Role of peroxidase inhibition by insulin in the bovine thyroid cell proliferation mechanism Leo ´ n Krawiec 1 , Ramo ´ n A. Pizarro 2 , Paula Aphalo, Elena M. V. de Cavanagh 3 , Mario A. Pisarev 1,2,4 , Guillermo J. Juvenal 1,2 , Lucı ´ a Policastro 2 and Laura V. Bocanera 2 1 Argentine National Research Council (CONICET); 2 Department of Radiobiology, Argentine Atomic Energy Commission (CNEA); 3 Department of Physical Chemistry, School of Pharmacy and Biochemistry, University of Buenos Aires and 4 Department of Human Biochemistry, School of Medicine, University of Buenos Aires, Buenos Aires, Argentina Monolayer primary cultures of thyroid cells produce, in the presence of insulin, a cytosolic inhibitor of thyroid peroxi- dase (TPO), lacto peroxidase (LPO), horseradish peroxidase (HRPO) and glutathione peroxidase (GPX). The inhibitor, localized in the cytosol, is thermostable and hydrophylic. Its molecular mass is less than 2 kDa. The inhibitory activity, resistant to proteolytic and nucleolytic enzymes, disappears with sodium metaperiodate treatment, as an oxidant of carbohydrates, supporting its oligosaccharide structure. The presence of inositol, mannose, glucose, the specific inhibition of cyclic AMP-dependent protein kinase and the disap- pearance of peroxidase inhibition by alkaline phosphatase and a-mannosidase in purified samples confirms its chemical structure as inositol phosphoglycan-like. Purification by anionic interchange shows that the peroxidase inhibitor elutes like the two subtypes of inositol phosphoglycans (IPG)P and A, characterized as signal transducers of insulin action. Insulin significantly increases the concentration of the peroxidase inhibitor in a thyroid cell culture at 48 h. The addition of both isolated substances to a primary thyroid culture produces, after 30 min, a significant increase in hydrogen peroxide (H 2 O 2 ) concentration in the medium, concomitantly with the disappearance of the GPX activity in the same conditions. The presence of insulin or anyone of both products, during 48 h, induces cell proliferation of the thyroid cell culture. In conclusion, insulin stimulates thyroid cell division through the effect of a peroxidase inhibitor, as its second messenger. The inhibition of GPX by its action positively modulates the H 2 O 2 level, which would produce, as was demonstrated by other authors, the signal for cell proliferation. Keywords: insulin; IPG; inhibitor; peroxidase; proliferation. Thyroid peroxidase (TPO) organifies iodide in the presence of H 2 O 2 and is responsible for the synthesis of the thyroid hormones. Bocanera et al. demonstrated that monolayer primary cultures of thyroid cells, in the presence of insulin, lose their iodide organification capacity several days before the disappearance of TPO activity, due to the presence of a cytosolic inhibitor, not detectable in fresh tissue [1]. However, the inhibitor is active on TPO isolated from fresh tissue and inhibits the iodide organification when it is added to free follicles. This inhibitor is thermostable, dialyzable, has a molecular mass of less than 2 kDa and it has no species specificity, as was shown in our previous work [1]. Several authors have demonstrated the remarkable role of insulin in many aspects of cell metabolism. The observed effects were ascribed to the modulation exerted by the hormone on key enzymes of different metabolic pathways. During the past decades it has been demonstrated that inositol phosphoglycan (IPG) acts as a mediator of insulin and insulin growth factor (IGF-I), mimicking their effects [2–8]. Water-soluble IPG, results from the hydrolysis of glycosyl phosphatidylinositol (GPI) [8,9]. Two related substances, able to act as insulin mediators, were isolated from hepatic membranes [10]. Both compounds have a molecular mass of 1000–2000 Da. The isolation and partial characterization, in bovine [11] and human liver [12], of two subtypes of IPG (P and A) which act as signal transducers of the insulin action, confirmed these findings. The aims of the present study were to establish: (a) if insulin regulates the peroxidase activity through the action of the cytosol inhibitor as a second messenger; (b) if the inhibitor and IPG are alike and (c) the implications of this regulatory pathway in the mechanisms of cell proliferation. Materials and methods Cell cultures Monolayer primary and free follicle cultures of bovine thyroid cells were obtained according to Bocanera et al.[1]. Briefly, bovine thyroids were obtained from a local slaugh- terhouse immediately after death and transported to the laboratory in ice-cold saline containing 100 UÆmL )1 penicillin and 100 lgÆmL )1 streptomycin. The glands were Correspondence to L. Krawiec, Divisio ´ n Bioquı ´ mica Nuclear, Unidad de Actividad Radiobiologı ´ a, Comisio ´ n Nacional de Energı ´ aAto ´ mica, Avenida del Libertador 8250, 1429-Buenos Aires, Argentina. Fax: + 54 11 67727121; Tel.: + 54 11 6772 7185; E-mail: krawiec@cnea.gov.ar Abbreviations: GPX, glutathione peroxidase; HRPO, horseradish peroxidase; LPO, lacto peroxidase; TPO, thyroid peroxidase. (Received 5 April 2004, accepted 28 April 2004) Eur. J. Biochem. 