Inhibition of Aurora kinase A suppresses trained immunity in macrophages.

(A) BMDMs were trained with β-glucan at a dosage of 50 μg/mL in the presence of different concentration of alisertib for 24 h. The viability of BMDMs was measured by CCK8 assay. (B) Supernatant levels of IL-6 (left) and TNF (right) in trained BMDMs with alisertib (0.5 μM or 1 μM), followed by restimulation with LPS (100 ng/mL) for 24 h. (C) qPCR analysis of relative mRNA expression of Il6 and Tnf in trained BMDMs with alisertib (0.5 μM or 1 μM), followed by restimulation with LPS (100 ng/mL) for 6 h. Actb served as a reference gene. (D) Immunoblotting analysis of AurA phosphorylation after the treatment of β-glucan (50 μg/mL) with or without alisertib (1 μM) for 90 min. (E) Immunoblotting analysis of AurA in BMDMs transfected with siRNA targeting AurA for 48 h. (F) The BMDMs were firstly transfected with siRNAs for 48 hours and then stimulated with β-glucan (50 μg/mL). Supernatant levels of IL-6 and TNF were detected by ELISA after 3 days rest and restimulation with LPS (100 ng/mL) for 24 h. (G) The BMDMs was stimulated with β-glucan (50 μg/mL) together with AurA knockdown or alisertib, followed by a rest for 3 days and restimulation with MC38 cell culture supernatant for 48 h. (H) Graphical outline of in vivo training model (3 mice per group). (I) Supernatant levels of IL-6 (left) and TNF (right) in trained BMDMs as shown in H. Data are presented as the mean ± SEM (except mean ± SD for in vivo training in I). P values were derived from one-way ANOVA with Dunnett’s multiple-comparison test (A, F, G, I), compared with only β-glucan stimulation group; or two-way ANOVA with Tukey’s multiple-comparison test (B, C), compared with every other group. In D, similar results were obtained from three independent experiments. Related to Figure 1— figure supplement 1, Figure 1—source data 1-2.

Aurora kinase A inhibition remodels chromatin landscape of inflammatory genes.

(A) Principal component analysis (PCA) of gene peaks in ATAC-seq (n=2 mice per group). (B) GO enrichment analysis of erased peaks by alisertib in trained BMDMs. (C) Representative motifs in the erased (n=15,431) and written (n=19,733) peaks respectively. (D and E) Genome browser views of ATAC-seq signal of representative genes inhibited by alisertib including Cxcl2, Il1a, Tnf and Il6 (D) and representative genes enhanced by alisertib including Mrc1 and Chil3 (E). (F and G) KEGG enrichment of differentially expressed genes in trained BMDMs rechallenged with LPS; alisertib downregulated genes (F) and upregulated genes (G) were mapped into KEGG respectively. Related to Figure 2—figure supplement 1.

Aurora kinase A inhibition decreases glycolysis and SAM level. (A and

B) Extracellular acidification rate (ECAR) in BMDMs with different treatments after a glycolysis stress test upon sequential addition of glucose (Gluc, 10 mM), oligomycin (Oligo, 1 μM), and 2-deoxyglucose (2-DG, 50 mM), as indicated (A); basal glycolysis and maximal glycolysis (B). (C, F and G) LC–MS/MS measurements of fumarate (C), serine and SAM (F), SAH and HCY (G) in trained BMDMs treated with vehicle or alisertib for 24 h. (D) BMDMs were trained with β-glucan (50 μg/mL) with or without alisertib (1 μM) for 24 h. The BMDMs were collected for RNA extraction and followed by RNA-seq (n=2 mice per group). The TOP 10 enriched pathways identified by KEGG enrichment analysis of differentially expressed genes (Foldchange >1.2, FDR<0.05) by comparing trained BMDMs with or without alisertib. (E) Intracellular levels of glutathione in trained BMDMs with or without alisertib for 24 h. The level was normalized to untrained BMDMs. (H) Western blot analysis of GNMT in trained BMDMs treated with vehicle or alisertib for 24 h. β-actin was used as a loading control; * showed the position of GNMT protein. (I) Western blot showing GNMT protein levels in wild type BMDMs that were transfected with siRNA targeting GNMT. (J) LC–MS/MS measurements of SAM and SAH in trained BMDMs treated by alisertib together with or without knockdown of GNMT. The SAM/SAH ratio is calculated by SAH normalization. P values were derived from two-tailed t-test. (K) Supernatant levels of IL-6 and TNF in trained BMDMs with AurA inhibition by alisertib or by siRNAs targeting AurA or GNMT. N=3 per group. Data are presented as the mean ± SEM. P values were derived from one-way ANOVA with Tukey’s multiple-comparison test (B, K) compared with every other group or with Dunnett’s multiple-comparison test (C, E-G) compared with only β-glucan stimulation group. In H and I, similar results were obtained for three independent experiments. Related to Figure 3—figure supplement 1, Figure 3—source data 1-2.

