The study, led by Elise N. Erickson of Oregon Health and Science University together with colleagues at Emory University and Imperial College London, set out to answer two deceptively simple questions. First, do existing epigenetic clocks, mathematical models that estimate biological age from chemical tags on DNA, actually track maternal age in the myometrium at all? Second, and more importantly for future clinical use, can a simple blood draw stand in for a tissue that can only be obtained during surgery? The answer to both, the team reports, is a cautious yes, with one clock in particular standing out from the pack.

Epigenetic clocks rest on a phenomenon called DNA methylation, the attachment of methyl groups to specific positions in the genome that changes in reproducible patterns as cells age. Because methylation patterns shift predictably over the lifespan, researchers can compute a weighted average across hundreds or thousands of CpG sites, the cytosine-guanine sequences where these tags accumulate, to produce an age estimate in years. The team measured methylation across more than 850,000 CpG sites using the Illumina MethylationEPIC BeadChip, then calculated epigenetic age with four well-known clocks: the pan-tissue Horvath clock, the blood-derived Hannum clock, the PhenoAge clock, and GrimAge, which incorporates methylation surrogates of plasma proteins linked to disease and mortality.

The samples came from 27 pregnant individuals undergoing planned cesarean birth at term, of whom 21 had complete, quality-controlled matched specimens of myometrium and blood. Their chronological ages spanned an impressive range, from 20 to 50 years, with a median of 35.5. During each cesarean, surgeons excised a small full-thickness piece of myometrium from the edge of the uterine incision immediately after delivery, which was then flash-frozen in liquid nitrogen. Blood was collected by venipuncture one to two hours before delivery. This paired design, the first of its kind to compare epigenetic age across blood and myometrium in the same individuals, is what gives the findings their unusual value.

The results were strikingly tissue-specific. Three of the four clocks, Horvath, Hannum, and GrimAge, produced epigenetic ages that correlated significantly with chronological age in both tissues, meeting the researchers’ benchmark of a moderate Spearman correlation of at least 0.5. In blood, the correlations were strong, ranging from 0.72 to 0.87. In myometrium, they were somewhat weaker but still significant, between 0.62 and 0.70. The PhenoAge clock, however, failed to track age in uterine tissue at all, showing no significant correlation, and was dropped from subsequent analyses. The pattern suggests that clocks trained on blood or on broad tissue panels do not transfer equally well to the pregnant uterus.

Just as intriguing were the absolute age estimates, which diverged from chronological age in opposite directions depending on the clock. Horvath’s pan-tissue clock judged the myometrium to be dramatically older than the women themselves, estimating every single uterine sample at age 40 or older even though the youngest donor was 20, with a median myometrial estimate of 49.4 years. In blood, the same clock also overestimated age, yielding a median of 39.2 years. By contrast, the Hannum and GrimAge clocks returned estimates younger than chronological age in both tissues. The authors caution that these discrepancies likely reflect how each clock was built and validated, and that pregnancy itself, a state of profound hormonal and immunological change, may reshape methylation patterns in ways the clocks were never designed to capture.

The most clinically consequential finding concerned the relationship between the two tissues. If blood is to serve as a proxy for the uterus, epigenetic age measured in the two must agree. GrimAge delivered the strongest concordance by far, with a Spearman correlation of 0.81 between blood and myometrial epigenetic age, while Horvath and Hannum showed moderate correlations of 0.60 and 0.62. Even more telling, when the researchers calculated age acceleration, the gap between biological and chronological age that signals faster or slower aging, only GrimAge’s acceleration measure correlated significantly across tissues, with a coefficient of 0.49. In other words, if you want to know how quickly a woman’s uterus is aging, GrimAge applied to her blood currently offers the best window available.

Why does this matter? The myometrium is the engine of childbirth, and its function during labor, in the third stage of delivery, and during postpartum involution depends on coordinated smooth muscle contraction. Population studies consistently show that older parturients experience longer labors, need more oxytocin, and face higher rates of labor dystocia, unplanned cesarean, and postpartum hemorrhage. Yet laboratory studies of myometrial contractility by age have produced conflicting results, and many women over 35 deliver without any difficulty while some younger women experience severe dysfunction. Chronological age, in short, is a poor predictor of the biology that matters. A biomarker of uterine biological age could eventually identify, before labor begins, which pregnancies are at risk of myometrial failure.

The study also surfaced a provocative secondary observation: myometrial GrimAge acceleration was moderately correlated with pre-pregnancy body mass index, with a Spearman coefficient of 0.52. The association did not survive correction for multiple testing, and the authors are careful to flag it as potentially spurious, but it aligns with earlier work showing that obesity accelerates epigenetic aging in liver and adipose tissue. Given that higher body mass is independently linked to poor birth outcomes, future studies will need to disentangle the contributions of age and body composition to uterine aging, a task that will require far larger cohorts than the 21 participants analyzed here.

