As cancer continues to affect over 18 million lives yearly, causing over 9.8 million mortalities worldwide as of 2020, there is an urgent need to find a cure ^[1]. Numerous cancer therapies have been and continue to be researched and developed. These include classical therapies, such as chemotherapy; radiotherapy; and surgery, and more modern, local therapies, such as focused ultrasound and specific drugs targeting tumor cells ^[2]. Although widely used, conventional methods like chemotherapy are limited due to their toxic side effects, low specificity, and resistance to the treatment, whether inherently present or developed later after successful initial treatments ^[3]. Hence, alternative treatments with reduced side effects have been explored, such as smart drug delivery systems (SDDSs).

SDDSs are generally defined as particles with a typical size between 1–500 nm that can carry one or multiple therapeutics either by encapsulating them in a matrix via covalent linkages, adsorption, or simple dispersion; these agents are later released at the target site ^[4][5]. These SDDSs have emerged as promising anti-tumor agents due to their enhanced solubility, bioavailability, and targeting; their ability for controlled delivery; and their potential to reduce drug resistance ^[3][6]. Furthermore, these drug delivery systems have the added advantage of allowing for the simultaneous delivery of drugs and genes (co-delivery), allowing for synergistic effects and enhanced anti-tumor activity ^[7][8]. Therefore, these delivery systems have a greater preference over conventional anticancer therapies due to their more specific and less toxic characteristics.

2. Disulfide Bonds

Disulfide bonds (S–S), covalent bonds between two sulfur atoms, are the most widely researched linkers used in the design and synthesis of SDDSs for cancer treatment. This is because they are susceptible to GSH and are easily and rapidly reduced to thiol groups in their presence, via a “thiol-disulfide exchange” reaction. In detail, the thiol-disulfide exchange redox reaction involves the donation of two hydrogen atoms by surrounding GSH molecules to the sulfur atoms of the disulfide bond, leading to the dissociation of the disulfide bond to form two thiol groups (–SH) [reduction] and the simultaneous oxidation of GSH to GSSG [oxidation] ^[9]. Hence, disulfide bonds are frequently used as linkers or crosslinkers in nanocarriers because, upon exposure to GSH at the tumor site, the disulfide bond will break, leading to complete disintegration of nanocarriers and effective cargo release ^[10][11]. Additionally, the use of disulfide bonds as linkers is favored because they are highly stable during blood circulation and will only break down when entering the tumor environment, thereby facilitating targeted delivery of chemotherapeutics to the tumor site alone and minimizing unwanted side effects ^[9]. Disulfide groups have been incorporated into different types of nanocarriers, and these will be explored in further detail in a later section. Moreover, the position of disulfide linkers in the drug delivery nanocarriers greatly influences the stability and effectiveness of the drug delivery system, allowing for the flexible design of various redox-responsive nanocarriers. Disulfide linkers have been utilized in the backbone of polymers used in the synthesis of nanocarriers ^[12][13]. Furthermore, they have also been used as side chain linkers ^[14] and as linkers on the surface of nanoparticles ^[15]. Disulfide linkers have also been used widely to link two chemical moieties, i.e., in making copolymers that can later aggregate and fold into micelles ^{[16][17][18][19]}. Additionally, disulfide bonds have been extensively used as crosslinkers in nanogels ^[20] and micelles, where, in micelles, the disulfide bonds were used to crosslink the inner core ^{[21][22][23][24]} or outer shell ^[25][26] of polymeric micelles. 

