Framework nucleic acids as promising reactive oxygen species scavengers for anti-inflammatory therapy

Abstract

Reactive oxygen species (ROS) are an array of derivatives of molecular oxygen that participate in multiple physiological processes under the control of redox homeostasis. However, under pathological conditions, the over-production of ROS often leads to oxidative stress and inflammatory reactions, indicating a potential therapeutic target. With the rapid development of nucleic acid nanotechnology, scientists have exploited various DNA nanostructures with remarkable biocompatibility, programmability, and structural stability. Among these novel organic nanomaterials, a group of skeleton-like framework nucleic acid (FNA) nanostructures attracts the most interest due to their outstanding self-assembly, cellular endocytosis, addressability, and functionality. Surprisingly, different FNAs manifest similarly satisfactory antioxidative and anti-inflammatory effects during their biomedical application process. First, they are intrinsically endowed with the ability to neutralize ROS due to their DNA nature. Therefore, they are extensively involved in the complicated inflammatory signaling network. Moreover, the outstanding editability of FNAs also allows for flexible modifications with nucleic acids, aptamers, peptides, antibodies, low-molecular-weight drugs, and so on, thus further strengthening the targeting and therapeutic ability. This review focuses on the ROS-scavenging potential of three representative FNAs, including tetrahedral framework nucleic acids (tFNAs), DNA origami, and DNA hydrogels, to summarize the recent advances in their anti-inflammatory therapy applications. Although FNAs exhibit great potential in treating inflammatory diseases as promising ROS scavengers, massive efforts still need to be made to overcome the emerging challenges in their clinical translation.

---

1 Introduction

Deoxyribonucleic acid (DNA) is one of the most important hereditary materials in all living systems, playing a central role in storing and encoding genetic information that lies in the sequence and quantity of four kinds of composing deoxynucleotide monomers.¹ This unique constitution and the base-pairing principle together determine the possibility of artificial design and fabrication of various single- and double-stranded DNAs (ssDNA and dsDNA).² Therefore, the naturally outstanding biocompatibility, programmability, and structural stability make DNA a promising biological nanomaterial.^3,4 After Ned Seeman first proposed the concept of DNA nanostructures in 1982,⁵ the pioneering DNA nanotechnology has ushered in a booming development. According to the Watson–Crick base-pairing rules, properly designing complementary sticky ends in different ssDNAs allows accurate and stable branch junctions to be automatically formed in dsDNA-based nanostructures.^4,6 Consequently, with different designing thoughts and fabricating methods, scientists have made major breakthroughs in constructing various DNA nanostructures, ranging from one-dimensional DNA tubes and ribbons to two-dimensional DNA tiles and DNA origami, and three-dimensional DNA polyhedral nanostructures.^7,8

Among numerous DNA nanostructures, the special framework nucleic acids (FNAs) come into the limelight due to their excellent biological performance. FNAs are a large category of DNA nanomaterials that can self-assemble into specific skeleton-like sizes and shapes, and serve as frameworks to organize functional materials, which include DNA tiles, DNA origami, DNA bricks, triangular prisms, tetrahedra, icosahedra, DNA hydrogels and so on.^1,6,9,10 Despite the diverse sizes and morphologies, considering the costs and yields, only several simple but practical FNA structures are selected and extensively studied. These FNAs mainly contain tetrahedral framework nucleic acids (tFNAs), DNA origami, and DNA hydrogels.⁹ Overall, apart from the aforementioned universal properties of DNA, the above three FNAs also particularly excel in controllable self-assembly, facilitated cellular endocytosis, precise addressability, and tailorable functionality.^11,12 First, adenine (A)–thymine (T) and guanine (G)–cytosine (C) pairing can be spontaneously formed inside the complementary ssDNA chains via two and three hydrogen bonds respectively, thus constructing a stable double-helical structure. Since the length and composition of each strand have been previously decided, the specific scales and morphologies of these self-assembled FNAs are also controllable and predictable.¹³ Second, it is well documented that regardless of their polyanionic characteristic, tFNAs can still easily penetrate the cell membrane in a caveolin-mediated way, achieving outstanding cellular endocytosis.¹⁴ Additionally, since each DNA double helix has definite diameters (2 nm) and pitches (3.4 nm and about 10.5 base pairs per turn), it is convenient to calculate the geometric parameters of different FNAs. Then, guided by a specific oligonucleotide sequence, we can precisely locate the address of functionalized groups and achieve subtle interactions or regulations.^1,15–18 Finally, the excellent editability of nucleotide sequences also enables flexible modifications with nucleic acids, aptamers, peptides, antibodies and low-molecular-weight drugs, hence endowing simple FNAs with specialized functions.^11,19 Taking advantage of the above characteristics, researchers have used FNAs and FNA-based drug delivery systems to treat various disease models, including ischemia–reperfusion (I/R) injuries, diabetic wound healing, osteoarthritis, sepsis, etc.

Among most of these physical diseases, uncontrolled inflammation is acknowledged as a pivotal pathological process, which is mediated by four major mechanisms, including reactive oxygen species (ROS)-induced oxidative stress, the JAK-STAT pathway, the NF-κB pathway, and the MAPK pathway.²⁰ Surprisingly, while exploring the biomedical application of FNAs, researchers noticed that different FNAs can exhibit similar therapeutic effects in suppressing inflammatory reactions, especially those characterized by ROS-scavenging properties. First, the unadorned FNAs alone can directly neutralize excessive ROS and directly or indirectly modulate the complicated inflammatory signaling network. Besides, based on their prominent addressability and functionality, FNAs can be flexibly modified by multiple drugs, further strengthening their inflammation targeting and therapeutic ability.

In this review, we first briefly demonstrated the major regulatory mechanisms of ROS-involved inflammatory signaling pathways to provide a clear outline of the close connections between antioxidation and anti-inflammation strategies. Then we summarized three representative FNAs, including tetrahedral framework nucleic acids (tFNAs), DNA origami, and DNA hydrogels, as promising ROS scavengers to resolve pathological inflammation and their respective advances in anti-inflammatory applications (Fig. 1). Finally, we also discussed the potential challenges in the clinical translation of FNAs, hoping to provide some instructive research references for future investigations.

Fig. 1 Classifications and main structures of three major ROS-scavenging framework nucleic acid nanomaterials. Adapted with permission from ref. 104. Copyright 2021, American Chemical Society. Adapted with permission from ref. 75. Copyright 2022, American Chemical Society. Adapted with permission from ref. 77. Copyright 2020, Springer Nature. Adapted with permission from ref. 105. Copyright 2020, American Chemical Society. Adapted with permission from ref. 86. Copyright 2018, Springer Nature. Adapted with permission from ref. 89. Copyright 2022, American Chemical Society. Adapted with permission from ref. 87. Copyright 2021, American Chemical Society. Adapted with permission from ref. 101. Copyright 2023, American Chemical Society. Adapted with permission from ref. 97. Copyright 2022, Wiley-VCH GmbH.

2 ROS in inflammatory signaling

ROS are a series of derivatives of molecular oxygen, mainly consisting of hydrogen peroxide (H₂O₂), the superoxide anion radical (O₂˙⁻), and the hydroxyl radical (˙OH). Under physiological conditions, the intracellular concentration of ROS is stably maintained at a low nanomolar level thanks to the balance between the generation and elimination systems.²¹ However, once the ROS production activities pathologically exceed the scavenging abilities, the redox homeostasis will be disrupted and inclined to ‘oxidative stress’. Then over-accumulated ROS will not only induce antioxidant responses but also activate a series of pro-inflammatory signaling pathways. Herein, we focused on several major signaling pathways to clarify the pleiotropic role of ROS under inflammatory circumstances, hoping to clearly explain the tight connections between antioxidation and anti-inflammation strategies (Fig. 2).

Fig. 2 The interplay between reactive oxygen species (ROS) and the inflammatory signaling network. First, the ROS-induced oxidative stress will switch on the innate antioxidative Nrf2 defense system, which activates the transcription of a series of cytoprotective genes. However, over-produced ROS can also extensively participate in the activation of pro-inflammatory NF-κB, MAPK, and NO-related signaling pathways, which can inversely impair redox homeostasis and even further promote the over-production of ROS. Therefore, the vicious circle formed between ROS and inflammation indicates the significance of antioxidative strategies in anti-inflammatory therapy.

2.1 NF-κB signaling

Nuclear factor-κB (NF-κB) proteins are a family of transcription factors, including p65 (RelA), RelB, c-Rel, p105/p50 (NF-κB1), and p100/p52 (NF-κB2), playing a critical role in inflammatory and immune responses. In the resting state, NF-κB is tightly associated with its inhibitory proteins IκBs, and sequestered in the cytoplasm.^22,23 However, in the stressed state, a wide range of stimuli can activate the NF-κB pathway, including inflammatory cytokines, bacterial components, ultraviolet light, oxidative stress, etc. Intermediated by the generation of ROS,²⁴ these agonists signal through the canonical pathway and the noncanonical pathway, finally leading to the translocation of p50/RelA or p52 to the nucleus and transcription of multiple pro-inflammatory genes.^23–26 First, activated NF-κB can induce the polarization of macrophages towards the M1 subtype via the TLR4/MyD88/NF-κB axis, inducing mass secretion of pro-inflammatory molecules, such as TNF-α, IL-6, and IL-1β, directly promoting inflammatory reactions.²⁷ Additionally, it also triggers the assembly of the NLRP3 inflammasome, a multiprotein complex that is closely associated with the activity of caspase-1. Caspase-1 will further catalyze the maturation of pro-inflammatory IL-1β, thus forming a vicious circle to aggravate the inflammatory reaction and finally result in cell pyroptosis.^22,28–30

2.2 Nrf2 signaling

Nuclear factor erythroid 2-related factor 2 (Nrf2) is a transcription factor that regulates the expression of multiple cytoprotective genes, involving antioxidation, anti-inflammation, anti-apoptosis, and so on.³¹ Under physiological conditions, Nrf2 is integrated with its repressor protein, Kelch-like ECH-associated protein 1 (KEAP1), and only exists in the cytoplasm with low abundance and a short half-life.³² When exposed to excessive ROS, cells will initiate antioxidative Nrf2 signaling through modification of Keap1 or activation of the PI3K-Akt pathway, both contributing to the dissociation and translocation of Nrf2 from KEAP1 to the nucleus to promote the expression of heme oxygenase-1 (HO-1), glutathione peroxidase (GPx-1) and NAD(P)H.^32–35 Besides, Nrf2 can also inhibit inflammation by directly competing with the NF-κB pathway to prevent the generation of IL-6 and IL-1β.³¹ However, if the innate Nrf2 system can't sufficiently neutralize the superfluous ROS, it is possible for the BAX/caspase-3 axis-mediated apoptosis to aggravate inflammatory damage.^36,37 Therefore, it is rather necessary to strengthen Nrf2 signaling when treating inflammatory diseases.

