University of Southern California Dissertations and Theses

TfR-mediated oral delivery of protein drugs: Oral delivery of recombinant G -CSF -Tf fusion protein and its spacer optimization

(USC Thesis Other)

TfR-mediated oral delivery of protein drugs: Oral delivery of recombinant G -CSF -Tf fusion protein and its spacer optimization

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content TFR-MEDIATED ORAL DELIVERY OF PROTEIN DRUGS:
ORAL DELIVERY OF RECOMBINANT G-CSF-TF FUSION PROTEIN AND ITS
SPACER OPTIMIZATION
Copyright 2006
by
Yun Bai
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALLIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PHARMACEUTICAL SCIENCES)
May 2006
Yun Bai
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3233786
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI
UMI Microform 3233786
Copyright 2006 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Acknow ledgem ents
I would like to express my gratitude to my advisor, Dr. Wei-Chiang Shen, for his
encouragement and support through my dissertation research. He helped me grow in
every aspect as a researcher. His enthusiasm and meticulous care to details deeply
influenced my attitude towards research.
My thanks also go to all my colleagues in Dr. Shen’s lab. Their support and advices
helped me overcome many obstacles in my research.
Dr. David K. Ann in the Department of Pharmacology and Toxicology is our
collaborator in this project. He gave me tremendous help. Other members in my
committee, Dr. Austin Yang, Dr. Stan G. Louie and Dr. James J.H. Ou also gave me a
lot of invaluable advices.
My husband, Tianwei Yu, is the one who made this dissertation possible. His
understanding, encouragement and constant care brought me through the tough times
when I feel unconfident about myself and my work. My parents, Bingzhe Bai and
Ying Zhong receive my deepest gratitude for their love and support.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table o f Contents
Acknowledgements ii
List of Tables vii
List of Figures viii
Abstract xii
Preface xiv
Introduction 1
1. The recent advances in intestinal macromolecular drug delivery
via receptor - mediated transport pathways. 2
2. Transferrin is a good vehicle for drug delivery 3
2.1 The general information of Tf protein 3
2.2 The transferrin receptor 6
2.3 The binding of Tf to the Tf receptor 6
2.4 Utilization of Tf for drug delivery 11
2.5 TfR-mediated transepithelial protein-drug delivery 12
3. The granulocyte colony stimulating factor (G-CSF) as a drug 14
3.1 The G-CSF protein 14
3.2 The signal transduction for G-CSF myeloid differentiation
effect 16
3.3 The pharmacokinetics of G-CSF 18
3.4 The modified recombinant G-CSF. 19
4. The advantages of recombinant fusion protein over chemical
conjugates. 19
4.1 The expression of the recombinant fusion protein 20
4.2 The protein free HEK293 mammalian expression 21
5. The linkages between moieties in the fusion protein 21
Materials and methods 23
1. Cell culture 23
1.1 HEK293 cell culture 23
1.2 Caco-2 cell culture 24
1.3 NFS60 cell culture 25
1.4 Human bladder carcinoma 5637 cell culture 25
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.5 The BHK cell culture 26
1.6 Pretreatment of cells 26
2. Plasmid construction 27
2.1 Plasmid isolation 27
2.2 Total RNA isolation by TRIZOL reagent (GIBCO-BRL) 28
2.3 RT-PCR 29
2.4 Gene clean 32
2.5 PCR product cloning 34
2.6 Transformation 35
2.7 The mutation of Tf Ser288 to Tf Cys288 36
3. Protein expression 37
3.1 Transfection (Lipofectamine 2000) 37
3.2 CD 293 medium 38
3.3 The G-CSF protein expression in DH5-a E. coli 38
3.4 The purification of inclusion body 39
3.5 The purification of GST-G-CSF fusion protein using GST
beads from inclusion body 40
3.6 Ammonium sulfate precipitation 40
4. Fusion protein analysis 42
4.1 SDS-PAGE analysis 42
4.2 Western-blotting analysis 42
4.3 The development of stable transfected cell lines 43
5. The linker formation 44
5.1 The primer design of the linkers 44
5.2 Primer annealing 44
5.3 The linearization fusion protein plasmid by digestion with
X hol 46
5.4 Phenol - chloroform extraction 47
6. The studies of bioactivity of fusion protein 48
6.1 The in vitro activity test 48
6.1.1 Iodination of transferrin 48
6.1.2 Caco-2 cell surface binding study 49
6.1.3 NFS-60 cell proliferation - MTT test 49
6.2 The in vivo activity test 50
7. Statistical analysis 52
Results 53
1. Conjugation of recombinant T f with a cysteinyl residue 53
1.1 Expression of Tf protein in mammalian cell 54
1.2 The mutation of Tf sequences 60
2. The cloning and expression of G-CSF 62
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.1 The cloning of G-CSF cDNA from Human Bladder
Cacinoma cell line 5637 62
2.2 Expression of G-CSF protein in BL21 bacteria 66
2.3 The purification of GST-G-CSF fusion protein with GST-
affmity beads from inclusion body 68
2.3.1 Purification of inclusion body 68
2.3.2 The renaturing of GST-G-CSF from inclusion body and
isolation of G-CSF by thrombin-cutting 68
3. Construction of the G-CSF-Tf fusion protein plasmid 71
3.1 The subcloning of G-CSF sequence 72
3.2 Subcloning the Tf sequence 74
3.3 The fusion protein expression plasmids construction 76
3.4 The expression of the fusion proteins 79
3.5 The expression of fusion protein in large scale and the fusion
protein purification 85
4. The biological activity of the G-CSF-Tf fusion protein 89
4.1 In vitro G-CSF and Tf activity of the fusion protein. 89
4.2 In vivo studies. 91
4.2.1 The in vivo activity by injection 91
4.2.2 The oral availability of fusion protein sG-CSF-Tf 91
4.2.3 The dose response study of the orally administered G-
CSF-Tf 94
4.2.4 The competition of G-CSF-Tf fusion protein oral
absorption by Tf and albumin 96
5. The linker optimization 99
5.1 The design of linkers 99
5.2 Linker insertion 99
5.3 The linkers sequence formation by annealing and insertion 102
5.4 Expression of fusion proteins and Westem-blot 111
5.5 Linker selection by in vitro G-CSF activity 112
6. Comparison of the biological activity between H4-2 and LE
fusion proteins 114
6.1 In vitro activity comparison 114
6.2 In vivo activity comparison 117
6.2.1 In vivo characterization of the fusion proteins by
subcutaneous injection 117
6.2.2 In vivo myelopoietic activity of the fusion proteins by
oral administration 121
Discussion 125
1. Conjugation of recombinant Tf with a cysteinyl residue 125
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2. Recombinant granulocyte colony-stimulating factor-transferrin
fusion protein as an oral myelopoietic agent 126
3. The linker optimization 131
4. The sustained increase of ANC observed in orally administered
G-CSF-Tf fusion protein 140
5. Summary 146
Future prospective 149
1. What is the underlying mechanism of the in vivo myelopoietic
activity of the orally administered G-CSF-Tf fusion protein? 149
2. Can the bioactivity and their therapeutic efficacy of the G-CSF-
Tf fusion protein be further improved? 151
3. Can the fusion protein strategy be applied to other therapeutic
protein/peptide drugs? 153
References 155
v i
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List o f Tables
Table 1. Design of primers for the linker 45
Table 2. Molecular weight of the linkers 111
Table 3. The EC50 of the fusion proteins in the NFS60 proliferation assay 112
v ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List o f Figures
Figure 1. Types of ligands transported by receptor-mediated endocytosis
mechanisms 4
Figure 2. The structure of human serum transferrin 5
Figure 3. Atomic model of the TfR-dTf Complex 8
Figure 4. The endocytosis cycle of Tf 9
Figure 5. The model of Tf binding to TfR 10
Figure 6. Ribbon diagram of rhG-CSF 15
Figure 7. Model of signal transduction at the G-CSF receptor 17
Figure 8. The Map of TOPO -TA cloning vector from invitrogen 33
Figure 9. The mechanism of Topo-TA cloning technology 35
Figure 10. Molecular model of HST showing the location of S288 54
Figure 11. The subcloning of Tf from TFR27A plasmid 56
Figure 12. The map of the pCMV-Tag vector 57
Figure 13. Selection of positive pCMV-Tag -T f plasmid by endonuclease
digestion 58
Figure 14. The BHK cell expressed Tf recognized by both anti- Tf and
anti-Flag antibodies 59
Figure 15. Site-specific mutation of recombinant Tf 60
Figure 16. The isolation of total RNA from 5637 cells 63
Figure 17. The RT-PCR of the G-CSF gene with and without secretion
signal 65
Figure 18. The GST-G-CSF protein expression 67
v iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 19. The GST-G-CSF fusion protein expressed in the inclusion body 69
Figure 20. Thrombin cutting of the GST-G-CSF fusion protein 70
Figure 21. The maps of the fusion protein constructs 72
Figure 22. The PCR products of G-CSF with (sG-CSF) or without
secretion signal (G-CSF) on agarose gel 73
Figure 23. The PCR products of Tf on agarose gel 75
Figure 24. The map of pcDNA3.0 77
Figure 25. Verifying the insert size using endonuclease digestion 78
Figure 26. The Westem-blotting test for the conditioned medium and cell
lysates of transfected BHK cells 81
Figure 27. The comparison of biological activity of the fusion proteins,
sG-CSF-Tf and sTf-G-CSF, using NFS-60 proliferation assay 82
Figure 28. The expression of fusion protein sG-CSF-Tf and sTf-G-CSF in
HEK293 cells 83
Figure 29. The 3-D structure model of G-CSF-Tf 84
Figure 30. Test of the fusion protein expression in SDS-PAGE 86
Figure 31. Expression and identification of G-CSF-Tf fusion protein 87
Figure 32. In vitro study of G-CSF-Tf fusion protein activity 90
Figure 33. Myelopoietic effect of s.c. administered the fusion protein, G-
CSF, or the control 92
Figure 34. Myelopoietic effect of orally administered fusion protein, G-
CSF, or control 93
Figure 35. The dose-dependency of orally administered recombinant
fusion protein 95
Figure 36. Competition by free Tf on the absorption of orally administered
G-CSF-Tf fusion protein 97
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 37. Competition by Albumin on the absorption of orally
administered G-CSF-Tf fusion protein 98
Figure 38. The six fusion constructs, in comparison with the original no
linker construct (denoted LE) 100
Figure 39. The linearization of the plasmid by Xhol cutting 101
Figure 40. Confirming the size of the linearized plasmid 102
Figure 41. The digestion of fusion constructs with linker- inserted
plasmids by Pst 1 103
Figure 42. The design of PCR experiment to test the copy number of the
linker 104
Figure 43. The selection of the GS fusion protein plasmid 105
Figure 44. The selection of the H3 fusion protein plasmids 108
Figure 45. The selection of H4 plasmid 109
Figure 46. Recognition of the fusion proteins by both anti-Tf and anti-G-
CSF antibodies 110
Figure 47. Evaluation of G-CSF activity of the G-CSF-Tf fusion proteins.
Proliferation of the murine myeloblastic cell line NFS-60 113
Figure 48. The SDS-PAGE analysis of the purified fusion proteins LE and
H4-2 114
Figure 49. In vitro characterization of the fusion proteins H4-2 and LE.
Evaluation of G-CSF activity of the purified G-CSF-Tf fusion
protein 115
Figure 50. The TfR -binding activity of the fusion proteins determined in
Caco-2 cell monolayers 116
Figure 51. Myelopoietic effect of subcutaneously administered fusion
proteins H4-2, LE, and G-CSF in BDF1 mice 118
Figure 52. Myelopoietic effect of subcutaneously administered different
dose of LE 119
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 53. Myelopoietic effect of subcutaneously administered different
dose of H4-2
Figure 54. Myelopoietic effect of subcutaneously administered different
dose of G-CSF
Figure 55. The dose -response curve of the in vitro activity of fusion
protein LE and H4-2 and G-CSF
Figure 56. Myelopoietic effect of orally administered fusion proteins H4-
2, LE, and G-CSF in BDF1 mice
Figure 57. Myelopoietic effect of subcutaneously administered fusion
proteins G-CSF, LE and H4-2 at different doses in BDF1 mice
Figure 58. A hypothetical diagram on the absorption of Tf-fusion proteins
by intestinal epithelial cells
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Abstract
An expression construct harboring granulocyte colony-stimulating factor (G-CSF)-
transferrin (Tf) fusion protein (G-CSF-Tf) was engineered by fusing human cDNAs
encoding G-CSF and Tf to explore the feasibility of using T f as a carrier moiety for
oral delivery of therapeutic proteins. The recombinant protein, G-CSF-Tf, was
harvested from protein-free, conditioned medium of transfected HEK293 cells. The in
vitro studies demonstrated that the purified G-CSF-Tf fusion protein possesses the
activity of both Tf receptor (TfR) binding in Caco-2 cells and G-CSF-dependent
stimulation of NFS-60 cell proliferation. Subcutaneous administration of G-CSF-Tf
fusion protein to BDF1 mice demonstrated a pharmacological effect comparable to the
commercial G-CSF on the increase of absolute neutrophil counts (ANC). However,
the fusion protein elicited a significant increase in ANC upon oral administration to
BDF1 mice, whereas G-CSF had no effect. This study also showed that orally
administered G-CSF-Tf elicits a sustained myelopoietic effect up to 3 days, whereas
the effect of subcutaneously administered G-CSF or G-CSF-Tf lasts only 1 day.
Furthermore, coadministration of free T f abolished the increase of ANC by orally
delivered G-CSF-Tf, suggesting that the recombinant protein is absorbed via a TfR-
mediated process in the gastrointestinal tract.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
To improve the activity, six fusion proteins with two types of linkers between the G-
CSF and T f moieties were developed, including the flexible linker (GGGGS)3,
denoted GS, and helix-forming linkers A(EAAAK)n A (n=2~4), denoted H2, H3, H4.
To make longer linkers, two or three copies of H4 linker were also inserted (denoted
H4-2, H4-3). In the NFS-60 cell proliferation assay, the activity of the fusion protein
with H4-2 linker, which is the most potent, is ten-fold higher than the fusion protein
with short LE linker. On the other hand, all fusion proteins show similar binding
affinity in Caco-2 TfR binding assay. Nevertheless, the fusion protein with H4-2
linker showed higher myelopoitic effect than the protein with LE linker in BDF1 mice
model, both by subcutaneous and oral administration.
Taken together, we conclude that the Tf-based recombinant fusion protein technology
represents a promising approach for future development of orally effective peptide and
protein drugs.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Preface
The application of biotechnological products as therapeutics for treatment of diseases
is limited to invasive delivery, due to the poor absorption of proteins and peptides
across epithelial cells. Since most of the recombinant proteins are used to treat
chronic diseases, frequent injections can cause inconvenience, poor compliance, and
adverse sides. Therefore, the development of the oral delivery systems for protein and
peptide drugs has long been sought, by the pharmaceutical industry.
The use of glycoprotein transferrin (Tf) as a carrier molecule for the receptor-mediated
transcytosis of protein drugs is a novel way to deliver these large molecules across the
gastro-intestinal (GI) epithelium. In our laboratory, chemically linked insulin-Tf and
G-CSF-Tf conjugates have been demonstrated to be feasible for protein drug oral
delivery. However, current conjugate synthesis technique results in the formation of
products of heterogenic composition. It is also limited by the ability to produce large
amounts of pure product in a cost effective manner.
In this project, we developed a series of orally bioavailable G-CSF-Tf recombinant
fusion proteins, that consist of the transcytotic receptor binding domain (Tf) and a
therapeutic domain (G-CSF), for the treatment of human diseases. And the success of
xiv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
this project also demonstrated that Tf could be used as a delivery vehicle to improve
the GI absorption of other peptide and protein drugs.
xv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Introduction
During the past two decades, recombinant DNA technology has made great progress
in producing a large number of recombinant human peptides and proteins that possess
therapeutic potential (Gosse et al. 1996). A recent survey lists 324 biotechnology
medicines, either in human clinical trials or under review by regulatory agencies. It is
estimated that the protein therapeutic market will grow rapidly at a compound annual
growth rate of 10.5%, and will double in market value from 2003 to 2010. Oral
administration of medicines is the preferred and most widely used route of
administration. However, most peptide and protein drugs are administered
parenterally, due to their large and bulky size, charge and hydrophilicity, and
sensitivity to digestive enzymes (Goldberg & Gomez-Orellana 2003).
Using receptors as targets and receptor-binding ligands as vectors for transcellular
transport is a novel way of achieving selective delivery of peptide and protein drugs
across the GI epithelium (W idera et al. 2003c). The iron-binding glycoprotein
transferrin (Tf) presents an ideal vehicle for the receptor-mediated delivery to across
the intestine epithelium.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1. The recent advances in intestinal macromolecular drug delivery
via receptor - mediated transport pathways.
Oral bioavailability of the peptide and protein drug is poor due the low cellular
penetration and high susceptibility to metabolic enzymes present within the
gastrointestinal tract. The receptor - mediated system is different from the carrier -
mediated system, which cross the membrane via either an energy dependent (ATP or
cotransport) flip-flop mechanism or a facilitated diffusion process. While the
receptor-mediated systems are internalized via vesiclular transport system after
binding to respective receptors on the plasma membrane. These processes of
membrane deformation are named ‘endocytosis’, which includes phagocytosis,
pinocytosis, receptor-mediated endocytosis (clathrin-mediated), and potocytosis (non-
clathrin-mediated RME). Types o f ligands transported by receptor-mediated
endocytosis mechanisms are shown in figure 1 (adapted from Swaan, 1998).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2. Transferrin is a good vehicle for drug delivery
2.1 The general information of Tf protein
Transferrin (Tf), a natural protein that delivers iron into cells, contains 679 amino acid
residues and has a molecular weight of approximately 79 kD (Gomme et al. 2005).
Serum Tf is a member of a homologous family of proteins, known collectively as the
transferrins, which also includes ovotransferrin (oTf), of avian egg white, and
lactoferrin (LTf) found in many secretory fluids and in white blood cells (Brock 1985;
Thorstensen & Romslo 1990). A melanotransferrin (also called P97) was recently
found anchored to the membrane surface of melanocytes and other cells (Brown et al.
1982). The serum T f is the most studied among the T f family. In my dissertation,
unless noted otherwise, Tf means the serum transferrin.
The protein structure of T f is stabilized by 19 intra-chain disulfide bonds and is
protected by three carbohydrate side chains. The Tf molecule has two lobes, the N-
lobe (336 amino acids) and the C-lobe (343 amino acids), which are connected by a
short spacer. In each lobe, there is a deep, hydrophilic metal ion-binding site, which
has four conserved amino acids including two tyrosines, one aspartic acid and one
histidine (N-terminal lobe- Asp-63, Tyr-95, Tyr-188, and His-249). The X-ray crystal
structure of T f is shown in figure 2.
3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Tf is widely distributed in various body fluids including plasma, bile, amniotic,
cerebrospinal, lymph and breast milk (Qian et al. 2002). Tf, which is synthesized
predominantly by hepatocytes (Beutler et al. 2000), has the in vivo half-life of about
eight days. The plasma concentration of Tf ranges from 2 to 3 grams per liter.
Serum transport
proteins and antibodies
IgE
IgG, via Fc receptors
Low density lipoprotein
Maternal IgG
Polymeric IgA
T ranscobalamin
Transferrin
Vitamins and metal ions
Iron/transferrin
Folate
Riboflavin
Vitamin B 1 2
Hormones and Toxins and lectins Viruses and bacteria
growth factors_________________________________________________________
Calcitonin Cholera toxin Adenovirus
Catecholamines Concanavalin A Rous sarcoma virus
Epidermal growth Diphtheria toxin Semliki forest virus
factor Pseudomonas toxin Vesicular stomatitits
Glucagon E. coli heat labile virus
Growth hormone toxin Rotavirus
Insulin Staphylococcal Varicella zoster
Interferon enterotoxin A & B Adenovirus
Leuteinizing hormone Toxic shock syndome Reovirus
Nerve growth factor toxin I Potato leafroll virus
Platelet derived growth Ribosome-inactivating L. plantarum
factor proteins: V. cholerae
Prolactin Ricin E. coli
Thyroid stimulating Saporin Y. pseudotuberculosis
factor Viscumin Klebsiella strains
Thyroid hormone Modeccin Enterobacter strains
Nigrin b
Alpha-sarcin
Serratia strains
Figure 1. Types of ligands transported by receptor-mediated endocytosis mechanisms
(Swaan 1998).
4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
N-lobe
His585 (249)
Asp392 (63)
X ’ V ' 0 F e 3*
\ j j \ C 0 32
Tyr517 (188) ^
Tyr426 (95)^
B
Figure 2. The structure of human serum transferrin. (A) X-ray crystal structure of Tf
human serum transferrin. The C-lobe, with Fe3 + bound, is shown in a closed form and
the apo N-lobe is in an open form. (B) The metal binding sites of human serum
transferrin with residue numbers of the N-lobe in brackets (Li & Qian 2002).
5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.2 The transferrin receptor
The transferrin receptors (TfR) are highly expressed in rapidly proliferating cells. For
example, there could be up to 100,000 TfR molecules per carcinoma cell. There are at
least two distinct types of TfRs, TfRl and TfR2 (Hemadi et al. 2004). The TfR is a
transmembrane homodimer that consists of two identical monomers with a molecular
mass of approximately 90 kDa (Jing & Trowbridge 1987). The protein HFE also
binds to TfRl. The mutation of HFE will cause hereditary hemochromatosis. The
mechanism by which HFE regulates iron uptake into the body is unknown. TfR2, a
new TfR-like family member, is primarily expressed in the livers of humans and mice
(F lem ing et al. 2000; Kawabata et al. 2001). Deficiency of TfR has been
demonstrated to cause an hereditary hemochromatosis-like phenopyte, which indicates
that TfR2 participates in iron homeostasis (Girelli et al. 2002). The structure of Tf-
TfR complex is shown in figure 3 (Richardson 2004).
2.3 The binding of Tf to the Tf receptor
The encocytosis cycle of T f has been well studied. Binding of Tf to the TfRl leads to
receptor-mediated endocytosis, as shown in figure 4. Iron - laden transferrin binds to
TfR on the surface. These complexes localize to clathrin-coated pits. The protein
pump decreases the pH within the endosomes, leading to conformational changes in Tf
6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
that result in the release of iron. The iron transporter DMTI moves iron across the
endosomal membrane. The Apo-Tf-TfR complex is recycled to the cell surface. In
the neutral pH environment, the Apo-Tf dissociates from the receptor, and can be
used in the next cycle after reloading of iron in the liver (Andrews 1999).
It was proved that a conformational change in Tf is thought to occur upon binding to
the TfRl that promotes Fe release (Baker et al. 2003; Richardson & Ponka 1997).
Furthermore, acidification of the endosome by membrane bound proton pumps and the
reduction of Fe(III) to Fe(II) are also involved in release and transport of this metal ion
into the cell (Morgan 1981; Richardson & Ponka 1997). Once released from Tf, Fe
then passes through the endosomal membrane via the divalent metal transporter 1
(DMTI) (Fleming & Andrews 1998).
The cellular uptake of Fe mediated by Tf is controlled by TfRl expression that is
modulated by intracellular Fe levels via iron-regulatory proteins (Dupic et al. 2002).
To add further complexity, the interaction of T f with the TfRl is also regulated by the
competitive binding of the protein, HFE, that is mutated in the Fe-loading disease,
hemochromatosis. This molecule assembles with |32-microglobulin (|32-m) to form a
complex that binds to the TfRl at a site involved in Tf-binding (Lebron & Bjorkman
1999).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3. Atomic Model of the TfR-dTf Complex. Modeled Structure of the TfR-
apoTf Complex, and Interactions of TfR with HFE and the T f N-Lobe. (A) Face-on
view and (B) modeled structure of a ternary complex of the TfR dimer with one HFE
and one Tf molecule bound.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
C lathrin-
c o ated pit
Mitochondria,
Ferritin
A Non-erythroid/
cells /
Proton
Haemosiderin
Acidified
endosome
N ature Reviews | G enetics
Figure 4. The endocytosis cycle of Tf (Andrews 2000).
Recently, By implementing cryo-electron microscopy (cryo-EM) and single particle
averaging techniques, Cheng and colleagues (2004) elucidated the structure of the Tf-
TfRl complex and also provided a new model of TfRl function that results in a more
plausible mechanism of Fe release (Cheng et al. 2004).
The diagram in figure 5 illustrates the lateral interaction of the C-terminal lobe of Tf
with the TfRl and the fact that the N-terminal extends into the gap between the bottom
of the receptor ectodomain and the membrane (A). (B) Upon binding to the TfRl
9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
there is a change in Tf conformation that aids Fe release. (C) Upon the release of Fe,
the C- and N-terminals open and may provide access to other proteins such as the
ferrireductase and/or the divalent metal ion transporter 1 (DMT1). These molecules
may potentially interact with the TfRl at its apical surface. (D) After removal of Fe
and the formation of apoTf, there is probably a further alteration in T f conformation
leading to decreased affinity for the TfRl and its release. (E) Interaction of the HFE-
(32-microglobulin (HFE-p2-m) complex with the TfRl competes with Tf at the same
