Int. J. Cancer: 94, 438 – 443 (2001)
© 2001 Wiley-Liss, Inc.
Publication of the International Union Against Cancer
ORAL CYTOKINE GENE THERAPY AGAINST MURINE TUMOR USING
ATTENUATED SALMONELLA TYPHIMURIUM
Li YUHUA1, Guo KUNYUAN1*, Chen HUI2, Xie YONGMEI2, Song CHAOYANG1, Tang XUN1 and Ren DAMING2
1
Department of Hematology, Zhujiang Hospital, First Military Medical University, Guangzhou, China
2
Institute of Genetics, School of Life Science, State Key Laboratory of Genetic Engineering, Fudan University, Shangshai, China
An attenuated strain of Salmonella typhimurium was used as
a vehicle for oral gene therapy against murine tumor. Eukaryotic expression vectors containing genes of human interleukin-12 (hIL-12), human granulocyte/macrophage colonystimulating factor (hGM-CSF), mouse (m)IL-12, mGM-CSF
and green fluorescent protein (GFP) were used to transform
attenuated Salmonella (SL3261), and such transformants
were administered orally to BALB/c and C57BL/6 mice. As a
reporter gene, GFP expression in murine liver, spleen, tumor, intestine and kidney was confirmed by confocal and flow
cytometry. Soluble cytokines were detected in murine sera,
and the concentrations were much higher than those of the
control, which contributed to the increased number of cytotoxic T cells and prolongation of survival. Oral cytokine gene
therapy using live attenuated Salmonella demonstrated a significant protection against the development of two unrelated
murine tumors. These results suggest that such gene therapy
has the potential to be simple, effective and (above all) safe
against tumor.
© 2001 Wiley-Liss, Inc.
Key words: gene therapy; attenuated Salmonella; interleukin-12;
granulocyte/macrophage colony-stimulating factor
Gene therapy for cancer is the object of intense investigation.
Both GM-CSF and IL-12 are considered to have antitumor effects.1–3 GM-CSF is a naturally occurring growth factor that is
produced by a variety of cells in response to various immune or
inflammatory stimuli. GM-CSF activates neutrophils, eosinophils
and macrophages to lyse tumor cells directly or to mediate antibody-dependent cellular cytotoxicity; it can increase MHC expression, thereby increasing recognition of tumor antigen.4 In a comparative study in mice, the most potent stimulator of specific
antitumor immunity was tumor cells engineered to secrete GMCSF.5
IL-12 is a heterodimeric cytokine produced predominantly by
antigen-presenting cells in response to infection agents. IL-12
stimulates the proliferation of activated NK and T cells and induces IFN-␥ production from them. IL-12 can also enhance the
cytotoxic activity of NK cells and promote the generation of
CTLs.6
The systemic administration of recombinant IL-12 or GM-CSF
protein has shown profound antitumor effects in animal models,
but their clinical applications have been associated with significant
side effect.6,7 In our research, we planned to develop a new gene
delivery method using live attenuated Salmonella typhimurium
(ST), which might avoid the costly need for ex vivo culture and
administration of virus. Using live but attenuated ST allows the
bacterium to penetrate the epithelial M cells and reach the Peyer’s
patches. During the process of phagocytosis of ST in Peyer’s
patches, a plasmid containing the encoded gene in the ST would be
released into the cytosol and integrated into the nucleus. This could
then lead to expression of the encoded gene in the host cells under
the control of the eukaryotic promoter.8,9 This opens up the exciting possibility of oral gene therapy against tumor using ST as
vehicle.
In the current study, we transferred the IL-12 and GM-CSF
genes into STs and administered them orally to mice to induce
cytokine expression in murine tissues. This oral gene therapy
resulted in a high degree of protection against the development of
two unrelated murine tumors. These results suggest that oral ad-
ministration of IL-12 and GM-CSF genes using attenuated ST may
work as a simple, effective and safe gene therapy against tumor
MATERIAL AND METHODS
Cells and mice
Lewis lung cancer cells derived from C57BL/6 mice and 4T1
mammalian cancer cells derived from BALB/c mice were cultured
in 90% RPMI-1640 with L-glutamine medium supplemented with
10% FBS at 37°C in a humidified 5% CO2 atmosphere. Four to
6-week-old female BALB/c (H-2d) and C57BL/6 (H-2b) mice
were purchased from the Animal Institute of the Academia Sinica
and kept under standard conditions according to institutional
guidelines.
