IJC International Journal of Cancer

S100A9 has a protective role in inflammation-induced skin carcinogenesis Eileen McNeill and Nancy Hogg

Cancer Cell Biology

Leukocyte Adhesion Laboratory, Cancer Research UK London Research Institute, London, United Kingdom

The S100A8/A9 heterodimer is expressed by myeloid cells where its function has been extensively investigated. Immune cell S100A8/A9 promotes proinflammatory effects, and its absence is often associated with lack of leukocyte recruitment resulting in protection in terms of disease progression. S100A8/A9 is also expressed by certain epithelia, either constitutively as in mucosal epithelia or following stimulation as in skin keratinocytes. The role of the heterodimer in this context has not been as frequently explored. In this study, the incidence of skin papillomas induced by 7,12-dimethylbenz(a)anthracene (DMBA)/ 12-O-tetradecanoylphorbol-13-acetate (TPA) in S100a92/2 mice has been investigated. Unlike the immune disorders and certain models of cancer, absence of S100A8/A9 caused an increased incidence in skin of papillomas and, subsequently, squamous cell carcinomas. Although associated in S100a92/2 mice with increased recruitment of neutrophils and T cells, a bone marrow chimera experiment revealed the major defect to be primarily due to the absence of S100A8/A9 in the skin keratinocytes. S100a92/2 skin displayed enhanced Ki-67 expression over the time period of appearance of the papillomas suggesting an effect of S100A8/A9 in regulating proliferation in the epidermal layer. Thus, despite immune cell recruitment in S100a92/2 mouse skin that might have been predicted to promote tumor growth, it was the absence of S100A8/A9 in skin keratinocytes that dominated in terms of papilloma formation. The study highlights the importance of the S100A8/A9-expressing skin epidermal layer in controlling skin tumor formation and suggests that the influence of the heterodimer is dependent on the tissue context in which it is expressed.

The S100A8/A9 heterodimer (MRP-8/MRP-14, calprotectin and calgranulin A/B) belongs to a family of 20 small Ca21binding proteins that are coded for in the epithelial differentiation complex on human chromosome 1q21 and 3qF1-F2, the orthologous region in the mouse.1–4 The heterodimer is constitutively expressed in myeloid cells such as neutrophils, monocytes, dendritic cells and platelets, with the level in neutrophil cytosolic protein constituting an abundance of 45%.5 S100A8/ A9 is also constitutively expressed in certain mucosal epithelia and in other epithelial tissues such as skin under conditions of inflammation or injury.6,7 Complex carbohydrates such as carboxylated glycans and heparan sulfate are bound by S100A8/ Key words: S100A9, MRP-14, skin carcinogenesis, inflammation, neutrophils Additional Supporting Information may be found in the online version of this article. Grant sponsor: Cancer Research UK Eileen McNeill’s current address is: Division of Cardiovascular Medicine, University of Oxford, John Radcliffe Hospital, Oxford, OX3 9DU, United Kingdom DOI: 10.1002/ijc.28725 History: Received 25 Oct 2013; Accepted 8 Jan 2014; Online 16 Jan 2014 Correspondence to: Nancy Hogg, Leukocyte Adhesion Laboratory, Cancer Research UK London Research Institute, 44 Lincoln’s Inn Fields, London WC2A 3LY, United Kingdom, Tel.: 144–207-269– 3255, Fax: 44–207-269–3417, E-mail: [email protected]

