Introduction

Tumor-relevant angiogenesis as a clinical target in cancer management became a reality with the work of the Folkman laboratory [1]. Metabolic differences between normal cells and cancer cells have been appreciated in terms of tolerance of hypoxia, dependence on aerobic glycolysis and preference for lowered extracellular pH in the immediate cellular microenvironment, but reduction in blood supply has been conceived to be as destructive to solid tumors as it is to normal tissues. However, the pharmacology of anti-angiogenesis in tumors that achieves clinical goals has been hampered by the complex molecular physiology of angiogenesis and redundancy of the modulators of blood vessel formation available to tumor cells [2, 3]. Initiation and maintenance of vascular supply involves local release of vascular growth factors—e.g., vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF; FGF2), platelet-derived growth factor (PDGF)—and discrete receptors for each of these proteins on the cell surface that transduce specific growth factor signals into intracellular and extracellular angiogenesis-related events. Targeting of one or two of these growth factors or receptors pharmacologically does leave other vascular growth factor options to cancer cells and these include epidermal growth factor (EGF) [4, 5] and insulin-like growth factor-1 (IGF-1) [6–8], in addition to the factors listed above. Function of the growth factor receptors may be modulated by adjacent plasma membrane integrins, such as αvβ3 [9–12], a structural protein of the plasma membrane [13].

Nonpeptide hormones may also support angiogenesis. Estrogen and progesterone are pro-angiogenic in their normal target tissues and estrogen may be a factor in tumor-relevant blood vessel formation in nuclear estrogen receptor (ER)-bearing cancers, such as lung [14, 15]. It is now clear that thyroid hormone is pro-angiogenic via its cell surface receptor on integrin αvβ3. Because this integrin is expressed generously by cancer cells and rapidly dividing blood vessel cells, thyroid hormone (l-thyroxine, T4; 3, 5, 3′-triiodo-l-thyronine, T3) may be an important contributor to blood supply maintenance about cancers. This pro-angiogenic activity may be blocked by tetraiodothyroacetic acid (tetrac), the deaminated metabolite of T4, which is not, itself, pro-angiogenic, but blocks binding of T4 and T3 at the hormone receptor site on αvβ3 [16]. Tetrac and its nanoparticulate formulation have also been shown in the absence of T4 and T3 to act via the integrin to inhibit activity of multiple vascular growth factors. These actions of tetrac and nanoparticulate tetrac are reviewed below.

There are other strategies to consider as points of attack in the vascular infrastructure of solid tumors. Vascular microtubule formation is a function of endothelial cell motility and cell-to-cell adhesion. Interference with these endothelial cell functions in the environment of tumors would be therapeutically desirable. Neovascularization around the tumor requires that the intercellular matrix structure of the existing vascular bed be loosened to permit vascular budding and vessel formation. Factors such as angiopoietin-2 (Ang-2) are premonitory contributors to such changes in vascular beds and when paired with VEGF promote angiogenesis [17]. Thrombospondin 1 (TSP1) gene expression is almost invariably suppressed in cancer cells; the gene product is an endogenous anti-angiogenic factor [18]. Relief of suppression of transcription of TSP1 is desirable in chemotherapy. While it is feasible to block vascular growth receptor function with monoclonal antibody to specific receptors or by interference with crosstalk between integrin αvβ3 and vascular growth factor receptors clustered with the integrin, it would also be desirable to decrease specific vascular growth factor receptor gene expression or local cancer cell release of just-synthesized protein growth factors. Partially successful anti-angiogenesis selects for hypoxia-tolerant tumor cells that are radioresistant. Radioresistance requires an intact cellular mechanism for rapid repair of double-strand DNA breaks induced by radiation; it would be desirable for the anti-angiogenic process to include suppression of radiation defense mechanisms that depend, for example, on cancer cell expression of genes such as hypoxia-inducible factor-1α (HIF-1α) [19]. HIF-1α protein is a nuclear transactivator whose activity is also relevant to VEGF gene expression and other important tumor cell functions [20].

A few inhibitors of vascular growth factors or their receptors that are currently available do interfere with the actions of more than one growth factor or angiogenesis-relevant protein. Ziv-aflibercept complexes with VEGF-A/VEGF-B and placental growth factor (PlGF) [21] and tyrosine kinase inhibitors such as sunitinib affect activities of both VEGF and PDGF [22]. It has seemed unreasonable, however, to expect to develop anti-angiogenic agents for use against cancer that affect the full complement of vascular growth factor defenses mentioned above that are available to blood vessels and tumor cells co-engaged in angiogenesis.

