Galeterone Prevents Androgen Receptor Binding to Chromatin and Enhances Degradation of Mutant Androgen Receptor

Prostate cancer is the second leading cause of cancer-related death in men in the United States. The androgen receptor (AR) plays a central role in prostate cancer, and the standard treatment for metastatic prostate cancer is androgen deprivation therapy by medical or surgical castration. Although most patients initially respond, they invariably relapse despite castrate androgen levels (castration-resistant prostate cancer, CRPC). Previous studies have identified increased intratumoral androgen synthesis from precursor steroids generated by the adrenal glands or possibly de novo androgen synthesis from cholesterol, as a mechanism of castration resistance (1–6). CYP17A1 is the critical enzyme required for the conversion of C21 steroids to C19 steroids such as dehydroepiandrosterone (DHEA) that can be further reduced to the potent androgens testosterone and dihydrotestosterone (DHT). CYP17A1 inhibitors can thereby further markedly decrease the levels of residual androgens and precursor steroids that remain after castration, and the CYP17A1 inhibitor abiraterone is now approved by the FDA for treatment of CRPC (7, 8).

The direct AR antagonist enzalutamide has also recently been approved for treatment of CRPC (9, 10). Previous AR antagonists used for prostate cancer (flutamide, nilutamide, and bicalutamide) do not effectively prevent AR binding to chromatin and may thereby have weak agonist properties that limit their efficacy in CRPC (11–13). In contrast, the enzalutamide-liganded AR does not bind to chromatin, making this drug a purer AR antagonist with improved efficacy in CRPC (10). However, the survival advantages for abiraterone and enzalutamide therapy in CRPC post-chemotherapy are only about 4 months (7, 9), and mechanisms of intrinsic or acquired resistance to these agents remain to be established (14).

Galeterone (previously known as VN/124-1 or TOK-001) was developed as a CYP17A1 inhibitor, but similar to related compounds, it has AR antagonist activity and was also found to promote AR degradation (15–17). However, its effects on AR binding to chromatin have not been examined. Moreover, further studies indicated that galeterone at high concentrations could induce an endoplasmic reticulum (ER) stress response (18) and may decrease AR translation through direct or indirect effects on mTOR (19), suggesting that some of its effects on AR expression may be indirect. Galeterone is currently in phase II clinical trials for CRPC, and responses in these trials may be related to both its activities toward CYP17A1 and its direct effects on AR. Therefore, this study was undertaken to determine the molecular basis for galeterone actions as a direct AR antagonist and for its effects on AR protein expression.

LNCaP, VCaP, LAPC4, CWR22RV1, PC3, and HEK293T cells were purchased from ATCC. LAPC4-CR and C4-2 cells were derived from castration resistant xenografts of LAPC4 and LNCaP, respectively. Cells were cultured in RPMI-1640 or DMEM supplemented with 10% FBS. Androgen deprivation was conducted by culturing cells in RPMI-1640 or DMEM supplemented with 5% CSS for at least 48 hours. Whole-cell lysates (WCL) were prepared using lysis buffer containing 2% SDS and subjected to immunoblotting. Nuclear and cytoplasmic fractions of the cell were prepared using the NE-PER nuclear and cytoplasmic extraction kit (Thermo Scientific) following manufacturer's instructions. The antibody against human AR (N20) and lamin A/C was obtained from Santa Cruz Biotechnology. The antibody against V5 epitope tag was from Invitrogen. Antibodies against β-actin (AC-15) and β-tubulin were from Millipore. Antibody against PSA was from Meridian Life Science. Antibodies against eIF2α and phospho-eIF2α were from Cell Signaling. The results from a minimum of 2 experiments were subjected to densitometry and normalized to β-actin or β-tubulin loading control and the mean values relative to vehicle control (set to 1.0) given.