271, 2607–2614 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04189.x dissected carefully, cut into small pieces under sterile condi- tionsanddigestedwith1 mgÆmL )1 collagenase type 1 A and DNAse 20 lgÆmL )1 in 199 medium (Sigma Chemical Co.) with 2.2 mgÆmL )1 NaHCO 3 ,5 lgÆmL )1 transferrin, 0.85 l M bovine insulin, 2.5 lgÆmL amphotericin B, 100 UÆmL )1 penicillin and 100 lgÆmL )1 streptomycin, in a relation of 10 mLÆg )1 tissue, at 39 °C, during 90 min. The digested material was filtered through sterile gauze and centrifuged at 700 g during 30 s, washed twice with the medium and resuspended in the complete M-199 medium containing 5% fetal bovine serum. The follicles were seeded in standard plates for monolayer cultures or in dishes supplied with a hydrophobic layer, to keep the follicles in suspension. After 48 h, the medium was replaced by another containing 0.1% fetal bovine serum. Previous to each treatment, insulin was suppressed for 48 h. The cultures were stopped at different days and assays were performed as indicated below. All cells were grown at 37 °C, under an atmosphere of 95% air and 5% CO 2 , in a humidified incubator. Bovine TPO isolation Thyroid peroxidase was obtained according to Bocanera et al. [1]. Briefly, fresh glands were dissected, cut into small pieces and homogenized in 0.1 M phosphate buffer, 0.1 m M KI, pH 8.0 [1 : 5 (w/v)]. The homogenate was filtered through gauze and centrifuged at 900 g during 10 min at 4 °C. The resulting supernatant was centrifuged at 105 000 g during 60 min. The pellet was resuspended [1/1, (w/v)] in the same buffer containing 0.6% Triton X-100 and kept on ice, with frequent agitation, during 180 min. After centrifugation at 105 000 g, during 60 min, the resulting supernatant was utilized as the source of solubilized TPO. Isolation of cytosolic inhibitor The 105 000 g supernatant obtained from monolayer cells cultures, homogenized in milli Q water, was analyzed for its inhibitory activity on TPO extracted from fresh glands. Thyroid peroxidase (TPO), lacto peroxidase (LPO), horseradish peroxidase (HRPO), glutathione peroxidase (GPX), catalase and cyclic AMP-dependent protein kinase activity assays TPO, LPO and HRPO activities were determined following the tyrosine iodination assay, described by DeGroot and Davis [13]. GPX was assayed according to Flohe ´ and Gunzler [14]. Catalase was determined by the method of Aebi [15] and cyclic AMP-dependent protein kinase according to Villalba et al. [16]. Hydrogen peroxide assay Hydrogen peroxide was determined according to Ravindra- nath [17]. Effect of the peroxidase inhibitor on glucose transport in free follicles The free follicles, cultured as described above, were washed three times with Krebs–Ringer 20 m M Hepes (KRH), pH 7.4, without glucose. The experiments were performed in 24-well plates, with 0.1 mL of the follicles suspension in KRH (5 mg of proteinÆmL )1 ) and aliquots corresponding to 0; 10; 20; 50 lg of cytosolic protein, obtained from confluent monolayer thyroid cell culture, in each well. After preincubation for 60 min at 37 °C,theuptakeof 2-deoxyglucose (nonmetabolic substrate) was assessed according to Krawiec et al.[18]. Solubility of the peroxidase inhibitor in organic solvents The cell supernatant (2 mL corresponding to 1 mg of protein) was mixed with 2 mL of chloroform, by stirring the mixture vigorously, at room temperature. Both phases, discarding the interphase, were separated by centrifugation and evaporated to dryness. The dry residues were resus- pended in the original volume of water and the TPO activity assayed as described above. Action of proteolytic enzymes on the peroxidase inhibitor Trypsin and chimotrypsin, resuspended in 200 m M ammo- nium bicarbonate, pH 8.0, were incubated with the cyto- solic fraction (1 lg of enzyme per 15 lgofprotein)at37 °C, during 12 h. Proteinase K (1 lg of enzyme per 10 lgofprotein)was incubated with the cytosolic fraction in a medium contain- ing (final concentration): 10 m M Tris/HCl; 0.1 M NaCl; 1m M EDTA, pH 7.5, at 56 °C, during 12 h. In all cases, the proteolytic enzymes were inactivated at 95 °C, during 10 min and the TPO control assay was performed in the presence of the inactivated enzymes. Effects of DNAse and RNAse on the peroxidase inhibitor The incubations were performed in both cases in NaCl/P i , pH 7.4, at 37 °C during 12 h, with an enzymatic concen- tration 1 : 10 of the protein present in the assay. The reactions were continued as indicated for the proteolytic assays. Purification of the peroxidase inhibitor by liquid–liquid partition Purification was performed according to the method of Folch et al. [19] modified, handled with a rotary evaporator (e.g. Heidolph WB 2000), under reduced pressure and low N 2 stream, at 65 °C in a water bath. The cell supernatant was extracted first, in a separating funnel, with 20 volumes of chloroform/methanol [2 : 1 (v/v)] and the partition was made with water (1 : 6 of the total volume). After centrif- ugation, the hydrophilic upper phase was removed. The hydrophobic lower phase was re-extracted with theoretical upper phase from the mixture chloroform/methanol/water [8 : 4 : 3 (v/v)]. The second upper phase was combined with the first for further processing. The evaporated upper and lower phases were resuspended in water (50% of the original volume) and the inhibitory effects were compared. The combined upper phases were submitted to a liquid–liquid partition in a column of Sephadex G-25 coarse (dry bead diameter 100–300 lm), according to Siakotos and Rouser 2608 L. Krawiec et al. (Eur. J. Biochem. 