Proangiogenic Action of Thyroid Hormone Is Fibroblast Growth Factor–Dependent and Is Initiated at the Cell Surface
The effects of thyroid hormone analogues on modulation of angiogenesis have been studied in the chick chorioallantoic membrane model. Generation of new blood vessels from existing vessels was increased 3-fold by either l-thyroxine (T4; 10−7 mol/L) or 3,5,3′-triiodo-l-thyronine (10−9 mol/L). T4–agarose reproduced the effects of T4, and tetraiodothyroacetic acid (tetrac) inhibited the effects of both T4 and T4–agarose. Tetrac itself was inactive and is known to block actions of T4 on signal transduction that are initiated at the plasma membrane. T4 and basic fibroblast growth factor (FGF2) were comparably effective as inducers of angiogenesis. Low concentrations of FGF2 combined with submaximal concentrations of T4 produced an additive angiogenic response. Anti-FGF2 inhibited the angiogenic effect of T4. The proangiogenic effects of T4 and FGF2 were blocked by PD 98059, a mitogen-activated protein kinase (MAPK) pathway inhibitor. Endothelial cells (ECV304) treated with T4 or FGF2 for 15 minutes demonstrated activation of MAPK, an effect inhibited by PD 98059 and the protein kinase C inhibitor CGP41251. Reverse transcription–polymerase chain reaction of RNA extracted from endothelial cells treated with T4 revealed increased abundance of FGF2 transcript at 6 to 48 hours, and after 72 hours, the medium of treated cells showed increased FGF2 content, an effect inhibited by PD 98059. Thus, thyroid hormone is shown to be a proangiogenic factor. This action, initiated at the plasma membrane, is MAPK dependent and mediated by FGF2.
Control of angiogenesis is a complex process involving local release of vascular growth factors,1 the extracellular matrix, adhesion molecules, and metabolic factors.2 Mechanical forces within blood vessels may also play a role.3 The principal classes of endogenous growth factors implicated in new blood vessel growth are the fibroblast growth factor (FGF) family and vascular endothelial growth factor (VEGF).4 The mitogen-activated protein kinase (MAPK) signal transduction cascade (extracellular signal-regulated kinases 1 and 2 [ERK1/2]) is involved in VEGF gene expression and in control of proliferation of vascular endothelial cells.4
In the setting of l-thyroxine (T4)-induced myocardial hypertrophy in the rat, coordinated angiogenesis and cardiac growth have been described by Chilian et al5 and by Tomanek.6 In the latter report, the appearance of increased capillary numerical density preceded hypertrophy. The same group also showed that diiodothyropropionic acid, a thyroid hormone analog,7 promoted angiogenesis in the hypertrophied heart that was subjected to experimental myocardial infarction.8 The evidence for a proangiogenic effect of systemically administered T4 has been based on histological evidence developed in animal models,5 and the possible mechanisms involved have not been described.
The availability of a chick chorioallantoic membrane (CAM) model of angiogenesis9 provided us with a system to quantitate angiogenesis and to study mechanisms involved in induction of angiogenesis by thyroid hormone. In this article, we describe a proangiogenic effect of T4 at physiological concentrations that approximates that of basic fibroblast growth factor (FGF2) in the CAM model. This effect can enhance the action of submaximal concentrations of FGF2. We also provide evidence that the hormone effect is initiated at the endothelial cell plasma membrane and is mediated by activation of the ERK1/2 signal transduction pathway. The mechanism also involves increased FGF2 expression by endothelial cells and release of FGF2 protein into the cell medium.
Materials and Methods
T4, 3,5,3′-triiodo-l-thyronine (T3), tetraiodothyroacetic acid (tetrac), T4–agarose, and 6-N-propyl-2-thiouracil (PTU) were obtained from Sigma; PD 98059 from Calbiochem; and CGP41251 was a gift from Novartis Pharma (Basel, Switzerland). Polyclonal anti-FGF2 and monoclonal anti-β-actin were obtained from Santa Cruz Biotechnology and human recombinant FGF2 from Invitrogen. Polyclonal antibody to phosphorylated ERK1/2 was from New England Biolabs and goat anti-rabbit IgG from DAKO.
