Designing Biomaterials to Modulate Notch Signaling in Tissue Engineering and Regenerative Medicine

Notch modulation in vasculogenesis and angiogenesis

Vasculogenesis is the differentiation of precursor cells (angioblasts) into endothelial cells (ECs) and the de novo formation of a primitive vascular network, whereas angiogenesis refers to the growth of new capillaries from preexisting blood vessels. Notch signaling is critical in both vasculogenesis during development and in angiogenesis.^24,25 It protects the endothelium, ²⁶ controls vascular smooth muscle cell (VSMC) phenotype,^27–30 promotes neoangiogenesis,^31–33 tip–stalk cell patterning, ³⁴ and regulates arteriovenous specification.^35,36 The various roles of Notch signaling in the vasculature, including its influence on ECs and smooth muscle cells (SMCs), have been reviewed previously.^8,37 Herein we highlight important Notch receptor-ligand pairings in the vasculature. VSMCs dominantly express Notch1, Notch2, and Notch3 receptors, while ECs express Notch ligands Jag1, 2, and Dll4 and to some extent Dll1 in the remodeling vasculature.^38,39 The context in which these ligand and receptors occur is essential; however, the anatomical location within the vascular bed and the associated physiological forces (flow and stress) are likely to impact the context-dependent Notch activation. ³¹

The development of the vascular system involves tip cell selection, sprout formation, tip cell migration, stalk cell proliferation, and ultimately vascular stabilization, ⁴⁰ which are collectively influenced by Notch signaling. Vascular sprouting is guided by the migration of tip cells in response to a matrix-bound vascular endothelial growth factor (VEGF) gradient,^34,41 with Dll4 acting downstream as a negative regulator.^42,43 Tip–stalk cell fate plays a significant role in angiogenesis since tip cells direct new blood vessel growth. Interestingly, the role of Notch in both tip and stalk cells is evident within the distribution patterns of Notch signaling ligands and receptors. VEGF signaling induces Dll4 in tip cells, and tip cells then suppress tip–cell features in adjacent stalk cells through Dll4/Notch-mediated lateral inhibition. ³⁴ Simultaneously Jag1 antagonizes Dll4-mediated Notch activation in stalk cells to increase tip cell number, which consequently enhances vessel sprouting. ⁴⁴ Hence it is this Jag1 and Dll4 “salt-and-pepper” pattern that dictates the tip–stalk cell phenotypes within this niche.

The function of Notch signaling varies greatly depending on the location of the vascular bed and also the type of the vessel. The arteriovenous specification is established early in development through a variety of transcription factors. A key Notch-defining factor in this specification is the Notch3 receptor found in arterial SMCs, which is notably absent in veins, and Jag1 expression by ECs is responsible for this maturation. The transcription factors Foxc1 and Foxc2 and VEGF signaling are primarily responsible for arterial fate^36,45; the upregulation of these transcription factors results in increased expression of Dll4. In contrast, vein identity is regulated by the repression of Notch1. ³⁶

Recruitment of mural cells (VSMC and pericytes) and the formation of a fully endothelialized lumen are hallmarks of arterial vessel maturation during development. ³⁶ Upon activation of Notch in VSMC, ligand-receptor signaling is initiated throughout the VSMC lamellae by the process of lateral induction. The propagation of Notch signaling is crucial for regulating VSMC phenotype throughout the vascular wall and, hence, is a critical phenomenon for inducing differentiation of the complete VSMC layer toward the homeostatic contractile phenotype. ⁴⁶ Notch3 targeting is more prominent for SMC control and differentiation^27,47,48; Notch1, in contrast, has been shown to regulate EC metabolism, proliferation, and monolayer regeneration. ⁴⁹ Communication between ECs and SMCs in the vascular wall is essential; thus, diminished Jag1 expression in ECs leads to abnormal smooth muscle development. Endothelial expression of Jag1 is required for activation of Notch3 on VSMC maturation, differentiation, and contraction. ⁵⁰

The impact of physiological forces within the vascular system on Notch signaling becomes apparent when considering the ability of the endothelium to respond to force patterning. ECs modulate its Notch component expression in response to hemodynamic forces. For instance, Notch1 activation in EC is sensitive to blood flow, where high shear stress has a critical role in the acquisition and maintenance of arterial identity through its role in endothelial quiescence. ⁵¹ The indication that EC appears to respond to predetermined arterial or venous patterns is evident through current research, and there is a certain level of plasticity cues of the local niche suggested to be imposed by physical forces, such as hemodynamics. ⁵² Taken together, ligand specificity and selective Notch activation regulate differential phenotypes and function within the vascular tissue.^46,53,54 Understanding tissue distribution of Notch signaling components is highly relevant to translate them into tissue engineering and regenerative medicine strategies.

Notch modulation in cardiac tissue specification and repair

Notch signaling also has a critical role in both cardiac development and cardiac repair. Notch signaling is very active during embryonic cardiac development and has roles in both the myocardium and the endocardium. Notch regulates ventricular trabeculation, ⁵⁵ cardiomyocyte differentiation and proliferation, ⁵⁶ and valve formation. ⁵⁷ We do not delve too much into Notch-mediated cardiac development, disease, and regeneration as it has been recently reviewed^58–61 ; however, we provide a brief discussion relevant to tissue engineering. Notch1 and Jag1 are expressed in immature cardiomyocytes to protect them from apoptosis and stimulate proliferation and differentiation into fully differentiated and beating phenotypes.^56,62,63 Although much attention is focused on the role of Notch in the myocardium, there is an essential function of Notch in the endocardium. Early endocardial Notch activity contributes to the patterning of the embryonic endocardium for the development of valve and chamber formation, ⁶⁰ and then later, it regulates signaling critical for outflow tract valve morphogenesis^64,65 and ventricular chamber development. ⁶⁶ The endocardium is a specialized endothelial lining that closely juxtaposes to the myocardium to communicate. At a cellular level, Notch1 expression is downregulated in the cardiomyocyte lineage and is enriched in the endocardial lineage, corresponding to antimyogenic activity and its functional requirement in the endocardium.^58,67 Notch plays a critical role in endocardium–myocardium communication that underlies trabeculation ⁶¹ ; Dll4 is found in the endocardium where trabeculae are formed. ⁶⁸ In contrast, Jag1 is expressed in the cardiomyocytes forming the trabeculae and is also present in the compact myocardium. Although Notch is required for trabeculation, it is not active in the ventricular endocardium during trabeculation, but otherwise controls trabeculation through intermediate signaling pathways. ⁶⁹ The Dll4-Notch1 pairing actively regulates gene expression, encoding signals that connect the ventricular endocardium and myocardium, but also promotes cardiomyocyte proliferation, differentiation, and metabolism.^58,70

