CELL STRUCTURE
AND FUNCTION
16:
503-510
(1991)
© 1991 by Japan Society for Cell Biology
A Potential Approach for Gene Therapy Targeting Hepatoma Using a Liver-Specific Promoter on a Retroviral Vector
Shigeki Kuriyama^, Masahide Yoshikawa1, Shigeaki Ishizaka1, Tadasu Tsujii1, Kazuhiro Ikenaka2, Tetsushi Kagawa2, Noriyuki Morita2, and Katsuhiko Mikoshiba2 3 13rd Department of Internal Medicine, Nara Medical University, 840 Shijo-cho, Kashihara, Nara 634, institute for Protein Research, Osaka University, 3-2 Yamadaoka, Suita, Osaka 565, and ^Institute of Medical Science, University of Tokyo, 4-6-1 Shiroganedai, Minato-ku, Tokyo 108, Japan. Keywords: Hepatoma/Liver-specific promoter/Genetherapy/Retroviral vector//3-galactosidase/Albumin enhancer and promoter ABSTRACT. Recent technological advances made in molecular biology and in vitro culture of humanand other mammalian cells have led to broad medical and scientific acceptance of the feasibility of gene therapy for genetic diseases. Cancer might practically be one of the attractive targets for such therapy. For the treatment of cancer, it is important to manipulate the gene of interest such that it is expressed solely in cancer cells. Wehave developed a tissue-specific gene expression system, based on a tissue-specific promoter on a retroviral vector. A murine ecotropic retroviral vector was constructed in which the Escherichia coli ^-galactosidase gene served as a reporter; it was expressed under control of the albumin enhancer element and promoter. The tissue specificity of this vector was first assessed in vitro, and /3-galactosidase activity was detected exclusively in hepatoma cell lines. This recombinant retrovirus was injected directly into a subcutaneous tumor composedof transplantable murine MH-134hepatoma cells, and expression of the gene was observed in vivo. Then this recombinant retrovirus was injected via the spleen or directly into the liver, resulting in the gene expression in dividing hepatocytes in partially hepatectomized mice, but not in nondividing hepatocytes in normal mice. Gene transfer specific to dividing hepatocytes and expression by means of retroviral vectors should possess high potential for selective elimination of hepatoma cells surrounded by nondividing normal hepatocytes.
Remarkable
developments
in molecular
biology
during the past several decades have provided a promising new treatment, gene therapy, for clinical use (7, 18). The most useful and popular model vectors for the efficient introduction of foreign genes into target mammalian cells have been derived from murine and avian retroviruses (6). Gene transfer by retroviral vectors has already shown that normal genes can be transferred to and expressed in cells derived from animals and humans (5, ll, 12, 13, 17, 19, 20). Cancer, as well as congenital genetic diseases, is an attractive target for gene therapy using retroviral vectors. Conveniently, retroviruses provide a potential means of selectively infecting cancer cells because they can only integrate into the genomeof dividing cells; most normal cells surrounding cancers are in a quiescent,
nonreceptive
stage of cell growth.
Abbreviations: neo, neomycin phosphotransferase; LTR, long terminal repeat; /3-gal, /3-galactosidase; DMEM,Dulbecco's modified Eagle's
medium;
FCS, fetal
calf
serum; PBS, phosphate-buffered
saline; cfu, colony-forming units; X-Gal, 5-bromo-4-chloro-3-indolyl-jQ-galactoside; Alb e/p, albumin enhancer and promoter. * To whomcorrespondence should be addressed.
Many cancer chemotherapeutic agents have become commercially available, and their killing effect has a definite selectivity for cancer cells over normal cells. Commonly used anticancer drugs such as antimetabolites, alkylating agents, plant alkaloids, and antitumor antibiotics, however, have almost no defined tissue specificity. The mechanismsof their antitumor effects mostly utilize the differences in cycle times between normal and tumor cells. For instance, antimetabolites act primarily during the DNAsynthetic phase of the cell cycle by preventing further entry of cells into the DNAsynthetic phase.
