Translational and HIF-1α-Dependent Metabolic Reprogramming Underpin Metabolic Plasticity and Responses to Kinase Inhibitors and Biguanides

Abstract

There is increasing interest in therapeutically exploiting metabolic differences between normal and cancer cells. We show that kinase inhibitors (KIs) and biguanides synergistically and selectively target a variety of cancer cells. Synthesis of non-essential amino acids (NEAAs) aspartate, asparagine, and serine, as well as glutamine metabolism, are major determinants of the efficacy of KI/biguanide combinations. The mTORC1/4E-BP axis regulates aspartate, asparagine, and serine synthesis by modulating mRNA translation, while ablation of 4E-BP1/2 substantially decreases sensitivity of breast cancer and melanoma cells to KI/biguanide combinations. Efficacy of the KI/biguanide combinations is also determined by HIF-1α-dependent perturbations in glutamine metabolism, which were observed in VHL-deficient renal cancer cells. This suggests that cancer cells display metabolic plasticity by engaging non-redundant adaptive mechanisms, which allows them to survive therapeutic insults that target cancer metabolism.

Keywords: HIF-1α; biguanide; cancer; kinase inhibitor; mRNA translation; mTORC1; metabolic plasticity; metabolism; non-essential amino acids.

Figure 1:. KIs and phenformin synergistically inhibit cell proliferation and suppress tumor growth in vivo

A. NMuMG and NT2197 cells were treated as indicated for 72h, proliferation was monitored by BrdU incorporation and expressed as a percentage of non-lapatinib treated cells. B. Combination Index 50 (CI₅₀) for lapatinib and phenformin was determined by isobologram method (corresponding proliferation curves are shown in Figure 1A and Figures S1C–E). (C) A375, (D) K562 and (E) HCT116 cells were treated with phenformin and/or PLX4032 (C), imatinib (D) or lapatinib (E), as indicated for 72h. Proliferation was estimated by BrdU incorporation. PLX4032 (C), imatinib (D) or lapatinib (E) interactions with phenformin were evaluated by the isobologram method. Corresponding proliferation curves are shown in Figure S1M–R, Figure S2A–B, F–G. A. –E. The data are presented as means +/− the SD (3 independent experiments-3 technical replicates averaged in each). F–G. 50,000 NT2197 cells were injected into two mammary fat pads of Nu/Nu mice. When tumors reached 100–200 mm³ mice were randomly distributed into 4 groups of 5 mice and treated with phenformin (50mg/kg), lapatinib (50mg/kg) or both every 24h for 8 days. (F) Tumor growth was measured as described (Ursini-Siegel et al., 2007) every two days until the control tumors reached 500mm³. The data are presented as means +/− the SEM (n=10 tumors) [*p<0.05,** <0.01 and ***<0.001 (two-way ANOVA)]. Mice were sacrificed on day 8, 4h after the last treatment. (G) Tumor sections from 7 tumors/group were stained with Ki67 antibody. (**P<0.001; ANOVA). Representative images are shown in Figure S2I. See also Figure S1 and Figure S2.

Figure 2:. Kinase inhibitors oppose phenformin-induced metabolic changes

A. Levels of indicated metabolites in NT2197 cells treated for 24 hours with phenformin (600μM), lapatinib (600nM) or combination thereof were determined by GC–MS. Data are shown as mean +/− the SD. *P<0.05 (2-way ANOVA; 2 independent experiments; in each 3 technical replicates were averaged). (B-D) NT2197 cells treated with vehicle (DMSO), phenformin (250μM) and/or lapatinib (600nM) for 24h. B. Glucose uptake was monitored by subtracting glucose concentration in media with cells vs. cell-free media, and normalized to DMSO control. Data are shown as mean +/− the SD. *p<0.05 and ** P<0.01 (2-way ANOVA; 2 independent experiments-3 technical replicates averaged in each). C, D. Intracellular lactate/pyruvate (C) and αKG/citrate ratios (D) were determined by GC/MS and normalized to control. Data are shown as mean +/− the SD. *p<0.05, ** P<0.01 and ***<0.001 (2-way ANOVA; 2 independent experiments-3 technical replicates averaged in each). E. NT2197 cells treated as in (B–D) for 24h and incubated with ¹³C₅-glutamine. Absolute malate (m+3) ion levels are shown. Data are representative of 2 independent experiments. AU: arbitrary units. Right: schematic of the ¹³C incorporation into metabolites. F–H. A375 cells were treated with phenformin (1.5mM) and/or PLX4032 (200nM) for 24h. F. Glucose uptake was normalized to DMSO control. Data are shown as mean +/− the SD. *p<0.05 (2-way ANOVA; 3 independent experiments-3 averaged technical replicates each). G. Intracellular lactate/pyruvate and H. αKG/citrate ratio. G, H. Metabolites were measured by GC/MS and normalized to DMSO control. Data are shown as mean +/− the SD. *p<0.05, ** P<0.01 and ***<0.001 (2-way ANOVA; 2 independent experiments-3 technical replicates averaged in each). I. NT2197 cells treated with vehicle (DMSO), 3-PO (15μM) or lapatinib (300nM) for 24h, were incubated with ¹³C₆-glucose for 90min. Absolute lactate (m+3) ion amounts are shown, normalized to control. Data are shown as mean +/− the SD from 3 independent experiments. J. NT2197 cells were treated as indicated for 72h. Proliferation was monitored by BrdU incorporation and expressed as percentage of non-phenformin treated cells. The data are presented as mean values +/− the SD (2 independent experiments-3 technical replicates averaged in each). See also Figure S2 and S3.

