Single-cell architecture of purinergic signaling in human cord blood hematopoietic stem and progenitor cells

Abstract

Purinergic signaling has emerged as a key regulator of hematopoietic stem and progenitor cell (HSPC) trafficking, metabolism, and innate immune responsiveness. Our previous studies demonstrated that extracellular ATP promotes HSPC mobilization, homing, and engraftment by activating P2X purinergic receptors and engaging the Nlrp3 inflammasome downstream, whereas enzymatic conversion of ATP to extracellular adenosine exerts opposite, anti-inflammatory effects. Subsequently, we postulated that purinergic signaling is an evolutionarily ancient regulatory system that remains intrinsically embedded within the hematopoietic stem cell program. However, its transcriptional organization across distinct human HSPC subsets remains unknown. We applied single-cell RNA sequencing to human umbilical cord blood–derived CD133⁺Lin⁻CD45⁺ and CD34⁺Lin⁻CD45⁺ cells enriched for HSPCs. We identified transcriptionally distinct clusters representing primitive progenitors and lineage-primed intermediates, and we demonstrated a hierarchical organization of purinergic receptors and nucleotide-metabolizing enzymes across these populations. Primitive HSPCs exhibited a restricted purinergic repertoire coupled with intracellular nucleotide recycling machinery, consistent with a tightly regulated metabolic–immune state. Lineage-biased clusters showed selective enrichment of receptors and ectonucleotidases associated with inflammatory activation, migration, and fate commitment. Together, these findings establish purinergic signaling as a fundamental, cell-intrinsic regulator of early hematopoiesis and highlight how this ancient signaling pathway shapes human stem cell fate decisions.

Introduction

Purinergic signaling is an evolutionarily conserved cell–cell communication system that regulates inflammation, immunity, metabolism, and stem cell behavior [1–5]. In the hematopoietic system, extracellular nucleotides such as ATP and their metabolites signal through seven ionotropic P2X receptors, eight G protein–coupled P2Y receptors, and four adenosine-activated P1 receptors to modulate hematopoietic stem and progenitor cell (HSPC) quiescence, activation, migration, and lineage priming [6, 7]. The extracellular availability of these ligands is tightly controlled by ectonucleotidases, including CD39 and CD73, as well as intracellular enzymes that recycle or degrade adenosine, thereby shaping local purinergic tone within the bone marrow microenvironment [8–12]. CD39 (ENTPD1) hydrolyzes extracellular ATP and ADP to AMP, whereas CD73 (NT5E) further converts AMP to adenosine, thereby controlling the balance between pro-inflammatory ATP signaling and anti-inflammatory adenosine signaling [13]. By integrating metabolic cues with innate immune responses, purinergic signaling contributes to both steady-state and stress hematopoiesis and is also involved in the pathogenesis of leukemia [14–20].

Over the past two decades, our group has systematically demonstrated that purinergic signaling is a vital regulatory axis governing hematopoietic stem and progenitor cell (HSPC) trafficking [10, 21–30]. We showed that extracellular ATP (eATP) promotes HSPC mobilization, homing, and engraftment by activating P2X purinergic receptors, including P2X1, P2X4, and P2X7 [13, 31, 32], and that these effects critically depend on activation of the intracellular Nlrp3 inflammasome [13, 27, 33–35]. In parallel, we identified a counter-regulatory pathway in which enzymatic conversion of eATP to extracellular adenosine by CD39 and CD73 ectonucleotidases impairs HSPC trafficking by engaging the A₂B adenosine receptor and inducing heme oxygenase-1 (HO-1), thereby inhibiting Nlrp3 inflammasome activity [36, 37]. Building on these functional studies, and given that eATP is the most important alarmin of innate immunity, we proposed that purinergic signaling and innate immunity are evolutionarily ancient, metabolically coupled systems that remain hard-wired into hematopoietic stem cells and regulate their fate decisions, stress responses, and interactions with the bone marrow microenvironment [23, 30, 38]. However, despite this extensive functional and conceptual framework, it remains unknown how individual components of purinergic signaling are expressed across transcriptionally distinct human HSPC subsets, particularly at the earliest stages of hematopoietic development.

