Specific Co-Activator for Human Androgen Receptor

The present invention relates to a cloned gene for a protein which co-activates an important hormonal receptor in humans and relates, in addition, to the use of the co-activator protein as an important constituent in clinical tests for diagnoses of human clinical conditions.

The class of compounds known as androgens are the hormonal signals responsible for maleness in mammals in general and human beings in particular. As with most hormonal signals, androgens interact with their targets by binding to a receptor, known as the androgen receptor. Recognition of androgens by the androgen receptor initiates a series of transcriptional events giving rise to male-associated processes in certain tissues and organs. The binding of androgens to the androgen receptor is also important in many androgen related diseases and conditions, such as baldness and acne, as well as important clinical diseases such as prostate cancer. The androgen receptor belongs to the steroid receptor super family that plays an important role in male sexual differentiation and prostate cell proliferation. Mutations or abnormal expressions of the androgen receptor in prostate cells may play an important role in the progression of prostate cancer.

When bound to androgens and androgen responsive elements, the androgen receptor can up-regulate or down-regulate the expression of androgen target genes through a complicated process that may involve multiple adaptors or co-activators. Adler et al., Proc. Natl. Acad. Sci. USA 89, 6319-6325 (1992). A fundamental issue in the field of steroid hormone regulation is the question or how specific androgen-activated transcription can be achieved in vivo when several different receptors recognize the same DNA sequence. For example, the androgen receptor (AR), the glucocorticoid receptor (GR) and the progesterone receptor (PR) all recognize the same sequence but activate different transcription activities. It has been speculated by some that accessory factors may selectively interact with the androgen receptor to determine the specificity of the androgen receptor target gene activation.

One of the uses for the androgen receptor is to detect the androgenic or anti-androgenic effects of specific candidate human pharmaceutical molecules. The androgenic effect of pharmaceuticals is usually an attribute of potential candidate therapeutic medicines that must be evaluated during the process of total evaluation of a molecule for human therapeutic value. Accordingly, the androgen receptor is used in screens to determine the frequency and specificity by which specific molecules bind to such receptors.

The present invention is summarized in that a specific co-activator for the human androgen receptor has been isolated and the gene for that co-activator has been cloned sequenced and is presented below.

The present invention is also summarized in that the cloning and reproduction of the androgen receptor activator gene permits new laboratory tests to be made to test the androgen specificity of candidate therapeutic molecules.

Other objects, advantages, and features of the present invention will become apparent from the following specification.

FIG. 1 is a schematic illustration of the use of the yeast two-hybrid system as used to identify ARA.sub.70.

The present invention is enabled by a discovery of a new regulatory protein in humans. This regulatory protein is the androgen receptor associated protein, here designated ARA.sub.70, which is a co-activator for the androgen receptor in human prostate cells. The ARA.sub.70 factor is a ligand dependent protein that functions as a specific co-activator to enhance the transcriptional effect of androgen binding to the androgen receptor and also facilitates binding and activation of the androgen receptor by molecules previously not thought to have androgenic character.

Using a yeast two hybrid system, as described below, a cDNA encoding the ARA.sub.70 molecule has been recovered from human brain cells. The recovered ARA.sub.70 cDNA encodes a protein of 614 amino acids, with a calculated molecular weight of 70 kilodaltons. The full length cDNA has fully been sequenced, and the sequence is presented as SEQ ID NO 1 below. A search of the GenBank indicates that the ARA.sub.70 cDNA shares a high degree of homology (99%) with a previously identified cDNA clone (RET-fused gene RFG), isolated from human thyroid as reported by Santoro et al., Oncogene 9, 509-516 (1994). Santoro et al. were unable to identify the main biological functions of the protein designated RFG, although the expression of the RFG in thyroid tumor suggests a potential role for the RFG molecule in thyroid carcinogenesis.

