Aptamer-Driven Nanozyme-SERS Bifunctional Sensors with Enrichment Separation Effect for Ultrasensitive Detection of Liver-Fibrosis-Related Biomarkers

Abstract

Liver fibrosis represents the earliest and most critical pathological stage in the progression of various liver diseases. Its accurate detection is paramount for preventing transformation into cirrhosis and hepatocellular carcinoma. Here, we report an aptamer-driven nanozyme-catalyzed surface-enhanced Raman scattering (SERS) dual-functionality sensor. By integrating magnetic enrichment and separation effects, we achieve multiple signal amplification, enabling the ultrasensitive detection of fibrosis-associated miR-34a. In this study, we synthesized gold–platinum nanoparticles (Au@Pt NPs) exhibiting high peroxidase activity and intrinsic surface-enhanced Raman scattering (SERS) properties. These were surface-modified with the complementary strand (H₁) of an aptamer to serve as nanozyme probes. Fe₃O₄ modified with an aptamer (cDNA) was employed as the capture substrate. The assembly and dissociation of Au@Pt NPs with Fe₃O₄ are driven by the competitive binding of the target and H₁ to the cDNA. The degree of aggregation of Au@Pt NPs on Fe₃O₄ directly enhances the catalytic oxidation of oxTMB. Moreover, the synergistic interaction between the local surface plasmon resonance effect of nanozyme probes and capture substrates with external magnetic enrichment achieves exponential amplification of oxTMB signals. Based on this strategy, the dual-functional sensor demonstrates robust detection performance, with a limit of detection (LOD) as low as 0.27 fM and a detection time of merely 12 min. Moreover, this nanozyme-SERS dual-functional sensor exhibits strong specificity and resistance to interference, with detection results demonstrating high concordance with qRT-PCR. This research provides a robust method for detecting serum miRNAs in liver fibrosis and paves new pathways for the clinical translation of the SERS-nanozyme combined technology.

1. Introduction

Liver fibrosis is an abnormal deposition process of the extracellular matrix (ECM) induced by multiple chronic liver injuries, Persistent liver damage (such as viral infections and alcoholic liver disease) can lead to significant fibrosis, ultimately progressing to cirrhosis and hepatocellular carcinoma. Approximately two million people worldwide die from chronic liver disease each year, constituting a significant public health burden. Liver fibrosis is regarded as a reparative response in the progression of chronic liver disease, and thus, its screening holds significant value in the early diagnosis of hepatic disorders. However, current commonly used clinical diagnostic methods, such as liver biopsy and serological markers, exhibit limitations including poor specificity, low sensitivity, high invasiveness, and risks of hemorrhage and infection. Imaging techniques, meanwhile, are constrained by drawbacks such as high cost and inadequate staging capabilities. The existing integrated assessment system combining serological, imaging, and pathological approaches also struggles to meet the demands of large-scale screening programs. Consequently, it is imperative to explore potential biomarkers closely associated with liver fibrosis, thereby providing novel avenues for early diagnosis.

In recent years, microRNAs (miRNAs) have been regarded as an ideal “molecular fingerprint” for the diagnosis and monitoring of liver disease due to their tissue-specific expression patterns and high stability in bodily fluids. Among these, miR-34a has garnered significant attention due to its markedly abnormal expression during hepatic fibrosis. Existing research indicates that miR-34a may promote hepatic stellate cell (HSC) activation by regulating the TGF-β1/Smad signaling pathway or drive the fibrotic process by mediating the Sirt1/p53 pathway. , In a clinical study involving 59 subjects, serum miR-34a levels above 0.548 and 0.611 served as diagnostic thresholds for mild and significant liver fibrosis, respectively. The AUC value for mild fibrosis was 0.8274 (95% CI: 0.681–0.973; sensitivity 71.4%, specificity 83.3%); the AUC for significant fibrosis was 0.8904 (95% CI: 0.789–0.992; sensitivity 77.8%, specificity 83.3%), showing statistically significant differences compared to the control group and demonstrating good diagnostic performance. Therefore, serum miR-34a expression levels hold potential as a noninvasive biomarker for the early diagnosis of liver fibrosis and for monitoring disease progression. Current widely employed miRNA detection methods, such as real-time quantitative polymerase chain reaction (qRT-PCR), Northern blot, microarray, and next-generation sequencing (NGS), all suffer from limitations including restricted detection efficiency, complex operational procedures, and inadequate clinical scalability. Given the extremely low abundance of miRNAs in serum, the development of novel techniques combining high sensitivity with rapid detection capabilities has become an urgent requirement in this field.