In this study, we address this gap by applying single-cell RNA sequencing to human umbilical cord blood–purified CD133⁺Lin⁻CD45⁺ and CD34⁺Lin⁻CD45⁺ HSPCs. The CD34⁺Lin⁻CD45⁺ and CD133⁺Lin⁻CD45⁺ fractions represent enriched populations of hematopoietic stem and progenitor cells, with CD133⁺ cells corresponding to more primitive stem-like states, whereas CD34⁺ cells comprise a broader and more heterogeneous progenitor population that includes cells at more advanced stages of lineage priming, thereby enabling the analysis of early versus more differentiated stages of hematopoietic development. This approach provides single-cell resolution of purinergic receptors and nucleotide-metabolizing enzymes, thereby establishing a mechanistic link between our prior functional studies and the cellular heterogeneity underlying human hematopoiesis.

Although previous studies have firmly established purinergic signaling as a critical regulator of HSPC trafficking and bone marrow niche dynamics, they have relied primarily on bulk transcriptomic and functional assays [31, 39–44]. Such approaches, while informative, obscure the cellular heterogeneity and transcriptional complexity inherent in the HSPC compartment. In particular, the distribution and regulation of purinergic receptors and ectonucleotidases across transcriptionally distinct HSPC subpopulations remain poorly understood. This knowledge gap is especially pronounced in umbilical cord blood (UCB), a rich source of naïve and developmentally early HSPCs, where the single-cell architecture of purinergic signaling has not been systematically defined.

To address this, we applied high-throughput single-cell RNA sequencing (scRNA-seq) to dissect purinergic signaling pathways in two immunophenotypically distinct UCB-derived HSPC subsets: CD34⁺Lin⁻CD45⁺ and CD133⁺Lin⁻CD45⁺ cells. Building on our prior work profiling very small embryonic-like stem cells (VSELs) using scRNA-seq [45]—which focused primarily on pluripotency and early developmental regulators—we now provide the first comprehensive single-cell analysis mapping purinergic signaling components in human HSPCs. This study employed scRNA-seq to i) resolve the expression patterns of purinergic P1 and P2 receptors at single-cell resolution; ii) profile the spatial distribution of ectonucleotidases and nucleotide-metabolizing enzymes (e.g., CD39, CD73, ENPPs, ADA, ADK); iii) identify transcriptionally defined HSPC clusters with distinct purinergic signaling profiles; and iv) infer how differential purinergic signaling may influence lineage priming, migratory behavior, and responsiveness to environmental cues.

Beyond its role in steady-state hematopoiesis, purinergic signaling has emerged as a key regulator of hematologic malignancies and stem cell–niche interactions. Altered extracellular ATP levels, P2X7 receptor signaling, and ectonucleotidase activity have been implicated in leukemic stem cell survival, inflammatory niche remodeling, and therapy resistance [46, 47]. Moreover, ATP-dependent Nlrp3 inflammasome activation and adenosine-mediated immunomodulation influence both normal and malignant hematopoietic progenitors within the bone marrow microenvironment [41, 42]. Defining the transcriptional architecture of purinergic signaling in normal human HSPCs is therefore essential not only for understanding physiological stem cell regulation but also for providing a reference framework for interpreting purinergic dysregulation in leukemic states.

Materials and methods

Sample collection and initial processing

Human umbilical cord blood (hUCB) sample was collected immediately after delivery from healthy term birth at the Department of Obstetrics and Gynecology, Medical University of Warsaw. All procedures were performed with informed consent and approved by the Institutional Bioethics Committee (approval no. KB/50/2022). Blood was collected directly from the umbilical vessels into EDTA-coated tubes to prevent coagulation. hUCB unit was diluted 1:1 with phosphate-buffered saline (PBS), carefully layered over Ficoll-Paque (GE Healthcare), and centrifuged at 400 × g for 30 min at 4 °C without brake. The mononuclear cell (MNC) layer, containing leukocytes and platelets, was harvested at the interface, washed twice in PBS, and immediately processed for downstream immunophenotyping and cell sorting.

