EXPERIMENTAL
CELL
RESEARCH
192,
41-45
(1991)
Vectorial Transcytosis of lmmunoglobulin G by Human Term Trophoblast Cells in Culture S. R. SOORANNA AND S. F. CONTRACTOR’ Department of Obstetrics and Gynaecology, West London Hospital, Hammersmith
Charing Road,
ing culture, it is not possible to determine whether uptake or release occurs at the apical and/or basal cell surfaces. It has been shown that epithelial cells (e.g., Madin-Darby canine kidney cells) grown on microporous membrane filters adhere to them, maintaining their polarity. This allows tissue culture medium access to the microvilli of the apical surface, whereas the basolateral surface which sits on the substratum has access to medium through the pores of the filters [13]. We have tried to emulate this system with trophoblast cells and present our attempts at studying vectorial transport of IgG across these cells adhering to filters sealed into polystyrene chambers.
Trophoblast cells were grown on filters that allow access to apical and basal surfaces of cells. Using this experimental system, IgG transport was shown to be specific and to occur primarily in the apical to basal direction. This transport was time- and temperaturedependent, with approximately 10% of added IgG appearing on the basal side within a 60-min incubation at 37°C. Other substances such as heparin were transported only minimally, whereas horseradish peroxidase was transported to the same degree in both directions. Greater than 90% of the transported IgG was precipitable by trichloroacetic acid and 81% was capable of binding to protein G. Such a rapid transport of large amounts of IgG in trophoblast cells is consistent with a (0 1991 receptor-mediated process of transcytosis. Academic
Press,
MATERIALS
Materials were obtained as described previously [2]. Human IgG, HRP (type II), and Azure A were obt.ained from Sigma Chemical Co., Poole, Dorset. In addition, Sephadex G-25 and protein G-Sepharose 4B were from Pharmacia LKB Biotechnology, Sweden, and sodium heparin was from Paines and Byrne Ltd., Greenford. Twelve-millimeter Millicell-HA culture plate inserts consisting of a surfactantfree 0.45-Frn pore size microporous membrane filter sealed to a cylindrical polystyrene holder were obtained from Millipore Corp., Bedford, U.S.A. Trophoblast cells were prepared using a Trophoblast cell culture. discontinuous Percoll gradient as described previously 1121. Cells obtained by this procedure were dispersed into Ham’s F10 medium containing 15% fetal calf serum (FCS), 10 mM Hepes, 100 units/ml penicillin, 100 fig/ml streptomycin, and 100 units/ml nystatin at a concentration of 2 X lOa live cells/ml, and 0.45ml portions were added to the inner chamber of Millicell inserts. These were placed in Sterilin 24well plates containing 0.75 ml of culture medium. The different volumes were used in order to have the same level of fluid in the inner and outer chamber. Cells were incubated at 37°C in an atmosphere of air:CO, (95%:5%). At the same time cells were also cultured in 96.well plates. These allowed the cell types present in individual cultures to be identified by using epithelial cell-specific and trophoblast-specific monoclonal antibodies [a]. Only cultures consisting of greater than 90% keratinand trophoblast-positive cells were used for transport studies. The number of cells on filters was counted after fixation with 4% paraformaldehyde in phosphate-buffered saline (PBS) and stained with Harris’ haematoxylin (Raymond A. Lamb, North Acton, London). On Day 1 approximately 1 X lo5 nuclei (based on four different cell preparations) were counted on filters although gaps without cells were visible on all filters examined. ZgG transport. IgG was labeled with izsI using 1,3,4,6-tetrachloro36diphenyl glycoluril (Iodogen) by the method of Sullivan and Wil-
Transcellular transport of ligands, or transcytosis as it is known, occurs in several cell types including epithelial [13] and endothelial cells [lo]. This process is important for delivery of macromolecules such as hormones, growth factors, and immunoglobulins across a cellular barrier. Immunoglobulin G (IgG) is known to be transported intact across the placenta in order to confer passive immunity to the fetus. The mechanism whereby this macromolecule crosses the trophoblastic layer remains unknown, although the initial steps probably involve its binding to Fc receptors on the syncytiotrophoblast surface [20]. The steps involved in studying the progress of this molecule through the trophoblast have been hindered for lack of a suitable experimental system. Recently Kliman et al. [ 121used Percoll gradient centrifugation of trypsinized placental cells to obtain relatively pure trophoblasts for primary culture. These cells have been characterized by various studies [4,5]. However, when cells adhere to plastic or glass surfaces dur-
To whom
AND METHODS
Inc.
INTRODUCTION
’
Cross and Westminster Medical School, London W6 7DQ, United Kingdom
reprint
requests
should
be addressed. 41
All
Copyright 0 1991 rights of reproduction
0014.4827/9153.00 by Academic Press, Inc. in any form reserved.
