A BBB-permeable β₃-AR probe enables redox imaging, natural inhibitors discovery, and therapeutic monitoring in glioblastoma

Abstract

β₃-Adrenergic receptors (β₃-ARs), as a subclass of G protein–coupled receptors (GPCRs), play a pivotal role in regulating oxidative stress. However, the dynamic interplay between their microenvironmental fluctuations and glioma mechanisms remains poorly understood. Here, we report the development of GSHP, a blood-brain barrier (BBB)–permeable probe that simultaneously visualizes β₃-ARs and reversibly monitors the surrounding redox status in real-time. This dual-responsive probe enables reversible dynamic imaging of redox homeostasis around β₃-ARs in living cells under stress conditions, providing direct visual evidence for redox adaptation. Using GSHP for high-throughput screening, we identified and validated baicalin as a potent β₃-AR natural inhibitor that induces glutathione depletion and triggers oxidative stress–mediated apoptosis via the Gα_i/o-extracellular signal-regulated kinase (ERK)–nuclear factor erythroid 2–related factor 2 (Nrf2)–glutamate–cysteine ligase catalytic subunit (GCLc) signaling pathway in U251 glioblastoma cells. In orthotopic U251 glioma mouse models, GSHP penetrated the brain and enabled dual-channel imaging of β₃-AR overexpression and redox imbalance in vivo, allowing for effective glioma discrimination and demonstrating its potential for therapeutic monitoring. GSHP thus serves as a versatile platform for studying β₃-AR–related redox biology and facilitating therapeutic discovery in brain diseases.

A BBB-permeable β₃-AR probe bridges molecular imaging and therapeutic discovery in glioblastoma.

INTRODUCTION

Maintaining redox homeostasis is critical for the survival, progression, and therapeutic resistance of aggressive cancers such as glioblastoma (GBM) (1–5). GBM is among the most common and lethal primary brain tumors, characterized by rapid proliferation, invasiveness, and resistance to conventional therapies (6, 7). A defining feature of glioma cells is their ability to maintain redox homeostasis under oxidative stress, primarily through up-regulation of endogenous antioxidant systems, particularly glutathione (GSH), which detoxifies reactive oxygen species (ROS), maintains proliferative signaling, and confers resistance to therapy (6, 8). This redox regulatory system is tightly coupled to membrane receptor signaling networks, among which G protein–coupled receptors (GPCRs) play a central role. GPCRs constitute the largest family of membrane receptors and are essential for transducing extracellular signals into intracellular responses (9–11).

β-Adrenergic receptors (β-ARs), a major subclass of GPCRs (12–14), serve as critical regulators of cardiovascular function and metabolic homeostasis (15). Of the three β-AR subtypes (β₁, β₂, and β₃) (16, 17), β₃-AR has emerged as a critical modulator of redox biology, which is widely distributed in various tissues (18, 19), including the brain (20–23), and is increasingly recognized for its association with diverse physiological processes, particularly inflammation and oxidative stress (23–26). Activation of β₃-AR has been shown to modulate cellular redox homeostasis by up-regulating NADPH (reduced form of nicotinamide adenine dinucleotide phosphate) oxidase activity and promoting the expression of antioxidant enzymes, thereby establishing a functional connection with cellular oxidative stress responses (27, 28). Notably, β₃-AR is overexpressed in high-grade gliomas, implicating it in tumor adaptation to oxidative stress and malignant progression (29–32). These observations raise a compelling hypothesis that glioma cells may exploit the known role of β₃-AR in redox regulation to sustain their survival, suggesting that targeting β₃-AR to disrupt the redox balance holds therapeutic potential (33, 34). However, the precise mechanistic links between β₃-AR signaling and redox homeostasis in gliomas remain poorly understood. A major obstacle lies in the lack of blood-brain barrier (BBB)–permeable molecular tools that allow simultaneous, real-time, in situ visualization of β₃-AR distribution and monitoring of the surrounding redox state, particularly GSH levels, within the native intact brain environment. Conventional imaging approaches are inadequate for probing both spatial distribution and dynamic redox fluctuations in the complex tumor microenvironment of the brain.

Here, we report GSHP, a rationally designed dual-functional fluorescent probe with BBB permeability that enables simultaneous imaging of β₃-AR distribution and real-time monitoring of local GSH fluctuations (Fig. 1). GSHP’s design integrates a β₃-AR–targeting motif derived from the carazolol warhead into a BBB-permeable quinoline fluorescent backbone, along with a molecular rotor unit that generates enhanced fluorescent signals upon binding to the receptor, thereby offering a spatial readout for β₃-AR distribution. Concurrently, its cyano-containing unsaturated alkenyl moiety undergoes a reversible Michael addition reaction with local GSH, yielding notable blue-shifted fluorescent emission. This dual-channel fluorescence response enables us to specifically assess GSH fluctuations within the β₃-AR–enriched glioma microenvironment, thereby providing unique insights into localized redox dynamics. Moreover, through high-throughput screening (HTS) with GSHP, the identified compound, baicalin, was validated as a β₃-AR natural inhibitor that depletes GSH, induces oxidative stress, and triggers apoptosis through the Gα_i/o-ERK-Nrf2-GCLc signaling pathway in U251 glioma cells. In an orthotopic glioma model, leveraging GSHP’s BBB penetration capability, we used dual-channel GSHP imaging to noninvasively map β₃-AR expression patterns and local GSH gradients, allowing for high-contrast delineation of tumors and offering a tool for the noninvasive monitoring and assessment of glioma treatment efficacy in a preclinical model. This platform has the potential to open new avenues for elucidating the role of β₃-AR in glioma biology and redox regulation, as well as for facilitating the discovery of natural modulators or related therapeutic molecules targeting redox-sensitive pathways in brain tumors.

Fig. 1. Schematic diagram of the design and application of GSHP, a redox-reversible probe targeting β₃-AR with BBB permeability.

(A) GSHP is a dual-responsive fluorescent probe capable of simultaneously targeting and imaging β₃-AR while reversibly monitoring GSH homeostasis changes in the receptor’s local microenvironment. (B) This multifunctional probe can be applied to dual-channel glioma diagnosis, monitoring of redox states, high-throughput drug screening, and evaluation of glioma treatment efficacy. In-depth mechanistic studies reveal that baicalin, identified through GSHP screening, acts as a natural inhibitor of β₃-AR. By inhibiting β₃-AR and its downstream Gα_i/o-ERK-Nrf2-GCLc signaling pathway, it depletes GSH in U251 GBM cells, thereby inducing oxidative stress–mediated apoptosis.

RESULTS

Design of a dual-responsive probe with BBB permeability for β₃-AR mapping and redox monitoring

To achieve visual imaging of β₃-ARs while simultaneously monitoring their surrounding redox environment, we set out to develop a β₃-AR–localized and redox-reversible fluorescent probe. Inspired by β-AR–targeted probe JE1319 developed by Annibale’s group, which uses a carazolol-based ligand for β-AR imaging (35), our previous work demonstrated that quinoline derivatives incorporating the carazolol scaffold can selectively label β₃-ARs with high fluorescence specificity, thereby establishing a foundation for subtype-selective receptor imaging (24). However, while such probes effectively visualize β₃-ARs, they cannot reversibly probe the local microenvironment surrounding the β₃-AR receptor. To address this limitation, we drew inspiration from the GSH-responsive probes developed by Yoon’s and Wang’s groups, namely, the RT-GSH and QG series, which use cyano-substituted electrophilic groups to achieve reversible and real-time detection of intracellular GSH throughout whole cells (fig. S1) (36, 37). Nevertheless, the absence of specific targeting in these probes prevents precise localization to β₃-ARs, thereby limiting accurate monitoring of local redox dynamics surrounding membrane receptors. Notably, due to BBB constraints, no probes currently exist for specifically detecting β₃-AR in the living brain (table S1 and fig. S2) (24, 35–48). Building on these insights, we propose designing a BBB-permeable redox-reversible probe specifically targeting β₃-ARs. Crucially, this single-molecular platform enables both precise localization imaging of β₃-ARs and real-time monitoring of redox dynamics within their specific local microenvironments, offering a powerful tool for investigating chemical-biological interplay surrounding β₃-ARs. Quinoline derivatives were selected as the fluorescent scaffold due to their potential BBB permeability and superior photophysical properties, including high quantum yields, two-photon excitation capability, and near-infrared emission characteristics (49–52).

Guided by these principles, we designed and synthesized GSHP, a dual-functional fluorescent probe constructed on a quinoline scaffold. Specifically, GSHP integrates a carazolol-derived β₃-AR ligand with a cyano-activated alkene unit, which can undergo reversible Michael addition with GSH. Upon binding to β₃-AR, the molecular rotor within GSHP exhibits enhanced red fluorescence emission, whereas reaction with GSH triggers a pronounced blue shift, enabling dual-channel, ratiometric imaging of both receptor localization and its surrounding redox status (Fig. 2A). Straightforwardly, GSHP was achieved through a one-step aldol condensation between a quinoline aldehyde and an active α-hydrogen site within the carazolol-derived warhead, ensuring efficient targeting of β₃-ARs. To validate the targeting specificity, a nontargeting control probe, GSHC, was also synthesized, which retains the same fluorophore but lacks the β₃-AR binding moiety (Fig. 2A). Detailed synthetic procedures and structural characterization of both GSHP and GSHC are provided in the Supplementary Materials.