binding site on the receptor.
DMT1 or
ferrireductase
Figure 5. The schematic model of T f binding to TfR (Cheng et al. 2004).
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.4 Utilization of Tf for drug delivery
The special endocytosis cycle of Tf-TfR make T f a good carrier for drug delivery. Tf
is generally studied in targeting for anti-cancer therapeutics and of carriers for drug to
across blood brain barrier. Great variety of cytotoxic agents have been conjugated to
Tf and investigated as potential anti-cancer therapeutics, for example, methotrexate
(Kratz et al. 1998), doxorubicin (Kratz et al. 2000), cisplatin (Iinuma et al. 2002),
ricin A (Kornfeld et al. 1991), daunorubicin (Bejaoui et al. 1991), and so on. The
monoclonal antibody for rat TfR, 0X26, demonstrated an ability to preferentially
target the TfR of the blood brain barrier (BBB) (Fishman et al. 1987). Recently, the
new type of Tf, named as P97 has shown the same effect for crossing BBB (Demeule
et al. 2002). The anti-TfR antibody, rather than Tf, was chosen as the carrier for this
blood to central nervous system transport model due to the high level of endogenous
T f in blood. However, for oral drug administration, since there is very little
endogenous T f in the gastrointestinal (GI) tract, the construction of fusion proteins
with a T f , rather than anti-TfR moiety should be suitable for the development of
protein drug for oral drug delivery.
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.5 TfR-mediated transepithelial protein-drug delivery
Using receptors as targets and receptor-binding ligands as vectors for transcellular
transport is a novel way to achieve the selective delivery of protein drugs across
epithelium. Several absorptive epithelia have been demonstrated to express TfR, such
as the small intestine (Shah & Shen 1996) and type II pneumocyte (Widera et al.
2003a). There are several advantages of T f to be the drug carrier. First, the
transcytosis of TfR does not adversely affect the structure of the membrane, and will
have less side effects. On the other hand, most methods currently used to enhance
epithelial absorption of protein drugs, such as the use of penetration enhancers like
bile salts and lipids, will affect the integrity of the epithelial membrane. Second,
diferric T f is relatively stable in GI tract. It has been reported that diferric T f is
partially resistant to chymotrypsin, which is responsible for the degradation of the
majority of the proteins in GI (Azari & Feeney 1958). Third, T f is a natural transport
protein. Therefore, Tf is going to have less side effect on cell biological function.
Fourth, TfR has been found to be present at high levels within the rat and and human
GI epithelium. However, there are also some dilemma when considering the use of
TfR for the transcytotic delivery of ligands. TfR are predominantly expressed on the
basolateral membranes (Wan et al. 1992), from which they will be endocytosed. And
the majority of the endocytosed TfRs will recycle back to the same side of the
membrane. These would tend to preclude the ability to achieve apical-to-basolateral
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
transcytosis of Tf-drug conjugates. However, there is indication that the apical- to
-basolateral transepithelial transport may still be achievable, especially in the areas
where large numbers of semi-differentiated cells are present, for example the crypt of
Lieberkuhnin of the small intestine. Furthermore, it has been reported that the
transcytosis of T f can be enhanced in the presence of Brefeldin A (BFA), a fungal
metabolite which has profound effects on the structure and function of the Golgi
apparatus. In Caco-2 cell monolayer the BFA causes a marked decrease in the number
of basolateral T f receptor along with a slight increase in the number of apical TfR
(Shah & Shen 1994). It has also been reported that BFA could enhance the TfR-
mediated transcytosis of both 1 2 5 I labled Tf in both apical to basolateral and basolateral
to apical directions (Prydz et al. 1992). Similarly, tyrphostin 8 (AGIO) , a small
GTPase inhibitor, has also been shown to increase TfR-mediated transcytosis both in
vitro an in vivo(Xia & Shen 2001). The mechanism of AGIO -enhanced TfR
-transcytossi is still unknown.
Hypoglycemic effect was observed with the orally administered Insulin-Tf chemical
conjugate (Xia et al. 2000). The oral administration of the Insulin-Tf conjugate to
streptozotocin-induced diabetic rats significantly reduced plasma glucose level by
about 80 %. The G-CSF-Tf chemical conjugate are orally bioavailable in BDF1 mice
(Widera et al. 2004).
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3. The granulocyte colony stimulating factor (G-CSF) as a drug
3.1 The G-CSF protein
The granulocyte colony-stimulating factor (G-CSF)is a member of the family of
glycoprotein growth factors that control the survival, proliferation and differentiation
of hematopoietic progenitor cells and functionally activate mature hematopoietic cells
(Demetri & Griffin 1991). G-CSF was first purified from conditioned fluid in the
lungs of mice injected with bacterial endotoxin (Nicola et al. 1983). The recombinant
human granulocyte colony-stimulating factor (rh G-CSF) is a member of the CSF
family of hormone-like glycoproteins that regulate hematopoietic cell proliferation and
differentiation and has been widely used for treating congenital and acquired
neutropenias (Ishiguro et al. 1993).
G-CSF is a glycoprotein consisting of 174 amino acids, with a molecular weight of
about 20 kDa (Souza et al. 1986). The crystal structure of G-CSF is shown in figure
6. G-CSF was first cloned from the bladder carcinoma cell line 5637 and expressed in
E coll., with the commercial name Filgrastim. It was approved as a medicine for
chemotherapy-induced neutropenia, by the US FDA in 1991 and successfully
commercialized by Amgen, Inc. (Thousand Oaks, CA). There is another recombinant
G-CSF cloned from human squamous carcinoma CHU-II and expressed in COS cells.
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Its commercial name is Lenograstim (produced by Chugai Pharmaceuticals, Tokyo,
Japan). Both forms have similar bioavailability and bioactivities following
subcutaneous or intravenous administration (Watts et al. 1997).
Figure 6. Ribbon diagram of rhG-CSF. The main bundle helices A (residues 11-39), B
(71-91), C (100-123), and D (143-172) are labeled near their N termini (Souza et al.
1986).
G-CSF, produced mainly by macrophages, induces proliferation of neutrophil colonies
and differentiation of precursor cells to neutrophils. In addition, it stimulates the
activity of mature neutrophil. It has been shown that the human G-CSF is protective
in a broad variety of animal infection models in terms of improved survival, reduced
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
bacterial loads, enhanced neutrophil action and immigration into infected sites(Lyman
et al. 2002).
G-CSF is an important drug for the treatment of several immune disorders and
complications associated with chemotherapy. In many indications, such as
neutropenia, the G-CSF needs to be injected chronically. The administration of G-
CSF is at doses of 1 -20 pg/kg per day. Higher doses may be required in patients with
severe congenital neutropenia (SCN).
3.2 The signal transduction for G-CSF myeloid differentiation effect
G-CSF induced myeloid differentiation has been demonstrated to be mediated by
STAT3 activation (W ard et al. 1999). The G-CSF receptor contains single
extracellular, transmembrane, and cytoplasmic domains. Seven different human G-
CSF receptor isoforms have been described which are generated by alternative
splicing (Fukunaga et al. 1993).
The interaction of G-CSF with its receptor a subunit is the first step in the formation
of a signaling-competent receptor complex. This process involves the oligomerization
o f the G-CSF bound subunit with another G-CSF bound ( 3 subunit. This
oligomerization initiates the process of signal transduction by activation of the
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
receptor-associated Janus family tyrosine kinases (JAKs) through cross
phosphorylation (Fig. 7).
Receptor
Ligand
\ /
. tr t\ 1 1 (i
< w "
j ^ \ . -7
}
n ii ii 'i
p J A K
S T A T '
[p v
' J \ Plasma
T V M ? m b / a n e
' I' II I f
v , i a t ^ # • *
'Sy#g |
STAT
S T A T Phosphorylation
v p
S T A T S T A T
Y iBHpr
STAT
Dimerization
Nucleus
DNA
P v
S T A T !sTAT
i v
Nuclear Translocation
DNA Binding
Gene Transcription
Figure 7. Model of signal transduction at the G-CSF receptor (Benekli et al. 2003).
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.3 The pharmacokinetics of G-CSF
Administration of pharmacological doses of G-CSF results in a triphasic response
(Watts et al. 1997): First, market neutropenia occurs approximately 5 to 60 min after
injection. Second, Leukocytosis due to release o f neutorphils, eosinopils and
monocytes from the bone marrow beginning after 3 to 4 hours. Third, proliferation
and differentiation of hematopoietic precursors in the bone marrow.
Absorption and clearance of filgrastim follows first-order pharmacokinetic modeling
without apparent concentration dependence. Continuous IV infusion of 20 mcg/kg of
filgrastim over 24 hours results in mean and median serum concentrations of
approximately 48 and 56 ng/mL, respectively. Subcutaneous administration of 3.45
mcg/kg and 11.5 mcg/kg results in maximum serum concentrations of 4 and 49
ng/mL, respectively, within 2 to 8 hours. The elimination half-life, in both normal
subjects and cancer patients, is approximately 3.5 hours. Clearance rate of filgrastim
is approximately 0.5 to 0.7 mL/minute/kg. Single parenteral doses or daily IV doses,
over a 14-day period, resulted in comparable half-lives. The elimination half-lives
are similar for IV administration (231 minutes, following doses of 34.5 mcg/kg) and
for SC administration (210 minutes, following doses of 3.45 mcg/kg) (Watts et al.
1997).
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.4 The modified recombinant G-CSF.
The recombinant G-CSF fused to human albumin (Albugranin) was reported to have
prolonged myelopoietic effects in mice and monkeys (Halpern et al. 2002).
Albugranin had a longer terminal half-life (T m) and mean residence time (MRT), and
slower clearance (CL/F) in mice. The T 1/2, MRT, and CL/F of Albugranin following
SC administration to BDF1 mice were 5.6-5.7 h, 16.7-20.7 h, and 6.37-12.2 mL/h/kg,
respectively, compared with 2.54 h, 4.9 h and 164mL/h/kg of G-CSF. Pegylated G-
CSF has been reported to have longer half life (Watts et al. 1997). The results of a
randomized, double-blind, phase 3 trial showed that the single dose of pegylated
filgrastim (SD/01) is as effective as daily Filgrastim for hematologic support of
chemotherapy in breast cancer patients. Even though the modified versions of G-CSF
exhibited better pharmacokinetic profiles, they still need to be injected.
4. The advantages o f recombinant fusion protein over chemical
conjugates.
TfR-mediated transcytosis in oral delivery of insulin-transferrin (In-Tf) has been
demonstrated in cultured epithelial cells and in animal models. (Xia, 2000). The G-
CSF-Tf chemical conjugates have also been proved to be bioavailable via oral
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
administration(W idera et al. 2004). However, there are several drawbacks of the
conjugation products. First, using current technique, the product is not structurally
defined. It is a heterogeneous mixture of variable molecular entities. Second,
chemical conjugation is not suitable for large scale production in industry. Third, it
will be costly for the loss of product through the purification process and the
inefficiency of the conjugation method itself.
4.1 The expression of the recombinant fusion protein
The recombinant fusion protein can be expressed in Escherichia coli (E. coli),
mammalian cells, insect cells (baculovirus) and yeast. The majority of therapeutic
proteins have been produced in either mammalian cell-culture system or in Ecoli.
Escherichia coli offer a means for rapid and economical production of recombinant
proteins. Until the mid-90s, E. coli was the predominant host for the production of
protein pharmaceuticals. However, in recent years, it has been taken over by
mammalian cell production (Chu & Robinson 2001). The reason is that E. coli
doesn’t always fold products into their proper conformation. In addition, E. coli
cannot prepare or attach mammalian glycosylation chains. The glycosylation is
important for the Tf bioactivity. Transient eukaryotic systems, in which the isolation
of stable transfectants is bypassed so that protein expression is obtained rapidly but
only for a limited period of time, have also been shown capable of producing
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
reasonable protein levels (10 to 20 mg/L) in reduced time (Meissner et al. 2001). In
my work, the T f fusion proteins will be expressed in the eukaryotic expression system.
4.2 The protein free HEK293 mammalian expression
The suspension culture of HEK293 cells in protein - free system was developed by
invitrogen (GIBCO). HEK 293 cells were originally derived from human embryonic
kidney and subsequently demonstrated to be a useful cell type to produce adenovirus,
other vial vectors, and effectively glycosylated human recombinant proteins. CD 293
Medium is a protein-free, chemically defined medium optimized for the growth of
suspension cultures of HEK293 cells. CD 293 Medium contains no proteins or
peptides of animal, plant, or synthetic origin. There are no undefined lysates or
hydrolysates in the formulation either. The lack of serum in the cell culture medium
can greatly simplify purification processes and reduce regulatory concerns by
improving purification and final product quality.
5. The linkages between moieties in the fusion protein
The construction of a fusion protein involves the linking of two proteins or domains of
proteins by a peptide linker. The selection of linkers is particularly important for
expressing the bi-functional fusion proteins. Most studies showed that the flexibility
21
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and hydrophilicity of the linker were important not to disturb the functions of the
domains (Crasto & Feng 2000; Robinson & Sauer 1998; Stanislawska-Sachadyn et al.
2006).
The flexible linker (GGGGS)3, which was designed by Huston et al, is frequently
used in single-chain antibody, sFv, and other recombinant fusion proteins (Huston et
al. 1988). However, some studies showed the fusion protein with the flexible linker
did not retain the bioactivity of the domain (Maeda et al. 1997). The helix linker
A(EAAAK)nA (n= 2-5) was reported to form about 80% helicity. And the helix linker
could control the distance and reduce the interference between the domains (Arai et al.
2001).
In this project, we first design a fusion protein with a very short linker. Based on the
known structures of the two proteins, there will be no steric hindrance between the two
moieties even if no linker is inserted. Later, we investigated the effect of inserting
flexible and helix linkers in the G-CSF-Tf fusion protein, to further improve the
biological activity , and thus the efficacy of the products.
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Materials and methods
1. Cell culture
1.1. HEK293 cell culture
HEK 293 cells, originally derived from human embryonic kidney, were abtained from
American Type Culture Collection, ATCC, Rockville, MD. They were cultured as a
cell monolayers in Dubacco’s Minimum Essential Medium (DMEM) or Minimum
Essential Medium (MEM), supplemented with 100 units/ml penicillin G sodium, 100
pg/ml streptomycin, 4 mM L-glutamine, and 10% fetal bovine serum (Sigma).
Cells were split every 2 - 4 days when they reach 80 - 90% confluency. To split the
cells, the medium was removed first, then the cells were washed twice with sterile
PBS (containing no Ca2 + or Mg2 + ). Cells were detached by adding 1 - 2 ml of trypsin -
EDTA solution and treat for 1-3 min, Then 5-10 ml of complete growth medium (to
stop trypsinization) was added the cells was resuspended gently but thoroughly. The
desired number of cells were transferred to a T-75 tissue flask containing 10 ml of
medium. Gently rock the plate to distribute cells.
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Before transfection, the HEK293 cells were seeded in 6-wells plates. And after
transfection, the cells were cultured in CD293 protein free medium (GIBCO). CD293
is a chemical defined protein free medium, without any animal-origin components,
undefined lysates, or hydrolysates. It was reported used in HEK293 cell suspension
culture. However, it was found in this study that CD293 mediuum is able to keep
HEK293 cells in adhesive culture.
1.2 Caco-2 cell culture
The Caco-2 cell line, obtained from ATCC, was cultured in DMEM supplemented
with 10% fetal bovine serum (Sigma). Newly confluent Caco-2 cells growing on T-25
tissue culture flask were rinsed twice by 5 ml PBS ( pH 7.4). Then 0.25 ml of Trypsin
- EDTA (0.5% Trypsin, 5.3 ml EDTA from Gibco BRL) was added to the cells and
the flask were incubated at 37 °C for 10 min. After the incubation, cells were
detached from the flask and 5 ml 10% Fetal Bovine Serum (FBS) DMEM with 0.1
mM nonessential amino acids, 2 mM L-glutamine, 50 U/ml penicillin-G, and 50 pg/ml
streptomycin, was added. The cell suspension was pipetted up and down with 10 ml
pipeter for 10 times to break down the cells, and the cell density was counted using
the Coulter counter (Coulter Electronics) or cell hemocytometer. The cells were
incubated in an atmosphere of 5% CO2 in air and 90 % humidity at 37 °C.
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Culture medium was changed every 3 days. For surface binding experiment, cells
were seeded in 6-well plate (Costar). Cells was seeded at density of about 4.2x104
cells/cm2. Caco-2 cells were incubated in an atmosphere of 5% CO2 in air and 90%
humidity at 37 °C. Confluent cell monolayers were obtained in 7 to 8 days.
1.3 NFS60 cell culture
Murine myeloblastic NFS60 cells, kindly provided by Dr. James Ihle (St. Jude
Children's Research Hospital, Memphis, TN), were grown in RPMI medium 1640
with 10% (vol/vol) FBS and 10% (vol/vol) WEHI-3 cells conditioned growth media.
NFS60 cells grow in suspension. Cells were passed every three days. The seeding
dilution is from 1:4 to 1:8.
1.4 Human bladder carcinoma 5637 cell culture
The human bladder carcinoma cell line 5637, obtained from ATCC (NO. HTB-9), was
grown in RPMI 1640 medium with 2 mM L-glutamin, 1.5 g/L sodium bicarbonate, 4.5
g/L glucose, 10 mM HEPES, 1.0 mM sodium pyruvate, and supplemented with 10%
FBS. The cells were subcultured twice or three times per week. For subculturing, the
cell monolayer was washed with PBS twice, and then treated w ithlml of trypsin. The
25
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
flask was allowed to sit at 37 °C until the cells detached. The fresh culture medium
was added to stop the trypsinization, and the cells were resuspended in fresh medium.
1.5 The BHK cell culture
BHK-21 (Baby Hamster Kidney Cells) adherent cell line was grown in the medium of
MEM (Eagles with non-essential AA’s, Sigma) plus 10% FBS. The medium was
changed every other day. The cells were passed after reaching 90% confluence. The
seeding ratio is from 1:4 to 1: 8.
1.6 Pretreatment of cells
In the TfR. binding study, the Caco-2 cells need to be pretreated to deplete endogenous
Tf. The confluent cell monolayers were rinsed with serum free -medium twice, then
incubated in serum -free medium containing 1 mg/ml bovine serum albumin (BSA)
(Sigma) at 37 °C for 1 hour. The medium was replaced before the experiment.
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2. Plasmid construction
2.1 Plasmid isolation
Single colonies from freshly streaked selective agar plate were picked and inoculated
in 5 ml (or 30 ml for Midi) LB buffer with antibiotics for 12-16 hours at 37 °C with
vigorous shaking (225 rpm). If the bacteria were taken from -8 0 °C, the bacteria
needed one hour for recovery. Pick some frozen bacterial from the vial using
autoclaved toothpick, culture in prewarmed LB buffer. After one hour the antibiotics
was added to the cultures.
The purification procedure is according the QIAGEN plasmid purification handbook.
(Mini and Midi). The QIAGEN plasmid purification protocol is based on a modified
alkaline lysis procedure, followed by binding of plasmid DNA to QIAGEN anion-
exchange resin under low salt and pH conditions. RNA, proteins, and other impurities
are removed by a medium-salt wash. Plasmid DNA is eluted in a high-salt buffer and
then concentrated and desalted by isopropanol precipitation.
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.2 Total RNA isolation by TRIZOL reagent (GIBCO-BRL)
The 5637 bladder carcinoma cells were seeded in 100 mm plate. After the cell
reached confluence, the total number of cells was about 13.7xl06. The procedure is
according to the vendor ‘s instruction (Invitrogen).
RNA (about 1 pg/ml) 2 pi
DNAase buffer (x 10) 2 pi
DNAase 2 pi
DEPC H 20 14 pi
Total 20 pi
First, after the cell monolayer was washed with 4°C PBS, 8 ml Trizol reagent was
loaded to the cells directly. Going through a pipette for several times was needed for
cell total homogenization. The homogenized samples were allowed to sit at 15 to 30
°C for 5 minutes to permit the complete dissociation of nucleoprotein complex. Then
1.6 ml chloroform was added to the tube. The tubes were shaken vigorously by hand
for 15 seconds and incubated at 30°C for 3 minutes. The two phases were separated
after centrifugation (9000 rpm for 20 minutes). The RNA, which remains in the
colorless aqueous phase, was transferred to a fresh tube, and was precipitated by
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
adding 0.5 ml isopropyl alcohol to 1 ml of TRIZOL reagent. The pellets were further
washed with 75% ethanol and air-dried.
2.3 RT-PCR
RT-PCR was performed with the following procedure:
(1) DNA free : DNAase was added to remove the DNA contamination from the total
RNA. The mixture was incubated for 30 min at 37 °C.
(2) Inactivation: 2 pi inactivation reagent was added to the DNA freed RNA and
mixed well. The mixture was incubated at room temperature for 2 minutes. Then the
mixture was spun down.
(3) Reverse transcription: the following integrants was added
DNAase treated RNA lOpl
50 pM oligo dT lpl
The mixture was incubated at 65 °C for 5 minutes.
After incubation put on ice, the following integrants were added to the mixture.
29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5xbuffer 4 pi
0.1 M DTT lp l
lOmMdNTPs 2pl
RNAase out lp l
Thermoscript lpl
Total 20 pi
Samples were incubate at 55 °C for 50 minutes, then were incubated at 85 °C for 5
minutes.
lpl RNAase H was added , and the mixture was incubated at 37 °C for 20 minutes.
(4) PCR
The following integrants were added to make mixture:
lOxamplification buffer 10 pi
10 mM dNTP mixture 3 pi
50 mM MgS04 2pl
5 ’ primer (20 pmol/pl) 1.5pl
3’ primer (20 pmol/pl) 1.5pl
Template cDNA 3 pi
Autoclaved dH20 78pl
Total 100 pi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Gently vortex the sample and briefly centrifuge to collect all drops from walls of tube.
Overlay the sample with half volume of mineral oil or add an appropriate amount of
wax. This step may be omitted if the thermal-cycler is equipped with a heated lid.
Place samples in a thermo-cycler and start PCR.
* Estimation of the melting and annealing temperatures of primer:
Tm= 4 (G + C) + 2 (A + T)
G, C, A, T - number of respective nucleotides in the primer.
Annealing temperature should be approx. 5°C lower than the melting temperature.
In the initial denaturation step, the DNA template was completely denatured in 95°C
for 5 min. In every cycle, the denaturation step is 0.5 min at 95°C. The primer
annealing temperature is 5°C lower than the melting temperature of primer-template
DNA duplex, and for 0.5 min. In the extending step, Pfx DNA polymerase works at
68°C. The rate of DNA synthesis by Pfx DNA Polymerase is highest at this
temperature. Recommended extending time is 1 min for the synthesis of PCR
fragments of up to 1 kb. When larger DNA fragments are amplified, the extending
time is usually increased by 1 min for each 1000 bp. For RT-PCR in this study, 40
cycles were performed due to the low number of the DNA template. In the final
extending step, the sample was treated at 68 °C for another 7 min to fill-in the
protruding ends of newly synthesized PCR products.
31
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.4 Gene clean
The Geneclean II kit (BIO 101, Inc) was used in the purification of PCR product and
the fragment after endonuclease digestion. The PCR product or the plasmid after
endonuclease digestion were loaded to the agarose gel. To minimize damage to DNA
sample, we used long-wave UV for short period. The DNA band was cut off under
UV. Three volumes of Nal stock solution was added to the cut agarose piece, and the
gel was dissolved by incubation at 45 °C to 55 °C. Then 5pi GLASSMILK was added
to the stock solution. The mixture was incubated on ice for 15 min to allow the DNA
binding to the GLASSMILK. The silica matrix with the band DNA was pelleted by
spinning for 5 seconds. When the centrifuge reaches 10,000 rpm, count to 5 and turn
off and remove the Nal solution. The pellet was washed with new solution for three
times. And the pellet was dried under vaccum for 10 minutes. The DNA was eluted
from GLASSMILK by adding 10 pi TE buffer and incubating at 45°C to 55°C for 2 to
3 minutes. The tube was centrifuged for about 30 s to make a solid pellet.
32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
M 13R ovoise P r i ^ o r ^ l AT° S a c i a ^ l Sf>o I
CAG GAA A C A G C T ATG A c t: ATG A T T ACG CCA AGC TTC- GTA GCG AGC TCG GAT CCA CTA
GTC CTT TG T CGA TAC TG P TAG TAA TGC GGT TCG AAC CAT GGC TCG AGC CTA C-GT GAT
(I
Bf
GTA ACG GCC GCC A GT GTG C
CA T TGC CGG CGG TCA CAC GAC
frcoR I
CTG GAA T TC GCC
AAG CGG
Edj-RI
GGC GAA T T C TGC
CCG C T T AAG ACG
EcoR V BstX. I Not I Xho I M s I Xba ( Apa I
I I I I II I