42
SOORANNA
AND
liams [19]. iz51-1gG was passed through a Sephadex G-25 column and the void volume containing intact IgG was collected just prior to use. Ninety-six percent of this iz5 I-IgG was precipitable by 10% trichloroacetic acid. Cells on Day 1 of culture were used for experiments. After aspiration of culture medium from both cell surfaces, each side of the chamber was washed with fresh medium without FCS and then reincubated at 37OC in this medium for 60 min prior to the start of the experiment. Removal of this medium was followed by the addition of 0.75 and 0.45 ml of fresh medium to the outer (basal surface) and inner (apical surface) sides of the chamber with only one side containing 12.5 pg/ml labeled IgG (spec act, 500,000 dpm/pg). The different volumes were used in order to have the same level of fluid in the inner and outer chambers to avoid pressure differences. Inserts were incubated at 37°C and sampled from the side opposite that to which the labeled ligand had been added at the designated time points. Fifty-microliter portions were sampled and replaced with fresh medium. Samples were mixed with 0.35 ml of 0.05% bovine serum albumin (added as carrier protein) and precipitated with 0.1 ml of 50% trichloroacetic acid (TCA). The TCA-precipitated and TCA-soluble activities were separated by centrifugation and counted in a Packard 5650 gamma counter. In some experiments 0.2-ml portions were sampled at 30 and 60 min. One-tenth-milliliter portions were passed through a protein G-Sepharose 4B column in order to determine the percentage of activity capable of binding to this substance. IgC binding. After aspiration of growth medium and washing of cells, fresh medium without FCS was added and then reincubated at 37°C for 60 min followed by a further 30 min at 4°C. iz51-IgG (500,000 dpm) was added to either the apical or the basal surface together with variable amounts of unlabeled IgG ranging from 5 to 5000 fig/ml. Incubation for 2 h at 4°C was followed by washing six times with PBS. Filters were detached from chambers and counted. To determine nonspecific binding to filters, cells from some chambers were trypsinized off by incubating at 37°C with 0.2% trypsin for 30 min at room temperature. Cells were removed by washing with PBS and the filters counted. In the presence of 12.5 &ml IgG nonspecific binding to filters was 0.017% of the initial ligand added on the apical surface and 0.019% on the basal surface. Heparin trcnsport. Cells were incubated as above except that culture medium containing 2500 units/ml sodium heparin was added instead of labeled IgG. Samples were assayed for heparin by the method of Klein et al. [ll] by measuring the decrease in absorbance of Azure A at 620 nm. HRP transport. HRP was labeled by the Iodogen method and “‘IHRP was passed through a Sephadex G-25 column prior to use. 92.1% of the void volume was TCA-precipitable. Cells were incubated as above except that culture medium contained 12.5 fig/ml labeled HRP instead of IgG. Fifty-microliter portions were sampled and counted. Statistical assessment. Statistical significance of data was analyzed by Student’s t test. RESULTS
Transport of human IgG was monitored across trophoblast cells adhered to nitrocellulose filters. Approximately 10% of cells added to filters adhered to them. It was decided that an excess of cells should be added to filters to allow the maximum possible number of cells to attach. Both apical to basal and basal to apical transport were evaluated and expressed as percentages of the initial activity present (Fig. 1). The passive diffusion of IgG across filters without cells was low when measured under the same conditions. After a 45-min incubation in the presence of 12.5 pg/ml of IgG, transport of the li-
CONTRACTOR
0
15
30
Incubation
45
time
60
(mid
FIG. 1. Transport of IgG across trophohlast cells. “%IgG was added to either the apical or the basal surface of trophoblast cells grown within chambers (solid lines). Cells were incubated at 37°C and 50.~1 samples were taken at 15.min intervals from the other surface and replaced with an equal volume of fresh medium. The IgG transported was determined by calculating total cumulative counts transported as a percentage of the total tracer added. Each point represents the mean of 16 experiments and the bars are the SEM. Twelve experiments were also conducted using chambers without cells (dotted lines). Solid symbols, basal to apical transport; open symbols, apical to basal transport.
