Paper - Review

DOI: 10.1038/s41588-020-0648-8

Abstract

⭐ DNA methylation
→ a key regulator ← of gene expression

Discovered → alterations affecting driver genes
← which were detectable only ← with integrated whole-genome approaches
← Through whole-genome bisulfite sequencing ← of castration-resistant prostate metastases

Observed that → 22% of tumors exhibited → a novel epigenomic subtype
← associated with hyper-methylation & somatic mutations
← in 1⃣ TET2 2⃣ DNMT3B 3⃣ IDH1 4⃣ BRAF

Identified → intergenic regions
← where methylation is associated
← with RNA expression ← of the oncogenic driver genes

Showed that → differential methylation
← during progression preferentially
← occurs at 1⃣ somatic mutational hotspots 2⃣ putative regulatory regions

Introduction

DNA methylation
→ is a pervasive epigenomic mechanism ← of gene regulation

Most CpG dinucleotides
→ are methylated
→ frequently mark → gene regulatory loci

Aberrant methylation
→ has been implicated ← in oncogenesis
→ differences ← in methylation patterns ← between tumors & benign tissues

Many studies → compared DNA methylation patterns
← between primary PCa (Prostate Cancer) & benign prostate
← between subtypes of primary PCa

mCRPC
← metastatic castration-resistant prostate cancer
→ is the lethal form of the disease

Complete epigenetic landscape
→ remains largely unknown
← Although 1⃣ the genomic 2⃣ the trascriptomic landscape (← of mCRPC) → has been well characterized

Many important regulatory regions
→ are outside the profiled areas
WGBS
← whole-genome bisulfite sequencing
→ is required → to systematically study → the entire genome ← at single-base-level resolution

Describe
→ a WGBS study ← in a metastatic cancer
← which integrated with matched deep WGS & RNA-seq ← in the same samples

Results

Obtained → fresh-frozen core biopsies ← of metastases
←from 100 patients with mCRPC

WGBS
→ was performed
← on 1⃣ 100 biopsy samples 2⃣ 10 matched benign tissue samples

1⃣ Bone 2⃣ Lymph node 3⃣ Liver
→ were represented ← in these benign-adjacent samples
← which exhibited distinct methylation patterns ← from the tumor samples

Novel CpG methylation subtype of metastatic castration-resistant prostate cancer

The total number of HMRs
→ ranged from 24388 to 85474 per sample

Most inter-sample variation
→ was outside promoters & (CpG islands &shores & shelves)
→ manifested in gene bodies and regulatory regions

Tumors with more HMRs
→ had significantly higher ← genome copy number alteration frequencies

HMR frequencies → was NOT ❌ associated
←with mutation & structural variant frequenc

DNA methylation
→ has been best characterized
← at the CpG islands
← in the promoter regions of genes

Recurrent intergenic HMRs
→ would be associated ← with regulatory loci

Un-supervised hierarchical clustering (← of rHMRs)
→ identified subgroups of tumors ← with distinct patterns of methylation

Identified → a novel subtype of mCRPC
← with significantly higher methylation levels ← at rHMRs

The CMP subtype
→ was NOT ❌ significantly associated
← with the anatomic site of the biopsy

(Benign prostate tumors) & (t-SCNC tumors)
→ formed separate clusters

Several CMP tumors harbored
→ mutually exclusive mutations
← in 1⃣ TET2 2⃣ IDH1 3⃣ BRAF

Mutations ← in these genes
→ have been increased → CpG methylation in other tumor types

Confirmed → the absence (← of these mutations) ← in peripheral blood germline DNA
← using both WGS & Sanger sequencing
← to assess the potential ← for mis-attribution of somatic mutations

Computational prediction (← of mutation consequences) ← by FATHMM
→ predicted 1⃣ (p.His1380Leu, p.Tyr1421His and p.Arg1808Thr) → to be deleterious 2⃣ p.Thr1499Arg → to be benign

CMP status
→ was independently associated
← with HMR number in 1⃣ CpG islands 2⃣ shores 3⃣ shelves 4⃣ CpG open seas
← after adjusting for tumor purity

Regional analysis of methylation

Long-range epigenetic activation & repression
→ is a phenomenon ← which large regions containing (multiple gene) → are concomitantly (activated & repressed)
∵ Concordant epigenetic changes

Identified
→ 14 candidate long-range interactions

PMDs
← Partially Methylated Domains
→ are (genomic regions) ← with incomplete loss of methylation

Modest correlation
← between 1⃣ PMD frequency 2⃣ HMR frequency
Fraction of the genome harboring PMDs
→ was NOT ❌ significantly different ← between 1⃣ Benign prostate 2⃣ Primary PCa 3⃣ mCRPC samples

Genome PMD fraction
→ was NOT ❌ significantly correlated
← with 1⃣ tumor purity 2⃣ total number of mutations 3⃣ percentage copy number altered in mCRPC

Level of PMD methylation
→ was significantly higher
← in CMP subtype > in non-CMP subtype

Identified → DMVs
← DNA methylation valleys
→ broad regions of hypo-methylation → which are associated with 1⃣ the activating histone mark 2⃣ the repressive histone mark

DMV frequencies
→ are more frequently associated ← with the activating histone mark
→ Tumors (← with many DMVs) → coincided with a nearly equal proportion

t-SCNS
← treatment-induced small cell neuroendocrine carcinoma
→ harbored distinct genome-wide methylation patterns
← which used enhanced reduced-representation bisulfite sequencing

