation of both RARß2 (middle) and CRBP1 (bottom) transcription compared with the cognate control clones. B, Si-{alpha}-1 clone showing significant histone H4 hypoacetylation in response to RA (P < 0.05) at both RARß2-RARE and CRBP1-RARE. C, left, distribution of CpG methylated sites in 10 RARß2 (top) and 10 CRBP1 (bottom) nonpermissive hypermethylated alleles, respectively, out of the nonpermissive alleles of S1-{alpha}-1 shown in the pie graphs (right). D, evidence of RA resistance (top, left), anchorage-independent growth (top, right), and impaired epithelial polarization in three-dimensional cultures (bottom) in both Si-{alpha}-1 and DN {alpha}4 clones.

 
The phenotypes acquired by both Si-{alpha}-1 and DN {alpha}4 clones fully recapitulated the phenotypes observed after RARß2 and CRBP1 silencing, including resistance to RA, anchorage-independent growth, and aberrant epithelial polarization when cells were grown as acini in three-dimensional cultures (Fig. 4D). Interestingly, these phenotypes were described previously in a human mammary epithelial cell strain transfected with the RAR{alpha}403 mutant (26, 27).

We conclude that an impaired RA signal through RAR{alpha} can cause a coordinated, epigenetic repression of RARß2 and its direct target CRBP1.


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
In this study, we show that an interference with RA signal at RAR{alpha} in RA-sensitive, untransformed, human mammary epithelial cells converts in a coordinated succession two RA-responsive genes, RARß2 and its direct target CRBP1, into a status nonpermissive for transcription. Interestingly, a fraction of nonpermissive RARß2 and CRBP1 alleles acquires also chromatin repressive modifications, including de novo DNA methylation. This is the first evidence of a repressive epigenetic domino effect involving two RA-responsive genes consequent to an impaired RA signal. As a consequence of this epigenetic domino effect, human breast epithelial cells acquire in a succession RA resistance and distinct phenotypes of breast epithelial transformation, such as loss of epithelial polarization, inability to form a hollow lumen in three-dimensional basement membrane cultures, and anchorage-independent growth.

Recently, we showed that an impaired RA signal through RAR{alpha} in RA-sensitive, transformed breast cancer cells leads to RARß2 epigenetic silencing (8). Apparently, an interference with RA signal through RAR{alpha} would trigger the silencing of genes that are normally epigenetically regulated by RA. Further, transcriptional silencing is clearly marked by aberrant repressive chromatin changes not only in transformed but also in untransformed cells. We do not know yet by which mechanism the transcriptionally inactive status created by lack of integration of RA signal through RAR{alpha} can invoke repressive modifying enzymes capable of imposing both histone and DNA repressive changes at RAR-regulated genes. We are tempted to speculate that, as it happens during the heterochromatinization of the transcriptionally inactive chromosome X (28, 29), an RNA-mediated mechanism may play a role in the heterochromatinization of transcriptionally inactive autosomal gene regions. Apparently, only a fraction of nonpermissive RARß2 or CRBP1 alleles develops aberrant de novo DNA methylation, which would indicate that it is the nonpermissive status that recalls DNA methylation and not the other way around. As for chromosome X, DNA methylation would "lock in" the silent status of nonpermissive alleles.

Interestingly, here we show that RAR{alpha}, via RARß2, controls the transcription of CRBP1, a gene involved both in retinol transport and breast epithelial cell morphogenesis. We found recently that RAR{alpha}, via RARß2, controls the transcription of another RA-responsive gene, CYP26A1, an enzyme implicated in RA catabolism and neural morphogenesis (ref. 30 and references within). Thus, RA-RAR{alpha} signal seems to control, through a concerted epigenetic mechanism, at least two RA-responsive genes involved in retinol/RA metabolism as well as morphogenesis. The gradient of retinoids is critical for symmetry and morphogenesis during organismal development (31–33). This could also be true during postnatal development of the mammary gland. The results shown here lend support to the hypothesis that retinoid metabolism is intertwined with retinoid-dependent morphogenesis. Specifically, RA itself would epigenetically regulate through its own metabolism critical morphogenetic programs.

Our study also provides useful information for improving a breast cancer prediction test, thus far based on RARß2 hypermethylation analysis alone (34). In light of what we showed here, RARß2 hypermethylation cannot distinguish between cells heterozygous for RARß2 nonpermissive alleles, thus still sensitive to RA, from cells homozygous for RARß2 nonpermissive alleles, thus RA resistant. Only the latter would have tumor phenotypic features. Combining RARß2 and CRBP1 epigenetic analysis might provide a better test for predicting breast cancer susceptibility.


