Anoctamin-1/TMEM16A is the major apical iodide channel of the thyrocyte.

Iodide is captured by thyrocytes through the Na(+)/I(-) symporter (NIS) before being released into the follicular lumen, where it is oxidized and incorporated into thyroglobulin for the production of thyroid hormones. Several reports point to pendrin as a candidate protein for iodide export from thyroid cells into the follicular lumen. Here, we show that a recently discovered Ca(2+)-activated anion channel, TMEM16A or anoctamin-1 (ANO1), also exports iodide from rat thyroid cell lines and from HEK 293T cells expressing human NIS and ANO1. The Ano1 mRNA is expressed in PCCl3 and FRTL-5 rat thyroid cell lines, and this expression is stimulated by thyrotropin (TSH) in rat in vivo, leading to the accumulation of the ANO1 protein at the apical membrane of thyroid follicles. Moreover, ANO1 properties, i.e., activation by intracellular calcium (i.e., by ionomycin or by ATP), low but positive affinity for pertechnetate, and nonrequirement for chloride, better fit with the iodide release characteristics of PCCl3 and FRTL-5 rat thyroid cell lines than the dissimilar properties of pendrin. Most importantly, iodide release by PCCl3 and FRTL-5 cells is efficiently blocked by T16Ainh-A01, an ANO1-specific inhibitor, and upon ANO1 knockdown by RNA interference. Finally, we show that the T16Ainh-A01 inhibitor efficiently blocks ATP-induced iodide efflux from in vitro-cultured human thyrocytes. In conclusion, our data strongly suggest that ANO1 is responsible for most of the iodide efflux across the apical membrane of thyroid cells.

Diphenyleneiodonium, an inhibitor of NOXes and DUOXes, is also an iodide-specific transporter.

NADPH oxidases (NOXes) and dual oxidases (DUOXes) generate O2 (.-) and H2O2. Diphenyleneiodonium (DPI) inhibits the activity of these enzymes and is often used as a specific inhibitor. It is shown here that DPI, at concentrations similar to those which inhibit the generation of O2 derivatives, activated the efflux of radioiodide but not of its analog (99m)TcO4 (-) nor of the K(+) cation mimic (86)Rb(+) in thyroid cells, in the PCCl3 rat thyroid cell line and in COS cell lines expressing the iodide transporter NIS. Effects obtained with DPI, especially in thyroid cells, should therefore be interpreted with caution.

Does NAD(P)H oxidase-derived H2O2 participate in hypotonicity-induced insulin release by activating VRAC in ß-cells?

NAD(P)H oxidase (NOX)-derived H(2)O(2) was recently proposed to act, in several cells, as the signal mediating the activation of volume-regulated anion channels (VRAC) under a variety of physiological conditions. The present study aims at investigating whether a similar situation prevails in insulin-secreting BRIN-BD11 and rat ß-cells. Exogenous H(2)O(2) (100 to 200 µM) at basal glucose concentration (1.1 to 2.8 mM) stimulated insulin secretion. The inhibitor of VRAC, 5-nitro-2-(3-phenylpropylamino)-benzoate (NPPB) inhibited the secretory response to exogenous H(2)O(2). In patch clamp experiments, exogenous H(2)O(2) was observed to stimulate NPPB-sensitive anion channel activity, which induced cell membrane depolarization. Exposure of the BRIN-BD11 cells to a hypotonic medium caused a detectable increase in intracellular level of reactive oxygen species (ROS) that was abolished by diphenyleneiodonium chloride (DPI), a universal NOX inhibitor. NOX inhibitors such as DPI and plumbagin nearly totally inhibited insulin release provoked by exposure of the BRIN-BD11 cells to a hypotonic medium. Preincubation with two other drugs also abolished hypotonicity-induced insulin release and reduced basal insulin output: 1) N-acetyl-L-cysteine (NAC), a glutathione precursor that serves as general antioxidant and 2) betulinic acid a compound that almost totally abolished NOX4 expression. As NPPB, each of these inhibitors (DPI, plumbagin, preincubation with NAC or betulinic acid) strongly reduced the volume regulatory decrease observed following a hypotonic shock, providing an independent proof that VRAC activation is mediated by H(2)O(2). Taken together, these data suggest that NOX-derived H(2)O(2) plays a key role in the insulin secretory response of BRIN-BD11 and native ß-cells to extracellular hypotonicity.