Nanozymes, as artificial enzyme mimics, integrate enzyme-like catalytic activity with the unique properties of nanomaterials, thereby offering significant potential for biosensing applications. Compared to natural enzymes, nanozymes offer advantages such as low cost, high stability, and ease of large-scale production, overcoming inherent limitations of natural enzymes such as susceptibility to inactivation under extreme conditions and difficulty in purification. However, existing nanozyme detection methods predominantly rely on colorimetric substrate-based label-free systems. Nevertheless, their limited specificity and interference from the sample’s inherent coloration restrict their widespread application. In recent years, surface-enhanced Raman scattering (SERS) technology has achieved ultrasensitive detection of trace amounts or even single molecules by significantly amplifying Raman signals through the localized surface plasmon resonance (LSPR) effect on nanostructured surfaces. This unique molecular fingerprint recognition capability holds vast application prospects in cutting-edge fields, such as chemistry, materials science, and biomedicine. The integration of nanozymes with SERS, coupled with aptamer modification, not only overcomes the limitations of insufficient specificity and color interference in conventional detection methods but also enables multilevel amplification of target signals. This approach maintains a high efficiency and sensitivity even in complex sample environments. Au@Pt nanospheres (Au@Pt NPs), as highly regarded bimetallic nanoenzymes in recent years, play a vital role in detection and analysis. They enable “explosive” cascade amplification of target signals, providing critical technological support for the highly sensitive detection of trace biomolecules. This unique material structure combines the distinctive efficacy of Au nanomaterials in LSPR for SERS enhancement with the highly efficient catalytic activity of Pt particles, enabling dual signal amplification for label-free detection as a nanozyme probe. Furthermore, the incorporation of magnetic nanoparticles as capture substrates enables efficient miRNA capture within complex samples and subsequent signal amplification. This is achieved through their unique magnetic response enrichment capability and superior plasmonic enhancement effect. Based on this, the construction of an aptamer-driven nanozyme-SERS dual-functional sensor holds promise for achieving ultrasensitive detection of liver-fibrosis-associated miR-34a.

Here, we have constructed a nanozyme-SERS dual-functional sensor driven by aptamer competitive recognition combined with a magnetic enrichment separation strategy to achieve secondary signal amplification, enabling ultrasensitive detection of miR-34a in serum. As depicted in Figure A, Au@Pt NPs were first synthesized by encapsulating a platinum shell around a gold nanoparticle core, thereby achieving both outstanding catalytic performance and SERS effects. Subsequently, a complementary strand (H₁) of nucleic acid aptamers modified with gold–sulfur bonds was immobilized onto their surface to serve as a nanozyme probe. On the other hand, Fe₃O₄ magnetic nanoparticles modified with streptavidin were conjugated with biotin-modified aptamers (cDNA) to form Fe₃O₄@cDNA. The detection principle is illustrated in Figure B. Following the hybridization reaction, the formation of a double-stranded structure through hybridization between H₁ and cDNA anchors the Au@Pt NPs@H₁ onto the Fe₃O₄@cDNA surface. This constructs an Fe₃O₄@cDNA-Au@Pt NPs@H₁ composite system with controllable assembly characteristics, namely, a nanozyme-SERS dual-functional sensor. When the target miRNA is present in the serum sample, its high affinity for cDNA causes the cDNA-H₁ double strand to unwind. Competitive hybridization then forms a cDNA-target double strand, triggering a strand displacement reaction. This “unlocks” H₁ from the cDNA, thereby driving the release of Au@Pt NPs from the Fe₃O₄ carrier. This binding–dissociation mechanism driven by competitive recognition of aptamers ensures a high degree of detection specificity and effectively minimizes interference from nonspecific molecules in serum. The “hot spots” generated by the nanoscale spacing when Au@Pt NPs combine with Fe₃O₄, coupled with the LSPR effect inherent to the nanoparticles themselves, can significantly amplify the SERS effect. Specifically (Figure C), serum from liver fibrosis mice containing target miRNA was reacted with the sensor. Following signal enrichment via an applied magnetic field, H₂O₂ and TMB were added as reaction substrates. Unreleased Au@Pt NPs catalyze the conversion of TMB to oxTMB, which exhibits a strong Raman signal. In contrast, the pure capture substrate shows a weaker Raman signal (Figure D). The intensity of this SERS response enables quantitative analysis of serum microRNAs via linear equation fitting with results validated against qRT-PCR data. This approach, driven by aptamers, responsive to the TMB, and amplified through an enrichment and separation strategy, not only significantly enhances detection sensitivity and specificity but also offers a novel technical pathway and promising applications for the early noninvasive diagnosis of liver fibrosis.

1.

(A) Preparation process of SERS nanozyme probes and capture substrates. (B) Schematic illustration of the recognition competition strategy triggered by target miRNAs. (C) Detection of liver-fibrosis-related miRNAs by nanozyme SERS sensors using an enrichment separation strategy. (D) Principle of signal response in nanozyme SERS sensors.

