RNA Pol II inhibition activates cell death independently from the loss of transcription

Abstract

RNA Pol II-mediated transcription is essential for eukaryotic life. Although loss of transcription is thought to be universally lethal, the associated mechanisms promoting cell death are not yet known. Here, we show that death following the loss of RNA Pol II activity does not result from dysregulated gene expression. Instead, it occurs in response to loss of the hypophosphorylated form of Rbp1 (also called RNA Pol IIA). Loss of RNA Pol IIA exclusively activates apoptosis, and expression of a transcriptionally inactive version of Rpb1 rescues cell viability. Using functional genomics, we identify the mechanisms driving lethality following the loss of RNA Pol IIA, which we call the Pol II degradation-dependent apoptotic response (PDAR). Using the genetic dependencies of PDAR, we identify clinically used drugs that owe their lethality to a PDAR-dependent mechanism. Our findings unveil an apoptotic signaling response that contributes to the efficacy of a wide array of anti-cancer therapies.

Keywords: BCL2L12; DNA damage; PTBP1; RNA polymerase II; apoptosis; cancer therapy; cell death; chemotherapy; cisplatin; transcription.

Figure 1.. RNA Pol II inhibition activates apoptosis prior to passive decay of apoptotic mRNA and protein

(A) Immunoblot of Rpb1 in U2OS following 1 or 0.1 μM triptolide (TRP). Blots representative of 3 biological replicates. (B) Quantification of (A). (C and D) Live/dead cell kinetics following 1 μM TRP in Nuc::mKate2-expressing U2OS cells. mKate+ = live; Sytox+ = dead. (C) U2OS. (D) U2OS^{BAX−/−/BAK1−/−} (BAX/BAK DKO). Scale bar, 100 μm. (E) Fractional viability (FV) following drugs ± pathway-specific inhibitors: z-VAD-FMK (z-VAD) inhibits apoptotic caspases; VX765 inhibits pyroptotic initiator, caspase-1; TTM inhibits cuproptosis (cupro); Rucaparib (Rucap.) inhibits parthanatos (parth); Nec-1 inhibits necroptosis (necro); and Fer-1 inhibits ferroptosis (ferro). RSL3, elesclomol (Eles.), MNNG, and TSZ are canonical activators of listed death pathways. M-157, MDA-MB-157. Heatmap scaled to the mean of 3 biological replicates. (F and G) Long-term response to 1 μM TRP. (F) Representative images from 3 biological replicates. Scale bar, 275 μm. (G) Lethal fraction (LF). (H) mRNA levels following 1 μM TRP, relative to untreated cells. mRNA was quantified for times in which lethality had not yet occurred (4 h for WT; 10 days for DKO). Range for DKO represents 50 longest/shortest half-life mRNAs. Dashed lines denote extrapolation from fit. For WT, the median, 25^th, and 75^th percentiles are shown. (I) (Top) Immunoblot of Rpb1 in HAP1, or HAP1-RPB1-AID, following 500 μg/mL auxin (IAA, 3-indoleacetic acid). (Bottom) LF for HAP1-RPB1-AID cells +IAA, ±z-VAD. Data are mean ± SD of 9 biological replicates. For panels with error bars, data are mean ± SD of 3 biological replicates unless otherwise noted. See also Figures S1 and S2 and Tables S1 and S2.

Figure 2.. Loss of hypophosphorylated RNA Pol II correlates with onset of apoptotic death

(A) Immunoblots of Rpb1 following flavopiridol (10 μM), THZ1 (1 μM), TRP (0.1 μM), or actinomycin-D (Act-D, 0.1 μM). Blots representative of 3 biological replicates. (B) (Top) Quantification of decay kinetics for RNA Pol IIO (hyperphosphorylated) and RNA Pol IIA (hypophosphorylated) from immunoblots in (A). t_1/2 denotes time to 50% decay. (Bottom) Cell death kinetics. LF₅₀ denotes time to 50% max observed death. (C and D) Relationship between loss of RNA Pol IIO, or loss of transcriptional activity, and cell death. (C) Cell death kinetics for 17 unique drug-dose combinations, aligned to t_1/2 for loss of RNA Pol IIO (dashed vertical line). Mean of 3 biological replicates shown. (D) Correlation between LF₅₀ and t_1/2 for loss of RNA Pol IIO (left) or t_1/2 for loss of transcriptional activity measured using EU incorporation (right). Solid gray line denotes fit to a linear regression model. Black dashed line denotes direct x = y relationship, shifted by a constant to account for a lag time between drug activity and cell death. (E and F) Relationship between loss of RNA Pol IIA and cell death. (E) Same as (C), except aligned to t_1/2 for loss of RNA Pol IIA (dashed vertical line). (F) Correlation between t_1/2 for loss of RNA Pol IIA and LF₅₀. For panels with error bars, data are mean ± SD of 3 biological replicates. See also Figure S3 and Table S3.