Northern Blot analysis indicated that the ARA.sub.70 co-activator transcript is present in many tissues, including prostate, testis, adrenal gland, and thymus. Most human cell lines tested positive for the ARA.sub.70 co-factor, with the significant exception of a human prostate cancer DU145 cell line, which did not express the ARA.sub.70 molecule.

A specific ligand is necessary to actuate the co-activator role of ARA.sub.70 as an enhancer of androgen receptor transcriptional activity. The most potent ligand yet identified is dihydrotestosterone (DHT). Using the yeast two hybrid model system, it has been demonstrated that the ARA.sub.70 molecule will enhance the transcriptional activity of the androgen receptor near 10 fold, as measured in the presence of 10.sup.-10 M DHT. Furthermore, as described in greater detail below, the transcriptional activity of AR was activated by ARA.sub.70 in the presence of 10.sup.-8 M 17.beta.-estradiol (E2) in human prostate cells, but did not have the same enhancement of transcriptional activity in the presence of 10.sup.-6 M diethylstilbestrol (DES) an estrogen thought to be more potent. This data suggests that co-activators, such as ARA.sub.70, for androgen receptor activity can mediate transcriptional activation of molecules previously thought to be essentially non-androgenic in a manner not previously detectable.

The availability of the ARA.sub.70 cDNA clone described below enables the production of ARA.sub.70 in foreign hosts. By joining the ARA.sub.70 coding region to a promoter effective to initiate transcriptional activity in a desired host, whether eukaryotic or prokaryotic cells, quantities of ARA.sub.70 can be manufactured in a foreign host for the uses described here and for other uses.

It is specifically envisioned that ARA.sub.70 will have particular use as a constituent in a drug testing or screening protocol. It is a general practice in the evaluation of new clinical compounds for pharmaceutical utility that the compounds be tested for androgenic activity. Androgenic or antiandrogenic side effects can be important in the administration of some pharmaceutical agents. Previously, one of the methods used to test androgen activity was testing for binding and activation of the androgen receptor transcriptional activity. As the data herein suggests, the presence of ARA.sub.70 in the presence of the androgen receptor greatly alters both the magnitude and the specificity of the transcriptional effect of the androgen receptor elicited by specific androgens. In addition, as evidenced with the estrogen E2 indicates, some molecules previously thought not to have androgenic activity will, in the presence of ARA.sub.70, initiate transcriptional activity when bound to the androgen receptor and some molecules previously thought not to have inhibitory effect will limit or oppose the activity of the androgen receptor activated by ARA.sub.70. Accordingly, in testing potential pharmaceutical molecules for androgenic or antiandrogenic effect, it would be important to include ARA.sub.70 in the assay for androgenic/antiandrogenic activity to fully test androgenic effects actuated by the candidate molecule in vivo.

It is also anticipated that ARA.sub.70 will serve as a clinical indicator of significant important for androgen related diseases. Significant androgen related diseases, such as prostate cancer, baldness, acne, and androgen insensitive syndromes, such as Tfm syndrome, may be due to defects in the co-activation mechanism between the androgen receptor and the ARA.sub.70 molecule. Accordingly, it becomes a reasonable possibility, given the data presented herein, to assay the relative ratios of AR and ARA.sub.70 in patients with such conditions. Such ratios may be measured by raising antibodies to both ARA.sub.70 and to AR in performing quantitative methods to adjudge the relative quantity of the two molecules in a particular patient. Several methods exist for measuring such comparative ratios, including radio immunoassay, ELISA, immunostaining, or Western Blot. In addition, it would be possible to use the ARA.sub.70 cDNA so as to construct probes for PCR assay for the presence of mutations of the normal DNA sequence in particular patients, or to generate transcript for Northern Blot assay or DNA for in situ hybridization assays.