Interestingly, Notch signaling is silenced in the postnatal heart and only restored after injury. Notch1 has been identified as a critical determinant receptor in the heart, as it interacts with various ligands in cardiac repair following injury. ⁵⁹ Because Notch signaling is directly linked to repairing mechanisms to drive regeneration, ⁷¹ rather than scar formation, it is an interesting target for therapeutic cardiac injury intervention. The inhibition of endothelial Dll4 in the heart resulted in impaired fatty transport, leading to metabolic and vascular remodeling. ⁷⁰ The upregulation of Notch1 in a hypertrophic heart controls the adaptive response of the heart to stress, which limits the extent of hypertrophic response and contributes to cardiomyocyte survival. ⁷² This demonstrates the cardioprotective role of Notch1, controlling the response to injury in the adult heart.

Notch modulation in the developing bone and repair

Skeletal tissues, specifically bone, is another tissue where Notch signaling plays an important role during development and disorder.^73,74 The roles of Notch signaling include control of bone regeneration,^75,76 skeletal remodeling, bone fracture repair,^77,78 congenital bone defects, and malignancies of the bone. ⁷⁹ Since Notch is tissue specific, Notch directed cellular response is dictated by the specific cells targeted and its stage of maturation. A coordinated effort by both chondrocytes and osteoblasts during embryogenesis leads to skeletal tissue specification. It is then continuously regenerated by the action of bone-forming osteoblasts and bone-resorbing osteoclasts to maintain homeostasis.

In early-stage skeletal development, Notch activation inhibits cell differentiation and causes cancellous bone osteopenia because of impaired bone formation. ⁸⁰ Notch ligands, specifically Jag1, are involved in cell commitment; Notch inhibitors can control intramembranous ossification where mesenchymal tissue is replaced directly by bone through osteoblastogenesis. ⁸¹ Although initial studies have implicated both Notch1 and Notch2 receptors in the regulation of osteoblastogenesis, ⁸² subsequent studies concluded that Notch2, instead of Notch1, plays the predominant role. ⁸³ Cell commitment becomes an important factor in Notch control since the activation of Notch signaling inhibits terminal differentiation of osteoblast progenitors (through Jag1) but does not affect mature osteoblasts. ⁷⁶ Synergistically, osteoclasts work with osteoblasts to control bone homeostasis, and this is influenced by Notch. Osteoclastogenesis from macrophage precursors is similarly regulated by Notch signaling; however, Notch1 is the predominant receptor. Similar to the vasculature Dll and Jag ligands have opposing functions on bone tissue response but are again context dependent. The Notch1/Jag1 axis suppresses osteoclastogenesis, whereas the Notch2/Dll1 axis promotes the effect. ⁸⁴ In bone remodeling, the disruption of Notch in osteocyte differentiation in the mineralized matrix can affect calcium phosphate deposition ⁷⁵ and alter crystal mineral structures, linking Notch with dysregulation and its role in bone mineralization. ⁷⁴ Collectively, it is evident that Notch signaling regulates several aspects that control osteoclasts, osteoblasts, and osteocytes, which are critical in bone tissue.

Notch signaling is also important in bone repair following injury. Fracture healing is caused by endochondral ossification and intramembranous bone formation. ⁸¹ Therefore, achieving a balance between osteoblastogenesis and osteoclastogenesis, alongside with the infiltration of blood vessels, mediates proper fracture healing. The interaction of mesenchymal, hematopoietic, and vascular cells plays a Notch-dependent role. For example, in bone marrow-derived mesenchymal stem cells (MSCs) reduced Notch signaling demonstrated depletion of this population, leading to fracture nonunion.^74,78 The role of Notch in fracture healing is complex, with both upregulation and downregulation playing a part depending on ligand-receptor selectivity, and needs more research to fully understand the mechanisms. To highlight ligand-receptor specificity, Jag1 delivery has been shown to induce osteoblast commitment, ⁷⁶ mineralization deposition, ⁷⁵ and promote healing of craniofacial and appendicular defects. ⁸⁵ In contrast, disruption of Jag1 signaling has proven to decrease femoral trabecular bone formation, ⁸⁵ accelerate cartilage regeneration, ⁸⁶ and decrease joint inflammation. ⁸⁶

Diseases and disorders associated with Notch dysregulation

Notch signaling has distinct and diverse biological roles in many tissues; therefore, it is not surprising that its dysregulation has been implicated in human diseases and disorders. ⁸⁷ In many cases, mutations at the genetic level are attributed to disease and disorder; however, in others, the cause of the disease is less direct and associated with improper ligand-receptor distribution or dysregulation of receptor-ligand activation complexes. Both loss- and gain-of-function activity of Notch signaling are unique to each disease/disorder. Table 1 summarizes the roles of Notch signaling and Notch related diseases/disorders in selected tissues relevant to tissue engineering (the list is not exhaustive). Although Notch signaling may not be the primary cause for these diseases/disorders, Notch has a role in each of these diseases and could be a useful therapeutic target.

