The gastrointestinal mucosa contains an intact immune system that protects the host from pathogens and communicates with the systemic immune system. Absorptive epithelial cells in the mucosa give rise to malignant tumors although the interaction between tumor cells and the mucosal immune system is not well defined. The pathophysiology of colorectal cancer has been elucidated through studies of hereditary syndromes, such as familial adenomatous polyposis, a cancer predisposition syndrome caused by germline mutations in the adenomatous polyposis coli tumor suppressor gene. Patients with FAP develop adenomas and inevitably progress to invasive carcinomas by the age of 40. To better delineate the role of mucosal immunity in colorectal cancer, we evaluated the efficacy of intrarectal recombinant vaccinia virus expressing the human carcinoembryonic Ag (CEA) in a murine FAP model in which mice are predisposed to colorectal cancer and also express human CEA in the gut. Mucosal vaccination reduced the incidence of spontaneous adenomas and completely prevented progression to invasive carcinoma. The therapeutic effects were associated with induction of mucosal CEA-specific IgA Ab titers and CD8+ CTLs. Mucosal vaccination was also associated with an increase in systemic CEA-specific IgG Ab titers, CD4+ and CD8+ T cell responses and resulted in growth inhibition of s.c. implanted CEA-expressing tumors suggesting communication between mucosal and systemic immune compartments. Thus, intrarectal vaccination induces mucosal and systemic antitumor immunity and prevents progression of spontaneous colorectal cancer. These results have implications for the prevention of colorectal cancer in high-risk individuals.
The mucous membranes covering the gastrointestinal tract provide a chemical and mechanical barrier between the external environment and the internal milieu of host organisms. The mucosa also contains a highly specialized innate and adaptive immune system that can protect against potentially harmful pathogens while promoting tolerance against possibly beneficial matter, such as food particles. The cells of the mucosa-associated lymphoid tissues continuously monitor the influx of Ags across the mucosal barrier and function independently from the systemic immune apparatus (1, 2). Ags taken up by absorptive or specialized (M) epithelial cells in the mucosa are often captured by professional APCs or directly shuttled to inductive sites for priming of αβ CD4+ and CD8+ T cells. Recent studies have highlighted the ability of Peyer’s patches and mesenteric lymph node dendritic cells (DC)3 to present Ag to cognate T cells and also induce selective T cell homing back to the gut through up-regulation of α4β7 and CCR9 receptors (3, 4). This compartmentalization has suggested that mucosal vaccination might be useful for preventing infection with pathogens that use the mucous membranes for colonization or host entry (5, 6, 7). Gut epithelial cells are also a frequent source of malignant transformation and, because the initial encounter with tumor Ags occurs within the mucosal compartment, mucosal vaccination represents a logical strategy for immunization.
Colorectal cancer is the second leading cause of cancer-related mortality in the Western world (8). The identification of specific germline mutations in DNA repair genes and Wnt signaling pathway inhibitors in hereditary forms of colorectal cancer has resulted in an improved understanding of colorectal cancer progression and provided novel targets for genetic testing and chemoprevention. Among the familial colorectal cancer syndromes, familial adenomatous polyposis (FAP) is perhaps the best characterized and is caused by germline mutations and/or polymorphisms in the adenomatous polyposis coli (Apc) gene. FAP is characterized by the appearance of numerous adenomatous polyps at a young age which inevitably progress to invasive carcinomas by the age of 40 (9). We have previously reported a mouse model of FAP generated by inserting an expression cassette at codon 1638 of the murine Apc gene (Apc1638N) resulting in an unstable truncated protein (10, 11). Apc1638N mice exhibit aberrant crypt foci at 2 mo and develop adenomatous polyps that progress to invasive carcinomas at 6–8 mo, providing a relevant model of spontaneous colorectal cancer for prevention studies.
Carcinoembryonic Ag (CEA) is a 180 kD protein that belongs to the Ig gene superfamily. Although CEA is expressed at low levels in normal adult colonic epithelium, it is highly overexpressed in malignant colorectal tumor cells (12). In addition, tumor-associated CEA differs from normal CEA in that it is aberrantly glycosylated, loses its apical localization, and is actively secreted by phospholipases resulting in high circulating serum levels in cancer patients. Increased CEA levels have been associated with increased tumor volumes and poor prognosis suggesting that metastatic colorectal cancer patients have developed peripheral tolerance to CEA (13, 14, 15). This concept is further supported by evidence of clonal T cell inactivation after exposure to tissue-specific self Ags in the absence of costimulatory signals (16, 17).
