Tunable and self-reporting esterase-activated cyanide prodrugs exert cytoprotective effects Authors: Xidan Tong, Wen Peng, Jinkang Feng, Shengping Huang, Jianru Liu, Wei Li, Jiaxuan Chen, Weiwei Guo, Yueqin Zheng Published: 2026 ABSTRACT Hydrogen cyanide (HCN) is increasingly recognized as a mammalian gasotransmitter with biphasic biological effects, but its study is limited by the lack of controllable delivery tools. Here we report esterase-activated, self-reporting cyanide donors with tunable release kinetics and fluorogenic readout. Upon esterase activation, the donors synchronously release cyanide and regenerate an ICT-based fluorophore. Steric modulation at the cyanohydrin α-carbon controls the release rate over one order of magnitude (kobs = 0.04–0.83 min−1). In cell-based assays, biological outcomes correlate with cyanide generation rate rather than nominal donor concentration, producing proliferation at low rates and toxicity at higher rates. Cytoprotection occurs within an experimentally observed range of 0.3–0.7 nmol mg−1 protein h−1 under these conditions. In the MCAO model, the lead donor Et-1 administered at 1 μmol kg−1 (0.3 mg kg−1) reduces infarct volume and attenuates lipid peroxidation and inflammation. Introduction Hydrogen cyanide (HCN) is widely recognized as an exceptionally toxic industrial pollutant, as evidenced by numerous poisoning incidents resulting in large-scale fatalities1–3. Paradoxically, cyanide was abundant in the early Earth’s atmosphere and is indispensable to prebiotic chemistry4–6. Although well-established biosynthetic pathways exist in bacteria, fungi, and plants7–10, cyanide has long been viewed in mammalian systems solely as a metabolic poison acting through inhibition of complex IV (Cytochrome C oxidase, CCOx)11–15. Recent discoveries, however, have overturned this long-standing assumption16–20. Cyanide exhibits a biphasic regulatory effect on CCOx: nanomolar to low micromolar levels stimulate respiration, whereas higher concentrations cause the classical inhibitory phenotype. Mammalian cells can endogenously generate low levels of cyanide through glycine-promoted, lysosome-localized, peroxidase-dependent pathways16. Cyanide also mediates S-cyanylation of cysteine residues, modulating key protein activities. At appropriate rates, cyanide enhances mitochondrial bioenergetics and proliferation; when elevated, it suppresses metabolism, producing a characteristic bell-shaped dose response. Notably, pathological accumulation occurs in disorders such as nonketotic hyperglycinaemia and Down syndrome16,17. These findings collectively support cyanide’s classification as an emerging mammalian gasotransmitter, joining NO, CO, and H2S21–28. Despite these emerging insights into cyanide biology, experimental investigation in this area remains challenging because suitable chemical tools are still relatively limited19. Traditional inorganic cyanide salts such as NaCN and KCN are highly toxic, and release HCN instantaneously, making them difficult to use for controlled biological studies. Naturally occurring cyanogenic glycosides exhibit comparatively lower acute toxicity, but function as indirect prodrugs that require β-glucosidase-mediated cleavage to generate cyanohydrin intermediates29,30. Because β-glucosidase activity varies substantially across mammalian tissues, cyanide release from these systems is often difficult to predict or control. Moreover, the resulting cyanohydrins (e.g., mandelonitrile) generally undergo rapid spontaneous decomposition under physiological conditions, limiting opportunities for sustained or tunable cyanide delivery. Recently, the Pluth and Feringa groups independently reported synthetic cyanide donor systems activated by esterase or light stimuli, representing important advances for the field31,32. These studies elegantly demonstrated the feasibility of controllable intracellular cyanide release. At the same time, questions regarding how cyanide release kinetics influence biological responses, therapeutic windows, and in vivo efficacy remain largely unexplored. In addition, methods capable of simultaneously achieving controlled cyanide delivery together with real-time monitoring of release behavior are still scarce. Given the emerging view of cyanide as a potential mammalian gasotransmitter with highly concentration- and rate-dependent biological effects, the development of donor systems integrating controllable release with self-reporting capability may provide valuable opportunities for mechanistic investigation. Such platforms could facilitate precise correlation of cyanide release kinetics with biological outcomes and help enable systematic evaluation of cyanide’s therapeutic and toxicological effects in complex biological settings. To help address these challenges, we developed a series of esterase-activated cyanide donors with tunable release kinetics. The donor activation mechanism was systematically investigated using multiple orthogonal approaches and validated with appropriate negative controls, while the donors themselves remain stable in the absence of enzymatic triggering. Importantly, the donor design couples cyanide release with synchronous fluorophore generation, enabling a self-reporting process in which cyanide liberation can be monitored in real time through fluorescence output. This strategy allows direct visualization and kinetic analysis of cyanide release under physiological conditions while minimizing interference from nonspecific signals. Using this platform, we found that donors administered at the same concentration but exhibiting different release rates produced markedly distinct biphasic cellular responses, highlighting the importance of release kinetics in cyanide biology. Furthermore, we explored the biological effects of controlled cyanide delivery in models of cerebral ischemia–reperfusion injury and observed significant protective effects against oxidative stress. Notably, the optimized donor Et-1 exhibited potent in vivo efficacy at a dose as low as 0.3 mg kg−1 in the mouse ischemia–reperfusion model. Collectively, these findings further support the biological relevance of regulated cyanide signaling and suggest that controllable, self-reporting cyanide donor systems may provide valuable tools for investigating cyanide biology and exploring its potential therapeutic applications. Results Design and validation of a fluorescent self-reporting cyanide donor H-1 The design of our self-reporting cyanide donor was inspired by the cyanohydrin motif prevalent in plant cyanogenic glycosides (Fig. S1), whose cyanide release is intrinsically coupled to the formation of the corresponding aldehyde29,30. Because this transformation entails a substantial change in electronic structure, we envisioned integrating this process into a fluorogenic system operating through an intramolecular charge transfer (ICT) mechanism33–35. Starting from 2-methoxynaphthalene, we synthesized the ICT-active fluorophore H-Flu via sequential bromination and formylation. The corresponding cyanohydrin intermediate H-CNOH was efficiently obtained via a silyl ether intermediate under TMSCN/ZnI2 conditions, followed by acidic hydrolysis. Given the intrinsic instability of cyanohydrins under neutral or basic conditions and in protic media, donor H-1 was assembled under strictly aprotic conditions through an EDCI/DMAP-mediated condensation, affording a 65% overall yield (see Supplementary Information). In H-1, the ICT pathway is effectively suppressed, resulting in negligible fluorescence. Upon esterase activation, the masked cyanohydrin is liberated and undergoes rapid hydrolysis to generate H-Flu concomitantly with cyanide release (Fig. 1A). This transformation restored the ICT effect, producing a pronounced 397-fold fluorescence increase, whereas no significant enhancement occurred in the absence of esterase (Fig. 1B). We next validated the donor-to-fluorophore conversion by 1H NMR spectroscopy. As shown in Fig. 1C, the disappearance of peaks at δ 2.18 (a) and 6.55 (b), together with the emergence of the aldehydic signal at δ 10.09 (c), provides direct evidence for the formation of H-Flu upon esterase-triggered activation of H-1. LC–MS analysis further confirmed the generation of the expected release product (Fig. S2).Fig. 1Design, structure, and release