Myeloid PTEN loss affects the therapeutic response by promoting stress granule assembly and impairing phagocytosis by macrophages in breast cancer

Abstract

Breast cancer (BRCA) has become the most common type of cancer in women. Improving the therapeutic response remains a challenge. Phosphatase and tensin homologue deleted on chromosome 10 (PTEN) is a classic tumour suppressor with emerging new functions discovered in recent years, and myeloid PTEN loss has been reported to impair antitumour immunity. In this study, we revealed a novel mechanism by which myeloid PTEN potentially affects antitumour immunity in BRCA. We detected accelerated stress granule (SG) assembly under oxidative stress in PTEN-deficient bone marrow-derived macrophages (BMDMs) through the EGR1-promoted upregulation of TIAL1 transcription. PI3K/AKT/mTOR (PAM) pathway activation also promoted SG formation. ATP consumption during SG assembly in BMDMs impaired the phagocytic ability of 4T1 cells, potentially contributing to the disruption of antitumour immunity. In a BRCA neoadjuvant cohort, we observed a poorer response in myeloid PTEN^low patients with G3BP1 aggregating as SGs in CD68+ cells, a finding that was consistent with the observation in our study that PTEN-deficient macrophages tended to more readily assemble SGs with impaired phagocytosis. Our results revealed the unconventional impact of SGs on BMDMs and might provide new perspectives on drug resistance and therapeutic strategies for the treatment of BRCA patients.

Fig. 1. Myeloid PTEN is related to SGs in macrophages and the antitumour immune response in BRCA patients.

A Survival curve for patients grouped according to the optimal cut-off for PTEN expression. The cut-off points used are displayed in Supplementary Fig. 1A and 1B. B Analyses of the correlation between therapeutic sensitivity and tumour PTEN expression. n = 27 patients (16 patients in the sensitive group and 11 in the resistant group based on the Miller–Payne grading system). PTEN and G3BP1 expression in CD68-positive macrophages was assessed by the MOD (immunohistochemical modification) value (Supplementary Table 2). For the histograms, the data are presented as means ± SDs; p values were determined by t tests; asterisks indicate significant differences from the control; * p < 0.05, ** p < 0.01, *** p < 0.001. For the correlation curves, the R² and p value were determined by linear regression analysis. C Analyses of the correlation between therapeutic sensitivity and myeloid PTEN expression. D IHC staining of CD68 and PTEN in tissue sections from BRCA patients in the neoadjuvant treatment cohort. Macrophages were identified as CD68-positive cells. The scale bar is 25 μm. E Immunofluorescence staining of G3BP1 and CD68 in sections from the BRCA cohort. The macrophages were labelled with CD68 (red); arrowheads indicating G3BP1 (green) aggregation indicate the formation of SGs. The scale bar is 10 μm.

Fig. 2. SG assembly is accelerated in PTEN-deficient BMDMs.

A Immunofluorescence staining of SGs (indicated by the aggregation of G3BP1) in BMDMs from Pten^f/f or Pten^mKO mice. BMDMs were stimulated with LPS (100 ng/mL) for 4 h and induced with arsenite (90 μM) for 1 h. G3BP1 was labelled with Alexa 488. The scale bar is 10 μm. B and C SG counts in BMDMs. The data are presented as means ± SDs; n = 3 independent experiments, and p values were determined by t tests. D Protein levels of eIF2α and P-eIF2α in BMDMs. E The protein levels of several conventional nucleating proteins in BMDMs without pretreatment. F Immunohistochemical staining for CD68, PTEN and TIAL1 in patients with different myeloid PTEN statuses. The scale bar represents 25 μm. G Correlation analysis between the MOD of myeloid PTEN and myeloid TIAL1 in the BRCA cohort. H Correlation analysis between the MOD of myeloid TIAL1 and the Miller–Payne score in the BRCA cohort.

Fig. 3. PTEN deficiency is accompanied by upregulated EGR1 expression in BMDMs.