Bacterial strains, plasmids and media
The auxotrophic Salmonella typhimurium LB5000 (LT2 Trp
MetE rpsl flaA ect, R⫹M⫺ for all three systems) and SL3261
(WARY hisG 46 aroA del 407 Fusaricres etc R⫹M⫹) were kindly
provided by B.A.D. Stocker (Stanford University School of Medicine, Stanford, CA). Bacterial strains were routinely grown at
37°C in LB broth or agar supplemented with 100 ␮g/ml of ampicillin or 50 ␮g/ml of kanamycin when required.
The eukaryotic green fluorescent protein expression vector
EGFPN1 (Clontech, Palo Alto, CA) was used to characterize gene
transfer in vivo and also as a control treatment. Plasmids pCMVhIL-12, pCMVhGM-CSF, pCMVmIL-12 and pCMVmGM-CSF
were gifts from N.S. Yang (Central Research Institute of Taiwan)
and were used for gene therapy in tumor protection studies. To
create the human GM-CSF expression plasmid pCMVhGM-CSF,
the cDNA of hGM-CSF was subcloned into a pUC19 vector
containing the immediate/early promoter intron A region of human
CMV and the bovine growth hormone gene polyadenylation region. The pUC19 plasmid backbone was derived from a bluescript
SK (⫹) vector with an ampicillin-resistant gene. Plasmid pCMVmGM-CSF was created in the same way.
Construction of plasmids pCMVhIL-12 and pCMVmIL-12 was
similar: the p35 and p40 subunits of hIL-12 (mIL-12) were linked
tandemly in the same direction and each was driven by its own
CMV immediate/early promoter/enhancer, an SV40 sd/sa intron
sequence and a bovine growth hormone polyadenylation sequence.
The LB5000 was grown in 100 ml L broth (Sigma, St. Louis,
MO) to an A 600 of 0.5, chilled on ice and harvested by centrifAbbreviations: cfu, colony-forming unit; CMV, cytomegalovirus; CTL,
cytotoxic T lymphocyte; FITC, fluorescein isothiocyanate; GFP, green
fluorescent protein; GM-CSF, granulocyte/macrophage colony-stimulating
factor; IFN-␥, interferon-␥; IL-12, interleukin-12; NK, natural killer; ST,
Salmonella typhimurium; SV40, simian virus 40.
*Correspondence to: Department of Hematology, Zhujiang Hospital,
First Military Medicial University, Guangzhou 501282, China.
Fax: ⫹86-20-8430980. E-mail: [email protected]
Received 16 November 2000; Revised 6 April, 8 June 2001; Accepted
18 June 2001
Published online 18 August 2001; DOI 10.1002/ijc.1489
439
GENE THERAPY USING SALMONELLA TYPHIMURIUM
ugation (10 min, 1000g at 4°C). The pellet was suspended in a final
volume of 200 ␮l in 10% glycerol. Aliquots (50 ␮l) were mixed
with 1–2 ␮g DNA in a chilled microcentrifuge tube and transferred
to chilled cuvettes (0.2 cm electrode gap). A single pulse of
12.5KV/cm (2.5 kV, 200 ⍀, 25 ␮F) was applied and 0.5 ml of
prewarmed LB was immediately added. The bacteria were transferred to 17 ⫻ 100 mm polypropylene tubes and shaken for 1 hr at
37°C before being plated onto LB agar with 100 ␮g/ml of ampicillin or 50 ␮g/ml of kanamycin.8,10 After amplification of the
colony, the plasmids were extracted from the transferred LB5000
and then introduced into SL3261, in the same fashion as for
LB5000.
Oral administration protocols
Bacteria were grown overnight until they reached mid-log
phase. They were then harvested by centrifugation (3,000g) and
resuspended in 10% sodium bicarbonate buffer. Three times at
14-day intervals mice were gently fed the bacterial suspension
(1 ⫻ 108 cfu/mouse) in a volume of approximately 30 ␮l.11 Each
group of mice received the corresponding strain of recombinant
Salmonella. Groups (C57BL/mIL-12 ⫹ mGM-CSF, C57BL/hIL12 ⫹ hGM-CSF, BALB/mIL-12 ⫹ mGM-CSF and BALB/hIL12 ⫹ hGM-CSF received the mixture (30 ␮l) of equal volumes (15
␮l) of the suspension of both strains. Control mice received the
SL3261/EGFPN1 strain or buffer only.
Transfer of the expression plasmids to host cells
On the day of primary oral administration, tumor cells were
inoculated subcutaneously in the shaved area of the right flank of
2 mice from each group. BALB/c mice received 3 ⫻ 103 4T1
tumor cells; C57BL/6 mice received 1 ⫻ 104 Lewis tumor cells.