mechanism validation of the self-reporting cyanide donor H-1.A Proposed mechanism of cyanide release from H-1. Esterase hydrolysis generates an unstable cyanohydrin intermediate, which spontaneously decomposes to release cyanide and the fluorescent reporter H-Flu. ICT: intramolecular charge transfer. B Time-dependent fluorescence turn-on of H-1 in the presence or absence of esterase (porcine liver esterase, PLE). Conditions: 100 µM H-1, 10 mM PBS (pH 7.4), 25 °C, with or without 1 U/mL PLE. C 1H NMR analysis confirming the conversion of H-1 to H-Flu. D Cyanide generation validated using specific fluorescent probes CP and CSP after 45 min incubation. cobalt(II) edetate (CoE): 10 µM; bis(4-nitrophenyl) phosphate (BNPP): 100 µM. E, F Myoglobin (5 μM) spectral assay demonstrating cyanide release. Created with BioRender.com. G Cyanide ion-selective electrode recording showing voltage drop upon cyanide binding to Ag+ sites. Conditions: 100 µM H-1, 10 mM PBS with 2.5% DMSO (pH 7.4), 37 °C, without PLE (0–15 min), with 1 U/mL PLE (15–30 min). Data are presented as mean ± sd (n = 3 independent experiments). Source data are provided as a Source Data file. Design, structure, and release mechanism validation of the self-reporting cyanide donor H-1. A Proposed mechanism of cyanide release from H-1. Esterase hydrolysis generates an unstable cyanohydrin intermediate, which spontaneously decomposes to release cyanide and the fluorescent reporter H-Flu. ICT: intramolecular charge transfer. B Time-dependent fluorescence turn-on of H-1 in the presence or absence of esterase (porcine liver esterase, PLE). Conditions: 100 µM H-1, 10 mM PBS (pH 7.4), 25 °C, with or without 1 U/mL PLE. C 1H NMR analysis confirming the conversion of H-1 to H-Flu. D Cyanide generation validated using specific fluorescent probes CP and CSP after 45 min incubation. cobalt(II) edetate (CoE): 10 µM; bis(4-nitrophenyl) phosphate (BNPP): 100 µM. E, F Myoglobin (5 μM) spectral assay demonstrating cyanide release. Created with BioRender.com. G Cyanide ion-selective electrode recording showing voltage drop upon cyanide binding to Ag+ sites. Conditions: 100 µM H-1, 10 mM PBS with 2.5% DMSO (pH 7.4), 37 °C, without PLE (0–15 min), with 1 U/mL PLE (15–30 min). Data are presented as mean ± sd (n = 3 independent experiments). Source data are provided as a Source Data file. To verify cyanide liberation, we first employed two established cyanide-responsive probes, CP and CSP16,36,37, which selectively detect cyanide even in complex biological matrices (Fig. S3). H-1 induced 7.6-fold and 4.3-fold fluorescence enhancements in CP and CSP, respectively, within 45 min following esterase exposure (Fig. 1D). The dependence of this response on cyanide was validated using either the cyanide scavenger CoE or the esterase inhibitor BNPP (Fig. S4)16,38–40. CoE reduced fluorescence by 3.8-fold and 1.6-fold (CP and CSP), while BNPP suppressed donor activation, lowering the corresponding signals by 3.3-fold and 2.0-fold. These results confirm that the observed fluorescence responses originate from cyanide generation. We further employed myoglobin (Mb) as an orthogonal reporter31,41. Cyanide binding to the ferrous heme center induces a characteristic red shift of the absorption maximum from 409 to 423 nm (Fig. 1E). Esterase-triggered activation of H-1 produced a clear decrease at 409 nm accompanied by growth of the 423 nm band, consistent with cyanide release (Fig. 1F). No shift occurred in the absence of esterase (Fig. S5). Moreover, a structurally analogous but non-releasing control cH-1 (Fig. S6) failed to alter the Abs423/Abs409 ratio, further confirming that the spectral shift is exclusively attributable to cyanide liberated from H-1. Finally, cyanide generation was directly monitored using a cyanide ion-selective electrode32,42. As shown in Fig. 1G, H-1 alone produced a stable potential (~ −25 mV) for the first 15 min. Upon addition of esterase, the potential dropped rapidly and stabilized near −60 mV, providing unambiguous electrochemical confirmation of cyanide release. Collectively, these results establish the molecular activation mechanism of donor H-1 and demonstrate cyanide release using multiple orthogonal analytical approaches. H-1 is stable in the absence of esterase and does not undergo spontaneous cyanide liberation. Importantly, the synchronous formation of H-Flu and release of cyanide forms the basis of a built-in, real-time, false-positive–resistant reporting mechanism, enabling accurate quantification of cyanide release kinetics even in complex environments. The discrepancies observed among traditional detection methods further highlight the unique advantage of this self-reporting design, which allows direct, temporally resolved monitoring of cyanide generation through fluorogenic readout. Structural optimization for modulated release kinetics Having established the activation mechanism of H-1, we next sought to modulate its cyanide-release kinetics to better emulate the slow, endogenous formation of HCN. Drawing on our previous experience with enzyme-responsive gasotransmitter donors, we initially hypothesized that altering the steric bulk of the ester moiety could tune the rate of esterase-mediated hydrolysis43–45. Accordingly, a series of analogues (H-2 to H-6) was synthesized by replacing the acetyl group of H-1 with propionyl, isobutyryl, isovaleryl, benzoyl, and pivaloyl substituents, respectively (Fig. 2A). Their activation kinetics were evaluated by monitoring the fluorescence turn-on of the reporter H-Flu (Fig. 2B, C).Fig. 2Structural optimization and release profiles of cyanide donors.A Chemical structures of cyanide donors H-1 to H-6. B, C Cyanide release kinetics of H-1 to H-6 (100 µM) in the presence of esterase (PLE, 1 U/mL), and corresponding observed rate constants (Kobs) and half-lives. Data are presented as mean ± sd (n = 3 independent experiments). D Structures of rationally optimized donors Me-1, Et-1, and iPr-1. E Release kinetics of Me-1, Et-1, and iPr-1 (100 µM) with PLE (1 U/mL), including Kobs and t1/2. Data are presented as mean ± sd (n = 3 independent experiments). Source data are provided as a Source Data file. Structural optimization and release profiles of cyanide donors. A Chemical structures of cyanide donors H-1 to H-6. B, C Cyanide release kinetics of H-1 to H-6 (100 µM) in the presence of esterase (PLE, 1 U/mL), and corresponding observed rate constants (Kobs) and half-lives. Data are presented as mean ± sd (n = 3 independent experiments). D Structures of rationally optimized donors Me-1, Et-1, and iPr-1. E Release kinetics of Me-1, Et-1, and iPr-1 (100 µM) with PLE (1 U/mL), including Kobs and t1/2. Data are presented as mean ± sd (n = 3 independent experiments). Source data are provided as a Source Data file. Increasing the steric bulk of the ester acyl substituent produced only a limited effect on slowing hydrolysis. Notably, donor H-6, bearing the most sterically hindered pivaloyl group, released cyanide slightly faster than H-1 (Kobs = 0.39 vs. 0.36 min−1). This behavior may reflect the broad substrate tolerance of porcine liver esterase (PLE), whose flexible and hydrophobic active site could potentially accommodate bulky esters and facilitate their hydrolysis through favorable hydrophobic interactions. Given the limitations of R1 modification, we turned to a mechanistically informed strategy aimed at slowing the intrinsic hydrolysis of the transient cyanohydrin intermediate. We reasoned that introducing steric bulk at the α-carbon (R2 position) adjacent to the cyano group would directly attenuate cyanohydrin breakdown and thus prolong cyanide release (Fig. 2D). Based on this rationale, three analogues—Me-1, Et-1, and iPr-1—were prepared with progressively increasing steric encumbrance at R2. Before assessing their release kinetics, we confirmed that the fluorophores generated upon activation (Me-Flu, Et-Flu, and iPr-Flu) exhibit excitation and emission spectra comparable to those of H-Flu (Fig. S7), ensuring that fluorescence intensity provides a quantitatively reliable readout of cyanide release across all donors. Furthermore, we verified that donor activation is strictly esterase dependent and unaffected by common biological nucleophiles or unrelated enzymes (Figs. S8 and S9). Cyanide release kinetics were then quantified by