Genetically Modified Dendritic Cell Vaccine Is a Much More Potent ...

ARTICLE

doi:10.1016/j.ymthe.2005.10.018

Development of Cell-Based Tuberculosis Vaccines: Genetically Modified Dendritic Cell Vaccine Is a Much More Potent Activator of CD4 and CD8 T Cells Than Peptide- or Protein-Loaded Counterparts Janet I. Malowany, Sarah McCormick, Michael Santosuosso, Xizhong Zhang, Naoko Aoki, Patricia Ngai, Jun Wang, Jaina Leitch, Jonathan Bramson, Yonghong Wan, and Zhou Xing* Department of Pathology and Molecular Medicine and Division of Infectious Diseases, Centre for Gene Therapeutics, McMaster University, Hamilton, ON, Canada L8N 3Z5 *To whom correspondence and reprint requests should be addressed at the Department of Pathology and Molecular Medicine, Room 4012-MDCL, McMaster University, 1200 Main Street West, Hamilton, ON, Canada L8N 3Z5. Fax: +1 905 522 6750. E-mail: [email protected].

Available online 15 December 2005

Genetically modified dendritic cell (DC)-based vaccines have not been explored for immunization against tuberculosis. A gene-modified DC vaccine expressing Mycobacterium tuberculosis (M.tb) antigen 85A (Ag85A) was developed by using a recombinant replication-deficient adenoviral gene transfer vector (AdAg85A). AdAg85A-transduced DC vaccine (AdAg85/DC) expressed higher levels of IL-12 and was much more immunogenic than Ag85 protein-loaded (pro/DC) or CD4/CD8 T cell peptide-loaded (pep/DC) DC vaccines. Compared to pro/DC or pep/DC, AdAg85/DC elicited a remarkably higher level of ex vivo IFN-; production by CD4 and CD8 T cells at weeks 2, 6, and 12 postimmunization, which was coupled with higher frequencies of antigen-specific T cells. By an in vivo CD8 or CD4 T cell cytotoxicity (CTL) assay, AdAg85/DC was shown to provoke much higher and more sustained levels of CD8 and CD4 CTL activity up to 12 weeks postimmunization. Intramuscular (im) AdAg85/DC immunization was more potent than the iv route of AdAg85/DC immunization. Such stronger immunogenicity of im AdAg85/DC vaccination was corroborated with better protection from M.tb challenge. Our results thus suggest that genetically modified DC-based TB vaccine is superior to subunit DC vaccines and has the potential for therapeutic applications.

INTRODUCTION Tuberculosis (TB) is one of the leading infectious causes of death worldwide [1]. The TB epidemic has been worsened by HIV infection [2]. Bacille Calmette Gue´rin (BCG) is currently the only TB vaccine available and is administered to humans shortly after birth. While BCG is effective in protecting from childhood TB, it is ineffective in protecting from adult TB [3,4]. Thus, one of the challenges to TB vaccinologists is to develop effective prophylactic TB vaccines able to confer long-term protection against pulmonary TB [3,4]. Furthermore, since about 1/3 of the world population has been latently infected by Mycobacterium tuberculosis (M.tb) and there is an ever-increasing incidence of drug-resistant TB, particularly in HIVinfected hosts, there is also an urgent need to develop therapeutic TB vaccines [3,4]. In this regard, since BCG may not be safe for immune-compromised hosts and has

766

proven ineffective in boosting immune anti-TB responses in BCG vaccinees [5,6], a therapeutic TB vaccine will likely be a recombinant genetic vaccine that is not only effective but also safe and repeatable. To date, recombinant DNA TB vaccines have been evaluated for their therapeutic effects [7] but recent results indicate that this form of vaccine is ineffective for therapeutic purposes [8,9]. Dendritic cells (DC) are the most potent professional antigen-presenting cells. At the interface of innate and adaptive immunity, these cells are critical to antigen sampling and activation of naRve T cells, having the power to polarize the immune response toward the type 1 or type 2 phenotype [10]. Thus, exploiting their potent immune activating properties, DC have been differentiated in vitro and manipulated as vaccines to deliver antigens for cancer immunotherapy [11,12] or prevention of infectious diseases [13]. While the most com-

MOLECULAR THERAPY Vol. 13, No. 4, April 2006 Copyright C The American Society of Gene Therapy 1525-0016/$30.00

doi:10.1016/j.ymthe.2005.10.018

monly used method to prepare DC vaccines is to load DC with immunogenic peptide or protein (bsubunit DC vaccinesQ) [14], recent studies, particularly those carried out in cancer models, have suggested that the DC vaccines virally transduced to express immunogenic peptides or proteins are more potent than subunit DC vaccines. Of several viral vector systems, recombinant adenoviral vector has been used effectively to transduce tumor antigen-coding genes into DC, and such adenovirally transduced DC cancer vaccines have been found to be not only more effective than subunit DC vaccine but also safe and repeatable [15,16]. While relatively few studies thus far have explored such viral-gene-modified DC vaccines and compared their potency to subunit DC vaccines for generating immune protection against infectious disease, emerging evidence supports their potential to be used as effective, safe, and repeatable cell-based vaccines to prevent a variety of diseases that are significant causes of morbidity and mortality. DC vaccines loaded with M.tb immunogenic proteins or peptides were previously evaluated in murine TB models with variable results [17,18]. However, gene-

ARTICLE

modified DC TB vaccines remain to be explored and/or compared with peptide/protein DC counterparts. In the past several years, we have been using adenoviral gene transfer vectors to transduce cancer antigen-coding genes into DC and found that such genetically modified DC cancer vaccines are potent in eliciting anti-cancer immune responses [19–22]. To develop adenoviral-transduced DC TB vaccines, we constructed a recombinant replication-deficient adenoviral TB vaccine expressing an immunodominant M.tb antigen, Ag85A (AdAg85A), and demonstrated its potency in protection against TB [23,24]. In our current study, we have used this adenoviral vector to develop an M.tb gene-modified DC vaccine and compared this DC vaccine with DC vaccines loaded with Ag85 proteins or Ag85A immunodominant CD4/ CD8 T cell peptides. We found that AdAg85A-transduced DC vaccine was much more potent than peptide- or protein-loaded DC counterparts in eliciting long-lasting immune activation of both CD4 and CD8 T cells. We believe that our findings hold implications in the future design of DC-based vaccines for TB as well as other intracellular infectious diseases.

