Integrin Receptor-Mediated Actions of Thyroid Hormone

Evidence that thyroid hormone can act primarily outside the cell nucleus has come from studies of mitochondrial responses to T₃ (5) or T2 (6), from rapid onset effects of the hormone at the cell membrane (7-9) and from actions on cytoplasmic proteins (10, 11). The recent description of a plasma membrane receptor for thyroid hormone on integrin αVβ3 (12-14) has provided some insight into effects of the hormone on membrane ion pumps, such as the Na⁺/H⁺ antiporter (9, 15), and has led to the description of interfaces between the membrane thyroid hormone receptor and nuclear events that underlie important cellular or tissue processes, such as angiogenesis (16, 17) and proliferation of certain tumor cells (18, 19).
Circulating levels of thyroid hormone are relatively stable; therefore, membrane-initiated actions of thyroid hormone on neovascularization or on cell proliferation or on membrane ion channels—as well, of course, as gene expression effects of the hormone mediated by TR mentioned above—may be assumed to contribute to ‘basal activity’ or setpoints of these processes in intact organisms. The possible clinical utility of cellular events that are mediated by the membrane receptor for thyroid hormone may reside in inhibition of such effect(s) in the contexts of neovascularization or tumor cell growth. Indeed, we have shown that blocking the membrane receptor for iodothyronines with tetraiodothyroacetic acid (tetrac), a hormone-binding inhibitory analogue that has no agonist activity at the receptor, can arrest growth of glioma cells (19) and of human breast cancer cells in vitro (18). Tetrac is a useful probe to screen for participation of the integrin receptor in actions of thyroid hormone. In this review we will briefly summarize some of the known effects of thyroid hormone that are mediated by the integrin receptor and then concentrate on new directions to explore in the area of membrane receptors for the hormone.
Integrin αVβ3 binds thyroid hormone near the Arg-Gly-Asp (RGD) recognition site of the protein; the RGD site is involved in the protein-protein interactions linking the integrin to extracellular matrix (ECM) proteins such as vitronectin, fibronectin and laminin (13). The intact integrin is structurally very plastic (20). Its conformational changes in response to ligand-binding may underlie its ability to transduce cell surface signals into discrete intracellular messages, as well as the ability to expose new surfaces for interactions. The integrin also generates crosstalk with other cell surface receptors. The thyroid hormone signal at the integrin is transduced into mitogen-activated protein kinase (MAPK) activity via phospholipase C and PKC (21). MAPK (ERK1/2) activation is associated with increased Na⁺/H⁺ antiporter activity locally at the plasma membrane in response to thyroid hormone (15) and we speculate that hormone effects on other ion pumps at the cell surface relate to MAPK or PKC activation. Hormone-activated MAPK also is directed rapidly to the cell nucleus where it may phosphorylate TRβ1 at Ser-142 (22), leading to disruption of the corepressor protein-TR complex and recruitment of co-activators. The fact that this can be achieved with agarose-T₄ that does not cross the cell membrane means that ‘de-repression’ of TR can be instigated from the cell surface without T₃ (or T₄) in the cell nucleus. However, only low (‘basal’) levels of transcription appear to be achieved in this manner and the natural ligand, T₃, apparently must be present in the nucleus in order to achieve multiple-fold increases in transcriptional activity of TR.
Also initiated at the cell surface integrin receptor is the complex process of angiogenesis, monitored in either a standard chick blood vessel assay (16) or with human endothelial cells in a sprouting assay (S Mousa, PJ Davis: unpublished observations). This hormone-dependent process requires MAPK activation and elaboration of basic fibroblast growth factor (bFGF; FGF2) that is the downstream mediator of thyroid hormone’s effect on angiogenesis (16). Tetrac blocks this action of T₄ and T₃, as does RGD peptide and small molecules that mimic RGD peptide. It is possible that desirable neovascularization can be promoted with local application of thyroid hormone analogues, e.g., in wound-healing, or that undesirable angiogenesis, such as that which supports tumor growth, can be antagonized in part with tetrac.
Thyroid hormone can also stimulate the proliferation in vitro of certain tumor cell lines (13). Murine glioma cell lines have been shown to proliferate in response to physiological concentrations of T₄ (19) by a mechanism initiated at the integrin receptor and that is MAPK-dependent. In what may be a clinical corollary, a prospective study of patients with far advanced glioblastoma multiforme (GBM) in whom mild hypothyroidism was induced by propylthiouracil showed an important survival benefit over euthyroid control patients (23). We reported in 2004 that human breast cancer MCF-7 cells proliferated in response to T₄ by a mechanism that was inhibited by tetrac (18). A recent retrospective clinical analysis by Cristofanilli et al. (24) showed that hypothyroid women who developed breast cancer did so later in life than matched euthyroid controls and had less aggressive, smaller lesions at the time of diagnosis than controls. Thus, the trophic action of thyroid hormone on in vitro models of both brain tumor and breast cancer appears to have clinical support.
The cellular or tissue actions of thyroid hormone that are known to be initiated at integrin αVβ3 and that require transduction of the hormone signal via MAPK are summarized in Fig. 1.