3. Diselenide Bonds

Diselenide bonds are covalent bonds between two selenium atoms (Se–Se); when reduced, two selenol (–SeH) groups form. This reduction occurs via the donation of a hydrogen atom from GSH to each selenium atom in the bond, leading to bond cleave and –SeH formation. Selenium and sulfur atoms are very similar in chemistry, as both belong to the same chemical group in the periodic table ^[27]. However, because selenium atoms are bigger than sulfur atoms, the diselenide bonds and carbon-selenium bonds (C–Se) are of lower dissociation energy and less stable than disulfide bonds, due to the longer bond length. While this characteristic of diselenide bonds is excellent when exposed to the reducing environment of tumors, as it enhances drug release, it also means that diselenide bonds have lower stability and can lead to the leakage of drugs when incorporated into a drug delivery system due to disintegration during circulation ^[10][27]. Diselenide-containing materials have also been known to have poor solubility, which hinders the effectiveness of diselenide-containing DDSs ^[28][29]. Nonetheless, diselenide bonds have excellent sensitivity to the reducing environments of tumor tissues (GSH) and have been utilized to develop redox-responsive drug delivery systems. Diselenide bonds have been incorporated into different nanocarriers for efficient and effective drug delivery. Generally, diselenide bonds are used as linkers in the backbone of polymers that can later self-assemble into micelles ^{[30][31][32][33][34][35]}, aggregate into stable nanoparticles ^[27][28], or form hydrogels ^[36][37] that can all deliver anti-tumor agents. To elaborate further, the diselenide bonds in the backbone could be within or between the polymer chains, as done in ^[31][37]. In ^[31], a polyurethane segment containing diselenide was polymerized with PEG [PEG–PESeSe–PEG] to form stable micelles. However, diselenide bonds can also be used as linkers between different polymer segments to form a block copolymer, as was demonstrated in ^[35][36]. In ^[35], 2 methoxypoly(ethylene glycol) (PEG) segments were linked to both ends of a polycaprolactone (PCL) segment via diselenide bonds to form a redox-sensitive triblock copolymer [CH₃–PEG–SeSe–PCL–SeSe–PEG–CH₃] that assembled into micelles which effectively encapsulated and delivered DOX. In addition to their role as linkers, diselenide bonds have also been used as crosslinkers in the development of redox-sensitive micelles ^[38], hydrogels ^[39], and nanogels ^[40]. In ^[38], PEG-b-PBSe block copolymers were irradiated with visible light during self-assembly to induce the formation of Se–Se crosslinking bonds in the micelle’s core and simultaneously encapsulate DOX and camptothecin (CPT). The resulting crosslinked micelle was biocompatible and demonstrated no significant side effects, making them potential candidates for clinical applications ^[38]. Moreover, selenocystamine molecules (containing –SeH) bonds were used to crosslink N-hydroxysuccinimide modified PEG molecules via the formation of Se–Se bonds to form injectable, redox-sensitive hydrogels for effective drug delivery ^[39]. Meanwhile, poly(2-methacryloyloxyethyl phosphorylcholine) (PMPC) based nanogels were crosslinked with Se–Se containing N,N′-bis(acryloyl) selenocystamine (BMASC) via