Molecular Medicine

, Volume 13, Issue 1–2, pp 69–78 | Cite as

Interplay of Genes Regulated by Estrogen and Diindolylmethane in Breast Cancer Cell Lines

  • Laura Mulvey
  • Alamelu Chandrasekaran
  • Kai Liu
  • Sarah Lombardi
  • Xue-Ping Wang
  • Karen J. Auborn
  • Leslie GoodwinEmail author
Open Access
Research Article


Diindolylmethane (DIM), a biologically active congener of indole-3-carbinol (I3C) derived from cruciferous vegetables, is a promising agent for the prevention of estrogen-sensitive cancers. Both DIM and estrogen affect transcription of genes by binding receptors, such as aryl hydrocarbon receptor (AhR) or estrogen receptors (ER). Gene regulation by DIM and estradiol (E2) can be very complex. While DIM typically binds the AhR, this complex can directly associate with the ER, recruit co-activators that bind to estrogen-responsive promoters, and activate transcription. Alternately, DIM can bind the ER directly. In this study, we have analyzed gene expression using microarray profiling and quantitative real time-polymerase chain reaction in MCF7 breast cancer cells treated with E2 (1 nM) or DIM (25 µM) alone or in combination for 16 h. The interplay of E2 and DIM was reflected in the expression of a subset of genes (<90) in which the combination of E2 and DIM acted either additively or antagonistically to alter gene expression.


I3C, DIM, and Estrogen-Enhanced Cancers

Indole-3-carbinol (I3C) and its chief condensation product, diindolylmethane (DIM), are naturally occurring phytochemicals from cruciferous vegetables that stimulate a number of cellular responses that are proapoptotic, anti-proliferative, and anti-estrogenic, i.e., processes incompatible with tumor development (1, 2, 3). Conversely, estrogen-initiated activity can lead to increased replication (4, 5, 6, 7) and inhibition of apoptosis (8), processes that are amenable with the development of tumors. Many animal studies indicate that I3C has anti-tumor efficacy for breast, cervical, and endometrial cancers (9, 10, 11, 12), indicating that I3C holds promise for the prevention of estrogen-enhanced cancers. Both I3C and DIM abrogate estrogen-enhanced cell proliferation (1,2), and the amount of apoptosis depends on the relative amount of I3C versus stradiol (E2) (13). Studies indicate that I3C and DIM affect estrogen, for example, by inducing enzymes that modulate estrogen metabolism (14, 15, 16). I3C induces the expression of the tumor suppressor gene BRCA1 (17), which inhibits ER α-regulated gene expression (18). I3C and DIM bind to estrogen receptors (ER) (19) and can compete with estrogen in reporter gene assays to inhibit estrogen dependent gene expression (20).

Estrogen, I3C, DIM, and Gene Expression

Estrogen, I3C, and DIM regulate multiple genes, and this regulation is not always mutually exclusive. 17β-Estradiol (E2) regulates genes by binding to either ER-α or ER-β, forming a complex, which binds the estrogen-responsive elements (ERE) in the promoter of estrogen-dependent genes (21,22). I3C or DIM binding to the aryl hydrocarbon receptor (AhR) activates it, resulting in nuclear translocation and complex formation with the basic helix loop helix region of the aryl hydrocarbon nuclear translocator protein (ARNT). The AhR/ARNT complex serves as a transcriptional unit, binding to highly conserved enhancer sequences termed xenobiotic response elements in promoters of relevant genes (23). However, I3C and DIM can also bind the ER acting as agonist (18,19) as well as competing with E2 for this binding (24). DIM-dependent AhR/ARNT-complex associates with and co-activates unliganded ER, modulating the estrogen-driven transcriptional signature, and thus, may act to enhance estrogenic activity (25). In contrast, DIM or its precursor I3C, have been found to inhibit estrogen-induced genes (26), carcinogen-induced rat mammary tumor formation, as well as the growth of estrogen-dependent tumors in a mouse xenograft model (27,28,11).

In this paper, we report the interplay between DIM and E2 upon the gene expression profile of the E2-responsive breast cancer cell line MCF-7.

Materials and Methods


17β-Estradiol and dimethyl sulfoxide (DMSO) were purchased from Sigma Chemical (St. Louis, MO, USA). DIM was a gift from Dr. M. Zeligs, BioResponse, Boulder, CO, USA.

Cell Lines and Cultures

The breast cancer cell line MCF-7 was purchased from the American Type Culture Collection (Manassas, VA, USA). All cells were maintained as monolayer cultures at 37°C in 5% CO2 and were grown in Dulbecco’s modified Eagle’s medium (DMEM) that contained 4.5 g glucose and bicarbonate/L (GIBCO-BRL, Gaithersburg, MD, USA) supplemented with 110 mg of sodium pyruvate/L, 200 mmol glutamine/L, 100 mL of fetal bovine serum/L, and 100,000 U each of penicillin and streptomycin/L.


The experiment was performed on estrogen responsive breast cancer MCF7 cells treated with combinations of DIM and E2 for 16 h. Four different sets of culture conditions were used on 1 × 105 MCF7 cells, and the cultures were done in triplicate. Cells grown in culture supplemented with 1 nM E2, cells supplemented with 25 µM DIM, and cells supplemented with 1 nM E2 and 25 µM DIM combination. As E2 and DIM were solubilized with DMSO, it served as vehicle control.

Microarray profiling

Total RNA was prepared from MCF-7 cells treated in the 4 culture conditions using the Qiagen RNeasy Kit according to manufacturer’s instructions (Qiagen, Valencia, CA, USA). Total RNA quality was assessed using the Agilent 2100 Bio-analyzer (Agilent Technologies, Palo Alto, CA, USA) and the 260/280 ratio was determined using the BioSpec-1601 Spectrophotometer (Shimadzu, Columbia, MD, USA). cDNA and cRNA were processed following the Ambion message AMP cRNA kit (Ambion, Austin, TX, USA) employing a T-7oligo-dT primer according to the manufacturer’s instructions. Biotinylated 16-UTP (Roche-Boehringer-Mannheim, Palo Alto, USA) and 11-CTP (Perkin Elmer, Wellesley, MA, USA) were used during the in vitro transcription process, and the cRNA was fragmented and hybridized to Affymetrix Human Genome U133A chips (Affymetrix, Santa Clara, CA, USA), which contains over 20,000 known genes. The hybridized chips were stained with phycoerythrinstreptavidin and washed to remove nonspecific signal according to the Affymetrix protocol. The chips were scanned using a laser confocal scanner manufactured by Hewlett Packard. The expression levels were calculated by Affymetrix’s Microarray Suite 5.0, and the overall intensity values were all scaled to the same value, 1500.