Fig. 1. Membrane-initiated actions of thyroid hormone that involve the hormone receptor on integrin αVβ3. The integrin is a signal transducing protein connecting signals from extracellular matrix (ECM) proteins to the cell interior (outside-in) or from cytoplasm and intracellular organelles to ECM (inside-out). Binding of L-thyroxine (T4) or 3,5,3’-triiodo-L-thyronine (T₃) to heterodimeric αVβ3 results in activation of mitogen-activated protein kinase (MAPK; ERK1/2). Activated MAPK (phosphoMAPK, pMAPK) translocates to the cell nucleus where it phosphorylates transactivator proteins such as thyroid hormone receptor-β1 (TRβ1), estrogen receptor-α (ERα) or signal transducer and activator of transcription–lα (STAT1α). Among the genes consequently transcribed are basic fibroblast growth factor (bFGF), that mediates thyroid hormone-induced angiogenesis) and other proliferation factors important to cell division of tumor cells. Depicted in this figure in red is the ability of tetraiodothyroacetic acid (tetrac) to inhibit the action of T₄ and T₃ at the integrin; tetrac blocks the binding of iodothyronines to the integrin receptor. Also shown is crosstalk between the integrin and epidermal growth factor receptor (EGFR). Here, the presence of thyroid hormone at the cell surface alters the function of EGFR to allow the latter to distinguish EGF from TGF-α, another growth factor that binds to EGFR.

When studied as an isolated heterodimeric protein and in contrast to TR, the integrin αVβ3 thyroid hormone receptor has a higher affinity for T₄ than for T₃ (12). Consistent with this observation, T₄ may be more potent than T₃ in MAPK (ERK1/2) activation (25). But physiological concentrations of T₃ are active in MAPK-dependent models of angiogenesis (16) and, in contrast, T₄ is effective only when converted to T₃ in stimulating Na⁺/H⁺ antiporter activity (9). We know this action of the hormone is also MAPK-requiring (15). This spectrum of results suggest that affinities of the integrin for hormone analogues may be different when the integrin is studied as an isolated protein and when it is imbedded in the plasma membrane in experiments involving the intact cell.

New Directions in Characterization of Membrane-Initiated Actions of Thyroid Hormone
Definition of Thyroid Hormone Actions that are Initiated Outside of the Nucleus.
Although a cell surface receptor for iodothyronines has been described, this finding does not exclude the possibility that other mechanisms exist for actions of the hormone that begin or are consummated outside the cell nucleus. For example, plasma membrane transporters for thyroid hormone (27, 28) could conceivably be linked to specific intracellular events. In addition, TR is now appreciated to exist in the cytoplasm, as noted above (29, 30), and TR family members could bind cytoplasmic T₄ or T₃ to initiate effects that are exclusively extranuclear or a premonitory step to genomic actions.
It is also possible that more than one integrin contains a thyroid hormone-binding site. At least seven integrins include an RGD domain and could be candidate heterodimeric receptor proteins for thyroid hormone. We are currently pursuing the possibility that an integrin, clearly not αVβ3, contains a receptor that binds T₃ preferentially. The issue here is whether such a site may support activation of phosphatidylinositol 3-kinase (PI 3-K) activation by T₃. PI 3-K activation by T₃ has been reported by several laboratories (31, 32), but may be initiated by the hormone after it has achieved the cytoplasmic space.