However, the appreciation of the existence of crosstalk between αvβ3 integrin and multiple growth factor receptors—either physically via their extracellular domains or chemically at the level of the plasma membrane or signal transduction systems available to the endofacial surface of the plasma membrane—has encouraged a search for αvβ3-targeted agents that are anti-angiogenic. A specialized domain of the integrin designated the Arg-Gly-Asp (RGD) recognition site can be disordered with cyclic RGD peptides and affect integrin-vascular growth factor receptor communication. In this review, however, we concentrate on a small molecule receptor for thyroid hormone analogues on αvβ3 and links of this receptor to angiogenesis. At this receptor, thyroid hormone analogues relate to transcription of a spectrum of angiogenesis-relevant genes, as well as to crosstalk between the integrin and receptors for at least five vascular growth factors (see next section).

Thyroid hormone and hormone analogues on actions of vascular growth factor cytokines

The mechanism of thyroid hormone-induced neovascularization was described in part a decade ago in the chick chorioallantoic membrane (CAM) assay of angiogenesis [23, 24]. Initiated at the hormone receptor on αvβ3 in blood vessel cells, the effect was shown to depend at least in part upon the transcription of basic fibroblast growth factor (bFGF; FGF2) gene and increased cell release of bFGF with autocrine stimulation of angiogenesis. Unmodified tetrac blocked this agonist action of T4 and T3. We subsequently showed that, in the absence of agonist thyroid hormone, tetrac blocked the activity of VEGF and bFGF added to the CAM assay [25]. It was proposed that disordering of crosstalk between the integrin and VEGFR and bFGFR explained this effect of tetrac. However, VEGF gene expression is also regulated by HIF-1α [26], a transcription factor whose abundance in cells is regulated by thyroid hormone via αvβ3 [27]. We have also shown that PDGF is inactive in the CAM assay in the presence of tetrac (SA Mousa: unpublished). We now know that function of the IGF1R is modulated by thyroid hormone analogues [28], as are activities of the EGFR [29] and transcription of the EGFR gene [30]. Our survey of plasma membrane proteins extracted from a variety of cells for binding of radiolabeled thyroid hormone has disclosed binding only by αvβ3 [24].

Studies of the effects in the CAM assay of a nanoparticulate formulation of tetrac that restricts tetrac to the extracellular space and αvβ3 have confirmed its anti-angiogenic activities. In this formulation, tetrac is covalently bound through a linker to the surface of a 200 nm nanoparticle that precludes cell entry [31]. Thus, a component of the anti-angiogenic properties of Nanotetrac and tetrac relate to integrin-dependent actions on functions of multiple vascular growth factors and of growth factor receptors. Actions of T4, nanoparticulate T4 and Nanotetrac on angiogenesis in the CAM are shown in Fig. 1.

Fig. 1
figure 1

Angiogenic activities of thyroid hormone (T4), nanoparticulate T4 (as T4-agarose, T4-ag) and tetrac in the chick chorioallantoic membrane (CAM) model. CAM assay methodology is as we have previously described [23]. Nanoparticulate T4 cannot enter the cell. The pro-angiogenic actions of T4 and T4-ag are similar in magnitude, indicating initiation of the effects at integrin αvβ3 on the cell surface. Tetrac eliminates the vascular activities of T4 and T4-ag

Full size image

Thyroid hormone analogues and endogenous small pro-angiogenic molecules

We have recently examined the pro-angiogenic activities of bradykinin, angiotensin II (Ang II) and lipopolysccharide (LPS) in the CAM and their responses to Nanotetrac (NT) (Table 1) (see Discussion). Quantitation of vascular branch points in the CAM permits comparison of effectiveness of different agents or of concentrations of single agents. In the case of each of the substances included in the Table, the pro-angiogenic properties were markedly diminished by Nanotetrac. The ability of Nanotetrac to oppose the pro-angiogenic activity of VEGF and bFGF confirmed a previous report [25]. The combination of VEGF, bFGF and tumor necrosis factor-α (TNF-α) was also examined and exhibited the most vigorous neovascularization in this set of experiments. This activity of the combination was inhibited by Nanotetrac.

Table 1 Inhibition of activities of pro-angiogenic factors in the CAM assay by Nanotetrac (NT) (2 μg/CAM)
Full size table

It should also be noted that T3 was more potent than T4 in the assay (Table 1). However, T3 will not contribute materially to angiogenesis when the nonthyroidal illness (NTI) syndrome has complicated clinical cancer and circulating endogenous T3 levels are low [32].

