Successive site translocating inoculation potentiates DNA/recombinant vaccinia vaccination

DNA vaccines have advantages over traditional vaccine modalities; however the relatively low immunogenicity restrains its translation into clinical use. Further optimizations are needed to get the immunogenicity of DNA vaccine closer to the level required for human use. Here we show that intramuscularly inoculating into a different limb each time significantly improves the immunogenicities of both DNA and recombinant vaccinia vaccines during multiple vaccinations, compared to repeated vaccination on the same limb. We term this strategy successive site translocating inoculation (SSTI). SSTI could work in synergy with genetic adjuvant and DNA prime-recombinant vaccinia boost regimen. By comparing in vivo antigen expression, we found that SSTI avoided the specific inhibition of in vivo antigen expression, which was observed in the limbs being repeatedly inoculated. Employing in vivo T cell depletion and passive IgG transfer, we delineated that the inhibition was not mediated by CD8+ T cells but by specific antibodies. Finally, by using C3−/− mouse model and in vivo NK cells depletion, we identified that specific antibodies negatively regulated the in vivo antigen expression primarily in a complement depended way.

Longer intervals between immunizations often seem to enhance quality and duration of immune response 18 . Herein, we immunized female mice with pSV-OVA SFI or SSTI in a 4-week interval (Fig. 2D), and 2 weeks after the last vaccination, immune responses were measured. The average percentage of OVA 257-264 specific IFN-γ secreting cells in CD8 + T cells elicited by SSTI was significantly higher than SFI (1.382 ± 0.2086% vs 0.6268 ± 0.03014%,  (Fig. 2E). And the mean OVA-specific antibody titer induced by SSTI was also significantly higher than that of SFI (p = 0.0001) (Fig. 2E).
The improvement of immunogenicity was not due to the redistribution of specific CD8 + T cells and was not limited to OVA antigen. To investigate whether the enhancement of cell-mediated immune response was due to the redistribution of specific CD8 + T cells, we collected the draining lymph nodes and spleens 2 weeks post the final vaccination. For each mouse, inguinal and iliac LNs were pooled together and the lymphocytes were isolated for ICS assay. Our data showed that SSTI induced significantly higher OVA 257-264 specific CD8 + T cell responses in both draining LNs (p = 0.0155) and spleen (p = 0.0433) (Fig. 3A,B). The improvement seemed even more obvious in draining LNs.
We proceeded to explore whether our observations have a wide application. Immunization of mice with a DNA vaccine encoding full length Pol protein derived from HIV-1 AE2f also showed that total T cell responses elicited by SSTI were significantly higher than SFI (732 ± 495 SFCs/10 6 splenocytes vs 325 ± 193 SFCs/10 6 splenocytes, p = 0.048) (Fig. 3C). Moreover, our data showed that specific CD8 + T cell responses induced by SSTI could be further augmented by a genetic adjuvant (cholera toxin subunit A, CTA) (1638 ± 800 SFCs/10 6 splenocytes), which was shown to be able to improve the immunogenicity of DNA vaccine in our previous work 11 (Fig. 3C). This suggested that SSTI could work in synergy with other immunogenicity enhancing approaches.
SSTI improved the magnitude of CD8 + T cell responses induced by both recombinant vaccinia alone and DNA-prime/recombinant vaccinia-boost, but not by protein vaccination. To broaden our understanding of SSTI further, we tested whether it could also enhance the immunogenicities of vaccines in modalities other than DNA. Mice were immunized with three different regimens: recombinant vaccinia alone, DNA-prime/recombinant vaccinia-boost and OVA protein alone. For each regimen, two parallel groups were designed for the comparison of SSTI and SFI (Fig. 4A). 2 weeks after the 3 rd vaccination, the mice were euthanized and specific IFN-γ secreting CD8 + T cells were measured by ICS. Compared to SFI, SSTI augmented specific CD8 + T cell responses induced by recombinant vaccinia alone by 2.25-fold (3.751 ± 1.077% vs 1.664 ± 0.462%, p = 0.0004) (Fig. 4B) and augmented specific CD8 + T cell responses elicited by DNA-prime/recombinant vaccinia-boost by 2.37-fold (3.919 ± 0.7164% vs 1.656 ± 0.794%, p = 0015) (Fig. 4D). Very interestingly, when being immunized with purified OVA protein, the average specific CD8 + T cell responses induced by SSTI was significantly lower than SFI (0.1038 ± 0.05868% vs 0.5870 ± 0.3604%, p = 0.0382) (Fig. 4C). The same tendency was also observed in binding antibody titer, although no statistical significance was reached (Fig. 4C).