FIG. 1. Immune phenotypes of dendritic cell TB vaccines. In vitro-differentiated dendritic cells were washed 5 h after being incubated without (DC) or with AdAg85A (AdAg85/DC), Ag85 complex proteins (Pro/DC), T cell peptides (Pep/DC), or LPS (LPS/DC). The cells were then incubated for an additional 19 h before immunostaining, FACS analysis, and/or ICCS. FACS analysis was carried out by gating on MHC class II- and CD11c-positive cells.

MOLECULAR THERAPY Vol. 13, No. 4, April 2006 Copyright C The American Society of Gene Therapy

767

ARTICLE

doi:10.1016/j.ymthe.2005.10.018

RESULTS Immune Characteristics of DC Vaccines To examine whether different modes of DC antigen loading may differentially alter the property of DC, we prepared DC vaccines as described under Materials and Methods and analyzed them by FACS at 24 h posttreatment. We used lipopolysaccharide (LPS) as a positive control for stimulation. We gated on DC stained positive for MHC class II and subsequently examined them for CD11c and surface activation markers or cytokine expression by FACS analysis software. Untreated DC, compared with AdAg85A-infected DC (AdAg85/DC) and Ag85 protein-loaded DC (pro/DC), similar to LPS/DC, had strikingly increased cell surface expression of B7.1 (Fig. 1) and CD40 (data not shown), whereas CD4/CD8 T cell peptide-loaded DC (pep/DC) did not demonstrate such increased expression. In comparison, untreated DC already expressed similarly high levels of B7.2, and AdAg85/DC, pro/DC, and LPS-treated DC demonstrated only mildly increased B7.2 or MHC class II expression (data not shown). By using intracellular cytokine staining (ICCS) we examined cytokine production by the different DC vaccine formulations. We found that compared to untreated DC, AdAg85/DC, pro/DC, or pep/DC showed little increase in TNFa production (Fig. 1). While LPS treatment did not increase TNFa production above the level seen in other DC at 24 h (Fig. 1), it did markedly increase TNFa responses at an earlier time of 5 h (data not shown). In sharp contrast to TNFa responses, compared to untreated DC, AdAg85/DC had a remarkably higher level of IL-12 production than pro/DC or pep/DC and such increased IL-12 expression in AdAg85/DC was comparable to that induced by LPS stimulation (Fig. 1). These data suggest that while viral gene-modified DC retain other properties of dendritic cells, they assume a much more pronounced type 1 immune-activating phenotype than protein- or peptide-loaded counterparts. Intramuscular AdAg85/DC Vaccination Leads to Potent T Cell Activation To compare the potential of immune activation by various DC vaccines, we immunized mice im by AdAg85/DC, pro/DC, or pep/DC and sacrificed them at weeks 2, 6, and 12 postvaccination. By using ICCS and FACS techniques, we first analyzed the number of antigen-specific, IFN-g-releasing CD4 and CD8 T cells in the spleen. At all time points compared to pro/DC and pep/DC vaccines, AdAg85/DC vaccination generated a much greater number of splenic IFN-g-producing CD8 or CD4 T cells upon stimulation by respective immune-dominant peptides (Figs. 2A and 2B). The frequency of antigen-specific CD8 T cells was greater than that of CD4 T cells. The level of increased CD8 T cells appeared to be better sustained than the CD4 counterpart. The control DC vaccine (DC infected only

768

FIG. 2. Percentage of IFN-g-producing, Ag85A-sepcific CD8 and CD4 T cells in the spleen following intramuscular immunization with pep/DC, pro/DC, or AdAg85/DC. Mice were im immunized with various DC vaccines and sacrificed at weeks 2, 6, and 12. The whole splenocytes were cultured for 6 h with or without (A) CD8 or (B) CD4 T cell peptide. The cells were then subjected to immunostaining, ICCS, and FACS analysis. The data are expressed as the average % F SEM (IFN-g-positive cells of all splenocytes) from three or four mice/DC vaccine/time.

with an empty adenoviral vector) elicited little response (data not shown). Since AdAg85A/DC vaccine was prepared in such a way that DC were infected with an adenoviral vector expressing Ag85A, there was a possibility that the lack of efficiency by pep/DC or pro/DC vaccine was due to the lack of DC activation caused directly by viral infection per se. To rule it out, we prepared pep/DC and pro/DC vaccines that were also infected with a control adenoviral vector (pepAddl/DC and proAddl/DC) and compared the immunogenicity of these two additional control vaccines with that of pep/ DC, pro/DC, or AdAg85A/DC. We found that adenoviral infection of pep/DC or pro/DC still failed to activate Ag85A-specific CD8 or CD4 T cells effectively (Figs. 3A and 3B). By using different culture conditions and ELISA, we examined the IFN-g production capacity of T cells and found that overall such increased numbers of antigen-specific T cells by AdAg85/DC vaccine were accompanied by proportionately increased levels of IFN-g protein release into culture supernatants (Figs. 4A and 4B). Of note, at 12 weeks, the IFN-g production level was lower than at earlier times, suggesting that the


doi:10.1016/j.ymthe.2005.10.018

ARTICLE

was an appreciable level of target cell lysis in mice elicited by all three forms of DC vaccine, animals vaccinated with AdAg85/DC were 4.5 times more effective (more than 80% target lysis) in killing CD8 peptidepulsed target cells in vivo than pro/DC and pep/DC groups (Figs. 5A and 5B). At both weeks 6 and 12, the level of CD8 T-cell-mediated killing markedly declined to a minimal level in both pro/DC and pep/DC groups, whereas the level of CD8 CTL was sustained by AdAg85/ DC vaccination, maintaining 50% antigen-specific target lysis (Fig. 5A). In contrast, the control DC vaccine (Addl/ DC) failed to elicit any antigen-specific target lysis at any time points (data not shown). By pulsing the target cells with a CD4 T cell peptide, we also examined the level of CD4 T-cell-mediated CTL in vivo. CD4 T-cell-mediated CTL has been found to play a role in host defense against mycobacterial infection [28,29]. To this end, we carried out an in vivo CD4 CTL assay as described under Materials and Methods and sacrificed the mice 24 h post-target cell delivery. Compared to CD8 CTL, the magnitude of CD4 CTL was lower.

