Evaluation of the Effect of CD70 Co-Expression on CD8 T Cell Response in Protein-Prime MVA-Boost Vaccination in Mice

Here, we investigate the potential of CD70 co-expression during viral vector boost vaccination to improve an antigen-specific T cell response. To determine the chance of activating antigen-specific T cells by CD70, we used the HBV core antigen as a model antigen in a heterologous protein-prime, Modified Vaccinia virus Ankara (MVA) boost vaccination scheme. Both the HBV core and a CD70 expression cassette were co-expressed upon delivery by an MVA vector under the same promoter linked by a P2A site. To compare immunogenicity with and without CD70 co-expression, HBV-naïve, C57BL/6 (wt) mice and HBV-transgenic mice were prime-vaccinated using recombinant HBV core antigen followed by the MVA vector boost. Co-expression of CD70 increased the number of vaccine-induced HBV core-specific CD8 T cells by >2-fold and improved their effector functions in HBV-naïve mice. In vaccinated HBV1.3tg mice, the number and functionality of HBV core-specific CD8 T cells was slightly increased upon CD70 co-expression in low-viremic, but not in high-viremic animals. CD70 co-expression did not impact liver damage as indicated by ALT levels in the serum, but increased the number of vaccine-induced, proliferative T cell clusters in the liver. Overall, this study indicates that orchestrated co-expression of CD70 and a vaccine antigen may be an interesting and safe means of enhancing antigen-specific CD8 T cell responses using vector-based vaccines, although in our study it was not sufficient to break immune tolerance.


Introduction
Persistent viral infections and malignant diseases threaten millions of people worldwide and impair quality of life, cause deadly complications and lead to high costs in health care. Therapeutic vaccination offers a promising immunotherapeutic treatment option for both viral infections and malignant diseases. It has the potential to elicit an effective immune response when a suitable antigen has been defined. Chronic hepatitis B is an example of such a disease caused by persistent Hepatitis B virus (HBV) infection. The number of chronically infected individuals is high, with almost 300 million HBV carriers worldwide at high risk to develop liver cirrhosis or liver cancer causing more than 880.0000 deaths per year [1,2]. The current therapeutic options comprise treatments with nucleos(t)ide analogues or interferon alpha. However, both are rarely curative and require long-term drug administration, which is cost-intensive and prone to cause side effects, respectively [3][4][5][6].
In patients with self-limiting HBV infection, the elimination or final control of the virus is associated with strong and polyclonal CD4 and CD8 T cell responses [7][8][9][10][11]. In persistent chronic HBV infection however, HBV-specific T cells are only detected at a very low frequency and-at least partially-in an immunotolerant state, described as T cell exhaustion, resulting in a failure to eradicate the virus [12][13][14][15][16][17][18].
Therapeutic vaccination aims at empowering the immune system and restoring the functions that control the virus. It has been applied as a potential approach for the treatment of chronic hepatitis B (reviewed in [19][20][21][22]). These studies, however, have shown that therapeutic vaccination is challenging and requires well-formulated antigens, delivery systems and most likely a combination with either antiviral treatment, immune stimulators or a simultaneous blocking of inhibitory molecules [21,23,24].
We have recently developed a therapeutic heterologous prime-boost vaccination regimen, termed TherVacB. Based on a prime vaccination with particulate HBV antigens followed by boost vaccination with a recombinant modified vaccinia virus Ankara (MVA) expressing HBV antigens, TherVacB stimulates a strong anti-HBV immunity [25]. The approach elicits both anti-HBs seroconversion and strong HBV surface-and HBV core-specific CD4 and CD8 T cell immunity [25]. We have demonstrated that TherVacB is capable of breaking HBV-specific immune tolerance in HBV transgenic mice with low to intermediate HBV antigen concentrations, but at high HBV antigen concentrations the therapeutic vaccine failed to induce HBV specific CD8 T cell immunity [25]. Thus, the TherVacB strategy was successfully combined with additional treatments before, during and after vaccination in preclinical studies: namely siRNA mediated knockdown of virus antigen expression prior to vaccination [26], knockdown of PD-L1 expression during vaccination [27] and post-treatment with CpG [28]. The therapeutic effect could be significantly increased in preclinical mouse models with these treatments-even in animals with a high antigen load-up to a permanent elimination of the virus in the AAV-HBV mouse model [26].
The CD70 molecule is a surface antigen that is physiologically expressed mainly on activated mature dendritic cells (antigen presenting cells). Its ligand, the CD27 molecule, is constitutively expressed on the surfaces of resting and naive T cells. Ligation of CD70 on antigen presenting cells with CD27 on T cells promotes T cell activation [29] and survival of antigen-specific activated T cells [30][31][32][33], and is tightly regulated to prevent overexpression and excessive lymphocyte activation. CD70 co-stimulation induces proliferation of T cells and contributes to the activation of their effector function such as the increase in cytotoxic CD8+ T cell activity [32,[34][35][36][37][38][39]. Further, it influences CD4+ T cell subset differentiation [36,40] and contributes to the formation of the effector and memory T cell pool [39,41]. Recent data have even shown that CD70 expressed on regulatory T cells inhibits their suppressor function and provides co-stimulation to conventional T cells [42]. All in all, CD70 is an interesting candidate to flip the switch from T cell tolerance to immunity and is therefore often the subject of investigation in immunotherapeutic approaches against cancer and chronic viral infections. However, CD70 co-stimulation in immunotherapy bears the potential risk of the over-activation or unspecific stimulation of immune cells and, therefore, promoting autoimmunity [43,44].
In order to investigate the role of CD70 co-expression in vaccine-induced CD8 T cell immunity we chose the HBV core protein as a model antigen, a strong inducer of HBVspecific CD8 T cell response. In the current study we omitted the simultaneous expression of the HBV surface protein-the most effective protective HBV vaccine candidate-in order to firstly exclude interference across specific responses and secondly avoid a robust antibody response with significant effects on HBV parameters (which would be the case by anti-HBs [26,45], but not by anti-HBc antibodies [28]). The use of the HBV core as a model antigen therefore allowed the scientific question of whether CD70 would increase T cell immunity to be addressed.
In this study, we investigated whether and how CD70 co-expression may improve the CD8 T cell response boosted by a viral vector vaccine. We generated an MVA-vector expressing the HBV core protein and integrated the co-stimulatory molecule CD70 under control of the same promotor. This system offers the opportunity of delivering the target antigen in direct spatial and temporal conjunction with the co-stimulatory CD70 protein, further reducing the risk of misdirected co-stimulation, which was already minimized by the upstream protein prime in our vaccination scheme. In this model, we showed that expression of and co-stimulation by CD70 can selectively increase and improve the CD8 T cell response against the viral model antigen in HBV-naive but not in immune-tolerant, HBV-transgenic mice.

