*Review* **Delivery of Natural Agents by Means of Mesoporous Silica Nanospheres as a Promising Anticancer Strategy**

**Khaled AbouAitah 1,2,\* and Witold Lojkowski 1,\***


**Abstract:** Natural prodrugs derived from different natural origins (e.g., medicinal plants, microbes, animals) have a long history in traditional medicine. They exhibit a broad range of pharmacological activities, including anticancer effects in vitro and in vivo. They have potential as safe, cost-effective treatments with few side effects, but are lacking in solubility, bioavailability, specific targeting and have short half-lives. These are barriers to clinical application. Nanomedicine has the potential to offer solutions to circumvent these limitations and allow the use of natural pro-drugs in cancer therapy. Mesoporous silica nanoparticles (MSNs) of various morphology have attracted considerable attention in the search for targeted drug delivery systems. MSNs are characterized by chemical stability, easy synthesis and functionalization, large surface area, tunable pore sizes and volumes, good biocompatibility, controlled drug release under different conditions, and high drug-loading capacity, enabling multifunctional purposes. In vivo pre-clinical evaluations, a significant majority of results indicate the safety profile of MSNs if they are synthesized in an optimized way. Here, we present an overview of synthesis methods, possible surface functionalization, cellular uptake, biodistribution, toxicity, loading strategies, delivery designs with controlled release, and cancer targeting and discuss the future of anticancer nanotechnology-based natural prodrug delivery systems.

**Keywords:** mesoporous silica nanoparticles; controlled release; drug delivery systems; anticancer natural prodrugs; natural products; cancer targeting; nanoformulations/nanomedicine applications

#### **1. Introduction**

In 2001, Vallet-Regi et al. [1] introduced a mesoporous silica material called MCM-41 that can be used as a drug carrier. The nanostructure (e.g., pore size) of MCM-41 can be optimized using different surfactants. Since then, many efforts and attempts have been made to synthesize versatile mesoporous silica nanoparticles (MSNs) with different nanostructures and morphologies to meet the demand for pharmaceutical and medical applications. The history of the synthesis of mesoporous silica materials dates back to 1992, when they were discovered by the Mobile Oil Corporation [2]. Silica is one of the most abundant minerals in the Earth's crust and is also found in the food chain and the human body [3]. As a biomaterial, silica is extensively used in many applications such as dentistry, orthopedics, and dermatology. MSNs have a characteristic mesoporous nanostructure that offers many advantages for medical applications in disease diagnosis and therapy [4]. The unique features include easy synthesis, the possibility of various surface modifications, the ability to obtain a tunable particle size, uniform pore size, high surface area to pore volume, good biocompatibility, and chemical stability [5–9]. In addition, easy functionalization to achieve magnetic, fluorescent, and photothermal properties increases the chance of using MSNs in bioimaging. MSN nanostructures can provide excellent nanoplatforms to fabricate smart drug delivery systems (DDSs) with a high drug loading capacity and

**Citation:** AbouAitah, K.; Lojkowski, W. Delivery of Natural Agents by Means of Mesoporous Silica Nanospheres as a Promising Anticancer Strategy. *Pharmaceutics* **2021**, *13*, 143. https://doi.org/ 10.3390/pharmaceutics13020143

Academic Editors: Hassan Bousbaa and Stefano Giovagnoli Received: 30 November 2020 Accepted: 14 January 2021 Published: 22 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

stimuli-responsive drug release effect compared to other nanocarriers [6,10]. Several nanocarriers have been used to deliver and control drug release, including niosomes, liposomes, dendrimers, lipid nanoparticles, and polymeric nanoparticles, but most of them have low stability and need external stabilization during synthesis. In contrast, MSNs have a strong Si-O bond that makes them stable (chemically and mechanically) to external responses in the surrounding environment [11–13]. It is generally accepted that encapsulation of drugs or therapeutic agents into MSNs can enhance their therapeutic activity, solubility, and bioavailability, as indicated by many studies [14–20].

A consequence of these advantages is that MSNs have gained much attention and popularity in DDSs during the last few decades for the delivery of cargo to specific sites in the organism. A large number of in vivo studies indicate the high biocompatibility/safety profile and low toxicity of MSNs if they are synthesized using an optimized way [21–23]. A careful optimization process is needed because many details of the nanostructure of engineered MSNs, i.e., size, shape, surface, presence of surfactant, and other factors like dose, administration route affect the safety profile. According to many animal studies, the toxicity of MSNs can be diminished by optimizing the synthesis parameters and surface modification, resulting in safe nanoparticles [24,25].

The administration route is an important characteristic for constructing any DDS. MSNs can be applied via different routes, including oral and intravenous injection [26–30]. Many choices in the development of pharmaceutical formulations depend on the target tissues and organs in the human body. An important advantage of DDS-based MSNs is that the amorphous forms of silica and silicates are generally recognized as safe materials for use as oral delivery ingredients (up to 1500 mg per day) according to the US Food and Drug Administration and the European Food Safety Authority [27]. MSNs are promising materials because they exhibit low toxicity levels in animals when applied, i.e., orally, injection [31].

