(1)
Wyeth Research, Women's Health Research Institute, 500 Arcola Rd, RN-3163, 19426 Collegeville PA, USA
Email: harrish@wyeth.com
Abstract
It has now been over 10 years since Jan-Ake Gustafsson revealed the existence of a second form of the estrogen receptor (ERβ) at a 1996 Keystone Symposium. Since then, substantial success has been made in distinguishing its potential biological functions from the previously known form (now called ERα) and how it might be exploited as a drug target. Subtype selective agonists have been particularly useful in this regard and suggest that ERβ agonists may be useful for a variety of clinical applications without triggering classic estrogenic side effects such as uterine stimulation. These applications include inflammatory bowel disease, rheumatoid arthritis, endometriosis, and sepsis. This manuscript will summarize illustrative data for three ERβ selective agonists, ERB-041, WAY-202196, and WAY-200070.
1 Introduction and Review of Compounds Discussed
Estrogens are classically described as ligand-activated transcription factors that modulate gene transcription via two intracellular receptors, ERα and ERβ. Although other mechanisms of signaling have been described, this path is the best described. ERα was the first ER cloned and a knockout mouse was made approximately 13 years ago. Although not all aspects of estrogen biology were neatly solved by these discoveries, the field generally accepted that there was but a single ER. When ERβ was unexpectedly discovered in a rat prostate cDNA library 10 years ago, it rejuvenated the field of estrogen research with the new goal of attributing estrogen's panoply of effects to the appropriate ER. A number of approaches were taken to investigate this question, including examining receptor distribution, in vitro activity/interaction studies, construction of knockout mice, and design of selective agonists/antagonists. Given the similarity of ERα and ERβ ligand-binding pockets, it was a particular challenge to develop highly selective ligands. In fact, although a variety of selective agonists have been developed (Veeneman 2005), highly selective antagonists have not been designed.
Data from three selective agonists synthesized by Wyeth Research (Malamas et al. 2004; Mewshaw et al. 2005) will be presented in an effort to outline our current understanding of ERβ biology. The majority of material discussed here was presented at the Ernst Schering Symposium on Tissue-Specific Estrogen Action in March 2006. The biology of other ER selective agonists (both ERα and ERβ) has been recently reviewed elsewhere (Harris 2007).
The structure of compounds discussed in this article and their binding affinities (as measured by IC50) are shown in Fig. 1. All three compounds are nonsteroidal agonists and bind to human ERβ ligand binding domain with roughly the same affinity as 17β-estradiol. Their selectivity in this assay varies from approximately 70- to more than 200-fold. In our experience, these and other structurally diverse ERβ agonists have similar in vivo profiles, although we cannot be certain that these compounds are capable of eliciting all ERβ-mediated effects. As more in vivo data is published on other ERβ selective agonists, the full spectrum of ERβ biology will be elaborated. It should be noted that the activity of other ERβ selective agonists is discussed elsewhere in this volume (see Chaps. 4, 6, 7 and 8).
Fig. 1
The structure of ERβ agonists discussed in this article, their binding affinity (as measured by IC50) and selectivity as assessed using the human ligand-binding domain in a competitive radioligand-binding assay
2 ERβ Agonists Lack Classic Estrogenic Effects
Key to ERβ's attractiveness as a drug target was the expectation that selective agonists would have reduced impact on the uterus and mammary gland. This assumption was supported by tissue distribution studies and the phenotype of the ERα knockout mouse. Indeed, in our hands, these three compounds (as well as many others) are nonuterotrophic in the sexually immature rat. However, given subcutaneously at high doses (~90 mg/kg), to ovariectomized adult rats, WAY-200070 does have some mild uterine stimulatory activity (unpublished observations). As part of our safety assessment of ERB-041 and WAY-202196 in preparation for clinical trials, higher doses have been tested orally in two species and no uterine stimulatory activity has been seen.
