MedicineLakex

Sladek-mh.fm

Meeting Highlights Nuclear Receptors as Drug
Targets: New Developments in
Coregulators, Orphan Receptors
and Major Therapeutic Areas
2. Why nuclear receptors are good 19 – 21 May 2003, Philadelphia, Pennsylvania, USA
3. New drugs for classical nuclear receptors University of California, Riverside, Riverside, CA 92521, USA 4. Orphan receptors as drug targets Nuclear receptors (NRs) are ideal targets for drug discovery. Not only do they 5. Use of X-ray structure analysis control a myriad of biological and disease processes, but they are also regu- to de-orphanise receptors lated by small lipophilic molecules that can be easily exchanged with a drug 6. Non ligand-based drugs of choice. All 48 of the NR genes in the human genome have been identified, 7. Other developments and many of their structures have been solved and their ligands identified. Their future directions mechanism of action has been elucidated and many of their target genes 8. Conclusion and expert opinion have been identified. Nonetheless, presentations at the recent conferencesponsored by IBC Life Sciences indicated that while many NRs already havemarketable drugs, the latest tools in robotics, genomics, proteomics, andinformatics are helping to identify more selective drugs.
Keywords: bioinformatics, cancer, cofactors, drug discovery, human disease, nuclear receptors
(NRs), partial agonists, promoter selectivity, structure–activity relationship
Expert Opin. Ther. Targets (2003) 7(5):
Nuclear receptors (NRs) constitute a large superfamily of ligand-dependent tran-
scription factors that are involved in nearly every aspect of vertebrate development
and are linked to a wide range of human diseases [1] (see Table 1). NRs bind spe-
cific DNA elements in the regulatory regions of genes via a highly-conserved
DNA-binding domain (DBD) and specific ligands via another highly-conserved
domain, the ligand-binding domain (LBD), which consists of a series of approxi-
mately 12 helices that form a hydrophobic pocket. The binding of ligands in the
pocket induces conformational changes in the receptor that affect the recruitment
of coregulatory molecules (cofactors) that stimulate (co-activators) or repress
(corepressors) transcription, typically via modification of the chromatin and inter-
action with the basal transcription machinery [2,3]. A detailed understanding of
this mechanism of action is what is driving drug discovery for NRs, as discussed in
the recent conference on ‘Nuclear Receptors: New Developments in Coregulators,
Orphan Receptors and Major Therapeutic Areas,' the highlights of which are
summarised in this report.
2. Why nuclear receptors are good drug targets
In addition to ligands that act as pure agonists and antagonists, NRs can also bindpartial (or mixed) agonists (or antagonists), which is what makes NRs really gooddrug targets. These compounds typically bind with a lower affinity than pure ago-nists and induce unique conformational changes in the receptor such that a spe- cific cohort of cofactors are recruited, resulting in certain biological responses, butnot others. The ability to selectively modulate the receptors has led to the search 2003 Ashley Publications Ltd ISSN 1472-8222 Nuclear receptors as drug targets: new developments in coregulators, orphan receptors and major therapeutic areas
Table 1. Nuclear receptors as drug targets.
Full name
Thyroid hormone receptor Hypothyroidism, obesity Retinoic acid receptor Inflammatory skin disorders, leukaemia Peroxisome proliferator-activated receptor Diabetes, coronary heart disease, obesity Retinoic acid-related orphan receptor Atherosclerosis, immunological disorders, neurological disorders, osteoporosis Farnesoid X receptor Dyslipidaemia, liver disease Vitamin D receptor Osteoporosis, calcium homeostasis, cancer prevention Pregnane X receptor Xenobiotic metabolism Constitutive androstane receptor Xenobiotic metabolism Hepatocyte nuclear factor 4 Diabetes, haemophilia, lipid metabolism Retinoid X receptor Leukaemia, coronary heart disease Oestrogen receptor Breast cancer, osteoporosis, atherosclerosis, CNS Early embryo development Glucocortioic receptor Immunological disorders, metabolic disorders Hypertension, myocardial hypertrophy Progesterone receptor Breast cancer, infertility, pregnancy maintenance Androgen receptor Prostate cancer, X-linked androgen insensitivity, spinal/muscular atrophy Nerve growth factor-induced-B Neurological disorders, immunological disorders, cancer Germ cell nuclear factor Bold: NRs with marketed drugs.