AGA TAT CCA TCA CAC TGC- CGG CCG CTC GAG CAT GCA TCT AGA GGG CCC AAT TCG
T C T ATA GGT ACT GTG ACC GCC GGC GAG CTC GTA CGT AGA T C T
CCC TAT
GGG TTA AGC GGG ATA
M13 Foiward (-20) Primer
;TC GTT TTA CAA CG T CGT GA:
:AG CAA A A T CTT GCA GCA Z V
TGG GAA AAC
ACC C T T TTG
: AAT TCA
“ TTA AGT
TA T TA
A TA AT
A GT GAG TC:
TCA CTC AG;
pCR®2.1-TOPO®
3.9 kb
Figure 8. The Map of TOPO - TA cloning vector from Invitrogen (Invitrogen manual).
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.5 PCR product cloning
TOPO® Technology (Invitrogen) presents a faster, more efficient way to clone PCR
products. The TOPO® vectors (as shown in figure 8) are provided linearized with 3’ -T
overhangs and topoisomerase I-activated to readily accept PCR products with 3 ‘-A
overhangs. The TA cloning technology makes it possible to easily clone PCR
products produced by Taq polymerase. Taq has a teminal transferase activity that adds
a single 3’-A overhang to each end of the PCR product. TOPO TA Cloning vector
contains 3’-T overhangs that enable the direct linagation of Taq-amplified PCR
products.
This kit is designed for the PCR products with Taq DNA polymerase. However, the
pfx DNA polymerase, which has the 3’ proof reading function, cleaves the 3’-A.
Therefore, in this study, one step to add 3’-A was needed. The PCR pfx mixture was
inactivated in 65 °C for 10 minutes. Then 1 U Taq DNA polymerase was added to the
PCR mixture and incubated at 72 °C for 10 min. The treated PCR products were
ready to be cloned by TOPO-TA cloning kit.
For cloning the PCR product, the PCR product and vector were mixed at ratio 50:1
and ligated by DNA ligase at room temperature for 5 min. And the ligation mixture
could be transformed to bacteria.
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
vector
Topolsomerase I
Topol somerase I
+
Taq-amplified
PCR product
vector
Add 1/d of a PCR-TOPO vector to 1 /d of Taq-
amplified PCR product
1
vector
Taq-amplified
PCR product
Incubate for 5 minutes
Topolsomerase I
vector
Topolsomerase I
Figure 9. The mechanism of Topo-TA cloning technology.
2.6 Transformation
The DH5-a E. coli bacteria from Invitrogen was thawn on ice and was aliquoted each
50 pi cells in pre-chilled 1.5 ml tubes. DNA (l-5pl) was added to the bacteria, and
was mixed by gently typing. The mixture was incubated on ice for 20 min, then was
heated shock at 37 °C for exact 40 sec, then 450 pi pre-warmed LB medium was
added. The bacteria was rescued at 37 °C, 225rpm shaking incubator for 1 hour, then
was spread 10% and 90% of the bacteria on two different LB antibiotic agar plates.
The plates were incubate in 37 °C over night.
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.7 The mutation of Tf Ser288 to Tf Cys288
6pg DNA was dissolve in 20 id dH20. 2 pi 2M NaOH/2mM EDTA was added to the
DNA. After the mixture was incubated at room temperature for 5 minutes, it was
chilled on ice. 7 pi ice-cold water, 5.5 pi 3M NaOAC (pH 4.7), and 75 pi ice-cold
ethanol was added in order quickly, and mixed well, theb keeped on ice. 1 pi mixture
was taken to check the pH value. If the pH is over 7.0, another 1 pi 3M NaOAC was
add to the mixture. The mixture was kept on dry ice for 15 minutes, then was Spined
down for 15 minutes at 4 °C. the precipitate was washed with 70% ethanol, then was
vacuum dried. The DNA was dissolved in 100 pi TE buffer (pH 8.0). The denatured
DNA can be stored in -20 °C. 10 pi of the denatured template DNA was mixed with 1
pi bottom selection oligonucleaotide and 1.25 pi mutagenic primer and 2 pi 1 Ox
annealing buffer in dH2 0 to make the total volume 20 pi. the mixture was heated to
75 °C for 5 minutes. And allow them to cool down to 37 °C.
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3. Protein expression
3.1 Transfection (Lipofectamine 2000)
One day before transfection, cells were seeded in the appropriate amount of growth
medium in 6-well plates without antibiotics such that they will be 80-90% confluent at
the time of transfection. For each well, 2 pg plasmid was diluted to 250 pi serum-free
opti-MEM medium. Lipofectamine™ 2000 was gently mixed before use. 5.5 pi
Lipofectamine was diluted to 250 pi Opti-MEM I Medium without serum. The
mixture was incubated for 5 minutes at room temperature. After the 5 minute
incubation, the diluted DNA was combined with the diluted Lipofectamine™ 2000.
The mixture was mixed gently and incubated for 20 minutes at room temperature to
allow complex formation to occur. The solution may appear cloudy, but this will not
impede the transfection, the mixture was added to cell dishes drop by drop. The
plates were incubated in 37 °C for 5 to 5.5 hours. CD293 medium was added to
replace the transfection mixture for protein expression.
37
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.2 CD 293 medium
CD293 is a chemically defined, protein and serum-free medium, which has been
formulated without human-origin components. CD293 has been developed for the
suspension adaptation and high-density culture of HEK293 cells. We used it for the
attached HEK293 cell. 5 hours after transfection, the CD293 protein free medium was
remounted. And the transfected cells were cultured in CD293 medium for 5 days,
before the conditioned medium was collected.
3.3 The G-CSF protein expression in DH5-a E. coli
Bacterial cultures were grown with shaking in 2 liter flasks containing 500 ml of
2xYT (16g of tryptone, 10 g of yeast extract, and 5 g of NaCl per liter) at 37 °C until
the absorbance was 0.6 at 600 nm. IPTG was added to induce the protein expression
at the concentration from 0.1 to 1 mM. The cultures were incubated an additional 3 to
3.5 hours. The bacteria were harvested by centrifugation at 5000g for 20 minutes in
Beckman JA10 rotor. The bacteria were re-suspended in ice cold PBS, and lysed by
Sonifier cell disruptor. The broken cell suspension was centrifuged at 10,000 g for 30
minutes. The supernatant was discarded and the pellet drained. The pellet was
washed with ice cold PBS once.
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.4 The purification of inclusion body
The pellet was re-suspended in 2x10 ml RA buffer (100 mM Tris-HCl, pH 8.0, 5mM
EDTA). lysozyme, DNAase and RNAase, was added to the solution. The solution
was frozen and thawn twice by using dry ice/ ethanol bath. The mixture was sonicated
until partially clear. The pellets was spin down at 35,000 g for 30 minutes. The pellet
was re-suspended in 2x12.5 ml R3 buffer (100 mM Tris-HCl, pH 8.0, 1% Triton X-
100). The debris was spin down at 35,000g for 30 minutes. The pellet was re
suspended in 2x12.5 ml R4 (1M NaCl) with sonication, then spin down at 35,000g for
20 minutes. The pellet can be kept at -20 °C. The pellet was re-suspended in 2x3 ml
of buffer R5A (8M GdnCl, 50 mM Tris-HCl, pH 8.0) or buffer R5 (same as R5A with
5mM DTT) with sonication. It was incubated at room temperature for 1 hour. The
pellet was spin down at 35,000g for 30 minutes. The supernatant was diluted in 34 ml
of buffer R7 (6 M urea, 0.1 M glycine, 50 mM Tris-HCl, pH 8.0) or buffer R6 (same
as R7 with 5mM DTT). The sample was dialyzed against 1L buffer RA with 1.3 M
urea and 5mM DTT, then dialysed against 1L buffer RA with 0.1 mM GSH and 0.1
mM GSSG. At last, the protein was dialyzed against 2 liter PBS (pH 7.0) twice.
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.5 The purification of GST-G-CSF fusion protein using GST beads from
inclusion body
The procedure for GST-G-CSF purification as follows:
(1) Appropriate amount of the 50 % slurry of glutathione-agarose beads was added to
the dialyzed supernatant.
(2) The protein beads mixture was rotated at 4 °C over night.
(3) The beads were washed with cold PBS for 3 times. Using 10 times PBS of the
volume of the pellet.
(4) Digest with thrombin. 50 units of the appropriate thrombin was used in 1 ml PBS
for each ml of resin volume. The solution was mixed by inverting the tube several
times and the mixture was incubated with shaking for 2 to 16 hours at room
temperature.
(5) The protein was eluted and e GST beads were regenerated. The buffer (50 mM
Tris-HCl pH 8.0, 150 mM NaCl, with 10 mM reduced glutathione) was added to the
beads, and rotated at 4 °C for 20 minutes.
3.6 Ammonium sulfate precipitation
The conditioned medium was collected after a 5 -dayculture. The protein was isolated
by 50% ammonium-sulfate precipitation. First, the precipitate was centrifuged at 900
40
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
rpm for 15 min to remove the debris. Prechilled conditioned medium was put on ice,
and solid ammonium sulfate was added little by little with stirring, until reaching 50%
saturation. The mixture was centrifuged at 8000 rpm for 20 min. After washing with
5M NaCl twice, the precipitated protein was dissolved in PBS.
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4. Fusion protein analysis
4.1 SDS-PAGE analysis
SDS-PAGE was performed accoding to the method described by Laemmli (Laemmli
1970). Briefly, the protein sample was mixed with 6x loading buffer and boiled for 10
minutes. 4% stacking gel and 10% and 7.5% separating gel were generally used. The
electrophoresis was started at 15 mA (current constant) for stacking gel, and after the
protein entered the separating gel, the current was changed to 20 mA. After the dye
band arrived at the bottom of the gel, the electrophoresis was stopped. The
polyacrylamide gel was stained by Coomassie blue (staining for 15 h, destaining I for
1 hour and destaining II until no blue background). And the molecular weight was
compared with the protein standard.
4.2 Western-blotting analysis
The protein samples were separated on 10% polyacrylamide gel. Three blotting
papers were cut as the same size as the gel and presoaked in towbin buffer. One Cut
PVDF membrane (Millipore Immobion-P) was first activated by soaking in methanol
for 1 minute and then was washed with dEbO. The "sandwich" was assembled with
the presoaked sponge, blotting paper, membrane. The transfer was performed for 2
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
hours at 200 mA. The pre-stained marker was labeled first, and the membrane was
blotted with 5% milk in TBS-T buffer (25 ml 1M Tris, 25 ml 5N NaCl, 1ml Tween-20
with 950 ml dTBO) for one hour at room temperature. The membrane was blotted
with primary antibody in 5% milk at 4 °C over night. After washing with TBS-T for
four times, the membrane was blotted with the horseradish peroxidase - conjugated
secondary antibody. The peroxidase activity was detected by enhanced
chemiluminescence (ECL, Amersham).
4.3 The development of stable transfected cell lines
First, the G418 selective concentration should be tested. Both BHK and HEK293 cells
were seeded into 24-well plate. The different concentrations of G418 sulfate (GIBCO)
from 200 to 1000 pg/ml were added to the normal culture medium. The cell death
was observed after one week. The concentration at 400 pg/ml was found to kill 50%
cells. The cells were transfected with the fusion plasmids by Lipofecctamine 2000.
Then the transfected cells were selected in 400 pg/ml G418 for several weeks.
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5. The linker formation
5.1 The primer design of the linkers
Table 1 shows the design of the primers. Note that longer helix linkers were obtained
by multiple inserts of the H4 sequence. Thus the primer sets were the same. The
primers were all 5’ end phosphate modified.
5.2 Primer annealing
The double strand DNA o f the linker was formed by annealing of the
oligonucleaotides. The procedure is as follows:
(a) Add the following components to a sterile microfuge tube:
Forward primer 2pl (10 pmole/pl)
Reverse primer 2pi (lOpmole/pl)
Annealing buffer (lOx) 2 pi
Deionized water 14pl
20 pi (1 pmole/pl duplex)
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(b) The mixture was incubated at 95 °C for 10 min
(c) The tube gradually cool to room temperature.
Table 1. Design of primers for the linker.
Flexible linker:
(GGGGS)3
Primer
(upstream)
BP=51
5’— TCGAG GGT GGA GGT GGC
TCA GGC GGT GGT GGA TCG GGT
GGC GGT GGA TCT C— 3’
Primer
(downstream)
BP=51
5’— TCGAG AGA TCC ACC GCC
ACC CGA TCC ACC ACC GCC TGA
GCC ACC TCC ACC C—3’
Helix linker:
A(EAAAK)2 A
Primer
(upstream)
Bp=42
5’— TCGAG GCT GAA GCT GCA
GCC AAA GAG GCC GCA GCT AAG
GCT C—3’
Primer
(downstream)
Bp=42
5’— TCGAG AGC CTT AGC TGC
GGC CTC TTT GGC TGC AGC TTC
AGC C—3’
Helix linker:
A(EAAAK)3 A
Primer
(upstream)
Bp=57
5’— TCGAG GCT GAA GCT GCA
GCC AAA GAG GCC GCA GCT AAG
GAA GCC GCA GCA AAA GCT C—3’
Primer
(downstream)
Bp=57
5’— TCGAG AGC TTT TGC TGC GGC
TTC CTT AGC TGC GGC CTC TTT
GGC TGC AGC TTC AGC C—3’
Helix linker:
A(EAAAK)4 A
Primer
(upstream)
Bp=72
5’— TCGAG GCT GAA GCT GCA
GCC AAA GAA GCT GCA GCC AAA
GAG GCC GCA GCT AAG GAA GCC
GCA GCA AAA GCT C—3’
Primer
(downstream)
Bp=72
5’— TCGAG AGC TTT TGC TGC GGC
TTC CTT AGC TGC GGC CTC TTT
GGC TGC AGC TTC TTT GGC TGC
AGC TTC AGC C
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.3 The linearization fusion protein plasmid by digestion withXhol
The plasmid was digested by Xho 1 endonuclease
as follows:
The G-CSF-Tf plasmid
Xho 1 (10 U/pl)
xlO buffer
dH20
50 pi
- 5 pi on gel
45 pi
The reaction mixture was incubated at 37 °C for 1 hour and 10 min.
The mixture was incubate at 65 °C for 15 min to inactivate Xho 1.
Dephosphate:
The linearized plasmid needs to be dephosphated to prevent it from forming a circle.
The procedure is as follows:
Product from previous step 45 pi
dH20 5 pi
46
to be linearized. The procedure is
2.5pi (about 5pg)
2.5pl
5 pi
40 pi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
phosphatase (CIP)
51 pi
- 5 pi on gel
46 pi
TE 314pl
5N NaCl 40 pi
400 pi
5.4 Phenol -chloroform extraction
10% 5N NaCl was added to the mixture. Equal volume of phenol/chloroform was
added also. The sample was mixed by hand or vortex until it forms homogenous
mixture. It was Spin at room temperature, 10,000 rpm for 2 min. aqueous phase (top
layer) was transferred to a new tube. The interface should not be disturbed. 2 to 2.5
times volume of absolute EtOH at -20 °C was added , and mixed well. The tube was
kept in -80 °C for 15 min, then the supernatant was removed. 0.5 ml 70% EtOH (at -
20 °C) was added for extraction. The sample was centrifuged at 4 °C, 10,000 rpm for
2 min. The supernatant was removed completely. The pellet was dried by speed-
vacuum for 7 to 10 min.
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6. The studies o f bioactivity o f fusion protein
6.1 The in vitro activity test
6.1.1 Iodination of transferrin
Human Tf was radiolabeled with 1 2 5 I (ICN, Irvine, CA) using chloramines-T catalyzed
iodination, followed by purification using Sephadex G-50 column chromatography,
and subsequently dialyzed in phosphate buffered saline (PBS, pH 7.8). Soak and pack
the G50 Sephadex column. For iodination of lmg Tf, add 50 pi Iodide. Keep the tube
on ice, then add 50ul chloramines-T (8 mg/ml), and incubate for 1.5 min. Then add
50pl Na2S2C >5 (4.8 mg/ml) immediately. Incubate for another 1.5 min on ice for the
reduction. And add 100 pil KI (10 mg/ml) to stop the reaction. Load the sample to the
G50 column, and elute with PBS. Collect the elution 1ml per tube. Take 5pl liquid
from each tube to be counted by gamma counter. Collect the sample with highest
reading. Further dialyze the iodinated protein against PBS (pH 7.4) in 4 °C for 18
hours.
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6.1.2 Caco-2 cell surface binding study
Human T f was radiolabeled with 1 2 5 I (ICN, Irvine, CA) using chloramines-T catalyzed
iodination, followed by purification using Sephadex G-50 column chromatography,
and subsequently dialyzed in phosphate buffered saline (PBS, pH 7.8). Caco-2 cells
were seeded in 12-well cluster plates until fully differentiated. Caco-2 monolayers
were washed with cold PBS three times, and then incubated in serum-free D-MEM
supplemented with 0.1 mg/ml BSA at 37 °C for 30 min to remove the endogenous Tf.
A mixture of 3 pg/ml 1 2 5 I-Tf with 3-, 10- or 30-fold of unlabeled fusion protein or Tf
in D-MEM with 1 mg/ml BSA was added to different wells. After 30 min of
incubation at 4 °C, the medium was removed, and the cell monolayers were washed
with cold PBS three times. The cells were then dissolved in 1 M NaOH, and the
lysates counted in a gamma counter.
6.1.3 NFS-60 cell proliferation - MTT test
The G-CSF activity of the fusion protein was measured by NFS-60 cell proliferation
assay. NFS-60 cells were washed three times with RPMI-1640 with 10% FBS and
aliquoted into 96-well microtiter plates at a density of lxlO5 cells/ml. Subsequently
10 pil of 10-fold serial dilutions of the G-CSF and fusion proteins were added. The
plates were incubated at 37 °C in a 5% C 02 incubator for 48 h. A MTT (3-(4,5-
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide) assay was performed
essentially as described. Briefly, the cells were treated with 1 mg/ml MTT in serum-
free and phenol red-free RPMI 1640 media for 4 h. The formazan crystals were
dissolved in isopropanol and absorbance was measured at 570 nm on a TECAN
GENios Plus microplate reader.
6.2 The in vivo activity test
Male BDF1 mice (Charles River Laboratories, Wilmington, MA), 6-8 weeks of age,
were used in all animal experiments described in my dissertation. The mice were
allowed to acclimate for 5 days. Animal experiments were compliant with the
‘Principles of Laboratory Animal Care’ (NIH Publication # 85-23) and has been
approved by the Institutional Animal Care and Utilization Committee at the University
of Southern California.
Prior to dosing, the mice were fasted for 12 h. The treatment groups received a single
dose on day 0. Due to the difference in molecular weight, i.e., 20 KD for G-CSF and
100 KD for the fusion protein, animals received the dose based on equivalent pmoles.
For subcutaneous administration, 5 mg/kg (0.05 qmol/kg) of the fusion protein or 1
mg/kg (0.05 pmol/kg) of G-CSF was injected. For oral administration, 50 mg/kg
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(0.5pmol/kg) fusion protein or 10 mg/kg (0.5 mol/kg) G-CSF were given via a gavage
needle.
Blood samples were collected daily from the tail vein, diluted 40-times and lysed in an
acidic crystal-violet solution (0.1 % crystal violet, 1 % acetic acid, in water). The total
white blood neutrophils (PMN) among the leukocytes were determined manually by
using Wright-stained blood smear glass slides that were examined under an Olympus
BH-2 microscope. The absolute neutrophil count (ANC) was determined by
multiplying the total WBC count by the PMN percentage.
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7. Statistical analysis
The statistical significance of the differences between experimental groups was
determined by using the unpaired Student t test. Findings with two-tailed P < 0.05
were regarded as significant.
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Results
1. Conjugation o f recombinant T f with a cysteinyl residue
The heterogenicity of the chemically conjugated insulin-transferrin (In-Tf) and G-
CSF-Tf conjugates was largely due to the multiple SPDP (N-Succinimidyl 3-[2-
pyridyldithioj-propionamido) - modification sites on Tf, a glycoprotein with 54 Lys
residues. There are at least 3 to 5 lysyl residues on the surface of a T f molecule that
are equally accessible and sensitive for SPDP-modification(Singh & Curtiss 1994).
The multiple sites of modification produce not only different forms of conjugate
regarding the site of linkage and the number of drug molecules per Tf.
In order to obtain a modified Tf with a defined modification site, to prepare a mutant
recombinant T f containing only one single cysteinyl residue, T f (Cys), on the surface
of the protein molecule (Fig. 10), was attempted mutagenized Ser-288 to Cys-288 to
generate S288C-Tf. Ser-288 is next to Phe-289 and His-290 where others have
previously inserted a nanopiptide without losing the biological activity(Ali et al.
1999). This Ser-Cys mutation will have minimum effect on the formation of
hydrogen-bonding and steric arrangement on the peptide structure (Ali et al. 1999).
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
221
110
Figure 10. Molecular model of HST showing the location of S288.
1.1 Expression of Tf protein in mammalian cell
The TFR27A plasmid was purchased from ATCC in frozen bacterium. After
recovery, the bacterium was cultured over night at 37 °C, with shaking 225 rpm. The
plasmid was purified with Mini-prep plasmid purification kit (Invitrogen). We cloned
the Tf gene using the forward primer 5’-GC GTC GAC GTC CCT GAT AAA ACT
GTG -3’, and the reverse primer 5’- GC GGG CCC TTA AGG TCT ACG GAA AGT
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
GCA C -3’. The PCR product was analysed on 1% agarose. The result is shown in
figure 11.
The PCR product was ligated into TOPO-TA cloning vector (Invitrogen), then
transformed to DH5-a bacteria. 10% and 90% of the bacteria were spread respectively
on ampicilin selective agar plates, and Incubated at 37 °C overnight. There were 33
positive clones grown in the 10% plate, and over 100 positive clones in the 90 % plate.
We chose 12 positive clones and cultured them in LB overnight separately. The
plasmids were isolated with Mini-prep plasmid purification kit.
The plasmids were digested by endonuclease EcoR 1 and Pst 1, respectively, to test
the positive clones. The digested products were analysed on 1% agarose gel. And the
Tf band was cut out and purified by gel cleaning kit. The T f cDNA was ligated into
the mammalian expression vector pCMV-Tag2b vector, using the EcoR 1 and Pst 1
cutting sites. The map of the pCMV-tag vector is shown in figure 12. And the
plasmids were digested by EcoRl/BamHl or A pal/Pstl. The digestion products were
loaded on agarose gel. The result was shown on figure 13. There was a Tag peptide at
the N-terminal of Tf. The pCMV-tag-Tf plasmid were double digested by EcoR 1 and
BamHl (Fig. 13a). By calculation, there should be three bands, 4799 bp, 672 bp and
270 bp. Clones 7 and 8 had the expected digestion pattern. To further verify the
composition, the plasmids were double digested by Apa 1 and Pst 1. Clones 7 and 8
55
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
showed three bands, 4276bp, 1449 bp and 629 bp, which matched the expectation.
The sequence of the plasmids were then verified by DNA sequencing.
T f
— 2.0 kB
Figure 11. The subcloning of T f from TFR27A plasmid. Only one band about 2.0
kilo-base was detected.
We used Midi-prep kit (Invitrogen) to isolate the positive pCMV-Tag 2b-Tf plasmid.
The plasmid of pCMV-tag-Tf were transfected into BHK cells. After 48 hours, the
cells were lysed. And same amount of protein was loaded in each lane in the westem-
blot assay (Fig. 14). In cells that were not transfected with the pCMV-tag-Tf plasmid
(the control lane), no band could be detected. On the other hand, the cells transfected
with the plasmid expressed the protein, which could be recognized by both anti-Tf and
the anti- flag antibody.
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
pUC ori i
tkpaiB pCM
MCS
SV40 pA
4
n e o /k a n ] - ^ • ^ P bla
- p
ifl
bl
PSV40
ori
T3 p rom oter
AA TTA ACC CTC ACT AAA G G G AAC AAA AGC TGG
A rt
AGC TCC ACC GCG GTG GCG GCC GCC ACC ATG. . .
1_________________I
K O Z A K
_____________________ FLA G t o g ____________________________
Srf I
D Y K D D D D K 1 i
. . .G A T TAC A A G GAT GAC GAC G AT A AG * GCC CGG
Bam H I Pst I EcoR I
I I I
GCG GAT CCC CCG GGC TGC AGG AAT TC. . .
EcoR V Hind III Acc 1/5 a I I Xho I
I I I I
. . .G ATA TCA AGC TTA TCG ATA CCG TCG ACC TCG
A p a I
I
AGG GGG GGC CCG GTA C C T . . .
. . . TAATTAATTAAGGTACCAGGTAAGTGTAC CCAA
I ___________________I
MULTIPLE S T O P C O D O N S
T7 promoter
TTC GC CC TATAGTGAGTC GTATTA
Figure 12. The map of the pCMV-Tag vector.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(a)
1 2 3 4 5 6 7 8 9 10 1 1 12
(b)
Figure 13. Selection of positive pCMV-Tag -T f plasmid by endonuclease digestion,
(a) The pCMV-Tag -T f plasmids were double digested with EcoRl and BamHl. (b)
The pCMV-Tag -T f plasmids were double digested with Apa 1 and Pst 1.
58
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(b)
Figure 14. The BHK cell expressed Tf could be recognized by both anti- T f and anti-
Flag antibodies. The cells were transfected with pCMV-tag 2b -T f plasmid. After 48
hours, the cells were lysed with boiling 6x loading buffer. The protein assay was done
to measure the protein concentration. Same amount of protein was loaded in each
lane, (a) Recognization by the anti-Tf antibody, (b) Recognization by the anti-flag
antibody.
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.2 The mutation of Tf sequences
In order to produce a Tf recombinant with single sulfhydryl group from a cysteinyl
residue, as described in the beginning of this section, we first mutagenize Ser-288 to
Cys-288 to generate S288C-Tf.
842 875
Original nucleotide sequence: GAA TTC CAA CTA TTC AGC | TCT| CCT CAT GGG AAG
Corresponding amino acid sequence: E F Q L F S S P H J K
Nucleotide sequence after mutagenesis: GAA TTC CAA CTA TTC AGC | TGT| CCT CAT GGG AAG
Corresponding amino acid sequence: E F Q L F S C P H J K
Figure 15. Site-specific mutation of recombinant Tf.
The mutant primer: CTA TTC AGC TGT CCT CAT GGG was used. And the 5’ -end
was modified with phosphorylation. The oligonucleotide site direct mutagenesis kit
was used (Stratagene). The procedure was described in the methods part. The mutant
Tf sequence was verified by DNA sequencing. The result showed only one mutantion
in the whole Tf sequence as intended.
After the mutant T f plasmid was successfully made, we did not continue the study to
make the T f chemical conjugate. In stead, moved from the chemical conjugation
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
method to the development of the recombinant fusion proteins consisting Tf and G-
CSF. However, the plasmid with mutated Tf can be used in future research when
chemical conjugation is preferred over the recombinant DNA method.
61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2. The cloning and expression o f G-CSF
In this part of the research, we intended to clone and express the G-CSF protein, for
the following three reasons: (1) the cloned cDNA sequence can be used for subsequent
