Optically tunable catalytic cancer therapy using enzyme-like chiral plasmonic nanoparticles

Characterization of enzyme-like chiral plasmonic nanoparticles

The chiral L-/D-Au were prepared followed by seed-mediated method with the introduction of L-/D-GSH ligand during the growth process as referred to previous literature^32,33,37. The morphological characterization was proceeded by measuring transmission electron microscopy (TEM) images (Supplementary Fig. S1a, c) and scanning electron microscopy (SEM) images (Supplementary Fig. S1b, d). TEM images displayed the cube-like L/D-Au with surfaces split in opposite directions with an average diameter of ~100 nm. As-prepared D-Au was selected for GOD-mimic enzyme-like nanoparticles, facilitating the enantioselective interaction with the biological D-glu substrate. Following the synthesis of L-Au, a Pd shell was introduced by reducing the Pd precursor in the presence of ascorbic acid (Supplementary Fig. S2a). The successful incorporation of Pd was verified with SEM images (Supplementary Fig. S2b), TEM images (Fig. 1b and Supplementary Fig. S2c), and high-resolution (HR)-TEM images (Supplementary Fig. S2d–f). Compared to the bare L-Au (Supplementary Fig. S1c, d), bumpy structures were observed at the L-AuPd surface, demonstrating the successful deposition of the Pd NP shell on the L-Au surface. The d-spacing of deposited Pd NPs is around 0.23 nm, corresponding to the (111) plane (Supplementary Fig. S2f). Energy dispersive X-ray (EDX) mapping results showcased the homogeneous distribution of the Pd element (red) at the Au surface (yellow), further suggesting the well-prepared L-AuPd for POD-mimic enzyme-like nanoparticles (Supplementary Fig. S3a). X-ray photoelectron spectroscopy (XPS) spectroscopy was surveyed to validate the oxidation state of Pd at the L-Au surface (Supplementary Fig. S3b, c). The binding energy of Pd 3d_5/2 and 3d_3/2 was obtained at 334.5 and 339.8 eV, respectively, for metallic Pd, while partially oxidized Pd existed at the Au surface with increased binding energy at 336.5 and 342 eV. Existing Pd (II) ions would be reduced to metallic Pd, serving as a catalytic active site by a localized hot electron from L-Au, which is feasible for POD-like activity under LC irradiation. The optical properties of prepared enzyme-like chiral plasmonic nanoparticles were then assessed (Supplementary Fig. S4). D-Au and L-Au showed mirror symmetry peaks at 610 nm and 800 nm in the circular dichroism (CD) spectra (Supplementary Fig. S4a) and strong localized surface plasmon resonance (LSPR) bands were monitored in the range of visible and near-infrared region as depicted in Uv-vis-NIR spectra (Figure S4c). The g-factor of L-Au at 808 nm is around 0.0329 while the value of D-Au at 808 nm is -0.0302, suggesting similar CPL-responsiveness of D- and L-Au (Supplementary Fig. S4b). These results confirmed the opposite chiroptical properties of D- and L-Au in the range of vis-NIR region. The LSPR peaks, CD peaks, and g-factor of L-AuPd exhibited noticeable red-shift compared to those of L-Au (Supplementary Fig. S4d–f), attributed to changes in the refractive index surrounding L-Au and the large imaginary number of Pd dielectric function. The g-factor of L-AuPd at 808 nm is 0.0281, indicating well-maintained L-AuPd chirality for using L-AuPd as chiroptical-dependent POD-like enzyme-like nanoparticles (Supplementary Fig. S4e). Despite the small g-factors, hot charge carriers generated under the continuous light irradiation can produce steady-state population of ROS. This can induce differences in the number of ROS molecules photogenerated over time, thereby explaining the significant variations in catalytic activity kinetics between the two light polarizations³⁴.