The global market for nanomedicine accounts for 5% when novel nanomedicines translated from the lab to the clinics are concerned [32]. Recently, the first clinical trial in humans was conducted with oral delivery of fenofibrate formulation based on the ordered mesoporous silica [33].

Despite these promising results for nanotechnology application in building DDSs, most research for targeted cancer therapy has been focused on drugs and therapeutic molecules of a synthetic nature. Combating cancers with synthetic drugs is an established therapy, however, progress in this area of medicine is slow and the treatments are frequently associated with undesirable effects: side effects and also insufficient patient compliance. For this reason, extensive research is carried out to apply natural prodrugs (known also as natural products and natural agents) in anticancer therapies.

Nature is a huge source of therapeutic substances, which can be derived from plants, microbes, and animals. Natural medicines account for 60% of anticancer agents used in clinical applications [34]. For example, vincristine, taxanes, and camptothecin are used in the treatment and prevention of cancer. There are still hundreds of promising new active natural anticancer agents to be discovered and renewed for cancer therapy [35–37]. The main advantages to using and developing natural prodrugs are that they offer safe, cost-effective, and have versatile pharmacological properties [38]. The main limitations for their use in cancer therapy are their poor water solubility, low bioavailability, short half-life, and non-specific targeting.

Nanotechnology offers many ways to overcome these obstacles [39–44]. Natural pro-drugs can be embedded into MSNs, which can serve as effective nanocarriers for the delivery of anticancer natural prodrugs to target cancers. In this review, we present an overview of synthesis methods, surface functionalization, as well as biodistribution, biocompatibility, toxicity, biological performance. Additionally, drug loading and release strategies, and active targeting approaches for MSNs will be addressed. We also discuss delivery and controlled release systems for selected prodrugs using MSNs.

Available data provide considerable evidence that MSNs allow the limitations associated with prodrugs, such as poor water solubility, poor bioavailability, and low specific targeting ability, to be overcome. Compared to organic delivery systems (e.g., lipid nanoparticles, polymeric nanoparticles) [45,46], the delivery of natural prodrugs by means of MSNs allows high drug loading and permits multifunctional delivery or co-delivery systems. Generally, MSNs allow long-term release compared to organic nanoparticles. This is because the prodrugs are trapped inside nano-pores. In the case of encapsulation of prodrugs into organic nanoparticles, fast degradation of the organic substance leads to quick prodrug release. The MSN-based nanomedicine technology is mature enough to be extended to thousands of prodrugs not yet investigated in clinical applications.

To the best of our knowledge, this is the first review considering MSNs as delivery systems for anticancer natural prodrugs. The need for such a review is a consequence of rapid development in the field. This review may help researchers accelerate research and development of this important field of nanomedicine and, ultimately, clinical applications.

#### **2. Synthesis of Mesoporous Silica Nanostructures**

Numerous synthesis methods have been developed to obtain MSNs with different morphological, structural, and pore geometry. Particular attention was paid to the production of biocompatible MSNs for medicine. Figure 1 presents the number of scientific publications (research articles, review articles, and book chapters) as an indicator of the growth in MSN synthesis methods due to their emergence as nanostructures for various promising applications.

**Figure 1.** Number of scientific publications (research papers, reviews, book chapters) during the period 2001–2020 found by entering key words "mesoporous silica nanoparticles and synthesis". The search was performed in ScienceDirect 10 September 2020.

#### *2.1. Discovery, Synthesis, and Properties of MSNs*

Porous materials (natural or artificial) are characterized by the presence of pores, including cavities, channels, or interstices. The properties of these materials vary depending on the characteristics of their pores: size, arrangement/structure, shape, porosity, and chemical composition. They have been extensively studied in different areas, including water purification, gas separation, catalysts, energy storage, adsorbents, electronics, engineering, tissue engineering, and drug delivery systems, among others [47]. Depending on the predominant pore size, the International Union of Pure and Applied Chemistry (IUPAC) classifies porous materials into three categories as shown in Table 1 [48,49].


**Table 1.** Classification of porous materials by pore size.

The history of MSN materials dates back to the early 1990s, when the Kuroda group at Waseda University and researchers from the Mobil Company discovered Mobil crystalline materials (MCMs), nanoparticles with a hexagonal porous structure [2]. In 1992 with the discovery of MCM-41, a material prepared using the cooperative assembly of surfactant with silicates, a breakthrough in the area of ordered mesoporous structures and their successful preparation occurred [50,51]. In addition, an ionic template, such as cetyltrimethylammonium bromide (CTAB), could be employed as a structure-directing agent to produce MCM-41 and MCM-48 with pore sizes of 2 to 10 nm [50,51]. MCM-41 has a hexagonal pore shape and MCM-84 has a cubic pore shape. For DDSs purpose, MCM-41 is considered to be one of the most widely explored materials. The synthesis mechanism for MCM-41 is shown in Figure 2 and electron microscope images in Figure 3.

**Figure 2.** The formation mechanism for mesoporous materials by structure-directing agents. (**a**) True liquid–crystal template mechanism. (**b**) Cooperative liquid–crystal template mechanism. Reproduced with permission from [52], WILEY-VCH Verlag GmbH and Co. KGaA, 2006.

