Microengineered Gradient Hydrogels for Mechanobiology

ABSTRACT

Extracellular biophysical gradients have a profound impact on cell behavior and biological processes during development, wound healing, and disease progression. Stiffness gradient hydrogels engineered to mimic the native extracellular matrix have emerged as powerful tools to probe how cells respond to these varied biophysical cues within their material surroundings, which gives rise to the concept of designer hydrogels to control cell and tissue function. The convergence of material science and microfabrication technologies provides exciting opportunities to identify new mechanobiological understanding and mechanomodulatory therapies by improving the physiological relevance of biomaterials‐based experimental platforms. This review reflects on the motivation to investigate stiffness gradients, describes the key considerations for developing application‐specific stiffness gradient hydrogel, and summarizes the various approaches that have been introduced to pattern the microscale material properties. The current and emerging applications of stiffness gradient platforms are presented, with particular emphasis on fundamental durotaxis studies, high‐throughput screening of cell‐material interactions, in vitro disease modeling, and tissue regeneration. The review concludes by discussing the standing challenges and positive trends with the aim of providing directions for future research in microengineered gradient hydrogel technologies.

Gradient hydrogels are used to mimic the mechanical heterogeneity in native tissues, offering powerful in vitro platforms to study cell‐material interactions in diverse pathophysiological contexts. Here, we present a comprehensive review of the design and experimental considerations for stiffness gradient hydrogels, discussing exemplary achievements and standing challenges for mechanobiology research and its potential translational applications.

1. Introduction

The behavior of nearly all cells in the human body is influenced by biophysical signals within their cellular microenvironment. At the core of this microenvironment is the extracellular matrix (ECM), a scaffold that not only provides structural integrity to tissues but also plays important instructive roles in various biological processes [1]. While ECM is made of the same family of sugars and proteins (e.g., collagen, laminin, glycosaminoglycan, proteoglycan), their composition and spatial organization are highly specialized across individual tissues throughout the body, endowing them with distinct functions and, naturally, biophysical properties [2].

Cells interact directly with the ECM primarily via receptor‐mediated mechanosensing through integrins and other transmembrane receptors, which physically connect the ECM and the cellular cytoskeleton [3]. This cell‐matrix coupling allows cells to probe their surrounding matrix via cytoskeletal contractility and apply traction forces at the cell adhesion sites. Then through a process called mechanotransduction, mechanical and structural cues from the ECM can trigger intracellular signaling cascades that alter gene transcription events which ultimately affect cell proliferation, migration, and differentiation, among other activities [4]. Notably, integrin receptor clustering is dependent on the ECM mechanics and is more pronounced at higher matrix stiffness [5]. In turn, this promotes cytoskeletal tension and nuclear translocation of mechanosensitive transcriptional regulators, such as yes‐associated protein (YAP) and transcriptional co‐activator with PDZ‐binding motif (TAZ), which are key to regulating gene expression, protein synthesis, and downstream cellular behaviors. The functions of YAP and TAZ are also dependent on both the subcellular abundance and distribution, which are modulated by ECM stiffness and topography. At the tissue level, such physical interactions mediated by cell‐ECM and cell‐cell adhesions drive diverse, complex processes such as epithelial morphogenesis during embryogenesis, ECM remodeling in cancer and fibrosis, and tissue repair after injury [6, 7]. In short, the way cells perceive their biophysical environment plays important roles in health and disease.

Research in mechanobiology has long benefited from the persistent development and refinement of hydrogels‐based experimentation platforms that allow spatiotemporal control over the cellular microenvironment [8, 9]. By making it possible to recreate the biophysical and biochemical properties of the native ECM, these in vitro platforms offer the promise to expand our capabilities to better understand how cells sense and respond to matrix cues in specific pathophysiological contexts. An archetypal example was shown in seminal studies where stem cell lineage commitment can be directed simply by matching the substrate stiffness to the target tissue type (e.g., neurogenic at 1 kPa, adipogenic at 3 kPa, myogenic at 12 kPa, and osteogenic at 30 kPa), establishing an important design principle in creating biomaterial scaffolds for tissue regeneration [10, 11, 12]. More recently, researchers used hydrogels of variable stiffness to demonstrate that tissue mechanics (400 Pa vs. 60 kPa) impact embryonic stem cell response to Wnt factors during mesoderm differentiation [13], and to interrogate the influence of substrate stiffness (7–50 kPa) on T cell activation in the context of adoptive immunotherapy [14]. In addition, an overarching goal of many of mechanobiological studies is to identify new therapeutic targets, such as recent developments in cancer, where ECM‐targeted drugs aim to inhibit the production of fibrotic ECM components or specific matrix interactions that have been implicated in cancer cell invasion [15]. Therefore, hydrogel culture systems where the mechanics can be tailored to replicate aspects of healthy and diseased tissues have been harnessed to understand fundamental biological processes, and how to control cell behavior by modulating their microenvironment.

Early mechanobiology studies have relied on uniform stiffness hydrogels and focused on examining cell responses to the substrate bulk properties [11, 13]. However, many native tissues exhibit spatially heterogeneous mechanical landscape, often in the form of microscale gradients [2]. This heterogeneity is prominent not only in the interface region of healthy tissues such as at neuromuscular junctions and muscle‐tendon boundaries, but also in disease processes like fibrosis. Modeling such micro‐scale spatial heterogeneity using traditional methods of fabricating bulk hydrogels poses difficult and often insurmountable challenges. This limitation provided the impetus for leveraging advances in microfabrication strategies to create stiffness gradient hydrogels that mimic the stiffness range and gradient profile characteristic of the target tissue type or biological process. A key aspect of these gradient materials is the continuous and tunable variation in mechanical stiffness across a defined spatial range, from hundreds of micrometers to a few centimeters in size, reflecting the high precision requirements.

The concept of precision engineering of stiffness gradient hydrogels to model the physiologically relevant tissue heterogeneities is not new, but has gained renewed traction in recent years. This growth is a result of the recognition that gradient gels offer a unique combination of high‐content analysis and the potential to gain additional insights into the complex cell‐material interplay, compared to testing discrete parameter values. For instance, the ability to screen stiffness‐mediated behavior on continuum of conditions can reveal potential threshold effects or nonlinear responses [16, 17]. Meanwhile, gradient hydrogels have become imperative tools for addressing the fundamental question of how cells sense ECM stiffness to guide their migration, with diverse implications ranging from limb budding to cancer metastasis, but is technically very difficult to study in vivo [18, 19]. In this regard, stiffness gradient hydrogels lend themselves as a bridging technology between simplistic bulk hydrogels and sophisticated animal models, providing an efficient means to systematically probe various aspects of the complex tissue landscape, while still enabling direct observations of the cell and tissue behavior.

This review aims to present some of the key ongoing innovations in biomaterials and microfabrication technologies that are setting the basis for future biophysical gradient platforms in mechanobiology research and applications. We will discuss the spatiotemporal biophysical complexities in the native ECM and how these can be mimicked using biomaterials approaches, focusing on the major area of stiffness gradients (Figure 1). We will also describe critical advances in our understanding of the cell‐ECM interplay that has been enabled using microengineered stiffness gradient hydrogel systems and provide an outlook on the emerging directions in gradient fabrication. By encouraging broader adoption of these platforms within the mechanobiology community, we expect to open new avenues for scientific inquiry, offering a pathway to identify potential mechanomodulatory strategies for clinical treatment of diverse diseases.

FIGURE 1.

Microengineering stiffness gradient hydrogels for mechanobiology. (A) The stiffness of hydrogels can be tuned by either varying the polymer concentration or degree of crosslinking. Hydrogels can also be designed to mimic in vivo ECM biological cues, often by incorporating adhesive and degradability motifs found in the native ECM. (B) Design and fabrication of the stiffness gradient system need to factor in the stiffness range, gradient strength, and gradient size, depending on the cell/tissue type and readout functionality required. (C) Choice of cell culture format will need to balance between biological relevance and feasibility for practical implementation. (D) Engineered gradient hydrogels have found utility as tools for the study of fundamental cell‐matrix interactions and biomedical applications, including to create in vitro models and in tissue regeneration.

2. Matrix Stiffness From a Tissue and Materials Perspective

2.1. Definition of Stiffness

“Stiffness” is a general term often used to describe the mechanical property of substrates and matrices for cell culture. The basic definition of stiffness from a materials science point of view is based on the force required to achieve a certain deformation of a structure: [20]

$$\mathrm{Stiffness}=\mathrm{Load}/\mathrm{Deformation}$$

Although the stiffness of a tissue or substrate can be quantitatively analyzed, there are many possible configurations of “load” (e.g., compressive force, tensile force) and measurable points of deformation (e.g., deformation, strain). For this reason, the reported stiffness values in literature should be interpreted in the context of the measurement technique. Moreover, the concepts of moduli and stiffness are intimately related, and thus are often used interchangeably. However, when comparing studies in the literature, it is important to note the distinction that stiffness is a property of a structure that depends on its shape. In contrast modulus is solely dependent on the intrinsic properties of the materials that compose that construct [2].

2.2. Stiffness Gradients in Living Tissues

In the context of mechanobiology, it is the most common practice to define stiffness as the material's Young's modulus (kilopascals, kPa), which is a measure of its ability to undergo non‐permanent, elastic deformation. The Young's modulus of local tissues is thought to be largely governed by the density of fibrillar type 1 collagen, being the most abundant ECM protein [2]. In general, adult tissues exhibit a characteristic Young's modulus that increases from soft brain and liver (several hundreds of pascals to several kPa), to slightly stiffer striated muscle (around 10 kPa), and to much stiffer cartilage and bone matrices (over hundreds of kPa) [2, 21]. While Young's modulus is widely adopted for its simplicity and comparability, it incompletely captures the time‐dependent and non‐linear mechanics of many biological ECM and tissues. This has led to recent push for the development of viscoelastic culture substrates to better recapitulate the in vivo microenvironment [22, 23]. Though outside the focus of this review, it is important to note that viscoelastic materials dissipate applied forces, which can dramatically change how the surrounding matrix responds and stores cell‐generated traction forces, as well as how cells perceive external mechanical forces [24]. Majority of literature agrees with elastic mechanical properties being the major regulator in cellular mechanotransduction behaviors, whilst acknowledging that it will be necessary to better understand and consider viscoelasticity in engineered hydrogel matrices [3].

Beyond basic variation in the range of mechanical stiffness across different tissue types, localized spatial gradients in stiffness are necessary for physiology (Table 1). For instance, the stiffness of the white matter in the healthy brain is 39% higher than that of the adjacent gray matter, a property critically dictating neurodevelopment [25]. The hyaline cartilage in osteochondral tissue comprises a continuous stiffness transition from ∼600 kPa near the subchondral bone (deep zone) to ∼240 kPa near the surface (superficial zone), which is essential for functional force transmission and stable joint motion [26]. Furthermore, some gradient tissue structures natively bridge two tissue types with drastically different stiffnesses. Notable examples include the cartilage‐to‐bone, muscle‐to‐tendon, tendon‐to‐bone, and ligament‐to‐bone tissue interfaces [27, 28, 29, 30]. Recent research on anterior cruciate ligament (ACL)‐bone interface has revealed that the sub‐30 µm scale stiffness gradient profile is uniquely different between the femoral and tibial junctions, significantly changing the way we think about the cell microenvironment in soft‐to‐hard interfaces during injury and regeneration [31].

TABLE 1.

Mechanical properties of biological tissue stiffness gradients reported in literature.