Twenty-eight days later, spleens, livers, kidneys, tumors and intestines were removed from these mice, and genomic DNA was
extracted by the SDS/proteinase K/phenol method, according to
Maki et al.12 The primers used for amplification of the CMV
promoter were sense 5⬘-cccagtacatgaccttatggg-3⬘ and antisense
5⬘-ggagacttggaaatccccgt-3⬘. The expected size of the PCR product
was 141 bp.13
PCR was performed in a programmable thermal cycler (PerkinElmer, Norwalk, CT) as follows: denaturation at 94°C for 30 sec,
annealing at 55°C for 30 sec and polymerization at 72°C 45 sec for
30 cycles. Specimens of GFP control and blank control groups
were also prepared, made into frozen sections and detected for
GFP under confocal microscopy.
12 hr. After coagulation, sera were collected by centrifugation (5
min, 2000g at 4°C). All samples were stored at ⫺20°C. At the end
of the experiment, samples of the same group that had been
collected at the same time were pooled and tested by hGM-CSF,
hIL-12, mGM-CSF, mIL-12 and mIFN-␥ sandwich ELISA kits
(Endogen, Woburn, MA) with triplicate wells.
Statistics
The percentages of spleen cells expressing GFP and the
CD3⫹CD8⫹/CD3⫹CD4⫹ ratios were analyzed for significance by
one-way ANOVA, comparing among the different groups of mice.
If the p-value was ⬍ 0.05, the comparisons between two groups
were analyzed by the Student-Neuman-Keuls method.
RESULTS
GFP expression and distribution in murine tissues
We first examined expression of the GFP gene in murine tissues.
Such expression could be detected under confocal microscopy in
livers, kidneys, spleens, tumors and intestines of the SL3261/
EGFPN1 orally treated mice. Spleens and tumors expressed more
GFP than other tissues. The percentages of spleen cells expressing
GFP were 3.01 ⫾ 0.29% and 3.03 ⫾ 0.36% in C57BL/6 mice and
BALB/c mice, respectively, significantly higher than that of the
blank control (F ⫽ 55.421, p ⫽ 0.000).
From confocal microscopic examination it could also be seen
that GFP expression seemed to gather in one site, especially in
tumor. The reason might be that macrophage cells are the most
important cells in the phagocytosis of ST and expression of GFP,
which secreted certain chemotactic factors that kept the cells at one
site.
Characterization of in vivo gene transfer
To confirm the transduction of the somatic cells, genomic DNA
was isolated from livers, spleens, intestines, kidneys and tumors.
Because all the eukaryotic expression vectors used in this research
were constructed with a CMV promoter, PCR was performed to
amplify a portion of the CMV promoter as the proof of transfer. As
had been expected, the PCR products of genomic DNA of the
livers, spleens, intestines, kidneys and tumors of the treated groups
and GFP groups were all 141 bp segments (Fig. 1); PCR products
were not detected in DNA from blank control mice tissue.
GFP expression in murine tissues
On day 28 after primary administration, spleens were removed
from 5 mice of each group from the GFP control and blank control
groups and made into simple cell suspension. After lysis of the red
blood cells, GFP-expressing cells were detected by a flow cytometer (Coulter, Hialeah, FL).
Tumor cell challenge
Six weeks after the first oral administration, 4 ⫻ 105 4T1 tumor
cells or 1 ⫻ 106 Lewis tumor cells were inoculated subcutaneously
in the shaved area of the right flank of BALB/c or C57BL/6 mice.
Tumor establishment was determined by palpation. Each group
consisted of 10 –15 mice.
CD3⫹CD8⫹/CD3⫹CD4⫹ lymphocyte subclone analysis
On day 42, after primary administration, 20 ␮l blood was
collected from 5 mice of each group. Heparin was added as an
antithrombin agent. After lysis of red blood cells, antibody-coated
cells were enumerated by flow cytometry (Coulter). The following
antibodies were used: FITC-conjugated rat anti-mouse CD3 (rat
IgG2bK), FITC-conjugated rat anti-mouse CD8 (rat IgG2a) and
FITC-conjugated rat anti-mouse CD4 (rat IgG2aK) (all from
PharMingen, San Diego, CA).
Cytokine assay
Blood was collected from all mice before and after oral administration every 2 weeks. The blood samples were stored at 5°C for
FIGURE 1 – Transduction of expression plasmids from Salmonella to
host cells. On day 28 after the first oral administration of the vectors,
genomic DNA was extracted from spleens, livers, intestines, kidneys
and tumors of each group of mice by SDS/proteinase K/phenol
method. A 141 bp product was amplified with DNA (200 ng) and the
CMV primers (400 nmol) CMV1(5⬘-cccagtacatgaccttatggg-3⬘) and
CMV2 (5⬘-ggagacttggaaatccccgt-3⬘). The PCR product was detected
in DNA of almost all tissues in C57BL/mIL-12 mice, as shown here.
