Transcriptional targeting of dendritic cells in gene gun-mediated DNA immunization favors the induction of type 1 immune responses

ARTICLE

doi:10.1016/S1525-0016(03)00242-9

Transcriptional Targeting of Dendritic Cells in Gene Gun-Mediated DNA Immunization Favors the Induction of Type 1 Immune Responses Stephan Sudowe,* Isis Ludwig-Portugall, Evelyn Montermann, Ralf Ross, and Angelika B. Reske-Kunz Clinical Research Unit Allergology, Department of Dermatology, Johannes Gutenberg-University, Obere Zahlbacher Strasse 63, D-55101 Mainz, Germany *To whom correspondence and reprint requests should be addressed. Fax: ⫹49 6131 3933360. E-mail: [email protected].

Cutaneous dendritic cells (DC) are pivotal for the elicitation of antigen-specific immune responses following gene gun-mediated biolistic transfection of the skin. We transcriptionally targeted transgene expression to DC using vectors containing the murine fascin promoter (pFascin) to control antigen production and compared the immune response elicited with conventional DNA immunization using plasmid constructs with the ubiquitously active CMV promoter (pCMV). Biolistic transfection with pFascin initiated a marked type 1 immune response characterized by the occurrence of a large population of IFN-␥-producing T helper (Th) cells in spleen and draining lymph nodes. Consistently, immunoglobulin production was dominated by IgG2a antibodies. In contrast, the humoral response after repeated administration of pCMV was strongly enhanced and characterized by a type 2-like isotype pattern (IgG1 > IgG2a). Cytokine production analysis in vitro indicated compartmentalization of the immune response, revealing large numbers of IL-4-producing Th cells in the lymph nodes and dominant presence of IFN-␥-producing Th cells in the spleen. Biolistic transfection with pFascin, like immunization with pCMV, led to potent induction of cytotoxic T cells as was assessed by JAM test. Thus gene gun immunization with plasmids that focus transgene expression and antigen production specifically to DC propagates type 1-biased cellular immune responses. Key Words: gene therapy, DNA vaccines, biolistics, gene gun technique, dendritic cells, fascin, cytotoxic T lymphocytes, Th1 cells, Th2 cells, mouse, BALB/c

INTRODUCTION Genetic immunization, i.e., in vivo transfection of somatic cells with antigen-encoding DNA, effectively induces MHC class I-restricted cell-mediated immunity in the form of CD8⫹ cytotoxic T lymphocytes (CTL) and elicits humoral immune reactions dependent on MHC class IIrestricted activation of T helper (Th) cells [reviewed in 1]. DNA vaccination has, therefore, been used for protective and therapeutic purposes in numerous preclinical animal models of viral, bacterial, and parasitic infections, but also as a novel form of immunotherapy against cancer and allergic diseases [reviewed in 2]. Naked plasmid DNA as the most frequently used nonviral gene vector was administered by various modes and routes, the most popular ones being intramuscular and intradermal needle injection of DNA in saline and bombardment of skin with DNA-coated gold particles using the helium-powered gene gun. Biolistic transfection allows for the direct intro-

MOLECULAR THERAPY Vol. 8, No. 4, October 2003 Copyright © The American Society of Gene Therapy 1525-0016/03 $30.00

duction of DNA into epidermal cells and requires 10 to 100 times lower amounts of DNA to evoke antigen-specific immune responses than intramuscular or intradermal immunization [3,4], in which plasmids have to be actively taken up by muscle or skin cells. All these routes of DNA inoculation induce significant numbers of CTL. However, following intramuscular and intradermal immunization the humoral response is characterized by the predominant production of specific IgG2a antibodies (Ab), whereas gene gun immunization yields a preponderance of IgG1 Ab, suggesting Th1- and Th2-biased responses, respectively [5]. Dendritic cells (DC) as professional antigen-presenting cells (APC) play a central role in the initiation of transgene-specific immune responses by all methods of DNA delivery due to their unique migratory behavior and primary stimulatory capacity [reviewed in 6]. DC transfected in vitro with antigen-encoding DNA induced a potent CTL response and variable antibody production as well as the

567

ARTICLE

development of specific CD4⫹ T cells with a Th1-polarized phenotype following transfer into recipient mice [7–9]. Given the critical role of DC in the elicitation of immune responses to plasmid-encoded antigens, the skin represents a favorable site for DNA immunization because of its tight network of Langerhans cells (LC) in the epidermis and the presence of dermal DC in considerable numbers. To control transgene expression, conventional plasmid vectors contain the ubiquitously active cytomegalovirus immediate early promoter (pCMV), which allows for antigen production by every successfully transfected cell. To restrict reporter gene expression to DC, the essential cell type for inducing specific immune responses, we used the promoter of the murine fascin gene. The actin-bundling protein fascin is expressed by DC [10,11] and represents a crucial structural component, required for the formation of dendrites [12,13]. Fascin is not produced in immature DC as represented by epidermal LC, but during maturation into primary stimulatory DC fascin gene expression is strongly upregulated [12,13]. Apart from maturing and mature DC, expression of fascin is restricted to only a few nonhematopoietic cell types such as neuronal/glial cells and capillary endothelial cells and to some transformed cells [12,14,15]. Recently we demonstrated that transfection in vitro and in vivo with reporter gene constructs using the fascin promoter (pFascin) allowed for transcriptional targeting of transgene expression specifically to mature DC and thus ensured DC-focused production of antigen [16]. The immune response following particle bombardment with pFascin in vivo was characterized by the induction of large numbers of antigen-specific IFN-␥producing CD8⫹ effector T cells, but weak production of specific IgG antibodies [16]. Here we compare the immune response initiated by gene gun and intradermal immunization, respectively, with plasmid vectors encoding the model antigen ␤-galactosidase (␤Gal) under control of the fascin promoter (pFascin-␤Gal) or the CMV promoter (pCMV-␤Gal). We reveal differences following biolistic transfection in the isotype pattern of the secreted IgG Ab as well as in the cytokine release of lymph node (LN) cells stimulated in vitro, suggesting that gene gun immunization with pFascin-␤Gal, contrary to pCMV␤Gal, leads to marked type 1 immune responses.