Tissue type	Species	Technique	Modulus value (condition)	Ref.
Intra‐tissue variation
Brain gray‐white matter boundaries	Bovine	Flat‐punch indentation	1.389 ± 0.289 kPa (gray matter), 1.895 ± 0.592 kPa (white matter)	[25]
Hyaline cartilage	Human	AFM indentation	240 ± 50 kPa (superficial zone), 600 ± 50 kPa (deep zone)	[26]
Tissue interfaces
Periodontal ligament (PDL)‐alveolar bone (AB) enthesis	Rat	AFM indentation	Healthy: 57.1 ± 13.5 kPa (PDL), 25.9 ± 1.5 MPa (AB) Periodontitis: 35.6 ± 12.4 kPa (PDL), 12.4 ± 1.5 MPa (AB)	[30]
Anterior cruciate ligament (ACL)‐bone enthesis	Human	AFM indentation	Femoral: 40 ± 5 MPa (ACL), 760 ± 40 MPa (femur bone) Tibial: 27 ± 5 MPa (ACL), 1262 ± 218 MPa (tibial bone)	[31]
Injury/Disease
Heart, post‐myocardial infarction	Mice	AFM indentation	Healthy: 17.94 ± 0.39 kPa Infarcted: 4.21 ± 0.16 kPa (Day 1), 31.38 ± 0.75 kPa (Day 7), 53.23 ± 0.75 kPa (Day 14), 90.22 ± 2.97 kPa (Day 28)	[36]
Lung, idiopathic pulmonary fibrosis	Human	AFM indentation	1.96 ± 0.13 kPa (healthy), 16.52 ± 2.25 kPa (fibrotic)	[37]
Liver, injury/diet‐induced fibrosis	Mice	AFM indentation	∼1 kPa (healthy), ∼6 kPa (fibrotic)	[35]
Visceral adipose, obesity‐related diabetes	Human	AFM indentation	4.48 ± 4.81 kPa (healthy), 11.50 ± 16.79 kPa (fibrotic)	[34]
Breast, cancer	Human	AFM indentation	Healthy: 1.13‐1.83 kPa Malignant tumor: 0.31‐0.75 kPa (core, peak 1), 1.54‐1.99 kPa (core, peak 2), 2‐20 kPa (periphery)	[40]
Colorectal, peritoneal metastasis	Human	AFM indentation	Average across 5 patients: ∼7.1 kPa (healthy), ∼16.6 kPa (neoplastic)	[43]
Glioma	Human	AFM indentation	Normal: ∼10‐180 Pa (gliosis) Malignant: ∼50‐1400 Pa (low grade glioma), ∼70‐13500 Pa (glioblastoma)	[44]
Pancreas, cancer	Human	Magnetic Resonance Elastography	2.47 ± 0.11 kPa (healthy), 6.06 ± 0.49 kPa (cancer)	[45]
Thyroid, cancer	Human	Shear wave elastography	Elasticity index: 15.9 ± 7.6 kPa (healthy thyroid gland), 150 ± 95 kPa (malignant nodules)	[46]
Vulva, cancer	Human	AFM indentation	3.07 ± 1.6 kPa (healthy), 1.58 ± 0.85 kPa (carcinoma)	[47]

Tissue stiffness gradients are also prominent in disease or injury (Table 1). In the context of fibrosis, tissue stiffening is caused by abnormal distribution of ECM resulting from an imbalanced matrix deposition and degradation, along with the release of inflammatory signals as well as other cellular events (e.g., myofibroblast activation, increased contractility) [32]. This process is the body's normal response to facilitate wound healing, but disruptions to ECM homeostasis can lead to permanent changes in tissue biochemical and biophysical properties which directly affect the residing cells via mechanobiological pathways [33]. Since the development of tissue scarring and fibrosis typically initiates in localized regions, this creates a microscopic stiffness gradient across the boundaries of healthy and fibrotic tissues. This condition can affect nearly every tissue in the body, including the heart, lungs, liver, and visceral fat [34, 35, 36, 37, 38]. In the cancer microenvironment, a higher tumor ECM stiffness is linked to dense overgrowth of packed basement‐membrane‐like connective tissue, which forms the basis of metastatic behavior in cancer cells [39]. For example, the Young's modulus of healthy mammary tissue (1.13–1.83 kPa) can increase by a full order of magnitude up to 20 kPa in patients with malignant breast cancer (Table 1) [40]. Similarly, cancerous lesions are typically stiffer than the adjacent normal tissues [41]. Finally, in a developmental context, a microscale stiffness gradient (100‐300 Pa) was observed in vivo and shown to guide neural crest migration along the gradient provided by the mesodermal tissue [42]. These cellular migratory responses to spatial stiffness gradients, collectively known as durotaxis, are thought to play a crucial role in various biological phenomena, making it a subject of intense research interest in mechanobiology [18].

So far, a remarkable degree of the bidirectional cell‐matrix coupling has been revealed using bulk mechanical substrate models. As mechanical characterization of biological tissues becomes more routine, and combined with the growing interest in the mechanobiology of diseases, microengineered stiffness gradient hydrogel platforms are now well‐positioned to expand our spatial understanding of the ECM mechanics and its role in pathophysiology.

2.3. Hydrogels as ECM Mimics

In the context of biomaterials, hydrogels constitute an ideal framework to organize and present various biophysical and biochemical cues to cells. Broadly, hydrogels are hydrophilic 3D polymeric networks that can absorb and hold large amounts of water, resembling many of the properties of biological tissues. Various tools are available that enable researchers to modify these materials, allowing them to control the bioactivity and tune the physical properties. Indeed, this versatile class of materials have already proved invaluable in mimicking specific features of the in vivo microenvironment, providing a highly controlled platform to study fundamental cell‐matrix interactions and facilitate tissue engineering applications [8, 48]. Hydrogels used for cell culture can be formed from a vast array of natural, synthetic, and semi‐synthetic materials. Each hydrogel system has its own set of advantages and disadvantages. Thus, the choice of hydrogel system will primarily be dependent on the specific application and objectives.

Natural hydrogels have a long history for cell culture and in vitro model applications due to their superior cytocompatibility, but their batch‐to‐batch variability can pose challenges in experimental reproducibility. The most widely used of these are purified ECM proteins (e.g., gelatin [49], collagen [50], laminin [51], Matrigel [52]), which offer a high level of biochemical ECM mimicry for in vitro studies. Alternatively, decellularized ECM may better provide the contextual biochemical and biophysical cues required to model specific tissue niches [53, 54]. Carbohydrate‐based hydrogels (e.g., alginate [24], hyaluronic acid [55], Gellan gum [56, 57]) have also been widely explored. These materials form the basic building blocks for many 3D culture systems, though chemical functionalization is typically needed for carbohydrate‐based systems to impart cell adhesivity. Natural hydrogels especially those of protein‐based tend to have lower stiffnesses (Pa to kPa range), because their mechanical properties are limited by the amount of polymer needed for network formation, as well as by the solubility in water.

Nevertheless, natural polymers can be chemically modified to provide greater control and flexibility over their mechanical properties, creating semi‐synthetic polymers. The most common strategy is to adapt the natural base polymers for photochemical reactions through functionalization with methacrylate, allyl, or norbornene groups, for instance [58, 59]. To illustrate an example, physically crosslinked gelatin results in hydrogels in the sub‐kPa to low‐kPa range that are also highly thermosensitive [49]; whereas methacrylated gelatin can be photopolymerized to yield mechanically stable hydrogels, with tunable stiffness spanning ∼1 kPa to over 100 kPa [60, 61].

On the other hand, fully synthetic hydrogels (e.g., polyacrylamide [62, 63], polyethylene glycol [64]) provide a chemically defined system with mechanical properties that can be modulated over a wide range (from Pa to MPa), making them a popular choice for mechanobiology studies. In contrast to naturally derived hydrogels, the stiffness of synthetic hydrogels can be modulated by altering the polymer crosslinking density without simultaneously changing the density of cell adhesion ligands [65]. Such property is especially important in mechanobiology studies to parse out the individual effect of mechanical cues on cell responses. However, these systems typically involve synthetic chemistry methods for hydrogel crosslinking and biofunctionalization to allow cell adhesion, which can be inaccessible and daunting to many laboratories.

2.4. Defining the Cell Microenvironment

While matrix stiffness has been regarded as a key metric for how mechanotransduction processes work to regulate cell behaviors, it is important to recognize that cell adhesion and ECM remodeling play important roles in mechanotransduction.

To that end, synthetic hydrogels are often rendered suitable as cell culture substrates through covalent coupling of ECM proteins or peptides that bind to integrin or other cell adhesive receptors. The RGD binding sequence found in fibronectin is often required for cells to spread and grow properly [55]. More fundamentally, such adhesive cues that mimic the native cell‐ECM contacts are essential for traction‐based mechanosensing [66]. Furthermore, cells can remodel their surrounding matrix through degradation and deposition of ECM components. In turn, this process can modulate adhesive interactions, alter matrix stiffness, and differentially affect cytoskeletal organization and downstream signaling pathways [53, 67]. Most of the naturally derived hydrogels are degraded by hydrolysis or enzyme‐mediated processes. On the other hand, degradability is a factor that must be engineered into synthetic and non‐mammalian‐derived materials. This is commonly achieved through the incorporation of peptide crosslinks that can be cleaved by matrix‐degrading enzymes, such as matrix metalloproteinases (MMPs) [64], in which the rate of degradation could further be controlled to assess their impact on cell mechanoresponses. For example, spatiotemporal stiffness patterning can be combined with on‐demand release of biochemical factors through different stages of a complex biologic process, such as pluripotent stem cell differentiation [68].

Therefore, an ideal hydrogel culture platform to model and manipulate the cell microenvironment should meet several key criteria, including 1) defined matrix composition that interacts with specific cell receptors and mechanotransduction pathways, 2) ability to independently control the various biochemical, structural, and mechanical cues, and 3) spatial and temporal programmability. Although the focus of this review is on stiffness gradient hydrogels, it is anticipated that future advances will extend beyond customizing the physical landscape, especially in view of growing interest in their utility for creating in vitro models and tissue regeneration [69, 70]. Leveraging the plethora of direct chemical modifications or post‐processing strategies such as plasma treatment, both natural and synthetic materials can be functionalized with user‐preferred bioactive molecules [71, 72, 73]. Thus, engineered hydrogels provide countless opportunities for mechanobiology exploration and translation.

2.5. Mechanical Characterization at Cell‐Relevant Scales

The ECM resists cell‐generated forces with its stiffness, and the level of resistance that cells experience in turn regulates cell phenotype in healthy and diseased tissues. As such, quantification of this mechanical property is important for developing gradient hydrogel platforms to model tissue spatial heterogeneities.

In order to study cell‐substrate interactions in a biologically relevant manner, it is important to characterize the substrate mechanics in a similar manner to how cells would probe their surrounding matrix in tissues. Since cells typically experience forces on the order of pico‐ to nano‐Newtons and in micrometer length scales (i.e., the size of a single cell), highly sensitive measurement techniques have been developed, including microsphere indentation, micro‐indenters, and atomic force microscopy (AFM) [74]. Among those, AFM has been the most widely employed to measure the cellular scale stiffness of tissue and hydrogel surfaces via local indentation using a probe with sizes ranging from <1 µm up to 50 µm [75]. Nano‐sized tips are best‐suited to probe the bending properties of individual polymer chains. On the other hand, micro‐sized probes capture the mechanical properties of a local network of polymer chains in a reasonably similar manner to how cells sense their surroundings via integrins. More advanced AFM systems are also equipped with confocal imaging capabilities, enabling additional correlation of stiffness measurements with localization of fluorescently labeled cells, mechanomarkers, or ECM components [76].

