Loss of epigenetic information as a cause of mammalian aging

In the paper

“Reduced levels of H3K122ac extend the lifespan of budding yeast (Sen et al., 2015) and levels of H3K27ac and H3K56ac decrease in many human immune cell types (Cheung et al., 2018; Dang et al., 2009).”

What the sources say

• ?Could not verify

  Sen et al. is cited for the budding-yeast H3K122ac lifespan claim, not the separate human immune-cell claim. The Results section on histone mutations and the Discussion report lifespan effects involving H3K36 methylation, but the passages found by searching for H3K122ac and H3K122 do not address whether reduced H3K122ac extends lifespan. (LLM, conf=0.76)

  cites Sen et al., 2015 sourceExperimental — accuracy not yet measured
• Disputed

  Cheung et al. report the opposite overall age trend: the Results (EpiTOF Reveals Aging-Associated Changes; page unavailable) say the vast majority of chromatin-mark measurements were elevated in older subjects, and the Discussion says most marks increased across a broad array of immune cell subtypes. The Results identify reduced marks in central memory CD8+ T cells as an exception, not a decrease across many immune cell types. The budding-yeast claim is attributed to Sen et al., not Cheung et al. (LLM, conf=0.82)

  Checked by a reasoning model — may be wrong.

  cites Cheung et al., 2018 sourceExperimental — accuracy not yet measured
• Partly supported

  The “Histone modifications change in old cells” section (page unavailable) reports that H3K56ac decreased with age in yeast mother cells. It does not establish a decrease in H3K56ac or H3K27ac across human immune cell types. The H3K122ac lifespan claim is attached to Sen et al., not to Dang et al. (LLM, conf=0.87)

  cites Dang et al., 2009 sourceExperimental — accuracy not yet measured

In the paper

“Old mice develop opaque lenses and lose retinal ganglion cells (RGCs) in the innermost retinal layer (Calkins, 2013; Downs, 2015; Samuel et al., 2011; Wolf et al., 2000) (Figure 7J).”

What the sources say

• ?Not checked — no full text of the sourcecites Calkins, 2013 source
• Partly supported

  Section 3.A (Aging; page unavailable) discusses retinal ganglion cell loss in an experimental mouse model of glaucoma. It does not say that aging itself causes this loss in mice, and the passages found on lens opacity do not support the claim about opaque lenses. (LLM, conf=0.82)

  cites Downs, 2015 sourceExperimental — accuracy not yet measured
• Disputed

  The Discussion mentions lens opacification as an age-related optical factor, but does not establish that the study's old mice developed opaque lenses. More decisively, the Results section, “Neurons are preserved in old retina,” reports little change in retinal neuron numbers and no detectable age-related increase in cell death, contrary to the claim that old mice lose RGCs. (LLM, conf=0.90)

  Checked by a reasoning model — may be wrong.

  cites Samuel et al., 2011 sourceExperimental — accuracy not yet measured

In the paper

“The type of DNA damage that is most consistently linked to aging is the double-stranded DNA break (DSB), occurring at a rate of 10 to 50 per cell per day (Tian et al., 2019; Vilenchik and Knudson, 2003).”

What the sources say

• Partly supported

  The Results section reports that DSB repair, but not nucleotide excision repair, coevolves with longevity, supporting a distinctive link between DSBs and aging. Neither that section nor the searched passages gives a rate of 10–50 breaks per cell per day, so the numerical part is unsupported by our copy. (LLM, conf=0.86)

  cites Tian et al., 2019 sourceExperimental — accuracy not yet measured
• Partly supported

  The Background Occurrence of DSBs section estimates about 50 endogenous double-strand breaks per human cell per cell cycle, supporting the claim that these breaks occur frequently. It does not give a rate of 10–50 per cell per day or identify double-strand breaks as the damage type most consistently linked to aging; the paper discusses their implications for cancer instead. (LLM, conf=0.88)

  cites Vilenchik and Knudson, 2003 sourceExperimental — accuracy not yet measured

In the paper

“Our Relocalization of Chromatin Modifiers or “RCM” Hypothesis and subsequent Information Theory of Aging propose that aging in eukaryotes is due to the loss of transcriptional networks and epigenetic information over time, driven by a conserved mechanism that evolved to co-regulate responses to cellular damage, such as a DSB or a crush injury (Mills et al., 1999; Oberdoerffer et al., 2008; Oberdoe”

What the sources say

• ?Not checked — no full text of the sourcecites Mills et al., 1999 source
• Partly supported

  Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  cites Oberdoerffer et al., 2008 sourceExperimental — accuracy not yet measured
• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Sinclair and LaPlante, 2019

In the paper

“Overexpression of SIR2, histones, or deletion of the histone methyltransferase gene SET2, extends yeast lifespan, indicating that epigenetic changes are not merely a biomarker, but a cause of yeast aging (Dang et al., 2009; Feser et al., 2010; Hu et al., 2014; Kaeberlein et al., 1999; Ryu et al., 2014).”

What the sources say

• Partly supported

  The section ‘Sas2 has opposing effects on H4K16ac and lifespan’ (page not supplied) reports that SIR2 overexpression extends yeast lifespan. The Discussion (page not supplied) links histone acetylation and telomeric chromatin to longevity, supporting the point that epigenetic changes can affect aging. This paper does not establish lifespan extension from histone overexpression or SET2 deletion; it reports lifespan extension from SAS2 deletion instead. (LLM, conf=0.88)

  cites Dang et al., 2009 sourceExperimental — accuracy not yet measured
• Supported

  Feser et al. support the histone-overexpression part of this multi-citation sentence: the Results report that extra H3/H4 expression extends yeast replicative lifespan, and the Discussion calls histone loss causal for aging. The paper also reports lifespan extension from an extra copy of SIR2; the SET2-deletion claim is not the part this histone-focused citation is offered to support. Page numbers are unavailable in our copy. (LLM, conf=0.90)

  cites Feser et al., 2010 source
• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Hu et al., 2014

In the paper

“A major driver in yeast is the DSB (Park et al., 1999), the repair of which requires epigenetic regulators Sir2, Hst1, Rpd3, Gcn5, and Esa1 (Martin et al., 1999; McAinsh et al., 1999; Mills et al., 1999; Tamburini and Tyler, 2005).”

