Spatiotemporally Resolved Protein Tagging for Dynamic and Tissue‐Specific Secretome Profiling In Vivo

ABSTRACT

The dynamic and tissue‐specific nature of protein secretion underlies a wide range of physiological and pathological processes, yet tools for profiling the in vivo secretome with high spatial and temporal resolution remain limited. Here, we present STePTag (Spatio‐Temporal Protein Tagging), a conditional proximity labeling (PL) system for profiling secreted proteins in live animals. STePTag integrates the rapid labeling kinetics of the PL enzyme TurboID with a destabilized dihydrofolate reductase (DHFR) domain, enabling tight post‐translational control of labeling activity through the small‐molecule stabilizer trimethoprim (TMP). Targeting STePTag to the endoplasmic reticulum (ER) confines labeling to the secretory pathway and enables robust labeling within 10 min of TMP administration. Furthermore, we design tissue‐specific lipid nanoparticles (tsLNPs) to enable programmable, in vivo delivery of STePTag to the mouse liver. Using this approach, we identified 93 liver‐derived secretory proteins under physiological conditions and uncovered 40 dynamically regulated proteins in a model of acetaminophen‐induced acute liver injury (ALI), including Aldh1a1, which we functionally validated as a protective factor in ALI. Together, STePTag provides a versatile platform for spatiotemporally resolved secretome profiling, enabling the discovery of context‐dependent biomarkers and tissue‐derived signaling molecules in native physiological environments.

STePTag is a conditional proximity labeling platform enables spatiotemporally resolved in vivo secretome profiling. By detecting circulating proteins to their origin, STePTag mapped homeostatic secretomes and uncovered a protective factor in acute liver injury, offering a versatile tool to decode inter‐organ communication and discover biomarkers.

1. Introduction

Secreted proteins are central mediators of intercellular communication, immune surveillance, tissue homeostasis, and systemic adaptation to physiological and pathological cues [1]. The secretome‐the ensemble of proteins actively released from cells‐represents a rich but underexplored source of disease biomarkers and therapeutic targets [2, 3]. Despite its biological and clinical significance, comprehensive and dynamic characterization of the secretome in vivo, particularly with tissue‐specific and temporally resolved precision, remains challenging. Mass spectrometry‐based analysis of conditioned media from cultured cells or ex vivo tissues has provided a high‐throughput means to profiling protein secretion across diverse cell types and conditions [4, 5]. However, in vitro secretome signatures often diverge markedly from those observed in intact organisms, limiting their physiological relevance and translational utility [5].

Proximity labeling (PL) has recently emerged as a powerful strategy to resolve subcellular proteomes and map secreted proteins with spatial precision in living systems [6, 7, 8, 9, 10]. Engineered enzymes such as TurboID and APEX enable covalent tagging of proteins in proximity, facilitating their enrichment and identification by mass spectrometry [11, 12, 13, 14, 15]. Although PL has been successfully used for secretome profiling in vivo, most existing approaches label secretory proteins constitutively, leading to elevated background and limited temporal resolution [16, 17, 18, 19, 20, 21, 22]. Inducible PL variants, including split‐TurboID [23], LOV‐Turbo [24], and chemically triggered systems [25], offer improved control but remain constrained by spatial limitations, such as poor tissue penetration of light‐based activation [24, 25, 26] or insufficient tissue specificity of small‐molecule inducers (e.g., rapamycin) [27, 28]. Consequently, a PL strategy that combines tight temporal regulation with tissue‐specific control is still lacking, hindering precise interrogation of dynamic, context‐dependent secretion events in vivo.

