Gene gun-mediated DNA vaccination stimulates an immune response characterized by the activation of IgG-secreting B cells and IFN-γ-secreting T cells. To monitor the contribution of cells at the site of vaccination to this process, transfected skin was periodically removed and grafted onto naive recipients. Immediate removal of vaccinated skin abrogated the development of an immune response. Low-level IgG production was stimulated when the vaccination site was left in place for ≥5 h, with the strength of this response increasing the longer the site remained intact (for up to 2 wk). Measurable primary T cell responses were observed in animals whose vaccination site remained in place for ≥1 day. Skin grafts transferred 0 to 24 h postvaccination stimulated a primary immune response in naive recipients. Memory B and T cells were generated in animals whose site of vaccination remained intact for 5 to 12 h. Skin transferred within 12 h of vaccination triggered memory B and T cell development in graft recipients, while the removal of skin >12 h postvaccination did not reduce memory in vaccinated mice. These findings suggest that 1) primary immunity is induced by cells that migrate rapidly from the site of immunization, 2) nonmigratory cells influence the magnitude of this primary response, and 3) migratory cells alone are responsible for the induction of immunologic memory.
Priming the host to mount a protective memory response against pathogenic microorganisms is a central goal of vaccination. Vaccine development has been revolutionized by the finding that DNA plasmids can induce memory responses against infectious agents (1, 2, 3, 4, 5, 6). DNA vaccines are composed of an Ag-encoding gene regulated by a strong mammalian promoter incorporated into a plasmid of bacterial DNA (1, 7, 8). Host cells take up plasmid, transcribe and translate the encoded gene, and produce protein that stimulates an immune response when presented to the immune system in the context of self-MHC (1, 3, 6, 7).
When DNA vaccines are administered by gene gun bombardment of the skin or by i.m. injection, a majority of the plasmid is taken up by keratinocytes and muscle cells, respectively (7, 9, 10, 11). Recent evidence suggests that these nonmigratory cells do not contribute to the development of immunity, since primary Ab and CTL responses can be generated despite immediate removal of the site of vaccination (12, 13). Some postulate that primary responses are therefore induced by plasmid or plasmid-containing cells that rapidly leave the site of vaccination. Evidence suggests that bone marrow-derived APC acting far from the site of vaccination contribute to this process (13, 14, 15, 16). Consistent with that hypothesis, dendritic cells in the skin were recently shown to take up DNA-coated beads and migrate within 24 h to the draining lymph nodes (14), where primary immunity develops (17).
Despite ongoing progress in understanding the location and type(s) of cells that contribute to DNA vaccine-mediated primary immune responses, the role of cells at the site of injection to the magnitude of the primary response, and the generation of immunologic memory remains unclear. We therefore analyzed Ab and cytokine production resulting from gene gun administration of a plasmid encoding the circumsporozoite protein (CSP)3 of malaria to BALB/c mice. In this well-defined system, primary IgG anti-CSP Ab production arises and can first be detected 2 to 3 wk postimmunization, while secondary humoral responses achieve near-maximal levels within 1 wk of boosting (17). The plasmid-encoding CSP (pCSP) vaccine also activates IFN-γ-secreting T cells with a memory component that can be detected by restimulation in vitro with Ag (18). Using this model system, we probed the contribution of gene gun-vaccinated skin to the development of primary and secondary immune responses.
Materials and Methods
Animals and immunizations
Female BALB/c mice were obtained from Charles River Associates (Wilmington, MA) and maintained in our specific pathogen-free facility. Mice were immunized by gene gun (Bio-Rad, Hercules, CA) administration of 1 μg of pCSP to shaved abdominal skin through an 8-mm hole cut through a Tegaderm mask (3M, St. Paul, MN). Animals were boosted 10 wk later. Serum was collected 1, 3, and 6 wk postimmunization and 1 wk postboost by retro-orbital puncture and stored at −20°C until use. In some experiments, mice were killed by cervical dislocation 4 wk postimmunization and their organs removed aseptically.
The CS.1 protein used in ELISA assays is an immunoaffinity-purified fusion protein produced in Escherichia coli, and consists of amino acids 64 to 321 of the intact PyCSP protein fused to 81 amino acids of the nonstructural protein of influenza A (2, 19).
