Selective Lipolysis by Photoactivation of Chaperone‐Mediated Autophagy Using Adipocyte Membrane‐Coated Nanoparticle in Hydrogel

Abstract

Despite the clinical importance of obesity, treatments have been confined to surgery or drugs, both of which carry adverse effects. Herein a nanomaterial consisting of adipocyte membrane‐coated and rosiglitazone‐loaded polydopamine nanoparticles embedded in hydrogel (ARNP‐H) is designed aiming to induce selective lipolysis in adipocyte through activation of chaperone‐mediated autophagy (CMA) by mild photothermal heat stress. Adipocyte membrane‐coated and rosiglitazone‐loaded polydopamine nanoparticles (ARNP) are taken up by adipocyte to a significantly greater extent than other cell membrane‐coated nanoparticles. Adipocyte interactions with ARNP are significantly higher than those with other cells such as macrophages and T cells. ARNP‐H enabled selective activation of CMA in adipocytes, increasing levels of heat shock cognate protein 70, a key component of CMA. Colocalization of heat shock cognate protein 70 and breakdown of CMA substrate perilipin 2, led to enhancement of adipose triglyceride lipase access to lipid droplets, initiating lipolysis. In high‐fat diet‐fed mice, ARNP‐H retention at the local injection site is greater than in other groups, maintaining photothermal responsiveness for over 3 days. ARNP‐H treatment significantly enhanced adipocyte lipolysis and led to substantial weight loss. While this study focuses on obesity model, ARNP‐H for CMA photoactivation holds potential for treating other diseases associated with autophagy dysregulation.

Adipocyte membrane‐coated nanoparticles loaded with rosiglitazone in hydrogel offer an alternative approach to conventional obesity treatments. This system selectively targets adipocytes, inducing lipolysis by activating chaperone‐mediated autophagy through mild photothermal stimulation. Demonstrating significant weight loss in animal models, this nanomaterial‐based strategy holds promise for obesity management and other autophagy‐related disorders.

1. Introduction

Obesity, a chronic and relapsing condition, presents a significant global public health challenge, affecting ≈39% of the adult population worldwide.^{[ 1 ]} Obesity is closely associated with various diseases, including insulin resistance, type 2 diabetes, dyslipidemia, fatty liver disease, cardiovascular disease, and atherosclerosis.^{[ 2 ]} Although anti‐obesity medications such as glucagon‐like peptide 1 receptor agonists^{[ 3 ]} are clinically available, their effectiveness remains limited, with ongoing concerns about adverse effects.^{[ 4 ]}

Chaperone‐mediated autophagy (CMA) has recently gained attention as a promising therapeutic target for multiple diseases due to its key role in lysosomal degradation of proteins containing the KFERQ motif.^{[ 5} , ^{6 ]} Notably, CMA has been identified as a facilitator of lipolysis, targeting lipid‐associated proteins such as perilipin 2 (PLIN2) for degradation upon activation.^{[ 7 ]} PLIN2 is broadly expressed across various tissues and cells, including adipocytes.^{[ 8 ]} Previous studies report that PLIN2 knockout or downregulation via antisense oligonucleotide prevents high‐fat diet (HFD)‐induced obesity, underscoring its potential as a target for anti‐obesity treatments through CMA activation.^{[ 9 ]}

While the potential of CMA as a target in obesity treatment is recognized, there have been limited efforts to leverage nanomaterials for the photoactivation of CMA. Studies have shown that mild heat stress can elevate the levels of heat shock cognate protein 70 (HSC70) and activate CMA.^{[ 10 ]} HSC70 has been reported to play a pivotal role in CMA by mediating the delivery of PLIN2 to lysosomes and enhancing the access of adipose triglyceride lipase (ATGL) to lipid droplets.^{[ 7 ]} Nanomaterials such as graphene and polydopamine have shown promise as heat‐generating agents under near‐infrared (NIR) irradiation.^{[ 11} , ¹² , ^{13 ]} However, their use has predominantly focused on anticancer phototherapy, leaving their potential for CMA activation in obesity treatment largely unexplored.

In this study, we proposed that selective activation of CMA and subsequent lipolysis in adipocyte could be achieved using adipocyte membrane‐coated and rosiglitazone (RG)‐loaded polydopamine nanoparticles embedded in hydrogel (ARNP‐H). Adipocyte membrane‐coated and rosiglitazone‐loaded polydopamine nanoparticles (ARNP) were specifically coated with adipocyte membrane to enhance the uptake by adipocyte, leveraging the homotypic attraction between adipocyte facilitated by adipocyte adhesive molecules. Polydopamine nanoparticles (PNP) served as the core photoresponsive element, generating controlled heat in adipose tissues upon NIR exposure. Since RG has been reported as a lipolytic agent^{[ 14} , ^{15 ]} and is known to increase ATGL expression,^{[ 16} , ^{17 ]} RG‐loaded PNP (RNP) were formulated to further enhance lipolysis. To prolong in vivo retention, nanoparticles were embedded in a hyaluronic acid‐tyramine conjugate‐based hydrogel (HAT). We report that ARNP‐H could effectively activate CMA in adipocytes via mild heat stress (Figure  1 ), leading to considerable lipolysis and subsequent weight loss.

Figure 1.

Construction of ARNP‐H and proposed CMA‐mediated anti‐obesity mechanism. A) Adipocyte membranes were isolated from adipocytes and used to coat the surfaces of RNP. The resulting ARNP was proposed for facilitating uptake by adipocytes over other cells in adipose tissue. B) A schematic illustration depicts the mechanism of ARNP‐H treatment for a CMA‐mediated anti‐obesity effect. ARNP was embedded in a hydrogel to enhance retention at the local injection site. ARNP was selectively taken up by adipocytes, providing NIR‐responsive heat stress specifically to these cells. Upon NIR irradiation, the mild heat stress upregulated HSC70, which then co‐localized with PLIN2 and facilitated its lysosomal trafficking. Consequently, the degradation of PLIN2, which served as a protective barrier for lipid droplets, allowed cytosolic ATGL to access the lipid droplets and promoted lipolysis.

2. Result

2.1. Characteristics of Nanoparticles and Hydrogels

The physical and biochemical properties of nanoparticles and their hydrogel formulations were characterized. ARNP‐H was constructed by coating RNP with adipocyte cell membranes. Stochastic optical reconstruction microscopy (STORM) imaging revealed the specific presence of prohibitin, indicative of the adipocyte membrane, within ARNP, distinguishing it from RNP (Figure  2A). SDS‐PAGE analysis showed that the protein profiles in ARNP closely matched those of adipocyte membranes (Figure 2B). Quantitative measurements indicated a protein concentration of 108.3 ± 1.0 µg mg⁻¹ in ARNP (Figure 2C).

Figure 2.

Characteristics of nanoparticles and nanoparticle‐loaded hydrogel. A) Representative STORM images displayed Cy3‐labeled RNP and ARNP, with prohibitin on the adipocyte membrane stained using an Alexa Fluor 647‐conjugated antibody. Scale bar: 100 nm. B) SDS‐PAGE analysis was performed for adipocyte membrane (AM), RNP, and ARNP. C) Protein content of nanoparticles was quantified using a BCA assay. Statistical values are represented as mean ± SD (n = 5; ***p < 0.001). D) A schematic illustration depicted the elemental composition, showing phosphorus from the adipocyte membrane, nitrogen and carbon from PNP, and sulfur from RG. E,F) XPS spectra revealed P 2p peaks for phosphorus in the adipocyte membrane of RNP (E) and ARNP (F). G) XPS narrow scanning of P 2p was conducted for ARNP. H) EDS‐TEM analysis examined the elemental composition of ARNP, identifying carbon, nitrogen, phosphorus, and sulfur, represented in blue, red, green, and yellow, respectively. Scale bar: 50 nm. I) Phosphorus content in RNP and ARNP was measured by a phosphate assay. Statistical data are represented as mean ± SD (n = 5; ***p < 0.001). J,K) Hydrodynamic diameter (J) and zeta potential (K) of nanoparticles were determined using dynamic light scattering and laser‐Doppler micro‐electrophoresis, respectively. Statistical data are represented as mean ± SD (n = 5; n.s., not significant, ***p < 0.001). L,M) High resolution‐TEM was used to observe lattice structure in ARNP (L) and diffraction pattern of the crystalline phase (M). N) A schematic illustrated HAT and ARNP‐H, with ARNP incorporated into the HAT‐based hydrogel for sustained in vivo release. O) Representative photographs displayed ARNP, HAT, and ARNP‐H in vials. P) SEM images presented ARNP within the HAT gel matrix, highlighted with a pseudocolor image (red). Scale bar: 1 µm. Q) STORM images captured Cy5‐labeled ARNP embedded in Cy3‐labeled HAT. Scale bar: 200 nm. R) Thermal images were taken post 808 nm irradiation using an infrared thermal camera.

