Estrogen Receptor Alpha Mediates Progestin-Induced Mammary Tumor Growth by Interacting with Progesterone Receptors at the Cyclin D1/MYC Promoters Free

Breast cancer is the most frequently diagnosed cancer and a leading cause of cancer death in women worldwide (1). Although most of the evidence suggests estrogens as the major etiologic factor in breast cancer (2), experimental and epidemiologic evidence, reviewed recently (3–5), also points to the involvement of progesterone receptors (PR) in breast cancer development and progression. However, the mechanisms by which PR participate in tumor growth are not yet well understood. Considering that PR is usually used as a marker of estrogen receptor alpha (ERα) functionality (6), it may be intuitive to think that there is a sequential effect on ERα inducing PR expression. It has been reported that an early cytoplasmic interaction between ERα and PR isoform B (PR_B) is necessary to activate c-Src/p21^ras/Erk cascade by progestins (7), which in turn phosphorylates PR. Moreover, the regions through which both receptors interact have been identified (8). Conversely, Boonyaratanakornkit and colleagues have proposed that a polyproline motif in the amino-terminal domain of PR is sufficient to mediate c-Src tyrosine kinase activation by progestins (9).

Using a progestin-dependent murine mammary carcinoma, C4-HD (10) and the human T47D breast cancer cells, which are also stimulated by progestins (11, 12), we show that a genomic interaction between ERα and PR is essential for progestin-induced gene expression and tumor cell proliferation. Chromatin immunoprecipitation (ChIP) using T47D cells, confirms that PR is activated in the absence of ERα. However, the presence of both activated receptors at the MYC or CYCLIN D1 (CCND1) promoters is required to trigger gene expression and cell proliferation. Moreover, the nuclear colocalization of both receptors in human breast cancer samples suggests that a genomic interaction between activated ERα and PR may be a common event in breast cancer growth.

PR (C-19 and H-190X), Erk1/2 (sc-94), ERα (MC-20 and HC-20X), AIF (sc-5586), BAX (sc-493), BCL/XL (sc-634), and IgG (sc-2027) are rabbit polyclonals (Santa Cruz Biotechnology); PR (Ab7) and ERα (Ab10) are mouse monoclonals and ERα (SP1, #RM-9101) a rabbit polyclonal (Thermo Scientific); CCND1 (#2978), pSer118 ERα (#2515), pSer167 ERα (#2514), MYC (#5605) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; #2118) are rabbit polyclonals (Cell Signaling Technology). Mouse monoclonal pSer162 PR_B, pSer190 PR, and pSer294 PR were a gift from Dr. D. Edwards (BCM, Houston, TX); ERα (M7047) and PR (M3568) are mouse monoclonals (Dako); pSer294 PR (Ab61785) and Ki67 (Ab15580), are rabbit polyclonals (Abcam). Secondary antibodies were obtained from Vector Labs.

4′, 6-Diamidino-2-phenylindole (DAPI), medroxyprogesterone acetate (MPA, 10 nmol/L), and RU-38486 (RU, 10 nmol/L) were purchased from Sigma. ICI 182.780 (ICI) was a gift from AstraZeneca.

Two-month-old virgin female BALB/c mice (IByME-Animal Facility) were used. Animal care and manipulation were in agreement with institutional and reference guidelines (13).

Depot MPA (20 mg) was used as a progestin. C4-HD tumors were subcutaneously transplanted into MPA-treated BALB/c mice as previously described (10). When tumors reached a size of approximately 50 mm², 6 mice were treated subcutaneously, as described (14), with Fulvestrant (FUL; AstraZeneca), 6 received no other treatment, and the MPA depot was removed in another 6 mice.

Breast cancer resection specimens from 15 patients immediately frozen at −70°C were provided by Bancario Hospital, Buenos Aires. The study was approved by the Institutional Review Board.

Human T47D cells obtained from American Type Culture Collection were validated by Genetica DNA Laboratories Inc. by short tandem repeat profiling and maintained as described (15). Passages lower than 15 were used.

