The process of right junction degradation was followed in a cell line of human colon adenocarcinoma.

Tight junctions are degraded by 2 mechanisms: (1) breakdown of junctional elements to intramembrane particles; (2) bleb formation by which tight-junctional elements are internalized into the cytoplasm or excluded into the medium. It is suggested that the first mechanism allows preservation of membrane particles for re-use, whereas the second is a mechanism by which the cells eliminate unneeded junctional elements.

Tight junctions are the outermost elements of the junctional complex in mammalian epithelia (Staehelin, 1974). By freeze-fracturing it was shown that such junctions are composed of a continuous meshwork of branching and anastomosing fibrils shared by the plasma membrane of 2 adjacent cells. This structure contributes to the tight junctions’ functions as cell contacts and as sealing elements (Wade & Karnovsky, 1974 a).

Studies on intact tissues have revealed that tight junctions are not disrupted by trypsin (Amsterdam & Jamieson, 1974). Moreover, it was found that proteolytic enzymes sometimes induce extensive development of tight-junctional elements (Orel et al. 1973; Shimono & Clementi, 1977).

Recently we have shown that HT29 cells, a cell line of human colon adenocarcinoma, which lack tight junctions when grown to confluency, developed such junctions very rapidly if treated with trypsin (Polak-Charcon, Friedberg, Shoham & Ben-Shaul, 1976; Polak-Charcon, Shoham & Ben-Shaul, 1978). The events involved in the assembly of the junctional elements were found to be similar in general to those described for other systems (Montesano, Friend, Perrelet & Orci, 1975; Wanson, Drochmans, Mosselmans & Ronveaux, 1977). However, particle fusion was preceded by the elevation of particle-free membrane regions.

We have also reported that in cells treated with trypsin and then transferred back into fresh medium, no tight junctions were observed. Since vesicles containing tight-junctional elements were found within the cytoplasms of these cells, it was suggested that an endocytosis-like mechanism is at least one process involved in the disassembly of these junctions (Polak-Charcon et al. 1978). Lysosome-like vesicles containing elements of such occluding junctions have been detected in intestinal epithelia and were assumed to be correlated with tight-junction degradation (Staehelin, 1973, 1974). However, no detailed information on such a mechanism is available.

As previously found in the study of tight-junction assembly, HT29 cells are also found to be suitable systems to follow tight-junction degradation. We have used freeze-fracturing to demonstrate stages in this disassembly process.

HT29 cells were grown in Dulbeco’s modified Eagle medium supplemented with 10% foetal calf serum. Stock cultures were grown in dishes (Falcon) and subcultured every 4 days. Cells were kept at 37 °C in a humidified atmosphere containing 5 % CO, in air.

Trypsin, twice-crystalized (Sigma), was used at a concentration of 0 · 25 % in Ca- and Mg-free phosphate buffer saline (PBS). Trypsin was added for 15 min to cells grown to confluency. Cells were then washed and transferred to Petri dishes with fresh medium and 10% foetal calf serum. After 1, 2, 3 and 6 h in the fresh medium the cells were fixed for 10 min with 5 % glutaraldehyde in 0·1 M cacodylate buffer, pH 7 · 4 at 37 °C. Then the cells were collected by gentle scraping and fixed for an additional period of 1 h. After several washings the cells were suspended overnight in 30 % glycerol in cacodylate buffer at 4 °C. Samples were frozen in freon 22, immediately transferred to liquid nitrogen and fractured in a Balzer’s freezeetching unit according to the standard procedure (Moor & Mühlethaler, 1963). Washed replicas were mounted on 300-mesh uncoated copper grids and photographed in a Jeol-iooB electron microscope.

Freeze-fractured HT29 cells grown to confluency and removed mechanically from the dish, show no membrane specializations. However, when treated for 15 min with trypsin, large areas of anastomosing ridges and complementary furrows were observed on their PF and EF faces respectively (Fig. 1).

Fig. 1.

Freeze-fractured plasma membrane of HT29 cells treated with trypsin for 15 min. Note well-developed network of tight junctional elements. Ridges (r) on PF, furrows (f) on EF face, × 39600. Circled arrow in the lower right comer indicates the angle of shadowing in all figures. EF and PF, EF and PF faces respectively.