Differential prostate cancer gene promoter methylation

Genes ← with higher expression
→ had more frequent ← 1⃣ promoter hypo-methylation 2⃣ gene-body hyper-methylation

Negative correlation
← of CPG methylation & gene expression
→ peaked at the gene promoter
Positive correlation
→ peaked in the gene body

Identified → rHMRs correlated with (expression of genes)
∴ eHMRs
← expression-associated hypo-methylated regions

Strongest positive correlations
← 30% of the total
→ fell at the 3' end of the gene body

Expanded → our analysis → to test for association in candidate enhancer regions
HMRs → identified in a 1-Mb windows around the TSS
← TSS: transcription start site

Candidate enhancers → were identified
← by the presence of H3K27ac peaks ← in primary prostate tumors

The association
← between methylation levels & expression
→ tended to be stronger ← in regions ← that were physically close to the TSS

Key androgen-response genes
→ demonstrated promoter hypo-methylation in mCRPC
← compared to benign prostate samples

NOT ❌ observe
→ promoter hyper-methylation of tumor suppressors
← in mCRPC tumors ← compared to benign prostate samples

Many gene ← with PCa-specific expression
→ lack PCa-specific DNA-sequence alterations

Performed → an unbiased analysis
→ compared eHMR correlation strength ← in all genes
→ to their expression variability
→ to test the model ← which methylation influences disease-specific expression

Novel intergenic regulatory regions of the AR gene

DNA methylation → may operate
← in tandem with other somatic DNA alterations
← which influence gene expression

Gene expression → was significantly associated
← with 1⃣ local DNA copy number alterations 2⃣ mutations 3⃣ structural variants

Methylation improved
→ the fit of a model
→ form gene expression → beyond DNA alteration alone

The top enriched → MSigDb hallmark pathway
← Molecular Signatures Database
→ for genes with improved fit
∴ Androgen response

∴ This finding supports → the role of methylation (← in androgen pathway activity) ← in mCRPC

Identified → a distal AR enhancer region
← which DNA copy number amplifications are associated ← with elevated AR expression

Identified → multiple eHMRs near AR
← including adjacent to the AR promoter

The AR promoter
→ was hypo-methylated ← in all tissues evaluated
Other eHMRs
→ was identified only in mCRPC samples
→ NOT ❌ in 1⃣ benign-adjacent tissue 2⃣ benign prostate 3⃣ primary PCa samples

AR & ERG ChIA-PET data
→ indicated that → long-range chromatin interactions exists
← between many of these loci
← which supports (the potential) → for physical interactions ← between these loci

AR gene body & the upstream enhancer
→ were amplified in a total of (81 % of mCRPC samples)

These data → are compatible with a model
← in which (selective pressure of androgen deprivation therapy) ← favors broad amplification
← that span (multiple enhancer) → to drive AR expression in mCRP

This analysis identified
→ focal genome loci
← which may represent novel intergenic regulatory regions of AR

Methylation associated with TMPRSS2-ERG and MYC expression

Half of prostate cancers
→ are defined by over-expression of the oncogenic transcript factor
← which encoded by ERG

ERG expression
→ is negligible in PCa
← unless it is activated ← by gene fusions that bypass the ERG promoter

ERG expression levels
→ vary widely
← within 1⃣ TMPRSS2-ERG fusion-positive tumors 2⃣ a linear model (← which predicts ERG expression) 3⃣ mutation status (← which provided a poor fit)

Identified rHMRs upstream TMPRSS2
← which co-localized with the TFBS of 1⃣ HOXB13 2⃣ FOXA1 3⃣ AR 4⃣ ERG

Methylation (← at these loci)
→ was negatively associated
← with ERG expression in only the fusion-positive samples
← which is consistent with a model ← in which TFBS methylation modulates expression

Prediction of ERG expression
→ was significantly improved
← by (the addition of methylation) at all rHMRs upstream of TMPRSS2

Methylation ← at regulatory regions (upstream of TMPRSS2)
→ contribute to this subtype

Oncogene MYC
→ is amplified
← in 38% of mCRPC samples

DNA interactions
← between PVT1 & MYC
→ were present ← in the VCaP ChIA-PET data

Methylation and prostate cancer progression

WGBS data
← on benign prostate & localized PCa samples
→ to identify DMRs
← Differentially Methylated Regions

mCRPC samples
→ were pre-dominantly (less methylated)
← than primary PCa

Global hypo-methylation (← in cancer)
→ may contribute → to genomic instability

Regions ← with more differential hypo-methylation in mCRPC
→ had an elevated somatic mutation rate in mCRPC

Discussion

Global analysis of methylation (← in mCRPC)
← with 1⃣ on 100 tumor samples 2⃣ 10 matched benign-adjacent metastatic regions
← integrated with 1⃣ matched deep WGS 2⃣ RNA-seq of the same samples

Data → identified →
1⃣ a novel epigenetic subtype of mCRPC
2⃣ new intergenic regulatory regions of AR
3⃣ the interplay ← between somatic & epigenetic alterations

Demonstrated → global methylome changes
→ which distinguish 1⃣ benign prostate 2⃣ primary PCa 3⃣ mCRPC
Found that → somatic mutations & putative regulatory regions

Identified → a new epigenetic CMP subtypes of mCRPC
← which is characterized ← by hyper-methylation both (within & outside) 1⃣ CpG islands 2⃣ shores 3⃣ shelves

The present study → cannot ❌ determine
← whether any mutations (← which we observed) → could drive methylation changes

---