Redox‐Responsive Polymeric Nanoparticle for Nucleic Acid Delivery and Cancer Therapy: Progress, Opportunities, and Challenges

Abstract

Cancer development and progression of cancer are closely associated with the activation of oncogenes and loss of tumor suppressor genes. Nucleic acid drugs (e.g., siRNA, mRNA, and DNA) are widely used for cancer therapy due to their specific ability to regulate the expression of any cancer‐associated genes. However, nucleic acid drugs are negatively charged biomacromolecules that are susceptible to serum nucleases and cannot cross cell membrane. Therefore, specific delivery tools are required to facilitate the intracellular delivery of nucleic acid drugs. In the past few decades, a variety of nanoparticles (NPs) are designed and developed for nucleic acid delivery and cancer therapy. In particular, the polymeric NPs in response to the abnormal redox status in cancer cells have garnered much more attention as their potential in redox‐triggered nanostructure dissociation and rapid intracellular release of nucleic acid drugs. In this review, the important genes or signaling pathways regulating the abnormal redox status in cancer cells are briefly introduced and the recent development of redox‐responsive NPs for nucleic acid delivery and cancer therapy is systemically summarized. The future development of NPs‐mediated nucleic acid delivery and their challenges in clinical translation are also discussed.

Redox‐responsive polymeric nanoparticles (NPs) are widely used for nucleic acid drug delivery and cancer therapy. Herein, the important signaling pathways regulating tumor redox status are summarized and the recent development of redox‐responsive NPs for nucleic acid delivery and cancer therapy is systemically reviewed. The future development of NPs‐mediated nucleic acid delivery and their challenges in clinical translation are also discussed.

1. Introduction

Cellular redox homeostasis refers to the dynamic balance between oxidizing and reducing species in cells. It is an internal defense mechanism that playing an important role in maintaining the physiological activities of cells, including cellular survival, growth, differentiation, and senescence. Oxidative stress, the relative excess of reactive oxygen species (ROS) and reactive nitrogen species, is closely associated with age‐related diseases such as neurodegenerative diseases and cardiovascular diseases as well as malignant tumor. Particularly, oxidative stress is a noteworthy biological feature of cancer cells because cancer cells tend to be characterized with higher levels of ROS, causing severe oxidative stress compared with normal cells. Classified by electrons status, the highly reactive molecules ROS family mainly can be divided into two types, free radicals and nonfree radicals. Free radicals with unpaired electrons include superoxide (O₂ ⁻), hydroxyl (•OH), alkoxyl radical (RO•⁻), and peroxyl radical (ROO•⁻).^{[ 1 ]} Hydrogen peroxide (H₂O₂), singlet oxygen (¹O₂), ozone (O₃), organic peroxide (ROOH), hypochlorous acid (HOCl), and hypobromous acid (HOBr) without unpaired electrons constitute nonfree radicals.^{[ 2 ]}

ROS are mainly generated from mitochondria, NADPH oxidase, peroxisomes, and endoplasmic reticulum. Mitochondrial electron transport chain (ETC) is the primary source of endogenous ROS. Complex I (NADH dehydrogenase) and complex III (ubiquinone cytochrome C reductase) in ETC mediate the redox‐related electron potential changes, which can generate little amounts of free electrons. The combination of O₂ with free single electrons results in the formation of O₂ ⁻, subsequently converted by superoxide dismutase (SOD) to the less cytotoxic product H₂O₂. H₂O₂ can be further reduced by catalase to the nontoxic product H₂O and O₂. However, in tumor cells with high concentrations of ROS, H₂O₂ is tended to involve in Fenton‐(like) reaction and converted to cytotoxic product •OH. Besides, intracellular ROS located on the cell membranes, nuclear membrane, and endoplasmic reticulum membranes are generated from NADPH oxidases (NOXs, including subtypes NOX1, NOX2, NOX3, NOX4, NOX5, DOUX1, and DOUX2). NOXs generate ROS through transferring an electron from NADPH to FAD and crossing the biological membrane to reduce O₂ to O₂ ⁻.^{[ 3 ]} Next, ROS are also generated from peroxisomes, endoplasmic reticulum nitric oxide synthase (eNOS), P450 enzymes, cyclooxygenases, and lipoxygenases.^{[ 4} , ^{5 ]}

Tumor cells are particularly sensitive to oxidative stress. On the one hand, the ROS level under a certain oxidative stress threshold exhibits the promoting effect in cancer development including tumor initiation, progression, angiogenesis as well as metastasis. ROS levels higher than the threshold display cytotoxicity to cancer cells, leading to progression inhibition and apoptosis. On the other hand, ROS also acts an important role in the occurrence and development of cancer by mediating oxidative DNA damage, proto‐oncogene activation, oncogene silencing, and DNA repair inhibition.^{[ 6 ]} If the gene mutations occur in critical gene sites for cell proliferation such as oncogenes and tumor suppressor genes, ROS‐mediated DNA damage or repair prevention always inspires the initiation and progression of cancer.^{[ 7 ]} Therefore, disruption of redox balance is instrumental in the development of cancer during the initiation and initial stages of cancer development. Under physiological conditions, the balance between oxidants and antioxidants is stably maintained, even though ROS are constantly produced. High oxidizing species level of tumor cells also activates the intracellular antioxidant system to relieve oxidative stress for maintaining intracellular redox homeostasis and promoting tumor cell survival. This intracellular antioxidant system mainly consists of enzyme systems including catalase (CAT), SOD, peroxiredoxin (Prx), paraoxonase (PON), glutathione peroxidase (GSH‐Px), glutathione sulfotransferase (GSTP1), thioredoxin reductase (Trx‐Rs), heme oxygenase (HO), and glutaredoxin (Grx) and nonenzyme systems including GSH/GSSG, NADPH/NADP⁺, uric acid, vitamin C, and vitamin E.^{[ 5 ]} Among above antioxidant system, homeostasis achievement of the intracellular redox state is largely maintained by reduced glutathione (GSH), which is usually present in a reductive state to protect cells from oxidative stress. Therefore, the redox state of cancer cell is primarily influenced by cellular ROS generation and buffering of cellular GSH.^{[ 8 ]}

The cellular redox homeostasis not only promotes tumor progression but also provides an effective therapeutic target in tumor therapy, including anticancer drugs targeting cellular redox homeostasis and redox‐responsive vehicles for intracellular ROS/GSH‐triggered drug release. In recent years, nucleic acid drugs have attracted more attention compared to traditional chemotherapeutic drugs, because they can effectively regulate gene expression to facilitate tumor inhibition. Besides, cationic polymers, as an emerging nucleic acid delivery vector, have displayed ideal biodegradability, biocompatibility, and modifiability compared to traditional liposome. Based on the abnormal ROS/GSH levels in tumor cells, some functional groups with ROS/GSH‐triggered cleaved ability have been introduced into polymer vector to obtain the redox‐responsive function. In this review, we will focus our attention mainly on the most substantial basic and translational cancer research involving redox‐responsive NPs for nucleic acid delivery and cancer therapy. First, we will systematically introduce the importance of redox homeostasis in cancer cells and how do cancer cells maintain redox homeostasis. Next, we will briefly introduce the universal nucleic acid drugs and nucleic acid vectors. At last, various redox‐responsive polymer vectors for nucleic acid drug delivery were comprehensively reviewed.

2. The Importance of Redox Homeostasis in Cancer Cells

Active ROS producing ability is known as the Achilles' heel of cancer cells. It is only under redox homeostasis condition that ROS perform the functions in favor of cancer development. However, comparatively high level of ROS always extensively attacks proteins, amino acids, RNA, DNA, and lipids in cancer cells, resulting in the disruption of normal metabolism. Once breaking through the affordable ROS threshold of cancer cells, redox imbalance will drive the cancer cells to apoptosis.

2.1. Cancer Cells Desperately Seek Redox Homeostasis

Due to tumor tissue always accompany by hypoxia, inflammation, and abnormal cell metabolism condition, high levels of ROS directly induce oxidative stress apoptosis in tumor cells. To avoid this hazard, cancer cells have to moderate oxidative stress by activating antioxidant genes and inducing metabolic reprogramming to maintain a redox homeostasis condition. In addition to avoid cell death triggered by excessive high‐intensity oxidative stress, it is only in redox homeostasis conditions can ROS promote cancer development through regulating various signaling pathway and transcription factors. For example, oncogene silencing, proto‐oncogene activating, and continuous hypoxic environment in tumor cells all contribute to the continuous generation of ROS, resulting in adverse events like mitotic signaling activation, cell cycle checkpoint crossing, proliferation promotion, neovascularization, and apoptosis inhibition. Next, moderate level of ROS in cancer cells causes the oxidation of protein and lipid, DNA damage, and gene mutation, facilitating subsequent tumor deterioration (Figure  1 ).

Figure 1.

Under certain redox homeostasis, excessive ROS promotes tumor development through various signaling pathways including PI3K/Akt, PKC, and MAPK. ROS also damages DNA molecules and proteins to disturb normally cellular metabolism. Some antioxidation genes are activated to clear excessive ROS.

Adequate amount of ROS in cancer cells activates a variety of proteins and signaling pathways including PI3K/Akt, MAPKs (JNK/ERK/p38), NF‐κB, Protein Kinase C, and cyclinD1.^{[ 9} , ^{10 ]} Such activated signals are instrumental in enhancing malignant tumor such as promoting tumorigenic development, inhibiting apoptosis, promoting proliferation, promoting neovascularization, acquisition of drug resistance, and promoting epithelial mesenchymal transition (EMT).

2.2. How Do Cancer Cells Maintain Redox Homeostasis?

Highly oxidative stress and huge concentrations of ROS engender oxidative damage and death of tumor cells. To avoid this hazard, cancer cells actively manipulate intracellular antioxidant enzyme and nonenzyme system to consume ROS. GSH/GSSG, Grx, and Trx/Trx‐Rx are the mainly functional actors in antioxidant system to consuming ROS.^{[ 11 ]} Overall, upregulating antioxidant genes and inducing metabolism reprogramming to enhance GSH generation for relieving excessive ROS is vital in alleviating oxidative stress.

2.2.1. Cancer Cells Maintain Redox Homeostasis via Metabolic Reprogramming

According to the Warburg effect, cancer cells actively tend to inhibit oxidative phosphorylation and bridge the energy gap by enhancing other metabolic fluxes with producing metabolites in favor of their proliferation. Such phenomenon is known as metabolic reprogramming. To relieve oxidative stress, GSH and NADPH generation is upregulated through metabolic reprogramming in cancer cells (Figure  2 ).

Figure 2.

Metabolic reprogramming for antioxidant effect in cancer cells. (Abbreviations: xCT: Cystine/glutamate antiporter system Xc⁻; GCL: Glutamate cysteine ligase; OAA: Oxaloacetic acid; TCA cycle: Tricarboxylic acid cycle; α‐KG: α‐ketoglutarate; GDH1: Glutamate dehydrogenase 1; GLS: Glutaminase; PSAT1: Phosphoserine Aminotransferase 1; GPT: Alanine aminotransferase; GLUT: Glucose transporter; HK: Hexokinase; G‐6‐P: Glucose‐6‐phosphate; F‐6‐P: Fructose‐ 6‐phosphate; FBPase: Fructose‐1,6‐bisphosphatase; PFK1: Phosphofructokinase‐1; F‐1,6‐BP: Fructose‐1,6‐bisphosphate; PEP: Phosphoenolpyruvate; PEPCK: Phosphoenolpyruvate carboxykinase; PK: Pyruvate kinase; LDH: Lactic dehydrogenase; PDH: Pyruvate dehydrogenase; PDK: Pyruvate dehydrogenase kinase; ACAC: Acetoacetic acid; HMG‐CoA: 3‐Hydroxy‐3‐MethylGlutaryl‐COenzyme A; MTP‐β: Mitochondrial trifunctional protein‐β; CPT: Carnitine palmitoyltransferase; CS: Citrate synthase; SOD2: Superoxide dismutase‐2; UCP2: Uncoupling protein‐2; STC1: Stanniocalcin‐1; FA: Fatty acid; PRPP: 5‐phosphorlbosyl α‐pyrophosphate; PRA: β−5‐phosphoribasylamine; Gln: Glutamine; Glu: Glutamate).

Inhibition of aerobic respiration is the first step of the inhibition of ROS generation in cancer cells. Mitochondria ROS (mito‐ROS) produced by mitochondrial ETC is the most important source of ROS in cancer cells. Oxidative phosphorylation of glucose is inhibited with the enhancement of anaerobic glycolysis according to the Warburg effect. The pyruvate generated by anaerobic oxidation of glucose is preferentially processed to lactate, avoiding further production of mito‐ROS.^{[ 12} , ^{13 ]} The low pH level, proinflammatory, and proangiogenic effects generated by lactate accumulation also facilitate tumor development.^{[ 14} , ^{15 ]} Overall, cancer cells pursue a high‐throughput conversion of glucose metabolism from oxidative phosphorylation to glycolysis. Early studies indicated cancer cells enhance glycolytic activity through high expression of lactate dehydrogenase A (LDHA, a glycolytic key enzyme) and glucose transporter 1 (GLUT, a glycolytic substrate transport enzyme) induced by c‐Myc and HIF‐1, respectively.^{[ 16} , ^{17 ]} Besides, pyruvate dehydrogenase (PDH), a key enzyme in the entrance of the TCA cycle, is inhibited by the highly expressed pyruvate dehydrogenase kinase (PDK) in cancer cells. Therefore, the metabolic entry of oxidative phosphorylation is restricted and the pyruvate generated by anaerobic oxidation easily enters into the glycolytic pathway. Next, the highly expressed protein stanniocalcin‐1 (STC1) improves the survival of lung cancer cells through oxidative phosphorylation uncoupling by STC1‐dependent uncoupling protein 2 (UCP2) upregulation. UCP2 is able to shift the glucose metabolism to glycolysis, resulting in the inhibition of ROS production and the alleviation of oxidative stress.^{[ 18 ]} In details, uncoupling effect of UCP2 is achieved by the proton leak‐mediated decrease of transmembrane electrochemical gradient that H⁺ is directly transferred from intermembrane space of mitochondria to mitochondrial lumen rather than used in adenosine triphosphate (ATP) synthesis. At last, breast cancer cells are also indicated to convert the highly toxic ROS to the mild H₂O₂ through upregulating manganese superoxide dismutase (MnSOD/SOD2). H₂O₂ in mitochondria mediates activation of adenosine monophosphate (AMP)‐activated kinase (AMPK) signaling, the main switch of mitochondrial metabolic transition, resulting in the effective strengthening of glycolysis for better survival of tumor cells.^{[ 19 ]}

The enhancement of pentose phosphate pathway (PPP) is another step for relieving oxidative stress. Glucose‐6‐phosphate (G‐6‐P), an intermediate product of glycolysis, can enter the PPP via G‐6‐P dehydrogenase (G‐6‐PD) to generate reducing equivalents of NADPH and ribose 5‐phosphate. As a promoter and gatekeeper of malignant tumor,^{[ 20 ]} highly active hexokinase (HK) maintains the continuous production of G‐6‐P in cancer cells, facilitating the high throughput maintenance of PPP. Besides, ribose 5‐phosphate requires the involvement of glutamine in the subsequent de novo synthesis of purine nucleotides with the generation of glutamate as by‐product. Both glutamate and NADPH act as important raw materials for the synthesis of GSH to counteract oxidative stress. For example, high level of P‐cadherin is considered as a poor prognostic factor in breast cancer which is associated with hypoxia, glycolysis, and acidosis biomarkers because it upregulates G‐6‐PD to upwardly adjust the flux of PPP. At the same time, the antioxidant enzyme SOD2 was also activated by P‐cadherin to alleviate oxidative stress.^{[ 21 ]} Next, active sites like Cys358 of the glycolytic enzyme pyruvate kinase M2 (PKM2) and Cys152 of glyceraldehyde 3‐phosphate dehydrogenase (GAPDH) are rapidly oxidized under oxidative stress, causing the inactivation of PKM2 and GAPDH and high PPP‐dependent production of NADPH.^{[ 22} , ²³ , ^{24 ]} Furthermore, glucose‐6‐phosphate dehydrogenase (G‐6‐PD) activated by the antioxidant‐sensitive gene nuclear factor (erythroid‐derived 2)‐like 2 (Nrf2) and TP53‐inducible regulator of glycolysis and apoptosis (TIGAR) have been shown to exert antioxidant effects via regulating metabolic shifts to PPP.^{[ 25} , ²⁶ , ²⁷ , ^{28 ]}

Maintaining fatty acid β‐oxidation (FAO) facilitates consuming excessive ROS. Cancer cells also maintain a certain level of FAO to produce acetyl CoA followed by condensing with oxaloacetate to form citric acid. Such metabolic process is accompanied by the release of NADPH as a reducing equivalent for relieving oxidative stress. FAO is so vital for cells to produce NADPH that its intensity largely determine whether cells can maintain the redox homeostasis. As an example, carnitine palmitoyltransferase (CPT), a key enzyme for fatty acid oxidation, was shown to be highly expressed in tumor cells.^{[ 29} , ^{30 ]} Therefore, aiming CPT as a target for FAO inhibition has become a good option for cancer therapy which may cause a significant decrease of NADPH production and amplify the oxidative stress in tumor cells.^{[ 31} , ^{32 ]} On the other hand, the β subunit of mitochondrial trifunctional protein (MTP‐β) is another key rate‐limiting enzyme in FAO which mediates the second, third, and fourth steps of the process of removing two carbon atoms from lipoyl coenzyme A in mitochondria. Its vital active site Cys458 is so sensitive to the redox state that excessive oxidative stress makes it easy to inactivate MTP‐β, resulting in abnormal production of NADPH and GSH. For example, nuclear receptor 77 (Nur77) could protect MTP‐β from oxidative inactivation by entering mitochondria during glucose starvation to facilitate cell survival in melanoma.^{[ 33 ]} Meanwhile, Chen et al.^{[ 34 ]} reported the stem cell markers NANOG activated by Toll‐like receptor 4 (TLR4), which can suppress oxidative phosphorylation and ROS production, conferring significant chemotherapy‐resistant property to hepatocellular carcinoma cells. Therefore, targeted silencing of the NANOG gene and effectively restoring oxidative phosphorylation to reduce FAO are hopeful therapeutic strategies for preventing redox imbalance from being relieved by cells.

Glutamine catabolism, another vital source for the production of GSH precursor glutamate, is significantly upregulated by cancer cells to provide adequate levels of GSH. It is well known that cancer cells are generally addicted to glutamine to compensate for the limited TCA cycle. The reduction of GSSG to GSH is one of the pathways for generating antioxidant GSH, which is catalyzed by GSH reductase and supplied with hydrogen by NADPH. Another pathway of generating GSH is the de novo synthesis mediated by β‐glutamylcysteine synthetase (GCS). Glutamate from glutamine catabolism is involved in the de novo synthesis of GSH as a substrate. Besides, glutamate also acts as a driver to promote intracellular transport of cysteine, another synthetic raw material of GSH. Therefore, the enhanced catabolism of glutamine often represents a metabolic reprogramming and stable redox homeostasis of tumor cells.^{[ 35 ]} Such behavior can be induced by transcription factor c‐Myc via upregulating glutamine transferring and strengthening glutamine catabolism. It is attributed to the c‐Myc‐dependent upregulation of glutamine transporter SN2 and ASCT2, and glutamine catabolic catalase GLS and GDH1,^{[ 3 ]} following by the production of the glutamine, glutamate, and α‐ketoglutarate. Such process not only provides supplemental substrates for the inhibited TCA cycle but also proposes raw materials for GSH synthesis.

Last but not least, integrating cellular input of glucose derivatives and amino acids by serine/glycine metabolism also helps cancer cells regulate redox homeostasis, because it effectively promotes the transfer of one‐carbon units for the biosynthesis of nucleotides, S‐adenosylmethionine, NADPH, and GSH.^{[ 36 ]} The activities of key rate‐limiting enzymes of serine/glycine metabolism and folate cycle such as methylenetetrahydrofolate dehydrogenase 2 (MTHFD2), phosphoserine aminotransferase (PSAT), and serine hydroxyltransferase (SHMT) are activated by the antioxidant gene Nrf2 under oxidative stress.

2.2.2. Cancer Cells Maintain Redox Homeostasis via Activating Antioxidant Genes

In addition to metabolic reprogramming, a large number of antioxidant genes net play an important role in the redox buffering system of tumor cells such as Nrf2, p53, HIF, NF‐κB, PKM2, GAPDH, NUAK1, FOXO, ANGPT1, NMN, ME, AP‐1, HSF‐1, BRCA1, ULK1, and PTPN12.^{[ 3} , ¹⁰ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ^{44 ]} In this section, several important and typical antioxidation genes regulating oxidative stress in cancer cells will be systematically described (Figure  3 ).