The safety and ability to generate CEA-specific T cell immune responses in colorectal patients treated with recombinant poxviruses expressing CEA has been reported but therapeutic responses have generally been limited (18, 19, 20, 21, 22, 23). These results are not unexpected given that vaccination was performed in patients with advanced disease where tolerance to CEA has likely already been established. Furthermore, these trials uniformly used s.c. or i.m. routes of immunization and did not explore the effects of such vaccinations on local or mucosal immune responses. Thus, we sought to determine whether early mucosal immunization could prevent colorectal cancers from developing by promoting local immune surveillance.
To test a preventative vaccine strategy, a spontaneous colorectal cancer murine model was needed because transplantable systems do not use the actual microenvironment in which tumors evolve. Thus, Apc1638N mice were crossed with a human CEA transgenic mouse to generate hybrid Apc1638N/CEA mice (11). Similar to Apc1638N mice, hybrid mice show signs of polyp formation at 2 mo and the adenomas progress to invasive carcinomas by 8 mo with a corresponding increase in serum CEA levels, as observed in human colorectal cancer. Herein, we show that early intrarectal immunization with recombinant vaccinia virus expressing CEA decreases adenoma formation, prevents progression to invasive colorectal cancer and induces local and systemic CEA-specific humoral and cell-mediated immunity.
Materials and Methods
Female C57BL/6 mice were purchased from Charles River Laboratories. Apc1638N/CEA hybrid mice were generated as previously described (11). All animals were 6–8 wk old when vaccination was initiated. Animals were maintained in filter top cages with access to food and water in the Association for Assessment of Laboratory Animal Care-accredited Animal Institute of Columbia University Medical Center.
Cell lines and recombinant viruses
The murine colon adenocarcinoma cell line MC38 was provided by Dr. N. Restifo (National Cancer Institute, Bethesda, MD). The CEA-expressing cell line MC32 was generated by transducing the MC38 cell line with full-length human CEA cDNA using retroviral expression vector pBNC (24). CEA expression was confirmed by flow cytometry using a mouse anti-human CEA Ab (clone COL-1, provided by Dr. J. Schlom, National Cancer Institute, Bethesda, MD), and a fluorescein-labeled goat anti-mouse IgG. CEA was expressed on >90% of the MC32 cells, while no expression was detected on MC38 cells. The BSC-1 cell line, derived from African green monkey kidney, and the human HeLa cell line were used for expansion of virus. All cell lines were cultured in complete DMEM containing 10% heat-inactivated FBS, 2 mM l-glutamine, penicillin (100 U/ml), and streptomycin (100 μg/ml), 1 mM nonessential amino acids, and 5 × 10−5 M 2-ME (all from Life Technologies).
rV-CEA is a recombinant vaccinia virus (strain Western Reserve) expressing the full-length human CEA gene produced by homologous recombination of the pSC-11 plasmid, containing the human CEA gene, into the thymidine kinase gene of the Western Reserve vaccinia virus, and control rV-LacZ expressing LacZ (25). Wild type vaccinia virus (strain Western Reserve) was purchased from American Type Culture Collection. Viral expansion and titering were performed as previously described (25).
DCs were prepared from bone marrow cells of C57BL/6 mice. In brief, cells from femurs and tibiae were enriched by plastic adherence and differentiated into DCs by culture in complete RPMI 1640 medium supplemented with recombinant murine GM-CSF and rmIL-4 (PeproTech) both at 25 ng/ml for 7 days. Media was replaced with fresh complete RPMI 1640 containing the same supplements every other day. To mature DCs, LPS (1 μg/ml) was added for 24 h. To characterize DCs, cells were stained with fluorescent-conjugated mAb (all from BD Pharmingen) for class I, class II, CD11c, CD80, CD83, and CD86, and then analyzed on a FACSCalibur flow cytometer (BD Biosciences) and FlowJo software (Tree Star).
