Colon cancer cells evade drug action by enhancing drug metabolism

Abstract

Colorectal cancer (CRC) is the second leading cause of cancer deaths worldwide. One key reason is the lack of durable therapies that target KRAS-dependent disease, which represents approximately 40% of CRC cases. Here, we use liquid chromatography/mass spectrometry (LC/MS) analyses on Drosophila CRC tumour models to identify multiple metabolites in the glucuronidation pathway-a toxin clearance pathway that impacts most drugs-as upregulated in trametinib-resistant RAS/APC/P53 ("RAP") tumours compared to trametinib-sensitive Ras^G12V single mutant tumours. Genetic inhibition of different steps along the glucuronidation pathway strongly reversed RAP resistance to trametinib; conversely, elevating glucuronidation pathway activity was sufficient to direct trametinib resistance in Ras^G12V animals. Mechanistically, pairing oncogenic RAS with hyperactive WNT activity strongly elevated PI3K/AKT/GLUT signalling, which in turn directed elevated glucose uptake and glucuronidation; our data also implicate the pentose phosphate pathway in this process. We provide evidence that this mechanism of trametinib resistance is conserved in a KRAS/APC/TP53 mouse CRC tumour organoid model. Finally, we identify two clinically accessible approaches to inhibiting drug glucuronidation: (i) blocking an initial HDAC1-mediated deacetylation step of trametinib with the FDA-approved drug vorinostat; (ii) reducing blood glucose by the alpha-glucosidase inhibitor acarbose. Overall, our observations demonstrate a key mechanism by which oncogenic RAS/WNT activity promotes increased drug clearance in CRC and provides a practical path towards abrogating drug resistance in CRC tumours.

Fig. 1. Glucuronidation pathway induces trametinib resistance in Drosophila.

Glucuronidation was upregulated in RAP tumours compared with Ras^G12V tumours and led to trametinib resistance. a, d, e, f, g, h Percent survival of transgenic flies to adulthood relative to control flies was quantified in the present or absence of trametinib (1 μM) or UDP-Glc as indicated. a Control (n = 16), Ras^G12V (+DMSO n = 18; +tram n = 37) and RAP ( + DMSO n = 12; +tram n = 10); d, e RAP + GFP (control, +DMSO n = 12; +tram n = 9), RAP +Hex-C-RNAi (+DMSO n = 5; +tram n = 9), RAP + UGP-RNAi (+DMSO n = 6; +tram n = 9), RAP +Sgl-RNAi (+DMSO n = 12; +tram n = 13), RAP +GlcAT-P-RNAi (+DMSO n = 9; +tram n = 8); f Control (+DMSO n = 19; +tram n = 19), Sgl-RNAi (+DMSO n = 15; +tram n = 15) and GlcAT-P-RNAi (+DMSO n = 14; +tram n = 14); g Control (without UDP-Glc n = 12; +0.1 mM n = 12; +0.5 mM n = 6); h Ras^G12V (without UDP-Glc n = 17; +0.1 mM n = 10; +0.5 mM n = 14), error bar is a standard deviation (SD). b A heatmap of LC/MS showed top 50 metabolites. c An overview of the glucuronidation pathway. Transgene expression was induced in Drosophila hindguts by a byn-GAL4 driver. Experiments were performed at 27 °C (a, b, g, h) or 29 °C (d, e, f). Drug concentrations indicate final food concentrations. Each data point represents a replicate. N.S P( > 0.12), *P(0.033), **P(0.002), ***P(0.001), and ****P( < 0.0001). P-values ≤ 0.033 were considered significant. All statistical data were summarized in Supplementary Table S2.

Fig. 2. Pentose phosphate pathway enhanced trametinib resistance by promoting glucuronidation pathway.