Figure 3.

The vital antioxidation genes regulating oxidative stress and their function in cells. Among them, Nrf2 acts as the most vital role in upregulating downstream antioxidant genes.

Antioxidant Gene Nrf2

Nrf2, the most critical antioxidant gene in cancer cells, is extremely sensitive to the redox status of cancer cells. Overexpression of Nrf2 is known to be detected in a variety of cancers. In a nonoxidative stress environment, Nrf2 always be inactivated via binding to a substrate adaptor for the Cullin‐3 (Cul‐3)‐dependent E3 ubiquitin ligase complex (Kelch‐like ECH‐associated protein 1, KEAP1) following by being degraded in the 26S proteasome. Under the oxidative stress condition, the disulfide bond between Nrf2 and KEAP1 is destroyed. Nrf2 no longer binds to the ubiquitin ligase of Cul‐3, allowing Nrf2 to be released into the cytosolic and transited into cell nucleus, where Nrf2 conjugate small protein Maf and form a transcription factor complex. This complex next activates antioxidant response elements and promotes the transcription of downstream target protein for relieving oxidative stress.^{[ 45} , ^{46 ]} Nrf2 is at least a marker of poor prognosis in cancer treatment which mediates tumor progression, invasion, angiogenesis, chemoresistance, and radiation resistance.^{[ 47} , ⁴⁸ , ^{49 ]} Several typical downstream genes of the Nrf2‐KEAP1 signaling pathway have been found to involve in relieving oxidative stress.

NAD(P)H: quinone oxidoreductase 1 (NQO1) is a typical effector protein downstream of Nrf2 signaling. It neutralizes ROS on the cytoplasmic membrane and induces an elevated NAD (P)⁺/NAD(P)H ratio through reducing the active quinone to hydroquinone to alleviate ROS stress.^{[ 50} , ^{51 ]}

Heme oxygenase‐1 (HO‐1), as the induced form of heme oxygenase, is another important antioxidant gene in Nrf2 downstream. As endogenous protective compounds, HO‐1 is produced during heme‐breakdown mediated by biliverdin reductase (BVR) and the generation of CO, Fe²⁺, and biliverdin, which are extremely important in maintaining cellular redox homeostasis and relieving ROS pressure.^{[ 52 ]}

xCT (also known as SLC7A11) is an important target for maintaining redox homeostasis which can be activated by Nrf2 signaling. Na⁺/Cl⁻‐dependent amino acid transporters xCT consists of the light chain SLC7A11 and the chaperone heavy chain SLC3A2 subunits, acting as a transporter to convey extracellular cystine into cells for GSH synthesis. This process is driven by the intracellular and extracellular glutamate gradient. Since cancer cells failed to possess the key enzyme in cysteine synthesis—γ‐cystinase, xCT is a vital intracellular source of cystine/cysteine. In addition, xCT has also been shown to promote the breakdown of glutamine into glutamate, enhancing the driving power of intracellular transport of cystine.^{[ 53 ]} xCT has been found upregulated in various human tumor to facilitate GSH generation and oxidative stress relieving.^{[ 54} , ⁵⁵ , ^{56 ]}

According to the above‐mentioned, metabolic reprogramming relating to PPP, glutamine catabolism, and serine/glycine metabolism is a common effective pathway for cancer cells to generate ROS and relieve oxidative stress. It is partly due to the regulatory effect of Nrf2 on the various key enzymes in different metabolic pathways in cancer cells.^{[ 57 ]}

G‐6‐PD controls the metabolic entry of PPP whose activation by Nrf2 can significantly increase the PPP flux to produce sufficient NADPH. Zhang et al.^{[ 26 ]} reported that Nrf2 can activate G‐6‐PD and PPP flux to link Notch pathway for facilitating proliferation, migration, and metastasis in breast tumor. In head and neck squamous cell carcinoma (HNSCC), c‐Myc‐dependent Nrf2 exceeded activation can promote malignant development through G‐6‐PD‐induced PPP enhancement.^{[ 58 ]}

GLS, a key rate‐limiting enzyme of glutamine catabolism is also a target protein of Nrf2, functioning effectively in controlling glutamate production rate and regulating oxidative stress.^{[ 59 ]} Some reports suggested using GLS as an inhibitory therapeutic target significantly suppresses Nrf2/KEAP1/GLS signaling in promoting glutamine catabolism, effectively reactivating chemotherapy sensitivity, and controlling migration and proliferation in cancer cells.^{[ 60} , ^{61 ]}

One‐carbon metabolism can also be involved in antioxidant reactions under the regulation of Nrf2 because the key enzymes of serine/glycine metabolic metabolism including methylenetetrahydrofolate dehydrogenase 2 (MTHFD2), phosphoserine transaminase‐1 (PSAT‐1), phosphoglycerate dehydrogenase (PHGDH), and serine hydroxymethyl‐transferase‐2 (SHMT‐2) are regulated by Nrf2 signaling.^{[ 9 ]} DeNicola et al. also reported that Nrf2 is always activated abnormally to support GSH production through activating transcription factor 4 (ATF4)‐dependent upregulation of PHGDH, SHMT‐2, and PSAT‐1 in nonsmall cell lung cancer (NSCLC), which links serine/glycine metabolism to the clinical aggressiveness of cancer cells.^{[ 62 ]} Another example is the intracellular redox homeostasis maintaining function of Nrf2 in hepatocellular carcinoma (HCC). The SUMOylation of Nrf2, a modified manner that small ubiquitin‐like protein being added to Lysine residue 110 (K110) of Nrf2, can effectively clean up ROS via upregulating glutathione peroxidase (GPX) in HCC. The Nrf2‐dependent highly expressed PHGDH is also detected, indicating SUMOylation of Nrf2 promotes a metabolic shift from glucose metabolism to serine synthesis.^{[ 63 ]}

p53 Gene

The antioncogene p53, acting as a cell gatekeeper and gene guardian, has been found to be activated in more than half of human cancers. Although a large number of reports have indicated that p53 promotes the apoptosis of cancer cells in high oxidative stress, p53 is also a key antioxidant gene activated by mild oxidative stress to facilitate tumor development. Due to the reduction‐dependent transcriptional activation ability of the key transcription factor encoded by wild‐type p53 gene after binding to the corresponding DNA promoter, p53 tends to mediate the activation of antioxidant signaling under moderate concentrated ROS level rather than the pro‐oxidation and pro‐apoptotic signaling.^{[ 64 ]} Therefore, p53 not only stimulates the expression of antioxidant survival genes but also regulates the expression of pro‐apoptotic and pro‐oxidant genes which depends on the redox condition of cancer cell.

MnSOD (SOD2), a p53‐regulated downstream gene widely involving in early and advanced stages of cancer,^{[ 65 ]} has been considered as a p53‐dependent oncogenic target with antioxidant function localizing in the mitochondrial matrix.^{[ 66} , ^{67 ]} In melanoma cells, mutant p53 induces Sirtuin3, a mitochondrial NAD‐dependent deacetylase, to catalyze the deacetylation of MnSOD and mediate its activation. ROS‐scavenging ability of MnSOD functions as a defense mechanism to relieve the oxidation promoting effect from mutant p53.^{[ 68 ]}

TIGAR is another p53‐dependent antioxidation gene in cancer cells, which promotes PPP flux and NADPH production via suppressing PFK‐1 to block glycolysis. It exhibited similar catalytic activity to fructose‐2,6‐phosphatase (FBPase) that promotes the degradation of fructose‐2, 6‐bisphosphate.^{[ 69} , ^{70 ]} Due to fructose‐2, 6‐bisphosphate is the most effective allosteric activator of PFK‐1 in glycolysis, bio‐activity of PFK‐1 and glycolysis are significantly suppressed, leading to the avoidance of high ROS production. Therefore, the presence of TIGAR can perfectly maintain mitochondrial function and mitigate oxidative stress damage in cancer cells through mediating NADPH generation in various cancer cells.^{[ 66} , ⁷¹ , ⁷² , ^{73 ]} However, recent reports indicate that in p53‐deficient tumor cells, Nrf2 can also activate TIGAR at the transcriptional level.^{[ 28 ]}

Nrf2 also partly undertakes the antioxidation function of p53. The antioxidant effect of Nrf2 can be stably activated by p53‐dependent Sestrins1/2 and p21. Sestrins1/2 induces nuclear translocation of Nrf2 by binding to and degrading KEAP1.^{[ 74 ]} p21 escorts Nrf2 to the nucleus by competitively binding to Nrf2 with KEAP1.^{[ 75 ]} Sestrins1/2 also induces reactivating of per‐oxidized Prx, indicating p53 is required for the regeneration of Prx‐containing cysteine sulfuric acid.^{[ 76} , ⁷⁷ , ^{78 ]}

HIF

Hypoxia‐inducible factor (HIF), an important regulator of glycolysis, is upregulated under oxidative stress condition and recruited in the hypoxia response elements (HREs) of the HIF‐1 targeted antioxidation genes to enhance their expression.^{[ 79} , ^{80 ]} ROS mainly stabilizes HIF through PI3K/Akt,^{[ 81} , ⁸² , ^{83 ]} NF‐κB,^{[ 84} , ^{85 ]} and MAPKs^{[ 82} , ^{86 ]} signaling pathway in cancer cells. Therefore, upregulation of HIF can be considered as an effective method for cancer cells to counteract oxidative stress induced by hypoxic environment.

Metabolic reprogramming avoiding oxidative phosphorylation is the first pathway for HIF restoring oxidative homeostasis, because HIF is able to mediate a metabolic switch from oxidative phosphorylation to anaerobic oxidation for reducing ROS generation in mitochondrial ETC.^{[ 87} , ⁸⁸ , ⁸⁹ , ^{90 ]} First, pyruvate generated from anaerobic oxidation of glucose is introduced into anaerobic oxidation rather than the TCA cycle by the highly expressed HIF‐dependent LDHA.^{[ 91} , ^{92 ]} Besides, HIF also reduces mitochondrial respiration by upregulating hexokinase 2 and pyruvate dehydrogenase kinase 1 (PDK1) in the first step of glycolysis to inactivate PDH. It results in the decreased intensity of acetyl coenzyme A entry into the TCA cycle and the shifting metabolism to anaerobic oxidation.^{[ 93} , ⁹⁴ , ^{95 ]} Next, PPP and serine anabolism, as two main energy shunt pathways for glucose metabolism, also act as targets for HIF‐mediated metabolic reprogramming.^{[ 89 ]} For example, both key rate‐limiting enzymes G‐6‐PD in PPP and phosphoglycerate dehydrogenase (PHGDH) in serine anabolism are upregulated by HIF in various cancer cells.^{[ 96} , ⁹⁷ , ^{98 ]} Last, HIF also relieves oxidative stress by mediating glutamine catabolism and promoting de novo synthesis of GSH. The stable HIF signaling can induce glutaminase‐mediated breakdown of glutamine to glutamate,^{[ 99 ]} which also facilitates intracellular cystine conveying by xCT. Cystine is reduced to cysteine after being transported into cancer cells by xCT and linked with glutamate with the help of glutamate cysteine ligase (GCL) to form GSH.

Autophagy is another method for HIF to induce oxidative stress resistance. Autophagy serves as a defense mechanism to protect cellular integrity by sequestering damaged organelles or protein components in autophagic vesicles for subsequent degradation in lysosomes. High ROS level is a significant characteristic in hypoxic tumor cells, improving autophagy flux allows timely phagocytosis of damaged organelles and prevents toxic damage by ROS.^{[ 100} , ^{101 ]} Due to mitochondrial ETC acting as the main ROS‐producing factory in cancer cell, mitochondria become the most vulnerable organelle being attacked by mito‐ROS. Mitochondrial DNA damage and ROS production constitute a mutually beneficial vicious cycle, resulting in severe ROS accumulation. It eventually leads to the loss of mitochondrial membrane potential and membrane rupture, resulting in cytochrome c spillover‐triggering programmed cell death. Therefore, mitochondrial autophagy (mitophagy) undertakes the task of removing damaged mitochondria to control intracellular ROS production and maintain redox homeostasis.^{[ 102} , ^{103 ]} HIF upregulates mitophagy to stabilize redox homeostasis by binding HREs of multiple target genes. For example, mitochondrial ribosome protein L52 (MRPL52) is upregulated in HIF‐1‐dependent pathway in breast cancer cells. MRPL52 is able to mediate mitophagy through activating PTEN‐induced putative kinase 1 (PINK1)/Parkin signaling to suppress uncontrollable ROS generation.^{[ 104 ]} Besides, a typical BH3 domain of hypoxia‐induced BCL2 and adenovirus E1B 19 kDa interacting protein 3 (BNIP3) signaling recently has been accepted as a sensitive regulator of mitophagy in downstream of HIF.^{[ 105} , ¹⁰⁶ , ¹⁰⁷ , ^{108 ]} For instance, overexpressed adipose triglyceride lipase in Hela cell line can upregulate FAO level in mitochondria, resulting in high production of ROS and stabilized HIF‐1α. HIF‐1α subsequently strengthens glycolysis and BNIP3‐mediated mitophagy to restrict further cell damage from ROS.^{[ 109 ]}

Other Antioxidative Genes

NF‐κB is a coordinative factor between inflammatory and metabolic pathways, which can be upregulated by oxidative stress signaling in cancer cells to activate downstream antioxidation genes including SOD1, SOD2, catalase, Trx1, Trx2, GST, NQO1, and GPX.^{[ 110 ]} Besides, oxidative stress‐mediated NF‐κB activation not only upregulates GCL for GSH generation but also modulates the expression of cIAP1 and cFLIP (inhibitor of apoptosis proteins) to facilitate tumor cells survival and therapeutic resistance.^{[ 111} , ¹¹² , ^{113 ]} Next, NF‐κB is also effective in inducing the Warburg effect in cancer cells to switch metabolism from oxidative phosphorylation to glycolysis through upregulation of glycolytic enzymes and suppression of key enzymes in mitochondria via p53‐dependent pathway.^{[ 114} , ^{115 ]}

PKM2, an isozyme of pyruvate kinase, catalyzes the conversion of phosphoenolpyruvate to pyruvate. Functional inhibition of PKM2 is critical for the transferring of glucose metabolism to PPP for alleviating the oxidative stress. For example, Cys358 of PKM2 is vulnerable to acute increase of ROS level through oxidative inactivation, leading to a higher PPP flux for preparing enough reduction potential for buffering ROS, which gives lung cancer cells an extra survival advantage.^{[ 22 ]} Besides, after translocating to mitochondria under conditions of oxidative stress, PKM2 phosphorylates the threonine (T) 69 site of Bcl‐2 by binding to the chaperone protein HSP90α1. Such process blocks the binding of Cul3‐based E3 ligase to Bcl2 and the subsequent degradation of Bcl2, which prepares good resistance to apoptosis and oxidative stress adaptation for glioblastoma.^{[ 116 ]} Last, PKM2 can also be detected in the exosomes of NSCLC in a hypoxic environment. The reduction equivalents produced by PKM2‐mediated hyperactive glycolysis counteract the toxic ROS produced from chemotherapeutic agent cisplatin, resulting in significant cisplatin‐resistance.^{[ 117 ]}

NUAK1 (also known as AMPK related‐kinase 5), a serine/threonine kinase member of the AMPK‐α family, is accepted as an important part in antioxidation effect in cancer cells. Under oxidative stress condition in colorectal cells, NUAK1‐mediated Akt signaling and PP1β suppression can silence GSK3β‐dependent inhibition of Nrf2 nuclear translocation which amplifies Nrf2 signaling for relieving ROS level.^{[ 118 ]} Besides, significant nuclear translocation of NUAK1 under oxidative stress is detected, which mediated the metabolic conversion to glycolysis to decrease ROS generation.^{[ 119 ]} Inhibitors targeting NUAK1 have shown promising therapeutic effect in animal tumor models by disrupting redox homeostasis.^{[ 120 ]}

Breast cancer susceptibility gene 1 (BRCA1) deficiency triggers a dramatic increase of H₂O₂ level causing oxidative damage in hereditary ovarian cancer cells, which can be eliminated by the replacement of BRCA1 gene. This observation indicated the important function of BRCA1 in antioxidant stress.^{[ 121 ]} In fact, the antioxidant effect of BRCA1 is achieved by stabilizing and amplifying p53 and Nrf2 signaling directly or indirectly. BRCA1 can activate transcription by interacting with the C‐terminus of p53 to strengthen its transcriptional activity at the promoter of p21 or BAX. p21 next inhibits KEAP1‐dependent Nrf2 signaling silencing by directly binding to Nrf2 and forming a competitive relationship with KEAP1.^{[ 122} , ^{123 ]} BRCA1 also combines with Nrf2 directly to protect Nrf2 from KEAP1‐induced ubiquitinated degradation.^{[ 124 ]} Some observations have indicated that BRCA1 upregulates Nrf2 downstream genes including NQO1, GST, oxidoreductase, and HO‐1 in prostate cancer and breast cancer as well as maintaining a higher ratio of GSH/GSSG balance under oxidative stress condition.^{[ 38} , ^{125 ]}

3. Nucleic Acid Drugs for Cancer Therapy

Nucleic acid drugs‐mediated tumor therapy has been accepted as an up‐and‐coming strategy to regulate genes expression for suppressing tumor therapy. In early years, attenuated RNA virus vaccines and DNA recombinant virus‐mediated nucleic acids therapy represented the main application of nucleic acid drugs. Recently, the strategies of upregulating/downregulating target genes expression and filling in the defective genes in host cells are mainly used to achieve the therapeutic effect. Overall, nucleic acid drug therapy has evolved from the traditional viral vector model to a safe, low‐immunogenicity nonviral vector model, including protein‐expressing plasmid (pDNA), messenger RNA (mRNA), small interfering RNA (siRNA), and microRNA (miRNA).^{[ 126} , ^{127 ]} In this section, we have systemically introduced the universal types of nucleic acid drugs and reviewed their modes in regulating target gene expression (Figure  4 ).

Figure 4.

The common nucleic acid drugs include pDNA, mRNA, miRNA, and siRNA. Both pDNA and mRNA serve as an enhancer in transcription of targeted gene. miRNA and siRNA disturb translation of targeted gene in different model.

3.1. pDNA Drugs

pDNAs (plasmid DNA) are the cytoplasmic double‐stranded DNA molecules outside of the chromosomes or nucleoids of organisms such as bacteria and yeast, which have become a typical tool for genetic engineering. pDNAs effectively compensate for the gene deficiency or low‐expression in tumor cells to regulate deficient gene expression or regulate RNA interference facilitating tumor development.^{[ 128 ]} For example, the pDNA encoding murine interferon alpha (mIFN‐alpha) was injected into mice tumor model to regulate the antitumor effect on primary and metastatic tumors. The upregulation of mIFN‐alpha in tumor microenvironment (TME) induced T helper 1 (Th1)‐type immune response to inhibit tumor metastasis.^{[ 129 ]} In recent years, pDNAs were designed to be carried by polymeric NPs to prevent the degradation by serum nucleases for facilitating tumoral accumulation. A reduction‐sensitive linear cationic click polymer was prepared to carry the pDNA for expressing P‐glycoprotein shRNA, which could silence the expression of P‐glycoprotein, a key protein in keeping multidrug resistance in breast cancer cells.^{[ 130 ]} Moreover, pDNA was also used as DNA vaccine to regulate tumor‐associated antigens and induce the adaptive immunity aiming at tumor therapy. Cerullo et al. reported the delivery strategy of combination between poly‐epitope pDNA vaccine encoding melanoma tumor‐associated antigens and B16F1‐specific neoantigens, and an intratumorally administrated oncolytic virus.^{[ 131 ]} The synthesized pDNA vaccine produced a large number of antigen‐specific T cells in the spleen to efficiently infiltrate tumor tissue. All in all, pDNA has exhibited long‐time and multiple therapeutic effect in tumor cells. However, the risk of pDNA integrating into the host cell genome and the relatively slow speed in regulating gene expression still influences its further application.