Statistical Analysis

Microarray data analysis with Genespring (Agilent Technologies) was done by importing Affymetrix MAS5.0 data. Principle component analysis with the different treatment types indicated two predominant patterns showing additive up- and down-regulation of genes in cells treated with E2 + DIM (E + D) compared with E2 or DIM alone. Genes of these two expression patterns were filtered on flags and on confidence with a P value of less than 0.05. This list was further filtered with a fold change cut off of 1.5 for E + D group for both up- and down-regulation. This gave a final list of 32 additively up-regulated genes and 46 additively down-regulated genes for E + D group. Important genes from these two lists were confirmed by quantitative polymerase chain reaction (QPCR).

Pathway Analysis

Pathway analysis was conducted using Pathway Studio Central version 1.1 (Ariadne Genomics, MD, USA). We imported the Genespring list of up- or down-regulated genes to initiate database mining. The software retrieves the most relevant networks that are differentially perturbed in a disease or provides insights into the common regulatory mechanisms of the set of genes. The networks are graphically displayed and allow validation by referral to the original abstracts or articles the facts were drawn from. The database can be queried for all known interaction or pathways that involve a specific protein or target molecules as well as for common relationships among a group of proteins. The software uses ResNet, a comprehensive molecular network database compiled by MedScan containing more than 500,000 events of regulation, interaction, and modification among thousands of proteins, cell processes, and small molecules. This is displayed as a global network of molecular interactions with pathways being sub-networks. This allows the building of individual as well as interplays among pathways.

Quantitative real time (RT)-PCR

RT-PCR was performed using TaqMan chemistry. TaqMan primers and probes were designed using Primer Express software version 1.5 (Applied Biosystems, Foster City, CA, USA) and synthesized at Feinstein Institute core facility. The probes were labeled with FAM at the 5′ end and TAMRA at the 3′ end. The relative expression of the various genes was determined using the Eurogentec RTqPCR mastermix (Eurogentec, Belgium) and ABI PRISM 7700 Sequence Detection System. The PCR mix contained 1 X master mix and 0.125 µL of Euroscript + RT and Rnase Inhibitor (RT-0.125 U/µL and Rnase Inhibitor 0.05 U/µL). Optimal concentrations of primers, probes, and the RNA were standardized. The final concentrations of forward and reverse primers and probes for PGDH, LDLR, CXCR4, and IFIP were 500 nm and 200 nm, respectively. The concentration of B-actin primers and probe, served as the internal housekeeping gene control were 500 nm and 100 nm, respectively. Fifty nanograms of total RNA were used per 25 µL reaction with all samples run in duplicate. The thermal cycler conditions were 48°C for 30 min, 95°C for 10 min, and 45 cycles of 95°C for 0.15 min, and 60°C for 1 min. Data was analyzed using Sequence Detection System (SDS) software version 1.9.1. Results were obtained as Ct (threshold cycle) values. Ct is inversely proportional to the starting template copy number. Relative expression in all samples was calculated in comparison with untreated control samples using delta delta Ct method. Results were expressed as change with respect to the experimental control.

Primers and Probes










































H-SCD (solute carrier family 7)








P8 (P8 protein-candidate of metastasis)








TGFb1 (Camurati-Engelmann disease)








Microarray Analysis of Cell Lines

MAS5 analysis. We were interested in genes whose expression was modulated by treatment with E2 and DIM when compared with control treatment with DMSO. A list of genes was compiled, and these same genes were examined for additive increases or decreases in expression levels (fold changes) when E2 and DIM were used in combination. We found four different sets of gene lists: (a) genes changed by E2, but not by DIM or the combination; (b) genes changed by DIM but not by E2 or the combination; (c) genes changed by E2 or DIM and the expression of the genes were enhanced by the combination of both E2 and DIM; (d) genes modulated by E2 or DIM and the expression dampened by the combined effect. We have focused on genes where we found additive effects when E2 and DIM were present together in the same culture, as in sets three and four (data not shown).

We analyzed gene expression profiles from three replicates in which MCF-7 cells were treated for 16 h with concentrations of E2 (1 nM) or DIM (25 µM) or the combination of both E2 and DIM. We imported our data as a tab-delimited file into GeneSpring software, where we were able to filter for genes of interest, i.e., genes whose expression reflected interplay of E2 and DIM.

Genes Significantly Up- or Down-regulated by Treatment with E2 and DIM but Not E2 or DIM Alone

Table 1 is a list of genes whose expression was significantly increased in the E2 + DIM combination when compared with the DMSO control, but not significantly changed in E2 or DIM alone when compared with the DMSO control. The up-regulated genes were 32 in number, and the fold change ranged from 3.93 for NM_000104, Cytochrome P450, CYP1B1 to 1.5 for AF070587.1 deleted in liver cancer 1. The genes whose expression were significantly decreased by the combination of E2 + DIM but not by E2 or DIM alone are shown in Table 2. We identified 46 down-regulated genes whose fold changes ranged from 2.21 for NM_006156 NEL to about 2 to 1.5 for NM_015950 mitochondrial ribosomal protein L2.
Table 1

Fold change of genes whose expression is increased by the combination of E2 + DIM

Gene title


EvsC ± SE

DvsC ± SE


Cytochrome P450 CYP1B1


1.389 ± 0.12

3.101 ± 0.22

3.93 ± 0.25

Cytochrome P450, CYP1B1


1.436 ± 0.15

2.738 ± 0.26

3.32 ± 0.32

Cytochrome P450, CYP1B1


1.581 ± 0.13

2.617 ± 0.53

2.86 ± 0.66

Cytochrome P450 CYP1A1


0.773 ± 0.13

2.605 ± 0.61

2.84 ± 0.51

Stearoyl-CoA desaturase (delta-9-desaturase)