What the sources say

• Partly supported

  The Abstract says spontaneous DNA damage, perhaps DSBs, may be an initial step of aging in wild-type yeast. The Discussion likewise argues that DSBs probably occur during the yeast life span, but describes ERC accumulation—not DSBs themselves—as the proposed driver of gradual aging. Calling DSBs a major driver states this proposed role more definitively than the source does. The regulator list is attached to other citations and is not assessed here. (LLM, conf=0.88)

  cites Park et al., 1999 sourceExperimental — accuracy not yet measured
• ?Not checked — no full text of the sourcecites Martin et al., 1999 source
• ?Not checked — no full text of the sourcecites McAinsh et al., 1999 source

In the paper

“Many types of old cells have a paucity of mutations (De Majo et al., 2021; Kaya et al., 2015), strains of mice or people with higher mutation rates show little to no evidence of premature aging (Narayanan et al., 1997; Robinson et al., 2021), and mammals can be cloned from old somatic cells to produce new individuals with normal lifespans (Burgstaller and Brem, 2017).”

What the sources say

• Partly supported

  De Majo et al. supports the first clause for naturally aged mouse hearts: the Discussion reports negligible genomic instability in physiologically aging myocardium. It does not establish that many types of old cells have few mutations; the Results report increased mutational burden with age in mouse intestine, lung, and liver. The later clauses are attributed to other citations, not De Majo et al. (LLM, conf=0.86)

  cites De Majo et al., 2021 sourceExperimental — accuracy not yet measured
• Supported

  The Results and discussion report a low number of mutations in old yeast cells—about 0.4 per average cell lifespan. Kaya et al. supports the old-yeast-cell example in the first clause; the mouse, human, and cloning claims have separate citations. (LLM, conf=0.91)

  cites Kaya et al., 2015 source
• ?Not checked — no full text of the sourcecites Narayanan et al., 1997 source

In the paper

“Mammalian aging is associated with a decline in central nervous system function (Figure 3A) (Johnson et al., 2018; Ungvari et al., 2017).”

What the sources say

• Supported

  The Introduction states that the central nervous system is affected by aging and describes cognitive impairment in older adults. The Discussion also reports age-associated changes in hippocampal cognitive and behavioral function in mice. These passages support the broad claim, though the study itself focuses on hippocampal function rather than every aspect of central nervous system function. (LLM, conf=0.92)

  cites Johnson et al., 2018 source
• ?Not checked — no full text of the sourcecites Ungvari et al., 2017 source

In the paper

“At 10-months post-treatment, the ICE mice had 6-fold fewer COX-positive myofibers (Figure S5G and S5H), about half the capillary to fiber ratio (Figure 4G and 4H), a loss of silencing at repetitive elements and the transcription of retrotransposons, paralleling normal aging (Das et al., 2019; De Cecco et al., 2019; Oberdoerffer et al., 2008; Wenz et al., 2009) (Figure S5I).”

What the sources say

• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Das et al., 2019
• Supported

  The section ‘Activation of L1 in human and mouse tissues’ reports that L1 mRNA was progressively upregulated with age in several mouse tissues. ‘Mechanisms of L1 activation’ describes reduced repressive marks at L1 elements in senescent cells. These passages support De Cecco et al. as a citation for the retrotransposon and aging parallel; the sentence does not identify it as the source of the ICE-mouse measurements. (LLM, conf=0.86)

  cites De Cecco et al., 2019 source
• Partly supported

  The Results section, “SIRT1-associated genes are deregulated in the aged brain” (page not provided), reports loss of silencing at repetitive DNA after genotoxic stress and increased major satellite repeat transcripts with age. It does not report retrotransposon transcription in ICE mice, the COX-positive myofiber count, or the capillary-to-fiber ratio. Those measurements cannot be attributed to this source. (LLM, conf=0.88)

  cites Oberdoerffer et al., 2008 sourceExperimental — accuracy not yet measured

In the paper

“Hippocampal function, critical for spatial and memory consolidation, declines with age (Gallagher et al., 2010; Miller and O’Callaghan, 2005; Park and Reuter-Lorenz, 2009) and is often measured by fear-conditioning, which measures short-term memory.”

What the sources say

• Supported

  The Introduction (page not provided) describes age-related alterations in hippocampal encoding and diminished pattern separation, supporting the claim of declining hippocampal function with age. The Gallagher et al. citation precedes the separate statement about fear conditioning, so that statement is not attributed to this source. (LLM, conf=0.90)

  cites Gallagher et al., 2010 source
• ?Not checked — no full text of the sourcecites Miller and O’Callaghan, 2005 source
• Supported

  In “Compensation for What?”, the paper states that evidence suggests hippocampal function is deficient with age. The citation is attached to the age-related decline clause; the subsequent fear-conditioning statement is not attributed to this source by that citation. (LLM, conf=0.92)

  cites Park and Reuter-Lorenz, 2009 source

In the paper

“Molecular hallmarks of muscle aging include reduced ATP, mitochondrial DNA, and alterations in subsarcolemmal and intermyofibrillar morphology (Demontis et al., 2013; Leduc-Gaudet et al., 2015).”

What the sources say

• Partly supported

  The section “Decline in mitochondrial function and metabolic homeostasis with age” (page unavailable) reports lower ATP levels in aged muscle, but describes increased mitochondrial DNA mutations rather than reduced mitochondrial DNA. The Fig. 2 caption in “Intrinsic defects leading to loss of muscle function during aging” (page unavailable) describes abnormal mitochondrial size and distribution, but does not specify subsarcolemmal and intermyofibrillar morphology. (LLM, conf=0.88)

  cites Demontis et al., 2013 sourceExperimental — accuracy not yet measured
• Partly supported

  The Discussion reports altered subsarcolemmal and intermyofibrillar mitochondrial morphology in aged mouse muscle. The copy does not establish reduced ATP or mitochondrial DNA: the Introduction mentions impaired energetics, while the Results report no significant change in ATP synthase subunit content, which is not a measure of ATP levels. Searches for mitochondrial DNA and reduced ATP found no supporting passage. (LLM, conf=0.88)

  cites Leduc-Gaudet et al., 2015 sourceExperimental — accuracy not yet measured

In the paper

“As mice age, they undergo characteristic physical and physiological changes, including alopecia, hair greying, kyphosis, decreased body weight, motion in dark phase, and a reduced respiratory exchange ratio (RER) (Ackert-Bicknell et al., 2015; Harkema et al., 2016; Houtkooper et al., 2011; Koks et al., 2016).”

