THE INTRACELLULAR TRAFFICKING OF TRANSFERRIN IN CACO-2 CELLS AND
ITS IMPLICATION TO TRANSFERRIN-MEDIATED ORAL DRUG DELIVERY
by
Fariba Norouziyan
________________________________________________________________________
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PHARMACEUTICAL SCIENCES)
May 2007
Copyright 2007 Fariba Norouziyan
ii
Dedication
This dissertation is dedicated to my family and my husband Keith.
iii
Acknowledgments
I thank my advisor, Dr. Wei-Chiang Shen for years of support. I would like to
thank my committee members, prof. Austin Mircheff, Curtis Okamoto, Judy
Garner, and Sarah Hamm-Alvarez for their help and guidance. My special thanks
to Dr. Sarah Hamm-Alvarez for providing the opportunity for me to use the
confocal microscopy.
I would also like to thank Mrs. Daisy Shen for her help with in vivo study.
Finally, I would like to thank my family who have always been there for me. My
special thanks to my beloved parents who always encouraged me. My thanks to
my husband Keith for his support and understanding.
iv
Table of Contents
Dedication ii
Acknowledgments iii
List of Tables viii
List of Figures ix
List of Schemes xii
Abbreviations xiii
Abstract xiv
Preface xvii
Significance xvii
Chapter I. Background Information and
Review of the Literature 1
1. Structure and function of transferrin 2
2. Structure and function of transferrin receptor 11
3. Regulation of expression and tissue distribution
of transferrin receptor 15
4. Intracellular trafficking of transferrin 17
5. Rab proteins involved in transferrin trafficking 20
6. Utilization of transferrin receptor for
drug delivery and drug targeting 24
7. Enhancers for transferrin receptor-mediated
drug delivery 28
a. Brefeldin A 28
b. Tyrphostin A8 (AG10) 31
Chapter II. The selectivity of the effect of AG10;
AG10 increases the transferrin transcytosis
in Caco-2 cells, but not in MDCK cells 33
1. Abstract 33
2. Introduction 35
v
3. Experimental 36
a. Cell culture 36
b. Preparation of diferric and radiolabeled Tf 36
c. Transport assay using 125I-Tf 37
d. Pulse-chase experiment using 125I-Tf 38
4. Results 41
a. AG10 increased the TfR-mediated
transcytosis of Tf in Caco-2 cells, but it did not
show any effect on Tf transcytosis in MDCK cells 41
b. AG10 did not show any effect on the
transcytosis and recycling of Tf in MDCK cells 44
5. Discussion 51
Summary 56
Chapter III. Involvement of Rab11 in the trafficking of apically
endocytosed transferrin; AG10 stimulates a novel trafficking
pathway of the apically endocytosed transferrin through
Rab11-enriched compartment in Caco-2 cells 57
1. Abstract 57
2. Introduction 59
3. Experimental 61
a. Cell culture 61
b. Preparation of diferric, radiolabeled, and
fluorescein isothiocyanate labeled Tf 61
formula III-1: calculation of protein
concentration and the degree of FITC labeling 62
c. Pulse-chase experiment using 125I-Tf 64
d. 125I-Tf uptake assay 65
e. Confocal fluorescence microscopy 66
f. Image processing 67
4. Results 72
a. FITC-Tf was colocalized with Rab5 and Rab11 in
Caco-2 cells pulsed at 16°C and 37°C, respectively. 72
b. FITC-Tf was colocalized with Rab5 and Rab11 in two
different intracellular locations above the tight junction. 79
c. The effect of AG10 on transcytosis, recycling, and cell-association
of Tf was only observed in the cells pulsed
at 37°C. 86
vi
d. AG10 did not show any effect on the integrity of
tight junction. 93
e. AG10 increased colocalization between FITC-Tf
and Rab11. 95
f. BFA did not show any effect on the apical
recycling of 125I-Tf. 98
g. BFA did not show any effect on the colocalization
between FITC-Tf and Rab11 in Caco-2 cells. 100
5. Discussion 107
Summary 114
Chapter IV. Involvement of Rab11 in the basolateral
trafficking of transferrin; AG10 increases the
transport of the basolaterally endocytosed
transferrin to the Rab11-containing endosomal
compartment in Caco-2 cells 116
1. Abstract 116
2. Introduction 118
3. Experimental 119
a. Cell culture 119
b. Preparation of diferric, radiolabeled, and
fluorescein isothiocyanate labeled Tf 119
c. Pulse-chase experiment using 125I-Tf 119
d. Confocal fluorescence microscopy 120
e. Image processing 120
4. Results 121
a. AG10 increased the transcytosis and
recycling of the basolaterally endocytosed Tf
in Caco-2 cells. 121
b. The basolaterally endocytosed FITC-Tf
was colocalized with Rab11 to an extent in
Caco-2 cells treated with AG10. 124
5. Discussion 134
Summary 138
Chapter V. AG10 increases the serum concentration of the orally
administered Biotin-Tf conjugate in CF-1 mice 139
1. Abstract 139
2. Introduction 140
vii
3. Experimental 142
a. Materials 142
b. Preparation of biotin labeled Tf 142
formula V-1: calculation of the
degree of biotin labeling 143
c. Animal model 145
d. Animal study 145
e. ELISA assay 146
4. Results 148
a. The serum concentration of Biotin-Tf
conjugate reached the maximum at 8 to 12 hours
after the oral administration of the conjugate. 148
b. AG10 increased the serum concentration of
the Biotin-Tf conjugate in CF-1 mice. 150
5. Discussion 152
Summary 155
Chapter VI. Summary and future perspectives 157
1. Summary 157
2. Future perspectives 162
References 163
viii
List of Tables
I-1. Residues in the C-lobe of hTf that interact with the hTfR. 7
II-1. The recycling and transcytosis of Tf in the absence
and presence of AG10 or BFA in MDCK cells. 47
III-1. Summary of the involvement of Rab5 and Rab11
in Tf intracellular trafficking. 60
ix
List of Figures
I-1. The crystal structure of apo-hTf. 4
I-2. The interface between the two-lobes of apo-hTf. 5
I-3. The classical endocytosis pathway of TfR. 10
I-4. X-ray crystal structure of the ectodomain of the TfR. 12
I-5. The Rab protein cycle. 21
I-6. Structure of BFA. 29
I-7. Structure of AG10. 32
II-1. A diagram for pulse-chase assay. 40
II-2. The effect of AG10 on TfR-mediated transcytosis
in Caco-2 and MDCK cells. 42-43
II-3. The effect of AG10 and BFA on transcytosis and
recycling of Tf in MDCK cells. 45-46
II-4. At early time points, AG10 did not show any effect
on Tf recycling in MDCK cells. 49
II-5. At early time points, a statistically significant increase in Tf
recycling was observed in Caco-2 cells in the presence of AG10. 50
III-1. FITC labeling reaction diagram. 63
III-2. A diagram for confocal z-sectioning. 68
III-3. A schematic illustration of the quantitative analysis of
colocalization between FITC-Tf and Rab proteins. 70
III-4. Colocalization between FITC-Tf and Rab5 in
Caco-2 cells pulsed at 16°C or 37°C. 74
x
III-5. Colocalization between FITC-Tf and Rab11 in
Caco-2 cells pulsed at 16°C or 37°C. 75
III-6. The quantitative analysis of the colocalization between
FITC-Tf and Rab5. 77
III-7. The quantitative analysis of the colocalization between
FITC-Tf and Rab11. 78
III-8. The location of colocalization between FITC-Tf and
Rab5 relative to the tight junction. 81
III-9. The location of colocalization between FITC-Tf and
Rab11 relative to the tight junction. 82
III-10. The level of colocalization between FITC-Tf and Rab5 was
maximal in the z-sections localized 5.6 μm to 6.4 μm above
the tight junction of Caco-2 cells pulsed at 16°C. 84
III-11. The highest level of colocalization between FITC-Tf and
Rab11 was observed in the z-sections localized 2.0 μm to
2.8 μm above the tight junction of Caco-2 cells pulsed at 37°C. 85
III-12. AG10 did not show any significant effect on the recycling,
transcytosis, and cell association of Tf at 16°C. 87-88
III-13. AG10 increased both recycling and transcytosis of Tf but
decreased cell association of Tf significantly at 37°C. 89-90
III-14. The effect of AG10 on the Tf uptake. 92
III-15. The effect of AG10 on the integrity of the
tight junction in Caco-2 cells. 94
III-16. The effect of AG10 on the colocalization
between FITC-Tf and Rab11. 96
III-17. The quantitative analysis of the colocalization between
FITC-Tf and Rab11 in the absence and presence of AG10. 97
xi
III-18. The effect of AG10 and BFA on the apical recycling of Tf
in Caco-2 cells. 99
III-19. The effect of BFA on the colocalization between FITC-Tf
and Rab11 in Caco-2 cells. 101-103
III-20. The quantitative analysis of the colocalization between FITC-Tf
and Rab11 in the absence and presence of BFA or AG10. 105-106
IV-1. The effect of AG10 on the recycling, transcytosis, and cell
association of the basolaterally endocytosed Tf in Caco-2 cells. 122-123
IV-2. Colocalization between the basolaterally endocytosed FITC-Tf
and Rab11 in the absence and presence of AG10 in Caco-2 cells. 126-128
IV-3. The presence of the apically endocytosed FITC-Tf in the Rab11-
containing endosomal compartment in Caco-2 cells. 129
IV-4. The basolaterally endocytosed FITC-Tf was colocalized
with Rab11 in AG10-treated Caco-2 cells. 130
IV-5. The quantitative analysis of the colocalization between the
basolaterally endocytosed FITC-Tf and Rab11 in the
absence and presence of AG10. 132-133
V-1. Biotin labeling reaction diagram. 144
V-2. Determination of Biotin-Tf concentration in serum. 147
V-3. The serum concentration of Biotin-Tf conjugate reached the
maximum at 8-12 hours after the oral administration. 149
V-4. The serum concentration of Biotin-Tf conjugate increased
in AG10-treated CF-1 mice. 151
xii
List of Schemes
III-1. Proposed model for the intracellular trafficking pathways
of the apically endocytosed Tf. 108
IV-1. Proposed model of the TfR trafficking pathways. 136
xiii
Abbreviations
AEE Apical early endosome
AG10 4-Hydroxybenzylidenemalononitrile; ∝-Cyano-(4-
hydroxy)cinnamonitrile; Tyrphostin A8
ARE Apical recycling endosome
BEE Basolateral early endosome
BFA g,4-Dihydroxy-2-[6-hydroxy-1-heptenyl]-4-cyclopentanecrotonic
acid 1-lactone; Cyanein; Brefeldin A
BHK Baby hamster kidney
Caco-2 Colorectal adenocarcinoma
CE Common endosome
DAPI 4′,6-diamidino-2-phenylindole, dilactate
ELISA Enzyme-linked immunosorbent assay
FITC Fluorescein isothiocyanate
HABA 4-Hydroxyazobenzene-2-carboxylic acid
HAC Acetic acid
Hela Human cervix carcinoma
HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
IRE Iron-response elements
IRP Iron-regulatory proteins
LDL Low-density lipoprotein
MDCK Madin-Darby Canine Kidney
PBS Phosphate-buffered saline
PIgR Polymeric immunoglobulin receptor
Tf Transferrin
TfR Transferrin receptor
xiv
Abstract
Transferrin receptor (TfR) targeted therapy has emerged as a drug delivery tool
which can lead to the delivery of therapeutic agents across the endothelial and
epithelial barriers of choice. The potential application of the TfR as a delivery
vehicle for the transport of macromolecular drugs across intestinal epithelial cells
is limited by several factors, including the low level of TfR-mediated transcytosis.
The GTPase inhibitor, AG10, has been shown previously to increase the apical-to-
basolateral transcytosis of Tf in Caco-2 cells.
My goal has been to improve the GI absorption of the orally administered Tf-protein
conjugates. For this reason the trafficking pathways of Tf was studied in
order to get insight into the cellular machinery that is involved in the intracellular
trafficking of Tf. Elucidating the role of the intracellular controllers of the
recycling and transcytosis of Tf, such as Rab proteins, is potentially important in
developing Tf as a vehicle for oral delivery of proteins.
In this thesis the intracellular trafficking of Tf through the endosomal
compartments and the effect of AG10 on Tf trafficking in Caco-2 cells are
described. The biochemical studies parallel with confocal fluorescence
microscopy studies revealed new findings regarding the Tf trafficking pathways
xv
in Caco-2 cells. The results from these studies indicate that the Rab5-containing
endosomal compartment and the Rab11-containing endosomal compartment are
involved in Tf trafficking at 16°C and 37°C, respectively. The Rab5-containing
endosomal compartment is localized far from the tight junction and close to the
apical membrane. While the Rab11-containing endosomal compartment, which
appeared as a population of cup-shaped vesicles, is localized above the tight
junction. This compartment may be analogous to the apical recycling endosome
(ARE) in MDCK cells. Furthermore, Rab11 is involved in the recycling of the
apically endocytosed Tf in Caco-2 cells. The basolaterally endocytosed Tf does
not have access to the Rab11-containing endosomal compartment. AG10
stimulates the transport of the apically and basolaterally endocytosed Tf to the
Rab11-containing endosomal compartment. The existence of an internal pool for
the TfR in Caco-2 cells may explain the effect of AG10 on the Tf trafficking in
Caco-2 cells. AG10 may increase the Tf transport from the TfR internal pool to
the late endosomal compartments involved in the transcytosis and recycling of Tf
such as the Rab11-containing endosomal compartment. This could result in an
enhancement of both recycling and transcytosis of Tf in Caco-2 cells.
Furthermore, the ability of AG10 to inhibit the GTP-utilizing enzymes raises the
possibility of its inhibitory effect on GTPase proteins such as Rab proteins. AG10
may inhibit the intrinsic GTPase activity of Rab11, resulting in a GTP-bound
active form of Rab11. The presence of an active form of Rab11, which is
xvi
involved in the apical recycling of Tf, may result in the enhancement of the Tf
recycling in the presence of AG10.
In addition, the effect of AG10 on the Tf transport was studied. The data showed
that AG10 is a selective enhancer of the TfR-mediated transcytosis in Caco-2
cells. AG10 is also a considerable enhancer of the intestinal absorption in CF-1
mice. Low potential side effects have been associated with the use of AG10 as an
enhancer of Tf transport both in vitro and in vivo.
Overall, the results described in this thesis provide information regarding the
intracellular processing of Tf in Caco-2 cells, data related to the effects of AG10
on the Tf intracellular processing in Caco-2 cells, data supporting a possible
mechanism for the action of AG10 in Caco-2 cells, and information supporting
AG10 as an enhancer of choice for the TfR-mediated transcytosis in vitro and in
vivo.
xvii
Preface
Significance
This research project is focused on transferrin (Tf) and transferrin receptors
(TfRs), which have been considered as a potential target to enable transepithelial
delivery of large bioactive molecules that would normally have negligible
bioavailability when administered orally. One of the problems associated with
this approach is the fact that most of the surface TfR in differentiated epithelial
cells lies at the basolateral membrane. To overcome this problem enhancers of
Tf-conjugate transport have been examined. AG10 has been shown previously to
increase the apical-to-basolateral transcytosis of Tf in Caco-2 cells. To further
understand its mechanism, the intracellular trafficking of Tf was defined. The
elucidation of the Tf trafficking pathways and the role of the intracellular
controllers of the recycling and transcytosis of Tf, such as Rab proteins, is
potentially important in developing Tf as a vehicle for oral delivery of proteins.