Elemental analysis (Figure 2D) identified key components of ARNP, including phosphorus from the adipocyte membrane, nitrogen from PNP, and sulfur from RG. XPS narrow scanning distinctively detected phosphorus peaks in ARNP, which were absent in RNP (Figure 2E–G). Energy‐dispersive X‐ray spectroscopy coupled with transmission electron microscopy (EDS‐TEM) imaging highlighted the spherical morphology of ARNP and confirmed the presence of phosphorus, carbon, nitrogen, and sulfur (Figure 2H). A phosphate content analysis measured 13.6 ± 0.7 µg of phosphate per mg of ARNP (Figure 2I). Coating RNP with the adipocyte membrane did not significantly affect its size or polydispersity index (PDI) (Figure 2J; Figure S1A–C, Supporting Information). SEM imaging revealed the spherical morphology of both RNP and ARNP (Figure S1D, Supporting Information). Zeta potential measurements recorded −39.0 ± 2.8 mV for RNP and −21.4 ± 2.2 mV for ARNP (Figure 2K).

High resolution‐TEM imaging revealed the lattice structure of ARNP, showing a hybrid composition of predominantly amorphous regions with areas of partial crystallinity. In crystalline regions, lattice spacings of ≈0.2 and 0.45 nm were observed. Additional lattice spacings under 0.21 nm indicate in‐plane graphitic domains rich in π electrons, while the 0.34 nm spacing suggests interlayer graphitic spacing.^{[ 18 ]} The 0.45 nm distance indicates increased carbon spacing, possibly due to interstitial effects such as doping (Figure 2L,M).

High‐performance liquid chromatography (HPLC; Dionex Ultimate 3000 system, Dionex, Waltham, MA, USA) analysis showed that ARNP had an RG loading capacity of 12.15 ± 1.27 wt.% (Figure S2, Supporting Information). Under acidic condition, RG release from ARNP reached 26.29 ± 1.18% (Figure S3, Supporting Information).

HAT were synthesized by conjugating tyramine to hyaluronic acid through EDC/NHS coupling (Figure 2N). HAT hydrogels were subsequently loaded with nanoparticles (Figure 2O). Scanning electron microscopy (SEM) images revealed the incorporation of ARNP within the HAT matrix, shown with pseudocoloring (Figure 2P). STORM further visualized ARNP distribution within the HAT hydrogel (Figure 2Q). The addition of hyaluronidase enabled a controlled release of nanoparticles from the HAT hydrogel over time (Figure S4B, Supporting Information). Mechanical testing demonstrated comparable elastic modulus (G′) and loss modulus (G″) between the plain HAT hydrogel and ARNP‐H (Figure S4C,D, Supporting Information). The swelling ratios for HA and ARNP‐H hydrogels were 14752% and 12276%, respectively (Figure S4E, Supporting Information). Upon exposure to 808 nm NIR irradiation, the temperatures of RNP, ARNP, and ARNP‐H formulations increased (Figure 2R).

2.2. Adipocyte Uptake of Nanoparticles and Regulation of Lipolytic Proteins

We investigated the impact of the adipocyte membrane coating on nanoparticle uptake by adipocytes and the ability of RG to modulate proteins involved in lipolysis (Figure  3A). Adipocytes treated with RNP showed no significant color change, while those treated with ARNP appeared dark black, reflecting the color of the PNP core (Figure 3B). Cy5‐labeled nanoparticle treatments revealed distinct uptake patterns, with ARNP showing greater uptake by 3T3‐L1 adipocytes compared to RNP (Figure 3C).

Figure 3.

Nanoparticle uptake by adipocytes and other cell types. A) Illustrations depict the structure of various nanoparticles, and uptake test by adipocytes, dendritic cells, T cells, and macrophages. Uptake of nanoparticles by cells was assessed by confocal microscopy, flow cytometry, and functional test of adipocyte proteins. B) Photographs display adipocyte pellets post‐treatment with RNP or ARNP. C) 3T3‐L1 adipocytes were treated with fluorescent nanoparticles, and the mean fluorescence intensity values of adipocytes were recorded at different time points using flow cytometry. (n = 3; ***p < 0.001). D) 3T3‐L1 adipocytes, BMDCs, BMDMs, and T cells were treated with Cy5 (red)‐labeled nanoparticles and stained with BODIPY dye for lipid droplets (green). Scale bar: 20 µm. The cells were examined by confocal microscopy. E–G) Cells were treated with various nanoparticles for 2 h and then incubated for an additional 24 h. The populations of cells with high Cy5 fluorescence were assessed by flow cytometry (E) and quantified (F). (n = 5; n.s., notsignificant, ***p < 0.001). G) The relative ratio of the Cy5 high population in the ARNP‐treated group compared to the E‐RNP‐treated group is presented. (n = 5; n.s., not significant, ***p < 0.001) H,I) 3T3‐L1 adipocytes treated with different nanoparticles were stained with an anti‐ATGL antibody. The expression of ATGL (red) was visualized by confocal microscopy (H) and quantified by flow cytometry (I). Scale bar: 10 µm. (n = 5; **p < 0.01, ***p < 0.001). J) In vivo cellular uptake of Cy5‐labeled RNP and Cy5‐labeled ARNP following injection into ingWAT was assessed. Adipocytes and immune cells were isolated from ingWAT, and the uptake of Cy5‐labeled nanoparticles was measured. K) The proportion of Cy5 high adipocytes, macrophages, dendritic cells, T cells, and neutrophils isolated from ingWAT was analyzed by flow cytometry. (n = 5; n.s., not significant, ***p < 0.001). All Statistical values are represented as mean ± SD.

ARNP exhibited significantly greater uptake in adipocytes compared to other nanoparticles, with no increased uptake observed in other cell types. Confocal microscopy showed no notable differences in Cy5 fluorescence intensity among adipocytes treated with RNP, E‐RNP, or N‐RNP (Figure 3D). However, ARNP treatment produced intense red fluorescence, indicating substantial uptake of Cy5‐labeled nanoparticles. Flow cytometry confirmed greater ARNP uptake in adipocytes compared to other treatments (Figure 3E), with ARNP achieving 3.1‐ and 1.8‐fold higher uptake than RNP and E‐RNP, respectively, in 3T3‐L1 adipocytes (Figure 3F). T cells, bone marrow‐derived dendritic cells (BMDCs), and bone marrow‐derived macrophages (BMDMs) showed no significant differences in uptake between ARNP and E‐RNP (Figure 3G).

ARNP‐treated adipocytes exhibited upregulated expression of ATGL and uncoupling protein‐1 (UCP1). Confocal imaging showed increased ATGL levels in ARNP‐treated cells compared to those treated with RNP (Figure 3H). FACS analysis indicated a 1.8‐fold increase in ATGL expression in ARNP‐treated adipocytes relative to the RNP group (Figure 3I; Figure S5, Supporting Information). UCP1 expression was highest in ARNP‐treated adipocytes, as shown by confocal imaging (Figure S6A, Supporting Information), with flow cytometry further confirming elevated UCP1 levels compared to RNP‐treated cells (Figure S6B,C, Supporting Information).

To further investigate the effect of differentiated 3T3‐L1 adipocyte membrane coating on ARNP delivery, ARNP coated with either 3T3‐L1 adipocytes or primary adipocytes were compared following subcutaneous administration. Both 3T3‐L1 adipocyte membrane coating and primary adipocyte membrane coating did not significantly differ in binding to adipocytes at inguinal white adipose tissue (ingWAT), as shown in Figures S7 and S8 (Supporting Information). Following subcutaneous administration to ingWAT, in vivo cellular uptake of ARNP was significantly higher in primary adipocytes compared to other immune cells, such as macrophages, dendritic cells, T cells, and neutrophils (Figure 3J,K). The adipocyte uptake of ARNP was 10.6‐fold higher, compared to that of RNP.

2.3. Photoactivation of CMA and Lipolysis

To determine the optimal temperature for adipocyte photoactivation of CMA, we measured HSC70 levels at various temperatures. HSC70 expression in adipocytes incubated at 42 °C was 2.1‐fold higher than in those maintained at 37 °C (Figure S9A,B, Supporting Information). The viability of 3T3− L1 adipocytes was maintained at temperatures up to 42 °C but decreased significantly at higher temperatures (Figure S10A,B, Supporting Information). Treating adipocytes with ARNP in hydrogel significantly increased lipolysis upon NIR irradiation. The PNP core of ARNP induced temperature changes in adipocytes, potentially enhancing lipolysis (Figure  4A). Among the tested formulations, ARNP‐H caused the most significant temperature rise, reaching ≈42 °C within 1 min of NIR irradiation (Figure 4B,C).

Figure 4.

