EXPERIMENTAL RESEARCH 194,202-209 (1991)

Transcytosis in Follicle Cells: Regulation and Implications for Thyroglobulin Transport

P. ROMAGNOLI’ AND V. HERZOG

With the technical assistance of S. Fuchs htitut fiir ZeUbiologie, University of Bonn, Federal Republic of Germany

common gating step for both pathways. o 1981 Academic In order to analyze quantitatively the translocation Press, Inc. of plasma during and transcyto- sis and the regulation of these processes in thyroid folli- cle cells, the apical cell surfaces of resting and TSH- INTRODUCTION stimulated inside-out follicles were labeled with cation- ized ferritin. Morphometric analyses showed that the Transcytosis, i.e., the vesicular transfer of molecules rates of endocytosis and transcytosis are TSH-depen- through the cytoplasm of polarized cells from one dent. More interestingly, whereas the effect of TSH on plasma membrane domain to another, is a common endocytosis was transient (with a maximum at 16 min), transport activity in eukaryotic cells which has been an- the effect on transcytosis continued to increase until alysed in endothelial [l-4] and several epithelial cell the end of the experiment (i.e., 70 min). During 1 h of types, such as the epithelial cells of the small intestine endocytosis, the fraction of membrane involved in [E&3], epididymis [9], choroid plexus [lo, 111, mammary transcytosis increased by a factor 4 upon TSH stimula- glands [6, 121, yolk sac entoderm [13], trophoblast [14, tion, corresponding to about 12% of the internalized 151, hepatocytes [ 16,171, MDCK cells [ 181, and capsule apical plasma membrane area. Cooling to 15“C slowed cells of vertebrate muscle spindles [19]. down, but did not block endocytosis entirely, whereas Transcytosis in thyroid follicle cells is of particular transcytosis and transfer to lysosomes were totally in- hibited. In order to quantitate transcytosis of thyroglob- interest, because it has been recognized as the cellular ulin (TG) and to ascertain whether this molecule under- prerequisite for the appearance of thyroglobulin (TG) goes cleavage during transcytosis, inside-out follicles in the circulation [20-221. Since thyroid function, in gen- were incubated in a medium containing 3H-labeled TG eral, and the appearance of TG outside the thyroid in the presence of TSH; upon washing and reopening of gland are regulated by TSH [22-241, the question arose follicles, the luminal fluid containing TG after transcy- as to whether transcytosis in follicle cells is also a TSH- tosis was found to contain about 10% of the total radio- regulated process. Because it has been shown before activity taken up by follicle cells. Transcytosed TG that cationized ferritin follows the same pathway in thy- proved to be unmodified with respect to its electropho- roid follicle cells as TG [21], we used this tracer to deter- retie mobility. We conclude that (i) the fraction of mine the amount of membrane transcytosed by thyroid transcytosed TG corresponds approximately to the follicle cells and to analyze the regulation of this trans- fraction of membrane involved in this process, (ii) TG port by TSH. For this purpose, the apical surface of does not undergo cleavage during transcytosis, (iii) en- inside-out follicles was labeled with cationized ferritin docytosis and transcytosis are regulated by TSH but (CF) and the kinetics of endocytosis and transcytosis differ in their kinetics after stimulation, and (iv) tran- were calculated from the amount of tracer translocated scytosis is affected by temperature in a similar way as by resting and by TSH-stimulated cells. transfer to lysosomes, suggesting the existence of a The observations show that endocytosis is transiently stimulated by TSH whereas transcytosis continues to be enhanced for prolonged periods of time. The experi- 1 To whom correspondence and reprint requests should be ad- ments also indicate that TG does not undergo cleavage dressed at: Dipartimento di Anatomia Umana e Istologia, Sezione “E. Allara,” Viale Pieraccini 6, I-50139 Firenze (Italia). Fax during transcytosis. Our findings support the hypothe- (055)4379384. sis [21] that transcytosis in thyroid follicle cells is selec-