The theory for such measurements of relative ratios AR to ARA.sub.70 is that androgen insensitive related disease may be due to an imbalance between androgen receptor and androgen ARA.sub.70 prevalence in target cells. Too much ARA.sub.70 might over-sensitize the androgen receptor system, so as to be responsive to molecules not intended to have androgenic effect. Under sensitivity due to absence or non-function of ARA.sub.70 may lead to androgen insensitivity at any levels. If too much ARA.sub.70 was found to be present in a particular patient, that would suggest the use of down regulation mechanisms such as antisense or other similar mechanisms, in clinical system so as to reduce the levels of ARA.sub.70 prevalent in a particular patient. If a particular patient had too little ARA.sub.70, then it would be possible to deliver ARA.sub.70 cDNA, protein, or DNA, into a patient by a variety of delivery mechanisms to increase levels of active ARA.sub.70 in the patient.

In addition to testing potential pharmaceutical uses, the ARA.sub.70 molecule would be useful for testing non-pharmaceutical compounds for potential androgenic/antiandrogenic activity. It is currently believed that many contaminants present within the environment at low samples have androgenic/antiandrogenic or estrogenic/antiestrogenic activity on various parts of the population. Since the ARA.sub.70 increases androgen receptor specificity by over 10 fold, the sensitivity of androgen receptor tests can be greatly enhanced by the use of ARA.sub.70 in such assay systems. As demonstrated by the fact that the addition of ARA.sub.70 causes compounds classically thought to be only estrogenic, such as 17.beta. estradiol, to exhibit androgenic activity, and by the fact that compounds thought to be only antiestrogenic, such as tamoxifen, can exhibit antiandrogenic activity, tests for androgenic/antiandrogenic activity would be incomplete without the use of ARA.sub.70 as a co-factor in such reactions.

To test samples for androgenic/antiandrogenic activity, genetic constructions including expression cassettes for both the androgen receptor and ARA.sub.70 would be transformed into host cells, such as a prostate cell line, in vitro. Also an easily detectable and quantifiable detector gene would be transformed in the cells as well. A suitable detector gene would be chloramphenicol acetyltransferase, or CAT, or luciferase the expression of which can be detected photometrically. The cells are then exposed to the pharmaceutical agent or environmental sample. Samples with androgenic/antiandrogenic activity will actuate increased or decreased detectable levels of CAT or luciferase activity.

Identification of the Androgen Receptor Specific-Associated Protein, ARA.sub.70. To understand the mechanism of androgen-AR action, a yeast two-hybrid system, using the GAL4AR fusion protein as bait, was used to isolate a cDNA encoding ARA.sub.70 which interacts specifically with AR. The fusion protein GAL4AR contains the GAL4 DNA binding domain (GAL4DBD) fused to the C-terminus of the androgen receptor. The fusion protein was used to screen for His-synthase gene positive clones from 3.times.10.sup.6 transformants of the MATCHMAKER human brain library. Two of the initial 41 putatively positive clones clearly reacted with the AR fusion protein, by liquid assays performed by the method of Durfee et al. Genes & Dev. 7, 555-569 (1993).

In this yeast two-hybrid system, illustrated schematically in FIG. 1, yeast will survive when GAL4AR is co-expressed with ARA.sub.70 in the presence of DHT. Neither GAL4AR nor ARA.sub.70 was active when ARA.sub.70 was expressed alone or when ARA.sub.70 was co-expressed with GAL4RAR or GAL4TR4, Chang et al. Proc. Natl. Acad. Sci. USA, 91, 6040-6044 (1994), (GAL4 fusion proteins with two other members of the steroid receptor superfamily). These data, therefore, clearly suggest that ARA.sub.70 can interact specifically with AR in the yeast cells.

We then tested whether the interaction of ARA.sub.70 with AR in yeast was ligand-dependent. It was found that DHT (5.times.10.sup.-10 M) can promote the interaction between ARA.sub.70 and GAL4AR. Testosterone (T), a less potent androgen in the prostate, can also promote this interaction at higher concentrations (5.times.10.sup.-9 M). Hydroxyflutamide (HF), an antiandrogen used in the treatment of prostate cancer, had no activity even at very high concentrations (10.sup.-5 M).