   Acknowledgments
 
Grant support: U.S. Army Award DAMD17-02-01-0432, the Roswell Park Alliance Foundation, and the Susan Komen Foundation (N. Sacchi) and the Graduate Program of Molecular Medicine, University of Milan (Milan, Italy), and the Predoctoral U.S. Army Award No. W81XWH0510222 (G. Bistulfi).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Pasha Apontes for the help with the editing of the article.

Received 3/21/06. Revised 8/14/06. Accepted 9/ 1/06.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Chambon P. A decade of molecular biology of retinoic acid receptors. FASEB J 1996;10:940–54.[Abstract]
  2. Montesano R, Soulie P. Retinoids induce lumen morphogenesis in mammary epithelial cells. J Cell Sci 2002;115:4419–31.[Abstract/Free Full Text]
  3. Wang YA, Shen K, Wang Y, Brooks SC. Retinoic acid signaling is required for proper morphogenesis of mammary gland. Dev Dyn 2005;234:892–9.[CrossRef][Medline]
  4. Dilworth FJ, Chambon P. Nuclear receptors coordinate the activities of chromatin remodeling complexes and coactivators to facilitate initiation of transcription. Oncogene 2001;20:3047–54.[CrossRef][Medline]
  5. Brand N, Petkovich M, Krust A, et al. Identification of a second human retinoic acid receptor. Nature 1988;332:850–3.[CrossRef][Medline]
  6. Sucov HM, Murakami KK, Evans RM. Characterization of an autoregulated response element in the mouse retinoic acid receptor type ß gene. Proc Natl Acad Sci U S A 1990;87:5392–6.[Abstract/Free Full Text]
  7. Husmann M, Lehmann J, Hoffmann B, Hermann T, Tzukerman M, Pfahl M. Antagonism between retinoic acid receptors. Mol Cell Biol 1991;11:4097–103.[Abstract/Free Full Text]
  8. Ren M, Pozzi S, Bistulfi G, Somenzi G, Rossetti S, Sacchi N. Impaired retinoic acid (RA) signal leads to RARß2 epigenetic silencing and RA resistance. Mol Cell Biol 2005;25:10591–603.[Abstract/Free Full Text]
  9. Sirchia SM, Ren M, Pili R, et al. Endogenous reactivation of the RARß2 tumor suppressor gene epigenetically silenced in breast cancer. Cancer Res 2002;62:2455–61.[Abstract/Free Full Text]
  10. Andreola F, Giandomenico V, Spero R, De Luca LM. Expression of a smaller lecithin:retinol acyl transferase transcript and reduced retinol esterification in MCF-7 cells. Biochem Biophys Res Commun 2000;279:920–4.[CrossRef][Medline]
  11. Guo X, Ruiz A, Rando RR, Bok D, Gudas LJ. Esterification of all-trans-retinol in normal human epithelial cell strains and carcinoma lines from oral cavity, skin and breast: reduced expression of lecithin:retinol acyltransferase in carcinoma lines. Carcinogenesis 2000;21:1925–33.[Abstract/Free Full Text]
  12. Mira-y-Lopez R, Zheng WL, Kuppumbatti YS, Rexer B, Jing Y, Ong DE. Retinol conversion to retinoic acid is impaired in breast cancer cell lines relative to normal cells. J Cell Physiol 2000;185:302–9.[CrossRef][Medline]
  13. Rexer BN, Zheng WL, Ong DE. Retinoic acid biosynthesis by normal human breast epithelium is via aldehyde dehydrogenase 6, absent in MCF-7 cells. Cancer Res 2001;61:7065–70.[Abstract/Free Full Text]
  14. Farias EF, Ong DE, Ghyselinck NB, Nakajo S, Kuppumbatti YS, Mira-y-Lopez R. Cellular retinol-binding protein I, a regulator of breast epithelial retinoic acid receptor activity, cell differentiation, and tumorigenicity. J Natl Cancer Inst 2005;97:21–9.[Abstract/Free Full Text]
  15. Smith WC, Nakshatri H, Leroy P, Rees J, Chambon P. A retinoic acid response element is present in the mouse cellular retinol binding protein I (mCRBPI) promoter. EMBO J 1991;10:2223–30.[Medline]
  16. Husmann M, Hoffmann B, Stump DG, Chytil F, Pfahl M. A retinoic acid response element from the rat CRBPI promoter is activated by an RAR/RXR heterodimer. Biochem Biophys Res Commun 1992;187:1558–64.[CrossRef][Medline]
  17. Fisher GJ, Reddy AP, Datta SC, et al. All-trans retinoic acid induces cellular retinol-binding protein in human skin in vivo. J Invest Dermatol 1995;105:80–6.[CrossRef][Medline]