By utilizing the knowledge of the intracellular sorting and recycling pathways of
the TfR, including Rab protein-mediated processes, the delivery of peptide-based
therapeutics can be maximized. Additionally, AG10 was demonstrated as a
considerable enhancer of the TfR-mediated transcytosis in vitro and intestinal
absorption in vivo.
1
Chapter I. Background Information and Review of the
Literature
To study the intracellular trafficking of transferrin (Tf) and its implication in drug
delivery, it was necessary for me to review some aspects related to the structure
and function of Tf and the transferrin receptor (TfR). The regulation of the TfR
expression, the tissue distribution of the TfR, the intracellular trafficking of Tf,
and the utilization of the TfR in the targeting and delivery of therapeutic agents
are also discussed in this chapter. Due to the low level of TfR-mediated
transcytosis, enhancers of TfR-mediated transcytosis such as brefeldin A (BFA)
and tyrphostin A (AG10) have been used to increase Tf transcytosis in vitro and
in vivo. The effects of BFA on different intracellular organelles have been studied
extensively, but the mechanism of increased transcytosis in the presence of AG10
is largely unknown. I have been using AG10 to investigate the intracellular
trafficking of Tf in polarized cells, particularly in Caco-2 cells. My findings also
provide information to explain the effect of AG10 on the intracellular trafficking
of Tf in Caco-2 cells. Information provided by this study is potentially important
in developing Tf as a vehicle for oral delivery of protein-based therapeutics.
2
1. Structure and function of transferrin
Tfs with a molecular size of 80 kDa are a structurally related class of metal-binding
glycoproteins (4). They have been classified into three major sub-classes:
serum Tf, lactoferrin, and ovotransferrin. Serum Tf is responsible for binding and
transporting iron through many of the biological fluids including blood, lymph,
cerebrospinal fluid, colostrum, bile, amniotic fluid, and breast milk (15; 16; 34;
35; 65). Lactoferrin is found in breast milk, saliva, and tear secretions (8; 99).
Unlike serum Tf, lactoferrin does not contribute iron to erythrocytes and is
cleared by the reticulo-endothelial system (18). It is thought to function as an iron
chelator; sequestering iron to act as a bacteriostatic agent (99). Ovotransferrin is
found in reptile, avian oviduct secretions and avian egg white (44). It primarily
serves as a bacteriostat via the sequestration of iron (47; 137). In addition to the
three major types of Tf, melanotransferrin (p97) can also be considered to be a
part of the Tf family. This fourth member of the Tf family is a structural
homologue, distinctly different from the others in that it seems to have little to do
with the receptor-specific cellular uptake of iron (40; 141). Melanotransferrin is
predominantly localized in the surface of human melanomas (19; 89). It may
assist in the rapid cell proliferation via iron scavenging and the prevention of lipid
peroxidation (89).
3
Serum Tf has been identified as a potential drug carrier candidate, and has also
been used extensively in the research focusing on endocytosis. The serum Tf
from different species shares a high degree of homology. It is typically composed
of a single chain polypeptide with approximately 700 amino acids with varying
degrees of carbohydrate content between species (50). The carbohydrate chains
of Tf do not affect the ability of Tf to bind to TfR, nor do they affect the ability of
the Tf-TfR complex to be internalized into the cells (30; 115). The Tf molecule is
composed of two homologous globular domains known as the amino-terminal
lobe (N-lobe, residues 1-331 in human Tf [hTf]) and carboxy-terminal lobe (C-lobe,
residues 339-679 in hTf). Each lobe is further separated into two
subdomains; N1 (1-92, 247-331), N2 (93-246), C1 (339-425, 573-679), and C2
(426-572) (Figure I-1) (163). The lobes are connected by a short linear peptide
region (residues 323-338 in hTf) (163). In addition to a loose, predominantly
hydrophobic interface between the lobes, there are salt bridges between the N-lobe
and C1-subdomain (Figure I-2).
6
As shown in Figure I-2, Arg308 in the N1-subdomain interacts with Asp376 in
the C1-subdomain and Asp240 in the N2-subdomain interacts with Arg678 in the
C1-subdomain. Table I-1 shows the sequence alignments for the regions of hTf
which bind the human transferrin receptor (hTfR) with high affinity. The exact
orientation of Tf binding to TfR is not fully understood, however the primary
recognition site appears to reside on the C-lobe of Tf. In addition, there is a
sequential synergistic effect between the N-lobe and C-lobe of Tf that enhances
affinity for TfR. In the C1-subdomain, 11 residues were identified in the cryo-
EM model of hTf bound to TfR (Table I-1) (27). These residues, which are in
contact with the helical region of the TfR, lie between His349 and Glu372, and
include Glu367, Val369, Ser370, and Glu372, all of which are part of an epitope
recognized by a monoclonal antibody (Mab) designated E-8 (27; 105; 106).
Interestingly, the MAb blocks binding of hTf to the TfR on HeLa S3 cells, and it
binds to diferric hTf with two-fold higher affinity compared to apo-hTf (105;
163).
8
These results are consistent with a conformational change in this region of the C-lobe,
which is required for the residue His349 to come into the contact with TfR
(163). In the N2-subdomain, the sequence of residues Pro142-Pro145 is
completely conserved in the Tfs that bind to the TfR, and is poorly conserved in
the Tfs that do not. The residues Pro142-Arg143-Lys144-Pro145, appear to make
contact with the residues Tyr123-Trp124-Asp125 of TfR (163).
Both lobes of the Tf molecule are capable of binding one iron atom with similar
affinities. The binding of Fe3+ alters the shape of the Tf molecule. This
conformational change, which seems to occur in a stepwise fashion as each iron
molecule is bound to Tf (25; 78; 103), affects the binding affinity of Tf for its
receptor. The iron-loaded Tf, known as holo-Tf, binds to the TfR with greater
affinity than mono-Tf and apo-Tf (66; 182). Binding of iron to the Tf molecule
also increases the stability of the Tf molecule, as a result holo-Tf, relative to the
apo-Tf, exhibits an increased resistance to the thermal and proteolytic degradation
(122; 181). In addition, the presence of bicarbonate anions is necessary for the
binding of iron to Tf (3; 11; 139). The bicarbonate anions allow Fe3+ to tightly
coordinate with the amino acids that are suspected of comprising the iron binding
pocket (i.e., two Tyr, one Lys, and one His for each lobe) of both lobes of the Tf
molecule (80; 102).
9
Presence of the bicarbonate anions also affects the ability of Tf to properly release
iron once it is bound. The bicarbonate-free Tf demonstrates poor iron release
capabilities (33; 140). The Tf molecule also undergoes a conformational change
as a result of a trigger-mechanism that is induced by the decrease in pH during the
endocytosis cycle. This allows the release of iron from the Tf-TfR complex, and
the subsequent recycling of the apo-Tf-TfR complex to the cell surface. The
lowered affinity of the apo-Tf for the receptor at extracellular pH causes the
release of Tf from its receptor (51; 58; 140). A schematic of the endocytosis and
the recycling cycle for the Tf-TfR complex is shown in Figure I-3.
The liver is the principle site of production and catabolism of Tf. This can be
demonstrated in the case of human liver transplant patients. After receiving a new
liver, they obtain the electrophoretic profile of the donor’s serum Tf, which is
determined by the inter-individual variability in post-translational modification of
Tf (2; 5).
11
2. Structure and function of transferrin receptor
The TfR is a homodimer composed of two identical subunit monomers with an
approximate mass of 90 kDa each (57). The monomers are covalently linked by
two disulfide bonds at residues Cys 89 and Cys 98 (81). The TfR also has a
single transmembrane domain, which is comprised of residues 68-88 of the
polypeptide followed by a short cytoplasmic tail (81). The cytoplasmic tail,
which is 5 kDa in size, is the site of the phosphorylation and acylation of the
receptor (81). The site-directed mutagenesis studies have shown that the central
helical domains of TfR, formed by the union of the two monomers, are the key to
the ability of the receptor to recognize Tf (32). The ectodomain of the receptor is
composed of three distinct domains; the protease-like, the apical, and the helical
domains (Figure I-4). The helical domains of the TfR face each other and interact
to form the dimeric receptor (92).
13
The inspection of the computational union of Tf and the TfR crystal structures has
indicated that much favorable interaction occurs between the outward facing
apical domains of the TfR and the Tf molecule (92). The extracellular portion of
the TfR is subject to an extensive post-translational modification with three sites
of mannose rich N-linked glycosylation at residues Asn251, Asn317, and Asn727,
and one O-linked glycosylation on Thr104 (121). Mutations that eliminate the
sites of potential N-linked glycosylation severely limit the ability of the receptor
to bind to Tf.
The TfR was first recognized through the formation of monoclonal antibodies
raised in mice that were immunized against various neoplastic cell lines. Several
of the antibodies recognized antigenic portions of the protein which we now know
as TfR1 (49; 120; 158). Within the past few years a new member of the TfR
family was discovered and named as TfR2. In contrast to TfR1, TfR2 is primarily
expressed in the liver. Analysis of the amino acid sequence suggests that TfR2 is
a type II transmembrane glycoprotein, like TfR1, sharing 45% sequence identity
and 66% similarity in the extracellular domain (84). TfR2 interacts with Tf in a
similar fashion as TfR1. Specifically, TfR2 exhibits changes in binding affinity to
Tf as a result of environmental pH and iron saturation status of Tf. TfR2 does not
bind very well to apo-Tf at physiological pH, demonstrating a preference for
14
holo-Tf, while an acidic milieu results in preferential binding to the apo-Tf (82).
However, unlike TfR1, TfR2 has a much lower affinity for Tf (25 fold lower than
TfR1) and it seems to be primarily involved with regulation and maintenance of
iron homeostasis in the body (84). Mutations in TfR2 have been associated with a
form of hereditary hemochromatosis that is not associated with forms of
hemochromatosis that arise from mutations in the hemochromatosis protein
(HFE) (22; 29; 135; 136), suggesting a role for TfR2 in excessive iron uptake. In
addition, TfR2 has also been implicated as a regulator of hepcidin expression; a
modulator of dietary iron uptake (39). TfR2 also lacks iron-regulatory elements
in the coding and noncoding regions of its mRNA, indicating that the expression
of TfR2 is not regulated by the iron-responsive protein feedback mechanism as
seen for TfR1 (82).
15
3. Regulation of expression and tissue distribution of transferrin receptor
The expression of TfR1 is regulated at the post-transcriptional level by the
interactions between iron-regulatory proteins (IRP) and iron-response elements
(IRE) (1; 24). The IRE are portions of the 3´-untranslated portion of TfR mRNA
that form secondary structural features in the shape of five hairpins (1; 24). The
IRE of TfR mRNA are recognized and bound by the IRP, thus controlling the
extent of TfR mRNA translation and stability (94). Under iron deprivation
conditions, IRP bind to the IRE increasing the stability of the mRNA and thus up-regulating
the expression of TfR (36; 118; 132).
There are two distinct IRP, known as IRP1 and IRP2, that interact with IRE of
TfR mRNA. They respond to variations in iron concentration via different
mechanisms (59; 94). IRP1 has been considered to be a bifunctional enzyme.
When the cellular iron level is high, it has no RNA binding ability. It switches to
a predominantly RNA binding protein when cellular iron levels are low (10; 17;
28; 56; 119). IRP1 also shows dual functionality in regard to the regulation of
intracellular iron concentrations. IRP1 up-regulates the TfR production by
binding to the IRE of the TfR, but there is evidence suggesting that it also binds
the 5′-region of ferritin mRNA to inhibit the complete translation of the ferritin
16
protein in iron-rich conditions (13; 60; 61; 168). A down-regulation of ferritin
would conceivably result in the complementary liberation of free iron from
intracellular storage. In contrast to IRP1, IRP2 synthesis is directly affected by
the intracellular iron levels. When the iron stores are plentiful, IRP2 rapidly
undergoes ubiquitination and is degraded in the cellular proteosome machinery
(77). IRP2 is up-regulated by the renewed production of non-proteosome directed
protein in conditions where intracellular iron levels are low (54; 55).
The regulation of TfR2 expression is not well understood at the cellular level. In
contrast to TfR1, chelation of iron does not up-regulate the expression of TfR2
(155), and in several instances TfR2 behaves in a manner that is opposite to TfR1.
For example, in liver development, TfR2 is up-regulated, while TfR1 is down-regulated.
But during erythrocytic differentiation of murine erythroleukemia
cells, expression of TfR1 steadily increases, while TfR2 expression is down-regulated
(83).
17
4. Intracellular trafficking of transferrin
The intracellular trafficking of proteins such as Tf has been studied extensively.
Tf internalized from the plasma membrane via TfR-mediated endocytosis is first
delivered to the peripheral early sorting endosomes via clathrin-coated vesicles in
non-polarized cells. In the acidic endosomal environment, the iron is dissociated
from Tf and the apo-Tf-TfR complex is targeted to the recycling compartment.
From this compartment, the apo-Tf-TfR complex is recycled back to the plasma
membrane where apo-Tf is released from the TfR (112; 138). The endocytosis
and recycling cycle for the Tf-TfR complex is illustrated in Figure I-3. The
recycling compartment contains recycling receptors including the TfR and the
low-density lipoprotein receptor (LDLR), but not lysosomally directed ligands
such as LDL or α2-macroglobulin. This endosomal compartment has been
characterized as a tubular network, in contrast to the tubulovesicular sorting
endosome, and is concentrated in the perinuclear area of non-polarized cells
including CHO, AtT20, and Hela cells (46; 52; 111; 113; 156; 180). It has been
reported that Tf-TfR complexes recycle to the plasma membrane with the same
kinetics as certain lipids in CHO cells (110). Temperatures lower than 20°C have
been shown to block the transport of the endocytosed ligand to the perinuclear
region of the cells (38).
18
In polarized cells, the cell surface is separated into apical and basolateral plasma
membrane domains. They have different lipid and protein compositions, and
endocytosis can occur from both domains (116; 125). In polarized cells, the
basolaterally endocytosed TfR is efficiently sorted from the transcytotic pathway
and recycled back to the basolateral membrane. Polarized epithelial cells contain
distinct populations of apical early endosomes (AEE) and basolateral early
endosomes (BEE). The majority of Tf (65%) was found to be recycled from
BEEs within 10 minutes after internalization from the basolateral membrane in
Madin-Darby Canine Kidney (MDCK) cells (143). No direct fusion between the
AEEs and BEEs was observed using an in vitro assay (64). The existence of an
endosomal compartment, which is accessible to both apically and basolaterally
endocytosed ligands, has been reported in the polarized epithelial cells (68; 87;
166). In addition, transcytosing proteins such as IgA and recycling proteins such
as Tf have been found in a "common" endosome (CE) which partly consists of
tubules (117). This compartment which seems to be an ubiquitous feature of the
polarized epithelial cells is termed the common recycling endosome in MDCK
cells (144) or the subapical compartment (SAC) in hepatocytes (76; 131; 151;
160).