In this study, we report STePTag (Spatio‐Temporal Protein Tagging), a conditional PL system that enables high‐resolution profiling of the in vivo secretome. STePTag integrates the rapid labeling kinetics of TurboID with a destabilized dihydrofolate reductase (DHFR) domain, allowing post‐translational control of enzyme activity via the small molecule trimethoprim (TMP) [29]. Targeting STePTag to the endoplasmic reticulum (ER) lumen confines labeling to the secretory pathway (Figure 1a). In parallel, we achieve liver‐specific secretome profiling through targeted delivery of STePTag‐encoding mRNA using lipid nanoparticles (LNPs), providing a non‐viral and programmable platform for in vivo secretome labeling (Figure 1b). We show that STePTag enables robust labeling of secretory proteins within 10 min of TMP administration in cultured cells and captures over 90 endogenous secreted proteins in the mouse liver. Moreover, application of STePTag to a model of acetaminophen‐induced acute liver injury (ALI) revealed 40 dynamically regulated liver‐derived secretory proteins, including Aldh1a1, which we functionally validated as a protective factor in ALI. Together, these results establish STePTag as a versatile and scalable platform for spatiotemporally resolved secretome profiling in live animals, with broad applications in biomarker discovery, disease modeling, and therapeutic development.

FIGURE 1.

Design and in vivo application of STePTag for spatiotemporally resolved secretome profiling. (a) Schematic illustration of the STePTag design, in which a destabilized DHFR domain confers TMP‐dependent control of TurboID activity, enabling spatiotemporally resolved proximity labeling of secretory proteins within the endoplasmic reticulum. (b) Overview of liver‐specific secretome profiling achieved by LNP‐mediated delivery of STePTag‐encoding mRNA, followed by TMP‐induced activation and enrichment of biotinylated liver‐derived secretory proteins from circulation.

2. Results and Discussion

2.1. Design and Validation of STePTag for Temporally‐Resolved Secretome Labeling

To enable temporal control over PL of secretory proteins, we designed STePTag by genetically fusing a destabilization domain (DD) derived from E. coli DHFR to the N‐terminus of TurboID, a fast‐acting biotin ligase (Figure 1a). In the absence of the small‐molecule stabilizer TMP, the DHFR‐fused STePTag undergoes rapid proteasomal degradation. TMP administration stabilizes the DD, restores TurboID activity of STePTag, and initiates proximity‐dependent biotinylation. To confine labeling to secretory proteins, we appended an ER‐anchoring sequence (Sec61b) to localize STePTag to the ER lumen. This modular design enables conditional, spatially restricted, and temporally precise labeling of secretory proteins in live cells.

We first validated TMP‐dependent PL activity of STePTag in HepG2 human liver carcinoma cells. To this end, cells were transfected with mRNA encoding STePTag (mSTePTag, 400 ng/mL) bearing a V5 epitope tag and treated with TMP (1 µM) and biotin (50 µM) 12 h post‐transfection. As a control, we used a constitutively active ER‐anchored TurboID construct lacking the DHFR domain (Sec61b‐TurboID, or Sec‐Tb). In the absence of TMP, V5 immunofluorescence was undetectable from mSTePTag‐transfected cells, indicating the rapid proteasomal degradation of STePTag (Figure 2a). TMP treatment led to a robust V5 signal that colocalized with streptavidin‐FITC staining, indicating efficient biotinylation of proximal proteins. The labeling pattern closely resembled that of Sec‐Tb, confirming TMP‐dependent stabilization and functional PL by STePTag.

FIGURE 2.

Validation of TMP‐dependent activation and temporal control of STePTag in cultured cells. (a) Immunofluorescence analysis of STePTag or Sec61b‐TurboID (Sec‐Tb) expression (anti‐V5) and biotinylated proteins (streptavidin‐FITC) in HepG2 cells transfected with mSTePTag or mSec‐Tb, followed by treatment with TMP (1 µM) and biotin (50 µM). Scale bars, 20 µm. (b, c) Immunoblot analysis of biotinylated proteins (streptavidin‐HRP) and STePTag (anti‐V5) in cell lysates (b) and culture supernatants (c) from HEK293T cells transfected with m STePTag for 12 h and subsequently incubated with increasing concentrations of TMP (0.01–1 µM) for an additional 16 h. (d, e) Immunoblot analysis of biotinylated proteins (streptavidin‐HRP) and STePTag (anti‐V5) in cell lysates (d) and culture supernatants (e) from HEK293T cells transfected with mSTePTag and incubated with TMP (1 µM) for the indicated times.