Sampling of NRs as drug targets ordered by subfamilies, modified from the information presented by T Burris (Eli Lilly).
NR: Nuclear receptor.
for selective oestrogen receptor (ER) modulators (SERMs), allowing a quick, and relatively inexpensive, classification of selective androgen receptor (AR) modulators (SARMs), the test compound.
selective liver X receptor (LXR) modulators (SeLRMs), The other key to developing SNuRMs is promoter selectiv- selective peroxisome proliferator-activated receptor (PPAR) ity or predictivity, i.e., finding a drug that will activate a given modulators (SPARMs) etc. (globally referred to as selective NR on certain target genes but not others. Much of the selec- NR modulators [SNuRMs]). For example, the ideal SERM tivity is thought to be derived from the fact that, like ligands, would activate ER in the bone to fight osteoporosis, but not the specific DNA sequence of a response element can also in breast or endometrial tissue, which might lead to cancer.
induce allosteric changes in the NR [5]. Although much less M Koegl (PheneX Pharmaceuticals AG, Heidelberg, Ger- well-documented than ligand-induced changes, this notion of many) noted that the key to identifying SNuRMs is to iden- DNA sequence affecting the profile of cofactors recruited by a tify all of the cofactors that bind NRs and to determine their NR to a given promoter permeated the talks at the conference tissue distribution and expression profile during develop- and is expected to be broadly applicable to all NRs. Specific ment. To this end, his company has screened the LBDs of all examples were provided by B Haendler (Schering AG, Berlin, NRs using the yeast two-hybrid system. They found typi- Germany) and L Freedman (Merck, West Point, PA, USA) for cally 50 – 100 different interacting proteins for each NR, the AR and vitamin D receptor (VDR), respectively. The with certain cofactors, such as steroid receptor co-activator author's and colleagues' own findings on the target genes of (SRC)-1, binding nearly all NRs, whilst others, such as per- orphan receptor HNF4 also indicate that there are at least oxisome proliferator-activated receptor-γ co-activator 165 unique DNA sequences to which HNF4 binds in the (PGC)-1, prefer just a few. This finding was also reflected in human genome, indicating an incredible amount of heteroge- a computer-assisted review of the published literature [4].
neity in response elements and therefore a great potential for The concept of each ligand-bound NR recruiting a different gene-specific responses. All told, combining the ability of profile of cofactors is also the basis of the Molecular Braille both partial agonists and specific DNA response elements to Technology presented by R Carlson and R Evans-Storm induce specific conformational changes in NRs enhances tre- (Karo Bio, Durham, NC, USA). This powerful new tech- mendously the possibility of finding a drug that has the nique uses peptides identified in phage display to detect appropriate desired effects without negative side effects.
changes in NR surface conformation in response to ligandbinding. The binding by a set of 4 – 16 peptides to ER, for 3. New drugs for classical nuclear receptors
example, can determine whether a test compound induces aconformational change similar to that of given reference Whereas most of the classical NRs have had drugs for some compounds, such as oestradiol or tamoxifen, thereby time, nearly all of those drugs have unwanted side effects. For Expert Opin. Ther. Targets (2003) 7(5)
example, glucocortioic receptor (GR) has long been a target macrophages. However, while LXR agonists can reduce for anti-inflammatory drugs but chronic use leads to bone plaque formation associated with atherosclerosis, they also loss, diabetes, myopathy, and hypertension. To avoid these tend to increase triglyceride levels and induce adipogenesis, and other negative effects of GR drugs, L Buckbinder (Pfizer, both of which are pro-atherogenic. D Lala (Pharmacia, Groton, CT, USA) screened for dissociated agonists (DAGRs) St Louis, MO, USA) identified a SeLRM (PHA-769956) that that regulate certain anti-inflammatory genes, but not others, increases the activity of LXRα but not that of LXRβ. Whereas and found two compounds (DAGR1 and DAGR2) that do it is much less potent than a previously identified LXR ligand not promote the differentiation of osteoblasts nor stimulate (T-0901317), it does not increase the expression of SREBP1c gluconeogenic enzymes in cell-based assays. J Baxter (Univer- or FAS, both of which are associated with the negative effects sity of California San Francisco, San Francisco, CA, USA) of LXR agonists. R Heyman (X-Ceptor Therapeutics, San reported on a new mode of action for thyroid hormone recep- Diego, CA, USA) presented in vivo data about another excit- tor (TR) agonists. Like previous agonists, these compounds ing SeLRM (XCT-628) that decreased the arterial lesions in a (GC-24II, GC-1, KB-000141) decrease obesity, cholesterol mouse model of atherosclerosis by 50%. The compound not and triglyceride levels. However, unlike traditional agonists, only prevented the lesions from getting worse, it actually they bind not in the ligand-binding pocket but in the dimer made them regress from the baseline, something that has interface and seem to have fewer unwanted effects compared never before been seen with the current drugs on the market.