2. Experimental Section

2.1. Materials and Reagents

Hydrochloric aqua regia (HAuCl₄), polyvinylpyrrolidone (PVP), trisodium citrate (C₆H₅Na₃O₇·2H₂O), hydrochloroplatinic acid (H₂PtCl₆), anhydrous ethanol, phosphate-buffered saline (PBS), and bovine serum albumin (BSA) were all procured from Sinopharm Chemical Reagent Co., Ltd. Ascorbic acid (AA), tris­(2-carboxyethyl)­phosphine hydrochloride (TCEP), and 3,3′,5,5′-tetramethylbenzidine (TMB) were purchased from Aladdin Biochemical Technology Co., Ltd. All reagents used in the experiment were of analytical grade (AR). Sstreptavidin-modified Fe₃O₄ nanoparticles (400 nm) and nucleotides mentioned in Table , including miR-34a, single-base mismatch sequence (MT1), triple-base mismatch sequence (MT3), random sequence, nucleic acid aptamer sequence (cDNA), and nucleic acid aptamer complementary sequence (H₁), were all procured from Sangon Biotech (Shanghai) Co., Ltd. (China). The quantitative real-time polymerase chain reaction (qRT-PCR) kit was procured from Jiedan Biotechnology Co., Ltd. Deionized water with a resistivity of 18.2 MΩ·cm was employed throughout this study. Male C57BL/6 mice used in the experiments were provided by the Department of Clinical Medicine at Yangzhou University.

1. Nucleotide Sequences Used in the Experiment.

name	sequences
miR-34a	UGGCAGUGUCUUAGCUGGUUGU
cDNA	biotin-ACAACCAGCTAAGACACTGCCA
H1	SH-TGGCAGTGTCTT
MT1	TGGCAGTGTGTTAGCTGGTTGT
MT3	TGGCACTCTCTTAGGTGGTTGT
random sequence	GGUUUACUUUAGAGACUAGAUC

2.2. Instrument

To characterize the nanomaterials, we employed a field emission scanning electron microscope (Hitachi S-4800 II, 10.0 kV) to observe the morphology of gold hexagonal plates and acquire scanning electron microscopy (SEM) images; transmission electron microscopy (TEM) images were collected using a Philips Tecnai 12 instrument at 120 kV; high-resolution transmission electron microscopy (HRTEM) images were obtained using a field emission transmission electron microscope (FEI Tecnai G2 F30 S-TWIN, 200 kV), with diffraction patterns acquired via selected area electron diffraction (SAED) techniques. Ultraviolet–visible spectra of the samples were recorded by using an ultraviolet–visible spectrophotometer (Agilent Cary 60). Raman spectroscopy measurements were performed using a DXRxi micro-Raman spectrometer (Thermo Fisher, USA).

2.3. Preparation of Au@Pt Nanoparticles

We have made improvements upon the work of Yang et al. First, 0.5 mL of HAuCl₄ (1 wt %) was added to 50 mL of ultrapure water. The mixture was brought to a rolling boil, and then 0.8 mL of sodium citrate solution (1 wt %) was promptly added. The boiling conditions were maintained while stirring continuously for 10 min. Subsequently, 1.0 mL of AA (0.1 M) and 1.25 mL of H₂PtCl₆ (1 wt %) were added in sequence, and the mixture continued to boil for half an hour. Heating ceased when the solution color transitions from wine red to dark brown. The precipitate was centrifuged at 10,000 rpm and then washed three times with ultrapure water. Finally, the Au@Pt NPs were redispersed in ultrapure water and sonicated for 5 min. Sonication was performed using an ultrasonic bath with a fixed power setting to ensure uniform dispersion while avoiding structural damage or aggregation of the nanoparticles.

2.4. Preparation of Fe₃O₄@cDNA-H₁@Au@Pt NPs

To obtain the nanozyme probe, we added H₁ (0.1 mM) activated with 1 mM TCEP to the Au@Pt NP solution and incubated it at 37 °C for 2 h. To block nonspecific binding sites on the particle surface, we added 30 μL of 1 wt % BSA solution to the mixture and continued the reaction for 180 min. Finally, impurities were removed by centrifugation at 9000 rpm for 25 min, yielding the final Au@Pt NPs@H₁ complex.

To obtain functionalized Fe₃O₄ (Fe₃O₄@cDNA), streptavidin-modified Fe₃O₄ was first washed three times with ultrapure water and sonicated for 10 min to ensure uniform dispersion. A 100 μM cDNA solution, preincubated at 95 °C for 300 s, was then added and shaken continuously for 30 min. The resulting product (Fe₃O₄@cDNA) was enriched and separated using a circular magnet (300 G) for 5 min (Table S2) and then washed multiple times with ultrapure water to remove excess cDNA. The final Au@Pt NPs@H₁ was mixed with Fe₃O₄@cDNA and incubated at 37 °C for 1.5 h. Excess Au@Pt NPs@H₁ was removed by washing three times with PBS, ultimately yielding the Fe₃O₄@cDNA-H₁@Au@Pt NP complex.

