Reduction in the resident intestinal myelomonocytic cell population occurs during ApcMin/+ mouse intestinal tumorigenesis.

With its significant contribution to cancer mortality globally, advanced colorectal cancer (CRC) requires new treatment strategies. However, despite recent good results for mismatch repair (MMR)-deficient CRC and other malignancies, such as melanoma, the vast majority of MMR-proficient CRCs are resistant to checkpoint inhibitor (CKI) therapy. MMR-proficient CRCs commonly develop from precursor adenomas with enhanced Wnt-signalling due to adenomatous polyposis coli (APC) mutations. In melanomas with enhanced Wnt signalling due to stabilized β-catenin, immune anergy and resistance to CKI therapy has been observed, which is dependent on micro-environmental myelomonocytic (MM) cell depletion in melanoma models. However, MM populations of colorectal adenomas or CRC have not been studied. To characterize resident intestinal MM cell populations during the early stages of tumorigenesis, the present study utilized the ApcMin/+ mouse as a model of MMR-proficient CRC, using enhanced green fluorescent protein (EGFP) expression in the mouse lysozyme (M-lys) lys-EGFP/+ mouse as a pan-myelomonocytic cell marker and a panel of murine macrophage surface markers. Total intestinal lamina propria mononuclear cell (LPMNC) numbers significantly decreased with age (2.32±1.39×107 [n=4] at 33 days of age vs. 1.06±0.24×107 [n=8] at 109 days of age) during intestinal adenoma development in ApcMin/+ mice (P=0.05; unpaired Student's t-test), but not in wild-type littermates (P=0.35). Decreased total LPMNC numbers were associated with atrophy of intestinal lymphoid follicles and the absence of MM/lymphoid cell aggregates in ApcMin/+ mouse intestine, but not spleen, compared with wild-type mice. Furthermore, during the early stage of intestinal adenoma development, there was a two-fold reduction of M-lys expressing cells (P=0.05) and four-fold reduction of ER-HR3 (macrophage sub-set) expressing cells (P=0.05; two tailed Mann-Whitney U test) in mice with reduced total intestinal LPMNCs (n=3). Further studies are necessary to determine the relevance of these findings to immune-surveillance of colorectal adenomas or MMR-proficient CRC CKI therapy resistance. Copyright: © Faluyi et al.

Introduction

Colorectal cancer (CRC), which accounts for nearly a million global deaths each year, remains a major cause of cancer mortality due to the limited efficacy of currently available systemic treatment for advanced disease (1). Consequently, improved understanding of the disease is required to further optimize systemic treatment strategies including immunotherapy. CRC is known to develop from intestinal epithelium following progressive accumulation of genetic alterations, which include mutations of the adenomatous polyposis coli (APC) gene (2). In the highly evolutionary conserved canonical Wnt signalling pathway, APC is known to target β-catenin for cytoplasmic degradation, thus preventing its nuclear translocation to promote tumorigenesis (3). Most sporadic CRCs are thought to acquire APC mutations as an early event during tumorigenesis, prior to the development of adenomas (4). Furthermore, patients with familial adenomatous polyposis (FAP) carry a germ-line APC mutation, which predisposes the individual to intestinal adenomas and CRC (5). The Apc mouse is a model of FAP that possess a germline heterozygous Δ850 Apc mutation (6). Such mice develop predominantly small intestinal adenomas and die at ~130 days of age from the intestinal adenoma burden (6). In addition, haematopoietic defects, including the development of generalized atrophy of lymphoid tissue, occur during the early stages of intestinal tumorigenesis at ~80 days of age (7), while myeloid defects have been reported at an advanced age in the Apc mouse and another mouse model that is haplo-insufficient for Apc (8,9).