The pCSP DNA vaccine (kind gift of VICAL, Inc., San Diego, CA; company designation VR2507) was constructed by cloning the coding sequence of the Plasmodium yoelii CSP from the nkCMVintPyCSP.1 plasmid (2) into the VR1012 vector, as previously described (20). Expression of pCSP was tested by in vitro transfection of HeLa cells and immunoblot analysis of cell lysates.
DNA-coated gold beads were prepared by combining 100 μg of pCSP and 100 μl of 0.1 M spermidine with 50 mg of gold beads. Plasmid was precipitated onto the beads by slowly adding 200 μl of CaCl2 while vortexing, as previously described (21). Coated beads were centrifuged at 1,000 × g and washed three times in cold 100% ethanol. Beads were resuspended and bound to the inner surface of Gold-Coat Tubing (Bio-Rad) as recommended by the manufacturer. A total of 1 μg of DNA was then delivered to the shaved abdominal skin of adult BALB/c mice using a Helios gene gun (Bio-Rad).
Skin excision/grafting experiments
Donor skin at the vaccination site was removed to the level of the panniculus carnosus by excisional biopsy (10 mm). Subcutaneous tissue was removed from donor skin, which was then grafted heterotopically onto the flanks of prepared recipients. The transplant site was covered with sterile gauze and a pressure bandage for 9 days.
Cytokine ELISA and ELIspot assays
Ninety-six-well Immulon 2 microtiter plates (Dynatech Laboratories, Alexandria, VA) were coated with 10 μg/ml of anti-IFN-γ (clone RA6a2; Lee Biomolecular Research Laboratories, San Diego, CA) in 0.1 M carbonate buffer (pH 9.6) for 3 h at room temperature (22). The plates were blocked with PBS-5% BSA for 1 h and washed with PBS-0.025% Tween-20. Culture supernatants or spleen cell suspensions prepared in complete medium (RPMI 1640 supplemented with 10% heat-inactivated FCS, 1.5 mM l-glutamine, and 100 U/ml of penicillin/streptomycin) starting with 1 × 106 cells/well, were incubated on anticytokine-coated plates for 8 to 10 h at 37°C in a humidified 5% CO2 incubator (23). Plates were then washed with PBS-Tween and overlaid with 1 μg/ml of biotinylated anti-IFN-γ (clone XMG 1.2; PharMingen, San Diego, CA) at 4°C and then treated with a 1:2000 dilution of avidin-conjugated alkaline phosphatase (Vector Laboratories, Burlingame, CA) for 2 h at room temperature. After a final wash, cytokine or ELIspot formation was detected by the addition of a solution of BCIP/NBT (Kirkegaard and Perry Laboratories, Gaithersburg, MD).
CS.1-specific ELISA assays
Ninety-six-well Immulon 1 microtiter plates were coated with 10 μg/ml of immunoaffinity-purified CS.1 protein in carbonate buffer, pH 9.5 (17). Plates were blocked with PBS-1% BSA, overlaid with serially diluted mouse serum, washed, and reacted with phosphatase-conjugated anti-mouse IgG (Southern Biotechnologies, Birmingham, AL). The concentration of specific Ab was determined by comparison to a standard curve generated using a high-titrated antiserum, as previously described (25).
In all experiments, the GMT of the test group was compared with that of a control group of similarly vaccinated mice. The % normal IgG anti-CSP response was calculated by the formula:
Measurement of primary and secondary immune responses
Previous studies showed that pCSP vaccination induces a primary IgG anti-CSP Ab response first detectable 2 wk postvaccination and peaking at 3 wk (17). Memory IgG responses peak at approximately 1 wk postboost. Primary cytokine responses are characterized by a significant increase in the number of spleen cells actively secreting IFN-γ in vivo 3 to 4 wk postimmunization (18). Memory T cells were detected by restimulating spleen cells from immunized mice for 4 days in vitro in the presence of 10 μg/ml of P16.
All studies were performed on groups of three mice (studied independently), and were repeated at least once with similar results. Data were analyzed by ANOVA using Fisher’s Exact test as described in the SigmaStat software package.