Cell staining and sorting

Cells were stained with antibodies and sorted according to the procedure described previously [48]. Briefly, sample was incubated with hematopoietic lineage marker cocktail consisting of FITC-conjugated anti-CD235a (clone GA-R2 [HIR2]; Cat No: 561017), anti-CD2 (clone RPA-2.10; Cat No: 555326), anti-CD3 (clone UCHT1; Cat No: 561807), anti-CD14 (clone M5E2; Cat No: 561712), anti-CD16 (clone 3G8; Cat No: 560996), anti-CD19 (clone HIB19; Cat No: 560994), anti-CD24 (clone ML5; Cat No: 560992), anti-CD56 (clone NCAM16.2; Cat No: 340410), and anti-CD66b (clone G10F5; Cat No: 561927) (all BD Biosciences, San Jose, CA, USA); together with PE-Cy7-conjugated anti-CD45 (clone HI30; Cat No: 304016; BioLegend, San Diego, CA, USA), PE-conjugated anti-CD34 (clone 581; Cat No: 343506 BioLegend), and APC-conjugated anti-CD133 (clone AC133; Cat No: 130-110-963 Miltenyi Biotec, Bergisch Gladbach, Germany). Antibodies were used according to the manufacturers’ recommendations. Small events, ranging from 4 to 12 µm, were included in the gate and analyzed for expression of CD45, lineage markers, CD34, and CD133. Hematopoietic stem and progenitor cell (HSPC) populations, specifically CD133+Lin-CD45+ and CD34+Lin-CD45 + , were sorted on the MoFlo Astrios EQ cell sorter.

Single-cell RNA sequencing and computational analysis

Single-cell libraries were pooled and run on an Illumina NextSeq 1000/2000 (Illumina, San Diego, CA, USA) with P2 flow cell chemistry (200 cycles) in paired-end mode (read 1–28 bp, read 2–90 bp, index 1–10 cycles, index 2–10 cycles), assuming 25 000 reads per cell. Single-cell transcriptomic profiling of sorted CD133⁺Lin⁻CD45⁺ and CD34⁺Lin⁻CD45⁺ HSPCs was performed as previously described [45, 48, 49]. Raw sequencing reads were processed with the 10x Genomics Cell Ranger pipeline for alignment, UMI quantification, and gene-barcode matrix generation. Downstream analysis was conducted in R (v4.5.1) using the Seurat package (v5.1.0). Initial quality control excluded cells with fewer than 200 detected genes, more than 2500 genes, or >5% mitochondrial RNA content to ensure quality and remove dead/dying cells. Doublets were identified and excluded using a manual filtering strategy based on outlier detection in the nFeature_RNA and nCount_RNA distributions. After quality control and doublet exclusion, a total of 3 602 CD133⁺ cells and 4811 CD34⁺ cells were retained for downstream analysis. The median number of detected genes per cell was 1540 for the CD133⁺ population and 1332 for the CD34⁺ population. The median number of UMIs per cell was 3360.5 and 2628 for CD133⁺ and CD34⁺ cells, respectively. Gene expression values were log-normalized and scaled before dimensionality reduction. The top 2000 most variable genes were selected for principal component analysis (PCA), followed by uniform manifold approximation and projection (UMAP) for visualization. A shared nearest-neighbor graph-based clustering algorithm was used to identify transcriptionally distinct subpopulations. Clustering resolution was optimized using data-driven heuristics, yielding 14 clusters in each dataset.

Cluster-specific marker genes were identified using the Wilcoxon rank-sum test implemented in Seurat, with p-values adjusted for multiple comparisons using the Benjamini–Hochberg method (adjusted p-value < 0.05) and log2 fold change >1. The test is based on a two-sided comparison of gene expression between groups. Although formal tests for equality of variance were not performed, this is consistent with established scRNA-seq analysis practices, where variability is intrinsically cell- and cluster-dependent. Within-cluster variation is instead directly reflected in the distribution of single-cell expression values shown in violin and dot plots. For this study, only positively expressed genes were considered for manual cluster annotation and functional characterization, including Gene Ontology, KEGG, and Reactome pathway analyses. To further define the transcriptional heterogeneity and functional specialization of progenitor populations, the top 50 most highly expressed genes in each cluster were subjected to detailed Reactome and Gene Ontology annotation, enabling assignment of key biological processes such as proliferation, innate immune response, and adaptive immunity. All visualizations were generated in R using ggplot2 (v3.4.4) and Seurat’s integrated plotting functions.