The limitations are real and the authors do not hide them. The sample was small, overwhelmingly of European ancestry, and composed entirely of women who never entered labor, since all samples came from planned cesareans. No in vitro contraction testing was possible with the preserved tissue, and limited clinical data prevented adjustment for confounders. But the consistency of the correlations across three clocks and two tissues, despite these constraints, gives the findings credibility. The study establishes proof of concept that blood-derived epigenetic age can plausibly stand in for the inaccessible uterus, opening the door to studies that follow vaginal births, sample earlier in pregnancy, and test whether epigenetic age acceleration predicts prolonged labor, hemorrhage, or even preeclampsia. It may also, the authors suggest, become a tool for measuring whether interventions such as diet, exercise, or stress reduction genuinely slow biological aging in pregnant women, and for examining how the cumulative wear of social disadvantage, which widens racial disparities in maternal outcomes with advancing age, is written into the genome of the uterus itself.

Subject of Research: Epigenetic clock-based estimation of biological age in maternal blood and uterine myometrium during pregnancy

Article Title: Advancing understanding of maternal age: correlating epigenetic clocks in blood and myometrium

Article References: Erickson, E. N., Knight, A. K., Smith, A. K., & Myatt, L. (2022). Advancing understanding of maternal age: correlating epigenetic clocks in blood and myometrium. Epigenetics Communications, 2(1), Article 3. https://doi.org/10.1186/s43682-022-00010-0

Image Credits: AI Generated

DOI: 10.1186/s43682-022-00010-0

Keywords: epigenetic clocks, DNA methylation, maternal age, myometrium, pregnancy, GrimAge, Horvath clock, biological aging, age acceleration, parturition, obstetrics, biomarkers

The study, published in Reproductive Sciences, builds on a long line of attempts to model human uterine muscle in the laboratory. Since 1984, when researchers first established methods for growing human myometrial smooth muscle cells in monolayer culture, scientists have wrestled with the limitations of isolated cells. Primary cultures eventually reach replicative senescence, dividing only a finite number of times before they stop. Immortalized lines—created first with viral oncogenes from human papillomavirus type 16 and later with the telomerase enzyme hTERT—can divide indefinitely, but they carry the fingerprints of their engineering: altered cell size, changed expression of smooth muscle actin filaments, and, in some cases, genomic instability. Cell lines derived from pregnant myometrium, such as the PHM1-41 line, retained oxytocin responsiveness and key smooth muscle markers, yet none of these two-dimensional systems preserve the multicellular composition, extracellular matrix, and three-dimensional architecture of intact uterine tissue.

Three-dimensional culture technologies have narrowed that gap. Protein-based gels such as Matrigel and collagen let cells self-organize and remodel their surroundings; bioprinted myometrial rings express contraction markers including alpha-smooth muscle actin, connexin-43, and the oxytocin receptor, and can even contract; polymer scaffolds and microfluidic devices model the uterine wall in increasingly sophisticated ways. But these systems are expensive, slow to prepare, and often demand specialized equipment. Whole-tissue explants—small cubes of biopsied myometrium kept alive in serum-free medium—remain the most faithful stand-in for the real thing. A whole-genome comparison by Georgiou and colleagues found that explants cultured for 30 hours differed from fresh myometrium in only 1,444 genes, whereas primary smooth muscle cells differed in 3,840 and telomerase-immortalized cells in 4,603. Explants, in short, stay closest to the tissue they came from.

Yet even explants are not frozen in time. The Newcastle group had previously shown that during 48 hours of culture, expression of the oxytocin receptor gene OXTR falls while ESR1, PTGS2, PR-A, and AKR1C1 all rise—changes strikingly similar to those documented at the onset of human labor. That observation prompted a bold hypothesis: perhaps non-laboring myometrium does not simply drift in culture but actively transitions toward a labor-like phenotype. If true, the culture dish itself could serve as an ex vivo model of labor onset, allowing researchers to watch the molecular choreography of birth unfold in tissue from a single donor, sampled at baseline and again two days later.

To test the idea, the team collected myometrial biopsies from 30 women undergoing elective cesarean sections at term, between 38.3 and 39.5 weeks of gestation, none of whom were in labor. A small piece of each sample was snap-frozen immediately to capture the true non-laboring baseline. The rest was cut into roughly two-millimeter cubes and incubated for 48 hours in serum-free medium, then frozen for analysis. Using quantitative real-time PCR, the researchers measured the abundance of 30 messenger RNAs spanning contraction-associated proteins, inflammatory mediators, steroid and prostaglandin enzymes, calcium regulators, small heat shock proteins, and phosphodiesterase isoforms. Selected proteins were quantified by Western blotting, with each participant’s fresh tissue serving as the paired control for their own cultured tissue—a design that neatly cancels out individual biological variation.