Glycerol metabolism alteration in adipocytes from n3-PUFA-depleted rats, an animal model for metabolic syndrome.

Aquaglyceroporin 7 (AQP7) is a glycerol transporter expressed in adipocytes. Its expression has been shown to be modulated in obesity. Metabolic syndrome is characterized by abdominal obesity, insulin resistance, dyslipidemia, and hypertension. An animal model displaying several features of metabolic syndrome was used to study the AQP7 expression at both mRNA and protein level and glycerol flux in adipocytes. Second generation n3-PUFA depleted female rats is a good animal model for metabolic syndrome as it displays characteristic features such as liver steatosis, visceral obesity, and insulin resistance. Our data show a reduced expression of AQP7 at the protein level in adipose tissue from n3-PUFA-depleted rats, without any changes at the mRNA levels. [U-(14)C]-Glycerol uptake was not modified in adipocytes from n3-PUFA-depleted animals.

Functional role of aquaglyceroporin 7 expression in the pancreatic beta-cell line BRIN-BD11.

Both mouse and rat pancreatic islet beta-cells were recently found to express aquaglyceroporin 7 (AQP7). In the present study, the expression and role of AQP7 in the function of BRIN-BD11 cells were investigated. AQP7 mRNA and protein were detected by RT-PCR and Western blot analysis, respectively. In an isoosmolar medium, the net uptake of [2-(3)H]glycerol displayed an exponential time course reaching an equilibrium plateau value close to its extracellular concentration. Within 2 min of incubation in a hypotonic medium (caused by a 50 mM decrease in NaCl concentration), the [2-(3)H]glycerol uptake averaged 143.2 +/- 3.8% (n = 24; P < 0.001) of its control value in isotonic medium, declining thereafter consistently with previously demonstrated volume regulatory decrease. When isoosmolarity was restored by the addition of 100 mM urea to the hypotonic medium, [2-(3)H]glycerol uptake remained higher (112.1 +/- 2.8%, n = 24; P < 0.001) than its matched control under isotonic conditions, indicating rapid entry of urea and water. Insulin release by BRIN-BD11 cells was 3 times higher in hypotonic than in isotonic medium. When glycerol (100 mM) or urea (100 mM) were incorporated in the hypotonic medium, the insulin release remained significantly higher than that found in the control isotonic medium, averaging respectively 120.2 +/- 4.2 and 107.0 +/- 3.8% of the paired value recorded in the hypotonic medium. These findings document the rapid entry of glycerol and urea in BRIN-BD11 cells, likely mediated by AQP7.

K+ secretion activated by luminal P2Y2 and P2Y4 receptors in mouse colon.

Extracellular nucleotides are important regulators of epithelial ion transport, frequently exerting their action from the luminal side. Luminal P2Y receptors have previously been identified in rat distal colonic mucosa. Their activation by UTP and ATP stimulates K+ secretion. The aim of this study was to clarify which of the P2Y receptor subtypes are responsible for the stimulated K+ secretion. To this end P2Y2 and P2Y4 knock-out mice were used to measure distal colonic ion transport in an Ussing chamber. In mouse (NMRI) distal colonic mucosa, luminal UTP and ATP with similar potency induced a rapid and transient increase of the transepithelial voltage (V(te)) (UTP: from -0.81 +/- 0.23 to 3.11 +/- 0.61 mV, n = 24), an increase of equivalent short circuit current (I(sc)) by 166.9 +/- 22.8 microA cm(-2) and a decrease of transepithelial resistance (R(te)) from 29.4 +/- 2.4 to 23.5 +/- 2.0 Omega cm2. This effect was completely inhibited by luminal Ba2+ (5 mm, n = 5) and iberiotoxin (240 nm, n = 6), indicating UTP/ATP-stimulated K+ secretion. RT-PCR analysis of isolated colonic crypts revealed P2Y2, P2Y4 and P2Y6 specific transcripts. The luminal UTP-stimulated K+ secretion was still present in P2Y2 receptor knock-out mice, but significantly reduced (DeltaV(te): 0.83 +/- 0.26 mV) compared to wild-type littermates (DeltaV(te): 2.08 +/- 0.52 mV, n = 9). In P2Y4 receptor knock-out mice the UTP-induced K+ secretion was similarly reduced. Luminal UTP-stimulated K+ secretion was completely absent in P2Y2/P2Y4 double receptor KO mice. Basolateral UTP showed no effect. In summary, these results indicate that both the P2Y2 and P2Y4 receptors are present in the luminal membrane of mouse distal colonic mucosa, and stimulation of these receptors leads to K+ secretion.