Fig. 1.

Freeze-fractured plasma membrane of HT29 cells treated with trypsin for 15 min. Note well-developed network of tight junctional elements. Ridges (r) on PF, furrows (f) on EF face, × 39600. Circled arrow in the lower right comer indicates the angle of shadowing in all figures. EF and PF, EF and PF faces respectively.

By optical microscopy it was observed that cells treated with trypsin for 15 min and then washed and transferred to fresh medium for 1 h, are either single or in small clusters. The number of cells within a cluster and the number of clusters decreased progressively with time in fresh medium. After 6 h most of the cells were single, rounded and attached to the substrate.

The appearance of the tight junctions after 1 h in medium, as revealed by freezefracturing, was unaltered in some cells, but in most of them striking changes were observed. These changes were expressed in 2 ways: (a) The mesh work of anastomosing tight-junction elements ‘opened’ and many of the ridges were broken into arrays of particles (Figs. 2, 3). It appears that the opening of the network and the disruption of ridges into separate particles is part of a mechanism leading to the redistribution of tight-junction components into the surrounding membrane. Often very small straight segments, small ‘looped’ structures or a few particles still arranged as a line on elevated membrane crests were observed (Figs. 2, 3 arrows). (6) Groups of still anastomosing elements were separated from the complexed network and concentrated in defined areas of the membrane, which occasionally were seen to be somewhat elevated (Fig. 4). This appears to be an early stage of bleb formation by which tight-junctional elements are internalized into the cytoplasm or excluded into the medium. A most striking stage in this process was observed after 1–4 h in fresh medium; large areas of the plasma membrane were observed to have ‘blebbed’ or formed vesicular structures carrying within them tight-junction elements (Fig. 5). Some of these structures were oriented toward the cytoplasm (Fig. 6), whereas others were protruding out from the cell surface (Fig. 7). It seems that the protruding bleb-like structures are shed into the surrounding medium, whereas the vesicle-like structures oriented towards the cytoplasms are involved in endocytosis-like processes which result in internalization of degraded elements. Indeed, already after 1 h in fresh medium many cells contained within their cytoplasm vesicles with tight-junction elements (Fig. 8, arrows). These elements covered either most of the vesicle membrane (Fig. 9 A) or only a limited part of it (Fig. 9B). The limiting membrane of the vesicle contained on its PF or EF face particles in an arrangement which is typical of plasma membranes (Figs. 9A, B). Lysosome-like structures containing other membrane fragments, in addition to junctional elements, were also observed (Figs. 9 c, 10).

Fig. 2, 3.

Freeze-fractured plasma membrane of HT29 cells treated with trypsin for 15 min and then transferred to fresh medium for 1 h. Fig. 2. Open network of tight junctional elements. Note looped (f). circular (c), small (r) ridges and a chain of anastomosing particles on an elevated membrane region (arrows), × 39600.

Fig. 2, 3.

Freeze-fractured plasma membrane of HT29 cells treated with trypsin for 15 min and then transferred to fresh medium for 1 h. Fig. 2. Open network of tight junctional elements. Note looped (f). circular (c), small (r) ridges and a chain of anastomosing particles on an elevated membrane region (arrows), × 39600.

Fig. 3.

Short strands of tight junction on PF face. Note lines of particles on elevated membrane crest (arrows), × 39600.

Fig. 3.

Short strands of tight junction on PF face. Note lines of particles on elevated membrane crest (arrows), × 39600.

Fig. 4, 5.

Freeze-fractured plasma membrane of HT29 cells treated with trypsin for 15 min and then transferred to fresh medium for 1 h. Fig. 4. Small groups of anastomosing ridges (a) on PF faces, × 33000.

Fig. 4, 5.

Freeze-fractured plasma membrane of HT29 cells treated with trypsin for 15 min and then transferred to fresh medium for 1 h. Fig. 4. Small groups of anastomosing ridges (a) on PF faces, × 33000.

Fig. 5.

Bleb-like or vesicular structures carrying tight junctional elements. Note well developed furrows (f) on EF face and ridges (r) on PF face, × 33000.

Fig. 5.

Bleb-like or vesicular structures carrying tight junctional elements. Note well developed furrows (f) on EF face and ridges (r) on PF face, × 33000.