19
The CE has been identified as a tubularized endosomal system with a diameter of
60 nm in MDCK cells, and it is from there that the basolaterally endocytosed Tf-
TfR complex recycles to the basolateral membrane (7; 41). On the other hand, the
basolaterally internalized complex formed by the polymeric immunoglobulin
receptor (pIgR) and dimeric IgA (dIgA) on its route from the basolateral to the
apical surface crosses the CE and moves to a population of 100-150 nm cup-shaped
vesicles above the CE which constitutes the apical recycling endosome
(ARE) in MDCK cells (20; 48). It may reach the apical plasma membrane
through the Tf recycling pathway. In hepatocytes, the SAC has been
characterized as a subcompartmentalized endosomal compartment which is
composed of 60-100 nm tubulovesicular structures (analogous to the CE in
MDCK ) and 150-200 nm cup-shaped vesicles in the apical region (analogous to
the ARE in MDCK) (131). The relation between the early endosomes (EEs) and
the recycling endosomes (REs) is poorly understood. However, the REs are
competent to communicate with both apical and basolateral early endosomes (64).
20
5. Rab proteins involved in Tf trafficking
In the past decades Rab proteins, which belong to a family of monomeric
GTPases with a molecular mass of 20-40 kDa, have been studied extensively.
Rab proteins have two GDP-bound inactive and GTP-bound active forms (153).
An upstream signal stimulates the dissociation of GDP from Rab proteins. This is
followed by the binding of GTP. The GDP/GTP exchange reaction is catalyzed
by GEF (guanine exchange factor). The GTP-bound form is converted to the
GDP-bound form by the action of the intrinsic GTPase activity of Rab proteins
(153). A schematic of the Rab cycle is shown in Figure I-5.
22
Rab proteins are involved in different intracellular processes including regulation
of vesicular trafficking pathways (153). The existence of three endosomal
populations in A431 cells have been reported. These endosomal compartments
contain only Rab5, both Rab5 and Rab4, and both Rab4 and Rab11, respectively
(150). The biochemical analysis of Rab5 and Rab11 enriched endosomes in CHO
cells has revealed that Tf traffics through these endosomes (157). Rab5 regulates
the early endocytic trafficking events in the cells and is localized to the EE (21;
26; 38). Tf has been shown to colocalize with Rab5 in clathrin-coated vesicles
and within Rab5-positive endosomes (150; 157). On the other hand, Rab11 is
involved in the recycling of internalized Tf in non-polarized cells (159), and both
dominant positive (Q70L) and dominant negative (S25N) mutant forms of Rab11
inhibit Tf recycling from perinuclear RE in CHO cells (134).
In polarized epithelial cells, certain Rab proteins (Rab4, Rab17 and Rab25)
localize to the CE or to specific regions of the SAC. Rab4 is involved in
regulating TfR recycling in MDCK cells (114). Rab17 associated with the CE,
which is insensitive to microtubule disruption, may play a role in the recycling
and transcytosis of membrane proteins. Rab17 has been shown to colocalize with
the apically and basolaterally endocytosed pIgR-dIgA complexes in MDCK cells.
The overexpression of Rab17 impairs the basolateral-to-apical transcytosis of
23
dIgA (70). Rab17 has also been shown to colocalize with the basolaterally
internalized Tf in polarized Eph4 cells, and expression of Rab17 mutant proteins
(Q77L and N132I) increases the basolateral-to-apical transcytosis of Tf (183). In
non-polarized BHK-21 cells, Rab17 colocalizes with internalized Tf in the
perinuclear RE (183). In MDCK cells, Rab11a and Rab25 are associated with the
ARE that is dependent on intact microtubules for its integrity (23; 167). The
basolaterally internalized Tf gains access to the Rab11a and Rab25-positive ARE
upon the treatment of MDCK cells with BFA (167). BFA missorts Tf to ARE and
blocks Tf recycling to the basolateral cell membrane.
24
6. Utilization of transferrin receptor for drug delivery and drug targeting
The large size and charge of protein drugs prevent the passage across epithelial
barriers. The TfR has emerged as a potential mediator to enable the transport of
these large molecules across the endothelial and epithelial barriers, such as the
blood brain barrier (BBB). The BBB effectively excludes many therapeutic drugs
including almost all peptide and protein-based therapeutics. Since the capillaries
of the brain have been shown to have relatively high levels of the TfR, targeting
of the TfR for drug delivery has been seen as an attractive strategy. However,
under physiologic conditions, the TfR of the BBB are postulated to be nearly
saturated with Tf due to the relatively high amounts of Tf in the serum (67).
Therefore, utilizing a Tf-based drug conjugation approach is undesirable. As a
result antibodies such as OX26, which specifically recognize the TfR with or
without the presence of Tf on the receptor, are considered as a viable alternative
to target the TfR of the BBB. OX26, a monoclonal antibody for rat TfR, has the
ability to preferentially target the TfR of the BBB (79). It binds to an extra-cellular
region that is distinct from the binding pocket for Tf. Therefore it is
postulated to not interfere with the normal functioning of iron uptake via the TfR
(79). Upon binding to the TfR at the BBB, OX26 subsequently exhibits superior
transport abilities into the CNS. Studies performed with isolated bovine brain
25
capillaries demonstrated that nearly 50% of bound radiolabeled OX26 was taken
up via endocytosis during a two hour incubation period (123). Several promising
OX26 conjugates have been created using vasoactive intestinal peptide analog
(14), brain-derived neurotrophic factor (BDNF) (124; 184; 185), and basic
fibroblast growth factor (bFGF) as the therapeutic cargo (149; 175). When
BDNF-PEG2000-SA-OX26 was administered to rats for one week after an
administered episode of transient forebrain eschemia, a neuroprotective effect was
observed. Specifically, the neuronal density in the hippocampus decreased 68%
compared to the control (174). TfR based transcytosis has also been explored for
targeted gene delivery to the CNS. High levels of reporter gene expression have
been observed when plasmids encoding for luciferase or β-galactosidase were
enclosed in liposomes which were conjugated to OX26 via a PEG2000-thio-ether
linkage and administered intravenously to rats (73; 74; 145).
In addition to the use of Tf-drug conjugates to target therapeutics to certain tissues
from the blood stream, Tf based conjugates have also been considered for the
systemic non-invasive delivery of therapeutics. Several absorptive epithelia
including the small intestine (142), and type II pneumocytes (170) have been
shown to express the TfR which provides a potential target to enable
transepithelial delivery of large bioactive molecules that would normally have
negligible bioavailability when administered orally.
26
The transport of an insulin-Tf conjugate via TfR-mediated transcytosis was
demonstrated across Caco-2 cells (142). The results from the cell culture
experiments have been further validated by the in vivo uptake studies. In
streptozotocin-induced diabetic rats, the orally administered insulin-Tf conjugate
demonstrated an extended duration of action, with hypoglycemic effects observed
11 hours after administration (177). One of the problems associated with this
approach is that the majority of the surface TfR in differentiated epithelial cells is
present at the basolateral membrane (164). This would tend to preclude the
ability to achieve apical-to-basolateral transcytosis of Tf-drug conjugates. While
there are indications that transepithelial transport may still be feasible, especially
in areas where large numbers of semi-differentiated cells may occur such as the
villus-crypt of the small intestine, enhancers of Tf-conjugate transepithelial
transport have been examined as a means to overcome the problem of a low level
of apical-to-basolateral transcytosis.
The TfR is more abundantly expressed in rapidly dividing cells than quiescent
cells (90; 152) because of its pivotal role in iron uptake, and the absolute
requirement for iron in rapid cell proliferation. High levels of the TfR expression
have also been demonstrated in many tumors (12; 43; 98; 128; 130; 133) and
importantly, studies have shown that the TfR is expressed more abundantly in
27
malignant tissues than their normal counterparts (37; 43; 128; 146). Therefore, the
TfR expressed on tumor cells has been seen as a suitable target for the delivery of
therapeutics into cancer cells by receptor-mediated endocytosis. A great variety
of cytotoxic agents including methotrexate, doxorubicin, ricin A, daunorubicin,
and toxin CRM107 have already been conjugated to Tf and investigated as
potential anti-cancer therapeutics (75; 88; 91; 154). Conjugation of anti-cancer
therapeutics to Tf reduces the toxicity in the undesired tissues, and also enhances
the effectiveness of these agents in many multi-drug resistant cell lines. For
example, Tf-doxorubicin conjugate was 5 to 10 times more effective than
doxorubicin alone in killing doxorubicin-resistant cell lines, whereas in
doxorubicin-sensitive cell lines the Tf-doxorubicin conjugate was only 4 to 5
times more effective than doxorubicin (75; 148).
28
7. Enhancers for transferrin receptor-mediated drug delivery
To overcome the limitation of TfR-mediated transcytosis at such a low rate of
transport, enhancers of the apical-to-basolateral Tf transcytosis such as BFA and
AG10 have been studied.
a. Brefeldin A
Brefeldin A (BFA), which is a fungal metabolite, has been extensively used in the
study of intracellular trafficking events (85). Figure I-6 shows the structure of
BFA from Penicillium brefeldianum. BFA enhances the transcytosis of Tf, and it
has been shown to have an effect in modulating the trans-Golgi network. It
causes a rapid rearrangement of the Golgi through microtubule dependent
processes to redistribute in a retrograde fashion back to the endoplasmic reticulum
(31; 96). In addition, the trans-Golgi network (TGN) tubulates in the presence of
BFA forming a fused trans-Golgi/endosomal network (97; 173). While the
lysosomes do not appear to fuse with the trans-Golgi/endosomal tubulates, they
do appear to form an extensive microtubule dependent network in the presence of
BFA (97). BFA also blocks the anterograde transport of sorted proteins from the
endoplasmic reticulum (31; 96). BFA does not cause any change in intra-lysosomal
or intra-endosomal pH (97).
30
In polarized epithelial cells, BFA treatment increases the basolateral-to-apical
transcytosis of the internalized TfR and the bulk plasma membrane (129).
However, the Golgi of some cell lines, including MDCK cells, seem to be
resistant to the effects of BFA (71). Nonetheless, BFA causes significant
alteration of trans-Golgi and endosomal function in MDCK cells (71). The fact
that BFA treatment results in a significant enhancement of TfR-mediated
transcytosis in MDCK cells would indicate that the means by which this
enhancement occurs are dependent upon the sorting events at the TGN/endosome
level. BFA also alters the intracellular sorting of internalized proteins in
polarized epithelial cells. For example, BFA treatment causes the missorting of
basolateral LDL receptors (108) and reduces the efficiency of pIgRs to the cell
surface (6). The effects of BFA on protein sorting appear to be much more
sensitive at the apical surface as the concentration of BFA required to see effects
at the apical surface is much lower than for the basolateral surface (100; 162).
This is important when considering BFA as an enhancer of TfR-mediated drug
delivery. An enhancement of TfR-mediated transcytosis in the apical-to-basolateral
direction has been demonstrated in two highly polarized cell lines,
namely MDCK cells (164) and primary cultured rat type II pneumocytes (171).
In both cell lines, BFA treatment results in the specific enhancement of apical-to-basolateral
transport of Tf-insulin and Tf-GCSF (filgrastim) conjugates.
31
b. Tyrphostin A8 (AG10)
Tyrphostins are a group of low molecular weight organic compounds which have
been used to specifically inhibit the GTPase activity of protein tyrosine kinases
(45; 101; 179). AG10 (tyrphostin A8, or 4-hydroxybenzylidenemalononitrile) has
been shown previously to increase the apical-to-basolateral transcytosis of Tf in
Caco-2 cells. Figure I-7 shows the structure of AG10. The effect of AG10 on the
transport of the insulin-Tf conjugate has been demonstrated both in vivo and in
vitro (176). The enhancement of the transport of the insulin–Tf conjugate was
determined to be dependent upon TfR-mediated processes since an insulin-albumin
conjugate demonstrated no transport across the monolayers, and
treatment with AG10 had no effect on its transport properties (176). In addition,
AG10 exhibited the ability to enhance the hypoglycemic effect of the orally
administered insulin-Tf conjugate relative to the BFA-treated and enhancer-free
insulin-Tf treatment groups (176). In addition to inhibition of protein tyrosine
kinases, AG10 was also reported to inhibit the GTPase activity of transducin and
calcineurin (104; 172). The mechanism involved in the enhancement of TfR-mediated
transcytosis in the presence of AG10 is still largely unknown. However,
the ability of tyrphostins to inhibit the GTP-utilizing enzymes raises the
possibility of their inhibitory effect on GTPase proteins such as Rab proteins.
33
Chapter II. The Selectivity of the Effect of AG10: AG10 increases
the transferrin transcytosis in Caco-2 cells, but not in MDCK cells
1. Abstract
In this chapter, the effect of AG10 on the trafficking of transferrin in Caco-2 cells
and MDCK cells was investigated. Transport studies showed that AG10 increases
the transferrin transcytosis only in Caco-2 cells. The effect of AG10 and BFA on
transcytosis and recycling of transferrin in MDCK cells was determined in the
cells which were pulsed at 37°C for 2 hours and chased at 37°C for 2 hours.
AG10 did not show any effect on the transferrin trafficking in MDCK cells, which
is consistent with the results obtained from the transport studies. BFA decreased
the transferrin recycling and increased the transferrin transcytosis in MDCK cells,
suggesting that BFA enhances the transferrin receptor-mediated transcytosis in a
manner that is different from the method of action of AG10. The effect of AG10
on transferrin recycling was studied in both Caco-2 and MDCK cells which were
pulsed at 37°C for 30 minutes and chased for 30 minutes. AG10 did not show
any effect on the transferrin recycling in MDCK cells. The transferrin recycling
increased in the presence of AG10 in Caco-2 cells. The differences in the results
obtained from pulse-chase assays, performed in MDCK and Caco-2 cells in
34
the absence or presence of AG10, may be due to the morphological differences
between these two cell lines. The existence of an internal pool for the transferrin
receptor in Caco-2 cells may explain the enhancement in transcytosis and
recycling of transferrin in AG10-treated Caco-2 cells. The existence of different
sorting mechanisms for the transferrin receptor in these two cell lines may also be,
but less likely, the cause of the differences in the data obtained in AG10-treated
MDCK cells versus AG10-treated Caco-2 cells.