Chick CAM Model of Angiogenesis
Neovascularization was examined by methods described previously.9–12 Ten-day-old chick embryos were purchased from SPAFAS (Preston, CT) and incubated at 37°C with 55% relative humidity. A hypodermic needle was used to make a small hole in the shell concealing the air sac, and a second hole was made on the broad side of the egg, directly over an avascular portion of the embryonic membrane that was identified by candling. A false air sac was created beneath the second hole by the application of negative pressure at the first hole, causing the CAM to separate from the shell. A window ≈1.0 cm2 was cut in the shell over the dropped CAM with a small-crafts grinding wheel (Dremel, division of Emerson Electric Co.), allowing direct access to the underlying CAM. FGF2 (1 μg/mL) was used as a standard proangiogenic agent to induce new blood vessel branches on the CAM of 10-day-old embryos. Sterile disks of No. 1 filter paper (Whatman International) were pretreated with 3 mg/mL cortisone acetate and 1 mmol/L PTU and air dried under sterile conditions. Thyroid hormone, hormone analogues, FGF2 or control solvents, and inhibitors were then applied to the disks and the disks allowed to dry. The disks were then suspended in PBS and placed on growing CAMs. Filters treated with T4 or FGF2 were placed on the first day of the 3-day incubation, with antibody to FGF2 added 30 minutes later to selected samples as indicated. At 24 hours, the MAPK cascade inhibitor PD 98059 was also added to CAMs topically by means of the filter disks.
Microscopic Analysis of CAM Sections
After incubation at 37°C with 55% relative humidity for 3 days, the CAM tissue directly beneath each filter disk was resected from control and treated CAM samples. Tissues were washed 3× with PBS, placed in 35-mm Petri dishes (Nalge Nunc), and examined under an SV6 stereomicroscope (Zeiss) at ×50 magnification. Digital images of CAM sections exposed to filters were collected using a 3-charge–coupled device color video camera system (Toshiba) and analyzed with Image-Pro software (Media Cybernetics). The number of vessel branch points contained in a circular region equal to the area of each filter disk were counted. One image was counted in each CAM preparation, and findings from 8 to 10 CAM preparations were analyzed for each treatment condition (thyroid hormone or analogues, FGF2, FGF2 antibody, PD 98059). In addition, each experiment was performed 3 times. The resulting angiogenesis index is the mean±SEM of new branch points in each set of samples.
ECV304 endothelial cells were cultured in M199 medium supplemented with 10% fetal bovine serum. ECV304 cells (106 cells) were plated on 0.2% gel-coated 24-well plates in complete medium overnight, and the cells were then washed with serum-free medium and treated with T4 or T3 as indicated. After 72 hours, the supernatants were harvested and assays for FGF performed without dilution using a commercial ELISA system (R&D Systems).
ECV304 endothelial cells were cultured in M199 medium with 0.25% hormone-depleted serum13 for 2 days. Cells were then treated with T4 (10−7 mol/L) for 15 minutes to 6 hours. In additional experiments, cells were treated with T4 or FGF2 or with T4 in the presence of PD 98059 or CGP41251. Nuclear fractions were prepared from all samples by our method reported previously,13 the proteins separated by polyacrylamide gel electrophoresis, and transferred to membranes for immunoblotting with antibody to phosphorylated ERK 1/2. The appearance of nuclear phosphorylated ERK1/2 signifies activation of these MAPK isoforms by T4.13–15
Reverse Transcription–Polymerase Chain Reaction
Confluent ECV304 cells in 10-cm plates were treated with T4 (10−7 mol/L) for 6 to 48 hours and total RNA extracted using guanidinium isothiocyanate (Biotecx Laboratories). RNA (1 μg) was subjected to reverse transcription–polymerase chain reaction (RT-PCR) using the Access RT-PCR system (Promega). Total RNA was reverse transcribed into cDNA at 48°C for 45 minutes, then denatured at 94°C for 2 minutes. Second-strand synthesis and PCR amplification were performed for 40 cycles with denaturation at 94°C for 30 s, annealing at 60°C for 60 s, and extension at 68°C for 120 s, with final ex-tension for 7 minutes at 68°C after completion of all cycles. PCR primers for FGF2 were as follows: FGF2 sense strand 5′-TGGTATGTGGCACTGAAACG-3′, antisense strand 5′-CTCAATGACCTGGCGAAGAC-3′; the length of the PCR product was 734 bp. Primers for GAPDH included the sense strand 5′-AAGGTCATCCCTGAGCTGAACG-3′, and antisense strand 5′-GGGTGTCGCTGTTGAAGTCAGA-3′; the length of the PCR product was 218 bp. The products of RT-PCR were separated by electrophoresis on 1.5% agarose gels and visualized with ethidium bromide. The target bands of the gel were quantified using LabImage software (Kapelan), and the value for [FGF2/GAPDH]×10 calculated for each time point.