Generation and Production of MVA Vectors
The gene expression cassettes for recombinant MVAs were cloned utilizing the MVA shuttle plasmid pLW-73 [46]. The coding sequence of human CD70, which was demonstrated to bind to murine CD27 [47], was derived from cDNA from an Epstein-Barr Virus (EBV)-immortalized human B-cell line. For generation of the respective recombinant MVA (rMVA), DF-1 cells were transfected with the MVA shuttle plasmid and infected with wildtype MVA (MVAwt). After recombination, amplification of recombinant MVAs (rMVAs) was performed in DF-1 cells (Institute of Virology, Helmholtz Center Munich). Amplified rMVAs were purified via 36% sucrose cushion ultracentrifugation (13,500 rpm at 4 • for 90 min, a second round was performed after resuspension). rMVAs were titrated in serial dilutions on DF-1 cells. The titers were determined by half-maximal median tissue culture infection dose (TCID 50 ) of two repetitions. The rMVAs were stored at −80 • . Three MVA vectors were used in this study: MVAwt, rMVA-HBc and rMVA-HBc-CD70.

Evaluation of Gene Expression in Cell Culture
HEK293 cells were transfected with the plasmids containing the gene expression cassette (CMV-Pr.-HBc[-P2A-CD70]-IRES-GFP) by using Lipfectamin2000 transfection reagent (Thermo Fisher Scientific, Waltham, MA USA 02451) according to the manufacturer's instructions. After two days, cells were harvested and analyzed. The expression of CD70 and GFP was determined by FACS analysis of the transfected cells after LIVE/DEAD staining with the Near-IR Dead Cell Stain Kit (Life technologies GmbH, Invitrogen, Darmstadt, Germany) utilizing anti-CD70-PE antibody (BioLegend, San Diego, CA, USA). Expression of the HBV core protein was detected by Western Blot/SDS-Page of a lysate of transfected HEK293 cells. After transfer to nitrocellulose, the blotted bands were immunodetected with rabbit anti HBV-core (DAKO Carpinteria, CA, USA) as the primary antibody and subsequently visualized with peroxidase-labeled mouse anti-rabbit IgG antibodies.