**Figure 3.** (**A**) Scanning electron microscopy (SEM) and (**B**) transmission electron microscopy (TEM) of MCM-41 material. Reproduced with permission from [23], Wiley-VCH Verlag GmbH and Co. KGaA, 2010.

In 1996, another kind of MSN was discovered that has a non-ordered pore structure, named KIT-1 (Korea Advanced Institute of Science and Technology Number 1) [53]. The KIT family currently has many members, such as KIT-6, which has a hexagonal arrangement of pores [54], and KIT-5, which has a cubic ordered structure [55]. In 1998, the SBA-15 type (pore size 4–6 nm) MSNs introduced by Santa Barbara Amorphous (SBA), which have a hexagonal or cubic pore structure, were developed by means of nonionic surfactants in acidic conditions [56]. The cubic SBA-11, 3D hexagonal SBA-12, hexagonal SBA-15, and SBA-16 are mainly prepared based on non-ionic triblock copolymers, such as alkyl poly(ethylene oxide) (PEO) oligomeric surfactants and poly(alkylene oxide) block copolymers [10]. The typical synthesis of SBA-15 is dependent on tetramethyl-orthosilicate (TMOS) or tetraethyl-orthosilicate (TEOS) as the silica precursor reacting with a series of block-copolymer surfactants as structure-directing agents. The MCM and SBA materials are recognized as the first generation of hexagonally ordered pore structures and are the common MSNs used in research. A variety of strategies have been designed to attain tunable pore sizes (from less than 2 nm up to 30 nm). In this scenario, the adjustments are made depending on the surfactant template's properties [57], pore swelling agents, such as mesitylene [50], or hydrothermal treatments [58].

Importantly, in 2010, high surface-area silica nanospheres with a fibrous morphology and non-ordered pore structure were discovered by a research group of the Catalysis Center at King Abdullah University of Science and Technology (KAUST Catalysis Center, KCC) [59]. This material, KCC-1, features a high surface area due to the presence of dendrimeric silica fibers and their respective channels, making KCC-1 a first-of-its-kind material. It is a spherical particle with 3D tomography, a uniform size ranging from 250 nm to 500 nm, high surface area, and large pore size in a non-ordered structure (Figure 4). Synthesis of KCC-1 [59] was accomplished by a microwave-assisted, templated, solvothermal strategy using cetylpyridinium bromide (CPB) or cetyltrimethylammonium bromide (CTAB) as a surfactant (template), 1-pentanol as a co-surfactant, TEOS as the silica source, urea (catalyst-hydrolyzing agent), and a mixture of the cyclohexane solvent and water (as the reaction solvent). The chemicals were introduced to the reaction system stepwise with mixing and microwave-assisted heating applied (in a closed vessel >1200 ◦C) for a predetermined time for the reaction. Finally, the solution was filtered or centrifuged, washed, and the obtained material calcinated at high temperature (>550 ◦C). Many research groups changed the surface of substances used in the synthesis in addition to the parameters. For example, Bayal et al. [60] showed that changing the concentrations of urea, surfactant (CTAB instead of CPB), or solvent (1-pentanol), the reaction time, or temperature can result in various particle sizes, fiber densities, surface areas, and pore volumes for KCC-1. Such easy manipulation and controlled synthesis of this material make KCC-1 a good solution for versatile applications in the environment, energy, biology, medicine, and other fields [42,43,61–68]. KCC-1 could be recommended for different small or large drug/therapeutic agents, possibly for any design and pathological disorder due to KCC-1 s unique physicochemical features. Our research team is among the first to study KCC-1 for DDSs [42,43,68,69], and we think that research on KCC-1 will increase soon. In the literature, there are references to "spherical wrinkled mesoporous silica" (WMS) [70–72] and KCC-1 is known also "dendritic fibrous nano-silica" (DFNS) [73]. They were all obtained based on changing the synthesis conditions and parameters of the original synthesis method for KCC-1 particles.

Unlimited opportunities exist for the synthesis of MSNs in pure, doped, composite, and modified forms by employing different templates (soft and hard), conditions, and methods [74].

**Figure 4.** Electron microscope images of prepared KCC-1 material. (**A**,**B**) Scanning electron microscopy (SEM). (**C**,**D**) Transmission electron microscopy (TEM). Note, the dendritic fibrous 3D mesopore structure is clearly seen by SEM in B. A and C reproduced from [42,43], Impact Journals, 2018 and MDPI, 2020. B and D reproduced with permission from [59], WILEY-VCH Verlag GmbH and Co. KGaA, 2010.

Due to the unique properties of the KCC-1 family, they offer a wide range of possible applications. It seems that KCC-1 has comparable potential as the commonly used members of the MCM and SBA families, as well as Stober silica, solid silica discovered before all the families [73]. Table 2 presents the major physicochemical properties for fibrous KCC-1, MCM-41, SBA-15, and others. Below, we highlight the common and promising families that could be favored for drug delivery and medical applications. Numerous interesting review articles have been published on MSN synthesis strategies and applications that we recommend for further reading [10,22,32,73,75–83].


**Table 2.** The physicochemical properties of the most common mesoporous silica nanoparticles (MSNs) synthesized by various approaches.

