Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Sep 11;41(9):1567-1585.e7.
doi: 10.1016/j.ccell.2023.07.013. Epub 2023 Aug 14.

Integrative multi-omic cancer profiling reveals DNA methylation patterns associated with therapeutic vulnerability and cell-of-origin

Wen-Wei Liang  1 , Rita Jui-Hsien Lu  1 , Reyka G Jayasinghe  1 , Steven M Foltz  1 , Eduard Porta-Pardo  2 , Yifat Geffen  3 , Michael C Wendl  4 , Rossana Lazcano  5 , Iga Kolodziejczak  6 , Yizhe Song  1 , Akshay Govindan  1 , Elizabeth G Demicco  7 , Xiang Li  1 , Yize Li  1 , Sunantha Sethuraman  1 , Samuel H Payne  8 , David Fenyö  9 , Henry Rodriguez  10 , Maciej Wiznerowicz  11 , Hui Shen  12 , D R Mani  13 , Karin D Rodland  14 , Alexander J Lazar  5 , Ana I Robles  10 , Li Ding  15 ; Clinical Proteomic Tumor Analysis Consortium
Collaborators, Affiliations

Integrative multi-omic cancer profiling reveals DNA methylation patterns associated with therapeutic vulnerability and cell-of-origin

Wen-Wei Liang et al. Cancer Cell. .

Abstract

DNA methylation plays a critical role in establishing and maintaining cellular identity. However, it is frequently dysregulated during tumor development and is closely intertwined with other genetic alterations. Here, we leveraged multi-omic profiling of 687 tumors and matched non-involved adjacent tissues from the kidney, brain, pancreas, lung, head and neck, and endometrium to identify aberrant methylation associated with RNA and protein abundance changes and build a Pan-Cancer catalog. We uncovered lineage-specific epigenetic drivers including hypomethylated FGFR2 in endometrial cancer. We showed that hypermethylated STAT5A is associated with pervasive regulon downregulation and immune cell depletion, suggesting that epigenetic regulation of STAT5A expression constitutes a molecular switch for immunosuppression in squamous tumors. We further demonstrated that methylation subtype-enrichment information can explain cell-of-origin, intra-tumor heterogeneity, and tumor phenotypes. Overall, we identified cis-acting DNA methylation events that drive transcriptional and translational changes, shedding light on the tumor's epigenetic landscape and the role of its cell-of-origin.

PubMed Disclaimer

Conflict of interest statement

Declaration of interests The authors declare no competing interests.

