Download Induction of antibody response to human tumor antigens by

Gene Therapy (1997) 4, 969–976
 1997 Stockton Press All rights reserved 0969-7128/97 $12.00
Induction of antibody response to human tumor
antigens by gene therapy using a fusigenic viral
liposome vaccine
T Okamoto1, Y Kaneda2, D Yuzuki1, SKS Huang1, DDJ Chi1 and DSB Hoon1
1
Department of Molecular Oncology, John Wayne Cancer Institute, Santa Monica, CA, USA; and 2Institute for Molecular and
Cellular Biology, Osaka University, Osaka 565, Japan
Development of effective cancer vaccines would help prevent and control tumor progression. A novel approach of
immunizing against tumor antigens is in vivo gene vaccination. We have developed a fusigenic viral liposome vector using HVJ (hemagglutinating virus of Japan) and liposome to deliver human tumor antigen genes effectively to
cells in vivo. Plasmids containing the human tumor antigen
genes MAGE-1 and MAGE-3 were encapsulated in fusigenic viral liposomes and injected into mice intramuscularly. MAGE-1 and -3 recombinant proteins were used in
Western blotting and affinity ELISA for assessment of antibody responses. Mice immunized with MAGE-1 and -3
gene vaccine individually were shown to produce anti-
MAGE-1 and -3 IgG antibody responses, respectively. Animals immunized with plasmid alone did not induce antiMAGE-1 or -3 IgG responses. Antibody responses could
be enhanced on reimmunization with the gene vaccines.
Muscle biopsies taken after vaccine injection were verified
to express gene-specific mRNA transcripts. Mice immunized with MAGE-1 or -3 gene vaccines were shown to
induce antibodies that could cross-react with the respective
recombinant proteins. This study demonstrates that in vivo
immunization using HVJ–liposome containing human
tumor antigen genes can effectively deliver and induce
immune responses to the respective whole proteins.
Keywords: vaccine; cancer; antibody; liposome; gene therapy
Introduction
DNA vaccines have recently been shown to be a promising approach for immunization against a variety of infectious diseases.1–3 Delivery of naked DNAs containing
microbial antigen genes can induce antigen-specific
immune responses in the host. The induction of antigenspecific immune responses using DNA-based vaccines
has shown some promising effects.4 Overall the DNAbased vaccination principle is similar to live attenuated
virus immunization however with a greater degree of
control of antigen expression, toxicity and pathogenicity,
although improvements in in vivo delivery and transgene
expression are needed. The majority of studies reported
to date on DNA immunization are with microbial antigen
genes that are known to be immunogenic in animals or
humans. DNA immunization strategy may be useful to
develop new effective cancer vaccines. In general, studies
on tumor antigen protein/peptide vaccines in human
and animal studies to date have been encouraging but
not as efficient as microbial antigen vaccines.5 Human
tumor antigens are in general weakly immunogenic. The
method of antigen presentation and the route of antigen
delivery are important to induce effective immunity as
well.
Correspondence: DSB Hoon, Department of Molecular Oncology, John
Wayne Cancer Institute, 2200 Santa Monica Boulevard, Santa Monica,
CA 90404, USA
Received 21 February 1997; accepted 8 May 1997
At present there are multiple approaches being
assessed in immunizing cancer patients with tumor antigens.5 Purification of natural or recombinant tumor antigens is tedious and not always logistically practical. To
date there are only a few human cancer vaccines showing
consistent and significant effective augmentation of
immunity with soluble antigens (proteins/peptides). This
problem is partially due to antigen presentation and host
major histocompatibility complex (MHC) polymorphism.
Studies to date suggest that cellular cancer vaccines
expressing the antigens may be more effective than soluble antigens alone.5 This may be due to tumor antigen
expression, processing and presentation by the cell.
Using the host normal cells (nonhemopoietic) to express
and present the tumor antigens to the immune system
may be an effective immunization strategy. In general,
when a cytopathic virus infects a host normal cell, the
viral proteins are endogenously processed and presented
on the cell surface, or in fragments by MHC molecules.
Foreign defined DNA transfected and expressed by normal cells can mimic viral infections.