Photoactivation of CMA and cascading lipolysis. A) Treatment of 3T3‐L1 adipocytes with ARNP‐H is illustrated. Adipocytes differentiated from 3T3‐L1 cells were cultured in 24‐well plates. Formulations were added to hanging inserts placed above the wells. To mimic in vivo nanoparticle clearance, the medium was replaced at 30 min after initial treatment and then every 12 h. At 72 h post‐treatment with ARNP‐H, NIR irradiation was applied to the transwell system. B) An infrared thermal camera was used to visualize the temperatures of untreated or ARNP‐H‐treated adipocytes. C) The temperatures of samples were presented. D) After treatment with ARNP‐H, cells were stained with BODIPY before NIR exposure, enabling tracking of lipid droplet size for 15 h. Scale bar: 15 µm. Real‐time video movies are shown in Videos S1 and S2 (Supporting Information). E,F) BODIPY levels were assessed by flow cytometry 24 h post‐NIR irradiation (E), and quantified (F). G) Triglyceride levels were quantified using a triacylglycerol assay. H) Flow cytometry analysis reveals HSC70 expression in adipocytes following ARNP‐H treatment and NIR laser exposure. I) Quantification indicates HSC70‐positive cell populations in adipocytes post‐ARNP‐H treatment and NIR exposure. J) A representative fluorescence images show colocalization of HSC70 (red) and PLIN2 (yellow) in adipocytes treated with ARNP‐H and NIR. Scale bar: 10 µm. K) Cellular colocalization of HSC70 and PLIN2 was quantified using ImageJ software. L) Confocal microscopy reveals colocalization of PLIN2 (yellow) and LAMP1 (green). Scale bar: 15 µm. M) ImageJ software was used to quantify PLIN2 and LAMP1 colocalization. N) Confocal fluorescence microscopy images display ATGL colocalization with BODIPY. Scale bar: 15 µm. O) Colocalization of ATGL and BODIPY was quantified using ImageJ software. All statistical data are represented as mean ± SD (n = 5; n.s., not significant, **p < 0.01, ***p < 0.001).

Real‐time imaging captured intracellular lipolysis upon NIR irradiation (Figure 4D). Upon NIR irradiation, untreated adipocytes showed no changes in the size and number of lipid droplets over time (Video S1, Supporting Information). In contrast, adipocytes treated with ARNP‐H demonstrated time‐dependent alterations in lipid droplets following NIR irradiation (Video S2, Supporting Information). Importantly, no significant change in cell viability was observed in ARNP‐H‐treated adipocytes compared to the untreated group after NIR irradiation (Figure S11A,B, Supporting Information). Post‐treatment with ARNP‐H, adipocytes exhibited the lowest levels of the lipid‐staining marker BODIPY (Figure 4E,F) and the lowest triglyceride levels following NIR irradiation (Figure 4G)

Adipocytes exposed to ARNP‐H and NIR irradiation at 42 °C exhibited a significant rise in HSC70 levels, reaching 2.6 times higher than in the control group (Figure 4H,I). Confocal microscopy showed colocalization of HSC70 and PLIN2 following ARNP‐H treatment and NIR irradiation (Figure 4J,K). Further analysis of HSC70 and PLIN2 trafficking, based on colocalization with lysosomal‐associated membrane protein 1 (LAMP1), revealed a 1.8‐fold increase in PLIN2 and LAMP1 colocalization in the ARNP‐H and NIR‐irradiated group compared to untreated control group (Figure 4L,M). The PLIN2 downregulation was transient, with levels returning to baseline within 24 h after photothermal irradiation (Figure S9C, Supporting Information). ARNP‐H treatment enhanced ATGL association with lipid droplets (Figure 4N,O).

2.4. In Vivo Anti‐Obesity Effect

The in vivo anti‐obesity effects of ARNP‐H were evaluated using the protocol illustrated in Figure  5A, which involved weekly local injections into the ingWAT followed by NIR irradiation. The ARNP‐H group exhibited prolonged retention of Cy7‐labeled ARNP in ingWAT, as evidenced by significantly higher fluorescence intensity of Cy7 on day 9 compared to the ARNP‐alone group (Figure 5B). Notably, only the ARNP‐H group demonstrated a temperature increase following NIR irradiation, whereas no comparable response was observed in the other groups (Figure 5C,D).

Figure 5.

In vivo anti‐obesity effect in HFD mouse model. A) The treatment regimen of ARNP‐H in the HFD mouse model is depicted schematically. Various formulations were locally administered to ingWAT in HFD mice, followed by NIR irradiation on days 1, 3, and 5 after injection. After 5 weeks of this repeated treatment, the mice were sacrificed for further analysis. B) HFD mice received local injections of Cy7‐labelled ARNP or ARNP‐H. In vivo imaging showcased the retention of Cy7‐labelled ARNP at the injection site, with fluorescence measured over 9 days post‐dose. C) Thermal imaging captured temperature changes in mice at various times following the initial treatment and NIR irradiation at 1, 3, 5 days post‐first treatment. D) Temperatures in ingWAT were monitored post‐NIR laser irradiation. Statistical values are represented as mean ± SD (n = 3). E) Photographs of mice were taken after completing 5 weeks of treatment. F) Body weights of the HFD mice were recorded following the initial administration of the various formulations (n = 5).

After 5 weeks of treatment, mice in the ARNP‐H and NIR‐treated group displayed significantly lower body weight compared to the other groups (Figure 5E,F). Food intake was monitored throughout in vivo study, with no significant difference observed between untreated mice and those treated with ARNP‐H, irrespective of NIR irradiation (Figure S12, Supporting Information).

ARNP‐H treatment had no adverse effects on liver or kidney function, regardless of NIR irradiation. Serum analysis revealed no significant differences in liver function markers, including alkaline phosphatase (ALP, Figure S13A, Supporting Information), alanine aminotransferase (ALT, Figure S13B, Supporting Information), and aspartate aminotransferase (AST, Figure S13C, Supporting Information), between untreated and ARNP‐treated groups (Figure S13A–F, Supporting Information). Similarly, kidney function markers, including creatinine (CRE, Figure S13D, Supporting Information), blood urea nitrogen (BUN, Figure S13E, Supporting Information), and creatine kinase‐MB (CKMB, Figure S13F, Supporting Information), remained unchanged across all groups.

2.5. In Vivo Fat‐Lowering Effect

The fat‐reducing impact of ARNP‐H was assessed through body weight, lipid droplet size, and triglyceride level analyses (Figure  6A). Microcomputed tomography (micro‐CT) demonstrated a significant diminution in both subcutaneous white adipose tissue (sWAT) and visceral white adipose tissue (vWAT) volumes in ARNP‐H‐treated mice (Figure 6B). Specifically, the sWAT and vWAT volumes in the ARNP‐H group were 7.0 and 5.7 times lower, respectively, than those in the untreated group (Figure 6C,D). Excised ingWAT also exhibited a marked size reduction post‐ARNP‐H and NIR treatment (Figure 6E), with the ingWAT weight in the ARNP‐H plus NIR group being the lowest, showing a 2.7‐fold decrease compared to the ARNP‐H group without NIR (Figure 6F).

Figure 6.

In vivo imaging of adipose tissues. A) Mice received 5 weeks of treatment with various formulations before sacrifice for body weight assessment, lipid droplet size measurement, and triglyceride quantification. B) Micro‐CT imaging analyzed subcutaneous (red) and visceral (yellow) fat. C,D) The volumes of subcutaneous white adipose tissue (C) and vWAT(D) were calculated using micro‐CT data. (n = 3; *p < 0.05, **p < 0.01, ***p < 0.001). E) Excised ingWAT was photographed. F) The weights of ingWAT for each group are presented. (n = 5; ***p < 0.001). G) H&E staining of ingWAT. Scale bar: 50 µm. H,I) The area of lipid droplets was calculated from H&E‐stained images of ingWAT for groups without (H) and with (I) NIR irradiation. (n = 150; *p < 0.05, **p < 0.01, ***p < 0.001). J) Triglyceride levels in ingWAT were quantified and normalized per 100 µg of protein. (n = 5; n.s., not significant, *p < 0.05, **p < 0.01, ***p < 0.001). All statistical values are represented as mean ± SD.

Furthermore, 808 nm NIR irradiation led to a significant reduction in lipid droplet average size in the ARNP‐H group, which was 6.7‐fold lower than in the untreated group (Figure 6G–I). The triglyceride content, normalized to 100 µg of protein, was notably lower in the ingWAT of the ARNP‐H treated group, with an 8.7‐fold reduction compared to the untreated group (Figure 6J). Following repeated administration and NIR irradiation (Figure S14A, Supporting Information), plasma adiponectin levels were significantly higher in the ARNP‐H group compared to the other groups (Figure S14B, Supporting Information).

2.6. CMA Photoactivation of ARNP‐H

As shown in Figure  7A, HFD mice received treatment with various formulations and NIR irradiation for 5 weeks, followed by an evaluation of protein markers linked to the CMA pathway. Utilizing immunofluorescence for perilipin 1 (PLIN1), a marker for lipid droplets, we visualized lipid droplets within ingWAT. PLIN1 staining revealed that lipid droplet sizes were notably smaller in the group treated with ARNP‐H compared to other groups (Figure 7B). Levels of HSC70 were significantly higher in ingWAT from the ARNP‐H‐treated group versus others (Figure 7B,C). Conversely, PLIN2 levels were the lowest in ingWAT of ARNP‐H‐treated mice (Figure 7D,E). ARNP‐H treatment resulted in the highest levels of ATGL, critical for adipose triacylglycerol lipolysis (Figure 7F,G). Moreover, UCP1, indicative of thermogenesis, was significantly increased in the ARNP‐H group, exhibiting a 14.3‐fold rise compared to the untreated control and a 1.3‐fold increase over the ARNP‐treated group (Figure S15A,B, Supporting Information).

Figure 7.