What the sources say

• Partly supported

  The Introduction reports age-related declines in physical function, and Commentary > Critical Parameters identifies sudden weight loss and hunched posture as signs of terminal decline in aging mice. Our copy does not substantiate alopecia, hair greying, kyphosis, a change in dark-phase motion, or reduced respiratory exchange ratio; it also does not present weight loss as a universal consequence of aging. The cited passages have no page numbers in our copy. (LLM, conf=0.84)

  cites Ackert-Bicknell et al., 2015 sourceExperimental — accuracy not yet measured
• ?Not checked — no full text of the sourcecites Harkema et al., 2016 source
• Partly supported

  The Results section (“Metabolic characterization and physical performance”; page not provided) reports lower respiratory exchange ratio and less running during the night phase in old mice. It reports that old mice were significantly heavier, not lighter, and does not report alopecia, hair greying, or kyphosis. (LLM, conf=0.92)

  cites Houtkooper et al., 2011 sourceExperimental — accuracy not yet measured

In the paper

“Astrocytes and microglia, mediators of the innate immune response of the central nervous system, become hyper-activated with age (Baruch et al., 2014; Norden and Godbout, 2013).”

What the sources say

• Partly supported

  The passage discussing Fig. 4F–G reports that blocking IFN-I in aged mice markedly decreased astrogliosis and microgliosis, which the authors describe as age-related neuroinflammation. This supports glial activation in aged mice, but the cited passage does not directly show that astrocyte and microglial activation increases from young to old age. (LLM, conf=0.82)

  cites Baruch et al., 2014 sourceExperimental — accuracy not yet measured
• Partly supported

  The section “Astrocytes also have a more inflammatory profile with age” reports age-related changes resembling an activated astrocytic profile. The Abstract and “Innate immune challenge leads to prolonged and exaggerated neuroinflammation” describe aged microglia as primed, with hyper-activation occurring following immune challenge. The claim omits that condition and overstates what the review says about activation with age alone. (LLM, conf=0.86)

  cites Norden and Godbout, 2013 sourceExperimental — accuracy not yet measured

In the paper

“Exercise endurance, strength, muscle mass, vascularization, and mitochondrial function decrease in skeletal muscle with age (Das et al., 2019; Demontis et al., 2013) (Figure 4A).”

What the sources say

• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Das et al., 2019
• Partly supported

  The Introduction reports age-related losses of muscle mass and strength, and its Fig. 1 caption describes decreased capillarization in aging mammalian muscle. The section ‘Decline in mitochondrial function and metabolic homeostasis with age’ reports mitochondrial defects but notes conflicting findings about respiratory function in muscle fibers. The paper does not clearly state that exercise endurance declines with age, so it does not support the full list as written. (LLM, conf=0.84)

  cites Demontis et al., 2013 sourceExperimental — accuracy not yet measured

In the paper

“Based on ChIP-seq and Assay for Transposase Accessible Chromatin using sequencing (ATAC-seq), H3K27ac was enriched in accessible regions, consistent with the known association of H3K27ac with active promoters and enhancers (Heinz et al., 2015; Klemm et al., 2019).”

What the sources say

• Partly supported

  The Enhancer characteristics section (page unavailable) identifies H3K27ac as a feature of active enhancers and describes its correlation with cell-type-specific enhancer activity at an accessible locus. The passages do not establish its association with active promoters. The ChIP-seq and ATAC-seq enrichment result is presented as the citing authors’ own finding, not a result attributed to this review. (LLM, conf=0.82)

  cites Heinz et al., 2015 sourceExperimental — accuracy not yet measured
• ?Not checked — no full text of the sourcecites Klemm et al., 2019 source

In the paper

“These data imply that the muscle of post-treated ICE mice exdifferentiate towards an immune signature, consistent with data from aged mouse tissues (Benayoun et al., 2019).”

What the sources say

• Supported

  The citation supports the comparison to aged mouse tissues, not the citing authors’ inference about post-treated ICE muscle. In the Results section, Benayoun et al. report that immune-signaling pathways, especially interferon responses, are up-regulated across aged mouse tissues. Page numbers were not supplied in our copy. (LLM, conf=0.91)

  cites Benayoun et al., 2019 source

In the paper

“Studies of these proteins indicate that repeated triggering of the RCM response disrupts youthful gene expression patterns and liberates retrotransposons, such as those silenced by SIRT1/SIRT6 (De Cecco et al., 2019; Oberdoerffer et al., 2008; Simon et al., 2019).”

What the sources say

• Partly supported

  The Abstract (page unavailable) reports that LINE-1 retrotransposons become derepressed during cellular senescence, supporting the retrotransposon portion of the claim. The paper does not establish that repeated triggering of an RCM response causes this or disrupts youthful gene-expression patterns. Its References list names a study of SIRT6-mediated LINE-1 repression, but this paper does not show that SIRT1 or SIRT6 silences the elements it examined. (LLM, conf=0.87)

  cites De Cecco et al., 2019 sourceExperimental — accuracy not yet measured
• Partly supported

  The Results and Discussion report that oxidative stress redistributes SIRT1, alters expression of SIRT1-associated genes, and increases transcription of normally silenced major satellite repeats; they also link some of these changes to aging. The paper treats the effects of sustained RCM triggering as a possibility, not an established result, and the cited passages do not identify the repeats as retrotransposons or establish a role for SIRT6. (LLM, conf=0.88)

  cites Oberdoerffer et al., 2008 sourceExperimental — accuracy not yet measured
• Partly supported

  The Discussion reports elevated LINE1 expression in aged wild-type and SIRT6-deficient mice and suggests that loss of silencing or redistribution of chromatin modifiers such as SIRT1 and SIRT6 may explain its activation. It does not establish that repeated triggering of an RCM response disrupts youthful gene-expression patterns or causes retrotransposon release. The indexed passages provide section labels but no page numbers. (LLM, conf=0.87)

  cites Simon et al., 2019 sourceExperimental — accuracy not yet measured

In the paper

“In mammals, cellular identity is established prenatally and maintained by H3K27ac-enriched super-enhancers (Hnisz et al., 2013; Whyte et al., 2013).”