RESULTS Biolistic Transfection Transcriptionally Targeting DC Induces a Type 1-Biased Antibody Isotype Pattern To analyze the humoral response after biolistic transfection with ␤Gal-expressing plasmids, we immunized BALB/c mice by a single or several consecutive administrations of pFascin-␤Gal or pCMV-␤Gal via the gene gun. After 42 days we detected low ␤Gal-specific IgG titers in sera of the mice as the result of a single immunization with both expression constructs, yet they were significantly higher following transfection with pCMV-␤Gal

568

doi:10.1016/S1525-0016(03)00242-9

(Fig. 1A, top). Repeated immunization with pCMV-␤Gal at weekly intervals considerably boosted the production of ␤Gal-specific IgG, whereas the increase in IgG Ab titers after three or five applications of pFascin-␤Gal was weak. Consequently, the amount of IgG Ab in sera of pCMV-␤Gal-immunized mice extensively exceeded the IgG Ab content in sera of animals transfected with pFascin-␤Gal (P ⬍ 0.001). With regard to the IgG isotype profile arising following gene gun immunization with pFascin-␤Gal or pCMV␤Gal we obtained divergent results (Fig. 1A, bottom). Biolistic transfection with pCMV-␤Gal elicited high titers of ␤Gal-specific IgG1 as well as IgG2a Ab, with IgG1 being the predominant isotype and thus representing a more Th2-biased isotype pattern. Particle bombardment with pFascin-␤Gal resulted in the secretion of specific IgG2a Ab after repeated administration. However, in contrast to the situation following biolistic transfection with pCMV␤Gal, production of specific IgG1 Ab was very low. The relative levels of IgG1 and IgG2a Ab documented preponderance of IgG2a Ab production, indicating a Th1-directed immune response. The IgG isotype pattern did not depend on the amount of plasmid DNA inoculated with the gene gun, since immunization with DNA doses over a broad range (0.5 to 12 ␮g) did not alter the quality of the humoral response following biolistic transfection with pFascin-␤Gal and pCMV-␤Gal, respectively (data not shown). In addition, the Ab isotype profile did not vary with the number of immunizations (Fig. 1A, bottom) or time after immunization (data not shown). To validate these results we performed gene gun immunizations with plasmids expressing enhanced green fluorescent protein (EGFP) as transgene under the control of the fascin (pFascin-EGFP) or the CMV (pCMV-EGFP) promoter. Similar to the experiments with ␤Gal-encoding vectors, production of EGFP-specific Ab was strongly increased after repeated transfection with pCMV-EGFP in contrast to gene gun immunization with pFascin-EGFP (Fig. 1B). Likewise, specific IgG Ab production after transfection with pCMV-EGFP was dominated by IgG1 Ab, whereas immunization with pFascin-EGFP induced more IgG2a Ab than IgG1 Ab (Fig. 1B). We further analyzed immunoglobulin production after genetic immunization with the two promoter constructs via an alternative cutaneous route, namely intradermal injection into the ear pinna. The majority of ␤Gal-specific IgG Ab in immune sera recovered after three intradermal applications of plasmids was of IgG2a isotype irrespective of whether pFascin-␤Gal or pCMV-␤Gal was used, thus pointing toward a Th1-type immune response (Fig. 2). Biolistic Transfection Transcriptionally Targeting DC Initiates Th1 Responses To investigate whether the outcome of contrasting IgG isotype profiles after gene gun immunization with pFascin-␤Gal or pCMV-␤Gal was dependent on activation of oppositely oriented Th cell populations, we tested for

MOLECULAR THERAPY Vol. 8, No. 4, October 2003 Copyright © The American Society of Gene Therapy

doi:10.1016/S1525-0016(03)00242-9

ARTICLE

FIG. 1. Quantitative and qualitative differences in the production of specific Ab following biolistic transfection with pFascin or pCMV. (A) BALB/c mice (n ⫽ 6) were immunized by a single (1⫻), three (3⫻), or five (5⫻) weekly administrations of pFascin-␤Gal or pCMV-␤Gal via the gene gun. 42 days after the first immunization sera were recovered and levels of ␤Gal-specific IgG, IgG1, and IgG2a were determined by ELISA. Data are representative of at least three independent experiments. (B) BALB/c mice (n ⫽ 3) were immunized by a single (1⫻) or three weekly (3⫻) administrations of pFascin-EGFP or pCMV-EGFP via the gene gun. After 35 days sera were recovered and EGFPspecific IgG1 and IgG2a titers were determined by ELISA. All data are presented as means ⫾ SD of individual sera. The ratio of IgG1 versus IgG2a for each experimental group is depicted above the bars. Asterisks indicate significant differences (*P ⬍ 0.01, **P ⬍ 0.001) between mice immunized with pFascin or pCMV.

cytokine production in vitro by splenocytes and LN cells from immunized mice. The cells were restimulated with ␤Gal protein, which led to MHC class II-restricted antigen presentation and thus induced stimulation of and cytokine release by CD4⫹ Th cells. After 3 days of culture we determined the frequencies of IL-4 and IFN-␥-secreting cells with an EliSpot assay. In cultures of spleen cells prepared from animals primed by a single or several immunizations with either

FIG. 2. Production of specific Ab following DNA immunization with pFascin␤Gal or pCMV-␤Gal via different cutaneous routes. BALB/c mice (n ⫽ 4) were immunized with pFascin-␤Gal or pCMV-␤Gal by three administrations via the gene gun (g.g.) or by three intradermal injections (i.d.) at weekly intervals. 35 days after the first immunization sera were recovered and ␤Gal-specific IgG1 and IgG2a titers were determined by ELISA. Data are presented as means ⫾ SD of individual sera. The ratio of IgG1 versus IgG2a for each experimental group is depicted above the bars. Asterisks indicate significant differences (*P ⬍ 0.01, **P ⬍ 0.001) between mice immunized with pFascin-␤Gal or pCMV-␤Gal. Data are representative of two independent experiments.

MOLECULAR THERAPY Vol. 8, No. 4, October 2003 Copyright © The American Society of Gene Therapy

pFascin-␤Gal or pCMV-␤Gal, the frequency of Th cells producing IFN-␥ considerably exceeded the frequency of IL-4-secreting Th cells (Fig. 3A). This was unexpected in the case of immunization with pCMV-␤Gal in light of our finding of a type 2-biased IgG isotype profile. However, the number of IL-4-producing Th cells among spleen cells following five consecutive administrations of pCMV-␤Gal was significantly enhanced compared with immunization with pFascin-␤Gal (P ⬍ 0.05). In vitro analysis of cytokine release by lymphocytes isolated from the draining (inguinal and axillary) LN revealed major differences primarily in the generation of IL-4-producing Th cells. A single bombardment with pFascin-␤Gal or pCMV-␤Gal elicited only low numbers of cytokine-producing cells. However, repeated gene gun immunization with pFascin-␤Gal strongly enhanced the priming of IFN-␥-secreting Th cells in particular, resulting in a predominance of IFN-␥-producing Th cells over IL-4producing Th cells (Fig. 3B). The ratio of the frequency of IFN-␥- versus IL-4-producing Th cells after three (5.2) and five (4.1) immunizations indicated a clear-cut Th1-biased response. In contrast, repeated gene gun immunization with pCMV-␤Gal induced IL-4-producing Th cells in high numbers in addition to IFN-␥-secreting cells (Fig. 3B), suggesting a more Th2-biased immune response in LN of these animals. These findings were supported by data obtained by measurement of additional Th2-associated cytokines accumulated in the culture supernatants of antigen-stimulated splenocytes and LN cells using sandwich ELISA. A polarization into different Th cell subsets again