CMA‐related proteins in adipose tissues. A) Mice were subjected to a 6‐week HFD regimen, during which ingWAT received local treatments and NIR irradiation on days 1, 3, and 5, repeated across 5 weeks. After 6 weeks, ingWAT was isolated to analyze CMA‐related proteins immunohistochemically. B,C) Isolated ingWAT was stained with Alexa Fluor 647‐labeled anti‐PLIN1 and Alexa Fluor 488‐labeled anti‐HSC70 antibodies, revealing PLIN1 (green) and HSC70 (red). A representative image of ingWAT stained for PLIN1 and HSC70 (B), and the mean fluorescence intensity of HSC70 (C) were quantified using ImageJ. Scale bar: 50 µm. D,E) The isolated ingWAT was fixed and stained with Alexa Fluor 568‐labeled anti‐PLIN2 antibody. A representative image of PLIN2‐stained ingWAT (D), and the mean fluorescence intensity of PLIN2 (E) were quantified using ImageJ. Scale bar: 50 µm. F,G) The isolated ingWAT was stained with Alexa Fluor 647‐labeled anti‐ATGL antibody, shown in yellow. A representative image of ATGL‐stained ingWAT (F), and the mean fluorescence intensity of ATGL (G) were quantified using ImageJ. Scale bar: 50 µm. Statistical values are represented as mean ± SD. (n = 5; n.s., not significant, *p < 0.05, **p < 0.01, ***p < 0.001).

2.7. Mechanism of Lipolysis and CMA Photoactivation

To elucidate the mechanisms underlying lipolysis and photoactivation of CMA by ARNP‐H, we investigated the translocation of PLIN2 to lysosomes, both in the presence and absence of the lysosomal protease inhibitor leupeptin (Figure  8A). Confocal microscopy visualized significant NIR‐induced accumulation of PLIN2 in protease‐inhibited lysosomes in adipocytes treated with ARNP‐H (Figure 8B,C). FACS analysis indicated that the photothermal effect of ARNP‐H facilitated an increase in CMA activity and PLIN2 degradation within lysosomes. In lysosomes from untreated adipocytes, PLIN2 levels remained unchanged, irrespective of NIR exposure.

Figure 8.

In vivo photoactivation of CMA by ARNP‐H. A) HFD mice were locally treated with ARNP‐H and subcutaneously injected with the lysosomal protease inhibitor leupeptin. Following NIR irradiation, lysosomes from ingWAT were analyzed. B,C) Immunostaining reveal PLIN2 in lysosomes, which were marked with LAMP1. Scale bar: 2 µm. D) From isolated ingWAT adipocyte lysosomes, PLIN2 expression was represented by flow cytometry data. E) Flow cytometry data provided mean fluorescence intensity quantification of PLIN2. F) Western blot analysis was performed to assess PLIN2, LAMP1, and β‐actin expression. G) HFD mice were locally administered ARNP‐H without lysosomal protease inhibitor. Post‐NIR irradiation, THUNDER imaging captured PLIN2 in adipose tissues, stained with Alexa 568‐conjugated anti‐PLIN2 antibody and BODIPY for lipid droplets. H) Using ImageJ, the mean fluorescence intensity of PLIN2 in adipose tissues was quantified. Statistical values are represented as mean ± SD (n = 5; n.s., not significant, **p < 0.01, ***p < 0.001).

Conversely, in mice treated with the lysosomal protease inhibitor, localized administration of ARNP‐H and NIR irradiation led to a notable increase in PLIN2 accumulation within lysosomal degradation compared to mice treated with ARNP‐H alone, suggesting a role for ARNP‐H and NIR irradiation in enhancing lysosomal degradation of PLIN2 (Figure 8D,E). Consistently, western blot analysis showed that adipocytes treated with ARNP‐H and subjected to NIR irradiation exhibited enhanced translocation of PLIN2 into lysosomes (Figure 8F). In cases where mice were not administered the lysosomal protease inhibitor, THUNDER imaging demonstrated reduced PLIN2 levels in adipose tissues, likely a result of lysosomal PLIN2 degradation following CMA photoactivation by ARNP‐H and NIR irradiation (Figure 8G,H).

3. Discussion

In this study, we developed a light‐responsive hydrogel designed to release nanoparticles coated with adipocyte membranes, enhancing targeted delivery to adipocytes. This hydrogel system, ARNP‐H, triggered CMA in adipocytes, promoting lipolysis and producing an anti‐obesity effect. Specifically, we applied mild photothermal therapy via NIR irradiation following ARNP‐H treatment to boost HSC70 expression. This approach led to the lysosomal degradation of PLIN2, a lipid droplet‐associated protein, activating cytosolic ATGL and catalyzing lipid droplet breakdown, which reduced adipocyte size.

3T3‐L1 cells have been widely used as a model for adipose tissue studies due to their well‐characterized differentiation process, transitioning from fibroblasts to preadipocytes and subsequently into mature adipocytes under appropriate conditions. In this study, differentiated 3T3‐L1 adipocytes were used to coat nanoparticles to enhance ARNP uptake by adipocytes. In this study, enhanced uptake of ARNP by adipocytes was observed both in vitro and in vivo. In vitro uptake of ARNP was significantly higher in 3T3‐L1 adipocytes compared to T cells, dendritic cells, and macrophages (Figure 3D). In contrast, nanoparticles coated with erythrocyte membranes or NIH‐3T3 fibroblast membranes did not show enhanced uptake by adipocytes. Although 3T3‐L1 adipocytes do not fully replicate the structural complexity of primary adipocytes, they are widely used in adipocyte research due to their morphological and functional similarity to primary adipocytes.^{[ 19 ]}

Consistent with these in vitro findings, in vivo cellular uptake studies revealed that ARNP preferentially accumulated in adipocytes within ingWAT, with significantly lower distribution to other cell types in the tissue (Figure 3J,K). Previous studies have shown that adipocytes engage in homotypic interactions mediated by adhesion molecules.^{[ 20 ]} For instance, adipocyte‐specific adhesion molecules regulate actin polymerization and are essential for maintaining adipocyte function and metabolism.^{[ 21 ]} Moreover, inter‐adipocyte adhesion and signaling are modulated by extracellular matrix components such as collagen IV.^{[ 22 ]} These mechanisms may contribute to the preferential uptake of ARNP by adipocytes, likely through homotypic membrane recognition and adhesion‐mediated interactions.

Beyond enhanced uptake, we also observed notable activation of CMA only in adipocytes following treatment with ARNP‐H and NIR irradiation in HFD mice. Adipocytes exposed to ARNP‐H and NIR light showed significantly increased HSC70 expression (Figure 4H,I), indicative of CMA activation. This was further supported by the colocalization of HSC70 with PLIN2 (Figure 4J,K). Notably, PLIN2 expression was reduced only after NIR irradiation (Figure S9C, Supporting Information), correlating with the onset of lipolysis and indicating that light‐triggered activation is essential for lipid droplet breakdown. Similar patterns of CMA activation were observed in vivo in HFD mice treated with ARNP‐H. Taken together, these results support that ARNP‐H enables adipocyte‐targeted delivery of RG and promotes selective activation of CMA and lipolysis upon NIR laser exposure, supporting its potential as a precision nanotherapy for obesity.

We also evaluated the immunogenic effects of ARNP following repeated injections. TNFα levels in ingWAT and adipose tissue‐associated macrophages were assessed using ELISA and flow cytometry (Figure S16, Supporting Information). ARNP was administered weekly for five weeks at the same dosage used in the in vivo lipolytic efficacy test. No significant differences were observed in the population of TNFα‐expressing macrophages or TNFα levels in adipose tissue, indicating minimal immunogenic effects.

The use of allogeneic cell membranes has been reported to be well tolerated in various murine models. Previous studies have shown that the administration of Raw264.7 macrophage‐derived membrane‐coated nanoparticles (BALB/c origin) to C57BL/6 mice did not induce significant immunogenic responses.^{[ 23} , ^{24 ]} Similarly, Raw264.7 macrophage‐derived membrane‐coated nanoparticles did not trigger immunogenic side effects in other non‐syngeneic Kunming mice (Swiss mouse lineage), or even Sprague‐Dawley rats.^{[ 25} , ^{26 ]} Furthermore, B16F10 cell membrane nanoparticles (C57BL/6 origin) have demonstrated cross‐strain compatibility in vitro when used with Raw264.7 cells of BALB/c origin.^{[ 27 ]} NIH/3T3 fibroblast‐derived cancer‐associated fibroblast membranes (Swiss mouse lineage) have also been used in studies involving both BALB/c and C57BL/6 mice.^{[ 28} , ^{29 ]}

To evaluate the enhanced delivery of RG using adipocyte membrane‐coated nanoparticles, we selected a dose of RG that had previously been reported to be ineffective. Studies have shown that RG administration at doses of 4 mg kg⁻¹ or higher can induce weight loss in mice.^{[ 30 ]} However, a lower dose of 2.5 mg kg⁻¹, when locally injected into ingWAT, had no significant effect on body weight.^{[ 15 ]} Based on these findings, we selected a dose of 1 mg kg⁻¹ per injection for our in vivo study. Consistent with previous reports, no weight loss was observed in the group treated with RG alone (Figure 5F).