2.5. Establishment of a Mouse Model of Liver Fibrosis and Serological Testing

To establish a mouse model of liver fibrosis, 32 male C57BL/6 mice aged 6–8 weeks were randomly divided into four groups of eight mice each. The experimental group received subcutaneous injections of a 20% CCl₄ oil solution (CCl₄/olive oil = 1:4) at a dose of 5 mL/kg of body weight administered every 3 days for 2, 4, or 6 consecutive weeks. The control group received only an equal volume of olive oil. Housing conditions were maintained at 20–26 °C, 40–70% relative humidity, with a 12 h light-dark cycle. Throughout the experiment, mice were monitored for body weight, activity levels, mental state, and fur changes. (All mice were housed individually in single-cage systems with unique identification numbers. The housing environment was strictly controlled throughout the study to ensure technical replicates within and between groups.) Serum samples were collected prior to injection and at 2, 4, and 6 weeks postinjection for SERS detection. After centrifugation, these were stored at −80 °C for future use. All mice were euthanized after the final blood collection, with liver tissue harvested for immunohistochemical, hematoxylin and eosin (HE), and Masson’s trichrome staining. All animal procedures adhered to Yangzhou University’s Guidelines for the Care and Use of Laboratory Animals and were approved by the Yangzhou University Animal Ethics Committee.

2.6. SERS Signal Acquisition

A 1 mL portion of the SERS sensor solution was transferred into a centrifuge tube, and the sample to be tested was added and mixed thoroughly. The mixture was incubated at 37 °C for 1 h to form a stable miR-34a/cDNA double-stranded complex. Subsequently, the complex was separated and enriched using a micromagnet, washing several times with PBS to remove nonspecific components. Freshly prepared H₂O₂ and TMB solutions were added to the complex. It was then incubated at a constant temperature in pH 4.5 PBS buffer for 12 min. Data were collected using a calibrated Raman spectrometer. Detection parameters are as follows: Raman shift scan range 200–2000 cm^–1, laser power 5 mW, and single acquisition time 5 s.

2.7. MiRNA Differential Analysis and Screening of Target Genes among Differentially Expressed Genes

Searching the GEO (http://www.ncbi.nlm.nih.gov/geo) database for “Liver Fibrosis (All Fields) AND miRNA (All Fields)” yielded the noncoding RNA profile GSE77271, comprising miRNA information from three groups of CCl₄-induced liver fibrosis mouse models and three groups of corn-oil-treated normal mouse liver samples. The data files were processed via the GEO2R online platform and subjected to differential expression analysis. Differentially expressed miRNAs were identified using the screening criteria |log₂ fold change| ≥ 1 and P < 0.05. Volcano plots of differentially expressed miRNAs in the data set were generated using the Origin software. A search of the GEO database for “Liver Fibrosis (All Fields) AND gene (All Fields)” yielded the gene chip data set GSE222576 (15 groups of liver fibrosis mice, 3 groups of normal controls). Using the GEO2R online data analysis tool, differentially expressed genes were identified with the screening criteria of |log₂ fold change| ≥ 1 and P < 0.05. Subsequent analyses were conducted on the overlapping genes between the target genes and the differentially expressed genes.

2.8. MiRNA Target Gene Prediction and Enrichment Analysis

Using the miRWalk online database (http://mirwalk.umm.uni-heidelberg.de/), with species selected as mouse, we input the predicted target genes of mmu-miR-34a-5p and exported those intersecting with the Targetscan database. GO and KEGG pathway enrichment analyses were performed on differentially expressed target genes via the MicroBioinformatics platform (https://www.bioinformatics.com.cn/). The GO enrichment analysis annotates and classifies genes according to the biological process (BP), cellular component (CC), and molecular function (MF). The top 10 most significantly enriched GO functional categories and 10 pathways from each domain were visualized.

3. Results and Discussion

3.1. Characterization of Fe₃O₄ and Au@Pt NPs

As shown in Figure , we employed transmission electron microscopy for detailed characterization of Fe₃O₄ and Au@Pt NPs. In the TEM and HRTEM images presented in Figure A,B and C, the Fe₃O₄ particles exhibit a regular spherical morphology with a radius of approximately 220 nm. They are uniformly dispersed with a stripe spacing of 0.263 nm corresponding to the {311} crystal plane of Fe₃O₄. TEM and HRTEM images shown in Figure D,E reveal that the Au@Pt NPs exhibit a core–shell structure with platinum encapsulating gold, featuring an average particle size of approximately 30 nm. Its 0.227 nm lattice spacing corresponds to the {111} crystal plane of Pt. The mapping image in Figure F clearly demonstrates elemental mapping for both gold and platinum.

2.

Schematic characterization of Fe₃O₄ and Au@Pt NPs. (A, B) TEM image and (C) HRTEM image of Fe₃O₄. (D) TEM image and (E) HRTEM image of Au@Pt NPs. (F) Mapping image of Au@Pt NPs.