Over the past decade, immune checkpoint inhibitor (CKI) therapy including targeting programmed cell death protein-1 (PD-1), has emerged as an effective therapeutic strategy against several types of cancer, such as lung, melanoma and renal cell cancer (10). More recently, CKI therapy has been shown to be effective for the mismatch repair (MMR)-deficient or microsatellite instability (MSI)-high subset of various malignancies, including CRC (11,12). However, this strategy has proven ineffective so far in the management of the majority of MMR-proficient CRC that represent >90% of sporadic CRCs (13). For melanoma, higher neoantigen burden (14) and a greater extent of T lymphocyte infiltration (15) are correlated with enhanced responses to PD-1 inhibition. However, T lymphocyte and MM cell infiltration have been inversely correlated with enhanced β-catenin pathway signalling in melanoma (16). Furthermore, a study in autochthonous mouse melanoma models with constitutive β-catenin signalling has demonstrated the dependence of T lymphocyte infiltration on the MM cell population (16). Therefore, characterization of the intestinal MM cell population during intestinal adenoma development in the Apc model could yield insight into any early MM population changes associated with enhanced Wnt signalling.

A review on MM cells highlighted their heterogeneity, with no pan-MM cell marker defined, which has compounded previous studies on MM cell populations (17). However, a mouse with EGFP knocked into the mouse-lysozyme (M-lys) locus by homologous recombination (M-lyslys-EGFP) was previously generated, which utilizes EGFP expression to facilitate studies on murine MM cells (18). In this model, EGFP fluorescence has been observed in multiple surface marker-defined peripheral MM sub-populations (18). The present study had two aims. First, to determine if there is reduction of the total resident intestinal LPMNC population during the early stage of intestinal tumorigenesis. Furthermore, utilizing Apc and wild-type mice bred onto the M-Lys background, it was investigated if there is a reduction in the resident intestinal MM cell population during the early stage of intestinal tumorigenesis.

Materials and methods

Mice

C57BL/6J and C57BL/6J-Apc mice were obtained from The Jackson Laboratory. C57BL/6J/Sv129-M-lys mice (18) were obtained from Albert Einstein College of Medicine. All mice were bred in-house, kept under isolator conditions at temperatures between 19 and 23°C on a 12-h light-dark cycle. They were pathogen-free by regular bacteriological and serological testing. Relative humidity was kept between 45 and 55%. The mice were fed on mouse complete maintenance diet with free access to food and water. For the experiments described here, the following mice were utilized: Apc (n=20), Apc (n=23), Apc (n=9) And Apc (n=12).

Mouse breeding and genotyping

Apc and Apc mice on the M-Lys background were obtained by mating male Apcs mice with female Apcs mice. Furthermore, Apc and Apcmice on the M-Lys background were obtained by mating male Apcs mice with female ApcM-lys lys-EGFP/lys-EGFP mice. The offspring were genotyped for Apc and M-lys at ~30 days of age by PCR analysis of genomic DNA (18,19).

Resident peritoneal cell, splenic and intestinal tissue collection

Each mouse was sacrificed between 30 and 138 days of age for experiments described below by cervical dislocation, immediately after which the peritoneal cavity was opened and peritoneal mononuclear cells (when required) were obtained by lavage with sterile phosphate buffer saline (PBS) at room temperature (25°C), prior to dissection of the spleen or whole intestine. Splenic and intestinal tissues were collected in ice cold PBS, then stored in ice cold PBS for ≤5 min until processed as described in subsequent sections.

Intestinal adenoma count

Following lavage for peritoneal cells and dissection of the spleen, intestine was removed from the pylorus to the anus. The small intestine was then separated from the caecum and colon. Intestines were flushed out gently with PBS until no luminal content remained, after which they were carefully cut and opened out longitudinally to avoid adenoma disruption. For five Apc at 92±23 days of age, adenomas were counted by naked eye examination of small intestinal tissue.

Histology

Tissue from the small intestine and spleen of three pairs of Apc and Apc mice at 93±23 days of age underwent histological evaluation and immunohistochemistry for EGFP localisation as described below. Tissue from an SW480 human CRC xenograft transfected with an EGFP-expressing herpes saimiri viral vector served as positive control due to previously demonstrated EGFP expression (20). Tissue from age matched Apcs mice served as negative control. Sections were fixed in 4% (w/v) paraformaldehyde in PBS for 6 h at 25°C and paraffin wax embedded. Sections were cut to 5–7-µm thickness and underwent haematoxylin and eosin staining or immunohistochemistry (as described in the next section). Sections were viewed using fluorescence and phase contrast microscopy by a NIKON Eclipse E1000 fluorescence microscope (Nikon Corporation). Images were then captured using LUCIA GF imaging software (version 4.60) (Nikon Corporation).