Targeting ligand–receptor interactions

To utilize Notch signaling during development as inspiration for engineered tissues, an understanding of ligand–receptor binding and specificity is necessary^20–22 ; therefore, one approach is to modulate the ligand–receptor interaction. As stoichiometric ligand–receptor interactions are necessary for pathway modulation, a distinction between cis- and trans-interactions among the ligands and receptors becomes crucial. ²⁰ All ligands, except for Dll3, can activate receptors in trans ⁶ and trans-activation, whereby the signaling ligand and its receptor are distinct, but contacting cells is the most common mechanism. Delivery of Notch ligands in their soluble form will not activate Notch signaling and rather inhibits activation.^28,112,133

An alternative strategy is if both the ligand and the receptor are expressed on the same cell surface. There are, in fact, cells that express both the Notch receptor and the ligand. As an example, in embryonic rat VSMCs, Notch3 and Jag1 are expressed in equal abundance. ¹³⁴ Despite this, Notch signaling was not activated when these cells were cultured alone or cocultured with Jag1 presenting ECs. ¹³⁴ Depending on the relative abundance of receptor to its ligand on the same cell surface, it can either be a ligand presenting (signal sending) or signal-receiving cell. However, when the Notch receptor binds to a ligand present on the same cell, signaling is inhibited. This phenomenon of cis-inhibition is an emerging feature of Notch signaling and relies on the relative abundance of the ligand and receptor.^135,136 Although it remains to be elucidated further, cis-activation is implied when trans-ligands are effectively competing with cis-ligands for a common pool of Notch receptors. ¹³⁶ Currently, Notch therapies targeting Notch ligand–receptor selectivity and binding include the use of binding proteins,^137,138 monoclonal antibodies (mAbs),^139–141 and engineered Notch decoy ligands.^142,143 These strategies are used to preferentially block Dll or Jag type binding specific binding. For example, in engineered soluble peptides composed of different Notch EGF repeats 11-24 (N11–N24) fused to IgG γ-heavy chain, the Notch decoy N11-13 preferentially blocked Dll-type ligands, whereas N10-24 was specific to Jag-type inhibiton. ¹⁴² The use of soluble Notch factors may also be used as a binding inhibitory mechanism.^28,112,133

Controlling Notch receptor processing

Notch receptor processing is another avenue that can be harnessed for enhancing or inhibiting proteolytic cleavages within the signaling cascade to control various cellular responses. S1 cleavage is controlled by Furin convertase; therefore, Furin inhibitors are ideal targets to control the abundance of receptors on the cell surface. Notch2 is insensitive to Furin inhibitors, whereas the surface expression of Notch1 is mildly reduced when S1 processing is impeded. ¹⁴⁴ Furin convertase targeting has been suggested to direct but not inhibit signaling and has some other target proteins such as transforming growth factor β1 (TGFβ1); therefore, there is an increased risk for off-target effects ⁶ and is not advantageous. The second cleavage found in the Notch cascade is the S2 cleavage, which is engaged upon ligand–receptor binding and is controlled by ADAM10 or ADAM17. ADAM inhibition is a possible therapeutic target for ligand processing and Notch modulation ¹⁴⁵ ; however, off-target risks remain. ⁶ The third receptor processing step occurs at the S3 cleavage site by γ-secretase and has been suggested as a possible avenue for intervention. Gamma-secretase inhibitors (GSIs) are used to inhibit S3 cleavage, attenuating the response of all Notch receptors. Targeting through GSIs diminishes the release of Notch from the plasma membrane and subsequent generation of NICD. The effectiveness of GSI treatments (e.g., DAPT) has shown promising results, and some have advanced into clinical trials, ¹⁴⁶ including its use as a tumor suppressor in cancer therapy.^147,148 Limitations of GSI treatments to inhibit the effects of Notch signaling include low treatment effectiveness, toxicity, and off-target effects blocking numerous other protein functions, ¹⁴⁹ suggesting a need for additional therapeutic S3 cleavage approaches. mAb inhibitors are potentially advantageous due to their specificity allowing for the targeting of individual Notch receptors. mAbs have been directed toward Notch receptors, both targeting the extracellular negative regulatory region or blocking the conformation change that allows the ADAM protease cleavage to occur, and clinical attempts have been made.^150,151 The process of endocytosis and endosomal sorting of Notch receptors is another therapeutic target through membrane-transport networks. Because of several excellent reviews on this topic,^9,152–154 it is not discussed here.

Notch targeting through signaling cross talk

The ability for Notch signaling components to interact, impact, and cross talk with multiple signaling pathways adds another opportunity in controlling Notch signaling and developing new therapies. It is clear that Notch signaling is not an isolated pathway and interacts with signaling pathways mediated by TGFβ/bone morphogenic proteins (BMPs),^155–158 VEGF, ¹⁰⁴ Wnt, ¹⁵⁹ and YAP/TAZ. ¹⁶⁰ Many proteins have been identified that interact with the Notch ICD (these have been summarized previously ¹⁶¹ ). Although our current understanding of the molecular mechanisms involved in Notch cross talk is limited and emerging, it is evident that the pathways are integrated and can cause an attenuated or amplified behavior with a purpose to influence tissue specification. This integration is interesting because it might be able to explain the initial asymmetries found in Notch activity before ligand–receptor binding.

TGFβ/BMP signal convergence with Notch signaling is a well-established cross-talk interaction. Notch is known to have an influence controlling the TGFβ pathway with the expression of R-Smads and the synergistic regulation and interactions with the NICD and Smad3 to regulate a subset of Smad3 target genes.^126,134 In addition, the integration of Notch and Smad1/5 signaling might be important. ³⁴ These interactions, however, are cell dependent; in some contexts such as within keratinocytes and epithelial cells, there is a cooperative amplified behavior, and conversely, in other cases, the NICD was found to antagonize TGFβ growth arrest and transcription. ¹⁶² Although other pathways, for example, YAP/TAZ interactions have been studied and hypothesized to be linked, for example, with Notch related to the Hippo pathway signaling,^163,164 a strong mechanistic insight has not been given as to the exact interactions of these related pathways. The cooperative effort and cross talk of various signaling pathways with Notch have been demonstrated in many cells and tissues, including cardiac tissues,^57,61 epithelial cells, ⁵⁷ in the epidermis, ¹⁶⁵ angiogeneses,^34,165 and myogenic differentiation^162,166 to name a few. It is evident that Notch interacts and cross talks with various signaling pathways, and the response frequently differs between cell types and tissues. A further understanding of the mechanisms present in this interaction is needed to further understand from a regulation perspective of how this information can be used in various tissue-engineered therapies. Notch signaling alone or combined interaction with other pathways could be harnessed to create proper homeostasis within various tissues and similarly regulate physiological and pathological processes.