Statistical analysis was performed by 1-way ANOVA comparing experimental with control samples.
Effect of Thyroid Hormone on Angiogenesis
As seen in Figure 1A and summarized in Figure 1B, both l-T4 and l-T3 enhanced angiogenesis in the CAM assay. T4, at a physiologic total concentration in the medium of 0.1 μmol/L, increased blood vessel branch formation by 2.5-fold (P<0.001). T3 (1 nmol/L) also stimulated angiogenesis 2-fold. The possibility that T4 was only effective because of conversion of T4 to T3 by cellular 5′-monodeiodinase was ruled out by the finding that the deiodinase inhibitor PTU had no inhibitory effect on angiogenesis produced by T4 (data not shown). In the studies included in this article, PTU was applied to all filter disks used in the CAM model. Thus, T4 and T3 promote new blood vessel branch formation in a CAM model that has been standardized previously for the assay of growth factors.
Effects of T4–Agarose and Tetrac
We have shown previously that T4–agarose stimulates cellular signal transduction pathways initiated at the plasma membrane in the same manner as T4 and that the actions of T4 and T4–agarose are blocked by a deaminated iodothyronine analogue, tetrac, which is known to inhibit binding of T4 to plasma membranes.13–15 In the CAM model, the addition of tetrac (0.1 μmol/L) inhibited the action of T4 (Figure 2A), but tetrac alone had no effect on angiogenesis (Figure 2C). The action of T4–agarose, added at a hormone concentration of 0.1 μmol/L, was comparable to that of T4 in the CAM model (Figure 2B), and the effect of T4–agarose was also inhibited by the action of tetrac (Figure 2B; summarized in 2C).
Enhancement of Proangiogenic Activity of FGF2 by a Submaximal Concentration of T4
Angiogenesis is a complex process that usually requires the participation of polypeptide growth factors. The CAM assay requires at least 48 hours for vessel growth to be manifest; thus, the apparent plasma membrane effects of thyroid hormone in this model are likely to result in a complex transcriptional response to the hormone. Therefore, we determined whether FGF2 was involved in the hormone response and whether the hormone might potentiate the effect of subphysiologic levels of this growth factor. T4 (0.05 μmol/L) and FGF2 (0.5 μg/mL) individually stimulated angiogenesis to a modest degree (Figure 3). The angiogenic effect of this submaximal concentration of FGF2 was enhanced by a subphysiologic concentration of T4 to the level caused by 1.0 μg FGF2 alone. Thus, the effects of submaximal hormone and growth factor concentrations appear to be additive.
To define more precisely the role of FGF2 in thyroid hormone stimulation of angiogenesis, a polyclonal antibody to FGF2 was added to the filters treated with either FGF2 or T4, and angiogenesis was measured after 72 hours. Figure 4 demonstrates that the FGF2 antibody inhibited angiogenesis stimulated either by FGF2 or by T4 in the absence of exogenous FGF2, suggesting that the T4 effect in the CAM assay was mediated by increased FGF2 expression. Control IgG antibody has no stimulatory or inhibitory effect in the CAM assay.12
Stimulation of FGF2 Release From Endothelial Cells by Thyroid Hormone
Levels of FGF2 were measured in the media of ECV304 endothelial cells treated with either T4 (0.1 μmol/L) or T3 (0.01 μmol/L) for 3 days. As seen in the Table, T3 stimulated FGF2 concentration in the medium 3.6-fold, whereas T4 caused a 1.4-fold increase. This finding indicates that thyroid hormone may enhance the angiogenic effect of FGF2, at least in part, by increasing the concentration of growth factor available to endothelial cells.
Role of the ERK1/2 Signal Transduction Pathway in Stimulation of Angiogenesis by Thyroid Hormone and FGF2
A pathway by which T4 exerts a nongenomic effect on cells is the MAPK signal transduction cascade, specifically that of ERK1/2 activation.13–15 We know that T4 enhances ERK1/2 activation by epidermal growth factor.16 The role of the MAPK pathway in stimulation by thyroid hormone of FGF2 expression was examined by the use of PD 98059 (2 to 20 μmol/L), an inhibitor of ERK1/2 activation by the tyrosine–threonine kinases MAPK kinase-1 (MEK1) and MEK2.13–15 The data in the Table demonstrate that PD 98059 effectively blocked the increase in FGF2 release from ECV304 endothelial cells treated with either T4 or T3.