recombinant fusion protein construction; (2) the expressed G-CSF can be used as
control in the T f -mediated transport study; (3) the expressed G-CSF can be used for
future chemical conjugation research.
2.1 The cloning of G-CSF cDNA from Human Bladder Cacinoma cell
line 5637
Human bladder carcinoma 5637 cell line was purchased from ATCC. To confirm that
the cell line secrets G-CSF, we applied the NFS60 cells proliferation assay to the
conditioned medium. The results showed that the conditioned medium did contain the
factor that could stimulate NFS-60 proliferation.
The total RNA was isolated from 5637 cells as described. To confirm the quality of
the isolation, the total RNA was loaded onto agarose gel. As shown in Fig. 15, three
major bands were seen on the gel, the bands of tRNA,18s rRNA, and 28s rRNA.
There was a smear of mRNAs in the background. To remove the DNA contamination
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
from the total RNA, the product was treated with DNAase. The DNA free samples
were continued with the RT-PCR procedure, using the following primers.
28 S rRNA
18 S rRNA
tRNA
Figure 16. The isolation of total RNA from 5637 cells. Lane A: 1Kb DNA ladder.
Lane B: total RNA.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
For G-CSF without secretion signal:
Forward: 5’ -G C GAA TTC ACC CCC CTG GGC CCT -3’
Reverse: 5’ -G C CCC TCG AGT CAG GGC TGG GC-3’
For G-CSF with secretion signal:
Forward: 5’ -GC GAA TTC ATG GCT GGA CCT GCC-3’
Reverse: 5’ -G C CCC TCG AGT CAG GGC TGG GC-3’
The G-CSF gene obtained from RT-PCR, both with and without signal sequence were
shown in figure 17(a), and the purified RT-PCR product was shown on figure 17(b).
From the RT-PCR results, the major band had the correct size. For G-CSF, the band
was 537 bp, and for the G-CSF with secretion signal (sG-CSF) the major band was
about 621 bp. There were several other bands. After gel purification, there is only
one band each lane in the agarose gel.
The purified RT-PCR products were cloned by TOPO-TA cloning kit. And the
plasmids isolated from single clones were digested by endonuclease EcoRl and Xhol
together. The plasmids containing the right size insert were chosen, and the sequence
was further verified by DNA sequencing.
64
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A B C
A B C
Figure 17. The RT-PCR of the G-CSF gene with and without secretion signal, (a) The
RT-PCR product of G-CSF. Lane A: the DNA lkb ladder; Lane B: the RT-PCR of
G-CSF; Lane C: the RT-PCT of sG-CSF; (b) the purified RT-PCR products; Lane A:
the DNA lkb ladder; Lane B: the RT-PCR of G-CSF; Lane C: the RT-PCR of sG-
CSF.
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.2 Expression of G-CSF protein in BL21 bacteria
The G-CSF cDNA was inserted into pGEX-4T-l vector. The ligated plasmids were
digested by EcoRl and Xho 1 to test whether they contain the G-CSF insert. The
selected plasmid was transformed into BL21 bacteria. Single positive clone was
chosen using ampicilin resistance selection. The G-CSF-GST fusion protein was
expressed in BL21 E. coli. The molecular weight of GST, and G-CSF should be about
26kD and 18.7 kD, respectively. So the GST-G-CSF fusion protein should be about
45 kD. Fig. 18 shows the difference of the protein expression before and after IPTG
inducing, and the identification of the production of GST-G-CSF.
As shown in Fig 18 there was a band of about 45 kD which becames more intense
with longer induction, in both the supernatant and the pellet. The protein express
amount was increased with time in the first two hours. However, most of the protein
was expressed in the pellets. With large amount of expression and improper folding,
the protein formed inclusion body. Thus the next step was to purify the GST-GCSF
from the precipitated inclusion body.
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
N o
inducing - ^ m 'n ^Omin 90m in 120m in 30m in 60niin 90m in 120min
supcrnatcnt
Figure 18. The GST-G-CSF protein expression. The bacteria were sampled before
IPTG inducing and after inducing 30, 60, 90, 120 min. The bacteria were precipitated
down and broken by sonication. After centrifuge, the supematent and pellet were
loaded in SDS-PAGE. The gel was stained with Coomassie brilliant blue.
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.3 The purification of GST-G-CSF fusion protein with GST-affinity
beads from inclusion body
2.3.1 Purification of inclusion body
The pellets were washed with different buffers for several times to purify the inclusion
body. The purified inclusion body was dissolved in the 6 times SDS loading buffer.
From SDS-PAGE, we estimated that, the purity of the fusion protein is over 80% in
the inclusion body (Fig. 19). In addition, from figure 18, the amount of the GST-
GCSF in the inclusion body continued increasing from 30 minutes to 120 minutes.
2.3.2 The renaturing of GST-G-CSF from inclusion body and isolation of G-CSF by
thrombin-cutting
The proteins in the inclusion body were denatured and renatured. The dissolved
protein was allowed to bind to the GST beads. After the beads were washed with ice
cold PBS, the fusion protein was eluted out and cut by thrombin.
As shown in Fig 20 in lane B, the fusion protein band turned lighter and thinner after
binding to the GST beads, indicating that the GST-G-CSF was bound to the beads. In
lane C and D, no band was detected. Therefore the fusion protein could not be washed
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
off by PBS. In lanes E and F, the fusion protein band became thinner with digestion,
and two bands, about 18 kD and 25kD , were detected. Indicating the fusion protein
was digested by thrombin into two proteins G-CSF and GST. However, the cutting is
not very specific. Under the cutting condition, the GST was also washed off the
beads. Further purification from the mixture is needed to obtain pure G-CSF.
No inducing 1 2 3 4 5
50kD
37kD
25kD
Inclusion body
Figure 19. The GST-G-CSF fusion proteinwere expressed in the inclusion body. The
bacteria were sampled before IPTG inducing and after inducing 120 min. The bacteria
were precipitated down and broken by sonication. And the inclusion bodies were
purified and loaded in SDS-PAGE. The gel was stained with Coomassie blue. Lanes
1~5: different clones.
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A B C D E F
GST-G-
CSF
GST
G-CSF
Figure 20. The GST-G-CSF fusion protein thrombin cutting. A. The re-dissolved
inclusion body; B. The sample after GST beads affinity chromatography; C. The
elution with PBS (first wash); D. The elution with PBS (second wash); E. After
digestion with thrombin (20 U); F. After digestion with thrombin (40 U).
70
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3. Construction o f the G-CSF-Tf fusion protein plasmid
The preparation of fusion proteins consisting of Tf domain is the major goal of this
research project. Our intention is to produce a fusion protein containing both
functioning moieties of G-CSF and Tf. The configuration is important because both
moieties must be correctly folded to maintain their biological activity. However, in a
fusion protein, the two domains are folded together in the cell which may interfere the
folding with each other.
To simplify the downstream purification procedure, the secretion signal is added at the
N-terminal of the fusion protein. Three constructs of fusion proteins were designed as
shown in Fig. 21. The fusion proteins were created in two orientations, with the G-
CSF domain inserted at the N-terminus portion or at the C-terminus. In the two
constructs, the secretion signal of either the G-CSF or T f was retained, depending on
which moiety was at the N-terminus. An additional construct without secretion signal
was also tested in case the secretion signal in the other two constructs fails.
There is no Xho 1, EcoRl and Xba 1 cutting sites in G-CSF sequence. There are
EcoRl cutting sites in Tf. So the G-CSF need to be inserted first. Details of which
endonucleases have been used is in figure 21.
71
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
EcoR V Xhol Xbai
Figure 21. The maps of the fusion protein constructs. The small dark gray box at the
N-terminal domain represents the respective leader peptide signal sequences for G-
CSF and Tf. The solid bar represents the sequence of Tf. The light grey bar represents
the sequence of G-CSF.
3.1 The subcloning of G-CSF sequence
The primers used to subclone G-CSF with secretion signal were:
Upstream: GCG AAT TCA CCA TGG CTG GAC CTG CCA CCC AGA
Downstream: GCT CTA GAT CAG GGC TGG GCA AGG TG
72
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The primer used to subclone G-CSF without secretion signal were:
Upstream: GCC TCG AGA CCC CCC TGG GCC CTG C
Downstream: GCT CTA GAT CAG GGCTGG GCA AGG TG
The PCR products were shown in Fig. 22. The PCR products were amplified in
TOPO-cloning kit. The plasmids were digested to test whether they contained the right
insert. The sequences of the fusion proteins were then verified by DNA sequencing.
sG-CSF G-CSF
531 bp
624 bp
(a)
(b)
Figure 22. The PCR products of G-CSF with (sG-CSF) or without secretion signal (G-
CSF) were tested on agarose gel. (a) The PCR product for sG-CSF; (b) The PCR
product for G-CSF.
73
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.2 Subcloning the Tf sequence
The primers used for cloning the Tf sequence for fusion protein with T f at the N-
terminal (with secretion signal sequence) were:
The upper stream: 5’-GCG ATA TCA CCA TGA GGC TGC CCG TGG GA-3’
The down stream: 5’-GCC TCG AGA GGT CTA CGG AAA GTG CAG G-3’
The primers used for cloning the T f sequence for fusion protein with T f at the C-
terminal were:
The upper stream: 5’-GCC TCG AGA CCC CCC TGG GCC CTG C-3’
The down stream: 5’-GCT CTA GAT TAA GGT CTA CGG AAA GT-3’
The T f with or without secretion signal were subcloned from TfR27A plasmid
(ATCC). The different primers were designed for the different endonuclease cutting
sites. The PCR products were analysed on 0.8 % agarose gel. T f with Xho 1 and Xba
1 cutting site was going to be inserted in the C-terminus of the fusion protein. Tf with
EcoRV and Xho 1 cutting site, which also contains the secretion signal of Tf, was
going to be inserted in the N-terminus.
74
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Tf(Xho l,Xbal) Tf(EcoRV.Xho 1)
Tf(XhoI X b a l) Tf(EcoR V. Xho I )
2.0 kb
(b)
Figure 23. The PCR products of Tf were tested on agarose gel. (a) The raw PCR
products of T f gene with and without secretion signal, (b) The purified PCR products
of Tf genes with and without secretion signal.
75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
From the figure 23, we can see there was a by-product in the PCR product of the Tf
with Xho 1 and Xba 1 endonuclease cutting sites. The PCR products were purified by
gel-purification kit. After purification, only a single band was detected for both Tf
cDNA digestions.
3.3 The fusion protein expression plasmids construction
The PCR products of Tf and G-CSF were ligated into the linear TOPO TA cloning
vector for amplification (Invitrogen). The TOPO -T f and TOPO-G-CSF ligation
mixtures were transformed to DH5a bacteria and selected on ampicilin argarose
plates. After incubation at 37 0 C for over night, the positive clones were cultured over
night at 37 0 C with shaking at 225 rpm. The plasmids were isolated using Mini-prep
kit. The plasmids were digested by endonuleases, and the sequence of the insert was
verified by DNA sequencing. The cDNA was cut out from the TOPO plasmid, and
inserted into pcDNA3.0 mammalian expression vector. The map of the pcDNA3.0 is
shown in figure 24.
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
pcDNA3
5.4 kb
B sm l
Figure 24. The map of pcDNA3.0.
The pcDNA3.0-fusion protein plasmid was digested with the edonucleases, the
digested mixture were loaded on 1% agarose gel to verify that they contain the correct
inserts. Figure 25 shows the sG-CSF-Tf-pcDNA3.0 construct as an example. With
the other two constructs, similar results were obtained.
77
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
vector
— Tf
_ G-CSF
Figure 25. Verifying the insert size using endonuclease digestion. The sG-CSF-Tf-
pcDNA3.0 plasmids were isolated from different clones. And the plasmids were
digested by EcoRl, Xho 1 and Xba 1. Three bands appeared in the gel, with sizes of
623 bp, 2.0 kb and 5.0 kb respectively. As shown in this figure, there were six clones
containing both G-CSF and Tf inserts.
The clones which consisted of cDNA of both G-CSF and T f cDNA were amplified in
large amount, and isolated by Midi-prep.
78
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.4 The expression of the fusion proteins
The plasmids were purified in large scale by Midi-prep (Invitrogen) and used in the
transient transfection with BHK cells. After two days of culture, the fusion protein
expression was tested in both the cell lysate and the conditioned medium by Western-
blotting using anti-Tf monoclonal antibody. The Western-blotting results showed the
fusion proteins could be detected from both the cell lysates and the conditioned
medium (Fig. 26). Note that the medium contained transferrin, which caused a band
of 80 kD in the conditioned medium results.
We conclude alone from the result that the fusion protein with secretion signal could
be secreted into the medium. It is easier to purify protein in the medium compared to
purifying from the cell lysate. The expression amount of fusion protein is higher in 48
hours compared to 24 hours, and is higher by using 2 pg plasmid than using 1 pg
plasmid.
Because the fusion proteins with secretion signal were found in significant amount in
the conditioned medium, we continued our work with the two constructs with
secretion signal. The fusion protein with T f at the N-terminus expressed only in low
levels. Preliminary result using similar amount of unpurified fusion proteins showed
that the G-CSF biological activity of sTf-G-CSF was also much lower than that of sG-
79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CSF-Tf in NFS-60 proliferation assay (Fig. 27). In addition, in transfected HEK293
cells, more sG-CSF-Tf was expressed at much higher level than sTf-G-CSF (Fig. 28).
Therefore, we chose the fusion protein with G-CSF at the N-terminus (sG-CSF-Tf) as
the focus of our further study. As shown if figure 29, the 3-dimensional model of the
fusion protein, the G-CSF and T f moieties are well-separated and can presumably
retain their respective biological function.
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1 2 3 4 5 6 7 8 9
lO O kD
80 KD
lO O kD
Figure 26. The Western-blotting test for the conditioned medium and cell lysates of
transfected BHK cells. Upper panel: the protein expression in conditioned cell
medium. Lane 1: control; Lane 2: (24 hours)l pg sG-CSF-Tf vector transfection; Lane
3: (24 hours) 2pg sG-CSF-Tf vector transfection; Lane 4: (24 hours)lpg sTf-G-CSF
vector transfection; Lane 5: (24 hours) 2 pg sTf-G-CSF vector transfection; Lane 6:
(48 hours)l pg sG-CSF-Tf vector transfection; Lane 7: (48 hours) 2 pg sG-CSF-Tf
vector transfection; Lane 8: (48 hours)lpg sTf-G-CSF vector transfection; Lane 9: (48
hours) 2 pg sTf-G-CSF vector transfection. Lower panel: the fusion protein
expression in cell lysates. The samples were loaded in the same order as in the upper
panel.
81
1 2 3 4 5 6 7 8 9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
800000
700000 -
600000
- 500000 -
400000 -
8 300000
200000 - 1
100000
0
C ?
/
V
&
#
*
, Q >
$
S '
&
Figure 27. The comparison of biological activity of the fusion proteins, sG-CSF-Tf
and sTf-G-CSF, using NFS-60 proliferation assay.
82
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
sG-CSF-Tf 0.001 (jg 0.1 |jg
sTf-G-CSFj 0.01 gg
i
Tf standard
Figure 28. The expression of fusion protein sG-CSF-Tf and sTf -G-CSF in HEK293
cells. The HEK293 cells were seeded in 6-well plates. After reaching 90%
confluence, each well was transfected with 2 p,g plasmids. After 48 hours, the
conditioned medium was collect. Samples were loaded on SDS-PAGE, and
transferred to PDFV membrane. The proteins were blotted by anti-Tf antibody.
83
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 29. The 3-D structure model o f G-CSF-Tf.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.5 The expression of fusion protein in large scale and the fusion protein
purification
The stable-transfected HEK293 cell line was developed. The advantage of the
stabilized transfected cells us that it can continuously secret the fusion protein to the
medium. However, it was found that another protein at the molecular weight between
50 and 70 kD was also detected (Fig. 30), since the transient-transfected cells express
the fusion protein as the only major protein, the transiant-transfected cells were used
to express the fusion proteins to simplify the purification procedure. The fusion
protein was expressed in CD293 protein - free medium and a chemical defined
medium. Unless specified the conditioned medium was loaded on the gel without
further concentration. Figure 30 shows the SDS-PAGE analysis results.
Lanes A and B are the T f control at different concentration, the molecular weight of
which is about 80 kD. Lane C is the protein markers. Lanes D, E and F, are the sG-
CSF-Tf fusion protein from conditioned medium, or concentrated by lyophilisation
and ammonium sulfate precipitation. The protein precipitated by ammonium sulfate
retained more biological activity than the protein concentrated by lyophilisation.
85
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A : Transferrin (20 ug)
B: Transferrin (2 tig)
C : Marker
D: Stable transfection
E: Transient transfection
F: Concentrated by
lyophilisation
G : Concentrated by A m onium
Sulfate precipitation
Figure 30. The fusion protein expression was tested in SDS-PAGE.
After transfection, HEK293 cells were cultured in CD293 medium for 5 days, and the
fusion protein was detected by performing SDS-PAGE analysis of the collected
conditioned medium (Fig. 31a). One major band of approximately 90% abundance,
with a molecular mass of approximately 100 kDa, was visualized by Coomassie blue
staining after enrichment with 50% ammonium sulfate precipitation. The estimated
molecular mass of the protein is consistent with that of the fusion protein, i.e., 80 kDa
for Tf plus 19.6 kDa for G-CSF.
86
ABC D E F G
t
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a A B
2 5 0 k D —
1 5 0 k D — - k m
1 0 0 k D
PWf *
— 1 0 0 k D
7 5 k D —
— 8 0 k D
5 0 k D — « M
3 7 k D — m
b A B
— 1 0 0 k D
f P — 8 0 k D
Figure 31. Expression and identification of G-CSF-Tf fusion protein, (a) SDS/PAGE
of purified recombinant fusion protein. Lane A, fusion protein; lane B, Tf. (b)
Recognition of the recombinant fusion protein by anti-Tf antibodies using Western
blot. Lane A, fusion protein; lane B, Tf. (c) Recognition of the recombinant fusion
protein by anti-G-CSF antibodies using Western blot. Lane A, fusion protein; lane B,
G-CSF control.
87
C A B
1 0 0 k D
2 0 k D
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The identity of the secreted fusion protein was confirmed by using both anti-Tf and
anti-G-CSF antibodies in Western blot as illustrated in Fig. 31 b and c. Fig. 31b shows
that the fusion protein (lane A) was recognized by anti-Tf antibody. Fig. 31c shows
that the fusion protein (lane A) was also recognized by an anti-human G-CSF
monoclonal antibody.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4. The biological activity o f the G-CSF-Tf fusion protein
4.1 In vitro G-CSF and Tf activity of the fusion protein.
The biological activity of the purified fusion protein was assayed for G-CSF activity
by determining its ability to stimulate NFS-60 proliferation. Different amounts of
fusion protein, which was sterile-filtered and normalized for G-CSF equivalency, was
included in NFS-60 cell culture medium to replace G-CSF as a cell growth factor. As
shown in Fig. 32a, the biological activity of the fusion protein was approximately
1/1 Oth of the commercial G-CSF, filgrastim. The EC50 of the G-CSF control was
approximately 0.1 ng/ml, whereas the EC50 of the fusion protein was approximately 1
ng/ml as a G-CSF equivalent.
The TfR binding ability of the G-CSF-Tf fusion protein was also determined. As
shown in Fig. 32b, addition of unlabeled fusion protein caused a decrease in binding
of 1 2 5 I-labeled Tf to TfR in cultured Caco-2 cells, indicating that the fusion protein
maintained specific binding ability to TfR, even though the binding affinity was only
approximately 1/16th of that of Tf.
89
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.8
1 . 6 — G-CSF standard
G-CSF-Tf 1. 4
1 . 2
1
0.8
0.6
0. 4
0 . 2
0 1 ---------- 1 —
0.00001 0.0001 0.001 0 . 01 0 . 1 1 10 100
G-CSF e q u iv a le n t (ng/m l)
b 120
100
- G-CSF-Tf
4 0 -
1 10 100 1000
Transferrin Equivalent ( ug/m l)
Figure 32. In vitro study of G-CSF-Tf fusion protein activity, (a) Evaluation of G-CSF
activity of the purified G-CSF-Tf fusion protein. Proliferation of the murine
myeloblastic cell line NFS-60 was measured via MTT assay. The concentration of the
fusion protein was expressed as the G-CSF equivalence. Error bars represent SD. (b)
Evaluation of T f activity of the purified G-CSF-Tf recombinant fusion protein. 1251-
labeled Tf (3 pg/ml in serum-free medium with 1 mg/ml BSA) was added to Caco-2
monolayers. Different concentrations of unlabeled fusion protein were added to
compete for TfR binding. Error bars represent SD. N=4.
90
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.2 In vivo studies.
4.2.1 The in vivo activity by injection
The biological activity of the fusion protein was measured by subcutaneous (s.c.)
injection to mice. BDF1 mice were injected s.c. with 1 mg/kg G-CSF, 5 mg/kg G-
CSF-Tf, or PBS control. The molecular mass of the fusion protein is approximately
five times higher than G-CSF itself (G-CSF is 20 kDa, whereas G-CSF-Tf is 100
kDa); therefore, the final dosage for each is 5 pmol/kg. The day of dosage
administration was denoted as day 0. As shown in Fig. 33, the fusion protein exhibited
a comparable therapeutic effect to that of G-CSF. The time-effective curves of the
fusion protein and G-CSF were similar; both G-CSF and G-CSF-Tf conferred
maximum effect at day 1 (P > 0.1).
4.2.2 The oral availability of fusion protein sG-CSF-Tf
For the oral dosage experiments, BDF1 mice were given 10 mg/kg G-CSF, 50 mg/kg
fusion protein, or the PBS vehicle via a gavage needle (day 0). As shown in Fig. 34,
oral administration of G-CSF did not result in a statistically significant change (P >
0.9) in neutrophil level compared to the control (219 ± 85 cells per pi and 311 ± 9 7
cells per mm3 at day 1 for G-CSF treatment and control, respectively). However, mice
91
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
that were administered orally the fusion protein demonstrated a significant elevation in
ANC, 1,112 ± 232 cells per mm3 at day 1 (fusion protein vs. control, P < 0.1 at day 1).
Furthermore, the ANC of the group treated with fusion protein increased to 1,643 ±
575 cells per mm3 at day 2 days after the oral administration and only returned to the
baseline as long as 5 days after the administration.
4500
4000
—- P la ceb o (P B S )
• * ■ G-CSF
— - G -C SF-Tf
3500
3000
T — H
^ 2500
i- H
o 2000
5 1500
1000
500
5 0 1 2 4 3
time (days)
Figure 33. Myelopoietic effect of s.c. administered fusion protein, G-CSF, or control.
G-CSF (1 mg/kg) and fusion protein (5 mg/kg) were injected s.c. into BDF1 mice.
Blood samples were tested to determine ANC daily. Error bar represents SEM (n = 3
for control and G-CSF; n = 4 for fusion protein). Student's t test (at day 1): placebo
group vs. G-CSF group, P = 0.04; placebo group vs. fusion protein group, P = 0.004;
G-CSF group vs. fusion protein group, P = 0.44.
92
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2500 i
Placebo(PBS)
■— G-CSF
* G-CSF-Tf
2000
a,
M
0 )
1000
o
z
<
IF
500
0 4
Time (days)
Figure 34. Myelopoietic effect of orally administered fusion protein, G-CSF, control.
G-CSF (10 mg/kg) or fusion protein (50 mg/kg) was given orally via gavage needle to
BDF1 mice. Blood samples were tested to determine ANC every 24 h. Error bar
represents SEM (n = 3 for control and G-CSF; n = 4 for fusion protein). Student's t test
(at day 2): placebo group vs. fusion protein group, P = 0.01; placebo group vs. G-CSF
group, P = 0.9.
93
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.2.3 The dose response study o f the orally administered G-CSF-Tf
To make sure the myoloeitic effect in Figure 34 specifically induced by G-CSF-Tf
fusion protein, the dose-dependency of the recombinant fusion protein was tested in
oral administration. G-CSF-TF fusion protein was administered orally to BDF1 mice
by gavage needle at doses of 50 mg/kg, 25 mg/kg and 10 mg/kg. Myelopoietic effect
was determined by daily ANC measurement. Figure 35a shows the ANC of the
BDF1 mice at every 24 hours oral administration of different doses of G-CSF-Tf.
Figure 35b shows the ANC of each group at 48 hours after administration, which
corresponds to the peak in the time course. From the figures, it appears that the
increase of ANC, in BDF1 mice with oral administration o f G-CSF-Tf is dose-
dependent, and is not detectable ate the dose below 10 mg/kg.
94
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
,3
j/>
flj
O
z
<
«■ "G-CSF(10mg/kg)
^ "Fusion protein
(10mg/kg)
"Fusion protein
(25mg/kg)
Fusion protein
(50mg/kg)
50 100
TIME (Hours)
150!
(a)
3500
3000
I 5 2500
! I 2000
I J 1500
< 1000
500
0
control 10 (mg/kg) 25 (mg/kg) 50(mg/kg)
(PBS)
Amount of fusion protein
Figure 35. The dose-dependency of orally administered recombinant fusion protein,
(a) G-CSF-TF fusion protein was administered orally to BDF1 mice by gavage needle
at doses of 50 mg/kg, 25 mg/kg and 10 mg/kg. Myelopoietic effect was determined by
daily absolute neutrophil counts, (b) The absolute neutrophil count (ANC) at 48 hours
of each group. Error bar represent SEM (n=4).
95
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.2.4. The competition of G-CSF-Tf fusion protein oral absorption by Tf and albumin
To confirm that the in vivo transport of the fusion protein across the GI epithelia into
