Induction of cytotoxic T lymphocyte responses against hepatitis delta virus antigens which protect against tumor formation in mice

Vaccine 20 (2002) 170–180

Induction of cytotoxic T lymphocyte responses against hepatitis delta virus antigens which protect against tumor formation in mice Christian Mauch a , Christian Grimm a , Stefan Meckel a , Jack R. Wands c , Hubert E. Blum a , M. Roggendorf b , Michael Geissler a,∗ a

Department of Medicine II, University Hospital Freiburg, Hugstetter Strasse 55, D-79106 Freiburg, Germany b Institute of Virology, University Essen, Essen, Germany c Liver Research Center, Brown University School of Medicine, Rhode Island Hospital, Providence, USA Received 16 January 2001; received in revised form 29 May 2001; accepted 18 June 2001

Abstract The cellular immune response is a crucial defense mechanism against hepatotropic viruses and in chronic viral hepatitis prevention. Moreover, hepatitis delta virus (HDV) immunogenicity may be an important component in the development of prophylactic and therapeutic vaccines. Therefore, we evaluated the immunogenicity of the small (HDAg) or large delta antigen (LHDAg) to be used as a DNA-based vaccine. We immunized different mouse haplotypes, determined cellular immune responses, and tested protection of animals against tumor formation using syngeneic tumor cells stably expressing the delta antigens. Both LHDAg and HDAg primed CD4+ and CD8+ T cell immunity against both forms of delta antigens. CD8+ T cell frequencies were about 1% and antigen-specific CD8+ T cells remained detectable directly ex vivo for at least 35 days post-injection. No anti-delta antibody responses could be detected despite multiple detection systems and varied immunization approaches. We observed protection against syngeneic tumor formation and growth in mice immunized with DNA plasmids encoding secreted or intracellular forms of HDAg and LHDAg but not with recombinant HDAg establishing the generation of significant cellular immunity in vivo. Both CD4+ and CD8+ T cells were required for antitumoral activity as determined by in vivo T cell depletion experiments. The results indicate that DNA-based immunization with genes encoding LHDAg and HDAg induces strong T cell responses and, therefore, is an attractive approach for the construction of therapeutic and prophylactic T cell vaccines against HDV. © 2001 Elsevier Science Ltd. All rights reserved. Keywords: CTL; HDV; Tumor

1. Introduction The hepatitis delta virus (HDV) is a defective RNA virus which can replicate only in the presence of hepatitis B virus (HBV). HDV/HBV co-infections or HDV superinfections of HBV infected patients are responsible for a wide range of acute and chronic liver diseases. Outbreaks of frequently fatal epidemics have been reported in several parts of the world, including developed countries. Although the frequency of HDV infection has declined in Italy and Taiwan [1–3], HDV infection is still a major problem in some parts of South America and in high risk populations such as intravenous drug abusers. Superinfection of HBV carriers often progresses rapidly to liver cirrhosis and hepatocellular carcinoma. Moreover, therapy of chronic delta hepatitis is ∗ Corresponding author. Tel.: +49-761-270-3509; fax: +49-761-270-3610. E-mail address: [email protected] (M. Geissler).

problematic. Interferon-␣ and other anti-viral agents, such as ribavirin, levamisole, and lamivudine appear ineffective against HDV replication. In addition, anti-viral therapy against chronic HBV infection is effective in only 30–50% of patients and blocking HBV replication by the nucleoside analogue lamivudine does not affect HDV replication [4]. Importantly, the activity of liver disease and cell injury is not affected by lamivudine because it has little effect on hepatitis B surface antigen (HBsAg) synthesis that is required for HDV packaging and export. Therefore, alternative approaches for the specific inhibition of HDV replication are necessary to prevent HDV superinfection of HBV carriers [5]. HDV consists of a single-stranded, circular RNA genome of 1.7 kb that is closely associated with the only virus-encoded protein, hepatitis delta antigen. The delta antigen is a ribonucleoprotein which is enclosed within an envelope consisting of HBsAgs, which are provided by the helper HBV. Hepatitis delta antigen is usually present

0264-410X/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 2 6 4 - 4 1 0 X ( 0 1 ) 0 0 2 5 2 - 3

C. Mauch et al. / Vaccine 20 (2002) 170–180

as two related protein species, a 24-kDa small form (hepatitis delta antibodies (HDAg)) and a 27-kDa large form (LHDAg), that are identical except for a 19-amino acid carboxy-terminal extension in LHDAg [6]. This 19-amino acid extension arises from a specific RNA editing event during genome replication, resulting in elimination of a termination codon and extension of the open reading frame [7–9]. These two proteins have similar biochemical properties but differ in their function during virus replication [6,10–12]. Some observations indicate that a specific immune response may play a role in protection against HDV infection as well as in the pathogenesis of liver injury. In this regard, it has been reported that HDV reinfection in HBV carrier chimpanzees that had recovered from HDV infection was characterized by a significantly reduced HDV replication [13]. This suggests that the first contact with HDV may be followed by a partial immunity that is able to control viral replication. Anti-hepatitis delta antibodies (anti-HDAg) as a specific humoral immune response do not protect even though these antibodies are commonly detected in infected subjects. Indirect evidence for the protective role of T cells in HDV infection has been obtained from both animal models and patients. It has been shown that woodchucks infected with woodchuck hepatitis virus (WHV) are partially protected from subsequent challenge with HDV when immunized with recombinant forms of HDAg in the absence of any detectable humoral response [14]. Furthermore, HDV-infected woodchucks show an increased level of viremia after treatment with cyclosporin A, a specific inhibitor of T cell-mediated responses [14]. First direct evidence that T cells play an important part in controlling HDV infection in humans was provided recently by Nisini et al. who demonstrated that CD4+ T cell activity in chronically HDV infected patients correlates with the decrease of HDV-induced disease activity [15]. The same group demonstrated that HDAg may undergo extracellular processing suggesting that the generation of immunogenic epitopes directly by serum proteases could play a role in the immune response against HDV during infection [16]. The induction of cytotoxic T cell (CTL) responses to HDAgs in patients, however, has not been demonstrated yet. Furthermore, the role of CTL responses against HDAgs in the recovery of HDV superinfection is unknown. Definition of the immunogenicity of HDAg and LHDAg at the CTL level, however, is required for the development of new immunotherapeutic and prophylactic approaches. Therefore, we performed experiments using the DNAbased immunization approach [17] to induce and characterize HDV-specific CTL and helper T (TH) cell responses in different mouse haplotypes and to determine whether cellular immune responses generated by this immunization would protect animals against tumor formation using syngeneic tumor cells stably transfected with cDNAs encoding HDAg and LHDAg, respectively.

