Keywords

1 Introduction

Thyroid hormones are essential for normal development, differentiation, and metabolism of the majority of organs. Their synthesis requires intact follicles, which form the functional units of the gland, several regulated biochemical steps, and an adequate nutritional iodide uptake (Kopp 2012; Pesce and Kopp 2014). At the basolateral membrane of thyroid follicular cells, iodide uptake is mediated by the sodium-iodide symporter (NIS) (Portulano et al. 2014). The function of NIS is dependent on the sodium gradient generated by the Na+/K+-ATPase (Fig. 7.1), and a constitutively active potassium channel consisting of the KCNQ1 and KCNE2 subunits, which promotes K+ efflux (Frohlich et al. 2011; Roepke et al. 2009). At the apical membrane, iodide enters the follicular lumen where the actual thyroid hormone synthesis occurs. Within the follicular lumen, iodide is oxidized by the membrane-bound heme enzyme thyroid peroxidase. The oxidation requires the presence of hydrogen peroxide (H2O2), which is generated by the dual oxidase (DUOX) system (Moreno and Visser 2007). Subsequently, oxidized iodide is organified into defined tyrosyl residues of thyroglobulin, which serves as the scaffold for thyroid hormone synthesis. Thyroglobulin is a heavily glycosylated and very large protein (330 kDa) that forms dimers (Di Jeso and Arvan 2016). In a first step, referred to as organification, the iodination leads to the formation of mono- and diiodotyrosines (MIT, DIT). A donor and an acceptor iodotyrosine are then fused in the coupling reaction, which generates T4 or T3. Iodinated thyroglobulin is digested by several endopeptidases, both in the follicular lumen and after uptake into the cells through macro- and micropinocytosis (Di Jeso and Arvan 2016). T4 and T3 are secreted into the bloodstream at the basolateral membrane, at least in part by the monocarboxylate transporter MCT8 (Di Cosmo et al. 2010). Remarkably, MIT and DIT are sorted and then deiodinated by an intracellular iodotyrosine dehalogenase (DEHAL1), which permits to recycle the released iodide into the follicular lumen (Moreno et al. 2008; Kopp 2012).

Fig. 7.1
figure 1

Cellular localization (top) and putative structure (bottom) of pendrin (PDS/SLC26A4) and anoctamin 1 (ANO1/TMEM16A). Pendrin contains a carboxyterminal STAS (sulfate transporter and antisigma factor antagonist) domain that contains a protein kinase A phosphorylation site. The predicted calmodulin- and calcium-binding sites of ANO1 are indicated. The sodium-iodide symporter (NIS) transports iodide into the follicular cells at the basolateral membrane against the electrochemical gradient. NIS is dependent on the sodium gradient generated by the Na+/K+-ATPase. At the apical membrane, pendrin is thought to function as a iodide/chloride exchanger, and anoctamin is a channel that may function in conjunction with other channels such as the transient receptor potential channel 2 (TRPC2)

Full size image

2 How Does Iodide Enter the Follicular Lumen?

In electrophysiological studies performed more than two decades ago with inverted thyroid membrane vesicles, Golstein et al. suggested that iodide efflux is mediated by two distinct channels (Golstein et al. 1992). The first one, referred to as the iodide channel, exhibited a high permeability and specificity for iodide with an approximate Km of 70 μM. The second channel was found to be about fourfold more permeable to iodide than chloride with a Km of about 33 mM. The authors postulated that the iodide channel is restricted to the apical membrane and transports iodide from the cytosol into the colloid space, whereas the second one may mediate predominantly chloride transport under physiological conditions. So far, the molecular identity of these two conductances, which could be anion channels or transporters, has however not been identified. Currently, two candidates have been considered: pendrin and anoctamin (Silveira and Kopp 2015) (Table 7.1 and Fig. 7.1). Chloride channels such as the cystic fibrosis transmembrane conductance regulator (CFTR) and the voltage-gated chloride channel 5 (CLCN5) are also expressed in thyroid follicular cells (Li et al. 2010; van den Hove et al. 2006). While they have affinity for iodide, they are not thought to play a role in mediating iodide efflux under physiological conditions. SLC5A8, a homologue of NIS, which was initially called human apical iodide transporter (hAIT) (Rodriguez et al. 2002), is clearly not involved in apical iodide efflux as formally demonstrated in functional studies in oocytes and polarized MDCK cells (Paroder et al. 2006).