271) Ó FEBS 2004 [20]. The beads were soaked overnight in methanol/water [1 : 1 (v/v)] and then rinsed four or five times with the same solvent. The Sephadex was packed into a 1.5 · 20 cm column and successively washed twice in the same order and with the same volumes of the four mixtures of solvents then used for the elution. After seeding the sample, the elution was made in the following order of increasing polarity: (a) Fifty mililiters chloroform/methanol 19:1 (v/v), saturated with 0.5 percentage water for the elution of non-ganglioside lipids. (b) One hundred mililiters chloroform/methanol 19:1 (v/v) five volumes, acetic acid one volume, saturated with 2.5 percentage water, for the elution of gangliosides. (c) Fifty mililiters chloroform/methanol 9:1 (v/v) five volumes, acetic acid one volume, saturated with 4.2 percentage water, for the elution of traces of gangliosides. (d) One hundred mililiters methanol/water 1:1 (v/v) for the elution of non-lipid polar compounds soluble in water. After evaporation the fractions were resuspended in water (50% of the volume submitted to the partition). Effect of sodium metaperiodate (NaO 4 I) on the activity of the peroxidase inhibitor The inhibitor, partially purified by liquid–liquid partition, was evaporated to dryness, resuspended in NaCl/P i with 2m M NaO 4 Iin0.1 M phosphate buffer 0.1 m M KI, pH 7.0, and maintained at room temperature in the darkness during 12 h. The remnant inhibitory action was assayed on the TPO enzyme as described above. The control activity of TPO was determined in the presence of NaO 4 I. Action of alkaline phosphatase on the inhibitor The inhibitor purified as described above, was evaporated to dryness and resuspended in 50 mm Tris/HCl (pH 9.3 at 25 °C), 1 m M MgCl 2 ,0.1 m M ZnCl 2 , 1 mm spermidine, and maintained at 37 °C during 12 h. After inactivating the enzyme at 95 °C, during 10 min, in the presence of 5 mm EDTA, the remnant inhibitory action was assayed on the TPO enzyme as described above. The control activity of TPO was determined in the presence of inactivated alkaline phosphatase. Action of a-mannosidase on the inhibitor The inhibitor purified as described above, was evaporated to dryness and resuspended in 100 m M acetate buffer (pH 5.0), and incubated in the presence of 100 mU of a-mannosidase at 37 °C during 12 h. After inactivating the enzyme at 95 °C, during 10 min, the mixture was neutralized with KOH. The remnant inhibitory action was assayed on the TPO enzyme as described above. The control activity of TPO was determined in the presence of inactivated a-mannosidase. Thin layer chromatography (TLC) of the peroxidase inhibitor and its components Samples, obtained by liquid–liquid partition, were applied in duplicate plates of silica gel 60 on polyester. The developing solvent consisted in a mixture of chloroform/ methanol/0.2% calcium chloride [60 : 40 : 9 (v/v/v)]. After drying the plates, one of them was (a) submerged in a solution of sulfuric acid 5% in ethanol; (b) dried and (c) heated at 120 °C in a stove, until the appearance of brown spots, typical of carbohydrates. The silica of the second plate, with no detection reagent, was scraped at the position where the brown spots had appeared on the first plate. This material was then resuspended in water and extracted by shaking. After centrifugation, the supernatant was assayed on the TPO activity. The inhibitor components were determined in the sam- ples, previously hydrolized in 6 M HCl, during 8 h at 100 °C and evaporated under nitrogen. The samples were applied in a plate of silica gel 60 on polyester. The developing solvent consisted of a mixture of propanol/ethanol/water (7 : 1 : 2). The procedure continued as was described above. D -Mannose, D -galactose, D -trehalose, inositol and D -glucose were run in parallel, as standards. High performance liquid chromatography (HPLC) for protein isolation The protein content of the same fraction purified by TLC was analyzed by reverse-phase chromatography, utilizing: column C8, 30 · 2.1 mm (Brownlee). Solvent A, trifluoro- acetic acid (TFA) 0.1%; Solvent B, acetonitrile 80% in TFA 0.08%; Gradient, 5 min at 10% of solvent B and 30 min from 10% to 100% of solvent B. The isolated protein fractions were assayed for the inhibitory effect on the TPO activity. Isolation of peroxidase inhibitor as inositol phosphoglycan (IPG)-like compound The samples, purified by liquid–liquid partition, were processed following the protocols described previously by Caro et al. [12]. Cell proliferation Bovine thyroid cells were cultured as described previously. After the different treatments (48 h in the presence of insulin, IPG P or IPG A-like) the cultures were stopped. Cell counting. Dishes of cells were detached from the monolayer with trypsin and counted in a hemocytometer. [ 3 H]Thymidine incorporation. To measure thymidine incorporation into TCA-precipitable material, [ 3 H]thymi- dine (5 lCiÆmL )1 ) was added together with insulin, IPG P or IPG A-like during all the treatment. At the end of the culture period the medium was removed and the cells were washed with phosphate-buffered saline (NaCl/P i ). TCA (0.5 mL, 10%) was added to the wells for 1 h at 4 °C. After the removal of TCA the treatment was repeated. The TCA- precipitable material was dissolved using 0.3 M NaOH, overnight at 37 °C and the radioactivity was counted in a liquid scintillation counter. Protein was determined according to Lowry et al.