Eur. J. Immunol. 2006. 36: 875–886
Immunity to infection
Immunity to infection
Long-term persistence of hepatitis B surface antigen and antibody induced by DNA-mediated immunization results in liver and kidney lesions in mice Xiao-Yuan Zi*1, Yu-Cheng Yao*1, Hai-Ying Zhu1, Jun Xiong1,4, Xiu-Ju Wu2, Nan Zhang1, Yue Ba1,5, Wen-Lin Li1, Xin-Ming Wang1, Jian-Xiu Li1, Hong-Yu Yu3, Xu-Ting Ye1, Joseph T. Y. Lau6 and Yi-Ping Hu1 1 2 3 4 5 6
Department Department Department Department Department Department
of of of of of of
Cell Biology, Second Military Medical University, Shanghai, P.R China Immunology, Second Military Medical University, Shanghai, P.R China Pathology, Second Military Medical University, Shanghai, P.R China Histology and Embryology, Second Military Medical University, Shanghai, P.R China Environmental Health, College of Public Health, Zhengzhou University, Zhengzhou, P.R China Molecular and Cellular Biology, Roswell Park Cancer Institute, Buffalo, NY, USA
DNA-mediated immunization has been recognized as a new approach for prevention and treatment of hepatitis B virus (HBV) infection. However, the side effects of this approach have not been well described. Here we report that DNA-mediated immunization by intramuscular injection of plasmid DNA encoding HBV surface antigen (HBsAg) induced long-term persistence of HBsAg and HBsAg-specific antibody (anti-HBs) in the sera of the immunized BALB/c mice and resulted in liver and kidney lesions. The lesions persisted for 6 months after injection. Lesions were also found in normal mice injected with the sera from immunized mice, and in HBV-transgenic mice injected with anti-HBs antibody, or sera from immunized mice. Furthermore, lesions were accompanied by deposition of circulating immune complex (CIC) of HBsAg and anti-HBs antibody in the damaged organs. These results indicate that long-term persistence of HBsAg and anti-HBs in the immunized mice can result in deposited CIC in liver and kidney, and in development of lesions. The use of DNA containing mammalian replication origins, such as the plasmids used in this study, is not appropriate for human vaccines due to safety concerns relating to persistence of DNA; nevertheless, the safety of DNA-mediated immunization protocols still needs to be carefully evaluated before practical application.
Received 6/9/05 Revised 31/12/05 Accepted 9/2/06 [DOI 10.1002/eji.200535468]
Key words: DNA vaccination Hepatitis B virus Transgenic mice
See accompanying commentary: http://dx.doi.org/10.1002/eji.200635986
Introduction * These authors contributed equally to this study. Correspondence: Yiping Hu, Department of Cell Biology, Second Military Medical University, 800 Xiang-Yin Road, Shanghai 200433, P.R China Fax: +8621-2507-0291 e-mail:
[email protected] Abbreviations: anti-HBs: HBsAg-specific antibody CIC: circulating immune complex EGFP: enhanced GFP HBsAg: HBV surface antigen HBV: hepatitis B virus sALT: serum alanine aminotransferase sAST: serum aspartate aminotransferase f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Hepatitis B virus (HBV) infection is the leading cause of chronic hepatitis, cirrhosis and hepatocellular carcinoma. Approximately 5% of the world's population are chronic HBV carriers [1–3]. Currently, immunization with recombinant protein vaccines against HBV is the most effective and safe means to prevent HBV infection [4]. In addition to conventional vaccination, DNAmediated immunization has been developed as an alternative approach in the prevention and treatment www.eji.de
875
876
Xiao-Yuan Zi et al.
of different infectious diseases, including HBV infection [5–12]. Previous reports indicated that DNA-mediated immunization in mice by a single intramuscular injection of plasmid DNA expressing HBV surface antigen (HBsAg) induces a strong and sustained humoral response for at least 74 weeks [13, 14]. Furthermore, it has also been demonstrated that direct intramuscular injection of plasmid DNA encoding HBsAg can lead to not only expression of the HBsAg protein in the injected muscle tissue, but also secretion of the HBsAg protein into the sera of immunized mice for more than 60 days [15, 16]. Although DNA-mediated HBV immunization can induce both humoral and cellular immune response, the side effects associated with this approach remain poorly understood [5, 17, 18]. Long-term persistence of HBsAg particles and/or its accompanying antibodies are generally considered potential causes of side effects associated with HBV immunization [12, 19, 20]. Furthermore, evidence shows that sedimentation of circulating immune complex (CIC) of HBsAg and HBsAg-specific antibody (anti-HBs) is the pathological basis of organ damage. An example is glomerulonephritis, a human nephritis with CIC sedimentation on the glomerular basement membrane [20, 21]. Fully understanding the pathological basis of DNA-mediated HBV immunization will lead to the development of novel procedures to alleviate the side effects associated with DNA-mediated HBV immunization. To dissect the side effects associated with DNAmediated immunization and its mechanisms, systemic pathology changes were evaluated in mice after intramuscular injection with plasmid DNA containing the HBsAg gene driven by CMV promoter. Pathological lesions were found in the liver and kidney of the injected mice. Sedimentation of CIC was also found in the damaged organs. Using a transgenic (Tg) model developed in our laboratory [22], we further demonstrated that both the presence of HBsAg and anti-HBs antibody in the forms of CIC is the pathological basis of the liver and kidney lesions in the mice immunized with HBV DNA.