Table 7.1 Observations for and against a physiological role of pendrin and anoctamin as mediators of iodide efflux in thyroid follicular cells
Full size table

Thyroid-stimulating hormone (TSH) rapidly activates the efflux of iodide efflux at the apical membrane through the protein kinase A and C pathways (Weiss et al. 1984; Nilsson et al. 1990, 1992; Iosco et al. 2014). These pathways may be differentially regulated depending on the physiological status.

3 Pendrin

The classic phenotype of the autosomal recessive Pendred syndrome (OMIM #274600) consists of sensorineural deafness associated with inner ear malformations, especially enlarged vestibular aqueduct (EVA), goiter, and a partial iodide organification defect (PIOD) (Pendred 1896; Morgans and Trotter 1958; Bizhanova and Kopp 2010). Pendred syndrome is caused by biallelic (homozygous and compound heterozygous) mutations in the PDS/SLC26A4 gene (Everett et al. 1997). Functionally, pendrin belongs to the SLC26 family of multifunctional anion transporters (Alper and Sharma 2013), and it was found to have affinity for iodide (Scott et al. 1999).

Because of the human phenotype with a partial iodide organification defect (PIOD), goiter, and in a subset of subjects congenital or acquired hypothyroidism (Gonzalez Trevino et al. 2001; Ladsous et al. 2014), the expression of pendrin at the apical membrane of thyroid follicular cells (Royaux et al. 2000) (Fig. 7.1), and its affinity for iodide (Scott et al. 1999), it is a plausible candidate for mediating iodide efflux at the apical membrane.

Initial studies in Xenopus oocytes demonstrated that pendrin is able to mediate uptake of anions such as chloride and iodide in a sodium-independent manner (Scott et al. 1999). In transfected unpolarized mammalian cells, pendrin was then shown to mediate iodide release (Yoshida et al. 2002). Next, studies performed in polarized cells expressing NIS at the basolateral membrane and pendrin at the apical membrane in a bicameral system demonstrated that pendrin can mediate vectorial iodide efflux at the apical membrane (Gillam et al. 2004). Importantly, more recent studies performed in oocytes have shown that pendrin functions as a coupled, electroneutral iodide/chloride, iodide/bicarbonate, or chloride/bicarbonate exchanger with a 1:1 stoichiometry, and that it has a preferential affinity for iodide, even in the presence of high chloride concentrations (Shcheynikov et al. 2008). Moreover, it has been shown that pendrin is expressed at the apical membrane of parotid gland ducts where it can mediate luminal iodide secretion (Shcheynikov et al. 2008).

Some of the more than hundred naturally occurring mutations have been tested functionally, mainly after transfection into heterologous mammalian cells. Disease-causing mutations result in a complete or partial loss in iodide efflux (Gillam et al. 2004; Taylor et al. 2002; Pera et al. 2008; Dossena et al. 2009, 2011), and many mutated proteins are retained in intracellular compartments such as the endoplasmic reticulum secondary to misfolding (Rotman-Pikielny et al. 2002). The human phenotype, which is characterized by goiter development under conditions of scarce iodide intake (Gonzalez Trevino et al. 2001), as well as the PIOD, is suggestive for a physiological role of pendrin in mediating or participating in iodide efflux in humans. In contrast, however, Slc26a4 −/− mice do not develop a goiter or abnormal thyroid hormone levels, even under conditions of iodine deficiency (Calebiro et al. 2011; Iwata et al. 2011). Although this has been used as an argument against a physiological role of pendrin in mediating iodide efflux in thyrocytes (Twyffels et al. 2011), it is currently unclear whether this simply reflects a species difference (Bizhanova and Kopp 2011).