FIG. 3. Percentage of IFN-g-producing, Ag85A-specific CD8 and CD4 T cells in the spleen following intramuscular immunization with pep/DC, pepAddl/DC, pro/DC, proAddl/DC, or AdAg85/DC. Mice were im immunized with various DC vaccines and sacrificed at week 2. The whole splenocytes were cultured for 6 h with or without (A) CD8 or (B) CD4 T cell peptide. The cells were then subjected to immunostaining, ICCS, and FACS analysis. The data are expressed as the average % of two mice/vaccine (IFN-g-positive cells of all splenocytes).

majority of antigen-specific T cells have entered a memory phase at this time. Overall, these observations indicate that AdAg85/DC vaccine is very potent in activating both CD8 and CD4 T cells, whereas pro/DC and pep/DC vaccines are weaker stimulators. Intramuscular AdAg85/DC Vaccination Leads to the Development of Potent CD8 and CD4 T-Cell-Mediated Cytotoxicity in Vivo We next set out to evaluate whether enhanced T cell activation elicited by AdAg85/DC vaccination was accompanied by enhanced antigen-specific T cytotoxic activity in vivo. To address this, we employed an in vivo CD8 or CD4 T cell cytotoxicity (CTL) assay [24,27]. To examine CD8 CTL, we immunized mice im with Addl/ DC, AdAg85/DC, pro/DC, or pep/DC vaccine, and at weeks 2, 6, or 12 postvaccination, we injected the mice iv with CD8 T cell Ag85A peptide-pulsed, carboxyfluorescein diacetate succinimidyl ester (CFSE)-labeled target cells and sacrificed them 6 h post-target delivery. We then analyzed the splenocytes by FACS for the presence of CFSE-labeled cells and calculated the percentage of lysis of the peptide-pulsed targets. At week 2, while there


FIG. 4. Quantitation of IFN-g production by Ag85A-specific CD8 or CD4 T cells in the spleen following intramuscular immunization with pep/DC, pro/ DC, or AdAg85/DC. Mice were im immunized with various DC vaccines and sacrificed at weeks 2, 6, and 12. The whole splenocytes pooled from three or four mice/vaccine/time were cultured for 24 h with or without (A) CD8 or (B) CD4 T cell peptide stimulation. The amount of IFN-g protein released into the culture supernatant was quantified by ELISA. Results are expressed as the means F SEM of triplicate cultures.

769

ARTICLE

doi:10.1016/j.ymthe.2005.10.018

FIG. 5. In vivo CD8 and CD4 T-cell-mediated cytotoxic activity following intramuscular immunization with pep/DC, pro/DC, or AdAg85/DC. Mice were im immunized with pep/DC, pro/DC, or AdAg85/DC vaccine. (A) At weeks 2, 6, and 12 postimmunization, CFSE-labeled Ag85A-specific CD8 T cell peptide-pulsed target cells were transferred intravenously into the immunized mice and the splenocytes were isolated 5 h after target cell transfer and examined by FACS analysis for the measurement of CD8 T-cell-mediated cytotoxicity (CTL), with the representative histogram of FACS analysis presented (B; week 2). (C) In separate experiments Ag85A-specific CD4 T cell peptide-pulsed target cells were transferred intravenously into the immunized mice and the splenocytes were isolated and examined by FACS analysis for the measurement of CD4 T-cell-mediated CTL. Results represent % loss of peptide-pulsed target cells and are expressed as the means F SEM of three or four mice/DC vaccine/time.

At both weeks 2 and 6 postvaccination, pro/DC and pep/ DC induced a small level of CD4 CTL activity, which decreased to a minimal level by 12 weeks postimmunization (Fig. 5C). In contrast, AdAg85/DC elicited higher levels of CD4 CTL at all time points, with a peak activity detected at week 6 (Fig. 5C), which differed from AdAg85/ DC-induced CD8 CTL, peaking at week 2 (Fig. 5A). The above CD8 and CD4 CTL data indicate that when given intramuscularly, AdAg85/DC vaccine is a much stronger

770

activator of Ag85A-specific CD8 and CD4 effector T cells than pro/DC or pep/DC vaccine. Intramuscular Route of AdAg85/DC Vaccination Is More Potent Than Intravenous Vaccination Since intravenous (iv) immunization with a DC TB vaccine loaded with CD4/CD8 T cell peptides was previously found to be effective [18], we compared the level of immune activation by im and iv vaccination with


doi:10.1016/j.ymthe.2005.10.018

ARTICLE

pro/DC, pep/DC, or AdAg85/DC. Similar to im immunization, iv immunization with pro/DC or pep/DC vaccine induced a minimum level of activation of CD4 and CD8 T cells at both weeks 2 and 6 as assessed by ICCS (data not shown). In contrast, at week 2, iv immunization with AdAg85/DC induced a significantly increased number of CD8 T cells accompanied by a relatively small number of CD4 T cells (Figs. 6A and 6B). By 6 weeks, the number of CD8 T cells markedly increased. In comparison, im AdAg85/DC vaccination triggered peak responses of CD8 or CD4 T cell activation at week 2, which was sustained up to week 6 (Figs. 6A and 6B). Since Elispot immunoassay is more sensitive than ICCS, we also used the IFN-g Elispot assay to assess the frequency of antigen-specific CD4 and CD8 T cells. Similar to im vaccination, iv vaccination with the control DC vaccine (Addl/DC), pep/DC, or pro/DC caused some basal levels of antigen-specific IFN-g release; it elicited little antigen-specific response at either week 2 or week 6 (Figs. 7A and 7B). In basic agreement with the results obtained by ICCS, however, iv vaccination with AdAg85/DC elicited enhancement of CD4 and CD8 T cell

FIG. 6. Percentage of IFN-g-producing, Ag85A-sepcific CD8 and CD4 T cells in the spleen following intramuscular or intravenous immunization with AdAg85/DC. Mice were im or iv immunized with AdAg85/DC and sacrificed at weeks 2 and 6. The whole splenocytes were cultured for 6 h with or without (A) CD8 or (B) CD4 T cell peptide. The cells were then subjected to immunostaining, ICCS, and FACS analysis. The data are expressed as the average % F SEM (IFN-g-positive cells of all splenocytes) from three or four mice/vaccination route/time.