3.2. Characterization of Fe₃O₄@cDNA-H₁@Au@Pt NPs

In the SEM images shown in Figure A,B, the composite structure of Fe₃O₄@cDNA-H₁@Au@Pt NPs with a radius of approximately 260 nm is clearly visible, exhibiting a relatively uniform distribution. Au@Pt NPs are randomly distributed on the surface of magnetite, forming numerous hotspots. Figure C displays the particle size distribution of the composite with an average particle size of 400 ± 50 nm, demonstrating the overall dimensional uniformity of the composite material. The corresponding thickness distribution (Figure D) shows an average particle thickness of 50 ± 5 nm, with the thickness distribution similarly exhibiting a high concentration. Figure S1 displays the UV–visible absorption spectra, demonstrating the successful modification of H₁ and cDNA onto the nanomaterial and their subsequent complexation. To further investigate the elemental distribution within the Fe₃O₄@cDNA-H₁@Au@Pt NP composite, Figure E presents EDX analysis revealing that its composition primarily comprises Fe, Au, and Pt. Moreover, Fe₃O₄@cDNA-H₁@Au@Pt NPs exhibit exceptional SERS enhancement, as demonstrated in Figure F. The characteristic peak of DTNB labeled Fe₃O₄@cDNA-H₁@Au@Pt NPs is markedly higher than that of pure DTNB. Calculated using the formula EF = (I _SERS/C _SERS)/(I _Raman/C _Raman), the EF value is determined to be 1.08 × 10⁸ (I _SERS and C _SERS denote the SERS intensity and analyte concentration on the assembled complex, respectively, while I _Raman and C _Raman represent the unamplified SERS intensity and analyte concentration, respectively).

3.

Schematic characterization of Fe₃O₄@cDNA-H₁@Au@Pt NPs. (A) SEM image of a single Fe₃O₄@cDNA-H₁@Au@Pt NP. (B) SEM image of multiple Fe₃O₄@cDNA-H₁@Au@Pt NPs. (C) Particle size distribution of Fe₃O₄@cDNA-H₁@Au@Pt NPs. (D) Shell thickness distribution of Fe₃O₄@cDNA-H₁@Au@Pt NPs. (E) EDX spectrum of Fe₃O₄@cDNA-H₁@Au@Pt NPs. (F) Comparison of SERS spectra between DTNB-labeled Fe₃O₄@cDNA-H₁@Au@Pt NPs and pure DTNB.

3.3. Correlation between Targets and Diseases

Building upon prior research, to further investigate the correlation between miR-34a and hepatic fibrosis, we retrieved the miRNA data set GSE77271 from the GEO database. This data set comprises liver samples from CCl₄-induced fibrotic mice and corn-oil-treated normal mice. Following GEO2R analysis and screening with |log₂ fold change| ≥ 1 and P < 0.05, 111 differentially expressed miRNAs were identified (60 downregulated, 51 upregulated), among which mmu-miR-34a-5p exhibited significant differential expression (Figure A). By predicting 185 mouse target genes for this miRNA via the miRWalk database and then intersecting these with 23,745 differentially expressed genes screened from the human liver fibrosis gene data set GSE222576 (15 liver fibrosis samples, 3 normal controls) in the GEO database, 181 common genes were obtained (103 downregulated, 78 upregulated) (Figure B). GO enrichment analysis of these 181 genes revealed their enrichment in biological processes central to hepatic fibrosis, including extracellular matrix deposition and fibroblast activation/proliferation. Cellular components focused on the extracellular space and collagen trimers, while molecular functions encompassed protein binding and cytokine activity. The red highlighted box indicates pathways that typically induce liver fibrosis, where small GTPase-mediated signal transduction regulates the activation and migration of hepatic stellate cells, with Rac1 directly participating in the migration process of these cells (Figure C). KEGG pathway analysis revealed that these genes were enriched in classical hepatic fibrosis regulatory pathways such as TGF-β, PI3K-Akt, MAPK, and NF-κB. The Ras signaling pathway, as the upstream core signaling pathway of the MAPK pathway, plays a major role in regulating liver fibrosis. It drives the activation of hepatic stellate cells, with Raf1 promoting HSC activation and MEK1 regulating collagen synthesis, among other pathways (Figure D). The aforementioned findings indicate that miR-34a exerts a significant regulatory role in the development and progression of liver fibrosis by targeting genes involved in the key pathological processes and pathways.

4.

Correlation analysis of target miRNAs with disease. (A) Volcano plot of differentially expressed miRNAs in liver samples from three groups of CCl₄-induced liver fibrosis mice and three groups of corn-oil-treated normal mice. (B) Intersection genes between liver fibrosis differentially expressed genes and mmu-miR-34a-5p target genes. (C) GO enrichment analysis bubble chart. (D) KEGG enrichment analysis Sankey bubble chart.

