Direct Injection of Hepatitis B Virus DNA into Liver Induced Hepatitis in Adult Rats*

Hepatitis B virus (HBV) surface antigen (HBsAg) genes were injected directly into the liver of adult rats with non-histone chromosomal protein high mobility group 1 by the hemagglutinating B virus of Japan (Sendai virus)-liposome method (Kato, K., Nakanishi, M., Kaneda, Y., Uchida, T., and Okada, Y. (1991) J. Biol. Chem. 266, 3361-3364). Immunohistochemical analysis showed that HBV surface antigen was ex- pressed by the hepatocytes in vivo. On successive injections of the HBsAg genes, the antibody to HBV surface polypeptides was produced in the rats, and characteristic pathological changes of lymphocytic fo- cal necrosis and denaturation of hepatic cells were observed in the liver of all the rats. We conclude that hepatitis is caused by the direct injection of HBsAg genes.


Hepatitis B virus (HBV)' infection is one
of the most widespread viral infections of humans, and HBV causes acute and chronic hepatitis and hepatocellular carcinoma (1). There have been many studies on HBV infection, replication, and virion production in cultured cells (in uitro) (2)(3)(4)(5). However, it is generally thought that viral hepatitis is mediated primarily by immune defense mechanisms (6,7), so the pathogenesis of HBV hepatitis should be examined in living animals (in uiuo). Woodchucks, ground squirrels, and Peking ducks are infected with their respective HBV-like viruses (hepadnauiruses), but those viruses are different from human HBV biologically (8). Transgenic mice with the HBV whole genome (9) have been produced, but they showed neither immune responses nor liver inflammation. Transgenic animals usually do not show an immune reaction against products of an introduced gene because of the immunological tolerance. Recently, HBV surface protein transgenic mice primed with spleen cells from a donor that had been immunized with a * This work was supported by grants from the Ministry of Education, Science and Culture of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. I To whom correspondence should addressed. recombinant vaccinia virus containing the coding region for the HBV major envelope protein were found to display biochemical evidence of liver cell injury (10). However, it is still unknown whether HBV surface antigen can induce immune and inflammatory responses in host animals without transplantation.
Another novel approach is direct delivery of the HBV gene into the liver in uiuo, and there are several recent reports of administration of functional genes to living animals by direct injection (11). With this system, systemic responses can be observed in host animals when exogenous DNAs are expressed in their organs. We established an efficient gene delivery system by direct injection using HVJ-liposomes that showed high gene transfer activity without cytotoxicity in vivo (12)(13)(14). Functional DNA and non-histone chromosomal protein high mobility group 1 (HMG1) co-encapsulated in liposomes were introduced into the cytoplasm of rat liver cells by HVJmediated membrane fusion. The migration of HMGl contributed to expression of the genes efficiently in hepatocytes of adult rats (12,14).
In this work, the HBV surface antigen genes were introduced into adult rat liver by the HVJ-liposome method, and expression of HBsAg in the liver was shown to induce immune and inflammatory responses in the animals.

