Perilipin‐2‐Anchored Lipid Droplet‐Targeting Phototheranostics

ABSTRACT

Conventional photothermal agents are readily taken up by normal cells, which may cause off‐target damage. Lipid droplet (LD) targeting can enable photothermal agents to accumulate in tumor cells enriched in LDs, a central hub for lipid storage. We herein report an LD‐targeted theranostic agent, C‐BDP‐OMe, through a dual‐anchoring strategy that combines the principle of “like dissolves like” with specific docking to the PLIN2 protein. Structurally, C‐BDP‐OMe features an extended conjugated system and an asymmetric coumarin‐fused BODIPY scaffold, enabling NIR‐region optical properties, a large Stokes shift, and high photostability. The probe exhibits polarity‐sensitive behavior, with turn‐on fluorescence and a high signal‐to‐noise ratio under no‐wash conditions. Furthermore, C‐BDP‐OMe demonstrates robust photothermal performance, with a photothermal conversion efficiency of 62.6%, enabling localized photothermal ablation and resulting in effective tumor suppression while avoiding systemic side effects. The therapeutic mechanism was systematically investigated and shown to involve the concurrent activation of two distinct cell death pathways—apoptosis and ferroptosis. This work highlights the therapeutic value of subcellular organelle‐level precision interventions and underscores the importance of exploiting subtle microenvironmental features in the design of next‐generation anticancer theranostic agents.

This study reports two NIR theranostic dyes, C‐BDP‐OH and C‐BDP‐OMe. C‐BDP‐OMe targets lipid droplets via a dual‐anchoring mechanism with PLIN2, enabling wash‐free tumor imaging. With 62.6% photothermal conversion efficiency, it achieves tumor ablation by downregulating PLIN2, disrupting LD‐mitochondria crosstalk, and triggering apoptosis and ferroptosis.

1. Introduction

Lipid droplets (LDs) are dynamic organelles ubiquitously present in eukaryotic cells, serving not only as central hubs for intracellular neutral lipid storage but also as essential regulators of energy metabolism, membrane biosynthesis, and redox homeostasis [1, 2, 3]. Accumulating evidence indicates that LDs are not inert lipid reservoirs but rather highly dynamic facilities that undergo profound morphological and functional remodeling under various pathological conditions [4, 5]. As recent studies have uncovered how LDs mediate fatty acid translocation from storage to oxidation sites, the biological importance of LDs has attracted increasing attention [6]. In the tumor microenvironment (TME), cancer cells primarily rely on glycolysis, as described by the Warburg effect, leading to elevated acetyl‐CoA levels and enhanced de novo fatty acid synthesis [7, 8]. This metabolic shift drives robust LD biogenesis, characterized by increased numbers and sizes of LDs and extensive remodeling of their surface proteome [9, 10, 11, 12, 13, 14]. Perilipin 2 (PLIN2), a signature protein localized to the LD membrane, is one of the proteins specifically overexpressed in multiple malignant tumors [15, 16]. It supports sustained proliferation and survival of tumor cells by suppressing lipolysis, thereby shielding LDs from catabolic degradation under nutrient‐deprived, oxidative‐stress conditions [17, 18]. This tumor‐associated accumulation makes PLIN2‐enriched LDs not only biomarkers of tumor metabolic reprogramming but also promising targets for cancer theranostics.

In this context, numerous fluorescent probes have been developed to visualize LDs and monitor distinctive metabolic processes in tumor cells compared with those in normal cells [19, 20, 21, 22]. While classical dyes such as BODIPY 493/503 and Nile Red are commercially available and widely utilized, their practical utility in biological imaging remains limited by several intrinsic drawbacks (Scheme 1A) [23]. Nile Red lacks inherent specificity to LDs and exhibits non‐selective accumulation in hydrophobic cellular environments, leading to off‐target labeling. Meanwhile, BODIPY 493/503 is constrained by its short excitation wavelength and minimal Stokes shift, rendering it highly susceptible to interference from endogenous autofluorescence and unsuitable for deep‐tissue imaging [24, 25, 26, 27, 28]. In general, the LD‐targeting strategy focuses on lipophilic accumulation within LDs, while microenvironment‐sensitive fluorescence provides an additional advantage for imaging contrast. Moreover, for red‐/NIR‐region optical responses, structural tuning is a common approach, such as extending π‐conjugation or incorporating donor‐rich aromatic units into the fluorophore scaffold [6, 22]. Although these modifications could shift the absorption maximum to longer wavelengths, they may render the structure more vulnerable to chemical degradation by reactive oxygen and sulfur species in the TME, undermining the accuracy and reproducibility of long‐term imaging [29]. Therefore, developing LD‐targeted probes that simultaneously achieve high targeting precision, a favorable long‐wavelength response, and robust structural stability remains highly desirable.

SCHEME 1.

Design strategy of LD‐targeted theranostics and its application in bioimaging and PTT.

From a therapeutic perspective, LD‐targeted photodynamic therapy (PDT) has shown promise in multiple studies, primarily by light‐triggered ROS generation to induce tumor cell death [6, 30, 31]. However, the hypoxic microenvironment, commonly present in solid tumors, significantly diminishes the effectiveness of PDT, which greatly limits its clinical translatability [32, 33, 34, 35, 36]. In contrast, photothermal therapy (PTT) achieves tumor ablation by generating localized hyperthermia upon light irradiation, leading to protein denaturation and membrane disruption [37, 38, 39]. Recent studies further suggest that PTT‐induced hyperthermia could accelerate lipid peroxidation (LPO) by promoting Fenton/Fenton‐like reactions, thereby enhancing ferroptosis [40, 41]. These oxygen‐independent mechanisms allow highly efficient tumor eradication even under hypoxic conditions, offering superior therapeutic robustness. Notably, when PTT agents can specifically target LDs, they could potentially influence multiple cell death pathways by disrupting a central hub of lipid metabolism, directly impairing the energy supply and antioxidant defense systems of cancer cells [39, 42]. Despite these advantages, reports of LD‐targeted PTT that combine high‐fidelity imaging guidance and efficient photothermal conversion remain scarce to date.