Immunohistochemistry for the detection of EGFP

Steps of the procedure were performed at room temperature (25°C) except otherwise stated. Sections were de-waxed progressively in xylene for 1 min each three times, then absolute (100%) ethanol for a minute each (×3), then washed in distilled water before endogenous tissue peroxidase activity was blocked by immersion of slides in 2% (v/v) H2O2 in absolute methanol for 15 min. Subsequently, slides were washed in distilled water for 10 min. Immunohistochemical staining of intestinal sections was carried out as previously described (21). Serum block was with 1.5% (v/v) goat serum (Dako; Agilent Technologies, Inc.) in PBS for 30 min. Sections were then incubated with rabbit anti-Aquorea anti-EGFP primary antibody (1:4,000; cat. no. A-6455; Invitrogen; Thermo Fisher Scientific, Inc.) for 20 h at 4°C, after which slides were washed in PBS four times for 5 min each. Sections were then incubated for 30 min with HRP/dextran polymer-conjugated goat anti-rabbit secondary antibody (ready to use; cat. no. K4002; Dako; Agilent Technologies, Inc.) at room temperature. Subsequently, sections were washed in PBS four times for 5 min each. Sections were then incubated for 10 min with 0.1% (v/v) diaminobenzidine solution [Dako; Agilent Technologies, Inc.) in Tris-buffer (0.05 M Tris (pH 7.6 with HCl)], containing 0.03% (v/v) H2O2 at room temperature. Sections were washed in tap water four times for 5 min each, counterstained with Mayer's Haematoxylin (cat no. MHS32; Sigma-Aldrich; Merck KGaA) for 90 sec at room temperature, followed by immersion in Scott's tap water (0.167 M and MgS04 and 0.042 M NaHCO3 in distilled water) for 1 min. Slides were then rinsed in distilled water and the stained sections were dehydrated by immersion in absolute ethanol for 3 min three times, then immersed in xylene for three times 3 min each followed by mounting in DPX mountant (cat. no. 100579; Millipore Sigma). Sections were, viewed and images were captured as aforementioned.

Isolation and enumeration of small intestinal lamina propria mononuclear cells

Utilizing Apc mice as controls, intestinal LPMNCs were isolated from Apc mice at the following ages: Weaning (~30 days of age; n=4 pairs), prior to the appearance of macroscopically visible adenomas (~70 days of age; n=8 pairs), prior to death from macroscopic adenoma burden (~100 days of age; n=8 pairs). LPMNCs were isolated from mouse small intestine at room temperature (25°C) with viability and numbers determined as previously described (22,23) ensuring complete enzymatic digestion of the intestines of mice at different ages.

Flow cytometric analysis of small intestinal lamina propria mononuclear cells

Flow cytometric analysis of small intestinal LPMNCs has been described previously for EGFP, phycoerythrin (PE) and propidium iodide (PI) expression (23). Flow cytometry was performed at room temperature (25°C) on intestinal LPMNCs of six Apc mice and seven Apcmice at 74±2 days of age. Monoclonal primary antibodies utilized were as follows: F4/80 (1:100; cat. no. MCA497G; clone A3-1; BioRad Laboratories, Inc.), BMDM-1, ER-HR3 (1:10), ER-MP23, ER-MP58, ER-TR9, MOMA-1 (all 1:10) and MOMA-2. Antibodies (with the exclusion of F4/80) were hybridoma-conditioned supernatant, a kind gift from Professor Pieter Leenen (24). A PE-conjugated goat anti-rat antibody was utilised (cat no. 305009; Bio-Rad Laboratories, Inc.) at 1: 100 dilution. Cells were then washed in 10% foetal calf serum (FCS) labelled with propidium iodide (cat. no. P4864; Sigma-Aldrich; Merck KGaA) at 1:2000 dilution. Flow cytometry was performed using 5×104 LPMNCs and a FACS Vantage cytometer (Becton-Dickinson and Company) with Cell Quest™ software version 3.3 (Becton-Dickinson and Company). The R1 gate was set up to exclude PI-positive cells and cells with low forward scatter (deemed non-viable). EGFP-positive cells (fluorescence level greater than that obtained by <0.1% of wild-type LPMNCs) and macrophage marker positive cells ([M] with fluorescence level greater than that obtained by <0.1% of cells with non-specific labelling by control rat IgG) were analysed as expressing the relevant marker. Some cells were labelled with a cocktail of ER-HR3, F4/80 and MOMA-1 primary antibodies defined as the ‘Mac-mix’(at the same concentration as utilized for single markers). BMDM-1, ER-MP23, ER-MP58, ER-TR9 and MOMA-2 were not evaluated further, due to the level of expression of undiluted hybridoma supernatant being no higher than for control rat IgG (data not shown).

