Extracellular vesicle cargo dynamics in the bone marrow microenvironment: from hematopoietic homeostasis to malignant transformation

Abstract

The bone marrow (BM) microenvironment relies on extracellular vesicle (EV)-mediated communication to maintain hematopoietic homeostasis and contribute to cellular responses in malignant transformation. EV function as molecular shuttles carrying miRNAs, proteins, and lipids that regulate hematopoietic stem cell (HSC) self-renewal, quiescence, and lineage commitment. HSC-derived EVs can stimulate stem cell factor (SCF) expression in recipient HSCs through autocrine/paracrine signaling, while mesenchymal stem cell (MSC)-EVs modulate HSC differentiation via TLR4 activation and miRNA transfer. Regulated EV biogenesis pathways, involving tetraspanins, ESCRT components, and lipid-sorting mechanisms, control cargo selection and secretion in both HSCs and BM cells. During malignant transformation, EV cargo composition shifts dramatically: leukemic cells release EVs enriched in immunosuppressive factors, pro-survival signals, and drug resistance mediators that reprogram the BM microenvironment to support tumor growth. These changes—driven by hypoxia, inflammatory signaling, metabolic reprogramming, and chemotherapeutic pressure—enable tumor-derived EVs to induce HSC quiescence, polarize macrophages toward immunosuppressive phenotypes, and promote stromal cell transformation. The distinct protein and miRNA profiles of EVs from malignant versus healthy cells offer diagnostic and prognostic value, positioning EVs as both biomarkers and therapeutic targets. This review examines EV cargo composition and functional roles in normal and malignant hematopoiesis, emphasizing dynamic changes that accompany disease progression and their clinical implications.

1. Introduction

Hematopoiesis is the process that generates all types of blood and immune cells throughout life, originating from hematopoietic stem cells (HSCs).¹ The balance between self-renewal and differentiation of the HSC directly impacts hematopoietic homeostasis.² This process is maintained and regulated by numerous factors including the bone marrow (BM) microenvironment, also called “niche,” which is a dynamic organ composed of many cellular players besides HSCs, including their progeny, endothelial cells, and cells originating from mesenchymal stem cells (MSCs) such as osteoblasts and adipocytes.^3–8 Under both physiological and pathological conditions, all of the aforementioned cells affect and transform the BM microenvironment through cell-to-cell communication.⁹ This communication is driven by the release of chemical signals, such as proteins, small molecules, and ions, from the donor cells, which are recognized by receptors on the recipient cells. The exchange of these mediators occurs via four mechanisms: cytoplasmic bridges, direct cell-cell interactions through membrane proteins, secreted molecules, and transfer of extracellular vesicles (EVs).¹⁰ Nowadays, EV communication represents a promising field of research both for their involvement in healthy/malignant conditions and their potential clinical application.

EVs are membranous vesicles released into the extracellular space, containing bioactive molecules, including nucleic acids, lipids, metabolites, and proteins.¹¹ EVs are classified by origin and size into i) microvesicles (100–1000 nm), formed by plasma membrane budding; ii) exosomes or small EVs (30–150 nm), generated within multivesicular bodies (MVB) containing premature EVs in the form of intraluminal vesicle and released upon MVB fusion with the plasma membrane¹²; and iii) apoptotic bodies (50–5000 nm), produced during programmed cell death¹³ (Fig. 1A). EVs are efficient mediators of intercellular communication by releasing their content into target cells through three main mechanisms—endocytosis, EV-plasma membrane fusion and receptor interaction—thereby influencing biological processes across multiple levels and affecting cell fate^14,15 (Fig. 1B). Importantly, the cargo composition of EVs can critically influence how they modulate recipient cell behavior and determine functional outcomes. EVs are present in nearly all body fluids, exhibit high stability, and are easier to store and manipulate than parent cells, making them attractive candidates for biomedical applications, including biomarkers identification for disease prognosis and diagnosis.¹⁶ EV-based strategies have entered several clinical trials since the late 20th century,^17–21 with clinical applications in hematological disorders or malignancies anticipated. EVs can be engineered to carry specific cargoes using bioengineering techniques, enabling delivery of chemical drugs, therapeutic proteins, and nucleic acids, owing to their protection of cargoes from nuclease and protease degradation.^22,23 In hematological malignancies, EVs hold promise as a non-invasive “liquid biopsy,” providing molecular profiling for disease stratification and monitoring.²⁴

Fig. 1.

Extracellular vesicle (EV) biogenesis, release, and uptake mechanisms. (A) Classification of EVs by size and origin: apoptotic bodies (50–5000 nm), microvesicles (100–1000 nm) generated by evagination of the cell membrane, and small EVs (30–150 nm) generated by the endosomal compartment. Intracellular biogenesis and secretion of EVs showing the endosomal pathway, where early endosomes mature into multivesicular bodies (MVBs) containing intraluminal vesicles (ILVs). Rab proteins facilitate MVB trafficking and fusion with the plasma membrane, resulting in the release of EVs into the extracellular space. (B) Principal uptake mechanisms and internalization of EVs by the recipient cell: (I) endocytosis; (II) direct fusion with the plasma membrane; (III) receptor-mediated interaction at the cell surface, enabling intercellular communication and cargo transfer. Following internalization, EVs may follow several fates: (1) recycling to the plasma membrane for re-secretion or shedding of membrane components; (2) fusion with mitochondria to release mitochondrial mediators; (3) fusion with the endoplasmic reticulum (ER); (4) fusion with a lysosome for degradation. Post-fusion with the plasma membrane, EV content can be digested in a lysosome (5) or processed in the cytosol (6). Receptor interaction between EV and the plasma membrane triggers downstream signaling in the recipient cell. ER, endoplasmic reticulum. Image created with BioRender.

EVs derived from BM-resident cells mediate diverse biological activities, including HSC differentiation, homing, immune modulation, and support of allogeneic transplantation,^25,26 yet how these functions are coordinated among different niche populations, and how EV cargo specificity is achieved, remain open questions. Recent findings from our laboratory illustrate this complexity: HSC-derived EVs upregulate stem cell factor (SCF) expression specifically in recipient HSCs through an autocrine and/or paracrine mechanism, enhancing HSC function without lineage bias.²⁷ This observation underscores that EV-mediated signaling in the BM operates through cell-type-specific mechanisms that warrant systematic characterization.

This review examines EV-mediated communication in the BM microenvironment through three integrated lenses. First, we describe the molecular machinery governing EV biogenesis in HSCs and stromal cells, highlighting how regulated pathways control cargo selection. Second, we characterize EV cargo—including miRNAs and proteins—and their roles in maintaining HSC quiescence, differentiation, and mobilization under physiological conditions. Third, we examine how EV composition and function become dysregulated in hematological malignancies, focusing on the dynamic changes that promote immune evasion, drug resistance, and microenvironmental reprogramming. Throughout, we emphasize the clinical potential of EVs as biomarkers and therapeutic targets.

2. BM-derived EV-mediated cell-to-cell communication in normal hematopoiesis

Intercellular communication within the BM plays a fundamental role in healthy hematopoiesis. EVs have been implicated in regulating key HSC properties, including self-renewal,²⁸ differentiation, quiescence, apoptosis, and mobilization, through interactions among multiple BM-resident cell populations (Fig. 2).^29–31

Fig. 2.

Graphical representation of the extracellular vesicle (EV) involvement in the physiological properties of hematopoietic stem cells (HSCs) in the bone marrow (BM) microenvironment that support healthy hematopoiesis. (a) Autocrine EV signaling supports HSC self-renewal capacity through upregulation of stem cell factor (SCF). (b) EVs participate in the multilineage differentiation of hematopoietic stem and progenitor cells; megakaryocytes enhance their own expansion via EVs; hypoxia induces erythroblastic EV release, promoting differentiation along the erythroid lineage. (c) Mesenchymal stem cell (MSC)-derived EVs decrease apoptosis of HSCs, helping maintain homeostasis in the BM microenvironment. (d) Various stromal cells release EVs under G-CSF (granulocyte colony-stimulating factor) stimulation, promoting trafficking and mobilization of HSCs from BM to the circulation. (e) MSC-derived EVs can induce HSC quiescence via upregulation of Egr1. Image created with BioRender.