To this end, we report herein a polarity‐responsive coumarin‐fused BODIPY series that exhibits high LD selectivity, NIR optical properties, and photothermal conversion via a fine‐tuned structure. Through the strategic incorporation of electron‐donating substituents—methoxy and hydroxyl groups—the molar extinction coefficient in the NIR region was increased, along with enlarged Stokes shift and improved photostability, enabling the rational design and synthesis of two novel probes, C‐BDP‐OH and C‐BDP‐OMe. Interestingly, C‐BDP‐OMe demonstrated outstanding LD selectivity, likely arising from the combined effects of the “like dissolves like” principle and favorable PLIN2‐associated interactions. Although PLIN2 is widely recognized as a hallmark LD‐associated protein in tumor cells, its potential as a molecular recognition handle for small‐molecule theranostic agents remains relatively underexplored [43, 44, 45]. Motivated by these possibilities, we investigated PLIN2‐associated interactions using docking studies in conjunction with in vitro imaging analyses. These overall characteristics identified C‐BDP‐OMe as the optimal candidate for LD imaging, owing to its superior photophysical properties, including turn‐on fluorescence response to the LD environment. Moreover, the asymmetric molecular architecture with freely rotatable functional moieties—diethylamino and p‐methoxyphenyl groups—allows non‐radiative decay pathways, rendering it an effective mild photothermal agent. Such characteristics enable C‐BDP‐OMe to serve as a theranostic agent, and the following in vitro and in vivo results demonstrate its effectiveness for precise imaging and therapeutic applications. Collectively, this work establishes a novel design paradigm for microenvironment‐responsive, LD‐targeted theranostic agents and provides a versatile technical platform for integrated tumor imaging and photothermal therapy.

2. Results and Discussion

2.1. Synthesis of the C‐BDP Series Dyes

C‐BDP‐H was prepared according to a previously reported procedure [46]. C‐BDP‐OMe was synthesized using 4‐chloro‐7‐diethylaminocoumarin‐3‐aldehyde as the starting material. First, a coumarin‐fused pyrrole intermediate bearing a p‐methoxybenzoyl substituent was prepared via a sequence of transformations, including a Wittig reaction, a nucleophilic substitution with an azide, and an intramolecular cyclization. This intermediate was subsequently demethylated with BBr₃, yielding the corresponding coumarin‐fused pyrrole intermediate bearing a p‐hydroxybenzoyl substituent. Next, each of these intermediates underwent condensation with 2,4‐dimethylpyrrole under POCl₃ mediation to form asymmetric dipyrromethene frameworks. Finally, these frameworks were complexed with BF₃·OEt₂ in the presence of triethylamine to afford the target compounds, C‐BDP‐OMe and C‐BDP‐OH (Scheme S1). The structures of all C‐BDP dyes and their key intermediates were systematically characterized by ¹H NMR, ¹³C NMR, and high‐resolution mass spectrometry.

2.2. Design Rationale of the C‐BDP Series

Although C‐BDP‐H exhibits excellent photophysical properties as a fluorescent probe, its biological applications have not been systematically investigated in prior studies. Based on its polarity‐responsive behavior and favorable lipophilicity, we hypothesized that this compound may hold promise for LD‐targeted imaging, a hypothesis subsequently confirmed by LDs co‐localization experiments (Figure S1). After confirmation, we further designed molecularly engineered derivatives to exhibit altered electronic architectures. In the rational design of theranostic probes, the molar extinction coefficient (ε) represents a critical parameter governing light energy absorption efficiency [47]. Maximizing ε within the near‐infrared biological window not only improves the sensitivity of optical imaging but also provides a structural basis for effective phototherapy with low‐dose treatment regimens. Given that molecular photoelectronic properties primarily depend on chemical structure, strong electron‐donating groups (e.g., ─OMe, ─OH) were introduced to the central benzene ring of the C‐BDP scaffold, thereby constructing a dual‐chromophore centered donor‐acceptor‐donor (D‐A‐D′) structure (Figure 1A).

FIGURE 1.

(A) Chemical structures of C‐BDP‐H, C‐BDP‐OH, and C‐BDP‐OMe; (B–D) UV–Vis absorption spectra of the C‐BDP series dyes (10 µM) in 1,4‐dioxane/water mixed solvents with varying water volume fractions; (E–G) corresponding fluorescence emission spectra. λ _ex = 600 nm.

2.3. Spectral Characteristics of C‐BDP Series

Building upon the pronounced solvent polarity‐dependent behavior of C‐BDP‐H, this study systematically investigated the optical response properties of C‐BDP‐OH and C‐BDP‐OMe across solvents with varying polarities. The photophysical properties of these dyes were evaluated in low‐polarity solvents (1,4‐dioxane, THF, acetone) and high‐polarity solvents (CH₃CN, DMF, DMSO), and their UV–vis absorption and fluorescence emission spectra were analyzed to elucidate their sensitivity to microenvironmental polarity. Although the absorption maxima showed minimal dependence on solvent polarity, fluorescence intensity decreased progressively with increasing solvent polarity, accompanied by a significant redshift in the emission wavelength (Figure S2). These findings demonstrate that the luminescence characteristics of this dye series are highly sensitive to local environmental polarity.

