Synthetic cleavage-resistant TREM2 boosts macrophage efferocytosis to treat inflammatory diseases

Summary

Triggering receptor expressed on myeloid cells 2 (TREM2), a critical sensor of cell debris, regulates macrophage efferocytosis to maintain tissue immune homeostasis. However, inflammatory mediators upregulate the sheddase ADAM17, leading to TREM2 cleavage, which impairs apoptotic cell clearance and exacerbates inflammation. We here report a synthetic cleavage-resistant TREM2 (CRT) to boost TREM2-dependent efferocytosis and alleviate inflammation associated with aberrantly accumulated apoptotic cells. CRT integrates the ligand-binding domain of TREM2 with its intracellular signaling adaptor DAP12 via a custom-engineered stalk and transmembrane segment. Our data demonstrate that CRT amplifies TREM2 signaling even in the presence of ADAM17. Customized lipid nanoparticles efficiently introduce CRT mRNA into macrophages, generating CRT-engineered macrophages (CRT-Ms) in situ. CRT-Ms effectively reduce apoptotic cell burden and alleviate inflammation in mouse models of metabolic-dysfunction-associated steatohepatitis and atherosclerosis. In sum, our findings establish that CRT strengthens TREM2-mediated macrophage efferocytosis and mitigates inflammation, with broad potential for apoptotic-cell-associated diseases.

Graphical abstract

Highlights

• •A cleavage-resistant TREM2 receptor (CRT) restores efferocytosis under inflammation
• •CRT resists ADAM17-mediated shedding and amplifies intracellular signaling
• •Phosphatidylserine-functionalized LNPs enable macrophage-selective mRNA delivery
• •In situ generated CRT macrophages reduce inflammation in MASH and atherosclerosis

To restore TREM2 signaling during inflammation, Dong et al. engineer a cleavage-resistant synthetic TREM2 receptor (CRT). In situ generated CRT macrophages via the LNP-mRNA system resist ADAM17 proteolysis, amplify TREM2 signaling, enhance efferocytosis, and reduce apoptotic cell accumulation and inflammation in mouse models of MASH and atherosclerosis.

Introduction

Efficient macrophage-mediated apoptotic cell clearance is a critical process for maintaining tissue homeostasis.^1,2 Triggering receptor expressed on myeloid cells 2 (TREM2) acts as a cell debris sensor,^3,4,5,6 orchestrating efferocytosis of apoptotic cells to prevent secondary necrosis and mitigate inflammation.^7,8,9,10,11 Upon binding to ligands such as aminophospholipids exposed on apoptotic cells, TREM2 induces activation of the DAP12 immunoreceptor tyrosine-based activation motif (ITAM) to initiate downstream signaling pathways that drive phagocytosis and inflammatory resolution.^12,13,14,15 TREM2 impairment compromises macrophage phagocytosis, skews them toward a pro-inflammatory phenotype, and results in the pathological accumulation of apoptotic cells and sustained inflammation.^{14,16,17,18,19,20,21} In Alzheimer's disease (AD) models, elevating TREM2 in microglia enhanced amyloid phagocytosis, reduced neuroinflammation, and improved cognitive performance, underscoring its critical role in neural homeostasis.^22,23 In peripheral inflammatory settings such as sepsis, myocardial infarction, and atherosclerosis, TREM2 upregulation promotes macrophage efferocytosis, attenuates tissue injury, and improves overall outcomes.^{17,20,24,25,26} Collectively, these studies highlight the broad protective functions of TREM2 signaling and illustrate that targeting and modulating TREM2 signaling in macrophages presents a promising advancement for inflammation.

However, the therapeutic potential of TREM2 is limited by its susceptibility to proteolytic cleavage. Kleinberger et al. identified TREM2 as a substrate of the sheddases ADAM10 and ADAM17 and demonstrated that inhibition of this proteolysis enhances macrophage phagocytosis.²⁷ Under inflammatory conditions, ADAM17 is upregulated, cleaving TREM2 within the stalk region and releasing soluble TREM2 (sTREM2) while leaving a truncated C-terminal fragment.^10,28,29,30 This proteolytic process disrupts sustained TREM2 signaling and has been linked to impaired efferocytosis and persistent inflammation. Elevated circulating sTREM2 levels have been reported in multiple inflammatory diseases,^{10,20,31,32,33,34,35} although their interpretation remains complex, as increased sTREM2 may reflect both enhanced receptor expression and concomitant loss of membrane-bound TREM2 through cleavage.²⁸ These observations emphasize the importance of approaches that preserve TREM2 signaling in macrophages to maintain their phagocytic function and promote effective inflammatory resolution.

Several agonistic antibodies targeting TREM2 have been developed to stabilize the receptor and prolong signaling by limiting proteolytic shedding, such as AL002 and 4D9.^36,37 These antibodies target the stalk region of TREM2, where they block proteolytic shedding and stabilize the receptor at the cell surface, thereby prolonging TREM2 signaling and enhancing macrophage phagocytosis.^20,38 However, the phase II clinical trial of AL002 failed due to limited efficacy,³⁹ highlighting the restricted applicability and insufficient function of current antibody-based strategies. These limitations underscore the urgent need for more precise, effectual approaches to unlock the therapeutic potential of TREM2 in inflammatory diseases. Synthetic receptors enable precise manipulation of natural receptor signaling via designing modular protein domains that seamlessly integrate ligand-binding and effector functions.^40,41,42 Thus, we hypothesize that a rationally engineered TREM2-based synthetic receptor that integrates the ligand-binding domains of TREM2 with its downstream signaling adaptor may prevent sheddase-driven cleavage, thereby enhancing efferocytosis during inflammation.

Here, we generated a cleavage-resistant TREM2 receptor (CRT) and evaluated its function in murine models of metabolic-dysfunction-associated steatohepatitis (MASH) and atherosclerosis. Specifically, we designed CRT by fusing the ligand-binding domain of TREM2 with the intracellular signaling domain of DAP12 via a custom-engineered stalk and transmembrane segment. We demonstrated that CRT could prevent proteolytic shedding and amplify TREM2/DAP12 signaling by eliminating the sheddase cleavage site and bypassing the conventional adaptor recruitment process, respectively. Moreover, we employed “eat me” phosphatidylserine-functionalized lipid nanoparticles (pLNPs) to optimize macrophage uptake of CRT mRNA and generate CRT-engineered macrophages (CRT-Ms). In HFMCD- and HFD-CCl₄-induced MASH models, as well as in high-fat diet (HFD)-fed LDLr-KO atherosclerosis mice, tailored surface-modified pLNPs successfully delivered CRT mRNA to macrophages within fatty livers and atherosclerotic plaques of model mice. CRT-Ms exhibited superior efferocytic capacity, reduced the burden of apoptotic cells, and polarized to the anti-inflammatory phenotype, effectively mitigating inflammation. These findings establish that CRT amplifies TREM2-mediated signaling to promote efferocytosis and effectively attenuates inflammation in apoptotic-cell-associated diseases.

Results

Design and screening of synthetic cleavage-resistant TREM2

TREM2 consists of an immunoglobulin (Ig)-like domain, a flexible stalk region, a transmembrane helix, and a short intracellular tail that requires recruitment of the adaptor protein DAP12 to initiate intracellular signaling. ADAM17 cleaves TREM2 within the stalk region, resulting in the loss of the ligand-binding domain and subsequent functional impairment (Figure 1A). To overcome this limitation, we tend to develop a synthetic cleavage-resistant TREM2 (CRT) by linking the ligand-binding domain of TREM2 to the ITAM-containing intracellular domain of DAP12 through engineered stalk and transmembrane regions, enabling direct activation of downstream signaling upon ligand engagement. Four CRT variants (types I–IV) were engineered using distinct stalk and transmembrane segments (Figures 1B and 1C). The type I variant, designed as an initial prototype, retains the entire extracellular domain of TREM2 (amino acids 19–174) and links it directly to the transmembrane and intracellular domains of DAP12. While this preserves the key domains of TREM2/DAP12 signaling pathway, it does not prevent ADAM17 cleavage at the His157-Ser158 site.²⁹ Type II variant fused the sTREM2 fragment (aa 19–157) to DAP12, eliminating the cleavage site but partially truncating the stalk region (aa 131–174),⁴³ which may affect signal transduction. Thus, type III variant was engineered by introducing an additional flexible hinge region between the sTREM2 fragment and the transmembrane domains of DAP12, refining the design of type II to enhance structural flexibility and optimize signal transduction.⁴⁴ Type IV variant simplifies the design by retaining only the Ig-like domain (aa 19–130) of TREM2 and incorporating the hinge and transmembrane regions of CD8, which are commonly utilized in chimeric antigen receptor (CAR) designs, to link to the intracellular signaling domain of DAP12.⁴⁴

Figure 1.

Construction and screening of four CRT variants

(A) Schematic showing TREM2 recruitment of DAP12 for downstream signaling and ADAM17-mediated cleavage rendering TREM2 nonfunctional.

(B and C) Domain organization and structural composition of the four CRT variants.

(D) Stable expression of each CRT variant in Jurkat NFAT-Luci reporter cells transduced with lentiviral vectors express little endogenous TREM2.

(E) Flow cytometry analysis of myc tag expression as a surrogate for each CRT variant.

(F) Western blot analysis of the expression levels of CRT variants.

(G) Coincubation of reporter cells with apoptotic cells for the specified duration.

(H) Luminescence intensity measurement following coincubation for the indicated time (n = 6).

(I) Coincubation for 10 h in the presence of different concentrations of the TREM2 antibody, followed by luminescence intensity measurement (n = 6).

(J) ELISA quantification of sTREM2 levels in culture supernatants with or without ADAM17 (n = 6).

(K) Western blot analysis of full-length TREM2 in cell lysates and sTREM2 in supernatants after PMA treatment. The data are shown as the means ± standard deviations (SDs) from biological replicates. Statistical analysis was carried out by two-way analysis of variance (ANOVA).

To assess TREM2/DAP12 signaling, CRT variants were expressed in Jurkat NFAT-Luci reporter cells, in which activation of the PKC-Ca²⁺-calcineurin pathway drives NFAT-dependent luciferase expression^8,15 (Figure 1D). Meanwhile, reporter cells expressing wild-type TREM2 and DAP12 (TREM2&DAP12) were generated as a positive control. Stable expression of the four CRT variants was initially confirmed by flow cytometry using the myc-tag as a reporter (Figure 1E). Western blot analysis further revealed the maturation status of the CRT variants (Figure 1F). Surface biotinylation followed by western blotting demonstrated the presence of mature CRT on the cell surface (Figure S1A), and deglycosylation analysis using myc-tag immunoblotting allowed accurate determination of the actual molecular weights of the CRT proteins (Figure S1B). These results demonstrated that CRT variants were stably expressed and showed proper maturation in cells. We next coincubated the reporter cells with palmitic-acid-induced apoptotic HHL-5 cells (Figure S1C) and monitored the luminescence intensity over time (Figure 1G). All CRT variants successfully activated the reporter cells, whereas the control group exhibited no luminescence. Reporter cells transfected with TREM2&DAP12 reached peak luminescence intensity at approximately 14 h, whereas CRT variants, except type II, peaked earlier and exhibited higher signal intensity. Notably, type IV group exhibited the highest peak luminescence in the shortest duration and maintained significant luminescence for up to 6 h (Figure 1H). At 10 h after coincubation, the phosphorylation levels of Syk in reporter cells paralleled the luminescence readouts, both reflecting a similar trend of TREM2/DAP12 pathway activation (Figure S1D). Reporter cells pre-incubated with TREM2-blocking antibodies exhibited a dose-dependent reduction in luminescence, confirming the critical role of TREM2 Ig-like domain in mediating CRT functionality (Figure 1I). These findings demonstrate that CRT effectively activates TREM2/DAP12 signaling upon binding to apoptotic cells, exhibiting superior signaling efficiency and intensity compared to the native TREM2-DAP12 complex. This enhanced functionality likely arises from the direct engagement of DAP12 signaling domain, bypassing the DAP12 recruitment and bridging steps required for native receptor activation.