MATERIALS AND METHODS
Construction of Plasmids-pAct-LMS-pUC-Act-c-myb plasmid (14,15) was restricted with NcoIIXbaI to remove c-myb and inserted into an XhoI site with linker. The HindIII/EcoRI fragment containing HBsAg (ORF S, nucleotides 2687-0/3215-1856) of pSV,-LMS (14) was inserted into the XhoI/EcoRI site in this vector by filling Hind111 and XhoI ends with T 4 DNA polymerase (7.0 kilobases). pAct-LMS contains the entire envelope region open reading frame consisting of three translation initiation codons, which represent the N termini of the large, middle, and major surface (S) polypeptides downstream of the chicken 0-actin promoter. pAct-MS (6.5 kilobases) was described previously (14).
Western Analysis of Intrahepatic H B V Envelope Proteins in Rat Hepatoma Cell-H4-II-E-C3 cells (rat hepatoma; ATCC CRL 1600) were inoculated into 100-mm Petri dishes (2 X lo6 cells/dish) and incubated for 1 day in Dulbecco's modified Eagle's minimum essential medium supplemented with 5% horse serum and 5% fetal calf serum. pAct-LMS was added a t 10 pg/dish according to the method reported previously (16).
The transfected cells were harvested, incubated a t 4 'C for 30 min with gentle shaking in 2% Nonidet P-40-phosphate-buffered saline (PBS; 137 mM NaCl, 3 mM KCI, 8 mM Na2HP04, 1 mM KH,P04) containing a protease inhibitor mixture composed of pepstatin, chymostatin, leupeptin, and aprotinin a t 10 pg/ml and 1 mM phenylmethanesulfonyl fluoride, 1 pg/ml o-phenanthroline, and 10 ,LM benzamidine (Sigma), and centrifuged a t 15,000 rpm a t 4 "C for 5 min. The supernatant was denatured in Laemmli's sample buffer and subjected to electrophoresis in 10% polyacrylamide gel. Proteins were transferred to nitrocellulose filters (Schleicher & Schuell; BA85) by semi-dry electroblotting (ATTO), and the filters were washed twice with 0.05% Tween 20-PBS a t room temperature for 30 min. The filters were then treated with anti-pres, antibody (5520, mouse monoclonal antibody, Japan Tanner) and with Iz5I-labeled rabbit antimouse IgG (1 pCi (1.3 pg of IgG)/ml), iodinated by the Bio-Rad enzymatic procedure. The filters were then washed three times with 5% skim milk, 0.05% Tween 20-PBS, twice with 0.05% Tween 20-PBS, and once with PBS. Anti-pres, antibody recognized large and middle surface proteins, but not major surface protein.
Detection of H B V Envelope Protein by Immunohistochemistry-HVJ-liposomes containing pAct-MS, pAct-LMS, and non-histone chromosomal protein HMGl were injected under the perisplanchnic A. K b , kilobases. B, Western analysis of intrahepatic HBV envelope proteins in rat hepatoma cells. pAct-LMS was transfected into H4-11-E-C3 cells (rat hepatoma; ATCC CRL 1600) by the calcium phosphate precipitation method (16). HBV surface proteins in the transfected cells were analyzed by immunoblotting using anti-pres2 antibody followed by '@I-rabbit anti-mouse IgG. Lane I , cytoplasmic fraction derived from H4-II-E-C3 cells transfected with pAct-LMS; lane 2, cytoplasmic fraction derived from H4-II-E-C3 cells transfected with pUC19. membrane of the liver of rats (Sprague-Dawley, 7 weeks old). The rat liver was perfused with 4% paraformaldehyde containing 0.1 M phosphate buffer, pH 7.4. After fixation, the tissue was sectioned at 6-pm thickness in a cryostat (Miles). Sections were developed with anti-presI antibody (5520) and examined with a commercial alkaline phosphatase immunoreaction kit (Vectastain, Vector Laboratories, Inc.) and alkaline phosphatase substrate kit (Black; Vector Laboratories, IC.).
Preparation of HVJ-Liposomes-HVJ-liposomes were prepared as described previously (14). Phosphatidylserine, phosphatidylcholine, and cholesterol were mixed in a weight ratio of 1:4.   Table I for rat numbers) were determined with a commercial enzyme immunoassay kit (Abbott). of lipids and 10-40 gg of encapsulated DNA) in BSS with 1 mg/ml glucose and 1 mM CaCll were injected under the perisplanchnic membrane of the liver of rats (Sprague-Dawley, 7 weeks old). Animal care was in accordance with institutional guidelines, and rats were anesthetized with ether before all treatments.
Assay of HBsAg and Anti-HBs Antibody in Rat Serum-After injection of HVJ-liposomes containing pAct-LMS, pAct-MS, and

TABLE I
Peak levels of HBsAg and anti-HBs antibody detected in the serum during 7 days after HVJ-liposome injection A HVJ-liposomes containing pAct-LMS, pAct-MS, and HMGl were injected into adult rat liver on day 0 (lst), day 7 (2nd), day 14 (3rd), and day 21 (4th). In the intervals between injections, the levels of HBsAg (ng/ml) and anti-HBs (pglml) antibody secreted into the rat serum were monitored. Maximum levels detected in these intervals are shown. The serum levels of HBsAg and anti-HBs antibody in rats treated with HVJ-liposomes containing pUC19 and HMGl were < 0.2 ng/ml and < 0.5 pg/ml, respectively. HMG1, all rats were bled at the indicated intervals, and the sera were analyzed for the amounts of HBsAg and anti-HBs antibody using a commercial enzyme immunoassay kit (Abbott). Purified HBsAg and anti-HBs antibody (HB7-2, mouse monoclonal antibody, the Chemo-Sero-Therapeutic Research Institute, Japan) were used as standards.
Pathological Anulysis-HVJ-liposomes containing pAct-LMS, pAct-MS, and HMGl were injected under the perisplanchnic membrane of the liver two to four times at 7-day intervals. Seven days after the last injection, the rats were perfused with 4% paraformaldehyde containing 0.1 M phosphate buffer, pH 7.4. Rat liver was embedded in paraffin, sectioned at 3 pm, and stained with hematoxylin and eosin by standard methods.