[Insulin and arterial hypertension: the role of the kidney]. / Insuline et hypertension artérielle: le rôle du rein.

We have demonstrated that insulin stimulates sodium reabsorption in the distal nephron by stimulating the phosphatidylinositol 3-kinase (PI 3-kinase) pathway and that any stimulation of this enzyme (e.g. by EGF, by H2O2 or by exogenous PIP3, added apically) leads to a parallel increase in sodium reabsorption. We therefore suggest that hyperinsulinemia leads to hypertension through increased renal sodium reabsorption in the distal nephron.

Hypotonic cell swelling stimulates permeability to cAMP in a rat colonic cell line.

This study characterized the membrane permeability to cAMP in a cell line derived from the rat colon (CC531(mdr+)) by comparison of fluxes of 3H-cAMP, 3H-8-bromo-cAMP, 3H-taurine, 3H-adenosine and 3H-5'AMP under various experimental conditions including cell membrane depolarization and hypotonic cell swelling. Cell volume was modified by changing the osmolality and composition of the extracellular medium. Incubation in iso- and hypotonic KCl media induced graded increases in cell volume and stable activation of volume-sensitive channels that was reflected in an increased efflux of 3H-taurine. Incubation in hypotonic KCl solution also enhanced the efflux of 3H-8-Br-cAMP (a non-hydrolysable analogue of cAMP). Both the efflux of 3H-taurine and of 3H-8-Br-cAMP were inhibited by 5-nitro-2-(3-phenylpropylamino)benzoate (NPPB, 100 microM) suggesting the involvement of volume-sensitive anion channels. To gain further insight into the route mediating cAMP permeability, the uptakes of 3H-cAMP, 3H-8-Br-cAMP and 3H-taurine were determined over short (5-min) periods. Uptakes of these substrates demonstrated close similarities: comparable increases were observed that correlated with the increases in cell volume in iso- and hypoosmotic KCl media; they were inhibited strongly by NPPB (100 microM) and metabolic inhibitors (deoxyglucose, 20 mM together with the mitochondrial uncoupler carbonylcyanide p-(trifluoromethoxy)phenylhydrazone, FCCP, 10 microM) while barely reduced by dipyridamole (100 microM) and they were not affected by adenosine (1 mM). In contrast, the uptakes of 3H-adenosine and 3H-5'AMP had strikingly different properties; they were insensitive to cell swelling; barely inhibited by NPPB (100 microM) and metabolic inhibitors (deoxyglucose and FCCP) while strongly reduced by dipyridamole (100 micro M). Unlike the uptakes of 3H-cAMP, 3H-8-Br-cAMP and 3H-taurine, the uptakes of 3H-adenosine and 3H-5'AMP were reduced in Na(+)-free media, suggesting the presence in this cell line of two different adenosine carriers, one sodium-dependent and one sodium-independent. Taken together the present data show that in this rat colonic cell line, cAMP permeability is increased by cell swelling in hypotonic KCl medium and inhibited by NPPB and metabolic inhibitors. The similarity of these characteristics to those of taurine permeability suggests the involvement of a volume-sensitive anion pathway.

Anion selectivity by the sodium iodide symporter.

The iodide transporter of the thyroid (NIS) has been cloned by the group of Carrasco. The NIS-mediated transport was studied by electrophysiological methods in NIS-expressing Xenopus oocytes. Using this method, the anion selectivity of NIS was different from that previously reported for thyroid cells, whereas perchlorate and perrhenate were found not transported. In this study we compared the properties of human NIS, stably transfected in COS-7 cells to those of the transport in a thyroid cell line, the FRTL5 cells, by measuring the transport directly. We measured the uptake of (125)I(-), (186)ReO(4)(-), and (99m)TcO(4)(-) and studied the effect on it of known competing anions, i.e. ClO(4)(-), SCN(-), ClO(3)(-), ReO(4)(-), and Br(-). We conclude that the properties of the NIS transporter account by themselves for the properties of the thyroid iodide transporter as described previously in thyroid slices. The order of affinity was: ClO(4)(-) > ReO(4)(-) > I(-) >/= SCN(-) > ClO(3)(-) > Br(-). NIS is also inhibited by dysidenin (as in dog thyroid).

