|Home | About | Journals | Submit | Contact Us | Français|
Human immunodeficiency virus type 1 (HIV-1) infection is characterized by the rapid onset of intestinal T-cell depletion that initiates the progression to AIDS. The induction of protective immunity in the intestinal mucosa therefore represents a potentially desirable feature of a preventive AIDS vaccine. In this study, we have evaluated the ability of an enteric adenovirus, recombinant adenovirus 41 (rAd41), to elicit intestinal and systemic immune responses by different immunization routes, alone or in combination with rAd5. rAd41 expressing HIV envelope (Env) protein induced cellular immune responses comparable to those of rAd5-based vectors after either a single intramuscular injection or a DNA prime/rAd boost. Oral priming with rAd41-Env followed by intramuscular boosting with rAd5-Env stimulated a more potent CD8+ T-cell response in the small intestine than the other immunization regimens. Furthermore, the direct injection of rAd41-Env into ileum together with intramuscular rAd5-Env boosting increased Env-specific cellular immunity markedly in mucosal as well as systemic compartments. These data demonstrate that heterologous rAd41 oral or ileal priming with rAd5 intramuscular boosting elicits enhanced intestinal mucosal cellular immunity and that oral or ileal vector delivery for primary immunization facilitates the generation of mucosal immunity.
Most infectious pathogens enter the body through mucosal surfaces. The mucosa therefore represents a first line of defense against infection. HIV in particular invades through the gastrointestinal, lower genital, and rectal mucosa (13). Several studies indicate that human immunodeficiency virus (HIV) depletes most of the CD4+ T cells in gut-associated lymphoid tissue in patients with AIDS. The virus targets CD4+ effector memory T cells bearing CCR5 (CD4+ CCR5+ TEM) in extralymphoid effector sites in the intestinal mucosa, as well as below and within the gastrointestinal tract epithelial layers. It infects TEM cells through a receptor and coreceptor and induces the apoptosis of target cells. This depletion of TEM occurs predominantly within the first few weeks after HIV infection (7, 20-22, 25, 26), so the early establishment of mucosal immunity, especially in the gut, is important for protection against HIV infection.
Adenovirus 41 (Ad41) is a human serotype F Ad that exhibits tropism for the gastrointestinal tract. It is associated with gastrointestinal disease: an estimated 2 to 6% of gastroenteritis cases are caused by this virus (2, 6, 10, 37). Ad41 possesses two distinct fibers, long and short, which are present in the virion in equal ratios (15). The long fiber, similarly to fibers of respiratory adenovirus Ad5, binds to the coxsackie and Ad receptor, but the function of the short fiber is unknown (31, 34). Unlike Ad5, Ad41 does not contain the integrin-binding RGD motif in its penton base protein for entry to host cells (1).
Recently, replication-deficient recombinant Ad41 (rAd41) vector expressing HIV envelope protein (Env) has been developed that stimulated HIV Env-specific humoral and cellular immune responses after prime and boost immunizations. Heterologous prime-boost with rAd41-rAd5 immunization induced significantly higher levels of cellular immune responses systemically than rAd5-only vector-based immunization through an intramuscular route of administration (18). In this study, we characterized a rAd41 vector expressing HIV Env for its ability to transduce different cell types in vitro and evaluate this vector in different vaccination regimens, specifically analyzing whether priming by an oral delivery can stimulate mucosal immunity without reducing the systemic response.
Six- to 8-week-old female BALB/c mice were purchased from NCI/DCT (Frederick, MD) and housed in the experimental animal facility of the Vaccine Research Center, National Institute of Allergy and Infectious Diseases (NIAID), NIH (Bethesda, MD). All animal experiments were reviewed and approved by the Animal Care and Use Committee, Vaccine Research Center, NIAID, NIH, and were performed in accordance with all relevant federal and NIH guidelines and regulations.
All cell lines were obtained from the American Type Culture Collection (Manassas, VA) and grown in the recommended media. The cells were plated in 96-well plates overnight and then transduced with rAd vectors encoding luciferase at the indicated titers for 1 h in medium containing 2% fetal bovine serum (FBS). The transduced cells were grown in fresh medium containing 10% FBS for 24 h after transduction and assayed using a luciferase assay kit (Promega Corporation, Madison, WI).