Resistance to proteolytic cleavage was evaluated following ADAM17 exposure. Luminescence was markedly reduced in cells expressing TREM2/DAP12 or type I, accompanied by increased sTREM2 release, whereas types II–IV maintained signaling activity without detectable sTREM2 elevation (Figures 1J and S1E). Remarkably, type IV maintained the highest activation potential after ADAM17 treatment, with luminescence intensity reaching ∼3.5-fold that of TREM2&DAP12 under baseline conditions. In addition, we coincubated reporter cells with DiO-labeled acetylated low-density lipoprotein (DiO-AcLDL) to examine CRT-mediated TREM2/DAP12 signaling activation, as TREM2 is known to independently induce AcLDL uptake in heterologous cells.⁴⁵ Cells expressing CRT variants or TREM2&DAP12 showed increased DiO-AcLDL uptake. Following exposure to ADAM17, the uptake capacity of cells expressing type II, III, and IV remained unaffected, which was consistent with the observed luminescence (Figure S1F). Cells expressing type IV demonstrated consistent and enhanced Ac-LDL uptake, with phagocytosis observed in up to 60% of the cell area, regardless of ADAM17 cleavage (Figure S1G). We further treated cells with PMA, a potent inducer of ADAM17, to trigger proteolytic cleavage. Western blot analysis of cell lysates and culture supernatants revealed a pronounced reduction in full-length protein levels in reporter cells expressing type I or TREM2/DAP12, accompanied by an increase of sTREM2 in the supernatant (Figure 1K). Consistently, CRT-derived C-terminal fragments (CTFs) were readily detected in type I and TREM2/DAP12 cells after PMA stimulation, whereas little to no CTF signal was observed in types II, III, and IV (Figure S1H). ELISA quantification further confirmed that PMA markedly increased sTREM2 release in the type I and TREM2/DAP12 groups, whereas types II, III, and IV showed no significant changes (Figure S1I). These observations were consistent with the trends directly observed upon ADAM17 treatment. Collectively, type IV exhibited the most robust and sustained TREM2/DAP12 signaling through resistance to proteolytic cleavage and was selected for subsequent experiments.

“Eat me” phosphatidylserine-functionalized lipid nanoparticles efficiently introduce CRT mRNA into macrophages

Recent advances in lipid nanoparticle (LNP)-based mRNA therapeutics have enabled efficient cell engineering through transient protein expression without genomic integration.⁴⁶ To achieve selective delivery of CRT mRNA to macrophages, we established a phosphatidylserine-functionalized LNP (pLNP) platform for in situ generation of CRT-Ms. To address the limited transfection efficiency and safety concerns of conventional LNPs,⁴⁷ we developed an ionizable lipid library derived from natural amino acids, incorporating tailored structural features to optimize delivery efficiency, functionality, and biocompatibility. These lipids were built on a 2,5-diketopiperazine (DKP) scaffold, a hydrophilic core with precise stereochemistry formed by the condensation of two L-amino acids, providing a unique structure that ensures superior biocompatibility and biodegradability. Unsaturated fatty acid tails, such as oleic, elaidic, and linoleic acids, are attached to the DKP core via amide bonds to terminal amino groups, with varying linker lengths to optimize hydrophobicity (Figure 2A), aiming to enhance endosomal escape and improve mRNA delivery efficiency.⁴⁸ The ionizable lipid library was synthesized by condensing two L-pyroglutamic acid molecules to form the DKP core, followed by amide coupling to introduce linkers and reductive amination to attach unsaturated fatty acid tails (Figures S2A and S2B).

Figure 2.

Synthesis, screening, and characterization of macrophage-specific mRNA delivery systems

(A) Construction of the ionizable lipid library.

(B) pLNP formulation library.

(C) Workflow for preparation and screening of EGFP-pLNPs.

(D) Heatmap of EGFP-positive percentages in RAW264.7 cells treated with EGFP-pLNPs.

(E) Chemical structure of DKP-2-O.

(F) Molar composition of the optimized pLNP formulation.

(G) Representative confocal images of RAW264.7 cells and BMDMs treated with free CRT mRNA, CRT-LNPs, or CRT-pLNPs. Cy5-labeled CRT mRNA (red) indicates the intracellular distribution of the delivered mRNA, while the cell membrane and nuclei were stained with DIO (green) and DAPI (blue), respectively. Cy5, Cyanine 5; DIO, 3,3′-dioctadecyloxacarbocyanine perchlorate; DAPI, 2-(4-amidinophenyl)-6-indolecarbamidine dihydrochloride. Scale bars, 10 μm.

(H) Representative flow cytometry analysis of mRNA uptake.

(I) Quantification of fluorescence intensity by flow cytometry (n = 5). MFI, mean fluorescence intensity.

(J) Particle size distribution, PDI, and EE of CRT-pLNPs (n = 3).

(K) Zeta potential of CRT-pLNPs (n = 3).

(L) TEM images of CRT-pLNPs. The data are shown as the means ± SDs from biological replicates. Statistical analysis was carried out by two-way ANOVA.

All lipids were structurally characterized by proton nuclear magnetic resonance (¹H NMR) spectroscopy and further validated by high-resolution mass spectrometry (HRMS) (Data S1). To enable macrophage targeting, dioleoyl phosphatidylserine (DOPS), a well-established “eat-me” signal that promotes macrophage recognition and phagocytosis,⁴⁹ was incorporated into a conventional four-component LNP formulation, yielding a library of 150 total LNPs with 25 distinct molar ratios (Figure 2B). We next employed a rapid mixing technique to generate phosphatidylserine-functionalized LNPs encapsulating EGFP reporter genes (EGFP-pLNPs) and incubated them with RAW264.7 cells (Figure 2C). Among these formulations, R17 formulation containing DKP-2-O showed the highest transfection efficiency, achieving 84.5% EGFP-positive cells (Figures 2D–2F).

CRT mRNA-loaded pLNPs (CRT-pLNPs) were subsequently formulated using microfluidics with the optimized R17 composition. Gel electrophoresis demonstrated rapid degradation of free mRNA in serum-containing buffer, whereas CRT mRNA encapsulated in pLNPs remained intact, indicating effective protection from nuclease-mediated degradation (Figure S2C). Cy5-labeled CRT mRNA encapsulated within pLNPs exhibited efficient uptake by RAW264.7 cells within 3 h, showing significantly higher internalization compared with DOPS-free LNPs (CRT-LNPs) (Figure 2G). To further validate the suitability of pLNPs for in vivo applications, bone-marrow-derived macrophages (BMDMs) were isolated from mice and employed for uptake analysis (Figure S2D). Confocal microscopy confirmed pronounced intracellular localization of CRT-pLNPs within BMDMs (Figure 2G). Consistently, flow cytometry analysis revealed markedly higher mean fluorescence intensity (MFI) in RAW264.7 cells and BMDMs treated with CRT-pLNPs, highlighting that the pLNP delivery system efficiently promoted cellular uptake (Figures 2H and 2I). Physicochemical characterization revealed that CRT-pLNPs exhibited an encapsulation efficiency (EE) of 87.3% ± 4.6%, a mean hydrodynamic diameter of 114.7 ± 5.8 nm with a polydispersity index (PDI) of 0.18 ± 0.05, and a slightly negative zeta potential (−3.82 ± 1.46 mV) attributable to DOPS incorporation (Figures 2J and 2K). Transmission electron microscopy (TEM) analysis provided further validation, revealing well-defined, spherical nanoparticles with an average diameter of approximately 110 nm, consistent with the dynamic light scattering measurements (Figure 2L). Collectively, the pLNP delivery system on the basis of DKP ionizable lipids demonstrates favorable physicochemical characteristics and superior targeting efficiency, significantly enhancing the selective delivery of CRT mRNA to macrophages.

CRT enhances macrophage efferocytosis and inflammatory resolution in vitro

To evaluate CRT function in macrophages, BMDMs were isolated from TREM2-KO C57BL/6 mice to exclude interference from endogenous TREM2 (Figure S3A). Following incubation with CRT-pLNPs, Cy5-labeled CRT mRNA efficiently escaped lysosomes and localized in the cytoplasm (Figures S3B and S3C). Flow-cytometric analysis further revealed a time-dependent increase in CRT protein surface expression following incubation with CRT-pLNPs, confirming progressive translation and membrane presentation of CRT (Figure S3D). After 12 h of incubation with CRT-pLNPs, flow cytometry analysis showed robust myc-tagged CRT expression in macrophages (Figures 3A and 3B). Confocal microscopy and immunoblotting further confirmed membrane localization, expression, and maturation of CRT in macrophages (Figures 3C and 3D). These findings demonstrate that pLNPs effectively deliver CRT mRNA into macrophages, generating CRT-Ms in vitro (Figure 3E). Macrophages were then cocultured with palmitate-induced apoptotic AML12 cells, and the engulfment capacity was quantitatively evaluated via fluorescence confocal microscope. We observed a significant reduction in phagocytic efficiency in TREM2-KO macrophages, while CRT-Ms exhibited markedly improved apoptotic cell engulfment efficiency, even surpassing that of control macrophages (Figures 3F and 3G). To model inflammatory suppression of efferocytosis, we pretreated macrophages with interleukin (IL)-1β, a cytokine known to trigger ADAM17-mediated TREM2 shedding (Figure S3E). This resulted in a notable decrease in apoptotic cell uptake in control macrophages, which dropped to levels similar to those observed in TREM2-KO macrophages. In contrast, CRT-Ms remained largely unaffected (Figures 3F and 3G). Flow cytometry further supported these findings, demonstrating a significant decline in phagocytosis in control macrophages following IL-1β treatment, while CRT-Ms maintained stable efferocytic activity (Figures 3H and 3I). Western blot analysis further confirmed that CRT protein levels in CRT-Ms remained essentially unchanged, and sTREM2 in the culture supernatant was nearly undetectable following IL-1β or PMA treatment, whereas endogenous TREM2 in control macrophages was markedly reduced, accompanied by a pronounced increase of sTREM2 in the supernatant (Figures 3J and S3F). As TREM2 has been reported to promote macrophage proliferation and survival,⁶ we next examined whether CRT could exert a similar effect. Indeed, expression of CRT enhanced macrophage proliferation and survival (Figures 3K–3M). Together, these results demonstrate that CRT can significantly enhance the efferocytosis capacity of macrophages and preserve their function under ADAM17-mediated cleavage.

Figure 3.

In vitro generation and functional assessment of CRT-Ms

(A) Representative flow cytometry analysis of CRT expression in TREM2-KO macrophages after 12-h treatment.

(B) Quantitative analysis of CRT expression levels (n = 5).

(C) Representative confocal images showing plasma membrane localization of myc-tagged CRT (green); cell membrane stained with DiD (red); nuclei stained with DAPI (blue). DiD, 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindodicarbocyanine, 4-chlorobenzenesulfonate salt. Scale bars, 10 μm.

(D) Western blot analysis of CRT expression.

(E) CRT-pLNP-mediated delivery of CRT mRNA promotes macrophage efferocytosis. CRT-pLNPs enable efficient uptake of CRT mRNA, followed by lysosomal escape and subsequent translation into membrane-expressed CRT protein, which enhances the efferocytic activity of macrophages.

(F) Representative confocal images of macrophages coincubated with apoptotic cells. Apoptotic cells labeled with CellTracker Red; macrophages labeled with CellTracker Green. Scale bars, 50 μm.

(G) Quantification of efferocytosis based on confocal images (n = 5).

(H) Representative flow cytometry analysis of pHrodo-labeled apoptotic cell uptake.

(I) Quantitative analysis of apoptotic cell engulfment by flow cytometry (n = 5).

(J) Western blot analysis of sTREM2 in cell culture supernatants treated with IL-1β and PMA.

(K) Flow cytometry analysis of macrophage proliferation using the BeyoClick EdU-594 Cell Proliferation Detection Kit.

(L) Quantification of proliferating macrophages (n = 5).

(M) Cytotoxicity quantified by lactate dehydrogenase (LDH) released into the macrophage culture supernatant after overnight incubation (n = 5).

(N and O) Representative flow cytometry analysis of CD86 (N) and CD206 (O) expression in CRT-Ms after the indicated treatment.

(P and Q) Quantification of CD86-positive (P) and CD206-positive (Q) cells (n = 5).

(R) Quantification of macrophage polarization by flow cytometry. The proportions of CD86-positive and CD206-positive cells were used as markers to represent M1 and M2 macrophages, respectively (n = 5).