Gene expression and functional annotation of purinergic signaling

Gene expression and functional annotation procedures are described in the Supplementary Methods.

Results

Single-cell transcriptomic profiling identifies primitive and lineage-primed HSPC subsets

Single-cell RNA sequencing of human umbilical cord blood–derived CD133⁺Lin⁻CD45⁺ and CD34⁺Lin⁻CD45⁺ cells revealed transcriptionally distinct clusters corresponding to primitive progenitors and lineage-primed intermediates (Figs. 1 and 2). Uniform manifold approximation and projection (UMAP) analysis identified multiple clusters within each population, underscoring the inherent heterogeneity of early hematopoiesis. Cluster annotation with established hematopoietic markers confirmed the presence of multipotent progenitors, early lineage-biased states, and cells primed for myeloid, lymphoid, erythroid, and megakaryocytic lineages.

Fig. 1. Single-cell transcriptomics resolves primitive and lineage-primed states within CD133⁺Lin⁻CD45⁺ cordblood HSPCs.

A UMAP representation of transcriptionally defined clusters within CD133⁺Lin⁻CD45⁺ hematopoietic progenitor cells. Single-cell RNA sequencing identified multiple transcriptionally distinct clusters, color-coded and annotated by lineage-associated marker expression (right panel). Red-highlighted clusters correspond to the most primitive progenitors, characterized by high expression of CD133, CD34, and KIT (c-Kit). The remaining clusters display transcriptional signatures indicative of lineage priming toward monocytes (CD14, CD68), T cells (CD2, CD3, CD4), natural killer cells (CD16, CD2L), B cells (CD19, CD22), erythroid cells (GYPA), and megakaryocytes (ITGA2B/CD41). The term dim denotes relatively low transcript abundance compared with marker-high populations. B Expression of lineage-associated markers across CD133⁺Lin⁻CD45⁺ clusters. Dot plot illustrating the expression of selected lineage-associated genes in transcriptionally defined clusters. Primitive progenitor clusters are highlighted with red circles. Dot size represents the proportion of cells within each cluster expressing a given gene, while color intensity (blue to red) indicates increasing average expression levels. C Violin plots of lineage-associated marker expression across CD133⁺Lin⁻CD45⁺ clusters. Violin plots show the distribution of transcript expression levels for selected lineage-associated genes across transcriptionally defined clusters. Each dot represents an individual cell, with the position along the y-axis corresponding to normalized transcript abundance (higher values indicate higher expression).

Fig. 2. Single-cell transcriptomics reveals the hematopoietic hierarchy within CD34⁺Lin⁻CD45⁺ cord blood HSPCs.

A UMAP representation of transcriptionally defined clusters within CD34⁺Lin⁻CD45⁺ hematopoietic progenitor cells. Single-cell RNA sequencing identified multiple transcriptionally distinct clusters, color-coded and annotated by lineage-associated marker expression (right panel). Red-highlighted clusters correspond to the most primitive progenitors, characterized by high expression of CD34, CD133, and KIT (c-Kit). The remaining clusters display transcriptional signatures indicative of lineage priming toward monocytes (CD14, CD68), T cells (CD2, CD3, CD4), natural killer cells (CD16, CD2L), B cells (CD19, CD22), erythroid cells (GYPA), and megakaryocytes (ITGA2B/CD41). The term dim denotes relatively low transcript abundance compared with marker-high populations. B Expression of lineage-associated markers across CD34⁺Lin⁻CD45⁺ clusters. Dot plot illustrating the expression of selected lineage-associated genes in transcriptionally defined clusters. Primitive progenitor clusters are highlighted with red circles. Dot size represents the proportion of cells within each cluster expressing a given gene, while color intensity (blue to red) indicates increasing average expression levels. C Violin plots of lineage-associated marker expression across CD34⁺Lin⁻CD45⁺ clusters. Violin plots show the distribution of normalized transcript expression levels for selected lineage-associated genes across transcriptionally defined clusters. Each dot represents an individual cell, with the position along the y-axis corresponding to transcript abundance (higher values indicate higher expression).

