Paracrine Interactions between Adipocytes and Tumor Cells Recruit and Modify Macrophages to the Mammary Tumor Microenvironment: The Role of Obesity and Inflammation in Breast Adipose Tissue

Abstract

The relationship between obesity and breast cancer (BC) has focused on serum factors. However, the mammary gland contains adipose tissue (AT) which may enable the crosstalk between adipocytes and tumor cells contributing to tumor macrophage recruitment. We hypothesize that the breast AT (bAT) is inflamed in obese females and plays a major role in breast cancer development. The effects of this interplay on macrophage chemotaxis were examined in vitro, using co-cultures of mouse macrophages, mammary tumor cells and adipocytes. Macrophages were exposed to the adipocyte and tumor paracrine factors leptin, CCL2 and lauric acid (alone or in combinations). In cell supernatants Luminex identified additional molecules with chemotactic and other pro-tumor functions. Focus on the adipokine leptin, which has been shown to have a central role in breast cancer pathogenesis, indicated it modulates macrophage phenotypes and functions. In vivo experiments demonstrate that mammary tumors from obese mice are larger and that bAT from obese tumor-bearers contains higher numbers of macrophages/CLS and hypertrophic adipocytes than bAT from lean tumor-bearers, thus confirming it is more inflamed. Also, bAT distal from the tumor is more inflamed in obese than in lean mice. Our results reveal that bAT plays a role in breast cancer development in obesity.

Figure 1

Free fatty acid (FFA) concentration in mammary tumors may be modulated by cells of the mammary tumor microenvironment. Co-culture of adipocytes, macrophages and mammary tumor cells results in reduced amounts of FFA detected in the co-culture cell supernatant, as compared with the amounts released by adipocytes only through lipolysis, as determined by acyl-CoA oxidase-based enzymatic colorimetric methodology in 48 h supernatants from 3T3-L1 murine adipocytes in vitro differentiated for 11 days and co-cultured with macrophages, with mammary tumor cells or with both. Data are representative of one of three independent experiments showing similar results; each sample was tested in duplicates.

Figure 2

In vitro differentiated adipocytes significantly differ from adipocytes ex vivo isolated from visceral obese adipose tissue in their leptin secretion: leptin is a major marker of obesity and not of adipocyte differentiation. 48 h supernatants from in vitro differentiated (A) and ex vivo isolated adipocytes, (B) co-cultured with macrophages and tumor cells were analyzed for leptin production using ELISA. Data are representative of one of three independent experiments showing similar results; each sample was tested in duplicates. One-way ANOVA analysis showed differences among all the samples in both figures (A: *** p < 0.001 and B: **** p < 0.0001); TMCT comparisons are depicted in the figures.

Figure 3

Co-culture of adipocytes, tumor cells and macrophages reveals the highest production of CCL2 (MCP-1). 48 h supernatants from in vitro differentiated (A) and ex vivo isolated (B) adipocytes co-cultured with macrophages and tumor cells were analyzed for CCL2 production using ELISA. Data are representative of one of three independent experiments showing similar results; each sample was tested in duplicates. One-way ANOVA analysis showed differences among all the samples in both figures (**** p < 0.0001), and TMCT comparisons are depicted in the figures.

Figure 4

Migration studies with human monocytic cell line THP1. Chemotaxis experiments were conducted for 2.5 h at 37 °C. Top Chamber (ThinCert 8 µm pores): 1 × 10⁶ THP1cells in 100 µL vol. Bottom chamber contained the chemoattractant cells supernatants or recombinant/synthetic factors in 600 µL volume. (A1) CCL2 is a major monocyte chemoattractant, anti-CCL2 antibody (25 µg/mL) significantly reduces its activity. (A2) Increasing leptin concentrations has weak and ns chemotactic activity. (A3) Lauric acid has no chemoattractive properties on monocyte. (B) Effect of the different paracrine factors: The combination of CCL2, leptin and lauric acid does not significantly modify the monocyte chemoattraction comparable to the one exerted by CCL2, but in absence of CCL2 there is no chemotaxis. (C) Effects of 48 h supernatants from the different cells of interest (adipocytes, macrophages and tumor cells) in monocyte recruitment, without and with the paracrine factors inhibitors. CCL2 was used at 50 ng/mL, leptin at 10 ng/mL and LA at 5 µM in B and C. Anti-CCL2 blocking antibody was used at 25 µg/mL, leptin receptor inhibitor LPrA2 was used at 100 nM and TLR4 inhibitor Eritoran was used at 10 ng/mL. Data are representative of one of three independent experiments showing similar results; each sample was tested in triplicates. One-way ANOVA analysis showed significant differences among all the samples in the different figures: A1, A2: *** p < 0.001; A3: **** p < 0.0001; B: ** p < 0.01, and, C1, C2 and C3 **** p < 0.0001; TCMT comparisons are depicted in the figures.