Plant alkaloids
are cell-cycle
specific
and pro-
duce metaphase arrest in dividing cells. Although alkylating agents act on cells in all phases of the cell cycle, proliferating cells are muchmorevulnerable to their effects. Therefore, it is easily understood that the major side effect of all present anticancer drugs is bone marrow suppression, resulting in neutropenia, anemia, and thrombocytopenia, because bone marrowis also a rapidly proliferating tissue. Some studies on gene therapy of cancer have already
been done, introducing
the
tumor necrosis factor gene (1), granulocyte colony-stim503
S. Kuriyama et al.
ulating
factor
gene (3), and thymidine
kinase gene (4)
by retrovirus-mediated gene transfer. These attemptswithout concern for tissue specificity-may result in the sameside effects as the present cancer chemotherapies because retroviruses seem to be able to infect all kinds of cells as long as they are in a proliferating stage. We have already reported a novel promoter assay using a retroviral vector (9), pIP200 (Fig. 1A). It contains the neomycin phosphotransferase (neo) gene to confer G418-resistance on the transfected cells and the Escherichia coli /3-galactosidase (/5-gal) gene as a reporter with-
surgical
operations can seldom be performed due to the complication of liver cirrhosis. For this therapy, it is essential to kill the hepatoma cells selectively without causing
any damage to actively dividing normal cells, such as
p-pIP200.
bone marrowcells. In the present study, a retroviral vector was generated in which the /3-gal gene was regulated albumin
Ecotropic psi2 packaging cells, kindlyprovided by Dr. Mul-
enhancer
ligan (10), were grown in Dulbecco's modified Eagle's me-
dium (DMEM)c ontaining 10%fetal calf serum (FCS). Trans-
and promoter. Wefirst examined whether the albumin enhancer and promoter could still maintain its appropriate liver-specific expression, without being affected by
5 fL T R H
cells in vivo, were infected.
rus. All plasmids were constructed by standard recombinant DNAtechniques (16). The Alb e/p-pIP200 vector (Fig. IB) was constructed from the previous pIP200vector, whichcontains the neo gene and the /3-gal gene between the LTRs. A 2.2-kb fragment containing the albuminpromoterand its enhancerelement,located between-8.5 kb and -10.4kb upstreamof the promoter, was kindly provided by Dr. Palmiter (14). Conveniently, the fragmentcontains a HindIII site on eachend. It wastherefore digested by Hind III and inserted into the unique Hind III site whichis located between the neo gene and the /3-gal gene in the pIP200vector, resulting in construction of Alb e/
This study is of gene therapy of hepatoma because it
of the liver-specific
retrovirus by a direct gene transfer systeminto the liver, while regenerating hepatocytes, whichmimichepatoma
Vectorconstructionandproductiono frecombinantretrovi-
ase activity.
under control
that quiescent hepatocytes were not infected by the
M A T E R I A L SA N D M E T H O D S
in the two Moloneymurine leukemia virus long terminal repeats (LTRs). Foreign promoters can be subcloned into the Hind III site, located between the two genes. Foreign promoters can also be inserted between the two Sal I sites, deleting a short DNAfragment without any detectable effects on neomycinphosphotransferis a common cancer and, most importantly,
the upstream LTRpromoter, in a variety of cell lines. Wethen investigated whether the /3-gal gene under the control of the albuminenhancerand promotercould be expressed in hepatoma in vivo. We finally confirmed
fection was by the calcium phosphatecoprecipitation method as previously described (8). Stable transformants were se-
lected by addition of G418(Gibco) to a final concentration of 1 mg/ml. Independentcolonies were isolated and expanded. Theculture supernatantwas removed,f iltered througha 0.45jum pore-size filter (Nalgene) and stored at -70°C. It served
S ail S a i l Hind HI I ne o H -H /? -g a l
as a source for infectious retroviral particles.
