Upregulation of the angiogenic factors PlGF, VEGF and their ... - Nature

Oncogene (1997) 15, 2687 ± 2698  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

Upregulation of the angiogenic factors PlGF, VEGF and their receptors (Flt-1, Flk-1/KDR) by TSH in cultured thyrocytes and in the thyroid gland of thiouracil-fed rats suggest a TSH-dependent paracrine mechanism for goiter hypervascularization Giuseppe Viglietto1, Annunciata Romano1, Giovanni Manzo1, Gennaro Chiappetta1, Iole Paoletti1, Daniela Califano1, Maria Giulia Galati1, Valeria Mauriello1, Paola Bruni1, Carmine T Lago2, Alfredo Fusco3 and M Graziella Persico2 1

Istituto Nazionale dei Tumori `Fondazione Senatore Pascale', Via M Semmola, 80131 Naples; 2International Institute of Genetics and Biophysics, CNR, Via G Marconi, 12, 80125 Naples; 3Dipartimento di Medicina Sperimentale e Clinica, FacoltaÁ di Medicina e Chirurgia di Catanzaro, UniversitaÁ degli Studi di Reggio Calabria, Italy

Placenta growth factor (PlGF) and vascular endothelial growth factor (VEGF) represent two closely related angiogenic growth factors active as homodimers or heterodimers. Since goiters of the thyroid gland are extremely hypervascular, we investigated the expression of PlGF, VEGF and their receptors, Flt-1 and Flk-1/ KDR, in a small panel of human goiters from patients with Graves's disease, in an animal model of thyroid goitrogenesis and in in vitro cultured thyroid cells. Here we report that the mRNA expression of PlGF, VEGF and their receptors is markedly enhanced in biopsies of goiters resected from Graves's patients. In vivo studies demonstrated that in the thyroid gland of thiouracil-fed rats, increased mRNA and protein expression of PlGF, VEGF, Flt-1 and Flk-1/KDR occurred subsequent to the rise in the serum thyroid stimulating hormone (TSH) levels and in parallel with thyroid capillary proliferation. In vitro studies con®rmed the existence of such TSHdependent paracrine communication between thyroid epithelial cells and endothelium since the conditioned medium collected from TSH-stimulated thyrocytes acquired mitogenic activity for human umbilical vein endothelial (HUVE) cells. Altogether, these data suggest that PlGF and VEGF, released by thyrocytes in response to the chronic activation of the TSH receptor pathway, may act through a paracrine mechanism on thyroid endothelium. Keywords: PlGF; VEGF; TSH; thyroid goiter; hypervascularization

Introduction In the thyroid gland, thyroid enlargement due to nontoxic goiters or diuse goiters of Graves's disease is often characterized by follicular cell hyperplasia and extensive hypervascularity with abnormally enlarged blood vessels (Wilson and Foster, 1989). Moreover, experimental increase of the circulating thyroid stimulating hormone (TSH) by dietary-induced iodide de®ciency and/or anti-thyroid drugs determines thyroid cell hyperplasia and marked hypervascularization likely Correspondence: MG Persico Received 2 May 1997; revised 25 July 1997; accepted 25 July 1997

to compensate iodide de®ciency (Wollman et al., 1978; Wynford-Thomas et al., 1982; Smeds and Wollman, 1983). Previous studies have shown that [3H]thymidine labeling is observed in thyroid capillaries, veins, arteries and lymphatic vessels within a few days of propylthiouracil (PTU) administration (Smeds and Wollman, 1983); however, the most prominent eects observed are enlargement and fusion of interfollicular vessels which result in increased blood ¯ow rate and enhanced vascular permeability (Wollman et al., 1978). A paracrine mediator produced by thyrocytes in response to the increased blood levels of TSH aecting the growth and/or permeability of endothelial cells (EC) was postulated by Dumont and co-workers (1992). The recent identi®cation of the EC mitogens vascular endothelial growth factor (VEGF) and placenta growth factor (PlGF) in normal adult human (Sato et al., 1995; Viglietto et al., 1995) and mouse (DiPalma et al., 1996) thyroid suggests that these growth factors may be potentially involved in the remodeling of thyroid vasculature during goiter development. VEGF is a 43 ± 46 kDa secreted, dimeric Nglycoprotein that is chemoattractant and mitogenic for EC in vitro (Ferrara and Henzel, 1989; Keck et al., 1989), elicits in vivo angiogenesis and increases the permeability of the vascular endothelium (Connolly et al., 1989). PlGF is a secreted, dimeric glycoprotein of 46 ± 50 kDa, highly similar to VEGF, shown to be chemotactic and mitogenic for EC in vitro (Maglione et al., 1991; Hauser and Weich, 1993; Sawano et al., 1996) and angiogenic in vivo (Ziche et al., 1997). Heterodimers between VEGF and PlGF have also been isolated from the conditioned medium collected from several human and rat tumor cell lines (DiSalvo et al., 1995; Cao et al., 1996). VEGF homodimers bind to and induce autophosphorylation of both the high anity Flt-1 (kd, 16 pM) and Flk-1/KDR (kd, 410 pM) (Shibuya et al., 1990; Terman et al., 1991; De Vries et al., 1992) tyrosine kinase receptors; conversely, PlGF homodimers bind only to the Flt-1 receptor (kd, 160 pM) (Kendall et al., 1994; Terman et al., 1994; Cao et al., 1996; Sawano et al., 1996). Puri®ed heterodimeric VEGF/PlGF has been shown to bind to the Flk-1/KDR receptor (Cao et al., 1996).

VEGF and PIGF in goiter hypervascularization G Viglietto et al

2688

To investigate the potential role of PlGF and VEGF in human goiters, we studied the mRNA expression of these angiogenic factors in a small panel of thyroid goiters resected from patients with Graves's disease. Furthermore, to elucidate the molecular mechanisms underlying vascular remodeling during thyroid goitrogenesis, we investigated the expression of PlGF, VEGF and their receptors, Flt-1 and Flk-1/KDR, in an animal model of thyroid goitrogenesis and in vitro cultured thyroid cells. The ®ndings reported in this study indicated that, in addition to VEGF (Sato et al., 1995), PlGF may play a pivotal role during thyroid hormone-dependent angiogenesis, likely through a paracrine mechanism.