2.3 MAPK signaling

Mitogen-activated protein kinases (MAPKs) are a family of protein serine/threonine kinases that mainly mediate the signal transduction process. The most studied conventional mammalian MAPKs include extracellular signal-regulated kinases (ERKs), c-Jun amino (N)-terminal kinases (JNKs), and p38.³⁸ Typically, ERKs mainly control cell proliferation, while JNKs and p38 are responsive to environmental stress, including ROS-induced oxidative stress, and pro-inflammatory cytokines TNF-α and IL-1β, thus extensively participating in inflammation and apoptosis processes.^39–41 Each MAPK axis signals through its specific sequential phosphorylation cascade and induces different downstream effects.⁴² Hence, based on the MAPK signaling, abundant studies have utilized the ROS-inducing medicine to kill malignant tumor cells^43,44 and the ROS-scavenging medicine to alleviate inflammation.^45,46

2.4 NO signaling

Nitric oxide (NO), the smallest known signaling molecule, is also an indispensable part of the oxidative-inflammatory signaling network. NO is synthesized by nitric oxide synthase (NOS), including neuronal NOS (nNOS), inducible NOS (iNOS), and endothelial NOS (eNOS),⁴⁷ among which only iNOS is expressed upon stimulation. Activated by various factors such as LPS, IFN-γ, IL-1β, TNF-α, and oxidative stress, macrophages will upregulate the expression of iNOS via complicated signaling pathways involving NF-κB, NLRP3, Akt, HO-1, and so on. Over-synthesized NO mainly acts as a detrimental factor in inflammation aggravation.⁴⁸ First, NO can react with ROS (O₂˙⁻) and produce peroxynitrite (ONOO⁻), a potent oxidant, directly damaging cellular components.⁴⁷ Besides, with the activation of iNOS⁺ M1 macrophages, the over-expression of NO is usually accompanied by mass production of TNF-α, IL-1β, and ROS, jointly resulting in serious inflammatory injury.²⁷ Moreover, ROS and NO also play a central role in the complicated redox regulation of the immune system, including macrophage polarization, neutrophil chemotaxis, and T/B lymphocyte differentiation, therefore indirectly mediating the inflammatory process.⁴⁹

In conclusion, oxidative stress and inflammatory responses tightly interweave with each other, forming a synergistic relationship. On the one hand, ROS can extensively participate in the activation of major pro-inflammatory signaling pathways. On the other hand, various inflammatory signaling pathways also involve the overproduction of ROS. This positive feedback loop contributes to the aggravation of destructive inflammatory reactions. From this perspective, the emerging antioxidative FNAs can effectively interrupt the vicious circle by scavenging excessive ROS, providing a promising strategy for anti-inflammatory therapy.

3 ROS-scavenging tFNAs in anti-inflammatory therapy

3.1 Fundamentals of tFNAs

Originally proposed by Turberfield and colleagues in 2004,⁵⁰ the elegant tetrahedral framework nucleic acids (tFNAs) quickly stand out from various DNA polyhedral nanostructures due to their extraordinary merits, thereby triggering a new era of FNA nanomaterials. A typical tFNA nanostructure basically consists of four precisely designed isometric (∼63 nt) single-stranded oligonucleotides. Every single strand should contain 3 complementary blocks (∼20 nt) to hybridize with the other three strands and 2 intermediate unhybridized ‘hinges’ (1–2 nt) to maintain 60° angles between adjacent edges of the final tetrahedron.^50,51 The most commonly adapted fabrication protocol of tFNAs is the convenient single-step synthesis. When the four kinds of ssDNA are equimolarly added into TM buffer, denatured at 95 °C for 10 minutes, and then annealed by quickly cooling to 4 °C for 20 minutes, they will self-assemble into tetrahedra through the base-pairing of previously designed complementary blocks.^3,11 After the self-assembly, each tFNA should possess 6 double helical DNA edges and 4 vertexes hanging a single oligonucleotide,^3,6 laying a solid foundation for further modifications with various oligonucleotides, small-molecular-weight drugs, macromolecular proteins, peptides, etc.

Oligonucleotides contain short DNA, RNA, aptamer molecules, etc. They can be loaded on tFNAs in two major ways. One is to be directly extended at the 5′- or 3′-ends of composing ssDNA via the bridges of several A or T nucleotides before the self-assembly process. The other is to be anchored to the sticky ends of the 5′- or 3′-ends of ssDNA after the synthesis of tetrahedral structures. Although both methods seem to be equal in hanging and releasing the medicine, the latter one is believed to avoid more interference and can broaden the sources of nucleic acids, which can functionalize as probes for biomarker sensing or disease diagnosis.^3,6 Besides, small-molecular-weight drugs such as traditional Chinese medicine monomers, anticancer drugs, metal complexes, etc. can be embedded into the helix of dsDNA. As for macromolecular proteins and peptides, researchers have also managed to synthesize peptide nucleic acids (PNAs) by replacing a short DNA sequence with a functional motif.⁵¹ Moreover, the tetrahedral DNA nanostructure naturally forms an interior caged space, which can be used to stably accommodate and deliver suitable functional molecules.

In addition to the abovementioned easy and productive fabrication as well as modifications, tFNAs and tFNA-derived complexes also possess other biological merits, such as excellent biocompatibility and structural stability. However, the preeminent membrane permeability particularly differentiated tFNAs from the other FNA nanomaterials. Traditionally, the polyanionic characteristic of DNA was considered hard to cross the hydrophobic cytoplasmic membrane. However, using the single-tracking technique, researchers found that tFNAs can be autonomously endocytosed via the caveolin-mediated pathway and then degraded in lysosomes, therefore achieving satisfactory cellular uptake.¹⁴ Hence, tFNAs have been widely used in the biomedical fields to regulate cell behavior, promote tissue regeneration, facilitate precise drug delivery, and so on.⁵¹

3.2 Innate anti-inflammatory effect of tFNAs via ROS scavenging

After Zhang et al. performed the first comprehensive analysis of the immunoregulatory effects of tFNAs on RAW264.7 cells in 2018,⁵² the attractive innate antioxidative and anti-inflammatory potential of tFNAs has aroused profound interest and investigation. First, although the antioxidative ability is one universal characteristic of all DNA-based structures, tFNAs still exhibit the best intracellular ROS-scavenging performance because of their outstanding cellular endocytosis.¹² What's more, due to the definite association between antioxidation and anti-inflammation, numerous studies have confirmed that tFNAs can extensively participate in the aforementioned signaling pathways as potent ROS scavengers and contribute to alleviating excessive inflammatory reactions.

First, tFNAs can reduce the high concentration of intracellular ROS to a normal level by upregulating the expression of HO-1, thus reversing the inflammatory state. Researchers have explored the therapeutic effects of tFNAs in a series of ROS-dominated disease models, including myocardial ischemia–reperfusion injury (MIRI),⁵³tert-butyl hydroperoxide (TBHP)-induced oxidative stress,⁵⁴ hepatic insulin resistance,⁵⁵ and osteoarthritis (OA) models.⁵⁶ These studies all detected less ROS in experimental groups than in the corresponding control groups and together attributed this effect to the activation of the Akt/Nrf2/HO-1 signaling pathway by tFNAs. Additionally, the Akt signaling pathway-mediated ROS scavenging can also contribute to the downregulated expression of pro-inflammatory cytokines, including IL-1β, IL-6, and TNF-α, hence preventing the further progression of inflammatory responses. Taking the wound healing process as an example, although the early inflammation phase plays a vital role in recruiting sufficient fibroblasts to guarantee the later epithelialization process, uncontrolled oxidative stress and inflammatory responses usually result in delayed healing or fibrosis. However, the emergence of tFNAs provides a promising anti-inflammatory therapy proposal. Lin et al. utilized tFNAs to treat diabetic mucosal wounds and achieved satisfactory Akt/Nrf2/HO-1 pathway activation, which stimulated the SOD activity to 1.22-fold and mediated ROS scavenging with decreased IL-6 (1.39-fold), TNF-α (2.10-fold), VCAM-1 (1.64-fold), ICAM-1 (1.62-fold), and selectin-E (1.45-fold), thus effectively elevating the wound recovering ability of animal models from 59.47% to 77.30% after 4 days.⁵⁷ Similarly, Zhu et al. used tFNAs to repair cutaneous wounds and also detected the activation of the Akt pathway as well as the reduction of IL-1β and TNF-α proteins, accelerating inflammation attenuation and scarless wound healing.⁵⁸ Since deficient ROS-scavenging competence still increases the risk of mitochondrial apoptosis, Xie et al. investigated the activity of the BCL-2/BAX/caspase-3 pathway before and after the treatment of tFNAs in a radiation-induced salivary gland inflammation model. As expected, they observed significantly decreased pro-apoptotic BAX (1.51-fold) and caspase-3 proteins (1.12-fold), increased pro-survival BCL-2 proteins (1.27-fold), and downregulated pro-inflammatory TNF-α, IL-6, and IL-1β (1.27-fold, 1.34-fold, and 1.89-fold respectively),⁵⁹ further complementing the anti-inflammation mechanism of ROS-scavenging tFNAs.