the blood circulation was mediated via TfR, BDF1 mice were orally administered 50
mg/kg fusion protein in the presence or absence of a 10-fold excess of Tf (500 mg/kg).
As shown in Fig. 36, a slight decrease of myelopoietic effect at day 1 was observed in
the presence of a 10-fold excess of Tf. However, at day 2, the resultant ANC for
fusion protein alone increased to 2,200 ±120 cells per mm3, whereas the ANC for
fusion protein plus 10-fold excess of T f decreased to 642 ± 100 cells per mm3 , similar
to the PBS control (on day 2, with Tf vs. without Tf: P < 0.01). Additionally, there
was no significant difference between treatment with the G-CSF-Tf and G-CSF-Tf
together with an excess of serum albumin (Fig. 37). Reduction of fusion protein
absorption by Tf, but not by albumin, suggests that the fusion protein is absorbed in
the GI tract via a TfR-mediated process.
96
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2500
- - - ♦ — Placebo (PBS)
— ■— G-CSF-Tf
— A— G-CSF-Tf+Tf
- -H - - Tf only
2000
/ —N
I — I
\ 1500
W
0 3
O
y 1000
4 5 3 1 2 0
Time (days)
Figure 36. Competition by free Tf on the absorption of orally administered G-CSF-Tf
fusion protein. Fusion protein (50 mg/kg), fusion protein (50 mg/kg) plus Tf (500
mg/kg), T f alone (500 mg/kg), and PBS as placebo were orally administered to BDF1
mice via a gavage needle. The myelopoietic effect was determined by daily ANCs.
Error bar represents SEM (n = 3). Student's t test (at day 2): placebo group vs. fusion
protein group, P = 0.01; placebo group vs. T f competition group, P = 0.1; placebo
group vs. T f only group, P = 0.3.
97
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4000
3500
f 3000
1 2500
8 2000
g 1500
< 1000
500
■ fusion (50mg/kg))
~ k — fusion(50m g/kg)+Tf(500mg/kg)
x fusion(50m g/kg)+BSA(500m g/kg
100
Time (hours)
Figure 37. Competition by albumin on the absorption of orally administered G-CSF-Tf
fusion protein. Fusion protein (50 mg/kg), fusion protein (50 mg/kg) plus albumin
(500 mg/kg), and fusion protein (50 mg/kg) plus T f (500 mg/kg) were orally
administered to BDF1 mice via a gavage needle. The myelopoietic effect was
determined by daily ANCs. Error bar represents SEM (n = 3).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5. The linker optimization
5.1 The design of linkers
The in vitro study indicates that Tf and G-CSF domains in the fusion protein with
short spacer (-L-E-) maintained less than 10% of the original activity of each
respective protein. We believe that the low activity is partially caused by the steric
hindrance of the two receptor binding sites in the fusion protein. In this study, we
prepared several fusion proteins by inserting flexible linker (GS) or helix-forming
peptide linkers A(EAAAK)n A, (n=2 to 4), between the two domains, as show in figure
38. The helical linkers were denoted as H2, H3, and H4 according to the number of
repeats. Two longer helix linkers, H4-2 (two copies of H4 linker) and H4-3 (three
copies of H4 linker) were also constructed. In the following description the linker
symbols will be used to denote the corresponding fusion proteins.
5.2 Linker insertion
There is only one X h o l cutting site in the plasmid. So after digesting with Xho 1
endonuclease, the plasmid is linear with the sticky end (Fig. 39). To prevent self
religation, the linear plasmid was treated with CIP to be dephospharylated.
99
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
As shown in figure 40, after Xhol digestion there was only one band shown on the
gel. And after the treatment with CIP, there was no size change detected on the gel.
Lb. —
----
H 4 - 3 :
LEGGGGSGGGGSGGGGSLE
LEAEAAAKEAAAKALE
LEAEAAAKEAAAKEAAAKALE
LEAEAAAKEAAAKEAAAKEAAAKALE
(LEAEAAAKEAAAKEAAAKEAAAKA)2 LE
(LEAEAAAKEAAAKEAAAKEAAAKA)3LE
Figure 38. The six fusion constructs, in comparison with the original no-linker
construct (denoted LE). The intergenic linkages are shown. The small grey box at the
N-terminus represents the secretion signal of G-CSF; the black box represents G-CSF;
the empty box represents Tf.
100
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Xho I
Xba I
EcoR V
TCGA
G-CSF vector Tf AGCT
Figure 39. The linearization of the plasmid by Xhol cutting.
101
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A B
Figure 40. Confirming the size of the linearized plasmid. Lane A, The plasmid was
cut with Xho 1. Lane B. the X hol cut plasmid was treated with CIP.
5.3 The linkers sequence formation by annealing and insertion
Two single strand oligonucleotides were designed for each linker, the sense and anti
sense strand, and the sticky ends, which match the Xhol cutting site was designed at
the both ends. The mixture of the two oligonucleotides was heated to 95 °C, then
slowly cooled down to allow the oligonucleotides to anneal to form double strand
102
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DNA which encode the linker sequence. The detailed sequence can be referred to the
methods section.
H2 LE H4
H4 LE H3
Figure 41. The fusion constructs with linker insert plasmids were digested with Pst 1.
The annealed double stranded DNA was ligated to the Xho 1 cut plasmid. The
ligation mixture was transformed into competent cells, and the positive clones were
then selected. The plasmids were tested by endonuclease Pst 1 digestion (Fig 41).
103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
There is one Pst 1 cutting site on the helix linker. From the figure, we can compare
the digestion pattern with the plasmid without insert.
From the agarose gel, the different patterns of the plasmids with linkers can be
distinguished from the original plasmid. Since there are two directions of insertion,
the size of the digestion product also showed the difference.
However, there were some plasmids having more than one copy of the insert. To
select the plasmids with multiple inserts, we designed a PCR method to test the length
of the linker, as shown in figure 42.
G-CSF lin k er" Transferrin
Figure 42. The design of PCR experiment to test the copy number of the linker. The
upstream primer: 5’-GGC AGC AGA TGG AAG A A C - 3 ’; The downstream primer:
5’-GGT CGC GGA AAC TCT GG - 3 ’.
Without linker, the PCR product should be 243bp. With linker, the size should be
243bp + linker bp. With N copies of linker, the size should be 243 bp + N times linker
bp.
104
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The further selection was done by both PCR and double endonuclease digestion. The
purpose of the PCR was to confirm the size of the insert; the purpose of the digestion
was to confirm the direction of the insert. Although some size information can be
obtained from digestion, it is not as sensitive to detect the number of copies.
control
300b p
220bp
(a)
Figure 43. The selection of the GS fusion protein plasmid, (a) The GS plasmids were
digested with Bam HI and Pst 1. (b) The PCR to measure the GS linker sequence.
105
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1 2 3 4 5 6 7 8 9 10
500 bp
400 bp
300 bp
200 bp
100 bp
(b)
Figure 43, continued.
The GS plasmids were double digested by BamHl and Pstl. By calculation, for the
LE plasmid (without linker), there should be a fragment of about 315 bp. After
insertion of the GS linker, this fragment should be 366 bp. Figure 43a, shows the shift
of the band from 315bp to 366 bp for some plasmids. Clones 3 and 9 showed band
shift to above 400 bp. This may be caused by double insertion.
106
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The PCR product with one copy of GS linker in the correct direction should be 294 bp.
(GS linker is 51 bp, other part is 243 bp). From both gels, the clones 1, 6, 7, 8,10 are
possibly correct (Fig. 43a & 43b). The clone 10 was sent for sequencing and was
verified to contain one copy of the correct GS sequence.
The same method was used to select the H3 fusion protein plasmids. The PCR
product should be 310 bp (243 bp +57 bp). From figure 44a, the clones 2, 3, 5, 6 and
11 have the right size of PCR product.
The H3 plasmids also were double digested by Bam HI and Pst 1. There is a Pst 1
cutting site in the H3 linker. The pattern was different when the linker was inserted in
different direction (Fig. 44b). The LE plasmid was digested by the two endonucleases
as control, there was a band of 315 bp. After inserting a H3 linker in the right
direction, the band showed a shift to 305 bp. When insert in the reverse direction, the
band was 277 bp.
107
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5 6 7 8 9 10 12 13 14
500 bp
400 bp
300 bp
200 bp
100 bp
— H3
(a)
3 4 5 6 7 8 9 10 11 12 13 14 LE control
1158bp
315bp
275bp
(b)
Figure 44. The selection of the H3 fusion protein plasmids, (a) The PCR products of
the H3 plasmids, (b) The H3 plasmids were double digested by Pst 1 and Bam HI.
108
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
control
(b)
Figure 45. The selection of H4 plasmid, (a) The double digestion of Bam HI and Pst
1. (b) The PCR products of the linker.
109
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The same methods were used to select the plasmid of other fusion proteins. An
example H4 is shown in Fig. 45. The selected plasmids were sent for sequencing.
The plasmids of G-CSF-Tf with linker GS, H2, H3, H4, H4-2 and H4-3 were
constructed. H4-2 and H4-3 were selected from the H4 verification, picking the ones
with double- and triple- inserts.
H 2 H 3 H 4 H4-2 H4-3 G S
— 100K D
(A)
— 100K D
(B)
Figure 46. The fusion proteins were recognized by both anti-Tf and anti-G-CSF
antibodies. (A) Westem-blot using anti-Tf antibody. (B) Westem-blot using anti-G-
CSF antibody.
110
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.4 Expression of fusion proteins and Westem-blot
HEK293 cell monolayer was transfected with different plasmids by using
Lipofectamine 2000 (Invitrogen). After 5 hours of incubation, the medium was
replaced with protein- free medium CD293. After a 5-day culture, the conditioned
medium was collected and subjected to 10% SDS/PAGE analysis. The proteins were
transferred to cellulose nitrate membrane and Westem-blot was performed using both
anti-Tf and anti-G-CSF antibodies. Horseradish peroxidase activity was detected by
the enhanced chemiluminescence (ECL) method. The Westem-blotting results were
shown in figure 46.
Table 2. Molecular weight of the linkers.
Linker Sequence Molecular
Weight (Dalton)
GS LEGGGGSGGGGSGGGGSLE 1430
H2 LEAEAAAKEAAAKALE 1567
H3 LEAEAAAKEAAAKEAAAKALE 2038
H4 LEAEAAAKEAAAKEAAAKEAAAKALE 2508
H4-2 LEAEAAAKEAAAKEAAAKEAAAKA
LEAEAAAKEAAAKEAAAKEAAAKALE
4775
H4-3 LEAEAAAKEAAAKEAAAKEAAAKA
LEAEAAAKEAAAKEAAAKEAAAKA
LEAEAAAKEAAAKEAAAKEAAAKALE
7041
I l l
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The six fusion proteins with different linkers all could be successfully expressed, and
all could be recognized by anti-G-CSF and anti-Tf antibodies. The insertion of the
linkers caused changes in the molecular weight of the fusion protein (Table 2), which
was also detected on the gel.
Table 3. The EC50 of the fusion proteins in the NFS60 proliferation assay.
Fusion protein with linkers EC50 (ng/ml)
GS 0.20
H2 0.19
H3 0.17
H4 0.20
H4-2 0.14
H4-3 0.30
LE 2.34
5.5 Linker selection by in vitro G-CSF activity
The purified fusion proteins were assayed for in vitro G-CSF activity by determining
their ability to stimulate NFS-60 cell proliferation. Different amount of fusion
proteins, which were sterile-filtered and normalized for G-CSF equivalency, were
included in NFS-60 cell culture medium to replace G-CSF as a cell growth factor.
The EC50 was calculated by fitting the data with sigmoidal curves. As shown in Table
3 and figure 47, among all the fusion proteins, the fusion protein G-CSF-(H4)2-Tf with
H4-2 demenstrated the highest activity, with a lowest EC5 0 about 0.143 ng/ml.
112
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Therefore, in the follow up study, the fusion protein with the H4-2 linker was used to
compare with G-CSF-Tf and native G-CSF.
H4-2
H4-3
0. 1 1 10
(a)
h 4-2
(b)
Figure 47. Evaluation of G-CSF activity of the G-CSF-Tf fusion proteins. The
proliferation of the murine myeloblastic cell line NFS-60 was measured via MTT
assay. The concentration of the fusion protein was expressed as the G-CSF
equivalence, (a) The bioactivity of six fusion protein with different linker were
compared in the NFS60 proliferation assay, (b) The proliferation curve of H4, H4-2
and GS. Error bars represent SD.
113
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6. Comparison o f the biological activity between H4-2 and LE fusion
proteins
Both LE and H4-2 fusion proteins were expressed in HEK293 cells in protein-free
medium. The proteins were purified using ammonium sulfate precipitation meathod
as described (Fig. 48).
6.1 In vitro activity comparison
The G-CSF activity of the two fusion proteins were compared by using NFS60 cell
proliferation assay (Fig. 49). A significant shift to the lower concentration was
observed for H4-2 compared to LE.
LE H4-2
100 kD
Figure 48. The SDS-PAGE analysis of the purified fusion proteins LE and H4-2.
114
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The TfR binding ability of the G-CSF-Tf fusion proteins was also determined. As
shown in figure 50, addition of unlabeled fusion protein caused a decrease in binding
of 1 2 5 I-labeled T f to TfR in cultured Caco-2 cells, indicating that the fusion protein
maintained specific binding ability to TfR, even though the binding affinity was only
about 1/16th of that of Tf. Unlike in the G-CSF assay, the fusion proteins H4-2 and
LE did not show any improved Tf receptor binding affinity.
0.9
LE
-H4-2
-G-CSF
0.8
0.7
0.6
0.3
0.2
o.i
0.0001 0.001 0.01 0.1 1 10 1000 100
G-CSF equivalence (ng/ml)
Figure 49. In vitro characterization of the fusion proteins H4-2 and LE. . Evaluation of
G-CSF activity of the purified G-CSF-Tf fusion protein. Proliferation of the murine
myeloblastic cell line NFS-60 was measured via MTT assay. The concentration of the
fusion protein was expressed as the G-CSF equivalence. Error bars represent SD.
115
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
140 i
— B --H 4 -2
- a - -Tf
p 120 -
I— I
in
I N
100 ifc
ra 80
60 -
40 -
-I
20
1 10 100
Tf equivalence (qg/m l) (
Figure 50. The TfR binding activity of the fusion proteins on Caco-2 cells, with
recombinant T f as positive control. 1 2 5 I-labeled T f (3 pg/ml in serum-free medium
with 1 mg/ml BSA) was added to the medium of Caco-2 monolayers. Different
concentrations of unlabeled fusion protein were added to compete for TfR binding.
Error bars represent SD.
116
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6.2 In vivo activity comparison
6.2.1 In vivo characterization o f the fusion proteins by subcutaneous injection
To determine whether the in vivo activity of the fusion protein H4-2 correlated with
the in vitro study, O.lmg/kg (0.5pmol/kg), 0.5 mg/kg (2.5 pmol/kg) and lmg/kg (5
pmol/kg) of G-CSF, and the equivalent molar amounts of LE, and H4-2 proteins were
subcutaneously injected to BDF1 mice. The molecular mass of the fusion protein is
approximately five times higher than G-CSF itself (G-CSF is 20 kDa, whereas G-
CSF-Tf is 100 kDa). The day of dosage administration was denoted as day 0. At the
maximum dose, 144-2 showed a higher myelopoietic effect than either LE or G-CSF
(Fig. 51).
117
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6000
- * - P B S
• C-CSF(lmg/kg)
LE (5mg/kg)
^-H 4 -2 (5 m g /k g )
5000
* = ■ 4000
D
I A
3000
u
< 2000
1000
0 20 40 60 80
Time (hours)
Figure 51. Myelopoietic effect of subcutaneously administered fusion proteins H4-2,
and LE, and native protein G-CSF in BDF1 mice. PBS was used as the vehicle
control. The doses of fusion proteins were adjusted to be equivalent to that of the
native G-CSF. The ANC were determined every 24 hours. Error bars represent SEM,
n=4 for all groups.
118
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4500
— -PBS
— LE (5mg/kg)
-'-■IE (2.5mg/kg)
— LE (0.5mg/kg)
4000
3500 -
3000 -
2500 -
2000 -
1500 -
1000 -
500 -
20 40
Time (hours)
60 80
Figure 52. Myelopoietic effect of subcutaneously administered different doses of LE.
PBS was used as vehicle controls. The ANC were determined every 24 hours. Error
bars represent SEM, n=4 for all groups.
Both H4-2 and LE showed effect at all three concentrations tested (Fig. 52 & Fig. 53).
For G-CSF, at the concentration of 0.1 mg/kg, the effect is significantly lower (Fig.
54). By comparing the peak effect at 24 hours, G-CSF and LE have similar
myelopoietic effect at doses of 0.5 mg/kg and 1 mg/kg G-CSF equivalence, which is
consistent with our previous report (Fig. 55). However, at these doses, subcutaneously
injected fusion protein H4-2 exhibited a higher effect than either G-CSF or LE in
absolute neutrophil count (ANC). At the low dose of 0.1 mg/kg G-CSF equivalence,
the group injected with fusion protein LE also showed a higher ANC compared to that
with G-CSF (Fig. 55).
119
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6000
-•--PBS
-G-CSF(1 mg/kg)
LE (5mg/kg)
H4-2(5mg/kg)
5000
4000
J/)
]§ 3000
u
< 2000
1000
60 80 20 40 0
Time (hours)
Figure 53. Myelopoietic effect of subcutaneously administered different doses of H4-
2. PBS was used as vehicle controls. The ANC were determined every 24 hours.
Error bars represent SEM, n=4 for all groups.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4000
- PBS
-G-CSF(1 m g/kg)
-G-CSF(0.5mg/kg)
-G-CSF (0.1 m g/kg)
3500
3000
= 2500
_ c / >
"aj
o
~ 1500
2000
z
<
1000
500
0 20 40 60 80
Time (hours)
Figure 54. Myelopoietic effect of subcutaneously administered different doses of G-
CSF. PBS was used as vehicle controls. The ANC were determined every 24 hours.
Error bars represent SEM, n=4 for all groups.
6.2.2 In vivo myelopoietic activity of the fusion proteins by oral administration
The BDF1 mice were randomly divided into different groups. The mice were fasted
16 hours before the oral administration of the proteins. The proteins were given at
time 0 by using gavage needle. To make a clear distinction among the proteins, a low
121
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
dose of the proteins (G-CSF 4mg/kg, LE 20 mg/kg, H4-2 20mg/kg) were chosen for
this oral experiment, which were lower than the concentration in our previous paper.
6000
G-CSF
5000
H4-2
40 0 0
J f)
J 300 0
u
z
< 2000
1000
0.5 0 1.5
dose G -CSF equivalent (mg/kg)
Figure 55. The dose -response curve of the in vitro activity of fusion protein LE and
H4-2 and native G-CSF. Myelopoietic effect of subcutaneously administered fusion
proteins H4-2 and LE, and native G-CSF in BDF1 mice. PBS was used as vehicle
controls. The doses of fusion proteins were adjusted to be equivalent to that of the
native G-CSF. The ANC were determined at 24 hours after administration. Error bars
represent SEM, n=4 for all groups.
As shown in figure 56, only the H4-2 group showed significantly increased ANC at
this low dose. In oral administration of the time the response was different from the
122
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
subcutaneous curves, with the peak of ANC appeared at 48 hours, instead of 24 hours.
This finding is consistant with our previous study the fusion protein has the delayed
myelopoitic effect when it was orally administered. To confirm that the fusion
proteins LE, H4-2 retained the G-CSF activity, the subcutaneous injection of these
proteins at this low dose was done using the same batch of fusion protein at the same
time, and the results agree with the previous section (Fig. 57).
3000
P B S
G-CSF(4mg/kg)
LE(20mg/kg)
- » - H4-2(20mg/kg)
2500
2000
U
z 1500
<
1000
500
0 20 40 60 80
Time (hours)
Figure 56. Myelopoietic effect of orally administered fusion proteins H4-2 and LE,
and native G-CSF in BDF1 mice. PBS was used as vehicle controls. The doses of
fusion proteins were adjusted to be molarly equivalent to that of the native G-CSF.
The ANC were determined every 24 hours. Error bars represent SEM, n=4 for all
groups.
123
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7 0 0 0
-PBS
- G-CSF(0.04mg/kg)
- LE(0.20mg/kg)
-H 4-2(0.20m g/kg)
6 0 0 0
5000
4 0 0 0
3 0 0 0
2000
1000
80 20 40
Time (hours)
Figure 57. Myelopoietic effect of subcutaneously administered fusion proteins, LE and
H4-2, as well as native G-CSF, at different doses in BDF1 mice. PBS was used as
vehicle controls. The ANC were determined every 24 hours. Error bars represent
SEM, n=4 for all groups.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
D iscussion
1. Conjugation o f recombinant T f with a cysteinyl residue
The heterogenicity of the chemically conjugated In-Tf and G-CSF-Tf was shown in
the previous work (Widera et al. 2004; Xia et al. 2000). This was largely due to the
multiple SPDP (N-Succinimidyl 3-[2-pyridyldithio]-propionamido) modification sites
on Tf. To avoid the problem of heterogenicity of the conjugates for the potential
application as therapeutic drugs, we originally planned to take two different
approaches. The first approach is to prepare recombinant T f containing only one
single cysteinyl residue, Tf (Cys), on the surface of the protein molecule. We have
successfully mutagenized Ser-288 to Cys-288 in cDNA of T f and sequence has been
verified. However, there is another inherent drawback of chemical conjugation
methods that is the loss of product through the purification process and inefficiency of
the conjugation method itself. The reaction process also needs to be carefully
monitored and controlled. The problem of the chemical conjugation can not be
avoided even if we use S288C-Tf for the reaction. Therefore, we decided to persue the
second approach, to prepare the recombinant G-CSF-Tf fusion protein. The
production of a recombinant G-CSF-Tf fusion protein will result in the ability to
obtain a large amount of a chemically defined conjugate at a relatively low cost.
However, there is a chance that the fusion protein may lose the bioactivity of G-CSF
125
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
or Tf. If we could not obtain an active recombinant fusion protein, we would have to
go back to make the 1:1 ratio chemical conjugation.
2. Recombinant granulocyte colony-stim ulating factor-transferrin
fusion protein as an oral myelopoietic agent
To achieve our goal recombinant fusion protein consisting of both human G-CSF and
human Tf moieties was engineered and demonstrated to confer myelopoietic effect
when orally delivered in an animal model. Recombinant fusion proteins have been
reported to be effective in the development of oral vaccines (Fitches et al. 2004;
Russmann et al. 2001). However, this is the first demonstration, to our knowledge,
that Tf moiety in Tf-based recombinant technology enhances the GI absorption of a
therapeutic protein drug.
It was reported that the fusion protein nerve growth factor - transferrin with the Tf
moiety at the C-terminus kept almost 100% the Tf receptor binding affinity and the
bioactivity of the nerve growth factor, while the construct with T f at the N-terminus
showed much lower activity (Park et al. 1998). It was also reported that the G-CSF-
albumin recombinant fusion protein (Albugranin), with G-CSF genetically fused at its
N-terminus to the C-terminus of recombinant human albumin, is functionally active
126
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(Halpern et al. 2002). The two studies reported different orientations of the carrier
domain and the functional domain in the fusion protein.
In this study the recombinant fusion proteins were designed at two orientations: G-
CSF at the N-terminal or Tf at the N-terminal. The orientation is very important for
the fusion protein to keep the bioactivities of both moieties. Our results showed that
the fusion protein G-CSF-Tf, which contains G-CSF at the N-terminus, has
maintained partially the G-CSF and T f receptor binding activity. Even though it is
less than 10% in vitro, it is much higher compared to the fusion protein Tf-G-CSF,
which contains the Tf moiety at the N-terminus.
By testing the activity of mutant proteins using an in vitro proliferation assay and an
ELISA binding competition assay, Young et al showed that there are no surface-
exposed residues on helix D of G-CSF that are critical for receptor interaction. Instead,
there is a region on helices A and C of G-CSF that is involved in receptor binding and
activation. The most important residue in this binding patch is G lul9 (Young et al.
1997) . This means that the N-terminal of G-CSF is critical for its bioactivity. In
addition, from the model of T f and T f receptor binding, the C-lobe of Tf binds to the
receptor, and the N-lobe binds to the cell membrane to stabilize the complex (Cheng et
al. 2004). These could explain that why the fusion protein G-CSF-Tf is better
compared with the fusion protein Tf-G-CSF. However, the result of the in vitro NFS-
127
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
60 proliferation study showed that the bioactivity of fusion protein G-CSF-Tf is less
than 10 % of G-CSF. This may be caused by the steric hindrance. G-CSF is only
about 20 kD, while T f is abount 80 kD. The G-CSF could be easily blocked by the big
T f moiety. And the loss of the bioactivity also could be caused by the interference of
folding between the two moieties.
In the introduction, we have already discussed why the fusion protein could not be
expressed by bacteria. In eukaryotic expression, the recombinant fusion protein with
the secretion signal of either G-CSF or Tf could be secreted into the medium. While
the fusion protein without secretion signal stayed inside the cell.
The expression level of Tf-G-CSF is lower compared to G-CSF-Tf. This may be
caused by the mis-folding o f Tf-G-CSF fusion protein, which could induce
degradation by the cell. It also could be caused by the different efficiency of
transfection. Because the Tf-G-CSF exhibits much lower expression and G-CSF
bioactivity, the G-CSF-Tf fusion protein was used for the study.
We tried multiple methods to isolate the secreted fusion protein from the conditioned
medium, such as gel filtration and ion exchange column chromatography (data not
show). It was very difficult to isolate the fusion protein from the serum-containing
medium, since large amounts of serum proteins interfere with the purification.
128