Thyroid hormone analogues and TSP1

A survey of the actions of unmodified tetrac and Nanotetrac on gene expression in breast cancer cells revealed that basally suppressed TSP1 expression in such cells was reversed by these hormone analogues [30]. This is a desirable quality in an agent with other anti-angiogenic properties.

Thyroid hormone analogues and cell motility

Endothelial cell motility is essential to new blood vessel formation. We have examined human endothelial cell motility in a modified Boyden chamber and have found that T4 and T3 (each at 0.1 μM total concentration) increase migration rate of the cells towards a chemical cue (vitronectin) by more than twofold and threefold, respectively (Fig. 2). The T4 concentration yields a physiologic free level of hormone, whereas the T3 level is supraphysiologic. Tetrac (1 μM) blocks cell migration induced by T4 and T3. Integrity of the cytoskeleton is essential to efficient motility and thyroid hormone, specifically T4, has been shown by Farwell and Leonard to support the conversion of soluble actin to F-actin [33]. This activity was described prior to the description of the iodothyronine receptor on integrin αvβ3, but the integrin has been implicated in control of the actin cytoskeleton [34].

Fig. 2
figure 2

Effects of T4 and T3, with and without unmodified tetrac, on human dermal endothelial cell (EC) migration towards a vitronectin (VN) cue. Cell migration assay was by our previously reported method, using a modified Boyden chamber apparatus [55]. Cells were placed in the upper chamber and VN was located in the lower chamber. Relative fluorescence units (RFU) track accumulation of cells in the lower chamber. Unmodified tetrac inhibits the actions of agonist T4 and T3 on cell migration. Standard error bars represent the means of duplicate studies carried out X3

Full size image

Thyroid hormone analogues and vascular microtubule formation and budding

Primordial blood vessel formation involves vascular microtubule formation. Thyroid hormone stimulates microtubule formation in an in vitro model of human endothelial cells [35]. Vascular budding is also demonstrable in ischemic rabbit hindlimb vessels when these vessels are perfused with T4 [36]. Such studies have not specifically been extended to the tumor vascular microenvironment, but these hormone-dependent activities are assumed to occur there.

Thyroid hormone and angiogenesis-relevant chemokines

The pro-angiogenic chemokine ligands CXCL2 and CXCL3 [37] are regulated from integrin αvβ3 by thyroid hormone analogues (GV Glinsky: unpublished). The CX3C chemokine family consists of the CX3CL1/fractalkine ligand and its receptor, CX3CR1, and the genes for these proteins are also differentially regulated by thyroid hormone analogues from the hormone receptor on the integrin [38]. The CX3C axis is relevant to angiogenesis via multiple pathways that ensure maturation and structural integrity of newly-formed microvessels [37]. In the absence of this chemokine ligand and its receptor, neovascularization results in undersized, leaking vessels [39]. Nanotetrac decreases transcription of CX 3 CL1 and CX 3 CR1 in cancer cells [38].

Thyroid hormone analogues and angiogenesis in situ in tumor xenografts

Pro-angiogenic properties of thyroid hormone demonstrated in model systems such as the CAM have also been studied with formulations of the anti-angiogenic thyroid hormone analogue, tetrac, in a variety of human tumor xenografts. The latter include lung [40] and pancreas [41] cancers, renal cell carcinoma (RCC) [42], medullary thyroid carcinoma [43] and follicular thyroid cancer [44]. In all such xenografts, unmodified tetrac and nanoparticulate tetrac have rapidly decreased tumor-related blood vessel formation. Tetrac formulations are anti-angiogenic by several mechanisms, as indicated above, but a critical action of these compounds is antagonism of the pro-angiogenic activity of thyroid hormone agonists, T4 and T3, at the plasma membrane hormone receptor site on integrin αvβ3 [16].

Tumor cell proliferation directed by thyroid hormone that may be relevant to effectiveness of specific anti-angiogenesis therapy

In the presence of physiologic concentrations of thyroid hormone, there is proliferation of a variety of tumor cell lines we have studied [16]. In the context of observations we have reviewed above, pharmacological angioinhibitory cancer strategies directed at single vascular growth factors appear handicapped. That is, monoclonal antibody to VEGF, alone, or TKI treatment that reduces effectiveness only of VEGF and bFGF must work clinically in a pro-angiogenic tumor environment supported by endogenous factors such as thyroid hormone. The latter supports most of the vascular growth factor cytokine axes by several mechanisms, stimulates multiple pro-angiogenic chemokines, increases endothelial cell motility and stabilizes the cytoskeleton of motile and adherent cells. These hormonal effects may be sufficient in the setting of aggressive and/or recurrent tumors to oppose specific anti-angiogenic therapy.