The major limitation of AFM stiffness testing of gradient hydrogels for mechanobiological studies is scale. Given the labor‐intensive and time‐consuming nature of AFM measurements, most of the studies have been optimized to capture a series of single‐point indentation at relatively large spatial intervals (typically 50 µm to 1000 µm) across a substrate to determine the presence and magnitude of stiffness gradients. Subsequent analysis of cellular stiffness responses is typically performed by averaging across discrete stiffness segments, each spanning hundreds of micrometers to several millimeters, often dictated by the gradient sample size. This approach facilitates a more straightforward data interpretation and works well for screening mechanical dose‐dependent cell behaviors. However, it does not fully capture how cells likely experience subcellular heterogeneities of native tissue stiffness, which have been reported in fibrosis and separately found to differentially affect stem cell fate [77, 78]. Nevertheless, it is important to recognize the unique proposition of gradient hydrogels in enabling high‐throughput analysis of cell mechanotransduction across the broad physiological stiffness landscape, whilst recognizing the necessity to better understand and harness the micro‐ to nano‐scale ECM properties toward translational mechanobiology [65, 79].

3. Methods to Pattern Microscale Stiffness Gradients in Hydrogels

The basic premise for biomimicry of stiffness gradient hydrogels is that they represent the stiffness range and gradient strength characteristic of the target tissue type or biological process. To that end, empirical stiffness measurements of ex vivo tissues are commonly performed to provide a physiological reference to guide the development of stiffness gradient hydrogels for in vitro cell studies [80, 81]. Broadly, the stiffness gradient strengths can range from ∼1 kPa mm⁻¹ in physiological tissues to ∼10 kPa mm⁻¹ in pathological tissues (e.g., fibrosis, tumors), and >100 kPa mm⁻¹ at tissue interfaces [82]. Of note, the drastically steeper stiffness gradients across regions of healthy and diseased tissues often arise from aberrant ECM remodeling, where imbalanced matrix turnover leads to excessive deposition of connective‐tissue‐like ECM, disorganized fibrous network, and even altered cell mechanics, which altogether drive pathological changes in tissue stiffness.

Over the past decades, numerous methods to fabricate gradient biomaterials for tissue engineering and regenerative medicine have been described in literature [69, 83, 84]. Each method has a unique set of tradeoffs between achievable stiffness ranges, gradient shape fidelity, ease of operation, and scalability. Here, special attention is given to the creation of microscale stiffness gradients in vitro relevant to mechanobiology applications, which can be classified into four overarching strategies: diffusion‐based, controlled polymerization, component redistribution, and micro‐molding. Representative examples are summarized in Table 2.

TABLE 2.

Representative examples of different methods for fabricating stiffness gradient hydrogels for mechanobiology applications. Unless stated otherwise, the stiffness range (E) is reported in terms of Young's modulus measured by AFM, whereas slope denotes the steepness of gradient.

Method	Material	Culture system	Gradient properties	Cell type	Major results	Ref.
Diffusion‐based
Sandwiching	PA	2D cell seeding	Continuous linear gradient at mixed interface of two drops E: ∼0.5‐60 kPa Slope: ∼5‐15 kPa mm−1	human telomerase‐immortalized fibroblasts	Incorporation of fluorescent microbeads in polymer solution allowed stiffness readout from bead density; growth of cell fibrillar adhesions is mechanosensitive in a tensin‐dependent manner	[16]
Microfluidic gradient generator	PA	2D cell seeding	Continuous linear gradient across planar substrate E: ∼1‐12 kPa Slope: ∼10 kPa mm−1	hMSCs	MSCs migrated from soft‐to‐stiff region (positive durotaxis); migration velocity is dependent on stiffness gradient slope	[82]
Sequential layering	GelMA	3D cell encapsulation	Discrete stiffness layers blended at interface E (compressive): ∼5‐80 kPa Slope: ∼15 kPa mm−1	NIH‐3T3 fibroblasts	Fibroblasts displayed differential morphology and migrated along gradient axis	[85]
Two‐step polymerization	PA	2D cell seeding	Continuous linear gradient across planar substrate (24x20x1 mm) E: ∼0.1‐159.2 kPa Slope: ∼0.5‐8.2 kPa mm−1	hADSCs; C2C12 mouse myofibroblasts	Stiffness gradient can be tuned to impose durotactic or non‐durotactic effect on cells; shallower gradients enable observation of dose‐dependent (instead of ON‐OFF response) of YAP nuclear translocation	[86]
Controlled polymerization
Gradient photomask	GelMA	3D cell encapsulation	Continuous linear gradient across planar substrate (12x12x1 mm) E: ∼5‐38 kPa Slope: ∼3.4 kPa mm−1	hADSCs	hADSCs encapsulated in soft regions exhibited increased cellular and nuclear volume, with increased expression of osteogenic markers; study suggests that volume is sufficient to direct 3D stem cell differentiation	[87]
Sliding photomask	PA	2D cell seeding	Continuous linear gradient across planar substrate E: ∼1‐240 kPa Slope: ∼7.5‐115 kPa mm−1	NIH‐3T3 fibroblasts; SY5Y neuroblastoma cells	Wide‐range stiffness gradient slopes achievable; less steep gradients with narrower stiffness ranges also possible with this method	[88]
Sliding photomask	PA	2D cell seeding	Continuous linear gradient embedded within a microfluidic device E: ∼1‐40 kPa Slope: ∼5 kPa mm−1	U87‐MG glioma cells	Substrate stiffness promoted cell chemotaxis, while epidermal growth factor gradient can accelerate cell migration	[89]
Repeated gradual freezing‐thawing	Polyvinyl alcohol/ Hyaluronic acid (PVA/HA)	2D cell seeding on cut sections of gel cylinder	Continuous gradient along height of gel column (25 mm diameter, 80 mm height) E (compressive): ∼20‐200 kPa Slope: ∼2.2 kPa mm−1	hMSCs (bone marrow)	Favorable stiffness range for hMSC differentiation: ∼20 kPa for nerve cell, ∼40 kPa for muscle cell, ∼80 kPa for chondrocyte, and ∼190 kPa for osteoblast	[90]
Temperature gradient during curing	PDMS	2D cell seeding	Continuous gradient across 12 mm distance on PMDS sheet Indentation modulus: ∼0.19‐3.1 MPa Slope: ∼242.5 kPa mm−1	rMSCs (bone marrow)	Osteogenic differentiation is influenced by stiffness, but also depends on the availability of appropriate adhesion molecules	[91]
Component redistribution
Temperature field‐driven	GelMA; Gellan gum	2D cell seeding	Linear gradient across planar substrate (1000 µm width, 120 µm depth) E: ∼5‐140 kPa Slope: ∼10‐100 kPa mm−1	3T3‐L1 fibroblasts	FITC labeling of hydrogel allowed stiffness readout from fluorescence intensity; fibroblasts displayed differential migration and morphological responses on GelMA vs. Gellan gum gradient gel with similar stiffness profile	[61]
Electric field‐driven	β‐sheet rich silk nanofibers (BSNF)	2D cell seeding on cut sections of gel cylinder	Continuous gradient along gel column (2 mm diameter, 10 mm height) E (compressive): 20‐130 kPa Slope: ∼11 kPa mm−1	rMSCs (bone)	BSNF can be incorporated into different hydrogel systems; tunable gradients of multiple bioactive cargos can also be achieved simultaneously	[92]
Micro‐molding
Hydrogel on topographically defined stiff supports	PA	2D cell seeding	Linear and nonlinear gradient pattern depending on topography of underlying support structure (e.g., step, beads, grooves)	CCL‐92 mouse embryonic fibroblasts	15 µm is the critical depth below which cells will respond to the underlying substrate; this critical thickness is independent of gel's bulk shear modulus	[93]

hADSC (human adipose‐derived stem cell); hMSC (human mesenchymal stem cell); rMSC (rat mesenchymal stem cell)

3.1. Diffusion‐Based

Diffusion is among the most common strategy to create microscale mechanical gradients in soft hydrogels. This approach covers a broad family of techniques that, at some stage of the fabrication process, harness the passive movement of hydrogel precursor molecules from regions of high to low concentration to generate spatially varying compositions.

The most rudimentary implementation involves polymerizing two adjacent drops of polyacrylamide solution with different concentrations of acrylamide and its crosslinker, whereby diffusive mixing at the boundaries results in a continuous linear gradient. Originally presented by Lo et al. in 2000 [94], this coverslip “sandwiching” technique is still widely employed today given its simplicity and cost‐effectiveness, but the resulting stiffness gradients can be extremely variable (Figure 2A) (Table 2) [16, 95].

FIGURE 2.

Representative examples of diffusion‐based strategies to create stiffness gradient hydrogels. (A) The simplest approach involves polymerizing two adjacent drops of different hydrogel precursor solution between two flat surfaces. Reproduced with permission [95]. Copyright 2022 Springer Nature. (B) Microfluidic mixing of different hydrogel precursor solutions can be used to create narrow, steep gradients. Adapted with permission [96]. Copyright 2004 American Chemical Society. (C) Two‐step polymerization of polyacrylamide (PA) hydrogels; the compositions of both layers could be either the same or varied to tune the final stiffness gradient properties. Reproduced with permission [86]. Copyright 2017 National Academic of Sciences.

To enable more precise patterning, researchers subsequently explored microfluidic gradient generators with photopolymerization process (Figure 2B) [96, 97]. Using this technique, a microfluidic network is designed to split and mix the precursor solutions in a controlled manner, generating stable streams of varying composition that combine at a single broad channel. By carefully tuning the system flow rates, the initial step profile observed at the inlet of the broad channel blurs into a smooth gradient by diffusion as the streams flow down the channel, which is then polymerized into stiffness gradient hydrogels via UV exposure. Although this method allows for the fabrication of well‐defined and relatively steep (∼10 kPa mm⁻¹) gradients, it is more restricted in the achievable stiffness range and can be cumbersome to operate, limiting its widespread adoption (Table 2) [82].

A more straightforward approach to create microscale stiffness gradients is by sequential layering of partially polymerized hydrogels in a predefined mold. In this method, each gel layer has a distinct stiffness (controlled by varying the polymer and/or crosslinker concentration), and a new layer is deposited before the previous layer has fully polymerized, allowing for diffusive mixing at the layer interface. For example, Ko et al. demonstrated this layer‐by‐layer stacking technique using photopolymerized hydrogels such as GelMA and PEG, showcasing their potential for 3D cell encapsulation applications (Table 2) [85]. Since the gradient gel construct is still essentially composed of discrete layers, the final gradient can be a stepwise profile, or the UV intensity can be tuned to adjust the degree of diffusion at the interface, resulting in continuous gradients. In another example, Hadden et al. employed a ramp‐shaped mold and a two‐step polymerization technique (Figure 2C) (Table 2) [86]. A first gel layer of varying thickness is first polymerized, on top of which a second gel layer is polymerized to form a 2D planar substrate. In this case, a stiffness gradient is created due to the differential diffusion of unreacted acrylamide and its crosslinker across the varying thickness of the first gel layer. An inherent limitation of this technique is that the resulting gradient slope is relatively shallow (∼0.5 to 8 kPa mm⁻¹), while the overall gel thickness (∼1 mm) may pose challenges for high‐magnification imaging. Nevertheless, these layering techniques are easy to implement without the need for clean‐room facilities, making it an attractive option for biological laboratories.