3.4. Material Performance Testing

To evaluate the intrinsic properties and catalytic efficacy of the nanomaterials, we characterized their magnetic properties and peroxidase-like activity. As shown in Figure A, when an external magnetic field is applied to the composite material, it rapidly aggregates from a dispersed state to the magnetic field region. This facilitates better separation from unbound Au@Pt NPs, simplifying the product purification and separation process. Consequently, precise measurement and secondary signal amplification are achieved. After the addition of TMB, buffer solution, and H₂O₂ to the composite material, the solution color changed from colorless to blue upon incubation (Figure B), and an ultraviolet absorption peak appeared at 657 nm. However, when only TMB and H₂O₂ were added, no peak was observed. This indicates that TMB underwent an oxidation reaction, confirming the POD enzyme activity of the composite material. To evaluate the catalytic performance of the nanozyme, we maintained a constant H₂O₂ concentration while adjusting the TMB concentration and keeping the solution pH unchanged. Five replicate wells were set up for each group to analyze the enzymatic reaction kinetics (Figures C,D). The study employed the Michaelis–Menten enzyme kinetic model in nonlinear fitting, yielding a mean Vmax of 0.68269 μM s^–1 with a standard error of 0.03923 and a mean Km of 0.28019 with a standard error of 0.04472 mM. These results indicate that the complex exhibits high affinity for the substrate and outstanding catalytic efficiency, making it more competitive compared to currently studied nanocatalysts.

5.

(A) Fe₃O₄@cDNA-H₁@Au@Pt NPs under an applied external magnetic field and without an external magnetic field. (B) UV–visible absorption spectra of Fe₃O₄ alone, Au@Pt NPs alone, and Fe₃O₄@cDNA-H₁@Au@Pt NPs in the presence of TMB and H₂O₂. The inset shows a photograph. (C) Michaelis–Menten curve of the nanozyme and (D) Lineweaver–Burk plot.

3.5. Conditional Optimization

This study employs a sensor-based, label-free detection method utilizing POD-like enzyme activity to catalyze the oxidation of TMB, thereby generating oxTMB with a strong SERS signal. The selected Raman characteristic peak at 1602 cm^–1 corresponds to the CC stretching vibration arising from the conjugation of the two benzene rings in oxTMB. First, we optimized the volume of the nanozyme probe. As shown in Figure A, when 3 μL of the nanozyme probe was added, the SERS signal peaked at 1602 cm^–1. This is attributed to the insufficient utilization of substrate adsorption sites at low volumes, whereas increasing volumes led to multilayer adsorption and disruption of the substrate’s LSPR effect. Moreover, the catalytic efficiency of nanozymes is constrained by multiple factors. Therefore, to enhance the detection performance of the nanozyme SERS sensing system, this study also optimized factors such as time, temperature, and substrate concentration. As shown in Figure B–F, the catalytic activity of the nanozyme SERS sensor reached its optimum under incubation conditions of 35 °C, pH 4.5, 1 mM TMB concentration, and 0.6 mM H₂O₂ concentration for 12 min. The substrate concentration selected in Figure F was determined based on Michaelis–Menten analysis, specifically operating at concentrations equivalent to or slightly higher than the extracted Km value. Under these conditions, the reaction rate remains sensitive to changes in catalytic activity while avoiding substrate saturation effects that could obscure performance differences.

6.

(A) Volume of nanozyme labels. (B) Incubation temperature. (C) Incubation time. (D) Solution pH. (E) H₂O₂ concentration. (F) TMB concentration.

3.6. Sensor Performance Validation

To evaluate the efficacy of this nanozyme SERS sensor in SERS detection, we monitored the SERS intensity of oxTMB over 100 min (Figures A,B). As shown in the figure, the characteristic peaks at 1183, 1330, and 1602 cm^–1 are attributed to oxTMB, while other minor peaks may be attributed to interfering molecules, such as nucleotides and lipids. This is consistent with previous reports in the literature. During detection, oxTMB-related peaks gradually intensified with the reaction time before reaching a stable plateau, indicating sustained catalytic activity and signal stability. The overall spectral profile remained unchanged, ensuring the reliability of the nanozyme SERS sensor for quantitative biomarker detection. Based on this, we also tested changes in signal intensity over a 24 h period (Figure S2). Furthermore, as shown in Figure (C), we measured the SERS spectra of oxidized TMB for five different batches of sensors. The results indicate that these sensors exhibit no significant differences and demonstrate excellent reproducibility. To investigate the sensor’s stability, we measured sensors from the same batch over varying storage periods. Results indicate that the Raman signal stabilizes after 10 days (Figure D), with this consistent output ensuring the sensor’s stability during storage. When verifying material homogeneity, we selected 20 random sampling points for SERS spectral analysis (Figures E,F). The results demonstrated relatively stable Raman intensities across all points, with a relative standard deviation of 3.98%, indicating the sensor’s uniform stability. During sample analysis, exposure to multiple biomarkers and serum interferences occurs. To evaluate specificity and interference resistance, this study demonstrated in Figure G–I the effects of interfering sequences (including single-base mismatches (MT1), triple-base mismatches (MT3), and random sequences), nontarget molecules (glucose, lipoproteins, cholesterol), and mismatched miRNAs on detection results as serum component concentrations increased. Additionally, Figure S3 demonstrates the comparative calibration between the buffer and serum, ruling out the interference of matrix effects on the detection. The sensor demonstrated robust specificity and interference resistance, with minimal signal disruption from contaminants. This renders it highly suitable for quantitative target detection within complex serum samples.

