In vivo shedding of apical plasma membrane in the thyroid follicle cells of the mouse

Cell Tissue Res (1984) 236:87-97

a n d T'msue R e s e a l r . h �9 Springer-Verlag 1984

In vivo shedding of apical plasma membrane in the thyroid follicle cells of the mouse Mikael Nilsson, Torsten Ofverholm, and Lars E. Ericson Department of Anatomy, University of G6teborg, G6teborg, Sweden

Summary. Clusters of luminal dense bodies, limited by a triple-layered membrane, were found in all follicle lumina in thyroid glands of mice. After thyroxine treatment the number of luminal dense bodies increased, especially in the periphery of the lumen, where the intraluminal bodies often displayed a striking resemblance to microvilli. In hyperplastic goiters, obtained by feeding mice with propylthiouracil, luminal dense bodies were replaced by intraluminal vesicles. During goiter involution the vesicles were gradually replaced by luminal dense bodies; the presence of intermediate forms suggests that vesicles and dense bodies are basically the same formations. Luminal dense bodies were observed in colloid droplets indicating their removal by endocytosis. As demonstrated by electron-microscopic cytochemistry, luminal dense bodies contain a membrane- bound peroxidase, and electron-microscopic autoradiography after administration of 1251 indicate that they possess an iodinating capacity.

Our observations on mouse thyroid glands suggest that the luminal dense bodies, which appear as vesicles in hyperplastic glands, are formed by shedding of the apical plasma membrane of the follicle cell. The shedding process might be of importance for the turnover of plasma-membrane material.

Key words: Iodination - Membrane shedding - Peroxidase - Thyroid follicle cell - Ultrastructure

The turnover of the apical plasma membrane of the thyroid follicle cell is rapid (Ericson and Johanson 1981). Mem- brane material is constantly added to the apical plasma membrane by exocytosis and removed by endocytosis. Ster- eological measurements strongly indicate a close functional relationship between exocytosis and endocytosis in acutely TSH-stimulated thyroid follicle cells of rats (Ericson 1981). However, other mechanisms of turnover of the apical plasma membrane are also possible. One such mechanism could be that plasma-membrane material is shed into the follicle lumen. In fact, Tachiwaki and Wollman (1982) recently obtained evidence of shedding in thyroid follicle cells in rats during involution of goitrogen-induced thyroid hyperplasia.

Send offprint requests to: Dr. Mikael Nilsson, Dept. of Anatomy, Univ. of G6teborg, Box 33031, S-40033 G6teborg, Sweden

This study was supported by Grant No. t2X-537 from the Swedish Medical Research Council.

Here we present evidence that plasma-membrane shedding occurs in normal mouse thyroid follicles, and also that experimental manipulations of the thyroid activity influence the amount of shed material accumulated in the follicle lumen.

Plasma-membrane shedding is defined as shedding of whole portions of the plasma membrane to the extracellular environment. Plasma-membrane shedding, synonymous to vesiculation (Scott 1976; Lutz etal. 1977), exfoliation (Koch and Smith 1978; Trams et al. 1981) and clasmatosis (Vitetta and Uhr 1972; Biberfeld 1973), can be considered as a subtype of the general term "cell-surface shedding", which in addition to plasma-membrane shedding also com- prises shedding of different kinds of single molecules (glyco- proteins, phospholipids, glucosaminoglycans) as well as molecular complexes (antigen-antibody complexes, recep- tor-ligand complexes) of membrane origin. The general phenomenon "cell-surface shedding" has been recently exten- sively reviewed (Doljanski and Kapeller 1976; Cone 1977; Black 1980) but a similar review detailing plasma-membrane shedding is not, to our knowledge, available. Never- theless, plasma-membrane shedding has been observed in a number of cell types under different physiological and pathological conditions, both in vivo and in vitro; in the discussion of the present results, we include a brief summary of studies in which plasma-membrane shedding has been reported.

Materials and methods

Animals. Male albino mice, weighing 15-35 g and (if not otherwise stated) maintained on a standard pellet diet (As- tra-Ewos, S6dert/ilje, Sweden) containing 2.6 I~g iodine/g and tap water, were used. In addition to normal mice of different ages of the N M R I strain (Anticimex, S6dert~ilje, Sweden) and Balb strain (Bomholdtgaard, Aarhus, Den- mark), groups of NMRI mice (5-10/group) pretreated in different ways were studied. The following pretreatments were given: 1) L-thyroxine (T4; 20 ~tg) injected subcutane- ously 48 h and 24 h before fixation. 2) Propylthiouracil (PTU; 0.5% in the diet) for 14 or 21 days. 3) PTU in diet and in addition T 4 for 1 or 2 days before fixation. 4) PTU and T 4 as in the previous group and in addition TSH (Actyron, Ferring AB, Malm6, Sweden; 50 mU) injected in a tail vein 30 min before fixation.

Electron microscopy. The animals, anesthetized with pento- barbital, were fixed by perfusion via the left heart ventricle

88

with 2.5% glutaraldehyde in 0.05 M sodium cacodylate, pH 7.2, for 5 min at 37 ~ C. After excision, the thyroid lobes were immersed in a similar glutaraldehyde solution for 1-2 h and then postfixed in 1% osmium tetroxide in 0.1 M sodium cacodylate for 1 h. After dehydra t ion in ethanol, specimens were embedded in Epon. Sections, contrasted with uranyl acetate and lead citrate, were examined in a Philips 300 electron microscope.

Electron-microscopic autoradiography. Mice, pretreated as in groups 1 (4 mice) and 3 (6 mice) described above, were injected intraveneously with 200 gCi sodium lzsI (carrier- free, N E N Chemicals, Dreieich, F R G ) 10 and 60 min before fixation by perfusion with 2.5% glutaraldehyde in 0.05 M sodium cacodylate. To inhibit further incorpora t ion of la-

Fig. t. Normal mouse thyroid follicle. Luminal dense bodies, mostly forming small clusters, are distributed throughout the follicular lumen (FL). Dense bodies are also present in colloid droplets (arrows) in the cytoplasm of follicle cells, x 5500. Inset: The luminal dense bodies are limited by a membrane and have a homogeneous content, x 40000

Fig. 2. Mouse injected with T 4 for 2 days. The majority of the luminal dense bodies are located in the periphery of the follicular lumen within or close to the microvillous region of the follicle cell. No colloid droplets are present in the follicle cells. x 6300

bel, the fixative contained I m M resorcinol and 10 m M KI. The thyroid tissue was further processed for electron microscopy as described above. Sections, about 100 nm thick, were picked up on Formvar -coa ted copper grids, contrasted with lead citrate and uranyl acetate and then covered by a layer of carbon by vacuum evaporat ion. The emulsion (Ilford L4) was applied by means of a wire-loop. After exposure for 4 weeks, the preparat ions were developed in K o d a k D-19 for 2 min and fixed in K o d a k F-24 fixer for 2 min.

