Kinome-wide CRISPR-Cas9 screen.
22Rv1 cells were transduced with lentiCas9-Blast (addgene, #52962) and then selected with 10 μg ml-1 blasticidin for 3 weeks to generate a stable cell line expressing Cas9 (22Rv1-Cas9). Single clones were picked and further determined Cas9 expression and ability to edit genes knockout.
Human Kinome CRISPR pooled library (Brunello)41 was a gift from Drs. John Doench and David Root (addgene #75314). The library was amplified as previously described41. Briefly, the Kinome library was diluted to 50 ng ul-1 in water and then electroporated using Endura electrocompetent cells (Lucigen, #60242). The transformations were plated onto pre-warmed agar plates and incubated for 14 hrs at 32 °C. Colonies were harvested and the library plasmids were purified with a maxi scale plasmid prep (Qiagene, #12162).To make lentivirus, the library plasmids were co-transfected with packaging plasmids pCMV-VSV-G (addgene, #8454) and psPAX2 (addgene, #12260) into HEK293T cells with Lipofectamine 2000 (Thermo Fisher). After 6 hrs, the medium was changed and further cultured for an additional 60 hrs, then viral supernatants were harvested and centrifuged at 2,000 rpm at 4 °C for 10 min to pellet cell debris. The supernatant was then filtered through a 0.45 µm membrane and viruses were concentrated with a high speed centrifuge (24, 000 rpm, 2 hrs). The titer of the lentiviruses was determined with 22Rv1 cells in a functional assay by measuring puromycin resistance after transduction.
Infections were set up at 1500-fold coverage of the library. 22Rv1-Cas9 cells were infected at a low multiplicity of infection (MOI = 0.3) with the above prepared kinome library viruses in the presence of 8 μg ml−1 polybrene, followed by selection with puromycin for 3 days, then 1/3 of the infected cells (500 × coverage) were harvested as the baseline for deep sequencing and the remaining cells were continuously cultured for an additional two weeks waiting for genes edition. Then the cells were randomly split and treated with 20 µM ENZA or DMSO for an additional one week. The sgRNAs incorporated into the cells were amplified from genomic DNA and sequenced on Illumina HiSeq 2500.
Analysis of CRISPR-Cas9 screen data
Screen data analysis was performed by MAGeCK42 software. Raw sequencing reads were processed by FastQC for quality control, 21-27 bp were trimmed from the 5’ end until the protospacer sequence was reached, then mapped to the expected sgRNA sequence with zero mismatches tolerated. The read counts of sgRNAs were normalized by non-targeting control sgRNAs. Using the MAGeCK MLE algorithm, the normalized sgRNA read counts in ENZA and DMSO samples were compared to the initial baseline sgRNA distribution and negative control sgRNAs to estimate a beta score for each targeted gene in the treatment and control groups separately. The beta scores of all genes were normalized by the median beta score of the essential genes to make the cell proliferation rates comparable between treatment and control samples and to reduce false-positive hits 17. The top negatively selected genes were those with the smallest negative values in the differential beta score, which was calculated by subtracting the control beta score from the treatment beta score.
RNA sequencing and analysis
Total RNA derived from 22Rv1, 22Rv1 with CK1α knockout and 22Rv1 cells with CK1α overexpression was purified using an RNeasy Mini kit (Qiagene, #74104). The RNA sequencing was run as a paired-end 150bp in length with 30 million reads, on the Illumina HiSeq 2500 (Novogene). Raw sequences were mapped via HISAT243 (v2.1.0) to the human reference genome GRCh38. Read counting was performed using featureCounts44 from the subread package (v1.5.1). Differential gene expression was analyzed by R package limma45 (v3.42.2). All p-values were corrected for multiplicity by the Benjamini-Hochberg method. Gene Set Enrichment Analysis (GSEA) statistical analysis was carried out with publicly available software from the Broad Institute (http://www.broadinstitute.org/gsea/index.jsp). Gene expression results were ranked and used to conduct GSEA. GSEA was performed to evaluate enrichment of differential expression patterns curated from Hallmarks, Reactome, BioCarta, and KEGG database within the MSigDBv7.4.
RNA sequencing of C4-2 and C4-2R cells has been described46. Briefly, extracted total RNA from C4-2 and C4-2R cells was sequenced as a paired-end 100bp in length with 30 million reads, on the Illumina HiSeq 2500 (Purdue Genomics Facility). Tophat2 was used to align reads to the Ensembl Homo sapiens genome database version GRCh38.p5. The htseq-count script in HTSeq v.0.6.1 was run to count the number of reads mapping to each gene. DESeq2, edgeR and Cufflinks2 were used for differential expression analysis. Gene expression results were ranked and used to conduct GSEA. GSEA was performed accordingly.