Results

1 PIGF

VEGF

Flk-1/KDR

2

3

4

5

6

7

8 — 18 S

— 28 S

— 28 S

Flt-1 — 28 S

— 28 S

Analysis of PlGF and VEGF mRNA expression in human thyroid goiters The role of soluble angiogenic factors in the development of thyroid goiters was analysed in 7 goiters from patients with Graves's disease for PlGF, VEGF, Flt-1 and Flk-1/KDR expression by Northern blot hybridization. The results are reported in Figure 1. The normal thyroid tissue expressed low levels of PlGF and VEGF transcripts of 1.8 and 3.7 Kb, respectively (Figure 1, lane 1 in the ®rst and second panels, respectively), and almost undetectable levels of Flt-1 and Flk-1/KDR mRNAs (Figure 1, lane 1 in the third and fourth panels, respectively). Conversely, we observed steady-state mRNA level, as measured by densitometry, of PlGF (2 ± 6-fold) in 6/7 goiters, of VEGF (1.6 ± 4-fold) in 5/7 goiters, of Flt-1 (1.7 ± 4-fold) and of Flk-1/KDR (1.6 ± 2.6-fold) in 4/7 goiters. These ®ndings suggest that a simultaneous overexpression of PlGF and VEGF and their receptors occurs in human thyroid goiters. Eect of PTU treatment on PlGF, VEGF, Flt-1 and Flk-1/KDR mRNA expression in rat thyroid gland In Graves's disease, normal thyroid regulatory mechanisms are overridden by the stimulatory action of abnormal auto-immunoglobulins (TSAb), which results in hyperfunction of the thyroid and mimics TSH action (Wilson and Foster, 1989). Therefore, to investigate the role of PlGF and VEGF in the vascular remodeling that accompanies thyroid goiter development, we used PTU-fed rats as an in vivo model, since such an experimental model has been successfully used in a number of thyroid goitrogenesis studies, including vascular remodeling (Wollman et al., 1978; WynfordThomas et al., 1982; Smeds and Wollman, 1983; Rognoni et al., 1984; Dumont et al., 1992; Sato et al., 1995). In PTU-fed rats the TSH serum level increased signi®cantly (day 0, 0.70+0.12 ng/ml; day 3, 2.60+ 0.21 ng/ml; day 8, 5.2+0.15 ng/ml; day 16, 5.6+ 0.11 ng/ml), accompanied by a marked enlargement and weight increase of the thyroid glands (day 0, 16.7+1.8 mg; day 3, 20.4+3.1 mg; day 8, 29.2+ 4.7 mg; day 16, 48.2+3.6 mg), and a marked reduction in the level of thyroid hormones T3 (from 0.7 ng/ml at day 0 to 0.2 ng/ml at day 16) and T4 (from 5.1 mg/dl at

— 18 S Figure 1 Analysis of PlGF, VEGF, Flt-1 and Flk-1/KDR gene expression in human thyroid goiters by Northern blot. Twenty micrograms of total RNA derived from normal thyroid (lane 1) and goiters from patients with Graves's disease (lanes 2 ± 8) was loaded in each lane. The position of 18S and 28S ribosomal RNA is shown. The bottom portion of the Figure shows the ethidium bromide staining of the gel

day 0 to 1.1 mg/dl at day 16). Figure 2 is a representative hematoxylin-eosin staining of thyroid sections, showing the vascular bed of the thyroid gland from rats fed with PTU for 16 days. The marked proliferation and enlargement of thyroid vessels evident from day 3 (Figure 2b) became more pronounced at day 8 (Figure 2c) and day 16 (Figure 2d). These results are in agreement with previous studies that have demonstrated that thyroid EC growth mainly occurs at days 3 ± 8 of PTU treatment (Wynford-Thomas et al., 1982; Smeds and Wollman, 1983). Thyroid glands of control rats expressed low levels of PlGF and VEGF (Figure 3, lane 1) mRNAs. PlGF mRNA expression started to increase at day 3 of PTU treatment with a peak at day 8 and remained elevated until day 16 (Figure 3, lanes 2 ± 4). Compared to untreated control rats, the mRNA expression of VEGF was increased by day 8 of PTU treatment (Figure 3, lanes 2 ± 4). However, the steady-state level of VEGF mRNA was gradually reduced at day 16 despite the high levels of serum TSH (see legend to Figure 3). The kinetics of VEGF and PlGF upregulation induced by increased levels of TSH in PTU-fed rats overlapped the timing of EC proliferation. These results may indicate that both VEGF and PlGF play a key role in the vascular remodeling occurring during thyroid goitrogenesis. We also determined the mRNA expression of Flt-1 and Flk-1/KDR receptors in the thyroid glands of PTU-fed and control rats. The steady-state levels of both Flt-1 and Flk-1/KDR (Figure 3) transcripts were increased at day 8 of PTU-stimulation and gradually decreased by day 16, though remaining higher than untreated thyroid.


2689

Figure 2 Haematoxylin-eosin staining of thyroid sections from normal rats (a), and after 3 (b), 8 (c) and 16 (d) days of PTU treatment. Blood vessel enlargement begins 3 days after PTU treatment and becomes more evident 8 and especially 16 days after PTU treatment

Immunostaining of thyroids of PTU-fed rats with antiFlt-1 and anti-Flk-1/KDR antibodies The enhanced mRNA expression of VEGF and PlGF receptors may re¯ect a marked increase in the number of vessels of PTU-stimulated thyroid glands or a positive regulation of Flt-1 and Flk-1/KDR mRNAs in EC indirectly exerted by TSH or, more likely, both. We therefore performed immunohistochemical staining of blood vessels in thyroids from normal, from PTUfed and from PTU-fed and subsequent iodide refed rats with anti-Flt-1 and anti-Flk-1/KDR IgGs. Figure 4 shows a representative experiment. Flt-1 and Flk-1/ KDR receptors are expressed at low levels in the vessels of untreated rats (Figure 4a and b, respectively). Moreover, staining is not homogeneous since several capillaries do not show positive reaction. In the hypervascular thyroids of PTU-treated rats, not only staining for Flt-1 and Flk-1/KDR appeared much more intense but most of the vessels stained positive for both antibodies at 8 days (Figure 4c and d, respectively) and 16 days of PTU treatment (Figure 4e and f, respectively). Staining with anti-Flt-1 and antiFlk-1/KDR IgGs was essentially restricted to EC although sometimes staining for both receptors was detected in thyrocytes. Control reactions, performed for each experiment by omitting the primary antibodies, showed no staining (data not shown). These

results strongly suggest that increase in the TSH level determines, most likely as an indirect consequence, an increased amount of mRNA and proteins encoding the Flt-1 and Flk-1/KDR receptors in the EC of stimulated thyroid. Mitogenic response of cultured endothelial HUVE cells to the conditioned medium collected from TSHstimulated PC Cl 3 thyrocytes A paracrine communication involving growth of blood vessels has been shown to play an important role during follicle maturation in the ovary and in the endometrium (Shweiki et al., 1996). Furthermore, previous work by Goodman and Rone (1987) has postulated the existence of a paracrine communication mechanism mediated by soluble growth factors between parenchymal thyroid cells and capillaries in the stroma. Accordingly, our in vivo data imply that during goiter development a paracrine mechanism, possibly mediated by PlGF and VEGF, may be responsible for TSH-dependent angiogenesis. To determine the existence of such a putative paracrine communication mechanism between thyroid follicular cells and capillaries of the vascular stroma, we investigated whether TSH stimulation of thyroid cells could enhance the secretion of endothelial mitogens. To this aim we measured [3H]thymidine


