Bone fractures caused by traumatic injury or common age-associated disorders, such as osteoporosis, present an enormous healthcare and socioeconomic burden worldwide (Baroli, 2009; Dimitriou et al., 2011), with 10 million people suffering with musculoskeletal disorders in the United Kingdom alone (Chance-Larsen et al., 2019). This is anticipated to increase further in the future, due to the growing ageing population and the demand for sustained quality of life in the older years (Amini et al., 2012). Autologous bone grafts have been seen as the gold standard clinical therapy for many years, as they are histocompatible and non-immunogenic. However, studies have reported the considerable issues associated with this treatment including donor site morbidity, limited availability and risk of infection (Swan and Goodacre, 2006; Oryan et al., 2014). Consequently, there is an urgent demand for novel therapeutic approaches for bone augmentation strategies. Engineered bone tissue has been seen as a potential alternative to conventional bone graft treatments, however, there are limitations regarding the translation of cell-based therapies to the clinical setting including their inherent heterogeneity, uncontrolled differentiation, immune rejection, functional tissue engraftment and neoplasm formation (Amariglio et al., 2009; Herberts et al., 2011). Moreover, the translation of cell-based treatments are hampered due to issues such as intensive cost, government regulations and ethical issues (Izadpanah et al., 2006; Heathman et al., 2015; Volarevic et al., 2018). Hence, there is growing precedence to develop cell-free or acellular technologies to promote bone regeneration (Burdick et al., 2013; Wang et al., 2018).

Cells are known to secrete a range of bioactive factors, which have been shown to regulate biological processes such as proliferation and differentiation through autocrine and paracrine signalling (Baraniak and McDevitt, 2010; Raposo and Stoorvogel, 2013; de Miguel-Gómez et al., 2020). One of these factors, Extracellular Vesicles (EVs), have gained increasing attention as novel acellular tools for regenerative medicine. EVs are cell-derived lipid nanoparticles that contain a diverse biological cargo including nucleic acids, proteins and bioactive molecules (Zaborowski et al., 2015; van Niel et al., 2018). The multitude of bioactive factors delivered by EVs has been shown to promote stem cell mineralisation when compared to the use of single growth factor treatment (Davies et al., 2017). Although the therapeutic efficacy of these cell-derived nanoparticles has been reported (Martins et al., 2016; Qin et al., 2016; Man et al., 2021a), to date EVs are commonly harvested from cells cultured on 2D tissue culture plastic. Conventional 2D culture limits the surface area for cell growth as well as providing an artificial cell-cell and cell-extracellular matrix interactions, which do not replicate in situ conditions mechanically or biologically (Abbott, 2003; Baker and Chen, 2012). By conferring a more physiologically relevant phenotype, it is hypothesised that this increased biomimicry will enhance the production of EVs exhibiting a more in vivo-like composition, improving their therapeutic potency for regenerative medicine (Thippabhotla et al., 2019; Man et al., 2020). Several studies have reported the influence of 3D culture on increasing the differentiation capacity of cells (Baraniak and McDevitt, 2012; Yamaguchi et al., 2014), however, limitations on current scaffold manufacturing hinders their implementation. Conventional scaffold fabrication techniques including solvent casting, freeze-drying and gas foaming provide limited control on scaffold architecture (Tabata, 2009), ultimately impacting the reproducibility of EVs from these systems. Moreover, there are issues associated with the procurement of EVs isolated from non-porous biomaterials, such as hydrogels (Patel et al., 2018), which may require extensive processing to extract these nanoparticles, ultimately damaging the integrity and functionality of the vesicles. Hence, there is a significant unmet need to refine the culture conditions to enhance EV therapeutic potency and yield for bone augmentation strategies.

