The spatial and multi-omic landscape of Down syndrome leukemogenesis: moving beyond cellular heterogeneity

Abstract

Down syndrome (DS) confers a developmentally rooted predisposition to both myeloid and lymphoid leukemias, particularly myeloid leukemia associated with DS (ML-DS) and acute lymphoblastic leukemia associated with DS (ALL-DS). While trisomy 21-driven gene dosage imbalance is central to this risk, DS leukemogenesis cannot be fully explained by recurrent mutations alone; it reflects a dynamic interplay between altered hematopoietic development, cell-intrinsic programs, and tissue microenvironmental cues. In this perspective, we argue that the field should move beyond cataloging cellular heterogeneity and adopt a topographic, multi-omic framework of DS leukemogenesis. We discuss how fetal niche biology shapes pre-leukemic evolution in ML-DS, including the developmental context of GATA1-mutant clones, and how therapy-driven bottlenecks may promote persistence of spatially protected residual disease in ALL-DS. We further highlight the translational potential of integrating spatially resolved transcriptomics with single-cell and protein-aware multi-omics to identify compartment-specific signaling programs and clinically actionable vulnerabilities. A spatially informed model of DS leukemia may improve biological stratification, clarify mechanisms of relapse and toxicity, and support the development of more effective and less toxic therapeutic strategies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40164-026-00815-y.

To the Editor,

Down syndrome (DS) is a paradigmatic human model of constitutional gene dosage imbalance. Trisomy 21 (T21) perturbs hematopoiesis through deregulation of chromosome 21 dosage-sensitive genes and downstream transcriptional networks, creating a permissive background for leukemogenesis [1, 2]. Individuals with DS are at markedly increased risk of both myeloid and lymphoid leukemias, most notably ML-DS and ALL-DS [3]. This predisposition is best understood not as a single-gene phenomenon but as systems-level reprogramming of hematopoietic development, in which T21 acts as a constitutive priming state that alters stem and progenitor behavior before overt leukemogenic mutations are acquired [4]. Human fetal hematopoiesis atlases reveal selective disruption of B-lymphoid, erythroid and myeloid development in T21 and an altered stromal environment marked by inflammatory signaling and perturbed endothelial–HSC communication [5]. These observations shift the focus from mutation-centric models toward an integrated view in which dosage imbalance, developmental timing, and tissue microenvironment jointly shape leukemia risk. Here we argue that the field should move beyond cataloging cellular heterogeneity and adopt a topographic, multi-omic framework of DS leukemogenesis.

In the myeloid setting, ML-DS is almost invariably preceded by transient abnormal myelopoiesis (TAM), driven by acquired GATA1 mutations (typically GATA1s) that emerge during fetal life [6]. Their developmental timing suggests that leukemogenic competence is established in a permissive fetal environment, not by mutation acquisition alone. Fetal bone marrow (BM) profiling in T21 shows increased inflammatory programs and altered endothelial–HSC signaling, including predicted enhancement of NOTCH-axis interactions and enrichment of TNF- and interferon/cytokine-response programs in stromal compartments [5]. These features provide a testable substrate for spatially restricted inflammatory hotspots and vascular niches that may shape clonal fitness. Progression from TAM to overt ML-DS requires cooperating lesions (e.g., cohesin or JAK/STAT hits), and a topographic model predicts that such second hits confer fitness within specific fetal vascular/stromal microdomains as hematopoiesis transitions from fetal liver to BM [7].

In contrast to the generally favorable outcomes of DS-adapted treatment in ML-DS, ALL-DS remains a major challenge because of combined relapse risk and treatment-related toxicity [4, 7]. ALL-DS is frequently associated with CRLF2 alterations and JAK/STAT pathway lesions, yet genetic events alone do not explain heterogeneity in treatment response [8]. Increasing evidence indicates that chemotherapy acts as an evolutionary bottleneck, enriching for stress-adapted, relatively quiescent states that persist in protective marrow niches [9]. Spatial and niche biology bridge these persister states and relapse: leukemia maintenance can depend on vascular endothelial niches, and leukemic clones can rewire BM architecture and build self-reinforcing stromal niches that shelter blasts from therapy [10]. Relapse is therefore not only a genetic event but also a topographic one, defined by which BM regions and vascular/immune configurations sustain persistence. Within this framework, rare subclones or persister-like populations may occupy microanatomical niches enriched for survival cues and reduced drug exposure. The key questions become which niches are permissive in DS, how they change with therapy, and which niche-derived dependencies are targetable without increasing systemic toxicity (see Table 1, Fig. 1).