In the CD133⁺Lin⁻CD45⁺ dataset, several clusters exhibited high expression of stemness-associated markers, including CD133, CD34, and KIT, and low expression of lineage-restricted genes, consistent with a primitive progenitor identity (Fig. 1A–C). Functional annotation linked these clusters to developmental, proliferative, and metabolic programs. The remaining clusters displayed transcriptional signatures indicative of commitment to monocytes, lymphocytes, natural killer cells, erythroid cells, or megakaryocytes.

A comparable hierarchical organization was observed in the CD34⁺Lin⁻CD45⁺ population (Fig. 2A–C). Primitive clusters expressed CD34 and early progenitor markers and were enriched for metabolic and developmental pathways, whereas lineage-primed clusters displayed gene expression profiles characteristic of innate and adaptive immune lineages and of erythro-megakaryocytic differentiation. Together, these analyses establish a consistent framework of primitive and fate-decision states across both immunophenotypically defined HSPC populations.

Purinergic receptor expression is hierarchically organized across HSPC subsets

To define the architecture of purinergic signaling across transcriptionally distinct HSPC subsets, we examined the expression of P1 (ADORA), P2X, and P2Y purinergic receptor genes in both datasets (Figs. 3 and 4). Purinergic receptor expression was highly structured and non-uniform, revealing a conserved core receptor repertoire in primitive progenitors and a progressively diversified receptor landscape in lineage-primed clusters.

Fig. 3. Purinergic receptor expression is hierarchically organized across human cord blood HSPC subsets.

Dot plots show mean expression levels and the proportion of receptor-positive cells for P1 (ADORA), P2X, and P2Y purinergic receptor family genes across transcriptionally defined clusters. Primitive progenitor clusters are indicated by red circles. A CD133⁺Lin⁻CD45⁺ cells. B CD34⁺Lin⁻CD45⁺ cells.

Fig. 4. Single-cell expression patterns reveal cluster-specifi c purinergic receptor repertoires in human HSPCs.

A Expression of purinergic receptor genes in CD133⁺Lin⁻CD45⁺ cells. Violin plots show the distribution of normalized transcript expression levels for P1 (ADORA), P2X, and P2Y purinergic receptor genes across transcriptionally defined clusters. Each dot represents an individual cell, with the y-axis position corresponding to transcript abundance. The width of each violin reflects the density of cells at a given expression level. Receptors not detected above threshold in any cluster were excluded from the analysis. B Expression of purinergic receptor genes in CD34⁺Lin⁻CD45⁺ cells. Violin plots show the distribution of normalized transcript expression levels for P1 (ADORA), P2X, and P2Y purinergic receptor genes across transcriptionally defined clusters. Each dot represents an individual cell, with the y-axis position corresponding to transcript abundance. The width of each violin reflects the density of cells at a given expression level. Receptors not detected above threshold in any cluster were excluded from the analysis.

Primitive HSPC clusters in both CD133⁺ and CD34⁺ populations consistently expressed a restricted set of purinergic receptors, most prominently P2RX1, P2RX4, P2RX7, and ADORA2A and ADORA2B. This conserved receptor profile suggests a basal purinergic signaling state associated with metabolic regulation and immune surveillance rather than overt inflammatory activation. In contrast, lineage-biased clusters showed selective enrichment for receptors previously linked to inflammatory signaling, chemotaxis, and stress responses, including P2RY11, P2RY13, and P2RY14.

Several receptors showed pronounced lineage specificity. P2RX5 expression was selectively enriched in B-cell–associated clusters across both datasets, whereas P2RY12 was confined to platelet-associated clusters, consistent with the known roles of these receptors in immune and megakaryocytic biology. Importantly, purinergic receptor expression patterns aligned with transcriptional state and lineage commitment rather than with CD133 or CD34 immunophenotype per se, indicating that purinergic signaling architecture is primarily dictated by differentiation status.

