Combined MEK and PARP inhibition enhances radiation response in rectal cancer

Abstract

Rectal cancer is frequently diagnosed at a locally advanced stage and treated by neoadjuvant chemoradiation. Current efforts to improve treatment outcome are focused on intensifying neoadjuvant chemotherapy, which is associated with higher levels of toxicity. To discover alternative strategies, we establish patient-derived rectal cancer organoids that reflect clinical radiosensitivity and use these organoids to screen 1,596 drug-radiation combinations. We find that inhibitors of rat sarcoma virus/mitogen-activated protein kinase (RAS-MAPK) signaling, especially mitogen-activated protein kinase kinase (MEK) inhibitors, strongly enhance radiation response. Mechanistically, MEK inhibitors suppress radiation-induced activation of RAS-MAPK signaling and selectively downregulate RAD51, a component of the homologous recombination DNA repair pathway. Through testing drug-drug-radiation combinations in organoids and cell lines, we identify that a combined poly ADP-ribose polymerase (PARP) and MEK inhibition can further enhance radiosensitivity of colorectal cancers, which we confirm in mouse xenograft models. Our data support clinical testing of MEK and PARP combination therapy with radiation in locally advanced rectal cancers as an alternative to chemoradiation.

Keywords: DNA repair; MEK inhibitor; PARP inhibitor; RAD51; RAS signaling; drug combination screen; organoids; radiation therapy; radiosensitizer; rectal cancer.

Graphical abstract

Figure 1

An organoid platform recapitulates essential aspects of rectal cancer (A) Schematic illustration of the rectal cancer organoid platform and approach for association of patient and organoid response. (B) Driver mutations identified in patient-derived organoids. (C) Response of rectal cancer organoids to increasing doses of radiation. (D) Analysis of organoid response to radiation according to donor patients’ rectal cancer response to radiation therapy assessed by MRI-based regression grading (left, 11 evaluable cases, two-tailed Student’s t test) and histopathological examination (right, 12 evaluable cases, two-tailed Welch’s t test). (C and D) Each data point represents the mean of 3–6 biological replicates tested per organoid line. (E) Representative bright-field images of rectal cancer organoid cultures undergoing radiation. Scale bars: 100 μM. (F) Representative endoscopy (left) and MRI (right) images from selected patients pre-treatment and post radiation therapy, sorted according to organoid response to radiation therapy. Green dotted lines (left) and red arrows (right) indicate location of rectal cancer. See also Table S1; Figure S1.

Figure 2

Drug screening identifies RAS-MAPK pathway inhibitors to synergistically enhance radiation in rectal cancer organoids (A) Schematic representation of the drug-radiation screening workflow. Rectal cancer organoids were seeded in 384-well plates; drug perturbations with 4–5 concentrations and radiation (2–4 Gy) were performed on day 3, before viability was measured on day 9 after seeding. Interactions of drugs and radiation were analyzed by calculating the difference between areas under the dose-response curves (ΔAUC values) between irradiated and non-irradiated conditions, each normalized to respective irradiated and non-irradiated DMSO controls on the same plates. (B) Composition of the kinase drug library with 224 drugs, tested in 4 concentrations in two organoid lines. (C) Ranking of differential effects of kinase inhibitors with or without radiation in tumor organoid D080T. (D) Top 10 hits with highest radiation enhancement (ΔAUC values) in the kinase library screen of organoid line D080T. (E) Ranking of differential effects of kinase inhibitors with or without radiation in cancer organoid D007T. (F) Top 10 hits with strongest radiation enhancement (ΔAUC values) in the kinase library screen of organoid line D007T. (C–F) Mean values of two biological replicates are shown. (G) Composition of the clinical cancer library consisting of 140 drugs. The drugs were administered in 5 concentrations, and 10 organoid lines were tested. (H) Mean area under the curve of all drugs tested in the clinical library in radiated vs. non-radiated conditions. (I) Ranking the mean differential effects of clinical cancer drugs with or without radiation in 10 rectal cancer organoids. (J) Top 10 hits with strongest radiation enhancement (mean ΔAUC values) in the clinical library screen with ten rectal cancer organoids. (K) Distribution of ΔAUCs of selected groups of inhibitors. (L) ΔAUCs of individual organoid lines, ΔAUCs of MEKi and PARPi are highlighted. (I–K) Mean values of 10 tested organoid lines are shown. (H and L) Each data point represents the mean ΔAUC value of two biological replicates tested for each organoid line. See also Figures S2–S7.