0014-4827/91$3.00 202 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved. TRANSCYTOSIS THROUGH THYROID FOLLICLE CELLS 203 tively regulated by TSH and that lysosomes are by- photographed. Twelve to 15 cell sections were photographed for each passed during this process. time point and experimental condition. The magnification of photo- graphs was controlled by the use of a calibrating grid with a line MATERIALS AND METHODS distance of 463.5 nm (Balzers, Liechtenstein). Besides the apical and basolateral plasma membrane, the following Tissue. Thyroid glands from 6- to 8-month-old (80-90 kg) pigs compartments were identified in the cytoplasm: vesicles, , multivesicular bodies, secondary lysosomes, Golgi cisternae. Vesicles were removed 5-10 min after anesthesia by electroshock and killing by bleeding. The glands were collected on ice and carried within 20 were round or oval, with a homogeneous, intermediately electron- min to the laboratory for isolation of follicle segments. dense content; their transverse diameter was never found to exceed 110 nm; this compartment included primary endocytic vesicles and Materials. India ink Russ-dispersion C11/1431a was from transcytotic vesicles. Primary endocytic vesicles were located less Gunther Wagner Pelikan-Werke (Hannover, FRG); India ink was than 1 pm distant from the apical plasma membrane, whereas trans- dialyzed overnight against PBS before use. Sodium dodecyl sulphate cytotic vesicles were closely associated with the basolateral cell sur- (SDS), EDTA, Coomassie brilliant blue R, and Na-p-phenylmeth- face. Endosomes were defined as vesicular compartments of variable anesulfonyl fluoride (PMSF) were from Sigma Chemie GmbH (Dei- size with an electron lucent matrix; CF particles were attached to the senhofen, FRG) and aprotinin was from Bayer (Leverkusen, FRG). inner surface of the limiting membrane. Multivesicular bodies usually TSH was obtained from Ferring Arzneimittel GmbH (Kiel, FRG) and carried CF attached to the inner surface of the membrane and, in cationized ferritin (p.1. 8.5) from Miles Laboratories, Inc. (Research part, also to the of the vesicles. Secondary lysosomes, Products Div., Elkhart, IN). Eagle’s minimum essential medium which were recognized by their electron-dense, sometimes inhomo- (MEM) with Earle’s salts, amphotericin B (250 pg/ml), penicillin geneous content and which have been identified before by the pres- (50,000 U/ml), streptomycin (50 mg/ml), soybean trypsin inhibitor, ence of lysosomal hydrolases [30], carried CF in their matrix, but not and collagenase were from Boehringer GmbH (Mannheim, FRG); attached to their membrane. leucine-free MEM and fetal calf serum from Grand Island Biological Co. (Karlsruhe, FRG); and L-[4,5-sH]leucine (sp act, 60 Ci/mmol), Morphmetry. The number of CF particles (Nc,) overlaying the [3H]toluene (2.4 Ci/g), and X-ray films (Hyperfilm-Betamax) for profiles of basolateral plasma membrane and those within the pro- fluorographic visualization of 3H-labeled TG from Amersham files of all cytoplasmic membrane-bound compartments were Buchler GmbH (Braunschweig, FRG). A&amide was from Serva counted at magnification X260,000 using a binocular Leitz-Wild ster- Feinbiochemica GmbH and Co. (Heidelberg, FRG) and 2,5-diphenyl- eomicroscope M3 (Leitz, Munich, FRG). oxazol and toluene were from Packard Instruments Int. (Zurich, The length of the apical plasma membrane (I,,,) was measured at Switzerland). Nybolt filters were purchased from Schweizerische Sei- magnification X18,000 with a semiautomatic image-analyzing system dengaze-Fabrik (Zurich, Switzerland). MOP/AM 03 and incorporated microprocessor Z 80 (Kontron, Mun- Preparation of inside-out follicles. Inside-out follicles from thyroid ich, FRG). The mean length of this profile per cell section was defined glands were prepared as described previously [20]. The impermeabil- as L& . ity of the intercellular tight junctions of the follicles was tested using The area equivalent diameter [31] of 100 primary endocytic vesicles India ink as a tracer [20]. (O,,) from resting follicles and of 100 ones from stimulated follicles, Application