Primary cultures of C4-HD tumors were carried out as described previously (16). Cell proliferation was evaluated by either [³H]-thymidine uptake (16) or cell counting. C4-HD and T47D cells were plated with Dulbecco's Modified Eagle's Medium/F12 (Sigma) plus 10% fetal calf serum (FCS; BioSer) for 48 hours. After starving for 24 hours with 1% steroid-stripped FCS (chFCS), the cultures were incubated with the experimental solutions.

T47D cells were seeded in 12- or 96-well plates and transfected with short interfering RNAs (siRNA) to human ERα (ESR1_8 and ESR1_10, QIAGEN), human CCND1 (ON-TARGETplus SMARTpool CCND1 from Thermo, or a pool of CCND1_5 and CCND1_6 from QIAGEN), or a nonspecific siRNA (SI03650318, QIAGEN) using HiPerFect transfection reagent (QIAGEN). Cells were used 48 hours posttransfection.

Sections of formalin-fixed, paraffin-embedded tissues were reacted with different antibodies using the avidin–biotin peroxidase complex technique (Vector Lab) and counterstained with hematoxylin (17). Positive cells were counted in 10 high-power fields (HPF, 1,000×) of each section and expressed as the mean ± SEM of the percentage of the ratios between the number of events and the cell number/HPF.

Frozen tumor sections were fixed in formalin, postfixed in 70% ethanol, blocked, and incubated with the primary antibodies and fluorescein isothiocyanate/TX-conjugated secondary antibodies, and counterstained with DAPI as described previously (18). Images were obtained using a Nikon Eclipse E800 Confocal Microscope and Nikon DS-U1 with ACT-2U software.

Cultures growing on chamber slides were fixed in 70% ethanol and processed as described previously (18). To quantify nuclear colocalization of PR and ERα, we used the Pearson's correlation coefficient (R_r). Nuclei (200) of selected samples were analyzed by using PSC Colocalization plug-in (ImageJ-NIH; ref. 19). R_r ranges between −1 (perfect negative correlation) to +1 (perfect positive correlation) with 0 meaning no correlation.

Tumors were homogenized and processed to obtain nuclear purified fractions (20) and total cell extracts prepared using M-PER mammalian protein extraction reagent (Pierce). Nuclear cell culture extracts were obtained and proteins quantified as described previously (21).

Nuclear extracts containing 0.5 to 1 mg of proteins were subjected to immunoprecipitation (IP) using 2 μg of PR or ERα antibodies and rocked overnight at 4°C. The immunocomplexes were then captured by adding protein A-agarose (Santa Cruz) processed as described (18) and subjected to Western blots.

Tumor, cell extracts (100 μg proteins/lane), or immunoprecipitated proteins were separated on discontinuous polyacrylamide gels and detected as previously described (20).

The PRE-Luc vector used was a gift from Dr. C. Gardmo (Karolinska Institutet, Stockholm, Sweden; ref. 22) and assays were conducted as described previously (18).

Total RNA was isolated from cultures with TRIzol Reagent (Invitrogen) and converted to cDNA as described previously (18). Specific oligos for human MYC (NM_002467.4) and CCND1 (NM_053056.2) were designed using Primer-Blast (NCBI; Supplementary Table S1). GAPDH (NM_002046.3) expression was used as a normalization control. Data from 3 experiments were combined to determine gene expression changes using 2^(–ΔC_t) formula. A melting curve was generated for every run to confirm assay specificity.

After treatment, cells were fixed with 1% paraformaldehyde for 30 minutes; ChIP assays carried out as recommended by Diagenode using the HighCell# ChIP kit. Specific oligos for human CCND1 and MYC promoters were designed using Primer-Blast (NCBI; Supplementary Table S1). The data from each immunoprecipitate (IgG, PR, and ERα) was normalized to the corresponding inputs of chromatin before IP, normalized to IgG/input data, and expressed as relative to the control. Five experiments were combined to determine receptor binding to gene promoters. Sequential ChIP (ChIP-reChIP) was carried out using the Re-ChIP-IT kit (Active Motif). Data from each sequential immunoprecipitates (PR/ERα and ERα/PR) were normalized to the corresponding inputs before IP, normalized to IP IgG/IgG data, and expressed as relative to the control.