We next assessed the TMP dose dependence of STePTag stabilization and labeling. HEK293T cells transfected with mSTePTag (200 ng/mL) were treated with increasing concentrations of TMP (10 nM to 1 µM) along with 50 µM biotin. Immunoblot analysis revealed a TMP dose‐dependent increase in intracellular STePTag levels (Figure 2b). Correspondingly, streptavidin‐HRP detection showed robust biotinylation of proteins in both cell lysates and culture supernatants across the studied TMP concentrations (Figure 2c). Notably, stabilization and labeling were detectable at TMP concentrations as low as 10 nM, indicating the high sensitivity of STePTag. This low TMP requirement is advantageous for in vivo applications, where minimizing systemic exposure is critical for achieving temporally resolved labeling with minimal physiological perturbation.

We next characterized the kinetics of TMP‐mediated STePTag stabilization and secretory protein labeling. In HEK293T cells transfected with mSTePTag (200 ng/mL), TMP (1 µM), and biotin (50 µM) were added at time intervals ranging from 0 to 18 h. Immunoblot analysis revealed stabilization of STePTag and detectable biotinylation of proteins within 10 min after TMP addition (Figure 2d,e), highlighting the rapid activation and suitability of STePTag for monitoring dynamic secretory events. In addition, protein biotinylation increased over time and sustained throughout the 18‐h labeling period, indicating the stability and effectiveness of STePTag for secretome labeling once activated. Notably, after 6 h of TMP treatment (1 µM), STePTag labeling efficiency was comparable to that of the constitutively active Sec‐Tb control (Figure 2d), indicating that TMP‐stabilized STePTag achieves full functionality for protein labeling. Compared to previously reported light‐activated or bioorthogonal‐triggered PL systems, STePTag offers comparable temporal resolution without the limitations of light penetration or chemical specificity in vivo. Furthermore, it outperforms existing small‐molecule‐activated split‐TurboID systems in activation speed, labeling efficiency, and experimental practicality (Table S1). Together, these results establish STePTag as a tunable, rapid, and efficient platform for temporally gated labeling of secretory proteins, well‐suited for capturing dynamic secretome events in living systems with high spatiotemporal precision.

2.2. Design of Biodegradable LNPs for Tissue‐Specific Delivery of STePTag

In vivo secretome profiling with PL enzymes has traditionally relied on viral delivery or the generation of transgenic animal models driven by tissue‐specific promoters [16, 17]. Although these approaches can achieve spatial targeting, they are often limited by constitutive labeling, restricted temporal resolution, and dependency on well‐characterized tissue‐specific promoters (Table S2). To overcome these limitations, we developed a non‐viral, biodegradable LNPs platform for delivering mRNA encoding STePTag [30]. This strategy enables flexible, tissue‐specific, and temporally controlled secretome labeling without the need for genetic modification or viral transduction.

LNPs typically consist of an ionizable lipid, helper lipid, polyethylene glycol (PEG)‐lipid, and cholesterol; they have demonstrated broad utility for mRNA delivery [31], including clinical applications such as mRNA‐based COVID‐19 vaccines [32, 33]. Among these components, the chemical structure of the ionizable lipid is the primary determinant of mRNA encapsulation, endosome escape, and delivery efficiency [34]. To systematically optimize LNPs for tissue‐specific delivery of mSTePTag, we synthesized a series of thioketal‐bearing ionizable lipids (AMPA‐TKn) via Michael addition of N,N‐bis(3‐aminopropyl)methylamine (AMPA) to biodegradable acrylates with varying hydrophobic tail lengths (Figure 3a) [35]. AMPA was selected based on previous findings that AMPA‐containing lipids exhibit high cellular uptake and endosomal escape efficiency in delivering biologics [36]. Moreover, the introduction of the thioketal moiety renders the lipids reactive oxygen species (ROS)‐degradable, facilitating intracellular mRNA release in oxidative environments and enhancing delivery efficiency [30].

