Limits...
Androgen receptor profiling predicts prostate cancer outcome.

Stelloo S, Nevedomskaya E, van der Poel HG, de Jong J, van Leenders GJ, Jenster G, Wessels LF, Bergman AM, Zwart W - EMBO Mol Med (2015)

Bottom Line: Biomarkers for outcome prediction are urgently needed, so that high-risk patients could be monitored more closely postoperatively.These differential androgen receptor/chromatin interactions dictated expression of a distinct gene signature with strong prognostic potential.By combining existing technologies, we propose a novel pipeline for biomarker discovery that is easily implementable in other fields of oncology.

View Article: PubMed Central - PubMed

Affiliation: Division of Molecular Pathology, The Netherlands Cancer Institute, Amsterdam, The Netherlands.

Show MeSH

Related in: MedlinePlus

Genomewide profiling of chromatin accessibility in prostate cancer specimensOverview of the experimental design. Representative examples of H&E-stained slides are shown. DNA and proteins were cross-linked using formaldehyde, followed by shearing of the chromatin and phenol–chloroform extraction. The organic phase contains compacted DNA (protein–DNA complexes), while DNA recovered from the aqueous phase represents accessible regulatory regions. DNA from the aqueous phase is further purified and sequenced.Snapshots of accessible chromatin regions as assessed through FAIRE-seq in normal prostate tissue (blue), primary tumor (green), treatment-resistant tumor (red), and metastatic (gray) tissue. Reads are normalized to millions of sequenced reads per sample. Genomic coordinates are indicated.Genomic distribution of FAIRE peaks in normal and tumor samples.Boxplots depicting normalized FAIRE-seq read counts (RPKM) in different tissues across the peaks found in at least three samples. Read counts in benign tissue (blue) are lower than in primary tumor (green), therapy-resistant tumor (red) or metastasis (gray) (P < 2.2e−16, paired t-test).Venn diagram, visualizing shared and unique FAIRE peaks in normal prostate samples (blue) and primary prostate tumor samples (green).Heatmap showing raw read count intensity in FAIRE-seq peaks enriched in either normal or tumor samples. The window represents 5 kb around the FAIRE-seq peak.Heatmap showing raw read count intensity of ChIP-seq signal from multiple cell line datasets (Appendix Table S12 for references and GEO accession numbers) at accessible regions identified in primary tumors (peaks present in at least two specimens) ranked on peak intensity. Top panel depicts promoter regions, and the bottom panel, all other regions. The window represents 5 kb around the FAIRE-seq peak.Ingenuity pathway analyses, illustrating one of the networks based on motifs found in FAIRE-seq peaks that were present in at least two out of four primary tumors. Genes previously described to be involved in prostate cancer are highlighted in blue (other networks in Appendix Fig S3).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4644377&req=5

fig02: Genomewide profiling of chromatin accessibility in prostate cancer specimensOverview of the experimental design. Representative examples of H&E-stained slides are shown. DNA and proteins were cross-linked using formaldehyde, followed by shearing of the chromatin and phenol–chloroform extraction. The organic phase contains compacted DNA (protein–DNA complexes), while DNA recovered from the aqueous phase represents accessible regulatory regions. DNA from the aqueous phase is further purified and sequenced.Snapshots of accessible chromatin regions as assessed through FAIRE-seq in normal prostate tissue (blue), primary tumor (green), treatment-resistant tumor (red), and metastatic (gray) tissue. Reads are normalized to millions of sequenced reads per sample. Genomic coordinates are indicated.Genomic distribution of FAIRE peaks in normal and tumor samples.Boxplots depicting normalized FAIRE-seq read counts (RPKM) in different tissues across the peaks found in at least three samples. Read counts in benign tissue (blue) are lower than in primary tumor (green), therapy-resistant tumor (red) or metastasis (gray) (P < 2.2e−16, paired t-test).Venn diagram, visualizing shared and unique FAIRE peaks in normal prostate samples (blue) and primary prostate tumor samples (green).Heatmap showing raw read count intensity in FAIRE-seq peaks enriched in either normal or tumor samples. The window represents 5 kb around the FAIRE-seq peak.Heatmap showing raw read count intensity of ChIP-seq signal from multiple cell line datasets (Appendix Table S12 for references and GEO accession numbers) at accessible regions identified in primary tumors (peaks present in at least two specimens) ranked on peak intensity. Top panel depicts promoter regions, and the bottom panel, all other regions. The window represents 5 kb around the FAIRE-seq peak.Ingenuity pathway analyses, illustrating one of the networks based on motifs found in FAIRE-seq peaks that were present in at least two out of four primary tumors. Genes previously described to be involved in prostate cancer are highlighted in blue (other networks in Appendix Fig S3).

