Does CEBPA–ALOX15B Drive Osteoblast Ferroptosis?

Research Question

Does the CEBPA–ALOX15B axis provide pathway-level evidence that osteoblast ferroptosis contributes to postmenopausal osteoporosis, and can the study distinguish ferroptosis from nonspecific cell-death readouts? This question matters because reduced osteoblast activity is central to estrogen-deficiency bone loss, while measurements such as Annexin V/PI positivity or TUNEL labeling cannot, on their own, identify whether cells died by apoptosis, ferroptosis, or another process.

The study by Wang et al. proposes that CEBPA transcriptionally activates ALOX15B, which then alters AMPK/mTOR signaling, increases ferroptosis-associated stress in osteoblasts, and impairs bone formation. The paper supplies several types of evidence for this model: patient and public-dataset associations, genetic perturbations, biochemical ferroptosis measurements, epistasis experiments, pathway pharmacology, primary-cell validation, and an ovariectomized mouse model. The interpretation is strongest where these measurements converge, but it remains more cautious than a claim that every death event measured in the study was ferroptotic.

Study Design and Methods

The authors first integrated four GEO expression datasets relevant to postmenopausal osteoporosis or bone-mineral-density differences: GSE56116, GSE56814, GSE56815, and GSE230665. These datasets used human blood-derived expression profiles. The analysis identified genes repeatedly associated with the disease and intersected them with ferroptosis-related genes. The investigators then tested candidate expression in blood from 28 women with postmenopausal osteoporosis and 13 age-matched postmenopausal controls. ALOX15B expression was compared with serum estradiol, rather than being treated as a direct measurement of bone-cell expression.

Mechanistic cell experiments used the hFOB 1.19 human osteoblast line. The design included ALOX15B knockdown, CEBPA knockdown or overexpression, ALOX15B overexpression, and exposure to Erastin as a ferroptosis-inducing challenge. Readouts included viability; Annexin V/PI labeling; Fe2+, MDA, GSH, GSSG, and ROS; GPX4 protein; PTGS2 mRNA; and osteogenic endpoints such as ALP, Alizarin Red S staining, RUNX2, and OCN. These measurements connect cell survival, redox and lipid-peroxidation stress, and osteoblast differentiation, but they do not all have equal specificity for ferroptosis.

To test the proposed transcriptional hierarchy, the authors used promoter-prediction analysis, CEBPA ChIP-qPCR, and a dual-luciferase promoter assay. The luciferase experiment was performed in HEK293T cells using promoter plasmid DNA and a mutated CEBPA-binding site; it was not an ALOX15B reporter-mRNA experiment. The pathway analysis used immunoblotting for phosphorylated and total AMPK and mTOR, followed by Compound C, an AMPK inhibitor, or MHY1485, an mTOR activator, in ALOX15B-knockdown cells. Notch- and Wnt-associated proteins were also examined as alternative pathway candidates.

The study extended key observations to primary mouse osteoblasts. For in vivo testing, the authors generated global ALOX15B-knockout mice using Cas9 mRNA and sgRNA-mediated editing, then subjected adult female mice to bilateral ovariectomy or sham surgery. The animal design compared wild-type sham, wild-type ovariectomized, knockout sham, and knockout ovariectomized groups. Bone mineral density and trabecular structure were assessed by micro-CT, while bone iron, GPX4, 4-HNE, AMPK/mTOR signaling, osteogenic markers, and TUNEL labeling provided supporting tissue measurements.

Main Findings

ALOX15B was associated with estrogen-deficient osteoporosis

Across the discovery analysis, ALOX15B was one of four ferroptosis-related genes selected for experimental follow-up. In the clinical blood cohort, ALOX15B mRNA was higher in the osteoporosis group than in the normal-bone-mass group. Its expression correlated inversely with serum estradiol, with the reported correlation of r = −0.4039 and p = 0.0088. This supports an association between ALOX15B dysregulation and the postmenopausal state, but it does not establish that circulating ALOX15B expression originates from osteoblasts or causes bone loss.

Genetic ALOX15B reduction countered an Erastin-associated ferroptosis profile

In hFOB 1.19 cells, Erastin compared with sh-NC reduced viability and increased the study’s Annexin V/PI cell-death readout. ALOX15B knockdown compared with sh-NC increased viability and reduced that readout. More importantly for ferroptosis classification, ALOX15B knockdown compared with Erastin treatment lowered Fe2+, MDA, GSSG, and ROS, while increasing GSH, the GSH/GSSG ratio, and GPX4. PTGS2 mRNA changed in the opposite direction: it decreased after ALOX15B knockdown compared with the Erastin condition.

This combination is more informative than a viability result alone. Iron accumulation, oxidative stress, lipid-peroxidation-associated MDA, glutathione depletion, GPX4 loss, and PTGS2 induction form a convergent ferroptosis-associated pattern. The same ALOX15B perturbation also improved ALP staining, mineralized nodule formation, and ALP, RUNX2, and OCN expression compared with Erastin-treated cells. The implication is that the proposed death pathway is functionally linked to impaired osteoblast differentiation rather than being an isolated biochemical signature.