3.2. mRNA Drugs

mRNAs, the single‐stranded nucleic acid carrying genetic information, are transcribed from the template DNA to directly guide the rapid translation of target proteins. The low risk of gene integration mutations and low immunogenicity make mRNA drug a safer choice compared to pDNA drugs. Besides, mRNA directly binds to ribosomes in cytoplasm to initiate translation process without nuclear translocation with higher transfection efficiency. However, mRNA drugs still suffer from a short half‐period and easily degradable instability. Carriers such as liposomes or NPs have successfully improved the stability and targeted ability of mRNA drugs for the facilitation of intravenous administration of mRNA drugs. Besides, appropriate structure modification including codon optimization, nucleotide modification, and self‐amplifying mRNAs, and formulation methods like polymers and peptides also have been widely explored in improving mRNA conveying efficiency.^{[ 132} , ^{133 ]} Next, mediating immunotherapy based on antigen‐presenting cells (APCs) is the primary function of mRNA drugs. After the mRNAs transcribed in vitro were transfected into APCs, they rapidly mediate the translational expression of tumor‐associated antigens to induce the activation of innate immune response or adaptive immune response. Different from peptides vaccine, mRNA drugs are able to encode complete tumor antigens, promoting APCs to present multiple epitopes and stimulating broader T cells activation. Extensive reports have identified various tumor‐associated antigens to provide targets for the development of mRNA vaccines‐induced immune therapy.^{[ 134} , ¹³⁵ , ¹³⁶ , ¹³⁷ , ¹³⁸ , ^{139 ]}

3.3. Noncoding RNA Drugs

Central dogma holds that RNA transcribing from DNA is a key factor in directing protein translation. The important role of mRNA and transfer RNA (tRNA) in protein translation has been focused on for a long period. However, such processes only account for less than 2% of the whole genome.^{[ 140 ]} It has gradually been recognized that a large number of noncoding RNAs are not junks and by‐products of transcription products, but they are even responsible for broader regulation. In the application of nucleic acid drugs, the RNA interference (RNAi) mediated by miRNA and siRNA and the pre‐transcription RNA silencing mediated by shRNA are widely explored to disturb gene expression for suppressing tumor development. RNAi is the efficient and specific homologous mRNA degradation and expression interference induced by a highly conserved double‐stranded RNA, based on Watson Crick's principle of base complementary pairing. It was originally discovered in antisense RNA studies of the Caenorhabditis elegans.^{[ 141 ]} With the discovery of RNAi in various eukaryotes and technological innovations, it has been developed as a far‐reaching tool for studying genes and diseases. Moreover, some tumor development‐associated genes can also be effectively silenced to inhibit tumor progress through RNAi.

miRNA is a double‐stranded noncoding RNA molecule with a length of 20–24 nt with high homology and conservatism in animals. Pri‐miRNA in short hairpin structure is the earliest forms transcribed by DNA, then it is processed into pre‐miRNA and conveyed out of nucleus through transporter exportin‐5. In cytoplasm, pre‐miRNA is cut and processed into final double‐stranded short miRNA molecule by Dicer. RNA‐induced silencing complex (RISC), consisting of Argonaute‐2 (AGO2) and guidance RNA strands, will sensitively recognize the mature miRNA to hydrolyze the sense chain in miRNA and search for target mRNA chain in the guidance of the antisense chain in miRNA.^{[ 142} , ^{143 ]} After combining with target mRNA, miRNAs next mediate the cleavage degradation effect or translation decay effect depending on whether the miRNA/mRNA complex is fully complementary, without influencing the bioactivity. Based on this mechanism, various miRNA drugs have been explored to inhibit tumor growth through RNAi. Various vectors including NPs carriers, micelles, liposomes, and dendrimers have been applied for conveying tumor suppressor miRNA.^{[ 144 ]} For instance, miR‐34 has been indicated to be dysfunction in various kinds of tumors, which could be regulated by p53 directly. In mechanism, miR‐34 serves as an effective tumor inhibitor by negatively regulating the EMT in tumor development and suppressing the metastasis. Therefore, miR‐34 has been used as a nucleic acid drug to suppress different tumor cells through lentiviral vector, lipid‐based vector, and polymeric vector.^{[ 145 ]} Besides, miR‐182‐3p drug loaded in lipid NPs was used in the treatment of triple‐negative breast cancer through miR‐182‐3p inhibiting activity of telomeric repeat‐binding factor 2 (TRF2), resulting in DNA damage at telomeric and pericentromeric sites.^{[ 146 ]} Next, a liver‐specific miRNA, miR‐122 was carried by galactose‐targeted lipid calcium phosphate (Gal‐LCP) nanoformulation to suppress colorectal cancer liver metastasis. The inhibitor effect of miR‐122 was related with the downregulation of several proinflammatory factors, matrix metalloproteinases, and extracellular matrix degradation enzymes. The increased CD8⁺/CD4⁺ T‐cell ratio and decreased immunosuppressive cell infiltration were also induced by miR‐122 to promote antitumor immunity in liver.^{[ 147 ]}

siRNA, a dsRNA molecule of 21–23 nt in length, can perfectly combine as a pair with target mRNAs and mediate the degradation of mRNAs for RNAi. Exogenous siRNA enters target cells by viral transfection or artificial introduction. While endogenous siRNAs are synthesized by RNA‐dependent RNA polymerase using the RNA single strand as a template. Similar to the action process of miRNA, after siRNA entering cytoplasm, it is recognized by Dicer and cleaved into short fragments and exposes the 3' and 5' ends groups on both sides. After further binding to the AGO‐RISC complex, the sense chain of siRNA will be degraded, leaving the antisense chain to carry AGO‐RISC complex and recognize target mRNA for mediating mRNA silencing. In 2018, the siRNA drug, Patisiran, was approved by the Food and Drug Administration for the treatment of transthyretin‐mediated amyloidosis, which achieved the siRNA‐targeted delivery therapy in humans first. Theoretically speaking, the expression of any target genes can be interfered by siRNA through synthesizing siRNA according to target genes’ nucleic acid sequences, especially some “undruggable” genes, which highlights the extensive treating horizon of siRNA drugs compared to normal drugs in clinical. What is more, siRNA drugs‐induced transient genetic disruption could elicit controllable and durable therapeutic safety, rather than producing the risk of changing the genome of patients induced by shRNA‐mediated nucleic acids therapy and CRISPR‐Cas9 system‐mediated gene editing. Currently, siRNA therapy has been widely used in cancer treatment, with the representative applicational direction including inhibiting EMT,^{[ 148 ]} inhibiting angiogenesis,^{[ 149 ]} inhibiting invasion and metastasis,^{[ 150 ]} enhancing chemotherapy sensitivity,^{[ 151 ]} and immune therapy.^{[ 152 ]} However, the undesirable stability induced by ubiquitous RNase in blood circulation and the potential off‐target effects are the deficiency that needed to be solved in siRNA drug therapeutic strategy. Some classic siRNA drug delivery systems such as lipid NPs and organic polymers have been widely reported to compensate for the shortcomings.^{[ 153 ]}

All in all, the delivery strategy of exogenous nucleic acid drugs into tumor cells to fill in defective genes or correct aberrantly expressed genes has been widely accepted. However, the instability of nucleic acid molecules in blood circulation system is still an urgent shortcoming in nucleic acid therapy. Due to the ubiquitous high concentrated DNase and RNase, nucleic acid drugs without any modifications only elicit a short half‐period with the risk of easily cleared by kidneys. Besides, the negative charges, huge molecular weight, and hydrophilicity of nucleic acid molecules give rise to poor cell membrane penetration, creating a significant challenge for nucleic acid drugs uptake of tumor cells. Carriers including liposomes, polymers, and metal NPs can facilitate the long circulation time of nucleic acid drugs and enhance the uptake rate by tumor cells. Furthermore, the explorations of some targeted modification and environment‐responsive strategies efficiently enhance nucleic acid drugs delivery in vivo. Nucleic acid delivery systems response to some specific abnormal characteristics in tumor cells including high ROS/GSH concentration, low pH, and high fluid osmolarity can significantly improve the therapeutic efficiency of nucleic acid drugs.

At last, nucleic acids drugs have been widely used in clinical transformation to treat tumor patients with gene level regulation. Recent typical clinical trials of nucleic acid drugs in tumor therapy are listed in Table  1 .

Table 1.

Recent typical clinical trials of nucleic acid drugs in tumor therapy.

Reference	Nucleic acids	Application	Phage	Trial registration number
[154]	Endogenous T cell receptor siRNA	Highly expressed NY‐ESO‐1 solid tumor	I	NCT02366546
[155]	Bcl2L12 siRNA	Glioblastoma	0	NCT03020017
[156]	Polo‐like kinase 1 siRNA	Advanced hepatocellular carcinoma	I	NCT02191878
[157]	Self‐amplifying mRNA neoantigen	Microsatellite‐stable colorectal cancer	I	NCT03639714
[158]	Four nonmutated RNA	Melanoma	I	NCT02410733
[159]	mRNA neoantigen	Pancreatic cancer	I	NCT02316457
[160]	BNT162b2 mRNA vaccine	Chronic lymphocytic leukemia	‐	NCT04862806
[138]	mRNA vaccine	Gastrointestinal cancer	I/II	NCT03480152
[161]	CT7, MAGE‐A3, and WT1 mRNA	Antimyeloma	I	NCT01995708
[162]	IL‐12 pDNA	Triple‐negative breast cancer	‐	APB‐3266‐NT0220
[163]	IL‐12 pDNA	Advanced epithelial ovarian cancer	I/II	NCT03393884
[164]	hTERT/IL‐12 pDNA	Pancreatic cancer	I	NCT04367675
[165]	IL‐12 pDNA	Advanced melanoma	II	NCT01502293
[166]	IL‐12 pDNA	Quiescent Melanoma	II	NCT02493361
[167]	RB94 pDNA	Genitourinary cancers	I	NCT01517464

4. Engineered NPs for Nucleic Acids Delivery in Cancer Therapy

Engineered NPs are extremely suitable for carrying nucleic acid drugs to protect them from DNase/RNase hydrolysis and glomerular filtration and convey them to tumor cells with high tumor tissue accumulation and cellular uptake rate. Besides, negatively charged nucleic acids exhibit resistance to targeted negatively charged cell membrane, leading to difficulty in intracellular delivery. Overall, ideal vectors for nucleic acid drugs should elicit the characteristics in i) protecting nucleic acid drugs from nuclease hydrolyzation and endothelial reticular capture in vivo, ii) promoting intracellular penetration to enhance transfection rate, iii) helping nucleic drugs escape from endosome intactly, iv) avoiding nonspecific adsorption and targeting tumor cells effectively. The typical physiological barriers that obstruct the engineered NPs from getting into tumor cell cytoplasm are mentioned in Figure  5 .

Figure 5.

The physiological barriers that engineered NPs need to overcome during delivering nucleic acid drugs into tumor cell. a) Having excellent biocompatibility without activating serious immune response, long circulation period, and protecting nucleic acid drugs from nuclease degradation. b) Escaping from blood circulation through the endothelium channel. c) Highly accumulating at tumor site. d) Showing excellent ability in penetrating tumor cell membrane. e) Mediating endosomal/lysosomal escape to protect cargoes from being degraded. f) Degrading in response to stimuli and rapidly release nucleic acid drugs.

To overcome the obstructions in delivering nucleus drugs, various engineered carriers have been explored for nucleic acids therapy mainly including virus vector and nonvirus vector. The high transfection rate virus including retrovirus, adenovirus, adeno‐associated virus (AAV), and herpes simplex virus have been modified to induce nucleic acids therapy in multiple diseases.^{[ 168} , ¹⁶⁹ , ¹⁷⁰ , ^{171 ]} However, the viral genes also elicit the risks of activating severe immune response and integrating into patient's genome, causing gene mutations or proto‐oncogene activation. The limited nucleic acid loading rate and expensive operation cost also affect the further development of viral vectors. Therefore, the low toxicity and low immunogenicity nonvirus NPs vector including liposome,^{[ 172} , ¹⁷³ , ¹⁷⁴ , ^{175 ]} membrane‐coated NPs (CNPs),^{[ 176 ]} inorganic NPs (such as mesoporous silica NPs,^{[ 177} , ¹⁷⁸ , ¹⁷⁹ , ¹⁸⁰ , ¹⁸¹ , ^{182 ]} oxide NPs,^{[ 183 ]} and gold NPs^{[ 184} , ^{185 ]}), lipid NPs,^{[ 186} , ^{187 ]} and polymeric NPs^{[ 188} , ¹⁸⁹ , ^{190 ]} delivering systems have attracted much attention in nucleic acids therapy. The above nucleus acid drugs vectors elicit durable stability in blood circulation, providing a relative security conveying environment for nucleus acid drugs. Next, on accounts of the abnormal permeability of lymphatic vessels and blood vessels‐induced high interstitial fluid pressure (IFP) around solid tumor tissue and the enhanced permeability and retention effect (EPR effect), NPs can be easily detached from the vascular bed and accumulated highly in the tumor mesenchyme. Furthermore, the specific characteristics in TME such as low pH level, high concentration of ROS/GSH, and high IFP can be considered as the response condition for designing nucleus acid drugs carrier. Some molecular antigen on tumor cells or immunocytes can be designed as targets through decorating NPs with their ligands to achieve efficient targeted delivery. Herein in this part, we will focus on the common nucleic acids therapy nanovectors and review their advantages and shortcomings, respectively (Figure  6 ).

Figure 6.

Model diagram of common nanovectors for nucleic acid delivery.

4.1. Liposome‐Based Vector

Liposomes were the most universal vector applied in nucleic acid drugs over the years.^{[ 191} , ^{192 ]} The polar head groups and nonpolar tail phospholipids‐induced amphiphilicity lead to its capability in spontaneous formation of lipophilic inner compartment. Bilayer liposomes are conventional carriers used to deliver small molecules, proteins, and nucleic acid drugs, allowing hydrophilic cargo encapsulated inside the liposome inner core and hydrophobic cargo wrapped inside the liposomal bilayer hydrocarbon chain. After loading the therapeutic cargo, liposomes can also be maintained at the appropriate nanometer size, which is the vital parameter deciding the drug loading rate and circulating half‐period. Liposomes in nanoscale are able to escape from the clearance of phagocytes and target tumor cell efficiently.

Among various kinds of applied liposomes, cationic liposomes (CLs) with the sensitivity to anionic polyelectrolytes are popular in encapsulating negatively charged therapeutic molecules, especially DNA/RNA molecules. CLs, consisting of cationic head, spacer, linker bond, and hydrophobic tail, are easily combined with both nucleic acid drugs and negatively charged plasma membrane to adsorb to tumor cell membrane and mediate intracellular uptake. The electrostatic attraction between the cationic head in liposomes and the negatively charged phosphate in nucleic acid molecules mediated the solid loading effect. Due to the positive charge overload of liposomes, the negative charge in nucleic acid molecules is completely offset, endowing the liposome‐nucleic acid complex positive charge to target tumor cell membrane. For example, cationic (2,3‐dioleoyloxy‐propyl)‐trimethylammonium‐chloride has been widely explored for delivering nucleic acid drugs in cancer therapy.^{[ 193} , ¹⁹⁴ , ^{195 ]} Furthermore, modifying cationic liposomes with polyethylene glycol (PEG) is effective to further prolong the half‐life of liposomes in blood circulation,^{[ 196} , ¹⁹⁷ , ^{198 ]} because PEGylation avoids the serum protein nonspecific binding‐induced dysfunction and IL‐6/TNF‐α‐induced immune reaction through activating Toll like receptor‐4. All in all, liposomes have been widely used for nucleic acid or drug delivery due to their positive charge conferred nucleic acid molecules binding ability, targeted cell membrane function, and endosomal escape ability. However, the complex preparation processes and expensive usage costs always limit their clinical development. The unpromising drug loading rate also leads to the excessive need of CLs concentrations in higher drug doses therapy, resulting in the indiscriminate cytotoxicity including disruption of cell membrane integrity, reduced cellular activity, and severe phagocytic response.^{[ 199 ]}

4.2. Lipid NPs‐Based Vector

Lipid NPs (LNPs) are a popular vector in nucleic acid drugs therapy which yield extraordinary effect in the current mRNA vaccine development for COVID‐19.^{[ 132} , ^{200 ]} To avoid the cell membrane cytotoxicity caused by sustained positive charge in liposome‐based carrier, LNPs with acquiring positive charge response to the acidic environment were designed to deliver nucleic acid drugs. The ionizable lipids compose the basic framework of LNPs to combine with the negative charge on nucleic acid and keep the LNPs–nucleic acid complex in electric neutrality, facilitating adsorbing to tumor cell membranes and escape from endosome. Amphiphilic phospholipids, cholesterol, and PEGylated lipids serve as auxiliary components to fill the ionizable lipid framework. On account of the amine groups in ionizable lipids are protonated to elicit cationic phenotype, LNPs in acidic environment can recognize and catch the phosphate group in nucleic acid drugs and become stable nanostructure with electric neutrality in physiological solutions.^{[ 201 ]} Therefore, the low toxicity and immunogenicity LNPs were prepared through rapidly mixing the lipid ethanol solution with acidic nucleic acid drug buffer and dialyzing the mixtures with the buffer in pK _a higher than the ionizable lipid. And the pK _a of the ionizable lipid was also accepted as a hopeful reorientation to minimize the risk of off‐target in LNPs‐nucleic acids therapy.^{[ 202 ]} Furthermore, after intravenous administration, LNPs are able to adsorb to the tumor cell membrane for inducing efficient cellular internalization and are encapsulated by early endosome. After the endosome maturation, the acidic environment generated by the fusion of late endosomes with lysosomes induces the re‐protonation of ionizable lipids on the surface of LNPs, causing the depolymerization of LNPs and the efficient release of nucleic acid drugs. As the first LNPs‐mediated RNAi drug, Onpattro, introduced in 2018 indicated wide applicational prospects, LNPs have been widely developed as delivery vehicles for nucleic acid drugs in tumor therapy.^{[ 203} , ²⁰⁴ , ²⁰⁵ , ^{206 ]} Of course, LNPs are not infallible. Pain and inflammatory reactions after intravenous administration of LNPs are not a rare case, which is regulated by interleukin 1 (IL‐1)‐interleukin 1 receptor antagonist (IL‐1ra) axis.^{[ 207 ]} Next, LNPs displayed poor thermal stability, which is susceptible to thermodynamic factors, resulting in serious loss of nucleic acid drugs and decreased transfection efficiency. Last but not least, large amounts of LNPs finally accumulate in liver after injection due to low‐density lipoprotein receptor‐induced absorption into live cell. The deficiency of excellent organ targeting has always been a roadblock to efficient delivery of LNPs. To overcome the flaw, Siegwart et al.^{[ 208 ]} reported a selective organ targeting strategy to exclusively edit extrahepatic tissues through supplying molecules and change the internal charge to enhance the molar compositions in LNPs, which successfully altered LNPs’ fate of distribution and achieved systematically engineered LNPs targeted to organs they aimed at.

4.3. Membrane‐Coated Vector

Membrane‐coated NPs (CNPs) consist of a synthetic nanointernal core and a natural bio‐membrane outer layer. The editable membrane coat was designed to load functional proteins or lipids via gene engineering, artificial modification, and metabolism engineering, allowing CNPs to interact with proteins, cells, and molecules in vivo. At the same time, the long half‐period in circulation system and immune escape ability better than PEGylation modifications are also obtained.

Currently, biological membranes such as erythrocyte membranes, platelet membranes, immune cell membranes, tumor cell membranes, and bacterial outer membranes have been developed for the synthesis of CNPs. The erythrocyte membrane coat provides a circulating retention capacity up to 72 h. Platelet membrane coat endows good circulatory capacity and tumor targeting ability in vivo. Tumor cell membrane coat provides an excellent targeting ability to the homologous tumor tissue. Bacterial outer membranes coat also activates immune response. Most importantly, due to the outer membranes of CNPs are homologous to target cells, the membrane coats elicit powerful cloaking and immune escape capabilities.^{[ 176} , ^{209 ]}

As an example of CNPs in delivering nucleic acid drugs, Zhang et al. reported a platelet cell membrane‐coated metal‐organic framework (MOF) nanocarrier to target deliver anti‐survivin siRNA. Platelet cell membrane coat brought excellent tumor target ability for MOF‐siRNA NPs, significantly improving gene transfection efficiency.^{[ 210 ]} Noteworthy, bacterial outer membranes coat‐mediated mRNA vaccines have been widely explored to induce tumor immune therapy.^{[ 211} , ²¹² , ^{213 ]} However, inflammatory response activated by xenogenic proteins on the outer cell membrane and biosafety issues also limit CNPs' further development. Instability in functional state and difficulty in long‐period storage of CNPs are the drawbacks needed to address.^{[ 209} , ²¹⁴ , ^{215 ]}

4.4. Inorganic NPs‐Based Vector

Inorganic NPs in low bio‐toxicity like gold NPs, mesoporous silica NPs, and iron oxide NPs are popular in delivering nucleic acids with safety and efficacy.^{[ 199 ]} Inorganic NPs carrier exhibits good biosecurity, stable chemical property, and modifiability in combining nucleic acids, facilitating delivering nucleic acid drugs to tumor tissue efficiently.