monitoring the time-dependent fluorescence increase of the corresponding fluorophore from each donor, taking advantage of the validated 1:1 conversion between fluorophore formation and HCN release (Fig. 1). As expected, introducing steric bulk at R2 substantially slowed cyanide release, yielding Kobs values of 0.18, 0.10, and 0.04 min−1 for Me-1, Et-1, and iPr-1, respectively (Fig. 2E). Notably, iPr-1 exhibited a nine-fold reduction in release rate relative to H-1, establishing a clear negative correlation between R2 steric size and cyanide-release kinetics. To further evaluate the generality of this observation, we synthesized a set of Me-1 analogues (Me-2 to Me-5) incorporating variations at the R1 ester position. Consistent with the behavior observed in the H-1 to H-6 series, modifications at R1 again failed to alter release rates (Fig. S10). This result confirms that steric modulation at the cyanohydrin α-carbon (R2)—rather than at the ester acyl moiety—is the key determinant of tunable cyanide-release kinetics in this platform. Cellular evaluation of rate-dependent biological responses With a panel of donors exhibiting distinct release kinetics in hand, we next examined their functional consequences in living cells. Cytotoxicity was first assessed for H-1, Me-1, Et-1, and iPr-1 across multiple cell lines (Fig. 3A and S11, S12). Strikingly, donors administered at identical nominal concentrations—but differing in release rates—elicited markedly different effects on cell viability. The fastest-releasing donor, H-1, showed the greatest cytotoxicity, with an IC50 of 11.5 μM. As the release rate decreased, cytotoxicity was progressively attenuated, yielding IC50 values of 35.2 μM (Me-1), 69.9 μM (Et-1), and 74.9 μM (iPr-1) (Fig. S13). These results clearly establish a negative correlation between release rate and cellular tolerance. At lower concentrations, all donors promoted cell proliferation (Fig. 3B). At higher concentrations, however, this proliferative effect was markedly attenuated and even reversed to inhibition, consistent with the excessive cyanide generated under these conditions (Fig. 3C). This proliferative effect was absent when cells were treated with the corresponding fluorophore, confirming that the phenotype originates from controlled cyanide release rather than from donor scaffold chemistry. Notably, H-1, despite possessing the lowest viability margin at higher concentrations, exhibited the strongest proliferative activity, with measurable enhancement observed at concentrations as low as 0.1 μM.Fig. 3Cytotoxicity and proliferation profiles of cyanide donors in PC12 cells.A Concentration-response cytotoxicity of donors H-1, Me-1, Et-1, and iPr-1 measured by CCK-8 assay. IC50 values were derived using GraphPad Prism 8.0. Data are presented as mean ± sd (n = 8, biological replicates). B Concentration-dependent proliferation of donors assessed by EdU assay. Data are presented as mean ± sd (n = 6, biological replicates). C Intracellular cyanide generation rates over 24 h, with the toxic range highlighted in red. Protein levels were normalized using the BCA assay. BF = Background fluorescence. Data represent mean ± sd (n = 6, biological replicates). Statistical significance was assessed using a two-tailed unpaired Student’s t-test. N.S., not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Source data are provided as a Source Data file. Cytotoxicity and proliferation profiles of cyanide donors in PC12 cells. A Concentration-response cytotoxicity of donors H-1, Me-1, Et-1, and iPr-1 measured by CCK-8 assay. IC50 values were derived using GraphPad Prism 8.0. Data are presented as mean ± sd (n = 8, biological replicates). B Concentration-dependent proliferation of donors assessed by EdU assay. Data are presented as mean ± sd (n = 6, biological replicates). C Intracellular cyanide generation rates over 24 h, with the toxic range highlighted in red. Protein levels were normalized using the BCA assay. BF = Background fluorescence. Data represent mean ± sd (n = 6, biological replicates). Statistical significance was assessed using a two-tailed unpaired Student’s t-test. N.S., not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Source data are provided as a Source Data file. Collectively, these findings demonstrate that the biological outcome of cyanide exposure is highly dependent on both concentration and release kinetics, reflecting the classical biphasic behavior of cyanide in mammalian systems. Importantly, the tunable donors developed here display substantially improved safety and controllability relative to conventional cyanide sources such as KCN, enabling precise dissection of rate-dependent cyanide signaling in cellular environments. We next evaluated the lead candidate Et-1 for its ability to release cyanide and support self-reporting fluorescence imaging in live cells. Cells were divided into three treatment groups: Group A, incubated with the cyanide probe CP alone; Group B, co-treated with Et-1 and CP; and Group C, treated with Et-1, CP, and the esterase inhibitor BNPP (Fig. 4A). As shown in Fig. 4B, Et-1 elicited a modest 1.7-fold increase in CP-derived red fluorescence, indicating intracellular cyanide generation. In striking contrast, the built-in self-reporting channel produced an over 89-fold fluorescence enhancement upon donor activation, highlighting its dramatically superior sensitivity relative to external probe–based detection. Esterase inhibition by BNPP (100 μM) markedly reduced both CP and self-reporting fluorescence signals, confirming that intracellular esterase activity is required for Et-1 activation and cyanide liberation (Fig. 4C). Notably, even without a wash step, no detectable background signal was observed, demonstrating that Et-1 exhibits efficient cellular uptake, minimal off-target activation, and clean intracellular fluorogenic turn-on. These results showcase the capability of Et-1 to function as a highly sensitive, self-reporting cyanide donor suitable for real-time imaging in live-cell environments.Fig. 4Live-cell self-reporting imaging and cyanide delivery by donor Et-1.A Schematic of the experimental protocol for intracellular cyanide detection and self-reporting. Created with BioRender.com. B Quantification of fluorescence signals from the external cyanide probe CP (red) and the self-reporting channel (blue). Data represent mean ± sd (n = 8, biological replicates). Statistical significance was assessed using a two-tailed unpaired Student’s t-test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. C Representative confocal microscopy images of cells under each treatment condition. Scale bar: 20 μm. Source data are provided as a Source Data file. Live-cell self-reporting imaging and cyanide delivery by donor Et-1. A Schematic of the experimental protocol for intracellular cyanide detection and self-reporting. Created with BioRender.com. B Quantification of fluorescence signals from the external cyanide probe CP (red) and the self-reporting channel (blue). Data represent mean ± sd (n = 8, biological replicates). Statistical significance was assessed using a two-tailed unpaired Student’s t-test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. C Representative confocal microscopy images of cells under each treatment condition. Scale bar: 20 μm. Source data are provided as a Source Data file. Specific-rate cyanide supplementation exerts cytoprotective effect Low-dose cyanide is increasingly recognized to exert cytoprotective effects analogous to those of established gasotransmitters such as NO, CO, and H2S. To investigate whether cyanide supplied at defined release rates could mitigate oxidative injury, we assessed the protective efficacy of the donor Et-1 in models of chemically induced cardiotoxicity. Doxorubicin (DOX) is a frontline chemotherapeutic with potent anticancer activity, yet its clinical utility is constrained by dose-limiting cardiotoxicity driven by excessive ROS generation during DOX metabolism46–48. Cardiomyocytes are especially vulnerable owing to their comparatively low antioxidant capacity. Mitoxantrone (MIT), though generating less ROS than DOX, similarly induces clinically relevant oxidative injury49–51. Using H9c2 cardiomyocytes, we exposed cells to DOX (5 μM), MIT (5 μM), or H2O2 (100 μM) as representative pro-oxidants and assessed