In recent years, Additive Manufacturing (AM) techniques have facilitated the reproducible fabrication of scaffold systems that may more closely replicate the increased complexity of in situ environments (Gu et al., 2016; Li et al., 2016). Titanium-based materials have been broadly implemented in orthopaedic reconstruction as a result of their osteoinductivity and mechanical strength (Chen et al., 2017; Ilea et al., 2019). Selective Laser Melting (SLM) has allowed for the processing of titanium into 3D printed structures, increasing its clinical utility (Majumdar et al., 2018; Elsayed et al., 2019). Thus, the use of 3D printed titanium scaffolds could provide a more physiologically relevant platform for the manufacture of pro-regenerative EVs compared to conventional 2D tissue culture plastic. Although titanium has been extensively used for numerous orthopaedic applications, there have been numerous investigations aiming to improve its functionality by applying different surface finishes to promote osteoinduction (Damiati et al., 2018; Kligman et al., 2021). Consequently, several studies have explored the influence of applying surface coatings to 3D printed constructs to further promote their biomimetic nature (Kaur and Singh, 2019; Kazimierczak and Przekora, 2020). For example, Eichholz et. al. reported the use of a nano-needle hydroxyapatite (nnHA) coating to enhance the osteoinductive capacity of polycaprolactone (PCL) scaffolds fabricated by melt electrowriting (Eichholz et al., 2020a). Hence, modifying the scaffold’s surface composition could further improve the production of EVs. Together, this study aims to address pertinent issues hindering the clinical application of EVs, scalability of manufacture and therapeutic efficacy, thus facilitating the development of these instructive acellular tools to promote bone repair.

In this present study, we investigated the influence of 3D printed scaffold architecture and surface composition on the therapeutic efficacy of osteoblast-derived EVs for bone repair. Titanium scaffolds were fabricated via SLM with differing pore sizes and shapes and determined their effects on osteoblast mineralisation (Figure 1). EVs were isolated from mineralising osteoblasts on different scaffold designs and administered to human bone marrow-derived mesenchymal stem cells (hBMSCs) to determine their efficacy in promoting osteogenic differentiation. Additionally, the effects of nnHA scaffold coating on osteoblast-derived EV yield and potency was evaluated.

There is growing precedence to utilise 3D culture platforms to more closely replicate in vivo environments, ultimately improving the biomimicry of EVs isolated from in vitro systems (Man et al., 2020). In this regard, harnessing AM methods such as 3D printing would allow for the large-scale, reproducible fabrication of 3D constructs with precise control on scaffold parameters, overcoming limitations of conventional scaffold systems for EV procurement (Tabata, 2009; Patel et al., 2018). Moreover, several studies have reported the osteoinductive capacity of titanium substrates (Xu et al., 2013; Salou et al., 2015), thus possibly providing a pro-osteogenic platform for EV manufacture. Therefore, this study aimed to address pertinent issues hindering the clinical translation of EVs, scalability of manufacture and therapeutic efficacy, hence facilitating the development of EVs as novel acellular therapies to promote bone repair.

