A Tissue Engineering's Guide to Biomimicry

ABSTRACT

Biomimicry is a broadly used term in many fields such as in architecture and industrial design to pharmacology and many others. Biomimicry tries to replicate a product or process that occurs in the natural environment. However, due to the broad use of the term biomimicry it becomes unclear what is exactly being mimicked. Specifically, in tissue engineering and regenerative medicine (TERM) where research is focussed on mimicking complex native tissue, the term biomimicry is often used to designate a single aspect. Therefore, in TERM biomimicry can be clustered into to three different categories correlated to the aspect that is being mimicked: i) mechanical ‐ that has a focus on obtaining the correct mechanical properties of a tissue; ii) morphological ‐ that aims at recreating a scaffold that has a similar morphology to its native counterpart; iii) biological ‐ that has a prime focus on recreating the biological microenvironment that is found in the targeted tissue. This review discusses the strategies and methods how these different forms of biomimicry can be achieved with the current techniques available.

This review categorizes biomimicry in tissue engineering into mechanical, morphological, and biological strategies. It explores how each aspect contributes to replicating native tissue function and structure, highlighting current techniques and emerging approaches. By clarifying the scope of biomimicry in regenerative medicine, this work offers a framework to guide future research toward more holistic and effective scaffold design.

Abbreviations

2D

two‐dimensional

α‐Gal

galactose‐α‐1,3‐galactose

ADAMTS4

ADAM metallopeptidase with thrombospondin type 1 motif 4

ALP

alkaline phosphatase

AQP1

aquaporin‐1

BMP‐2

bone morphogenetic protein 2

BSP

bone sialoprotein

CD31

cluster of differentiation 31

COX‐2

cyclooxygenase‐2

ECM

extracellular matrix

EDAC

1‐ethyl‐3‐(3‐dimethylaminopropyl) carbo‐diimide hydrochloride

EGF

epidermal growth factor

EGFR

epidermal growth factor receptor

EtOH

ethanol

FDM

fused deposition modeling

FGF

Fibroblast growth factors

GA

gum Arabic

GelMA

gelatin methacryloyl

HA

hydroxyapatite

HUVECs

human umbilical vein endothelial cells

IL‐1

Interleukin 1

KLF2

Krüppel‐like Factor 2

MMP3

matrix metalloproteinase‐3

MSCs

mesenchymal stem cell

NGF

nerve growth factor

OPN

osteopontin

PCL

poly(ε‐caprolactone)

PDA

polydopamine

PDLLA

poly(D,L‐lactide)

PDS

(poly(p‐dioxanone)

PDTC

pyrrolidine dithiocarbamate

PEGDA

polyethylene glycol diacrylate

PEOT/PBT

poly(ethylene oxide terephthalate)/poly(butylene terephthalate)

PGCL

poly(glycolide‐co‐ɛ‐caprolactone)

PLCL

poly(L‐lactide‐co‐ɛ‐caprolactone)

RGD

arginine, glycine, and aspartate

rGO

reduced graphene oxide

Runx2

Runt related transcription factor 2

SDS

sodium dodecyl sulfate

SEM

scanning electron microscopy

sGAG

sulfated glycosaminoglycans

SOX9

SRY‐box transcription factor 9

TERM

tissue engineering and regenerative medicine

TGF‐β1

transforming growth factor beta 1

TNC

tenascin‐C

TNF‐α

tumor necrosis factor alpha

TPP

tripolyphosphate

VE‐Cadherin

vascular endothelial cadherin

VEGF‐B

vascular endothelial growth factor B

ZO‐1

zonula occludens‐1

1. Introduction

The term “biomimicry” or “biomimetic” was first described by Otto Schmitt in 1957 when he attempted to mimic the electrical action of a nerve using a physical device [1]. Over the course of the years, scientists from different fields used the term to describe a product or process that tries to imitate a certain aspect from the natural environment, or taken more literally as the imitation of life and nature [2, 3, 4, 5]. Biomimicry ranges from biomimicking furniture design from natural elements [6] to efficient car design [7] to improved platforms for toxicity studies [8]. In the field of tissue engineering and regenerative medicine (TERM) similar efforts were made to emulate the natural microenvironment [9, 10], ranging from mimicking the morphology [11, 12], to physiological conditions such as electrical stimulation to improve muscle contractility [13], or the mechanical properties of a specific tissue [14, 15].

Studies in regenerative medicine, where a biomimicry strategy is aimed for, often focus on only one single characteristic that researchers aim to mimic [16]. The different types of biomimicry can be classified into three main categories: 1) mechanical, where researchers focus on obtaining mechanical properties similar to those of the target tissue [17]; 2) morphological, where the aim is to create a scaffold that is morphologically similar to the native counterpart [18]; 3) biological, where the prime focus is on recreating the exact biological environment that is found in the specific tissue [19].

Therefore, the aim of this review is to discuss strategies and methods to obtain the previously mentioned mechanical, morphological and biological biomimicry. We will focus on 3D tissue engineered scaffolds and hydrogels, which are relevant in the context of regenerative medicine, since other reviews have focused more on a specific tissue [20], single aspect of biomimicry [21], or two‐dimensional (2D) substrates [22, 23].

2. Mechanical Biomimicry

Each tissue is characterized by specific mechanical properties, from soft brain tissue in the range of 0,1‐16 kPa [24] to stiff cortical bone with 19,3 GPa [25]. Mimicking the mechanical properties of a targeted tissue with a designed 3D environment has been the subject of many studies [26, 27, 28]. The reason to mimic the mechanical properties is to reduce the risk of a mechanical mismatch, which often leads to graft failure [29, 30, 31]. This is especially the case where the mechanical properties play a pivotal role in tissue development and homeostasis, such as the supporting capabilities of bone [32], the ability to withstand physiological pressure conditions and prevent stenosis in arteries [33] or the elasticity of skin [34]. Strategies to obtain mechanical properties of 3D scaffolds mimicking the tissue of interest can vary by choosing the right material [35], by varying the porosity or the architecture of the pore network [36, 37], by changing the cross‐linkers to tailor the mechanical properties of the scaffold [38], or using fillers to create composite materials to improve its performance [39, 40]. Mimicking the mechanical properties of a targeted tissue has also important implications from a mechanobiological perspective. Depending on the mechanical forces at play at the interface between materials and cells, a different degree of mechanostransduction can be expected, which is known to steer cell morphology and fate.

2.1. Bulk Material

One major influence on the mechanical properties of a scaffold is the bulk material. The materials that are used in TERM can range from soft hydrogels to stiff metals. The range in stiffness can vary by as much as a factor 1.000.000 [41, 42]. Therefore, the mechanical bulk properties are often one of the reasons why to create a scaffold out of a certain material [35, 43] (Figure 1). For example, Janke et al. described how the mechanical properties of collagen type I scaffolds for ureteral tissue engineering improved significantly with the addition of poly(L‐lactide‐co‐ɛ‐caprolactone) (PLCL), poly(glycolide‐co‐ɛ‐caprolactone) (PGCL) or (poly(p‐dioxanone) (PDS) [44]. The requirements for the material was that it is biocompatible and biodegradable. The ultimate tensile strength increased from 1.8 ± 0.8 to 160 ± 20 kPa with the addition of PDS. In addition, the stress strain curve became “J”‐shaped displaying the toe‐region, which is also observed in the native tissue. It is worth to note that even though the study of Janke et al. reported that the mass and compressive strength of PDS did not decrease after 48 days, a different study by Ping Ooi et al. showed both the strain and load at break in tension decreased over time even after 15 days of degradation [45]. To overcome the issue of degradation Yoon et al. used poly(ε‐caprolactone) (PCL) to increase the mechanical properties [46]. PCL is known to be a slow degrading polymer and can hold its shape even after 2 years of implantation [47]. The study showed that the resilience of the scaffold increased by introducing an electrospun PCL mesh inside an alginate hydrogel. The PCL mesh also helped retain the shape of the alginate hydrogel after cross‐linking and compression tests. Besides the mechanical support, the mesh was still permeable for the cells to migrate through. A different application for PCL was found in skin regeneration, where Rad et al. combined PCL with Zein and gum Arabic (GA) in electrospun meshes [48]. During the study it was noted that the formulations with the highest concentration of PCL (20%) and a lower concentration of Zein (15%) had the highest tensile strength (2.9 MPa) while the samples with the lowest tensile strength (1.3 MPa) contained less PCL (15%) and more Zein (20%). Interestingly to note is that the presence of GA enhanced the hydrophilicity and gave the scaffold antibacterial properties. A different approach to alter the mechanical properties of a scaffold by adding materials was taken by Khalili et al. [49]. During the study elastin was added to electrospun meshes, which slightly decreased the Young's modulus, but significantly increased the elongation at break under tensile characterization, thus making this an ideal candidate for skin tissue engineering where skin is able to stretch up to 100% without break [50].