cell viability using the CCK-8 assay (Fig. 5A). To exclude potential interference of cyanide with the tetrazolium-based readout, cell survival was independently validated by propidium iodide (PI) staining. As expected, each pro-oxidant markedly reduced viability and increased PI-positive cell counts. Remarkably, pretreatment with Et-1 conferred concentration-dependent cytoprotection at lower concentrations (0–10 μM), significantly attenuating both viability loss and PI-defined cell death (Fig. 5B, C). Crucially, neither the non-releasing control cEt-1 nor the isolated fluorophore Et-Flu provided any protective benefit, demonstrating that the observed cytoprotection arises specifically from cyanide released at controlled rates rather than from donor scaffold chemistry or photophysical byproducts. At higher concentrations (30–100 μM), however, the protective effect was markedly diminished and even reversed toward toxicity. These findings establish Et-1 as an effective tool for supplementing cyanide at biologically relevant rates and reveal a protective role for controlled cyanide delivery against oxidative stress–induced cardiotoxicity, further reinforcing the biphasic biological effects of cyanide.Fig. 5Cyanide donor Et-1 protects H9c2 cardiomyocytes against pro-oxidant-induced injury.A Schematic of the experimental timeline. Cells were pretreated with Et-1 (0–100 µM), the negative control cEt-1 (10 µM), or the fluorophore Et-Flu (10 µM) for 6 h, followed by exposure to pro-oxidants (DOX, 5 µM; MIT, 5 µM; H2O2, 100 µM) for an additional 6 h. Created with BioRender.com. B Cell viability was assessed by CCK-8 assay. C Cell survival quantification by propidium iodide (PI) staining. Data represent mean ± sd (n = 8, biological replicates). Statistical significance was assessed using a two-tailed unpaired Student’s t-test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Source data are provided as a Source Data file. Cyanide donor Et-1 protects H9c2 cardiomyocytes against pro-oxidant-induced injury. A Schematic of the experimental timeline. Cells were pretreated with Et-1 (0–100 µM), the negative control cEt-1 (10 µM), or the fluorophore Et-Flu (10 µM) for 6 h, followed by exposure to pro-oxidants (DOX, 5 µM; MIT, 5 µM; H2O2, 100 µM) for an additional 6 h. Created with BioRender.com. B Cell viability was assessed by CCK-8 assay. C Cell survival quantification by propidium iodide (PI) staining. Data represent mean ± sd (n = 8, biological replicates). Statistical significance was assessed using a two-tailed unpaired Student’s t-test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Source data are provided as a Source Data file. Neuroprotective effects in an ischemia–reperfusion injury model Given that oxidative stress is a central pathological driver of ischemia–reperfusion injury in ischemic stroke, and motivated by our findings that controlled cyanide supplementation confers cytoprotection, we next evaluated the neuroprotective potential of donors with distinct release kinetics. An oxygen–glucose deprivation (OGD) model was established in NGF-differentiated PC12 cells—a widely used neuronal injury paradigm—to mimic ischemic stroke–associated oxidative damage (Fig. 6A)52–54. Therapeutic efficacy was assessed by measuring cell viability and lactate dehydrogenase (LDH) release, two complementary indicators of neuronal injury. As shown in Fig. 6B, C, all donors alleviated OGD-induced cytotoxicity, though the optimal protective concentrations varied with the cyanide-release rate. The fastest-releasing donor, H-1, exhibited maximal efficacy at 1 μM, whereas the slower donors Me-1, Et-1, and iPr-1 required 3–10 μM for comparable protection. Among these, Et-1 showed the most pronounced effect, increasing cell viability by 12% and reducing LDH leakage by >26% at 10 μM. Interestingly, the slowest-releasing analogue iPr-1 displayed reduced efficacy in this acute injury model—although its kinetic profile may be advantageous in chronic conditions where sustained low-level cyanide signaling is beneficial. The observed protection was verified to originate specifically from cyanide, as it was abolished by the cyanide scavenger CoE (Fig. S14) and was not reproduced by the corresponding released fluorophores (Fig. S15), confirming that neither the donor scaffold nor the fluorophore is responsible for the effect.Fig. 6Tunable cyanide donors rescue OGD-induced neuronal injury in a release-kinetic-dependent manner.A Schematic of the OGD injury model and drug treatment protocol in NGF-differentiated PC12 cells. Cells were exposed to ischemic buffer (20 mM 2-deoxy-D-glucose, 20 mM sodium lactate, 2.5 mM Na2S2O4 in 10 mM PBS) for 30 min, followed by treatment with cyanide donors H-1, Me-1, Et-1, or iPr−1 at varying concentrations for 24 h. Created with BioRender.com. B Cell viability measured by CCK-8 assay. Data are presented as mean ± sd (n = 6, biological replicates). C LDH release assay quantifying cellular injury. Data are presented as mean ± sd (n = 6, biological replicates). D Intracellular cyanide generation rates over 24 h, with the effective cytoprotective range (0.3–0.7 nmol mg−1 protein h−1) highlighted in green. Protein levels were normalized using the BCA assay. BF = Background fluorescence. Data represent mean ± sd (n = 12, biological replicates). Statistical significance was assessed using a two-tailed unpaired Student’s t-test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Source data are provided as a Source Data file. Tunable cyanide donors rescue OGD-induced neuronal injury in a release-kinetic-dependent manner. A Schematic of the OGD injury model and drug treatment protocol in NGF-differentiated PC12 cells. Cells were exposed to ischemic buffer (20 mM 2-deoxy-D-glucose, 20 mM sodium lactate, 2.5 mM Na2S2O4 in 10 mM PBS) for 30 min, followed by treatment with cyanide donors H-1, Me-1, Et-1, or iPr−1 at varying concentrations for 24 h. Created with BioRender.com. B Cell viability measured by CCK-8 assay. Data are presented as mean ± sd (n = 6, biological replicates). C LDH release assay quantifying cellular injury. Data are presented as mean ± sd (n = 6, biological replicates). D Intracellular cyanide generation rates over 24 h, with the effective cytoprotective range (0.3–0.7 nmol mg−1 protein h−1) highlighted in green. Protein levels were normalized using the BCA assay. BF = Background fluorescence. Data represent mean ± sd (n = 12, biological replicates). Statistical significance was assessed using a two-tailed unpaired Student’s t-test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Source data are provided as a Source Data file. Leveraging the built-in self-reporting capability of our donor platform, we quantified cyanide generation in the cell-based system under each treatment condition. As shown in Fig. 6D, donors with different release kinetics produced markedly different levels of cyanide generation even at identical nominal concentrations. We identified an experimentally observed cytoprotective range of 0.3–0.7 nmol mg−1 protein h−1, within which cyanide generation correlated with optimal cytoprotection. Notably, a low concentration of a fast-releasing donor generated cyanide amounts comparable to those produced by higher concentrations of slow-releasing analogues. These findings demonstrate that the cyanide generation rate under the specified cell-based assay conditions, rather than the nominal donor concentration, is a key determinant of cytoprotective efficacy. Together, they support a bell-shaped, rate-dependent cytoprotective profile of cyanide and validate our tunable donor system as a powerful tool for dissecting the kinetics–response relationships that govern cyanide signaling in neuronal injury. It should be noted that diffusion, redistribution, or efflux of HCN/CN− and/or the released fluorophore may affect the measured signal. Accordingly, these data should be interpreted comparatively rather than as absolute in situ intracellular cyanide concentrations. Cyanide donors mitigate ischemic stroke-induced oxidative damage in vivo Encouraged by the pronounced neuroprotection observed in the OGD cellular model, we next evaluated the protective effects of the donors in vivo. We first assessed the pharmacokinetic behavior and biosafety profile of Et-1. Taking advantage of the donors’ built-in