C 2014 UICC Int. J. Cancer: 135, 798–808 (2014) V

A9.8,9 The glycans decorate receptor of advanced glycation end Products (RAGE) that acts as one well-studied type of target ligand.10 S100A8/A9 also serves as a coreceptor of the toll-like receptor TLR4–MD2 complex enhancing the response to endotoxin lipopolysaccharide (LPS).11 Mechanistic analysis has focused on myeloid cells with studies making use of the S100a92/2 mouse, which lacks not only S100A9 but also the protein of its partner S100A8 which is unstable in the absence of S100A9.12,13 In these models, S100A8/ A9 is frequently associated with proinflammatory effects. Thus, S100a92/2 mice are protected from Gram-negative bacteria endotoxin-induced shock11 and acute pancreatitis.14 They are also protected from vascular injury in an atherosclerosis model, where there is a lack of leukocyte recruitment,15 but not from Tcell-mediated cardiac allograft rejection, where antigen presentation is enhanced.16 S100A8 and also S100A9 have been reported to be directly chemotactic for myeloid cells.17,18 However, the proteins operate intracellularly as well as extracellularly.1–4 Upregulation of S100A8/A9 is associated with cancer progression in mouse tumors.2,4,19 In models involving the pancreas,14 colon20 and the breast,21 the absence of S100A8/A9 results in a decrease in infiltrating leukocytes and a subsequent reduction in tumor burden. In the solid tissue tumors, the tumor cells themselves express increased S100A8/A9, suggesting that the heterodimer may have distinct roles in each cell type. In the lymphoid EL4 and multiple myeloma models, antitumor responses are suppressed by S100A8/A9 through its ability to promote accumulation of myeloid-derived suppressor

799

McNeill and Hogg

What’s new? Expression of the proinflammatory protein S100A8/A9 in epithelial tissue can be induced following stress such as that caused by skin carcinogen exposure. Here, the ability of the carcinogens 7,12-dimethylbenz(a)anthracene (DMBA) and 2-O- tetradecanoylphorbol-13-acetate (TPA) to induce skin papillomas was investigated in S100a9-/- mice. The experiments show that the absence of S100A8/A9 results in increased incidence of skin papillomas, apparently due to a loss of S100A8/A9 regulation resulting in over proliferation in the epidermal layer. The findings suggest that S100A8/A9 plays a protective role against skin papillomas specifically in skin keratinocytes, indicating that the heterodimer acts in a tissue-dependent context.

Material and Methods Mice

Mice were housed in a specified pathogen-free environment according to the UK Home Office Legislation. S100a92/2 mice were backcrossed against C57BL6/J mice for ten generations.12 S100a92/2 mice were crossed with Lfa-12/2 (Ref. 26) or Rag-22/2 (Ref. 27) mice (both backcrossed to C57Bl6/J for more than ten generations) to produce mice that were homozygous for both alleles. Control C57BL6/J mice were exactly age-matched with the genetically altered mice. Skin Carcinogenesis

A standard protocol for carcinogen-induced skin cancer was followed as previously described.28 Mice were shaved over a 2 cm 3 3 cm dorsal area of skin at 7 weeks. An initiating dose of 25 mg of 7,12-dimethylbenz(a)anthracene (DMBA; Acros Organics/Thermal Fisher Scientific) in 100 ml of acetone was applied 1 week later. Then, 4 mg of 12-O-tetradecanoylphorbol-13-acetate (TPA; Sigma Aldrich) in acetone was applied thrice weekly for 15 weeks. For time course studies, mice were harvested within 22–24 hr of the final TPA treatment. Mice were observed daily, and the number of palpable tumors per mouse was recorded at twice weekly intervals. Mice were euthanized at the experimental endpoint of tumor reaching 1 cm in any dimension or becoming invasive or ulcerated. The conversion of papillomas to carcinoma was determined by technical staff expert in this model. Tumors were scored as carcinoma when they showed macroscopic evidence of invasion into the skin and diagnostic histopathology. Mouse Monoclonal Antibody Treatment

Rat mAb selective for neutrophils (RB6-8C5-Ly6G and IgG2b) and isotype control mAb PyLT1 were produced C 2014 UICC Int. J. Cancer: 135, 798–808 (2014) V

under low endotoxin conditions (CR UK London Research Institute Cell Services Unit). The mAbs were assessed to have minimal endotoxin using the QCL-1000 assay (Lonza). Mice were injected intraperitoneally with 0.2 mg of mAb twice per week, starting the week prior to initiation with DMBA. MAb treatments were maintained throughout the TPA treatment phase. Bone Marrow Chimeras