gand in the apical to basal direction was significantly lower at 4°C than at 37°C (2.65 + 0.46% (4 experiments) and 8.80 f 1.20% (14 experiments), respectively; p < 0.01). Basal to apical transport of the ligand was not significantly increased by temperature (0.62 f 0.44% at 4°C and 1.17 + 0.44% at 37°C). At 4°C IgG transport from the apical to basal direction was inhibited by increasing concentration of the ligand. After a 45-min incubation transport at 5,50,500, and 5000 pg/ml of IgG was 2.97 t 0.69, 2.82 -t 0.46,1.64 ? 0.58, and 1.11 t 0.52%, respectively, of the initial ligand concentration (means and SEM of four experiments). Transport from basal to apical was 0.43, 0.62, 0.65, and 0.95% of the initial IgG concentration, respectively. Table 1 shows that almost all the activity transcytosed at 37°C from the apical to basal direction was TCA-precipitable. Binding to protein G was used as an indicator of intactness of the IgG molecule and approximately 9% of the added ligand was broken down during transcytosis from the apical to basal direction. In the opposing direction more than 40% was degraded (Table 1). From Table 1 it can be seen that 3.0 + 0.2% of the incubation medium not exposed to cells consisted of TCA-soluble products which are mainly free radioactive iodine and some low molecular compounds. In the absence of cells it is possible to measure the transfer of TCA-soluble products across microporous filters. After a 60-min incubation the transfer of these low-molecular-weight compounds was 13.9 * 2.1% of the initial ac-
TRANSCYTOSIS
TABLE
OF
IMMUNOGLOBULIN
43
G
1
Analysis of Transported IgG Incubation time (min) 30
Condition With
With
ap to bas bas to ap ap to bas bas to ap ap to bas bas to ap ap to bas has to ap
cells cells
Without Incubation cells
from
cells
Without 60
Transport
cells
medium
not exposed
oioof protein 5% of TCAG-bound precipitable activity activity 81.3 48.2
0.0 0.0 81.0
49.4 3.1
0.0
ir + + t +++*
0.0
89.2 k 70.3 f 39.7 * 13.6 f 92.0 f 79.6 31 32.6 t 31.2 t
0.8
97.0 f
3.7",& 12.2
0.0 0.0 2.7".b 13.2 3.1
3.0a,* 7.5 11.4 0.6 1.2" 8.8
12.9 16.1
to 89.5 +
0.2
Note. Samplesof transportedligandafter 30. and 60.minincubations were either passed through protein G-Sepharose 4B or precipitated with trichloroacetic acid. The results refer to the percentage of these samplesthat either bindsto protein G or is precipitatedby TCA, and the results are the means and SEM of five experiments. ’ P < 0.05 for apical to basal transfer versus incubation medium not exposed to cells. * P < 0.05 for apical to basal transfer versus basal to apical transfer.
tivity in the apical to basal direction and 10.2 & 1.6% in the basal to apical direction (means and SEM of five experiments). Binding of IgG to trophoblast cells was determined by addition of different concentrations of labeled ligand either to the apical or to the basolateral side and incubating for 2 h at 4°C (Fig. 2). Preliminary experiments showed that this was the minimum time needed for maximal binding. Binding of ligand to the basal surface of trophoblast cells was substantially less than that at 0.11 r
T
0.09 -0 5 2
0.07
u ul
0.04
s
0.02
0.00
'-LULL 4
'
j
L ' "'I
10
100
IgG
Cont.
loo0
lug/ml)
FIG. 2. IgG binding to trophoblast cells. iz51-IgG was added to either the apical or t.he basal surface of trophoblast cells grown within chambers. Binding studies were performed in the presence of cold IgG ranging between 5 and 5000 fig/ml. After incubation at 4°C for 2 h, filters containing cells were washed with cold PBS, detached from chambers, and counted. The results refer to the percentage of IgG bound to cells and are the means of four experiments. The bars represent SEM. (0) Basal binding; (0) apical binding.
0
15
30
Incubation
45
time
60
(mln)
FIG. 3. Transport of heparin across trophoblast cells. Sodium heparin was added to either the apical or the basal surface of trophoblast cells grown within chambers (solid lines). Cells were incubated at 37°C and 50-~1 samples were taken at 15.min intervals from the other surface and replaced with an equal volume of fresh medium. The heparin transported was determined by calculating the total cumulative amount transported as a percentage of the initial amount added. Each point represents the mean of five experiments and the bars represent SEM. The dotted lines represent experiments conducted using chambers without cells. Solid symbols, basal to apical transport; open symbols, apical to basal transport.