Figures

Figure 1.
Figure 1.. Correlations between promoter DNA methylation, transcriptome, and proteome
(A) Left: A schematic of tumor types collected for this study. Right: Density plot showing the distribution of adjusted R values between protein abundance and RNA expression (yellow), promoter methylation and RNA expression (orange) or protein abundance (green). (B) Scatter plot showing adjusted R values distribution for genes with promoter methylation correlated with RNA (left, orange), protein (middle, green), or both RNA and protein (right, blue). (C) Upsetplot showing the breakdown of genes based on their correlation with promoter methylation and RNA/protein expression. (D) Examples with distinct correlations between promoter methylation, RNA expression, and protein abundance. Each dot represents one tumor, with the solid line representing the correlation between scaled RNA and promoter methylation, and the dashed line representing the correlation between scaled protein abundance and promoter methylation. Significant correlations are highlighted (orange: RNA only; green: protein only, blue and purple: RNA and protein). (E) Examples of anti-correlated genes with promoter hypermethylation, exhibiting upregulation at the RNA level and downregulation at the protein level. (F) Pathway enrichment analysis of the 31 anti-correlated genes with promoter hypermethylation, showing upregulation at the RNA level and downregulation at the protein level. Pathway with FDR P-value <0.05 is highlighted in bold. See also Figures S1, S2, and Table S1.
Figure 2.
Figure 2.. The cancer methylome landscape associated with transcriptomic and proteomic change
(A) RNA expression (upper) and protein abundance (lower) changes in genes between aberrant and normal samples. Y axis corresponds to the statistical significance of aberrant DNA methylation with changed expression and x axis to the median difference of gene expression of samples with or without aberrant DNA methylation. Representative genes are colored based on methylation status: yellow, hypermethylation; blue, hypomethylation. Dot size indicates the number of CGIs associated with expression changes. (B) Venn diagrams showing the number of hypermethylated (upper) and hypomethylated (lower) genes having significant RNA expression and/or protein abundance changes. (C) Common and cancer type-specific aberrant methylations of cancer-associated genes. Shading of the filled circle indicates the median methylation difference (upper), RNA expression difference (middle), and protein abundance difference (lower) between aberrant and normal samples at significant CpG sites. Dot size is proportional to the number of samples harboring indicated aberrant DNA methylation events in the cancer cohort. See also Figure S2 and Table S2.
Figure 3.
Figure 3.. Characterization of aberrant methylation in driver genes
(A) Distribution of the number of transcription factor binding sites for functional hypermethylation (yellow) and hypomethylation (blue). Enriched motif was highlighted in the inset. (B) Mutual exclusivity and co-occurrence of genomic and epigenomic alterations in driver genes in LSCC. (C) Violin plot comparing histone H3 acetylation levels between highly hypermethylated tumors and lowly hypermethylated tumors. Boxes represent the interquartile range (IQR), with the median frequency indicated by the horizontal line. Whiskers extend from the boxes to indicate the data range. Statistically significant differences between groups were determined using Wilcoxon rank sum test. (D) Heatmap of histone sites exhibiting significant differential acetylation (FDR < 0.1) among immune subtypes. The grayscale color scale denotes the hypermethylation frequency in each tumor. (E) RNA expression (upper) and protein abundance (lower) levels stratified by IDH2 genomic alterations and IDH2 hypomethylation (blue) versus IDH2 normal methylation (gray), with ** denoting Wilcoxon P<0.005. Median values are shown as solid black lines, and first and third quartiles are represented by dashed lines. (F) Pathway diagram illustrating the average difference in RNA expression (left square) and protein abundance (right square) between IDH2 hypomethylated samples and normal methylated samples in LSCC. The shading of the filled squares indicates the extent of the differences. (G) Positive (upper) and negative (lower) correlation coefficients of histone acetylation levels and methylation levels at α-KG target genes among IDH1/IDH2 wild-type, IDH1 mutant, IDH2 hypomethylated samples, and IDH2 mutant. The breakdown of each group was shown in the pie chart below. Boxes represent the IQR, with the median correlation value indicated by the horizontal line. Whiskers extend from the boxes to show the data range. Statistically significant differences between groups were determined using FDR-corrected P-values, with **** indicating P < 2.2e-16. See also Figure S3 and Table S2.
Figure 4.
Figure 4.. Collaborative effects of FGFR2 mutations and hypomethylation on FGFR2 upregulation