Vaccination of mice
Three immunizations were given 14 days apart through intrarectal (IR) route. Animals were inoculated with 107 PFU of either rV-CEA or rV-LacZ in a total volume of 100 μl. For IR inoculation, animals were instigated to defecate, and immunization was performed using a 1.5 mm diameter silicon catheter introduced 1 to 3 cm in the rectum. Animals were monitored for 5 min, and animals that evacuated or expelled any liquid through the anus during this time were excluded. Five to 7 days after the last booster, the CEA and vaccinia virus-specific immune responses were evaluated in the vaccinated mice (n = 12). The remaining animals (n = 10 each) were randomly divided and received one or more s.c. injections of either 2 × 105 MC32 or MC38 tumor cells. Tumors were measured by calipers in two dimensions every second day. All mice were sacrificed by CO2 asphyxia once tumors reached 300 mm2 or on day 22 after tumor implantation to examine CEA- and vaccinia virus-specific immune responses. To determine the effect of vaccination on tumor development in Apc1638N/CEA mice, IR vaccination was performed as described above at 2 mo, and tumor growth was evaluated 2, 4, 6, 8, and 10 mo after the last vaccination by counting the number of neoplastic lesions in the gastrointestinal organs using a dissecting microscope (11).
Histology and CEA immunohistochemistry
The histology and CEA immunohistochemistry of the gastrointestinal organs of the Apc1638N/CEA hybrid mice were performed as previously described (11). Individual organs were removed, fixed in formalin and examined for neoplastic lesions under a dissecting microscope. The number, location, and macroscopic features of the tumors, including shape and ulcer formation, were recorded. Tissues were embedded in paraffin, and sections were stained with H&E. Each sample was examined microscopically by a pathologist in a blinded fashion and tumors classified according to the WHO criteria. To examine CEA expression, paraffin-embedded tissues were deparaffinized and stained with an anti-CEA Ab (clone: COL-1) or control mouse serum using the Dako Peroxidase kit (DakoCytomation).
Determination of Ab titers
Serum Ab against CEA or vaccinia was evaluated using ELISA. Serum samples and rectal washes were collected after the third vaccination, both before and ∼22 days after tumor implantation, and analyzed for the presence of Abs to CEA or vaccinia virus. In brief, 96-well plates were coated overnight at 4°C with 200 ng/well purified human CEA (International Enzymes) or 107 PFU of wild type vaccinia lysate. Wells were blocked with PBS containing 5% BSA overnight. Mouse serum or rectal wash samples were diluted in PBS containing 0.1% BSA (1/5 to 1/1000) and incubated in the wells for 1–2 h. Plates were washed, and Abs bound to the wells were detected by incubation with HRP-conjugated sheep anti-mouse IgG (1/3000) or IgA (1/1500) for 1 h followed by addition of the HRP substrate, O-phenylenediamine (Amersham Biosciences). Absorbance was measured using an ELISA microplate reader at A450 nm (Bio-Rad). All samples were analyzed in triplicate.
Detection of CEA-specific T cell responses
Lymphocytes were isolated from spleen, inguinal lymph nodes (LN), and intestinal Peyer’s patches. Single cell suspensions were obtained from spleen and LN by mechanical dispersion and filtration through 0.45-μm nylon cell strainers. RBC were removed in ACK lysis buffer (0.15 M NH4Cl, 10 mM KHCO3, 0.1 mM Na2EDTA). The Peyer’s patches were carefully excised from the intestinal wall without disrupting the mucosal layer, mechanically disrupted, and incubated with 300 U/ml Collagenase VIII (Sigma-Aldrich) for 1 h at 37°C. The single cell suspension was filtered through a 0.45-μm nylon cell strainer, and lymphocytes were isolated using Ficoll gradient.
To evaluate CEA-specific T cell proliferation, splenocytes (2 × 105 cells/well) were labeled with 2.5 μM CFSE in 0.5% BSA in PBS according to manufacturer’s recommendations (Molecular Probes). CFSE-labeled cells were then incubated for 3 days in the presence or absence of CEA protein (2 μg/ml) (Sigma-Aldrich), OVA as a control (2 μg/ml), or lysates containing apoptotic bodies of MC38, MC32, or a CEA Kb-restricted peptide CAP-M8 (CEA526–533), synthesized at the Albert Einstein College of Medicine (Bronx, NY), and CFSE dilution was analyzed by flow cytometry with cell surface staining of CD4 and CD8.
