Background
Colorectal cancer (CRC) is the fourth leading cause of cancer related death worldwide [
1]. In most cases, the disease occurs sporadically but can also be preceded by inflammatory bowel disease or a familial genetic predisposition, for example, familial adenomatous polyposis (FAP) [
2,
3]. As with other cancers, the immune system plays a large role in the control of disease. Work by Galon et al has shown that a large infiltrating population of CD8
+ T cells predicts improved patient outcome [
4]. More mixed results have been obtained regarding other immune cell populations in tumour tissue and their relation to patient outcome [
5‐
8].
The gut is an immunosuppressive environment and control of local disease may also be altered. For example, in many cancers an increased frequency of infiltrating regulatory T cells (Tregs) correlates with poor outcome but in colorectal cancer, the infiltrate of Tregs has been associated with both good and poor outcomes [
7]. It is therefore important to consider this altered environment and to perform observations at this distinct site when considering tissue targeted immunotherapies. Orthotopic models of colorectal cancer have been described [
9‐
14], but we wished to investigate the ability to modulate the
systemic immune response to prevent local tumour growth in the gut.
Previously, we showed that vaccination with chitosan gel could generate a population of CD8
+ memory T cells in both peripheral and gut-associated lymphoid sites [
15]. Furthermore, vaccination with chitosan gel also provided protection in a subcutaneous tumour challenge model, both prophylactically and therapeutically [
15]. However due to the specialised nature of the gut immune system the ability of systemic immunisation to protect against a gut tumour is unknown. We showed that vaccination with chitosan gel was protective in an intracaecal mouse model of cancer, while vaccination with DCs was not. The gel-mediated protection was associated with an increase in antigen specific T cells and T cells producing IFN-γ.
Methods
Mice
C57BL/6 and OT-I transgenic mice were obtained from the HTRU (University of Otago, Dunedin, NZ) and were bred and housed under specific pathogen free conditions. All experimental procedures were approved by the University of Otago Animal Ethics Committee. No changes in weight or general health (positive or negative) were observed in mice following tumour injection with or without vaccination.
Cell lines
B16-OVA and B16-luc cell lines were cultured in complete RPMI (with 100 μg/mL Penicillin, 100 μg/mL Streptomycin, 55 μM 2-mercaptoethanol, (all from Invitrogen, Carlsbad, USA), 5 % fetal calf serum (PAA laboratories, Morningside, QLD, AU) at 37 °C, 5 % CO2. B16-OVA cells were grown in 5 % complete RPMI with the addition of geneticin (Invitrogen) at 500 μg/mL to prevent loss of ovalbumin protein expression.
Chitosan (1 % w/v) (Sigma-Aldrich-Aldrich, St Louis, MO, USA) and methylcellulose (0.5 % w/v) (Sigma-Aldrich) were added to 0.05 mol/L hydrochloric acid (VWR, Radnor, PA, USA) and stirred at 4 °C overnight. Glycerol 2-phosphate disodium hydrate (Sigma-Aldrich) was added drop wise to give the solution thermosensitive properties and stirred for a further hour. Ovalbumin protein (OVA; Sigma-Aldrich) and Quil-A (QA; Brenntag Biosector, Denmark) were added for a final concentration of 100 μg/mL and 200 μg/mL, respectively, then stirred for a further 30 min to ensure uniform distribution in solution.
Bone marrow derived dendritic cell generation
Bone marrow harvested from the leg bones of naïve C57BL/6 mice was cultured for 7 days in complete RPMI and 20 ng/mL GM-CSF (MyBiosource, San Diego, CA, USA) as described [
15]. For OVA vaccinations, cells were pulsed with OVA protein overnight on day 5 at a final concentration of 200 μg/mL followed by 1 μg/mL of lipopolysaccharide (Sigma-Aldrich) overnight on day 6. Dendritic cells were MHCII
+ and CD80
hi.
Vaccination and adoptive transfer of cells
Two hundred microlitres of chitosan hydrogel containing both OVA and QA (Gel + OVA) or QA alone (Gel), or 2 × 10
5 OVA-pulsed DCs (DC + OVA) in 200 μl phosphate buffered saline (PBS; NaCl – Biolab, Australia, 137 mM; KCl – VWR, 2.7 mM; Na
2HPO
4 – VWR, 4.3 mM; KH
2PO
4 – Merck, 1.4 mM), were injected subcutaneously into the flank of C57BL/6 mice. 2 × 10
5 naïve OT-I lymphocytes [
16] were injected intravenously in 200 μl of PBS into the tail vein of mice at the time of vaccination.