Taken together, the above data revealed that SSTI could augment the specific T cell responses induced by DNA and recombinant vaccinia vaccines, but not by protein vaccine. SFI significantly reduced the in vivo antigen expression compared to SSTI, in an antigen specific way. The opposite influences on protein versus DNA/recombinant vaccinia vaccination implied that the in vivo antigen expression might be the key step that determined the difference between SSTI and SFI. To test this hypothesis, we inoculated mice with a luciferase expressing plasmid either by SSTI or SFI (Fig. 5A). Each mouse received two injections (one injection for each hind limb) for the 3 rd immunization. Three days after the 3 rd inoculation, tibialis anterior muscles of both hind limbs were isolated and homogenized for luciferase activity assay. Our data showed that the average luciferase activity in the repeatedly inoculated limbs of SFI group was significantly lower than that in the non-repeatedly inoculated limbs (paired t-test, p = 0.0462) (Fig. 5B). And it was also significantly lower than the average luciferase activity in the repeatedly inoculated hind limbs of control group, which received empty plasmids for the first two shots (p = 0.0218) (Fig. 5B). The average luciferase activities were similar between two hind limbs of SSTI group, both of which were significantly higher than the repeatedly inoculated limbs of SFI group (Fig. 5B). As there was no significant difference between the non-repeatedly inoculated limbs of SFI group and SSTI group, we then used the non-repeatedly inoculated hind limb of SFI group as the normal expression control in the following in vivo expression assays. In vivo live imaging also showed that the signal intensities of the repeatedly inoculated (right) limbs were obviously weaker than those of the non-repeatedly inoculated (left) limbs in SFI group (Fig. 5C). The observation was repeated by replacing the luciferase DNA with plasmids encoding luciferase-OVA or eGFP-OVA fusion protein and measured the in vivo expression by IP-WB. We found the expression of luciferase-OVA and eGFP-OVA were obviously decreased as well in the repeatedly inoculated limbs (Fig. 5D).
To further explore whether the reduction of antigen expression in the repeatedly inoculated limbs of SFI group was due to the pre-existing antigen-specific immune responses, we inoculated mice with OVA plasmid for the 1 st and the 2 nd time, but with luciferase plasmid for the last inoculation (Fig. 5E). Our data showed that the pre-vaccination of OVA didn't interfere with the expression of luciferase, which indicated that SFI inhibited DNA vaccine expression in an antigen specific way (Fig. 5F).

Compensating antigen expression by escalation the DNA dosage during site-fixed inoculation improved specific T cell responses to the level similar with SSTI.
To clarify whether the reduction of in vivo antigen expression was the direct cause of decreased immunogenicity in SFI group, we designed a dose-escalation vaccination experiment (Fig. 6A), in which, each mouse received 50μ g DNA for the first two shots and received 100 μ g DNA for the 3 rd shot. In vivo antigen expression was monitored at 2, 4 and 6 days post the last vaccination. Ranging from high to low, the order of AUC (area under dynamic curve of in vivo expression) was: moc k > SSTI > dose-escalation > SFI (Fig. 6B), suggesting that dose-escalation improved the in vivo antigen expression of the 3 rd shot in SFI group, but it was still lower than SSTI group. ICS assay revealed that dose-escalation improved the specific CD8 + T cell responses induced by SFI, although no statistical significance was reached (Fig. 6C).