NF-κB and its related inflammatory signaling pathways are also the targets of ROS-scavenging tFNAs. As aforementioned, abnormally elevated intracellular ROS will activate the NF-κB pathway, leading to massive secretion of pro-inflammatory factors and the formation of the NLRP3 inflammasome. However, evidence shows that tFNAs can effectively prevent the above events. For example, Chen et al. established an in vitro sepsis model on RAW264.7 cells and treated it with tFNAs. They not only observed nearly 43% of the scavenging of ROS mediated by Nrf2, but also detected nearly 30% decreased NF-κB p65 and NF-κB p-p65 in the tFNA pre-treated group, consistent with the downregulated neutrophils infiltration in vivo, indicating the inhibition of the NF-κB pathway.⁶⁰ Considering the possible downstream signaling of NF-κB, Jiang et al. further detected the level of pyroptosis-related proteins in a skin fibrosis model and found that tFNAs can inhibit 35% of the expression of the NLRP3 inflammasome, with 12% caspase-1, 15% IL-1β, and 24% IL-18, thus helping to prevent pro-inflammatory pyroptosis.⁶¹ Moreover, given the close interrelations among ROS, NF-κB, macrophages, and NO, it is also necessary to figure out the immunoregulatory effect of ROS-scavenging tFNAs. Zhao et al. adapted LPS and IFN-γ to induce RAW264.7 cells polarizing to the pro-inflammatory M1 phenotype and then investigated the therapeutic role of tFNAs in osteonecrosis models. Statistics showed that tFNAs not only significantly reduced the excessive ROS again, but also reversed the polarization direction of macrophages. By attenuating the phosphorylation of STAT1 and enhancing the activation of STAT6 signaling pathways, tFNAs successfully transformed the pro-inflammatory M1 macrophages characterized by iNOS into anti-inflammatory M2 macrophages, with declined secretion of 35% NO, 44% TNF-α, 70% IL-6, and 31% IL-1β.⁶²

Inhibiting the MAPK signaling pathway is another anti-inflammatory mechanism of tFNAs. Zhou et al. developed periodontitis models and then used tFNAs for therapy. They not only detected downregulated ROS, TNF-α, IL-6, and IL-1β levels in in vitro cellular experiments, but also observed decreased inflammatory cell infiltration and pro-inflammatory cytokine secretion in in vivo periodontal tissues. Additionally, they also further explored the underlying mechanism and found an obviously inhibited expression of MAPK proteins,⁶³ consistent with the discovery of Zhang et al.,⁵² jointly verifying the ROS-scavenging and anti-inflammatory functions of tFNAs.

Given the aforementioned multi-facet antioxidative and anti-inflammatory properties, tFNAs have been used to treating numerous neuroinflammation diseases, in which the over-accumulated ROS act as a central initiator. For instance, ischemic stroke is a typical neural system disorder that involves ischemia/reperfusion-induced oxidative stress and inflammatory responses, causing severe motor and cognitive dysfunction. However, Zhou et al. innovatively explored the therapeutic potential of tFNAs in this disease by using oxygen–glucose deprivation/reoxygenation (OGD/R) models and ischemic stroke tMCAo rat models. Compared with the OGD/R group, the tFNA-incubated group significantly eliminated the excessive ROS in both neurons (SHSY-5Y cells) and astrocytes (RA cells) from 4.64 times to 1.66 times and from 1.73 times to 0.29 times, respectively, via inhibiting the TLRs/NF-κB signaling pathway, accompanied by the lower levels of pro-inflammatory factors like IL-1β, TNF-α, and IL-6.^64,65 In addition, tFNAs also contributed to the transformation of reactive astrocytes from the pro-inflammatory A1 phenotype to the neuroprotective A2 phenotype, further attenuating the inflammatory damage in ischemic hemispheres.⁶⁵ Besides, the pathogenesis of epilepsy is also closely associated with neuroinflammation and hyperexcitability caused by oxidative stress. Zhu et al. used a tFNA-based antiepileptic strategy and effectively inhibited nearly 43% of the production of ROS in the inflamed microglia/astrocyte cocultures in vitro, along with decreased concentrations of TNF-α, IL-6, and IL-1β from M1 microglia and IL-1α, TNF-α, and C1q from A1 astrocytes in vivo.⁶⁶ Moreover, Cui et al. found that tFNAs can also play a remarkable ROS-scavenging and neuroprotective role in Parkinson's disease (PD) models via activating the protein expression of Akt up to 2-fold and inhibiting the BAX/caspase-3 signaling pathways from 2.2 times to 1.7 times,⁶⁷ indicating the bright perspective of the anti-inflammatory applications of tFNAs.

3.3 Anti-inflammatory applications of ROS-scavenging tFNA-based drug delivery systems

Although unadorned tFNAs are already endowed with outstanding ROS-scavenging and anti-inflammatory abilities, there is still a considerable improvement space for targeted therapy and functional reinforcement. Taking advantage of their excellent editability, structural stability, and tissue permeability, researchers have exploited plentiful tFNA-based drug delivery systems, combining the innate merits of tFNAs with other specialized medicines to achieve more efficient and secure treatment. Fig. 3 summarizes the representative ROS-scavenging and inflammatory signaling modulation properties of tFNAs and tFNA-based drug delivery platforms. Hitherto, in terms of anti-inflammatory applications, tFNAs mainly cooperated with RNAi, aptamers, traditional Chinese medicine monomers, and therapeutic peptides, which will be explained in further detail below (Table 1).

Fig. 3 ROS-scavenging and inflammatory signaling modulation properties of tFNAs and tFNA-based drug delivery platforms. (A) Western blot and immunofluorescence staining results, showing that tFNAs can activate the Akt/Nrf2/HO-1 signaling pathway to alleviate oxidative stress induced by AGEs. Adapted with permission from ref. 57. Copyright 2020, American Chemical Society. (B) Immunofluorescence staining results, showing that tFNAs can upregulate the expression of the HO-1 protein. Adapted with permission from ref. 56. Copyright 2020, American Chemical Society. (C) Treatment of tFNAs significantly scavenging the intracellular ROS in RGC-5 cells. Adapted with permission from ref. 54. Copyright 2019, Royal Society of Chemistry. (D) Schematic diagram of the anti-inflammatory effects of tFNAs by participating in the modulation of Akt signaling pathways. Adapted with permission from ref. 67. Copyright 2019, American Chemical Society. (E) Schematic diagram of the extensive participation of tFNA-based drug delivery platforms in regulating inflammatory signaling pathways. Adapted with permission from ref. 71. Copyright 2022, American Chemical Society. (F) Schematic diagram illustrating the anti-inflammatory and chondroprotective effects of TWC. Adapted with permission from ref. 77. Copyright 2020, Springer Nature.

Table 1 Anti-inflammatory applications of ROS-scavenging tFNAs and tFNA-derived drug delivery platforms