Aristolochic acid impedes endocytosis and induces DNA adducts in proximal tubule cells.

BACKGROUND: Aristolochic acid (AA), present in Aristolochia plants, appears to be the toxin responsible for Chinese herbs nephropathy (CHN), a rapidly progressive tubulointerstitial nephritis. One of the earliest sign of CHN is the urinary excretion of low-molecular-weight proteins (LMWP), suggesting that AA is toxic to proximal tubules (PT). METHODS: The effects of AA on PT functions including reabsorption of LMWP were investigated on the well-established opossum kidney (OK) cell line, a model for PT, and compared with those of the classical PT toxin cadmium chloride (CdCl2). RESULTS: OK cell monolayers internalized albumin and beta2-microglobulin by receptor-mediated endocytosis, both proteins apparently competing for the same receptor, a complex of megalin and cubulin. The process was significantly impaired by 24-hour preincubation with AA (10 or 20 micromol/L) or CdCl2 (15 micromol/L). Furthermore, 24-hour exposure to AA followed by its removal during one to six days led to a persistent inhibition of the uptake of albumin, in contrast to the substantial recovery observed after CdCl2 removal. Neither AA nor CdCl2 affected cell viability, Na+-glucose cotransport or total rate of protein synthesis. AA significantly decreased megalin expression and formed specific DNA adducts in OK cells, similar to those found in kidneys from CHN patients. CONCLUSIONS: The present data support the involvement of AA in the early PT dysfunction found in CHN; furthermore, they suggest a causal relationship between DNA adduct formation, decreased megalin expression, and inhibition of receptor-mediated endocytosis of LMWP.

Inhibition of basolateral cAMP permeability in the toad urinary bladder.

1. The effect of sulphonylurea drugs on hydrosmotic flow across toad urinary bladder epithelium was re-evaluated in the present study. Glibenclamide, added to the basolateral medium, significantly enhanced the osmotic flow induced by low doses of antidiuretic hormone (ADH) or forskolin (FK), while it inhibited the effect of exogenous cyclic adenosine monophosphate (cAMP) or its non-hydrolysable bromo derivative, 8-Br-cAMP, added to the basolateral medium. These opposite effects of glibenclamide on the transepithelial osmotic flow can be explained by a reduction of cAMP permeability across the basolateral membrane of the epithelium. The decrease in cAMP permeability leads, according to the direction of the cAMP gradient, to firstly an enhanced osmotic flow when cAMP is generated intracellularly by addition of ADH and FK, glibenclamide reducing cAMP exit from the cell, and secondly a decreased osmotic flow in response to cAMP (and 8-Br-cAMP) added to the basolateral medium, glibenclamide inhibiting, in this case, their entry into the cell. 2. The demonstration that glibenclamide actually inhibits the basolateral cAMP permeability rests on the fact that firstly it decreases the release of cAMP into the basolateral medium by about 40 %, at each concentration of ADH or forskolin tested, secondly it increases the cAMP content of paired hemibladders incubated in the presence of ADH or FK, when intracellular degradation was prevented by phosphodiesterase inhibition, and thirdly it decreases also the uptake of basolateral 8-Br-[3H]cAMP into paired toad hemibladders. 3. Taken together, the present data demonstrate that glibenclamide inhibits the toad urinary bladder basolateral membrane permeability to cAMP, most probably by a direct interaction with a membrane protein not yet indentified but distinct from the sulphonylurea receptor.

Expression and intracellular processing of chimeric and mutant CFTR molecules.

The cystic fibrosis transmembrane conductance regulator (CFTR) is a cyclic AMP-activated chloride channel comprising two membrane-spanning domains (MSDs), two nucleotide-binding domains (NBDs) and a unique regulatory (R) domain. The most frequent cystic fibrosis (CF) mutation, a deletion of Phe508 in NBD1, results in the retention of the DeltaF508 CFTR in the endoplasmic reticulum, as do many other natural or constructed mutations located within the first NBD. In order to further define the role of NBD1 in CFTR folding and to determine whether the higher frequency of mutations in NBD1 with respect to NBD2 results from its position in the molecule or is related to its primary sequence, we constructed and expressed chimeric CFTRs wherein NBD domains were either exchanged or deleted. Synthesis, maturation and activity of the chimeras were assessed by Western blotting and iodide efflux assay after transient or stable expression in COS-1 or CHO cells respectively. The data showed that deletion of NBD1 prevented transport of CFTR to the cytoplasmic membrane whereas deletion of NBD2 did not impair this process but resulted in an inactive chloride channel. On the other hand, substituting or inverting NBDs in the CFTR molecule impaired its processing. In addition, while the NBD1 R555K mutation is known to partially correct the processing of CFTR DeltaF508 and to increase activity of both wild-type and DeltaF508 individual channels, it showed no positive effect when introduced into the double NBD1 chimera. Taken together, these observations suggest that the proper folding process of CFTR results from complex interactions between NBDs and their surrounding domains (MSDs and/or R domain).

