JOURNAL OF CELLULAR PHYSIOLOGY 143391-395 (1990)

Quantification of Protein Transcytosis in the Human Colon Carcinoma Cell Line CaCo-2 MARTINE HEYMAN,* ANNE-MARIE CRAIN-DENOYELLE, SAMlR KUMAR NATH, JEHAN-FRANCOIS DESJEUX

lnserrn U. 290,

HApitdl Saint-Lazare, 750 10 Paris,

AND

France

The transepithelial absorption of food-type proteins has been shown to proceed by endocytosis along two functional pathways: a minor direct pathway allowing transport of intact protein and J major lysosomal degradative pathway. The human colon carcinoma cell line CaCo-2 grown on Millipore filters was used here further to characterize these pathways by measuring HRP transport across the cell monolayer in Ussing chambers. In t h e apical-basal direction,this transport mainly occurred along the degradative pathway and was inhibited at 4°C (7.41 i 1.26 pnioles/h-cm’ vs. 27.40 -+ 8.91 at 37°C). T h e amount conveyed via the direct pathway was very small (0.89 0.35 prnoles/h.cm’) and did not diminish at 4°C (1.43 0.59 pmoles/h-cm’). In the basal-apical direction, HRP transport along the degradative pathway at 37°C was similar to the apical-basal value and wa5 inhibited at 4°C (16.40 2 4.05 vs. 2.72 5 2.52 prnoles/h.cm’), but along the direct pathway, it was eight times the apical-basal value (8.36 * 3.1 1 pmoIrDs/ h.cm2)and was inhibited at 4°C (2.43 * 0.78 pmoles/h.cm2).Intact HRP fluxes were not correlated with the electrical resistance of the filters, indicating transport via a transcellular route. Monensin at l o p 5 M did not affect direct or degradative transport in the apical-to-basal direction. These results suggest that in CaCo-2 cells HRP undergoes bidirectional transcytosis by a fluid-phase mechanism, but the extent of degradation during that transport varies according to t h e membrane (apical or basal) where it is presented.

*

*

Over the past ten years, numerous studies have shown that in polarized epithelia, vesicular transport of different peptides and proteins occurs via distinct mechanisms, including receptor-mediated endocytosis (Gonnella et al., 1987; Hasegawa et al., 1987; MaratosFlier et al., 1987; Rodewald and Kraehenbuhl, 1984; Solari and Kraehenbuhl, 1985) and non-specific endocytosis (Hidalgo et al., 1989; Von Bonsdorff et al., 1985). All these mechanisms involve the formation of endosomes which are sorted into different intracellular pathways, with or without passage through the lysosoma1 system. In previous studies of intact intestinal mucosa from different mammalian species, including man (Heyman et al., 1988, 1986, 1982), we used as protein models horseradish peroxidase (HRP) and also milk proteins (Marcon-Genty et al., 1989) to show that their transepithelial transport is transcellular. We also showed that a large proportion of these proteins is transported in a degraded form after passing through the lysosomal system, but that a small proportion (< 10%)is transported in a n intact or antigenically active form. However, the exact location of these direct and degradative pathways is still not known. The minute quantities of proteins crossing the tissue in intact form might either escape proteolysis in the lysosomal system or follow a distinct independent route, parallel to the lysosomal route. Furthermore, the intestinal mucosa consists of at least four types of epithelial cells (goblet, absorptive, endoG

1990 WILEY-LISS, INC

crine and Paneth) which themselves undergo different stages of maturation along the villi. It is therefore not possible to discern whether the direct and degradative transport pathways are similarly involved in each cell type, or whether a particular type preferentially uses a particular pathway. To quantitate and characterize the pathways of protein absorption, the human colonic carcinoma cell line CaCo-2 was used here as a transport system of homogeneous composition, since these cells were previously reported to undergo spontaneous epithelial differentiation in vitro (Pinto et al., 1983).

METHODS

Cell culture The human cell line CaCo-2 (Fogh et al., 1977) was obtained from Dr A. Zweibaum, INSERM U. 178, Paris, at passage 64. Cells were routinely grown as previously described (Pinto et al., 1983). Briefly, they were seeded at 1.2 lo4 cellskm’ in Dulbecco’s modified Eagle’s minimum essential medium (Eurobio, Paris, France) and grown at 37°C in a 10% C02-90% air atmosphere. The medium was supplemented with 20%’ fetal bovine serum (Boehringer, Mannheim, FRG), 1% Received July 26, 1989; accepted January 29, 1990. :#Towhom reprint requestsicorrespondence should be addressed.