Mechanism of chiral- and chiroptical-dependent enhanced catalytic performance

We then performed the theoretical analysis to investigate the chiroptical-dependent enhancement of catalytic performance in our enzyme-like chiral nanoparticles (Fig. 2). Light-matter interaction of D-Au and L-AuPd were simulated to confirm the chiroptical properties via finite-difference time-domain (FDTD) simulation under the CPL illumination³³ (Fig. 2a). The differences in these fields under the LC and RC excitation illustrated the asymmetric responses of the electric fields. In the case of D-Au, it exhibited the locally enhanced electric field with a right-handed direction under RC. In contrast, the opposite direction of electric field enhancement was observed in L-Au or L-AuPd under LC irradiation (Fig. 2a, Supplementary Fig. S5). These results emphasized the chiroptical responses of both chiral plasmonic nanostructures, which facilitate the generation of hot charge carriers under the corresponding CPL illuminations (RC for D-Au and LC for L-AuPd). We then performed density functional theory (DFT) calculations to verify the impact of hot charge carriers on the catalytic performance of both GOD- mimic D-Au and POD- mimic L-AuPd under CPL illumination³⁸ (Fig. 2b–d). Firstly, the influence of hot electrons on the catalytic performance of L-AuPd was investigated. As illustrated in Fig. 1d, LC illumination induced the generation of hot electrons at the Au site and generated hot electrons were concentrated at the Pd catalytic site to decompose the H₂O₂. Previous research demonstrated the enrichment of electrons on Pd through hot electron transfer from the Au site, as evidenced by X-ray absorption near-edge structure (XANES) spectroscopy measurements under light illumination³⁹. These findings support the hypothesis that hot electrons are transferred from plasmonic Au to Pd. Thus, for L-AuPd, the calculation was carried out by dividing it into the L-Au surface and the Pd surface. Figure 2b shows the adsorption energy of H₂O₂ homolytic dissociation on the Pd surface as a function of hot electron level (σ)^40,41. With increasing the electron density at the Pd site, the adsorption energy of homolytic dissociation (HOOH*)^TS decreased, confirming that the enrichment of electrons at the Pd site facilitates the decomposition of H₂O₂. The impact of hot holes at glucose oxidation reaction was then simulated for both D-Au and L-Au surfaces as a function of hot hole level (σ), as illustrated in Fig. 2c, d. The reaction mechanism of noble metal-mediated GOD-like activity is the same as that of natural GOD, except for the role of OH^–, which is used to abstract the H⁺ from glucose. The H⁺ in glucose (C₆H₁₂O₆ + 2OH^–) is abstracted by the surrounding OH^–, forming an intermediate (first step: C₆H₁₁O₆ + H₂O + e^–). Then, the direct hydride transfer occurs from the glucose to the Au surface, forming C₆H₁₀O₆ and releasing 2e^– (second step). Finally, generated electrons reduce the oxygen to produce H₂O₂⁴². Based on the GOD mentioned above reaction pathway, the initial adsorption of D-glu on each of D-Au, achiral Au, and L-Au was simulated by DFT calculations (Supplementary Fig. S6a–c). The initial adsorption energy of C₆H₁₂O₆ + 2OH^– on the D-Au (−6.536 eV) surface is lower compared to that on the L-Au (−6.499 eV) and achiral Au (−6.410 eV) surface, suggesting that the most stable adsorption of D-glu occurred at the D-Au surface. After confirming the initial adsorption of D-glu, we calculated the adsorption energy for each oxidation reaction step of D-glu at both D-Au and L-Au with increasing the hole level at the Au surface. In the case of D-Au, by increasing the hole level at each step, the adsorption energy was dramatically reduced, demonstrating that the plasmonic hot holes facilitated the oxidation of glucose to gluconic acid. In contrast, in the case of L-Au, a moderate decrease in adsorption energy was observed in the first step, and no significant trend was observed in the second step. These might result from the lower substrate specificity of D-glu for L-Au compared to D-Au, as described in Supplementary Fig. S6. These findings emphasized that the adsorption of D-glu on D-Au was much more stable than on L-Au at each oxidation step, promoting the hole-mediated glucose oxidation reaction to form H₂O₂ and an acidic environment by producing gluconic acid. To further demonstrate the enantioselective GOD-like reaction of our enzyme-like chiral plasmonic nanoparticles, isothermal titration calorimetry (ITC) was conducted at 25 °C. ITC measurement clarified the thermodynamic profile of the binding and dissociation behavior between D-glu substrates and both D- and L-Au catalysts^27,43 (Fig. 2e–h). The thermogram was obtained by mixing the D-glu substrates solution with D- or L-Au, respectively, as displayed in Fig. 2e, g. Exothermic responses were observed for both D- and L-Au, suggesting that the binding between D-glu and D-/L-Au generates heat. Based on raw ITC thermogram, the integrated heat as a function of molar ratio of D-glu to D-Au or L-Au was obtained and fitted with a simple one-site binding model to investigate the thermodynamic profiles and determine the dissociation constant (K_d) (Fig. 2f, h). The dissociation constant of D-glu with D-Au was 28.2μM, while it was found to be 60.9μM in the L-Au and D-glu system. These results suggested that the binding affinity between D-Au and D-glu was 2 times higher than that of L-Au and D-glu system, emphasizing the enantioselective GOD-like reaction was more facilitated on the D-Au surface than L-Au. These findings were consistent with the above DFT calculation results (Supplementary Fig. S6). In addition, thermodynamic profiles were obtained from fitting data, and the difference of Gibbs free energy (ΔG = ΔH-TΔS, T = 278 K) was calculated. ΔG during the binding of D-glu on D-Au was −6.3 ±  0.3 kcal mol⁻¹ (ΔH = −7.039 ±  0.33 kcal mol⁻¹; ΔS = −2.78 cal mol⁻¹deg⁻¹), whereas −6.0 ±  0.5 kcal mol⁻¹ (ΔH = −8.802 ±  0.57 kcal mol⁻¹; ΔS = −10.2 cal mol⁻¹deg⁻¹) was obtained for the binding of D-glu on L-Au surface. These results indicated that the binding process of D-glu on the D-Au surface was slightly more favorable than on the L-Au surfaces. Taken together, enantioselective substrate specificity and chiroptical-dependent hot carrier generation were likely to promote the enzymatic reactions of both D-Au mediated GOD-like and L-AuPd mediated POD-like activities.