the apical surface. Moreover, no inhibition of binding was seen at high levels of unlabeled ligand at the basal surface. The ratio of apical to basal binding of IgG at 5, 12.5, and 50 pg/ml was 5.2, 4.4, and 6.1, respectively. Using Scatchard [16] analysis on the binding of IgG to the apical surface of trophoblast cells a linear correlation was obtained (r = -0.94). The number of binding sites/nucleus (2.8 X 106) was calculated for IgG with a Kd of 3.9 x lop5 M. The transport of two other molecules was also evaluated using this experimental system. Figure 3 shows the transfer of sodium heparin (lo-15 kDa), which is known to cross the placenta only minimally in uiuo [6]. A high transfer of heparin is seen with blank filters. The presence of cells does not significantly alter the transfer of this compound. HRP (44 kDa) is used to demonstrate nonspecific endocytosis in cells [ 181and when examined in this experimental system shows transport in both apical to basal and basal to apical directions to approximately the same extent (Fig. 4). The data from Figs. 1, 3, and 4 are summarized in Table 2. By subtracting the no cell transfer values from those in the presence of cells and presenting the ratio of apical to basal transfer, it can be seen that there is almost no overall transport of heparin across trophoblast cells. IgG is preferentially transported from the apical to basal surface of trophoblast cells, whereas HRP is transported equally well in both directions. DISCUSSION Transcytosis has been established in a variety of cell types [7,8,14, 171. One of the best characterized exam-
44
SOORANNA 10
r
AND
T
0
15
30
Incubation
45
time
60
(mid
FIG. 4. Transport of HRP across trophoblast cells. “%HRP was added to either the apical or the basal surface of trophoblast cells grown within chambers (solid lines). Cells were incubated at 37°C and 50-~1 samples were taken at 15.min intervals from the other surface and replaced with an equal volume of fresh medium. The HRP transported was determined by calculating total cumulative counts transported as a percentage of the total tracer added. Each point represents the mean of three experiments and the bars represent SEM. The dotted lines represent experiments conducted using chambers without cells. Solid symbols, basal to apical transport; open symbols, apical to basal transport.
ples of transcytosis involves the transport of IgA across epithelial cells, where it appears in many secretions including milk, bile, saliva, tears, as well as respiratory and intestinal secretions [ 151. Although the mechanism of this process is not yet fully understood, some specialized tissues, such as the trophoblast, probably utilize a similar mechanism for the transfer of maternal IgG molecules from the apical to the basal surfaces of this cellular layer in order to confer passive immunity to the fetus. Trophoblast cells separated by Percoll gradient centrifugation of trypsinized term placental cells were cultured on microporous filters within plastic inserts.
CONTRACTOR
These allowed independent access to apical and basal surfaces of cells and the transport of radiolabeled IgG was monitored in both directions. The data presented here are consistent with rapid transport of IgG in the apical to basal direction by trophoblast cells. This process is both time- and temperature-dependent. The initial lag in appearance of IgG with time is indicative of cellular interaction of the ligand during transcytosis [13]. The passive diffusion of labeled IgG across filters without cells was negligible during the time period examined. As gaps existed on filters when cells were present, passive diffusion of IgG across these gaps contributed only minimally to the total amount of transported ligand measured. The large amount of IgG transported in the presence of cells suggests a receptor-mediated process of transcytosis by trophoblast cells in culture. The transport process appears to be unidirectional with the concentration of receptors for IgG being approximately five times greater on the apical surface when compared to the basal surface. In addition, most of the transported IgG is precipitable by trichloroacetic acid and is able to bind protein G and therefore must escape lysosomal breakdown within the cell. The small percentage that is broken down probably enters the cell by nonspecific endocytosis and is directed to the lysosomes. The much greater percentage of degraded IgG found during basal to apical transport of the ligand suggests that a large proportion of this pathway is nonspecific. Transcytosed IgA is released extracellularly with part of the receptor (the secretory component) still attached to the ligand [ 151. As transcytosed IgG binds to protein G, it indicates that the Fc portion of the molecule is intact, suggesting that the ligand is transcytosed by an intracellular route different from that of IgA. Although the evidence indicates the involvement of
TABLE
2
Transport of IgG, Heparin,
and
HRP o/c transfer
Incubation time (min) 15 30 45 60
Transport from Apical Basal Apical Basal Apical Basal Apical Basal
to basal to apical to basal to apical to basal to apical to basal to apical
Heparin
W + i ++8.79 +1.62 +10.48 t 1.92 f 2.47 0.32 5.65 1.04
0.66"
0.19 (7.7) 0.81' 0.51 (5.4) 1.13* 0.71 (5.4) 1.22b
0.74 (5.5)
to ligand
k 0.10 0.0 * 0.0 (0.0) 0.11 * 0.11 0.05 + 0.05 (2.2) 0.06 ? 0.04 0.02 -+ 0.02 (3.0) 0.24 -+ 0.24 0.0 ? 0.0 (0.0)
0.27
HRP 2.49 1.25 3.73 2.75 5.32 5.45 7.53 5.73
f k k I!I -+ +i
0.66 0.32 0.87 0.32 1.53 1.46 1.71
(2.0) (1.4)
(1.0)
f 0.97 (1.3)
Note. The values of transported ligand were obtained by subtracting values in the absence of cells from those in the presence of cells. The results are from Figs. 1,3, and 4 and are the means and SEM of 16,5, and 3 experiments for IgG, heparin, and HRP, respectively. The ratios of apical to basal transfer are in parentheses. a and b represent P < 0.01 and
CELL
RESEARCH
192,
41-45
(1991)
Vectorial Transcytosis of lmmunoglobulin G by Human Term Trophoblast Cells in Culture S. R. SOORANNA AND S. F. CONTRACTOR’ Department of Obstetrics and Gynaecology, West London Hospital, Hammersmith
Charing Road,
ing culture, it is not possible to determine whether uptake or release occurs at the apical and/or basal cell surfaces. It has been shown that epithelial cells (e.g., Madin-Darby canine kidney cells) grown on microporous membrane filters adhere to them, maintaining their polarity. This allows tissue culture medium access to the microvilli of the apical surface, whereas the basolateral surface which sits on the substratum has access to medium through the pores of the filters [13]. We have tried to emulate this system with trophoblast cells and present our attempts at studying vectorial transport of IgG across these cells adhering to filters sealed into polystyrene chambers.