Cell culture, retroviral infection, and X-Gal staining of the cells. All cell lines were grownin DMEM c ontaining 10% FCS, 100 units/ml of penicillin, and 100 //g/ml of streptomy-
A . p l P 20 0
cin.
5 'L T R H
ne o
H
A lb e /p
B. Alb e/p-PIP
0 -g a l
H 3 'LT R
200
Fig. 1. Structure of the vectors. The pIP200 vector (A) contains the neo gene and /3-gal gene within the two Moloneymurineleukemiavirus LTRs.A Hind III site is located betweenthe two genes, whereforeign promoterscan be subcloned. Foreign promoterscan be also inserted betweenthe two Sal I sites. The longer transcript initiating from the 5' LTRcontains both neo and /3-gal regions; however, /3-gal will not be produced. On the other hand, the ATGcodon of the /3-gal gene in the transcript initiated from the internal promoterwill initiate translation efficiently. Therefore, /3-galactosidasep roductionshowsthe sametissue specificityas that of the promoter inserted, which can be detected by X-Gal staining. The albuminenhancer, located between-8.5 kb and - 10.4 kb upstream, and promoterwas cloned into the HindIII site of the pIP200 vector, resulting in the plasmid Alb e/p-pIP200 (B). 504
Cells were infected by incubating them in the viral supernatant with Polybrene (final cone, 8 //g/ml) at 37°C for 3 to 6 h; next the mediumwasreplaced with virus-free fresh medium, followedby incubation overnight at 37°C.The cells were then fixed with 0.5% glutaraldehyde for 5 min at room temperature. After washing twice with phosphate-buffered
saline
(PBS) with 1 mMMgCh,they were incubated overnight in an X-Gal (5-bromo-4-chloro-3-indolyl-/3-galactoside) reaction mixture consisting of 0.1% X-Gal, 0.01% sodium deoxycholate, 0.02% NP-40, 1 mMMgCl2, 5 mMK3[Fe(CN)6], and 5 mMK4[Fe(CN)6] in PBS without calcium or magnesium (2). Retroviral infection in vivo and histochemistry. Trans-
plantable murine MH-134hepatomacells in ascites (4 x 107 in 0.2 ml of DMEMwith 10% FCS) were injected into the backs of C3H/He mice. Whena subcutaneoustumorexceeded10
mmin diameter, it directly received 0.5 ml of retroviral super-
An Approach for GeneTherapy Targeting Hepatoma. natant with 10//g/ml of Polybrene. Mice were sacrificed 7 days after the injection, and tumors were removed into the
RESULTS AND DISCUSSION
Construction of liver-specific retroviral vector and analysis in vitro. The albumin promoter, which is contained within the 0.3-kb fragment, directs tissue-specific transcription; however, it is not expressed efficiently unless an additional 5' sequence is present. The precise location of the albumin enhancer-like elements has been already examined in transgenic mice (14). It was proven that a region located between -8.5 kb and -10.4kb upstream of the albumin promoter is indispensable for
fixative consisting of 4% paraformaldehyde, 0.2% glutaralde-
hyde, and 0.02% NP-40 in 0.1 M sodium phosphate buffer (pH 7.4). After fixation for 10 min, they were cut into 200-jum sections with a microslicer, followed by post-fixation for 50 min in the same fixative. These slices were then rinsed twice in PBS with 1 mMMgCl2 and incubated for 4h at 37°C in the X-Gal reaction mixture. Partial hepatectomy was carried out in 6- to 7-week-old male C3H/Hemice under ether anesthesia. Thirty-six h after two-thirds partial hepatectomy, 0.3 ml of retroviral supernatant with 10 jug/m\ of Polybrene was injected into the inferior tip of the spleen or directly into the liver. The mice were then sacrificed 7 days later and their livers were removed. Thereafter the same procedure as described for the X-Gal staining of subcutaneous hepatoma was followed, except that the livers were sliced into 100-^m sections instead of 200-^m.
liver-specific
strong expression. This albumin enhancer
and promoter region was cloned into the Hind III site of the pIP200 vector (Fig. 1A). The integrity and orientation of this region in the resulting plasmid, Alb e/ppIP200 (Fig. IB), were confirmed by restriction map-
ping. The longer transcript
initiated
from the 5' LTR
contains both the neo and /3-gal genes; however, the /3-gal gene utilizes the second translational initiation
codon within the transcript for its translational
initia-
;.