Overall, the simplicity of diffusion‐based methods enables faster and higher throughput gradient formation, making them ideally suited for large‐scale biological experiments. Despite their advantages, however, the process of passive molecular diffusion is inherently sensitive to external environmental factors such as temperature and humidity, which are major sources of batch‐to‐batch variability. Because each step of the fabrication process is highly time‐dependent, even a slight deviation could also result in inconsistencies in the stiffness gradient outcomes. As a potential solution for addressing this issue, researchers have reported strategies to incorporate fluorescent probes into the hydrogel mix, which is designed to display a fluorescence signal proportional to the hydrogel stiffness [16, 95]. Not only is this useful for quickly identifying erroneous gradient formation, it also allows researchers to more precisely orient the gradient for cell analysis.

3.2. Controlled Polymerization

Besides spatial modulation of the hydrogel precursor concentration, another main route to creating substrates with mechanical or architectural gradients relies on spatially controlling the degree of crosslinking during the crosslinking process.

Graded light exposure in combination with photopolymerizable hydrogels (e.g., polyacrylamide, GelMA) has been widely explored to generate microscale stiffness gradients in mechanobiology studies. This is achieved either by using photomasks patterned with varying degrees of transparency [87, 98, 99] or sliding an opaque photomask across the polymer solution at a controlled speed (Table 2) [88]. The sliding mask technique is one of the most flexible methods, allowing the creation of shallow (1–10 kPa mm⁻¹) to very steep (>200 kPa mm⁻¹) stiffness gradients, though its widespread adoption remains limited by the need for specialized equipment, including a stepper motor. In an interesting example, Dou et al. recently used the sliding mask technique to create polyacrylamide hydrogels with combined stiffness gradient and growth factor gradient in the lateral direction to probe glioma cell migration (Figure 3A) (Table 2) [89]. Compared to sliding photomasks, masks with patterned gray‐intensity allow for a simpler fabrication process and a more well‐defined stiffness profile (Figure 3B) [98]. However, the resolution of conventional photomask printing techniques limits these gels to a softer substrate range of ∼1–20 kPa and shallow gradient strength of ∼1 kPa mm⁻¹ in practice [88].

FIGURE 3.

Representative examples of creating stiffness gradient hydrogels through controlled crosslinking. (A) Polyacrylamide substrates with combined stiffness and chemical (epidermal growth factor, EGF) gradients; the stiffness gradient was formed using a sliding photomask technique whereas the chemical gradient was generated via microfluidics. Reproduced with permission [89]. Copyright 2020 American Chemical Society. (B) Gradient hydrogel fabricated using a patterned gray‐intensity photomask for polymerization of GelMA. Reproduced with permission, under terms of the CC BY‐NC 4.0 license [98]. Copyright 2022 The Author(s), published by Wiley Periodicals LLC. (C) Repeated free‐thawing method to introduce transitions in crystallinity of polyvinyl alcohol/hyaluronic acid (PVA/HA) hydrogels. Reproduced with permission [90]. Copyright 2016 Elsevier.

Alternatively, temperature control‐based techniques have also been explored, including the application of a gradient temperature during the curing process of PDMS [91], and the repeated gradual freeze‐thawing method (Table 2) [90]. Of note, stiffness gradients can be created in PDMS since the curing temperature directly affects the degree of PDMS crosslinking. On the other hand, freeze‐thawing method works by creating ice fronts that travel through the polymer solution to induce the formation of localized polymer crystallites that, in turn, facilitate network crosslinking among the polyvinyl alcohol (PVA) chains upon thawing. This is possible due to the unique hydrogelation mechanism of PVA. Oh et al. demonstrated this technique using a PVA/HA composite to generate a wide stiffness gradient range of 20–200 kPa (Figure 3C) [90]. Gradual freezing of the PVA/HA solution within a cylindrical mold was accomplished via direct contact with liquid nitrogen at the bottom, allowing the freeze front to progress upward. The stiffness gradient properties can be tuned by controlling the freezing temperature, freezing time, and the number of freeze‐thawing cycles, thus offering a useful hydrogel patterning approach that does not require any specific laboratory equipment.

3.3. Component Redistribution

Recently, there has also been a growing number of reports that create stiffness gradients by leveraging an external force to guide the redistribution of solutes within an initially homogenous precursor solution. Accordingly, regions of higher polymer concentration result in higher local stiffness in the final hydrogel, and vice versa. By decoupling the gradient generation mechanism from the hydrogel crosslinking process, this family of techniques offer the promise of greater control and reproducibility compared to the previous methods (Table 2) [100].

One of the earliest concepts exploited capillary flow in microfluidics to develop a compositional gradient, which can be preserved through subsequent photopolymerization [101]. He et al. demonstrated this technique using a straight, reversibly bonded PDMS microchannel, which was first pre‐filled with a low concentration precursor solution, then a second high concentration solution was introduced at the outlet [101]. The second solution spontaneously flowed into the microchannel, and the desired gradient was developed through a combined effect of evaporation‐induced backflow and molecular diffusion. The study employed fully non‐mechanical means of driving flow within the microfluidic channel, which although greatly lowered the barrier of entry to microfluidic‐based techniques, are subject to a high degree of inconsistency.

A more recent development has been to take advantage of the physical phenomenon of thermophoresis to enable a flexible yet reliable control over the gradient pattern and type of hydrogel (Table 2) [56, 61, 102]. In contrast to the gradual freeze‐thawing method which relies on gradient temperatures to induce spatial variations in crosslinking density, the thermophoresis technique uses a controlled temperature field to actively direct the movement of precursor components in solution before gelation. In principle, thermophoresis occurs in any aqueous system when subjected to a non‐uniform temperature field, making this technique highly versatile across different crosslinking chemistries. First reported by Vigolo et al. in 2019, stiffness gradient alginate hydrogels were fabricated in a two‐layer microfluidic device designed for on‐chip temperature patterning (concentration gradient generation step) and slow release of calcium ions to facilitate eventual hydrogel crosslinking, demonstrating steep (10‐100 kPa mm⁻¹) gradients over 600 µm [102]. Subsequently, the same group further developed this technique to create a modular platform that can be easily configured to provide both linear and complex spatial gradients (Figure 4A) [100]. Notably, the platform was designed to be readily adaptable for photocrosslinked and thermosensitive hydrogels, as demonstrated in the study using GelMA and Gellan gum. Such streamlined platforms can offer great advantages for rapid design‐to‐application iteration, especially in allowing researchers to expand the library of gradient materials beyond classical elastic materials for investigating cell responses to more complex substrate cues [61].

FIGURE 4.

Representative examples of external force‐driven component redistribution approach for stiffness gradient hydrogel fabrication. (A) Temperature field‐driven approach using a microfluidic platform with configurable heating and cooling channels which allows customization of the stiffness gradient profile; thermophoresis effect causes the hydrogel precursor molecules to migrate toward colder regions and form a concentration gradient within the solution, which translates into a stiffness gradient upon crosslinking. Reproduced under terms of the CC BY 4.0 license [100]. Copyright 2025 The Author(s), published by Wiley‐VCH GmbH. (B) Electric field‐driven migration of negatively charge β‐sheet rich silk nanofibers (BSNF) in amorphous silk nanofiber (ASNF) hydrogel solution; the BSNF blocks can be functionalized to create bioactive mechanical gradients for tissue engineering applications. Reproduced under terms of the CC BY license [92]. Copyright 2020 The Author(s), published by Oxford University Press.

Other emerging methods have sought to functionalize the biomaterial components, which can then be actively manipulated using externally applied forces via magnetic or electric fields, for example [92, 103]. In 2020, an interesting approach by Xu et al. reported the simultaneous formation of compositional and mechanical gradients through electric‐field‐driven protein chain migration (Figure 4B) (Table 2) [92]. This method relies on the progressive movement of negatively charged 𝛽‐sheet rich silk nanofibers (BSNF) which serve to fine‐tune the mechanical gradients in a variety of base polymers (e.g. GelMA, N‐isopropylacrylamide, amorphous silk nanofiber solution (ASNF)). Interestingly, the possibility to incorporate bioactive cargos such as growth factors and gold nanoparticles into the BSNF blocks can open potential new avenues to simultaneously define the biophysical and biochemical cues at micron‐scale resolution, enhancing our capabilities to study cell mechanobiology behavior in a more physiologically relevant and dynamic context.

3.4. Micro‐Molding

Micro‐molding is an alternative variation to the previously described hydrogel layering methods. This approach is based on creating abrupt changes in the gel thickness using microfabricated molds. It works on the principle that the extent to which an underlying (stiff) substrate is felt through a second layer of (compliant) material will vary as a function of the depth of the superficial compliant layer [93, 104]. This method is particularly useful to generate step stiffness gradients of up to 300 kPa mm⁻¹. For example, Choi et al. demonstrated the use of micro‐patterned SU‐8 photoresist molds to generate alternating soft and stiff stripes in polyacrylamide hydrogels, which can be used to simulate the striations in muscle tissues [105]. Furthermore, the base substrate could also be patterned with different topographies to generate more complex gradient patterns, first proposed by Kuo et al. using a dispersed array of polystyrene microbeads which resulted in radial stiffness gradients (Figure 5) (Table 2) [93]. A main advantage of this approach is that gels with spatially varying stiffness but uniform pore size can be created. However, the differential hydrogel swelling may cause a considerable variation in the overall substrate thickness along the gradient, raising the question of whether cells are responding to stiffness or topographical cues. Moreover, it is important to note that when the gel thickness is <20 µm, cells can sense the underlying substrate [93]. This makes it difficult to evaluate the effective matrix stiffness that cells are sensing, which can explain why ultra‐thin hydrogels are generally avoided in stiffness gradient studies.

FIGURE 5.

Representative examples of micro‐molding method to create mechanically patterned polyacrylamide (PAA) hydrogels. The apparent hydrogel stiffness varies as a function of the gel thickness (H). Adapted with permission [93]. Copyright 2012 John Wiley and Sons.

4. Applications of Stiffness Gradient Hydrogels in Mechanobiology

The range of applications of stiffness gradient hydrogels has increased over the past two decades, driven not only by advances in precision engineering of microscale material properties, but also mirrors the rapid conceptual shifts in our understanding of the active mechanical interactions between cells and their ECM. Applications have matured from using a simple 2D linear gradient for single‐cell analysis through to the deployment of 3D cell cultures and simultaneous control of multiple spatiotemporal cues. Based on this trend, this section highlights representative examples of stiffness gradient hydrogels developed for the study of cellular processes at the fundamental level, and how the gradient features can be harnessed in disease modeling and tissue engineering.

4.1. Studying In Vitro Durotaxis

Controlled cell migration underlies many biological processes in life. Whether during development, disease progression, or in response to tissue injury, cells within the body must be able to migrate at coordinated speeds and trajectories. The phenomenon that cells can directionally migrate along spatial stiffness gradients (Figure 6A), broadly referred to as durotaxis, was first discovered in 2000 by Lo et al. with the observation that single‐cell fibroblasts tend to move from a softer to stiffer region on ECM‐coated polyacrylamide gels [94]. The ability of cells to modulate their actomyosin contractility depending on the matrix stiffness was postulated to underlie this phenomenon of directed cell migration. Since then, this finding has also been shown in various other cell types, such as vascular smooth muscle cells [106], mesenchymal stem cells [82, 107], immune cells [108] and cancer cells [95, 109].

FIGURE 6.

Application of gradient hydrogels for in vitro durotaxis studies. (A) Cells can exhibit different modes of durotaxis in response to stiffness gradients. Reproduced with permission [116]. Copyright 2024 The Company of Biologists Ltd. (B) U‐215MG glioblastoma cells on polyacrylamide gels were shown to migrate toward regions of intermediate stiffness, corresponding to where cells would exhibit maximal traction at 5‐10 kPa range. Reproduced with permission [95]. Copyright 2022 Springer Nature. (C) A431 carcinoma cell clusters exhibit positive durotaxis on E‐cadherin‐coated polyacrylamide gels, with maximal velocity at the intermediate stiffness region. Reproduced with permission [119]. Copyright 2022 Springer Nature.