The RACE-PCR technique (10,11) was then used to clone the full-length ARA.sub.70 cDNA, encoding a protein of 615 amino acids with a calculated molecular weight of 70K, (SEQ ID NO 1 & 2). A search of GenBank indicated that ARA.sub.70 shares 99% homology (three different amino acids in the coding region) with one identified cDNA clone (RET-fused gene, RFG) isolated from human thyroid. Although the biological functions of RFG are mostly unknown, the expression of RFG in thyroid tumor may suggest some potential roles of RFG in thyroid carcinogenesis.

The Tissue Distribution of ARA.sub.70. Northern blot analysis in mouse indicated that ARA.sub.70 is expressed as an mRNA of -3600 bp in many tissues, including prostate, testis, adrenal gland, and thymus. The relative expression of ARA.sub.70 in the following mouse tissues, using adrenal gland as 100%, are: testis, 77%; prostate, 97%; preputial gland, 64%, thymus, 214%; submaxillary gland, 24%; muscle, 41%, heart, 73%; kidney, 37%; lung, 49%; fat pad, 20%; seminal vesicle and spleen undetectable. Among the cell lines (LNCaP, MCF-7, CHO, HeLa and DU145) tested, the human prostate cancer cell line, DU145, proved to be the only cell line that did not express ARA.sub.70, and therefore was chosen for further functional study.

The In Vitro Interaction Between AR and ARA.sub.70. To further confirm that the interaction that occurred in yeast cells is due to a direct interaction between AR and ARA.sub.70, we applied an in vitro immunoprecipitation assay with an anti-AR antibody designated CW2. We demonstrated that CW2 can co-precipitate the AR and ARA.sub.70 when in vitro transcribed/translated full-length human AR and ARA.sub.70 were incubated with it in a lysate mixture. This precipitation is specific, as CW2 did not precipitate the ARA.sub.70 in the absence of AR and CW2 did not precipitate two other proteins (RXR and TR4 orphan receptors) incubated with AR. A Far-Western assay also demonstrated that ARA.sub.70 can bind to immobilized AR peptide containing DNA binding domain and hormone binding domain (AR-DBD/HBD), but not the BL21 protein lysate or the AR peptide containing the N-terminal and DNA binding domain of AR (AR-N/DBD). This data indicates that the association is due to a direct interaction between AR and ARA.sub.70.

To perform the Far-Western assay, AR-N/DBD and AR-DBD/HBD were expressed, as polyhistidine fusion proteins by inserting the N-terminal or C-terminal fragments into pET 14b (Novagen). Proteins were separated on 10% polyacrylamide gel. .sup.35 S-labeled ARA.sub.70 was diluted into hybridization buffer and the titers were hybridized overnight in the presence of 1 .mu.M DHT. After three washings, filters were dried and autoradiographs made.

Stimulation of the Transcriptional Activity of AR by ARA.sub.70. DU145 cells were co-transfected with ARA.sub.70 and AR under the control of a eukaryotic promoter. Ligand-free AR was found to have minimal MMTV-ARE CAT reporter activity, with or without the presence of ARA.sub.70. Addition of DHT resulted in a 6-fold increase of AR activity. This transcriptional activity was increased 58 (.+-.3.2)-fold (mean.+-.SEM; n=4) by the co-transfection of ARA.sub.70 cDNAs in a dose-dependent manner. The induced activity reached a plateau at 4.5 .mu.g of co-transfected ARA.sub.70 cDNA. Additional ARA.sub.70, beyond 4.5 .mu.g, (up to 6 .mu.g) did not affect the induced activity of AR in DU145 cells. To rule out any indirect effects on the basal activity of the MMTV-ARE CAT reporter, we removed the ARE DNA fragment from the reporter (MMTV-.DELTA.ARE-CAT). The results showed that ARA.sub.70 induced no activity on this reporter in the presence or in the absence of DHT.