The pentose phosphate pathway was increased in RAP tumours compared with Ras^G12V tumours and enhanced glucuronidation of trametinib. a An overview of the pentose phosphate pathway. b–e Percent survival of control or tumour flies to adulthood relative to control fly was quantified in the present or absence of trametinib (1 μM), D-ribulose-5p (50 μM). b Pgd-RNAi (+DMSO n = 16; +tram n = 16) or RAP +Pgd-RNAi (+DMSO n = 10; +tram n = 9), c RAP + Zw-RNAi (+DMSO n = 12; +tram n = 11) or RAP +Rpi-RNAi (+DMSO n = 19; +tram n = 19), d Ras^G12V (+DMSO n = 12; +D-ribulose-5p n = 12; +tram n = 12; +tram/D-ribulose-5p n = 12), e Control (+tram n = 12; +tram/D-ribulose-5p n = 12). f–h Released UDP analysis of Ras^G12V (+tram n = 14; +tram/D-ribulose-5p n = 14) (f), RAP + GFP (+tram n = 14) or RAP +Pgd-RNAi (+tram n = 14) (g), RAP + GFP ( + DMSO n = 21) or RAP +Pgd-RNAi (+DMSO n = 21) (h) in the present or absence of trametinib (1 μM), D-ribulose-5p (50 μM) in fly hindguts. Experiments were performed at 29 °C (b, c) or 27 °C (d–h). Error bar is a standard deviation (SD). N.S P( > 0.12), *P(0.033), **P(0.002), ***P(0.001), and ****P( < 0.0001). P-values ≤ 0.033 were considered significant. All statistical data were summarized in Supplementary Table S2.

Fig. 3. Pi3K/Akt signalling induces trametinib resistance by enhancing glucuronidation in Drosophila.

Upregulation of Wnt signalling enhanced glucose uptake in a Pi3K/Akt-dependent manner, thereby increasing glucuronidation of trametinib. a, c, f, h, k Percent survival of transgenic flies to adulthood relative to control flies was quantified in the present or absence of trametinib (1 μM), sucrose, error bar is a standard deviation (SD). a Ras^G12V (+DMSO n = 24; +sucrose n = 18; +trametinib n = 24; +tram/sucrose n = 12); c Ras^G12V + GFP (+tram/sucrose n = 20), Ras^G12V+Sgl-RNAi (+DMSO/sucrose n = 21; +tram/sucrose n = 17) and Ras^G12V +GlcAT-P-RNAi (+DMSO/sucrose n = 14; +tram/sucrose n = 15); f RAP + GFP ( + DMSO n = 5; +tram n = 8), RAP +Akt-RNAi (+DMSO n = 8; +tram n = 10); h RAP + AS160-RNAi (+DMSO n = 12; +tram n = 12); k Ras^G12V +Arm^CA (+DMSO n = 20; +tram n = 20) were induced in Drosophila hindguts. d, e Western blot analysis of Drosophila hindguts pAkt and Akt levels in Ras^G12V, RAP, Control, Arm^CA, or Ras^G12V +Arm^CA with or without sucrose. e’ The median normalised intensity of anti-pAKT in hindguts in each genotype compared with Control, error bar is a 95% confidence interval for the median, n = 3. g, i, j Released UDP analysis of RAP + GFP (+tram n = 5), RAP +Akt-RNAi (+tram n = 5) (g); RAP + GFP (+tram n = 6), RAP + AS160-RNAi (+tram n = 6) (i); Ras^G12V (+tram n = 4), Arm^CA (+tram n = 4) or Ras^G12V +Arm^CA (+tram n = 4) (j) with trametinib in Drosophila hindguts, error bar is a standard deviation (SD). Transgene expression was induced in Drosophila hindguts by a byn-GAL4 driver. Increased dietary sugar led to increased glucuronidation and reduced trametinib activity, while targeting glucuronidation enzymes or Pi3K pathway activity strongly potentiated trametinib activity. Experiments were performed at 27 °C (a–e, g and i–k) or 29 °C (f and h). N.S P( > 0.12), *P(0.033), **P(0.002), ***P(0.001), and ****P( < 0.0001). P-values ≤ 0.033 were considered significant. All statistical data were summarized in Supplementary Table S2.

Fig. 4. Deacetylation, glucuronidation lead to trametinib resistance in mouse AKP organoids.

Glucuronidation contributed to trametinib resistance in mouse AKP organoids. a Released UDP analysis of mouse AKP organoids in the present of trametinib. b, c, e, f Percent survival of AKP organoids relative to control was quantified in the present or absence of trametinib (5 nM), fasentin (30 μM), LY294002 (8 μM), vorinostat (0.5 μM) or phenacetin (100 μM). Error bar is a standard deviation (SD), n = 4. Data points display technical replicates. d Representative images showing the impact of drugs on AKP organoids, magnification 10X. Targeting glucuronidation led to increased effectiveness of trametinib. N.S P( > 0.12), *P(0.033), **P(0.002), ***P(0.001), and ****P( < 0.0001). P-values ≤ 0.033 were considered significant. All statistical data were summarized in Supplementary Table S2.