569

ARTICLE

doi:10.1016/S1525-0016(03)00242-9

measured the proliferative capacity. The proliferative response of splenic CD4⫹ T cells isolated from mice immunized by a single and three (data not shown) as well as five administrations (Fig. 4) of pFascin-␤Gal or pCMV-␤Gal was equivalent between the two immunization groups.

FIG. 3. Development of divergent cytokine profiles in draining LN cells following biolistic transfection with pFascin-␤Gal or pCMV-␤Gal. BALB/c mice (n ⫽ 4) were immunized by a single (1⫻), three (3⫻), or five (5⫻) weekly administrations of pFascin-␤Gal or pCMV-␤Gal via the gene gun. 56 days after the first immunization (A) spleens and (B) inguinal/axillary LN were removed and separately pooled and single-cell suspensions were prepared. Cells (5 ⫻ 106/ ml/well) were cultured with (closed/open bars) or without (hatched bars) 25 ␮g/ml recombinant ␤Gal. After 72 h frequencies of IL-4- (open bars) and IFN-␥- (closed bars) producing cells in the cultures were determined by EliSpot assay. Data are presented as means of triplicates ⫾ SD. Asterisks indicate significant differences (*P ⬍ 0.05, **P ⬍ 0.01, ***P ⬍ 0.001) between mice immunized with pCMV-␤Gal or pFascin-␤Gal. Data are representative of at least two independent experiments.

was obvious in cultures of LN cells. After 3 days of culture only lymphocytes from mice immunized by five consecutive administrations of pCMV-␤Gal had released high amounts of IL-5, IL-10, and IL-13 into the medium (Table 1). This phenomenon was not due to a generally decreased activation of CD4⫹ T cells in cultures from mice immunized with pFascin-␤Gal, because the production of the Th1 cytokine IFN-␥ was slightly higher in these cultures (Table 1). Thus, the opposite cytokine profiles produced by LN cells reflected the type 1 and type 2 antibody isotype patterns arising following biolistic transfection with pFascin-␤Gal and pCMV-␤Gal, respectively.

Biolistic Transfection Transcriptionally Targeting DC Effectively Primes CD8ⴙ CTL We also addressed the question whether the mode of transgene expression (ubiquitously versus transcriptionally targeted to DC) and the route of cutaneous gene delivery had a major impact on MHC class I-restricted induction of CD8⫹ effector T cells, representing the CTL population. For that reason we analyzed spleen cells ex vivo using a modified EliSpot assay, in which activated ␤Gal-specific CD8⫹ T cells were detected and enumerated by IFN-␥ production following incubation of splenocytes with the H-2Ld-restricted ␤Gal-derived nonamer peptide TPHPARIGL (␤Gal876 – 884). A single administration of pFascin-␤Gal or pCMV-␤Gal via the gene gun already primed for high numbers of CD8⫹ effector T cells, which were considerably increased by subsequent immunization (Fig. 5A). The frequencies of CD8⫹ effector T cells arising following biolistic transfection with pFascin-␤Gal or pCMV-␤Gal were in comparable ranges. Likewise, intradermal injection of pFascin-␤Gal or pCMV-␤Gal induced similar quantities of IFN-␥-producing CD8⫹ effector T cells (Fig. 5B). To evaluate whether the ␤Gal-specific CD8⫹ effector T cells induced by DNA immunization with the two ␤Galencoding vectors are functional CTL, we measured the cytolytic activity of spleen cells from immunized mice by JAM test. Since the cytolytic potential of unstimulated splenocytes was only marginal (data not shown), we preincubated spleen cells for a period of 6 days with the ␤Gal876 – 884 peptide to activate ␤Gal-specific CD8⫹ T cells and thereby increase the number of CTL. Stimulated

TABLE 1: Cytokine production in spleen and lymph node cell cultures Cytokine production (pg/ml) Spleen Cytokine

18 (9)

IL-10

333 (3)

IL-13 IFN-␥

Proliferative Responses of Splenocytes Following Biolistic Transfection with pFascin or pCMV Are Equivalent To support our finding that DNA immunization transcriptionally targeting DC is efficient in priming CD4⫹ T cells, we stimulated splenocytes in vitro with ␤Gal protein and

570

pFascin-␤Gal

IL-5

20 (⬍3) 4982 (137)

Lymph node

pCMV-␤Gal 81 (54)

pFascin-␤Gal

pCMV-␤Gal

36 (9)

1423 (45)

472 (7)

428 (⬍3)

1709 (⬍3)

41 (7)

41 (⬍3)

684 (⬍3)

10,709 (⬍8)

6836 (⬍8)

1899 (61)

BALB/c mice (n ⫽ 4) were immunized by five weekly administrations of pFascin-␤Gal or pCMV-␤Gal via the gene gun. 66 days after the first immunization spleens and inguinal/ axillary LN were removed and separately pooled. Cells (5 ⫻ 106/ml/well) were cultured in quadruplicates with or without 25 ␮g/ml recombinant ␤Gal. After 72 h culture supernatants were pooled and cytokines were quantified in duplicates using a sandwich ELISA. Numbers in parentheses represent cytokine production of cells cultured in medium alone. Data are representative of three independent experiments.

MOLECULAR THERAPY Vol. 8, No. 4, October 2003 Copyright © The American Society of Gene Therapy

doi:10.1016/S1525-0016(03)00242-9

FIG. 4. Proliferative response of splenic T cells following biolistic transfection with pFascin-␤Gal or pCMV-␤Gal. BALB/c mice (n ⫽ 4) were immunized by five weekly administrations of pFascin-␤Gal or pCMV-␤Gal via the gene gun. 56 days after the first immunization pooled spleen cells (5 ⫻ 105/well) were cultured on 96-well flat-bottomed microtiter plates alone or with the indicated concentrations of recombinant ␤Gal. Proliferation of splenocytes was measured after 72 h by incubation with [3H]thymidine (1 ␮C/well) for another 18 h. Data are presented as means of six wells ⫾ SD. Data are representative of two independent experiments.

spleen cells from mice biolistically transfected with pFascin-␤Gal or pCMV-␤Gal showed comparable, high cytolytic activity against ␤Gal-expressing P13.1 target cells, whereas parental P815 cells, which do not produce ␤Gal, were not killed (Fig. 5C, top). Similarly, intradermal injection of pFascin-␤Gal or pCMV-␤Gal induced equivalent high CTL responses (Fig. 5C, bottom).