Results Expression of the recombinant HBsAg gene in cultured cells Mouse L cells were transfected with the pHBS2-S plasmid DNA or the pcDNA3-EGFP, and the expression of HBsAg in the culture supernatant was assayed by ELISA. HBsAg was detected only in cells transfected with pHBS2-S, but not in cells transfected with pcDNA3EGFP, or in that of untransfected cells. The concentraf 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Immunol. 2006. 36: 875–886
tion of HBsAg protein reached approximately 1000 ng/ mL at 72 h after transfection. Long-term expression of a reporter gene in DNA injected muscle To estimate the persistence in expression of an injected gene in muscle, pcDNA3-EGFP plasmid DNA (100 lg in PBS) was injected into the anterior tibialis muscles of 16 BALB/c mice (6–8 weeks old). Muscle tissue was excised from the anesthetized mice for the detection of enhanced GFP (EGFP) at different time intervals from 2 days to 6 months after injection. At 2 days after the DNA injection, EGFP expression was easily detectable. Maximal expression was observed at 7 and 14 days after the injection. Distribution of the EGFP persists in some muscles at a steady level for 3–9 weeks and then declines to a low but easily detectable level at 3–6 months (Fig. 1). These results showed that expression of the reporter gene EGFP persisted in mouse muscle for more than 6 months after plasmid DNA injection. In vivo expression of HBsAg gene after DNA-mediated immunization To evaluate the persistence and expression of the recombinant HBsAg gene in vivo, 100 lg pHBS2-S plasmid DNA was injected into the anterior tibialis muscle of 16 mice with or without bupivacaine treatment. Total DNA was extracted from the injected muscle tissues at different time points after DNA administration. HBsAg gene sequence persisted in the DNA injected muscle for at least 6 months (Fig. 2A). When sera of the mice injected with pHBS2-S plasmid DNA were collected and the expression levels of the HBsAg protein were analyzed by ELISA, HBsAg protein could be detected as early as 2 days after injection and reached peak levels (2 ng/mL) at 3 weeks after injection. At 10 weeks after injection, the HBsAg protein remained detectable (Fig. 2B). These results showed that the injected HBsAg DNA could remain in mouse muscle, and that the expression of the HBsAg in serum persisted for at least 10 weeks, with fluctuating concentration in the range of 0.1–0.5 ng/mL. In contrast, HBsAg persisted in serum for less than 14 days if the purified HBsAg protein was tail-vein injected directly (Fig. 2C). This result confirmed that the longterm expression of HBsAg protein was induced by the pHBS2-S plasmid DNA.
www.eji.de
Eur. J. Immunol. 2006. 36: 875–886
Immunity to infection
Figure 1. Expression of EGFP in muscle fibers after pcDNA3-EGFP plasmid DNA injection. pcDNA3-EGFP plasmid DNA was injected into muscles of BALB/c mice (n=16) and the entire muscles were excised at different time intervals over a period of 6 months.
Production of circulating anti-HBs antibody in DNA-immunized mice To assess the efficacy of injected pHBS2-S plasmid DNA in inducing a humoral immune response, anti-HBs and anti-preS2 antibodies in the sera were monitored at different times over a period of 6 months by ELISA. AntiHBs antibody was detected as early as 7 days after injection and peaked (over 200 mIU) at 28 days after the injection (Fig. 3, upper panel). The level and kinetics of the anti-preS2 antibody were similar to those of the anti-HBs antibody (Fig. 3, lower panel). Liver and kidney lesions in pHBS2-S DNA-immunized mice To assess the safety of DNA immunization with pHBS2-S DNA, serological analysis was performed for serum alanine aminotransferase (sALT), serum aspartate f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
transaminase (sAST), and serum urea nitrogen. The levels of these markers increased significantly in the pHBS2-S DNA-immunized mice with or without pretreatment of bupivacaine (Fig. 4A). sALT increased significantly at 3 weeks after DNA transfer, and during the following 5 weeks, maintained at a high level that peaked at 9 weeks. sAST also increased at 3 weeks after injection but reached a peak immediately and then dropped gradually. The serum urea level started to increase at 5 weeks and continued to increase gradually with no obvious peak until 11 weeks after the DNA injection. These results indicated that the liver and kidney of the immunized mice were damaged with time as a result of the DNA immunization. At 10 weeks after the DNA injection, liver, kidney, spleen, heart and lung were obtained, and paraffinsections were prepared to assess histological changes. It was found that liver and kidney of the pHBS2-Simmunized mice developed histological lesions, includwww.eji.de
877
878
Xiao-Yuan Zi et al.
