De novo MYC addiction as an adaptive response of cancer cells to CDK4/6 inhibition

Abstract Cyclin‐dependent kinases (CDK) are rational cancer therapeutic targets fraught with the development of acquired resistance by tumor cells. Through metabolic and transcriptomic analyses, we show that the inhibition of CDK4/6 leads to a metabolic reprogramming associated with gene networks orchestrated by the MYC transcription factor. Upon inhibition of CDK4/6, an accumulation of MYC protein ensues which explains an increased glutamine metabolism, activation of the mTOR pathway and blunting of HIF‐1α‐mediated responses to hypoxia. These MYC‐driven adaptations to CDK4/6 inhibition render cancer cells highly sensitive to inhibitors of MYC, glutaminase or mTOR and to hypoxia, demonstrating that metabolic adaptations to antiproliferative drugs unveil new vulnerabilities that can be exploited to overcome acquired drug tolerance and resistance by cancer cells.

A D B C Figure EV2. Intracellular glutathione, ROS, NADPH levels and fatty acid synthesis in CDK4/6kd and control cells.
A Total intracellular glutathione content normalized to cell number. B Intracellular ROS levels determined by flow cytometry. Data are expressed as percentages of mean fluorescent intensity (MnX) relative to control cells. C NADP and NADPH levels quantified by a colorimetric assay using the NADP/NADPH Quantification Kit (MAK038, Sigma-Aldrich) and normalized to cell number. D Dynamic accumulation of isotopologues in palmitate and stearate after 24 h incubation with 10 mM [1,2-13 C 2 ]-glucose (top) or 2 mM [U-13 C]-glutamine (bottom), suggesting an impaired fatty acid synthesis in CDK4/6-kd cells.
A Mass spectrometry analysis of the two c-MYC peptides that were found to be phosphorylated by CDK4/6-Cyclin D1 complexes. Kinase assays were performed with full-length recombinant human c-MYC adding CDK4-Cyclin D1 or CDK6-Cyclin D1 complexes and using (right) or not using (left) unlabeled ATP followed by a mass spectrometry phosphosite profiling assay. Samples from the kinase assay were digested with trypsin and analyzed with LC-MS/MS. Data were blindly assessed. B Glucose and glutamine consumption and lactate and glutamate production rates in HCT116 cells expressing a T58A MYC mutant. Media were collected after 24 h of incubation with fresh media and results are normalized to cell number.
Data information: Bars correspond to mean AE SD (n = 3). Statistically significant differences between T58A MYC mutant-expressing and control cells were determined by two-tailed independent sample Student's t-tests and are indicated as P < 0.01 (**), and P < 0.001 (***). A Western blotting analysis of total protein fractions of CDK4/6-inhibited and control cells. Protein extracts were obtained after incubating HCT116 cells with PD0332991 (2 lM) or vehicle for 96 h. B Western blotting analysis of total protein fractions of CDK4/6-kd and control cells. C Polyamine metabolism in CDK4/6-kd and control cells. Top, schematic representation of the polyamine metabolism and the urea cycle. The metabolites and enzymes which were found to be increased are depicted in red and those found to be downregulated in green. Bottom left, quantification of spermidine and putrescine polyamines by GC/MS. Results are shown as percentage of polyamine metabolite relative to control cells. Bottom right, qRT-PCR measures of ODC and SAT1 gene expression. Data are normalized to cyclophilin A and expressed as fold change of mRNA relative to non-targeting siRNA-treated cells. Altogether, polyamine accumulation and changes in expression of enzymes involved in polyamine metabolism suggest an increase in polyamine synthesis in CDK4/6-kd cells. D PIK3R3 and CDKN1B gene expression levels determined by qRT-PCR in CDK4/6-kd and control cells 96 h after transfection. Results are normalized to cyclophilin A and expressed as fold change of mRNA relative to expression of non-targeting siRNA-treated cells. E SLC25A13 gene expression levels determined by qRT-PCR. Results are normalized to cyclophilin A and expressed as fold change of mRNA relative to expression of non-targeting siRNA-treated cells. F HIF-1a, HIF-2a, and PHD2 protein levels were determined by Western blotting in CDK4/6-kd and control cells. G Cell cycle distribution of NCM460 cells after PD0332991 treatment or serum deprivation. H Western blotting analysis of total protein fractions of control, PD0332991-treated and serum-deprived NCM460 cells. I Comparative extracellular metabolic fluxes for control, PD0332991-treated, and serum-starved NCM460 cells. Glucose and glutamine consumption and lactate and glutamate production rates were obtained after 24 h of incubation with fresh media and normalized to cell number.

Molecular
Data information: Bars represent mean AE SD (n = 3). All experiments were performed 96 h after siRNA transfection or PD0332991 treatment, as specified in each case. b-actin was employed as a protein loading control in all Western blotting experiments. Statistically significant differences between CDK4/6-kd and control cells or serum-starved and control cells were determined by two-tailed independent sample Student's t-tests and are indicated as P < 0.01 (**) and P < 0.001 (***). n.s., statistically non-significant. Source data are available online for this figure. A The combination of PD0332991 with the GLS1 inhibitor CB-839 is not toxic to BJ fibroblasts. BJ normal human fibroblasts, HCT116 colorectal cancer cells, and MCF-7 and SK-BR-3 breast cancer cells were cultured at the indicated concentrations of inhibitors for 96 h, and cell proliferation was determined by Hoechst staining.
Results are shown as percentage of proliferation relative to untreated cells (mean AE SD of n = 6). B Effects on spheroid growth of MCF-7 and SK-BR-3 breast cancer cells by CDK4/6 and GLS1 inhibitors. Cells were grown in low-attachment plates and incubated with the indicated combinations of PD0332991 (2 lM) with BPTES (10 lM) or CB-839 (10 lM) for 10 days. Top, representative images of spheroids after staining with MTT. Bottom, quantifications of spheroid volume (mean AE SD of n = 4), represented as percentages of total spheroid volume relative to untreated cells. Spheroids were scored by image acquisition and through spheroid area and volume quantification with ImageJ software. Significance was determined by Kruskal-Wallis and twotailed independent sample Student's t-tests. Statistically significant differences between CDK4/6-inhibited and control cells were indicated at P < 0.001 (***), while differences between treatment (BPTES or CB-839) and the corresponding control (PD0332991-treated cells or untreated cells in complete medium) were shown at P < 0.001 (###) for CDK4/6-inhibited cells and at P < 0.01 ( ¶ ¶) for control cells.