of tracers for studies on endocytosis and transepithelial all incubated with CF for 30 min was measured with a calibrated ruler transport. Inside-out thyroid follicles were washed extensively in at magnification X260,000. The boundary length of each primary vesi- MEM to remove fetal calf serum which could interact with the cle profile (L,) was computed from J& = ~0~. Lm was distributed tracers. The follicles used either remained unstimulated or were stim- approximately log-normally and was well correlated with lnNcm (P ulated (50 mU TSH/ml) 10 min before the addition of the tracer. < 0.001 that r = 0). Therefore, the number of CF particles internal- Cationized ferritin (CF; 10 pg/ml) was added to culture medium ized at each time point was taken as a measure of the amount of containing stimulated or resting follicles. Aliquots of CF-labeled in- membrane internalized during the formation of primary endocytic side-out follicles were fixed for electron microscopy 6, 11, 16,30, and vesicles. Vesicle profiles contained a median number of CF particles 60 min after the addition of CF to the culture medium. In some exper- WCFpv = antiln( C lnNcFpev 1100) and had a mean boundary length Lw iments, CF was added to stimulated follicles at 25”, 15’, 12’, lo’, or = 2 L,,/lOO. Each particle of CF was considered to express the inter- 4°C to assess the effect of temperature on endocytosis and transcyto- nalization of a fraction of apical plasma membrane profile k = L&l sis. In these experiments the follicles were fixed for electron micros- MXPW x L&J. copy 60 min after the addition of CF to the culture medium. The sum of the number of CF particles overlaying the basolateral 3H-labeled TG was prepared biosynthetically from inside-out folli- plasma membrane and those in each membrane-bound compartment cles as described previously [21]. The activity of [3H]TG was 310 pCi/ was transformed into percentage surface of apical plasma membrane mg as measured with a Tri-Carb liquid scintillation spectrometer internalized (S,,) by the equation Sispm = k X 100 X Nip. The num- (Packard Instruments Co., Inc., United Technologies, Downers ber of CF particles on the basolateral plasma membrane profile was Grove, IL) using [3H]toluene as an internal standard. transformed into percentage surface of apical plasma membrane Electron microscopy. Samples of inside-out follicles were fixed in transcytosed (S,,,) by the equation S,,, = k X 100 X &r. Karnovsky’s fixative [25] for 2 h at room temperature, pelleted in a Analysis of endocytosed and transcytosed radiolabeled TG (see Fig. microfuge (Beckman Instruments Inc.), postfixed in 1% unbuffered 1). Freshly prepared inside-out follicles, resting or stimulated with 0~0, for 1 h at 4”C, and embedded in Epon. Thin sections were cut TSH, were resuspended in a medium containing 300 pCilm1 [‘H]TG with diamond knives, stained with lead citrate, and examined in a prepared as described above. Aliquots were removed after 5 h. The Siemens Elmiskop I or 102 (Siemens AG, Munich, FRG) at 80 kV. follicles were washed in MEM and opened by treatment with EDTA Follicles with an average cell height of 10 pm were analyzed for mor- (10 mg/ml) followed by pipeting. After centrifugation (30 s at 7Og), phometry, because they allowed a clear distinction between lateral the supernatant (containing the transeytosed material) and the sedi- and basal plasma membranes. Follicles with attenuated epithelial ment (containing follicle cells) were separated from each other. The cells as described by Wollman and co-workers [26, 271 were also ob- precipitates were lysed with 10 mM Tris-HCl buffer, pH 7.6, contain- served but excluded from morphometric analyses because the lateral ing 1% Triton X-100,2 mM PMSF and 400 U/ml aprotinin. and basal cell surfaces did not appear as structurally distinct do- The transcytosed material and the cell lysates were analyzed by mains. Randomly selected whole vertical cell sections [28, 291 com- SDS-polyacrylamide gel electrophoresis (PAGE) [32] followed by flu- prising the apical and basal plasma membrane and the nucleus were orography [33]. 204 ROMAGNOLI AND HERZOG