The distr ibution of autoradiographic grains was evalu- ated quanti tat ively by counting the number of grains over- lying intraluminal material and colloid proper. The areas of these compartments were determined by a digitizer and the grain density calculated.

89

Electron-microscopic cytochemistry. For the demonstration of peroxidase activity thyroids from 6 mice, pretreated as in group 3 described above, were fixed by perfusion with 2.5% glutaraldehyde in 0.05 M sodium cacodylate, pH 7.2, for 5 min at room temperature and then immersed in the same fixative for 30 min. Thyroid slices, produced by use o f a razor blade, were thoroughly washed in 0.05 M Tris- HC1 buffer, pH 7.6, and then incubated at room temperature in a medium containing 0.05% 3-3'-diaminobenzidine hydrochloride (Sigma, St. Louis, MO, USA) and 0.001% H202 in 0.05 M Tris-HC1. Controls were incubated in the same medium without H202 or with added resorcinol (1 mM) or catalase (600 units/ml; Sigma). After incubation the slices were postfixed in osmium tetroxide and further processed for electron microscopy as described above, with the exception that sections were not contrasted.

Results

Normal mice

The thyroid follicle cells in normal mice were in general flat to cuboidal. The apical plasma membrane formed typical microvilli and occasionally pseudopods. Colloid droplets were frequent (Fig. 2).

All follicle lumina, without exception, contained dense bodies (Fig. 1). These luminal dense bodies (LDB) were most often elongated, with a length of about 400 nm and a width of about 150 nm. Droplet-shaped profiles were also common (Fig. 1). LDB were limited by a triple-layered membrane, about 9 nm thick, and contained a homogeneous material (Fig. 1, inset). LDB most often formed small aggregates, which were spread throughout the lumen. Occa- sionally, LDB were present in colloid droplets (Fig. 1). The number, size and distribution of LDB appeared to be roughly the same irrespective of the age and strain of the mouse.

T 4-treatment

Injection of T 4 for 1--2 days caused an inhibition of thyroid macropinocytosis as indicated by the absence o f pseudo-

Fig. 3. T4-treated mouse. An aggregate of luminal dense bodies is located in the periphery of the lumen close to microvilli (my) of the follicle cell. The dimensions and shapes of the dense bodies are similar to those of the microvilli. In some profiles of dense bodies an internal striation can be distinguished (arrows). x 29000

Fig. 4. T4-treated mouse. Several structures (arrows) with a dense core and limited by a triple-layered membrane located in the microvillous region. The dense structures cannot be distinguished definitely as either luminal dense bodies or microvilli with a dense core. Exocytotic vesicles (arrowheads) are present in the apical cytoplasm, x 41000

pods and colloid droplets (Fig. 2). The apical cytoplasm contained numerous vesicles, about 180 nm in diameter, with a moderately dense content (Figs. 2-4). Similar vesicles were also present in normal mice but in lower numbers. They are probably exocytotic in function, transferring newly synthesized thyroglobulin to the follicle lumen (Bjrrkman et al. 1976).

As compared to normal mice, the LDB were more numerous in the T4-treated mice and their distribution dif- fered. In addition to their presence in the central regions o f the lumen, they were also frequently located in the periphery o f the lumen close to or between the microvilli formed by the apical plasma membrane of the follicle cell (Figs. 2, 3). At this site many LDB had the same dimensions as microvilli and sometimes they also contained longitudi- nally arranged filaments (Fig. 3). Microvilli with a dense matrix and with a limiting membrane in continuity with the apical plasma membrane were also observed (Fig. 4). Thus, a general impression was obtained that LDB were formed from microvilli.

Treatment with propylthiouracil

PTU inhibits thyroid peroxidase, which leads to a decreased synthesis and decreased plasma levels of thyroid hormones and a compensatory increment in TSH secretion. In the present study feeding PTU for 2 or 3 weeks caused a hyper- trophy of the follicle cells and a decrease in the size of the follicle lumen, which often attained a narrow, elongated shape. Pseudopods and colloid droplets were present in hy- pertrophic follicle cells, while exocytotic vesicles were few.

The colloid in the majority o f the follicles had a low density and in such follicles typical LDB were not present. Instead, membrane-limited vesicles with a content of lower density than that o f LDB were observed (Figs. 5-8). These vesicles, single or in small clusters, were located close to the apical plasma membrane (Figs. 5, 8), and they were generally not present in the central parts of the follicle lumen. However, some luminal profiles also contained large aggregates of vesicles (Fig. 5). The diameters of the vesicles ranged between 100 nm and 300 nm. A few vesicle profiles contained one or two smaller granules of higher density

90

Fig. 5. Mouse fed PTU for 3 weeks. Spherical vesicles with a rather dense content are present along the apical surface of the follicle cell and form a large aggregate in the lumen. The colloid has a rather low density. x 7500

Fig. 6. Detail of the aggregate of intraluminal structures shown in Fig. 5. The size of the vesicles varies. Several of the vesicles contain dense granules. Small granules (arrows) with a diameter of about 30 nm are present among the larger vesicles, x 23000

Fig. 7. PTU for 3 weeks. Detail of intraluminal aggregate. The indicated structure (*) is an intermediate form between a typical luminal dense body and a vesicle, and is limited by a triple- layered membrane, x 48 000

Fig. 8. PTU for 3 weeks. An aggregate, consisting of numerous small granules and vesicles, is located close to the apical surface of the follicle cell. x 14000

Fig. 9. PTU for 3 weeks and T 4 for 1 day. The section passes obliquely through the apical end of a follicle cell. Numerous, mostly elongated luminal dense bodies, as well as small granules are located close to or between the microvilli. In the right upper corner an aggregate is present containing luminal dense bodies (arrowhead), as well as vesicles (arrows) and small granules. x 12000

(Fig. 6). In addition, vesicle aggregates also contained large amounts of small granules which were limited by a membrane and had a diameter of about 30 nm (Fig. 8). These granules were exclusively found in PTU-treated mice.

