STAR Methods

CRISPR screens identify gene targets and drug repositioning opportunities at breast cancer risk loci

Resource availability

Further information and requests for resources and reagents should be directed to and will be fulfilled by Joseph Rosenbluh (sefi.rosenbluh{at}monash.edu).

Materials availability

All unique reagents generated in this study will be made available upon request. An agreement with our Institute’s Materials Transfer Agreement (MTA) may be required.

Data and code availability

RNA-Seq and ATAC-Seq data generated in this study will be made available at GEO. All CRISPR functional screening raw and analyzed data is available in the Supplementary Tables of this paper.

Cell lines

Human mammary epithelial cells (HMLE) used in this study was a gift from Prof. William Hahn (Dana Farber Cancer Institute), the B80 cell lines (B80-T17 and B80-T5) are in-vitro immortalized mammary cell lines previously described (Toouli et al., 2002). K5+/K19- and K5+/K19+ cell lines are immortalized progenitor mammary stem cells (Zhao et al., 2010). HMLE were induced to undergo epithelial to mesenchymal transition (EMT) to obtain a mesenchymal phenotype (mesHMLE) by culturing cells in DMEM:F12 media (1:1) supplemented with 10µg/ml insulin, 20ng/ml EGF, 0.5µg/ml hydrocortisone, 5µg/ml gentamycin, 5% FBS treated with 2.5ng/ml TGFβ1 for a minimum of 14 days (Pillman et al., 2018). HMLE and B80-T17 were propagated in mammary epithelial growth medium (MEGM) (Sigma). B80-T5 were cultured in RPMI 1640 (Sigma) supplemented with 10% FBS, 1% Penicillin and Streptomycin and 1% Glutamine. K5+/K19- and K5+/K19-cells were maintained in DFCI medium containing: MEM⍺/Ham’s F12 nutrient mixture (1:1, vol/vol) supplemented with 0.1M HEPES, 1µg/ml insulin, 1µg/ml hydrocortisone, 12.5ng/ml epidermal growth factor, 10µg/ml transferrin, 14.1µg/ml phosphoethanolamine, 0.545ng/ul β-Estradiol, 2mM glutamine, 2.6ng/ml sodium selenite, 1ng/ml cholera toxin, 6.5ng/ml triiodothyronine, 0.1 mM ethanolamine, 35µg/ml bovine pituitary extract, 10µg/ml gentamycin and 10µg/ml freshly prepared ascorbic acid. All cell lines were maintained in a humidified incubator at 37⁰C with 5%CO2.

Generation of stable cell lines

Lentiviral vector expressing a gene or sgRNA of interest, along with pMD2.G (Addgene#12259) and psPAX2 (Addgene#12260) were transfected into HEK293FT packaging cells. Lentiviral supernatant was harvested after 48-hour incubation in DMEM containing 30% FBS and passed through a 0.45µm Milli-hex filter. For oncogenic potential: K5+/K19- and K5+/K19+ cells were transduced with pLENTI-Hygro-PGK-TP53-DD and selected using 100µg/ml hygromycin. For colony formation assays and in vivo assays: HMLE, mesHMLE, B80-T5, B80-T17 and K5+/K19+ were transduced with pLX311-GFP-MEKDD and selected for GFP using the BD Influx™ cell sorter. K5+/K19-was transduced with pRRLsin-SV40 T antigen-IRES-mCherry (Addgene #58993) and positive cells were sorted using the BD Influx™ cell sorter. For CRISPR screens and validations all cell lines were transduced with following lentiviral vectors: Lenti-Cas9-2A-Blast (Addgene #73310), Lenti-dCas9-KRAB-Blast (Addgene #89567) and Lenti-dCas9-VP64-Blast (Addgene #61425). Cells were selected and maintained in blasticidin (5µg/ml to 10µg/ml). For single gene perturbation, 3 sgRNAs were cloned into BsmBI-digested lenti-Guide-Puro vector (Addgene# 52963) for CRISPRko and pXPR502 vector (Addgene #96923) for CRISPRa. Cells were infected with sgRNAs, selected and maintained in puromycin (1µg/ml to 2µg/ml).

RNA-seq

Transcriptome profiling was carried out using strand-specific TruSeq kit. Following RNA extraction (RNeasy, Qiagen) mRNA was enriched using polyT beads (Genewiz) and sequencing libraries were prepared using Illumina strand-specific TruSeq kit (Genewiz). Samples were sequenced on an Illumina HiSeq machine (PE 150bp). RNA-seq were aligned to Ensembl v70 gene models with STAR v2.7.1a. Duplicate reads were marked with PicardTools v2.19, then reads mapping to transcriptome using featureCounts in subread v1.6.0, count matrix generated using RSEM v1.3.1. Differential expression analysis was performed using DESeq2 in R v3.6.2.