FIGURE 3.

Design and validation of biodegradable lipid nanoparticles for tissue‐specific mRNA delivery. (a) Chemical structures of biodegradable ionizable lipids (AMPA‐TKn) used for formulation of tissue‐specific lipid nanoparticles for delivery of mLuc, mSec‐Tb, and mSTePTag. (b) Schematic illustration of tissue‐specific mRNA delivery mediated by LNPs. (c, d) Representative whole‐body bioluminescence images (c) and quantitative analysis of in vivo mLuc delivery efficiency (d) following intravenous administration of mLuc (0.2 mg/kg) encapsulated in liver‐targeting AMPA‐TKn LNPs or DLin‐MC3‐DMA LNPs, measured by IVIS imaging 6 h post‐injection (n = 3). (e) Immunoblot analysis of Sec‐V5‐TurboID (anti‐V5) expression in tissue lysates from mice administered liver‐targeting mSec‐V5‐Tb/AMPA‐TK4 LNPs for 6 h.

To evaluate the efficacy of these lipids for in vivo mRNA delivery, firefly luciferase mRNA (mLuc) was encapsulated into LNPs formulated with each AMPA‐TKn and administered intravenously to BALB/c mice at a mRNA dosage of 0.2 mg/kg. The LNPs were formulated using AMPA‐TKn, cholesterol, 1,2‐dioleoyl‐sn‐glycero‐3‐phosphoethanolamine (DOPE), and DMG‐PEG₂₀₀₀ at a molar ratio of 50:38.5:10:1.5 (Figure 3b). Meanwhile, the clinically validated ionizable lipid, DLin‐MC3‐DMA, was used as a positive control. All mLuc‐formulated LNPs were homogeneous (PDI < 0.15) and exhibited a zeta potential between ‐5 and 0 mV, showing a particle size ranging from 87.5 to 159.5 nm (Table S3). Whole‐body bioluminescence imaging performed 6 h after systemic administration revealed predominant luciferase expression in the liver for all AMPA‐TKn formulations. Notably, mRNA delivery efficiency was strongly influenced by hydrophobic tail length, with shorter‐tail lipids showing higher luciferase expression (Figure 3c). Among the studied five lipids, AMPA‐TK4 exhibited the highest liver‐specific luciferase signal, outperforming DLin‐MC3‐DMA as well as AMPA‐TK6 and AMPA‐TK8 (Figure 3d).

We next validated the utility of the optimized LNPs for liver‐specific delivery of PL enzymes, given that the role of the liver as the primary secretory organ. To verify this potential, mRNA encoding Sec61b‐TurboID (mSec‐Tb) was encapsulated in AMPA‐TK4 LNPs and administrated systemically to BALB/c mice. Liver‐targeted delivery of mSec‐Tb (1 mg/kg mRNA) resulted in robust Sec‐Tb expression in the liver (Figure 3e), with minimal signal detected in other tissues. Together, these results establish a programmable and biodegradable LNPs platform for liver‐specific delivery of PL enzymes. By coupling LNP‐mediated tissue‐specific mRNA delivery with conditional PL, this approach may enable precise spatial control of labeling activity and support dynamic, in vivo secretome profiling without reliance on viral vectors or genetic animal models.

2.3. Spatiotemporally Resolved Profiling of the Liver Secretome Using STePTag

To evaluate the effectiveness of STePTag for temporally controlled secretome labeling in vivo, mSTePTag was encapsulated in AMPA‐TK4 LNPs (hereafter referred to Liver‐STePTag) and administrated systematically to BALB/c mice (1 mg/kg mSTePTag). Transmission electron microscopy (TEM) characterization confirmed uniform nanoparticle morphology of Liver‐STePTag (Figure S1). Thirty minutes after Liver‐STePTag administration, mice were treated intraperitoneally with TMP and biotin (50 mg/kg each) to activate STePTag. Liver tissues and serum were collected 12 h later to assess STePTag expression and secretory proteins labeling, respectively.