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The protein-free medium solved the problem. This is a medium designed for HEK293
cell growth in suspension. Growing in suspension, the cell could reach higher density
in the same volume of medium. However, the condition is very hard to control. And
the cells need to grow in the CO2 incubator with shaking. We found the HEK 293 cell
was still attached to plates in protein free medium. In the absence of the cell attach
factor, the cell grows in multiple layers, instead of monolayer, which increased the cell
numbers expressing the fusion proteins. The expression amount of the fusion protein
could reach 800 pg/ml. After ammonium sulfate precipitation, the purity of the fusion
protein could reach 80%.
Even though the in vitro biological activity of the recombinant G-CSF-Tf approached
only 1/10th of that by individual G-CSF or Tf, the fusion protein exhibited a
comparable myelopoietic effect to that of G-CSF when administered s.c. in mice. This
elevated in vivo effectiveness may be due to a prolonged plasma half-life and a
decrease of clearance as previously described in a recombinant G-CSF-albumin fusion
protein (Halpern et al. 2002). Conceivably, G-CSF-Tf may have a clearance rate
similar to Tf, instead of G-CSF. This prolonged plasma half-life may compensate for
the lower biological activity of the fusion protein. The fact that G-CSF-Tf exhibited a
comparable in vivo pharmacological effect even though a much lower in vitro
biological activity was observed suggests that a significant improvement of the
129
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
therapeutic efficacy of G-CSF-Tf could be achieved if the biological activity of
recombinant fusion protein could be augmented by other means.
To achieve this goal, one possibility is to alter the length and/or amino acid
composition of the linker between the G-CSF and Tf moieties. In the fusion protein
described in this article, only a short linker of two amino acids (-Leu-Glu-) was
inserted between the Tf and G-CSF moieties. Conceivably, such a short linker may
decrease the chance for immunogenicity of the fusion protein; but it may also block
the accessibility of the functional moiety for its receptor binding. It is noteworthy that
a recombinant nerve growth factor-Tf fusion protein has been produced previously, in
which a linker of IgG3 hinge region was inserted between the two protein moieties to
retain the in vitro biological activity (Fitches et al. 2004).
The most striking finding in this study is that oral administration of the fusion protein
elicited a clinically relevant increase in ANC, whereas oral G-CSF was totally
ineffective. The fusion protein, at an oral dose of 50 mg/kg, elicited a 5-fold increase
of ANC in BDF1 mice. Because G-CSF has been shown clinically to have a nonlinear
dose-response (Hestdal et al. 1993; McCullough et al. 1999), a comparison of the area
under the curve or bioavailability cannot be drawn here at this time. However, this
study clearly demonstrates that the orally administered fusion protein does indeed
have pharmacological effect. In addition, the orally administered fusion protein
130
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
exhibited a prolonged effect on the neutrophil proliferation in mice. The increase of
ANC for the oral administration of 50 mg/kg fusion protein in BDF1 mice was
observed up to 4-5 days, whereas only up to 2 days for s.c. administration (Figs. 3 and
4). Interestingly, this sustained myelopoietic activity was not observed in the oral
delivery of the G-CSF-Tf chemical conjugate (Widera et al. 2003b).
3.The linker optimization
The G-CSF-Tf fusion protein can be absorbed across gastrointestinal epithelia, as
indicated by its myelopoietic activity when orally administered in BDF-1 mice (Bai et
al. 2005). However, the fusion protein only maintained a small fraction of the cell
prolifeative and TfR-binding activity in in vitro study. Since T f and G-CSF in the
fusion protein was linked by only a short LE dipeptide spacer, it is unlikely that these
two protein domains can be separated in the body to release the fully active form of G-
CSF. Therefore, the increase of ANC in BDF-1 mice most likely correlates directly to
the in vitro cell proliferative activity of the fusion protein. The low in vitro activity
may be an indication of the interference of the two domains with each other in the
fusion protein. Such interference will most likely decrease the binding of the fusion
protein to G-CSF and Tf receptors. Conceivably, if the in vitro activity of the fusion
protein could be improved, the in vivo efficacy would be further enhanced.
131
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In order to optimize the intrinsic bioactivity of the fusion protein, we investigate the
spacers between G-CSF and Tf domains that will keep them at a distance. Several
studies have reported that the flexibility and hydrophilicity of the peptide linkers were
important for the preservation of the functions of the individual domains in a fusion
protein (Argos 1990; Robinson & Sauer 1998). The flexible linker, (GGGGS)3, is the
most commonly used linker between fusion protein domains, especially for single
chain Fv (Huston et al. 1988; Trinh et al. 2004). However, study showed that other
fusion proteins may lose their activity using the flexible spacer to link the two
moieties (Maeda et al. 1997). Furthermore, for orally administrated protein, a flexible
linker without secondary structures will be easily accessed by the digestive enzyme in
the GI tract. On the other hand, a peptide linker, A(EAAAK)n A, was reported to form
an a-helix conformation, which could control the distance and reduce the interference
between the domains (Arai et al. 2001). It was reported that the helix linker could
effectively separate bifunctional domains of the fusion protein (Arai et al. 2001).
In order to increase the in vitro biological activity and, consequently, the in vivo
myelopoietic effect of G-CSF-Tf, we inserted different spacers to separate the Tf and
G-CSF domains in the fusion protein. We chose six different peptide sequences to
construct the spacer between G-CSF and Tf domains, including peptides with random
(GS), short a-helical (H2, H3, and H4), and long a-helical (H4-2 and H4-3)
conformation. From the NFS-60 proliferation results, all the fusion proteins with
132
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
inserted spacer appeared more potent than G-CSF-LE-Tf (Fig. 47). Among the
spacers, the fusion protein with H4-2 spacer showed the highest cell proliferative
activity in NFS-60 cell assay. It has been reported that spacers with helical
conformation worked better than that with random conformation for maintaining the
distance between domains in fusion proteins (Maeda et al. 1997). In the case of G-
CSF-(H4 )2-Tf, however, H4-2 not only provide a rigid, but also a longer spacer
between G-CSF and T f than in fusion proteins with H I, H2 and H4 spacers.
Consequently, there is an approximately 10-fold decrease of the EC50 from that of G-
CSF-LE-Tf, even though G-CSF-(H4 )2-Tf is still less effective than native G-CSF in
in vitro proliferative assay (Fig. 49).
Annealed synthetic phosphorylated oligonucleotides were used to create the linker
consisting of double strand DNA between G-CSF and Tf. The linkers were designed
with sticky ends that are complimentary with the Xho 1 cutting site. Because the
linkers were inserted with the same cutting site, there are two possible directions of the
linkers. It causes the trouble to select the plasmids with the right insert. And the
linker can also self-ligate by the Xho 1 cutting site first, then inserted into the plasmid.
More than one copies of the linker were inserted in this situation.
In the helix linkers, the Pst 1 digestion site were designed. The cutting site is not in
the middle of the linker. So the direction of linker could be tested by the Pst 1
133
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
digestion, since the size of the cut fragment is different when the linker was inserted in
the reverse direction.
To test the copy number of the linkers, a PCR experiment was designed. The two
primers were designed to PCR 243 bp from the G-CSF and T f sequences in total.
After the insertion of the linker, the PCR products grew longer, by the length of the
inserted sequence.
The six fusion proteins with different linkers were successfully developed and
expressed. Results showed that all 6 fusion proteins could be recognized by anti-Tf
antibody (Fig. 46a) and anti-G-CSF antibody (Fig. 46b). The molecular weight of the
six fusion proteins correlated well with the size of the linker peptides, from the
smallest GS (1.4 kD), to the largest H4-3 (7.0 kD).
The G-CSF bioactivities of the fusion proteins were compared in the NFS-60
proliferation assay. As show in Table 3, the fusion protein with the H4-2 spacer
exhibited the highest effect in cell proliferation. And as show in figure 47b, the H4-2
linker works much better compare with the flexible linker GS.
According to EC 50 values, all the linkers could increase the G-CSF bioactivity of the
fusion proteins. The best one, H4-2, exhibited aboutlO fold higher activity compared
134
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
to the fusion protein with only LE linker. This means the distance between the two
moieties of the fusion protein is important to maintain the functions. However, the
fusion protein with the linker H4-3 only has 50% G-CSF bioactivity compared to H4-
2. This may suggest that the maintaining of the function of a fusion protein is not only
related to the length of the linker.
To our surprise, when we compared the Tf receptor binding affinity by the surface
binding assay, H4-2 did not show improvement over LE, which retains only about
1/16th of the activity of the native Tf protein (Fig.50). The lack of improvement of
TfR binding by inserting the spacer indicates that the decrease of TfR binding in the
fusion protein is most likely due to the modification of the N-terminus of T f domain
rather than a simple steric hindrance of the receptor-binding site in the fusion protein.
Attempt was made to reverse the sequence of G-CSF and T f in the fusion protein, but
the Tf-G-CSF-plasmid expressed in much lower amount in HEK293 cells. Therefore,
the low affinity to TfR binding remains to be a problem for further developing the Tf-
fusion proteins into oral bioavailable protein drugs.
The male BDF1 mice was used in all animal experiments described in this report.
This is because BDF1 mice is relatively sensitive to human G-CSF. In addition,
unlike other commercially available chemotherapy- or radiation induced neutropenia
mouse models (e.g. Perry Sceientific, Inc., San Diego, CA), BDF1 mice are normal
135
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
animals and will be good model to study the physiological GI absorption without the
interference of the complications associated with drug or radiation treatment. The
current application focuses on the optimization and the mechanism of the GI
absorption of Tf-fusion protein. The neutropenia mouse models could be used in the
future when the therapeutic efficacy of the fusion proteins are evaluated.
We found that the in vivo myelopoietic activity of the fusion proteins of G-CSF and Tf
correlate with the in vitro biological activity. The in vivo myelopoietic effectiveness in
both subcutaneous injection (Fig. 55) and oral administration (Fig. 56) correlated very
well with the in vitro cell proliferative activity in cultured NFS-60 cells (Fig. 49). For
the subcutaneous injection study, the mice were given three different doses of G-CSF,
fusion protein LE and H4-2 (0.1 mg/kg, 0.5 mg/kg and 1 mg/kg G-CSF equivalance).
From figure 52-54, the increase of ANC peaked at 24 hours. And there is a ceiling
effect of the subcutaneous injected G-CSF and G-CSF-Tf fusion proteins. By
comparing the response at 0.1 mg/kg G-CSF equivelance, H4-2 is more potent than
LE, and LE is more potent compared to G-CSF, even though the in vitro NFS60 study
showed that both H4-2 and LE were not as potent as G-CSF.
In subcutaneous administration, the superiority of G-CSF-(H4)2-Tf over native G-CSF
in myelopoietic activity in BDF-1 mice is possibly due to a prolonged plasma half-life
and a decrease of clearance as previously described in a recombinant G-CSF-albumin
136
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
fusion protein (Halpern et al. 2002). The advantage in pharmacokinetic parameters
supersedes the fact that the fusion protein is less active than the native G-CSF in in
vitro cell proliferation assay.
When the three forms of G-CSF and its Tf-fusion proteins were orally administered to
BDF1 mice at a dose equivalent to 4 mg/kg of G-CSF, only the G-CSF-(H4)2-Tf-
treated mice group showed a significant increase of the ANC (Fig. 56). The dose in
this study was significantly lower than that in our previous report, i.e., 10 mg/kg of G-
CSF equivalence, and therefore no significant myelopoiesis was detected in mice that
were orally administered with G-CSF-LE-Tf. However, at this low dose, a delayed
onset of the myelopoietic effect was observed in orally administered G-CSF-(H4)2-Tf
(Fig. 56), which is consistent with our previous report on oral administration of a
higher dose of G-CSF-LE-Tf (Bai, et. al. 2005).
The pharmacokinetics of G-CSF is very complicated. G-CSF is rapidly removed from
the body by a combination of renal and active neutrophil clearance processes. Many
different half-life of G-CSF were reported, depending on the dose, animal model, and
so on. It was known that increasing neutrophil counts results in faster clearance of G-
CSF. In patients undergoing stem cell transplantation, the half-life of G-CSF
exceeded 15 hours during severe neutropenia but decreased during the recovery of
neutrophils (Shimazaki et al. 1995).
137
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The mechanism of enhancing of neutrophil count by G-CSF is also very complicated.
G-CSF is an unusual cytokine, which is a protein mediator formed primarily by
leukocytes, and It acts on leukocytes but not only on a local level (Hartung et al.
2003). G-CSF was formed at the site of infection or inflammation but exerts its
primary action at the remote organ, for example, bone marrow. And G-CSF resulted
in neutrophilia by increasing the production of neutrophils in the bone marrow and
hastening the release of less mature forms into the circulation (Tamura et al. 1987;
Tkatch & Tweardy 1993; Ulich et al. 1989).
The pharmacokinetics will be even more complicated for the G-CSF-Tf fusion protein.
The fusion protein, when orally administered, may be released slowly. As discussed
above, the fusion protein given by oral administration may have longer half-life
compared with that of the fusion proteins given by injection, while having a milder
effect in terms of neutrophil count at the peak. The neutrophil level reach the peak at
48 hours in BDF1 mice orally administered the G-CSF-Tf fusion protein. It was
reported that after G-CSF injection, neutrophil increased to the peak levels at about 24
hours (Lord et al. 2001). And in our experiments, the subcutaneous injection of G-
CSF-Tf fusion proteins also induced the neutrophil peak level at about 24 hours.
These results suggest that with slower release of the G-CSF fusion protein, the
clearance process is delayed, allowing longer effect of the protein.
138
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Unlike the fusion proteins, chemical conjugates of Tf and G-CSF induced peak effect
at 24 hours (Widera et al. 2004). There are two possibilities that may account for the
difference in the duration of the oral myelopoietic activity between the fusion protein
and the chemical conjugates. First, the non-reducible characteristic of the fusion
protein may lead to a longer half-life in serum, whereas free G-CSF may be released
from the disulfide-linked conjugates after entering into the blood circulation (Widera
et al. 2004; Widera et al. 2003b). Secondly, there may be a phenomenon such as a
depot effect of the orally administered fusion protein in some organs, for instance, the
liver or pancreas. These organs are targeted via the portal vein after GI absorption and
are responsible for Tf storage, metabolism, and secretion (Aisen 1984). A depot effect
could explain why the prolonged effect on myelopoiesis was not observed with s.c.
administered fusion protein (Fig. 57). It is conceivable that the fusion protein entering
the circulation from the injection site would have to compete with the high
endogenous T f concentration in the blood for receptor binding in various organs.
Additionally, co-administration of a large excess of free T f abolished the sustained
myelopoietic effect of G-CSF-Tf and caused a shift of the ANC peak to day 1 (Fig.
36), a response that bears a resemblance to that of the s.c. injection (Fig. 33). The
effect of free T f on the myelopoietic effect of G-CSF-Tf suggests that the GI
absorption of the fusion protein is a TfR-mediated process. However, the shortening
139
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
of the duration of the myelopoietic effect may also suggest that a phenomenon such as
a depot effect occurs in the liver or other organs because of the binding of the G-CSF-
Tf to tissue-associated TfR. In this case, the excess Tf would compete with the fusion
protein to the binding sites at the depot organ. Further work is needed to demonstrate
such a competition between T f and G-CSF-Tf.
4. The sustained increase o f ANC observed in orally administered G-
CSF-Tf fusion protein
The oral administration of G-CSF-Tf maintained an increased ANC in BDF1 mice for
3 to 4 days. Since the life span for neutrophils is only about 12 hours, our finding
implies that either the half-life of G-CSF-Tf is significantly longer than that of G-CSF,
or there is a sustained release mechanism of G-CSF-Tf transport from the intestine to
blood stream. The fact that subcutaneously injected G-CSF and G-CSF-Tf have a
similar effect on neutrophil counts may suggest that the prolonged effect of orally
administered G-CSF-Tf is most likely due to a sustained release rather than the plasma
half life.
Our hypothesis is that orally administered G-CSF-Tf is transported across the GI
epithelium, and subsequently, to the liver via the portal vein. The fusion protein will
accumulate either in the intestinal epithelum or in the liver, possibly as an apo-Tf
140
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
form, followed by a slow release into blood circulation as the diferric form. At the
present time, we believe that intestinal epithelium, rather than the liver, is more likely
the retention site for the sustained release of orally absorbed G-CSF-Tf (Fig. 58). The
reason is that, once delivered into the portal vein, the fusion protein will be mixed with
a high concentration of endogenous Tf in the blood before reaching the liver. Such a
dilution effect will unlikely make G-CSF-Tf selectively retained in the liver. There
was the evidence that Tf accumulation in GI epithelium. Our lab’s study suggested
that apically-intemalized T f is retained longer in an intracellular compartment in
Caco-2 cells, the enterocyte-like cell, and this retention was not seen in MCF-7 cells
(Lim & Shen 2005).
141
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TfR
A
A p o-T f
A D iferric-T f
T f in M ucosal Fluid ^
Y7 ▼ ▼
▼ A pical TfR V
Apical
A Y Y Y Y Y Y Y Y W Y Y W Y Y Y Y A
I tjI (4)/ DM T1 | T J
BE (3) ' “
%k ■ «
Recycling
< 70%
Transcytosis
(5)
Basal TfR
A
A
T f in Blood A
Figure 58. A hypothetical scheme of the regulatory mechanism for the transport of Tf
from the mucosal side of the intestinal epithelial cells to the blood.
As shown in Fig 58, there are several possible passway for T f transported from apical
side to basolateral side. 1. Miss-sorting of basolateral membrane in basolateral
endosomes (BE) would allow a small number of TfR to appear on the apical surface.
2. Orally administered T f would bind to apical TfR and internalized to apical
endosomes (AE), where differric T f would be converted to apo-Tf due to the
acidification. 3. Apo-Tf in AE would be transported to a common endosome (CE) by
a similar process that has been described for apo-Tf in BE(Hughson & Hopkins 1990).
142
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4. Iron uptake from the mucosal surface via divalent metal transporter 1 (DMT1)
could reach CE due to the endocytosis of DMTl(Ma et al. 2002). 5. The conversion
of Apo-Tf to diferric Tf in CE would accelerate the transport of differric T f from CE
to the basolateral membrane via exocytosis, and eventually to be released to the
blood(Ma et al. 2002).
Both intestinal epithelium and the liver could be the retention site of Tf-fusion
proteins. However, the transport of Tf-fusion protein from the intestine to the liver
will be encountered with a large excess of endogenous T f of the blood in mesenteric
and portal veins. Therefore, it is less likely that the Tf-fusion will be retained in the
liver. The figure is produced by Dr. Jennica L. Zaro.
There is another concern whether the fusion protein remains as a whole protein or is
digested into two moieties. From the current data, we believe that the fusion proteins
were not cleaved. There is no cleavage site on the spacers, and in subcutaneous
injection experiments, the fusion proteins showed higher myelopoitic effect than G-
CSF in the mice model, while in in vitro experiments the activity of the fusion protein
was much lower. This is likely to be caused by the longer half-life. So there is a
higher chance that the fusion protein remains as a whole protein after it was released
to the plasma.
143
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Generally speaking, the prolonged effect of the orally delivered fusion protein should
be beneficial in clinical applications, since less frequent doses of the recombinant
fusion protein will be needed to maintain the increased level of circulating neutrophil
number. However, the increase of ANC induced by orally administered fusion protein
in the animal model did not reach the level induced by s.c. injected fusion protein, and
the peak was reached only after a prolonged period, i.e. 48 hours compared to 24
hours by s.c. injection. Therefore, the clinical conditions require an immediate
increase of neutrophil number, and/or a relatively high level of ANC, the orally
administered G-CSF-Tf fusion protein may not be effective. The oral doseage form of
the fusion protein may be more desireable for the treatment of chronic neutropenia, in
which the normal level of neutrophil needs to be maintained.
In this study, the effect of G-CSF-Tf fusion protein was tested in the mouse model,
which is less sensitive to human G-CSF. From the current data, it is hard to draw
solid conclusions as to how the fusion protein will perform in human. In addition,
only the effects of a single dose of G-CSF-Tf fusion protein have been studied. It was
found that the effect of orally administered fusion protein has a different profile
compare to that of s.c. administered G-CSF. However, the effect of multiply doses of
the fusion protein, in both oral and s.c. route of administration, remains to be
determined;
144
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The G-CSF-Tf fusion protein increases both the total white blood cell number and the
neutrophil count. Indicating that the treatment also increases the number of other
lymphocytes. In bone marrow transplanted patients, both the proliferation and the
differentiation of the G-CSF effect will be desired. However, in the patients with
neutropenia, the neutrophil proliferation is more important, since only the increase of
the circulating neutrophil is needed. In this case, increasing the monocyte, eosinophil,
basophil and platelet may cause undesired side effects.
In this study, the ANC was measured, which consists both the segmented neutrophils
and the matured band cells. Since the band cells account for a small fraction of all
matured neutrophils, their percentage is difficult to be determined reliably. To better
understand the biological mechanism of the fusion protein, it will be helpful to
identify the specific effects of G-CSF-Tf on proliferation and differentiation, of each
cell type.
145
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5. Summary
The investigation was aimed at achieving oral delivery of poorly absorbed peptides
and protein therapeutics. As a continuous effort to investigate transferrin receptor
(TfR) - mediated transcytosis in the gastrointestinal (GI) tract, the G-CSF-Tf fusion
proteins with a series of linkers were developed by DNA recombinant technology.
Both the in vitro models and the animal model were used in this study. Murine
myeloid leukemic cell line, NFS60, which is known to express the G-CSF receptor,
was used to determine the EC 50 o f the fusion proteins by the dose-dependent
stimulation of the proliferation of NFS-60 cells. The human colonic carcinoma cell
line Caco-2 was used to determine the binding affinity of the fusion proteins to TfR.
The BDF1 mice model was used to study the in vivo activity of the fusion proteins.
The major findings of this investigation are summarize as follows:
(1) A 2 8 8 Ser-Cys mutantion Tf-pCMV-tag-2B plasmid was constructed. The mutant
Tf could be used to prepare 1:1 ratio G-CSF-Tf chemical conjugate.
146
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(2) The study developed a series of novel G-CSF fusion proteins that could be used for
treating various diseases.
(3) The fusion proteins containing a therapeutic peptide/protein and a T f domain could
be transported across the epithelium layer which expresses TfR.
(4) The design of the fusion protein is important. The orientations of the two moieties
are critical for the fusion proteins to keep the bi-functional activity.
(5) The recombinant fusion proteins can be expressed in HEK 293 cell in protein free
medium, with the expression amount of up to 80pg/ml. And the purity of fusion
protein after ammonium sulfate precipitation is 80 %.
(6) The ability of a recombinant G-CSF-Tf fusion protein with Leucine and