35
2. Introduction
The targeting of the TfR for the delivery of large bioactive molecules that would
normally have negligible bioavailability when administered orally, has been seen
as an attractive strategy. Several absorptive epithelia including the small intestine
(142) have been shown to express the TfR which provides a potential target to
enable the transepithelial delivery of the large bioactive molecules. The transport
of an insulin-Tf conjugate via TfR-mediated transcytosis was demonstrated across
Caco-2 cells (142). The results obtained from the cell culture experiments have
been further validated by the in vivo studies (177).
As I described in Chapter I, one of the problems associated with the targeting of
the TfR for the transepithelial delivery of large bioactive molecules is that the
majority of the TfR localized on the surface of differentiated epithelial cells is
present at the basolateral membrane (165). To overcome this problem, enhancers
of the Tf conjugate transepithelial transport such as BFA and AG10 have been
examined. In this chapter, I will discuss the effect of AG10 on the intracellular
trafficking of Tf in Caco-2 and MDCK cell lines.
36
3. Experimental
a. Cell Culture: Caco-2 cells and strain I MDCK cells were obtained from the
American Type Culture Collection (ATCC). The cells were grown in Dulbecco’s
Modified Eagle Medium (DMEM, Invitrogen) supplemented with 50 units/ml
penicillin, 50 μg/ml streptomycin, non-essential amino acids, L-glutamine, and
fetal bovine serum (FBS, 10%). AG10 and BFA were obtained from Calbiochem
and Sigma, respectively.
b. Preparation of diferric and radiolabeled Tf: To prepare diferric Tf, 100 μl
ferric ammonium citrate in H2O, with a stock concentration of 10 mg/ml, was
added to 1 ml of apo-Tf in PBS solution (20 mg/ml) under constant stirring. After
incubation of the mixture at 37°C for 2 hours, the iron loaded Tf was dialyzed
against 2 L PBS at 4°C overnight.
To radiolabel Tf, the diferric Tf was added to Na125I in PBS, followed by the
addition of freshly made chloramine T and sodium metabisulfite in the PBS
solution. The reaction was terminated by the addition of potassium iodide to the
reaction tube. The radiolabeled Tf was separated by gel filtration through a
10×0.5 cm Sephadex G-50 column.
37
c. Transport Assay using 125I-Tf: The two week old confluent Caco-2 or MDCK
cells grown on 24 mm Transwell filters (Costar, Cambridge, MA) were incubated
with serum free DMEM containing 0.1% BSA and 20 mM HEPES for 1 hour at
37°C to deplete endogenous Tf. Caco-2 and MDCK cells displayed a
transepithelial electrical resistance (TEER) of at least 0.1 Kohm/cm2 and 1.0
Kohm/cm2, respectively, which was measured by using an epithelial
voltohmmeter, EVOM (World Precision Instrument, West Haven CT). The cell
monolayers were further incubated in the presence and absence of AG10 (500
μM) or BFA (1.6 μg/ml,) for 2 hours at 37°C. 125I-Tf (1.5 μg/ml) was added to
the apical medium of the transwells in the presence or absence of AG10 or BFA.
At various time intervals, a fraction of the basal media containing the transcytosed
125I-Tf was collected and replenished with an equal volume of DMEM. The
collected samples containing 125I-Tf were subjected to 15% trichloroacetic acid
(TCA) precipitation and the radioactivity level in the protein precipitates and filter
membranes recovered in the end of the experiment was measured by using a
Packard gamma counter. The results were corrected for non-specific transcytosed
125I-Tf which was determined in the parallel experiments in the presence of 100-
fold excess of unlabeled Tf.
38
d. Pulse-chase Experiment using 125I-Tf: The two week old confluent Caco-2
and MDCK cells grown on 24 mm Transwell filters (Costar, Cambridge, MA),
which displayed a TEER of 0.1 Kohm/cm2 or 1.0 Kohm/cm2, were incubated with
serum free DMEM containing 0.1% BSA and 20 mM HEPES for 1 hour at 37°C
to deplete endogenous Tf. The cell monolayers were further incubated in the
presence and absence of AG10 (500 μM) or BFA (1.6 μg/ml) for 2 hours at 37°C.
125I-Tf (1.5 μg/ml) was bound to either the apical or basolateral surface of the
cells in the presence and absence of AG10 or BFA on ice for 15 minutes. This
was followed by an incubation at 37°C for 2 hours. After removing the unbound
and cell surface-bound 125I-Tf with ice-cold PBS (containing 0.1 mM Ca2+ and
0.05 mM Mg2+) and acid (0.15 M NaCl, 0.5 M HAC, pH; 2.2) washes on ice, the
cells were chased with or without AG10 or BFA and in the presence of 100-fold
excess of unlabeled Tf for up to 2 hours at 37°C. At the end of the chase period,
the samples from apical and basal media were collected. The collected samples
were subjected to 15% TCA precipitation and the radioactivity level in the protein
precipitates and filter membranes recovered in the end of the chase period was
measured by using a Packard gamma counter.
39
125I-Tf recovered from the basal chamber was taken as a measurement of Tf
transcytosis, while the radioactive protein recovered from the apical chamber was
used as a measurement of Tf recycling. The radioactivity in the filter membrane
reflected the internalized and cell membrane-bound 125I-Tf in the end of the chase
period. In the experiments in which the radiolabeled Tf was bound to the
basolateral membrane, the radioactivity recovered from the basal chamber was
taken as a measurement of Tf recycling. The results are presented as a percentage
of the total internalized 125I-Tf.
To look at the effect of AG10 and BFA at the early time points of the Tf
recycling, the cells were pulsed for 30 minutes at 37°C, and then chased for up to
30 minutes at 37°C. At indicated time points, the samples from apical and
basolateral media were collected. The collected samples were subjected to 15%
TCA precipitation and the radioactivity level in the protein precipitates and filter
membranes recovered in the end of the chase period was measured.
The results were corrected for non-specific recycled, transcytosed, and cell-associated
125I-Tf which were determined in the parallel experiments in the
presence of 100-fold excess of unlabeled Tf during the pulse period. The results
are presented as a percentage of the total internalized 125I-Tf. A schematic of
pulse-chase assay is illustrated in Figure II-1.
41
4. Results
a. AG10 increased the TfR-mediated transcytosis of Tf in Caco-2 cells, but it did
not show any effect on Tf transcytosis in MDCK cells. The TfR-mediated
transcytosis was measured in Caco-2 and MDCK cells in the absence or presence
of AG10 (500 μM). As shown in Figure II-2A, Tf transcytosis increased in
AG10-treated Caco-2 cells. The effect of AG10 was more prominent after 6
hours of treatment. On the other hand, no change in the extent of Tf transcytosis
was observed in AG10-treated MDCK cells (Figure II-2B). As shown in Figure
II-2C,D, the integrity of the tight junction, which was tested by measuring TEER,
was not affected by AG10 treatment in either cell line.
44
b. AG10 did not show any effect on the transcytosis and recycling of Tf in MDCK
cells. To further study the effect of AG10 on Tf intracellular trafficking in
MDCK cells, the recycling and transcytosis of Tf were measured in the absence or
presence of AG10 (500 μM) or BFA (1.6 μg/ml). In MDCK cells, which were
pulsed with 125I-Tf at either the apical or the basolateral membrane for 2 hours at
37°C and then chased for 2 hours at 37°C, no significant change in the recycling
and transcytosis of Tf was observed in the presence of AG10 (Figure II-3A-D).
On the other hand, BFA decreased both apical and basolateral recycling of Tf in
MDCK cells (Figure II-3B,D). BFA also increased the TfR-mediated transcytosis
in both directions in MDCK cells (Figure II-3A, C). The results are also
presented in Table II-1.
48
To obtain more information regarding the effect of AG10 on Tf intracellular
trafficking, Tf recycling was measured in MDCK and Caco-2 cells pulsed with
125I-Tf for 30 minutes at 37°C in the absence or presence of AG10 (500 μM) and
chased for 30 minutes at 37°C in the continued absence or presence of AG10.
Since AG10 did not show any effect on Tf recycling in MDCK cells pulsed with
125I-Tf at either the apical or basolateral membrane, 125I-Tf was internalized only
at the apical membrane of the cells. As expected, AG10 did not show any effect
on the recycling of Tf in MDCK cells (Figure II-4). On the other hand, Tf
recycling increased significantly in AG10-treated Caco-2 cells (Figure II-5).
51
5. Discussion
To study the selectivity of the effect of AG10, Tf intracellular trafficking was
studied in two different cell lines, Caco-2 and MDCK cells. AG10, which
increased TfR-mediated transcytosis in Caco-2 cells, did not show any effect on
TfR-mediated transcytosis in MDCK cells (Figure II-2A,B). This indicates the
selectivity of the effect of AG10.
The enhancement of Tf transcytosis was not due to the presence of a leaky tight
junction. AG10 had no significant effect on the integrity of the tight junction.
This was determined by measuring TEER in MDCK and Caco-2 cells, indicating
a low potential toxicity associated with AG10 treatment in these cell lines (Figure
II-2C,D).
The pulse-chase assay confirmed the results obtained from transport studies in
MDCK cells. In the cells pulsed with 125I-Tf at either the apical or basolateral
membrane for 2 hours at 37°C, no change in the extent of the transcytosis of Tf
was observed in the presence of AG10 (Figure II-3A,C). AG10 also did not show
any effect on Tf recycling in MDCK cells (Figure II-3B,D). On the other hand,
BFA increased the TfR-mediated transcytosis in MDCK cells (Figure II-3A,C).
52
Both apical and basolateral recycling of Tf decreased in BFA-treated MDCK cells
(Figure II-3B,D). These results suggest that AG10 and BFA may affect different
intracellular trafficking pathways for Tf. In MDCK cells, these pathways are not
affected by the treatment with AG10.
The mechanism involved in the enhancement of TfR-mediated transcytosis in the
presence of AG10 in Caco-2 cells is largely unknown. The morphological
changes caused by BFA have been studied extensively. In Chapter I, I have
reviewed the effects of BFA on the cellular organelles. Golgi apparatus, TGN,
and endosomes take a tubulated appearance in the presence of BFA. Golgi
redistributes to ER and the membrane traffic out of ER is blocked, while an
enhancement of membrane traffic from the Golgi apparatus to ER is observed. It
is believed that these morphological changes are the results of the effect of BFA
on the membrane proteins. BFA prevents the binding of β-COP and mono-ADP-ribosylation
factor (ARF) to the Golgi membrane. In spite of these observations,
the Golgi of some cell lines including MDCK cells seem to be resistant to the
effects of BFA (72). BFA releases γ-adaptin and clathrin coats from TGN and
endosomes respectively. In MDCK cells, after 2 minutes incubation with BFA,
the γ-adaptin was dispersed on both TGN and endosomes (42). Longer incubation
with BFA caused an extensive tubulation on the endosome compartments and
53
some tubulation of TGN, but the Golgi stack remained largely intact (42). The
fact that BFA causes significant enhancement of TfR-mediated transcytosis in
MDCK cells would indicate that this enhancement is dependent upon the sorting
events at the TGN/endosome level. From CE the basolaterally endocytosed Tf-
TfR complex recycles to the basolateral membrane. The 60-nm endosomal
tubules have coated buds containing clathrin and γ-adaptin (42). It has been
reported that BFA removes the coats from these buds resulting in the disruption of
polarized sorting of the TfR (42). This may indicate the involvement of γ-adaptin
containing buds in the concentrating and packaging of the TfR into the basolateral
vesicles (42). The disruption of the TfR polarized sorting decreases the
basolateral recycling of Tf (Figure II-3D, Table II-1). The enhancement of the Tf
transcytosis may be at the expense of the Tf recycling which decreases in the
presence of BFA.
Sheff et al. described two Tf recycling pathways with different kinetics in MDCK
cells (143). In their analysis they used a model which was defined by a series of
first order rate constants for the transfer of the ligands, such as Tf between
different compartments. In both polarized and non-polarized cells the half life of
the slow phase of Tf recycling was reported to be 15-30 minutes (53; 62; 63; 110;
127; 167). To study the effect of AG10 on the early time points of Tf recycling in
MDCK and Caco-2 cells, the cells were pulsed with 125I-Tf at the apical
54
membrane for 30 minutes at 37°C and chased for 30 minutes. Consistent with the
previous results, Tf recycling was not affected in MDCK cells (Figure II-4). On
the other hand, AG10 increased the recycling of Tf in Caco-2 cells (Figure II-5).
To explain the different outcomes of AG10 treatment in Caco-2 and MDCK cells,
the differences in the morphology and trafficking pathways present in these two
cell lines were studied. In many cell types, including Caco-2 cells, the existence
of an internal TfR pool has been suggested (86; 109). Therefore, a potential effect
of AG10 on the TfR internal pool can be considered. This may explain the effect
of AG10 on Tf intracellular trafficking in Caco-2 cells. In the presence of AG10,
an increase in the Tf transport from the TfR internal pool to the late endosomal
compartments involved in the transcytosis and recycling of Tf may result in the
enhancement of both recycling and transcytosis of Tf in Caco-2 cells. AG10 did
not show any effect on the Tf intracellular trafficking in MDCK cells, which may
be due to the absence of the TfR internal pool in this cell line.
In polarized epithelial cell lines, such as MDCK and Caco-2 cells, the apical and
basolateral membrane proteins are sorted at TGN. It has also been postulated that
in MDCK cells the polarized sorting of the membrane proteins which consists of
the signal dependent apical and basolateral pathways occurs from an endosomal
compartment in the cells (69; 147). Residues 19-41 in the human TfR
cytoplasmic tail are required for the basolateral transport of the newly synthesized
55
proteins, as well as for the basolateral recycling of the internalized TfR in MDCK
cells. In MDCK cells, the apical and basolateral membrane proteins, after the
sorting, are transported directly to the appropriate plasma membrane domain. In
Caco-2 cells, in addition to the direct pathway, the membrane proteins destined
for the apical surface can be targeted to the basolateral plasma membrane. They
may be endocytosed to the BEE and later segregated away from the basolateral
proteins destined for recycling or degradation in the late endosomes/lysosomes.
They are then packed into transcytotic carrier vesicles which deliver their contents
to the apical plasma membrane through the indirect pathway (93; 107). It is less
likely, but the effect of AG10 on Tf trafficking in Caco-2 cells can also be due to
the alteration of the membrane protein sorting, which may occur at the endosomal
compartments of the cells. In the presence of AG10, a missorting of the apical
TfR to the basolateral membrane may increase the transport of Tf to the
basolateral cell membrane from where they are redirected back to the apical
membrane via the indirect pathway present in Caco-2 cells. This may result in an
increase in the basolateral-to-apical transcytosis of Tf in AG10-treated Caco-2
cells. The enhancement of the basolateral-to-apical transcytosis of the TfR could
be a transient state, and the TfR might be internalized and redirected back again to
the basolateral membrane. This results in an increase in the apical-to-basolateral
transcytosis of Tf.