Gold NPs (Au NPs) exhibit easy surface modification capability and high specific surface area, indicating the excellent capability of adsorbing nucleic acid drugs through mercaptan group covalent connection.^{[ 216 ]} Laser‐activated photothermal effect based on localized surface plasmon resonance achieves the auxiliary therapy along with the nucleic acid drugs. Based on the excellent optical properties, Rotello et al. designed photolabile AuNPs to provide light‐triggered nucleic acid release from Au NPs carrier, which mediated high level of target gene expression recovery.^{[ 217 ]} Some ligands to target cell antigens can also be modified with Au NPs surfaces to improve the efficiency of targeted delivery in vivo.^{[ 218} , ^{219 ]}

Besides, mesoporous silica NPs (MSNs) exhibit superior nucleic acid drug loading capacity due to the noncovalent interaction between nucleic acid and MSNs. The load rate and release effect of nucleic acid drugs can be regulated by the adjustable mesopore structure, void size, and surface function. Broader modification with stimuli‐responsive groups, polymers, and proteins further enhances the loading capacity and releasing properties.^{[ 220 ]} According to different nucleic acid drugs, the pore size of MSNs can be adjusted to mediate appropriate loading capacity and release efficiency. For example, the small‐hole MSNs are suitable for delivering small molecules like siRNA, the large‐hole MSNs are suitable for delivering large molecules like pDNA. Functional molecules such as cationic organics, inorganic substances, and polymers can be edited onto MSNs to provide a personalized prospect for more intelligent designment, including targeted conveying, on‐demand release, and synergistic therapy.^{[ 221} , ²²² , ^{223 ]} Through cationic phenotype decoration like amination modification, cationic polymer functionalization, and metal cations co‐delivered vector, the ionization of surface silanol groups‐generated negative charges in MSNs was offset, enhancing the capacity of loading nucleic acids drugs.^{[ 220 ]} However, further MSN application is blocked by their difficultly controllable particle size uniformity, poor dispersibility, and nondegradability in vivo.^{[ 224} , ^{225 ]}

Next, iron oxide NPs including Fe₂O₃ NPs and Fe₃O₄ NPs also serve as superior nucleic acids carrier with super paramagnetism. After engineered modification with iron oxide NPs for obtaining cationic phenotype, nucleic acids can be loaded in iron oxide NPs through electrostatic adsorption.^{[ 226} , ²²⁷ , ^{228 ]} Based on the properties like high transverse relaxation, thermal energy conversion capacity, and Fenton reaction catalytic potential, various auxiliary function including magnetic resonance imaging (MRI)‐monitored biodistribution, magnetic field‐guided targeted conveying can be introduced in iron oxide NPs. Nucleic acids therapy combing with MRI monitoring,^{[ 229} , ²³⁰ , ²³¹ , ²³² , ^{233 ]} lncRNA therapy combing with photothermal therapy,^{[ 234 ]} and siRNA therapy combining with Fenton‐reaction therapy and MRI monitoring^{[ 235 ]} have been reported to exert efficient multimodal nucleic acids strategy in cancer therapy. All in all, inorganic NPs have been widely used for nucleic acid delivery based on their good biocompatibility and multifunctional modifiability. However, the poor degradability and nonnegligible toxicity of inorganic materials limit the further application of high dose nucleic acid delivery in vivo. Some biodegradable nucleic acid vectors need to be developed to undertake a wider range of drug delivery.

4.5. Polymer‐Based Vector

Polymeric NPs (PNPs) have been applied in delivering nucleic acids drugs due to the excellent biodegradability, biocompatibility, and functional modification potential. PNPs can be classified into natural polymer carriers (chitosan, glucan, peptides. and proteins) and synthetic polymers (polyethylenimine (PEI), polylactic acid (PLA), poly(lactic‐co‐glycolic acid) (PLGA), and poly(amido amine) (PAMAM)). Based on the structure, PNPs could also be divided into polyplex, polymersome, and dendrimer.

4.5.1. Natural Polymers‐Based Nucleic Acid Vector

Natural polymers are considered as potential candidates for delivering nucleic acids drug in tumor therapy, due to their excellent bio‐compatibility, easy surficial modification, and low immunogenicity. For example, low molecular weight peptide was used for efficient delivery through conjugating nucleic acids drug via covalent affixation and noncovalent complexation. The peptide‐induced nucleic acid delivery system significantly enhances membrane penetration and targeting ability, and the complex preparation process and off‐target effects are also needed to be optimized. Besides, chitosan is a kind of natural polymer with amine groups in structure, thus having positive charges to absorb negatively charged nucleic acids, especially mRNA drugs. Its outstanding biodegradability, mucoadhesive, and permeability‐enhancing properties lead to fewer accumulations in body and better security. However, the transfection efficiency of chitosan is indicated to depend on various formulation parameters like molecular weight, degree of deacetylation, and positive/negative charge ratio of chitosan and nucleic acids. Further chemical modification and additive incorporation of chitosan need to be explored to increase bad solubility of polyplex, which mainly impacts delivery efficiency.^{[ 236} , ^{237 ]} Last, the classical transfection enhancement protein, protamine is a natural polymer that has been widely used in delivering mRNA for 60 years due to its outstanding ability in stabilizing RNA and penetrating cell membrane.^{[ 238 ]} However, protamine was indicated to conjugate mRNA so tightly that it impacted the intracellular translation efficiency of mRNA.

4.5.2. Synthetic Polymers‐Based Nucleic Acid Vector

Synthetic polymer is the most common vector applied in the nucleic acid delivery. The polymer with cationic property or obtained cationic property through modification can effectively adsorb nucleic acid molecules with negative charges via electrostatic action and form stable hybrids in nanoscale under physiological solution, thereby encapsulating nucleic acids in the polymer matrix and protecting drugs molecules from nucleases.^{[ 239 ]} With the wrapping of cationic polymer, nucleic acid molecules also become electrically neutral for more conducive internalization of target cells. Smarter delivery system can also be constructed through modulating the pK _a of amino groups and linking functional groups amide and carbamate via covalent bonds.^{[ 240 ]} Next, in order to solve the problems of instability and bad targeted ability, the hydrophilic polymer PEG is designed to link to the polymer carrier such as PEI during the preparation process. The synthesized PEG‐PEI NPs display good biocompatibility, long circulation period, and low cytotoxicity. What is more, some modifications with antibodies and ligands can also improve the targeting ability and transfection efficiency of PNPs vector. After the polymer‐based carrier internalized by the target cells, the ability of escaping from endosomes is another challenge needed to be considered. At the end of endosomal maturation, lysosomes will fuse with the late endosomes, causing the decreased pH values and nucleases‐mediated degradation of nucleic acid drugs.^{[ 241} , ^{242 ]} Regulation of the pK _a of amino groups in polymers and the proton sponge effect induced by protonation lysosomes have been accepted as the basic mechanism of cationic polymers‐mediated endosome escape.^{[ 240} , ^{243 ]} In conclusion, the synthesized polymer‐based nucleic acid drug delivering system exhibits superior stability in circulation, accurate targeted ability, endosome escape, and degradation response to outer stimuli to release intact nucleic acid drug molecules.

Recently, the synthesized polymer‐based nucleic acid vectors have been widely used in tumor therapy, significantly in enhancing delivery efficiency and exerting additional smart function. For example, a PLGA‐based vector with the decoration of tumor cell membrane proteins was designed to deliver mRNA drugs and sonosensitizer chlorin e6 (Ce6) into APCs in lymph nodes at high efficiency. The tumor cell membrane proteins and mRNA translation products provided neoantigens of metastatic cancer in advance. Under ultrasound irradiation, Ce6 facilitated endosomal escape of mRNA and the augments antigen presentation was enhanced. Thus, antitumor immunity was elicited in advance to inhibit tumor metastasis.^{[ 244 ]} Besides, fluoroalkane‐grafted polyethylenimine (F‐PEI) has been synthesized to deliver tumor antigen‐encoding mRNA into dendritic cells (DCs), facilitating their maturation and amplifying tumor immune. F‐PEI not only exerts the ability of penetrating across the lipid bilayer of cell membranes as well as endosomal/lysosomal membranes, but also acts as an agonist for the Toll‐like receptor 4 (TLR4)‐mediated signaling pathway to trigger DCs activation.^{[ 245 ]} Next, an acidity‐triggered NP (NPs) with size reduction and charge switchable features was synthesized to deliver indoleamine 2,3‐dioxygenase 1 siRNA (IDO1 siRNA) into tumor cells. The component 2,3‐dimethylmaleic anhydride‐grafted poly(ethylene glycol)‐poly(L‐lysine) copolymer was detached from nagatively charged large‐size NPs to enhance tumor penetrating ability. The formed positively charged small size NPs subsequently mediated a rapid lysosomes escape and efficiently released mitoxantrone (MIT) and IDO1 siRNA.^{[ 246 ]}

Our team have developed a series of TME‐triggered polymeric vectors for nucleic acid delivery (Figure  7 ). For example, we prepared PEG‐PLGA to encapsulate and deliver siRNA/mitoxantrone (MTO) complex based on the interaction between the positively charged MTO and the negatively charged siRNA. The CD47‐targeted siRNA (siCD47) could silence the CD47/SIRP α axis to shield the “do not eat me” signaling between macrophages and tumor cells, resulting in the increase of macrophages‐killing effect. MTO also mediated the surface exposure of calreticulin (CRT) which serves as an “eat‐me” signaling to active macrophages. In this PEG‐PLGA delivering system, PLGA was a suitable nucleic acid carrier with good biocompatibility, biodegradability, and controlled degradation. The modification of PEG brought the carrier to better colloidal stability in vivo.^{[ 247 ]}

Figure 7.

Pattern diagram of the TME‐triggered polymeric vectors for nucleic acid delivery in our team. Response to various stimulation such as abnormal hypoxia, pH value, and esterase, the polymeric vector rapidly degraded and released the cargoes in tumor cells.

For example, we developed sensitive hypoxia‐responsive NPs (HRNPs) with efficient tumor targeting ability through self‐assembly of the cationic lipid‐like compound and 2‐nitroimidazole‐modified polypeptide for the delivery of CDC20‐targeted siRNA. In the hypoxic TME, the hydrophobic 2‐nitroimidazole (NI) group was converted to a hydrophilic 2‐aminoimidazole (AI) due to a series of nitroreductases catalyzing a single‐electron reduction, which mediated the rapid disassembly of HRNP and release of intact siRNA. Through the conveyance by HRNPs, oncogene cell division cycle 20 (CDC20) in breast tumor cells was sufficiently silenced, eliciting inhibition of anaphase‐promoting complex/cyclosome (APC/C) E3 ubiquitin ligase to suppress breast cancer development.^{[ 248 ]}

Besides, we have reported a polymeric nanoplatform for Survivin siRNA (siSurvivin) delivery which consists of the ultra‐pH‐responsive polymer Meo‐PEG‐b‐P(DPA‐co‐GMA‐TEPA‐C14) and the tumor‐penetrating polymer iRGD‐PEG‐b‐PDPA. The synthesized siRNA nanoplatform exhibited internalizing RGD (iRGD)‐induced tumor targeting and tissue penetration through targeting integrin αv and neuropilin‐1 on tumor cells, while PEG induced prolonged blood circulation and sharp pH‐responsive ability for endosome escape.^{[ 249 ]}

Similarly, we also prepared a cationic membrane‐penetrating oligoarginine grafts‐functionalized sharp pH‐responsive copolymer consisting of Meo‐PEG‐b‐P(DPA‐co‐GMA‐Rn) for condensing siRNA and ACUPA‐PEG‐b‐PDPA for targeting prostate‐specific membrane antigen (PSMA) receptor. The synthesized polymeric vector could specifically deliver siRNA into PSMA‐expressing prostate cancer cells for effective RNAi therapy.^{[ 250 ]}

Moreover, we synthesized co‐delivery systems for carrying Polo‐like kinase 1 (PLK1)‐targeted siRNA (siPLK1) and MTO prodrug to mediate synergistic treatment. After MTO prodrug co‐assembling with a TME pH‐responsive polymer methoxyl‐poly(ethylene glycol)‐b‐poly(2‐(pentamethyleneimino)ethyl methacrylate) (Meo‐PEG‐b‐PPMEMA), siRNA could be encapsulated into the polymer‐prodrug hybrid NPs through the positive charges on siRNA surface and intrinsic cationic characteristic in MTO prodrugs. Protonation of the hydrophobic PPMEMA segment triggered by low pH value in TME induced the accurate release of siRNA‐prodrug complexes. The complexes exposed in tumor cells were dissociated into intact siRNA and MTO by the esterase‐induced hydrolyzing effect on the amphiphilic structure in prodrug. siPLK1‐mediated oncogene silencing inhibited the mitosis of tumor cells and enhanced the therapeutic effect of chemotherapy drugs.^{[ 188 ]}

Last, we used a pH‐responsive PEGylated polymer Meo‐PEG‐b‐PHMEMA and tumor cell‐targeting and ‐penetrating peptide‐amphiphile (TCPA) to synthesize nanoplatform for conveying bromodomain 4 (BRD4)‐targeted siRNA. In summary, the PEG outer shell endowed siRNA long circulation period for enhancing tumoral accumulation. Hydrophobic poly(2‐(hexamethyleneimino) ethyl methacrylate) (PHMEMA) gave rise to the rapid release of siRNA in TME low pH level. TCPA not only enhanced the targeted ability and intratumoral uptake of siRNA attributed to its RGD ligand segment, but also facilitated cell‐penetrating ability attributed to the cationic polyarginine segment. Above ingredients synergistically enhance siRNA targeting capacity to achieve the efficient BRD4 gene silencing to disturb the recruitment of androgen receptor to target gene loci and suppress prostate cancer progression.^{[ 251 ]}

5. Redox‐Responsive Polymeric NPs for Nucleic Acids Delivery in Cancer Therapy

Degradable polymeric vectors structure for nucleic acid have been widely developed for cancer applications, which display long circulation period, EPR effect‐mediated passive targeting of tumor cells, and considerable biodegradability. However, they fail to disassemble rapidly and release drug efficiently after reaching the tumor lesion, which is the key shortcoming need to be addressed. In order to achieve the high concentration of nucleic acid drug release and ideal tumor therapy effect, we always use a large dose of polymer to deliver nucleic acid drugs, which not only inhibited tumor cells uncontrollably, but also caused unignorable toxicity to nontumor cells. In recent years, the conditional degradation of polymeric vectors‐induced nucleic acid drug release has been widely explored.

Comparing TME to normal tissue, some significantly different physical and chemical states can be designed as activators of polymer degradation, including hypoxia, weak acid, chaotic vasculature, overexpressed enzymes, and redox state.^{[ 252} , ^{253 ]} Based on the abnormal redox response status of cancer cells discussed in Section 2, tumor cells have higher concentration of ROS and GSH. Redox‐responsive nucleic acid drug vectors exhibit rapid drug release and efficient accumulation in tumor tissues with reduced toxicity in normal tissues. Due to the more significant differences in redox states between tumor cells and normal cells, it is appropriate to select ROS and GSH as stimulatory signals triggering drug release from responsive PNPs carriers, rather than other types of stimuli such as pH values and intracellular enzymes.^{[ 254} , ^{255 ]} The redox‐responsive polymeric vectors also exhibit better antihydrolysis ability in circulation system and faster degradation after being stimulated to mediate drug release in high efficacy compared to other stimuli‐responsive nanovectors. Herein, we first focus on the redox‐responsive groups that have been explored in the polymer‐mediated drug delivery (Figure  8 ), then we will detailly discuss the redox‐responsive groups in polymer‐mediated nucleic acid drugs delivery in cancer therapy. All the strategies of mentioned reports in this part are listed in Table  2 , based on their names, polymers, loading nucleic acids, redox‐responsive groups, and therapeutic application.

Figure 8.

The redox‐responsive groups that have been widely applied in the polymer‐mediated drug delivery. Thioketal, chalcogen ether, aminoacrylate, and arylboronic eater break down in response to excessive ROS. Both disulfide and ditelluride break down in response to excessive GSH. Both high level of ROS and GSH trigger breaking of diselenide.  

Table 2.

All the strategies of mentioned reports in this part have been listed, based on structure, polymers, loading nucleic acids, redox‐responsive groups, and therapeutic application.

Nanoplatform	Redox‐responsive groups	Polymer	Nucleic acid	Therapeutic application	Reference
DSPE‐PEG‐PDSA	Disulfide	PDSA	siKIF11/siMYC	Prostate cancer	[256]
			lncAFAP1‐AS1 siRNA	Breast cancer	[257]
			siMGLL/siCB‐2	Pancreatic cancer	[258]
mPEG‐b‐PLA‐PHis‐ss‐bPEI	Disulfide	PLA/PEI	siBcl‐2	Breast cancer	[259]
RGD‐g‐PGA/β‐CD‐ss‐PEI	Disulfide	PGA/PEI	TRAIL pDNA	Colorectal cancer	[260]
Fe3O4‐PEI/HA‐ss‐LWM PEI	Disulfide	PEI	miR‐30a‐5p	Melanoma	[261]
PEG2k‐PEI/PEG2k‐PEI‐ss/PEG2k‐CMPEI‐ss	Disulfide	PEI	Anti‐miR‐21	Ovarian cancer	[262]
pOEI‐PEG‐DHA	Disulfide	PEI	Anti‐miR‐21	Breast cancer	[263]
PSP@MB	Disulfide	PEI	ALDH1 shRNA	Ovarian cancer	[264]
CBA‐DAH‐BG	Disulfide	PAMAM	pDNA	Glioblastoma	[265]
dMSNs/PAMAM‐CD	Disulfide	PAMAM	siBcl‐2	Breast cancer	[266]
mPDV/PDV/DOX/siL	Disulfide	PAMAM	siLDHA	Breast cancer	[267]
CAPL‐ssPBAE	Disulfide	PAE	pDNA	Liver cancer	[268]
PAENs	Disulfide	PAE	EGFP‐shRNA	Glioblastoma	[269]
PAP	Disulfide	PAE	iMdr‐1‐shRNA/iSurvivin shRNA	Breast cancer	[270]
PBAEs	Disulfide	PAE	siGFP	Glioblastoma	[271]
PEI‐ss‐PCL‐ss‐PEI	Disulfide	PEI/PCL	P53 pDNA	Liver cancer	[272]
poly(PEGMA)‐ss‐PCL	Disulfide	PCL	siPLK1	Breast cancer	[273]
NTA‐ss‐PEG‐PCL	Disulfide	PCL	Nrf2 sgRNA	Nasopharyngeal cancer	[274]
PEI‐PLA‐LA	Disulfide	PLA	siSTAT3	Breast cancer	[150]
mPEG‐b‐PLA‐PHis‐ssOEI	Disulfide	PLA	siMDR1	Breast cancer	[275]
OBAE‐PEG‐co‐PLGA	Disulfide	PLGA	siTUBB3	Lung cancer	[276]
PLGA/C16‐S‐S‐PEI	Disulfide	PLGA	siP‐gp	Breast cancer	[277]
Meo‐PEG‐S‐S‐PLGA	Disulfide	PLGA	siCFL1	Liver cancer	[278]
PEG‐ss‐PLL	Disulfide	PLL	siVEGF	Liver cancer	[279]
RGD‐PEG‐PLys‐ss	Disulfide	PLL	siRNA	Glioblastoma	[280]
LMWPEI‐ss‐chitosan	Disulfide	Chitosan	P‐glycoprotein shRNA	Breast cancer	[281]
Trimethyl chitosan‐ss‐nona‐arginine	Disulfide	Chitosan	siVEGF	Liver cancer	[282]
GCS‐PDP/PEI‐SH	Disulfide	Chitosan	pSur‐fLuc	Prostate cancer	[283]
cRGD‐PEG‐PAsp(MEA)‐PAsp	Disulfide	Poly(aspart‐amide)	siPLK1	Prostate cancer	[284]
PBR	Disulfide	Poly(disulfide amine)	siKRAS	Pancreatic cancer	[285]
siRNA‐SS‐PNIPAM	Disulfide	Poly(N‐isopropylacrylamide)	siSTAT3	Glioblastoma	[286]
CA‐PLL‐TK	Thioketal	PLL	siFGL1/siPD‐L1	Lung cancer	[287]
TKPFH	Thioketal	PEI	shGPX4/shMTHFD2	Melanoma/Ovarian cancer	[288]
TK‐PEI	Thioketal	PEI	p53 pDNA	Melanoma	[289]
TKPEI‐Ce6	Thioketal	PEI	siPLK1	Breast cancer	[290]
iRGD‐PEG‐PAsp(TKQA)‐PPhe	Thioketal	PAsp(TKQA)	siGPX4	Breast cancer	[291]
SP‐Crosslinked‐BPEI	Boronic acid ester bond	PEI	siPLK1	Breast cancer	[292]
B‐PDEAEA	Boronic acid benzyl	PDEAEA	pTRAIL	Lung cancer	[293]
B‐PDEAEA	Boronic acid benzyl	PDEAEA	pTRAIL	Lung cancer	[294]
BP–PDM–PG(TPE)	Boronic acid pinacol ester benzyl	BP	CAG‐Luc pDNA	Cervical cancer	[295]
B‐PDEAEA	Boronic acid benzyl	PDEAEA	pTRAIL	Pancreatic cancer	[296]
TAEA‐S‐xF	Thioacetal	TAEA‐S‐xF	pEGFP‐N1	Ovarian cancer/Prostate cancer	[297]
Thioacetal‐linked LMW PEI	Thioacetal	PEI	pGL3	Cervical cancer	[298]
P(TPECM‐AA‐OEI)‐g‐mPEG	Aminoacrylate	PEI	pEGFP	Cervical cancer	[299]
OEI800‐SeSex	Diselenide	PEI	pEGFP /pGL3	Melanoma/Cervical cancer	[300]
OEI800‐SeSex	Diselenide	PEI	pEGFP /pGL3	Melanoma/Cervical cancer/Breast cancer	[301]
OEI‐SeSex/ HA‐SS‐COOH	Diselenide/Disulfide	PEI	pDNA	Cervical cancer	[302]
DSe‐PEI‐F	Diselenide	PEI	pLuciferase	Cervical cancer	[303]
UCNPs‐PEIRB‐PEISeSe	Diselenide	PEI	siPLK1	Liver cancer	[304]
SP‐SeSe‐SP	Diselenide	dPSPs	pEGFP	Breast cancer	[304]
HA‐SeSe‐COOH/siR‐93C@PAMAM	Diselenide	PAMAM	siR‐93C	Lung cancer	[305]