Fig. 2. Design of dual-responsive fluorescent probe GSHP for β₃-ARs mapping and redox monitoring.

(A) GSHP integrates β₃-AR targeting with GSH-responsive signaling, enabling simultaneous visualization of receptor distribution and local redox dynamics. GSHC, lacking the targeting moiety, serves as a nonspecific control probe. (B and C) Fluorescence (F.L.) spectra of GSHP (B) and GSHC (C) (10 μM) in PBS/glycerol mixtures with varying viscosity. (D and E) Fluorescence spectra of GSHP (D) and GSHC (E) (1 μM) in the presence of cytoplasmic or membrane protein (1 μg/ml each) extracted from β₃-AR overexpressing HEK293 cells (λ_ex = 470 nm, λ_em = 655 nm, slit = 5/5 nm, pH 7.4, 37°C). (F and G) Fluorescence spectra of GSHP (F) and GSHC (G) (1 μM) upon increasing concentrations of membrane proteins (0 to 5 μg/ml). (H) Molecular docking of GSHP with β₃-AR (PDB: 7XJH), showing overall binding pose and interaction with residue Asn²⁰¹. (I) Two-dimensional interaction map of GSHP within β₃-AR binding pocket. (J) Fluorescence imaging of HEK293T cells and β₃-AR–transfected HEK293T cells incubated with GSHP or GSHC (100 nM), in the absence or presence of the β₃-AR antagonist carazolol (λ_ex = 488 nm and λ_em = 620 to 680 nm). (K) IF imaging of β₃-AR–transfected HEK293T cells incubated with GSHP or GSHC (100 nM). The β₃-AR channel (ADRB3) was displayed in pseudocolor magenta (λ_ex = 561 nm and λ_em = 580 to 620 nm). a.u., arbitrary units.

In vitro spectral response of GSHP to β-ARs and fluorescence imaging of β₃-ARs in living cells

Following the abovementioned design, GSHP is anticipated to emit a robust fluorescence signal upon binding to the ligand-binding domain of β₃-ARs, whereas the nontargeting control probe, GSHC, should not exhibit such a response. To mimic the viscosity of membrane protein–rich microenvironments, we chose the phosphate-buffered saline (PBS)–glycerol system as the initial in vitro testing system. The test results revealed that the emission of GSHP and GSHC at 655 nm was negligible in the pure PBS buffer system. As the glycerol ratio increased, the emission gradually enhanced, reaching a maximum in the pure glycerol system (Fig. 2, B and C). This behavior likely results from restricted molecular rotation of GSHP or GSHC, thereby diminishing the dissipation of capacity associated with the rotation of the molecular rotor and subsequently releasing a strong fluorescence signal. We envisioned that the binding interaction between GSHP and β₃-AR might be affected by the unique microenvironment. To evaluate this, we investigated the response of GSHP or GSHC to β-ARs in biological samples by extracting membrane and cytoplasmic proteins from β₃–human embryonic kidney (HEK) 293 cells stably expressing β₃-ARs. Remarkably, the enhancement in signal intensity became apparent with the addition of only 1 μM GSHP to the membrane protein extract, while GSHC remains unchanged (Fig. 2, D and E). Furthermore, the fluorescence intensity of GSHP at 655 nm exhibited a concentration-dependent enhancement as the amount of membrane protein increased (Fig. 2F). In contrast, the control probe, GSHC, exhibited a negligible fluorescence response to extracted membrane protein fractions, even at high concentrations (Fig. 2G), which strongly demonstrated that the enhanced emission of GSHP at 655 nm arises from specific binding interactions with β₃-AR.

Moreover, molecular docking studies between GSHP and β₃-AR were conducted, expanding our theoretical insight. GSHP exhibits specific interactions with amino acid residues associated with the seven-transmembrane helix microenvironment of β₃-AR through van der Waals forces and hydrogen bonding interactions (Fig. 2, H and I). The calculated binding energy for the GSHP/β₃-AR complex was determined to be −8.9 kcal/mol. Compared to the binding energies of GSHP/β₁-AR (−6.7 kcal/mol), GSHP/β₂-AR (−7.3 kcal/mol), or carazolol/β₃-AR (−5.6 kcal/mol) (figs. S3 and S4), GSHP exhibits a notable advantage, confirming its efficient targeting of β₃-AR. This property highlights its potential as a powerful tool for visualizing and functionally analyzing β₃-AR.

Given the high environmental sensitivity of GSHP observed in vitro, we next evaluated its performance in living cells. Initially, MTT assays confirmed that both GSHP and GSHC exhibited low cytotoxicity toward HEK293T cells at concentrations up to 10 μM (fig. S5). Subsequent live-cell imaging experiments were conducted using a concentration of 100 nM, a level at which probe-induced toxicity was negligible. We then used a HEK293T cell model with stable β₃-AR overexpression. Following incubation, these cells exhibited strong fluorescence signals upon GSHP treatment, in stark contrast to the negligible signal observed with the nontargeting control probe GSHC, or in wild-type cells (Fig. 2J and fig. S6). Pharmacological studies further confirmed the probe’s high selectivity for β₃-AR. The fluorescence signal in HEK293T cells overexpressing β₃-AR was markedly suppressed by the β₃-AR antagonist SR59230A (Fig. 2J). Conversely, in cells transfected with β₁-AR or β₂-AR (where endogenous β₃-AR expression is comparatively low) (fig. S7), the signal showed no discernible effect from the corresponding antagonists, underscoring the probe’s specific interaction with β₃-AR. Moreover, to establish the spatial correlation between signal and receptor localization, we performed immunofluorescence (IF) analysis, which showed substantial colocalization between GSHP and anti-β₃-AR antibody signals, yielding a high Pearson correlation coefficient of 0.86 (Fig. 2K). These results collectively demonstrate that the fluorescence “turn-on” effect is a direct result of GSHP selectively binding to β₃-AR.

After thoroughly validating the specificity of GSHP in engineered cellular systems, we next addressed the central goal of this study: enabling the visualization of endogenous β₃-AR in GBM. As a prerequisite, we sought to identify an appropriate GBM cell model with high endogenous expression levels of the target β₃-AR. To comprehensively assess β₃-AR expression, we expanded our analysis to include multiple human brain cancer cell lines: four mesenchymal-subtype GBM models (U251, U87, LN229, and T98G), the neuroblastoma line SH-SY5Y, and the murine neuronal line HT22. The results demonstrated that β₃-AR expression was higher in the human GBM cells than in the murine neuronal cells, with U251 cells having the highest endogenous expression level (fig. S8, A to D). In addition, the specificity of the signal was confirmed through competitive binding assays. Treatment with the β₃-AR antagonist SR59230A significantly attenuated the GSHP signal in the four GBM cell lines, while not affecting the weak fluorescence observed in the control HT22 and SH-SY5Y cells (fig. S8, E and F). Besides, we verified that GSHP displayed minimal cytotoxicity (fig. S9). Together, these results establish GSHP as a reliable and specific tool for visualizing endogenous β₃-AR in living glioma cells and identify U251 cells as the ideal cellular model for subsequent mechanistic investigations and drug-screening applications.

Reversible detection of GSH via ratiometric fluorescence of GSHP

To assess the feasibility of GSHP for detecting GSH within physiological systems, we conducted a comparative analysis of the fluorescence emission spectra of GSHP in the presence and absence of GSH (Fig. 3A). Following the incubation of GSHP with GSH, a discernible blue shift in the maximum excitation wavelength of GSHP from 470 to 395 nm was observed, indicating the generation of the adduct product GSHP-SG (Fig. 3B). Concomitantly, a dose-dependent relationship between GSHP spectra and GSH concentration was identified. As GSH concentration increased, the emission peak at 525 nm gradually enhanced (fig. S10). Simultaneously, upon GSH addition, the absorption at 395 nm gradually increased while the absorption at 460 nm gradually decreased (fig. S11). When the excitation wavelength was kept at 470 nm, the emission at 525 nm gradually increased with the addition of GSH, whereas the emission at 650 nm gradually decreased (Fig. 3C). This ratiometric fluorescence change establishes a robust method for GSH quantification, greatly improving detection reliability by minimizing potential interferences. In addition, we found that the control probe, GSHC, also exhibited a blue-shifted green fluorescence emission after responding to GSH (fig. S12), demonstrating the generalizability of this strategy. To better understand the spectral responses of GSHP, we conducted density functional theory (DFT) calculations. As shown in Fig. 3D, the lowest unoccupied molecular orbital (LUMO) → highest occupied molecular orbital 1 (HOMO-1) transition of GSHP contributes most to its emission, having the largest oscillator strength and a calculated energy gap (ΔE) of 5.11 eV. Notably, this ΔE (5.11 eV) is lower than that of the product formed after reaction with GSH (ΔE = 5.78 eV), which was mainly ascribed to the radiation transition of the LUMO to the HOMO. These computational results are in good agreement with the observed blue-shift change of fluorescence, which clearly explains the potential mechanism of the notable spectral change of GSHP after interaction with GSH.

