Hypothesis

**EPO therapy in aged bone marrow triggers a macrophage‑mediated M-CSF surge that expands myeloid‑derived suppressor cells (MDSCs) and osteoclasts, establishing a self‑reinforcing inflammatory loop that blunts erythropoiesis and drives bone loss.

Rationale

• Aged marrow shows elevated TNFα/IL‑6 from MDSCs that suppress erythroid progenitors via ROS‑caspase‑3 [https://pubmed.ncbi.nlm.nih.gov/29263441/].
• High or single doses of EPO induce bone loss through an M-CSF‑dependent osteoclast expansion [https://pmc.ncbi.nlm.nih.gov/articles/PMC12720310/].
• EPO receptors (EPOR) are expressed not only on erythroid progenitors but also on bone‑marrow macrophages and mesenchymal stromal cells, a fact documented in human BM scRNA‑seq atlases.
• In aging, EPOR expression on myeloid cells is increased relative to young marrow, shifting EPO signaling toward M-CSF production rather than erythropoiesis.

Novel Mechanistic Insight

We propose that EPO binding to EPOR on aged marrow macrophages activates JAK2/STAT5 signaling that upregulates CSF1 (M-CSF) transcription. The resulting M-CSF surge:

1. Expands osteoclast precursors → bone resorption.
2. Polarizes monocytes toward an MDSC phenotype, amplifying TNFα/IL‑6 output.
3. Heightens ROS production in erythroid progenitors, deepening EPO hyporesponsiveness.

Thus, the very hormone intended to stimulate red blood cell production inadvertently fuels the niche defects that cause anemia in older individuals.

Testable Predictions

1. Biomarker rise – In aged mice (20–24 mo) receiving weekly low‑dose EPO (5 U/g), BM interstitial fluid will show a ≥2‑fold increase in M-CSF concentration within 6 h post‑injection compared with saline controls (measured by ELISA).
2. Cellular expansion – Flow cytometry will reveal a significant rise in CD11b⁺Ly6C⁺Ly6G⁻ monocytic MDSCs and CD11b⁺F4/80⁺CSF1R⁺ macrophages after 3 days of EPO treatment.
3. Functional rescue – Co‑administration of a CSF1R inhibitor (e.g., PLX3397) with EPO will:
   • Prevent the EPO‑induced increase in osteoclast numbers (TRAP⁺ staining).
   • Reduce MDSC frequency and TNFα/IL‑6 levels.
   • Improve hemoglobin recovery and erythroid colony‑forming unit‑erythroid (CFU‑E) counts relative to EPO alone.
4. Human correlate – In a retrospective cohort of elderly patients receiving EPO for anemia of chronic disease, baseline serum M-CSF will predict poorer hemoglobin response and greater decline in bone mineral density over 6 mo.

Falsifiability

If EPO administration in aged marrow does not elevate M-CSF, or if CSF1R blockade fails to mitigate MDSC expansion and bone loss despite EPO treatment, the hypothesis would be refuted. Conversely, confirming the predicted molecular and cellular changes would support the proposed feedback loop.

Experimental Approach (outline)

• Model: C57BL/6 mice, young (3 mo) vs aged (20‑24 mo).
• Treatment groups (n=8 per group):
  1. Aged + saline
  2. Aged + low‑dose EPO (5 U/g, twice weekly)
  3. Aged + EPO + CSF1R inhibitor (PLX3397, 30 mg/kg diet)
  4. Young + EPO (control for age‑specific effects)
• Readouts (at 24 h, 72 h, and 2 weeks):
  • BM fluid M-CSF, TNFα, IL‑6 (ELISA)
  • Flow cytometry for MDSCs, macrophages, osteoclast precursors
  • Serum hemoglobin, reticulocyte count
  • Bone histology (TRAP, osteoblast surface)
  • CFU‑E assays from flushed BM

Implications

Should the hypothesis hold, it would redefine EPO therapy in the elderly: rather than simply dosing higher, strategies that temper macrophage‑derived M-CSF (CSF1R antagonists, EPOR biased agonists, or intermittent low‑dose regimens) could uncouple erythropoietic stimulation from niche damage, improving efficacy and safety.