Classification	Drug	Disease	Model		Pathways	Anti-inflammatory efficiency	Ref.
			In vitro	In vivo
		Acute myocardial infarction	H9c2 cells	N/A	Akt/Nrf2/HO-1↑	ROS↓	53
						Cell viability↑
		Retinal ischemia–reperfusion (I/R) injuries	Retinal ganglion cells (RGC-5s)	N/A	Akt/Nrf2/HO-1↑	ROS↓	54
						Cell viability↑
		Insulin resistance (IR)	HepG2 cells	T2DM C57BL/6L mice	PI3K/Akt↑	ROS↓	55
						Cell viability↑
		Osteoarthritis (OA)	Chondrocytes	N/A	Nrf2/HO-1↑	ROS↓	56
					BCL2/BAX/caspase-3↓	Cell viability↑
		Diabetic wound healing	Human umbilical vein endothelial cells (HUVECs)	Oral mucosal wounded diabetic Wistar rats	Akt/Nrf2/HO-1↑	ROS↓ NO↓	57
						IL-6↓ TNF-α↓
						VCAM-1↓ ICAM-1↓ electin-E↓
Simple tFNAs		Cutaneous wound healing	HaCaT cells	Skin-wounded SD rats	AKT↑	IL-1β↓ TNF-α↓	58
			HSF cells			VEGF↑ bFGF↑
		Sepsis	RAW264.7 cells	Septic BALB/c mice	Nrf2↑	ROS↓ NO↓
					NF-κB↓	NF-κB p65↓ NF-κB p-p65↓	60
						NLRP3 inflammasome↓
						IL-1β↓ IL-18↓
		Bisphosphonate-related osteonecrosis of the jaw (BRONJ)	HUVECs	BRONJ Wistar rats	STAT↑	ROS↓ NO↓
			RAW264.7 cells			IL-1β↓ IL-6↓ TNF-α↓	62
						M1 macrophages↓ iNOS↓
						M2 macrophages↑ TGFβ1↑ IL-10↑
		Periodontitis	PDLSCs	Ligature-induced periodontitis SD rats	MAPK/ERK↓	ROS↓	63
						IL-1β↓ IL-6↓ TNF-α↓
						Inflammatory cells↓
		Ischemic stroke	SHSY-5Y cells	Ischemic stroke t MCAo rats	TLRs/NF-κB↓	ROS↓
			Reactive astrocytes (RA)		BCL2/BAX/caspase-3↓	IL-1β↓ TNF-α↓ IL-6↓	64 and 65
						Cell viability↑
	BACE1-targeted aptamer (Bapt)	Alzheimer's disease (AD)	SH-SY5Y cells	APP-PS1 transgenic mice	N/A	ROS↓	75
						IL-1β↓ IL-6↓ caspase-3↓
						Microglia and astrocytes↓
	miRNA-22	Peripheral nerve injury (PNI)	Schwann cells (SCs)	Facial nerve (FN) crush-injured SD rats	ERK1/2↑	ROS↓ NO↓
			RAW264.7 cells		Nrf2↑	IL-1β↓ IL-6↓ TNF-α↓	70
					NF-κB↓	M1 macrophages↓ iNOS↓
						M2 macrophages↑ TGFβ1↑ IL-10↑
tFNAs + oligonucleotides	TLR2-targeted siRNA	Sepsis	RAW264.7 cells	Gout CD-1 mice	iNOS↓	ROS↓ NO↓
				LPS-induced abdominal inflammation CD-1 mice	TLR2/MyD88/NF-κB↓	IL-1β↓ IL-6↓ TNF-α↓	71
					TLR2/MyD88/MAPK↓	Neutrophils infiltration↓
	TNF-α-targeted siRNA	Inflammation	RAW264.7 cells	BALB/c mice	N/ATLR2/MyD88/PI3K/Akt↓	ROS↓ NO↓
			Murine peritoneal macrophages	LPS-treated C57BL/6 mice		TNF-α↓ IL-1β↓ IL-6↓	72
						iNOS↓
	C5a-targeted aptamer (aC5a)	Ischemic stroke	Primary SD rat neurons	Cerebral IRI SD rats	C5a/C5aR binding↓	ROS↓
			Primary SD rat polymorphonuclear neutrophils (PMNs)			TNF-α↓ IL-1β↓ caspase-3↓	74
			Primary SD rat microglial			C5a in vivo↓
						PMNS and microglial chemotaxis↓
	Wogonin	Osteoarthritis (OA)	Chondrocytes	Knee joint OA Wistar rats	NF-κB↓	TNF-α↓ IL-1β↓	77
					BAX/caspase-3↓	BCL-2↑
	Curcumin (Cur)	Acute gouty arthritis	RAW264.7 cells	Gout CD-1 mice	NF-κB↓Nrf2/HO-1↑	ROS↓ NO↓	76
tFNAs+ small-molecular-weight drugs						IL-1β↓ IL-6↓ TNF-α↓
						CD68 + macrophages↓
	Resveratrol (RSV)	Insulin resistance (IR)	RAW264.7 cells	High-fat diet (HFD)-induced IR C57BL/6L mice	N/A	M1 macrophages↓	78
						iNOS↓ IL-6↓ TNF-α↓
						M2 macrophages↑
						TGF-β↑ IL-10↑ Arg-1↑ CD206↑
						Th1 cells↓ Th17 cells↓
						Th2 cells↑ Treg cells↑
tFNAs+ peptides	REGRT healing peptide	Diabetic wound healing	HUVECs	Skin-wounded T2D db/db mice	ERK1/2↑	ROS↓	79
					HO-1↑	Cell viability↑

---

Small interfering RNAs (RNAi) such as miRNAs and siRNAs are important posttranscriptional gene expression regulators. They can specifically complement with targeted mRNAs and induce gene silencing.^68,69 Despite their powerful and subtle regulatory potential, their application is still limited by the unstable structures. However, the emerging tFNAs provide an ideal drug delivery platform. As a special kind of nucleic acid in nature, tFNAs tend to integrate with RNAi, simultaneously achieving maximal structural stability and therapeutic functionality. For example, Li et al. combined neural-protective miR-22 with tFNAs via the cohesive terminus and synthesized an MiD system, aiming to diminish inflammatory responses in RAW264.7 cells, thus facilitating the neuronal regeneration in a peripheral nerve injury (PNI) model. Statistics showed that compared to merely incubated with miR-22 or tFNAs, the MiD-treated group exhibited the best ROS-scavenging performance, reaching 11.0%, 13.0%, and 25.2%, respectively, which was paralleled with the downregulated trend of pro-inflammatory factors IL-1β, IL-6, TNF-α, and CCL2 and upregulated anti-inflammatory factors IL-10 and TGF-β, indicating the repolarization of macrophages from the M1 to M2 subtype.⁷⁰ Similarly, Zhang et al. developed a tFNAs-based TLR2-targeted siRNA delivery platform to cope with severe inflammation in the sepsis model. Through the synergistic modulation of tFNA-mediated elimination of ROS and siRNA-mediated inhibition of TLR2/MyD88/NF-κB and TLR2/MyD88/MAPK signaling pathways, this novel drug delivery system successfully achieved around 60% decrease of IL-1β and IL-6 and up to 70% decrease of TNF-α in vitro, showing the maximal anti-inflammatory effect compared with unadorned tFNAs.⁷¹ Moreover, enlightened by the dynamic DNA nanostructures, researchers have made further progress in the responsive and sequential drug release strategies. Gao et al. utilized tFNAs as a nanobox to encapsulate TNF-α-targeted siRNA and constructed a smart lysosome-activated pH-responsive nanobox-siR system. When researchers incubated LPS-pretreated macrophages with different nucleic acid medicines, the nanobox-siR system achieved the most percentages of TNF-α gene silencing (∼75%) and ROS scavenging (∼33%). Therefore, the expressions of other relevant pro-inflammatory molecules, such as IL-6, IL-1β, iNOS, and NO, are also inhibited to different extents, proving the powerful anti-inflammatory function of the tFNA-based drug delivery system.⁷² Moreover, Tian et al. synthesized a CCR2-targeted tFNA-siCcr2 system and also verified its anti-inflammatory function in a liver cirrhosis model through a specific modulation of macrophage and neutrophil phenotypes via the NF-κB signaling pathway.⁷³

Aptamers are short oligonucleotide sequences that can specifically bind with their corresponding ligands, therefore holding a profound prospect for precise drug delivery. Like RNAi, the combination of aptamers and tFNAs is also intrinsically convenient and compatible, thus becoming a promising anti-inflammatory therapeutic strategy. In practice, considering the potent pro-inflammatory function of C5a in ischemic stroke, Li et al. incorporated bipyramidal FNAs with C5a-targeted aptamers (aC5a), synthesizing an aC5a-FNAs system. In an in vitro experiment, aC5a-FNAs not only scavenged about 40% of intracellular ROS, but also specifically blocked the C5a/C5aR binding in a concentration-dependent way, thus inhibiting 1–2-fold of the chemotaxis of pro-inflammatory microglia and polymorphonuclear neutrophils (PMNs). Consistently, after the treatment of aC5a-FNAs, in vivo cerebral ischemic penumbra also showed that the concentration of C5a in the systemic plasma is reduced from ∼7.5 ng ml⁻¹ to ∼5 ng ml⁻¹, along with ∼44% decreased TNF-α and ∼60% decreased IL-1β, further confirming the synergistic antioxidative and anti-inflammatory functions of FNAs and C5a-targeted aptamers.⁷⁴ In Alzheimer's disease (AD), however, oxidative stress and neuroinflammation are mainly induced by the abnormal deposition of amyloid β (Aβ) proteins, which are generated by β-site APP cleavage enzyme 1 (BACE1). Therefore, by targeting the production process of Aβ, Wang et al. innovatively loaded BACE1-targeted aptamers (Bapt) to tFNAs and constructed a tFNA–Bapt complex. When put into application, both tFNAs and the tFNA–Bapt complex achieved satisfactory elimination of excessive ROS, while the tFNA–Bapt complex better inhibited the activation of microglia and astrocytes, and decreased the subsequent secretion of IL-1β and IL-6 by nearly 55% and 25%, respectively, compared with that of Bapt alone, indicating an enhanced anti-inflammatory function.⁷⁵

Traditional Chinese medicine is well known for its mild but effective therapeutic effects. However, the poor water solubility and short biological half-life still greatly hamper its extensive application. To resolve the above problems, Zhang et al. used tFNAs as an excellent drug delivery vehicle to load the natural polyphenol curcumin (Cur) and synthesized Cur-tFNA nanoparticles, hoping to investigate its inflammation preventive capacity in gout. Compared with merely treated with tFNAs, the Cur-tFNA therapy exhibited a more powerful and stable effect on eliminating ROS and pro-inflammatory factors. Statistics showed that Cur-tFNAs effectively downregulated the protein expression level of NF-κB from 1.7-fold to only 0.5-fold, and meanwhile upregulated the blocked protein expression level of Nrf2 from 0.4-fold to 1.3-fold and that of HO-1 from 0.9-fold to 1.2-fold, indicating that Cur-tFNAs can inhibit the expression of pro-inflammatory factors through blocking the NF-κB signaling pathway and activating the Akt/Nrf2/HO-1 signaling pathway.⁷⁶ Likewise, utilizing the multi-facet therapeutic function and special DNA incorporation trait of wogonin, Shi et al. fabricated a tFNA/wogonin complex (TWC) to prevent the progression of osteoarthritis (OA). By inhibiting the NF-κB signaling pathway, pre-treatment with the TWC can more potently reduce the expression levels of TNF-α and IL-1β and meanwhile upregulate the gene level of BCL-2, thus effectively protecting chondrocytes against inflammatory damage.⁷⁷ Moreover, Li et al. designed a resveratrol (RSC)-loaded tFNA (tFNA-RSV) platform to treat insulin resistance, in which chronic low-grade tissue inflammation is established as an ultimate reason. Statistical analysis revealed that in terms of modulating the transformation of macrophages and T cells from the pro-inflammatory phenotype to the anti-inflammatory phenotype, the tFNA-RSV platform performed better than RSV or tFNAs alone, thus successfully breaking the links between obesity and IR through attenuating inflammation.⁷⁸

Moreover, tFNAs also provide an attractive resolution for the membrane-crossing delivery of therapeutic peptides or proteins, avoiding enzymatic degradation and promoting bioavailability. For instance, Lin et al. fabricated a novel tFNA-based healing peptide (REGRT) delivery system called p@tFNA to help in controlling the oxidative stress in diabetic cutaneous wounds. Statistics showed that modification with healing peptide not only further strengthened the ROS-scavenging ability of simple tFNAs from 0.03-fold to 0.67-fold, but also elevated the phosphorylation of ERK1/2 to 1.49-fold, therefore improving the angiogenesis and wound closure processes.⁷⁹

Although extensive research studies have confirmed that stable and biocompatible tetrahedral DNA nanostructures are naturally endowed with the ability to neutralize ROS and trigger a rather wide range of cytoprotective pathways, including Akt/Nrf2/HO-1, BCL2/BAX/caspase-3, MAPK/ERK, and NF-κB signaling, the specific trigger mechanism still requires further investigation. First, due to the complicated interaction between ROS and inflammation, the elimination of ROS by DNA self-oxidation can be considered as an important regulatory factor. Second, since the basic metabolic process of tFNAs undergoes cell uptake, lysosome degradation, and fragment release, it is instructive to figure out whether tFNAs possess general breaking sites or patterns and trace different fragments to observe whether they can act on the aforementioned signal molecules.