To further evaluate the potential applicability of the C‐BDP series of dyes for polarity sensing, a gradient‐polarity model was established using a 1,4‐dioxane/water binary solvent system [48]. Spectral analyses revealed that with increasing solvent polarity, the UV‐Vis absorption spectra of these dyes exhibited a progressive blue shift (Figure 1B–D), while their fluorescence emission displayed pronounced polarity dependence: as water content increased from 0% to 50%, fluorescence intensity gradually decreased, and the maximum emission wavelength underwent a significant red shift (Figure 1E–G). This behavior is highly consistent with the polarity‐responsive trends previously observed in single‐solvent systems, further supporting the feasibility of employing C‐BDP series dyes as fluorescent probes. To quantitatively elucidate the relationship between fluorescence response and solvent polarity, the Lippert‐Mataga polarity function (Δf) was adopted as a theoretical descriptor [49]. Linear regression analysis of experimental data yielded a correlation model relating maximum fluorescence intensity (F _max) to Δf. Both C‐BDP‐OMe and C‐BDP‐OH exhibited strong linear correlations (R ² > 0.98), indicating robust and reproducible response characteristics (Figure S3). These results demonstrate that the C‐BDP series dyes exhibit excellent quantitative sensing performance, highlighting their potential for real‐time, reliable monitoring of dynamic polarity changes in biological microenvironments.

Furthermore, a comparative analysis of the dyes’ optical properties was performed in the nonpolar solvent 1,4‐dioxane, which serves as a useful nonpolar model for the LD microenvironment. As summarized in Table 1, the introduction of electron‐donating substituents at the central phenyl ring led to a gradual blue shift in both absorption and emission maxima. Meanwhile, owing to the asymmetric fused‐ring derivatization strategy employed during synthesis, all dyes in this series show large Stokes shifts, averaging about 88 nm—significantly greater than that of the commercial LD stain BODIPY 493/503 (∼10 nm). Notably, both the molar extinction coefficient and fluorescence quantum yield increased upon EDG substitution.

TABLE 1.

Photophysical properties of the C‐BDP series dyes in 1,4‐dioxane.

	Abs. (nm)	FL. (nm)	Φ (%)	ε (104 M−1 cm−1)	Stokes shift (nm)
C‐BDP‐H	616	712	52.14	0.25	96
C‐BDP‐OH	612	693	66.71	0.67	81
C‐BDP‐OMe	613	697	70.99	0.47	84

Φ represents the absolute quantum yield in 1,4‐dioxane.

To better understand these spectral trends, DFT calculations were performed for each compound (Figure 2A). The results indicated a slightly enlarged HOMO‐LUMO gap in the presence of EDG, consistent with the blue‐shift observed in spectral measurements. Upon photoexcitation, the dimethylamino group within the coumarin moiety serves as the primary donor (D), while the substituent at the phenyl ring serves as an auxiliary donor (D′), together establishing a D–A–D architecture with the coumarin‐fused BODIPY core as the acceptor (A). Such an electronic arrangement promotes intramolecular charge transfer (ICT) between the two peripheral donors and the acceptor, thereby influencing the transition dipole moment, consistent with the experimentally observed increase in the molar extinction coefficient. In addition, the reduced dihedral angle in the excited state suggests partial planarization of the EDG‐substituted framework, which may help suppress torsion‐assisted non‐radiative decay, thereby enhancing radiative transition efficiency (Figure 2B).

FIGURE 2.

(A) HOMO and LUMO energy levels (E _HOMO and E _LUMO) and energy gap (ΔE, ΔE = E _LUMO—E _HOMO). (B) The dihedral angle of the median benzene ring relative to the parent nucleus in the ground state and excited state.

To further evaluate the selectivity and anti‐interference performance of C‐BDP‐OMe and C‐BDP‐OH in polarity sensing, the fluorescence response behaviors of these dyes were systematically examined under varying viscosity, pH conditions, and in the presence of potential interfering substances. In the viscosity interference study, MeOH and THF—solvents with comparable viscosities but markedly different polarities—were employed as comparative media. The results revealed that both dyes exhibited significantly higher fluorescence intensity in THF than in MeOH, demonstrating that their fluorescence response is minimally affected by solvent viscosity and primarily governed by solvent polarity (Figure S4). In pH interference experiments, the dyes displayed consistently low fluorescence intensities in aqueous solutions across the pH range of 2–10; in contrast, a pronounced fluorescence enhancement was observed in a 1,4‐dioxane system containing 10% water, confirming their insensitivity to pH variations (Figure S5). To assess detection selectivity, various common interferents—including cations, anions, amino acids, and reactive species—were individually introduced into PBS for evaluation. No significant changes in fluorescence intensity were detected upon addition of these species, whereas a strong fluorescence signal emerged exclusively in the low‐polarity environment of 1,4‐dioxane (Figure S6). These findings collectively demonstrate that C‐BDP‐OMe and C‐BDP‐OH exhibit excellent selectivity and robust resistance to interferences during polarity recognition, making them well suited for the specific and reliable detection of polarity changes in complex biological or chemical environments.

2.4. In Vitro Validation of C‐BDP Series as a LD‐Specific Probe

The biological safety of the C‐BDP series is a fundamental prerequisite for their application in biological systems. To evaluate the cytotoxicity of the C‐BDP series dyes, cell viability was assessed using the CCK‐8 assay prior to imaging experiments. Results demonstrated that even at a concentration of 30 µM and after 12 h of incubation, the survival rate of HeLa cells remained above 80%, indicating minimal dark toxicity for this dye series (Figure S7). To gain mechanistic insights into the role of lipophilicity in subcellular organelle targeting, theoretical calculations of the octanol‐water partition coefficients (log P) were performed using ChemDraw 20.0. The analysis revealed that the C‐BDP series exhibits higher lipophilicity compared to commercially available LDs stains, including BODIPY 493/503 and Nile Red (Figure 3A). Subsequent co‐localization studies were conducted to evaluate their LDs targeting capability in HeLa cells. Despite uniformly high lipophilicity across the series, distinct differences in subcellular localization were observed: C‐BDP‐OMe displayed characteristic punctate red fluorescence and exhibited strong co‐localization with BODIPY 493/503, yielding a Pearson correlation coefficient of 0.83. In contrast, C‐BDP‐OH, although similarly lipophilic, failed to achieve specific LDs labeling, showing a much lower Pearson coefficient of 0.22 (Figure 3B). These findings indicate that while high lipophilicity facilitates initial accumulation in lipid‐rich regions, it alone is insufficient to ensure selective targeting of LDs. Additional molecular factors must act in concert to govern precise subcellular localization.