Relevance of Notch therapeutics to tissue engineering and regenerative medicine

Most Notch targets to date are focussed on cancer therapies, but there are other applications in tissue engineering and regenerative medicine (Fig. 3). Notch signaling can be amplified or attenuated to differentiate stem cells into a target lineage for regenerative purposes. Delivery of Notch ligand Jag1 to full-thickness wounds significantly enhances wound healing ¹⁶⁷ suggesting the need to incorporate the ligand to wound dressing biomaterials. Activation of Notch1 in injectable cardiomyocytes enhances cardiac repair following infarction. Delivery of Jag1-containing hydrogels also counteracts cardiac fibrosis, primarily through inhibition of myofibroblast differentiation, thus accelerating cardiac repair. ¹⁶⁸ Impaired VSMC differentiation following stent deployment and the subsequent excessive proliferation causes restenosis due to endothelial denudation; therefore, Notch ligand delivery using stents could be useful to modulate restenosis. ¹⁶⁹ Similarly, activation of Notch signaling is needed to repair nonunion bone fracture. ⁷⁸ Instead of activation, attenuation of Notch by either soluble ligands or by one of S1–S3 cleavage inhibitors promotes injectable cell survival and chondrogenic differentiation, thereby increasing the therapeutic potential. ⁸⁶ These illustrative examples underscore the relevance of Notch therapeutics to tissue engineering and regenerative medicine and the need for Notch activation through appropriate ligand presentation strategies.

OPEN IN VIEWER

FIG. 3.

Translation of Notch signaling developmental biology principles to tissue engineering and regenerative medicine. Notch signaling plays a role in tissue development and pathophysiological principles. This signaling pathway is prominent in a wide range of tissues and organs within the body, which include but are not limited to cardiac, vascular, and bone tissue. Here we present selected examples of how Notch signaling in developmental can be translated into a tissue engineering and regenerative medicine approach. Color images are available online.

Notch ligand presentation strategies in tissue engineering

In general, signaling molecules can be presented to cells in one of three ways: (i) adding them as a soluble factor, (ii) direct conjugation or conjugation through a flexible molecular arm to a scaffold, (iii) and affinity immobilization to the scaffold through antibody-binding proteins. As discussed above, adding the soluble form of the ligand does not seem to be effective in activating Notch signaling; hence, attachment of Notch ligands to biomaterial surfaces has been explored and evaluated based on ligand bioactivity on the material surface.^170,171 Covalent conjugation chemistries utilize free reactive groups on the surface scaffolds in a random and uncontrolled manner. Numerous covalent conjugation chemistries can be used to immobilize Notch ligands depending on the reactive groups presented on the scaffold surface (Fig. 2B): SH-targeting maleimide chemistry, OH-targeting carbonyldiimidazole chemistry (CDI), and COOH-targeting carbodiimide chemistry (EDC/NHS). If not already present on the biomaterial surface, reactive groups might also be introduced to the surface of the scaffolds through surface modifications like chemical vapor deposition, plasma vapor deposition, and self-assembled monolayers (SAMs), among others. ¹⁷² These allow for chemical alterations to the surface of materials without significantly affecting its bulk properties. Covalent immobilization assures ligand presence, but the active site of the ligand is not always accessible for binding due to this random attachment to the surface.

Immobilization through a flexible molecular arm has been applied using polymeric spacers such as polyethylene glycol (PEG),^76,173–175 poly(acrylic acid), ¹⁷⁶ and disuccinimidyl suberate ¹⁷⁵ to allow for better accessibility and activity of the immobilized ligand. Spacer immobilization linked with Notch signaling needs to be investigated further because the interplay between tether length, ligand spacing, and ligand accessibility has not been elucidated. Optimal ligand surface coverage can be maximized with spacers due to the ability of the polymer-bound proteins to form a layer and disperse the ligands in space to optimize binding and minimize lateral repulsions. ¹⁷⁷ In fact, the added mobility afforded by longer spacer arms enhanced the formation of focal adhesion due to the ability of cells to reorganize tether peptides on the nanometer length scale. ¹⁷⁸ In translation to the Notch signaling system, there is some evidence for target selectivity of specific Notch receptors by which the ability of Notch ICDs to form dimers might influence the activation of downstream targets, including receptor recruiting of gene response. ¹⁶¹ Interestingly, the configuration of CSL binding sites appearing as monomer or dimers has influenced the likelihood of recruiting Notch1 or Notch3 ICD. ¹⁷⁹ Therefore, when designing Notch biomaterials, the flexibility of the chemical spacers of varying lengths may enhance the dynamic nature of the ligands. To mimic the dynamic regulation of signaling ligands, polymer chemistry can be harnessed to create chemical spacers to improve biomolecular recognition, ligand accessibility, and increase the dynamic behavior of immobilization. ¹⁸⁰

Notch orientation-regulated immobilized scaffolds have been engineered using indirect affinity immobilization strategies. Antibody binding proteins such as Protein A/G, Streptavidin/Biotin binding, as well as anti-Fc antibodies, have been harnessed to control ligand accessibility and orient the ligands with the active site available for binding (Fig. 2C). Compared to covalent immobilization strategies, affinity binding strategies allow indirect immobilization where the active site is oriented controllably for maximized receptor binding capability. Most Notch engineered biomaterials utilize affinity immobilization as it most effectively controls attachment and enhances cellular behavior. Although widely used, affinity immobilization strategies are limited by the binding strength, thus having a higher possibility of dissociating compared to the covalent binding. Protein A and Protein G bind Notch ligands by their biospecific targeting of the Fc domain on recombinant proteins (protein G: k_d∼ 10⁻¹⁰, protein A: k_d∼ 10⁻⁹ M).^181,182 Streptavidin/Biotin binding (k_d∼ 10⁻¹⁵ M) ¹⁸³ immobilizes streptavidin to a biomaterial surface through its reactive amine group where the compound affinity binds up to four biotinylated ligands per molecule of streptavidin. Finally, anti-human IgG Fc capture has a strong affinity for the human Fc domain (k_d∼ 10⁻⁶ M). ¹⁸² Using different ligand attachment and presentation strategies, the effect of Notch signaling on cellular function can be studied in both two-dimensional (2D) and three-dimensional (3D) microenvironments to gain insight into various aspects of Notch signaling biomaterial design.