Mentions: We assessed chromatin accessibility in multiple prostate tissue specimens as well as the changes thereof in prostate cancer development and progression. Four normal prostate tissue samples, four primary tumors, and three ADT-resistant prostate tumors were assessed, as well as three prostate cancer metastases (Fig2A). FAIRE-seq was applied to identify accessible chromatin regions with gene-regulatory functions on a genomewide scale (Giresi & Lieb, 2009). FAIRE is based on phenol–chloroform mediated sample separation, in which accessible chromatin fragments can be separated from the condensed state, effectively enriching for regulatory genomic regions (schematically visualized in Fig2A). Metastases and prostate adenocarcinomas contained more than 70% tumor cells with a Gleason score ranging from 7 (3 + 4) to 10 (5 + 5), while all normal prostate tissues were derived from a healthy region from prostatectomy specimens. Tumor and normal tissues were validated by our pathologists. Clinicopathological parameters are shown in Appendix Table S1. The number of FAIRE peaks identified was highly variable between the tissues, ranging from 50 peaks up to over 13,000 peaks (Appendix Table S2). Figure2B shows four randomly selected representative coverage profiles of accessible chromatin at promoter regions. Over 50% of accessible chromatin sites in healthy and tumor specimens were found at promoter regions (Fig2C), and average signal for each specimen showed clear enrichment of reads at transcription start sites (Appendix Fig S1). Tumor samples showed more enriched chromatin accessibility at both intron and distal intergenic regions, as opposed to normal prostate tissue where FAIRE signal was mainly found at promoters (Fig2C).


Androgen receptor profiling predicts prostate cancer outcome.

Stelloo S, Nevedomskaya E, van der Poel HG, de Jong J, van Leenders GJ, Jenster G, Wessels LF, Bergman AM, Zwart W - EMBO Mol Med (2015)

Genomewide profiling of chromatin accessibility in prostate cancer specimensOverview of the experimental design. Representative examples of H&E-stained slides are shown. DNA and proteins were cross-linked using formaldehyde, followed by shearing of the chromatin and phenol–chloroform extraction. The organic phase contains compacted DNA (protein–DNA complexes), while DNA recovered from the aqueous phase represents accessible regulatory regions. DNA from the aqueous phase is further purified and sequenced.Snapshots of accessible chromatin regions as assessed through FAIRE-seq in normal prostate tissue (blue), primary tumor (green), treatment-resistant tumor (red), and metastatic (gray) tissue. Reads are normalized to millions of sequenced reads per sample. Genomic coordinates are indicated.Genomic distribution of FAIRE peaks in normal and tumor samples.Boxplots depicting normalized FAIRE-seq read counts (RPKM) in different tissues across the peaks found in at least three samples. Read counts in benign tissue (blue) are lower than in primary tumor (green), therapy-resistant tumor (red) or metastasis (gray) (P < 2.2e−16, paired t-test).Venn diagram, visualizing shared and unique FAIRE peaks in normal prostate samples (blue) and primary prostate tumor samples (green).Heatmap showing raw read count intensity in FAIRE-seq peaks enriched in either normal or tumor samples. The window represents 5 kb around the FAIRE-seq peak.Heatmap showing raw read count intensity of ChIP-seq signal from multiple cell line datasets (Appendix Table S12 for references and GEO accession numbers) at accessible regions identified in primary tumors (peaks present in at least two specimens) ranked on peak intensity. Top panel depicts promoter regions, and the bottom panel, all other regions. The window represents 5 kb around the FAIRE-seq peak.Ingenuity pathway analyses, illustrating one of the networks based on motifs found in FAIRE-seq peaks that were present in at least two out of four primary tumors. Genes previously described to be involved in prostate cancer are highlighted in blue (other networks in Appendix Fig S3).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4644377&req=5