Fig. 3. Reversible fluorescence response of GSHP and its imaging application in glioma cells and tissues.

(A) Schematic of the reversible reaction between GSHP and GSH. (B to F) In vitro characterization of the ratiometric fluorescence response of GSHP (10 μM) to GSH. (B) Normalized excitation/emission spectra before (red) and after (green) GSH (10 mM) addition. (C) Fluorescence titration spectra showing a ratiometric shift with increasing GSH concentration (0 to 30 mM). (D) Frontier molecular orbitals of GSHP and its adduct calculated by DFT. (E) The fluorescence ratio (F₅₂₅/F₆₅₀) increased with GSH concentration, showing excellent linearity in the 0 to 30 mM range (inset, R² = 0.991). (F) Kinetic analysis confirms reaction reversibility. (G to I) Live-cell imaging in a GSHP-loaded U251 cell demonstrates dynamic ratiometric response to sequential H₂O₂ (500 μM), GSH (1 mM), and NEM (100 μM) treatments (G); quantification of the fluorescence intensity in each channel (H) and ratiometric trace (I) are shown. Data are means ± SEM (n = 3). DIC, differential interference contrast. (J) Imaging of a human glioma section (Patient 1) costained with GSHP [green (G)/red (R) channels] and an ADRB3 antibody (magenta). Quantitative analysis confirmed significantly higher individual channel intensities (K) and a higher fluorescence ratio (L) in the tumor. Data are means ± SEM (n = 5 fields). Scale bars, 10 μm (G) and 50 μm (J). Statistical significance by two-tailed unpaired t test, **P < 0.01; ***P < 0.001.

Notably, a robust linear relationship has been established between the fluorescence intensity ratio (F_{525 nm}/F_{650 nm}) and GSH concentrations ranging from 0.1 to 30 mM (Fig. 3E). This detection range is well-suited for monitoring physiological GSH levels, offering a broad dynamic window for cellular or in vivo applications. The limit of detection of GSHP for GSH was calculated to be 0.24 μM according to the 3σ/m formula, indicating that it has a good affinity binding ability and high sensitivity. Further investigations were conducted to assess the specific selectivity of GSHP. The results revealed that cysteine (Cys), a structurally similar biothiol, exhibited the strongest competitive interference; however, its effect was negligible under physiological GSH concentrations (fig. S13). Moreover, the fluorescence signal of GSHP remained stable across a pH range of 5.0 to 9.0, confirming its suitability for physiological environments (fig. S14). To further validate its applicability, the performance of GSHP was assessed in complex systems containing membrane lysates. Consistent with earlier results, fluorescence enhancement at 525 nm and attenuation at 650 nm were observed in these environments (fig. S15). These findings demonstrate that GSHP is a reliable tool for sensitive and ratiometric detection of dynamic GSH fluctuations under complex biological conditions.

Encouraged by the above results, we subsequently investigated whether GSHP could enable a reversible detection of GSH. A time-dependent spectroscopy test was initially conducted to assess the dynamic response. Upon introducing 10 mM GSH, the fluorescence intensity at 525 nm was enhanced and reached a plateau within a relatively short period (fig. S16). Subsequent treatment with the thiol scavenger N-ethylmaleimide (NEM) resulted in a marked decrease in fluorescence intensity, stabilizing again within a comparable time span. These results indicate that GSHP responds sensitively and reversibly to GSH under physiological conditions. To further characterize the reversibility of the GSHP-GSH interaction, the forward and reverse reaction rates were investigated by the concentration-jump relaxation assays. As shown in Fig. 3F, following equilibrium attainment, the addition of GSH induced a rapid spectral shift, with a new equilibrium promptly attained. Furthermore, we investigated the reaction kinetics of GSHP in the presence of GSH. The results show that the GSHP probe exhibits robust operational stability, retaining over 80% of its initial peak intensity through six consecutive cycles (fig. S17). Collectively, these findings unequivocally demonstrate the complete reversibility of the reaction between GSHP and GSH, suggesting its potential as a robust chemical probe for real-time, ratiometric monitoring of dynamic GSH fluctuations in biological systems with high sensitivity and reversible response characteristics.

Imaging the redox status of the microenvironment surrounding β₃-ARs with GSHP

Given the satisfactory performance of GSHP in both β₃-AR localization imaging and in vitro reversible GSH response assays, we next sought to assess its capability to monitor dynamic redox fluctuations in the local microenvironment of β₃-AR in live cells. We first validated the probe’s responsiveness to physiological stimuli using the β₃-AR–overexpressing HEK293T cell model (fig. S18). Changes were tracked in real-time via confocal fluorescence intensity ratio analysis, with the GSHP channel displayed in red and the GSHP-SG channel in green. We tested glutamate (Glu), a known abundant excitatory neurotransmitter that induces oxidative stress, to observe whether it disrupts GSH levels in the β₃-AR local microenvironment. Upon stimulation with 2 mM Glu, cells loaded with GSHP exhibited a rapid elevation in the fluorescence ratio (F_{488 nm}/F_{405 nm}) within 120 s. This result indicates that Glu stress temporarily disrupts redox homeostasis around β₃-ARs. Notably, this oxidative shift was fully reversible. Subsequent addition of exogenous GSH caused the ratio to plummet, green fluorescence (F_{405 nm}) markedly increased, while red signal (F_{488 nm}) significantly diminished. This reduced state gradually reverted over time, reflecting the intrinsic ability of live cells to restore redox equilibrium. Last, the introduction of the thiol scavenger NEM rapidly depleted cellular GSH, causing the ratio to rise again. These preliminary experiments in a controlled cellular model confirm that GSHP serves as a potent tool for monitoring dynamically reversible redox events in the local microenvironment of β₃-AR.

Having validated the probe’s dynamic responsiveness, we further explored its applicability in disease-relevant contexts. Given its high receptor expression and dependence on elevated intracellular GSH for antioxidative stress resistance, we tracked the dynamic changes in the local redox microenvironment around β₃-AR receptors in U251 glioma cells using real-time imaging technology. As shown in Fig. 3G, U251 cells loaded with GSHP exhibited a dynamic and reversible response to a series of redox stimuli. Upon stimulation with 500 μM H₂O₂, cells exhibited a mild oxidative shift. Notably, the subsequent addition of 1 mM exogenous GSH reversed this effect, triggering a strong reductive response. The green fluorescence intensity (F_{405 nm}) significantly increased, while the red fluorescence signal (F_{488 nm}) markedly decreased (Fig. 3H). Consequently, the fluorescence ratio (F_{488 nm}/F_{405 nm}) dropped significantly, as quantified in Fig. 3I. In contrast, the introduction of 100 μM NEM, a thiol scavenger, effectively depleted GSH, restoring the red fluorescence signal and rapidly elevating the fluorescence ratio. Altogether, these observations demonstrate that GSHP enables dynamic, reversible, and spatially resolved monitoring of GSH fluctuations in the microenvironment surrounding β₃-AR within living U251 cells, offering a valuable tool for investigating localized oxidative stress and real-time redox signaling associated with β₃-AR events in glioma cells.

To bridge these findings to clinical reality, we further performed fluorescence imaging on tissue sections from human glioma patients. As shown in Fig. 3J, a tissue section from a World Health Organization (WHO) Grade 4, isocitrate dehydrogenase (IDH)–wild-type GBM, with its diagnosis and distinct tumor-paracancerous boundaries validated by corresponding hematoxylin and eosin (H&E) staining (fig. S19 and table S2), showed distinct fluorescence patterns in tumor regions compared with adjacent paracancerous tissue. Both the GSHP-SG (green) channel, reflecting GSH abundance, and the GSHP (red) channel, indicating β₃-AR expression, displayed markedly stronger signals in tumor areas (Fig. 3K). IF staining with an anti-β3-AR (ADRB3)–specific antibody (magenta) further confirmed β₃-AR up-regulation, showing strong spatial correlation with the GSHP signaling within the cancerous tissue. Ratiometric analysis revealed a significantly elevated GSHP/GSHP-SG ratio in glioma regions compared to the paracancerous areas (Fig. 3L). Moreover, the reproducibility of these findings was confirmed in two additional patient samples: a molecularly distinct WHO Grade 3 tumor and a WHO Grade 2 tumor with diagnoses and grades also confirmed by histopathology (fig. S19 and table S2). Both samples similarly exhibited significantly higher signals for GSH, β₃-AR, and the ratiometric value in the tumor core compared to adjacent tissue (fig. S20). These results indicate that glioma tissues are characterized by GSH enrichment and β₃-AR overexpression, together defining a distinctive redox-receptor biochemical signature. Collectively, these data highlight the potential of GSHP as a diagnostic imaging probe capable of discriminating malignant from normal tissue, suggesting its possible translational promise for clinical glioma evaluation.