a Schematic illustration showing the plane above the nanoparticle surface utilized for simulation under CPL irradiation and simulated electric field differences under LC and RC excitation for D-Au and L-AuPd, respectively. The color scale represents the magnitude of the field difference, with red indicating positive difference and blue indicating negative difference. b Adsorption energy for H₂O₂ homolytic dissociation on the Pd surface as a function of the electron level (σ). The adsorption configuration of reaction intermediates is also shown. The navy, red, and white balls represent Pd, O, and H atoms, respectively. Adsorption energy for the D-glucose oxidation on the (c) D-Au surface and (d) L-Au surface as a function of the hole level (σ). The adsorption configuration of reaction intermediates is also shown. The yellow, red, gray and white balls represent Au, O, C and H atoms, respectively. Isothermal Titration Calorimetry (ITC) data of the D-glu interaction with (e), (f) D-Au (K_d= 28.2 μM, N = 1 (Fix)) and (g), (h) L-Au (K_d= 60.9 μM, N = 1 (Fix)). K_d represents dissociation constant, and N represents binding stoichiometry. The experimental data are shown as solid squares, and the least squares’ best-fit curves derived from a simple one-site binding model (fixed) are shown as red lines. Insets are the side view of D-glu on Au (321) surface. The yellow, red, gray and white balls represent Au, O, C and H atoms, respectively. Source data are provided as a Source Data file.

Light-dependent GOD- and POD-like enzymatic performance

Inspired by the mechanism study of our system, enzymatic performance under the CPL was then investigated, as illustrated in Fig. 3. To conduct the CPL-dependent GOD- and POD-like enzymatic performance, we manually set up the light source by integrating a linear polarizer and quarter wave plate with an 808 nm laser, as shown in Supplementary Fig. S7. Under the RC illumination, D-glu selectively interacted with D-Au, which was oxidized to generate gluconic acid by injecting hot holes. In contrast, the hot electrons were injected into O_2, leading to the generation of H₂O₂. Subsequently, L-Au generated the hot electrons under LC illumination, which were feasibly concentrated at the Pd active site to facilitate the effective decomposition of H₂O₂ to produce the cytotoxic ∙OH, while the remaining hot holes at the Au site participate in the oxidation of H₂O to generate further the cytotoxic ∙OH (Fig. 3a, b and Supplementary Fig. S8). Before evaluating catalytic performance, we monitored the stability of nanoparticles in the acetate buffer (pH 4.5) by measuring DLS, zeta-potential, and SEM images (Supplementary Fig. S9). These results supported that the nanoparticles were well dispersed in the acidic condition without forming aggregation to validate the POD-like activities of L-AuPd further. To assess the chiroptical-dependent POD-like activity of L-AuPd, the colorimetric assay was performed using 3,3’,5,5’-tetramethylbenzidine (TMB) as substrate, and the absorbance at 650 nm was then monitored^44,45. Colorless TMB was oxidized by ∙OH to form a blue-colored diamine/diimine charge-transfer complex (ox-TMB), which appeared absorbance at 650 nm. L-AuPd exhibited POD-like activity even without light; however, its catalytic activity improved under light illumination, as confirmed by higher absorbance compared to both neat L-AuPd and L-AuPd + H₂O₂ (Fig. 3c and Supplementary Fig. S10a).

a Schematic illustration of D-Au and L-AuPd cascade enzymatic reaction under RC and LC. b Energy level diagram illustrating the hot charge carrier mediated-catalytic reaction and their reaction pathway. E_vacuum represents vacuum energy scale and E_NHE represents electrochemical energy scale relative to a normal hydrogen electrode. (Fermi level of Au and Pd is 5.3 eV and 5.6 eV, redox potential of (O₂/H₂O₂) = 0.68 V and (H₂O₂/H₂O) = 1.77 V)^36,51. c UV-vis absorption spectra of ox-TMB upon adding L-AuPd and H₂O₂ in acetate buffer (pH 4.5) under RC, LC, and L irradiation, demonstrating CPL-dependent catalytic activity with the highest intensity under LC. DW (deionized water) was used for the control group. d Photoluminescence spectra of resorufin upon adding D-Au and glucose under RC, LC, and L irradiation, showing CPL-dependent catalytic activity with the highest intensity under RC. DW (deionized water) was used for the control group. e Photoluminescence spectra of terephthalic acid (TA) solution under different NPs and light irradiation conditions. f Electron paramagnetic resonance (EPR) spectra of DMPO/∙OH adducts generated under different NPs and light irradiation conditions. Source data are provided as a Source Data file.