Notch immobilization on 2D biomaterial surfaces

Notch ligands have been immobilized extensively to 2D surfaces to study the behavior of Notch-directed signaling of several cell types, including hematopoietic stem cells, ECs, SMCs, pericytes, MSCs, keratinocytes, and various cell lines (Table 2). With the exception of Dll3, all other Notch ligands have been immobilized to various 2D surfaces. Dll3 is presumed to not activate receptors in trans; therefore, it is not relevant in heterotypic cell signaling. ⁶ Jag1 is the most commonly immobilized ligand and is used to control a number of cell fate decisions, including angiogenesis, ⁵³ osteogenic differentiation,^76,184,185 epithelial differentiation, ¹⁸⁶ immunoprotection, ¹⁸⁷ blood cell expansion, ¹⁸⁸ and cardiac differentiation, ¹⁸⁹ demonstrating its various functional roles in multiple tissue types.

Two-dimensional systems allow us to study culture characteristics, including dose, activation kinetics, and cell behavior, without the complexity of the 3D microenvironment. As discussed in Notch Signaling in Developmental Biology Section Notch signaling is driven by stoichiometric interactions (rather than being driven enzymatically). Therefore, immobilized dose and ligand density are extremely important in activating or restricting signaling between interacting cells. Dose and ligand density can help to modulate Notch signaling in terms of its signaling mechanism through cis- and trans-relationships, based upon stoichiometric ligand–receptor levels. The response to trans-Delta activation is graded, while the response to cis-Delta exhibits a sharp, switch-like response at a fixed threshold. ²⁰⁵ Based on ligand–receptor interactions, there is an ultrasensitive switch between mutually exclusive sending (high Delta/low Notch) and receiving (high Notch/low Delta) signaling states. ²⁰⁵ Overall this alludes to the influence of dose and density of Notch ligands, which can control cell-fate outcomes in different organ systems to regulate cellular outcomes.

Dose and activation kinetics can also control cellular patterning; thus spatial control incorporating patterning will be expanded on in Spatial cues to control Notch ligands section. Immobilization dose dependency has been linked to precoat solution concentration. ¹⁸⁹ Within a lymphoid cell fate model dose dependence was explored to evaluate cell fate outcomes of hematopoietic precursors to a self-renewed or differentiated cell fate. Lower Dll1 density enhanced generation of CD34⁺ cell types consistent with myeloid and lymphoid differentiation, whereas increased amounts of Dll1 induced apoptosis of CD34⁺ cells. ¹⁹⁹ In addition, dose dependence of Jag1 that mediated enhancement of store-operated Ca2⁺ entry in human pulmonary arterial SMCs was evaluated. ¹⁰⁰

The combination of nano/microbead platforms and protein immobilization has led to several applications to enhance nanomedicine. Nanoparticles are often engineered for drug delivery applications, but are limited by their ability to target specific locations and have biodistribution problems. In the application of Notch signaling, nanoparticles with peptide functionalization have been suggested as carriers for targeted delivery of GSIs to block Notch signaling.^206,207 In addition, nanoparticles can also be used to deliver immobilized ligands to cell surface receptors and act as cell-surrogate biomaterials replacing the signaling cell to direct Notch signaling. Evidently, bone marrow hematopoietic stem cells were directed to a T cell lineage through Dll4 functionalized microbeads. ²⁰⁰ Furthermore, in human embryonic stem cells (hESCs), Jag1 beads were able to significantly increase Notch signal activation, encouraging ectodermal formation after 2 days, as well as cardiomyocyte differentiation of KDR⁺ progenitor cells. ¹⁸⁹ These, as well as other approaches summarized above, demonstrate the success of Notch ligand immobilization to a biomaterial platform to control cell fate decisions. Translation from nonclinically relevant 2D cell culture models into a clinically translatable 3D biomaterial construct is the next step in developing engineered materials to control cell behavior.

Notch ligand immobilization to 3D scaffolds

Immobilization of ligands to 3D biomaterial surface offers an approach to study cellular response, which can better recapitulate the native 3D environments experienced through cell–cell and cell–matrix interactions. Most Notch signaling biomaterials in 3D systems focused on hydrogel systems, although new biomaterials have started to emerge as summarized in Table 3.

Hydrogel based systems are attractive because the mechanical properties of the material can be tuned and controlled to recapitulate the proper cellular niche. Hydrogels are also advantageous due to their biocompatibility and tunable biodegradability. ²¹⁶ PolyHEMA is perhaps the most commonly used biomaterial as it is resistant to protein adsorption, thus providing a “blank slate” for covalent immobilization studies. ¹⁸⁶ PolyHEMA has been used to direct both epithelial ¹⁸⁶ and keratinocyte ²⁰⁹ differentiation, demonstrating the versatility of this hydrogel system. Hydrogel systems have been developed to engineer bone tissue. Notch signaling activation during mid-to-late tissue development drove osteoblast differentiation and promoted the anabolic activity of committed osteoblasts. ⁸⁵ Therefore, from a tissue engineering perspective, activating Notch signaling postinjury could be a potential target for increased bone formation. In addition, Notch signaling biomaterials utilizing Jag1 immobilized to a PBAE hydrogel was used as a novel osteoconductive scaffold for bone tissue engineering. ²¹¹

Other 3D materials that have been developed include applications in cardiac tissue engineering, epithelial differentiation, and immunity. For cardiac tissue engineering, an elastomeric poly(acrylic acid) brush grown on a substrate of poly(L-lactide-co-caprolactone) was developed. This elastomer upregulated myogenic but not osteogenic differentiation of human MSCs on the engineered surface, suggesting a new strategy for bioengineering a cardiac patch in myocardial infarction repair. ¹⁷⁶ Controlling epithelial differentiation has direct application in tissue-engineered organs where a competent epithelial barrier is required. ¹⁸⁶ Promoting squamous cell stratification and rapid differentiation could reduce tissue culture time and improve initial performance. This would also be applicable in percutaneous devices where rapid migration of epithelial cells to the biomaterial surface and cornification/growth arrest upon contact could prevent common modes of failure. An engineered perivascular microenvironment mimicked the endothelial–mesenchymal cell relationship by interacting encapsulated MSCs with surface-immobilized endothelial Jag1. ¹⁷⁴

Another application for Notch 3D biomaterials is improving immune response for diabetes and other diseases. Surface modified biomaterials have been used to shift the immune system from an inflammatory to the anti-inflammatory environment by utilizing pegylated Jag1 islet surfaces to direct splenocyte response and improve immunoprotection of pancreatic islets. ¹⁸⁷ Collectively, these examples demonstrate attempts at designing 3D biomaterials and strategies to recapitulate the native biology. Although progress has been made, further development in biomaterials is needed. Three dimensional microenvironments allow for incorporating both biochemical and biomechanical signaling that can integrate additional signaling tools within an engineered system. The following section will discuss ways to incorporate complementary cues (rather than just biochemical signals) at the biomaterial interface to better mimic 3D microenvironments.