(S–U) IL-6 (S), IFN-γ (T), and TNF-α (U) levels in cell culture supernatants (n = 5). The data are shown as the means ± SDs from biological replicates. Statistical analysis was carried out by one-way ANOVA in (B), (P), (Q), (R), (S), (T), and (U); two-way ANOVA in (G) and (I); and an unpaired two-tailed Student’s t test in (L) and (M).

Given that TREM2 signaling promotes anti-inflammatory polarization of macrophages,^18,19,50,51 we next interrogated whether CRT-Ms could resolve inflammation through phenotypic reprogramming. To simulate the inflammatory conditions of macrophages, we induced the pro-inflammatory M1 phenotype in macrophages. After treatment with lipopolysaccharide (LPS), macrophages predominantly maintained a pro-inflammatory phenotype, as evidenced by high expression of the M1 marker CD86 and low levels of the M2 marker CD206. However, when transfected with CRT-pLNPs, CRT-Ms showed a decrease in CD86 expression and an increase in CD206 expression (Figure S3G). The phenotype shifting in CRT-Ms was further quantified by flow cytometry (Figures 3N and 3O), showing a reduction in CD86 expression from 62.7% ± 10.6% to 30.6% ± 3.8% and an increase in CD206 expression from 2.5% ± 1.0% to 15.8% ± 4.4% (Figures 3P and 3Q), leading to a significant reduction in the M1/M2 ratio (Figure 3R). Additionally, ELISA assays revealed a decrease in pro-inflammatory cytokines, including, IL-6, interferon gamma (IFN-γ), IL-1β, and tumor necrosis factor alpha (TNF-α), in macrophage supernatants in CRT-pLNP-treated macrophages. (Figures 3S–3U and S3H).

To further assess the application potential of CRT, we validated its function in the human monocytic cell line THP-1. Our pLNPs efficiently released CRT mRNA from lysosomes into the cytoplasm within 6 h (Figures S3I and S3J) and mediated robust CRT expression by 12 h of incubation (Figures S3K and S3L). Consistent with the findings in BMDMs, flow cytometry and confocal microscopy demonstrated that TREM2-KO THP-1 cells exhibited markedly impaired efferocytosis, which was restored upon CRT expression (Figures S3M–S3P). Moreover, CRT expression increased the proportion of CD206⁺ subsets and reduced the frequency of CD86⁺ subsets in LPS-stimulated THP-1 cells (Figures S3Q–S3U), accompanied by a marked decrease in IL-6 concentration in the culture supernatant (Figure S3V), indicating a shift toward an anti-inflammatory state. These results demonstrate that CRT potently restores efferocytosis and reinforces anti-inflammatory programming in human macrophage-like cells.

In summary, CRT delivery via pLNPs enhances apoptotic cell clearance and reduces inflammation under inflammatory conditions. Changes in CD86/CD206 expression and cytokine secretion support this, serving as indicators of reduced inflammatory activation. The main findings highlight the improved efferocytosis and maintained macrophage function, even under ADAM17-mediated cleavage.

Tailored surface-modified pLNPs effectively generate CRT-engineered macrophages in situ

To evaluate the therapeutic potential of in situ generated CRT-engineered macrophages, we investigated the application of CRT-pLNPs in mouse models of MASH and atherosclerosis, targeting the liver and aortic lesions, respectively. Biodistribution of intravenously injected DiR-labeled nanoparticles in healthy mice revealed that unmodified CRT-LNPs accumulated predominantly in the liver and spleen, whereas DOPS-functionalized CRT-pLNPs showed enhanced splenic accumulation, likely due to the negative surface charge introduced by DOPS⁵² (Figures S4A and S4B). Thus, to improve tissue-specific delivery while maintaining macrophage targeting, we further modified the surface of CRT-pLNPs with targeting ligands to change tissue distribution while preserving their macrophage-specific targeting properties.

For liver targeting in MASH, CRT-pLNPs were modified with soluble Pirb (sPirb) via a reactive oxygen species (ROS)-responsive linker.^53,54 sPirb binds ANGPTL8, which is highly expressed in the liver during MASH, enabling preferential hepatic accumulation following intravenous administration.⁵⁵ The ROS-responsive linker cleaves in the inflammatory microenvironment of the liver, exposing DOPS on the nanoparticle surface to ensure macrophage-specific targeting. Specifically, we incorporated DSPE-PEG-TK-Mal, featuring a thioketal (TK) bond that cleaves in response to ROS, into the pLNP formulation. The nanoparticles were then post-modified to conjugate sPirb with the maleimide (Mal) group⁵⁶ (Figures S4C and S4D). sPirb conjugation increased nanoparticle size and slightly elevated zeta potential, both of which reverted to baseline after ROS treatment, consistent with linker cleavage and ligand detachment (Figures S4E and S4F). Consistently, fractionation by purification column showed that, after ROS treatment, nanoparticles and sPirb eluted in distinct fractions, further validating ROS-responsive release (Figure S4G). TEM imaging showed that the modified nanoparticles maintained a uniform, spherical morphology (Figure S4H). CRT-sPirb/pLNPs exhibited good biocompatibility in BMDMs and remained stable at 4°C for up to 1 week (Figures S4I and S4J). To evaluate the in vivo distribution and mRNA delivery performance of CRT-sPirb/pLNPs within an inflammatory MASH microenvironment, C57BL/6 mice were fed a high-fat and methionine-choline-deficient (HFMCD) diet for 8 weeks to induce an MASH model prior to nanoparticle administration. Next, we intravenously administered Luci mRNA-loaded nanoparticles to MASH model mice. IVIS imaging revealed a significant increase in hepatic Luci expression in the sPirb/pLNPs group compared to the unmodified pLNP group, accompanied by a marked reduction in luminescence within the spleen (Figures S4K and S4L). In the single-dose Luci-sPirb/pLNPs treatment group, the bioluminescence signal decreased to near baseline levels by the fourth day post-administration (Figures S4M and S4N). Notably, the repeated administration of Luci-sPirb/pLNPs on the first and fourth days of the week effectively sustained elevated mRNA expression levels (Figures S4O–S4Q). Flow cytometry and immunofluorescence analyses of liver tissue 24 h post-injection demonstrated a significant upregulation of CRT expression in hepatic macrophages following the incorporation of DOPS. Moreover, the conjugation of sPirb resulted in a marked amplification of CRT expression across the liver. The combined action of DOPS and sPirb enabled pLNPs to achieve successful editing of hepatic macrophages in MASH mice (Figures S4R–S4X). Furthermore, we confirmed that inflammation-associated monocyte-derived macrophages (F4/80⁺ TIM-4^- CLEC4F⁻) recruited to the liver during MASH also achieved robust CRT expression within 48 h of nanoparticle administration (Figures S4Y and S4Z). Collectively, these findings demonstrate that CRT-sPirb/pLNPs effectively target both resident and newly recruited hepatic macrophages, enabling efficient CRT expression within the inflammatory microenvironment of MASH.

For atherosclerosis targeting, CRT-pLNPs were surface-modified with the integrin-binding peptide cRGDfk,⁵⁷ which binds ανβ3 integrin highly expressed in atherosclerotic plaques.⁵⁸ This modification significantly augmented the local accumulation of nanoparticles near the arterial plaques and facilitated their transendothelial migration, thereby enhancing their targeted infiltration into the atherosclerotic lesions. The successful conjugation of cRGDfk to DSPE-PEG-NHS was confirmed by ¹H NMR (Figures S5A and S5B). CRT-cRGDfk/pLNPs were subsequently formulated by microfluidics (Figure S5C). The modified nanoparticles exhibited appropriate size, zeta potential, uniform morphology, biocompatibility, and stability (Figures S5D–S5H). To evaluate the distribution of CRT-cRGDfk/pLNPs within atherosclerotic plaques, we intravenously administered the nanoparticles to atherosclerotic model mice and subsequently assessed Luci expression in aortic tissue collected from mice 24 h post-injection. Regardless of DOPS incorporation, cRGDfk-modified nanoparticles showed significantly improved Luci expression within the aorta (Figures S5I and S5J). Furthermore, immunofluorescence staining of the aortic root revealed that DOPS facilitated enhanced macrophage colocalization with CRT, while cRGDfk notably increased overall CRT expression within the aorta tissue (Figures S5K–S5M). These results demonstrate that CRT-cRGDfk/pLNP selectively targets macrophages in atherosclerotic plaques and enhances the expression of CRT.

CRT-Ms amplify efferocytosis of apoptotic cells and alleviate inflammation in HFMCD-induced MASH model mice

In MASH, prolonged nutrient overload induces inflammatory cytokines that trigger TREM2 shedding, resulting in the accumulation of dead cells and subsequent inflammation.¹⁰ In the establishment of an MASH model using an HFMCD diet, TREM2 expression in liver macrophages was assessed during disease progression. At 4 weeks, corresponding to early steatosis, both TREM2 mRNA and protein levels were elevated. By contrast, at 8 weeks, during the steatohepatitis phase, TREM2 protein levels were markedly reduced despite sustained increases in mRNA expression, accompanied by elevated circulating sTREM2, indicating enhanced proteolytic shedding (Figures S6A–S6C). We then treated MASH model mice with various formulations via tail vein injection twice a week for four consecutive weeks (Figure 4A). sPirb/pLNPs encapsulating TREM2 mRNA (TREM2-sPirb/pLNPs) were prepared to compare the efficacy of CRT. After 2 weeks of treatment, immunofluorescence analysis revealed extensive apoptotic cell accumulation in MASH livers compared with normal-diet-fed controls (Figure 4B). Both TREM2-sPirb/pLNPs and CRT-sPirb/pLNPs increased macrophage-apoptotic cell colocalization, with CRT-sPirb/pLNPs inducing an approximately 3-fold increase, indicating enhanced efferocytosis during inflammation (Figures 4B and 4C). Consistently, serum sTREM2 levels were markedly elevated in TREM2-sPirb/pLNP-treated mice but remained comparable to PBS controls in the CRT-sPirb/pLNP group, confirming the cleavage resistance of CRT during inflammation (Figure 4D).

Figure 4.

Therapeutic effects of CRT-Ms mediated by sPirb/pLNPs in HFMCD-induced MASH model mice

(A) Schematic illustration of MASH model establishment and treatment strategy in C57BL/6 mice.

(B) Immunofluorescence images of macrophages (green) and apoptotic cells (red) in liver tissues after 2 weeks of treatment with the indicated reagents. The nuclei were stained with DAPI. TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling. The white arrows indicate the colocalization of macrophages and apoptotic cells. Scale bars, 50 μm.

(C) Quantification of efferocytosis represented as the ratio of macrophage-associated apoptotic cells to total apoptotic cells per liver area (n = 5).

(D) Serum sTREM2 levels in mice after 2 weeks of treatment.

(E) Representative liver tissue sections stained with oil red O, H&E, and α-SMA after 4 weeks of treatment. H&E, hematoxylin-eosin staining; α-SMA, alpha-smooth muscle actin. Scale bars, 50 μm.

(F) Quantitative analysis of oil-red-O-positive areas (n = 5).

(G) Quantitative analysis of α-SMA-positive areas (n = 5).

(H) Representative TUNEL staining (red) of liver sections from mice treated for 4 weeks. The nuclei were stained with DAPI. The white arrows indicate TUNEL⁺ apoptotic cells. Scale bars, 50 μm.

(I) Quantitative analysis of apoptotic cells relative to the total number of hepatocytes (n = 5).

(J) Representative flow cytometry analysis of liver macrophages showing the expression of the M1 marker CD86 and the M2 marker CD206 expression in mice treated for 4 weeks.

(K) The ratio of M1 to M2 macrophages was quantified as the proportion of CD86-positive cells to CD206-positive cells (n = 5).

(L–N) Serum levels of inflammatory cytokines TNF-α (L), IL-1β (M), and IL-6 (N) after the indicated treatments (n = 5). The data are shown as the means ± SDs from biological replicates. Statistical analysis was carried out by one-way ANOVA.