Ethical Statement
Animal experiments were conducted in strict accordance with the regulations of the German Society for Laboratory Animal Science (GV-SOLAS) and the health laws of the Federation of European Laboratory Animal Science Associations (FELASA). Experiments were approved by the local Animal Care and Use Committee of Upper Bavaria (Permit No. 55.2-1-54-2532-103-12). For animal welfare reasons, clear trends in individual transgenic animals were described by observation without raising the groups to a statistically relevant size. Mice were kept in a specific pathogen-free facility under appropriate biosafety level following institutional guidelines.

Experimental Animals
Wild-type C57BL/6 mice (haplotype H-2 b/b ) were purchased from Charles River Laboratories (Schulzfeld, Germany). Fourteen to sixteen weeks old female and male HBV transgenic mice (StrainHBV1.3.32, HBV genotype D, subtype ayw, kindly provided by F. Chisari, The Scripps Institute, La Jolla, CA, USA [7]) were bred at the AVM Animal Facility, Helmholtz Center Munich under specific pathogen-free conditions following institutional guidelines.

Heterologous Protein-Prime MVA-Boost Vaccination
Mice were immunized with a particulate protein prime and subsequently boosted with a recombinant Modified Vaccinia Ankara virus (MVA) expression HBV core protein. As protein prime, we used recombinant HBV core antigen (HBcAg) expressed in Escherichia coli that consist of 180 or 240 subunit of the HBV core protein and spontaneously forms particles (APP Latvijas Biomedicinas, Riga, Latvia). Briefly, for the protein prime at day 0, mice were immunized subcutaneously with 30 µg of recombinant HBcAg adjuvanted with 32 µg of synthetic phosphorothioated CpG ODN 1668 and 25 µg poly[di(sodiumcarboxylatoethylphenoxy)phosphazene] (PCEP) dissolved in 50 µL PBS (group wt received adjuvants only). For the MVA-boost at day 21, mice received 10 8 infectious units (IU) of MVAwt (group wt), rMVA-HBc (group HBc) or rMVA-HBc-CD70 (group HBc-CD70) by intraperitoneal injection in 250 µL PBS [25]. Mice were sacrificed at day 35 of the experiment for analysis of splenocytes +/− liver associated lymphocytes +/− liver sections + serum.

Isolation of Lymphocytes from the Spleen and Liver
Splenocytes were isolated as previously described [48]. Liver-associated lymphocytes (LAL) were isolated and purified by density gradient centrifugation as described in [12]. Briefly, mouse livers were perfused with pre-warmed PBS and pressed through a 100 µm nylon cell strainer (BD Falcon, Franklin Lakes, NJ, USA). After washing, cell pellets were suspended in 10 mL of a pre-warmed enzyme solution containing 1 mg/mL collagenase type IV (Worthington, Lakewood, NJ, USA) in RPMI 1640 medium containing 10% fetal bovine serum (Gibco, Thermo Fischer Scientific, Darmstadt, Germany) and digested at 37 • C for 30 min. The cell pellets were then resuspended in 40% Percoll solution (GE Healthcare, Munich, Germany), layered on 80% Percoll solution and centrifuged at 1600× g for 20 min without brakes for density separation, resulting in the LAL fraction concentrating between the two Percoll layers.

Detection of HBV Core-Specific CD8 T Cells by Multimers and Intracellular Cytokine Staining
HBV core-specific CD8 T cells were detected by staining with MHC class I multimers conjugated with HBV core-derived H-2K b -restricted peptide C 93-100 (C 93 , MGLKFRQL), as described previously [28].

Heterologous Protein-Prime MVA-Boost Vaccination
Mice were immunized with a particulate protein prime and subsequently boosted with a recombinant Modified Vaccinia Ankara virus (MVA) expression HBV core protein.

