Ultrastructural aspects on iodination and hormone secretion in the thyroid gland

Colloidal aggregates of insoluble inclusions in human goiters

To shed some light on the physicochemical properties of the thyroid follicular colloid, we have screened retrospectively the autoradiographs of 60 human nodular goiters labeled 17 h preoperatively with 100 microCi 125I for evidence of colloid compartmentalization. In 87% (52/60) of all goiters examined, sporadic or multiple colloidal inclusions ('colloid stones') not mixing with newly labeled Tg were detected. The detailed analysis of 17 goiters revealed a mean incidence of 0.09+/-0.11 'colloid stones' of variable size per follicle ranging from 0.02+/-0.01 (10) to 0.43+/-0.09 (5) (mean values +/- S.D., number of sections examined in brackets). In this study we did not find a clear-cut association of incidence of 'colloid stones' with sex, age or nosologic group (hyperthyroid, preclinically hyperthyroid, euthyroid). The existence of different colloidal compartments as demonstrated in this and other studies is of considerable importance for thyroid function, interpretation of iodine kinetics, and studies on the role of iodine on growth and function of the thyrocytes. Different thyroidal iodine compartments could well be of functional relevance, for example in the adaptation of thyroid hormone secretion to antithyroid drugs or in severe and prolonged iodine deficiency, when very slow compartments become an important source of minimal quantities of iodine and thyroid hormone. 'Colloid stones', for example, may well explain the repeatedly observed, surprisingly large total iodine store in human endemic goiters, even in the presence of severe iodine deficiency. It is evident that the existence of multiple iodine compartments and, in particular, of particulate slow-turnover pools complicates the interpretation of total glandular iodine measurements with modern techniques such as X-ray fluorescence and positron emission tomography.

Studies on functional and morphological aspects in human multinodular simple goiter tissues

Samples from 2 different locations within the same euthyroid multinodular goiters (SMG) and normal (N) human thyroids were assayed for their content of DNA, thyroglobulin (Tg), and stable iodine (¹²⁷1), and determined the response of adenosine 3',5'-cyclic monophosphate (cAMP) to TSH and NaF. Quantitative morphological estimation of histological components in the thyroid was performed and correlated with functional parameters. Regardless the zonal evaluation, in SMG the mean (± SD) DNA content (⧎g/mg tissue) (1.04 ± 0.86) was not statistically different from that in N (1.13 ± 0.21). The mean¹²⁷I concentration (⧎g/⧎g DNA) in N tissues (0.357 ± 0.091) was greater than that in SMG (0.176 ± 0.074). In these tissues, the Tg mean level (± SD) (⧎g/⧎g DNA) was lower (28.3 ± 21.5) than that in N (75.6 ± 41.1). The mean relative proportion (Vv) of epithelial cells in SMG (range, 6.0-30.6%) was statistically different (p <0.00) from that observed in N tissues (range, 10.4-18.2%). The meanbasal (± SD) cAMP level (pmol/⧎g DNA) in these tissues (0.11 1 ± 0.036) was different (p < 0.05) from that in SMG (0.231 ± 0.026). In response to TSH (10 mU), both SMG and N increased their cAMP contents to 0.454 ± 0.045 and 0.572 ± 0.020, respectively. A further elevation in cAMP levels was observed in N (1.154 ± 0.210) after 75 mU TSH, whereas in SMG tissues, no consistent increase (0.609 ± 0.496) occurred. Goiter and normal thyroid slices were unable to increase their cAMP concentrations in response to NaF in vitro. No correlation was found between functional and morphological data in SMG samples. In contrast, this relation was quite uniform in normal thyroids. The results are concordant with the intrathyroidal pathogenic processes often cited for the heterogeneity in human goiter.