ANOVA and Tukey multiple post t test were used to evaluate differences of means of multiple samples, and Student t test was used to compare means of 2 different groups. In all graphs, the mean ± SEM is shown, and experiments were repeated at least 3 times. Significant differences between control and treated cells were indicated with asterisk (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

We have previously shown that C4-HD tumors that express ERα and PR grow in MPA- or progesterone (Pg)-treated female mice (10) and that the blockade of PR induces complete tumor regression (23). This experimental system provided an opportunity to explore the role of ERα in progestin-induced tumor growth by using the pure antiestrogen FUL. Surprisingly, FUL induced a complete regression of tumors growing in the presence of MPA (Fig. 1A), and this was associated with a decrease in both PR isoforms and ERα expression, as evaluated by Western blot (Fig. 1B) and immunohistochemistry (Fig. 1C). Expression of ERα after MPA withdrawal was negligible, however, a significant increase in PR was observed after MPA removal, suggesting that, in the progestin-dependent C4-HD tumor, although MPA downregulates PR expression, it may be required to maintain high levels of ERα expression in vivo (Fig. 1C). Moreover, activated ERα (pSer167 and pSer118 ERα) was also high in MPA-treated tumors (Supplementary Fig. S1A). FUL-induced tumor regression was associated with a cytostatic effect, as shown by a decrease in the mitotic index (Ki67 quantification, Supplementary Fig. S1B), and in the expression of 2 progestin-regulated proteins, CCND1 and MYC (Fig. 1C). In addition, in FUL-treated tumors, an increase in apoptosis (Supplementary Fig. S1B), associated with a decrease in BCL/XL, and an increase in BAX and AIF (Supplementary Fig. S1C) expression were observed. These results indicated that activated ERα contributes to progestin-dependent tumor growth.

The fact that high levels of PR, but not of ERα, were observed in the nuclei of C4-HD tumors after MPA removal, led to hypothesize that both receptors participated in growth stimulation. Therefore, we investigated the effect of the blockade of ERα on MPA-induced cell proliferation and the role of MPA on ERα and PR expression in vitro. In C4-HD cultures, ICI inhibited MPA-induced proliferation as shown by [³H]-thymidine uptake (16), cell counting (Fig. 2A), or bromodeoxyuridine staining (Supplementary Fig. S2A). Similarly, blocking ERα expression using siRNAs also inhibited the MPA-induced increase in [³H]-thymidine uptake (Supplementary Fig. S2B). A time course analysis of ERα and PR expression after ICI treatment showed an early downregulation of ERα (6 hours), although high levels of PR were still detected after 24 hours (Supplementary Fig. S2C), indicating that the blockade of MPA-induced cell proliferation by ICI was not associated with PR downregulation.

An increase in both nuclear ERα and PR immunoreactivity and nuclear colocalization was observed in MPA-treated cells (Fig. 2B). A time course analysis of the interaction revealed that they start colocalizing as early as 5 minutes after MPA incubation with a decrease after 1 hour (Fig. 2B). In cells treated for 30 minutes with MPA+ICI, there was a decrease in nuclear and an increase in cytosolic ERα staining (Fig. 2C left, arrows). These results suggested that ICI disrupts the molecular interaction induced by MPA. Similar incubations were done with the corresponding phospho-receptor antibodies. Phospho-Ser118 ERα staining increased after 30 minutes of MPA treatment and colocalized with pSer162 PR_B (Fig. 2C, middle) or pSer294 PR (Fig. 2C, right). These observations suggested that ERα and PR may be forming part of the same complexes in their active state (24). No cytosolic or membrane colocalization of PR and ERα was observed in MPA-treated cells and no staining was observed in hormone receptor–negative murine LM3 (25) breast cancer cells (data not shown). Moreover, using frozen samples from C4-HD tumors growing in MPA-treated mice, we confirmed the nuclear colocalization between PR/ERα in vivo (Fig. 2D, left). Finally, we corroborated the interaction between both receptors by co-IP assays using nuclear extracts from MPA-treated C4-HD tumors. Proteins were immunoprecipitated with 2 different PR or ERα antibodies and blotted accordingly (Fig. 2D, right). These results suggested that both PR isoforms can participate in a nuclear complex with ERα.

