Introduction: Prostate cancer research spans genetics, proteomics, metabolomics, cell signaling, and tumor microenvironment domains. While this data is valuable, integrating clinical and basic research findings presents challenges. The Prostate Cancer Cartography Initiative (PCCI) offers a unified data model for efficient research across clinical, co-clinical, and mechanistic studies. The data model divides cases into anatomic and biological groups, facilitating cross-domain knowledge translation and hypothesis generation through efficient data integration. Materials & Methods: Our database incorporates data from 972 patients across three complementary sources: TCGA (primary prostate cancer), SU2C (metastatic patients), and MDA trials (intermediate-risk prostate cancer). For reproducibility and data consistency, we use NF core pipelines to harmonize data and derive copy loss, Single Nucleotide Variants/Insertions/Deletions (SNV & Indel), and gene expression. After analyzing this information, we group patients based on clinical information such as Gleason scores or genomic information such as copy loss and SNV/Indel to evaluate how these conditions affect gene expression and pathways. Results: In our initial work, we focused on TCGA patients with Gleason scores, categorizing them into three groups (3+3, 3+4/4+3, and 8 or higher). We further subdivided each category based on genetic alterations: patients with no copy loss, patients with at least one copy loss, and patients with Single Nucleotide Variants or Insertions/Deletions (SNV & Indel). We analyzed key genes including TP53 across these subgroups. Our initial results showed that medium-risk patients (Gleason scores 3+4/4+3) had significantly lower TP53 gene expression when they had both copy loss and SNV & Indel. Patients with SNV & Indel alone also showed lower TP53 gene expression, though this finding was not statistically significant (p = 0.06). High-risk patients (Gleason scores 8 or higher) demonstrated similar patterns. When analyzing how copy loss or SNV & Indel affect genes near TP53, we found reduced expression in nearby genes—ALOX15B (logFC: -1.0153, p<0.001), TMEM220 (logFC: -0.719, p<0.001), and MYO1C (logFC: -0.348, p<0.001)—in patients with TP53 copy loss or SNV & Indel. This pattern was confirmed in both medium-risk and high-risk groups. In our pathway analysis, including AD-target and homologous recombination pathways, patients without TP53 copy loss or SNV & Indel showed higher Enrichment Scores. Discussion: The PCCI provides a platform to integrate these data and generate hypotheses. Our results show that genetic alterations—such as copy loss and SNV & Indel—affect not only the TP53 but also nearby genes ALOX15B, TMEM220, and MYO1C. Notably, ALOX15B and MYO1C are tumor suppressor genes, which raises the question of how TP53 alteration leads to network changes in tumor suppressor genes. These findings further demonstrate the necessity of PCCI in facilitating hypothesis generation. Citation Format: Kang-Lin Hsieh, Sankar Maity, Ganiraju Manya, Bradley Broom, Christopher Logothetis, Patrick Pilie. Post-doctoral fellow [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 1061.

In contrast to stage-specific transcription factors, the role of ubiquitous transcription factors in neuronal development remains a matter of scrutiny. Here, we demonstrated that a ubiquitous factor NF-Y is essential for neural progenitor maintenance during brain morphogenesis. Deletion of the NF-YA subunit in neural progenitors by using nestin-cre transgene in mice resulted in significant abnormalities in brain morphology, including a thinner cerebral cortex and loss of striatum during embryogenesis. Detailed analyses revealed a progressive decline in multiple neural progenitors in the cerebral cortex and ganglionic eminences, accompanied by induced apoptotic cell death and reduced cell proliferation. In neural progenitors, the NF-YA short isoform lacking exon 3 is dominant and co-expressed with cell cycle genes. ChIP-seq analysis from the cortex during early corticogenesis revealed preferential binding of NF-Y to the cell cycle genes, some of which were confirmed to be downregulated following NF-YA deletion. Notably, the NF-YA short isoform disappears and is replaced by its long isoform during neuronal differentiation. Forced expression of the NF-YA long isoform in neural progenitors resulted in a significant decline in neuronal count, possibly due to the suppression of cell proliferation. Collectively, we elucidated a critical role of the NF-YA short isoform in maintaining neural progenitors, possibly by regulating cell proliferation and apoptosis. Moreover, we identified an isoform switch in NF-YA within the neuronal lineage in vivo, which may explain the stage-specific role of NF-Y during neuronal development.

PDF file - 150K, A: Mitotic figures in the xenografts; left panel: graphical presentation of the number of mitotic figures in SCPC/LCNEC xenografts (mean 11 mitoses per high power field) and adenocarcinoma xenografts (mean 2 per high power field); right panel: pictures from H&E stained slides of a SCPC and an adenocarcinoma (AdCa) xenograft. Arrows point to the mitotic figures. B: Higher expression levels of ki67 are noted in the SCPC/LCNEC xenografts and clinical samples compared to adenocarcinomas C:Flow cytometry of SCPC xenograft 146-10 and adenocarcinoma xenograft 180-30. (SCPC, small-cell prostate carcinoma; LCNEC, large-cell neuroendocrine carcinoma).