Mice treated with Liver‐STePTag and TMP exhibited robust hepatic expression of STePTag (Figures 4a and S2), accompanied by the efficient biotinylation of serum proteins (Figure 4b). In contrast, mice that did not receive TMP showed negligible biotinylation, demonstrating tight temporal control of secretory protein labeling using Liver‐STePTag. Notably, biotinylated proteins were only detected when TMP and biotin were co‐administered (Figure 4b), indicating that STePTag‐mediated labeling is both inducible and temporally restricted in vivo. However, delivery of a constitutively active ER‐targeted mSec‐Tb using AMPA‐TK4 LNPs resulted in persistent labeling of serum proteins, even in the absence of exogenous biotin, reflecting basal TurboID activity and endogenous biotin availability, which further highlighted the enhanced temporal specificity achieved with STePTag for low‐background labeling. Alanine aminotransferase (ALT), aspartate aminotransferase (AST), interleukin‐6 (IL‐6), and interleukin‐1β (IL‐1β) levels were not significantly altered in mice treated with Liver‐STePTag or the biotinylation labeling group compared to controls. This indicates that neither the administration of Liver‐STePTag nor the subsequent TMP‐mediated stabilization induced noticeable hepatic toxicity or systemic inflammatory responses, demonstrating the high biocompatibility of Liver‐STePTag for in vivo secretome labeling (Figure S3). Notably, while our macroscopic biochemical assessments confirm the absence of overt liver injury and systemic inflammation, as evidenced by the stable serum profiles, we explicitly recognize the potential intrinsic stress and potential toxicity associated with LNP delivery vehicles, which may remain a limitation when profiling native physiological secretomes.

FIGURE 4.

Spatiotemporally resolved profiling of the liver secretome using STePTag. (a) Immunoblot analysis of STePTag expression (anti‐V5) in tissue lysates from mice administered Liver‐ STePTag for 12 h. (b) Streptavidin‐HRP detection of biotinylated proteins in serum from mice administered Liver‐ STePTag or mSec‐Tb LNPs, followed by intraperitoneal injection of TMP and biotin (50 mg/kg each). Experiments were performed as biological triplicates with similar results. (c, d) Time‐course immunoblot analysis of STePTag expression (anti‐V5) in liver lysates (c) and biotinylated proteins (streptavidin‐HRP) in serum (d) from mice administered Liver‐ STePTag and subsequently treated with TMP and biotin (50 mg/kg each). (e) Volcano plot of streptavidin bead‐enriched serum proteins from mice treated with Liver‐ STePTag, TMP, and biotin. (f) Number of biotinylated proteins identified in the Liver‐ STePTag dataset that contain predicted signal peptides, as determined by SignalP 6.0. (g) Gene ontology enrichment analysis of liver‐derived secretory proteins identified by Liver‐ STePTag. (h) Venn diagram comparing enriched biotinylated secretory proteins identified by Liver‐ STePTag with those detected using previously reported TurboID‐based secretome profiling methods.

To assess the secretory protein labeling kinetics using STePTag, we administered BALB/c mice intravenously with Liver‐STePTag (1 mg/kg mSTePTag), followed by intraperitoneal injection of TMP and biotin (50 mg/kg each) 30 min later. Serum samples were collected at 0.5, 1, 3, and 6 h post‐TMP injection, and biotinylated proteins were enriched for analysis. Both hepatic STePTag expression (Figure 4c) and serum biotinylation levels (Figure 4d) increased in a time‐dependent manner. Notably, biotinylated proteins were detectable as early as 30 min after TMP administration (Figure 4d), indicating the rapid activation and labeling capability of STePTag. This temporal resolution represents a substantial improvement over transgenic or viral delivery approaches, which typically require days to weeks to achieve effective PL in vivo (Table S1). To further demonstrate the temporal dynamics of STePTag, we performed secretory labeling at various time points following TMP administration after mSTePTag delivery. The results indicated that STePTag activation yields consistent secretory protein labeling at different time intervals, ranging from 0.1 to 4 h, confirming that STePTag maintains robust activation kinetics across a wide range of time windows for dynamic analysis of secretory protein (Figure S4).