Although in vivo DNA vaccination protocols are available, there are still problems in effective in vivo delivery
of DNA to the cells and expression of genes. One of the
problems is introduction of DNA into specific cells without degradation of the DNA by endosomes or lysosomes.6 Current approaches of in vivo delivery of DNA
by retroviral or adenoviral vectors have problems related
to efficacy, viral gene integration, potential pathogenic
activity and immune response to viral vector encoding
Human tumor antigen gene vaccination
T Okamoto et al
970
proteins. To get around some of these problems, we have
developed an in vivo delivery system using a nonviral
DNA vector. The delivery system also includes protein
encapsulated in a fusigenic viral liposome that is a hybrid
vector of viral and nonviral reagents. Briefly, anionic liposomes are fused with HVJ (Sendai virus member of paramyxovirus family).7–10 HVJ proteins are known for their
unique fusing properties to cell membranes of nonhemopoietic nucleated cells.11,12 This hybrid vector allows targeting to normal nonhemopoietic cells and delivery of
encapsulated DNA into the cytoplasm of cells without
lysosome/endosome degradation. Another advantage of
the fusigenic viral liposome system is that it can deliver
DNA and protein molecules together. Most gene transfer
methods are limited to the low level of expression of the
transfected gene.8,11 To facilitate gene transcription, we
incorporated a high mobility group-1 protein (HMG)
which is a non-histone DNA-binding protein. HMG-1
protein enhances gene transcription. This delivery system
has been demonstrated to be highly effective in in vivo
gene transfer and delivery of genes and antisense DNA
oligomers.13,14
Our objective was to utilize the HVJ–liposome DNAmediated transfer technique in vivo to immunize against
human tumor antigens MAGE-1 and MAGE-3. The
MAGE-1 and -3 antigens were first described in melanoma and subsequently demonstrated in various other
cancers. MAGE-1 and -3 genes are expressed in .30% of
melanomas and carcinomas such as lung, breast, liver
and gastrointestinal cancers,15–17 but not in normal tissues
except testes. The MAGE-1 and -3 antigens were assessed
to study tumor antigen DNA vaccination primarily
because they have been shown to be immunogenic,
expressed by a wide variety of human cancers and not
expressed by normal tissues.15–17 These are important criteria in the overall design of an effective vaccine against
multiple cancers. Studies have demonstrated that the
MAGE-1 and MAGE-3 proteins are immunogenic in
which antigen-specific cell-mediated and humoral
immune responses have been demonstrated in
humans.14,16 However the MAGE-1 and -3 are generally
weakly immunogenic. Previously, we have demonstrated
that anti-MAGE-116 and -3 (Huang et al, submitted) IgG
antibodies responses can be induced in melanoma
patients immunized with a melanoma cell vaccine. In this
study we demonstrate that by immunization with HVJ–
liposome containing MAGE-1 and -3 genes, anti-MAGE1 or -3 IgG antibodies could be induced. The antibody
responses could be enhanced by reimmunization and
maintained for several weeks.
Results
Production of HVJ–liposomes
A schematic diagram of the production of HVJ–liposomes encapsulated with plasmid pcDNA3 containing
MAGE-1 (pMAGE-1) or MAGE-3 (pMAGE-3) and HMG1 protein is shown in Figure 1. In the study we will refer
to the HVJ–liposome complex containing pMAGE-1 and
pMAGE-3 plus HMG-1 protein as MAGE-1 or -3 vaccine,
respectively. Preparation of HVJ–liposomes with plasmid
DNA was carried out on the day or within 24 h of injection into animals. The starting amount of plasmid DNA
used in liposome encapsulation was 100 mg per tube. The
Figure 1 DNA immunization protocol: schematic diagram of HVJ–liposomes plus plasmid DNA preparation and immunization of mice.
amount of encapsulated plasmid DNA injected per animal was approximately 7–10 mg. This dose was the most
efficient amount of HVJ–liposome DNA complex needed
for in vivo gene expression. This was also the most logistically feasible dose of vector to produce a significant
rapid and consistent immune response to the MAGE proteins. Optimization and control experiments with individual components of the fusigenic viral liposome for in
vivo delivery of genes has been previously reported.8,9,11
Immunization of mice immediately after HVJ–liposome
preparation was performed to obtain optimal results.
Through experimentation an optimal vaccine immunization schedule was determined.
No necrosis, inflammation or visible injury was
observed after multiple injections at the same site.
Groups of animals injected with plasmid DNA were kept
for over 12 months and no visible injury or change in
physical activity was observed compared with nonimmunized animals. HVJ–liposome solution did not
affect the well-being of the animals or hind leg function
(site of injections).
Detection of anti-MAGE-1 and MAGE-3 antibody by
Western blot
To assess the induction of anti-MAGE-1 and -3-specific
antibody responses after vaccination recombinant human
MAGE-1 (rMAGE-1) and recombinant human MAGE-3
(rMAGE-3) were produced and purified for immunoassays (Figure 2). MAGE-1 protein has 309 amino acids
with a predicted molecular weight of 34 kDa. The purified rMAGE-1 run on SDS-PAGE under reducing conditions was approximately 42 kDa. MAGE-3 protein has
314 amino acids with a predicted molecular weight of
35 kDa. The purified rMAGE-3 run on SDS-PAGE under
reducing conditions was approximately 48 kDa. The
possible reasons for the differences in protein calibrated
weight by gel electrophoresis versus predicted weight has
been previously discussed.16 Both proteins have
additional amino acids used as affinity tags. Whole protein was essential to assess the overall immune response
since antigenic peptide sequences in mice are not known.
Western blots with rMAGE-1 and rMAGE-3 were performed with sera of mice immunized at least twice to
verify specificity of antibody responses (Figure 2). No
antibody (IgG) responses to rMAGE-1 or -3 were detected
Human tumor antigen gene vaccination
T Okamoto et al
971
Figure 2 Analysis of rMAGE proteins and representative antibody
responses. Purified rMAGE-1 and rMAGE-3 run under reducing condition in a 10% SDS-PAGE and stained with coomassie blue (left side).
Lanes: A, rMAGE-1; B, rMAGE-3; M, molecular weight standards (kDa).