171

2. Materials and methods 2.1. Plasmids Plasmids expressing HDAg and LHDAg under the control of a CMV promotor and designated pSS32 and pSS33, respectively, were a generous gift from D. Lazinski, Tufts University, Boston, MA, USA. In addition, the cDNA encoding HDAg was cloned into the eukaryotic expression plasmid pcDNA3 (Invitrogen, San Diego, CA, USA) as follows: the infectious clone of HDV (genotype 1) was obtained by American Type Culture Collection (ATCC, Rockville, MD, USA; accession number M21012) and amplified by PCR with primers HDA-2 (nucleotide (nt) 986–1005, 5 -TGC CGC CTC TAG CCG AGA TG-3 ) and HDA-1 (nt 1615–1596, 5 -GCG GAT CGG CTG GGA AGA GT-3 ). The PCR products were cloned into pCRII vectors (Invitrogen, San Diego, USA) according to the manufacturer’s instructions. One clone was selected and sequenced to verify the correct nucleotide sequence of the PCR product. The fragment containing the gene for HDAg (p24) was isolated by digestion with HindIII and XhoI and inserted between the corresponding sites of the pcDNA3 vector (Invitrogen). The generated plasmid, peHDAg, contained the gene encoding HDAg (p24) under the control of the cytomegalovirus promotor. The integrity of the clone was verified by sequencing. In order to obtain potentially more immunogenic delta antigens, the expression plasmids pSec␦33 and pSec␦32 were created by PCR-subcloning the HDAg and LHDAg encoding cDNA-fragments in frame into the eukaryotic secretion vector pSecTagB (Invitrogen, San Diego, CA) which contains the signaling sequence from the V-J2-C region of the mouse Ig-kappa light chain. A plasmid expressing IL-12p70 (pApIL-12p70) and the GM-CSF expressing plasmid pRJB-GM have been described recently [18,19]. IL-18 expressing plasmid (pCI-1sIL-18) was a generous gift from J. Reimann, University of Ulm, Germany. 2.2. Mice Female Balb/c and DBA-2 (both H-2d ) and C57BL/6 (H-2b ) mice obtained from Charles River Labs (Wilmington, MA, USA) were maintained according with the National Association for Laboratory Animal Sciences and used between the age of 10–25 weeks. 2.3. Genetic immunization Facilitated DNA immunization was performed by injecting plasmid constructs into the tibialis anterior muscle of mice at five different sites in a final volume of 100 ␮l 0.25% bupivacaine or after pre-treatment with cardiotoxin at the indicated time points. For gene gun immunization, mice were anesthetized, abdominal hair was removed, and each mouse received two shots of gold particles coated with 0.5 ␮g of

172

C. Mauch et al. / Vaccine 20 (2002) 170–180

different delta antigen/cytokine plasmid DNA combinations using a gene gun (BioRad, Munich, Germany). Mice were sacrificed 10 days after the last immunization and sera and spleen cells were examined for HDV-specific antibody and T cell responses.

cell culture supernatant were measured using commercial ELISAs (Endogen, Cambridge, MA, USA). GM-CSF was detected in the cell culture supernatant by ELISA technique as previously described [20]. 2.6. Cytotoxicity assay

2.4. Generation of target cells expressing HDAgs The P815 mastocytoma cell line syngeneic to DBA-2 and congeneic to Balb/c mice was used to generate target cells to measure CTL activity in vitro. Two P815 transfectants were established after transfection of P815 cells with pSS32 or pSS33, one expressing HDAg (P815␦32), the other LHDAg (P815␦33). These cell lines were also used for the in vivo tumor model in DBA-2 mice. All cell lines were obtained from the ATCC (Rockville, MD, USA). 2.5. HDAg expression by DNA plasmids A murine myoblast (G8) cell line was transiently transfected with expression constructs to assess intracellular levels and secretion into the culture supernatant of HDV antigens and IL-18, IL-12, and GM-CSF cytokines. pSS32 and pSS33 were transfected into G8 cells, respectively. After 48 h of transfection, cells were analyzed for the presence of HDAg and LHDAg by Western blot analysis. In brief, cell lysates were prepared in RIPA buffer (0.15 M NaCl, 1% NP-40, 50 mM Tris, 0.5% DOC and 1% SDS) or a buffer containing 1% NP-40, 100 mM NaCl, 50 mM Tris, separated by 12% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and electrotransferred onto Immobilon-P membranes (Millipore, Bedford, MA, USA). After blocking with 3% non-fat dry milk and 1% BSA, membranes were incubated with a rabbit anti-HDAg antibody (kindly provided by D. Lazinski, Tufts University, Boston, MA, USA), followed by detection of a bound antibody with a horse radish-peroxidase-labeled anti-rabbit antibody (Amersham, Arlington Heights, IL, USA). Lysates prepared from cells transfected with an “empty” plasmid (mock) were used as negative controls. In addition, cell lysates and cell culture supernatant were examined for presence of HDAg and LHDAg using a commercial ELISA format (DiaSorin, Cologne, Germany). For immunofluorescence studies transfected cells were grown on sterile glass slides and fixed with 95% ethanol/5% acetic acid at −20◦ C. Slides were incubated with polyclonal rabbit anti-HDAg antibody at a dilution of 1:8000 or with human anti-HDV positive sera for 1 h at 37◦ C. After washing with PBS in 3% BSA, slides were incubated with a 1:1000 dilution of fluorescein-conjugated anti-rabbit or anti-human antibody (Cappel, Durham, NC, USA) for 1 h at 37◦ C, respectively. The pcD/3-IL12, pCI-1sIL-18, and pRJB-GM vectors were transfected into G8 cells. Cell culture medium was collected 2 days later. IL-12 and IL-18 secretion into the