FIGURE 1.

Changing bulk material of a scaffold to match the mechanical properties of a native tissue. (A) Schematic illustration to describe the change in materials influencing the mechanical properties. (B) Examples of scaffolds with different mechanical properties for soft (Material A) and hard (Material B) tissue regeneration; left panel: the mechanical properties of different Polyvinyl alcohol scaffolds for neurogenic differentiation, reproduced with permission [41]; right panel: the mechanical properties of different Polylactic Acid and nano HydroxyApatite blend scaffolds for bone tissue engineering, reproduced with permission [55].

As mentioned earlier, the bulk material choice is pivotal for the mechanical properties of the final scaffold and is often the starting point of a study. However, the human body remodels biomaterials over time and materials can degrade resulting in a change in mechanical properties [51, 52]. To overcome this, the degradation time of the material has to match the speed of which the biological tissue is regenerated. Another issue that might arise is that even though the material properties can be right from a mechanical biomimicry perspective, the material might not facilitate biological integration or even elicit an immune response [53, 54]. In summary, the choice of the bulk material is the major player in the mechanical properties of a scaffold and the first step in obtaining mechanical biomimicry.

2.2. Scaffold Design

Not only the choice of material has an influence, but also the geometrical design can play a pivotal role. For instance, the porosity of the scaffolds has a direct effect on their mechanical properties [56, 57]. Melchels et al. showed that by varying the design within poly(D,L‐lactide) (PDLLA) functionalized with methacrylate groups from a cube to a gyroid architecture the elastic modulus decreased from 324 MPa to 169 MPa, respectively [58]. The von Mises stress distribution was also improved with the gyroid design. This in turn could expose attached cells to a more equal mechanical stimulation throughout the whole scaffold. Other designs can introduce a negative Poisson's ratio (auxetic) as Jin et al. showed [59]. The auxetic properties were introduced by creating a periodic star‐shaped re‐entrant lattice structure on the scaffold through fused deposition modeling (FDM). This was combined with a melt‐electrospun mesh to increase the surface area for cells to attach to. The Poisson's ratio could be varied by changing the angle of the unit cell. In addition, it was found that the elastic modulus in the scaffolds increased with the decreasing re‐entrant angle of the unit cell from 100 to 40 MPa with a 140° and 110° angle, respectively. Unfortunately, only a static culture with cells was performed and the auxetic effect was not investigated in this study. A different study by Di Luca et al. used PCL, to fabricate scaffolds in different designs [60]. By changing the deposition angle of the fibers, scaffolds were fabricated in a gradient pattern and compared to their non‐gradient counterparts. The results showed that the gradient design were stiffer compared to a non‐gradient scaffold. Instead of changing the deposition pattern, Huebner et al. showed that by varying the fiber spacing from 100 to 400 µm within the scaffolds, the elastic modulus after four weeks of implantation in a porcine model for meniscus regeneration was 380 and 108 kPa respectively [61]. Yet, after 12 weeks of implantation the elastic modulus was similar between both conditions and the collagen alignment within the strands of the scaffold had a significant influence on the elastic modulus rather than the density of the deposited matrix. This demonstrates that the mechanical properties can be precisely controlled pre‐implantation, though after implantation there is little to no control over the scaffolds mechanical properties as degradation and tissue formation have a major influence. Besides proving the importance of design, Zamani et al. showed that the direction in which the scaffold was fabricated in a layer‐by‐layer fashion can change drastically the mechanical properties [62], especially if there is tension which is perpendicular to the fabrication direction of the scaffold. This results in an increased chance of delamination, which is also true if there is a compression collinear to the fabrication direction. An entirely different strategy described by Jing et al. used scaffold's design to prevent kinking of vascular grafts [63] (Figure 2). Testing various designs, having an outer spiral architecture on a tubular scaffold had an improved effect on the retention of the scaffold's diameter compared to a ring or cross design. Even with a small bending radius of 4 mm, the spiral design prevented a major diameter reduction (< 20%).

FIGURE 2.

Changing the scaffold design to match the mechanical behavior of a native tissue. (A) Illustrated example describe the change in design to prevent kinking of the scaffold when bent around a tight corner. (B) Bending test of polycarbonate bisurea scaffolds with two different designs around a radius of 30 mm; reproduced with permission [63].

As discussed in these studies, the mechanical properties can be tuned to a large extent depending on the design. This allows the use of the same material for a wider range of applications. For instance, the design of a scaffold can be changed for materials that have stiff properties in bulk. In addition, with the relatively new field of meta‐materials, complex mechanical properties can be introduced such as negative Poisson's ratio designs or origami‐inspired meta‐materials on materials that do not have those intrinsic properties [64, 65, 66]. However, these meta‐material designs require the unit cells to be relatively open and only display these properties when there is room for the design to bend or twist [67], which opens the question whether they would retain their properties when implanted in‐vivo. As discussed earlier with the material choice, the design can change the mechanical properties on a scaffold level, but it cannot change the bulk material properties. This might result in a scaffold design with the right mechanical properties but suboptimal biological properties (e.g. cell attachment and immune response) if the material chemistry is not adequate to allow cell‐material interactions [53, 54]. In summary, scaffolds’ design plays a key role in obtaining the right mechanical properties, whether it is to fine‐tune the mechanical properties to match exactly those of the tissue or to prevent the collapse of a scaffold.

