CRISPR/Cas9 Screens Implicate RARA and SPNS1 in Doxorubicin Cardiotoxicity

Figures

• 

  Central Illustration. Interrogating the Genetic Basis of Doxorubicin-Induced Cardiotoxicity via a Forward-Genetics Screening Framework We report the use of genome-wide loss-of-function CRISPR screens to nominate and interrogate genetic determinants of doxorubicin (DOX)-induced cardiotoxicity. This framework has the potential to shed light on mechanisms, define novel therapeutic entry points, and discover genetic risk factors. DAPI = 4′,6-diamidino-2-phenylindole; DMSO = dimethyl sulfoxide; iCM = induced pluripotent stem cell–derived cardiomyocyte; RARA = retinoic acid receptor-⍺; sgRNA = single-guide RNA.

• 

  Figure 1. Defining Modifiers of DOX Cytotoxicity Via Genome-Wide CRISPR/Cas9 KO Screens (A and B) Experimental overview. HL-1 (A) and AC16 (B) cells were transduced with genome-wide CRISPR/Cas9 knockout (KO) libraries at approximately 300× coverage, expanded, and treated with doxorubicin (DOX) (1 μmol/L for 3 days) or dimethyl sulfoxide (DMSO), after which single-guide RNA (sgRNA) representation was quantified by next-generation sequencing. (C and D) Ranked dot plots showing the relative abundance of sgRNAs aggregated per gene in HL-1 (C) and AC16 (D) cells treated with DOX versus DMSO. Each point represents a gene. Red points indicate sgRNAs differentially represented in DOX- vs DMSO-treated cells (false discovery rate [FDR] < 0.25). Select hits are labeled. (E and F) Top gene ontology (GO) terms and corresponding P values (DAVID; see Statistical methods) associated with perturbations leading to DOX desensitization in HL-1 (E) and AC16 (F) cells.

• 

  Figure 2. Validation of RARA Status as a Determinant of DOX Cardiotoxicity (A) Experimental overview. The top hits from the HL-1 and AC16 screens were used to design a targeted CRISPR/Cas9 KO library, which was transduced into SV207 induced pluripotent stem cells (iPSCs) at >1,000× coverage. Following differentiation, induced pluripotent stem cell–derived cardiomyocytes (iCMs) were treated with DOX or DMSO, after which sgRNA representation was quantified by next-generation sequencing. (B) Ranked dot plot showing the relative abundance of sgRNAs in iCMs treated with DOX vs DMSO. Each point represents an sgRNA. Red points are sgRNAs targeting RARA. (C) CRISPR-based disruption of RARA in the hB53 iPSC line. Two sites in exon 3 were targeted simultaneously. Multiple indel-bearing clones were recovered, including the homozygous 44 bp deletion mutant shown. (D) Wild-type and RARA KO iPSCs were differentiated into iCMs and treated with the indicated DOX concentrations for 72 hours. Cell viability was measured by alamarBlue assay (see Methods section). (E) Effect of tamibarotene (TBT) pretreatment on DOX toxicity in wild-type iCMs. Wild-type iCMs were pretreated with 1 μmol/L TBT or DMSO for 24…

• 

  Figure 3. TBT Mitigates DOX-Mediated Suppression of Mitochondrial Gene Expression (A) Principal component analysis plot based on transcript profiles in iCMs treated with DMSO, DOX, TBT, or DOX plus TBT. (B) Pairwise volcano plots of RNA-Seq data. Each point represents a transcript. Red points indicate transcripts differentially expressed by quasi-likelihood F tests (FDR <0.05). (C) GO term analysis (clusterProfiler) showing top cellular component terms associated with activated and suppressed transcripts in iCMs treated with DOX compared with DMSO. (D) As in C, but for iCMs treated with DOX plus TBT compared with DOX. In C and D, GeneRatio is the number of genes in the input list annotated to a given GO term divided by the total number of genes in the input list; color indicates the adjusted P value. Abbreviations as in Figures 1 and 2.