4 ROS-scavenging DNA origami in anti-inflammatory therapy

4.1 Fundamentals of DNA origami

In 2006, Paul W. K. Rothemund first described the concept of ‘scaffolded DNA origami’, marking a brand new milestone in the development of DNA nanotechnology.⁸⁰ As its name suggests, the DNA origami technique refers to using numerous short ssDNAs as staples to complement with and fold a major long ssDNA into a specific nanostructure.^80,81 By flexibly introducing sticky ends, multi-arm junctions, curvatures, etc., scientists have exploited various DNA origami frameworks, including two-dimensional rectangles, triangles, five-point stars, smiley faces, maps of China, etc., and three-dimensional cuboidal boxes, spheres, footballs, prisms, tetrahedra, even nanosized flasks, and so on.^81,82

These complicated structures not only enrich the geometric shapes and spatial orders of the DNA origami but also provide a broad space for further modifications. Choosing addressable and quantifiable short staple strands as major carriers, researchers have integrated multiple functional molecules such as fluorophores, lipids, aptamers, peptides, and polymers into the DNA origami via covalent or non-covalent bonds.⁸³

Meanwhile, diverse DNA origami nanostructures (DONs) can also exhibit precise addressability and flexible cargo loading capacity.^81,84 Especially, 50–400 nm-scaled homogeneous DNA origami nanostructures have been reported to exhibit optimal drug delivery properties. Therefore, as an excellent framework nucleic acid nanomaterial, DNA origami has been extensively used in multiple biological fields, including biosensing, biophysics, biomedicine and so on.^16,85

4.2 Anti-inflammatory applications of ROS-scavenging DNA origami therapeutic systems

As a DNA molecule in nature, DNA origami has the ability to scavenge ROS and resist oxidative stress, which was systematically demonstrated by Jiang and colleagues in 2018. In this study, researchers established an acute kidney injury (AKI) model and fabricated a series of DONs, including rectangular, triangular, and tubular DNA origami, to explore their therapeutic effect. First, using positron emission tomography (PET) and confocal imaging techniques, researchers successfully tracked the distribution and excretion of the three kinds of DONs and found that all of them can preferentially accumulate in the kidneys, laying a solid foundation for DON-based renal disease therapy. Since the rhabdomyolysis-induced AKI is usually accompanied by an overproduction of ROS, researchers further tested the antioxidative ability of DONs. Both extracellular and intracellular experiments showed that DONs can neutralize about 30% ROS after 0.5 hours and about 50% after 2 hours of treatment, possibly via directly oxidizing their own composing DNA bases, thus restoring the superoxide dismutase (SOD) levels and protecting important cellular components against oxidative damage. Moreover, immunogenicity tests also confirmed that these DONs won't increase the levels of pro-inflammatory factors IL-6 and TNF-α, thus being a biocompatible and secure ROS-scavenging nanomedicine.⁸⁶ As a result, given the tight connections between ROS and inflammation, the excellent antioxidative properties of DNA origami undoubtedly make it a promising anti-inflammatory candidate.

To further strengthen the structural stability and anti-inflammatory functionality of DNA origami, researchers designed various modification proposals. For example, still focusing on the AKI therapy, Chen et al. proposed a C5a-targeted aptamer (aC5a)-functionalized rectangular DNA origami nanostructure and developed a dynamic sequential drug-releasing nanodevice named aC5a-rDONs. At the beginning 4 hours of renal ischemia–reperfusion (I/R), also called stage I, the aC5a-rDONs first decreased nearly 3.6 folds of the intracellular ROS, alleviating oxidative stress. Then after another 4 hours, as the pathological process progressed to stage II, the well-retained renal aC5a-rDONs started to competitively bind to C5a receptors, thereby blocking the downstream pro-inflammatory responses. Besides, researchers also observed that the aC5a-rDONs can effectively inhibit the secretion of pro-inflammatory factors at 24 h post-surgery, with a reduction of TNF-α reaching 24.5%, IL-6 reaching 48.5%, and IL-1β reaching 72.7%, respectively.⁸⁷ From another perspective, Li et al. adopted cytokine immunotherapy. Using the rectangular DNA origami as a nanoraft to carry and continuously release IL-33, they significantly promoted the expansion of renal anti-inflammatory immune cells, including group 2 innate lymphoid cells (ILC2s), M2 macrophages, and Tregs, jointly contributing to the protection of IRI kidneys.⁸⁸

In other diseases, various DNA origami-derived therapeutic systems also show prominent ROS-scavenging and anti-inflammatory potential. For instance, using folic acid (FA) to modify triangular DNA origami nanostructures (tDONs), Ma et al. fabricated FA-tDONs, which can successfully attenuate the inflammatory progression of rheumatoid arthritis (RA). Statistics showed that when directly exposed to extracellular O₂˙ and ˙OH with the same number of DNA bases, the ROS-neutralizing ability of FA-tDONs and tDONs didn't exhibit obvious differences, both surpassing the incompletely folded M13 DNA nanostructure, indicating the innate advantages of the compact triangular structure. However, when applied to activated pro-inflammatory M1 macrophages, FA-tDONs scavenged about 20% more intracellular ROS than simple tDONs, proving the strengthened targeting ability of FA. Further investigations revealed that compared with tDONs, FA-tDONs can also significantly reduce the expression of pro-inflammatory biomarkers iNOS to only 26%, TNF-α to 55%, and IL-6 to 44%, while promoting the expression of anti-inflammatory biomarkers CD206 to 3.6 folds and IL-10 to 10.3 folds, thus facilitating the repolarization of macrophages from the M1 to the M2 subtype and blocking the inflammatory responses.⁸⁹ Moreover, targeting neuroinflammation, Zhu et al. synthesized a topotecan (TPT)-loaded DNA origami nanostructure called TopoGami, hoping to inhibit the myeloid-specific topoisomerase 1 (TOP1) in microglia. This special complex also effectively downregulated around 50% gene expression of TNF-α, 93% IL-1β, and 86% IL-6, therefore providing a new therapeutic strategy for resolving inflammatory diseases in the neural system.⁹⁰

The most widely studied DNA origami nanostructures are rectangular and triangular nanostructures, which exhibited similar ROS-scavenging capacity in the aforementioned articles. However, since the sizes and shapes of DNA origami are highly changeable, we believe that it will arouse profound interest to investigate the structural stability of different origami geometrical shapes. Moreover, to confirm whether DNA nanomaterials can universally scavenge ROS by sacrificing their deoxyribonucleotides, comparisons of the ROS-scavenging efficiency among DNA origami nanostructures in different scales are especially necessary because they provide different numbers of reductive units.

5 ROS-scavenging DNA hydrogels in anti-inflammatory therapy

5.1 Fundamentals of DNA hydrogels

DNA hydrogels are a kind of biological material that contains physically or chemically cross-linked DNA chains in a hydrophilic polymeric framework. According to the composition and fabrication strategies, DNA hydrogels can be sorted into two categories. One is hybrid DNA hydrogels, in which DNA functions as cross-linking points of other backbone materials, such as polyacrylamides, polypeptides, proteins, graphene oxides, etc. The other is pure DNA hydrogels, which completely consist of self-assembled or rolling circle amplified branched and linear DNA scaffolds.^91,92

The introduction of DNA imparts traditional hydrogels with unique programmability and tunable properties and researchers have developed various modification strategies to strengthen the biomedical functions of DNA hydrogels. First, by properly designing branched DNA scaffolds, 2D lattices and 3D dendrimers can be successfully constructed to connect the desired molecules like oligonucleotides.⁹³ Second, the porous structure in hybrid hydrogels can be utilized to stably encapsulate and release small-molecular-weight drugs. Moreover, the polyanionic characteristic of DNA hydrogels also allows for electrostatic interactions with cationic medicines.⁹⁴

Apart from the excellent editability, the combination of DNA and hydrogels also increases other advantages. On the one hand, the secondary structures of DNA can make hydrogels highly bio-responsive to various stimuli, including temperature, pH, metal ions, proteins, DNA, RNA, etc.,⁹² therefore becoming a smart and dynamic system. On the other hand, interior hydrogen bonding, π–π stacking, and hydrophilic/hydrophobic interactions can also intensify the mechanical properties and meanwhile restore the self-healing and thixotropic properties of DNA hydrogels.⁹³ Hence, due to these strengths, DNA hydrogels are becoming more and more popular in biomedical and biosensing fields.^95,96