FIGURE 3.

(A) Calculated octanol‐water partition coefficients (Clog P) of BODIPY 493/503, Nile Red, C‐BDP‐H, C‐BDP‐OH, and C‐BDP‐OMe, along with predicted binding free energies to the LDs marker protein PLIN2. (B) Fluorescence co‐localization imaging analysis of C‐BDP‐OH and C‐BDP‐OMe with BODIPY 493/503 in HeLa cells. (C) Fluorescence co‐localization imaging of C‐BDP‐OMe with Mito‐Tracker Green, ER‐Tracker Green and Lyso‐Tracker Green in HeLa cells. (D) Confocal microscopy images of C‐BDP‐OMe (10 µM) and BODIPY 493/503 (10 µM) under unwashed and washed conditions, and (E) corresponding signal‐to‐noise ratio (S/N) statistics (n = 3). (F) Photostability comparison between C‐BDP‐OMe and BODIPY 493/503, and (G) fluorescence intensity decay profiles during continuous laser excitation under confocal microscopy. (H) Time‐lapse confocal imaging of C‐BDP‐OMe‐labeled LDs in living HeLa cells and their pseudo‐colored maximum‐intensity projection overlays. Excitation and emission wavelengths: green channel, λ _ex = 488 nm, λ _em = 500–540 nm; red channel, λ _ex = 633 nm, λ _em = 680–740 nm; (I) Schematic representation of molecular docking between C‐BDP‐OMe and PLIN2 (UniProt ID: Q9BYF1). (J) Three‐dimensional structural details of the interactions between C‐BDP‐OMe and neighboring amino acid residues within the PLIN2 binding pocket.

Through systematic co‐localization experiments, the specific targeting capability of C‐BDP‐OMe toward LDs in HeLa cells was further evaluated. Cells were co‐incubated with C‐BDP‐OMe and a panel of commercial organelle‐specific fluorescent probes, including Mito‐Tracker Green, Lyso‐Tracker Green, and ER‐Tracker Green. As shown in Figure 3C, the red fluorescence signal from C‐BDP‐OMe exhibited minimal overlap with the green fluorescence signals of the aforementioned organelle markers. Quantitative co‐localization analysis revealed no significant spatial co‐distribution between C‐BDP‐OMe and the respective control organelles, providing strong evidence for its high selectivity toward LDs. In contrast to conventional LDs dyes, C‐BDP‐OMe exhibits negligible fluorescence emissions in highly polar environments and undergoes selective fluorescence activation exclusively within the low‐polarity microenvironment of LDs. This property enables high‐precision imaging without the need for washing steps (Figure 3D). Signal‐to‐noise ratio (S/N) analysis further demonstrated that C‐BDP‐OMe maintains a high S/N even under no‐wash conditions, thereby enhancing spatial resolution in biological imaging (Figure 3E). Furthermore, the fluorescence emission properties of C‐BDP‐OMe in LD‐mimicking systems provide additional spectroscopic evidence (Figure S8). Moreover, photostability assessments revealed that C‐BDP‐OMe retains stable fluorescence intensity under continuous laser irradiation, whereas the commercial dye BODIPY 493/503 displays pronounced photobleaching (Figure 3F,G). The superior photostability of C‐BDP‐OMe is likely attributable to its highly conjugated and relatively rigid framework, which effectively suppresses photo‐induced degradation pathways, such as double‐bond isomerization or cleavage. This excellent photostability was further corroborated in 1,4‐dioxane, a LD‐mimicking low‐polarity solvent (Figure S9). These properties suggest that C‐BDP‐OMe is well‐suited for real‐time monitoring of LD dynamics in living systems.

To directly assess its utility, real‐time tracking of LD dynamics was performed in living cells using C‐BDP‐OMe. Fluorescence images were acquired at 1‐min intervals, and signals from each time point were encoded with distinct pseudo‐colors. Subsequent superimposition of multiple frames generated composite images visualizing the movement trajectories of LDs. The resulting merged images clearly illustrate the spatial distribution and trajectories of LDs within the cell (Figure 3H). These results demonstrate that C‐BDP‐OMe enables efficient, stable monitoring of LD dynamics at the cellular level, highlighting its potential for investigating the physiological functions of LDs and their pathological relevance to disease progression.

The LD surface contains various proteins, including PLIN2, and serves as a critical interface for molecular recognition and information exchange between the LDs and their surrounding cellular environment [50]. Given that a subtle functional‐group modification led to a marked difference in LD‐targeting ability, we hypothesize that C‐BDP‐OMe may engage in specific interactions with LD‐associated surface proteins. To elucidate the molecular basis of its efficient targeting, molecular docking simulations were performed using AutoDock Vina with PLIN2, a key structural protein associated with LDs. The results revealed a binding free energy of −8.6 kcal/mol for C‐BDP‐OMe, whereas C‐BDP‐OH exhibited a significantly weaker interaction with a binding free energy of only −3.8 kcal/mol.