Note: The above-mentioned characteristics of these materials can be controlled and can vary (more or less) from these values. Reproduced with permission from [75], Wiley-VCH Verlag GmbH and Co. KGaA, 2017. NA = not available.

#### *2.2. Surface Modification of MSNs for Drug Delivery*

The keystone in the development of DDSs is to functionalize their surface [84,85] to increase their drug loading and release, leading to high therapeutic effects. The surface chemistry modulates the interaction of MSNs with the surrounding media. The MSNs have a high density of silanol groups (Si-OH) on their surface, allowing surface modification by various organic functionalities (e.g., silanes, polymers, proteins, and targeting moieties). Thus, MSNs can load various drugs with high capacity and release them in a sustained or controlled manner. A variety of functional groups can be used, such as amine, carboxylate, phosphonate, polyethylene glycol, octadecyl, thiol, carboxylic acid, and octadecyl groups. To introduce functional groups on the surface of MSNs, covalent bonding and electrostatic interactions are generally used [86]. The common approach to modify MSNs is to use organic silane groups via direct covalent attachment by means of co-condensation or post-synthetic grafting.

The co-condensation method is referred to as a one-pot synthesis method [87,88] as presented in Figure 5A. The desired functional group of silanes, such as 3-aminopropyltriethoxysilane (APTES "NH2") is added during the sol-gel synthesis process together with the silica source (e.g., TEOS). Next, the template is removed (Figure 5A) [52,87,89]. To remove the surfactant template, an extractive method using alcoholic/acidic solution under reflux can be used [90]. Removing the template anchors the organic residue covalently to the porous walls of the MSNs. This approach has the advantages of easy preparation, more homogeneous distribution of organic units, and high drug loading [52,83]. Despite these advantages, disadvantages are a potential change in the mesoscopic order, disordering the porosity and reducing the pore diameter, pore volume, and specific surface areas [52].

A post-synthetic approach refers to the subsequent modification of the inner/outer surface of MSNs by covalent and electrostatic interactions. The modification is usually achieved after surfactant removal from MSNs (Figure 5B). The most remarkable advantages of this approach are selective functionalization (either external or internal surfaces) and retention of the mesostructure of MSNs during synthesis. The major disadvantages include reduced pore size and non-homogeneous distribution of functional groups into/onto pores [52,91,92].

**Figure 5.** A schematic presentation of the organic functionalization methods for mesoporous silica materials. (**A**) Cocondensation method and (**B**) grafting method. Reproduced with permission from [52], WILEY-VCH Verlag GmbH and Co. KGaA, 2006.

#### *2.3. The Biological Performance of MSNs*

#### 2.3.1. Cellular Uptake

Any nanocarriers have to cross the cell membrane boundary to enter cells, allowing the therapeutic effects of the delivered drugs. The internalization of nanoparticles carrying therapeutic agents into cells represents the initial step in successful drug delivery [93,94]. The acting mechanisms and surface chemistry of nanocarriers are the major parameters in designing a preferred DDS for any pathological disease [78]. Nanoparticles mainly access the cell interior via simple diffusion or translocation as an energy-dependent process [95]. The most common mechanism of their internalization is the energy-dependent endocytosis, which allows the uptake of nanoparticles and submicron particles from an extracellular environment to the cell plasma membrane [96]. The mechanisms can generally be classified into phagocytosis, pinocytosis, micropinocytosis, receptor-mediated endocytosis, clathrinmediated endocytosis, caveolin-mediated endocytosis, and others (e.g., Arf-6, Rho-A or IL2Rb-dependent pathway, flotillin, or CDC42 (CLIC/GEEC)-dependent endocytosis) [93]. The intracellular uptake and trafficking mechanisms by which nanoparticles are internalized in cells vary broadly depending on many factors, including size, shape, charge, and surface modification. Therefore, these factors should be taken into consideration in constructing DDSs.

#### Size of MSNs

Particle size determines the intracellular uptake of MSNs (Figure 6) [97]. It is generally accepted that particles with the smaller size of 50 nm can internalize into cells via non-phagocytosis [98]. Nanoparticles up to 150 and 200 nm in size are internalized by pinocytosis, such as clathrin-mediated endocytosis and caveolin-mediated endocytosis, respectively [99,100]. In contrast, particles from 250 nm to 3 µm in size can internalize the cells by macropinocytosis and phagocytosis [101]. It is also accepted that the microparticles are efficiently taken up through phagocytosis but the process depends also on other parameters, such as geometry, surface charges, and functional groups of microparticles [102]. Particles with sizes ranging from 30 to 50 nm internalize also efficiently via receptormediated endocytosis [103]. Despite extensive investigations exploring the relationship between particle size and uptake pathways, the results are inconsistent [101,104–106]. The main reason for such contradictions can be attributed to the complexity of control of structural parameters, such as shape and surface charges. For successful internalization, particles should avoid degradation (within endosomal/lysosomal vesicles) and release their cargo in the cytoplasm [107]. Therefore, particle size is important in tailoring DDSs. It is also important for their intersections with the reticulo-endothelial system (RES), which is responsible for elimination of nanoparticles from the body, and prolong the circulation time in the blood. In this context, several studies have shown that increasing the particle size increases clearance from the body, reducing the therapeutic impact [108–112].