Discussion

It is clear that multiple components of neovascularization are supported by thyroid hormone. In the setting of cancer chemotherapy that is primarily angioinhibitory and directed at one or no more than several vascular growth factor systems—growth factors, themselves, or their receptors—endogenous thyroid hormone is proposed by us to be a confounding influence. That is, actions of the hormone may contribute to diminished drug effectiveness in certain patients whose solid tumors are subjected to anti-angiogenesis therapy. Such patients may be those with circulating free thyroxine (FT4) levels that are in the upper quartile of the reference range or are elevated in the NTI state [32]. Patients with suppressed serum thyrotropin (TSH) levels that are consistent with NTI or with subclinical hyperthyroidism may have blood FT4 concentrations within the reference range that are interpreted by at least one organ, the pituitary gland, as elevated. We propose that such free hormone concentrations are sufficient in certain patients to be multifactorially pro-angiogenic in tumor-relevant vasculature, regardless of the presence of anti-angiogenic agents directed at one or two vascular growth factors. There is some clinical evidence to support this proposition. For example, highly vascular glioblastoma multiforme (GBM) has been shown to respond favorably to induction of mild (‘subclinical’) hypothyroidism [45]. In this GBM study, patients had already exhausted standard therapeutic measures. The vascularity of RCC has encouraged the use of anti-angiogenic therapy in this condition [46]. Interestingly, trials of TKI therapy in RCC patients has shown promise particularly when TKI-induced hypothyroidism has complicated management [47–49]. This raises the possibility that loss of thyroid hormone support for tumor cell and blood vessel cell proliferation is important to management of this condition.

Nanoparticulate tetrac is an antagonist of actions of thyroid hormone at the hormone receptor on αvβ3. The agent has been shown experimentally to have multiple anti-angiogenic actions [16], but has not undergone clinical trial. Nanotetrac inhibits binding of T4 and T3 to αvβ3, but has been shown to have a variety of anti-angiogenic properties in the absence of T4 and T3. That is, it blocks expression of pro-angiogenic chemokine CX 3 C ligand and receptor [38], stimulates expression of anti-angiogenic TSP1 gene that is almost invariably suppressed in cancer cells [30] and inhibits expression of EGFR [30]. As noted above, Nanotetrac disorders the crosstalk between αvβ3 and nearby receptors for VEGF, bFGF and PDGF. Endogenous substances in addition to vascular growth factors that are pro-angiogenic in the tumor microenvironment include Ang-II [50] and bradykinin [51]. The angiogenic activity of these agents is inhibited by Nanotetrac (Table 1). Neovascularization that is induced by LPS is also blocked by Nanotetrac. The pro-angiogenic effect of LPS is relevant to inflammation [52] and inflammation-associated cancer [53, 54].

The preclinical evidence is extensive that thyroid hormone is pro-angiogenic by a variety of mechanisms. We postulate that in some cancer patients in whom specific anti-angiogenic therapy has been ineffective, the angiogenic activity of endogenous thyroid hormone contributes to suboptimal cancer response. As noted above, spontaneous or induced hypothyroidism can favorably change the clinical courses of GBM, breast and RCC. These clinical observations may reflect proliferative effects of thyroid hormone on tumor cells, themselves [16]. However, when we have examined the vascularity of human cancer xenografts, the prompt decrease in tumor volume with anti-thyroid hormone (Nanotetrac) action at integrin αvβ3 has been associated with a 50–60 % decrease in tumor vascularity [40–44]. This is consistent with, but does not establish, the concept that endogenous thyroid hormone may be acting at αvβ3 in the host animals, in concert with tumor secretion of vascular growth factors, to support angiogenesis. However, αvβ3 is expressed generously by dividing blood vessel cells and tumor cells [16], as stated above, and the integrin is not prominent in quiescent endothelial cells in nonmalignant, normal tissues. Vascular scanning of tumors in patients in whom subclinical hypothyroidism has been medically induced—for example, with TKI therapy—will be an initial test of the contribution of thyroid hormone to tumor vascularity, when compared with euthyroid TKI-treated patients.