Mechanomedicine-guided mechanical preconditioning of dental-derived stromal cells for tissue regeneration

Abstract

Dental-derived stromal cells (DSCs), including periodontal ligament stem cells, dental pulp stem cells, stem cells from the apical papilla, and stem cells from human exfoliated deciduous teeth, are promising candidates for oral and craniofacial regeneration because of their accessibility, expandability, and functional relevance to periodontal, dentin–pulp, bone, and neurovascular repair. However, the therapeutic performance of DSC-based products remains inconsistent, partly because conventionally expanded cells may be insufficiently adapted to the mechanical cues encountered in vivo. In this Review, we present mechanical preconditioning as a mechanomedicine-guided strategy for ex vivo functional priming of DSCs. We summarize how major DSC populations respond to defined biophysical cues such as tensile and compressive forces, fluid shear stress, hydrostatic pressure, matrix stiffness, and surface topography, and we discuss the principal mechanotransduction pathways involved. We further outline representative quantitative loading windows and consider how these may support subtype-specific and indication-specific preconditioning design. Finally, we highlight key translational barriers, including stromal cell heterogeneity, donor variability, senescence, uncertain persistence of mechanically induced states, safety concerns, and the lack of standardized manufacturing workflows. Overall, clinical translation will require a shift from descriptive mechanobiology toward parameter-defined, indication-specific, and good manufacturing practice-compatible preconditioning strategies for DSC-based regeneration.

Graphical Abstract

Introduction

Mesenchymal stromal cells (MSCs) are heterogeneous adherent stromal cell populations in which bona fide stem cells constitute only a subset. Standard isolation and expansion procedures generally yield non-clonal mixtures that also include lineage-committed progenitors and more differentiated stromal cells [1, 2]. Among them, dental-derived stromal cells (DSCs) originate from oral tissues and include periodontal ligament stem cells (PDLSCs), dental pulp stem cells (DPSCs), stem cells from the apical papilla (SCAP), and stem cells from human exfoliated deciduous teeth (SHED) [3]. Although these populations are conventionally referred to as “stem cells” in the dental literature, they are more accurately regarded as heterogeneous, non-clonal stromal/progenitor mixtures rather than uniformly defined stem-cell entities [4, 5]. Nevertheless, DSCs remain highly attractive for regenerative applications because they are relatively accessible, can be obtained across multiple developmental stages, expand efficiently ex vivo, and are functionally relevant to periodontal, dentin–pulp, craniofacial, and neurovascular repair [6]. Their generally favorable immunological profile further supports their therapeutic potential [7, 8]. However, despite encouraging preclinical findings, the performance of DSC-based therapies remains inconsistent. One plausible explanation is that conventional expansion is typically performed under mechanically simplified culture conditions that do not recapitulate key biophysical features of the target tissue, potentially limiting subsequent adaptation after implantation [9, 10].

DSCs operate within microenvironments shaped by both biochemical and biophysical cues. A landmark study by Engler et al. established matrix elasticity as a determinant of lineage bias in MSCs, providing a conceptual foundation for mechanical regulation in regenerative biology [11]. This principle is directly relevant to DSCs, whose tissues of origin are embedded in mechanically dynamic oral environments. However, the biomechanical context of oral tissues should be distinguished from the actual biophysical inputs experienced by cells during ex vivo conditioning. Processes such as tooth eruption, occlusion, implantation, and orthodontic tooth movement underscore the mechanically active nature of oral tissues. However, isolated DSCs in experimental systems do not directly experience these macroscopic mechanical events. Rather, they respond to defined physical cues such as tensile and compressive forces, fluid shear stress (FSS), hydrostatic pressure, matrix stiffness, and surface topography [9, 10, 12, 13]. These inputs are sensed through mechanotransduction systems involving integrins, ion channels, focal adhesion complexes, and cytoskeletal–nuclear coupling, and can induce downstream changes in gene expression and cell state that regulate lineage bias and cellular function [14, 15]. Through these processes, mechanical regulation contributes to tissue homeostasis, adaptive remodeling, and regenerative behavior [16].

Against this background, it is important to distinguish mechanobiology from mechanomedicine. Mechanobiology addresses how cells sense, transmit, and respond to physical forces, whereas mechanomedicine refers to the translational application of these principles to guide therapeutic design and improve function [17, 18]. Within this framework, mechanomedicine-guided preconditioning can be understood as the intentional application of controlled ex vivo physical cues to functionally prime cells before implantation. Several studies already support the feasibility of this approach. For example, geometrical confinement through micropatterning has been shown to reorganize the cytoskeleton of DPSCs while preserving stemness-related features, and such mechanically informed conditioning was associated with improved peripheral nerve regeneration in vivo [19]. Likewise, centrifugation-based loading has been reported to increase SCAP proliferation and to enhance osteo-/odontogenic marker expression and mineralization [20]. Although these findings do not define a universal mechanical recipe for all DSC populations, they support the broader premise that defined physical cues can be used to bias DSCs toward therapeutically relevant states.

In this review, we focus on mechanical preconditioning as a mechanomedicine-guided ex vivo functional priming strategy for DSCs. Specifically, we compare how major DSC populations respond to the principal biophysical cues relevant to preconditioning. We then examine the mechanotransduction pathways most relevant to preconditioning and discuss how defined physical inputs may be converted into functionally relevant cell states, such as matrix-adaptive remodeling, osteo-/odontogenic bias, reparative activation, neuro-supportive plasticity, and immunomodulatory tuning. We also summarize representative quantitative loading ranges, where available, and discuss how these parameters may support subtype- and indication-tailored preconditioning design. Finally, we evaluate the major challenges that currently limit translation, including stromal-cell heterogeneity, donor variability, senescence, durability of mechanical memory, safety concerns, and the lack of standardized manufacturing workflows. Given that the current evidence base remains predominantly preclinical and highly heterogeneous, limiting direct cross-study comparison and precluding the definition of a universally validated conditioning strategy across DSC subtypes or application scenarios, this review is intended as a conceptual and translational synthesis rather than a prescriptive framework. Through this framework, we aim to provide a more coherent basis for developing parameter-defined, indication-matched, and good manufacturing practice (GMP)-compatible preconditioning strategies for DSC-based regeneration and regenerative applications.

Responses of DSCs to defined biophysical cues

DSCs respond to a broad spectrum of physical inputs that can influence proliferation, lineage bias, reparative activation, and matrix-adaptive behavior. In experimental settings, these responses are driven not by macroscopic oral events per se, but by defined biophysical cues at the cellular level, including tensile and compressive forces, FSS, hydrostatic pressure, matrix stiffness, viscoelasticity, and surface topography [9, 10]. Because these cue types can be precisely controlled and manipulated during ex vivo culture, they provide the practical basis of mechanomedicine-guided DSC preconditioning [16, 18].

Importantly, DSC populations do not respond uniformly to these cues. Their mechanoresponsiveness appears to vary with tissue source, developmental status, native matrix context, and baseline functional state [3, 21]. Consequently, the same class of physical input may produce distinct outcomes across PDLSCs, DPSCs, SCAP, and SHED [6, 22]. Accordingly, the following sections are organized according to the major categories of biophysical cues experienced by cells, with emphasis on subtype-specific response patterns under each condition (Fig. 1).

Fig. 1.

