Thyroglobulin

* Objective.-The composition of sugar chains on thyroglobulin (Tg) produced in thyroid carcinoma cells (C-Tg) is different from Tg produced in normal thyroid tissues (NTg). In this study, we designed a new method for detecting Tg derived from thyroid carcinoma based on the differences between C-Tg and N-Tg in the reactivity with lectins.

Materials and Methods.-Thyroglobulin preparations obtained from various thyroid tissues were incubated with lectins, and the amount of lectin-unbound Tg (ub-Tg) in the supernatant relative to Tg untreated with lectin was determined by enzyme-linked immunosorbent assay and expressed as ub-Tg(%). In addition, to study further the differences in glycosylation between C-Tg and N-Tg, concanavalin A binding to Tg digested with Staphylococcus aureus VS protease was analyzed on nitrocellulose membrane after Western blotting.

Results.-The ub-Tg(%) in C-Tg from papillary carcinoma was significantly higher than in Tg from Graves' dis

ease, benign goiter, and normal thyroid tissue for both concanavalin A and ricinus communis agglutinin-120. Concanavalin A did not appear to bind to Tg from papillary carcinoma after V8 treatment by Western blot analysis. The ub-Tg(%) in Tg from follicular adenoma was significantly higher than C-Tg from follicular carcinoma, whereas there were no differences in ub-Tg(%) between follicular carcinoma and normal thyroid tissue in concanavalin A treatment.

Conclusions-These results suggest our new methods can distinguish both between C-Tg from papillary carcinoma and N-Tg, and between follicular carcinoma and follicular adenoma in thyroid tissue specimens. Thus, this type of analysis may be applicable to differentiate C-Tg from NTg in thyroid aspirates for the adjunctive cytodiagnosis of thyroid carcinoma.

(Arch Pathol Lab Med. 1998;122:715-720)

The composition of sugar chains on thyroglobulin (Tg) produced in thyroid carcinoma cells (C-Tg) is known to be different from that produced in normal thyroid tissues (N-Tg).1-5 Miscellaneous lectins have been used to study the differences in carbohydrate structures on Tg from thyroid carcinoma and normal thyroid tissues. Tarutani and Ui6 demonstrated that Tg from thyroid carcinoma was unretarded on concanavalin A (Con A) columns, in contrast to Tg from normal thyroid tissue. Consequently, we designed a new method for detecting Tg derived from thyroid carcinoma based on the differences of the reactivity of Tg to lectins.

MATERIALS AND METHODS Subjects

Thyroid tissue specimens were obtained at surgery, and the diagnoses of the thyroid specimens were confirmed by pathologic examination as follows: 21 specimens of papillary carcinoma, 4 specimens of follicular carcinoma, 7 specimens of follicular adenoma, 5 specimens of benign goiter, 5 specimens of Graves' disease, and 15 specimens of normal tissues (areas surrounding thyroid carcinoma). Tissues were stored at -80 deg C until use. Production of Antibodies

Rabbits were injected subcutaneously with a mixture containing 50 (mu)g of purified Tg derived from normal thyroid tissue in complete Freund's adjuvant on day 0. This was followed by three more injections on days 14, 28, and 42 in incomplete Freund's adjuvant. After test-bleed on day 56, 10 to 15 more injections with 50 (mu)g Tg in incomplete Freund's adjuvant were administered. Sera were tested for antibody binding activity by a passive particle agglutination method for Tg (Serodia-ATG; Fujirebio Inc, Tokyo, Japan), and those sera with a titer of 10 x 2^sup 10^ or higher were selected. Anti-Tg antibodies were then further purified by filtration on a Tg-sepharose 4-B column (Pharmacia LKB Biotechnology, Bucks, England) for use in Western blot analysis.

Preparation of Tg From Various Thyroid Tissues

A piece of frozen tissue was homogenized in phosphate-buffered saline (PBS, pH 7.5) containing 1 mmol/L phenylmethylsulfonyl fluoride (Sigma Biosciences, Dorset, England) and was centrifuged at 30000g for 5 minutes at 4 deg C. The supernatant was mixed with an equal volume of saturated ammonium sulfate for 30 minutes at 4 deg C, and the mixture was centrifuged at l0000g for 30 minutes at 4 deg C. Thyroglobulin-containing precipitate was dissolved in PBS and dialyzed against PBS overnight at 4 deg C. Macromolecular fractions were then purified by gel filtration on an ACA34 column (Pharmacia LKB) from the precipitate. The accurate concentration of Tg in these fractions was determined by the previously reported enzyme-linked immunosorbent assay for Tg (Tg-ELISA)7 and adjusted to a concentration of 0.1 mg/mL for ELISA and 5 mg/mL for Con A binding.