For Ad only, mice were immunized with 109 virus particles (VP) of rAd-Env by the intramuscular or oral route. For DNA prime/rAd boost immunizations, mice were primed with an intramuscular injection of 50 μg of DNA into the hind leg three times at 2-week intervals and boosted intramuscularly with 109 VP of rAd-Env 2 weeks later. For rAd prime/rAd intramuscular boost immunizations, 109 VP of rAd-Env was delivered either orally (by oral gavage) or intramuscularly, and animals received an intramuscular boost with 109 VP of rAd5-Env 3 weeks later. The immunized mice were sacrificed 3 weeks after the Ad-only immunization or 2 weeks after the boost immunization. Mice were fasted overnight before oral immunizations.
Following 16 h of fasting, mice were anesthetized using ketamine-xylazine (25 and 5 mg per kg of body weight, respectively) administered intramuscularly. The animals were carefully monitored and kept warm throughout the surgical procedure. Using an aseptic technique, a midline abdominal incision was made, and the ileum was readily identified by locating the cecum. Two small atraumatic serrated prewetted clamps (catalog no. 18055-02; Fine Science Tools) were placed at the ileum-cecum junction and 5 cm upstream of the cecum. VP (1010) in a total volume of 0.1 ml of rAd41 vector encoding HIV-1 gp140B (rAd41gp140B) was injected into the isolated ileum. Following 20 min of incubation, the clamps were released and the abdominal cavity was closed. The animals then received 0.05 mg/kg buprenorphine for discomfort following the surgery. Gene expression and immunogenicity were determined after the ileal injection of rAd5 encoding luciferase or rAd5 encoding HIV-1 gp140B, respectively, with and without clamping, and the optimal time for the vector to infect intestinal cells was determined. Transgene expression measured by luciferase activity was higher in the groups with clamps, and 20 min was sufficient to infect intestinal cells (data not shown), indicating that clamping facilitated the contact of virus with the intestinal cells by separating the vector from digestive contents and by distending the ileum.
Bone marrow (BM)-derived DC were obtained from the BM of BALB/c mice and cultured as previously described (14). More than 80% of these cells incubated in the presence of murine granulocyte-macrophage colony-stimulating factor for 1 week expressed DC surface markers CD11b and CD11c, as measured by flow cytometry. Lymphoid DC (CD8+ DC) and plasmacytoid DC (B220+ DC) were isolated from mouse spleens by magnetic cell sorting according to the manufacturer's instructions (Miltenyi Biotec, Auburn, CA). More than 90% of these purified cells expressed CD8 or B220, as measured by antibody staining of the cells.
Peripheral blood mononuclear cells (PBMC) were purified using Lympholyte (Cedarlane Laboratories Ltd., Burlington, Ontario, Canada) according to the manufacturer's instructions. Single cells from mouse spleen and mesenteric lymph nodes (MLN) were separated by mincing them using a nylon mesh screen. For the measurement of cellular immune responses in the small intestine, total lymphocytes (intraepithelial and lamina propria lymphocytes) from the jejunum were prepared at the same time by enzyme digestion. Briefly, Peyer's patches were removed and the small intestine was opened longitudinally. The intestine was flushed with phosphate-buffered saline (PBS) and cut into 2-mm-long pieces in an enzyme solution of RPMI 1640 containing 50 mg/100 ml collagenase II (Sigma, St. Louis, MO) and 10% FBS and then incubated at 37°C in a shaking incubator at 300 rpm for 30 min. Lymphocytes in the supernatant were purified by 40 and 75% Percoll gradient centrifugation at 1,200 × g for 15 min, and the purified lymphocytes were collected from above the 75% Percoll layer.
Lymphocytes or PBMC were stained with phycoerythrin-conjugated Dd/PA9 tetramer and then fluorescein isothiocyanate-conjugated anti-mouse CD3 monoclonal antibody (MAb) (clone 145-2C11; BD Pharmingen), peridinin chlorophyll protein-Cy5.5-conjugated anti-mouse CD8α MAb (clone 53-6.7; BD Pharmingen), and allophycocyanin-conjugated anti-mouse CD19 MAb (clone 6D5; Biolegend). The stained cells were examined by using a BD LSR-II (BD Pharmingen), and the data were analyzed by FlowJo software (Tree Star Inc.).