A GSHP-powered high-throughput screen unlocks baicalin as a β₃-AR natural inhibitor

Given the exceptional performance of GSHP in detecting abnormally up-regulated β₃-AR in clinical glioma tissue samples and U251 live cells, we hypothesize that this probe could serve as a HTS tool for rapidly finding natural modulators capable of targeting and regulating the β₃-AR receptor, thereby advancing therapeutic interventions for glioma. To this end, we further developed a fluorescence-based HTS platform using GSHP to identify natural modulators that competitively bind to β₃-AR at the live cell level. The screening principle relies on competitive ligands competing with GSHP from the receptor binding site, resulting in a quantifiable decrease in red fluorescence intensity.

Control experiments confirmed that most tested natural products did not significantly affect cellular redox balance under short-term incubation (fig. S21), indicating that fluorescence changes primarily resulted from competitive binding rather than nonspecific redox perturbation. Using this strategy, we screened a natural product library in live cells with GSHP (Fig. 4A). Heatmap analyses revealed that several compounds, particularly those with flavonoid or alkaloid scaffolds, substantially reduced GSHP fluorescence signals (53, 54). Among them, the classical flavonoid baicalin produced the most pronounced signal decrease at the β₃-AR binding site, identifying it as a potent competitive ligand candidate for β₃-AR. To validate this finding in disease-relevant models, we further examined whether baicalin could competitively inhibit GSHP detection of endogenous β₃-AR in an expanded panel of brain cancer cell lines. As shown in Fig. 4 (B and C), baicalin treatment markedly diminished GSHP fluorescence intensity in all four tested GBM cell lines (U251, U87, LN229, and T98G), but not in the low-expressing control line SH-SY5Y. Collectively, these results demonstrate that the GSHP-based high-throughput screening strategy provides a robust and rapid platform for ligand discovery, enabling the identification of natural β₃-AR modulators such as baicalin. The ability of this technology to screen functional compounds in living cells highlights GSHP’s potential as an interdisciplinary chemical-biomedical research and translational tool for finding therapeutic candidates targeting β₃-AR in GBM.

Fig. 4. High-throughput screening identifies baicalin as a β₃-AR inhibitor that modulates noncanonical CREB signaling across multiple glioma cell lines.

(A) Fluorescence-based high-throughput screening of a natural product library in β₃-AR–overexpressing HEK293T cells. Cells were pretreated with various compounds (20 μM) before incubation with GSHP (100 nM). Representative fluorescence images are shown with compound structures overlaid. The heatmap below summarizes the relative fluorescence intensity (normalized to the DMSO control), with lighter colors indicating stronger inhibition of the GSHP signal. Baicalin (position 2d) was identified as the most potent inhibitor. Scale bar, 25 μm. (B) Confocal images showing that baicalin (25 μg/ml) competitively displaces GSHP (100 nM) in an expanded panel of GBM cell lines (U251, U87, LN229, and T98G). Scale bar, 20 μm. (C) Quantification of relative fluorescence intensity from (B). The signal from the U251 group (GSHP only) was set to 1. (D and E) Western blot (D) and quantification (E) showing baicalin induces p-CREB in U251 cells in a dose-dependent manner. (F and G) Western blot (F) and quantification (G) confirming baicalin acts as a β₃-AR inhibitor, as its effect on p-CREB mirrors the antagonist SR59230A and opposes the agonist BRL-37344. Data are presented as means ± SEM (n = 3). Statistical significance was determined by a one-way analysis of variance (ANOVA) with multiple comparisons, *P < 0.05; **P < 0.01; ***P < 0.001. ns, not significant.

Having confirmed baicalin as a direct β₃-AR ligand, we next sought to validate its modulatory function by examining its impact on downstream signaling. While β-ARs can signal through the classical Gs-cAMP-PKA-CREB pathway, where inhibition would be expected to decrease cAMP response element–binding protein (CREB) phosphorylation (p-CREB) (55–57), our results revealed a counterintuitive outcome. In U251 cells, baicalin treatment led to a dose-dependent increase in p-CREB levels (Fig. 4, C and D). To confirm that this unexpected effect was genuinely mediated by β₃-AR inhibition, we used specific pharmacological tools. As shown in Fig. 4E, the known β₃-AR antagonist SR59230A perfectly mimicked baicalin’s effect, also causing a robust increase in p-CREB. Conversely, the β₃-AR agonist BRL-37344 led to a decrease in p-CREB. These opposing effects between the agonist and the inhibitors (baicalin and SR59230A) strongly suggest that in U251 glioma cells, β₃-AR tonically suppresses CREB phosphorylation through a noncanonical pathway (Fig. 4F). All these findings trace a clear discovery pathway originating from our innovative GSHP screening platform, which successfully identified baicalin from a natural product library as a promising β₃-AR–targeting modulator. We further confirmed baicalin as a β₃-AR inhibitor rather than an agonist, establishing a solid foundation for subsequent anti-GBM studies.

Imaging detection of baicalin’s effects on β₃-AR–mediated redox signaling and GBM invasion

Next, we attempted to use GSHP to detect whether baicalin directly induces oxidative stress in U251 GBM cells. To ensure that the probe itself was biologically inert, we confirmed that up to 10 μM GSHP treatment did not significantly affect the expression levels of key signaling proteins (p-CREB, Nrf2, and β₃-AR), thus validating it as a noninterfering reporting tool for these experiments (fig. S22). As shown in Fig. 5 (A and B), both the GSHP and GSHP-SG channels exhibited strong fluorescence signals in untreated U251 cells. Upon addition of baicalin, the fluorescent signal of GSHP decreased dramatically within half an hour due to competitive inhibition, while cellular GSH levels were largely unaffected at this early time point. Notably, after 12 hours of baicalin treatment, the GSHP-SG signal significantly decreased, but this signal was restored by the addition of N-acetylcysteine (NAC; an antioxidant) (Fig. 5C). These results indicate that baicalin can influence cellular redox homeostasis in U251 cells. Subsequently, we examined the levels of ROS and the degree of apoptosis in U251 cells after different durations of baicalin treatment. The findings indicated that with the prolongation of baicalin treatment time (3, 6, 9, and 12 hours), the level of ROS in U251 cells significantly and continuously increased, along with a gradual increase in the degree of apoptosis (Fig. 5D and fig. S23). These observations strongly suggest that baicalin induces severe oxidative stress in U251 cells, consequently triggering apoptosis.

Fig. 5. Imaging revealed that baicalin induces oxidative stress and promotes apoptosis in U251 GBM cells by inhibiting the β3-AR–ERK–Nrf2 axis.

(A) Confocal fluorescence images of U251 cells incubated with GSHP (100 nM) after treatment with baicalin (Bai; 25 μg/ml) for different durations (0.5 and 12 hours) or cotreatment with NAC (5 mM) for 12 hours. The red channel (GSHP) indicates β₃-AR occupancy, and the green channel (GSHP-SG) reports local GSH levels. Scale bar, 15 μm. (B and C) Quantification of the fluorescence signals from (A), showing the relative fluorescence intensity of GSHP and GSHP-SG (B), and the GSHP/GSHP-SG ratio (C). (D) Fluorescence images of ROS levels (DCFH-DA stain) and apoptosis (TUNEL stain) in U251 cells treated with baicalin (25 μg/ml) for the indicated times (3, 6, 9, and 12 hours). Scale bar, 50 μm. (E) Western blot analysis showing the effect of baicalin on nNrf2 levels in the different brain cancer cell lines. Lamin B serves as a nuclear loading control. (F) Western blot analysis elucidating the signaling pathway in U251 cells treated with various pharmacological modulators for 24 hours. Treatments include baicalin (25 μg/ml), β₃-AR agonist BRL-37344 (10 μM), ERK inhibitor PD98059 (20 μM), curcumin (20 μM), Nrf2 inhibitor brusatol (10 μM), and the antioxidant NAC (5 mM). (G) Representative images from a Transwell invasion assay, showing the effect of baicalin (0, 25, and 50 μg/ml) on cell lines with different β₃-AR expression levels (SH-SY5Y, U87, and U251). (H) Corresponding quantification for (E). (I) Quantification of invasion rate for (G). (J) A schematic summary of the mechanism of action of baicalin, created with Figdraw.com. Data are presented as means ± SEM (n = 3). Statistical significance was calculated using a one-way ANOVA with a multiple comparisons test, *P < 0.05; **P < 0.01; ***P < 0.001.