Polydopamine, a key component of ARNP, is known to be pH‐sensitive and degrade in the oxidative and acidic lysosomal environment.^{[ 31} , ^{32 ]} Based on these findings, it is likely that RG release is facilitated by nanoparticle degradation under acidic lysosomal conditions. We observed a pH‐dependent release of RG from ARNP (Figure S3, Supporting Information). After incubation in a lysosome‐mimicking acidic buffer (pH 4.5), ARNP exhibited a significantly higher release of RG (26.29 ± 1.18%) compared to neutral conditions. This pH‐dependent release suggests that RG can be effectively released within lysosomes following endocytosis.

The observed enhancement in lipolysis following ARNP‐H treatment can be attributed to increased ATGL expression, elevated UCP1 expression, and greater ATGL access to lipid droplets. We found that RG encapsulated in ARNP upregulated ATGL expression in adipocytes (Figure 7E,F). RG has previously been reported to increase ATGL expression in both mouse WAT and 3T3‐L1 preadipocytes.^{[ 16} , ^{17 ]} Moreover, ATGL overexpression has been shown to mitigate diet‐induced obesity.^{[ 33 ]} ATGL expression facilitated by RG could provide synergistic effect on the lipolysis by ARNP‐H and light irradiation.

Increased lipolysis was also associated with UCP1 upregulation. ARNP treatment significantly enhanced UCP1 expression in adipocytes, with the highest expression levels observed in cells treated with ARNP (Figure S6, Supporting Information). In vivo experiments further confirmed that UCP1 levels were elevated in adipose tissues following ARNP‐H treatment and light irradiation (Figure S15, Supporting Information). These findings suggest that higher adipocyte uptake of RG delivered via ARNP‐H promotes UCP1 expression, facilitating the conversion of white adipose tissue into a thermogenic phenotype, thereby promoting lipolysis and weight reduction.^{[ 14} , ^{15 ]}

Following ARNP‐H treatment and NIR irradiation, PLIN2 degradation facilitated greater ATGL access to lipid droplets, further promoting lipolysis. Intracellular PLIN2 levels significantly decreased after ARNP‐H treatment and NIR irradiation (Figure S9C, Supporting Information). PLIN2 is known to protect lipid droplets by preventing ATGL‐mediated triglyceride hydrolysis.^{[ 34} , ^{35 ]} ATGL plays a crucial role in triglyceride hydrolysis, and its absence leads to excessive lipid accumulation and metabolic dysregulation.^{[ 36} , ^{37 ]}

Chronic consumption of HFD is recognized for its inhibitory effects on CMA activity.^{[ 38} , ^{39 ]} Additionally, impairment in CMA function is known to lead to the buildup of lipid droplets, an increase in adipocyte progenitors, and consequent hyperplasia, inflammation, and fibrosis in adipose tissue.^{[ 40} , ^{41 ]} The anti‐obesity effect in HFD mouse model might be contributed due to the photoactivation of CMA by ARNP‐H and NIR treatment.

To trigger CMA‐mediated lipolysis in adipocytes, we utilized mild photothermal therapy through NIR laser irradiation (Figure 4A). HSC70, a critical component of CMA, is essential for the CMA‐mediated breakdown of targeted proteins.^{[ 42 ]} Moreover, mild thermal shock has been observed to boost the expression of heat shock proteins and facilitate their relocation to lipid droplets.^{[ 10} , ^{43 ]} Notably, an increase in HSC70 expression induced by starvation has been reported to be associated with the degradation of PLIN2.^{[ 44 ]} It is known that CMA activation leads to the degradation of PLIN2, a prevalent protein that encases intracellular lipid droplets, thereby preventing lipase access to triglycerides.^{[ 45 ]} PLIN2 acts as a key protective layer on lipid droplet surfaces, shielding lipid droplets from lipases and triacylglycerol hydrolysis.^{[ 46 ]}

Previous studies have reported that CMA activation, followed by the degradation of PLIN proteins, plays a crucial role in stimulating hepatic lipolysis. A decrease in PLIN2 levels is known to be associated with the reduction of lipid droplet size.^{[ 44 ]} Specifically, starvation‐induced CMA activation has been shown to lead to the lysosomal degradation of PLIN2, enhancing lipase access to lipid droplets and facilitating lipolysis.^{[ 44} , ^{47 ]} Moreover, the absence of PLIN2 in myotubes has been shown to promote lipolysis.^{[ 48 ]} Mice lacking PLIN2 have been found to be resistant to obesity, adipose tissue inflammation, fatty liver disease, and exhibit higher insulin sensitivity and browning.^{[ 49} , ^{50 ]} These findings underscore the potential of targeting PLIN2 as a means of activating CMA for the treatment of metabolic diseases.

Encapsulating nanoparticles within the hydrogel is expected to enhance therapeutic efficacy while minimizing off‐target effects. Hyaluronic acid has been reported to be biodegradable in vivo through both hydrolysis and enzymatic degradation by hyaluronidases. Previous studies have shown high levels of hyaluronidase expression in skin tissues,^{[ 51 ]} which specifically target the high‐molecular‐weight hyaluronic acid backbone.^{[ 52 ]} Similarly, mimicking the in vivo environment by supplementing the media with hyaluronidase enhanced the diffusion rate of ARNP in the ARNP‐H hydrogel (Figure S4, Supporting Information). The optimal HAT concentration in ARNP‐H was determined based on the structural stability of the hydrogel network, which supports sustained release at the injection site. HAT doses below 100 µg led to the rapid release of ARNP (Figure S17, Supporting Information). Encapsulation of ARNP within HAT (100 µg) significantly prolonged in vivo nanoparticle retention up to day 9, allowing for at least three rounds of NIR irradiation per injection (Figure 5B; Figure S17, Supporting Information).

The dose of the PNP core was selected based on its ability to maintain the temperature required for mild hyperthermia following ARNP‐H administration with NIR irradiation (Figure S18, Supporting Information). The temperature of NIR‐irradiated sites varied with the PNP dose. At 100 µg PNP, the maximum recorded temperature was 39.4 °C after 6 min of NIR irradiation. At 200 µg PNP, temperatures reached a plateau of 40.0–42.0 °C from 2 min of irradiation. At 400 µg PNP, temperatures exceeded 45 °C. For mild photothermal effect, we selected a PNP dose of 200 µg for the ARNP‐H formulation. Following ARNP‐H (at a dose of 200 µg PNP) administration with NIR irradiation, we observed a significant upregulation of HSC70 and ATGL, along with a downregulation of PLIN2 (Figure 7B–G and Figure 8G,H). Additionally, ARNP‐H combined with NIR treatment promoted lysosomal engulfment of PLIN2 (Figure 8). These findings suggest that mild thermal stress is an effective strategy for increasing HSC70 levels while simultaneously reducing PLIN2 expression.

In this study, we tested the peripheral effect of ARNP‐H under mild photothermal conditions. Various nanoparticles were locally administered to ingWAT, followed by NIR irradiation at the injection sites. A mild photothermal effect, with temperatures kept below 42 °C, was applied. Under these conditions, liver and kidney functions remained unaffected (Figure S13, Supporting Information). Previous studies have shown that targeted NIR irradiation has minimal impact on healthy tissues and can help mitigate peripheral side effects.^{[ 53} , ⁵⁴ , ^{55 ]}

We observed that ARNP‐H treatment combined with NIR irradiation significantly increased plasma adiponectin levels (Figure S14, Supporting Information). This elevation may contribute to systemic metabolic regulation and body weight management, including the modulation of subcutaneous adipose tissue.^{[ 56 ]} Adiponectin is known to exert both paracrine and systemic effects^{[ 57} , ^{58 ]} and is primarily secreted by adipose tissue, where it plays a key role in glucose metabolism and fatty acid oxidation.^{[ 57 ]} Adiponectin has been reported to inhibit hepatic lipogenesis and promotes free fatty acid oxidation, helping to prevent liver steatosis and insulin resistance.^{[ 58 ]}

Unlike existing phototherapies for obesity, which primarily rely on adipocyte apoptosis and photodynamic oxidative stress, ARNP‐H introduces a new approach by activating CMA‐mediated lipolysis and stimulating the expression of browning‐related genes, all at mild temperatures. Optimizing the dosing regimen is crucial for clinical translation, particularly when compared with current clinical protocols. Enhancing factors such as light penetration, photothermal efficiency, and the ease of administering photothermal treatments may benefit from integrating additional methodologies. These findings emphasize the complexity of lipid metabolism and the diverse roles of CMA across different tissues, highlighting the need for further exploration of the underlying mechanisms.

In conclusion, ARNP‐H effectively induced lipolysis and an anti‐obesity effect through the photoactivation of CMA. Using an HFD‐induced obese mouse model, we demonstrated the anti‐obesity potential of ARNP‐H, driven by CMA‐mediated lipolysis. The use of adipocyte membrane coating presents a versatile platform for delivering various chemical and biological agents directly to adipocytes. Furthermore, the potential of ARNP‐H to photoactivate CMA may pave the way for novel approaches in treating diseases associated with CMA deficiency, such as neurodegenerative disorders, kidney diseases, aging, and diabetes.