with traditional TR agonists, such as tachycardia, arrhythmia A decrease in macrophage foam cell content and an increase and heart failure. Freedman (Merck) also presented data on in plaque stability were also observed. T Burris (Eli Lilly, Indi- 20-epi VitD3 analogues (MC-1627 and MC-1288) that may anapolis, IN, USA) presented a potential new role for LXR in avoid the extreme hypercalcaemic effects of the more tradi- diabetes, showing that it repressed the expression of gluconeo- tional 1,25-dihydroxy VitD3 by causing VDR to selectively genic enzymes such as PEPCK and G6P in the liver. He also bind certain co-activators. showed that LXR and PPARα shared the same ‘chemical T Mirzadegan (Roche Biosciences, Palo Alto, CA, USA) space', in that they both bind fenofibrates, although LXR pre- presented an elegant example of how computational chemis- fers the ester form, which act as antagonists, whilst PPARα try can be used to design a drug specific to a given NR iso- prefers the acid form, which act as agonists. Therefore, it form. Mirzadegan and colleagues compared the structure of appears that fenofibrates that are used clinically to lower trig- RARγ, which is implicated in emphysema, to that of RARα lyceride levels may be acting via LXR, in addition to PPARα.
and RARβ and used molecular dynamic simulations This would make fenofibrates the first clinically used com- (SANDER, FlexX and Locally Enhanced Sampling [LES]) to pound to act via LXR. identify first a unique residue in RARγ that contacts the lig-and and then a new ligand specific for RARγ. Ro-3300074 5. Use of X-ray structure analysis to
has a different ligand scaffold and fewer side effects than the de-orphanise receptors
traditional all-trans-retinoic acid. It promotes the regenerationof lung tissue, decreases emphysema and restores lung func- Another common theme of the conference was the use of tion. and is soon to enter Phase II trials.
X-ray crystallographic structure analysis and mass spectometry Finally, B Cheskis (Wyeth Women's Health Research Insti- to de-orphanise receptors and, in conjunction with molecular tute, Collegeville, PA, USA) presented exciting data on a modelling, to rationally design drugs to increase the specificity potentially new drug target that links the ER to another well- or affinity of known ligands. M Geiser (Novartis Biomedical characterised family of regulators, the Src kinases. Cheskis Research Institute, Basel, Switzerland) presented the structure identified modulator of non-genomic activity of oestrogen of the LBD of RORα which, when expressed in insect cells, receptor (MNAR) in a liver cancer cell line, which increases was found to contain cholesterol in the ligand-binding pocket the ability of ER to activate transcription by forming a com- [8]. Additional space in the pocket accommodated derivatives plex with ER and Src via ten LXXLL motifs and three PXXP of cholesterol such as cholesterol sulfate, epicholesterol and (SH3) motifs, respectively. MNAR is found in many tissues 7-dehydroxycholesterol, although it is not yet clear whether but it is amplified in breast cancer cells [6].