To determine whether STePTag selectively labels liver‐derived secretory proteins, we performed label‐free quantitative proteomic analysis based on liquid chromatography and tandem mass spectrometry (LC‐MS/MS) of biotinylated proteins enriched from the serum of mice treated with Liver‐STePTag (1 mg/kg mSTePTag), TMP, and biotin (50 mg/kg each). We identified 93 proteins that were selectively enriched in the serum of Liver‐STePTag‐treated mice (Figure 4e). A comprehensive list of these proteins is provided in the Supporting Information S2: Data 1. Furthermore, 82.8% (77/93) of these proteins were found to be covalently biotinylated when profiled by directly measuring the biotinylated peptides using previously reported methods [37] (Figure S9; Supporting Information S2: Data 1 and 2). Signal peptide analysis using SignalP 6.0 [38] revealed that 98.9% (92/93) of these proteins contain ER‐targeting signal peptides (Figure 4f), consistent with their classification as secretory proteins. Using BioGPS [39] transcriptomic data from 191 tissues and cell types, we confirmed that all 93 proteins were expressed in the liver, with 65.6% (61/93) showing liver‐predominant expression (≥ 25% of total transcript abundance; Figure S5). These findings confirm that Liver‐STePTag enables selective and temporally precise labeling of liver‐derived secretory proteins in circulation. Gene ontology (GO) analysis revealed that these proteins are involved in biological processes such as complement activation, acute‐phase response, blood coagulation, and innate immune response (Figure 4g), consistent with known liver‐secreted protein roles. Of the 93 identified proteins, 71.0% (66/93) overlapped with liver‐derived secretory proteins previously reported using traditional PL approaches (Figures 4h and S6), including albumin (Alb), apolipoproteins (Apoa1, Apob, Apoh, Lcat), coagulation factors (F5, F9, F12, F13b), and complement factors (C2, C5, C8a, C8b, C9). Importantly, 27 proteins were uniquely identified by STePTag compared to prior methods (Figure 4h, Supporting Information S2: Data 1). These include hepatocyte‐derived secretory proteins such as apolipoproteins (Apoc1, Apoc2, Apoa4), coagulation factor F7, complement protein C1rl, acute‐phase proteins (Apcs, Orm1, Saa1, Saa2), and lipopolysaccharide‐binding protein (Lbp). Additionally, STePTag captured secreted proteins typically considered multi‐tissue or immune‐derived, such as ficolin‐1 (Fcn1), immunoglobulin J chain (Igj), macrophage colony‐stimulating factor (Csf1), and (Metalloproteinase inhibitor 3) Timp3, indicating that the liver is also the source of these proteins. Notably, STePTag also labeled membrane‐associated proteins, including multiple MHC‐I molecules such as H‐2 class I histocompatibility antigens (H2‐D1, H2‐K1, H2‐Q8) [40], which are involved in antigen presentation and immune responses in the liver (Figure S7). These proteins are typically anchored at the cell surface and mediate antigen presentation and immune surveillance in the liver. The extracellular peptides detected by LC‐MS/MS mapped to annotated extracellular domains, consistent with ectodomain shedding rather than full‐length protein release. This indicates that STePTag can capture post‐translational cleavage events and trace their tissue origin in vivo, offering insights into dynamic immune signaling and surface protein turnover. Together, these results established Liver‐STePTag as a rapid, inducible, and highly specific platform for spatiotemporally resolved profiling of the liver secretome. This approach not only recapitulates known liver‐secreted factors but also reveals previously undetected or poorly characterized proteins, including immune‐related and membrane‐shed components.