2690

PTU lodide Days

– –

+ –

+ –

+ –

0

3

8

16 — 18 S

PIGF

— 28 S VEGF

Flk-1/KDR — 28 S

Flt-1 — 28 S

— 28 S

— 18 S Figure 3 Analysis of the mRNA expression of PlGF, VEGF, Flk1/KDR and Flt-1 in the thyroid gland of PTU-treated goitrous rats by Northern blot. Lane 1: normal thyroid gland from control rats; lanes 2 ± 4: thyroid gland from rats treated with PTU for 3, 8 and 16 days, respectively. Twenty micrograms of total RNA was loaded in each lane. The position of 18S and 28S ribosomal RNA is shown. The bottom portion of the Figure shows the ethidium bromide staining of the gel. The ratio of PlGF, VEGF, Flt-1 and Flk-1/KDR mRNAs to 18S RNA in PTU-fed rats compared to control rats (1 at day 0) was 1.8, 6.4 and 4.1-fold for PlGF; 1.6, 4.1 and 1.9-fold for VEGF; 1.7, 3.4 and 2.7-fold for Flt-1; 1.5, 3 and 1.8-fold for Flk-1/KDR, on day 3, 8 and 16, respectively

incorporation in EC in response to the conditioned medium collected from TSH-treated thyrocytes (CMTSH), and determined the growth rate of EC after stimulation with CM-TSH. In this set of experiments we used the rat thyroid TSH-dependent PC Cl 3 cell line (Fusco et al., 1987) and HUVE cells. Usually, HUVE cells are grown in DMEM supplemented with 20% FCS and endothelial cell growth supplement (ECGS); however, when seeded in DMEM supplemented with 10% FCS without ECGS, HUVE cells arrest their growth (Figure 5a and b). We found, in a [3H]thymidine incorporation assay, that the addition of increasing amounts (10%, 25%, and 50% vol/vol to DMEM, respectively) of CM-TSH to

subcon¯uent HUVE cells stimulated DNA synthesis in a dose-dependent manner (Figure 5a). Compared to conditioned media collected from untreated PC Cl 3 cells (CM), addition of 25% and 50% of CM-TSH to culture medium stimulated [3H]thymidine incorporation in HUVE cells up to 200%. In a complementary set of experiments, we determined the growth rate of HUVE cells after stimulation with CM or CM-TSH. A representative result, with the mean of three experiments performed in quadruplicate is shown in Figure 6b. After stimulation with 50% CM-TSH, the number of HUVE cells almost doubled after 4 days' treatment as compared to cells treated with 50% CM. The eect exerted by CM-TSH on the growth rate of HUVE cells was dose-dependent, since addition of 25% of CMTSH to the culture medium induced an increase in the cell number of 162% (data not shown). These results demonstrate that cultured rat thyrocytes synthetize and secrete endothelial growth factors in response to TSH stimulation, thus con®rming the existence of a TSHdependent paracrine communication between thyroid epithelial cells and EC. To demonstrate that among the plethora of angiogenic growth factors, PlGF could be involved in the TSH-dependent angiogenic response of the thyroid gland, we ®rst determined whether PlGF and VEGF proteins were present in CM-TSH but not in CM. Since the role of VEGF but not that of PlGF in the TSH-dependent vascularization of the thyroid gland has already been investigated (Sato et al., 1995), we subsequently investigated whether anti-PlGF antibodies could abrogate the endothelial mitogenic activity shown by CM-TSH. Western blot experiments demonstrated that increased levels (threefold) of PlGF protein were detected in the CM-TSH compared to the CM. In fact, anti-PlGF IgGs detected a 26 ± 28 kDa protein corresponding to rat PlGF monomer in CM-TSH but not in CM collected from untreated cells (Figure 6a). In addition, immunoprecipitation experiments performed with anti-VEGF antibodies demonstrated increased levels of two VEGF isoforms (approximately 21 and 25 kDa, likely corresponding to VEGF120 and VEGF164, respectively) in CM-TSH compared to CM (Figure 6b); no signal was detected when preimmune rabbit antibodies in the immunoprecipitation assay were used (data not shown). [3H]thymidine incorporation assays and growth rate determinations with CM and CM-TSH, respectively, were performed in the presence and absence of antibodies elicited to recombinant mouse PlGF protein. The results reported in Figure 6a and b demonstrate that a consistent portion of the endothelial mitogenic activity of CM-TSH could be abrogated by anti-PlGF antibodies, both in the [3H]thymidine incorporation and growth rate assays. The ®nding that the anti-PlGF antibody had no eect on the [3H]thymidine incorporation and growth rate of HUVE cells induced by 20 ng/ml of basic ®broblast growth factor (bFGF, data not shown), demonstrated that the results obtained with the anti-PlGF antibody was not due to any nonspeci®c eect. As an additional control, we used a dierent unrelated antibody (elicited against CRIPTO protein; Brandt et al., 1994) to mockneutralize the endothelial mitogenic activity of CMTSH and found that anti-CRIPTO antibody had no


2691

Figure 4 Analysis of the expression of the KDR and Flt-1 receptors by immunostaining in the thyroid gland of normal and PTUtreated rats. Paran sections from normal thyroid tissue were analysed using anti-Flt-1 (a, c, e) and anti-Flk-1/KDR (b, d, f) antibodies. Magni®cation was 4006. (a) In normal thyroid tissue, some vessels showed staining for Flt-1 antibodies. However, staining was heterogeneous and some adjacent small vessels did not stain with anti-Flt-1 antibody. (b) In normal thyroid tissue, weak staining for Flk-1/KDR was detected in some interfollicular blood vessels. Eight days after PTU treatment, most of the vessels in the stimulated thyroid gland stained consistently either with anti-Flt-1 (c) or anti-Flk-1/KDR (d) antibodies. Sixteen days after treatment, staining was further increased either with anti-Flt-1 (e) or anti-KDR (f) antibodies.

eect on the endothelial mitogenic activity of CM-TSH (data not shown). Eects of TSH on VEGF and PlGF mRNA expression in rat cultured thyrocytes To investigate the molecular mechanisms underlying TSH regulation of VEGF and PlGF mRNA expression, we used the rat thyroid cell line PC Cl 3 as an in vitro model. Cells were treated for 3 h with 0.01, 0.1 and 1 mU/ml of TSH. Figure 7b shows that both PlGF (Figure 7b, upper panel) and VEGF (Figure 7b, middle panel) transcripts were induced by TSH treatment in a dose-dependent manner. The time-course of TSH induction (1 mU/ml) of VEGF and PlGF appeared to be very dierent (Figure 7a). The increase in VEGF

mRNA levels occurred very early (Figure 7a, middle panel), with a peak after 30 ± 60 min, whereas PlGF transcript levels started to increase slowly and peaked at 3 ± 6 h from the beginning of the treatment (Figure 7a, upper panel), suggesting that dierent molecular mechanisms may regulate the mRNA expression of VEGF and PlGF. TSH modulates thyroid functions by binding to the high anity receptor (TSHR) on the thyrocyte surface; receptor activation triggers the cAMP-dependent protein kinase pathway (PKA) (Rognoni et al., 1984; Yun et al., 1986; Van Sande et al., 1975). In agreement with the results obtained with TSH, the mRNA expression of VEGF and PlGF was increased in thyroid cells also by treatment with forskolin, at a concentration of 25 mM (data not shown). Since