ing degeneration of hepatocytes (balloon cells), disappearance of sinusoid, disarray of lobules in the liver (Fig. 4B), disruption of the glomerular basement membrane, sedimentation of immunocomplex on basement membrane, and swelling of the renal tubular epithelial cells in the kidney (Fig. 5). No obvious histological change was detected in other organs of the immunized mice or in any of the organs examined in control mice (Fig. 4B, and data not shown). These
Figure 2. Long-term persistence of pHBS2-S DNA in muscle fibers and expression of HBsAg in vivo. (A) After gene transfer, muscle fibers were obtained at different times. DNA from fiber cells was used as template for PCR. A specific 444-bp band of the S gene was amplified out of muscle fibers from mice immunized with pHBS2-S plasmids DNA. Lane 1: 100-bp DNA ladder; lanes 2–8: muscle fibers from mice immunized after 2, 4, 6, 8, 9, and 10 weeks and 6 months; lane 9: pHBS2-S plasmid DNA; lane 10: muscle fibers from mice injected with pcDNA3EGFP; lane 11: muscle fibers from normal mice. (B) Detection of HBsAg expression in serum of the mice injected with plasmid pHBS2-S (n=16). HBsAg was measured at different times after injection in a period of 6 months (in triplicate at each time point). Each point represents mean titer value. Error bars represent SEM. (C) Longevity of purified HBsAg in vivo. Recombinant HBsAg was injected into five mice via tail vein, and HBsAg was detected at different times after injection over a period of 3 weeks. The persistence of recombinant HBsAg was less than 14 days. f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Immunol. 2006. 36: 875–886
results demonstrated that liver and kidney lesions developed in pHBS2-S-immunized mice, in which serum HBsAg and anti-HBs antibodies remained in serum for a long time. Pathological lesions result from HBsAg and anti-HBs antibody complex To explore the cause of liver and kidney lesions in pHBS2-S DNA-immunized mice, purified anti-HBs antibody, mouse IgG2b isotype control, sera from normal mice and from pHBS2-S and pcDNA3-EGFP DNAimmunized mice at 4 weeks after immunization were injected into HBV-Tg mice or normal mice. Purified HBsAg and a mixture of HBsAg and anti-HBs antibodies were also injected into normal BALB/c mice. The levels of sALT and sAST in these mice were analyzed after daily injection for 9 days. The results indicated that both sALT and sAST started to increase at 3 days after the antibody injection, and remained at high levels for the remaining
Figure 3. Antibodies produced in pHBS2-S-immunized mice. Serum obtained from 16 mice injected with plasmids pHBS2-S or pcDNA3-EGFP DNA at different times over a period of 6 months and diluted (1:50). Antibodies were screened with adr subtype HBsAg or preS2 antigen bound to the solid phase in triplicate at each time point. Each point represents mean titer value. Error bars represent SEM. www.eji.de
Eur. J. Immunol. 2006. 36: 875–886
Immunity to infection
Figure 4. Liver and kidney lesions in mice injected with pHBS2-S DNA. (A) Kinetics of sALT, sAST and serum urea (sUrea) in the sera of the pHBS2-S-immunized mice and the pcDNA3-EGFP-injected mice. Each point represents mean values of eight mice. Error bars represent SEM. (B) Ballooning degeneration of hepatocytes is observed in mice immunized with pHBS2-S DNA pretreated with bupivacaine (panel ii) or without treatment (panel iii) and normal in mice injected with pcDNA3-EGFP plasmid DNA (panel i) at 10 weeks after injection (original magnifications: 400).