shown in Figs. 2-4. Morphological details on endocyto- sis and transcytosis in thyroid follicle cells have been described previously [20, 21). Two morphologically distinct states of thyroid folli- cles could be distinguished: follicles with attenuated epi- thelial cells and, most common, follicles with cuboidal thyrocytes. Both forms may be related to periodically 1 pipeting occurring water accumulation in the central follicle cavi- ties followed by collapse with water release [35]. Transcytosis of CF was observed in both follicular tronscytosed TG states, but follicles with cuboidal epithelial cells were selected for morphometrical analyses. + intracellular TG <

= internalized TG E cl Effect of TSH and Temperature on Endocytosis and Transcytosis FIG. 1. Experimental model to analyze transcytosis of TG. In inside-out follicles the apical cell surface (thick line) is directed to- In 1 h of endocytosis, about 460% of the apical plasma wards the culture medium; the cavity inside the follicle is homologous membrane was internalized in resting conditions (Table to the interstitial space in the intact gland. Inside-out follicles were I). TSH stimulated the rate of endocytosis only at early incubated in the presence of [3H]TG (dotted area) for 5 h (A, B), then washed in MEM to remove extrafollicular [SH]TG (C), and opened by time points (Fig. 5) with a maximum at 16 min (Fig. 6). EDTA and pipeting (D), thereby releasing the transcytosed [SH]TG. This effect was transient and followed by an inhibition After centrifugation (E), the supernatant and the sediment were ana- below resting rates (Fig. 6). The labeling of lysosomes lyzed to obtain qualitative and quantitative data on the transcytosed with CF increased progressively with time until the end and the total internalized TG (for results see Fig. 11 and text). of the experiment, in both resting conditions and upon TSH (Fig. 7). In 1 h, 16% of the apical plasma membrane was Quuntitation of th.e transcytosis of TG. The fraction of transcy- tosed TG (TG,) was determined for each of four different experi- transcytosed in resting conditions. TSH stimulated the ments by two distinct procedures: (i) the radioactivity of the superna- rate of transcytosis by a factor of about 4 (Table I). This tant from reopened inside-out follicles (corresponding to the amount effect increased progressively until the end of the exper- of TG,) and the radioactivity in the cell lysate (corresponding to the iment (Figs. 6 and 8). In both resting conditions and amount of TG internalized but not transcytosed: TG,) were measured upon TSH, primary endocytic vesicles and transcytotic directly with a Tri-Carb liquid scintillation spectrometer (Packard Instruments) as described above; and (ii) after SDS-PAGE, the dif- vesicles were the same size; transcytotic and the vast ferent lanes were suspended in toluene and the radioactivity was majority of primary endocytic vesicles were uncoated measured. The amount of transcytosed TG was expressed as a per- (not shown). centage: TG,% = 100 X TG,/(TG, + TGi). Analysis of the data showed In both resting and stimulated conditions, endosomes that the overall mean of the four experiments did not differ signifi- cantly depending on the method used for quantitation; therefore, the were labeled in all cells at 6 min of endocytosis of CF. mean between the values obtained with the two methods was as- Labeling of the basolateral cell surfaces domains oc- sumed as representative for each experiment. curred in all cells 30 min after beginning of endocytosis. Statistics. Student’s t test for unpaired values, with one tail, chi The amount of CF internalized at 15 or 25°C was sig- square test, and regression analysis were used as appropriate [34]. In all the evaluations, probability levels that the observed differences were due to chance, P < 0.05, P < 0.01, or P < 0.001, were recorded; P TABLE 1 < 0.05 was chosen as significance level. In the results, the median and standard error are given. Translocation of Apical Plasma Membrane (APM) during 60 min of Endocytosis

RESULTS %APM %APM endocytosed transcytosed The walls of inside-out follicles were tight against the f&38 diffusion of particulate tracers such as India ink or Control 457 16 ‘; -74 ferritin (not shown), indicating that the follicle frag- +1otl 540 65 +7 ments had closed again to form tight epithelial walls. On Stimulated -90 this structural basis paracellular diffusion was excluded P(Student’s t test) n.s. to.01 and inside-out follicles could confidently be used in ex- periments on selective endocytosis from the apical Note. The relative amount of apical plasma membrane (* standard error) internalized by resting and by TSH-stimulated cells. Note that plasma membrane. Inside-out follicles in suspension, transcytosis is enhanced by a factor of about 4 upon stimulation with selective labeling of apical plasma membrane, and en- TSH and that about 12% of the apical membrane internalized is docytosis in resting and TSH-stimulated conditions are transferred to the basolateral cell surface. FIG. 2. Inside-out follicles in suspension culture. Light micrograph, X400. FIG. 3. Thyrocyte from an inside-out follicle 60 min after the addition of CF to the incubation medium in resting conditions. The apical cell surface (top) is labeled with CF, which does not permeate intercellular tight junctions. X8000. FIG. 4. Thyroid follicle cell 60 min after the addition of CF to the culture medium in the presence of TSH. Besides the apical cell surface (top), vesicles (VE), endosomes (ES), lysosomes (LY), and the basolateral cell surface (arrows) are heavily labeled (compare with Fig. 3). x20,000. 205 26.