Lane 1, DL2000 molecular marker; lane 2, spleen genomic DNA; lane
3, tumor genomic DNA; lane 4, liver genomic DNA; lane 5, intestinal
genomic DNA; lane 6, kidney genomic DNA. Other groups of mice
showed similar results, except for the blank control group.
440
YUHUA ET AL.
TABLE I – INCREASE IN CYTOTOXIC T CELLS IN PERIPHERAL BLOOD OF
C57BL/6 MICE ORALLY ADMINISTERED SALMONELLA CARRYING
EUKARYOTIC EXPRESSION PLASMIDS1
Group
Number
of mice
CD3⫹CD8⫹/CD3⫹CD4⫹
(mean ⫾ SE%)
C57BL/blank
C57BL/GFP
C57BL/mIL-12
C57BL/mGM-CSF
C57BL/mIL-12 ⫹ mGM-CSF
C57BL/hIL-12
C57BL/hGM-CSF
C57BL/hIL-12 ⫹ hGM-CSF
5
5
5
5
5
5
5
5
79.86 ⫾ 3.24
73.57 ⫾ 6.43
109.71 ⫾ 12.60*
201.65 ⫾ 31.25*
93.41 ⫾ 17.97
159.32 ⫾ 28.99*
175.58 ⫾ 24.36*
97.39 ⫾ 19.44
1
C57BL/6 mice were orally administered either 10% sodium bicarbonate buffer (C57BL/blank), SL3261/EGFPN1(C57BL/GFP),
SL3261/pCMVmGM-CSF(C57BL/mGM-CSF), SL3261/pCMVmIL12(C57BL/mIL-12), both SL3261/pCMVmIL-12 and SL3261/pCMVmGM-CSF (C57BL/mIL-12 ⫹ mGM-CSF), SL3261/pCMVhGMCSF (C57BL/hGM-CSF), SL3261/pCMVhIL-12 (C57BL/hIL-12) or
both SL3261/pCMVhIL-12 and SL3261/pCMVhGM-CSF (C57BL/
hIL-12 ⫹ hGM-CSF), 3 times at 14-day intervals. Six weeks after the
first oral administration of 20 ␮l, blood was collected from 5 mice of
each group. After lysis of the red blood cells, antibodies were added.
CD3⫹CD8⫹ cells and CD3⫹CD4⫹ cells were enumerated by flow
cytometer and the CD8⫹/CD4⫹ ratios were calculated. The oral application experiment was performed with 5 animals per group and
repeated twice. Data shown here are means ⫾ SE, representing 1
experiment of 2. The corresponding groups of BALB/c mice showed
similar results.–*Compared with blank control group, p ⬍ 0.05.
Increase of cytotoxic T cells in peripheral blood
It is known that CD3⫹CD8⫹ is the phenotype of CTLs, which
play an important role in mediating the cellular immune response
against tumor, especially in IL-12-treated mice. In vivo depletion
of CD3⫹CD8⫹ T cells, but not of CD3⫹CD4⫹ T cells, abrogated
the effect of IL-12 gene therapy.14 Six weeks after the first oral
administration, the karyocytes in the peripheral blood had not
increased significantly compared with the control groups (data not
shown). However, flow cytometry analysis showed that the
CD3⫹CD8⫹/CD3⫹CD4⫹ ratios of the treated groups of C57BL/6
mice were much higher than those of the GFP and blank control
(F ⫽ 27.99, p ⫽ 0.000; Table I), as were those in the BALB/c mice
(F ⫽ 4.05, p ⫽ 0.003). There was no significant difference
between the GFP and blank control groups. These results suggest
that cytokine genes harboring Salmonella contributed to the differentiation of T cells toward the cytotoxic phenotype, which
might result in improvement of the cellular immune response,
helping to kill the invading tumor cells. ST alone had little effect.
Prolongation of survival time
Three days after inoculation with 1 ⫻ 106 Lewis cancer cells, all
C57BL/6 mice developed tumor, but 10 days after inoculation, the
tumors of 7 mice had disappeared. Five of those mice had been
treated with SL3261/mGM-CSF, and another 2 were given both
SL3261/mGM-CSF and SL3261/mIL-12. To test whether the mice
possessed the memory T cells to counteract the tumor cells, the 7
tumor-free mice received a second challenge (2 ⫻ 106 Lewis
cancer cells), None of them developed tumor. The survival time of
the treated groups was much longer than that of the GFP and blank
control groups (Fig. 2). No tumor in BALB/c mice disappeared,
but the survival time also tended to be prolonged. It seemed that
BALB/c mice had a lower response to cytokine genes than
C57BL/6 mice. Of both C57BL/6 mice and BALB/c mice, the
murine cytokine gene-treated groups tended to have a higher
survival rate. To our surprise, the combination of GM-CSF and
IL-12 did not show a stronger effect than IL-12 or GM-CSF alone
in the prolongation of the survival time. They seemed to counteract
each other. The GFP control and blank control groups showed no
significant difference in survival time. This might mean that ST
alone could not protect the body from tumor invasion.
