Organoid cultures derived from patients with advanced prostate cancer

Abstract

The lack of in vitro prostate cancer models that recapitulate the diversity of human prostate cancer has hampered progress in understanding disease pathogenesis and therapy response. Using a 3D organoid system, we report success in long-term culture of prostate cancer from biopsy specimens and circulating tumor cells. The first seven fully characterized organoid lines recapitulate the molecular diversity of prostate cancer subtypes, including TMPRSS2-ERG fusion, SPOP mutation, SPINK1 overexpression, and CHD1 loss. Whole-exome sequencing shows a low mutational burden, consistent with genomics studies, but with mutations in FOXA1 and PIK3R1, as well as in DNA repair and chromatin modifier pathways that have been reported in advanced disease. Loss of p53 and RB tumor suppressor pathway function are the most common feature shared across the organoid lines. The methodology described here should enable the generation of a large repertoire of patient-derived prostate cancer lines amenable to genetic and pharmacologic studies.

Figure 1. Copy number landscape of patient derived prostate cancer organoid lines

A. Top: Significant genomic aberrations in the prostate oncogenome from the MSKCC CGH dataset adopted from (Taylor et al., 2010). Bottom: Copy number landscape of seven patient derived organoid lines and six publically available prostate cancer cell lines using array CGH data. Shades of red and blue represent level of gain and loss. B. Copy number changes at chromosome 3p14 containing FOXP1, SHQ1, and RYBP, the CHD1 locus, the PTEN locus, the TMPRSS2-ERG locus and the AR locus. Arrows point to highly focal regions of biallelic deletion. See also Figure S1 for zoomed view together with corroborating RNA-Seq and WES reads.

Figure 2. Mutational landscape of patient-derived prostate cancer organoid lines

A. Quantification of number of non-synonymous single nucleotide variations (SNV, pink) and insertion-deletions (Indel, orange) in each organoid sample compared to average total SNV and Indels of published whole-exome datasets (green). The light blue line represents the average mutations of the seven organoid lines. B. The alteration rate (green: mutation, including both non-synonymous SNV, blue: homozygous deletion, red: amplification, grey: multiple alterations) of genes mutated in the 7 organoid samples in prostate cancer datasets. C. Table of genes that are mutated in more than one sample as well as genes with putative functional oncogenic significance. The genes are color-coded by RNA-Seq based mRNA expression quantified as reads per kilobase mapped (RPKM). See also Figure S2 for view of TP53 and SPOP mutations. See also Table S1 for quality metrics of WES and Table S2 for all a list of all somatic mutations associated gene expression. D. Scatterplot of allele frequencies of point mutations (single-nucleotide variant, SNV) determined by whole-exome DNA sequencing between organoid sample and adjacent tumor tissue of MSK-PCa2 and MSK-PCa7. Most points lie on the regression line, suggestion preservation of the mutational landscape. E. Scatterplot of allele frequencies of expressed SNVs determined by RNA-Seq sequencing between organoid sample and adjacent tumor tissue. Among the expressed subset of SNV's, the allele frequencies lie on the regression line suggesting preservation of the mutational landscape. F. Scatterplot of allele frequencies of SNVs determined by whole-exome DNA sequencing between organoid sample and FFPE lymph node metastasis from ∼1 year prior of MSK-PCa1 and MSK-PCa5. The red dots indicate mutations found only in the organoid line. Some mutations lie above the regression line of the common mutations, indicating that they have gained allele frequency from primary tumor to organoid line. The slope of the regression line of ∼0.6 suggests 60% tumor purity of the lymph node specimen. See also Figure S2, Tables S1, S2 and S3.