Preparation of Various Lectins and Sugar Solutions

The following lectins were used in this study: Con A, ricinus communis agglutinin-120 (RCA-120), and wheat-germ agglutinin (WGA), all obtained from Honen Oil Ltd, Tokyo, Japan. The lectins were diluted in PBS at concentrations of 0.1, 0.2, 0.3, 0.5, and 1 mg/mL before use. Sugars used for testing the specificity of binding of lectins with Tg were a-mannose (Man), 0-galactose (Gal), N-acetyl-D-glucosamine (GluNAc), N-acetyl-D-galactosamine (GalNAc), ct-fucose (Fuc), and N-acetylneuraminic acid (NANA), all obtained from Seikagaku Kogyo Co Ltd, Tokyo, Japan. The concentrations of sugar in solutions were 0, 0.01, 0.1,1, 10, and 100 mg/mL.

Treatment of Specimens With Lectins and Measurement of Tg

Fifty microliters of the partially purified Tg preparation obtained from the surgical specimens (papillary carcinoma, n = 11; follicular carcinoma, n = 4; follicular adenoma, n = 7; benign goiter, n = 5; Graves' disease, n = 5; and normal thyroid tissue, n = 5) were mixed with 50 (mu)L of various lectin solutions and incubated at 4 deg C overnight. For control experiments, samples with addition of PBS only (without lectin) were incubated in the same way. The mixtures were then centrifuged at 3000g for 20 minutes at 4 deg C to remove the lectin-bound Tg. The residual unbound-Tg (ub-Tg) in the supernatant or Tg untreated with lectin in the supernatant was determined by Tg-ELISA.7 In brief, microplates (Nalge-Nunc, Rochester, NY) were coated with 5 (mu)g of mouse anti-thyroglobulin monoclonal antibody (TgAb) in PBS and then blocked with 1% fetal calf serum in PBS. One hundred microliters of 1% fetal calf serum in PBS followed by 50 (mu)L of Tg preparation or standard Tg solutions were added to the wells, and the samples were incubated for 1 hour at 20 deg C. The plates were washed three times and then incubated for 1 hour at 20 deg C with a 1:104 dilution of TgAb horseradish peroxidase conjugate. After washing, 100 (mu)L of 3,3',5,5'-tetramethylbenzidine in a microwell peroxidase substrate system (Kirkegaard & Perry Labs Inc, Gaithersburg, Md) was added, and the plates were incubated for 30 minutes at 20 deg C. Finally, 50 (mu)L of 1 mol / L phosphoric acid (H^sub 3^PO^sub 4^) was added to stop the reaction, and the optical density at 450 nm was measured. Background values for control wells without Tg, which were subtracted from test wells, were always less than 0.05 optical density units.

Treatment of Tg With V8 Protease

One hundred microliters of Tg preparation (5 mg/mL) obtained from surgical specimens (papillary carcinoma, n = 10; normal thyroid tissue, n = 10) was incubated at 37 deg C for 1 hour with 10 (mu)L of Staphylococcus aureus V8 protease (V8; Sigma) solution (20 mg/mL) in PBS; the Tg to V8 protease ratio was 25:1.

Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis and Western Blot Analysis

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was carried out according to the method of Laemmli.8 The samples were run on the 9% SDS-PAGE; molecular weight standards (Bio-Rad, Tokyo, Japan) were included on each gel. Electrophoresis was carried out for 2 hours at constant 25 mA. Western blot analysis was performed according to the method of Towbin et al.9 The separated proteins were transferred electrophoretically onto nitrocellulose membrane for 1 hour at 150 mA and blocked for 1 hour in 1:5 diluted BlockAce (Dainihon Seiyaku Ltd, Osaka, Japan) in PBS.

Mouse Anti-Thyroglobulin Monoclonal Antibody and Con A Binding to Tg Blotted onto Nitrocellulose Membrane and Protein Staining

The membranes were incubated with a 1:50 dilution of rabbit polyclonal Tg antibody for 1 hour at 37 deg C. After rinsing, the membranes were incubated for 1 hour at 37 deg C with horseradish peroxidase-conjugated goat anti-rabbit IgG (Kirkegaad & Perry). The Tg-TgAb complex was chemically visualized using diaminobenzidine (Nakarai Kagaku Ltd, Kyoto, Japan). Or, in separate experiments, the membranes were incubated with horseradish peroxidase-labeled Con A (Dakopatts, Glostrup, Denmark) diluted 1: 300 and incubated for 1 hour at 20 deg C. After rinsing, the signal was chemically developed with diaminobenzidine. Furthermore, proteins transferred onto the membrane were stained with 1% Amido Black 10B (Wako Pure Chemical Industries, Osaka, Japan) for no more than 5 minutes and then rinsed in 7% acetic acid.