Analysis of antibody response to rMAGE-1 (a) and rMAGE-3 (b) by
Western blot (right side). Representative experiments of analysis of antiMAGE-1 IgG and anti-MAGE-3 IgG in sera. Lanes 1 and 2, analysis of
non-immunized mice, respectively. For MAGE-1 vaccine immunization
(a): lanes 3 and 4, sera from mice immunized three times biweekly with
MAGE-1 vaccine. For MAGE-3 vaccination (b): lanes 3 and 4, sera from
mice immunized three times biweekly with MAGE-3 vaccine. Representative serum analysis for two mice for each group at 1:100 dilution is shown.
in non-immunized mice. Mice immunized several times
with HVJ–liposome complex containing pcDNA3 without a MAGE-1 or -3 gene insert also had no antigenspecific antibody response. Naked DNA given in vivo
and taken up by cells degrades rapidly in
lysosomes/endosomes (data not shown). Mice immunized with MAGE-1 vaccine three times and assessed 6
weeks after immunization showed an induction of a
strong antibody response to rMAGE-1 (Figure 2a). Similarly mice immunized with MAGE-3 vaccine three times
and followed-up after 6 weeks showed a strong antibody
response to rMAGE-3 (Figure 2b). These studies indicated that antigen-specific IgG responses could be
induced rapidly and maintained for at least 8 weeks after
immunization. IgM anti-MAGE-1 and -3 responses were
also assessed after various numbers of immunization.
The level of response was weak or non-detectable in sera
tested; therefore we did not pursue further analysis of
IgM responses.
Detection of anti-MAGE-1 and MAGE-3 IgG antibody by
ELISA
Anti-MAGE-1 IgG antibodies were detected after
immunization at various time-points by a MAGE-1 affinity ELISA. The selective affinity binding of Ni2+ and hexaHis tagged protein insured higher specificity and less
steric conformation variation as opposed to conventional
absorbance of proteins to ELISA plates. A representative
experiment of antibody response after MAGE-1 immunization is shown in Figure 3: mice were immunized on
day 0, 7 and 14, and then reimmunized approximately 1
year after. IgG antibody levels were enhanced after 7
days of booster immunization and maintained at day 14.
Figure 3 Representative experiments of serial bleed analysis after MAGE1 vaccine immunization. Affinity ELISA with rMAGE-1 was performed
to detect IgG responses at 1:100 dilution of sera. Representative experiment of mice (four per group) immunized on days 0, 7, 14 and reimmunized 1 year later and assessed 28 days later. Mean absorbency of triplicates
± s.e. of all mice shown at specific time-points. Experiments were performed several times to confirm results.
Anti-MAGE-1 IgG antibody was induced and could be
elevated with subsequent immunizations. Immunized
mice were also assessed at different dilutions of serum
ranging from 1:50 to 1:800. In general, after two immunizations on day 14 IgG anti-MAGE-1 antibody titers were
1/200 to 1/400. A positive antibody dilution was considered when it was 2 s.d. above negative serum background at a dilution of 1:100.
One year after three immunizations antibody levels
went back down to background level as expected. Animals were reimmunized and evaluated after 4 weeks.
Anti-MAGE-1 IgG responses were shown to be boosted
to early immunization levels. These studies indicated that
repeated immunization with the HVJ–liposome system
does not induce any significant inhibiting immune
response to the HVJ–liposome components or plasmid
vector to prevent antigen-specific immunization.
Cross-reactivity between MAGE-1 and MAGE-3
antibodies
Mice were immunized with MAGE-1 or -3 vaccine on
days 0, 14 and 28. The objective of the study was to determine whether anti-MAGE-1 IgG antibodies induced by
MAGE-1 vaccine immunization cross-reacted with
MAGE-3, and whether anti-MAGE-3 IgG antibodies
induced by MAGE-3 vaccine immunization cross-reacted
with MAGE-1. Western blot analyses with rMAGE-1 and
rMAGE-3 were performed with sera from MAGE-1 and
MAGE-3 vaccine immunized mice (Figure 4). The studies
demonstrated cross-reactivity of anti-MAGE-1 and
-MAGE-3 sera with both rMAGE-1 and rMAGE-3.
Further analysis of the cross-reactivity was shown by
ELISA (Figure 5). There were no significant differences
between the antibody responses against MAGE-1 (Figure
5a). However, anti-MAGE-3 IgG activity in MAGE-1
immunized mice was lower than that of MAGE-3 immunized mice (Figure 5b). Although both MAGE-1 and
MAGE-3 have been shown to have distinct antigenic
determinants in humans, and antigenic determinants
have not been defined in the murine system.
Human tumor antigen gene vaccination
T Okamoto et al
972
Figure 4 Analysis of cross-reactivity of anti-MAGE-3 and anti-MAGE1 antibody response to rMAGE-1 and rMAGE-3, respectively by Western
blotting. (a) Lanes 1 and 2 are representative of experiments with mice
immunized three times biweekly with MAGE-3 and assessed for antibody
against rMAGE-1. (b) Lanes 1 and 2 are representative of experiments
with mice immunized three times biweekly with MAGE-1 and assessed
for antibody against rMAGE-3. Representative serum analysis for two
mice for each group at 1:100 dilution is shown.