CD8 + T cells migrated to the inoculation site after the 3 rd immunization, but didn't contribute to the reduction of antigen expression. Three days after the 3 rd immunization, the tibialis anterior muscles of inoculation site were harvested for immunohistochemistry assay (Fig. 7A). Compared to SSTI, SFI attracted more infiltrated CD8 + T cell (Fig. 7B). And the infiltration could be largely blocked by administering FTY720 (Fig. 7B), indicating that the infiltrated cells were mainly migrated from the draining LNs. Further, we isolated the lymphocytes from the muscles of injection site 1 day before and 3 days after the 3 rd inoculation. Flow cytometry analysis showed that the OVA (SIINFEKL) tetramer + CD8 + T cells were rare before the 3 rd vaccination, and their number increased sharply after the 3 rd vaccination. But, no significant difference was observed between SFI and SSTI (Fig. 7C).
To verify whether the infiltrated CD8 + T cells contributed to the reduction of in vivo antigen expression, we depleted mouse CD8 + T cells by giving depletion antibody intravenously one day before the 3 rd vaccination. Our data showed that the expression of luciferase in the leg that received 3 shots was still significantly lower than its expression in the other leg that received only one shot (p = 0.0065) (Fig. 7D). In a parallel experiment, mice were treated with FTY720, before the 3 rd vaccination to block the migration of CD8 + T cells. Consistently, we found the in vivo expression of luciferase wasn't improved (Fig. 7E). These data implied that the reduction of in vivo antigen expression was not caused by specific CD8 + T cells.
Specific antibodies led to the suppression of in vivo antigen expression in a complement dependent way. Ruling out the contribution of CD8 + T cell, then we were curious about whether specific antibody responses played a role in suppressing in vivo antigen expression, since the OVA specific antibody titers in the The expression of Luciferase-OVA was measured by detecting luciferase activity, which showed the expression was significantly lower in the limb being repeatedly inoculated by SFI (n = 4). (C) Live imaging also showed the expression of Luciferase was much lower in the limb being repeatedly inoculated by SFI for 3 times. (D) The expression of Luciferase-OVA and eGFP-OVA were detected by western blotting after immunoprecipitation (IP-WB). The IP antibody (10 μ g/sample) and the WB first antibody (diluted at 1:1000) was rabbit anti-OVA, and the WB second antibody was HRP-labeled goat anti rabbit (diluted at 1:5000). (E,F) The in vivo antigen expression was inhibited in an antigen specific way. Mice were immunized according to the schedule shown in (E). Luciferase expression was comparable between the repeatedly (right) and non-repeatedly (left) inoculated limbs in OVA-OVA-luc group, while, in luc-luc-luc group, the expression of Luciferase was much higher in the non-repeatedly inoculated limb(left) compared to repeatedly inoculated limb(right)(n = 5) (F).
lysate tibialis anterior muscles of injection site were obviously higher in SFI group (Fig. 8A & Table 1). And this observation was supported by transcriptomic microarray analysis, which showed that the transcriptions of several Ig genes were significantly upregulated at the injection site of SFI group (Fig. 8E).
To test this hypothesis, we inoculated mice with the mixture of pSV-Luc-OVA DNA and purified serum IgG (Fig. 8B). For each mouse, 50μ g pSV-Luc-OVA were mixed with 100μ g total IgG freshly purified from either OVA vaccinated or empty plasmid vaccinated mice sera. Our data showed that total IgG purified from OVA vaccinated mice sera significantly suppressed the expression of luciferase-OVA (Fig. 8C), the average expression of luciferase decreased by > 1 log10 (paired t test, p = 0.0409). The suppression could only be slightly counteracted by NK depletion, as our data showed that the average expression of luciferase still decreased by 0.76 log10 (paired t test, p = 0.005) in the limbs inoculated with the mixture of pSV-Luc-OVA and total IgG purified from OVA vaccinated mice sera (Fig. 8D). While, in C3 −/− mice, the suppression of in vivo luciferase expression could be almost completely counteracted (Fig. 8D), suggesting that the specific antibodies mainly functioned through a complement dependent way under this setting.