Optimization of vaccine delivery by gene gun
Initial experiments examined the effect of varying the injection pressure and size of vaccine-coated beads on the magnitude of the primary immune response. Gene gun administration of pCSP-coated beads to the abdominal skin of BALB/c mice at pressures ranging from 150 to 450 psi uniformly elicited strong IgG anti-CSP Ab responses. No response was induced when beads were administered at pressures below 100 psi. Optimal IgG anti-CSP Ab production was elicited by immunizing BALB/c mice with 1.6-micron gold beads delivered at 200 psi, and these parameters were used in all subsequent experiments.
Kinetics of the induction of primary and secondary immune responses
Condon et al. recently showed that dendritic cells are triggered to migrate from the skin to the draining lymph nodes following entrance of plasmid-coated beads (14). Whether nonmigratory cells in the epidermis also contribute to the induction/perpetuation of vaccine-induced responses has not been elucidated. To examine this issue, pCSP-vaccinated skin was removed and grafted onto syngeneic recipients at discrete time points following immunization. Donor and recipient BALB/c mice were then studied for the development of humoral (Ab) and T cell-specific (cytokine) responses.
As seen in Figure 1, neither Ab nor cytokine production was elicited in mice whose injection site was removed immediately after immunization. “Boosting” these animals resulted in an IgG anti-CSP response indistinguishable in magnitude or kinetics from that of naive mice immunized simultaneously, suggesting that the immediate removal of vaccinated skin also abrogated the development of immunologic memory. By comparison, primary IgG anti-CSP responses were elicited in animals whose vaccination site was left in place for only 5 h, although this response was 15-fold lower than that of mice whose vaccination site was never removed. Progressively stronger humoral responses developed as the vaccination site was left intact for longer periods, with near maximal levels achieved at ≥2 wk (Fig. 1, left panel).
Maximal B cell memory was generated over a much shorter period. Animals whose injection site was removed 12 h postvaccination mounted nearly optimal secondary responses, as demonstrated by the presence of high titrated IgG anti-CSP Abs 1 wk postboost (Fig. 1, right panel). In contrast, the primary Ab response generated in these mice was fivefold lower than that of mice whose vaccination site was never removed.
Primary T cell responses were evaluated by monitoring cytokine-producing cells in vivo. A significant increase in the number of spleen cells secreting IFN-γ was observed in mice whose vaccination site was left in place for ≥1 day, but not for ≤5 h (Fig. 2). Maximal cytokine production resulted when the site of immunization was left undisturbed for ≥1 wk. To monitor the generation of “memory” T cells, splenocytes from immunized mice were cultured in vitro with P16, a stimulatory T cell peptide present on the CSP protein. We previously showed that this 4-day culture activated quiescent Ag-specific T cells (18). Consistent with results involving memory B cells, the generation of memory T cells took place quite rapidly, requiring only 5 h (Fig. 2, right panel). Of note, no primary T cell response was detected at this 5-h time point (Fig. 2, left panel). Coupled with evidence that transfected dendritic cells migrate from the skin to the draining lymph nodes (14, 26), these findings suggest that immunologic memory is induced by dendritic cells that migrate from the skin 5 to 12 h postvaccination. Alternatively, it might take 5 to 12 h for nonmigratory cells to transcribe, translate, and secrete the encoded CSP protein required for the induction of the responses measured above.
Contribution of migratory and nonmigratory cells to the development of primary and secondary immune responses
To differentiate between these alternatives, vaccinated skin was removed from donor mice and grafted onto naive recipients. Skin transferred within 5 h of vaccination stimulated IFN-γ production by the graft recipient, consistent with the induction of a primary T cell response (Fig. 2). Yet no primary B cell response was detected in any skin graft recipient. Given that IgG anti-CSP Ab production was not induced, and that the magnitude of the primary T cell response was 40% lower than that of optimally vaccinated donors, we conclude that the immunogenicity of transfected epidermal cells was reduced by the process of excision and grafting.
Nevertheless, both B and T cell memory was generated when donor skin was grafted onto naive recipients within 12 h of vaccination. As seen in Figure 3, recipients of such skin grafts mounted an IgG anti-CSP response 1 wk post-pCSP “boost,” in contrast to recipients of skin grafted ≥1 day postimmunization, whose anti-CSP response developed 3 wk postboost. The timing and magnitude of the latter response is indicative of a primary rather than secondary humoral response. Moreover, memory T cells were present in recipients of skin vaccinated for ≤5 h, but not ≥1 day (Fig. 2). Since it is well documented that cells at the site of DNA vaccination continue to transcribe, translate, and secrete the encoded protein for many days postimmunization (27, 28), our results support the conclusion that migratory cells (rather than secreted protein) are responsible for the induction of immunologic memory in both T and B cell subsets, and that such cells are critical for inducing primary T cell responses.