Table 1.

The table summarizes key clinical, genetic, developmental, and microenvironmental differences between ML-DS and ALL-DS, highlighting distinct pre-leukemic trajectories, recurrent genomic drivers, and divergent therapeutic challenges, and emphasizing the emerging relevance of niche-specific, spatially resolved analyses

Comparative features of myeloid and lymphoid leukemias in Down Syndrome
Feature	Myeloid Leukemia associated with DS (ML-DS)	Acute Lymphoblastic Leukemia associated with DS (ALL-DS)
Increased risk (vs. non-DS)	~ 150-fold	~ 20-fold
Pre-leukemic phase	Transient abnormal myelopoiesis (TAM)	Not well defined (putative quiescent/persister reservoirs)
Key genetic drivers	GATA1 (GATA1s); cohesin genes (STAG2, SMC3); CTCF; JAK/STAT	CRLF2 aberrations; JAK2 (R683); RAS-pathway mutations
Peak age of diagnosis	Early childhood (< 5 years)	Childhood/adolescence (later peak than non-DS)
Overall prognosis	Generally favorable (OS > 80%)	Poor (OS ~ 35.7% vs. 75.8% in non-DS)
Spatial niche focus	Fetal liver/early BM; endothelial NOTCH axis (e.g., NOV/DLK1–NOTCH)	BM; protected quiescent niches (most informative in remission/early-treatment BM)
Therapeutic challenge	Targeting pre-leukemic clones and early evolutionary trajectories	High chemotherapy toxicity and relapse risk; niche-protected residual disease

Fig. 1.

Conceptual multi-omic and spatial framework of Down syndrome leukemogenesis. The schematic illustrates how T21 establishes a primed hematopoietic state predisposing to leukemic transformation, depicting a myeloid route characterized by fetal GATA1-mutant pre-leukemia (TAM/ML-DS continuum) and niche-dependent expansion, and a lymphoid route marked by clonal branching, therapy-driven bottlenecks, and persistence within protective BM niches in ALL-DS. The lower panel highlights the analytical integration of spatially resolved transcriptomic, single-cell RNA sequencing, and proteogenomic approaches to identify compartment-specific programs and clinically actionable vulnerabilities

Epigenetic deregulation adds a further layer. Increased dosage of chromatin-associated factors such as HMGN1 can globally amplify transcription and influence lineage programs, and HMGN1 is a functional dependency in CRLF2-driven DS-ALL [11]. Integrating chromatin state, protein phenotypes and spatial context may help localize when and where epigenetically primed progenitors gain a selective advantage.

These hypotheses can be operationalized with feasible sampling anchored to clinically accessible time points, comparing TAM/early ML-DS and paired diagnosis/remission ALL-DS specimens to age-matched controls, with spatial assays prioritized in remission or early-treatment marrow where microanatomical context is interpretable. Because overt leukemia disrupts normal architecture, such specimens may better reveal microanatomical reservoirs than densely infiltrated marrow. Details of proposed workflows, protein-aware multi-omic platforms, and BM-specific technical considerations are provided as Supplementary Material [12].

T21 creates a developmentally primed hematopoietic landscape that leukemogenic mutations exploit in a niche- and stage-dependent manner. A topographic, multi-omic approach, integrating spatial transcriptomics with single-cell and protein-aware multi-omics, is thus not only biologically informative but clinically strategic, locus regit actum. For DS-associated leukemias, where cure must be balanced against host vulnerability, this integrated framework may represent a key translational shift in the coming years.

Abbreviations

DS

Down syndrome

T21

Trisomy 21

ML-DS

Myeloid leukemia associated with Down syndrome

ALL-DS

Acute lymphoblastic leukemia associated with Down syndrome

BM

Bone marrow

TAM

Transient abnormal myelopoiesis

HSPC

Hematopoietic stem and progenitor cells

HSC/MPP

Hematopoietic stem cell/multipotent progenitor

Author contributions

EP developed the study concept and prepared the manuscript. RA, MB, MF and AR critically revised the manuscript and provided important intellectual input. All authors approved the final version.

Funding

This research received “Ricerca Corrente” funding from the Italian Ministry of Health to cover publication costs.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

All authors have read and approved the final version of the manuscript.

Competing interests

The authors declare no competing interests.