Rethinking how bone marrow works

One huge reason why the world of medicine hasn’t yet found “the cure” for hard-to-treat malignancies like acute myeloid leukemia (AML) and many other forms of cancer is that the world of science is still striving to fully understand how our bone marrow forms the many types of cells within the stuff we call “blood.”

While many dramatic improvements have been achieved in childhood cancer survival rates across the decades, a collection of blood cancer variations and subtypes still take too many young lives. The hunt to cure more of these elusive conditions is bringing together teams of experts in exotic topics such as high-dimensional flow cytometry, advanced computational medicine, single-cell genetic analysis, transcriptomics, and more.

These experts have made exciting progress along several fronts. Now, a new study led by researchers at Cincinnati Children’s proposes a way to combine previously separate data gathering approaches and concepts about blood cell formation into a more unified whole.

Details about the team’s bioinformatics approach to combine different single-cell technologies and new methods for flow cytometry were published online Oct. 22, 2025, in the journal Nature Immunology.

“This integrative approach marks a significant advance in stem cell biology,” says corresponding author H. Leighton “Lee” Grimes, PhD, director of the Cancer Pathology Program at Cincinnati Children’s. “Beyond hematopoiesis, this computational and experimental framework sets the stage for elucidating developmental hierarchies in a variety of complex tissues, including organs, cancer tumors and other systemwide regenerative processes.”

Co-corresponding authors for this study were Harinder Singh, PhD, formerly with Cincinnati Children’s, now with the University of Pittsburgh, and Nathan Salomonis, PhD, Division of Biomedical Informatics at Cincinnati Children’s. They say the new approach has expanded capacity to identify, isolate, and manipulate populations of cells within larger tissues based on detecting key markers of underlying gene regulatory networks these cells use as they move between different “discrete states” of development. The advance opens new avenues for developing targeted cancer therapies and for new progress in stem cell engineering.

“If you want to make blood cells in a dish or correct congenital defects or malignancy you need to know what ‘normal’ programs look like,” Grimes says. “This is an important first step toward that goal.”

A stepwise view of normal bone marrow formation

The new research uses deep single-cell data sets to resolve underlying transcriptional programming that controls a number of transitory cell states, some of which have not been previously identified.

“What is revealed is a model of hematopoiesis in which cells move between “buckets” of varying stability to reprogram to the next step of differentiation in a stepwise manner. These discrete cell states contrast significantly with other models of hematopoiesis,” Grimes says.

The study generated a new atlas of murine bone marrow progenitor cells that reflects the combined output of three cutting-edge technologies:

• CITE-seq: a tool that combines surface protein detection with transcriptomics.
• TEA-seq: which adds chromatin accessibility data alongside RNA and surface markers.
• InfinityFlow: which uses high-dimensional flow cytometry with imputation algorithms to analyze vast numbers of surface markers across millions of cells.

Key Breakthroughs

The team identified and isolated a set of rare bone marrow cell populations—termed "MultiLin"—that are the last marrow cells to possess multilineage potential. These cells are distinct from previously categorized progenitors and act at pivotal decision points as cells begin restricting their menu of potential final fates.

The researchers also “flipped” a classical method of identifying gene regulatory networks that underpin cell states. By analyzing transcription factor (TF) activity down to base-pair resolution within open chromatin regions, researchers determined that specific TFs modulate entire gene expression programs to drive cells to commit to their ultimate fate.

These MultiLin progenitors can give rise to diverse mature lineages, including erythroid, myeloid, eosinophil, basophil, and mast cells.

“Notably, these progenitors and the lineages they produce respond dynamically to physiological stress, such as parasitic infection, which underscores their biological importance,” Grimes says.

Implications and Future Directions
Contrary to the idea of a smooth continuum of cell development, the findings support a model where discrete progenitor cell states serve as key regulatory nodes. These states are governed by complex, but specific gene regulatory architectures.

A key next step will be translating the work focused on blood formation in mice to humans.

Understanding how these architectures work in concert opens the possibility of manipulating the architectures as whole pieces, potentially resulting in flexible ways to develop targeted therapies and engineer improved forms of stem cells that could help more bodies heal themselves.

About the study

Funding sources for this work include grants from the National Institutes of Health (RC2DK122376, R01HL122661, and T32CA117846) and financial support from Cytek Biosciences for cytometer services provided at the Ohio State University Comprehensive Cancer Center (OSUCCC).