Discussion
Over the past 20 years, DNA vaccines have proved effective in animal models 7 and have been intensively evaluated in human trials 3 . Compared to conventional vaccines, DNA vaccines have a number of advantages, including ease of development and production, the ability to present antigen by both MHC class I and class II molecules 19 , and good safety record. However, an obvious disadvantage shown in human DNA vaccine trials has been their suboptimal immunogenicity when compared with traditional vaccines 7 . In spite of great efforts that have been made 4 , further optimizations are still needed to get the immunogenicity of DNA vaccine closer to the level required for human use. In this study, we report for the first time that changing the inoculation site during multiple vaccinations (designated as successively site translocated inoculation, SSTI) can significantly enhance specific T cell responses elicited by DNA vaccines in both spleen and draining lymph nodes, compared to conventional anatomical site-fixed inoculation (SFI). While most previous efforts focused on the optimizations of plasmid design, codon preference, adjuvants, delivery tools or heterologous prime-boost regimens, our finding represents a new way to improve the immunogenicity of DNA and viral vector vaccines. Moreover, our data highlight that SSTI can work in synergy with genetic adjuvant (such as CTA) and DNA prime-recombinant vaccinia boost regimen. Several previous studies 18,20 showed the immunization intervals significantly influenced the specific immune responses. Our data showed that SSTI also improved the immunogenicities of DNA vaccines that were administrated at an interval of 4 weeks.
Live attenuated and purified proteins are two of the major vaccine formalities under clinical use 21,22 , therefore, we also tested whether their immunogenicities could also be augmented by SSTI. A recombinant Tiantan vaccinia expressing OVA was used as an example of live attenuated vaccine and OVA protein was used as an example of purified protein vaccine. We found that SSTI significantly enhanced the immunogenicity of the recombinant Tiantan vaccinia. But to our surprise, the immunogenicity of OVA protein was weakened by SSTI, compared to SFI. These observations highlight the necessity of inoculating protein vaccine at the same site (limb), while imply that the DNA and viral vector vaccines should be injected at different limbs during multiple vaccinations.
Since the major difference between DNA/viral vector vaccines and protein vaccine is that the former need to express antigen in vivo, thus we postulated that giving DNA/viral vaccines by SFI may inhibit the in vivo antigen expression. And our data proved this hypothesis by showing that the expression of DNA vaccines after the 3 rd inoculation was significantly lower in the limbs that were repeatedly immunized. And we confirmed this is the leading cause of immunogenic difference between SSTI and SFI, because increasing the DNA dosage for the 3 rd inoculation could bring the immune responses induced by SFI to the level similar with SSTI. A reasonable explanation for this phenomenon is that the locally existed specific immune responses may inhibit the expression of "non-self " antigens encoded by DNA vaccines and our data supported this conjecture by showing that immunizing mice with pSV-OVA for the first two vaccinations did not interfere with the expression of luciferase for the third inoculation. It is quite similar with the mechanism that has been proposed for the immune surveillance of tumor 23 . But unexpectedly, our data showed that the inhibition is not mediated by T cells, but by specific antibodies. High titers of specific antibodies existed in the muscles which were repeatedly inoculated for 2 times or more,

Figure 7. CD8 + T cells migrated to the inoculation sites, but didn't contribute to the inhibition of antigen expression. 3 days post the last vaccination (A)
, tibialis anterior muscles were collected for hematoxylineosin (H&E) and histochemical staining. Much more CD8 + T cells migrated to the muscles in SFI group, and the migration could be blocked by administration of FTY720 (B). (C) leukocytes were isolated from tibialis anterior of the limbs that received the 3 rd immunization. Specific CD8 + T cells responses were detected by OVA tetramer + staining. Depleting CD8 + T cells (D) or blocking the migration (E) before the 3 rd immunization did not counteract the reduction of antigen expression in the limbs being repeatedly inoculated. (D: n = 6/7; E: n = 5) and these antibodies inhibited the later on expression of DNA vaccines in a complement dependent way. Of note, in addition to antigen specific antibody response, the anti-vector antibody response may also be an important factor that can restrict the replication of vaccinia and thus reduce its expression, which has been shown by several previous studies [24][25][26][27] .