1.533 ± 0.14

1.96 ± 0.34

2.40 ± 0.28

Caldesmon 1


1.708 ± 0.32

1.75 ± 0.33

2.37 ± 0.31

Solute carrier family 7/ member 11


1.444 ± 0.17

1.566 ± 0.19

1.90 ± 0.13

p8 protein (candidate of metastasis 1)


1.118 ± 0.12

1.395 ± 0.45

1.92 ± 0.27

Interferon-induced protein/ tetratricopeptide repeats 1


0.93 ± 0.17

1.438 ± 0.13

1.88 ± 0.20

Nucleophosmin/nucleoplasmin, 3


1.72 ± 0.22

1.775 ± 0.16

1.85 ± 0.16

Amyloid beta (A4) precursor protein (protease nexin-II, Alzheimer disease)


1.717 ± 0.39

1.624 ± 0.31

1.84 ± 0.32

Zinc finger protein 557


1.324 ± 0.12

1.559 ± 0.21

1.83 ± 0.20

Like Bifunctional methylenetetrahydrofolate dehydrogenase/cyclohydrolase, mitochondrial precursor


1.263 ± 0.13

1.055 ± 0.20

1.81 ± 0.23

Asparagine synthetase


1.267 ± 0.16

1.303 ± 0.32

1.81 ± 0.25

GTP bp overexpressed in skeletal muscle


1.666 ± 0.12

1.133 ±0.19

1.78 ± 0.21

TBC1 domain family, member 5


1.486 ± 0.11

1.395 ± 0.16

1.76 ± 0.12

Fatty acid binding protein 5 (psoriasis-assoc)


1.564 ± 0.12

1.33 ± 0.12

1.75 ± 0.12

Clathrin, heavy polypeptide (Hc)


1.406 ± 0.24

1.43 ± 0.29

1.75 ± 0.85

Solute carrier family 33 (acetyl-CoA transporter)1


1.338 ± 0.15

1.348 ± 0.13

1.69 ± 0.21

Hypothetical protein FLJ21148


1.469 ± 0.17

1.424 ± 0.17

1.68 ± 0.16

Thyroid receptor interacting protein 15


1.088 ± 0.15

1.226 ± 0.13

1.68 ± 0.20

FK506 binding protein 14, 22 kDa


1.165 ± 0.15

1.247 ± 0.21

1.65 ± 0.16

Stromal antigen 2


1.359 ± 0.43

1.553 ± 0.81

1.65 ± 1.29

Kelch-like 9 (Drosophila)


1.495 ± 0.17

1.38 ± 0.13

1.64 ± 0.26

Phosphoglycerate dehydrogenase


1.167 ±0.10

1.399 ± 0.17

1.64 ± 0.13

Diacylglycerol kinase, epsilon 64kDa


1.119 ± 0.25

1.53 ± 0.16

1.62 ± 0.24

KIAA0746 protein


1.227 ± 0.10

1.361 ± 0.19

1.62 ± 0.16

Aldo-keto reductase family 1, member C3


0.901 ± 0.13

1.309 ± 0.18

1.62 ± 0.13

Insulin induced gene 1


1.395 ± 0.11

1.41 ± 0.27

1.56 ± 0.14

Esterase D/formylglutathione hydrolase


1.306 ± 0.15

1.078 ± 0.16

1.55 ± 0.15

Transducin (beta)-like 1X-linked receptor 1


1.42 ±0.15

1.341 ± 0.13

1.53 ± 0.14

Ataxin 1


1.361 ± 0.36

1.361 ± 0.35

1.53 ± 0.24

Similar to 0610010D24Rik protein


1.121 ± 0.12

1.337 ± 0.18

1.52 ± 0.12

Deleted in liver cancer 1


1.361 ± 0.15

1.108 ± 0.12

1.50 ± 0.14

Table 2

Fold change of genes whose expression is diminished by the combination of E2 + DIM

Gene Title


EvsC + SE

DvsC + SE

E+D + SE

Mitochondrial ribosomal protein L2


0.758 ± 0.09

1.035 ± 0.22

0.666 ± 0.10

Tumor protein p53 (Li-Fraumeni disease)


0.873 ± 0.11

0.889 ± 0.15

0.663 ± 0.10

HERV-H LTR-associating 3


0.942 ± 0.12

0.929 ± 0.16

0.662 ± 0.14

Lipin 2


0.983 ± 0.09

0.82 ± 0.16

0.659 ± 0.10

ATP-binding cassette, D (ALD), 1


0.894 ± 0.09

0.911 ± 0.13

0.653 ± 0.10

Cytoplasmic FMR1 interacting protein 2


0.903 ± 0.09

0.844 ± 0.17

0.653 ± 0.11

Endothelin converting enz 1


0.858 ± 0.11

0.723 ± 0.19

0.648 ± 0.11

Aldehyde dehydrogenase 3 B1


0.84 ± 0.09

0.718 ± 0.15

0.648 ± 0.12

ATP-binding cassette,G 1


0.754 ± 0.12

0.809 ± 0.14

0.648 ± 0.11

Kelch domain containing 3


0.949 ± 0.16

1.11 ± 0.14

0.639 ± 0.11

Erythropoietin receptor


0.782 ± 0.09

0.788 ± 0.15

0.636 ± 0.12

Tripartite motif-containing 3


0.764 ± 0.10

0.743 ± 0.15

0.633 ± 0.12

GPI-anchored metastasis-associated proteinhomolog


0.693 ± 0.09

0.88 ± 0.13

0.631 ± 0.10

DEAD box polypeptide 54


1.185 ± 0.17

0.863 ± 0.25

0.619 ± 0.18

Distal-less homeobox 4


0.777 ± 0.11

0.944 ± 0.15

0.618 ± 0.12

Slingshot homolog 3 (Drosophila)