The results were striking. Genes encoding contraction-associated proteins surged: GJA1, which encodes connexin-43, the gap junction protein that electrically couples uterine muscle cells into a synchronized whole, rose significantly, as did PTGS2, the cyclooxygenase enzyme that drives prostaglandin synthesis, increasing an extraordinary 171-fold on average. ESR1, the estrogen receptor alpha that acts as a master upstream switch for myometrial activation, climbed alongside it. Meanwhile, genes encoding relaxation-associated proteins moved in the opposite direction: PLCL1, a brake on intracellular calcium signaling, dropped at both the mRNA and protein levels, and the small heat shock proteins HSPB1 and HSPB6 also declined. The oxytocin receptor gene OXTR fell to roughly a third of its baseline abundance, echoing findings from laboring tissue in vivo, even though OXTR protein showed only a non-significant downward trend.

The inflammatory arm of the transition was equally dramatic. Nuclear factor kappa B subunits NFκB1 and NFκB3, the cytokines TNFα, IL1β, IL6, and IL8, and the calcium channel TRPV4 all rose significantly, with IL1β increasing 31-fold, IL6 nearly 123-fold, and IL8 more than six-thousand-fold. This sterile inflammatory cascade is a hallmark of labor onset, recruiting immune cells and amplifying prostaglandin production. The team also documented widespread upregulation of phosphodiesterase genes, including all four PDE4 variants and PDE5A, PDE7A, PDE7B, and PDE8A—enzymes that degrade the cyclic nucleotides cAMP and cGMP, which normally keep the pregnant uterus relaxed. Their rise suggests a weakening of the biochemical brakes on contraction, though the roles of most PDE isoforms in human labor remain poorly charted.

Weighing the evidence, the authors found that of 17 targets that could be compared against established in vivo labor-associated changes, 14 shifted in the same direction as genuine labor: ESR1, GJA1, OXTR, PTGS2, PR-A, PR-B, AKR1C1, TNFα, IL1β, IL6, IL8, PLCL1, HSPB1, and HSPB6. Only PTGS1, SRD5A1, and PDE8B behaved in ways that may diverge from the in vivo picture, and these genes are considered secondary players in the labor cascade. AKR1C1, which converts active progesterone into its inactive metabolite 20α-hydroxyprogesterone and thereby enacts a functional progesterone withdrawal, rose in both transcript and protein—consistent with the local progesterone withdrawal thought to unlock the contractile machinery at term.

What drives this transformation in the dish? The authors are careful not to overclaim. Removing tissue from the body abruptly withdraws circulating progesterone and other endocrine signals, cuts off vascular perfusion, alters oxygen and nutrient gradients, and eliminates the mechanical stretch of the gravid uterus, while biopsy and dissection themselves inflict injury that can ignite inflammatory stress. The precise mechanisms remain unresolved, and the study measured only two time points, relied on a targeted gene panel, and did not directly assess tissue viability or contractile function. The researchers acknowledge that future work must map the timeline of the transition with shorter and longer incubations, test whether the molecular shift is accompanied by functional changes in contractility, and validate the model with transcriptome-wide comparisons against genuine laboring tissue.

Even with those caveats, the implications are twofold and significant. On one hand, the findings expose an under-recognized pitfall: experiments that culture supposedly resting myometrial explants for two days or more may no longer be studying the quiescent state they intended to probe, since the tissue has already begun marching toward labor. On the other hand, the reproducible acquisition of labor-associated molecular features offers something genuinely new—an experimentally accessible system in which the transition toward labor can be observed within tissue from the same woman, opening the door to identifying early-response genes, testing candidate triggers and inhibitors, and asking whether the shift toward labor can be slowed, prevented, or reversed. What was once dismissed as culture-induced drift may, it turns out, be a window into the moment life’s most fundamental muscular performance begins.

Subject of Research: An ex vivo model of the molecular transition of term human myometrium toward a labor-like phenotype

Article Title: An Ex Vivo Model for the Onset of Human Labor

Article References: Paul, M., Barreda, A. P., Gregson, A., King, M., Hossain, M. R., Hussein, W. M., Walker, F. R., Smith, R., Zakar, T., & Paul, J. W. (2026). An Ex Vivo Model for the Onset of Human Labor. Reproductive Sciences. https://doi.org/10.1007/s43032-026-02214-4

Image Credits: AI Generated

DOI: 10.1007/s43032-026-02214-4

Keywords: myometrium, labor onset, ex vivo culture, contraction-associated proteins, progesterone withdrawal, inflammation, prostaglandins, oxytocin receptor, preterm birth, reproductive physiology, explants, parturition