Figure 4

Migration studies with human monocytic cell line THP1. Chemotaxis experiments were conducted for 2.5 h at 37 °C. Top Chamber (ThinCert 8 µm pores): 1 × 10⁶ THP1cells in 100 µL vol. Bottom chamber contained the chemoattractant cells supernatants or recombinant/synthetic factors in 600 µL volume. (A1) CCL2 is a major monocyte chemoattractant, anti-CCL2 antibody (25 µg/mL) significantly reduces its activity. (A2) Increasing leptin concentrations has weak and ns chemotactic activity. (A3) Lauric acid has no chemoattractive properties on monocyte. (B) Effect of the different paracrine factors: The combination of CCL2, leptin and lauric acid does not significantly modify the monocyte chemoattraction comparable to the one exerted by CCL2, but in absence of CCL2 there is no chemotaxis. (C) Effects of 48 h supernatants from the different cells of interest (adipocytes, macrophages and tumor cells) in monocyte recruitment, without and with the paracrine factors inhibitors. CCL2 was used at 50 ng/mL, leptin at 10 ng/mL and LA at 5 µM in B and C. Anti-CCL2 blocking antibody was used at 25 µg/mL, leptin receptor inhibitor LPrA2 was used at 100 nM and TLR4 inhibitor Eritoran was used at 10 ng/mL. Data are representative of one of three independent experiments showing similar results; each sample was tested in triplicates. One-way ANOVA analysis showed significant differences among all the samples in the different figures: A1, A2: *** p < 0.001; A3: **** p < 0.0001; B: ** p < 0.01, and, C1, C2 and C3 **** p < 0.0001; TCMT comparisons are depicted in the figures.

Figure 5

Cytokines, chemokines and growth factors detected in cell supernatants from adipocytes, mammary tumor cells, macrophages and their co-cultures. 48 h supernatants from the three different cell types and their co-cultures were analyzed for cytokines, chemokines and growth factors using Luminex analysis. (A) Comparison of the proteins secreted by in vitro differentiated 3T3-L1 adipocytes and ex vivo isolated adipocytes, showing statistical differences (Student’s t-test) between the two types of adipocytes compared to each other. (B) Cytokines/chemokines/chemokines/growth factors detected in E0771 cell supernatants. (C) Cytokines/chemokines/growth factors secreted by peritoneal elicited macrophages. (D) Cytokines/chemokines/growth factors secreted by co-cultures between ex-vivo isolated adipocytes, macrophages and E0771 mammary tumor cells. Data are representative of one of two independent experiments showing similar results; each sample was tested in duplicates.

Figure 5

Cytokines, chemokines and growth factors detected in cell supernatants from adipocytes, mammary tumor cells, macrophages and their co-cultures. 48 h supernatants from the three different cell types and their co-cultures were analyzed for cytokines, chemokines and growth factors using Luminex analysis. (A) Comparison of the proteins secreted by in vitro differentiated 3T3-L1 adipocytes and ex vivo isolated adipocytes, showing statistical differences (Student’s t-test) between the two types of adipocytes compared to each other. (B) Cytokines/chemokines/chemokines/growth factors detected in E0771 cell supernatants. (C) Cytokines/chemokines/growth factors secreted by peritoneal elicited macrophages. (D) Cytokines/chemokines/growth factors secreted by co-cultures between ex-vivo isolated adipocytes, macrophages and E0771 mammary tumor cells. Data are representative of one of two independent experiments showing similar results; each sample was tested in duplicates.