To further elucidate the underlying molecular mechanisms, we next investigated alternative β₃-AR–regulated proteins. Nrf2, a master regulator of redox homeostasis, is closely associated with both antioxidant defense and invasive behavior. We therefore investigated whether baicalin affected the nuclear translocation of Nrf2 in neuronal cells. Western blotting analysis revealed minimal effects in SH-SY5Y cells, moderate reduction in U87, LN229, and T98G cells, and a marked decrease in nuclear Nrf2 (nNrf2) levels in U251 cells, even at 25 μg/ml (Fig. 5, E and H, and fig. S24). This pattern closely paralleled β₃-AR expression and baicalin’s inhibitory potency across the three cell lines. To clarify how baicalin regulates nNrf2 via β₃-AR to induce oxidative stress and apoptosis, we analyzed key signaling and apoptotic proteins by Western blotting (Fig. 5F). The results showed that baicalin treatment significantly down-regulated phosphorylated ERK1/2, nNrf2, and its downstream key antioxidant enzyme GCLc, as well as up-regulated apoptotic markers cleaved caspase-3 and Bcl-2 associated X protein (BAX). These findings suggest that baicalin suppressed important prosurvival antioxidant pathways and triggered apoptosis in U251 cells. In contrast, cells treated with β₃-AR agonist BRL-37344 exhibited the opposite effects. Moreover, cells treated with the β₃-AR agonist BRL-37344 reversed baicalin-induced suppression of nNrf2 axis proteins (nNrf2 and GCLc). These observations demonstrate a functional relationship between nNrf2 signaling and β₃-AR activation. In addition, treatment with baicalin and the ERK inhibitor PD98059 resulted in a slight increase in p-p65 levels, suggesting that the activation of p-p65 might not be a direct downstream of ERK (58), but rather a secondary or compensatory response triggered by the decreased antioxidant capacity and increased oxidative stress due to the inhibition of the ERK-Nrf2 pathway (fig. S24). However, this p65 activation was insufficient to reverse baicalin-induced apoptosis, highlighting the predominant role of the β₃-AR–ERK–Nrf2 axis in maintaining redox homeostasis and survival of U251 cells. Furthermore, the Nrf2 inhibitor brusatol, when administered alone, significantly down-regulated nNrf2 and GCLc, while up-regulating apoptotic markers, similar to the effects observed with baicalin. This suggests that the loss of Nrf2 function alone is sufficient to induce apoptosis. Subsequently, the ERK inhibitor PD98059 replicated the phenotypic effects of baicalin by blocking the upstream signaling pathway, thereby confirming the hierarchical regulation of Nrf2 by ERK. The antioxidant NAC markedly diminished apoptosis by enhancing the antioxidant capacity (59, 60), thereby functionally validating that oxidative stress is the primary driver of baicalin-induced apoptosis in U251 cells (fig. S25). As a critical control, curcumin significantly inhibited phosphorylated p65 and activated Nrf2 independently of the ERK pathway (61). This finding showed that the anti-inflammatory and antioxidant properties of curcumin contrasted with the mechanism of action of baicalin, further underscoring the β₃-AR–ERK–Nrf2 axis as a pivotal hub for antioxidant defense and cell survival in U251 glioma cells, and emphasizing the significance of baicalin in specifically targeting this axis. Together, these results demonstrate that baicalin, a natural modulator of β₃-AR identified through GSHP screening, impairs the antioxidant defense capacity of U251 glioma cells by inhibiting the β₃-AR–ERK–Nrf2 pathway and induces apoptosis mediated by oxidative stress. This provides reliable experimental evidence for targeting β₃-AR in U251 glioma therapy.

Aberrant invasion and metastasis are defining features of high-grade gliomas and major determinants of their poor prognosis (62, 63). As an important class of cell surface receptors, GPCRs play crucial roles in regulating cell proliferation, migration, and invasion (32, 55, 64). Through the aforementioned GSHP imaging studies, we not only detected significant up-regulation of β₃-AR in U251 GBM cells but also confirmed that baicalin, the modulator identified by GSHP screening, induces stress-mediated apoptosis in U251 cells. These findings further prompted us to examine whether baicalin could inhibit GBM invasion. To address this, we compared the effects of baicalin on five cancer cell lines with distinct β₃-AR expression levels. As expected, Transwell migration assays revealed that baicalin suppressed invasion in a dose-dependent manner across all four GBM cell lines (U251, U87, LN229, and T98G), with the inhibitory effect correlating with β₃-AR expression levels (Fig. 5G and fig. S26). In contrast, the effect was negligible in the low-expressing SH-SY5Y cells (Fig. 5I). To verify that baicalin does not directly modulate β₃-AR expression, we performed quantitative polymerase chain reaction (qPCR) analysis. After 24-hour treatment with baicalin (25 and 50 μg/ml), β₃-AR transcriptional levels remained unchanged across the cell lines (fig. S27), suggesting that baicalin primarily affects β₃-AR functional activity. Collectively, these findings confirm that baicalin, identified by GSHP screening, acts as a potent GBM cell inhibitor. Mechanism of action is conserved across multiple GBM cell lines, involving the disruption of the β₃-AR–ERK–Nrf2 signaling axis to induce apoptosis and attenuate downstream proinvasive responses (Fig. 5J).

GSHP dual-channel in vivo imaging for glioma diagnosis and evaluation of baicalin’s anti-glioma efficacy

To evaluate the potential and preclinical value of the GSHP in glioma detection and therapeutic efficacy assessment, we established an orthotopic U251 glioma xenograft (U251 OX) model by intracranial injection of U251 cells into nude mice (Fig. 6A). Given the intracranial location of gliomas, BBB permeability is a prerequisite for effective in vivo diagnostic imaging. Following intravenous administration, GSHP effectively crossed the BBB and rapidly accumulated in the brain tumor region of U251 tumor-bearing mice. In vivo fluorescence imaging revealed specific and time-dependent fluorescence signals in both the GSHP and GSHP-SG channels, with significantly elevated fluorescence intensities in tumor-bearing brains compared to normal controls (Fig. 6, B and C). Ex vivo imaging of dissected brains further confirmed the specific accumulation of the probe within glioma tissue (Fig. 6, B and D). To assess its pharmacokinetics and brain accumulation, we performed an in vivo imaging time-course study using the GSHP probe. The results demonstrate that the probe rapidly crosses the BBB, selectively accumulates in orthotopic U251 tumors, and is largely cleared from the body within 1.5 hours postinjection (fig. S28). These findings illustrate its key attributes for imaging applications: efficient brain delivery, tumor-specific targeting, and rapid systemic clearance. Together, these observations demonstrate that the β₃-AR–targeted probe GSHP exhibits excellent BBB permeability and enables dual-channel, high-contrast visualization of U251 glioma lesions in vivo.

Fig. 6. In vivo dual-channel imaging for glioma diagnosis and therapeutic monitoring of baicalin efficacy using GSHP.

(A) Schematic illustration of the orthotopic U251 glioma model, GSHP-based dual-channel imaging, and therapeutic monitoring workflow, created with Figdraw.com. (B) In vivo fluorescence imaging of GSHP-SG and GSHP signals between normal and U251 orthotopic (OX) glioma–bearing mice. Bottom: Ex vivo brain images at 60 min postinjection. (C and D) Quantification of GSHP-SG and GSHP signals in (B). (E) In vivo fluorescence imaging of GSHP-SG and GSHP signals in different treatment groups at day 20. (F) Quantitative analysis of GSHP-SG and GSHP signals in (E). (G) Representative MRI images showing tumor progression over 20 days in mice treated with PBS, TMZ, or baicalin. (H) Quantification of tumor cross-sectional area derived from MRI images in (G). (I) Kaplan-Meier survival curves of U251 OX mice from different treatment groups. (J) Confocal images of brain slices from normal, untreated U251 OX, and baicalin-treated U251 OX mice. Scale bar, 50 μm. (K and L) Quantitative analysis of GSHP or GSHP-SG and the ratio from (J). (M) H&E staining (top) and TUNEL staining (bottom) of brain sections from each treatment group. (N) Quantification of the tumor area from H&E-stained sections. (O) Quantification of the apoptosis rate based on TUNEL staining. Data are presented as means ± SEM (n = 5 per group for in vivo experiments). Statistical significance was determined using two-tailed Student’s t test (D) and one-way ANOVA with multiple comparisons test [(F), (K), (L), (N), and (O)], *P < 0.05; **P < 0.01; ***P < 0.001.