The AR-WT, T878A, and W742C cDNA were subcloned from previously described constructs (20, 21) to pCMV-3xFLAG vector and to pDONR223 entry vector. For AR-F877L mutant, site-directed mutagenesis using pDONR223-AR-WT as template was performed. The primer sets used were: sense 5′-cgagagagctgcatcagctcacttttgacctgcta-3′, antisense 5′-tagcaggtcaaaagtgagctgatgcagctctctcg-3′. pLenti6.3 lentiviral expression vectors carrying AR-WT, T878A, or F877L mutant sequences were then generated using the Gateway Cloning System (Invitrogen). For lentivirus production, pLenti6.3-AR-WT, T878A, or F877L were co-transfected with pCMV-VSVG and pCMV-dR8.91 (Addgene) into HEK293T (ATCC) using Lipofectamine 2000 (Invitrogen). LNCaP and PC3 cells were infected with lentivirus expressing AR-WT or AR-T878A in the presence of polybrene before subsequent selection for cells with stable ectopic gene expression with blasticidin (Invitrogen).

COS-7 cells were transfected with a luciferase reporter gene construct driven by 4 androgen-responsive elements (ARE), Renilla luciferase gene construct, and pCMV-FLAG-AR or pCMV-VP16-AR in RMPI-1640 with 5% CSS. Forty-eight hours later, transfected cells were treated with DHT or vehicle (ethanol) in the presence of galeterone (Tokai Pharmaceuticals), abiraterone, bicalutamide, or enzalutamide. Transfection was done in triplicates. Luciferase and Renilla luciferase activity was measured using the Dual-luciferase Reporter Assay (Promega). For quantitative RT-PCR (qRT-PCR), RNA was isolated using RNeasy Mini Kit (Invitrogen). The TaqMan primer probe sets for PSA were: Forward, 5′-gatgaaacaggctgtgccg-3′, reverse, 5′-cctcacagctacccactgca-3′, probe, 5′-FAM-caggaacaaaagcgtgatcttgctggg-3′. The primer probe sets for FKBP5 and GAPDH were from Life Technologies. The qRT-PCR was performed in an ABI7900 thermal cycler.

For anti-AR chromatin immunoprecipitation (ChIP), androgen-deprived VCaP cells were treated with DHT or antagonists for 4 hours. DNA–protein crosslinking was induced by formaldehyde treatment and the cell lysates were subjected to sonication. Immunoprecipitation was done with anti-AR antibody or control IgG (Santa Cruz). The bound DNA was quantified by qRT-PCR using the following primer sets: PSA enhancer: Forward, 5′-tcgattgtccttgacagtaaaca-3′, reverse, 5′-tctcagatccaggcttgctt-3′; KLK2 enhancer: Forward, 5′-gcctttgctcagaagacaca-3′, reverse, 5′-acaagagtggaaggctctgg-3′; TMPRSS2 enhancer: Forward, 5′-tggtcctggatgataaaaaaagttt-3′, reverse, 5′-gacatacgccccacaacaga-3′; and NKX3.1 enhancer: Forward, 5′-ctggcaaagagcatctaggg-3′, reverse, 5′-ggcacttcctgagcaaactt-3′.

We first examined the effects of galeterone on AR signaling in prostate cancer cells. Androgen-deprived LNCaP and VCaP cells were stimulated with DHT at 0.1, 1, or 10 nmol/L in the presence of 1 or 10 μmol/L galeterone, and AR activity was assessed by PSA protein expression. A dose-dependent repression of PSA expression by galeterone was observed in both LNCaP and VCaP cells (Fig. 1A). This repression could be partially rescued with higher concentrations of DHT, consistent with galeterone binding to the steroid-binding pocket and acting as a competitive antagonist.

PSA mRNA levels were also examined in androgen-deprived LNCaP and VCaP cells treated with galeterone or vehicle control, in the presence or absence of DHT. The effect of galeterone was compared with those of abiraterone (CYP17A1 inhibitor) and of 2 other direct AR antagonists (enzalutamide and bicalutamide). Galeterone and enzalutamide decreased basal PSA mRNA in both cell lines in the absence of androgen (Fig. 1B). In contrast, abiraterone had minimal effects, consistent with galeterone directly suppressing basal AR activity in these cells independently of its CYP17A1 inhibitor activity. As expected, treatment with DHT was able to partially rescue PSA expression repressed by galeterone or enzalutamide. Bicalutamide more modestly decreased PSA expression in both cell lines treated with DHT but did not decrease basal PSA levels in the absence of ligand, consistent with it being a less potent AR antagonist with weak partial agonist activity.