Among these, self-renewal has been most extensively characterized. Vacuolar protein sorting-associated protein 33B (VPS33B) is a protein involved in EV maturation and secretion, and Gu H et al. demonstrated that Vps33b KO mice (a tamoxifen-inducible conditional Vps33b mouse model) showed significantly impaired repopulation capacity, confirming the critical role of EV-mediated signaling in HSC biology.³² Loss of VPS33B in CD34⁺ HSPCs from human umbilical cord blood resulted in reduced HSPC numbers, disrupted quiescence, and elevated apoptotic cell death, providing additional evidence for the fundamental role of EV-mediated signaling in HSC physiology³³ (Fig. 2-a). Treatment of HSPCs with GW4869, a widely used inhibitor of EVs generation, markedly reduces their self-renewal capacity in vitro. Additionally, Rab27a silencing - targeting the docking of multivesicular endosome at the cell membrane and thereby impairing EV secretion - leads to a loss of reconstitution potential.²⁷

EVs are also involved in the HSPC differentiation process. For instance, megakaryocyte (MK)-derived EVs can induce HSC differentiation into new MKs by transferring MK RNA upon internalization, facilitated by interactions with intercellular adhesion molecule 1 (ICAM-1), CD43, CD18 and CD11b epitopes.^29,34 Hypoxic conditions also modulate EV secretion to support erythroid differentiation²⁹ (Fig. 2-b). Additionally, MSC-derived EVs through toll-like receptor 4 (TLR4) signal can induce myeloid-biased expansion of the HSCs interfering with the multipotent progenitor (MPP) expansion.³⁵ Regarding apoptosis in HSCs, injections of MSC-EVs have been shown to restore murine HSC engraftment capacity after irradiation by decreasing programmed cell death³⁶ (Fig. 2-c). EVs influence HSC mobilization in response to signals such as granulocyte colony-stimulating factor (G-CSF)^29,37,38 (Fig. 2-d). Finally, MSC-derived EVs positively affect HSC quiescence capacity through the Egr1/Cdkn1a axis to promote the expansion of long-term HSC (LT-HSC) (Fig. 2-e).³⁹ MSC-derived EVs also support BM microenvironment integrity by promoting angiogenesis.^40,41 Under hypoxic culture, MSCs secrete EVs with enhanced Jagged-1 content that stimulates Notch signaling in HSCs, promoting self-renewal capacity, quiescence maintenance, and clonogenic potential.⁴²

Extensive research has demonstrated the crucial function of EVs in preserving HSC characteristics, including studies on embryonic stem cells (ESCs) and MSCs. EVs function as molecular shuttles, delivering bioactive factors to stem cells, progenitor cells, and terminally differentiated cells.⁴³ HSPCs respond to EV-mediated intercellular signaling (reviewed in²⁸). For example, ESC-derived EVs can transfer protein or nucleic cargo to HSPCs in co-culture systems, improving cell viability, expanding progenitor populations, and upregulating HSC-specific markers.⁴⁴ Similar effects have been reported for EVs from BM niche cell populations, particularly MSCs. Cells within the BM microenvironment secrete numerous paracrine factors – such as stem cell factor, thrombopoietin (TPO), WNT signaling molecules, tumor growth factor β (TGF-β), angiopoietin 1, fibroblast growth factor 1/2 (FGF1/2), and angiopoietin-like proteins (ANGPTLs) – which play essential roles in maintaining HSC stemness properties (reviewed in⁴⁵). Several of these proteins, including TPO, ANGPTL2, and ANGPTL3, have been detected within purified EV preparations from cell culture supernatants and human plasma samples.³² Furthermore, findings by Hurwitz et al. demonstrated that brief ex vivo treatment of HSPCs with GW4869 (a neutral sphingomyelinase inhibitor commonly employed to suppress EV production⁴⁶ enhanced in vivo repopulation capacity by promoting multipotent progenitor cell expansion, despite inducing preferential myeloid lineage differentiation.⁴⁷

Emerging evidence highlights the diversity of EV-mediated crosstalk beyond classical niche interactions: BM adipocyte-derived EVs can transfer adipogenic transcripts such as PPARγ and C/EBPα to osteoblasts, shifting their gene expression toward adipogenesis at the expense of osteogenesis,⁴⁸ while skeletal muscle cell-derived EVs have been shown to promote osteogenic differentiation of BM-MSCs and inhibit monocyte-derived osteoclast formation.^49,50 Whether adipocyte-derived EVs directly influence HSC function remains unexplored, and a full characterization of EVs from all resident BM cell populations and their interactions is an important open challenge, underscoring the need for systematic investigation of EV-mediated communication within the BM microenvironment.

The diverse functional effects of EVs on HSC biology described above raise fundamental questions about how cargo specificity is achieved. The selective incorporation of proteins, RNAs, and lipids into EVs depends on highly regulated biogenesis pathways that control both vesicle formation and cargo sorting. We next examine these molecular mechanisms, which are essential for understanding how different BM cell populations generate functionally distinct EV populations.

2.1. Regulated pathways of EV production in hematopoietic stem cells and bone marrow stroma

EV formation is initiated by the clustering of specific integral membrane proteins into microdomains that organize the local membrane environment. In microvesicles, these microdomains form at the plasma membrane, whereas in small EVs they are established on the limiting membrane of multivesicular bodies (MVBs). This organization is facilitated by tetraspanins, a family of four-pass transmembrane proteins that includes CD9, CD53, CD63, CD81, and CD82.⁵¹ Tetraspanins act as scaffolds that assemble tetraspanin-enriched microdomains (TEM), contributing to the selective loading of proteins, lipids, and nucleic acids into EVs. In HSCs, CD53 and CD63 were originally identified as markers of asymmetric segregation.⁵² More recent studies have demonstrated that both proteins are enriched in HSCs and that their deletion leads to loss of quiescence and HSC exhaustion.^53,54 Similarly, CD82 is highly expressed in HSCs, where it supports the maintenance of quiescence, at least in part by supporting TGF-β signaling.^55,56 CD9 is enriched in human cord blood–derived HSCs, where it supports homing and repopulating capacity⁵⁷; in adult murine HSCs, high CD9 expression is associated with priming toward megakaryopoiesis.⁵⁸

TEM platforms recruit the molecular machinery required to induce membrane budding, typically characterized by an outward curvature relative to cytoplasm. Membrane budding is mediated, at least in part, by components of the endosomal sorting complexes required for transport (ESCRT). Although the full repertoire of ESCRT components involved in EV formation is not yet completely defined, key contributors include ALG-2–interacting protein X (ALIX) and tumor susceptibility gene 101 (TSG101).⁵⁹ ALIX has been proposed to recruit syndecan heparan sulfate proteoglycans and their cytoplasmic adaptor syntenin, as well as to interact with TSG101, thereby coordinating ESCRT-dependent EV formation.¹⁵ The contribution of ESCRT-dependent pathways to HSC biology is still not fully understood. Notably, syndecan-2 is highly expressed in HSCs, and its deletion significantly impairs repopulation capacity. Interestingly, syndecan-2 appears required for the maintenance of quiescence in HSCs, similar to CD53, CD63, and CD82.⁶⁰ Moreover, deletion of syndecan-2 in BM-MSCs also impairs the repopulation capacity of wild-type HSC, suggesting that ESCRT-dependent EV biogenesis supports bidirectional communication between HSCs and BM-MSCs.⁶¹ Intriguingly, HSCs lacking TSG101 display improved long-term reconstitution capacity in vivo. TSG101 plays a role in cargo selection,⁶² and its deletion in HSCs leads to the accumulation of mitochondrial proteins and mtDNA in EVs,⁴⁷ clearly indicating that cargo selectivity is as critical as biogenesis itself for EV-mediated regulation of HSC function.

Microvesicles are pinched off directly from the plasma membrane with the involvement of ARF6 and ARF1, which promote phosphorylation of myosin light chain (MLC). This phosphorylation triggers actomyosin contraction, facilitating scission of the nascent vesicle.^63,64 Currently, there is no direct evidence that microvesicles exert distinct functional effects compared with small EVs in mediating communication between HSCs and their niche. However, knock-in of inactivating mutations in ARAP3, a natural inhibitor of ARF6, impairs HSC self-renewal in a non–cell-autonomous manner.⁶⁵ This observation raises the possibility that microvesicles participate in negative-feedback signaling in the HSC niche and highlights the need for a more refined characterization of EV subtypes in bone-marrow communication.