the cholesterol acts as a structural cofactor or as a traditionalligand that induces an activated form of the receptor. None- 4. Orphan receptors as drug targets
theless, since RORα has been linked to bone metabolism, thisraises the issue of whether drugs used to treat high cholesterol, Much attention at the conference was paid to several of the such as statins, might also increase osteoporosis. receptors that are involved in intermediary metabolism and Another orphan receptor also linked to lipid metabolism for which ligands have recently been identified [7]. Chief and diabetes, HNF4, was similarly de-orphanised by T Will- among these was LXR, which binds oxysterols. It plays a role son and colleagues (GlaxoSmithKline, Research Triangle Park, in lipid transport by increasing cholesterol efflux via activa- NC, USA) when the structure of its LBD was solved [9,10]. In tion of ABCA1, Cyp7A and ApoE genes, and decreases its ligand-binding pocket were a mixture of fatty acids but, inflammation by inhibiting cytokine gene expression in surprisingly, the fatty acids could not be removed from the Expert Opin. Ther. Targets (2003) 7(5)
Nuclear receptors as drug targets: new developments in coregulators, orphan receptors and major therapeutic areas
native protein, suggesting that HNF4 may not be a good drug xenobiotics and drug metabolism. Not only do NRs such as target. Upon crystallisation of the Drosophila orthologue of HNF4, LXR, FXR, PPAR and RXR play a pivotal role in RXR, ultraspiracle protein (USP), D Moras (IGBMC, regulating the expression of the Phase I and II genes that regu- Illkirch, France) found that it also unexpectedly, and late xenobiotic and drug metabolism, NRs PXR and constitu- irreversibly, bound phospholipids [11]. Moras went on to show tive androstane receptor (CAR) do as well, and they respond that not all NR LBDs expressed in heterologous systems bind directly to many of those same compounds. JT Moore (Glax- ligands, which may or may not be fortuitous. The agonist oSmithKline, Research Triangle Park, NC, USA) and M Red- structure of the ERRγ LBD revealed an empty ligand-binding inbo (University of North Carolina, Chapel Hill, NC, USA) pocket, consistent with the fact that known ERRγ ligands, both discussed the role of PXR in drug–drug interactions.
such as 4-hydroxy tamoxifen (4-OHT) and diethylstilbesterol Hyperforin, the active ingredient in St John's Wort, which is (DES), are antagonists and are not found in bacteria [12].
popularly used to ward off mild depression, is a ligand forPXR and stimulates its activity on the Cyp3A gene, which 6. Non ligand-based drugs
metabolises the majority of prescription drugs [17]. Therefore,transplant patients taking cyclosporin for immune suppres- The structure of yet another NR LBD showed that not all NRs sion or women taking ethyloestradiol for birth control can have ligand-binding pockets. Willson's group showed that the drastically affect the metabolism of their medication if they space where one would expect to see the pocket in NGFI-B and also take St John's Wort. Since PXR has a very large ligand- its Drosophila orthologue, DHR-38, was filled with phenyla- binding pocket (1300 Å3) it can accommodate many different lanines, leaving a volume of < 30 Å3 compared to a typical vol- types of compounds, such as the antibiotic rifampacin and the ume of ∼300 Å3 for receptors that have ligands [13]. Since phyto-oestrogen coumestrol found in soy products, alfalfa and NGFI-B is linked to manic depression and schizophrenia, this infant formula, increasing greatly the potential for unwanted finding, like that of the non-exchangeable ligands, is very disap- drug interactions. Similar effects are possible with CAR, pointing. Nonetheless, P Ordentlich (X-Ceptor Therapeutics, which is stimulated by phenobarbital, although its mecha- San Diego, CA, USA) showed that the activity of the related nism of action is different from that of PXR (Moore; M Negi- Nurr1, which also plays a role in the CNS as well as in immu- shi [NIEHS, Research Triangle Park, NC, USA]) [18].
nity and cancer, could be modulated by something other than a The conference also discussed the issue of how to deal with ligand. Upon screening a library of 500,000 compounds, they the deluge of data coming not just from genomics experi- found one compound, 6-mercaptopurine (6-MP), that stimu- ments with DNA microarrays containing tens of thousands of lated the transcriptional activity of Nurr1 via the N-terminal genes, but also from automated cell- and in vitro-based assays portion of the protein, far from the LBD [14,15]. 6-MP has been that spew out hundreds of thousands of data points per assay used since the 1950s to treat leukaemia and more recently to per day. As PY Yim (Rosetta Inpharmatics, Kirkland, WA, treat inflammatory diseases such as Crohn's disease.