Stiffness gradient hydrogels have already proven to be instrumental in shaping our fundamental understanding of durotaxis, especially given the technical difficulties associated with investigating this process in vivo [18]. Notably, majority of the in vitro studies to date reported a preferred migration toward stiffer environments (i.e., positive durotaxis) through coordinated cell‐ECM adhesion, contractility and cytoskeletal signaling, representing the long‐standing dogma of stiffness‐guided directional cell migration. At the molecular level, integrin clustering activates focal adhesion kinase (FAK) and triggers the formation of larger, more stable focal adhesion protein complexes, generating a front‐rear traction asymmetry that facilitates preferential migration from soft to stiffer environments [110]. This migratory response is typically context dependent, with key influencing parameters including the steepness of gradient [82, 111, 112], absolute stiffness of the gradient [109, 113], and ECM composition [114]. Durotaxis may also be cell‐type specific. For example, many human cancer cell lines can migrate up stiffness gradient ranging from 2 to 7 kPa but exhibit limited responsiveness on the stiffer end from 7 to 18 kPa [109]. On the other hand, the durotactic speed of mesenchymal stem cells correlated positively with the gradient strength [82], whereas Schwann cells were found to migrate equally well on both shallow and steep stiffness gradients [115].

While initial studies primarily sought to confirm the phenomenon of durotaxis and characterize the stiffness responsiveness across different cell types, more recent work has placed emphasis on mechanistic depth and physiological relevance, revealing that this cellular process is far more nuanced than previously thought. Among the emerging concepts, elegant in vitro studies have shown the ability of cells to directionally migrate from stiffer to softer environments (i.e., negative durotaxis), or be insensitive to stiffness gradients (i.e., adurotaxis) [116]. In an interesting example, U251‐MG glioblastoma cells were observed to migrate toward an intermediate substrate stiffness (∼5 – 10 kPa), and the switch between positive and negative durotaxis appear to depend on the optimal balance between the intracellular cytoskeletal tension and extracellular stiffness sensing via focal adhesions to achieve maximal traction force (Figure 6B) [95]. This biphasic behavior is consistent with established molecular clutch models in which actomyosin “motors” pull against integrin‐ECM “clutches”; the generated traction forces increase with stiffness up to an optimal point, beyond which excessive clutch slippage or over‐mature adhesions reduce effective traction and favoring migration toward softer regions [117]. It turns out that that choice of durotaxis phenotype could also be predetermined in specific tumor subtypes [95, 109], highlighting the need for a better understanding of these diverse scenarios to elucidate how mechanical gradients promote cancer cell migration from the stiff tumor to the softer surrounding stroma during cancer progression. Furthermore, evidence is emerging that points to collective durotaxis as a new mode of directed migration, wherein multicellular clusters exhibit more efficient durotaxis than their individual constituent cells (Figure 6C) [113, 118, 119]. The current consensus is that the larger surface area covered by a group of cells and force‐transmission via cell‐cell junctions can enable an increased mechanosensitivity of clusters to gradients, simply by virtue of a greater difference in extracellular stiffness experienced between their front and rear ends. Of note, α‐catenin‐containing cell‐cell adhesions has been postulated to be the key molecular machinery for collective durotaxis, but can be independent of cell‐ECM coupling [119]. In vivo, it is believed that this phenomenon could guide tissue organization and adaptability in processes such as epithelial migration during development [120], with ongoing speculations on its contribution to cancer metastasis and wound healing [121]. Meanwhile, from a biophysical viewpoint, researchers have garnered interest in exploring additional cellular structures beyond focal adhesions and actin cytoskeleton that regulate mechanically directed migration, such as nuclear repositioning [122] and focal‐adhesion independent durotaxis [123].

Studying durotaxis in vitro requires the analysis of cell behavior on a wide range of stiffness gradients slopes (<1 kPa mm⁻¹ to ∼400 kPa mm⁻¹), but achieving this in a way that is precise, reproducible, and tunable according to experimental needs remains a persistent challenge [19]. Only until recently, studies have largely been based on examining cells at the soft‐to‐stiff transition of a “step” gradient, many of which were fabricated using the rudimentary method of polymerizing two adjacent drops of soft and stiff polyacrylamide solutions [94]. Nowadays, photopolymerization techniques are preferred for better control and flexibility. For example, Yeoman et al. used a high‐resolution chrome patterned photomask to create a step‐gradient polyacrylamide gel with alternating 200 µm soft (0.3–10 kPa) and 100 µm stiff (1.5–30 kPa) stripes, which was applied to probe single cancer cell migration in 2D [124]. In another approach, Pallarès et al. employed the sliding mask method to fabricate 2D linear gradient polyacrylamide hydrogels covering stiffnesses from ∼0.1 kPa up to ∼22 kPa (shallow gradient) or ∼135 kPa (steep gradient) over 3 mm, enabling systematic investigations of long‐range continuous gradients [119].

Despite the wealth of 2D studies, however, durotaxis in 3D context has remained disproportionately underexplored, primarily due to the lack of suitable biofabrication strategies that can modulate the stiffness of a matrix without changing other properties (e.g., pore size, ligand density) [125, 126]. Yet, considering that most tissues provide a 3D environment where cells may encounter additional directional cues such as spatial confinement, the shift from 2D to 3D is likely to be a next major frontier in this research area. Nevertheless, continued development both 2D and 3D stiffness gradient hydrogel systems will be essential to complement in vivo models toward a more holistic understanding of the role of durotaxis in pathophysiological conditions, and how to modulate this process as a potential therapeutic strategy for diseases like cancer and fibrosis [127].

4.2. Reductionist Screening of Cell‐Material Interactions

Evaluating the mechanical responsiveness of cells in a reductionist, yet physiologically relevant manner represents an active pursuit in mechanobiology research to which gradient hydrogel technology is well‐poised to make unique contributions. Reductionist analogs of the native ECM are deliberately simplistic, intended to help researchers parse out the effects of specific ECM properties on cell behaviors. Traditionally, such investigations are performed by comparing cells cultured on separate hydrogels with multiple stiffness levels in parallel, or by transferring the cells between different substrates to simulate temporally dynamic microenvironments. In contrast, because gradient hydrogels present a user‐defined continuum of mechanical cues, each sample replicate effectively captures a series of stiffness‐response relationships. A significant advantage of this screening approach is the opportunity to uncover previously unappreciated insights that might be missed from testing discrete values, such as mechanosensitivity thresholds or non‐linear effects (Figure 7A) [16, 17]. Another important advantage gained with the use of gradient hydrogels over discrete samples is the reduced fabrication and biological variability, making them an attractive solution for scaling complex biological experiments.

FIGURE 7.

Application of gradient hydrogels as reductionist platform for screening cell‐matrix interactions. (A) Correlation analysis between matrix stiffness and α5β1‐integrin of fibroblasts cultured on polyacrylamide stiffness gradient hydrogels revealed a non‐linear mechanoresponse of fibrillar adhesions. Reproduced under terms of the CC BY 4.0 license [16]. Copyright 2020 The Author(s), published by The Company of Biologists Ltd. (B) Combined biophysical (stiffness) and biochemical (fibronectin density) gradients allowed identification of optimal conditions for stem cell lineage commitment; adipogenic differentiation (red stain) was promoted across the soft regions, whereas osteogenic differentiation (dark purple stain) was strongest in areas of high stiffness and fibronectin density. Reproduced under terms of the CC BY license [132]. Copyright 2015 The Author(s), published by Springer Nature. (C) GelAGE hydrogel crosslinking density is phototunable enabling a uniform soft gel to be formed with short and homogenous light exposure, then on‐demand stiffness gradient patterning can be induced by a secondary exposure through graded photomask; MCF7 breast cancer spheroids exhibited stiffness‐mediated size and β1‐integrin expression when encapsulated directly in a gradient environment but showed no difference when the spheroids were only exposed to a stiffness gradient after 5 days, highlighting the role of dimensionality in 3D mechanosensation. Reproduced under terms of the CC BY 4.0 license [138]. Copyright 2023 The Author(s), published by Wiley‐VCH GmbH.

Microengineered stiffness gradient hydrogels have been especially useful in identifying conditions that promote specific lineage commitment programs and distinct cellular behaviors. In this context, the bioinertness of synthetic hydrogels is favorable in order to decouple the biochemical complexity of natural ECM, whilst still providing a facile means to define the ECM protein composition. In an early example using PDMS stiffness gradients coated with fibronectin or gelatin, Wang et al. found that rat mesenchymal stem cells exhibit a stiffness‐dependent morphological and osteogenesis behavior that is also influenced by the ECM protein type [91]. This highlighted the impact of ECM composition on cell stiffness mechanosensing, which was consistent with findings in other studies [114, 128, 129]. Indeed, much of our foundational understanding of mechanobiology was developed using adult stem cells, establishing the classical mechanisms of integrin‐mediated biophysical cue sensing and intracellular signaling. Over the years, efforts in this direction has also broadened the scope of investigation into the role of substrate stiffness on cellular signaling and behavior in several different cell types, such as fibroblasts [16, 53, 100], neural cells [130], cardiac cells [131], and cancer cells [132, 133]. In an interesting example of nonlinear mechanosensation (Figure 7A), Barber‐Pérez et al. showed that fibroblasts form longer fibrillar adhesions with increasing substrate stiffness, particularly in the low‐to‐intermediate range (0.5–7 kPa), with slower growth and eventual plateau at higher stiffness (7–22 kPa) when cultured on polyacrylamide gradient gels [16]. Importantly, this mechanosensitive adhesion elongation was found to be dependent on tensin‐1 (a focal adhesion protein), which stabilizes active α5β1‐integrin at fibrillar adhesions and traction force transmission in fibroblasts to regulate fibronectin fibrillogenesis during ECM remodeling. In another excellent study, Chin et al. used stiffness gradient hydrogels to comprehensively map cardiomyocyte mechanosensation under the simulated scenarios of necrotic, healthy, and infarcted tissue within a single platform. While the observed trends in cell morphology (cell size, cell shape, nuclear size) and YAP mechanomarker validated a classical stiffness‐dependent cellular response, these trends were maintained under disrupted actin cytoskeleton function, and therefore suggest that mechanisms other than actomyosin contractility may underpin cardiac cell mechanophenotype [131].

It is worth noting that gradient hydrogels for screening cell‐material interactions are typically >1 cm wide, rendering the capability for high‐content analysis [86]. More recently, researchers have extended this approach to the development of orthogonal spatial gradients that can enable high‐throughput, combinatorial screening of biophysical and biochemical signals within a single experimental platform. The work of Rape et al. is a good example that leverages hyaluronic acid hydrogel modified with chemical handles that can be manipulated using distinct wavelengths of light to achieve dually patterned matrix stiffness and fibronectin density (Figure 7B) [132]. The platform was successfully used to reveal non‐linear dependencies of both properties on mesenchymal stem cell differentiation and oncogenic microRNA expression in glioma cells, showcasing a significant advance beyond iterative testing of individual combinations. In addition to matrix‐bound ligands, new workflows continue to emerge from different laboratories describing the orthogonal modulation of surface wettability [134], environmental pH [133], and chemical concentration [89, 135], which are often facilitated by the integration with microfluidic platforms.