Isolation of Lymphocytes from the Spleen and Liver
Splenocytes were isolated as previously described [48]. Liver-associated lymphocytes (LAL) were isolated and purified by density gradient centrifugation as described in [12]. Briefly, mouse livers were perfused with pre-warmed PBS and pressed through a 100 µm nylon cell strainer (BD Falcon, Franklin Lakes, NJ, USA). After washing, cell pellets were suspended in 10 mL of a pre-warmed enzyme solution containing 1 mg/mL collagenase type IV (Worthington, Lakewood, NJ, USA) in RPMI 1640 medium containing 10% fetal bovine serum (Gibco, Thermo Fischer Scientific, Darmstadt, Germany) and digested at 37 °C for 30 min. The cell pellets were then resuspended in 40% Percoll solution (GE Healthcare, Munich, Germany), layered on 80% Percoll solution and centrifuged at 1600× g for 20 min without brakes for density separation, resulting in the LAL fraction concentrating between the two Percoll layers.

Detection of HBV Core-Specific CD8 T Cells by Multimers and Intracellular Cytokine Staining
HBV core-specific CD8 T cells were detected by staining with MHC class I multimers conjugated with HBV core-derived H-2K b -restricted peptide C93-100 (C93, MGLKFRQL), as described previously [28].

Statistical Analysis
Statistical analyses were performed using GraphPad Prism version 5 (GraphPad Software Inc., San Diego, CA, USA). Results are depicted as means ± standard deviations. Differences between groups were analyzed for statistical significance using a one-way ANOVA to compare all groups. If significant, unpaired t-tests were performed as post hoc tests (95% confidence level). p-values < 0.05 were considered significant.
For the calculation of the gain of functionality data, results of the ICS and C 93 multimer stainings were used: the proportion of IFNγ+ CD8 T cells was divided by the proportion of C 93 multimer-positive CD8 T cells for each mouse separately.

Generation of Recombinant MVA Encoding CD70
A gene expression cassette was designed encoding the HBV core antigen and the human B-cell-derived protein CD70. Simultaneous and equimolar expression of HBV core and CD70 was accomplished utilizing a P2A site [49]. The corresponding control construct was generated in analogy but without the P2A-CD70 cassette. During the production process, co-expression of GFP coupled to the HBV core gene with or without CD70 via an internal ribosomal entry site served for both recombinant vectors as reporter ( Figure 1A). For both MVA vectors, transgene expression was under control of the early promoter mH5 [50]. The correct expression of the transgenes HBV core and CD70 was confirmed in HEK293 cells by Western blot and flow cytometry analyses ( Figure 1B,C) resulting in the generation of MVAs rMVA-HBc-CD70 and rMVA-HBc ( Figure 1D). Both vectors were amplified in DF-1 cell culture, purified and titrated using serial dilution ( Figure 1E).

CD70 Co-Expression during Boost Vaccination Exclusively Increases the Number of Transgene-Specific CD8 T Cells in wt Mice
Employing the well-established protein-prime MVA-boost vaccination scheme of Ther-VacB [25,26], we aimed at investigating whether CD70 co-expression during boost impacts the HBV core-specific CD8 T cell response. C57BL/6 mice were vaccinated subcutaneously with recombinant particulate HBcAg adjuvanted with a mixture of CpG and PCEP, followed by an intraperitoneal rMVA boost injection using rMVA-HBc-CD70 or control rMVA-HBc at day 21 ( Figure 2A). T cell responses were analyzed 14 days after the rMVA-boost. A group of mock vaccinated C57BL/6 mice (group wt) were primed with adjuvants only and boosted with wild-type MVA (MVAwt). Splenocytes were isolated from vaccinated animals 14 days after the MVA-boost and stained with major histocompatibility complex class I multimers loaded with the C 93 peptide MGLKFRQL.