Fig. 6-8.

Freeze-fractured HT29 cells treated with trypsin for 15 min and then transferred to fresh medium for 1 h. Fig. 6. A vesicular-like structure of the plasma membrane oriented toward the cytoplasm (cy), and a bleb-like structure protruding out (0). Note tight-junctional elements. × 63 000.

Fig. 6-8.

Freeze-fractured HT29 cells treated with trypsin for 15 min and then transferred to fresh medium for 1 h. Fig. 6. A vesicular-like structure of the plasma membrane oriented toward the cytoplasm (cy), and a bleb-like structure protruding out (0). Note tight-junctional elements. × 63 000.

Fig. 7.

A bleb-like structure of the plasma membrane protruding out of (0) the cell surface. Note tight-junctional elements, × 45 000.

Fig. 7.

A bleb-like structure of the plasma membrane protruding out of (0) the cell surface. Note tight-junctional elements, × 45 000.

Fig. 8.

Phagocytic-like vesicles (arrows) bearing remnant of tight-junctional elements in the cytoplasm (cy) of HT29 cells, × 27600.

Fig. 8.

Phagocytic-like vesicles (arrows) bearing remnant of tight-junctional elements in the cytoplasm (cy) of HT29 cells, × 27600.

Fig. 9.

Freeze-fractured vesicles in the cytoplasm of HT29 cells treated with trypsin for 15 min and transferred to fresh medium for 1 h. A. Phagocytic-like vesicle covered with tight junctional elements. Note ridges (r) on its PF face, × 45 000. B. Phagocytic-like vesicle carrying few remnants of tight-junctional elements. Note furrows (f) on its EF face, × 45 000. C. Lysosome-like structure containing degraded membranal fragments. Note remnant of tight-junctional elements (arrow) on a small vesicle within the lysosome, × 72 600.

Fig. 9.

Freeze-fractured vesicles in the cytoplasm of HT29 cells treated with trypsin for 15 min and transferred to fresh medium for 1 h. A. Phagocytic-like vesicle covered with tight junctional elements. Note ridges (r) on its PF face, × 45 000. B. Phagocytic-like vesicle carrying few remnants of tight-junctional elements. Note furrows (f) on its EF face, × 45 000. C. Lysosome-like structure containing degraded membranal fragments. Note remnant of tight-junctional elements (arrow) on a small vesicle within the lysosome, × 72 600.

Fractured faces of HT29 cell membranes after 6 h in fresh medium were found to be free of any junctional structure and look very much like plasma membranes of cells which were not induced to form tight junctions by trypsinization (Fig. 10). The cytoplasm of these cells, however, contained many vesicles with remnants of junctional elements.

Fig. 10.

Freeze-fractured plasma membrane of HT29 cells treated with trypsin for 15 min and transferred to fresh medium for 6 h. Note absence of tight-junctional elements in PF and EF faces, mu, cross-sectioned microvilli, × 33000.

Fig. 10.

Freeze-fractured plasma membrane of HT29 cells treated with trypsin for 15 min and transferred to fresh medium for 6 h. Note absence of tight-junctional elements in PF and EF faces, mu, cross-sectioned microvilli, × 33000.

DISCUSSION

It was previously established that HT29 cells grown to confluency as monolayers do not have tight or gap junctions (Polak-Charcon et al. 1976). When treated with trypsin, however, these membrane specializations develop rapidly, especially tight junctions.

In this study we have followed the breakdown of tight junctions after washing out the trypsin and transferring the cells to fresh medium. The data obtained suggest that 2 coexisting mechanisms may be involved in the process of degradation.

