Image Map

buff old looking guy standing

By Karl Hoffman

Despite decades of research, there is as yet no universally agreed upon model that explains the observed actions of androgens and anabolic steroids (AAS) when these agents are administered to animals and humans. The classical model of AAS action posits that after an anabolic/androgenic compound binds to the androgen receptor (AR), the AAS/AR complex enters the nucleus of target cells, recruits additional transcription factors, and initiates the transcription of androgen responsive genes. In a recent issue of Mind & Muscle we described how one model attributes the spectrum of effects caused by AAS to activation of different genes by different steroids

We also briefly discussed in that article how some androgens may exert their effects by inhibiting the action of glucocorticoids, catabolic hormones produced by the adrenal glands. It is this proposed mechanism that I would like to look at in more detail in this article.


Androgens and Glucocorticoids are generally accepted as having opposing actions in skeletal muscle. The former breaks down muscle during periods of stress in order to provide a supply of amino acids that can be used as fuel substrates. It does this primarily by increasing the activity of the ubiquitin-proteasome proteolytic system in muscle. Glucocorticoids also inhibit the growth of muscle, possibly by increasing levels of myostatin (1). Androgens on the other hand promote muscle hypertrophy and according to some studies, muscle hyperplasia (growth of new muscle fibers) via a number of proposed mechanisms.

Research has shown that at least two AAS, nandrolone and R 1881 (methyltrienolone) competitively bind to the glucocorticoid receptor (GR) in the cytosol of cells. (2) It was observed that androgens were able to displace dexamethasone, a potent synthetic GR agonist, from the GR in a concentration dependent manner. So here we have an example of how AAS might exert an anabolic effect independently of any binding to the androgen receptor (AR). In some respects this result is not surprising, since the GR and the AR share a high degree of sequence homology, meaning simply that the DNA sequences that code for the two receptors are quite similar.

Competitive inhibition of glucocorticoid binding to the GR by androgens could explain how certain androgens with a very weak binding affinity for the AR (e.g. oxymetholone) nevertheless act as potent anabolic agents: they may have a relatively high affinity for the GR. Of course, we also have an alternate model described in the issue of Mind & Muscle linked to above, where low AR binding-affinity AAS may still be capable of activating genes that strongly promote anabolism. Yet a third model, described below, attributes the inhibition of glucocorticoid action by androgens (and vice versa, the inhibition of androgen action by glucocorticoids) to a form of “crosstalk” between the AR and the GR. This interaction between the two receptors inhibits the binding of the receptors to their respective hormone response elements on target genes.


Interestingly, some androgens and synthetic AAS such as DHT and oxandrolone are unable to displace such potent endogenous glucocorticoids as cortisol from the GR, yet in culture these androgens inhibit the expression of cortisol dependent genes in the presence of cortisol (3, 4). Clearly this observation is incompatible with the simple model of competitive inhibition of the GR by AAS described above. Evidence for lack of competitive inhibition is bolstered by the observation that the cellular presence of the AR is required for the observed glucocorticoid antagonism, at least in the case of oxandrolone (4). Clearly, had the effect been observed in cells lacking the AR, there would have been no possibility of any GR/AR interaction.

Normally, in order for androgens to activate target genes, they must bind to the androgen receptor, which then binds (dimerizes) to another androgen receptor forming a so-called homodimer. The receptors joined in this way are then able to bind stably to the target gene, recruit various transcription factors, and initiate gene transcription, which eventually leads to protein synthesis. What Chen et. al. (1) discovered was that the androgen receptor and the glucocorticoid receptor are capable of forming so called heterodimers that repress transcription of target genes instead of activating the genes, as in the case of androgen or glucocorticoid homodimers. See fig 1.