When designing reductionist gradient hydrogels for these studies, one should take into consideration the mechanobiological effects of dimensionality. Increasingly, experimental evidence has supported the notion that whether cells are cultured on 2D surfaces or within 3D matrices could alter cellular responses toward the same stiffness cues. Of note, studies using 2D stiffness gradient hydrogels have shown that both cell and nuclear areas increase with substrate stiffness [80, 86]. However, this trend was not followed in 3D gradient hydrogel cultures, instead, cell volume was observed to decrease with stiffness whereas nuclear volume remains the same [87]. Another study using encapsulated spheroids further reported that cells in 3D environments exhibit expression of canonical mechanomarkers (e.g., YAP, MRTF‐A, Lamin‐A) that better correlate with cellular volume rather than substrate stiffness, indicating volume adaptation‐based mechanotransduction [136]. Accordingly, this relates to the cell's capacity to adjust its internal volume to the surrounding matrix, and in 3D contexts, will depend on the degradation and remodeling of the surrounding ECM [53, 137]. Thus, despite the popular use of polyacrylamide and PDMS substrates for 2D mechanobiology studies, they are unsuitable for cell encapsulation due to inherent cytotoxicity and restricted capacity for cell‐mediated matrix remodeling [86, 91]. To overcome this challenge, researchers have turned to alternative hydrogel types, such as GelMA and decellularized ECM [53, 98, 138]. In a representative study, Vahala et al. demonstrated the culture of MCF‐7 breast cancer spheroids within homogenous allylated gelatin (GelAGE) hydrogels for 5 days before exposing them through a graded UV photomask to induce a linear stiffness gradient, simulating both dynamic spatial and temporal changes encountered in tumor microenvironments (Figure 7C) [138]. In the context of cancer mechanobiology, this platform was successfully used to reveal how the expression of Transient receptor potential cation channel subfamily V member 4 (TRPV4), integrin β1, F‐actin, and E‐cadherin was regulated by the spheroid volume expansion rather than the matrix stiffness cue itself. The fact that 2D mechanomarkers, including YAP and Lamin A/C, did not show a clear trend in 3D environments further prompts the need for engineered biomaterials to understand the role of dimensionality. Indeed, many of the recent gradient systems are based on photocrosslinking chemistries, which are expected to be a key enabler in future developments toward a better spatiotemporal understanding of mechanobiology [58].

4.3. Disease Modeling

In vitro models can serve as a tool for understanding disease mechanisms and preclinical assessments of drug efficacy. Driven by this trend, increasing efforts are being made to harness the potential of stiffness gradient hydrogels to overcome the intrinsic material limitations of tissue culture plastic (∼3 GPa) or glass (∼70 GPa) substrates for modeling biological tissue stiffnesses (∼0.1 kPa to 1 MPa). It should be noted that stiffness gradients are not necessarily needed to confer in vivo relevance, but as discussed in this section, could be useful as an enabling research platform to better recapitulate the complexity of biological systems.

Intuitively, the previously described reductionist gradient systems can also be used to study the mechanobiological pathways in specific pathological contexts. A relevant example is the case in chronic obstructive pulmonary disease (COPD), whereby characteristics of abnormal ECM deposition and local stiffness changes have been well‐documented, but a question remains whether the resident airway smooth muscle (ASM) cells are more mechanosensitive to extracellular stiffness or ECM protein types [139, 140]. On the basis of this question, Tan et al. used a 2D linear stiffness gradient polyacrylamide system (Young's modulus ∼2–30 kPa; gradient slope ∼4 kPa mm⁻¹) that mirrored empirical measurements of ex vivo ASM tissue, and focused specifically on understanding the effect of collagen I, fibronectin, and laminin on cellular stiffness responses (Figure 8A) [80]. Human ASM cells’ mechanosensitivity was demonstrated through a strong correlation between matrix stiffness and cellular morphology and contractile marker alpha‐smooth actin (αSMA) expression, though mechanistically YAP and Lamin‐A nuclear translocalization only exhibited stiffness‐mediated behavior on substrates coated with collagen I and fibronectin, highlighting the complex interplay between substrate stiffness and ECM protein type. A similar approach has been used by Zhang et al. to model the naturally graded stiffness across the periodontal ligament‐alveolar bone (PDL‐AB) enthesis (Figure 8B) [30]. The study employed microfabricated gradient hydrogels of different slopes (50 kPa mm⁻¹ vs. 250 kPa mm⁻¹) to mimic the altered mechanical environment in periodontitis disease compared to healthy state. With cellular‐scale feature resolutions, the study provided new insights into the molecular mechanisms underlying mesenchymal stem cell immunoregulation relevant to PDL‐AB enthesis. The shallower stiffness gradient was found to minimize cell polarization in guiding an anti‐inflammatory phenotype, likely driven by the reduced polarization of integrin β1 clusters and myosin IIB, leading to Histone H4 acetylation at lysine 16 (H4K16ac) localization‐dependent chromatin opening that facilitates Toll like receptor 3 (TLR3) gene expression. Overall, this study highlighted the value of gradient hydrogels to enable in vitro studies of soft‐hard tissue interfaces where the stiffness gradient typically spans from 50 to 500 µm [29, 56, 141].

FIGURE 8.

Application of gradient hydrogels for modeling diseases. (A) Linear stiffness gradient polyacrylamide hydrogels functionalized with ECM proteins to recapitulate airway smooth muscle (ASM) tissue, used to study the interplay between stiffness and ECM protein type on human airway smooth muscle cell (hASMC) mechanosensation; nuc/cyto YAP exhibited stiffness‐dependent behavior only on collagen I and laminin coated gels. Reproduced under terms of the CC BY 4.0 license [80]. Copyright 2024 The Author(s), published by Wiley‐VCH GmbH. (B) Periodic stiffness gradient polyacrylamide hydrogels fabricated with high or low stiffness gradient strength (SGS) to mimic the periodontal ligament (PDL)‐alveolar bone (AB) enthesis in health and periodontitis; low SGS was found to minimize cell polarization and influence chromatin organization, which regulate immunophenotype development of mesenchymal stem cells. Adapted with permission [30]. Copyright 2023 John Wiley and Sons. (C) Radial stiffness gradient created in allyl‐modified gelatin (GelAGE) hydrogels with varying stiffness gradient profiles; encapsulated MCF‐7 breast cancer cells exhibited an epithelial‐to‐metastatic phenotypic switch in response to both the local stiffness and gradient strength. Reproduced under terms of the CC BY 4.0 license [143]. Copyright 2024 The Author(s), published by IOP Publishing Ltd.

Moving beyond purely static and planar stiffness gradients, the next decade is likely to push these systems toward greater spatiotemporal complexity of biological processes. For example, Cao et al. recently proposed a gradient platform based on magnetorheological elastomers for dynamic and reversible tuning of stiffness gradients [142]. The platform was applied to study the combined effects of antifibrotic drug dosing and stiffness changes on cardiac fibroblasts under post‐myocardial infarction conditions, demonstrating its preclinical relevance for drug screening. To further enable modeling of the structure‐to‐function relationships in native tissues, Soliman et al. exploited droplet microfluidics to create radial stiffness gradients in shape‐controlled microgels (Figure 8C) [143]. A breast cancer model in which MCF‐7 cancer cells were encapsulated within the microgels with either “low” or “high” gradient steepness revealed a stiffness gradient‐induced phenotypic plasticity in cell aggregation. Interestingly, despite having similar stiffness values at the core region of both test groups, a shift in cell‐cel adhesion protein expression from E‐cadherin (indicating epithelial phenotype) to N‐cadherin (indicating breast cancer phenotype) was only observed with the “high” stiffness gradient, implying a contributing role of the steepness of the stiffness gradient in cancer metastasis.

Another opportunity for innovation lies in the integration of gradient hydrogels into microphysiological systems like organ‐on‐chips (OoCs). OoCs are microfabricated cell culture devices that are engineered to recapitulate human organ‐level physiology and cellular microenvironments [144]. These systems are especially attractive as a bridging technology for modeling diseases and drug responses, because they can better mimic the native biological complexities compared to static 2D and 3D cell cultures, yet are in many ways more accessible than direct investigations in animals or humans. The most relevant approaches for probing aspects of cell‐ECM interactions have been developed using PDMS or hydrogels (e.g., GelMA, polyacrylamide) as device construction materials whereby the substrate stiffness can be precisely modulated [145, 146, 147], and often surface treated with ECM protein coatings to promote cell attachment [72, 148]. In the field of cardiovascular research, there are already some demonstrations where microfluidic OoC systems have been able to elucidate the role of mechanics in the manifestation of clinically relevant disease phenotypes. For example, a novel in situ photopolymerization protocol used to create flexible poly(ethylene glycol) diacrylate (PEGDA) valve leaflets in a vein‐on‐a‐chip model allowed Baksamawi et al. to capture the dependance of leaflet stiffness on platelet accumulation in the context of venous thrombosis [149]. Another example fabricated a single‐channel OoC system with PDMS wall stiffness of 40 and 200 kPa representing young and aged arteries, respectively, which was successfully used to uncover a new role of Piezo1 mechanosensitive ion channel in endothelial responses to vascular stiffening under hemodynamic shear [150]. The expression of Piezo 1 was crucial in positively regulating endothelial cell morphology and inflammatory responses (assessed through ICAM‐1 expression and THP‐1 cell adhesion) to both atheroprotective (10 dyn cm⁻²) and atherogenic (2 dyn cm⁻²) shear stress regardless of substrate stiffness. Through γ‐H2AX labeling of DNA double‐strand breaks and β‐galactosidase activity, the authors further showed that Piezo 1 negatively regulates endothelial senescence under a combined soft substrate and atheroprotective shear level. Altogether, the study showed that Piezo 1's contribution to endothelial senescence is determined by the synergistic combination of matrix stiffness and hemodynamic shear stress, while its role in regulating inflammation is predominantly affected by shear stress.

Higher fidelity models are created by embedding hydrogel matrices into the OoC devices, either to present tissue‐specific cues or to provide a 3D microenvironment required to simulate the biological functionality in certain tissue types [151, 152]. So far, these devices are generally fabricated with uniform hydrogel composition for simplicity, but do not capture the heterogeneity at natural tissue transition zones [70]. This issue can be addressed with recent progress made in micromanufacturing and biofabrication techniques, enabling an emerging wave of development in engineering spatially patterned biomaterials within OoC platforms [153, 154, 155]. In a related example, Haack et al. have demonstrated the feasibility of leveraging capillary pinning as a microfluidic flow control technique to pattern multi‐region tissue constructs with spatially controlled hydrogel and cell composition [156]. Using models of fibrotic heart tissue border and bone‐ligament interface as exemplars, they were able to show remarkably different tissue contraction behavior compared to that observed in single‐region constructs. Other promising spatial patterning techniques include photomask patterning [157], extrusion‐based bioprinting [158], and digital light processing [159], presenting countless opportunities for establishing new workflows to translate biophysical signaling pathways into engineered pathophysiological models. Conversely, efforts in further development of OoC are expected to make critical contributions toward our understanding of fundamental mechanobiology. With future OoC design and optimization based on how well the model recapitulates the target biological functionality instead of a specific hypothesis of how cells interact with their matrix properties, this may push researchers into previously unexplored regions of the hydrogel design space, to discover new means of cellular mechanosensing and mechanotransduction in complex processes.

4.4. Tissue Engineering and Regenerative Medicine

As the last decade has included major advances in our ability to design hydrogels for spatiotemporal modulation of cell responses in vitro, the next decade is likely to drive these systems toward therapeutic cell‐based applications. Accordingly, the concept of programming cell fate and behaviors via mechanics of the scaffold materials for cell delivery is gaining momentum [160]. Given the conserved functional role of gradients in tissue development and homeostasis, the development of multiphasic hydrogel scaffolds is especially attractive, a major upcoming trend being continuous gradient tissue engineering.