2.4. Liver‐Specific Secretome Profiling in Drug‐Induced ALI

To validate the capabilities of STePTag for secretome profiling under pathological conditions, we applied Liver‐STePTag to profile secretory proteins during acetaminophen (APAP)‐induced ALI, a well‐established model of drug‐induced hepatotoxicity [41]. APAP overdose causes oxidative stress, mitochondrial dysfunction, and hepatocellular necrosis, and is accompanied by profound alterations in liver‐derived secretory proteins, which represent potential biomarkers and therapeutic targets in ALI [42].

BALB/c mice were administered Liver‐STePTag (1.5 mg/kg mSTePTag), followed by a single intraperitoneal dose of APAP (600 mg/kg) to induce ALI (Figure 5a). Thirty minutes later, TMP and biotin (50 mg/kg each) were administered to initiate PL. Serum was collected 7 h after APAP challenge, and biotinylated proteins were enriched for proteomic analysis. Using this approach, we identified 40 secretory proteins that were significantly upregulated in APAP‐treated mice compared with untreated controls (Figure 5b). Among these, 20 proteins have previously been implicated in ALI (Supporting InformationS2: Data 3) [43, 44, 45, 46], including known protective factors such as nicotinamide phosphoribosyltransferase (Nampt) [47], cytosolic non‐specific dipeptidase (Cndp2) [48], and dehydrogenase/reductase SDR family member 1 (Dhrs1) [45]. Notably, insulin‐like growth factor‐binding protein 1 (Igfbp1) and cathepsin L1 (Ctsl) were markedly upregulated in the liver secretome of APAP‐challenged mice [43], despite previous reports describing their downregulation in primary hepatocyte cultures following APAP exposure. This discrepancy underscores the importance of studying secretome dynamics in intact physiological systems. In addition, 20 secretory proteins not previously associated with ALI were identified as upregulated using Liver‐STePTag (Supporting Information S2: Data 3), including chaperones Hspa5 and Hspa8, and nicotinate phosphoribosyltransferase (Naprt). GO analysis linked these proteins to biological processes including NAD⁺ biosynthesis, retinoid metabolism, oxidative stress response, and xenobiotic stimulus pathways (Figure 5c).

FIGURE 5.

Liver‐specific secretome profiling reveals Aldh1a1 as a protective factor in acute liver injury. (a) Experimental schematic for Liver‐STePTag administration, TMP‐induced activation, and biotin labeling in the APAP‐induced acute liver injury (ALI) model. (b) Volcano plot of streptavidin‐enriched serum proteins from mice subjected to Liver‐STePTag secretome labeling with or without APAP treatment. (c) Gene ontology enrichment analysis of upregulated liver‐derived secretory proteins identified in APAP‐induced ALI using Liver‐STePTag. (d) Schematic illustration of intracellular delivery of mAldh1a1/AMPA‐TK4 LNPs to mitigate ALI. (e, f) Serum ALT (e) and AST (f) levels in APAP‐challenged mice treated with mAldh1a1/AMPA‐TK4 LNPs (n = 3). (g) Representative H&E‐stained liver sections showing reduced centrilobular necrosis in mice treated with mAldh1a1/AMPA‐TK4 LNPs following APAP overdose. Livers were harvested 4 h after APAP administration. (h) Serum MDA levels in APAP‐challenged mice treated with mAldh1a1/AMPA‐TK4 LNPs, measured 4 h after APAP administration (n = 3). Mice were challenged with APAP at a dose of 350 mg/kg. Statistical analysis was performed using one‐way ANOVA followed by Tukey's multiple‐comparison test. ***p < 0.001, ****p < 0.0001. Data are presented as mean ± SD.