FIG. 7. Frequencies of Ag85A-specific CD8 and CD4 T cells in the spleen following intramuscular or intravenous immunization with pep/DC, pro/DC, or AdAg85/DC. Mice were immunized iv with pep/DC, pro/DC, AdAg85/DC, or Addl/DC as a control or immunized im with AdAg85/DC. Mice were then sacrificed at (A) week 2 or (B) week 6. The whole splenocytes were pooled from three or four mice/group and cultured for 24 h with or without CD8 or CD4 T cell peptide. The cells were then subjected to Elispot assay. The number of IFN-g-positive spots/million cells was enumerated. The data are expressed as the means F SEM from triplicate wells/DC vaccine/time. (C) In separate experiments, mice were immunized iv or im with AdAg85/DC. At weeks 2 and 6 postimmunization, CFSE-labeled Ag85A-specific CD8 T cell peptide-pulsed target cells were transferred intravenously to the immunized mice and the splenocytes were isolated 5 h after the target cell transfer and examined by FACS analysis for the measurement of CD8 T-cell-mediated CTL. Results represent % loss of peptide-pulsed target cells and are expressed as the means F SEM of three or four mice/DC vaccine/time.

771

ARTICLE

responses. Intramuscular AdAg85/DC vaccination induced further increased numbers of CD4 and CD8 T cell responses at both time points (Figs. 7A and 7B); this was particularly evident for CD8 T cells. Accompanying greater numbers of CD4 and CD8 T cells by im AdAg85/ DC were higher levels of IFN-g protein in culture supernatant (data not shown). Of interest, while im AdAg85/ DC elicited a higher level of in vivo CD8 CTL activity than iv AdAg85/DC at week 2, the CTL activities were similar between iv and im by week 6 (Fig. 7C). These data suggest that, first, regardless of the route of vaccination (im or iv), pro/DC and pep/DC are much weaker DC vaccines than gene-modified DC and, second, gene-modified DC vaccine, when im administered, elicited stronger immune activation than when it was iv administered. Immune Protection from M.tb Challenge by DC Vaccination To examine and compare the immune protective effect mediated by DC vaccines, we immunized groups of mice im or iv with the three forms of DC-based vaccines or sc with BCG, the current gold standard for protection against M.tb challenge. We then challenged the mice intraperitoneally with M.tb 6 weeks postvaccination and sacrificed them 4 weeks postchallenge for evaluation of bacterial burden in the spleen and lung. Mice immunized im or iv with pro/DC or pep/DC vaccine had relatively

FIG. 8. Immune protection from M.tb challenge by DC vaccination. Mice were immunized iv or im with pep/DC, pro/DC, or AdAg85/DC or sc with BCG as a control. At 6 weeks postimmunization, mice were ip infected with M.tb and the level of M.tb infection in the spleen was assessed by colony formation assay 4 weeks after M.tb challenge. Results (colony-forming units/spleen) are expressed as the means F SEM of seven or eight mice/group. The difference between im AdAg85/DC and im pep/DC, im pro/DC, iv AdAg85/DC, iv pep/ DC, or iv pro/DC is statistically significant (P = 0.008, P = 0.01, P = 0.0002, P = 0.04, or P = 0.04, respectively).

772

doi:10.1016/j.ymthe.2005.10.018

high levels of colony-forming units (cfu) in the spleen (Fig. 8). The mice immunized iv with AdAg85/DC also had similarly high levels of cfu. In comparison, the im AgAg85/DC group had much lower levels of M.tb infection, which were comparable to the levels following BCG vaccination (Fig. 8). The level of infection in the lung was much lower than that in the spleen and the trend was similar to that in the spleen (data not shown). These protection data are in keeping with the immunogenicity data, suggesting that gene-modified DC vaccine, when given im, triggers better immune protection against systemic M.tb challenge.

DISCUSSION While antigenic peptides or proteins have often been used to prepare DC-based vaccines, our previous work has demonstrated the potency of adenoviral-mediated gene transfer to DC for the purpose of cancer vaccine development [19–22]. Our recent development of a novel recombinant adenoviral TB vaccine expressing Ag85A [23,24] has thus provided us a unique opportunity to compare the immune-activating capabilities of three types of dendritic cell-based TB vaccines: AdAg85A gene modified (AdAg85/DC), Ag85A CD4/CD8 peptide-loaded (pep/DC), and Ag85 protein-loaded (pro/DC). We found that AdAg85/DC vaccine after intramuscular inoculation was able to elicit a long-lasting immune activation characterized by a type 1 polarization of both CD4 and CD8 T cells exemplified by markedly increased frequencies of antigen-specific IFN-g-releasing CD4 and CD8 T cells and markedly increased CD8 and CD4 T-cellmediated in vivo CTL activities. We also found that the intramuscular route of delivery of AdAg85/DC was more potent than iv delivery. In comparison, pep/DC or protein/DC vaccine led to activation of a very small number of antigen-specific, IFN-g-releasing CD4 and CD8 T cells at all time points examined and was able to elicit only a small and transiently raised level of CD8 or CD4 Tcell-mediated CTL activities in vivo. The superiority of gene-modified DC vaccine over peptide or protein-loaded DC vaccines corroborated with its better immune protection from M.tb challenge. However, the longevity of protection by such gene-modified DC vaccine requires further investigation. We represent the first group to have developed a genemodified DC TB vaccine by using a replication-deficient, Ag85A-expressing adenoviral vector, and we compared this form of DC vaccine with peptide- or protein-loaded DC vaccines in their ability to activate CD4 and CD8 T cells. An Ag85 protein DC vaccine was previously reported not to have protective efficacy [17]. McShane and colleagues have developed a DC TB vaccine loaded with both CD4 and CD8 Ag85A peptides and found that iv vaccination with this vaccine was able to provide protection from systemic M.tb challenge [18]. Of note, repeated