^^^ zr+>
%
-*s*
j^
V»
'*/!
T-
à"
-'^ià"""' à"
^
^*&. . %r-iN*~
s*
~ å *v
M
**
Vå
\«w
t:.
^ «*r
Fig. 3. Expression of /3-galactosidase detected by X-Gal staining in non-hepatoma cell lines. NIH 3T3 fibroblasts were infected with the psi2-KS-6 retrovirus (A) or BAGretrovirus (B). Murine B lymphoma cells, A20-2J, were infected with the psi2-KS-6 retrovirus (C) or BAGretrovirus (D). Magnification: A and B, x 140; C and D, x280. 506
An Approach for Gene Therapy Targeting Hepatoma.
onstrate that expression of the /3-gal gene was strictly regulated by the albumin enhancer and promoter within the pIP200 vector. The A20-2J B lymphoma cells, originating from the Balb/c mouse, were also infected with each retrovirus. Again, none of the cells infected with the psi2-KS-6 retrovirus were X-Gal positive (Fig. 3C), while approximately 5 to 10% of the cells infected with the BAGretrovirus were stained blue (Fig. 3D). It is important that the psi2-KS-6 retrovirus did not express /3gal in the B lymphoma cell line, because this suggests that bone marrow cells mayescape the killing effect of the anti-hepatoma retrovirus as long as it is regulated by a tissue-specific promoter. It was thus proven that the al-
bumin enhancer and promoter still retained its tissuespecific expression, even on a retroviral vector. These results indicate that one can easily construct a desirable tissue-specific retroviral vector, which is of great advantage for cancer gene therapy.
å
å
Gene expression in hepatoma in vivo. Experiments were performed to determine whether the /3-gal gene under control of the albumin enhancer and promoter would be expressed in hepatomas in vivo. Transplantable murine MH-134hepatoma cells were injected into the backs of C3H/Hemice, resulting in subcutaneous hepatoma formation. The psi2-KS-6 retrovirus was directly injected into the subcutaneous tumor when it ex-
>- fe
m
X
B
*
"V'V^I fc.
Fig. 4. Expression of /3-galactosidase in hepatoma infected with psi2-KS-6 retrovirus. Transplantable murine MH-134hepatoma cells were injected into the backs of C3H/Hemice, resulting in subcutaneous hepatoma formation. The psi2-KS-6 retrovirus was directly injected into the solid tumor, followed by microslicing to 200-^m thickness 7 days later. The cells expressing the /3-gal gene were detected by X-Gal staining. Magnification: A, 35; B, x 140. 507
S. Kuriyama et al.
ceeded 10 mmin diameter. One week after infection, the tumors were removed and microsliced into 200-^m
sections, followed by X-Gal staining. Several hepatoma cells from the backs of mice receiving the psi2-KS-6
retrovirus injection were X-Gal positive (Fig. 4). However, we could not detect the /3-gal gene expression in connective tissues surrounding hepatoma (data not shown). It was thus shown that we could force expression of a gene of interest under the control of the albu-
-*.