Isolation of resident peritoneal cells

Lavaged peritoneal cells from three pairs of Apc and Apc mice at 93±23 days were treated with 5 mM EDTA in Hank's Balanced Salt Solution (Invitrogen; Thermo Fisher Scientific, Inc.) for 75 min and collagenase/dispase enzymes for 90 min at 37°C as for small intestinal LPMNCs (23). Peritoneal cells were then washed twice in an excess of 10% (v/v) foetal calf serum in PBS and re-suspended at 25°C in the same culture medium as for LPMNCs prior to been viewed by fluorescence microscopy.

Statistical analysis

The mean ± standard error of mean (SEM) was calculated for numbers or proportions of cells belonging to different small intestinal LPMNC sub-populations of different groups of mice. Analysis of the difference between total LPMNCs of the oldest and youngest mice of either genotype (parametric data) was compared using unpaired two-tailed Student's t-tests. Statistical comparison of the myelomonocytic sub-population and adenoma numbers between mice was conducted using two-tailed Mann-Whitney U tests. Statistical analysis utilized Minitab software version 13 (Minitab, LLC). P≤0.05 was considered to indicate a statistically significant difference.

Results

Small intestinal LPMNC numbers decrease with age in ApcMin/+ mice

Initially, the total small intestinal LPMNC population was characterised during intestinal tumorigenesis in the Apc mouse. It was determined if there were any differences in small intestinal LPMNC numbers with age. There were no significant differences in LPMNC viability between groups of mice (Table I). There was no significant difference in total intestinal LPMNC numbers with age in Apc mice (33±1 vs. 109±2 days old; P=0.35), though older mice had a trend to reduced total LPMNCs (Table I). By contrast, total intestinal LPMNCs numbers significantly reduced with age in Apc mice (33±1 vs. 109±2 days old; P=0.05; Table I). This suggested significantly reduced intestinal LPMNCs with age in Apc mice.

Table I.

Decreased numbers of small intestinal LPMNCs with age in Apc mice.

	Mouse strain
	Apc+/+		ApcMin/+
Age, (days)	LPMNCs, ×106	Viability, %	LPMNCs, ×106	Viability, %
33±1	25.0±6.0	81.5±2.1	23.2±13.9	81.3±4.3
71±1	17.1±2.6	78.2±1.9	12.3±2.4	77.1±2.4
109±2	18.7±3.3[a]	76.8±1.6	10.6±2.4[b]	76.1±1.5

P=0.35 vs. Apc mice at 33±1 days of age

P=0.05 vs. Apc mice at 33±1 days of age.

No significant effect of M-lys hemizygosity on ApcMin/+ mouse small intestinal tumorigenesis

A previous study reported no difference in the proportion of EGFP-expressing cells in the blood and bone marrow of Apc compared with Apc mice (18). Therefore, Apc and Apc mice were bred onto the M-Lys background to facilitate the study of MM cell populations with intact M-Lys function. It was determined if there was any impact of heterozygous M-Lys deletion on intestinal tumorigenesis by counting macroscopic adenomas from Apc intestine. Adenoma numbers were 45.8±10 (mean ± SEM; n=5) from the small intestine of Apc mice (data not shown), similar to that of Apc mice previously bred in our facility that had 53±4 tumours (mean ± SEM; n=25) (25). Consequently, hemizygous deletion of M-Lys appeared not to have significant impact on intestinal tumorignenesis in the Apc mouse.

