Barcode, Passenger, or Driver? Mitochondrial Heteroplasmy in the Somatic Evolution of Human Haematopoiesis

Technical Summary: Barcode, Passenger, or Driver? Mitochondrial Heteroplasmy in the Somatic Evolution of Human Haematopoiesis

Problem Statement
Human haematopoiesis is increasingly recognized as a lifelong somatic evolutionary process driven by clonal competition among haematopoietic stem and progenitor cells (HSPCs). While the evolutionary framework for this process is well-established regarding the nuclear genome (e.g., mutations in DNMT3A, TET2, JAK2), the role of the mitochondrial genome remains unsettled. Mitochondrial DNA (mtDNA) presents a unique evolutionary substrate due to its multicopy nature, high mutation rate, and the phenomenon of heteroplasmy (the coexistence of multiple mtDNA genotypes within a single cell). The central problem addressed by this review is determining the evolutionary role of somatic mtDNA variants in human blood: are they merely neutral markers of ancestry, passengers hitchhiking on nuclear-driven clones, active modifiers of cellular fitness, or independent drivers of clonal expansion? Current literature often conflates these states, leading to ambiguity in interpreting whether mitochondrial heteroplasmy is a cause or a consequence of clonal haematopoiesis (CH).

Methodology
This work is a critical narrative review synthesizing peer-reviewed literature from the past decade, with a focus on human population genomics, single-cell lineage tracing, and experimental models of mitochondrial dependency. The author does not present new primary data but rather constructs a conceptual framework to reconcile contradictory evidence. The methodology involves:

1. Literature Synthesis: Analyzing large-scale population studies (e.g., UK Biobank, 750,000 individuals), single-cell multi-omics studies, and studies on mitochondrial diseases.
2. Conceptual Categorization: Defining distinct evolutionary states for mtDNA variants based on biological role, expected observations, and evidentiary thresholds.
3. Critical Evaluation of Evidence: Explicitly addressing methodological limitations, such as the distinction between bulk sequencing signals and single-cell heteroplasmy, and the confounding effects of lineage-specific selection and cellular composition.
4. Framework Development: Proposing a classification system that accounts for mitonuclear epistasis (interaction between nuclear and mitochondrial genotypes) and the multi-level nature of selection (intracellular vs. intercellular).

Key Contributions and Proposed Framework
The paper proposes a four-state framework (plus a superimposed epistatic state) to classify mitochondrial heteroplasmy in haematopoiesis, moving beyond a binary "neutral vs. functional" dichotomy:

1. Barcode: mtDNA variants that faithfully record cellular ancestry without measurably altering fitness. These are useful for lineage tracing but require high-confidence molecular support to distinguish from pre-existing heteroplasmies or artifacts.
2. Passenger: Variants that increase in observed abundance solely because the cellular clone carrying them expands due to an independent nuclear driver. Large population studies suggest the majority of age-associated mtDNA variants fall into this category, accumulating via stochastic mutation and being revealed by clonal expansion (genetic hitchhiking).
3. Fitness Modifier: Variants that do not initiate clonal expansion independently but alter the selection coefficient of an existing clone. Evidence suggests mitochondrial metabolism can modify the fitness of clones driven by specific nuclear mutations (e.g., DNMT3A R882 mutants show increased dependence on oxidative phosphorylation).
4. Driver: Variants that directly confer a competitive advantage sufficient to initiate or sustain clonal expansion. The author argues that current evidence for this state is extremely limited and requires a stringent evidentiary threshold (recurrent selection, temporal precedence, and experimental sufficiency).
5. Mitonuclear Epistasis: A state where the fitness effect of a nuclear mutation depends on the mitochondrial genotype, and vice versa. This suggests that clonal fitness is an emergent property of the combined nuclear-mitochondrial genotype, lineage, and cellular context.

Key Results and Evidence Synthesis

• Passenger Dominance: Analysis of ~750,000 individuals indicates that most age-associated mtDNA single nucleotide variants (SNVs) in blood are likely neutral passengers that accumulate due to replication errors and become detectable only when the host clone expands.
• Lineage-Specific Selection: Studies of mitochondrial disease demonstrate that pathogenic variants undergo strong purifying selection in specific lineages (e.g., T cells) but may be tolerated in others, proving that mitochondrial genotype can alter cellular representation in a lineage-dependent manner.
• Metabolic Dependencies: Experimental data (e.g., Gozdecka et al.) show that DNMT3A R882 mutant HSPCs exhibit specific metabolic dependencies (increased oxidative phosphorylation) that can be pharmacologically targeted, suggesting mitochondrial state modifies the fitness of nuclear-driven clones.
• Risk Association: Heteroplasmy is associated with higher risk of myeloid neoplasia in individuals with CH, but this association may reflect the age/size of the clone (passenger) or a modifying effect rather than direct causality.
• Methodological Challenges: Bulk sequencing cannot distinguish between a small clone with high intracellular heteroplasmy and a large clone with low heteroplasmy. Furthermore, "selection of an mtDNA mutation" is ambiguous without defining the level of selection (intracellular drift vs. cellular expansion).

Significance and Claims
The paper claims that the field must shift from asking if mitochondrial heteroplasmy is important to defining how it functions in specific contexts. The primary significance of this review is the establishment of a rigorous, context-dependent classification system that prevents the premature labeling of mtDNA variants as "drivers."

The author argues that:

• Mitochondrial heteroplasmy is not a uniform class: It ranges from effectively neutral barcodes to context-dependent fitness modifiers.
• The "Driver" label requires high evidence: Current observational data (association with age, clone size, or disease risk) is insufficient to claim driver status; experimental perturbation is required to prove causality.
• Mitonuclear interaction is central: The evolutionary unit is not the nuclear or mitochondrial genome in isolation, but the combined mitonuclear genotype. The fitness landscape is conditional, not fixed.
• Clinical Implications: While many variants may be passengers, they can serve as valuable biomarkers for risk stratification. Furthermore, even if a variant is a passenger or modifier, the mitochondrial pathways it influences (e.g., in DNMT3A mutants) may be viable therapeutic targets for "evolutionary interception."

The review concludes that distinguishing these states requires longitudinal, lineage-resolved, and experimentally perturbable models. It posits that in defined contexts, mitochondrial heteroplasmy may form part of the evolutionary system itself, rather than serving merely as an instrument for observing somatic evolution.