Spleen cells from immunized mice were suspended in complete IMEM with 10% FCS and 2-mercaptoethanol (5× 10−5 M) and analyzed for cytotoxic activity following 6 days of in vitro stimulation. In vitro stimulation was performed by adding recombinant murine IL-2 once at a concentration of 2 U/ml and Balb/c or DBA-2 responder cells (4 × 107 ) were co-cultured with 2.5 × 106 irradiated (10,000 rad) P815␦32 or P815␦33 cells. As controls for specificity of CTL activity, we stimulated in vivo primed effector cells with parental P815 cells or with another P815 cell line [21] stably expressing the large HBsAg (P815-LS). Cytotoxic effector lymphocyte populations were harvested after incubation for 5 days. A 5 h 51 Cr-release assay was performed in 96-well round bottom plates using 51 Cr-labeled P815␦32, P815␦33, parenteral P815 or P815-LS target cells. In vitro stimulated effector cells were tested at E:T ratios of 100:1, 20:1, and 5:1. Results were expressed according to the formula: percent specific lysis = (experimental release − spontaneous release)/(maximum release − spontaneous release). Experimental release represents the mean counts/min released by target cells in presence of effector cells. Total release represents the radioactivity released after lysis of target cells with 5% Triton X-100. Spontaneous release represents the radioactivity present in medium derived from target cells alone. For T cell depletion experiments, we employed low toxicity rabbit complement (Cedarlane Laboratories, Ont., Canada), anti-Ly-2 and anti-L3T4 mAbs purchased from Boehringer (Indianapolis, IN, USA). 2.7. IFN-γ ELISPOT assay Multiscreen-HA 96-well filter plates were coated with 4 ␮g/ml rat anti-mouse IFN-␥ antibody (PharMingen, San Diego, CA, USA, clone R46A2) at 4◦ C overnight. CD8+ T cells (1 × 105 per well) derived from DNA-immunized mice were separated from spleen cells using magnetic beads according to the manufacture’s instructions (MACS, Myltenyi Biotech, Bergisch–Gladbach, Germany). CD8+ T cells (1× 105 per well) derived from immunized mice were cultured in triplicates for 24 h with 1 × 104 irradiated stimulator cells (P815␦33) per well in 200 ␮l medium. After culture, the cells were washed out and 2 ␮g/ml biotinylated rat anti-mouse IFN-␥ (PharMingen, San Diego, CA, USA, clone XMG1.2) was added, and the plates were incubated for 3 h at room temperature. The plates were again washed, incubated with a 1:100 dilution of streptavidin–alkaline phosphatase polymer (Mabtech, Köln, Germany) for 30 min at room temperature and then developed with ALP conjugate substrate solution (BCIP/NBT, BioRad, Richmond, USA). The spots in each

C. Mauch et al. / Vaccine 20 (2002) 170–180

173

well were counted under a microscope, and the values are expressed as numbers of spot-forming cells relative to the number spleen cells added to each well at the start of the culture. As a control for specificity, primed spleen cells and the different irradiated stimulator cells were cultured alone.

immunoblotted against P815␦33 lysates or lysates of G8 cells transiently transfected with pSS33 or pSS32 using the Western blot technique described above.

2.8. T cell proliferation assay

CD4 and CD8 T cell subpopulations were depleted by intraperitoneal injection of purified hybridoma supernatant. A total of 1 mg per mouse per injection of anti-CD8 (clone YTS 169) or anti-CD4 (clone YTS 191.1) (9, 24, 28) was injected on days 5, 3, and 1 before tumor challenge and every 5 days thereafter. FACS analysis of peripheral blood mononuclear cell populations at different time points demonstrated that more than 95% of the CD4 and CD8 T cells were deleted.

Mice were anesthetized with isoflurane (Aerrane, Anaquest, NJ, USA) and spleen cells were harvested. Red blood cells were removed by incubation in 8.3% NH4 Cl/0.17 M Tris pH 7.4 for 10 min at 37◦ C. Spleen cells were cultured in triplicate using 96-well flat bottom plates at 5 × 105 cells per well in 100 ␮l complete IMEM (Mediatech, Washington, DC, USA) containing 10% FCS. Spleen cells were stimulated with a recombinant bacterially derived HDAg (kindly provided by D. Lazinski, Tufts University, Boston, MA, USA) protein at different concentrations. The HDAg protein was estimated to be 99% pure by Coomassie staining and immunoblot analysis and was tested negative for LPS contamination (data not shown). As negative controls, effector cells were stimulated with 10 ␮g/ml recombinant HBsAg. Finally, 2-mercaptoethanol was added to a final concentration of 5 × 10−5 M. Spleen cells were stimulated for 3 days and BrdU-incorporation was measured using a commercial kit (BrdU Cell Proliferation ELISA, Boehringer Mannheim, Germany). Incorporation of BrdU was corrected for background activity, e.g. stimulation with medium only ( OD 450 nm). 2.9. Cytokine assays In vivo primed spleen cells were co-cultured for 48 h with LHDAg and HBsAg in parallel to the proliferation assays as described above. IL-2, IL-4, and interferon-␥ levels were measured in the culture supernatant derived from proliferating spleen cells in 96-well-plates by commercial kits according to manufacturer’s instructions (Endogen, Boston, MA, USA). The lower limit of sensitivity of these assays was 5 pg/ml. 2.10. Anti-HDAg antibodies