Interestingly, L-AuPd under the LC exhibited the most significant absorption peak compared to the linear light (L) and RC, indicating the chiroptical-dependent POD-like activities (Fig. 3c). To evaluate the underlying reaction mechanism of L-AuPd, time-dependent ox-TMB absorption change profiles were monitored with various H₂O₂ concentrations (0–80 mM) under LC, L, and RC irradiation. From these profiles, the Michaelis−Menten kinetic model and Lineweaver-Burk model of L-AuPd were obtained (Supplementary Fig. S11 and Supplementary Table S1). Based on the Lineweaver-Burk model, the Michaelis-Menten constant (K_m) of L-AuPd was calculated to be 60.083 mM, 67.557 mM, and 51.412 mM for RC, L, and LC irradiation, respectively⁴⁶. Meanwhile, the maximum reaction velocity (V_max) values of L-AuPd under RC, L, and LC illumination were obtained at 5.865, 7.775, and 10.539 × 10^–8 M s⁻¹_, respectively, for H₂O₂ (Supplementary Table S1). The K_m value for the H₂O₂ substrate under LC irradiation was significantly reduced compared to L or RC, suggesting that the substrate’s binding affinity was increased in a chiroptical-dependent manner. The V_max value of L-AuPd under LC irradiation was the highest compared to RC or L. These results demonstrated that the fast decomposition of H₂O₂ occurred at the L-AuPd active site under LC irradiation, further emphasizing the chiroptical-dependent enhancement of POD-like activity.

Next, the chiroptical-dependent GOD-like activity of D-Au was verified. During the glucose oxidation reaction, O₂ obtains 2 electrons from the Au catalyst to produce the H₂O₂, while the glucose is oxidized to gluconic acid to form an acidic environment that facilitates the POD-like activity of L-AuPd. Firstly, the production of H₂O₂ was confirmed by colorimetric Amplex red assay under light illumination. Amplex red is a representative indicator for validating GOD-like activity, converted to red fluorescent resorufin (peaks at 584 nm) in the presence of H₂O₂ and horseradish peroxidase (HRP)⁴⁷. To optimize the substrate concentration, the fluorescence intensity of Amplex red was monitored with various concentrations of D-glu (0.01–1 M). As the D-glu concentration increased, the fluorescence intensity gradually increased (Supplementary Fig. S12). As shown in Fig. 3d and Supplementary Fig. S10b, the D-Au + D-glu case exhibited a slight fluorescence peak at 580 nm compared to the control group (deionized water, D-Au, and D-glu solution), suggesting a moderate GOD-mimic activity of D-Au. Meanwhile, improved fluorescence intensity was observed in the cases of D-Au + D-glu under light illumination. Importantly, the most significant fluorescence intensity was monitored in the case of D-Au + D-glu + RC (red line) compared to the L (green line) or LC (blue line), demonstrating the effective generation of H₂O₂ from D-glu at D-Au active site in a chiroptical-dependent manner (Fig. 3d). To further verify whether substrate selectivity varies based on chirality, we conducted the Amplex red assay using both matched and non-matched handedness of Au catalyst and glucose, respectively (Supplementary Fig. S13a–e). The cases of D-Au + D-glu (Supplementary Fig. S13a) and L-Au + L-glu (Supplementary Fig. S13d) showed significantly enhanced fluorescence signals compared to the control group (deionized water), while negligible fluorescence signals were observed in the cases of D-Au + L-glu (Supplementary Fig. S13c) and L-Au + D-glu (Supplementary Fig. S13b) compared to the control group, which were consistent with DFT calculations and ICT measurements. These findings further emphasized that integrating chirality with enzyme-like nanoparticles could significantly improve substrate selectivity. Also, light irradiation resulted in a further increase in fluorescence intensity compared to control groups in all cases. This suggests that the plasmon-mediated hot charge carriers promote the overall glucose oxidation reaction for generating H₂O₂ (Supplementary Fig. S13a–d). Then, gluconic acid production was monitored by measuring the pH under light irradiation in a time-dependent manner, as shown in (Supplementary Fig. S14). The pH of solutions gradually decreased with increasing reaction time in all cases, indicating the effective generation of gluconic acid by D-Au and D-glu under light irradiation. Notably, the decrease in pH was the greatest under the RC illumination, followed by moderate decrease in L, and the least significant decrease in LC. These results originated from the chiroptical-dependent GOD-like performance, which is consistent with previous demonstrations.