self-reporting fluorogenic system, ex vivo fluorescence imaging clearly revealed the donor’s biodistribution across major organs, showing predominant accumulation and metabolism in the liver and kidneys (Fig. S16). Histopathological examination (H&E and Masson staining) of major organs showed no detectable pathological alterations following Et-1 administration (Fig. S17). Furthermore, hematological and serum biochemical parameters remained within normal ranges (Fig. S18), indicating that Et-1 possesses favorable biocompatibility and well-controlled pharmacokinetics in vivo. We then employed the middle cerebral artery occlusion (MCAO) model in C57BL/6 J mice to mimic ischemic stroke and evaluate in vivo neuroprotective efficacy. Mice underwent 1 h 20 min of MCAO, and Et-1 was administered intravenously at only 1 μmol kg−1 (0.3 mg kg−1) during the first 20 min of occlusion. (Fig. 7A)55–60. TTC staining revealed a substantial reduction in infarct volume in Et-1–treated animals compared with the model group (Fig. 7B, C). Consistent with reduced tissue damage, serum LDH levels decreased by >43% relative to the untreated model group (Fig. 7D). To further quantify oxidative stress in vivo, we measured malondialdehyde (MDA)—a key lipid peroxidation marker—and superoxide dismutase (SOD) activity. Et-1 treatment decreased MDA levels by >34% while elevating SOD activity by 23%, demonstrating suppression of lipid peroxidation and enhancement of endogenous antioxidant defenses (Fig. 7E, F). Histological analyses provided additional insight into neuronal protection. H&E staining of MCAO brain sections revealed extensive neuronal necrosis in the untreated model, whereas Et-1 treatment markedly reduced cell death (Fig. 7G). Nissl staining further confirmed pronounced structural preservation in the Et-1 group: neurons in the peri-infarct region retained well-defined morphology and visible Nissl bodies, in contrast to the collapsed and degenerating neurons in the MCAO model (Fig. 7G). These results collectively indicate that controlled cyanide release from Et-1 effectively mitigates ischemia-induced oxidative injury.Fig. 7Cyanide donor Et-1 mitigates cerebral ischemia–reperfusion injury in a mouse MCAO model.A An experimental timeline. Created with BioRender.com. B, C Representative TTC-stained brain sections and quantitative analysis of infarct volume (n = 6). D Serum LDH levels (n = 8). E, F Malondialdehyde (MDA) content (n = 7) and superoxide dismutase (SOD) activity (n = 6) in brain tissue. G Representative H&E- and Nissl-stained brain sections (scale bars are provided in the images). Similar results were obtained from n = 3 biologically independent mice per group. Data are presented as mean ± sd from biological replicates. Statistical significance was assessed using a two-tailed unpaired Student’s t-test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Source data are provided as a Source Data file. Cyanide donor Et-1 mitigates cerebral ischemia–reperfusion injury in a mouse MCAO model. A An experimental timeline. Created with BioRender.com. B, C Representative TTC-stained brain sections and quantitative analysis of infarct volume (n = 6). D Serum LDH levels (n = 8). E, F Malondialdehyde (MDA) content (n = 7) and superoxide dismutase (SOD) activity (n = 6) in brain tissue. G Representative H&E- and Nissl-stained brain sections (scale bars are provided in the images). Similar results were obtained from n = 3 biologically independent mice per group. Data are presented as mean ± sd from biological replicates. Statistical significance was assessed using a two-tailed unpaired Student’s t-test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Source data are provided as a Source Data file. To further elucidate the neuroprotective mechanisms of Et-1, we performed transcriptomic analysis on brain tissues from C57BL/6 J MCAO mice. Differential gene expression analysis revealed that Et-1 treatment resulted in 205 significantly upregulated and 135 downregulated genes relative to the MCAO model (p = 0.05, |log2FC| ≥ 1.5) (Fig. 8A). Gene Set Enrichment Analysis (GSEA) identified pronounced activation of the TNF-α/NF-κB inflammatory pathway in the MCAO group (Fig. S19). In contrast, Et-1 administration significantly downregulated natural killer (NK) cell-mediated cytotoxicity pathway genes compared with the model group. Moreover, Et-1 effectively suppressed the TNF-α/NF-κB signaling axis relative to the non-releasing control cEt-1, and its transcriptomic signature also displayed marked downregulation of NK cytotoxicity genes when compared with the fluorophore control Et-Flu (Fig. 8B). The mild fluctuations observed in the control groups are likely attributable to weak non-specific effects of the donor scaffold or vehicle, whereas Et-1 produced the most pronounced and specific pathway inhibition. Pathway-level key gene analysis further confirmed that Et-1 robustly counteracted ischemia-induced immune activation, outperforming both control groups (Fig. 8C). The observed downregulation of the NK cell cytotoxicity pathway by Et-1 is consistent with the established detrimental role of NK cells in cerebral ischemia–reperfusion injury. NK cells infiltrate the ischemic brain at an early stage and exacerbate neurovascular damage through the release of granzymes, perforin, and pro-inflammatory cytokines, thereby promoting blood–brain barrier disruption and neuronal death61–65. The precise molecular mechanism underlying this regulation may involve direct signal transduction modulation, bioenergetic reprogramming, or S-cyanylation of specific proteins, and represents an important direction for future investigation.Fig. 8Transcriptomic and histochemical analysis of inflammatory marker expression after cerebral ischemia.A Volcano plot of differentially expressed genes between the Et-1 and MCAO model groups. Differential expression analysis was performed using the exactTest function in edgeR v3.36.0. Genes with P < 0.05 and |log2FC| ≥ 1.5 were considered differentially expressed. B Gene set enrichment analysis (GSEA) was performed using the R package clusterProfiler, with genes ranked according to their differential expression. C Heatmap of gene expression in brain tissue after drug treatment. D, E Representative immunohistochemical staining shows expression levels of IL-1β (D) and Iba1 (E) in drug-treated groups following cerebral ischemia. Scale bar: 100 μm. F Representative immunofluorescence staining depicts inflammatory factor expression post-ischemia (DAPI: blue, NeuN: pink, iNOS: green, TNF-α: red). Scale bar: 1000 μm. Similar results were obtained from n = 3 biologically independent mice per group. Source data are provided as a Source Data file. Transcriptomic and histochemical analysis of inflammatory marker expression after cerebral ischemia. A Volcano plot of differentially expressed genes between the Et-1 and MCAO model groups. Differential expression analysis was performed using the exactTest function in edgeR v3.36.0. Genes with P < 0.05 and |log2FC| ≥ 1.5 were considered differentially expressed. B Gene set enrichment analysis (GSEA) was performed using the R package clusterProfiler, with genes ranked according to their differential expression. C Heatmap of gene expression in brain tissue after drug treatment. D, E Representative immunohistochemical staining shows expression levels of IL-1β (D) and Iba1 (E) in drug-treated groups following cerebral ischemia. Scale bar: 100 μm. F Representative immunofluorescence staining depicts inflammatory factor expression post-ischemia (DAPI: blue, NeuN: pink, iNOS: green, TNF-α: red). Scale bar: 1000 μm. Similar results were obtained from n = 3 biologically independent mice per group. Source data are provided as a Source Data file. To validate these transcriptomic findings, immunohistochemical staining was conducted for key inflammatory markers. Et-1 treatment substantially reduced expression of the pro-inflammatory cytokine IL-1β and markedly attenuated immunoreactivity of the microglia/macrophage marker Iba1, in sharp contrast to the elevated levels observed in both cEt-1 and Et-Flu groups (Fig. 8D, E). The concurrent reduction of IL-1β and Iba1 suggests that Et-1 mitigates neuroinflammation by inhibiting microglial overactivation. Immunofluorescence staining for the mature neuronal marker NeuN