It has become increasingly apparent that the mechanobiological interactions between cells and their substrates play a vital role in augmenting the cell’s functionality (Guilak et al., 2014). Numerous studies have shown the importance of pore geometry in modifying the differentiation capacity of cells (Van Bael et al., 2012; Xu et al., 2014; Rubert et al., 2021). In this study, constructs exhibiting a triangle pore conformation significantly improving osteoblast differentiation and mineralisation when compared to square pored scaffolds of both sizes. It has been similarly reported in the literature that a more acute 3D printed fibre angle promotes cellular contact, increasing the cell’s osteogenic capacity (Guvendiren et al., 2017; Ji and Guvendiren, 2019; Rüdrich et al., 2019). For example, Yilgor et al. described the fabrication of 3D printed porous PCL scaffolds with either a basic or crossed fibre orientation, where enhanced MSC osteogenesis was observed in the crossed fibre scaffolds (Yilgor et al., 2008). It has been proposed that substrates exhibiting a high curvature or acute fibre angle augments differentiation and tissue deposition by enhanced cellular tension at those interfaces (Rumpler et al., 2008). For instance, Bidan et al. showed a positive correlation between tissue formation and substrates with high curvatures, where osteoblasts deposited extracellular matrix on cross-shaped pores twice as fast when compared to square-shaped pores (Bidan et al., 2013). Hence, the increased substrate curvature of the triangle pore scaffolds may increase cellular tension within osteoblasts and promote their differentiation capacity. We found that cells cultured on scaffolds exhibiting a larger pore size increased osteoblast differentiation (ALP activity, collagen production and calcium deposition) when compared to smaller pores in both conformations, consistent with previous reports (Van Bael et al., 2012; Di Luca et al., 2016). Scaffolds possessing a triangle pore geometry and larger pore sizes exhibited increased porosity and permeability, likely enhancing the cell seeding efficiency of these scaffolds and media diffusivity during culture. Improving scaffold permeability through increasing pore size, has been shown to promote osteoblast differentiation via elevated fluid flow/shear stress (Ouyang et al., 2019). An increasing body of evidence has reported the influence of biomaterial cues as potent regulators of cell fate via altering epigenetic functionality (Li et al., 2011; Tan et al., 2015; Lv et al., 2018a). It is well known that during osteogenic differentiation, the epigenetic landscape of cells is altered. For example, Lee et al. showed that HDAC activity is reduced during osteogenesis, which resulted in hyperacetylation induced gene activation (Lee et al., 2006). As such, researchers have harnessed epigenetic modifying compounds including HDAC inhibitors (HDACis) to accelerate osteogenic differentiation both in vitro and in vivo (Huynh et al., 2016; Man et al., 2021c). Moreover, several studies have demonstrated the role of biomaterials substrates in augmenting key epigenetic functions via the modification of the cells actin cytoskeleton which transmits forces to the nuclear protein lamin A/C, ultimately resulting in enhanced transcriptional activation (Li et al., 2011; Downing et al., 2013). In this present study, our findings showed that osteoblast cultured on triangle pore scaffolds exhibited substantial reduced HDAC activity, resulting in increased histone acetylation levels. The enhanced transcriptional permissiveness induced by hyperacetylation has been reported to accelerate differentiation on biomaterial substrates via increased gene activation (Li et al., 2011; Lv et al., 2018b), consistent with the upregulated osteogenic gene expression observed in this study. Thus, these findings indicate the influence of biomechanical cues induced by the different scaffold architectures on augmenting osteogenic differentiation capacity via epigenetic regulation. Therefore, altering the scaffolds architecture through augmenting pore size/geometry has a significant impact in regulating osteoblast differentiation (Abbasi et al., 2020). Moreover, modifying the scaffolds pore geometry and size likely plays a pivotal role in the enhanced regenerative capacity of these constructs used for bone fracture healing.

Due to issues regarding EV diffusion efficiency from conventional scaffold systems and the lack of control in scaffold properties (Patel et al., 2018), 3D printing allows for the reproducible manufacture of constructs with controllable architectural parameters. The present findings show a significant increase in EV production for scaffold-cultured osteoblasts compared to 2D culture, consistent with several studies demonstrating elevated EV yield from 3D model systems (Zhang et al., 2017; Cha et al., 2018). Interestingly, we observed substantial differences in the quantity of EVs isolated from these constructs during osteoblast culture. Scaffolds possessing a triangle pore conformation secreted significantly enhanced EV yield (>2.15-fold) when compared to square pore scaffolds. Moreover, we showed that larger pore sizes resulted in a slight enhancement in EV yield in both pore conformations, although not significant. This EV release profile correlated with the differentiation status of the parental cell, likely impacting osteoblast differentiation and mineralisation observed during scaffold culture via autocrine and paracrine signalling. Interestingly, it has been reported that scaffolds exhibiting an acute pore angle promote osseointegration and de novo tissue formation following implantation (Deng et al., 2021). Hence, the influence of the scaffold geometric architecture likely impacts EV yield and potency in vivo stimulating fracture healing, although this would require further investigation. Through characterisation of the constructs, we showed that scaffolds with a triangle pore conformation and larger pore sizes exhibited an increased permeability/porosity. The improved scaffold permeability will likely enhance the media diffusivity during culture, impacting both osteoblast differentiation and EV procurement. Several studies have reported the influence of increased fluid flow on osteoblast differentiation, through the activation of mechanotransductive signals (Yourek et al., 2010) and the enhanced bioactivity of secreted EVs (Eichholz et al., 2020b). Hence, augmenting scaffold permeability through architectural modifications could provide a tailorable method to improve EV manufacture. In addition to the influence of scaffold architecture on EV procurement, 3D culture-induced differentiation may also increase EV biogenesis and production through increased intracellular calcium (Savina et al., 2003; Taylor et al., 2020). It has been reported that intracellular calcium is causatively linked to osteogenic differentiation. For example, Wu et al. showed that intracellular calcium levels within BMSCs increased during osteogenic differentiation, and the introduction of Thapsingargin, an inhibitor to endoplasmic reticulum calcium ATPase, reduced osteogenesis (Wu et al., 2020). Interestingly, we showed that osteoblasts cultured on triangle pore scaffolds exhibited a significantly elevated intracellular calcium content when compared to cells cultured on other scaffolds designs and 2D. These findings correlated with the quantity of EVs procured during culture. Therefore, EV procurement from these titanium lattices is likely influenced by the scaffold effects on osteoblast differentiation, the cell secretion phenotype and scaffold permeability/porosity.