Figure 1. Schematic illustration of AR/GR heterodimer bound by ionic bond between Arginine (+) and Aspartate (-) ARDBD = androgen receptor DNA Binding Domain; GRDBD = glucocorticoid receptor DNA binding domain. (Adapted from Chen


While Zhao et. al. in their study of glucocorticoid antagonism by oxandrolone did not look specifically for the presence of AR/GR heterodimers, their data are consistent with this model. Nevertheless, as the authors pointed out, potential alternative mechanisms may exist. In order for the AR and the GR to form homo or heterodimers, they must first enter the nucleus bound to their respective ligands. Zhao et. al. constructed a mutant AR that was incapable of translocating into the nucleus. When oxandrolone was added to cell cultures containing the normal GR and the mutant AR, GR dependent gene transcription was still repressed. To quote the authors: “while the mechanisms underlying this surprising finding remain to be defined, these data indicate that repression can occur even when the two receptors are localized in different subcellular compartments”. So in this admittedly artificial system where the engineered AR was unable to enter the nucleus and interfere with GR binding, the mutant AR was nevertheless able to communicate with the normal GR to interfere with its action.

Another possible mechanism for mutual repression is so-called squelching, where the two receptors compete for a limited number of transcriptional co-activators. For example, in the case of excess androgen relative to cortisol, more androgen receptors will translocate to the nucleus, effectively using up the co-activators common to both receptors (4). Squelching has been observed in a number of systems. One in particular may have important physiological and environmental implications. The action of estrogen is antagonized by a number of xenobiotic estrogenic compounds (e.g. DDE, a breakdown product of DDT) that bind only very weakly to the estrogen receptor but bind strongly to the so called xenobiotic orphan receptor CAR. Ligand bound CAR has been shown to inhibit the action of estrogen via squelching (5). The possibility exists that residues from xenobiotic estrogens may interfere with female reproductive physiology. Some researchers have attributed the worldwide decline in amphibians to CAR/ER squelching.


Heterodimer formation may have physiological implications as well. The classic model of receptor dimerization posits that receptor dimers are required for stable binding of the receptor to DNA. Chen et. al. suggest another possibility, namely that the ability to form homodimers is an effect secondary in importance to the ability of different receptors to form heterodimers. They suggest that heterodimer formation is a regulatory mechanism that allows the body to respond to changing environmental conditions.

For example, during periods of stress such as induced by starvation, the body’s cortisol to testosterone ratio increases. This would lead to the formation of GR/AR heterodimers and an excess of GR homodimers. The result would be a shutdown of energy consuming anabolic processes and an increase in catabolic processes to provide energy substrates. Squelching could also represent such a regulatory system.


Before leaving the subject of heterodimerization, it’s logical to ask whether such a phenomenon occurs with other receptor types, and whether there are physiological or clinical implications for such heterodimerization. It turns out the answer is yes on both counts. For example, bradykinin and angiotensin II are compounds in the body that are involved in the regulation of blood pressure. The former binds to the B (2) receptor and acts as a vasodepressor, lowering blood pressure. The latter binds to the AT(1) receptor as a vasopressor, elevating blood pressure (There may be readers with hypertension who use a so called ACE inhibitor such as lisinopril that inhibits the formation of angiotensin II, or an angiotensin receptor antagonist like valsartan, that blocks the action of angiotensin at its receptor). Heterodimerization between AT (1) and B (2) has been observed, and the combined AT (1)/B (2) receptor is hypersensitive to the effects of angiotensin II, leading to high blood pressure (6). The hypertension that accompanies preeclampsia during pregnancy may be the result of such heterodimerization (7).

About half of all Europeans are either homo or heterozygous for an isoform of the growth hormone (GH) receptor gene that lacks exon 3, so called (d3-GHR). In people who possess both the regular GH receptor and the d3 isoform (heterozygous individuals), the two receptor types are capable of forming heterodimers, increasing the sensitivity of the combined receptor to GH. People who are homozygous for d3 are even more sensitive to GH. For example, in children administered GH for short stature, those children homozygous for the d3 allele experience 1.7 to 2 times the growth acceleration than do children that possess the normal allele (8). So it’s very likely that athletes and bodybuilders who self-administer GH experience a degree of responsiveness dependent on whether they are homo or heterozygous for d3, and dependent as well on the extent of heterodimerization of the two isoforms.

