Metabolomics analysis of MDA-MB-231 cells upon WMK-1 treatment

MDA-MB-231 cells were treated with compounds, including DMSO (Vehicle), ML210 (for 2h, 4 μM), WMK-1 (10 μM, for 5h), ZL-1-180 (10 μM, for 5h), in the presence or absence of Lip-1 (1 μM). For metabolomics analysis, 3 µL of the prepared tissue lipidomics samples were loaded into an ACQUITY UPLC BEH amide column (100 × 2.1 mm, 1.7 μm, 130 Å, Waters, Cat# 186004741), and separated chromatographically with a 25-min gradient elution at a flow rate of 0.3 mL/min. The mobile phase for the positive ion acquisition mode contains H2O with 0.1% (v/v) formic acid and 10 mM ammonium formate for Phase A, and ACN/H2O (95:5) with 0.1% (v/v) formic acid and 10 mM ammonium formate for Phase B. The mobile phase for the negative ion acquisition mode is similar to the buffers for the positive ion mode except excluding formic acid in the buffer. Metabolomics data analysis: The Compound Discoverer v3.2 software (Thermo Fisher Scientific) was used for LC-MS/MS metabolomics data processing including peak alignment, background filtering, peak calling, and compound identification. We applied an adaptive curve alignment model and filtered out metabolites with intensities lower than the minimum cutoff of 100,000 in both positive and negative ion mode datasets. The remaining metabolites were used for further analysis. Adduct ions including [M+H]+, [M+Na]+, [M+K]+, [M+NH4]+and [M-H]- were identified by assigning predicted chemical composition to each ion based on the exact mass within a ±5 ppm deviation and the isotopic pattern with intensity tolerance < 30%. For metabolite annotation, mzCloud search was used for mapping the MS/MS spectra with the endogenous metabolite library. A cosine-based identity search procedure with match activation type was employed during the DDA search method to facilitate accurate matching of fragmentation patterns. During the mapping of activation energy, a tolerance within 20, and a minimum intensity threshold for filtering low-intensity signals were employed. Similarity search was turned off during annotation. We used a match factor threshold of 60 during the metabolite identification. Metabolites not detected in the samples were substituted with background noise signals. A minimum of 50% QC coverage and a maximum QC area RSD of 30% were exerted during the QC correction. Lastly, metabolites whose intensities are lower than five times that of the corresponding blank samples were filtered out. Metabolomics data normalization: For downstream analysis, the raw intensities of metabolites were normalized via dividing by the median metabolite intensity among the corresponding sample and scaled with a scaling factor. Then, the fold change of each metabolite between compound treatment (ML210, WMK-1, ZL-1-180 or Lip-1 rescue) and vehicle treatment groups was calculated. P-value was calculated using a two-tailed, non-parametric, unpaired Student’s t-test. Metabolites with fold change >= 2 and P-value <= 0.05 were considered as significantly changed metabolites after WMK-1 treatment and were then used for plotting.