After 4 weeks of treatment, histological examination of liver tissue revealed pronounced hepatic steatosis in MASH mice. Treatment with CRT-sPirb/pLNPs diminished the oil-red-O-positive area and less abundant lipid droplets (Figures 4E and 4F). In addition, CRT-sPirb/pLNPs treatment led to a reduction in the positive area of α-SMA immunohistochemical staining, indicating decreased hepatic fibrosis (Figure 4G). Serum biochemical analysis showed that CRT-sPirb/pLNPs markedly reduced alanine aminotransferase (ALT), aspartate aminotransferase (AST), and hepatic hydroxyproline levels compared with PBS-treated MASH mice (Figures S6D–S6F). TUNEL staining of the liver confirmed an approximately 6-fold reduction in apoptotic cells in the CRT-sPirb/pLNPs-treated group, demonstrating the potent performance of CRT in contributing to the clearance of dead cells (Figures 4H and 4I). To investigate whether the macrophage phenotype in the liver changed after treatment, we quantitatively assessed the expression levels of CD86 and CD206 by flow cytometry. A notable downregulation of CD86 and upregulation of CD206 expression was observed after treatment with CRT-sPirb/pLNPs (Figures 4J and S6G–S6I), resulting in a marked shift toward M2 phenotype, as evidenced by the reduced M1/M2 ratio (Figure 4K). ELISA quantification of circulating inflammatory cytokines demonstrated a more significant reduction of pro-inflammatory cytokines, including TNF-α, IL-1β, and IL-6 in the CRT-sPirb/pLNPs group (Figures 4L–4N). TREM2-sPirb/pLNPs exhibited moderate therapeutic effects in all conditions, but these were weaker than those observed with CRT-sPirb/pLNPs, showing superior anti-inflammatory efficacy of CRT.

In summary, liver-specific expression of TREM2 and CRT via sPirb/pLNPs ameliorated steatosis and fibrosis, reduced apoptotic-cell burden, and attenuated inflammation in the HFMCD-induced MASH model. Notably, treatment with CRT-sPirb/pLNPs did not increase sTREM2 in the circulation, and its superior therapeutic activity was associated with reduced susceptibility to ADAM17-mediated cleavage under inflammatory conditions.

In addition, CRT-sPirb/pLNP treatment did not induce detectable toxicity in healthy C57BL/6 mice, with no significant changes in organ morphology, body weight, serum biochemical parameters, or hematological indices after 4 weeks of administration (Figures S6J–S6X).

CRT-Ms reduce hepatic apoptotic cell accumulation while broadly mitigating inflammation and fibrosis in HFD-CCl₄-induced MASH mice

The protective role of TREM2 in MASH has been widely documented. To further demonstrate the application potential of our approach, we assessed the therapeutic efficacy of CRT therapy based on pLNPs in another well-established MASH model. Following previously reported protocols, we induced an MASH model characterized by marked hepatic fibrosis by feeding C57BL/6 mice an HFD for 12 weeks while administering CCl₄ injections for an additional 4 weeks, during which HFD feeding was maintained. Consistent with the prior model, treatments were initiated after successful model establishment and administered twice weekly for four consecutive weeks (Figure 5A). To benchmark the efficacy of CRT-sPirb/pLNPs against an established therapy, resmetirom, the first approved treatment for MASH, was administered via oral gavage twice weekly as a comparator.

Figure 5.

Therapeutic evaluation of steatosis, fibrosis, and inflammation in HFD-CCl₄-induced MASH model mice treated with CRT-Ms mediated by sPirb/pLNPs

(A) Schematic illustration of MASH model induction and treatment regimen.

(B) Representative H&E, oil red O, and Sirius Red staining of liver sections from mice subjected to the indicated treatments.

(C) Quantitative analysis of NAS scoring based on H&E-stained sections (n = 8).

(D) Quantitative analysis of oil-red-O-positive areas (n = 8). Scale bars, 100 μm.

(E) Quantitative analysis of Sirius-Red-positive areas in liver sections (n = 8).

(F) Representative TUNEL staining (red) of liver sections from mice treated for 4 weeks. The nuclei were stained with DAPI. Scale bars, 100 μm.

(G) Quantitative analysis of apoptotic cells relative to the total number of hepatocytes (n = 8).

(H) Western blot analysis of SCD1, α-SMA, COL1A1, TNF-α, and IL-1β in liver tissues from mice receiving the indicated treatments.

(I–L) RT-qPCR quantification of Col1a1 (I), Scd1 (J), Tnf (K), and Il1b (L) expression levels in liver tissues.

(M) Representative flow cytometry analysis of the proportion of monocyte-derived macrophages within the hepatic macrophage population.

(N) Quantitative analysis of the proportion of monocyte-derived macrophages within the hepatic macrophage population (n = 8).

(O) Representative flow cytometry analysis of neutrophil populations in the liver.

(P) Quantitative analysis of neutrophil populations in the liver (n = 8). The data are shown as the means ± SDs from biological replicates. Statistical analysis was carried out by one-way ANOVA.

After 4 weeks of treatment, histological assessment by H&E staining and NAS scoring revealed that TREM2-sPirb/pLNPs, resmetirom, and CRT-sPirb/pLNPs all reduced NAS scores relative to the PBS group (Figures 5B and 5C). Both resmetirom and CRT-sPirb/pLNPs improved steatosis, inflammation, and ballooning, whereas TREM2-sPirb/pLNPs mainly reduced inflammation and ballooning with limited effect on steatosis. Consistently, oil red O staining confirmed reduced lipid accumulation in resmetirom- and CRT-sPirb/pLNP-treated mice, while only a modest reduction was observed in the TREM2-sPirb/pLNP group (Figure 5D). Sirius Red staining further showed that all three treatment groups reduced fibrotic deposition compared with PBS (Figure 5E). Quantification of liver hydroxyproline content indicated that CRT-sPirb/pLNPs and resmetirom achieved comparable fibrosis attenuation, both outperforming TREM2-sPirb/pLNPs (Figure S7A). TUNEL staining demonstrated that CRT-sPirb/pLNPs most effectively reduced apoptotic cell accumulation (Figures 5F and 5G).

At the molecular level, western blot analysis demonstrated that the expression levels of fibrosis-associated proteins (α-SMA and COL1A1), lipid-metabolism-related SCD1, and pro-inflammatory cytokines (TNF-α and IL-1β) were reduced across all treatment groups (Figure 5H). Notably, CRT-sPirb/pLNPs consistently achieved reductions comparable to or greater than those observed with resmetirom, whereas TREM2-sPirb/pLNPs showed only partial improvement. Further RT-qPCR quantification of Col1a1, Scd1, Tnf, and Il1b expression revealed patterns largely consistent with the western blot findings. Both resmetirom and CRT-sPirb/pLNPs elicited marked reversal of fibrotic, metabolic, and inflammatory alterations, with CRT-sPirb/pLNPs showing significantly greater suppression of inflammatory cytokine expression compared to resmetirom (Figures 5I–5L). Monocyte-derived macrophages recruited to the liver during steatohepatitis are key drivers of inflammation, and their relative abundance reflects the overall state of the hepatic macrophage compartment. Analysis of hepatic macrophage subsets across treatment groups showed that CRT-sPirb/pLNPs induced the most pronounced reduction in monocyte-derived macrophages, indicating effective depletion of inflammatory macrophages (Figures 5M, 5N and S7B). In parallel, neutrophil infiltration, a hallmark of hepatic inflammation, was also most substantially decreased in the CRT-sPirb/pLNPs group, further confirming robust attenuation of liver inflammation (Figures 5O, 5P, and S7C).

Together, these results demonstrate that CRT-sPirb/pLNPs achieve comprehensive protection against steatosis, fibrosis, apoptosis, and inflammation in MASH, primarily by efficiently enhancing TREM2 signaling under inflammatory conditions, yielding superior efficacy compared with TREM2-sPirb/pLNPs and offering broader therapeutic benefits than the clinically approved drug resmetirom.

CRT-Ms increase atherosclerotic plaque stability and reduce inflammation in atherosclerosis model mice

We further examined the therapeutic potential of CRT-Ms in an atherosclerosis model, based on the critical role of TREM2 in maintaining plaque stability through sustained clearance of cell debris by macrophages. This protective function was reported by Piollet et al., who demonstrated that administration of the TREM2 agonist antibody 4D9 improved plaque stability and reduced overall plaque burden in murine atherosclerosis models.²⁰ Building on these findings, we investigated whether in situ generated CRT-Ms could achieve comparable or enhanced benefits. Given that elevated circulating sTREM2 levels correlate with plaque progression in patients, we hypothesized that preventing TREM2 shedding may improve plaque stability.³¹ Atherosclerosis was induced by feeding LDL receptor-knockout (LDLr-KO) mice an HFD for 12 weeks, followed by bi-weekly intravenous administration of the indicated formulations for 4 weeks (Figure 6A). After 2 weeks of treatment, immunofluorescence analysis of aortic roots revealed substantial apoptotic cell accumulation in all HFD groups compared with normal diet controls (Figure 6B). Both TREM2-cRGDfk/pLNPs and CRT-cRGDfk/pLNPs enhanced macrophage-apoptotic cell colocalization relative to PBS treatment, with CRT-cRGDfk/pLNPs producing a markedly stronger effect (Figures 6B and 6C). After 4 weeks of treatment, mice were euthanized for histological analysis. The total atherosclerotic plaque area was quantified in the aortic root, aortic arch, and brachiocephalic artery (Figure 6D). LDLr-KO mice fed an HFD and treated with PBS, TREM2-cRGDfk/pLNPs, or CRT-cRGDfk/pLNPs exhibited significantly larger arterial plaques compared to the normal diet control (Figures 6E–6G). While the plaque areas in the aortic root, aortic arch, and brachiocephalic artery were similar between the TREM2-cRGDfk/pLNPs and PBS groups, CRT-cRGDfk/pLNPs administration led to a noticeable reduction in plaque size in the aortic arch and brachiocephalic artery, with no notable change observed in the aortic root (Figures 6E–6G). Whole aorta staining confirmed the reduced overall disease burden in the CRT-cRGDfk/pLNPs treatment group (Figures S8A and S8B). Assessment of plaque composition revealed that PBS-treated mice exhibited large necrotic cores and thin fibrous caps. In contrast, CRT-cRGDfk/pLNP treatment markedly reduced necrotic core size and increased fibrous cap thickness, as shown by H&E and Sirius Red staining (Figures 6H–6K). These findings suggest that CRT-cRGDfk/pLNPs enhanced plaque stability, probably due to the augmented collagen synthesis and fibrous tissue remodeling. Consistently, TUNEL staining showed that treatment with CRT-cRGDfk/pLNP reduced the number of apoptotic cells within the aortic root plaque by approximately 5-fold. (Figures 6L and 6M). In parallel, immunofluorescence analysis revealed reduced expression of pro-inflammatory cytokines, including TNF-α, IL-1β, and IL-6, within atherosclerotic lesions of CRT-cRGDfk/pLNP-treated mice (Figures S8C–S8F). These findings were corroborated by significantly lower serum levels of the same cytokines (Figures 6N–6P).

Figure 6.

In vivo therapeutic effects of CRT-Ms mediated by cRGDfk/pLNPs in atherosclerotic model mice

(A) Schematic illustration of atherosclerosis model establishment and treatment strategy in LDLr-KO mice.

(B) Representative immunofluorescence images of apoptotic cells (red) and macrophages (green) in the aortic root of mice after three administrations. The nuclei were stained with DAPI. The white arrows indicate the colocalization of macrophages and apoptotic cells. Scale bars, 50 μm.

(C) Quantification of macrophage engulfment capacity within plaques, represented as the ratio of macrophage-associated apoptotic cells to total apoptotic cells.

(D) Representative oil red O staining of frozen sections from the aortic root, aortic arch, and brachiocephalic artery of mice after 4 weeks of treatment. Scale bars, 200 μm.

(E–G) Quantitative analysis of oil-red-O-positive areas in sections of the aortic root (E), aortic arch (F), and brachiocephalic artery (G).

(H) Representative H&E staining of sections from the aortic root of mice after 4 weeks of treatment. The upper image shows the entire plaque area outlined by black dashed lines, while the lower image highlights the necrotic core area within the plaque. Scale bars, 200 μm.

(I) Quantification of the ratio of the necrotic core area to the plaque area in H&E-stained sections of the aortic root (n = 5).