What the sources say

• Partly supported

  The Results section, “Super-enhancers in many cell types” (page unavailable), says super-enhancers drive genes that control and define cell identity and identifies H3K27ac as a useful mark for finding them. The paper does not say that mammalian cellular identity is established prenatally or demonstrate that H3K27ac-enriched super-enhancers maintain it thereafter. (LLM, conf=0.88)

  cites Hnisz et al., 2013 sourceExperimental — accuracy not yet measured
• Partly supported

  The Results report that super-enhancers have substantially higher H3K27ac levels than typical enhancers, and the Discussion links super-enhancers to genes that establish and maintain cellular identity. The paper does not say that mammalian cellular identity is established prenatally; its statement about establishment and maintenance is conditional. (LLM, conf=0.88)

  cites Whyte et al., 2013 sourceExperimental — accuracy not yet measured

In the paper

“Based on yeast work performed in our lab and others, we proposed the RCM and Information Theories of Aging, which state that a loss of epigenetic information driven in part by cellular repair events, are a cause of aging (Oberdoerffer et al., 2008; Sinclair and Guarente, 1997; Sinclair and LaPlante, 2019; Sinclair et al., 1997).”

What the sources say

• Partly supported

  The Discussion (page unavailable) explicitly says yeast work led the authors to propose an RCM model in which DNA-damage-induced chromatin reorganization partly causes aging. The paper also discusses lasting changes at repair sites in its Perspective, but our copy does not name or propose an “Information Theory of Aging.” (LLM, conf=0.84)

  cites Oberdoerffer et al., 2008 sourceExperimental — accuracy not yet measured
• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Sinclair and Guarente, 1997
• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Sinclair and LaPlante, 2019

In the paper

“While it is incontrovertible that certain types of DNA damage mimic aging, including chemotherapy, radiation, smoking, and progeria”

• Also stated in: § DISCUSSION · plain

  “evidence to the contrary includes the fact that cloned animals or mice engineered to accumulate mutations have normal health and lifespans (Narayanan et al., 1997; Wakayama et al., 2013).”

What the sources say

• Partly supported

  The sections “Defects in DNA repair, and premature ageing” and “Increased DNA damage and defective DNA repair” link DNA damage and impaired repair to ageing and progeroid syndromes, and mention susceptibility to ionizing radiation. The copy does not establish the claim’s specific examples of chemotherapy or smoking. Its “Mouse models of progeroid syndromes” section reports some mutant mice without accelerated senescence or altered lifespan, but does not support the broader assertion that mutation-accumulating or cloned animals have normal health and lifespans. Page numbers are unavailable in the indexed copy. (LLM, conf=0.85)

  cites Carrero et al., 2016 sourceExperimental — accuracy not yet measured
• Partly supported

  The Results and Discussion (pages unavailable) report an association between smoking and accelerated epigenetic age in normal breast tissue, supporting the smoking-related part of the claim. They do not establish that DNA damage causes aging-like changes, or support the claims about chemotherapy, radiation, progeria, cloned animals, or mutation-accumulating mice. The Methods mention chemotherapy and radiation only to exclude patients who had received them before surgery. (LLM, conf=0.86)

  cites Hofstatter et al., 2018 sourceExperimental — accuracy not yet measured
• Partly supported

  The Discussion links HIV-induced DNA damage to epigenetic age acceleration only as a possible explanation, conditional on viral integration causing epigenomic instability. It does not establish that DNA damage mimics aging or address the listed examples of chemotherapy, radiation, smoking, and progeria. The cited copy also does not address the cloned-animal or mutation-accumulating-mouse claim in the supplied quote. (LLM, conf=0.76)

  cites Horvath and Levine, 2015 sourceExperimental — accuracy not yet measured

In the paper

“The cyclic expression of Yamanaka factors Oct4, Sox2, Klf4 and Myc (OSKM) (Takahashi and Yamanaka, 2006) alleviates the symptoms and extends the lifespan of progeroid mice and other aspects of aging (Ocampo et al., 2016; Sarkar et al., 2020).”

What the sources say

• ?Not checked — no full text of the sourcecites Takahashi and Yamanaka, 2006 source
• Supported

  The Discussion (page not provided) states that cyclic OSKM induction ameliorated aging hallmarks and extended lifespan in a mouse model of premature aging. It also reports improved pancreatic and muscle regeneration in aged mice, supporting the reference to other aspects of aging. (LLM, conf=0.93)

  cites Ocampo et al., 2016 source
• Partly supported

  The Abstract reports prior evidence that transient reprogramming improves aging hallmarks and extends lifespan in progeroid mice, and identifies Ocampo et al. as the source of the mouse finding. Sarkar et al.’s own Results concern other aspects of aging in cells treated with transient OSKMLN expression, not cyclic OSKM treatment or lifespan in mice. (LLM, conf=0.87)

  cites Sarkar et al., 2020 sourceExperimental — accuracy not yet measured

In the paper

“Studies of simple organisms, such as budding yeast, worms and flies, indicate that a loss of epigenomic information, may be an underlying cause of aging in those species (Imai and Kitano, 1998; Jiang et al., 2013; Lu et al., 2014; Mills et al., 1999; Oberdoerffer et al., 2008).”