In contrast, vesicles budding into the lumen of MVBs give rise to intraluminal vesicles (ILVs). MVBs then associate with microtubules and are transported toward the plasma membrane through the activity of several RAB GTPases, most notably RAB7, RAB27a, and RAB27b. Our group previously demonstrated that the low oxidative metabolism of HSCs favors cholesterol accumulation and EV biogenesis, a process that can be completely abrogated by down-regulation of RAB27a.²⁷ Consistent with this, the recruitment of cholesterol into ILVs—a cargo-sorting mechanism involving CD63⁶⁶—promotes their secretion pathway over lysosomal degradation.⁶⁷ Beyond cholesterol, ceramide represents a second lipid with a functionally important role in EV biogenesis and cargo regulation in HSPCs. Ceramide-dependent EV secretion, driven by neutral sphingomyelinase-2 (nSMase2), acts as a negative regulator of HSPC repopulation. Its blockade not only reduces EV output but selectively depletes from EVs proteins involved in myeloid differentiation and immune activation, ultimately enhancing long-term reconstitution capacity after transplantation.⁴⁷ Thus, lipid composition is not merely a structural feature of BM-derived EVs; rather, it actively shapes cargo selectivity and hematopoietic function.

Overall, these findings support the existence of evolutionary conserved and actively regulated mechanisms governing EV biogenesis in HSCs and stromal bone-marrow cells. They provide a critical framework for a detailed understanding of EV-based communication within the BM niche.

2.2. EV-derived miRNAs in the healthy bone marrow microenvironment

Among cargo components, miRNAs have emerged as particularly well-characterized mediators of EV function due to their potent gene-regulatory capacity and stability within vesicles.⁶⁸ We next examine the repertoire of miRNAs found in BM-derived EVs and their roles in orchestrating hematopoietic processes.

MiRNAs are a large family of small non-coding RNAs that regulates post-transcriptional gene regulation.^69,70 Incorporation of miRNAs into EVs enables their circulation in blood while protecting them from RNase-mediated degradation. In the context of the BM microenvironment, miRNAs are involved in regulating the proliferation and differentiation of the HSCs.³⁸ Currently, only a few specific miRNAs are reported to be involved in physiological functions of the BM microenvironment that contribute to the maintenance of normal hematopoiesis (Fig. 3). miR-486, for instance, has the function of stimulating the commitment and expansion of HSPCs toward erythroblast differentiation by targeting the protein SIRT1.⁷¹ Along with miR-486, miR-144 and miR-451 have also been reported to stimulate erythroid differentiation of HSCs.^38,72 EV-derived miR-126 on the other hand, facilitates HSC mobilization by modulating adhesion molecules.^37,38 Stromal cell-derived EVs are involved in the BM microenvironment for the hematopoietic homeostasis through abundant miRNA secretion, including miR-10a (inducing HSC megakaryocyte differentiation),^73,74 miR-223 (involved in the granulopoiesis process),³⁸ miR-196b (pushing HSCs toward myelopoiesis),^75,76 and miR-424, miR-150, and miR-181, which are respectively responsible for inducing monocyte differentiation, B lymphocyte differentiation, and T lymphocyte differentiation.⁷⁷ Platelets have been reported to be capable of influencing HSC differentiation toward megakaryocyte formation via miR-1915–3p.⁷⁸ Osteoblast-derived EVs carrying miR-29a can also trigger the expansion of HSPCs.⁷⁹

Fig. 3.

HSC biological responses (dotted arrows) to EV-derived miRNAs (solid black arrows) produced by different resident cell populations in the BM microenvironment. MSC, mesenchymal stem cell: HSPC, hematopoietic stem and progenitor cell. Image created with BioRender.

Liang M et al. reported that osteoclast-derived EVs enriched with miR-324 can promote the osteogenic differentiation of stromal cells by targeting ARHGAP1, a negative regulator of osteogenic differentiation.⁸⁰ On the other hand, osteoblasts can be enriched with miR-143–3p, which inhibits osteoblast differentiation and stimulates osteoclast formation and expansion by targeting CBFB mRNA.^81,82 BM MSC-derived EVs enriched with miR-22–3p can promote osteogenic differentiation in an autocrine manner through fat mass and obesity-associated protein by inhibiting the MYC/PI3K/AKT pathway.⁸³ Additionally, specific miRNAs with established roles in HSC biology have been identified in EVs from different sources. For instance, miR-221 and miR-150, which are known to be involved in the maintenance of HSC quiescence and multipotency,^84,85 have been found in EVs derived from CD34⁺ cells in umbilical cord blood.⁸⁶ These findings strongly support the concept that EV-mediated miRNA transfer represents a crucial mechanism for intercellular communication within hematopoietic environments. Table 1 summarizes the known EV-derived miRNAs secreted by different cell populations within the BM microenvironment (Table 1).

Table 1.

miRNA EVs-derived function in bone marrow microenvironment.

Donor cell	miRNA	Target cell	Function	Reference
BM-Stromal cells under G-CSG factor	miR-126	HSCs	Trafficking and Mobilization	37,38
MSCs	miR-10a	HSCs	Megakaryocytes differentiation	De Luca et al., 2015;73
MSCs	miR-22–3p	MSCs	Osteogenic differentiation	83
MSCs	miR-183–5p	MSCs	Inducing of senescence process	87
MSCs	miR-223	HSPCs	Granulopoiesis	38
MSCs	miR-196b	HSCs	Myelopoiesis	75,76
MSCs	miR-424	Myeloid Progenitors cell	Monocytes differentiation	77
MSCs	miR-150	Myeloid Progenitors cell	B lymphocytes differentiation	77
MSCs	miR-181	Myeloid Progenitors cell	T lymphocytes differentiation	77
MSCs	miR-21	HSCs	Preserve HSCs from irradiation-induced damage; regulation of hematopoiesis	77
MSCs	miR-125–5p, miR-210	HSCs	Increased self-renewal and proliferation	Sarvar et al., 2015; Lazare et al., 2014
BM microenvironmentunder hypoxia condition	miR-486, miR-144 and miR-451	HSCs	Erythroblastic differentiation	38,72,88
Osteoclast	miR-324	MSCs	Promotion of osteogenic differentiation	81
Osteoblast	miR-29a	HSPCs	HSPCs expansion	79
Osteoblast	miR-143–3p	MSCs	Promotion of osteoclasts formation	81
BM resident Macrophage M1	miR-155	MSCs	Osteogenic inhibition	Kang M. et al., 2020
BM resident Macrophage M2	miR-5106	MSCs	Osteogenic differentiation	Xiong Y. et al., 2024
Platelets	miR-1915–3p	HSCs	Megakaryocyte differentiation	78

Abbreviation: HSC, hematopoietic stem cell; HSPCs, hematopoietic stem/progenitor cell; MSC, mesenchymal stem cell.

Although the involvement of BM EV-derived miRNAs in homeostasis has become increasingly recognized over the past decade, many questions remain regarding the complete EV-derived miRNA profile across all BM cell populations and how this communication mode is utilized, warranting further investigation.

2.3. EV-derived protein markers in the healthy bone marrow microenvironment

EV-associated proteins can be broadly divided into two functionally distinct compartments: surface-exposed proteins, accessible on the outer leaflet of the EV membrane, and intravesicular proteins, enclosed within the EV lumen. This topological distinction carries direct functional consequences. Surface proteins, including tetraspanins, integrins, adhesion molecules, and receptor ligands, mediate EV interactions with the extracellular milieu, determine target cell selectivity, and can trigger receptor-dependent signaling in recipient cells without requiring EV internalization.^10,89 In the hematopoietic context specifically, surface-exposed tetraspanins and integrins on BM-derived EVs interact with hematopoietic-specific receptors including MHC-I/II, integrin α4β1, and co-stimulatory CD2, positioning surface proteins as key determinants of niche cell targeting.²⁸ By contrast, intravesicular proteins, encompassing cytosolic chaperones, ESCRT machinery components, and transcription-regulatory factors, constitute the active cargo delivered upon EV uptake and require internalization followed by endosomal processing to exert their effects in recipient cells.⁸⁹ This distinction is particularly relevant in the BM microenvironment, where the specificity of HSC-niche communication depends not only on which proteins are packaged into EVs, but also on whether they act at the cell surface or after internalization.

