Establishing a human bone marrow single cell reference atlas to study ageing and diseases

Area of Science:

• Immunology
• Hematology
• Computational Biology

Background:

• Human bone marrow aging is linked to immune decline, increasing elderly vulnerability to diseases.
• A comprehensive reference atlas is needed to understand age-related immunological shifts and identify abnormal cell states.

Purpose of the Study:

• To construct a healthy human bone marrow atlas using single-cell transcriptomic data.
• To characterize age-associated changes in bone marrow cell populations and gene expression.
• To develop a machine learning model for predicting biological age and identifying abnormal cell states in diseases.

Main Methods:

• Collected and analyzed single-cell transcriptomic data from 145 healthy human bone marrow samples (ages 2-84).
• Constructed a bone marrow atlas with 673,750 cells and 54 annotated cell types.
• Developed cell type-specific aging gene signatures and a machine learning model for biological age prediction.

Main Results:

• Identified significant age-associated changes in lymphoid lineage cells, including naive CD8+ T cell shrinkage and effector/memory CD4+ T cell increase.
• Observed an age-correlated decline in common lymphoid progenitors, consistent with myeloid skew in the elderly.
• Successfully identified abnormal plasma cells and erythroblasts in multiple myeloma, and abnormal cells in acute myeloid leukemia.

Conclusions:

• The developed healthy bone marrow atlas serves as a valuable reference for studying bone marrow aging and related diseases.
• The atlas facilitates the identification and investigation of abnormal cell states by mapping disease samples.
• This resource can be mined for novel discoveries in hematopoiesis and age-related immune function.