[21]. All reagents were obtained from Sigma Chemical Co (St. Louis, MO, USA). Na 125 Iand[ 3 H]thymidine were purchased from New England Nuclear. Each experiment was repeated at least four times and each point was Ó FEBS 2004 Peroxidase inhibition and proliferation (Eur. J. Biochem. 271) 2609 run in quadruplicate. Statistical significance of the differences among groups was calculated according to Dunnet [22]. Results The 105 000 g supernatant (cytosolic fraction) obtained from monolayer bovine thyroid cells cultured in the presence of 0.85 l M insulin, usual concentration for thyroid cultures, causes significant inhibitory effects on the semi- purified TPO activity from fresh tissue. Values of TPO activity in the monolayer thyroid cell cultures were below the blanks. In order to isolate the peroxidase inhibitor, the 105 000 g supernatant was submitted to a purification procedure by a liquid–liquid partition method. In the first stage, the hydrophylic upper phase contained the total original activity of the peroxidase inhibitor. In the second stage of the purification, 70% of the inhibitory activity was recov- ered in fraction 4 (F4) (methanol/water 1 : 1), which contains nonlipid water-soluble compounds. Fractions 3 and 2 contributed with 25 and 5%, respectively, of total inhibitory activity, while no activity was recovered in fraction 1, which extracts lipids (results not shown). The physicochemical nature of the peroxidase inhibitor was assessed by different methods. The biological activity, as a TPO inhibitor, was used as an end-point to monitor the effect of different treatments. Table 1 shows the effect of several enzymes, including trypsin, chymotrypsin and proteinase K, on the 105000·g supernatant from the thyroid monolayer cell culture. None of these proteolytic enzymes affected the inhibitory power. These results were confirmed by the HPLC analysis of the semipurified fraction F4 demonstrating the absence of inhibitory effect on TPO in the protein fractions (data not shown). The absence of lipids in the structure of the peroxidase inhibitor, discernible throughout the purification process, was confirmed by the fact that all the inhibitory activity remained in the aqueous phase after emulsifying the 105 000 g supernatant fraction with an equal volume of chloroform. DNA and RNA were ruled out as possible components of the inhibitor, as DNAse and RNAse did not affect the inhibitory capacity (data not shown). Finally, as can be seen in Table 1, peroxidase inhibition disappeared when the isolated fractions were preincubated with sodium metaperiodate, which produces oxidative ruptures of carbohydrates. Furthermore, the presence of carbohydrates in the inhibitor structure was confirmed by TLC. In order to achieve a more accurate identification of the peroxidase inhibitor, fraction F4 was submitted to the method of inositol phosphoglycans (IPG) isolation. Figure 1 shows the inhibition on TPO elicited by the different fractions successively isolated from the original cytosolic supernatant. Both fractions IPG-like: type P and type A significantly inhibited TPO activity. Fraction A was more powerful (100%) than fraction P (61%), as indicated by comparing the inhibition percentages of equal volumes eluted of both fractions. The disappearance of the inhibitory effect on the peroxidase activity, when fraction F4 was submitted to alkaline phosphatase (Table 1), or to a-mannosidase (not shown) suggested that its chemical structure would Table 1. Effects of the different treatments on the physicochemical properties of the TPO (thyroid peroxidase) inhibition. The peroxidase inhibitor (PI) was isolated from monolayer thyroid cell cultures per- formed in the presence of insulin. The remnant inhibitory activity, after the treatments is the measure of stability. Other details are given in the text. TPO, thyroid peroxidase; PI, peroxidase inhibitor. Each value is the average ± SEM of four samples from four different experiments. *P<0.01. Assay TPO activity (pmol IÆmin )1 Æ mg protein )1 ) Sodium metaperiodate TPO 208 ± 12 TPO + PI 113 ± 6* TPO + (PI + NaO 4 I) 187 ± 13 Alkaline phosphatase TPO 383 ± 24 TPO + PI 116 ± 15* TPO + (PI + alkaline phosphatase) 394 ± 32 Trypsin TPO 391 ± 35 TPO + PI 34 ± 2* TPO + (PI + trypsin) 55 ± 4* Chymotrypsin TPO 168 ± 14 TPO + PI 3 ± 05* TPO + (PI + chymotrypsin) 1 ± 02* Proteinase K TPO 325 ± 28 TPO + PI 96 ± 12* TPO + (PI + proteinase K) 78 ± 20* Fig. 1. TPO inhibition caused by the insulin mediator with different purity grades. Comparison of the inhibitory effect of: supernatant from bovine primary thyroid cell cultures (B-MLS); upper phase isolated by the method of Folch et al. [19] (UP); fraction 4 of liquid–liquid partition (F4); inositol phosphoglycan-like P and A (IPG P and IPGA) on the tyrosine iodinating activity of fresh thyroid peroxidase (TPO). Each value is the average of four samples from four different experiments. 