EGFP fluorescence in isolated small intestinal LPMNCs of M-Lyslys-EGFP/+ mice

M-Lys mRNA expression has previously been demonstrated in mouse intestine (26). The presence of EGFP protein was tested for and fluorescence in isolated intestinal LPMNCs of M-Lys mice. EGFP fluorescence in intestinal LPMNCs of Apc mice in situ as well as in the LPMNC isolate was observed (Fig. S1).

EGFP expression in the spleen and small intestine of Apc+/+M-Lyslys-EGFP/+ and ApcMin/+M-Lyslys-EGFP/+ mice

Utilizing an EGFP-expressing SW480 human CRC xenograft as positive control, it was determined if there was any difference in the resident intestinal MM cell localization between Apc and Apc mice at ~100 days of age (n=3 pairs) by immunohistochemistry for EGFP (Figs. 1 and 2). The spleen was studied as independent lymphoid tissue with a sentinel MM cell population. Even though there was some fibrotic distortion of Apc splenic tissue, EGFP-expressing cells were localized to the marginal zone of the spleen in both types of mice (Fig. 1B and C). In the intestine, EGFP-expressing cells were localized to intestinal villi and in particular lymphoid follicles of Apc mice (Fig. 2Aa-c). While EGFP-expressing cells were also localized to the intestinal villi of Apc mice and the periphery of adenomas, lymphoid follicles were absent from the intestine of Apc mice (Fig. 2Ba-c). This suggested loss of myelomonocytic cells and immune cell aggregates in Apc mouse intestine

Figure 1.

Localisation of EGFP expression in Apc and Apc mouse spleen. Sections of the spleen of an Apc mouse and an age-matched Apc mouse at 138 days of age were labelled with an antibody to EGFP (brown cells). Brown staining in (A) an EGFP-expressing SW480 cell line, (B) the spleen of an Apc mouse and (C) the spleen of an Apc mouse. In sections with omission of the primary antibody, absence of brown staining in (D) EGFP-expressing SW480 cell line, (E) the spleen of an Apc mouse and (F) the spleen of an Apc mouse. (G) Absence of brown staining in a section from the spleen of an Apc mouse. Arrows point to EGFP-expressing cells. Scale bars, 100 µm. APC, adenomatous polyposis coli; EGFP, enhanced green fluorescent protein; M-lys, mouse lysozyme.

Figure 2.

Localisation of EGFP expressing cells in Apc mouse small intestine. (A) Sections of the small intestine of an Apc mouse at 138 days of age labelled with an antibody to EGFP (brown cells). With the primary antibody to EGFP, brown staining in (a) proximal small intestine of the Apc mouse, (b) distal small intestine of the Apc mouse and (c) an intestinal lymphoid follicle of the distal small intestine of an Apc mouse (higher magnification, ×600). In sections with omission of the primary antibody, absence of brown staining in (d) proximal small intestine and (e and f) distal small intestine. (g) Absence of brown staining in distal small intestine of an Apc mouse stained with the EGFP antibody. (B) Sections of the distal small intestine of an Apc mouse labelled with an antibody to EGFP. With the primary antibody to EGFP, brown staining in (a) LPMNCs within the villi and lamina propria of an Apc mouse, (b and c) LPMNCs within an intestinal adenoma of an Apc mouse. Inset are EGFP-expressing cells within intestinal adenomas, shown at higher magnification (magnification, ×600). In sections with omission of the primary antibody, absence of brown staining in (d) distal small intestine and (e and f) adenomas. Arrows point to EGFP-expressing cells. Scale bars, 100 µm. APC, adenomatous polyposis coli; EGFP, enhanced green fluorescent protein; LPMNCs, lamina propria mononuclear cells; M-lys, mouse lysozyme.