Trophoblast cells were grown on filters that allow access to apical and basal surfaces of cells. Using this experimental system, IgG transport was shown to be specific and to occur primarily in the apical to basal direction. This transport was time- and temperaturedependent, with approximately 10% of added IgG appearing on the basal side within a 60-min incubation at 37°C. Other substances such as heparin were transported only minimally, whereas horseradish peroxidase was transported to the same degree in both directions. Greater than 90% of the transported IgG was precipitable by trichloroacetic acid and 81% was capable of binding to protein G. Such a rapid transport of large amounts of IgG in trophoblast cells is consistent with a (0 1991 receptor-mediated process of transcytosis. Academic
Press,
MATERIALS
Materials were obtained as described previously [2]. Human IgG, HRP (type II), and Azure A were obt.ained from Sigma Chemical Co., Poole, Dorset. In addition, Sephadex G-25 and protein G-Sepharose 4B were from Pharmacia LKB Biotechnology, Sweden, and sodium heparin was from Paines and Byrne Ltd., Greenford. Twelve-millimeter Millicell-HA culture plate inserts consisting of a surfactantfree 0.45-Frn pore size microporous membrane filter sealed to a cylindrical polystyrene holder were obtained from Millipore Corp., Bedford, U.S.A. Trophoblast cells were prepared using a Trophoblast cell culture. discontinuous Percoll gradient as described previously 1121. Cells obtained by this procedure were dispersed into Ham’s F10 medium containing 15% fetal calf serum (FCS), 10 mM Hepes, 100 units/ml penicillin, 100 fig/ml streptomycin, and 100 units/ml nystatin at a concentration of 2 X lOa live cells/ml, and 0.45ml portions were added to the inner chamber of Millicell inserts. These were placed in Sterilin 24well plates containing 0.75 ml of culture medium. The different volumes were used in order to have the same level of fluid in the inner and outer chamber. Cells were incubated at 37°C in an atmosphere of air:CO, (95%:5%). At the same time cells were also cultured in 96.well plates. These allowed the cell types present in individual cultures to be identified by using epithelial cell-specific and trophoblast-specific monoclonal antibodies [a]. Only cultures consisting of greater than 90% keratinand trophoblast-positive cells were used for transport studies. The number of cells on filters was counted after fixation with 4% paraformaldehyde in phosphate-buffered saline (PBS) and stained with Harris’ haematoxylin (Raymond A. Lamb, North Acton, London). On Day 1 approximately 1 X lo5 nuclei (based on four different cell preparations) were counted on filters although gaps without cells were visible on all filters examined. ZgG transport. IgG was labeled with izsI using 1,3,4,6-tetrachloro36diphenyl glycoluril (Iodogen) by the method of Sullivan and Wil-
Transcellular transport of ligands, or transcytosis as it is known, occurs in several cell types including epithelial [13] and endothelial cells [lo]. This process is important for delivery of macromolecules such as hormones, growth factors, and immunoglobulins across a cellular barrier. Immunoglobulin G (IgG) is known to be transported intact across the placenta in order to confer passive immunity to the fetus. The mechanism whereby this macromolecule crosses the trophoblastic layer remains unknown, although the initial steps probably involve its binding to Fc receptors on the syncytiotrophoblast surface [20]. The steps involved in studying the progress of this molecule through the trophoblast have been hindered for lack of a suitable experimental system. Recently Kliman et al. [ 121used Percoll gradient centrifugation of trypsinized placental cells to obtain relatively pure trophoblasts for primary culture. These cells have been characterized by various studies [4,5]. However, when cells adhere to plastic or glass surfaces dur-
To whom
AND METHODS
Inc.
INTRODUCTION
’
Cross and Westminster Medical School, London W6 7DQ, United Kingdom
reprint
requests
should
be addressed. 41
All
Copyright 0 1991 rights of reproduction
0014.4827/9153.00 by Academic Press, Inc. in any form reserved.