To investigate whether the colocalization between ERα and PR was unique for our murine model, we evaluated the expression of ERα, PR, and pPR in 15 frozen breast cancer samples. In 4 of them (2 ductal and 2 lobular carcinomas), we found a high degree of nuclear colocalization (Fig. 3A). We found a mild colocalization in 3 samples and a sporadic colocalization in other 2 samples. No staining was observed in receptor negative tumors (Fig. 3B). Co-IP assays carried out using purified nuclear extracts from 2 positive samples and a negative control confirmed the nuclear interaction between ERα and PR (Fig. 3C). These results suggested that the interaction between ERα and PR has an important and yet unexplored role in human breast cancer.

To further investigate the role of ERα in MPA-induced cell proliferation we used T47D cells. MPA increased the nuclear colocalization between ERα and PR during the first 5 to 10 minutes and then a decrease was observed after 30 minutes of treatment (Fig. 4A). No cytosolic or membrane colocalization of PR and ERα was observed. Using phospho-specific antibodies, we showed that pSer162 PR_B and pSer294 PR colocalized with ERα after 10 minutes of MPA incubation (Supplementary Fig. S3A and S3B). Purified nuclear extracts from untreated or MPA-treated cells were immunoprecipitated with PR or ERα antibodies. We observed a significant increase in pSer294 PR (P < 0.01) and in ERα (P < 0.05), or in total PR (P < 0.01), respectively, as compared with immunoprecipitates from untreated cells (Fig. 4B). Cellular fractionation was controlled by Western blot using anti-tubulin or anti-Sp1 antibodies (Supplementary Fig. S3C). These results showed that both PR isoforms interact with ERα in the cell nuclei of human progestin-treated cells. We then explored the role of ERα in MPA-driven proliferative responses. ICI (0.1 and 1 μmol/L) dramatically inhibited DNA synthesis to levels similar to those of the antiprogestin RU (Fig. 4C). In addition, we used 2 different siRNAs that decreased ERα expression (Fig. 4D, left) and also inhibited MPA-induced [³H]-thymidine uptake (Fig. 4D, right).

As part of their proliferative activity, progestins induce the expression of CCND1 (18, 26–30) and MYC (18, 31, 32) mRNA in T47D cells. We analyzed their time-dependent expression in response to MPA. We observed an early increase (15 minutes) after MPA incubation that lasted 24 hours, except for a decrease observed 1 hour (CCND1) or 3 hours (MYC) after treatment (Fig. 5A). The increase in mRNA correlated with an early and gradual increase in protein expression (Supplementary Fig. S4). The knockdown of CCND1 using siRNAs prevented DNA synthesis triggered by MPA (Fig. 5B). We therefore used ICI or siRNAs to analyze the contribution of ERα to gene transcription activated by MPA. The inhibition of ERα blocked the MPA-dependent transcription of both CCND1 and MYC genes (Fig. 5C and D). All this data suggested that ERα activity, presumably through its ability to interact with PR by forming nuclear complexes, can control the expression of key proliferative genes in response to progestins.

To further understand the role of ERα mediating MPA transcriptional activities, we evaluated the effect of ICI on the activation of a reporter luciferase assay controlled by the progesterone response element (PRE) sequence in T47D cells. ICI inhibited MPA-induced PRE-luc expression (Fig. 6A) and induced the downregulation of ERα, whereas PR was still expressed even after 48 hours of ICI incubation (Supplementary Fig. S3D). Moreover, MPA induced a higher PRE-luc activity in MDA-MB-231 cells stably transfected with PR_B, when they were cotransfected with ERα (Supplementary Fig. S5). These results strongly suggested a role for the PR/ERα complexes in the regulatory elements of MPA-regulated genes. To confirm the binding of both receptors to the same promoter regions, we used ChIP analysis on CCND1 and MYC regulatory sequences. In Fig. 6B we show a schematic representation of both gene promoters, highlighting the PRE and estrogen response element (ERE) sites in each case, as well as the primers used in ChIP/qPCR analysis. Cells were incubated with MPA (10 minutes) and the chromatin subjected to IP with PR- or ERα-specific antibodies. DNA fragments were amplified by qPCR with 3 pairs of primers for each gene, previously used by others to report PR binding to those sequences (refs. 31, 33, 34; Fig. 6B). The recruitment of ERα and PR to the sites at +5 to 6 Kb (ChIP primers C) was used as a negative control of receptor binding (Fig. 6B). Specific binding of both receptors was detected at the same promoter regions in each gene (ChIP primers A and B) after MPA treatment (Fig. 6C and D, left and middle panels). We then evaluated whether PR and ERα were simultaneously bound to the CCND1 and MYC gene promoters by using a sequential ChIP assay. PR or ERα antibodies were used in the first IP, and ERα or PR antibodies in the sequential ChIP (reChIP). qPCR analysis clearly showed that PR and ERα co-occupy the CCND1 and MYC promoters after 10 minutes of MPA stimulation (Figs. 6C and D, right panels). These findings suggested that progestins induce the assembly of PR/ERα protein complexes at both promoters to control its transcriptional activation in breast cancer cells.