2.3. Cross‐Linkers

The cross‐linking of materials is another method to modulate the mechanical properties of a scaffold, especially since the cross‐linking density can drastically affect the mechanical properties. As described by Suo et al. [68], an interpenetrating network of gelatin methacryloyl (GelMA) and chitosan was formed by covalent bonds and hydrophobic interactions through both photo‐cross‐linking and basification through sodiumhydroxide as cross‐linking methods. The gelatin from GelMA was derived from porcine skin and had a methacrylation degree of 95% [69]. This resulted in a significant increase in tensile strength and compressive modulus compared to chitosan entrapped in the UV cross‐linked GelMA network. The lowest compressive modulus of 3.4 kPa was with chitosan gels that were cross‐linked through basification. The strongest gels with a compressive modulus of 116.1 kPa were made out of a GelMA‐ chitosan blend that was cross‐linked through both photo‐cross‐linking and basification. In addition to the increased mechanical properties, the degradation time also increased with cross‐linking. Whereas the non‐cross‐linked GelMA degraded within 48 hours, the cross‐linked GelMA remained with about 75% of its mass after 4 days. Besides the increase in mechanical properties, the method of cross‐linking has also an influence. For instance, Martínez et al. compared three different cross‐linking methods on Chitosan‐collagen sponges [70]. A comparison between 1‐ethyl‐3‐(3‐dimethylaminopropyl) carbo‐diimide hydrochloride (EDAC), sodium tripolyphosphate (TPP) and a combination of the two methods was made. The results showed, in line with the other papers, that the Young's modulus change with the cross‐linking conditions. The non‐cross‐linked scaffold had the lowest Young's modulus (4.5 ± 0.3 kPa) while the TPP cross‐linked scaffold had the highest Young's modulus (20 ± 0.2 kPa). Interestingly, the combined cross‐linking method resulted in the lowest Young's modulus of 7.7 ± 0.5 kPa with EDAC followed right after at 9.9 ± 0.4 kPa. The authors explained that specifically chitosan cross‐links more efficiently with TPP instead of EDAC, implying that the combination between material choice and cross‐linking method matters. Besides the cross‐linking method, the duration of cross‐linking also influences mechanical properties as shown by Naghieh et al. [71], where printed alginate scaffolds were subjected to calcium chloride cross‐linking for different periods (Figure 3). Directly after printing the Young's modulus was only 39.8 kPa, whereas after 1 day of cross‐linking the Young's modulus increased up to 273.4 kPa. In addition, the results showed that having a bigger cross‐linker volume led to a higher Young's modulus due to the larger amount of calcium ions available. Though this method can increase the mechanical properties, long exposure to Ca²⁺ ions at high concentrations can lead to induced apoptosis [72].

FIGURE 3.

Changing cross‐linking density in a scaffold to match the mechanical properties of a native tissue. (A) Schematic illustration of increasing the cross‐linker density to increase the stiffness. (B) Bioplotted bulk gel in the left panel and the effect of cross‐linking duration and volume on the elastic modulus of the scaffold in the right panel [71].

Certain cross‐linkers, however, can be toxic especially when used in higher concentrations due to the increased amount of unreacted cytotoxic compounds [73, 74]. An alternative is to switch to non‐toxic cross‐linking system, which in turn can change the mechanical properties [70]. In addition, a tradeoff has to be made with hydrogels: on one side having a mechanically stable and stiff hydrogel that is easy to process and shape, while on the other side a softer hydrogel that allows cell proliferation but degrades within an appropriate application‐specific time frame [75]. A solution to this problem could be offered by reversible hydrogels that change mechanical properties with either light or temperature [76, 77]. In general, the mechanical properties of cross‐linked hydrogels range from a storage modulus of less than 100 Pa [78] up to 300 MPa [79], which makes them suitable for soft tissues such as lung and brain [80], or hard ones such as cancellous bone, respectively [81]. In summary, using cross‐linkers is one of the strategies to modulate the mechanical properties of a hydrogel to obtain mechanical biomimicry.

2.4. Fillers

There is a variety of reasons why the use of fillers inside a bulk material is investigated. One of them is to change the mechanical properties. For example, Janfada et al. described that when PCL was electrospun in combination with mesoporous silica particles the elastic modulus increased according to the amount of particles added [82] (Figure 4). The explanation given by the author is that the physical connection between the particles and the polymer restricts the movement of the polymeric chains, resulting in an increase in elastic modulus and ultimate strength. In addition, the ultimate tensile strength also increased. However, the strain at break significantly decreased, because of the brittleness of the ceramic particles that are included in the polymer. Similar findings were observed by Du et al. who used hydroxyapatite as a filler in a polyurethane composite scaffold [83]. Both compressive strength and elastic modulus increased significantly from 0.6 ± 0.1 MPa and 4.4 ± 0.4 MPa respectively without fillers, to 4.6 ± 0.3 and 36.9 ± 8.7 MPa respectively when 40% wt. hydroxyapatite was added. A solution to reduce the brittleness that occurs when using a high filler content is to create gradients in composition during scaffold fabrication [84]. This continuous hydroxyapatite composition gradient within a poly(ethylene oxide terephthalate)/poly(butylene terephthalate) (PEOT/PBT) resulted in scaffolds with an improvement of the strain at break by 50% compared to their discrete gradient counterpart. By having a discrete gradient, the difference in mechanical properties between layers is bigger compared to having a continuous gradient, which results in reduced stress at these interfaces. Increasing the amount of fillers does not always increase the mechanical properties. A study by Seyedsalehi et al. found that adding 0.5% wt. reduced graphene oxide (rGO) in a 3D printed scaffold increased its compressive strength and modulus [85]. However, when a higher percentage rGO was added the compressive modulus significantly decreased even when compared to the PCL scaffold without rGO. An explanation for this was that increased rGO concentrations caused aggregation and re‐stacking of the rGO sheets, whereas at a lower concentration the rGO was dispersed homogenously throughout the polymer matrix contributing to the mechanical properties.

FIGURE 4.

Adding fillers inside a bulk material to match the mechanical properties of a native tissue. (A) Schematic illustration to describe the added effect of fillers on the Young's modulus of a scaffold. (B) SEM images of the PCL scaffold with and without mesoporous silica particles and the stress strain curves of the tested scaffolds; reproduced with permission [82].

Introducing fillers to a material can increase the stiffness of a scaffold. However, not always does this lead to an increase in mechanical properties as discussed earlier. Adding more fillers might contribute to an increased Young's modulus, but it also causes the scaffold to be more brittle, which poses a problem when used for more flexible tissues such as cartilage, muscle or tendons [86, 87]. Another more practical issue with the use of fillers is the difficulty to work with them; mixing them inside the polymer matrix can be difficult, since the fillers are prone to form micrometer sized aggregates leading to a heterogeneous distribution [88, 89]. Replicating the anatomical features with precision requires high resolution biofabrication techniques. These techniques use nozzles that range in the micrometers. This in turn makes the use of fillers more prone to clogging of the needle, resulting in producing heterogeneous scaffolds with defects [90, 91]. Nonetheless, the use of fillers can be ideal for stiffer tissues such as bone and dental applications, due to the mechanical properties of the filler and the inorganic component found in both the tissue and the filler that is used [92].

3. Morphological Biomimicry

The morphology of a tissue is normally related to its function. In the trachea, for example, the C‐shaped rings of cartilaginous tissue have their opening facing the side of the esophagus [93]. The cartilage within these rings helps maintaining the tissue structural integrity, yet the opening is made out of flexible tissue that extends whenever food travels through the esophagus. Another example is the shape of the dental root, which can cause a difference in force distribution and in turn can impact alveolar bone resorption [94]. There are multiple strategies to obtain a morphological biomimicking scaffolds, ranging from mimicking interfaces between tissues [95], to mimicking the hierarchical pore architecture that is present in various tissues [96], to using a decellularized tissue [97], or through templating techniques [98].