0.861 ± 0.11

0.936 ± 0.13

0.617 ± 0.14

Synaptobrevin 2


0.675 ± 0.10

1.09 ± 0.26

0.617 ± 0.12

Ribosomal protein S6 kinase, 90kDa, polypeptide 4


0.813 ± 0.12

0.912 ± 0.19

0.615 ± 0.10

Chromodomain helicase DNA binding protein 3


0.751 ± 0.15

0.848 ± 0.12

0.615 ± 0.10

TGFB1-induced anti-apoptotic factor1


0.858 ± 0.09

0.915 ± 0.17

0.612 ± 0.12



0.849 ± 0.10

1.187 ± 0.18

0.607 ± 0.12

Kynurenine 3-monooxygenase


0.744 ± 0.11

0.697 ± 0.13

0.601 ± 0.11

Chr 20 ORF 149


0.796 ± 0.14

1.169 ± 0.31

0.596 ± 0.11

SH3 domain binding glutamic-acid-rich protein like


0.787 ± 0.13

0.748 ± 0.14

0.594 ± 0.11



0.697 ± 0.10

0.655 ± 0.12

0.594 ± 0.10

ATPase, Ca++ transporting, ubiquitous


0.816 ± 0.12

0.777 ± 0.14

0.591 ± 0.12

KIAA0912 protein


0.843 ± 0.13

0.837 ± 0.17

0.588 ± 0.14

E74-like factor 5 (ets like)


0.77 ± 0.11

0.748 ± 0.12

0.586 ± 0.11

Tyr 3-/Trp 5-monooxygenase activation protein,?


0.864 ± 0.11

0.965 ± 0.20

0.585 ± 0.16

TGFbeta 1


0.725 ± 0.10

0.863 ± 0.16

0.583 ± 0.10

Homeo box A6


0.615 ± 0.11

0.712 ± 0.16

0.575 ± 0.13



0.86 ± 0.11

0.87 ± 0.13

0.572 ± 0.12

Chondroitin polymerizing factor


0.837 ± 0.12

1.068 ± 0.15

0.569 ± 0.10

B-cell CLL/lymphoma 6


0.707 ± 0.09

0.739 ± 0.14

0.568 ± 0.10

CCR4-NOT transcription complex, 3


0.803 ± 0.09

1.044 ± 0.22

0.562 ± 0.12

Collagen, type VII, alpha 1


0.786 ± 0.09

0.912 ± 0.13

0.552 ± 0.10

Aldehyde oxidase 1


0.631 ± 0.15

0.664 ± 0.28

0.552 ± 0.12

Solute carrier family 4


1.021 ± 0.10

0.835 ± 0.16

0.551 ± 0.15

Oxidised low density lipoprotein R1


0.614 ± 0.12

0.76 ± 0.16

0.546 ± 0.15

B-cell CLL/lymphoma 3


0.754 ± 0.17

0.853 ± 0.16

0.532 ± 0.12

TCR gamma variable 9


0.703 ± 0.11

0.569 ± 0.13

0.525 ± 0.11

Actin, gamma 2, smooth muscle


0.946 ± 0.10

0.619 ± 0.13

0.503 ± 0.10

Neurofibromin 1


0.655 ± 0.16

0.576 ± 0.14

0.497 ± 0.17

Keratin 15


0.657 ± 0.10

0.642 ± 0.12

0.486 ± 0.17

SH3-domain GRB2-like endophilin B2


1.109 ± 0.18

0.917 ± 0.16

0.482 ± 0.11

NEL-like 2 (chicken)


0.814 ± 0.09

0.723 ± 0.13

0.451 ± 0.10

Increased Gene Expression by Combination Treatment with E2 + DIM or Inhibition by E2 + DIM Confirmation by Quantitative Real Time-Reverse Transcriptase PCR (Q-RT-PCR)

We selected genes of interest from the microarray data that were highly expressed or suppressed in the E2 + DIM combination and showed a minor or no change in the E2 or DIM only samples. We have confirmed these microarray results by Q-RT-PCR analysis with an aliquot of the RNA extracted from the MCF7 cells treated under four experimental conditions (DMSO [control], E2 [1 µM] alone, DIM [25 µM] alone, or the combination of E2 and DIM) and used in the processing for the microarray analysis. We selected six genes that showed increased expression with the combined treatment of E2 and DIM: CYP1A1, CYP1B1, PGDH, AS, P8, Caldesmon, and SCD. We also looked at three genes that demonstrated decreased expression when treated with the combination of E2 and DIM, CXCR4, TGFb-1, and BCLL-6. All genes chosen from the microarray experiment and subsequently confirmed by QPCR were analyzed for correlation between the two gene expression technologies. The correlation coefficient is shown in each figure for comparison. The results of this analysis are shown in Figure 1.
Figure 1

Validation of microarray data by quantitative real time PCR with the corresponding correlation coefficients. The figure is comprised of two sets of data, up-regulated genes, indicated by the heading showing seven genes with increased fold changes for cells treated with E2 + DIM, and down-regulated genes heading three genes whose expression is decreased with treatment of E2 + DIM. The vertical bars indicate relative mRNA expression in E2, DIM, and E2 + DIM. The expression levels were normalized against DMSO treated MCF-7 cells. Microarray fold change is shown as a line along the bars. The one-step QRT-PCR was performed by TaqMan chemistry. B-actin was used as endogenous control.

Pathway analysis

We imported our Genespring lists of additively up-regulated genes in cells treated with a combination of DIM and E2 into PathwayStudio. We sought to analyze the relationships and biological interactions among the up-regulated genes identified by combination treatment. Figure two displays the pathways retrieved from the PathwayStudio software when we put in three of our up-regulated genes.
Figure 2

Pathway analysis was performed using PathwayStudio software by importing three of the 32-up-regulated genes in our list. The key to color identification of network components is as follows: yellow, cell processes; orange, functional class; red, protein; red with blue border, genes up-regulated; red connection lines, negative regulation; green connection lines, positive regulation; grey lines, unknown regulation type.

In a like manner, we also analyzed the common regulators and/or biological interactions among our incrementally suppressed or down-regulated genes by querying the relationships between several of the genes showing the most significantly diminished gene expression, data not shown.