  18. Balmer JE, Blomhoff R. Gene expression regulation by retinoic acid. J Lipid Res 2002;43:1773–808.[Abstract/Free Full Text]
  19. Blomhoff R, Blomhoff HK. Overview of retinoid metabolism and function. J Neurobiol 2006;66:606–30.[CrossRef][Medline]
  20. Debnath J, Muthuswamy SK, Brugge JS. Morphogenesis and oncogenesis of MCF-10A mammary epithelial acini grown in three-dimensional basement membrane cultures. Methods 2003;30:256–68.[CrossRef][Medline]
  21. Tsai S, Bartelmez S, Heyman R, Damm K, Evans R, Collins SJ. A mutated retinoic acid receptor-{alpha} exhibiting dominant-negative activity alters the lineage development of a multipotent hematopoietic cell line. Genes Dev 1992;6:2258–69.[Abstract/Free Full Text]
  22. Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB. Methylation-specific PCR: a novel PCR assay for methylation status of CpG islands. Proc Natl Acad Sci U S A 1996;93:9821–6.[Abstract/Free Full Text]
  23. Napoli JL, Boerman MH, Chai X, Zhai Y, Fiorella PD. Enzymes and binding proteins affecting retinoic acid concentrations. J Steroid Biochem Mol Biol 1995;53:497–502.[CrossRef][Medline]
  24. Napoli JL. Interactions of retinoid binding proteins and enzymes in retinoid metabolism. Biochim Biophys Acta 1999;1440:139–62.[Medline]
  25. Kuppumbatti YS, Rexer B, Nakajo S, Nakaya K, Mira-y-Lopez R. CRBP suppresses breast cancer cell survival and anchorage-independent growth. Oncogene 2001;20:7413–9.[CrossRef][Medline]
  26. Seewaldt VL, Caldwell LE, Johnson BS, Swisshelm K, Collins SJ, Tsai S. Inhibition of retinoic acid receptor function in normal human mammary epithelial cells results in increased cellular proliferation and inhibits the formation of a polarized epithelium in vitro. Exp Cell Res 1997;236:16–28.[CrossRef][Medline]
  27. Dietze EC, Caldwell LE, Marcom K, et al. Retinoids and retinoic acid receptors regulate growth arrest and apoptosis in human mammary epithelial cells and modulate expression of CBP/p300. Microsc Res Tech 2002;59:23–40.[CrossRef][Medline]
  28. Heard E, Clerc P, Avner P. X-chromosome inactivation in mammals. Annu Rev Genet 1997;31:571–610.[CrossRef][Medline]
  29. Heard E. Delving into the diversity of facultative heterochromatin: the epigenetics of the inactive X chromosome. Curr Opin Genet Dev 2005;15:482–9.[CrossRef][Medline]
  30. Pozzi S, Rossetti S, Bistulfi G, Sacchi N. RAR-mediated epigenetic control of the cytochrome P450 Cyp26a1 in embryocarcinoma cells. Oncogene 2006;25:1400–7.[CrossRef][Medline]
  31. Kawakami Y, Raya A, Raya RM, Rodriguez-Esteban C, Belmonte JC. Retinoic acid signalling links left-right asymmetric patterning and bilaterally symmetric somitogenesis in the zebrafish embryo. Nature 2005;435:165–71.[CrossRef][Medline]
  32. Vermot J, Gallego Llamas J, Fraulob V, Niederreither K, Chambon P, Dolle P. Retinoic acid controls the bilateral symmetry of somite formation in the mouse embryo. Science 2005;308:563–6.[Abstract/Free Full Text]
  33. Vermot J, Pourquie O. Retinoic acid coordinates somitogenesis and left-right patterning in vertebrate embryos. Nature 2005;435:215–20.[CrossRef][Medline]
  34. Bean GR, Scott V, Yee L, et al. Retinoic acid receptor-ß2 promoter methylation in random periareolar fine needle aspiration. Cancer Epidemiol Biomarkers Prev 2005;14:790–8.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
S. N. Vasilatos, G. Broadwater, W. T. Barry, J. C. Baker Jr., S. Lem, E. C. Dietze, G. R. Bean, A. D. Bryson, P. G. Pilie, V. Goldenberg, et al.
CpG Island Tumor Suppressor Promoter Methylation in Non-BRCA-Associated Early Mammary Carcinogenesis
Cancer Epidemiol. Biomarkers Prev., March 1, 2009; 18(3): 901 - 914.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bistulfi, G.
Right arrow Articles by Sacchi, N.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bistulfi, G.
Right arrow Articles by Sacchi, N.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online