CoChat
Sign Up
Cited by 328Epigenetics and DNA MethylationCancer Genomics and DiagnosticsTelomeres, Telomerase, and Senescence

The DOI matches its Crossref record; 71 of the 71 references it cites were found in a registry or an index.

Correlation of an epigenetic mitotic clock with cancer risk

Open access
Yang, Zhen, Wong, Andrew, Kuh, Diana, Paul, Dirk S., Rakyan, Vardhman K., Leslie, R. David, Zheng, Shijie C., Widschwendter, Martin, Beck, Stephan, Teschendorff, Andrew E.·2016·Genome Biology
DOI 10.1186/s13059-016-1064-3PMC PMC5046977
Open this paper in CoChat — free accountOpen PDF

Abstract

BACKGROUND: Variation in cancer risk among somatic tissues has been attributed to variations in the underlying rate of stem cell division. For a given tissue type, variable cancer risk between individuals is thought to be influenced by extrinsic factors which modulate this rate of stem cell division. To date, no molecular mitotic clock has been developed to approximate the number of stem cell divisions in a tissue of an individual and which is correlated with cancer risk. RESULTS: Here, we integrate mathematical modeling with prior biological knowledge to construct a DNA methylation-based age-correlative model which approximates a mitotic clock in both normal and cancer tissue. By focusing on promoter CpG sites that localize to Polycomb group target genes that are unmethylated in 11 different fetal tissue types, we show that increases in DNA methylation at these sites defines a tick rate which correlates with the estimated rate of stem cell division in normal tissues. Using matched DNA methylation and RNA-seq data, we further show that it correlates with an expression-based mitotic index in cancer tissue. We demonstrate that this mitotic-like clock is universally accelerated in cancer, including pre-cancerous lesions, and that it is also accelerated in normal epithelial cells exposed to a major carcinogen. CONCLUSIONS: Unlike other epigenetic and mutational clocks or the telomere clock, the epigenetic clock proposed here provides a concrete example of a mitotic-like clock which is universally accelerated in cancer and precancerous lesions.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1186/s13059-016-1064-3