• 

  Figure 4. Discovery of Factors Regulating DOX Accumulation via Genome-Wide CRISPR/Cas9 KO Screens in AC16 Human Cardiomyocytes (A) Flow cytometry histograms showing a dose-dependent increase in DOX fluorescence in AC16 cells incubated with DOX (0, 100, 500, and 1000 nmol/L). (B) ABCC1 KO AC16 cells exhibit higher DOX fluorescence than WT AC16 cells after DOX incubation. (C) Experimental overview. AC16 cells were transduced with the Brunello genome-wide CRISPR/Cas9 KO library at approximately 300× coverage, expanded, and incubated with DOX (1 μmol/L for 24 hours); cells were then separated by FACS into a low DOX-content bin (bottom 15.0%) and a high DOX-content bin (top 15.0%). sgRNA representation was quantified by next-generation sequencing. (D and E) Ranked dot plots showing the relative abundance of sgRNAs aggregated per gene in the low DOX-content bin (D) and high DOX-content bin (E) vs unsorted cells. Each point represents a gene. Red points indicate differentially represented sgRNAs (FDR <0.25); select hit genes are labeled. (F and G) Top GO terms and corresponding P values (DAVID; see Methods section) associated with perturbations leading to low (F) and high (G) DOX accumulation in…

• 

  Figure 5. Validation of SPNS1 as a Modifier of DOX Accumulation (A and B) Flow cytometry analysis of DOX accumulation in SPNS lysolipid transporter 1 (SPNS1) WT and SPNS1 knockout (KO) AC16 cells. (A) Representative flow cytometry histogram. (B) Background-corrected median fluorescence after incubation with the indicated concentrations. (C) Time-lapse confocal microscopy of DOX accumulation in SPNS1 WT and SPNS1 KO AC16 cells; a 12-hour timepoint is shown. Dashed lines demarcate nuclear membranes for clarity; N indicates nucleus; arrows indicate perinuclear DOX aggregates. Scale bar: 20 μm. (D) Live-cell fluorescence imaging of SPNS1 WT and SPNS1 KO AC16 cells after DOX treatment (1 μmol/L for 24 hours) followed by LysoTracker Green (75 nmol/L; 1 hour). Scale bar: 20 μm. (E) Fluorescence imaging of microtubule-associated protein 1A/1B-light chain 3 (LC3) distribution in SPNS1 WT and SPNS1 KO AC16 cells after DOX treatment (1 μmol/L for 24 hours). Scale bar: 20 μm. Abbreviations as in Figures 1 and 2.

• 

  Figure 6. SPNS1 Deficiency Exacerbates DOX Accumulation and Toxicity in iPSC-Derived Cardiomyocytes (A) SPNS1 gene structure, WT genomic DNA (gDNA) sequence, region targeted by sgRNA, and Sanger traces from SPNS1 WT and SPNS1 KO iPSCs. (B) Flow cytometry analysis of DOX accumulation in SPNS1 WT and SPNS1-KO iCMs, shown as a representative histogram (left) and background-corrected median fluorescence after incubation with the indicated DOX concentrations (right). (C) Immunofluorescence microscopy of DOX accumulation in SPNS1 WT and SPNS1-KO iCMs. Scale bar: 25 μm. (D) Image cytometry of DOX localization in SPNS1 WT and SPNS1 KO iCMs: representative image captures (top), DOX fluorescence vs maximum pixel intensity (middle; maximum pixel intensity is proportional to aggregation, visible as puncta), relative quantification of DOX puncta (bar charts), and LysoTracker Green fluorescence vs DOX fluorescence (bottom). Scale bar: 25 μm. (E) LC3 and phosphorylated H2AX in SPNS1 WT and SPNS1 KO iCMs after treatment with the indicated DOX concentrations for 24 hours. (F) Flow cytometry of DOX accumulation in SPNS1 WT and SPNS1 KO iCMs cotreated with rapamycin (50 nmol/L) or Torin 1 (50 nmol/L). Bars…

• 

  Figure 7. Loss of Ribosome Genes and Nutrient Deprivation Reduce DOX Accumulation and Toxicity (A) Top gene ontology (GO) terms with corresponding normalized enrichment score (NES) and P values for cell survival data (DOX vs DMSO; left) and DOX accumulation data (low DOX content vs unsorted; right). (B) Flow cytometry histograms of DOX accumulation in iCMs exposed to DOX in full medium, Hanks balanced salt solution (HBSS), or HBSS supplemented with B27, MEM amino acids (AAs), and/or fetal bovine serum (FBS), as indicated. Data shown are from 1 experiment and are representative of 2 independent experiments. (C) Quantification of DOX accumulation by iCMs in specified media conditions. Bars show background-corrected median fluorescence intensity (mean ± SD) from 2 independent experiments. (D) Effect of nutrient deprivation on DOX toxicity in iCMs. iCMs were exposed to the indicated DOX concentrations in full medium or HBSS for 3 days, after which viability was measured by alamarBlue (see Methods section). Bars show mean normalized alamarBlue signal ± SD; ∗∗P < 0.01 (2-sided unpaired Student's t-test). Asterisks denote comparison between full media and HBSS at specified dose. Data are…