5.2 Anti-inflammatory applications of ROS-scavenging DNA hydrogel therapeutic systems

Impaired wound healing ability in diabetes mellitus is considered to be closely related to the advanced glycation end products (AGEs)-induced overproduction of ROS and chronic inflammation, which urgently requires effective and symptomatic treatment approaches. Fortunately, novel ROS-scavenging DNA hydrogels with satisfactory biocompatibility, soft texture, and adaptable shapes can perfectly meet the requirements for wound dressing to promote anti-inflammatory recovery. Wang et al. proved that DNA hydrogels alone can already neutralize up to 4-folds of intracellular ROS, probably through oxidizing and breaking their DNA strands. Meanwhile, they also detected an obviously decreased mRNA level of pro-inflammatory factor IL-1β from 2.6 times to less than 1 times, indicating the latent anti-inflammatory function of DNA hydrogels via ROS scavenging.⁹⁷ To further enhance the anti-inflammatory properties, researchers used DNA hydrogels to encapsulate immunomodulatory cytokine IL-33 and fabricated a new dressing material called IL-33-cytogel. As expected, the IL-33-cytogel effectively induced anti-inflammatory immune cells such as ILC2s from 6% to 12% and M2 macrophages from 25% to 50% with a decrease of pro-inflammatory factors IL-1β, TNF-α, and iNOS, thus contributing to accelerated wound closure.⁹⁷ Similarly, Kim et al. incorporated salmon sperm-derived DNA and silica particles into alginate hydrogels and innovatively synthesized a 3D-printed dressing biomaterial called DNA-bSi30@FSA. In this system, DNA functions as a ROS scavenger mainly through activating adenosine A_2A receptors and thereby regulating the mitochondrial activity to inhibit ROS generation and the inflammatory reaction.⁹⁸ Furthermore, enlightened by the similar double-stranded structures of siRNA and dsDNA, Lei et al. used a hydrogel compound to load tannic acid (TA) and MMP9-targeted siRNA, forming a smart ROS-responsive PHTB(TA-siRNA) hydrogel. Both extracellular and intracellular experiments showed this hydrogel can preferentially oxidize and disintegrate its interior borate ester bonds, thus releasing therapeutic TA-siRNA nanoparticles to further modulate the polarization of macrophages and attenuate inflammatory responses.⁹⁹ Apart from cutaneous wounds, the repair of diabetic bone defects is also facing anti-inflammatory challenges. To address this problem, Li et al. combined anti-inflammatory cytokine IL-10 and DNA hydrogels to fabricate a new bio-scaffold called ILGel, hoping to promote bone regeneration in diabetic periodontitis. Statistics showed that after 3-day treatment with ILGel, the damaged sites exhibited nearly 5-fold expansion of anti-inflammatory M2 macrophages, consistent with 90% downregulation of TNF-α, 95% downregulation of MCP-1, and 70% downregulation of IL-1β, demonstrating the promising therapeutic properties of ILGel under inflammatory circumstances.¹⁰⁰

In other inflammatory diseases, DNA hydrogels also play a satisfactory ROS scavenger role. Targeting the osteoarthritis (OA) treatment, Zhang et al. developed a DNA supramolecular hydrogel (DSH)-based metformin (MET) delivery platform and named it MET@DSH. Compared to the MET group, MET@DSH not only rendered the best ROS-scavenging effect, but also alleviated local inflammation with around 1.65-folds of TNF-α and more than 5-folds of IL-6 mRNA reduction in vitro.¹⁰¹Fig. 4 summarizes the multi-facet anti-inflammatory applications of ROS-scavenging DNA origami and DNA hydrogel nanosystems (Table 2).

Table 2 Anti-inflammatory applications of ROS-scavenging DNA origami and DNA hydrogel nanomaterials

Classification	Drug	Disease	Model		Anti-inflammatory efficiency	Ref.
			In vitro	In vivo
DNA origami	Folic acid (FA)	Rheumatoid arthritis (RA)	RAW264.7 cells	Collagen-induced arthritis (CIA) mice	ROS↓ NO↓	89
					M1 macrophages↓
					iNOS↓ IL-6↓ TNF-α↓
					M2 macrophages↑
					CD206↑ IL-10↑
					Paw swelling↓ Bone erosion↓
	C5a-targeted aptamer	Acute kidney injury (AKI)	HK-2 cells	Renal ischemia–reperfusion (I/R) C57BL/6 mice	ROS↓	87
			HEK293 cells		IL-1β↓ IL-6↓ TNF-α↓ C5a↓
					Cell viability↑
	IL-33	Acute kidney injury (AKI)	N/A	Renal ischemia–reperfusion (I/R) C57BL/6 mice	M1 macrophages↓	88
					iNOS↓ IL-1β↓ TNF-α↓
					M2 macrophages↑
					IL-10↑ Arg-1↑
DNA hydrogels	IL-33	Diabetic wound healing	HaCaT cells	Streptozotocin (STZ)-induced diabetic skin wounded C57BL/6 mice	ROS↓	97
					ILC2s↑ M2 macrophages↑ Tregs↑
					IL-1β↓ TNF-α↓ iNOS↓ IL-10↑
					Granulation tissue regeneration↑
					Diabetic wound closure↑
	Tannic acid (TA)	Diabetic wound healing	RAW264.7 cells	LPS-induced skin tissue inflammation in SD rats	ROS↓	99
	siRNA			Streptozotocin (STZ)-induced diabetic SD rat's skin defects	TNF-α↓ IL-6↓ IL-10↑
				Streptozotocin (STZ)-induced diabetic SD rats skin burns	Inflammatory cells↓
					Tissue regeneration↑
	IL-10	Diabetic alveolar bone defect	Peritoneal macrophages	Streptozotocin (STZ)-induced diabetic alveolar bone-defected C57BL/6 mice	M2 macrophages↑	100
			Primary bone marrow stromal cells (BMSCs)		TNF-α↓ MCP-1↓ IL-1β↓
					IL-10↑ Arg-1↑

---

Fig. 4 Multi-facet anti-inflammatory applications of ROS-scavenging DNA origami and DNA hydrogel nanosystems. (A) DONs can selectively accumulate in the kidneys and locally scavenge ROS to alleviate oxidative stress and alleviate AKI. Adapted with permission from ref. 86. Copyright 2018, Springer Nature. (B) Schematic diagram of the preferential renal accumulation of rDONs, which facilitates the sequential therapy of AKI in multi stage. Adapted with permission from ref. 87. Copyright 2021, American Chemical Society. (C) Schematic diagram of the rational design and the proposed RA therapeutic mechanism of FA-tDON nanomedicine. Adapted with permission from ref. 89. Copyright 2022, American Chemical Society. (D) Schematic diagram of the repair of diabetic chronic wounds using the combination of ES therapy and adaptive conductive PHTB(TA–siRNA) hydrogels. Adapted with permission from ref. 99, Copyright 2022, Wiley-VCH GmbH.

Compared to tFNAs and DNA origami, the composition of DNA hydrogels appears more complex and variable. Therefore, we suggest more investigations into the drug release and delivery processes of DNA hydrogels to better demonstrate the metabolic characteristics and biosafety grade. Additionally, the underlying regulatory function and the mechanism of DNA hydrogels on multiple signaling pathways still lack enough research, which also require further exploration.

6 Conclusion and perspective

The overproduction of ROS is always tightly associated with the intricate inflammatory signaling network, synergistically forming a vicious circle that aggravates inflammatory injuries in multiple tissues and organs. Fortunately, with the advent of DNA nanotechnology, various FNA nanoparticles emerge and function as effective ROS scavengers, providing a promising strategy to interrupt the detrimental feedback and contribute to resolving excessive inflammation and promoting recovery.

Although different FNAs have been widely proven to scavenge excessive ROS under inflammatory conditions and trigger various signaling pathways, the specific mechanisms remain ambiguous and require further investigation. However, thanks to the complicated crosstalk between ROS and inflammation, we consider the elimination of ROS as an important regulatory factor. To demonstrate the process of ROS scavenging, researchers have proposed several potential hypotheses. First, since all of these DNA nanomaterials are basically composed of deoxyribonucleotides, they can directly sacrifice themselves to be oxidized by ROS and consequently protect other important intracellular or extracellular substances from oxidative damage. Second, they are found to regulate the mitochondrial activity via adenosine A_2A receptors, thus effectively controlling the generation of intracellular ROS. What's more, due to the diverse composition, most DNA hydrogels can preferentially oxidize their interior chemical or physical bonds, timely responding to oxidative stress. Finally, these highly editable FNAs can also flexibly collaborate with other powerful antioxidative and anti-inflammatory drugs to achieve more efficient and targeted therapy.

Despite the competitive ROS-scavenging potential of FNAs for anti-inflammatory therapy, massive efforts still need to be made to overcome the emerging challenges in their clinical translation. First, since FNAs neutralize excessive ROS partly by self-oxidation and disintegration, it remains to clarify whether these reaction products or broken fragments are toxic or harmful to the human body. Second, most aforementioned studies only focused on the transient pharmacologic actions of FNAs, while neglecting the long-term biocompatibility. Therefore, we highly recommend increasing investigations into the long-run pharmacokinetics and possible adverse reactions of various FNA systems. Additionally, since there are not enough horizontal comparisons among the therapeutic effects of different FNA nanomaterials, it is also difficult to select the relatively best material for further development. Finally, the high cost and relatively low yield of chemically synthesized DNA still greatly limited the large-scale production of therapeutic FNAs. Although researchers have made some breakthroughs in the mass production of precursor ssDNA using bacteriophages¹⁰² and plasmid DNA by fermentation,¹⁰³ the specific practicability and immunogenicity are still unsubstantiated. As DNA technology-based targeting therapy becomes a research hotspot, we hopefully expect that there will emerge more and more innovative proposals to address these issues and bring these promising ROS scavengers into clinical practice.

Author contributions

Y. J. Zhu and R. J. H. Shi conceived and wrote the manuscript. W. T. Lu, S. R. Shi, and Y. Chen reviewed and edited the manuscript. All authors contributed to the article and approved the submitted version.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (82101077), the Sichuan Science and Technology Program (2023NSFSC1516), the Postdoctoral Science Foundation of China (2021M692271, 2023T160455), West China School/Hospital of Stomatology Sichuan University, No. RCDWJS2023-5, the Fundamental Research Funds for the Central Universities, and the Research and Develop Program, West China Hospital of Stomatology Sichuan University.