Six-week-old mice were irradiated with 23 5 Gy twice, separated by 4 hr and injected intravenously with bone marrow from donor animals within 24 hr of the last radiation dose. Bone marrow from one donor mouse was used to reconstitute five recipient animals, with each receiving at least 5 3 106 cells. Mice were maintained on prophylactic antibiotics for 6 weeks. Histology and Immunohistochemistry

Tumors or skin samples were mounted on nitrocellulose to preserve orientation and then immediately fixed in 10% neutral buffered formalin. Samples were paraffin embedded, sectioned and stained with hemotoxylin and eosin for morphological analysis. The tumors were staged by a pathologist blinded to the experimental group. Paraffin-embedded sections were stained with the following mAbs specific for mouse neutrophils (Ly-6G; BD Biosciences/Pharmingen), mononuclear phagocytes (Mac-3; BD Biosciences/Pharmingen), T cells (rabbit anti-CD3; Dako), S100A9 (2B10)12 and proliferating cells (Ki-67; Dako).29 Secondary antibody biotinylated rabbit anti-rat Ig or goat antirabbit Ig (Vector Laboratories) was applied for 45 min at room temperature, and then sections were washed in PBS and incubated in ABC (PK-6100 Elite ABC Kit; Vector Laboratories) for 30 min. Following washing in PBS, DAB solution (SK4100; Vector Laboratories) was applied for 2–5 min followed by hematoxylin staining. Sections were imaged using a Leica DMRBE microscope, Media Cybernetics Cool Snap Pro Color camera and Image Pro Express acquisition software (Media Cybernetics), and the area of positive staining per section (203 objective field of view) was quantified using Image Pro Plus software (Media Cybernetics). In the case of neutrophils, the number of cells per section was quantified manually. Ki-67-expressing nuclei were quantified by automatic cell counting using the NIS Elements software (Nikon).

Cancer Cell Biology

cells (MDSCs) and the subsequent blocking of T-cell responses that are reduced by delayed dendritic cell maturation.22–24 In this study, we have used the S100a92/2 mouse to investigate a role for S100A8/A9 in the skin papilloma model of carcinogenesis where S100A8 and S100A9 are known to be induced.25 In contrast to other mouse models of cancer, we find that S100A8/A9 has a suppressive effect on the incidence of papilloma tumors and that S100A8/A9 expressed in the epidermal compartment predominates.

800

Isolation of Epidermal RNA

Skin sections were immersed in diethylpyrocarbonate-treated water at 65 C for 7 sec. The heated sections were then stretched onto card, flattened between two pieces of card and twisted to loosen the epidermal layer. The top card was removed, and the epidermis was gently separated from dermis using a spatula. The epidermal sample was immediately transferred into ice-cold Solution D (4 M guanidine isothiocyanate, 25 mM sodium citrate, 0.5% N-laurosylsarcosine, 0.1 M 2-mercaptoethanol, pH 7.0) and homogenized. RNA was isolated using phenol chloroform extraction and was further purified using the RNeasy kit according to the manufacturer’s instructions (Qiagen).

Cancer Cell Biology

Quantitative RT-PCR

Epidermal RNA samples were reverse-transcribed to cDNA using Superscript II according to the manufacturer’s instructions (Invitrogen). Cytokine gene expression was assessed using Taqman Gene Expression Assay probes which were obtained from Applied Biosystems with assays performed according to the manufacturer’s instructions and analyzed using an ABI 7700 Sequence Detection System instrument (Applied Biosystems). Statistical Analysis