Glutamate short spacer to achieve oral bioavailability was proved by this study.
(7) The orally administered fusion protein elicits sustained elevation of the absolute
number of neutrophils in the BDF1 mice model. However, an elucidation of the exact
mechanism of the absorption and the sustained effect is needed to expand and advance
this approach to other protein drugs and, ultimately, to therapeutic applications.
147
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(8) The peptide spacers, which can maintain a distance between the two domains to
minimize the mutual interference, can improve the G-CSF bioactivities at different
levels. The insertion of the spacer peptide in the fusion protein with an alpha-helical
conformation, H4-2, results in a 10-fold increase of the in vitro cell proliferative
activity without significant change in TfR binding in culture cells. Furthermore, the
fusion protein with H4-2 spacer shows an improved efficacy for both subcutaneous
and oral administration in BDF1 mice. Therefore, the design of spacers is important
for the future studies of fusion proteins as either injectable or oral therapeutic agents.
(9) In general, this study suggests Tf could be used as a delivery vehicle to improve
the GI absorption of peptide and protein drugs. The recombinant G-CSF-Tf fusion
proteins demonstrated that it is feasible to prepare a recombinant protein with both
oral absorption and therapeutic effect. These findings provide an approach for the
future development of orally efficacious protein drugs.
148
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Future prospective
1. What is the underlying mechanism o f the in vivo m yelopoietic
activity o f the orally administered G-CSF-Tf fusion protein?
To understand and to further improve the therapeutic efficacy and bioavailability, the
pharmacokinetic properties of the fusion protein need to be studied. The method to
detect the blood concentration of orally administered fusion protein in mice needs to
be developed. There are commercial RIA and ELISA kits available for both G-CSF
and Tf that are highly specific for human G-CSF and Tf. On the other hand, it had
already been shown that the fusion proteins we developed could be recognized by both
anti-Tf and anti-G-CSF antibodies. The fusion protein specific ELISA (sandwich
ELISA using anti-G-CSF and anti-Tf as capture- and detection-antibodies
respectively) could be used to detect the plasma concentration. Both the
pharmacokinetics (PK) and bioavailability can be obtained by measurements over
multiple time points. The pharmacokinetic parameters, such as area under the curve
(AUC), apparent plasma half-life (fi/2), mean residence time (MRT), maximum plasma
concentration ( C m ax) and time to reach maximum plasma concentration (tMAx) can be
fitted using a computer program (e.g. WinNonlin). The biodistribution of the fusion
protein is another important issue of the fusion protein. It is necessary for the
149
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
explanation of the prolonged effect on the increased of ANC increase by orally
administered G-CSF-Tf fusion protein. The G-CSF-Tf fusion protein maintained an
elevated level of ANC for 3 to 4 days (Figure 34), while the injected fusion protein
only maintain the effect for one day (Figure 33). Since the life span for neutrophils is
only about 12 hours, this finding implies that either the plasma half-life of the fusion
protein is significantly longer compared to G-CSF or there is a sustained release
mechanism of G-CSF-Tf transport from the intestine to the blood stream. There are
several ways of studying the biodistribution of the fusion protein: (1) tissue immuno-
staining; (2) fusion protein concentration determination using ELISA from tissue
lysates; (3) imaging using radio or photo-labeled fusion protein or (4) concentration
determination using radio-labeled fusion protein from separated tissues.
In our study, the BDF1 mice model was used. In our future study the inflammation
mice model and the neutropenia mice models that are induced either by radiation or
chemicals can be used to test the G-CSF-Tf fusion protein’s therapeutic effects. Those
disease animal models can give more direct proof of the effect of the fusion protein
and possibly provide deeper insights into the mechanism.
150
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2. Can the bioactivity and their therapeutic efficacy o f the G-CSF-Tf
fusion protein be further improved?
For the long term goal of achieving a clinically useful oral protein drug, we need to
make further improvements on the G-CSF-Tf fusion protein. There are still some
limitations of the G-CSF-Tf fusion proteins for the therapeutc uses. First, the in vitro
activity of G-CSF is decreased in the fusion proteins with or without the linker.
Furthermore, the TfR binding affinity was not improved even with the linker insertion.
Second, as we have observed, the G-CSF-Tf fusion protein is not as stable as Tf native
protein. Tf is a very stable protein, with a serum half-life of over 24 hours. In mice,
thereas the recombinant G-CSF is highly hydrophobic with low solubility. The
concentration of Neupogen, the injection formulation, is 300 mcg/mL. After frozen
the activity of recombinant G-CSF lost about 70%. The fusion protein was also found
to lose a lot of activity after being frozen and thawed. Therefore, a proper formulation
is needed to prolong the shelf-life of the fusion protein. To achieve enhanced
therapeutic effect, we can consider improving the formulation of the fusion protein.
For example, the fusion protein could be given together with proteinase inhibitors,
which could increase the fusion protein stability, or transcytosis enhancers (AG 10,
BFA), which can increase the absorption of the fusion protein.
151
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Based on the prolonged ANC stimulation effect, the fusion protein with the helix
linker, likely still worked as a whole protein after crossing the GI epithelium. The Tf
moiety hampers the G-CSF moiety to achieve 100% activity as shown in in vitro
results. In addition, for protein/peptide drugs that act on specific action sites, the Tf
moiety may also affect the biodistribution of the whole protein, causing less amount of
protein drug reach the action site. If the fusion protein can be cleaved at the linker
between two protein domains after entering the blood stream, it is likely that the
therapeutic protein can achieve higher therapeutic activity, and tissue targeting.
Conceivably, fusion proteins with cleavable linkers can be designed. For example,the
linker can contain the specific amino acid sequence recognized by a protease, such as
Factor Xa(Guama et al. 2000), enterokinase, thrombin and tobacco etch virus protease
(TEVP)(Shih et al. 2005). A fusion protein with an inactivated protease at one end,
which could be activated by the enzyme in the liver or during the process across GI
epithelium, can be cleaved at the linker between Tf and protein drug. The released
protein drug can then achieve optimal activity.
The expression of the fusion protein could be modified. Currently, the fusion protein
is expressed by mammalian cells after transient transfection, which is tedious and
costly. The stable transfected cell line has been developed. However, another protein
besides the fusion protein was also secreted and caused problem for purification. A
152
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
better purification procedure can be designed to isolate the fusion protein from the
mixture. Also, it is possible that the Tf-fusion protein can be expressed by insect cells
(Lim et al. 2004) or yeast system (Bewley et al. 1999), in which it was already
reported that T f could be successfully expressed.
3. Can the fusion protein strategy be applied to other therapeutic
protein/peptide drugs?
The concept of oral protein drug delivery through TfR mediated transcytosis can be
applied to other protein or peptide drugs. Not all the protein and peptide drugs are
suitable for TfR mediated oral delivery. First, the therapeutic index should be high
enough, since the oral absorption of protein drug is irregular compared to injection and
is affected by many factors, such as the effect of food or concomitant medication.
(The therapeutic index of a drug is defined as the ratio of the toxic dose to the
therapeutic dose. It is large when the toxic dose is much higher than the therapeutic
dose.) Second, single chained peptides and proteins are easier to be developed into
recombinant fusion proteins, for example G-CSF and nerve growth factor (McGrath et
al. 1997; Park et al. 1998). On the other hand, insulin, which is formed by two chains,
is more difficult to be developed into a fusion protein. In this case, the post-expression
modification may be acquired.
153
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The Tf protein has a molecular weight of 80 kD. The over-sized fusion protein will be
difficult to be expressed and stabilized. Efforts have been made to mimic the TfR
binding with short peptides. Until now, no promising results have been reported yet.
However, with better understanding of the binding between T f and TfR (It was
believed only one lobe of Tf is involved with TfR binding), it is possible to design
proteins or peptides that can bind to TfR and be internalized as a complex with TfR,
with a smaller molecular weight. This approach will greatly help for the application of
the TfR-mediated fusion proteins as therapeutic drugs in the future.
154
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
References
Aisen P. (1984) Transferrin metabolism and the liver. Semin Liver Dis 4: 193-206.
Ali S. A., Joao H. C., Hammerschmid F., Eder J. & Steinkasserer A. (1999)
Transferrin trojan horses as a rational approach for the biological delivery of
therapeutic peptide domains. J Biol Chem 274: 24066-24073.
Andrews N. C. (1999) Disorders of iron metabolism. N Engl J Med 341: 1986-1995.
Arai R., Ueda H., Kitayama A., Kamiya N. & Nagamune T. (2001) Design of the
linkers which effectively separate domains of a bifunctional fusion protein. Protein
Eng 14: 529-532.
Argos P. (1990) An investigation of oligopeptides linking domains in protein tertiary
structures and possible candidates for general gene fusion. J Mol Biol 211: 943-958.
Azari P. R. & Feeney R. E. (1958) Resistance of metal complexes of conalbumin and
transferrin to proteolysis and to thermal denaturation. J Biol Chem 232: 293-302.
Baker H. M., He Q. Y., Briggs S. K., Mason A. B. & Baker E. N. (2003) Structural
and functional consequences of binding site mutations in transferrin: crystal structures
of the Asp63Glu and Argl24Ala mutants of the N-lobe of human transferrin.
Biochemistry 42: 7084-7089.
Bejaoui N., Page M. & Noel C. (1991) Cytotoxicity of transferrin-daunorubicin
conjugates on small cell carcinoma of the lung (SCCL) cell line NCI-H69. Anticancer
Res 11:2211-2213.
Beutler E., Gelbart T., Lee P., Trevino R., Fernandez M. A. & Fairbanks V. F. (2000)
Molecular characterization of a case of atransferrinemia. Blood 96: 4071-4074.
Boneberg E. M. & Hartung T. (2002) Molecular aspects of anti-inflammatory action
of G-CSF. Inflamm Res 51: 119-128.
Brock D. W. (1985) Taking human life. Ethics 95: 851-865.
Brown J. P., Hewick R. M., Hellstrom I., Hellstrom K. E., Doolittle R. F. & Dreyer W.
J. (1982) Human melanoma-associated antigen p97 is structurally and functionally
related to transferrin. Nature 296: 171-173.
155
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Cheng Y., Zak O., Aisen P., Harrison S. C. & Walz T. (2004) Structure of the human
transferrin receptor-transferrin complex. Cell 116: 565-576.
Chu L. & Robinson D. K. (2001) Industrial choices for protein production by large-
scale cell culture. Curr Opin Biotechnol 12: 180-187.
Crasto C. J. & Feng J. A. (2000) LINKER: a program to generate linker sequences for
fusion proteins. Protein Eng 13: 309-312.
Demetri G. D. & Griffin J. D. (1991) Granulocyte colony-stimulating factor and its
receptor. Blood 78: 2791-2808.
Demeule M., Poirier J., Jodoin J., Bertrand Y., Desrosiers R. R., Dagenais C., Nguyen
T., Lanthier J., Gabathuler R., Kennard M., Jefferies W. A., Karkan D., Tsai S., Fenart
L., Cecchelli R. & Beliveau R. (2002) High transcytosis of melanotransferrin (P97)
across the blood-brain barrier. J Neurochem 83: 924-933.
Dupic F., Fruchon S., Bensaid M., Loreal O., Brissot P., Borot N., Roth M. P. &
Coppin H. (2002) Duodenal mRNA expression of iron related genes in response to
iron loading and iron deficiency in four strains of mice. Gut 51: 648-653.
Fishman J. B., Rubin J. B., Handrahan J. V., Connor J. R. & Fine R. E. (1987)
Receptor-mediated transcytosis o f transferrin across the blood-brain barrier. J
Neurosci Res 18: 299-304.
Fitches E., Edwards M. G., Mee C., Grishin E., Gatehouse A. M., Edwards J. P. &
Gatehouse J. A. (2004) Fusion proteins containing insect-specific toxins as pest
control agents: snowdrop lectin delivers fused insecticidal spider venom toxin to insect
haemolymph following oral ingestion. J Insect Physiol 50: 61-71.
Fleming M. D. & Andrews N. C. (1998) Mammalian iron transport: an unexpected
link between metal homeostasis and host defense. J Lab Clin Med 132: 464-468.
Fleming R. E., Migas M. C., Holden C. C., Waheed A., Britton R. S., Tomatsu S.,
Bacon B. R. & Sly W. S. (2000) Transferrin receptor 2: continued expression in
mouse liver in the face of iron overload and in hereditary hemochromatosis. Proc Natl
Acad Sci U S A 97: 2214-2219.
Fukunaga R., Ishizaka-Ikeda E. & Nagata S. (1993) Growth and differentiation signals
mediated by different regions in the cytoplasmic domain of granulocyte colony-
stimulating factor receptor. Cell 74: 1079-1087.
156
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Girelli D., Bozzini C., Roetto A., Alberti F., Daraio F., Colombari R., Olivieri O.,
Corrocher R. & Camaschella C. (2002) Clinical and pathologic findings in
hemochromatosis type 3 due to a novel mutation in transferrin receptor 2 gene.
Gastroenterology 122: 1295-1302.
Goldberg M. & Gomez-Orellana I. (2003) Challenges for the oral delivery of
macromolecules. Nat Rev Drug Discov 2: 289-295.
Gomme P. T., McCann K. B. & Bertolini J. (2005) Transferrin: structure, function and
potential therapeutic actions. Drug Discov Today 10: 267-273.
Gosse M. E., DiMasi J. A. & Nelson T. F. (1996) Recombinant protein and therapeutic
monoclonal antibody drug development in the United States from 1980 to 1994. Clin
Pharmacol Ther 60: 608-618.
Halpem W., Riccobene T. A., Agostini H., Baker K., Stolow D., Gu M. L., Hirsch J.,
Mahoney A., Carrell J., Boyd E. & Grzegorzewski K. J. (2002) Albugranin, a
recombinant human granulocyte colony stimulating factor (G-CSF) genetically fused
to recombinant human albumin induces prolonged myelopoietic effects in mice and
monkeys. Pharm Res 19: 1720-1729.
Hartung T., Von Aulock S., Schneider C. & Faist E. (2003) How to leverage an
endogenous immune defense mechanism: the example of granulocyte colony-
stimulating factor. Crit Care Med 31: S65-75.
Hemadi M., Kahn P. H., Miquel G. & El Hage Chahine J. M. (2004) Transferrin's
mechanism of interaction with receptor 1. Biochemistry 43: 1736-1745.
Hestdal K., Welte K., Lie S. O., Keller J. R., Ruscetti F. W. & Abrahamsen T. G.
(1993) Severe congenital neutropenia: abnormal growth and differentiation of myeloid
progenitors to granulocyte colony-stimulating factor (G-CSF) but normal response to
G-CSF plus stem cell factor. Blood 82: 2991-2997.
Hughson E. J. & Hopkins C. R. (1990) Endocytic pathways in polarized Caco-2 cells:
identification o f an endosomal compartment accessible from both apical and
basolateral surfaces. J Cell Biol 110: 337-348.
Huston J. S., Levinson D., Mudgett-Hunter M., Tai M. S., Novotny J., Margolies M.
N., Ridge R. J., Bruccoleri R. E., Haber E., Crea R. & et al. (1988) Protein engineering
of antibody binding sites: recovery of specific activity in an anti-digoxin single-chain
Fv analogue produced in Escherichia coli. Proc Natl Acad Sci U S A 85: 5879-5883.
157
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Iinuma H., Maruyama K., Okinaga K., Sasaki K., Sekine T., Ishida O., Ogiwara N.,
Johkura K. & Yonemura Y. (2002) Intracellular targeting therapy of cisplatin-
encapsulated transferrin-polyethylene glycol liposome on peritoneal dissemination of
gastric cancer. Int J Cancer 99: 130-137.
Ishiguro A., Nakahata T., Shimbo T., Amano Y., Yasui K., Koike K. & Komiyama A.
(1993) Improvement of neutropenia and neutrophil dysfunction by granulocyte
colony-stimulating factor in a patient with glycogen storage disease type lb. Eur J
Pediatr 152: 18-20.
Jing S. Q. & Trowbridge I. S. (1987) Identification of the intermolecular disulfide
bonds of the human transferrin receptor and its lipid-attachment site. Embo J 6: 327-
331.
Kawabata H., Nakamaki T., Ikonomi P., Smith R. D., Germain R. S. & Koeffler H. P.
(2001) Expression of transferrin receptor 2 in normal and neoplastic hematopoietic
cells. Blood 98: 2714-2719.
Komfeld S. B., Leonard J. E., Mullen M. D. & Taetle R. (1991) Assessment of ligand
effects in intracellular trafficking of ricin A chain using anti-ricin hybridomas. Cancer
Res 51: 1689-1693.
Kratz F., Beyer U., Roth T., Tarasova N., Collery P., Lechenault F., Cazabat A.,
Schumacher P., Unger C. & Falken U. (1998) Transferrin conjugates of doxorubicin:
synthesis, characterization, cellular uptake, and in vitro efficacy. J Pharm Sci 87: 338-
346.
Kratz F., Roth T., Fichiner I., Schumacher P., Fiebig H. H. & Unger C. (2000) In vitro
and in vivo efficacy of acid-sensitive transferrin and albumin doxorubicin conjugates
in a human xenograft panel and in the MDA-MB-435 mamma carcinoma model. J
Drug Target 8: 305-318.
Laemmli U. K. (1970) Cleavage of structural proteins during the assembly of the head
of bacteriophage T4. Nature 227: 680-685.
Lebron J. A. & Bjorkman P. J. (1999) The transferrin receptor binding site on HFE,
the class I MHC-related protein mutated in hereditary hemochromatosis. J Mol Biol
289: 1109-1118.
Li H. & Qian Z. M. (2002) Transferrin/transferrin receptor-mediated drug delivery.
Med Res Rev 22: 225-250.
158
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Lim C. J. & Shen W. C. (2005) Comparison of monomeric and oligomeric transferrin
as potential carrier in oral delivery of protein drugs. J Control Release 106: 273-286.
Lord B. I., Woolford L. B. & Molineux G. (2001) Kinetics of neutrophil production in
normal and neutropenic animals during the response to filgrastim (r-metHu G-CSF) or
filgrastim SD/01 (PEG-r-metHu G-CSF). Clin Cancer Res 7: 2085-2090.
Lyman G. H., Kuderer N. M. & Djulbegovic B. (2002) Prophylactic granulocyte
colony-stimulating factor in patients receiving dose-intensive cancer chemotherapy: a
meta-analysis. Am J Med 112: 406-411.
Ma Y., Specian R. D., Yeh K. Y., Yeh M., Rodriguez-Paris J. & Glass J. (2002) The
transcytosis of divalent metal transporter 1 and apo-transferrin during iron uptake in
intestinal epithelium. Am J Physiol Gastrointest Liver Physiol 283: G965-974.
Maeda Y., Ueda FI., Kazami J., Kawano G., Suzuki E. & Nagamune T. (1997)
Engineering of functional chimeric protein G-Vargula luciferase. Anal Biochem 249:
147-152.
McCullough J., Clay M., Elerr G., Smith J. & Stroncek D. (1999) Effects of
granulocyte-colony-stimulating factor on potential normal granulocyte donors.
Transfusion 39: 1136-1140.
Meissner P., Pick H., Kulangara A., Chatellard P., Friedrich K. & Wurm F. M. (2001)
Transient gene expression: recombinant protein production with suspension-adapted
HEK293-EBNA cells. Biotechnol Bioeng 75: 197-203.
Morgan E. H. (1981) Inhibition of reticulocyte iron uptake by NH4C1 and CFF3NH2.
Biochim Biophys Acta 642: 119-134.
Nicola N. A., Metcalf D., Matsumoto M. & Johnson G. R. (1983) Purification of a
factor inducing differentiation in murine m yelomonocytic leukem ia cells.
Identification as granulocyte colony-stimulating factor. J Biol Chem 258: 9017-9023.
Park E., Starzyk R. M., McGrath J. P., Lee T., George J., Schutz A. J., Lynch P. &
Putney S. D. (1998) Production and characterization of fusion proteins containing
transferrin and nerve growth factor. J Drug Target 6: 53-64.
Prydz K., Hansen S. H., Sandvig K. & van Deurs B. (1992) Effects of brefeldin A on
endocytosis, transcytosis and transport to the Golgi complex in polarized MDCK cells.
J Cell Biol 119: 259-272.
159
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Qian Z. M., Li H., Sun H. & Ho K. (2002) Targeted drug delivery via the transferrin
receptor-mediated endocytosis pathway. Pharmacol Rev 54: 561-587.
Richardson D. R. (2004) Mysteries of the transferrin-transferrin receptor 1 interaction
uncovered. Cell 116: 483-485.
Richardson D. R. & Ponka P. (1997) The molecular mechanisms of the metabolism
and transport of iron in normal and neoplastic cells. Biochim Biophys Acta 1331: 1-
40.
Robinson C. R. & Sauer R. T. (1998) Optimizing the stability of single-chain proteins
by linker length and composition mutagenesis. Proc Natl Acad Sci U S A 95: 5929-
5934.
Russmann H., Igwe E. I., Sauer J., Hardt W. D., Bubert A. & Geginat G. (2001)
Protection against murine listeriosis by oral vaccination with recombinant Salmonella
expressing hybrid Yersinia type III proteins. J Immunol 167: 357-365.
Shah D. & Shen W. C. (1994) The establishment of polarity and enhanced transcytosis
of transferrin receptors in enterocyte-like Caco-2 cells. J Drug Target 2: 93-99.
Shah D. & Shen W. C. (1996) Transcellular delivery of an insulin-transferrin
conjugate in enterocyte-like Caco-2 cells. J Pharm Sci 85: 1306-1311.
Shimazaki C., Uchiyama H., Fujita N., Araki S., Sudo Y., Yamagata N., Ashihara E.,
Goto H., Inaba T., Haruyama H. & et al. (1995) Serum levels of endogenous and
exogenous granulocyte colony-stimulating factor after autologous blood stem cell
transplantation. Exp Hematol 23: 1497-1502.
Singh V. & Curtiss R., 3rd (1994) Effect of amino group modification of ovine
luteinizing hormone (oLH) by N-succinimidyl 6-[3-(2-
pyridyldithio)propionate]hexanoate, a long chain N-succinimidyl-3-(2-pyridyldithio)
propionate (SPDP) on immunological and biological properties: a comparative study
with SPDP modified oLH. Mol Cell Biochem 130: 83-90.
Souza L. M., Boone T. C., Gabrilove J., Lai P. H., Zsebo K. M., Murdock D. C.,
Chazin V. R., Bruszewski J., Lu H., Chen K. K. & et al. (1986) Recombinant human
granulocyte colony-stimulating factor: effects on normal and leukemic myeloid cells.
Science 232: 61-65.
Stanislawska-Sachadyn A., Sachadyn P., Ihle K., Sydorczuk C., Wiejacha K. & Kur J.
(2006) The construction of bifunctional fusion proteins consisting of MutS and GFP. J
Biotechnol 121: 134-143.
160
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Swaan P. W. (1998) Recent advances in intestinal macromolecular drug delivery via
receptor-mediated transport pathways. Pharm Res 15: 826-834.
Tamura M., Hattori K., Nomura H., Oheda M., Kubota N., Imazeki I., Ono M.,
Ueyama Y., Nagata S., Shirafuji N. & et al. (1987) Induction of neutrophilic
granulocytosis in mice by administration of purified human native granulocyte colony-
stimulating factor (G-CSF). Biochem Biophys Res Commun 142: 454-460.
Thorstensen K. & Romslo I. (1990) The role of transferrin in the mechanism of
cellular iron uptake. Biochem J 271: 1-9.
Tkatch L. S. & Tweardy D. J. (1993) Human granulocyte colony-stimulating factor
(G-CSF), the premier granulopoietin: biology, clinical utility, and receptor structure
and function. Lymphokine Cytokine Res 12: 477-488.
Trinh R., Gurbaxani B., Morrison S. L. & Seyfzadeh M. (2004) Optimization of codon
pair use within the (GGGGS)3 linker sequence results in enhanced protein expression.
Mol Immunol 40: 717-722.
Ulich T. R., del Castillo J., Guo K. & Souza L. (1989) The hematologic effects of
chronic administration of the monokines tumor necrosis factor, interleukin-1, and
granulocyte-colony stimulating factor on bone marrow and circulation. Am J Pathol
134: 149-159.
Wan J., Taub M. E., Shah D. & Shen W. C. (1992) Brefeldin A enhances receptor-
mediated transcytosis of transferrin in filter-grown Madin-Darby canine kidney cells. J
Biol Chem 267: 13446-13450.
Ward A. C., Smith L., de Koning J. P., van Aesch Y. & Touw I. P. (1999) Multiple
signals mediate proliferation, differentiation, and survival from the granulocyte
colony-stimulating factor receptor in myeloid 32D cells. J Biol Chem 274: 14956-
14962.
Watts M. J., Addison I., Long S. G., Hartley S., Warrington S., Boyce M. & Linch D.
C. (1997) Crossover study of the haematological effects and pharmacokinetics of
glycosylated and non-glycosylated G-CSF in healthy volunteers. Br J Haematol 98:
474-479.
Widera A., Bai Y. & Shen W. C. (2004) The transepithelial transport of a G-CSF-
transferrin conjugate in Caco-2 cells and its myelopoietic effect in BDF1 mice. Pharm
Res 21: 278-284.
161
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Widera A., Beloussow K., Kim K. J., Crandall E. D. & Shen W. C. (2003a)
Phenotype-dependent synthesis of transferrin receptor in rat alveolar epithelial cell
monolayers. Cell Tissue Res 312: 313-318.
Widera A., Kim K. J., Crandall E. D. & Shen W. C. (2003b) Transcytosis of GCSF-
transferrin across rat alveolar epithelial cell monolayers. Pharm Res 20: 1231-1238.
Widera A., Norouziyan F. & Shen W. C. (2003c) Mechanisms o f TfR-mediated
transcytosis and sorting in epithelial cells and applications toward drug delivery. Adv
Drug Deliv Rev 55: 1439-1466.
Xia C. Q. & Shen W. C. (2001) Tyrphostin-8 enhances transferrin receptor-mediated
transcytosis in Caco-2- cells and inreases hypoglycemic effect of orally administered
insulin-transferrin conjugate in diabetic rats. Pharm Res 18: 191-195.
Xia C. Q., Wang J. & Shen W. C. (2000) Hypoglycemic effect of insulin-transferrin
conjugate in streptozotocin-induced diabetic rats. J Pharmacol Exp Ther 295: 594-600.
Young D. C., Zhan H., Cheng Q. L., Hou J. & Matthews D. J. (1997) Characterization
of the receptor binding determinants of granulocyte colony stimulating factor. Protein
Sci 6: 1228-1236.
162
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Transferrin and its oligomer: Intracellular processing and applications as drug carriers