(J) Representative images of Sirius-Red-stained frozen sections from the aortic root after 4 weeks of treatment. The lower image shows black dashed lines outlining the plaque area, and black solid lines indicate the fibrous cap thickness. Scale bars, 200 μm.

(K) Quantification of the fibrous cap area to plaque area ratio in Sirius-Red-stained sections of the aortic root.

(L) Representative TUNEL immunofluorescence staining (red) indicating apoptotic cells in frozen sections of the aortic root from mice after 4 weeks of treatment. The nuclei were labeled with DAPI. The white dashed lines in the lower image outline the plaque area. Scale bars, 200 μm.

(M) Quantification of the ratio of apoptotic cells to total cells in the aortic root from TUNEL-stained sections (n = 5).

(N–P) Plasma concentrations of TNF-α (N), IL-1β (O), and IL-6 (P) in mice after 4 weeks of treatment (n = 5). The data are shown as the means ± SDs from biological replicates. Statistical analysis was carried out by one-way ANOVA.

Taken together, these results demonstrate that the in situ generated CRT-Ms within atherosclerotic plaques via CRT-cRGDfk/pLNPs significantly reduce arterial lesion areas and enhance plaque stability, leading to a reduction in necrotic core size and thickening of the fibrous cap. Moreover, CRT expression reduces pro-inflammatory cytokine levels, promoting inflammation resolution. In contrast, in situ generated TREM2-engineered macrophages (TREM2-Ms) via TREM2-cRGDfk/pLNPs provided limited therapeutic benefits. Although TREM2-Ms reduced apoptotic cell accumulation and provided modest plaque stabilization, it was insufficient to effectively mitigate the inflammatory processes within the plaques.

Finally, safety evaluation of CRT-cRGDfk/pLNPs in LDLr-KO mice maintained on a normal diet revealed no histological abnormalities in major organs and no significant changes in organ indices, serum biochemical parameters, or hematological profiles following 4 weeks of treatment (Figures S8G–S8T).

Discussion

Efferocytosis is a subtle yet crucial process through which macrophages remove apoptotic cells, thereby resolving inflammation and maintaining tissue homeostasis.⁵⁹ TREM2 has emerged as an important regulator in this process, promoting the ingestion of dying cells while simultaneously driving anti-inflammatory pathways, by binding to ligands present in apoptotic cell debris, such as phospholipids and lipoproteins.^6,17,19 This functionality is particularly critical in diseases including MASH, atherosclerosis, ischemic stroke, and myocardial infarction, where TREM2-mediated efferocytosis and immune modulation contribute significantly to disease resolution.^9,10,17,60 However, TREM2 is susceptible to proteolytic cleavage, particularly within inflammatory environments. Inflammatory mediators such as TNF-α, IFN-γ, and IL-1β have been shown to induce ADAM17-mediated shedding of TREM2, resulting in the release of sTREM2. This process compromises the function of TREM2, thereby exacerbating inflammatory responses and accelerating disease progression.^10,15 Elevated circulating levels of sTREM2 in the blood have been implicated in the worsening of several diseases, highlighting the pathological consequences of TREM2 cleavage in the context of sustained inflammation.^32,61

In this study, a cleavage-resistant TREM2 receptor is engineered to preserve TREM2 signaling under inflammatory conditions. By linking the ligand-binding domain of TREM2 directly to the intracellular signaling adaptor DAP12 through customized stalk and transmembrane elements, the engineered receptor eliminates the canonical cleavage site and bypasses the requirement for adaptor recruitment. This design enables sustained activation of TREM2/DAP12 signaling despite inflammatory stimuli that otherwise promote receptor shedding. Efficient and selective expression of CRT in macrophages is achieved through delivery of CRT mRNA using tailored lipid nanoparticles. The transient nature of mRNA-based expression allows temporal control of receptor activity while avoiding permanent genetic modification. In both in vitro and in vivo settings, CRT-Ms macrophages exhibit enhanced efferocytosis and reduced inflammatory activation.

In MASH mice, treatment with CRT-sPirb/pLNPs improves liver inflammation, reduces fibrosis, and markedly decreases apoptotic cell accumulation. In atherosclerotic mice, CRT-cRGDfk/pLNPs reduce plaque inflammation, decrease necrotic core size, increase fibrous cap thickness, and improve plaque stability. Notably, Piollet et al. demonstrated that treatment with the TREM2 agonist antibody 4D9 improved plaque stability and reduced aortic lesion burden in murine models.²⁰ While CRT achieved comparable improvements in plaque stability, our approach further demonstrated reductions in apoptotic cell accumulation and inflammatory activity. Compared with antibody-based approaches that stabilize endogenous TREM2, CRT provides a cell-intrinsic strategy to sustain TREM2 signaling directly within macrophages, thereby limiting systemic exposure and enhancing therapeutic precision.

In sum, our study demonstrates that engineered CRT effectively amplifies TREM2 signaling, thereby enhancing macrophage-mediated apoptotic cell clearance and attenuating inflammation across multiple disease models. Further investigations extending these findings to additional pathological contexts, including neurodegenerative diseases, and elucidating the broader mechanisms of CRT action will support its potential for broader translational application.

Limitations of the study

Despite these findings, several limitations warrant consideration. First, CRT was evaluated using mRNA-based transient expression via pLNP delivery. TREM2 overexpression has already shown beneficial effects in AD models,^23,37,62 and inducible knockin (KI) models could offer further insights into the long-term effects of CRT under physiological expression levels. Second, our mechanistic validation primarily focused on efferocytosis and inflammatory resolution; however, TREM2 signaling also modulates lipid metabolism, cell growth, and survival.^3,6,63,64 Investigating these functions may uncover additional pathways underlying CRT-mediated benefits. Third, macrophage subpopulation profiling in MASH is complex, and the M1/M2 paradigm provides only limited resolution. Subset-level profiling of macrophage populations will be critical to delineate how CRT influences macrophage diversity and polarization. Finally, the potential of CRT in CNS disorders remains unexplored, largely due to challenges in delivering LNPs across the blood-brain barrier, despite TREM2’s well-established protective role in neurodegenerative diseases such as AD.^62,65

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Kun Zhao (kzhao2013@sdu.edu.cn).

Materials availability

Materials described in this manuscript may be made available to qualified, academic, noncommercial researchers through a materials transfer agreement (MTA). Requests for materials should be directed to the lead contact.

Data and code availability

• •All data supporting the findings of this study are available within the article and its supplementary information.
• •This study did not generate any unique code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon reasonable request.

Acknowledgments

We acknowledge the technical support from L.W., Y.Y., M.W., J.Z., and X.Y.in the Advanced Medical Research Institute/Translational Medicine Core Facility of the Advanced Medical Research Institute, Shandong University. We are also grateful for X.W. in the Pharmaceutical Biology Sharing Platform, School of Pharmaceutical Sciences, Shandong University for technical support. This work was supported by the National Key Research and Development Program of China (2024YFA0918400), National Natural Science Foundation of China (82425056, 82350125, and 82173763), National Natural Science Foundation of China (22101154), the Fundamental Research Funds of Shandong Province (ZR2022ZD18), and Department of Science and Technology of Shandong Province (2022HWYQ-008 and ZR2021QB038).

Author contributions

X.D., X.Z., J.G., K.Z., and Y.S. conceived and designed the experiments. X.D. and J.G. synthesized and characterized the materials. X.D., J.G., Y.Z., Q.X., and Z.L. conducted in vivo and in vitro experiments and analyzed the experimental data. Z.K., W.S., C.L., and X.X. contributed to experimental methods and data interpretation. X.D., L.B., and Y.S. contributed to the interpretation of core datasets underlying the main conclusions and structured the manuscript. X.D. and L.B. wrote the manuscript and designed the figures. X.Z., J.G., K.Z., N.L., J.L., and X.J. edited and revised the manuscript. All authors approved the final version of the manuscript.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal anti-Myc tag, FITC-conjugated	Abcam	Cat# ab1394; RRID: AB_300767
Mouse monoclonal anti-Myc tag	Abcam	Cat# ab18185; RRID: AB_444307
Rabbit polyclonal anti-TREM2	Invitrogen	Cat# PA5-119690; RRID: AB_2913263
Rabbit monoclonal anti DAP12	Abcam	Cat# ab283679 RRID: AB_3086739
Goat polyclonal secondary antibody to Mouse IgG1 heavy chain, HRP-conjugated	Abcam	Cat# ab97240; RRID: AB_10695944
Goat polyclonal secondary antibody to Rabbit IgG (H&L), HRP-conjugated	Abcam	Cat# ab6721; RRID: AB_ 955447
Rabbit polyclonal anti-β-Actin	Abcam	Cat# ab8227; RRID:AB_2305186
Rabbit polyclonal anti-TREM2	Invitrogen	Cat# PA5-87933; RRID: AB_2804516
Rat monoclonal anti-mouse F4/80 (Brilliant Violet 421)	Biolegend	Cat# 123131; RRID: AB_10901171
Rat monoclonal anti-mouse TREM2, FITC-conjugated	Thermo Fisher Scientific	Cat# MA5-28223; RRID: AB_2745193
Rabbit polyclonal anti-Myc tag	Abcam	ab9106; RRID: AB_307014
Goat polyclonal anti-Rabbit IgG (H&L), Alexa Fluor® 488-conjugated	Abcam	Cat# ab150077; RRID: AB_2630356
Rat monoclonal anti-mouse CD86	Thermo Fisher Scientific	Cat# 14-0862-82; RRID: AB_467368
Rabbit polyclonal anti-CD206	Proteintech	Cat# 18704-1-AP; RRID: AB_10597232
Rat monoclonal anti-mouse CD86, PE/Cyanine7-conjugated	Biolegend	Cat# 105115; RRID: AB_493601
Rat monoclonal anti-mouse CD206, PE/Dazzle™ 594-conjugated	Biolegend	Cat# 141731; RRID: AB_2565931
Rabbit monoclonal anti-mouse/rat F4/80	Abcam	Cat# ab300421; RRID: AB_2936298
Rabbit polyclonal anti-alpha smooth muscle Actin (α-SMA)	Abcam	Cat# ab5694; RRID: AB_2223021
Rabbit monoclonal anti-SCD1	Abcam	Cat# ab236868; RRID: AB_2928123
Rabbit anti-Collagen I	Abcam	Cat# ab270993; RRID: AB_2927551
Rabbit anti- TNF alpha	Abcam	Cat# ab183218; RRID: AB_2889388
Rabbit anti- IL-1 beta	Abcam	Cat# ab315084; RRID: AB_3105874
Rat monoclonal anti-mouse CD45, FITC-conjugated	Invitrogen	Cat# 103108; RRID: AB_312973
Rat monoclonal anti-mouse Ly6g, PerCP/Cy5.5-conjugated	Abcam	Cat# ab210207
Rat monoclonal anti-mouse/rat CD11b, APC/Cy7-conjugated	Abcam	Cat# ab79096; RRID: AB_1604124
Rat monoclonal anti-mouse Tim-4, PE-conjugated	BioLegend	Cat# 129905; RRID: AB_1227799
Mouse monoclonal anti-mouse CLEC4F, Alexa Fluor® 647-conjugated	BioLegend	Cat# 156803; RRID: AB_2814081
Rabbit recombinant monoclonal anti-Syk (phospho Y352) + ZAP70 (phospho Y319)	Abcam	Cat# ab300398; RRID: AB_3095853
Rabbit polyclonal anti-SYK	Abcam	Cat# ab155187
Rabbit recombinant monoclonal anti-TREM2	Abcam	Cat# ab305103; RRID: AB_3086793
Rabbit recombinant monoclonal anti-IL-6	Abcam	Cat# ab290735; RRID: AB_3064891
Chemicals, peptides, and recombinant proteins
L-Pyroglutamic acid	Bide Pharmatech Ltd.	CAS: 98-79-3
EDCI	Bide Pharmatech Ltd.	CAS:7084-11-9
HOBT	Bide Pharmatech Ltd.	CAS: 2592-95-2
DIPEA	Bide Pharmatech Ltd.	CAS: 7087-68-5
NaHB(OAc)3	Bide Pharmatech Ltd.	CAS: 56553-60-7
TEA	Bide Pharmatech Ltd.	CAS: 121-44-8
TFA	Bide Pharmatech Ltd.	CAS: 76-05-1
N-Boc-Ethylenediamine	Bide Pharmatech Ltd.	CAS: 57260-73-8
DOPE	MedChemExpress	CAS: 4004-05-1
DSPE-PEG	MedChemExpress	CAS: 892144-24-0
Cholesterol	MedChemExpress	CAS: 57-88-5
DOPS	AVT Pharmaceutical Tech Co., Ltd.	CAS: 6811-55-8
DSPE-PEG-NHS	RuixiBiotechCo.Ltd	CAS: 1445723-73-8
sPirB	Cusbio	CSB-MP012941MO2
cRGDfk	RuixiBiotechCo.Ltd	R-RG0219
Critical commercial assays
BCA Protein Assay Kit	Beyotime	P0012
Luciferase Assay System	Promega	E1500
sTREM2 ELISA kit	Absin	abs551362
sTREM2 ELISA kit	Jonlnbio	JL20435
T7 High Yield RNA Transcription Kit	Novoprotein	E131-01A
Cap 1 Capping System	Novoprotein	M082-01A
HyperScribe™ T7 High Yield Cy5 RNA Labeling Kit	APE×BIO	K1062
Quant-it™ RiboGreen RNA Assay Kit	Thermo Fisher Scientific	R11490
BeyoClick™ EdU-594 Cell Proliferation Kit	Beyotime	C0078S
LDH Release Assay Kit	Beyotime	C0016
IFN-γ ELISA Kit	Beyotime	PI508
TNF-α ELISA Kit	Beyotime	PT512
IL-1β ELISA Kit	Beyotime	PI301
IL-6 ELISA Kit	Beyotime	PI326
One Step TUNEL Apoptosis Assay Kit	Beyotime	C1090
HYT Content Assay Kit	Solarbio	BC0255
GOT Assay Kit	Solarbio	BC1560
GPT Assay Kit	Solarbio	BC1550
Experimental models: Cell lines
Jurkat reporter cells	This study	N/A
HEK293T	Procell	CL-0005
RAW264.7	Procell	CL-0190
RAW264.7 TREM2 KO	Ubigene	YKO-MT0098
HHL-5	Bluefbio	BFN6072012687
AML12	Procell	CL-0602
BMDMs	This study	N/A
THP-1	Procell	CL-0233
THP-1 TREM2 KO	Ubigene	RRID: CVCL_E1DK
Experimental models: Organisms/strains
C57BL/6	Vital River	213
C57BL/6J-Ldlr KO	Cavens	C000114
C57BL/6J-Trem2-KO	Cyagen	C001207
Software and algorithms
ImageJ	ImageJ	https://imagej.net/
GraphPad Prism	Dotmatics	https://www.graphpad.com/
Adobe Illustrator	Adobe	https://www.adobe.com/products/illustrator/
Microsoft Powerpoint	Microsoft	https://powerpoint.cloud.microsoft/
FlowJo	FlowJo LLC	https://www.flowjo.com
MestReNova	Mestrelab Research	https://mestrelab.com/
ChemDraw	Revvity Signals	https://revvitysignals.com/products/research/chemdraw