Structural and functional aspects of the thyroid follicular epithelium

The thyroid epithelium is morphologically and functionally polarized, with an apical surface facing the follicular lumen containing colloid and a basolateral surface facing the interstitium. Iodination and thyroid hormone synthesis occur in the colloid at the apical plasma membrane. The introduction by Mauchamp et al. of primary cultures of porcine thyroid cells grown as a polarized, confluent monolayer on a filter in a bicameral chamber system has now made it possible to study in more detail the barrier function and vectorial ion transport in the thyroid epithelium. The follicular cells form a very tight monolayer (transepithelial resistance > 6000 ohm cm2) and establish a transepithelial potential difference (apical medium negative) of about 20 mV. These parameters are rapidly influenced by TSH, mainly by an action on apical sodium channels, and by EGF. The integrity of the barrier is, as in other epithelia, dependent on extracellular calcium. A calcium-dependent cell adhesion molecule, uvomorulin, is expressed at the lateral plasma membrane surface. EGF induces cell proliferation as well as migration of some of the epithelial cells to a position below the monolayer, which however maintains its polarity and barrier function. In contrast, during TPA-induced proliferation the barrier function is disrupted. Iodide is vectorially transported in basoapical direction while the epithelial layer is virtually impermeable for iodide transfer in the opposite direction. Iodide is concentrated in the cell by the basolateral "iodide-pump" and its efflux across the apical plasma membrane is rapidly and selectively increased by TSH via cAMP. EGF inhibits vectorial basoapical iodide transport mainly by reducing the iodide permeability of the apical plasma membrane. Together, these recent observations indicate that the ion content of the follicular lumen is strictly controlled by the thyroid epithelium.

Formation of molecular iodine during oxidation of iodide by the peroxidase/H2O2 system. Implications for antithyroid therapy

The first step in the biogenesis of thyroid hormones is the oxidation of iodides taken up by the thyroid gland. Oxidation of I- by the H2O2/peroxidase system leads to the formation of iodinium ions I+ which bond to thyroglobulin by electrophilic substitution. However, it is not clear whether I- is transformed directly to I+ or whether it passes through a molecular iodine intermediate. This latter possibility is indicated by the oxidation potentials of the reactions. I2 can be detected in vitro from the formation of I3- ions, although this has yet to be confirmed in vivo. The present study was designed to determine, albeit indirectly, whether this reaction occurs in vivo. If I2 is produced, it may form charge transfer complexes with numerous drugs. We also investigated the action of various drugs on lactoperoxidase and assessed their antithyroid activity in the rat by assay of plasma levels of T3, T4, and TSH. We found a good correlation between the value of Kc, the formation constant of the complex of the drug with molecular iodine, and the antithyroid activity in vivo. This correlation was observed in four different classes of compound. The possibility that molecular iodine is produced in the thyroid gland has implications for antithyroid therapy.

Intrathyroidal iodine heterogeneity of iodocompounds and kinetic compartmentalization

In the normal thyroid, but not necessarily in the goitrous gland, the bulk of iodine is bound to thyroglobulin. Even in the normal thyroid-and much more so in goiters-iodine is contained in many different compartments with widely differing kinetics, biochemical composition, localization, and physiologic significance. Any change of thyroid function profoundly affects intrathyroidal iodine kinetics and produces a redistribution of stored iodine. This must be taken into account whenever the impact of a global change in intrathyroidal iodine stores on thyroid function and growth is studied in vivo.

Immunocytochemical study of localization and traffic of thyroid peroxidase/microsomal antigen