Statistical Analysis

A two-sample t test (Welch's method) was applied for the statistical analysis.

RESULTS Determination of Optimum Amount of Lectin for Incubation With Tg

In preliminary experiments, Tg preparations were incubated with lectins at different ratios (from 1:0 to 1:10), and the ub-Tg(%) in the case of papillary carcinoma ranged from 0.6% to 100% for Con A, 5% to 100% for RCA-120, and 62.4% to 100% for WGA (Table). In normal thyroid tissue, the amount of ub-Tg(%) ranged from 0.1% to 100%, 0.8% to 100%, and 91.9% to 100% for Con A, RCA-120, and WGA, respectively (Table). The highest ubTg ratio (63.0) between papillary carcinoma and normal thyroid tissue was found when Tg and Con A were incubated at a ratio of 1:3 (Table). For RCA-120, the best ubTg ratio (8.3) was found at a Tg-RCA-120 ratio of 1:3. For WGA, there was no difference between binding of lectin to Tg from papillary carcinoma and normal thyroid tissue. Consequently, in all subsequent experiments a ratio of 1:3 was used for incubations of Tg with Con A or RCA-120, and incubations with WGA were discontinued.

Effect of Lectins on Tg Determination in ELISA

The three lectins (Con A, RCA-120, and WGA) had no effect on Tg determination in ELISA up to 50 mg/mL (data not shown).

Analysis of ub-Tg(%) in Various Thyroid Tissues Results of analysis of ub-Tg(%) in samples from papillary carcinoma, benign goiter, Graves' disease, and normal thyroid tissues are shown in Fig 1. In Con A-treated samples, the ub-Tg(%) was 4.48 +/- 2.73% (mean +/- SD; n = 11) for papillary carcinoma, 0.21 +/- 0.18% (n = 5) for benign goiter, 0.36 +/- 0.20% (n = 5) for Graves disease, and 0.18 +/- 0% (n = 5) for normal tissues. The values for papillary carcinoma tissue were significantly higher than those for Graves' disease, benign goiter, and normal tissues (P

Results of analysis of ub-Tg(%) in Con A-treated samples from follicular carcinoma and follicular adenoma are shown in Fig 2. The ub-Tg(%) was 0.83 +/- 0.32% (n = 4) for follicular carcinoma and 3.82 +/- 2.12% (n = 7) for follicular adenoma. The values for the follicular adenoma were significantly higher than those for follicular carcinoma (P

Analysis of the Specificity of Tg-Lectin Binding

Addition of increasing concentrations (up to 100 mg/ mL) of Man to the reaction mixture of Tg and Con A resulted in the inhibition of binding of Tg to Con A (ie, ub-Tg increase) (Fig 3). The effects of NANA and GluNAc were smaller whereas Gal, GalNAc, and Fuc had no effect on Tg binding to Con A (Fig 3). Increasing concentrations of Gal (up to 100 mg/mL) inhibited binding of Tg to RCA120 in a dose-dependent manner, whereas Man, GluNAc, NANA, GalNAc, and Fuc had no effect (Fig 3).

Analysis of Con A Binding to Tg Blotted Onto Nitrocellulose Membrane

There were no differences between intact Tg from papillary carcinoma tissue and normal thyroid tissue in protein staining pattern, reactivity with TgAb, or Con A binding. When Tg after treatment with V8 protease was analyzed, the protein staining pattern and the reactivity with TgAb were not different for either C-Tg from papillary carcinoma tissue or N-Tg. However, Con A did not appear to bind to C-Tg after VS treatment, whereas it reacted well with N-Tg treated in the same way (data not shown).

Similar results were obtained in each case of Tg from 10 papillary carcinomas and Tg from 10 samples of normal thyroid tissue surrounding the carcinoma.

COMMENT

Thyroglobulin is a macromolecular glycoprotein of 660 kd with a sugar content of approximately 10%.10-12 There are two types of sugar chains on N-Tg, namely, Unit A and Unit B. The former is characterized by high-mannosetype sugar residues13 and the latter by mixed type.14 In contrast, the sugar chains in C-Tg have three to five branches and are composed of complex-type sugar residues.4 The characteristics of Tg in patients' serum and metastatic foci reflect characteristics of Tg in primary thyroid carcinoma.ls Therefore, the ability to distinguish CTg from N-Tg will contribute to more accurate diagnosis of thyroid carcinoma.