These results indicate that there are a subset of genes in MCF7 whose expression can be modulated by the combination of E2 and DIM. The expression of 32 genes was significantly up-regulated or enhanced by combination treatment with E2 and DIM. Expression of 46 genes was down-regulated by the combination of E2 and DIM in the MCF7 cultures. The subset of genes analyzed by QRT-PCR confirmed the results of microarray profiling. Clearly, the expression of many genes is interdependent on both E2 and DIM in cells that respond to estrogen.

In general, the effect of the combination treatment of E2 and DIM on MCF7 cells was antagonistic, reflected in the greater number of down-regulated genes observed in our microarray analysis. This finding is consistent with previous reports that showed interactions that were primarily inhibitory in studies both in vitro and in vivo (27,28). Many of the changes are subtle, but show an effect over a wide variety of affected genes, including transcription factors as well as numerous metabolic gene products whose expression is diminished in comparison to the findings of treatment with E2, or for that matter DIM alone. Where the changes in gene expression are modest, it is important to note that, while intriguing, these gene modulations are only possibilities and will need additional studies to determine whether or not they significantly impact the behavior of the cell culture.

Previous studies by Chen et al. (26) using subtractive hybridization indicated that DIM could decrease the expression of a number of genes up-regulated by E2 and hypothesized cross-talk between AhR and ER. Our studies support these findings, but additionally showed that expression of a number of genes up-regulated in the presence of 25 µM DIM could be modulated in the presence of 1 nM E2, with expression either dampened by E2 or enhanced by E2. Alternately, some genes whose expression was up-regulated by E2 could be further enhanced by DIM. Finally, the expression of some genes was only detectable when the cells were exposed to the combination of E2 and DIM. These observations dramatically increase the complexity of the interplay by E2 and DIM and how they might affect the microenvironment of a cell. The study raises questions as to imbalances of the relative amount of these compounds. It is clear that the risk of breast cancer is inversely proportional to the amount of cruciferous vegetables in diet (the natural source of DIM) (29), and that more estrogen increases risk of breast cancer (30). Relative amounts of these compounds in combination may prove crucial to the protective or preventative benefits of DIM (or cruciferous vegetables).

As one example, DIM induces the expression of CYP1A1 and CYP1B1 members of the P450 superfamily. This study indicated that a further enhancement of expression of these genes occurred in the presence of E2. CYP1A1, a phase I enzyme, can be the first step in the detoxification of number of carcinogens. Conversely, it is known to convert many procarcinogens to carcinogens. Importantly, CYP1A1 increases 2-hydroxylation of estrone leading to 2-hydroxyestrone (not estrogenic), and which is rapidly O-methylated into compounds that are anti-proliferative, pro-apototic, and anti-angiogenic. However, induction of CYP-1B1 shifts metabolism toward 4-hydroxyestrone, which can be carcinogenic. An imbalance of estrogen metabolism is indicated in breast (31), cervical (32), and endometrial (unpublished results) cancers, which are all estrogen-enhanced cancers where I3C and DIM appear to be preventative. Systemic lupus erythematosis and rheumatoid arthritis, diseases predominantly affecting women, also have abnormal estrogen metabolism (33). An animal study indicated that a diet rich in I3C ameliorated the lupus disease and changed estrogen metabolism (34).

An interesting aspect of altered estrogen metabolism is found in the up-regulation of aldo-keto reductase family 1, member 3 (AKR1C3), an isomer of the AKR superfamily that is found most prominently in prostate and mammary glands, in samples treated with the combination of E2 and DIM. This enzyme has the ability to interconvert testosterone with 4-androstene-3–17-dione but inactivate 5a-DHT and, therefore, eliminate active androgens from the prostate. In the mammary gland, AKR1C3 may function predominantly as a reductase to produce testosterone from 4-androstene-3–17-dione in an intracrine manner and to reduce estrone to estradiol (35). Recently, it has been shown that levels of these steroid-metabolizing genes are diagnostic of tumor versus normal breast tissue (36). Levels of AKR1C3 are reduced in tumor versus normal tissue, which reflect our findings in E2 treatment alone compared with treatment with both E2 and DIM. The induction of this gene, which has been correlated well with its enzyme activity, suggests a rebalancing to normal steroid-metabolism homeostasis.

One can predict numerous effects of how genes affected by DIM and E2 change the microenvironment of a cell. For instance, the DEAD box polypeptide 4 is a member of a family of genes characterized by the conserved motif Asp-Glu-Ala-Asp (DEAD). These proteins are putative RNA helicases that mediate nucleoside triphosphate-dependent unwinding of double-stranded RNA. They are thought to be involved in a variety of cellular processes that involve modification of RNA secondary structure, i.e., translation initiation, nuclear and mitochondrial splicing, ribosome and spliceosome assembly. Members of this family are alleged to be involved in embryogenesis, spermatogenesis, and cellular growth and division.

It is of interest that the combination of DIM and E2 leads to a decrease in expression of genes such as Bcl-6, a zinc finger nuclear phosphoprotein, normally expressed in the germinal center B cells and some intrafollicular T cells. This gene codes for a DNA-binding transcriptional repressor that exerts an important role in the development of normal germinal centers. Its constitutive expression has been associated with suppression of p53 expression as well as phenotypic changes in germinal center cells by affecting differentiation and/or apoptosis. Mature germinal center B cells that leave the germinal center environment generally down-regulate Bcl-6 expression. A block in normal down-regulation of Bcl-6 has been postulated to cause genetic instability in the germinal center cells, and subsequently leads to neoplastic transformation (37). Additionally the down regulation of Bcl3 in cells treated in combination with E2 and DIM disrupts a signaling pattern observed in MCF7 cell (38) where Bcl3 complexes with phosphorylated Bcl10 and translocates to the nucleus, where it alters transcription.

The mixture of E2 and DIM treatment dampens the expression of cytoplasmic FMR1 interacting protein 2, which associates with FMRP (Fragile X mental retardation protein) as well as FMRP-related proteins FXRIP and FXR2P. The protein is cytoplasmically colocalized with FMRP and ribosomes, and is thought to interact with RAC1. RAC1 is a small GTPase that stimulates actin polymerization toward lamellopodia formation, whose overexpression in tumor cells has been associated with invasion and metastasis in human tumor cells (39,40).