min enhancer and promoter exclusively in hepatoma by direct administration of the retrovirus. Gene expression specific to dividing hepatocytes by direct in vivo gene transfer. Wecontinued experiments to examine whether wecould directly introduce a foreign gene into the liver in vivo. Weinjected psi2-KS-6 retroviruses into murine spleen so that they could reach the liver by way of the portal vein. Seven days later, the mice were sacrificed; their livers were microsliced to
$
»'à"
x"**
å1 x
AND FUNCTION
16:
503-510
(1991)
© 1991 by Japan Society for Cell Biology
A Potential Approach for Gene Therapy Targeting Hepatoma Using a Liver-Specific Promoter on a Retroviral Vector
Shigeki Kuriyama^, Masahide Yoshikawa1, Shigeaki Ishizaka1, Tadasu Tsujii1, Kazuhiro Ikenaka2, Tetsushi Kagawa2, Noriyuki Morita2, and Katsuhiko Mikoshiba2 3 13rd Department of Internal Medicine, Nara Medical University, 840 Shijo-cho, Kashihara, Nara 634, institute for Protein Research, Osaka University, 3-2 Yamadaoka, Suita, Osaka 565, and ^Institute of Medical Science, University of Tokyo, 4-6-1 Shiroganedai, Minato-ku, Tokyo 108, Japan. Keywords: Hepatoma/Liver-specific promoter/Genetherapy/Retroviral vector//3-galactosidase/Albumin enhancer and promoter ABSTRACT. Recent technological advances made in molecular biology and in vitro culture of humanand other mammalian cells have led to broad medical and scientific acceptance of the feasibility of gene therapy for genetic diseases. Cancer might practically be one of the attractive targets for such therapy. For the treatment of cancer, it is important to manipulate the gene of interest such that it is expressed solely in cancer cells. Wehave developed a tissue-specific gene expression system, based on a tissue-specific promoter on a retroviral vector. A murine ecotropic retroviral vector was constructed in which the Escherichia coli ^-galactosidase gene served as a reporter; it was expressed under control of the albumin enhancer element and promoter. The tissue specificity of this vector was first assessed in vitro, and /3-galactosidase activity was detected exclusively in hepatoma cell lines. This recombinant retrovirus was injected directly into a subcutaneous tumor composedof transplantable murine MH-134hepatoma cells, and expression of the gene was observed in vivo. Then this recombinant retrovirus was injected via the spleen or directly into the liver, resulting in the gene expression in dividing hepatocytes in partially hepatectomized mice, but not in nondividing hepatocytes in normal mice. Gene transfer specific to dividing hepatocytes and expression by means of retroviral vectors should possess high potential for selective elimination of hepatoma cells surrounded by nondividing normal hepatocytes.
Remarkable
developments
in molecular
biology
during the past several decades have provided a promising new treatment, gene therapy, for clinical use (7, 18). The most useful and popular model vectors for the efficient introduction of foreign genes into target mammalian cells have been derived from murine and avian retroviruses (6). Gene transfer by retroviral vectors has already shown that normal genes can be transferred to and expressed in cells derived from animals and humans (5, ll, 12, 13, 17, 19, 20). Cancer, as well as congenital genetic diseases, is an attractive target for gene therapy using retroviral vectors. Conveniently, retroviruses provide a potential means of selectively infecting cancer cells because they can only integrate into the genomeof dividing cells; most normal cells surrounding cancers are in a quiescent,
nonreceptive
stage of cell growth.
Abbreviations: neo, neomycin phosphotransferase; LTR, long terminal repeat; /3-gal, /3-galactosidase; DMEM,Dulbecco's modified Eagle's
medium;
FCS, fetal
calf
serum; PBS, phosphate-buffered
saline; cfu, colony-forming units; X-Gal, 5-bromo-4-chloro-3-indolyl-jQ-galactoside; Alb e/p, albumin enhancer and promoter. * To whomcorrespondence should be addressed.
Many cancer chemotherapeutic agents have become commercially available, and their killing effect has a definite selectivity for cancer cells over normal cells. Commonly used anticancer drugs such as antimetabolites, alkylating agents, plant alkaloids, and antitumor antibiotics, however, have almost no defined tissue specificity. The mechanismsof their antitumor effects mostly utilize the differences in cycle times between normal and tumor cells. For instance, antimetabolites act primarily during the DNAsynthetic phase of the cell cycle by preventing further entry of cells into the DNAsynthetic phase.