l%ne (mln) 8 time (min)

i

temperatwe ( OC) 6 9

37-c

ve ea ly blm ve ee ly blm ve 00 ly blm

7 time (min) 10 FIG. 5. Effect of TSH on endocytosis. TSH results in an increased uptake of apical plasma membrane (PM), with a maximum in the first 6 min after the addition of CF. At later time points, the TSH effect decreases progressively until the rate of uptake becomes lower than that in resting cells and both curves converge about 40 min after the addition of CF, i.e., 50 min after stimulation with TSH. See also Fig. 6. Median values and standard errors are indicated. Best fits for y = 2.62 t1~2620 (resting cells) and for y = 12.12 toa3** (TSH-stimulated cells) are shown; t indicates the time (in min) from the addition of CF. FIG. 6. Time course of the TSH-dependent increase of endocytosis- and transcytosis-rates. The effect of TSH on endocytosis is ex- hausted in approximately 50 min. Thereafter, the endocytosis rate of stimulated follicles decreases below that of unstimulated follicles. The effect of TSH on transcytosis, on the contrary, increases progressively until the end of experiment. The curves were derived from the best fits indicated in Figs. 5 and 8, in the time interval between 16 and 70 min from the addition of TSH to the culture medium (continuous line). Hypothetical extrapolations between 0 and 16 min after addition of TSH are also indicated (discontinuous line). FIG. ‘7. Effect of TSH on the labeling of lysosomes with CF. TSH results in increased labeling of lysosomes until the end of the experiment (60 min after the addition of CF) without saturation of this compartment. Median values and standard errors are indicated, standard errors are shown only when at least 11 cells could be used for computations. Best fits for y = 0.202 t’.- (resting cells) and y = 0.065 t2.“2s (TSH-stimulated) are also shown; t indicates the time (in min) from the addition of CF. FIG. 8. Effect of TSH on transcytosis. TSH results in an increased uptake of the apical plasma membrane (PM) until the end of the experiment. The TSH-induced rate of transcytosis increases progressively until the end of the experiment (compare with Fig. 5). Median values and standard errors are indicated; standard errors are shown only when at least 11 cells could be used for computations. Best fits for y = 0.066 t’.29*6 (resting cells) and y = 0.007 p no’ (TSH-stimulated) are also shown; t indicates the time (in min) from the addition of CF. FIG. 9. Effect of temperature on endocytosis. Lowering of temperature results in a decrease of CF uptake at the apical plasma membrane (P < 0.01 at 25’C; P < 0.001 at 15°C; Student’s t test). Complete cessation of CF uptake occurs at 12°C and below. Median values and standard errors are given. Best fit for y = -1260 + 902’ (derived from In(y) = 4.5 + ln(T - 14)) is also shown. T indicates the temperature in “C. FIG. 10. Effect of temperature on the intracellular distribution and transcytosis of CF. The cells were TSH-stimulated and incubated for 60 min with CF. Note the block of transport to lysosomes (ly) and to basolateral cell surfaces (blm) at 15°C. es: Endosomes, including multivesicular bodies; ve: Vesicles. The total number of CF particles counted for this figure was 2020 particles at 15”C, 14,635 at 25”C, and 28,962 at 37°C. The differences in the distribution of tracer depending on the temperature were significant (P < 0.001; chi square test).