Statistical analysis was performed using Prism software (Graph Pad). The difference between genotypes in papilloma multiplicity was tested using two-way ANOVA and in mouse survival by Kaplan–Meier analysis. For the quantification of histological sections, three randomly selected fields of view were imaged for each stained tissue section, and the mean data for each of the four individual animals per time point were used for statistical analysis. The effect of genotype on cell infiltration and cytokine production was assessed using two-way ANOVA. When ANOVA demonstrated significant differences between genotypes, a modified T-test was used to identify the differences. A p-value of 80 weeks, respectively) when survival to the defined humane endpoint was assessed (Fig. 1e). This difference in carcinoma development was statistically significant, with tumor-free survival being greater in the wild-type mice. When the carcinomas elicited by this model were examined histologically, the majority were either lowor high-grade dysplastic papillomas, but a significant percentage had the characteristics of invasive squamous cell carcinoma (SCC; Fig. 1f). However, a key point was that there were no differences in proportions of each class of tumor between wild-type and S100a92/2 mice. S100A9 was expressed by the keratinocytes comprising the papillomas (Fig. 1g) as previously described.25 Therefore, the absence of S100A9 increased the incidence of papilloma and SCC formation in both female and male mice and appeared to affect the early stages of tumor development. Increased Neutrophils and T Cells in S100a92/2 Skin of DMBA/TPA-Treated Mice

To discover whether the absence of S100A8/A9 was influencing the inflammatory phase of papilloma growth, we investigated the leukocyte infiltration of the dermis and epidermis of wild-type and S100a92/2 female mice. Control skin showed only a minimal population of neutrophils, T cells and monocyte/macrophages (Supporting Information Fig. S1). However, analysis of skin samples taken from mice postDMBA initiation revealed recruitment of all three cell types to the dermis, and also to the epidermis in the case of T cells, of both wild-type and S100a92/2 mice (Fig. 2a). A significantly greater number of neutrophils and T cells were recruited to the skin of S100a92/2 than wild-type mice, with no difference observed in monocyte/macrophages. Maximum recruitment occurred at the 3-week postinitiation time point which preceded the appearance of papillomas (Figs. 2a and 2b). Thus, there was a significant difference in leukocytes migrating into S100a92/2 and control skin at a key stage during exposure to DMBA/TPA. C 2014 UICC Int. J. Cancer: 135, 798–808 (2014) V

801

Cancer Cell Biology

McNeill and Hogg

Figure 1. Genetic deletion of S100A9 increases susceptibility to DMBA/TPA skin carcinogenesis. (a) Experimental scheme of TPA-enhanced skin carcinogenesis protocol with a single dose of DMBA followed by 15 weeks of TPA application. The average number of papillomas per mouse was recorded weekly for the course of the experiment; at later time points, the numbers of papillomas displaying grossly dysplastic features were recorded as carcinoma. Mean numbers of papillomas per mouse were quantified up to the time that the first animal reached its humane endpoint. (b) Female and male S100a92/2 mice show enhanced tumor multiplicity compared to wild-type controls (n 5 15 mice per group; female and male, p < 0.01). (c) S100a92/2 compared to wild-type mice show a more rapid onset of papilloma development (female, p < 0.05; male, p < 0.01). (d) Carcinoma development (female and male, p < 0.01) and (e) decreased survival to a defined endpoint (female and male, p < 0.05). (f) Histopathological analysis of tumor type on termination of the experiment showed a similar proportion of each grade in wild-type and S100a92/2 mice; pooled male and female mice, nine tumors from eight wild-type mice and 25 tumors from S100a92/2 mice. (g) Immunohistochemistry of a papilloma from wild-type mouse skin at 12 weeks post-DMBA/TPA initiation showing tumor keratinocytes labeled with anti-S100A9 mAb. Scale bar 5 100 mm.