The antagonist activity of galeterone was further assessed using COS-7 cells co-transfected with an AR expression construct and a luciferase reporter driven by 4 consecutive ARE (ARE4-LUC) and treated with increasing concentration of DHT. As shown in Fig. 1C, galeterone potently blocked DHT-induced AR transcription activity to a level comparable to that seen with bicalutamide. To directly evaluate the effect of galeterone on expression of an endogenous AR target gene, PC3 cells stably expressing AR-WT were treated with bicalutamide, enzalutamide, or galeterone. Expression of FKBP5 mRNA was tested as an indicator of AR transcription activity. Galeterone and enzalutamide both potently blocked FKBP5 mRNA expression induced by up to 10 nmol/L DHT, whereas this inhibitory effect was partially reversed by the treatment with 100 nmol/L DHT. Overall, these results confirm that galeterone can function as a direct AR antagonist.

An important distinction between bicalutamide and enzalutamide is that the bicalutamide-liganded AR can associate with chromatin, which may contribute to weak partial agonist activity (10, 12). To test whether galeterone stimulates AR binding to DNA, we used a plasmid encoding the full-length AR fused to the VP16 transactivation domain (VP16-AR), which can stimulate transcription independently of coactivator recruitment by AR. COS-7 cells were transfected with this plasmid and with the ARE4-LUC reporter construct and then stimulated with a series of ligands (Fig. 2A). As a positive control, DHT induced ARE-LUC activity in a dose-dependent fashion, indicating the binding of VP16-AR to the ARE4-LUC construct (Fig. 2A). As we showed previously, the VP16-AR similarly could be stimulated by bicalutamide, indicative of AR binding to the reporter gene (12). In contrast, also as shown previously, enzalutamide did not stimulate the VP16-AR (10). Significantly, galeterone similarly failed to activate the VP16-AR, indicating that it does not support DNA binding.

We next used ChIP to test whether galeterone can induce recruitment of endogenous AR to its target genes. When androgen-deprived VCaP cells were treated with galeterone or enzalutamide, no increases in AR recruitment to the ARE sites on the PSA, KLK2, NKX3.1, and TMPRSS2 genes were observed compared with vehicle-treated cells (Fig. 2B). As a positive control, DHT treatment led to substantial binding of AR at these sites. Consistent with previous studies, bicalutamide stimulated AR binding to the PSA gene (Fig. 2C). These data together indicate that galeterone, similar to enzalutamide, does not stimulate AR binding to chromatin.

To test whether galeterone has an effect on nuclear AR levels, VCaP and LNCaP cells were treated overnight with DHT, galeterone, enzalutamide, or vehicle. Nuclear and cytoplasmic fractions were separated, and AR protein expression in these fractions was tested. As a positive control, DHT treatment led to substantial increases in nuclear AR protein expression in both VCaP and LNCaP cells (Fig. 2D). Consistent with previous findings, enzalutamide somewhat decreased nuclear AR expression in VCaP or LNCaP cells, without substantial effects on cytoplasmic AR (10). The effects of galeterone in VCaP cells were similar to enzalutamide. In contrast, while galeterone markedly decreased nuclear AR in LNCaP cells, there was also a decrease in cytoplasmic AR. These data together show that galeterone does not induce AR accumulation in the nucleus or binding to chromatin and may also decrease overall AR expression.

The unliganded AR is degraded by a proteasome-dependent pathway, and its stability is increased by androgen binding (22–24). Galeterone has been shown previously to decrease AR protein expression in prostate cancer cell lines (LNCaP and LAPC4) and in LAPC4 xenografts (16, 17). However, these decreases in AR protein and increases in AR degradation may, in part, be due to competitive displacement of androgen from the AR ligand domain by galeterone. Moreover, galeterone at higher concentrations (generally over 10 μmol/L) has been reported to activate cellular stress pathways that may indirectly decrease AR protein expression (18, 19). Therefore, to determine whether galeterone binding to the AR could directly enhance AR degradation, we assessed the effects of galeterone at up to 10 μmol/L in androgen-deprived cells. Importantly, on the basis of the results in Fig. 1, we anticipate that most of the AR would be liganded to galeterone at 10 μmol/L, so that effects at higher concentration would be indirect and mediated by off-target mechanisms.