Figure 6

Leptin regulates expression of cytokines and chemokines in macrophages. Peritoneal elicited macrophages were pre-treated with increasing concentrations of leptin (0, 10 and 100 ng/mL) and activated with LPS (10 µg/mL) or with and LPS+IFNγ (IFNγ at 100 u/mL) for 20 h, and supernatants were collected and tested for IL-12p70 and IL-10 expression by ELISA and for nitric oxide (NO) production via the Griess reaction (A). In other experiments, constitutive and LPS-activated macrophages (B) were treated with increasing concentrations of leptin (0, 20 and 100 ng/mL) for 20 h and their supernatants examined by Luminex. Data are representative of one of three independent experiments in each case, which yield similar results, where each sample was tested in duplicates. One-way ANOVA analysis showed significant differences among all the samples in the figures: in A *** p < 0.0001 and in B ** p < 0.0001; TCMT comparisons are depicted in the figures.

Figure 7

Leptin impairs expression of myeloid differentiation markers in macrophages. (A) Flow cytometric analysis of CD11b, F4/80, CD115 and Gr-1 expression in resting N-PEMs pre- treated with leptin at 10 ng/mL for 20 h shows that macrophages exposed to leptin downregulate the percentage of cells that express F4/80 and CD11b. Unstained macrophages were compared with untreated (control) and leptin-treated macrophages stained with the antibodies for those markers. Flow cytometry results were analyzed using Flow Jo software; percentages (%) of cells expressing the respective markers as well as the mean fluorescence intensity value (MFI) of each marker were analyzed. The figure presents the results of a representative experiment with macrophages from an individual mouse; three different experiments were made with five mice per experiment; (B) Histograms showing the statistical analysis made using Student’s t-test to compare the means of the percentages of cells expressing the analyzed markers in macrophages from n = 15 mice used in the three experiments.

Figure 8

Leptin regulates the expression of pro-inflammatory/tumor-promoting transcription factors, membrane receptors and signaling intermediates, as well as of its own receptor in macrophages. N-PEMs (and also TAMs in C) were pre-treated with increasing concentrations of leptin for 20 h with or without LPS, whole cell protein was isolated and Western blots were performed. Densitometries of the autoradiograms are shown.

Figure 9

In vivo experiments: weight control and tumor progression. (A) Experimental design of the in vivo experiments. (B) Feeding 33%Kcal/fat HFD results in a rapid splitting of the mice weights into obese, overweight and obese resistant animals in approximately 7–8 weeks after the start of the diets. (C) Weight progress in all the mice fed both HF and LF diets during the first eight weeks before tumor cells were implanted. (D) General effect of body weight on tumor growth. (E) Effect of treatment with leptin antagonist peptide on tumor growth in overweight mice. (F) Effect of treatment with leptin antagonist peptide on tumor growth in obese mice. (G) Association between body weight and systemic (serum) levels of leptin in tumor bearers with different body weights untreated or treated with the leptin inhibitor. Each experimental subgroup was comprised of ten (10) mice.

Figure 10

Mammary adipose tissue is inflammatory in obese mammary tumor-bearing females: comparison of mammary fat in the tumor microenvironment vs. mammary fat distant from the tumor. IHC was performed in sections of mammary tumors and in tumor-distal mammary adipose tissues of obese, overweight and lean C57 BL6 tumor bearer female mice. IHC- positive cells appear in brown (20× magnification in all the slides). (A) mammary tumor microenvironments of lean, overweight and obese mice showing F4/80⁺ macrophages and crown- like structures (CLS, depicted by red arrows). (B) Mammary adipose tissue biopsied distant from the tumor site showing F4/80⁺ macrophages in lean, overweight and obese tumor bearers. (C) IL-10⁺ cells in the tumor microenvironments of lean, overweight and obese tumor bearers. (D) IL-12+ cells in the tumor microenvironments of lean, overweight and obese tumor bearers. (E) Ob-R expression is reduced in the mammary tumor microenvironments of obese and overweight females. IHC was quantified by three independent observers and their average is shown in the figures: scores were 1+, 2+, 3+, where 1+ represents from 1–150 positive cells in the whole slide, 2+ represents 151–300 positive cells in the whole slide and 3+ represents more than 300 positive cells in the whole slide. Each experimental subgroup was comprised of ten (10) mice.