However, the Annexin V/PI result should not be converted into a diagnosis of apoptosis. Annexin V/PI detects changes associated with loss of membrane asymmetry and membrane integrity, not caspase dependence. The paper did not measure caspase activation or cleavage, nor did it use a caspase inhibitor to establish caspase-dependent or caspase-independent apoptosis. Thus, the study provides stronger evidence for an ALOX15B-sensitive ferroptosis-associated state than for a separate apoptotic mechanism.

CEBPA acted upstream of ALOX15B in the proposed hierarchy

CEBPA overexpression compared with vector increased ALOX15B mRNA and protein, whereas CEBPA knockdown compared with sh-NC reduced ALOX15B. ChIP-qPCR detected CEBPA enrichment at the ALOX15B promoter. In the luciferase assay, CEBPA increased activity from the wild-type promoter construct compared with the vector control, but not from the construct with the CEBPA-binding site mutated. Together, these results support direct transcriptional regulation rather than a relationship inferred only from co-expression.

The epistasis experiments added a functional test. CEBPA knockdown compared with the control increased GPX4 and reduced PTGS2, Fe2+, MDA, GSSG, and ROS while improving osteogenic staining and marker expression. ALOX15B overexpression compared with the control produced the opposing pattern. When ALOX15B was overexpressed in CEBPA-knockdown cells, the protective changes were partially reversed: GPX4 decreased, PTGS2 increased, ferroptosis-associated biochemical measures shifted toward the ALOX15B-overexpression state, and osteogenic effects became intermediate. Partial reversal is consistent with ALOX15B acting downstream of CEBPA, although it also indicates that CEBPA has additional targets or that ALOX15B does not account for the entire phenotype.

AMPK/mTOR perturbation supplied pathway-level support

ALOX15B knockdown compared with the control increased the p-AMPK/AMPK ratio and decreased the p-mTOR/mTOR ratio. ALOX15B overexpression produced the reverse pattern, while combined CEBPA knockdown and ALOX15B overexpression gave intermediate values. CEBPA did not directly increase luciferase activity from AMPKα1 or mTOR promoter reporters, supporting an indirect relationship in which CEBPA acts through ALOX15B rather than directly transcribing these pathway components.

Pathway interference strengthened the causal interpretation. Compared with ALOX15B knockdown alone, adding Compound C or MHY1485 reduced viability, increased the Annexin V/PI readout, raised Fe2+, MDA, GSSG, and ROS, lowered GSH and the GSH/GSSG ratio, reduced GPX4, increased PTGS2, and impaired ALP and mineralization results. This pharmacological reversal connects ALOX15B loss to AMPK/mTOR-dependent protection. It is not proof of a uniquely linear pathway because pharmacological agents can have off-target effects, but it is stronger evidence than a pathway expression change alone. Notch- and Wnt-related protein measurements did not show significant group differences, making them weaker explanations in this experimental system.

Mouse and primary-cell results extended the model

In ovariectomized mice, ALOX15B knockout compared with wild-type ovariectomized mice increased BMD, BV/TV, and trabecular thickness and reduced trabecular separation. Histology and osteogenic markers also improved relative to the wild-type ovariectomized group. These values did not return to the knockout-sham levels, so ALOX15B loss attenuated, rather than completely prevented, the ovariectomy phenotype.

Bone tissue from knockout ovariectomized mice compared with wild-type ovariectomized mice contained less iron and 4-HNE and more GPX4. The knockout also increased the p-AMPK/AMPK ratio and reduced CEBPA and the p-mTOR/mTOR ratio relative to wild-type ovariectomized mice. TUNEL-positive cells were reduced by knockout compared with the wild-type ovariectomized group, but TUNEL is a general DNA-fragmentation readout and cannot independently identify ferroptosis. In primary mouse osteoblasts, CEBPA knockdown again reduced ferroptosis-associated measures and enhanced osteogenesis, while ALOX15B overexpression reversed those effects. This cross-model consistency strengthens the proposed axis.

Limitations

The clinical cohort was small and drawn from a single center, and the discovery datasets measured blood rather than bone tissue. The inverse estradiol association therefore supports clinical relevance but does not establish osteoblast-specific regulation in patients. The main human cell model was the immortalized hFOB 1.19 line; primary mouse osteoblast validation improves physiological relevance but does not replace validation in primary human osteoblasts.

The global ALOX15B knockout and ovariectomy model show that genotype influences the disease phenotype, but they do not identify which bone or immune cell populations are required. Species differences also limit direct translation. The study did not test a therapeutic intervention after osteoporosis was established, so the results are target-validation evidence rather than clinical or treatment efficacy evidence.

Most importantly for cell-death interpretation, the study did not report a ferroptosis-selective inhibitor rescue, direct caspase-dependence testing, or a caspase-independent apoptosis analysis. Annexin V/PI and TUNEL should therefore remain secondary, nonspecific death measurements. The ferroptosis conclusion is supported by the combined iron, redox, lipid-peroxidation, GPX4, PTGS2, Erastin, genetic, epistasis, and pathway-interference results, but the authors did not measure every proposed downstream mediator, including SLC7A11 or TFR1. The AMPK and mTOR drugs also support pathway involvement without proving that these are the only relevant intermediates.

Source

Wang T, Li J, Li C, et al. Targeting CEBPA/ALOX15B attenuates postmenopausal osteoporosis by inhibiting ferroptosis through the AMPK/mTOR signaling pathway. Inflammation Research. 2026;75:54. https://doi.org/10.1007/s00011-026-02202-z