We studied the distribution of binding sites for anti-peroxidase monoclonal antibody and anti-microsomal antibodies on isolated human thyroid follicles and a human thyroid cell line. Both open follicles and cells were incubated first with antibodies at +4 degrees C, then with colloidal gold labelled protein A. The topography of the binding sites for monoclonal anti-peroxidase antibody corresponded closely to the expected cell surface distribution of endogenous thyroid peroxidase since labelling was observed at the apical cell surface of the follicles. Furthermore, labelling was restricted to the microvilli level; while smooth membrane territories were devoid of binding sites. In some cases, incubations at 4 degrees C were followed by warming the follicles and cells up to 37 degrees C for 20 minutes in order to study internalization of ligands. Ligands were then observed in intracellular organelles: endosomes and lysosomes. Essentially the same results were observed when human antibodies to the microsomal antigen were used. Controls with microsomal antibodies depleted in anti-peroxidase were negative. In conclusion these findings show that: 1) thyroid peroxidase is present in limited areas on the apical cell surface, 2) labelling of follicles and cells by the anti-microsomal antibodies had the same pattern of distribution as the monoclonal anti-peroxidase antibody, thus suggesting that they recognize the same apical antigens, and 3) TPO/MIC antigen traffics from the cell surface towards lysosomes when the cells are incubated at 37 degrees C.

Plasma membrane shedding and colloid vacuoles in hyperactive human thyroid tissue

The ultrastructural appearance of colloid vacuoles, considered to be a typical sign of hyperactivity in the human thyroid gland, was studied in human thyroid tissue transplanted to nude mice and in human thyroid tissue fixed directly after surgical removal in patients with thyrotoxicosis. Transplanted normal thyroid tissue and toxic diffuse goiter (TDG) tissue was fixed by vascular perfusion with glutaraldehyde 5 or 12 weeks after transplantation. Light microscopic quantification showed that daily injections for 2 weeks of a gamma globulin fraction of patient sera containing thyroid-stimulating immunoglobulins (TSI) greatly increased the number of colloid vacuoles in both types of transplants. The vacuoles were mainly located in the periphery of the follicle lumen, giving the colloid a scalloped appearance. Electron microscopy of TSI-exposed tissue revealed, in addition to colloid vacuoles, the presence of large amounts of membrane material in the follicle lumen. Only sparse amounts of intraluminal membrane material were present in controls. The colloid vacuoles were almost invariably associated with such membrane material, which lined the border between the vacuole and the surrounding colloid. The intraluminal material consisted of spherical and elongated formations, each structure limited by a triple-layered membrane and often containing a dense interior. The elongated structures were often of the same dimensions as microvilli. The apical surface of follicle cells in TSI-exposed tissue expressed numerous microvilli, of which many showed a similar dense interior as the intraluminal membrane structures. The intraluminal membranes frequently showed, like the apical plasma membrane of the follicle cells, a positive reaction for peroxidase. Organelles, such as mitochondria, lysosomes or rough endoplasmic reticulum, were not encountered among the intraluminal membrane structures. These observations indicate that the intraluminal membrane material is derived from the apical plasma membrane of the follicle cells, presumably by shedding of microvilli. A similar association between colloid vacuoles and membrane material was also found in thyroid tissue from patients with thyrotoxicosis fixed directly at operation. It is suggested that the presence of membrane material in the follicle lumen precipitates the formation of colloid vacuoles in hyperactive thyroid tissue. The possible involvement of intraluminal membrane material in the development of microsomal autoantibodies in Graves' disease, i.e. exposure and presentation of thyroid microsomal antigen (identical to thyroperoxidase) to the immune system, is discussed.

Forskolin-induced elevation of cyclic AMP does not cause exocytosis and endocytosis in rat thyroid follicle cells in vivo