Subcutaneous and intracaecal tumour challenge
Mice were injected with 2 × 10
5 B16-OVA or B16-
luc cells subcutaneously in the flank in 100 μL PBS 30 days following vaccination. Surgery for intracaecal injection was carried out according to Tseng et al [
17]. Briefly, mice were anaesthetised one at a time using a combination of ketamine, domitor and atropine injected subcutaneously. Pre-operative carporfen was also administered subcutaneously. The abdomen of the mouse was shaved and oil was applied to the eyes. An incision was made through the skin and peritoneum. The caecum was externalised and 25 μL PBS containing 1×10
5 cells was injected under a 10x surgical microscope into the subserosa of the caecum. The caecum was repositioned in the abdominal cavity and peritoneum and skin were resutured separately. The mouse was injected with amphoprim antibiotic and antisedan to reverse the anaesthetic effect. Mice were monitored twice daily for 5 days to assess recovery and carporfen and amphoprim were administered twice daily for 2–3 days. All drugs and anaesthetics were distributed by the Animal Welfare Office, University of Otago.
Bioluminescence
Mice were injected intraperitoneally with 200 μL of luciferin (Pure Science, Porirua, New Zealand) and placed in the induction chamber of an isofluorane based gas anaesthetic device. At 5 min following administration of luciferin, mice were x-rayed for 30 s and imaged to detect bioluminescence for 5 min. Bruker MI (Bruker, Billerica, MA, USA) and ImageJ (NIH, Bethesda, MD, USA) were used for image capture and analysis.
Flow cytometry
Single cell suspensions were resuspended in 1 mL PBS and incubated with titrated concentrations of Live/Dead Fixable Red Dead Cell Stain (Invitrogen) for 30 min at 4 °C in the dark. Samples were washed with FACS buffer (PBS + 0.5 % Fetal Calf Serum and 0.01 % NaN3 – VWR, Radnor, PA, USA) and incubated with titrated concentrations of the following fluorescently labelled anti-mouse antibodies for 10 min at 4 °C in the dark: CD8-PerCPCy5.5 (53–6.7), CD45.1-BV421 (A20), CD122-PE (5H4; all from BioLegend, San Diego, CA, USA), CD4-APCH7 (GK1.5), CD19-APCH7 (1D3), Vα2-APC (B20.1), Vβ5-FITC (MR9-4), CD44-V500 (1 M7; all from BD Biosciences, Franklin Lakes, NJ, USA). Following this, samples were washed in FACS buffer, fixed in 1 % paraformaldehyde (Sigma-Aldrich) for 30 min then resuspended in FACS buffer for acquisition on an LSR Fortessa (BD Biosciences) and analysed using FlowJo (Treestar, Ashland, OR, USA). For cytokine detection cells were restimulated with phorbol-12-myristate-13-acetate (PMA) and ionomycin, and brefeldin-A (all from Sigma-Aldrich) was added to samples 2 h before harvesting. Samples were incubated with IFN-γ-PE (XMG1.2; BioLegend) antibodies in permeabilisation buffer for 30 min in the dark at 4 °C and washed three times in permeabilisation buffer.
Statistical analysis
GraphPad Prism (GraphPad, La Jolla, CA, USA) was used for all graphs and statistical analysis. Significance was calculated using one-way ANOVA and a Tukey post-hoc test as indicated in figure legends.
Discussion
Colorectal cancer (CRC) is the third most common cancer in men and the second in women worldwide [
1]. In order to evaluate new immune therapies in a preclinical setting, appropriate animal models are required. Here we have described an intra-caecal model of colorectal cancer in mice and shown that immune modulation of the systemic response can result in tumour protection associated with antigen specific T cells producing IFN-γ, consistent with the human disease [
4,
22].
Interestingly we also demonstrated that biological location has an effect on the makeup of immune infiltrate into a tumour. It is well known that in many solid tumours infiltration of high numbers of T cells correlates with improved prognosis [
4] and it has been reported that both B and T cells are important in human colorectal cancer [
23]. Interestingly, in this study we found a difference in CD3+ T cell frequencies between intra-caecal and subcutaneous tumours, but no difference in the frequencies of CD4+ or CD8+ T cells. There are many subsets of T cells that may be reflected in this result, including pro-inflammatory “pathogenic” T cells [
24], naturally occurring T cells that have down-regulated co-receptor expression [
20,
25], NKT cells and MAIT cells [
19], γδ T cells [
26], and T cells in the context of immune deficiency [
27]; all of which may be involved in anti-tumour immune responses, especially in the gut. Based on these results, further experiments should include an expanded panel of molecules to determine which of these CD3
+ populations may have a significant effect on the local tumour immune response.