Fig. 2a–c
Activity of ERB-041 in the 7-day mouse mammotrophic assay. a Whole mount images of mammary glands from animals treated with 17β-estradiol (1 μg/kg) + progesterone (30 mg/kg). b Whole mount images of mammary glands from animals treated with ERB-041 (50 mg/kg) + progesterone (30 mg/kg). c Defensin β1 mRNA expression in mammary glands from animals treated with 17β-estradiol (E2, 1 μg/kg), 17β-estradiol (1 μg/kg) + progesterone (30 mg/kg) (E2 + P4), ERB-041 (50 mg/kg), ERB-041 + progesterone (30 mg/kg) (ERB-041 + P4) or 17β-estradiol (1 μg/kg) + progesterone (30 mg/kg) + ERB-041 (50 mg/kg) (E2 + ERB-041 + P4)
The mouse mammary gland responds to the combination of an estrogen and a progestin by elaboration/development of ducts and formation of lobuloalveolar endbuds. Typically these steroids are administered for approximately 3 weeks in order to see full development of mammary gland morphology. However, to facilitate compound evaluation, a 7-day model was developed (Crabtree et al. 2006). Under this abbreviated regimen, morphological changes still occur, although they are less pronounced than those seen with longer exposure. In addition, we measured defensin β1 mRNA expression. This gene is uniquely upregulated by the combination of estrogen and progestin; neither compound alone elevates its expression. ERB-041 and WAY-202196 (50 mg/kg, PO) were evaluated alone and in combination with estradiol, estradiol + progesterone, and progesterone in this model and did not appreciably alter morphology or significantly change defensinβ1 mRNA expression (Fig. 2; Crabtree et al. 2006). Therefore, they are inactive as estrogens, antiestrogens, progestins, or antiprogestins in this model.
ERB-041 and WAY-200070 (10 mg/kg SC) have been evaluated for their ability to prevent bone mineral density loss after ovariectomy in the rat (Harris et al. 2003; Malamas et al. 2004). Unlike estrogens, selective estrogen receptor modulators (Miller 2002) or ERα selective agonists (Harris et al. 2002; Hillisch et al. 2004), these compounds were inactive. Finally, ERB-041 failed to inhibit ovulation in rats (Harris et al. 2003), another point of divergence between ERβ selective agonists and nonsubtype selective estrogens.
3 Improving Intestinal Function: A Common Theme Among Three Vivo Efficacy Models
3.1 HLA-B27 Transgenic Rat
The first in vivo activity seen with these ERβ selective agonists was in a model of inflammatory bowel disease, the HLA-B27 transgenic rat (Harris et al. 2003). These rats experience chronic diarrhea from about 8–10 weeks of age until their death. Daily oral doses of ERB-041 (Harris et al. 2003) and WAY-202196 (Mewshaw et al. 2005) given to male rats rapidly normalized stool character and improved intestinal histology. In this model, doses of 1 mg/kg (PO) and greater were fully effective. The mechanism of action of ERβ is unclear, as is the target cell; ERβ is expressed in the colonic epithelium as well as cells of the immune system, making multiple sites of action possible.
3.2 Mdr1aKO Mouse Model of Colitis
Mdr1a (P-glycoprotein) knockout mice spontaneously develop colitis due to defects in intestinal barrier epithelial function, and males develop earlier and more severe disease than females (Resta-Lenert et al. 2005). Because of the effects seen in the HLA-B27 transgenic rat, a group of male and female mdr1aKO mice were treated orally with either vehicle or WAY-200070 (2 mg/kg) for 10 days, at which time colonic epithelium was removed for functional studies and expression of inflammatory mediators (Barrett and Resta-Lenert, unpublished observations). When placed in Ussing chambers to assess epithelial integrity, colonic epithelium from vehicle-treated mdr1aKO mice had elevated baseline current, indicating a defect in barrier function, and this deficit was significantly reversed when mice were treated with WAY-200070. In addition, when stimulated by forskolin, colonic epithelium from vehicle-treated mice responded poorly to this secretagogue, whereas epithelium from mice treated with WAY-200070 responded normally. Finally, epithelium from vehicle-treated mice had elevated COX-2 and iNOS protein expression and this increase was largely blunted by WAY-200070. Because WAY-200070 has poor systemic exposure upon oral dosing, these data suggest the compound affects the colonic epithelium directly.