Based on the catalytic performance of both D-Au and L-AuPd, we then assessed the cascade enzymatic properties of our enzyme-like chiral plasmonic nanoparticle system under the RC and LC illumination (Fig. 3e, f and Supplementary Fig. S10c). Terephthalic acid (TA) was used to determine the efficiency of ∙OH generation from D-glu through a series of enzymatic reactions. Non-fluorescent TA reacts with ∙OH, inducing its transformation into fluorescent hydroxyl-TA with 425 nm emission when excited at 315 nm⁴⁸. Without D-Au and L-AuPd catalysts, negligible peaks at 425 nm were observed (Fig. 3e, dark gray line), indicating that the TA solution remains stable under light irradiation. No peak was observed without D-glu, implying that the cascade enzymatic activity occurred only in the presence of D-glu substrate. (Supplementary Fig. S10c, light gray). D-Au + L-AuPd without light illumination showed a moderate FL peak compared to the illumination conditions, suggesting light illumination effectively enhanced the catalytic performance for both D-Au and L-AuPd (Supplementary Fig. S10c, green line). Also, negligible fluorescence peaks were detected in D-Au (Fig. 3e, f, purple line) and L-AuPd (Fig. 3e, f, blue line) in the presence of D-glu under RC + LC illuminations, suggesting that the presence of both catalysts was mandatory for generating ∙OH from D-glu. D-glu incubated with both D-Au and L-AuPd under sole RC or LC irradiation showcased the moderate fluorescence signal compared to under both RC + LC illumination, indicating the ∙OH generated effectively through the sequentially activated enzymatic reaction. Importantly, D-Au + L-AuPd under RC + LC illumination showcased the 1.3-fold enhanced fluorescence signal compared to the D-Au + L-AuPd under RC or LC illumination, demonstrating the catalytic efficiency could be improved by manipulating the CPL. These results further supported the chiroptical-dependent properties of our enzyme-like chiral plasmonic nanoparticles, facilitating the ∙OH generation using D-glu as a substrate molecule for the cascade enzymatic reaction. Electron paramagnetic resonance (EPR) measurements were then carried out to corroborate further the effective generation of ∙OH in various control groups, as shown in Fig. 3f. 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) molecules were used as spin-trapping agents⁴⁹. After spin trapping of ∙OH using DMPO, the characteristic EPR signals of DMPO/∙OH adduct appeared with 1:2:2:1 of intensity ratio. D-Au + L-AuPd + RC or D-Au + L-AuPd + LC presented characteristic EPR peaks of DMPO/∙OH adducts with moderate intensity, while D-Au + L-AuPd + RC + LC exhibited the most intense EPR signals. These results further emphasized the effective generation of ∙OH in the presence of our enzyme-like chiral plasmonic nanoparticles under both LC and RC illumination.

The above series of experiments persistently verify the excellent ROS-generating properties of our enzyme-like chiral plasmonic nanoparticles through the sequential chiroptical-dependent generation of hot charge carriers. This facilitates the oxidation of D-glu with strong substrate selectivity to produce H₂O₂, followed by the reduction of H₂O₂ at the L-AuPd active site to generate ∙OH.

In vitro cascade enzymatic activities of enzyme-like chiral plasmonic nanoparticles by CPL