We also replaced ARA.sub.70 with another nuclear orphan receptor-associated protein, TR4AP, in the AR: MMTV-ARE CAT reporter assay and found this protein had no effect in our assay. Furthermore, when we replaced DU145 cells with CHO cells, which express a relative abundance of ARA.sub.70, we found that although the exogenously transfected ARA.sub.70 did not show a dramatic effect on induction of AR transcriptional activity, the transfection of antisense ARA.sub.70 did partially block the AR transcriptional activity. Together, these data strongly suggest that stimulation of AR transcriptional activity by ARA.sub.70 occurs through a specific ligand-bound AR and the relative amount of AR vs ARA.sub.70 in cells plays an important role for the activation of AR.

The effect of ARA.sub.70 on transactivation of AR bound to different concentrations of testosterone (T), dihydrotestorenone (DHT) and hydroxy flutamide (HF) in DU145 cells was also tested. Whereas 10.sup.-10 M DHT maximized induced transcriptional activity of AR, with T a 10-fold higher concentration (10.sup.-9 M) was needed for maximum activity. HF induced very low at a pharmacological concentration (10.sup.-6 M). These results are consistent with the data generated from yeast cells and previous reports, which indicated DHT is more potent androgen in the prostate. In fact, the greater potency of DHT to modulate the interaction between AR and ARA.sub.70 may actually provide the reason why DHT is a more potent androgen in prostate.

The enhancement of AR transcriptional activity from 6-fold to 58-fold by ARA.sub.70 may explain androgen activity in the prostate that androgen-AR alone cannot explain. Since we detected ARA.sub.70 in AR-positive LNCaP prostate cancer cells, but not in AR-negative DU145 cells, it will be important to determine if the expression of ARA.sub.70 and its ability to interact properly with androgen-AR changes during the progression of prostate cancer from an androgen-dependent to an androgen-independent state.

ARA.sub.70 Functions As a Specific Activator to Enhance the Transcriptional Activity of AR. We also examined the effect of ARA.sub.70 on the transcriptional activity of several other steroid receptors through their cognate DNA response elements. While ARA.sub.70 induces the transcriptional activity of AR up to 10-fold, ARA.sub.70 can only slightly enhance (up to 2-fold) the transcription activity of other steroid receptors, such as GR, PR, and ER. These results clearly indicate that ARA.sub.7. is a very specific co-activator for AR.

Several proteins have been demonstrated to interact with other steroid receptors in a ligand-dependent or ligand-independent manner. However, none of these proteins have been shown to enhance specifically AR-mediated transcriptional activity; therefore, it is likely that ARA.sub.70 has a different mechanism for interacting with AR.

In summary, our data demonstrated that ARA.sub.70 is the first identified ligand-dependent associated protein for AR which may function as a specific co-activator for inducing the transcriptional activity of AR in human prostate cells. Further studying the potential role of ARA.sub.70 may therefore help us to understand better the molecular mechanism of androgen action.

Transcriptional Activity of AR Induced by 17.beta.-estradiol

Tests in both DU145 cells and yeast cells demonstrated that 17.beta.-estradiol, at a concentration of 10.sup.-8 M or higher, stimulated the transcriptional activity of AR in the presence of ARA.sub.70. By contrast, diethylstilbestrol (DES), even at concentrations of 10.sup.-6 M, did not increase AR transcriptional activity. This result may explain why DES, but not 17.beta.-estradiol, has fewer side effects when used by clinicians to treat prostate cancer patients.

Antiandrogenic Activity of Tamoxifen and ICI.sub.162780

Similar protocols were repeated but, instead of adding an androgen or estrogen, tamoxifen and ICI.sub.162780 were added, both compounds known to be antiestrogenic. The data revealed that both compounds inhibited AR initiated transcriptional activity in human prostate cells. This demonstrates the ability to assay for antiandrogenic effects using this same style of assay.