ARTICLE

doi:10.1016/j.ymthe.2005.10.018

DC vaccinations appeared to be required for immune protection. In comparison, in our current study, we demonstrated that a single delivery of an Ag85A genecarrying DC vaccine led to very potent immune activation, contrasting with CD4/CD8 peptide-loaded or Ag85 protein-loaded DC counterparts that resulted in only weak immune activation. In addition to its potency, gene-modified DC vaccine is more economical than peptide- or protein-loaded DC vaccine. Although DC TB vaccine may also be prepared by using live attenuated replication-competent M. bovis BCG, BCG/DC vaccine was found to elicit a very short-lived protective immunity (mice had to be challenged by 1 week postvaccination) [30] and this form of DC vaccine may not be safe for immune-compromised hosts. Although our current study indicates that AdAg85A/DC confers a level of immune protection similar to that of BCG, it is both unlikely and impractical to replace the current BCG vaccine with a gene-modified DC vaccine as a mainstream vaccine. However, as gene-modified DC vaccines have the advantage of being both safe and effective over BCG or BCGinfected DC, they are attractive immunotherapeutics for immune-compromised hosts with multidrug-resistant TB. The conclusion from our study is further supported by findings from cancer models in which adenoviral gene vector-modified DC vaccines were found to be much more potent than peptide- or protein-loaded counterparts [31]. Although there is a lack of reports in which gene-modified DC vaccine was compared with both peptide-loaded and protein-loaded DC vaccines in infectious disease models, it has been shown that viral gene transfer vector-modified DC vaccine is more effective than peptide-loaded counterparts [32,33]. Similar to what has been found in a murine TB model [17], chlamydial protein-loaded DC provided no protection against bacterial challenge [34]. Compared to other viral gene transfer vector systems, adenoviral gene transfer vector stands out as an attractive vehicle for generating gene-modified DC vaccines against intracellular infectious diseases [13,16]. It is believed that the potency of adenoviral-modified DC vaccines results from several mechanisms: (1) DC is the most potent T-cell-activating cellular component, serving as a mobile immune adjuvant capable of shuttling microbial antigens directly to the site of lymphocyte activation; (2) compared to peptide or antigenic protein loading, adenoviral gene transfer to DC leads to much prolonged microbial antigen expression and presentation of multiple antigenic epitopes to both CD4 and CD8 T cells; and (3) adenoviral backbone serves as an additional potent type 1 immune adjuvant. Indeed, we found that, while compared to peptide or protein loading, AdAg85A transduction only moderately enhanced B7.2 and CD40 expression of DC, it most potently enhanced IL-12 production by DC. Since IL-12 is an important Th1activating cytokine [35], this has been considered to be a


critical feature of effective DC vaccine [36]. In our current study, in addition to comparing gene-modified, peptide-loaded, and protein-loaded DC vaccines, we have also compared the intramuscular route with the intravenous route of AdAg85/DC vaccination and found that im DC vaccination led to better immune activation and protection than iv vaccination. Increasing evidence suggests that the route of DC vaccination is critical to the effectiveness of immune activation and protection and the optimal route of vaccination varies depending on the nature of the infectious disease. For example, the iv route, but not the intradermal or intraperitoneal route, of DC vaccination was protective from leishmaniasis [36]. On the other hand, the subcutaneous route, but not the iv route, of DC vaccination protected from Candida infection [37]. Our current findings hold important implications in the development of therapeutic TB vaccines for treating drug-resistant TB, which represents a unique challenge to TB vaccinologists. Although there was a great deal of initial optimism, recent experimental assays of plasmid DNA-based vaccines for TB therapy have failed [8,9]. We believe that viral gene-modified DC vaccine brings forward a renewed hope in this regard. In addition to its potency as we have demonstrated herein, adenoviraltransduced DC vaccine has been shown not to induce significant antibody or T cell responses against adenoviral antigens, thus allowing repeated vaccinations and repeated expression of the transgene coding for the microbial antigen of interest, in contrast to direct inoculation of adenovirus [31,38].

MATERIALS AND METHODS Mice and reagents. Six- to ten-week-old female Balb/c mice were purchased from Harlan Laboratories (Indianapolis, IN, USA). The animals were housed in a specific-pathogen-free facility in the Central Animal Facility at McMaster University and cared for in accordance with the McMaster Animal Research Ethics Board. The construction and amplification of a replication-deficient (E1/E3-deleted) recombinant adenovirus encoding the gene for antigen 85A has recently been described [23] and it was used to transduce bone marrow-derived dendritic cells. An adenoviral vector (Addl70-3) was used as control (the replication-deficient adenovirus that does not express any foreign gene). All viruses were purified and stored at 708C until needed. Purified M.tb antigen 85 complex protein was provided by Colorado State University through funds from the National Institute of Allergy and Infectious Diseases (Contract 1-AI75320). Two synthetic Ag85A peptides specific for Balb/c background mice (H-2d) were used. The MHC class I-specific peptide (MPVGGQSSF) and the MHC class II-specific peptide (LTSELPGWLQANRHVKPTGS) [24] were synthesized by Dalton Chemical Laboratories (Toronto, ON, Canada). As control peptides, the H-2d-restricted peptide from h-galactosidase876–884 (TPHPARIGL) was used for CD8 CTL experiments, and I-Adrestricted ovalbumin peptide (OVA323–339; ISQAVHAAHAEINEAGR) was used for CD4 CTL experiments. All proteins were dissolved in dimethyl sulfoxide and stored at 208C until needed. Generation of bone marrow-derived dendritic cells and preparation of DC TB vaccines. Bone marrow cells were harvested from the femurs and tibiae of naRve Balb/c mice and cultured in RPMI 1640 containing 10% fetal bovine serum (FBS), 100 U/ml penicillin, 100 Ag/ml streptomycin, 2

