How many iodide transporters are there ?
How many true iodide transporter do we know ?

Iodide entering thyroid gland and its functional units, the follicles, must sequentially cross two lipid bilayers: the basolateral and then the apical plasma membrane of thyrocytes (Step1 and Step 2 – Fig.1) to reach finally the lumen of follicles. In this compartment, iodide is oxydized and used for the generation of iodothyronine residues inside thyroglobulin molecules. Under certain circumstances such as an alteration of iodide oxidation…., iodide can leave the lumen of follicles by crossing the same plasma membrane domains but in the reverse direction (Step 3 and Step 4 – Fig. 1).

Fig.1: Schematic representation of iodide fluxes in the thyroid follicle. [ I ^-]_E , [ I ^-]_C and [ I ^-]_L are iodide concentration in extracellular fluid, cytoplasm and lumenal compart-ment, respectively. Numbers identify the different steps of the thyroid iodide transport. The dotted line illustrates the re-utilization of iodide coming from the deiodination of iodotyrosines generated by proteolysis of thyroglobulin (Tg).

1- Iodide transport – General considerations
As iodide and small charged molecules cannot easily diffuse through cell membranes, its transfer from one side to the other side of the plasma membrane requires a membrane protein. As the cell interior is negatively charged with respect to the extracellular milieu or to the follicle lumen, the transport of iodide inside thyroid cells (iodide influx) at either pole requires energy. Thus, the membrane protein involved in either Step 1 or Step 3 must be an active transporter i.e. an ATP-driven pump or a Na+ gradient-dependent transporter. On the opposite, the movement of iodide from the cytoplasm of thyrocytes to either the follicle lumen or to the extracellular (extrafollicular) milieu (iodide efflux) is expected to be a passive process; it takes advantage of the favorable electrochemical gradient. Thus, Step 2 and Step 4 would require a membrane protein with a function of "permease" or a function of ion channel. From these considerations and experimental data, it is clear that the influx and efflux reactions occuring at a pole of thyrocytes (either basolateral or apical) are not mediated by a given protein capable of transporting iodide in either direction.
It is thus reasonnable to think that thyrocytes could express four distinct membrane proteins with either a function of iodide transporter or a function of iodide channel ; a high selectivity towards iodide would only be required for some of them. At present, how many of these proteins do we know ?

2- Iodide transport – Step 1.
The nature and the main properties of the active iodide transport system located at the basolateral plasma membrane of thyrocytes allowing the uptake and the concentration of iodide in the thyroid is known since several decades (1,2). It has formally been identified ten years ago (3) and named NIS for Na+/Iodide Symporter. Since that time, NIS expression and activity has been the subject of a large number of studies and reviews (4). NIS belongs to the solute linked carrier (SLC) transporter family and more precisely to the SLC5A subfamily of Na+-dependent transporters as the SLC5A5 member. NIS is characterized by a high selectivity for iodide especially towards chloride; it transports complex anions such as perchlorate which exhibits a size comparable to that of iodide. Thus, perchlorate ion is a competitive inhibitor of the iodide transport by NIS (5,6).

3- Iodide transport – Step 2.
The transport of iodide from the cytoplasm to the follicle lumen has been elegantly studied on polarized porcine thyrocytes cultured as tight monolayers in bicameral devices by Nilsson and his colleagues (7,8); they proposed that the apical efflux of iodide, which is rapidly increased in response to TSH, could occur through a perchlorate-insensitive, cAMP-regulated iodide channel. Independently, it was reported that an inhibitor of anion channels, DIDS, completely blocks apical iodide efflux from polarized thyrocytes (9). A study performed on thyroid membrane vesicles (10) also concluded to the existence of an iodide channel in the thyroid; however, the authors could not assigned a precise location (in relation with the cell polarity) to this channel.
These data diriving from fuctional studies indicate that the membrane protein insuring the apical iodide efflux (expected to be a passive process) might be an anion channel.