42
SOORANNA
AND
liams [19]. iz51-1gG was passed through a Sephadex G-25 column and the void volume containing intact IgG was collected just prior to use. Ninety-six percent of this iz5 I-IgG was precipitable by 10% trichloroacetic acid. Cells on Day 1 of culture were used for experiments. After aspiration of culture medium from both cell surfaces, each side of the chamber was washed with fresh medium without FCS and then reincubated at 37OC in this medium for 60 min prior to the start of the experiment. Removal of this medium was followed by the addition of 0.75 and 0.45 ml of fresh medium to the outer (basal surface) and inner (apical surface) sides of the chamber with only one side containing 12.5 pg/ml labeled IgG (spec act, 500,000 dpm/pg). The different volumes were used in order to have the same level of fluid in the inner and outer chambers to avoid pressure differences. Inserts were incubated at 37°C and sampled from the side opposite that to which the labeled ligand had been added at the designated time points. Fifty-microliter portions were sampled and replaced with fresh medium. Samples were mixed with 0.35 ml of 0.05% bovine serum albumin (added as carrier protein) and precipitated with 0.1 ml of 50% trichloroacetic acid (TCA). The TCA-precipitated and TCA-soluble activities were separated by centrifugation and counted in a Packard 5650 gamma counter. In some experiments 0.2-ml portions were sampled at 30 and 60 min. One-tenth-milliliter portions were passed through a protein G-Sepharose 4B column in order to determine the percentage of activity capable of binding to this substance. IgC binding. After aspiration of growth medium and washing of cells, fresh medium without FCS was added and then reincubated at 37°C for 60 min followed by a further 30 min at 4°C. iz51-IgG (500,000 dpm) was added to either the apical or the basal surface together with variable amounts of unlabeled IgG ranging from 5 to 5000 fig/ml. Incubation for 2 h at 4°C was followed by washing six times with PBS. Filters were detached from chambers and counted. To determine nonspecific binding to filters, cells from some chambers were trypsinized off by incubating at 37°C with 0.2% trypsin for 30 min at room temperature. Cells were removed by washing with PBS and the filters counted. In the presence of 12.5 &ml IgG nonspecific binding to filters was 0.017% of the initial ligand added on the apical surface and 0.019% on the basal surface. Heparin trcnsport. Cells were incubated as above except that culture medium containing 2500 units/ml sodium heparin was added instead of labeled IgG. Samples were assayed for heparin by the method of Klein et al. [ll] by measuring the decrease in absorbance of Azure A at 620 nm. HRP transport. HRP was labeled by the Iodogen method and “‘IHRP was passed through a Sephadex G-25 column prior to use. 92.1% of the void volume was TCA-precipitable. Cells were incubated as above except that culture medium contained 12.5 fig/ml labeled HRP instead of IgG. Fifty-microliter portions were sampled and counted. Statistical assessment. Statistical significance of data was analyzed by Student’s t test. RESULTS
Transport of human IgG was monitored across trophoblast cells adhered to nitrocellulose filters. Approximately 10% of cells added to filters adhered to them. It was decided that an excess of cells should be added to filters to allow the maximum possible number of cells to attach. Both apical to basal and basal to apical transport were evaluated and expressed as percentages of the initial activity present (Fig. 1). The passive diffusion of IgG across filters without cells was low when measured under the same conditions. After a 45-min incubation in the presence of 12.5 pg/ml of IgG, transport of the li-
CONTRACTOR
0
15
30
Incubation
45
time
60
(mid
FIG. 1. Transport of IgG across trophohlast cells. “%IgG was added to either the apical or the basal surface of trophoblast cells grown within chambers (solid lines). Cells were incubated at 37°C and 50.~1 samples were taken at 15.min intervals from the other surface and replaced with an equal volume of fresh medium. The IgG transported was determined by calculating total cumulative counts transported as a percentage of the total tracer added. Each point represents the mean of 16 experiments and the bars are the SEM. Twelve experiments were also conducted using chambers without cells (dotted lines). Solid symbols, basal to apical transport; open symbols, apical to basal transport.