RESULTS AND DISCUSSION
The outer membrane of HB virus consists of host lipid and HBV major, middle, and large envelope proteins within a large coding region that has three in-phase translation start codons. The mRNAs of large, middle, and major S are transcribed from the pres1, pres2, and S initiator codons, respectively. When pBRneo-LMS, carrying the BglII fragment encoding surface proteins of native human HBV (ORF S , nucleotides 2296-0/3215-1856) cloned into the BarnHI site of pBRneo, was transcribed into BNL CL.2 cells (normal mouse liver; ATCC TIB 73), H4-II-E-C3 cells, and Huh-7-cl4 cells (human hepatoma) by the calcium phosphate method (16), the level of expression of HBsAg in rodent cells was less than one-twentieth of that in Huh-7-cl4 cells (8). So we constructed pAct-LMS, which includes a large coding region containing the pres1, pres2, and S initiator codon, under the control of chicken 8-actin promoter (Fig. Ut).
Western blot analysis revealed that rat hepatoma cells 0 22073 I transfected with pAct-LMS by the calcium phosphate precipitation method produced HBsAg in the cytoplasm in vitro (cultured cells). When anti-preSz antibody was used, the large (p39, gp42) and middle (gp33, gp36) HBV envelope proteins were produced (Fig. lB), while the major surface proteins were not detected by immunoblotting with anti-major S antibody. When pAct-MS (14) was transfected into rat hepatoma cells, major S proteins were expressed (data not shown).
For in vivo gene delivery, we constructed HVJ-liposomes containing pAct-MS, pAct-LMS, and HMGl for expression of all three HBV envelope proteins in liver cells of rats. As described previously (14), we succeeded in expressing the pgalactosidase gene under the control of the chicken &actin promoter in hepatocytes of adult rats in vivo by this method.
Two days after the injection of HVJ-liposomes containing pAct-LMS, pAct-MS, and HMGl under the perisplanchnic membrane of rat liver, the liver was fixed, and sections of the tissue were developed with anti-pres, antibody. Immunohistochemical analysis showed that in these rats HBsAg was present both on the cell surface and in the cytoplasm of the hepatocytes (Fig. 2 A ) . The immunopositive hepatocytes appeared to be intact.
The level of HBsAg in rat serum was monitored by enzyme immunoassay after introduction of pAct-LMS and pAct-MS. HBsAg was detected during the first 3 days with a maximum of 1 ng/ml on day 2 after the injection (Fig. 3, open symbols). The levels of HBsAg in the serum after injection of pAct-LMS and pAct-MS were lower than that after injection of pAct-MS alone, described previously (14). Next the antibodies to HBV surface proteins in rat serum were assayed. Antibodies were detected on day 4 after injection of HVJ-liposomes (Fig. 3, closed symbols). Seven days after the first injection of pAct-LMS and pAct-MS, both plasmids were again injected into the same rats by the HVJ-liposome method. After the second injection, the secretion of anti-HBs antibody into the serum continued until the time of death with a maximum of 13.3 pg/ml. In the first period (day 0 to day 7 after the first injection), the serum HBsAg levels reached 0.5-1.0 ng/ml in all the rats. After the second period (day 8 to day 28), anti-HBs antibody was detected in most of the serum samples, while HBsAgs were not detected in any (Table I). It is likely that anti-HBs antibody functions as a competitor of the detection of HBsAg in rat serum. When HVJ-liposomes containing pUC19 were injected by the same method, no HBsAg nor antibody was detected in the sera.
The pathological changes of liver tissues in the rats were investigated. For this, rats were put to death and perfused with 4% paraformaldehyde 7 days after the second injection (day 14). Liver tissue of rats sectioned and stained with hematoxylin and eosin showed focal necrosis characterized histologically by infiltration of lymphocytes and other mononuclear cells and hepatocytes denaturation (arrowhead) in the parenchyma (Fig. 4, A and B ) . Infiltration of mononuclear cells was observed in all the rats examined. In liver tissues of rats A-1 and A-2 (Table I) 7 days after the fourth injection, the late phase (day 28) of liver cell injury was characterized mainly by inflammation of Glisson's sheaths (Fig. 4, C and  D). These pathological changes in the liver were not observed in rats treated with pUC19 by the same method. Transition from focal necrosis (after two injections) to glissonitis (after four injections) resembled the histological change from acute to chronic inflammation in man.
Thus, without transplantation of spleen cells immunized with HBV surface antigen, the expression of HBsAg in rat liver in situ was able to produce anti-HBs antibody and to induce pathological change of the liver. We think that the difference in the quantities of antibody in the sera of the rats may reflect differences in the level of activation of major histocompatibility complex class I-restricted cytotoxic T lymphocytes. Therefore, we are now examining the activation of cytotoxic T lymphocytes in rats transfected with HBV surface antigen genes.
In these rats, liver inflammation was restricted to the lobules that received injections and was not distributed through total liver. In fact, the serum glutamic-pyruvic transaminase activities in the sera of these rats were slightly higher than those in rats treated with pUC19 but lower than that observed in human hepatitis. However, the pathological changes in the liver, such as focal necrosis, inflammation of Glisson's sheaths, and liver cell degeneration, mimicked some changes in human HBV infection. Therefore, these rats treated by the HVJ-liposome method and expressing HBV surface proteins are a potential model of human HBV-induced hepatitis. This system will be useful for elucidating the mechanism of HBV-induced liver injury and for developing suitable therapeutic treatments.
Thus, this direct gene delivery system will provide a new way for studying cellular functions in various organs, establishing animal models of human diseases and postnatal gene therapy in vivo (11).