References

1. L. Chen, J. Zhang, Z. Lin, Z. Zhang, M. Mao, J. Wu, Q. Li, Y. Zhang and C. Fan, Acta Pharm. Sin. B, 2022, 12, 76–91.
2. Q. Hu, H. Li, L. Wang, H. Gu and C. Fan, Chem. Rev., 2019, 119, 6459–6506.
3. Y. Lin, Q. Li, L. Wang, Q. Guo, S. Liu, S. Zhu, Y. Sun, Y. Fan, Y. Sun, H. Li, X. Tian, D. Luo and S. Shi, Int. J. Oral Sci., 2022, 14, 51.
4. W. Ma, Y. Zhan, Y. Zhang, C. Mao, X. Xie and Y. Lin, Signal Transduction Targeted Ther., 2021, 6, 351.
5. N. C. Seeman, J. Theor. Biol., 1982, 99, 237–247.
6. T. Zhang, T. Tian, R. Zhou, S. Li, W. Ma, Y. Zhang, N. Liu, S. Shi, Q. Li, X. Xie, Y. Ge, M. Liu, Q. Zhang, S. Lin, X. Cai and Y. Lin, Nat. Protoc., 2020, 15, 2728–2757.
7. Y. Dong, C. Yao, Y. Zhu, L. Yang, D. Luo and D. Yang, Chem. Rev., 2020, 120, 9420–9481.
8. Q. Hu, S. Wang, L. Wang, H. Gu and C. Fan, Adv. Healthc. Mater., 2018, 7, e1701153.
9. F. Yang, Q. Li, L. Wang, G.-J. Zhang and C. Fan, ACS Sens., 2018, 3, 903–919.
10. Z. Ge, H. Gu, Q. Li and C. Fan, J. Am. Chem. Soc., 2018, 140, 17808–17819.
11. S. Li, T. Tian, T. Zhang, X. Cai and Y. Lin, Mater. Today, 2019, 24, 57–68.
12. T. Zhang, T. Tian and Y. Lin, Adv. Mater., 2022, 34, e2107820.
13. M. R. Jones, N. C. Seeman and C. A. Mirkin, Science, 2015, 347, 1260901–1260901.
14. L. Liang, J. Li, Q. Li, Q. Huang, J. Shi, H. Yan and C. Fan, Angew. Chem., Int. Ed., 2014, 53, 7745–7750.
15. Y. Wang, X. Lu, X. Wu, Y. Li, W. Tang, C. Yang, J. Liu and B. Ding, Innovation, 2022, 3, 100217.
16. Q. Jiang, S. Liu, J. Liu, Z. G. Wang and B. Ding, Adv. Mater., 2019, 31, e1804785.
17. A. Keller and V. Linko, Angew. Chem., Int. Ed., 2020, 59, 15818–15833.
18. B. Zhu, L. Wang, J. Li and C. Fan, Chem. Rec., 2017, 17, 1213–1230.
19. Y. Guo, J. Tang, C. Yao and D. Yang, Wiley Interdiscip. Rev.: Nanomed. Nanobiotechnol., 2022, 14, e1753.
20. R. Han, Y. Xiao, Q. Bai and C. H. J. Choi, Acta Pharm. Sin. B, 2023, 13, 1847–1865.
21. H. Sies and D. P. Jones, Nat. Rev. Mol. Cell Biol., 2020, 21, 363–383.
22. H. Yu, L. Lin, Z. Zhang, H. Zhang and H. Hu, Signal Transduction Targeted Ther., 2020, 5, 209.
23. G. Gloire, S. Legrand-Poels and J. Piette, Biochem. Pharmacol., 2006, 72, 1493–1505.
24. S. Nakajima and M. Kitamura, Free Radicals Biol. Med., 2013, 65, 162–174.
25. M. J. Morgan and Z.-G. Liu, Cell Res., 2010, 21, 103–115.
26. P. J. Jost and J. R. Ruland, Blood, 2006, 109, 2700–2707.
27. S. Chen, A. F. U. H. Saeed, Q. Liu, Q. Jiang, H. Xu, G. G. Xiao, L. Rao and Y. Duo, Signal Transduction Targeted Ther., 2023, 8, 207.
28. T. Bergsbaken, S. L. Fink and B. T. Cookson, Nat. Rev. Microbiol., 2009, 7, 99–109.
29. D. Capece, D. Verzella, I. Flati, P. Arboretto, J. Cornice and G. Franzoso, Trends Immunol., 2022, 43, 757–775.
30. M. M. McDaniel, L. C. Kottyan, H. Singh and C. Pasare, Cell Rep., 2020, 31, 107604.
31. A. Cuadrado, A. I. Rojo, G. Wells, J. D. Hayes, S. P. Cousin, W. L. Rumsey, O. C. Attucks, S. Franklin, A.-L. Levonen, T. W. Kensler and A. T. Dinkova-Kostova, Nat. Rev. Drug Discovery, 2019, 18, 295–317.
32. M. Yamamoto, T. W. Kensler and H. Motohashi, Physiol. Rev., 2018, 98, 1169–1203.
33. A. Loboda, M. Damulewicz, E. Pyza, A. Jozkowicz and J. Dulak, Cell. Mol. Life Sci., 2016, 73, 3221–3247.
34. B. Li, M. I. Nasser, M. Masood, S. Adlat, Y. Huang, B. Yang, C. Luo and N. Jiang, Biomed. Pharmacother., 2020, 126, 110074.
35. L. Baird and A. T. Dinkova-Kostova, Arch. Toxicol., 2011, 85, 241–272.
36. H. Blaser, C. Dostert, T. W. Mak and D. Brenner, Trends Cell Biol., 2016, 26, 249–261.
37. E. Vringer and S. W. G. Tait, Cell Death Differ., 2022, 30, 304–312.
38. M. Cargnello and P. P. Roux, Microbiol. Mol. Biol. Rev., 2011, 75, 50–83.
39. T. Takata, S. Araki, Y. Tsuchiya and Y. Watanabe, Int. J. Mol. Sci., 2020, 21, 8750.
40. L. R. Coulthard, D. E. White, D. L. Jones, M. F. McDermott and S. A. Burchill, Trends Mol. Med., 2009, 15, 369–379.
41. K. L. Jeffrey, M. Camps, C. Rommel and C. R. Mackay, Nat. Rev. Drug Discovery, 2007, 6, 391–403.
42. E. K. Kim and E.-J. Choi, Biochim. Biophys. Acta, Mol. Basis Dis., 2010, 1802, 396–405.
43. S. Li, J. Chen, Y. Fan, C. Wang, C. Wang, X. Zheng, F. Chen and W. Li, Signal Transduction Targeted Ther., 2022, 7, 49.
44. S.-Y. Hung, W.-F. Chen, Y.-C. Lee, J.-H. Su, Y.-S. Juan, I.-P. Lin, Y.-H. Zhang, M.-K. Chang, M.-Y. Lin, C.-Y. Chen and C.-H. Lee, Phytomedicine, 2021, 92, 153720.
45. G. Yang, C.-C. Chang, Y. Yang, L. Yuan, L. Xu, C.-T. Ho and S. Li, J. Agric. Food Chem., 2018, 66, 12953–12960.
46. C. Liu, F. Hu, G. Jiao, Y. Guo, P. Zhou, Y. Zhang, Z. Zhang, J. Yi, Y. You, Z. Li, H. Wang and X. Zhang, J. Nanobiotechnol., 2022, 20, 65.
47. U. Forstermann and W. C. Sessa, Eur. Heart J., 2011, 33, 829–837.
48. C. Bogdan, Trends Immunol., 2015, 36, 161–178.
49. G. Morris, M. Gevezova, V. Sarafian and M. Maes, Cell. Mol. Immunol., 2022, 19, 1079–1101.
50. R. P. Goodman, R. M. Berry and A. J. Turberfield, Chem. Commun., 2004, 1372–1373,  10.1039/b402293a.
51. T. Zhang, W. Cui, T. Tian, S. Shi and Y. Lin, ACS Appl. Mater. Interfaces, 2020, 12, 47115–47126.
52. Q. Zhang, S. Lin, S. Shi, T. Zhang, Q. Ma, T. Tian, T. Zhou, X. Cai and Y. Lin, ACS Appl. Mater. Interfaces, 2018, 10, 3421–3430.
53. M. Zhang, J. Zhu, X. Qin, M. Zhou, X. Zhang, Y. Gao, T. Zhang, D. Xiao, W. Cui and X. Cai, ACS Appl. Mater. Interfaces, 2019, 11, 30631–30639.
54. X. Qin, N. Li, M. Zhang, S. Lin, J. Zhu, D. Xiao, W. Cui, T. Zhang, Y. Lin and X. Cai, Nanoscale, 2019, 11, 20667–20675.
55. Y. Li, Y. Tang, S. Shi, S. Gao, Y. Wang, D. Xiao, T. Chen, Q. He, J. Zhang and Y. Lin, ACS Appl. Mater. Interfaces, 2021, 13, 40354–40364.
56. S. Shi, T. Tian, Y. Li, D. Xiao, T. Zhang, P. Gong and Y. Lin, ACS Appl. Mater. Interfaces, 2020, 12, 56782–56791.
57. S. Lin, Q. Zhang, S. Li, T. Zhang, L. Wang, X. Qin, M. Zhang, S. Shi and X. Cai, ACS Appl. Mater. Interfaces, 2020, 12, 11397–11408.
58. J. Zhu, M. Zhang, Y. Gao, X. Qin, T. Zhang, W. Cui, C. Mao, D. Xiao and Y. Lin, Signal Transduction Targeted Ther., 2020, 5, 120.