3.1. Interfaces between Tissues

Biphasic or bilayered scaffolds are generally used in tissue engineering where there is a tissue transition, such as the interface between bone and cartilage [99]. Liu et al. described how a biomimetic biphasic osteochondral scaffold was made by fabricating a porous 3D plotted hydroxyapatite scaffold as the solid bone part and integrating it with a hyaluronic acid methacryloyl hydrogel as chondrogenic layer on top [100] (Figure 5). The results showed a higher expression of cartilage specific markers, such as collagen type II and aggrecan in the chondrogenic side of the scaffold, and an upregulated expression of osteogenic markers such as alkaline phosphatase (ALP) and Runt Related Transcription Factor 2 (Runx2) in the osteogenic side of the scaffold. In addition, by integrating a transition zone where both the hydrogel and solid scaffold were combined, both sides remained connected after two months of subcutaneous implantation. Instead of using a solid polymer and hydrogel combination, Mesallati et al. either used two hydrogels or a hydrogel combined with a self‐assembled cell layer. The bone compartment was made out of alginate in both conditions while the chondral part was made out of agarose or a scaffold free cartilage layer of bone marrow derived mesenchymal stem cells (MSCs) [101]. A higher sulfated glycosaminoglycans (sGAG) and collagen amount was found in the self‐assembled chondral layer compared to the agarose layer. Furthermore, the integration and stability between both bone and cartilage layers seemed to improve when a co‐culture was performed with chondrocytes and bone marrow stem cells. This technique could also be easily scaled up to an anatomically shaped osteochondral construct, which covers an entire condyle. After 8 weeks in vivo, a layer of cartilage remained on the surface of the construct as was shown by collagen type II staining. However, the same histological analysis also showed the presence of collagen type I or III. Collagen type I in cartilage is associated with osteoarthritis [102], while collagen type III indicates that the chondrocytes are dedifferentiated [103]. In addition, some degree of mineralization was found in the osseous part of the scaffold. Another example of an anatomically shaped osteochondral construct was made by Alhadlaq et al. [104]. Here, the mandibular condyle of an adult human cadaver was used to fabricate a mold out of acrylic. MSC derived chondrogenic cells were suspended in a polyethylene glycol diacrylate (PEGDA) hydrogel, the chondral part of the mold was filled with this hydrogel and photocross‐linked. Afterwards, mesenchymal stem cell derived osteogenic cells were suspended in PEGDA and cast on top of the previous layer. The results showed a clear distinction between the cartilage part and the bone part of the hydrogel with collagen type II in the cartilage part and collagen type I in the bone part of the scaffold, mimicking the osteochondral tissue. Another more recent attempt to make an anatomically correct condyle shape was by Peiffer et al. [105]. In this study, a mesh fabricated through MEW was deposited on a curved surface that matched the curvature of the diarthrodial condyle. This mesh was then placed on top of an extrusion based PCL scaffold with similar curvature. The surface of the scaffold was seeded with articular cartilage progenitor cells while the bone part of the scaffold was left without cells. The study showed the presence of collagen type II in the chondral part and no cell infiltration in the bone part of the scaffold. However, the interaction between the osteo and chondral cells in their respective part of the scaffold was not known due to the lack of osteoprogenitor cells in the study. In addition, the adherence between the two different zones was not tested, which is a known problem [101].

FIGURE 5.

Mimicking tissue interfaces such as the bone cartilage interface with a hydrogel and solid polymer combination. (A) Schematic illustration of a bilayered scaffold mimicking the morphology of the surrounding tissue. (B) Bilayered osteochondral scaffold on the left, before implantation and after 2 months subcutaneous implantation; reproduced with permission [100].

The tendon‐bone interface has similar challenges when it pertains to the integration between these two different tissues [106]. Instead of using 2 different hydrogels, Echave et al. applied magnetic alignment to create a biphasic gelatin hydrogel with anisotropic features [107]. The bone layer was loaded with hydroxyapatite (HA) microparticles to promote osteogenesis, while cellulose nanocrystals in a magnetic field were used to create aligned pores in the hydrogel for the tendon layer. The results showed the possibility to create an anisotropic hydrogel and the addition of HA significantly increase the Young's modulus from 17.4 ± 6.8 to 38.4 ± 4.9 kPa, and stimulated the ALP activity and osteopontin (OPN) expression, which is beneficial for osteogenesis. In addition, the magnetically stimulated nano‐cellulose crystals were able to orient in a controlled direction, which resulted in increasing of the tendon‐related marker Tenascin‐C (TNC) expression. Even though the Young's modulus increased significantly with the addition of HA, the mechanical properties of bone are still several orders of magnitude higher [25]. In addition, the presence of HA does mimic the osseous part, but the particles themselves did not show any hierarchical structure as seen in bone [108].

The morphology at interfaces of tissues has caught the attention of many researchers trying to replicate it. Because of the dissimilarity between these targeted tissues in terms of morphology, mechanical properties and extracellular matrix (ECM) composition, many attempts have been made to create a scaffold that is comprised of a stiff and a soft part made out of two materials [100, 109]. The disadvantage of this approach is that there is a risk of delamination at the interface of the scaffold [101, 110]. To overcome this challenge, these interface scaffolds could be made out of a single base material with one or both sides functionalized to promote the differentiation toward their end‐goal application [111, 112]. However, in most interfaces, especially when bone is one of the tissues, the mechanical properties are several orders of magnitude different, which makes this approach of using a single material challenging from a mechanical point of view [25]. In addition, the ECM composition between the tissues at the interfaces are very dissimilar and complex, which raises the question whether adding or removing a single component on one side of the scaffold could lead to successful in‐vivo integration [113, 114]. Future studies could focus on trying to create a better integration between the two dissimilar parts of the scaffold. In addition, in the specific case of skeletal articular tissues, the tidemark area is often an overlooked part of the interface that could be recreated. In spite of the challenges, biphasic scaffolds are an elegant strategy to create scaffolds suitable for interfaces of tissues.

3.2. Architecture of Pores and Porosity

Pore architecture, size and porosity in tissues are important properties to mimic. The range of pore sizes associated to a specific tissue are generally attributed to a specific function. For example, in bone the largest pore or cavity are called the medullary cavities, where the mesenchymal stem cells reside and are produced, which can range between 4 mm up to 10.5 mm [115, 116]. The smaller pores come from the Harversian and Volkmann's canals that provide the blood supply for the osteon's with a size of around 200 µm; even smaller pores of around 7 µm called canaliculi, which spread out radially from the Haversian canals, are used by osteocytes to communicate with each other [117]. Pore size is intrinsically correlated to total porosity. Both are important structural properties to take into consideration in tissues. Observing this organization from a macro perspective, a gradient in pore morphology and total porosity can be observed from a more dense and compact tissue outside to a more porous tissue inside [118]. With this observation in mind, Di Luca et al. fabricated a scaffold with a radial gradient with 3 distinct regions, ranging from the most porous medulla substitute part with a porosity of 77.6% ± 3.2% to the densest outer cortical mimicking bone part with 29.6% ± 5.0% [119]. The study showed an increase in ALP activity in the areas with pores of 1000 µm, yet not directly correlated to osteogenic differentiation of the seeded human MSCs. Additional findings showed that specific osteogenic markers Runx‐2 and bone sialoprotein (BSP) were upregulated in the regions with smaller pores of 500 µm. Bittner et al. used a similar extrusion‐based 3D printing technique to create a gradient scaffolds out of PCL; instead of having a radial gradient, the gradient was created by increasing the spacing of each fiber after several layers ranging from 200 µm up to 900 µm [120]. In addition, each section of the gradient had a certain concentration of HA mixed inside the PCL scaffold, with highest concentration of 30% in the layers with the largest spacing. Though from a morphological point of view the scaffold mimics the porosity found at an osteochondral transition zone by having the largest pores at the center of the bone, from the previously discussed study we could infer that having smaller pores of 500 µm instead of 1000 µm improved osteogenic gene expression. However, Bittner et al. did not show any in vitro work during this study. A different approach to create porosity described by Sola et al. [121] was by solvent‐casting poly(methyl methacrylate) or polyurethane with NaCl or NaHCO₃ with a subsequent salt‐leaching step. By varying the different ratios between the polymer and the salt, different size in pores could be obtained ranging from 600 µm up to 1000 µm. The resulting surface of the pores was irregular, which could be advantageous since that is known to enhance cellular attachment and proliferation [122]. Initial cell studies showed that some of the seeded cells managed to migrate in the central part of the scaffolds after 48 hours, though in limited number; furthermore, the long term effect of cell differentiation in these scaffolds has not been studied. Davidenko et al. demonstrated an alternative method to produce scaffolds with morphologically similar pores while avoiding the use of salts [123](Figure 6). During this study, a gradient in porosity was created through ice crystal formation by adding a collagen solution inside a custom‐made mold that would create a gradient in temperature. The porosity of the scaffold was 98.9% ± 0.05, which corresponds to the 1 wt.% collagen suspension used. This resulted in a gradient in pores of between 70–100 µm and 130–180 µm when the solution was frozen and lyophilized. However, the cell response on these small pores was not studied, which might be too small and, together with their tortuosity and incomplete interconnectivity and accessibility, is known to cause limited nutrient supply and oxygen availability at the center of the scaffold depending on the application [124].