206 TRANSCYTOSIS THROUGH THYROID FOLLICLE CELLS 207 nificantly less than at 37°C (Fig. 9). Endocytosis was 6). This might be the expression of a poststimulatory not detected at temperatures below 15’C. Labeling of inhibition of endocytosis below the basal rate. Alterna- lysosomes and basolateral plasma membrane occurred tively, the effect might be the consequence of loss of only at 25’C or above, whereas multivesicular bodies CF-particles after recycling. were labeled, together with endosomes, also at 15°C. The observations show that transcytosis involves Therefore, the morphometrical data on multivesicular about 12% of the apical plasma membrane and about bodies are included in those on endosomes (Fig. 10). 10% of TG internalized by TSH-stimulated cells. Pri- mary endocytic and transcytotic vesicles are of the same Transcytosis of [3H] TG size and principally able to carry the same amount of When the luminal content of TSH-stimulated inside- TG molecules. This might explain the fact that the same out follicles was analyzed by fluorography after SDS- proportion of membrane and of TG internalized is PAGE, a single band in the range of intact TG was de- transcytosed. tected. Intact TG and low molecular weight bands repre- Transcytosis is also regulated by TSH, but differs senting degradation products were only detected in the from endocytosis in being stimulated above resting val- cell lysates from the same experiments (Fig. 11). Quan- ues by a factor of about 4 for prolonged periods of time titative analysis showed that the amount of TG transcy- (see Fig. 4). This effect of TSH exerted on transcytosis tosed was 10.5% of that internalized (SE f 1.4; range 7 may explain clinical and experimental observations to 13). which show that after injection of TSH the serum con- centrations of TG in humans and various mammalian DISCUSSION species are enhanced by a factor of about 4.8 (for review see [23]). The question arises as to whether transcytotic Endocytosis and Transcytosis Are Regulated by TSH vesicle formation is directly regulated by TSH or Our morphometric observations indicate that in rest- whether the enhanced appearance of vesicles along the ing conditions membrane corresponding to about the basolateral plasma membrane results from a saturation 4.6 fold area of apical cell surface is removed during 60 of the pathway to lysosomes. Apparently, lysosomes are min of endocytosis. This observation is in agreement not saturated because their labeling increases progres- with the results of others obtained by the use of differ- sively with time. Therefore, we favor the idea of selec- ent experimental approaches (for review see [36]). Sur- tive regulation of transcytosis by TSH. prisingly, the effect of TSH is only transient and, about Inside-out follicules from pig thyroid glands have 50 min after the addition of TSH, the rate of endocyto- been found to be tight against the diffusion of macromol- sis decreases below the rate in resting follicles (see Fig. ecules [20,21] and stimulated by the addition of TSH to the culture medium. Unless TSH is not transcytosed with subsequent action on the basolateral cell surfaces, we have to conclude that a sufficient number of TSH- receptor molecules is located on the apical cell surfaces. However, the localization of the TSH-receptor on thy- rocytes is still a matter of debate. This paper describes a TSH-response from the apical cell surface, but it ex- ceeds the purpose of this study to include analyses on kDa the localization of the receptor. - 330 Transcytosis and Transfer to Lysosomes Share a Common Temperature-Dependent Gating Step

