Activated Thyroglobulin Possesses a Transforming Growth Factor-β Activity*

Thyroglobulin (Tg), the thyroid hormone precursor, is a major protein component in the thyroid gland and may have other important functions. Here, we show that bovine Tg inhibited125I-labeled transforming growth factor-β1 (125I-TGF-β1) binding to cell-surface TGF-β receptors in mink lung epithelial cells with an IC50 of ∼300 nm. After disuccinimidyl suberate (DSS) modification, reduction/alkylation, treatment with 8m urea, 0.1% SDS, or acidic pH (pH 4–5), Tg exhibited a ∼5–10-fold increase of 125I-TGF-β1 binding inhibitory activity with IC50 of ∼30–60 nm. This inhibitory activity was an intrinsic property of the Tg and could not be segregated from Tg protein by 5% SDS-polyacrylamide gel electrophoresis or by immunoprecipitation using antiserum to Tg. Untreated Tg did not affect DNA synthesis but blocked the TGF-β-induced inhibition of DNA synthesis in mink lung epithelial cells. After DSS activation, Tg possessed TGF-β agonist activity and inhibited DNA synthesis of mink lung epithelial cells and rat thyroid cells. The activated Tg also exerted a small but significant TGF-β agonist activity in transcriptional activation of plasminogen activator inhibitor-1. These results suggest that Tg possesses an authentic TGF-β activity which can be induced by chemical modifications and treatments with denaturing agents and acidic pH.

Thyroglobulin (Tg), the thyroid hormone precursor, is a major protein component in the thyroid gland and may have other important functions. Here, we show that bovine Tg inhibited 125 I-labeled transforming growth factor-␤ 1 ( 125 I-TGF-␤ 1 ) binding to cell-surface TGF-␤ receptors in mink lung epithelial cells with an IC 50 of ϳ300 nM. After disuccinimidyl suberate (DSS) modification, reduction/alkylation, treatment with 8 M urea, 0.1% SDS, or acidic pH (pH 4 -5), Tg exhibited a ϳ5-10-fold increase of 125 I-TGF-␤ 1 binding inhibitory activity with IC 50 of ϳ30 -60 nM. This inhibitory activity was an intrinsic property of the Tg and could not be segregated from Tg protein by 5% SDS-polyacrylamide gel electrophoresis or by immunoprecipitation using antiserum to Tg. Untreated Tg did not affect DNA synthesis but blocked the TGF-␤-induced inhibition of DNA synthesis in mink lung epithelial cells. After DSS activation, Tg possessed TGF-␤ agonist activity and inhibited DNA synthesis of mink lung epithelial cells and rat thyroid cells. The activated Tg also exerted a small but significant TGF-␤ agonist activity in transcriptional activation of plasminogen activator inhibitor-1. These results suggest that Tg possesses an authentic TGF-␤ activity which can be induced by chemical modifications and treatments with denaturing agents and acidic pH.
During investigation on the TGF-␤ agonist activities of several protein conjugates of TGF-␤ peptide antagonists (8), we found that Tg, but not other proteins (bovine serum albumin and carbonic anhydrase), exhibited a potent TGF-␤ antagonist activity after modification with a bifunctional reagent, disuccinimidyl suberate (DSS). This preliminary result prompted us to investigate the TGF-␤ antagonist/agonist activities of Tg treated with various chemical reagents and acidic pH. In this report, we show that, prior to treatment, Tg moderately inhibited 125 I-TGF-␤ 1 binding to cell-surface TGF-␤ receptors in mink lung epithelial cells. After activation by DSS modification, reduction/alkylation or treatment with 8 M urea, 0.1% SDS, or acidic pH, the 125 I-TGF-␤ 1 binding inhibitory activity of Tg increased ϳ5-10-fold with IC 50 of ϳ30 -60 nM. Untreated and DSS-activated Tg possessed TGF-␤ antagonist and agonist activities as determined by DNA synthesis and by Northern blot analysis of plasminogen activator inhibitor-1 (PAI-1).