3.3 Rodent Models of Sepsis
Sepsis can be generally described as a maladaptive response to infection, and two hallmarks of this disease are epithelial and endothelial barrier dysfunction (Buras et al. 2005; Vincent and Abraham 2005). The advantage that females have over males in conditions of trauma and shock have been well described (Angele et al. 2000). Recent work has specifically focused on the intestine as a target organ for damage and shown that female rats have improved barrier function and mount less of a pro-inflammatory response (as measured by IL-6 and TNFα) than do males after hypoxic or acidic insult (Homma et al. 2005).
Several animal models of sepsis exist, and WAY-202196 has been evaluated in two: Pseudomonas infection of the neutropenic rat and mouse cecal ligation and puncture (Cristofaro et al. 2006). In the first model, rats are rendered neutropenic by cyclophosphamide and normal gut flora is disrupted by an antibiotic. They then receive an oral bolus of Pseudomonas aeruginosa. WAY-202196 was studied twice. In the first study, rats were dosed on days 4 and 6 after Pseudomonas inoculation. In the second study, rats were dosed from days 4–11 after Pseudomonas inoculation. In both studies, histological signs of injury were significantly reduced in the intestinal epithelium. Moreover, in the longer-term study, survival was significantly increased by administration of the ERβ selective agonist.
In the second model, puncturing the cecum and expressing a small amount of bowel contents into the peritoneal cavity induces a peritoneal infection. WAY-202196 was administered at the time of surgery, 24 and 48 h afterwards. Consistent with the observation from the rat model, WAY-202196 increased survival, and histological signs of injury were significantly reduced in the ileum. Follow-up studies indicate that the compound is equally effective in males and females, that the minimum effective oral dose of WAY-202196 is 1 mg/kg and that intravenous dosing is as effective as oral dosing (Cristofaro et al. 2006; Opal et al., unpublished data).
4 Endometriosis and Inflammatory Pain
Although endometriosis is undoubtedly an estrogen-responsive disease, it is now appreciated that immune system dysfunction may explain why only a subset of women with retrograde menstruation develop the disease. Because of the anti-inflammatory activity seen with ERB-041 in other models, we evaluated it in a rodent model of endometriosis. The model chosen was a xenograft model using normal human endometrial fragments implanted into nude mice. These tissue fragments adhere, implant, establish a blood supply, and form lesions that are histologically similar to human disease (Bruner et al. 1997; Grummer et al. 2001). Dosing with ERB-041 began approximately 2 weeks after tissue implantation, and continued for about 2 weeks. Spontaneous lesion regression was not seen in vehicle-treated mice, but 40%–75% of mice treated with ERB-041 were completely lesion-free (depending on the study) (Harris et al. 2005). Interestingly, ERB-041 seemed more effective at causing lesion regression when implants were established inside the peritoneal cavity than subcutaneously. As with the other models, the mechanism of action is uncertain; however, because lesions recovered at the end of the study express ERα and not ERβ, the compound is likely acting on the host (e.g., an immunomodulatory effect) rather than on the implants (a pro-apoptotic effect). The fact that ERB-041 is active in a model of this disease illustrates that an ERβ agonist's profile is not just a subset of estradiol's activity or that of other nonselective ER agonists.
Pain is a central feature of several diseases for which ERβ selective agonists have been effective in preclinical models. We examined whether ERB-041 had antinociceptive activity in several models, including a model of inflammatory pain (Leventhal et al. 2006). When prostaglandin E2 is injected into a rat's tail, the tail becomes hypersensitive to warm water. Acute oral administration of ERB-041 (10 mg/kg) can partially reverse this effect, and its action is blocked by the ER antagonist ICI-182780. Similar results are seen when capsaicin is used as the sensitizer. However, ERB-041 was not effective in other models of pain, including postsurgical pain and neuropathic pain. Again, not understanding the mechanism of action impedes an explanation for these patterns of activity.