To explain the difference of DNA vaccine expression between SFI and SSTI, we measured the antibody responses at the injection sites and our data showed that both the specific antibody titer and the transcriptions of Ig genes in SSTI group were much lower than SFI group (Table 1 & Fig. 8E), which we think is the main reason for the better expression of DNA vaccines in SSTI group. The presence of high Ig transcription level at the injection site of SFI group implied the existence of plasma B cells in the local environment, which might be due to the long lasting residual expression of DNA vaccine 28 . Besides, our microarray data also showed significantly different transcriptomic profiles at the injection sites of SFI group and SSTI group (Fig. 8E), which raised the possibility mice and in vivo NK cell depletion were used to investigate the mechanism of antibody mediated expression reduction. C3 −/− mice showed comparable Luc-OVA expression between repeatedly (right) and non-repeatedly (left) inoculated limbs (n = 6), while NK cell depletion did not counteract antibody dependent suppression (n = 6). (E) Three days post the 3 rd vaccination, the injected muscles were isolated and total RNA was extracted for microarray analysis. Total 336 differentially transcribed genes were identified, including 67 upregulated (red ID) and 269 downregulated genes (green ID) (SFI vs SSTI, absolute fold change > 2, p < 0.05, n = 3).
that the local microenvironment might contribute to the complement response. Another interesting finding of our study was that depleting NK cells only partially counteracted the suppression of in vivo antigen expression. Potential explanations for this phenomenon include: first, in addition to NK cell, myeloid cell (e.g. macrophage) can also mediate ADCC 29,30 ; second, the suppression mediated by NK cells could be compensated to some extent by complements, as the complements are soluble factors and readily available in many tissues.
As most clinically used vaccines are injected into either side of two arms, we then tried to mimic this situation by optimizing SSTI into two limbs. However, we found that the expression of antigen was still obviously suppressed even when the DNA vaccine was inoculated into two different muscles of the same hind limb (data not shown), which was presumably due to the small figure size of mouse. Hence, we chose vaccination in three limbs as an alternative for this study. In spite of this limitation, our study suggests that changing the inoculation site (limb) successively (SSTI) during multiple vaccinations can improve the immunogenicity of both DNA and viral vector vaccines. But for protein vaccines and inactivated vaccines, inoculation at the same site is better, since they do not need to express antigens in vivo. We believe this strategy is of practical significance, especially when it can be properly used in combination with other immunogenicity enhancing approaches.

Materials and Methods
Ethics statement. Animal care and experiments were reviewed by the Institutional Animal Care and Use Committee (IACUC) of Shanghai Public Health Clinical Center and were performed in strict accordance with the approved protocol (Permit Number: 2013-E013). All experiments were performed at least two times with similar results. One representative result is shown.

Reagents and vaccines. Mouse H-2K b matched OVA peptide (SIINFEKL) was synthesized by Shanghai
Science Peptide Biological Technology Co. ltd. HIV-1 AE2f Pol peptides (15-mers overlapping by 11 amino acids with its next peptide) were synthesized by GL Biochem (Shanghai) Ltd. The purity of all peptides was equal to or higher than 95%. H-2K b OVA tetramer-SIINFEKL-APC was purchased from MBL (Cat# TS-5001-2C).