Plant alkaloids
are cell-cycle
specific
and pro-
duce metaphase arrest in dividing cells. Although alkylating agents act on cells in all phases of the cell cycle, proliferating cells are muchmorevulnerable to their effects. Therefore, it is easily understood that the major side effect of all present anticancer drugs is bone marrow suppression, resulting in neutropenia, anemia, and thrombocytopenia, because bone marrowis also a rapidly proliferating tissue. Some studies on gene therapy of cancer have already
been done, introducing
the
tumor necrosis factor gene (1), granulocyte colony-stim503
S. Kuriyama et al.
ulating
factor
gene (3), and thymidine
kinase gene (4)
by retrovirus-mediated gene transfer. These attemptswithout concern for tissue specificity-may result in the sameside effects as the present cancer chemotherapies because retroviruses seem to be able to infect all kinds of cells as long as they are in a proliferating stage. We have already reported a novel promoter assay using a retroviral vector (9), pIP200 (Fig. 1A). It contains the neomycin phosphotransferase (neo) gene to confer G418-resistance on the transfected cells and the Escherichia coli /3-galactosidase (/5-gal) gene as a reporter with-
surgical
operations can seldom be performed due to the complication of liver cirrhosis. For this therapy, it is essential to kill the hepatoma cells selectively without causing
any damage to actively dividing normal cells, such as
p-pIP200.
bone marrowcells. In the present study, a retroviral vector was generated in which the /3-gal gene was regulated albumin
Ecotropic psi2 packaging cells, kindlyprovided by Dr. Mul-
enhancer
ligan (10), were grown in Dulbecco's modified Eagle's me-
dium (DMEM)c ontaining 10%fetal calf serum (FCS). Trans-
and promoter. Wefirst examined whether the albumin enhancer and promoter could still maintain its appropriate liver-specific expression, without being affected by
5 fL T R H
cells in vivo, were infected.
rus. All plasmids were constructed by standard recombinant DNAtechniques (16). The Alb e/p-pIP200 vector (Fig. IB) was constructed from the previous pIP200vector, whichcontains the neo gene and the /3-gal gene between the LTRs. A 2.2-kb fragment containing the albuminpromoterand its enhancerelement,located between-8.5 kb and -10.4kb upstreamof the promoter, was kindly provided by Dr. Palmiter (14). Conveniently, the fragmentcontains a HindIII site on eachend. It wastherefore digested by Hind III and inserted into the unique Hind III site whichis located between the neo gene and the /3-gal gene in the pIP200vector, resulting in construction of Alb e/
This study is of gene therapy of hepatoma because it
of the liver-specific
retrovirus by a direct gene transfer systeminto the liver, while regenerating hepatocytes, whichmimichepatoma
Vectorconstructionandproductiono frecombinantretrovi-
ase activity.
under control
that quiescent hepatocytes were not infected by the
M A T E R I A L SA N D M E T H O D S
in the two Moloneymurine leukemia virus long terminal repeats (LTRs). Foreign promoters can be subcloned into the Hind III site, located between the two genes. Foreign promoters can also be inserted between the two Sal I sites, deleting a short DNAfragment without any detectable effects on neomycinphosphotransferis a common cancer and, most importantly,
the upstream LTRpromoter, in a variety of cell lines. Wethen investigated whether the /3-gal gene under the control of the albuminenhancerand promotercould be expressed in hepatoma in vivo. We finally confirmed
fection was by the calcium phosphatecoprecipitation method as previously described (8). Stable transformants were se-
lected by addition of G418(Gibco) to a final concentration of 1 mg/ml. Independentcolonies were isolated and expanded. Theculture supernatantwas removed,f iltered througha 0.45jum pore-size filter (Nalgene) and stored at -70°C. It served
S ail S a i l Hind HI I ne o H -H /? -g a l
as a source for infectious retroviral particles.
