Classical HCR was firstly developed by Dirks and Pierce (Dirks and Pierce, 2004), which involves an initiator and two metastable hairpins. The initiator could trigger the alternate opening of hairpins, forming nicked double-stranded DNA structures. As a simple and powerful amplification method, HCR has been incorporated into a number of sensing platforms. A high-fidelity amplified FISH system was reported by Marras et al. (2019), in which the target mRNA interacted with the binary probe to release the HCR initiator. This system could distinguish single nucleotide variations within a single mRNA. In another study, Yang et al. (2019) realized the simultaneous detection of miRNA and pre-miRNA in living cells through two pairs of HCR hairpin probes, showing detection limits of 680 pM and 820 pM, respectively (Figure 2A). Later, Tang et al. (2020) integrated HCR with the dual-split aptamer probe (DSAP) for activatable cancer cell detection. In this design, DSAP could specifically recognize cancer cells and release the trigger for HCR, leading to in-situ HCR on the cell membrane and amplified fluorescence. This strategy could reach a detection limit of 20 cells/200 uL. Overall, these studies proved the potential of HCR in biosensing and bioimaging.

Branched hybridization chain reaction (bHCR) involves hierarchical reactions of HCR, producing branched DNA polymers and higher levels of amplification efficiency. For example, Wang’ group (Wei et al., 2018a) developed a concatenated HCR system for miRNA detection. As shown in Figure 2B, target RNA could trigger the first layer of HCR, exposing the initiator for the second layer of HCR, resulting in a shortened distance between FRET donor and acceptor. The detection limit was calculated to be 3 pM. Soon after, the same group integrated the concept of cascade HCR with logic gate operation for intracellular miRNA imaging. In this study, the presence of miRNA was transduced to the initiator of cascading HCR through biocomputing operations. By rationally designing the sensing module, binary or advanced logic gates were achieved. (Gong et al., 2019). To obtain better sensitivity, electrochemical detection has also been implemented with branched HCR for miR-141 detection, achieving a LOD of 0.18 fM (Li et al., 2019). Additionally, by integrating aptamers into the branched HCR system, biomolecules, such as ATP, can be detected. In a study by Zhou et al. (2019), the binding of ATP aptamer to the target released the initiator for branched HCR, leading to FRET signals with a detection limit of 170 nM.

The kinetics of typical HCR is limited since the working principle is based on random collision of hairpins. To accelerate the reaction, localized HCR (LHCR) was developed based on the spatial-confinement effect. The reaction time could be reduced by anchoring the reactants to DNA nanostructures in order to reach higher local concentration of hairpins. For instance, Bui et al. (2017) immobilized the HCR hairpins along a 1D linear DNA track, which enabled the hairpins to hybridize along the track. The kinetics of this strategy was six times faster than traditional HCR due to the locality effect. To further apply the concept of LHCR to sensing applications; Wu et al. (2020a) developed a tripartite Y-shaped DNA probe for mRNA imaging in living mice. Y-shaped scaffolds were functionalized with HCR hairpins and a folate (FA) ligand that could selectively enter tumor cells, leading to HCR assemblies on the scaffold in response to miR-21. Moreover, taking advantage of the stability of Y-shaped structure, the probe was successfully applied for high-contrast miRNA imaging in living mice. In addition, 3D DNA nanostructures have also been employed for LHCR. As illustrated in Figure 2C, Wang et al. (2019a) reported hyperbranched HCR mediated by tetrahedral DNA nanostructures (qTDN). By modifying HCR hairpins on the vertices of qTDN, the reaction orientation was extended, as well as the collision probability and reaction kinetics. The reaction rate is about 70 times faster than that of traditional HCR, and the detection limit was 2.14 pM. These reports confirmed the acceleration effect of LHCR.