Fig. 5. HDAC1 is required for glucuronidation of trametinib in Drosophila.

Trametinib was deacetylated by HDAC1 in Drosophila. a Released UDP analysis of RAP + GFP (+tram n = 6) or RAP + HDAC1-RNAi (+tram n = 6) with trametinib in Drosophila hindguts. b–e Percent survival of adult tumour flies relative to control flies was quantified in the present or absence of trametinib (1 μM), vorinostat (0.5 μM) or phenacetin. b RAP + HDAC1-RNAi (+DMSO n = 12; +tram n = 10), c–e RAP ( + DMSO n = 18; +vorinostat n = 12; +tram n = 18; +tram/vorinostat n = 18) (c); RAP (+tram n = 12; +tram/5 μM phenacetin n = 12; +tram/10 μM phenacetin n = 12; +tram/15 μM phenacetin n = 12; +tram/20 μM phenacetin n = 12) (d); RAP (+DMSO n = 12; +5 μM phenacetin n = 12; +10 μM phenacetin n = 12; +15 μM phenacetin n = 12; +20 μM phenacetin n = 12) (e). Reduced HDAC activity led to reduced trametinib-dependent UDP release. f–j Images of the digestive tract of third instar larvae in the present or absence of trametinib (1 μM), vorinostat (0.5 μM) which include the hindgut proliferation zone (HPZ). Nuclei are visualized with 4′,6-diamidino-2-phenylindole (DAPI) staining, hindgut is marked by GFP. Scale bar 1 mm. k The average of hindgut proliferation zone (HPZ) size was measured by Fiji ImageJ and quantified as relative size to control hindgut. Control (n = 7), RAP (+DMSO n = 12; +vorinostat n = 7; +tram n = 12; +tram/vorinostat n = 11). Experiments were performed at 27 °C (a, c and d–k) or 29 °C (b). Error bar is a standard deviation (SD). N.S P( > 0.12), *P(0.033), **P(0.002), ***P(0.001), and ****P( < 0.0001). P-values ≤ 0.033 were considered significant. All statistical data were summarized in Supplementary Table S2. Transgene expression was induced in Drosophila hindguts by a byn-GAL4 driver. Reducing deacetylation/glucuronidation with vorinostat increased trametinib’s ability to rescue hindgut size.

Fig. 6. Reducing blood glucose levels suppressed drug resistance in RAP tumours.

Reducing blood glucose levels by feeding acarbose suppressed the glucuronidation of trametinib. a–c Percent survival of control or tumour flies to adulthood relative to control fly was quantified in the present or absence of binimetinib (8 μM), selumetinib (8 μM), or acarbose (15 μM). a RAP ( + DMSO n = 7; +binimetinib n = 20; +selumetinib n = 18) or RAP +Sgl-RNAi (+DMSO n = 8; +binimetinib n = 21; +selumetinib n = 20), b RAP (+selumetinib n = 12; +selumetinib/acarbose n = 12; +binimetinib n = 12; +binimetinib/acarbose). c RAP ( + DMSO n = 12; +acarbose n = 12). a–c Experiments were performed at 27 °C. Error bar is a standard deviation (SD). N.S P( > 0.12), *P(0.033), **P(0.002), ***P(0.001), and ****P( < 0.0001). P-values ≤ 0.033 were considered significant. All statistical data were summarized in Supplementary Table S2. d Schematic summary. Trametinib (tram) is a potent MEK inhibitor with the demonstrated preclinical ability to block RAS pathway signalling and oncogenic transformation. Pairing activated RAS and WNT activities leads to activation of PI3K/AKT signalling, AS160, and GLUT1/4 to increase glucose flux into cells. The result is elevated glucuronidation and elimination of trametinib. Potential therapeutic targets include HDAC1: deacetylation is an obligatory pre-step required for glucuronidation of some drugs including trametinib. This reaction can be reversed by P300.