FIGURE 2 – Protection against tumor challenge with Salmonellacarrying cytokine genes. (a) Murine cytokine gene-treated groups. (b)
Human cytokine gene-treated groups. C57BL/6 mice were orally given
either 10% sodium bicarbonate buffer (C57BL/Blank), with SL3261/
EGFPN1 (C57BL/GFP), SL3261/pCMVmGM-CSF (C57BL/mGMCSF), SL3261/pCMVmIL-12 (C57BL/mIL-12), both SL3261/pCMVmIL-12 and SL3261/pCMVmGM-CSF (C57BL/mIL-12 ⫹ mGMCSF),SL3261/pCMVhGM-CSF(C57BL/hGM-CSF),SL3261/pCMVhIL12 (C57BL/hIL-12) or both SL3261/pCMVhIL-12 and SL3261/
pCMVhGM-CSF (C57BL/hIL-12 ⫹ hGM-CSF), 3 times at 14-day
intervals. Six weeks after the first oral administration, C57BL/6 mice
were challenged with Lewis cancer cells. Tumor growth and survival
were checked every day after inoculation. Two oral application experiments showed similar results. Data shown here represent 1 experiment
of 2, consisting of 7 mice per group. Corresponding groups of BALB/c
mice showed similar results.
Detection of hIL-12 and hGM-CSF in sera
The concentration of hIL-12 (and hGM-CSF) in BALB/c mice
was almost the same as that in C57BL/6 mice (Fig. 3). Both
concentrations reached a peak at 2– 4 weeks after oral administration and decreased gradually. At 10 –12 weeks, such concentrations reached normal levels. These variations might be the results
of recognition and clearance of the heterogenous antigens, presumably operating as follows: after oral administration the exogenous gene was expressed in epithelial cells of the intestine. When
the Salmonella reached the lymphatic system, spread to tissues all
over the mice and was expressed there, cytokine serum concentrations increased continuously. When they reached a certain level,
they were recognized, and antibodies were produced to clear them.
Detection of mGM-CSF and mIL-12 in sera
The variations in mGM-CSF and mIL-12 concentrations in
C57BL/6 mice were much more complicated than those of hGMCSF and hIL-12. They increased, decreased and then increased
again (Fig. 4). mGM-CSF and mIL-12 were murine cytokines. It
was assumed that their production was under control of the negative response of the body and that their antibodies were not easily
induced. Therefore, mGM-CSF and mIL-12 concentrations remained at high levels even at 10 –12 weeks after primary oral
441
GENE THERAPY USING SALMONELLA TYPHIMURIUM
FIGURE 3 – Presence of transduced hGM-CSF and hIL-12 cytokines
in the sera of C57BL/6 mice. (a) Presence of transduced hGM-CSF
cytokine in the sera of C57BL/6 mice. (b) Presence of transduced
hIL-12 cytokine in the sera of C57BL/6 mice. C57BL/6 mice were
orally administered either 10% sodium bicarbonate buffer (C57BL/
Blank), SL3261/EGFPN1 (C57BL/GFP), SL3261/pCMVhGM-CSF
(C57BL/hGM-CSF), SL3261/pCMVhIL-12 (C57BL/hIL-12) or both
SL3261/pCMVhIL-12 and SL3261/pCMVhGM-CSF (C57BL/hIL12 ⫹ hGM-CSF) 3 times at 14-day intervals. Sera samples were taken
from these mice every 2 weeks. At the end of the experiment, samples
from the same group collected at the same time were pooled and
quantified using hGM-CSF and hIL-12 cytokine ELISA kits. Data
shown are mean ⫾ SE from 2 independent oral application experiments with triplicate wells. The corresponding groups of BALB/c mice
showed similar results.
administration. The concentration patterns of mGM-CSF and
mIL-12 in BALB/c mice were similar.