392

HEYMAN ET AL

A

5 10 15 21 days after passage

cathepsin B

Fig. 1. A. Evolution of cathepsin B (01 and D (A)activities during the growth of CaCo-2 cells, compared to sucrase activity ( 0 ) .B: Values of cathepsin B activities previously found in neonatal and adult mouse intestine and in isolated villus and crypt cells in mouse (results from Heyman et al., 1986, 1989).

*

non-essential amino acids, 2 mM glutamine, and 100 was 168.9 k 15.3dL/cm2,with a mean PD of 0.27 0.07 U/ml of penicillin and streptomycin (Gibco, Paisley, mV (n = 18 filters). Apical-to-basal and basal-to-apical Scotland). The cells were grown in 25 cm2 plastic HRP fluxes were measured a s previously described dishes and passaged every 5 days, confluency being (Heyman et al., 1988). HRP (Sigma, type VI) was added reached 6-7 days after plating. For Ussing chamber to the apical or basal bathing compartment a t a final experiments cells were seeded a t a density of 4.5 lo4 concentration of 0.4 mg/ml. 'H-HRP (10 Ciimmole), lacells per cm' on HAHY Millipore filters with a diam- belled as previously described (Heyman et al., 19821, For 110 min, 800 eter of 13 mm (Millipore, Molsheim, France) which was also added as a tracer (1.35 $3). were fixed at three points to the bottom of 2 cm' four- pl samples were taken from the opposite bathing comwell dishes (Nunc, Denmark) with a soldering iron. partment at 20 min intervals and replaced by fresh Ringer's. The medium was changed daily. The rate of intact HRP transfer across the filter was Lysosomal proteolytic activities in CaCo-2 cells determined by enzymatic assay of 200 pl aliquots acThe activities of the lysosomal proteolytic enzymes cording to Maehly and Chance (1954). Results were cathepsin B and cathepsin D were measured during the expressed in pmoles/h.cm2. Total (intact and degraded) growth of CaCo-2 cells and compared with the activity HRP fluxes were assessed on 500 pl aliyuots by tritium of the brush-border enzyme sucrase. Cathepsin B and D counting, using liquid scintillation photometry. These were assayed as previously described (Heyman et al., total fluxes represented "H-HRP-equivalent, as previ1986),and the results were, respectively, expressed in ously described (Heyman et al., 1982). Degraded HRP nmoles paranitroanilineiminimg protein and nmoles fluxes were then calculated as total minus intact tyrosine-equivalentsiminimg protein. Proteins were as- fluxes. In some experiments, the temperature was sayed according to Lowry et al. (1951). Sucrase activity maintained a t 4°C using a n ice bath, and in others, the was assayed according to Messer and Dahlquist ( 1966). effect of a 2-hour preincubation of the monolayers with M monensin was also tested on HRP fluxes. Transport and degradation of HRP across Statistical analyses CaCo-2 monolayers Statistical analyses were made using the SAS proThe filter-grown CaCo-2 cell monolayers were used for HRP transport studies after complete maturation, gramme (SAS Institute Inc., 1985).Student's t-test was i.e., as from day 21 after plating. Filters were gently used to compare means and ranges. detached from the wells and mounted in suitable UsRESULTS sing chambers with a n exposed area of 12 mm'. Both sides were bathed with 1.5 ml Ringer's solution, pH 7.4, Lysosomal proteolytic activities in developing containing 10 mM glucose thermostated at 37°C and CaCo-2 monolayers oxygenated with 95% O,, 5% COz. The lysosomal system is mainly responsible for the Electrical parameters (potential difference [PD I and transepithelial electrical resistance [R]) were recorded degradation of the proteins endocytosed at the apical or regularly as a n index of cell viability and monolayer basal membrane of the enterocyte. To measure lysosointegrity. When R was less than 100 Wcm2, filters were ma1 proteolytic activity during the differentiation of discarded, a s were filters whose resistance dropped by CaCo-2 cells, we compared the development of cathepmore than 15% during the 2-hour experiments in the sin B and D activities with the known increase in the chambers. Under basal conditions, mean resistance activity of the brush border enzyme sucrase (Pinto et

