SUMOylation regulates the aggressiveness of breast cancer-associated fibroblasts

Abstract

Background: Cancer-associated fibroblasts (CAFs) are the most abundant stromal cellular component in the tumor microenvironment (TME). CAFs contribute to tumorigenesis and have been proposed as targets for anticancer therapies. Similarly, dysregulation of SUMO machinery components can disrupt the balance of SUMOylation, contributing to tumorigenesis and drug resistance in various cancers, including breast cancer. We explored the role of SUMOylation in breast CAFs and evaluated its potential as a therapeutic strategy in breast cancer.

Methods: We used pharmacological and genetic approaches to analyse the functional crosstalk between breast tumor cells and CAFs. We treated breast CAFs with the SUMO1 inhibitor ginkgolic acid (GA) at two different concentrations and conditioned media was used to analyse the proliferation, migration, and invasion of breast cancer cells from different molecular subtypes. Additionally, we performed quantitative proteomics (SILAC) to study the differential signalling pathways expressed in CAFs treated with low or high concentrations of GA. We confirmed these results both in vitro and in vivo. Moreover, we used samples from metastatic breast cancer patients to evaluate the use of GA as a therapeutic strategy.

Results: Inhibition of SUMOylation with ginkgolic acid (GA) induces death in breast cancer cells but does not affect the viability of CAFs, indicating that CAFs are resistant to this therapy. While CAF viability is unaffected, CAF-conditioned media (CM) is altered by GA, impacting tumor cell behaviour in different ways depending on the overall degree to which SUMO1-SUMOylated proteins are dysregulated. Breast cancer cell lines exhibited a concentration-dependent response to conditioned media (CM) from CAFs. At a low concentration of GA (10 µM), there was an increase in proliferation, migration and invasion of breast cancer cells. However, at a higher concentration of GA (30 µM), these processes were inhibited. Similarly, analysis of tumor development revealed that at 10 µM of GA, the tumors were heavier and there was a greater degree of metastasis compared to the tumors treated with the higher concentration of GA (30 µM). Moreover, some of these effects could be explained by an alteration in the activity of the GTPase Rac1 and the activation of the AKT signalling pathway. The results obtained using SILAC suggest that different concentrations of GA affected cellular processes differentially, possibly influencing the secretome of CAFs. Treatment of metastatic breast cancer with GA demonstrated the use of SUMOylation inhibition as an alternative therapeutic strategy.

Conclusion: The study highlights the importance of SUMOylation in the tumor microenvironment, specifically in cancer-associated fibroblasts (CAFs). Targeting SUMOylation in CAFs affects their signalling pathways and secretome in a concentration-dependent manner, regulating the protumorigenic properties of CAFs.

Keywords: Breast cancer; CAFs; Ginkgolic acid; SILAC; SUMOylation.

Fig. 1

Modification of CAFs at different levels of SUMOylation affects the aggressiveness of breast cancer cells. a Representative Western blot showing SUMO1 levels in different breast cancer-derived CAF subtypes: NF (normal fibroblasts), Lum (luminal), HER2+ (enriched-HER2+), and TNBC (triple negative breast cancer). b Densitometric analysis of SUMO1 from (a) was performed, and the data were quantified from at least three independent experiments using ImageJ software (ANOVA; Tukey’s post hoc test). c Expression of SUMO2/3-conjugated proteins in different molecular CAF subtypes of breast cancer. d Densitometric analysis of SUMO2/3 from (c) was performed, and the data were quantified from at least three independent experiments using ImageJ software (ANOVA; Tukey’s post hoc test). e Expression of SUMO-conjugated proteins (nSUMO1) in luminal CAFs exposed to GA at different concentrations, was analysed by immunoblotting. f Densitometric analysis of SUMO1 from (e) was performed, and the data were quantified from at least three independent experiments using ImageJ software (t test). g Cell viability after 72 hours of treatment with GA (below) and cell images (above) compared to those of the control group recorded at 10× magnification. h Densitometric analysis of SUMO1 expression after GA treatment was performed, and the results were compared with the levels of SUMO1 in TNBC CAFs (ANOVA; Tukey’s post hoc test). i MCF7, BT474, and MDA-MB-231 breast cancer cell lines were treated with GA, and cell proliferation was assessed using crystal violet staining (ANOVA; Tukey’s post hoc test). j Schematic representation of the experimental design used to collect conditioned media (CM) from CAFs. k MCF7, BT474, and MDA-MB-231 breast cancer cell lines were treated with CM from the indicated fibroblasts, and cell proliferation was assessed using crystal violet staining (ANOVA; Turkey’s post hoc test). The data are presented as the means ± s.d (n ≥ 3), and the significant differences between the groups are indicated as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 and ****P ≤ 0.0001