In the first presumed mechanism the order of events seems to be as follows: first, the large meshwork of junctional elements is opened. This is followed by separation and displacement of smaller groups of still anastomosing elements. These are then broken to shorter segments which eventually are degraded to single particles. Breakdown of tight-junctional elements into short disconnected segments has been previously reported for osmotically disrupted zonula occludens of toad urinary bladder (Wade & Karnovsky 19746), in pancreatic acinar cells isolated by collagenase, hyaluronidase EDTA and mild shearing forces (Amsterdam & Jamieson, 1974) or by EGTA (Galli, Brenner, De Camilli & Meldolesi, 1976). Disintegration of tight junctions was also found to occur in rapidly growing foetal liver (Montesano et al. 1975) and in early chick embryos (Revel, Yip & Chang, 1973). The appearance of non-particulated crest-like areas on which the single particles were observed at the presumably last stage of tight junction degradation, is similar to the early stage of tight-junction assembly (Polak-Charcon et al. 1978). It seems that this process of tight-junction degradation is the reverse of tight-junction assembly. The crest-like elevated structures were assumed to be more rigid lipid regions of the plasma membrane and to be involved in the initial contact between the adhering cells (Polak-Charcon et al. 1978). It appears that these crests are also the last to hold the cells together while recovering from the effects of trypsin. It has been shown that fusion of membranes occurs in areas devoid of particles (Lucy, 1975; Zakai, Kulka & Loyter, 1977; Lawson et al. 1977). Short particle rows in continuation with fibril segments were also observed in disrupted occludens elements of guinea pig pancreas lobules incubated with Ca-free medium and EGTA (Galli et al. 1976).

In the second presumed mechanism the initial stage is also splitting of anastomosing strands. However, the disrupted groups of branching elements remain compact and are found concentrated in certain regions of the plasma membrane. These regions in a later stage form bleb-like structures which carry remnants of tight junctions. Such structures were observed in thin sections of isolated pancreatic exocrine cells (Amsterdam & Jamieson, 1972). The presence within the cytoplasm of recovered HT29 cells of many endocytic-like vesicles containing broken elements of tight junctions indicates that these blebs became internalized into the cells; this does not exclude the possibility that some of the blebs are shed into the surrounding medium. The occurrence of both ridges and furrows on the ‘blebs’ membranes or in the endocytic-like vesicles, indicates that when cells are detached from each other, tight-junction elements are not split into halves, as is the case with desmosomes. Lysosomes, containing elements of tight junctions, were observed in normal intestinal epithelial cells in vivo (Staehelin, 1973), but were not observed in a vitamin A system which induces mucose metaplasia (Elias & Friend, 1976). Endocytosis is known to be the process by which non-functional dissociated desmosomes are eliminated (Overton, 1968). A similar mechanism was also proposed for gap junctions (Albertini, Fawcett & Olds, 1975; Coons & Espay, 1977).

Summarizing, it seems that 2 mechanisms are involved in the degradation of tight junctions; the first is the breakdown of these occluding elements to single particles; the second, elimination of such elements by formation of blebs, which are either internalized or possibly shed off. Since similar mechanisms have been suggested for different experimental systems we do not think that the mechanisms described are limited only to cancer cells or even to intestinal epithelial cells. It is not known whether these 2 mechanisms are independent, or whether there is a master-mechanism which regulates the two (or more) sub-mechanisms. It is reasonable to assume that the 2 mechanisms coexist (a) to ensure that sufficient ‘junction particles’ will be left in the membrane to be re-used by the cells in case of re-induced cell attachment and junction formation without immediate protein synthesis, and (b) to eliminate unnecessary junctional elements by endocytosis or shedding. No data are yet available to distinguish between intramembrane particles which may become part of a tight (or gap) junction and other membrane particles, if indeed such a distinction exists.