FIGURE 6.

(A) Schematic illustration of obtaining a porosity gradient scaffold and (B) SEM images of pore‐gradient collagen type I scaffolds produced in a mold that creates a temperature gradient to mimic the porosity gradient found in bone; reproduced with permission [123].

Mimicking the porosity of a tissue is important not only for replicating the morphology of a tissue, but can also dictate whether cells would differentiate toward their targeted tissue [119]. Populating in a uniform way a scaffold with cells depends on the seeding methods and the time in culture [125]. Having larger pores requires more time to entirely populate the scaffold; on the other hand having small pores might take less time but can cause a lack of oxygen and nutrient supply [124]. To tackle oxygen supply limitation, a possible solution is by incorporating oxygen‐releasing materials, such as calcium peroxide incorporated in a polyurethane scaffold [126]. Another alternative could be by promoting vascularization prior and after implantation by including preformed vessels or incorporating VEGF in the scaffold respectively, which could solve both the nutrient and oxygen supply problems [127, 128]. In summary, having and controlling porosity is vital for morphological mimicry and ultimately nutrient supply and de novo tissue homeostasis.

3.3. Decellularized Tissues

Using a decellularized tissue to retain extracellular matrix morphology has been an investigated strategy for many years, where these templates are typically seeded with cells from another host [129, 130, 131]. Certain decellularized products are already clinically applied such as pig heart valves [132], porcine and human dermis for rotator cuff repair [133, 134], and bovine decellularized bone for skeletal repair [135]. Pors et al. described how human ovarian tissue was decellularized and repopulated with human follicles and implanted in mice [136]. The results showed that the decellularized scaffolds were able to support the survival and promote the growth of human follicles both in vitro and in vivo. Similarly, Giraldo‐Gomez et al. investigated a rapid decellularization process of a porcine trachea with guanidine, trypsin and ultrasounds [137]. Besides the architecture, also some essential ECM components, such as collagen type II and glycosaminoglycans, were preserved after the decellularization process. The decellularized scaffolds were recellularized with human adipose‐derived mesenchymal stem cells, implanted in mice and assessed for their immunogenicity. The results showed that although there was a presence of the inflammatory cytokine tumor necrosis factor alpha (TNF‐α), this was also shown in the sham control group and in both groups the concentration decreased over time. In addition, some cells in the scaffold showed early differentiation toward chondrocytes by the presence of SRY‐Box Transcription Factor 9 (SOX9) after 15 days of implantation. Even though the study used low concentrations, it is known that trypsin can cause ECM degradation [138]. Casali et al. proposed a different method where supercritical CO₂ was used in combination with sodium dodecyl sulfate (SDS) for the decellularization of porcine aortas [139] (Figure 7). Whereas other methods mainly use harsh chemicals or physical treatments, supercritical CO₂ is relatively inert as the results showed that the combination with SDS resulted in a similar morphology as the native tissue, but without the presence of cells. However, the method still has to be optimized and tested in vivo to check whether it is biologically comparable to other methods. Besides animal or human organs, also decellularized plants are potential candidates for skeletal muscle tissue as studied by Cheng et al. [140]. During this study, several fruits and vegetables were decellularized and their morphology was evaluated with scanning electron microscopy (SEM). As final candidate, scaffolds produced from green onion were chosen since the resulting decellularized structure presented highly aligned pores similar to those of the skeletal muscle tissue. Once seeded with human skeletal muscle cells, these aligned in the direction of the pores. Even though the availability and costs of green onions or any other vegetable is better compared to human or animal tissue, their use is somewhat limited to certain tissues and for now only limited to just a few cell layers thick tissue. Furthermore, the degradation of the cellulose present in vegetables pose a foreseeable problem due to the absence of cellulase in the human body.

FIGURE 7.

(A) Schematic illustration of decellularizing a tissue to mimic the morphology of a native tissue. (B) Haematoxylin and Eosin stained section of a porcine aorta (a) before decellularization, (b) after SDS treatment and (c) after supercritical CO₂ treatment; reproduced with permission [139]. Scale bar represents 50 µm.

Since there is almost no better way to exactly mimic a tissue morphologically than through the use of a decellularized tissue, researchers have tried for over 25 years to use this method to mimic the native tissue morphology [141]. However, some issues such as removal of chemical remnants in the decellularization process [142] and the immunogenic response that could be elicited by the presence of DNA, collagen type VI or galactose‐α‐1,3‐galactose (α‐Gal) [143], remain unsolved challenges. In addition, there is a trade‐off where either valuable ECM components are lost or unwanted components, such as DNA, remain in the tissue after processing [144]. To summarize, using decellularized tissue for morphological biomimicry is a viable strategy when current techniques are unable to replicate the morphology of the tissue.

3.4. Templating

Another way to morphologically mimic a scaffold is by using a template for cells to grow in. These templating techniques have been the main strategy in creating conduits for nerve regeneration or vasculature [145, 146]. Especially in recent years smaller and more complex microstructures were developed [147, 148]. For example, Jenkins et al. described how a conduit with microchannels was created by electrospinning a blend of poly(serinol hexamethylene urea) and PCL onto sucrose fibers using a dual electrospinning setup [149]. The resulting fibers contained highly aligned microchannels of less than 500 µm, which had a similar morphology to a nerve and showed to stimulate neurite growth and extension by the presence of ß‐III tubulin staining throughout the microchannels. A similar approach of using sugar as a template was performed by Wu et al., where sugar fibers where winded around a mandrel and dissolved afterward to create circumferential oriented microchannels envisioning a vascular graft [150] (Figure 8). The size of the circumferentially aligned channels was approximately 56 µm and much like the native situation the seeded vascular smooth muscle cells showed a circumferential alignment. A different approach performed by Huling et al. was to make a negative template of the whole kidney vasculature [151]. This was done by injecting a rat kidney with PCL. After solidification, the renal tissue was digested leaving the PCL vasculature cast behind. That cast was coated in collagen and then washed with acetone to remove PCL leaving a hollow collagen microvascularized scaffold behind. SEM analysis showed the preservation of the delicate and complex structures, such as the glomerulus ultrastructure. The seeded cells inside the cast formed a uniform and complete monolayer in the channels. Gershlak et al. showed that templates for vasculature can also be made from spinach leaves, since it shares a similar branching structure [152]. Perfusion tests within the leave template showed that the vascular network remained intact and seeded human umbilical vein endothelial cells (HUVECs) were able to populate the network. The use of spinach leaves is a simple and cheap solution that could easily be up‐scaled for mass production to create templates, although plant vasculature does not have a venous system like mammalians do, which might present a drawback of such solution.

FIGURE 8.

(A) Creating a template that mimics the architecture of a tissue. (B) Resulting sugar fibers after being winded around a mandrel and the resulting scaffold after dissolving the sugar fibers; reproduced with permission [150]. Scale bar represents 1 mm in the overview images and 200 µm in the close‐up images.

Depending on what sacrificial material or solvent is used, there is always the chance that remnants of the template remain that can be cytotoxic or impartial dissolution that can block the template itself from being washed away from the scaffold [153, 154]. In addition, the majority of the published studies that use templating techniques are somewhat limited to the complex shapes that they can achieve, resulting in channels that run parallel or simple vascular network structures, which do not contain any micro capillaries [128, 155]. On the other hand, complex structures such as the kidney vasculature have been made through templating, showing the possibility to achieve this complexity [151]. Nonetheless, using templates to morphologically mimic organs remains a valuable and often pursued strategy.