Collectively, these data suggest that purinergic receptor expression is hierarchically organized during early human hematopoiesis, with primitive progenitors maintaining a constrained signaling repertoire that expands and diversifies during fate specification. These patterns align with receptors previously shown by our group to regulate HSPC mobilization and homing, thereby providing a transcriptional basis for their functional heterogeneity.

Dynamic remodeling of purinergic regulatory enzymes during hematopoietic differentiation

We next analyzed the expression of enzymes that regulate extracellular nucleotide metabolism, including ectonucleotidases, adenosine-metabolizing enzymes, and components of intracellular nucleotide recycling (Figs. 5 and 6). Across both datasets, enzyme expression patterns showed clear state- and lineage-dependent organization, mirroring the hierarchical distribution of purinergic receptors.

Fig. 5. Purinergic regulatory enzymes are diff erentially expressed across CD133⁺Lin⁻CD45⁺ HSPC states.

Violin plots show the distribution of normalized transcript expression levels for key enzymes in nucleotide metabolism, including ectonucleotidases (ENTPD1/CD39, NT5E/CD73, ENPP1), adenosine deaminase (ADA), adenosine kinase (ADK), and mitochondrial ATP synthase (ATP5F1A), across transcriptionally defined clusters. Each point represents an individual cell, with the y-axis position indicating transcript abundance and violin width reflecting the density of cells at each expression level.

Fig. 6. Purinergic regulatory enzymes undergo state-dependent remodeling across CD34⁺Lin⁻CD45⁺ HSPCs.

Violin plots show the distribution of normalized transcript expression levels for key enzymes involved in nucleotide metabolism, including ectonucleotidases (ENTPD1/CD39, NT5E/CD73, ENPP1), adenosine deaminase (ADA), adenosine kinase (ADK), and mitochondrial ATP synthase (ATP5F1A), across transcriptionally defined clusters. Each point represents an individual cell, with the y-axis position indicating transcript abundance. Violin width reflects the density of cells at each expression level.

Primitive HSPC clusters preferentially expressed enzymes involved in intracellular nucleotide recycling and mitochondrial ATP production, including adenosine kinase (ADK) and ATP synthase subunits, consistent with a metabolically regulated purinergic state. In contrast, lineage-primed clusters showed selective enrichment for ectonucleotidases such as ENTPD1 (CD39) and NT5E (CD73), indicating a progressive acquisition of extracellular ATP-degrading and adenosine-generating capacity during differentiation.

Adenosine-degrading enzymes, including adenosine deaminase (ADA), were detected in both primitive and lineage-biased populations, suggesting that tight control of adenosine availability is a general feature of early hematopoiesis. Members of the ENPP family showed cluster-restricted expression, indicating specialized roles in nucleotide metabolism within specific progenitor states. Together, these findings indicate that purinergic enzyme expression is dynamically remodeled during hematopoietic differentiation, shifting from intracellular nucleotide maintenance in primitive progenitors to extracellular signal modulation in lineage-committed cells.

Functional annotation links purinergic architecture to hematopoietic hierarchy

Functional annotation of the most highly expressed genes in each cluster using Reactome and Gene Ontology analyses further supported the hierarchical organization of HSPCs (Figs. 7 and 8). Primitive clusters were enriched for developmental, proliferative, translational, and metabolic pathways, whereas lineage-biased clusters were associated with innate immune responses, adaptive immunity, erythroid differentiation, or megakaryocytic programs.

Fig. 7. Functional annotation defi nes primitive and lineage-primed states within CD133⁺Lin⁻CD45⁺ HSPCs.

Reactome and Gene Ontology analyses identified the general biological processes associated with each cluster, including proliferation, innate immune system, and adaptive immune system (annotated on the plot). The plot also shows lineage markers with the highest expression per cluster. Clusters highlighted in red represent truly primitive progenitors, while the remaining clusters correspond to cells in distinct fate-decision states.

Fig. 8. Functional annotation maps developmental and lineage-specifi c programs within CD34⁺Lin⁻CD45⁺ HSPCs.