56
Summary: AG10 increases the TfR-mediated transcytosis only in Caco-2 cells,
indicating that AG10 may be an appropriate enhancer to increase the absorption
of orally administered Tf-protein conjugates across the GI epithelia. AG10 did
not cause any disruption in the integrity of the tight junction of the cells. It was
shown to have low potential side effects in vitro. The different outcomes related
to the effects of AG10 in Caco-2 and MDCK cells may be related to the presence
of an internal TfR pool in Caco-2 cells. The existence of two sorting pathways,
direct and indirect, in Caco-2 cells may also explain the differences in the
obtained data in these two cell lines.
57
Chapter III. Involvement of Rab11 in the Trafficking of Apically
Endocytosed Transferrin: AG10 stimulates a novel trafficking pathway
of the apically endocytosed transferrin through the Rab11-enriched
endosomal compartment in Caco-2 cells
1. Abstract
The GTPase inhibitor AG10 has been shown previously to increase the apical-to-basolateral
transcytosis of transferrin in Caco-2 cells. However, the mechanism
of the increased transcytosis has not been established. In this chapter, the effect
of AG10 on the trafficking of endocytosed transferrin among different endosomal
compartments as well as the involvement of Rab11 in the recycling of transferrin
was investigated. Confocal microscopy studies showed a high level of
colocalization of FITC-transferrin with Rab5 and Rab11 in Caco-2 cells pulsed at
16°C and 37°C. This indicates the presence of apically endocytosed FITC-transferrin
in early endosomes and apical recycling endosomes at 16°C and 37°C,
respectively. The effect of AG10 and BFA on the accumulation of transferrin
within different endosomal compartments was studied. An increase in the
58
transcytosis and recycling of internalized 125I-transferrin, as well as a decrease in
cell-associated 125I-transferrin, was observed in AG10-treated Caco-2 cells pulsed
at 37°C and chased for 30 minutes. Recycling of endocytosed transferrin involves
its passage through different endosomal compartments. In addition, confocal
microscopy showed that FITC-transferrin exhibits a higher level of colocalization
with Rab11 in the presence of AG10, suggesting an effect of AG10 on the later
steps of transferrin receptor trafficking which are involved in subsequent
recycling, and possibly transcytosis, of apically endocytosed transferrin in Caco-2
cells. The apical recycling of transferrin was not affected by the treatment with
BFA. BFA also did not show any effect on the extent of colocalization between
FITC-Tf and Rab11. These data indicate the involvement of Rab11 in the apical
recycling of transferrin in Caco-2 cells.
59
2. Introduction
The intracellular trafficking of Tf has been studied extensively in polarized
epithelial cells such as MDCK cells. Published studies have focused on
trafficking of Tf internalized at the basolateral membrane, and there is little
information available regarding the intracellular trafficking pathways of the
apically endocytosed Tf. Due to the importance of enhancement of the apical-to-basolateral
TfR-mediated transcytosis of Tf conjugates, the intracellular
trafficking of the apically endocytosed Tf is the focus of this study.
In this chapter, the effect of temperature and AG10 on Tf intracellular trafficking
was used to investigate the intracellular processing of the apically endocytosed
fluorescein isothiocyanate (FITC)-Tf conjugate in order to identify endosomal
compartments involved in Tf trafficking in Caco-2 cells. I also study the effect of
AG10 and BFA on the colocalization of Rab11 with endocytosed Tf during its
intracellular trafficking in Caco-2 cells. The involvement of Rab5 and Rab11 in
Tf intracellular trafficking is summarized in Table III-1.
61
3. Experimental
a. Cell Culture: C2BBe1, a subclone of Caco-2 cells was obtained from the
ATCC. Rabbit polyclonal antibodies against Rab5 and Rab11 were purchased
from Santa Cruz Biotechnology. The rabbit polyclonal antibody against ZO-1
and Alexa-Fluor (568) labeled secondary antibody were purchased from
Molecular Probes. The mouse monoclonal antibody against ZO-1 and Alexa-
Fluor (633) labeled secondary antibody were purchased from Invitrogen. More
information provided in Chapter II.
b. Preparation of diferric, radiolabeled, and fluorescein isothiocyanate labeled
Tf: Preparation of diferric and radiolabeled Tf has been described in Chapter II.
FITC (Sigma) was conjugated to Tf with a molar ratio of 50 to 1 in 1 M sodium
bicarbonate pH 9.0 under constant stirring for 1 hour at room temperature (Figure
III-1). The untreated FITC was removed by dialysis against 2 L PBS at 4°C
overnight.
62
Protein concentration and the degree of labeling were calculated using the
formula provided by Molecular Probes.
Formula III-1:
protein concentration (M)=[A280-(A494×0.30)]×dilution factor/89,400,
dye per protein molecule=A494×dilution factor/68,000×protein concentration
where 0.30 is a correction factor, 89,400 is the molar extinction coefficient of Tf
(cm-1M-1) at 280 nm, and 68,000 is the molar extinction coefficient of the dye
(cm-1M-1) at pH 8.0 at 494 nm.
64
c. Pulse-chase Experiment using 125I-Tf: The pulse-chase experiment has been
described extensively in Chapter II (Figure II-1). 125I-Tf (2-4 μg/ml, 2×106
cpm/ml) was bound to the apical surface of the cells in the presence and absence
of AG10 or BFA on ice for 15 minutes, which was followed by incubation at
either 16°C or 37°C for 30 minutes. After removing the unbound and cell
surface-bound 125I-Tf with ice-cold PBS (containing 0.1 mM Ca2+ and 0.05 mM
Mg2+) and acid (0.15 M NaCl, 0.5 M HAC, pH; 2.2) washes on ice, the cells were
chased with or without AG10 or BFA and in the presence of 100-fold excess of
unlabeled Tf for up to 30 minutes at 37°C. Student’s t-test was used to analyze
the significance, and p corresponds to 0.05 or 0.01.
65
d. 125I-Tf Uptake Assay: Caco-2 cells cultured for 2 weeks grown on 24 mm
plates, were depleted of endogenous Tf and preincubated in serum free DMEM
containing 0.1% BSA and 20 mM HEPES in the presence or absence of AG10
(500 μM) for 2 hours at 37°C. The medium was replaced with a fresh DMEM
(0.1% BSA, 20 mM HEPES) containing 125I-Tf (with a final concentration of 2-4
μg/ml, 2×106 cpm/ml) in the continued presence or absence of AG10 for 30
minutes at either 16°C or 37°C. The unbound and surface-bound 125I-Tf was
removed as indicated above. The cells were chased in the presence of 100-fold
excess of unlabeled Tf and in the continued presence or absence of AG10 for 10
or 30 minutes at 37°C. At the end of the chase period, the surface-bound 125I-Tf
was removed as indicated above and the PBS and acid washes were collected.
The cells were lysed in 1 M NaOH, and the radioactivity level in the cells, the
collected PBS, and acid fractions was counted by the gamma counter. Protein
assay was performed by using BCA (Pierce). The data were corrected for non-specific
uptake of 125I-Tf determined in parallel experiments in which an excess of
unlabeled Tf was present during the pulse period. The internalized Tf is
expressed as ng per mg total protein.
66
e. Confocal Fluorescence Microscopy: Caco-2 cells grown on 12 mm
polystyrene filters (Costar, Cambridge, MA), were subjected to pulse-chase assay
as described above except that FITC-Tf was used to replace 125I-Tf. To obtain a
better visualization, the cells were used at about >90% confluency. However, the
fields with 100% confluency were selected for confocal analysis. After binding
of FITC-Tf (45 ng/ml) to the apical surface of the cells for 45 minutes on ice, the
cells were incubated at either 16°C or 37°C for 30 minutes. The surface-bound
FITC-Tf was removed by sequential PBS (containing 0.1 mM Ca2+ and 0.05 mM
Mg2+) and acid (0.15 M NaCl, 0.5 M HAC, pH; 2.2) washes and then the cells
were chased for 0 or 30 minutes in the presence of 10-fold Tf at 37°C. At the end
of the chase period the unbound FITC-Tf was removed by multiple PBS washes,
and the cells were fixed using 3.7% para-formaldehyde (Sigma) in PBS, which
was followed by quenching and permeabilization with 0.2% saponin for 5
minutes. The cells were then blocked with 10% FBS in PBS and incubated with
the appropriate primary antibodies diluted in 1.5% FBS in PBS. Alexa-Fluor 568
goat anti-rabbit antibody was used for the secondary detection of the proteins.
Unless ZO-1 was labeled with the mouse anti-ZO-1 primary antibody, the Alexa-
Fluor 633 goat anti-mouse antibody was used for the secondary detection of the
ZO-1. The nucleus was labeled by incubation with 4′,6-diamidino-2-
phenylindole, dilactate (DAPI, dilactate, Invitrogen) dye for 15 minutes on a
platform shaker.
67
The efficiency of "acid washes" in removing the surface-bound FITC-Tf was
demonstrated in control experiments, in which the cells were incubated with
FITC-Tf for 30 minutes at 37°C and were not subjected to the acid washes at the
end of the pulse period.
f. Image Processing: The images were collected from different fields of the
prepared slides using a confocal laser scanning microscope (LSM 510 Meta NLO
imaging system, Carl Zeiss) equipped with Argon, red HeNe and green HeNe
lasers for standard confocal fluorescence microscopy and a Ti-Sapphire tunable
Coherent Chameleon laser (720-950 nm) for multiphoton excitation of DAPI.
The ability of this system to acquire fluorescence emission signals resolved within
narrow ranges in multitrack mode, and the use of singly-labeled control samples
ensured the validity of colocalization studies. The acquisition parameters for the
collection of images were kept constant between samples to allow comparison of
the images taken from different fields and slides. Caco-2 cell monolayers were
optically sectioned in the z-axis from the apical to the basolateral side of the cell
(Figure III-2).
69
The first z-section from the apical surface of the cells was captured when FITC-Tf
(green) was observed. The step size in the z-axis was 0.4 μm. Analysis of the
extent of colocalization between markers of interest in confocal microscopy
images was done using the Enhanced Colocalization tool available with the Zeiss
LSM510 software. To quantify the colocalization between FITC-Tf and Rab
proteins, the image fields were thresholded. The whole image field was defined
as a region of interest (ROI) and the system calculated mean intensities +2x
standard deviation to threshold the images. Colocalization coefficients
representing the relative number of colocalizing pixels in either channel as
compared to the total number of pixels above the threshold irrespective of their
intensity were measured through the image sequence in the z-series. To analyze
the extent of the colocalization between FITC-Tf and the Rab proteins, we used
the colocalization coefficient of FITC determined in each z-section of a z-series
which shows the relative number of colocalizing FITC pixels compared to the
total number of FITC pixels above the threshold in the z-sections. The average
colocalization coefficient of FITC-Tf at any specific height in the cells was
obtained using the FITC-Tf colocalization coefficients of the z-sections which
were collected from different fields of the prepared slide and localized at the same
height above either the nucleus or the tight junction (Figure III-3).
71
For statistical analysis, the standard error of the mean (s.e.m.) was calculated. A
student’s t-test was used to analyze the significance (p≤0.05). In the images
labeled with the same fluorophore to detect ZO-1 and Rab protein staining, the
single cells were thresholded twice, with and without ZO-1 staining. To obtain
the colocalization coefficient, ZO-1 staining was omitted and the inner part of the
cells were defined as the ROI. To look at the distribution of the tight junction, the
mean intensity for Alexa-Fluor 568 pixels or Alexa-Fluor 633 pixels in the cells
thresholded with ZO-1 was considered. Three-dimensional views of the z-series
were obtained by using the Projection function, which calculated a series of
projections after the rotation of the data package around the Y-axis. The separate
profile for each channel along a three-dimensional view was obtained by using the
Profile function. Two-dimensional deconvolution views of the yz and xz-plane
were generated by using the Ortho function. All images were compiled using
Adobe Photoshop 7.0 (Adobe Systems Inc., Mountain View CA).
72
4. Results
a. FITC-Tf was colocalized with Rab5 and Rab11 in Caco-2 cells pulsed at 16°C
and 37°C, respectively. The transport of endocytosed synaptic vesicle proteins to
the perinuclear region has been shown to be inhibited at temperatures lower than
20°C (38). To study the effect of the temperature on the intracellular location of
the internalized Tf in Caco-2 cells, the cells were pulsed at the apical plasma
membrane with FITC-Tf for 30 minutes at two different temperatures; at 16°C
and 37°C. The surface-bound FITC-Tf was removed by sequential neutral and
acid washes, and then the cells were labeled with antibodies against Rab5 or
Rab11. The acid washes contributed to the removal of FITC-Tf, which may bind
to microvilli, from the apical cell surface, which have a length of 1.32 μm (data
not shown). The average short and long microvilli in C2BBe1 subclone cells was
reported to be 0.6 μm and 1.6 μm, respectively (126).
Consistent with the previous observations showing the effect of the temperature
on the accumulation of Tf in different endosomal compartments, FITC-Tf showed
a high level of colocalization with Rab5 in the cells pulsed at 16°C.
73
Galleries of the optical sections from the apical to the basolateral plasma
membrane of the cells are presented in Figures III-4A,B and III-5A,B. The three-dimensional
views of these galleries were obtained by using the Projection
function (Figures. III-4C,D, III-5C,D). Comparing the galleries of the optical
sections of the cells pulsed at 16°C and 37°C, I found FITC-Tf colocalized with
Rab5 to a high degree in the first three z-sections of the gallery of the cells pulsed
at 16°C (Figure III-4A,B), which is more noticeable in one of the cells presented
in this gallery. These z-sections are localized about 4.6 μm to 6.7 μm above the
intracellular level at which the nucleus begins to appear. An apically peripheral
distribution of FITC-Tf, which was absent in the cells pulsed at 37°C, is shown in
the three-dimensional view of the gallery of the Caco-2 cells pulsed at 16°C and
immunolabeled with the antibody against Rab5 (Figure III-4C, arrow). On the
other hand, a high degree of colocalization between FITC-Tf and Rab11 was
observed in the cells pulsed with FITC-Tf at 37°C. As shown in Figure III-5B,
FITC-Tf was colocalized with Rab11 in the z-sections localized only 2 μm to 3
μm above the nucleus of the cells pulsed at 37°C. The three-dimensional views of
the galleries confirm also the FITC-Tf presence above the nucleus in the cells
pulsed at 37°C (Figures III-4D, III-5D). The colocalization between FITC-Tf and
Rab11 in Caco-2 cells pulsed at 37°C is shown in Figure III-5D (arrows).
76
The average colocalization coefficient of the FITC-Tf colocalized with the Rab
proteins, which was obtained analyzing the images collected from different fields
of the prepared slides, was plotted versus the z-sections arranged based on their
distance from the nucleus towards the apical surface of the cells (Figures III-6,
III-7). In the collected z-series, the first z-section from the apical surface in which
the nucleus was first observed was set to zero, negative numbers move towards
the basolateral membrane (BLM) while positive move towards the apical
membrane (APM).