4. Biological Biomimicry

Many organs have a specific biological microenvironment. Examples are the low oxygen tension in cartilage [156], the electric conductivity of the atrioventricular bundle in the heart [157], the high blood pressure of the glomerular capsule in the kidney [158], or the presence of high amounts of growth factors in the bone marrow [159]. This natural environment for each of these organs is vital for their functioning. Therefore, attempting to replicate that specific biological environment has also been a topic of many studies. Strategies to obtain biological mimicry can range from improving cellular attachment by mimicking cellular anchoring points [160], conditioning the scaffold by placing it in a more naturally mimicking environment [161], adding growth factors in order to promote rapid growth and maturation of the tissue [162], or using decellularized organs as a base material for creating scaffolds [163].

4.1. Improving Cellular Attachment

The lack of cell attachment can have a great influence on the success of a scaffold [164]. High cell adhesion determines the initial amount of cells that attach right after seeding and has also a strong influence in the longer term when cell‐seeded scaffolds are subjected to dynamic culturing conditions [165, 166]. Therefore, several studies aimed at increasing the cell attachment properties of scaffolds through various methods. Tsai et al. covalently cross‐linked chitosan, a natural polymer known for its lack in cell attachment, through azide photocross‐linking to an Arginine, Glycine, and Aspartate (RGD) cell‐binding peptide sequence for bone tissue engineering [167]. This resulted in an improved amount of cells attached from day 1, which led to increased calcium deposition after 14 days in culture. Another method studied by Kao et al. used a polydopamine (PDA) coated PLA scaffold for bone tissue engineering [168]. The results showed that higher concentrations of PDA coating on the scaffold improved adipose derived stem cell attachment and stimulated proliferation. Besides the increased attachment, PDA coated membranes also stimulated angiogenesis as shown by the increased expression of angiogenesis related proteins angiopoietin‐1 and von Willebrand factor. Instead of chemical treatments, plasma surface treatment with inert gases can also be used to increase cell attachment as shown by Wang et al. [169]. PLA scaffolds were treated with cold atmospheric plasma for bone regeneration, resulting in improved cellular attachment. Atomic force microscopy analysis showed that the plasma treatment had an influence on the roughness of the fibers and that the contact angle decreased based on the duration of plasma exposure. A low contact angle indicates that the surface is hydrophilic. The hydrophilic surface in turn improved the MSC attachment. However, the contact angle increased from 25° at day 0 to 40° at day 10 as compared to 70° of the untreated scaffolds, indicating that the effect of plasma treatment decreases over time. A different approach without altering the structure or surface of the scaffold to increase the cell attachment was described by Cámara‐Torres et al. [125] (Figure 9). Media density and viscosity were changed to counteract the gravitational force to which the cells are subjected when seeding on PEOT/PBT scaffolds. The results showed that the seeding efficiency was higher and a homogenous cell distribution throughout the scaffolds was obtained when either density or viscosity was changed compared to the conventional seeding method. This initial improvement in seeding efficiency led to increased mineralization in osteogenic differentiation conditions, suggesting a specific role of cell density in differentiation mechanisms.

FIGURE 9.

(A) Schematic illustration of different seeding methods to improve seeding efficiency. (B) Fluorescent staining of cells in the bottom sides (top) and cross sections (bottom) of the scaffolds seeded with a conventional method (CS), with dextran to match the viscosity (MS‐Dextran) or with Ficoll to match the density (MS‐Ficoll‐Pq); reproduced with permission [125]. Scale bar represents 1mm.

Adding RGD peptides to scaffolds is known to have a beneficial biological effect in terms of cell adhesion and differentiation, it is also observed that the addition of RGD could decrease the mechanical properties [170]. Wang et al. even observed that spacing RGD‐motifs too tightly together at 37 nm can actually decrease proliferation, but increasing this spacing to 87 nm increased growth rate [171]. In summary, controlling the attachment of cells to a scaffold is important for the success of a tissue engineered graft and a valuable parameter for biologically mimicry.

4.2. Mimicking the Cellular Niche Biophysical Properties

Conventional culturing methods rely on mono culturing in a static manner on polystyrene plates with a steady supply of nutrients and oxygen. However, many organs are dynamic and have a specific biophysical environment specific to their function, which greatly differ from the stiff and flat environment offered by conventional culture plates. Examples comprises: i) an air‐liquid interface characteristic of the lungs to control carbon dioxide and oxygen exchange from the inhaled air [172]; ii) the necessity to apply a load to bone, which is otherwise remodeled [173]; iii) the critical need for oxygen in the kidney with specifically the proximal tubule cells requiring large amounts of oxygen in order to facilitate sodium transport [174], to mention a few. Therefore, some studies aim at mimicking a specific local biophysical environment to maturate their scaffolds. For example, Munir et al. compared adipose‐derived stromal cells cultured under 21% O₂ or 5% O₂ and confirmed that a reduced oxygen tension mimicking native cartilage conditions increased the deposition of cartilage specific markers such as SOX9 and collagen type II [175]. Besides, also the hypertrophic cartilage marker collagen type X decreased, indicating that the formed cartilage in reduced oxygen conditions is neither fibrocartilage nor hypertrophic. Doryab et al. mimicked the natural lung environment in multiple ways [176] (Figure 10). During the study, epithelial cells were cultured on top of a PCL gelatin membrane, which was lifted to create an air‐liquid interface. The resulting membrane was then stretched to mimic also the natural condition in the lung. The results showed that under dynamic physiological culturing the formation of cell‐cell tight junctions was increased compared to static culture conditions, as shown by the presence of Zonula occludens‐1 (ZO‐1). Stretching the samples beyond physiological conditions led to a decrease in tight junction formation, as well as an increase in apoptosis because of a lower cell number after stimulation. A different study performed by Pennings et al. used a custom‐made bioreactor to mature a bioengineered vascular graft [177]. The results showed that under dynamic culturing conditions the endothelial cells had a higher expression of the endothelial markers vascular endothelial cadherin (VE‐cadherin) and cluster of differentiation 31 (CD31). Shear stress‐induced genes cyclooxygenase‐2 (COX‐2) and Krüppel‐like Factor 2 (KLF2) were also highly expressed when subjected to flow. A recent study, however, showed that dynamically overstimulating scaffolds increased the expression of inflammatory markers [178]. Takeda et al, dynamically stimulated cartilage cells in a collagen hydrogel, which was entrapped in a collagen sponge, and found that inflammatory and cartilage matrix‐degrading enzymes ADAM Metallopeptidase With Thrombospondin Type 1 Motif 4 (ADAMTS4), Interleukin 1 (IL‐1) and matrix metalloproteinase‐3 (MMP3) were upregulated when the cyclic load was at a physiological load of 40 kPa or 10% deformation as compared to no loading or only 20 kPa loading. In addition, reactive oxygen species accumulation was observed when samples were dynamically loaded. When the antioxidant pyrrolidine dithiocarbamate (PDTC) was administered during mechanical loading, the upregulation of ADAMTS4 and IL‐1 receptor was no longer observed, thus partially compensating for the overloading conditions.

FIGURE 10.

(A) Schematic illustration of mimicking the natural environment of a lung by stretching the scaffold and using an air‐liquid interface to further maturate the cells. (B) ZO‐1 expression in respiratory epithelial cells in (left) static condition, (middle) cyclic stretch at 21% strain and (right) cyclic stretch at 35% strain; reproduced with permission [176]. Scale bar represents 20 µm.