773

ARTICLE

mM l-glutamine, 50 AM h-mercaptoethanol, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, as well as 40 pg/ml recombinant murine granulocyte–macrophage colony-stimulating factor (GM-CSF) (Schering– Plough Research Institute, NJ, USA) [21,22]. Three and six days after initial culture, cells were replenished with fresh medium supplemented with GM-CSF. On day 6, some DC were incubated with Ag85 complex proteins overnight at the concentration of 7 Ag/ml and used as Ag85 proteinloaded DC vaccine (pro/DC). On day 7, other DC were either infected with AdAg85A (100 pfu/cell) or loaded with both CD4 and CD8 Ag85A peptides (1 Ag/ml each) for 5 h. As additional controls, some pro/DC and pep/DC were infected with an empty control adenoviral vector (Addl703). At the end of incubation with antigens, all DC were harvested, washed three times with PBS, and used to immunize mice or characterized phenotypically by FACS analysis. Immunization with dendritic cell TB vaccines. Vaccines were prepared immediately prior to immunization and kept on ice until use. Intramuscular immunization was carried out by injecting 0.5 106 DC/mouse in 100 Al PBS into the hind legs (50 Al each leg). Intravenous immunization was carried out by injecting 0.5 106 DC/mouse in 200 Al PBS into the tail vein. Live M. bovis BCG (0.5 106 cfu/mouse; Connaught Laboratories, North York, ON, Canada) was injected sc over both hips, as a control in M.tb challenge experiments. Phenotypic characterization of DC vaccines. The phenotypic analysis of DC was performed by FACS, after 5 and 24 h of stimulation of DC cultures. Vaccine DC were stimulated as described above; additional stimuli included an adenoviral vector expressing ovalbumin (AdOVA.1) and LPS (0.1 Ag/ml). After 5 h of stimulation, DC were removed, washed, and resuspended in a bovine serum albumin (BSA)-based FACS buffer (0.2% BSA/0.1% NaN3/PBS) and kept at 48C until immunostaining. The remaining incubating cells were also gently removed and washed to remove any remaining stimuli, resuspended in 3 ml of GM-CSFsupplemented DC medium, and returned to the appropriate dish for an additional 19 h. Cells were then stained as previously described [23,24] with the following anti-mouse antibodies (Pharmingen): allophycocyanin-conjugated CD11c; fluorescein isothiocyanate (FITC)-conjugated anti-I-Ad or anti-TNFa; phycoerythrin (PE)-conjugated anti-CD80, antiCD86, anti-IL-6, or anti-IL-12; biotinylated anti-I-Ad or anti-CD40; and cychrome (Cyc)-conjugated streptavidin (SA). Lymphocyte isolation and culture for cytokine ELISA or Elispot assay. Splenocytes were isolated as previously described [23,24]. Briefly, spleens were removed into PBS. Splenocyte suspension was filtered through a 55Am nylon membrane before being centrifuged at 1500 rpm for 5 min. Pellets were resuspended in 2 ml/spleen of red blood cell lysis buffer (R&D Systems, Inc.) and incubated at room temperature for 12 min. Approximately 30 ml of PBS was added after the 12-min incubation to stop the red blood cell lysis. The whole splenocytes were filtered through a 55-Am nylon membrane, centrifuged at 1500 rpm for 5 min, and resuspended at 2 106 cells/ml in complete culture medium (RPMI 1640 supplemented with 10% FBS, 100 U/ml penicillin, 100 Ag/ml streptomycin, 2 mM lglutamine, and 50 AM h-mercaptoethanol). These cells were subsequently cultured for the purpose of in vitro antigen stimulation and ELISA or Elispot assay as previously described. For ELISA, 0.5 106 cells were plated in each well of a 96-well plate and cultured at 378C/5% CO2 for 24 h with or without antigen stimulation (4 Ag/ml CD4 or CD8 T cell peptide). The culture supernatants were collected and stored at 208C until IFN-g ELISA (R&D Systems; the sensitivity of detection was 2 pg/ml). For Elispot assay, splenocytes (0.25 106 and 0.5 106/well) were seeded into a 96-well PVDF microplate (Millipore Corp., Bedford, MA, USA) precoated overnight with a mouse IFN-g capture antibody (R&D Systems, Inc.; 1:60 dilution). Cells were cultured at 378C/5% CO2 for 24 h with or without stimulation by Ag85A CD4 peptide (10 Ag/ml), Ag85A CD8 peptide (10 Ag/ml), and Ag85 complex protein (10 Ag/ml). The plate was then developed by the standardized SA-conjugated alkaline phosphatase and chromogen method (R&D Systems, Inc.) and the number of IFN-greleasing cells was determined under a dissecting microscope aided by an enumeration computer software [25].