showed increased neuronal survival in Et-1–treated mice (Fig. 8F, pink channel). Additional staining of TNF-α (red channel) and iNOS (green channel) revealed strong immunofluorescence signals in the cEt-1 and Et-Flu groups, indicative of pronounced inflammatory activation and M1-type oxidative stress. In contrast, Et-1 treatment led to a marked reduction in both TNF-α and iNOS signals, demonstrating significant suppression of neuroinflammation and oxidative injury. Together, these transcriptomic and histological results establish that Et-1 confers neuroprotection by attenuating the cerebral ischemia–induced immune storm, suppressing inflammatory signaling pathways, reducing microglial activation, and preserving neuronal integrity. Discussion In this work, we introduce a generalizable chemical strategy for controlled and self-reporting cyanide delivery, overcoming a major long-standing limitation in cyanide biology. By integrating an esterase-activated trigger with an ICT-based fluorogenic reporter, we developed a library of cyanide donors that couple endogenous-mimetic, rate-tunable release with real-time, interference-free self-reporting. Mechanistic analysis established that steric modulation at the α-carbon of the cyanohydrin intermediate—not at the ester moiety—is the key determinant of release kinetics, enabling precise control over cyanide liberation across nearly an order of magnitude. This platform allowed us to demonstrate that cyanide’s biological activity is dictated not by its nominal concentration but by its intracellular generation rate. Donors with different kinetic profiles produced distinct biphasic cellular responses, consistent with cyanide’s emerging role as a mammalian gasotransmitter. Remarkably, controlled cyanide supplementation conferred potent cytoprotection at extremely low effective cyanide levels, revealing a narrow therapeutic window (0.3–0.7 nmol mg−1 protein h−1) within which cyanide supports cellular resilience and antioxidant defenses. The lead donor Et-1 further demonstrated excellent biocompatibility, well-behaved pharmacokinetics, and neuroprotective effects at the acute 24-h endpoint in the MCAO mouse model. Even at a very low systemic dose, Et-1 significantly reduced cerebral infarct volume, attenuated lipid peroxidation and neuroinflammation, preserved neuronal morphology, and improved molecular markers of neuronal survival in the MCAO mouse model. These results establish that minute, precisely controlled amounts of cyanide can produce marked therapeutic benefit—an observation with important clinical implications given cyanide’s historical reputation as a toxin. Building on pioneering studies that have observed cyanide-mediated cytoprotection in models of ischemic and metabolic injury, our work confirms and extends these observations, most notably by providing the proof-of-concept in vivo validation of the neuroprotective efficacy of controlled cyanide delivery in a cerebral ischemia–reperfusion injury model. However, because the present study assessed only acute histological, biochemical, and molecular outcomes at 24 h after MCAO, the durability of these effects and their impact on long-term neurological recovery remain to be established. It should also be noted that this validation was conducted exclusively in male mice to minimize potential confounding by the neuroprotective effects of estrogen. Therefore, future studies are warranted to further validate these findings using longer-term neurological assessments and in female mice. Collectively, this work provides a powerful chemical toolkit for deconvoluting cyanide signaling with unprecedented temporal precision and establishes the fundamental principle that rate-specific, low-level cyanide delivery is both biologically meaningful and therapeutically beneficial. Beyond deepening our understanding of cyanide as a strong candidate gasotransmitter, these findings open new avenues for the rational design of cyanide-based cytoprotective therapies, particularly for oxidative stress–driven diseases such as ischemic stroke and myocardial injury. Method Chemical synthesis Synthesis protocol and additional characterization can be found in the Supplementary Information. LCMS experimental method LC-MS analysis was performed on a Waters LC-e2695 system coupled with a Waters SQ Detector 2 mass spectrometer. Separation was achieved using a Waters Symmetry® C18 column (2.1 mm × 100 mm, 1.7 μm) maintained at 35 °C. The mobile phase consisted of solvent A (water with 0.1% formic acid) and solvent B (acetonitrile with 0.1% formic acid). A gradient elution program was applied as follows: 5% B initially, increased to 95% B over 12 min, held for 3 min, then returned to initial conditions in 3 min and re-equilibrated for 2 min. The flow rate was 0.5 mL/min, and the total run time was 20 min. The injection volume was 5.0 μL. Mass detection was operated in positive electrospray ionization (ESI+) mode with the following settings: capillary voltage 3.0 kV, cone voltage 30 V, source temperature 150 °C, desolvation temperature 350 °C, cone gas flow 50 L/h, desolvation gas flow 800 L/h, and scan range 100–1000 m/z. Data acquisition and processing were performed using Waters MassLynx V4.2 software. Fluorescence measurements All fluorescence measurements were performed on the Gangdong F-300 spectrofluorometer (Tianjin, China). Unless otherwise specified, both the excitation and emission slit widths were set at 5 nm, and the photomultiplier tube voltage was maintained at 600 V. Donors and control compounds were prepared as 100 mM stock solutions in DMSO, while esterase (PLE) was prepared as a 60 U/mL stock solution in PBS; all stock solutions were stored at room temperature. For kinetic analysis of donor release, 3 μL of donor or control compound stock (final conc: 100 μM) was added to a sealed quartz fluorescence cuvette containing 3 mL of PBS (with 1% DMSO), followed by 50 μL of the esterase stock (final conc: 1 U/mL). The mixture was homogenized immediately, and time-scan measurements were initiated using the fluorometer’s kinetic mode, recording data at 1-min intervals over the specified duration. The excitation and emission wavelengths (Ex/Em) were set to 350/460 nm. All recorded values were blank-corrected against control samples without donor or enzyme. Validation of cyanide generation Fluorescent probe method: Cyanide-specific probes CP and CSP were prepared according to literature procedures36,37 (chemical structures and sensing mechanisms are provided in Fig. S3). For the assay, stock solutions of probes CP and CSP (10 mM in DMSO), donor H-1 (100 mM in DMSO), and esterase (PLE, 60 U/mL) were prepared. In a sealed vial containing 3 mL of PBS (with 1% DMSO), 3 μL of either probe CP or CSP was added (final conc: 10 μM), followed by 3 μL of donor H-1 (final conc: 10 μM). Then, 50 μL of the esterase stock solution (final conc: 1 U/mL) was introduced, and the mixture was incubated at room temperature in the dark. After 45 min, an aliquot of the solution was transferred to a quartz fluorescence cuvette, and fluorescence intensity was immediately recorded using a Gangdong F-300 spectrofluorometer (Tianjin, China). Excitation and emission slit widths were set to 5/5 nm, and the photomultiplier tube voltage was maintained at 600 V. The excitation/emission wavelengths (Ex/Em) were set at 365/610 nm for probe CP and 400/490 nm for probe CSP. Data are presented as mean ± sd from three independent experiments. Myoglobin spectral assay: The assay was adapted from reported literature with minor modifications. Myoglobin (from horse skeletal muscle, CAS: 100684-32-0, molecular weight: 16,700) was obtained from Solarbio Life Sciences (Beijing, China) and prepared as a 10 mM stock solution in PBS daily before use, then stored at 4 °C. Donor H-1 and esterase (PLE) were prepared as 100 mM (in DMSO) and 60 U/mL stock solutions, respectively, with the latter stored at room temperature. In a sealed UV cuvette containing 3 mL of PBS (with 1% DMSO), 1.5 μL of the myoglobin stock (final conc: 5 μM) was added, followed by 3 μL of donor H-1 (final conc: 10 μM). Then, 50 μL of the esterase stock (final conc: 1 U/mL) was introduced and mixed thoroughly. UV-vis spectra were recorded on a Shimadzu UV-1800 spectrophotometer. The absorption spectra from 300 to 550 nm were recorded