Detection of MAGE-1 DNA
Muscle biopsies were analyzed for presence and
expression of pcDNA3 and pMAGE-1 after selected periods of immunization. Tissue biopsies at the sites of vaccine injections were dissected out and nucleic acids were
extracted. RT-PCR and DNA PCR plus Southern blots
were performed on RNA and DNA, respectively, from
muscle tissue of immunized mice and non-immunized
mice (Figure 6). Muscle biopsies of mice immunized with
pcDNA3 alone were assessed 7, 14 and 28 days after
immunization. Two mice from each time period were
killed and assessed. A 218 bp RT-PCR and DNA PCR
cDNA product was detected in all animals indicating the
presence of pcDNA3. RT-PCR analysis of muscle biopsies
of mice immunized with MAGE-1 vaccine was carried
out on days 7 (two mice), 14 (three mice) and 28 (two
mice) after immunization. All tissues examined were
shown to express specific MAGE-1 mRNA 7, 14 and 28
days after immunization. MAGE-1 gene expression was
further verified using a separate set of specific primers to
MAGE-1 (RT-PCR cDNA product 920 bp). RT-PCR analysis of muscle biopsies at the site of MAGE-1 vaccine injection also showed the presence of the pcDNA3 plasmid
alone (218 bp). This may be due to incomplete transcription of the MAGE-1 gene insert in pcDNA3 or contamination of pcDNA3 without MAGE-1 gene insert in vector
preparations. This was observed in only a few animals
injected with different HVJ–liposome. Normal muscle
and other body organs (lung, kidney, liver) RNA of nonimmunized mice were evaluated by RT-PCR and Southern blotting and were shown not to express MAGE-1
mRNA.
MAGE-1 DT DNA immunization
A plasmid containing MAGE-1 plus diphtheria toxin
peptide DNA hybrid (MAGE-1 DT) was constructed to
determine if genetically engineering an immunogenic
determinant to MAGE-1 gene product would enhance
antibody responses to MAGE-1. Mice were immunized
Figure 5 Representative experiments of anti-MAGE-1 and anti-MAGE3 IgG after three biweekly immunizations. Affinity ELISA with rMAGE1 (a) and rMAGE-3 (b) was performed with sera at 1:100 dilution. Assessment of mice immunized with MAGE-1 vaccine (P) or MAGE-3 vaccine
(p) is shown. Data are shown as mean absorbency of triplicates ± s.e. of
all mice.
Figure 6 DNA PCR and RT-PCR analysis of mice muscle biopsies after
immunization of pcDNA3 and MAGE-1 vaccine. RT-PCR or DNA PCR
was performed on isolated RNA and DNA from tissues, respectively, followed by Southern blotting with specific cDNA probes. Lanes 1 and 2
represent DNA PCR analysis 14 days after immunization with pcDNA3.
Lanes 3 and 4 represent RT-PCR analysis 14 and 28 days after immunization, respectively, with pcDNA3. Lanes 5, 6 and 7 represent RT-PCR
analysis 7, 14 and 28 days after immunization, respectively, with MAGE1 vaccine. Lane 8 represents DNA PCR of pcDNA3 (standard control;
218 bp). Lane 9 represents DNA PCR of MAGE-1 vaccine (standard
control; 1136 bp). Lane 10 represents RT-PCR of pMAGE-1(standard
control; 1136 bp). These are representative Southern blot analyses of several experiments.
Human tumor antigen gene vaccination
T Okamoto et al
Figure 7 Analysis of antibody response to rMAGE-1 by Western blotting.
Comparison of the anti-MAGE-1 IgG induction after immunization with
MAGE-1 or MAGE-1 DT vaccines. Lanes 1 and 2, sera from nonimmunized mice. Lanes 3 and 4, sera from mice immunized three times biweekly
with MAGE-1 vaccine. Lanes 5 and 6, mice immunized with MAGE-1
DT vaccine. Representative serum analysis for two mice for each group
at 1:100 dilution is shown.
(three times) with MAGE-1 DT vaccine as for MAGE-1
vaccine. The 16mer diphtheria toxin (DT) peptide
sequence was tagged with a his-tag to allow peptide
affinity ELISA to be performed. The antibody response
to rMAGE-1 assessed by Western blot after immunization
with MAGE-1 DT vaccine was not better than MAGE-1
vaccine immunization (Figure 7). In assessment of mice
immunized with MAGE-1 DT by affinity ELISA there
was some enhancement (not significant) of antibody
response to DT peptide above control MAGE-1 vaccine
(Figure 8). Overall, adding a second immunogenic determinant did not improve antibody responses to rMAGE-1.
Discussion
In this study we have demonstrated a novel efficient
approach using a fusigenic viral liposome system to
deliver human tumor antigen genes in vivo to immunize
Figure 8 Representative example of anti-DT peptide IgG detection by
affinity ELISA after three biweekly immunizations with MAGE-1 DT vaccine (l) or MAGE-1 vaccine (P). Mice serum was analyzed as mean
absorbency triplicates ± s.e. of all mice.
the host. Human cancer vaccines being studied today
require culturing tumor cells, purifying tumor antigens,
or producing specific peptide/recombinant proteins. The
HVJ–liposome system avoids many of the tedious processes involved in vaccine preparation. Quality control
and purity of plasmid DNA can be more easily monitored. Gene expression in muscle has been successful by
injection of naked DNA of microbial antigens.4 However,
we found that HVJ–liposome-mediated gene transfer was
30–100 times more efficient in gene expression in skeletal
muscle than naked DNA transfer (unpublished data).
Fusigenic viral liposomes can provide an efficient vehicle
to package, deliver and direct DNA to specific targets
and at the same time protect against nucleic acid degrading enzymes in body fluids and cytoplasmic organelles.8
The delivery system using the HVJ–liposome is safe and
to date there have been no side effects observed in
animals.