Among the five most highly enriched proteins were Dhrs1, Igfbp1, Ctsl, aldehyde dehydrogenase 1A1 (Aldh1a1), and elongation factor 1‐delta (Eef1d) (Figure 5b). Notably, Aldh1a1 is an enzyme involved in lipid and retinoid metabolism as well as aldehyde detoxification [49]. It catalyzes the oxidation of lipid peroxidation‐derived aldehydes, such as malondialdehyde (MDA), into less‐toxic acids (Figure 5d). To definitively confirm its secretion, we performed orthogonal validations by enriching both biotinylated proteins from the serum of STePTag‐treated mice following APAP induction. Immunoblotting successfully corroborated the specific presence of Aldh1a1 within the biotinylated secretome fraction (Figure S8). Given its robust induction in the liver secretome following APAP challenge, we hypothesized that Aldh1a1 may exert a protective role in ALI by mitigating oxidative stress. To verify this hypothesis, Aldh1a1 mRNA (mAldh1a1) was encapsulated in liver‐targeting AMPA‐TK4 LNPs and administered intravenously at a dose of 1.5 mg/kg 4 h prior to APAP exposure (350 mg/kg). Liver injury was assessed 4 h after APAP exposure. Indeed, APAP treatment alone resulted in marked elevations in serum ALT and aspartate aminotransferase (AST) levels, indicating hepatocellular injury (Figure 5e,f). In contrast, mice pretreated with mAldh1a1/AMPA‐TK4 LNPs showed a ∼65% reduction in ALT and a ∼60% reduction in AST compared with APAP‐treated controls (Figure 5e,f). Histopathological analysis using hematoxylin and eosin (H&E) staining further confirmed the protective effects, revealing extensive hepatic necrosis in APAP‐treated mice, which was substantially attenuated in animals receiving mAldh1a1 pretreatment (Figures 5g and S10).

To further validate the protective role of Aldh1a1, we examined the changes in oxidative stress markers, such as MDA [50], in the serum of APAP‐challenged mice with and without mAldh1a1/AMPA‐TK4 LNPs pre‐treatment. Pretreatment with mAldh1a1/AMPA‐TK4 LNPs reduced serum MDA levels by up to 80%, restoring them to near‐baseline levels observed in untreated mice (Figure 5h). Together, these results demonstrate that STePTag enables dynamic, liver‐specific secretome profiling in vivo under pathophysiological conditions. The identification of Aldh1a1 as a protective secreted factor highlights the utility of STePTag for discovering mechanistically informative biomarkers and therapeutic candidates in ALI and other disease contexts.

3. Conclusion

In summary, we have developed STePTag, a conditional PL system that enables spatiotemporally resolved secretome profiling in vivo, allowing direct attribution of circulating proteins to their tissue origin in both physiological and pathological contexts. By integrating conditional PL with tissue‐specific mRNA delivery via lipid nanoparticles, we applied STePTag to the mouse liver and successfully mapped its physiological secretome while identifying dynamically regulated proteins during acute liver injury, including Aldh1a1, which we functionally validated as a protective factor. Nevertheless, the STePTag approach facilitates the discovery of context‐dependent biomarkers and functionally relevant secreted proteins, thereby advancing the study of inter‐organ communication, disease mechanisms, and therapeutic target identification. More broadly, STePTag establishes a versatile platform for systematically dissecting organ‐specific protein secretion dynamics in complex living systems.

Author Contributions

Qizhen Zheng: conceptualization, methodology, data curation, investigation, validation, writing – original draft. Rui Yao: methodology, investigation. Tianyu Ma: methodology. Lijuan Li: methodology. Ying Jiang: conceptualization, writing – review and editing. Ming Wang: conceptualization, methodology, supervision, writing – review and editing, writing – original draft, funding acquisition, investigation, validation.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: anie72890‐sup‐0001‐SuppMat.docx.

Supporting File 2: anie72890‐sup‐0002‐Data.zip.

Data Availability Statement

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