It was previously reported that target gene could be expressed in
a few hours in the intestine after oral administration with Salmonella, but about 28 days elapsed between the first oral administration and the start of gene expression in other tissues,10 which might
explain why the concentrations of mGM-CSF and mIL-12 start to
rise at about 6 – 8 weeks after the first oral administration, although
the peak mIL-12 concentration came later in the C57BL/mIL-12
group than in the C57BL/mIL-12 ⫹ mGM-CSF group. Some of
the supposed reasons for this possibly complex mechanism are as
follows: perhaps in the C57BL/mIL-12 group IFN-␥, which was
upregulated by mIL-12, acted in synergy with other stimuli, suggesting an amplification loop for IL-12 production,15 whereas in
the C57BL/mIL-12 ⫹ mGM-CSF group, mIL-12 production
might have been influenced by even more factors. For example,
GM-CSF could have induced the secretion of IL-1, which could
synergize with IL-12 to induce considerable IFN-␥ production by
T cells;16,17 then IFN-␥ in turn could have upregulated IL-12
production,15 whereas GM-CSF itself could have inhibited IL-12
production.18 All these influences might have caused the difference
in mIL-12 variations between C57BL/mIL-12 ⫹ mGM-CSF and
C57BL/mIL-12 mice.
FIGURE 4 – Presence of transduced mGM-CSF and mIL-12 cytokines in the sera of C57BL/6 mice. (a) Presence of transduced mGMCSF cytokine in the sera of C57BL/6 mice. (b) Presence of transduced
mIL-12 cytokine in the sera of C57BL/6 mice. C57BL/6 mice were
orally administered either 10% sodium bicarbonate buffer (C57BL/
Blank), SL3261/EGFPN1 (C57BL/GFP), SL3261/pCMVmGM-CSF
(C57BL/mGM-CSF), SL3261/pCMVmIL-12 (C57BL/mIL-12), or
both SL3261/pCMVmIL-12 and SL3261/pCMVmGM-CSF (C57BL/
mIL-12 ⫹ mGM-CSF) 3 times at 14-day intervals. Sera samples were
taken from these mice every 2 weeks. At the end of the experiment,
samples of the same group collected at the same time were pooled and
quantified using mGM-CSF and mIL-12 cytokine ELISA kits. Data
shown are mean ⫾ SE from 2 independent oral application experiments with triplicate wells. The corresponding groups of BALB/c mice
showed similar results.
Detection of mIFN-␥ in sera
IFN-␥ is the effector molecule that mediates the antitumor effect
of IL-12. Therefore, IFN-␥ activity might reflect that of IL-12. In
both C57BL/mIL-12 and C57BL/mIL-12 ⫹ mGM-CSF mice, the
mIFN-␥ concentration began to rise at about 6 – 8 weeks after the
first oral administration, the same time as mIL-12. Variations in
mIFN-␥ concentration were similar in the corresponding groups of
BALB/c mice. However, in SL3261/hIL-12-treated and SL3261/
hIL-12 ⫹ SL3261/hGM-CSF-treated C57BL/6 mice, mIFN-␥ levels changed in a different fashion, corresponding to that of the
hIL-12 group (Fig. 5).
The concentrations of the five cytokines detected were much
higher in treated groups than in GFP and blank control groups,
possibly showing that cytokine secretion upregulated the immune
function of the mice and mediated the antitumor effect. These
results indicated the efficiency of oral gene therapy using live
attenuated Salmonella.
DISSCUSSION
After chemotherapy or irradiation there are still 106–109 tumor
cells left, which are the source for relapse. However, chemotherapy or irradiation to eliminate all these tumor cells would do more
442
YUHUA ET AL.
FIGURE 5 – Presence of mIFN-␥ cytokine in the sera of C57BL/6
mice. (a) mIFN-␥ cytokine in the sera of C57BL/6 mice given Salmonella-harboring murine cytokine genes. (b) mIFN-␥ cytokine in the
sera of C57BL/6 mice given Salmonella-harboring human cytokine
genes. C57BL/6 mice were orally given either 10% sodium bicarbonate buffer (C57BL/Blank), SL3261/EGFPN1 (C57BL/GFP), SL3261/
pCMVmGM-CSF (C57BL/mGM-CSF), SL3261/pCMVmIL-12 (C57BL/
mIL-12), both SL3261/pCMVmIL-12 and SL3261/pCMVmGM-CSF
(C57BL/mIL-12 ⫹ mGM-CSF), SL3261/pCMVhGM-CSF (C57BL/
hGM-CSF), SL3261/pCMVhIL-12 (C57BL/hIL-12), or both SL3261/
pCMVhIL-12 and SL3261/pCMVhGM-CSF(C57BL/hIL-12 ⫹ hGMCSF) 3 times at 14-days intervals. Sera samples were taken from these
mice every 2 weeks. At the end of the experiment, samples of the same
group collected at the same time were pooled and quantified using an
mIFN-␥ cytokine ELISA kit. Data shown are mean ⫾ SE from 2
independent oral application experiments with triplicate wells. The
corresponding groups of BALB/c mice showed similar results.
harm to normal tissues. Immunotherapy is perhaps the most promising way to remove minimal residual disease.