2.11. In vivo monoclonal antibody ablation

2.12. Tumor protection model It could be demonstrated that a reliable tumor growth of P815␦33 and P815␦32 cells in 100% of mice was achieved by injection of 1 × 106 cells in 10 ␮l serum-free DMEM medium into the right flank of mice. After 10 days, tumors were visible and reached a size of about 3000 mm3 by day 40. This tumor size was used as endpoint in our study and mice were subsequently sacrificed. As a HDAg negative control cell line, the P815-LS cell line was used. Tumor incidence and volume were assessed two times weekly using calipers. Data are presented as mean volume ± S.E. Strong MHC class I expression in P815␦33, P815␦32, and P815-LS cells was observed by FACS analysis using an anti-mouse H-2Kd specific antibody (# MCA1055, Biozol, Eching, Germany) and a subsequent PE-labeled anti-mouse antibody (data not shown). Two identical cohorts of different groups of DBA-2 mice were intramuscularly immunized two times at a 14 day interval with the following plasmid combinations: Groups

Therapy

1 2

No therapy 75 ␮g pcDNA3 + 25 ␮g + 25 ␮g pCI-1sIL-18 75 ␮g pSS␦32 + 25 ␮g + 25 ␮g pCI-1sIL-18 75 ␮g pSS␦33 + 25 ␮g + 25 ␮g pCI-1sIL-18 75 ␮g pSec␦32 + 25 ␮g + 25 ␮g pCI-1sIL-18 75 ␮g pSec␦33 + 25 ␮g + 25 ␮g pCI-1sIL-18 75 ␮g pSS␦32 + 25 ␮g + 25 ␮g pCI-1sIL-18 75 ␮g pSec␦32 + 25 ␮g + 25 ␮g pCI-1sIL-18

3 4

The following assays were developed and employed to measure anti-HDAg antibodies in sera of immunized mice. First, a qualitative commercial ELISA for detection of both HDAg and LHDAg was used (Anti-Delta-EIA, ABBOTT Laboratories, Chicago, IL, USA). Second, P815␦33 or P815␦32 were starved in methionine-free medium, and then labeled with 250 uCi of 35 S-methionine/cysteine. Cells were lysed, pre-cleared, and immunoprecipitated using 1:20 dilutions of sera in PBS derived from immunized or untreated mice. Immune complexes were bound to staphylococcal protein A cells and subsequently separated by a 12% SDS-PAGE. Third, sera of immunized mice were

5 6 7 8

pRJB-GM pRJB-GM pRJB-GM pRJB-GM pRJB-GM pRJB-GM pRJB-GM

The IL-12 expression plasmid was not included in this protocol, because studies using HCV antigens demonstrated

174

C. Mauch et al. / Vaccine 20 (2002) 170–180

that co-immunizations using IL-18 and IL-12 DNA plasmids were not superior to immunizations with IL-12 or IL-18 expressing plasmids alone (data not shown). Each experimental group contained 10 animals. At day 10 after the last immunization, mice were inoculated with 1 × 106 P815␦33 (groups 1–6 of cohort 1) or P815␦32 (groups 1–6 of cohort 2) or P815-LS (groups 7 and 8) cells into the right flank. Each experimental group contained 10 animals. 2.13. Statistical analysis For comparison of results between the different groups, we used a non-parametric Mann–Whitney U-test. P-values <0.05 were considered statistically significant. Each group in the tumor therapy experiments included five mice. Tumor appearance and growth to 3000 mm3 was calculated by the Kaplan–Meier method, presented as standard deviation of the mean for each group, and differences between immunized and control mice were calculated by the Mantel–Haenszel test.

3. Results 3.1. HDAg and cytokine expression by DNA expression plasmids Western blot analysis demonstrated delta antigen expression after transfection of G8 cells with pSS32, pSec32, pSS33, pSec33, and in P815␦33 or P815␦32 cells (Fig. 1A). Immunofluorescence studies clearly demonstrated substantial cytoplasmic accumulation of HDAgs in transfected cells (data not shown). No staining pattern was observed in Mock transfected cells. In order to determine whether delta antigens encoded by pSec␦32 or pSec␦33 were secreted into the culture medium by G8 cells, an HDAg-specific ELISA was used. HDAg and LHDAg were detectable in both cell lysate and supernatant (data not shown). G8 cells transfected with pApIL-12p70, pRJB-GM, or pCI-1sIL-18 secreted high amounts of p70-IL-12 dimer, GM-CSF, and mature IL-18 into cell culture medium, respectively, as determined by ELISAs (Fig. 1B). 3.2. Induction of cellular immune responses against HDAgs Three times DNA-based immunizations (gene gun + i.m.) using Balb/c and C57BL/6 mice at a 2 week interval induced significant CD4+ helper T cell proliferative responses against HDAg. No differences in strength of proliferative T cell activity against HDAg were observed between mice immunized with constructs expressing LHDAg and HDAg (Fig. 2). Due to the lack of recombinant LHDAg we were not able, however, to determine proliferative re-