What the sources say

• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Imai and Kitano, 1998
• Supported

  The Introduction says age-related changes in chromatin can release normally silenced genes and have been postulated to harm cellular physiology. The Discussion says declining heterochromatin repression in metazoans could contribute to cellular aging. Those hedged statements support the claim that loss of epigenomic information may underlie aging; this paper supplies evidence from flies, not worms. (LLM, conf=0.88)

  cites Jiang et al., 2013 source
• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Lu et al., 2014

In the paper

“The fact that the DSB repair efficiency of mammals correlates with lifespan, but not other types of repair, such as NER and BER, indicates that DSB are a special case, perhaps because only severe threats to cell survival disrupt the epigenome sufficiently to cause aging (Brown and Stuart, 2007; Tian et al., 2019).”

What the sources say

• ?Not checked — no full text of the sourcecites Brown and Stuart, 2007 source
• Partly supported

  The Results sections report that double-strand break repair efficiency correlates with maximum lifespan in the rodent species studied, whereas nucleotide excision repair (NER) does not. The Discussion suggests that base excision repair (BER) may vary little between species, but the paper does not report a BER–lifespan correlation test. The claim therefore states the BER comparison more definitively, and for mammals more broadly, than the evidence supports. (LLM, conf=0.86)

  cites Tian et al., 2019 sourceExperimental — accuracy not yet measured

In the paper

“During normal aging, levels of H3K9me3 in kidneys decrease and H3K36me2 in muscle increase (Tvardovskiy et al., 2017).”

What the sources say

• Partly supported

  The Discussion table (page unavailable) lists H3K9me3 as decreasing with age in kidney, supporting the first part. The Results section on global H3 methylation (page unavailable) reports an increasing trend for H3K36me2 in heart, but the paper does not report this finding for muscle generally or for skeletal muscle. (LLM, conf=0.88)

  cites Tvardovskiy et al., 2017 sourceExperimental — accuracy not yet measured

In the paper

“The spatial organization of chromatin, including a variety of long-range DNA interactions, plays a key role in controlling gene expression and maintaining cellular identity (Gorkin et al., 2014; Kundu et al., 2017; Narendra et al., 2015; Schoenfelder et al., 2015; Zirkel et al., 2018).”

What the sources say

• Supported

  The Introduction (page not provided) states that genome organization plays an essential role in lineage-specific gene expression programs underlying cellular identity. The “Genome organization and pluripotency” section (page not provided) discusses long-range contacts as part of that organization, supporting the cited paraphrase. (LLM, conf=0.91)

  cites Gorkin et al., 2014 source
• Partly supported

  The Results (page unavailable) link changes in PRC1-associated chromatin domains to Hox gene derepression during differentiation. The Discussion (page unavailable) hypothesizes that these domains might help silence genes and maintain their silenced state; it does not establish that spatial organization plays a key role in maintaining cellular identity. (LLM, conf=0.84)

  cites Kundu et al., 2017 sourceExperimental — accuracy not yet measured
• Supported

  The paper’s main text (page unavailable) reports 4C-seq evidence that interactions within the HoxA cluster reorganize into distinct spatial domains during motor-neuron differentiation. Its abstract states that disrupting CTCF-dependent domain boundaries changes Hox gene activation and that this insulation helps produce discrete positional identities. (LLM, conf=0.90)

  cites Narendra et al., 2015 source

In the paper

“GO biological processes were ranked by HyperFdrQ and only GO terms made up of at least 5 genes were included. ChIP-seq data were also analyzed using ChIP-Enrich (Welch et al., 2014). GO terms with at least 5 genes were ranked by FDR.”

What the sources say

• Supported

  The citation is attached to the sentence identifying ChIP-Enrich as a tool used to analyze ChIP-seq data. The Results (‘ChIP-Enrich method’) describe its gene-set enrichment analysis using GO terms. The adjacent five-gene cutoff is stated as the citing authors’ method, not attributed to Welch et al. (LLM, conf=0.90)

  cites Welch et al., 2014 source

In the paper

“The best practices recommended by the Broad Institute for variant calling were then followed (Van der Auwera et al., 2013).”

What the sources say

• Supported

  The Introduction describes a variant-calling workflow, and Strategic Planning, ‘Calling Variants,’ gives recommendations for choosing a GATK variant caller. The citation identifies the best-practices guidance the citing authors say they followed; the source is that guidance, though it cannot independently verify what the citing authors did. (LLM, conf=0.91)

  cites Van der Auwera et al., 2013 source

In the paper

“We removed peaks with significance (signal relative to noise) of p > 10−5. Peaks that fell into the ENCODE blacklist regions were removed (Landt et al., 2012).”

What the sources say

• ?Could not verify

  The citation is attached to the sentence about removing peaks in ENCODE blacklist regions. The Peak calling section discusses significance thresholds and false positives, but the passages found in our copy do not mention blacklist regions or their removal (searched: “blacklist,” “blacklist regions,” and filtering problematic genomic regions). (LLM, conf=0.76)

  cites Landt et al., 2012 sourceExperimental — accuracy not yet measured

In the paper

“Equal amounts of D. Melanogaster DNA were spiked-into ChIP-seq samples. In addition to aligning to the mouse genome, we aligned reads to the D. Melanogaster dm6 genome (Cunningham et al., 2019). To provide a sense of total ChIP-seq signal strength, the proportion of reads aligning to the dm6 genome were compared to the proportion aligning to the mouse genome (Orlando et al., 2014).”

What the sources say

• ?Could not verify

  Cunningham et al. is cited specifically for the D. melanogaster dm6 genome used in read alignment, not for the spike-in or read-ratio method. The Introduction and “Supporting annotation of all species” discuss Ensembl genome resources, but searches for “dm6” and “Drosophila melanogaster” found no passage identifying that assembly in our copy. (LLM, conf=0.74)

  cites Cunningham et al., 2019 sourceExperimental — accuracy not yet measured
• Partly supported

  The Results describe using the proportion of reads mapping to a fixed Drosophila reference to assess changes in ChIP-seq signal relative to the test genome. But Orlando et al. tested human cells and aligned reads to human hg19 and Drosophila dm3, not mouse and dm6 (Discussion, “Library Construction, Sequencing, and Data Collection”). The cited paper supports the comparison’s rationale, not those specific genomes. (LLM, conf=0.86)

  cites Orlando et al., 2014 sourceExperimental — accuracy not yet measured

In the paper

“The techniques described for processing the ChIP-seq and ATAC-seq reads are based on ENCODE/Roadmap guidelines with a few modifications (Gjoneska et al., 2015; Landt et al., 2012; Roadmap Epigenomics et al., 2015).”