(Heyman et al., 1986, 1989). As shown in Figure lA, there was a progressive increase in the activity of both cathepsin B and D which was parallel to the increase in sucrase activity. The fully mature monolayer (from day

which was faster than the 60 minutes required tb establish such fluxes in the full intestinal mucosa of different animal species and of man (Heyman et al., 1982, 1988). Intact HRP fluxes along the direct pathway were very small (0.89 ? 0.35 pmoles/h.cm2) but significantly different from zero (P < 0.02). However, these fluxes were not inhibited a t 4°C (1.43 0.59 pmolesi h.cm2), indicating that they might reflect some extracellular leakage. By contrast, degraded HRP fluxes were inhibited by 72% a t 4°C (27.40 ? 8.91 and 7.41 t 1.26 pmoles/h-cm" a t 37 and 4"C, respectively) indicating the existence of a n energy-dependent transcellular pathway.

*

E

2

;;;

12

-

time (min) CI

hl

E c!

B

I

C

I

aplcal t o b a r e l

80

100

\ 0

30

Ea

Y

cn

-: 2 0 3

r

Basal-to-apicalHRP fluxes along the direct and degradative pathways As Figure 2A,B and Table 1 indicate, intact basalto-apical HRP fluxes were ten times greater than apical-to-basal fluxes (P < 0.04). This large difference was probably not due to paracellular leakage, because the electrical resistance of the filters used in the basalto-apical flux measurements was not different from the resistance of those used in the apical-to-basal measurements (151 -t 10 vs. 176 * 19 Wcm", respectively), and because a t 4"C, these intact HRP fluxes were inhibited by 66%. However, due to the skewed distribution of intact HRP fluxes, this inhibition just failed to reach significance ( P = 0.09). Basal-to-apical degraded HRP fluxes (Fig. 2B) were similar to the apical-to-basal fluxes and at 4°C were inhibited by 88%. Effect of monensin on apical-to-basalHRP fluxes (Table 2) Monensin, a drug known to inhibit receptor-mediated endocytosis (Brown et al., 19831, was added a t 10 M to the culture medium 2 hours before and during the flux measurements in the Ussing chambers. Monensin significantly raised the PD of the cell monolayers but had no effect on HRP transcytosis from the apical to the basal side along either the direct or degradative pathway, indicating that in the present model, HRP was probably conveyed by non-specific endocytosis.

'

20

40

60

time (mln) Fig. 2. A Intact HRP fluxes as a function of time in Ussing chambers. Apical to basal direction (0)and basal to apical direction (nl. Values are means of 7-10 filters. B: Degraded HRP fluxes as a function of time in Ussing chambers. Apical to basal direction ( 0 ) and basal to apical direction (m).Values are means of 7-10 filters.

DISCUSSION The present results confirm t h a t the food-type protein HRP (MW = 40 kD) is transported by endocytosis, mainly along a degradative pathway that probably involves the epithelial lysosomal system. The direct pathway observed in the intact mucosa in previous studies (Heyman et al., 1982, 1988) during mucosal to serosal transport (i.e. apical to basal) was not present in the model of filter-grown CaCo-2 monolayers studied here. However, this direct pathway was responsible for much of the transport in the basal-to-apical direction (i.e., serosal to mucosal). In a recent work in which the CaCo-2 cell line was used as a model of intestinal epithelial permeability,

394

HEYMAN ET AL

TABLE 1. Effect of temperature on electrical parameters and steady state HRP fluxes across filter-grown CaCo-2 monolayers

PD (mV) R (f1/crn2) Intact HRP fluxes (prnoleslhan') Degraded HRP fluxes (prnoles/h.cm')

HRP fluxes' Apical to basal Basal to apical 37°C 4°C 37°C 4°C 0.33 t 0.12 0.05 i 0.04:': 0.20 0.04 0.02 5 0,02:':.: (10) (6) (7) (7) 268 29:%3; 176 19 299 + 30::.$. 151 i 10 (7) (91 (10) (6) 0.89 2 0.35 1.43 i 0.59 8.36 3.11";"" 2.43 2 0.78. (7) (9) (10) (6) 16.40 4.05 2.72 5 2.52:':::; 7.41 i 1.26::' 27.40 2 8.91 (9) (10) (6) (7) +

+

+

'Significantly different from values at 37°C I' P

"

0.05; '!.-:P.-. 0.01; P

=

0.09,.Different from apical to basal value at

37°C I ' ' P -:.0.041.