Reactome and Gene Ontology analyses identified the general biological processes associated with each cluster (annotated on the plot). The plot also shows lineage markers with the highest expression per cluster. Clusters highlighted in red represent truly primitive progenitors, whereas the remaining clusters correspond to cells in distinct fate decision states.

Notably, the functional stratification of clusters closely mirrored the organization of purinergic receptor and enzyme expression, indicating a coordinated relationship among transcriptional state, lineage commitment, and purinergic signaling configuration. Primitive progenitors combined developmental and metabolic programs with a constrained purinergic repertoire, whereas fate-decision clusters integrated immune and lineage-specific pathways with expanded purinergic signaling capacity.

Integrated view of purinergic signaling during early human hematopoiesis

Collectively, these results suggest that purinergic signaling components are not uniformly distributed across the human HSPC compartment but instead exhibit a hierarchical, subset-specific organization. Primitive progenitors maintain a constrained purinergic program linked to metabolic regulation and immune surveillance, whereas lineage-primed clusters progressively acquire receptor and enzyme configurations that enable inflammatory responsiveness, chemotaxis, and fate commitment. This single-cell framework provides a mechanistic basis for previously observed context-dependent effects of purinergic signaling on HSPC trafficking and function.

Notably, this hierarchical organization of purinergic receptors and nucleotide-metabolizing enzymes was consistently observed in independently purified CD133⁺Lin⁻CD45⁺ and CD34⁺Lin⁻CD45⁺ populations, indicating that differentiation state, rather than immunophenotypic selection, governs purinergic architecture.

Discussion

Single-cell architecture of purinergic signaling explains functional heterogeneity in HSPC trafficking

The present single-cell analysis provides a mechanistic context for our previous functional studies demonstrating a central role for purinergic signaling in HSPC trafficking. Earlier in vivo and in vitro experiments showed that genetic deficiency or pharmacological inhibition of P2X1, P2X4, or P2X7 receptors impaired HSPC mobilization, homing, and engraftment in an Nlrp3 inflammasome-dependent manner [13, 31, 32]. However, these studies could not determine whether these receptors were uniformly expressed across the HSPC compartment or confined to specific functional subsets. It is conceivable that differential expression of P2X family members across developmental stages may provide functional redundancy, as P2X4 and P2X7 activate overlapping downstream pathways [50–52]. Whether such mechanisms operate in early HSPCs will require future investigation. As reported, both receptors activate similar pathways, and it has even been postulated that they can dimerize [53, 54].

Our single-cell data suggest that purinergic signaling is hierarchically organized. The most primitive CD133⁺ and CD34⁺ progenitors preferentially express a limited set of purinergic receptors and enzymes involved in intracellular nucleotide recycling, consistent with a tightly regulated metabolic and immune surveillance state. The preferential expression of selected purinergic receptors in primitive HSPCs likely reflects specific functional requirements of early hematopoietic states, including metabolic sensing, stress responsiveness, and immune surveillance. Rather than indicating the evolutionary “age” of individual receptors, this pattern may suggest that a conserved subset of purinergic signaling components is preferentially utilized at early stages to maintain stem cell integrity under fluctuating environmental conditions. This interpretation is consistent with the observed restriction of receptor diversity in primitive clusters and its progressive expansion during lineage priming. In contrast, lineage-primed and fate-decision clusters show selective enrichment of receptors and ectonucleotidases previously implicated in inflammatory activation, chemotaxis, and trafficking. This organization provides a cellular explanation for why perturbation of purinergic signaling produces strong functional effects despite relatively modest changes in bulk gene expression.

Importantly, these findings align with our prior observations that extracellular adenosine acts as a negative regulator of HSPC trafficking via the A2B receptor and by inhibiting the intracellular pattern recognition receptor Nlrp3 inflammasome [37]. Although primitive HSPC clusters showed limited expression of ectonucleotidases such as CD39 and CD73, they retained expression of adenosine receptors (A2A and A2B). This apparent discrepancy may reflect the dependence of primitive HSPCs on extracellular adenosine generated by neighboring niche cells, including stromal and immune populations. In addition, circulating adenosine and inducible expression of ectonucleotidases under stress or inflammatory conditions may provide alternative sources of ligand availability. Thus, adenosine signaling in primitive HSPCs is likely context-dependent and may not be fully captured by steady-state transcriptomic profiling [13, 55]. The preferential expression of ectonucleotidases and adenosine-related signaling components in specific transcriptional subsets suggests that the balance between ATP-driven activation and adenosine-mediated suppression is dynamically tuned during early hematopoietic fate decisions.