Using a newly developed infusion technique, the in vivo effects of forskolin and dibuturyl cyclic AMP (dbcAMP) on exocytosis and endocytosis in thyroid follicle cells were studied in thyroxine-treated rats and mice. Reactants were selectively infused via the superior thyroid artery to one thyroid lobe. The contralateral lobe served as a control. In the rat, a supramaximal i.v. dose of thyrotropin (TSH, 500 mU) induced a slight increase in thyroidal tissue levels of cyclic adenosine monophosphate (cAMP) while TSH 50 mU i.v. had no effect on cAMP levels. On the other hand both doses of TSH stimulated exocytosis, signified by a decrease in the number of exocytotic vesicles and endocytosis, signified by the appearance of pseudopods and colloid droplets. Selective thyroid infusion of dbcAMP (5 mM) or forskolin (25 microM), which induced a 10-fold increase in thyroid cAMP levels, did not induce any morphological sign of exocytosis or endocytosis in the follicle cells. The morphological response to TSH given i.v. was quantitatively unaltered by simultaneous infusion of forskolin. In contrast to the findings in rats, infusion of forskolin and dbcAMP in mice induced endocytosis. In conclusion, our findings in the mouse are in agreement with earlier studies in this and other species, indicating that cAMP mediates the effects of TSH on endocytosis and probably also on exocytosis. In contrast, our observations in the rat thyroid in vivo lead to the conclusion that cAMP is not the main intracellular mediator of exocytosis and endocytosis in this species. This conclusion is at variance with previous reports, mostly from in vitro studies.(ABSTRACT TRUNCATED AT 250 WORDS)

Synthesis and apical and basolateral secretion of thyroglobulin by thyroid cell monolayers on permeable substrate: modulation by thyrotropin

In the thyroid glands, thyroglobulin (Tg) is specifically synthesized by follicular cells and then secreted into the apical lumen where it is concentrated and used as a substrate for thyroid hormone synthesis. The presence of Tg in the circulation has been reported in normal and pathological situations. To determine the domains of the plasma membrane, apical and/or basolateral, involved in Tg secretion, porcine thyroid epithelial cells were cultured as a monolayer on the porous bottom of a culture chamber in which both apical and basal media are independently accessible. Control experiments using labeled Tg ascertained the tightness of the monolayer and showed that within 48 h only 0.2-0.5% of the Tg introduced in the apical medium was transferred through the cell layer into the basal compartment. For kinetic studies of Tg synthesis and secretion, monolayers were cultured for up to 72 h in the presence of 35S-methionine and with or without 100 microU/ml thyrotropin (TSH) in the basal medium. Labeled Tg was measured by double immunoprecipitation and by fluorography of polyacrylamide gel electrophoresis. We showed that 80-95% of total secreted Tg was recovered in the apical medium. The remainder was secreted through the basolateral membranes in the basal medium. The amount of tg secreted into the apical compartment was stimulated two- to threefold by TSH whereas no TSH effect was observed on secretion in the basal compartment. Moreover, measuring apical and basal volumes, we observed a net water flow from the apical to the basal side. It was stimulated threefold by TSH, increasing the Tg concentration in the apical compartment of the stimulated cell layer. During the culture time, the amount of Tg synthesized and secreted was increased by TSH, as was the Tg mRNA content, as determined by the dot-blot hybridization method.

Effects of TSH on iodination in rat thyroid follicles studied by autoradiography

The possible functional relation between TSH-stimulated exocytosis and TSH-stimulated iodination in the thyroid gland was studied using quantitative EM autoradiography and cytochemistry. The study was performed in rats pretreated with thyroxine for 2 days. TSH, giving i.v. 10 min before sacrifice to thyroxine-treated rats, increased organification of 125I by about 50%. TSH decreased the number of peroxidase-positive vesicles in the apical cytoplasm and increased the width of the peroxidase reaction at the apical plasma membrane, suggesting a redistribution of peroxidase. EM autoradiography after labelling with [3H]leucine showed that TSH caused a rapid redistribution by exocytosis of newly synthesized protein to the follicle lumen. The protein deposited in the lumen remained to a large extent in the microvillus region. 10 min after injection of 125I, newly iodinated protein was distributed in a gradient in the lumen periphery. TSH, given 5 min before 125I, caused a significant increase in the labelling of the colloid in the microvillus region, indicating a selective incorporation of iodine into newly synthesized molecules deposited in this region by stimulated exocytosis. Our results confirm and extend earlier observations on a functional link between exocytosis and iodination. Redistribution of peroxidase as well as newly synthesized protein to the site of iodination might be of importance.