In one set of experiments, we used LAPC4 cells and its subline LAPC4-CR cells, which both express wild-type (WT) AR. The LAPC4-CR line was established from a relapsed castration-resistant LAPC4 xenograft. The cells were cultured in androgen-depleted medium for 48 hours and then treated with galeterone at 0.5 to 10 μmol/L for 24 hours. In LAPC4 cells, galeterone at concentrations of 0.5 to 5 μmol/L did not lead to decreased AR expression but did lead to a 48% decrease at 10 μmol/L (Fig. 3A). To more directly assess AR protein stability, we then treated the androgen-deprived and vehicle or galeterone-stimulated LAPC4 cells with cycloheximide to block new protein synthesis. Significantly, the rate at which AR protein decayed in these cells was not altered by the treatment with galeterone compared with vehicle control (Fig. 3B). In LAPC4-CR cells, galeterone had no clear effect on overall AR protein levels or stability (Fig. 3C and D).

We next similarly examined VCaP cells, which express an amplified WT AR. In androgen-deprived VCaP cells, galeterone led to a moderate decrease of AR protein levels (Fig. 3E). However, as there is substantial de novo androgen synthesis in VCaP cells (2), any decrease in AR protein may reflect decreased intracellular androgen levels due to the inhibition of CYP17A1 by galeterone. To clarify whether any decrease in AR expression is a direct effect of galeterone on AR rather than an indirect effect via its CYP17A1 activity, we treated VCaP cells with abiraterone at up to 5 μmol/L, which effectively blocks CYP17A1 activity in these cells (2). Abiraterone treatment alone did not have a clear effect on AR protein, but addition of galeterone to the abiraterone did modestly decrease AR protein, suggesting a direct effect of galeterone on AR (Fig. 3F). The addition of galeterone at 10 μmol/L to the abiraterone similarly decreased AR protein in VCS2 cells, which were derived from a castration-resistant VCaP xenograft (2). However, when we next used cycloheximide treatment to directly assess for effects of galeterone on AR stability in VCaP cells, we found that galeterone at 10 μmol/L did not increase the rate of AR degradation (Fig. 3G). On the basis of these results, we conclude that galeterone at concentrations up to 10 μmol/L does not substantially enhance AR degradation in VCaP cells but may decrease AR protein by indirect mechanisms at higher concentrations or in combination with other agents.

Similar results were obtained when we examined CWR22Rv1 cells. The AR in the parental CWR22 cells has a mutation in codon 875 that broadens its ligand specificity, and the CWR22Rv1 cells have in addition a duplication of exon 3 encoding the DNA-binding domain (25). These cells also express an alternatively spliced AR isoform that deletes the ligand-binding domain (26). As observed in VCaP cells, AR protein was slightly decreased at 10 μmol/L galeterone, whereas bicalutamide, enzalutamide, or abiraterone treatment did not change AR levels (Fig. 3H). As expected, the full-length AR protein was slightly increased by DHT, with no clear effect on the AR splice variant (AR-V). Overall, these results indicate that galeterone does not have marked direct effects on AR protein stability but may decrease AR protein by indirect mechanisms at concentrations at or above 10 μmol/L.

In contrast to the results above, galeterone at concentrations well below 10 μmol/L in androgen-deprived C4-2 cells markedly decreased AR protein (Fig. 4A). This decrease was from 36% to 87% at 0.5 to 10 μmol/L, whereas there was no decrease in AR protein with abiraterone at 2.5 or 5 μmol/L (Fig. 4A). Using cycloheximide to block new protein synthesis, we found that galeterone decreased the half-life of endogenous AR in androgen-deprived C4-2 cells from about 6 to 2 hours (Fig. 4B). The parental line of C4-2, LNCaP, was also tested. Galeterone at 5 or 10 μmol/L similarly dramatically decreased AR protein expression in androgen-deprived LNCaP cells (Fig. 4C), and treatment with cycloheximide showed that galeterone was increasing AR degradation in these cells (Fig. 4D).

To confirm that the decrease in AR protein was through proteasome-mediated degradation, we used a proteasome inhibitor, MG132. Treatment of LNCaP cells with MG132 prevented the decrease in AR protein expression caused by galeterone (Fig. 4E). To determine whether the effects of galeterone were direct and mediated by AR binding, we also assessed whether DHT could prevent the decline in AR protein. As shown in Fig. 4F, DHT could partially block the galeterone-mediated decrease in AR protein.