Figure 3.. Loss of hypophosphorylated RNA Pol II initiates apoptosis

(A) Schematic describing approaches for uncoupling RNA Pol II activity from protein levels (Ai–Aiii). (B) Absolute mRNA fold change following 1 μM TRP in U2OS expressing sgRNA-NUP93 or nontargeting sgRNA. Two-sided KS test p value shown. (C) LF for cells in (B) following TRP. Wilcoxon rank sum p value shown (n.s. p > 0.05). (D) Immunoblot of Rpb1 in U2OS following 1 μM TRP, 0.316 μM THZ1, or a combination of the two. Blots representative of 3 biological replicates. (E and F) Quantification of Rpb1 after 4-h drug exposure, associated with (D). (E) Hypophosphorylated (RNA Pol IIA). (F) Hyperphosphorylated (RNA Pol IIO). (G) LF at 24 h, measured in identical conditions to (D)–(F). Data are mean ± SD for 6 biological replicates. (H–L) RNA Pol II switchover system (H) diagram of RNA Pol II construct. (I) Experimental logic of switchover system. (J) Immunoblot of Rpb1 following 10 μM ⍺-amanitin. (K) Absolute mRNA fold changes following ⍺-amanitin. (L) LF following ⍺-amanitin. Mean ± SD for 5 biological replicates. Unless otherwise noted, error bars represent mean ± SD of 3 biological replicates. See also Figure S3 and Table S4.

Figure 4.. Genome-wide screen identifies regulators of TRP-induced apoptosis

(A) Schematic of chemogenetic profiling to identify death regulatory genes. (B) Gene-level chemogenetic profiling data. Key death regulatory genes and nontargeting controls highlighted. (C) Simplified schematic of the cell-intrinsic apoptotic pathway. Bcl-2* denotes anti-apoptotic Bcl-2 family members. BH3* denotes pro-apoptotic BH3-only proteins. Regulators in blue/yellow were identified in chemogenetic profiling data. (D) Running enrichment scores (ESs) from gene set enrichment analysis on rank-ordered death rates, evaluating short-lived RNAs or long-lived RNAs. Nominal p value shown (n.s. p > 0.05). (E) Enrichment for stress response pathways among death regulatory genes identified by chemogenetic profiling. Significance was determined using a false discovery rate (FDR)-corrected one-tailed Fisher’s exact test. (F) Chemogenetic profiling data, highlighting genes associated with stress response pathways: Biocarta mitochondrial pathway (Mito. Pathway), unfolded protein response (UPR), integrated stress response (ISR), and DDR. Odds ratio for one-tailed Fisher’s exact test shown. Gray dashed lines: FDR = 0.1. (G) Immunoblot of phosphorylated-H2A.X(Ser139) following 31.6 μM etoposide or 1 μM TRP in U2OS. (H) LF following TRP in U2OS or U2OS-TP53-KO. Data are mean ± SD for 9 biological replicates. See also Figure S4 and Table S5.