59. X. Xie, W. Ma, G. Li, Y. Zhan, L. Quan and X. Cai, Cell Proliferation, 2023, 56, e13381.
60. X. Chen, J. He, Y. Xie, T. Zhang, S. Li, Y. Zhao, N. Hu and X. Cai, Cell Proliferation, 2023, e13424,  DOI:10.1111/cpr.13424.
61. Y. Jiang, S. Li, T. Zhang, M. Zhang, Y. Chen, Y. Wu, Y. Liu, Z. Liu and Y. Lin, ACS Appl. Mater. Interfaces, 2022, 14, 15069–15079.
62. D. Zhao, W. Cui, M. Liu, J. Li, Y. Sun, S. Shi, S. Lin and Y. Lin, ACS Appl. Mater. Interfaces, 2020, 12, 44508–44522.
63. M. Zhou, S. Gao, X. Zhang, T. Zhang, T. Zhang, T. Tian, S. Li, Y. Lin and X. Cai, Bioact. Mater., 2021, 6, 1676–1688.
64. M. Zhou, T. Zhang, B. Zhang, X. Zhang, S. Gao, T. Zhang, S. Li, X. Cai and Y. Lin, ACS Nano, 2022, 16, 1456–1470.
65. M. Zhou, T. Zhang, X. Zhang, M. Zhang, S. Gao, T. Zhang, S. Li, X. Cai, J. Li and Y. Lin, ACS Appl. Mater. Interfaces, 2022, 14, 37478–37492.
66. J. Zhu, Y. Yang, W. Ma, Y. Wang, L. Chen, H. Xiong, C. Yin, Z. He, W. Fu, R. Xu and Y. Lin, Nano Lett., 2022, 22, 2381–2390.
67. W. Cui, Y. Zhan, X. Shao, W. Fu, D. Xiao, J. Zhu, X. Qin, T. Zhang, M. Zhang, Y. Zhou and Y. Lin, ACS Appl. Mater. Interfaces, 2019, 11, 32787–32797.
68. A. Fire, S. Xu, M. K. Montgomery, S. A. Kostas, S. E. Driver and C. C. Mello, Nature, 1998, 391, 806–811.
69. Q. Lu, R. Wu, M. Zhao, A. Garcia-Gomez and E. Ballestar, Trends Pharmacol. Sci., 2019, 40, 853–865.
70. J. Li, Y. Yao, Y. Wang, J. Xu, D. Zhao, M. Liu, S. Shi and Y. Lin, Adv. Mater., 2022, 34, e2202513.
71. X. Zhang, M. Zhang, M. Zhou, T. Zhang, Y. Gao, S. Li, Y. Lin and X. Cai, ACS Appl. Mater. Interfaces, 2022, 14, 6442–6452.
72. Y. Gao, X. Chen, T. Tian, T. Zhang, S. Gao, X. Zhang, Y. Yao, Y. Lin and X. Cai, Adv. Mater., 2022, 34, e2201731.
73. T. Tian, C. Zhao, S. Li, Z. Huang, Y. Guo, W. Dai, R. Bai, C. Tang, Y. Lin and J. Gao, ACS Appl. Mater. Interfaces, 2023, 15, 10492–10505.
74. S. Li, D. Jiang, Z. T. Rosenkrans, T. E. Barnhart, E. B. Ehlerding, D. Ni, J. W. Engle and W. Cai, Nano Lett., 2019, 19, 7334–7341.
75. Y. Wang, J. Zhu, W. Jia, H. Xiong, W. Qiu, R. Xu and Y. Lin, ACS Appl. Mater. Interfaces, 2022, 14, 44228–44238.
76. M. Zhang, X. Zhang, T. Tian, Q. Zhang, Y. Wen, J. Zhu, D. Xiao, W. Cui and Y. Lin, Bioact. Mater., 2022, 8, 368–380.
77. S. Sirong, C. Yang, T. Taoran, L. Songhang, L. Shiyu, Z. Yuxin, S. Xiaoru, Z. Tao, L. Yunfeng and C. Xiaoxiao, Bone Res., 2020, 8, 6.
78. Y. Li, S. Gao, S. Shi, D. Xiao, S. Peng, Y. Gao, Y. Zhu and Y. Lin, Nano-Micro Lett., 2021, 13, 86.
79. S. Lin, Q. Zhang, S. Li, X. Qin, X. Cai and H. Wang, Cell Proliferation, 2022, 55, e13279.
80. P. W. Rothemund, Nature, 2006, 440, 297–302.
81. P. Zhan, A. Peil, Q. Jiang, D. Wang, S. Mousavi, Q. Xiong, Q. Shen, Y. Shang, B. Ding, C. Lin, Y. Ke and N. Liu, Chem. Rev., 2023, 123, 3976–4050.
82. F. Hong, F. Zhang, Y. Liu and H. Yan, Chem. Rev., 2017, 117, 12584–12640.
83. G. A. Knappe, E.-C. Wamhoff and M. Bathe, Nat. Rev. Mater., 2022, 8, 123–138.
84. C. J. Kearney, C. R. Lucas, F. J. O'Brien and C. E. Castro, Adv. Mater., 2016, 28, 5509–5524.
85. S. Loescher, S. Groeer and A. Walther, Angew. Chem., Int. Ed., 2018, 57, 10436–10448.
86. D. Jiang, Z. Ge, H. J. Im, C. G. England, D. Ni, J. Hou, L. Zhang, C. J. Kutyreff, Y. Yan, Y. Liu, S. Y. Cho, J. W. Engle, J. Shi, P. Huang, C. Fan, H. Yan and W. Cai, Nat. Biomed. Eng., 2018, 2, 865–877.
87. Q. Chen, F. Ding, S. Zhang, Q. Li, X. Liu, H. Song, X. Zuo, C. Fan, S. Mou and Z. Ge, Nano Lett., 2021, 21, 4394–4402.
88. W. Li, C. Wang, H. Lv, Z. Wang, M. Zhao, S. Liu, L. Gou, Y. Zhou, J. Li, J. Zhang, L. Li, Y. Wang, P. Lou, L. Wu, L. Zhou, Y. Chen, Y. Lu, J. Cheng, Y. P. Han, Q. Cao, W. Huang, N. Tong, X. Fu, J. Liu, X. Zheng and P. O. Berggren, ACS Nano, 2021, 15, 18237–18249.
89. Y. Ma, Z. Lu, B. Jia, Y. Shi, J. Dong, S. Jiang and Z. Li, ACS Nano, 2022, 16, 12520–12531.
90. K. Zhu, Y. Wang, H. Sarlus, K. Geng, E. Nutma, J. Sun, S. Y. Kung, C. Bay, J. Han, J. H. Min, I. Benito-Cuesta, H. Lund, S. Amor, J. Wang, X. M. Zhang, C. Kutter, A. O. Guerreiro-Cacais, B. Hogberg and R. A. Harris, EMBO Rep., 2022, 23, e54499.
91. Y. Wei, K. Wang, S. Luo, F. Li, X. Zuo, C. Fan and Q. Li, Small, 2022, 18, e2107640.
92. D. Wang, Y. Hu, P. Liu and D. Luo, Acc. Chem. Res., 2017, 50, 733–739.
93. Y. Shao, H. Jia, T. Cao and D. Liu, Acc. Chem. Res., 2017, 50, 659–668.
94. Q. Wang, Y. Qu, Z. Zhang, H. Huang, Y. Xu, F. Shen, L. Wang and L. Sun, Gels, 2022, 8, 400.
95. F. Mo, K. Jiang, D. Zhao, Y. Wang, J. Song and W. Tan, Adv. Drug Delivery Rev., 2021, 168, 79–98.
96. F. Li, D. Lyu, S. Liu and W. Guo, Adv. Mater., 2019, 32, e1806538.
97. Z. Wang, W. Li, L. Gou, Y. Zhou, G. Peng, J. Zhang, J. Liu, R. Li, H. Ni, W. Zhang, T. Cao, Q. Cao, H. Su, Y. P. Han, N. Tong, X. Fu, E. Ilegems, Y. Lu, P. O. Berggren, X. Zheng and C. Wang, Adv. Healthc. Mater., 2022, 11, e2200782.
98. N. Kim, H. Lee, G. Han, M. Kang, S. Park, D. E. Kim, M. Lee, M. J. Kim, Y. Na, S. Oh, S. J. Bang, T. S. Jang, H. E. Kim, J. Park, S. R. Shin and H. D. Jung, Adv. Sci., 2023, 10, e2300816.
99. H. Lei and D. Fan, Adv. Sci., 2022, 9, e2201425.
100. W. Li, C. Wang, Z. Wang, L. Gou, Y. Zhou, G. Peng, M. Zhu, J. Zhang, R. Li, H. Ni, L. Wu, W. Zhang, J. Liu, Y. Tian, Z. Chen, Y. P. Han, N. Tong, X. Fu, X. Zheng and P. O. Berggren, ACS Appl. Mater. Interfaces, 2022, 14, 25173–25182.
101. C. Zhang, H. Huang, J. Chen, T. Zuo, Q. Ou, G. Ruan, J. He and C. Ding, ACS Appl. Mater. Interfaces, 2023, 15, 16369–16379.
102. F. Praetorius, B. Kick, K. L. Behler, M. N. Honemann, D. Weuster-Botz and H. Dietz, Nature, 2017, 552, 84–87.
103. T. Nöll, S. Wenderhold-Reeb, H. Schönherr and G. Nöll, Angew. Chem., Int. Ed., 2017, 56, 12004–12008.
104. S. Gao, Y. Li, D. Xiao, M. Zhou, X. Cai and Y. Lin, Nano Lett., 2021, 21, 4437–4446.
105. Y. Liu, Y. Sun, S. Li, M. Liu, X. Qin, X. Chen and Y. Lin, Nano Lett., 2020, 20, 3602–3610.

---

Footnote

† These authors contributed equally to this work.

---