We next evaluated GSHP as a real-time imaging tool for assessing therapeutic efficacy. First, we performed a detailed pharmacokinetic study of baicalin using a quantitative liquid chromatography–mass spectrometry (LC-MS) method. Following a single intraperitoneal injection (50 mg/kg), we tracked baicalin concentrations over time in plasma, normal brain tissue, and glioma tissue. As shown in fig. S29, the plasma concentration peaked at ~60 min, while the highest level in tumor tissue was observed around 240 min postinjection. Baicalin demonstrated preferential accumulation and extended retention in glioma tissue relative to the adjacent normal brain. Tumor-bearing mice were then treated with PBS, temozolomide (TMZ), or baicalin, followed by in vivo fluorescence imaging. Both GSHP and GSHP-SG signals were significantly reduced in the tumor regions of the TMZ- and baicalin-treated groups relative to controls (Fig. 6, E and F). Remarkably, baicalin treatment resulted in a greater reduction in GSHP signal intensity, suggesting suppressed tumor proliferation. To further validate this, we used magnetic resonance imaging (MRI) and provided complementary confirmation (Fig. 6G). Tumors in the PBS group grew rapidly, while TMZ treatment modestly slowed progression. In contrast, baicalin treatment markedly inhibited tumor growth (Fig. 6H), consistent with the GSHP fluorescence imaging results. Notably, monitoring of body weight throughout the treatment period revealed that mice in the baicalin group maintained stable weight comparable to the PBS group, whereas the TMZ group exhibited a weight loss trend, indicating lower systemic toxicity of baicalin (fig. S30). Moreover, hematological and histopathological analyses confirmed its favorable safety profile. Compared to TMZ, which induces potential toxicity, baicalin, as a natural product, exhibited markedly lower toxicity, with no significant signs of hematological toxicity or organ damage observed (table S3 and fig. S31). The final Kaplan-Meier survival curve analysis further validated baicalin’s therapeutic benefit, showing significantly prolonged survival in the baicalin group (Fig. 6I).

Ex vivo fluorescence imaging of brain sections revealed a significantly elevated GSHP/GSHP-SG ratio in tumors from the baicalin-treated group (Fig. 6, J to L), consistent with reduced local GSH levels and enhanced oxidative stress. Concurrent histological analysis corroborated these findings. H&E staining revealed substantially smaller tumor areas in the baicalin group compared to both the PBS and TMZ groups (Fig. 6, M and N). terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL) staining further demonstrated widespread apoptosis in tumors from the baicalin-treated group (Fig. 6, M and O). To address β₃-AR modulation in tumor tissue after baicalin dosing, we performed Western blot analysis on protein lysates from dissected orthotopic tumors of mice from our four treatment groups (control, U251 OX, U251 OX + baicalin, and U251 OX + TMZ). The results from the tumor tissue were highly consistent with our in vitro findings. In vivo baicalin treatment did not substantially alter β₃-AR expression, but led to a significant reduction in ERK1/2 phosphorylation and in the levels of its downstream targets Nrf2 and GCLc, accompanied by a marked increase in the apoptosis marker cleaved caspase-3. As a mechanistic control, TMZ induced apoptosis via DNA damage, as indicated by a strong γ-H2AX signal, but did not inhibit the ERK-Nrf2 pathway. Together, these data indicate that the therapeutic effect of baicalin primarily stems from its inhibition of signal transduction along the β₃-AR–ERK–Nrf2 axis within U251 tumors, rather than from directly down-regulating β₃-AR expression (fig. S32). Collectively, these results establish GSHP as a sensitive dual-channel imaging probe capable of noninvasive, redox-responsive monitoring of U251 glioma progression and therapeutic response. Baicalin, identified through GSHP-based screening, exerts potent anti-U251 glioma effects by inducing redox imbalance and promoting U251 glioma cell apoptosis, highlighting its potential as a β₃-AR–targeted therapeutic candidate. Furthermore, the excellent safety profile observed for baicalin underscores the potential of targeting β₃-AR as a reliable and well-tolerated anti-glioma strategy.

DISCUSSION

To facilitate redox-based biological studies of β₃-ARs and to explore the potential of targeting β₃-ARs for glioma diagnosis and therapeutic efficacy assessment, this study reports the innovative design and characterization of GSHP, a BBB-permeable, dual-functional probe. By integrating a carazolol-derived recognition unit with a reversible GSH-responsive moiety, GSHP achieves both specific receptor binding and ratiometric detection of GSH dynamics. This dual-functional design allows simultaneous imaging of β₃-AR distribution and real-time monitoring of redox states across various models, from in vitro systems to in vivo orthotopic U251 GBM mouse models. GSHP not only visualizes receptor localization but also captures dynamic GSH variations, providing a powerful tool for investigating redox biology involving β₃-AR. Using the GSHP-based high-throughput screening platform, we identified baicalin as a competitive β₃-AR ligand. Through in-depth molecular mechanism studies, we further revealed baicalin as a natural inhibitor of β₃-AR. It disrupts the β₃-AR–ERK–Nrf2 signaling cascade, leading to reduced nuclear accumulation of Nrf2 and down-regulation of its downstream target GCLc. Inhibition of this pathway enhances oxidative stress and promotes apoptosis in U251 glioma cells. Functional studies confirm that baicalin significantly suppresses the invasive capacity of β₃-AR–overexpressing U251 GBM cells, suggesting the therapeutic potential of β₃-AR modulation in U251 glioma treatment.

Beyond its role in molecular screening, GSHP shows promise for diagnostic and therapeutic evaluation capabilities. Imaging of various brain tumor cells and human glioma tissues revealed that GSHP has the potential to distinguish cancerous from normal cells/tissues based on differences in β₃-AR expression levels and redox status. Crucially, in vivo imaging in U251 glioma orthotopic mouse models confirmed that GSHP efficiently crosses the BBB, accumulates at tumor sites, and enables dual-channel imaging for tumor visualization and therapeutic response monitoring. Using this platform, we confirmed that baicalin, screened by GSHP, inhibits U251 glioma progression and prolongs survival, with its efficacy positively correlated with the enhanced tumor oxidative stress detected by GSHP. Despite promising prospects, this study has limitations. Most findings were derived from the U251 cell line and a single orthotopic mouse model, thus requiring future validation across multiple glioma subtypes and clinically relevant models. Moreover, although the β₃-AR–ERK–Nrf2 pathway has been identified as a key regulatory mechanism in this study, further investigation is warranted to explore other signaling pathways or off-target effects of baicalin. Future work should aim to optimize GSHP’s pharmacokinetics and stability and expand its applications in preclinical imaging, drug discovery, and efficacy assessment, thereby advancing baicalin or related β₃-AR modulators through the drug development pipeline. In summary, GSHP represents a potent chemical tool that bridges molecular imaging, redox biology, and drug discovery/efficacy assessment, offering a new avenue for β₃-AR–targeted diagnosis and therapy in U251 GBM. The integration of GSHP imaging with β₃-AR–directed therapeutic strategies holds promise for future applications in precision diagnosis, treatment monitoring, and mechanistic exploration in glioma and broader biomedical fields.

MATERIALS AND METHODS

Experimental design

This study aimed to develop and validate a novel BBB-permeable dual-response fluorescent probe, GSHP, capable of simultaneously visualizing β₃-ARs in the brain and monitoring their surrounding redox microenvironment (particularly GSH levels) in real time. This design concept is based on the critical role of β₃-ARs and redox homeostasis in major brain diseases such as GBM. Existing technologies lack synchronous imaging capabilities and BBB permeability, hindering their study in relevant biological systems. Building upon our prior work on BBB-permeable probes and β₃-ARP, this study integrates a β₃-AR targeting motif (derived from carazolol) with a GSH-responsive fluorescent unit (cyano-activated alkene) onto a BBB-permeable quinoline fluorescent scaffold. The study design involved a systematic evaluation of probe performance: (i) in vitro spectroscopic characterization of GSHP; (ii) validation of its specificity and imaging capabilities for β₃-ARs and GSH in various cell lines (HEK293T subtypes and glioma cell lines); (iii) application of GSHP for high-throughput screening to identify natural modulators of β₃-ARs; (iv) identification and validation of baicalin as a natural β₃-AR inhibitor, along with its effects on glioma cell biology mechanisms, focusing on redox homeostasis, invasion, and apoptosis; and (v) translation of GSHP and lead compound into an orthotopic glioma mouse model for in vivo diagnosis, imaging of the tumor microenvironment, and assessment of therapeutic efficacy. Prespecified components included the probe’s dual-channel detection strategy (receptor binding and GSH response), reversible responsiveness to GSH, and the use of standard cell and animal models for validation and application.