Lumina containing only LDB but no vesicles were also observed in PTU-treated thyroids, but exclusively in follicles with a colloid density similar to that of normal glands.

Treatment with propylthiouracil and thyroxine

Administrat ion of T 4 for 1 or 2 days to mice given PTU for 2 or 3 weeks induced an accumulat ion of colloid, which

led to an increase in the size of the follicle lumina and an increase of the colloid density, being most pronounced after 2 days. Pseudopods and colloid droplets were absent, while exocytotic vesicles were numerous in the apical cytoplasm (Fig. 10). Desquamated cells were observed in about 25% of the luminal profiles.

Injection of T 4 to PTU-treated mice caused the reap- pearance of LDB, which were associated with the apical plasma membrane, but also formed large aggregates in the lumen (Figs. 9, 10). Thus, the distribution of LDB was similar to that observed in mice given T 4 only, although their number in each luminal profile was appreciably higher. A

91

small number o f follicles still contained colloid of lower density, and such follicles contained luminal vesicles exclusively or together with LDB (Fig. 9).

Injection of TSH

TSH given to normal mice or mice pretreated with PTU and T 4 30 min before fixation induced exocytosis, as indicated by a pronounced decrease in the number o f exocytotic vesicles, and endocytosis, as signified by the appearance of pseudopods and colloid droplets. Newly formed colloid droplets located in pseudopods or in the cytoplasm often contained LDB or vesicles of the same type as present in the follicle lumen (Fig. 11).

Electron-microscopic autoradiography

In mice pretreated with T 4 and given radioiodide 10 rain before fixation, the autoradiographic grains were distributed in most follicles in the periphery of the lumen forming a typical ring-reaction. A similar distribution was also present in some follicles 60 min after exposure to radioiodide. In the lumina with a ring-reaction an association between the low number of grains present over central portions of the lumen and aggregates of LDB, which in general were rather small (cf. Fig. 2), was not apparent. However, in the lumina that lacked a ring-reaction and in which label

Fig. 10. PTU for 3 weeks and T 4 for 2 days. The colloid has a rather high density and contains several aggregates of luminal dense bodies. The apical cytoplasm contains numerous exocytotic vesicles, some of which are indicated (arrows). x 8500

Fig. 11. Mouse pretreated with PTU and T 4 and then injected with TSH 30 min before fixation. Colloid droplets (CD) in the cytoplasm containing dense bodies and small granules, x 19000

was more uniformly distributed such an association was in general evident (Figs. 12, 13).

In mice pretreated with PTU and T 4 ring-reactions were usually not present. As described above, the follicular lumen contained large aggregates of LDB and vesicles and the concentration of silver grains was in general appreciably higher over aggregates than over the colloid proper both 10 and 60 min after injection of radioiodide (Figs. 14-16). This was confirmed by quantitative measurements. In two mice injected with radioiodide for 60 min the relative density (setting the grain density over colloid proper to 1) over intraluminal aggregates was 6 .3+1.67 and 6.0___0.69 (mean _+ SE for 20 follicles).

Electron-microscopic eytochemistry

Under the incubation conditions used, a pronounced accumulation of reaction product was found along the apical border of the follicle cell, while the cytochemical reaction in cytoplasmic structures, such as rough endoplasmic reticulum and exocytotic vesicles, was notably weaker although present (Fig. 17). The reaction was most pronounced in follicles close to the surface of the slice, but even in such follicles not all follicle cells had a reaction product along the apical border. The dense reaction product formed a gradient extending from the outer side of the apical plasma membrane into the lumen (Fig. 17). Intraluminal aggregates

92

also displayed a positive cytochemical reaction. Also at this location the reaction product formed a gradient on the outer side of the membrane limiting each single structure, the content of which was unreactive. All follicular lumina containing reactive aggregates also displayed unreactive aggregates (Fig. 17). Cells present in the lumen were generally unreactive.

No reaction product along the apical plasma membrane or in intraluminal aggregates was found in slices incubated in the absence of H 2 0 / or in the presence of resorcinol or catalase.

Discussion

Morphology and origin of intraluminal structures

In this study we have described the presence of membrane- limited structures, dense bodies and vesicles, in the lumen

Fig. 12. Electron-microscopic autoradiography. Mouse given thyroxine for 2 days and injected with radioiodine 60 min before fixation. Many of the silver grains present in the lumen are associated with aggregates of luminal dense bodies. In this follicle no grains are present at the apical surface of the follicle cell (FC). • 6000

Fig. 13. Detail from previous figure showing accumulation of grains over intraluminal dense bodies. • 15000

Fig. 14. Mouse given PTU for 3 weeks and thyroxine for 2 days. Ten min after radioiodide. Autoradiographic grains are concentrated over aggregates of intraluminal dense bodies but are also present over the colloid, x 7000

Fig. 15. Mouse given PTU for 3 weeks and thyroxine for 2 days. Electron- microscopic autoradiograph 60 min after injection of radioiodide. Silver grains are to a large extent associated with aggregates ofintraluminal structures but are not present at the apical surface of the follicle cell (FC). x 9000

Fig. 16. Same pretreatment as in Fig. 15; 60 min after radioiodide. Silver grains are concentrated over intraluminal aggregates located in the central portion of a follicular lumen. x 6000

of mouse thyroid follicles. The intraluminal structures were present in normal and T4-treated mice as well as in mice given PTU for 2-3 weeks. The ultrastructure was related to the density of the colloid in which the material was embedded as evident in mice given PTU for 2 or 3 weeks. After PTU-treatment only, the colloid had a low density and the intraluminal material appeared mainly in the form of vesicles. After administration of T 4 to PTU-treated mice the density of the colloid increased and the intraluminal material now appeared mainly as nonspherical, dense bodies similar to those present in follicles of normal mouse thyroids. In some colloids both types of structures as well as intermediate forms appeared simultaneously. These observations suggest that the dense bodies and the vesicles are basically the same formations, the structure of which depends on the physico-chemical properties, as, for example, the protein concentration of the colloid. However, the

93

Fig. 17, Thyroid tissue of a mouse given propylthiouracil for 3 weeks and thyroxine for 2 clays processed for the demonstration of peroxidase. A dense reaction product surrounding microvilli is present along the apical plasma membrane (arrows). A reaction product is also present in an aggregate of intraluminal material (**). Other aggregates (*) are unreactive or have a very weak reaction. Colloid droplets (CD) in the apical cytoplasm are unreactive, x t0000

small granules, about 30 nm in diameter, and the granule- containing vesicles observed in PTU-treated mice do not seem to have their counterparts in other conditions.