continuously over 30 min at 1-min intervals. Myoglobin exhibited an absorption maximum at 409 nm, which shifted to 423 nm upon cyanide binding. A progressive redshift in the spectrum was observed during the reaction. The ratio of absorbance at 423 nm to that at 409 nm (Abs423/Abs409) was used to model cyanide release kinetics. Data are presented as mean ± sd from three independent experiments. Cyanide-selective electrode method: Cyanide levels (detected as CN−) were measured using a cyanide-selective electrode (Shanghai Leici, model PCN-1-01 micro-ion electrode, detection range: 26 μg/L–0.26 g/L). Although this ion-selective electrode is typically used for cyanide detection in alkaline solutions, it remains operational at pH 7.4, albeit with reduced sensitivity due to lower anion concentration. An ion meter was used to record the electrode potential in real time with a resolution of 0.1 mV. Prior to formal testing, a standard curve was established using a KCN stock solution (0.1 M) diluted to various concentrations (0.1–100 μM). For the assay, compound H-1 was prepared as a 100 μM solution in PBS containing 2.5% DMSO (pH 7.4) and placed in a sealed vial. The measurement was conducted at 37 °C. The electrode was fully immersed in the sample, and the potential (mV) was recorded every second for 30 min at a resolution of 0.1 mV. Esterase (PLE, final conc: 1 U/mL) was added at the 15 min time point after the start of measurement (without PLE: 0–15 min; with 1 U/mL PLE: 15–30 min). All measured values were subtracted from those of a blank control. Cyanide concentration was calculated based on the KCN standard curve. Data are presented as mean ± sd from ten independent experiments. Cell culture PC12, HeLa, and H9c2 cell lines (American Type Culture Collection, ATCC) were cultured in high-glucose Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin. All cells were maintained at 37 °C in a humidified Thermo Scientific HeracellTM VIOS 160i CR CO2 incubator with 5% CO2. Cell viability assay Cell viability was assessed using the Cell Counting Kit-8 (CCK-8, Beyotime Biotechnology Co., Ltd.) according to the manufacturer’s instructions. PC12, HeLa, and H9c2 cells were seeded in 96-well plates at a density of 4 × 103 cells per well and allowed to adhere for 24 h under standard culture conditions. The cells were then treated with cyanide donors at a series of concentrations (0.1, 0.3, 1, 3, 10, 30, and 100 μM) for 24 h. After treatment, 10 μL of CCK-8 solution was added to each well, and the plates were incubated at 37 °C for 45 min. The absorbance at 450 nm (OD450) was measured using a Molecular Devices SpectraMax i3x microplate reader. Viability was expressed as a percentage relative to untreated control cells. All experiments were performed with at least three biological replicates, and data are presented as mean ± sd. Cell survival assay To exclude potential interference of cyanide with tetrazolium-based CCK-8 readouts, cell survival was independently evaluated using propidium iodide (PI) staining. Following the same treatment protocol as described for the CCK-8 assay, cells seeded in black 96-well plates dedicated to fluorescence measurement were incubated with PI solution (Beyotime Biotechnology Co., Ltd.) at a final concentration of 1 μg/mL for 45 min at 37 °C in the dark. After staining, cells were gently washed with PBS and immediately analyzed by flow cytometry (Agilent Technologies NovoCyte, excitation/emission: 535/617 nm). PI-positive cells (indicating dead or membrane-compromised cells) were quantified and expressed as a percentage of the total cell population. All experiments were performed with eight biological replicates, and data are presented as mean ± sd. Confocal imaging HeLa cells were seeded on 15 mm glass-bottom confocal dishes at a density of 2.5 × 104 cells per dish and allowed to adhere overnight. Before imaging, adherent cells were gently washed three times with PBS and incubated with serum-free DMEM containing cyanide probe CP (5 μM). Cells were divided into two groups: Group C was co-treated with the esterase inhibitor BNPP (100 μM), while Group B received no inhibitor. Following 30 min incubation at 37 °C, the medium was carefully removed, and the cells were washed three times with PBS. Donor Et-1 (10 μM in PBS) was then added, and cells were further incubated for 60 min at 37 °C. Without additional washing, live-cell imaging was immediately conducted using an Olympus FV-3000 confocal laser scanning microscope (CLSM). Images were acquired as single optical sections at a resolution of 1024 × 1024 pixels. The blue channel was configured with 405 nm excitation and 440–470 nm emission collection (HV: 530 V, Gain: 1.000×, Offset: 1%). The red channel used 405 nm excitation and 600–630 nm emission collection (HV: 650 V, Gain: 1.000×, Offset: 2%). Additional PMT settings included: Mode: VBF, Average: Line, Sequential Scan: Line, Confocal Aperture: Auto (151 μm), and Airy Disk: 1.00×. Quantitative analysis of cellular images was performed using ImageJ software. LDH release assay The LDH Release Assay Kit (WST-8 method; Beyotime Biotechnology, Cat. #C0019M) was used to quantify LDH activity released from damaged cells according to the manufacturer’s instructions. This colorimetric assay enables sensitive detection of LDH release as an indicator of membrane integrity loss. Prior to formal testing, a standard curve was established using purified LDH. For sample preparation, cells were collected and centrifuged at 10,000 g for 10 min at 4 °C. The pellet was resuspended in lysis buffer (1 mL per 5 × 106 cells) and sonicated on ice (200 W, 3 s pulse, 10 s interval, 30 cycles). The lysate was then centrifuged at 10,000 × g for 10 min at 4 °C, and the supernatant was collected and kept on ice. Subsequent steps were performed strictly following the kit protocol. Absorbance was measured at 450 nm using a ‌‌Molecular Devices SpectraMax i3x microplate reader, and all readings were blank-corrected. All experiments were performed with six biological replicates, and data are presented as mean ± sd. Cell proliferation Cell proliferation was assessed using the BeyoClick™ EdU Cell Proliferation Kit with TMB (Beyotime Biotechnology, Cat. #C0085S) following the manufacturer’s protocol. This assay utilizes the incorporation of 5-ethynyl-2’-deoxyuridine (EdU), a thymidine analog, into newly synthesized DNA during cell proliferation. The incorporated EdU is then detected through a click reaction with biotin labeling, followed by incubation with horseradish peroxidase-conjugated streptavidin (HRP-Streptavidin) and subsequent color development. For quantitative analysis, the blue product generated by the TMB reaction was measured spectrophotometrically at either 370 nm or 620-650 nm. Alternatively, following reaction termination with 2 M H2SO4, absorbance was measured at 450 nm after the solution turned yellow. All experiments were performed with six biological replicates, and data are presented as mean ± sd. Determination of cyanide generation in cells To validate the 1:1 stoichiometry of fluorophore formation to cyanide release upon donor activation, a cell-free conversion assay was first performed. Each donor compound was incubated with excess porcine liver esterase (PLE, 10 U/mL) in PBS (pH 7.4) at 37 °C, and fluorescence intensity at Ex/Em = 350/460 nm was monitored until a stable plateau was reached. The plateau fluorescence values were compared against a calibration curve constructed from authentic X-Flu standards of known concentrations. This confirmed a near-complete conversion efficiency (>95%) for all donors, establishing the reliability of the fluorogenic readout for indirect cyanide quantification under the assay conditions (Fig. S20). For cellular experiments, PC12 cells were seeded at a density of 1.2 × 104 cells per well. To correct for potential matrix effects on fluorescence intensity in a cellular environment, a cell-based calibration curve was established (Fig. S21). Specifically, cells were treated with a concentration gradient of X-Flu (the fluorophore corresponding to each donor) in serum-free medium for 24 h, after which fluorescence intensity was measured at Ex/Em = 350/460 nm using a Molecular Devices SpectraMax i3x microplate reader. To confirm that the observed fluorescence signal in cells arises