What the sources say

• Partly supported

  The Materials and Methods sections on ChIP-seq read mapping and differential analysis say the procedures were consistent with ENCODE standards and describe an adjustment for limited sequencing depth. The paper does not describe ATAC-seq read processing or say that its read-processing procedures followed Roadmap guidelines. Page numbers are not available in our copy. (LLM, conf=0.88)

  cites Gjoneska et al., 2015 sourceExperimental — accuracy not yet measured
• Supported

  Landt et al. is an identification citation for the ENCODE ChIP-seq guidelines: the Abstract says the consortia developed working standards for ChIP experiments, and the Peak calling section discusses processing mapped reads. In this multi-citation sentence, Landt et al. supports the ENCODE ChIP-seq guidance, not the citing authors’ modifications or the separately cited Roadmap guidance. (LLM, conf=0.87)

  cites Landt et al., 2012 source
• Partly supported

  Online Methods §1.2 describes the Roadmap consortium’s ChIP-seq read-mapping and processing procedures, supporting its use as a reference for ChIP-seq processing. That section covers ChIP-seq and DNase-seq, not ATAC-seq; the paper also cannot establish whether the citing authors followed those procedures with only a few modifications. No page number is available in our copy. (LLM, conf=0.85)

  cites Roadmap Epigenomics et al., 2015 sourceExperimental — accuracy not yet measured

In the paper

“Primer sequences were analyzed using the UCSC genome browser in silico PCR tool (genome.ucsc.edu/cgi-bin/hgPcr) to determine the number of genomic elements that contribute to the amplification products (De Cecco et al., 2013).”

What the sources say

• ?Could not verify

  The Experimental procedures section on PCR primer design refers readers to supplemental methods, while the Results say that some primers were empirically verified to amplify individual elements with single-copy kinetics. Neither passage says primer sequences were analyzed with the UCSC in silico PCR tool to count the genomic elements contributing to amplification products; searches for that tool and purpose found no such passage in our copy. (LLM, conf=0.72)

  cites De Cecco et al., 2013 sourceExperimental — accuracy not yet measured

In the paper

“Indeed, rDNA instability and nucleolar size influence the lifespan of lower organisms (Sinclair et al., 1997; Tiku et al., 2017).”

What the sources say

• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Sinclair et al., 1997
• Supported

  Tiku et al. support the nucleolar-size part of the sentence: the Results report that reducing FIB-1 shrank nucleoli and extended lifespan in C. elegans, and the Discussion describes nucleolar regulation as affecting lifespan. The sentence separately cites Sinclair et al. for the rDNA-instability part. (LLM, conf=0.90)

  cites Tiku et al., 2017 source

In the paper

“Using a system called “ICE” (Inducible Changes to the Epigenome), we find the act of faithful DNA repair advances aging at physiological, cognitive, and molecular levels, including erosion of the epigenetic landscape, cellular exdifferentiation, senescence, and advancement of the DNA methylation clock”

In the paper

“At 16 months of age, ICE mice had significantly less muscle mass (Figure 4B), reduced endurance (Figure 4C), greater lactate levels post-exercise (Figure S5A), reduced grip strength (Figure S5B) and other molecular hallmarks of muscle aging (Figure 4D–4F and S5C–S5F).”

In the paper

“In post-treated ICE cells, regions with higher accessibility lost H3K27ac and those with lower accessibility gained it (Figure S6H).”

In the paper

“Based on the two mouse clocks, the rate of epigenetic aging was ~50% faster in the ICE mice than Cre controls (p < 0.0001), closely paralleling treated ICE fibroblasts (Figure 4K–M).”

In the paper

“ICE mice moved ~50% less in the dark phase (Figure 3B) with an impaired gait (Figure S4N–S4P).”

In the paper

“Erosion was also seen across H3K27ac peaks genome-wide and in the top or bottom 40% quantile of H3K27ac signals.”

In the paper

“In skeletal muscle, genes that were significantly dysregulated in ICE mice correlated with changes in old mice (Figure 4I, 4J, S5M–S5O and Table S2), including Cdkn1a (Cyclin Dependent Kinase Inhibitor 1A or p21), a mediator of p53-mediated cellular senescence”

What the sources say

• ?Not checked — no full text of the sourcecites Beggs et al., 2004 source
• ?Not checked — no full text of the sourcecites Choudhury et al., 2007 source
• ?Not checked — no full text of the sourcecites Welle et al., 2004 source

In the paper

“During and after I-PpoI induction, there was no detectable change in cell cycle profile, apoptosis, or senescence (Figure S1E–S1G).”

In the paper

“At 10-months post-treatment, the ICE mice exhibited classic features of old age, including reduced body weight and fat mass, independent of food intake (Figure 2B–2E and S4A–S4E), a lower respiratory exchange ratio (RER) during the day (Figure 2F and S4F) and decreased motion in the dark phase (Figure 2G).”

In the paper

“At the rDNA in particular, there was no evidence of residual cutting (Figure S2J), changes in copy number (Figure S2O), 28S expression levels (Figure S2P) or protein synthesis (S2M and S2N).”

In the paper

“analysis of genes with significant increases in H3K27ac and H3K56ac and decreases in H3K27me3 by Gene Ontology (GO) analysis indicated that cell identity and developmental processes were disrupted in ICE cells. Of the top 20 processes, half of them were involved in development (Figure 5C, Table S5 and S6).”

In the paper

“The ICE:Cre ratio of H3K27ac signals was inversely correlated with basal H3K27ac signals (Figure 5B and S6J), demonstrating that the ICE treatment corrupted epigenetic information and increased Shannon entropy (Hannum et al., 2013).”

What the sources say

• Not checked yet

  Checked before — its earlier verdict is held back until a new check runs.

  cites Hannum et al., 2013 source

In the paper

“There was also no difference in mutation frequency at canonical or non-canonical I-PpoI recognition sites (Wittmayer et al., 1998) or across the genome (Figure S3G–S3I).”