4. Conclusion

In conclusion, the ARNP‐H hydrogel system represents a promising approach for adipocyte targeted anti‐obesity therapy. This strategy effectively activates CMA in adipocytes, thereby promoting lipolysis and reducing adipocyte size through the lysosomal degradation of PLIN2. The incorporation of adipocyte membrane into the nanoparticle system significantly enhances delivery efficiency and therapeutic outcomes, leading to a marked upregulation of key lipolytic proteins, such as HSC70 and ATGL, which in turn stimulate energy expenditure and fat breakdown. Furthermore, the use of adipocyte membrane coatings provides a versatile platform for the targeted delivery of various chemical and biological agents directly to adipocytes. Although we have utilized ARNP‐H for anti‐obesity treatment, the ability of ARNP‐H to photoactivate CMA holds potential for novel therapeutic strategies in diseases associated with CMA dysfunction, including neurodegenerative disorders, kidney diseases, aging, and diabetes.

5. Experimental Section

Cell Culture

To induce adipocyte differentiation, murine 3T3‐L1 preadipocytes (Korean Cell Line Bank, Seoul, Republic of Korea) were incubated in a differentiation medium. This medium comprised DMEM enriched with 10% fetal bovine serum (FBS, GenDEPOT, Barker, TX, USA), streptomycin (100 mg mL⁻¹, Capricom, Santa Clara, CA, USA), penicillin (100 units mL⁻¹, Capricom), dexamethasone (1.0 µM, Thermo Fisher Scientific Inc., Waltham, MA, USA), methyl isobutyl xanthine (0.5 mM, Thermo Fisher Scientific), and insulin (1 µg mL⁻¹, Sigma–Aldrich, St. Louis, MO, USA; cat. No. I6634; prepared at 10 mg mL⁻¹ in 25 mM HEPES, pH 8.2).

BMDMs were generated from monocytes isolated from the femurs and tibias of 5‐week‐old C57BL/6 mice.^{[ 11 ]} These monocytes were cultured for 7 days in Iscove's modified Dulbecco's medium, which was supplemented with 10% fetal bovine serum (FBS), streptomycin (100 mg mL⁻¹), penicillin (100 units mL⁻¹), recombinant mouse macrophage colony‐stimulating factor (20 ng mL⁻¹, GenScript, Piscataway, NJ, USA), and β‐mercaptoethanol (50 µM, Sigma–Aldrich).

BMDCs were from monocytes collected from the femurs and tibiae. Following red blood cell lysis, the monocyte pellet was resuspended in Iscove's modified Dulbecco's medium containing 10% FBS, penicillin (100 units mL⁻¹), streptomycin (100 mg mL⁻¹), recombinant mouse granulocyte‐macrophage colony‐stimulating factor (20 ng mL⁻¹, GenScript), recombinant mouse interleukin‐4 (20 ng mL⁻¹, GenScript), and β‐mercaptoethanol (50 µM, Sigma–Aldrich). The culture was maintained for 7 days, with medium refreshment every 3 days.

T cells were isolated from C57BL/6 mouse spleens using slight modifications to the previously described method.^{[ 59 ]} Spleens were harvested and gently passed through 70 µm cell strainers (SPL Life Sciences, Pocheon‐si, Republic of Korea) to obtain single‐cell suspensions. Red blood cells were lysed using ACK lysis buffer (Gibco), followed by centrifugation to isolate splenocytes. To purify T cells, splenocytes were loaded onto a nylon column (Polysciences, Inc., Warrington, PA, USA) and incubated at 37 °C for 1 h. After incubation, T cells were eluted from the nylon column.

Isolation of Cell Membrane

The cell membrane was isolated from adipocytes following previously reported methods.^{[ 60 ]} Adipocytes were mechanically homogenized in a buffer supplemented with a protease inhibitor cocktail (Roche, Basel, Switzerland), a phosphatase inhibitor cocktail (Sigma–Aldrich), and 0.5 mM ethylene glycol bis (beta‐aminoethyl ether)‐N,N,N“,N”‐tetraacetic acid (Sigma–Aldrich). After centrifugation at 10 000 × g for 25 min, the supernatant was ultracentrifuged at 150 000 × g for 35 min. The resulting pellet was resuspended in RNase/DNase‐free water containing 0.2 mM ethylenediaminetetraacetic acid (Sigma–Aldrich) and stored at −80 °C until use. This isolation procedure was similarly applied to NIH‐3T3 and erythrocyte membranes. Membrane protein concentration was quantified using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific).

Preparation of Nanoparticles Dispersed in Hydrogel

To construct ARNP, RNP was synthesized through self‐polymerization of dopamine, and further coated with adipocyte membranes. In RNP, RG was loaded to PNP during self‐polymerization process.^{[ 12 ]} The RNP pellet, post‐centrifugation, was resuspended in sodium bicarbonate buffer (pH 8.3). Adipocyte membrane and RNP suspension were combined at a cell membrane protein/particle weight ratio of 1:20 and sonicated using Power sonic 410 bath sonicator (Hwashin Technology, Seoul, Republic of Korea) for 5 s, followed by cooling in ice for 10 s. This process was repeated 20 times, and the resulting ARNP was collected by centrifugation, and resuspended for further use.

Nanoparticles were incorporated into a hydrogel based on HAT. For HAT synthesis, 100 mg of hyaluronic acid (1000 kDa) was reacted for 12 h with 100 mg of 1‐ethyl‐3‐(3‐dimethylaminopropyl) carbodiimide hydrochloride (TCI chemicals, Tokyo, Japan), 60 mg of N‐hydroxysuccinimide (NHS) (Sigma–Aldrich), and 225 mg of tyramine hydrochloride at pH 5.0. After dialysis for 2 days, the lyophilized HAT conjugates were stored at 4 °C until reconstitution.

To prepare HAT‐based hydrogels, equal volumes of HAT conjugates (2 mg mL⁻¹) and phosphate‐buffered saline (PBS) were mixed, followed by the addition of horseradish peroxidase (6.4 µg mL⁻¹) (Sigma–Aldrich) and hydrogen peroxide (0.5 µg mL⁻¹) (Sigma–Aldrich). For nanoparticle‐loaded hydrogels, an equivalent method was applied, where equal volumes of HAT (2 mg mL⁻¹) and a nanoparticle suspension in PBS (1 mg mL⁻¹) were combined, followed by the addition of horseradish peroxidase and hydrogen peroxide.

Characterization Study of Nanoparticles

Nanoparticles were characterized by their size, zeta potential, morphology, and chemical composition. Size was measured using dynamic light scattering at a 90° angle at 24.1 °C with a HeNe laser (10 mW) (ELS8000, Photal, Osaka, Japan). Zeta potential was monitored via laser Doppler microelectrophoresis (ELS8000, Photal). Scanning transmission electron microscopy (TEM) with a high‐angle annular dark field detector was used for imaging. Nanoparticles (0.2 mg mL⁻¹) were placed on a Formvar/Carbon 300 mesh copper grid with a 63 µm hole size (TED PELLA Inc., Redding, CA, USA; cat. No.01753‐F) and allowed to dry.

EDS‐TEM was conducted to observe morphology and perform elemental analysis using the Themis S TEM (Thermo Fisher Scientific). SEM imaging of the nanoparticles was performed using the ZEISS Auriga Crossbeam system (Carl Zeiss AG, Oberkochen, Germany). The crystalline and amorphous structures of the nanoparticles, along with their diffraction patterns, were visualized using the High‐Resolution TEM (JEM‐2200FS with Image Cs‐corrector, JEOL Ltd., Tokyo, Japan).

To measure photothermal properties, samples were centrifuged into a pellet and irradiated with an NIR diode laser (BWT Beijing Ltd., Beijing, China) at 1.5 W. Temperature changes were recorded with an infrared thermal camera (FLIR T420, FLIR System Inc., Wilsonville, OR, USA). Phosphate content in the nanoparticles was measured via a phosphate assay as described previously.^{[ 61 ]} The proteins of membrane‐coated nanoparticles were profiled by SDS‐PAGE. Phosphorus content was quantified using X‐ray photoelectron spectroscopy (XPS; AXIS‐HSi, Kratos Analytical, Manchester, UK) on lyophilized nanoparticle powder.

The loading capacity of RG in nanoparticles was determined using HPLC. To extract RG, ARNP were dispersed in methanol and shaken for 12 h. The nanoparticles were then removed by centrifugation at 27 000 × g for 20 min, and the RG‐containing supernatant was collected for analysis. To evaluate the RG release profile, ARNP were incubated for 3 days under either neutral (pH 7.4) or acidic (pH 4.5) conditions. HPLC analysis was performed using a mobile phase composed of methanol and 20 mM potassium dihydrogen phosphate buffer (78:22, v/v). Detection was conducted at 238 nm using a C18 reverse‐phase HPLC column (Phenomenex, Torrance, CA, USA).

Characterization Study of Hydrogels

To assess the physical properties of the hydrogel, tests for swelling ratio, morphology, and rheology were conducted. The swelling ratio was determined by placing 4 mg of lyophilized hydrogel in 1 ml of triple‐distilled water and measuring the weight of the swollen hydrogel at various time intervals. Rheological properties were evaluated using a rotational rheometer (DHR‐1; TA Instruments Ltd., New Castle, DE, USA) with a 3.5 cm diameter plate and a cone angle of 0.949°. Hydrogel morphology was examined by scanning electron microscopy (SEM) (AURIGA, Carl Zeiss AG, Oberkochen, Germany).