To further understand the molecular mechanism driving these effects, we evaluated ERα and PR binding to these regulatory sequences when we inhibited ERα. PR binding to both gene promoters was unaffected by the presence of ICI (Fig. 7A) or siRNA to ERα (Fig. 7B), although they did prevent ERα binding. These data indicated that both proteins need to interact at the CCND1 and MYC promoters to induce gene transcription and cell proliferation, supporting our hypothesis that the presence of ERα at those promoters is required to induce PR-mediated gene expression.

In this study we have shown that a progestin can induce a direct and transient nuclear interaction between ERα and both PR isoforms at the promoters of 2 progestin responsive proto-oncogenes, namely CCND1 and MYC. Moreover, this activity can have dramatic effects on breast cancer cell proliferation and seems to be dependent on ERα actions, as its inhibition with ICI induced complete regression of C4-HD tumors growing in the presence of the progestin. Thus, our results suggest that a combined treatment with antiestrogens and antiprogestins can be beneficial to breast cancer patients. As it has previously been reported (35), the cotreatment with antiprogestins plus selective estrogen receptor modulators may have an additive effect. Moreover, MPA-independent murine mammary carcinomas, C4-HI, respond better to a combination of tamoxifen and mifepristone than to both single agents (36).

We confirmed our observations in the murine model, using T47D cells in which the inhibition of ERα activity resulted in a complete blockade of MPA-dependent MYC and CCDN1 gene transcription and cell proliferation. The fact that progestins exerted growth inhibitory effects on MDA-MB-231 cells stably transfected with PR (37) but stimulated cell proliferation in models that coexpress ERα and PR (11, 12, 15, 16, 38), also suggests that both receptors cooperate to trigger cell proliferation. In this regard, it is known that the human MYC gene promoter contains a functional PRE that mediates the binding of activated PR (18, 31, 39) and we also identified other consensus PRE half sites (40) that might also bind PR (Fig. 6B). Moreover, it has recently been reported that ERE half sites at the MYC proximal promoter (Fig. 6B) are not responsive to estrogens (41). It may be possible that after progestin treatment, these sites might also bind ERα in complexes with PR. In addition, we have recently shown in T47D cells that MPA induces the binding of PR, transcription factors (TF), such as STAT5, and nuclear tyrosine kinase receptors (RTK), such as FGFR-2, to the same regions of the MYC promoter (18). The results reported herein indicate that activated ERα could be present in the same multimeric protein complexes as supported by NoShift electrophoretic mobility shift assays (ref. 18; Fig. 7C -iii-). The regulation of human CCND1 by progestins may be more complicated, as no canonical PRE sites have been described in its promoter and accordingly, it has been suggested that PR regulates CCND1 expression by nongenomic mechanisms (7, 9, 34). The 2 models of cytoplasmic signaling pathways activated by Pg are shown in Fig. 7C. Model -i- proposes that an early interaction between ERα and PR_B is necessary for c-Src/p21^Ras/Erk, PI3K/Akt, and JAK/STAT activation (7, 8, 42); conversely, model -ii- proposes that a polyproline motif in the amino-terminal domain of PR is sufficient to activate cell signaling pathways (9). Albeit a cytoplasmic as well as membrane localization of PR has been shown (21), we were not able to find PR colocalizing with ERα at these sites. Activated growth factor receptors, usually RTKs, may stimulate cytoplasmic signaling pathways, that in turn induce PR phosphorylation and activation in both the absence or presence of steroids (15, 43). Both models propose that these nongenomic effects of Pg-activated MAPKs use TF at the CCND1 promoter, inducing gene transcription and subsequently cell proliferation (43; Fig. 7C,-v-). However, it has been recently shown that PR may have genomic effects at the CCND1 promoter (34, 44), even as a coactivator of STAT3 (26; Fig. 7C,-iv-). In this study, we showed for the first time that PR and ERα share the same progestins-sensitive regions at CCND1 and MYC promoters (Fig. 7C,-iii-). Interestingly, we found distinct consensus PRE half sites (40) at the CCND1 promoter (Fig. 6B), which might bind activated PR, as shown for others genes (45, 46). Aligned with our observations, the hypothesis that both ERα and PR can interact at the gene promoter level has been proposed by other authors in different contexts (44). However, this is the first report showing that PR and ERα are recruited to the same sites at the CCND1 and MYC promoters after PR activation by MPA. It has been described that SRC (steroid receptor coactivator) proteins may also participate in this response (47), but we have not yet studied their involvement in this setting. Our results also show that antiestrogenic concentrations of ICI (≤ 1 μmol/L) block the formation of MPA-induced PR/ERα nuclear complexes, inhibiting gene transcription and cell proliferation, without affecting the activation and binding of PR at the gene promoter. This implicates a change in the paradigm that a rapid, nongenomic interaction between PR_B and ERα is necessary to activate the c-Src/p21^ras/Erk cascade and PR by progestins. Whether the genomic interaction described here also involves ERID domains (8) at PR, remains to be investigated.