Fig. 1. DNA expression constructs. (A) G8 myoblast cells were transfected with different DNA expression plasmids. Subsequently, cell lysates were analyzed for protein expression by Western blot analysis. Lanes 1: pcDNA3, 2: pSS33, 3: pSS32, 4: pSec32, 5: P815␦32, 6: pSecTagB, 7: recombinant HDAg, 8: pcDNA3, 9: pSecTagB, 10: pSS33, 11: pSec33, 12: P815␦33, 13 and 14: positive control plasmids expressing LHDAg. (B) Expression of cytokines by DNA plasmids. G8 cells were transfected with pApIL-12p70, pRJB-GM, or pCI-1sIL-18. Secretion of p70-IL-12 dimer, GM-CSF, and mature IL-18 into the culture medium was determined by ELISAs. No immunoreactivity was observed in media from cells transfected with a control plasmid expressing enhanced green fluorescent protein (pCMV-EGFP).

sponses against LHDAg of mice immunized with HDAg or LHDAg DNA. Combined gene gun/i.m. immunization was not superior to i.m. immunization alone (data not shown). T cell proliferative activity was similar between Balb/c and C57BL/6 mice (Fig. 2). Comparable CD8+ CTL responses against delta antigens were primed in all immunized Balb/c mice independent of the immunization mode. CTL activity in C57BL/6 mice could not be evaluated due to the lack of syngeneic target cell transfectants. Priming with HDAg, e.g. by immunization with pSS32, induced CTL responses which were comparable in strength against both LHDAg and HDAg. Similarly, no differences between CTL activity against HDAg and LHDAg was observed in mice immunized with pSS33. CTL activity was comparable between mice immunized with pSS32 and pSS33 (Fig. 3A). Cytokine co-immunizations using GM-CSF, IL-18, and IL-12 encoding plasmids together with pSS33 or pSec33 (Fig. 3A) induced significantly stronger CTL activity in immunized mice in comparison to mice immunized with pSS33 alone (P = 0.005).

C. Mauch et al. / Vaccine 20 (2002) 170–180

175

Fig. 2. Delta antigen-specific T cell proliferative responses. T cell proliferative response. Balb/c and C57BL/6 mice (n = 5 in each group) were immunized three times at 2 weekly intervals (50 ␮g of HDAg plasmids + 50 ␮g pApIL-12 + 50 ␮g pRJB-GM/intramuscular injection and 0.5 ␮g of each plasmid/gene gun shot) using a combined gene gun and i.m. approach. Spleen cells were harvested 8 days after the last immunization and stimulated for 3 days with recombinant HDAg at the concentrations indicated and BrdU-incorporation was measured. Incorporation of BrdU was corrected for background activity, e.g. stimulation with medium only ( OD 450 nm).

Using the ELISPOT technique, we determined the HDAg-specific in vivo CTLp frequency and the kinetics of the CD8+ T cell response (Fig. 3B). Following a single i.m. injection of Balb/c mice with 100 ␮g pSS33 or pSS32, only few IFN-␥ producing cells were detectable at day 5 but were easily identified by day 10 and peaked at 15 days post-immunization. From day 15 onward, the frequency of HDAg-specific CD8+ T cells decreased only slightly suggesting that memory CD8+ T cells remain detectable directly ex vivo for at least 30 days post-injection (0.5–0.6%

of all splenic CD8+ T cells). Interestingly, performing a booster immunization at day 14 did not result in increased CTLp frequencies suggesting that one immunization is sufficient for maximal expansion of HDAg-specific CD8+ T cell precursors (data not shown). 3.3. Failure to induce anti-HDAg antibody responses Initial experiments in Balb/c (n = 25) and C57BL/6 mice (n = 25) demonstrated that no anti-HDAg antibody

Fig. 3. CD8+ T cell responses. (A) CTL activity. Balb/c mice (n = 10 in each group) were immunized three times with the indicated plasmids as described above. Single spleen cell suspensions were assayed after in vitro stimulation with syngeneic P815␦33 cells for 5 days. The effector cells were then tested against P815␦33 or P815-LS target cells in a 51 Cr-release assay at the E:T ratios indicated. Values represent means of triplicate determinations. (B) Kinetics of ex vivo CD8+ T cell frequency following DNA immunization. Balb/c mice were immunized once with 100 ␮g plasmid DNA. CD8+ splenocytes were harvested at the indicated time points post-immunization and stimulated for 24 h with irradiated P815␦33 cells. Subsequently, IFN-␥ ELISPOT assays were performed. The spots in each well were counted under a microscope, and the values are expressed as percentage of spot-forming cells relative to the number of CD8+ spleen cells added to each well at the start of the culture.

176

C. Mauch et al. / Vaccine 20 (2002) 170–180

responses could be detected after two i.m. immunizations with pSS32 or pSS33. Independent experiments using pcD/3HDAg gene gun and cardiotoxin associated i.m. immunizations of Balb/c mice confirmed these negative results. To enhance immunogenicity of HDAg i.m., co-immunizatons with different plasmids encoding GM-CSF, IL-12, or IL-18 were, therefore, performed. In addition, gene gun immunizations and IL-4 co-immunizations [20] thought to polarize immune responses towards a TH2 subtype, were carried out to induce anti-HDAg antibody responses. Using this immunization approaches we were not able, however, to induce detectable anti-HDAg antibodies in a total of 122 Balb/c and C57BL/6 mice using a highly sensitive ELISA system. In addition, none of the animals immunized against HDAgs and rejecting HDAg expressing syngeneic tumors (see below) developed anti-HDAg antibodies. In order to exclude conformational changes of delta antigens encoded by pSS32 and pSS33, sera from immunized mice were immunoblotted against denatured HDAg and LHDAg, respectively, or immunoprecipitated with conformational intact HDAg or LHDAg, respectively. Again, no antibody responses could be detected. By contrast, a control rabbit anti-HDAg antibody recognized delta antigens in these assays (data not shown). Finally, immunizations with plasmids encoding secreted forms of HDAg also did not result in detectable anti-HDAg antibody responses in both Balb/c and C57BL/6 mice. Immunization of animals was successful because all mice developed helper T cell and CTL responses against HDAg. Anti-HDAg antibody responses with a titer of 1:2000 could be detected in Balb/c mice immunized once subcutaneously with 10 ␮g recombinant HDAg in complete Freund’s adjuvants (CFA)