The morphometric observations show that both trans- port to lysosomes and transcytosis are temperature-de- pendent. Both processes are not observed at 15°C al- though endocytosis continues at this temperature. At FIG. 11. Transcytosis of thyroglohulin through inside-out thy- roid follicles. Gel electrophoresis and fluorography of hiosyntheti- 25°C the transport to lysosomes and transcytosis are tally labeled 13H]TG administered to inside-out follicles (lane l), cell observed, but at a lower rate than at 37°C. Therefore, lysate (lane 2), and transcytosed material (lane 3) after 5 h incubation the fusion processes of endocytic vesicles with the lyso- in MEM containing biosynthetically radiolabeled thyroglobulin. Deg- somal compartment and the basolateral cell surfaces radation products of thyroglobulin are present in the cell lysate, show a striking similarity with regard to their tempera- whereas only intact thyroglobulin, with the same electrophoretic mo- bility as that administered to the inside-out follicles, is seen in the ture dependence. This points to the possibility that transcytosed material. these two processes share a common temperature-de- 208 ROMAGNOLI AND HERZOG pendent gating step. The labeling of endosomes at 15°C small fraction of degraded TG is released from lyso- suggests that this compartment is reached before the somes during this process. This, however, was not ob- temperature-dependent gating step and that the endo- served. Hence, the observations would suggest that thy- somal compartment is part of the transcytotic pathway. roid hormones are not released by a vesicular mecha- Our observations also show that endosomes are reached nism. Alternatively, hydrolyzed TG might be selectively before the basolateral cell surfaces become labeled. retained in lysosomes from which T3- and T4-transport- However, the experimental model used here does not ing vesicles detach. allow conclusions on an obligatory role of endosomes in membrane translocation during transcytosis. Participa- We thank Drs. T. Kehle and W. Neumtiller for discussions and Dr. tion of endosomes in transcytosis would also allow CF R. Stiemerling for valuable organizational help. This work was sup- particles to detach from the endosomal membrane at ported by the Alexander von Humboldt-Stiftung (P.R.) and by grants low pH and detachment of CF could have led to an un- from Deutsche Forschungsgemeinschaft, the Wilhelm Sander-Stif- tung, and the Chemische Industrie (V.H.). derestimation of the membrane fraction translocated during transcytosis. However, since no CF particle was seen in the matrix of endosomes (in contrast to lyso- somes), we favor the interpretation that only a minor REFERENCES fraction of CF might detach from membrane portions undergoing transcytosis and that our quantitative data 1. Bruns, R. R., and Palade, G. E. (1968) J. Cell Biol. 37,277-299. on the proportion of membrane participating in trans- 2. Simionescu, N., Simionescu, M., and Palade, G. E. (1975) J. Cell are correct. Biol. 64, 586-607. The effect of cooling on the transfer of endocytic vesi- 3. Palade, G. E., Simionescu, M., and Simionescu, N. (1979) Actu cles to lysosomes parallels the effect of temperature on Physiol. Stand. Suppl. 463, 11-32. the secretion of thyroid hormones [37-391. In other cell 4. Hachiya, H. L., Halban, P. A., and King, G. L. (1988) Am. J. types, a block in transport to lysosomes has been ob- Physiol. 255, C459-C464. served at 20°C and below [18, 401. Interestingly, the 5. Rodewald, R. (1973) J. Cell Biol. 68, 189-211. transfer to multivesicular bodies (MVB) still continues 6. Brandtzaeg, P. (1974) J. Immunol. 112, 1553-1559. in thyrocytes at 15°C. Apparently, MVB and secondary 7. Siminoski, K., Gonnelia, P., Bernanke, J., Owen, L., Neutra, M., lysosomes differ in their ability to fuse with incoming and Murphy, R. A. (1986) J. Cell Biol. 103,1979-1990. vesicles. These observations differ from results with 8. Abrahamson, D. R., and Rodewald, R. (1981) J. Cell Bill. 91, MDCK cells, in which the transport to lysosomes and 270-280. MVB is completely blocked at 18°C [18]. Also, in 9. Cooper, T. G., Yeung, C.-H., and Bergmann, M. (1988) Cell Tis- sue Res. 253,631-637. MDCK cells [ 181 a discrepancy has been found between 10. Van Deurs, B., van Btilow, F., and Moller, M. (1981) J. Cell Biol. transfer to lysosomes (which is totally inhibited at 89, 131-139. 18°C) and transcytosis (which still occurs at this tem- 11. Van Deurs, B., Petersen, 0. W., Olsnes, S., and Sandvig, K. perature). These observations show that MDCK cells (1989) Znt. Rev. Cytol. 117,131-177. and thyrocytes differ in the sensitivity of their transport 12. Kraehenbuhl, J.-P., Bron, C., and Sordat, B. (1979) Curr. Top. pathways towards low temperatures. P&201. 66, 105158. 13. Huxham, M., and Beck, F. (1981) Cell Biol. Znt. Rep. 5, 1073- Thyroglobulin Is Recovered after Transcytosis without 1081. Molecular Weight Changes 14. Wilson, J. M., and King, B. F. (1985) Anut. Rec. 211, 174-183. 15. Fleming, T. P., and Goodall, H. (1986) Anut. Rec. 216,490-503. The qualitative analysis of TG before and after 16. Renston, R. H., Jopnes, A. L., Christiansen, W. D., Hradek, transcytosis shows that there is no detectable molecular G. T., and Underdown, B. J. (1980) Science 208, 1276-1278. weight change during transport across the follicle wall. 17. Geuze, H. J., Slot, J. W., Strous, G. J. A. M., Peppard, J., von This suggests that transcytotic vesicles bypass the lyso- Figura, K., Hasilik, A., and Schwartz, A. L. (1984) Cell 37,195- somal compartment. Therefore, our results allow the 204. conclusion that the differences in molecular weight be- 18. Van Deurs, B., Hansen, S. H., Petersen, 0. W., Lokken Melby, E., and Sandvig, K. (1990) Eur. J. Cell Biol. 51, 96-109. tween TG from the follicular lumen and the circulation observed in vivo [41] are not due to partial hydrolysis 19. Fukami, Y., and Schlesinger, P. H. (1989) Bruin Res. 499,249- 257. during transcytosis, but occur after TG release from thy- 20. Herzog, V. (1983) in “Methods in Enzymology” (Fleischer, S., roid follicles, possibly within the circulation. and Fleischer, B., Eds.), Vol. 98, pp. 447-458, Academic Press, The release mechanism for T3 and T4 is still un- New York. known: it may occur by diffusion or be mediated by a 21. Herzog, V. (1983) J. Cell Biol. 97, 607-617. vesicular mechanism. If a vesicular mechanism operates 22. Herzog, V. (1984) Int. Rev. Cytol. 91, 107-139. in the transfer of T3 and T4 from lysosomes to the baso- 23. Van Herle, A. J., Vassart, G., and Dumont, J. E. (1979) N. En&. lateral cell surface, then one could expect that also a J. Med. 301,239-249,307-314. TRANSCYTOSIS THROUGH THYROID FOLLICLE CELLS 209