Fig. 2

The inhibition of SUMO1 activity in luminal CAFs disturbs the oncogenic activity of breast cancer cells. a Transwell migration assays and images of migrated MDA-MB-231 cells at 24 h. Migrated cancer cells were quantified as the number of migrating cells relative to the controls (ANOVA; Tukey’s post hoc test). b Expression of SUMO-conjugated proteins (nSUMO1) in CAF-Lum cells exposed to GA (10 µM and 30 µM), analysed by immunoblotting. c Densitometric analysis of SUMO1 from (b) was performed, and the data were quantified from at least three independent experiments using ImageJ software (ANOVA; Tukey’s post hoc test). d Invasion assay of MDA-MB-231 cells treated with CM from CAFs. The number of stained invasive cancer cells was quantified as the number of invading MDA-MB-231 cells relative to that of the controls (ANOVA; Tukey’s post hoc test). e Representative images of migrating MCF-7 cells. f The migration of MCF-7 cells treated with the indicated CM was determined via a wound healing assay (ANOVA; Sidak’s multiple comparisons test). g Representative Western blot showing Rac1 activity in the MDA-MB-231 BC cell line stimulated with CM from vehicle or GA-treated CAFs. h Relative normalized amounts of Rac1-GTP in (g) quantified by scanning densitometry. The results represent three independent experiments and are presented as the means ± SEMs (ANOVA; Tukey’s post hoc test). i Representative zymogram gels showing the degradation bands of MMP-2 and MMP-9 (metalloproteinases 2 and 9, respectively). j Band quantification from gelatin zymography in (i) as determined by scanning densitometry (ANOVA; Sidak’s post hoc test). The data are presented as the means ± s.e.m (n ≥ 3), and the significant differences between the groups are denoted as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 and ****P ≤ 0.0001

Fig. 3

Genetic depletion of SUMO1 is sufficient to induce the protumorigenic activity of CAFs. a SUMO1 protein levels in control and SUMO1-depleted CAFs were analysed by immunoblotting. b Densitometric analysis of SUMO1 from (a) and quantification of the data from at least three different experiments by scanning densitometry (t test). c Representative Western blot of CAF-Lum extracts after SUMO1 silencing and GA treatment (10 and 30 µM). d Total densitometry of SUMO-conjugated proteins was quantified and normalized to that of the control. e MCF7, BT474, and MDA-MB-231 breast cancer cell lines were treated with CM from SUMO1-depleted CAFs, and cell proliferation was assessed using crystal violet staining (t test). f Representative bright-field images showing that silencing SUMO1 in CAFs resulted in significantly increased migration of MCF7 cells compared with that of nontargeting siRNA controls. g Quantitative analysis of the wound closure rate of MCF7 cells with SUMO1 knockdown. Wound closure is expressed as the remaining area not covered by the cells. h Representative images of MDA-MB-231 migrating cells stimulated with CM from SUMO1-knockdown CAFs determined via the transwell method. i Cell migration was assayed in Boyden chambers and quantified as the number of migrated MDA-MB-231 cells treated with CM from SUMO1-knockdown CAFs relative to that from controls. j Representative zymogram gels showing the degradation bands of MMP-2 and MMP-9. k Band quantification from the gelatin zymography in (j) as determined by scanning densitometry (ANOVA; Sidak’s post hoc test). l SUMO2/3 protein levels in control and SUMO2/3-depleted CAFs were analysed by immunoblotting. m Densitometric analysis of SUMO2/3 from (l) was performed, and the data were quantified from at least three different experiments using scanning densitometry (t test). n MCF7 and o MDA-MB-231 breast cancer cell lines were treated with CM from the indicated fibroblasts, and cell proliferation was assessed using crystal violet staining (t test). p The migration of MDA-MB-231 cells treated with CM from SUMO2/3-knockdown CAFs was analysed in Boyden chambers, and the number of migrated MDA-MB-231 cells was quantified relative to that of the controls. q Representative bright-field images showing that silencing SUMO2/3 in CAFs did not affect the migration of MCF7 cells compared with that in cells transfected with nontarget siRNAs. r Quantitative analysis of the wound closure rate from (q). Wound closure is expressed as the remaining area not covered by the cells. The data are presented as the means ± s.e.m (n ≥ 3), and the significant differences between the groups are denoted as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 and ****P ≤ 0.0001