(A) Lolliplot showing missense mutations of FGFR2 in UCEC samples. The amino acids and types of mutations are labeled. Positions that are recurrently mutated are highlighted with the number of occurrences. The FGFR2 functional domains are colored. (B) Unsupervised clustering of UCEC tumors (upper) and normal adjacent tissues (lower) based on DNA methylation of the FGFR2 promoter. (C) Correlation of methylation with gene expression (upper) and protein abundance (lower). Samples are colored based on genetic and/or epigenetic alterations of FGFR2. Tumors harboring FGFR2 hypomethylation are highlighted by large dot size. (D) RNA expression (upper) and protein abundance (lower) levels stratified by FGFR2 genomic alterations and FGFR2 hypomethylation (blue) versus FGFR2 normal methylation (gray), with ** denoting Wilcoxon P<0.005. Median values are shown as solid black lines, and first and third quartiles are represented by dashed lines. See also Figure S4.
Figure 5.
Figure 5.. STAT5A hypermethylation associated with pervasive STAT5A regulon changes
(A) Unsupervised clustering of STAT5A regulon genes using Pearson correlation of scaled RNA sequencing data. Annotations denote STAT5A expression and methylation levels. Mean activity indicates the overall sum of regulon activity. The color scale is proportional to expression (red: upregulation; blue: downregulation). (B) Unsupervised clustering of STAT5A regulon genes using Pearson correlation of scaled global proteome data. (C) Violin plot comparing regulon activity in hypermethylated STAT5A (yellow) and normally methylated STAT5A (gray) samples. Median values are shown as solid black lines, and first and third quartiles are represented by dashed lines. Statistical significance was determined using a Wilcoxon rank-sum test. (D) Pathway members and interactions in the STAT5A regulon. The mean expression differences between STAT5A hypermethylated samples and normally methylated samples are indicated by shading of the filled squares. See also Figure S5 and Table S3.
Figure 6.
Figure 6.. Functional impact of STAT5A hypermethylation on immune cell depletion in HNSCC
(A) Heatmaps showing distinct immune subtypes of HNSCC tumors derived from xCell enrichment scores. The top panel shows the immune score, DNA methylation status of STAT5A, immune subtype, and tumor stage. (B) Violin plots comparing xCell enrichment scores of immune effectors and dendritic cells in hypermethylated STAT5A (yellow) and normally methylated STAT5A (gray) samples in HNSCC tumors. Median values are shown as solid black lines, and first and third quartiles are represented by dashed lines. Statistical significance was determined using a Wilcoxon rank-sum test. (C) Representative image of IHC (immunohistochemistry) staining of STAT5A protein (brown) in HNSCC tumor sample. Scale bar = 100 μm. (D) Correlation between the quantified STAT5A protein abundance versus the level of tumor-infiltrating lymphocytes (TILs, left panel) or peritumoral lymphocytes (right panel). Samples were colored by the DNA methylation status of STAT5A (yellow: hypermethylation, gray: normal methylation). Samples with representative IHC images are highlighted by large dot size. (E) IHC staining of STAT5A in HNSCC tumors. Representative tumor cells (stars) and lymphocytes (arrows) are shown. Tumor boundary is indicated by a black line. Scale bar = 100 μm. See also Figure S6 and Table S3.
Figure 7.
Figure 7.. Summary of the cancer methylome for cell-of-origin, tumor signatures, and therapy
(A) Projection of the cancer methylomes. Each point is a sample and is colored based on the cancer type (first column), sample type (second column), methylation subtype (third column), or multi-cancer methylation group (fourth column). (B) Heatmap of differentially methylated CpG sites at promoter regions in seven cancer cohorts compared to normal adjacent tissue. Selected promoters annotated with the number of differentially methylated CpG sites are shown. (C) Alluvial plot showing the per-cancer methylation subtypes (second row), their enriched significantly mutated genes (SMGs) (first row), enriched RNA expression signature (third row), and enriched protein signature (fourth row). The curved lines across panels correspond to different methylation subtypes. Signatures with FDR P-values < 0.05 are highlighted with *. See also Figure S7 and Tables S4-S6.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Feinberg AP, Ohlsson R, and Henikoff S. (2006). The epigenetic progenitor origin of human cancer. Nat. Rev. Genet. 7, 21–33. - PubMed
    1. Baylin SB, and Jones PA (2011). A decade of exploring the cancer epigenome - biological and translational implications. Nat. Rev. Cancer 11, 726–734. - PMC - PubMed
    1. Shen H, and Laird PW (2013). Interplay between the cancer genome and epigenome. Cell 153, 38–55. - PMC - PubMed
    1. Ehrlich M. (2002). DNA methylation in cancer: too much, but also too little. Oncogene 21, 5400–5413. - PubMed
    1. Berman BP, Weisenberger DJ, Aman JF, Hinoue T, Ramjan Z, Liu Y, Noushmehr H, Lange CPE, van Dijk CM, Tollenaar RA, et al. (2012). Regions of focal DNA hypermethylation and long-range hypomethylation in colorectal cancer coincide with nuclear lamina--associated domains. Nat. Genet. 44, 40. - PMC - PubMed

Publication types

LinkOut - more resources