Cell culture
Human PCa cell lines LNCaP, C4-2, MR49F, C4-2R and 22Rv1 were cultured in RPMI-1640 medium (Sigma-Aldrich). Among these cell lines, LNCaP is androgen-dependent, whereas C4-2 is an androgen-independent cell line derived from LNCaP. MR49F is an ENZA-resistant derivative of LNCaP obtained from Dr. Amina Zoubeidi (University of British Columbia, Vancouver, BC, Canada), whereas C4-2R is derived from C4-2 and kindly provided by Dr. Allen Gao (University of California, Davis, CA, USA). The concentration of ENZA for ENZA-resistance maintenance of MR49F and C4-2R was 10 or 20 μM, respectively. ENZA was removed for at least 48 hrs before any experiment. 22Rv1 is an established intrinsic ENZA resistant cell line. LNCaP, C4-2, and 22Rv1 cells were purchased from American Type Culture Collection. HEK293T cell was a gift from Dr. Andrea Kasinski (Purdue University, USA) and cultured in Dulbecco’s modified Eagle’s medium (Sigma-Aldrich). All medium contained 10% fetal bovine serum (Atlanta Biologicals, GA, USA), 100 U ml-1 of penicillin and 100 μg ml−1 streptomycin. Cells were cultured in a humidified atmosphere at 37 °C, with 5% CO2.
Mouse models and treatment
All animal experiments were approved by the Institutional Animal Care and Use Committee at University of Kentucky (KY, USA). Patient derived xenograft (PDX) experiments were conducted as previous described21. Male NSG (NOD scid gamma) mice (age 6-8 weeks) were castrated and then implanted subcutaneously with LuCaP 35CR tumor bits two weeks later. When the tumor volume reached 50 mm3, mice were pre-treated with ENZA (50 mg kg-1 in PBS containing 1% carboxymethyl cellulose, 0.1% Tween 80, 5% DMSO) by oral gavage daily, for a routine of 5 days on, 2 days off. After one week, ENZA resistance would have developed21. Mice were randomly assigned into four groups: Control (with vehicle); ENZA (20 mg kg-1 in PBS containing 1% carboxymethyl cellulose, 0.1% Tween 80, 5% DMSO, by oral gavage every 3 days); A51 (5 mg kg-1 in ddH2O containing 10% DMSO, 10% Solutol and 10% 2-hydroxy prolyl-b-cyclodextrin, by oral gavage, two days on and one day off); or combination with ENZA and A51 (Figure 2G). Tumor growth was monitored twice weekly by measuring the length and width of tumors. Tumor volume was evaluated using the formula: (length × width2)/2. The mice were sacrificed when tumor volume exceeded 1,500 mm3, when the diameter exceeded 20 mm, when the animals became compromised, or after 12 weeks, whichever developed first for the end point of the study. After humane euthanization, the tumors were harvested for paraffin embedding or frozen for subsequent analyses.
22Rv1 in vivo xenograft experiments were conducted by subcutaneous injection of 2 × 106 22Rv1 cells (wild-type or CK1α-knockout) (100 μl in 50% PBS and 50% Matrigel, Corning) into the flanks of pre-castrated male nude mice (Jackson Lab, MI, USA). When the tumor volume reached 100 mm3, daily gavage treatment with 20 mg kg-1 ENZA or vehicle (PBS containing 1% carboxymethyl cellulose, 0.1% Tween 80, 5% DMSO) was started and continued for 5 weeks (5 days on, 2 days off ) (Figure 2D). Tumor growth was monitored twice weekly by measuring the length and width of tumors. Tumor volume was evaluated using the formula: (length × width2)/2. After humane euthanization, tumors were harvested for paraffin embedding or frozen for subsequent analyses.