DISCUSSION In this study, we analyzed in detail the in vivo immune response initiated by cutaneous DNA immunization transcriptionally targeting antigen production specifically to DC. The data show that gene gun immunization with plasmids expressing transgenes under the control of the promoter of the fascin gene propagated strong type 1 immune responses, characterized by (1) high frequencies of IFN-␥-producing Th cells in spleen and draining LN, (2) predominant production of IgG2a Ab, and (3) potent induction of IFN-␥-secreting CD8⫹ effector T cells with antigen-specific cytolytic activity. In contrast, biolistic transfection with conventional plasmid vectors containing the ubiquitously expressed CMV promoter resulted in humoral responses characterized by a type 2-like isotype profile (IgG1 ⬎ IgG2a). Analysis of cytokine production by spleen and inguinal/axillary LN cells isolated from mice immunized with pCMV-␤Gal revealed compartmentalization of the immune response, being Th1-like in the spleen, but more Th2-biased in the draining LN. DC-focused transgene expression and antigen production is being considered to be of great interest, since it was demonstrated by several studies that following cutaneous DNA immunization directly transfected skin-derived DC are crucial for the initiation of immune responses in the draining LN [17–20]. By histochemical staining we de-

MOLECULAR THERAPY Vol. 8, No. 4, October 2003 Copyright © The American Society of Gene Therapy

ARTICLE

tected very few ␤Gal-positive cells in the regional inguinal and axillary LN following gene gun immunization with pFascin-␤Gal as well as with pCMV-␤Gal (data not shown). In a previous report we identified cells in LN expressing the transgene EGFP following biolistic transfection as LC by virtue of their expression of the LC marker molecule langerin/CD207 [16]. Our results obtained by biolistic transfection with DC-specific plasmids are consistent with data showing that transfection of a small number of DC is sufficient to initiate a wide variety of immune responses [9]. The immune reaction following gene gun immunization with pFascin revealed fundamental qualitative differences in comparison with responses elicited by gene gun immunization using pCMV vectors. The humoral response after particle bombardment of skin of BALB/c mice with pCMV-␤Gal was characterized by the production of IgG1 as well as IgG2a Ab, with a predominance of IgG1 Ab, however, whereas intradermal injection of pCMV␤Gal induced a preponderance of Ab of the IgG2a subclass. These findings are in line with reports from other groups [21–23]. Since it is generally accepted that the isotype profile serves as an indicator for the type of T cell help supporting a murine immune reaction, the results are concurrent with prior studies with various antigens, demonstrating that intradermal injection of plasmid DNA primes for a more Th1-biased response, whereas gene gun immunization is associated with the induction of Th2 responses [5,24,25]. In contrast, for two different antigens (␤Gal, EGFP) we could show that Ig production as the result of gene gun immunization with pFascin was dominated by Ab of the IgG2a subclass and thus was similar to the isotype pattern induced by intradermal injection of pFascin-␤Gal, suggesting that a Th1-oriented immune response was induced by both routes of application. However, Ab production after biolistic transfection with pFascin-␤Gal was low. This phenomenon, which has been recognized previously by us and others following selective genetic transduction of DC in vitro [26] and in vivo [16, 27], is probably due to reduced antigen production and thus limited availability of B cell epitopes following DCfocused DNA immunization. In vitro analysis of cytokine patterns of draining LN cells from immunized mice confirmed the assumption of oppositely oriented Th cell populations in animals gene gun-immunized with pCMV-␤Gal or pFascin-␤Gal. Biolistic transfection with pCMV-␤Gal induced IFN-␥-secreting Th cells as well as a large population of Th cells producing IL-4, a cytokine that is preferentially released by Th2 cells and that is the main supporter of IgG1 production [28]. In addition, these cells secreted considerable amounts of other Th2-associated cytokines like IL-5, IL-13, and IL-10. The observation that the induction of IL-4-producing Th cells was most pronounced following multiple consecutive genetic immunizations is consistent with the study of Pertmer et al. [29]. In sharp contrast, repeated gene gun

571

ARTICLE

doi:10.1016/S1525-0016(03)00242-9

FIG. 5. Induction of CTL responses following DNA immunization with pFascin-␤Gal or pCMV-␤Gal via different cutaneous routes. (A) BALB/c mice (n ⫽ 4) were immunized by a single (1⫻), three (3⫻), or five (5⫻) weekly administrations of pFascin-␤Gal or pCMV-␤Gal via the gene gun. 56 days after the first immunization frequencies of IFN-␥-producing CD8⫹ effector T cells among splenocytes were determined following incubation for 22 h without or with the H-2Ldspecific peptide ␤Gal876 – 884 by EliSpot assay. (B, C) BALB/c mice (n ⫽ 4) were immunized with pFascin-␤Gal or pCMV␤Gal by three administrations via the gene gun (g.g.) or by three intradermal injections (i.d.) at weekly intervals. 49 days after the first immunization frequencies of IFN-␥-producing CD8⫹ effector T cells among splenocytes were determined as described (B). All data are presented as means ⫾ SD of individual mice. Differences between mice immunized with pCMV-␤Gal or pFascin-␤Gal, did not reach statistical significance (P ⬎ 0.05). (C) Cytolytic activity of spleen cells from BALB/c mice immunized by gene gun (top) or intradermal injection (bottom) with pFascin-␤Gal or pCMV-␤Gal against target cells expressing ␤Gal (P13.1) or not (P815) was determined by JAM test. Data are presented as means of triplicates ⫾ SD.