2692

CM

CM-TSH

CM

CM-TSH

b

a

— 30

— 21

— 30

— 21

Figure 6 (a) Western blot analysis of PlGF protein in the conditioned medium collected from TSH-stimulated (CM-TSH) or unstimulated (CM) rat PC Cl 3 thyroid cells. Media from TSH-stimulated or unstimulated rat PC Cl 3 thyroid cells were analysed on SDS/12.5% polyacrylamide gel followed by immunoblot. Position and molecular masses are as follows: carbonic anhydrase, 30 kDa; trypsin inhibitor, 21 kDa. (b) Immunoprecipitation of 35S-labeled VEGF protein from the conditioned medium collected from TSH-stimulated (CM-TSH) or unstimulated (CM) rat PC Cl 3 thyroid cells. Aliquots of CMTSH and CM wre immunoprecipitated using either immune antiVEGF (lane 1) or preimmune (data not shown) antibodies. Samples were loaded on SDS/12.5% polyacrylamide gel under reducing conditions. Bands of approximately 21 and 25 kDa, likely corresponding to VEGF120 and VEGF164 were speci®cally detected in CM-TSH but not in CM. Position and molecular masses are as follows: carbonic anhydrase, 30 kDa; trypsin inhibitor, 21 kDa

forskolin activates the cAMP-dependent protein kinase pathway by elevating the intracellular levels of cAMP, the ®nding that cultured thyrocytes upregulate the mRNA expression of VEGF and PlGF in response to forskolin, in a time and dose-dependent manner, suggest that TSH stimulation of VEGF and PlGF gene expression may occur through the activation of the protein kinase A-dependent pathway.

Figure 5 Eects of TSH and neutralizing anti-PlGF antibody on the EC mitogenic activity of CM-TSH. (a) [3H]thymidine incorporation assay. Cells were treated as follows: black bars, increasing doses (0.1, 0.25 and 0.5 ml) of conditioned medium collected from untreated thyroid cells were added to DMEM medium containing a ®nal concentration of 10% FCS, and then incubated overnight with [3H]thymidine; hatched bars, increasing doses (0.1, 0.25 and 0.5 ml) of conditioned medium collected from TSH-stimulated PC Cl 3 cells were added to DMEM medium containing a ®nal concentration of 10% FCS in the absence of anti-PlGF antibodies, and then incubated overnight with [3H]thymidine; white bars, increasing doses (0.1, 0.25 and 0.5 ml) of conditioned medium collected from TSH-stimulated PC Cl 3 cells were added to DMEM medium containing ®nal concentration of 10% FCS in the presence of 200 ng/ml of antiPlGF antibodies, and then incubated overnight with [3H]thymidine. Radioactivity incorporated into the TCA-insoluble material was determined from three dierent experiments performed in duplicate. (b) Cell growth assay. A representative result, with the mean of three experiments performed in quadruplicate is shown. HUVE cells were stimulated with 0.5 ml of conditioned medium collected either from untreated or from TSH-stimulated PC Cl 3 cells in DMEM medium containing ®nal concentration of 10% FCS, in the presence or absence of anti-PlGF (200 ng/ml), and then counted after 2 and 4 days' treatment. (&) Treatment with the conditioned medium from unstimulated PC Cl 3 cells; (&) treatment with the conditioned medium from TSH-stimulated PC Cl 3 cells in the absence of antiPlGF IgGs; (*) treatment with the conditioned medium from TSH-stimulated PC Cl 3 cells in the presence of anti-PlGF IgGs

Eects of RNA and protein synthesis inhibitors on VEGF and PlGF mRNA To study further the molecular mechanisms of VEGF and PlGF mRNA upregulation by TSH, we measured the eects of the 5,6-dichloro-1-b-D-ribofuranosylbenzimidazole (DRB) RNA polymerase inhibitor (25 mg/ ml) on the transcription of VEGF and PlGF mRNAs exerted by TSH (Figure 8). In the presence of DRB, the TSH-induced increase in the steady-state levels of either PlGF (Figure 8a, upper panel) or VEGF (Figure 8a, middle panel) was almost completely abrogated, suggesting that TSH stimulation of VEGF and PlGF may occur at the transcriptional level. The rapid increase of VEGF mRNA induced by TSH and its subsequent decline in thyrocytes suggest a rapid turnover of VEGF mRNA, similar to that of early genes such as c-fos. Conversely, the delayed increase in PlGF mRNA induced by TSH in thyroid cells and subsequent slower mRNA turnover suggest a higher stability of PlGF mRNA compared to VEGF. Dierent mechanisms may be responsible for the regulation of the mRNA expression of these two genes in thyroid cells. To characterize further the mechanism for TSH upregulation of VEGF and PlGF mRNAs, we examined the eect of the protein synthesis inhibitor cycloheximide (Figure 8b). In PC Cl 3 cells, the


2693

a

b TSH (1 mU / ml) Hours

0

0.5

1

3

6

12

TSH ( mU / ml) 24

0

0.01

0.1

1

— 18 S

— 18 S

—28 S

—28 S

— 28 S

— 28 S

—18 S

—18 S

Figure 7 Eect of TSH on PlGF and VEGF mRNA expression in PC Cl 3 thyroid cells. (a) Time-course analysis of TSH regulation of PlGF and VEGF mRNA expression in the rat thyroid PC cell line. Twenty micrograms of total RNA was loaded in each lane. Upper panel: PlGF steady-state mRNA levels are strongly increased by treatment of PC cells with TSH (lane 1, untreated PC cells; lanes 2 ± 6, TSH treatment for the indicated time); middle panel: VEGF steady-state mRNA levels are strongly induced by treatment of 3-day-starved PC cells with TSH (1 mU/ml) (lane 1, untreated PC cells; lanes 2 ± 6, TSH treatment for the indicated time). (b) Dose-dependent analysis of TSH regulation of PlGF and VEGF mRNA expression in the rat thyroid PC cell line. Twenty micrograms of total RNA was loaded in each lane. Upper panel: PlGF steady-state mRNA levels are strongly increased by treatment of 3-day-starved PC cells with increasing amounts of TSH (0.01, 0.1 and 1 mU/ml (lane 1, untreated PC cells; lanes 2 ± 4, treatment with the indicated concentration of TSH); middle panel: VEGF steady-state mRNA levels are strongly increased by treatment of PC cells with TSH (lane 1, untreated PC cells; lanes 2 ± 4, treatment with the indicated concentration of TSH). The bottom portion of the Figure shows the ethidium bromide staining of the gel

increase in VEGF mRNA after 1 h of treatment with TSH was potentiated by cycloheximide (5 mg/ml), which alone had no gross eect on VEGF mRNA. Furthermore, at 5 h, when VEGF mRNA started to decline to basal levels, cycloheximide maintained increased VEGF mRNA levels, suggesting that cycloheximide prevented the degradation of VEGF mRNA (Figure 8b, middle panel). Conversely, cycloheximide treatment dramatically reduced PlGF mRNA upregulation induced by TSH (Figure 8b, upper panel), suggesting that in the case of PlGF, protein synthesis is necessary for PlGF gene transcription. The eects of the RNA polymerase inhibitor DRB on the transcription of VEGF and PlGF genes induced by TSH, suggest that TSH is able to stimulate the transcriptional rate of both the VEGF and PlGF genes. However, to determine whether TSH exerts a transcriptional control on VEGF and PlGF genes, we performed run-on transcription assays on nuclei isolated from cultured rat thyrocytes treated with TSH (Figure 9). As shown in the ®gure, PC Cl 3 cells express low basal level of PlGF. However, the transcriptional rate of the PlGF gene increases 7.3-fold after TSH treatment, thus con®rming the experiments performed with DRB, which suggested that TSH may regulate

