Using AI and Stem Cell Models for Targeted Treatments in Myelofibrosis

At the 2025 American Society of Hematology meeting, researchers presented a new approach to treat myelofibrosis by targeting the disease-driving cells more precisely. This early-stage study combines single-cell analysis, artificial intelligence, and lab-grown stem cells to identify small molecules that may treat the disease while protecting healthy blood production .

Why is myelofibrosis difficult to treat?

Myelofibrosis is a type of myeloproliferative neoplasm (MPN), a blood cancer where the bone marrow does not function normally. Over time, scarring in the bone marrow can lead to anemia, an enlarged spleen, fatigue, and low blood counts.

Many patients have a mutation called JAK2 V617F. This mutation causes blood-forming stem cells to grow and signal continuously. Current treatments can help with symptoms but often suppress healthy blood cells as well.

How do current JAK inhibitors work?

A type of medication called JAK inhibitors block the JAK signaling pathway. This reduces symptoms like spleen enlargement and inflammation. Examples of JAK inhibitors to treat myelofibrosis include: 

• Ruxolitinib (Jakafi, Incyte)
• Fedratinib (Inrebic, BMS)
• Pacritinib (Vonjo, Sobi)

However, these drugs affect both mutated and normal cells.Because of this, many patients experience dose-limiting side effects, including anemia and low platelet counts. These treatments usually do not eliminate the disease-causing stem cells.

How did researchers study myelofibrosis at the single-cell level?

The research team analyzed more than 150,000 individual cells from 50 myelofibrosis patients using single-cell RNA sequencing, a method that shows how genes behave in each cell.

This allowed scientists to create a detailed map of how mutated stem and progenitor cells differ from healthy ones. These differences were then used to guide drug discovery using artificial intelligence.

Integrating artificial intelligence into research and cell analysis

Researchers used a deep learning framework, a type of artificial intelligence (AI), to identify gene networks that are active only in cells with the JAK2 mutation.

By focusing on these unique disease signals, the AI system helped predict small molecules that could reverse abnormal cell behavior without directly blocking the JAK2 protein itself.

The team developed induced pluripotent stem cell (iPSC) models carrying the JAK2 mutation and converted them into blood-forming stem cells. These lab-grown cells closely matched what is observed in people with myelofibrosis.

Importantly, these mutant cells showed that (megakaryocytes) platelet-producing cells formed even without normal growth signals. This abnormal process is a hallmark of the disease.

The AI-identified molecules reduced abnormal platelet-related markers, and lowered the number of diseased platelet-producing cells. Unlike current JAK inhibitors, these compounds don’t bind directly to JAK2 and show very different gene-expression effects. They were also more than 10 times less likely to interfere with normal red blood cell production.

Why is this important?

Anemia and low blood counts greatly affect quality of life for people with myelofibrosis. Treatments that control the disease while preserving healthy blood production could reduce fatigue, transfusion needs, and treatment interruptions.

This approach aims to decouple mutant signaling from normal hematopoiesis, potentially lowering the risk of anemia seen with current therapies.

What does this research mean for myelofibrosis patients?

This study is preclinical, meaning the treatments are not yet available outside the research setting. However, it represents a promising step toward therapies that directly target the root of myelofibrosis while protecting healthy cells.

Research like this focuses on developing therapies that precisely treat the disease. Progress takes time, but every study like this brings the myelofibrosis community closer to better options and brighter possibilities ahead.

Keep reading more myelofibrosis and myeloproliferative neoplasms news at the healthtree news site. 

READ MORE NEWS

Explore more clinical trials, and personalize your search with our clinical trial finder, using filters based on your specific needs. 

CLINICAL TRIAL FINDER

Source: Identification of small molecules that selectively target JAK2V617F driven cytokineindependent megakaryopoiesis by leveraging single cell RNA sequencing maps of myelofibrosis patients samples and a deep learning framework