The central actor in this story is the senescence-associated secretory phenotype, or SASP. Senescent cells are not simply dormant; they secrete a cocktail of pro-inflammatory cytokines, chemokines, matrix metalloproteinases and growth factors. In the ageing bone microenvironment, this secretion comes from a surprising range of cell types. Senescent osteocytes, the mechanosensing cells entombed in mineralised matrix, senescent mesenchymal stem cells that would otherwise generate new osteoblasts, senescent osteoblasts themselves, immune cells resident in the marrow, and vascular endothelial cells lining the bone’s rich blood supply all contribute to the SASP. The result is a persistent, locally generated inflammatory input that remoulds skeletal biology from within.

What makes the review’s argument distinctive is its claim about convergence. The SASP is not a separate pathway running in parallel to the classical causes of osteoporosis. Instead, the factors released by senescent cells feed directly into the same signalling axes that estrogen deficiency, glucocorticoid excess and mechanical unloading exploit. These include the RANKL/OPG axis, which governs osteoclast formation and bone resorption; the Wnt/β-catenin pathway, which drives osteoblast differentiation and bone formation; and the NF-κB signalling network, the master regulator of inflammation. By flowing into these well-characterised channels, senescence amplifies resorption while suppressing formation, adding an accelerating force to mechanisms clinicians already recognise.

The molecular detail is striking. SASP cytokines such as interleukin-6, working through its soluble receptor, can potentiate RANKL-induced osteoclast differentiation via NF-κB, ERK and JNK signalling. Chemokines like CCL2 recruit inflammatory cells into the marrow, while CCL3 has been shown to cause bone loss and marrow adiposity in aged mice. SASP components also include Wnt antagonists such as sclerostin, which binds the LRP5/6 co-receptors and blocks the canonical Wnt signal that osteoblasts need to build bone. Sclerostin additionally stimulates osteocyte support of osteoclast activity through a RANKL-dependent pathway, meaning a single SASP constituent can simultaneously dampen formation and promote resorption. Osteocytes, long appreciated as endocrine cells and master orchestrators of remodelling, sit at the centre of this regulatory web, expressing RANKL and controlling both sides of the remodelling cycle.

The regulation of the SASP itself is equally intricate. NF-κB not only mediates the downstream effects of SASP factors but also drives the secretory programme within senescent cells. The inflammasome coordinates paracrine senescence, allowing senescent cells to spread their state to healthy neighbours. mTOR signalling promotes the translation of interleukin-1 alpha, a keystone of the pro-tumorigenic SASP, while p38 MAPK regulates the phenotype independently of the canonical DNA damage response. More recently, cGAS-STING signalling, triggered by leaking DNA in senescent cells, has emerged as an amplifier of the ageing process through remodelling of immune function. Epigenetic changes add another layer of control, and extracellular vesicles, tiny membrane-bound packages, have been identified as key carriers that ferry SASP cargo between cells, spreading senescence paracrinally through the marrow and beyond.

Different skeletal cell types contribute distinct SASP profiles. In osteoprogenitors expressing the transcription factor Osx1, DNA damage and senescence deplete the pool of cells available to form bone with age. Senescent mesenchymal stromal cells shift toward adipogenesis, contributing to the fatty marrow that accumulates with ageing and obesity and further impairing both blood cell and bone regeneration. Osteocyte senescence has been documented in type 2 diabetes, where accelerated senescence of these matrix-embedded cells is associated with skeletal fragility, and radiation has been shown to induce a primary osteocyte senescence phenotype that alters mesenchymal stem cell differentiation through paracrine signalling. Senescent osteocytes also undergo biophysical stiffening, potentially compromising their mechanosensing function. Even osteoclasts display a protective role in this landscape, shielding bone blood vessels from senescence through the angiogenin/plexin-B2 axis, a reminder that the senescence story is not uniformly destructive.

The review carefully distinguishes how senescence participates in age-related, postmenopausal and secondary osteoporosis. In natural ageing, senescent cell burden accumulates gradually and contributes to the slow deterioration of bone mass and microarchitecture, including de novo intracortical remodelling and porosity. In postmenopausal osteoporosis, estrogen deficiency and senescence operate as partly independent inputs; studies in young adult mice and older humans indicate that neither fully explains the other, yet both converge on the same oxidative stress and NF-κB-dependent mechanisms. In secondary forms, the evidence is particularly vivid. Glucocorticoid-induced bone loss can be prevented in mice by targeting cellular senescence through modulation of the DPP4-GLP-1 axis. Radiation-induced osteoporosis is driven specifically by p21-positive, rather than p16-positive, senescent cells, and clearing p21-high cells prevents both bone loss and marrow adiposity. Chemotherapy drives senescence in adipo-lineage cells that then promote bone loss, although one 2025 study found senolytics did not prevent chemotherapy’s adverse skeletal effects, underscoring that the picture remains nuanced.