Finally, as galeterone has been reported to induce an ER stress response that may indirectly affect AR expression (18), we addressed whether galeterone treatment in these cells was inducing an ER stress response. The ER stress response was assessed on the basis of phosphorylation of eukaryotic initiation factor 2α (eIF2α). Consistent with a previous report (18), an ER stress response could be induced by galeterone, which we observed at 20 μmol/L (Fig. 4G). However, galeterone at 10 μmol/L, which still markedly reduced AR protein, did not induce this stress response. These results together indicate that galeterone is directly enhancing AR protein degradation in LNCaP and C4-2 cells.

The marked effects of galeterone on AR protein expression in the C4-2 and LNCaP cells could reflect their T878A-mutant AR or other distinct features of these cell lines such as their PTEN deficiency. To initially address the role of the T878A mutation, LNCaP cells were transfected with a construct coding for FLAG-tagged WT AR and geneticin resistance, and short-term geneticin-resistant cells were selected. These cells were then androgen-deprived for 48 hours and treated with galeterone for 24 hours at increasing doses. Expression of ectopic WT AR was probed using anti-FLAG antibody. No decrease in the expression of ectopic WT AR was observed at up to 10 μmol/L galeterone, whereas the endogenous T878A AR protein in the parental LNCaP cells was decreased (Fig. 5A).

To rule out the possible confounding effect of the ectopic expression vector versus the endogenous AR, we established PC3 stable cells using lentiviral vectors expressing AR WT or AR T878A. Treatment of these stable PC3 cells with galeterone led to about 30% decrease in AR WT, but the decrease in AR T878A was substantially greater (∼70%; Fig. 5B). To further validate this result, LNCaP stable cells expressing epitope-tagged AR WT or T878A were also established. Similar to the PC3 stable cells, galeterone substantially decreased AR T878A expression, with minimal effect on AR WT levels (Fig. 5C).

The T878A AR can be strongly stimulated by the AR antagonists hydroxyflutamide and nilutamide. Therefore, we also tested whether galeterone enhances degradation of AR with other known mutations that can mediate resistance to AR antagonists such as enzalutamide or bicalutamide. Recent reports have found that enzalutamide can activate an F877L-mutant AR (27–29). However, in contrast to the transfected T878A AR, galeterone did not decrease expression of the F877L-mutant AR (Fig. 5D). Similarly, we found that expression of the W742C-mutant AR, which is stimulated by bicalutamide, was not decreased by galeterone (Fig. 5E). Taken together, these results show that galeterone markedly and directly enhances degradation of the T878A-mutant AR while having only modest effects on expression of WT AR that may be cell line–dependent and through direct or indirect mechanisms.

Galeterone suppresses AR activity both by functioning as a CYP17A1 inhibitor and by binding directly to AR (15, 16). It has clear activity against AR-dependent prostate cancer cells in vitro and in xenograft models, and its efficacy in patients is currently being assessed in phase II clinical trials. However, the molecular basis for its direct effects on AR remains to be fully established. In this study, we confirmed that galeterone functions as a direct AR competitive antagonist. Moreover, we found that galeterone, in contrast to bicalutamide but similar to enzalutamide, does not stimulate AR binding to chromatin. Previous studies have shown that bicalutamide functions as an AR antagonist because it does not effectively recruit coactivator proteins and may instead recruit corepressors (12, 30). However, this antagonist activity may be circumvented in CRPC through several mechanisms, such as increased AR or alterations in transcriptional coactivator or corepressor proteins, which may enhance the weak partial agonist activity of bicalutamide (14). In contrast, enzalutamide has a distinct mechanism of action as it does not stimulate AR binding to chromatin, which makes it a purer antagonist and likely contributes to its efficacy in CRPC (10). Therefore, our findings suggest that galeterone will function similar to enzalutamide in CRPC as a relatively pure direct AR antagonist, which in conjunction with its activity as a CYP17A1 inhibitor may enhance the ability of galeterone to suppress AR function in CRPC.