TABLE 2. Effect of monensin on apical to basal HRP fluxes (pmoles/h.cm')

Control monolayers Monolayers treated with 10

P

' M monensin

n PD (mV) R tllicrn") 7 0.33 i 0.12 176 i 19 6 1.05 2 0.12 161 i 20 < 0.002 NS

Hidalgo et al. (1989) showed by electron microscopy that in this cell line, HRP underwent endocytosis by a fluid phase mechanism, without any evidence of occluding junction or paracellular leakage. Here, the lack of inhibition of HRP transcytosis by monensin confirm that a fluid phase mechanism is involved. Moreover, the results indicate that CaCo-2 cells transcytose HRP a t a high rate and that the two pathways (direct and lysosomal) are not implicated to the same extent in both directions. Indeed, the degree of HRP degradation is modified according to the membrane where it is presented.

CaCo-2 cells have a large transcytotic capacity To compare the capacity for transcytosis of CaCo-2 monolayers with that of intact intestinal mucosa, we have to consider the number of absorptive cells per unit of filter (or serosal) surface. This number is indeed far larger in intact intestine, due to its villus architecture, than in a flat monolayer. We previously reported that the rate of HRP transport across the jejunum of healthy children was about 39 pmoles/htm2 (Heyman et al., 1988). Comparison of this rate with the one observed presently under the same conditions in CaCo-2 cells (27 pmoles/h.cm2)indicates that these cells have a larger transport capacity than intact intestine. Assuming that the presence of villi normally increases the absorptive surface roughly eightfold (Verzar and Mc Dougall, 1936), the present results suggest that CaCo-2 cells have a capacity for endocytosis five times greater than that of human jejunal epithelium. These results imply that CaCo-2 cells behave like immature crypt-like cells, a t least a s regards protein endocytosis. Despite the ability of CaCo-2 cells to differentiate partially into enterocyte-like cells, they also behave like intestinal crypt cells because a) they are mainly secretory cells (Grasset et al., 1985), b) they bear some crypt cell markers (Quaroni, 1986), and c) they display the high lysosomal proteolytic activity that characterizes neonatal or crypt intestinal cells

Intact HRP Degraded HHP fluxes fluxes 0.89 i 0.34 27.40 i 8.91 1.13 -+ 0.33 33.32 5 5.87 NS NS

(Heyman et al., 1986, 1989) and also high endocytotic activity, as reported recently for intestinal crypt cells (Heyman et al., 1989).

The unidirectional basal-to-apicaldirect pathway in CaCo-2 cells The present results show that CaCo-2 cells are capable of the unidirectional transfer of HRP in intact form, from the basal to the apical side of the monolayer. This transfer is unlikely to constitute paracellular leakage, because a ) the filter-grown monolayers used to measure basal-to-apical HRP fluxes had a n electrical resistance similar to the resistance of those used to measure apical-to-basal fluxes; b) a t 4"C, basal-to-apical intact HRP transfer was inhibited by 66%, indicating that this transfer is energy-dependent. Although the electrical resistance of CaCo-2 monolayers almost doubled at 4°C compared to 37"C, this increase had no effect on apical-to-basal intact HRP transport, indicating that restricted diffusion a t 4°C is unlikely, and c) there was no significant correlation between the electrical resistance of the filters and intact HRP fluxes, suggesting that the paracellular pathway is not involved. Such a transcellular secretion (basal-apical) of protein across a n epithelium has previously been described in different models of cell monolayers (Maratos-Flier et al., 1987; Von Bonsdorff et al., 1985) in hepatocytes (Renston et al., 1980) or in intestine (Gotze and Rothman, 1978). Why is there such a direct pathway in CaCo-2 cells? The transcytotic process requires polarized movement of the endocytotic vesicles, which are sorted during their transcellular migration and directed either to a lysosomal degradative pathway or conveyed via a direct pathway across the cell. It is increasingly likely that the paracellular diffusion is a n unusual pathway which is occasionally found to be significant in differ-