The amount of LDB increased following injection of T 4 to either normal or PTU-treated mice. In such mice LDB were frequently observed in close relation with the microvillous region of the follicle cell. LDB in this position often had the same dimensions and shapes as microvilli and also possessed internal filamentous cores similar to that seen in microvilli. Thus, at least a portion of the intraluminal structures seems to be formed by the pinching off of microvilli, implying that not only the membrane but also the core of the microvilli is lost from the follicle cell. Also the intraluminaI vesicles appearing in PTU-treated mice were located close to or within the microvillous region. Shedding of microvilli has also been observed in other cell types (Marcus 1962; Vitetta and Uhr 1972; Vitetta et al. 1974; Koch and Smith 1978; Liepins et al. 1978; Van Blit- terswijk et al. 1979; Black et al. 1980; Huggins et al. 1980; Misch et al. 1980; Nguyen and Woodard 1980).

Tachiwaki and Wollman (1982) recently reported the occurrence of dense cell fragments, with structural similari- ties to the dense bodies and vesicles observed in the present study, in rat thyroid follicle lumina at early stages of involution of hyperplastic glands. The fragments disappeared when normal steady state was obtained. Thus, in this study on the rat thyroid gland, shedding seems to be a temporary

phenomenon associated with a period of alteration in function of thyroid follicle cells, while our observations on mice indicate that plasma-membrane shedding occurs in the normal gland as well. It also appears possible that the temporary accumulation of dense vesicles in the periphery of follicle lumina of bat thyroids just before the arousal from hi- bernation observed by Nunez et al. (1974) might be another example of plasma-membrane shedding in the thyroid gland.

Peroxidase activity and iodination in intraluminal structures

Iodination of thyroglobulin normally takes place in the lumen at the apical plasma membrane of the follicle cell and is catalyzed by thyroperoxidase bound in this membrane (Ekholm and Wollman 1975; Ekholm 1981). In the present study we found a positive reaction for peroxidase in the apical plasma membrane as well as in aggregates of intraluminal structures. In both these locations the dense reaction product was located on the outer side of the membrane forming a gradient extending into the colloid. This gradient is most likely the result of diffusion of reaction product formed at the membrane surface by the action of membrane-bound peroxidase rather than an indication of the presence of soluble enzyme in the colloid. The presence of peroxidase activity in membranes of intraluminal struc-

94

tures as well as in the apical plasma membrane supports the idea that the former originates from the latter.

A positive reaction for peroxidase was also found in intracellular compartments of the follicle cell, such as in rough endoplasmic reticulum and exocytotic vesicles, and the possibility therefore exists that peroxidase-positive membranes might have been brought to the lumen by des- quamation and lysis of follicle cells. However, as discussed above, the reaction product in intraluminal aggregates was always located on the outer side of the limiting membrane, while that in intracellular compartments was present on the inside of the membrane more or less filling the interior. This observation makes an intracellular origin of the intraluminal peroxidase-positive membranes less likely. This is also supported by the observations that the cytochemical reaction in intracellular compartments was weak compared to that in the apical plasma membrane and intraluminal structures and that cells present in the lumen (in PTU- treated rats) in general did not contain any reaction product.

Following injection of radioiodide, autoradiographic grains were found to be concentrated over aggregates of intraluminal structures. The most direct interpretation of this finding is that intraluminal material possesses an iodinating capacity, which is in keeping with the presence of peroxidase in their membranes and an origin from the apical plasma membrane. The high concentration of grains over the intraluminal aggregates suggests that the labeled product is not freely diffusible. Whether this is due, for example, to trapping of newly iodinated thyroglobulin or whether constituents of the dense bodies and vesicles in the aggregates are iodinated is not known.

Other possible origins of intraluminal structures

Feeding with a goitrogen causes an increase in the number of thyroid cells, and cessation of this feeding induces a rapid decline in the cell number (Wollman and Breitman 1970). In accordance with this fact we observed dead cells in the follicular lumen of mice fed PTU and then injected with T 4 to inhibit TSH secretion. At the same time the number of LDB and vesicles in the lumen increased conspi- cuously. This may suggest an origin of the intraluminal structures from desquamated cells. However, the fact that

L D B were encountered also in normal mice in which free cells in the lumen are not usually present, as well as the absence of any structures that could be identified as organ- eltes in the aggregates of intraluminal structures, do not favor dead cells as the origin of the intraluminal material. In the preceding section we also discussed observations from electron cytochemistry supporting the idea that the apical plasma membrane rather than intracellular components of the follicle cell is the source of the intraluminal material.

Macrophages and erythrocytes were also, although rare- ly, observed in the lumen of hyperplastic thyroids. Both cell types can theoretically be responsible for supplying material causing a positive cytochemical reaction to the lumen, since macrophages contain a soluble peroxidase and hemoglobin is known to oxidize diaminobenzidine. However, in either case a location of the reaction product inside the limiting membrane would be expected, while our cytochemical observations indicate the presence of a membrane- bound peroxidase facing the colloid. Furthermore, in the

case of hemoglobin the cytochemical reaction is not expected to be influenced by addition of resorcinol, an inhibi- tor of peroxidase, or elimination of H20 2 by addition of catalase.

It should be pointed out that the formation of luminal dense bodies and vesicles by pinching off from microvilli as suggested by our ultrastructural observations bears some resemblance with the budding-off of virus particles from the host cell (Marcus 1962). However, the intraluminal structures differ greatly from virus particles, which makes this possibility unlikely.