2610 L. Krawiec et al. (Eur. J. Biochem. 271) Ó FEBS 2004 correspond to an inositol phosphoglycan (Table 1). The similarity between the TPO inhibitor and IPGs was confirmed by TLC, which demonstrated the presence of inositol, glucose and mannose. Finally, it is important to mention that the total inhibition caused by the peroxidase inhibitor on the cyclic AMP-dependent protein kinase (Table 2) is one of the specific characteristics of IPG as was demonstrated by Villalba et al.[16]. To determine whether the cytosolic inhibitor, obtained from the monolayer thyroid cell culture, has a selective effect on TPO, we assayed its action on LPO, HRPO and GPX activities. As shown in Table 2 all the enzymes were inhibited by the F4 preparation. Conversely, catalase activity was not affected at all. Table 3 shows the total inhibition on the GPX activity caused by both IPGs-like products as compared to the semipurified F4. The significant increase of the hydrogen peroxide release into the culture medium, 30 min after the addition of the inhibitor is depicted in Table 4. Table 5 shows the significant increase in concentration of the peroxidase inhibitor by the action of insulin, during 48 h, in thyroid primary cultures with 0.1% fetal bovine serum, totally depleted of the hormone, 48 h earlier. Table 6 depicts the effect produced on cell proliferation of the primary thyroid culture by the addition of insulin and both IPGs-like compounds, under equal conditions to those described for experiments of Table 5. In all cases, thyroid cell cultures underwent a significant increase in cell number and [ 3 H]thymidine incorporation into DNA compared to the controls. Discussion It has been demonstrated that insulin and IGF I have similar effects as growth promoting factors [23]. Even though insulin is 1000 times less active on the IGF I receptor than on its own, the effects of high levels of insulin may be explained by a dual action on both [24]. Besides, it has been identified hybrids formed by subunits of the insulin and the IGF I receptors [23]. Table 2. Comparative effects of the cytosolic inhibitor on thyroid per- oxidase (TPO), lactoperoxidase (LPO), horseradish peroxidase (HRPO), glutathione peroxidase (GPX), cAMP-dependent protein kinase (PKA) and catalase (CAT) activities. The cytosolic inhibitor, obtained from monolayer thyroid cell cultures, in the presence of insulin, was purified, as fraction F4 or IPG-like, by the methods of liquid–liquid partition and anionic interchange. Other details are given in the text. TPO, 0.5 mg of protein; lactoperoxidase, 50 mU; horse- radish peroxidase, 100 mU; glutathione peroxidase, 600 mU; catalase, 60 U and F4 equivalent to 15 lg of protein. PKA was totally inhibited by F4 and by the purified IPG-like fraction (not shown). Each value is the average of four samples from four different experiments. Assay % of inhibition TPO + F4 50 LPO + F4 82 HRPO + F4 100 GPX + F4 100 CAT + F4 0 PKA + F4 100 Table 3. Inhibition of glutathione peroxidase (GPX) activity by the cytosolic inhibitor. Technical details are given in the text. Each value is the average of four samples from four different experiments. F4, fraction 4 of the liquid–liquid partition, equivalent to 30 lgofprotein. IPGP-like and IPGA-like: inositol phosphoglycan-like P and A, equivalent to 6 lgofprotein. Assay GPX activity (mUÆmL )1 ) GPx control 77.5 ± 8.5 GPx + F4 0 GPx + IPGP-like 0 GPx + IPGA-like 0 GPx + H 2 O milli q 75.2 ± 6.7 Table 4. H 2 O 2 released by monolayer primary cultures of bovine thyroid cells in confluence, treated with IPGP-like or IPGA-like. H 2 O 2 was measured 30 min after the medium renewal in the presence or absence of IPG-like P or A (equivalent to 20 lg of protein in the extract). Each value is the average of four samples from four different experiments. *P < 0.01 vs. control. Assay H 2 O 2 released to the medium (nmolÆmin )1 Æmg )1 of cell protein) % Control 28.8 ± 4.7 100 IPGP-like 52.8 ± 7.4* 183 IPGA-like 99.2 ± 14.2* 344 Table 5. Effect of insulin on the peroxidase inhibitor production. Comparison of the inhibitory effect, on TPO activity, of the super- natants (80 lg protein) obtained from primary monolayer cell cultures of bovine thyroid (B-MLS), in the absence or presence of 3.4 l M insulin during 48 h. Each value is the average ± SEM of six samples from four different experiments. *P<0.05 compared to fresh TPO. **P<0.01 compared to fresh TPO. Assay TPO activity (pmol IÆmin )1 Æmg protein )1 ) % Inhibition Fresh bovine TPO control 439 ± 56 Fresh bovine TPO + BMLS 241 ± 42* 45 Fresh bovine TPO + (B MLS + Insulin) 87 ± 37** 80 Table 6. Effect of insulin and its mediators on the proliferation of bovine thyroid cells. Technical details are described in the text. Aliquot’s protein: Insulin, 3.4 l M ; IPG-like P and A, 1 lgÆmL )1 . The treatments were carried on during 48 h. Each value is the average ± SEM of six samples from six different experiments. *P < 0.01 vs. control. Treatment Cells · 10 3 per dish [ 3 H]Thymidine (c.p.m.