EV proteins are highly enriched in cytoskeletal, cytosolic, heat shock, and plasma membrane proteins, as well as proteins responsible for vesicle trafficking and membrane fusion.¹⁰ Although no universal markers uniquely identify all EVs, researchers rely on proteins that are enriched in different EV subpopulations. These markers include four key transmembrane markers, tetraspanins (CD9, CD63, CD81, and CD82), which are among the most well-characterized surface-exposed EV markers, functioning as scaffolds for target cell recognition and EV uptake. Additional markers encompass cytosolic chaperone proteins such as 14-3-3 proteins and heat shock proteins (HSP60, HSP70, HSP90) that assist in protein folding and cellular stress responses. EVs also contain multivesicular body (MVB)-related proteins including tumor susceptibility gene 101 (TSG101), Alix protein, and VPS4, which are part of the endosomal sorting complex required for transport (ESCRT) machinery involved in vesicle biogenesis and membrane abscission.^10,90,91 Furthermore, EVs are enriched in cytoskeletal proteins (such as ACTIN, TUBULIN, MYOSIN I and II), adhesion proteins and major histocompatibility complex molecules (MHCI and MHCII), as well as membrane fusion and transport-related proteins including annexin and Rab-GTPases. Reflecting the surface/intravesicular distinction introduced above, tetraspanins, MHC molecules, adhesion proteins, and annexins represent the surface-exposed fraction mediating cell recognition and uptake, whereas HSPs, ESCRT components, and cytoskeletal proteins constitute the intravesicular compartment delivered following internalization. Notably, many proteins of cytosolic origin are associated with the EV membrane, and several membrane proteins exhibit reversed orientation compared to their typical cellular localization.⁹² The main categories of EV markers reported so far are summarized in Fig. 4.

Fig. 4.

Schematic representation of extracellular vesicle (EV) protein composition and markers. The illustration summarizes major protein categories found in EVs, organized by cellular localization and function. Key EV markers include four transmembrane tetraspanins (CD9, CD63, CD81), cytosolic chaperon proteins (14-3-3 proteins, HSP60, HSP70, HSP90), multivesicular body (MVB)-associated proteins involved in vesicle biogenesis (Alix, TGS101, VPS4), cytoskeletal proteins (actin, tubulin, myosin I and II), adhesion proteins and MHC molecules (MHCI, MHCII, ICAM-1, integrin), and membrane fusion and transport-related proteins (annexin, Rab-GTPase). Image created with BioRender.

Saunderson SC et al. reported that the CD169 protein mediates the capture of B cell-derived EVs by macrophages.⁹³ Additionally, CD59 and CD55 were reported as EV-derived markers for immune system cells.⁹⁴ Another comprehensive study was conducted in 2020 by Lyden et al. This work was able to define EV markers from BM, increasing the number of possible markers useable for collecting and characterizing EVs from BM tissue. They reported heat shock 70 KDa protein 8 (HSPA8) (also found as Hsp70), ALIX, heat shock protein 90 (HSP90-AB1/AA1), CD9, and Flotilin-2 (FLOT2) as established and previously registered markers, while adding all the following as new EV markers for BM-derived EV populations: fibronectin 1 (FN1), Galectin 3 Binding Protein (LGALS3PB), alpha-2-macroglobulin gene (A2M), protein-coding gene located on chromosome 4 (JCHAIN), human beta-globin gene (HBB), gelsolin (GSN), beta-actin (ACTB), beta 2 Microglobulin protein (B2M), stomatin (STOM), moesin (MSN), peroxiredoxin (PRDX2), ras-related protein rap-1b (RAP1B), filamin A (FLNA), CD36, CAVIN2, and shiga toxin STX1.⁹⁵ Moreover, the presence of polarized macrophage-specific EV markers such as CD68, CD86, and CD206 for M0, M1, and M2, respectively, has been confirmed in the BM.^96,97 CD31 is reported as the main marker for endothelial cells-derived EVs also in the BM microenvironment.⁹⁸ Another work focused on the study of the bone calcification process defined EV protein markers such as sclerostin (SOST) for osteocytes, collagen type I alpha 1 (COL I) for osteoblasts, and collagen alpha1X (COL X) for chondrocytes.⁹⁹ Furthermore, osteoblast-derived EVs can promote osteogenic differentiation of BM-MSCs through the attachment between annexin EV-associated proteins and mineral accumulation and nucleation sites.^81,87 Delving deeper into the hematopoietic BM-resident cells, EVs derived from short-term HSC (ST-HSC) have been shown to carry the CD34 marker.^28,78 In contrast, BM-MSC-derived EVs have been reported to interact with HSPCs via TLR4, thereby activating TLR4 signaling that promotes myeloid expansion and skews hematopoietic repopulation potential of HSCs.³⁵ Collectively, these cell-type-specific EV protein signatures provide an emerging framework for distinguishing the cellular origin of BM-derived EVs and for understanding how distinct EV populations selectively engage their target cells within the niche.

3. Dynamic remodeling of EV cargo in response to microenvironmental stresses and its role in malignant transformation

Understanding the factors that drive EV cargo alterations during disease progression provides mechanistic insights into how EVs acquire pro-tumorigenic functions and offers potential therapeutic intervention points. During malignant transformation, the quantity and quality of EV secretion change dramatically, shifting from supporting normal hematopoiesis to promoting immune evasion, drug resistance, and microenvironmental reprogramming¹⁴ in response to microenvironmental stresses that can include hypoxia, inflammation, and therapeutic pressure.¹⁰⁰

Hypoxia represents fundamental microenvironmental stress in both pre-leukemic and established malignancies, driving adaptive changes in EV production and cargo composition. While hypoxic stress can promote erythroblastic expansion through increased EV secretion,²⁹ this response may shift toward pathology: in erythroleukemia cells, hypoxia rapidly upregulates miR-486 intracellularly and in secreted EVs, and EV-mediated transfer of miR-486 to recipient healthy HSPCs promotes erythroid differentiation through SIRT1 suppression, suggesting that hypoxia-driven EV cargo remodeling can actively contribute to malignant erythropoiesis (Fig. 5, panel A).⁷¹ The hypoxic tumor microenvironment induces HIF-1α stabilization, directly regulating EV biogenesis and cargo selection, and sustained hypoxia triggers pathological EV remodeling that supports tumor progression. In leukemia, hypoxia increases EV secretion and enriches cargo with pro-angiogenic miR-210, which enhances endothelial tube formation and vascular network formation.¹⁰¹ Similarly, hypoxia-resistant multiple myeloma cells release EVs enriched with miR-135b that target factor-inhibiting HIF-1 (FIH-1) in endothelial cells, thereby activating the HIF pathway and promoting angiogenesis.¹⁰⁰ Beyond angiogenesis, hypoxia directly upregulates PD-L1 expression on tumor cell-derived EVs through HIF-1α binding to the PD-L1 promoter, facilitating immune evasion by suppressing T cell activation.¹⁰²

Fig. 5.

Dynamic remodeling of extracellular vesicle cargo composition in response to microenvironmental stressors during hematological malignant transformation. (A) Hypoxia: Low O₂/HIF-1α stress in erythroleukemia cells drives upregulation of miR-486 in both the intracellular compartment and secreted EVs. EV transfer of miR-486 to recipient HSPCs suppresses Sirt1, promoting aberrant erythropoiesis and contributing to erythroleukemic cell expansion. (B) Inflammation: TNF-α/IFN-γ stress activates RAB27B-dependent pathways and NF-κB signaling in BM-stromal cells, remodeling their EV cargo to include immunomodulatory proteins and altered miRNA profiles that prime the pre-malignant state. Once malignancy is established, these stromal EVs enhance chemoresistance, promote M2 macrophage polarization, and drive stromal reprogramming to sustain tumor growth. (C) Chemotherapy: Chemotherapy stress triggers bidirectional EV-mediated crosstalk between leukemic cells and mesenchymal stromal cells. Stromal EVs deliver anti-apoptotic proteins (MCL-1), drug efflux pumps (MDR1, MRP1), and translation initiation factors (eIF4A) alongside pro-survival factors (IL-6, Gas-6, Galectin-3) to residual leukemic cells, conferring chemoresistance and sustaining malignant cell survival and proliferation. BM, bone marrow; EV, extracellular vesicle; HIF-1α, hypoxia-inducible factor 1-alpha; HSPC, hematopoietic stem and progenitor cell; MDR1, multidrug resistance protein 1; MRP1, multidrug resistance-associated protein 1; NF-κB, nuclear factor kappa B. Image created with BioRender.