Intriguingly, CXCR4 is also down-regulated by the combination treatment with E2 and DIM. CXCR4 appears to be necessary for breast cancer metastasis (41). This interesting gene was significantly suppressed in one gene list (data not shown), but not in the other, yet bears mention. CXCR4, a cytokine, is the cognate receptor for stromal cell derived factor 1, and the expression of this complex in breast cancer cells is associated with significant increases in invasiveness and faster migration of these cells to the lymph nodes (42). Silencing CXCR4 gene expression with siRNA blocks in vitro invasion and in vivo metastasis of breast cancer cells in animal models (41,43). Also of note is the down-regulation of CCR4-NOT complex 3 in the presence of both E2 and DIM, suggesting a suppressive effect on global transcription through regulation of transcription factor TFIID. As a master switch, dampening the expression of this factor would result in myriad gene effects both positive and negative (44).

Many of the noted gene changes and effects are subtle in form, with regulation modulated rather than radically altered. These findings suggest interesting perturbations in a biological system in the presence of physiologic concentrations of hormone and low concentrations of bioactive chemicals present in the environment. As much of the effects of DIM and E2 shift metabolism toward a proapoptotic, moderated, proliferative state, our results are consistent with DIM offering an effective preventive adjuvant in a healthy nutritional regime.

We have examined the overlap between lists of genes identified by each of the analysis programs. Using MAS5 and dChip (45,46), we identified 17 genes whose expression was increased or up-regulated when the cells were treated with a combination of E2 + DIM. This is in comparison to 32 genes that were identified using our filtering criterion with GeneSpring. Between these two groups, there are six genes that are found on both lists. The six overlapping genes are among the largest fold changes on the dChip list while they are found throughout the list with GeneSpring. However, the list does completely overlap with the clustering data identifying enhanced gene expression when treated with the combination of E2 + DIM (data not shown). In a similar manner, we identified 14 genes whose expression was suppressed or inhibited by treatment with E2 + DIM using dChip and the same filtering criterion as for gene induction, contrasted with 46 genes identified with GeneSpring. Between these two groups were six genes in common.

The pathway analysis is consistent with our in vitro observations as well as our expression analysis of treatment of MCF7 cells with E2 or DIM or both. Among the up-regulated gene pathway we found interrelated cell processes, which contain other genes found in our up-regulated list but not included in our query, such as nuclear receptors, cyclin dependent kinase inhibitor, and dicer (a ribonuclease essential for RNA interference of small temporal RNA pathways which represses gene expression). These genes are in keeping with a program focused on increasing cell differentiation, through cytoskeletal development and assembly, detection of DNA damage and cell cycle arrest, as well as apoptosis, post translation, and gene silencing. These results suggest there may be common regulators or signal transduction pathways among these genes.

When we examined cell systems affected by the expression of BCL-6 and TGF-b, the expression of both is down-regulated in treatment of cells with E2 and DIM, we find a multitude of processes that are common targets of these molecules but not necessarily overlapping in their modulation. In fact, there are almost an equal number of processes, which are coordinately regulated by these genes as those that receive opposite and antagonistic signals from these powerful molecules. The net result may be due to the relative concentrations of each protein.

We noted common regulators between TGF-b and BCL6. These include MAPK1 and 3, ABL-1, FOXO 3a (belonging to the forkhead family of transcription factors, which may function as a trigger for apoptosis), tumor suppressor gene EP300, a cyclin D-related transcription factor, as well as cyclin D itself. In many ways, the paradoxical effects of the regulated genes we found with the E2 + DIM-treated cells mirror the complicated physiologic findings of estrogens and exposure to toxins such as TCDD and dioxin.

Joint effects of DIM and E2 are complicated. One possibility is coordinate gene regulation due to the binding of DIM to its cognate AhR receptor as well as unliganded ER complex. The binding of AhR/ARNT complex to its response elements functions as a cis-acting enhancer in the regulatory domains of its target genes, a representative group identified as the AhR gene battery (47). The presence of proximal regulatory elements to AhR binding sites within discrete chromosomal locals suggests a mechanism by which this complex can activate a number of other transcription factors (not AhR targets) and clusters of genes as a secondary effect of binding (48). The juxtaposition of ligands dependent AhR-ER binding may have a concerted effect on the activation of additional transcription factors with both induction as well as repression of gene expression. The concentration of receptors as well as the concentration of DIM and E2 would matter and determine how much E2 or DIM bind to their cognate receptors (DIM binding to AhR and E2 binding to ERs) or compete for binding to the opposing receptors.

This work focused on the combined activity of E2 and DIM on gene expression, i.e., how the combination would be different than individual effects. It is well established that most activities of I3C and DIM are not related to estrogen per se. Nonetheless, this study confirms that there is a strong interplay of estrogen and DIM, which is reflected in gene expression changes when both compounds are present.

Note added in proof: A recent paper by Liu et al. (49) showed AhR agonists directly activate ERa in MCF-7 breast cancer cells.



Laura Mulvey and Alamelu Chandrasekaran contributed equally to this work. This work was supported by RO1-CA733850 to KJA from the National Institutes of Health.