Consistent with the results obtained from the galleries of the optical sections of
the cells, the highest degree of the colocalization between FITC-Tf and Rab5 was
found close to the apical membrane of the cells pulsed at 16°C. This was
determined in about 70% of the fields analyzed in the z-sections localized from
6.4 μm to 7.2 μm above the z-section containing the first sign of nucleus labeling
in the cells pulsed at 16°C (Fig. III-6). In these z-sections about 55% of FITC-Tf
was colocalized with Rab5 at 16°C. The highest level of the colocalization
between FITC-Tf and Rab11 was found in the z-sections localized from 2.0 μm to
3.2 μm in the cells pulsed at 37°C (Figure III-7). In the z-sections localized about
2.4 μm and 2.8 μm above the nucleus about 35-40% of total FITC-Tf was
colocalized with Rab11 at 37°C.
79
b. FITC-Tf was colocalized with Rab5 and Rab11 in two different intracellular
locations above the tight junction. In order to determine the intracellular location
of the z-sections containing the highest level of the colocalization between FITC-Tf
and the Rab proteins, the tight junction was labeled by using antibody against
ZO-1. Staining of ZO-1 allowed us to follow the FITC-Tf distribution in single
cells and also eliminated the effect of the uneven filter membrane and different
cell height.
Galleries of the optical sections from the apical to the basolateral plasma
membrane of the cells pulsed at 16°C and 37°C are shown in Figures III-8A and
III-9A. Consistent with the results obtained from the cells without any tight
junctional labeling, the intracellular location of the FITC-Tf colocalized with
Rab5 and Rab11 is different in the cells pulsed at 16°C and 37°C (Figures III-8A,
III-9A). In the cells pulsed at 16°C, a high level of colocalization between FITC-Tf
and Rab5 was observed in the z-sections localized apically and distal from the
tight junction. On the other hand, FITC-Tf was colocalized with Rab11 in the z-sections
localized above the tight junction in the cells pulsed at 37°C.
80
The intracellular location of the FITC-Tf colocalized with the Rab proteins
relative to the tight junction is also shown in the two-dimensional deconvolution
views of the yz and xz-plane (Figures III-8B, III-9B). These results were also
confirmed by analyzing the distribution of FITC-Tf, the Rab proteins and the ZO-
1 along the three-dimensional views of the galleries (Figures III-8C, III-9C). The
distribution profile for each channel was obtained using the Profile function. At
16°C, the green and red peaks that represent FITC-Tf and Rab5 overlapped far
from the peaks representing the ZO-1 and the nucleus (Figure III-8C). On the
other hand, an overlap between the peaks representing FITC-Tf and Rab11 was
observed adjacent to the ZO-1 in the cell pulsed at 37°C (Figure III-9C). The
animated three-dimensional projections are also available (supplementary
material Movies III-1, III-2).
83
The z-sections arranged based on their distance from the tight junction were
plotted versus both the average mean intensity of the red pixels, and the average
colocalization coefficient of the FITC-Tf, which were obtained analyzing a
number of single cells. After adjustment of the level of the collected z-sections
relative to the tight junction, the first z-section containing the highest mean
intensity of the red pixels was set to zero. The distribution of the red pixels
through the cells is shown (Figures. III-10, III-11), which represent both the Rab
labeling and the ZO-1 labeling, and the peak of the graph correlates with the
presence of the tight junction in the particular z-sections in the cells as detected in
the images collected by confocal microscope. The average thickness of the z-sections
containing ZO-1 staining shown in Figures III-10 and III-11 was 1.22
μm. Staining of ZO-1 did not affect the observed colocalization between FITC-Tf
and the Rab proteins (Figures III-6, III-7, III-10, III-11). The highest level of
colocalization between FITC-Tf and Rab5 at 16°C was observed in the z-sections
with an average distance of 6 μm above the tight junction (Figure III-8A, III-10).
On the other hand, the z-sections localized about 2.0 μm to 2.8 μm above the tight
junction contained the highest degree of the colocalization between FITC-Tf and
Rab11 at 37°C (Figure III-9A, III-11). The distribution of Rab5 and Rab11 was
not affected by the temperature (data not shown ).
86
c. The effect of AG10 on transcytosis, recycling and cell-association of Tf was
only observed in the cells pulsed at 37°C. The recycling, transcytosis, and cell-association
of Tf was measured in the presence and absence of AG10. As shown
in Caco-2 cells, which were pulsed with 125I-Tf at 16°C, the recycling and
transcytosis of Tf were not affected by the treatment with AG10 (Figure III-
12A,B), but the cells pulsed at 37°C showed a statistically significant increase in
the recycling and transcytosis of Tf in the presence of AG10 (Figures III-12A,B,
III-13A,B). The cell association of Tf did not change in the presence of AG10 in
the cells pulsed at 16°C for 30 minutes (Figure III-12C). On the other hand, a
statistically significant decrease in the amount of cell-associated Tf was observed
in the cells which were treated with AG10 and pulsed at 37°C for 30 minutes
(Figures III-12C, III-13C). In control cells, the transcytosis, recycling and cell-association
of Tf reached the plateau after a 10 minute chase, while AG10 showed
a continuous effect through the chase period (Figure III-12A-C). Consistent with
previously presented results (Figure II-2), the integrity of the tight junction which
was tested by measuring TEER was not affected by AG10 treatment.
91
Consistent with the results obtained from Pulse-Chase assays, the uptake of Tf
was decreased in the presence of AG10 in the cells pulsed with 125I-Tf at 37°C but
no change in Tf uptake was observed in the cells pulsed at 16°C (Figure III-14A).
There was no significant difference in the amount of CPM counted in the acid-wash
fractions which were recovered in the end of the chase periods from the
apical cell membranes (Figure III-14B).
93
d. AG10 did not show any effect on the integrity of tight junction. To ensure
that AG10 did not affect the tight junction, the integrity of the tight junction was
evaluated in the presence of AG10. The tight junction was labeled with the
antibody against ZO-1 in Caco-2 cells that were incubated in serum free DMEM
with or without AG10 (500 μM) for 2 hours at 37°C. As shown in Figure III-
15B, the labeling of ZO-1 was intact in the cells treated with AG10 for 2 hours.
An intact tight junction was also observed in the three-dimensional views
obtained from the galleries of optical sections of the control and AG10-treated
cells (Figure III-15C,D).
95
e. AG10 increased colocalization between FITC-Tf and Rab11. To define the
involvement of Rab11 in the effect of AG10 on Tf distribution in Caco-2 cells,
the cells were pulsed with FITC-Tf in the presence or absence of AG10 (500 μM)
for 30 minutes at 37°C followed by a chase in the continued presence or absence
of AG10 for 30 minutes at 37°C. The galleries of the optical sections from the
apical towards the basolateral cell membrane of the non-treated and AG10-treated
Caco-2 cells are shown in Figure III-16A,B. In the z-sections located above the
nucleus, a higher degree of colocalization between FITC-Tf and Rab11 was
observed in the cells treated with AG10 compared with that in non-treated cells
(Figure III-16A,B). Both colocalization coefficient and the mean intensity of
FITC-Tf obtained from the average of three different experiments were plotted
versus the z-sections arranged based on their distance from the nucleus (Figure
III-17A,B). The intracellular level at which the nucleus begins to appear was
determined using the data obtained from the experiments with the nucleus
labeling. As shown in Figure III-17A, the average colocalization coefficient of
the FITC-Tf increased significantly in the z-sections localized about 1.2 μm to 2.4
μm above the nucleus in the cells treated with AG10. An increase in the amount
of FITC-Tf in the more basolateral regions of the Caco-2 cells was also observed
in AG10-treated cells, suggesting an effect of AG10 on the transcytosis of Tf
(Figure III-17B).
98
f. BFA did not show any effect on the apical recycling of 125I-Tf. To study the
effect of BFA on intracellular trafficking of the apically endocytosed Tf in Caco-2
cells, the cells were pulsed at the apical membrane with 125I-Tf in the absence or
presence of AG10 (500μM) or BFA (3.2 μg/ml) for 30 minutes at 37°C. Then the
surface-bound Tf was removed by sequential neutral and acid washes and the
cells were chased in the continued absence or presence of AG10 or BFA for 30
minutes at 37°C. The recycling of 125I-Tf was measured in the absence or
presence of either AG10 or BFA. Consistent with the results shown previously
(Figures II-5, III-12A, and III-13A), the recycling of the apically endocytosed Tf
increased significantly in the presence of AG10 at 37°C (Figure III-18). There
was no significant change in the Tf recycling in BFA-treated Caco-2 cells (Figure
III-18).
100
g. BFA did not show any effect on the colocalization between FITC-Tf and Rab11
in Caco-2 cells. To further define the involvement of Rab11 in the apical
recycling of Tf, Caco-2 cells were pulsed with FITC-Tf at the apical membrane in
the absence or presence of either AG10 (500 μM) or BFA (3.2 μg/ml) for 30
minutes at 37°C, followed by the removal of the unbound and surface-bound
FITC-Tf by multiple PBS and acid washes. The cells were immunolabeled with
primary and fluorescent secondary antibodies against Rab11 (red) and ZO-1
(pink). Galleries of the optical sections from the apical to the basolateral plasma
membrane of the cells are presented in Figure III-19. As expected, the apically
endocytosed FITC-Tf was found in the z-sections localized above the tight
junction at 37°C. It had access to the z-sections which were located about 2.0 μm
to 2.4 μm above the tight junction in non-treated Caco-2 cells (Figure III-19A).
In AG10-treated cells, I found FITC-Tf in the z-sections localized more apically,
i. e. 2.4 μm to 2.8 μm above the tight junction (Figure III-19B). FITC-Tf was
also colocalized with Rab11 to a higher extent in the presence of AG10. I did not
see any difference in the intracellular distribution of FITC-Tf in BFA-treated
Caco-2 cells compared to the non-treated cells (Figure III-19A,C). There was
also no difference in the extent of the colocalization between FITC-Tf and Rab11
in the non-treated and BFA-treated Caco-2 cells (Figure III-19A,C). The
animated three-dimensional projections are also available (supplementary
material Movies III-3, III-4, and III-5).
104
The images, which were collected from different fields of the prepared slides,
were analyzed as previously described. The average mean intensity of ZO-1 and
the average colocalization coefficient of the FITC-Tf colocalized with Rab11
were plotted versus the z-sections which were arranged based on their distance
from the tight junction towards the apical surface of the cells (Figure III-20).
Consistent with the results obtained from the galleries of the optical sections of
the cells, FITC-Tf was colocalized with Rab11 in a higher extent in the z-sections
localized 1.2 μm to 2.8 μm above the tight junction of the cells treated with AG10
(Figure III-20B). In the z-sections localized about 2.4 μm to 2.8 μm above the
tight junction, FITC-Tf was colocalized with Rab11 only in AG10-treated Caco-2
cells. The extent of the colocalization between FITC-Tf and Rab11, compared to
the control, was not affected by the presence of BFA (Figure III-20B).
107
5. Discussion
The effect of the temperature on the intracellular trafficking of Tf, as well as its
colocalization with the Rab proteins, was studied in cultured Caco-2 cells.
Consistent with the published reports regarding the effect of the low temperatures
below 20°C on intracellular trafficking of the endocytosed ligands (38), I
observed that approximately 50-60% of FITC-Tf labeling was colocalized with
Rab5 in the apically located z-sections of the cells pulsed at 16°C (Table III-1,
Figure III-6). The transfer of Tf to the late endosomal compartments which
contain Rab11 was blocked (Scheme III-1: pathway 3), as evidenced by the
failure to observe marked colocalization of these two markers under these
conditions. Consequently, the recycling and transcytosis of Tf occurred mostly
from endosomal compartments which contain Rab5 (Scheme III-1: pathway 2).
109
As expected, Rab5 showed a vesicular staining and the endosomal compartment
containing Rab5, which was accessible to FITC-Tf bound to the apical surface of
the cells, was localized close to the apical membrane (Figures III-8, III-10). This
observation indicates the presence of FITC-Tf in the AEEs. The peripheral
colocalization between FITC-Tf and Rab5, which was observed through the
cytoplasm, may depend on the role of Rab5 in the vesicular transport among
surface membranes and the endosomal compartments (Table III-1). A population
of Rab5 observed in the level of the tight junction was not accessible to the FITC-Tf
bound to the apical or the basolateral surface of the cells (data not shown). On
the other hand, FITC-Tf showed a higher level of colocalization with Rab11 in the
cells which were pulsed at 37°C (Figures III-5, III-7). This observation indicated
the transfer of FITC-Tf from Rab5-containing endosomal compartments to a
Rab11-containing endosomal compartment in the cells pulsed at 37°C (Scheme:
III-1: pathway 3).
The existence of a population of cup-shaped vesicles containing Rab11 and
Rab25, which constitute the ARE, has been reported in the apical region of
MDCK cells (23; 48). I found Rab11 in the vesicles in the apical region of Caco-
2 cells and above the z-sections containing ZO-1 staining, which could be
analogous to the ARE in MDCK cells (Figures III-9, III-11). The involvement of
110
Rab11 in the recycling of internalized Tf has been shown in non-polarized cells
(Table III-1). Ren et al. reported the inhibition of Tf recycling from perinuclear
RE in TRVb cells expressing the dominant positive (Q70L) or dominant negative
(A25N) mutant forms of Rab11 (134).
To investigate the effect of AG10 on intracellular processing of Tf among the
endosomal compartments, I made use of the effect of temperature on the
accumulation of endocytosed Tf in the endosomal compartments containing Rab5
and Rab11 at 16°C and 37°C, respectively (Table III-1, Figures III-4-11). The
cells were pulsed with 125I-Tf at two different temperatures and consistent with
the results obtained from confocal microscopy, the passage of the endocytosed Tf
from endosomal compartments containing Rab5 to those containing Rab11 was
blocked at 16°C and the majority of Tf (about 90%) was chased out of the cells
within the first 10 minutes (Figure III-12A, Scheme III-1: pathway 2). As shown
in Figure III-13A, in the cells pulsed with 125I-Tf at 37°C, the recycling of Tf
followed the criteria of the slow phase recycling, i.e. only an average of 30% of
Tf was recycled in the first 10 minutes (Scheme III-1: pathway 4). The effect of
AG10 was pronounced at 37°C and the results obtained from pulse-chase and
uptake assays suggested that AG10 affects the late process of Tf transport, after
the exit of Tf from endosomal compartments containing Rab5.
111
The significant increase in the apical-to-basolateral transcytosis of Tf (Figure III-
13B) in the presence of AG10 may explain the effect of AG10 on the transport of
insulin-Tf conjugate in Caco-2 cells reported by Xia and Shen (176). The
increase in the mean intensity of the apically endocytosed FITC-Tf, which was
observed in the more basolateral regions of the Caco-2 cells treated with AG10
(Figure III-17B), may also be explained by the effect of AG10 on TfR-mediated
transcytosis. The amount of cell-associated Tf decreased by an average of about
30% after 30 minutes chase in the presence of AG10 (Figure III-13C). As shown
in Figure III-14A, AG10 decreased the uptake of Tf by 15% after 10 minutes
chase which can be explained by the increase in the amount of recycled and
transcytosed Tf in Caco-2 cells. From the uptake assay in the presence and
absence of AG10, there was no significant difference in the amount of surface-bound
Tf, as shown in the radioactivity recovered in the acid-wash fractions
(Figure III-14B).