Cytokines are very important in mediating the immune system.
Limitations associated with systemic administration of cytokines
include the short half-life of many cytokines and the severe side
effects commonly observed with direct administration.4 Gene
transfer might avoid some of the side effects, but it is costly,
time-consuming and complicated. A new, safe and convenient way
for clinical gene transfer is urgent.
It was previously demonstrated that Salmonella could mediate
GFP expression in dendritic spleen cells about 28 days after the
first oral administration.10 We therefore chose 28 days after the
first oral administration to test gene transfer and GFP expression in
murine tissues. To our excitement, GFP expression could not only
be detected in the spleen, but also in many other organs such as the
intestine, kidney, tumor and liver. However, our results did not
support the hypothesis that Salmonella possessed a tumor-targeting property, as reported by Low et al.19 and Pawelek et al.:20
tumor-targeted Salmonella exhibited tumor accumulation ratios in
excess of 1000:1 compared with normal tissues, since gene transduction and GFP expression could be detected in almost all mouse
organs, and GFP expression in the spleen was more wide-spread
than that of the tumor. The different strains of Salmonella used in
our research might have contributed to the difference in results, in
addition to different methods for comparison and different murine
tumor models, etc. At the same time, our results gave us an
exciting indication that attenuated Salmonella was the best vector
for cytokine gene expression, because it could cause expression at
the sites where they could exert their maximum effect on lymphocytes and act directly on tumor cells.
It was previously shown that both IL-12 and GM-CSF have
antitumor properties,5,21,22 and other experiments had revealed that
the combination of these two cytokines might amplify this effect.23
It has recently been reported that GM-CSF could inhibit IL-12 P40
production by anti-CD40/IFN-␥-stimulated mouse Langerhans
cells.18 Our research has shown that the combination of IL-12 and
GM-CSF did not increase the CD8⫹/CD4⫹ ratio as much as that
of IL-12 or GM-CSF alone. The survival time of IL-12 ⫹ GMCSF-treated groups was shorter than that of groups treated with
GM-CSF alone. It seemed that IL-12 and GM-CSF did not act in
synergy, but their counteracting effect was not significant either.
The reason was not clear and the explanations might be various.
Perhaps GM-CSF inhibited the secretion of IL-12, or vice versa,
IL-12 inhibited tumor angiogenesis while GM-CSF counteracted it
by stimulating endothelial cells to proliferate.24,25
According to Darji et al.,11 the CTL reaction induced by Salmonella started to increase 5 weeks after the first oral administration and continued to increase. Based on this result, we challenged
the mice with tumor cells 6 weeks after the first oral administration. The recombinant Salmonella harboring either murine cytokine genes or human cytokine genes displayed significant antitumor effect. However, hGM-CSF and hIL-12 concentrations
decreased to normal levels in a short time, whereas mGM-CSF and
mIL-12 concentrations remained at high levels for a longer time,
which might explain why there was a difference in survival time
between the murine and human cytokine gene-treated groups. The
regulatory mechanisms contributing to the variations in hGM-CSF,
hIL-12, mGM-CSF and mIL-12 concentrations should be tested
further.
The results of our research show that Salmonella mediated
cytokine gene expression in vivo and exerted an antitumor effect.
It was concluded that attenuated Salmonella could serve as a
vector for oral gene therapy and provide a new safe and efficient
way to treat cancer.
ACKNOWLEDGEMENTS
We are grateful to Dr. Z. Minghui of the Second Military
Medical University (Shanghai, China) and to Miss W. Feng and
Mr. X. Kuangcheng of the First Hospital of Shanghai (Shanghai,
China) for their help in cell culture and flow cytometry analysis.
REFERENCES
1.
2.
3.
Sumimoto H, Tani K, Nakazaki Y, et al. GM-CSF and B7-1 (CD80)
co-stimulatory signals co-operate in the induction of effective antitumor immunity in syngeneic mice. Int J Cancer 1997;73:556 – 61.
Nastala CL, Edington HD, McKinney TG, et al. Recombinant IL-12
administration induces tumor regression in association with IFNgamma production. J Immunol 1994;153:1697–706.
Nishioka Y, Hirao M, Robbins PD, et al. Induction of systemic and
4.
therapeutic antitumor immunity using intratumoral injection of dentritic cells genetically modified to express interleukin 12. Cancer Res
1999;59:4035– 41.