**Figure 6.** Different endocytosis pathways across the intestinal villus for particles of different sizes. Reproduced with permission from [97], Elsevier Inc., 2020.

Lu et al. [103] investigated the impact of various sizes (30, 50, 110, 170 nm) of MSNs on cellular uptake by HeLa cancer cells using MSNs labeled with fluorescein isothiocyanate (FITC) green fluorescence (MSN-FITC) and confocal laser scanning microscopy. They found that the MSNs were internalized as non-uniform green-fluorescent aggregates in the perinuclear region, and no MSNs penetrated the nucleus (Figure 7). Quantifying the internalization of MSNs, they concluded that the cellular uptake is highly particle sizedependent, observing the order 50 > 30 > 110 > 280 > 170 nm (Figure 8). Haddick et al. [113] demonstrated that MSNs with a size of 160 nm had the fastest cellular internalization in

T24 bladder cancer cells through receptor-mediated cellular internalization compared to 60, 80, 100, and 130 nm, leading to the highest level of gene knock-down for antitumoral effects. Yang et al. [114] tested different sizes of rod-shaped SBA-15 (from 80 to 200 nm) and spherical MCM-41 particles, as well as their intracellular uptake in human osteosarcoma cancer cells (KHOS). They found that the cellular uptake efficiency depends on the particle size and shape.

**Figure 7.** Confocal laser microscopy images of HeLa cells after incubation with different sizes of MSNs labeled with fluorescein isothiocyanate (FITC) green fluorescence (MSN-FITC) (100 µg mL−<sup>1</sup> , green) for 5 h at 37 ◦C. (**A**) 170 nm, (**B**) 110 nm, (**C**) 50 nm, and (**D**) 30 nm. The cell skeleton was stained with rhodamine-phalloidin (red), and the cell nucleus with 40 ,6-diamidino-2-phenylindole (DAPI; blue). Reproduced with permission from [103], WILEY-VCH Verlag GmbH and Co. KGaA, 2009.

**Figure 8.** Cellular uptake of FITC-MSN-x based on nanoparticle size. Reproduced with permission from [103], WILEY-VCH Verlag GmbH and Co. KGaA, 2009.

#### Surface Charges of MSNs

Another critical factor influencing the cellular uptake of nanoparticles is the surface charge. MSNs are characterized by silanol groups permitting to add different functional groups, modifying their surface to be either cationic or anionic [115]. Most cells have a negatively charged cell membrane, enhancing the uptake of positively charged nanoparticles. Several studies have shown that positively charged nanoparticles internalize with higher uptake than neutral and negatively charged nanoparticles [116–119]. Furthermore, neutral nanoparticles usually have lower cellular uptake compared to negatively charged nanoparticles [98,120]. As a result of the internalization of nanoparticles by cells,

their interaction with the cell membrane can occur by means of gelation of membranes (with negatively charged nanoparticles) or fluidity of membranes (with positively charged nanoparticles) [121,122]. On the one hand, the positively charged nanoparticles mainly enter cells via micropinocytosis; on the other hand, the negatively charged nanoparticles always enter cells by clathrin- or caveolae-independent endocytosis [123].

Positively charged MSNs generally exhibit higher endocytosis efficiency compared to negatively charged MSNs due to the higher affinity for the negatively charged cell membranes. Jambhrunkar et al. [124] prepared MCM-41 with negative and positive charges for delivering curcumin. They found that the positively charged MCM-41-NH<sup>2</sup> had more efficient uptake in the human squamous cell carcinoma cell line (SCC25) than negatively charged particles. Baghirov et al. [125] studied spherical and rod-shaped MSNs that were either non-modified or modified with a poly(ethylene glycol)-poly(ethylene imine) (PEG-PEI) block copolymer in in vitro models of the blood–brain barrier. The results showed that the modified MSN-PEG-PEI particles exhibited robust uptake in RBE4 rat brain endothelial cells and Madin–Darby canine kidney epithelial cells. Our group performed a comprehensive study of cellular uptake using two types of MSNs: KCC-1 and MCM-41 (non-modified, positive charges with -NH2, and folic acid ligands) [42]. The FA-conjugated MSNs exhibited higher cellular uptake than MSNs-NH<sup>2</sup> and non-modified MSNs.

#### Morphological Structures of MSNs

The morphological structures (i.e., different shapes) play an important role in the cellular uptake and trafficking of nanoparticles into cells or organs. Trewyn et al. [126] studied the impact of different MSN shapes on cellular uptake in vitro, finding that a tubular structure achieves more efficient uptake by both cancer and normal cells than those of spherical morphology. Huang et al. [127] investigated the effect of three differently shaped particles on non-specific cellular uptake by human melanoma (A375) cells. Their results proved that particles with a larger aspect ratio are efficiently internalized by cells in large amounts at faster rates. Another study tested the core–shell MSNs with spherical or rod-like shapes for cellular uptake, showing that a rod shape results in more internalization by cells than a spherical shape [128] It is generally accepted that this effect could be due to the larger contact area of the rod than a sphere, permitting high favored internalization of nanoparticles in cell membranes [116,128] Furthermore, rod-shaped MSNs exhibit superior intracellular uptake compared to spherical MSNs [129]. The shape of the nanoparticles can allow a specific mechanism of intracellular uptake. In this context, Hao et al. [130] reported that the spherical particles are taken up by cells via clathrin-mediated endocytosis, whereas the rod-shaped particles enter cells through caveolae-mediated endocytosis.