Parallel studies of ERK1/2 inhibition were performed in CAM assays, and representative results are shown in Figure 5. A combination of T3 and T4, each in physiologic concentrations, caused a 2.4-fold increase in blood vessel branching, an effect that was completely blocked by 3 μmol/L PD 98059 (Figure 5A). FGF2 stimulation of branch formation (2.2-fold) was also effectively blocked by this inhibitor of ERK1/2 activation (Figure 5B). This result is consistent with reports from other laboratories.17,18 Thus, the proangiogenic effect of thyroid hormone begins at the plasma membrane and involves activation of the ERK1/2 pathway to promote FGF2 release from endothelial cells. ERK1/2 activation is again required to transduce the FGF2 signal and cause new blood vessel formation.
Action of Thyroid Hormone and FGF2 on MAPK Activation
Stimulation of phosphorylation and nuclear translocation of ERK1/2 MAPKs was studied in ECV304 cells treated with T4 (10−7 mol/L) for 15 minutes to 6 hours. The appearance of phosphorylated ERK1/2 in cell nuclei occurred within 15 minutes of T4 treatment, reached a maximal level at 30 minutes, and was still apparent at 6 hours (Figure 6A). This effect of the hormone was inhibited by PD 98059 (Figure 6B), a result to be expected because this compound blocks the phosphorylation of ERK1/2 by MAPK kinase.13,14 The traditional protein kinase C (PKC)-α, PKC-β, and PKC-γ inhibitor CGP41251 also blocked the effect of the hormone on MAPK activation in these cells, as we have seen with T4 in other cell lines.13
Thyroid hormone enhances the action of several cytokines and growth factors, such as interferon-γ13 and epidermal growth factor.16 In ECV304 cells, T4 enhanced the MAPK activation caused by FGF2 in a 15-minute coincubation (Figure 6C). Applying observations made in ECV304 cells to the CAM model, we propose that the complex mechanism by which the hormone induces angiogenesis includes endothelial cell release of FGF2 and enhancement of the autocrine effect of released FGF2 on angiogenesis.
RT-PCR in ECV304 Cells Treated With Thyroid Hormone
The final question addressed in studies of the mechanism of the proangiogenic action of T4 was whether the hormone may induce FGF2 gene expression. Endothelial cells were treated with T4 (10−7 mol/L) for 6 to 48 hours, and RT-PCR–based estimates of FGF2 and GAPDH RNA (inferred from cDNA measurements; Figure 7) were performed. Increase in abundance of FGF2 cDNA, corrected for GAPDH content, was apparent by 6 hours of hormone treatment and was further enhanced by 48 hours.
The CAM assay has been used to validate angiogenic activity of several growth factors and other promoters or inhibitors of angiogenesis.9–12 In the present studies, this model was used to establish that thyroid hormone in a physiologic concentration is proangiogenic. The mechanism by which thyroid hormone causes angiogenesis in the CAM model is conceived by us to be a complex one that is initiated at the plasma membrane by the hormone and involves transduction of the hormone signal into an FGF2-dependent angiogenic response. The basis for the latter conclusion includes (1) increased abundance of FGF2 RNA in endothelial cells with T4 treatment; (2) increased medium content of FGF2 with T4 treatment of endothelial cells; (3) blocking of the angiogenic effect of T4 by addition of anti-FGF2 to the CAM assay; and (4) enhancement by T4 of the angiogenic activity of low concentrations of FGF2 in the CAM model. In studies of T4-induced myocardial hypertrophy in the rat, the hormone has been shown by other investigators to increase abundance of myocardial FGF2 mRNA in 1 to 2 days.19
T4 at 10−7 M total hormone concentration was comparable in angiogenic activity to the maximal FGF2 effect in this model. T3 at 10−9 M was also an effective proangiogenic factor in the CAM model. Although new blood vessel growth in the rat heart has been reported to occur concomitantly with induction of myocardial hypertrophy by a supraphysiologic dose of T4,6 thyroid hormone has not been regarded as an angiogenic factor.20 This article establishes that the hormone in physiologic concentrations is angiogenic in a setting other than the heart.