It is the therapeutic implications that give the review its urgency. When senescent cells were cleared in aged mice, either genetically or pharmacologically, age-related bone loss was prevented, providing proof of concept that the SASP is a distinct and independently addressable contributor to skeletal ageing. Senolytic drugs such as the dasatinib and quercetin combination, discovered by exploiting the survival dependencies of senescent cells, have since shown benefit in multiple preclinical models, including postmenopausal, glucocorticoid-induced, diabetic, hypoxia-associated and vitamin D insufficiency-related bone loss. Intermittent senolytic treatment has also been reported to improve responsiveness to mechanical loading in the skeletons of aged mice, linking senescence clearance to the well-known anabolic effects of exercise. A phase 2 randomised controlled trial has now tested intermittent senolytic therapy on bone metabolism in postmenopausal women, marking a critical step toward clinical translation.

Delivery remains a central challenge, and the field is responding with engineering ingenuity. Bone-targeted nanoparticles and nanomaterials are being developed to concentrate senolytics and other therapeutics in the skeleton, and liposomal senolytic formulations have alleviated senescence-induced bone loss in preclinical work. Apoptotic vesicle-mediated senolytic approaches have been shown to require mechanical loading, suggesting that drug therapy and physical activity may need to be combined. Alternative senomorphic strategies aim not to kill senescent cells but to quiet their secretions. JAK inhibition has been shown to alleviate the SASP and frailty in old age, rapamycin extends lifespan in mice partly through mTOR modulation, and glutaminolysis inhibition achieves senolysis through metabolic vulnerability. Rejuvenating senescent bone marrow stromal cells by enhancing TFEB-mediated autophagy has alleviated age-related bone loss in middle-aged mice, and metformin’s effects on bone metabolism are under active investigation.

The authors are candid about the gaps. Senolytics and senomorphics must be evaluated against established osteoporosis treatments, from bisphosphonates and denosumab to the anabolic agents teriparatide and romosozumab, which themselves reshape the RANKL/OPG and Wnt pathways. Denosumab discontinuation causes rebound bone turnover that requires careful sequencing strategies, and the same scrutiny of long-term skeletal outcomes must be applied to senescence-targeted approaches. Biomarker validation is a priority: proteomic atlases of senescence-associated secretomes are enabling the development of circulating biomarkers, and emerging techniques such as spatial transcriptomics are mapping where senescent cells reside within bone tissue. The review also highlights that p16-positive and p21-positive cells represent distinct senotypes with different functional consequences, meaning precision may matter when targeting them. As the population ages and osteoporosis imposes an ever-heavier burden of fragility fractures, the convergence of senescence biology with classical skeletal physiology offers a genuinely new therapeutic frontier, one where ridding bone of its zombie cells, or silencing their inflammatory chorus, could preserve skeletal strength deep into old age.

Subject of Research: Cellular senescence and the senescence-associated secretory phenotype in skeletal ageing and osteoporosis

Article Title: Cellular senescence and the SASP in skeletal ageing: convergent mechanisms of progressive bone loss in osteoporosis

Article References: Liu, H., & Lei, B. (2026). Cellular senescence and the SASP in skeletal ageing: convergent mechanisms of progressive bone loss in osteoporosis. Biogerontology, 27(5), Article 164. https://doi.org/10.1007/s10522-026-10508-y

Image Credits: AI Generated

DOI: 10.1007/s10522-026-10508-y

Keywords: cellular senescence, SASP, osteoporosis, skeletal ageing, senolytics, senomorphics, osteocytes, inflammaging, RANKL/OPG, Wnt signalling, bone remodeling, senescence biomarkers

The premise behind the work is deceptively simple. When a cell enters senescence, a stable state of growth arrest triggered by DNA damage, telomere shortening, oncogene activation, or other stresses, it does not merely switch genes on and off. Its entire chemical constitution shifts. Lipid droplets accumulate, proteins aggregate, mitochondrial metabolism falters, and the balance of nucleic acids, carbohydrates, and small metabolites changes in ways that define the senescent phenotype as much as any transcriptional signature. Yet these changes have been difficult to observe directly in intact, spatially resolved tissue, because conventional biochemical assays typically require cell lysis and bulk extraction, destroying the very spatial context that makes senescence biologically meaningful.

Raman spectroscopy offers a way around this limitation. The technique exploits the fact that when laser light scatters off a molecule, a small fraction of photons exchanges energy with the molecule’s vibrational modes, producing a spectral fingerprint that encodes the chemical bonds present in the sample. Because these fingerprints are characteristic of lipids, proteins, nucleic acids, and carbohydrates, Raman spectra can be used to infer the molecular composition of living or fixed cells without any labels, stains, or destructive preparation. Advances in coherent anti-Stokes Raman scattering and stimulated Raman scattering have dramatically accelerated acquisition speeds, making it feasible to image large cell populations and tissue sections at subcellular resolution in minutes rather than hours.