Chronic inflammation in the BM microenvironment profoundly impacts EV biogenesis and cargo composition in stromal cells. Proinflammatory cytokines such as TNF-α and IFN-γ, activate stromal cells and fundamentally alter their EV secretion patterns through upregulation of RAB27B-dependent pathways and NF-κB signaling.^103,104 This inflammatory priming of MSCs drives increased EV production and remodels their molecular cargo to include immunomodulatory proteins and altered miRNA profiles that can influence hematopoietic cell fate toward malignant transformation¹⁰⁵ (Fig. 5, panel B). Once established, the EV-mediated crosstalk is intensified in acute myeloid leukemia (AML), with stromal EVs carrying inflammatory miRNAs (miR-155, miR-375) enhancing chemoresistance and promoting invasiveness in leukemic cells.^106,107 Similarly, in chronic lymphocytic leukemia (CLL), BM-stromal cells exposed to chronic inflammatory signaling release EVs containing miR-202–3p, miR-146a, and miR-451, which reprogram the microenvironment to support tumor cell survival.^108–110 Furthermore, leukemia-derived EVs can polarize macrophages toward immunosuppressive M2 phenotypes through delivery of specific miRNA cargo, as demonstrated in chronic myeloid leukemia (CML) where tumor-derived EVs induce M2 polarization and suppress anti-tumor immunity.¹¹¹

Beyond inflammation, other clinical interventions can transiently alter the BM EV landscape. G-CSF mobilization, for instance, increases BM-derived EVs from activated neutrophils and monocytes and modifies stromal EV cargo profiles,^37,112,113 although the long-term significance of these changes for hematopoietic surveillance remains to be determined.

Chemotherapy exposure fundamentally reshapes EV cargo composition in both malignant and stromal cells.¹¹⁴ In AML, BM-MSCs respond to chemotherapy-stressed leukemic cells by releasing EVs that confer chemoresistance through multiple mechanisms: delivery of anti-apoptotic proteins (MCL-1), transfer of drug efflux pumps (MDR1, MRP1), and transmission of translation initiation factors (eIF4A) that sustain protein synthesis in residual leukemic cells.^109,115,116 Reciprocally, leukemia-derived EVs reprogram BM-MSCs to express pro-survival factors including IL-6, GAS-6, and GALECTIN-3, creating a chemotherapy-protective niche that shields minimal residual disease (MRD) (Fig. 5, panel C).^117,118 Critically, leukemia-derived EVs isolated from relapsed patients carry significantly higher levels of drug resistance mediators compared to those from newly diagnosed patients, and these “resistance EVs” can transfer chemoresistant phenotypes to treatment-naïve cells through cargo delivery of resistance genes and reactive oxygen species (ROS)-scavenging molecules.¹⁰¹ In CLL, stromal cell-derived EVs protect malignant B cells from eight different chemotherapeutic agents including fludarabine, ibrutinib, and venetoclax.¹⁰⁹ These therapy-induced EV changes create a self-reinforcing cycle where treatment itself remodels the microenvironment to favor survival of resistant clones, promoting progression to more aggressive disease.

4. The role of BM-derived EVs in cell-to-cell communication in hematological malignancies

EVs contribute to malignant transformation, drug resistance, and metastasis.^119–121 Leukemia-derived EVs can reprogram the BM microenvironment, suppressing both normal hematopoiesis and antitumor immunity, thereby fostering therapy-resistance.¹⁴ Levels of EVs in blood are elevated in many hematological malignancies compared to healthy individuals, and EV size varies across hematological diseases representing an interesting diagnosis/prognosis factor.¹²² The role of EVs in the crosstalk between tumor cells and tumor microenvironment (TME) remains incompletely understood.⁸⁸ For example, AML-derived EVs can induce quiescence in residual HSCs within a leukemic marrow, illustrating how malignant EVs disrupt normal hematopoiesis.¹²³ Growing evidence supports EVs’ capacity to mediate crosstalk between cancer cells and the surrounding TME.^31,124

Hematological malignancies represent an evolutionary and ecological process involving dynamic, reciprocal interactions between cancer cells and the TME,^125,126 which consists of immune cells, stromal cells, fibroblasts, all components belonging to the extracellular matrix, together with blood cells and lymphatic vessels.^125,127 The dynamic network and crosstalk between cancer cells and their TME allows tumor development and growth, influencing biological processes at multiple levels through various strategies. In vivo studies have observed that AML-derived EVs induce similar functional changes in the BM microenvironment compared to those observed after AML cell engraftment, particularly by targeting stromal and endothelial cells in the BM microenvironment. Furthermore, AML-derived EVs can impair the differentiation potential of BM-MSCs by enhancing the expression of negative regulators of osteoblast development (i.e., DKK-1) and suppressing normal hematopoiesis.¹²⁸ In CLL, MSC-derived EVs enhance anti-apoptotic protein expression (MCL-1, BCLXL, XIAP) and promote SDF-1-induced migration of CLL cells.¹²⁹ Additionally, multiple myeloma (MM) cell-derived EVs have been reported to be able to affect the surrounding microenvironment. MM cell-derived EVs inhibit osteoblastic differentiation and promote osteoclastogenesis through EV-mediated signaling, MM-derived EVs contain unfolded protein response (UPR) signaling molecules by activating the IRE1alpha/XBP1 axis, thereby contributing to bone resorption.¹³⁰ It has been demonstrated that acute lymphoblastic leukemia (ALL)-derived EVs can promote proliferation and survival of both leukemic and non-leukemic B cells in vitro by transferring proliferative, pro-survival, and anti-apoptotic factors.^131,132 Beyond protein and miRNA cargo, the lipid composition of leukemic EVs has emerged as a functionally relevant dimension of their pathological activity. Lipidomic profiling of ALL-derived EVs revealed enrichment in cholesterol, phosphatidylcholines, and sphingomyelins relative to EVs from non-leukemic cells; critically, it is the lipid fraction of ALL EVs, not the protein or RNA fraction alone, that directly disrupts HSPC quiescence and promotes mitochondrial activation in targeted stem cells within the BM.¹³³ These findings position EV lipid cargo as an underappreciated yet functionally active mediator of leukemia-driven niche disruption.

The complex roles of malignant EVs can be understood through three interconnected lenses: (i) their effects on immune cells that enable tumor escape, (ii) their influence on stromal cells that reshape the supportive microenvironment, and (iii) the specific cargo molecules, miRNAs and proteins, that mediate these pro-tumorigenic functions.

4.1. Effects of EVs derived from hematological malignancies cells on the immune system

The immune system plays a critical role in eliminating tumor cells. Tumor antigens, including mutated proteins or altered post-translational modifications, are exposed to enable immune recognition.¹³⁴ Loss or masking of tumor-associated antigens (TAAs) is a key immune-evasion mechanism that permits tumor growth.¹³⁵ EV-mediated communication also contributes to immune evasion (Fig. 6). Cancer-derived EVs deliver multiple immunosuppressive factors -miRNA, DNA, pro-apoptotic mediators, metabolites, and enzymes -modulating immune activation and response.^125,135,136 Leukemic-cell-derived EVs can affect CD8⁺ T cells by downregulating of T-cell receptor T3 zeta chain (CD3ζ) and tyrosine-protein kinase (JAK3) expression and promoting Fas/FasL-mediated apoptosis (Whiteside TL et al., 2013). Natural killer (NK) cells are likewise impacted by lymphoma-derived EVs that carry major histocompatibility complex (MHC), Apo2 ligand (APOL2), Fas ligand (FASL), T-cell receptor (TCR), and NK group-2-member-D (NKG2D), inhibiting NK cytotoxicity and impairing the antigen-presenting cells (APC) antigen processing.¹³⁷ CML-derived EVs polarize macrophages toward immunosuppressive M2 phenotypes¹¹¹; they can also induce malignant transformation of mononuclear cells via genome instability.¹³⁸ Specific EV-derived miRNAs can influence T cell migration; for example, leukemic-cell-released miR-363 modulates CD4⁺ T cells in CLL.¹¹⁰ Hematopoietic malignancy-derived EVs can also impair monocyte differentiation into dendritic cells.^139,140 Given these diverse mechanisms (Fig. 6), understanding EV-mediated tumor and immune cell communication may reveal therapeutically exploitable cross-talks.

Fig. 6.

Effects of EVs derived from hematological malignant cells (black arrows) on immune system cells. APC, antigen-presenting cell; MDSC, myeloid-derived suppressor cell. Image created with BioRender.

Additional examples of how hematological malignancies cells-derived EVs affect the immune cells are reported in Table 2.