Because the appearance of new blood vessel branch points in the CAM model requires several days, we assumed from preliminary studies with T4 that the effect of thyroid hormone was wholly dependent on a genomic mechanism. That is, T4 was very likely converted to T3 by cellular 5′-deiodinase activity, and T3 entered the cell and cell nucleus, formed a complex with nuclear thyroid hormone receptor (TRβ1), and initiated transcription of 1 or more genes relevant to angiogenesis. The presence of PTU in the assay system did not diminish the action of T4, however, indicating that production of T3 via deiodination of T4 did not contribute to the action of T4. This did not exclude the possibility that T4 entered the cell and activated TRβ1, but T4–agarose reproduced the effects of T4, and this derivative of thyroid hormone is thought not to gain entry to the cell interior, and has been used in our laboratory to examine models of hormone action for possible cell surface-initiated actions of iodothyronines.13,14 Further, experiments performed with T4 and tetrac also supported the conclusion that the action of T4 in the CAM model was initiated at the plasma membrane. We have shown previously that tetrac blocks membrane-initiated effects of T4 but does not itself have agonist activity at the membrane.13,14
Because we have shown previously that thyroid hormone nongenomically activates the MAPK (ERK1/2) signal transduction pathway, we examined the possibility that the action of the hormone on angiogenesis was MAPK mediated. When added to the CAM model, an inhibitor of the MAPK cascade, PD 98059,13,21 inhibited the proangiogenic action of T4. Although this result was consistent with an action on transduction of the thyroid hormone signal upstream of an effect of T4 on FGF2 elaboration, it is known that FGF2 also acts via a MAPK-dependent mechanism.17,18 In studies presented above, we demonstrate that T4 and FGF2 individually cause phosphorylation and nuclear translocation of ERK1/2 in endothelial cells, and when used in submaximal doses, combine to enhance ERK1/2 activation further. To examine the possibility that the only MAPK-dependent component of hormonal stimulation of angiogenesis related exclusively to the action of FGF2 on vessel growth, we measured cellular release of FGF2 in response to T4 in the presence of PD 98059. The latter agent blocked the hormone-induced increase in growth factor concentration and indicated that MAPK activation was involved in the action of T4 on FGF2 release from endothelial cells, as well as the consequent effect of FGF2 on angiogenesis.
That thyroid hormone is proangiogenic in the CAM model permits speculation about possible clinical significance of the finding. Because ambient concentrations of thyroid hormone in the intact organism are relatively stable, the hormone may be a tonic permissive factor in settings of new blood vessel growth. In the case of angiogenesis in the heart in the presence of arterial narrowing, the proangiogenic effect of thyroid hormone may be desirable. In contrast, in proliferative neovascularization in the eye (eg, diabetic retinopathy), a proangiogenic contribution of the hormone would be undesirable. If thyroid hormone were a permissive factor in this setting, its action might be subject to inhibition by administration of tetrac. Angiogenesis around primary or metastatic tumors might also be tonically stimulated by iodothyronines. This also is undesirable and might be pharmacologically inhibitable. Cellular models are available in which to test these possible consequences of thyroid hormone action.
This work was supported in part by the Department of Veterans Affairs (H.-Y.L., P.J.D.), the Candace King Weir Foundation, and the Charitable Leadership Foundation (Albany, NY).
Original received November 13, 2003; revision received April 21, 2004; accepted April 22, 2004.
Chilian WM, Wangler RD, Peters KG, Tomanek RJ, Marcus ML. Thyroxine-induced left ventricular hypertrophy in the rat. Anatomical and physiological evidence for angiogenesis. Circ Res. 1985; 57: 591–598.
Pennock GD. Cardiac effects of 3,5-diiodothyropropionic acid, a thyroid hormone analog with inotropic selectivity. J Pharmacol Exp Ther. 1992; 263: 163–169.
Davis PJ, Shih A, Lin H-Y, Davis FB. Thyroxine promotes association of mitogen-activated protein kinase (MAPK) and nuclear thyroid hormone receptor (TR) and causes serine phosphorylation of TR. J Biol Chem. 2000; 275: 38032–38039.
Lin H-Y, Shih A, Davis FB, Davis PJ. Thyroid hormone promotes phosphorylation of STAT3 and potentiates the action of EGF in cultured cells. Biochem J. 1999; 338: 427–432.
Collin-Osdoby P, Rothe L, Bekker S, Anderson F, Huang Y, Osdoby P. Basic fibroblast growth factor stimulates osteoclast recruitment, development, and bone pit resorption in association with angiogenesis in vivo on the chick chorioallantoic membrane and activates isolated avian osteoclast resorption in vitro. J Bone Miner Res. 2002; 17: 1859–1871.
Tomanek RJ, Doty MK, Sandra A. Early coronary angiogenesis in response to thyroxine: growth characteristics and upregulation of basic fibroblast growth factor. Circ Res. 1998; 82: 587–593.