Inhibition of Aurora kinase A impairs histone trimethylation at H3K4 and H3K36.

(A) Western blot analysis of histone methylation levels in trained BMDM treated with vehicles or alisertib. Histone 3 (H3) was used as a loading control. BMDMs were trained with β-glucan (50 μg/mL) with or without alisertib (1μM) for 24 h, then BMDMs were washed and cultured in fresh medium for 3 days, followed by protein extraction and Western blot analysis. (B) ChIP-qPCR analysis of H3K4me3 and H3K36me3 enrichment in Il6 and Tnf promoter regions in trained BMDMs treated with vehicles or alisertib for 24 h and rest for 3 days. N=3 per group.(C) Western blot analysis of total H3K4me3 and H3K36me3 levels upon GNMT knockdown in BMDMs. The BMDMs were transfected with siRNA targeting GNMT for 48 h, followed by β-glucan (50 μg/mL) with or without alisertib (1 μM) treatment for 24 h. Then the BMDMs were washed and cultured in fresh medium for 3 days and the protein was extracted for Western blot analysis. Data are presented as the mean ± SEM. P values were derived from one-way ANOVA with Tukey’s multiple-comparison test compared with every each other. In A and C, similar results were obtained for three independent experiments. Related to Figure 4—figure supplement 1, Figure 4—source data 1-2.

Aurora kinase A regulates GNMT through transcription factor FOXO3.

(A) Protein level of GNMT in BMDMs with FOXO3 knockdown was detected by Western blot; * showed the position of FOXO3 band. (B and C) Western blot analysis of GNMT downregulation by siFoxo3 in trained BMDMs with AurA inhibition. BMDMs were transfected with siRNA targeting FOXO3 for 48 h, followed by β-glucan (50 μg/mL) and alisertib (1 μM) for 24 h (B); BMDMs were co-transfected with siRNAs targeting FOXO3 and AurA for 48 h, followed by β-glucan (50 μg/mL) stimulation for another 24 h (C). (D) Supernatant levels of IL-6 in BMDMs. The cells were treated with β-glucan (50 μg/mL) and alisertib (1 μM) after transfection of siRNAs targeting FOXO3. (E) Western blot analysis of phosphorylation level of FOXO3 at ser 315 in BMDMs treated with β-glucan (50 μg/mL) with or without alisertib (1 μM) for 12 h. (F) Immunofluorescence staining of FOXO3 in BMDMs after 12 h β-glucan stimulation with or without alisertib. Scale bars: 10 μm (left). The nuclear localization of FOXO3 was compared by calculating the ratio of mean nuclear intensity to cytoplasmic intensity, and the representative data (right) showed the mean intensity of counted macrophages. (G) Western blot analysis of AKT-mTOR-S6 pathway in β-glucan-trained BMDM. BMDMs were transfected with siRNA targeting AurA for 48 h, followed by β-glucan stimulation for 6 h (left); BMDM was trained with β-glucan in the absence or presence of alisertib for 6 h (right). (H) Supernatant levels of IL-6 and TNF in BMDMs. The trained cells were treated with siRNA targeting AurA or alisertib in combination with an mTOR agonist, MHY1485 (2 μM), and restimulated with MC38 culture supernatant for 48 h. Data in D and H are representative of three independent experiments and presented as the mean ± SEM. P values were derived from one-way ANOVA with Tukey’s multiple-comparison test compared with each other in D and H, or with Dunnett’s multiple-comparison test in F compared with β-glucan only group. In A-C and E-G, similar results were obtained for three independent experiments. Related to Figure 5—source data 1-3.

Alisertib abrogates the anti-tumor effect induced by trained immunity.

(A) Experimental scheme of tumor model. 6∼8 weeks old mice was injected with β-glucan together with vehicle or alisertib, followed by subcutaneous inoculation of MC38 cells (1×106 cells/mouse, n=5 per group). (B) The MC38 tumor growth curves of mice in A. (C) Experimental scheme of bone marrow transplantation model. CD45.1+ mice (n=3 per group) as donor were treated as indicated, and CD45.2+ mice (n=7 per group) as recipient received 4×106 BM cells from CD45.1+ mice at day 8 post irradiation with 9 Gy. After 4 weeks, CD45.2+ mice were inoculated with MC38 cells. (D) The MC38 tumor growth curves of mice in C. (E) Flow cytometric analysis of myeloid cells (CD45+CD11b+) and macrophages (CD45+CD11b+F4/80+) in MC38 tumors in A. Gating strategy was shown in Figure 6-figure supplement 1D. (F) Co-immunofluorescence staining of DAPI, F4/80 and GNMT in tumor section from A; Scale bars: 20 μm. (G) Flow cytometry analysis for intracellular phospho-S6 in macrophages from tumors in A. (H) Tumor tissue was lysed and the lysates were collected for detection of IL-6 by ELISA. (I) FACS analysis of intracellular IL-6 in tumor-infiltrated myeloid cells and macrophages from tumors in A. Data are represented as the mean ± SD. P values were derived from one-way ANOVA (E, F, G, H, I) or two-way ANOVA (B and D) with Dunnett’s multiple-comparison test with data from mice treated with β-glucan only as control. Related to Figure 6—figure supplement 1, Figure 6—source data 1.