Experimental model and study participant details

Cell lines

Jurkat reporter cells were cultured in RPMI 1640 supplemented with 10% FBS and 1% Penicillin/Streptomycin. HEK293T Cells and RAW264.7 were cultured in DMEM supplemented with 10% FBS and 1% Penicillin/Streptomycin. HHL-5 cells were cultured in DMEM-F12 supplemented with 10% FBS and 1% Penicillin/Streptomycin. AML12 cells were cultured in DMEM-F12 supplemented with 10% FBS, 0.5% ITS-G (100×), 40ng/mL Dexamethasone and 1% Penicillin/Streptomycin. THP-1 cells were maintained in RPMI 1640 supplemented with 10% FBS and 1% Penicillin/Streptomycin. To induce macrophage differentiation, THP-1 cells were treated with 100 nM PMA for 24–48 h prior to subsequent assays.

Animal experiments

Six-week-old male TREM2-KO mice were purchased from Cyagen. Upon arrival, the mice were immediately euthanized, and BMDMs were isolated for subsequent analysis.

Six-week-old male C57BL/6 mice were purchased from Beijing Vital River and housed under controlled conditions (25°C ± 2°C, 40–60% humidity, 12-h light/dark cycle) with ad libitum access to food and water. For the HFMCD-induced MASH model, mice were fed an HFMCD (A06071301B) for 8 weeks, while age-matched mice fed a normal diet (D11112201) under the same housing conditions served as controls. For the HFD-CCl₄-induced MASH model, sex-balanced (male and female) C57BL/6 mice that had been fed an HFD (D09100310) for 12 weeks were purchased from Cavens. Upon arrival, the mice were maintained on the HFD and additionally received intraperitoneal injections of CCl₄ (0.5 mL/kg in corn oil) twice per week. Mice fed a normal diet (D11112201) under identical housing conditions were used as controls. All animal procedures were approved by the Ethical Committee of the School of Pharmaceutical Sciences of Shandong University (Approval No. 22074).

Six-week-old male LDLr-KO C57BL/6 mice were obtained from Cavens. The mice were maintained under standardized conditions with a temperature of 25°C ± 2°C, 40–60% humidity, and a 12-h light/dark cycle. Food and water were provided ad libitum. For the induction of the atherosclerosis model, the LDLr-KO mice were fed an HFD (D12079B) for 12 weeks. Control mice, under identical housing conditions, were fed a normal chow diet (D11112201). All animal procedures were conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals, as approved by the Ethical Committee of the School of Pharmaceutical Sciences of Shandong University (Approval Nos.22074).

Method details

Transfection of reporter cells

The domain information of the native protein was sourced from the UniProt database. A myc-tag was inserted downstream of the signal sequence in each chimeric gene to enable detection of the expression of CRT variants. Jurkat reporter cells, stably expressing an NFAT-driven luciferase reporter, were generated based on the Jurkat T cell line (clone H33HJ) as previously described.⁶⁶ Constructs were cloned into the pCDH-CMV-MCS-EF1α-Puro vector. Stable Jurkat reporter cell lines expressing each CRT variant were generated separately via lentiviral transduction. Specifically, HEK293T cells were co-transfected with the expression plasmid and viral packaging plasmids pMD2.G and psPAX2 (Jiangyuan Biotechnology, JY03061). After 36 h, lentiviral supernatant was collected and used to transduce Jurkat cells for 48 h at a multiplicity of infection (MOI) of 10, calculated from the lot-specific functional titer (10¹⁰ TU/mL). Transduced cells were selected with 1 μg/mL puromycin until sufficient transgenic populations were established. Cells were harvested by centrifugation and stained with an anti-myc-tag antibody (Abcam, ab1394), followed by flow cytometry (Gallios, Beckman Coulter) to assess the expression of CRT variants on the surface of reporter cells. TREM2-KO RAW264.7 reporter cells were generated by lentiviral transduction following the same procedure used for Jurkat NFAT-Luci cells.

For western blotting, reporter cells were lysed in RIPA buffer containing protease and phosphatase inhibitors. For samples requiring deglycosylation, lysates were treated with PNGase F (Yeasen, 20407ES01) according to the manufacturer’s instructions; samples not requiring deglycosylation were processed directly. For samples requiring surface protein biotinylation, cells were washed twice with ice-cold PBS and incubated with 1 mg/mL EZ-Link Sulfo-NHS-LC-Biotin (Thermo Fisher Scientific, A39257) in PBS for 30 min on ice with gentle agitation. The reaction was quenched by adding 50 mM Tris-HCl for 10 min on ice, followed by three washes with PBS. Cells were then lysed in RIPA buffer, and biotinylated proteins were captured using streptavidin-conjugated beads. Protein concentrations were quantified by a BCA Protein Assay Kit (Beyotime, P0012). Equal amounts of protein were resolved via SDS-PAGE and transferred onto nitrocellulose membranes. Membranes were blocked with rapid blocking buffer (Beyotime, P0252) and incubated with a primary anti-myc antibody (Abcam, ab18185) or a primary anti-TREM2 antibody (Invitrogen, PA5-119690), followed by a secondary antibody (Abcam, ab97240 or Abcam, ab6721). β-actin (Abcam, ab8227) was used as a loading control. Signal development was performed by SuperSignal West Pico PLUS (Thermo Fisher Scientific, 34577), and images were captured with an Amersham Imager 600 RGB system.

Generation of apoptotic cells

PA (0.2 mmol) was added to 2 mL of 0.1 M NaOH and reacted at 70°C for 20 min to undergo saponification. The reaction mixture was then quickly added to 18 mL of 20% BSA solution (preheated to 55°C and free of fatty acids) to yield a 10 mM PA stock solution. After sterilization by filtration, the stock solution was diluted with culture medium to a final concentration of 0.4, 0.8 and 1.2 mM for use. HHL-5 or AML12 cells were cultured to 80–90% confluence, at which point the medium was replaced with designated concentrations of PA. The cells were incubated for 16–24 h until extensive cell detachment was observed. Apoptosis was assessed using an Annexin V-PE/7-AAD Apoptosis Detection Kit (Yeasen, 40310ES20) and flow cytometry. Upon confirming successful apoptosis induction, the cells were transferred to fresh, blank medium for 12 h before use.

TREM2/DAP12 signaling reporter assay

Jurkat reporter cells stably expressing CRT variants, either untreated or pretreated with 300 nM ADAM17, 200 nM PMA, or anti-TREM2 antibody (Invitrogen, PA5-87933), were seeded in 96-well plates and cocultured with apoptotic HHL-5 cells at a 1:1 ratio, calculated based on the pre-apoptotic cell count. The expression level of Luci was then measured by the Luciferase Assay System (Promega, E1500), and luminescence intensity was read with a Multimode Plate Reader (EnSight, PerkinElmer). Cells pretreated with ADAM17 or PMA were lysed, and whole-cell lysates were analyzed by western blot to determine full-length TREM2 levels and the generation of CTFs using an anti-TREM2 antibody (Invitrogen, PA5-119690) or an anti-DAP12 antibody (Abcam, ab283679). Culture supernatants were collected and concentrated by ultrafiltration, and the resulting protein samples were subjected to western blot to assess sTREM2 levels with the same antibody. Specifically, equal amounts of protein from cell lysates and concentrated supernatants were resolved by SDS-PAGE and transferred onto nitrocellulose membranes, followed by immunoblotting as described above.

Quantification of sTREM2 by ELISA

Culture supernatants from Jurkat reporter cells treated as indicated were collected, clarified by centrifugation at 300 × g for 5 min to remove cell debris, and stored at −80°C until analysis. The concentration of sTREM2 in the supernatants was measured using a sTREM2 ELISA kit (Absin, abs551362 or Jonlnbio, JL20435) according to the manufacturer’s instructions. Absorbance was read at 450 nm with a Multimode Plate Reader (EnSight, PerkinElmer), and sTREM2 concentrations were calculated from a standard curve generated with recombinant sTREM2 provided in the kit.

Detection of Syk phosphorylation

Jurkat reporter cells were cocultured with apoptotic cells at a 1:1 ratio, calculated based on the pre-apoptotic cell count, for the indicated times. Cells were harvested, washed twice with ice-cold PBS, and lysed in RIPA buffer (Beyotime, P0013B) supplemented with protease and phosphatase inhibitors (Beyotime, P1045). Protein concentrations were determined using a BCA Protein Assay Kit (Beyotime, P0012). Equal amounts of protein were resolved by SDS-PAGE and transferred onto nitrocellulose membranes. Membranes were blocked with rapid blocking buffer (Beyotime, P0252) and incubated with an anti–phospho-Syk antibody (Abcam, ab300398) or an anti-Syk antibody (Abcam, ab155187), followed by incubation with an HRP-conjugated secondary antibody (Abcam, ab6721). β-actin (Abcam, ab8227) was used as a loading control. Signal development was performed by SuperSignal West Pico PLUS (Thermo Fisher Scientific, 34577), and images were captured with an Amersham Imager 600 RGB system.

AcLDL uptake assay

The reporter cells were seeded in six-well plates at a density of 10⁵ cells per well. Subsequently, the cells were incubated with 20 μg/mL DiO-AcLDL (NoninBio, NBS9914) for 4 h. Following incubation, the cells were washed and stained with 1 μM Cell Tracker Red for 30min. The samples were then imaged using ZEISS LSM 900 with Airyscan 2. Use ImageJ to calculate the positive area of Cell Tracker Red and DiO-AcLDL.