The principle of the method developed in this study is based on the observation that there are differences between reactivity of sugar chains in C-Tg and N-Tg with lectins. In other words, weaker binding of lectin to sugar chains in C-Tg compared with N-Tg may be detected as an increase in the ub-Tg(%). The ub-Tg could be regarded as fractions of C-Tg that were not absorbed on lectin af finity chromatography in former studies2,616,17 In the present experiment, although Tg preparations may have contained proteins other than Tg, the Tg concentration in the solutions before and after treatment with lectins could be confirmed accurately by Tg-ELISA.

First, the effect of lectins on Tg determination in ELISA was examined. The determination of standard Tg concentrations was not affected by high concentrations of lectins (up to 50 mg/ mL), therefore the presence of lectins did not interfere in the short time of reaction (1 hour) of TgAb to Tg in the first phase of Tg-ELISA.

The specificity of the reaction between the lectin and Tg was then analyzed using various concentrations of six different sugars. Concanavalin A binding was inhibited mainly by Man, but also to some extent by GluNAc and NANA (Fig 3), thus confirming Con A specificity for these types of sugar residues.18 RCA-120 appeared to react specifically with Gal (Fig 3). There were no differences between papillary carcinoma and normal thyroid tissue in the inhibition of different lectins' binding by addition of sugars, except that slightly elevated ub-Tg(%) in C-Tg without sugars was observed. These results suggest that although sugar chains on C-Tg are similar to those on NTg, there may be some heterogeneity in sugar residues associated with C-Tg. This is in agreement with the findings reported by Chang,19 who stated that the reactivity of various lectins with Tg in colloid studied by histochemical staining techniques showed heterogeneous patterns in thyroid carcinoma.

The ub-Tg(%) might depend on the reaction ratio of Tg to lectin. More binding of Tg to lectin may occur in proportion to the amount of lectin. Therefore, the optimum amount of lectin for incubation with Tg was determined as the ub-Tg ratio. The highest ub-Tg ratio (63.0) between papillary carcinoma and normal thyroid tissue was found when Tg and Con A were incubated at a 1:3 ratio (Table). In the case of RCA-120, the best ub-Tg ratio (8.3) was found at a Tg-RCA-120 ratio of 1:3.

On the basis of these preliminary experiments, ubTg(%) in Tg preparations obtained from various thyroid tissues was measured. The ub-Tg(%) in C-Tg from papillary carcinoma was significantly higher than that in other types of thyroid tissue with both Con A and RCA-120 (Fig 1).

In addition, to study further the differences in glycosylation between C-Tg in papillary carcinoma and N-Tg, Con A binding to Tg digested with V8 protease (V8 Tg) was analyzed on nitrocellulose membrane after Western blotting. VS protease, which was isolated from Staphylococcus aureus (V8 strain), specifically cleaves peptide linkages on the carboxyl terminal side of either asparatic acid or glutamic acid.20 Although there were no differences in reactivity with TgAb in the cases of both intact Tg and V8 Tg, Con A did not appear to bind to Tg from papillary carcinoma after VS treatment, whereas it reacted well with VS Tg from normal tissue. These results suggest that the analysis of reactivity with TgAb may be of less use to distinguish C-Tg from N-Tg (despite the findings that the primary protein structure as well as the conformational structures of C-Tg are different from those of N-Tg16), but the examination of reactivity with Con A in Tg after treatment with VS protease can practically contribute to the differentiation of C-Tg from N-Tg.

The differentiation of follicular carcinoma and follicular adenoma is dependent on finding capsular invasion on histopathologic examination.21 Consequently, this distinction cannot be reliably made by cytodiagnosis using fineneedle aspiration. Therefore, our findings that the ubTg(%) values for follicular adenoma are significantly higher than those for follicular carcinoma may be of use in differentiating between these entities when the aspirates obtained from thyroid nodules, showing any suspicion of follicular tumors by cytodiagnosis, are tested. Although there were no differences in ub-Tg(%) for Con A between follicular carcinoma and normal thyroid tissues, examining ub-Tg(%) with different kinds of lectins may be helpful to facilitate their differentiation.