24. Ericson, L. E. (1981) Mol. Cell. Endocrirwl. 22, l-24. 34. Lentner, C., Lentner, C., and Wink, A. (1982) Geigy Scientific 25. Karnovsky, M. J. (1965) J. Cell Bid. 27,137A-138A. Tables, Vol. 2,8th ed., Ciba-Geigy, Basel. 35. Nitsch, L., and Wollman, S. H. (1986) Ewp. Cell Res. 162,278- 26. Nitsch, L., and Wollman, S. H. (1986) J. Cell Bid. 86,875-880. 284. 27. Garbi, C., and Wollman, S. H. (1982) Exp. Cell Res. 136, 343- 36. Wollman, S. H. (1989) Eur. J. Cell Biol. 60, 247-256. 353. 37. Rocmans, P. A., Ketelbant-Balasee, P., Dumont, J. E., and 28. Romagnoli, P. (1984) Hi&o&em. J. 16,855-868. Neve, P. (1978) Endocrinology 103, 1834-1848. 29. Baddeley, A. J., Gundersen, H. J. G., and Cruz-Orive, L. M. 38. Van Sande, J., Ketelbant-Balasse, P., Grenier, G., and Dumont, (1986) J. Microsc. 142, 259-276. J. E. (1978) Endocrinology 102, 576-581. 30. Herzog, V., and Miller, F. (1979) Eur. J. Cell Biol. 19, 203-215. 39. Ring, P., Bjorkman, U., and Ekholm, R. (1987) Cell Tissue Res. 247,505-513. 31. Weibel, E. R. (1979) Stereological Methods, Vol. 1, Academic Press, London. 40. Griffiths, G., Hoflack, B., Simons, K., Mellman, I., and Korn- feld, S. (1988) Cell 52,329-341. 32. Laemmli, U., and Favre, M. (1973) J. Mol. Biol. 80,575-599. 41. Schneider, A. B. (1985) in Thyroglobulin-The Prothyroid Hor- 33. Bonner, W. M., and Laskey, R. A. (1974) Eur. J. Biochem. 46, mone (Eggo, M. C., and Burrow, G. N., Eds.), pp. 297-366, Ra- 83-88. ven Press, New York.

Received September 7,1996 Revised version received December 27, 1990