Basal-like breast cancer engages tumor-supportive macrophages via secreted factors induced by extracellular S100A4

Abstract

The tumor microenvironment (TME) may influence both cancer progression and therapeutic response. In breast cancer, particularly in the aggressive triple-negative/basal-like subgroup, patient outcome is strongly associated with the tumor's inflammatory profile. Tumor-associated macrophages (TAMs) are among the most abundant immune cells in the TME, shown to be linked to poor prognosis and therapeutic resistance. In this study, we investigated the effect of the metastasis- and inflammation-associated microenvironmental factor S100A4 on breast cancer cells (BCCs) of different subtypes and explored their further interactions with myeloid cells. We demonstrated that extracellular S100A4 activates BCCs, particularly the basal-like subtype, to elevate secretion of pro-inflammatory cytokines. The secreted factors promoted conversion of monocytes to TAM-like cells that exhibited protumorigenic activities: stimulated epithelial-mesenchymal transition, proliferation, chemoresistance, and motility in cancer cells. In conclusion, we have shown that extracellular S100A4 instigates a tumor-supportive microenvironment, involving a network of cytokines and TAM-like cells, which was particularly characteristic for basal-like BCCs and potentiated their aggressive properties. The S100A4-BCC-TAM interaction cascade could be an important contributor to the aggressive behavior of this subtype and should be further explored for therapeutic targeting.

Keywords: S100A4; breast cancer; cytokines; tumor microenvironment; tumor-associated macrophages; tumor-stroma interactions.

Figure 1

Association between S100A4 expression and abundance of stroma and immune cells in BC tissue. (A) In silico methods (specified in Materials and methods) were applied to estimate the percentage of tumor cells, stromal score, immune score, and macrophage abundance in the tumor samples from TCGA's breast cancer cohort (n = 1052). The correlation between the estimated parameters and S100A4 transcript levels is indicated (the colors specify BC subtypes: red: basal‐like; dark and light blue: luminal A and B, respectively; violet: HER2; and green: normal‐like). (B) The levels of S100A4 and infiltrated macrophages in the PDX models, MAS98.06 (luminal) and MAS98.12 (basal‐like). The representative images show IHC staining of S100A4 protein and the macrophage‐specific marker CD68. Scale bar: 250 μm for all the images.

Figure 2

Cytokines secreted by BCCs at control and S100A4‐stimulated conditions. Cytokine levels were assessed in a panel of BCCs, which include cell lines (MCF7, MDA468, MDA231, HMLE, MA11, SKBR3), ex vivo cultures from PDXs (MAS98.06 and MAS98.12), and primary cultures from patient biopsies (P1, P2, P3, and P4) that represented different BC subtypes as indicated. CMs were harvested from nonstimulated controls (CM‐Ctr) and cells stimulated with 2 μg·mL⁻¹ rS100A4 for 24 h (CM‐S100A4), and analyzed for cytokines by the multiplex immunoassay. (A) Heatmaps of 25 of 27 cytokines (IL‐1β and IL‐5 were excluded as their levels were zero in the majority of the samples) across different BCCs in CM‐Ctr (top panel) and CM‐S100A4 (bottom panel). The cytokine data (average from two independent experiments) were log10‐transformed and median‐centered, CM‐Ctr data were clustered, and the resulting order of the BCCs and cytokines was used for visualizing both CM‐Ctr and CM‐S100A4 data. Two clusters (a and b) are shown. (B) Average concentrations (pg·mL⁻¹) of five most abundant cytokines secreted by the individual BCC models in two independent experiments (error bars indicate SEM). (C) Gene expression levels of the most abundant S100A4‐inducible cytokines in luminal and basal‐like tumors (the data retrieved from TCGA's breast cancer cohort).

Figure 3

Effect of BCC‐derived CMs on primary human monocytes. (A) The experimental setup is depicted in the cartoon. Primary human monocytes were cultivated for 7 days in the presence of indicated factors: GM‐CSF or M‐CSF (both 50 ng·mL⁻¹), rS100A4 (2 μg·mL⁻¹), CM‐Ctr, and CM‐S100A4 from MCF7 or MDA468 BCCs. Changes in monocyte morphology and expression of M1 and M2 polarization markers, CD80 and CD206, were analyzed by microscopy and flow cytometry, respectively. Representative phase‐contrast images from each condition are shown in the upper panel. Histogram overlays indicate the levels of CD80 and CD206 under each treatment condition (filled histograms) compared to the untreated control (unfilled histograms). (B) Levels of CD80 and CD206 in primary monocytes cultured in CM‐S100A4‐0 (depleted for rS100A4) versus the respective control CM‐Ctr‐0 from MCF7 and MDA468.