Several studies have demonstrated the enhanced biological potency of EVs procured from cells cultured in 3D models compared to 2D (Cha et al., 2018; Otto et al., 2021). The use of 3D printing would allow for the fabrication of biomimetic scaffold systems, allowing for the procurement of EVs from cells with a more physiological relevant phenotype when compared to conventional 2D culture. The present results showed that EVs derived from titanium lattices significantly improved hBMSCs osteogenic differentiation when compared to EVs procured from 2D cultured cells. It is well known that culturing cells in more physiological relevant models promotes their osteogenic capacity, thus, it is likely the augmented EV yield and potency induced by modifying the cells substrate, plays a critical role in the enhanced differentiation capacity of cells observed in 3D culture. The data also showed that EVs derived from the triangular pore conformation displayed increased osteoinductive potency when compared to vesicles isolated from cells cultured on the square pore scaffolds. The enhanced osteoinductive efficacy of EVs procured from the triangle pore scaffolds correlated with the differentiation status of the parental cell during osteogenic culture. Increasing evidence has reported the influence of biophysical and biochemical stimulation on augmenting secreted EVs osteoinductive potency. For example, Davies et al. reported EVs procured from mineralising osteoblast were significantly enriched in proteins involved in osteoblast communication and extracellular mineralisation such as bridging collagens and annexin calcium channelling proteins when compared to EVs from non-mineralising cells (Davies et al., 2017). Eichholz et al. showed that mechanical stimulation of osteocytes accelerated its differentiation and the osteoinductive potency of their EVs via the enrichment of several pro-osteogenic proteins (Eichholz et al., 2020b). Thus, it is likely the accelerated mineralisation induced by the triangle pore scaffolds enriched the secreted EVs with proteins involved in osteoblast differentiation and mineralisation. Additionally, in this study, we reported that osteoblasts cultured on triangle scaffolds elicited substantially altered epigenetic functionality, resulting in increased osteogenic gene activation. Previously we demonstrated that altering the epigenome of mineralising osteoblast via HDACi induced hyperacetylation, significantly enriched secreted EVs with several pro-osteogenic microRNAs and transcriptional regulators which augmented the epigenetic landscape of recipient hBMSCs and accelerated its osteogenic capacity (Man et al., 2021a). Therefore, it is likely the EVs derived from osteoblasts cultured on triangle pore scaffolds were enriched with bioactive factors which modified the epigenetic landscape at recipient hBMSCs increasing its differentiation capacity. Future work elucidating the scaffold-derived EVs mechanisms of actions would be an important area for future studies. Together, these findings indicate the importance of EV parental cell phenotype on the biological potency of their secreted EVs, as these nanoparticles are essentially a fingerprint of their donor cell (Zhang et al., 2019; Sahoo et al., 2021). Moreover, the use of titanium implants exhibiting a triangle pore architecture in vivo could facilitate the production of EVs from recruited stem/progenitor cells, further potentiating the scaffolds role in stimulating endogenous bone repair by increasing pro-regenerative secretome signalling. Therefore, the use of AM techniques such as 3D printing allows for the tailorable fabrication of biomimetic culture platforms to enhance the osteoinductive potency of EVs.