immunization with pFascin-␤Gal did not induce substantial numbers of IL-4-producing Th cells among LN cells but primed primarily Th cells producing IFN-␥, a cytokine that is a surrogate marker for Th1 responses and that promotes the formation of IgG2a Ab [28]. Furthermore, the production of IL-5, IL-10, and IL-13 after in vitro restimulation was low. Strikingly, splenocytes from mice immunized with pFascin-␤Gal as well as with pCMV-␤Gal showed a Th1-biased cytokine profile with a majority of Th cells secreting IFN-␥. This observation corresponds to several other reports that document high IFN-␥ production by splenic cells from mice immunized via the gene gun [9,21,29,30], revealing a lack of correlation between IgG isotype ratios and cytokine production. Our data indicate the coexistence of Th1 and Th2 responses in different lymphoid compartments in mice immunized with pCMV-␤Gal as was previously reported by Lewis et al. [31]. However, since the initial priming of antigen-specific B cells following DNA immunization takes place in the draining LN rather than the spleen [31–33], the opposite Th1- and Th2-type isotype patterns arising after biolistic transfection with pFascin-␤Gal and pCMV-␤Gal, respectively, probably can be attributed to the divergent cytokine profiles exhibited by LN cells. The reasons why gene gun immunization with pFascin-␤Gal elicits a Th1-skewed immune response are not clear. Like in our DC-targeted gene gun immunization approach intradermal injection of plasmid DNA induces Th1 responses as well [5,23–25]. Since following injection the DNA is located extracellularly, successful transfection of DC requires the internalization of the plasmids. Activation of DC is facilitated by the specific binding of immunostimulatory CpG motifs within the backbone of the DNA vector to the toll-like receptor 9 within endocytic

572

vesicles, initiating a transduction signal cascade that promotes the creation of a Th1-biased environment through the secretion of proinflammatory cytokines, in particular IL-12, by the DC [reviewed in 34]. This pathway is probably bypassed for the most part following gene gun immunization because plasmid-coated gold particles penetrate directly into the cytoplasm or ideally into the nucleus of epidermal cells. Weiss and colleagues [35] claimed that biolistic transfection per se represents a Th2inducing “danger signal,” which is even able to overcome partially the Th1-promoting effects of intradermal injection. Our data show that this postulate cannot be held in general. We argue that the amount of protein released by or leaking from transfected cells and thus the antigen dose accessible for APC to prime T cells seems to have a major impact on the outcome of the immune response following DNA immunization. As was mentioned before, IgG Ab production and thus the activation of antigenspecific B cells was strongly enhanced following gene gun immunization with pCMV. Since abundant availability of B cells, acting as APC to sustain the T cell response, favors the development of Th2 cells [36], this might contribute to the generation of type 2-biased immune responses induced by biolistic transfection with pCMV. This notion might be substantiated by gene gun immunizations with plasmids expressing secretory or cell-associated forms of an antigen under the control of the fascin promoter, with the intention to investigate whether increasing the quantity of free protein results in a shift toward a more Th2polarized isotype profile or cytokine pattern. The phenomenon that the nature of the antigen (secretory vs membrane-bound/cytosolic) and with this the antigen dose dictates the type of immune response has been fre-

MOLECULAR THERAPY Vol. 8, No. 4, October 2003 Copyright © The American Society of Gene Therapy

ARTICLE

doi:10.1016/S1525-0016(03)00242-9

quently documented following intramuscular injection of DNA [32,37– 41]. We had shown before that pFascin as well as pCMV elicited a potent CTL response, as was evaluated by the presence of IFN-␥-producing CD8⫹ effector T cells among spleen cells of biolistically transfected mice [16]. In this report we demonstrate that these cells actually exhibit strong antigen-specific cytolytic potential. In addition we extended our studies to analysis of CTL induction by intradermal injection. Our data show that following intradermal immunization frequencies of CTL are equivalent following injection of pFascin-␤Gal and pCMV-␤Gal. Our findings support the notion that endogenous antigen production by directly transfected DC is sufficient to induce a CTL response [17,18,20] and that cross-presentation of exogenous antigens, either secreted by transfected cells or associated with apoptotic cells, represents a minor pathway of CD8⫹ T cell priming in this context. DC-targeted DNA vaccine approaches have become an attractive field of research activity [6]. The selective antigen production in DC by means of transcriptional targeting of gene expression to DC represents a promising strategy to increase the efficiency of genetic immunization and to tailor the type of the resulting immune response. Up to now, only few data have been published dealing with this concept of DC-focused DNA immunization, most likely because of the limited knowledge of DC-specific molecules and the respective genes, including their regulatory elements. The CD11c promoter was established by Brocker and co-workers to drive DC-specific transgene expression in a number of transgenic mouse lines [42,43]; however, it was not used in gene immunization studies. Bonkobara et al. [44] reported that transcriptional activity of the promoter of the gene for the C-type lectin dectin-2 is selective for LC in vitro and in vivo. Biolistic transfection of mice with vectors containing the dectin-2 promoter induced T cell proliferation and IFN-␥ production, but neither cytotoxic T cell responses nor Ab production were detected [27]. Notably, the dectin-2 promoter was more active in vitro in cell lines representing immature DC (XS52 [45]) than in cell lines with the phenotype of mature DC (XS106 [9]). This finding leaves the possibility open that presentation of endogenous antigen by immature DC leads to suboptimal stimulation of T cells or even to the suppression of T cell activation [46]. In contrast, the fascin promoter used in our experiments is induced during maturation of DC and highly active in mature DC, which allows optimal priming of T cell responses. In conclusion, the fascin promoter proved to be a suitable tool for DC-focused DNA immunization and thus provides new opportunities for immunotherapy and for improved vaccination strategies against cancer as well as infectious and allergic diseases.

MOLECULAR THERAPY Vol. 8, No. 4, October 2003 Copyright © The American Society of Gene Therapy