Supplementary Figure 1. Supervised hierarchical clustering of differentially expressed genes between typical CRPC PDX ("ADENO") and CRPC PDX with small cell carcinoma morphology ("SCPC"). Supplementary Figure 2. Morphologic spectrum of tumors of patients with the aggressive variant phenotype in the primary (A-D) and metastatic tumor sites (E-H) Supplementary Figure 3. Immunohistochemical analysis of all 59 available NCT00514540 samples and 8 PDX lines. Supplementary Figure 4. Principal component analysis of a16-serum-and-IHC-tumor-marker set in "Baseline" samples from 21 patients. Supplementary Figure 5. Total number of copy number alterations per sample. Supplementary Figure 6. mRNA levels of genes of interest. Supplementary Figure 7. Principal component analysis of a 10-CNA-marker set in TCGA, unsupervised CRPC and AVPC samples.

Supplementary Table 1 from BMP4 Promotes Prostate Tumor Growth in Bone through Osteogenesis

supplementary materials and methods, supplementary tables and supplementary figure legends Supplemental Table 1: Patient Characteristics. Supplemental Table 2: Targeting sequences for c-met. Supplemental Table 3: Antibodies and Sources. Supplemental Table 4: Oligonucleotides for PCR Amplification of Mouse Alkaline Phosphatase and Osteonectin.

PDF file - 270K, mRNA levels of A, CDC20 and B, FZR1/CDH1 in SCPC/LCNEC (MDA PCa 144-4, 144-13, 146-10, and 155-2) and adenocarcinoma (MDA PCa 170-4 and 180-30) xenografts, normalized to GAPDH. SCPC, small-cell prostate carcinoma; LCNEC, large-cell neuroendocrine carcinoma.

Supplemental Figure 1 (fig. S1). Representative whole body NaF-18 PET scan images after 44 days of (A) vehicle-treated and (B) cabozantinib-treated animals. Supplemental Figure 2 (fig. S2). Effect of cabozantinib on PDX tumor growth, animal weight, and survival. Supplemental Figure 3 (fig. S3). Representative images of viable islets of tumor cells in (A) Patients; (B) MDA PCa-118b grown intrafemurally; and (C) MDA PCa- 118b grown subcutaneously. Supplemental Figure 4 (fig. S4). Time-dependent effect of cabozantinib on expression and phosphorylation of primary and MET downstream targets. Supplemental Figure 5 (fig. S5). Relative expression of c-met mRNA in NT and knockdown cells as determined by qRT-PCR. Supplemental Figure 6 (fig. S6). MET expression and activity in NT and knockdown tumors. Supplemental Figure 7 (fig. S7): Quantitation of CD31 staining on bone tumors. Supplemental Figure 8 (fig. S8). Effects of cabozantinib on bone remodeling markers in the phase-2 clinical trial and growth and differentiation of primary osteoblasts. Supplemental Figure S9 (fig S9). Effect of cabozantinib on bone turnover in the Phase 2 trial and PDX grown intrafemurally. Supplemental Figure 10 (fig. S10). Quantitation of phospho-histone H3 staining.

SUPPLEMENTARY TABLE 1. AVPC ('ANAPLASTIC') ELIGIBILITY CRITERIA FOR NCT00514540, A PHASE II STUDY OF CARBOPLATIN AND DOCETAXEL IN PATIENTS WITH ANAPLASTIC PROSTATE CARCINOMA. SUPPLEMENTARY TABLE 2. CLINICOPATHOLOGICAL FEATURES OF THE PDX DONOR PATIENTS. SUPPLEMENTARY TABLE 3. PDX GROWTH RATE AND PSA PRODUCTION. SUPPLEMENTARY TABLE 4. ANTIBODIES USED FOR IMMUNOHISTOCHEMISTRY AND WESTERN BLOT. SUPPLEMENTARY TABLE 5. qRT-PCR PRIMER SEQUENCES. SUPPLEMENTARY TABLE 6. DONOR PATIENT CHARACTERISTICS. SUPPLEMENTARY TABLE 7. SERUM MARKER LEVELS AT TIME OF REGISTRATION TO NCT00514540. SUPPLEMENTARY TABLE 8. PRINCIPAL COMPONENT VARIATE WEIGHTINGS TCGA/UNSELECTED CRPC/AVPC SAMPLES. SUPPLEMENTARY TABLE 9. LINEAR DISCRIMINANT ANALYSIS WEIGHTS. SUPPLEMENTARY TABLE 10. Tp53 MUTATIONS DETECTED IN TCGA, UNSELECTED CRPC (21) AND AVPC SAMPLES.