Major biophysical cues and subtype-specific response tendencies of DSCs. A Tensile and compressive forces. Tensile loading generally promotes osteogenic responses in PDLSCs and DPSCs, whereas compressive loading is more commonly associated with osteoclast-related signaling or pressure-dependent remodeling. In SCAP, centrifugation-based loading promotes osteo-/odontogenic differentiation. B Fluid shear stress and hydrostatic pressure. Fluid-associated cues regulate both immunomodulatory and differentiation-related responses in DSCs. In PDLSCs, fluid shear stress is linked primarily to immunomodulatory regulation, whereas hydrostatic pressure in pulp-derived populations is more often associated with osteogenic or odontogenic responses. C Matrix stiffness and viscoelasticity. Matrix mechanics regulate DSCs in a subtype-dependent manner. Stiffer matrices generally favor osteogenic differentiation in PDLSCs, DPSCs, and SHED, whereas softer matrices in DPSCs are more closely associated with quiescence. SCAP show a more restricted and non-linear stiffness response, with softer matrices favoring chondrogenesis and stiffer matrices favoring osteogenesis. D Surface topography and geometrical confinement. Nanofibrous architectures, micropattern-induced confinement, and microgrooved substrates regulate cell alignment, cell state, and lineage tendency, thereby supporting odontogenic differentiation or neuro-supportive behavior in selected DSC populations. Overall, the figure summarizes representative cell-experienced biophysical cues and broad subtype-specific response trends relevant to preconditioning design. Its content is based on findings reported in references [19, 20, 24–41, 43–57, 61, 62]. Created with BioRender.com. PDLSCs Periodontal ligament stem cells, RUNX2 Runt-related transcription factor 2, ALP Alkaline phosphatase, OCN Osteocalcin, DPSCs Dental pulp stem cells, RANKL Receptor activator of nuclear factor-kappa B ligand, TNF-α Tumor necrosis factor-α, IL-6 Interleukin-6, SCAP Stem cells from the apical papilla, OSX Osterix, IDO Indoleamine 2,3-dioxygenase, Tregs Regulatory T cells, BMP-2 Bone morphogenetic protein-2, SHED Stem cells from human exfoliated deciduous teeth, PANX3 Pannexin 3, DSPP Dentin sialophosphoprotein

Tensile and compressive forces

Tensile and compressive regimens are among the most extensively studied mechanical inputs in DSC research, yet their biological effects depend strongly on cue direction, magnitude, frequency, duration, and cell type [23]. In PDLSCs, which arise from the naturally load-bearing periodontal ligament, tensile loading generally promotes osteogenic programs, whereas sustained compression more readily activates inflammatory and osteoclast-related signaling in a parameter-dependent manner [24, 25]. Static tension and cyclic stretch increase osteogenic markers such as runt-related transcription factor 2 (RUNX2), alkaline phosphatase (ALP), and osteocalcin (OCN), consistent with roles in matrix remodeling and alveolar-bone-associated repair [26–28]. By contrast, compressive regimens can increase receptor activator of nuclear factor-kappa B ligand (RANKL), tumor necrosis factor-α (TNF-α), and interleukin-6 (IL-6), consistent with a shift toward osteoclastogenic and resorptive remodeling under selected loading conditions [24, 29–31]. Compressive force can also activate autophagy-related adaptive responses in PDLSCs, indicating that compression regulates not only osteoclastogenic remodeling but also broader cellular stress-response programs [32]. Together, these findings indicate that PDLSCs respond in a direction-dependent manner to tensile versus compressive loading.

DPSCs also respond to tensile and compressive stimuli, but their response profile differs from that of PDLSCs. Cyclic tensile loading increases proliferation and promotes a more osteogenic, bone-like extracellular matrix (ECM) phenotype, as evidenced by increased collagen and osteopontin (OPN) expression together with upregulation of collagen type I, fibronectin, osteoprotegerin (OPG), and bone sialoprotein [33, 34]. Static equiaxial strain induces a broadly similar pro-osteogenic response under biochemical-reagent-free conditions, accompanied by increased expression of osteogenic markers such as ALP, osteopontin, and osteocalcin [35]. However, these effects are strongly parameter dependent. Uniaxial cyclic stretch has been reported to suppress osteo-/odontogenic differentiation and, in some settings, to enhance proliferation, suggesting a trade-off between cell growth and lineage commitment under specific loading conditions [36, 37]. Compression-dependent responses are likewise sensitive to both magnitude and duration. Moderate uniaxial pressure (0.09–0.18 MPa) enhances proliferation and early odontogenic differentiation, with increased ALP activity, matrix mineralization, and expression of dentin sialophosphoprotein (DSPP), dentin matrix acidic phosphoprotein 1 (DMP1), and RUNX2 observed on day 7 [38]. More broadly, mechanically applied loading within three-dimensional biomimetic scaffolds has also been reported to promote odontoblastic differentiation in DPSCs [39]. By contrast, stronger pressure-related stimulation in deciduous dental pulp-derived cells has been associated with secretory mammalian Ly-6 urokinase-type plasminogen activator receptor-associated protein-1 (SLURP-1)/α7 nicotinic acetylcholine receptor (α7nAChR)/nuclear factor kappa B (NF-κB)-mediated pro-resorptive signaling, suggesting that excessive loading may shift the cellular response away from reparative differentiation and toward resorption-related programs [35]. Collectively, these findings indicate that tensile and compressive cues can promote therapeutically relevant responses in DPSCs, but only within an appropriate parameter window.

SCAP also respond robustly to direct loading. In representative studies, centrifugation-based loading at approximately 200 g significantly increased proliferation and enhanced osteo-/odontogenic differentiation, as evidenced by elevated ALP activity, increased mineralized nodule formation, and upregulated expression of RUNX2, osterix (OSX), OCN, and DSPP [20]. These findings support the view that developmental dental populations can be biased toward mineralized-tissue programs under appropriately scaled direct force. By contrast, direct evidence for classical tensile or compressive responses in SHED remains comparatively limited, and most SHED-related studies have been conducted under hydrostatic or matrix-defined conditions [40–43].

Overall, tensile and compressive cues have been most thoroughly characterized in PDLSCs and DPSCs, whereas direct-loading evidence remains more limited in SCAP and classical tensile/compressive responses remain less well defined in SHED [20, 25, 34, 43]. For preconditioning design, these findings argue against treating mechanical loading as a single category and instead support distinguishing tension from compression, and moderate from excessive loading, when designing subtype-specific regimens.

FSS and hydrostatic pressure

Fluid-associated mechanical cues are particularly relevant to dentin–pulp-derived populations, although they also regulate periodontal-derived cells through adaptive and immunomodulatory programs [13, 44, 45]. In PDLSCs, FSS has been reported to increase indoleamine 2,3-dioxygenase (IDO) activity, promote regulatory T-cell differentiation, and restrain effector T-cell proliferation, thereby supporting local immune homeostasis [13]. Taken together, these findings suggest that fluid-associated cues in periodontal-derived cells may be particularly important for immunomodulatory regulation rather than solely for directing lineage commitment.

In DPSCs, FSS and hydrostatic pressure are highly relevant to the dentin–pulp microenvironment, where confined geometry and interstitial fluid movement are likely to shape cell behavior. Representative studies applying shear stress at 0.5 Pa before osteogenic induction have reported enhanced osteogenic differentiation, suggesting that fluid-flow-based conditioning may support reparative and mineralizing programs in DPSC populations [45, 46]. Consistent with these observations, dynamic hydrostatic pressure has been reported to promote DPSC differentiation, as evidenced by earlier calcium accumulation in vitro, enhanced responsiveness to bone morphogenetic protein-2 (BMP-2), and increased hard-tissue regeneration in vivo, despite reduced post-loading adhesion and survival [44]. More broadly, pressure-related cues appear well aligned with the role of DPSCs in dentin repair and may therefore be particularly useful for reparative priming.

SHED likewise respond to hydrostatic pressure. Hydrostatic pressure enhances calcium deposition, increases expression of odontogenic markers such as pannexin 3 (PANX3) and DSPP, and promotes Piezo1-dependent nuclear translocation of RUNX2 [41]. In addition, pressure-associated stimulation can trigger ATP release and modulate stemness-associated signaling through crosstalk involving IL-6 and Rex-1 [40]. These observations indicate that hydrostatic cues regulate not only differentiation-related programs but also stemness-associated states in SHED.

Compared with DPSCs and SHED, fluid-associated cues in SCAP remain less extensively characterized, although the developing root apex is likely to experience mixed force and fluid environments during morphogenesis. At present, the strongest evidence for fluid and hydrostatic pressure lies in dentin–pulp-derived populations [40, 41, 44, 45]. Overall, this class of cues appears particularly relevant to reparative priming, odontogenic commitment, and immunomodulatory regulation, rather than to generalized osteogenesis alone.

Matrix stiffness and viscoelasticity

Matrix mechanics are among the most experimentally tractable inputs in DSC studies, but their effects are distinctly subtype dependent. In PDLSCs, increasing substrate stiffness generally promotes proliferation and osteogenic differentiation. Studies spanning approximately 6–135 kPa suggest that periodontal-derived cells tend to favor the stiffer end of tested soft-to-intermediate matrices for osteogenic and matrix-remodeling programs [47–49]. Similarly, PDLSCs encapsulated within stiff, highly crosslinked gelatin methacryloyl (GelMA) hydrogels show enhanced osteogenic differentiation in vitro and promote bone formation in vivo [50]. These findings support stiffness- and matrix-dependent priming in PDLSCs.

DPSCs display a more clearly stratified stiffness response [51]. Defined elastic substrates at 1.5, 15, and 28 kPa have been associated with distinct functional states: softer matrices favor a more soft-tissue-like phenotype, intermediate stiffness supports a cytokine-rich trophic state, and higher stiffness biases cells toward a hard-tissue-like differentiation program [52]. Softer substrates have also been associated with maintenance of quiescence and reduced senescence, whereas increasing stiffness promotes proliferation and osteo-/odontogenic differentiation [53, 54]. These findings suggest that, in DPSCs, stiffness does not simply increase osteogenesis in a linear fashion; rather, it can be used to tune the balance among maintenance, trophic priming, and osteo-/odontogenic commitment.

SCAP exhibit a distinctly non-linear response to stiffness. In representative scaffold-based studies, stiffness values spanning approximately 16–131 kPa produced maximal osteogenic differentiation around 83 kPa, whereas more excessive stiffness dampened differentiation potential [55–57]. Stiffness also biases lineage choice in SCAP: softer matrices favor chondrogenic differentiation, whereas more rigid substrates favor osteogenesis, consistent with stiffness-dependent lineage bias in SCAP [55, 56]. These observations suggest that SCAP may operate within a relatively constrained stiffness window for osteogenic differentiation.

SHED also show stiffness-sensitive behavior, although the literature remains less extensive. In osteoinductive media, stiffer matrices enhance osteogenic differentiation relative to softer substrates [42]. However, most studies of SHED rely on scaffold-based or composite biomaterial systems in which composition, architecture, and bulk mechanical properties, potentially including viscoelastic behavior, are interdependent, making it difficult to isolate the contribution of individual matrix parameters [21, 58, 59].

Taken together, stiffness and viscoelasticity are not generic “more is better” variables. Instead, matrix mechanics appear to define subtype-specific permissive windows for proliferation, trophic function, lineage bias, and maintenance of less differentiated states [42, 47, 50, 53, 55, 60]. This makes them particularly important for rational preconditioning design.

Surface topography and geometrical confinement

Surface topography and geometrical confinement regulate DSC behavior by reshaping cell adhesion, cytoskeletal organization, nuclear geometry, and, potentially, mechanosensitive transcriptional programs. Compared with bulk stiffness cues, these inputs more directly bias cell alignment, polarity, morphology, and lineage tendency through spatial constraint.

In DPSCs, topography and confinement have been most clearly characterized. Nanofiber architectures potentiate Wnt/β-catenin signaling, increase ALP activity and mineralization, and promote odontoblastic differentiation [61]. By contrast, micropattern-induced geometrical confinement preserves stemness-related features and, through nuclear-shape remodeling and altered cytoskeletal organization, has been associated with improved peripheral nerve regeneration in vivo [19]. These findings suggest that topography can be used not only to promote differentiation, but also to preserve therapeutically useful cell states prior to implantation.

SCAP also respond to topographical guidance. Microgrooved substrates (60–120 μm) align and elongate cells and, when combined with appropriate stiffness, can synergistically enhance odontogenic differentiation, as indicated by increased ALP activity [62]. This is particularly relevant in the context of root and mineralized-tissue engineering, where spatial guidance and anisotropic tissue organization may be as important as lineage induction.

In PDLSCs, topographical cues are more often studied in biomaterial contexts linked to attachment, alignment, and matrix remodeling [63–66], whereas in SHED much of the available evidence comes from scaffold systems that mimic aspects of immature ECM architecture [59, 67]. In both cases, topography appears to contribute to functionally relevant cell states, although the evidence is generally less standardized than in stiffness-based studies.

Overall, these findings indicate that topography and geometrical confinement regulate DSC behavior in ways that are not fully captured by stiffness alone [19, 61, 62]. They are especially relevant when preconditioning aims to influence cell alignment, maintenance of plasticity, neural-supportive behavior, or spatially organized differentiation.

Quantitative synthesis and provisional design windows

Although the preceding subsections describe DSC responses according to cue type, the literature becomes more translationally informative when reported parameters are synthesized into comparable quantitative windows rather than presented as isolated experimental conditions (Table 1). Rather than supporting a single optimal mechanical condition across all DSC populations, the available evidence points to cue-specific and subtype-specific working windows in which biological outcomes depend on loading magnitude, temporal program, and matrix context.

Table 1.

Representative quantitative mechanical parameters and predominant reported output classified by biophysical cue in DSCs

Biophysical cue	DSC subtype	Representative quantitative parameters	Predominant reported output	References
Tensile force	PDLSCs	10–12% strain, 0.1–0.7 Hz, 6–24 h	Osteogenic differentiation	[23, 25, 28]
	DPSCs	3–10% cyclic strain, 0.03–0.2 Hz, 10–14 days	Osteogenic differentiation; bone-like ECM program	[33–35]
	SCAP	200 g, 30 min by centrifuge	Osteo-/odontogenic differentiation	[20]
Compressive force	PDLSCs	1.5–2.0 g/cm² static compression, 12–24 h	Compression-side remodeling window	[29, 30, 32]
	DPSCs	19.6 kPa, 0.083 Hz, 9 h; 0.09–0.18 MPa, 30 min/day for 7 days	Odontogenic/osteo-odontogenic differentiation	[38, 39]
FSS	PDLSCs	0.5–5 dyn/cm², 3 h	Immunomodulatory/adaptive response	[13]
	DPSCs	5–12.4 dyn/cm², 24 h–1 h/day, 5 days/week	Osteogenic differentiation	[45, 46]
Hydrostatic pressure	PDLSCs	0–45 to 0–180 kPa, 0.1 Hz, 2 h/day for 3 days	Osteoclast differentiation	[31]
Matrix stiffness	PDLSCs	6–135 kPa	Proliferation and osteogenic differentiation	[47–49]
	DPSCs	75–135 kPa	Proliferation and osteogenic differentiation	[51, 53]
	SCAP	1.7–16.4 kPa	Proliferation	[62]
Surface topography/geometrical confinement	PDLSCs	Nanotubes of 25–50 nm diameter; 30–40 μm grooves plus 100 nm pores	Osteogenic differentiation	[65, 66]
	DPSCs	1024 μm² micropatterns, 1:2 and 1:4	Stemness maintenance	[19]

Several quantitative patterns emerge from this comparison. First, tensile and compressive stimuli should not be treated as a single loading category. In PDLSCs, osteogenic responses are most commonly reported under cyclic strain of approximately 10–12% at 0.1–0.7 Hz for 6–24 h [23, 25, 28]. In DPSCs, by contrast, similar pro-osteogenic or bone-like extracellular matrix responses are reported across a broader tensile range of approximately 3–10% at 0.03–0.2 Hz over 10–14 days [33–35]. Compressive regimens show a different response pattern. In PDLSCs, static compression of approximately 1.5–2.0 g/cm² for 12–24 h is more commonly associated with compression-side remodeling [29, 30, 32], whereas in DPSCs pressure-related stimulation, such as 19.6 kPa or 0.09–0.18 MPa delivered under defined schedules, has been linked to odontogenic or osteo-odontogenic responses [38]. Together, these comparisons indicate that distinct magnitude–time combinations map onto different functional outputs.

Second, fluid-associated cues also show subtype-dependent quantitative tendencies. In PDLSCs, relatively low FSS ranges of approximately 0.5–5 dyn/cm² are associated with adaptive or viability-related responses [13], whereas in DPSCs higher FSS ranges of approximately 5–12.4 dyn/cm² are more often linked to osteogenic differentiation [45, 46]. These findings suggest that fluid-associated cues are not functionally uniform across DSC populations, but instead correspond to different biological objectives depending on cell source and loading context.

Third, matrix stiffness likewise shows a subtype-dependent quantitative pattern. Reported stiffness-associated proliferative or osteogenic windows span approximately 6–135 kPa in PDLSCs [47–49] and 75–135 kPa in DPSCs [51, 53], whereas SCAP appear to respond to a softer range of approximately 1.7–16.4 kPa in studies emphasizing proliferative outputs [62]. Thus, matrix stiffness should be regarded not as a generic “more is better” variable, but as a subtype-dependent design parameter whose biological effects vary according to the target function being prioritized.

Finally, not all mechanically relevant cues are best represented by a single scalar range. In topographical regulation, the key variables are structural dimensions rather than force magnitude. In the available studies, osteogenic tendencies in PDLSCs have been associated with nanotopographical features such as 25–50 nm nanotubes and hierarchical microgroove/nanopore structures [65, 66], whereas defined micropattern geometries in DPSCs have been linked to stemness maintenance [19]. Quantitative synthesis in DSC mechanobiology therefore needs to incorporate both classical loading parameters and geometry-based physical constraints.

Taken together, Table 1 provides representative comparative windows rather than prescriptive engineering rules. These ranges remain provisional because they are influenced by differences in scaffold composition, culture configuration, readout timing, and endpoint selection. Nevertheless, organizing the evidence in this way helps move the field beyond isolated parameter reporting toward a more reproducible, indication-oriented framework for mechanomedicine-guided DSC preconditioning.

Mechanisms of mechanical preconditioning

Mechanical preconditioning translates defined physical inputs applied ex vivo into biochemical and transcriptional responses through interconnected mechanotransduction modules. These modules transmit signals from the cell surface and cytoskeleton to the nucleus, where gene expression, chromatin organization, and potentially epigenetic state are remodeled to shape cell behavior. In the context of DSCs, the most relevant pathways include integrin/focal adhesion kinase (FAK)-dependent force sensing, yes-associated protein (YAP)/transcriptional co-activator with PDZ-binding motif (TAZ)-mediated nuclear mechanotransduction, Wnt/β-catenin signaling, and additional mechanically responsive modules such as Ras homolog gene family member A (RhoA)/Rho-associated kinase (ROCK), Piezo1, and mitogen-activated protein kinase (MAPK) [68–72] (Fig. 2). Rather than acting independently, these mechanotransduction pathways can bias DSCs toward distinct functionally relevant cell states, including periodontal remodeling, osteo-/odontogenic differentiation, and maintenance of neuro-supportive or stemness-associated features [19, 41, 73]. Importantly, these pathways should not be interpreted as uniformly beneficial targets: their activation is highly context dependent and may contribute to maladaptive responses if stimulus type, magnitude, or duration is not appropriately controlled.

Fig. 2.

Mechanotransduction modules linking defined biophysical cues to functionally relevant cell states in DSCs. Defined biophysical inputs are sensed through an integrated mechanotransduction network involving integrin/FAK signaling, actomyosin–RhoA/ROCK contractility, Piezo1-mediated Ca²⁺ influx, YAP/TAZ, Wnt/β-catenin, and MAPK-associated pathways. Through dynamic crosstalk, these modules shape functionally relevant cell states, including matrix-adaptive remodeling, osteo-/fibrogenic bias, osteo-/odontogenic commitment, reparative activation, neuro-supportive plasticity, immunomodulatory tuning, and potentially mechanical memory. Its content is based on findings reported in references [24, 41, 45, 68, 69, 71, 73, 74, 81, 85, 89, 91, 93, 95]. Created with BioRender.com. FRZ Frizzled receptor, LRP5/6 Low-density lipoprotein receptor-related protein 5/6, RTK Receptor tyrosine kinase, RGD Arginine–glycine–aspartic acid, FAT4 FAT atypical cadherin 4, APC Adenomatous polyposis coli, CK1 Casein kinase 1, GSK-3β Glycogen synthase kinase-3β, TCF T-cell factors, LEF Lymphoid enhancing factor, MAPK Mitogen-activated protein kinase, FAK Focal adhesion kinase, MEK MAPK kinase, MKK MAPK kinase kinase, ERK1/2 Extracellular signal-regulated kinase 1/2, JNK c-Jun N-terminal kinase, CaMK Calmodulin-dependent protein kinase, CaN Calcineurin, PI3K Phosphatidylinositol 3-kinase, Akt Protein kinase B, OPG Osteoprotegerin, mTOR Mechanistic target of rapamycin, RANKL Receptor activator of nuclear factor-kappa B ligand, RhoA Ras homolog gene family member A, ROCK Rho-associated kinase, MLC Myosin light chain, MAL Myocardin-related transcription factor, SRF Serum response factor, NPC Nuclear pore complex, YAP Yes-associated protein, TAZ Transcriptional co-activator with PDZ-binding motif, MST1/2 Mammalian sterile 20-like kinase 1/2, LATS1/2 Large tumor suppressor homolog 1/2, SAV1 Salvador homolog 1, MOB MOB kinase activator, TEAD Transcription enhanced association domain, EZH1/2 Zeste homolog 1/2, AP2 Activating protein 2, DNMT1/3A DNA methyltransferase 1/3A, DSPP Dentin sialophosphoprotein, DMP1 Dentin matrix acidic phosphoprotein 1, RUNX2 Runt-related transcription factor 2, ALP Alkaline phosphatase, OCN Osteocalcin, LEF1 Lymphoid enhancing factor

Integrin/FAK pathway

Integrins are heterodimeric ECM receptors that function as primary mechanotransducers linking extracellular mechanics to intracellular signaling [68]. Changes in matrix stiffness or externally applied loading promote ligand engagement, integrin clustering, and focal-adhesion assembly, leading to FAK recruitment and autophosphorylation at Tyr397 [52]. This converts physical cues into biochemical signaling and initiates downstream mechanotransduction cascades. For example, compressive force acting through the integrin β1/FAK axis increases macrophage colony-stimulating factor (M-CSF), RANKL, and TNF-α, while reducing OPG, thereby shifting the RANKL/OPG balance toward bone remodeling [74].

Activated FAK then serves as a central signaling hub that coordinates multiple downstream pathways. Through integrin–actin coupling, it promotes RhoA/ROCK-dependent cytoskeletal tension and focal adhesion maturation, which are commonly associated with osteogenic bias under stiff matrix conditions [68, 75]. In parallel, FAK transduces signals through phosphatidylinositol 3-kinase (PI3K)/protein kinase B (Akt)- and MAPK-related pathways to regulate proliferation and differentiation [76]. Under selected compressive regimens, activation of the integrin β1–FAK–p38 axis has been associated with increased expression of osteogenic markers such as RUNX2, ALP, and OCN, along with enhanced mineralized nodule formation [74, 77].

In the context of DSCs, integrin-mediated sensing provides a direct interface between ex vivo matrix design and cell-state regulation. Viscoelastic hydrogels and stiffness-tunable biomaterials that recapitulate key features of the native periodontal or pulp-associated microenvironment can activate mechanotransduction pathways and enhance osteogenic marker expression in PDLSCs, while promoting odontogenic differentiation and pulp-regenerative responses in dentin–pulp-derived populations such as DPSCs [50, 78]. From a translational perspective, these findings support the use of biomimetic materials not simply as passive carriers, but as controllable platforms for mechanically informed functional priming; however, current scaffold-based studies primarily show that biomaterial mechanics and matrix design can direct DSC behavior and support regenerative outcomes, rather than demonstrating that pre-activation of the integrin/FAK axis is an independent determinant of subsequent in vivo benefit [79, 80].

YAP/TAZ pathway

YAP and TAZ are key mechanosensitive transcriptional regulators in DSC preconditioning because they link defined physical cues to changes in cell state [81]. In general, rigid substrates, tensile loading, and cytoskeletal tension favor YAP/TAZ nuclear localization, whereas softer or mechanically permissive conditions more often promote cytoplasmic retention [49, 69]. Through this force-dependent shuttling behavior, YAP/TAZ link extracellular mechanics to transcriptional programs relevant to cell behavior.

In DSCs, YAP/TAZ are most informative when interpreted in relation to functional outcomes rather than as isolated pathway markers. In PDLSCs, tensile loading promotes actin polymerization, strengthens force transmission to the nucleus, and facilitates YAP nuclear entry, changes that are associated with proliferation and periodontal remodeling [73, 82]. In DPSCs, several physical cues, including mechanically instructive scaffold architectures, stiffness-associated cues, and static magnetic-field exposure, have been linked to YAP/TAZ activation together with changes in proliferation or osteo-/odontogenic differentiation [78, 83, 84]. By contrast, geometrical confinement through micropatterning has been associated with redistribution of YAP away from the nucleus while preserving stemness-related features, indicating that YAP/TAZ activity in DSCs is not simply proportional to regenerative value, but depends on the specific preconditioning goal [19]. Taken together, these observations suggest that YAP/TAZ help determine whether mechanical conditioning favors remodeling, osteo-/odontogenic commitment, or maintenance of a more plastic cell state.

YAP/TAZ function within a broader mechanotransduction network rather than as a standalone pathway. Their activity is closely coupled to integrin-mediated adhesions, actomyosin tension, and RhoA/ROCK signaling, and also interacts with Wnt/β-catenin and other mechanically responsive pathways [71, 85, 86]. For this reason, YAP/TAZ are best interpreted as integrative readouts of mechanical state rather than as binary switches. This distinction is important for translation, because increased pathway activity should not automatically be assumed to be beneficial. In non-dental systems, dysregulated YAP/TAZ activity has been linked to fibrosis, maladaptive remodeling, and tumor-associated processes [87, 88]. Accordingly, future work should define safe and indication-relevant windows of YAP/TAZ engagement during DSC preconditioning, rather than assuming that stronger activation will necessarily improve regenerative performance.

Wnt/β-catenin pathway

Wnt/β-catenin signaling is a mechanosensitive pathway that regulates stromal cell proliferation, lineage specification, and the maintenance of progenitor-like states [89]. Broadly, mechanical cues can promote β-catenin stabilization and nuclear translocation, thereby linking extracellular forces to transcriptional regulation [90]. In the context of mechanical preconditioning, this pathway is most relevant not as an isolated signaling module, but as a mediator through which defined physical inputs shape differentiation and cell-state maintenance.

Mechanically activated Wnt/β-catenin signaling is especially relevant to DSCs because it can couple physical inputs to both differentiation and maintenance of functionally important progenitor states. In PDLSCs, brief static pressure markedly activates Wnt/β-catenin signaling and enhances osteogenesis [24]. In SHED, hydrostatic pressure activates the mechanosensitive channel Piezo1, promotes the expression of WNT-related genes and odontogenesis-associated genes such as PANX3 and DSPP, and supports differentiation toward odontoblast-like cells [41]. At the same time, Wnt signaling is also implicated in maintenance of DPSC stemness-related features. Wnt activity cooperates with metabolic regulation and Notch signaling to support viability and multipotency, and reciprocal activation between Wnt and Notch has been reported to sustain expression of octamer-binding transcription factor 4 (OCT4), SOX2, p75, and activating protein 2 (AP2) in DPSCs [91]. Taken together, these observations indicate that Wnt/β-catenin does not simply drive DSCs toward one fixed differentiated state, but instead helps coordinate the balance between lineage commitment and preservation of therapeutically useful progenitor properties [24, 41, 91].

Crosstalk between Wnt/β-catenin and the integrin/FAK axis further links extracellular mechanics to stromal cell behavior. Through integrin signaling, ECM stiffness can tune the expression of Wnt ligands and inhibitors [90]. From a translational perspective, this pathway is therefore best viewed as a context-dependent regulator of DSC preconditioning rather than as a uniformly beneficial target for maximal activation [24, 92].

Other mechanosensitive pathways

RhoA/ROCK, Piezo1, and MAPK signaling

Additional mechanosensitive pathways, particularly RhoA/ROCK, Piezo1, and MAPK signaling, play complementary roles in DSC mechanoresponses by linking distinct physical cues to cytoskeletal remodeling, calcium-dependent signaling, and phosphorylation-mediated transcriptional control. Rather than functioning as independent pathways, they primarily modulate how DSCs interpret stretch, compression, and fluid-associated stimuli [93]. RhoA/ROCK is most closely associated with force-dependent actomyosin contractility and the regulation of cell shape, adhesion, and lineage bias [85, 94]. Piezo1 functions as a mechanically gated ion channel and is especially relevant under pressure-associated mechanical stimulation, where Ca²⁺ influx contributes to osteo-/odontogenic signaling [41, 95]. MAPK pathways, including extracellular signal-regulated kinase 1/2 (ERK1/2), c-Jun N-terminal kinase (JNK), and p38, further transmit mechanical information to regulate proliferation, differentiation, and inflammatory adaptation in a stimulus-dependent manner [20, 45]. Collectively, these pathways complement the major mechanotransduction modules by fine-tuning contractility, gene expression, and cell-state regulation under defined biophysical conditions.

Integration with canonical mechanotransduction pathways

These pathways are integrated into the canonical mechanotransduction network through several recurring points of crosstalk. RhoA/ROCK cooperates with integrin/FAK signaling to organize focal adhesions and maintain cytoskeletal tension [96]. In DSCs, Piezo1 can interact with YAP- and Wnt-related signaling, thereby providing a mechanism through which force-induced ion flux influences nuclear mechanotransduction and lineage-associated transcription [41, 95, 97]. MAPK signaling likewise intersects with Wnt-associated and other differentiation-related pathways in DSCs, contributing to lineage regulation and adaptive responses [98, 99]. Accordingly, RhoA/ROCK, Piezo1, and MAPK are best regarded as reinforcing modules within a broader signaling network rather than as isolated determinants of DSC fate. Collectively, these interactions help explain how DSCs translate defined physical inputs into coordinated changes in lineage commitment, functional adaptation, and reparative priming.

Integrative perspective

Taken together, these mechanotransduction pathways are best understood not as isolated signaling routes, but as an integrated network through which defined mechanical inputs are translated into functionally relevant cell states. Integrin/FAK signaling and cytoskeletal remodeling are most closely associated with matrix interaction and load adaptation [68]. YAP/TAZ provide a nucleus-oriented transcriptional axis that links mechanical state to cell-state regulation [73, 81]. Wnt/β-catenin integrates physical and developmental cues to balance differentiation with the maintenance of progenitor-related properties [24, 91, 98]. Auxiliary pathways, including Piezo1, RhoA/ROCK, and MAPK, further refine context-dependent outcomes [19, 41]. From a translational perspective, different therapeutic goals are likely to require different mechanistic emphases. Matrix remodeling and osteo-/fibrogenic programs may be most relevant to periodontal and mineralized tissue repair, whereas the maintenance of plasticity, neuro-supportive function, and immunomodulatory tuning may be more important in dentin–pulp and neural applications. Accordingly, the therapeutic value of these pathways depends not simply on whether they can be activated, but on whether they can be engaged in a controlled, safe, and indication-matched manner during ex vivo preconditioning.

Applications of mechanical preconditioning in tissue regeneration

Mechanical preconditioning has emerged as a promising strategy for improving the functional state of DSCs in regenerative applications. By influencing lineage bias, matrix remodeling, reparative priming, and, in some contexts, immunomodulatory function, controlled physical conditioning may enhance the translational potential of DSC-based products [19, 50, 79]. However, the strength of the supporting evidence varies substantially across indications. In some contexts, particularly periodontal regeneration, the rationale for mechanical preconditioning is relatively strong because the target tissue is inherently load-bearing [29, 100]. In other areas, particularly neural repair and broader disease-oriented applications, much of the current evidence derives from biomaterial-defined physical microenvironments rather than from rigorously isolated ex vivo mechanical preconditioning protocols [101–103]. Accordingly, the applications discussed below are organized with attention to the strength of evidence and to the distinction between cell-intrinsic preconditioning effects and combined cell–material system effects (Fig. 3).

Fig. 3.

Representative regenerative applications associated with mechanically preconditioned cell states in DSCs. Applications of mechanical preconditioning in DSCs can be broadly grouped into four domains: periodontal regeneration, bone regeneration, nerve repair, and broader disease-oriented applications. Conceptually, DSCs are exposed ex vivo to defined mechanical cues before being delivered, often within suitable biomaterial carriers, into relevant defect or disease models. Within this framework, mechanical preconditioning may bias lineage commitment and modulate other functionally relevant properties, thereby supporting regeneration or repair in a context-dependent manner. Beyond these core regenerative indications, such effects may also have broader disease-oriented relevance in settings including pathological bone loss and selected neurodegenerative or vascular-related conditions. Its content is based on findings reported in references [19, 25, 26, 34, 50, 79, 101, 102, 108, 116, 122, 126–129]. Created with BioRender.com. DSCs Dental-derived stromal cells, PDLSCs Periodontal ligament stem cells, DPSCs Dental pulp stem cells, SHED Stem cells from human exfoliated deciduous teeth, RGD-Alg Arginine–glycine–aspartic acid-modified sodium alginate, GelMA Gelatin methacryloyl, PCL Polycaprolactone

Periodontal regeneration

Among current application scenarios, periodontal regeneration provides one of the strongest biological rationales for mechanical preconditioning. Successful periodontal repair requires coordinated reconstruction of the soft–hard tissue interface linking cementum, periodontal ligament, and alveolar bone [100, 104]. Because PDLSCs arise from a naturally load-bearing niche and display strong mechanosensitivity, they are particularly relevant candidate cells for this purpose [29]. In principle, mechanically informed conditioning may help preserve matrix-remodeling capacity, improve lineage balance among fibrogenic, cementogenic, and osteogenic programs, and enhance compatibility with the mechanically dynamic periodontal environment after implantation.

Available studies support the regenerative value of PDLSC-based constructs in periodontal repair, although the underlying source of benefit should be interpreted carefully. For example, biomimetic scaffolds with defined viscoelastic or structural properties have been associated with improved soft–hard tissue reconstruction. In one representative approach, a Pluronic F127 diacrylate (F127DA) hydrogel with tunable viscoelasticity was engineered to approximate key mechanical features of the periodontal ligament. Fast-relaxing F127DA hydrogels enhanced the fibroblastic differentiation tendency of PDLSCs in vitro and promoted periodontal ligament repair in a rat delayed replantation model [79]. These findings strongly support the importance of biomimetic matrix mechanics in periodontal regeneration. However, they do not by themselves establish that ex vivo mechanical preconditioning alone directly determines the in vivo regenerative outcome.

Direct loading studies provide additional support for the relevance of mechanical conditioning to periodontal repair. Cyclic tensile loading at 12% has been shown to increase ALP activity and upregulate osteogenic markers such as RUNX2 and OCN in PDLSCs, consistent with activation of osteogenic pathways relevant to alveolar-bone-associated repair [25, 26]. Together, these findings suggest that mechanical conditioning may influence structural regeneration in the periodontal setting.

At the same time, this field would benefit from more rigorous separation of preconditioning effects from implant-environment effects. Many current studies use scaffolds or delivery systems whose own viscoelastic and structural properties continue to regulate cell behavior after implantation [50, 79, 105]. Thus, while the periodontal literature is among the strongest in this review, future work should explicitly compare preconditioned versus non-preconditioned cells under matched implantation conditions and should assess not only osteogenesis, but also fibrous attachment quality, load adaptation, and immunomodulatory potency. These issues are especially relevant given the early translational progress already achieved by cell-based and 3D-printed biomimetic approaches in periodontal regeneration [105–107]. Systematic incorporation of standardized mechanical preconditioning into the manufacturing workflow may therefore represent a realistic next step toward more stable and predictable regeneration of periodontal soft–hard tissue interfaces.

Bone regeneration

Mechanical preconditioning may also enhance the osteogenic potential of DSCs for maxillofacial and bone-defect repair. In vitro, cyclic loading of DPSCs cultured on scaffolds in a bioreactor (10% strain, 0.2 Hz) significantly upregulated osteogenesis-related genes and proteins, including collagen I and bone sialoprotein (BSP), and promoted matrix mineralization [34]. Likewise, dynamic loading generated by turbulent flow in a spinner-flask bioreactor increased osteogenic differentiation of DPSCs on silk fibroin scaffolds, as indicated by greater mineralized matrix formation and calcium deposition [108].

These studies support the idea that dynamic culture and loading conditions can enhance osteogenic bias in DSC populations, particularly when integrated with scaffold-based systems [34, 108, 109]. However, most bone-regeneration studies condition cells within mechanically active scaffold or bioreactor systems that continue to influence later outcomes. Accordingly, the observed benefit often reflects combined system effects rather than isolated preconditioning alone.

A representative example is the use of bioprinted GelMA hydrogels to support the osteogenic potential of PDLSCs [50, 110, 111]. In this system, PDLSC-laden GelMA constructs of different hydrogel concentrations were implanted into a mouse calvarial defect model [112]. Higher-concentration GelMA constructs exhibited greater stiffness, smaller pore size, lower swelling, and slower degradation. These properties were associated with enhanced osteogenic differentiation of PDLSCs in vitro and greater new bone formation in vivo [50]. These findings are highly encouraging, but they most directly support the value of mechanically informed construct design rather than proving that prior ex vivo priming alone is sufficient to accelerate bone healing.

Overall, the bone-regeneration literature suggests that mechanically dynamic culture systems can enhance the osteogenic competence of DSCs, particularly when integrated with biomaterial and bioreactor platforms [34, 50, 108, 109]. For translation, the next step will be to determine which loading regimens generate robust, reproducible osteogenic priming before implantation, and how these regimens can be standardized independently of scaffold-specific effects.

Nerve repair

DPSCs are derived from the neural crest and exhibit an intrinsic propensity for neurogenic differentiation. This property has generated considerable interest in their application to peripheral and central nervous system repair. In particular, DPSCs can promote axonal regeneration and remyelination, especially when combined with biomaterial systems such as collagen scaffolds, silicone conduits, or poly(lactic-co-glycolic acid) (PLGA) conduits that support axonal growth, myelination, and functional recovery after facial nerve injury [103, 113–115]. SHED, which are likewise neural crest-derived, also show robust neurogenic potential. For example, three-dimensional bioprinted arginine–glycine–aspartic acid (RGD)-modified sodium alginate/GelMA/polycaprolactone significantly enhanced neuroregeneration when implanted with Schwann-like cells induced from SHED [116]. Taken together, these findings support the importance of mechanically and structurally appropriate microenvironments in neural repair.

Direct evidence for ex vivo mechanical preconditioning remains limited, but the available findings are increasingly encouraging. In the 2023 micropattern-based “stem cell gym” study, adhesive-island training conferred durable mechanical memory on DPSCs, preserved stemness, reduced senescence, and improved functional and histological outcomes after transplantation into a rat sciatic nerve defect model [19]. Complementary scaffold-centered studies further highlight the importance of mechanically organized neural microenvironments. For example, aligned scaffold-free DPSC sheets promoted and guided axonal extension, scaffold-free conduits engineered from DPSCs and their endogenous aligned ECM enhanced facial nerve regeneration, and customized three-dimensional nanofibrous scaffolds supported DPSC neural differentiation for peripheral nerve repair [117–119]. These findings provide some of the clearest current evidence that ex vivo mechanical preconditioning itself may enhance DPSC survival and neuroregenerative function in a complex injury setting.

Additional studies provide complementary support, although most remain scaffold-centered rather than true preconditioning studies. For example, mechanically favorable nanofiber or composite scaffolds can enhance neurogenic marker expression or anti-inflammatory function in DPSCs [120, 121]. However, these findings primarily suggest the importance of physical microenvironment design rather than isolated ex vivo mechanical training. Accordingly, future work should distinguish more rigorously between conditioning history and implant microenvironment, define how long advantageous mechanically induced states persist after transplantation, and test whether multimodal strategies combining scaffold mechanics, electrical stimulation, and ex vivo priming provide synergistic benefit.

Broader disease-oriented applications

Beyond periodontal regeneration, bone regeneration, and nerve repair, DSCs have also been explored in broader disease-oriented applications [101, 102, 122]. In periodontitis-related bone loss, exosomes enriched in miR-200b/c released by mechanically stimulated PDLSCs were reported to be internalized by human mandibular bone marrow MSCs under tensile loading, thereby enhancing alveolar bone defect repair [101]. In neuroregeneration-related contexts, DPSCs respond to biomaterial topography and other matrix-associated cues by adopting neuronal or neuron-like phenotypes and by providing neurotrophic support [123–125]. These properties have generated interest in DPSC-based approaches for stroke, Alzheimer’s disease, Parkinson’s disease, and spinal cord injury [126–129]. In cutaneous repair, SHED encapsulated within rapidly gelling palmitoyl tetrapeptide-7/chitosan hydrogels promoted angiogenesis and accelerated diabetic wound healing [102].

These studies are encouraging; however, they remain exploratory and should not yet be regarded as mature translational indications for mechanical preconditioning [101, 102]. In most cases, the reported benefits are attributable to biomaterial design, paracrine activity, or proof-of-concept observations rather than to rigorously defined ex vivo conditioning protocols, which limits their value for establishing clear translational principles. Future progress will depend on disentangling cell-intrinsic conditioning effects from scaffold- and host-mediated influences and on defining indication-specific potency assays that accurately capture therapeutic function.

Taken together, the application literature remains uneven in maturity. Periodontal regeneration has the strongest tissue-specific rationale and the most coherent support for mechanically informed conditioning, whereas the evidence for bone regeneration and nerve repair remains largely preclinical and broader disease-oriented applications remain exploratory. A key limitation is that many reported benefits cannot yet be attributed unequivocally to prior ex vivo conditioning itself, because scaffold mechanics, delivery systems, and host microenvironments continue to influence cell behavior after implantation. This distinction should be considered when interpreting the current translational promise of mechanical preconditioning in DSCs.

Challenges and future perspectives

Despite growing interest in the mechanical preconditioning of DSCs, substantial barriers continue to limit its translation into reproducible clinical use. These challenges include incomplete standardization, donor- and population-level heterogeneity, cellular senescence, uncertainty regarding the persistence of mechanically induced phenotypes, safety considerations, and the lack of GMP-compatible manufacturing workflows and functionally relevant potency frameworks. Addressing these issues will be essential for advancing mechanically preconditioned DSCs from experimental promise to clinically meaningful regenerative products.

Standardization and parameter optimization

One of the most immediate barriers to translation is the lack of standardization in mechanical-conditioning protocols [130, 131]. Although the benefits of mechanically informed culture are increasingly evident, heterogeneity in stimulation parameters—including cue type, magnitude, duration, frequency, waveform, matrix context, and exposure schedule—continues to limit cross-study comparability and prevent the emergence of consensus practice [45, 55]. This problem is further compounded by differences in culture context and readout strategy. To improve reproducibility, future studies should report not only cell source and biological output, but also device calibration, matrix properties, loading waveform, exposure timing, and quantitative parameter ranges in a more systematic way [18, 132, 133].

Importantly, the goal should not be to define a single universal “optimal” condition for all DSCs. The current evidence instead supports the concept of cell-type-specific working windows, in which biological output depends on both cue category and the combination of magnitude, frequency, duration, and matrix context [25, 45, 55]. Standardization should therefore focus on comparability, reproducibility, and transparent reporting, while preserving the flexibility needed for subtype-specific and indication-specific design.

Conflicting findings, experimental limitations, and pathway-related risks

Although mechanical preconditioning is often presented as a beneficial strategy to enhance DSC function, the current literature is far from uniformly positive. Mechanical effects are highly context dependent and may become maladaptive when stimulus type, magnitude, duration, or temporal pattern are not properly matched to cell identity and culture conditions. In some settings, mechanical loading promotes osteo-/odontogenic differentiation, whereas in others it enhances proliferation at the expense of lineage commitment, triggers inflammatory signaling, or reduces stromal potency [36, 134]. Such bidirectional responses indicate that mechanical stimulation should not be interpreted as intrinsically regenerative, but rather as a powerful regulator whose effects depend on biological and engineering context.

A second limitation is that many studies differ substantially in cell source, donor age, passage number, scaffold composition, loading platform, and readout timing [130, 135, 136]. As a result, apparently similar regimens may produce divergent outcomes, and favorable results observed in reductionist in vitro systems may not translate directly to implantation settings. This issue is particularly important in studies that combine biomaterials with dynamic culture, because the final phenotype often reflects the integrated action of mechanical conditioning, scaffold mechanics, biochemical signaling, and post-implantation environmental cues, rather than a single isolated preconditioning effect [79, 108, 109].

Potential safety concerns also deserve explicit attention. Excessive or prolonged loading may enhance proliferation at the expense of lineage commitment, favor inflammatory or osteoclastogenic programs, and reduce regenerative potency [37, 134, 137]. Moreover, pathways frequently highlighted as favorable mechanotransduction targets, including YAP/TAZ and Wnt/β-catenin, are highly context dependent and have been linked in other systems to fibrosis, maladaptive remodeling, or tumor-associated processes when dysregulated [87, 88, 138]. For this reason, future studies should define safe and indication-relevant windows of pathway engagement, rather than assuming that stronger activation will necessarily improve regenerative performance.

Senescence, persistence, and long-term stability

Long-term stability and senescence remain under-addressed in the current DSC preconditioning literature. Mechanical stimulation can remodel cytoskeletal architecture, focal adhesions, nuclear mechanics, chromatin organization, and epigenetic state [15, 139]. These changes may be beneficially instructive, but they may also become maladaptive if excessively prolonged or applied to vulnerable cell populations. Future studies should therefore assess not only immediate efficacy, but also genomic stability, long-term phenotype maintenance, and whether mechanically induced states remain functionally advantageous after implantation.

Senescence represents a particularly important dimension of this stability problem. Depending on context, mechanical stimulation may preserve desirable functional features, but if improperly scaled it may also contribute to stress accumulation and senescence-associated phenotypes [137, 140, 141]. This concern intersects with donor age, ex vivo expansion history, oxidative burden, cryostorage, and the persistence of the preconditioned state [130, 135, 137, 140, 142]. Importantly, senescence-related risks are unlikely to affect all DSC populations equally and may differ across PDLSCs, DPSCs, SCAP, and SHED [135]. Rigorous future studies should therefore evaluate senescence-associated readouts and long-term functional stability under defined conditioning regimens, rather than limiting assessment to short-term proliferation or differentiation markers alone.

The persistence of mechanical memory is another unresolved issue. It remains unclear how long advantageous mechanically induced phenotypes endure after transplantation, whether these states are reversible, and to what extent they are overwritten by inflammatory, biochemical, and mechanical cues in the host environment [19, 143]. Resolving these questions will be essential for determining whether ex vivo preconditioning produces durable therapeutic benefit or only transient phenotypic drift.

Potency assessment, donor variability, and immunomodulatory function

A major translational limitation of the current literature is that many studies still rely heavily on lineage markers while paying comparatively less attention to application-relevant functional assays [144, 145]. Yet for regenerative therapy, the most important outputs may include not only osteogenic or neurogenic differentiation, but also host adaptation, matrix remodeling, secretory behavior, and immunomodulatory function [8, 79, 146, 147]. This is especially relevant in periodontal applications, where excessive inflammation can directly compromise regeneration [146].