a reflux precipitation polymerization reaction to form a stable and biocompatible GSH-responsive DDS ^[40]. Additionally, carboxymethyl chitosan-based nanoparticles have been crosslinked with 3,3′-diselenodipropionic acid di(N-hydroxysuccinimide ester) (DSeDPA-NHS) to form redox-responsive DDSs that could effectively load DOX and deliver it in-vitro and in-vivo ^[41]. Thus, diselenide bonds are also used as crosslinkers in the design of redox-responsive DDSs. An emerging type of redox-responsive nanocarrier that utilizes diselenide bonds are hybrid, mesoporous silica-based nanoparticles ^[42][43]. Mesoporous silica nanovehicles have been embedded with nanoparticles, coated with a protein gate of myoglobin or serum albumin via diselenide bonds, and loaded with DOX ^[42]. Upon exposure to the GSH, the diselenide bonds were cleaved, and DOX was released; additionally, the DDS was activated by pH and H₂O₂ in the tumor environment by changes in the protein conformation, releasing DOX ^[42]. Meanwhile, An et al. used Se–Se bonds in a silane coupling agent to clog the mesoporous silica pores and trap DOX in the developed hybrid drug carrier ^[43]. Both DDSs were biocompatible and demonstrated excellent cytotoxicity toward cancer cells ^[42][43]. Hence, diselenide bonds have been proven efficient in developing various redox-responsive DDSs. Moreover, diselenide bonds are versatile in that they not only respond to the reducing environment of the tumor microenvironment (GSH) but can also be cleaved upon exposure to reactive oxygen species (ROS), which are generally present at higher concentrations (up to 1 mM) in tumor microenvironments as compared to normal cells ^[35][42]. In this case, the diselenide bonds would be oxidized into selenic acids (RSeOOH) ^[35]. Common ROSs include hydroxyl radicals (OH•), hydroperoxy radicals (•HO₂), and hydrogen peroxide (H₂O₂), which is the reagent commonly used for in vitro experimentations ^{[31][35][37][42]}. Numerous studies have investigated the additional effect of the oxidative cleavage of redox-responsive DDSs containing diselenide bonds and whether it can be used as an additional internal stimulus to enhance tumor-targeted drug delivery ^{[31][35][36][37][40][41][42][44]}. In general, most of the designed, diselenide-containing DDSs disintegrated in the presence of varying concentrations of H₂O₂, with higher concentrations leading to quicker degradation and drug release at the tumor site ^[31][36][37]. When comparing the efficiency of oxidative cleavage via H₂O₂ against reductive cleavage via GSH, Yan et al. ^[42] found that the drug release from protein-gated nanoparticles in the presence of H₂O₂ and GSH was similar at equivalent concentrations of 1 mM. However, GSH concentrations in tumor environments are generally higher than H₂O₂ (up to 10 mM); thus, reductive cleavage of diselenide bonds is often the dominant cleavage taking place ^[45]. This is supported by various studies in which diselenide bond cleavage was significantly faster in the presence of GSH as compared to H₂O₂, although both were effective, with release percentages reaching up to 87% ^{[35][40][41][44]}. Therefore, diselenide bonds are useful chemical bonds that can be incorporated into various SDDSs for enhanced delivery via redox-responsiveness in the presence of GSH and ROS.