Æwell )1 ) Control 470 ± 29 4518 ± 309 Control + insulin 1070 ± 37* 7023 ± 234* Control + IPGP-like 970 ± 50* 6546 ± 315* Control + IPGA-like 810 ± 50* 6420 ± 129* Ó FEBS 2004 Peroxidase inhibition and proliferation (Eur. J. Biochem. 271) 2611 Our previous studies [1] demonstrated that monolayer primary cultures of bovine thyroid cells, maintained in the presence of insulin, produced a cytosolic inhibitor of TPO activity. Its molecular mass of less than 2 kDa coincides with that reported previously for the insulin mediators [10]. Under our experimental conditions the TPO inhibitor was neither detected in fresh tissue nor in free follicle cultures. The insulin concentration in the culture medium, which is 1000- to 10 000-fold higher than that of calf plasma, explains the undetectable levels of the TPO inhibitor in fresh thyroid. The fact that insulin mediators were isolated from 1 kg of fresh liver [11] supports this assumption. On the other hand, the absence of the inhibitor in free follicle cultures, may be due to the lack of cellular proliferation, absent in this type of culture. These facts support the hypothesis that insulin regulates negatively the activity of the thyroid peroxidase. The similar effects observed on TPO, LPO, HRPO and GPX show that the inhibition has not enzyme specificity, suggesting that insulin regulates the different peroxidase activities by means of a cytosolic inhibitor as mediator of its action. To rule out the possibility that the cytosolic inhibition of peroxidase, instead of being specific for the enzyme activity, could be part of a generalized action on the cellular metabolism, we assayed the influence of the cytosol on the glucose uptake by free thyroid follicles. There were no differences, when the cytosol was added, discarding this mechanism for the effect of insulin on glucose uptake (results not shown). These results agree with those obtained when the action of phospho-oligosaccharides on glucose uptake, by isolated rat adipocytes, was studied [25]. The purification procedure demonstrated the hydro- phylic nature of the peroxidase inhibitor and the absence of lipid constituents in its structure. The compound has no protein residues, as the activity persisted after the action of proteinases and this was confirmed by the absence of inhibitory activity in protein fractions isolated by HPLC (not shown). The loss of biological activity by treatment with sodium metaperiodate and the identifica- tion of carbohydrate residues after the inhibitor purifica- tion by TLC would suggest that the compound is an oligosaccharide. Purification of the cytosolic material by liquid–liquid partition and anionic interchange shows that the peroxidase inhibitory activity corresponds to two similar products to IPG (P and A) isolated and characterized as signal transducers of insulin action in bovine [11] and human liver [12]. This similarity was demonstrated by analytical meth- ods and confirmed by their characteristic inhibition caused on the cyclic AMP-dependent protein kinase, which is specific of IPG as was demonstrated by Villalba et al.[16]. Jones and Varela-Nieto [9] reported a long list of metabolic pathways and enzymatic activities that are affected by the insulin mimetic action of IPG P and IPG A. The observation that both compounds isolated by us inhibit the different peroxidase activities is a new finding that must be added to the previously mentioned list. Our results demonstrate that insulin enhances the peroxidase inhibitory action in bovine thyroid cell cultures, suggesting that the hormone modulates the peroxidase activities through a mediator such as the one reported for other enzyme activities [9]. IPG mimics the action of insulin in placenta [11], liver [12] and adipocytes [25]. Vasta et al. [26] working in cultures of human fibroblasts and Leo ´ n et al. [23] in organotypic cultures of chicken otic vesicles, reported the mitogenic effects of insulin. These effects were attributed to IPG, acting as its second messenger. In addition, anti-IPG Ig block the mitogenic effects of IGF-I on the otic vesicle, suggesting that the hydrolysis of GPI to produce IPG is an important pathway in the mechanism of action on cell proliferation by insulin and IGF- I [23]. Insulin acts through the interaction with its membrane receptors, producing long and short-term metabolic effects [27]. The long-term process involves, as mediators, phos- phorylated proteins produced by the tyrosine kinase acti- vation. On the other hand, the short-term effects are developed by the action of IPG stimulating a phosphopro- tein phosphatase [27]. Considering that both mechanisms mediate mitogenic effects of insulin, it may be licit to postulate that IPG could act as the transducer of the short-term mechanism, without discarding the long-term effect on the same process. Taking into account that IPG is a potent mitogen for a variety of tissues [28,29] it seems important to determine whether the same phenomenon occurs in thyroid, with the peroxidase inhibitor purified as IPG, verifying the possible relationship between the peroxidase inhibition and the proliferation process(Fig. 2). Our results, without discarding the anti-apoptotic effect of insulin [24], demonstrate that both insulin and the peroxidase inhibitor, in the absence of TSH, stimulate cell growth of thyroid cultures. Conversely, Petitfrere et al.