Although NT5E (CD73) transcripts were modest across primitive clusters, CD73 expression is dynamically regulated by microenvironmental cues and inflammatory signals. Thus, transcript levels in steady-state umbilical cord blood may not fully reflect protein abundance or inducible activity observed during mobilization or under stress conditions.

These results further support the emerging concept that purinergic signaling and innate immunity form an integrated regulatory network in early hematopoiesis [17, 56]. In our recent work on the intracellular complement - complosome, we demonstrated that intracellular complement activation operates within primitive HSPCs as a metabolic and immune regulatory module [57]. The present study extends this framework by showing that purinergic signaling components are similarly embedded within primitive human HSPCs at single-cell resolution. Together, these findings indicate that evolutionarily ancient danger-sensing pathways are not peripheral modulators but fundamental constituents of the hematopoietic stem cell program. An open question under investigation is the role of purinergic signaling, if it occurs similar to the complosome, in autocrine-, intrinsic-manner to modulate cell fate through interactions with intracellular purinergic receptors expressed on mitochondria or mitochondria-associated membranes (MAMs) [58–60].

A limitation of the present study is its reliance on transcriptomic data, which do not directly capture receptor protein abundance or functional activity. Therefore, the observed expression patterns should be interpreted as indicative of potential functional states rather than direct evidence of receptor activity. Although our findings are strongly supported by extensive prior functional work from our group, future studies integrating single-cell proteomics or functional perturbation approaches will be required to directly validate receptor activity at the protein level.

Importantly, the transcriptional framework defined here may serve as a reference for understanding purinergic dysregulation in leukemia. Leukemic stem cells have been shown to exploit purinergic signaling pathways, including elevated extracellular ATP levels, aberrant P2X7 receptor activation, and increased ectonucleotidase activity leading to adenosine-mediated immunosuppression. The hierarchical organization observed in normal HSPCs suggests that leukemic transformation may involve selective modulation or rewiring of these pathways, potentially contributing to niche remodeling, immune evasion, and therapy resistance [11, 55].

In summary, this study provides, for a first time a single-cell-resolved map of purinergic receptors and nucleotide-metabolizing enzymes across primitive and lineage-biased human HSPCs. By integrating these data with our prior functional data and conceptual work, we demonstrate that purinergic signaling is not merely an inflammatory pathway but a core regulator of early hematopoietic development. These findings position purinergic signaling as a fundamental component of the stem cell regulatory network and establish a cellular framework for understanding how metabolic and innate immune cues shape human hematopoiesis. This study may have implications for optimizing stem cell mobilization and transplantation strategies that rely on modulating purinergic and innate immune signaling pathways. It is also relevant to understand better pathogenesis of leukemia, where purinergic signaling plays an important role.

Author contributions

MM – Investigation, Methodology, Formal analysis, Writing – original draft. JJ – Investigation, Methodology, Formal analysis, Writing – original draft. PK - Investigation, Methodology, MZR – Supervision, Writing – review and editing. MK – Conceptualization, Project administration, Supervision, Investigation, Writing – original draft, Writing – review and editing

Data availability

The dataset(s) supporting the conclusions of this article are available in the Sequence Read Archive (SRA) (https://www.ncbi.nlm.nih.gov/sra) under the unique persistent identifier PRJNA1128409 (https://www.ncbi.nlm.nih.gov/sra/PRJNA1128409).

Code availability

Code used for analysis is available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.

Ethics statement

The study was conducted under the Declaration of Helsinki (ethical principles for medical research involving human subjects) and approved by the Medical University of Warsaw Bioethics Committee (permission number KB/50/2022). Informed consent was obtained from the donor’s legal guardian.

Supplementary information

The online version contains supplementary material available at 10.1038/s41375-026-03031-z.