Mimicking a complex environment of a tissue is not an easy task, since multiple elements could play a role in the environment. For example, in articular cartilage a complex anisotropic loading plays an important role besides low oxygen tension and different osmolarity of the local microenvironment [179, 180]. In vascular networks, the blood flow through the vessel lumen is regulated dynamically by the autonomic nervous system to contract or dilate [181]. Mimicking the local biophysical conditions and properties of the microenvironment of a tissue plays a key role in biological biomimicry and can help further tissue maturation.

4.3. Growth Factors

Growth factors are important in the natural regeneration process and are therefore widely used in regenerative medicine [182, 183, 184]. The general use for these growth factors is to enhance the natural healing process or to stimulate the differentiation of cells toward the targeted tissue (or organ) of interest. For example, Li et al. used nerve growth factor (NGF) to promote peripheral nerve regeneration [185]. Heparin was used to load NGF in chitosan scaffolds, resulting in an increase in initial Schwann cell attachment and proliferation after 5 days of culture. In addition, the number of neurites formed by the Schwann cells was increased in the presence of NGF. Instead of using heparin to load growth factors via electrostatic interactions, Park et al. chemically immobilized a Bone morphogenetic protein 2 (BMP‐2) mimetic peptide in a click‐cross‐linked hyaluronic hydrogel [186] (Figure 11). The bioactivity of the biomimetic BMP‐2 peptide showed similar results to BMP‐2 supplemented medium condition, as seen by ALP and bone specific markers osteonectin and osteocalcin expression. Because of the chemical immobilization, the release of BMP‐2 was more controlled over a longer period of time instead of an initial burst release. This in turn led to a more effective osteogenic differentiation. A different strategy was applied by Rezaii et al., who used curcumin nanoparticles incorporated in collagen‐chitosan scaffolds to regulate the release of Transforming growth factor beta 1 (TGF‐β1) and Smad7 for wound healing [187]. TGF‐β1 is known to enhance extracellular matrix production and Smad7 is an inhibitor of the TGF‐β family. The results of the in vivo study showed that with the addition of the curcumin nanoparticles, the mRNA expression of TGF‐β1 and Smad7 were upregulated after 3 days. This upregulation was no longer observed after 15 days and the wound area was significantly reduced. TGF‐β1 expression is known to decrease when the wound is closed, indicating that the particles decreased the duration of wound closing [188].

FIGURE 11.

(A) Schematic of using a mimetic protein that attaches to a hydrogel. Adding growth factors to boost the growth of a newly implanted scaffold. (B) ALP staining during the osteogenic differentiation of stem cells (left) in culture medium (control) or osteogenic medium containing BMP‐2 or biomimetic BMP‐2 (BP) and Osteonectin expression (right); reproduced with permission [186]. Scale bar represents 500 µm.

The addition of growth factors to promote regeneration is not valid for all cases. This was shown by Pan et al. by blocking a growth factor receptor to promote meniscus fibrocartilage formation [189]. During the study, an epidermal growth factor receptor (EGFR) inhibitor was injected in vivo at the implantation site of a collagen type I scaffold. After 16 weeks, the scaffolds injected with the EGFR inhibitor showed a similar matrix to that of a healthy meniscus, as observed from Safranin‐O staining for GAG production. In addition, there was more collagen type II expression in the scaffolds with EGFR inhibitor and had a lower Mankin's score, which is used to grade osteoarthritis.

Growth factors are vital in normal tissue morphogenesis and function, which makes them an interesting approach for biological biomimicry. However, uncontrolled or prolonged release of growth factors are associated with multiple forms of cancer [190, 191, 192]. Another issue is that simply adding the growth factors to a scaffold in vivo would result in a systemic response. Therefore, growth factors controlled release or immobilization is needed in designing bioactive scaffolds [193, 194, 195]. Even though the addition and release profile of growth factors to a scaffold is complex to control, it remains a viable strategy for biological biomimicry.

4.4. Decellularized Tissue for Hydrogels

Unlike the previously discussed use of decellularized ECM to mimic the tissue morphology, in this section decellularized ECM is discussed as bulk material for hydrogels and bioinks to mimic the biological microenvironment. For example, Seo et al. exploited a supercritical CO₂ ethanol (EtOH) treatment to decellularize heart tissue, which was then combined with collagen type I and used as an injectable hydrogel [196]. As comparison for the supercritical CO₂ EtOH, a common SDS Triton X‐100 detergent based method was used. Both methods proved to remove the DNA, but hematoxylin and eosin staining showed that the supercritical CO₂ EtOH treatment was better at retaining the collagen bundles of the native tissue. GAG and Collagen contents were better preserved, in addition to retaining Fibroblast growth factors (FGF), vascular endothelial growth factor B (VEGF‐B) and platelet factors. The decellularized ECM was able to form a gel. Finally, subcutaneous injection for 3 days showed that the supercritical CO₂ EtOH treated gels improved neovascularization as compared to the detergent based gels. Another study by Lewis et al. used decellularized liver ECM to create hydrogels for bile duct regeneration [197]. The results showed that increasing the concentration of decellularized liver ECM also increased the stiffness of the gels from less than 50 Pa at a concentration of 2 mg/mL to just below 1000 Pa at 20 mg/mL. Cholangiocytes cultured in the 2 mg/mL gels showed the highest amount of branching structures, thus indicating that the soft hydrogels allowed the infiltration and migration of cells, in turn resulting in better tissue morphogenesis. A lower concentration, however, could not be obtained due to the mechanical stability of the gel. In addition, the cholangiocytes were capable of forming complex duct formations after 7 days of culture. A different approach performed by Toprakhisar et al, was to use decellularized tendon to form a bioink for bioprinting [198]. The decellularization process proved to be successful in removing 97% of the DNA. The decellularized ECM was composed mostly out of collagen type I, whereas other proteins such as fibromodulin and prolargin at much smaller concentrations. In addition, during the decellularization process the collagen band patterns remained intact and could be used as a bioink for bioprinting. An additional bioprinting approach done by Ali et al. used decellularized kidney that was photo‐cross‐linkable [199] (Figure 12). The decellularized kidney ECM was chemically modified through a methacrylate reaction that allowed the ECM to cross‐link under ultraviolet light exposure. The results showed that growth factors such as BMP‐7, epidermal growth factor (EGF), NGF‐1 and TGF‐β1 remained active even after methacrylation. The bioink was able to create a 16‐layered scaffold that did not dissolve after 7 days unlike the unmethacrylatd decellularized ECM, which dissolved within 1 day after printing. Even after 2 weeks of culture, bioprinted primary human kidney cells maintained their tubular marker Aquaporin‐1 (AQP1).

FIGURE 12.

Using decellularized ECM as base material for scaffolds to mimic the natural environment. (B) Histological sections of a native porcine kidney before and after the decellularization process (left) and the scaffold printed with the ECM based bioink (right); reproduced with permission [199]. Scale bar represents 200 µm.

Using decellularized tissue as hydrogels or bioinks to recreate the targeted tissue biological microenvironment is a topic that many researchers pursue [200, 201, 202]. However, as mentioned earlier, some challenges need still to be tackled when using decellularized ECM, such as the removal of chemical remnants and the immunogenic response and the possible loss of valuable ECM components [142, 143, 144]. In addition, the ECM of various tissues is highly complex, organized and anisotropic, which when pulverized and blended to make an homogenous hydrogel will most likely lose this organization [203, 204, 205]. Summarizing, using decellularized tissue as bulk material is a valuable strategy to biologically mimic the complex cellular microenvironment, though an exquisite balance between retaining the ECM structural properties and bioactivity is needed.