Our approach of in vivo gene therapy using the host
normal cells is an effective attenuated form of human
tumor antigen gene vaccination. In this type of situation
tumor antigen is expressed on a background of normal
cell(s) as opposed to a tumor cell. The latter has many
negative type regulating elements that may suppress
immune responses as well as altered physiologic functions which can modify antigen expression. Gene vaccination provides an opportunity for molecular immunophysiologic manipulation of antigen expression that can
be a useful tool in cancer vaccine design. Intramuscular
injection has been shown in the past to be an important
delivery route for induction of immunity. Skeletal muscle
has properties such as vascularization, multi-nucleation,
nonreplicating and capability of expressing recombinant
proteins;18 these offer advantages to intramuscular site
injection for gene therapy. The mechanism of how muscle
presents the protein and induces immune response is still
unknown. One suggestion is that recombinant protein is
produced and released into the vascular network of the
muscle and eventually presented by professional antigenpresenting cells such as dendritic cells or macrophages
infiltrating the muscle.19 Another suggestion is that
at the injection site muscle injury induces myoblast
proliferation and activation of infiltrating macrophages/dendritic-like cells and they then present antigens through MHC class II antigens.18,19 Since IgG
responses were detected the mechanism of antigen presentation is most likely through a T helper2 subset
response and MHC class II antigen presentation.
Analysis of muscle biopsies at the site of immunization
showed that the plasmid DNA was being expressed after
28 days of immunization. The study verified that the
expression of the plasmid DNA was transient and maintained in the cells at the site of injection for at least 1
month. This potentially provides a continual exposure of
antigen to the host immune system as opposed to a single
antigen injection. The advantage of the HVJ–liposome
system over retroviral or adenoviral vectors is that it is
not incorporated into the genome. The inactivated HVJ
virus is safe and induces no pathogenic effect. The other
advantage of the HVJ–liposome system is that it does not
incorporate into hemopoietic-derived cells such as lymphocytes as do other liposome delivery systems.11,12 Also
repetitive immunization did not appear to induce strong
immune responses towards the vector system to inhibit
973
Human tumor antigen gene vaccination
T Okamoto et al
974
gene transfer. This is likely due to the rapid incorporation
of the fusigenic viral liposome into the host cells.
We and others have shown human cancer vaccines are
an effective method of immunization of patients against
tumor antigens to control tumor progression and
improve survival.5,20,21 Although there are many formulations of human cancer vaccines, there have been problems in efficiently inducing specific antitumor antigen
responses. Immunization protocols have not succeeded
very well in consistently inducing anti-MAGE immune
responses in humans. In this study we demonstrate consistent induction of anti-MAGE-1 and -3 antibody
responses shortly after vaccination and maintenance of
activity for several weeks. The antibody titers generated
were as expected for a weak immunogenic tumor antigen. There is a possibility that T cell responses were
induced, however, at this time we do not know the antigenic determinants of the antigens in mice to allow
assessment. Future studies will examine the induction of
MAGE-1 and/or -3-specific T cells. To date, anti-MAGE1 and -3 T cell responses in mice have not been documented. Two highly homologous sequences of MAGE-1,
referred to as Smage1 and Smage2 are expressed in the
murine testes but not normal murine tissue.22 There may
be host immune tolerance to the MAGE family proteins.
Studies reported on generation of anti-MAGE-1 and -3 in
mice by recombinant protein administration with adjuvants have in general not produced high antibody titers.23
We assessed the incorporation of an immunogenic
determinant DT as a ‘helper’ antigen on the MAGE-1
gene to improve its efficacy. Diphtheria toxin B fragment
COOH-terminal region has been shown to be immunogenic in mice.24 The failure to enhance anti-MAGE-1
responses could be due to the lack of full expression of
MAGE-1 DT hybrid protein in vivo, conformation of the
recombinant MAGE-1 DT protein, or lack of recognition
of DT peptide by the host as an immunogenic determinant due to presentation. A low level of response was
observed against DT peptide alone suggesting some
induction of immune response. Alternatively, a T cell
response may have been induced. Further work needs
to be carried out in the development of chimeric gene
constructs for gene vaccination to improve immunogenicity and polyvalency.
Both MAGE-1 and MAGE-3 have been shown to be
immunogenic in humans. There are other MAGE gene
family members with similar homologies.25 The strategy
of using two dominant immunogenic MAGE family antigens may be beneficial in that they can elicit immunity
to a wide spectrum of MAGE antigens expressed by different human tumors. The cross-reactivity of the antiMAGE-1 and MAGE-3 responses demonstrated is not
surprising since the proteins express 66% amino acids
homology.17 Cross-reactivity to MAGE-1 and MAGE-3
has not been previously well documented. The availability of the recombinant MAGE proteins and immunized mice sera has allowed us to assess these events. The
study also indicated that vaccination with MAGE-1 or
-3 could potentially induce immune responses to tumors
expressing either antigen. MAGE-1 and MAGE-3 are not
always coexpressed in the same tumor biopsy or cell
lines.
Multiple genes can be incorporated into our delivery
system to produce a polyvalent DNA cancer vaccine. It
is becoming clearer now that effective tumor vaccination
requires a polyvalent antigen vaccine to control human
tumor progression effectively. 5 Recent studies have demonstrated the potential feasibility of immunization using
a polynucleotide-mediated vaccine for carcinoembryonic
antigen (CEA) and MUC-1.26,27 Further work on cancer
gene vaccines will determine if such approaches are significantly effective in controlling tumor progression.