5.1. Advanced Redox‐Responsive Groups in Drug Delivery

5.1.1. Reduction (GSH)‐Responsive Groups in Drug Delivery

Sulfur bonds, including disulfide bonds and trisulfide bonds, are important components of reduction‐sensitive polymeric carriers due to the almost 90° bond angle in sulfur bonds, providing superior flexibility for carrier‐loaded drugs.^{[ 306} , ³⁰⁷ , ^{308 ]} Disulfide bond is a type of sensitive reduction‐response linkage which has been introduce in various drug delivering system. Besides, due to three sulfur atoms and two sulfur‐containing planes, trisulfide bonds exhibits better drug self‐assembly and redox responsiveness than disulfide bonds, inducing drug release rate in more excellent efficacy. Trisulfide bonds possess bigger number of GSH trigger sites, higher reduction potentials, and higher chemical valence in β‐sulfur compared to disulfide bonds.^{[ 309 ]} In sulfur bonds, disulfide bond (─S─S─) has become the most important chemical bonds for achieving the reduction‐responsive drug release. Disulfide bond is the covalent bond between two sulfur atoms formed by the oxidation of two sulfhydryl groups, which is susceptible to be broken by the reduction equivalent like GSH in tumor cells. Therefore, reduction‐responsive polymeric nanovectors are mainly synthesized by introducing disulfide bond linkages or functional groups containing disulfide bonds to achieve efficient drug release. Endogenous disulfide (cystamine, 3,3'‐dithiodipropionic acid, 2,2'‐dithiodiethanol, and cystamine bisacrylamide) and synthesized disulfide (based on lipoic acid (LA), cysteine, pyridyl disulfide group) are the mainly role inducing reduction‐responsive ability.^{[ 310 ]}

Ditelluride is another GSH‐responsive group having higher sensitivity than disulfide. Tellurium and sulfur are in the same group on the periodic table with similar chemical properties. Compared to sulfur atom, tellurium atom has bigger radius and weaker electronegativity, which results in the lower bond energy of the homonuclear and heteronuclear single bond of tellurium, suggesting the ditelluride‐containing polymers are more sensitive to GSH environment. The introduction of ditelluride into polymers to achieve highly sensitive carrier degradation and drug release is still in its infancy. Wang et al. first introduced ditelluride groups into water soluble copolymers poly(ether‐urethane) for the GSH‐responsive controlled release of DOX in tumor cells.^{[ 311 ]} Pang et al. also prepared a folic acid (FA)‐modified PEGylated polycaprolactone‐containing ditelluride linkage to deliver DOX for inhibiting breast cancer cells.^{[ 312 ]} However, ditelluride still has not been introduced in the construction of polymeric nucleic acid vector for enhancing gene therapy or gene silencing efficiency.

5.1.2. Oxidation (ROS)‐Responsive Groups in Drug Delivery

Chalcogen ethers mainly include thioether, selenoether, and telluroether, which are the sulfur, selenium, and tellurium analogs of ether, respectively. Chalcogen ethers exhibit ideal ROS responsiveness for being introduced into polymeric carriers to achieve ROS‐triggered drug release. In oxidant environment, the thioethers‐based polymeric vectors rapidly degrade and accurately release the cargoes because thioether is oxidized to sulfoxide and further to sulfone in oxidant environment. In light of the transition from hydrophobicity to hydrophilicity of thioethers in per stage of oxidation, the ROS‐triggered hydrophilicity switch is the main responsiveness mechanism employed by thioether‐based polymeric vectors rather than being cleaved. Such hydrophilic and hydrophobic transformation ability also endows polymer the ability of self‐assembling into hydrophobic core in physiological environment for loading hydrophobic drug, and rapid degradation in the oxidative environment of tumor cells. Currently, thioether‐containing polymers are mainly synthesized through step‐growth polymerization reactions based on nucleophilic substitution, nucleophilic or radical addition reactions, and chain‐growth reactions based on ring‐opening polymerization strategies. In aqueous solutions of physiological environments, thioether‐containing polymers exhibit self‐assembling ability to form the hydrophobic cores. After grafting hydrophilic groups such as PEG, the obtained block copolymers can form the drugs delivering vectors with ROS‐responsive destabilization and disassembling through hydrophobic/hydrophilic interactions. Next, through introducing thioethers in main chains, side chains, and tail chains, various amphiphilic ROS‐responsive copolymers were synthesized with tunable self‐assembling characters. Thioether has been more widely introduced in polymers and polypeptides to encapsulate anticancer drugs like doxorubicin (DOX), paclitaxel (PTX), and camptothecin (CPT) due to the outstanding hydrophobic property of thioether.^{[ 313 ]} However, the reports about using thioether‐containing polymers as ROS‐responsive vector for delivering nucleic acid drugs are still insufficient due to its limitation from the relative low ROS sensitivity compared to selenoether and telluroether. Selenide has better ROS sensitivity than sulfide compounds due to the larger atomic radius and lower electronegativity of selenium. Key energy of selenoether (244 kJ mol⁻¹) is lower than key energy of thioether (272 kJ mol⁻¹), resulting in selenoether being more susceptible to be oxidated by ROS. Similar to thioether, selenoether also responses to ROS environment through the hydrophilic and hydrophobic transformation in oxidation reaction. Last, tellurium exhibits the lowest electronegativity compared to sulfur and selenium, indicating the higher sensitivity of telluroether in oxidative conditions.

Thioketal is a type of sulfur‐based ROS‐responsive groups synthesized by the condensation between thiols and ketones, which exhibits the sensitivity to a broad spectrum of ROS including ¹O₂, H₂O₂, O₂ ⁻, and •OH. As a common protective group, thioketal is stable in enzyme‐containing environment, acidic environment, and alkaline environment. Only in the oxidative environment can thioketal be dissociated into an acetone molecule and two thiol fragments. Therefore, the thioketal‐containing ROS‐responsive polymer is based on the degradation of the polymer triggered by the breakage of thioketal in tumor cells. Similar to disulfide bonding, thioketal has been introduced in polymers in different sites to achieve various ROS‐responsive functions. Introducing thioketal in the polymeric backbone facilitates the preparation of poly‐prodrug. Grafting thioketal as pendant groups also forms the ROS‐triggered degradation sites. What is more, thioketal is also used in linking hydrophobic groups and hydrophilic groups for constructing ROS‐responsive amphiphilic copolymer as delivering system.^{[ 314} , ^{315 ]} The diseases associated with significant ROS production induced by metabolic reprogramming or excessive inflammatory response including tumors,^{[ 316} , ³¹⁷ , ^{318 ]} acute lung injury,^{[ 319 ]} and colitis^{[ 320 ]} have been the typical targets for the drug delivery therapy of thioketal‐containing polymer.

Boronic acid (BA) has been introduced in polymeric NPs as stimuli‐responsive functional groups and targeted ligands. BA chemistry not only offers abundant potentials of responding to abnormal level of ATP,^{[ 321 ]} ROS,^{[ 322 ]} reduced pH,^{[ 323 ]} and glucose,^{[ 324 ]} but also acts as ligands for targeting sialic acids‐expressed cells.^{[ 325 ]} Acting as a Lewis acid with the highly dictated chemistry by surrounding neighbors, BA has the ability of forming reversible covalent bonds through constructing various types of C─C and C─heteroatom bonds, facilitating the synthesis of bioactive polymers in drug delivery. Besides, BA features vacant p‐orbit center on the boron atom, so the reversible covalent bonds are easily constructed with oxygen and nitrogen nucleophiles.^{[ 326 ]} For example, Pandya et al. prepared polymer‐based NPs‐bearing pinacol‐type boronic ester groups for delivering paclitaxel, which exhibited ideal ROS‐triggered biodegradation and rapid drug release in tumor cells. Recently, Lu et al.^{[ 327 ]} also designed an ROS‐degradable branched poly(β‐amino ester) (PBAE) with built‐in phenylboronic acid in the backbone for the efficient cytosolic protein delivery, which could be degraded into small segments by the highly concentrated H₂O₂ in tumor cells, resulting in the rapid intracellular release of the targeted protein saporin.

The application of ROS stimuli‐responsive gene delivery system in cancer therapy has been investigated widely due to abnormal ROS concentration in tumor cells. In addition to maintaining the stability of drug circulation in the body, the ROS stimuli‐responsive gene delivery system can release gene drugs in tumor cells in response to the differences in ROS levels between tumor cells and normal cells. By selectively enriching drugs in tumor tissue and achieving environmentally responsive drug release in target cells, antitumor efficacy is improved and toxic side effects are reduced. Although the various types of ROS stimuli‐responsive gene delivery system have been successful in vitro development and testing, but many into clinical aspects are lack of the corresponding research and improvement. On the one hand, ROS‐responsive molecules are prone to REDOX side reactions during preparation and preservation to lose their responsive activity, which affects the therapeutic effect. In addition, the design of responsive vectors and the difficulty of scaling up preparation make the transformation of the ROS stimuli‐responsive gene delivery system challenging. The endogenous stimulus response nanometer carrier may be triggered because of the target tissue in patients with different levels caused great individual differences. Therefore, future research needs to solve these new agents' applicability and effectiveness when used.

5.1.3. Reduction (GSH)/Oxidation (ROS) Dual‐Responsive Groups in Drug Delivery

Diselenide has been introduced in polymer vector for ROS/GSH dual‐responsive delivery of nucleic acid drugs in cancer therapy. Diselenide not only can be oxidized to seleninic acid (‐SeOOH) by ROS but also be reduced to selenol (‐SeH) by GSH. Compared to sulfur, selenium has weaker electronegativity and bigger atomic radius, resulting in Se─Se bonds exhibit unique bond energy (172 kJ mol⁻¹) and dynamic character for making diselenides responsive to mild stimuli‐like redox condition in tumor cells.^{[ 328 ]} For example, Wei et al. prepared the diselenide‐centered biodegradable tri‐block copolymers methoxyl poly(ethylene glycol)‐b‐poly(e‐caprolactone)‐b‐methoxyl poly(ethylene glycol) (mPEG–PCL–Se)₂ through ring opening polymerization, which could self‐assemble to polymicelles and load hydrophobic anticancer drug DOX.^{[ 329 ]} In the highly concentrated GSH environment of tumor cells, (mPEG–PCL–Se)₂ rapidly degraded and efficiently released DOX due to the cleavage of the diselenide bond. Besides, Deepagan et al. developed the in situ diselenide‐crosslinked micelles to deliver DOX, which spontaneously derived from selenol‐bearing triblock copolymers consisting of PEG and polypeptide derivatives. The ROS‐rich environment triggered the rapid release of DOX because the hydrophobic diselenide bond was cleaved into hydrophilic selenic acid derivatives.^{[ 330 ]} Wang et al. reported a diselenide‐crosslinked carboxymethyl chitosan NPs (DSe‐CMC NPs) for DOX delivery, which could disintegrate and accelerate the release of DOX by the breakdown of diselenide bonds through oxidation by H₂O₂ or reduction by GSH in tumor cells.^{[ 331 ]}

5.2. Disulfide‐Based Polymer Vector for Nucleic Acid Drug Delivery

In the last decade, polymeric NPs formed by self‐assembly of amphiphilic copolymers have been widely developed in nucleic acid delivery for cancer therapy. The introduction of disulfide bond units endows such vectors' rapid dissociation and efficient cargoes release under reducing environments. In general, there are two methods to introduce disulfide bonds into polymeric hydrophobic cores in a typical core–shell polymer vector. The first idea is the introduction of disulfide bonds in the backbone of hydrophobic core to achieve reduction‐responsive core degradation and drug release, which facilitates endosomal escape and decreases the cytotoxicity. The other one is linking the inner core and outer shell via disulfide bonds or disulfide‐containing groups for the GSH‐triggered polymer degradation, which is designed to prolong blood circulation.^{[ 332 ]} Next, the dissociation products of such polymers also exhibit the advantage of excellent biodegradability and low toxicity. For example, cationic polymers containing disulfide bonds are degraded into cationic fragments with low cytotoxicity in reducing cell environment, and disulfide bonds are reduced to sulfhydryl groups. Compared to other cationic polymers with cell membrane toxicity, disulfide bonds containing polymeric cations exhibit low toxicity and excellent nucleic release rate after disulfide bonds being reduced by GSH.^{[ 240 ]} Introducing disulfide bonds to the internal backbone of the polymer can achieve the reduction‐responsive degradation of polymer vectors. Besides, disulfide bonds can also be used to synthesize the reduction‐responsive sheddable hydrophilic outer shells to protect inner core or the reduction‐responsive degradable inner core to encapsulate nucleic acid drugs through linking the hydrophilic chains and hydrophobic chains of star, block, and graft polymers.^{[ 333} , ^{334 ]} Next, after nucleic acid molecules are attached to PEG or hyaluronic acid through disulfide bonds to form nucleic acid conjugates, the complexes synthesized via electrostatical bonding between cationic polymers and nucleic acid conjugates exhibit better stability than the nucleic acid–cationic polymer complexes without disulfide bonds.^{[ 335} , ^{336 ]} The cellular uptake efficiency of siRNA drugs can also be improved by increasing the branching index of the polymer through the introduction of disulfide bonds.^{[ 337 ]} In this part, we will focus on the disulfide‐based reduction‐responsive polymeric nanovectors for nucleic acid delivery in cancer therapy.

5.2.1. Poly(Disulfide Amide)‐Containing Polymeric Vectors

Our team has developed novel l‐cysteine‐based poly(disulfide amide) (PDSA) for synthesizing reduction‐responsive NPs with considerable hydrophobic drug loading capacity and adjustable potential (Figure  9a). One disulfide part (l‐cystine ester) and one diacid part (aliphatic dicarboxylic acid) are the two typical repeating units without cytotoxicity in PDSA structure. Based on the disulfide bonds, l‐cystine ester displays primitive reduction‐sensitive characteristics. Through introducing diacid segments, polymeric properties including hydrophobicity and density of disulfide bonds can be turner for various applications. We finally chose the direct polycondensation route of cystine esters and diacid monomers to synthesize PDSA and demonstrated the excellent biocompatibility, reductive responsiveness, and superior drug loading capacity of PDSA in vivo and in vitro.^{[ 338 ]} On the basis of the tunable physiochemical properties of PDSA (e.g., structure, molecular weight, and hydrophobicity), we hypothesized that the rapid reduction‐responsive degradation of PDSA can be used to deliver siRNA to achieve accurate drug release in tumor cells and enhance gene silencing. We subsequently designed a PDSA‐based siRNA nanovector consisting of siRNA, amphiphilic cationic lipid G0‐C14, PDSA, and 1,2‐distearoyl‐sn‐glycero‐3‐phosphoethanolamine‐N‐[methoxy (poly ethylene glycol)−3000] (DSPE‐PEG_3k). Cationic lipid G0‐C14 for adsorbing siRNA due to electrostatic effect has been reported in our previous research.^{[ 339 ]} In detail, such polymeric nanovector was synthesized through mixing siRNA aqueous solution with dimethylformamide mixture of PDSA polymer, DSPE‐PEG_3k, and cationic lipid G0‐C14. Since both G0‐C14 and PDSA were hydrophobic, the nano self‐assembly precipitation promoted PDSA to encapsulate G0‐C14 adsorbed with siRNA to form a PDSA inner core in nanoscale. Subsequently, the amphiphilic DSPE‐PEG_3k were used to wrap around PDSA core as outer shell to provide long blood circulation ability and high tumor accumulation ability. After internalization and endosomal escape, the high concentration of GSH in tumor cells could break the disulfide bonds in PDSA to trigger the degradation of polymers and the release of siRNA for tumor therapy, which facilitated gene silencing through siRNA‐induced translation disturbance (Figure 9b).

Figure 9.

a) Synthesis process of PDSA. b) Schematic illustration of the reduction‐responsive NPs made with the cationic lipid‐like compound G0‐C14 and PDSA, DSPE‐PEG, and NPs‐mediated targeted gene silencing for tumor treatment.

For example, we prepared PDSA‐based vector for KIF11 (siKIF11) and MYC (siMYC) siRNA delivery to suppress prostate tumor. In order to achieve the most effective gene silencing efficiency, we also developed a library of PDSA polymers with different molecular weights and chemical structures, based on the values of n in ingredient fatty diacid structure of 2, 4, 6, and 8, respectively. With the increase of n value and PDSA polymeric chain, the siRNA encapsulation efficiency was enhanced but the zeta potential (ζ) and size were decreased, possibly because of that the increased hydrophobicity and the increased PDSA length caused more compact hydrophobic core. Finally, the most superior siRNA vector, PDSA8a‐based NPs, was chosen to respond to the cytosolic reduction stimuli and achieve rapid intracellular siRNA release, resulting in efficient gene silencing of target proteins in prostate cancer therapy.^{[ 256 ]} In another report, we identified a key lncRNA (actin filament‐associated protein 1 antisense RNA1, AFAP1‐AS1) regulating triple negative breast cancer cells radiosensitivity through canonical Wnt/β‐catenin signaling pathway. In order to reverse this resistance to radiotherapy and enhance the efficacy of breast cancer therapy, we delivered lncAFAP1‐AS1 siRNA via the PDSA‐based polymeric nanoplatform mentioned above. High levels of GSH‐triggered polymer degradation and rapid gene release induced an ideal lncAFAP1‐AS1 silencing in tumor cells to improve the radiotherapy efficacy in xenograft and metastatic tumor models.^{[ 257 ]} In our recent report, we successfully applied the PDSA‐based polymeric nanoplatform to co‐deliver two kinds of siRNA [monoacylglycerol lipase siRNA (siMGLL) and endocannabinoid receptor‐2 siRNA (siCB‐2)] simultaneously for synergistic therapy. Through effective gene silencing, the fatty acid oxidation was suppressed and tumor‐associated macrophages was reprogrammed to M1‐like phenotype macrophages to inhibit pancreatic cancer cells.^{[ 258 ]}

Besides nucleic acids delivery, the sensitive GSH‐responsive polymer PDSA also has been designed in other tumoral therapeutic program, indicating its wide application prospect. Ling et al.^{[ 340 ]} screened out the optimized PDSA NPs (referred to as CP5 NPs) to carry Pt(IV) prodrugs. CP5 NPs showed a small particle size (76.2 nm), high loading of Pt(IV) prodrugs (15.50% Pt), and the rapid release of platinum (Pt) ions, significantly inhibiting the growth of cisplatin‐resistant xenograft tumors while alleviating serious side effects associated with cisplatin. Besides, Wang et al.^{[ 341 ]} also used PDSA to deliver a histone methyltransferase G9a inhibitor (UNC0638), to simultaneously block GSH biosynthesis and clear cellular GSH levels in pancreatic ductal adenocarcinoma. UNC0638 nanodrug realized the GSH‐controlled drug release and better tumor inhibited ability, compared to UNC0630 alone. Next, PROteolysis TArgeting Chimeras (PROTACs) is a novel antitumor strategy through hijacking the ubiquitin‐proteasome system to selectively degrade intracellular proteins. To improve the delivery efficiency, Liu et al.^{[ 342 ]} reported nanoengineered PROTACs, Nano‐PROTACs, which were synthesized via encapsulating PROTACs into PDSA. As a proof of concept, nanoengineered BRD4 degrader ARV‐771 was conveyed by PDSA to enhance BRD4 degradation and downregulate the downstream oncogene c‐Myc, showing a superior antitumor efficacy with a low‐dose administration and good biocompatibility. Last but not least, Chen et al.^{[ 343 ]} innovatively encapsulated black phosphorus quantum dots into PDSA to develop a robust photothermal optical coherence tomography agent with rapid GSH‐responsive release and photothermal transduction‐mediated high‐resolution bioimaging. In conclusion, lots of studies have validated that PDSA is a promising nanoplatform to encapsulate various functional agents, achieving efficient delivery and conditional release. Although the use of PDSA nanoparticles for nucleic acid delivery has been successful in vitro development and testing, but many into clinical aspects are lack of the corresponding research and improvement. On the one hand, PDSA are prone to REDOX side reactions during preparation and preservation to lose their responsive activity, which affects the therapeutic effect. In addition, the difficulty of scaling up preparation of PDSA may make the transformation of PDSA nanoparticles for nucleic acid delivery challenging.