Table 2.

Effects of hematological malignancies cells-derived EVs on immune system cells.

Hematological Malignancies	Donor cells	Target cells	Mechanism	References
AML	Leukemic cells	CD8(+)T-cells	CD3ζ and JAK3 down-regulation with the promotions of Fas/FasL-mediated T-cell apoptosis	Whiteside TL. et al., 2013
CLL	Leukemic cells	CD4(+)T-cells	CD69 down-regulation with the inhibition of T-cell migration	110
DLBCL	Lymphoma cells	T-cells	T-cell apoptosis increasing inducing PD-1 expression	Cheng Z. et al., 2018
DLBCL	Lymphoma cells	Macrophages	Stimulation of NK-cells inflammation process through transferring MyD88	Manček-Keber M. et al., 2018
BCL	Lymphoma cells	T-cells	CD39 and CD73 surface marker to induce ATP hydrolyzes and inhibit T-cell proliferation	Ghiringhelli F. et al., 2018
BCL	Lymphoma cells	NK-cells, APCs	Carries MHC, APO2L, FASL, TCR, and NKG2D and inhibits NK-cells cytotoxicity and antigen processing of APCs	137
MM	Myeloma cells	MDSCs	Triggering growth and immunosuppressive activity	Wang J. et al., 2016; Wang J. et al., 2016
CML	Lymphoma cells	Macrophages	Macrophages M2 Polarization through TNF-α and IL-10 expression to inhibiting NO and ROS generation	111
CML	Leukemic cells (BCR-ABL +)	Mononuclear cells	EVs induce genome instability pushing toward a malignant transformation	138
Hematopoietic malignancy	/	Monocytes	Suppression of differentiation capacity in dendritic cells	Almeida et al., 2024).
Lymphoma	Lymphoma cells	MDSCs	HSP72 carrying with the promotion of suppressive functions	Chalmin F. et al., 2010

Abbreviation: AML, Acute myeloid leukemia; CLL, Chronic lymphocytic leukemia; DLBCL, Diffuse large B cell lymphoma; BCL, B-cell lymphoma; MM, multiple myeloma; CML, Chronic myeloid leukemia; MDSC, Myeloid derived suppressor cells.

4.2. Effects of hematological malignancy cell-derived EVs on stromal cells

Beyond suppressing antitumor immunity, malignant EVs also target non-hematopoietic cells within the BM microenvironment to create a supportive niche that sustains malignant cell survival, promotes angiogenesis, and contributes to therapy resistance (Table 3, Fig. 7). Several study groups have investigated specific interactions between tumor cells and stromal cells. In CLL, stromal cells can be affected by leukemia cell-derived EVs through the release of VEGF, promoting tumor survival and inducing cancer-associated fibroblast (CAF) transformation, which supports metastasis and angiogenesis.^{108,109,141–143} ALL-derived EVs induce metabolic reprogramming in stromal cells, contributing to chemotherapy resistance.¹⁴⁴ Conversely, stromal cell-derived EVs can influence ALL cells by inducing galectin-3 release, which promotes drug resistance,¹⁴⁵ or by enhancing proliferation, migration, and survival in multiple myeloma (MM) cells.¹⁴⁶ MM-derived EVs also promote endothelial tube formation and vascular remodeling through delivery of miR-135b and STAT3 pathway activation.^100,146 Within the BM niche, MM-derived EVs support osteoclast growth and migration.¹⁴⁷ Conversely, MM cells themselves respond to fibroblast-derived EVs, which foster their development.¹⁴⁸ Generally, leukemic cells modulate MSCs through EV-derived miRNA delivery, impacting their proliferation to support tumor development.^115,149,150 Myelodysplastic syndrome and AML cell-derived EVs can disrupt the hematopoiesis-supporting microenvironment by suppressing MSC osteo-lineage differentiation, a process mediated in part by EV-derived miR-128–3p and miR-574–5p¹⁵¹ (Fig. 7).

Table 3.

Hematological malignancies-derived EVs effect on stromal cells.

Hematological Malignancies	Donor cell	Target cells	Mechanism	References
CLL	Leukemic cells	SC	VEGF releasing to promote tumor survival	Boysen J. et al., 2019
CLL	Leukemic cells	SC	SC conversion in CAFs to induce metastasis and angiogenesis	108,109
ALL	Leukemic cells	SC	Induce of aerobic glycolysis for cancer progression	144
ALL	SC	Leukemic cells	SC-EVs trigger GAL3 expression in tumor cells inducing drug resistance development	145
AML/MDS	Leukemic/myelodysplastic cells	MSC	Inhibition of HSC support by suppressing MSC osteolineage differentiation	151
MM	Myeloma cells	EC	Endothelial STAT3 signaling activation promoting tumor angiogenesis and growth	152
MM	Myeloma cells	EC	Tumor-derived miR-135b targets HIF-1 promoting angiogenesis	100
MM	Myeloma cells	Osteoclast	MM-EVs support osteoclast migration and growth	147
ATL	Tumor cells	MSC	miR-155/21 transfer to induce MSC proliferation	149
TCL, CML	Leukemic cells	Fibroblast	Fibroblast developing in tumor-like phenotype	142,143
CML	Leukemic cells	MSC	Induce releasing of IL-8 from MSC leading to the migration of tumor cells	115

Abbreviations: CLL, Chronic lymphocytic leukemia; AML, Acute myeloid leukemia; MM, Multiple myeloma; ATL, Adult T-cell leukemia/lymphoma; TCL, T-cell leukemia; CML, Chronic myeloid leukemia; ALL, Acute lymphocytic leukemia; MDS, Myelodysplastic syndrome; SC, Stromal cells; EC, Endothelial cells; MSC, mesenchymal stem cell; CAF, cancer-associated fibroblast; GAL3, Galectin-3.

Fig. 7.

Effect of EVs derived from hematological malignant cells (black arrows) on stromal cells and the responses of stromal cell-derived EVs (dotted green arrows); MSC, mesenchymal stem cell; CAF, cancer-associated fibroblast; VEGF, vascular endothelial growth factor. Image created with BioRender.

4.3. EV-derived miRNAs in hematological malignancies

The panel of miRNAs is among the best-characterized mediators of the detrimental effects of EVs in hematological malignancies. These dysregulated miRNAs drive tumor progression, drug resistance, and microenvironmental reprogramming.

In AML, stromal EVs enriched in miR-155 and miR-375 promote chemoresistance and tumor invasiveness,¹⁰⁶ while miR-375 release enhances AML cell proliferation and mobilization, enhancing tumor invasiveness.¹⁰⁷ AML-derived EVs also affect the niche through miR-150 and miR-7977, disrupting normal hematopoietic support.^150,153 EV-derived miR-150 plays a different role in ALL, facilitating tumor development by blocking B-cell maturation from pro-B to pre-B cells.¹³¹ In CLL, EVs carry miR-202–3p, miR-146a, miR-451, and miR-363, which modulate stromal and immune cells.^108–110 MM cell-derived EVs affect endothelial cells, enhancing angiogenesis through miR-340 and miR-135.^100,154

CML-derived EVs can modulate endothelial cells through miR-126 and miR-210, promoting motility and angiogenesis.^101,155 In MDS, EV-derived miRNAs can also affect the BM microenvironment. miR-10a in MDS cells-derived EVs alters HSC viability and clonogenicity, impairing normal hematopoiesis.¹⁵⁶ Additional examples of EV-derived miRNA effects in hematological malignancies are summarized in Table 4.

Table 4.

Hematological malignancies-derived EVs miRNAs content effect on stromal and immune system cells.