Figure 3

Radiation induces activation of RAS-MAPK signaling (A) Phosphorylation of ERK1/2 in CRC lines at different time points after irradiation. (B) Expression of RAS-MAPK pathway target genes is induced in DLD1 and SW480 cell lines 6 days after irradiation. (C) RNA expression profiling of rectal cancer organoid line D007T, 96 h after irradiation treatment with 4 Gy. Volcano plot of differentially expressed genes in irradiated vs. non-irradiated organoids. Target genes of the EGFR signaling pathway according to PROGENY are highlighted. RNA expression profiling experiments of D080T and D160T can be found in Figure S7C. (D) Gene set enrichment analysis of HALLMARK gene sets in irradiated vs. non-irradiated organoids D007. Analysis of D080T and D160T can be found in Figure S7D. (E) Phosphorylation of ERK1/2 in CRC lines after irradiation is reduced by MEKi trametinib (TRA) treatment. (F) Transcriptional induction of target genes of the RAS-MAPK pathway after irradiation is suppressed by MEK inhibition in CRC cell lines. (G) Phosphorylation of ERK1/2 in rectal cancer organoids 2 days after irradiation is reduced by concomitant MEKi treatment. (A, E, G) Representative images of three independent biological replicates are shown. (C and D) Data from five independent biological replicates are shown. (B and F) Data from three independent experiments are presented as mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 two-tailed Student’s t test. See also Figure S8.

Figure 4

MEKi modulates DNA damage response by downregulating DNA repair protein RAD51 (A) Radiation-induced DNA damage as determined by immunofluorescence staining of p-γH2AX. Green, p-γH2AX; blue, DAPI; 63× magnification; scale bars: 20 μm. (B) Measurement of p-γH2AX foci per nuclei under different treatment conditions. (C) Immunoblot showing induction of cellular p-γH2AX levels upon radiation. (D) Global proteome profiling by mass spectrometry of SW480 cells after treatment with 100 nM trametinib vs. DMSO for 24 h, abundance of selected DNA damage response pathway proteins is depicted below. (E) RAD51 protein expression at different time points after irradiation and MEKi trametinib treatment in CRC cell lines. (F) RAD51 protein expression 2 days after irradiation and MEKi trametinib treatment in patient-derived rectal cancer organoids. (G) RNA expression levels of RAD51 in CRC cell lines 24 h after irradiation ± trametinib treatment as determined by qPCR. (H) Colony-forming assay with CRC cell lines after siRNA-mediated knockdown of RAD51 ± radiation. Staining of cell culture plates was performed 11 days post radiation. Knockdown efficiency of RAD51 after 48 h is shown by western blot (left). (I) Colony-forming assay in CRC cell lines after treatment with different concentrations of the RAD51 inhibitor RI-1 for 11 days ± radiation. Scans of complete wells of standard 6-well plates are shown (9.6 cm² per well) (H and I). (J) Proliferation of patient-derived rectal cancer organoids after treatment with RI-1 for 2 days and ± radiation, scale bars: 50 μm. (A, E, F) representative images of three independent biological replicates are shown. (B and G) Data from three independent experiments are presented as mean ± SD. ∗p < 0.05, ∗∗p < 0.01, two-tailed t test, p values are only shown in case of significant differences. See also Figures S9 and S10.