In vitro T cell cytotoxicity assay
Cytotoxicity activity against CEA and vaccinia virus following immunization was evaluated using splenocytes and Peyer’s patches lymphocytes in a 6 h 51Cr-release assay. To analyze the immune response against vaccinia, MC38 cells were infected with wild-type vaccinia viruses (MOI = 1 to 25) for 18 h before the assay. Two million target cells (MC32, MC38, or vaccinia-infected MC38) were labeled with 200 μCi of Na251CrO4 (Amersham Biosciences) for 1 h at 37°C and washed. Effector and target cells were combined at various E:T ratios in 96-well U-bottom plates in triplicates. The plates were centrifuged at 400 rpm for 5 min to initiate cell contact and were incubated at 37°C for 6 h. The plates were centrifuged at 400 rpm for 5 min again. The supernatants were collected, and radioactivity was read in a gamma scintillation counter (Wallac). Spontaneous release of 51Cr was determined using target cells without effectors, and maximum release was determined by adding 2.5% Triton X-100 to target cells. The assay was repeated three times with similar results. The percentage of specific lysis was calculated as the mean ± SE of triplicate wells according to the following formula: Specific lysis = Experimental cpm spontaneous cpm/spontaneous cpm × 100.
In vivo cytotoxicity assay
Splenocytes from syngeneic C56/BL6 mice were pulsed with MC32 (CEA+) or MC38 (CEA−) cell lysates for 2 h at 37°C. After washing, target and control cells (107/ml in PBS) were incubated with 5.0 μM or 0.5 μM carboxyfluorescein diacetate succinimidyl ester (Molecular Probes), respectively, at 37°C for 10 min followed by 5 min incubation on ice to stop the reaction. The cells were washed again and then mixed in a 1:1 ratio and resuspended in PBS. An aliquot of 107 cells in 0.2 ml PBS were injected i.v. by tail vein injection into mice. Spleen, LN, and PP were collected after 18 h and incubated for 10 min in RBC lysis buffer at 4°C. Cells were analyzed for CFSE expression by flow cytometry using a FACSCalibur (BD Biosciences). Cytotoxicity in naive controls was set to 0% and the specific cytolytic activity in vivo was calculated as (% CFSEhigh cells/% CFSElow cells) × 100.
For simultaneous detection of IFN-γ and CD107a surface expression, cells were stimulated with appropriate Ags as above in the presence of FITC-conjugated CD107a mAb for 24 h followed by CD8 expression. Cells were then fixed with 2% paraformaldehyde (eBioscience) and permeabilized for intracytoplasmic cytokine staining. Cells were incubated for 30 min at room temperature with PE-conjugated IFN-γ mAb in permeabilization buffer. Cells were washed two times with fixation buffer and data were collected on 10,000 CD8+ T cells on an FACSCalibur (BD Biosciences) and analyzed using the FlowJo software (Tree Star). The frequency of IFN-γ producing cells is depicted as percentage of lymphocyte subpopulations. Cells stimulated with PMA (50 ng/ml) plus ionomycin A (Sigma-Aldrich, 0.5 μg/ml) for 4 h was used as positive control. Isotype Ab control was included in all experiments.
Five million irradiated (2400 rad) syngeneic C56/BL6 splenocytes were incubated in 24-well plates in the presence of appropriate Ags as described above for 30 min at 37°C, 5% CO2. CD4+ T cells were sorted from splenocytes in vaccinated mice and added to the culture to a final volume of 2 ml and incubated for 24 h. Supernatants were collected from cultures at the indicated time points and the concentration of TNF-α, IFN-γ, IL-2, and IL-4 were simultaneously quantitated using a Mouse Th1/Th2 Cytokine Cytometric Bead Array Kit (BD Biosciences) and IL-10 level was determined by Quantikine ELISA kit according to manufacturer’s protocols.
All results are expressed as mean ± SD. Statistical significance of the Ab titers and T cell functional data (i.e., proliferation, cytotoxicity, intracellular cytokine production) was determined using Student’s two tailed t test. All p values are two-sided and p < 0.05 was considered significant.