gand in the apical to basal direction was significantly lower at 4°C than at 37°C (2.65 + 0.46% (4 experiments) and 8.80 f 1.20% (14 experiments), respectively; p < 0.01). Basal to apical transport of the ligand was not significantly increased by temperature (0.62 f 0.44% at 4°C and 1.17 + 0.44% at 37°C). At 4°C IgG transport from the apical to basal direction was inhibited by increasing concentration of the ligand. After a 45-min incubation transport at 5,50,500, and 5000 pg/ml of IgG was 2.97 t 0.69, 2.82 -t 0.46,1.64 ? 0.58, and 1.11 t 0.52%, respectively, of the initial ligand concentration (means and SEM of four experiments). Transport from basal to apical was 0.43, 0.62, 0.65, and 0.95% of the initial IgG concentration, respectively. Table 1 shows that almost all the activity transcytosed at 37°C from the apical to basal direction was TCA-precipitable. Binding to protein G was used as an indicator of intactness of the IgG molecule and approximately 9% of the added ligand was broken down during transcytosis from the apical to basal direction. In the opposing direction more than 40% was degraded (Table 1). From Table 1 it can be seen that 3.0 + 0.2% of the incubation medium not exposed to cells consisted of TCA-soluble products which are mainly free radioactive iodine and some low molecular compounds. In the absence of cells it is possible to measure the transfer of TCA-soluble products across microporous filters. After a 60-min incubation the transfer of these low-molecular-weight compounds was 13.9 * 2.1% of the initial ac-
TRANSCYTOSIS
TABLE
OF
IMMUNOGLOBULIN
43
G
1
Analysis of Transported IgG Incubation time (min) 30
Condition With
With
ap to bas bas to ap ap to bas bas to ap ap to bas bas to ap ap to bas has to ap
cells cells
Without Incubation cells
from
cells
Without 60
Transport
cells
medium
not exposed
oioof protein 5% of TCAG-bound precipitable activity activity 81.3 48.2
0.0 0.0 81.0
49.4 3.1
0.0
ir + + t +++*
0.0
89.2 k 70.3 f 39.7 * 13.6 f 92.0 f 79.6 31 32.6 t 31.2 t
0.8
97.0 f
3.7",& 12.2
0.0 0.0 2.7".b 13.2 3.1
3.0a,* 7.5 11.4 0.6 1.2" 8.8
12.9 16.1
to 89.5 +
0.2
Note. Samplesof transportedligandafter 30. and 60.minincubations were either passed through protein G-Sepharose 4B or precipitated with trichloroacetic acid. The results refer to the percentage of these samplesthat either bindsto protein G or is precipitatedby TCA, and the results are the means and SEM of five experiments. ’ P < 0.05 for apical to basal transfer versus incubation medium not exposed to cells. * P < 0.05 for apical to basal transfer versus basal to apical transfer.
tivity in the apical to basal direction and 10.2 & 1.6% in the basal to apical direction (means and SEM of five experiments). Binding of IgG to trophoblast cells was determined by addition of different concentrations of labeled ligand either to the apical or to the basolateral side and incubating for 2 h at 4°C (Fig. 2). Preliminary experiments showed that this was the minimum time needed for maximal binding. Binding of ligand to the basal surface of trophoblast cells was substantially less than that at 0.11 r
T
0.09 -0 5 2
0.07
u ul
0.04
s
0.02
0.00
'-LULL 4
'
j
L ' "'I
10
100
IgG
Cont.
loo0
lug/ml)
FIG. 2. IgG binding to trophoblast cells. iz51-IgG was added to either the apical or t.he basal surface of trophoblast cells grown within chambers. Binding studies were performed in the presence of cold IgG ranging between 5 and 5000 fig/ml. After incubation at 4°C for 2 h, filters containing cells were washed with cold PBS, detached from chambers, and counted. The results refer to the percentage of IgG bound to cells and are the means of four experiments. The bars represent SEM. (0) Basal binding; (0) apical binding.
0
15
30
Incubation
45
time
60
(mln)
FIG. 3. Transport of heparin across trophoblast cells. Sodium heparin was added to either the apical or the basal surface of trophoblast cells grown within chambers (solid lines). Cells were incubated at 37°C and 50-~1 samples were taken at 15.min intervals from the other surface and replaced with an equal volume of fresh medium. The heparin transported was determined by calculating the total cumulative amount transported as a percentage of the initial amount added. Each point represents the mean of five experiments and the bars represent SEM. The dotted lines represent experiments conducted using chambers without cells. Solid symbols, basal to apical transport; open symbols, apical to basal transport.