Based on the effectiveness of our enzyme-like chiral plasmonic nanoparticles, we then evaluated the therapeutic properties of our system in vitro. To enhance the biocompatibility and dispersity of our nanoparticles, we introduced polyethylene glycol conjugated with the thiol group (PEG-SH) via Au-thiol and Pd-thiol interactions. No morphological changes were observed after conjugating with PEG-SH to both D-Au and L-AuPd, as depicted in Supplementary Fig. S15. The average size distribution of all samples (D-Au, D-Au@PEG, L-Au, L-AuPd, and L-AuPd@PEG) was measured by dynamic light scattering (DLS). It was approximately 100 nm, suggesting their characteristics were preserved after introducing PEG (Supplementary Fig. S16). We then carried out flourier transform infrared spectroscopy (FT-IR) measurements to further confirm the conjugation of PEG (Supplementary Fig. S17). FT-IR spectra of pure PEG and L-Au@PEG showed the characteristic peak at 1100 cm⁻¹, corresponding to C-O stretching in PEG, demonstrating the successful incorporation of PEG-SH into the enzyme-like chiral plasmonic nanoparticles. We then monitor the catalytic effect of D-Au@PEG and L-AuPd@PEG by conducting TMB and Amplex red assay. As shown in Supplementary Fig. S18, the overall catalytic activity of both D-Au and L-AuPd after PEGylation was well preserved, suggesting surface modification did not reduce the overall therapeutic efficacy of our nanoparticles both in vitro and in vivo. In vitro therapeutic properties mediated by cascade GOD- and POD-like activities of our enzyme-like chiral plasmonic nanoparticles were evaluated with colon cancer cells (CT26 and HT29) (Fig. 4a and Supplementary Fig. S19a). Firstly, the dark cytotoxicity of D-Au and L-AuPd was assessed by water-soluble tetrazolium salts (WST-8) assays for various concentrations (optical density at 808 nm: 0–1) (Fig. 4b and Supplementary Fig. S19b). CT26 cells and HT29 cells were incubated with D-Au and L-AuPd for 24 h, respectively, following the absorbance at 450 nm was measured after incubated with WST-8 solution for 30 min. With increasing the concentration of both D-Au and L-AuPd, cell viability gradually decreased but remained insufficient due to the weak generation of ∙OH without light illumination. To optimize the light irradiation time, we conducted the time-dependent cell viability test under various irradiation times for CT26 and HT29 cells (Supplementary Fig. S20). RC and LC were irradiated sequentially for equal duration. For both CT26 and HT29, cell viabilities were above ~90% with 5 min of light illumination, suggesting the phototoxicity was negligible. Thus, we proceeded with subsequent experiments under 5 min of light illumination. To assess the light-induced cytotoxicity and demonstrate the CPL-dependent catalytic therapeutic effects in our system, we conducted WST-8 assays under both LC and RC illumination. The samples (OD at 808 nm = 0.1 of D-Au and L-AuPd with ~90% cell viability for CT26 and OD at 808 nm = 0.25 of D-Au and L-AuPd with ~90% cell viability for HT29) were incubated with CT26 and HT29 cells for 24 h under various irradiation conditions (ctrl, L, RC, LC, and RC + LC) as depicted in Fig. 4d and Supplementary Fig. S19d. No phototoxicity was observed without samples under light illumination of L, RC, and LC (Fig. 4c and Supplementary Fig. S19c, light yellow region). However, cells treated with samples exposed to L showed approximately 70% and 80% cell viability for CT26 and HT29, respectively, whereas the most significant cell cytotoxicity (approximately 40% and 60% cell viability for CT26 and HT29, respectively) was observed under both RC + LC illumination (Fig. 4c and Supplementary Fig. S19c, light blue region). Irradiation with RC or LC alone resulted in moderate cell cytotoxicity, falling between the levels observed when only L and RC + LC were irradiated. In the presence of either D-Au or L-AuPd under both RC and LC illumination, the cell survival rates were approximately 70 % and 80 %, respectively. This indicates that when only one of the GOD- or POD-like activities is activated, lower catalytic therapeutic effects are observed compared to when both activities are sequentially activated (Supplementary Fig. S21). These findings demonstrated the chiro-optical-dependent improvement of enzymatic activity in each enzyme-like nanoparticle, leading to an unprecedented decrease in cell viability. Live dead cell assays were conducted using Calcein-AM/PI, further coinciding with the above WST-8 assay results, as shown in Fig. 4e and Supplementary Fig. S19e. To unveil the in vitro ROS generation, 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) probe was employed to determine the intracellular ROS levels (Fig. 4f and Supplementary Fig. S19f). The non-fluorescent DCFH-DA reacted with intracellular ROS within living cells, forming the green-fluorescent 2′,7′-dichlorodihydrofluorescein (DCF). Notably, superior fluorescence intensity was disclosed under RC + LC illumination against the other control groups examined (Fig. 4f and Supplementary Fig. S22). Most importantly, these findings emphasized the intended efficiency of our enzyme-like chiral plasmonic nanoparticles and the proposed amplified cascade enzymatic activity, which facilitates catalytic therapeutic effects within the cellular environment. Cell death pathway was then revealed using flow cytometric analysis (Fig. 4g and Supplementary Fig. S19g). Representatively, the highest occurrence of late apoptosis was observed when cells were incubated with D-Au + L-AuPd under RC + LC illumination (73.3% for CT26 and 53.25 % for HT29, respectively) among the evaluated irradiation conditions. In contrast, L, RC, and LC illumination revealed ratios of only 50.9, 65.5, and 64.7% for CT26 and 20.80, 39.72, and 51.88 % for HT29, respectively. To further support the apoptotic cell death, we monitored the mRNA expression level of caspase-3 (Casp-3) through reverse transcription polymerase chain reaction (RT-PCR) (Fig. 4h, i and Supplementary Fig. S19h, i). Agarose gel electrophoresis of RT-PCR-amplified mRNA products and their quantitative analysis revealed the highest Casp-3 expression level in cells treated with D-Au + L-Au Pd under RC + LC illumination. This expression was notably higher when compared to the control groups treated with D-Au + L-AuPd under L, RC, and LC illumination, respectively. These findings strongly support the induction of apoptotic cell death, demonstrating that our system significantly enhances Cas-3 activation, further validating its role in promoting apoptosis in treated cells. We then monitored the cellular GOD-like activity by measuring the glucose level to evaluate the GOD-like activity of D-Au (Fig. 4j, Supplementary Figs. S19j and S23). Both CT26 and HT29 cells treated with D-Au under the RC illumination showed the most significant reduction in cellular glucose levels, while the cells without light illumination showed a moderate reduction, demonstrating the effective GOD-like activity of D-Au under RC irradiation. In summary, the aforementioned results solidify the superior therapeutic effectiveness achieved through the apoptosis pathway.