Cites

71 of 71 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 Flowchart of epiTOC. a The epiTOC score (also pcgtAge score) is estimated over 385 PCGT/PRC2-marked promoter CpGs that are constitutively unmethylated in over 37 fetal tissue samples from 12 tissue types and whose DNAm increases with chronological age in a large cohort of healthy individuals, as assessed in one tissue type (blood). The epiTOC/pcgtAge score of any sample represents the average DNAm from the ground fetal state over these 385 sites, representing the life-time accumulation of DNAm replication errors, and provides a relative estimate of the number of stem cell divisions per stem cell. The lower panel shows how this score varies linearly with chronological age (data as estimated in the Hannum et al. [28] blood dataset). b Validation of the epiTOC/pcgtAge score by correlation to the estimated intrinsic rate of stem cell division per stem cell in normal tissue samples from The Cancer Genome Atlas (TCGA) for which such estimates are available. Further validation of epiTOC/pcgtAge by correlation to an mRNA expression-based mitotic index in cancer tissues from TCGA. c Assessment of whether epiTOC/pcgtAge is accelerated in precancerous lesions (lung carcinoma in situ (LCIS) and ductal carcinoma in situ of the breast (DCIS) and cancer (TCGA) and whether it can predict the prospective risk of invasive lung cancer (ILC)
Fig. 1 Flowchart of epiTOC. a The epiTOC score (also pcgtAge score) is estimated over 385 PCGT/PRC2-marked promoter CpGs that are constitutively unmethylated in over 37 fetal tissue samples from 12 tissue types and whose DNAm increases with chronological age in a large cohort of healthy individuals, as assessed in one tissue type (blood). The epiTOC/pcgtAge score of any sample represents the average DNAm from the ground fetal state over these 385 sites, representing the life-time accumulation of DNAm replication errors, and provides a relative estimate of the number of stem cell divisions per stem cell. The lower panel shows how this score varies linearly with chronological age (data as estimated in the Hannum et al. [28] blood dataset). b Validation of the epiTOC/pcgtAge score by correlation to the estimated intrinsic rate of stem cell division per stem cell in normal tissue samples from The Cancer Genome Atlas (TCGA) for which such estimates are available. Further validation of epiTOC/pcgtAge by correlation to an mRNA expression-based mitotic index in cancer tissues from TCGA. c Assessment of whether epiTOC/pcgtAge is accelerated in precancerous lesions (lung carcinoma in situ (LCIS) and ductal carcinoma in situ of the breast (DCIS) and cancer (TCGA) and whether it can predict the prospective risk of invasive lung cancer (ILC)CC BY · PubMed Central
Fig. 2 epiTOC correlates with chronological age in purified cell and stem cell populations and is independent of blood cell type composition. a pcgtAge versus chronological age in three purified blood cell subtype sample sets, as indicated. Number of samples is indicated at the top of each panel. The green dashed line is a linear least squares fit. R2 value, Pearson correlation coefficient (PCC) and linear regression P values are given. b The fraction of the 385 PCGT CpGs that make up epiTOC which are significantly hypermethylated with age in each of these three purified sample sets, as well as the fraction of these 385 CpGs which are significantly hypermethylated with age in at least one of these three sets (ALL, red). c Gene set enrichment analysis odds ratios (OR) of immune and blood cell subtype terms that are highly enriched among age-associated promoter CpGs in Hannum et al. [28] without adjustment for cellular heterogeneity (Unadjusted, magenta bars). Adjusted P values (adjP) are given. ORs for these same biological terms among the 385 PCGT CpGs of epiTOC are also shown (cyan bars), indicating strong underenrichment. d Left panel: Correlation of the pcgtAge score with donor age of eight bone marrow-derived mesenchymal stem cell (MSC) populations (all of low passage). R2 value, Pearson correlation coefficient (PCC) and linear regression P value are given. Right panel: Correlation of pcgtAge score with donor age of 12 CD34+ hematopoietic progenitor cell (HPC) populations. Here we used a Wilcoxon rank sum test to derive a P value since there are two well defined groups of cord blood (n = 7, age zero) and adult peripheral blood (AdultPB, n = 5) samples
Fig. 2 epiTOC correlates with chronological age in purified cell and stem cell populations and is independent of blood cell type composition. a pcgtAge versus chronological age in three purified blood cell subtype sample sets, as indicated. Number of samples is indicated at the top of each panel. The green dashed line is a linear least squares fit. R2 value, Pearson correlation coefficient (PCC) and linear regression P values are given. b The fraction of the 385 PCGT CpGs that make up epiTOC which are significantly hypermethylated with age in each of these three purified sample sets, as well as the fraction of these 385 CpGs which are significantly hypermethylated with age in at least one of these three sets (ALL, red). c Gene set enrichment analysis odds ratios (OR) of immune and blood cell subtype terms that are highly enriched among age-associated promoter CpGs in Hannum et al. [28] without adjustment for cellular heterogeneity (Unadjusted, magenta bars). Adjusted P values (adjP) are given. ORs for these same biological terms among the 385 PCGT CpGs of epiTOC are also shown (cyan bars), indicating strong underenrichment. d Left panel: Correlation of the pcgtAge score with donor age of eight bone marrow-derived mesenchymal stem cell (MSC) populations (all of low passage). R2 value, Pearson correlation coefficient (PCC) and linear regression P value are given. Right panel: Correlation of pcgtAge score with donor age of 12 CD34+ hematopoietic progenitor cell (HPC) populations. Here we used a Wilcoxon rank sum test to derive a P value since there are two well defined groups of cord blood (n = 7, age zero) and adult peripheral blood (AdultPB, n = 5) samplesCC BY · PubMed Central