In addition to the strategies mentioned above, other HCR technologies have been developed. A recent study by Cheng et al. (2021) used proximity-dependent surface hybridization chain reactions to design an electrochemical biosensor for miRNA detection. MiRNA triggered HCR and the formation of duplexes with single-stranded dangling tails, which were labeled with ferrocene (Fc) and could be used to identify the sequence fixed on the electrode. After identification, the Fc tags were brought proximal to the electrode, which induced amplified electrochemical signals with a LOD of 3 fM. Furthermore, Jiang et al. (2021) presented a dumbbell hybridization chain reaction (DHCR), which improved the amplification performance by a novel morphology of the DNA nanostructure. This system consisted of four probes (A, B, C, D), in which probes A and B contained PSA split aptamers that could recognize PSA. However, in the absence of PSA, the unhybridized parts of probes A and B could alternatively open dumbbell probes C and D, forming a stacked DNA nanostructure and enhancing electrochemical signals. The LOD for PSA was 0.5 pg/mL. Moreover, HCR can be regarded as an ideal choice for target cell imaging and gene delivery. Ma et al. (2019) reported a bipedal hybridization chain (BiHCR) for activable cancer cell imaging (Figure 2D). A Y-shaped nanoprobe was constructed containing an i-motif sequence, an aptamer sequence and two trigger strands for HCR. After specific binding between the aptamer and cancer cells, the acidic extracellular environment triggered a conformational change in the i-motif structure, exposing the HCR initiators. Which further induced in-situ bipedal HCR and produced FRET signals. In addition, Peng et al. (2022) developed a double loop-stem mediated HCR (D-HCR) for label-free biosensing. The presence of miR-21 not only triggered the formation of DNA nanowires, but also released previously locked G4 aptamers. The LOD of 4.60 pM was achieved by combining D-HCR with thioflavin T (ThT) for fluorescent detection, and 13.60 pM was achieved by incorporating Hemin for colorimetric detection.

The CHA was firstly introduced by Yin’s group (Yin et al., 2008), in which an initiator could catalyze the duplex formation between two metastable hairpins, resulting in the amplification of double-stranded DNA products. Owing to the feature of simple operation, mild conditions, and flexibility, CHA has been widely applied in the development of sensors. For example, a universal CHA system (uniCHA) was proposed by Zhang et al. (2022) to detect various piRNAs and miRNAs. An adaptable starting hairpin containing the CHA initiator was employed in this system. The targets would interact with the starting hairpin to expose the initiator, which could trigger CHA and generate FRET signals. It is noteworthy that the sequence of the starting sequence can be adjusted depending on the target. Therefore, this system has been applied to the detection of various piRNAs and miRNAs. In addition, CHA can be integrated with lateral flow immunoassay (LFIA) to detect SARS-CoV-2 RNA (Zou et al., 2021). In this study, target RNA catalyzed the formation of a H1-H2 duplex, which was then detected by LFIA via digoxin/biotin recognition. The detection limit was calculated to be 2000 copies/mL. These studies demonstrate the potential of CHA for clinical diagnosis.

Branched CHA (bCHA) usually employ three or more hairpin probes to form three-way or four-way junctions upon the initiator. Due to the higher energy barriers, bCHA had a much lower background signal than traditional CHA, providing higher detection sensitivity. Yue et al. (2019) employed bCHA for miRNA imaging in living cells by anchoring three CHA hairpins on a Y-shaped DNA scaffold. After entering the cell, the probe could undergo bCHA through the catalysis of miR-155, leading to the formation of DNA dendrimers and in-situ imaging of miRNA. In another study, Liu et al. (2020) proposed a miRNA imaging method based on in-situ growth of DNA nanostructures by introducing palindromic sequences to bCHA hairpins. According to Figure 2E, upon miR-21, bCHA could be triggered to form Y-shaped structures, which could further assemble into 3D nanospheres owing to the palindromic fragments at the end of the hairpins. The process of bCHA also caused separation of fluorophore and quencher, leading to the recovery of fluorescence and reaching a LOD of 1.4 pM. Other than miRNA detection, bCHA has also found its application in antibiotic detection, Zhang et al. (2019) developed an aptasensor for the simultaneous detection of chloramphenicol (CAP) and kanamycin (Kana). In the presence of CAP or Kana, CHA hairpins can be catalyzed to form three-arm or four-arm branched junctions, which could be distinguished by microchip electrophoresis (MCE) based on migration time. And the detection limits for CAP and Kana were 0.49 pg/mL and 0.52 pg/mL, respectively.