Reagents
Enzalutamide (S1250), doxycycline (S5159), cycloheximide (S7418), MG132 (S2619) and D4476 (S7642) were purchased from Selleck Chemicals (TX, USA). A51 was synthesized by DC Chemicals (Shanghai, China). CK1α kinase (PV3850) was purchased from Thermo Fisher (MA, USA). ATM protein (14-933) and polybrene (TR-1003-G) were purchased from Sigma-Aldrich (MO, USA). Transfection Reagents Lipofectamine 2000 (11668019) and jetPRIME® (89129-924) were obtained from ThermoFisher (MA, USA) and VWR (PA, USA), respectively. For in vitro study, chemicals were dissolved in DMSO (D2650, Sigma-Aldrich). Enzalutamide used in animal studies was dissolved in PBS containing 5% DMSO (D2650, Sigma-Aldrich), 1% carboxymethyl cellulose (S6703, Selleck Chemicals) and 0.1% Tween 80 (P4780, Sigma-Aldrich) as previously described21. A51 was dissolved in ddH2O containing 10% DMSO (D2650, Sigma-Aldrich), 10% Solutol (HY-Y1893, MedChemExpress) and 10% 2-hydroxy prolyl-b-cyclodextrin (HY-101103, MedChemExpress) as previously described 26. Protein A/G Magnetic Beads (#88802) were purchased from ThermoFisher (MA, USA).
Plasmids
The following plasmids were obtained from Addgene: lentiCas9-Blast (#52962), lentiGuide-Puro (#52963), LentiGuidPuro-hTP53 (#88853), sgTP53_3 (#78164), pcDNA-FLAG-ATM (#43907), PRKDC gRNA1 (#77861), PRKDC gRNA2 (#77862), ATM gRNA1 (#77530), ATM gRNA2 (#77531), CHEK2 gRNA1 (#76487), CHEK2 gRNA2 (#76488), Tet-plko-puro (#21915), pLC-Flag-CSNK1A1-WT-Puro (#123319), CSNK1A1 gRNA1 (#76188), CSNK1A1 gRNA2 (#76189), pcDNA3.1-HA (#128034), psPAX2 (#12260) and pCMV-VSV-G (#8454). Myc-DDK-CK1α (#RC217936) was purchased from OriGene (MD, USA). GST-ATM fragments were gifts from Dr. Zhenkun Lou (Mayo Clinic, Rochester, USA) and site mutations were generated by a Q5 Site-Directed Mutagenesis Kit (New England Biolabs, E0554). PcDNA3.1-HA-CK1α, Tet-plko-puro-CK1α, sgAR-full length were generated in this study.
Antibodies
Antibodies for immunobloting: anti-Cas9 mAb (#14697), anti-AR mAb (#5153), anti-β-Actin mAb (#4970), anti-CK1α mAb (#2655), anti-ATM mAb (#2873), anti-CHK2 Ab (#2662), anti-phospho-CHK2 (T68) Ab (#2661), anti-γH2AX mAb (#9718), anti-H2AX mAb (#7631), anti-DNA-PKcs mAb (#38168), anti-Cleaved PARP (Asp214), anti-Myc-Tag (#2278), anti-HA-tag (#3724) were purchased from Cell Signaling Technology (MA, USA). Anti-vinculin mAb (v4505) and anti-Flag-tag (F1804) were bought from Sigma-Aldrich (MA, USA). Anti-CK1α mAb (sc74582) and anti-p53 mAb (sc126) were from Santa Cruz Biotechnology (CA, USA). Antibodies for immunoprecipitation: anti-HA Magnetic Beads (#88836) was obtained from ThermoFisher (MA, USA), anti-Flag Magnetic Beads (M8823) was purchased from Sigma-Aldrich (MA, USA). Anti-CK1α mAb (sc74582) was from Santa Cruz and anti-ATM mAb (ab201022) was purchased from Abcam (MA, USA). Antibodies for immunofluorescence (IF) and immunohistochemistry (IHC) assay: anti-γH2AX (#9718), anti-Cleaved Caspase-3 (#9661) and anti-phospho-CHK2 mAb (T68) (#2197) were obtained from Cell Signaling Technology. Anti-CK1α (TA313698) was obtained from Origene (MD, USA). Anti-ATM (ab32420) was from Abcam (Cambridge, UK). Goat anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 488(A-11034) was purchased from Thermo Fisher (MA, USA). Anti-phospho-ATM (S1270) pAb was made by Sino Biological (PA, USA).