The data presented here support one of the hypotheses discussed in Chapter II.
Since a high level of the cellular uptake of Tf at 37°C has been demonstrated in
many cell types, and the existence of an internal TfR pool has been suggested (86;
109), a potential effect of AG10 on the TfR internal pool can be considered. In
the presence of AG10, the internalized Tf-TfR complexes may be depleted from
112
Caco-2 cells resulting in an increase in both recycling and transcytosis of Tf. The
effect of AG10 on the late processes of Tf trafficking was confirmed by the
demonstration of a higher degree of colocalization between FITC-Tf and Rab11
in the presence of AG10 (Figures III-16, III-17) which indicates that AG10
stimulates the trafficking pathway of Tf through Rab11-containing endosomal
compartment.
In polarized epithelial cells, BFA treatment significantly up-regulates the
basolateral-to-apical transcytosis of the internalized TfR and bulk plasma
membrane (129). It has also been shown to increase the TfR-mediated
transcytosis from the apical to the basolateral side in Caco-2 cells. In Chapter II, I
showed that BFA does not have any significant effect on the apical recycling of
Tf in Caco-2. To study the involvement of Rab11 in the apical recycling of Tf, I
made use of the effect of AG10 and BFA on the recycling of the apically
endocytosed Tf in Caco-2 cells. Consistent with the previous observations
regarding the effect of AG10 and BFA on the apical recycling of Tf, AG10
increased the apical recycling of Tf in Caco-2 cells (Figure III-18).
BFA did not show any significant effect on the apical recycling of Tf in Caco-2
cells (Figure III-18). On the other hand, both AG10 (Figures III-12B, III-13B)
and BFA increased the TfR-mediated transcytosis in Caco-2 cells, suggesting the
113
involvement of different intracellular pathways in the recycling and transcytosis
of Tf in this cell line. The possible mechanism for BFA has already been
discussed in Chapters I and II.
I further studied the effect of BFA on the colocalization between FITC-Tf and
Rab11. Caco-2 cells were pulsed with FITC-Tf for 30 minutes at 37°C in the
absence or presence of either AG10 or BFA, followed by the removal of unbound
and surface-bound FITC-Tf with multiple neutral and acid washes. As expected,
FITC-Tf was colocalized with Rab11 to a higher extent in AG10-treated Caco-2
cells (Figures III-19A,B, III-20B). In the z-sections localized about 2.4 μm to 2.8
μm above the tight junction, FITC-Tf was colocalized with Rab11 only in AG10-
treated Caco-2 cells, indicating the presence of FITC-Tf in a higher degree in the
z-sections localized more apically relative to the tight junction. BFA did not
show any effect on the apical recycling of Tf in Caco-2 cells (Figure III-18). As a
result, no change in the extent of the colocalization between FITC-Tf and Rab11
was observed in BFA-treated cells, compared to non-treated cells (Figure III-
19A,C, III-20B).
These data confirm that the apically endocytosed Tf, after leaving the CE, crosses
the Rab11-containing endosomal compartment on its way towards the apical
membrane in Caco-2 cells (Scheme III-1: pathway 4). In Caco-2 cells, this
114
pathway is largely affected by AG10, while BFA does not show any significant
effect on the Tf transport through this pathway.
Summary: The effect of temperature on the Tf accumulation in different
endosomal compartments was studied. I found the apically endocytosed Tf in the
Rab5-containing endosomal compartment at 16°C. At 37°C, the majority of Tf
was transfered to the Rab11-containing endosomal compartment. Extensive
confocal microscopy studies revealed the location of these endosomal
compartments relative to the nucleus and tight junction in Caco-2 cells. The
Rab11-containing endosomal compartment was localized above the tight junction,
while the Rab5-containing endosomal compartment was localized apically and
distal from the tight junction.
I took advantage of the effect of the temperature on colocalization between the
apically endocytosed Tf and the Rab proteins, Rab5 and Rab11, to study the effect
of AG10 on the intracellular trafficking of Tf. The effect of AG10 on Tf
intracellular trafficking was observed only at 37°C, but not at 16°C. This
indicates that AG10 affects the trafficking of Tf after its exit from AEE. AG10
increased the transcytosis and recycling of the apically endocytosed Tf. It also
stimulated the transport of the apically endocytosed Tf to the Rab11-containing
endosomal compartment in Caco-2. I concluded that AG10 may increase the
115
transport of Tf-TfR complexes from intracellular compartments containing the
internalized Tf at 37°C, which may be localized prior to the Rab11-containing
endosomal compartment, to the endosomal compartments from which transcytosis
and recycling of Tf occur.
Furthermore, the presence of different intracellular pathways through which the
recycling and transcytosis of the apically internalized Tf occur was demonstrated.
After leaving the CE, where the Tf sorting may take place, the apically
endocytosed Tf crosses the Rab11-containing endosomal compartment on its way
towards the apical membrane (Scheme III-1: pathway 4). Tf, on its way from the
apical to the basolateral cell membrane, does not have access to the Rab11-
containing endosomal compartment. It may cross the CE on its way towards the
basolateral membrane. BFA did not show any effect on the transport of the
apically endocytosed Tf to the Rab11-containing endosomal compartment which
was largely affected in AG10-treated Caco-2 cells. This confirms the
involvement of Rab11 in the apical recycling of Tf in Caco-2 cells.
116
Chapter IV. Involvement of Rab11 in the Basolateral Trafficking
of Transferrin: AG10 increases the transport of the basolaterally
endocytosed transferrin to the Rab11-containing endosomal compartment in
Caco-2 cells
1. Abstract
The involvement of Rab11 in the apical recycling of transferrin has previously
been discussed. The GTPase inhibitor AG10 has been shown to stimulate the
trafficking pathway of the apically endocytosed transferrin through the Rab11-
containing endosomal compartment in Caco-2 cells. However, the involvement
of Rab11 in the intracellular trafficking of the basolaterally endocytosed
transferrin is largely unknown. In this chapter, the effect of AG10 on the
trafficking of transferrin endocytosed from the basolateral cell membrane was
investigated. AG10 increased the recycling and transcytosis of transferrin and
decreased the cell-associated transferrin in Caco-2 cells which were pulsed with
125I-transferrin at the basolateral membrane for 30 minutes at 37°C and chased for
45 minutes. Confocal microscopy studies showed that the basolaterally
endocytosed FITC-transferrin had no access to the Rab11-containing endosomal
117
compartment. Interestingly, the basolaterally endocytosed FITC-transferrin was
colocalized with Rab11 to some degree in AG10-treated Caco-2 cells. This
indicates that the basolaterally endocytosed transferrin, after the exit from the CE,
crosses the Rab11-containing endosomal compartment on its way towards the
apical cell membrane in Caco-2 cells.
118
2. Introduction
The involvement of Rab11 in the Tf recycling in non-polarized cells has been
reported (159). In Chapter III, I showed the involvement of Rab11 in the apical
recycling of the apically endocytosed Tf in Caco-2 cells. Some studies have
discussed the absence of the basolaterally endocytosed Tf in the Rab11-containing
endosomal compartment in MDCK cells (20; 167). In contrast to these studies, I
observed the apically endocytosed FITC-Tf to be colocalized with Rab11 at 37°C
in Caco-2 cells. In this chapter, I discuss the trafficking of Tf endocytosed from
the basolateral cell membrane through the Rab11-containing endosomal
compartment in Caco-2 cells.
119
3. Experimental
a. Cell Culture: C2BBe1, a subclone of Caco-2 cells was obtained from the
ATCC. More information provided in Chapter II.
b. Preparation of diferric, radiolabeled, and fluorescein isothiocyanate labeled
T f : To prepare diferric Tf, radiolabeled Tf, and fluorescein isothiocyanate
labeled Tf, I followed the procedures which have already been described in
Chapters II and III.
c. Pulse-chase Experiment using 125I-Tf: The pulse-chase assay was performed
as previously described in Chapter II. However, to study the effect of AG10 on
the trafficking of the basolaterally endocytosed Tf, 125I-Tf was added to the
basolateral chamber of the transwell. Caco-2 cells were pulsed for 30 minutes at
37°C and then chased for 45 minutes at 37°C. The radioactivity recovered from
the basolateral chamber was taken as a measurement of Tf recycling, while the
radioactive protein recovered from the apical chamber was used as a measurement
of Tf transcytosis.
120
d. Confocal Fluorescence Microscopy: Preparation of Caco-2 cells for confocal
fluorescence microscopy has been described in Chapter III. In order to study the
effect of AG10 on the intracellular trafficking of Tf internalized at the basolateral
membrane, FITC-Tf conjugate was added to the basolateral chamber of the
transwell.
e. Image Processing: The processing of the collected images has been explained
extensively in Chapter III.
121
4. Results
a. AG10 increased the transcytosis and recycling of the basolaterally
endocytosed Tf in Caco-2 cells. The recycling and transcytosis of the
basolaterally endocytosed Tf were measured in the absence or presence of AG10
(500 μM). As shown in Figure IV-1A,B, the recycling and transcytosis of Tf
increased in the presence of AG10 in Caco-2 cells which were pulsed with 125I-Tf
at the basolateral membrane for 30 minutes at 37°C and chased for 45 minutes at
37°C. The cell-associated Tf decreased in AG10-treated Caco-2 cells. (Figure
IV-1C). Consistent with the results shown in Chapters II and III, the integrity of
the tight junction, which was tested by measuring TEER, was not affected by
AG10 treatment.
124
b. The basolaterally endocytosed FITC-Tf was colocalized with Rab 11 to an
extent in Caco-2 cells treated with AG10. To study the involvement of Rab11 in
the intracellular trafficking of the basolaterally endocytosed Tf, Caco-2 cells,
which were pulsed with FITC-Tf at the apical or basolateral membrane of the
cells in the absence or presence of AG10 for 30 minutes at 37°C, were labeled
with the antibody against Rab11. To determine the exact location of FITC-Tf, the
cells were also labeled with the antibody against ZO-1. The nucleus was labeled
using DAPI. The galleries of the optical sections from the apical towards the
basolateral cell membrane of non-treated and AG10-treated Caco-2 cells are
shown in Figure IV-2. Rab11 staining appeared as a population of cup-shaped
vesicles which began above the tight junction and continued towards the apical
membrane of Caco-2 cells. This is more noticeable in one of the non-treated
Caco-2 cells pulsed with FITC-Tf at the apical membrane (Figure IV-3). In
Figure IV-4, the distribution of FITC-Tf in Caco-2 cells, pulsed with FITC-Tf at
the basolateral membrane, is shown in two single cells chosen from the galleries
which are presented in Figure IV-2. In Caco-2 cells, pulsed with FITC-Tf at the
basolateral membrane, the FITC-Tf was mainly found in the basolateral regions of
the cells. FITC-Tf did not reach the z-sections localized above the tight junction
and there was no colocalization between Rab11 and FITC-Tf (Figures IV-2A, IV-
4A,C).
125
As I have already shown (Figures III-9, III-11), in the cells pulsed with FITC-Tf
at the apical cell membrane at 37°C, FITC-Tf was found in the apical regions of
the cells and above the intracellular level at which the ZO-1 begins to appear
(Figures IV-2B). More FITC-Tf was present in the z-section localized about 1.6
μm to 2.4 μm above the tight junction (Figure IV-3). The colocalization between
the apically endocytosed FITC-Tf and Rab11 is shown in Figure IV-3 (arrows).
In the presence of AG10, the basolaterally endocytosed FITC-Tf had access to the
z-sections localized above the tight junction (Figures IV-2C, IV-4B). The two
individual channels (green and red) of the images which are shown in panels A
and B, and also the superimposed images, are shown in Figure IV-4C,D. The
images presented in panels C and D were collected from the intracellular level
localized above the tight junction of the cells. In the control cell monolayer, the
display for the green channel did not show any FITC-Tf in the z-section localized
about 1.0 μm above the tight junction (Figure IV-4C). On the other hand, I
observed that the FITC-Tf was colocalized with Rab11 in the z-section located
about 1.6 μm above the tight junction of the AG10-treated cell (Figure IV-4D).
The colocalization between the basolaterally endocytosed FITC-Tf and Rab11 is
shown in Figure IV-4D (arrows). The animated three-dimensional projections are
also available (supplementary material Movies IV-1, IV-2, and IV-3).
131
As described in Chapter III, the z-sections were arranged based on their distance
from the tight junction and were plotted versus the average mean intensity of the
green pixels and the average colocalization coefficient of the FITC-Tf. The level
of the collected z-sections was adjusted relative to the tight junction, and the first
z-section containing the highest mean intensity of the red pixels was set to zero,
and positive numbers move towards the APM (Figure IV-5A). Consistent with
the results obtained from the galleries of the optical sections of the cells, the
amount of FITC-Tf increased in the z-sections localized more apically in AG10-
treated Caco-2 cells (Figure IV-5B). FITC-Tf was found in the z-sections
localized about 2.0 μm to 2.4 μm above the tight junction only in the cells that
were treated with AG10 (Figure IV-5B). A higher degree of the colocalization
between the basolaterally endocytosed FITC-Tf and Rab11 was observed in the z-sections
localized from 1.6 μm to 2.4 μm above the tight junction of the AG10-
treated Caco-2 cells. This indicates the presence of FITC-Tf in the Rab11-
containing endosomal compartment (Figure IV-5C).
134
5. Discussion
The involvement of Rab11 in the intracellular trafficking of the apically
endocytosed Tf was studied in the cultured Caco-2 cells (Chapter III). I have
already shown that AG10 increases the transcytosis and recycling of Tf
internalized at the apical membrane at 37°C in Caco-2 cells (Figures III-12, III-
13). I have also presented data showing that the apically endocytosed FITC-Tf
was colocalized with Rab11 at 37°C in Caco-2 cells (Table III-1, Figures III-5,
III-7, III-9, and III-11), as well as the involvement of Rab11 in the apical
recycling of Tf. In the presence of AG10, a higher degree of colocalization
between the apically endocytosed FITC-Tf and Rab11 was observed, indicating
an enhancement of the Tf transport to the Rab11-containing endosomal
compartment in AG10-treated Caco-2 cells (Figures III-16, III-17). However,
some studies have discussed the absence of Tf in the endosomal compartments
containing Rab11 in MDCK cells (20; 167). To further define the role of Rab11
in Tf trafficking in Caco-2 cells, and explain the contrary data obtained from
different studies, I investigated the Tf intracellular pathways in Caco-2 cells
pulsed at the apical or the basolateral membrane in the absence or presence of
AG10. Interestingly, the recycling and transcytosis of the basolaterally
endocytosed Tf increased in AG10-treated Caco-2 cells which were pulsed for 30
135
minutes at 37°C and chased for 45 minutes (Figure IV-1A,B). As expected, the
amount of the cell-associated Tf decreased in the presence of AG10 (Figure IV-
1C). These data suggest that AG10 affects the trafficking of both apically and
basolaterally endocytosed Tf. In both cases, the recycling and transcytosis of Tf
increase, resulting in a decrease in the amount of cell-associated Tf in AG10-
treated Caco-2 cells (Figures III-13, IV-1).