a Schematic diagram of cascade enzymatic reaction of D-Au and L-AuPd under RC and LC in vitro (created in BioRender, https://Biorender.com/q04a380). b Dark cytotoxicity of CT26 cells with increasing concentrations (optical density at 808 nm) of D-Au and L-AuPd (n  =  3 independent experiments). c Cell viability of CT26 exposed to CPL without samples (yellow region) and with samples (L-AuPd and D-Au) under different CPL irradiation (blue region) (n  =  3 independent experiments). Statistical analysis was conducted by one-way ANOVA test (***p < 0.001, ****p < 0.0001). d Schematic illustration of different light irradiation conditions for cascade reaction. The control group was not irradiated. e Fluorescence image of Calcein-AM (green)- and PI (red)-contained CT26 cells treated with L-AuPd and D-Au under different CPL irradiation. f Fluorescence image of ROS with ∙OH probe (DCFH-DA) in CT26 cells treated with L-AuPd and D-Au under different CPL irradiation. g Flow cytometry analysis of CT26 cells treated with L-AuPd and D-Au with different CPL irradiation. h Agarose gel electrophoresis of RT-PCR amplified Casp-3 mRNA from CT26 cells treated with L-AuPd and D-Au under various CPL conditions. β-actin was used as an internal control. (n  =  2 independent experiments). i Quantified fluorescent intensity of Casp-3 mRNA staining in gels for each condition. The intensities were normalized to β-actin mRNA expression (n  =  2 independent experiments). Statistical analysis was conducted by one-way ANOVA test (**p < 0.01, ***p < 0.001). j Relative glucose concentrations in CT26 cells under different conditions (n  =  3 independent experiments). Absorbance intensities were normalized to control (untreated and unirradiated). Dark groups were treated with D-Au without irradiation, while RC groups were treated with D-Au and irradiated. Source data are provided as a Source Data file.

In vivo antitumor performance

Inspired by the efficient in vitro anticancer effects, we further assessed the in vivo chiroptical-dependent catalytic cancer treatment using our enzyme-like chiral plasmonic nanoparticles. Samples (OD = 1 at 808 nm for D-Au and L-AuPd) were injected intratumorally into CT26 and HT29 tumor-bearing balb/c nude mice. To optimize the duration of light exposure to mice, we monitored the infrared thermography of samples under various exposure times (Supplementary Fig. S24). Photothermal effects of our nanoparticles under the LC, RC, and L were negligible until irradiation reached 10 min. Based on the above thermography results, tumor-bearing mice were exposed to the RC followed by LC for 5 min, respectively (total 10 min irradiation), for subsequent in vivo experiments. The tumor suppression ability was then monitored for 18 days, after which the CT26 and HT29 tumor-bearing mice were sacrificed to measure the corresponding tumor weight and perform histological analysis (Fig. 5a and Supplementary Fig. S26a). Significant tumor suppression was evident exclusively in the light-exposed groups. In contrast, no alterations in tumor growth were exhibited in the group without light (Supplementary Fig. S25a–c). These findings were consistent with the in vitro cytotoxicity assessments (Fig. 4c and Supplementary Fig. S19c). In detail, some inhibition of tumor growth was displayed in L, RC, or LC-treated groups, inferring a moderate ROS generation effect to eradicate the tumor (Fig. 5b and Supplementary Fig. S26b). However, the RC + LC-treated group exhibited the complete elimination of tumor. Promoted generation of ROS could likely induce highly efficient catalytic cancer treatment via chiroptical-activated generation of hot charge carriers. Most importantly, no recurrence was observed in the RC + LC treated group, underscoring the exceptional therapeutic efficacy. The corresponding tumor weight and photographs further supported the aforementioned results (Fig. 5d, Supplementary Figs. S26d,  S27 and 28). Throughout the monitoring period, no changes in body weight were observed, reaffirming the in vivo safety of the PEG-coated chiral plasmonic structures (Fig. 5c, Supplementary Figs. S25d and S26c). We then examined hematoxylin and eosin (H&E)-stained tumor tissue sections to assess the tumor-treating efficacy of our nanoparticles under various light treatment conditions. The RC + LC-treated group exhibited the most severe tumor tissue damage attributed to the enhanced cascade enzymatic activity of D-Au and L-AuPd for catalytic cancer treatment (Fig. 5e and Supplementary Fig. S26e). This observation contrasts sharply with the moderate tumor damage observed in specimens resected from all other irradiation groups (L, RC and LC) and negligible tumor damage was observed in dark condition (Supplementary Fig. S25e, f).