Cell culture, retroviral infection, and X-Gal staining of the cells. All cell lines were grownin DMEM c ontaining 10% FCS, 100 units/ml of penicillin, and 100 //g/ml of streptomy-
A . p l P 20 0
cin.
5 'L T R H
ne o
H
A lb e /p
B. Alb e/p-PIP
0 -g a l
H 3 'LT R
200
Fig. 1. Structure of the vectors. The pIP200 vector (A) contains the neo gene and /3-gal gene within the two Moloneymurineleukemiavirus LTRs.A Hind III site is located betweenthe two genes, whereforeign promoterscan be subcloned. Foreign promoterscan be also inserted betweenthe two Sal I sites. The longer transcript initiating from the 5' LTRcontains both neo and /3-gal regions; however, /3-gal will not be produced. On the other hand, the ATGcodon of the /3-gal gene in the transcript initiated from the internal promoterwill initiate translation efficiently. Therefore, /3-galactosidasep roductionshowsthe sametissue specificityas that of the promoter inserted, which can be detected by X-Gal staining. The albuminenhancer, located between-8.5 kb and - 10.4 kb upstream, and promoterwas cloned into the HindIII site of the pIP200 vector, resulting in the plasmid Alb e/p-pIP200 (B). 504
Cells were infected by incubating them in the viral supernatant with Polybrene (final cone, 8 //g/ml) at 37°C for 3 to 6 h; next the mediumwasreplaced with virus-free fresh medium, followedby incubation overnight at 37°C.The cells were then fixed with 0.5% glutaraldehyde for 5 min at room temperature. After washing twice with phosphate-buffered
saline
(PBS) with 1 mMMgCh,they were incubated overnight in an X-Gal (5-bromo-4-chloro-3-indolyl-/3-galactoside) reaction mixture consisting of 0.1% X-Gal, 0.01% sodium deoxycholate, 0.02% NP-40, 1 mMMgCl2, 5 mMK3[Fe(CN)6], and 5 mMK4[Fe(CN)6] in PBS without calcium or magnesium (2). Retroviral infection in vivo and histochemistry. Trans-
plantable murine MH-134hepatomacells in ascites (4 x 107 in 0.2 ml of DMEMwith 10% FCS) were injected into the backs of C3H/He mice. Whena subcutaneoustumorexceeded10
mmin diameter, it directly received 0.5 ml of retroviral super-
An Approach for GeneTherapy Targeting Hepatoma. natant with 10//g/ml of Polybrene. Mice were sacrificed 7 days after the injection, and tumors were removed into the
RESULTS AND DISCUSSION
Construction of liver-specific retroviral vector and analysis in vitro. The albumin promoter, which is contained within the 0.3-kb fragment, directs tissue-specific transcription; however, it is not expressed efficiently unless an additional 5' sequence is present. The precise location of the albumin enhancer-like elements has been already examined in transgenic mice (14). It was proven that a region located between -8.5 kb and -10.4kb upstream of the albumin promoter is indispensable for
fixative consisting of 4% paraformaldehyde, 0.2% glutaralde-
hyde, and 0.02% NP-40 in 0.1 M sodium phosphate buffer (pH 7.4). After fixation for 10 min, they were cut into 200-jum sections with a microslicer, followed by post-fixation for 50 min in the same fixative. These slices were then rinsed twice in PBS with 1 mMMgCl2 and incubated for 4h at 37°C in the X-Gal reaction mixture. Partial hepatectomy was carried out in 6- to 7-week-old male C3H/Hemice under ether anesthesia. Thirty-six h after two-thirds partial hepatectomy, 0.3 ml of retroviral supernatant with 10 jug/m\ of Polybrene was injected into the inferior tip of the spleen or directly into the liver. The mice were then sacrificed 7 days later and their livers were removed. Thereafter the same procedure as described for the X-Gal staining of subcutaneous hepatoma was followed, except that the livers were sliced into 100-^m sections instead of 200-^m.
liver-specific
strong expression. This albumin enhancer
and promoter region was cloned into the Hind III site of the pIP200 vector (Fig. 1A). The integrity and orientation of this region in the resulting plasmid, Alb e/ppIP200 (Fig. IB), were confirmed by restriction map-
ping. The longer transcript
initiated
from the 5' LTR
contains both the neo and /3-gal genes; however, the /3-gal gene utilizes the second translational initiation
codon within the transcript for its translational
initia-
;.