774

doi:10.1016/j.ymthe.2005.10.018

Intracellular cytokine staining and FACS analysis. ICCS was carried out as previously described [23–25]. Briefly, 2 106 cells were plated in each well of a round-bottom 96-well plate. Cells were cultured at 378C/5% CO2 for 6 h with or without CD4 or CD8 peptide stimulation (5 Ag/ml) along with brefeldin A treatment. ICCS was carried out using Cytofix/Cytoperm Kits (BD PharMingen, Mississauga, ON, Canada) as specified by the manufacturer. Cells were washed, Fc receptors were blocked by incubation with rat anti-mouse CD16/CD32 (Fc Block; PharMingen) for 15 min, and the cells were then stained with Cyc-labeled anti-mouse CD8 and FITClabeled anti-mouse CD4 for 30 min. After fixation and permeabilization with Cytofix/CytoPerm for 20 min, the cells were stained with PE-labeled anti-mouse IFN-g Ab or an isotype control Ab (rat IgG1) for 30 min and analyzed by flow cytometry as described above. A FACScan LRSII instrument was used (Becton–Dickinson, Sunnyvale, CA, USA) to collect 250,000 total events from each sample for analysis. In vivo CD8 and CD4 T cell cytotoxicity assays. The whole splenocytes were isolated as above and resuspended at 20 106 cells/ml in complete medium. The in vivo CD8 and CD4 CTL assays were carried out following the protocols established by us and others [24,26,27]. Briefly, splenocytes from naRve female Balb/c mice were isolated the night prior to each in vivo cytotoxicity assay. For the in vivo cytotoxicity assay, the splenocytes were pulsed with either the CD8 or the CD4 Ag85A peptide or a corresponding control peptide (1 Ag/ml) and incubated overnight at 48C. Such splenocytes used as CTL target cells were then resuspended at 20 106 cells/ml in PBS containing 5% FBS. The Ag85A peptide-pulsed splenocytes were labeled with 5 AM CFSE and denoted CFSEhigh, and the splenocytes pulsed with the control peptide (h-galactosidase peptide for CD8 CTL; ovalbumin peptide for CD4 CTL) were labeled with 0.5 AM CFSE (CFSElow) for 5 min at room temperature under unlit conditions. The cells were washed twice to remove any free peptide with PBS/5% FBS and then once with PBS to obtain an ultimate concentration of 50 106 cells/ml. Both CFSEhigh and CFSElow cell populations (5 106 each/mouse) were mixed in a 1:1 ratio (in a total of 200 Al volume) and injected iv to DC-vaccinated mice. A naRve mouse was also injected with CFSE-labeled cells and used as an bunprimedQ control for calculation (see below). Six hours (for CD8 CTL assay) or 24 h (for CD4 CTL assay) following iv target cell injection, both splenocytes and lung mononuclear cells were isolated. Splenocytes were resuspended in FACS buffer. The in vivo lysis of the target cells in the spleen and lungs was determined according to the extent of loss of 5 AM CFSE dye by flow cytometry. Up to 1 106events were collected for analysis. To calculate antigen-specific, CD8 or CD4 T-cell-mediated lysis, the following formula was used: percentage of specific lysis = [1 (ratio unprimed/ratio primed) 100], where ratio is the percentage CFSElow/ percentage CFSEhigh. Immune protection from M. tuberculosis challenge. Groups of Balb/c mice (seven or eight mice per group) were challenged ip with live M.tb (H37Rv strain). For ip delivery, 2 106 cfu of M.tb was injected. The protective efficacy of vaccination with different immunization regimens was evaluated 4 weeks after M.tb challenge by plating serial 10-fold dilutions of lung and spleen tissue homogenates in quadruplicates on Middlebrook 7H10 agar plates containing OADC enrichment (Difco), as previously described [23–25]. Plates were incubated inside semisealed plastic bags at 378C for 3 weeks, and colonies in each plate were counted.

ACKNOWLEDGMENTS The authors are grateful to Anna Zganiacz, Duncan Chong, and Xueya Feng for their technical assistance and Dr. LinMing Liu for her helpful advice. This study is supported by funds from the Canadian Institutes for Health Research. RECEIVED FOR PUBLICATION JUNE 7, 2005; REVISED OCTOBER 4, 2005; ACCEPTED OCTOBER 27, 2005.

REFERENCES 1. TDR/WHO Reference Group on Tuberculosis (2004). Tuberculosis. Nat. Rev. Microbiol. 2: 930 – 932. 2. Corbett, E. L., et al. (2003). The growing burden of tuberculosis: global trends and interactions with the HIV epidemic. Arch. Intern. Med. 163: 1009 – 1021.


doi:10.1016/j.ymthe.2005.10.018

3. Kaufmann, S. H. (2000). Is the development of a new tuberculosis vaccine possible? Nat. Med. 6: 955 – 960. 4. Wang, J., and Xing, Z. (2002). Tuberculosis vaccines: the past, present and future. Expert Rev. Vaccines 1: 341 – 354. 5. Karonga Prevention Trial Group (1996). Randomized controlled trial of single BCG, repeated BCG or combined BCG and killed Mycobacterium leprae vaccine for prevention of leprosy and tuberculosis in Malawi. Lancet 348: 17 – 24. 6. Buddle, B. M., Wedlock, D. N., Parlane, N. A., Corner, L. A., De Lisle, G. W., and Skinney, M. A. (2003). Revaccination of neonatal calves with Mycobacterium bovis BCG reduces the level of protection against bovine tuberculosis induced by a single vaccination. Infect. Immun. 71: 6411 – 6419. 7. Lowrie, D. B., et al. (1999). Therapy of tuberculosis in mice by DNA vaccination. Nature 400: 269 – 271. 8. Turner, J., Rhoades, E. R., Keen, M., Belisle, J. T., Frank, A. A., and Orme, I. M. (2000). Effective preexposure tuberculosis vaccines fail to protect when they are given in an immunotherapeutic mode. Infect. Immun. 68: 1706 – 1709. 9. Repique, C. J., Li, A., Collins, F. M., and Morris, S. L. (2002). DNA immunization in a mouse model of latent tuberculosis: effect of DNA vaccination on reactivation of disease and on reinfection with a secondary challenge. Infect. Immun. 70: 3318 – 3323. 10. Kapsenberg, M. L. (2003). Dendritic-cell control of pathogen-driven T-cell polarization. Nat. Rev. Immunol. 3: 984 – 993. 11. Reichardt, V. L., Brossart, P., and Kanz, L. (2004). Dendritic cells in vaccination therapies of human malignant disease. Blood Rev. 18: 235 – 243. 12. O’Neill, D. W., Adams, S., and Bhardwaj, N. (2004). Manipulating dendritic cell biology for the active immunotherapy of cancer. Blood 104: 2235 – 2246. 13. Moll, H. (2003). Dendritic cells as a tool to combat infectious diseases. Immunol. Lett. 85: 153 – 157. 14. Brinkman, J. A., Fausch, S. C., Weber, J. S., and Kast, W. M. (2004). Peptide-based vaccines for cancer immunotherapy. Expert Opin. Biol. Ther. 4: 181 – 198. 15. Basak, S. K., Kiertscher, S. M., Harui, A., and Roth, M. D. (2004). Modifying adenoviral vectors for use as gene-based cancer vaccines. Viral Immunol. 17: 182 – 196. 16. Timares, L., Douglas, J. T., Tillman, B. W., Krasnykh, V., and Curiel, D. T. (2004). Adenovirus-mediated gene delivery to dendritic cells. Methods Mol. Biol. 246: 139 – 154. 17. Gonzalez-Juarrero, M., Turner, J., Basaraba, R. J., Belisle, J. T., and Orme, I. M. (2002). Florid pulmonary inflammatory responses in mice vaccinated with antigen-85 pulsed dendritic cells and challenged by aerosol with Mycobacterium tuberculosis. Cell. Immunol. 220: 13 – 19. 18. McShane, H., Behboudi, S., Goonetilleke, N., Brookes, R., and Hill, A. V. S. (2002). Protective immunity against Mycobacterium tuberculosis induced by dendritic cells pulsed with both CD8+- and CD4+-T-cell epitopes from antigen 85A. Infect. Immun. 70: 1623 – 1626. 19. Wan, Y., Bramson, J., Carter, R., Graham, F., and Gauldie, J. (1997). Dendritic cells transduced with an adenoviral vector encoding a model tumor-associated antigen for tumor vaccination. Hum. Gene Ther. 8: 1355 – 1363. 20. Wan, Y., Bramson, J., Pilon, A., Zhu, Q., and Gauldie, J. (2000). Genetically modified dendritic cells prime autoreactive T cells through a pathway independent of CD40L and interleukin 12: implications for cancer vaccines. Cancer Res. 60: 3247 – 3253. 21. Chen, Y., et al. (2001). Induction of ErbB-2/neu-specific protective and therapeutic antitumor immunity using genetically modified dendritic cells: enhanced efficacy by cotransduction of gene encoding IL-12. Gene Ther. 8: 316 – 323.


