The therapeutic potential of hematopoietic stem cells in bone regeneration

The repair process of bone fractures is a complex biological mechanism requiring the recruitment and in situ functionality of stem/stromal cells from the bone-marrow (BM). While BM mesenchymal stem/stromal cells have been widely explored in multiple bone tissue engineering applications, the use of hematopoietic stem cells (HSCs) has been poorly explored in this context. A reasonable explanation is the fact that the role of HSCs and their combined effect with other elements of the hematopoietic niches in the bone healing process is still elusive. Therefore, in this review we intend to highlight the influence of HSCs in the bone repair process, mainly through the promotion of osteogenesis and angiogenesis at bone injury site. For that, we briefly describe the main biological characteristics of HSCs, as well as their hematopoietic niches, while reviewing the biomimetic engineered BM niche models. Moreover, we also highlighted the role of HSCs in translational in vivo transplantation or implantation as promoters of the bone tissue repair.


Introduction
The high incidence of bone defects places the bone as the second most transplanted tissue after blood transfusion 1 . Although bone tissue has the peculiar ability to fully regenerate and restore its biomechanical function, about 10% of fractures fail to heal properly 2,3 . In these cases several surgical interventions are required, which potentiates the risk of infection, pain, and disability 4 . Currently, even the gold standard treatment option for bone regeneration, namely bone autografts, has shown some drawbacks [5][6][7] . In the last decade, the demonstration that HSCs can act through a variety of mechanisms to promote repair and tissue regeneration has dramatically broadened their clinical utility for the repair and regeneration of several nonhematopoietic tissues  . Considering that the development of effective bone therapy strategies continues a challenge, researchers in bone tissue regeneration have shown great interest in exploring the use of HSCs [14][15][16] .
The BM specialized microenvironments hold distinct cellular and non-cellular components where osteogenesis and hematopoiesis occur [17][18][19][20] . The dynamic BM microenvironment and the interactions between its components have been explored as novel therapeutic targets to facilitate the bone's regenerative capacity. Although most stem cell-based therapy approaches 3 aiming bone regeneration have focused on the role of BM-derived MSCs (BM-MSCs) and their interaction with endothelial cells (ECs), recently the plethora potential use of HSCs alone 10,[21][22][23][24][25][26] or in combination with other cells from the BM [27][28][29] has been recently proposed and investigated for bone healing strategies. The understanding of the cellular and molecular crosstalk involved in the HSCs regulation and cell fate, in both health and bone disorder, is of fundamental importance to precisely clarify, control, modulate and find new HSC-therapeutic strategies aiming towards bone regeneration. In such context, advanced models that mimic the BM-HSCs niches have been developed to explore its biological complexity and precisely control key components that facilitate the regenerative process [30][31][32][33] . However, the understanding of such privileged environment is still elusive.

The hematopoietic stem cells and their bone-marrow niches
Self-renewal and pluripotency properties ensure the maintenance of functional hematopoiesis during the lifetime [34][35][36] . This hierarchical process is represented in figure 1. The primitive longterm HSCs (LT-HSCs) self-renew and maintain the HSC pool, while the short-term HSCs (ST-HSCs) differentiate towards the multipotential progenitors (MPPs). The MPPs give rise to common myeloid progenitors (CMPs) and common lymphoid progenitors (CLPs), originating all the diverse mature and functional hematopoietic cell types in vivo 37,38 . The HSCs and their progenitors are subdivided according to presence or absence of specific cell membrane markers (CD34, CD38, CD90, and CDR45) 38,39 . These cells are dynamic with a nonrandom spatial orientation within their BM niches, the endosteal and the perivascular niche. These niches comprehend a complex network of cell-cell contact and interaction with non-cellular components, all necessary for the HSC maintenance, proliferation, activation, differentiation, and migration 18,34,40-42 ( Figure 2).