4. Succinimide-Thioether Linkages

Succinimide-thioether linkages, also known as maleimide thioether linkages, are a much more complex reduction-sensitive chemical entity when compared to disulfide and diselenide bonds. They are formed via the Michael addition reaction of maleimides with thiol groups. ^{[46][47][48][49][50]}. Upon exposure to tumor GSH concentrations, these succinimide-thioether groups undergo a retro Michael addition where the double bond is restored in the maleimide moiety and the thiol group is released (due to receiving a hydrogen atom from GSH). This is followed by a thiol exchange reaction where the thiol group in GSH attacks the double bond in the maleimide and forms a new succinimide-thioether ^[47][48]. The cleavage of the initial succinimide thioether linkage causes a gap in the nanocarrier structure, leading to subsequent disintegration and cargo release due to the reducing environment of tumors ^[48]. Hydrolysis may also occur, either initially or following the retro-Michael reaction and thiol exchange, leading to ring opening of the maleimide and more holes in the nanocarrier structure which can enhance cargo release ^[48]. It is important to note that only aryl thiol-based succinimide thioether linkages are GSH-sensitive, while alkyl thiols are not and will remain intact when exposed to GSH ^[8]. Furthermore, mechanistic studies investigating the GSH-susceptibility of succinimide thioether bonds found the type of thiol and N-substituent used to play key roles in the rate of cleavage of succinimide thioether linkages in reducing environments, where N-substituents capable of forming more hydrogen bonds and thiol groups of low pKas led to greater rates of bond cleavage in GSH ^[48]. Thus, the functional groups in the succinimide thioether linkages can be designed and modified to tune the reactivity and speed of release of cargo from DDSs containing these linkages. When compared to disulfide bonds, DDSs containing succinimide thioethers have been found to have release rates 10–100 times slower than those of disulfides but a longer circulation time, greater stability, and a more sustained drug release profile ^[47]. As such, succinimide thioether linkages are much more relevant in biomedical applications that require slower degradation profiles with prolonged drug delivery ^[47][48]. Research on succinimide thioether bonds in redox-responsive DDSs is much less common than disulfide and diselenide bonds, with fewer preseapersrch in the literature. Succinimide thioether bonds have most commonly been incorporated as crosslinkers in hydrogels ^{[49][51][52][53]} but have also been used in micelles ^[46] and hybrid hydrogel-liposome systems ^[8]. Hydrogels incorporating succinimide thioether linkages often utilized PEG as the building block polymer; however, different groups were used for PEG functionalization to form the succinimide thioether cross-linkages ^{[49][51][52][53]}. Baldwin et al. ^[49] formed hydrogels using thiolated-four-armed PEG polymers and crosslinked them with maleimide-functionalized heparin via the click Michael addition reaction. PEG was thiolated with either 3-mercaptopropanoice acid (MP), 4-mercaptophenylpropanoic acid (MPP), 3-mercaptoisobutyric acid (MIB), or 2,2-dimethyl-3-(4-mercaptophenyl)propionic acid (DMMPP), with MPP- and DMMPP-modified PEG-based hydrogels exhibiting high GSH-sensitivity while MP- and MIB-modified PEG-based hydrogels exhibiting significantly lower GSH sensitivity because of the absence of a retro-Michael cleavage as a result of the nonaromatic nature of the thiol groups. Subsequently, the MPP-PEG and DMMPP-PEG-based hydrogels were best at delivering low-molecular-weight heparin upon GSH exposure and hydrogel degradation, where MPP-PEG had the fastest heparin release profile ^[49]. These hydrogels are predicted to have anti-tumor properties; however, their capabilities have not been tested thoroughly. Similarly, a series of studies by Kharkar et al. developed several injectable hydrogels that consisted of multi-arm PEGs functionalized with aryl thiols and maleimide-functionalized PEGs or maleimide-functionalized heparin ^[51][52][53]. Aryl-thiol modifications have been done using moieties such as 4-mercaptophenylacetic acid (MPA) with additional ester modification ^[51][52] and MPA modified with photodegradable (PD) moieties ^[53] to allow for dual redox and light sensitivity. All hydrogels portrayed excellent GSH sensitivity at conditions mimicking in vivo GSH levels, with the hydrogels developed in ^[53], synthesized from a mix of PD-MPA-modified PEG, maleimide-modified PEG, and maleimide-modified heparin, exhibiting excellent delivery and release of bioactive proteins in vitro and in vivo due to the dual redox and light responsiveness of the hydrogels. In addition to hydrogels, succinimide thioether bonds have also been utilized to develop a hybrid hydrogels-liposome DDS, where the liposomes were functionalized with maleimides and used as crosslinkers by reacting with thiolated, 4-armed PEG molecules ^[8]. The PEG polymers were modified with 4-mercaptohydrocinnamic acid as the source of aryl thiols, and the resulting hydrogel showed minimal degradation (~15%) at 10µM GSH and a significant, rapid degradation (~97%) at 10 mM concentrations of GSH within seven days. Meanwhile, hydrogels with no succinimide-thioether crosslinkers showed no significant degradation in both GSH conditions. Moreover, these hydrogels were effective for the delivery of DOX alone and the co-delivery of DOX and cytochrome C, allowing for synergism and greater anti-tumor efficacy through a two-stage release process ^[8]. Furthermore, succinimide thioether linkages were also incorporated in the preparation of micelles for the delivery of the fluorescent probe, Nile Red ^[46]. Succinimide thioether bonds were used as linkers in the copolymers by reacting arylthiol-modified xyloglucan oligosaccharides with maleimide-modified polycaprolactone polymers, where arylthiol modification was performed using an alkyl thiol. Since an alkyl thiol was used, the observed cleavage and drug release in the presence of GSH were relatively slower. As such, the researchers proposed reacting bromomaleimide with thiols, instead of maleimides, which exhibited faster degradation when exposed to GSH due to the different mechanism ^[46]. From this, one may suggest using bromomaleimide-thiol linkages when interested in using alkyl thiols to overcome the previous restriction regarding the use of aryl thiols when forming succinimide thioether linkages for GSH-sensitivity. Nonetheless, succinimide thioether bonds are valuable, redox-sensitive moieties for the design of DDSs where slow and sustained drug release is preferred.