Figure 3. Histology of in situ prostate cancer, 3D organoids, and xengrafts

A. H&E of prostatectomy specimen of patient MSK-PCa1. B. H&E of organoids of MSK-PCa1. C. H&E of subcutaneous xenograft of MSK-PCa1. D. H&E of right acetabulum with metastatic prostate cancer from patient MSK-PCa2. E. H&E of organoids of MSK-PCa2. F. H&E of renal capsule xenograft of MSK-PCa2. G. H&E of retroperitoneal metastasis from patient MSK-PCa3. H. H&E of organoids of MSK-PCa3. I. H&E of subcutaneous xenograft of MSK-PCa3. J. Cytology of pleural effusion from patient MSK-PCa4. Arrows point to clusters of malignant cells. K. H&E of organoids of MSK-PCa4. L. H&E of subcutaneous xenograft of MSK-PCa4. Arrows point to mitotic figures. M. H&E of prostatectomy specimen of patient MSK-PCa5. N. H&E of circulating tumor cell-derived organoids of MSK-PCa5. O. H&E of subcutaneous xenograft of MSK-PCa5. P. H&E of extraprostatic prostate cancer from cystoprostatectomy specimen of patient MSK-PCa6. The field shown contains both adenocarcinoma and squamous differentiation highlighted in the dotted area. Q. H&E of organoids of MSK-PCa6. R. H&E of subcutaneous xenograft of MSK-PCa6. S. H&E of organoids of MSK-PCa7. T. H&E of organoids derived of MSK-PCa7. Scale bars represent 50 μM. See also Figure S3 for accompanying IHC.

Figure 4. Prostate cancer organoids exhibit diverse gene expression profiles

A. Heatmap of RNA expression of selected genes of seven organoid samples and three matched tumor samples (MSK-PCa2T, MSK-PCa6T, MSK-PCa7T) annotated in blue font. The genes are grouped by prostate cancer subclasses (ERG, SPINK1), prostate lineage transcription factors (FOXA1, HOXB13, AR), androgen receptor target genes (FKBP5, KLK2, KLK3, NKX3-1, TMPRSS2, STEAP1), a prostate lineage AR suppressed gene PSMA (FOLH1), epithelial markers (CK8, EPCAM, E-cadherin:CDH1), neuroendocrine markers (synaptophysin A:SYP, NCAM:CD56, chromogranin A:CHGA), genes whose expression is associated with epithelial to mesenchymal transition (FOXA2, N-cadherin:CDH2, vimentin: VIM, N-myc:MYCN, aurora kinase A: AURKA), squamous markers (TP63, KRT5, KRT6A). For each gene, the expression was log2 transformed and heatmap range is from minimum to maximum. B. Western Blot of selected proteins validating mRNA expression. Note pancytokeratin recognizes a pattern of low and high-molecular weight keratins. MSK-PCa6 displays a range similar to normal prostate organoids that contain both basal and luminal cells and MSK-PCa4 has lost expression of all cytokeratins. See also Figure S4 for hierarchical clustering of organoid mRNA expression and gene expression analysis of Michigan CRPC dataset. See also Figure S1 and S4.

Figure 5. Loss of tumor suppressors PTEN, TP53, RB1, and CDKN2A in prostate organoid lines

A. Oncoprint view of PTEN, TP53, RB1, and CDKN2A. The fill of the rectangle denotes copy number change (HOMOLOSS: dark blue, HETLOSS: light blue, GAIN: red). The border of the rectangle denotes gene expression change. Green square denotes mutation. B. Western Blot of PTEN, RB, phosphorylated RB (S807/S811), p16, and GAPDH of normal prostate organoids and prostate cancer organoid lines. C-E. Gene expression of PTEN, TP53, and RB1 in normal prostate tissue, primary cancer, and CRPC from the Michigan dataset. For CRPC, copy number and mutational data is available for most samples. Open circles indicates that mutational/copy number data is not available. Green fill indicates mutation. Light and dark blue fill indicates heterozygous and homozygous loss. F. Scatter plot of gene expression of RB1 and CDKN2A in benign (orange triangle), primary cancer (pink square) and CRPC (black circles). See also Figure S5 for RNA-Seq and Array-CGH tracks of CDKN2A.

Figure 6. Sensitivity to AR and PI3K inhibitor in organoid lines in vitro and in vivo

A-C. Dose-response curves of organoid lines to enzalutamide, everolimus, and BKM120. Growth was measured in quadruplet and viability was assayed 4 days after treatment at the indicated drug concentrations (Mean +/- SD). D-E. Growth of MSK-PCA1 and MSK-PCA2 xenografts. Treatment was started with tumors reach an average size of 400 mm³. Each tumor was normalized to the pre-treatment size. (N=10, Mean +/-SEM). P value derived from unpaired two-tailed T-test of normalized tumor size at the end of treatment. Ev: everolimus, Enz: enzalutamide, Cast: castration.