What the sources say

• ?Not checked — no full text of the sourcecites Wittmayer et al., 1998 source

In the paper

“In the Barnes maze test, a measure of long-term memory, the recall of ICE mice was about half that of Cre controls, similar to that of 24-month-old WT mice (Figure 3F and 3G).”

In the paper

“In skeletal muscle, liver and kidney, the Tmem56 intron and the 28S I-PpoI site were cut during TAM treatment and, as expected, no cutting was detectable 1- and 10-month later (Figure 1N and S3F).”

In the paper

“Validation sets had R2 = 0.915 and 0.944 for muscle and blood, respectively, demonstrating accurate age estimations using independent datasets.”

In the paper

“We could not detect changes in mutation frequency at the 28S rDNA (Figure S2A–S2C and 1K), RNA levels (Figure S2D and S2F), or overall translation efficiency (Figure S2K and S2L).”

In the paper

“The ICE mice had thinner left ventricular (LV) posterior walls, implying possible dilated cardiomyopathy (Figure S5K and S5L).”

In the paper

“At later times post-treatment (120–168 hours), ICE cells had lower Lamin B1 levels (Figure S1J and S1K) and increased indicators of cellular senescence (Figure 1F–1H).”

In the paper

“After the addition of TAM, HA-I-PpoI was detected in the nucleus of ICE cells, but not controls (Figure 1B), and the number of serine 139-phosphorylated H2AX (γH2AX) foci, a marker of DSBs, reached a maximum of 4-fold above background after 24-hours, with locus-dependent cutting (Figure 1C and 1D).”

In the paper

“Similar to the hippocampi of aged mice, ICE mice had greater numbers of activated astrocytes (1.6x) and microglia (3.5x) (Figure 3H–3K).”

In the paper

“Loci with higher H3K27ac peak intensities tended to lose the most signal and vice versa, consistent with a genome-wide erosion of the H3K27ac landscape, with no apparent changes specifically at I-PpoI recognition sites (Figure S6I).”

In the paper

“Key features of skin aging were apparent, including subepidermal thinning and hair greying due to a loss of KIT/CD117-positive melanocyte stem cells (Matsumura et al., 2016; Nishimura et al., 2005) (Figure S4J–S4M).”

What the sources say

• ?Not checked — no full text of the sourcecites Matsumura et al., 2016 source
• ?Not checked — no full text of the sourcecites Nishimura et al., 2005 source

In the paper

“On the second day, ~75% of the young mice and ~40% of old mice froze, indicating reduced contextual recall, with a similar difference between Cre and ICE mice (Figure 3C–3E).”

In the paper

“At 1-month post-treatment there was no significant difference in FI between ICE mice and controls, but at 10- and 12-months the ICE mice had a higher FI resembling that of 24-month-old WT mice (p=0.0006 and < 0.0001, respectively) (Figure 2H), along with other features of aging, including kyphosis, loss of cortical bone thickness and trabecular bone density in the inner layer”

In the paper

“In the ICE mice, lens opacity was greater (Figure S7N–S7P) and the number of RGC axons in the myelinated region were fewer than controls (Figure 7K–7N).”

In the paper

“Regions with lower H3K27ac (p < 0.01) in Cre vs. ICE showed the strongest enrichment for muscle tissue signatures (p=9.0×10−8) while regions with higher H3K27ac showed an enrichment for immune cell enhancers (p=9.3×10−28) (Figure 6H and S6M).”

In the paper

“From Hoxa1 to Hoxa6, levels of H3K27ac and H3K56ac decreased and from Hoxa9 to Hoxa13 they increased, with concomitant changes in H3K27me3 and mRNA levels (Figure 5F and S6K).”

In the paper

“After ICE treatment, spleen super-enhancers in muscle were derepressed, gaining H3K27ac (Figure 6J), as were the Class II major histocompatibility complex (MHC) cluster and the Nfkbid gene, a regulator of humoral immunity (Figure 6K).”

In the paper

“Consistent with the in vitro aging forward and reversal data, 7/10 of the top upregulated processes in the RGCs of 12-month-old mice were involved in development (Figure 7P), the majority of which (86% of nervous system developmental genes) were restored by OSK (Figure 7Q).”

In the paper

“Of the top 20 processes that were increased in ICE muscle, all were elevated in old WT mice, consistent with WT and ICE mice undergoing similar epigenetic changes (Figure 6I).”

In the paper

“To test this, we reprogrammed the post-treated Cre and ICE cells into neurons by chemical means (Li et al., 2015) and found that Neurod1 and Nefh were 8–15-fold more easily derepressed in the ICE MEFs (Figure 6D and 6E), coincident with increases in neuron count and the neuronal cell marker TUJ1 (Figure 6F and 6G).”

What the sources say

• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Li et al., 2015

In the paper

“Expression of OSK in treated ICE cells reversed age-associated mRNA changes, including those for Hmgb, Chaf1, Hoxa and canonical histone genes (Figure 7A–7C and S7E), similar to the effect of OSK expression in aged fibroblasts (Figure 7D–7F and S7F), and did not go so far back in age as to begin to erase cell identity or induce the pluripotency gene Nanog (Figure S7A–S7E).”

In the paper

“H3K27ac-positive posterior Hoxa gene promoters (Hoxa 9–13) formed new contacts with active enhancers in an adjacent TAD, with concomitant increases in mRNA levels consistent with weakened insulation and disordered promoter-enhancer (P-E) communication.”

In the paper

“Of the genes with decreased H3K27me3, 4/10 of the top GO processes and 6/10 tissue-specific transcriptional profiles were related to neuronal processes or neuronal tissue types (Figure 6A and 6B).”

In the paper

“More than 3.6 M CpG sites were covered in each sample and 2.7M were covered in all samples.”

In the paper

“Increase of the threshold for the CpG sites coverage decreased the number of the clock sites included in the analysis.”

In the paper

“That said, we saw no effects on rDNA mutations, rRNA levels, or protein translation, and I-SceI, which does not cut the rDNA, generated a I-PpoI-like gene expression pattern.”