TRANSCYTOSIS IN CaCo-2 CELLS

ent pathological situations (Madara, 1989, Phillips et al., 1987). In the present model, the direct transfer of HRP from the basal to the apical side might constitute a n opportunistic transfer linked to the vectorial transfer of some other compound, a possible candidate being the secretory component (SC). Indeed, polymeric IgA are known to be transferred from the serosal to the luminal side of the intestine by a receptor-mediated endocytotic process involving the SC as a receptor. It is now recognized that free SC internalization and transport are constitutive processes independent of the binding of the ligand (IgA) (Mostov and Deitcher, 1986). Although the existence of SC on the basolateral side of CaCo-2 cells has not actually been demonstrated, i t is probable, more especially as SC is expressed preferentially in intestinal crypt cells (Brandtzaeg, 1974) and is present in HT 29 cells (Nagura et al., 1979). Taken together, our results suggest that the direct and degradative pathways participating in protein transport could be unequally involved in the different cell types that make up the intestinal epithelium. In the present crypt-like CaCo-2 cells, the amount of protein transported along the direct pathway from the basal to the apical membrane was more than 30% of the total amount internalized. This finding may have some connection with the recent observation that fluid phase markers can be internalized by all endocytic pathways, including receptor-mediated endocytosis, which could be responsible for 20-50% of the transport of fluid phase markers (Doxsey et al., 1987, Sandvig et al., 1987).

ACKNOWLEDGMENTS The authors are grateful to M. Dreyfus for revising the manuscript and L. Chadufau for skillful secretarial assistance. LITERATURE CITED Brandtzaeg, P. (1974) Mucosal and glandular distribution of immunoglobulin components. Immunohistochemistry with a cold ethanol-fixation technique. Immunology, 26.1101-1 114. Brown, M.S., Anderson, R.G.W., and Goldstein, J.L. (1983) Recycling receptors: the round trip itinerary of migrant membrane proteins. Cell, 32.663-667. Doxsey, S.J., Brodsky, F.M.. Blank, G.S., and Helenius, A. (19871 Inhibition of endocytosis by anti-clathrins antibodies. Cell, 50:453463. Fogh, J., Fogh, J.M., and Orfeo, T. (1977) One hundred and twentyseven cultured human tumor cell lines oroducing tumors in nude mice J Natl Cancer I n s t , 59 221-226 Gonnella, P A , Siminoski, K , Murphy, R A., and Neutra, M A (1987) Transepithelial transport of epidermal growth factor by absorptive cells of suckling rat. J. Clin. Invest., 8022-32. Gotze, H.. and Rothman, S.S. (1978) Amylase transport across ileal epithelium in vitro. Biochim. Biophys. Acta, 512t214-220. Grasset, E., Bernabeu, J., and Pinto, M. (1985)Epithelial properties of human colonic carcinoma cell line CaCo-2: effect of secretagogues. Am. J. Physiol., 248tC410-C418. Hasegawa, H., Nakamura, A., Watanabe, K., Brown, W.R., and Nagura, H. (1987) Intestinal uptake of IgG in suckling rats. Distinction between jejunal and ileal epithelial cells demonstrated by simultaneous ultrastructural localization of IgG and acid phosphatase. Gastroenterology, 92:186-191. Heyman, M., Ducroc, R., Desjeux, J.F., and Morgat, J.L. (1982) Horseradish peroxidase transport across adult rabbit _ieiunum in vitro. _ Am. J. Physiol., 242:G558-G564.