Materials and methods
MAGE-1 gene plasmid
MAGE-1 gene was cloned and isolated as previously
described.16 The MAGE-1 gene was cloned into the
pcDNA3 plasmid vector designed for expression of proteins in eukaryotic cells (InVitrogen, San Diego, CA,
USA). The pcDNA3 has an enhancer–promotor sequence
from the immediate–early gene of the human cytomegalovirus, and a polyadenylation signal and transcription
termination sequence from the bovine growth hormone
gene. The MAGE-1 gene insert in pcDNA3 was verified
by DNA sequencing. The pcDNA3 containing MAGE-1
gene (pMAGE-1) was purified by equilibrium ultracentrifugation using CsCl-ethidium bromide gradient. Purified
plasmid DNA was dissolved in 10 mm Tris-HCl, pH 8.0,
0.1 mm EDTA, assessed by spectrophotometry and
shown to have an A260/A280 ratio of 1.9 or higher.
Further analysis by gel electrophoresis was carried out to
verify the purity of the purified plasmid.
Expression and purification of rMAGE-1
rMAGE-1 was expressed as previously described and
purified by an affinity column containing Ni2+–NTA resin
(Qiagen, Chatsworth, CA, USA).16 The rMAGE-1 construct was tagged at the amino terminal end with the
sequence Gly-Ser-hexaHis for affinity purification and
affinity ELISA. The recombinant protein was further
purified by gel electrolution and a Biologic FPLC system
(Bio-Rad, Richmond, CA, USA) for analysis. Recombinant proteins were run in Laemmli 10% SDS-PAGE gel
under reducing conditions.
MAGE-3 gene plasmid
cDNA (942 bp) encoding full-length MAGE-3 was amplified by PCR from a human testis library (Clontech, Palo
Alto, CA, USA). The PCR cDNA product obtained was
digested with restriction enzyme and the appropriate size
fragment was subcloned into the vector pET30b
(Novagen, Madison, WI, USA) with hexaHis on the Cterminus, resulting in a plasmid designated pSH007. The
full length MAGE-3 cDNA was sequenced by dideoxynucleotide sequencing method using T7 DNA polymerase
(USB, Cleveland, OH, USA). Plasmid pSH007 was
digested by EcoRV and NotI and the EcoRV-NotI fragment
containing MAGE-3 gene was then inserted into the
pcDNA3 plasmid. The pcDNA3 containing MAGE-3
gene (pMAGE-3) was purified and prepared as
pMAGE-1.
A DNA Expression and purification of rMAGE-3
DNA fragment of MAGE-3 gene with the terminal hexaHis affinity tag was removed from pSH007 and inserted
into the SmaI site of pGEX-2T (Promega, Madison, WI,
USA). This expression plasmid was transformed into E.
coli strain BL21 cells. Recombinant fusion protein was
Human tumor antigen gene vaccination
T Okamoto et al
induced by isopropyl-d-thiogalactoside (IPTG) and purified from bacterial lysates by affinity chromatography
using Glutathione Sepharose 4B (Pharmacia, Piscataway,
NJ, USA). Before the elution of fusion protein, rMAGE-3
was cleaved from the GST carrier by adding thrombin to
the bound fusion protein and eluted off. The FPLC purified rMAGE-3 was verified by SDS-PAGE gel and Western blotting using Ni2+–NTA conjugate (Qiagen), and
then used for ELISA and Western blotting.
MAGE-1 DT gene plasmid
The complementary oligonucleotide sequences (sense: 5′
AGCTTACGCAACCATTTCTTCATGACGGGTATGCTG
TCAGTTGGAACACTGTT G-3′; antisense: 5′TCGACAA
CAGTGTTCCAACTGACAGCATACCCGTCATGAAG A
AATGGTTGCGTA-3′), derived from diphtheria toxin
fragment B nucleotide sequence, corresponding to the Cterminal end 16 amino acid residue TQPFLHDGY
AVSWNTV were synthesized. The sequences were
hybridized and inserted into the HindIII–SalI sites of
pBluescript. cDNA encoding full length of MAGE-1 was
then inserted into BamHI–EcoRV sites, fusing it to the Nterminus of the DT peptide nucleotide sequence. The
cDNA clone was sequenced by the dideoxynucleotide
sequencing method using T7 DNA polymerase (USB).
The BamHI–XhoI fragment containing MAGE-1 DT gene
was then inserted into the pcDNA3 plasmid. The plasmid
MAGE-1 DT was purified as described for pMAGE-1.
PCR and Southern blot analysis
RT-PCR and DNA PCR plus Southern blot analyses of
MAGE-1 were performed as previously described.28 PCR
cDNA products of MAGE-1 were evaluated by gel
electrophoresis on a 2% agarose gel and visualized by
ethidium bromide staining under UV light. Southern blot
analysis of gel electrophoresed PCR cDNA product was
performed using a MAGE-1-specific cDNA probe. The
pcDNA3, pMAGE-1 and MAGE-1 gene insert alone PCR
cDNA products were 218 bp, 1136 bp and 920 bp,
respectively. The sequences of the primers were as follows: pcDNA3; 5′ primer was 5′-TAATACGACTCACTATAGGG-3′ and the 3′ primer was 5′-AGGGGCAAACAACAGATGGC-3′ that gave a 218 bp cDNA product for
pcDNA3 alone and a 1136 bp cDNA product for pcDNA3
with MAGE-1 insert.
Mouse tissue was obtained from animals killed at
designated time-points after vaccinations. The hind
muscle at the site of injection was aseptically dissected,
minced into small pieces and processed for RNA and
DNA. Total RNA and DNA from the tissue biopsies was
extracted, isolated and purified using Tri-Reagent
(Molecular Research Center, Cincinnati, OH, USA) protocol according to the manufacturer’s instructions. RNA
and DNA extraction was carried out in a designated sterile laminar flow hood with RNase and DNase-free labware. Purified RNA and DNA was quantified and
assessed for purity by UV spectrophotometry. Normal
muscle and other body organs were assessed from nonimmunized animals of the same sex and age group.