5 Hypothalamic–Pituitary–Adrenal Axis: A New Area of Investigation
One of the challenges we face regarding our compounds is the ambiguity of their mechanism of action. To date, all the in vivo activities described for this set of ERβ selective agonists relate to inflammation and/or the immune system. These activities may be anti-inflammatory, immunomodulatory, or potentially even immunostimulatory, but thus far, they share this common thread. There is the possibility that these compounds may affect the hypothalamic–pituitary–adrenal (HPA) axis and that this may explain their activity in some in vivo models. There is tremendous precedent for the actions of estrogens on the HPA axis. For example, sex differences in stress responses exist at both the behavioral and biochemical levels and across species (Bowman et al. 2002; Walf and Frye 2005; Kajantie and Phillips 2006). Estrogens affect a variety of rodent behavioral models of anxiety, although the effects vary with the model, dosing regimen and dose. These conflicting data likely result from the system's inherent complexity, but may also be influenced by the inherently different activity of ERα and ERβ and the fact that the best-studied estrogen, 17β-estradiol, can activate both ER subtypes.
ERβ has the potential to modulate the HPA axis in that it is expressed in the adrenal gland of a variety of species (Saunders et al. 1997; Albrecht et al. 1999) and is the dominant ER in the rat paraventricular nucleus (Shughrue et al. 1997). Moreover, one of the models where ERB-041 and WAY-202196 have profound effects is the Lewis rat adjuvant-induced arthritis model (Harris et al. 2003; Mewshaw et al. 2005; Follettie et al. 2006). Lewis rats are hypersensitive to inflammatory stimuli because this strain does not secrete appropriate corticotropin releasing hormone from the paraventricular nucleus. In fact, these rats have reduced basal ERα and ERβ expression in this nucleus (Tonelli et al. 2002). Finally, recent studies have begun to implicate ERβ in influencing the function of the HPA axis (Isgor et al. 2003; Miller et al. 2004; Lund et al. 2005, 2006).
We have examined ERβ mRNA and protein levels in a rat model of immobilization stress. Immobilization is a very potent stressor and leads to a large activation of the HPA axis as well as sympathoadrenal catecholaminergic systems. Pretreatment with estradiol benzoate has been previously shown to blunt stress-induced plasma ACTH and to modulate a variety of basal and stress-induced changes in gene expression in central and peripheral catecholaminergic locations (Serova et al. 2005). Repeated immobilization stress was found to dramatically upregulate ERβ in the adrenal medulla at both the mRNA and protein levels (Sabban E, unpublished observations). Interestingly, on a Western blot, several immunoreactive species are detected, with the predominant one being at 45 kD. The characterization of this isoform's sequence remains to be determined. Studies are underway to determine the effect of WAY-200070, compared to estradiol benzoate, on a variety of immobilization stress-induced biochemical responses.
6 Summary and Future Directions
The preclinical biology of selective ERβ agonists is as impressive as it is diverse. Using a set of selective agonists, we have discovered that manipulation of this receptor's activity may have value in treating human diseases such as inflammatory bowel disease, rheumatoid arthritis, endometriosis, and sepsis. In fact, phase II clinical trials are currently underway with ERB-041 for rheumatoid arthritis and endometriosis. Although there were hints that nonsubtype selective estrogens might affect some of these diseases (but not endometriosis), their potential beneficial effects were overshadowed by their effects on classic target tissues such as the uterus and mammary gland. Despite this progress, however, several key questions remain about ERβ selective agonists: first, and foremost from a scientific perspective, what are the molecular mechanism(s) behind the in vivo activities observed? Second, do these series of compounds reveal all aspects of ERβ biology or will other selective compounds have a different spectrum of activity? Lastly, and most important from a pharmaceutical perspective, what is the long-term safety profile of these compounds, and will their preclinical profile be mirrored in the human analog of these diseases? The next year or so will provide answers to some of these questions.
Acknowledgements
Our accomplishments in the field of ERβ research were enabled by the cooperative efforts of a number of scientists. First and foremost, Chris Miller, Mike Malamas, Rick Mewshaw, and Eric Manas led the chemistry effort to design and deliver highly selective ERβ agonists for biological characterization. Among the biologists, Jim Keith made key discoveries of in vivo activity; C. Richard Lyttle supported the team and suggested the endometriosis model evaluation. Several external collaborators have also contributed their expertise: Kaylon Bruner-Tran and Kevin Osteen (Vanderbilt University; endometriosis model), Kim Barrett and Silvia Resta-Lenert (University of California, San Diego; MDR1aKO mouse), Esther Sabban and Lydia Serova (New York Medical College, immobilization stress model), Steve Opal (Brown University; sepsis models). Finally I thank Rick Winneker for valuable suggestions on this manuscript.