To study the release properties of the hydrogel, ARNP‐H loaded with Cy5‐labeled ARNP was incubated in DMEM or DMEM containing hyaluronidase (0.1 U mL⁻¹). Fluorescence intensity of the supernatants was measured at various time points using a Multimode Plate Reader (SG/Ensight, PerkinElmer, Waltham, MA, USA). Temperature changes in the hydrogels were monitored by irradiating them with an NIR laser for 2 min, with temperatures recorded using an infrared thermal camera (FLIR System Inc.).

Stochastic Optical Reconstruction Microscopy

Stochastic optical reconstruction microscopy (STORM) was used to visualize the adipocyte membrane, nanoparticles, and hydrogel. For STORM imaging, fluorescently labeled RNP was prepared by adding 200 µg of Cy3‐NHS to 1 ml of a 1 mg mL⁻¹ RNP suspension in NaHCO₃ buffer (pH 8.3). The mixture was continuously stirred for 1 h, and Cy3‐labeled RNP was purified by centrifugation. To generate Cy3‐labeled ARNP, the Cy3‐labeled RNP was coated with 3T3‐L1 adipocyte membranes. Adipocyte membranes and Cy3‐labeled RNP were mixed at a membrane protein‐to‐particle weight ratio of 1:20 and subjected to sonication using a Power Sonic 410 bath sonicator for 5 s, followed by cooling on ice for 10 s. This sonication‐cooling cycle was repeated 20 times, and the resulting ARNP were collected by centrifugation.

The adipocyte membrane was labeled by incubating it with an anti‐prohibitin antibody (1:100, Abcam, Cambridge, UK; cat. No. ab154589, Lot. No. GR3433896‐1) for 1 h, followed by staining with an Alexa Fluor 647‐conjugated goat anti‐rabbit IgG antibody (1:300, Abcam; cat. No. ab150083, Lot No. GR3370563‐1). Fluorescently labeled ARNP (0.2 mg mL⁻¹) was applied to a poly‐L‐lysine‐coated glass coverslip (Corning) and examined using a stochastic optical reconstruction microscope (TiA1‐N‐STORM, Nikon, Tokyo, Japan).

In certain experiments, ARNP was tagged with Cy5 as described above, while the HAT scaffold was labeled with Cy3. A 10 mg mL⁻¹ hyaluronic acid solution was reacted for 12 h with final concentrations of 52 mM 1‐ethyl‐3‐(3‐dimethylaminopropyl) carbodiimide hydrochloride (TCI chemicals), 52 mM NHS (Sigma–Aldrich), 129 mM tyramine hydrochloride, and 13 mM Cy3‐NH2. The resulting hydrogel was then observed using the TiA1‐N‐STORM microscope.

In Vitro Cellular Uptake Assay

Cellular uptake of nanoparticles was assessed in 3T3‐L1 adipocytes, BMDCs, BMDMs, and T cells using confocal microscopy, flow cytometry, and fluorescence staining. For visualization, RNP was labeled with Cy5‐NHS (Lumiprobe, Cockeysville, MD, USA; cat. No. 13 020) by incubating 1 mg RNP with 300 µg Cy5‐NHS in 1 ml NaHCO3 (0.1 M, pH 8.3) for 1 h. The Cy5‐labeled RNP was then coated with 3T3‐L1 adipocyte cell membranes at a membrane protein‐to‐particle weight ratio of 1:20 to form Cy5‐labeled ARNP. The coating process involved 20 cycles of sonication (Power Sonic 410 bath sonicator; 5 s) alternating with cooling on ice (10 s). The Cy5‐labeled ARNP were isolated by centrifugation.

Murine 3T3‐L1 adipocytes, BMDCs, BMDMs, and T cells were incubated with cell membrane‐coated Cy5‐labeled nanoparticles for 2 h, stained with BODIPY (BODIPY 493/503, Thermo Fisher Scientific), and visualized using a TCS8 confocal microscope (Leica Microsystems, Wetzlar, Germany). For fluorescence‐activated cell sorting (FACS) analysis, these cell types were treated with membrane‐coated Cy5‐labeled nanoparticles (0.5 mg mL⁻¹) for 2 h, and fluorescence intensity was quantified for each cell using a BD FACSLyric flow cytometer (BD Bioscience, San Jose, CA, USA) with FACSuite software (v8.0.1, BD Bioscience).

To identify specific cell types that internalized the membrane‐coated nanoparticles, fluorescence staining was performed with BODIPY (Thermo Fisher Scientific) for 3T3‐L1 adipocytes, FITC‐conjugated anti‐CD11c antibody (1:100, BioLegend, San Diego, CA, USA; cat. No. 117 306, lot. No. B353146) for BMDCs, FITC‐conjugated anti‐F4/80 antibody (1:100, BioLegend; cat. No. 123 108, lot. No. B407719) for BMDMs, and FITC‐conjugated anti‐CD3 antibody (1:100, BioLegend; cat. No. 100 305, lot. No. B324851) for T cells.

In Vitro Lipolysis Assay

The lipolytic efficacy of various formulations was evaluated through flow cytometry, triglyceride quantification, and real‐time imaging of lipid droplets. Adipocytes derived from 3T3‐L1 cells were plated in 24‐well plates, and formulations were applied to hanging insert wells (pore size 3 µm; SPL Life Science, Heidelberg, Germany) positioned above. To simulate in vivo clearance of nanoparticles from the administration site, the initial medium was replaced with fresh medium 30 min after treatment and subsequently every 12 h for a total of 72 h. Following this treatment, the adipocytes were exposed to 2 min of irradiation from an 808 nm NIR diode laser (BWT Beijing Ltd.) at 1.5 W, maintaining a temperature of 42 °C, and were then incubated overnight in maintenance medium. In certain tests, 3T3‐L1 adipocytes were subjected to varying temperature exposures.

The cells were subsequently labeled with an Alexa Fluor 647‐conjugated anti‐PLIN1 antibody (1:100, Cell Signaling Technology, Danvers, MA, USA; cat. No. 52975S, lot. No. 1) for 1 h, followed by staining with BODIPY (Thermo Fisher Scientific) for 15 min before flow cytometry analysis on a BD LSRFortessa X‐20 (BD Biosciences). For triglyceride analysis, NIR‐treated cells were lysed in 10% Nonidet P40 substitute (Sigma–Aldrich) and heated at 90 °C for 20 min. Triglyceride levels were measured using a colorimetric/fluorometric kit (Sigma–Aldrich, cat. No. MAK266). To monitor lipolysis, NIR‐treated cell pellets were placed in a 96‐well plate and imaged in real‐time using an Operetta High‐Content Imaging system (PerkinElmer), recording immediately after NIR exposure for a duration of 15 h.

Cell Viability Test

Cell viability was evaluated using flow cytometry and the WST‐1 assay. 3T3‐L1 adipocytes were subjected to 20 min of temperature variation treatment, followed by overnight incubation. These cells were then stained with annexin V and propidium iodide for flow cytometry analysis. For the WST‐1 assay, heat‐stressed adipocytes were treated with WST‐1 reagent, incubated for 4 h, and the absorbance of the supernatant was measured at 450 nm using a SpectraMax M5 plate reader (Molecular Devices, San Jose, CA, USA). Additionally, adipocytes treated with various formulations for 3 days with consecutive medium replacement, were prepared as single‐cell suspensions and subjected to centrifugation. The resulting cell pellets were irradiated for 2 min with an 808 nm NIR diode laser (BWT Beijing Ltd.) at 1.5 W. After 24 h of incubation, cells were treated with WST‐1 for 4 h, and absorbance was measured using the SpectraMax M5 plate reader.

Assay of CMA Activation

CMA activation via the photothermal effect was tested with confocal microscopy and flow cytometry. 3T3‐L1 adipocytes were treated with various formulations for 3 days, with medium replacement at 30 min after treatment and subsequently exchanged every 12 h, and were then irradiated for 2 min using an 808 nm NIR laser. Post‐irradiation, cells underwent a 20 min incubation, fixation with True‐Nuclear Transcription Factor buffer set (BioLegend), and staining with anti‐HSC70 (1:100, Abcam; cat. No. ab51052, lot. No. GR336823‐4), Alexa Fluor 647‐conjugated anti‐PLIN1 (1:100, Cell Signaling Technology), anti‐PLIN2 (1:200, PROGEN Biotechnik GmbH, Heidelberg, Germany; cat. No. GP40, lot. No. FAK21107‐01), and anti‐LAMP1 antibodies (1:100, Abcam; cat. No. ab24170, lot. No. GR342889‐1). Secondary antibodies used were Alexa Fluor 647‐conjugated goat anti‐rabbit IgG (1:300, Abcam; cat. No. ab150083, Lot No. GR3415075‐1) and Alexa Fluor 568‐conjugated anti‐guinea pig IgG (1:300, Abcam; cat. No. ab175714, Lot. No. GR3413420‐1), followed by analysis with a BD LSRFortessa X‐20 (BD Biosciences).

Colocalization of CMA markers was quantitatively evaluated using a TCS8 confocal microscope (Leica Microsystems), with “JACoP” plugin in ImageJ for image analysis. BODIPY (Thermo Fisher Scientific) stained lipid droplets for 15 min, examining colocalizations such as HSC70 with PLIN2, PLIN2 with LAMP1, and ATGL with lipid droplets.