Shedding and functional state of the follicle cells

Administration of T 4 to normal or PTU-fed mice caused an accumulation of LDB. In rats, T4-treatment causes an inhibition of thyroid macropinocytosis and hormone release (Bjrrkman et al. 1974). The absence of pseudopods and colloid droplets observed in the present study indicates that a similar inhibition also occurs in the mouse thyroid. The presence of LDB in colloid droplets in normal and TSH- injected mice suggests that intraluminal material is removed by macropinocytosis. The accumulation of LDB observed in T4-treated mice might therefore be the result of a decreased rate of removal by macropinocytosis. T4-treatment also inhibits protein synthesis and the transfer of newly synthesized protein to the follicular lumen. However, the decrease in protein synthesis is relatively negligible and gradual as compared to the almost complete inhibition of macropinocytosis (Bjrrkman et al. 1974). Newly synthesized protein is transferred to the follicular lumen in exocytotic vesicles, which fuse with the apical plasma membrane (Bjrrkman et al. 1976), and T4-treatment therefore creates a situation in which membrane continues to be inserted into the apical plasma-membrane, while removal of membrane material by macropinocytosis is hampered. Although retrieval by micropinocytosis probably still occurs, it appears possible that T4-treatment causes an imbalance between the amount of membrane inserted by exocytosis and removed by endocytosis and that the excess of plasma membrane is shed to the follicle. Thus, the increased amounts of intraluminal material observed in T4-treated mice might be due to both an increased rate of shedding and a decreased rate of removal of shed membranes by endocytosis. It appears reasonable to assume that the very pronounced accumulation of intraluminal material observed in PTU-fed mice given T 4 reflects the higher rates of membrane turnover in hyperplastic cells and that the mechanism of accumulation of material in the lumen is basically the same as in mice given T 4 only. If the rather low amounts of intraluminal structures present in animals fed only PTU reflect a decreased rate of shedding or an efficient removal of shed material by pseudopod-colloid droplets, which were present in the hyperplastic cells, cannot be decided at present.

Plasma-membrane shedding in other cell systems

Reports concerning plasma-membrane shedding are few compared with reports of shedding of membrane-derived molecules, the latter mainly studied in vitro (Doljanski and Kapeller 1976; Cone 1977; Black 1980). However, in addition to the observations mentioned above of possible plasma-membrane shedding in the thyroid (Nunez et al. 1974; Tachiwaki and Wollman 1982), several studies have re-

95

vealed similar findings in other organs in vivo. Shedding of outer segment discs, following a circadian rhythm, seems to be a universal phenomenon in retinal photoreceptors (Young 1976). Pigment epithelial cells then phagocytose the shed plasma-membrane fragments. Another example, where plasma-membrane shedding might be of importance in vivo is the epithelium of the intestine. So-called membrane- bounded bodies are released by intestinal cells in hamsters (Misch et al. 1980) and rats (Jacobs 1981). These findings are confirmed by retrospective analysis (Misch et al. 1980) of earlier ultrastructural studies and results obtained by analysis of duodenal fluids from humans (DeBroe et al. 1977) and chick embryos (Black et al. 1980), as well as the medium from cultured chick embryo (Black et al. 1980) and fetal rat (De Ritis et al. 1975) duodenum. Interestingly, the release is releated to the intermittent administration of nu- trients (Misch et al. 1980), but whether the released vesicles only reflect normal plasma-membrane turnover (Misch et al. 1980) or possess a digestive (Black et al. 1980; Misch et al. 1980) or an immunological (Black et al. 1980) function has not been clarified. In human urine, membrane vesicles containing an enzyme profile identical to the cell membrane of the renal proximal tubule cell have been found (DeBroe et al. 1977), and since it is known that cell renewal is very slow in the proximal tubule (LeBlond 1972), a shedding mechanism was suggested. Ultrastructural observation on mammalian kidneys are in line with this assumption (Misch et al. 1980). Observations suggesting the occurrence of plasma-membrane shedding in vivo have also been made in the human liver (Cossel 1980), in the venom gland of the rattle snake (Warshawsky et al. 1973), in peritoneal wash- ings of the guinea pig (Dvorak et al. 1981), and in the thymus of rabbits (Roozemond and Urli 1981) and mice (Van Blitterswijk et al. 1977). Perhaps these in vivo observations in the thymus correspond to the spontaneous or antigen- induced shedding of plasma-membrane fragments known to occur in cultured thymocytes and lymphoctes of various species (Vitetta and Uhr 1972; Esselman and Miller 1977; Freimuth et al. 1978; Emerson and Cone 1981, 1982). These studies have revealed a number of lymphocyte-specific antigens, including immunoglobulins, as integral membrane components in these vesicles obtained in vitro. The vesicles have been suggested to have a possible role in vivo as lym- phokines (Esselman and Miller 1977; Emerson and Cone 1981, 1982). Recent findings of mononuclear leukocyte-derived vesicles promoting erythroid-stem cell proliferation in culture (Dainiak and Cohen 1982) further support this hypothesis. Another example where shedding of plasma- membrane-derived vesicles might have a physiological role in vivo is from studies of bone formation. Chondrocytes are believed to shed matrix vesicles from their cellular processes and the vesicles then serve as initial sites for hydroxy- apatite deposition (Bonucci 1970; Rabinovich and Ander- son 1976). The ability of viable chondrocytes in culture to shed Ca + +-binding vesicles has recently been observed (Golub et al. 1981 ; Shapiro et al. 1981). Vesicles also seem to be of importance in ectopic calcification. As mentioned above, the normal proximal tubule cell of the kidney probably has the ability of plasma-membrane shedding and this mechanism might be involved in the pathogenesis of neph- rocalcinosis in the rat (Nguyen and Woodard 1980). The well-known calcification of the aortic valve and aortic media might similarly be initiated, at least partly, by vesicles, presumably derived from degenerating fibroblasts and

smooth muscle cells (Kim 1976; Tanimura et al. 1983). This is further supported by Trams et al. (1981), who have described the capacity of plasma-membrane shedding in cul- tures of many different cells, amongst them cells from rat aorta. Other examples where matrix vesicles seem to play a role in calcification are human pulmonary alveolar micro- lithiasis (Bab etal. 1981), psammoma bodies of meningioma (Lipper et al. 1979), and experimental cutaneous calcinosis (Boivin 1975). Plasma-membrane shedding has also been observed in other pathological conditions. The alkaline phosphatase activity in serum of patients with cholestasis is bound to membrane vesicles, presumably derived from the hepatocyte plasma membrane (DeBroe et al. 1975), and perhaps this is an exaggerated form of the normal conditions in the liver (Cossel 1980). In vitro studies of erythrocytes (Reed and Swisher 1966; Weed and Bowdler 1966) have suggested that the pronounced hemolytic ten- dency in hereditary spherocytosis is due to fragmentation of the erythrocyte plasma membrane, rendering the cell more fragile. Studies on the infection process of Entamoeba histolytica (Takeuchi and Phillips 1976) and Myeoplasma hyorhinis (Stanbridge and Weiss 1978) indicate that micro- organisms might induce plasma-membrane shedding in infected cells.