This substantial difference indicates that the interaction between C‐BDP‐OMe and PLIN2 is thermodynamically favorable (Figure 3A). Detailed analysis of the binding site identified polar residues (e.g., ASN51, THR347), non‐polar residues (e.g., TRP271, VAL343), and ionizable residues (e.g., ARG273, HIS345) in close proximity to the ligand‐binding pocket. Notably, the ionizable residue ARG273 forms a non‐conventional C─H···O hydrogen bond with the ether oxygen atom of C‐BDP‐OMe. Additional non‐covalent interactions were also identified: electrostatic contributions include a cation–π interaction between ARG273 and the central benzene ring of C‐BDP‐OMe, as well as an anion–π interaction between GLU145 and the five‐membered ring of the dye. Hydrophobic interactions further stabilize the complex and include (1) π–π stacking between TRP271 and the aromatic ring of C‐BDP‐OMe, (2) π–π stacking between PHE504 and the two pyrrole rings, (3) alkyl–alkyl interactions involving VAL343 and the ethyl group, and (4) π–alkyl interactions between TRP349 and the ethyl moiety, as well as between TRP271/PHE274 and the ether‐linked carbon atoms of C‐BDP‐OMe (Figure 3I). To provide a comprehensive visualization of these intermolecular interactions, a detailed 3D interaction map was generated (Figure 3J).

Taken together, C‐BDP‐OMe was selected as the lead molecule for further studies based on its superior LD‐target ability, potentially associated with its favorable interaction with PLIN2, and its promising real‐time imaging performance.

2.5. Biological Imaging Applications in Cellular Systems

To further elucidate the correlation between intracellular LDs content and fluorescence intensity, oleic acid (OA) was employed to induce intracellular LDs accumulation [51, 52]. In OA‐pretreated HeLa cells (Figure 4A), the fluorescence intensity of C‐BDP‐OMe increased gradually with increasing OA concentration, confirming the high sensitivity and detection efficiency of C‐BDP‐OMe in monitoring LDs accumulation. Moreover, in starvation experiments, the fluorescence intensity of C‐BDP‐OMe decreased as LDs degraded, demonstrating that C‐BDP‐OMe can not only sensitively detect LD formation and accumulation but also accurately reflect LD dynamics during the regulation of energy metabolism (Figure 4B).

FIGURE 4.

(A) Schematic diagram of the experimental process for stimulating HeLa cells with OA gradient concentration. (B) Confocal fluorescence images of C‐BDP‐OMe in HeLa cells cultured under different concentrations of OA and nutrient deprivation conditions. Green channel: λ _ex = 488 nm, λ _em = 500–540 nm; Red channel: λ _ex = 633 nm, λ _em = 680–740 nm. (C) Schematic illustration of the microenvironmental differences between normal and tumor tissues. (D) Confocal fluorescence imaging comparison of C‐BDP‐OMe in normal versus tumor cells.

Numerous studies have demonstrated that tumor cells exhibit elevated LD levels and decreased cellular polarity compared to normal cells (Figure 4C) [53]. Based on these characteristics, we sought to evaluate the potential of C‐BDP‐OMe as a tumor imaging agent by leveraging differences in LD quantity and polarity between tumor and normal cells. Tumor cells (HeLa, HCT‐116, and HepG2) and normal cells (HaCat and 7702) were incubated with C‐BDP‐OMe (10 µM) for 10 min, followed by fluorescence imaging in the red channel. The fluorescence intensity in tumor cells treated with C‐BDP‐OMe was markedly higher than that in normal cells (Figure 4D). These results demonstrate that C‐BDP‐OMe can effectively distinguish tumor cells from normal cells by recognizing differences in LD quantity and polarity.

2.6. Photothermal Properties of C‐BDP‐OMe

The molecular design of C‐BDP‐OMe is centered on a coumarin‐fused BODIPY core functionalized with an N,N‐diethylamino group and a p‐methoxyphenyl rotor tethered through freely rotatable single bonds. This D–A–D architecture induces strong ICT. Notably, the p‐methoxyphenyl rotor of C‐BDP‐OMe can facilitate rapid intramolecular rotation and vibration, providing an effective non‐radiative relaxation pathway for dissipating absorbed photon energy as heat. The interplay between an environment‐sensitive ICT fluorescence response and rotor‐promoted non‐radiative decay provides a molecular basis for the photothermal performance of C‐BDP‐OMe, rendering it particularly suitable for LD‐targeted photothermal therapy. Motivated by this mechanistic rationale, we next conducted a systematic photothermal characterization of C‐BDP‐OMe (Figure 5A). First, the photothermal conversion efficiency was assessed by irradiating C‐BDP‐OMe solutions of varying concentrations (0–40 µM) with a 660 nm laser at a power density of 200 mW/cm² for 5 min. The results indicated that the system temperature increased significantly with increasing C‐BDP‐OMe concentration (Figure 5B).

FIGURE 5.

(A) Photothermal performance of C‐BDP‐OMe in an aqueous solution under irradiation with a 660 nm laser. (B) Time‐dependent temperature variation curves of C‐BDP‐OMe at varying concentrations (0–40 µM). (C) Temperature response of 40 µM C‐BDP‐OMe under varying laser power densities (0–200 mW/cm²) over time. (D) Heating–cooling curve of a 40 µM C‐BDP‐OMe solution under 200 mW/cm² irradiation and the linear fitting of cooling time versus −ln(θ). (E) Evaluation of photothermal stability through cyclic heating and cooling curves. (F) Infrared thermal images of C‐BDP‐OMe solutions at concentrations of 0 and 40 µM after 5‐min irradiation at 200 mW/cm².

Furthermore, at a fixed C‐BDP‐OMe concentration (40 µM), the system temperature showed a progressive increase with rising laser power, confirming its effective photothermal conversion capability (Figure 5C). The photothermal conversion efficiency of the 40 µM C‐BDP‐OMe solution was determined to be 62.6% through heating–cooling cycle measurements (Figure 5D). The photothermal stability of C‐BDP‐OMe was further validated by consistent performance across three consecutive heating–cooling cycles (Figure 5E). Notably, photothermal measurements showed that the C‐BDP‐OMe solution (40 µM in PBS) reached temperatures ranging from 29.3°C to 52.5°C, which are conducive to inducing cellular heat stress, demonstrating its promising potential for application in tumor photothermal therapy (Figure 5F). Collectively, these results demonstrate that C‐BDP‐OMe is not only a high‐performance LD‐imaging probe but also a promising photothermal agent, supporting its further evaluation as an LD‐targeted theranostic molecule in biological systems.