IR immunization protects against transplantable colorectal cancer
To determine whether IR mucosal immunization using rV-CEA inhibits tumor growth, a standard transplantable tumor model was evaluated first. Seven days after a series of three IR immunizations with either rV-CEA or rV-LacZ, mice were inoculated s.c. with 2 × 105 MC32 (CEA+) or MC38 (CEA−) tumor cells, and tumor growth was measured over 22 days (Fig. 1). IR immunization with rV-CEA significantly inhibited the growth of CEA-expressing MC32 tumors compared with control mice immunized with rV-LacZ (Fig. 1,A, p < 0.05). By contrast, rV-CEA did not inhibit the growth of CEA-negative MC38 tumors (Fig. 1 B). Thus, prophylactic IR immunization with rV-CEA induced a significant delay in tumor development. Notably, IR immunization was safe and no mice suffered from adverse effects related to vaccination, such as perforation, diarrhea, anorexia, or development of autoimmunity (data not shown). In subsequent repeat experiments, we also immunized a group of mice with rV-CEA (n = 5) or rV-LacZ (n = 5) by i.v. injection and found that rV-CEA could inhibit growth of CEA-bearing MC32 tumors to a similar degree as IR vaccination (Supplemental Fig. 1).4
IR immunization decreases spontaneous polyp formation and prevents invasive colorectal cancers
To see whether IR vaccination could prevent colorectal cancer, we used a hybrid transgenic mouse model wherein mice express human CEA in the gastrointestinal tract and develop spontaneous adenomatous polyps that progress to invasive carcinomas due to a truncated Apc gene (Apc1638N/CEA mice). Mice were sacrificed at various time points to determine the presence of benign or invasive tumors along the entire GI tract from stomach to rectum by a blinded pathologist. In control animals, routine H&E staining demonstrated multiple adenomas characterized by epithelial proliferation disrupting the glandular pattern, which could be detected by 2 mo of age (Fig. 2, A and B). Invasion of the basal layer was identified in adenocarcinomas, which appeared between 6 and 8 mo of age (Fig. 2, C and D). Immunostaining of a representative tumor shows the majority of cells strongly express CEA (Fig. 2 E). These histologic patterns suggest that the mice exhibit a similar polyp-to-carcinoma sequence as observed in human colorectal cancer.
To determine whether IR vaccination could prevent polyps or colorectal cancer progression, Apc1638N/CEA transgenic mice were vaccinated by the IR route with 107 PFU of rV-CEA, rV-LacZ (virus control) or PBS (injection control) every 2 wk. Mice were sacrificed every 2 mo after the third vaccination; the number of adenomas and adenocarcinomas was determined in a blinded manner by a single pathologist. In mice treated with PBS or LacZ, the incidence of adenomas and adenocarcinomas increased over time and by 8–10 mo of age nearly all mice had evidence of adenomas and adenocarcinomas (Fig. 3). Mice vaccinated with rV-CEA had a significantly lower incidence of adenomas (p < 0.01, Fig. 3,A) and progression to adenocarcinoma was completely blocked (p < 0.001, Fig. 3 B). These data demonstrate that IR immunization with rV-CEA is effective in decreasing the incidence of spontaneous polyps and completely prevents progression of polyps to invasive carcinomas.
IR immunization induces CEA-specific CD8+ T cell responses in the systemic and mucosal immune compartments
Mice were vaccinated by IR immunization as described as above. Splenocytes and Peyer’s patch T cells were collected 5 days after the last vaccination and CEA-specific cytotoxic T cell responses were analyzed in vitro by 51Cr-release assay using CEA-expressing (MC32) and nonexpressing (MC38) tumor cells as targets. Immunization with rV-CEA induced CEA-specific CTL responses in both spleen and Peyer’s patches (Fig. 4,A, top panels). Interestingly, the magnitude of the CTL response was higher in spleen-derived T cells compared with those from Peyer’s patches. Lymphocytes isolated from mice receiving rV-LacZ showed no recognition of either the MC32 or MC38 tumor cell lines. Mice treated with the rV-LacZ vaccine did, however, demonstrate robust CTL activity against wild-type vaccinia-infected MC38 cells in both spleen and Peyer’s patches (Fig. 4 A, bottom panels). Thus, IR immunization with rV-CEA induces CEA-specific CD8+ CTL responses in both the systemic and local compartments.