Preparation of HVJ–liposomes
HVJ–liposomes were prepared as previously described.11
Briefly, HVJ (Z strain) was prepared from chorioallantoic
fluid of virus inoculated embryonated chick eggs and the
hemagglutinating units (HAU) of the virus titer was
determined by spectrophotometry. Preparation of liposomes involved mixing bovine brain l-a-phosphatidylserine sodium salt (Avanti Polar-lipids, Alabaster, AL,
USA), egg yolk l-a-phosphatidylcholine (Sigma, St Louis,
MO, USA), and cholesterol (Sigma) in a glass tube at a
weight ratio of 1:4, 8:2 with tetrahydrofuran (Nakarai,
Kyoto, Japan). The tube of lipid mixture was evaporated
using a rotary evaporator while under 400 mm vacuum
pressure and immersed in a 45°C water bath. HMG-1
protein was purified from calf thymus as previously
described.11 Purified plasmid DNA (200 mg) was mixed
with HMG-1 (65 mg) and BSS (137 mm NaCl, 5.4 mm KCl,
10 mm Tris-HCl, pH 7.6) in nuclease-free tubes. The mixture was incubated in a water bath at 20°C for 1 h. The
HMG-1:DNA mixture was then added to the lipid mixture and agitated intensely by vortexing, water bath sonication and incubation in a 37°C water bath for eight
cycles. Balanced salt solution (BSS) was added to the unilamellar liposomes:HMG-1:DNA, incubated in a 37°C
water bath (shaking) and then put in an ice bath. Purified
HVJ was inactivated with UV irradiation 10 J/m2 /s for
3 min. Inactivated HVJ (30 000 HAU) was added to the
DNA:liposome solution, incubated on ice for 10 min and
then incubated in a 37°C shaking water bath for 1 h. Free
HVJ was then removed from HVJ–liposomes by sucrose
density gradient centrifugation. HVJ–liposomes were
then carefully removed from the density gradient,
washed, precipitated by centrifugation (27 000 g, 30 min),
resuspended in sterile BSS plus 2 mm CaCl 2 and kept on
ice until used.
Immunization protocol
The HVJ was treated with detergent, heat and UV light
inactivating any pathogenic effects in the host. C57BL/6
male mice 6 to 8-weeks old in groups of three or four
were used in each experiment. Mice were lightly anesthetized with ether vapor and injected (150 ml) with HVJ–
liposome solution in the right quadricep hind leg muscle
only with a sterile 26-gauge needle syringe. Animals
were bled before vaccination and at designated timepoints. Before bleeding animals were anesthetized using
ether vapor and subsequently retro-orbital bled using
sterile glass polished Pasteur pipettes. Blood (approximately 250 ml) was collected in 0.5 ml Eppendorf tubes,
clotted at 4°C and centrifuged at 700 g for 5 min. The
serum was carefully removed and aliquoted into Eppendorf tubes and stored at −30°C until used.
Western blotting
Affinity purified recombinant proteins (5 mg) were prepared and Western blotting was carried out as previously
described.16 The blots were blocked with SuperBlock
blocking buffer (Pierce, Rockford, IL, USA) overnight at
room temperature. Blots were tested with various mouse
serum dilutions in PBS. After washing several times blots
were incubated with alkaline phosphatase-conjugated
goat anti-mouse IgG (g-chain specific) or IgM (m-chain
specific) (Caltag, San Francisco, CA, USA). Blots were
washed and developed using 5-bromo-4-chloro-3-indolyl-1-phosphate and Nitro blue tetrazolium (Promega).
Negative controls were conjugate alone with developing
solution, non-immunized mice sera (age and sex
matched), and pcDNA3 alone immunized mice.
975
Human tumor antigen gene vaccination
T Okamoto et al
976
ELISA
An anti-rMAGE-1 or -3 affinity ELISA was developed to
detect mouse anti-MAGE-1 or -3 antibodies. Briefly, purified rMAGE-1 or rMAGE-3(4 mg per well) with hexaHis
tag in 10 mm Tris-HCl plus 50 mM NaCl buffered saline
pH 7.5 (TBS) was incubated overnight at room temperature in Ni2+ chelate coated ELISA microplates (Xenopore,
Hawthorne, NJ, USA). Microplates were washed three
times with TBS, incubated with 5% SuperBlock blocking
buffer for 2 h at room temperature and washed three
times with PBS. Mouse serum was diluted in PBS and
added to microwells and incubated at room temperature
for 2 h and subsequently washed three times with PBS.
Goat anti-mouse IgG (g-chain specific) horseradish peroxidase conjugate (Caltag) was added and plates were
incubated for 2 h at room temperature, washed five times
with PBS and developed with ortho-phenylenediamine
sodium citrate solution plus 6 n HCl. Plates were then
read at 490 nm by using an ELISA reader (Molecular
Devices, Palo Alto, CA, USA), and data were analyzed
using the instrument software.