Tumor microarray (TMA) construction and IHC staining
For TMA, formalin-fixed paraffin-embedded tissue blocks from prostate cancer patients who underwent radical prostatectomy at the University of Kentucky (KY, USA) were cored to construct a TMA through the Markey Cancer Center Biospecimen Procurement and Translational Pathology Shared Resource Facility. Approval for use of human prostate tissue was obtained from the University of Kentucky Institutional Review Board. Cancerous, adjacent benign epithelial, or benign prostatic hyperplasia tissue cores (2 mm) were used, with duplicate cores from each patient. Four-micron thick sections of the TMA were cut for IHC staining. For regular xenograft tumor, four-micron thick sections of formalin-fixed paraffin-embedded tumor tissue were cut for IHC staining. Staining was carried out on Ventana Discover Ultra. Antigen retrieval was performed using CC1 (Roche) (for ATM, γH2AX and Cleaved Caspase-3) or CC2 (Roche) (for CK1α) with standard conditions. Slides were incubated with primary antibodies (ATM at 1:50, γH2AX at 1:100, Cleaved Caspase-3 at 1:150 and CK1α at 1:50) for 1 hr at 37 °C, incubated with anti-Rabbit-HQ (Roche), followed by anti-HQ-HRP (Roche) and visualized with DAB prior to light counterstain with Mayer’s hematoxylin. The TMA staining was evaluated and scored by a pathologist. The regular staining of tumor xenografts was also evaluated by a pathologist by comparison to the positive control. Images were taken with a Nikon microscopic camera or scanned with the Aperio Digital Pathology Slide Scanner.
Immunofluorescence staining
Cells were fixed with 4% paraformaldehyde for 15 min at room temperature (RT), permeabilized in 0.2% Triton X-100 for 10 min and blocked with 10% FBS/PBS (v/v) for 30 min. Afterwards, the cells were incubated with appropriate primary antibodies overnight at 4°C and secondary antibodies for 1 hr at RT. All antibodies were diluted in PBS containing 1% BSA and 1% normal goat serum. The dilution ratios of the primary antibodies were 1:500 for ATM and γH2AX, 1:100 for p-CHK2. The secondary antibody was used at 1: 500. The nucleus was stained with Hoechst (Sigma-Aldrich, B2261). Images were captured with a confocal microscope (Nikon).
Immunoblot (IB) and immunoprecipitation (IP) analyses
Cell lysates were prepared as previously described47. Briefly, cultured cells were harvested and lysed with the lysis buffer (50 mM Tris, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.2% SDS, protease inhibitors, and phosphatase inhibitors). For tumor tissue, the lysates were prepared by homogenization using the above lysis buffer followed by sonication with a probe sonicator. After quantification with BCA assay, equal amounts of protein were loaded on SDS-PAGE gel and then transferred to nitrocellulose membranes, followed by IB with the indicated antibodies. For IP analysis, 1 mg protein collected with the lysis buffer (50 mM Tris, 150 mM NaCl, 0.5% sodium deoxycholate,1% NP-40, protease inhibitors) were incubated with the indicated primary antibodies overnight at 4°C followed by the protein A/G magnetic beads for 3 hrs at RT or antibody-conjugated beads overnight at 4 °C. The immunocomplexes were washed three times with the lysis buffer and analyzed by IB.
AquaBluer assay to determine cell viability
AquaBluer assay was performed according to the manufacturer’s instructions (MultiTarget, # 6015). Cells were seeded at 3000-8000 (dependent on the cell size) per well in 96-well culture plates for cells grow overnight. Cells were then treated as indicated and further cultured for 72 hrs. Dilute AquaBluer™ (1: 100) in the culture medium and mix well. The medium was then removed from the cell culture and 100 μl of the diluted AquaBluer™ was added to each well. The plate was returned to the incubator for another 4 hrs. Once placed in a fluorescence plate reader, the fluorescence intensity was read at 540ex/590em.
Colony formation assay
Cells were seeded at 500-2000 (dependent on cell size) per well in 6-well or 12-well culture plates. After growing for 24 hrs, cells were treated as indicated and further cultured for 3 weeks. The medium was then removed and the cells were rinsed carefully with PBS. Cells were fixed with 4% paraformaldehyde for 15 min at RT before they were stained with 0.5% crystal violet for 30 min at RT. After staining, the crystal violet solution was removed and rinsed carefully with tap water. Colonies were left to dry on the plates at RT. The images were taken with the ChemiDoc Imaging System (Bio-Rad).
In vitro kinase assay
In vitro kinase assays were performed with a kinase buffer (CST, #9804) (25 mM Tris-HCl (pH 7.5), 5 mM beta-glycerophosphate, 2 mM dithiothreitol (DTT), 0.1 mM Na3VO4 and 10 mM MgCl2) supplemented with 125 μM ATP and 10 μCi of [γ- 32P] ATP at 30 °C for 30 min in the presence of CK1a kinase and GST-ATM proteins. After the reaction, mixtures were resolved by SDS-PAGE, the gels were stained with Coomassie brilliant blue, dried, and subjected to autoradiography. In vitro kinase assays for mass spectrometry analysis were performed with the kinase buffer above supplemented with 200 μM ATP (CST, #9804) at 30 °C for 30 minutes in the presence of CK1a kinase and GST-ATM proteins. After the reaction, proteins were denatured, resolved on SDS–PAGE, the gels were stained with Coomassie brilliant blue, the protein bands were cut and subjected to mass spectrometry analysis.