USA) pointed out in the Pre-Conference Workshop on the Finally, one of the most exciting new drugs that was pre- topic, there are programmes to keep track of experiments, to sented at the conference does not even have a NR as a target, pull out specific data points upon request and to visualise although its identification was made possible by research on 500,000 data points in one presentation (e.g., Spotfire). This NRs. Activated NRs recruit co-activators which often contain allows for mega analyses across different doses, time points, histone acetylase transferase (HAT) activity, while inactive NRs tissues, and species as well as compounds that can provide form complexes with corepressors which recruit histone new insights into mechanisms that cannot be gained by more deacetylase (HDAC) activity. V Richon (Aton Pharmaceuti- traditional methods of experimentation or data analysis. cals, Tarrytown, NY, USA) reported that the HDAC inhibitor Finally, DR Artis (Plexxikon, Inc., Berkeley, CA, USA) pro- suberoylanilide hydroxamic acid (SAHA) is being used in an vided a superb example of how structure–activity relation- oral formula to treat advanced malignancies of both solid ships, automation and informatics will send drug discovery tumours and lymphomas and leukaemias. SAHA induces into overdrive in the near future. Artis described high apoptosis and differentiation, and blocks proliferation of trans- throughput crystallisation that resulted in an amazing formed cells by increasing the expression of cell cycle inhibitor 1300 co-crystals, including 250 unique structures solved, and p21 and decreasing the expression of cyclin D1. The effects are gave examples of two anti-diabetic compounds acting on presumably due to increased acetylation of histones H3 and PPARγ (PLX-101203, PLX-101204) that were identified in H4, which can be tracked in the blood of patients. The toxic the process, all accomplished during the course of 1 year. effects (anorexia, fatigue, diarrhoea) are manageable and revers-ible with some patients being on the drug for up to 2 years [16].
8. Conclusion and expert opinion
7. Other developments and future directions
The parameters for rational drug discovery for NRs havebeen identified – the receptors, their structure and mecha- Another area that was addressed that will become increasingly nism of action, the cofactors they recruit, and the target important to drug discovery is the interplay between herbals, genes they regulate. However, whereas we have detailed Expert Opin. Ther. Targets (2003) 7(5)
information about the first two, we have only begun to 4000 genes have been identified in the human genome that uncover the tip of the proverbial iceberg of the latter two.
contain verified HNF4-binding sites in their regulatory What is needed now is a systematic effort to identify and regions. Even if only one quarter of those genes are true tar- characterise all cofactors, all DNA response elements and gets, that still leaves two orders of magnitude more target all target genes for each NR in the human body, as well as genes than we are used to thinking about. Factor in poten- all of their isoforms and post translational modifications, tially hundreds of different cofactors, dozens of developmen- topics barely touched upon in the conference. Such infor- tal time points, several tissues and countless environmental mation, which would provide an indispensable atlas not conditions, and then consider that any one gene is also regu- just for drug discovery but also for basic research, should be lated by multiple transcription factors, many of which regu- a public resource.
late each other's expression [19]. A complex biological web The second area that needs more attention is the develop- quickly emerges, rivalling the intricacy of the worldwide web.
ment of new ways to organise and analyse the mountains of Clearly, we need to develop new methods to handle, organise, data being generated on a regular basis. Nearly every presenta- analyse and even present all the information that is being pro- tion in the conference relied on flow diagrams to keep track of duced, not just by the prodigious drug discovery efforts of the numerous target genes, many with contradictory effects, pharma, but by all the biological experiments in the post- for just one NR. In the author's laboratory, more than genomic age.
Metabolic Regulators. In: Hormones and ORDENTLICH P, YAN Y, ZHOU S, Papers of special note have been highlighted as Signalling. O'Malley BW (Ed.), HEYMAN RA: Identification of the either of interest (•) or of considerable interest (••) Academic Press (2000):23-87.
antineoplastic agent 6-mercaptopurine as an Excellent review on the so-called
activator of the orphan nuclear hormone receptor Nurr1. J. Biol. Chem. (2003) BURRIS TP, MCCABE ERB: KALLEN JA, SCHLAEPPI JM, BITSCH F Nuclear Receptors and Genetic Disease. et al.: X-ray structure of the hRORα LBD at WANSA KD, HARRIS JM, YAN G, Academic Press, San Diego (2001).