5. Future Outlook

This review provides an overview of the biomimicry strategies used in tissue engineering. We identified three main categories on which biomimicry criteria are pursued when designing scaffolds for tissue regeneration: mechanical, morphological and biological biomimicry. Each of these categories are important on their own and understanding how to mimic one aspect can give valuable insights on successfully regenerating a tissue. It is important to note that these subdivisions are not mutually exclusive to each other. Certain strategies or changing a single parameter can influence multiple forms of biomimicry. For example, the mechanical properties are influenced by changing the bulk material. Likewise, a different biological response would be implicitly expected [206]. Another example comes from decellularized tissues where the morphology is retained, but they can also trigger a biological response due to growth factors or ECM components left after the decellularization process [137].

Strategies for mechanical biomimicry have undergone a lot of development over the last years. New materials have been created or modified to alter the mechanical properties [207, 208], the geometrical complexity of the scaffolds have been increased [209] and complex mechanical behavior has been introduced and studied [210, 211]. Future development for mechanical biomimicry should focus, in our opinion, on four subjects that when combined could create a successful mechanically mimicking scaffold. First, new materials should be developed and characterized, thus leading to expanding the palette of potential candidates for scaffold fabrication. The second focus point is the development of new biofabrication techniques that allow the creation of higher resolution scaffolds and more complex designs. In line with this, thirdly the role of scaffold design with respect to the mechanical properties of the scaffold should be studied more in depth and correlated to the biological effects that are observed. Finally, researchers should focus on what happens to the mechanical properties of the scaffolds, and consequently of the biological constructs, over time. The scaffold will be remodeled over time and little is known of what happens to the mechanical properties after longer periods of in vivo and post in vivo implantation. Combining this knowledge could ultimately lead to the development of scaffolds that are tailored to the specific patient needs with a combination between the biomaterial used and the structural design that mimics the mechanical properties of the native tissue.

Also morphological biomimicry has undergone a lot of advances in terms of morphology [212, 213], improved decellularization protocols [214], better templating techniques [155], and improved biofabrication techniques such as single cell bioprinting that can accurately deposit single cells [215]. This could lead to the fabrication of small tissues mimicking the cellular organization of the native tissue. Future developments could look into how to create large‐scale biological constructs, specifically on how to vascularize and innervate them with the final goal of improving tissue integration and homeostasis after implantation. Many of the currently fabricated scaffolds, and hence biological constructs, are made on small scale. This brings a potential problem since the vascularization of small scaffolds is often neglected and not needed, although for clinically relevant defects it is the main requirement for success. Working out these obstacles would prove beneficial in creating a morphologically mimicking scaffold.

Many of the strategies that focus on biological biomimicry have improved, such as providing alternatives to static cultures [216, 217], improved understanding of the dynamic environment [218, 219] and enhancing cell scaffold interactions [220, 221]. Future development for biological mimicry should be focused on the dynamic culturing of scaffolds to standardize testing conditions. Further developments on bioreactors to make them more readily available at affordable costs should aid in creating a standardized dynamic culture protocol for each tissue. Cell‐material interactions should be further studied to increase the level of communication between synthetic and biological matters in order to reach a better functionality of the engineered tissue. In relation to decellularized ECM, it will be also important to achieve decellularization protocols able to avoid an immune response while keeping the ECM components. Solving these issues all together could aid in the fabrication of a scaffold that biologically mimics the native environment.

A future roadmap for the development of a fully biomimicking scaffold should not just focus on a single aspect of biomimicry. A more holistic approach toward biomimicry should be already conceived from the material development stage. We should also focus on how to get the morphology right, how to create a scaffold that is on scale of the native tissue, taking into consideration the multiscale properties of the targeted tissue, and how to tackle vascularization and innervation. In addition, to further mature and minimize the gap between in vitro and in vivo, the dynamicity of the local cellular environment should be recreated in vitro as close as possible to that of the native tissue. It is important to note that a lot of progress has been made in the development of a biomimicking scaffold and certain strategies are already at the preclinical evaluation stage in several animal models [222, 223].

A contradicting question arises whether how important it is to create a scaffold that fully mimics all aspects which after implantation are remodeled over time, or in other words how close should a scaffold mimic a tissue to achieve an optimal repair after implanted. If we learn from autologous tissue transplants, such as the reconstruction of the pharynx of cervical esophagus using the jejunum, success rates hcompared to an artery ave been reported between 80 and 90% [224]. Though, the epithelial cells of the jejunum have a different function [225] and the general morphology is different [226, 227]. A similar example is transplanting the saphenous vein for arterial reconstruction, it seems that only a resemblance is needed [228]. Yet, compared to an artery the saphenous vein has weaker mechanical properties [229], a thinner tunica media layer and not a smooth tunica intima [230], and the pressure and flow under physiological conditions is much lower [231], which account for a 58% survival rate in patients 5 years post‐implantation. This leaves a lot of room for improvement. One of these improvements could be grafts or scaffolds that better mimic the original tissue. This was also shown in a coronary artery bypass graft. In this meta analyses the replacement of the coronary artery with the radial artery had an improved outcome over the saphenous vein [232], indicating that it is important to mimic the original tissue as close as possible.

Future studies should also aim at applying these biomimetic strategies for advanced 3D in vitro human models of tissues and organs that could be used to better understand pathophysiological mechanisms and test new therapeutic approaches. Applying biomimicry strategies to these systems could significantly enhance their physiological relevance. For instance, the dynamic modulation of substrate stiffness has been shown to influence cardiac fibroblast activation, highlighting the importance of mechanical cues in tissue models [233]. Morphological biomimicry has been advanced through the creation of perfusable vascularized cancer niches using multi‐material 3D bioprinting, enabling the study of tumor progression in a more physiologically relevant context [234]. Biological biomimicry has been exemplified by the development of ultrathin extracellular matrix‐derived membranes in organ‐on‐a‐chip devices, facilitating tissue‐specific morphogenesis and barrier function [235]. Combining these three axes of biomimicry into engineered in vitro systems can result in next‐generation models that more accurately recapitulate organ‐level function and intercellular crosstalk, thus bridging the gap between simplified 2D cultures and complex in vivo biology.

A major challenge remains in translating such sophisticated biomimetic constructs into reproducible and scalable systems suitable for widespread adoption. Variability in biological materials, complexity in fabrication workflows, and dependence on manual interventions often limit throughput and standardization. Tackling these issues will require the integration of automation, quality control systems, and robust biofabrication protocols that accommodate variability without compromising functional fidelity. Furthermore, modular design principles could offer a pathway to scale up while preserving biological complexity [236]. Standardized bioinks, multi‐material printing platforms, and interoperable bioreactor systems will play a pivotal role in advancing reproducibility. Overcoming these barriers will be essential not only to validate in vitro models as reliable tools for research and regulation, but also to accelerate the clinical translation of biomimetic scaffolds.

6. Conclusion

Mimicking only one aspect of biomimicry proofs to be a challenge, whether it is to obtain the correct mechanical properties for a mechanically complex tissue such as bone [237], mimicking the exact structure and organization of a liver [238], or replicating the complex kidney environment for each kidney cell type [239]. To fully mimic all these aspects presents an outstanding challenge unattainable with our current knowledge. However, over the years research on biomimicry has improved from the early reports from Otto Schmitt in 1957 to biologically mimic the electrical action of a nerve [240] to complex hydrogel yarns that align cells, mimicking the native architecture of tendons [241]. For regenerative medicine, biomimicry is important in order to manufacture a fully functioning tissue or organ. Whenever a strategy successfully employs mechanical, morphological and biological biomimicry, a fully functioning regenerated tissue should be expected.

Conflicts of Interest

The authors declare no conflict of interest.

van Kampen K. A., Mota C., and Moroni L., “A Tissue Engineering's Guide to Biomimicry.” Macromol. Biosci. 25, no. 10 (2025): e00093. 10.1002/mabi.202500093

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.