For mucosal immunity, priming of T cells by DCs in gut associated lymphoid tissues preferentially up-regulates mucosal homing receptors on those cells; hence the rationale for testing DC-OVA as a vaccine control. In the Peyer’s patches and mesenteric lymph nodes, CD103
+ DCs cause up-regulation of CCR9 and α4β7 on lymphocytes, thus allowing them to migrate to gut mucosa [
28‐
30]. Mucosal vaccines typically take advantage of this mechanism through administration at sites allowing antigen presentation to occur in gut associated lymphoid tissue. However, there are difficulties in developing oral peptide and protein based vaccines in terms of delivery of antigen in an immunogenic form to gut-associated lymphoid tissue. In the intra-caecal mouse model of cancer, the sustained release vaccine, but not a DC vaccine, provided protection. This protection was associated with an increased frequency of tumour specific T cells at peripheral lymphoid but not gut-associated lymphoid tissues. However, the absolute number of cells present in Peyer’s patches and mesenteric lymph nodes was low and it is possible that a later time-point is needed to detect differences. The frequency of tumour specific memory CD8
+CD122
+CD44
+ T cells was significantly higher in mice vaccinated with the sustained release vaccine in peripheral lymphoid sites sampled consistent with our previous data prior to tumour challenge [
15]. Here, we show that this increase is maintained after tumour challenge and is likely to be a correlate of protection. Further we show that the frequency of cytokine producing (IFN-γ) CD8
+ and CD4
+ T cells detected in the spleen was higher in Gel + OVA vaccinated mice than others, implying an IFNγ-mediated anti-tumour effect.
It is unclear if the tumours generated in this model are susceptible to systemic immunity or if gut-associated immune protection is required. Vaccination with the sustained release gel generated a population of antigen-specific CD8
+ T cells in both Peyer’s patches and mesenteric lymph nodes, two gut-associated lymphoid sites, but this vaccine also stimulated stronger systemic responses than did the DC vaccine [
15]. Therefore it is also possible that protection was conferred through effector memory T cells. While not mucosa specific, these cells are able to track through multiple tissues [
31]. An effective effector memory T cell (T
em) population generated through vaccination with a sustained release gel may provide protection in multiple different biological locations through the body.
The magnitude of the immune response generated by the DC vaccine was in general smaller as compared to the gel vaccine, similar to subcutaneous vaccination [
15]. However the increased production of IFN-γ by OVA-specific CD8
+ T cells in mice vaccinated with the gel compared to those vaccinated with the DC vaccine indicated that these cells could also have a functional difference. In preliminary experiments we also showed an increased frequency of IL-2
+ OVA-specific CD8
+ memory T cells in mice vaccinated with the gel vaccine compared to other vaccine groups (data not shown). It is possible that these cells may also have an increased capacity for proliferation and survival. This would allow for development of a more robust secondary recall population of CD8
+ memory T cells upon tumour challenge.
These results also address a potential concern with the use of sustained release vaccines, namely that sustained release of vaccine may result in immune exhaustion rather than the development of effector or memory populations [
32]. It is likely that the kinetics of release as well as the strength of the immune stimulation will impact on the type of response generated therefore caution should be taken when developing sustained release formulations in order to ensure that such parameters are optimised.
Interestingly the gel itself, in the absence of antigen, induced an increase in the frequency and number of antigen specific IFN-γ producing T cells in the spleens of mice challenged with OVA expressing tumour cells. Chitosan has been reported to have adjuvant activity [
33,
34] and the gel, while not containing any antigen, was also loaded with the potent Th1 adjuvant Quil A [
35]. The gel would have created an inflammatory depot that may have non-specifically boosted immune reactivity, much the same way the CpG have been used as a therapy to boost anti-tumor immunity [
36]. Destruction of the tumour and release of OVA (a foreign antigen) by dying tumours would then have led to the development of some degree of antigen specific immunity. Mice immunised with the gel only did show some slight evidence of anti-tumor immunity. However as with the CpG studies, inclusion of a tumour antigen with the therapy increases the efficacy of the response [
36].
The B16 cell line used for intracaecal challenge is a melanoma cell line. While this allowed for direct comparison to subcutaneous B16 challenge ref, in this mouse model of colorectal cancer it may be beneficial to use a mouse colorectal cancer cell line. Further experiments using the CT26 colorectal cancer cell line expressing an endogenous peptide [
37,
38] or a MC38 colorectal cancer cell line expressing OVA [
39] may provide more insight into the disease by providing more accuracy to the model system.
Acknowledgements
We thank David Botman and Michelle Wilson for technical assistance.