7.

Performance evaluation of nanozyme SERS sensors: (A) SERS spectra at different time intervals and (B) two-dimensional intensity heatmap. (C) SERS spectra and local method plots of nanozyme SERS sensors from different batches. (D) SERS spectra of nanozyme SERS sensors stored for different days and a line chart of signal intensity at 1602 cm^–1. (E) Scatter plot of 20 randomly selected positions and (F) 3D surface heatmap with projections. (G) Bar chart demonstrating nanozyme SERS sensor selectivity for miR-34a. (H) Effect of small-molecule interferences in simulated serum on miR-34a detection. (I) The effect of mismatched miRNA and increased serum component concentrations on miR-34a detection.

3.7. Quantitative miRNA Detection

Under optimized experimental conditions, this study employed the sensor to assess its sensitivity toward detecting target microRNA (miR-34a) at varying concentrations. Figure A shows the SERS spectrum during detection, showing that the intensity of the SERS spectrum gradually increases with rising miR-34a concentrations. To provide a more intuitive observation of the intensity changes at the characteristic peak at 1602 cm^–1, we plotted a heatmap at this peak (Figure B), where the magnitude of the signal intensity directly reflects variations in the target miRNA concentration. In Figure C, a fitted linear curve was plotted using the characteristic peak intensity at 1602 cm^–1 against the logarithm of the miR-34a concentration. The concentration range established in this study was 1 fM–1 nM. The miR-34a concentration in serum from mice with liver fibrosis ranged from nanomolar to femtomolar levels and significantly increased with fibrosis progression, indicating that this concentration window holds physiological significance. The resulting linear regression equation was y = −3213.24x – 24595.94, with R ²=0.9841. The detection results confirmed that the nanozyme-SERS dual-functional sensor, incorporating aptamer modification and enrichment separation strategies, can quantitatively detect miR-34a concentrations in complex serum environments. Based on the formula LOD = 3SD/k (SD represents the standard deviation based on the blank sample reaction, n = 6; k denotes the slope of the standard calibration curve), a detection limit as low as 0.27 fM was achieved. Compared with other detection methods, this approach offers the advantages of a low detection limit and high specificity (Table ).

8.

SERS spectra of miR-34a at different concentrations in serum (A) and a heatmap at 1602 cm^–1 (B). Linear equation fitted between the characteristic peak intensity at 1602 cm^–1 and the logarithm of the miR-34a concentration (C).

2. Comparison of the Methodology Employed in This Study with Other MiRNA Detection Techniques.

method	target	LOD	ref
fluorescence	miR-155	2.18 fM
photoelectrochemical	miR-21	9 fM
electrochemical	miR-21	5.4 fM
SERS	miR-141	18.4 fM
SERS	miR-34a	0.27 fM	this work

3.8. Characterization of a Mouse Model of Liver Fibrosis

Through histological and biochemical analyses, we successfully established a mouse model of hepatic fibrosis. Figure A shows macroscopic histological observations of liver tissue. In the normal control group, the liver surface was smooth and well-defined, with the volume within the physiological normal range. In contrast, mice in the liver fibrosis induction group exhibited progressive hepatomegaly, with surface texture coarsening starting from week 2. The pathological phenotype continued to progress throughout the modeling process, demonstrating significant volume expansion and nodular protrusions by week 6, with clearly defined macroscopic pathological alterations. The quantitative analysis of liver weight/body weight ratio (LW/BW) in Figure E demonstrated a dynamic trend fully synchronized with the macroscopic morphological expansion observed in Figure A. This confirms the core phenotypic feature of proliferative hepatomegaly during the progression of liver fibrosis. Moreover, in Figure B–D, following induction of hepatic fibrosis, HE staining in mice revealed structural disruption and inflammatory infiltration as fibrosis progressed. α-SMA immunostaining demonstrated a marked increase in α-SMA-positive cells in perisinusoidal and periportal regions, indicating the progressive activation of hepatic stellate cells. Masson’s trichrome staining confirmed progressive collagen deposition over time, with bridging fibrosis evident by week 6. In contrast, the control group exhibited a well-preserved hepatic lobule architecture with minimal collagen deposition (Table S1). From the biochemical indicators (Figure F–H), ALT and AST levels showed a marked elevation commencing from the second week, reflecting hepatocyte injury in the mice. Subsequently, TBIL levels rose, indicating potential abnormalities, such as biliary obstruction or impaired excretory function. The above results demonstrate that the mouse model of liver fibrosis has been successfully established, enabling the collection of serum for SERS detection.

9.