PlGF gene expression by controlling PlGF transcriptional rate. In the case of VEGF, we observed that the increase in the transcriptional rate was only 2.6-fold, suggesting that, in addition to a transcriptional control in the VEGF gene, also post-transcriptional regulatory mechanisms may be involved in the regulation of the VEGF gene expression exerted by TSH. Control experiments (using pGEM 4Z and the plasmid carrying the insert for 18S rRNA) demonstrated that the increase in the hybridization signal observed in the lanes containing plasmids carrying the PlGF and VEGF inserts, hybridized with nuclear RNA extracted from TSH-treated cells compared to untreated cells, was not due to a nonspeci®c eect. Altogether, these data suggest that the regulation of the expression of the PlGF and VEGF genes induced by TSH in thyroid cells occurs at least in part at the transcriptional level. Discussion In patients with Graves's disease or with nontoxic goiters the thyroid gland is dramatically hypervascularized, so that a venous hum is usually heard (Wilson


2694

a

b TSH DRB Hours

– –

– +

+ –

+ –

+ + + +

0

0

3

6

3

TSH CHX Hours

6

– – 0

– + 0

+ + – – 0.5 1

+ – 5

+ + 0.5

+ + 1

+ + 5

— 18 S

— 18 S

— 28 S

— 28 S

— 28 S

— 28 S

— 18 S — 18 S

Figure 8 The eects of mRNA and protein synthesis inhibitors on PlGF and VEGF mRNA expression in PC Cl 3 thyroid cells. (a) The eect of DRB on TSH induction of PlGF and VEGF mRNA expression in PC Cl 3 thyroid cells. PC Cl 3 cells were treated with 1 mU/ml of TSH for 3 and 6 h in the presence and absence of DRB (25 mg/ml). Total RNA (20 mg/lane) was analysed by Northern blot using PlGF (upper panel) or VEGF (middle panel) as probes. Lane 1, untreated PC cells; lane 2, PC cells treated with 25 mg/ml of DRB; lanes 3 and 4, TSH treatment for the indicated time in the absence of DRB; lanes 5 and 6, TSH treatment for the indicated time in the presence of 25 mg/ml of DRB. (b) The eect of cycloheximide (CHX) on TSH induction of PlGF and VEGF mRNA expression in PC Cl 3 thyroid cells. PC Cl 3 cells were treated with 1 mU/ml of TSH for 0.5, 1 and 5 h in the presence and absence of cycloheximide (5 mg/ml). Total RNA (20 mg/lane) was subjected to Northern blot analysis using PlGF (upper panel) or VEGF (middle panel) as probes. Lane 1, untreated PC cells; lane 2, PC cells treated with cycloheximide; lanes 3 ± 5, TSH treatment for 0.5, 1 and 5 h in the absence of cycloheximide; lanes 6 ± 8, TSH treatment for 0.5, 1 and 5 h in the presence of cycloheximide. The bottom portion of the Figure shows the ethidium bromide staining of the gel

and Foster, 1989). Since PlGF and VEGF have been shown to be chemotractant and mitogenic for EC in vitro, and to elicit angiogenesis in vivo (Connolly et al., 1989; Ferrara and Henzel, 1989; Keck et al., 1989; Maglione et al., 1991; Hauser and Weich, 1993; Sawano et al., 1996; Ziche et al., 1997), we investigated the role of such angiogenic growth factors in the hypervascularization occurring during goiter development. In the present study, we demonstrate that the mRNA expression of VEGF and PlGF and their receptors is markedly enhanced, compared to normal thyroid, in a small panel of thyroid goiters from patients with Graves's disease suggesting that hypervascularity in human goiters may result from coordinate PlGF and VEGF secretion by thyrocytes and from enhanced expression of their receptors Flt-1 and Flk-1/KDR by thyroid capillaries. Hypervascularization of thyroid goiters has been suggested to result from an abnormal constitutive activation of the TSHR pathway, which may occur through several mechanisms: increased pituitary TSH production, increased susceptibility of thyrocytes to TSH induced by iodide de®ciency as in goitrous patients, or chronic stimulation of the TSHR by autoimmune TSAb as in Graves's disease (Wilson and Foster, 1989; Dumont et al., 1992). By using PTUfed rats as an in vivo model of TSH-dependent thyroid goitrogenesis, we demonstrated that the mRNAs for PlGF and VEGF were coordinately upregulated in a time-dependent manner by increased levels of TSH in

the thyroid of PTU-fed rats compared to the thyroid gland of control rats. We also show that the expression of the mRNAs encoding VEGF and PlGF receptors, the tyrosine kinases Flt-1 and Flk-1/KDR, was enhanced 3 ± 8 days after stimulation of the thyroid gland and decreased gradually by day 16. Immunostaining results demonstrated that a higher number of vessels stained positive for Flt-1 and Flk-1/KDR in stimulated thyroid compared to normal thyroid. Increased amounts of Flt-1 and Flk-1/KDR proteins were observed at days 8 ± 16 in thyroid capillaries of PTU-treated rats compared to vessels of normal thyroid in untreated rats. Thus, increased steady-state levels of the mRNAs of Flt-1 and Flk-1/KDR receptors in the thyroid of PTU-fed rats resulted in a higher expression of the corresponding proteins in the endothelium of hyperplastic thyroid in comparison to unstimulated thyroid. Such in vivo data indicated that the increase in the mRNA levels of both growth factors (PlGF and VEGF) and their receptors occurred 3 ± 8 days from the beginning of PTU treatment, thus overlapping with the timing of proliferation of ECs (3 ± 5 days) observed in previous histomorphological studies by Wollman (Wollman et al., 1978; Smeds and Wollman, 1983). Our results suggest that in response to TSH stimulation, PlGF and VEGF are secreted by thyrocytes, activate their endothelial receptors on thyroid capillaries and stimulate proliferation of blood vessels in a paracrine manner. Our in vitro results on cultured thyrocytes con®rm the existence of a paracrine communication between


PIGF

VEGF

rRNA18S

pGEM 4Z

Figure 9 Transcriptional activation of rat PlGF and VEGF genes after stimulation of thyroid cells by TSH. Plasmids pGEM 4Z carrying the PlGF, VEGF or 18S rRNA inserts bound to nylon ®lters were hybridized with 32P-labeled run-on transcripts from nuclei isolated after 0 h (left lane) or 24 h (right lane) from the beginning of TSH stimulation of PC Cl 3 cells. Filters were exposed to autoradiographic ®lm for 10 days in the case of pGEM 4Z, PlGF and VEGF probes, and 24 h in the case of 18S rRNA probe

follicular cells and thyroid endothelium, and strongly suggest that PlGF may represent an important element of hormone-dependent angiogenesis that occurs in the thyroid gland. PC Cl 3 thyroid cells produce endothelial growth factors in response to TSH, since the conditioned medium collected from TSH-stimulated PC Cl 3 cells acquired a strong mitogenic activity for HUVE endothelial cells. In addition, we provide clear evidences that PlGF may be one of the endothelial growth factors present in the CM-TSH. In fact, TSH treatment of cultured PC Cl 3 thyroid cells induced a time- and dose-dependent up-regulation of both PlGF and VEGF mRNA expression. Furthermore, increased levels of PlGF mRNA were paralleled by an increased amount of 26 ± 28 kDa PlGF protein found in the conditioned medium of TSHtreated PC Cl 3 cells compared to the conditioned medium of untreated PC Cl 3 cells. Finally, the neutralization of the conditioned medium collected by TSH-stimulated PC Cl 3 cells with anti-PlGF antibodies drastically reduced the mitogenic activity exerted by CM-TSH. Altogether these results strongly suggested that at least a part of the mitogenic activity

exerted by CM-TSH on endothelial cells was due to the presence of PlGF in the conditioned medium. At present, the regulatory mechanism of PlGF expression in endocrine tissues is poorly understood. We demonstrate that TSH and forskolin induce a signi®cant increase in the expression of PlGF and VEGF mRNA, suggesting that the mRNA expression of both growth factors is regulated by the cAMP-dependent PKA-mediated pathway. TSH-induced upregulation of PlGF and VEGF mRNA expression occurs at the transcriptional level, since the increase in the steadystate level of both genes was suppressed by simultaneous treatment of TSH-stimulated thyrocytes with the RNA polymerase inhibitor DRB. Furthermore, run-on transcriptional assay con®rmed that in cultured rat thyrocytes, TSH-induced regulation of the expression of PlGF and VEGF occur at the transcriptional level. Time-course studies and experiments with the cycloheximide protein synthesis inhibitor suggested that dierent molecular mechanisms may be involved in the transcriptional regulation of VEGF and PlGF. TSH-induced upregulation of VEGF mRNA occurs as early as 30 min and does not require protein synthesis whereas PlGF expression occurs later and appears to require, in part, protein synthesis. Since TSH modulates cAMP-dependent gene expression by binding to its high-anity TSHR on the thyrocyte cell surface and activating downstream the cAMP-dependent PKA pathway (Yun et al., 1986; Van Sande et al., 1975; Dumont et al., 1992), we also investigated whether the activation of the PKA-dependent pathway by forskolin exerted some eect on the mRNA expression of PlGF and VEGF. Accordingly, forskolin-stimulated increase in the intracellular cAMP level induced a dose- and time-dependent increase in the VEGF and PlGF mRNA expression, suggesting that the action of TSH on the VEGF and PlGF mRNA level is mediated through the PKA-dependent pathway. In conclusion, our results suggest a key role for PlGF in the hypervascularization of thyroid goiters. Although previous work has suggested that VEGF produced by thyroid cells in response to TSH and/or Graves's IgGs, might represent the most likely candidate for the angiogenesis factor produced in the thyroid gland (Sato et al., 1995), our study suggests the existence of a complex interplay between VEGF and PlGF in the stimulation of goiter vascularization. Our in vivo and in vitro ®ndings suggested that the chronic activation of the TSHR pathway (either induced by increased TSH levels or, possibly, by TSH mimicking agents such as Graves's stimulatory IgGs) may trigger and maintain the dramatic hypervascularization observed in experimental and human goiters by coordinately regulating the expression of both the growth factors of the vasculotrophins family (VEGF and PlGF) in thyrocytes and the Flt tyrosine kinase receptor family (Flt-1, Flk-1/KDR) in the thyroid endothelium.

Materials and methods Tissue samples Goiter samples were obtained after resection from 7 patients who had undergone surgery at the National Cancer Institute `Fondazione Pascale', Naples, Italy.

2695


2696

Bioptic specimens were immediately frozen in liquid nitrogen until RNA extraction was performed. All patients were diagnosed as having Graves's disease. In vivo studies in thiouracil-fed rats Thirty 8-week-old Fisher rats were fed with rat-chow with low-iodide content. The PTU goitrogen agent was added to the water at a concentration of 2 mg/ml. Thirty 8-week-old Fisher rats fed with normal chow were used as controls. After 1, 3, 8 and 16 days, rats were anesthetized with ether, blood samples taken from jugular vein and sacri®ced by cervical dislocation. TSH, T3 and T4 hormones were evaluated by radioimmunoassay using a rat-speci®c kit (Amersham Inc.). The thyroid glands were removed and divided in two equally representative parts: one half was weighed and subsequently frozen in liquid nitrogen until RNA extraction was performed; the other half was ®xed in 4% formaldehyde made from paraformaldehyde for 15 h at 48C, embedded in paran and then processed for immunoperoxidase staining. RNA extraction and Northern blotting hybridization Total cellular RNA was isolated from cultured cell lines as previously described (Chomczynski and Sacchi, 1987). RNA was extracted from frozen specimens by the CsCl cushion method (Chirgwin et al., 1979) with minor modi®cations. Northern blots were performed essentially as described (Sambook et al., 1989), using nylon HybondN membranes (Amersham Inc.), according to the instructions provided by the manufacturer. All cDNA probes were radiolabeled with a random primed synthesis kit (Multi-Prime, Amersham Inc.). Hybridization reactions were performed at 428C, in 50% formamide, 5% Denhardt's, 56SSPE, 0.2% SDS and 100 mg/ml of denatured sonicated salmon sperm DNA, with 26106 c.p.m./ml of hybridization solution. Filters were washed at 608C in 26SSC, 0.2% SDS twice for 30 min and subsequently, for the stringent washes, twice for 30 min each in 0.26SSC, 0.1% SDS. Filters were air-dried and exposed to autoradiographic ®lm for 4 days in the case of VEGF probe and 3 days in the case of PlGF probe. The VEGF probe used in this study was a 0.6 Kb EcoRIBamHI fragment speci®c for the human VEGF cDNA. The PlGF probe used in this study was the full-size mouse cDNA (DiPalma et al., 1996). The Flt-1 and Flk-1/KDR probes used in this study are human cDNAs previously described (Viglietto et al., 1996). Nuclear RNA isolation The run-on in vitro assay was used on nuclei isolated from the rat PC Cl 3 thyrocytes. PC Cl 3 cells were grown to semicon¯uence and starved for 48 h in F12 medium supplemented with 0.5% bovine serum albumin (BSA). Exogenous TSH was added to starved PC Cl 3 cells at a dose of 1 mU/ml in F12 with 0.5% BSA. After 24 h of stimulation with TSH, approximately 36107 cells were collected by centrifugation at 2000 r.p.m. for 5 min, resuspended in lysis buer (10 mM Tris-HCl pH 7.4, 10 mM NaCl, 3 mM MgCl2, 0.5% NP40) and incubated in ice for 20 min. Nuclei were isolated by centrifugation at 1600 r.p.m., counted and stored in storage buer (50 mM Tris-HCl pH 8.3, 40% (v/v) glycerol, 5 mM MgCl2 and 0.1 mM EDTA) at a concentration of 108 nuclei/ml in liquid nitrogen. Run-on transcription was performed using 56106 nuclei in a reaction buer containing 10 mM Tris-HCl pH 8.0, 300 mM KCl and 5 mM MgCl2, and 150 mCi of [32P]UTP incubated for 60 min at 288C. Nuclear RNA was prepared as previously described (Chirgwin et al., 1979).