Synthesis of probe GSHP

GSHP was synthesized via a one-step aldol condensation of quinolinic aldehyde (compound a3) and the active α-hydrogen site within the carazolol-derived warhead (compound c3) (fig. S1). To a stirred solution containing compound a3 (55 mg, 0.27 mmol) and c3 (100 mg, 0.27 mmol) in anhydrous ethanol (10 ml), catalytic piperidine was added under a nitrogen atmosphere. The reaction mixture was stirred at 60°C under a nitrogen atmosphere for 4 hours. After completion, the reaction was diluted with H₂O and extracted with CH₂Cl₂. The combined organic layers were dried over anhydrous Na₂SO₄, followed by filtration and concentration under reduced pressure. The crude product was purified by silica gel column chromatography (CH₂Cl₂/MeOH: 50/1) to yield GSHP as an orange solid (59.0 mg, 40% yield). The structure was characterized by ¹H nuclear magnetic resonance (NMR; 400 MHz, DMSO-d₆), detailed as follows: δ 11.25 (s, 1H), 8.44 (t, J = 5.6 Hz, 1H), 8.29–8.03 (m, 3H), 7.83 (dd, J = 10.0, 8.0 Hz, 2H), 7.56 (dd, J = 9.6, 2.8 Hz, 1H), 7.44 (d, J = 8.0 Hz, 1H), 7.30 (dt, J = 12.8, 7.6 Hz, 2H), 7.14 (t, J = 7.6 Hz, 1H), 7.06 (d, J = 8.0 Hz, 1H), 6.95 (d, J = 2.8 Hz, 1H), 6.70 (d, J = 8.0 Hz, 1H), 5.31 (s, 1H), 4.29–4.03 (m, 3H), 3.42 (s, 1H), 3.11 (s, 6H), 2.98 (d, J = 7.8 Hz, 1H), and 2.86 (t, J = 6.4 Hz, 3H).¹³C NMR (101 MHz, DMSO-d₆) results are as follows: δ 161.8, 155.4, 150.1, 149.5, 145.5, 141.6, 139.4, 134.4 (2C), 130.7, 126.9, 125.0, 123.9, 123.0, 122.2, 120.8, 119.0, 117.0, 112.0, 110.8, 107.2, 104.2, 100.9, 70.9, 68.7, 52.6, and 48.7. With electrospray ionization (ESI), the mass/charge ratio calculated for C₃₂H₃₂N₆O₃ is 549.2536 [M + H]⁺, and the actual value found was 549.2598. Synthesis of control probe GSHC and intermediates was also performed and is detailed in the Supplementary Materials.

In vitro spectroscopy measurements

Unless otherwise specified, all spectral measurements were performed in PBS (10 mM, pH 7.4) containing 0.1% (v/v) dimethyl sulfoxide (DMSO). The fluorescent probe stock solution (1 mM) was prepared in DMSO. All fluorescence measurements were recorded at 37°C. Standard excitation wavelengths were 395 or 470 nm, with slit widths set at 5 nm for both excitation and emission.

Viscosity response study

Fluorescence emission spectra of GSHP (10 μM) were recorded in PBS/glycerol mixtures with varying volume ratios to evaluate viscosity effects (Fig. 2B). The control probe, GSHC, was similarly tested (Fig. 2C).

GSH responsiveness and reversibility study

Fluorescence emission spectra of GSHP (10 μM) were recorded after incubation with different concentrations of GSH (0.1 to 30 mM) at 37°C for 30 min (Fig. 3C). Ultraviolet-visible absorption spectra were also measured upon addition of varying GSH concentrations (fig. S11). The ratiometric response (I_{525 nm}/I_{650 nm}) was calculated from fluorescence spectra using 470 nm excitation (Fig. 3E). Time-dependent fluorescence intensity changes at 525 nm were monitored after the addition of 10 mM GSH, followed by the addition of 500 μM NEM (fig. S16). The reversibility of the GSHP-GSH reaction was further assessed by monitoring the ratio of fluorescence intensities at F_{525 nm}/F_{650 nm} upon successive additions of GSH and NEM (fig. S17).

The pH stability study

The ratio of F_{525 nm}/F_{650 nm} after incubation of GSHP (10 μM) with GSH (10 mM) was measured in different pH buffers (pH 5.0 to 9.0) to determine pH dependence (fig. S14). For the selectivity test, fluorescence response of GSHP (10 μM) was measured after incubation with various biologically relevant analytes to assess selectivity toward GSH (fig. S13). Fluorescence spectra of GSHP (1 μM) or GSHC (1 μM) were measured in the presence of cytoplasmic or membrane protein extracts (0 to 5 μg/ml) from β₃-AR overexpressing HEK293 cells (Fig. 2, D to G).

Theoretical calculation

Molecular docking was performed using AutoDockTools (ADT) and AutoDock Vina. Ligand structures (GSHP, carazolol) were constructed and energy-minimized (Chem3D, MM2). The β₃-AR crystal structure [Protein Data Bank (PDB) ID: 7XJH] was prepared by removing water/ligands and adding hydrogens. The binding site was defined by a grid box centered on a secondary active site. Docking poses were ranked by binding energy, and the lowest-energy conformation was analyzed. Interactions were visualized using Schrödinger Maestro (Fig. 2I and fig. S3). DFT calculations were performed using Gaussian 09W. Geometry optimizations and single-point energy calculations for GSHP and GSHP-SG were carried out at the M06-2X/6-31G (d, p) level. Frontier molecular orbital energies, ΔE, and oscillator strengths were determined. Molecular orbital visuals were generated using GaussView 5.0 (Fig. 3D).

Cell culture methods

Unless otherwise specified, all cell culture reagents were obtained from Biochannel. All cells were cultured at 37°C in a humidified incubator with 5% CO₂.

HEK293T cells lines

Wild-type HEK293T cells and those stably transfected with β-AR subtypes (β₁, β₂, and β₃) were cultured in Dulbecco’s modified Eagle’s medium (DMEM; cat. no. BC-M-033) supplemented with 10% fetal bovine serum (FBS; cat. no. BC-SE-FBS07) and 1% penicillin-streptomycin (cat. no. BC-CE-007). The transfected cell lines were provided by M. Li.

Glioma and neuroblastoma cells lines

Human GBM cell lines U87, U251, LN229, and T98G and the human neuroblastoma cell line SH-SY5Y were all cultured in DMEM (cat. no. BC-M-033). This medium was supplemented with 10% FBS (cat. no. BC-SE-FBS07), 1 mM sodium pyruvate (cat. no. BC-P-012), 1% nonessential amino acids (cat. no. BC-CE-003), and 1% penicillin-streptomycin (cat. no. BC-CE-007). The LN229 and T98G cell lines were provided by X. Zhou.

Cellular confocal imaging

For fixed-cell imaging, β-AR–transfected HEK293T cells were seeded into 12-well plates with cover slides and cultured overnight. Cells were incubated with GSHP (100 nM) for 30 min at 37°C, either alone or in the presence of a β-AR antagonist (SR59230A for β₃-AR, metoprolol for β₁-AR, and ICI118551 for β₂-AR). Following treatment, cells were washed twice with PBS and fixed with 4% paraformaldehyde (PFA) in PBS for 10 min at room temperature. Fixed cells were washed and imaged using a Leica TCS SP8 MP confocal laser scanning microscope. Imaging parameters were λ_ex = 488 nm and λ_em = 620 to 680 nm (red channel). For real-time dynamic imaging, β₃-AR–transfected HEK293T cells or U251 cells were seeded onto confocal dishes. For U251 cells, treatments included baicalin (25 μg/ml) for 0.5 or 12 hours, or baicalin + NAC (5 mM) for 12 hours. Cells were incubated with GSHP (100 nM, with 1% DMSO in Hanks’ balanced salt solution) for 30 min at 37°C before imaging. Confocal images were acquired using a Leica TCS SP8 MP microscope with dual channels: green channel (GSHP-SG, λ_ex = 405 nm and λ_em = 490 to 550 nm) and red channel (GSHP, λ_ex = 488 nm and λ_em = 620 to 680 nm). Microscope settings were kept consistent. For dynamic redox monitoring, cells were treated sequentially with 2 mM Glu, 100 μM GSH, and 500 μM NEM while imaging. Local GSH level fluctuations surrounding β₃-AR were quantitatively analyzed by calculating the fluorescence ratio (F_{488 nm}/F_{405 nm}).

High-throughput screening

A high-throughput screening assay was performed using a high-content imaging system to identify natural products that competitively inhibit GSHP binding to β₃-ARs. β₃-AR–transfected HEK293T cells were cultured in 96-well plates. Cells were pretreated with 20 μM of various natural products (structures shown in Fig. 4A) for 1 hour at 37°C. Following pretreatment, cells were washed and incubated with GSHP (0.1 μM) for 30 min. Fluorescence images were acquired using a high-content imaging system with dual excitation/emission settings: 405 nm/500 to 550 nm for GSHP-SG (green channel) and 488 nm/640 to 670 nm for GSHP (red channel). Red fluorescence intensity (GSHP signal) was quantified for each well and normalized to the DMSO-treated control group (set as 1).

Transwell invasion assay

The Transwell invasion assay used 24-well Transwell chambers (8.0-μm pore size, Corning, USA), precoated with Matrigel (BD Biosciences, USA). Human GBM and neuroblastoma cells were pretreated with baicalin (0, 25, and 50 μg/ml) for 24 hours. Then, 2 × 10⁵ cells suspended in serum-free medium containing the corresponding concentrations of baicalin were seeded into the upper chamber. The lower chamber contained medium supplemented with 10% FBS as a chemoattractant. After 24 hours of incubation at 37°C, noninvasive cells on the upper surface were removed. Invaded cells on the lower surface were fixed with 4% PFA for 15 min and stained with 0.1% crystal violet for 15 min. Images were captured using an inverted microscope (Olympus IX71, Japan) at 100× magnification. Five random fields per well were quantified. Experiments were performed in triplicate (Fig. 5G).