Figure 5. Detection of CIC and HBsAg in the damaged kidneys. Immunohistochemistry and ultrastructural analysis show that the HBsAg and anti-HBs antibody form CIC, which accumulate on the glomerular basement membrane of the pHBS2-S DNAimmunized mice at week 10 after DNA injection with bupivacaine treatment (A panels i, iii) or without bupivacaine treatment (A panels ii, iv), HBV-Tg mice after injection of purified antibody injection (B panels i, iii) or serum from DNA-mediated immunized mice from week 4 (B panels ii, iv), normal BALB/c mice after injection of serum from DNA-mediated immunized mice from week 4 (C panels ii, iv) or complex of HBsAg and anti-HBs antibody (D panels ii, iv). None are seen in BALB/c mice injected with purified anti-HBs antibody (C panels i, iii) or purified HBsAg particles (panels i, iii). Bar represents 10 lm and arrows indicate CIC (panels i, ii). Original magnifications 400 (panels iii, iv). f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji.de
879
880
Xiao-Yuan Zi et al.
5 days of the protocol in HBV-Tg mice injected with the purified antibody. Elevated sALT and sAST were not observed in the normal mice injected with the purified antibody (Fig. 6A). We also found that sALT and sAST levels increased in both normal mice and Tg mice after injection of sera from pHBS2-S DNA-immunized mice (Fig. 6A). However, sALT and sAST levels increased much earlier in HBV-Tg mice than in normal BALB/c mice. The change in sALT and sAST levels in normal mice injected with the complex of purified HBsAg and antibody was similar to that in the mice injected with serum from DNA-immunized mice (Fig. 6A). Consistent with changes in sALT and sAST, the histological analysis confirmed liver damage in HBV-Tg mice injected either with serum from pHBS2-S DNA-
Eur. J. Immunol. 2006. 36: 875–886
immunized mice or with purified antibody, and in normal mice injected with serum from DNA-immunized mice or with complex of HBsAg and purified antibody. The disarray of liver cell plates, swollen (ballooned) hepatocytes and disappearance of sinusoid were also visible in the liver of these mice (Fig. 6B panels i, ii; C panels i, ii). These results demonstrated that pathological hepatic and kidney lesions, such as those observed in the pHBS2-S DNA-immunized mice, could be induced directly by the persistent presence of the HBsAg and anti-HBs antibody in serum, which could form the CIC. Pathological lesions and abnormal changes of sALT and sAST were not detected in: (i) the normal BALB/c mice injected with purified anti-HBs antibody or HBsAg
Figure 6. Liver lesions were induced in HBV-Tg mice and normal mice by injection of serum from DNA-mediated immunized mice. (A) Kinetics of sALT and sAST in HBV-Tg mice injected with sera from DNA-immunized mice (IST) or purified anti-HBs antibody (ABT) and normal mice injected with sera from DNA-immunized mice (ISN), purified anti-HBs antibody (ABN), HBsAg particles (AgN) or both HBsAg and purified antibody (AAN). Four mice were used every group and the experiments were performed in triplicate at each time point over a period of 9 days. Each point represents mean value. Error bars represent SEM. (B) Liver lesions are observed in HBV-Tg mice at 9 days after injection with serum from pHBS2-S DNA-immunized mice from week 4 (panel i) or purified anti-HBs antibody (panel ii), and none are seen in HBV-Tg mice injected with serum from pcDNA3-EGFP DNA-immunized mice (panel iii) or mouse IgG2b isotype control (panel iv) (n=4, original magnifications panels i, ii 400; panels iii, iv 200). (C) Ballooning degeneration of hepatocytes is observed in normal mice at 9 days after injection with serum from pHBS2-S DNAimmunized mice from week 4 (panel i) or mixture of 4 ng HBsAg and 400 mIU anti-HBs antibody (panel ii), and none is seen in mice injected with purified anti-HBs antibody (panel iii) or purified HBsAg particles (panel iv) (n=4, original magnifications 400). f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eji.de
Eur. J. Immunol. 2006. 36: 875–886
(Fig. 6A; C panels iii, iv), which indicated that HBsAg or anti-HBs antibody alone was not sufficient to cause organ lesions, (ii) the HBV-Tg mice injected with serum from pcDNA3-EGFP DNA-immunized mice (Fig. 6B panel iii, and data not shown), which indicated that the lesions were induced directly by an immune complex of HBsAg and anti-HBs antibody and that cytokines might have no or weak roles in this process, and (iii) the HBVTg mice injected with mouse IgG2b isotype control (Fig. 6B panel iv, and data not shown), which indicated that the lesions induced by the immune complex of HBsAg and anti-HBs antibody were specific. Detection of the CIC To detect CIC formation in the damaged organs, kidneys were obtained from pHBS2-S DNA-immunized mice, from normal mice injected with purified HBsAg, or antiHBs antibody, or complex of HBsAg and the antibody or sera of the immunized mice, and from HBV-Tg mice injected with antibody or antisera. Ultrastructural