References
Albrecht ED, Babischkin JS, Davies WA, Leavitt MG, Pepe GJ (1999) Identification and developmental expression of the estrogen receptor alpha and beta in the baboon fetal adrenal gland. Endocrinology 140:5953–5961PubMedCrossRef
Angele MK, Schwacha MG, Ayala A, Chaudry IH (2000) Effect of gender and sex hormones on immune responses following shock. Shock 14:81–90PubMedCrossRef
Bowman RE, Ferguson D, Luine VN (2002) Effects of chronic restraint stress and estradiol on open field activity, spatial memory, and monoaminergic neurotransmitters in ovariectomized rats. Neuroscience 113:401–410PubMedCrossRef
Bruner KL, Matrisian LM, Rodgers WH, Gorstein F, Osteen KG (1997) Suppression of matrix metalloproteinases inhibits establishment of ectopic lesions by human endometrium in nude mice. J Clin Invest 99:2851–2857PubMedCrossRef
Buras JA, Holzmann B, Sitkovsky M (2005) Animal models of sepsis: setting the stage. Nat Rev Drug Discov 4:854–865PubMedCrossRef
Crabtree JS, Zhang X, Peano BJ, Zhang Z, Winneker RC, Harris HA (2006) Development of a mouse model of mammary gland versus uterine tissue selectivity using estrogen- and progesterone-regulated gene markers. J Steroid Biochem Mol Biol 101:11–21PubMedCrossRef
Cristofaro PA, Opal SM, Palardy JE, Parejo NA, Jhung J, Keith JC Jr, Harris HA (2006) WAY-202196, a selective estrogen receptor-beta agonist, protects against death in experimental septic shock. Crit Care Med 34:2188–2193PubMedCrossRef
Follettie MT, Pinard M, Keith JC, Wang LL, Chelsky D, Hayward C, Kearney P, Thibault P, Paramithiotis E, Dorner AJ, Harris HA (2006) Organ messenger ribonucleic acid and plasma proteome changes in the adjuvant-induced arthritis model: responses to disease induction and therapy with the estrogen receptor-beta selective agonist ERB-041. Endocrinology 147:714–723PubMedCrossRef
Grummer R, Schwarzer F, Bainczyk K, Hess-Stumpp H, Regidor P, Schindler A, Winterhager E (2001) Peritoneal endometriosis: validation of an in vivo model. Human Reprod 16:1736–1743CrossRef
Harris HA (2007) Estrogen receptor-beta: recent lessons from in vivo studies. Mol Endocrinol, pp 2111–2113
Harris HA, Katzenellenbogen JA, Katzenellenbogen BS (2002) Characterization of the biological roles of the estrogen receptors ER alpha and ER beta, in estrogen target tissues in vivo through the use of an ER alpha-selective ligand. Endocrinology 143:4172–4177PubMedCrossRef
Harris HA, Albert LM, Leathurby Y, Malamas MS, Mewshaw RE, Miller CP, Kharode YP, Marzolf J, Komm BS, Winneker RC, Frail DE, Henderson RA, Zhu Y, Keith JC Jr (2003) Evaluation of an estrogen receptor-beta agonist in animal models of human disease. Endocrinology 144:4241–4249PubMedCrossRef
Harris HA, Bruner-Tran KL, Zhang X, Osteen KG, Lyttle CR (2005) A selective estrogen receptor-beta agonist causes lesion regression in an experimentally induced model of endometriosis. Human Reproduction 20:936–941PubMedCrossRef
Hillisch A, Peters O, Kosemund D, Muller G, Walter A, Schneider B, Reddersen G, Elger W, Fritzmeier KH (2004) Dissecting physiological roles of estrogen receptor alpha and beta with potent selective ligands from structure-based design. Mol Endocrinol 18:1599–1609PubMedCrossRef
Homma H, Hoy E, Xu D-Z, Lu Q, Feinman R, Deitch EA (2005) The female intestine is more resistant than the male intestine to gut injury and inflammation when subjected to conditions associated with shock states. Am J Physiol Gastrointest Liver Physiol 288:G466–G472PubMedCrossRef