[5] and Deleu et al. [30] considered the presence of insulin and TSH necessary to stimulate cell growth, ruling out the individual effects of both hormones. Petitfrere et al.’s [5] conclusions are not unequivocal, as they did not assay insulin and TSH separately. On the other hand, Deleu et al. [30] utilized bovine hormones for dog thyroid cultures but Leo ´ n et al. [23] stressed the importance of using homolog- ous species factors in in vitro studies, as we have carried out in the present studies. Finally, Deleu et al. [30] point out that thyroid cell regulation varies from one species to another, thus explaining our discrepancy with their results. Fig. 2. Proposed mechanism for the relationship between peroxidase inhibition and cell proliferation. 2612 L. Krawiec et al. (Eur. J. Biochem. 271) Ó FEBS 2004 The stimulation of cell growth by insulin was ascribed to the induction, mediated by IPG, of the nuclear proto-oncogene c-jun expression [23]. Our results show a clear relationship among the enhance- ment of cellular proliferation, the increase in the hydrogen peroxide concentration and the GPX inhibition caused by the addition of the insulin mediator. Hydrogen peroxide, at noncytotoxic concentrations, has been pointed out as a cell transducing signal for insulin and other hormones [31]. The normal level of hydrogen peroxide, in different tissues, ranges between 10 )9 and 10 )7 M [32] and an extracellular concentration of 10 n M H 2 O 2 promotes fibroblast growth [33]. Hydrogen peroxide has been postulated as the signal transduction for the induction of protooncogene c-jun expression [34]. The intracellular levels of H 2 O 2 are main- tained as a result of its metabolism by GPX and catalase [35]. In conclusion, the present results propose that the insulin mediator, as inhibitor of TPO activity, is responsible for the decrease of thyroid hormone biosynthesis in primary cultures in the presence of insulin. In addition, our data suggest that insulin stimulates cell division, in bovine thyroid cultures, promoting the hydrolysis of a membrane glycosyl-phosphtidyl inositol (GPI) which generates an inositol phosphoglycan-like substance as insulin’s second messenger, which mimics the insulin’s effect on cell prolif- eration by inhibiting glutathione peroxidase (GPX). This inhibition positively modulates the H 2 O 2 level and its derivative compounds, the reactive oxygen species. These changes would be the signal for the induction of a mitogenic mechanism. Acknowledgements The authors thank Dr Amanda Schwint for reviewing the English version of the manuscript, Dr R.M. de Lederkremer, and Dr Oscar J. Opezzo for the discussion of results. The contribution of Dr Marı ´ a. A. Dagrosa and Dr Silvia Moreno in some experiments and the excellent technical assistance of Ms. Gabriela Beraldi are acknowledged. These studies were supported by grants from the National Secretary of Science and Technology (SEPCYT-ANPCYT) and CONICET to M.A.P. References 1. Bocanera, L.B., Aphalo, P., Pisarev, M.A., Gartner, R., Sil- berschmidt, D., Juvenal, G.J., Beraldi, G. & Krawiec, L. (1999) Presence of a soluble inhibitor of thyroid iodination in primary cultures of thyroid cells. Eur. J. Endocrinol. 141, 55–60. 2. Larner, J. (1983) Mediators of postreceptor action of insulin. Am. J. Med. 74, 38–51. 3. Larner, J. (1988) Insulin-signaling mechanisms: lessons from the old testament of glycogen metabolism and the new testament of molecular biology. Diabetes 37, 262–275. 4. Saltiel, A.R. (1990) Second messengers of insulin actions. Diabetes Care 13, 244–256. 5. Petitfrere, E., Sartelet, H., Vivien, D., Varela-Nieto, I., Elbtaouri, H., Martiny, L. & Haye, B. (1998) Glycosyl phosphatidylinositol (GPI)/inositolphosphate glycan (IPG): An intracellular signaling system involved in the control of thyroid cell proliferation. Biochimie 80, 1063–1067. 6. Varela-Nieto, I., Leo ´ n, Y. & Caro, H.N. (1996) Cell signalling by inositol phosphoglycans from different species. Comp. Biochem. Physiol. 115b, 223–241. 7. Jones, D.R. & Varela-Nieto, I. (1998) The role of glycosyl-phos- phatidylinositol in signal transduction. J. Int. Biochem. Cell Biol. 30, 313–326. 8. Jones, D.R. & Varela-Nieto, I. (1999) Diabetes and the role of inositol-containing lipids in insulin signaling. Mol. Med. 5, 505–514. 9. Stralfors, P. (1997) Insulin second messengers. Bioessays 19, 327–335. 10. Saltiel, A. & Cuatrecasas, P. (1986) Insulin stimulates the gen- eration from hepatic plasma membranes of modulators derived from an inositol glycolipid. Proc.NatlAcad.Sci.USA83, 5793– 5797. 11. Nestler, J.E., Romero, G., Huang, L.C., Zhang, C. & Larner, J. (1991) Insulin mediators are the signal transduction system responsible for insulin’s actions on human placental steroido- genesis. Endocrinology 129, 2951–2956. 12. Caro, H.N., Kunjara, S., Rademacher, T.W., Leo ´ n, Y., Jones, D.R., Avila, M.A. & Varela-Nieto, I. (1997) Isolation and partial characterisation of insulin-mimetic inositol phosphoglycans from human liver. Biochem. Mol. Med. 61, 214–228. 13. DeGroot, L.J. & Davis, A.M. (1962) Studies on the biosynthesis of iodotyrosine: a soluble thyroidal iodide-peroxidase tyrosine- iodinase system. Endocrinology 70, 492–504. 