Fig. 4

Changes in SUMOylation levels affect tumor proliferation and metastasis in vivo. a Schematic representation of the experimental design. Breast cancer cells were treated with CM obtained from CAFs treated with GA at the indicated concentrations. b Representative images of tumors obtained from xenografts 7 days after inoculation. Scale bar: centimetres (cm). c Tumor weight 7 days after inoculation of the chorioallantoic membrane (CAM). Statistical analysis was performed using ANOVA followed by Tukey’s post hoc test. d Quantification of MDA-MB-231 cells in the bone marrow using ALU sequences (ANOVA; Tukey’s post hoc test). The data are presented as the mean ± SD, and each dot represents an independent animal. e Representative immunoblot showing Rac1 activity and the AKT and ERK1/2 signalling pathway components in tumors from the xenograft. T1–T3 are tumors derived from the same condition. f The Rac1-GTP bands were quantified, and the normalized intensities were calculated relative to those of the vehicle (n = 12 tumors per group) (ANOVA; Tukey’s post hoc test). g Quantification of Cyclin D1 expression in tumors from (e). h AKT activity was quantified, and the normalized intensities were calculated relative to those of the controls (n = 12 tumors per group) (ANOVA; Tukey’s post hoc test). i Representative images of patient samples before treatment (upper panel) and after tumorsphere formation (bottom panel). j Quantification of tumorsphere formation from (i). k Effect of GA inhibition on self-renewal ability (as determined by LDA) in two patient samples (n = 3). The data are presented as the means ± s.d (n ≥ 3), and the significant differences between the groups are indicated as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 and ****P ≤ 0.0001

Fig. 5

SILAC analysis identified differential targets that were altered in CAFs after treatment with GA. a Schematic representation of the major steps of SILAC analysis to identify proteins whose expression is altered in response to low and high concentrations of GA. b Volcano plot obtained from SILAC-based quantitative proteomics analysis. The plot shows the significantly differentially abundant proteins in tears according to quantitative proteomics analysis. The red spots show the upregulated proteins, the green spots show the upregulated proteins in CAFs, and the blue spots show the non-dysregulated proteins between the two groups. Abundance ratio (log2) > 1.5, P value < 0.05). c Venn diagram of the differentially expressed proteins that were upregulated in CAFs exposed to low (10 µM) or high (30 µM) concentrations of GA. The number within the overlapping section represents the number of shared differentially expressed proteins across the comparisons, and the nonoverlapping numbers specify the genes unique to each condition. d Network visualization illustrating the interconnected relationships of highlighted proteins, emphasizing the signalling pathways affected in response to GA obtained from the functional enrichment analysis. e, f Dotplot enrichment map showing biological processes associated with SILAC-10 µM GA (e) and SILAC-30 µM GA (f) DEPs. The colour of each dot corresponds to the adjusted p value (p.adj), while its size reflects the number of DEPs associated with the respective pathway. g, h Chord diagram showing the relationships between Gene Ontology (GO:BP) terms and DEPs for SILAC at 10 µM (g) and 30 µM (h). Enrichment analysis was performed on the set of DEPs identified in the SILAC-treated groups at 10 µM and 30 µM (p value cutoff = 0.01). i, j Separate analyses of upregulated and downregulated proteins and their participation in signalling pathways in response to 10 µM GA (i) and 30 µM GA (j): red columns represent signalling pathways associated with downregulated proteins, and blue columns represent signalling pathways associated with upregulated proteins

Fig. 6

Working model of the mechanism of SUMO1 inhibition in breast CAFs. The SUMOylation machinery is crucial for cancer-associated fibroblasts (CAFs) to regulate protumorigenic properties and functions. Reducing SUMOylation by treating CAFs with low concentrations of GA results in a secretome that enhances tumor development by promoting proliferation and dissemination. However, more extensive inhibition of SUMOylation results in a secretome that suppresses tumor development