2.7. In Vitro Validation of PTT Performance and Cell Death Pathways

Building upon the excellent photothermal properties of C‐BDP‐OMe, its antitumor efficacy under laser irradiation was systematically evaluated. HeLa cells were treated with varying concentrations of C‐BDP‐OMe (0–20 µM) for 30 min, followed by laser irradiation (660 nm, 200 mW/cm², 10 min). Cell viability was assessed after an additional 2‐h incubation. Notably, treatment with 15 µM C‐BDP‐OMe resulted in a significant decrease in cell viability to 23% after irradiation, reflecting a 3.5‐fold enhancement in cytotoxicity compared to the control group (Figure 6A), which was further confirmed by live/dead cell staining using Calcein‐AM/PI. As shown in Figure 6B, only the C‐BDP‐OMe + Laser group displayed prominent PI red fluorescence, whereas the control groups (PBS, PBS + Laser, C‐BDP‐OMe alone) exhibited predominantly Calcein‐AM green fluorescence, indicating minimal cytotoxic effects in the absence of combined treatment.

FIGURE 6.

Tumor cell death experiments. (A) Cell viability of HeLa cells treated with C‐BDP‐OMe (0–20 µM) in the absence or presence of 660 nm laser irradiation (200 mW/cm², 10 min). (B) Confocal imaging of HeLa cells under different treatments after co‐staining with Calcein‐AM (green) and PI (red) for 2 h. (C) Cell viability of HeLa cells treated with C‐BDP‐OMe (15 µM) and various cell death inhibitors following 660 nm laser irradiation (200 mW/cm², 10 min). (D) Apoptosis analysis and (E) apoptosis rate of HeLa cells under different treatments by flow cytometry after co‐staining with Annexin V‐FITC/PI dyes. (F) Western blot analysis of PLIN2 protein expression under different treatments. Data in (E) are presented as mean ± SD. Statistical significance was determined by one‐way ANOVA. ****p < 0.0001.

To elucidate the specific mode of programmed cell death (PCD) induced by C‐BDP‐OMe‐mediated LDs‐targeted PTT, various cell death inhibitors were employed to investigate the underlying mechanism [54, 55]. As shown in Figure 6C, distinct cell proliferation profiles were observed following irradiation of HeLa cells pretreated with C‐BDP‐OMe and various cell death inhibitors. Compared with the control group treated with C‐BDP‐OMe alone, both the apoptosis inhibitor z‐VAD‐fmk (61.72% inhibition) and the ferroptosis inhibitor ferrostatin‐1 (Fer‐1, 44.34% inhibition) significantly attenuated the suppression of cell proliferation. Notably, the combination of z‐VAD‐fmk and Fer‐1 synergistically increased cell viability to 80.32%. In stark contrast, neither the autophagy inhibitor 3‐methyladenine (3‐MA, 25.56%) nor the necroptosis inhibitor necrostatin‐1 (Nec‐1, 26.83%) effectively rescued cell survival. These findings collectively indicated that apoptosis and ferroptosis are the predominant PCD pathways activated by C‐BDP‐OMe‐mediated LDs‐targeted PTT. Moreover, the apoptosis induced by the photothermal effect of C‐BDP‐OMe has been confirmed by Annexin V‐FITC/PI flow cytometry. The results showed that the proportion of apoptotic cells in the C‐BDP‐OMe + Laser group increased significantly to 38.44%, whereas the apoptosis rates in the PBS (4.44%), PBS + Laser (6.75%), and C‐BDP‐OMe (5.14%) groups remained relatively low (Figure 6D,E).

To further elucidate the underlying mechanisms of apoptosis and ferroptosis, we systematically investigated the dynamic changes in LDs and key regulatory molecules involved in cell death pathways during photothermal therapy. An experimental phenomenon was observed whereby HeLa cells pre‐treated with C‐BDP‐OMe underwent progressive cellular shrinkage upon exposure to light irradiation, accompanied by a gradual decrease in LDs fluorescence intensity (Figure S10). This phenomenon may be attributed to the temperature elevation caused by photothermal effects, which leads to localized depletion of the phospholipid membrane surrounding LDs, thereby further promoting LDs ablation [56]. PLIN2 plays a critical role in maintaining LDs stability and metabolic regulation, as alterations in its expression levels can significantly influence LDs dynamics and cellular metabolism [18, 57]. Therefore, we performed Western blot analysis to examine changes in PLIN2 expression levels after PTT treatment. The results show that the PTT‐treated group exhibited lower PLIN2 levels than the control group, likely due to LD ablation following PTT (Figure 6F, uncropped blot, Figure S11). Given the protective role of PLIN2 on the LD surface, its reduction may further weaken LD stability and render LDs more susceptible to lipolytic enzymes such as ATGL, thereby potentially amplifying triglyceride hydrolysis and subsequent LD breakdown, in line with previous reports [58].