P-glycoprotein inhibition by glibenclamide and related compounds.

Glibenclamide is well known to interact with the sulphonylurea receptor (SUR) and has been shown more recently to inhibit the cystic fibrosis transmembrane conductance regulator protein (CFTR), both proteins that are members of the ABC [adenosine 5'-triphosphate (ATP)-binding cassette] transporters. The effect of glibenclamide and two synthetic sulphonylcyanoguanidine derivatives (dubbed BM-208 and BM-223) was examined on P-glycoprotein, the major ABC transporter responsible for multidrug resistance (MDR) in cancer cells. To this end, we employed different cell lines that do or do not express P-glycoprotein, as confirmed by Western blotting: first, a tumour cell line (VBL600) selected from a human T-cell line (CEM) derived from an acute leukaemia; second, an epithelial cell line derived from a rat colonic adenocarcinoma (CC531(mdr+)) and finally, a non tumour epithelial cell line derived from the proximal tubule of the opossum kidney (OK). Glibenclamide and the two related derivatives inhibited P-glycoprotein because firstly, they acutely increased [3H]colchicine accumulation in P-glycoprotein-expressing cell lines only; secondly BM-223 reversed the MDR phenomenon, quite similarly to verapamil, by enhancing the cytotoxicity of colchicine, taxol and vinblastine and thirdly, BM-208 and BM-223 blocked the photoaffinity-labelling of P-glycoprotein by [3H]azidopine. Furthermore, glibenclamide is itself a substrate for P-glycoprotein, since the cellular accumulation of [3H]glibenclamide was low and substantially increased by addition of P-glycoprotein substrates (e. g., vinblastine and cyclosporine) only in the P-glycoprotein-expressing cell lines. We conclude that glibenclamide and two sulphonylcyanoguanidine derivatives inhibit P-glycoprotein and that sulphonylurea drugs would appear to be general inhibitors of ABC transporters, suggesting an interaction with some conserved motif.

Intra-renal and subcellular distribution of the human chloride channel, CLC-5, reveals a pathophysiological basis for Dent's disease.

Dent's disease, which is a renal tubular disorder characterized by low molecular weight proteinuria, hypercalciuria and nephrolithiasis, is associated with inactivating mutations of the X-linked chloride channel, CLC-5. However, the manner in which a functional loss of CLC-5 leads to such diverse renal abnormalities remains to be defined. In order to elucidate this, we performed studies to determine the segmental expression of CLC-5 in the human kidney and to define its intracellular distribution. We raised and characterized antisera against human CLC-5, and identified by immunoblotting an 83 kDa band corresponding to CLC-5 in human kidney cortex and medulla. Immunohistochemistry revealed CLC-5 expression in the epithelial cells lining the proximal tubules and the thick ascending limbs of Henle's loop, and in intercalated cells of the collecting ducts. Studies of subcellular human kidney fractions established that CLC-5 distribution was associated best with that of Rab4, which is a marker of recycling early endosomes. In addition, confocal microscopy studies using the proximal tubular cell model of opossum kidney cells, which endogenously expressed CLC-5, revealed that CLC-5 co-localized with the albumin-containing endocytic vesicles that form part of the receptor-mediated endocytic pathway. Thus, CLC-5 is expressed at multiple sites in the human nephron and is likely to have a role in the receptor-mediated endocytic pathway. Furthermore, the functional loss of CLC-5 in the proximal tubules and the thick ascending limbs provides an explanation for the occurrences of low molecular weight proteinuria and hypercalciuria, respectively. These results help to elucidate further the patho-physiological basis of the renal tubular defects of Dent's disease.