PDF

Recombinant arginine deiminase (rADI) and arginine metabolism

PDF

Trans-epithelial transport of G -CSF -transferrin conjugate

PDF

Oxidation induced changes in conjunctival ion transport: Critical role of conjunctival GSH metabolism and transport in ocular surface health

PDF

Drug delivery to the posterior region of the eye: Characterization of organic cation transporters

PDF

Studies on the mechanism and regulation of the transport of 5-fluorouracil (5FU) into tumor cells

PDF

Design of a stealth liposome delivery system for a novel glycinamide ribonucleotide formyltransferase inhibitor

PDF

Time dependent survival analysis of Kaiser Permanente/USC pharmacists' consultation intervention study

PDF

Synthesis of carboplatin and studies on its biotransformation products

PDF

Effects of a formulary expansion of the use of SSRIs and health care services by depressed patients in the California Medicaid program

PDF

Endogenous regulatory factors in the inhibition and down-regulation of immune responses

PDF

Enhancement of transferrin receptor-mediated transcytosis: a novel approach for macromolecular drug delivery

PDF

Physician adherence to national hypertension guidelines in an elderly Medicaid population

PDF

The construct validity of the Health Utilities Index in patients with chronic respiratory disease in a managed care population

PDF

Noninvasive in vivo MRI measurements of tumor vascularization and relative tumor osmotic pressure

PDF

Assessing the cost implications of combined pharmacotherapy in the long term management of asthma: Theory and application of methods to control selection bias

PDF

Structure-function studies on the mammalian intestinal dipeptide transporter hPEPT1

PDF

Preferential translation of Drosophila heat shock protein 70 and heat shock protein 90 mRNAs during heat shock

PDF

Conjunctival epithelial uptake of biodegradable nanoparticles: Mechanism, intracellular distribution, and absorption enhancement

PDF

Insulin-transferrin conjugate as an oral hypoglycemic agent

Asset Metadata

Creator Bai, Yun (author)

Core Title TfR-mediated oral delivery of protein drugs: Oral delivery of recombinant G -CSF -Tf fusion protein and its spacer optimization

Degree Doctor of Philosophy

Degree Program Pharmaceutical Sciences

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Health Sciences, Pharmacy,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-621574

Unique identifier UC11336551

Identifier 3233786.pdf (filename),usctheses-c16-621574 (legacy record id)

Legacy Identifier 3233786.pdf

Dmrecord 621574

Document Type Dissertation

Rights Bai, Yun

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA