The relationship between substrate topography and stem cell differentiation in the musculoskeletal system

Abstract

It is well known that biomaterial topography can exert a profound influence on various cellular functions such as migration, polarization, and adhesion. With the development and refinement of manufacturing technology, much research has recently been focused on substrate topography-induced cell differentiation, particularly in the field of tissue engineering. Even without biological and chemical stimuli, the differentiation of stem cells can also be initiated by various biomaterials with different topographic features. However, the underlying mechanisms of this biological phenomenon remain elusive. During the past few decades, many researchers have demonstrated that cells can sense the topography of materials through the assembly and polymerization of membrane proteins. Following the activation of RHO, TGF-b or FAK signaling pathways, cells can be induced into various differentiation states. But these signaling pathways often coincide with canonical mechanical transduction pathways, and no firm conclusion has been reached among researchers in this field on topography-specific signaling pathways. On the other hand, some substrate topographies are reported to have the ability to inhibit differentiation and maintain the ‘stemness’ of stem cells. In this review, we will summarize the role of topography in musculoskeletal system regeneration and explore possible topography-related signaling pathways involved in cell differentiation.

Introduction

Stem cell differentiation is a complex process that involves interactive cross-talk between various biological molecules such as growth factors, hormones and their corresponding cell surface receptors. Apart from biological and chemical stimuli, physical stimuli can also play a crucial role in stem cell differentiation. In recent years, many researchers are beginning to focus on the effects of physical stimuli on cell differentiation. Physical stimuli can be categorized into externally applied forces such as hydrostatic pressure, fluid shear stress, bending, tension, and compression [1]; and internal forces including substrate rigidity [2], and topography [3, 4]. Substrate topography has attracted the most attention amongst these. It is conceptualized that ECM presents a variety of geometrically defined three-dimensional (3D) or two-dimensional (2D) physical cues at the micron and submicron scale [5–7]. In vivo (3D) topographies refer to the geometric features present in the extracellular matrix of the stem cell niche or surrounding cells, while in vitro (2D) topographies refer to the bottom-up geometric features presented by the substrate on which the stem cells are seeded on. However, topography, as a type of mechanical property, originates from the natural stem cell niche, with most researchers holding the opinion that biomimetic topography can facilitate stem cell differentiation into terminally differentiated cell phenotypes. Some substrate properties such as rough and smooth structures can also be classified as topography. Synthetic substrate topographies can be categorized into several basic types including gratings, posts and arrays [8]. The substrate topographies can be sensed by cells through a series of intracellular signaling and cascade amplification pathways, with topography being able to profoundly influence adhesion, proliferation, differentiation and other cellular functions. But the exact effects of substrate topography on cell differentiation remain to be established. In this review, we will critically examine the effects of topographies on stem cell differentiation within the musculoskeletal system, with the major focus being on the specific mechanisms of topography-induced cell differentiation.

Topography-induced stem cell differentiation

There has been remarkable progress in utilizing embryonic and adult stem cells in tissue engineering, given their extensive self-renewal and differentiation capacities [9–12]. However, the inability to precisely control stem cell fate has been a limitation in realizing the full therapeutic potential of these cells. The in vivo microenvironment of stem cells contains highly dynamic and intricate biochemical and biophysical cues, each playing their own role in synergizing and regulating stem cell differentiation [11–13].

Topographies that induce stem cells into osteocytes

Bones are highly vascularized tissues with an intrinsic capacity to remodel and undergo self-healing. It is relatively easy to induce stem cells to differentiate into bones as compared with tendon cells and cartilage cells. Upon comparing smooth surfaces to rough topographies, it was found that rough topographies are more conducive for promoting human osteogenic cell differentiation (Ra 3.2 μm vs. Ra 40 nm) [14]. Henceforth, various rough topographies including grating, grid [14], and porous topographies [15] have been successfully utilized to promote osteogenic differentiation. Grating is the most widely used to promote osteogenic differentiation amongst these topographies. It can induce osteogenic marker expression even in the absence of an osteogenic inducer. For example, Matthew et al. found that on groove topographies (5 μm or 50 μm width), the osteoprogenitor cells orientate along the grooves while an OC positive layer orientate perpendicularly to the grooves. In contrast, on pit topographies (30 μm or 40 μm width) mature nodule formation was observed with cells forming a multilayered rim around an intensely OC positive center [16]. Similarly, Vega et al. found that co-continuous (ribbons) topography can promote osteogenic marker expression by human mesenchymal stem cells, as compared to discontinuous (islands and pits) topography after 14 days of culture (16 ± 1 µm, 48 ± 5 µm, and 151 ± 13 µm, respectively) [17]. However, stem cell differentiation on different topological surface is quite complex, with varying results. Abagnale et al. used nanograting to induce MSC differentiation and found that 15 mm microridges promoted adipogenic differentiation whereas 2 mm microridges promoted osteogenic differentiation [18]. By contrast, in the study of Ahn et al. which used nano-posts to induce hMSC (human mesenchymal stem cells) differentiation, it was found that greater distances (5.6 μm) between nano-posts were more conducive to osteogenic (1.2 μm) differentiation, whereas the smaller post-to-post distances were more conducive to adipogenesis [19]. Zouani et al. found out that increasing the depth of nanogratings can promote the differentiation of human mesenchymal stem cells selectively into osteoblasts (10 nm vs. 100 nm) [20], but the specific mechanisms need to be further studied [21]. In summary, topographic dimensions that promote stem cell differentiation into osteocytes are relatively big, as compared with other components of the musculoskeletal system that are often in the micro-scale.

Topographies that induce stem cells into tendon cells/tenocytes

Topographical features have been shown to promote tendon-associated gene expression in various adult stem cells for example in tendon stem/progenitor cells (TSPCs), bone marrow mesenchymal stem cells (bMSC), encapsulated mesenchymal stem cells (eMSc) and even in culture medium without a corresponding inducer. Most of the tendon-associated topographies are aligned structures, which are different from the topological structures that are conducive to bone differentiation. This may be related to the natural structure of the tendon, which is aligned along the long axis of the ligament and display a wave-like pattern along their length [22]. The aligned topographies simulate the natural structure of the tendon, thus promoting tendon differentiation. The aligned structures can be fabricated in many ways. Most studies utilize the electrospinning technique to produce aligned topographies [23]. Yin et al. used aligned PLLA nanofibers to promote tendon regeneration (MSC) in vivo and align the topography successfully to promote tendon regeneration (1068 ± 190 nm), as compared to randomly oriented PLLA scaffolds (739 ± 129 nm) [24]. Because these techniques tend to use PLLA/PLGA as raw materials, these scaffolds often have poor biocompatibility and high hydrophobicity. Therefore, researchers often add other materials (such as gelatin) on these synthetic surfaces to increase their biocompatibility. Another widely used type of scaffolds for tendon tissue engineering is collagen-based scaffolds. These scaffolds have good biocompatibilities and high porosities and it is easy for cells to penetrate through them. Zheng et al. used freeze casting to produce an aligned collagen scaffold (parameter is not available) and found that MSCs seeded on those aligned collagen scaffolds became slenderer and appeared more mature in an in vivo study. This thus confirmed that aligned scaffolds can promote rotator cuff regeneration and the mechanism by which aligned collagen fibers can simulate natural anterior cruciate ligament (ACL) structures, thus promoting ACL regeneration [25]. The main drawbacks of these scaffolds are the poor mechanical properties and fast degradation rate. So, researchers often add some other materials such as silk, or they use some special weaving techniques to increase the mechanical strength [26]. Overall, topographies used to promote tendon differentiation are mostly aligned structures, with most of such topographies having parameters similar to natural collagen fibers (dimensions between 1 and 20 μm [27]) to promote tendon regeneration.

Topographies that induce stem cells into cartilage

As a non-vascularized tissue, self-regeneration of the cartilage is quite difficult. In the past few decades, various new techniques such as 3D printing have been used to promote cartilage regeneration and have achieved much success. These topographies have also been demonstrated to promote cartilage regeneration. Savaiano et al. found out that nanoscale (32 nm) PLGA/nanophase titanic composites can better promote protein-68 and alkaline phosphatase expression in cartilage chondrocytes, as compared to micro-scale (4120 nm) topography [28]. Some researchers have used chitosan–polybutylene succinate (CPBS) scaffolds to promote bovine articular chondrocytes cartilage regeneration and realized that the random topographies with large pore sizes (276.8 ± 52.5 μm) improved the expression of proteoglycans and collagen type II that are associated with cartilage differentiation, as compared with small pore sizes (199.3 ± 5.3 μm) [29]. Ferlin et al. also discovered that ordered cubic pore topographies (pore volume of 1 mm³) significantly enhanced the expression of genes associated with MSC adipogenesis and chondrogenesis of MSC, as compared to topography with ordered cylindrical pores (pore volume of 1 mm³) [30]. Da Silva et al. also used electrospun nanofibrous meshes and realized that Wharton’s jelly stem cells cultured on the electrospun nanofibrous meshes (pore size of 2.7 μm) displayed a higher expression level of cartilage-related genes [31]. Chen et al. also used the electrospinning technique to produce scaffolds with different surfaces that are rough and found out that a lower surface roughness of (Ra = 14.3 ± 2.5 nm) is more conducive for promoting chondrogenic gene expression of hMSCs on day 7, as compared to a higher surface roughness of (Ra = 71.0 ± 11.0 nm) [32]. Shafiee et al. used aligned and random electrospun poly-l-lactide (PLLA)/polycaprolactone (PCL) hybrid scaffolds and discovered that aligned topographies can promote chondrogenic differentiation of human nasal septum-derived progenitors (parameter is not available) [33]. Some researchers seeded mesenchymal stem cells (MSCs) on vertically aligned TiO₂ nanotubes and found out that 100 nm nanotubes can better promote chondrogenic differentiation, as compared to 15 nm nanotubes, which induced spreading and even dedifferentiation of chondrocytes [34]. Overall, various topographies can be used to promote cartilage differentiation, but pore geometries are most commonly used in cartilage differentiation, probably because pore geometries can mimic cartilage lacuna, which is very essential for cartilage development. Additionally, cartilage-related topographies are relatively small and are often in the nanoscale, as compared with topographies that induce stem cells into osteocytes.

Topographies that induce stem cells into myocytes

Muscle tissues can be categorized into three different types: skeletal, heart and smooth muscle cells. Unlike cartilage and tendons, the muscular system itself possesses some intrinsic regenerative capacity. This regenerative capacity has a certain threshold and some serious injuries resulting from road traffic accidents or firearms may induce massive muscle defects that cannot heal completely. Eventually this will affect the function of the joints. Therefore, the regeneration of the musculoskeletal system has very important clinical significance. Currently, many researchers have succeeded in enhancing regeneration of muscles using appropriate biomaterial topographies. Huang et al. found out that nanofibrous and micropatterned polymers can be utilized in the morphogenesis of myoblasts. These topographic features can regulate cellular and cytoskeletal alignment, myotube assembly, myotube striation and myoblast proliferation (the average diameter is approximately 500 nm with an average gap size of approximately 4 µm) [35]. Similarly, Murray et al. used tissue culture grade polystyrene (TCPS), glass and cured polydimethylsiloxane (PDMS) substrates, and it was observed that these TCPS substrates promoted the differentiation of myotubes better than PDMS substrates. Furthermore, myoblasts were found to be differentiating and elongating along the pattern length. Their alignment and differentiation were aligned along the grating patterns (parameter is not available) [36]. Nevertheless, much more research still needs to be done before clinical applications can be realized.

Nanoscale and micro-scale, which is better?

As seen above in most studies, micro-scale topographies have better effects in promoting stem cell differentiation, as compared with nanoscale topographies in our opinion. For example, Zhou et al. demonstrated that the expression of tendon differentiation marker genes (scleraxis and tenascin-C) of adipose-derived stem cells was only upregulated with micro-scale topography (groove width of 10 µm and groove depth of 3 µm) after 7 days of culture, as compared with the nanoscale (2211 ± 538 nm) control [37]. This may be partly due to the fact that the structural dimensions of stem cells are often at the micro-level. For example, mesenchymal stem cells, which are the most widely used cell type in tissue engineering, have diameters ranging from 15 to 30 μm [38], while the diameters of somatic cells such as cartilage cells range from 10 to 12 μm [39]. Furthermore, most of the biological macromolecules in ECM are also at the micro-level; for example, collagen fibers, which dimensionality is within the region of 1–20 μm [27]. Micro-scale topography is more similar to the natural ECM component, while nanoscale topography may be too small for filopodia and lamellipodium to sense and, thus, cannot stimulate downstream signaling pathways to promote further cell differentiation. Although many researchers have claimed that nanoscale topography can promote stem cell differentiation, most of these topographies are more than 1000 nm, and the dividing line between micro- and nanoscale topography is still unclear. Currently, 10 nm is the smallest dimension to be shown to affect cell behavior [40]. Gui et al. showed that cells can discern topographic pore sizes within the range of 100–350 nm [41]. Some researchers categorize topographic length scales between 10 nm and 3 µm as nanoscale topography for simplicity [42]. Many recently published topography-related studies utilize the term “submicron” to define the topography they use. May be, the bionic topography of several hundred nanometers is more suitable for stem cell differentiation. However, due to the differences in microenvironment between different components of the musculoskeletal system, the most suitable parameters for different lineages of the musculoskeletal system are different. Large topography may be more suitable for bone (1–10 μm) and muscle (200 nm–10 μm) regeneration, whereas cartilage (15 nm–3 μm) and tendon (600–1000 nm) can regenerate better with smaller topography (Table 1). In summary, we believe that biomimetic topography at the submicron scale is more suitable for musculoskeletal system regeneration. Interestingly, several papers published this year showed that micro- and nano-level topographies promote cell differentiation through different mechanisms. For example, a recently published article found that micron grid topography (246 ± 3 nm) was beneficial for cell colonization, whereas the nano-grid topography (66 ± 2 nm) was beneficial for cell locomotion [43]. Similarly, Chen et al. found that micron grooves (data is not available) can affect cellular alignment by modifying the initial cell polarization, whereas submicron grooves (data are not available) affected cellular extension [44]. Some researchers found that micro- and nano-structures may activate integrins in different ways, so that the combination of micro- and nano-structures has synergistic activation effects [45]. Based on this hypothesis, many researchers prefer to use micro/nanoscale hybrid topographies to promote stem cell differentiation and achieve good results [46, 47], but the specific mechanism needs to be further studied.

Table 1.

Characteristics of several topographies that induce musculoskeletal system related stem cell differentiation

Feature type	Methods	Material	Parameter	Cell type	Phenomenon	Title of article
Osteogenic differentiation
Smooth and rough substrates	NA	Ti	Smooth [Ra < 0.4 μm], Rough [Ra ≥ 3.4 μm]	Human mesenchymal stem cells, MG63 osteoblast-line cells	Rough surfaces promote osteogenic differentiation in comparison to smooth surfaces	[118]
Microgrid topography and non-patterned substrates	Photolithographic techniques	Polydimethylsiloxane (PDMS)	3 μm pattern width, 2 μm pattern interval length, 7 μm pattern height	MC3T3-E1 cells	GT surfaces promote osteogenic differentiation	[119]
Microgrooved substrates	Photolithography and hot embossing	Polystyrene (PS)	2/6 (2 μm width, 6 μm interval) and 4/11 (4 μm width, 11 μm interval)	MG-63 human Osteosarcoma cells	2/6 substrate significantly enhanced osteogenic differentiation, as compared to the 4/11 substrate	[120]
Groove/ridge structures	Reactive ion etching and multi-beam laser interference	Polyimide	Microgrooved PI chip	MSCs	15 μm ridges increased adipogenic differentiation whereas 2 μm ridges enhanced osteogenic differentiation	[18]
Aligned and random fibrous scaffolds	Wet-collection electrospinning method	Silk fibroin	N/A	MC3T3-E1 pre-osteoblasts	Aligned scaffold promoted cell proliferation and osteogenic differentiation	[21]
Aligned structures	Microfluidic setup	Type I collagen	N/A	MSC	Aligned ECM structures maintained and enabled multilineage (osteogenic/adipogenic/chondrogenic) differentiation of MSCs, but osteogenesis was most prominent	[121]
Chondrogenic differentiation
Smooth and rough substrates	3D electrospun	PEOT/PBT	Ra from 14.3 ± 2.5 nm to 71.0 ± 11.0 nm	hMSCs	Lower surface roughness (Ra = 14.3 ± 2.5 nm) supported chondrogenic gene expression, as compared to higher surface roughness (Ra = 71.0 ± 11.0 nm)	[32]
Aligned and random nanofibers	Dual electrospinning	(PLLA)/(PCL)	Nanometer scale	NSP	Aligned nanofibers promote chondrogenic differentiation	[33]
Nanoislands and nanopits	Polymer demixing	PCL/PMMA	Depth: 90–200 nm Diameter: 400–700 nm	hMSC	Nanoisland topography has the potential to induce differentiation into osteogenic, chondrogenic and adipogenic lineages	[122]
Vertically aligned nanotubes	Electrochemical treatments	TiO2	15 or 100 nm	Mesenchymal stem cells (MSCs)	100 nm BMP-2-coated nanotubes promote chondrogenic differentiation, 15 nm nanotubes induced de-differentiation of chondrocytes	[34]
Micron, submicron, and hybrid grooves	Femtosecond laser	Titanium (Ti)	N/A	Human mesenchymal stem cells (hMSCs)	Hybrid surface can promote osteogenic and chondrogenic differentiation	[123]
Nanofibrous meshes	Electrospinning	Polycaprolactone (PCL)	70–80% of porosity, average pore size of 2.7 μm	Wharton’s Jelly stem cells (hWJSCs)	Nanofibrous meshes promote cartilage-related gene expression	[31]
Myogenic differentiation
Micro/nano-surfaces	Photolithography and anodic oxidation	TiO2	The width of the micro-grooves/ridges on the two-surfaces were 10 μm/20 μm and 20 μm/10 μm	Mesenchymal stem cells	The micro/nano-interfaces induced MSCs differentiation into contractile SMCs	[124]
E-shell polymer	Laser material processing	Acrylate resin e-Shell	Grooves with period of 10 µm; or 30 µm; square gratings with period of 10 µm/30 µm	Human primary vascular smooth muscle cells	Topography induce VSMCs from synthetic to contractile phenotype and directs formation and maintenance of cell–cell communication and adhesion structures	[125]
Grooved substrates	Photolithography	Poly(dimethylsiloxane) (PDMS)	Lateral dimensions: 2–10 µm depths: 50–200 nm	Human fibroblast cells (FCs), endothelial cells (ECs) and smooth muscle cells (SMCs)	Contact guidance was significantly increased when the cells were cultured on substrates with smaller lateral dimensions or deeper grooves, with different cell types responding to topographies differently	[126]
Nano- and micro-alignment combined scaffolds	Electrospun nanofiber and PEG hydrogel micro-patterning	Poly(ethylene glycol) (PEG)	N/A	Skeletal muscle cells	Myosin heavy chain (MHC) expression was significantly higher on parallel versus randomly aligned fibers	[127]
Parallel grooves	Capillary force lithography	Poly urethane acrylate (PUA)	800 nm in width and 600 nm in height	C2C12	Biomimetic nanotopography enhanced the formation of aligned myotubes	[128]
Nanofiber topography	Electrospinning	PCL–PLLA	PCL has mean diameter of 630 nm. PLLA has mean diameter of 180 nm	Human Wharton’s Jelly-derived mesenchymal stem cells (hWJ-derived UC-MSCs)	The nanofiber topography significantly up-regulated the expression levels of calponin1 and SM22α	[129]
Tendon differentiation
Aligned and randomly oriented scaffolds	Electrospinning	PLLA	Aligned (1068 ± 190 nm) and randomly oriented scaffolds (739 ± 129 nm)	Mouse multipotent mesenchymal stem cells	The aligned group was observed to form more mature tendon-like tissues in the Achilles tendon injury model	[24]
Well-aligned ultrafine fibers	Stable jet electrospinning (SJES)	Chitosan	891 ± 71 nm	Human-induced pluripotent stem cells (hiPSCs)	The in situ tendon repair study confirmed that the aligned fiber scaffold seeded with hiPSC-MSCs had a significant effect in improving the structural and mechanical properties of tendon injury repair	[103]
Anisotropically aligned biotextiles	Electrophoretic compaction with macromolecular alignment	Collagen	3D-biotextile with porosity of 80%	Mesenchymal stem cells	Tendon related markers were significantly elevated in the aligned group, as compared to randomly oriented collagen gels	[130]
Nanofibers	Electrospinning technique	Poly(l-lactic acid) (PLLA)	Fiber diameter was approximately 600 nm	Adipose-derived stem cells (ADSCs)	Nanofiber alignment alone had no effects, but when combined with PDGF can have a synergistic effect with maximum expression of tenogenic markers	[131]

Mechanisms by which substrate topography promotes cell differentiation

Cell differentiation is a very intricate biological process. During the process of embryonic development, stem cells are subjected to temporal- and spatial-specific stimuli of various physicochemical and biological signals and eventually differentiate into functional tissues and organs. As mentioned earlier, in reference to scientific literature, it has been proven that topographies can promote cell differentiation. However, the number of studies focusing on topography-induced cell differentiation is still few, meaning that a lot of research still needs to be done in this area. The specific mechanism of topography-induced cell differentiation remains unclear. In this section, we will discuss this problem with main emphasis on the following: (1) the mechanism of topographic recognition, (2) the transmission of topographic cues within cells, and (3) the cross-talk between topographic and chemical signals.

Mechanism of topographic recognition

Initiation process of cells that sense and recognize substrate topography

Many stem cell types such as MSCs and TSPCs are contractile cells and their mechanotransduction can be explained by the contact guidance theory. Contact guidance is a leading example of a naturally occurring phenomenon which explains how cells sense substrate topographies at the micrometer and sub-micrometer scales. These cause changes in cellular mechanical transduction [48]. The formation of lamellipodia and filopodia is thought to be closely related to contact guidance [49, 50] and it is also a symbol of recognition between cells and topographies. The lamellipodia and filopodia are both actin-based cell structures that have a distinct architecture and internal organization. Lamellipodia are based on a thin sheet-like branched network of actin filaments, whereas filopodia are highly organized and tightly cross-linked long bundles of unidirectional and parallel actin filaments [51]. The outward extension of filaments can be observed immediately after cells are in contact with the topographic surface. Interestingly, filopodia when observed under a microscope are always parallel or vertical to the topography. After a transient formation of the filopodia, some of it falls off and the remaining ones eventually nucleate into lamellipodia [52]. Lamellipodia and filopodia when in contact would seem to have a sense of being in direct contact with the topography. After activation of a class of downstream signaling pathways, these mature into an adhesion structure (focal adhesion), which can input topographical information into cells through cascade amplification signaling pathways. Eventually cell function is affected. Contact guidance is the beginning of cell sensing and recognition of topography through the formation of filopodia and lamellipodia. Topographic features can be sensed by cells through a class of signal transduction pathway. These topographical signals can be transmitted within the cell and finally influence stem cell differentiation.

Assembly of integrins and the formation of focal adhesions

Integrins, a large homologous family of heterodimeric transmembrane receptors, function as connections between the cytoskeleton and topographic features on the substrate. Like other transmembrane receptors, integrins can move freely through the cell membrane and assemble into focal adhesions that are responsible for cellular adhesion to topographic features. The focal adhesions will connect the cytoskeleton to the ECM and provide physical links between topographic and intracellular mechanotransduction architecture. About 20 integrin homologs have so far been identified. These different integrins have different binding affinities with various ECM components like collagen, vitronectin, and fibronectin [53]. The formation of integrin heterodimers is the initial stage of cell adhesion. Integrins bind to the topographic feature, together with other adhesion-related proteins such as vinculin, talin, paxillin and phosphorylated FAK to form primary adhesion points (0.5–1.0 μm). Then, various focal adhesion (FA) proteins, as mentioned earlier, promote the integration of cytoskeleton with these integrins, which ultimately lead to the formation of mature adhesion points (3–10 μm) [54]. Some researchers also demonstrated that newly formed focal adhesions can phosphorylate focal adhesion kinase (FAK) and FAK activity can directly affect the topography-induced gene expression. However, morphological changes were observed not to affect the level of FAK at all [55]. Many studies found that topographic features can reduce the cell spreading and cells seeded on topographic substrates have more smaller focal adhesions [56]. Ayomiposi et al. found that flat and 200 nm nano-patterned Pt-BMGs (d Pt 57.5Cu14.7Ni5.3P22.5 bulk metallic glass) can promote osteogenic and adipogenic differentiation of MSCs, respectively, and they believed that osteogenic differentiation is associated with increased formation of focal adhesions [57]. Yin et al. found that aligned topographies can promote the expression of tendon-related genes and the level of integrin alpha1, alpha5, beta1 subunits and myosin II B was evaluated in aligned topographies [58]. Topographies can also promote focal adhesions encoding genes (such as Ezr and Msn) expression [59]. This phenomenon implies that there is a certain relationship between the focal adhesions and topography. Furthermore, integrins can influence the level of cytokines and growth factors including IL-3 and TGF-β, thus influencing cell function. The function of integrins can also be influenced by these cytokines and growth factors. This is a process of interaction between the integrin and cytokines [60, 61]. Different topographies have different effects on focal adhesion formation. For example, when osteoblasts were seeded on three different topographies, researchers found that nanopit arrays tend to disrupt the formation of focal adhesions. Compared with planar substrates, 10-μm groove/ridge arrays hindered osteogenic differentiation but promoted adipose differentiation and formed less focal adhesions. In contrast, a 100-μm groove/ridge could promote osteogenic differentiation and promote focal adhesion formation as well as ERK/MAPK signaling pathways that were thought to be responsible for this observed phenomenon in mesenchymal populations [62].

Transmission of topographical signals within cells

FAK

Focal adhesion kinase (FAK) is a non-receptor tyrosine kinase that has an important role in cell adhesion, migration and many other cellular functions [55]. After integrins come into physical contact with the topographical features, focal adhesion formation takes place. The intracellular level of phosphorylation of FAK also changes and, together with other focal adhesion-related proteins such as vinculin, talin and paxillin, promotes cell differentiation [63]. A study seeded hMSCs on polydimethylsiloxane nanogratings (with 250 nm line width), and the neurogenic (MAP2, nestin, thyroxine hydroxylase and neurofilament-light) and myogenic differentiation markers (PAX3, myogenin, collagen IV and desmin) were observed to be upregulated. The FAK inhibitor (pFAKi) PF573228 was found to inhibit topography-induced differentiation. FAK siRNA inhibited topography-induced hMSC neuronal differentiation. Hence, FAK can be considered to be directly involved in topography-induced cell differentiation. FAK has been demonstrated to promote cell differentiation by altering actomyosin and cytoskeleton contractility [55]. Some researchers found that when an osteoblast cell line was seeded on nanometric grooved surfaces, osteogenic maker expression dropped significantly. It is believed that this phenomenon is associated with the activation of the ERK negative feedback pathway caused by FAK [64]. FAK knockout experiments showed that FAK may be able to communicate directly with SYNE to transmit mechanical signals between the cytoplasm and the cell nucleus. This phenomenon may be independent of actomyosin cytoskeleton contractility [65]. Additionally, the Rho-ROCK pathway is considered to modulate FAK phosphorylation as well as its formation. Actin organization within the seeded cells is known to be dependent on the substrate topography [66]. Some researchers seeded rat osteoblasts on different surfaces including grooves, tapered pits, and gap-cornered boxes, and found that FAK and ERK 1/2 phosphorylation were increased on groove and gap-cornered boxes. By contrast, Src levels were higher on smooth surfaces. Inhibition of Src phosphorylation can inhibit FAK and ERK 1/2 phosphorylation on grooved topographies but had no detectable effects on either FAK or ERK 1/2 on smooth substratum. Thus, it was concluded that osteoblast response to substrata with specific topographical features requires FAK-Y397-Src-Y416 complexes for ERK 1/2 phosphorylation. On smooth surfaces, Src-independent methods of ERK 1/2 phosphorylation are activated [1]. The depth of topographic features has been demonstrated to affect the expression of FAK. Some researchers observed that 14 and 29 nm deep pit surfaces increased hFOB cell attachment, spreading and selective integrin subunit expression (i.e., alpha relative to alpha5, beta1 and beta3). Focal adhesion paxillin protein synthesis and paxillin co-localization with cytoskeletal actin stress fibers and focal adhesion kinase (FAK) and phosphorylated FAK (pY397) expression levels are higher with 45 nm deep pits versus flat PLLA surfaces [67].

Actin

After topographic features have been sensed by lamellipodia and filopodia, the integrin from the cell membrane integrates together to form focal adhesions. Focal adhesions receive some intracellular proteins and become mature. The specific topographic features sensed by focal adhesions need to be transmitted into the stem cells to promote cell differentiation. The cytoskeleton is responsible for transmitting the information of topographic features into cells and consists of many molecular networks. Actin is the major component of the cytoskeleton and it consists of non-muscle contractile stress fibers [8]. The actin cytoskeleton can influence a variety of biological functions in eukaryotic cells. Stress fibers start from focal adhesions and form a mechanoresponsive network that transduce mechanical forces into cells and trigger downstream signaling pathways that promote cell differentiation [9]. Cytoskeletal tension can regulate the architecture of the nucleus and the degree of tightness between nuclei chromatin affecting gene transcription and has some effects on gene differentiation [68]. Calcium-dependent myosin light chain kinase (MLCK) and calcium-independent Rho kinase can regulate stress fibers through MLC phosphorylation [69], cytoskeletal actin and intracellular mechanical transduction proteins such as members of the Rho family, thereby promoting cell differentiation. Besides providing a structural framework for the cells, it also acts as a “molecular engine” responsible for providing and directing forces for cells to move and divide. Therefore, the action of the cytoskeleton is responsible for the morphological change of the cells seeded on various topographies. There is no doubt that we can deepen our understanding of topography-induced stem cell differentiation if we fully understand the exact role played by the actin cytoskeleton.

Morphological changes

Cell shapes and structures have been shown to influence the lineage fate of stem cells. Cell shapes have been demonstrated to influence MSC osteogenic differentiation through the increase of extracellular-related kinase (ERK1/2) and c-Jun N-terminal kinase (JNK) activation in conjunction with elevated wingless-type (Wnt) signaling. Interestingly, after some cells have been identified through topography contact guidance, the most obvious change of these cells seeded on various topographies is their cellular morphological change. Many different cell types such as fibroblasts, endothelial cells, adult stem cells, smooth muscle cells, epithelial cells and some Schwann cells seeded on gratings respond to topographic features simultaneously by aligning and elongating in the direction of the grating axis [70]. For example, some studies showed that a parallel micro-grooved topography could transform and spread human dermal fibroblasts into a slender morphology. The expression level of tenogenic biomarkers was increased significantly [71]. Stem cells respond to topographic features swiftly and morphological change can be observed within a time range of 5 min [72]. In general, the depth and width of topographic features can affect the morphology of cells seeded on it [73]. However, the depth of topographies seems to have a stronger effect on cellular morphological changes [74]. Some studies have proven that some nanogratings can promote stem cell vertical arrangement better than horizontal nanograting axis [75]. Nano-posts and nanopits may have a weaker effect on cellular morphological changes as compared to nanograting. Cytoskeletal tension is thought to have an important effect on cell shape-mediated stem cell differentiation [76]. There is a hypothesis that when cell spreading is limited, cytoskeletal tension in stem cells is reduced and this promotes adipogenic differentiation. Full extension of the cell leads to higher cytoskeletal tension within cells that ultimately promote osteogenic differentiation [77, 78]. In other words, topographies can switch stem cell lineage fate between osteogenesis and adipogenesis through modification of the aspect ratio (length: breadth) and the curvature of the cells, regardless of whether there is a chemical inducer within the culture milieu [79]. Some researchers observed that focal adhesions were smaller and slenderer on patterned topographies, as compared to non-controlled patterns. This phenomenon relies actively on actomyosin contractility, thus suggesting a possible force-dependent mechanism [55].

The relationship between topographical and chemical signals

Rho family

Rho family is small guanosine triphosphatases, which have long been demonstrated to be key regulators of the actin cytoskeleton. Additionally, through their interaction with multiple target proteins, they ensure coordinated control of other cellular activities such as gene transcription and adhesion [80]. The RhoA–ROCK–CDC42 (cell division cycle 42) pathway is a canonical mechanotransduction pathway which has been validated and confirmed by many studies [81]. The RhoA downstream protein can affect many intracellular mechanics including stress-fiber formation, actin polymerization, actin–myosin complex contraction, focal adhesion assembly and many others [82]. Rho also acts as a molecular switch which can control signal transduction pathways that are associated with membrane receptors and the cytoskeleton [80]. The activation of Rho can induce the formation of contractile actin–myosin filament network, but its activation has no effects on focal adhesions [83, 84]. Other Rho GTPases such as Rac1 also act as mechanotransduction receptors and exert a profound influence on actin cytoskeletal reorganization [85]. Some researchers found that the RhoA and ROCK pathway control cell spreading and cell differentiation [77]. Some researchers also believe that Rac, Cdc42 and Rho may have antagonistic effects. There is a dynamic balance between these proteins. For example, cells with higher levels of Rac are typically flat and spread out, whereas those that are enriched in Rho have a circular and contracted morphology [83, 86, 87]. Similarly, osteogenic differentiation of cells need extensive cell spreading which means that the higher the actin cytoskeletal tension, the higher the increase in RhoA activity. Meanwhile, the adipogenic differentiation of stem cells requires a smaller cell spreading area and low RhoA activity. This implies that adipogenic differentiation needs lower actin cytoskeletal tension [88, 89]. Some studies also found out that RhoA-dependent actomyosin contractility can control stem cell differentiation into either the osteogenic or adipogenic lineages by varying their spreading areas [77]. Furthermore, the Cdc42 protein has proven to have a profound effect on filopodia formation and cellular polarity [90]. Some researchers also discovered that Cdc42 is an important component of Wnt/b-catenin signaling. This will be discussed in the latter part of the section on mechanical induced osteogenic differentiation and on the Cdc42 inhibition process. The inactivation of the GSK3b/APC/Axin complex leads to b-catenin accumulation, which translocates into the nuclei to promote cell differentiation [91]. Cadherin-based adherens mainly function to induce cell–cell adhesion and some researchers have found out that the RHO pathway may influence their assembly [92].

Wnt pathway

The Wnt pathway has long been thought to be the canonical signaling pathway that modulates osteogenic differentiation [93]. Similar to the Rho pathway, Wnt is also a signaling pathway that is closely related to cytoskeleton function. The actin cytoskeleton of stem cells seeded on various topographies can control Wnt/b-catenin signaling activation [94, 95]. Wallace et al. discovered that regulating modulators of Wnt pathway such as DDK1 and DDK2 are responsible for the early differentiation and late differentiation stages of stem cells on micro-structured titanium surfaces [96]. Recent studies have proven that during the osteogenic differentiation of mesenchymal progenitors, surface topography and chemistry can affect the Wnt pathway at the same time. Like the Rho pathway, the Wnt pathway can also affect cytoskeletal formation of cells seeded on various topographies [94]. The Wnt pathway may modulate formation of the primary cilium of cells seeded on various topographies, thereby affecting intracellular actin–myosin tension, which in turn modulate cell differentiation processes [97]. There is more cross-talk between the Wnt and Rho signaling pathways in stem cells seeded on various topographies. For example, rough topography enhances the activation of Wnt signaling, thus promoting cell differentiation. The Wnt signaling pathway can be modulated by rock signaling pathways in cells seeded on micropatterned titanium surfaces [98]. Yu et al. found that micro/nano hierarchically structured Ti samples (parameter not shown) can promote osteoblast differentiation through a Rock–Wnt5a feedback loop [47]. Some studies have clearly shown that topography-induced elongation alone is also effective in the upregulation of Wnt signaling pathways in adult neural stem cells. These effects can be enhanced by retinoic acid treatment. These mechanisms thus suggest that there is traffic between topological and chemical signaling pathways [99]. Interestingly, some researchers also realized that GSK3β can act as a cross-link between the Cdc42 and Wnt/β-catenin signaling pathways. Thus, GSK3β can promote the osteogenic differentiation of cells on various topographical surfaces [85].

TGF-b

TGF-b is an effective inducer of the tendon progenitor marker Scx. The absence of TGF-b signaling in Tgfb2-null and Tgfb3-null double-mutant embryos can lead to tendon and ligament developmental disorders [100]. An exogenous addition of TGF-b into the cell culture media can also promote and modulate cell differentiation [101]. Various topographies can affect the extracellular secretion of TGF-b from seeded stem cells [102]. TGF-b/Smad2/3-mediated signaling pathways may be responsible for the Scx expression mediated by the topographical mechanisms. It appears that integrins and cell cytoskeleton can affect TGF-b signaling pathways that activate SCX expression and promote the synthesis of tendon ECM [103]. A study found that exogenous 2 ng/ml TGF-1 could enhance tendinous gene expression. Interestingly, cell morphology seems to affect the sensitivity of cells to TGF-b, with enhancement being more significant in elongated cells compared with cell spreading. When a higher dosage of TGF-b was used, stem cells were differentiated into myoblasts rather than the tendon lineage and under these circumstances the effect of a high TGF-b dosage could overcome the effect of cell shape [71]. In addition to tendon differentiation promotion, the TGF-b signaling pathway has also demonstrated the ability to promote osteogenic differentiation. Some researchers used submicron and micro-scale topographies to promote osteogenic differentiation and found that submicron topography can alter the intracellular TGF-b level to promote osteogenic differentiation [104]. Another study on the other hand found out that topography-induced differential regulation of miRNAs can regulate the level of TGF-β/BMP and WNT/Ca(²⁺) that in turn modulate osteogenic differentiation [105]. The topography of calcium phosphate ceramics can regulate primary cilia length and recruit p-TGFβRII, thereby promoting bone formation [104].

Other related pathways

The topography of ECM can promote stem cell differentiation and many signaling pathways involved in these processes can interact with each other, eventually forming a complex interactive intracellular network. Besides canonical signaling pathways such as the RHO and Wnt pathways, some studies have also discovered that stem cells such as MSCs grown on various topographies express a higher level of growth hormone receptors (GHRs). Interestingly, ERK signaling can be affected by GH on topography but not on flat surfaces [106]. Shi et al. found that endoplasmic reticulum (ER) stress and PERK–eIF2α–ATF4 pathways may be responsible for substrate topography-induced osteogenic differentiation [85]. Furthermore, many researchers found that epigenetic regulation may be related to topography-induced cell differentiation [107]. Zhang et al. used TSA coupled with aligned topographies to promote tendon differentiation, which suggest that there are some relationships between epigenesis and substrate topographies [108]. Some researchers also found that nanoscale topography can promote osteoblast differentiation by increasing micro-RNA expression levels [109]. Other researchers found that some topographies may promote stem cell differentiation by affecting the electrical potential (EP) of the substrate [110].

In actual fact, substrate topographies often share signaling transduction pathways with chemical stimuli-induced cell differentiation. Some important components of signaling transduction pathways are cell differentiation inducers such as TGF-b and exogenous blockers. These chemical signals will block topography-induced differentiation. Thus, physical and chemical stimuli share many signaling pathways in the process of cell differentiation, particularly in the latter part of the signaling pathway. The main difference between chemically and physically induced cell differentiation is the cell signal recognition phase. Cells sense chemical stimuli mainly by ligands on cell membrane, while sensing physical stimuli such as topography through focal adhesions and utilizing cytoskeleton to transmit signal into the cells. However, how the extracellular topography is converted to intracellular biological processes remains unclear. The understanding of cytoskeletal reorganization that results in lineage-specific gene switch is not deep enough and the relationship between biochemical and topographic signals needs further investigation [111]. A study showed that topography together with chemical modification can synergistically promote osteogenic differentiation of MSCs [112]. Some studies have stated that nano-topographies might be important in the modulation of the inflammatory response in cultured human gingival fibroblasts [113]. Liu et al. studied in-depth various signaling pathways and found that the focal adhesion kinase, Wnt, and TGF-β pathways were involved in the activation of osteogenic differentiation of cells on random topographies, as compared with aligned topographies [114]. Although there are many similarities between topography-induced and chemical-induced stem cell differentiation, some researchers have found out that topographically induced stem cells have a distinct differentiation profile, as compared with those treated with chemically induced media [3].

Current problems and future directions

There are still some pertinent problems in previous studies. For example, most researchers seeded a certain type of stem cell in the same medium with different topography, and then compared the cell differentiation induced by the different materials, such as osteogenic and adipogenic differentiation. However, there is insufficient evidence to support the relationship between cell differentiation and topography (Table 1). For example, aligned topography has been demonstrated to have the capacity to promote tendon-, [24] bone- [21] and cartilage [33]-related gene expression. We think that this may be due to homology between the different components of the musculoskeletal system, so that even a single type of topography can induce stem cells to exhibit variable gene expression in a variety of different lineages. But most topographic structures used in current studies are simple synthetic structures, such as gratings, posts, and arrays [8], which are unable to mimic the natural topography of ECM. As a result, although cells seeded on these materials express tissue-specific genes, there is still a certain distance from physiological functional status and clinical applications. On the other hand, signal transduction pathways investigated in published studies on topography often overlap with classical mechanical transduction pathways such as stiffness. This may be partly due to the fact that when focusing on material topography, we sometimes overlook the inevitable influence of manufacturing process on other mechanical properties of the materials including stiffness. Thus, it is crucial to ensure the standardization of scaffold parameters during the production process, as well as the indispensable corresponding technical inspection afterwards. A recently published review expressed a similar opinion as us [115]. Although a growing body of evidence has shown the importance of topography in stem cell differentiation, some authors hold a different opinion. For example, some researchers found that topography with a high degree of roughness (Ra greater than 573 nm) can promote cell differentiation while a lesser degree of roughness (Ra less than 392 nm) can promote long-term maintenance of stemness [104]. Similarly, a study showed that micro–nano roughness (MN surfaces: Sa 919 nm) had the ability to maintain the stemness (> 3 weeks) of mouse embryonic stem cells for an extended duration [116]. Another group of researchers found that topography is also a critical factor for maintaining stemness [117]. In another study, topographic cues (15 nm) were found to induce the dedifferentiation of cartilage cells [34]. In summary, the existing methods based on topography alone cannot truly promote the differentiation of cell lineages, since cell marker expression induced by topographic cues is insufficient, and there are few studies that have investigated both the stemness and differentiation markers of cells to our knowledge. Paradoxically, there is a possibility that cells seeded on various topographies can express both stemness and differentiation markers at the same time, so we put forward the hypothesis that the morphology of stem cells seeded on the materials is altered, as well as the gene expression pattern, but those stem cells still have a certain degree of stemness at the same time. We deemed this special state as a “pre-differentiation” state, in which topographic cues together with other factors such as chemical factors and growth factors can promote stem cell differentiation into mature and functional cells (Fig. 1).

Fig. 1.

For now, most studies which focus on ECM topography are carried out in vitro with controversial results. Stem cells which seeded on 2D surface with various topographic features exhibit different tendencies during differentiation. Mechanistic studies found that topographic features can be sensed by cells through focal adhesions, which transmit those extracellular cues into cytoskeletal rearrangement and morphologic changes, which further regulate various intracellular signaling pathways. However, we proposed that topography alone does not dictate stem cell differentiation into any specific lineage, instead, it probably primed stem cells into a “pre-differentiation state”. Only when synergize with proper physical (such as stiffness of extracellular matrix) and chemical (such as growth factor in the stem cell niche) cues, can topographies promote stem cells differentiation into functional terminally differentiated cells

Conclusions

Topography of materials plays an important role in the field of tissue engineering. With the continuous development of manufacturing technology, the fabrication of more biocompatible and sophisticated scaffolds has become a reality. Many researchers used various topographies to promote stem cell differentiation. But the underlying mechanisms by which certain materials promote stem cell differentiation remain unclear. It is now widely acknowledged that the cell surface protein integrin plays a crucial role in sensing the topography of biomaterials. By sensing topographic features, it could induce changes in intracellular ligand proteins and promote the formation of mature focal adhesions that is linked to cytoskeleton and responsible for the transduction of mechanical signals to biological signals. Finally, nuclear transcription would initiate the differentiation of stem cells into different lineages.

Compliance with ethical standards

Conflict of interest

The authors hereby declare that they have no conflict of interest.