DIDS (4,4'-diisothiocyanatostilbene-2,2'-disulfonic acid) increases iodide trapping, inhibits thyroperoxidase and antagonizes the TSH-induced apical iodide efflux in porcine thyroid cells.

4,4'-Di-isothiocyanatostilbene-2,2'-disulfonic acid (DIDS), an inhibitor of several anionic channels and transporters including the band 3 protein of the red blood cell membrane was tested on iodide metabolism in cultured porcine thyroid cells. We used three experimental cell culture models: (i) forskolin-stimulated correctly inside-in polarized follicle-associated thyroid cells cultured onto plastic support (ii) suspensions of isolated cells derived from such cultures (iii) polarized monolayers in bicameral chambers. DIDS was observed to increase free-iodide trapping in all conditions. Organification of iodide by follicle-associated cell cultures incubated for 6 h decreased as a function of DIDS concentration with an IC50 of 5 x 10(-5) M. This block in organification is accounted for a block in thyroperoxidase activity as in vitro both purified lactoperoxidase and purified porcine thyroperoxidase were inhibited by DIDS with a similar dose-dependency the IC50 being also of 5 x 10(-5) M. Both control and DIDS-treated cells in suspension, actively trapped iodide and reached a steady concentration in about 50 min; however the plateau was 4.4-fold higher in (10(-3) M) DIDS-treated cells. Acute TSH-stimulation at this plateau of 125I-preloaded cells in suspension in the presence of 2 mM methimazole (MMI) induced a fast release of iodide from these cells as expected (first step of the TSH-biphasic effect). This TSH-induced iodide efflux was however completely inhibited by DIDS (10(-3) M). Furthermore, addition of DIDS to the apical compartment of TSH-prestimulated cell monolayers in bicameral chambers resulted in an increase in intracellular-iodide concentration and in an inhibition of iodide efflux into the apical medium. Taken together, the present results demonstrate that DIDS mainly interacts with two main components of the thyroid apical cell membrane: thyroperoxidase and a cAMP-sensitive iodide channel.

The Na+-I- cotransporter of the thyroid: characterisation of new inhibitors.

New inhibitors of the Na+-I- cotransporter of the thyroid in bovine thyroid slices and in bovine plasma membrane vesicles have been investigated. They include: (1) econazole; (2) 5-(N,N-hexamethylene)amiloride (HMA); and (3) dysidenin. In both systems, the kinetics of iodide transport yielded apparent Km values of 39 and 14 microM respectively. The possible interaction of each of these inhibitors with the iodide site of the Na+-I- cotransporter was tested by performing detailed transport kinetics analysis at varying iodide concentrations and at 150 mM NaCl. Econazole induced a non-competitive inhibition while dysidenin and HMA gave a mixed type of inhibition. The Ki values for dysidenin and econazole, computed from Dixon plots, were 5 and 2 microM respectively while the Ki value for HMA could not be determined. Each inhibition was reversible, indicating the absence of covalent binding of the inhibitor to the Na+-I- cotransporter.

Expression of CFTR in human and bovine thyroid epithelium.

The expression of cystic fibrosis transmembrane conductance regulator (CFTR) in the thyroid has not been documented to date, although a role for CFTR in the thyroid follicular epithelium is suggested both clinically, by the occurrence of subclinical hypothyroidism in patients with cystic fibrosis (CF), and physiologically, by the presence of low-conductance, adenosine 3',5'-cyclic monophosphate-activated Cl channels in the follicular cells. Using reverse transcriptase-polymerase chain reaction with nested primers derived from exons 13 and 14 of the human CF gene, we have now documented the presence of CFTR mRNA in the human thyroid. Western blot analyses using six antibodies directed against different domains of human CFTR showed that a 165-kDa band was present in membrane extracts from bovine and human thyroid. This protein has the predicted size of mature CFTR and was not detected with preimmune serum or preadsorbed antiserum. By immunofluorescence and immunoperoxidase, CFTR was located in the follicular cells, with a diffuse, intracellular labeling pattern. Quantitative analysis revealed that 64% of the follicles were CFTR positive, but only 16% of the follicular cells were stained per follicle. The number of CFTR-positive cells was inversely proportional to the size of the follicle. These results 1) demonstrate the expression of CFTR at the mRNA and protein levels in human and bovine thyroid follicular cells and 2) suggest that CFTR expression could be instrumental in follicular enlargement.