An RNA damage response network mediates the lethality of 5-FU in colorectal cancer

Abstract

5-fluorouracil (5-FU), a major anti-cancer therapeutic, is believed to function primarily by inhibiting thymidylate synthase, depleting deoxythymidine triphosphate (dTTP), and causing DNA damage. Here, we show that clinical combinations of 5-FU with oxaliplatin or irinotecan show no synergy in human colorectal cancer (CRC) trials and sub-additive killing in CRC cell lines. Using selective 5-FU metabolites, phospho- and ubiquitin proteomics, and primary human CRC organoids, we demonstrate that 5-FU-mediated CRC cell killing primarily involves an RNA damage response during ribosome biogenesis, causing lysosomal degradation of damaged rRNAs and proteasomal degradation of ubiquitinated ribosomal proteins. Tumor types clinically responsive to 5-FU treatment show upregulated rRNA biogenesis while 5-FU clinically non-responsive tumor types do not, instead showing greater sensitivity to 5-FU's DNA damage effects. Finally, we show that treatments upregulating ribosome biogenesis, including KDM2A inhibition, promote RNA-dependent cell killing by 5-FU, demonstrating the potential for combinatorial targeting of this ribosomal RNA damage response for improved cancer therapy.

Keywords: 5-FU; 5-FU-based chemotherapy; RNA damage; ribosomal RNA; ribosomal protein.

Graphical abstract

Figure 1

Sub-additivity and antagonism of tumor cell killing in clinically relevant 5-FU drug combinations (A) Viability of CRC cell lines after 72 h treatment with 5-FU, oxaliplatin, or irinotecan. (B) Drug responses in 11 CRC cell lines quantified by AUC. (C) 5-FU effects on DDR potentially alter responses to oxaliplatin and irinotecan. (D) Dose-response matrix of 5-FU with irinotecan (top doses 200 and 50 μM, respectively) in HT-29 cells (top). Observed lethality at 72 h compared to Bliss independence model (bottom). (E) Dose-response analysis of HU with irinotecan (top doses 4 mM and 50 μM, respectively) as in (D). (F) HCT116 apoptosis after 48 h treatment with 1 mM HU ± 50 μM 5-FU using flow cytometry. (G) Dose-response matrix of 5-FU with oxaliplatin (top doses 200 and 20 μM, respectively) in CRC cell lines as in (D). (H) DLD1 apoptosis after 48 h treatment with oxaliplatin (5 μM) or irinotecan (6.25 μM) ± 50 μM 5-FU. (I and J) Clinical trial analysis of 5-FU/LV, oxaliplatin, and FOLFOX (I) or 5-FU/LV, irinotecan, and FOLFIRI (J). Data shown as mean ± SD (A, D, E, F, H).

Figure 2

5-FU metabolites biased toward RNA and DNA damage reveal pathway-specific responses (A) 5-FU metabolism by the pyrimidine salvage pathway. (B) DDR activation in HCT116 cells using equal concentrations of 5-FU, 5-FdUR, or 5-FUR. (C) HCT116 apoptosis after 24 h treatment with 5-FUR or 5-FdUR. (D) HCT116 p53^−/− apoptosis after 48 h treatment with 5-FUR or 5-FdUR. (E) Apoptotic protein levels in SW48 cells following 16 h treatment with increasing doses of 5-FUR or 5-FdUR (10 μM top dose). (F) HCT116 apoptosis after CX-5461 treatment prior to 2 μM 5-FUR or 20 μM 5-FdUR for 24 h. ∗p < 0.05. (G) HCT116 apoptosis after CX-5461 or ML-60218 treatment prior to 5-FUR addition for 4 h, assayed at 24 h. (H) HCT116 apoptosis after 24 h treatment with 1 ng/mL actinomycin-D 1 h prior to 5-FU (80 μM), 5-FUR (2 μM), or 5-FdUR (20 μM). ∗p < 0.05; ∗∗p < 0.01. (I) Viability of CRC cell lines (Figure 7A) after 72 h treatment with 5-FUR or 5-F-5′dUR. (J and K) HCT116 WT and p53^−/− cells treated with increasing doses of PF-477736 (Chki) and 5-FdUR (left) or 5-FUR (right), stained with SYTO60 and photographed. Quantified data show mean ± SEM from three independent experiments.

Figure 3

Distinct signaling responses to 5-FU metabolites that cause DNA and RNA damage (A) Schematic for 11-plex TMT mass spectrometry assessing signaling response to 5-FUR and 5-FdUR. (B) Indicated protein levels in HCT116 cells treated with 1 μM 5-FUR or 5-FdUR for indicated times. (C) Volcano plot for p-SQ/p-TQ phosphopeptide enrichment. (D) KAP1 S824-p peptide example for 5-FdUR-specific p-SQ/p-TQ phosphopeptide. (E) Volcano plot for phosphopeptide enrichment. (F) RPL12 S38-p peptide example for 5-FUR-specific phosphopeptide. (G) Volcano plot for KGG IP-enriched peptides. (H) Quantification of KGG IP-enriched peptides from ribosomal subunits in 5-FUR or 5-FdUR-treated cells. (I) As in (C), showing protein expression changes following treatment. (J) Ribosomal protein levels following treatment with 5-FUR relative to 5-FdUR. (K and L) STRING protein association network of 5-FUR (K) or 5-FdUR (L)-induced protein ubiquitination. Quantified data show mean ± SEM. ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Figure 4