Figure 4

S100A4‐activated basal‐like BCCs stimulate THP1 monocyte differentiation and polarization toward M2 phenotype. The experimental setup is depicted in the cartoon. THP1 monocytes were cultured for 7 days in CM‐Ctr or CM‐S100A4 from luminal or basal‐like BCCs and analyzed for their morphological changes (A), number of immature cells that is a reverse measure of differentiation (B), and expression of genes reflecting M1 and M2 polarization (C). ‘rS100A4’ indicates treatment with 2 μg·mL⁻¹ rS100A4 alone. (A) Representative phase‐contrast images of THP1 cells under indicated conditions; THP1 cultured in CM‐Ctr did not demonstrate any morphological changes compared to ‘Untreated’ THP1 and are therefore not shown. (B) Relative number of immature THP1 cells under each condition. (C) Relative mRNA expression of M1 and M2 polarization markers, iNOS and CD206, respectively, in THP1 cells under each condition compared to the untreated control (set to 1); average ± SEM (n = 3; no SEM indicates n = 1); *P < 0.05.

Figure 5

S100A4‐activated basal‐like BCCs educate THP1 cells to produce cytokines. (A) THP1 cells were cultured for 7 days in CM‐Ctr or CM‐S100A4 from luminal or basal‐like BCCs (‘rS100A4’ – treatment with 2 μg·mL⁻¹ rS100A4 alone) and analyzed for expression of cytokine genes. Relative mRNA expression of indicated cytokines under each condition compared to untreated controls (set to 1). For CCL2, average ± SEM (n = 3, except MAS98.06, where n = 1); the other cytokines were measured in one representative sample from each BCC model. N.D. indicates below the detection limit. (B) THP1 cells were cultured for seven days in CM‐Ctr or CM‐S100A4 from MDA468, washed, and received an ordinary serum‐free medium, and after 3 days, their growth medium was analyzed for cytokines by multiplex immunoassay. The graph indicates the top 10 cytokines elevated in CM‐S100A4 compared to CM‐Ctr. Average concentrations (pg·mL⁻¹) of cytokines from two independent experiments ± SEM. The morphology of the resulting THP1 cells is depicted in the insert. (C) Heatmap with hierarchical clustering of all detectable cytokines (log₁₀‐transformed and median‐centered) in THP1 and MDA468 (Analyzed) in control and S100A4‐stimulated conditions (Stimuli). Two biological parallels for each condition are shown.

Figure 6

TAM‐like THP1 make MDA468 cells gain mesenchymal traits. The experimental setup is illustrated in the cartoon. Briefly, MDA468‐derived CM‐Ctr and CM‐S100A4 were prepared as before (step 1) and added to THP1 cells for 3‐day treatment to generate Ctr THP1 and TAM‐like THP1, respectively (step 2), which then were put in coculture together with naïve MDA468 cells labeled with GFP‐Luc for tracking purposes (step 3). The MDA468 cells from both cocultures were analyzed as follows: (A, B) Morphological changes were tracked by IncuCyte ZOOM ^® (A) and quantified at day 3 by measuring eccentricity (B); average ± SEM (n = 3). Scale bar: 50 μm. (C) Cellular localization of E‐cadherin (red) at day 6 was detected by IF [DAPI nuclear staining (blue)]. Scale bar: 50 μm. (D) Relative mRNA expression (at day 6) of indicated genes in the presence of TAM‐like THP1 versus Ctr THP1 (set to 1); average ± SEM (n = 3); *P < 0.05.

Figure 7

TAM‐like THP1 stimulate proliferation and reduce chemosensitivity of MDA468 cells. The cocultures of GFP‐Luc‐labeled MDA468 cells and Ctr THP1 or TAM‐like THP1 were prepared as illustrated in Fig. 6. (A) Cancer cell confluence/proliferation was tracked over time by IncuCyte ZOOM, recording the GFP signal. The confluence at 24 h postseeding was set to 1, and the increase in the confluence over time is indicated; average ± SEM (n = 5); *P < 0.05 for 48–72 h. (B) Both cocultures were treated with carboplatin or paclitaxel at the indicated concentrations for 3 days. The proliferation/viability of the treated cancer cells was scored by measuring Luc‐mediated bioluminescence and is presented as % from respective untreated controls (average ± SEM, n = 3–4).

Figure 8

TAM‐like THP1 stimulate motility of MDA468 cells. (A) GFP‐Luc‐labeled MDA468 cells were formed into spheroids and cultured alone (empty) or together with Ctr THP1 or TAM‐like THP1 as illustrated. Representative microscopy pictures (at day 5) under each condition are shown, indicating the total area covered by the MDA468 cells. Scale bar: 500 μm for all the images. (B) Migration area, that is, total area minus initial spheroid area, was measured in each condition and normalized to the migration area in the ‘empty’ control (set to 1); average ± SEM (n = 3); *P < 0.05.

Figure 9

An illustrative summary of the findings. Basal‐like BCCs, when exposed to S100A4 present in TME, increase secretion of numerous pro‐inflammatory cytokines (1). The secreted factors convert monocytes into TAMs (2), which stimulate EMT and promote aggressive functions (proliferation, chemoresistance, and motility) in BCCs (3).