Small intestinal myelomonocytic sub-populations of Apc+/+M-Lyslys-EGFP/+ and ApcMin/+M-Lyslys-EGFP/+ mice

To determine any differences in the intestinal MM population during the early stages of intestinal tumorigenesis, mice were studied at ~70 days of age, which is prior to overt Apc mouse lymphoid atrophy (7). This was also prior to ulceration, bleeding and potential secondary inflammation associated with advanced intestinal adenomas (6). As for mice on the M-Lys background, there was a trend towards reduced total intestinal LPMNCs in Apc mice which did not reach statistical significance (2.65±0.23×107 Apc vs. 1.72±0.39×107 Apc mice, P=0.12) Typical flow cytometry plots are shown in Fig. 3. There was no significant difference in the proportion of cells in the R1gate ([%] Apc 35. 40±4.00 vs Apc 38.00±4.60, P=0.52). Data on the three expressed macrophage marker antibodies, their mixture (Mac mix) and EGFP are displayed for Apc and Apc mice (Table II). A higher proportion of LPMNCs from Apc mice (n=6) expressed EGFP (P=0.11), ER-HR3 (P=0.11) or the Mac mix (P=0.18) compared with Apc littermates (n=7) (Table II). However, these differences were not statistically significant.

Figure 3.

Macrophage marker flow cytometry on M-lys intestinal LPMNCs. Representative flow cytometry plots from Apc mouse intestinal LPMNCs. Figures show the percentage of the total population of LPMNCs in relevant regions. EGFP fluorescence is on the abscissa, while PE fluorescence for macrophage markers is on the ordinate. R1, viable cell gate (not shown); R2, G+M+; R3, G+M-; R4, G-M-; R5, G-M+; R6, LPMNCs that do not express EGFP, but were labelled by the macrophage surface marker at a similar level to non-specific rat IgG; R7, LPMNCs that expressed EGFP and were labelled by macrophage surface markers at a similar level to non-specific rat IgG. LPMNCs, lamina propria mononuclear cells; M-lys, mouse lysozyme; PE, phycoerythrin; G+, EGFP-positive LPMNCs, G-: EGFP-negative LPMNCs. M+, LPMNCs labelled by macrophage markers above levels of non-specific binding by rat IgG. M-, LPMNCs not labelled by macrophage markers

Table II.

Small intestinal myelomonocytic cell populations of Apc and Apc mice at 74±2 days of age.

Myelomonocytic cell marker	Apc+/+	ApcMin/+	P-value
EGFP (M-lys)	4.29±0.68	2.90±0.47	0.11
ERHR-3
G-	0.11±0.03	0.13±0.04	0.83
G+	1.02±0.25	0.55±0.13	0.23
F4/80
G-	0.22±0.17	0.31±0.18	0.52
G+	0.24±0.10	0.14±0.07	0.20
MOMA-1
G-	0.04±0.01	0.07±0.02	0.12
G+	0.13±0.09	0.04±0.01	0.39
MAC-MIX
G-	0.16±0.10	0.15±0.05	0.48
G+	1.20±0.34	0.66±0.18	0.28

Pre-immune rat IgG served as the antibody-isotype control. G+, EGFP-positive LPMNCs; G, EGFP-negative LPMNCs.

Lower numbers of small intestinal myelomonocytic cells in the ApcMin/+M-Lyslys-EGFP/+ mice with the lowest total small intestinal LPMNC numbers

To clarify if there was an association between the trend to reduced Apc total intestinal LPMNC and MM cell numbers, the lamina propria MM cell population of mice with the lowest total LPMNC yield were studied (Apc; mice nos. 4, 5 and 7; Tables SI and SII). These three mice had an LPMNC yield <35% compared with the average LPMNC yield of Apc mice. For these three Apc mice, the proportion of EGFP-expressing cells (1.78±0.26) was >2-fold depleted compared with the mean for Apc mice (P=0.05) while the proportion of Mac-mix expressing cells (0.37±0.04) was ~4-fold depleted compared with the mean for Apc mice (P=0.05). This suggested that there was depletion of the myelomonocytic sub-population associated with reduced intestinal LPMNCs.