^^^ zr+>
%
-*s*
j^
V»
'*/!
T-
à"
-'^ià"""' à"
^
^*&. . %r-iN*~
s*
~ å *v
M
**
Vå
\«w
t:.
^ «*r
Fig. 3. Expression of /3-galactosidase detected by X-Gal staining in non-hepatoma cell lines. NIH 3T3 fibroblasts were infected with the psi2-KS-6 retrovirus (A) or BAGretrovirus (B). Murine B lymphoma cells, A20-2J, were infected with the psi2-KS-6 retrovirus (C) or BAGretrovirus (D). Magnification: A and B, x 140; C and D, x280. 506
An Approach for Gene Therapy Targeting Hepatoma.
onstrate that expression of the /3-gal gene was strictly regulated by the albumin enhancer and promoter within the pIP200 vector. The A20-2J B lymphoma cells, originating from the Balb/c mouse, were also infected with each retrovirus. Again, none of the cells infected with the psi2-KS-6 retrovirus were X-Gal positive (Fig. 3C), while approximately 5 to 10% of the cells infected with the BAGretrovirus were stained blue (Fig. 3D). It is important that the psi2-KS-6 retrovirus did not express /3gal in the B lymphoma cell line, because this suggests that bone marrow cells mayescape the killing effect of the anti-hepatoma retrovirus as long as it is regulated by a tissue-specific promoter. It was thus proven that the al-
bumin enhancer and promoter still retained its tissuespecific expression, even on a retroviral vector. These results indicate that one can easily construct a desirable tissue-specific retroviral vector, which is of great advantage for cancer gene therapy.
å
å
Gene expression in hepatoma in vivo. Experiments were performed to determine whether the /3-gal gene under control of the albumin enhancer and promoter would be expressed in hepatomas in vivo. Transplantable murine MH-134hepatoma cells were injected into the backs of C3H/Hemice, resulting in subcutaneous hepatoma formation. The psi2-KS-6 retrovirus was directly injected into the subcutaneous tumor when it ex-
>- fe
m
X
B
*
"V'V^I fc.
Fig. 4. Expression of /3-galactosidase in hepatoma infected with psi2-KS-6 retrovirus. Transplantable murine MH-134hepatoma cells were injected into the backs of C3H/Hemice, resulting in subcutaneous hepatoma formation. The psi2-KS-6 retrovirus was directly injected into the solid tumor, followed by microslicing to 200-^m thickness 7 days later. The cells expressing the /3-gal gene were detected by X-Gal staining. Magnification: A, 35; B, x 140. 507
S. Kuriyama et al.
ceeded 10 mmin diameter. One week after infection, the tumors were removed and microsliced into 200-^m
sections, followed by X-Gal staining. Several hepatoma cells from the backs of mice receiving the psi2-KS-6
retrovirus injection were X-Gal positive (Fig. 4). However, we could not detect the /3-gal gene expression in connective tissues surrounding hepatoma (data not shown). It was thus shown that we could force expression of a gene of interest under the control of the albu-
-*.
min enhancer and promoter exclusively in hepatoma by direct administration of the retrovirus. Gene expression specific to dividing hepatocytes by direct in vivo gene transfer. Wecontinued experiments to examine whether wecould directly introduce a foreign gene into the liver in vivo. Weinjected psi2-KS-6 retroviruses into murine spleen so that they could reach the liver by way of the portal vein. Seven days later, the mice were sacrificed; their livers were microsliced to
$
»'à"
x"**
å1 x