ATAC-seq

Profiling of regions of open chromatin using previously reported protocols (Buenrostro et al., 2015). Duplicate libraries were prepared for each cell type and paired-end sequenced (150bp) generating a minimum of 40 filtered reads per library. Adapters trimmed using Cutadapt v1.13 and reads aligned to GRCh37 using Bowtie v2.2.9. Duplicates marked with Picard MarkDuplicates v2.19. Peaks were called using MACS2 and cell type-specific replicating peaks identified using BedTools.

HiChIP

HiChIP libraries were generated with the Arima HiChIP kit using an antibody against H3K27ac (Active Motif AbFlex: 91193). Cells were counted using the Countess II automated cell counter (Thermo Scientific) and fixed with 2% formaldehyde using the Arima HiC+ Kit (Arima, A101020). 1e6 fixed cells were used in restriction enzyme digest, biotin end filling ligation reactions to the manufacturer’s protocol. Libraries were prepared using the KAPA Kit (KAPA, KK2620), according to the Arima-HiC kit protocol. Libraries were indexed using the Swift Biosciences indexing kit then paired-end sequenced (150bp) with Illumina Novaseq 6000 to generate >500M raw reads per library. Individual replicate reads were processed with HiC-Pro (v 2.11.4) and aligned to hg19. Replicate samples for each cell type were quality controlled and checked for genome-wide signal correlation before merging with HiC-Pro. Enriched regions representing H3K27ac peaks were detected using MACS2. Chromatin loops were detected in each cell type-specific dataset using FitHiChIP v8.1 at 2 kb resolution limiting to 2 Mb interaction distance. Peak-to-peak and peak-to-nonpeak loops were used for background modelling and a q < 0.01 threshold set to determine significant interactions.

Generation of pooled sgRNA library

sgRNA sequences in custom libraries are available in Supplementary Table 2. sgRNAs were designed using CRISPick algorithm (Doench et al., 2016). For each gene we chose top scoring 5 sgRNAs (based on CRISPick scores). Libraries were prepared as previously described (Davies et al., 2021; Rosenbluh et al., 2016; Rosenbluh et al., 2017). Briefly, oligonucleotide pools (CustomArray) contained the sgRNA sequence appended to BsmBI cutting sites and overhang sequences for PCR amplification. The final sequence obtained is: AGGCACTTGCTCGTACGACGCGTCTCACACCG20nt spacerGTTTCGAGACGTTAAGGTGCCGGGCCCACAT. Following PCR amplification with Fwd: 5’-AGGCACTTGCTCGTACGACG-3’, Rev: 5’- ATGTGGGCCCGGCACCTTAA-3’ primers the PCR product was cloned via Golden Gate assembly into BsmBI-digested lentiGuide-Puro vector (Addgene# 52963) for CRISPRko and CRISPRi libraries and into pXRP502 (Addgene #96923) for CRISPRqtl and CROPSeq oligos were cloned into CROPseq-Guide-Puro (Addgene#86708). Ligated libraries were electroporated into NEB5α electrocompetent cells (NEB), plasmid DNA was extracted using Qiagen Maxi Prep. For each library preparation, a 1000X representation was ensured.

2D and 3D proliferation screens

Mammary cell lines (HMLE, mesHMLE, B80-T5, B80-T17, K5+/K19- and K5+/K19+ cells) stably expressing Cas9 (Addgene# 73310), KRAB-dCas9 (Addgene# 89567) or dCas9-VP64 (Addgene# 61425) were established. Cells were then transduced with either the CRISPRko, CRISPRi or CRISPRa Library at a MOI of 0.3 to obtain 1,000 cells/sgRNA. Twenty-four hours post infection cells were selected using Puromycin (2µg/ml) for 7 days. Cells were then subdivided to assay for 3D proliferation by plating cells in low attachment conditions (Corning#4615) or on 2D plates. To ensure sgRNA and Cas9/dCas9 expression, cells were maintained with puromycin and blasticidin throughout the screen. Twenty-one days post-infection cells were washed in PBS and genomic DNA was extracted using NucleoSpin Blood XL kit (Clontech). For colonies grown in low attachment conditions, genomic DNA was extracted using the DNeasy Kit (Qiagen).

Olaparib synthetic lethal screens

Cell lines stably expressing Cas9 (Addgene# 73310), KRAB-dCas9 (Addgene# 89567) or dCas9-VP64 (Addgene# 61425), were transduced with the CRISPRko, CRISPRi or CRISPRa sgRNA libraries at a low MOI (0.3) at a coverage of 1,000 cells/sgRNA. Puromycin containing medium was added 24 hours post-infection and cells were allowed to undergo selection for 7 days. For all screens, following selection, cells were trypsinized and divided into two treatment groups: DMSO or Olaparib. HMLE, mesHMLE, B80-T17, K5+/K19- and K5+/K19+ cells were treated with 5uM of Olaparib and B80-T5 cells were treated with 2.5uM of Olaparib for 14 days. Media was replaced every 4 days with DMSO or Olaparib. Cells were harvested by centrifugation and genomic DNA was extracted using NucleoSpin Blood XL kit (Clontech).