Future studies should therefore incorporate indication-specific potency assays. For inflammation-associated periodontal repair, functionally informative immunomodulatory readouts—such as interferon-γ-inducible IDO activity—may be more useful than phenotypic marker panels alone [148]. In neural applications, neuro-supportive secretome activity or maintenance of neuro-supportive plasticity may be more relevant than neuronal marker expression alone [147, 149]. In bone and other mineralized-tissue contexts, potency assessment may need to include matrix deposition, mineralization competence, and integration-associated remodeling potential [34, 50].

Donor variability and stromal cell heterogeneity further complicate this issue. Since standard isolation procedures generate heterogeneous, non-clonal populations, baseline mechanosensitivity, secretory profile, and regenerative potential may vary substantially across donors and subpopulations [135, 150]. These sources of variability must be incorporated into future study design, potency testing, and manufacturing strategy if mechanically preconditioned DSC products are to achieve reproducible clinical performance [130].

Smart biomaterials and dynamic mechanoculture systems

Next-generation smart biomaterials provide an important opportunity to implement mechanically informed preconditioning with higher precision and temporal control. Programmable hydrogels can deliver phase-specific stiffness or stress profiles that emulate aspects of tissue repair, while responsive materials can dynamically alter cell-experienced mechanics during expansion or differentiation [151, 152]. Such systems are particularly attractive because they allow mechanical conditioning to be embedded within a more biomimetic and controllable culture platform.

Dynamic culture systems are equally important. Programmable bioreactors provide reproducible control of FSS, cyclic strain, compression, or multimodal loading, thereby enabling more standardized mechanical training [153]. Perfusion systems, spinner-flask platforms, and microfluidic devices may further improve nutrient transport, enable real-time modulation of conditioning regimens, and support integration of mechanical cues with biochemical or electrical stimulation [154]. Together, such mechanoculture platforms may improve the reproducibility, functional consistency, and scalability of mechanically preconditioned DSC products [132, 153]. However, future work should also distinguish carefully between the effects of true preconditioning history and those of continued mechanical regulation by the culture or delivery platform itself.

Toward GMP-compatible mechanobiomanufacturing

Successful clinical translation will require incorporation of mechanical preconditioning into GMP-compatible manufacturing as a controlled unit operation rather than as a loosely defined laboratory enhancement step [132]. This will require standardized cell sourcing, expansion conditions, conditioning devices, in-process monitoring, release criteria, and transport or delivery, together with explicit definition of critical process parameters (CPPs) and critical quality attributes (CQAs), management of donor variability and batch-to-batch inconsistency, post-thaw recovery assessment, and lot-to-lot comparability [130, 131, 155–157] (Fig. 4). For indication-specific products, release criteria should extend beyond phenotypic profiling to include indication-specific potency assessment [133, 144, 158]. Clinical implementation will also require regulatory-compliant process validation, predefined acceptance criteria for sterility, mycoplasma, endotoxin, viability, and functional potency, and scalable closed or semi-closed bioreactor platforms to improve reproducibility and reduce contamination risk [132, 155, 157].

Fig. 4.

Translational framework for GMP-compatible manufacturing of mechanically preconditioned DSCs. This figure illustrates a translational workflow for incorporating mechanical preconditioning into the standardized manufacturing of DSC products. The upper panel summarizes the major operational stages, including cell sourcing, ex vivo expansion, controlled mechanical conditioning, in-process monitoring, release testing, transport or delivery, and indication-specific potency assessment. The central panel highlights cross-cutting requirements for process control and standardization, including CPPs, CQAs, batch consistency, sterility assurance, and regulatory traceability. The lower panel depicts the progression toward scalable biomanufacturing, preclinical validation, and clinical translation. Collectively, the figure emphasizes that the clinical value of mechanomedicine-guided preconditioning will depend not only on biological efficacy, but also on whether the induced cell state can be generated reproducibly, functionally validated, and integrated into regulator-ready manufacturing workflows. Its content is based on findings reported in references [130–133, 155–158]. Created with BioRender.com. CPPs Critical process parameters, CQAs Critical quality attributes

Future optimization will likely require multimodal strategies, but clinical translation will depend on whether such combinations can be standardized, validated, and linked to mechanism-aligned potency outputs. At present, a major limitation of the field is the lack of direct clinical evidence showing that mechanically preconditioned DSCs consistently outperform conventionally expanded DSCs under standardized manufacturing and implantation conditions. Moreover, because the available studies remain highly heterogeneous, the quantitative ranges summarized here should be regarded as representative working windows rather than definitive therapeutic specifications.

Accordingly, future progress will depend less on demonstrating mechanosensitivity in DSCs per se than on defining which cues, at which magnitudes, in which cell populations, and for which therapeutic indications can reproducibly generate mechanically induced cell states that are durable, safe, functionally relevant, and clinically meaningful. Achieving this goal will require parameter-defined conditioning protocols, mechanism-aligned potency assays, longer-term in vivo evaluation, and potentially computational modeling to support parameter selection across broad design spaces.

Abbreviations

MSCs

Mesenchymal stromal cells

DSCs

Dental-derived stromal cells

PDLSCs

Periodontal ligament stem cells

DPSCs

Dental pulp stem cells

SCAP

Stem cells from the apical papilla

SHED

Stem cells from human exfoliated deciduous teeth

FSS

Fluid shear stress

GMP

Good manufacturing practice

RUNX2

Runt-related transcription factor 2

ALP

Alkaline phosphatase

OCN

Osteocalcin

RANKL

Receptor activator of nuclear factor-kappa B ligand

TNF-α

Tumor necrosis factor-α

IL-6

Interleukin-6

ECM

Extracellular matrix

OPN

Osteopontin

OPG

Osteoprotegerin

DSPP

Dentin sialophosphoprotein

DMP1

Dentin matrix acidic phosphoprotein 1

SLURP-1

Secretory mammalian Ly-6 urokinase-type plasminogen activator receptor-associated protein-1

α7nAChR

α7 nicotinic acetylcholine receptor

NF-κB

Nuclear factor-kappa B

OSX

Osterix

IDO

Indoleamine 2,3-dioxygenase

BMP-2

Bone morphogenetic protein-2

PANX3

Pannexin 3

GelMA

Gelatin methacryloyl

FAK

Focal adhesion kinase

YAP

Yes-associated protein

TAZ

Transcriptional co-activator with PDZ-binding motif

RhoA

Ras homolog gene family member A

ROCK

Rho-associated kinase

MAPK

Mitogen-activated protein kinase

M-CSF

Macrophage colony-stimulating factor

PI3K

Phosphatidylinositol 3-kinase

Akt

Protein kinase B

OCT4

Octamer-binding transcription factor 4

AP2

Activating protein 2

ERK1/2

Extracellular signal-regulated kinase 1/2

JNK

c-Jun N-terminal kinase

F127DA

Pluronic F127 diacrylate

BSP

Bone sialoprotein

PLGA

Poly(lactic-co-glycolic acid)

RGD

Arginine–glycine–aspartic acid

CPPs

Critical process parameters

CQAs

Critical quality attributes

Author contributions

QSS and YFR conceived and supervised the study. XYP and LY conducted the primary literature review, drafted the manuscript, prepared the figures and table, and led the revision process. JQZ contributed to manuscript revision and helped address the editor’s and reviewers’ comments. GXX, YXJ, GWL, YYW, XLW, and HXL assisted with literature collection, reference management, and manuscript review. All authors read and approved the final manuscript.

Funding

This work was supported by grants from the China Postdoctoral Science Foundation (No. 2025M781691), the Natural Science Foundation of Education Department of Henan Province, China (No. 26A130001, 23A310012), the Foundation for Young Scholar in School of Stomatology of Henan University (No. HUSSYS2024008), the National Natural Science Foundation of China (No. 82302350, 82404210) and the Foundation of Science & Technology Department of Henan Province, China (No. 242102311151).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.