MATERIAL

AND

METHODS

Vector construction and plasmid preparation. The construction of plasmid vectors encoding the transgenes ␤Gal or EGFP under control of the CMV promoter (pCMV-␤Gal, pCMV-EGFP) or the fascin promoter (pFascin-␤Gal, pFascin-EGFP) has been described in detail previously [16]. Briefly, the coding sequence of either ␤Gal (derived from plasmid pCMV␤, BD Biosciences Clontech, Heidelberg, Germany) or EGFP (derived from plasmid pEGFP-N1, BD Biosciences Clontech) was inserted into vector pCI (Promega, Mannheim, Germany) containing the CMV promoter. Subsequently, a 2.6-kb promoter fragment of the fascin gene was subcloned into these plasmids, thereby replacing the CMV promoter of pCI. The plasmids were propagated in Escherichia coli strain TOP10 and purified from overnight cultures using the GenElute endotoxin-free plasmid purification kit (Sigma–Aldrich, Deisenhofen, Germany) according to the manufacturer’s instructions. The DNA was precipitated in ethanol, resuspended in pyrogen-free water, quantified using a spectrophotometer, and stored at ⫺20°C. Prior to immunization the DNA was diluted in sterile PBS (1 mg/ml) for intradermal injection or, alternatively, in pyrogen-free water to suitable concentrations for preparation of gene gun cartridges. Endotoxin levels in plasmid preparations were ⬍20 EU/mg DNA as determined by Limulus amebocyte lysate assay (BioWhittaker, Walkersville, MD). Mice and immunization protocol. BALB/c mice were bred under specificpathogen-free conditions in the animal facilities of the University of Mainz according to the local guidelines for animal care. Female mice at 6 –10 weeks of age at the beginning of the experiments were used. Biolistic transfection was performed using the helium-driven Helios gene gun system (Bio-Rad, Munich, Germany). Coating of gold particles (1.6 ␮m) with plasmid DNA and preparation of cartridges were carried out according to the manufacturer’s instructions. Each cartridge contained 2 ␮g of DNA precipitated to 1 mg of gold. Cartridges were stored at 4°C for up to 2 weeks. Mice were immunized via the gene gun at a discharge pressure of 400 psi by two nonoverlapping shots on shaved abdominal skin resulting in a total immunization dose of 4 ␮g of DNA. For intradermal immunization mice were anesthetized by ip application of 250 ␮l of 2.5% avertin (1 mg/ml 2,2,2-tribromethanol in 2-methyl2-butanol; Sigma–Aldrich) in PBS. Animals were injected intrapinna with 100 ␮g of plasmid DNA in a volume of 30 ␮l using a 1-ml syringe and a 25-gauge needle. Determination of serum Ab titers. Mice were bled by puncture of the retro-orbital plexus. Sera were frozen at ⫺20°C until thawed for determination of antigen-specific IgG, IgG1, and IgG2a by antigen capture ELISA as reported [47] with the modification that recombinant EGFP was used as antigen as well. The Ab titer was defined as the reciprocal serum dilution, yielding an absorbance reading of OD-0.2 after linear regression analysis. To compare IgG1 and IgG2a Ab titers directly isotype-specific ELISA were performed in parallel. IgG content in serum samples was normalized using a high-titer standard reference serum, obtained from mice repeatedly immunized with ␤Gal protein in aluminum hydroxide [47]. EliSpot assay for enumeration of IFN-␥-producing CD8ⴙ effector T cells. The frequency of CD8⫹ effector T cells among splenocytes recognizing the H-2Ld-binding ␤Gal-derived nonamer peptide TPHPARIGL (␤Gal876 – 884) was determined by an EliSpot assay as described [47]. Spots were counted and evaluated using the EliSpot Reader System (AID, Strassberg, Germany). Cytotoxicity assay. The cytotoxic activity of spleen cells from immunized mice against LacZ-transfected P13.1 [48] and untransfected P815 target cells, respectively, was determined using the JAM test [49]. Splenocytes (5 ⫻ 106/well) were stimulated in 24-well tissue culture plates (Corning Costar, Bodenheim, Germany) in a volume of 1 ml culture medium [IMDM supplemented with 10% fetal calf serum (PAN Systems GmbH, Nu ¨ rnberg, Germany), 50 ␮m 2-mercaptoethanol (Carl Roth GmbH & Co.), 2 mM L-glutamine, and 100 U penicillin/streptomycin (Gibco, Paisley, UK)] containing 1 ␮g/ml ␤Gal876 – 884 peptide and 20 U/ml recombinant IL-2 (Cetus, Emeryville, CA). After 6 days effector cells were harvested, pooled, and recultured in graded numbers on 96-well V-bottom culture plates (Greiner, Frickenhausen, Germany) at various effector-to-target ra-

573

ARTICLE

tios with 1 ⫻ 104 target cells, previously labeled with [3H]thymidine (5 ␮Ci/ml for 6 h), in a total volume of 150 ␮l culture medium containing 10 U/ml recombinant IL-2. After 18 h of incubation at 37°C cells were harvested onto glass fiber filters. Recovery of radioactivity (in counts per minute) was assessed by liquid scintillation counting. The cytolytic activity was calculated as: % killing ⫽ (S ⫺ E)/S ⫻ 100, where S is retained target cell DNA in the absence of effector cells (spontaneous) and E is retained target cell DNA in the presence of effector cells (experimental). EliSpot assay for enumeration of cytokine-producing cells. Splenocytes or inguinal and axillary LN cells (5 ⫻ 106/well) were incubated for 72 h on 24-well tissue culture plates in a volume of 1 ml culture medium with or without recombinant ␤Gal (25 ␮g/ml). The frequency of IFN-␥- and IL-4producing cells in the cultures was determined by EliSpot assay as described [47]. Spots were counted and evaluated using the EliSpot Reader System. Determination of cytokines in culture supernatants. Cytokines were quantified in culture supernatants of spleen and LN cells, stimulated with recombinant ␤Gal for 3 days (see above), using a sandwich ELISA performed after a standard procedure as previously described [50]. Briefly, aliquots of supernatant were incubated overnight at 4°C on MaxiSorp ELISA plates previously coated with capture mAb as recommended by the manufacturer: anti-IL-5 clone TRFK5 (PharMingen), anti-IL-10 clone JES052A5 (R&D Systems, Wiesbaden, Germany), anti-IL-13 clone 38312.11 (R&D Systems), and anti-IFN-␥ clone R4-6A2 (PharMingen). Subsequently, bound cytokines were detected by addition of biotinylated detection Ab: anti-IL-5 clone TRFK4 (PharMingen), anti-IL-10 (BAF417; R&D Systems), anti-IL-13 (BAF 413; R&D Systems), and anti-IFN-␥ clone XMG1.2 (PharMingen). The ELISA was developed by successive addition of ExtrAvidin–peroxidase conjugate (Sigma–Aldrich) and o-phenylenediamine/H2O2 (Sigma–Aldrich) as substrate. The reaction was stopped by addition of 1 M H2SO4. The absorption was measured in a microplate reader, Emax (MWG-Biotech, Ebersberg, Germany), at 490 nm. Recombinant mouse cytokines (IL-5, IFN-␥ purchased from PharMingen; IL-10, IL-13 purchased from R&D Systems) were used as standards. T cell proliferation assay. Splenocytes isolated from immunized mice (5 ⫻ 105/well) were cultured on 96-well flat-bottomed microtiter plates (Corning Costar) in culture medium alone or with various concentrations of recombinant ␤Gal. After 3 days cells were incubated with [3H]thymidine (1 ␮C/well) for another 18 h. Levels of incorporated [3H]thymidine were determined by liquid scintillation counting. Statistical analysis of data. Statistical evaluation of the experimental data was performed using Student’s t test and SigmaStat software. A value of P ⬍ 0.05 was considered statistically significant.