Previous studies have also shown that galeterone treatment can decrease AR protein levels and may increase AR protein degradation (16). However, it has not been clear whether this is due to direct effects on AR, as galeterone at higher concentrations (∼20 μmol/L) has been found to induce an ER stress response (18) and to decrease phosphorylation of 4EBP1, possibly reflecting decreased TORC1 activity (19). Using galeterone at lower concentrations (up to 10 μmol/L) that could effectively suppress DHT-stimulated AR activity, but did not stimulate a stress response, we found that galeterone did not substantially enhance degradation of WT AR in androgen-starved LAPC4 or VCaP prostate cancer cells. In contrast, degradation of the T878A-mutant AR, which is expressed by LNCaP and C4-2 cells, was markedly increased by a proteasome-dependent mechanism in response to galeterone binding.

These findings are consistent with previous data showing that galeterone at 5 to 10 μmol/L could decrease AR protein expression in LNCaP cells in vitro (16). While galeterone has also been found previously to decrease AR protein in LAPC4 cells in vitro, this effect was observed at higher galeterone concentrations (15–20 μmol/L) that may have decreased AR expression by indirect mechanisms (16). Significantly, in vivo studies using LAPC4 xenografts have shown marked decreases in AR protein in response to galeterone (16, 17). Our in vitro data certainly do not rule out the possibility that galeterone binding may directly enhance AR degradation in vivo due to displacement of an endogenous ligand and/or subtle differences in mechanisms regulating AR degradation in vivo versus in vitro. Alternatively, the decreased AR protein in vivo may be mediated by indirect mechanisms or possibly synergy between direct and indirect mechanisms. Finally, a recent study found that galeterone and related compounds could enhance degradation of the full-length AR and truncated AR (ligand-binding domain deleted) in CWR22Rv1 cells, further establishing that galeterone has a mechanism of action that is independent of binding to the AR ligand-binding domain (31).

Structural studies have shown that T878 in helix 9 of the AR ligand-binding domain sits in the steroid-binding pocket and that the T878A mutation provides more room in this pocket (32, 33). As a result, binding of the AR antagonist hydroxyflutamide to the T878A-mutant AR does not distort the ligand-binding pocket and this drug thereby functions as potent agonist of the T878A AR (34). Galeterone is a steroidal compound modified with a bulky benzimidazole group at the C17 position, which in DHT has a hydroxyl group that contacts T878. However, as galeterone is still an antagonist for the T878A AR, it is apparent that the benzimidazole group must still distort the ligand-binding pocket. Nonetheless, we propose that the T878A mutation alters the positioning of helix 9 in the galeterone-liganded AR and may thereby enhance the ubiquitylation of one or more helix 9 lysines by an E3 ubiquitin ligase that normally mediates degradation of the unliganded AR or that mediates turnover of the agonist-liganded AR. Alternatively, the galeterone-liganded T878A-mutant AR may be degraded through a distinct mechanism that does not normally mediate AR degradation. In either case, these findings demonstrate that certain conformations of AR can markedly enhance its degradation and support efforts to identify novel antagonists that can similarly alter the structure of WT AR and thereby enhance its degradation.

S.P. Balk reports receiving a commercial research grant from TOKAI. No potential conflicts of interest were disclosed by the other authors.

Conception and design: Z. Yu, C. Cai, S. Gao, S.P. Balk

Development of methodology: Z. Yu, C. Cai, H.C. Shen, S.P. Balk

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Z. Yu, N.I. Simon, H.C. Shen, S.P. Balk

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Z. Yu, C. Cai, S. Gao, S.P. Balk

Writing, review, and/or revision of the manuscript: Z. Yu, C. Cai, S. Gao, S.P. Balk

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Z. Yu, S.P. Balk

Study supervision: S.P. Balk

The authors thank Tokai Pharmaceuticals for providing galeterone and Douglas Jacoby (Tokai Pharmaceuticals) for helpful suggestions.

This study was supported by a Department of Defense Prostate Cancer Research Program Postdoctoral Training Award (W81XWH-12-PCRP-PTA to Z. Yu) and Idea Development Awards (W81XWH-11-1-0295, W81XWH-08-1-0414, and W81XWH07-1-0443 to S. Balk), DF/HCC-Prostate Cancer SPORE P50 CA090381, and NIH K99 award CA166507 (C. Cai), NIH P01 CA163227-01A1 (S. Balk), a Prostate Cancer Foundation Challenge Award, and by a sponsored research grant from Tokai.

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.