In vivo screen

HMLE-MEKDD, K5+/K19+-MEKDD and B80-T5-MEKDD cells expressing Cas9 or dCas9-VP64 were infected at MOI=0.3 with CRISPRko or CRISPRa libraries. Following puromycin selection (2µg/ml) for 7 days, 2e6 cells/site were subcutaneously injected into NSG mice at 3 sites/mouse. Tumor growth was measured using a digital caliper every 48 hours and monitored continuously until tumor volume reached 1cm3 (sum of all three sites). Tumor volume was calculated using the formula length (mm) × width (mm) × height (mm). Mice were sacrificed once tumors reached 1cm3. Cells were dissociated using Bead Ruptor machine and glass beads and DNA was extracted using DNeasy Kit (Qiagen).

Library preparation, sequencing and analysis

High-throughput sequencing library was generated using one-step PCR to amplify the integrated sequence within the construct and the addition of a barcode as previously described (Davies et al., 2021; Rosenbluh et al., 2016; Rosenbluh et al., 2017). PCR products were then purified using AMPure beads and samples sequenced using HiSeq (Illumina). PoolQ was used for deconvolution and alignment of sgRNA reads.

Crystal violet proliferation assay

Cells were plated at 2,000 cells/well and allowed to propagate until confluent. Media was aspirated and washed twice in PBS followed by fixation in 10% formalin for 10 minutes at room temperature. Formalin was removed and 0.5% (w/v) of crystal violet solution (Sigma) was added and incubated for 20 minutes at room temperature. Plates were washed in dH20 and imaged. For quantification 10% acetic acid was added to each well and incubated at room temperature for 30 minutes. The crystal violet solution was quantified by measuring the OD at 590nm using the PHERAstar (BMG).

3D proliferation assays

Cells were plated at 8000 cells/well in a 24-well low attachment plate (Corning). Colonies were allowed to form for 21 days. Images were taken at 4X magnification using an EVOS M5000 microscope (Thermo). Quantification of colonies was done by adding Cell-Titer-Glo Reagent (Promega) to wells, followed by a 10-minute incubation at room temperature on a shaker. Cell lysates were transferred to a 96-well white plate and luminescence measured using the PHERAstar (BMG).

Western blot

Cells/tissue were harvested, washed in PBS and resuspended in RIPA buffer (CST-9806) containing proteinase inhibitors (Roche) and quantified using the Pierce BCA Protein Assay Kit (Thermo Fisher). Protein lysates diluted in 4 X Laemmli Sample Buffer (Bio-Rad 161-0747) were loaded onto Bio-Rad 4-20% precast gels. Following electrophoresis, proteins were transferred to a pre-activated PVDF membrane using the Trans-Blot®Turbo™ Transfer System and visualized using ECL (Bio-Rad Chemidoc). Antibodies used in this study are listed in STARS methods key resource table.

Animals

The Monash University Animal Ethics Committee approved all animal use in this study (AEC – approval number 2020-24197-49078). For these experiments, 5-7 week old female NSG mice were purchased from Australian Research Laboratories (WA, Australia) or were kindly gifted from Professor Gail Risbridger and A/Prof. Renea Taylor (Monash University).

Validation of in vivo screens

B80-T5-MEKDD cells stably expressing Cas9 were infected with lentiviruses containing sgRNA’s targeting AAVS1 (control), ATF7IP, DUSP4, TGFBR2, CREBBP. Twenty-four hours post-infection, cells underwent puromycin selection for 7 days and expanded. Cells were trypsinized, washed twice in PBS and injected into NSG mice subcutaneously under isofluorane anesthesia. For each sgRNA, we injected 2e6 cells/site, 3 sites per mouse. Tumor growth was measured using a digital caliper every 48 hours and monitored continuously until tumor volume reached 1cm3 (sum of all three sites). Tumor volume was calculated using the formula length (mm) × width (mm) × height (mm). Mice were sacrificed once tumors reached 1cm3.