A Rank of the log2-fold change in the expression of several conventional SG proteins in BMDMs from Pten^f/f or Pten^mko mice based on the RNA-seq results. B Transcription factor prediction for the potential regulation of TIAL1. C Venn diagram of the predicted TFs of TIAL1 in B and the total TFs among the DEGs in BMDMs from Pten^f/f or Pten^mko mice. D Ranking of log2-fold changes among the 6 potential TFs obtained in D based on the RNA-seq results. E Protein level of EGR1 in BMDMs from Pten^f/f and Pten^mko mice without pretreatment.

Fig. 4. Upregulated EGR1 promotes the transcription of TIAL1 in PTEN-deficient BMDMs.

A ChIP-seq of EGR1 in BMDMs from Pten^f/f and Pten^mko mice. B Diagram of plasmid construction. Wild-type (TIAL1-wt) and Egr1 binding region-mutated (TIAL1-mut) plasmids harbouring the TIAL1 promoter reporter gene were constructed using the PGL3.0 vector. C Dual-luciferase reporter gene assay of TIAL1 and EGR1. D and E ChIP-qpcr analyses, including % of input (D) and fold enrichment (F) using the EGR1 antibody and 3 different primers of TIAL1. F, G KEGG and GO enrichment analyses of peak-related genes identified via ChIP-seq analysis of EGR1 in BMDMs from Pten^f/f and Pten^mko mice.

Fig. 5. ATP consumption by SG assembly impairs phagocytosis by macrophages.

BMDMs were stimulated with LPS (100 ng/mL) for 4 h. SGs were induced by arsenite (90 μM) for 1 h and inhibited by anisomycin (25 µg/ml) for 20 min before arsenite treatment. A~C Coculture of BMDMs and 4T1 cells to assess phagocytosis. BMDMs were labelled with Cell Tracking Dye Kit-Red-Cytopainter, and 4T1 cells were labelled with CFSE. BMDMs and 4T1 cells were cultured at a ratio of 5:1 for 4 ~ 6 h. Cells in A (indicated by white arrows) or in B were defined as BMDMs that underwent phagocytosis. The scale bar is 50 μm. n = 3 independent experiments. The data are presented as means ± SDs; p values were determined by single factor analysis of variance. D Fluorescence of the cytoskeleton and stress granules. The Actin-Tracker Green probe was used to label myosin filaments (F-actin), and Alexa Fluor 594 was used to label G3BP1. The aggregation of G3BP1 indicates the formation of SGs. The arrow indicates cells with SGs. The scale bar represents 10 μm. E Intracellular ATP levels in BMDMs. Cells were treated with 5 mM 2DG for 1 h. The data are presented as means ± SDs; n = 10 wells and 3 independent experiments; the p value was determined by two-tailed unpaired Student’s t test and a multiple comparisons test. F Evaluation of the phagocytosis efficiency of BMDMs towards 4T1 cells. ATP was added at 4 M for 90 min. BMDMs were labelled with Cell Tracking Dye Kit-Red-Cytopainter, and 4T1 cells were labelled with CFSE. The cells were cocultured for 5 h. The data are presented as means ± SDs; n = 3 independent experiments; the p value was determined by a two-tailed unpaired Student’s t test and a multiple comparisons test.

Fig. 6. Phagocytosis impairment by SGs in PTEN-deficient BMDMs is alleviated by ATP supplementation or PAM pathway inhibition.

BMDMs were stimulated with LPS (100 ng/mL) for 4 hours. SGs were induced with arsenite (90 μM) for 1 h; the cells were treated with wortmannin (100 nM), rapamycin (100 nM) or AZD5356 (3 μM) 2 h before SG induction. A Quantification of SGs in BMDMs treated with a PAM pathway inhibitor. The p value was determined by one-way ANOVA; The data are presented as means ± SDs; n = 3 independent experiments. B Intracellular ATP levels in BMDMs from Pten^f/f or Pten^mko mice. The cells were treated with 2DG (5 mM) for 1 h. ATP (4 M) was added for 90 min. The p value was determined by one-way ANOVA; the data are presented as means ± SDs; n = 12 wells and 4 independent experiments in each group. C and D Phagocytosis efficiency of BMDMs in Pten^f/f mice. E and F Phagocytosis efficiency of BMDMs in Pten^mKO mice.