14. Flohe ´ , L. & Gunzler, W.A. (1984) Assays of glutathione perox- idase. Methods Enzymol. 105, 114–121. 15. Aebi, H. (1984) Catalase in vitro. Methods Enzymol. 105, 121–126. 16. Villalba, M., Kelly, K.L. & Mato, J.M. (1988) Inhibition of cyclic AMP-dependent protein kinase by the polar head group of an insulin-sensitive glycophospholipid. Biochimica Biophysica Acta 968, 69–76. 17. Ravindranath, V. (1994) Animal models and molecular markers for cerebral ischemia-reperfusion injury in brain. Methods Enzy- mol. 233, 610–619. 18. Krawiec, L., Chester, H.A., Bocanera, L.V., Pregliasco, L.B., Juvenal, G.J., Silberschmidt, D. & Pisarev, M.A. (1995) Thyroid autoregulation: action of two iodolactones on deoxyglucose and iodide uptake by FRTL-5 cells. Thyroidol. Clin. Exp 7, 73–78. 19. Folch, J., Lees, M. & Sloane-Stanley, G.H. (1957) A simple method for the isolation and purification of total lipids from animal tissues. J. Biol. Chem. 226,497. 20. Siakotos, A.N. & Rouser, G. (1965) Analytical separation of nonlipid water soluble substances and gangliosides from other lipids by dextran gel column chromatography. J. Am. Oil Chemists Soc. 42, 913–919. 21. Lowry, O.H., Rosebrough, N.J., Farr, A.L. & Randall, R.J. (1951) Protein measurement with the folin phenol reagent. J. Biol. Chem. 193, 265–275. 22. Dunnet, C.W. (1955) A multiple comparison procedure for com- paring several treatments with a control. J. Am. Stat. Assoc. 50, 1096–1121. 23. Leo ´ n, Y., Sanz, C., Gira ´ ldez,F.&Varela-Nieto,I.(1998) Induction of cell growth by insulin and insulin-like growth factor-I is associated with jun expression in the otic vesicle. J. Compar. Neurol. 398, 323–332. 24. De Meyts, P., Christoffersen, C.T., Ursoe, B., Wallache, B., Groenskov, K., Yakushiji, F. & Shymko, M. (1995) Role of the time factor in signalling specificity: application to mitogenic and metabolic signalling by the insulin and insulin-like growth factor I receptor tyrosine kinases. Glucose Metabolism and Growth Factors: Metabolism 44, 1–11. 25. Kelly, K.L., Merida, I., Wong, E.H.L., DiCenzo, D. & Mato, J.M. (1987) A phospho-oligosaccharide mimics the effect of insulin to inhibit isoproterenol-dependent phosphorylation of phospholipid methyltransferase in isolated adipocytes. J. Biol. Chem. 262, 15285–15290. Ó FEBS 2004 Peroxidase inhibition and proliferation (Eur. J. Biochem. 271) 2613 26. Vasta, V., Bruni, P., Clemente, R., Vannini, F., Ochoa, P., Romero, G., Farnararo, M. & Varela-Nieto, I. (1992) Role of the glycosylphosphatidylinositol/inositol phosphoglycan system in human fibroblast proliferation. Exp. Cell Res. 200, 439–443. 27. Litwack, G. (1993) Biochemistry of hormones I: peptide hor- mones, In Textbook of Biochemistry with Clinical Correlations (Devlin, T.M., ed.) 3rd edn, pp. 847–900, Wiley-Liss, New York. 28. Clemente, R., Jones, D.R., Ochoa, P., Romero, G., Mato, J.M. & Varela-Nieto, I. (1995) Role of glycosylphosphatidylinositol hydrolysis as a mitogenic signal for epidermal growth factor. Cell Signal. 7, 411–421. 29. Merida, I., Pratt, J.C. & Gaulton, G.N. (1990) Regulation of interleukin 2-dependent growth factor responses by glycosy- lphosphatidyl molecules. Proc. Natl Acad. Sci. USA 87, 9421– 9425. 30. Deleu, S., Pirson, I., Coulonval, K., Drouin, A., Taton, M., Clermont, F., Roger, P.P., Nakamura, T., Dumont, J.E. & Maenhaut, C. (1999) IGF-1 or insulin, and the TSH cyclic AMP cascade separately control dog and human thyroid cell growth and DNA synthesis, and complement each other in inducing mito- genesis,. Mol. Cell. Endocrinol. 149, 41–51. 31. Krieger-Brauer, H.I. & Kather, H. (1995) The stimulus-sensitive H 2 O 2 -generating system present in human fat-cell plasma membranes is multireceptor-linked and under antagonistic control by hormones and cytokines. Biochem. J. 307, 543–548. 32. Chance, B., Sies, H. & Boveris, A. (1979) Hydroperoxide meta- bolism in mammalian organs. Physiol. Rev. 59, 527–605. 33. Burdon, R.H. & Rice-Evans, C. (1989) Free radicals and the regulation of mammalian cell proliferation. Free Radic. Res. Comm. 6, 345–358. 34. Lee, S.F., Huang, Y.T., Wu, W.S. & Lin, J.K. (1996) Induction of c-jun protooncogene expression by hydrogen peroxide through hydroxyl radical generation and p60 src tyrosine kinase activation. Free Radic. Biol. Med. 21, 437–448. 35. Wolin, M.S., Mohazzab, H. & K.M. (1997) Mediation of signal transduction by oxidants. In Oxidative Stress and the Molecular Biology of Antioxidant Defenses (Scandalios, J.G., ed.), pp. 21–48. Cold Spring Harbor Laboratory Press, Cold Spring Harbor. 2614 L. Krawiec et al. (Eur. J. Biochem. 271) Ó FEBS 2004 . responsible for the decrease of thyroid hormone biosynthesis in primary cultures in the presence of insulin. In addition, our data suggest that insulin stimulates cell division, in bovine thyroid cultures,. significant increase in concentration of the peroxidase inhibitor by the action of insulin, during 48 h, in thyroid primary cultures with 0.1% fetal bovine serum, totally depleted of the hormone,. concentration in the medium, concomitantly with the disappearance of the GPX activity in the same conditions. The presence of insulin or anyone of both products, during 48 h, induces cell proliferation of