exclusively from intracellular esterase-mediated activation, a non-releasing control probe (cEt-1) was included in parallel experiments. Cells treated with cEt-1 exhibited only background fluorescence levels, confirming the absence of non-specific hydrolysis or off-target activation. In formal experiments, cells were treated with a concentration gradient of each donor in serum-free medium for 24 h. Total cyanide release was quantified based on fluorescence intensity, where each fluorophore molecule faithfully reports the concurrent release of one cyanide molecule according to the intrinsic 1:1 stoichiometry. It should be noted that this method quantifies the apparent total cyanide generation, representing the cumulative amount liberated over the incubation period, and does not directly report the instantaneous free cyanide concentration, which may be influenced by volatilization and cellular metabolism. Total protein content was determined using a BCA protein assay kit (Beyotime Biotechnology, Cat. #P0009). The average cyanide generation rate was calculated as total cyanide amount divided by incubation time and expressed as an averaged value over the 24 h period, normalized to protein content, and expressed in units of nmol mg−1 protein h−1. The experiment was independently repeated twelve times, and data are presented as mean ± sd. Ethics statement All animal procedures were reviewed and approved by the Ethics Committee of China Pharmaceutical University (approval no. YSL-202606043). We have complied with all relevant ethical regulations for animal use. Sex was considered in the study design. Only male mice were used to minimize potential confounding from the neuroprotective effects of estrogen; therefore, sex-based comparisons were not performed in this study. Animal experiments All animal experiments were conducted using male C57BL/6 J mice (Mus musculus; C57BL/6 J inbred background; 6–8 weeks old, 20–25 g, specific pathogen-free grade) obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China; Production License No. SCXK (Jing) 2021-0011). All animals were housed under specific pathogen-free conditions with a 12-h light/12-h dark cycle, an ambient temperature of 18–23 °C, and a relative humidity of 50 ± 10%, with ad libitum access to food and water. Unless otherwise stated, these animal characteristics and housing conditions applied to all animal experiments described below. Cerebral infarction experiments Male C57BL/6 J mice (6–8 weeks old, 20–25 g, specific pathogen-free grade) were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China; Production License No.: SCXK (Jing) 2021-0011) and randomly divided into the following groups (n = 6 per group): sham-operated group, MCAO model (vehicle, VEH) group, Et-1 treatment group (1 μmol kg−1), cEt-1 negative control group (1 μmol kg−1), and Et-Flu control group (1 μmol kg−1). The sham group underwent the same surgical procedure without filament insertion, while the remaining groups were subjected to MCAO. The dose of Et-1 (1 μmol kg−1) was determined through a two-stage preliminary screening. First, Et-1 was administered to healthy male C57BL/6 J mice (6–8 weeks old, 20–25 g; n = 12 per dose group) via tail vein injection at graded doses of 1, 3, 10, 30, and 100 μmol kg−1, and the 24-h survival rate was evaluated to assess acute tolerability under the tested conditions (Fig. S22). Doses of 10 μmol kg−1 or lower were tolerated during the 24-h observation period, whereas mortality occurred at higher doses. Based on these acute tolerability results, a second round of screening was conducted in male C57BL/6 J mice (6–8 weeks old, 20–25 g; at least 6 mice per dose group) using doses of 10, 5, 2.5, 1, 0.5, 0.25, and 0.1 μmol kg−1 in the MCAO model (Fig. S23). LDH levels in the infarcted hemisphere were used as a preliminary indicator of tissue injury. Based on the LDH-based dose-screening results, 1 μmol kg−1 was selected for subsequent experiments. Animals were anesthetized by inhalation of 3% isoflurane for induction and maintained under 0.6% isoflurane. After sterilization, a midline cervical incision was made, and the left common carotid artery (CCA), internal carotid artery (ICA), and external carotid artery (ECA) were carefully exposed by blunt dissection. The proximal CCA and distal ECA were ligated, followed by ligation of the distal CCA. A monofilament (Beijing Xinong Biological Technology Co., Ltd.) was inserted through the ICA and advanced to occlude the middle cerebral artery (MCA). Cyanide donor Et-1 or control compounds were administered via tail vein injection during the first 20 min of occlusion. After 1 h 20 min of occlusion, the filament was withdrawn to allow reperfusion. Twenty-four hours post-surgery, blood samples were collected, and mice were euthanized by cervical dislocation. Brain tissues were rapidly harvested for subsequent analysis. TTC staining Mouse brains were frozen at −20 °C for 20 min and sectioned into five 1-mm-thick coronal sections. Each section was immediately stained with 1% 2,3,5-triphenyltetrazolium chloride (TTC; obtained from Sigma-Aldrich) solution at 37 °C for 30 min, followed by fixation in 4% paraformaldehyde (PFA). Viable tissue stained red, while infarcted areas remained unstained. All sections were photographed, and infarct volume was quantified using ImageJ software according to the formula: Infarct volume = Total infarct area / Total brain slice area. Biochemical parameter assays (LDH, MDA, and SOD) Following blood collection, serum was separated by centrifugation at 3000 × g for 10 min at 4 °C. Brain tissues were homogenized in ice-cold PBS to prepare tissue lysates. Lactate dehydrogenase (LDH) activity in serum was measured using the LDH Release Assay Kit (WST-8 method; Beyotime Biotechnology, Cat. #C0019M). Malondialdehyde (MDA) content and superoxide dismutase (SOD) activity in brain homogenates were determined using the Lipid Peroxidation MDA Assay Kit (Beyotime, Cat. #S0131M) and the Total SOD Activity Assay Kit (WST-8 method; Beyotime, Cat. #S0101S), respectively. LDH, MDA, and SOD assays were performed with at least 6 biological replicates per group, and data are presented as mean ± sd. Ex vivo fluorescence imaging Male C57BL/6 J mice (6–8 weeks old, 20–25 g, specific pathogen-free grade) were intravenously administered donor Et-1 (20 μmol kg−1) via the tail vein. At designated time points (10 min, 30 min, 1 h, 6 h, and 12 h) post-injection, animals (n = 3 per time point) were euthanized by cervical dislocation. Major organs, including the heart, liver, spleen, lungs, and kidneys, were collected and subjected to ex vivo fluorescence imaging using a Tanon ABL-X1 imaging system. Images were acquired under blue light excitation with an emission filter set at 490 nm and an exposure time of 300 ms. Statistics and reproducibility All experiments were independently repeated at least three times, and data are presented as the mean ± sd Where group allocation was applicable, samples were randomly assigned to experimental groups. In animal experiments, mice were randomly allocated to the designated treatment groups before intervention; in cell-based experiments, independently cultured samples were randomly assigned to the respective treatment conditions. Statistical significance was determined using Student’s t-test, with data normality confirmed by the Shapiro–Wilk test (α = 0.05) and homogeneity of variances verified by Bartlett’s test (α = 0.05). Two-tailed tests were applied for single comparisons, and the Bonferroni correction was used to control the type I error rate in multiple comparisons. A threshold of P < 0.05 was considered statistically significant. All analyses were performed using GraphPad Prism 8.0, and raw data were retained to three decimal places to ensure computational accuracy. Moreover, no biological (cellular or animal) data points were excluded unless technical failures occurred during sample collection procedures (e.g., plasma volume) or they were identified as outliers following outlier detection using the ROUT (Robust Outlier Detection Test) method developed in GraphPad Prism 8.0. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Reporting Summary Transparent Peer Review file Supplementary Information Reporting Summary Transparent Peer Review file Source data Source data Source data