C 2014 UICC Int. J. Cancer: 135, 798–808 (2014) V

Cancer Cell Biology

802

Absence of S100A9 enhances skin tumor development

Figure 2. Inflammatory cell recruitment to the skin following DMBA/TPA treatment. Skin samples were taken from female mice following TPA treatment at 3 weekly intervals or from untreated controls. (a) Tissue sections were stained with mAbs targeted against neutrophils (Ly-6G), T cells (CD3) or monocyte/macrophages (Mac3); *p < 0.05. The number of Ly-6G1 neutrophils (some neutrophils marked with arrows) were counted for three fields of view per sample (n 5 4 samples per genotype, per time point). CD31 T cells and Mac31 macrophages were assessed by quantification of positive staining using Image Pro Plus software (n 5 4 samples per genotype, per time point). CD31 cells were assessed in both epidermis and dermis; Mac31 staining was quantified in the dermis alone as no specific staining of the epidermis was observed. (b) Representative images of neutrophils, T cells and monocyte/macrophages in skin of wild-type and S100a92/2 mice after 3 weeks of TPA treatment. Scale bar 5 50 mm.

The Contribution of T Cells and Neutrophils to Papilloma Formation

An issue was whether increased leukocyte recruitment contributed in general to papilloma growth and, second, whether there might be differences between wild-type and S100a92/2 mice. To answer the first question, we made use of Lfa-12/2 mice that lack the major leukocyte integrin that leukocytes use to migrate from the circulation into tissue spaces26 and compared recruitment in Lfa-12/2 mice and Lfa-12/2/ S100a92/2 mice. The absence of LFA-1 caused an equivalent reduction in papilloma development in both wild-type and S100a92/2 strains of mice (Fig. 3a). This decrease accounted for most of the papilloma formation in wild-type mice, but left a substantial complement of tumors in the S100a92/2 mice. The result indicates that the presence of leukocytes is important in this skin papilloma model, especially in wildtype mice, and that they drive disease in both the presence and absence of S100A8/A9. To further define the nature of the leukocytes that mediated papilloma formation, we next made use of Rag-22/2

mice that lack mature lymphocytes, but have competent myeloid cells.30 Rag-22/2 mice were almost entirely disease free, indicating that lymphocytes are essential for papilloma formation in wild-type mice (Fig. 3b). In contrast, although lack of lymphocytes in the S100a92/2 mice had some suppressive effect, it did not account for the elevation in their papilloma incidence. To address whether the increased recruitment of leukocytes might contribute to enhanced papilloma development, we performed a standard mAb depletion protocol using Gr-1 mAb RB68C5 that targets principally not only neutrophils (Ly6G) but also monocytes (Ly6C) and CD81 T cells (Ly6C). This treatment had no effect on the numbers of papillomas in the wild-type mice, suggesting that neither myeloid cells nor potentially CD8 T cells contribute to their formation (Fig. 3c). However, in the S100a92/2 mice, depletion of these leukocytes caused a significant, although not complete, decrease in papillomas. Altogether, the data indicate that myeloid cell recruitment has a specific role in promoting enhanced papilloma formation in S100a92/2 mice.

C 2014 UICC Int. J. Cancer: 135, 798–808 (2014) V

803

Cancer Cell Biology

McNeill and Hogg

Figure 3. Leukocyte recruitment controls susceptibility to DMBA/TPA skin carcinogenesis. S100a92/2 mice were crossed with Lfa-12/2 or Rag-22/2 mice to produce S100a92/2/Lfa-12/2 and S100a92/2/Rag-22/2 mice. (a) Female Lfa-12/2 mice displayed reduced papilloma multiplicity compared to wild-type controls on both wild-type and S100a92/2 backgrounds [wt (n 5 15), Lfa-12/2 (n 5 7), p < 0.0001; S100a92/2 (n 5 13), S100A92/2/Lfa-12/2 (n 5 13), p < 0.0001]. (b) Female Rag-22/2 mice displayed reduced papilloma multiplicity compared to wild-type controls, with a less evident effect of Rag-22/2 seen on an S100a92/2 background [wt (n 5 15), Rag22/2 (n 5 6), p

S100A9 has a protective role in inflammation-induced skin carcinogenesis.

The S100A8/A9 heterodimer is expressed by myeloid cells where its function has been extensively investigated. Immune cell S100A8/A9 promotes proinflam...
857KB Sizes 3 Downloads 0 Views