Synthesis of DKP-2-O lipids

The synthetic pathway for DKP-2-O can be found in Figure S2B in the supplementary information. 40 mL of acetic anhydride was mixed with 8 mL of pyridine, and L-Pyroglutamic acid (60.0 mmol) was added to this mixture. The reaction was refluxed at 110°C overnight. After cooling to 0°C in an ice bath, the mixture was filtered, and the precipitate was washed with pre-cooled methanol. The resulting crude intermediate 2a was dried. This crude product was then dissolved in 8 mL of concentrated sulfuric acid, followed by the addition of 15 mL of distilled water. The mixture was allowed to stand for 2 h, after which it was filtered, and the precipitate was washed with distilled water and dried to obtain the pure intermediate 2a with a yield of 31.7%. Next, compound 2a (5.00 mmol), EDCI (15.0 mmol), HOBT (15.0 mmol), and DIPEA (15.0 mmol) are added to 15 mL of DCM and stirred for 30 min until fully dissolved. N-Boc-Ethylenediamine (15.0 mmol) is then added to the mixture, which is stirred overnight. The reaction mixture is washed with 15 mL of saturated NaHCO₃ solution, and the organic phase is separated and concentrated under reduced pressure. The resulting residue is purified by column chromatography to yield pure compound 3a with a yield of 69.9%. Dissolve compound 3a (2.00 mmol) in 5 mL of DCM, and slowly add 5 mL of TFA under an ice bath. Stir the reaction mixture at 0°C for 4 h. Remove the solvent under reduced pressure to obtain compound 4a, which is used directly in the next step without purification. Dissolve compound 4a in 15 mL of DCM, add DIPEA (4.8 mmol), oleic aldehyde (8.80 mmol) and NaHB(OAc)₃ (8.80 mmol), and stir the reaction overnight. Quench the reaction mixture with 15 mL of saturated NaHCO₃ solution, separate the layers, and purify the organic layer by column chromatography to obtain DKP-2-O with a yield of 57.3%.

Isolation of mouse BMDMs

Femurs were harvested from 8-week-old male C57BL/6 mice or TREM2-KO mice (Cyagen, C001207), and the bone ends were clipped. Bone marrow cells were flushed out with 1× PBS, followed by red blood cell lysis and centrifugation. The resulting cell pellet was resuspended in medium supplemented with 20 ng/mL M-CSF. After 72 h, non-adherent cells were removed by washing, and adherent cells were cultured for an additional seven days in M-CSF-containing medium to promote BMDM differentiation. The purity of the BMDMs was confirmed by flow cytometry after staining with an F4/80 antibody (BioLegend, 123131). To detect TREM2 expression, BMDMs were stained with a TREM2 antibody (Thermo Fisher Scientific, MA5-28223) for flow cytometric analysis.

Synthesis of CRT mRNA

The plasmid encoding CRT was linearized by restriction enzyme digestion and transcribed into mRNA by the T7 High Yield RNA Transcription Kit (Novoprotein, E131-01A). A 5′ cap was added by the Cap 1 Capping System (Novoprotein, M082-01A). The Poly(A) tail sequence was pre-included in the plasmid sequence. For fluorescent labeling, Cy5-labeled CRT mRNA was synthesized by the HyperScribe T7 High Yield Cy5 RNA Labeling Kit (APE×BIO, K1062) in combination with the Cap 1 Capping System.

Cellular uptake of mRNA

CRT mRNA, CRT-LNPs and CRT-pLNPs were added to BMDMs and RAW264.7 cells seeded in 12-well plates at equivalent mRNA concentrations and incubated for 6 h. CRT mRNA was Cy5-labeled for tracking purposes. CRT-LNPs were prepared as described previously, with the omission of DOPS from the formulation. For confocal imaging, cells were washed with PBS and stained with DiO to label the cell membrane. After washing, cells were fixed with 4% paraformaldehyde in PBS and counterstained with DAPI to visualize nuclei. Imaging was performed by LSM 900 with Airyscan 2 (ZEISS). For flow cytometry (Gallios, Beckman Coulter), cells were directly harvested and analyzed to quantify mRNA uptake.

Preparation and characterization of CRT-pLNPs

CRT mRNA was diluted in citrate buffer (pH 4.0) to obtain the aqueous phase. DKP-2-O, DOPE, DSPE-PEG, DOPS, and cholesterol were dissolved in ethanol at specific molar ratios (DKP-2-O: cholesterol: DSPE-PEG: DOPE: DOPS = 45 : 25 : 4.5 : 30 : 5) to form the organic phase. By microfluidic technology, the aqueous phase and organic phase were mixed on a chip at a ratio of 3:1 to prepare CRT-pLNPs. The LNPs were dialyzed in 1×PBS solution for 12 h. The encapsulation efficiency was measured by the Quant-it RiboGreen RNA Assay Kit (Thermo Fisher Scientific, R11490). Briefly, prepare a series of RNA standards with different concentrations to construct a standard curve. Add CRT-pLNPs to 1×TE buffer, with and without Triton lysis buffer, and incubate for 15 min. Add Ribogreen staining working solution to both RNA standards and samples. Measure fluorescence by a Multimode Plate Reader (EnSight, PerkinElmer) in fluorescence mode (Ex/Em = 485/528 nm). Calculate the RNA concentration of the samples using the standard curve constructed from the RNA standards. The non-lysed samples represent free RNA content, while the Triton-lysed samples represent total RNA content. The encapsulation efficiency is calculated as the ratio of the difference between the total RNA content and the free RNA content to the total RNA content. The CRT-pLNP solution was diluted 100-fold, and the particle size, zeta potential, and PDI were measured by a Zetasizer Nano ZS90 (Malvern).

Lysosomal escape

CRT-pLNPs were coincubated with macrophages for 1 h, 3 h and 6 h to track the intracellular distribution of CRT mRNA labeled with Cy5. After coincubation, lysosomes and nuclei were stained using LysoTracker Green (Beyotime, C1047S) and DAPI, respectively. Images were captured using ZEISS LSM 900 with Airyscan 2, and the Manders’ coefficient was analyzed using ImageJ.

In vitro generation of CRT-Ms

TREM2-KO macrophages were seeded in 6-well plates at a density of 3×10⁵ cells per well and treated with either 1 μg of free CRT mRNA or 1 μg of CRT mRNA encapsulated in pLNPs for 12 h. The proportion of CRT-Ms was determined by quantifying myc-tag-positive cells by flow cytometry (Gallios, Beckman Coulter). For confocal microscopy, treated cells were fixed with 4% paraformaldehyde and blocked with 2% BSA. Cells were then incubated with a primary myc antibody (Abcam, ab9106) diluted in 2% BSA for 1 h at room temperature, followed by incubation with an FITC-labeled secondary antibody (Abcam, ab150077) diluted in 2% BSA for 45 min. Cells were subsequently stained with DAPI and DiD to visualize the nuclei and cell membranes, respectively. The colocalization of myc-tag with the cell membrane was imaged by LSM 900 with Airyscan 2 (ZEISS). For western blotting, cells were lysed and protein samples were analyzed with an anti-TREM2 antibody (Invitrogen, PA5-98213) and an HRP-conjugated secondary antibody (Abcam, ab6721), following the procedures described above.

Macrophage proliferation and viability assays

Macrophage proliferation was assessed 24 h after transfection with the BeyoClick EdU-594 Cell Proliferation Kit (Beyotime, C0078S). Cells were exposed to 10 μM EdU for 2 h at 37°C, fixed in 4% paraformaldehyde for 15 min at room temperature, and permeabilized with 0.3% Triton X-100 for 10 min. The Click reaction was performed according to the manufacturer’s instructions. After staining, cells were collected and analyzed by flow cytometry (Gallios, Beckman Coulter), and the percentage of EdU-positive cells was calculated as an index of proliferation.

Cell viability was evaluated 24 h after transfection by quantifying LDH activity in the culture supernatants with an LDH Release Assay Kit (Beyotime, C0016). Supernatants were harvested and cleared by centrifugation to remove cells. Absorbance at 490 nm was recorded using a Multimode Plate Reader (EnSight, PerkinElmer), and LDH release was expressed as a percentage of the maximal release obtained from lysed-cell controls.

In vitro efferocytosis assay

Macrophages were incubated with or without 5 ng/mL IL-1β for 6 h and incubated with pHrodo Red-labeled (Thermo Fisher Scientific, P36600) apoptotic AML12 cells for 4 h. After washing to remove unengulfed apoptotic cells, macrophages were stained with anti-F4/80 antibody (BioLegend, 123131) and the engulfment efficiency was analyzed by flow cytometry (Gallios, Beckman Coulter). To visualize efferocytosis, macrophages were labeled with Cell Tracker Green (Thermo Fisher Scientific, C2925) and coincubated with Cell Tracker Red-labeled (Thermo Fisher Scientific, C34552) apoptotic AML12 cells for 4 h. The cells were then imaged by LSM 900 with Airyscan 2 (ZEISS).

Phenotype shifting of macrophages in vitro

100 ng/mL LPS was added to the culture medium of macrophages for 24 h prior to transfection with CRT to induce M1 polarization. Macrophages were coincubated with apoptotic AML12 cells for 6 h. Apoptotic cells were then washed away, and the macrophages were continued to be cultured for 12 h. Then the cells were fixed with 4% paraformaldehyde. The fixed cells were permeabilized with 0.1% Triton X-100 and blocked with 2% BSA. Cells were then incubated with primary antibodies anti-CD86 (Thermo Fisher Scientific,14-0862-82) and anti-CD206 (Proteintech, 18704-1-AP) for 1 h at room temperature, followed by a 45-min incubation with secondary antibodies. Nuclei were stained with DAPI, and imaging was performed by LSM 900 with Airyscan 2 (ZEISS). For flow cytometry analysis, treated cells were stained with anti-CD86 (BioLegend, 105115) and anti-CD206 (BioLegend, 141731) antibodies to assess the proportions of M1 and M2 macrophage phenotypes (Gallios, Beckman Coulter). Additionally, the supernatant from treated cells was collected after centrifugation, and the concentrations of inflammatory cytokines IFN-γ, TNF-α, IL-1β, and IL-6 were measured by ELISA kits (Beyotime, PI508, PT512, PI301, PI326).

In vivo distribution of DiR-labeled nanoparticles

DiR was incorporated into an ethanol solution containing the lipid formulation for the preparation of DiR-labeled CRT-LNPs and CRT-pLNPs. Mice were injected via the tail vein with the DiR-labeled particles, and in vivo fluorescence imaging was performed 12 h post-injection using an IVIS spectrum in vivo imaging system to quantify the fluorescence intensity.

Synthesis of ROS linker DSPE-PEG-TK-Mal and DSPE-PEG-cRGDfk

DSPE-PEG-NHS (1.0 mmol) was dissolved in 5 mL of chloroform, and Mal-TK-NH₂ (1.0 mmol) and triethylamine (2.0 mmol) were added. The reaction was carried out at room temperature for 0.5 h. The reaction mixture was then concentrated under reduced pressure using a rotary evaporator. The residue was recrystallized with ethanol, precipitated with ice-cold diethyl ether, filtered, and the collected solid was vacuum-dried to obtain DSPE-PEG-TK-Mal with a yield of 43.7%.

DSPE-PEG-NHS (1.00 mmol) was dissolved in 3 mL of DMF, followed by the addition of cRGD peptide (1.10 mmol) and triethylamine (3.00 mmol) until fully dissolved. The reaction was carried out at room temperature for 12 h. The reaction mixture was then transferred to a dialysis bag (MWCO 1000 Da) and dialyzed against deionized water for 24 h. The dialysate was collected and lyophilized to obtain DSPE-PEG-cRGDfk with a yield of 39.9%.