Thus, the determination of ub-Tg(%) and analysis of lectin binding to Tg digested with VS protease (V8 Tg) should allow for the differentiation of C-Tg and N-Tg in thyroid aspirates for cytodiagnosis and consequently may be useful for the adjunctive diagnosis of thyroid carcinoma.

The authors thank J. Furmaniak, MD, FIRS Laboratories RSR Ltd (Cardiff, United Kingdom) for her critical review and helpful suggestions regarding our manuscript.

References

1. Hotta T, Ishii I, Ishihara H, Tejima S, Tarutani O, Takahashi N. Comparative study of the oligosaccharides of human thyroglobulin obtained from normal subjects and patients with various diseases. J Appl Biochem. 1985;7:98-103.

2. Nakajima H. Studies on the nature of thyroglobulin in human thyroid tumor tissue. Kitakantoh Med J. 1982;32:177-187.

3. Yamamoto K, Tuji T,Tarutani O, Osawa T. Structural changes of carbohydrate chains of human thyroglobulin accompanying malignant transformation of thyroid gland. Eur J Biochem. 1984;143:133-144.

4. Yamamoto K, Tuji T, Tarutani O, Osawa T. Phosphorylated high mannosetype and hybrid-type oligosaccharide chain of human thyroglobulin isolated from malignant thyroid tissue. Biochim Biophys Acta. 1985;838:84-91.

5. Sinadinovie J, Cvejic D, Savin S, Janic-Ziguricas M, Micic JV. Altered terminal glycosylation of thyroglobulin in papillary thyroid carcinoma. Exp Clin Endocrinol.1992;100:124-128.

6. Tarutani O, Ui N. Properties of thyroglobulins from normal thyroid and thyroid tumor on a concanavalin A-sepharose column. J Biochem. 1985;98:851857.

7. Kato R, Noguchi S, Noguchi A. Human serum thyroglobulin determination with monoclonal antibody one-step assay; minimum interference of autoantibody. Endocrinol Jpn. 1987;34:171-178.

8. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature. 1970;227:680685. 9. Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci U S A. 1979;76:4350-4354.

10. Arima T, Spiro MJ, Spiro RG. Studies on the carbohydrate units of thyroglobulin. J Biol Chem. 1965;240:1603-1610.

11. Spiro RG. The carbohydrate units of thyroglobulin. J Biol Chem.1965;240: 1603-1610.

12. Spiro MJ. Presence of a glucuronic acid-containing carbohydrate unit in human thyroglobulin. J Biol Chem.1977;252:5424-5430. 13. Tsuji T, Yamamoto K, Irimura T, Osawa T. Structure of carbohydrate unit A of porcine thyroglobulin. Biochem J.1981;195:691-699.

14. Yamamoto K, Tsuji T, Irimura T, Osawa T. Structure of carbohydrate unit B of porcine thyroglobulin. Biochem J. 1981;195:701-713. 15. Ishikita T. The existence of thyroglobulin in metastatic lymph nodes of thyroid carcinoma and significance of measurement of blood Tg level after surgery. Folia Endocrinol. 1995;71:105-114.

16. Tarutani O, Abe N, Hosono O, et al. Comparative studies on immunological property of thyroglobulins obtained from the thyroid tumor and the adjacent tissue. Folia Endocrinol.1985;61:1176-1181.

17. Sato K. The properties of serum thyroglobulin in patients with thyroid carcinoma. Kitakantoh Med J. 1988;38:353-360.

18. Ogata S, Muramatsu T, Kobata A. Fractionation of glycopeptides by affinity column chromatography on concanavalin A-sepharose. J Biochem.1975;78:678696.

19. Chang Y. Histochemical study of human thyroid; with special reference to oligosaccharide recognition by lectin stains. Shinshu Med J. 1990;38:13-39.

20. Drapeau GR, Boily Y, Houmard J. Purification and properties of an extracellular protease of staphylococcus aureus. J Biol Chem. 1972;247:6720-6726. 21. Hazard JB, Kenyon R. Atypical adenoma of the thyroid. Arch Pathol. 1954; 58:554-563.

Accepted for publication February 17, 1998. From the Department of Surgery, Shinshu University School of Medicine, Matsumoto, Japan (Drs Maruyama, Kobayashi, and Kasuga), and the Division of Medical Technology, Shinshu University School of Allied Medical Science (Mr Kato), Matsumoto, Japan.

Reprints: Masayuki Maruyama, MD, Department of Surgery 2, Shinshu University School of Medicine, Asahi 3-1-1, Matsumoto 390, Japan.

Copyright College of American Pathologists Aug 1998
Provided by ProQuest Information and Learning Company. All rights Reserved