demonstrating that anti-HDAg antibodies can be induced. Under the experimental conditions used, however, the DNA vaccines were unable to induce anti-HDAg antibody responses in mice. 3.4. Cellular immune responses against HDAgs protect against tumor challenge The in vivo function of CTLs and helper T cells induced by DNA-based immunization was assessed using a syngeneic tumor model. DBA-2 mice immunized with cDNAs encoding the different forms of delta antigen and challenged with syngeneic murine mastocytoma cell lines stably expressing either HDAg (P815␦32) or LHDAg (P815␦33) were protected in 80–100% (Fig. 4). Moreover, tumor size was significantly less (P < 0.0001) as determined by the measurement of tumor weight in comparison to mice immunized with empty vector DNA or mice challenged with the same syngeneic P815 cell line stably expressing large HBsAg (P815-LS, data not shown). Indeed, 100% of mice immunized with empty plasmid or challenged with P815-LS demonstrated tumor formation confirming the specificity of cellular immune responses against delta antigens. It is important to emphasize that immunization with recombinant HDAg in CFA did not protect animals against tumor challenge (Fig. 4) despite the presence of anti-HDAg antibodies (data not shown) and significant CD4+ T cell proliferative responses (Fig. 5) but in absence of HDAg-specific CTL activity (Fig. 6). These observations suggested that antitumoral immunity requires the participation of CD8+ CTL. In fact, in vivo depletion of either CD4+ or CD8+ T cell subsets in mice specifically immunized against HDAg or

Fig. 4. Survival of animals. Tumor appearance and growth to 3000 mm3 was calculated by the Kaplan–Meier method. Mice were immunized with the plasmids indicated. Note that all plasmids were co-delivered with cytokine expression plasmids as described in Section 2. Each DNA-immunization group (n = 10) and mice immunized subcutaneously with 10 ␮g of recombinant HDAg in CFA (n = 2) were challenged with P815␦32 or P815␦33 tumors. For control of specificity, additional groups of mice (n = 10) immunized with pSS32 or pSec32 were challenged with P815-LS cells stably expressing HBV large envelope protein.

C. Mauch et al. / Vaccine 20 (2002) 170–180

177

Fig. 5. CD4+ T cell proliferation. Mice were immunized with the plasmids indicated. Note that all plasmids were co-delivered with cytokine expression plasmids as described in Section 2. Each DNA-immunization group (n = 10) and mice immunized with recombinant HDAg (10 ␮g) in CFA (n = 2) were challenged with P815␦32 or P815␦33 tumors. Mice were sacrificed if tumor size had reached 3000 mm2 or at day 50 after the second immunization if no or smaller tumors were present. Spleen cells were stimulated for 3 days with recombinant HDAg at the concentrations indicated and BrdU-incorporation was measured. Incorporation of BrdU was corrected for background activity, e.g. stimulation with medium only ( OD 450 nm).

LHDAg resulted in tumor growth comparable to control animals (Fig. 7) suggesting that CD8+ T cell responses induced by DNA immunization are dependent on CD4+ T cell help with regard to both priming and maintaining the CD8+ T cell response [22]. On the other hand, priming only CD4+ T cells by rHDAg/CFA is not sufficient to induce protection against tumor challenge. CD8+ T cells, therefore, seem to be necessary to reject transplanted tumor cells. Priming immune responses with HDAg cDNAs (e.g. pSS32 and pSec32) protected against tumor challenge with P815 mastocytomas expressing either HDAg or LHDAg (Fig. 4). Similarly, HDAg or LHDAg expressing tumors

did not grow in mice immunized with LHDAg cDNAs (e.g. pSS33 and pSec33) suggesting comparable in vivo immunogenicity between HDAg and LHDAg. It was an important finding that in all tumor challenged DBA-2 mice specifically immunized against HDAg and LHDAg, significant CTL activity against the HDAg expressing P815␦32 but not against P815-LS could be detected (Fig. 6). In addition, no CTL activity against HDAg could be demonstrated in mice co-immunized with empty plasmid and cytokine expression plasmids suggesting that cytokines produced by DNA plasmids without a specific viral tumor antigen were not able to prime a protective immunity

Fig. 6. CTL activity in tumor challenged animals. Mice were immunized as described in the legend to Fig. 4. After 5 days of in vitro stimulation spleen cells were analyzed for cytotoxic activity against P815␦33 and P815-LS target cells at the E:T ratio indicated.

178

C. Mauch et al. / Vaccine 20 (2002) 170–180

Fig. 7. Identification of anti-tumoral immune reactivities in vivo. Anti-tumoral immunity required the participation of both CD4+ and CD8+ T cells. CD4 and CD8 T cell subpopulations were depleted by i.p. injection of purified hybridoma supernatant. Mice were immunized with plasmids indicated. Each indicated plasmid was co-injected with pRJB-GM and pCI-1sIL-18 as described in the legend to Fig. 4. At day 10, animals were challenged using 1 × 106 P815␦33 cells. If CD8+ (n = 2) or CD4+ T cells (n = 2) had been depleted in vivo prior and during tumor challenge tumor growth was comparable to non-immunized control animals (n = 2). By contrast, animals (n = 2) which had not been depleted of either CD4+ or CD8+ T cells were protected.

against HDAg expressing tumors. Significant CD4+ T cell proliferative responses were detectable in mice immunized with delta antigen cDNAs but not in mice immunized with empty plasmid (Fig. 5).