CRISPRqtl

CRISPRqtl was done as previously described (Gasperini et al., 2019). Briefly, K5+/K19+ cells stably expressing KRAB-dCas9 were infected with the CRISPRqtl library at MOI=5. 24h post infection, cells were selected with puromycin (2µg/ml) and cultured for 10 days. Cells were trypsinized washed with PBS and resuspended in PBS to reach a concentration of 1,200 cells/ml. Single-cell suspensions were loaded generated using the 10X Genomics Chromium Controller and Chromium Next GEM Single Cell 3′ GEM, Library and Gel Bead Kit v3.1 (10X Genomics cat #1000121), per manufacturer’s instructions (CG000204 Rev D) with the following modifications and variables. A single sample was loaded in two wells of the Next Gem Chip G, overloaded at 150% of the recommended cell input volume, with the corresponding volume of dH2O deducted at Step 1.2b (using the Cell Suspension Volume Calculator Table; p26). At Step 2.2d, cDNA was generated using 11 cycles of PCR. Samples were recombined 1:1 before Step 3.1. Prior to enzymatic shearing, 10% of the cDNA was used for sgRNA PCR enrichment. Specifically, A three-step nested PCR was used for gRNA enrichment (Hill et al., 2018).

PCR 1: 5 ng of 10x cDNA was amplified using NEBNext high fidelity 2x PCR mix (NEB # M0541) and the following primers: Rxn1_Fwd: TTTCCCATGATTCCTTCATATTTGC, Rxn1_Rev: ACACTCTTTCCCTACACGACG. Cycling conditions: 98°C for 30s, 14x (98°C for 10s, 50°C for 10s, 72°C for 20s), 72°C for 2min. PCR product was gel purified using the Qiagen MinElute Gel extraction kit (Qiagen # 28604). PCR 2: 5ng of PCR 1 was amplified using NEBNext high fidelity 2x PCR mix (NEB # M0541) and the following primers: Rnx2_Fwd: GTGACTGGAGTTCAGACGTGTGCTCTTCCGATCTTTGTGGAAAGGACGAAACA C, Rnx2_Rev: AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTC. Cycling conditions: 98°C for 30s, 7x (98°C for 10s, 64°C for 10s, 72°C for 15s), 72°C for 2min. PCR product was gel purified using the Qiagen MinElute Gel extraction kit (Qiagene # 28604). PCR 3: 5ng of PCR 1 was amplified using NEBNext high fidelity 2x PCR mix (NEB # M0541) and the following primers: Rnx3_Fwd: CAAGCAGAAGACGGCATACGAGATGACAGCATGTGACTGGAGTTCAGACGT, Rnx2_Rev (see PCR_2). Cycling conditions: 98°C for 30s, 11x (98°C for 10s, 64°C for 10s, 72°C for 15s), 72°C for 2min. PCR product was gel purified using the Qiagen MinElute Gel extraction kit (Qiagene # 28604) and then purified using AMPure beads (Beckman Coulter # A63881). cDNA and PCR product were pooled at a 1:10 ration and sequenced on two lanes of an MGISeq machine (Genewiz) using 150 PE-cycles (total of 569e6 reads).

CROPSeq

CROPSeq analysis was similar to CRISPRqtl with the following modifications. CROPSeq library was transduced into K5+/K19+ cells expressing WT Cas9 at a MOI of 0.1 ensuring 1 sgRNA/cell. Single cell isolation and library preparation was exactly as described for CRISPRqtl but only one 10x chromium lane was used (15,181 cell isolated) and sequencing was done on one MGISeq lane.

L1000 analysis

For each gene knockout we use the Z-score matrix (Supplementary Table 7) to define the up and down regulated genes. Top 150 up/down regulated genes were used as an input for the CMap for reference perturbagen signatures (https://clue.io/query) (Subramanian et al., 2017). We used version 1 of the CMap signature database for this analysis and collected both individual compound (PERT) and PCL for each of these signatures.

Article TitleCRISPR screens identify gene targets and drug repositioning opportunities at breast cancer risk loci

Abstract

Genome-wide association studies (GWAS) have identified >200 loci associated with breast cancer (BC) risk. The majority of candidate causal variants (CCVs) are in non-coding regions and are likely to modulate cancer risk by regulating gene expression. We recently developed a scoring system, INQUISIT, to predict candidate risk genes at BC-risk loci. Here, we used pooled CRISPR activation and suppression screens to validate INQUISIT predictions, and to define the cancer phenotypes they mediate. We measured proliferation in 2D, 3D, and in immune-deficient mice, as well as the effect on the DNA damage response. We performed 60 CRISPR screens and identified 21 high-confidence INQUISIT predictions that mediate a cancer phenotype. We validated the direct regulation of a subset of genes by BC-risk variants using HiCHIP and CRISPRqtl. Furthermore, we show the utility of expression profiling for drug repurposing against these targets. We provide a platform for identifying gene targets of risk variants, and lay a blueprint of interventions for BC risk reduction and treatment.


Login or Signup to leave a comment
Find your community. Ask questions. Science is better when we troubleshoot together.
Find your community. Ask questions. Science is better when we troubleshoot together.

Have a question?

Contact support@scifind.net or check out our support page.