Iodide efflux at the apical membrane is rapidly accelerated by TSH (Nilsson et al. 1990, 1992; Weiss et al. 1984). Stimulation of the protein kinase A pathway in rat thyroid cells results in a rapid increase in membrane insertion of pendrin and an increased iodide efflux (Bizhanova et al. 2011; Pesce et al. 2012). Stimulation with forskolin increases the membrane insertion of pendrin in PCCL3 rat thyroid cells, presumably through activation of a protein kinase A site in the intracellular carboxyterminus (Bizhanova et al. 2011). Deletion or targeted mutation of the protein kinase A site residing in the intracellular carboxyterminal tail containing the so-called STAS (sulfate transporter and antisigma factor antagonist) domain results in decreased basal function and membrane insertion of pendrin (Bizhanova et al. 2011). In addition, the protein kinase A site mutation (T717A) is partially functional, but it has a mitigated response to forskolin (Bizhanova et al. 2011). Trafficking of murine pendrin to the apical membrane is also activated by cAMP in microperfused mouse cortical collecting duct (CCD) and in polarized renal opossum kidney proximal tubule (OKP) cells via phosphorylation of S49 in the aminoterminal intracellular domain (Azroyan et al. 2012). In contrast to the rapid translocation mediated by the protein kinase A pathway, stimulation of the protein kinase C pathway in rat thyroid cells appears to result in a delayed translocation of pendrin to the plasma membrane (Muscella et al. 2008). Interestingly, rat thyroid cells incubated with supraphysiological amounts of iodide show an increased abundance of pendrin at the plasma membrane, the half-life of pendrin increases, and the efflux of iodide is enhanced (Calil-Silveira et al. 2016). These findings suggest that pendrin could have a role in mediating iodide efflux under conditions of iodide excess.

Under conditions of normal or abundant iodide intake, goiter development is unusual (Sato et al. 2001), but goitrous congenital and overt hypothyroidism developing later in life can be present in patients with Pendred syndrome (Gonzalez Trevino et al. 2001; Ladsous et al. 2014). In a study from Northern France, a region which has a normal to marginal iodide intake, Ladsous et al. characterized the thyroid phenotype in patients with Pendred syndrome and non-syndromic EVA (Ladsous et al. 2014). Fifteen out of the 19 patients with Pendred syndrome (79 %) presented with a goiter. Fifteen (79 %) subjects had hypothyroidism: 6/15 had congenital hypothyroidism, 5/15 had overt hypothyroidism, and 4/15 had subclinical hypothyroidism. Ten out of 16 (63 %) of these patients showed abnormal iodide organification as determined by a perchlorate test, a test that is unfortunately poorly standardized and subject to exogenous (e.g., iodine intake) and endogenous (e.g., autoimmune thyroid disease) modulators. The study by Ladsous et al. clearly demonstrates that there is a relatively wide spectrum in the thyroid phenotype among patients with Pendred syndrome, suggesting that it is influenced by genetic and environmental modifiers, including nutritional iodine intake (Ladsous et al. 2014). Intriguingly, biallelic mutations in the SLC26A4 gene have also been identified in two patients with thyroid hypoplasia and congenital hypothyroidism from two unrelated families (Kühnen et al. 2014). The mutations found in these subjects have been previously identified in patients with the classical form of Pendred syndrome or familial EVA (Kühnen et al. 2014; Kopp 2014). The reasons why these two patients developed thyroid hypoplasia, rather than a goitrous phenotype, are unclear. It has been speculated that either the retained misfolded proteins or an increased production of free radicals in response to sustained stimulation by TSH could have a toxic effect leading to cell death, or that the hypoplastic phenotype requires the presence of additional modifying (genetic) factors (Kopp 2014; Kühnen et al. 2014).