5.2.2. Poly(Ethyleneimine)‐Based Disulfide‐Containing Polymer Vectors

PEI, a type of efficient gene vector has been considered as a “gold standard” among all polycations, due to its outstanding nucleic acid loading capacity and endosomal escaping ability. Positively charged PEI can encapsulate nucleic acids and adsorb target cell membrane through electrical interaction between PEI and the negatively charged components like glycoproteins and proteoglycans. PEI also further induces endosomal escape through proton sponge effect. Besides, the condensation between DNA/RNA molecules and PEI also provides protection from nuclease digestion. The PEIs for nucleic acid vectors are mainly divided into branched PEI (bPEI) with the mixture of primary, secondary, and tertiary amines, and linear PEI (lPEI) with only secondary amines. bPEI has been explored more broadly than lPEI attributed to the low price and the easier modification of bPEI. However, the transfection efficiency of PEI is highly dependent on its branching structure. The efficiency of pure PEI‐mediated transfection is pretty low because the PEI/nucleic acid complexes will bind to erythrocytes, albumin, apolipoprotein and complement components in circulation system, resulting in subsequent blocking and swallowing by the reticuloendothelial system.^{[ 344 ]} Therefore, introducing PEG, fatty acids, and amino acids in PEI modification has been accepted as the functional modification strategies to prolong half‐period of PEI‐based gene vectors.^{[ 345} , ³⁴⁶ , ^{347 ]}

Linking PEI and modification components by the biodegradable crosslinking disulfide bonds not only further enriches the reduction‐responsive release ability of PEI carriers for mediating efficient nucleic acid delivery therapy, but also avoids the cytotoxicity caused by increasing the amount of PEI for ideal therapeutic efficacy. For example, a pH value and GSH dual‐responsive polymeric delivery system composed of PEG‐b‐PLA‐PHis and bPEI linked with disulfide bonds (PEG‐b‐PLA‐Phis‐ss‐bPEI) was prepared. The hydrophobic core of the synthesized polymer was composed of poly‐histidine (PHis), PLA segments, and bPEI. PHis‐mediated hydrogen bonding complexation and PEI‐mediated electrostatic adsorption were applied for adsorbing siRNA. Under acidic environment in lysosome, the protonation imidazole groups caused the hydrophobic to hydrophilic transition, which accelerated the degradation of polymer and siRNA release. Both PHis and PEI also mediated proton sponge effects to destabilize endosomal membrane and facilitate siRNA endosomal escape.^{[ 259 ]} Besides, a biocompatible self‐assembly nanovector for co‐delivering pDNA and monensin was constructed. The low‐molecular‐weight PEI (LMW‐PEI) crosslinked by sulfhydryl cyclodextrin through disulfide bonds was applied for adsorbing pDNA through electrical interaction and carrying monensin through host–guest inclusion, following by modifying the polyplex core with poly‐γ‐glutamic acid (γ‐PGA) decorated with RGD through charge interaction. Both RGD‐induced tumor homing and γ‐PGA‐induced tumor‐associated g‐glutamyl transpeptidase (GGT)‐targeted conveying enhanced the delivering accuracy. γ‐PGA also exhibited rapid pH sensitivity and was detached from the polymeric NPs, which resulted in the exposure and breakdown of disulfide bonds under highly concentrated GSH. Therefore, pDNA and monensin could be efficiently conveyed into the tumor cytoplasm with negligible loss.^{[ 260 ]}

Recently, Ma et al.^{[ 261 ]} reported a synergistic therapy against melanoma, which is shown in Figure  10 . The quasi‐mesoporous magnetic nanospheres (MMNs) with cationic polymer‐capped quasi‐mesoporous inner tunnels were constructed through polyelectrolytes‐mediated self‐assembling Fe₃O₄ nanocrystals, followed by cationic polymer capping. Then, a disulfide bond bridged very low molecular weight PEI network was used to modify the outer layer of the MMNs and form redox‐responsive nanospheres (rMMNs), which facilitates miRNA payload and realizes rapid miRNA release under GSH‐dominant TME. Next, miR‐30a‐5p was chosen to be loaded on rMMNs in high loading rate through positively charged PEI. After endocytosis and PEI‐induced endosomal escape, rMMNs rapidly degraded triggered by GSH and released miR‐30a‐5p for inhibiting E2F7 transcription and downregulating caspase‐3 expression, thus suppress tumor cells proliferation, migration, and triggering apoptosis. The thiol‐disulfide bond exchange reaction between disulfide bonds and GSH also depleted GSH in tumor cells, facilitating the Fenton reaction induced by the Fe²⁺ released from the rMMNs core. rMMNs also mediated M1 polarization of monocytes to remodel TME, promoting tumor cells apoptosis.

Figure 10.

Multifunctional rMMNs for gene therapy and immunoenhancement of melanoma.

Furthermore, Javanmardi et al.^{[ 262 ]} synthesized the chemical crosslink composed of branched PEI with disulfide bonds and PEGylated modification reaction (PEG_2k‐PEI‐ss), followed by the subsequent carboxymethylation (PEG_2k‐CMPEI‐ss) for modulating polymer pK _a. The steric shielding, redox‐sensitive PEG_2k‐CMPEI‐ss was used for stably delivering and rapidly releasing anti‐miR‐21 to suppress ovarian cancer cells. Attributed to the crosslinking and carboxymethylation reactions, the protonation of primary amines in PEI at physiologic pH was diminished, facilitating PEI‐induced endosomal escape. Stem from the breakdown of disulfide bonds response to reduction environment in tumor cell, the general cytotoxicity of PEG_2k‐CMPEI‐ss was also significantly decreased. Last, Gong et al.^{[ 264 ]} reported a ultrasound‐targeted microbubble destruction triggerable by poly(ethylene glycol)‐disulfide bond‐PEI‐loaded microbubble (PSP@MB). Local gene delivery for ovarian cancer stem cells (OCSCs) treatment was realized due to the GSH responsiveness, ultrasound triggering, and spatiotemporally controlled release manner. Combined with ultrasound, ALDH1 shRNA plasmid was successfully delivered into OCSCs in high efficiency by PSP@MB, inducing a significant cellular apoptosis.

Disulfide bonds were also introduced into the backbone structure of PEI to offer deeper breakdown and faster drug release response to highly concentrated GSH. Guo et al.^{[ 263 ]} reported a tumor‐accumulative and redox‐responsive PEI‐containing delivering system for anti‐miR‐21. After the disulfide‐bonded polyethyleneimine (pOEI) is conjugated with PEG and subsequently decorated with dehydroascorbic acid (DHA), the obtained polymer pOEI‐PEG‐DHA could condense and encapsulate anti‐miR‐21 nanosphere. In light of the structural similarity between DHA and glucose, the uptake of DHA in tumor cells by GLUT1 was a continuously enriched one‐way transportation, which induced an ideal tumoral accumulation for pOEI‐PEG‐DHA. Anti‐miR‐21 was protected safely from Dicer until the GSH‐triggered polymeric degradation. Therefore, pOEI‐PEG‐DHA successfully mediated an efficient RNAi therapy for triple negative breast cancer.

5.2.3. Poly(Amido Amine)‐Based Disulfide‐Containing Polymer Vectors

PAMAM is a typical dendrimer synthesized by the Michael addition reaction between amine compound and bisacrylamide. Due to the protonation of tertiary amines in the backbone of PAMAM molecules, PAMAM molecules wholly exhibit basicity and positively charged characteristics, facilitating easy loading of nucleic acid drugs and dissolution. The considerable water solubility is a key advantage of PAMAM acting as a drug carrier, induced by the degradation of the amide bonds on its side chains in water solution.^{[ 348 ]} Therefore, PAMAM has been widely developed for the delivery of nucleic acid drugs in cancer therapy. However, the assignable cytotoxicity and slower hydrolysis rate are the issues still need to be addressed. Achieving precise GSH‐triggered degradation by introducing disulfide bonds into the PAMAM structure could accelerate the hydrolysis and nucleic acid release in tumor cells, while simultaneously producing decomposing fragments in lower toxicity. For example, the disulfide‐containing PAMAM (ss‐PAMAM), synthesized by the Michael‐type polyaddition of primary or bis‐secondary amines to N,N’‐cystaminebisacrylamide (CBA), was an effective polymer for gene delivery. What is more, some functional groups or disulfide containing groups were linked to PAMAM including arginine/agmatine,^{[ 349 ]} boronic acid,^{[ 350 ]} arginine‐grafted poly(disulfide amine),^{[ 351 ]} cholesterol,^{[ 352 ]} PEI,^{[ 353 ]} flotillin,^{[ 354 ]} monomers‐containing guanidino groups,^{[ 355 ]} hyaluronic acid,^{[ 356 ]} folate,^{[ 357 ]} branched PEG,^{[ 358 ]} ethylene diamine,^{[ 359 ]} gambogic acid,^{[ 360 ]} cell‐penetrating peptides (TAT),^{[ 361 ]} and the hydrophobic benzoyl groups^{[ 362 ]} for additional features and higher gene delivering efficiency.

Recently, Xing et al.^{[ 265 ]} prepared the polymer scaffold composed of CBA and N‐Boc‐1,6‐diaminohexane (Boc‐DAH) through Michael addition polymerization, followed by N‐Boc deprotection. Next, the exposed primary amino groups were partly transformed into biguanide via addition reaction with dicyandiamide in order to synthesize the product CBA‐DAH‐biguanide (CBA‐DAH‐BG). CBA‐DAH‐BG polymer could condense pDNA molecules in nanoscale in the presence of dithiothreitol. The synthesized CBA‐DAH‐BG/pDNA complexes could degrade in tumor cells triggered by highly concentrated GSH value. The rich primary, secondary, and tertiary amino groups in CBA‐DAH‐BG also mediated ideal endosomal escape effect to enhance gene delivery efficiency.

Besides, Fei et al.^{[ 266 ]} reported a generally applicable strategy that hierarchically integrating the degradable large‐pore dendritic mesoporous silica NPs (dMSNs) and cyclodextrin‐modified PAMAM (PAMAM‐CD) dendrimers, which were designed for effective delivery of antitumor drug SN‐38 and Bcl‐2 siRNA into 4T1 cancer cells. In detail, the ROS‐responsive nitrophenyl‐benzyl‐carbonate (NBC) groups were first used to modify the orifice rim of the dMSN and the disulfide‐bonded azido ligands were grafted onto the inner channel walls through heterogenous functionalization. Next, PAMAM‐CD was interred into dendritic pores via click reactions as the first stage carrier for the hydrophobic drug and siRNA. After that, dMSN was further coated with 4T1 cancer cell membrane (CCM). Attributed to the intelligent designments, the CCM outer coat induced an efficient cancer cell internalization, and the ROS‐triggered NBC removement in lysosome also mediated lysosomal escape through electrostatic deshielding. The redox‐liable disulfide linkers were further cleaved by GSH in tumor cytoplasm, resulting in rapid cargo release. In this strategy, merits of both MSN and PAMAM were combined to achieve synergistic therapy. The inorganic dMSNs potentiated the immobilization of the PAMAM‐CD molecules. The shortcomings of low blood circulation stability and the weak binding force due to the negative charges of dMSNs were also eliminated.

Next, Xia et al.^{[ 267 ]} reported a nanoassembly self‐assembled by three GSH‐responsive polymers, employing poly(δ‐valerolactone) as hydrophobic segment and 3,3'‐dithiodipropionic acid as linkage to connect hydrophilic segment. DOX is encapsulated in the inner core and LDHA siRNA is effectively compressed by cationic PAMAM. Benefiting from introducing c(RGDfk) (RGD) ligand modification, cellular internalization and tumor‐homing are enhanced by recognizing an integrin (αvβ3) on tumoral cell membrane. After escaping from lysosomes, the nanoassembly could be degraded by thiol‐disulfide exchange under intracellular GSH environment, triggering rapid release of DOX and siLDHA. Downregulation of LDHA inhibited expression of tumor‐derived cytokines like granulocyte‐colony stimulating factor (G‐CSF) and granulocyte‐macrophage colony stimulating factor (GM‐CSF), further suppressed the recruitment of myeloid‐derived suppressor cells to remodel TME.

5.2.4. Poly(β‐Amino Esters)‐Based Disulfide‐Containing Polymer Vectors

Poly(β‐amino esters) (PAE) is a group of amino groups‐containing cationic polymer as an efficient nucleic acids vector in gene therapy because its cationic characteristic gives rise to the secure condensation between PAE and pDNA. PAE is synthesized from the Michael addition between an acrylate and an amine, so it inherently exhibits the properties of both tertiary amines and esters including pH‐responsiveness and biodegradability.^{[ 363 ]} Due to the pK _b value of about 6.5 of tertiary amines, PAE has a weak basic characteristic and becomes water‐soluble below pH 6.5, compared to its stable structure in neutral pH.^{[ 364 ]} Therefore, in acidic environments such as tumor cell cytoplasm, the tertiary amines in the PAE structure are transformed into hydrophilic structures due to protonation, which leads to the disintegration of the PAE core. Currently, PAE has been widely accepted as a pH‐responsive nucleic acids carrier. In particular, protonation of PAE tertiary amines in lysosomes also resulted in continuous pumping of protons and chloride ions in lysosomes, causing rupture of lysosome membrane and lysosome escape of the cargoes. However, relatively long degradation times of PAE (a few hours in pH 7 and a few days in pH 5) lead to inadequate intracellular release of cargoes, which limits the efficiency of tumor therapy. Introducing disulfide bonds in the backbone of PAE or linking other functional groups to PAE via disulfide bonds accelerates the degradation of PAE‐based polymeric vectors.^{[ 365 ]} For example, the disulfide bonds‐containing PAE‐based polymeric vectors have been developed recently to achieve the redox/pH dual‐responsive drug release, which has been shown to improve therapeutic effect in cancer cell inhibition.^{[ 364} , ³⁶⁶ , ³⁶⁷ , ³⁶⁸ , ³⁶⁹ , ³⁷⁰ , ³⁷¹ , ³⁷² , ^{373 ]}

In nucleic acid drug delivery for cancer therapy, disulfide bonds‐containing PAE vectors have been attempted and designed to condense various nucleic acid drugs including pDNA,^{[ 268 ]} shRNA,^{[ 269} , ^{270 ]} and siRNA.^{[ 271 ]} In particular, siRNA molecules were smaller in size and harder in rigidity than pDNA, so PAE vectors failed to keep stable after condensing siRNA. Encapsulating siRNA and achieving gene silencing effectively require further chemical modification to PAE. Kozielski et al.^{[ 271 ]} successfully prepared an effective PAE‐based siRNA vector through balancing polymeric bioreducibility and hydrophobicity, as hydrophobicity of PAE could enhance its stability. They also examined the influences of changing nanoparticle formulation and found higher polymer concentrations promoted the enhancement of siRNA delivery.

5.2.5. Poly(ε‐Caprolactone)‐Based Disulfide‐Containing Polymer Vectors

Poly(ε‐caprolactone) (PCL), a kind of hydrophobic aliphatic polyester based on hydroxyalkylic acid, is synthesized from caprolactone monomer by the ring‐opening polymerization reaction catalyzed by metal–anion complexes. This polymer exhibits good biocompatibility, low immunogenicity, and excellent biodegradability which will degrade into 6‐hydroxyhexanoic, a natural metabolite in human body. The freshly synthesized PCL is a nonfunctional hydrophobic solid without any modification potential. Therefore, processing PCL into block copolymers is a promising modification method. Regulating the block ratio and adding new functional blocks is a promising method to modulate amphiphilicity and physical properties.^{[ 374 ]} PCL has been developed to effectively combine with polymers such as chitosan, PLA, PEG, polyglycolide (PGA), polyvinyl alcohol, and polyethylene oxide to provide the functionality required for PCL in drug delivery.^{[ 375 ]} Next, PCL also possesses the ability to degrade gradually in vivo acting as a commonly used drug carrier. However, the degradation lasting for days to weeks makes PCL‐based vector fail to release the drug efficiently and achieve ideal therapeutic efficacy. The introduction of disulfide bonds into PCL‐based copolymers can achieve rapid release of nucleic acid drugs under redox environment.^{[ 376 ]}

For example, a straightforward reaction was proposed for preparing an amphiphilic copolymeric gene vector composed of low molecular weight PEI and PCL (PEI‐ss‐PCL‐ss‐PEI). The reaction process including activation of PCL‐diol hydroxyl end groups, cystamine attachment, and LMW‐PEI conjugation exhibited higher efficacy compared to the typical PCL copolymerization pathway like ring‐opening copolymerization. PEI‐ss‐PCL‐ss‐PEI could encapsulate antitumor drug DOX due to hydrophilic and hydrophobic interaction, and condense p53‐pDNA due to the positively charged surface. After endocytosis, the NPs could degrade sensitively under highly concentrated GSH in tumor cells and release intact DOX in core and p53‐pDNA on the surface.^{[ 272 ]} Besides, a short chain PEI modified with poly(PEG‐methacrylate)‐ss‐PCL copolymers synthesized through atom transfer radical polymerization (ATRP) was prepared to co‐deliver DOX and siRNA for synergistic therapy. After preparing PCL macroinitiator through ring‐opening copolymerization, the macroinitiator was used with PEG‐methacrylate (PEGMA) for ATRP to obtain the diblock polymer, poly(PEGMA)‐ss‐PCL. Next, the hydroxyl groups in poly(PEGMA)‐ss‐PCL were used to react with succinic anhydride for obtaining carboxyl groups, which were the sites for the modification of short chain PEI. After mixing the finally synthesized copolymers with DOX, the DOX‐loaded NPs were prepared through self‐assembly. Attributed to the positively charged PEI on the surface shell, siRNA adsorption and complexation were achieved. After intravenous administration and endocytosis, the siRNA/DOX‐loaded NPs were degraded aroused by GSH‐stimulated breakdown of disulfide bonds in the diblock polymers. Intact siRNA and DOX were released to induce synergistic therapy.^{[ 273 ]}

Recently, Deng et al.^{[ 274 ]} developed a near‐infrared (NIR)/reduction‐dual‐responsive PCL‐based NPs for co‐delivering Cas9/sgRNA and chlorin e6 (Ce6) to achieve synergistic therapy. They prepared nitrilotriacetic acid‐disulfanediyldipropionate‐polyethyleneglycol‐b‐polycaprolactone (NTA‐ss‐PEG‐PCL) copolymer to encapsulate Ce6 through anionic micelles self‐assembly. Next, Cas9/sgRNA was loaded on the nanomicelles attributed to the NTA‐terminated PEG block to form the Cas9/sgRNA‐loaded copolymer NPs (CC‐NPs), which also introduced negative charges from sgRNA on the surface. What is more, the cationic copolymer internalizing RGD‐PEG‐pAsp(DAB) (iRGD‐PD) was synthesized to interact with the negatively charged sgRNA on the CC‐NPs in order to introduce the tumor‐targeting ligand iRGD in CC‐NPs. After the iRGD induced tumor penetration and endocytosis, Ce6 generated ROS stimulated by NIR irradiation for disrupting endosomes/lysosomes, which facilitated Cas/sgRNA release in cytoplasm. Highly concentrated GSH in cytoplasm induced the breakage of disulfide bonds, leading to the disassembling of CC‐NPs and rapid release of Cas9/sgRNA. Cas9/sgRNA was designed to target Nrf2 gene silencing to prevent tumor cells escape from Ce6‐induced photodynamic therapy for tumor therapy.