To determine whether IR vaccination induced effective CTL in vivo, cytotoxicity was also determined by adoptive transfer of CFSE-labeled target cells. Mice were vaccinated three times with 107 PFU of rV-CEA or rV-LacZ by IR immunization. Syngeneic splenocytes pulsed with either MC32 or MC38 cell lysates were labeled with CFSE at two different concentrations and adoptively transferred to rV-CEA-vaccinated mice at 1:1 ratio. Spleens were harvested after 18 h and the degree of in vivo cell killing was analyzed by flow cytometry for CFSE+ cells (Fig. 4 B). Mice vaccinated with rV-CEA showed that 30.1 ± 11.5% of the CFSE-labeled target splenocytes were eliminated within 18 h after infusion while mice vaccinated with rV-LacZ indicated 2.56 ± 7.5% were eliminated (p < 0.05). Similar results are shown in LN with elimination of 36.4 ± 10.3% of target cells. These data demonstrate IR rV-CEA vaccination induced functional CEA-specific CTLs in vivo.
To determine whether CD8+ T cells could recognize Ag through cross priming, CD8+ T cells derived from spleen were exposed to autologous DCs pulsed with apoptotic MC32 tumor cell lysates, CEA protein, or OVA as an Ag control. CD8+ T cells from mice vaccinated with rV-CEA recognized apoptotic MC32 cells and CEA protein as demonstrated by robust proliferation with a minimum of five divisions, whereas mice receiving rV-LacZ or PBS did not proliferate (Fig. 4 C). This suggested that CD8+ T cells in vaccinated mice could recognize Ag through cross presentation.
To further explore the nature of the CD8+ T cell response generated by IR vaccination, the defined Kb-restricted murine CEA epitope (CEA526–533) was tested using CD8+ T cells isolated from splenocytes of mice vaccinated with rV-LacZ or rV-CEA. Vaccination with rV-CEA, but not rV-LacZ, induced specific CD8+ T cell proliferation upon exposure to peptide-pulsed autologous DC (Fig. 4,C, 75% vs 21%, p < 0.05). We further characterized effector CD8+ T cell function by evaluating concomitant expression of CD107a and IFN-γ upon in vitro stimulation with either autologous DC pulsed with apoptotic MC32 tumor cell lysate or CEA peptide for 24 h (Fig. 4,D). CD8+ T cells derived from spleen of rV-CEA vaccinated mice showed Ag-specific mobilization of CD107a and produced IFN-γ upon stimulation demonstrated as double positive T cells for CD107a and IFN-γ. In contrast, there were fewer CD8+ T cells from rV-LacZ immunized mice that were double positive (Fig. 4,D). We also identified a subset of CD8+ T cells in rV-CEA-vaccinated mice that produced IFN-γ but lacked CD107a expression (Fig. 4 D, left upper quadrant of far right panels). In mice that received IV vaccinations, although tumor growth inhibition was observed, we did not detect a significant increase in CD107a and IFN-γ CD8+ T cells (Supplemental Fig. 2) although IV vaccination with rV-CEA did promote CEA-specific proliferative responses (Supplemental Fig. 3). These data demonstrate that rV-CEA IR vaccination induces CD8+ T cells with cytotoxic potential and effector cell functional capacity in vitro and in vivo within both the mucosal and systemic immune compartments. Further, mucosal immunization induces a functionally distinct population of CD8+ T cells compared with IV vaccination with the same vector.
IR immunization induces local and systemic humoral responses
The induction of anti-vaccinia and anti-CEA Ab responses in the sera was evaluated by standard ELISA. Comparable vaccinia-specific titers were detected in the sera of mice immunized with rV-LacZ and rV-CEA, indicating a similar response to the vaccinia vector (Fig. 5,A). In contrast, mice immunized with rV-CEA had much higher serum CEA-specific titers compared with mice immunized with the control rV-LacZ (p < 0.001, Fig. 5,B). To determine whether local Ab responses were generated rectal washed were collected and subjected to ELISA for both IgG and IgA titers. Interestingly, the mice immunized with rV-CEA had CEA-specific IgG titers and IgA Ab titers that could be detected in the rectal wash (Fig. 5 C). Thus, IR vaccination with rV-CEA leads to the generation of IgG and IgA Ab titers within the mucosal compartment and a corresponding induction of IgG-specific Ab titers within the systemic compartment.