  1. 1.
    Ge X, Fares FA, Yannai S. 1999. Induction of apoptosis in MCF-7 cells by indole-3-carbinol is independent of p53 and bax. Anticancer Res. 19:3199–3203.PubMedGoogle Scholar
  2. 2.
    Chen DZ, Qi M, Auborn KJ, Carter TH. 2001. Indole-3-carbinol and diindolylmethane induce apoptosis of human cervical cells and in murine HPV-16 transgenic preneoplastic cervical epithelium. J. Nutr. 131:3294–3302.CrossRefGoogle Scholar
  3. 3.
    Bjeldanes LF, Kim JY, Grose KR, Bartholomew JC, Bradfield CA. 1991. Aromatic hydrocarbon responsiveness-receptor agonists generated from indole-3-carbinol in vitro and in vivo: comparisons with 2,3,7,8-tetrachlorodibenzo-p-dioxin. Proc. Natl. Acad. Sci. U.S.A. 88:9543–9547.CrossRefGoogle Scholar
  4. 4.
    Edwards DP, Adams DJ, McGuire WI. 1981. Estrogen regulation of growth and specific protein synthesis in human breast cancer cells in tissue culture. Adv. Exp. Biol. 138:133–149.CrossRefGoogle Scholar
  5. 5.
    Newfield L, Bradlow HL, Sepkovic DW, Auborn K. 1998. Estrogen metabolism and the malignant potential of human papillomavirus immortalized keratinocytes. Exp. Biol. Med. 217:322–326.CrossRefGoogle Scholar
  6. 6.
    Prall OWJ, Sarcevic B, Musgrove EA, Watts CK, Sutherland RL. 1997. Estrogen-induced activation of Cdk4 and Cdk2 during G1-S phase progression is accompanied by increased cyclin D1 expression and decreased cyclin-dependent kinase inhibitor association with cyclin E-Cdk2. J. Biol. Chem. 272:10882–10894.CrossRefGoogle Scholar
  7. 7.
    Lobenhofer EK, Lee Bennet P, Cable L, Li L, Bushel PR, Afshara CA. 2002. Regulation of DNA replication fork genes by 17b-estradiol. Mol. Endocrinol. 16:1215–1229.PubMedGoogle Scholar
  8. 8.
    Perillo B, Sasso A, Abbondanza C, Palumbo G. 2000. 17β-estradiol inhibits apoptosis in MCF-7 cells, inducing bcl-2 expression via 2 estrogen-responsive elements present in the coding sequence. Mol. Cell. Biol. 8:2890–2901.CrossRefGoogle Scholar
  9. 9.
    Jin L et al. 1999. Indole-3-carbinol prevents cervical cancer in human papilloma virus type 16 (HPV16) transgenic mice. Cancer Res. 59:3991–3997.PubMedGoogle Scholar
  10. 10.
    Bell MC et al. 2000. Placebo-controlled trial of indole-3-carbinol in the treatment of CIN. Gynecol. Oncol. 78:123–129.CrossRefGoogle Scholar
  11. 11.
    Bradlow HL, Michnovicz JJ, Telang NT, Osborne MP. 1991. Effects of dietary indole-3-carbinol on estradiol metabolism and spontaneous mammary tumors in mice. Carcinogenesis 12:1571–1574.CrossRefGoogle Scholar
  12. 12.
    Kojima T, Tanaka T, Mori H. 1994. Chemoprevention of spontaneous endometrial cancer in female donyru rats by indole-3-carbinol. Cancer Res. 54:1446–1449.PubMedGoogle Scholar
  13. 13.
    Chen D-Z, Carter TH, Auborn KJ. 2004. Apoptosis in cervical cancer cells: Implications for adjunct anti-estrogen therapy for cervical cancer. Anticancer Res. 24:2649–2656.PubMedGoogle Scholar
  14. 14.
    Chen I, Safe S, Bjeldanes L. 1996. Indole-3-carbinol and diindolylmethane as aryl hydrocarbon (Ah) receptor agonists and antagonists in T47D human breast cancer cells. Biochem. Pharmacol. 51:1069–1076.CrossRefGoogle Scholar
  15. 15.
    Bradlow HL, Telang NT, Sepkovik DW, Osborn MP. 1996. 2-Hydroxyestrone: the “good” estrogen. J. Endrocrinol. 150:S259–265.Google Scholar
  16. 16.
    LaVallee TM et al. 2002. 2-Methoxyestradiol inhibits proliferation and induces apoptosis independently of estrogen receptors a and beta. Cancer Res. 62:3691–3697.PubMedGoogle Scholar
  17. 17.
    Meng Q et al. 2000. Suppression of breast cancer invasion and migration by indole-3-carbinol: associated with up-regulation of BRCA1 and E-cadherin/catenin complexes. J. Mol. Med. 78:155–165.CrossRefGoogle Scholar
  18. 18.
    Fan S et al. 1999. BRCA1 inhibition of estrogen receptor signaling in transfected cells. Science 284:1354–1356.CrossRefGoogle Scholar
  19. 19.
    Liu H, Wormke M, Safe SH, Bjeldanes LF. 1994. Indolo[3,2-b]carbazole: a dietary-derived factor that exhibits both antiestrogenic and estrogenic activity. J. Natl. Cancer Inst. 86:1758–1765.CrossRefGoogle Scholar
  20. 20.
    Auborn KJ et al. 2003. Indole-3-carbinol is a negative regulator of estrogen. J. Nutr. 133:2470s–2475s.CrossRefGoogle Scholar
  21. 21.
    Evans RM. 1998. The steroid and thyroid hormone receptor superfamily. Science 240:889–895.CrossRefGoogle Scholar
  22. 22.
    Katzenellenobogen BS. 1996. Estrogen receptors: bioactivities and interactions with cell signaling pathways. Biol. Reprod. 54:287–293.CrossRefGoogle Scholar
  23. 23.
    Safe S, Krishnan V. 1995. Cellular and molecular biology of aryl hydrocarbon (Ah) receptor-mediated gene expression. Arch. Toxicol. Suppl. 17:99–115.CrossRefGoogle Scholar
  24. 24.
    Riby J, Chang G, Firestone G, Bjeldanes L. 2000. Ligand-independent activation of estrogen receptor function by 3,3′ diindolylmethane in human breast cancer cells. Biochem. Pharmacol. 60:167–177.CrossRefGoogle Scholar
  25. 25.
    Ohtake F et al. 2003. Modulation of oestrogen receptor signaling by association with the activated dioxin receptor. Nature 423:545–550.CrossRefGoogle Scholar
  26. 26.
    Chen I, Hsieh T, Thomas T, Safe S. 2001. Identification of estrogen-induced genes downregulated by AhR agonists in MCF-7 breast cancer cells using suppression subtractive hybridization. Gene 262:207–214.CrossRefGoogle Scholar
  27. 27.