Here, it is worth reiterating that while gradient hydrogels used in mechanobiology research are generally simpler with the aim of understanding how cells respond to isolated mechanical cues in a dose‐dependent manner, the overarching goal of tissue engineering and regenerative medicine is to build in sufficient physiological complexities that can enable therapeutically or technologically useful cellular programs. Nevertheless, this section highlights the growing synergy between mechanobiology and tissue regeneration.

Engineered gradient hydrogels that can more accurately mimic the native tissue architecture can be useful in many ways to study tissue regeneration process in vitro or facilitate in vivo tissue restoration. In the domain of osteochondral regeneration, for instance, gradient hydrogels are increasingly favored over traditional approaches based on joining multiple layers of polymeric scaffolds to mimic the soft‐to‐stiff tissue interface [26]. In addition to addressing delamination issues inherent in the layering technique, gradient hydrogels provide a more accurate representation of the continuous variation at the interface, which is critical to support proper tissue load‐bearing function. Among the commonly employed material systems include alginate, collagen, Gellan gum, GelMA, polycaprolactone (PCL), poly(lactic‐co‐glycolic acid) (PLGA), poly(lactic‐acid) (PLA) [161, 162, 163, 164]. Various strategies have been proposed to create biophysical gradients in osteochondral scaffolds, but are generally based on two main routes: i) direct modulation of hydrogel stiffness, or ii) incorporation of inorganic nanomaterials in a graded concentration. For example, Xiao et al. fabricated silk fibroin scaffolds by leveraging the directional movement of β‐sheet rich silk nanofibers (BSNF) in response to an electrical field, whereby higher BSNF content corresponded to higher hydrogel stiffness that enhanced osteogenic differentiation [165]. Using a similar material system, Singh et al. showed that higher BSNF regions promoted osteocyte maturation for bone formation, whereas lower BSNF regions were more conducive to chondrocyte growth thereby driving cartilage formation [166]. In another approach, Zhang et al. demonstrated magnetic‐field‐driven patterning of superparamagnetic hydroxyapatite nanorods within a GelMA‐based double hydrogel network (Figure 9A) [167]. Beyond mimicking the stiffness gradient, this method further presents the continuous low‐to‐high mineral content from the superficial zone to the subchondral bone zone of native osteochondral tissues, demonstrating their feasibility to guide site‐specific cell differentiation to repair osteochondral defects.

FIGURE 9.

Application of gradient hydrogels for tissue engineering and regenerative medicine. (A) Magnetic‐field‐patterned superparamagnetic hydroxyapatite nanorods (MagHA) within a double network hydrogel presenting continuous HA and stiffness gradient to facilitate repair of full‐thickness osteochondral defect. Adapted with permission [167]. Copyright 2023 John Wiley and Sons. (B) Stiffness gradient methacrylated hyaluronic acid (HAMA) hydrogel encapsulated with fibroblast spheroids accelerated full‐thickness wound healing compared to uniform gels. Adapted with permission [170]. Copyright 2025 Elsevier. (C) Gradient granular microgel suspension with tunable physical properties enabled spatially controlled adipose derived stem cells (ADSCs) differentiation into adipocytes and osteocytes, as indicated by FABP4 expression and mineralization levels, respectively. Reproduced with permission [180]. Copyright 2022 John Wiley and Sons.

Biophysical gradient hydrogels have also been explored for skin regeneration. Such studies are particularly valuable in elucidating how differential stiffness influence cell alignment and matrix remodeling [53, 168], with potential clinical translation for scar‐free wound healing and full‐thickness skin grafting. The recent work by Yee et al. is a good example that exploits silver ions (an established antimicrobial agent) as crosslinker to create gradient hydrogels using human hair keratin (HHK), effectively mimicking the heterogenous microstructure across the epidermis and dermis layers of native skin [169]. An in vivo study using a full‐thickness wound healing mouse model showed that the HHK gradient gels promoted greater reepithelization and collagen deposition compared to uniform GelMA hydrogels, which indicated a better healing quality, thus showcasing its enormous potential as an antibacterial dermal template for skin regeneration. In another interesting example, Kwon et al. proposed the integration of fibroblast spheroids into methacrylated hyaluronic acid (HAMA) hydrogels (Figure 9B) [170]. The presence of a matrix stiffness gradient matching that of the wound site in a rat model was reported to accelerate wound closure and hair follicle regeneration compared to the control or uniform hydrogel groups, further asserting the critical role of gradients in functional tissue regeneration.

Other promising areas where mimicking the gradient ECM stiffness and understanding cellular mechanoresponsiveness can guide the design of materials‐based therapies include nerve regeneration [130, 171], angiogenesis [172], and corneal healing [173]. In this regard, the structural complexities of biological tissues present an important challenge in the clinical deployment of gradient hydrogel technologies because the gradient can occur along different spatial axes or in a non‐linear pattern [174]. Thus, future development in this area is expected to be focused on improving the capabilities to generate complex gradients. Remarkable progress has been made using 3D bioprinting techniques [175], and could perhaps benefit from a combined implementation with other methods such as ice‐templating [176] or thermophoretic [100] gradients during the printing process to enable further control over the microscale features.

Further to these examples, there has been a noticeable attempt to form 3D tissue engineering scaffolds through the assembly of microgels (1–1000 µm), termed as microporous annealed particle (MAP) hydrogels [177]. A unique advantage of MAP hydrogels is their highly interconnected pore structure that enhances cell spreading and migration compared to conventional nanoporous hydrogels [178]. The inherent modular nature of the microgel assembly process also makes it uniquely suited to create gradient scaffolds, as microgels with varying sizes, composition, mechanics, and encapsulated content can be patterned within a single construct [179]. For example, Molley et al. recently described the 3D printing of size gradient GelMA microgels, demonstrating cellular and nuclear morphological adaptations to the corresponding variation in interstitial porosity, which in turn dictate stem cell lineage specification (Figure 9C) [180]. Specifically, large microgels without interstitial filler facilitated cell elongation on the microgel surface reminiscent of 2D adhesion on planar substrates which drove osteogenesis [181]. Contrastingly, small microgels with interstitial filler promoted round cells with preferential adipogenesis, consistent with findings from prior 3D mechanotransduction studies [182, 183]. Notably, this work reinforces the importance of biophysical cues in guiding initial cell adhesion and cytoskeletal tension to steer stem cell differentiation, and elegantly demonstrates the application of fundamental matrix structure‐cell function relationships in enabling hierarchical tissue biofabrication. In addition, Xin et al. utilized a microfluidic mixing and droplet generator platform to demonstrate stiffness and degradability gradients within PEG‐based MAP hydrogels to explore the interaction between mesenchymal stem cells and the material cues [17]. They also showed that the gradient stability could be maintained after implantation into a mouse femoral defect. Interestingly, studies have shown that cell spreading and proliferation in MAP matrices follows trends observed in 2D planar assays rather than conventional 3D hydrogel models, likely influenced by the size of individual microgels. Progress in this area is nascent, but interesting studies are emerging to showcase the versatility and potential offered by gradient MAP matrices for spatiotemporal programming of cell biology.

5. Outlook

The ability to present spatially organized mechanical cues at cell‐relevant scale has created many unique opportunities to probe cell mechanobiology in vitro, contributing to bridging our knowledge gap between simplistic well‐plate/bulk hydrogel studies and sophisticated in vivo animal models. Modern stiffness gradient platforms can reproduce several quantitative descriptors of biological stiffness landscapes remarkably well, including the absolute stiffness ranges and magnitude of the gradient slope. Even if they are still fairly basic and face numerous practical challenges that must first be resolved, 2D linear gradient gels have already been successfully applied to uncover previously unappreciated insights into stiffness‐mediated cell behavior with specific ECM substrates. Such approaches can offer tissue‐level cues that reflect the altered ECM amount and composition that are often associated with local stiffness changes in diseases.

Beyond fundamental inquiry of how mechanical cues impact cellular responses, gradient hydrogel studies are useful to establish the clinical relevance of mechanosensation and mechanotransduction. At present, the pursuit of minimal hydrogel complexity to achieve some target biological outcomes remains a key driving force for innovation. Yet, as the field continues to make important strides to better control the gradient dimensionality, temporal evolution, and multi‐scale matrix architecture, these systems remain far from truly replicating the dynamic ECM milieu of native tissues. Nevertheless, with the future of translational mechanobiology in mind, insights gained from stiffness gradient hydrogel studies will be a critical first step, not only to understand the pathophysiological impact of tissue ECM mechanics, but also to explore potential directions for improving disease therapies through increased fidelity of in vitro biomaterial platforms that include key mechanobiological concepts.

As interests grow in the development of biomaterials‐based platforms for mechanobiology, tissue engineering and regenerative medicine, microengineered stiffness gradient hydrogels are expected to keep exhibiting active vitality in inspiring investigations of fundamental cellular processes. In parallel, the rise of both advanced tissue models and mechanotherapy will contribute to the creation of novel techniques and preclinical applications using gradient hydrogel platforms, allowing for this seemingly niche research tool to expand utility across broad biomedical domains and applications.

5.1. Challenges and Limitations

5.1.1. Finding the Right Balance Between Technological Novelty and Biological Utility

As in the case with most technologies, the deployment of gradient hydrogels presents its own challenges. Achieving precise and reproducible fabrication of stiffness gradients is perhaps the greatest challenge. Presently, diffusion‐based methods appeal as the simplest among the suite of fabrication techniques, but the precise modulation of stiffness profile is majorly challenged by the variability of external environmental factors. Such limitations often impede the transfer of technological developments for more complex and higher throughput biological studies, restricting further evolution of gradient hydrogels. Additionally, flexible patterning congruent with the full mechanical landscape of physiological and pathological tissues is vital for acceptance in mainstream biological research. However, the fabrication capabilities of existing gradient hydrogel platforms in this regard are inadequate.

Here, it is important to emphasize that every gradient generation method has its own achievable stiffness ranges (Table 2). Consequently, it is common practice for a laboratory to adopt multiple different techniques, each covering a specific a specific range across the broad biological mechanical landscape [82, 95]. Conventional diffusion‐based methods are generally limited to producing low gradient strengths (<15 kPa mm⁻¹) due to molecular diffusion kinetics. In a similar principle, the graded photomask technique yields softer and shallower stiffness gradients due to diffusion of polymers and radicals over the time scale required for complete polymerization. By contrast, recent advances in external‐field‐based approaches that enable active manipulation of hydrogel precursor components have led to increased flexibility and precision of gradient formation (from ∼1 kPa mm⁻¹ up to 100 kPa mm⁻¹). In such cases, the material system becomes a main factor limiting the maximum stiffness range; for example, polyacrylamide hydrogel can be tuned from ∼0.2 kPa to 200 kPa, whereas collagen hydrogel only exhibits a narrow stiffness window of <5 kPa. Various strategies have also been proposed to further expand the fidelity of stiffness gradient hydrogels, including combined two‐step polymerization and UV photomask technique for anisotropic patterning [80], and development of a modularized platform using thermophoresis effect for customizing complex gradient shapes [100].

More sophisticated fabrication systems guided by an engineering‐oriented approach may facilitate substantial performance improvements, though widespread adoption of these technologies is often stifled by the requisite infrastructure and specialized knowledge. Thus, deliberate design choices should be made to lower the barrier of entry for new users, further highlighting the importance of a continued dialogue between engineers and mechanobiologists during platform development. The goal should be to create solutions that build on biologically validated approaches, while adding functionalities that enable new avenues of biological inquiry.