Confocal microscopy studies revealed the presence of different distribution
patterns for FITC-Tf in Caco-2 cells which were pulsed at the apical cell
membrane, compared to the cells pulsed with FITC-Tf at the basolateral cell
membrane. Caco-2 cells were pulsed with FITC-Tf for 30 minutes at 37°C at
either the apical or basolateral cell membrane. The basolaterally endocytosed
FITC-Tf had no access to the apical regions of Caco-2 cells (Figure IV-2A). It
was found mainly at the basolateral membrane of the cells. No colocalization
between FITC-Tf and Rab11 was observed suggesting that the majority of FITC-Tf
was recycled to the basolateral cell membrane (Scheme IV-1: pathway 8) and
did not have access to the Rab11-containing endosomal compartment. This
observation is consistent with the studies reporting the absence of the
basolaterally endocytosed Tf in the Rab11-containing endosomal compartment in
MDCK cells (20; 167).
137
Consistent with the previous observations shown in Chapter III (Figures III-9, III-
11), FITC-Tf was found in the Z-sections localized above the tight junction of
Caco-2 cells which were pulsed with FITC-Tf at the apical membrane (Figure IV-
2B). The apically endocytosed FITC-Tf was colocalized with Rab 11 in the
apical regions of the cells indicating the presence of FITC-Tf in the Rab11-
containing endosomal compartment (Figure IV-3). Despite the fact that the
basolaterally endocytosed Tf does not have access to the Rab11-containing
endosomal compartment in non-treated Caco-2 cells (Figures IV-2A, IV-4C), in
AG10-treated Caco-2 cells the FITC-Tf, internalized at the basolateral membrane,
was found in the z-sections localized above the tight junction (Figures IV-2C, IV-
4D). Analyzing the distribution of FITC-Tf also confirms the presence of FITC-Tf
in a higher degree in the z-sections localized about 0.8 μm to 2.4 μm above the
tight junction of the cells (Figure IV-5B). In the z-sections localized 2.0 μm to
2.4 μm above the tight junction, FITC-Tf was found only in AG10-treated Caco-2
cells (Figure IV-5B). The basolaterally endocytosed FITC-Tf was colocalized
with Rab11 in AG10-treated Caco-2 cells indicating the presence of FITC-Tf in
the Rab11-containing endosomal compartment (Figure IV-5C). In the z-sections
localized 2.0 μm to 2.4 μm above the tight junction, about 35% of the total FITC-Tf
was colocalized with Rab11 in the presence of AG10 (Figure IV-5C) This
may be due to the effect of AG10 on the transcytosis of the basolaterally
endocytosed Tf. In the presence of AG10, the basolateral-to-apical transcytosis of
138
the basolaterally endocytosed Tf increases (Figure IV-1B), resulting in the
presence of Tf in the apical regions of the cells and the colocalization between
FITC-Tf and Rab11 (Figures IV-4, IV-5)
Summary: The majority of the basolaterally endocytosed Tf recycles to the
basolateral membrane (Scheme IV-1: pathway 8) and does not have access to the
Rab11-containing endosomal compartment localized above the tight junction.
The Tf-TfR and pIgR-dIgA complexes, endocytosed from the basolateral
membrane, have been found in the CE. From the CE the Tf-TfR complex
recycles to the basolateral membrane, while the pIgR-dIgA complex on its route
from the basolateral to the apical surface crosses the CE and moves to the ARE
(Scheme IV-1: pathway 4) (7; 20; 41; 48). Interestingly, in the presence of AG10,
the transcytosis of the basolaterally endocytosed Tf increases resulting in the
colocalization between Tf and Rab11 in AG10-treated Caco-2 cells. This may
indicate that the basolaterally endocytosed Tf, after leaving the CE, crosses the
Rab11-containing endosomal compartment on its way towards the apical
membrane in Caco-2 cells (Scheme IV-1: pathway 4). The endosomal
compartment containing Rab11 may be accessible to both apically and
basolaterally endocytosed Tf in Caco-2 cells. A low level of the basolateral-to-apical
transcytosis of Tf could be the reason why the basoalterally internalized Tf
has not been found in this compartment.
139
Chapter V. AG10 Increases the Serum Concentration of the
Orally Administered Biotin-Transferrin Conjugate in CF-1
1. Abstract
The effect of AG10 on the intracellular trafficking of Tf has been studied
extensively in Caco-2 cells. It has been shown that AG10 increases TfR-mediated
transcytosis in both apical-to-basolateral and basolateral-to-apical directions in
Caco-2 cells. The results from cell culture studies were confirmed by in vivo
studies and it was shown that the hypoglycemic response to the orally
administered insulin-Tf conjugate in diabetic rats was more effective in the
presence of AG10. In this chapter, the effect of AG10 on the transport of Tf
through GI epithelia was studied. CF-1 mice were subjected to the Biotin-Tf
conjugate via oral administration. The blood, which was collected at different
time points, was centrifuged and the concentration of the Biotin-Tf conjugate was
determined in the serum by performing ELISA assay. AG10 significantly
increased the serum concentration of the orally administered Biotin-Tf conjugate
in CF-1 mice.
140
2. Introduction
In previous chapters, the effect of AG10 on the intracellular trafficking of Tf has
been studied extensively in Caco-2 cells. It has been shown that AG10 increases
TfR-mediated transcytosis in both apical-to-basolateral and basolateral-to-apical
directions in Caco-2 cells. The selectivity of the effect of AG10 has also been
studied. It was shown that AG10 increases TfR-mediated transcytosis only in
Caco-2 cells and not in MDCK cells. AG10 did not show any affect on the
integrity of the tight junction in Caco-2 cells. This indicates AG10 has low
potential side effects in vitro.
The absorption of insulin-Tf as an intact conjugate through the GI epithelia of
STZ-induced diabetic rats was reported by Xia et al. (178). The absorption of
insulin-Tf conjugate through GI epithelia was evaluated by its hypoglycemic
effect after the oral administration in STZ-induced diabetic rats. Insulin-Tf
conjugate exhibited a slow but prolonged hypoglycemic effect via oral
administration (178). However, a relatively high dose of the conjugate was
required in order to reduce the blood glucose to the normal level in diabetic rats
(178). Interestingly, the hypoglycemic response to the orally administered
insulin-Tf conjugate in diabetic rats was more effective in the presence of AG10
141
(176). The orally administered insulin together with AG10 did not show any
effect on the blood glucose level, indicating the effect of AG10 on TfR-mediated
transcytosis.
As mentioned above, the absorption of insulin-Tf conjugate, with or without
AG10, was evaluated by its hypoglycemic effect which could be affected by
either the transport of insulin-Tf conjugate across the epithelia, or the release of
the active insulin from insulin-Tf conjugate, or both. There were no data
provided to correlate the effectiveness of the hypoglycemic effect of the conjugate
in the presence of AG10 to the effect of AG10 on the transport of the the insulin-
Tf conjugate through the GI epithelia. In order to study the effect of AG10 on the
transport of Tf through the GI epithelia, the serum concentration of Biotin-Tf
conjugate was determined in CF-1 mice. Biotin is a relatively small molecule
with a molecular mass of 244 dalton, which can be conjugated to many proteins
including Tf without altering the biological activity of the protein.
142
3. Experimental
a. Materials: EZ-LinkTMSulfa-NHS-LC-Biotin was obtained from Pierce.
Human Tf and goat anti-human Tf antibody were purchased from Sigma.
Peroxidase-conjugated AffiniPure donkey anti-goat antibody was purchased from
Jackson ImmunoResearch. 4-Hydroxyazobenzene-2-carboxylic acid
(HABA)/Avidin reagent and TMB Microwell Peroxidase Substrate System were
purchased from Sigma and KPL, respectively. AG10 was obtained from
Calbiochem.
b. Preparation of biotin labeled Tf: EZ-LinkTMSulfo-NHS-LC-Biotin was
conjugated to holo-Tf with a molar ratio of 12 to 1 in PBS under constant stirring
for 1 hour at room temperature. The reaction of EZ-LinkTMSulfo-NHS-LC-Biotin
with protein is shown in Figure V-1. The unreacted Sulfo-NHS-LC-Biotin was
removed by dialysis against 2 L PBS at 4°C overnight. The level of the biotin
incorporation was measured by using HABA/Avidin reagent containing 0.3 mM
HABA and 0.45 mg/ml avidin. In a 1 ml cuvette, 100 μl of the biotin-Tf
conjugate was added to 900 μl HABA/Avidin reagent.
143
The absorbance at 500 nm was measured before and after the addition of the
biotin-Tf conjugate to the reagent. The degree of labeling was calculated using
the formula provided by Sigma.
Formula V-1:
ΔA500=0.9A(HABA/Avidin)500-A(HABA/Avidin+sample)500,
where 0.9 is the dilution factor of HABA/Avidin upon addition of sample
μmole biotin/ml=(ΔA500/34)(10),
34=mM extinction coefficient at 500 nm
10=dilution factor of sample into cuvette
mole biotin/mole protein=(μmole biotin/ml sample)/(μmole protein/ml sample)
In average, 4 mole biotins were incorporated in 1 mole Tf.
145
c. Animal Model: Adult male CF-1 mice 6-8 weeks old, weighing 20-25 g, were
obtained from Charles River Laboratories (Wilmington, MA). The animals were
housed under standard laboratory conditions (relative humidity 65±2%,
temperature 23±2°C and 12 hours light-dark cycle). They were fed with standard
rodent pellet diet and tap water. The study was approved by the Institutional
Animal Care and Use Committee of University of Southern California. The
experiments were conducted according to the guide for the Care and Use of
Laboratory Animals (Institute of Laboratory Animal Resources, Commission on
Life Sciences, National Research Council, Washington, DC, 1996).
d. Animal Study: The mice were fasted for 12 hours, then were orally
administered with Biotin-Tf conjugate (1.2 mg) together with and without AG10
(8.5 mg/kg) in sodium bicarbonate (NaHCO3, 30 mg/ml) solution by using a
gavage needle. The formulation with NaHCO3 was used to neutralize the
stomach acid. The mice were then kept in metabolic cages with free access to
water. They were fed 12 hours after the treatment. Blood was collected at certain
time points after the oral administration of the conjugate, with and without AG10.
The collected blood was subjected to the centrifugation at 3000 rpm for 30
minutes. The concentration of the Biotin-Tf conjugate in the serum was
determined by using a standard enzyme-linked immunosorbent assay (ELISA)
procedure.
146
e. ELISA Assay: The samples were incubated in streptavidin-coated wells
(StreptaWell, Roche) for one hour on a platform shaker at room temperature.
They were then washed with ice-cold PBS (0.1% Tween 20) four times, 5 minutes
each time, followed by an incubation with goat anti-human Tf antibody (Sigma)
diluted in 1% BSA in PBS for 1 hour on the shaker at room temperature. After
further washing with ice-cold PBS (0.1% Tween 20) for a total of 20 minutes, the
samples were incubated with peroxidase-conjugated AffiniPure donkey anti-goat
antibody (Jackson ImmunoResearch) for 2 hours on the shaker at room
temperature. The samples were then washed again as explained above and
subjected to TMB Microwell Peroxidase Substrate System (KPL). The reaction
was stopped by adding 30 μL of H2SO4 (1N), and the absorbance was measured at
450 nm by using a plate reader. The serum concentration was determined using a
standard curve obtained from known concentrations of biotin-Tf conjugate versus
their absorbance at 450 nm. The ELISA assay is summarized in Figure V-2.
148
4. Results
a. The serum concentration of Biotin-Tf conjugate reached the maximum at 8 to
12 hours after the oral administration of the conjugate. The blood of CF-1 mice
was collected at 4, 8, 12, and 24 hours after the oral administration of the Biotin-
Tf conjugate (1.2 mg). The blood was then subjected to centrifugation, and the
concentration of the Biotin-Tf conjugate was determined in the serum by
performing the ELISA assay. The serum concentration of the Biotin-Tf conjugate
reached the maximum level at 8 to 12 hours after the oral administration of the
conjugate (Figure V-3). The serum concentration of the conjugate decreased over
the next 12 hour period in CF-1 mice (Figure V-3).
150
b. AG10 increased the serum concentration of the Biotin-Tf conjugate in CF-1
mice. The serum concentration of the Biotin-Tf conjugate was determined at 8,
16, 24, and 48 hours after the oral administration of the conjugate (1.2 mg)
without (CON) or with AG10 (8.5 mg/kg). As shown in Figure V-4, the serum
concentration of the conjugate increased significantly in AG10-treated mice at 16
hours after the oral administration. Even at 24 hours after the oral administration
a slight increase in the serum concentration of the conjugate was observed in the
presence of AG10. The serum concentration of Biotin-Tf conjugate in the
absence of AG10 reached a minimum at 24 hours after the oral administration,
and no further change in the conjugate concentration was observed at 48 hours
(Figure V-4).
152
5. Discussion
It has been demonstrated that the orally administered insulin-Tf conjugate
significantly reduced the plasma glucose level in STZ-induced diabetic rats (178).
The hypoglycemic effect of the insulin-Tf conjugate was further potentiated by
AG10 treatment (176).
In order to study the effect of AG10 on the intestinal absorption of Tf conjugates,
the serum concentration of Tf in CF-1 mice was measured in non-treated and
AG10-treated mice. To determine the serum concentration of the Tf, it was
conjugated to the biotin molecule (Figure V-1). Conjugation to biotin, which is a
relatively small molecule, does not alter the biological activity of the Tf molecule.
Measurement of the serum concentration of Biotin-Tf conjugate at 4, 8, 12, and
24 hours after the oral administration of the conjugate revealed the presence of the
conjugate in a high level at 8 and 12 hours after the oral administration (Figure V-
3). The pharmacokinetic study of the intravenously administered Biotin-Tf
conjugate in CF-1 mice showed a half–life of about 24 hours for the conjugate
(Dr. Qi, unpublished observations). On the other hand, the results obtained from
the orally administered Biotin-Tf conjugate did not provide enough information to
determine the half-life of the conjugate (Figure V-3).
153
The orally administered Biotin-Tf conjugate may be subjected to the metabolism
which may be taking place in the GI, and particularly in the liver. The presence
of the biotin molecule in the conjugate may also increase the elimination rate of
the orally administered conjugate resulting in a shorter half-life for the conjugate.
Biotin is a vitamin found in tissue and blood. The presence of the biotin receptor
on the hepatic plasma membranes has been reported (161). This raises the
possibility of a higher transport of the Biotin-Tf conjugate to the liver resulting in
an increased metabolism of the conjugate. In addition, the presence of a sodium-dependent
multivitamin transporter (SMVT) for biotin has been established on
human intestinal and liver epithelial cells (9). However, we may consider that the
Biotin-Tf conjugate c