sgRNA or shRNA clones
To clone the sgRNA or shRNA to the vectors, lentiGuide-Puro (Addgene #52963) or Tet-plko-puro (Addgene #21915) was cut by restriction enzymes and dephosporylated with FastAP (Thermo Fisher). Oligonucleotides for the sgRNAs or shRNAs sequence (sgAR-full#1 5’GTTACACGTGGACGACCAGA3’ sgAR-full#2 5’GTGTCCAGCACACACTACACC3’, shCK1α 5’GCCACAGTTGTGATGGTTGTT3’) were phosphorylated using polynucleotide kinase and then annealed by heating to 95 °C for 5 minutes and cooling to 25 °C at 1.5 °C minute-1. Annealed oligos were then ligated into lentiGuide-Puro or Tet-plko-puro above using T4 ligase (New England Biolabs) at 16 °C overnight.
Statistics Methods
For comparisons between two groups of independent datasets, multiple t-tests were performed, and p-value and standard error of the mean (SEM) were reported. For comparisons among more than two groups, one-way or two-way ANOVA were performed, p-values and SEM were reported. For drug response curves, p-values were calculated by non-linear regression with paired t-tests. For all figures, ∗ represents p<0.05, ∗∗ represents p<0.01, ∗∗∗ represents p<0.001. All bioinformatic analysis and comparisons are described below.
Bioinformatic analysis
Correlation analysis with Spearman’s Correlation coefficient
The Spearman’s Correlation coefficient and its associated 95% confidence interval (CI) were used to quantify the correlation between ATM and CK1α expression in TMA IHC staining. A linear regression line helps visualize the negative correlation between the two genes.
Analysis of “SU2C PNAS2019” dataset
Scatter plot together with Spearman's rank correlation coefficient was applied to evaluate the correlation strength between the duration of ARSI treatment and FPKM-normalized ATM mRNA expression based on the “SU2C PNAS2019” datasets23. A linear regression line helped visualize the correlation. The analysis concerns only those patients whose ARSI treatment status is recorded as “Off treatment”=TRUE. SU2C PNAS-2019 mRNA FPKM expression values were downloaded from the cBioportal website (https://www.cbioportal.org/). The histogram of the FPKM-normalized gene levels of ATM of the SU2C PNAS-2019 samples was plotted to demonstrate the distribution of ATM levels. A dashed line indicates the median expression of ATM. Kaplan-Meier curve, together with the logrank p-value, was generated to evaluate the association between the possibility of patients staying on the line of ARSI treatment and the ATM expression levels (dichotomized as high versus low with median chosen as the cutoff). Forest plot associated with the statistics based on Cox proportional hazards model was generated to evaluate the association between the ARSI-treatment duration length and the ATM mRNA levels (dichotomized as high versus low with median chosen as the cutoff). An Hazard Ratio being less than 1 indicates that patients in the ATM-high group tended to remain on the line of ARSI treatment longer than those in the ATM-low group.
GSEA analysis with TCGA dataset (https://www.cancer.gov/tcga)
TCGA-PRAD mRNA FPKM expression values were downloaded from the GDC data portal and manually converted to TPM-normalized gene levels. The histogram of the interested gene level from the TCGA-PRAD tumor samples was plotted to demonstrate the distribution of gene expression. The dashed lines indicate the median or quartile expression of the interested genes. Differential comparison analysis between the indicated gene expression high (>= cut off, indicated in the figure legend) and the low (< cut off) subgroups was performed based on the quasi-likelihood F-test in R package edgeR. Genes were then ranked based on p-values together with the corresponding signs of log2-foldchange. Pre-ranked GSEA was then performed accordingly.
Analysis of “Dr. Alumkal PNAS2020” dataset
Dr. Joshi Alumkal kindly provided the raw count table from their RNA-seq dataset7. Raw read counts were log transformed and analyzed by limma45 to generate log2 fold change for non-responders versus responders, which was used to pre-rank the genes as the input for GSEA. A CK1α gene signature composed of 188 genes was generated by overlapping up regulated genes (adjusted p<0.05, knockout/control ratio >1) in CK1α knockout samples and down regulated genes (adjusted p<0.05, overexpression/control ratio <1) in CK1α overexpression samples. Heatmap was generated by R package pheatmap.