CD70 Co-Expression during Boost Vaccination Exclusively Increases the Number of Transgene-Specific CD8 T Cells in wt Mice
Employing the well-established protein-prime MVA-boost vaccination scheme of TherVacB [25,26], we aimed at investigating whether CD70 co-expression during boost impacts the HBV core-specific CD8 T cell response. C57BL/6 mice were vaccinated subcutaneously with recombinant particulate HBcAg adjuvanted with a mixture of CpG and PCEP, followed by an intraperitoneal rMVA boost injection using rMVA-HBc-CD70 or control rMVA-HBc at day 21 ( Figure 2A). T cell responses were analyzed 14 days after the rMVA-boost. A group of mock vaccinated C57BL/6 mice (group wt) were primed with adjuvants only and boosted with wild-type MVA (MVAwt). Splenocytes were isolated from vaccinated animals 14 days after the MVA-boost and stained with major histocompatibility complex class I multimers loaded with the C93 peptide MGLKFRQL. The staining revealed that CD70 co-expression significantly increased the proportion of HBV core-specific CD8 T cells upon heterologous vaccination (from 6.6% of all splenic CD8 T cells in vaccination group rMVA-HBc to 16.8% in vaccination group rMVA-HBc-CD70; p = 0.0077; Figure 2B,C). In a second step we re-stimulated the T cells by loading splenocytes ex vivo with the immune-dominant C 93 core-peptide (MGLKFRQL) recognized by the murine CD8 T cells and analyzed T cell activation by intracellular cytokine staining (ICS).
In the rMVA-HBc vaccination group we detected 5.9% core-specific CD8 T cells producing IFNγ, compared to 12.6% in the rMVA-HBc-CD70 vaccination group (p = 0.0086; Figure 2D,E). This confirmed that CD70 co-expression during boost vaccination fosters a functional CD8 T cell response. Of note, the results were confirmed by re-stimulation with an HBV core peptide pool at the individual mouse level (data not shown). A total of 2.6% (vaccination group rMVA-HBc) and 3.6% (vaccination group rMVA-HBc-CD70) of splenic CD8 T cells responded with IFNγ production to ex vivo stimulation with the dominant CD8 T cell epitope of the MVA-vector backbone B8R ( Figure 2E). In this experiment, CD70 co-expression during boost vaccination resulted in a 2.1-fold increase in mean HBV core-specific CD8 T cell responses (corresponding to a difference of 1.81 standard deviations between both vaccination groups) and no significant difference in B8R-specific CD8 T cell responses (difference 1.01-fold standard deviations). Thus, the co-stimulatory effect of CD70 during boost vaccination seemed to mainly support HBV core-specific CD8 T cell responses and only had a minor effect on the un-primed MVA-specific CD8 T cell response. Of note also, no significant difference in terms of IFNγ production of CD4 T cells after re-stimulation with an HBV core peptide pool was observed between the vaccination groups (data not shown).

Transgene-Specific CD8 T Cells Gain Functionality by Co-Expression with CD70
Next, we used ICS to examine the ability of CD8 T cells to express multiple cytokines simultaneously as a measure of the quality of the HBV core-specific CD8 T cell response and found that the proportion of multifunctional CD8 T cells with respect to the expression of IFNγ, IL2 and TNF was increased in the HBc-CD70 group compared to the HBc-group 14 days after boost vaccination ( Figure 2F). In order to investigate whether the proportion of C 93 -specific CD8 T cells producing IFNγ increased after CD70 co-expression, the percentage of IFNγ-producing CD8 T cells detected by ICS in relation to C 93 multimer-positive CD8 T cells was determined for all mice individually. We detected a significant difference between the vaccination groups rMVA-HBc and rMVA-HBc-CD70 (p = 0.0129, Figure 2G) indicating the increased functionality of C 93 -specific CD8 T cells. CD70 co-expression activated about 20% of previously non-reactive C 93 -specific CD8 T cells in our setting, indicating that CD70 efficiently antagonizes tolerogenic signaling during CD8 T cell boosting. Taken together, CD70 co-expression improved both the quality and quantity of vaccine-induced, HBc-specific CD8 T cells in C57Bl/6 mice.

CD70 Co-Expression during Boost Vaccination Does Not Affect Humoral Immunity in a Therapeutic Vaccination Setting
Next, we evaluated the effect of CD70 co-expression during boost vaccination in a therapeutic vaccination approach using the TherVacB regime in HBV1.3.32 transgenic (HBV1.3tg) mice, which is a well-established animal model for the simulation of a chronic HBV infection in mice [51].
Previous studies have demonstrated that therapeutic efficacy of TherVacB vaccination depends on the antigen load in the HBV1.3tg mouse model [25]. Therefore, we used both low-intermediate and high viremic HBV1.3tg mice to evaluate the co-stimulatory effect of CD70 depending on the viral load. Hereby, hepatitis B e antigen (HBeAg) levels prior to treatment were determined utilizing the Abbott Architect HBeAg 2.0 assay to classify the treated mice discriminating between mice with low/intermediate (HBeAg 1-8 S/CO) and high viremia (HBeAg levels > 8 S/CO). First, we evaluated the effects of CD70 co-stimulation on the induction of humoral immunity. As expected, we observed that antibodies against HBc evolved in both vaccination groups (HBc and HBc-CD70) already after the protein-prime vaccination, and the MVA boost vaccination had no significant additional effect ( Figure 3A). Further, no significant difference in the induction of HBc-specific antibodies was detectable between the two vaccination groups, indicating that there was no beneficial effect of CD70 in the production of anti-HBc. In addition, none of the vaccination regimes elicited a significant anti-HBs response, as expected ( Figure 3B). Our results therefore do not indicate that CD70 co-stimulation affected humoral immunity in our therapeutic vaccination setting.