Donor cell	miRNA	Target cell	Function	Reference
CML	miR-126	Endothelial cells	Regulation of adhesion and motility	155
CML	miR-210	Epithelial cells	Decreasing of angiogenesis	101
MM	miR-340	Endothelial cells	Enhancement of angiogenesis	154
MM	miR-135	Endothelial cells	Increasing angiogenesis	100
MM	miR-let-7c	Macrophages	Enhance of the M2 phenotype polarization	157
Bone marrow stromal cells	miR-155	AML cells	Chemoresistance	106
DLBCL	miR-99a, miR-125b	/	Drug resistance, prognosis factors	Feng Y. et al., 2019
Leukemia cell line K562	miR-92a	Endothelial cells	Migration and tube formation	Umezu T et al., 2013
MDS	miR-10a	CD34+ progenitor cells	Changing in HSCs viability and clonogenicity	156
AML/MDS	miR-7977	MSCs	Decreasing MSCs support to hematopoiesis	150
AML	miR-150	HSCs/HPCs	Mobilization of HSCs in the peripheral blood	153
AML	miR-375	Stromal cells	Increased proliferation, migration and invasion	107
AML	miR-34a-5p	MSCs	Inhibition of osteogenic differentiation	158
AML/MM	miR-125b	CD34(+) T-cell	Tumorigenesis	110
CLL	miR-202–3p	Stromal cells	Tumor suppressor and differentiation trigger	108
CLL	miR-146a	Stromal cells	Negative regulation of innate immune activation	109
CLL	miR-451	Stromal cells	Apoptosis and invasion	109
CLL	miR-363	CD34(+)T-cell	Migration	110
MSCs	miR-15a	MDS	Proliferation, apoptosis, clonogenicity capacity	Muntion S. et al., 2016
MSCs	miR-26a-5p	AML	Increased proliferation, migration and invasion	Barrera-Ramirez J. et al., 2017
MSCs	miR-101	AML	Enhance tumor suppressor activity and increasing of proliferation	Barrera-Ramirez J. et al., 2017
MSCs	miR-23b-5p	AML	Enhance tumor suppressor activity and increasing of proliferation	Cheng H. et al., 2021
MSCs	miR-339–3p	AML	Proliferation and apoptosis	Barrera-Ramirez J. et al., 2017
MSCs	miR-425–5p	AML	Inhibition apoptosis, proliferation migration and invasion	Zhang L. et al., 2021
NK cells	miR-155	AML	positive regulation of immunosuppression function	Otegbeye F. et al., 2018
ALL	miR-150	B-cells	Block of differentiation from pro-B to pre-B stage, arresting the maturation of B cells	131

Abbreviations: AML, Acute myeloid leukemia; MM, Multiple myeloma; DLBCL, Diffuse large B cell lymphoma; MDS, Myelodysplastic syndrome; CML, Chronic myeloid leukemia; CLL, chronic lymphocytic leukemia; ALL, acute lymphoblastic leukemia; MSCs, Mesenchymal stem cells.

4.4. EV-derived protein markers in hematological malignancies

EVs carry surface markers that help tumor cells evade immune surveillance. The surface exposure of these proteins is functionally critical: by acting directly on immune cell receptors without requiring internalization, surface-displayed immunosuppressive molecules such as PD-L1 can suppress T cell activation and bypass immune checkpoints in a contact-dependent manner.^89,159 This contrasts with intravesicular pro-survival and resistance proteins, such as MCL-1, BCLXL, and drug efflux pumps, which must be delivered into recipient cells following EV uptake to exert their effects. Understanding which proteins act at the surface versus intravesicularly therefore has direct implications for how EV-mediated immune evasion and drug resistance are mechanistically targeted.

Interestingly, EV markers can be used to identify specific types of hematological malignancies. AML patients exhibit increased levels of EVs compared to healthy individuals, and these vesicles carry blast-associated markers such as CD34, CD33, and CD117. These surface proteins could have potential utility in early diagnosis or in improving disease prognosis. Indeed, it has been shown that these proteins can impair NK cell cytotoxicity, potentially weakening immune surveillance and promoting disease progression.^136,160 Again, in leukemia of lymphoid origin, CD19 exposed on circulating EVs can support the identification of CLL and ALL origin cells.¹⁶¹ B-precursor cells (BCP) present CD19 within membrane lipid rafts. Johnson et al. observed that diagnostic samples enriched in BCP-ALL cells contained CD19-positive vesicles, confirming their leukemic origin.¹⁴⁴ Saidu et al. used a proteomic approach to identify a core protein signature of BCP-ALL-derived EVs, which may guide the development of EV biomarkers in acute leukemias. Notably, candidates include CD317, CD38, Insulin-like growth factor 2 mRNA-binding protein 1 (IGF2BP1), proliferating cell nuclear antigen (PCNA), cold shock domain-containing protein E1 (CSDE1), and G-protein coupled receptor 116 (GPR116), consistently detected in plasma and leukemic blasts.¹⁶² Numerous studies show that EVs are enriched in specific protein markers distinguishing healthy samples from hematological malignancies. In MM, for example, EVs overexpress CD55, CD59, and CD147, as well as CD38, which is targeted by daratumumab (DARA) and PD-L1 is elevated in MM EVs versus normal.⁹⁴ B-cell lymphoma-derived EVs carry CD20 and can contribute to therapeutic resistance.¹⁶³ Collectively, EV protein content yields clinically relevant proteomic markers that distinguish malignancies from healthy cells, underscoring the need for future investigation.

The comprehensive characterization of EV cargo in health and disease naturally raises the question of how this knowledge can be translated to patient benefit. The unique properties of EVs—their stability, cargo-carrying capacity, and natural role in intercellular communication—position them as promising tools for both diagnostic and therapeutic applications in hematological disorders.

5. Clinical applications, therapeutic potential and challenges of BM-derived EVs

Accumulated evidence shows that EVs actively shape BM microenvironment dynamics in health and disease, catalyzing efforts to translate this knowledge into clinical applications. BM-derived EVs hold promise as non-invasive biomarkers for detection and monitoring, as therapeutic agents leveraging regenerative properties, and as engineered drug delivery vehicles.¹⁶⁴ Realizing this potential requires addressing substantial technical and regulatory challenges from bench to bedside.

5.1. EVs as liquid biopsy tool

The molecular composition of circulating EVs reflects the cellular states and processes within the BM, positioning them as accessible liquid biopsy tools that can spare invasive BM aspirations. Compared with conventional BM biopsy, EV-based liquid biopsy offers several advantages: it is non-invasive, captures tumor heterogeneity more broadly, can be serially repeated across multiple time points, and EVs are more abundant and stable in peripheral blood than circulating tumor cells or cell-free DNA.¹⁶⁵

This diagnostic utility spans multiple hematological malignancies. In AML, elevated levels of CD34⁺/CD33⁺/CD117⁺ EVs in peripheral blood correlate with disease burden and may detect MRD with greater sensitivity than conventional flow cytometry.¹³⁶ Recent validations with larger cohorts have confirmed the clinical utility of EV-based MRD detection with improved sensitivity for low-level disease versus standard flow cytometry. The size distribution and concentration of circulating EVs also vary across disease subtypes, with specific patterns associated with disease subtype and prognosis.¹²² In AML, circulating EVs carrying specific miRNA signatures—miR-10b, miR-125b, and a panel of five upregulated serum EV miRNAs (miR-10a-5p, miR-155–5p, miR-100–5p, miR-146b-5p, and let-7a-5p)—have been identified as promising non-invasive prognostic/diagnostic biomarkers, with miR-10b levels in particular predicting poor patient outcomes.¹⁶⁶ Beyond miRNAs, surface marker profiling of AML-derived EVs (CD9, CD34, CD117, CD123, and CD135) plus lncRNA signatures (decreased LINC00265, LINC00467, and UCA1 and increased SNHG1 in plasma EVs) offers a multiparameter liquid biopsy framework for AML detection and monitoring, avoiding invasive BM aspiration.¹⁶⁶

MicroRNA profiling of circulating EVs has emerged as a promising biomarker strategy. In CLL, EV-derived miR-155, miR-146a, and miR-21 levels correlate with disease stage, treatment response, and overall survival.¹⁶ When targeted therapies like BTK inhibitors and BCL2 inhibitors are used together to both inhibit B-cell proliferation and induce apoptosis, EV-based prognostic signatures have been further refined to predict treatment response and progression-free survival.¹⁵⁷

In MM, EV numbers are significantly elevated compared to healthy volunteers, suggesting diagnostic utility at early disease stages.¹²² CD38⁺ EVs are more abundant in BM plasma of MM patients at diagnosis versus pre-active MM conditions like monoclonal gammopathy of undetermined significance (MGUS) and smoldering MM, and positively correlate with International Staging System stage.¹⁶⁷ Protein markers including CD38, the target of daratumumab, and PD-L1 add insight into treatment susceptibility and immune evasion.⁹⁴ CD138⁺ EVs rise in MM, increase with disease progression, and decrease in response to therapy, supporting longitudinal treatment monitoring.^168,169 Extending this multiparameter approach to B-cell precursor ALL, the proteomic signature of BCP-ALL-derived EVs, including CD317, CD38, IGF2BP1, PCNA, CSDE1, and GPR116, offers a framework for disease identification and monitoring.¹⁶² Together, these findings across hematological malignancies establish EVs as a broadly applicable liquid biopsy platform, with the common thread being that EV cargo composition dynamically mirrors disease state and treatment response.