Albertini
,
D. F.
,
Fawcett
,
D. W.
&
Olds
,
P. J.
(
1975
).
Morphological variations in gap junction of ovarian granulosa cells
.
Tissue & Cell
7
,
389
402
.
Amsterdam
,
A.
&
Jamieson
,
J. D.
(
1972
).
Structural and functional characterization of isolated pancreatic exocrine cells
.
Proc. natn. Acad. Sci. U. S. A
.
69
,
3028
3032
.
Amsterdam
,
A.
&
Jamieson
,
J. D.
(
1974
).
Studies on dispersed pancreatic exocrine cells, i. Dissociation technique and morphologic characteristics of separated cells
.
J. Cell Biol
.
63
,
1037
1057
.
Coons
,
L. W.
&
Espey
,
L. L.
(
1977
).
Quantitation of nexuses junctions in the granulosa cell layer of rabbit ovarian follicules during ovulation
.
J. Cell Biol
.
74
,
321
325
.
Elias
,
P. M.
&
Friend
,
D. S.
(
1976
).
An in vitro system for modulating tight and gap junction differentiation
.
J. Cell Biol
.
68
,
173
188
.
Galli
,
P.
,
Brenna
,
A.
,
De Camilli
,
P.
&
Meldolesi
,
J.
(
1976
).
Extracellular calcium and the organization of tight junctions in pancreatic acinar cells
.
Expl Cell Res
.
99
,
178
182
.
Lawson
,
D.
,
Raff
,
M. C.
,
Gomperts
,
B.
,
Fewtrell
,
C.
&
Gilula
,
N. B.
(
1977
).
Molecular events during
membrane fusions
:
A study of exocytosis in rat peritoneal Mast cells
.
J. Cell Biol
.
72
,
242
259
.
Lucy
,
J. A.
(
1975
).
Aspects of the fusion of cells in vitro without viruses
.
J. Reprod. Fert
.
44
,
193
205
.
Montesano
,
R.
,
Friend
,
D.
,
Perrelet
,
A.
&
Orci
,
L.
(
1975
).
In vivo assembly of tight junctions in fetal rat liver
.
J. Cell Biol
.
67
,
310
319
.
Moor
,
H.
&
Mühlethaler
,
K.
(
1963
).
Fine structure in frozen-etched yeast cells. J
.
Cell Biol
.
17
,
609
628
.
Orci
,
L.
,
Amherdt
,
M.
,
Henquin
,
J. C.
,
Lambert
,
A. E.
,
Unger
,
R. H.
&
Renold
,
A. E.
(
1973
).
Pronase effect on pancreatic Beta cell secretion and morphology
.
Science, N. Y
.
180
,
647
649
.
Overton
,
J.
(
1968
).
The fate of desmosomes in trypsinized tissue. J
.
exp. Zool
.
168
,
203
214
.
Polak-Charcon
,
S.
,
Friedberg
,
L
,
Shoham
,
J.
&
Ben-Shaul
,
Y.
(
1976
).
Effect of trypsin on HT29 an adenocarcinoma of the human colon cell line
.
In Electron Microscopy ‘76, Proc. 6th EUT. Congr. E.M
., vol.
11
(ed.
Y.
Ben-Shaul
), pp.
356
358
. TAL International Publishers.
Polak-Charcon
,
S.
,
Shoham
,
J.
&
Ben-Shaul
,
Y.
(
1978
).
Junctions formation in trypsinized cells of human adenocarcinoma cell line
.
Expl Cell Res
. (in press).
Revel
,
J. P.
,
Yip
,
P.
&
Chang
,
L. L.
(
1973
).
Cell junctions in the early chick embryo. A freeze-etch study
.
Devi Biol
.
35
,
302
317
.
Shimono
,
M.
&
Clementi
,
F.
(
1977
).
Intercellular junction of oral epithelium. II. ultrastructural changes in rat buccal epithelium induced by trypsin digestion. J’
.
Ultrastruct. Res
.
59
,
101
112
.
Staehelin
,
L. A.
(
1973
).
Further observations on the fine structure of freeze-cleaved tight junctions
.
J. Cell Sci
.
13
,
763
786
.
Staehelin
,
L. A.
(
1974
).
The structure and function of intercellular junction
.
Int. Rev. Cytol
.
39
,
191
283
.
Wade
,
J. B.
&
Karnovsky
,
M. J.
(
1974a
).
The structure of the zonula occludens. A single fibril model based on freeze-fracture
.
J. Cell Biol
.
60
,
168
180
.
Wade
,
J. B.
&
Karnovsky
,
M. J.
(
1974b
).
Fracture faces of osmotically disrupted zonula occludens
.
J. Cell Biol
.
62
,
344
351
.
Wanson
,
J. C.
,
Drochmans
,
P.
,
Mosselmans
,
R.
&
Ronveaux
,
M. F.
(
1977
).
Adult rat hepatocytes in primary monolayer culture. Ultrastructural characteristics of intracellular contacts and cell membrane differentiations
.
J. Cell Biol
.
74
,
858
877
.
Zakai
,
N.
,
Kulka
,
R. G.
&
Loyter
,
A.
(
1977
).
Membrane ultrastructural changes during calcium phosphate induced fusion of human erythrocyte ghosts
.
Proc. natn. Acad. Sci. U. S. A
.
74
,
2417
2421
.