Clonal memory of cell division in humans diverges between healthy haematopoiesis and acute myeloid leukaemia

This study reveals that human hematopoietic stem and progenitor cells inherit a "clonal memory" of division synchronicity and fate commitment across multiple generations, a regulatory mechanism that is disrupted in acute myeloid leukemia but can be partially restored through epigenetic modulation.

The Gist

The Big Picture: The "Family Memory" of Blood Cells

Imagine your body's blood system as a massive, bustling city. The Stem Cells are the master architects and builders living in the city's foundation. Their job is to constantly build new houses (blood cells) to replace old ones.

For a long time, scientists thought these builders worked like independent contractors: every time a builder made a decision to work or rest, it was a random, isolated event. But this new study reveals something fascinating: Blood cells have "family memories."

Just like how you might inherit your grandmother's habit of drinking tea at 4 PM or your father's tendency to wake up early, these cells inherit specific behaviors from their ancestors. This paper calls this "Clonal Memory."

The Two Types of Family Memories

The researchers discovered that blood cells actually carry two distinct types of memories down the family tree:

1. The "Work Schedule" Memory (Division):

   • The Analogy: Imagine a family of bakers. If the grandfather baked bread at 6:00 AM, and the father baked at 6:00 AM, the children are likely to bake at 6:00 AM too. They don't just bake bread; they bake it at the same time as their siblings.
   • The Science: In healthy blood cells, if a mother cell divides, her two "daughter" cells will usually divide again at almost the exact same time. Even "cousin" cells (grandchildren of the same great-grandmother) tend to keep this synchronized rhythm. They move through their life cycles in lockstep.

2. The "Career Choice" Memory (Fate):

   • The Analogy: Imagine a family where everyone tends to become doctors, or everyone becomes artists. Even if they are different people, they share a "family vibe" regarding what they want to become.
   • The Science: Cells from the same family tend to turn into the same type of blood cell (e.g., all becoming red blood cells or all becoming immune cells) rather than randomly picking different jobs.

The Healthy vs. The Sick: A Broken Clock

The researchers compared healthy blood cells to those from patients with Acute Myeloid Leukemia (AML), a type of blood cancer.

• Healthy Cells: They are like a well-rehearsed orchestra. Everyone plays the same note at the same time. The "family memory" is strong; the cells know exactly when to divide and what to become, keeping the system stable and predictable.
• Leukemia Cells: These are like a chaotic jazz jam session where everyone is playing different notes at different times. The "family memory" is broken. The cancer cells lose their synchronization. They divide at random times and don't stick to a family plan.
  • Why does this matter? This chaos makes the cancer cells very adaptable. Because they aren't stuck in a rigid schedule, they can survive changes in the environment (like chemotherapy) better than the orderly healthy cells.

The "Magic Reset" Button

The most exciting part of the study is that this broken memory isn't necessarily permanent.

The researchers used a drug (a bromodomain inhibitor called JQ-1) that acts like a "reset button" for the cell's internal software (epigenetics).

• When they gave this drug to the chaotic cancer cells, something amazing happened: The cells started remembering their family schedule again.
• They didn't stop dividing, but they started dividing together again, like a synchronized team.
• The Takeaway: This proves that the "memory" is controlled by how genes are switched on and off (epigenetics), not just by hard-coded DNA mutations. This means we might be able to use drugs to "tune" cancer cells, making them less chaotic and perhaps easier to treat.

Summary in One Sentence

This study shows that healthy blood cells act like a synchronized family with a shared rhythm and career path, while leukemia cells have lost this family memory and become chaotic; however, we can use drugs to help the cancer cells remember their rhythm again, opening new doors for treatment.

Technical Summary

1. Problem Statement

Stem cell heterogeneity is a critical factor in tissue homeostasis and disease progression. While "clonal memory" (the inheritance of cellular properties over at least two divisions) is known to drive differentiation bias in mice, its role in human hematopoietic stem and progenitor cells (HSPCs) remains poorly understood, particularly regarding cell division timing.

• Key Gaps: It is unclear if human HSPCs inherit a memory of division synchronicity, whether division memory and fate commitment memory are distinct or coupled, and how these mechanisms are altered in malignancies like Acute Myeloid Leukemia (AML). Furthermore, the plasticity of clonal memory and its potential for therapeutic modulation is unknown.

2. Methodology

The authors developed a multi-modal, high-resolution experimental and computational framework to track individual human HSPCs ex vivo.