IR immunization induces systemic CD4+ helper T cell responses
Because IR immunization resulted in elevated IgG titers in the systemic compartment, we sought to determine whether CD4+ helper T cells could be identified in the systemic T cell pool. Splenocytes were derived from mice immunized with rV-CEA or rV-LacZ and evaluated in a T cell proliferation assay using CEA protein or MC32 cell lysates (Fig. 6,A). To determine the degree of cell proliferation, splenocytes were CFSE labeled and cultured in the presence of tumor Ags for 72 h and analyzed for CFSE dilution on CD4+ T cells by flow cytometry (Fig. 6 A, histograms). CEA protein and MC32 lysate-stimulated CD4+ T cells derived from rV-CEA vaccinated mice went through at least four divisions in 3 days while less than one division was observed in CD4+ T cells derived from rV-LacZ vaccinated mice (p < 0.05). The response was specific to CEA because OVA failed to stimulate lymphocytes from the same mice.
Next, we determined the cytokine production profiles from CD4+ T cells derived from vaccinated mice and stimulated with CEA protein pulsed with allogeneic DC for 24 h (Fig. 6 B). We detected the highest level of Th1 cytokines, IL-2, IFN-γ, and TNF-α in rV-CEA vaccinated mice, whereas Th2 cytokines IL-4 and IL-10 were barely detectable. Cytokines were not detected when CD4+ T cells were stimulated in a similar manner using the hgp100209–217 peptide (data not shown). These data indicated that IR rV-CEA vaccination resulted in Th1 dominant immune responses associated with secretion of IL-2, IFN-γ, and TNF-α.
Colorectal cancer arises from epithelial cells within the mucosal layers of the colon and rectum. In this report, we have demonstrated that IR immunization using rV-CEA results in mucosal and systemic anti-CEA humoral and CD8+ T cell responses. These responses also correlated with a decrease in benign spontaneous adenomas and vaccination completely prevented progression to invasive carcinomas in a genetic mouse model of FAP. The observation that IR vaccination was more potent in halting tumor progression rather than blocking polyp formation suggests that clinical application of tumor vaccines might be more effective in the prevention setting, such as in patients with hereditary familial polyposis syndromes or in the adjuvant setting following surgical resection of high-risk colorectal cancers. Nonetheless, IR vaccination was also effective at delaying growth of tumors implanted into the systemic compartment (Fig. 1), suggesting the potential of this approach to impact metastatic disease as well.
The microenvironment in which T cells are primed is important in directing both the activation and homing of cells. Previous studies have shown that T cells activated by gut-associated DCs express unique sets of homing and chemokine receptors that allow them to traffic back to selected mucosal tissues wherein they exert their effector functions. Thus, while DCs from spleen, peripheral LNs, and Peyer’s patches can equally activate CD8+ T cells, only DCs derived from Peyer’s patches or mesenteric LNs can induce expression of α4β7 and CCR9 on CD8+ T cells (3, 4, 26, 27). Trafficking is further regulated by complementary mucosal tissue-specific receptors (addressins) on vascular endothelial cells and chemokines in the mucous membranes. T cells expressing α4β7 interact with MAdCAM-1 expressed on intestinal endothelium to help gain entry to the mucosal microenvironment. Similarly, the CCR9 ligand, CCL25, is abundantly expressed by epithelial cells and endothelial cells of small bowel and helps promote accumulation of activated T cells within the gut mucosa following activation (28, 29). In our studies, we found that IR immunization induced the local generation of CEA-specific CTLs within the Peyer’s patches. This is consistent with previous reports demonstrating that IR immunization induces a higher percentage of functionally active CD8+ T cells and higher avidity CTL in the gut-associated lymphoid tissues (30, 31). We have also demonstrated systemic induction of CEA-specific CD4+ Th1 cells by cytokine production, cytotoxic CD8+ T cell activation by Ag-specific in vivo cytotoxicity assay, and simultaneous detection of degranulation of CTLs by surface staining of CD107a and IFN-γ production. Together, these data suggest that IR vaccination can effectively break peripheral CEA-specific T cell tolerance within the mucosal and systemic compartments.