5. Tetrasulfide Bonds

Tetrasulfide bonds consist of four sulfur bonds linked to each other via covalent linkages (S–S–S–S). Although similar to disulfide bonds, tetrasulfide bonds are less commonly used in the synthesis of redox-responsive DDSs and have only been recently introduced ^{[54][55][56][57]}. When exposed to GSH, tetrasulfide bonds undergo multiple thiol-disulfide exchange reactions to eventually degrade into thiol groups (–SH) and produce hydrogen disulfide (H₂S) ^[56][58]. The initial attachment of GSH can either be on the α or β sulfur atom, leading in either case to the cleavage of the tetrasulfide bond and degradation of the nanocarrier ^[58]. Furthermore, the generated H₂S can play a role in killing cancer cells by damaging their mitochondria and reducing cellular respiration ^[57]. Due to these great properties, tetrasulfide bonds have been utilized in the development of redox-responsive DDSs. The most common nanocarriers that tetrasulfide bonds have been incorporated in are mesoporous organosilica nanoparticles (oMSNs) ^[54][55]. Tetrasulfide bonds are generally introduced into oMSNs using bis-[gamma-(triethoxylsilicon)propyl]tetrasulfide as the S–S–S–S-containing reagent ^[54][55][56]. For example, Song et al. developed a surfactant-free Stöber method to synthesize oMSNs containing tetrasulfide bridges in the silesquioxane framework ^[55]. The oMSN had good GSH sensitivity, with significant degradation observed after incubation in 5 mM GSH for 60 h. However, the rate of degradation was highly variable across the oMSNs due to unequal interactions with GSH because of the relatively high hydrophobicity of the oMSNs. Furthermore, the synthesized oMSN was successful in encapsulating methylene blue and curcumin (separately) ^[55]. Hence, these oMSNs have the potential to be used for cancer applications. Wang et al. ^[54] also developed oMSNs containing different concentrations of tetrasulfide bonds to induce redox responsiveness in the nanocarriers and to study the effect of the tetrasulfide content on the nanocarrier’s properties. Volume ratios of tetraethylorthosilicate (TEOS, the silica source) to BTEPTS (the S–S–S–S source) studied were 4:1, 2:1, and 1:1, with total volumes being constant. While one would assume that the greater amount of tetrasulfide bonds would lead to faster degradation due to more reduction-sensitive bonds, the study found oMSNs formulated at a 2:1 TEOS:BTEPTS ratio had the fastest degradation rate (in 5 mM DTT) and oMSNs with 1:1 TEOS:BTEPTS ratios had the slowest degradation rate, which was even slower than oMSNs made with no S–S–S–S bonds. These findings are quite significant and indicate the importance of using an optimal content of tetrasulfide bonds, as excess tetrasulfide bonds can lead to lower surface area oMSNs and slower degradation rates ^[54]. Further studies are needed to investigate their efficiency in vitro and in vivo. In addition to being incorporated in oMSNs alone, tetrasulfide bonds have been utilized in the development of hybrid oMSN nanocarriers, where oMSNs are mixed with other nanoparticles ^[56][57]. Song et al. ^[56] developed a core–shell hybrid nanocarrier, with gold nanorods (GNR) being the core structure and oMSNs containing tetrasulfide bonds being the shell. The oMSN coat was successful in encapsulating DOX in its mesopores, which was released rapidly upon GSH exposure due to the cleavage of tetrasulfide bonds. Meanwhile, the gold nanorod core was utilized for photothermal therapy and synergetic tumor eradication with DOX. The hybrid GNR-oMSN nanocarrier had excellent biocompatibility, rapid degradation in 10 mM GSH, and significant tumor inhibition in vivo, with no significant side effects—serving as a promising redox-responsive DDS for cancer treatment ^[56]. On the other hand, Liu et al. ^[57] developed a hybrid dendritic mesoporous organosilica (DMOS) DDS containing tetrasulfide bonds and doped with Mn²⁺ ions, Fe³⁺ ions, or Co²⁺ ions. The Mn²⁺-doped DMOS was effective in loading indocyanine green at high loading capacities while also being highly GSH-sensitive with rapid degradation rates and H₂S generation. This led to significant apoptosis of cancer cells due to the redox-triggered release of indocyanine green and subsequent photothermal effect by irradiating indocyanine green with near-infrared radiation ^[57]. Hence, tetrasulfide bonds exhibit good sensitivity to GSH and can be used in the development of redox-responsive DDSs. Further studies are needed to explore the use of these tetrasulfide moieties in other types of nanocarriers, such as micelles, liposomes, nanogels, and others.