The expression of MYC and CCND1 constitutes an early and transient event mediated by MPA, and it is quite conceivable that PR/ERα complexes driven effects are required to unwind the chromatin. This may be followed by the recruitment of other transcription factors, and full transcription of proliferative oncogenes. In addition, this activity could also be required for transcription events induced by other mitogens such as epidermal growth factor (48). On the other hand, MYC can also be involved in the activation of cyclins (D1, D2, E1, and A2), cyclin-dependent kinases (CDK4), and in the downregulation of cell-cycle inhibitors (49).

Finally, in this study we also showed that both receptors interact in the nuclei of selected human breast cancer samples, suggesting that ligand-independent hormone receptor activation may also be implicated in breast cancer tumor growth in patients. Thus, it is possible to speculate that patients showing higher levels of PR/ERα colocalization may have a better response to a combined antiprogestin–antiestrogen therapy.

No potential conflicts of interest were disclosed.

Conception and design: S. Giulianelli, J.P. Vaqué, A.A. Molinolo, J.S. Gutkind, C. Lanari

Development of methodology: S. Giulianelli, J.P. Vaqué, V. Wargon, L.A. Helguero, C.A. Lamb, J.S. Gutkind

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Giulianelli, R. Soldati, V. Wargon, S.I. Vanzulli, R. Martins, E. Zeitlin

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Giulianelli, J.P. Vaqué, A.A. Molinolo, C. Lanari

Writing, review, and/or revision of the manuscript: S. Giulianelli, J.P. Vaqué, A.A. Molinolo, L.A. Helguero, C.A. Lamb, J.S. Gutkind, C. Lanari

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Giulianelli

Study supervision: S. Giulianelli, C. Lanari

The authors thank Dr. C. Gardmo (Karolinska Institutet, Stockholm) for providing the PRE-Luc plasmid, Dr. D. Edwards (BCM, Houston, TX) for the pPR antibodies, AstraZeneca (UK) for providing Fulvestrant, Dr. María Gorostiaga, Pablo DoCampo, and Bruno Luna for excellent technical assistance, and Dr. A. Pecci for helpful assistance.

This work was supported by SECYT (PICT2007/932), Carrillo-Oñativia Fellowships 03/04, CONICET and Fundación Sales. J.S. Gutkind, A.A. Molinolo, and J.P. Vaqué are supported by the Intramural Research Program (NIDCR-NIH). S. Giulianelli and V. Wargon are fellows of CONICET. C.A. Lamb and C. Lanari are members of the Research Career, CONICET. S. Giulianelli received awards from Avon Foundation to present data at the AACR Meetings 2009–2011, and he received an ICRETT fellowship (UICC) to carry out ChIP studies in Dr. Gutkind's Laboratory.

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.