Figure 5

MEK and PARP inhibition have synergistic viability effects in colorectal cancer models (A) A combination drug screen was performed with MEKi trametinib vs. 4 other top candidates interacting with radiation, derived from the radiosensitization screening experiments shown in Figure 2 (PI3K inhibitor, PARP inhibitor, EGFR inhibitor, and CHK1 inhibitor) in matrices of 7 × 7 concentrations (8 × 8 including DMSO) using 7 organoid lines. (B) Synergy scores according to Bliss synergy, highest single agent (HSA), Loewe, and zero interaction potency (ZIP) models for the 4 drugs in combination with trametinib are shown. The overall scores represent the highest score of all dose combinations tested, 2 biological replicates were analyzed for D030T and D157T, 3 replicates were analyzed for D007T, D052T, D134T, and D160, and 4 replicates were analyzed for D080T. (C) Three-dimensional response (% inhibition) surface of the talazoparib and trametinib combination, exemplified by D134T organoids. The surface contains fitted values. (D) Bliss synergy surface of the talazoparib and trametinib combination in D134 organoids. The surface contains fitted values. (E) Heatmap of response (% inhibition) of trametinib-talazoparib combinations. The mean values of all seven tested organoid lines are shown; values for the individual lines were calculated as means of 2–4 biological replicates, as indicated in (B). Results of individual organoid lines are found in Figure S12. (F) Heatmap Bliss synergy score of trametinib-talazoparib combinations. The average of bliss synergy scores for each dose combination of all seven tested organoid lines is shown. Results of individual organoid lines are found in Figure S12. (G) Growth inhibition of 4 representative cancer organoid lines treated with increasing concentrations of trametinib in the presence of talazoparib at 0.039 and 0.16 μM. Expected response according to Bliss synergy model and observed response are shown. The mean values of 3 biological replicates are shown for D007T, D134T, and D160, and of 4 replicates for D080T. (H and I) Viability after combinatorial inhibition of MEK and PARP in short-term and long-term viability assay in CRC cell lines. SW480 (H) and DLD1 (I) were treated for 4 days with trametinib and talazoparib in a concentration matrix, followed by cell viability measurement. Results were normalized to the DMSO control. Means of 3 biological replicates are presented (left). Long-term colony formation assays showing the combinatorial inhibition of MEK and PARP on colony growth compared to any single reagent treatment in CRC cell lines (right). Representative scans of complete wells of standard 6-well plates are shown (9.6 cm² per well). See also Figures S11 and S12.

Figure 6

Radiation synergizes with MEK-PARP combination therapy (A) Response/inhibition matrix derived from talazoparib-trametinib combinations, averaged over all seven tested organoid lines: non-irradiated, irradiated, and Bliss expected response, according to a model of added radiation to fixed combinations of trametinib and talazoparib, as well as Bliss excess (observed response − expected response). Data were normalized to non-irradiated DMSO controls. The lowest 5–6 concentrations tested are shown for each drug. (B) Dose-inhibition relationships of trametinib-radiation in combination with talazoparib treatments. Bliss expected response was calculated by using trametinib-radiation as one perturbation and adding talazoparib as second perturbation. D007T, D030T, D080T, D134T, and D160T are shown as representative examples of seven tested organoid lines. D007T, D030T, D080T, and D160T were irradiated with 4 Gy, D134T as a more radiation-sensitive line was irradiated with 2 Gy. Mean values of 2 biological replicates are shown for D030T, 3 replicates for D007T, D134T, and D160, and 4 replicates for D080T. (C) Bar plots of Bliss expected response vs. observed response in organoid lines treated with radiation, 0.625 μM talazoparib and 0.6–9.8 nM trametinib. Bliss expected response was calculated according to the same model as described in (B). Mean values of 2 biological replicates are shown for D030T and D157T, 3 replicates for D007T, D052T, D134T and D160, and 4 replicates for D080T. (D) Long-term colony formation assays of radiation in combination with MEK inhibition and PARP inhibition on colony growth compared to any single reagent treatment in CRC cell lines. Representative scans of complete wells of standard 6-well plates are shown (9.6 cm² per well). See also Figures S11 and S13.

Figure 7

Radiation combined with MEK and PARP inhibition leads to significant tumor growth inhibition in vivo (A) Schematic overview of the in vivo experiments. (B–D) Tumor volume of DLD1 xenografts according to treatment condition over the course of the experiment (B) and at day 27 (C). (D) Images of the tumors (n = 5 per group) after sacrifice of mice, sorted by treatment condition. (E–G) Tumor volume of SW480 xenografts according to treatment condition over the course of the experiment (E) and at day 27 (F). (G) Images of the tumors (n = 5 per group) after sacrifice of mice, sorted by treatment condition. (H) Weight of tumor-bearing mice during the experiments, according to experimental conditions. Significant weight differences were observed for the MEKi-PARPi-radiation group versus the untreated and radiation only groups. (B–H) Two-way ANOVA with Tukey’s multiple comparisons test was used to test statistical significance. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. (I) Putative mechanisms of interaction of MEK-PARP-radiation combination therapy. Radiation leads to DNA damage, which induces the cellular DNA damage repair machinery to enable cancer cell survival. Additionally, the RAS-MAPK pathway is upregulated. MEKi block radiation-induced RAS-MAPK signaling and downregulate RAD51, a core protein of the DNA DSB homologous recombination repair pathway. Addition of PARPi further enhances the effect by targeting DNA damage response via a different target.