DNA vaccines expressing OVA, Luciferase, Luciferase-OVA, eGFP-OVA, HIV-1 AE2f Pol, or HIV-1 AE2f Pol-CTA were constructed previously by cloning the target genes into pSV-1.0 vector and all DNA vaccines used in this study were prepared by using the Endo-free Plasmid Giga Kit (Qiagen, Cat#12391) and reconstituted in sterile normal saline. Recombinant Tiantan vaccinia (rTTV) encoding OVA was also constructed in our previous work 10 . Purified OVA protein was purchased from Sigma (Sigma, Cat# A5378) and dissolved in sterile normal saline. Mice vaccination. 6 to 8 week-old female C57BL/6 mice were housed under specific-pathogen free environment. All vaccines were injected intramuscularly (i.m.) and the detailed vaccination schedules and regimens were shown along with the results of each experiment. The dose of DNA vaccine was 50 μ g/mouse except the 3 rd vaccination of the dosage escalation experiment in which 100 μ g/mouse was used. The dose of recombinant vaccinia vaccine was 1 × 10 6 pfu/mouse and the dose of protein vaccine was 20 μ g/mouse. Intracellular cytokine staining. Freshly isolated splenocytes or lymphocytes were plated into round-bottom 96-well plates (2 × 10 6 cells per well) and stimulated with OVA (257-264) (SIINFEKL) peptide at the final concentration of 5 μ g/ml. 1 hour later, brefeldin A and monesin were added to each well at final concentration of 1 μ g/ml and 1 μ M. Another 7 hours later, the stimulation were stopped by washing the plates with R10 medium and 4 cell surface markers were stained on ice with fluorescein labeled antibodies, including PerCP-Cy5.5-labeled anti-mouse CD3 (Clone: 17A2, Biolegend Cat# 100218), Pacific Blue labeled anti-mouse CD8 (Clone: 53-6.7, Biolegend Cat# 100725), FITC-labeled anti-mouse CD44 (Clone: IM7, eBioscience Cat# 11-0441-81), and APC-eFluor780-labeled anti-mouse CD62L (Clone: MEL-14, eBioscience Cat# 47-0621-80). Next, the cells were fixed and permeablized, and intracellular IFN-γ was stained with PE-conjugated anti-mouse IFN-γ (Clone: XMG1.2, Biolegend Cat# 505808). Stained samples were measured using BD FACS Aria I. Data analysis was done by using FlowJo X software (Tree Star, Inc). Gating strategy was shown in Fig. 1B. IFN-γ ELISPOT assays. Enzyme-linked immunosorbent spot (ELISPOT) assays for IFN-γ release were performed by using mouse IFN-γ ELISPOT kits (BD Bioscience, Cat# 551083). The 96-well ELISPOT plates were coated with purified anti-mouse IFN-γ monoclonal antibody overnight at 4 °C. The plates were then blocked, and 2 × 10 5 fresh splenocytes were added into each well and incubated with peptide for 20 h in a 37 °C 5% CO 2 incubator. The final concentration for each peptide was 5 μ g/ml. After incubation, detecting antibody and Avidin-HRP were added one after another. Subsequently, the plates were developed according to the manufacturer's manual. Spots representing IFN-γ producing cells were enumerated by using an automated ELISPOT plate reader (ChampSpot III Elispot Reader, Saizhi, Beijing, China). ELISA. The serum antibodies against OVA were assessed using an enzyme-linked immunosorbent assay (ELISA) as previously described 11 . Briefly, 96-well EIA/RIA plates were coated overnight with 1 μ g/ml of OVA protein at 4°C. Serially 2-fold diluted mouse sera were added to each well after being blocked with 200 μ l 5% powdered milk. Binding Abs were detected by adding 1:2000 diluted goat anti-mouse IgG (HRP labeled) (Santa Cruz Biotechnology, Cat# sc-2005). After 1 hour incubation, 100 μ l OPD substrate (Thermo Scientific, Cat# 34006) was added to each well. 15 minutes later, the reaction was stopped by adding 50 μ l of 2N H 2 SO 4 . The endpoint antibody titers were defined as the last reciprocal serial serum dilution at which the absorbance at 492 nm was greater than 2× (negative mean + SD).
Antibody titer at the inoculation site. was measured by ELISA. We isolated tibialis anterior muscle at 48 h post the last vaccination and homogenized with 600 μ l Pierce IP Lysis Buffer (Thermo Scientific, cat# 87788). Supernatant was used for serial dilution (started from 1:10), and performed with the same method described above.