Plasma-membrane shedding has also been suggested to be a mechanism of defence. For example, a parasite, Leishmania enriettii, sheds parts of its envelope subsequent to antibody exposition in vitro (Doyle et al. 1974), which could be interpreted as an escape mechanism. This possible function is also indicated from studies on tumor cells, where shedding of plasma-membrane vesicles is a common observation. It is most often recognized in malignant cells of the lymphoid system, both in vitro in the culture medium (Van Blitterswijk et al. 1977; Liepins et al. 1978; Raz et al. 1978a; Sachs et al. 1980) and in vivo, in samples of serum, pleural effusion and ascites fluid (Skinnider and Ghadially 1977; Petitu et al. 1978 ; Raz et al. 1978 b; Van Blitterswijk et al. 1979). Neoplastic epithelial cells in vitro (Koch and Smith 1978 ; Higgins et al. 1980) and in vivo (Nowotny et al. 1974; Dvorak et al. 1981), as well as malignant cells of other origins (Poskitt et al. 1976; Rittenhouse et al. 1978; Trams et al. 1981) also possess this property. T-lymphocytes (Biberfeld et al. 1973; Liepins et al. 1978) have been shown to induce the shedding process, and the vesicles then specifically adhere to the inductors, preventing them from interacting with the tumor cells (Liepins et al. 1978). How- ever, this is not the only functional role tumor-derived vesicles could play in malignancy. For example, such vesicles could be the causing agent of the hemostatic defects often occurring, possessing procoagulant (Dvorak et al. 1981) or platelet-aggregating activity (Gasic and Gasic 1982).

It should also be mentioned that various chemical and physical agents are potent inductors of plasma-membrane shedding, such as low-molecular weight aldehydes and di- sulfid-blocking agents (Scott 1976), cytochalasins (Godman etal. 1975; Liepins and Hillman 1981), vinca alkaloids (Krishan and Frei 1975; Liepins and Hillman 1981), some phospholipids (Billington and Coleman 1978; Ferber et al. 1980; Ott et al. 1981), bile salts (Holdsworth and Coleman 1976) and temperature shifts, e.g., cooling with subsequent warming (Liepins and Hillman 1981). Also ATP-depletion (Lutz et al. 1977) and increased Ca + + concentration (Allan et al. 1976), obtained by ionophore A23187, cause cultured erythrocytes to shed plasma-membrane vesicles.

96

Concluding remarks

P l a s m a - m e m b r a n e shedding is a p h e n o m e n o n observed in v ivo and in v i t ro in b o t h n o r m a l and pa tho log ica l cells o f va r ious origins. The m e c h a n i s m is suggested to have a func t iona l role in m a n y cases. The repor ts concern ing p l a s m a - m e m b r a n e shedding, especial ly in vivo, are ra ther rare c o m p a r e d to o the r aspects o f p l a s m a - m e m b r a n e turnover. O n e reason for this migh t be tha t the shed m e m b r a n e rapid ly d i sappears in the ext racel lu lar env i ronment , thus m a k i n g it diff icult to recognize. The m o r p h o l o g y o f the thy ro id favors the obse rva t ion o f p l a s m a - m e m b r a n e shed- d ing as the epi thel ia l cells face a closed lumen, in which m e m b r a n e c o m p o n e n t s are cap tured . Our present observa- t ions in the mouse thyro id indicate tha t p l a s m a - m e m b r a n e shedding occurs in the n o r m a l follicle and tha t the a m o u n t o f shed mate r ia l in the fol l icular l umen varies wi th the func- t ional s tate o f the follicle cell. This m e c h a n i s m might be o f i m p o r t a n c e in the ma in t enance o f a cons tan t surface area o f the apical p l a sma membrane .

References

Allan D, Billah MM, Finean lB, Michell RH (1976) Release of diacylglycerol-enriched vesicles from erythrocytes with increased intracellular (Ca 2+). Nature 26I:58 60

Bab I, Rosenmann E, N6eman Z, Sela J (1981) The occurrence of extracellular matrix vesicles in pulmonary alveolar microlith- iasis. Virchows Arch [Pathol Anat] 391 : 357-361

Biberfeld P, Biberfeld G, Perlmann P, Holm G (1973) Cytological observations on the cytotoxic interaction between lymphocytes and antibody-coated monolayer cells. Cell Immunol 7:60-72

Billington D, Coleman R (1978) Effects of bile salts on human erythrocytes. Plasma membrane vesiculation, phospholipid so- lubilization and their possible relationships to bile secretion. Biochim Biophys Acta 509:33-47

Bj6rkman U, Ekholm R, Elmqvist L-G, Ericson LE, Melander A, Smeds S (1974) Induced unidirectional transport of protein into the thyroid follicular lumen. Endocrinology 95:1506-1517

Bj6rkman U, Ekholm R, Ericson LE, Ofverholm T (1976) Trans- port of thyroglobulin and peroxidase in the thyroid follicle cell. Mol Cell Endocrinol 5:3 17

Black PH (1980) Shedding from the surface of normal and cancer cells. Adv Cancer Res 32:75-199

Black BL, Yoneyama Y, Moog F (1980) Microvillous membrane vesicle accumulation in media during culture of intestine of chick embryo. Biochim Biophys Acta 601:343-348

Boivin PG (1975) Cutaneous calcinosis induced by topical calciphy- laxis in the rat. 1. Ultrastructural aspects. Arch Anat Microsc 64: 183-205