Figure 5.. Apoptosis following RNA Pol II degradation depends on cytoplasmic translocation of PTBP1

(A) TRP-induced death in U2OS and PTBP1-KO. Data are mean ± SD for 6 biological replicates. (B) PTBP1-dependence for RNA Pol II degraders, or canonical apoptotic agents: 1 μM TRP, 10 μM ⍺-amanitin (⍺-ama.), 100 μM ABT-199 (ABT), and 0.5 μM staurosporine (STS). Drugs tested in U2OS, PTBP1-KO, or BAX/BAK DKO. LF kinetics with area under the curve colored: gray, WT; red, DKO or PTBP1-KO, as indicated. Areas based on mean of 6 biological replicates. (C) RNA-seq comparing drug-induced transcriptional changes in U2OS (WT) and PTBP1-KO, 8 h after TRP. Data normalized using ERCC spike-ins. Dashed line, x = y. Pearson correlation coefficient shown. Anti-apoptotic (anti-APO) and pro-apoptotic (pro-APO) regulators highlighted. (D) Chemogenetic profiling data, highlighting genes alternatively spliced in a TRP/PTBP1-KO dependent manner. Blue, increased in PTBP-KO; red, decreased in PTBP1-KO. See also Figure S5H. (E) CoIP of RNA Pol IIA with mCherry-PTBP1. (F) CoIP of PTBP1 with varied forms of RNA Pol II. (G) Localization of mCherry-PTBP1 ± TRP for 48 h. Scale bar, 25 μm. (H) Quantification of PTBP1 cytoplasm/nuclear ratio following TRP. (I and J) Fractionation to assess PTBP1 localization following exposure to ⍺-ama. W, whole-cell lysate; C, cytoplasmic; N, nuclear. (I) U2OS-RBP1-N792D/ΔCTD cells following 12-h 10 μM ⍺-amanitin. (J) As in (I), except with doxycycline (+Dox). See also Figures S5 and S6 and Table S6.

Figure 6.. BCL2L12 activates apoptosis following PTBP1 translocation

(A) Chemogenetic profiling data, highlighting BCL-2 family. (B) LF following TRP in U2OS and BCL2L12-KO. Data are mean ± SD of 6 biological replicates. (C) BCL2L12-dependence for RNA Pol II degraders or canonical apoptotic agents, as in Figure 5B. (D) Localization of HA-tagged BCL2L12 following TRP or vehicle for 24 h. Scale bar, 25 μm. (E) Quantification of BCL2L12 mitochondria:nuclear ratio, 24 h after TRP or vehicle. (F) Alignment of BH3 domains from established pro-apoptotic proteins and the BH3-like domain of BCL2L12. Conserved hydrophobic residues (h1–h4) and an invariant charged residue (blue) highlighted. (G) LF after the indicated compounds, in U2OS or BCL2L12-KO, expressing either BCL2L12 (WT B-L12) or mutant BCL2L12 (E220A B-L12). (H) iBH3 profiling of the BCL2L12 BH3-like domain. BIM is shown for comparison. (Left) Cytochrome c (CytoC) release in U2OS-BCL2L12-KO with 100 μM BIM-BH3, 100 μM BCL2L12-BH3-like peptide, or vehicle. (Right) iBH3 profiling across doses of BIM or BCL2L12 peptides. Data are mean ± SD of 3 biological replicates. (I) Schematic of PDAR. See also Figure S6.

Figure 7.. Commonly used anti-cancer drugs owe their lethality to RNA Pol II degradation

(A) Strategy to identify PDAR-dependent drugs. KO refers to genetic dependencies observed for RNA Pol II degraders. (B) Cell death in U2OS for a high-scoring drug (10 μM ⍺-ama) and low-scoring drug (3.16 μM STS). Data are mean ± SD of 3 biological replicates. (C) Heatmap depicting PDS score across a 7-point, half-log dose range for 46 compounds. Data collected in U2OS. Gray boxes, non-lethal doses. ***Clinically approved; *clinically investigated. (D) Binary classifier, ordered as in (C). Red circles denote PD-like mechanisms. (E) mCherry-PTBP1 cytoplasmic-nuclear (Cyto/Nuc.) ratio following indicated compounds. (F) Immunoblot of Rpb1 in U2OS following 1 μM abemaciclib (PD-like) or a 10-fold higher dose (low PDS). (G) As in (F), but for cisplatin. (H) PDS across a panel of cell lines for 1 μM TRP, 10 μM flavopiridol, 100 μM cisplatin, and 100 μM ABT-199. (I) Relationship between LF₅₀ and RNA Pol IIA t_1/2. Data in gray denote established transcriptional inhibitors, as shown in Figure 2F. Red and blue data represent PD-like and non-PD-like compounds, respectively. See also Figure S7 and Table S7.