In the paper

“These data are consistent with the Information Theory of Aging, which states that a loss of epigenetic information is a reversible cause of aging.”

In the paper

“In yeast, epigenetic information is lost over time due to the relocalization of chromatin modifying proteins to DNA breaks, causing cells to lose their identity, a hallmark of yeast aging.”

• Also stated in: § INTRODUCTION · boosted

  “We and others showed that relocalization of the silent information regulator complex (Sir2/3/4) away from silent mating-type loci to the unstable rDNA is the cause of sterility, a hallmark of yeast aging”

What the sources say

• ?Not checked — no full text of the sourcecites Smeal et al., 1996 source

In the paper

“Our data are consistent with aging in mammals being the equivalent of a software problem, the result of corrupted epigenetic information that can be restored from an existing back-up copy.”

In the paper

“Thus, the ICE system induces specific DNA breaks with no evidence of mutations or immediate deleterious effects.”

In the paper

“I-PpoI recognizes the DNA sequence CTCTCTTAA▼GGTAGC (Monnat et al., 1999), which occurs at 20 targets in the mouse genome, 19 of which are non-coding, including the rDNA, and none of which occur in mitochondrial DNA (Berkovich et al., 2007).”

What the sources say

• ?Not checked — no full text of the sourcecites Monnat et al., 1999 source
• ?Not checked — no full text of the sourcecites Berkovich et al., 2007 source

In the paper

“By introducing DSBs without causing mutations, we provide evidence that it is the cell’s reaction to damage and the resulting loss of epigenetic information that drives mammalian aging forward.”

In the paper

“In a parallel study to this one, we found it possible to safely reverse DNAme age and gene expression patterns of old and damaged neurons to cure blindness in mice, a process requiring DNA demethylation (Lu et al., 2020).”

What the sources say

• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Lu et al., 2020

In the paper

“Together with the neuronal priming of ICE cells and increased EMT in the ICE kidney, we conclude that the induction of non-mutagenic DSBs accelerates the DNAme clock and age-related changes to chromatin, gene expression, and cellular identity.”

In the paper

“In this paper, we show that non-mutagenic DSB repair causes the epigenetic landscape to erode and cells to drift into adjacent Waddington valleys, where they malfunction and cause aging, consistent with the Information Theory of Aging.”

In the paper

“Indeed, Wnt and Hox genes direct embryonic development and also facilitate DSB repair (Feltes, 2019; Liu et al., 2007; Rubin et al., 2007; Ye et al., 2007; Zhao et al., 2018).”

What the sources say

• ?Not checked — no full text of the sourcecites Feltes, 2019 source
• ?Not checked — no full text of the sourcecites Liu et al., 2007 source
• ?Not checked — no full text of the sourcecites Rubin et al., 2007 source

In the paper

“Over time, however, repeated activation causes chromatin factors to scatter and long-range DNA interactions to change over time, thereby aging cells and tissues.”

In the paper

“The effectiveness of the digestion was assessed using controls that omitted reverse transcriptase (RT). Digestion with DNase was repeated until the control lacking RT was negative for γ-satellite sequences.”

In the paper

“We did not perform epigenomic analyses at the single cell level and induced the ICE system in the whole body but not in a tissue-specific manner, so cannot make conclusions about cell autonomy.”

In the paper

“Sex chromosomes were excluded from analyses due to the inconsistency of the sexes of the MEFs.”

In the paper

“Mitochondrial network, size and number were quantified blindly using FUJI ImageJ.”

In the paper

“To reduce non-specific binging to beads, the diluted chromatin was mixed with Dynabeads protein A/G for 1 h at 4 °C and the beads were removed prior to incubating chromatin with 2 ug of the appropriate antibodies with Dynabeads protein A/G.”

In the paper

“Data were normalized using ComBat from the SVA package in R.”

In the paper

“Only CpG sites covered in all samples were considered for DNA methylation clocks application. This resulted in 89 out of 90 sites being covered for the blood DNA methylation clock.”

In the paper

“Matrix pearson’s correlations were calculated in R v 3.6.1 (Bunn, 2008, 2010) and used to perform a principal component analysis. The sign of the first eigenvector for each chromosome was adjusted to correlate with GC content, and were smoothed with loess smoothing using 1 megabase windows.”

What the sources say

• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Bunn, 2008, 2010

In the paper

“Tissue from 6 individual animals was analyzed in triplicate.”

In the paper

“The sorted binary alignments underwent post-processing to remove duplicates via Picard’s MarkDuplicates (v.2.01; http://broadinstitute.github.io/picard) before germline variants were identified using Genome Analysis Toolkit (GATK; v. 3.7) HaplotypeCaller (McKenna et al., 2010).”

In the paper

“For MEF experiments, we used a stringent threshold of adjusted p < 0.01.”

In the paper

“Low quality reads and reads (q < 20) that did not map to the genome were removed.”

In the paper

“In this study, we did not determine which chromatin factors are relocalized, nor did we study chromatin contacts in vivo.”

In the paper

“All primers were tested with serial dilutions of cDNA to ensure they amplified their target sequences quantitatively.”

In the paper

“Scorers were blinded to experimental groups.”

In the paper

“Pairwise interaction frequencies were binned in 50-kb nonoverlapping windows and intra-chromosome interaction frequencies were normalized by dividing each interaction by the average number of interactions observed with that distance.”

In the paper

“To compensate for the low complexity of RRBS libraries 10–20% of phiX was spiked in.”

In the paper

“1/4 number of Drosophila S2R+ cells relative to mouse cells were added as a spiked-in control and combined cells were treated as a single sample during the rest of the procedures.”

In the paper

“Fortunately, it is now apparent that mammals retain a backup copy of youthful epigenetic information that can safely restore the function of old tissues, akin to reinstalling software.”

In the paper

“RNA integrity was determined using an Agilent Bioanalyzer 2100 and an RNA-nano chip.”

In the paper

“For muscle ChIP-Seq experiments, very few peaks attained significance levels at that cutoff. Therefore, we restricted our analysis to looking at the group of peaks enriched at p < 0.01.”