HIV gp140B-specific immunoglobulin G (IgG) and IgA were examined as follows: 96-well enzyme-linked immunosorbent assay (ELISA) plates were coated with 2 μg/ml recombinant HIV gp140 clade B protein (VRC2801), incubated at 4°C overnight, and blocked with PBS containing 1% bovine serum albumin at 37°C for an hour. Sera from the immunized mice were diluted by twofold serial dilutions, and the diluted sera were added. The plates then were incubated at 37°C for 2 h. Horseradish peroxidase (HRP)-conjugated anti-mouse IgG (Jackson ImmunoRseseach Laboratories, Inc., West Grove, PA) and IgA (Southern Biotech, Birmingham, AL) were added and incubated at 37°C for an hour. 3,3′,5′,5-Tetramethylbenzidine (TMB; Sigma) in HRP substrate was added to each well, and yellow color that developed after the addition of 0.5 M H2SO4 was measured at 450 nm.
Sera from immunized mice were inactivated by being heated at 56°C for an hour and diluted with RPMI 1640 medium with 2% FBS and then mixed with the indicated rAd vector encoding luciferase for an hour at room temperature. The neutralized virus was used to infect 293 cells (at a multiplicity of infection of 100) based on the number of VP per cell for 2 h, and then the medium was replaced with RPMI 1640 containing 10% FBS and incubated overnight at 37°C. The infected 293 cells were lysed, and the supernatants were mixed with luciferin. The luminescence was measured using a microplate scintillation and luminescence counter (PerkinElmer, Shelton, CT).
Results are expressed as means ± standard errors. Statistical analyses were performed upon comparisons made between the control groups and treated groups or between treated groups using Student's t test.
rAd41 vectors encoding luciferase were prepared, and their ability to transduce different cell types was analyzed in vitro (Fig. (Fig.1).1). The rAd41 vector transduced 293T cells with higher efficiency than an rAd5 vector. It also readily transduced human intestinal cell lines, FHs74 intestine and Caco-2 cells (Fig. (Fig.1A).1A). In addition to intestinal epithelial cells, its lymphoid tropism was analyzed on murine lymphoid and plasmacytoid DC isolated from a mouse spleen and myeloid DC derived from mouse BM. Although the transduction efficiencies of the latter cells were lower than that of the rAd5 vector, gene transfer was readily observed (Fig. (Fig.1B).1B). rAd41 therefore was able to deliver transgenes into multiple cell types of epithelial and lymphoid origins.
To assess whether rAd41 vectors can induce immune responses comparable to those of rAd5 vectors after a single intramuscular immunization, mice were immunized intramuscularly with rAd41-Env or rAd5-Env. The immunized mice were sacrificed 3 weeks later, and Env-specific CD8+ T cells were examined in PBMC using a tetramer to an immunodominant epitope in the V3 loop, H-2Dd/PA9, a response that corresponded to functional activation, since cells showed an increase in intracellular cytokine staining for interleukin-2, gamma interferon, and tumor necrosis factor alpha after stimulation with the same peptide in vitro (M. Honda, R. Wang, W. Kong, M. Kanekiyo, W. Akahata, L. Xu, K. Matsuo, K. Natarajan, H. Robinson, T. E. Asher, D. A. Price, D. C Douek, D. H. Margulies, and G. J. Nabel, submitted for publication). Representative flow plots confirmed tetramer-specific CD8 responses for each vector (Fig. (Fig.2A),2A), and rAd41 vaccine-immunized mice showed levels of Dd/PA9+ CD8+ T cells that were similar to those of rAd5 vaccine-immunized mice (Fig. (Fig.2B).2B). The antibody response also was measured by the detection of serum IgG ELISA titers. Significantly higher levels of HIV Env-specific IgG were detected in rAd41-Env-immunized animals than in controls, though titers were lower than those in rAd5 vaccine-immunized mice (Fig. (Fig.2C).2C). An examination of the IgG and IgA responses in vaginal washes and fecal extracts revealed minimally detectable responses (data not shown) despite the systemic effects, suggesting alternative mechanisms of stimulation of these disparate compartments. These results demonstrate that the rAd41 vector induced substantial antigen-specific cellular and humoral immune responses after a single intramuscular immunization.