the apical surface. Moreover, no inhibition of binding was seen at high levels of unlabeled ligand at the basal surface. The ratio of apical to basal binding of IgG at 5, 12.5, and 50 pg/ml was 5.2, 4.4, and 6.1, respectively. Using Scatchard [16] analysis on the binding of IgG to the apical surface of trophoblast cells a linear correlation was obtained (r = -0.94). The number of binding sites/nucleus (2.8 X 106) was calculated for IgG with a Kd of 3.9 x lop5 M. The transport of two other molecules was also evaluated using this experimental system. Figure 3 shows the transfer of sodium heparin (lo-15 kDa), which is known to cross the placenta only minimally in uiuo [6]. A high transfer of heparin is seen with blank filters. The presence of cells does not significantly alter the transfer of this compound. HRP (44 kDa) is used to demonstrate nonspecific endocytosis in cells [ 181and when examined in this experimental system shows transport in both apical to basal and basal to apical directions to approximately the same extent (Fig. 4). The data from Figs. 1, 3, and 4 are summarized in Table 2. By subtracting the no cell transfer values from those in the presence of cells and presenting the ratio of apical to basal transfer, it can be seen that there is almost no overall transport of heparin across trophoblast cells. IgG is preferentially transported from the apical to basal surface of trophoblast cells, whereas HRP is transported equally well in both directions. DISCUSSION Transcytosis has been established in a variety of cell types [7,8,14, 171. One of the best characterized exam-
44
SOORANNA 10
r
AND
T
0
15
30
Incubation
45
time
60
(mid
FIG. 4. Transport of HRP across trophoblast cells. “%HRP was added to either the apical or the basal surface of trophoblast cells grown within chambers (solid lines). Cells were incubated at 37°C and 50-~1 samples were taken at 15.min intervals from the other surface and replaced with an equal volume of fresh medium. The HRP transported was determined by calculating total cumulative counts transported as a percentage of the total tracer added. Each point represents the mean of three experiments and the bars represent SEM. The dotted lines represent experiments conducted using chambers without cells. Solid symbols, basal to apical transport; open symbols, apical to basal transport.
ples of transcytosis involves the transport of IgA across epithelial cells, where it appears in many secretions including milk, bile, saliva, tears, as well as respiratory and intestinal secretions [ 151. Although the mechanism of this process is not yet fully understood, some specialized tissues, such as the trophoblast, probably utilize a similar mechanism for the transfer of maternal IgG molecules from the apical to the basal surfaces of this cellular layer in order to confer passive immunity to the fetus. Trophoblast cells separated by Percoll gradient centrifugation of trypsinized term placental cells were cultured on microporous filters within plastic inserts.
CONTRACTOR
These allowed independent access to apical and basal surfaces of cells and the transport of radiolabeled IgG was monitored in both directions. The data presented here are consistent with rapid transport of IgG in the apical to basal direction by trophoblast cells. This process is both time- and temperature-dependent. The initial lag in appearance of IgG with time is indicative of cellular interaction of the ligand during transcytosis [13]. The passive diffusion of labeled IgG across filters without cells was negligible during the time period examined. As gaps existed on filters when cells were present, passive diffusion of IgG across these gaps contributed only minimally to the total amount of transported ligand measured. The large amount of IgG transported in the presence of cells suggests a receptor-mediated process of transcytosis by trophoblast cells in culture. The transport process appears to be unidirectional with the concentration of receptors for IgG being approximately five times greater on the apical surface when compared to the basal surface. In addition, most of the transported IgG is precipitable by trichloroacetic acid and is able to bind protein G and therefore must escape lysosomal breakdown within the cell. The small percentage that is broken down probably enters the cell by nonspecific endocytosis and is directed to the lysosomes. The much greater percentage of degraded IgG found during basal to apical transport of the ligand suggests that a large proportion of this pathway is nonspecific. Transcytosed IgA is released extracellularly with part of the receptor (the secretory component) still attached to the ligand [ 151. As transcytosed IgG binds to protein G, it indicates that the Fc portion of the molecule is intact, suggesting that the ligand is transcytosed by an intracellular route different from that of IgA. Although the evidence indicates the involvement of
TABLE
2
Transport of IgG, Heparin,
and
HRP o/c transfer
Incubation time (min) 15 30 45 60
Transport from Apical Basal Apical Basal Apical Basal Apical Basal
to basal to apical to basal to apical to basal to apical to basal to apical
Heparin
W + i ++8.79 +1.62 +10.48 t 1.92 f 2.47 0.32 5.65 1.04
0.66"
0.19 (7.7) 0.81' 0.51 (5.4) 1.13* 0.71 (5.4) 1.22b
0.74 (5.5)
to ligand
k 0.10 0.0 * 0.0 (0.0) 0.11 * 0.11 0.05 + 0.05 (2.2) 0.06 ? 0.04 0.02 -+ 0.02 (3.0) 0.24 -+ 0.24 0.0 ? 0.0 (0.0)
0.27
HRP 2.49 1.25 3.73 2.75 5.32 5.45 7.53 5.73
f k k I!I -+ +i
0.66 0.32 0.87 0.32 1.53 1.46 1.71
(2.0) (1.4)
(1.0)
f 0.97 (1.3)
Note. The values of transported ligand were obtained by subtracting values in the absence of cells from those in the presence of cells. The results are from Figs. 1,3, and 4 and are the means and SEM of 16,5, and 3 experiments for IgG, heparin, and HRP, respectively. The ratios of apical to basal transfer are in parentheses. a and b represent P < 0.01 and