• Cell Sources: Primary HSPCs from umbilical cord blood (foetal) and adult bone marrow (healthy donors), as well as primary CD34+ blasts from AML patients.
• Live-Cell Imaging: Single cells were sorted into 96-well plates and cultured for 96 hours in two distinct media (Differentiation-promoting "Diff" and self-renewal-promoting "GT"). Division times for the first three generations were manually annotated using time-lapse microscopy to assess synchronicity between sister and cousin cells.
• MultiGen Assay: A lineage-tracing technique using four combinations of CFSE and CTV dyes. Four progenitors per well were labeled with unique dye combinations. After 72–96 hours, flow cytometry was used to simultaneously quantify:
  • Number of divisions per cell.
  • Phenotypic fate (using surface markers for 6 HSPC subsets).
  • Lineage relationships (family structure).
• Single-Cell Transcriptomics (scRNA-seq):
  • SMART-seq3: Individual cells from specific families (identified via live imaging) were re-sorted and sequenced to link transcriptomes with division history.
  • CITE-seq: Used for broader population profiling of surface markers and transcriptomes.
• Computational Modeling:
  • Stochastic Modeling & ABC: An agent-based model was built to simulate HSPC proliferation and differentiation. Approximate Bayesian Computation (ABC) was used to compare different scenarios (e.g., independent vs. correlated division/fate) against experimental data to determine the most likely biological mechanisms.
  • Network Analysis: The MIIC (Multivariate Information-based Inductive Causation) algorithm was used to reconstruct causal networks linking gene expression, family identity, and division properties.
• Pharmacological Modulation: AML cells were treated with JQ-1, a pan-bromodomain (BET) inhibitor, to test if epigenetic remodeling could restore clonal memory.

3. Key Contributions

1. Discovery of Dual Clonal Memories: The study establishes that human HSPCs possess two distinct, inherited forms of clonal memory: one governing division synchronicity and another governing fate commitment.
2. Decoupling of Division and Fate: Mathematical modeling proves that these two memories are distinct regulatory processes; one cannot be explained solely by the other.
3. Pathological Disruption: The study demonstrates that AML blasts exhibit a significant loss of division synchronicity (reduced clonal memory of division) compared to healthy HSPCs, despite retaining some fate memory.
4. Epigenetic Modulation: The authors show that clonal memory is a plastic trait that can be pharmacologically restored in leukemic cells using epigenetic inhibitors, without altering the total number of divisions.

4. Key Results

A. Clonal Memory in Healthy HSPCs

• Division Synchronicity: Sister cells (sharing a mother) and cousin cells (sharing a grandmother) divide synchronously. The correlation in division timing persists for at least two divisions (three generations).
• Fate Commitment: Cells within the same family tend to differentiate into similar lineages, showing significant intra-familial similarity in phenotype beyond what is expected by chance.
• Distinct Mechanisms: ABC modeling revealed that the best fit for the data requires a scenario where cells share both a memory of division timing and a memory of fate commitment. Models assuming only one type of memory or no memory were statistically rejected.
• Transcriptional Basis: scRNA-seq revealed that cells from the same family share unique transcriptional "fingerprints" (genes related to membrane trafficking, cytoskeleton, and metabolism) that are distinct from their differentiation state. A network analysis identified STAB1 as a central hub associated with family identity.

B. Disruption in AML

• Loss of Synchronicity: Primary AML blasts showed significantly lower correlation in division times between sisters and cousins compared to healthy controls. The "range" of divisions within a family was higher (more discordant).
• Independence from Genetics: This loss of synchronicity was observed across different AML patients and was not correlated with specific mutation profiles, suggesting it is a non-genetic, epigenetic, or regulatory defect.
• Fate vs. Division: While division memory was disrupted, the differentiation profile of AML cells remained relatively consistent with their healthy counterparts in terms of lineage bias, highlighting the specific loss of division coordination.

C. Pharmacological Restoration

• JQ-1 Treatment: Treating AML cells with the BET inhibitor JQ-1 (5 nM) significantly increased the synchronicity of division (restoring the correlation coefficient and the percentage of families with range = 0).
• Specificity: JQ-1 treatment did not change the total number of divisions or the overall clonogenicity, nor did it significantly alter the differentiation profile. This confirms that clonal memory of division is a tunable parameter regulated independently of cell cycle progression capacity.

5. Significance and Implications

• Fundamental Biology: This work redefines the understanding of stem cell heterogeneity, identifying clonal memory of division as a fundamental, conserved feature of human hematopoiesis that operates independently of differentiation fate.
• Disease Mechanism: The disruption of division synchronicity in AML suggests that leukemic cells may reduce clonal memory to increase adaptability and plasticity, allowing them to survive environmental stresses and evade therapies.
• Therapeutic Potential: The ability to pharmacologically "tune" clonal memory (restoring stability in cancer cells) opens a new avenue for therapy. By forcing leukemic cells to divide more synchronously, it may be possible to reduce intra-tumor heterogeneity and sensitize them to standard treatments.
• Methodological Advance: The integration of live-cell imaging, MultiGen lineage tracing, and single-cell transcriptomics provides a robust blueprint for studying cell lineage dynamics in other tissue systems.

In conclusion, the paper demonstrates that human HSPCs rely on distinct, heritable programs to coordinate division and fate. In leukemia, the division program is disrupted but remains epigenetically modifiable, offering a novel target for therapeutic intervention.