5.2. MSC-derived EVs as therapeutic agents

Beyond their diagnostic utility, EVs, especially those derived from MSCs, have attracted considerable interest as therapeutic agents.¹⁷⁰ MSC-derived EVs capture many therapeutic properties of parent MSCs, while offering practical advantages: reduced immunogenicity, easier storage and handling. Comparative studies have demonstrated that MSC-EVs can achieve similar therapeutic outcomes to intact MSCs in inflammatory and immune-mediated conditions, with a lower risk of adverse events.¹⁷¹ In hematopoietic recovery, preclinical studies indicate MSC-EVs can restore HSC engraftment capacity after irradiation by reducing apoptosis and promoting quiescence through the Egr1/Cdkn1a axis.^36,39 In graft-versus-host disease (GVHD) models, MSC-EVs modulate immune responses by delivering immunoregulatory miRNAs and proteins that suppress T cell activation and promote regulatory T cell expansion. Murine GVHD studies have shown that MSC-EV administration reduces disease severity, improves survival, and crucially preserves graft-versus-leukemia effects.¹⁷²

5.3. EV therapy challenges

Despite this preclinical promise, translating EVs to routine clinical use faces interconnected challenges in production, standardization, delivery, and safety. Producing at therapeutic scale remains a major bottleneck. Lab-scale methods yield less than 1 μg of EVs protein per mL of culture medium, and standard isolation techniques such as ultracentrifugation are costly, time-consuming, and prone to co-isolation of protein aggregates and other contaminants.^173,174 Emerging approaches including tangential flow filtration, size-exclusion chromatography, and microfluidics improve purity and scalability, even if each brings tradeoffs in yield, cost, and infrastructure requirements, and none has yet been validated at GMP-compliant scale for hematological indications.^175,176

EVs cargo composition is not fixed; it varies with MSC passage number, glucose concentration, and antibiotic use, causing batch-to-batch variability that complicates characterization and dosing.¹⁷⁵ 3D bioreactor cultures show promise for high yield and consistency while remaining under development.¹⁷⁷ Storage and stability add further hurdles: EVs are vulnerable to aggregation and degradation, and long-term cryopreservation protocols that preserve cargo without compromising function are still being refined.^178,179

Targeted delivery to hematopoietic cells within the BM microenvironment remains technically challenging. Unmodified EVs are rapidly cleared by the mononuclear phagocyte system, limiting in vivo half-life.¹⁸⁰ Surface engineering—ligand conjugation and lipid membrane modification—is being explored to improve homing efficiency; however, these modifications introduce additional regulatory complexity.^181,182 Using biomaterial scaffolds like hydrogels can prolong local retention at injury sites, though biocompatibility and degradation kinetics require further evaluation.¹⁸³

Safety considerations are nuanced. While MSC-EVs lack the tumorigenicity concerns of chromosomally abnormal live-cell transfers, residual cytokines and engineered surface proteins may provoke immune responses.¹⁸⁴ A more troubling point, as noted above, is that EVs derived from leukemic MSCs can actively promote disease progression and drug resistance, so therapeutic EV preparations must be rigorously validated for MSC-source integrity. Long-term toxicological data are limited, and robust preclinical models for chronic exposure are lacking.^178,185

To implement diagnostic and therapeutic EVs clinically, standardizing isolation methods, storage conditions, and analysis platforms is essential. The International Society for Extracellular Vesicles (ISEV) has established minimal information guidelines, MISEV2018, updated in MISEV2023, outlining requirements for EV studies including isolation, characterization, and functional assessment. Wider adoption of these standards by researchers and industry will be crucial for generating reproducible, comparable datasets needed for regulatory approval.¹³

Although BM-derived EVs hold clear clinical potential, realizing it requires addressing these fundamental gaps systematically. Moving from strong preclinical observations to validated clinical tools will depend on coordinated advances in the scientific field of EV biology, delivery engineering, safety profiling, and regulatory framework—areas where progress is real yet consensus on solutions is still evolving.

6. Conclusions

The expanding EV field has reshaped our understanding of cell-to-cell communication in the BM microenvironment.¹⁸⁶ This review integrates current knowledge on EV biology across four interconnected dimensions: the regulated biogenesis pathways that govern cargo selection in HSCs and stromal cells; the functional roles of miRNAs, surface-exposed and intravesicular proteins, and other cargo in maintaining hematopoietic homeostasis; the dynamic remodeling of EV composition driven by hypoxia, inflammation and therapeutic pressure during malignant transformation; and the emerging clinical applications of EVs as liquid biopsy tools and therapeutic agents in hematological disorders. In healthy BM, EVs serve as sophisticated molecular shuttles regulating fundamental HSC properties. The recent finding that HSC-derived EVs can enhance stem cell function through autocrine signaling represents a paradigm shift in HSC homeostasis.^27,43 Moreover, a complex network of EV-mediated communication between HSCs and their microenvironment, MSCs, osteoblasts, endothelial cells, and immune cells, orchestrates the balance required for lifelong hematopoiesis.

EV cargo specificity, particularly in miRNA and protein content, provides mechanistic insights into how BM populations communicate and coordinate their functions. Specific miRNAs, such as miR-486, miR-126, and miR-10a, regulate erythroid differentiation, HSC mobilization, and megakaryocyte development, respectively, illustrating the precision of EV signaling.^38,71 Surface proteins, including tetraspanins, HSPs, and cell-specific markers, have also advanced the identification and tracking of EVs from distinct BM populations.

Isolating adequate amounts of EVs from BM cell populations, especially HSC, remains challenging due to the limited number of stem cells and the need for in vitro culture, which can induce differentiation.¹⁸⁷ Current EV isolation techniques do not readily accommodate such low sample quantities, making the development of improved protocols for BM-derived EVs essential for reliable omics data. Indeed, proteomic sensitivity may be constrained by the low EV input concentration, and a core protein signature for each BM cell population is still lacking. A comprehensive definition and characterization of protein-protein interactions between EVs and target cells within the BM is required to better understand how this mode of communication operates under physiological and pathological conditions in this intricate microenvironment.

In hematological malignancies, EVs emerge as key players in tumor progression, immune evasion, and therapeutic resistance. Although blood EVs were described over three decades ago, consensus on their definition and role in hematological malignancies is still lacking. Tumor cells exploit EV release to foster therapy resistance, using vesicles enriched with leukemic markers to evade binding.¹⁶³ Cancer-derived EVs actively reshape the BM microenvironment to favor malignant cell survival while suppressing healthy hematopoiesis. Leukemic EVs can induce residual HSC quiescence, promoting drug resistance through specific miRNA cargo and aiding immune escape through surface protein modifications, underscoring their central role in disease pathogenesis.¹²³

Clinical implications are profound. EVs offer non-invasive means for disease monitoring and prognosis via liquid biopsy.¹⁶⁵ EV markers such as CD19 in lymphoid malignancies,¹⁶¹ CD34/CD33/CD117 in AML, and diverse surface proteins in multiple myeloma underpin EV-based diagnostics.¹³⁶ Therapeutically, EVs hold promise as drug delivery vehicles or intervention targets.

Nevertheless, the precise mechanisms of EV uptake and cargo delivery in the BM remain incompletely understood. A comprehensive protein signature for EVs from each BM population is still lacking, limiting our ability to map BM intercellular communication fully. Therapeutic targeting of EV-mediated pathways requires deeper insights into EV-cell protein interaction.

The field stands at a pivotal juncture where single-cell analyses, advanced imaging, and improved isolation methods are beginning to address these limitations. Future investigations should develop sensitive protocols for rare cell-derived EV isolation, build molecular atlases of BM EV populations, and translate these findings into clinical diagnostics and therapies. Understanding EV-mediated BM communication is key to new therapies for hematological malignancies and to advancing stem cell biology. As we decode the EV language, we move closer to harnessing these natural communication systems for patient benefit. The field’s rapid growth and accumulating clinical relevance suggest that EVs will continue to answer critical questions in both stem cell research and cancer biology, ultimately improving outcomes for patients with hematological diseases.

Data availability

Statement Data sharing is not applicable to this article as no new data were created or analyzed in this study.