Targeting aurora A inhibits β-glucan-induced trained immunity.

(A) Schematic of the assay protocol for the drug screening in BMDMs. (B) Fold change of IL-6 production showing the outcomes of the entire drug screening, including inhibitors targeting Aurora kinase, JAK/STAT pathway, PI3K/AKT/mTOR pathway, histone methyltransferases (HMT), lysine-specific demethylases (KDM), histone deacetylases (HDAC), Sirtuin (SIRT), Bromodomain and PARP families. (C) Fold change of IL-6 production inhibited by AurA inhibitors compared to that of β-glucan only group. (D) BMDMs were trained with β-glucan (50μg/mL) in the presence of Tozasertib (1 μM), alisertib (1 μM) or MLN8054 (1 μM) for 24 h, followed by LPS (100 ng/mL) stimulation for 6 h after a rest for 3 days. (E) Supernatant levels of IL-6 in trained J774A.1 cells and THP-1 cells. J774A.1 cells were transfected with AurA-specific siRNAs followed by β-glucan (50 μg/mL) stimulation. After rest for 3 days, trained J774A.1 cells was counted and seeded into cell culture plate with MC38 supernatant rechallenge for 48 h. THP-1 cells were trained with β-glucan (50 μg/mL) for 24 h, and were centrifuged and washed once to remove medium. The THP-1 cells were then cultured with fresh medium containing 10 ng/mL PMA for 48 h and rest 1 day, followed by LPS rechallenge (100 ng/mL). Data in D and E are representative of three independent experiments and presented as the mean ± SEM. P values were derived from one-way ANOVA with Dunnett’s multiple-comparison test, compared with β-glucan stimulation only group. Related to Figure 1.

Aurora kinase A inhibition suppresses the expression of transcription factors involved in inflammation activation.

(A) In vivo training model in C57BL/6J mice with intraperitoneal injection of β-glucan (2 mg per mice) and daily administration of alisertib (30 mg/kg/d) for 7 days (n=2 mice per group). (B) Heatmap from RNA-seq analysis showing the differentially expressed transcription factors (DETFs) from mice treated as described in A. (C) GO enrichment analysis of differentially expressed transcription factors (DETFs). (D) Multiplex immunoassay measuring 18 cytokines/chemokines in supernatant from trained BMDMs as described in A, which were rechallenged with LPS (100 ng/mL) for 6 h. Related to Figure 2.

Alisertib inhibits glucose incorporation into glycolysis and TCA cycle.

(A) Oxygen consumption rate (OCR) in BMDMs with different treatment after a mito stress test upon sequential addition of oligomycin (oligo, 1 μM), FCCP (1 mM), and rotenone/antimycin A (0.5 μM), as indicated; (B) Mass labelling of trained BMDMs with U-13C-glucose in the absence or presence of alisertib for 24 h. Training by β-glucan increased incorporation of 13C-glucose into glycolysis and TCA cycle intermediates; this was reversed by alisertib. (C) Peak area of tyrosine and comparing of the sum of peak areas for unlabeled and labeled tyrosine between different treatment groups. (D) Diagram illustrating the cross link between glycolysis, TCA cycle, glutathione and SAM. N=3 per group. Data are presented as the mean ± SEM. and P values were derived from one-way ANOVA (B) or two-way ANOVA (C) with Dunnett’s multiple-comparison test. Related to Figure 3.

Inhibition of Aurora kinase A impairs histone trimethylation at H3K4 and H3K36.

Quantification of H3K4me1, H3K9me3, H3K36me3, H3K4me3 and H3K27me3 protein levels was determined by Image Lab software; data are represented as mean ± SEM of three independent experiments. P values were derived from one-way ANOVA with Dunnett’s multiple-comparison test compared with β-glucan stimulation only group. Related to Figure 4.

Alisertib abrogates the anti-tumor effect induced by trained immunity.

(A) Individual tumor growth curves (n=5 per group). (B) Representative FACS analysis for CD45.1+ and CD45.2+ cells in the spleen of a chimeric mouse at the end of experiment (day 19 after tumor inoculation). (C) Individual tumor growth curves of chimeric mice (n=7 per group). (D) The gating strategy used for analyzing tumor-infiltrated myeloid cells and macrophages. (E) Tumor tissue from Figure 6A was lysed and the lysates were collected for detection of IL-1β and IL-12p70 (n=5). Data are represented as mean ± SD. P values were derived from one-way ANOVA with Dunnett’s multiple-comparison test with data from mice treated with β-glucan only as control. Related to Figure 6.