HPV associated tumor cells control tumor microenvironment and leukocytosis in experimental models

Abstract

Human papillomavirus (HPV) is the main etiological factor for cervical cancer development. HPV is also associated with other anogenital and oropharyngeal tumors. HPV associated tumors are frequent and constitute a public health problem, mainly in developing countries. Therapy against such tumors is usually excisional, causing iatrogenic morbidity. Therefore, development of strategies for new therapies is desirable. The tumor microenvironment is essential for tumor growth, where inflammation is an important component, displaying a central role in tumor progression. Inflammation may be a causal agent, suppressor of anti-tumor T cell responses, or may have a role in angiogenesis, drug resistance, and metastasis. The aim of this work was to investigate the role of HPV transformed cells in the tumor microenvironment and tumor effects on myeloid populations in lymphoid organs in the host. We used experimental models, where we injected cervical cancer derived cell lines in immunodeficient mice, comparing HPV positive, SiHa, and HeLa cells (HPV 16 and HPV18, respectively), with HPV negative cell line, C33A. Our data shows that HPV positive cell lines were more efficient than the HPV negative cell line in leukocyte recruitment to the tumor microenvironment and increase in myeloid cell proliferation in the bone marrow and spleen. We also observed that HPV positive cells lines expressed significantly higher levels of IL-6 and IL-8, while C33A expressed significantly higher levels of IL-16 and IL-17. Finally, in spite of cytokine secretion by tumor cells, leukocytes infiltrating SiHa and HeLa tumors displayed almost negligible STAT3 and no NFκB phosphorylation. Only the inflammatory infiltrate of C33A tumors had NFκB and STAT3 activated isoforms. Our results indicate that, although from the same anatomical site, the uterine cervix, these cell lines display important differences regarding inflammation. These results are important for the design of immunotherapies against cervical cancer, and possibly against HPV associated tumors in other anatomical sites.

Keywords: cervical cancer; cytokines; human papillomavirus; inflammation; leukocytosis.

Figure 1

Leukocyte infiltrate in HeLa, SiHa, and C33A tumors. A: Immunohistochemistry from tumor sections stained with anti-CD45, anti-CD11b, and anti-Gr1. Tissue cryo-sections were incubated with the indicated antibodies, which were detected with the Vectastain Elite kit. Sections were counterstained with hematoxylin and eosin before mounting with Permount. Images were acquired in an Olympus BX61 microscope with 100× magnification. The bar indicates 100 μm. B–D: Flow cytometry analyzes of tumor cell suspensions. After harvesting, tumors were digested with Collagenase. Single cells suspensions were washed, aliquoted, and stained with antibodies against the indicated cell surface markers. In B, we show the flow cytometry results of a representative experiment. Upper panels display the CD45⁺ population, within which, all other analyzes were performed (arrows indicate the CD45⁺ population gate that contained the other populations). C and D. Average of cell population frequency in tumors in RAG1^−/− mice (C) and Nude mice (D). Each experimental group had, at least, four mice. * indicates significant differences in a population compared with the same in the C33A tumors, ** indicates significant differences between HeLa and SiHa tumors.

Figure 2

Cytokine expression profiles in in vitro and in vivo cervical cancer cell lines. Cell lysates were prepared from cells in culture (in vitro, black diamonds) or from cells sorted from tumors (in vivo, white diamonds). Sorted cells were obtained from tumors single cell suspensions eluted from magnetic columns after labeling with anti-CD45 magnetic beads. For all incubations, we used 100 μg of protein. Each of the lysates was incubated with membranes from the Human Cytokine Proteome Array kit, according to manufacturer's instructions. Cytokine/antibody complexes were detected using the ECL Chemiluminescence kit. Autoradiograms were scanned and densitometry data was used to generate the results. The graphs show absolute densitometry from each cytokine minus the average of six background spots with the same area of all other spots. The black lines in each graph represent the determined threshold of detection. Red arrows indicate significant differences between cytokine expression in a given cell line in vitro and in vivo. Gray arrows indicate cytokines with significant higher expression in HPV positive cell lines. White arrows indicate cytokines with significant higher expression in C33A. Three independent experiments are represented in this figure.

Figure 3

Signaling proteins expression in cervical cancer cell lines. Western blotting analyzes of 40 μg of proteins from cells in culture or from sorted CD45⁻ and CD45⁺ tumor cells. Sorted cells were obtained from tumors single cell suspensions eluted from magnetic columns after labeling with anti-CD45 magnetic beads. The membranes were first incubated with anti-phosphorylated proteins, as indicated in the left side of each autoradiogram. After striping in acidic glycine, Tris solution, membranes were incubated with antibodies that recognize the indicated proteins independently of post-translational modifications. This is one representative experiment of two independent ones. Control corresponds to Akt expression, which we have observed to be constitutive among all our expression conditions. The last reaction depicted, NFκB* is a longer exposure of the same autoradiogram shown for NFκB.