Bonucci E (1970) Fine structure and histochemistry of "calcifying globules" in epiphyseal cartilage. Z Zellforsch 103:192-217

Cone RE (1977) Dynamic aspects of the lymphocyte surface. In: Marchalonis II (ed) The lymphocyte: structure and function vol 2. Dekker, NY, pp 565-592

Cossel L (1980) Peripheral cytoplasmic shedding around hepato- cytes-apocrine secretion, pathological cell reaction or prepara- tion artefact? Pathol Res Pract 170:298 327

Dainiak N, Cohen CM (1982) Surface membrane vesicles from mononuclear cells stimulate erythroid stem cells to proliferate in culture. Blood 60:583 594

DeBroe ME, Borgers M, Wieme RJ (1975) The separation and characterization of liver plasma membrane fragments circulat- ing in the blood of patients with cholestasis. Clin Chim Acta 59 : 369-372

DeBroe ME, Wieme R J, Logghe GN, Roels F (1977) Spontaneous shedding of plasma membrane fragments by human cells in vivo and in vitro. Clin Chim Acta 81 : 237 245

DeRitis G, Falchuk ZM, Trier JS (1975) Differentiation and matu- ration of cultured fetal rat jejunum. Dev Biol 45 : 304-317

Doljanski F, Kapeller M (1976) Cell surface shedding - the phenomenon and its possible significance. J Theor Biol 62 : 253-270

Doyle J J, Behin R, Manuel J, Rowe DS (1974) Antibody-induced movement of membrane components of Leishmania enriettii. J Exp Med 139:1061-1069

Dvorak HF, Quay SC, Orenstein NS, Dvorak AM, Hahn P, Bitzer AM (1981) Tumor shedding and coagulation. Science 212: 923-924

Ekholm R (1981) Iodination of thyroglobulin - an intracellular or extracellular process ? Mol Cel Endocrinol 24:141-163

Ekholm R, Wollman SH (1975) Site of iodination in rat thyroid gland deduced from electron microscopic autoradiographs. En- docrinology 97:1432-1444

Ekholm R, Engstr6m G, Ericson LE, Melander A (1975) Exocyto- sis of protein into thyroid follicle lumen: an early effect of TSH. Endocrinology 97 : 337-346

Emerson SG, Cone RE (1981) I-K k and H-2K k antigens are shed as supramolecular particles in association with membrane lipids. J Immunol 127:482-486

Emerson SG, Cone RE (1982) Absorption of shed I-A k and H-2K k antigen by lymphoid cells. Transplantation 33 : 36-40

Ericson LE (1981) Exocytosis and endocytosis in the thyroid follicle cell. Mol Cell Endocrinol 22: 1 24

Ericson LE, Johanson V (1981) Turnover of exocytotic vesicles in rat thyroid follicle cells. Ann d'endocrinologie A. Abstr 1 l th Annual Meeting of the European Thyroid Association, Pisa p 77

Esselman WJ, Miller HC (1977) Modulation of B cell responses by glycolipid released from antigen-stimulated T cells. J Im- munol 119:1994-2000

Ferber E, Schmidt B, Weltzien HU (1980) Spontaneous and deter- gent-induced vesiculation of thymocyte plasma membranes. Biochim Biophys Acta 595:244-256

Freimuth WW, Esselman W J, Miller HC (1978) Release of Thy-1.2 and Thy-l.1 from lymphoblastoid cells: partial characterization and antigenicity of shed material. J Immunol 120:1651-1658

Gasic G J, Gasic TB (1982) Plasma membrane vesicles as mediators of interactions between tumor cells and components of the hemostatic and immune systems. In: Interaction of platelets and tumor cells. Alan R Liss Inc, New York, pp 429-444

Godman GC, Miranda AF, Deitcb AD, Tanenbaum SW (1975) Action of cytochalasin D on cells of established lines. III. Zeio- sis and movements at the cell surface. J Cell Biol 64:644-667

Golub EE, Schattschneider S, McArthur WP, Burke A, Shapiro IM (1981) Induction of vesiculation by cultured choudrocytes. In: Matrix vesicles. Proceedings of the Third International Con- ference on Matrix Vesicles. Mondeluco, Spoleto, pp 41-45

Holdsworth G, Coleman R (1976) Plasma membrane components can be removed from isolated lymphocytes by the bile salts glycocholate and taurocholate without cell lysis. Biochem J 158 : 493-495

Huggins JW, Trenbeath TP, Yeitman DR, Carraway KL (1980) Restricted concanavalin A redistribution on the branched microvilli of an ascites tumor subline. Exp Cell Res 127:31-46

Jacobs LR (1981) Alterations in surface ultrastructure and anionic sites of rat dimethylhydrazine-induced intestinal tumors. Vir- chows Arch [Cell Pathol] 37:207-216

Kim KM (1976) Calcification of matrix vesicles in human aortic valve and aortic media. Fed Proc 35:156-162

Koch GLE, Smith MJ (1978) An association between actin and the major histoeompatibility antigen H-2. Nature 273:274-278

Krishan A, Frei E (1975) Morphological basis for the cytolytic effect of vinblastine and vincristine on cultured human leukemic lymphoblasts. Cancer Res 35:497-501

LeBlond CP (1972) Growth and renewal. In: Goss RJ (ed) Regula- tion of organ and tissue growth. Academic Press, New York, pp 13-39

Liepins A, Hillman AJ (1981) Shedding of tumor cell surface membranes. Cell Biol Int Rep 5:15-25

97

Liepins A, Faanes RB, Choi YS, de Harven E (1978) T-lymphocyte mediated lysis of tumor ceils in the presence of alloantiserum. Cell Immunol 36:331-344

Lipper S, Dalzell JC, Watkins PJ (1979) Ultrastructure of psam- morea bodies of meningioma in tissue culture. Arch Pathol Lab Med 103 : 670-675

Lutz HU, Liu S-C, Palek J (1977) Release of spectrin-free vesicles from human erythrocytes during ATP depletion. J Cell Biol 73 : 548-560

Marcus PI (1962) Dynamics of surface modification in myxovirus- infected cells. Cold Spring Harbor Syrup on Quant Biol 27:351-365