2.8. Mechanistic Investigation of LD‐Targeted Photothermal Therapy

Under conditions of fatty acid overload, the tricarboxylic acid (TCA) cycle and electron transport chain (ETC) are subjected to heightened metabolic stress, resulting in increased electron leakage and a subsequent surge of ROS in mitochondria, which further exacerbates mitochondrial dysfunction [59]. The overproduction of mitochondrial ROS was quantitatively assessed using the dihydroethidium (DHE)‐based fluorescent staining. As shown in Figure 7A, the C‐BDP‐OMe‐pretreated cells displayed a progressively increasing red fluorescence signal upon laser irradiation, suggesting a significant accumulation of ROS. This finding supports well‐established evidence that excessive fatty acid generation leads to substantial mitochondrial ROS production. To further confirm that ROS overproduction contributes to mitochondrial dysfunction, real‐time mitochondrial imaging was performed during photoirradiation using Mito‐Tracker Green‐based fluorescent staining. The results demonstrated a gradual decrease in mitochondrial green fluorescence intensity (Figure 7B), suggesting that mitophagy occurred due to mitochondrial dysfunction rather than photobleaching of the mitochondrial tracker dye. The latter possibility was excluded by conducting the same fluorescence staining in the absence of C‐BDP‐OMe (Figure S12).

FIGURE 7.

(A) Confocal imaging of ROS levels in HeLa cells under different treatments using DHE staining. (B) Confocal imaging of LDs and mitochondria at different time points after C‐BDP‐OMe treatment and laser irradiation. (C) Confocal imaging of MMP in HeLa cells under different treatments using JC‐1 staining. (D) Confocal imaging of LPO levels in HeLa cells under different treatments using BODIPY 581/591 C11 staining. Western blot analysis of (E) Caspase‐3 and Cleaved caspase‐3 and (F) GPX4 protein expression in HeLa cells under different treatments. (G) Schematic illustration of the proposed mechanism underlying C‐BDP‐OMe‐mediated LD‐targeted photothermal therapy, involving (i) apoptotic and (ii) ferroptotic pathways.

The decrease in mitochondrial membrane potential (ΔΨ _m) serves as both a biomarker of mitochondrial damage and an early indicator of apoptosis. Therefore, mitochondrial membrane potential was evaluated using JC‐1 fluorescence analysis. The results revealed a significant decrease in the red‐to‐green fluorescence ratio, indicating substantial depolarization of the mitochondrial membrane potential in the C‐BDP‐OMe + laser group (Figure 7C), thereby confirming mitochondrial dysfunction and activation of the apoptotic pathway. Furthermore, western blot analysis revealed elevated cleaved caspase‐3 expression in C‐BDP‐OMe‐pretreated cells following laser irradiation (Figure 7E, uncropped blot, Figure S13), providing additional evidence of apoptotic pathway activation.

Excessive mitochondrial‐derived ROS can induce LPO, which subsequently triggers ferroptosis in cells [60]. The intracellular LPO levels were assessed using the BODIPY 581/591 C11 fluorescence staining. A significant decrease in the red/green fluorescence ratio was observed in C‐BDP‐OMe‐pretreated cells following laser irradiation (Figure 7D), indicating elevated LPO levels and confirming the successful induction of ferroptosis during the therapeutic intervention. During ferroptosis, accelerated LPO contributes to the depletion of reduced glutathione and the downregulation of glutathione peroxidase [61], the key enzyme that prevents LPO and suppresses ferroptosis. Western blot analysis revealed decreased GPX4 expression in C‐BDP‐OMe‐pretreated cells after laser irradiation (Figure 7F, uncropped blot, Figure S14), providing definitive molecular evidence of ferroptosis induction. In conclusion, this study provides the first demonstration that LD‐targeted photothermal therapy can concurrently induce apoptosis and ferroptosis by modulating lipid metabolism, thereby affecting mitochondrial dysfunction (Figure 7G). This discovery provides a novel strategy for precisely regulating cell fate by modulating lipid metabolism, with important biological implications and potential clinical applications.

2.9. In Vivo Imaging and Therapeutic Evaluation of C‐BDP‐OMe

Building upon the cellular‐level tumor cell discrimination capability of C‐BDP‐OMe, we further investigated its ability to differentiate cervical cancer tumors from surrounding normal tissues at the tissue level using a murine xenograft model. HeLa cells, characterized by high LD content, were subcutaneously implanted into the right hind limb of female BALB/c nude mice to establish a xenograft tumor model. Once tumors reached an appropriate volume, equal volumes of C‐BDP‐OMe (500 µM, 50 µL) were administered via intratumoral injection and into the contralateral normal tissue as a control. In vivo time‐lapse fluorescence imaging was subsequently performed on anesthetized animals (Figure 8A). The results revealed that a robust fluorescence signal emerged in the tumor region within 30 min post‐injection, attributable to the rapid polarity‐responsive characteristics of C‐BDP‐OMe toward LDs, with signal intensity reaching a plateau by 60 min. Quantitative analysis demonstrated a tumor‐to‐normal tissue fluorescence ratio (T/N) of up to 2.3 (Figure 8B,C), providing strong evidence that C‐BDP‐OMe maintains high discriminatory efficacy in complex biological environments.

FIGURE 8.

(A) Schematic representation of image‐guided tumor resection via intratumoral injection of C‐BDP‐OMe in a nude mouse model bearing HeLa xenograft tumors. (B) Time‐series in vivo fluorescence images and (C) quantitative analysis of fluorescence intensity following intratumoral and normal tissue administration of C‐BDP‐OMe (n = 3). (D) Schematic depiction of image‐guided tumor resection using topical spraying of C‐BDP‐OMe in a nude mouse model bearing HepG2 xenograft tumors. (E) Bright‐field image prior to surgery; (F) fluorescence image; (G) fluorescence image acquired 30 min after spraying C‐BDP‐OMe; (H) surgical excision of the tumor tissue; (I) fluorescence validation image obtained 30 min after reapplication of C‐BDP‐OMe post‐resection.