Physical cues to control Notch signaling: matrix stiffness and topography

Matrix stiffness is a mechanical property that can cause effects on cellular adhesion, migration, proliferation, and differentiation. Cell response has lineage-specific commitments when cultured on substrates matching the stiffness of their corresponding tissues in vivo. Diseased microenvironments are often correlated with increased ECM stiffness and dysregulation of cellular response. ²¹⁷ Studies have demonstrated that soft substrates are beneficial for a neurogenic or adipogenic differentiation, substrates of intermediate stiffness favor a myogenic lineage commitment, and stiff substrates are beneficial for osteogenic differentiation.^218,219 Depending on the cell type seeded on the scaffold, the biomaterial substrate should be designed to properly promote cell behavior so that trans-differentiation into other undesirable cell types is minimized. Scaffold stiffness is, therefore, critical as materials can be designed to have various ranges of modulus, which will control cell behavior differently. ²²⁰ Therefore the intrinsic material properties of the scaffold are important for controlling Notch signaling (Fig. 4A-i).

There are only a few studies that investigated the relationship between matrix stiffness and Notch signaling. Within the liver cell niche, both biochemical cues (Notch signaling) and biophysical cues (substrate stiffness) have been linked to the control of liver progenitor cells into both hepatocyte and ciliary cell fates. ²²¹ Progenitor differentiation on stiff (30 kPa) substrates directed peripheral localization of Jag1, Dll1, and Notch2; conversely, there was a lack of expression for Jag1 and Notch2 on a soft (4kPa) substrate. ²²¹ The loss of ligand induction on soft substrates suggests higher responsiveness of Notch signaling within a stiffer substrate in a liver progenitor model. This cited study showed that the segregation of liver progenitor fates is dependent on both Notch signaling and substrate stiffness. ²²¹ In addition, hydrogel stiffness can activate Notch1 signaling and regulate the cardiac progenitor cell gene expression and function in vivo. ²¹² An increased hydrogel concentration in the presence of the Notch ligand Jag1, in turn, amplified Notch1 activation, indicating that stiffness can affect Notch1 activation in 3D. ²¹² Furthermore, higher concentration hydrogels increased Hes1 expression and its downstream target Jag1, which is a ligand that could sustain Notch1 activation. Interestingly culture on lower stiffness hydrogels resulted in increased endothelial and smooth muscle gene expression, controlled through Notch1 activation. ²¹² This suggests that substrate stiffness can control Notch signaling cell fate decisions.

In addition, the effect of substrate stiffness has been implied within the MSC niche. A biomaterial system has mimicked this niche by tailoring PEG hydrogels to best promote Notch directed MSC behavior. In combination with Notch ligand immobilization, the hydrogel properties were also tailored to various stiffnesses. Material stiffness triggered the formation of microcapillaries in soft PEG matrices (1–1.3% PEG with 0.074 − 0.276 kPa), but as the hydrogel stiffness increased (3% PEG, 2.157 kPa), networks were almost completely absent, ¹⁷⁴ indicating that hydrogel stiffness plays a role in promoting Notch directed cell behavior. Reciprocal interactions of cells within the biomaterial led to a dynamic reshaping of the initial microenvironment, which further allowed development and ECM deposition. Finally, material surface coatings, including layer-by-layer deposition, have been used to tune the stiffness and promote cell contacts. A clay/PDDA film was used to provide a stiff/rigid layer on the hydrogel matrix, followed by a DLL1 layer. ²¹³

In addition to matrix stiffness, various topographical cues created through pore size manipulation, surface treatments, and various scaffold fabrication methods (electrospinning, bioprinting, freeze-drying, solvent leaching, etc.) also influence Notch signaling (Fig. 4A-ii, iii). This is important because surface roughness and patterning can affect the way that cells adhere and orient to the surface. A polycaprolactone/hydroxyapatite (PCL/HA) material compared to PCL alone displayed increased surface roughness, which promoted osteogenic differentiation in periodontal defects directed by immobilized Jag1. ²⁰⁴ Although PCL/HA membranes displayed lower mechanical properties, the surface roughness improved cell attachment and spreading, the nanostructured/hydrophobic surface also increased osteogenic differentiation compared to PCL surfaces alone. To create the thymic stroma, a 3D inverted colloidal crystal (ICC) scaffold coated with DLL1 was fabricated. ²¹³ The ICC topography was tested in a rotary dynamic cell culture system, where convective media flow was the major driving force for cellular motility within the scaffold. It was also determined that the ICC combined with the layer-by-layer surface modification promoted cell adhesion.

To explore topography further, SAMs and supramolecular motifs have been investigated. Surface-modified gold surfaces utilizing SAMs have been explored for Jag1 immobilization, where they were used to create patterns on the surface that encouraged cell adhesion and orientation. ¹⁸⁸ In addition, supramolecular UPy-poly(ɛ-caprolactone) (UPy-PCL) modified with a UPy-conjugated Tz (UPy-Tz) additive was also studied using drop cast and electrospun films. ¹⁷⁰ Primary human VSMCs endogenously expressing the Notch3 receptor were cultured on UPy-PCL/UPy-Tz electrospun meshes modified with TCO-pG/Fc-Jag1 and found upregulated the expression of Hey1, Hes1, Jag, and Notch3 compared to drop cast films. The larger surface area and potential for higher ligand concentration of the surface provided by electrospun mats were superior to the drop cast films and were able to affect cell adhesion, viability, and Notch functionality. ¹⁷⁰

Topographical and intrinsic material cues could be important in Notch signaling because of a recent report that cell–cell contact area can affect Notch signaling and Notch-dependent patterning. ²²² Cell fate decisions where larger cell–cell contact communication resulted in a higher expression signal and opposingly smaller cell–cell contact created less signal, which is highly dependent on scaffolding conditions. ²²² The various biomaterial surfaces also created a better environment for ligand coating efficiency by forming higher electrostatic charges. ²¹³ Stiffness could also play a factor in Notch signaling cell fate decisions based on the hypothesis that a ligand-mediated pulling force must be exerted on the NECD for Notch activation. ¹³ Thus, controlling substrate stiffness can control the traction forces and, therefore, Notch activation.