4 Anoctamin

Three recent studies have suggested that anoctamin 1 (ANO1), also referred to as TMEM16A, could be involved in apical iodide efflux in thyroid cells (Viitanen et al. 2013; Iosco et al. 2014; Twyffels et al. 2014). ANO1 is a calcium-activated anion channel, which is expressed in numerous tissues, including thyroid follicular cells (Pedemonte and Galietta 2014; Ferrera et al. 2010). ANO1 is part of a family of ten paralogs (ANO1-10; TMEM16A-K) sharing a common transmembrane topology, but a wide spectrum of in part putative functional roles as ion channels, regulatory subunits of other channels or phospholipid scramblases, proteins responsible for the translocation of phospholipids between the two monolayers of the cell membrane lipid bilayer (Pedemonte and Galietta 2014; Picollo et al. 2015). The ANO1 gene generates several splice variants, and most of them have a higher affinity for iodide than chloride (Ferrera et al. 2010). Human and rat thyroid cells predominantly express the so-called abc and the ac isoforms (Ferrera et al. 2009; Iosco et al. 2014), whereas the rat thyroid cell lines PCCL3 and FRTL-5 predominantly express the ac isoform, which is more sensitive to calcium (Ferrera et al. 2009, 2010). A functional study by Viitanen et al. performed in native FRTL-5 rat thyroid cells suggested that ANO1, in conjunction with the transient receptor potential channel 2 (TRPC2), mediates iodide release (Viitanen et al. 2013). Iosco et al. then demonstrated that the ANO1 protein is localized at the apical membrane of human thyrocytes (Fig. 7.1), and that its expression is more abundant in active cells (Iosco et al. 2014). Functional studies determining the intracellular iodide content revealed that iodide release can be stimulated by adenosine triphosphate (ATP) in a calcium-dependent manner from FRTL-5 cells, whereas treatment with inhibitors or siRNA knockdown decreased the iodide efflux. Similarly, iodide efflux was also increased in transfected mammalian cells expressing both NIS and ANO1, and iodide release could be further stimulated by calcium (Iosco et al. 2014). Twyffels et al. demonstrated that Ano1 mRNA is stimulated by TSH and that the protein expression, which is relatively discrete under basal conditions, increases after stimulation by TSH (Twyffels et al. 2014). In unpolarized human embryonic kidney (HEK293) cells transfected with ANO1, the efflux of iodide is increased compared to untransfected cells. However, in HEK293 cells expressing pendrin, the iodide efflux is higher compared to ANO1 transfected cells, suggesting that pendrin is more efficient in mediating iodide efflux, at least in this model system (Fig. 7.2). Treatment with the calcium ionophore ionomycin was found to acutely stimulate the ANO1-mediated iodide release (Twyffels et al. 2014). The mechanism by which calcium activates ANO1 seems to involve calmodulin as well as direct calcium binding to ANO1 (Pedemonte and Galietta 2014). Rat thyroid cell lines and human primary thyroid cells treated either with an ANO1 inhibitor (T16Ainh-A01) or siRNA show a decrease in iodide release (Twyffels et al. 2014). Studies in oocytes have demonstrated that ANO1 is a calcium-activated chloride channel with a preference for iodide over chloride (Schroeder et al. 2008; Yang et al. 2008) indirectly suggesting that it could be involved in mediating iodide efflux. ANO1 is able to mediate iodide efflux from FRTL-5 cells after ATP stimulation in the absence of chloride, which suggests that it functions independently of pendrin, which functions as an anion exchanger. In aggregate, these results indicate that ANO1, which has a preferential affinity for iodide over chloride (Schroeder et al. 2008; Yang et al. 2008), is able to mediate iodide release from thyroid cells.

Fig. 7.2
figure 2

Release of radioiodide from 125I-preloaded PCCl3 cells (○) or HEK 293 T cells transfected to express Na+/I− symporter (NIS) in combination with GFP (●) or ANO1 (▲) or pendrin (■). The data are representative of more than five independent experiments (From Ref. Twyffels et al. (2014). With permission). Note that the release of iodide is more efficient in these pendrin-transfected cells compared to ANO1-transfected cells

Full size image

5 Future Directions

In conclusion, the current body of data suggests that the multianion exchanger pendrin (PDS/SLC26A4) and the calcium-dependent channel anoctamin 1 (ANO1/TMEM16A) can mediate apical iodide efflux in thyroid cells and several model systems. It is conceivable that they are part of a redundant system. The exact physiological role of pendrin and ANO1 awaits further characterization, and it may be variable between basal conditions and conditions of thyroid dysfunction as illustrated by the regulation of ANO1 expression by TSH, and the differential regulation of pendrin trafficking by the protein kinase A and C pathways. As suggested by the absence of a thyroid phenotype in Slc26a4 −/− knockout mice, which differs with the human phenotype that can include goiter and congenital or acquired hypothyroidism, there may be relevant differences in the expression pattern and physiological roles of pendrin and ANO1 between species.