Constituents of the Endosteal niche
The endosteal niche is functionally responsible for the regulation of bone formation and resorption 33,[43][44][45] . Localized at the inner surface of the bone cavity, this niche ensures the 4 maintenance of the HSCs numbers and trans-marrow migration 18, 34,46,47 . The LT-HSCs are found quiescent in this niche and once activated they self-renew to sustain hematopoiesis during lifetime 9,48 . Osteoblasts control HSC stemness, and quiescence by N-cadherinmediated adhesion 49. These cells secrete a range of cytokines, such as stem cell factor (SCF), thrombopoietin (TPO), osteopontin (OPN), angiopoietin-1 (Ang-1), CXC-chemokine ligand 12 (CXCL12), CXCL-4, and interleukin (IL)-6 32,45,50 . Together with osteoclasts, these cells are responsible for bone remodeling. Once stimulated, osteoclasts secrete enzymes that cleave osteoblast-expressed niche molecules. Consequently, HSCs and HPCs are released from the niche and mobilized to the periphery from BM 51 . This osteoclast-osteoblast close association suggests a delicate and controlled balance of the HSC-regulating cytokines which is responsible to support the endosteal niche and required for bone repair. BM-MSCs can differentiate towards osteoblasts, chondrocytes, and adipocytes. These cells support HSC regulation through the production of interleukins, TPO, SCF, macrophage colony-stimulating factor, Flt3 ligand, Ang-1, and CXCL12 33 . Furthermore, BM-MSCs produce and create a network of extracellular matrix (ECM) proteins including proteoglycans, fibronectin, collagen, laminin, and thrombospondin 52 . These components modulate the HSC behavior by promoting cell homing, viability, self-renewal, expansion, differentiation, and mobilization of the HSCs between their different niches 32,53,54 .

Constituents of the perivascular niche
The perivascular niche is placed adjacent to the blood vessels and secrete angiocrine factors required for the survival, maintenance, and self-renewal of the HSCs/HPCs 55  Despite the efforts of intensive investigation to better characterize and understand the HSCniche components, their interactions, and localization, a deep understanding of the mechanisms involving the HSC regulation and cell fate, in both health and bone disorder, is still crucial to identify key components that can facilitate and induce the bone regenerative process.

HSCs as powerful candidates aiming towards bone regeneration
After a bone fracture, loss of skeletal integrity, disruption of the bone vasculature, hematoma, and inflammation, occur locally 6,62 . The vascular restoration of bone forecasts its repair [63][64][65] . In fact, high vascularization of bone provides key cell players, including osteolineage cells, MSCs, HSC/HPCs, endothelial progenitor cells (EPCs), and ECs, for a proper bone formation, remodeling, repair, and homeostasis. Although these events compromise the hematopoietic niches, they also stimulate the HSCs to switch from quiescent to proliferative and differentiation state to quickly recover the hematopoiesis system, which is an essential HSC property to repair the bone tissue 10 . 6 The CD34 + cells population comprehend a plethora of cellular phenotypes, including EPCs and osteo precursors cells [66][67][68][69][70] . Several studies report the use of CD34 + cells either from PB, BM, or UCB to enhance fracture repair in animal and human models, through angiogenesis and osteogenesis 21,66,71,72 ( Figure 3). In the last decades, the ability of CD34 + cells to differentiate into ECs and osteoblasts were pointed out, suggesting a possible overlap between endothelial and osteoblast precursor cells 16,68 . Importantly, it is not clear yet whether HSCs are able to differentiate into ECs or osteoblasts, or a fraction of the CD34 + cells population could be instead accountable for that. Common BM progenitors with hematopoietic/endothelial and hematopoietic/osteoblastic differentiation potential may exist and be implicated in such endothelial and osteogenic phenotypic cells. Consequently, the mechanisms by which the bone healing potential of CD34 + is addressed have been extensively investigated, but it still requires additional clarities.

The angiogenic potential of HSCs
The contribution of the CD34 + cells for the restoration of bone vasculature at injury sites is indubitably evident. So far, these cells are believed to induce vascularization through endothelial differentiation and paracrine stimulation. With the discovery of EPCs in adults, the PB, BM and UCB-derived EPCs have gained attention for neovascularization therapies 73-76 .
It strongly suggests that in response to ischemia and cytokines, corresponding to the early/acute phase of bone healing, EPCs are recruited from BM into PB and are then mobilized to the fracture site, where they differentiate into mature ECs [77][78][79] . Histological studies have uncovered the occurrence of neovascularization at local fracture independent of vasculogenesis from BM-EPCs. This finding suggests that BM-EPCs may also have a paracrine effect on resident ECs and EPCs, resulting in angiogenesis and vasculogenesis orchestrated by the respectively resident cells 71 . Accordingly, transplanted CD34 + cells have shown to secrete vascular endothelial growth factor (VEGF), whose inhibition with a soluble antagonist showed both angiogenesis/vasculogenesis and intrinsic osteogenesis suppression, emphasizing the contribution of the paracrine mechanism 66 . 7