Isgor C, Cecchi M, Kabbaj M, Akil H, Watson SJ (2003) Estrogen receptor beta in the paraventricular nucleus of hypothalamus regulates the neuroendocrine response to stress and is regulated by corticosterone. Neuroscience 121:837–845PubMedCrossRef
Kajantie E, Phillips DIW (2006) The effects of sex and hormonal status on the physiological response to acute psychosocial stress. Psychoneuroendocrinology 31:151–178PubMedCrossRef
Leventhal L, Brandt MR, Cummons TA, Piesla MJ, Rogers KE, Harris KA. An estrogen receptor-β agonist is active in models of inflammatory and chemical-induced pain. European Journal of Pharmacology 553:146–148
Lund TD, Rovis T, Chung WCJ, Handa RJ (2005) Novel actions of estrogen receptor-beta on anxiety-related behaviors. Endocrinology 146:797–807PubMedCrossRef
Lund TD, Hinds LR, Handa RJ (2006) The androgen 5 alpha-dihydrotestosterone and its metabolite 5 alpha-androstan-3 beta,17 beta-diol inhibit the hypothalamo-pituitary-adrenal response to stress by acting through estrogen receptor beta-expressing neurons in the hypothalamus. J Neurosci 26:1448–1456PubMedCrossRef
Malamas MS, Manas ES, McDevitt RE, Gunawan I, Xu ZB, Collini MD, Miller CP, Dinh T, Henderson RA, Keith JC Jr, Harris HA (2004) Design and synthesis of aryl diphenolic azoles as potent and selective estrogen receptor-beta ligands. J Med Chem 47:5021–5040PubMedCrossRef
Mewshaw RE, Edsall RJ, Yang CJ, Manas ES, Xu ZB, Henderson RA, Keith JC, Harris HA (2005) ER beta ligands. 3. Exploiting two binding orientations of the 2-phenylnaphthalene scaffold to achieve ER beta selectivity. J Med Chem 48:3953–3979PubMedCrossRef
Miller CP (2002) SERMs: evolutionary chemistry, revolutionary biology. Curr Pharmaceut Des 8:2089–2111CrossRef
Miller WJS, Suzuki S, Miller LK, Handa R, Uht RM (2004) Estrogen receptor (ER)beta isoforms rather than ER alpha regulate corticotropin-releasing hormone promoter activity through an alternate pathway. J Neurosci 24:10628–10635PubMedCrossRef
Resta-Lenert S, Smitham J, Barrett KE (2005) Epithelial dysfunction associated with the development of colitis in conventionally housed mdr1a(-/-) mice. Am J Physiol Gastrointest Liver Physiol 289:G153–G162PubMedCrossRef
Saunders PTK, Maguire SM, Gaughan J, Millar MR (1997) Expression of oestrogen receptor beta (ER-beta) in multiple rat tissues visualised by immunohistochemistry. J Endocrinol 154:R13–R16PubMedCrossRef
Serova LI, Maharjan S, Sabban EL (2005) Estrogen modifies stress response of catecholamine biosynthetic enzyme genes and cardiovascular system in ovariectomized female rats. Neuroscience 132:249–259PubMedCrossRef
Shughrue PJ, Lane MV, Merchenthaler I (1997) Comparative distribution of estrogen receptor-alpha and -beta mRNA in the rat central nervous system. J Comp Neurol 388:507–525PubMedCrossRef
Tonelli L, Kramer P, Webster JI, Wray S, Listwak S, Sternberg E (2002) Lipopolysaccharide-induced oestrogen receptor regulation in the paraventricular hypothalamic nucleus of Lewis and Fischer rats. J Neuroendocrinol 14:847–852PubMedCrossRef
Veeneman GH (2005) Non-steroidal subtype selective estrogens. Curr Med Chem 12:1077–1136PubMedCrossRef
Vincent J-L, Abraham E (2005) The last 100 years of sepsis. Am J Respir Crit Care Med 173:256–263PubMedCrossRef
Walf AA, Frye CA (2005) Antianxiety and antidepressive behavior produced by physiological estradiol regimen may be modulated by hypothalamic-pituitary-adrenal axis activity. Neuropsychopharmacology 30:1288–1301PubMedCrossRef