Misch DW, Giebel PE, Faust RG (1980) Intestinal microvilli: responses to feeding and fasting. Eur J Cell Biol 21 : 269-279

Nguyen HT, Woodard JC (1980) Intranephronic calculosis in rats. Am J Pathol 100:39-56

Nowotny A, Grohsman J, Abdelnoor A, Rote N, Yang C, Watters- dorff R (1974) Escape of TA3-tumors from allogenic immune rejection: theory and experiments. Eur J Immunol 4:73-78

Nunez EA, Wallis J, Gershon MD (1974) Secretory processes in follicular cells of the bat thyroid. III. The occurrence of extracellular vesicles and colloid droplets during arousal from hiber- nation. Am J Anat 141:179-202

Ott P, Hope MJ, Verkleij AJ, Roelofsen B, Brodbeck U, Van Deenen LLM (1981) Effect of dimyristoyl phosphatidylcholine on intact erythrocytes. Release of spectrin-free vesicles without ATP depletion. Biochim Biophys Acta 641 : 79-87

Petitu M, Tuy F, Rosenfeld C, Mishal Z, Paintrand M, Jasnin C, Mathe G, Inbar M (1978) Decreased microviscosity of membrane lipids in leukemic cells: two possible mechanisms. Proc Natl Acad Sci USA 75:2306-2310

Poskitt PKF, Poskitt TR, Wallace JH (1976) Release into culture medium of membrane-associated tumor-specific antigen by B- 16 melanoma cells (39332). Proc Soc Exp Biol Med 152:76-80

Rabinovitch AL, Anderson HC (1976) Biogenesis of matrix vesicles in cartilage growth plates. Fed Proc 35 : 112-116

Raz A, Goldman R, Yuli I, Inbar M (1978a) Isolation of plasma membrane fragments and vesicles from ascites fluid of lymphoma-bearing mice and their possible role in the escape mechanism of tumors from host immune rejection. Cancer Immunol Immunother 4: 53-59

Raz A, Barzilai R, Spira G, Inbar M (1978b) Oncogenicity and immunogenicity associated with membranes isolated from cell- free ascites fluid of lymphoma-bearing mice. Cancer Res 38 : 248(~2485

Reed CF, Swisher SN (1966) Erythrocyte lipid loss in hereditary spherocytosis. J Clin Invest 45:777-781

Rittenhouse HG, Rittenhouse JW, Takemoto L (1978) Character- ization of the cell coat of Ehrlich ascites tumor cells. Biochemis- try 17:829-837

Roozemond RC, Urli DC (1981) Fluorescence polarization studies and biological properties of membranes exfoliated from the cell surface of rabbit thymocytes in situ. Biochim Biophys Acta 643 : 327-338

Sachs DH, Kiszkiss P, Kim KJ (1980) Release of Ia-antigens by a cultured B cell line. J Immunol 124:2130-2136

Scott RE (1976) Plasma membrane vesiculation: A new technique for isolation of plasma membrane. Science 194:743-745

Shapiro IM, Burke A, Schattschneider S, Golub EE (1981) Vesicu- lation of chondrocytes in vitro: a new technique for isolating vesicles from chondrocytes. In: Matrix vesicles. Proceedings of the Third International Conference on Matrix Vesicles, Mon- teluco, Spoleto, pp 33-39

Skinnider LF, Ghadially FN (1977) Ultrastructure of cell surface abnormalities in neoplastic histiocytes. Br J Cancer 35 : 657-667

Stanbridge EJ, Weiss RL (1978) Mycoplasma capping on lymphocytes. Nature 276:583-587

Tachiwaki O, Wollman SH (1982) Shedding of dense cell fragments into the follicular lumen early in involution of the hyperplastic thyroid gland. Lab Invest 47:91-98

Takeuchi A, Phillips BP (1976) Electron microscope studies of experimental Entamoeba histolytica infection in the guinea pig. Virchows Arch B [Cell Path] 20:1-13

Tanimura A, McGregor DH, Anderson HC (1983) Matrix vesicles in atherosclerotic calcification. Proc Soc Exp Biol Med 172:173-177

Trams EG, Lauter CJ, Salem N, Heine U (1981) Exfoliation of membrane ecto-enzymes in the form of micro-vesicles. Biochim Biophys Acta 645 : 63-70

Van Blitterswijk WJ, Emmelot P, Hilkmann HAM, Oomen-Meule- mans EPM, Inbar M (1977) Differences in lipid fluidity among isolated plasma membranes of normal and leukemic lymphocytes and membranes exfoliated from their surfaces. Biochim Biophys Acta 467:309-320

Van Blitterswijk WJ, Emmelot P, Hilkmann HAM, Hilgers J, Felt- kamp CA (1979) Rigid plasma membrane-derived vesicles, enriched in tumour-associated surface antigens (MLr), occurring in the ascites fluid of a murine leukaemia (GRSL). Int J Cancer 23 : 62-70

Vitetta ES, Uhr JW (1972) Cell surface immunoglobulin. V. Re- lease from murine splenic lymphocytes. J Exp Med 136 : 676-696

Vitetta ES, Uhr JW, Boyse EA (1974) Metabolism of H-2 and Thy-1 alloantigens in murine thymocytes. Eur J Immunol 4:276-282

Warshawsky H, Haddad A, Gongalves RP, Vareri V, DeLucca FL (1973) Fine structure of the venom gland epithelium of the South American rattlesnake and radioautographic studies of protein formation by the secretory cells. Am J Anat 138:79-120

Weed RI, Bowdler AJ (1966) Metabolic dependence of the critical hemolytic volume of human erythrocytes : Relationship to os- motic fragility and autohemolysis in hereditary spherocytosis and normal red cells. J Clin Invest 7:1137-1149

Wollman SH, Breitman TR (1970) Changes in D N A and weight of thyroid glands during hyperplasia and involution. Endocri- nology 86: 322-327

Wollman SH, Loewenstein JE (1973) Rates of colloid droplet and apical vesicle production and membrane turnover during thyroglobulin secretion and resorption. Endocrinology 93 : 248-252

Young RW (1976) Visual cells and the concept of renewal. Invest Ophthalmol 15: 700-725

Accepted December 16, 1983

Documents

In vivo shedding of apical plasma membrane in the thyroid follicle cells of the mouse