Localized CHA (LCHA) takes advantage of the spatial-confinement effect by limiting CHA hairpins in a compact area, leading to higher amplification efficiency and reaction rate. Similar to LHCR, the hairpins could be fixed on 1D, 2D, or 3D DNA nanostructures to realize LCHA, such as DNA origami (Dai et al., 2021), Y-shaped scaffolds (Wu et al., 2020b), DNA nanocubes (Liu et al., 2019; Feng et al., 2022), DNA triangular prisms (DTP) (Yang et al., 2020), and DNA nanolanterns (Wang et al., 2022a). Recently, a nanowire-based LCHA for miRNA imaging was reported by Wei et al. (2018b). Target RNA catalyzed the hybridization between adjacent hairpins on the nanowire, leading to the recovery of quenched fluorescence. The LOD of this method was 2.0 pM, which was about 11.8 times lower than typical CHA. In addition, Bai et al. (2022) developed a dispersion-to-localization of catalytic hairpin assembly (DL-CHA) strategy. In the presence of miR-21, CHA was induced between hairpins anchored to linker strands, which promoted the formation of DNA ladders. A high signal-to-background (S/B) ratio was reached at 37.8, with a wide linear range of 100 pM to 100 nM. Moreover, Qing et al. (2020) developed an intramolecular CHA strategy for mRNA imaging, where the distance between reactants were shortened by anchoring the CHA hairpins on a DNA tetrahedron (Figure 2F). The local concentration of hairpins was about 808 times higher than free CHA, which increased the initial reaction rate by 15.6 times. Later, an intermolecular CHA strategy for FRET imaging of miRNA-155 in cells was proposed by Wang et al. (2022b). By anchoring CHA hairpins on separate DNA tetrahedrons, the targets could trigger CHA between DNA nanostructures, resulting in polymerized products and amplified FRET signals.

Cascaded CHA could achieve exponential signal growth by interconnecting multiple CHA circuits. For example, Guk et al. (2019) reported a self-circulation CHA system for miRNA detection. According to their design, miR-34a initiated the first CHA cycle and released the catalyst for the second CHA cycle, which in turn produced the initiator for the first cycle. Owing to its versatility, this strategy has been adapted for interferon-γ (IFN-γ) detection, reaching a LOD of 0.6 pM (Wang et al., 2019b). Besides, another promising approach is the self-replicating catalytic hairpin assembly system (SRCHA), in which only two hairpins were required for the cascade reaction. As illustrated in Figure 2G, the target could trigger the first CHA cycle and generate duplexes with exposed trigger sequences. Therefore, the CHA product may serve as a new CHA initiator. This system has been successfully applied to the detection of metal ions, biomolecules, and nucleic acids (Dai et al., 2017; Pan et al., 2019).

The cascaded integration of CHA and HCR is expected to provide higher amplification efficiency than single amplification. As shown in Figure 2H, Wang et al. (2018) established a general CHA-HCR sensing platform, where the outputs of CHA induced by the target miRNA could serve as the initiators for HCR circuits, showing a LOD of 2 pM. Later, Li et al. (2022) integrated the concept of palindrome-based HCR (PHCR) with CHA for miR-21 detection. The presence of miR-21 catalyzed the formation of CHA products, which further initiated HCR between hairpins with palindromic domains, leading to the formation of crosslinking networks through the assembly of HCR products. With amplified FRET signals, target RNA could be detected down to 10 pM. The strategy of HCR-CHA has also been employed in sensors. Wei et al. (2016) integrated split CHA initiators into HCR hairpins. The presence of miR-144 could trigger HCR and co-localize the split sequences, which further initiated the CHA process and induced amplified fluorescence, reaching the LOD of 0.3 fM.