4. Discussion Nisini et al. demonstrated that the detection of a specific T cell response to HDAgs in the peripheral blood from HDV infected individuals was related to the resolution of HDV-induced disease activity [15]. Therefore, in individuals with chronic HDV infection the quality and strength of the anti-delta immune responses may not be sufficient to promote viral clearance and generate protective immunity. Patients infected with HDV have been shown to develop antibodies to different antigenic sites of HDAg. However, neutralizing anti-HDAg antibodies which are able to provide protection have not been found. It is not known whether the small or the large delta antigen is sufficiently immunogenic to generate broad-based and vigorous CTL responses in vivo. Furthermore, CTL responses to HDV have not yet been characterized in patients or mice. The DNA-based immunization approach, therefore, was employed to determine whether CTL responses against HDAgs might be induced. This technique has been previously shown to induce cellular immune responses of different levels against a variety of pathogens in animal model systems [17]. Recently, Huang et al. demonstrated the

induction of antibody and TH1 responses against LHDAg using this approach [23]. However, they did not study CTL responses and did not compare the immunogenicity of LHDAg and SHDAg. In this study, we present evidence that DNA-based vaccination with plasmids encoding secreted and intracellular forms of LHDAg and HDAg elicits a strong cellular immune response. Most importantly, a specific CD8+ CTL response was generated for both delta antigens. A single inoculation of plasmid DNA can induce CD8+ T cells which are detectable as early as days 5–10 post-inoculation and which peak at day 15. By day 35, approximately 0.6% of splenic CD8+ T cells are HDAg specific. Since no small animal model, except for woodchucks, is currently available for HDV replication and infection, we determined whether the CTL responses generated by DNA-based immunization would protect DBA-2 mice against tumor formation using syngeneic P815 tumor cells stably transfected with a cDNA encoding HDAg or LHDAg. Approximately 80–100% of immunized mice were protected against tumor formation, indicating that in vivo functional CTLs are induced by this immunization approach. We were not able to measure CTL activity in C57BL/6 mice due to the lack of delta antigen expressing target cells. However, Polo et al. [24] were not able to detect CTL responses in C57BL/6 mice using a vaccinia virus-based system. This does not preclude the possibility of generating CTL responses in this mouse haplotype using other systems. However, MHC class I restriction may have an impact on strength and presence of CTL responses against delta antigens.

C. Mauch et al. / Vaccine 20 (2002) 170–180

Despite the presence of HDAgs expressing tumors as an immunological danger signal CTL activity in tumor challenged DBA-2 mice was not stronger than in DNA-immunized naive Balb/c mice. This can be readily explained by the fact that P815 target cells are syngeneic to DBA-2 mice but H-2 congeneic to Balb/c mice. Therefore, minor MHC differences between Balb/c derived lymphocytes and P815 cells may increase killing activity in the in vitro CTL assay. It was noteworthy that none of more than 120 immunized mice of different haplotypes developed a detectable anti-HDAg antibody response. Failure to detect anti-delta antibodies was not due to the lack of T cell help, as a CD4+ helper T cell responses was demonstrated for all plasmids encoding secreted and intracellular forms of LHDAg and HDAg. Our data are also supported by the WHV/HDV infection model. No measurable humoral anti-HDAg immune response was observed in WHV carrier woodchucks following HDAg cDNA gene gun immunization. In contrast, all chronically WHV infected woodchucks, co-immunized with recombinant HDAg/CpG-oligonucleotides, elicited detectable anti-HDV antibodies. Following HDV superinfection of DNA vaccinated woodchucks, two remained anti-HDV negative up to week 52 after challenge (Fiedler et al., unpublished results). This lack of seroconversion is unusual since prior studies reported anti-HDAg seroconversion in 20 HDV superinfected woodchucks [25]. These data are in contrast to results obtained after i.m. immunization with a plasmid containing replication competent head-to-tail cDNA dimers of HDV that induced anti-HDAg antibody responses in C57BL/6 mice [24]. Although these differences may be partially due to different expression plasmids, protein detection systems, and unknown factors associated with intramuscular replication of HDV, our data strongly suggest that delta antigens are weak B cell immunogens independent of the mode and route of vaccine delivery. In addition, there is no evidence that anti-HDAg antibodies are protective. The development of therapeutic or prophylactic vaccines against HDV, therefore, should focus on T cell immunity. This study demonstrates the capability of assessing cellular immune responses against HDAgs in an animal model as measured by inhibition of tumor growth. Based on both previous clinical studies demonstrating the importance of cellular immune responses to delta antigens for disease activity, and the experimental results presented here, indicating that both delta antigens generate significant cellular immune responses in mice, we are led to believe that DNA-based immunization with genes encoding LHDAg and HDAg is an attractive approach for the development of a therapeutic vaccine against HDV. Since neutralizing antibodies seem to be absent after HDV infections in humans, the lack of detectable anti-HDAg antibody responses in our study does not preclude the possibility to employ DNA-based immunization for developing a prophylactic T cell-based vaccine against HDV. In addition, in the future, it will be important to develop anti-viral strategies aimed of the clearance of HBV

179

or suppression of HBsAg, thereby eliminating its helper function for HDV virion assembly, export, and infectivity.