Fig. 3 The pcgtAge/epiTOC model approximates a mitotic clock. a The x-axis labels the cumulative total number of stem cell divisions (TNSC) incurred per stem cell in the given normal tissue in a log2 scale. This number is the log2 of the product of the intrinsic rate of cell divisions per year per stem cell (tissue-dependent) with the chronological age of the sample (tissue-independent). The y-axis labels the pcgtAge score of the sample. Samples have been colored according to the normal tissue of origin (normal samples adjacent to TCGA cancer types, as indicated). The number of normal samples for each tissue is given. Samples are seen to cluster into three main groups: a low cellular turnover group, a medium cellular turnover group, and a high cellular turnover group (note that this grouping is inferred from the unlogged scale, as shown in b). b Same as panel a but with TNSC shown in the unlogged scale. The P value is from a linear regression adjusted for chronological age of samples. c As b but for Horvath’s age-acceleration measure (i.e., DNAmAge minus chronological age). The P value is from a linear regression (already adjusted for age from the definition of age acceleration). d pcgtAge scores for the three different turnover groups as shown in a and b. The P values are from comparing the neighboring groups using a one-tailed Wilcoxon rank sum test, adjusted for chronological age. e As d but for Horvath’s age-acceleration measure. f The Pearson correlation coefficient (PCC) between either pcgtAge or Horvath’s age-acceleration measure, with an mRNA expression-based mitotic index, as estimated over all cancer samples of a given cancer type for a total of 15 different TCGA cancer types. Thus, each boxplot has 15 data points. The P value is from a one-tailed paired Wilcoxon rank sum test. g Corresponding heat map indicating for which cancer types there was a significant positive PCC between pcgtAge or Horvath’s AgeAccel with the mRNA expression-based mitotic index. The binomial test P value indicates that pcgtAge correlates more strongly with the mitotic index than Horvath’s age-acceleration measure
Fig. 3 The pcgtAge/epiTOC model approximates a mitotic clock. a The x-axis labels the cumulative total number of stem cell divisions (TNSC) incurred per stem cell in the given normal tissue in a log2 scale. This number is the log2 of the product of the intrinsic rate of cell divisions per year per stem cell (tissue-dependent) with the chronological age of the sample (tissue-independent). The y-axis labels the pcgtAge score of the sample. Samples have been colored according to the normal tissue of origin (normal samples adjacent to TCGA cancer types, as indicated). The number of normal samples for each tissue is given. Samples are seen to cluster into three main groups: a low cellular turnover group, a medium cellular turnover group, and a high cellular turnover group (note that this grouping is inferred from the unlogged scale, as shown in b). b Same as panel a but with TNSC shown in the unlogged scale. The P value is from a linear regression adjusted for chronological age of samples. c As b but for Horvath’s age-acceleration measure (i.e., DNAmAge minus chronological age). The P value is from a linear regression (already adjusted for age from the definition of age acceleration). d pcgtAge scores for the three different turnover groups as shown in a and b. The P values are from comparing the neighboring groups using a one-tailed Wilcoxon rank sum test, adjusted for chronological age. e As d but for Horvath’s age-acceleration measure. f The Pearson correlation coefficient (PCC) between either pcgtAge or Horvath’s age-acceleration measure, with an mRNA expression-based mitotic index, as estimated over all cancer samples of a given cancer type for a total of 15 different TCGA cancer types. Thus, each boxplot has 15 data points. The P value is from a one-tailed paired Wilcoxon rank sum test. g Corresponding heat map indicating for which cancer types there was a significant positive PCC between pcgtAge or Horvath’s AgeAccel with the mRNA expression-based mitotic index. The binomial test P value indicates that pcgtAge correlates more strongly with the mitotic index than Horvath’s age-acceleration measureCC BY · PubMed Central
Fig. 4 The pcgtAge/epiTOC model predicts universal age acceleration in cancer. Boxplots comparing the pcgtAge of age-matched normal cancer samples across 15 TCGA cancer types. P values are from a paired (one-tailed) Wilcoxon rank sum test. BRCA breast adenocarcinoma; BLCA bladder adenocarcinoma; COAD colon adenocarcinoma; ESCA esophageal cancer; HNSC head & neck squamous cell carcinoma; KIRC kidney renal cell carcinoma; KIRP kidney renal papillary carcinoma; LIHC liver hepatocellular carcinoma; LSCC lung squamous cell carcinoma; LUAD lung adenocarcinoma; PAAD pancreatic adenocarcinoma; PRAD prostate adenocarcinoma; READ rectal adenocarcinoma; THCA thyroid carcinoma; UCEC uterine corpus endometrial carcinoma