5.2.6. Polylactic Acid‐Based and Poly(Lactic‐co‐Glycolic Acid)‐Based Disulfide‐Containing Polymer Vectors

Polylactic acids (PLA), known as polylactides, are synthesized through the polycondensation reaction between hydroxyl groups and carboxyl groups or the ring‐opening polymerization reaction in lactide. Ideal biodegradability, biocompatibility, and potential for the modification of PLA are the most appealing advantages in the application of drug delivery. After entering targeted tumor cells, PLA will be hydrolyzed to nontoxic hydroxy carboxylic acid by breaking the ester bond, which next participates in TAC and transforms into water and carbon dioxide. Therefore, PLA‐based vectors have little risk of significant accumulation or causing serious immune response in vivo. After modified with cationic groups, PLA could provide excellent loading capacity and efficient targeting ability for nucleic acid drugs.

PLGA is synthesized by condensing lactic acid (LA) and glycolic acid (GA) through ester bonds. In controlled drug release, the polymer related factors including molecular weight, LA/GA ratios, particle size, morphology, and surface modification of PLGA can be used to change PLGA properties.^{[ 377 ]} Initially, nucleic acid drugs were designed to be encapsulated in inner core of PLGA NPs. However, the anionic acid groups in PLGA have electrostatic repulsive forces with phosphates in nucleic acid molecules, resulting in the low encapsulating efficacy. Introducing the cationic groups like PEI and PEG as modification ingredients in PLGA‐based vector can adsorb nucleic acids on the surface of PLGA NPs. Next, in the acid environment of endolysosome, the charge reversal of PLGA also induces destabilization of endolysosomal membrane and helps the cargoes escape from lysosome.^{[ 378 ]}

Compared to other polymeric vectors, such polyester‐based polymers have significant low immunogenicity and better cargoes releasing curves that can completely degrade in natural pathways. However, PLA and PLGA always suffer from low gene delivering efficacy due to their anionic and hydrophobic characters. Pre‐cationization of PLA/PLGA brings the positively charged property and nucleic drugs can be adsorbed solidly on the surface of PLA/PLGA hydrophobic core. After grafting the hydrophilic groups to PLA/PLGA like PEG, PEI, dextran, chitosan, and lipid‐based coat, the outer shell will be formed through the self‐assembling in water solution. The outer shell also prolongs the circulation period and avoids the blockages by reticular system for PLA/PLGA polymeric vector.^{[ 379} , ³⁸⁰ , ^{381 ]} Besides, natural degradation of PLA/PLGA in vivo also leads to unideal drug release rate. The introduction of disulfide bonds achieves the accurate redox‐responsive drug release and rapid vector degradation. Linking PLA/PLGA and hydrophilic groups by disulfide bonds,^{[ 259} , ³⁸² , ^{383 ]} grafting PLA/PLGA with disulfide‐containing cationic groups^{[ 332 ]} and even directly conjugating siRNA to PLGA through disulfide bonds^{[ 384} , ³⁸⁵ , ^{386 ]} have been designed for rapid GSH‐triggered degradation and effective nucleic acids therapy in recent reports.

For example, disulfide‐containing group lipoic acid (LA) was introduced into PEI‐PLA complexes to prepare the redox‐responsive polymeric micelles, PEI‐PLA‐LA (PPL), as a co‐delivering vector for paclitaxel (PTX) and STAT3 siRNA (siSTAT3). PTX was first encapsulated in PPL micelles through hydrophobic interaction and self‐assembly, followed by the electronic adsorption of siSTAT3 on the PEI surface. After coating the hyaluronic acid (HA) for targeting the CD44‐overexpressing 4T1 cells, the synthesized NPs were injected into the orthotopic 4T1 tumor‐bearing mice and exhibited sensitive GSH‐triggered degradation and ideal synergistic therapy effect in tumor suppression.^{[ 150 ]} Besides, methoxypoly(ethylene glycol)‐polylactide‐polyhistidine‐ss‐oligoethylenimine (mPEG‐b‐PLA‐PHis‐ssOEI) was prepared to codeliver doxorubicin (DOX) and MDR1 siRNA (siMDR1). DOX was encapsulated in the hydrophobic core consisted of PLA‐PHis components. The positively charged OEI was used to adsorb siMDR1 and permeabilize the endosomal membrane for escaping from endosome. Outer shell mPEG ensured the biostability in blood circulation. After the co‐delivering NPs entered into tumor cells by phagocytosis, both PHis protonation in acidic environment and GSH‐triggered disulfide bond breakage realized a pH/redox dual‐response rapid release of siMDR1 and DOX.^{[ 275 ]} Next, a disulfide‐containing cationic oligo‐β‐aminoesters (OBAE) was designed to modify the disulfide‐containing polymer PEG‐co‐PLGA, for co‐delivering docetaxel (DTX) and an siRNA against TUBB3, a gene encoding for βIII‐tubulin. The synthesized redox‐responsive polymeric NPs were composed of outer PEG‐rich hydrophilic shell and inner PLGA‐rich hydrophobic core. DTX was encapsulated in PLGA core and siRNA was loading in cationic component OBAE. OBEA was able to condense siRNA solidly in a physiological environment and protect it to cross the cell membrane. Attributed to the reversible protonation of amine groups, acid/base‐sensitive ester bonds, and disulfide bonds in OBEA, it eventually degraded into nontoxic small molecules after being responsive to pH value and reduction environment.^{[ 276 ]} In order to realize doxorubicin and programmed P‐gp siRNA delivery, a reduction/photodual‐responsive device (RPDRD) was synthesized through loading o‐nitrobenzyl ester derivative caged DOX (DOC) into the inner PLGA core and adsorbing siP‐gp onto the cationic polymeric shell derived from a disulfide‐containing alkyl‐modified polyethylenimine (C16‐S‐S‐PEI). Disulfide broke down under enriched GSH cytoplasm and rapidly releases siP‐gp for the reversal of drug resistance by initially suppressing P‐gp protein expression in tumor cells.^{[ 277 ]}

Recently, our team designed NPs composed of a biodegradable methoxyl‐poly(ethylene glycol)‐b‐poly(lactic‐co‐glycolic acid) copolymer with a reduction‐responsive disulfide linker (denoted Meo‐PEG‐S‐S‐PLGA) and the amphiphilic cationic lipid‐like compound G0‐C14 to deliver cofilin 1 (CFL1) siRNA for enhancing the sensitivity of hepatocellular carcinoma (HCC) to sorafenib. In sorafenib‐resistant HCC cells, CFL1 upregulated phosphoglycerate dehydrogenase transcription and promoted serine synthesis and metabolism to accelerate the production of antioxidants for scavenging the excessive ROS generated by sorafenib. To avoid the serious side effect and block the clearance of ROS, the inner core composed of siCFL1 and G0‐C14, as well as sorafenib was encapsulated in Meo‐PEG‐S‐S‐PLGA outer shell. The disulfide bonds degradation not only achieved rapid GSH‐responsive cargo release, but also depleted GSH in cytoplasm, facilitating ROS accumulation. After being release in cytoplasm, siCFL1 disturbs CFL1 transcription, which results in decreased Nrf2 release and blocks the activation of antioxidant genes, especially PHGDH. Downregulation of PHGDH impairs serine synthesis and metabolism and inhibits the generation of antioxidant for scavenging excessive ROS produced by co‐delivered sorafenib, leading to reversed drug resistance and enhanced apoptosis^{[ 278 ]} (Figure  11 ).

Figure 11.

Schematic illustration of the molecular mechanism of co‐delivering sorafenib and siCFL1 by Meo‐PEG‐S‐S‐PLGA to reverse sorafenib sensitivity. CFL1 impairs the interaction between Keap1‐Nrf2 by depolymerizing F‐actin to promote the release of Nrf2, which enhances transcription of antioxidant genes, especially PHGDH. PHGDH plays a vital role in serine synthesis and metabolism to accelerate the generation of antioxidants, which act as ROS wiper, facilitating sorafenib resistance.

5.2.7. Polylysines‐Based Disulfide‐Containing Polymer Vectors

Polylysines (PLLs) is a type of cationic polymers condensed from lysine residues, whose differences arise from the changing in the attachment position and stereochemistry of the lysine residues. PLL can condense nucleic acids and target cell membrane via charged interaction, which significantly enhances nucleic acids encapsulating rates and cellular internalization.^{[ 387 ]} PLLs also have ideal biocompatible, biodegradable properties, and easily controlled molecular weight through n‐carboxylic anhydride polymerization.^{[ 388 ]} However, PLL is not good at mediating endosomal escape due to its lower amine group density. The modified components that induce proton sponge effects such as histidine residues and PEI can be attached to PLL to facilitate endosomal escape of nucleic acid drugs.^{[ 389} , ^{390 ]} Next, the large molecular weight PLL strongly traps the nucleic acid drug in the endosome due to its strong condensation to the nucleic acid drug, resulting in low efficiency of nucleic acid release. Therefore, the introduction of disulfide bonds was applied to the high nucleic acid loading rate polymer PLL for obtaining the ability of redox‐sensitive degradation, which further improves its efficiency as a nucleic acid carrier.

Currently, the methods of introducing disulfide bonds into PLL‐containing polymers include: i) linking PLL and hydrophilic groups like PEG through disulfide bonds; ii) adding disulfide bonds into the back bone of PLL; iii) directly conjugating nucleic acid molecules to PLL through disulfide bonds. For example, PEGylation of PLL via disulfide bonds was used to prepare the rapid GSH‐responsive and ideally biodegradable siRNA vector, which could effectively deliver VEGF siRNA (siVEGF) into HepG2 cells or HepG2 xenograft and release siVEGF triggered by the highly concentrated GSH to inhibit their development.^{[ 279 ]} Besides, disulfide bonds were attempted to directly attach siRNA and the units of PLL derivatives to form an siRNA‐conjugated polymer brush, followed by grafting the spacer component PEG and decoration group RGD for targeting αVβ3/αVβ5 integrin‐overexpressing tumor cells. Consequently, appreciable multivalent siRNA conjugates with high siRNA‐loading and GSH‐triggered siRNA releasing polymeric vector were obtained.^{[ 280 ]}

5.2.8. Chitosan‐Based Disulfide‐Containing Polymer Vectors

Chitosan, a type of natural polymeric alkaline polysaccharide, is composed of β−1,4‐linked‐d‐glucosamine and N‐acetyl‐d‐glucosamine units, which is well known for its ideal biocompatibility, biodegradable ability, bioadhesivity, and low immunogenicity in drug delivery.^{[ 391 ]} Currently, polyelectrolyte complexation and ionic gelation have been the two most common techniques in fabricating chitosan‐containing polymer for nucleic acids delivery. Due to the amino groups with pK _a 6.3 in chitosan structure, chitosan is naturally sensitive to the environment with pH value lower than 6. In acidic media such as lysosomes, chitosan can be protonated rapidly, causing dissolution of the polymer, endosomal escape, and rapid release of cargoes. The negatively charged nucleic acids drug can conjugate to the protonated amines in chitosan through charged interaction. The positive charges on chitosan also induce cellular internalization through adsorbing the negatively charged lipid bilayer or glycocalyx on the cell membrane. However, the hydrophilic structure of chitosan partially reduces the adsorption capacity of nucleic acid molecules, some cationic groups like PAMAM need to be grafted in chitosan for assisting nucleic acid condensation.^{[ 392 ]}

In light of its solubility and cationic properties, chitosan has been widely developed for nucleic acid delivery in cancer therapy. But it is noteworthy to optimize several important factors of chitosan to obtain more effective nucleic acid therapy effect, including molecular weight, derivatives properties, degree of deacetylation, and the N/P ratio. First, the molecular weight of chitosan can affect the formation of chitosan/nucleic acid polyplexes. Only chitosan with molecular weight five to ten times larger than siRNA can condense siRNA effectively and form stable NPs in 100–300 nm. Although higher molecular weight chitosan offers better stability and protection, its strong adsorption also limits the release of nucleic acid drugs. The ideal intracellular release and the molecular weight of chitosan need to be balanced when using chitosan as a vehicle to deliver nucleic acid drugs. Besides, the degree of modification in chitosan also influences its delivery efficiency. Typical hydrophilic modifications such as PEGylation on the chitosan backbone not only confers better solubility without limiting the nucleic acid condensation, but also reduces the size and ζ‐potential of the NPs, facilitating the internalization of target cells. Next, degree of deacetylation is also an important factor in altering chitosan property. The increased exposure of reactive amino groups induced by the increased deacetylation causes an enhanced positive charge of the chitosan, giving rise to the better nucleic acid loading capacity and cell membrane adsorbing ability.^{[ 393 ]} The degree of deacetylation of chitosan must reach 65% and 80% for stable complexation with pDNA and siRNA, respectively.^{[ 236} , ^{394 ]} Last, the N/P ratio is defined as the molar ratio between the nitrogen of chitosan and the phosphate of per nucleic acid. Higher N/P ratio indicates greatly concentrated and highly positive charged chitosan component. Similar to molecular weight, higher N/P ratio also leads to over‐adsorption of nucleic acid drugs resulting in lower delivery efficiency and therapeutic efficacy.^{[ 395 ]}

What is more, the rich amino and hydroxyl groups provide modification sites for conferring chitosan conditionally responsive release ability, which makes chitosan‐based GSH‐responsive vectors possible through introducing disulfide bonds in skeleton of the vectors. For example, Jia et al.^{[ 281 ]} grafted LMW‐PEI to polymeric backbone chitosan through disulfide bonds to synthesize a polymeric nucleic acid vector with suitable charge density/molecular weights for condensing P‐glycoprotein shRNA effectively and GSH‐triggered degradation ability for rapid drug release. Folic acid was next grafted to the synthesized copolymer for enhancing selective breast cancer cell uptake. The smart design solved the high viscosity and poor solubility of chitosan in physiological pH environments. Besides, to solve the problems of spontaneous formation of protein corona on chitosan‐based vector in blood circulation and further rapid blood clearance by reticuloendothelial system, Yang et al.^{[ 282 ]} used cyclic RGDyK peptide (cRGD)‐modified bovine serum albumin as a tumor‐targeting leader and a corona coat on redox‐responsive chitosan‐based nanocarriers (TsR NPs). For preparing TsR NPs, trimethyl chitosan (TMC) was synthesized via quaternization of chitosan for better solubility and Nona‐arginine (9R), a polycationic peptide for forming pores in the lysosomal membrane, was grafted to TMC skeleton through disulfide bonds. The synthesized polymer denoted as TMC‐ss‐9R could improve the loading capacity for nucleic acid drugs and induce the function of lysosomal escape, following the GSH‐triggered degradation and siRNA release. Next, Chen et al.^{[ 283 ]} used thiolated PEI (PEI‐SH) to complex with pDNA through charged interactions to form the inner core, followed by grafting the glycol chitosan modified with succinimidyl 3‐(2‐pyridyldithio) propionate (GCS‐PDP) on the surface as an outer shell through a thiolate‐disulfide interchange reaction. The synthesized GCS‐PDP/PEI‐SH/DNA NPs exhibited high colloid stability in the simulated physiological environment and redox‐responsive structural degradation and rapid DNA release for high transfection efficiency, which were useful candidate for genetic therapy in cancer.

5.2.9. Other Disulfide‐Containing Polymer Vectors

To overcome the drawback that cation‐associated toxicity limits the further clinical application of cationic polymer vector, a cation‐free polymer‐siRNA nanocapsule with disulfide‐crosslinked interlayer was developed to mediate tumor‐targeted RNAi therapy (Figure  12 ). Complexation of siRNA with a cationic block polymer cRGD‐poly(ethylene glycol)‐b‐poly[(2‐aminoethanethiol)aspartamide]‐b‐poly{N'‐[N‐(2‐aminoethyl)−2‐ethylimino‐1‐aminomethyl]aspartamide} was first developed, followed by forming interlayer crosslinking by disulfide bond in pH 7.4 solution. Finally, cationic DETA pendant was removed at pH 5.0 via breakage of imide bond. The synthesized siPLK1 nanocapsules showed significant siRNA delivery efficiency and tumor inhibiting ability without cation‐associated toxicity side effects.^{[ 284 ]}

Figure 12.

Preparation and intracellular fate of cation‐free disulfide bond‐crosslinked polymer‐siRNA nanocapsule T‐SS(‐). a) T‐SS(‐) was prepared via three steps including siRNA complexation by electrostatic interaction, disulfide bond crosslinking by blowing O2, and removal of cationic pendant DETA by breaking pH‐labile imide bond. b) cRGD‐mediated cancer cell targeting, GSH‐sensitive siRNA release for PLK1 silencing to induce apoptosis of tumor cell. Reproduced with permission.^{[ 284 ]} Copyright 2023, Elsevier.

Similarly, a novel cation‐free siRNA micelle was prepared by the self‐assembly of siRNA‐disulfide‐poly(N‐isopropylacrylamide) (siRNA‐SS‐PNIPAM) diblock copolymers to co‐deliver temozolomide and siSTAT₃. Due to being noncationic, these micelles showed low level of charge‐associated toxicity. The ideal characteristics including enhanced circulation periods, cellular uptake, and rapid siRNA release endow the siRNA micelles with outstanding growth inhibition of orthotopic U87MG xenografts.^{[ 286 ]}

An arginine‐modified poly(disulfide amine)/siRNA nanocomplex was developed to deliver siRNA targeting to oncogene KRAS. The polycation, poly((N,N'‐bis(acryloyl)cystamine‐co‐ethylenediamine)‐g‐Nω‐p‐tosyl‐l‐arginine) (PBR), was synthesized by aza‐Michael polyaddition followed by the introduction of arginine groups onto its backbone. Benefited from optimizing the grafting density of arginine groups and disulfide components, such nanocomplexes achieved excellent cellular uptake, endosomal/lysosomal escape, and rapid GSH‐responsive degradation in the cytoplasm.^{[ 285 ]}

5.3. Thioketal‐Based Polymer Vectors for Nucleic Acid Drug Delivery

The earliest attempt of thioketal‐containing polymers for delivering nucleic drugs dates back to 2010, Wilson et al.^{[ 396 ]} used ROS‐responsive poly‐(1,4‐phenyleneacetonedimethylene thioketal) to deliver TNF‐α siRNA to the intestinal inflammation sites with abnormally high concentration of ROS. Shim and Xia^{[ 318 ]} first reported the successful application of delivering nucleic acids to cancer cells via thioketal‐containing polymer. Synthesized by the polymerization of oligoamines with acrylamide thioketal crosslinkers, the ROS‐cleavable and cationic polymer, poly(amino thioketal), significantly enhanced the efficiency of delivering DNA into cancer cells.

Recently, Wan et al.^{[ 287 ]} prepared the ROS‐responsive polymer (CA‐PLL‐TK) with a polycationic PLL as the backbone, linking thioketal (TK) and cis‐aconitate (CA), followed by the conjugation of iRGD after CA‐PLL‐TK self‐assembling in water and loading siRNA drugs. Ibrinogen‐like protein 1 siRNA (siFGL1) and programmed death‐ligand 1 (siPD‐L1) were chosen to inhibit the FGL1/Lag‐3 signaling axis and PD‐1/PD‐L1 signaling axis to reverse T‐cell depletion. The iRGD‐induced cell penetrating, CA‐induced lysosomal escaping, and thioketal‐induced ROS‐triggered degradation significantly increased the gene silencing effect of siFGL1 and siPD‐L1, which facilitated the T cells‐mediated tumor immunity.

Similarly, Yang et al.^{[ 288 ]} also constructed a thioketal‐containing polymer (^TKPFH) for co‐delivering shGPX4 and shMTHFD2 plasmids for cancer treatment. ^TKPFH consisted of the inner thioketal‐crosslinked fluorinated polyethyleneimine 1.8K (^TKPF) core and outer hyaluronic acid (HA) shell. HA induced active delivery targeted to the CD44‐expressed tumor cells. After the ROS‐triggered thioketal cleaved and polymer degradation, shGPX4 plasmids and shMTHFD2 plasmids were released and mediated synergistic cancer therapy. Through producing ROS and lipid peroxides via downregulating GPX4, shGPX4 plasmid induced ferroptosis for killing cancer cells, followed by shMTHFD2 triggering apoptosis by regulating NADPH/NADP and GSH depletion in cancer cells. What is more, the shMTHFD2‐induced GSH consumption also indirectly inhibits GPX4 expression which further strengthened the ferroptosis (Figure  13 ).

Figure 13.

Schematic illustration of the fabrication processes of a) TKPFH NPs and b) anticancer mechanisms of triggering apoptosis and ferroptosis for cancer therapy. Reproduced with permission.^{[ 288 ]} Copyright 2023, Elsevier.