Administration of FTY720. From the day of the 3 rd vaccination through SFI, mice were treated intraperitoneally every other day, with doses of 20 μ g FTY720 (Cayman Chemical, Cat# 10006292) per mouse (1 mg/kg) in a final volume of 0.2 ml. The control mice were injected with 0.2 ml saline only 31 .
In vivo luciferase expression analysis. Three days post the final pSV-Luciferase or pSV-Luciferase-OVA vaccination, C57BL/6 mice were euthanized, and the tibialis anterior muscle was excised and homogenized with 600 μ l 1× luciferase cell lysis buffer (Promega, Cat# E1531). After 30 minutes incubation on ice, 50 μ l supernatant of each sample was transferred to 96-well plate, following with 50 μ l luciferase assay substrate (Promega, Cat#E1501 In vivo imaging of luciferase expression. The mice were vaccinated with luciferase DNA through either SFI or SSTI. Luciferin for in vivo use(Promega, Cat# P1043) was injected into tibialis anterior muscle on the 3 rd day post last vaccination and ten minutes later, the mice were anaesthetized and the in vivo imaging was performed by using the In-Vivo Xtreme (Bruker Corporation, Cat#1815538).
In vivo OVA expression assay. Three days after the final pSV-luciferase-OVA or pSV-eGFP-OVA vaccination, tibialis anterior muscles of the immunized mice were homogenized and lysed with Pierce IP Lysis Buffer. Rabbit anti-OVA polyclonal antibody was added to the supernatant, and the antibody/OVA complex was pulled down by Protein A+ G agarose beads (Beyotime, Cat# P2012). The expression of luciferase-OVA or eGFP-OVA was determined by western blotting.
Leukocytes isolation from tibialis anterior muscle. Three days post the final vaccination, C57BL/6 mice were euthanized and tibialis anterior muscles were excised and samples from mice in the same group were pooled together. Infiltrated lymphocytes were isolated according to a previously reported procedure with minor modification 32 . Briefly, the excised tibialis anterior muscle tissues were first washed with PBS and minced into small pieces. Then, the tissue pieces were digested in 0.2% collagenase IV (Life Technology, Cat# 17104-019) at 37 °C for 1 hour in a shaking incubator. Next, the tissue pieces were loaded and grinded in 70 μ m nylon strainers. The passing through single cells were washed with R10 medium (RPMI-1640 supplemented with 10%FBS and 1%PS) and resuspended with 10 ml EZ-Sep TM Mouse 1× lymphocytes isolation buffer (DAKEWE, Cat# DKW33-R0100). Isolation was completed following the manufacturer's instruction. After a 500 xg centrifuge for 30 min, the leukocytes were harvested and washed with R10 twice, and then applied for fluorescent antibody and OVA tetramer staining.
Serum IgG purification. Total IgG was purified from the sera of either pSV-OVA immunized mice or blank mice by using Protein G HP Spin Trap (GE Healthcare, Cat# 28-9031-34). And the OVA binding activities of the purified IgG were tested by ELISA.
In situ detection of infiltrated CD8 + T cells. Three days after the final OVA DNA vaccination, the tibialis anterior muscles were excised and freshly fixed for immunohistochemistry assay. For each sample, three consecutive slides (thickness 4 μ m) were stained with H&E, anti-CD3 and anti-CD8 mAb, respectively. The slides were scanned with Tissue FAXS Plus (Tissue Gnostics Gmbh, Austria).

Statistical analysis.
All statistical analyses were done by using GraphPad Prism 5.0 (GraphPad Software, Inc).
Comparisons between two groups were analyzed by the method of t test. First check whether data follow normal distribution. If yes, t test is an appropriate measure of mean comparison. If not, then the Mann-Whitney test will be more appropriate for the same objective. Comparisons among three or more groups were done by One-way ANOVA. Significant difference was defined as p < 0.05.