change was analyzed by transmission electron microscopy. Sedimentation of immune complex was detected on the basement membrane of the pHBS2-S DNAimmunized mice with or without bupivacaine treatment (Fig. 5A panels i, ii), the HBV-Tg mice injected with serum from the DNA immunized mice or with the purified antibody (Fig. 5B panels i, ii), and the normal mice injected with the serum from DNA-immunized mice or with mixture of HBsAg and purified antibody (Fig. 5C panel ii; D panel ii). No deposited immune complex was found on the glomerular basement membrane of the normal mice injected with antibody or HBsAg alone (Fig. 5C panel i; D panel i). To assess the composition of the deposited immune complex, immunohistochemistry was performed on the immune complex deposited in renal tissue. The results showed that the complex contained HBsAg (Fig. 5A panels iii, iv; B panels iii, iv; C panel iv; D panel iv). Furthermore, both HBsAg and anti-HBs antibody were detected in the renal tissue by the Western blot analysis. CIC was also present in the sera of mice with organ lesions (data not shown). These results confirmed that CIC was formed by HBsAg and anti-HBs antibodies, and that the liver and kidney lesions resulted from the deposition of CIC. Long-term persistence of liver and kidney lesions in HBV DNA-immunized mice To evaluate the persistence of liver and kidney lesions in pHBS2-S DNA-immunized mice, sALT, sAST and serum urea were measured at 6 months after DNA immunization. It was found that sALT, sAST and serum urea remained elevated despite significant decline in the f 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Immunity to infection
levels of serum HBsAg and anti-HBs antibody (Fig. 7A panels iii, iv; B panel iv). Moreover, the sALT level even reached a second peak higher than the initial peak at 9 weeks after DNA immunization. This suggested that the organ damage was sustained and perhaps even aggravated over time. Histological analysis showed that the histopathological lesions, including ballooned hepatocytes and deposited CIC on the renal basement membrane, were still present at 6 months after the DNA injection, and the damage was not significantly mitigated after week 10 (Fig. 7A panel ii; B panel ii). Western blot and immunohistochemistry assay also showed that the HBsAg was present at a low level in renal tissues (Fig. 7B panel i; B panel iii). These results indicated that the pathological lesions caused by pHBS2-S DNA immunization could persist for at least 6 months in the immunized mice.
Discussion DNA-mediated immunization can induce both humoral and cellular immune responses since the antigen can be continuously synthesized in vivo [14, 15, 23]. Therefore, the safety issues have to be carefully considered before this method can be used clinically [5, 17, 18]. In this study, we provide experimental evidence for a potential side effect associated with DNA immunization, and the mechanism explaining the cause of the side effect. We found that a single intramuscular injection with pHBS2-S plasmid DNA encoding HBsAg, containing the backbone of the pcDNA3, can induce long-term persistence of HBsAg protein and anti-HBs antibodies in the circulation of the immunized BALB/c mice, and result in liver and kidney lesions due to the deposition of the CIC. Based on the expression kinetics of the reporter gene EGFP, our results revealed that an injected gene can be continuously expressed in the muscle tissue for at least 6 months after DNA injection (Fig. 1). EGFP expression peaked at 7–14 days and declined to a low but easily detectable level at 9 weeks post injection. Previous studies have also showed that direct gene transfer into mouse muscle results in long-term and stable expression of an exogenous protein. For example, expression of luciferase lasts for over 19 months [24], of b-galactosidase for over 32 weeks [25], of CAT protein for over 4 months [24], of human erythropoietin (Epo) for over 40 weeks [25], of fibroblast growth factor 1 (FGF1) for over 9 months [26] and of factor IX protein for several months [27, 28] after DNA injection. These data indicate that muscles can take up plasmid DNA and express the foreign protein for a long time, which can therefore be used in DNA immunization. www.eji.de
881
882
Xiao-Yuan Zi et al.
Previous studies have showed that, after intramuscular injection of plasmid DNA encoding secreted proteins including HBsAg, the muscles can not only
Eur. J. Immunol. 2006. 36: 875–886
take up plasmid DNA and express foreign proteins, but also can secret the proteins into the serum [15, 29]. In the present study, we have shown that circulating HBsAg
Figure 7. Pathological analysis of mice 6 months after pHBS2-S DNA injection. (A) Ballooning change of hepatocytes in mice injected with pHBS2-S DNA (A panel ii) and normal in mice injected with pcDNA3-EGFP (A panel i). sALT and sAST level are higher in mice injected with pHBS2-S DNA than mice injected with pcDNA3-EGFP (A panels iii, iv; n=4,*p