The DT C-terminal peptide with a four histidine tag
TQPFLHDGYAVSWNTVHHHH was synthesized by
Research Genetics (Huntsville, AL, USA). The purity of
the his tagged peptide used in the affinity ELISA was
>97%. The peptide (100 pmoles per well) was incubated
in Ni2+ chelate coated ELISA microplates. Procedures for
ELISA were the same with anti-rMAGE-1 and -3 affinity
ELISA. Rabbit anti-DT peptide serum (1:104 dilution) was
used as a positive control for DT-his tag peptide binding
in the affinity ELISA plates.
Acknowledgements
This work was supported in part by the National Institutes of Health-NCI grant CA-12582 and Joseph Drown
Foundation (DSBH).
References
1 Michel ML et al. DNA-mediated immunization to the hepatitis B
surface antigen in mice: aspects of the humoral response mimic
hepatitis B viral infection in humans. Proc Natl Acad Sci USA
1995; 92: 5307–5311.
2 Huygen K et al. Immunogenicity and protective efficacy of a
tuberculosis DNA vaccine. Nature Med 1996; 2: 893–898.
3 Wang B et al. Gene inoculation generates immune responses
against human immunodeficiency virus type 1. Proc Natl Acad
Sci USA 1993; 90: 4156–4160.
4 Wolff JA et al. Long-term persistence of plasmid DNA and
foreign gene expression in mouse muscle. Hum Mol Genet 1992;
1: 363–369.
5 Hoon DSB, Irie RF. Current status of melanoma vaccines. Can
they control malignant melanoma? BioDrug 1997; 1: 66–84.
6 Cotten M, Wagner E, Birnstiel ML. Receptor-mediated transport
of DNA into eukaryotic cells. Meth Enzymol 1993; 217: 618–643.
7 Kaneda Y, Iwai K, Uchida T. Increased expression of DNA cointroduced with nuclear protein in adult rat liver. Science 1989;
243: 375–378.
8 Dzau VJ, Mann MJ, Morishita R, Kaneda Y. Fusigenic viral liposome for gene therapy in cardiovascular diseases. Proc Natl Acad
Sci USA 1996; 93: 11421–11425.
9 Yanagihara I et al. Expression of full-length human dystrophin
cDNA in mdx mouse muscle by HVJ–liposome injections. Gene
Therapy 1996; 3: 549–553.
10 Uchiyama A et al. Transfection of interleukin 2 gene into human
melanoma cells augments cellular immune response. Cancer Res
1993; 53: 949–952.
11 Kaneda Y. Virus (Sendai virus envelope)-mediated gene transfer. Cell Biology: A Laboratory Handbook. Academic Press: San
Diego, 1994, pp 50–57.
12 Okada Y. Sendai virus-induced cell fusion. Meth Enzymol 1993;
221: 18–39.
13 Isaka Y et al. Gene therapy by skeletal muscle expression of
decorin prevents fibrotic disease in rat kidney. Nature Med 1996;
2: 418–423.
14 Morishita R et al. A gene therapy strategy using a transcription
factor decoy of the E2F binding site inhibits smooth muscle proliferation in vivo. Proc Natl Acad Sci USA 1995; 92: 5855–5859.
15 van der Bruggen P et al. A gene encoding an antigen recognized
by cytolytic T lymphocytes on a human melanoma. Science 1991;
254: 1643–1647.
16 Hoon DSB, Yuzuki D, Hayashida M, Morton DL. Melanoma
patients immunized with melanoma cell vaccine induce antibody responses to recombinant MAGE-1 antigen. J Immunol
1995; 154: 730–737.
17 Gaugler B et al. Human gene MAGE-3 codes for an antigen
recognized on a melanoma by autologous cytolytic T lymphocytes. J Exp Med 1994; 179: 921–930.
18 Hohlfield R, Engel AG. The immunobiology of muscle. Immunol
Today 1994; 15: 269–274.
19 Kuhober A et al. DNA immunization induces antibody and cytotoxic T cell responses to hepatitis B core antigen in H-2b mice.
J Immunol 1996; 156: 3687–3695.
20 Morton DL et al. Prolongation of survival in metastatic melanoma after active specific immunotherapy with a new polyvalent melanoma vaccine. Ann Surg 1992; 216: 463–482.
21 Morton DL et al. Polyvalent melanoma vaccine improves survival of patients with metastatic melanoma. Ann NY Acad Sci
1993; 690: 120–134.
22 De Backer O et al. Structure, chromosomal location, and
expression pattern of three mouse genes homologous to the
human MAGE genes. Genomics 1995; 28: 74–83.
23 Carrel S et al. Monoclonal antibodies against recombinantMAGE-1 protein identify a cross-reacting 72 kDa antigen which
is co-expressed with MAGE-1 protein in melanoma cells. Int J
Cancer 1996; 67: 417–422.
24 Autran B et al. Monoclonal B-cell response to diphtheria toxoid:
evidence for cross-reactive epitopes. Immunol 1987; 60: 531–538.
25 De Plaen E et al. Structure, chromosomal localization, and
expression of 12 genes of the MAGE family. Immunogenetics
1994; 40: 360–369.
26 Conry RM et al. Characterization of a messenger RNA polynucleotide vaccine vector. Cancer Res 1995; 55: 1397–1400.
27 Graham RA, Burchell JM, Beverly P, Taylor-Papadimitriou J.
Intramuscular immunization with MUC1 cDNA can protect C57
mice challenged with MUC1-expressing syngeneic mouse tumor
cells. Int J Cancer 1996; 65: 664–670.
28 Hoon DSB et al. Detection of occult melanoma cells in blood
with a multiple-marker polymerase chain reaction assay. J Clin
Oncol 1995; 13: 2109–2116.