Preparation of Untreated Tg and Tg Treated with DSS Modification, Reduction/Alkylation, 8 M Urea, 0.1% SDS, and Acidic pH-Tg was isolated as described previously (10) from bovine thyroid, which was obtained fresh from a local meat-packing company. Since purified Tg and commercial Tg exhibited similar 125 I-TGF-␤ 1 binding inhibitory activities with IC 50 ϳ300 nM, commercial Tg was used for most experiments. Tg was modified with DSS at a ratio of 1:360 (Tg:DSS) as described previously (8), dialyzed against phosphate-buffered saline, pH 7.4 (PBS), and kept at 4°C prior to use. DSS-modified Tg was used as the activated Tg throughout most experiments unless otherwise indicated. The reduction and alkylation of Tg was performed using dithiothreitol and iodoacetamide as described previously (11) and then dialyzed against PBS. Tg was exposed to several pH levels (0.1 M Tris acetate, pH 7.4; 0.1 M acetate, pH 6.0; 0.1 M acetate, pH 5.0; and 0.1 M acetate, pH 4.2) at 4°C for 16 h. Following the treatment, the 125 I-TGF-␤ binding inhibitory activity of Tg was determined. Tg was also treated with 8 M urea in PBS, dialyzed against PBS, and used for experiments. Tg was also treated with 0.1% SDS in 0.1 M Tris-HCl, pH 8.3, at room temperature for 2 h and then analyzed for 125 I-TGF-␤ 1 binding inhibitory activity. The acidic pH and 0.1% SDS buffer solu-tions without Tg did not affect 125 I-TGF-␤ 1 binding to cell-surface TGF-␤ receptors after a 20 -1,200-fold dilution in the assay.
Specific Binding of 125 I-TGF-␤ 1 to Mink Lung Epithelial Cells-125 I-TGF-␤ 1 was prepared by iodination of TGF-␤ 1 with Na 125 I in the presence of chloramine T as described previously (8,12,13). The specific radioactivity of 125 I-TGF-␤ 1 was 1-3 ϫ 10 5 cpm/ng. Mink lung epithelial cells were grown on 24-well clustered dishes to near confluence in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. The monolayers were incubated with 0.1 nM 125 I-TGF-␤ in the presence of various concentrations of Tg and activated Tg in binding buffer (8,12,13). The specific binding of 125 I-TGF-␤ 1 to cell-surface TGF-␤ receptors was then determined at 0°C as described previously (8,12,13).
[methyl-3 H]Thymidine Incorporation and RNA Analysis-The [methyl-3 H]thymidine incorporation into cellular DNA and the RNA analysis of PAI-1 and glyceraldehyde-3-phosphate dehydrogenase (G3PDH) were carried out as described previously (8,13). The relative levels of PAI-1 mRNA were estimated based on the ratio of PAI-1 mRNA and G3PDH mRNA levels. The relative intensities of both mRNAs on the autoradiograms were quantitated by PhosphorImager.
Native Polyacrylamide Gel Electrophoresis-Approximately 1 ng of 125 I-TGF-␤ 1 was incubated with 10 g of untreated Tg, 10 g of DSSactivated Tg, or 20 g of human ␣ 2 M in 0.05 M Tris-HCl, pH 7.4, at room temperature for 0.5 h. An aliquot of the reaction mixture was analyzed by 5% native polyacrylamide gel electrophoresis as described previously (9). After electrophoresis, the gels were dried and autoradiographed. ␣ 2 M, untreated Tg, and DSS-activated Tg were located by Coomassie Brilliant Blue staining.
Assay for the 125 I-TGF-␤ 1 Binding Inhibitory Activities of Untreated and DSS-activated Tg after 5% SDS-Polyacrylamide Gel Electrophoresis-Approximately 100 g of untreated or DSS-activated Tg was subjected to 5% SDS-polyacrylamide gel electrophoresis (8,12,13). After electrophoresis, the gels were cut into slices (0.5 cm length). The sliced gels were then incubated with 0.2 ml of 0.2 M NH 4 HCO 3 , pH 7.8, at 4°C for 16 h. The extracts were analyzed for 125 I-TGF-␤ 1 binding inhibitory activity using mink lung epithelial cells. Additional untreated and DSS-activated Tg were run in parallel with the tested sample for location detection by Coomassie Brilliant Blue staining. 125 I-TGF-␤ 1 was also run as a control.
Immunoprecipitation of DSS-activated Tg-One hundred l of rabbit antiserum to Tg or non-immune serum was mixed with 20 l (gel volume) of protein A-Sepharose CL-4B at 4°C for 2 h. After washing, the immune IgG (anti-Tg IgG) affinity gel or non-immune IgG affinity gel was mixed with ϳ0.6 M DSS-activated Tg in 100 l of 0.05 M HEPES, 0.15 M NaCl, pH 7.4, at 4°C for 16 h. Twenty l of the supernatant was then assayed for 125 I-TGF-␤ 1 binding inhibitory activity.

I-TGF-␤ Binding Inhibitory Activities of Untreated and DSS-activated
Tg-To investigate TGF-␤ receptor binding activities of untreated and DSS-activated Tg, we first determined their abilities to inhibit 125 I-TGF-␤ 1 binding to cell-surface TGF-␤ receptors in mink lung epithelial cells. Mink lung epithelial cells have been used as a model system to examine TGF-␤ receptor binding activities (15,16). As shown in Fig To test the possibility that untreated and DSS-activated Tg might inhibit 125 I-TGF-␤ 1 binding to cell-surface TGF-␤ receptors by directly binding 125 I-TGF-␤ 1 , we investigated the interaction of both proteins with 125 I-TGF-␤ 1 using a 5% native polyacrylamide gel electrophoresis system (10). This gel system has been used to analyze the complex formation of 125 I-TGF-␤ and high molecular weight binding proteins, e.g. ␣ 2 M (9). In this system, the 125 I-TGF-␤ 1 -binding protein complex co-migrates with the binding protein toward the anode as free 125 I-TGF-␤ 1 migrates to the cathode (9). As shown in Fig. 2, no 125 I-TGF-␤ 1 radioactivity associated with untreated and DSSactivated Tg. (Fig. 2, A, lanes 1 and 2, and B, lanes 5 and 6). DSS-activated Tg migrated faster than untreated Tg toward the anode due to decrease of its positive charges after modification of ␣and ⑀-amino groups by DSS (Fig. 2B, lane 5 versus lane 6). As a positive control, 125 I-TGF-␤ 1 associated and comigrated with ␣ 2 M (Fig. 2, A, lane 3, and B, lane 8). The inability of untreated or DSS-activated Tg to form a complex with 125 I-TGF-␤ 1 was further confirmed by the observation that untreated and DSS-activated Tg did not affect the chromatographic migration of 125 I-TGF-␤ 1 on Sepharose 4B (data not shown). Tg has been shown to bind to an N-acetylglu- cosamine (GlcNAc) receptor (17), mannose 6-phosphate/insulin-like growth factor-II receptor (18), and megalin (19). However, inhibitors of this receptor binding including mannose 6-phosphate (5 mM), EDTA (5 mM), and lactoferrin (10 M) failed to have a significant effect on the 125 I-TGF-␤ 1 binding inhibitory activities of untreated and activated Tg (data not shown). These results suggest that untreated and activated Tg inhibit 125 I-TGF-␤ 1 binding by direct interaction with cell-surface TGF-␤ receptors but not by sequestering 125 I-TGF-␤ 1 from binding to the TGF-␤ receptors or by interaction with other known Tg-binding receptors.
Authenticity of 125 I-TGF-␤ 1 Binding Inhibitory Activity of Tg-TGF-␤ has been identified, primarily in a latent form, in platelets, tissues, and cultured cells (20 -22). We considered the possibility that the preparations of untreated and DSS-activated Tg were contaminated with active TGF-␤ and/or the latent form of TGF-␤. Under such circumstances, the chemical modification with DSS might release active TGF-␤ from the latent form of TGF-␤ in the Tg preparations, which could compete with 125 I-TGF-␤ 1 for binding to TGF-␤ receptors. To ex-clude this possibility, both preparations of untreated and DSSactivated Tg were subjected to 5% SDS-polyacrylamide gel electrophoresis under non-reducing conditions, eluted from the SDS-polyacrylamide gel, and analyzed for 125 I-TGF-␤ 1 binding inhibitory activity. As shown in Fig. 3, the 125 I-TGF-␤ 1 binding inhibitory activity was only detected where untreated and DSS-activated Tg migrated (close to the top of the separating gel, fraction 1). In this gel system, TGF-␤ migrated at the dye front because of its small molecular size (25 kDa). To further prove that the 125 I-TGF-␤ 1 binding inhibitory activity is an inherent property of DSS-activated Tg, we determined the effect of immunoprecipitation with antiserum to Tg on the activity of DSS-activated Tg. As shown in Table I, immunoprecipitation using antiserum to Tg but not non-immune serum completely depleted the 125 I-TGF-␤ 1 binding inhibitory activity from the activated Tg solution. These results suggest that the 125 I-TGF-␤ 1 binding activities of untreated and activated Tg are authentic.
Activation of Tg by Chemical Modifications and Treatments with Denaturing Agents and Acidic pH-As described previously, after modification of its ␣-amino and ⑀-amino groups with DSS, the 125 I-TGF-␤ 1 binding inhibitory activity of Tg was activated ϳ10-fold (IC 50 ϳ30 nM versus IC 50 ϳ300 nM prior to modification). This activation may result from a conformational change of Tg following modification by DSS. To test this possibility, we determined the effects of reduction and alkylation and treatments with 8 M urea and 0.1% SDS on the 125 I-TGF-␤ 1 binding inhibitory activity of Tg. As shown in Table II, Tg activated by 8 M urea and 0.1% SDS exhibited 125 I-TGF-␤ 1 binding inhibitory activities with IC 50 similar to that of DSSactivated Tg (Table II). Interestingly, reduction and alkylation also activated the 125 I-TGF-␤ 1 binding inhibitory activity of Tg (Table II). These results suggest that Tg can be activated by chemical modifications and treatments with denaturing agents. The activation of Tg by reduction and alkylation excludes the possibility that the 125 I-TGF-␤ 1 binding inhibitory activity of Tg might be contributed by contamination with TGF-␤ and/or latent TGF-␤ in the Tg preparations. Modification by reduction and alkylation is known to abolish the activity of TGF-␤ (20 -22).
To investigate the physiological relevance of the activation of Tg by denaturing agents, we examined the effect of acidic pH treatment on the 125 I-TGF-␤ 1 binding inhibitory activity of Tg. Acidic pH, like denaturing agents, may induce a conformational change of Tg. Furthermore, the acidic pH effect may be physiologically relevant since Tg encounters acidic intracellular compartments during its secretion, uptake, and recycling by FIG. 2. Native polyacrylamide gel electrophoresis of 125 I-TGF-␤ 1 after incubation with untreated Tg, DSS-activated Tg, and ␣ 2 -macroglobulin. Approximately 1 ng of 125 I-TGF-␤ was incubated with 10 g of untreated Tg, 10 g of DSS-activated Tg, or 20 g of human ␣ 2 M at room temperature for 0.5 h. An aliquot of the reaction mixture was subjected to 5% native polyacrylamide gel electrophoresis and autoradiography (A). Untreated Tg, DSS-activated Tg, and ␣ 2 M were located by Coomassie Blue staining (B). BSA* represents the aged (modified) product of bovine serum albumin, which was a carrier protein in the 125 I-TGF-␤ 1 preparations.

FIG. 3. Association of the 125 I-TGF-␤ 1 binding inhibitory activity with untreated and DSS-activated Tg on 5% SDS polyacrylamide gel electrophoresis.
Approximately 200 g of untreated or DSS-activated Tg was subjected to 5% SDS-polyacrylamide gel electrophoresis. Following electrophoresis, the gels were cut into slices (0.5 cm each) and extracted by incubation with 0.2 ml of 0.1 M NH 4 HCO 3 , pH 7.8, at 4°C for 16 h. Twenty l of the extracts was analyzed for 125 I-TGF-␤ 1 binding inhibition activity. The specific binding of 125 I-TGF-␤ 1 obtained in the presence and absence of 10 M ␤ 1 25 (41-65), a specific TGF-␤ peptide antagonist, were taken as 100% and 0% inhibition (220 Ϯ 40 and 4,932 Ϯ 352 cpm/well, respectively). Untreated and DSS-activated Tg were located at the top of the separating gel by Coomassie Blue staining. As a control, 125 I-TGF-␤ 1 migrated at the dye front. thyroid cells (23). As shown in Fig. 4, Tg was activated in a pH-dependent manner; treatment at decreasing pH enhanced Tg activity. The IC 50 of Tg treated with 0.1 M Tris acetate, pH 7.4, 0.1 M acetate, pH 6.0, 0.1 M acetate, pH 5.0 and pH 4.2 were estimated to be 250 Ϯ 50 nM (n ϭ 4), 100 Ϯ 20 nM (n ϭ 4), 36 Ϯ 6 nM (n ϭ 4), and 25 Ϯ 4 nM (n ϭ 4), respectively. These results suggest that Tg can be activated by physiologically relevant acidic pH levels (pH 5-6), which are found in organelles such as endosomes, lysosomes, and the trans-Golgi network.  The assay for 125 I-TGF-␤ 1 binding to cell-surface TGF-␤ receptors was carried out by incubating mink lung epithelial cells with 0.1 nM 125 I-TGF-␤ 1 and ϳ0.06 M Tg treated with non-immune IgG affinity gel or immune IgG (anti-Tg IgG) affinity gel. The specific binding was then determined.
b DSS-activated Tg was treated with non-immune IgG-affinity gel and immune IgG (anti-Tg IgG) affinity gel at 4°C for 16 h. After centrifugation, the supernate was assayed for 125 I-TGF-␤ binding inhibitory activity.
b Tg was treated with 8 M urea at room temperature for 2 h, and dialyzed against PBS.
c Tg was reduced and alkylated with ␤-mercaptoethanol and iodoacetamide. After reduction and alkylation, Tg was dialyzed against PBS.
d Tg was modified with DSS at a molar ratio of 1:360 (Tg:DSS). The 125 I-TGF-␤ 1 binding inhibitory activity of Tg was assayed after dilution with PBS containing 0.1% bovine serum albumin.
TGF-␤ Antagonist and Agonist Activities of Untreated and DSS-activated Tg-TGF-␤ has two prominent biological activities: growth inhibition and transcriptional activation (20 -22). To define the antagonist/agonist activities of untreated and DSS-activated Tg, we first determined the effects of both proteins on cellular growth as measured by [methyl-3 H]thymidine incorporation into cellular DNA. As shown in Fig. 5A, DSSactivated Tg inhibited the [methyl-3 H]thymidine incorporation into DNA of mink lung epithelial cells in a concentration-dependent manner with an IC 50 of ϳ30 nM. The DNA synthesis inhibition induced by 30 nM activated Tg was blocked in the presence of 10 M ␤ 1 25 (41-65), a specific TGF-␤ peptide antagonist (data not shown). In contrast, untreated Tg did not affect DNA synthesis at concentrations up to 120 nM (Fig. 5A). Since untreated Tg inhibited 125 I-TGF-␤ binding to TGF-␤ receptors ( Fig. 1) but did not affect DNA synthesis, we determined the antagonist activity for untreated Tg toward TGF-␤ 1 -induced inhibition of DNA synthesis. As shown in Fig. 5B, untreated Tg at 180 nM blocked the inhibition of DNA synthesis induced by 0.125 pM TGF-␤ 1 . Untreated Tg alone did not affect DNA synthesis. These results suggest that DSS-activated and untreated Tg can function as TGF-␤ agonist and antagonist, respectively.
The other prominent biological activity of TGF-␤ is transcriptional activation of fibronectin and PAI-1 (20 -22). We therefore determined the effect of DSS-activated Tg on the transcriptional expression of PAI-1 in mink lung epithelial cells. As shown in Fig. 5C, DSS-activated Tg exhibited a small but significant activity in transcriptional activation of PAI-1. At 0.025, 0.05, and 0.1 M, Tg exhibited 1.2 Ϯ 0.05-fold (n ϭ 6), 1.4 Ϯ 0.1-fold (n ϭ 6), and 1.5 Ϯ 0.1-fold (n ϭ 6) increase of the relative level of PAI-1(PAI-1 mRNA:G3PDH mRNA ratio), respectively. These results demonstrate that Tg exhibits two prominent biological activities of TGF-␤ (growth inhibition and transcription activation), although the transcription activation activity is relatively weak.
Growth Inhibition of Rat Thyroid Cells by DSS-activated Tg-As described previously, Tg may encounter intracellular acidic compartments during internalization and recycling in thyroid cells (2,23). The thyroid cells are presumably the target cells for activated Tg. We, therefore, examined the effect of DSS-activated Tg on cell growth as measured by DNA synthesis of rat thyroid cells (FRTL-5 cells). FRTL-5 cells are a cell line of thyroid-stimulating hormone-dependent thyrocytes that display distinct cell biological properties of thyroid cells (14). As shown in Table III

FIG. 7. Putative TGF-␤ active-site motifs in human and bovine Tg (hand b-Tg) and human and rat insulinlike growth factor-binding protein 3, 4, 5, and 6 (h-IGFBP-3, r-IGFBP-3, h-IGFBP-4, r-IGFBP-4, h-IGFBP-5, r-IG-FBP-5, h-IGFBP-6, and r-IGFBP-6).
The underlined 4 amino acid residues are putative TGF-␤ active-site motifs. The numbers represent the positions in the amino acid sequence of the molecule. receptor types including the type I, type II, type III, and type V receptors (Fig. 6). These results indicate that DSS-activated Tg is capable of inhibiting cell growth of its known target cells.
Putative TGF-␤ Active-site Motifs in Tg-DSS-activated Tg and TGF-␤ have similar biological activities (growth inhibition and transcriptional activation) although TGF-␤ is much more potent on a molar basis than DSS-activated Tg. Tg is a 660-kDa homodimeric glycoprotein and does not show sequence homology with TGF-␤, a 25-kDa homodimeric protein (2, 20 -22). The distinct difference in the sizes of Tg and TGF-␤ raises the question of the mechanism of action since heterodimerization of TGF-␤ receptors induced by dimeric TGF-␤ (24 -26) is believed to be the mechanism of TGF-␤ receptor activation (27). To determine how activated Tg exerts TGF-␤ agonist activity, we examined potential TGF-␤ active-site motifs in Tg. We have identified a putative TGF-␤ active-site motif (WXXD) in TGF-␤ isoforms (8). Synthetic pentacosapeptides containing this motif were found to exert potent TGF-␤ antagonist activities (8). Multiple conjugation of these pentacosapeptides to carrier proteins such as bovine serum albumin and carbonic anhydrase enhances the 125 I-TGF-␤ 1 binding inhibitory activities and confers partial TGF-␤ agonist activities (8). These studies led us to identify naturally occurring, structurally unrelated TGF-␤ agonists that contain this motif and also possess dimeric or oligomeric structures. Insulin-like growth factor-binding protein 3 (IGFBP-3) was the first to be identified as a TGF-␤ partial agonist. IGFBP-3 has a putative TGF-␤ active-site motif (WCVD) near its C terminus and forms a dimer in aqueous solution (13,28). Human and bovine Tg contains 5-7 WCVD motifs, which are identical to those of IGFBP-3 and other IGFBPs (Fig. 7) (3,4,28). Since untreated and DSS-activated Tg can function as TGF-␤ antagonist and agonist, respectively, we hypothesize that the availability of the WCVD motifs in Tg molecules determines the function of Tg as antagonist or agonist. We speculate that untreated Tg functions as an antagonist when a few of the WCVD motifs are available for interaction with TGF-␤ receptor but unable to induce appropriate heterodimerization or hetero-oligomerization of TGF-␤ receptors. Upon activation by DSS modification or acidic pH treatment, most of the WCVD motifs in the Tg molecule become available and are capable of inducing appropriate heterodimerization or hetero-oligomerization of TGF-␤ receptors. Thus, activated Tg can function as an agonist. DISCUSSION In this report, the authenticity of TGF-␤ activities of untreated and activated Tg has been supported by several lines of evidence. These include the following: 1) The 125 I-TGF-␤ 1 binding inhibitory activity could not be segregated from Tg proteins by 5% SDS-polyacrylamide gel electrophoresis and by immunoprecipitation using specific antiserum to Tg. 2) Reduced and alkylated Tg exhibited a TGF-␤ 1 binding inhibitory activity with an IC 50 close to those for Tg activated by DSS modification, 8 M urea, or acidic pH treatments. 3) Both untreated and activated Tg were incapable of binding TGF-␤ 1 . 4) The possibility that Tg obtained from commercial sources might contain activated Tg (potentially generated during purification), which might contribute to the TGF-␤ activity of untreated Tg, was ruled out by the observation that at 500 nM untreated Tg did not show any growth inhibitory activity whereas activated Tg inhibited growth at ϳ50 nM.
DSS-activated Tg shows TGF-␤ agonist activities in growth inhibition and transcriptional activation of PAI-1. The reason why DSS-activated Tg exerts a moderate effect on the transcriptional activation of PAI-1 is unknown. It would be inter-esting to determine qualitatively and quantitatively which types of TGF-␤ receptors are activated by DSS-activated Tg. Since the type I and type II TGF-␤ receptor heterodimerization is required for the transcriptional activation of PAI-1 (27), DSS-activated Tg would have a moderate effect on the type I/II receptor heterodimerization as demonstrated by its effect on the PAI-1 expression.
In the lumen of thyroid follicles, Tg can be found at a concentration up to 590 mg/ml (29). The levels of activated Tg which could be potentially generated during its receptor-mediated endocytosis and recycling are not known. It is possible that the steady-state activation of Tg may play a role in homeostasis of thyroid gland growth and regulation of immune status. Although the immunosuppressive activity of activated Tg has not been tested, activated Tg is presumed to possess such activity since TGF-␤ is known to be a potent immunosuppressor (30). We hypothesize that Tg may be involved in thyroid autoimmune diseases through the TGF-␤ antagonist activity of native Tg and/or through activation of its latent TGF-␤ agonist activity.