The efficacy of a vaccine depends upon its ability to prime the host to mount a strong protective response when exposed to an infectious microorganism. The ability of DNA vaccines to elicit such responses has been demonstrated in a variety of animal models (1, 2, 5, 6, 29). The current study was undertaken to improve our understanding of the contribution of cells at the site of gene gun-mediated DNA vaccination to the generation of immunologic memory.
Dendritic cells in the epidermis function as specialized APCs, efficiently inducing immune responses against Ags that come into contact with the skin (26, 30). Mor et al. showed that the initial immune response to a DNA vaccine takes place in the draining lymph nodes (17) while Condon et al. showed that dendritic cells transfected by plasmid-coated beads migrate within 24 h from the skin to the draining lymph nodes (14). The precise kinetics of this dendritic cell migration were not delineated, nor was their relative contribution to the generation of primary vs memory responses.
The contribution of Ag-expressing nonmigratory cells to the development of immunity remains controversial. Williams et al. reported that plasmid introduced by particle bombardment directed cells in the epidermis to produce the encoded protein for at least 2 wk (31). Yet since neither muscle cells nor keratinocytes constitutively express the class II MHC or costimulatory molecules required to activate CD4+ Th cells (such as CD40 or B7) (32), their role in this process has been suspect. Yet Ulmer et al. showed that myoblasts transfected in vitro could generate a primary immune response when transplanted into naive recipients (33). Those investigators concluded that Ag shed by the myocytes was taken up and presented by professional APCs, since the induced immune response was restricted by the MHC type of resident APCs rather than transplanted myocytes (34). Further evidence that cells at the vaccination site contribute to the induction of immunity was supplied by Torres et al. (13), who showed that immediate removal of gene gun-vaccinated skin prevented the development of T or B cell responses.
However, this set of findings is inconsistent with two other reports in which normal immune responses were detected despite the immediate removal of muscle and/or skin following DNA vaccination (12, 13). In the latter cases, it is possible that the injected plasmid gained access to the lymphatic or circulatory systems, obviating the need for transfection of cells at the site of immunization. Indeed, plasmid can be detected in serum several hours after i.m. injection of a DNA vaccine (28).
The current studies were undertaken to address these conflicting results and to better define the role of migratory and nonmigratory cells at the site of DNA immunization. Our results indicate that primary immune responses are induced by cells that migrate from the epidermis (presumably dendritic cells) within 24 h of immunization, and that Ag-expressing nonmigratory cells (presumably keratinocytes) influence the magnitude of this response. These conclusions are supported by our finding that: 1) immediate removal of the vaccination site abrogated immunity, 2) primary B and T cell responses were induced when the vaccination site was left intact for 5 to 24 h, 3) transfer of the vaccination site within 5 h of immunization induced a primary T cell response in naive recipients, and 4) the magnitude of this primary response increased the longer the vaccination site was left in place (for up to 2 wk). It should be noted that transfected keratinocytes continue to produce Ag for 2 wk (although maximal protein production occurs within the first 3 days), while transfected dendritic cells emigrate from the epidermis within 24 h (14).
We also found that rapidly migrating cells were solely responsible for the development of immunologic memory. Leaving the site of vaccination intact for 5 to 12 h was sufficient for the host to generate memory B cells that rapidly produced high titrated IgG anti-CSP Abs when re-exposed to Ag in vivo, and memory T cells that secreted IFN-γ when restimulated by Ag in vitro. Similarly, memory T and B cells were generated when skin vaccinated within 12 h was grafted onto naive recipients, unlike grafts taken >1 day postvaccination. These data support the conclusion that T and B cell memory is triggered by cells that migrate from the epidermis 5 to 12 h postimmunization. These experiments demonstrate that gene gun vaccination combined with skin excision/grafting provides a novel method for analyzing the effect of Ag dose and duration on the induction of primary and memory responses.
This work was supported in part by a grant from the National Vaccine Program.
Abbreviations used in this paper: CSP, circumsporozoite protein; GMT, geometric mean titer; pCSP, plasmid-encoding CSP.