For autophagy inhibition studies, 3T3‐L1 adipocytes were treated with leupeptin hemisulfate (100 µM, Selleckchem; cat. No. S7380) and ammonium chloride (20 mM, Merck; cat. No. 0 9718) for 16 h, followed by a 2 h treatment with various formulations. Cell pellets were then exposed to a 20 min NIR laser irradiation at 808 nm, maintaining 42 °C, monitored by an infrared thermal camera (FLIR System Inc.). Lipolysis monitoring in 3T3‐L1 adipocytes via flow cytometry involved resuspension, a 24 h incubation, and staining with BODIPY and Alexa Fluor 647‐conjugated anti‐PLIN1 antibody (1:100, Cell Signaling Technology).

Measurement of Browning and Lipase Expression

The upregulation of UCP1 and ATGL was assessed using FACS analysis. 3T3‐L1 adipocytes were treated with various formulations for 72 h, with consecutive medium exchanges. Adipocytes were fixed and stained with anti‐UCP1 antibody (1:100, Cell Signaling Technology; cat. No. 72298S) and anti‐ATGL antibody (1:100, Cell Signaling Technology; cat. No. 2439S). The cells were then incubated with Alexa Fluor 647‐conjugated goat anti‐rabbit IgG antibody (1:300, Abcam) for analysis on a BD LSRFortessa X‐20 cell analyzer (BD Biosciences).

Animal Model

The in vivo anti‐obesity efficacy was evaluated using a HFD animal model. Six‐week‐old wild type male C57BL/6 mice were obtained from Raon bio (Yongin, Republic of Korea). The HFD model was established by feeding the mice with a Rodent Diet containing 60 kcal% Fat (D12492, Research Diets, New Brunswick, NJ, USA) for 6 weeks. All animal care and experimental procedures were performed following the Guidelines for the Care and Use of Laboratory Animals of the Institute of Laboratory Animal Resources, Seoul National University (Seoul, Republic of Korea; approved animal experimental protocol number, SNU‐220227‐1).

Cellular Distribution Study of Nanoparticles in ingWAT

The distribution of RNP, 3T3‐L1 adipocyte membrane‐coated ARNP, and primary adipocyte membrane‐coated nanoparticles among various cell types in ingWAT was assessed. Primary adipocytes, dendritic cells, macrophages, neutrophils, and T cells were isolated from ingWAT. ARNP was labeled with Cy5‐NHS. To prepare Cy5‐labeled RNP, 1 mg of RNP was mixed with 300 µg of Cy5‐NHS (Lumiprobe) in 1 ml of sodium bicarbonate buffer (0.1 M NaHCO₃, pH 8.3) and stirred for 1 h. The labeled RNP was then coated with cell membranes. Cy5‐labeled RNP, 3T3‐L1 ARNP, and Primary ARNP were locally injected into the ingWAT of mice. After 24 h, adipocytes and immune cells were isolated as previously described.^{[ 62 ]}

The weight of excised ingWAT was recorded before enzymatic digestion. The tissue was minced and incubated in a collagenase solution (PBS with 1% bovine serum albumin and 0.2% collagenase) at 37 °C on a shaker for 1 h. Following digestion, the sample was centrifuged at 300 × g for 5 min at 22 °C to separate cellular fractions. Mature adipocytes were collected from the top layer of the centrifuge tube. The remaining infranatant was centrifuged at 500 × g for 5 min at 4 °C to obtain the immune cell pellet.

Fluorescence intensity was quantified for each cell using a BD FACSLyric flow cytometer (BD Biosciences, San Jose, CA, USA) with FACSuite software (v8.0.1, BD Biosciences). Fluorescence staining was performed using BODIPY (Thermo Fisher Scientific) for primary adipocytes. Dendritic cells were stained with an FITC‐conjugated anti‐CD11c antibody (1:100, BioLegend; cat. no. 117 306, lot no. B353146). Macrophages were labeled with an FITC‐conjugated anti‐F4/80 antibody (1:100, BioLegend; cat. no. 123 108, lot no. B407719). T cells were stained with an FITC‐conjugated anti‐CD3 antibody (1:100, BioLegend; cat. no. 100 305, lot no. B324851). Neutrophils were labeled with an FITC‐conjugated anti‐Ly6G antibody (1:100, BioLegend).

In Vivo Lipolytic Efficacy Assessment

The lipolytic efficacy of diverse formulations was assessed in HFD mice through tissue analysis and triglyceride content measurement. Mice received local injections into ingWAT at four sites weekly for 5 weeks. They were exposed to an 808 nm NIR laser for 2 min at the injection sites thrice weekly. During each NIR session, spot temperatures were recorded using an infrared thermal camera (FLIR System Inc.), and body weights were monitored. CT imaging utilized a Quantum FX micro‐CT scanner (PerkinElmer), with image segmentation conducted semi‐automatically via Volume Edit tools in Analyze 14.0 (AnalyzeDirect, Overland Park, KS, USA) to calculate subcutaneous and vWAT volumes.

Post 5 weeks from the initial NIR treatment, mice were euthanized; brown adipose tissue (BAT), vWAT, and ingWAT were harvested and weighed. Tissue sections underwent Hematoxylin and Eosin (H&E) staining, analyzed by a Vectra tissue analysis system (PerkinElmer), with lipid droplet sizes quantified using InForm v2.4.11 software (PerkinElmer). Triglyceride levels in adipose tissue homogenates were determined using a triglyceride quantification kit (Sigma‐Aldrich). Lipolysis marker expression in ingWAT was assessed via immunohistochemistry.

For epitope retrieval, tissues were treated with Target Retrieval Solution (pH 6) (Agilent Dako, Santa Clara, CA, USA) and stained with antibodies against UCP1, ATGL, HSC70 (all from Cell Signaling Technology), PLIN2 (PROGEN Biotechnik GmbH), and PLIN1. Secondary antibodies included Alexa Fluor 647‐conjugated goat anti‐rabbit IgG, Alexa Fluor 568‐conjugated anti‐guinea pig IgG, and Alexa Fluor 488‐conjugated goat anti‐rabbit IgG (all from Abcam). Imaging was performed using a Thunder imager 3D assay (Leica Microsystems).

Measurement of Adiponectin Levels

Plasma adiponectin levels were measured in HFD mice following treatment with various formulations and three repeated NIR irradiations. Blood samples were collected, and plasma was isolated. Adiponectin levels were quantified using an ELISA kit (R&D Systems; cat. no. PMRP300).

Food Intake Analysis

HFD mice were treated with ARNP‐H, followed by NIR laser irradiation two days later. The mice were individually housed in metabolic cages, and their physiological and metabolic parameters were assessed using the Comprehensive Lab Animal Monitoring System (CLAMS; Columbus Instruments, OH, USA). After body weight measurement, food intake was continuously monitored for 48 h under alternating light and dark cycles.

In Vivo Test of Biochemical Markers

ARNP‐H were administered into ingWAT of mice at 2 sites bilaterally. NIR was irradiated 3 times on day 1, 3, 5 from the mouse injected with ARNP‐H. Serum was isolated from collected blood by centrifuging 3000 g for 30 min at 4 °C. The level of ALP, ALT, AST, CRE, BUN, CKMB in the serum were determined by DRI‐CHEM 3500S chemistry analyzer (Fujifilm, Tokyo, Japan).

Evaluation of CMA in Fat Tissue Lysosomes

HFD mice received localized injections into their ingWAT with various formulations. After 48 h, subcutaneous leupeptin injections (2 mg per 100 g body weight) were administered. Four h post‐leupeptin, mice underwent 808 nm NIR laser irradiation, keeping the temperature at 42 ± 1 °C. The ingWAT was collected 2 h post‐NIR exposure. Adipocyte lysosomes from these tissues were isolated following a previously described method.^{[ 63 ]} Tissue homogenates were filtered and centrifuged, with lysosomes isolated from Percoll gradients for CMA activation assessment.

Lysosomal PLIN2 expression was analyzed via western blot. For immunostaining, lysosomes were labeled with anti‐PLIN2 (1:200, PROGEN Biotechnik GmbH) and anti‐LAMP1 (1:100, Abcam) antibodies, followed by staining with Alexa Fluor 488‐conjugated goat anti‐rabbit IgG (1:300, Abcam; cat. No. ab150077, lot. No. GR324468‐1) and Alexa Fluor 568‐conjugated anti‐guinea pig IgG (1:300, Abcam), and examined through flow cytometry and confocal microscopy. Whole‐mount adipose tissue PLIN2 expression was assessed using a THUNDER imager 3D assay (Leica Microsystems).

Statistical Analysis

Statistical comparisons between two groups utilized Student's t‐tests, while one‐way ANOVA with Tukey's post hoc test was applied for multiple group comparisons. Analyses were performed using GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA), with a p‐value of less than 0.05 indicating significance.

Conflict of Interest

The authors declare no conflict of interest.

Supporting information

Supporting Information

Supplemental Video 1

Supplemental Video 2

Choi J., Byun J., Kim D., et al. “Selective Lipolysis by Photoactivation of Chaperone‐Mediated Autophagy Using Adipocyte Membrane‐Coated Nanoparticle in Hydrogel.” Adv. Mater. 37, no. 49 (2025): e18445. 10.1002/adma.202418445

Data Availability Statement

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