a Schematic illustration of in vivo experiments with CT26 tumor-bearing mice (created in BioRender, https://Biorender.com/t51f293). b Tumor suppression profiles of the control group (PBS-treated) and sample treated groups under different light irradiation conditions (n  =  4 mice per group). Statistical analysis was conducted by one-way ANOVA test (**p < 0.01, ***p < 0.001). c Body weight change profiles of the control group (PBS-treated) and sample treated groups under different light irradiation conditions (n  =  4 mice per group). d Weight of the dissected tumors of the control group (PBS-treated) and sample treated groups under different light irradiation conditions on day 18. (n  =  4 mice per group). Statistical analysis was conducted by one-way ANOVA test (****p < 0.000). Inset: corresponding tumor photograph. scale bar = 1 cm. Histochemical analyses of (e) H&E and (f) TUNEL of tumor tissues harvested from control group (PBS-treated) and sample treated groups under different light irradiation. g TUNEL positive fluorescence intensity of tumor sections with control group (PBS-treated) and sample treated groups under different light irradiation. (n  =  3 independent experiments). Statistical analysis was conducted by one-way ANOVA test (*p < 0.1,**p < 0.01, ***p < 0.001, ****p < 0.0001). h Blood biochemical analysis of CT26 tumor-bearing mice after different conditions of CPL irradiation in the presence of D-Au and L-AuPd. The Dark group was not irradiated with CPL in the presence of D-Au and L-AuPd. The control group was treated with PBS. Source data are provided as a Source Data file.

Moreover, we conducted terminal deoxynucleotidyl transferase (TdT)-mediated deoxyuridine triphosphate (dUTP) nick end labeling (TUNEL) staining and the quantitative analysis on tumor sections to ascertain the apoptosis level of tumor cells (Fig. 5f, g). As expected, the highest level of TUNEL-positive signals was exhibited from the RC + LC illuminated group, which aligns well with the H&E staining results for CT26 tumor-bearing mice. In contrast, mice treated with D-Au + L-AuPd under dark condition exhibited no significant TUNEL and Ki67 signals, similar to the control group (Supplementary Fig. S25g–i). For the HT29-bearing mice group, we performed the Ki67 staining and the quantitative analysis on tumor sections to confirm the proliferation level of tumor cells (Supplementary Fig. S26f, g). As expected, the lowest level of Ki67-positive signals was exhibited from the RC + LC illuminated group, demonstrating efficient suppression of tumor proliferation in our system. To investigate the biocompatibility of our nanoparticles, we conducted the additional H&E staining of five major organs (heart, lung, kidney, liver, and spleen) for both CT26 and HT29 tumor-bearing mice. When treated with either PBS or our enzyme-like nanoparticles under RC + LC illumination, there were no signs of abnormal morphological disruption and notable inflammation in major organs, suggesting the biosafety of our enzyme-like nanoparticles (Supplementary Figs. S29 and 30) for both CT26 and HT29 tumor-bearing mice.

Furthermore, the results of blood biochemical analysis, which included measurements of glutamic-pyruvate transaminase (GPT), alkaline phosphatase (ALP), total protein (TP), blood urea nitrogen (BUN), and creatine phosphokinase (CPK), all fell within normal ranges. This indicates minimal systemic toxicity and underscores the biosafety of the PEG-coated chiral plasmonic structures (Fig. 5h and Supplementary Fig. S26h). These results further prove the significant potential of our enzyme-like chiral plasmonic nanoparticles as a CPL-activated catalytic nanomedicine.

In summary, we integrated the chiral plasmonic feature with enzyme-mimicking nanoparticles as a promising CPL-dependent enzyme-like nanoparticle system with highly efficient tumor inhibition ability via adjusting cascade GOD- and POD-like activity. Intrinsic handedness of D-Au enantioselectivity interacted with D-glu substrate, likely natural enzymes, exhibiting a 2-fold stronger binding affinity compared to L-Au. Incorporated with plasmonic features, D-Au exhibited significantly enhanced GOD-like activities via LSPR-driven hot charge carriers under the chiral-matched light illumination, i.e., RC. After the irradiation of LC, the POD-like activity of L-AuPd was promoted to decompose the H₂O₂ to ∙OH from the D-glu substrate. The overall catalytic performance was enhanced when GOD and POD reactions were sequentially triggered by irradiation with RC followed by LC, demonstrating that chiroptically modulated cascade reactions could lead to effective therapeutic outcomes. Comprehensive in vitro and in vivo experiments have shown that our enzyme-like chiral plasmonic nanoparticles and the underlying strategy achieved successful tumor suppression in the group irradiated with both LC + RC compared to only LC, RC or L irradiated. We believe that our enzyme-like chiral plasmonic nanoparticles, which depend on CPL and enantioselective substrates, create a unique approach to catalytic therapy via cascade reactions.