In addition to the influence of scaffold architecture on improving material functionality, several studies have investigated the effects of different surface coatings on promoting the biomimetic nature of biomaterials (Kaur and Singh, 2019; Kazimierczak and Przekora, 2020). For example, Eichholz et al. developed a bone-mimetic nnHA coating, which enhanced the osteoinductive capacity of PCL scaffolds (Eichholz et al., 2020a). Moreover, it has been reported that the elevation of extracellular calcium levels transiently increases intracellular concentrations through the activation of calcium-sensing receptors (Yanai et al., 2019). Therefore, the application of a calcium-based scaffold coating could promote osteoblast differentiation and EV production. We demonstrated that the application of a nnHA coating significantly enhanced mineralisation of osteoblasts during culture, consistent with several studies in the literature (Habraken et al., 2016; Eichholz et al., 2020a). Interestingly, we observed a significantly increased quantity of EVs procured from the nnHA coated scaffolds (>3.5-fold) when compared to the uncoated groups. This profile correlated with the differentiation status of the parental cell and its secretory phenotype (intracellular calcium levels). The enhanced quantity of EVs released during osteoblast culture on the nnHA coated scaffolds, likely played a pivotal role in promoting the mineralisation observed on the constructs. Interestingly, there was no significant difference in the osteoinductive potency of EVs procured from the nnHA scaffolds compared to the untreated groups. Thus, there is likely a delicate balance between promoting EV yield and the biological potency of these cell-derived nanoparticles. It is important to note, that given the dramatic increase in EV yield induced by nnHA coated scaffolds, the direct implantation of nnHA coated constructs in situ would likely promote osteogenesis as a result of enhanced EV production, rather than augmenting the potency at an individual vesicle level. These findings suggest that the application of a nnHA coating would improve the bone formation capacity of titanium substrates, consistent with several studies in the literature (Oonishi et al., 1989; Popkov et al., 2017). Although titanium alloys exhibit a poor degradation profile, the nnHA coating has demonstrated its versatility to improve the osteoinductive potency of non-metallic scaffolds (Eichholz et al., 2020a). Therefore, these findings show that the application of a simple nnHA scaffold coating can significantly augment the production of pro-osteogenic EVs, further enhancing the scalable manufacture of these nanoparticles for clinical use.

In this study, we investigated the influence of 3D printed scaffold architecture on the osteoinductive potency of osteoblast-derived EVs for bone repair. Scaffolds exhibiting a triangle pore conformation and larger pore size displayed increased porosity and permeability. Osteoblast differentiation was significantly accelerated when cultured on triangle pore scaffolds, with increased EV production from these scaffold systems. EV procurement yield correlated with osteogenic phenotype, intracellular calcium levels and scaffold porosity. EVs derived from the triangle pore scaffolds significantly promoted hBMSCs osteogenic differentiation when compared to EVs derived from other scaffold designs. Interestingly, we demonstrated that a simple nnHA coating significantly improved osteoblast osteogenic differentiation and EV yield when compared to the uncoated groups. Future studies would be required to investigate the influence of different scaffold parameters on EV composition, which may be linked to their mechanism of action. A limitation of this study is the lack of in vivo assessment which would provide increased pre-evidence into the biological potency of these scaffold-derived EVs in stimulating bone regeneration. Moreover, there is tremendous scope to investigate the influence of combining tailored 3D printed scaffolds with bioreactors systems to further increase the scalable manufacture of therapeutic relevant EVs for clinical applications.

Together, these findings demonstrate the impact of 3D printed scaffold architecture and surface composition on improving EV manufacture and their therapeutic efficacy, indicating the considerable potential of harnessing bone-mimetic culture platforms to enhance the production of pro-regenerative EVs for bone repair.