Figure 3:. Kinase inhibitors and phenformin collaboratively inhibit mTORC1 NT2197

(A, D), A375 (B, E), and K562 (C) cells were treated as indicated. Levels and phosphorylation status of indicated proteins were determined using western blot. β-actin was a loading control. These data are representative of at least 2 independent replicates.

Figure 4.. 4E-BP1/2 determine the efficacy of the lapatinib/phenformin combination

A. Control (CTR) or CRISPR 4E-BP1 and 2 depleted (4E-BP DKO clone 17; cl17) NT2197 cells were treated as indicated for 6h. The expression and phosphorylation status of indicated proteins were determined by western blot. β-actin was a loading control. Data are representative of 2 independent replicates. B, C. CTR or 4E-BP DKO cl17 NT2197 cells were treated as indicated for 72h. Proliferation was monitored by BrdU incorporation and was expressed as a percentage of a vehicle (DMSO) treated cells. The data are presented as mean values +/− the SD (3 independent experiments-3 technical replicates averaged in each). D,E. CTR or indicated clones of 4E-BP DKO NT2197 cells were treated with phenformin (250μM), lapatinib (600nM) or combination thereof for 4h. The interactions between eIF4E/eIF4G1 and eIF4E/4E-BP1 were assessed by proximity ligation assay (D) and m⁷GDP pull-down (E). D. ** p<0.001 (ANOVA; posthoc Tukey test from 10 different fields). Representative images are shown in Figure S5A and Data S1. E. Amounts of the indicated proteins in the input or pull-down were determined by western blot; β-actin was a loading control (input) and was used to exclude contamination (m⁷GDP pull-down). F. CTR or 4E-BP DKO cl17 cells were treated as shown. The expression of indicated proteins were determined by western blot. β-actin was a loading control. (E, F) experiments are representative of 2 independent replicates. See also Figure S4 and S5.

Figure 5.. Phenformin/lapatinib combination suppresses translation of mRNAs which encode NEAA biosynthetic enzymes

A. CTR or 4E-BP DKO cl17 NT2197 cells were treated with vehicle (DMSO) or combination of phenformin (250μM) and lapatinib (600nM) for 4h. Sub-polysomal, light-and heavy-polysome fractions were obtained by ultracentrifugation using 5–50% sucrose gradients. Position of 40S and 60S ribosomal subunits, monosome (80S) and polysomes in the absorbance profiles (254nm) are shown. B. Amount of indicated mRNAs in polysome fractions were determined by RT-qPCR [as described; (Miloslavski et al., 2014)]. Corresponding total mRNA levels are shown in Figure S5B–C. Data are representative of 3 independent experiments (3 technical replicates each) and are expressed as a percentage of mRNA in each fraction +/− SD. See also Figure S5.

Figure 6.. 4E-BP1/2 determine the lapatinib/phenformin combination efficacy

A. Control (CTR) or 4E-BP DKO cl17 NT2197 cells were treated with phenformin (250μM), lapatinib (600nM) or combination thereof for 24h. Levels of indicated metabolites were determined by GC–MS. Additional metabolites are shown in Figure S5D. Data represent means +/− SD (2 independent experiments-3 technical replicates averaged in each). B. Cells from (A) were incubated with ¹³C₆-glucose for 20 min. Stable isotope tracer analyses were performed and the absolute levels of aspartate (m+3) and serine (m+3) ions are shown. Data are representative of 2 independent experiments. AU: arbitrary units. More isotope tracing data are shown in Figure S5E. Right: schematic of the ¹³C incorporation into metabolites. C, D. NT2197 CTR cells were treated as indicated for 72h. Where indicated media were supplemented with aspartate (10mM) (C) or asparagine (2mM) (D). Proliferation was monitored by BrdU incorporation and expressed as a percentage of control treated cells. The data are presented as means +/− the SD (2 independent experiments-3 technical replicates averaged in each). See also Figure S5 and S6.

Figure 7.. The phenformin/lapatinib combination efficacy is influenced by HIF-1α

A. VHL-null RCC4 cells expressing either empty vector (RCC4-mock) or VHL (RCC4-VHL) were treated as shown for 6h. Expression of indicated proteins was determined by western blot. β-actin was a loading control (representative data from 2 independent experiments). B. Cells from (A) were treated with the indicated concentration of phenformin and lapatinib for 72h. Proliferation was monitored by BrdU incorporation and was expressed as a percentage of vehicle (DMSO)-treated cells. The data are presented as means +/− the SD (3 independent experiments-3 technical replicates averaged in each). C. Cells described in (A) were incubated with ¹³C₅-glutamine for 30min, 1h, and 2h. Stable isotope tracer analysis was performed, and the levels of the indicated ion amounts (fraction of the total pool) are shown. Data are representative of 2 independent experiments. Bottom right panel: schematic of the ¹³C incorporation into metabolites. For additional data see Figure S7E. D. Schematic showing the utilization for glutamine in RCC4 cells depending on VHL status/HIF-1α levels. See also Figure S6 and S7.