Fig. 4 The pcgtAge/epiTOC model predicts universal age acceleration in cancer. Boxplots comparing the pcgtAge of age-matched normal cancer samples across 15 TCGA cancer types. P values are from a paired (one-tailed) Wilcoxon rank sum test. BRCA breast adenocarcinoma; BLCA bladder adenocarcinoma; COAD colon adenocarcinoma; ESCA esophageal cancer; HNSC head & neck squamous cell carcinoma; KIRC kidney renal cell carcinoma; KIRP kidney renal papillary carcinoma; LIHC liver hepatocellular carcinoma; LSCC lung squamous cell carcinoma; LUAD lung adenocarcinoma; PAAD pancreatic adenocarcinoma; PRAD prostate adenocarcinoma; READ rectal adenocarcinoma; THCA thyroid carcinoma; UCEC uterine corpus endometrial carcinomaCC BY · PubMed Central
Fig. 5 pcgtAge/epiTOC is aggravated in pre-cancerous lesions whereas Horvath’s age-acceleration measure is not. a Left panel: the pcgtAge between normal lung tissue (N), lung carcinoma in situ (LCIS), and LCIS samples from patients who later progressed to an invasive lung cancer (LCIS- > LC). P values between respective groups are from a one-tailed Wilcoxon rank sum test. P value from a linear regression of pcgtAge versus group adjusted for age is also given at the top. Right panel: corresponding receiver operating characteristic (ROC) for discriminating the normal and all LCIS samples. AUC and 95 % CI is given. b Horvath’s age-acceleration measure between normal lung tissue (N), lung carcinoma in situ (LCIS), and LCIS samples from patients who later progressed to an invasive lung cancer (LCIS- > LC). P values between respective groups are from a one-tailed Wilcoxon rank sum test. c, d Exactly as a, b but now in a different Illumina 450 k data set, encompassing normal breast tissue, ductal carcinoma in situ (DCIS), and DCIS samples from women who later developed an invasive breast cancer (BC). e, f Exactly as a, b but now in an Illumina 450 k set which profiled normal breast tissue from healthy women (N), normal breast tissue adjacent to a breast cancer (NADJ), and breast cancers (BC). The number of samples in each group is given below
Fig. 5 pcgtAge/epiTOC is aggravated in pre-cancerous lesions whereas Horvath’s age-acceleration measure is not. a Left panel: the pcgtAge between normal lung tissue (N), lung carcinoma in situ (LCIS), and LCIS samples from patients who later progressed to an invasive lung cancer (LCIS- > LC). P values between respective groups are from a one-tailed Wilcoxon rank sum test. P value from a linear regression of pcgtAge versus group adjusted for age is also given at the top. Right panel: corresponding receiver operating characteristic (ROC) for discriminating the normal and all LCIS samples. AUC and 95 % CI is given. b Horvath’s age-acceleration measure between normal lung tissue (N), lung carcinoma in situ (LCIS), and LCIS samples from patients who later progressed to an invasive lung cancer (LCIS- > LC). P values between respective groups are from a one-tailed Wilcoxon rank sum test. c, d Exactly as a, b but now in a different Illumina 450 k data set, encompassing normal breast tissue, ductal carcinoma in situ (DCIS), and DCIS samples from women who later developed an invasive breast cancer (BC). e, f Exactly as a, b but now in an Illumina 450 k set which profiled normal breast tissue from healthy women (N), normal breast tissue adjacent to a breast cancer (NADJ), and breast cancers (BC). The number of samples in each group is given belowCC BY · PubMed Central
Fig. 6 epiTOC’s tick rate is increased in buccal tissue from smokers. a Density scatterplot of pcgtAge (y-axis) versus smoking pack-years for a total of 647 buccal samples from women all aged 53 years at the time of sampling. Pearson correlation coefficient (PCC) and one-tailed P value from a linear regression are given. b As a but for Horvath’s age-acceleration measure. c Density scatterplot of pcgtAge (y-axis) in whole blood (WB) versus pcgtAge in buccal (BUC) tissue for a total of 152 women all aged 53 years who provided both a WB and a buccal sample. PCC and associated one-tailed P value are given. d Density scatterplots of pcgtAge in WB and BUC tissue against smoking pack years for 122 women with smoking pack year information. P values are from a one-tailed correlation test. e, f As c, d but for Horvath’s age acceleration measure. Note that because all samples are of the same age, the age-acceleration measure is just a constant shift of the predicted DNAm age
Fig. 6 epiTOC’s tick rate is increased in buccal tissue from smokers. a Density scatterplot of pcgtAge (y-axis) versus smoking pack-years for a total of 647 buccal samples from women all aged 53 years at the time of sampling. Pearson correlation coefficient (PCC) and one-tailed P value from a linear regression are given. b As a but for Horvath’s age-acceleration measure. c Density scatterplot of pcgtAge (y-axis) in whole blood (WB) versus pcgtAge in buccal (BUC) tissue for a total of 152 women all aged 53 years who provided both a WB and a buccal sample. PCC and associated one-tailed P value are given. d Density scatterplots of pcgtAge in WB and BUC tissue against smoking pack years for 122 women with smoking pack year information. P values are from a one-tailed correlation test. e, f As c, d but for Horvath’s age acceleration measure. Note that because all samples are of the same age, the age-acceleration measure is just a constant shift of the predicted DNAm ageCC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Cited by