Discussion

Selective EGFP fluorescence of MM cells without significant impact on tumorigenesis in the novel Apc mouse in the present study prompted investigation into the MM cell population. No pan-myelomonocytic cell marker has yet been fully validated for the murine MM cell population, which the present observation of EGFP-negative Mac mix expressing cells corroborates. While a range of myelomonocytic cell markers were used in the present study, rare subsets not evaluated in the present study should not be ignored. However, the murine M-lys-expressing MM sub-population has been ascribed roles in phagocytosis and antigen presentation in the intestinal micro-environment, which are crucial to the immune response (27). A notable observation of the current study was a reduction in the Apc lamina propria MM cell population early during intestinal adenoma development, associated with loss of intestinal lymphoid/MM cell aggregates but with retention of splenic marginal zone MM cells. It is possible that similar to the reduction in total intestinal LPMNC numbers observed, MM cell depletion is progressive through adenoma development. Conventional dendritic cells that function as antigen presenting cells in the intestinal micro-environment are known to reside in intestinal lymphoid follicles (28). Consequently, the loss of lymphoid/MM cell interaction during intestinal adenoma development could impair the development of an immune response to tumour antigens, as previously shown in an autochthonous melanoma model (16), thus compromising immunosurveillance to prevent adenoma growth. In this respect, the prognosis following resection of early CRC is positively correlated with the extent of intra-tumoral lymphocytic infiltration (29). Furthermore, response to PD-1-inhibition in melanoma is associated with the presence of a primed peri-tumoral cytotoxic lymphocyte population (15).

Apc mice ≤84 days of age did not have evidence of reduced peripheral MM cells or lymphocytes (8). Consequently, total intestinal lamina propria and intestinal MM cell depletion in 70-day old Apc mice appeared to be due to factors in the intestinal micro-environment. After weaning in mice, F4/80-positive resident intestinal macrophages of embryonic origin are now known to be replaced by F4/80-negative macrophages, which are constantly re-populated from the peripheral circulation and adopt an anti-inflammatory phenotype in the intestinal microenvironment (30,31). We previously demonstrated that the resident intestinal MM population is the highest PGE2-secreting intestinal LPMNC sub-population in mice. Furthermore, there is a trend for higher PGE2 production by the Apc mouse resident intestinal MM population compared with the Apc mouse (23). PGE2 has previously been ascribed an immunosuppressive role (32,33), which would be consistent with an anti-inflammatory MM cell phenotype. It is noteworthy that MM cell conversion to an immunosuppressive phenotype has been shown to take up to 96 h after MM cell migration into the intestinal micro-environment; with dendritic cells typically surviving for several days while macrophages typically survive for several weeks (28,34). Consequently, in this respect, enhanced Wnt signalling may be physiologically relevant to signal exclusion of MM cells prior to their conversion to an immunosuppressive phenotype in the Apc intestinal milieu. However, other factors, such as cyclooxygenase-2 (35), TGF-β and T regulatory cells (36) may play a role in MM cell depletion from the intestinal micro-environment.

Out of the murine macrophage surface markers evaluated in the present study, ER-HR3 was the most widely expressed by intestinal lamina propria macrophages, irrespective of Apc genotype. ER-HR3 is known to be expressed by a sub-set of peripheral and resident tissue macrophages including those in sentinels, such as the skin, spleen and lymph nodes (37). Functionally, ER-HR3-expressing macrophages have been shown to be phagocytic and associated with immune latent milieu, such as the granulomata of BCG infected mice and with the relative exclusion of F4/80-expressing macrophages, which are more common in other murine tissue (38).

In conclusion, the present study demonstrated that the loss of MM/lymphoid interaction occurs during the early stages of Apc intestinal tumorigenesis, which may be due to enhanced micro-environmental Wnt signalling. Subsequent anergy to emerging tumour neoantigens could compromise tumour immunosurveillance in the Apc model. The relevance of this to colorectal adenomas and cancer requires evaluation, as well as the possibility that, similar to advanced melanoma, immune changes secondary to enhanced Wnt signalling may persist in MMR-proficient CRC. Furthermore, the mechanism to MM cell depletion requires further evaluation. Such studies may contribute to understanding the resistance to CKI therapy of MMR-proficient CRC despite a tendency for higher neoantigen burden compared with other malignancies, such as renal or bladder cancer (14).