The osteogenic potential of HSCs
Similarly, an osteo precursors population enriched in CD34 + cells and osteogenic paracrine mechanisms are believed to be accountable for the osteogenic potential of the CD34 + cells.
PB and BM-derived CD34 + cells osteoblastic differentiation is reported in in vitro studies, when cultured in the presence of osteogenic medium containing dexamethasone, resulting in mineral matrix secretion and increase in alkaline phosphatase (ALP) activity 23,69 . PB-CD34 + cells were found to express osteoblastic genes such as osteocalcin, collagen type I and bone Considering all these evidences, it is clear that CD34 + cells exert an exceptional effect in the process of bone healing by promoting adequate conditions to angiogenesis and osteogenesis occur. Although, the precise cellular and molecular mechanisms are still not entirely clarified; similarly, the identity and characterization of the specific CD34 + cells accountable for the outcomes remains poorly discerned.

Bioengineering strategies for recapitulation of HSC niches
Bioengineered BM niches allow a better understanding of the dynamic signaling between their elements 81 . Particularly, the hematopoietic niches have recently received special attention to clarify the role of HSCs/HPCs in bone regeneration (Table 1) Correspondingly, HSCs showed migration-dependence towards the high levels of the chemokine CXCL12 secreted by MSCs. Indeed, the addition of AMD-3100, a CXCR4 antagonist, triggered mobilization of the primitive HSCs from the co-culture matrix to the supernatant 83 . CXCL12 is reported to be implicated in the maintenance of HSCs quiescence [102][103][104][105] . Moreover, a large percentage of the retained HSCs in the co-culture revealed an expression of the protein p21 by qPCR assay 87 , which is also recognized as an important regulator of HSCs quiescence 106 . Altogether, the co-culture with MSCs have proven to provide an adequate stromal support which, similarly to the native HSC niche, are able to foster a large pool of quiescent HSCs whereas expansion and self-renewal are ensured.
Heretofore, these types of 3D engineered hematopoietic niches in static systems have shown to mimic part of key features of the native BM niches, such as bone architecture, ECM secretion, osteogenesis stimulation by the HSC-MSC interactions and cell signaling molecules which ultimately support the maintenance of HSCs with preservation of the primitive phenotype, quiescent state, and multi-lineage differential potential.

Dynamic culture systems
To more accurately recapitulate the physiologic conditions of BM-HSC niches, bioengineered hydrogels and scaffolds have been designed through 3D in vitro-culture approaches combining advanced biomaterials and dynamic culture systems. In fact, such innovative systems are becoming a promising bioengineering tool to approximate the bench research to the in vivo physiology. The capability to control key physicochemical parameters such as oxygen levels and mechanical forces would permit the conceiving of more complex BM models. Moreover, these approaches are also suitable for drug and toxicity testing 107,108 .
Microfluidic techniques are the utmost sophisticated systems that have been ultimately applied to produce bone-marrow-on-a-chip (BMoC) units for dynamic systems [107][108][109]  Although the cells were not in direct contact, the authors did not explore the paracrine stimulation between endothelial and the co-cultured cells.
In recent years, a biomimetic perfusion bioreactor containing a porous hydroxyapatite scaffold functionalized with hUCB-HSCs and hBM-MSCs was purposed as an advantageous BM niche comparing to microfluidic systems 31

Bone regenerative properties of HSCs in translational in vivo models
The potential of CD34 + cells in bone regeneration is recognized in both animal and human in vivo models. HSCs are mainly transplanted/implanted monocultured or co-cultured with MSCs ( Table 2) HSC implantation has also been performed in human models to assess its therapeutical potential in bone healing for mandible defects, and tibial or femoral nonunion fractures.
Accordingly, patients with 6-to 8-continuity defects of the mandible received an in situ tissueengineered graft with a combination of hBM-derived CD34 + and osteoprogenitor cells, together with BMP-2 in an absorbable collagen sponge 72

Conflicts of interest
The authors declare no conflict of interest.