However, not all kinds of cancer cells produce highly concentrated ROS enough for breaking thioketal. The uneven distribution of ROS concentrations in varies cancer cell also results in the different intensity drug release. Some strategies that actively generate ROS for cleaving thioketals and releasing nucleic acids rapidly need to be explored.

Wang et al.^{[ 289 ]} reported an ROS‐degradable polycation for co‐delivering p53 gene and photosensitizer, which realized the photo‐programmable gene release in tumor cells. Cationic polymer thioketal‐crosslinked PEI (TK‐PEI) was prepared to condense p53 gene and form the nanocomplexes, followed by the decoration of HA modified with the photosensitizer, pheophytin a (Pha), for enhancing their colloidal stability and enabling cancer cell targeting. After the gene delivery system being internalized in tumor cell and entering into endosome, the nonlethal concentration of ROS was generated by the far‐red light irradiation (661 nm) at low optical power density, which triggered the thioketal‐containing polymer degradation and p53 gene release. In particular, the generated ROS also further oxidized and destroyed the endosomal membranes to induce endosomal escape of the p53 gene, which was denoted as “photo‐chemical internalization.” Similarly, Wang et al.^{[ 290 ]} also reported the polymeric siRNA carrier, ^TKPEI‐Ce6, synthesized by the crosslinking reaction between bPEI and thioether followed by the conjugation of PEG and chlorin e6 (Ce6). Cationic ^TKPEI‐Ce6 could condense siRNA effectively and generate ROS by Ce6 component under red‐light irradiation (660 nm), which significantly accelerated endosomal escape, polymer degradation, and gene silencing efficacy of siRNA.

Nevertheless, the combination strategy of thioether and the cationic polymer like PEI still has some shortcomings. The secondary and tertiary amines in PEI could also promote endosomal escape through proton sponge effect, which caused the undesirable siRNA release and gene regulation in the normal cells. In conclusion, due to the secondary and tertiary amines‐induced nonspecific lysosomal escape, such cationic polymers failed to prevent the unwanted siRNA‐mediated gene silencing in nontarget cells. In comparison, the quaternary ammonium group‐based cationic polymers with highly positive charged property are unprotonatable under any pH condition, which could limit siRNA drugs in the endosomes of nontarget cells without the risk of unwanted gene silencing. For example, Deng et al.^{[ 291 ]} developed a multifunctional quaternary ammonium group‐based cationic polymer iRGD‐poly(ethylene glycol)‐b‐poly(aspartic acid‐ethylthioketal quaternary amino)‐b‐polyphenylalanine, denoted as iRGD‐PEG‐PAsp(TKQA)‐PPhe, for co‐delivering siRNA and Ce6. iRGD‐PEG‐PAsp(TKQA)‐PPhe was first self‐assembled with Ce6 and forms positively charged T‐Ce6, following complexing with siGPX4. After being injected in the body, siRNA molecules were limited and degraded in endolysosome and showed the siRNA activity “off” state, due to the unprotonatable property of quaternary ammonium in normal cells. After applying NIR irradiation to the tumor tissue, the Ce6‐generated ROS triggered siRNA endolysosomal escape, rapid cytosolic release and showed the siRNA activity “on” state, due to the photochemical internalization effect and cleavage of thioketal (Figure  14 ).

Figure 14.

Preparation of NIR‐ and ROS‐sensitive NPs incorporating photosensitizer Ce6 and siRNA and their distinct intracellular fates in the normal and tumor cells. Reproduced with permission.^{[ 291 ]} Copyright 2023, Wiley.

5.4. Boronic Acid‐Based Polymer Vector for Nucleic Acid Drug Delivery

The boronic acid‐based ROS‐responsive functional groups used in polymer vector for nucleic acid delivery mainly include phenylboronic ester (PBE) and poly[(2‐acryloyl) ethyl (p‐boronic acid benzyl) diethylammonium bromide] (B‐PDEAEA).

PBE is an important BA‐containing functional group with high selectivity and sensitivity to H₂O₂, which has been widely applied for ROS‐triggered polymer degrade and drug release. During the reaction of PBE with H₂O₂, H₂O₂ first attacks the boron center via a nucleophilic addition reaction, followed by the migration of the phenyl group to the oxygen atom via 1,2‐insertion and the final hydrolysis of the boronic ester and the release of phenol.^{[ 315 ]} In conclusion, PBE was mainly designed in three strategies to mediate ROS‐triggered polymeric degradation. PBE could be installed as a pendant group to the backbone of polymers for revealing phenols stimulated by H₂O₂, resulting in the polymer degradation through electronic transition and quinone methide rearrangement. Second, PBE was also used as the pendant group to cover the reactive groups of cleavable backbone of polymers. In highly concentrated H₂O₂ environment, the oxidation of PBE induced the exposure of reactive groups and the chemical amplification‐accelerated polymer degradation, which significantly improved the ROS sensitivity. Next, directly grafting hydrophobic PBE or the hydrophobic pendant PBE‐containing blocks into the backbone of hydrophilic polymer also achieved the hydrophilicity switch‐triggered degradation in H₂O₂ condition.^{[ 397 ]} For example, Ruan et al.^{[ 292 ]} prepared an ROS‐biodegradable cationic siRNA vector derived from bPEI through a crosslinking reaction by boric acid ester bonds, followed by the functionalization of substance P peptide through PEG. The synthesized Pololike Kinase 1 siRNA vector had boric acid ester bonds in backbone, exhibited ROS cleavability and ideal targeting ability as well as gene silencing efficiency for inhibiting breast cancer cells.

Furthermore, charge‐reversal is another pathway of the ROS‐triggered degradation of boronic acid‐containing polymer. The charge‐reversal polymeric NPs are able to shade their positive surface charges in blood circulation and re‐expose at tumor sites in response to specific stimuli changing their surface charges, including pH value, redox condition, and some concentrated enzymes.^{[ 398 ]} Liu et al.^{[ 293 ]} first proposed the positive to negative charge‐reversal concept that the cationic polymer can turn to negatively charged completely in response to H₂O₂ in tumor cells. They first prepared tertiary amine‐based polymer poly[2‐(N,N‐diethylamino)ethyl acrylate] (PDEAEA) through radical polymerization, which was reacted with an excess of 4‐(bromomethyl)phenylboronic acid for synthesizing water‐soluble poly[(2‐acryloyl) ethyl (p‐boronic acid benzyl) diethylammonium bromide] (B‐PDEAEA) through quaternization reaction. Cationic polymer B‐PDEAEA was used to condense DNA molecules for obtaining polyplex, which was coated by a lipid envelope for fusing with cell membrane like a paramyxovirus and ejecting the polyplex into the cytosol. In the oxidization of boronic acid in highly concentrated ROS environment, the quaternary ammonium released p‐quinone methide (p‐hydroxylmethylenephenol (HMP)) and became tertiary amine, followed by the self‐catalyzed fast hydrolysis of the ester group with the production of negatively charged poly(acrylic acid). Next, the negatively charged polymer released DNA rapidly without the risk of interfering gene transcription compared to the cationic gene vector with degraded cationic fragments. Besides, their team also coated the ROS‐responsive B‐PDEAEA polymers with cationic lipids to form the ROS‐responsive lipopolyplexes with integrated properties to overcome multiple delivery barriers. The cationic lipids endowed the polyplex with better serum stability, lysosomal escape, and amplified ROS generation for charge reversal and DNA release.^{[ 294 ]}

In order to achieve more precise charge conversion for rapidly releasing nucleic acid drugs, the strategies for actively generating ROS to mediate the conversion of positive to negative electrical properties of polymers have been developed. Zhu et al.^{[ 295 ]} attached a β‐cyclodextrin‐functionalized poly(glycerol methacrylate) (PG) segment and a quaternary amine‐functionalized poly[(2‐acryloyl)‐ethyl‐(p‐boronic acid pinacol ester benzyl)diethylammonium bromide] (BP) segment for complexing DNA molecules to synthesize the ROS‐responsive polymeric vector, followed by the medication of hydrophobic poly[2‐(5,5‐dimethyl‐1,3‐dioxan‐2‐yloxy)ethyl acrylate] (PDM) for enhancing complex stability. The light‐stimulated ROS‐producing 4,40‐(1,2‐diphenylethene‐1,2‐diyl)bis(1,4‐phenylene)diboronic acid (TPE) was further installed in the cavities of cyclodextrin. Upon light irradiation, the actively generated ROS by TPE not only destroyed the endosomal membrane for polyplex escaping, but also induced the charged transformation for DNA release. Recently, Zhang et al.^{[ 296 ]} reported an ultrasound‐enhanced ROS‐triggered charge‐reversal polymeric gene vector through wrapping B‐PDEAEA/DNA complex in IR780 (a typical sonosensitizer)‐loaded liposomes. Upon the ultrasound treatment, IR780 produced abundant intracellular ROS to mediate the ROS‐responsive cationic polymer B‐PDEAEA release p‐hydroxybenzylalcohol (HMP) and switch to a tertiary amine, facilitating the polymeric charge‐reversal and rapid DNA release.

5.5. Thioacetal‐Based Polymer Vector for Nucleic Acid Drug Delivery

Endogenous ROS in tumor cells can induce thioacetal breakage, which has been introduced to construct ROS‐responsive polymer carriers for rapid drug release.^{[ 399} , ^{400 ]} Zhang et al.^{[ 297 ]} prepared a series of fluorobenzene‐substituted and thioacetal‐contained polycations (TAEA‐S‐xF) to explore novel strategies for efficient gene delivering vectors. A diepoxide ring‐opening polymerization was used to synthesize the fluorobenzene‐substituted polycations with ROS‐responsive thioacetal linkage, which exhibited ideal DNA condensation ability, serum stability, cell uptake, and endosomal escape ability attributed to the fluorination. ROS‐triggered thioacetal breakdown induced the degradation of cationic polymers, leading to an efficient DNA release. Besides, Lin et al.^{[ 298 ]} reported a series of polycations synthesized by conjugating diepoxide‐containing ROS‐responsive thioacetal groups to low molecular weight PEI through epoxide ring‐opening polymerization. Four diglycidyl ethers with various electron effects were also employed in the side chain for exploring the structure–activity relationships. The synthesized polymers exhibited ideal DNA condensing ability with moderate size and ζ potential, which could rapidly degrade in highly concentrated ROS and release DNA efficiently for cancer therapy.

5.6. Aminoacrylate‐Based Polymer Vector for Nucleic Acid Drug Delivery

Aminoacrylate group is a ROS‐responsive group that cannot be oxidated by other superoxide radical but only by ¹O₂. Yuan et al.^{[ 299 ]} reported the polymer vector consisted of oligoethylenimine (OEI) conjugated through aminoacrylate (AA) groups and photosensitizer (PS) with aggregation‐induced emission. After the polymer self‐assembles into the bright red fluorescent NPs in aqueous media, DNA can be effectively condensed through electric interaction. Under the irradiation of visible light, the PS‐generated ¹O₂ on the one hand destroyed the endosomal membrane to facilitate endosomal escape, on the other hand broke the aminoacrylate group in polymeric vector to induce polymer degradation and rapid DNA release in cytosolic. The synthesized ROS‐responsive polymer exhibited on average over 50% increase of transfection efficiency, compared to commercial PEI_25k.

5.7. Diselenide‐Based Polymer Vector for Nucleic Acid Drug Delivery

The heterogeneity of tumor cells results in the uneven distribution and the insufficient concentration of ROS/GSH in different species or in the same species of tumor cells, which indicated the challenge of redox‐responsive vectors’ universality and the efficiency limitation of single‐responsive vectors. Based on the redox dual‐responsiveness of diselenide, it has been widely used for nucleic acid drug delivery to address the low degradation sensitivity of conventional polymeric carriers. Diselenide can be introduced in the backbone or side chain of polymer structure as well as used to crosslink the polymer chains for achieving ROS/GSH dual‐responsive nucleic acid drug release to enhance therapeutic efficiency.

Diselenide bonds were originally developed to replace disulfide bonds as GSH‐responsive degradation groups because of their smaller bond energies. Chen et al.^{[ 300} , ^{301 ]} synthesized a polycationic carrier, OEI800‐SeSe_x, through the addition of branched oligoethylenimine 800 Da (OEI₈₀₀) via an active ester containing diselenide bonds. OEI800‐SeSe_x exhibited higher transfection efficiency, lower cytotoxicity, and enhanced GSH‐responsive ability as efficient as that for disulfide bonds crosslinked with OEI₈₀₀‐SS_x. He et al.^{[ 302 ]} also synthesized diselenide‐conjugated oligoethylenimine (OEI‐SeSe_x) as a nanoscale core for condensing pDNA drugs, followed by the shielding of disulfide‐modified hyaluronic acid derivatives (HA‐SS‐COOH). The ternary polyplexes (pDNA/OEI‐SeSe_x/HA‐SS‐COOH) could respond in a stepwise fashion to the dual reduction gradients in the tumor site and intracellular conditions through chemically defined mechanisms with the degradation induced by disulfides and diselenide, respectively. The ROS‐responsive ability of diselenide was first applied in nucleic acids delivery in 2017, Deng et al.^{[ 303 ]} attempted to synthesize ROS‐responsive polycations based on diselenide for enhancing the transfection performance in cancer therapy. LMW PEI was crosslinked with a diselenide‐containing linker, followed by the modification with fluorocarbon chains. In the highly concentrated ROS environment of cancer cells, diselenide was sensitively oxidized to seleninic acid, resulting in the degradation of polycations and rapid release of pDNA with low cytotoxicity. Recently, He et al.^{[ 304 ]} constructed upconversion nanoonions with upconversion NPs (UCNPs) core with inner coating layer of photosensitizer rose Bengal (RB)‐conjugated PEI₆₀₀, middle coating layer of ROS‐responsive diselenide‐linked PEI₆₀₀ (PEI₆₀₀‐SeSe) with therapeutic siRNA loading and cell‐penetrating peptide R8 decoration, and outer coating layer of HA. After HA‐induced targeting to the tumor sites and positively charged R8‐induced escaping from tumoral endosomes, the 808 nm irradiation mediated the emissions about 540 nm generated from the UCNPs core, which activated RB for producing ROS. ROS‐triggered diselenide breakdown facilitated the degradation of PEI₆₀₀‐SeSe, resulting in the efficient and rapid release of therapeutic siRNA for gene silencing cancer therapy.

Recently, the ROS/GSH dual‐responsive property of diselenide has been introduced in polymer nucleic acid vector for more efficient cancer therapy. Fang et al.^{[ 401 ]} reported a high molecular weight polymer dPSPs synthesized through crosslinking endogenous spermine by diselenide, which exhibited complete DNA compression, low cytotoxicity, and ideal proton buffering effect. Next, therapeutic pDNA and photosensitizer indocyanine green (ICG) were loaded in the synthesized polymeric vector to synthesize the polyplexes. After the polyplexes escape from endosome induced by spermine, the NIR in 808 nm was applied for tumor sites irradiation, resulting in the ROS generation from ICG for initial vector degradation. The highly concentrated ROS and GSH in tumor cells next completely broke the diselenide in dPSPs, leading to the rapid DNA release. What is more, Chen et al.^{[ 305 ]} constructed a novel ROS/GSH dual HA‐modified nanohybrid HA‐SeSe‐COOH/siR‐93C@PAMAM for efficiently delivering siRNA into tumor cells with redox‐induced intracellular degradation. Polycationic PAMAM was used as basic structure to condense siRNA, followed by shielding the excessive positive charges by diselenide‐modified HA. Introduction of diselenide provided an ROS/GSH‐triggered dis‐shielding effect to hide the positive charges of PAMAM through the breakdown of diselenide and significantly decreased the cytotoxicity of PAMAM vector. Therefore, diselenide‐modified HA not only could promote the stability and safety of polymer vector, but also enhanced intracellular behavior of siRNA due to ROS/GSH dual‐sensitivity.

6. Conclusion and Challenges

The application of redox‐sensitive polymer vector for delivering nucleic acid drugs in cancer therapy has been investigated widely due to abnormal ROS and GSH concentration in tumor cells. Up to now, the ROS/GSH‐responsive function groups introduced in polymer are limited, which mainly consist of disulfide, thioketal, boronic acid, and diselenide. Such functional groups can be introduced in the main chain of the polymer as a backbone and installed in the side chains to link other groups or drugs, as well as be used to crosslink polymer chains. The synthesized polymer vectors always load nucleic acid drugs through electric interaction and protect them from nucleases in blood circulation. After internalization of tumor cells, some cationic polymers also induce proton sponge effect to promote lysosomal membrane destruction and lysosomal escape. In cytosolic with high concentrated ROS/GSH, the functional groups are cleaved resulting in the polymer degradation and nucleic acid drugs release for further regulation of gene expression.

Despite the rapid development of redox‐responsive polymeric carriers, the ultimate goal of these materials for clinical applications is still difficult to achieve. A prerequisite for the clinical application of redox‐responsive polymeric vectors is their sensitivity to the abnormal redox environment of tumor cells in human body. Due to heterogeneity of tumor cells, various tumor cells possess different GSH/ROS concentration distributions, resulting in the inhomogenous drug release from polymeric vector. In some cancer cells, the polymer vector even fails to degrade completely because of the relatively low ROS level. Therefore, it is necessary to design the multiresponsive polymer vector for enhancing degradation efficiency. Loading photosensitizers in polymer to generate highly concentrated ROS for actively inducing polymer degradation is also a hopeful strategy. Besides, the increased production of ROS is also found in acute organ inflammation, hypoxia, neurodegenerative diseases, cardiovascular disease, and diabetes, resulting in the ROS‐responsive polymers degrading in noncancerous cells and unwanted nucleic acid drugs release. The tumor‐targeted ability of polymer vector mainly depends on EPR effect. Hence, it is worthwhile to explore how to improve the tumor tissue‐specific targeting ability of ROS‐responsive polymer for avoiding unwanted drug release and toxic side effect.

Next, in the selection of nucleic acid drugs, the strategies that targeting the genes associated with redox‐homeostasis in tumor cells needed to be explored. The current reports of nucleic acid therapy were still focused on typical antitumor gene, which failed to combine with the abnormal redox status of tumor cells to achieve synergistic therapeutic effect. For example, the effective silencing of crucial antioxidation gene like Nrf2 may result in the significant disturbance of metabolic reprogramming and in turn amplified oxidant stress, facilitating more effective degradation of redox‐responsive polymers and rapid drug release.

In addition, it cannot be ignored that the strategy for controllable and reproducible synthesis of the designed NPs is the first need to meet clinical translation. The successful and repeatable preparation of ideal NPs depends on their controllable optimal physicochemical parameters. Because the physicochemical parameter of NPs can influence their immune escape, stability in blood circulation, tumor penetration and accumulation, tumor cell internalization, and controlled cargoes release in individual factors. Ideal batch‐to‐batch reproducibility in size distribution and tunable physicochemical is crucial when screening targeted NPs. The transition from laboratory to clinical application needs the optimization of formulation parameters for achieving large‐scale and reproducible preparation. Next, most of the developed NPs in researches are only trapped in preclinical stage with huge difficulty and cost in clinical translation and commercial production. In detail, experiments in vitro limited to multiwell plates cannot effectively mimic the biological complexity including immune system reaction, serum protein adsorption, and phagocytic effects of the endothelial reticular system in vivo, which fails to reflect the intricate interaction between NPs and physiological barriers. Meanwhile, current animal tumor models including subcutaneous tumor models and xenograft tumor models also fail to perfectly replicate all aspects of human homologous tumors. Therefore, in addition to work on designing more efficient polymer vector, we should also focus on the fundamental factors that affect the clinical translation of NPs.

Conflict of Interest

The authors declare no conflict of interest.

Biographies

Lei Xu received his Ph.D. in polymer chemistry and physics at Wuhan University in 2020. From 2021 to 2022, he worked at Sun Yat‐Sen University and worked as a postdoctoral research fellow. In 2023, he was promoted to the position of associate professor at Sun Yat‐Sen University. His main research interests focus on targeted drug delivery and cancer therapy.

Yuan Cao obtained his Medical Bachelor in basic medicine at Southern Medical University in 2021. He is now studying for MS degree at Sun Yat‐Sen University under the direction of Prof. Xiaoding Xu. His main research interests focus on cancer nanomedicine.

Xiaoding Xu obtained his B.Sc. in chemistry and Ph.D. in polymer chemistry and physics at Wuhan University in 2006 and 2011, respectively. After finishing 2 years' postdoctoral research at Wuhan University, he moved to Harvard Medical School and worked as a postdoctoral research fellow. After 3 years' postdoctoral research training, he was promoted to the position of Instructor at Harvard Medical School in 2017. Since 2018, he has been a professor at Sun Yat‐Sen University. His main research interests focus on cancer pathogenesis and cancer nanomedicine.