ACKNOWLEDGMENTS This work was supported by the Deutsche Forschungsgemeinschaft, SFB 548. We thank Hans-Georg Rammensee (University of Tu¨bingen, Germany) for his kind gift of P13.1 cells. This study was done in partial fulfillment of the requirements of the doctoral thesis of I.L.-P. RECEIVED FOR PUBLICATION MAY 17, 2003; ACCEPTED JULY 8, 2003.

REFERENCES 1. Donnelly, J. J., Ulmer, J. B., Shiver, J. W., and Liu, M. A. (1997). DNA vaccines. Annu. Rev. Immunol. 15: 617– 648. 2. Reyes-Sandoval, A., and Ertl, H. C. (2001). DNA vaccines. Curr. Mol. Med. 1: 217–243. 3. Fynan, E. F., Webster, R. G., Fuller, D. H., Haynes, J. R., Santoro, J. C., and Robinson, H. L. (1993). DNA vaccines: protective immunizations by parenteral, mucosal, and gene-gun inoculations. Proc. Natl. Acad. Sci. USA 90: 11478 –11482. 4. Pertmer, T. M., Eisenbraun, M. D., McCabe, D., Prayaga, S. K., Fuller, D. H., and Haynes, J. R. (1995). Gene gun-based nucleic acid immunization: elicitation of humoral and cytotoxic T lymphocyte responses following epidermal delivery of nanogram quantities of DNA. Vaccine 13: 1427–1430. 5. Feltquate, D. M., Heaney, S., Webster, R. G., and Robinson, H. L. (1997). Different T helper cell types and antibody isotypes generated by saline and gene gun DNA immunization. J. Immunol. 158: 2278 –2284.

574

doi:10.1016/S1525-0016(03)00242-9

6. Takashima, A., and Morita, A. (1999). Dendritic cells in genetic immunization. J. Leukocyte Biol. 66: 350 –356. 7. Manickan, E., Kanangat, S., Rouse, R. J., Yu, Z., and Rouse, B. T. (1997). Enhancement of immune response to naked DNA vaccine by immunization with transfected dendritic cells. J. Leukocyte Biol. 61: 125–132. 8. Tu¨ ting, T., DeLeo, A. B., Lotze, M. T., and Storkus, W. J. (1997). Genetically modified bone marrow-derived dendritic cells expressing tumor-associated viral or “self” antigens induce antitumor immunity in vivo. Eur. J. Immunol. 27: 2702–2707. 9. Timares, L., Takashima, A., and Johnston, S. A. (1998). Quantitative analysis of the immunopotency of genetically transfected dendritic cells. Proc. Natl. Acad. Sci. USA 95: 13147–13152. 10. Mosialos, G., et al. (1996). Circulating human dendritic cells differentially express high levels of a 55-kd actin-bundling protein. Am. J. Pathol. 148: 593– 600. 11. Pinkus, G. S., et al. (1997). Fascin, a sensitive new marker for Reed–Sternberg cells of Hodgkin’s disease. Evidence for a dendritic or B cell derivation? Am. J. Pathol. 150: 543–562. 12. Ross, R., Ross, X. L., Schwing, J., La¨ ngin, T., and Reske-Kunz, A. B. (1998). The actin-bundling protein fascin is involved in the formation of dendritic processes in maturing epidermal Langerhans cells. J. Immunol. 160: 3776 –3782. 13. Ross, R., et al. (2000). Expression of the actin-bundling protein fascin in cultured human dendritic cells correlates with dendritic morphology and cell differentiation. J. Invest. Dermatol. 115: 658 – 663. 14. Mosialos, G., et al. (1994). Epstein–Barr virus infection induces expression in B lymphocytes of a novel gene encoding an evolutionarily conserved 55-kilodalton actinbundling protein. J. Virol. 68: 7320 –7328. 15. Edwards, R. A., Herrera-Sosa, H., Otto, J., and Bryan, J. (1995). Cloning and expression of a murine fascin homolog from mouse brain. J. Biol. Chem. 270: 10764 –10770. 16. Ross R. (2003). Transcriptional targeting of dendritic cells for gene therapy using the promoter of the cytoskeletal protein fascin. Gene Ther. 10: 1035–1040. 17. Porgador, A., Irvine, K. R., Iwasaki, A., Barber, B. H., Restifo, N. P., and Germain, R. N. (1998). Predominant role for directly transfected dendritic cells in antigen presentation to CD8⫹ T cells after gene gun immunization. J. Exp. Med. 188: 1075–1082. 18. Condon, C., Watkins, S. C., Celluzzi, C. M., Thompson, K., and Falo, L. D. Jr. (1996). DNA-based immunization by in vivo transfection of dendritic cells. Nat. Med. 2: 1122– 1128. 19. Bouloc, A., Walker, P., Grivel, J. C., Vogel, J. C., and Katz, S. I. (1999). Immunization through dermal delivery of protein-encoding DNA: a role for migratory dendritic cells. Eur. J. Immunol. 29: 446 – 454. 20. Bot, A., Stan, A. C., Inaba, K., Steinman, R., and Bona, C. (2000). Dendritic cells at a DNA vaccination site express the encoded influenza nucleoprotein and prime MHC class I-restricted cytolytic lymphocytes upon adoptive transfer. Int. Immunol. 12: 825– 832. 21. Raz, E., et al. (1996). Preferential induction of a Th1 immune response and inhibition of specific IgE antibody formation by plasmid DNA immunization. Proc. Natl. Acad. Sci. USA 93: 5141–5145. 22. Tu¨ ting, T., Gambotto, A., Robbins, P. D., Storkus, W. J., and DeLeo, A. B. (1999). Co-delivery of T helper 1-biasing cytokine genes enhances the efficacy of gene gun immunization of mice: studies with the model tumor antigen ␤-galactosidase and the BALB/c Meth A p53 tumor-specific antigen. Gene Ther. 6: 629 – 636. 23. Hochreiter, R., Hartl, A., Freund, J., Valenta, R., Ferreira, F., and Thalhamer, J. (2001). The influence of CpG motifs on a protein or DNA-based Th2-type immune response against major pollen allergens Bet v 1a, Phl p 2 and Escherichia coli-derived ␤-galactosidase. Int. Arch. Allergy Immunol. 124: 406 – 410. 24. Weiss, R., et al. (2000). Genetic vaccination against malaria infection by intradermal and epidermal injections of a plasmid containing the gene encoding the Plasmodium berghei circumsporozoite protein. Infect. Immun. 68: 5914 –5919. 25. McCluskie, M. J., et al. (1999). Route and method of delivery of DNA vaccine influence immune responses in mice and non-human primates. Mol. Med. 5: 287–300. 26. Steitz, J., Bru¨ ck, J., Knop, J., and Tu¨ ting, T. (2001). Adenovirus-transduced dendritic cells stimulate cellular immunity to melanoma via a CD4⫹ T cell-dependent mechanism. Gene Ther. 8: 1255–1263. 27. Morita, A., et al. (2001). Development of a Langerhans cell-targeted gene therapy format using a dendritic cell-specific promoter. Gene Ther. 8: 1729 –1737. 28. Snapper, C. M., and Paul, W. E. (1987). Interferon-␥ and B cell stimulatory factor-1 reciprocally regulate Ig isotype production. Science 236: 944 –947. 29. Pertmer, T. M., Roberts, T. R., and Haynes, J. R. (1996). Influenza virus nucleoproteinspecific immunoglobulin G subclass and cytokine responses elicited by DNA vaccination are dependent on the route of vector DNA delivery. J. Virol. 70: 6119 – 6125. 30. Toda, M., et al. (2000). Inhibition of immunoglobulin E response to Japanese cedar pollen allergen (Cry j 1) in mice by DNA immunization: different outcomes dependent on the plasmid DNA inoculation method. Immunology 99: 179 –186. 31. Lewis, P. J., van Drunen Littel-van den Hurk, S., and Babiuk, L. A (1999). Altering the cellular location of an antigen expressed by a DNA-based vaccine modulates the immune response. J. Virol. 73: 10214 –10223. 32. Mor, G., et al. (1995). Complexity of the cytokine and antibody response elicited by