CD70 Co-Expression during Boost Vaccination Has No Influence on HBeAg and HBsAg Levels
Next, we examined the effect of CD70 co-expression during boost vaccination on serum HBeAg and HBsAg levels, which serve as markers for HBV replication. In most vaccinated mice (from both groups, HBc and HBc-CD70), HBeAg levels decreased to low or undetectable levels after boost vaccinations in mice with low viremia ( Figure 3C). In high viremic mice the levels of HBeAg significantly decreased after immunization. Assessment of the vaccination effects on the serum HBsAg was only possible in the high-viremic mice, since most of the mice with low viremia demonstrated undetectable HBsAg at the beginning of the experiments (Figure 3D), confirming our previous reports [25]. In contrast to controls (wt), vaccination of high-viremic mice with HBc and HBc-CD70 regimens resulted in an approximately 1-log10 drop in HBsAg levels at day 35 as compared to the baseline values. However, no additional effect of CD70 co-expression on lowering neither HBeAg, nor HBsAg levels was observed.

CD70 Co-Expression during Boost Vaccination Has No Influence on HBeAg and HBsAg Levels
Next, we examined the effect of CD70 co-expression during boost vaccination on serum HBeAg and HBsAg levels, which serve as markers for HBV replication. In most vaccinated mice (from both groups, HBc and HBc-CD70), HBeAg levels decreased to low or undetectable levels after boost vaccinations in mice with low viremia ( Figure 3C). In high viremic mice the levels of HBeAg significantly decreased after immunization. Assessment of the vaccination effects on the serum HBsAg was only possible in the high-viremic mice, since most of the mice with low viremia demonstrated undetectable HBsAg at the beginning of the experiments ( Figure 3D), confirming our previous reports [25]. In contrast to controls (wt), vaccination of high-viremic mice with HBc and HBc-CD70 regimens resulted in an approximately 1-log 10 drop in HBsAg levels at day 35 as compared to the baseline values. However, no additional effect of CD70 co-expression on lowering neither HBeAg, nor HBsAg levels was observed.
3.6. CD70 Co-Expression during Boost Vaccination Enhances the Number of Transgene-Specific CD8 T Cells in a Therapeutic Vaccination Setting Using rMVA-HBc-CD70 We next analyzed the number of HBV core-specific CD8 T cells derived from the spleens and livers of vaccinated HBV1.3tg mice by C 93 multimer staining ( Figure 4). As expected, an inverse correlation of viremia and HBV core-specific CD8 T cell responses was observed in our vaccination experiments ( Figure 4A,C) corresponding to earlier results [25]. expected, an inverse correlation of viremia and HBV core-specific CD8 T cell responses was observed in our vaccination experiments ( Figure 4A,C) corresponding to earlier results [25]. The functionality of C93-specific CD8 T cells was assessed by ICS ( Figure 4B,D). In low viremic mice, CD70 co-expression during boost vaccination resulted in a slightly higher proportion of IFNγ-producing CD8 T cells in the liver and spleen (not significant), comparable to our results in antigen-naïve C75Bl/6 mice. In high viremic mice, however, no increase in HBV core-specific T cell immunity was detected and CD70 co-expression may even have had an unwanted effect such as a reduced IFNγ production in the spleen (see Figure 4D, splenocytes).
Examination of the vector-specific immune response by intracellular cytokine staining after stimulation with the MVA-derived peptide B8R ( Figure 4E) showed that CD70 co-expression also had no noticeable effect on the strength of the vector-specific CD8 T cell response.
In summary, CD70-costimulation increased the number of HBV core-specific CD8 T cells as well as the number of IFNγ-producing core-specific T cells upon vaccination in low-medium viremic mice. However, HBV core-specific immune tolerance could not be overcome in high viremic, HBV-transgenic mice by CD70 co-expression. The functionality of C 93 -specific CD8 T cells was assessed by ICS ( Figure 4B,D). In low viremic mice, CD70 co-expression during boost vaccination resulted in a slightly higher proportion of IFNγ-producing CD8 T cells in the liver and spleen (not significant), comparable to our results in antigen-naïve C75Bl/6 mice. In high viremic mice, however, no increase in HBV core-specific T cell immunity was detected and CD70 co-expression may even have had an unwanted effect such as a reduced IFNγ production in the spleen (see Figure 4D, splenocytes).
Examination of the vector-specific immune response by intracellular cytokine staining after stimulation with the MVA-derived peptide B8R ( Figure 4E) showed that CD70 co-expression also had no noticeable effect on the strength of the vector-specific CD8 T cell response.
In summary, CD70-costimulation increased the number of HBV core-specific CD8 T cells as well as the number of IFNγ-producing core-specific T cells upon vaccination in low-medium viremic mice. However, HBV core-specific immune tolerance could not be overcome in high viremic, HBV-transgenic mice by CD70 co-expression.