Further, it’s been recognized for some time that the estrogen receptor exists in two forms, ER-alpha and ER-beta. The two are capable of forming heterodimers affecting estrogen signaling (9). In fact, it has been suggested that the primary role of ER-beta may be to modulate the effect of ER-alpha, the primary estrogen receptor. ER-beta seems to exert a bodywide repressing effect on ER-alpha dependent transcription (10). In bone, for example, ER-beta activation inhibits the anabolic effect normally exerted by estrogen (10).

These are just a few examples of receptor heterodimerization that may be of clinical importance. The classical picture of a ligand binding to its receptor and the complex initiating gene transcription is blurred now that we see that receptors are able to crosstalk with one another via heterodimerization and other mechanisms like squelching. Perhaps most intriguing is the idea that the requirement for dimerization of hormonal receptors in order that they be active evolved to allow for heterodimerization as a way for receptors to modulate the activity of one another, rather than as simply a way to stabilize DNA binding, as the classical model of receptor action postulates.


(1) Ma K, Mallidis C, Bhasin S, Mahabadi V, Artaza J, Gonzalez-Cadavid N, Arias J, Salehian B.Am J

Glucocorticoid-induced skeletal muscle atrophy is associated with upregulation of myostatin gene expression. Physiol Endocrinol Metab. 2003 Aug;285(2):E363-71

(2) Mayer M, Rosen F.Am J Physiol. Interaction of anabolic steroids with glucocorticoid receptor sites in rat muscle cytosol. 1975 Nov;229(5):1381-6.

(3) Chen S, Wang J, Yu G, Liu W, Pearce D.J Androgen and glucocorticoid receptor heterodimer formation. A possible mechanism for mutual inhibition of transcriptional activity.

Biol Chem. 1997 May 30;272(22):14087-92.

(4) Zhao J, Bauman WA, Huang R, Caplan AJ, Cardozo C. Oxandrolone blocks glucocorticoid signaling in an androgen receptor-dependent manner Steroids. 2004 May;69(5):357-66.

(5) Min G, Kim H, Bae Y, Petz L, Kemper JK Inhibitory cross-talk between estrogen receptor (ER) and constitutively activated androstane receptor (CAR). CAR inhibits ER-mediated signaling pathway by squelching p160 coactivators.J Biol Chem. 2002 Sep 13;277(37):34626-33.

(6) AbdAlla S, Lother H, Quitterer U AT1-receptor heterodimers show enhanced G-protein activation and altered receptor sequestration. Nature. 2000 Sep 7;407(6800):94-8.

(7) AbdAlla S, Lother H, el Massiery A, Quitterer U. Increased AT(1) receptor heterodimers in preeclampsia mediate enhanced angiotensin II responsiveness. Nat Med. 2001 Sep;7(9):1003-9

(8) Dos Santos C, Essioux L, Teinturier C, Tauber M, Goffin V, Bougneres P. A common polymorphism of the growth hormone receptor is associated with increased responsiveness to growth hormone. Nat Genet. 2004 Jul;36(7):720-4

(9) Pettersson K, Grandien K, Kuiper GG, Gustafsson JA. Mouse estrogen receptor beta forms estrogen response element-binding heterodimers with estrogen receptor alpha. Mol Endocrinol. 1997 Sep;11(10):1486-96.

(10) Lindberg MK, Moverare S, Skrtic S, Gao H, Dahlman-Wright K, Gustafsson JA, Ohlsson C Estrogen receptor (ER)-beta reduces ERalpha-regulated gene transcription, supporting a “ying yang” relationship between ERalpha and ERbeta in mice. Mol Endocrinol. 2003 Feb;17(2):203-8