MOLECULAR THERAPY Vol. 8, No. 4, October 2003 Copyright © The American Society of Gene Therapy

ARTICLE

doi:10.1016/S1525-0016(03)00242-9

33.

34. 35.

36.

37.

38.

39.

40.

immunizing mice with Plasmodium yoelii circumsporozoite protein plasmid DNA. J. Immunol. 155: 2039 –2046. Boyle, C. M., Morin, M., Webster, R. G., and Robinson, H. L. (1996). Role of different lymphoid tissues in the initiation and maintenance of DNA-raised antibody responses to the influenza virus H1 glycoprotein. J. Virol. 70: 9074 –9078. Wagner, H. (2002). Interactions between bacterial CpG-DNA and TLR9 bridge innate and adaptive immunity. Curr. Opin. Microbiol. 5: 62– 69. Weiss, R., Scheiblhofer, S., Freund, J., Ferreira, F., Livey, I., and Thalhamer, J. (2002). Gene gun bombardment with gold particles displays a particular Th2-promoting signal that over-rules the Th1-inducing effect of immunostimulatory CpG motifs in DNA vaccines. Vaccine 20: 3148 –3154. Bradley, L. M., Harbertson, J., Biederman, E., Zhang, Y., Bradley, S. M., and Linton, P. J. (2002). Availability of antigen-presenting cells can determine the extent of CD4 effector expansion and priming for secretion of Th2 cytokines in vivo. Eur. J. Immunol. 32: 2338 –2346. Boyle, J. S., Koniaras, C., and Lew, A. M. (1997). Influence of cellular location of expressed antigen on the efficacy of DNA vaccination: cytotoxic T lymphocyte and antibody responses are suboptimal when antigen is cytoplasmic after intramuscular DNA immunization. Int. Immunol. 9: 1897–1906. Haddad, D., Liljeqvist, S., Stahl, S., Perlmann, P., Berzins, K., and Ahlborg, N. (1998). Differential induction of immunoglobulin G subclasses by immunization with DNA vectors containing or lacking a signal sequence. Immunol. Lett. 61: 201–204. Higgins, T. J., Herold, K. M., Arnold, R. L., McElhiney, S. P., Shroff, K. E., and Pachuk, C. J. (2000). Plasmid DNA-expressed secreted and nonsecreted forms of herpes simplex virus glycoprotein D2 induce different types of immune responses. J. Infect. Dis. 182: 1311–1320. Piechocki, M. P., Pilon, S. A., and Wei, W. Z. (2001). Complementary antitumor

MOLECULAR THERAPY Vol. 8, No. 4, October 2003 Copyright © The American Society of Gene Therapy

41.

42.

43. 44.

45. 46. 47.

48. 49. 50.

immunity induced by plasmid DNA encoding secreted and cytoplasmic human ErbB-2. J. Immunol. 167: 3367–3374. Hasan, U. A., Harper, D. R., Wren, B. W., and Morrow, W. J. (2002). Immunization with a DNA vaccine expressing a truncated form of varicella zoster virus glycoprotein E. Vaccine 20: 1308 –1315. Brocker, T., Riedinger, M., and Karjalainen, K. (1997). Targeted expression of major histocompatibility complex (MHC) class II molecules demonstrates that dendritic cells can induce negative but not positive selection of thymocytes in vivo. J. Exp. Med. 185: 541–550. Kurts, C., Cannarile, M., Klebba, I., and Brocker, T. (2001). Dendritic cells are sufficient to cross-present self-antigens to CD8 T cells in vivo. J. Immunol. 166: 1439 –1442. Bonkobara, M, Zukas, P. K, Shikano, S., Nakamura, S., Cruz, P. D. Jr., and Ariizumi, K. (2001). Epidermal Langerhans cell-targeted gene expression by a dectin-2 promoter. J. Immunol. 167: 6893– 6900. Xu, S., et al. (1995). Successive generation of antigen-presenting, dendritic cell lines from murine epidermis. J. Immunol. 154: 2697–2705. Jonuleit, H., Schmitt, E., Steinbrink, K., and Enk, A. H. (2001). Dendritic cells as a tool to induce anergic and regulatory T cells. Trends Immunol. 22: 394 – 400. Sudowe, S., Montermann, E., Steitz, J., Tu¨ ting, T., Knop, J., and Reske-Kunz, A. B. (2002). Efficacy of recombinant adenovirus as vector for allergen gene therapy in a mouse model of type I allergy. Gene Ther. 9: 147–156. Carbone, F. R., and Bevan, M. J. (1990). Class I-restricted processing and presentation of exogenous cell-associated antigen in vivo. J. Exp. Med. 171: 377–387. Matzinger, P. (1991). The JAM test. A simple assay for DNA fragmentation and cell death. J. Immunol. Methods 145: 185–192. Sudowe, S., Specht, C., Kolbe, L., and Ko¨ lsch, E. (1995). In situ dormancy of B lymphocytes programmed for an IgE antibody response and their sudden release from unresponsiveness under in vitro conditions. Int. Immunol. 7: 1799 –1807.

575