Although rAd5 vectors are known to readily boost DNA-primed animals (32), the ability of rAd41 vectors to boost DNA-primed responses is not known. To address this question, mice were immunized intramuscularly with DNA encoding HIV Env three times at 2-week intervals and boosted intramuscularly or orally immunized with rAd41-Env or rAd5-Env 2 weeks later. Two weeks after boosting, immunized mice were sacrificed; HIV-1 Env-specific CD8+ T cells were examined in PBMC, spleen, MLN, and the small intestine; and HIV Env-specific IgG titers were measured in serum. Representative flow plots confirmed tetramer-specific responses in both spleen and intraepithelial intestinal lymphocytes (Fig. (Fig.3A).3A). In both systemic and intestinal tissues, mice boosted orally with rAd5-Env or rAd41-Env did not generate substantial Env-specific Dd/PA9+ CD8+ T-cell tetramer responses (Fig. (Fig.3A,3A, columns 5 and 6). However, mice that were boosted with rAd vectors intramuscularly showed substantially higher levels of cellular immune responses than the negative control group in both compartments (Fig. (Fig.3B,3B, compare lanes 3 and 4 to lane 1), and both rAd41 and rAd5 vaccine-boosted groups elicited significantly higher tetramer responses than single rAd5 vaccine-immunized mice in the spleen (Fig. (Fig.3B,3B, spleen panel, compare lanes 3 and 4 to lane 2). These results indicate that the rAd41 vaccine induced cellular immunity similar to that induced by rAd5 in both systemic and mucosal compartments after DNA prime/intramuscular rAd vaccine boost immunization. The IgG analysis revealed differences in the antibody response between the two vectors. rAd41 vectors stimulated a substantial increase in ELISA titers, and they were greater after intramuscular than after oral boosting (Fig. (Fig.3C,3C, compare lane 4 to lane 6), but rAd5 stimulated these responses more effectively (Fig. (Fig.3C,3C, compare lanes 3 and 5 to lanes 4 and 6). Interestingly, rAd41 and rAd5 vaccine-boosted mice using the oral route of administration produced significantly higher levels of IgG than the negative control mice (Fig. (Fig.3C,3C, compare lanes 5 and 6 to lane 1) despite the fact that the vaccinations did not induce high levels of cellular immune responses (Fig. (Fig.3B,3B, lanes 5 and 6). Taken together, these data indicate that following DNA priming, rAd41 and rAd5 vectors induce comparable cellular immune responses in both mucosal and systemic compartments and stimulate systemic antibody responses.
The low pH of the stomach represents a barrier for the introduction of rAd and other vaccine vectors for oral immunization. To determine whether different diluents could affect the efficacy of vaccination, we compared a high-pH solution with low buffering capacity, sodium bicarbonate, to a neutral-pH diluent with higher buffering capacity for their ability to stimulate rAd-mediated cellular and humoral immune responses. A single oral immunization did not elicit HIV Env-specific cellular immune responses in the tetramer assay (Fig. (Fig.4A),4A), so we boosted orally immunized mice with rAd5-Env intramuscularly. Mice primed orally with rAd41-Env in PBS produced higher cellular immune responses in the small intestine than those of bicarbonate, although no differences were noted in the spleen (Fig. (Fig.4B,4B, lanes 2 and 3). These findings suggest that PBS is the preferred diluent to prime for an rAd41-stimulated mucosal immune response.
We next compared rAd41 to rAd5 for its ability to prime an rAd5 vaccine boost. Intramuscular priming with rAd41-Env induced an increased antigen-specific CD8+ T-cell response in both systemic and mucosal compartments compared to that of rAd5-Env (Fig. (Fig.5A,5A, lanes 5 and 6). Notably, oral priming with rAd41-Env also stimulated a higher-magnitude response than rAd5-Env, more so in the small intestine than the spleen (Fig. (Fig.5A,5A, lanes 3 and 4). Similarly, intramuscular rAd41/rAd5 elicited the highest-titer ELISA responses, which were greater than those of intramuscular rAd5/rAd5 or by oral priming with either vector (Fig. (Fig.5B).5B). These results demonstrate that priming with rAd41 vectors induced CD8+ T-cell immunity in the small intestine by an oral route as well as stimulated systemic IgG responses to a greater extent than did intramuscular priming.