Spatial cues to control Notch ligands

Spatial patterning allows researchers to explore the effects of clustering and density on Notch activation ²⁰⁸ and also precise control of stem cell fate. Biologically, clustering and ligand density are driven dynamically within the cell membrane, and it is known that ligands form clusters and dimers, which are needed as a structural requirement for ligand stabilization.^190–192 Extending this to engineered systems, the notion of clustering and oligomerization using spatial patterning is suggested in controlling cell fate decisions.^{171,203,208,209} Spatial patterning and gradients of various ligands and soluble factors extend past this and are highly important in the development of complex multicellular organisms to emulate tissues. Cellular patterning is important in biological tissues, including retinal patterning and sprouting angiogenesis. The patterning complexity of various ligands, ECM molecules, growth factors, and other factors to control and regulate the cellular niche is important when designing engineered tissue. Multimolecular patterning has thus started to emerge in the biomaterials field to direct cellular behavior and has been done with a variety of molecules, including cell-adhesive RGD peptide, ²²³ fibronectin, ²²⁴ albumin, ²²⁵ and others as summarized previously. ²²⁶

Up to this point, we discussed the activation of Notch signaling using engineered biomaterials, where immobilization of ligands was homogeneously coated on a biomaterial surface. Since Notch signaling is contact dependent and operates through both lateral activation and inhibition, the patterning of specific ligands influencing the sender/receiver state of cells could be a beneficial technique to control cell fate decisions. Therefore, spatial patterning could include the use of ligand density, ligand clustering, or specific patterning to control Notch-directed cell fate decisions (illustrated in Fig. 4Bi–iii). Patterning of biomaterial surfaces can be integrated during both fabrication (i.e., stereolithography) or postprocessing of a uniform biomaterial using techniques, including microcontact printing. ²²⁶

In the vasculature, as discussed previously, Notch ligands Dll4 and Jag1 have opposing effects on angiogenesis. ⁵³ VEGF signaling induces Dll4 in tip cells; tip cells then suppress tip–cell features in adjacent stalk cells through Dll4/Notch-mediated lateral inhibition. ³⁴ Conversely, Jag1 antagonizes Dll4-mediated Notch activation in stalk cells to increase tip–cell number and enhances vessel sprouting. ⁴⁴ It is this Jag1 and Dll4 salt and pepper pattern that dictates the phenotypes of cells within this niche into tip–stalk phenotypes. Using microcontact printing, patterning of Dll4 ligand immobilized to pink fluorescence beads in parallel lines controlled angiogenic sprouting. ²⁰¹ Regulation of the location and direction of endothelial sprouting was achieved using micropatterns to locally inhibit angiogenic sprouting and direct the formation and location of new sprouts using the Dll4 negative regulator.

Similarly, microarray patterning in circular patterns of ECM copresented with Notch ligands provided a biochemically and biophysically defined microenvironment for liver progenitor differentiation. Patterning of Notch ligand Dll1, Dll4, and Jag1 showed that cells were spatially localized based on the patterning with segregated differentiation of progenitors toward biliary fates peripherally and hepatocytic fates centrally. ²²¹ Spatial presentation of ligands on biomaterial surfaces consequently has promise in manipulating and controlling cell fate decisions for more controlled regenerative medicine applications. Future directions include exploring various other ligand–receptor patterns, as well as various combinations of ligands.

Temporal cues to control Notch signaling

To further expand the capabilities of Notch signaling biomaterials, temporal control of ligand delivery should be evaluated. Delivery of signals can be controlled temporally by modulated frequency, dose dependence, duration of cultures in a Notch activated state, as well as temporal control of cell life cycle. Temporal frequency in a pulsatile or sustained delivery could affect the cell fate of myoblast cells. ²¹ Single-cell imaging revealed that Dll1 activated Notch1 in discrete frequency-modulated pulses that specifically upregulate Notch target gene Hes1, promoting myogenesis. In contrast, Dll4 had sustained activation dynamics activating target gene Hey1, inhibiting myogenesis. It has also been indicated that Hes1 activation is relatively insensible to the duration of Notch activation and could be induced strongly by brief pulses or by sustained activation; however, the activation of Hey1 and HeyL was more sensitive to the duration and accumulation as long as Notch activation was maintained (illustrated in Fig. 4C-i). This indicates that temporal delivery of Notch ligands in cellular therapies could be used to selectively activate Hes1 versus Hey1/L activation, which would result in different cell fate determinations. Interestingly increased dose of Dll1 increased the pulse frequency, whereas an increased Dll4 dose caused an increased amplitude by still a steady-state response. Dll1 signaling was deemed as frequency modulated, and Dll4 was amplitude modulated. Although interesting to consider when evaluating the effectiveness of dynamic Notch control using engineered biomaterials, it is still unclear how this translates to other Notch ligand–receptor relationships.

Not only do various tissues respond differently to the ligand identity but additionally the stimulation of cells within a niche at different times in their cellular specification can lead to different cell fates. Biphasic effects of Notch activation have been demonstrated in cardiac differentiation. Jag1 stimulation and early activation of undifferentiated hESCs promoted ectodermal differentiation and, in contrast, Notch activation in specified cardiovascular progenitor cells increased cardiac differentiation in these cells. ¹⁸⁹ Within the pancreatic field of pluripotent derived cell niche, the duration of time cells spent in a Notch active state was also indicated to affect lineage specification. ²²⁷ Together, this may indicate a link between temporal delivery and dynamics of Notch signaling in cell fate lineage specification.

