**1. Introduction**

The extracellular matrix (ECM) is the noncellular component presenting within all tissues and organs, and it usually has three-dimensional porous architecture. It provides not only essential physical sca ffolding for the cellular constituents but also initiates crucial processes that are required for tissue morphogenesis, di fferentiation, and homeostasis [1]. Each tissue has its own ECM with a unique composition and topology, generated during tissue development [2]. Cells interact with biochemical and biophysical cues in their ECM in highly dynamic and reciprocal manners, and such cell–ECM interactions play a critical role in cell behavior mediation, cell function normality and even cell fate decisions, involving quite complicated processes from quiescence to activation and progenitor state to terminal di fferentiation [1,2]. Accordingly, it is pivotal to harness the interactions between ECM and the resident cells in developing strategies for e ffectively regenerating tissue or regulating disease [1,2]. To data, many kinds of hard biomaterials have been employed in tissue engineering for various purposes, but their applications in ECM biomimicry have been limited because these hard biomaterials are lack of ability to adequately mimic the structure and properties of ECM in body tissues [3]. Polymer hydrogels, which behave like soft and elastic objects, are usually constructed by physically or chemically crosslinked macromolecules. They contain large amounts of water while having highly porous architecture with tailorable physiochemical properties and easy di ffusivity of small molecules. These features make them attractive candidates for mimicking the dynamic ECM [3,4].

Injectable polymer hydrogels having in situ gelling properties under physiological conditions have received a grea<sup>t</sup> deal of attention in tissue engineering owing to their two advantageous characteristics [5]. One is that they can be conveniently used to deliver cells, bioactive compounds, and therapeutic drugs alone or in combination by simply mixing these consignments with polymer solutions prior to gelation; and another is that they can be injected to the defect site via minimally invasive surgery followed by formation into solid-like fillers with discretional shapes [5,6]. Nowadays, varied kinds of natural polymers have been commonly used in the form of hydrogels. Among them, alginate has been extensively investigated for its hydrogel applications. Physically crosslinked alginate hydrogels can be easily built by using certain divalent cations, typically calcium ions (Ca<sup>2</sup>+), as a crosslinker. Despite the convenience of preparation, so constructed alginate hydrogels often undergo progressive disintegration in vivo due to the ionic exchange between Ca2+ in the gels and monovalent cations (such as Na<sup>+</sup> and K<sup>+</sup>) coming from the host tissue surrounding the applied gels, which often results in their unstable dimension and uncontrolled properties [7,8]. Another type of physically crosslinked alginate hydrogel was engineered by grafting a type of thermosensitive polymer, poloxamer (Pluronic F127), onto alginate backbone, and the sol-gel transition of alginate–poloxamer (ALG–POL) hydrogels can be trigged by thermosensitive action arisen from the poloxamer component [9]. Despite easy-handled and safe advantages, ALG–POL gels appear to be weak and brittle in nature, and are apt to disintegrate due to its high percentage of Pluronic F127 [10]. Thus, ALG–POL gels are incompetent for certain applications in cartilage or bone tissue engineering where su fficient strength and persistent dimension stability are concomitantly required.

Many studies have revealed that hydrogels with dual or multiple networks could be largely enhanced in their stability and mechanical performance when compared to the single network gel [11]. In spite of the mentioned advantages, multiple network gels are not all suitable for tissue engineering applications as they are usually fabricated via chemical crosslinking, and the involved crosslinking reactions could possibly impair the loaded cells or the host tissue surrounding the applied gel [5,12]. Silk fibroin (SF) is a kind of natural fibrous protein and it can be processed into hydrogels via enzyme-catalyzed crosslinking of amino acid phenolic groups by the aid of H2O2, and the obtained SF hydrogels show tunable strength and elasticity [13]. In the case of in vivo usage, the applied amount of H2O2 for crosslinking SF gels has to be controlled at a safe level since the resulting SF gels could be cytotoxic if the applied dose of H2O2 is higher than certain thresholds [14]. As a result, so prepared SF gels were usually weak [15]. Taking into account the gelable characteristics of ALG–POL and SF through their respectively independent gelling mechanisms, it is feasible to construct a new type of ALG–POL/SF gel with dual network structure while having enhanced performance.

Some growth factors have been proved to be highly e ffective for promoting bone repair, especially taking advantage of the controlled factor release by way of proper carriers [16]. Among various kinds of growth factors, insulin-like growth factor-1 (IGF-1) is considered to be crucial for longitudinal bone growth, skeletal maturation, and bone mass acquisition not only in the bone growth phase of young individuals but also in the maintenance of bone in adult life [17]. In the situation of bone repair, in addition to promoting cell proliferation and matrix synthesis, the applied IGF-1 at the defect site can also induce the chemotactic migration of osteoblasts to the repair site via local concentration gradients established by factor di ffusion [18]. Like many other growth factors, IGF-1 has a short half-life when exposed to the circulatory system [19]. Therefore, when the need arises to maintain sustained release of IGF-1 at the local site in vivo, one of the practical strategies for its administration is to encapsulate IGF-1 into certain vehicles to preserve its activity while managing to modulate its dose and action duration. It is generally realized that directly encapsulating growth factors into hydrogels would result in their burst release because hydrogels commonly have rather porous structures with high water content [7,20]. It has been suggested that release kinetics of growth factors delivered by a hydrogel system could be mediated to varied degrees by encapsulating the growth factor into certain microcarriers such as microspheres (MPs) and nanoparticles (NPs) first, and then, embedding the factor-loaded microcarrier into the hydrogel system [20–22].

Bioactive glasses (BGs) have now been widely used as an attractive inorganic biomaterial for bone repair since they have the ability to strongly bond bone tissue via a hydroxycarbonate apatite interface layer with composition and function similar to naturally occurring bone hydroxyapatite [23]. Mesoporous BG microspheres or nanoparticles can also serve as a reservoir for delivering therapeutic drugs or bioactive molecules besides their functions for acting as bone repair material [24]. In addition to regular BG MPs or NPs, many of them have been doped with di fferent kinds of compounds, and their ionic dissolution products are capable of inducing osteogenesis or angiogenesis at the bone defect site, depending on the variety of doped elements [25]. In this context, it would be rational to load IGF-1 into porous BG NPs, and then, to incorporate the IGF-1-loaded BG NPs into above mentioned dual network ALG–POL/SF gels. On the basis of such designed strategy, a multifunctional composite hydrogel system with mechanically strong features and capabilities for administering the release of IGF-1 could be obtained. In this study, an attempt was made to achieve this goal. Some formulated IGF-loaded BG/ALG–POL/SF gels were found to be injectable and mechanically strong, and to have affirmative abilities to control the release of IGF-1 while preserving its bioactivity.