Since priming with rAd vectors may generate vector-specific antibodies that can affect the efficiency of the subsequent rAd boosts, we analyzed mouse sera for the presence of neutralizing antibodies after the administration of Ad vectors (Fig. (Fig.5C).5C). The intramuscular administration of rAd5 vectors elicited antibodies that specifically neutralized rAd5 vector transduction by more than 100-fold compared to that of nonimmune mouse sera (Fig. (Fig.5C,5C, left, compare bars 1 and 5). As expected, this sera did not neutralize a rAd41 reporter vector (Fig. (Fig.5C,5C, right, bar 1), consistently with the known designation of these viruses as distinct serotypes. Similarly, the intramuscular administration of rAd41 vectors elicited rAd41-specific neutralizing antibodies that did not neutralize rAd5 vectors (Fig. (Fig.5C,5C, left, bar 2). Notably, the oral administration of Ad5 did not elicit neutralizing antibodies to the homologous vector (Fig. (Fig.5C,5C, left, bar 3). Similarly, the oral administration of Ad41 did not stimulate the production of neutralizing antibodies to Ad41 (Fig. (Fig.5C,5C, right, bar 4). These results suggest that heterologous prime/boost vaccination regimens circumvent vector-induced immunity from priming and that oral immunization, while less potent than intramuscular priming, can prime for a secondary immunization without inducing systemic antivector humoral immunity.
To circumvent the degradation of vectors by low gastric pH or the presence of degradative enzymes in the stomach or small intestine, we directly injected rAd41-Env into the ileum surgically. Three weeks later, animals received a booster vaccination with an intramuscular rAd5 vector. The injection of rAd41-Env alone into the ileum induced a substantial Dd/PA9+ CD8 T-cell response in PBMC prior to boosting (Fig. (Fig.6A).6A). After the boost with intramuscular rAd5-Env, it induced remarkably high levels of CD8+ T cells in the spleen (Fig. (Fig.6B)6B) and especially high levels in the small intestine, presumably a mix of both intraepithelial and laminar propria lymphocytes (Fig. (Fig.6C).6C). These results indicate that the direct delivery of rAd41 into the lower small intestine effectively primes mucosal and systemic cellular immune responses.
Due to the rapid development of antibody neutralization escape mutations (29) or through its ability to evade neutralization because of glycan shielding (38), HIV has evaded attempts to control infection by eliciting a humoral immune response. The potential importance of CD8+ T cells in the control of HIV infection has been demonstrated in a variety of studies. Simian immunodeficiency virus (SIV) replication is markedly increased in monkeys when CD8+ lymphocytes are depleted (33), and SIV-specific central memory CD8+ T cells and linked CD4+ T cells are responsible for protection against disease progression in a monkey model (36). It is well known that replication-deficient virus-based vaccines predominantly induce cytotoxic T lymphocytes and CD4+ T-cell responses (24) and that rAd5-based vectors are among the most potent stimulators of systemic responses (35). Such vectors have been tested in intramuscular immunization alone and in DNA-rAd vaccine regimens in human clinical trials (8, 12, 27). Due to the high prevalence of neutralizing antibodies against Ad5, alternative serotypes and other vectors are being actively pursued.
Here, we evaluated a rAd41 vector to determine if it can prime for a rAd5 boost immunization and evaluated its ability to induce cellular immune responses in the gut by different routes of delivery. We have previously demonstrated that rAd41 prime/rAd5 boost was more effective than rAd5 prime/rAd41 boost in stimulating cellular immunity (18). In the present study, we examined the route of the administration of Ad41 vector as a priming agent. rAd41-Env induced similar cellular immune responses to rAd5-Env after both a single intramuscular immunization and DNA prime/Ad boost immunization, which suggests that rAd41 is a good substitute for rAd5 in systemic immunization for the induction of cellular immune responses. In in vitro studies, sera from intramuscular rAd5-immunized mice, but not orally immunized mice, displayed neutralizing activity against rAd5. Similar results were observed with a rAd41-based regimen with respect to Ad41 neutralization (Fig. (Fig.5C).5C). The effect of the neutralizing activity could be seen in the lack of a T-cell boost in the Ad5 homologous intramuscular prime-boost group (Fig. (Fig.5A).5A). However, this same regimen induced significantly higher humoral immune responses than single intramuscular rAd5 vaccine immunization (Fig. (Fig.5B).5B). Intramuscular administration with low-dose vector stimulated humoral immune responses to the transgene that could be boosted further despite the presence of anti-vector neutralizing antibodies in the homologous prime-boost immunization. These results are in agreement with those generated in a Dengue virus vaccine model (28).