#### Other Features of MSNs

One significant characteristic of any nanocarrier delivery system is hydrophobicity. Nanoparticles that have a hydrophobic nature exhibit a higher affinity for interacting with the cell membrane than those with a hydrophilic nature, contributing to improved cellular uptake [94].

#### 2.3.2. Biocompatibility and Biodistribution of MSNs

Any DDSs introduced into clinical investigations should exhibit biocompatibility with body tissues and organs. The biocompatibility is dependent on many MSN characteristics, such as size, shape, surface functionality, porosity, route of administration, and structure (Figure 9) [131].

**Figure 9.** Schematic illustration of the biocompatibility and biotranslocation of MSNs and the main physical–chemical characteristics. These highly influence the cellular uptake, intracellular translocation, and cytotoxicity on the in vitro level, and the biodistribution, biodegradation, excretion, and toxicity on the in vivo level. Reproduced with permission from [131], WILEY-VCH Verlag GmbH and Co. KGaA, 2012.

Most animal studies indicate the high biocompatibility and safety of MSNs [31,132,133]. The degree of biocompatibility of MSNs can vary according to many factors such as synthesis conditions, suitable structural features, and appropriate route at the right dosage [8,133–137]. As with other nanomaterials, for future translation to clinical applications, the safety aspects of MSNs should be considered carefully for each type [133]. Below, we present some studies highlighting the biocompatibility of MSNs in vitro and in vivo. Park et al. [138] investigated the biodistribution and biocompatibility of MSNs intravenously injected into mice at 20 mg/kg. The histopathological examination showed no significant toxicity compared to the control group. Their studies also indicated that MSNs are mostly cleared from the liver, spleen, heart, kidneys, brain, and lungs after 4 weeks. Hudson et al. [139] examined the biocompatibility of non-modified MSNs with particle sizes of ~150 (pores about 3 nm), 800 nm (pores about 7 nm), and ~4 µm (pores about 16 nm) at different does/concentrations. In vitro results in mesothelial cells showed that the cytotoxicity depends on the concentration; increasing concentration increases cytotoxicity towards cells. For in vivo studies, mcice were injected (intra-peritoneal, intra-peritoneal, and subcutaneous) at single dose of 30 mg/mL per mouse. The biocompatibility of MSNs in vivo depends on the dose and the route of administration. The subcutaneous injection of MSNs in rats indicates good biocompatibility, whereas intraperitoneal and intravenous injections at very high dose ~1.2 g/kg is lethal for mice due to toxicity or distress necessitating euthanasia, but at dose of ~40 mg/kg is safe. This severe systemic toxicity can be mitigated by further surface modification of the MSNs. Lu et al. [23] evaluated various doses of MSNs intravenously injected in mice (twice per week) for 14 days, they concluded that dose at 50 mg/kg is well tolerated in mice, no toxicity, no apparent abnormalities on the histopathological level or lesions were observed. They also revealed that this dose is adequate for the pharmacological application in cancer therapy.

Huang et al. [30] evaluated the biocompatibility of differently shaped and PEGylated MSNs (Figures 10–12), measuring various blood and serum biochemical indicators 24 h and 18 days after injection of MSNs at a dose of 20 mg/kg. All hematology markers were within normal ranges without any considerable toxicity, showing excellent biocompatibility. The results indicated that these particles do not influence liver function, and other parameters were also in the normal range. Concerning the quantitative determination of biodistribution and clearance, approximately 80% of MSNs are trapped in RES of the liver, spleen, and lung after 2 h of administration. Comparing the Si contents of different organs (at 2 h, 24 h, and 7 days), the Si content obviously decreased over time, indicating the possible degradation and clearance of MSNs from the liver, spleen, lung, and kidney. Moreover, the circulation time of MSNs in blood shows that long rod MSN (NLR) has a longer blood circulation time than short rod MSN (NSR), and the effect of surface modification by PEGylation is partially dependent on the shape.

**Figure 10.** Characterization of short rod MSN labeled with FITC (NSRFITC) and long rod MSN labeled FITC (NLRFITC). (**A**) TEM image of NSRFITC. (**B**) TEM image showing the mesostructure of NSRFITC. (**C**) TEM image of NLRFITC. (**D**) TEM image showing the mesostructure of NLRFITC. Arrows denote FITC embedded in a particle. Reproduced with permission from [30], American Chemical Society, 2011.

**Figure 11.** Biodistribution of differently shaped and poly(ethylene glycol) (PEG)ylated MSNFITC in liver, spleen, and lung observed by confocal microscopy 2 h after intravenous injection. Arrows denote NLRFITC distribution in the lung. Reproduced with permission from [30], American Chemical Society, 2011.