    Chen I, McDougal A, Wang F, Safe S. 1998. Aryl hydrocarbon receptor-mediated antiestrogenic and antitumorigenic activity of diindolylmethane. Carcinogenisis 19:1631–1639.CrossRefGoogle Scholar
  28. 28.
    Chang X et al. 2005. 3,3′-Diindolylmethane inhibits angiogenesis and the growth of transplantable human breast carcinoma in athymic mice. Carcinogenesis 26:771–778.CrossRefGoogle Scholar
  29. 29.
    Terry P, Wolk A, Persson I, Magnusson C. 2001. Brassica vegetables and breast cancer risk. J. Am. Med. Assoc. 285:2975–2977.CrossRefGoogle Scholar
  30. 30.
    Russo J, Hasan Lareef M, Balogh G, Russo IH. 2003. Estrogen and its metabolites are carcinogenic agents in human breast epithelial cells. Steroid Biochem. Mol. Biol. 87:1–25.CrossRefGoogle Scholar
  31. 31.
    Rogan EG et al. 2003. Relative imbalances in estrogen metabolism and conjugation in breast tissue of women with carcinoma: potential bio-markers of susceptibility to cancer. Carcinogenesis 24:697–702.CrossRefGoogle Scholar
  32. 32.
    Jin L et al. 1999. Indole-3-carbinol prevents cervical cancer in human papilloma virus type 16 (HPV16) transgenic mice. Cancer Res. 59:3991–3997.PubMedGoogle Scholar
  33. 33.
    Lahita RG, Bradlow HL, Kunkel HG, Fishman J. 1979. Alterations of estrogen metabolism in systemic lupus erythematosus. Arthritis Rheum. 22:1195–1198.CrossRefGoogle Scholar
  34. 34.
    Auborn KJ et al. 2003. Lifespan is prolonged in autoimmune-prone (NZB/NZW) F1 mice fed a diet supplemented with indole-3-carbinol. J. Nutr. 133:3610–3613.CrossRefGoogle Scholar
  35. 35.
    Penning TM et al. 2000. Human 3 α-hydroxysteroid dehydrogenase isoforms (AKR1C1-AKR1C4) of the aldo-keto reductase superfamily: functional plasticity and tissue distribution reveals roles in the inactivation and formation of male and female sex hormones. Biochem J. 351:67–77.PubMedPubMedCentralGoogle Scholar
  36. 36.
    Lewis MJ, Wiebe JP, Heathcote JG. 2004. Expression of progesterone metabolizing enzyme genes (AKR1C1, AKR1C2, AKR1C3, SRD5A1, SRD5A2) is altered in human breast carcinoma. BMC Cancer 4:27–38.CrossRefGoogle Scholar
  37. 37.
    Phan RT, Dalla-Favera R. 2004. The BCL6 proto-oncogene suppresses p53 expression in germinal-center B cells. Nature 432:635–639.CrossRefGoogle Scholar
  38. 38.
    Yeh PY et al 2006. A pathway for tumor necrosis factor-a-induced Bcl10 nuclear translocation. J. Biol. Chem. 281:167–175.CrossRefGoogle Scholar
  39. 39.
    Bouzahzah B et al. 2001. Rho family GTPases regulate mammary epithelium cell growth and metastasis through distinguishable pathways. Mol. Med. 7:816–830.CrossRefGoogle Scholar
  40. 40.
    Baugher PJ, Krishnamoorthy L, Price JE, Dharmawardhane SF. 2005. Rac1 and Rac3 isoform activation is involved in the invasive and metastatic phenotype of human breast cancer cells. Breast Cancer Res. 7:R965–R974.CrossRefGoogle Scholar
  41. 41.
    Liang Z et al. 2005. Silencing of CXCR4 blocks breast cancer metastasis. Cancer Res. 65:967–971.PubMedPubMedCentralGoogle Scholar
  42. 42.
    Kang H et al. 2005. Stromal cell derived factor-1: its influence on invasiveness and migration of breast cancer cells in vitro, and its association with prognosis and survival in human breast cancer. Breast Cancer Res. 7:R402–R410.CrossRefGoogle Scholar
  43. 43.
    Li YM et al. 2004. Upregulation of CXCR4 is essential for HER2-mediated tumor metastasis. Cancer Cell 6:459–430.CrossRefGoogle Scholar
  44. 44.
    Collert MA, Struhl K. 1994. NOT1(CDC39), NOT2(CDC36), NOT3, and NOT4 encode a global-negative regulator of transcription that differentially affects TATA-element utilization. Genes Dev. 8:525–537.CrossRefGoogle Scholar
  45. 45.
    Cheng L, Wing HW. (2001a). Model-based analysis of oligonucleotide arrays: Expression index computation and outlier detection. Proc. Natl. Acad. Sci. 98:31–36.CrossRefGoogle Scholar
  46. 46.
    Cheng L, Wing HW. (2001b). Model-based analysis of oligonucleotide arrays: model validation, design issues and standard error application. Genome Biol. 2:research0032.1–0032.11Google Scholar
  47. 47.
    Nebert DW et al. 2000. Role of the aromatic hydrocarbon receptor and [Ah] gene battery in the oxidative stress response cell cycle control and apoptosis. Biochem. Pharmacol. 59:65–85.CrossRefGoogle Scholar
  48. 48.
    Reymann S, Borlak J. 2006. Transcriptome profiling of human hepatocytes treated with Aroclor 1254 reveals transcription factor regulatory networks and clusters of regulated genes. BMC Genomics 7:217–235.CrossRefGoogle Scholar
  49. 49.
    Liu S et al. 2006. Aryl hydrocarbon receptor agonists directly activate estrogen receptor α in MCF-7 breast cancer cells. Biol. Chem. 387:1209–1213.CrossRefGoogle Scholar

Copyright information

© Feinstein Institute for Medical Research 2007

Authors and Affiliations

  • Laura Mulvey
    • 1
  • Alamelu Chandrasekaran
    • 1
  • Kai Liu
    • 1
  • Sarah Lombardi
    • 1
  • Xue-Ping Wang
    • 1
  • Karen J. Auborn
    • 1
    • 2
    • 3
  • Leslie Goodwin
    • 1
    Email author
  1. 1.Feinstein Institute for Medical ResearchManhassetUSA
  2. 2.Department of Otolaryngology, Long Island Jewish Medical CenterThe Long Island Campus of Albert Einstein College of MedicineNew Hyde ParkUSA
  3. 3.Department of Microbiology and ImmunologyAlbert Einstein College of MedicineBronxUSA

Personalised recommendations