Figure 4

HPV positive tumors induce accumulation of myeloid cells in the spleen of mice. A: Number of nucleated cells in the spleen of control mice (naïve) or mice bearing C33A, HeLa, and SiHa tumors, respectively. After spleen dissociation and red cell lysis, nucleated cells were counted in Neubauer chambers to obtain the number of nucleate cells/spleen. All groups had significantly more cells than control naïve mice (*), HeLa and SiHa tumor bearing mice had significantly more cells than C33A tumor bearing mice (§). The inset over the graph shows a representative photograph of the spleens of tumor bearing mice, in the same order they are represented in the graph. B: Cell morphology in the spleens of control mice (naïve) and tumor bearing RAG1^−/− mice (lineages indicated above each image). Frozen and fixed spleen sections were stained with hematoxylin prior to mounting the slides. 400× magnification images were acquired with an Olympus BX61 microscope. The insets show nuclei details under 1000× magnification. Scale bar indicates 100 μm. Arrows indicate mononuclear cells and arrowheads indicate cells with doughnut shaped nuclei. C. Representative experiment of flow cytometry analyzes of myeloid CD11b⁺Gr1⁺splenocytes (left) and CD11b⁺F4/80⁺splenocytes (right) from RAG1^−/− mice. In the right panels, gates indicate the CD11b + Gr1int and Gr1+ populations as indicated. In the right panels, gates indicate the macrophage populations, independently of F4/80 staining intensity. D. Quantification of splenocyte populations analyzed by flow cytometry, within the nucleated population, as represented in C (right side splenocytes from RAG1^−/− mice, left side from Nude mice). E: The graphs in the right side show the percentage of BrdU incorporating cells within each specific population in C (right side splenocytes from RAG1^−/− mice, left side from Nude mice). * indicates significant differences in comparison to the control (naïve) group. Mac corresponds to CD11b⁺F4/80⁺, which we assume are macrophages.

Figure 5

HPV positive tumors induce accumulation of myeloid cells in the bone marrow of mice. Bone marrow single cell suspensions from RAG1^−/− (A) and Nude (B) mice were stained with antibodies against the indicated antigens and analyzed by flow cytometry. CD11b⁺Gr1⁺ population was split in Gr1int and Gr1⁺ for our analysis. The graphs in the left side show the frequency of each population within the nucleated cell population in control, C33A (C33), HeLa and SiHa tumor bearing mice. The graphs in the right side show the percentage of BrdU incorporating cells within each specific population. * indicates significant differences in comparison to the control group. Mac corresponds to CD11b⁺F4/80⁺, which we assume are macrophages. C: Splenocytes from mice injected or not with BrdU (BrdU +, solid line and BrdU-, dashed line, respectively) were processed the same way as cells in all other experiments and incubated with anti-BrdU in the same exact conditions. This experiment demonstrates that the antibody anti-BrdU is specific and our results are not a staining artifact. D. Splenocytes from control RAG1^−/− mice (RAG1^−/−) or RAG1^−/− mice with SiHa tumor (SiHa) were stained with anti-CD11b and anti-Gr1, as described before. Cells were then fixed, permeabilized and incubated with 10 μg/ml DAPI for 30 min, washed and ressuspended in the same DAPI concentration right before acquisition [22]. The histograms represent frequency of cells in each cell cycle phase, G1, S and G2/M, as specified in the right side of each graph, within the CD11b⁺Gr1^int population.

Figure 6

Schematic diagram of local and systemic tumor effects. In HPV associated tumors, we observed the presence of macrophages (MØ) and CD11b⁺Gr1⁺, here represented as polymorphonuclear cells. HPV associated tumor cells express IL-6, IL-8, and CXCL1, which may have systemic effects, promoting cell proliferation, accumulation, and possibly influence differentiation of myeloid cells in the bone marrow and spleen. We have not tested whether myeloid cells migrate from the bone marrow to the spleen, but due to the correlation of the type of cells expanding in the lymphoid organs and infiltrating the tumors, we believe these cells migrate to the tumor, where they have a suppressor like phenotype with low phosphorylated STAT3 expression, and no phosphorylated NFκB expression. In red, we pointed out possible mechanisms to inhibit tumor growth by inhibiting cell proliferation and recruitment, neutralizing cytokines or generating inflammatory signals that could activate infiltrating leukocytes.