Yildirim et al. [140] evaluated the interactions of MSNs with different surface functional groups (ionic, polar, neutral, and hydrophobic) on blood parameters (hemolytic activity, thrombogenicity, and adsorption of blood proteins) to understand their biocompatibility. They concluded that the blood compatibility of MSNs positively improves with surface functional groups. Table 3 shows some data reported on the biocompatibility, biodistribution, and clearance of MSNs in vitro and in vivo.

**Table 3.** The biocompatibility, biodistribution, and clearance of MSNs with different shapes, sizes, and surface modifications in vitro or in vivo (injection or oral administration).



**Table 3.** *Cont.*

#### 2.3.3. Toxicity of MSNs

For preclinical and further clinical investigations, nanocarriers should be optimized to avoid undesirable characteristics (e.g., toxicity, side effects, non-specific interactions) and to allow good biological performance [131]. As one of the most abundant materials on Earth, silica (or silicon dioxide) in crystalline form can be found in nature as sand or quartz [149]. In contrast, the amorphous form is present in biological materials, including plants, cells, microbes (e.g., bacteria), vertebrates, and invertebrates [150]. Silica is also endogenous to human tissues, such as cartilage and bone [151]. Several efforts are underway to identify the toxicity of both the crystalline and amorphous forms of silica in different methods of application [10]. Crystalline silica mainly results in toxicity as a result of breathing fine crystalline powders created by the extraction of stone materials in soil [86]. Because it is found in vegetables, whole grains, and seafood, silica is a considerable part of the human diet (approximately 20–50 mg silicon/day for Western populations and reaching 200 mg/day for people whose diet is mainly plant-based as in China and India) [152]. Furthermore, after ingestion of silica, it circulates in the blood plasma and is absorbed in the form of silicic acid; up to 41% of silicic acid is excreted in the urine [153]. Silica nanomaterials are hydrolytically unstable and dissolve into the soluble form of silicic acid (Si(OH)4, pKa 9.6) [152]. This can occur through three different processes: hydration, hydrolysis, and ion-exchange [154]. A schematic representation of silica degradation is shown in Figure 13 [155]. Silicic acid has good bioavailability, contributing many health benefits, such as maintaining bone health [154,156,157]. The FDA has approved silica as "generally recognized as safe" for use in food additives and pharmaceutical products [86,155]. Silica nanoparticles have also been approved by the FDA for cancer imaging in clinical trials [158] and MSNs being developed with high potential for DDSs in clinical investigations [159].

The biosafety of engineered MSNs has been confirmed by several studies. As shown in the literature, MSNs have insignificant toxicity, and the degree of toxicity is identified as low from in vivo studies. Additionally, even such insignificant toxicity can be reduced with the optimization of the synthesis process. However, a few reported data [160–163] provide contrary reports. The plausible reason for this is that there are many factors affecting the biocompatibility and safety of MSNs (e.g., shape, size, surface functional groups, physicochemical properties). For example, the method of removing the surfactant/template after MSNs synthesis (by calcination or by refluxing) influences the final cytotoxicity [139]. According to a number of in vivo experiments, a coherent message regarding the toxicity of MSNs is that that the toxicity depends on the dose/concentration used. For example, Hudson et al. [139] investigated the toxicity for MSNs (single dose) in vivo, they evaluated various doses and administration routes. They concluded that a very high dose (1.2 g/kg) is lethal for mice compared to the half dose which is well-tolerated and safe when applied by intraperitoneal or intravenous injection. Liu et al. [164] studied the single and repeated dose

of MSNs via intravenous administration in mice. In the single-dose toxicity investigations, they found that the LD50 is higher than 1000 mg/kg. They also demonstrated that the groups that received low doses of MSNs did not show any behavioral, hematology, and pathological changes, whereas the groups that received high doses (1280 mg/kg) did not survive. In the repeated dose toxicity experiments, the mice groups were given continuously for 14 days followed by observation for a month. The results display that no mortality and no remarkable changes (in pathology or blood parameters) were detected. They also reported that the treatment of MSNs at daily doses (80 mg/kg) for 14 days is safe without any adverse effects in animals. Fu et al. [29] evaluated toxicity of MSNs (110 nm) in ICR mice treated by different routes: hypodermic, intramuscular, intravenous injections, and oral administration. They found that the oral route is well tolerated in mice even when increased to 5000 mg/kg compared to the intravenous route which shows the least threshold. As such results and others available from literature generated evidence to show that MSNs are well tolerated and safe in animals by various routes of administrations, i.e., oral, and intravenous injections [29,133,164,165]. However, there is no doubt that optimized production of MSNs and the final nanoformulation can achieve good biocompatibility and safe nanoparticles for treating diseases. Table 4 lists some studies that have explored the toxicity of MSNs and their delivery systems. For more reading concerning the toxicity and biosafety of MSNs, there are several extensive reviews [10,137,151,166–168]. The toxicity of any material/object, including MSNs, in a given environment is dependent on the dose [168]. As reviewed by Croissant et al. [168], there are mainly two mechanisms governing the toxicity of MSNs on the cellular level [88]. The first mechanism is surface silanolates that lead to membranolysis after the electrostatic interactions between MSNs and phospholipids of the cell membrane occur [169]. The second mechanism is reactive oxygen species (ROS) generation, which leads to cell death (by necrosis or apoptosis) by means of membranolysis [170]. Reducing the possible toxicity and improving the biosafety of MSNs can be achieved by optimizing the synthesis properties of MSNs for drug delivery and biomedical applications.