Hybridization to immobilized plasmid DNA Fifteen micrograms of the plasmid pGEM 4Z (Promega Inc.) and of the plasmids carrying human VEGF, mouse PlGF and 18S rRNA, were linearized, denatured with 0.2 N NaOH, neutralized and applied to a slot blot apparatus. Nylon ®lters (Huybond-N, Amersham Inc.) were subsequently ®xed by u.v. irradiation. Hybridization reactions were performed at 428C, in 50% formamide, 5% Denhardt's, 50 mM NaPhosphate pH 7.0, 56SSC, 0.1% SDS and 100 mg/ml of denatured sonicated salmon sperm DNA, and 2.5 mg/ml of pGEM 4Z plasmid, with 56106 c.p.m./ml of hybridization solution. Filters were washed at room temperature in 26SSC, 0.2% SDS twice for 30 min and subsequently, for the stringent washes, twice for 30 min each in 26SSC, 0.1% SDS at 608C, and for 30 min at 378C in 10 mg/ml of RNAse in 26SSC. Filters were air-dried and exposed to autoradiographic ®lm for 10 days in the case of pGEM 4Z, and of the PlGF and VEGF probes, and 1 day in the case of 18S rRNA probe. Cell lines and culture The rat thyroid cell line PC Cl 3 used in this study is a thyroid epithelial cell line derived from 18-month-old Fisher rats previously described (Fusco et al., 1987). PC Cl 3 cells were grown in F12 medium (Sigma) suplemented with 5% calf serum (Flow Laboratories) supplemented with a mix of the six hormones (TSH, somatostatin, insulin, growth hormone, glycyl-hystidyl-lysine and transferrin, Sigma). Experiments with PC Cl 3 cells were performed by growing the cells to semicon¯uence and subsequently starving them in serum-free medium supplemented with 0.5% BSA. Bovine TSH (Sigma) or forskolin (Sigma) were diluted at the concentrations indicated in the Results section in serum-free F12 medium supplemented with 0.5% BSA. The HUVE cells used in this study were derived from fresh umbilical cord and were at passage 4 ± 6 when used. HUVE cells were grown on 0.5% gelatin-treated plates in DMEM medium supplemented with 20% FCS, 70 mg/ml of ECGS (Sigma) and 100 mg/ml of heparin (Sigma). [3H]thymidine incorporation assay HUVE cells were seeded at a density of 36104 in 24-well dishes in DMEM 20% FCS and ECGS. The medium was replaced after 24 h with 10% FCS in DMEM. After 24 h the cells were treated with increasing amounts (10%, 25%, and 50%, vol/vol) of conditioned medium collected either from untreated PC Cl 3 cells (CM) or from 15 h-TSHstimulated PC Cl 3 (CM-TSH), in the presence or absence of the anti-PlGF or of a control antibody (anti-CRIPTO). The stimulation was repeated after 24 h. Subsequently, 0.5 mCi/well of [3H]thymidine was added and the cells incubated overnight at 378C. Cells were then ®xed in 5% TCA at 48C for 5 min, washed with ethanol and lysed with 20 mM NaOH, 1% SDS. Incorporated [3H]thymidine was counted in a liquid scintillation counter. Each experiment was performed three times in duplicate. Cell proliferation experiments Cells were seeded in 24-well dishes (1.56104/well). After 24 h, the medium was replaced with DMEM plus 10% FCS containing 0%, 25% and 50% (vol/vol) conditioned medium collected from untreated (CM) or from TSHstimulated PC Cl 3 (CM-TSH), in the presence or absence of 200 ng/ml of anti-PlGF antibody. Control antibodies used were the anti-CRIPTO used at the same concentration. Cell number was determined after 2 and 4 days. Each experiment was performed in quadruplicate for three times.


Immunoperoxidase staining

Cell labeling and immunoprecipitation

Immunohistochemistry was performed using puri®ed antiFlt-1 IgG, anti-Flk-1/KDR IgG and anti-PCNA antibodies at a concentration of 1 mg/ml. Control reactions were performed using control normal rabbit IgGs at a concentration of 1 mg/ml. Incubation with anti-rabbit IgG and avidin-biotin-peroxidase complex was carried out according to the supplier's conditions (Vectastain) followed by counterstain with Harris Haematoxylin.

PC Cl 3 cells were grown in F12 medium supplemented with 5% calf serum and the mix of the six hormones; the medium was then changed and cells were incubated for 7 h in Eagle's minimal essential medium (EMEM) devoid of methionine, in the presence and in absence of 1 mU/ml of TSH. Subsequently, 35S-methionine (800 mCi/mmole, Dupont/NEN) was added (150 mCi/ml) and cells were incubated for 6 h at 378C. After centrifugation to remove cellular debris, the conditioned medium from unstimulated (CM) and TSH-stimulated (CM-TSH) PC Cl 3 cells was collected and concentrated by ultra®ltration (Amicon). Aliquots of either CM or CM-TSH were immunoprecipitated overnight at 48C with anti-VEGF antibodies and protein A-Sepharose or with normal rabbit preimmune antiserum and protein A-Sepharose. Immunoprecipitated proteins were collected by centrifugation and electrophoresed in reducing conditions on 12.5% polyacrylamide gel. Gels were ®xed in 10% acetic acid, enlightened for 30 min, dried and exposed for 7 days.