Although vaccination through the mucosal route has been explored, especially in the development of oral vaccines, IR vaccination may offer a distinct advantage because the IR route avoids Ag degradation in the stomach and the potential induction of oral tolerance (32, 33). Although we directly compared IR and IV vaccination (Supplemental Fig. 1) and found similar antitumor effects, we also observed an increase in CD107a+IFN-γ+ CD8+ T cells in mice vaccinated through the IR route, suggesting that IR vaccination might alter the functional capacity of the effector cells induced by vaccination with recombinant vaccinia vectors. Furthermore, because colorectal cancer develops from mucosal epithelial cells, the mucosal compartment may be especially important in early decisions related to immune protection or tolerance. Thus, while the mechanisms of peripheral tolerance induction during colorectal cancer progression are not clearly defined, early vaccination within the mucosal compartment could influence the subsequent immune response as tumors develop. The large commensal pathogen load within the large intestine also suggests that the mucosa-associated lymphoid tissues system may be better equipped to induce immune responses in the lower GI tract compared with sites within the upper GI tract (7, 34, 35). We observed an increase in CEA-specific T cell responses Peyer’s patches as well as CEA-specific IgA titers in rectal washes of vaccinated mice, consistent with other reports of mucosal vaccines (36). This suggests that IR immunization induces potent local humoral and cellular response that might be particularly important during the early progression of colorectal tumors.
In addition to local immunity, we also observed significant induction of systemic CEA-specific humoral and cell-mediated immunity following IR vaccination as evidenced by an increase in CEA-reactive CD8+ CTLs, CD4+ helper T cells, and CEA-specific Ab titers. The induction of such responses has been seen in previous preclinical models using recombinant poxviruses delivered through systemic routes, including i.v., s.c., and i.m. immunization protocols (20, 37, 38, 39, 40, 41, 42). Furthermore, clinical trials of recombinant poxviruses coexpressing CEA and T cell costimulatory molecules in patients with advanced colorectal and pancreatic cancer have shown a correlation between CEA-specific CD8+ T cell responses in peripheral blood and objective clinical responses in small subsets of patients (43, 44). The low frequency of therapeutic responses observed in these studies have been ascribed to the advanced nature of the disease in eligible patients, the poor immunogenicity of the Ag and vaccine, and the presence of complex regulatory suppressive mechanism in patients with established tumors (45, 46, 47, 48, 49, 50, 51, 52). Although we did not directly test the effects of systemic vaccination on spontaneously arising colorectal cancers in this study, previous studies of a recombinant vaccinia virus expressing the HIV gp160 Ag suggested that IR immunization induced trafficking of gp160-specific T cells into the systemic compartment but i.v. vaccination did not influence mucosal gp160-specific immunity (53). Thus, the use of the IR route of administration offers significant benefits for induction of both local and systemic immune responses.
IR vaccination has been previously examined in both infectious disease and tumor models demonstrating the feasibility of the approach and the ability to generate local and systemic Ag-specific immunity (30, 31, 53, 54, 55, 56). This report, however, is the first to evaluate this approach in a spontaneous tumor model wherein the tumor cells are derived from the mucosal compartment. The ability to completely block tumor progression, coupled with the appearance of local and systemic humoral and cellular immune responses, suggests that IR vaccination approaches might be especially interesting for the prevention of colorectal cancer. These results also provide support for evaluating mucosal vaccination in patients with hereditary colorectal cancer syndromes, such as FAP, where the genetic predisposition is established and prevention strategies are needed to avoid certain progression to cancer or major prophylactic surgical procedures.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by NIH K08 CA79881 (to H.L.K.).
Abbreviations used in this paper: DC, dendritic cell; FAP, familial adenomatous polyposis; Apc, adenomatous polyposis coli; CEA, carcinoembryonic Ag; IR, intrarectal; LN, lymph node; rv-CEA, recombinant vaccinia virus expressing CEA.
The online version of this article contains supplementary material.