Co-Expression by CD70 during Boost Vaccination Increases the Number of T Cell Clusters in Livers of Vaccinated HBV1.3tg Mice without Promoting Immunopathology
To investigate whether CD70 co-expression may promote liver damage, we measured serum alanine transaminase (ALT) activity in serum. In some of the vaccinated HBV1.3tg mice we found a minor ALT increase independent of the vaccination group at day 35 ( Figure 5A). This was most probably caused by the MVA boost vaccination per se. In the HBc-CD70 group, no elevated ALT-levels in general or increased numbers of mice with high ALT-levels (>100 U/L) were detected.
Finally, we investigated the infiltration of T cells into the liver in the therapeutic vaccination setting. By immunohistochemistry staining for CD3 in liver biopsies we observed similar absolute numbers of T cells in HBV1.3tg mice in all vaccination groups (wt, HBc and HBc-CD70). The number of T cells showed no correlation with the viremia of mice ( Figure 5B). The number of T cell clusters (≥3 or ≥6 cells in direct contact) appeared to be increased in vaccination groups HBc and HBc-CD70 but did also not correlate with the viremia of mice ( Figure 5D). In a double staining of liver sections (CD3 for T cells and Ki67 for proliferating cells) we noticed that T cells in cell clusters frequently expressed Ki67 ( Figure 5E). As both the number of single T cells and T cells in clusters did not correlate with changes of viremia in the HBV1.3tg mice, the statistical analysis of T cell clusters in liver sections was performed on all mice together ( Figure 5F); compared to the control group (wt), the number of T cell clusters was increased in both vaccination groups HBc and HBc-CD70 to 0.33 and 0.45 clusters/mm 2 , respectively, compared to 0.2 clusters/mm 2 in the control group (wt) ( Figure 5E, clusters of ≥ 6 cells). The improvement was highest in the vaccination group HBc-CD70. Taken together, these results suggest that therapeutic vaccination using a protein-prime, MVA-vector boost scheme leads to T cell proliferation in the liver, which may be enhanced by the co-expression of CD70.