5-FUR causes proteasome-dependent degradation of ribosomal proteins (A) HCT116 ribosome profiling after 18 h treatment with 2 μM 5-FUR or 5-FdUR. (B) Indicated protein levels after 18 h treatment with 5-FUR. (C) Quantified ribosomal protein levels after 18 h treatment with 2 μM 5-FUR ± 20 μM chloroquine and/or 5 μM MG132. Mean ± SEM from three independent experiments. ∗p < 0.05. (D) Nucleolin immunostaining after 6 h treatment with DMSO, 80 μM 5-FU, 2 μM 5-FUR, or 20 μM 5-FUR. Arrows indicate necklace-like nucleolin distribution; arrowheads indicate punctate nucleolin foci. Scale bar, 10 μm. (E) Nucleolin immunostaining after 6 or 24 h of 5-FUR treatment.

Figure 5

5-FUR causes autophagy-dependent degradation of ribosomal rRNA (A and B) rRNA abundance in HCT116 cells after 18 h treatment with 40 μM 5-FU, 2 μM 5-FUR, or 2 μM 5-FdUR, by agarose electrophoresis (A) or qPCR (B). (C) LC3 immunostaining following 18 h treatment with 2 μM 5-FUR or 5-FdUR. Scale bar, 10 μm. (D) Autophagic flux in HCT116 cells treated with 5-FUR or 5-FdUR. (E) rRNA abundance in si-control or si-ATG5-transfected HCT116 cells after 18 h treatment with 2 μM 5-FUR. Data normalized to DMSO-treated cells. (F) Autophagosome staining following 12 h treatment with DMSO or 2 μM 5-FUR. Scale bar, 10 μm. (G) Viability following 18 h treatment with 2 μM 5-FUR ± 20 μM chloroquine (CQ) using trypan blue exclusion, normalized to DMSO-treated cells ± CQ. (H) Apoptosis quantification following 24 h treatment as in (G). (I and J) Viability and apoptosis following si-ATG5 knockdown in cells after 18 h treatment with 5-FUR. All data shown as mean ± SEM from three independent experiments. ns, not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗p < 0.001.

Figure 6

Modulating rRNA synthesis enhances 5-FU RNA damage and cytotoxicity (A) Viability of HCT116 cells in DMEM ± glucose after 48 h treatment with 5-FUR. (B) RNA synthesis assessed after 24 h in DMEM ± glucose by EU incorporation. Dots indicate averaged nuclear EU intensity for all cells in a field. Scale bar, 10 μm. (C) Viability following 24 h treatment with 0.4 μM 5-FUR ± KDM2A inhibitor daminozide (Dam), assessed by trypan blue exclusion. Scale bar, 10 μm. (D) RNA synthesis as in (B) following 24 h treatment with DMSO or 2 μM daminozide. (E) PARP-1 cleavage following 24 h treatment with DMSO or 5-FUR ± daminozide. (F) Viability of si-control or si-KDM2A-transfected cells after 24 h treatment with 0.4 μM 5-FUR, as in (C). (G) PARP-1 cleavage in cells treated in (F). (H) Colony formation assay of cells transfected as in (F) after 11 days treatment with 5-FU, 5-FUR, and 5-FdUR, normalized to DMSO controls. (I) Apoptosis of cells as in (F) following 24 h treatment with DMSO, 40 μM 5-FU, 0.4 μM 5-FUR, or 20 μM 5-FdUR. (J) Apoptosis of cells pretreated with actinomycin-D (1 ng/mL) 1 h prior to 5-FUR (0.4 μM) for 24 h. All data shown as mean ± SEM from two (B, J) or three (C, D, F, I) independent experiments. ∗p < 0.05; ∗∗p < 0.01; ∗∗p < 0.001.

Figure 7

Variance in response to 5-FU RNA and DNA metabolites predicts patient response to 5-FU-based therapy (A) Viability of CRC cell line panel treated with 5-FUR and 5-FdUR shown as log₂(GI₅₀ 5-FUR/GI₅₀ 5-FdUR). Top: SW48 dose titrations. (B) Log₂(GI₅₀ 5-FUR/GI₅₀ 5-FdUR) for NCI60 cell lines with representative dose-response curves from publically available data. Each bar indicates a single cell line. (C) NCI60 Z scores of log₂(GI₅₀ 5-FUR/GI₅₀ 5-FdUR) correlated with their gene expression Z scores reveal genes highly correlated with sensitivity to 5-FU RNA or DNA metabolites. (D) RNA Pol I promoter opening gene set correlates with sensitivity to 5-FU RNA damage. (E) A gene signature correlating with patient sensitivity to FOLFOX indicates non-responders are resistant to 5-FU-mediated RNA damage. (F–H) Viability of patient-derived colon cancer organoids after 5 days treatment with 5-FUR or 5-FdUR. Mean ± SEM from six (F) or 3 (G and H) independent experiments are shown. (I) Nucleolin immunostaining and quantification in PDM2 organoids following 6 h drug treatment. Scale bar, 10 μm. ∗∗∗p < 0.001.