ARTICLE

22. Lane, C., Leitch, J., Tan, X. H., Hadjati, J., Bramson, J. L., and Wan, Y. H. (2004). Vaccination-induced autoimmune vitiligo is a consequence of secondary trauma to the skin. Cancer Res. 64: 1509 – 1514. 23. Wang, J., et al. (2004). Single mucosal, but not parental, immunization with recombinant adenoviral-based vaccine provides potent protection from pulmonary tuberculosis. J. Immunol. 173: 6357 – 6365. 24. Santosuosso, M., Zhang, X., McCormick, S., Wang, J., Hitt, M., and Xing, Z. (2005). Mechanisms of mucosal and parenteral tuberculosis vaccinations: adenoviral based mucosal immunization preferentially elicits sustained accumulation of immune protective CD4 and CD8 T cells in the airway lumen. J. Immunol. 174: 7986 – 7994. 25. Wang, J., Santosuosso, M., Ngai, P., Zganiacz, A., and Xing, Z. (2004). Activation of CD8 T cells by mycobacterial vaccination protects against pulmonary tuberculosis in the absence of CD4 T cells. J. Immunol. 173: 4590 – 4597. 26. Yang, T., Dayball, K., Wan, Y.-H., and Bramson, J. L. (2003). Detailed analysis of the CD8+ T cell response following adenovirus vaccination. J. Virol. 77: 13407 – 13411. 27. Jellison, E. R., Kim, S. K., and Welsh, R. M. (2005). Cutting edge: MHC class II-restricted killing in vivo during viral infection. J. Immunol. 174: 614 – 618. 28. Schluger, N. W., and Rom, W. N. (1998). The host immune response to tuberculosis. Am. J. Respir. Crit. Care Med. 157(3 Pt. 1): 679 – 691. 29. Silva, C. L., and Lowrie, D. B. (2000). Identification and characterization of murine cytotoxic T cells that kill Mycobacterium tuberculosis. Infect. Immun. 68: 3269 – 3274. 30. Demangel, C., Bean, A. G., Martin, E., Feng, C. G., Kamath, A. T., and Britton, W. J. (1999). Protection against aerosol Mycobacterium tuberculosis infection using Mycobacterium bovis Bacillus Calmette Guerin-infected dendritic cells. Eur. J. Immunol. 29: 1972 – 1979. 31. Brossart, P., Goldrath, A. W., Butz, E. A., Martin, S., and Bevan, M. J. (1997). Virusmediated delivery of antigenic epitopes into dendritic cells as a means to induce CTL. J. Immunol. 158: 3270 – 3276. 32. Nakamura, Y., et al. (2003). Induction of protective immunity to Listeria monocytogenes with dendritic cells retrovirally transduced with a cytotoxic T lymphocyte epitope minigene. Infect. Immun. 71: 1748 – 1754. 33. Behboudi, S., Moore, A., Gilbert, S. C., Nicoll, C. L., and Hill, A. V. (2004). Dendritic cells infected by recombinant modified vaccinia virus Ankara retain immunogenicity in vivo despite in vitro dysfunction. Vaccine 22: 4326 – 4331. 34. Shaw, J., Grund, V., Durling, L., Crane, D., and Caldwell, H. D. (2002). Dendritic cells pulsed with a recombinant chlamydial major outer membrane protein antigen elicit a CD4(+) type 2 rather than type 1 immune response that is not protective. Infect. Immun. 70: 1097 – 1105. 35. Trinchieri, G. (2003). Interleukin-12 and the regulation of innate resistance and adaptive immunity. Nat. Rev. Immunol. 3: 133 – 146. 36. Flohe, S. B., Bauer, C., Flohe, S., and Moll, H. (1998). Antigen-pulsed epidermal Langerhans cells protect susceptible mice from infection with the intracellular parasite Leishmania major. Eur. J. Immunol. 28: 3800 – 3811. 37. Bacci, A., et al. (2002). Dendritic cells pulsed with fungal RNA induce protective immunity to Candida albicans in hematopoietic transplantation. J. Immunol. 168: 2904 – 2913. 38. Brown, K., et al. (2003). Adenovirus-transduced dendritic cells injected into skin or lymph node prime potent simian immunodeficiency virus-specific T cell immunity in monkeys. J. Immunol. 171: 6875 – 6882.

775