Discussion
In this study, we investigated the potential of co-expressing CD70 in the context of an MVA-vector vaccine to enhance the quantity and quality of the CD8 T cell response directed against a transgenic target utilizing a protein-prime, MVA-boost vaccination strategy [25,26]. We therefore integrated a CD70 coding sequence into an MVA-vector under the same promotor as an HBV core open reading frame serving as a model antigen and showed the co-expression of HBV core protein and CD70. In HBV-naïve mice, coexpression of CD70 during boost vaccination doubled the number of HBV core-specific CD8 T cells and improved their functionality indicated by the secretion of IFNγ, TNFα and IL2, while T cell immunity specific for the MVA-vector was not significantly affected. Furthermore, CD70 co-expression was able to trigger IFNγ production in a higher portion of HBV core-specific CD8 T cells after vaccination in comparison to mice receiving vaccination without CD70 co-expression resulting in increased levels of functional T cells. In HBV1.3tg mice, a minor effect was observed on T cells derived from the spleen and liver of lowviremic mice after vaccination. In high viremic mice, however, the immune response after vaccination was reduced independently of the co-expression of CD70. Additionally, no benefit of CD70 co-expression could be detected.
Liver histology of the HBV1.3tg mice showed that CD70 co-expression increased the number of vaccine-induced, proliferative T cell clusters. We thus conclude that T cells in proliferative cell clusters are most likely HBV core-specific and that CD70 co-expression has a beneficial effect on T cell proliferation. The overall number of CD3+ cells in the liver was not influenced by vaccination. This was expected as the majority are single T cells that traffic through the liver but are not specific for HBV core [16]. Interestingly, the number of T cell clusters was independent of the viremia of mice while the analysis of HBV core-specific T cells by flow cytometry showed the previously observed negative correlation (lower CD8 T cell responses in high viremic mice). In conclusion, in low-medium viremic mice, CD70 co-expression during boost vaccination seems to provide benefits for T cell function and proliferation, especially in the liver, while in high viremic mice, these effects could not be confirmed. A high antigen load is assumed to be one factor which leads to immune tolerance enabling chronic infection and is one of the main hurdles in preclinical studies of therapeutic vaccination against chronic HBV infection [25,26]. At this place, co-expression with CD70 was not able to overcome this obstacle and it has been shown that other strategies such as the knockdown of antigens prior to the therapeutic vaccination are more promising [26].
The current study is limited by the fact that only one antigen was used and only one prime and one boost vaccination was applied in contrast to more elaborate vaccination schemes such as the TherVacB scheme [26]. Whether the potential of CD70 co-expression in a therapeutic vaccination setting against hepatitis B is underestimated by the choice of the HBV core instead of the more potent HBV surface antigen as a model was not assessed and would be an interesting subject for further investigation. An additional limitation was the model used because HBV1.3tg mice have a depleted core-specific T cell repertoire because they start expressing the HBV core antigen already before birth. This limits the predictive value of this study for a therapeutic hepatitis B vaccine.
However, our data provide valuable information on how CD70 can be used as a safe component in a boost-vector design to precisely enhance transgene-specific CD8 responses to a vaccine antigen. In a physiological setting, CD70 expression on immune cells is highly regulated [39] and overexpression and prolonged exposure to CD70 can have detrimental effects on cellular immunity leading to exhaustion of the naïve T cell pool [52]. Hence, it was important to show that delivery by MVA is a safe application of this co-stimulator: infected cells undergo apoptosis within a few days [53], but this time-frame seemed to be sufficient to selectively improve the desired CD8 T cell recall response. There is evidence that MVA preferentially infects antigen-presenting dendritic cells, which upon infection with MVA are impaired in their maturation due to the shutdown of cellular protein production in MVA infection [53][54][55]. CD70 is naturally upregulated in mature dendritic cells and helps to form the effector and memory T cell pool [30,39]. Transgenic CD70 provided by rMVA could compensate this the lack of maturation and cause the observed improvements in CD8 T cell effector function. It has been shown that the expression of CD70 by immature dendritic cells is capable of converting CD8 T cells from a tolerogenic to an immunogenic state which interestingly occurs independently of CD4 T cell help [56,57].
Bathke et al. [58] designed rMVA-CD70 constructs for homologous prime-boost experiments and observed that CD8 T cell responses against the MVA epitope B8R were stronger after CD70 co-expression via two rMVA immunizations than those against the target transgene-ovalbumin (OVA). Such misdirected effects could lead to unspecific and harmful immune reactions. With our vaccination scheme and vector design, we obtained a stronger improvement of CD8 T cell responses against HBV core as the model antigen than that against the MVA-specific epitope B8R. The major difference in the study design was our protein prime that raised the target-specific T cell responses over those directed against that of the viral backbone [25]. At the time of boost immunization, HBV core-specific CD8 T cells were already present in our mice. The boost itself was described to shape the immunodominance hierarchy [59] and appeared in our experiments even to be significantly improvable by CD70 co-expression-at least in naive mice. However, MVA-specific T cells were not yet present in our mice at the time of boost, so the MVA antigens in our experiments represented a baseline immunization that has already been shown to play no role in antigen competition [59]. Further, Bathke et al. expressed CD70 under the pHyb promoter while the target antigen OVA was expressed under the pS promoter. We spatially and temporally coupled the transgene expression of CD70 and the target gene utilizing only one promoter mH5 to generate one type of transcript, which via the P2A site results in equimolar expression of CD70 protein and the vaccine antigen. We consider this essential to drive the co-stimulation of cellular immunity against the target antigen.
In summary, we present here a safe system to selectively enhance the vector-vaccineinduced CD8 T cell response to the target antigen through CD70 co-expression during boost immunization. In this respect, our study may contribute to the development and improvement of therapeutic vaccines revealing novel generic design principles and thus foster advances of current strategies to fight malignancies and persistent viral infections.