Chemotherapy Destroys Bones by Triggering Senescence in Marrow Fat Cells

Chemotherapy-induced bone loss affects virtually all cancer patients, increasing fracture risk and reducing survival, yet its cellular mechanisms have remained poorly understood. This study, published in Nature Communications, identifies the specific bone marrow cells responsible and demonstrates that targeting them can prevent skeletal damage without undermining cancer treatment.

Using mouse models treated with doxorubicin or paclitaxel, researchers showed that even a single chemotherapy dose triggers rapid trabecular bone loss within 9 days, reducing bone volume/total volume (BV/TV), trabecular number and thickness, while increasing trabecular spacing. Critically, cortical bone was spared. Both female and male mice were affected, and the findings extended to paclitaxel, establishing generalizability across chemotherapy agents.

The team discovered that senescence — marked by p16 expression, SA-β-galactosidase activity, and a robust SASP — was not distributed broadly across bone cell types as might be expected from systemic drug exposure. Instead, it was highly restricted to two adipo-lineage populations: CAR cells (Cxcl12-abundant reticular cells, key mesenchymal stromal regulators) and bone marrow adipocytes (BMAds). Senescent CAR cells and BMAds (senCARs and senBMAds) secreted elevated RANKL, which drove osteoclastogenesis and increased osteoclast numbers and bone surface coverage. Simultaneously, osteoblast numbers and bone formation rates were suppressed, effectively decoupling the normally tightly coordinated processes of bone resorption and formation.

Using the INK-ATTAC transgenic mouse — which allows p16+ senescent cells to be selectively eliminated by AP20187 drug treatment — the researchers confirmed that removing senescent cells restored bone density, normalized osteoclast and osteoblast numbers, and rescued bone formation rates. A vossicle transplant model, where vertebral bodies from INK-ATTAC mice were implanted into wild-type recipients, confirmed that it is specifically the bone-resident senescent cells (not systemic senescence elsewhere) driving the skeletal damage. Pharmacologically, the senolytic combination dasatinib and quercetin (D+Q) replicated these protective effects, selectively eliminating senCARs and senBMAds and preventing bone loss. Importantly, neither genetic nor pharmacologic senolytic approaches interfered with the anti-tumor efficacy of chemotherapy in tested models.

Inhibiting the p38MAPK-MK2 pathway, which post-transcriptionally regulates SASP factor production, also suppressed RANKL secretion and prevented bone loss, providing an orthogonal therapeutic strategy. Together, the data establish a mechanistic framework: chemotherapy → adipo-lineage cell senescence → RANKL-driven osteoclastogenesis + impaired osteoblastogenesis → bone loss. Caveats include reliance on mouse models and the need for clinical validation, particularly regarding whether dasatinib and quercetin dosing regimens safe in cancer patients produce similar bone protection.