What Zhang and colleagues realized is that these chemical fingerprints could serve as a complementary data modality to the spatial transcriptomic maps that have transformed modern biology. Spatial transcriptomics, which measures gene expression while preserving information about each transcript’s location within a tissue, has become a cornerstone technology in studies of development, disease, and aging. But the technology measures RNA, not the biochemical end products of gene activity. By acquiring Raman spectra from the same tissue regions that are profiled transcriptomically, the researchers could ask a question that neither modality can answer alone: how do the genes a senescent cell expresses relate to the actual chemical state it occupies?

The answer, according to the study, is that the relationship is rich but nontrivial. Senescent cells display distinctive Raman signatures dominated by lipid-associated peaks, consistent with the well-documented accumulation of lipid droplets in senescent cells across multiple cell types and species. But the RamanOmics framework goes beyond confirming known markers. By training computational models to associate specific spectral features with transcriptomic states, the authors could identify subpopulations of senescent cells that differ biochemically even when their gene-expression profiles appear similar, and conversely, to detect chemical changes that precede or accompany particular transcriptional programs. This multimodal integration effectively doubles the information available from a single tissue section, capturing both the intent of the cell, as written in its transcripts, and the consequence, as written in its chemistry.

The significance of this approach extends well beyond technical novelty. Senescent cells are central players in aging and age-related disease. They accumulate in tissues over time, secrete inflammatory and matrix-remodeling factors through the senescence-associated secretory phenotype, and contribute to conditions ranging from osteoarthritis and atherosclerosis to neurodegeneration and cancer. Therapies designed to eliminate senescent cells, known as senolytics, are now in clinical trials, and interventions that modulate senescent cell behavior, termed senomorphics, are under intense development. Yet the field has struggled with a persistent problem: how to identify and characterize senescent cells reliably in real tissues, where they are rare, heterogeneous, and scattered among healthy neighbors. No single marker is universal, and transcriptional profiles vary with the senescence trigger and the cell type involved.

A chemical fingerprint offers a potentially powerful addition to the senescence-detection toolkit. Because Raman spectra reflect the integrated biochemical state of a cell rather than the expression of any one gene, they may capture aspects of senescence that marker-based and transcriptomic approaches miss. The authors demonstrate that Raman-based classification can distinguish senescent from non-senescent cells in complex biological samples, and that the spectral signatures carry information about the functional state of the cells, including their secretory behavior. If validated broadly, such label-free chemical phenotyping could allow researchers to survey senescence burden in tissues without relying on a panel of imperfect markers, and to track how senescent cells respond to senolytic or senomorphic interventions at the level of their actual biochemistry.

The study also speaks to a broader trend in aging research: the move toward multimodal, spatially resolved atlases of the aging body. Building on large-scale single-cell efforts that have catalogued cell types across mammalian organs and lifespan, researchers are increasingly combining technologies, transcriptomics, proteomics, metabolomics, and now chemical imaging, to construct layered portraits of aging tissues. Each modality reveals a different facet of the aging process, and the correlations and contradictions among them are often where the most interesting biology lies. RamanOmics fits squarely into this program, adding a modality that is unusually direct: rather than inferring chemistry from gene expression, it measures the chemistry itself.

There are, of course, caveats and challenges ahead. Raman spectra are high-dimensional and influenced by factors beyond senescence, including cell type, culture conditions, tissue preparation, and instrument calibration, so robust computational models and careful validation across tissues and species will be essential before the approach becomes standard practice. The spatial registration between Raman imaging and transcriptomic profiling must be precise, and the interpretation of spectral features in terms of specific molecular species remains an active area of chemometric research. Nonetheless, the conceptual advance is clear: senescence is not only a transcriptional state but a chemical one, and the tools now exist to read both simultaneously in the same tissue.

For a field racing to translate senescence biology into therapies, the ability to see the chemical fingerprint of aging cells in place could prove transformative. It may sharpen the identification of target cells, refine the evaluation of anti-aging interventions, and ultimately deepen the understanding of what it means, at the level of molecules and bonds, for a cell to grow old. As the authors suggest, the hidden biochemical landscape of senescence is now coming into view, one spectrum at a time.

Subject of Research: Multimodal Raman chemical imaging and spatial transcriptomics of cellular senescence

Article Title: The chemical fingerprint of cellular senescence

Article References: The chemical fingerprint of cellular senescence. (n.d.). https://doi.org/10.1038/s43587-026-01230-y

Image Credits: AI Generated

DOI: 10.1038/s43587-026-01230-y

Keywords: cellular senescence, Raman spectroscopy, RamanOmics, spatial transcriptomics, aging, senolytics, biophysical chemistry, metabolomics, lipid droplets, chemical imaging, senescence biomarkers, Nature Aging