Western blot Conditioned media were obtained by starving con¯uent PC Cl 3 cells in F12 medium supplemented with 0.5% (BSA) for 3 days. After medium replacement, the cells were incubated for 18 h in F12 medium devoid of serum in the presence or absence of 10 mU/ml of TSH (Sigma). Conditioned media were concentrated by acetone precipitation. Fifty mgs of protein were electrophoresed in reducing conditions on a 12.5% polyacrylamide gel and transferred onto PVDF membranes (Immobilon, PDVF). Membranes were blocked in 5% non-fat dry milk, 50 m M Tris pH 8, 0.05% Tween 20, for 2 h at room temperature and subsequently incubated with anti-PlGF polyclonal antibodies in 0.5% non-fat dry milk in TBS (150 mM NaCl, 20 mM Tris-HCl pH 8) for 1 h at room temperature. After several washings with TBS and TTBS (150 mM NaCl, 20 mM Tris-HCl pH 8 and 0.05% Tween 20), for 5 min each, the membranes were incubated with secondary antirabbit antibodies (Amersham, Inc.) for 1 h at room temperature. After several washings with TBS and TTBS, the membranes were incubated with the avidin-biotinperoxidase complex (Dako, Glostrup, Denmark) and developed with the ECL system (Amersham, Inc.).

Acknowledgements We thank Mrs M Terracciano for technical assistance, Miss A Secondulfo for correcting and typing of manuscript and Dr H Weich for the generous gift of the VEGF probe. This work was supported by grants of the Progetto Speciale `Angiogenesi' from the Associazione Italiana Ricerca sul Cancro (AIRC) to GV and MGP; the Progetto speciale `Oncosoppressori' from AIRC to AF and the `Progetto A.I.D.S. 1995 ± Roma ± Italia' from the Ministero della SanitaÁ, Istituto Superiore di SanitaÁ to MGP.

References Brandt R, Normanno N, Gullick WI, Lin JH, Harkins R, Schneider D, Wallace-Jones B, Ciardiello F, Persico MG, Armenante F, Kim N and Salomon DS. (1994). J. Biol. Chem., 269, 17320 ± 17328. Cao Y, Chen H, Zhou L, Chiang M-K, Anand-Apte B, Weatherbee JA, Wang Y, Fang F, Flanagan JG and LikShing Tsang M. (1996). J. Biol. Chem. 271, 3154 ± 3162. Chirgwin JM, Pryziyba AE, MacDonald RJ and Ruttner WJ. (1979). Biochemistry, 18, 5294 ± 5299. Chomczynski P and Sacchi N. (1987). Anal. Biochem., 162, 156 ± 159. Connolly DT, Olander JV, Heuvelman D, Nelson R, Monsell R, Siegel N, Haymore BL, Leimgruber R and Feder J. (1989). J. Biol. Chem., 264, 20017 ± 20024. De Vries C, Escobedo JA, Ueno H, Houck K, Ferrara N and Williams LT. (1992). Science, 255, 989 ± 991. DiPalma T, Tucci M, Russo G, Maglione D, Lago CT, Romano A, Saccone S, DellaValle G, De Gregorio L, Dragani TA, Viglietto G and Persico MG. (1996). Mamm. Genome, 7, 6 ± 12. DiSalvo J, Bayne ML, Conn G, Kwock PW, Trivedi PG, Sodermann DD, Palisi TM, Sullivan KA and Thomas KA. (1995). J. Biol. Chem., 270, 7717 ± 7723. Dumont JE, Lamy F, Roger P and Maenhaut C. (1992). Physiol. Rev., 72, 667 ± 697. Ferrara N and Henzel WJ. (1989)., Biochem. Biophys. Res. Commun., 161, 851 ± 858. Fusco A, Berlingieri MT, Di Fiore PP, Portella G, Grieco M and Vecchio G. (1987). Mol. Cell. Biol., 7, 3365 ± 3370. Goodman AL and Rone JD. (1987). Endocrinology, 121, 2131 ± 2140. Hauser S and Weich HA. (1993). Growth Factors, 9, 259 ± 268.

Keck PJ, Hauser SD, Krivi G, Sanzo K, Warren T, Feder J and Connolly DT. (1989). Science, 246, 1309 ± 1312. Kendall RL, Wang G, DiSalvo J and Thomas KA. (1994). Biochem. Biophys. Res. Commun., 201, 326 ± 330. Maglione D, Guerriero V, Viglietto G, Delli-Bovi P and Persico MG. (1991). Proc. Natl. Acad. Sci. USA, 88, 9267 ± 9271. Rognoni JB, Penel C and Ducret F. (1984). Acta Endocr., 105, 40 ± 48. Sambrook J, Fritsch EF and Maniatis T. (1989). Molecular Cloning: A Laboratory Manual. 2nd ed, Cold Spring Harbour, New York, Cold Spring Harbor Laboratory Press. Sato K, Yamazaki K, Shizume K, Kanaji Y, Obara T, Ohsumi K, Demura H, Yamaguchi S and Shibuya M. (1995). J. Clin. Invest., 96, 1295 ± 1302. Sawano A, Takahashi T, Yamaguchi S, Aonuma M and Shibuya M. (1996). Cell Growth and Dierentiation, 7, 213 ± 221. Shweiki D, Itin A, Neufeld G, Gitay-Goren H and Keshet E. (1996). J. Clin. Invest., 91, 2235 ± 2243. Shibuya M, Yamaguchi S, Yamane A, Ikeda T, Tojo A, Matsushime H and Sato M. (1990). Oncogene, 5, 519 ± 524. Smeds S and Wollman SH. (1983). Lab. Invest., 48, 285 ± 291. Terman BI, Carrion ME, Kovacs E, Rasmussen BA, Eddy RL and Shows TB. (1991). Oncogene, 6, 1677 ± 1683. Terman BI, Khandke L, Dougher-Vermazan M, Maglione D, Lassam NJ, Gospodarowicz D, Persico MG, BoÈ hlen P and Eisinger MG. (1994). Growth Factors, 11, 187 ± 195. Van Sande J, Grenier G, Willems C and Dumont JE. (1975). Endocrinology, 96, 781 ± 786.

2697


2698

Viglietto G, Maglione D, Rambaldi M, Cerutti J, Romano A, Trapasso F, Fedele M, Ippolito P, Chiappetta G, Botti G, Fusco A and Persico MG. (1995). Oncogene, 11, 1569 ± 1579. Viglietto G, Romano A, Maglione D, Rambaldi M, Paoletti I, Lago CT, Califano D, Monaco C, Mineo A, Santelli G, Manzo G, Botti G, Chiappetta G and Persico MG. (1996). Oncogene, 13, 577 ± 587. Wilson JD and Foster DW. (1989). The Thyroid. Textbook of Endocrinology. 7th ed, Press of Saunders Company, Philadelphia, USA.

Wollman SH, Herveg JP, Zeligs JD and Ericson LE. (1978). Endocrinology, 103, 2306 ± 2314. Wynford-Thomas D, Stringer MJ and Williams ED. (1982). J. Endocr., 94, 131 ± 140. Yun K, Yamashita S, Izumi K, Yonemitsu N and Sugihara H. (1986). J. Endocr., 111, 397 ± 405. Ziche M, Maglione D, Ribatti D, Morbidelli L, Lago CT, Battisti M, Paoletti I, Barra A, Tucci M, Parise G, Vincenti V, Granger HJ, Viglietto G and Persico MG. (1997). Lab. Investigation, 76, 517 ± 531.