Is there a Requirement for Membrane Integrin Receptor-Directed Posttranslational
Modification of TR prior to Genomic Action of T₃?
Current concepts of nuclear actions of T₃ include the shedding by TR of corepressor proteins and recruitment of coactivators as a consequence of intranuclear complexing of T₃ with TR, resulting in transcriptionally active TR-T₃. As noted above, we have shown that T₄ at the cell surface can cause specific serine phosphorylation of TR and de-repression of the receptor that could be premonitory to the binding of T₃ and full activation of TR. Through knockdown experiments involving the integrin and pharmacologic inhibition of MAPK, we are investigating the possibility that T₄ and T₃ may work cooperatively to promote TR-based transcription.

Life Cycle of the Integrin Receptor for Thyroid Hormone.
Integrin αVβ3 is recycled from the plasma membrane to endosomes by mechanisms that can involve protein kinase B (PKB)/Akt (33) or PKD1 (34) under the direction of platelet-derived growth factor (PDGF). It is not yet known if thyroid hormone can induce integrin recycling. Preliminary studies we have carried out of abundance of αVand of β3 mRNAs in T₄-treated CV-1 cells indicates no increase in either compared to untreated cells (12). This indicates that the hormone does not affect transcription of the monomeric genes, as do DNA microarray surveys of gene transcription in thyroid hormone-treated cells (35, 36).

Clustering of Growth Factor Receptors on the Cell Surface and the Integrin Receptor for Thyroid Hormone.
We have concluded that there is cross-talk between the integrin receptor and the epidermal growth factor receptor (EGFR), based on the ability of T₄ to potentiate the activation of MAPK by transforming growth factor-α (TGF-α), a ligand of EGFR (37), and the ability of tetrac to block the thyroid hormone effect. It will be useful to examine the possibility that signals of insulin-like growth factors and of PDGF are modified by iodothyronines.

Is Integrin Signaling to ECM Proteins Affected by Thyroid Hormone?
In a set of interesting studies a decade ago, Farwell and co-workers showed in vitro that the laminin-integrin interaction of astrocytes was affected by T₄, but not T₃ (38). The process was shown to be RGD peptide sensitive at a time when the existence of an integrin receptor for thyroid hormone was not suspected. It is possible that the action of thyroid hormone on the integrin-laminin interaction required intracellular signal transduction concluding with an inside-out message, but we feel that it is just as likely that the binding of T₄ by the integrin directly induced a conformational change in the integrin that favored interaction with laminin. Such studies could be repeated today with tetrac and with inhibitors of MAPK or of PKC activities to 1) confirm that binding by the integrin of T₄ is the basis of the hormonal effect on laminin and 2) determine whether intracellular signaling is involved.

Actions of Thyroid Hormone on Cell Migration.
Farwell, Leonard and co-workers have recently reported that the rate of migration of neurons is increased by T₄ (39). It is not yet known where in the cell this action of the hormone is initiated. The same group has shown that fibrous actin content of glial cells is increased by T₄ treatment (40), as is that of neurons (39). This stabilization of contractile elements would support cell migration.

Conclusions
A recently-described cell surface receptor for thyroid hormone near the Arg-Gly-Asp (RGD) recognition site on integrin αVβ3 transduces the hormone signal into mitogen-activated protein kinase (MAPK) activation. Consequent MAPK-dependent events occur at the plasma membrane and in the nucleus. At the cell membrane, basal activities or set points of certain ion pumps or channels may be regulated in part by thyroid hormone-directed MAPK. The nuclear thyroid hormone receptor (TRβ1) may be de-repressed from the cell surface by physiological concentrations of L-thyroxine (T₄) and complex cellular activities, such as angiogenesis, may be initiated at the integrin by T₄ or 3,5,3’-triiodo-L-thyronine (T₃). Proliferation of certain tumor cell lines may be stimulated by iodothyronines through the integrin receptor, notably, glioma cells and breast cancer cells. Tetraiodothyroacetic acid (tetrac) inhibits the binding of thyroid hormone to the integrin receptor and thereby blocks the pro-angiogenic effect of the hormone and actions of the latter on tumor cells. In this review we identify areas of further investigation regarding cell surface actions of iodothyronines. These include the possibility of additional receptors in the plasma membrane, proof of cooperativity between derepression of TR from the cell surface receptor by T₄ and important enhancement by nuclear T₃ of transcriptional activity of TR and crosstalk between the hormone receptor and polypeptide growth factor receptors on the cell surface.