**Figure 13.** Schematic representation of the intact and degraded structures of silica material nanoparticles with the mechanisms and regulating factors underlying degradation. Reproduced from [155], WILEY-VCH Verlag GmbH and Co. KGaA, 2017.




**Table 4.** *Cont.*

#### **3. Drug Loading and Release Strategies**

*3.1. Drug Loading Strategies*

A unique feature of MSNs (e.g., large pore volume, high surface, pores, stability) makes them one of the most common nanocarriers exploited for drug delivery with a high drug loading capacity for a variety of drugs. Generally, drugs or therapeutic molecules can be loaded into MSNs with or without pore-capping. In the first technique without pore-capping, hydrophobic or hydrophilic therapeutic agents directly load MSNs with covalent or noncovalent bonding or electrostatic interactions. Loading of drugs or therapeutic agents into the mesopore network of MSNs delivers them to target tissues while simultaneously saving them from undesirable factors found in the surrounding environment (e.g., enzymatic degradation in the body) [9]. To load a suitable amount of drug, MSNs are immersed in the desired stock solution of the drug or therapeutic agent under stirring/shaking, during which the drug loading is highly driven by the concentration gradient, the competition between drug (adsorbate) and MSNs (adsorbent), adsorbate and solvent, and adsorbent and solvent [177,178]. As such, a loading process has been reported with a variety of drugs, such as camptothecin (hydrophobic anticancer molecule) [90], doxorubicin (Dox) hydrochloride [179], curcumin [69], quercetin [68], 5-fluorouracil (5FU) [180], erythromycin [181], alendronate [182], silymarin [183], and paclitaxel (PTX) [184]. Importantly, the degree of drug loading can be maximized by choosing the desired solvent for the drug, modifying the MSN surface, and adjusting the loading parameters (e.g., time, temperature) [10,86,185].

In the second strategy with capping as the "gatekeeper" for the pore openings of MSNs [168], the first stage is to engineer the outer surface of MSNs via many techniques: molecular or supramolecular functionalization, capping with nanoparticles, and coating with polymer, protein, or lipid. This approach can control the release and delivery of therapeutic agents. In the molecular or supramolecular approach, caps are mainly rotaxanes, pseudorotaxanes, and others consisting of a long chain-like molecule that is threaded via a cyclic molecule [186]. Under certain conditions (e.g., pH, redox), the cyclic molecule can attract rotaxane (to one end of it), with the presence of a stimulus allowing it to slide to the other end. By attaching the thread near the pore opening on MSNs, the sliding cyclic molecule blocks the pore when it is near the particle or opens if it slides away. The idea of nanoparticles as gatekeepers was pioneered by Lin and co-workers [187–190] with many nanoparticles, such as iron oxide nanoparticles and gold nanoparticles. These small nanoparticles can graft on top of MSNs loaded with cargos through chemical bonding upon cleavage of the chemical bonds linking the nanoparticles with MSNs. Consequently, under certain conditions (pH, redox), external stimuli can trigger the release of cargos in a controlled manner. Next, in the coating strategy, different types of biomaterials, such as polymer, proteins, and lipids, can be introduced onto the surface of MSNs loaded with drugs. Drug release can occur upon degradation of these biomaterials or changing the surrounding environment stimuli, either external or internal [191–193]. Table 5 lists some examples of reported studies on prodrug loading in MSNs and their loading capacity. Table 6 provides the different loading strategies and their relationship to stimulate release under various conditions for MSNs, showing the connection between loading and release effects.


**Table 5.** Loading capacity for natural prodrugs into MSNs established as recent drug delivery systems for natural medicinal substances.


**Table 6.** Different loading strategies and their relationships to stimuli release under various conditions for MSNs.

#### *3.2. Drug Delivery Strategies*

In this section, we provide a summary of delivery strategies that have been developed to treat cancer. This topic is well discussed in several reports for MSN delivery systems, and the readers are referred to these selected reviews [15,32,79,168,218–221]. Open pores on MSNs, the so-called cavities due to their porous structure, are not only used to load therapeutic agents, but also allow them to diffuse out in the surrounding solution. Closing these pores loaded with therapeutic agents is an essential step to avoid their premature release into the blood vessels, protecting from several side effects because of non-specific release [221]. Much effort has been made in controlled delivery systems with the stimulated or responsive release of therapeutic agents under certain conditions. Two major common strategies for delivering drugs have been reported by internal stimuli release (typical of the treated pathology), such as pH, redox potential, and enzymes, or by external stimuli (remotely applied by the clinician), such as magnetic fields, ultrasound, and light (Figure 14) [32].

**Figure 14.** (**A**) Schematic representation of stimuli-responsive release of drugs from MSNs. (**B**) Internal stimuli-responsive release. (**C**) External stimuli-responsive release. Reproduced from [81], MDPI, 2017.