Real-time qPCR

Approximately 2 × 10⁵ cells were seeded into 60-mm dishes and subjected to the different treatments (baicalin 25 or 50 μg/ml for 24 hours). Total RNA was extracted using TRIzol reagent (Ambion). Reverse transcription (RT) was performed using a commercial kit (Vazyme Biotech, cat. no. R323-01) to synthesize cDNA. RT-qPCR was conducted using a SYBR Green Master Mix kit (Vazyme Biotech, cat. no. Q711) according to the manufacturer’s protocol. Primers used were glyceraldehyde-3-phosphate dehydrogenase (GAPDH; internal control): forward 5′-GTCTCCTCTGACTTCAAC AGCG-3′, reverse 5′-ACCACCCTGTTGCTGTAGCCAA-3′; and β₃-AR: forward 5′- GCTCGACGGGGCTTCTT-3′, reverse 5′-TCTGAACAGAGGCCAGAGGT-3′. Relative expression levels were normalized to GAPDH.

Patient primary specimen collection

Human glioma tissue samples of varying grades (including GBM and astrocytoma) were obtained from treatment-naïve patients undergoing surgical resection at Linyi Central Hospital. Immediately following resection, fresh tumor specimens were snap-frozen in liquid nitrogen or embedded in optimal cutting temperature compound and stored at −80°C until sectioning. The detailed clinicopathological characteristics, including WHO grade and IDH mutation status of the samples used in this study, are summarized in table S2. All procedures involving human subjects were approved by the Medical Ethics Committee of Linyi Central Hospital (approval no. ME-LYCH-2026050) and were conducted in strict accordance with the ethical standards of the institutional review board and with the 1964 Helsinki declaration and its later amendments. Written informed consent was obtained from all patients prior to surgery.

IF staining

For β₃-AR staining (cells and tissue) (Fig. 2K), β₃-AR–transfected HEK293T cells incubated with GSHP (100 nM) were fixed with 4% PFA, permeabilized, blocked, and incubated with a primary antibody targeting β₃-AR (anti-ADRB3 antibody, Abcam, ab300483; dilution not specified for cell IF, but 1:500 for tissue IF). After washing, cells were incubated with a secondary antibody [Alexa Fluor CY3-conjugated Goat Anti-Rabbit IgG (H + L), Yeasen, cat. no. 33108ES60; dilution not specified for cell IF, but 1:200 for tissue IF]. Confocal imaging was performed, and colocalization with GSHP was analyzed (Pearson correlation coefficient reported). For tissue staining (Fig. 3J), frozen GBM tissue sections (collected with patient informed consent and ethical approval from Linyi Central Hospital) were processed similarly: equilibrated, fixed (4% PFA, 10 min), washed (PBS), permeabilized (0.3% Triton X-100, 10 to 30 min), blocked (3% bovine serum albumin in PBS, 30 min), incubated overnight at 4°C with primary β₃-AR antibody (1:500), washed (0.1% Tween 20 in PBS), incubated with CY3-conjugated secondary antibody (1:200) for 1 hour at 37°C in the dark, washed, and mounted with anti-fade medium. Confocal imaging (Leica TCS SP8 MP) was performed (CY3 channel: λ_ex = 561 nm and λ_em = 580 to 620 nm). Tissue sections stained with GSHP were imaged simultaneously (GSHP red/green channels). For TUNEL assay (apoptosis), U251 cells were cultured, treated with baicalin (25 μg/ml) for indicated times (3, 6, 9, and 12 hours), fixed (4% PFA, 30 min), permeabilized (0.1% Triton X-100 in 0.1% sodium citrate, 5 min on ice), and incubated with the TUNEL reaction mixture. Nuclei with fragmented DNA emitted green fluorescence and were visualized by confocal microscopy (Fig. 5D).

Western blotting

Protein (20 to 30 μg) was separated by 8 or 10% SDS–polyacrylamide gel electrophoresis (80 V stacking, 120 V separation). Proteins were transferred to polyvinylidene difluoride (Bio-Rad wet transfer, 200 mA, 60 to 90 min) in transfer buffer (25 mM tris, 192 mM glycine, and 20% MeOH, pH 8.3). Membranes were blocked (5% nonfat milk in PBST, 1 hour at room temperature). These were incubated with primary antibodies [1:1000 dilution unless noted: ADRB3 (Abcam, ab300483), p-ERK1/2 (Abcam, ab76299), ERK1/2 (Abcam, ab184699), nNrf2 (Abcam, ab137550), GCLc (Abcam, ab190685), nuclear factor κB (NF-κB) p-p65 (Abcam, ab86299), NF-κB p65 (Abcam, ab16502), cleaved caspase-3 (CST, no. 9664), BAX (Abcam, ab32503), p-CREB (Abcam, ab32096), CREB (Abcam, ab32515), Lamin B (Abcam, ab16048, 1:2000), Flotillin-1 (CST, no. 18634), Flotillin-2 (CST, no. 3436), GAPDH rabbit polyclonal antibody (ABclonal, AC001, 1:20,000), and β-actin mouse monoclonal antibody (Proteintech, 66009-1-Ig, 1:5000)] in 5% milk/PBST (1.5 hours at room temperature or overnight at 4°C). These were washed (PBST, three times for 10 min) and then incubated with horseradish peroxidase–conjugated secondary antibodies (Proteintech, SA00001-2 or SA00001-1, 1:5000 to 1:10,000 in 5% milk/PBST, 1 hour at room temperature). Bands were detected using enhanced chemiluminescence (Beyotime) and imaged (Tanon 5200 or Bio-Rad ChemiDoc MP). Band intensity was quantified with ImageJ and normalized to loading controls (GAPDH and Lamin B).

Establishment of orthotopic human GBM xenograft model

Orthotopic human GBM xenografts were established in female Balb/c nude mice [6 to 8 weeks old, obtained from SiPeiFu (Beijing) Biotechnology Co., Ltd.]. U251 cells (2 × 10⁵ cells in 5 μl of sterile PBS) were stereotaxically implanted into the right striatum of anesthetized mice (2% isoflurane). A burr hole was drilled at coordinates 1.5 mm anterior and 2.5 mm lateral to the bregma. The microsyringe was lowered to 4 mm ventral from the skull surface and then retracted to 3.5 mm for cell delivery at a rate of 0.25 μl/min. The microsyringe was slowly withdrawn after injection. The burr hole was sealed with bone wax, and the skin incision closed with surgical adhesive (Vetbond). Postoperative pain management included subcutaneous injection of meloxicam (1 mg/kg). The model is reported to recapitulate key histopathological characteristics of human GBM. All animal care and experimental procedures were performed in accordance with the protocols approved by the Institutional Animal Care and Use Committee (IACUC) of Nanjing Normal University (approval no. IACUC-2024270) and complied with the national guidelines for the care and use of laboratory animals.

In vivo imaging and MRI

MRI data were acquired using a 9.4 T Bruker BioSpec 94/20 USR MRI scanner. Radiofrequency excitation was via an 86-mm volume coil, and signal reception was with a surface coil. Mice were anesthetized (medetomidine + isoflurane) and positioned in a stereotaxic frame. For in vivo fluorescence imaging, GSHP was administered intravenously. In vivo fluorescence imaging of GSHP-SG (green) and GSHP (red) signals was performed on normal and U251 tumor-bearing mice at specified time points postinjection. Imaging was also performed on treatment groups (PBS, TMZ, and baicalin) at day 20 post–treatment initiation (Fig. 6G). Ex vivo imaging of brain tissues was performed after sacrifice (Fig. 6, B, J, and M). The in vivo imaging was performed using an IVIS Lumina XR imaging system (PerkinElmer, Nanjing University). For ratiometric fluorescence detection, images were acquired sequentially using two distinct filter sets. The green channel (GSHP-SG adduct) was captured with an excitation of 430 nm and an emission filter of 500 to 600 nm. The red channel (GSHP) was captured with an excitation of 500 nm and an emission filter of 600 to 680 nm. The system’s standard high-intensity illumination setting was used for excitation power in all experiments.

Statistical analysis

All quantitative data are presented as means ± standard error of the mean (SEM) from at least three independent biological replicates or animals, as specified in the figure legends. For tissue quantification, data from each subject (animal or human) were averaged from at least three nonoverlapping fields of view across three separate sections. Statistical analyses were performed using GraphPad Prism 9.0. A two-tailed unpaired Student’s t test was used for two-group comparisons. One-way or two-way analysis of variance (ANOVA), followed by Tukey’s or Sidak’s multiple comparisons test, respectively, was used for multigroup analyses. Survival curves were analyzed using the log-rank (Mantel-Cox) test. A P value of <0.05 was considered statistically significant. Significance is denoted as ns (not significant), *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Supplementary Materials

This PDF file includes:

Figs. S1 to S32

Tables S1 to S3

Supplementary Text

NMR and ESI-MS Spectra