The in situ visualization capability of C‐BDP‐OMe for delineating liver tumor margins under a topical spray application was further assessed (Figure 8D). Experimental results demonstrated that direct topical administration of C‐BDP‐OMe enabled precise fluorescent labeling of hepatic tumor tissues, allowing clear visualization of their complete anatomical boundaries (Figure 8E–G). Based on this imaging guidance, the fluorescently marked tumor region was surgically resected (Figure 8H), and the surgical bed was subsequently resprayed with the probe to assess potential residual disease. Notably, no detectable fluorescence signal was observed in the resection site (Figure 8I), a result consistent with the negative control group, thereby confirming complete removal of the tumor tissue. These findings collectively demonstrate that C‐BDP‐OMe has strong potential as an effective intraoperative contrast agent and may serve as a reliable tool for real‐time visual guidance during precision tumor resection in clinical surgery.

Finally, we assessed the therapeutic efficacy of C‐BDP‐OMe in treating human cervical cancer (HeLa) via localized LD‐mediated PTT. Tumor‐bearing mice were randomly divided into four groups: PBS group, PBS + laser group, C‐BDP‐OMe group, and C‐BDP‐OMe + laser group. To ensure the effective accumulation of C‐BDP‐OMe in tumor tissues, C‐BDP‐OMe was administered via intratumoral injection directly into the tumor sites of the mice (Figure 9A). Experimental results revealed that only the C‐BDP‐OMe + laser group exhibited a temperature increase to 51°C at the tumor site, whereas the PBS + laser group showed no significant temperature elevation, demonstrating the pronounced photothermal effect of C‐BDP‐OMe (Figure 9B). During the treatment, the body weight of the mice was monitored every 2 days. The results showed no significant differences in body weight across the experimental groups (Figure 9C), confirming the favorable biocompatibility of C‐BDP‐OMe. Furthermore, the tumor volume in the C‐BDP‐OMe + laser group gradually decreased over time; by Day 12, tumors were nearly eliminated, and the wounds were largely healed. In contrast, the tumor volumes in the other groups (PBS, PBS + laser, and C‐BDP‐OMe) continued to increase progressively throughout the observation period (Figure 9D). After the treatment, representative tumor images (Figure S15) and post‐resection tumor photographs (Figure 9E) from the four groups were analyzed, further supporting the conclusion that C‐BDP‐OMe effectively inhibits tumor proliferation via localized LD‐mediated photothermal therapy.

FIGURE 9.

In vivo anti‐tumor performance of C‐BDP‐OMe: (A) Establishment of the BALB/c mouse tumor model and schematic diagram of PTT. (B) Infrared thermal imaging and temperature variation in the PBS + laser and C‐BDP‐OMe + laser groups during photothermal treatment. (C) Body weight and (D) absolute tumor volume of mice under different treatments during PTT (n = 3). (E) Digital photographs of ex vivo tumors and (F) H&E staining of heart, liver, spleen, lung, kidney, and tumor tissues obtained from different treatment groups after 12 days of intervention.

To systematically evaluate the therapeutic efficacy and biocompatibility of C‐BDP‐OMe, histopathological analyses were performed on tumor tissues and major organs (heart, liver, spleen, lung, and kidney) from experimental mice using hematoxylin and eosin (H&E) staining (Figure 9F). The histopathological results revealed that tumors in the C‐BDP‐OMe + laser group exhibited extensive cytoplasmic vacuolization and characteristic nuclear changes, including pyknosis, karyorrhexis, and karyolysis. In contrast, no notable histopathological abnormalities were observed in either tumor tissues or major organs from the three control groups. Furthermore, the tissue architecture of all examined organs in the treatment group remained structurally intact. These results provide compelling evidence for the significant antitumor efficacy of C‐BDP‐OMe‐mediated photothermal therapy and further confirm its excellent biosafety profile.

3. Conclusion

In conclusion, this study reports the successful development of two NIR theranostic fluorescent dyes, C‐BDP‐OH and C‐BDP‐OMe, through an electronic modulation strategy. Notably, C‐BDP‐OMe achieves specific targeting of LDs by combining the “like dissolves like” principle with a dual‐anchoring mechanism involving molecular docking between the probe and the PLIN2 protein. Its high sensitivity to microenvironmental polarity enables high S/N fluorescence imaging of LDs under no‐wash conditions. Moreover, due to the elevated LD content and their intrinsically low‐polarity environment in tumor cells, C‐BDP‐OMe effectively discriminates between tumor and normal cells, as well as between tumor tissues and adjacent healthy tissues. Importantly, C‐BDP‐OMe functions not only as a diagnostic agent for tumor detection but also exhibits outstanding photothermal performance, achieving a photothermal conversion efficiency of up to 62.6%. It enables tumor ablation via LD‐mediated photothermal therapy, with the underlying mechanism systematically elucidated. The localized photothermal effect within LDs induces downregulation of PLIN2, a key structural protein located at the LD surface. This disruption impairs metabolic crosstalk between LDs and mitochondria and triggers two distinct cell death pathways: first, ROS‐mediated activation of caspase‐3 initiates mitochondrial apoptosis; second, the accumulation of LPO products suppresses GPX4, thereby inducing ferroptosis. This work not only establishes a novel design paradigm for precise LD labeling but also innovatively integrates metabolic reprogramming with site‐specific photothermal ablation of LDs in tumor tissues, offering a promising therapeutic strategy for malignant cancers.

Author Contributions

Lipeng Zhang: conceptualization, investigation, writing – original draft, visualization, validation. Jiyoung Yoo: conceptualization, investigation, writing – original draft, visualization, validation. Juan Bai: investigation, validation, methodology, writing – original draft. Sunghyun Kim: validation, visualization, data curation. Yesim Bakan: validation, visualization, data curation. Yongbin Zhang: data curation, validation, visualization. Fangjun Huo: validation, visualization, data curation. Jingying Zhou: validation, visualization, data curation. Caixia Yin: supervision, writing – review and editing, project administration, funding acquisition. Jong Seung Kim: supervision, project administration, writing – review and editing, funding acquisition.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: anie72862‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data supporting the findings of this study are available from the corresponding author upon reasonable request.