fig02: Genomewide profiling of chromatin accessibility in prostate cancer specimensOverview of the experimental design. Representative examples of H&E-stained slides are shown. DNA and proteins were cross-linked using formaldehyde, followed by shearing of the chromatin and phenol–chloroform extraction. The organic phase contains compacted DNA (protein–DNA complexes), while DNA recovered from the aqueous phase represents accessible regulatory regions. DNA from the aqueous phase is further purified and sequenced.Snapshots of accessible chromatin regions as assessed through FAIRE-seq in normal prostate tissue (blue), primary tumor (green), treatment-resistant tumor (red), and metastatic (gray) tissue. Reads are normalized to millions of sequenced reads per sample. Genomic coordinates are indicated.Genomic distribution of FAIRE peaks in normal and tumor samples.Boxplots depicting normalized FAIRE-seq read counts (RPKM) in different tissues across the peaks found in at least three samples. Read counts in benign tissue (blue) are lower than in primary tumor (green), therapy-resistant tumor (red) or metastasis (gray) (P < 2.2e−16, paired t-test).Venn diagram, visualizing shared and unique FAIRE peaks in normal prostate samples (blue) and primary prostate tumor samples (green).Heatmap showing raw read count intensity in FAIRE-seq peaks enriched in either normal or tumor samples. The window represents 5 kb around the FAIRE-seq peak.Heatmap showing raw read count intensity of ChIP-seq signal from multiple cell line datasets (Appendix Table S12 for references and GEO accession numbers) at accessible regions identified in primary tumors (peaks present in at least two specimens) ranked on peak intensity. Top panel depicts promoter regions, and the bottom panel, all other regions. The window represents 5 kb around the FAIRE-seq peak.Ingenuity pathway analyses, illustrating one of the networks based on motifs found in FAIRE-seq peaks that were present in at least two out of four primary tumors. Genes previously described to be involved in prostate cancer are highlighted in blue (other networks in Appendix Fig S3).
Mentions: We assessed chromatin accessibility in multiple prostate tissue specimens as well as the changes thereof in prostate cancer development and progression. Four normal prostate tissue samples, four primary tumors, and three ADT-resistant prostate tumors were assessed, as well as three prostate cancer metastases (Fig2A). FAIRE-seq was applied to identify accessible chromatin regions with gene-regulatory functions on a genomewide scale (Giresi & Lieb, 2009). FAIRE is based on phenol–chloroform mediated sample separation, in which accessible chromatin fragments can be separated from the condensed state, effectively enriching for regulatory genomic regions (schematically visualized in Fig2A). Metastases and prostate adenocarcinomas contained more than 70% tumor cells with a Gleason score ranging from 7 (3 + 4) to 10 (5 + 5), while all normal prostate tissues were derived from a healthy region from prostatectomy specimens. Tumor and normal tissues were validated by our pathologists. Clinicopathological parameters are shown in Appendix Table S1. The number of FAIRE peaks identified was highly variable between the tissues, ranging from 50 peaks up to over 13,000 peaks (Appendix Table S2). Figure2B shows four randomly selected representative coverage profiles of accessible chromatin at promoter regions. Over 50% of accessible chromatin sites in healthy and tumor specimens were found at promoter regions (Fig2C), and average signal for each specimen showed clear enrichment of reads at transcription start sites (Appendix Fig S1). Tumor samples showed more enriched chromatin accessibility at both intron and distal intergenic regions, as opposed to normal prostate tissue where FAIRE signal was mainly found at promoters (Fig2C).

Bottom Line: Biomarkers for outcome prediction are urgently needed, so that high-risk patients could be monitored more closely postoperatively.These differential androgen receptor/chromatin interactions dictated expression of a distinct gene signature with strong prognostic potential.By combining existing technologies, we propose a novel pipeline for biomarker discovery that is easily implementable in other fields of oncology.

View Article: PubMed Central - PubMed

Affiliation: Division of Molecular Pathology, The Netherlands Cancer Institute, Amsterdam, The Netherlands.

Show MeSH
Related in: MedlinePlus