395

Heyman, M., Crain-Denoyelle, A.M., Corthier, G., Morgat, J.L., and Desjeux, J.F. (1986)Postnatal development of protein absorption in conventional and germ-free mice. Am. J. Physiol., 251:G326-G331. Heyman, M., Grasset, E., Ducroc, R., and Desjeux, J.F. ( 1988 I Antigen absorption by the jejunal epithelium of children with cow’s milk allergy. Pediatr. Res., 24:197-202. Heyman, M., Crain-Denoyelle, A.M., and Desjeux, J.F. ( 19891 Endocytosis and processing of protein by isolated villus and crypt cells of the mouse small intestine. J. Pediatr. Gastroenterol. Nutr., 9.238245. Hidalgo, I.J., h u b , T.J., and Borchardt, R.T. f 1989) Characterization of the human colon carcinoma cell line (CaCo-2)a s a model system for intestinal epithelial permeability. Gastroenterology, 96,736749. Jonard, P.P., Rambaud, J.C., Dive, C., Vaerman, J.P., Galian, A,, and Delacroix, D.L. (1984) Secretion of immunoglobulins and plasma proteins from the jejunal mucosa. Transport rate and origin of polymeric IgA. J. Clin. Invest., 74r525-535. Lowry, O.H., Rosebrough, N.J., Farr, A.L., and Randall, R.J. 119511 Protein measurement with t h e Folin phenol reagent. J. Biol. Chem., 193:265-275. Madara, J.L. (1989) Loosening tight junctions. Lessons from the intestine. J. Clin. Invest., 83t1089-1094. Maehly, A.C., and Chance, B. (1954) The assay of catalases and peroxidases. In: Methods of Biochemical Analysis. Glick, D. ed. Wiley Interscience, New York, Vol. I pp. 357-398. Maratos-Flier, E., Yang Kao, C.Y., Verdin, E.M., and King G.L. f 1987) Receptor-mediated vectorial transcytosis of epidermal growth factor by Madin-Darby Canine Kidney cells. J. Cell Biol., 105t1595-1601. Marcon-Genty, D., Tome, D., Kheroua, O . ,Dumontier, A.M., Heyman, M., and Desjeux, J.F. (1989) Transport of p-lactoglobulin across rabbit ileum in vitro. Am. J. Physiol., 256:G943-948. Messer, M., and Dahlquist, A. (1966)A one step ultramicromethod for the assay of intestinal disaccharidases. Anal. Biochem., 14r376392. Mostov, K.E., and Deitcher, D.L. (1986) Polymeric immunoglobulin receptor expressed in MDCK cell transcytoses IgA. Cell, 46:613621. Nagura, H., Nakane, P.K., and Brown, W.R. (19791 Translocation of dimeric IgA through neoplastic colon cells in vitro. J. Immunol., 123t2359-2363. Pinto, M., Robine-Leon, S., Appay, M.D., Kedinger, M., Triadou, N., Dussaulx, E., Lacroix, B., Simon-Assmann, P., Haffen, K., Fogh, J., and Zweibaum, A. (1983)Enterocyte-like differentiation and polarization of the human colon carcinoma cell line CaCo-2 in culture. Biol. Cell, 47:323-330. Phillips. T.E., Phillips, T.L., and Neutra, M.R. (19871 Macromolecules can pass through occluding junctions of r a t ileal epithelium during cholinergic stimulation. Cell Tissue Res., 247t547-554. Quaroni, A. (1988) Crypt cell antigen expression in human colon tumor cell lines: analysis with a panel of monoclonal antibodies to CaCo-2 luminal membrane components. J. Natl. Cancer Inst., 76: 571-585. Renston, R.H., Maloney, D.G., Jones, A.L., Hradec, G.T., Wong, K.Y., and Goldfine, I.D. (1980) Bile secretory apparatus: Evidence for a vesicular transport mechanism for proteins in the rat, using horseradish peroxidase and ““I-insulin. Gastroenterology, 78:13731388. Rodewald, R., and Kraehenbuhl, J.P. (1984)Receptor-mediated transport of IgG. J. Cell Biol., 99:159-164. Sandvig, K., Olsnes, S.,Petersen, O.W., and Deurs, B.V. (1987) Acidification of the cytosol inhibits endocytosis from coated pits. J. Cell Biol., 1051679-689. SAS Institute Inc (1985)Version G Edition. Guide for personal computers. Cary, North Carolina: SAS Institute Inc., p. 429. Solari, R., Kraehenbuhl, J.P. (1985) The biosynthesis of secretory component and its role in the transepithelial transport of IgA dimer. Immunol. Today, 6.17-20. Verzar, F., Mc Dougall, E.J. (1936) Absorption From the Intestine. London: Longmans, Green. Von Bonsdorff, C.H., Fuller, S.D., and Simons, K. (1985, Apical and basolateral endocytosis in Madin-Darby Canine Kidney (MDCKJ cells grown on nitrocellulose filters. EMBO J., 4/11): 2781-2792. Wilson, T.H. (1962) Intestinal Absorption. W.B. Saunders Co., Philadelphia.