Acknowledgements This work was supported by a grant from the Max-PlanckFoundation. M.G. is supported by BMBF-Grant 01GE9611 from the German Center for Aeronautics and Space Travel (DLR), by Grants Ge824/6-1 and SFB364 from the Deutsche Forschungsgemeinschaft, and by Grant 10-1662-Re from the Deutsche Krebshilfe. References [1] Sagnelli E, Stroffolini T, Ascione A, Chiaramonte M, Craxi A, Giusti G, et al. Decrease in HDV endemicity in Italy. J Hepatol 1997;26:20–4. [2] Huo TI, Wu JC, Lin RY, Sheng WY, Chang FY, Lee SD. Decreasing hepatitis D virus infection in Taiwan: an analysis of contributory factors. J Gastroenterol Hepatol 1997;12:747–51. [3] Gaeta GB, Stroffolini T, Chiaramonte M, Ascione T, Stornaiuolo G, Lobello S, et al. Chronic hepatitis D: a vanishing disease? An Italian multicenter study. Hepatology 2000;32:824–7. [4] Lau DT, Doo E, Park Y, Kleiner DE, Schmid P, Kuhns MC, et al. Lamivudine for chronic delta hepatitis. Hepatology 1999;30:546–9. [5] Malik AH, Lee WM. Hepatitis B therapy: the plot thickens. Hepatology 1999;30:579–81. [6] Chao M, Hsieh SY, Taylor J. Role of two forms of hepatitis delta virus antigen: evidence for a mechanism of self-limiting genome replication. J Virol 1990;64:5066–9. [7] Casey JL, Bergmann KF, Brown TL, Gerin JL. Structural requirements for RNA editing in hepatitis delta virus: evidence for a uridine-to-cytidine editing mechanism. Proc Natl Acad Sci USA 1992;89:7149–53. [8] Casey JL, Gerin JL. Hepatitis D virus RNA editing: specific modification of adenosine in the antigenomic RNA. J Virol 1995;69:7593–600. [9] Luo GX, Chao M, Hsieh SY, Sureau C, Nishikura K, Taylor J. A specific base transition occurs on replicating hepatitis delta virus RNA. J Virol 1990;64:1021–7. [10] Chang FL, Chen PJ, Tu SJ, Wang CJ, Chen DS. The large form of hepatitis delta antigen is crucial for assembly of hepatitis delta virus. Proc Natl Acad Sci USA 1991;88:8490–4. [11] Kuo MY, Chao M, Taylor J. Initiation of replication of the human hepatitis delta virus genome from cloned DNA: role of delta antigen. J Virol 1989;63:1945–50. [12] Guilhot S, Huang SN, Xia YP, La Monica N, Lai MM, Chisari FV. Expression of the hepatitis delta virus large and small antigens in transgenic mice. J Virol 1994;68:1052–8. [13] Negro F, Shapiro M, Satterfield WC, Gerin JL, Purcell RH. Reappearance of hepatitis D virus (HDV) replication in chronic hepatitis B virus carrier chimpanzees rechallenged with HDV. J Infect Dis 1989;160:567–71. [14] Karayiannis P, Saldanha J, Jackson AM, Luther S, Goldin R, Monjardino J, et al. Partial control of hepatitis delta virus superinfection by immunisation of woodchucks (Marmota monax) with hepatitis delta antigen expressed by a recombinant vaccinia or baculovirus. J Med Virol 1993;41:210–4. [15] Nisini R, Paroli M, Accapezzato D, Bonino F, Rosina F, Santantonio T, et al. Human CD4+ T cell response to hepatitis delta virus: identification of multiple epitopes and characterization of T-helper cytokine profiles. J Virol 1997;71:2241–51.

180

C. Mauch et al. / Vaccine 20 (2002) 170–180

[16] Accapezzato D, Nisini R, Paroli M, Bruno G, Bonino F, Houghton M, et al. Generation of an MHC class II-restricted T cell epitope by extracellular processing of hepatitis delta antigen. J Immunol 1998;160:5262–6. [17] Lewis PJ, Babiuk LA. DNA vaccines: a review. Adv Virus Res 1999;54:129–88. [18] Grimm C, Ortmann D, Mohr L, Michalak S, Mohr L, Krohne TU, Meckel S, et al. Mouse alpha-fetoprotein specific DNA-based immunotherapy of hepatocellular carcinoma leads to tumor regression in mice. Gastroenterology 2000;119:1104–12. [19] Xiang Z, Ertl HC. Manipulation of the immune response to a plasmid-encoded viral antigen by coinoculation with plasmids expressing cytokines. Immunity 1995;2:129–35. [20] Geissler M, Gesien A, Tokushige K, Wands JR. Enhancement of cellular and humoral immune responses to hepatitis C virus (HCV) core protein using DNA-based vaccines augmented with cytokine expressing plasmids. J Immunol 1997;158:1231–7.

[21] Geissler M, Bruss V, Michalak S, Ortmann D, Wands JR, Blum HE. Intracellular retention of hepatitis B virus surface proteins reduces the immune response augmenting effects of IL-2 after cytokine genetic coimmunizations. J Virol 1999;73:4284–92. [22] Wild J, Grusby MJ, Schirmbeck R, Reimann J. Priming MHC-I-restricted cytotoxic T lymphocyte responses to exogenous hepatitis B surface antigen is CD4+ T cell dependent. J Immunol 1999;163:1880–7. [23] Huang YH, Wu JC, Tao MH, Syu WJ, Hsu SC, Chi WK, et al. DNA-based immunization produces TH1 immune responses to hepatitis delta virus in a mouse model. Hepatology 2000;32:104–10. [24] Polo JM, Lim B, Govindarajan S, Lai MM. Replication of hepatitis delta virus RNA in mice after intramuscular injection of plasmid DNA. J Virol 1995;69:5203–7. [25] Schlipkoter U, Ponzetto A, Fuchs K, Rasshofer R, Choi SS, Roos S, et al. Different outcomes of chronic hepatitis delta virus infection in woodchucks. Liver 1990;10:291–301.