Hogge GS, Burkholder JK, Culp J, et al. Development of human
granulocyte-macrophage colony-stimulating factor–transfected tumor
cell vaccines for the treatment of spontaneous canine cancer. Hum
Gene Ther 1998;9:1851– 61.
GENE THERAPY USING SALMONELLA TYPHIMURIUM
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Armitage JO. Emerging application of recombinant human granulocyte-macrophage colony-stimulating factor. Blood 1998;92:4491–
508.
Simmons SJ, Tjoa BA, Rogers M, et al. GM-CSF as a systemic
adjuvant in a phase II prostate cancer vaccine trial. Prostate 1999;39:
291–7.
Gambotto A, Tuting T, McVey DL, et al. Induction of antitumor
immunity by direct intratumoral injection of a recombinant adenovirus vector expressing interlukin-12. Cancer Gene Ther 1999;6:45–53.
Urashima M, Suzuki H, Yuza Y, et al. An oral CD40 ligand gene
therapy against lymphoma using attenuated Salmonella typhimurium.
Blood 2000;95:1258 – 63.
Courvalin CG, Goussard S, Courvalin P. Bacteria as gene delivery
vectors for mammalian cells. Curr Opin Biotech 1999;10:477– 81.
Pagilia P, Medina E, Arioli I, et al. Gene transfer in dendritic cells
induced by oral DNA vaccination with Salmonella typhimurium results in protective immunity against a murine fibrosarcoma. Blood
1998;92:3172– 6.
Darji A, Guzman A, Gerstel B, et al. Oral somatic transgene vaccination using attenuated S. typhimurium. Cell 1997;91:765–75.
Maki R, Roeder W, Traunecker A, et al. The role of DNA rearrangement and alternative RNA processing in the expression of immunoglobulin delta gene. Cell 1981;24:353– 65.
During MJ, Symes CW, Lawlor PA, et al. An oral vaccine against
NMDAR1 with efficacy in experimental stroke and epilepsy. Science
2000;287:1453– 64.
Rekehmilevich AL, Turner J, Ford MJ, et al. Gene gun-mediated skin
transfection with interleukin 12 gene results in regression of established primary and metastatic murine tumors. Proc Natl Acad Sci
USA 1996;93:6291– 6.
Skeen MJ, Miller MA, Shinnick TM, et al. Regulation of murine
macrophage IL-12 production, activation of macrophages in vivo,
restimulation in vitro, and modulation by other cytokines. J Immunol
1996;156:1196 –206.
443
16. Skeen MJ, Ziegler HK. Activation of gamma delta T cells for production of IFN-gamma is mediated by bacteria via macrophagederived cytokines IL-1 and IL-12. J Immunol 1995;154:5832– 41.
17. Mahvi DM, Sondel PM, Yang NS, et al. Phase I/IB study of immunization with autologous tumor cells transfected with the GM-CSF
gene by particle-mediated transfer in patients with melanoma or
sarcoma. Hum Gene Ther 1997;8:875–91.
18. Tada Y, Asahina A, Nakamura K, et al. Granulocyte/macrophage
colony-stimulating factor inhibites IL-12 production of mouse Langerhans cells. J Immunol 2000;164:5113–9.
19. Low B, Ittensohn M, Le T, et al. Lipid A mutant Salmonella with
suppressed virulence and TNF␣ induction retain tumor-targeting in
vivo. Nature Biotech 1999;17:37– 41.
20. Pawelek J, Low KB, Bermudes D. Tumor-targeted Salmonella as a
novel anti-cancer vector. Cancer Res 1997;57:4537– 44.
21. Ogawa M, Umehara K, Yu WG, et al. A critical role for a
peritumoral stromal reaction in the induction of T-cell migration
responsible for interlukin12-induced tumor regression. Cancer Res
1999;59:1531– 8.
22. Boggio K, Carlo ED, Rovero S, et al. Ability of systemic interleukin-12 to hamper progressive stages of mammary carciogenesis in
HER2/neu transgenic mice. Cancer Res 2000;60:359 – 64.
23. Aruga A, Tanigawa K, Aruga E, et al. Enhanced adjuvant effect of
granulocyte-macrophage colony-stimulating factor plus interlukin-12
compared with either alone in vaccine-induced tumor immunity. Cancer Gene Ther 1999;6:89 –95.
24. Gee MS, Koch CJ, Evans SM, et al. Hypoxia-mediated apoptosis from
angiogenesis inhibition underlies tumor control by recombinant interleukin 12. Cancer Res 1999;59:4882–9.
25. Mueller MM, Fusenig NE. Constitutive expression of G-CSF and
GM-CSF in human skin carcinoma cells with functional consequence
for tumor progression. Int J Cancer 1999;83:780 –9.