Changes in liver morphology and biochemical indicators over time in a mouse model of liver fibrosis. (A) Macroscopic appearance of mouse livers at 0, 2, 4, and 6 weeks. (B) Hematoxylin and eosin (H&E) staining results, (C) immunohistochemical staining results, and (D) Alcian blue staining results of mouse liver tissue sections. (E–H) Liver-to-body weight ratio, ALT, AST, and TBL levels, respectively, with W denoting weeks.

3.9. Analysis of Serum Samples from Mice with Liver Fibrosis

To validate the sensor’s performance in complex serum samples, we collected serum samples from mice at different stages of liver fibrosis (0, 2, 4, and 6 weeks; n = 8 per group). This study treats serum samples from individual mice as independent biological replicates. Should samples become invalid due to hemolysis, contamination, or other issues, replacement samples will be obtained from mice under identical conditions to ensure that each group maintains a minimum of eight valid biological replicates. Each sample underwent three technical replicates of SERS detection with the average value recorded. This approach eliminated technical operational errors while accounting for individual biological variations, ensuring the reliability of the detection results. Before performing the assays, we first carried out a preprocessing step to remove interfering molecules in the serum samples (such as proteins and lipids). The sensors were then thoroughly rinsed with PBS to minimize the impact of nontarget species on sensor performance. As shown in Figure A, spectral intensities in liver-fibrotic mice were markedly lower than those in control mice, progressively decreasing with worsening fibrosis severity. Student’s t test evaluated intensity variations at the characteristic peak of 1602 cm^–1 (Figure B), revealing significant differences in signal intensity between groups. Furthermore, this experiment employed qRT-PCR results as the gold standard for detection. These were compared with outcomes obtained by substituting SERS signal intensities into a fitted linear equation, thereby ensuring the accuracy of the detection results (Table ). The results demonstrated consistency between the two methods employed, confirming the reliability and practicality of this sensor. This lays a solid foundation for its future application in clinical testing.

10.

Detection of miR-34a in a mouse model of liver fibrosis. (A) Average SERS spectra from serum samples of mice with different degrees of liver fibrosis. (B) Quantitative differentiation of SERS intensity at 1602 cm^–1 in liver tissue sections from mice with varying fibrosis stages via t test.

3. Comparison of miR-34a Expression Levels in Serum from Mice with Liver Fibrosis Detected by SERS and qRT-PCR.

week	SERS (mol/L)	qRT-PCR (mol/L)	relative error (%)
0	(1.87 ± 0.21) × 10–12	(1.82 ± 0.16) × 10–12	2.75
2	(2.95 ± 0.67) × 10–11	(2.75 ± 0.06) × 10–11	7.27
4	(3.58 ± 0.47) × 10–10	(3.63 ± 0.23) × 10–10	–1.38
6	(9.38 ± 0.59) × 10–10	(9.15 ± 0.25) × 10–10	2.51

4. Conclusions and Discussion

In summary, this study successfully constructed a nanocatalyst-SERS dual-functional sensor driven by aptamer recognition competition. Combined with an enrichment and separation strategy, it achieved highly specific and ultrasensitive detection of miR-34a in serum. In this system, Au@Pt NPs serve as nanozyme probes. Through complementary pairing between aptamers and targets, recognition competition is triggered, causing the probes to detach from the Fe₃O₄ surface. This results in a decrease in the sensor’s peroxidase catalytic activity and SERS signal. Following enrichment and separation, the signal undergoes secondary amplification, demonstrating outstanding detection performance with a limit of detection as low as 0.27 fM. Concurrently, building upon prior research, this study conducted bioinformatics analysis of miR-34a associated with hepatic fibrosis to further validate its reliability as a detection target. Following multifactor optimization, we performed correlation analysis between this sensor and the gold standard qRT-PCR results from this study, confirming that the detection outcomes possess high reliability and accuracy. In summary, this aptamer-driven nanozyme-SERS dual-functional sensor demonstrates broad application prospects for the early screening of liver fibrosis. In the future, this system may be integrated with microfluidic chips or portable POCT platforms to further enable high-throughput detection and simultaneous quantitative analysis of multiple miRNAs, thereby providing a robust foundation for clinical translation.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c12530.

• Ultraviolet–visible absorption spectra of Au@Pt NPs, Fe₃O₄, and H₁/cDNA-modified composite nanomaterials (Figure S1); SERS spectra measured within 24 h and a line graph of signal intensity at 1602 cm^–1(Figure S2); SERS spectra of miR-34a in serum and PBS environments and bar chart of miR-34a signal intensity at 1602 cm^–1 in PBS and serum environments (Figure S3); collagen area fraction in Masson-stained sections of mice at different stages of liver fibrosis (Table S1); and recovery efficiency of magnetic enrichment separation (Table S2) (PDF)

#.

Y.B. and B.Z. contributed equally to this work.

This work was supported by the General Program of Nantong Municipal Health Commission Research Project (MS2024111, MS2024112), Nantong University Special Research Fund for Clinical Medicine­(2024LZ028), and 2024 Haimen District-Level Social Welfare and People’s Livelihood Technology Plan Projects.

The authors declare no competing financial interest.