Sorted by citations

See all 328

Related

See all

Metadata

Authors
Yang, Z., Wong, A., Kuh, D., Paul, D. S., Rakyan, V. K., Leslie, R. D., Zheng, S. C., Widschwendter, M., Beck, S., & Teschendorff, A. E.
Affiliations
  • Chinese Academy of Sciences (CN)
  • MRC Unit for Lifelong Health and Ageing (GB)
  • University College London (GB)
  • Queen Mary University of London (GB)
Corresponding author
Teschendorff, Andrew E.
ORCID
Venue
Genome Biology
Publisher
Springer Science and Business Media LLC
Published
2016-10-03
Volume
17
Issue
1
Pages
205-205
ISSN
  • 1465-6906 (ISSN-L)
  • 1474-760X
DOI
10.1186/s13059-016-1064-3
PMC
PMC5046977
OpenAlex
W2529317472
Cited by
325
References
71
Citations by year
2026: 31 · 2025: 52 · 2024: 43 · 2023: 40 · 2022: 53 · 2021: 32 · 2020: 25 · 2019: 17 · 2018: 18 · 2017: 12 · 2016: 1
FWCI
8.27
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • Genome biology (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • UCL Discovery (University College London) (submittedVersion)
  • Queen Mary Research Online (Queen Mary University of London) (submittedVersion, cc-by)
  • Apollo (University of Cambridge) (submittedVersion, cc-by)
  • Apollo (cc-by)
Funders
  • Royal Society (NAF 164914)
  • National Science Foundation of China (NSFC 31571359)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Epigenetics and DNA Methylation
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Cancer Research › Cancer Genomics and Diagnostics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Cell Biology › Telomeres, Telomerase, and Senescence
Keywords
epigenetic clocks, DNA methylation, stem cell division, cancer risk, promoter CpG sites, fetal tissue, carcinogen exposure, precancerous lesions, cancer risk stratification
MeSH terms
Aging, Humans, Mitosis, Mitotic Index, Models, Theoretical, Neoplasms, Promoter Regions, Genetic, Neoplastic Stem Cells, Gene Expression Regulation, Neoplastic, CpG Islands, DNA Methylation, Epigenesis, Genetic, High-Throughput Nucleotide Sequencing, Polycomb-Group Proteins

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review