Supplementary Table S1. PCR primer sequences in this study. Supplementary Table S2. AR IHC and FISH result Supplementary Table S3. AR mutation detail result in this study.

IHC, qRT-PCR and Western blot methods.

PDF file - 1.1MB, A, Chromosomal regions commonly amplified (red) and deleted (blue) in the SCPC/LCNEC (left panel) and adenocarcinoma (AdCa, right panel) xenografts. B, AR, cyclin D1, and RB1 promoter methylation of the AR-positive LNCaP prostate cancer cell line, the AR-negative PC-3 and DU145 prostate cancer cell lines, and the SCPC/LCNEC MDA PCa 144-4, 144-13, 146-10, and 155-2 xenografts. SCPC, small-cell prostate carcinoma; LCNEC, large-cell neuroendocrine carcinoma; AR, androgen receptor; RB, retinoblastoma.

Supplementary Figure S1. Distribution of RET981 cell line. Supplementary Figure S2. Detection of AR splice variants in SDCs. Supplementary Figure S3. Cleaved PARP increased by AR siRNAs.

Supplementary Figures 1-4 from BMP4 Promotes Prostate Tumor Growth in Bone through Osteogenesis

PDF file - 180K, A, Growth rate of MDA PCA 144-13, 146-10, 155-2, 170-4 and 180-30 xenografts. Note that the MDA PCA 144-13 xenograft line displays a large amount of necrosis. B. Diagram illustrates unsupervised hierarchal clustering using complete linkage and Pearson's correlation coefficient analysis of the raw expression profiles of androgen receptor-positive adenocarcinoma (MDA PCA 130, 117-9, 79) and androgen receptor-negative SCPC/LCNEC (MDA PCA 146-10, 155-2, 155-12, 144-13, 144-4) xenografts obtained with Affymetrix HGU133Plus2 array. In red are samples assessed on 1 day and in blue, those assessed on a different day. B, diagram illustrates Gene Ontology analysis results showing enrichment in biologic-process subtrees related to mitosis among the differently expressed genes. The biological processes that reached statistical significance are highlighted in red. SCPC, small-cell prostate carcinoma; LCNEC, large-cell neuroendocrine carcinoma.

Bioinformatic analysis of 94 patient-derived xenografts (PDXs), cell lines, and organoids (PCOs) identifies three intrinsic transcriptional subtypes of metastatic castration-resistant prostate cancer androgen receptor (AR) pathway + prostate cancer (PC) (ARPC), mesenchymal and stem-like PC (MSPC), and neuroendocrine PC (NEPC). A sizable proportion of castration-resistant and metastatic stage PC (M-CRPC) cases are admixtures of ARPC and MSPC. Analysis of clinical datasets and mechanistic studies indicates that MSPC arises from ARPC as a consequence of therapy-induced lineage plasticity. AR blockade with enzalutamide induces (1) transcriptional silencing of TP53 and hence dedifferentiation to a hybrid epithelial and mesenchymal and stem-like state and (2) inhibition of BMP signaling, which promotes resistance to AR inhibition. Enzalutamide-tolerant LNCaP cells re-enter the cell cycle in response to neuregulin and generate metastasis in mice. Combined inhibition of HER2/3 and AR or mTORC1 exhibits efficacy in models of ARPC and MSPC or MSPC, respectively. These results define MSPC, trace its origin to therapy-induced lineage plasticity, and reveal its sensitivity to HER2/3 inhibition.

Unsupervised clustering and deconvolution analysis identifies a novel subtype of M-CRPC endowed with hybrid epithelial/mesenchymal (E/M) and luminal progenitor-like traits (Mesenchymal and Stem-like PC, MSPC). Analysis of patient datasets and mechanistic studies indicate that MSPC arises as a consequence of therapy-induced lineage plasticity. AR blockade instigates two separate and complementary processes: 1) transcriptional silencing of TP53 and hence acquisition of hybrid E/M and stem-like traits; and 2) inhibition of the BMP signaling, which promotes resistance to the pro-apoptotic and anti-proliferative effects of AR inhibition. The drug-tolerant prostate cancer cells generated through reprogramming are rescued by neuregulin and generate metastases in mice. Combined inhibition of HER2/3 and AR or mTORC1 exhibit efficacy in preclinical models of mixed ARPC/MSPC or MSPC, respectively. These results identify a novel subtype of M-CRPC, trace its origin to therapy-induced lineage plasticity, and reveal its dependency on HER2/3 signaling.