From a biomaterial perspective, Jag1 modified 3D microenvironments have induced target genes in MSCs that were found upregulated in cocultures with ECs, indicating that biofunctional biomaterials can be used to reprogram cells in the absence of ECs. There is some indication that biomaterial driven control of Notch signaling can induce a Jag1 mediated ECM switch in MSCs that was reversible when exposed to control IgG-modified microenvironments. Temporal delivery of Jag1 for 7 days followed by a subsequent 7-day exposure to igG microenvironments reverted the BM-MSC ECM from an induced perivascular phenotype back toward a vascular basement membrane phenotype ¹⁷⁴ (illustrated in Fig. 4C-ii). This indicates that Notch signaling biomaterials can be involved in reversibly reprogramming cellular environments, and signaling dynamics are involved in various cell fate switches in the body. When the ligand–receptor dynamics is controlled, this results in proper homeostasis; however, when dysregulated ligand–receptor dynamics can cause diseased states. Temporal delivery of ligands can thus be incorporated into the design of Notch biomaterial cellular therapies. Further research should thus be done to investigate the behavior and controlled activation of Notch signaling utilizing dynamic encoding with biomaterials.

Cellular mechanosensing by bioreactor conditions

Notch signaling has been shown to regulate its behavior by a mechanical stimulus in the cellular microenvironment. Various bioreactor conditions can be controlled and maintained to mature and direct cellular response in engineered tissue (Fig. 4D-i). Mechanical loading of bone primarily through strain forces is sensed by bone cells resulting in biochemical signals that result in the activation of mechanosensors. In vitro, Notch signaling has been activated by mechanical strain loading through cyclic stretching, ²³⁰ intermittent compressive stresses,^231,232 and pulsed electromagnetic fields. ²³³ This indicates a functional link between Notch activation and mechanotransduction caused by mechanical or electromagnetic loading, which could be utilized for in vitro Notch activation.

In addition, within the vasculature Jag1 and Notch3 target genes were downregulated with strain magnitude defined by the vascular geometry and blood pressure upon increased or decreased wall thickening caused by vascular morphogenesis. ⁴⁶ Notch signaling has also been shown to be modulated through shear stress caused by blood flow through the body due to cycles of contraction and relaxation of heart tissue. Endothelial Notch1 is responsive to shear stresses and is necessary for the maintenance of junctional integrity, cell elongation, and suppression of proliferation phenotypes induced by laminar shear stress. ²³⁴ It has been shown that both angiogenesis and osteogenesis are defective when blood flow is impaired in vessels, which is coincident with downregulated Notch signaling in ECs. ²³⁰ Notch signaling also controls the expression of fluid flow responsive genes in ECs and modulates the formation of fluid flow sensing primary cilia. ²³⁴ Mechanotransduction forces help maintain homeostasis of various tissues as discussed; however similar to how ligand–receptor dysregulation can occur, dysregulation can also be driven by mechanotransduction forces. Evidently, mechanotransduction and Notch signaling have been linked in tumor progression and disease. ²²⁸ Together these findings substantiate that Notch signaling is influenced by its mechanical environment and is a mechanosensing signaling pathway. Incorporating mechanical signals to cell-seeded scaffolds using bioreactors capable of exerting compressive or shear forces may accelerate our understanding of Notch dynamics in an in vitro system.

Molecular level mechanosensing

There is also evidence suggesting that Notch signaling driven by cell–cell interactions are mechanosensitive at the molecular level (Fig. 4D-ii). Crystal structures have revealed the overall Notch receptor–ligand conformation indicating that the S2 binding site is deeply embedded within the Notch heterodimer Lin12-Notch repeats domain and, thus, is protected from metalloprotease cleavage and creates an autoinhibited conformation.^152,235 The Notch pulling model indicates that the Notch regulatory region of the receptor acts as a force sensor that is unfolded by a threshold level of mechanical tension generated across the ligand/receptor bridge. This tensional force in a cell–cell signaling system is proposed to be caused by the endocytosis force at the ligand–receptor complex ¹³¹ ; however, this is highly debated. ²³⁶ Nonetheless, in the absence of endocytosis, Notch ligands were shown to accumulate on the cell surface and failed to activate Notch signaling on neighboring cells. ¹⁰

Various molecular force sensors have been developed to study cellular forces acting through single mechanosensitive receptors. Force quantification techniques for cell biology include 2D traction microscopy, micropillars, cantilevers, atomic force microscopy, fluorescence resonance energy transfer, optical tweezers, magnetic tweezers, and tension gauge tethers (TGTs). ²³⁷ TGTs are force sensors which have been widely used to study Notch signaling. Protein G based TGTs were used to study the force magnitude needed for signal activation. Recombinant ligands with IgG-Fc fusion are assembled using DNA tethers with different tension tolerances immobilized through a glass surface passivated with PEG. Notch ligand Dll1 was tethered to the surface, and using a reporter cell line, it was determined that under 12pN of force was required for signal activation.^238,239 Force-induced Notch activation has also been studied at the ligand–receptor bridge using Notch ligand immobilized to magnetic biomaterial nanoparticles. Magnets have been utilized using a magnetic tweezer assay to apply a range of pN-scale forces to the Notch receptor on the cell surface, and Dll4-loaded magnetic beads induced Notch signal activation with the addition of this mechanical tension force. ²⁴⁰ This further suggests that force must be applied to bead-tethered ligands to further induce the canonical proteolytic steps responsible for Notch activation. It also indicates that the Notch signaling system is controlled through the interplay of both biochemical and biomechanical signals.

The role of mechanotransduction within the Notch signaling pathway is still controversial and a highly debated topic of research currently. Considering these concepts, the interplay between mechanical force transduction and its interaction with ligand immobilized to the surface of biomaterials poses another important challenge in the integration of these biomaterials into complex cellular processes, including the mechanical conditions and molecular mechanotransduction. Mechanotransduction within Notch signaling thus far has been very context dependent; therefore, various ligand–receptor combinations could prove to be more mechanosensitive than others. In addition, the delivery of chemical signals through a biomaterial surface presents the absence of the mechanical tension force found biologically through the pulling force caused by cell endocytosis of the ligand–receptor complex. Therefore, methods must be investigated in how to provide accurate traction and tension forces within biomaterial-based Notch signaling systems to allow for Notch activation to occur. The application of mechanotransduction to the activation of Notch presenting biomaterials could be a possible target to enhance signaling and repair of various processes.