5.1.2. Enhancing Correlation of Cellular Responses With Spatial Gradients

The ability to probe a continuum of material cues on gradient hydrogels is a double‐edged sword because it limits the use of many molecular assays such as qPCR and Western blot. Instead, most studies based on gradient hydrogels rely on optical imaging approaches to gather information on the cell state and behavior (e.g, cell morphology, mechanosensitive protein expression). However, an intrinsic drawback of this approach is the lack of a direct stiffness indicator on the substrate, making it difficult to spatially map and correlate cellular responses with stiffness gradients. The current gold standard AFM technique for stiffness characterization requires physical nanoindentation on the hydrogel surface, which is often prone to sample damage and very time consuming, making it practically intractable to measure sample replicate for large‐scale cell studies. To address this problem, increasing attention is being paid to alternative fluorescence imaging strategies that enable in situ and real‐time measurement of the hydrogel matrix mechanics [184]. As demonstrated by several papers examined in this review, this could be implemented with a range of different staining methods, such as diffusion of fluorescein within the hydrogel precursor mix [185], incorporation of fluorescent microbeads [16, 95], and the use of fluorescently tagged polymers [61]. Further exploration should focus on the adaptability and ease of integration of these methods into existing experimentation workflows, which is critical to enhance the practical value of gradient hydrogel systems.

5.1.3. Improving In Vivo and Clinical Relevance

Despite promising progress in gradient hydrogel technology, many studies remain in the nascent stages, grappling with obstacles to validate the physiological and clinical relevance of findings obtained from these in vitro hydrogel culture studies. Today, the most robust approach to platform validation is to perform comparative assessments using data generated by ex vivo tissue samples or in vivo animal models [30, 80, 81, 185]. However, these specimens are often hard to obtain at the target tissue site or condition, especially for direct comparison with human‐based systems. Ideally, microscale mechanical characterization of the tissue samples and hydrogel substrates should also be performed using the same technique, but this has proven to be technically very challenging with uneven tissue surfaces. Overcoming this challenge will require technological innovations for new tools to probe the stiffness in living tissues. On a related note, other analytical modalities could be used on the tissue samples to extract additional relevant biological signals. For example, in the previously described study by Zhang et al. [30], they used Raman microprobe spectroscopy to show that the lower stiffness gradient of PDL‐AB enthesis in periodontitis is associated with lower mineral content gradient across the tissue interface. Progress in this direction could further enhance confidence in the causal relationships deduced from in vitro studies and will be critical in enabling deeper mechanistic insights into cell mechanobiology.

Cells typically reside in a 3D environment in vivo. However, disproportionately few studies of stiffness gradient hydrogels have been conducted in 3D, precisely because of the technical challenges of replicating the spatiotemporal complexity of 3D microenvironmental cues in vitro. Early studies predominantly focused on probing cells in a 2D monolayer context using polyacrylamide‐based gradient gels, but this material does not permit cell encapsulation. With the advancement of hydrogels that support 3D cell growth, studies using homogenous model systems have demonstrated markedly different mechanotransduction behavior because of the additional factors at play, including confinement, cell density, matrix degradation, and stress relaxation [186, 187]. For instance, mesenchymal stem cells in 2D culture display enhanced spreading and nuclear localization of YAP with increasing substrate stiffness, whereas spreading in stiffer 3D environment is limited by the cell's ability to mechanically deform the matrix [137]. Matrix degradability has also been shown to influence the capacity of cells to remodel their local environment and generate traction to steer stiffness‐mediated stem cell differentiation [188]. Such complex biophysical regulation of cell responses is a critical area of research focus given the translational potential for developing regenerative therapies or creating functional disease models. Currently, the most promising approaches to create 3D stiffness gradients are those based on passive diffusion and photomask patterning, but the low reproducibility and consistency issue remains problematic. On the other hand, transferability of external‐field‐driven methods from 2D to 3D gradient platforms has been majorly hindered by the difficulty in maintaining cell viability during the gradient formation process (e.g., when exposure to extreme temperatures is required). To expand the impact of in vitro stiffness gradient platforms, an immediate practical challenge is to identify optimal combinations of hydrogels and processing techniques that meet the requirements for the intended application, including spatiotemporal control, cell compatibility, and production. For example, protocols based on droplet microfluidics for photocrosslinked hydrogels can draw on the existing wealth of cell microencapsulation literature to achieve shape‐controlled hydrogel engineering and higher throughput processing [143]. We can also expect the development of fundamentally new gradient generation principles that will be imperative as new hydrogel chemistries continue to emerge [100], such as alternative gelation strategies to avoid potentially harmful UV light and cytotoxic crosslinkers, or very soft hydrogels demanded for modeling some tissue types. Finally, it is important to note that AFM is generally a surface characterization technique, which means that this technique alone does not provide a direct measurement of the pericellular matrix stiffness. Thus, as the community moves into 3D stiffness gradient hydrogel cultures, it will also be imperative to have complementary approaches to quantify the underlying matrix mechanics and biological responses, with emerging efforts including in optical imaging being promising [189].

5.2. Trends and Opportunities

5.2.1. Probing Beyond Stiffness

As these technical limitations are resolved and new applications are identified, microengineered gradient hydrogels hold great promise to advance the frontiers of research in mechanobiology. Although ECM stiffness is established as a major regulator of cell behavior, recent studies have highlighted viscoelasticity [24, 190] and fibrous microstructure [191] as key emerging matrix cues influencing cell function and behavior. Thus, it is expected that future development of gradient hydrogel systems will transition away from purely elastic and planar substrates (e.g., polyacrylamide, PDMS), toward more complex hydrogel chemistries and biofunctionalization approaches that can emulate these intrinsic features of the native ECM. Realistically, living tissues exhibit a range of stress‐relaxation properties (i.e., reduced resistance to mechanical deformation over time) [2]. Studies have postulated that this viscous dissipation is critical in permitting cells to apply cellular traction forces and to remodel their surrounding matrix [66, 192]. Crucially, these phenomena may partially explain why more degrading, physically crosslinked hydrogels (e.g., collagen, supramolecular assemblies) have been amongst the most successful biomaterials in regenerative medicine to date [193]. Moreover, the viscoelastic behavior of native ECM is also related to its fibrous architecture. The forces exerted by cells on the surrounding fiber structures can lead to transient nonlinear strain stiffening of the matrix, which has been implicated in self‐generated stiffness gradients [194] and long‐range cell communication [195]. In light of recent success in recapitulating these properties using bulk hydrogels, future studies should explore the extension of this concept to engineering gradient hydrogel systems enabling a more accurate interpretation of dynamic cell‐matrix interactions.

5.2.2. Advanced In Vitro Tissue Models

Beyond investigations of cell mechanosensing and mechanotransduction processes, it is believed that the future impact of gradient hydrogel technology may also be closely linked to recent, significant expansion in advanced microphysiological systems and personalized medicine. The OoCs examples discussed in this review highlighted the advantages of incorporating hydrogel‐based ECM to improve their physiological relevance for tissue modeling. Yet, current efforts in the field mostly employ uniform substrates to mimic a single tissue region. While the synergistic combination of spatially controlled stiffness gradients with OoC models may bring non‐trivial design and engineering challenges, the outcome could provide new advanced capabilities to simulate the dynamic pathophysiological interactions at tissue interfaces in ways that has generally been impossible using conventional organ‐on‐chip approaches alone [156]. In a similar vein, organoids are a leading in vitro technology that allows researchers to study a simplified version of organ systems in the context of tissue development, disease models, and drug testing [196]. Toward realizing the goal of complex models that can more faithfully recapitulate the in vivo structure and function, multiple studies have demonstrated the potential of using hydrogel platforms to investigate how physical cues affect organoid behavior, though they have only been explored at discrete levels [197, 198]. Given the intrinsic advantages of microfabricated stiffness gradient hydrogels, one can envision that these materials could similarly be applied as a screening tool to study organoid mechanoresponses in a dose‐dependent manner. In another promising future function, localized stiffness patterns could be leveraged to guide deterministic organoid formation, as demonstrated by a recent study of spatially controlled crypt formation [199]. Ultimately, the success of microfabricated gradient hydrogel organotypic models hinges not only on the ability to recreate a physical microenvironment representative of the native conditions, but also on compatibility with characterization methods that can resolve the spatiotemporal effects of the engineered microenvironmental cues on cellular mechanochemical signaling [200].

5.2.3. Data‐Driven Innovations

Gradient hydrogels created using soft hydrogels are promising, but their design and optimization remain a complex task due to the multidimensional variable space and multiscale structure‐property relationships. Traditionally, development of gradient hydrogel systems has been progressed by experimental trial and error, though it is often limited by the reliance on informally encoded “expertise” on specific materials and fabrication methods. In this context, the field has much to benefit from the ongoing revolution of data‐driven methodologies, particularly in driving predictive design frameworks [201, 202, 203]. This idea was recently demonstrated in the adjacent domain of granular hydrogel matrices, where machine learning techniques were coupled with experimental optimization at each design step to enable efficient exploration of compositional/processing parameter spaces and accurate forecasting of the resulting material characteristics [204]. Perhaps most notably, the smart use of both quantitative and qualitative experimental records in the modeling pipeline was shown to be a powerful means of capturing batch‐to‐batch variabilities and removing biases associated with human analysis of multifaceted processes. In the case of gradient hydrogels, examples of quantitative metrics may include Young's modulus and pore size distribution. Qualitative indicators, on the other hand, are for instance, whether a stiffness gradient was successfully formed and the gradient linearity. Despite huge potential in accelerating the design and fabrication of gradient matrices, it is important to recognize that such data‐driven modeling remains limited in its ability to distinguish between correlation and causation [205]. Future work to address this issue will likely require developing physics‐informed hybrid frameworks that can more directly embed core mechanistic principles into the predictive process [206]. Nevertheless, as an important first step, concerted efforts to map the data infrastructure and machine learning landscape with problem‐specific guidelines could ease the adoption of data‐driven strategies by the hydrogels research community, thus promising greater speed and impact of technological contributions to the mechanobiology field. To generate more predictive models, the quality and consistency of input data are also critical, reinforcing the need for all the model development and measurement improvements proposed throughout the review.

6. Concluding Remarks

With the advancement of gradient hydrogel technology, new opportunities are emerging to harness its unique capabilities and potential for mechanobiology research and its translational applications. Microengineered biomimetic stiffness gradients of human tissues can enable more physiologically relevant and predictive in vitro platforms to explore the impact of isolated biophysical and biochemical cues on cellular responses in ways that have not been possible using simple bulk hydrogels. The achievements we have highlighted in this review help to illustrate how new technology platforms continue to emerge and improve over time as biomaterials and microfabrication techniques develop. At present, gradient hydrogels largely focus on static linear stiffness gradients as a reductionist approach to study single‐cell mechanobiology, but also increasingly to probe multicellular structures such as spheroids. However, with rapidly growing research interests in translating basic mechanobiological knowledge into functional disease models and regenerative therapies, it is not difficult to foresee that future platforms will begin to build in more spatiotemporally complex and tissue‐specific matrix cues beyond substrate stiffness alone. Building this complexity requires compatibility of the hydrogel chemistry and processing techniques with the application of interest, and many efforts have been directed toward identifying the right problems in biology and medicine to solve using gradient hydrogel technology. It is certainly an exciting area to monitor as further progress is made with new hydrogel chemistries, fabrication methods, and mechanobiology applications.

Author Contributions

S.W.C. and D.V. conceptualized the manuscript. S.W.C. wrote the original draft. D.A. assisted in the manuscript writing. D.V., A.W., and M.M.M.B. supervised and revised the manuscript. All authors have reviewed and approved the final version of the manuscript.

Conflicts of Interest

The authors declare no competing interest.

Data Availability Statement

The authors have nothing to report.