Several approaches have been taken to elicit immune responses against HIV infection in mucosal compartments, including intestinal, vaginal, and rectal mucosa. Intranasal immunization with a rAd5 vaccine generated stronger IgA responses in systemic and mucosal compartments than intramuscular immunization in mice, but safety concerns may limit the use of this route of administration for rAd5 vaccines (19). Intrarectal immunization with peptide vaccines composed of HIV or SIV antigens with adjuvant can induce mucosal cytotoxic T lymphocytes more efficiently and control RNA levels better in mucosal and systemic compartments than subcutaneous immunization in macaques (4). Oral priming with enteric coated rAd5-based HIV vaccines followed by intranasal boosting with an envelope peptide cocktail with adjuvant induced HIV-specific cellular immune responses in the intestine of rhesus macaques (23). In addition, several studies have demonstrated that the mucosal administration of replication-competent Ad-based vaccines can generate immune responses against HIV antigens. Intranasal priming with replication-competent Ad4, Ad5, and/or Ad7 vaccines followed by intramuscular boosting with Env elicited mucosal immune responses and protection against HIV challenge in chimpanzees (30). Intranasal-oral or oral-oral priming with a replication-competent Ad5-HIV vaccine followed by a protein vaccine induced mucosal immunity in rhesus macaques (40). These studies demonstrate the possibility of generating mucosal and systemic immune responses to HIV antigens, but the potency and vaccination regimens need to be further optimized before advancing these vaccines into human clinical trials.
Novel vectors that possess natural mucosal tropism may have advantages over these reported vectors in terms of administration, safety, and vaccine potency. When antigen or vector is administered by an oral route, it typically is inefficient compared to the efficiency of injection, often requiring 100- to 1,000-fold higher levels of protein or virus due to loss by degradation or mucosal clearance. The actual delivered dose therefore was likely much less than the injected amount. This finding suggests that antigen delivery is relatively efficient, possibly due in part to the gut tropism of Ad41. In the present study, oral rAd41-Env priming induced the highest HIV Env-specific CD8+ T-cell responses, as determined by a tetramer response in the small intestine. Sera from orally immunized animals did not block the entry of the homologous serotype vectors in neutralization assays, indicating that oral rAd immunization induces either weak or no neutralizing antibodies against the administered rAd vector. These findings are consistent with studies of a different route of delivery, intranasal inoculation, which also does not elicit Ad5 neutralizing antibody (39). It is widely accepted that the systemic immunization of protein- or peptide-based vaccine induces only systemic immune responses, whereas mucosal immunization induces mucosal as well as systemic immune responses (3, 16). In this study, heterologous oral rAd41 prime-intramuscular rAd5 boost induced the highest levels of antigen-specific tetramer-positive CD8+ T-cell responses in the small intestine, whereas heterologous intramuscular rAd41-intramuscular rAd5 boost induced the highest systemic response. The increased cellular immune responses by heterologous intramuscular rAd41 priming-intramuscular rAd5 boosting may be related not only to the evasion of preexisting immunity to rAd41 but also to intrinsic characteristics of the virus vector. Possibly, antigen-presenting cells activated by intramuscular rAd41 vaccination educate immune cells, including T cells, in the draining lymph nodes and direct these cells to the mucosal compartment in the gut. The microenvironments of various secondary lymphatic tissues are very different: antigen-presenting cells of the same phenotype are able to respond differently to the antigen based upon where the immune responses are initiated. For example, CD8α− DC present orally delivered antigens in MLN, while CD8α+ DC in the spleen present intravenously delivered antigens (11). The types of antigens, in addition to immune cells, also affect the nature of the reaction. CD8α+ DC mediate antiviral immunity to viral infections by subcutaneous or intravenous infections, whereas CD8α− DC are activated against protein antigen with adjuvant delivered intranasally (5, 17). rAd41 vector vaccines delivered to the intestinal mucosa likely utilize such mechanisms to induce mucosal cellular immune responses in the gut in combination with Ad5 boosting and represent a potential approach to elicit protective immune responses to HIV. While direct injection into the ileum represents an experimental tool to demonstrate this effect, the development of appropriate formulations such as enteric coatings or nanoparticles could allow the vector to avoid the degradative environment of the upper gastrointestinal tract and be amenable to clinical applications.
We thank Ati Tislerics for assistance with manuscript preparation, Brenda Hartman for figure preparation, members of the Nabel laboratory for helpful discussions and advice, and Srinivas Rao and colleagues for help with anesthetic and surgical techniques.
This work was supported by the Intramural Research Program of the National Institutes of Health, Vaccine Research Center, NIAID, and by the Bill and Melinda Gates Foundation.
Published ahead of print on 5 November 2008.