6. Platin Conjugation

Platin conjugation involves the integration of platinum atoms into the structure of the drug delivery system. In general, octahedral Pt (IV) is incorporated into the structure of DDSs, which can then be reduced in the presence of GSH to Pt (II). This leads to the release of the platinum-based drugs and the disintegration of the DDS ^[59][60]. Platinum-based (Pt-based) complexes have the added advantage of exhibiting anticancer properties themselves; hence, they can be simultaneously used as prodrugs and as a redox-responsive unit in drug delivery systems ^[61]. An example of a Pt-based prodrug used in cancer treatments is Pt(IV) Cisplatin which, upon reduction to Pt(II) Cisplatin, releases free cisplatin to the tumor ^[59]. He et al. ^[59] developed a redox and pH dual sensitive micelle by reacting octahedral Pt(IV) cisplatin functionalized with two carboxylic acid groups [Pt(NH₃)₂Cl₂(OOCCH₂CH₂CO₂H)₂] with an orthoester monomer [2,2′-((4,4′-(Oxybis(methylene)) bis(1,3-dioxolane-4,2-diyl)) bis(Oxy)] to form the polymeric micelle backbone. The micelle was then loaded with DOX, at a high capacity and efficiency, for enhanced anti-tumor efficacy via simultaneous delivery of Pt(II) cisplatin and DOX. In vitro studies used dithiothreitol (DTT) as a reducing agent and found that the reduction of Pt(IV) in the micelle led to over 80% release of both drugs within 9–12 h at a pH of 5.0, mimicking tumor microenvironments. The micelles also exhibited longer circulation times, allowing for higher accumulation at the tumor site via the enhanced permeation retention effect (EPR), were stable, and had minimal side effects on other organs ^[59]. Hence, this redox-responsive Pt-based micelle exhibited excellent properties as an effective and rapid drug delivery system for cancer treatment. In addition to being incorporated into the backbone of micelles, platin conjugations have also been utilized to form redox-sensitive nanoparticles for drug delivery ^[62][63]. Ling et al. ^[62] utilized an octahedral Pt(IV) prodrug and coated it with lipid-PEG to form a biocompatible nanoparticle. Upon exposure to GSH, Pt(IV) is reduced to square planar Pt(II) and the nanoparticle disintegrates, releasing the Pt(II)-based cisplatin at the tumor site. The nanoparticle was found to have enhanced circulation time and better accumulation at the tumor site, allowing for successful drug delivery with fewer side effects ^[62]. Meanwhile, Wang et al. ^[63] integrated platin conjugation into nanoconjugates made up of a gamma-polyglutamic acid backbone and citric acid side chains, where octahedral succinic acid axially-functionalized Pt (IV) cisplatin was conjugated to the side chains. The nanoconjugate was dually responsive to pH and GSH, where the polymer underwent hydrolysis due to low pH, then the Pt (IV) group was reduced to Pt(II) cisplatin by GSH. In vitro and in vivo studies showed the nanoconjugate DDS to have excellent anti-tumor activity while also exhibiting significantly lower toxic side effects ^[63]. Thus, platin conjugation is another useful redox-responsive moiety that allows for the design of redox-responsive DDSs as well as Pt-based prodrugs.