Information about the molecular identity of the protein insuring apical iodide transport has been provided by genetic and genomic analyses. The discovery that pendrin, the protein encoded by the gene altered in the Pendred syndrome i.e. the PDS gene, is a membrane protein located at the apical pole of thyrocytes with an ion transport activity rapidly gave rise to over-interpretation (11). Indeed, the scientific community interested in the thyroid field was waiting for the identification of the apical iodide transporter ; it was claimed that pendrin might be an iodide transporter; pendrin was very rapidly considered as the apical iodide transporter despite the uncertainties and/or discrepancies which were apparent since the first reports. The initial proposal of pendrin as a sulphate transporter (12) was rapidly abandoned (13). Most subsequent studies performed on different experimental systems assigned a function of anion exchanger to pendrin exchanging chloride for OH-, for bicarbonate or for formate (14-17). Some studies concluded on an activity of chloride/iodide transporter(18,19) or chloride/iodide exchanger (20). If pendrin transports chloride and iodide, it is difficult to envisage a physiological contribution of pendrin to apical iodide efflux since the cytoplasmic chloride concentration is likely to be more than 1000- fold greater than that of iodide. An exchange of cytoplasmic iodide with lumenal chloride on a 1 to 1 stoechiometry seems very unlikely considering the iodide and chloride concentrations. Still more confusing are the reports claiming that pendrin mediates iodide uptake (not efflux) by MCF-7 cells (21,22) or that pendrin has an activity of sugar transporter (23).
At present, there is no experimental data showing that pendrin activity accounts for the "transfer" of iodide from thyrocytes to the apical compartment in a polarized thyroid cell system. The result of the study performed on NIS expressing polarized MDCK cells transfected with the PDS gene (24) although interesting should not be extrapolated and considered as the demontration of the function of pendrin in the thyroid. Several other arguments make the hypothesis of pendrin as an apical iodide transporter very uncertain. The knock out of the pds gene in the mouse, which causes functional alterations in the ear, does not induce any thyroid phenotype (25). It has often been quoted that many patients with the Pendred syndrome having a non-functional pendrin do not exhibit any thyroid alteration.
As already discussed by one of the pioneer and main contributor of the "iodide transport" field (26), it seems difficult to admit that a given protein could have different functions in different organs. As a simple suggestion, pendrin (SLC 26 A4 in the SLC transporter nomenclature) expressed in the thyroid, could exert (as in the kidney) a function of chloride/bicarbonate exchanger (the most admitted function) and play, for exemple, a role in the control of lumenal pH with a possible incidence on iodide oxidation.
Indecision about the identity of the apical iodide transporter : pendrin or nor pendrin ?, has been complicated by a report describing a potentiel "competitor" for the same function named AIT for Apical Iodiode Transporter (27). As pendrin, AIT appears selectively located at the apical plasma membrane of thyrocytes and, as pendrin, was reported to cause iodide discharge in transfected cell systems. The first problem arose when it was found, using colon cell lines, that this protein, structurally related to NIS was a Na+-dependant transporter (28) belonging to the SLC5 family as the SLC5A8 member. Indeed, both previous studies and prediction (on biological considerations) do not point to a Na+ dependency for the apical iodide transport. Soon after, two independent groups (29-31) described SLC5A8 as a Na+- coupled transporter for short-chain fatty acids or as a Na+/ monocarboxylate co-transporter. Analyses of the recent studies on SLC5A8 indicate a consensus on this function (32,33) . Attempts by several investigators to confirm an activity of SLC5A8 in iodide efflux have been unsuccessful (personal communication).
At this stage, one has to conclude that the identity of the protein(s) insuring the Step 2 of the iodide transport in the thyroid is not known.

Interestingly, SLC5A8 appears to be endowed with a tumor suppressor function, which was convincingly demonstrated in colon cell lines. SLC5A8 expression is largely reduced (by more than one order of magnitude) in colon cancers (28), gastric cancers (34), gliomas (35) but also in thyroid cancers (36,37). The silencing of SLC5A8 gene in thyroid cancer appears restricted to the group of papillary thyroid carcinomas of classical type (37). In thyroid tumors as in other cancers, the silencing of the tumor suppressor gene SLC5A8 results from an epigenetic event, the methylation of a promoter region (28,37).

4- Iodide transport – Step 3 and Step 4 – A statement of ignorance
These two steps should be operative to allow iodide to go out of follicles. There is often a confusion in the literature concerning iodide efflux from the thyroid. In the intact follicle, iodide efflux means Step 3 and Step 4. Using isolated thyroid cells, Step 4 occurs together with Step 2. Only thyrocytes cultured as polarized monolayer cells in bicameral devices allow to discriminate Step 3 (uptake from the upper or apical chamber) from Step 4 (release of iodide in the lower or basolateral chamber). Studies on this priviledged model have generated important information on Step 1 and Step 2 (as already mentioned) but only few data on Step 3 and Step 4. In fact, measurements of apical iodide uptake by adding radioactive iodide in the upper chamber were made but uptake values were very low as compared to the uptake values obtained by adding radioiodide in the lower chamber (38); similarly, measurements of basolateral efflux after cell loading with radioiodide yielded very low values as compared to values of apical efflux (7,8). These observations indicate either that the reaction rates of Step 3 and Step 4 are low as compared respectively to those of Step 1 and Step 2 or that Step 3 and Step 4 mainly occur in certain circumstances which were not fulfilled in the experimental system.
As mentioned in the introductory part, the flux of iodide from the lumen to the cell cytoplasm (Step3) should be energy-dependent because of an unfavourable electrochemical gradient whereas Step 4 is probably a passive process. There is no NIS on the apical plasma membrane but a protein functionning as NIS could be operative; indeed, the Na+ concentration in the follicle lumen is close to that of the extrafollicular milieu. One way to learn about a biological process is to know whether it is regulated or not. Information about TSH action on Step 3 is scarce but, we can bring back to light rather old studies reporting the in vivo action of TSH on iodide fluxes (39) analyzed by measurements of the [T]/[S] ratio (or thyroid to serum iodide concentration ratio). TSH injection to rats led to the activation of iodide efflux before the activation of iodide influx (now known to result from TSH-induced activation of NIS gene transcription). Thus, thyroid iodide efflux, likely Step 3, is positively controlled by TSH.
Proteins involved in both Step 3 and Step 4 and mediating thyroid iodide efflux are completely unknown.

5- Concluding remarks
The bidirectional trancellular transport of iodide inside thyroid follicles brings into play several distinct membrane proteins endowed with a function of ion transporter, either active or passive or with a function of ion channel; at present, only one of these molecular species is identified, the Na+/ Iodide Symporter.