1.63 A: structural and functional data that ORDENTLICH P, MUSCAT GE: Book reviewing NRs and their role in
cholesterol or a cholesterol derivative is the The AF-1 domain of the orphan nuclear human disease, with emphasis on mutations.
natural ligand of RORα. Structure (2002) receptor NOR-1 mediates trans-activation, STEINMETZ AC, RENAUD JP, coactivator recruitment, and activation by the MORAS D: Binding of ligands and WISELY GB, MILLER AB, DAVIS RG purine anti-metabolite 6-mercaptopurine. activation of transcription by nuclear et al.: Hepatocyte nuclear factor 4 is a J. Biol. Chem. (2003) 278:24776-24790.
receptors. Annu. Rev. Biophys. Biomol. Struct. transcription factor that constitutively MITSIADES N, MITSIADES CS, binds fatty acids. Structure (2002) RICHARDSON PG et al.: FREEDMAN LP: Increasing the complexity Molecular sequelae of histone deacetylase of coactivation in nuclear receptor signaling. DHE-PAGANON S, DUDA K, inhibition in human malignant B cells. Cell (1999) 97:5-8.
IWAMOTO M, CHI YI, SHOELSON SE: Blood (2003) 101:4055-4062.
ALBERT S, GAUDAN S, KNIGGE H Crystal structure of the HNF4α ligand WATKINS RE, MAGLICH JM, et al.: Computer-assisted generation of a binding domain in complex with MOORE LB et al.: 2.1 A crystal structure of protein-interaction database for nuclear endogenous fatty acid ligand. J. Biol. Chem. human PXR in complex with the St. John's receptors. Mol. Endocrinol. (2003) wort compound hyperforin. Biochemistry BILLAS IM, MOULINIER L, ROCHEL N, A global view of protein interactions
MORAS D: Crystal structure of the HONKAKOSKI P, SUEYOSHI T, ligand-binding domain of the ultraspiracle NEGISHI M: Drug-activated nuclear THOMPSON EB, KUMAR R: protein USP, the ortholog of retinoid X receptors CAR and PXR. Ann. Med. (2003) DNA binding of nuclear hormone receptors receptors in insects. J. Biol. Chem. (2001) influences their structure and function. LEVINE M, TJIAN R: Biochem. Biophys. Res. Commun. (2003) GRESCHIK H, WURTZ JM et al.: Transcription regulation and animal Structural and functional evidence for diversity. Nature (2003) 424:147-151.
WONG CW, MCNALLY C, ligand-independent transcriptional A global, yet concise, view of transcription
NICKBARG E, KOMM BS, CHESKIS BJ: activation by the estrogen-related receptor 3. regulation in animals.
Estrogen receptor-interacting protein that Mol. Cell. (2002) 9:303-313.
modulates its nongenomic activity-crosstalk BAKER KD, SHEWCHUK LM, with Src/Erk phosphorylation cascade. KOZLOVA T et al.: The Drosophila orphan Proc. Natl. Acad. Sci.USA (2002) nuclear receptor DHR38 mediates an NucleaRDB: an Information System for atypical ecdysteroid signaling pathway. Nuclear Receptors.
SLADEK R, GIGUERE V: Orphan Nuclear Cell (2003) 113:731-742.
Emphasis on sequence alignments,
Receptors: An Emerging Family of structures and mutations.
Expert Opin. Ther. Targets (2003) 7(5)
Nuclear receptors as drug targets: new developments in coregulators, orphan receptors and major therapeutic areas
The Nuclear Receptor Resource.
General resource on NRs.
nurebase.htmlNuclear Receptor Database.
Nomenclature and evolution of
NR superfamily.
Affiliation
Frances M Sladek PhD
Department of Cell Biology and Neuroscience,
5429 Boyce Hall, University of California,
Riverside, CA 92521-0314, USA
Tel: +1 909 787 2264; Fax: +1 909 787 3087;
E-mail: [email protected]
Expert Opin. Ther. Targets (2003) 7(5)

Source: http://www.sladeklab.ucr.edu/papers/ExpertOpinion.NR.drugs.2003.pdf