Preparation of CRT-sPirb/pLNPs and CRT-cRGDfk/pLNPs

To 1 mL of sPirb in PBS, add SATA dissolved in DMF at a molar ratio of 1:10. Allow the reaction to proceed at room temperature for 30 min. Subsequently, incubate the mixture in PBS containing 0.5 M hydroxylamine and 25 mM EDTA at room temperature for 2 h. Purify the thiol-modified sPirb from the reaction mixture using a PD MidiTrap G-25 column (Cytiva, 28918008). CRT-pLNPs functionalized with DSPE-PEG-TK-Mal were synthesized using microfluidic technology. In the optimized formulation, 50% of DSPE-PEG was replaced with DSPE-PEG-TK-Mal. The functionalized nanoparticles were incubated with thiol-modified sPirb at room temperature for 1 h. Unreacted proteins were removed by purifying the nanoparticles using a qEV/70nm isolation column (IZON), and CRT-sPirb/pLNPs were obtained.

CRT-cRGDfk/pLNPs nanoparticles were prepared using microfluidic technology with a molar ratio of DKP-2-O:cholesterol: DSPE-PEG: DOPE: DOPS: DSPE-PEG-cRGDfk = 45 : 25 : 2.25 : 30 : 5 : 2.25.

Evaluation of in situ programming efficiency of liver or aortic root macrophages by nanoparticles

For the liver, nanoparticles encapsulating Luci mRNA were administered via tail vein injection to MASH mice. After 24 h, D-Luciferin potassium salt was administered intraperitoneally at a dose of 0.15 mg/g. Mice were then euthanized, and the heart, liver, spleen, lungs, and kidneys were harvested. Ex vivo imaging of the isolated organs was performed using the IVIS system to assess Luci mRNA expression levels. Moreover, mice were injected with nanoparticles loaded with CRT mRNA via tail vein injection. After 24 h, the mice were euthanized, and the abdominal cavity was exposed. A portion of the liver was excised, finely minced, and digested with collagenase at 37°C for 1 h with continuous shaking. Cells were filtered through a 70 μm mesh, and red blood cell lysis was performed. Myc-tag (Abcam, ab1394) and F4/80 (BioLegend, 123131) antibodies were used for flow cytometry to label CRT-expressing cells and macrophages, respectively (Gallios, Beckman Coulter). Additionally, a portion of the liver tissue was collected for paraffin embedding and sectioning. Immunofluorescence staining for F4/80 and myc-tag was conducted on the sections, and images were captured under bright field using the VS120-S6-W system.

For the aortic root, nanoparticles encapsulating Luci mRNA were injected via the tail vein into AT mice. After 24 h, D-Luciferin potassium salt (0.15 mg/g) was injected intraperitoneally, followed by euthanasia. The aorta and heart were completely isolated, and ex vivo imaging was performed using IVIS system to evaluate Luci mRNA expression. Additionally, following tail vein injection of CRT mRNA-loaded nanoparticles for 24 h, the mice were euthanized, and frozen sections of the aortic root were prepared. Immunofluorescence staining for F4/80 and myc-tag was performed on the sections, and images were captured using the VS120-S6-W system.

Preparation of tissue sections

For paraffin sections, tissues were fixed in neutral paraformaldehyde for 24 h, dehydrated, and embedded in paraffin. Sections were cut at 5 μm thickness and stored at room temperature. Before staining, the sections were deparaffinized and rehydrated. For frozen sections, tissues were fixed in neutral paraformaldehyde for 24 h, dehydrated, and embedded in OCT. Sections were cut to a thickness of 8–10 μm and stored at −20°C. Before staining, sections were thawed and fixed in formalin for 15 min.

Immunofluorescence staining of the sections

Sections were blocked with 3% hydrogen peroxide and 3% BSA, followed by overnight incubation with the primary antibody targeting F4/80 (Abcam, ab300421), TNF-α (Abcam, ab183218), IL-1β (Abcam, ab315084) or IL-6 (Abcam, ab290735) at 4°C. After washing, the sections were incubated with secondary antibody (Abcam, ab150079 or Abcam, ab150077) at room temperature for 45 min, then TSA reagent was applied for 10 min at room temperature, followed by washing. If no additional antigen staining is required, nuclei were counterstained with DAPI, autofluorescence was quenched, and coverslips were applied for subsequent imaging. Otherwise, antigen retrieval was performed by heating in citrate buffer using a microwave for 10 min. The sections were then incubated overnight at 4°C with the second primary antibody targeting myc-tag (Abcam, ab18185), washed, and incubated with secondary antibody (Abcam, ab150115) at room temperature for 45 min. After washing, TSA reagent was added for 10 min. Finally, sections were counterstained with DAPI, autofluorescence was quenched, and coverslips were applied for subsequent imaging.

Western blot of TREM2

After centrifugation, BMDMs or liver macrophages were resuspended in RIPA buffer containing a protease inhibitor and phosphatase inhibitor cocktail. Proteins were separated by SDS-PAGE and transferred onto a nitrocellulose membrane. The membrane was blocked at room temperature for 15 min using a rapid blocking buffer (Beyotime P0252). It was then incubated overnight at 4°C with the TREM2 antibody (Abcam, ab305103). Subsequently, the membrane was incubated with a secondary antibody (Abcam, ab6721) at room temperature for 1 h, and developed with SuperSignal West Pico PLUS (Thermo Fisher Scientific). Images were recorded with an Amersham Imager 600 RGB.

In situ efferocytosis assay

Liver or aortic root paraffin sections were blocked with 3% BSA and incubated with anti-F4/80 antibody (Abcam, ab300421) for 1 h, followed by incubation with a fluorescent secondary antibody (Abcam, ab150077) for 45 min. After washing, the sections were stained according to a One Step TUNEL Apoptosis Assay Kit (Beyotime, C1090). DAPI was used for counterstaining. Confocal images were captured by LSM 900 with Airyscan 2 (ZEISS). Efferocytosis capacity was evaluated by calculating the ratio of TUNEL-positive apoptotic cells colocalized with F4/80-positive macrophages to total apoptotic cells.

Evaluation of therapeutic efficacy in MASH

Frozen liver tissue sections were stained with Oil Red O to assess the degree of steatosis, while paraffin sections were stained with hematoxylin and eosin (H&E) to evaluate the general tissue morphology. The severity of steatosis, lobular inflammation, and hepatocyte ballooning was scored on H&E sections according to the NAS system. To detect the extent of liver fibrosis, paraffin sections were incubated in 3% hydrogen peroxide solution for 25 min to block endogenous peroxidase activity. After washing, the sections were blocked with 2% BSA for 30 min, then incubated with primary anti-α-SMA antibody (Abcam, ab5694) for 1 h. Subsequently, the sections were incubated with an HRP-conjugated secondary antibody (ab6721) at room temperature for 45 min. Color development was performed by 3,3′-diaminobenzidine (DAB), and the nuclei were counterstained with DAPI. Additional paraffin sections were stained with Sirius Red to visualize collagen deposition, and the collagen-positive area was quantified by ImageJ software. The sections were then dehydrated and mounted. Images of the sections were captured under bright field by VS120-S6-W system (Olympus). Image quantification was performed by ImageJ software. Tissue hydroxyproline levels were measured by commercial assay kits (Solarbio, BC0255).

For the evaluation of blood parameters, blood samples were collected from mice via retro-orbital bleeding and allowed to clot at room temperature for 1 h. The samples were then centrifuged, and the supernatant serum was collected. Serum concentrations of ALT and AST were measured by commercial assay kits (Solarbio). The concentrations of TNF-α, IL-1β, and IL-6 in serum were measured by ELISA kits (Beyotime).

Liver tissues from MASH mice were collected and homogenized in RIPA buffer supplemented with protease and phosphatase inhibitors. Protein concentrations were determined using a BCA Protein Assay Kit (Beyotime, P0012). Equal amounts of protein were resolved by SDS-PAGE and transferred onto nitrocellulose membranes. Membranes were blocked with rapid blocking buffer (Beyotime, P0252) and incubated with primary antibodies against SCD1 (Abcam, ab236868), α-SMA (Abcam, ab5694), COL1A1 (Abcam, ab270993), TNF-α (Abcam, ab183218), and IL-1β (Abcam, ab315084), followed by an HRP-conjugated secondary antibody (Abcam, ab6721). Signals were developed using SuperSignal West Pico PLUS chemiluminescent substrate (Thermo Fisher Scientific, 34577) and captured with an Amersham Imager 600 RGB system.

Moreover, after dissection, mouse liver tissues were immediately immersed in RNAsolid RNA Stabilization Solution for Tissue (Servicebio) to preserve RNA integrity (Servicebio, G3019) to preserve RNA integrity. Primer design, RNA extraction, reverse transcription, and subsequent RT-qPCR analysis were performed by Servicebio. The mRNA levels of Col1a1, Scd1, Tnf, and Il1b were quantified and normalized to housekeeping genes to evaluate hepatic fibrosis, lipid metabolism, and inflammation.

Flow cytometric analysis of hepatic myeloid cell populations

Liver immune cell populations were analyzed by flow cytometry. After euthanizing the mice, the abdominal cavity was opened, and a portion of liver tissue was excised and finely minced. The tissue was digested with 0.75 mg/mL collagenase A at 37°C for 1 h with constant shaking. The resulting cell suspension was filtered through a 70 μm strainer, followed by red blood cell lysis.

For macrophage analysis, cells were stained with anti-CD45 (Invitrogen, 103108), anti-LY6G (Abcam, ab210207), anti-CD11b (Abcam, ab79096), and anti-F4/80 (BioLegend, 123131) antibodies to identify the macrophage population, and anti-CD86 (BioLegend, 105115) and anti-CD206 (BioLegend, 141731) antibodies were used to distinguish M1 and M2 macrophages, respectively (Gallios, Beckman Coulter). To determine the proportion of monocyte-derived macrophages, cells were additionally stained with anti-TIM-4 (BioLegend, 129905) and anti-CLEC4F (BioLegend, 156803) antibodies to label liver-resident macrophages. TIM4⁻CLEC4F⁻ double-negative cells within the F4/80⁺CD11b⁺ population were considered monocyte-derived macrophages.

To assess hepatic neutrophil content, the digested liver cell suspensions were stained with anti-CD45, anti-LY6G, and anti-CD11b antibodies as above, and neutrophils were identified as CD45⁺LY6G⁺CD11b⁺ cells.

Evaluation of therapeutic efficacy in atherosclerosis

Frozen sections of the aortic root, aortic arch, and brachiocephalic artery were stained with Oil Red O, and the total plaque area was assessed by quantifying Oil Red O-positive regions. H&E staining was performed on frozen aortic root sections to measure the necrotic core size within the plaques. Sirius Red staining was also applied to the aortic root sections to evaluate collagen content and fibrous cap thickness in the plaques. Images of the sections were captured under bright field by VS120-S6-W system (Olympus). Image quantification was performed by ImageJ software. To evaluate serum inflammatory cytokine, TNF-α, IL-1β, and IL-6 concentrations were measured by ELISA kits (Beyotime, PT512, PI301, PI326). Furthermore, immunofluorescence staining was performed on aortic root sections for TNF-α, IL-1β, or IL-6. Images of the sections were captured by VS120-S6-W system (Olympus) and image quantification was performed by ImageJ software.

TUNEL apoptosis assay

Liver and aortic root sections were stained following the protocol provided in the One Step TUNEL Apoptosis Assay Kit (Beyotime, C1090). Nuclei were counterstained with DAPI, and the sections were washed and mounted. Images were captured by the fluorescence mode of VS120-S6-W system (Olympus).

Safety evaluation

Male C57BL/6 mice and LDLr-KO C57BL/6 mice were treated with the designated reagents for 4 weeks. At the end of the treatment, blood samples were collected, and the mice were sacrificed. The heart, liver, spleen, lung, and kidney were harvested and weighed, then the organ coefficients were calculated. Portions of the tissues were sectioned and stained with H&E to assess potential organ damage. Blood samples were analyzed for safety-related biomarkers, including ALT, AST, blood urea nitrogen (BUN), creatinine (CRE), red blood cell count (RBC), white blood cell count (WBC), platelets (PLT), and hemoglobin (Hb).

Quantification and statistical analysis

All statistical analyses and figure generation were conducted by GraphPad Prism software. Data are presented as mean ± standard deviation, with error bars representing the standard deviation of independent sample means. For significance analysis between two groups, a Student’s t test was employed. Comparisons among multiple groups were performed by one-way or two-way ANOVA. The p values were indicated in the figure, with no statistical significance when p ≥ 0.05.

Published: January 29, 2026