Pump-free SERS microfluidic chip based on an identification-competition strategy for ultrasensitive and efficient simultaneous detection of liver cancer-related microRNAs

Abstract

In this study, we present a pump-free SERS microfluidic chip capable of detecting liver cancer-related miR-21 and miR-155 concurrently with ultra-sensitivity and high efficiency. We employed a Fe₃O₄@cDNA-AuNPs@Raman reporter@H composite structure and a recognition competition strategy. When the target miRNAs (miR-21 and miR-155) are present in the test liquid, they specifically compete with the nucleic acid complementary strand(H) of Fe₃O₄@cDNA-AuNPs@Raman reporter@H, causing AuNPs to competitively detach from the surface of Fe₃O₄, resulting in a decrease in the SERS signal. Consequently, this pump-free SERS microfluidic chip enables the detection of the target miRNAs more rapidly and accurately in complex environments. This method offers an approach for the simultaneous and efficient detection of miRNAs and holds promising applications in the early diagnosis of liver cancer.

1. Introduction

Cancer is a major disease that endangers human health and is associated with a high mortality rate. Among these, liver cancer is one of the most common malignant tumors [1,2], and it is the fourth leading cause of cancer-related deaths worldwide, and is characterized by a high mortality rate [3]. Its five-year survival rate is approximately 15% - 17% [4]. Liver cancer encompasses a variety of subtypes, including hepatocellular carcinoma and cholangiocarcinoma, among others [5]. As liver cancer often exhibits no obvious symptoms in its early stages, most patients are already in the advanced stage by the time they are diagnosed. Despite advancements in diagnostic and treatment methods, the mortality rate of liver cancer continues to rise rapidly [6,7]. Therefore, early diagnosis of liver cancer is crucial for providing patients with better treatment options and higher survival rates.

Currently, the primary screening methods for liver cancer include imaging tests, serological markers and tissue biopsy [8]. Imaging examinations, such as X-ray, CT and MRI, have limitations, as the accuracy of the results depends significantly on the operator's experience and the precision of the equipment, potentially leading to oversights and misdiagnoses [9]. As for serological markers, alpha-fetoprotein (AFP) is a widely utilized biomarker for liver cancer. However, in the early stages, AFP's sensitivity is low, rendering it less effective in diagnosing liver cancer [10]. Tissue biopsy, although more reliable, is an invasive procedure that may not always yield representative tissue samples and poses risks of bleeding and infection [8]. Consequently, there is an urgent need to develop a rapid, efficient, and easy diagnostic method for liver cancer.

In recent years, the use of microRNAs (miRNAs) as biomarkers for the early diagnosis of liver cancer has demonstrated great potential in clinical applications. MiRNAs are a class of small endogenous non-coding RNAs with a length of 22-25 nt, they are one of the gene regulatory molecules in multicellular organisms, have an important influence on many protein-coding genes, and participate in tumor cell proliferation, differentiation, and metabolism, playing an important role in tumor development [11,12]. Some studies have shown that miR-21 is associated with the liver cirrhosis and the staging and poor prognosis of liver cancer [13]. In a study conducted by Liu et al., researchers examined different liver cancer cells and found that miR-21 expression levels were significantly upregulated in cancer cells compared to normal cells [14]; Studies carried out by Wagenaar et al have confirmed that miR-21 plays an important role in the occurrence and development of liver cancer, and it can promote tumor growth and proliferation [15]. MiR-155 also plays an important regulatory role in tumor development. Liu et al. found that miR-155 expression was upregulated in hepatocellular carcinoma cells [16]. It has been shown that miR-155 promotes tumor progression by regulating the expression of proteins such as H3F3A [17]. Studies have reported that miR-155 can promote the activity of HCC cells by inhibiting the PTEN/PI3K-AKT pathway [18]. At present, the main detection methods for miRNAs are PCR, northern-blotting, electrochemical detection and microarray [19]. However, these methods have some limitations in practical application, such as complex operation, high cost, low sensitivity and so on [20,21]. Moreover, in order to improve the accuracy of detection and reduce the misdiagnosis rate, the simultaneous detection of multiple miRNAs is a promising method that can provide more comprehensive and accurate diagnostic information. Therefore, there is an urgent need to develop a rapid, efficient, ultra-sensitive, and simultaneous detection of method for multiple miRNAs for the early detection of cancer.

Surface-enhanced Raman scattering (SERS) is a new spectral enhancement technique for improving molecular Raman signals that can provide unique molecular fingerprint information. Its sensitivity can reach the level of a single molecule and can reflect subtle changes in tissue biochemistry. It has a wide range of applications in biochemistry, clinical diagnosis and other fields [22]. The local surface plasmon resonance (LSPR) effect on the surfaces of precious metal nanomaterials is the main source of SERS enhancement. The electromagnetic field region generated by these plasmon resonances is called “hot spot”, and its enhancement effect is closely related to the structure and element composition of the nanomaterials [23]. In recent years, magnetic nanoparticles coated with gold nanoparticles have shown great potential for SERS-based analysis methods [24–28]. Compared with other SERS detection methods, due to the strong enrichment ability of Fe₃O₄ and the many nanogaps formed by Au particles gathering on the magnetic core, more hot spots can be formed, thus enhancing the Raman signal even more [29]. Nanomaterials are used to separate and enrich target analytes using a magnetic field. It is non-invasive, highly sensitive, and stable for detection, and is a promising SERS substrate [29,30].

Owing to its complicated process and high cost, SERS still lacks the capability of instant detection [31]. As a “laboratory on a chip”, the microfluidic system has the advantages of high efficiency, fast speed, and portability [32]. It can integrate functions such as sample preparation, enrichment, and detection in a small chip, and has been widely used in chemical analysis, biological detection, chemical synthesis, and other fields [33–36]. To overcome the limitations of traditional microfluidic chips that rely on external pumps, researchers have used water-soluble and lubricated polyethylene glycol (PEG) to conduct hydrophilic treatment on microfluidic chips, thereby improving the hydrophilicity of microfluidic chip channels. Furthermore, the comb-structured channels were designed to enhance the capillary force of microchannels, allowing liquids to flow freely in the microchannel [37–40].

In this study, by using a composite structure of Fe₃O₄@cDNA-AuNPs@Raman reporter@H and an identification-competitive strategy, a pump-free SERS microfluidic chip was developed for the ultra-sensitive and efficient simultaneous detection of liver cancer-related miR-21 and miR-155. Two complementary chains (H₁ and H₂) and two Raman reporters (4-mercaptobenzoic acid (4-MBA) and 5,5'-dithiobis (2-nitrobenzoic acid) (DTNB)) were modified on the surfaces of Au nanoparticles (AuNPs) to prepare two types of SERS labels (AuNPs@DTNB@H₁ and AuNPs@4-MBA@H₂), respectively. cDNA-functionalized Fe₃O₄ nanoparticles (Fe₃O₄@cDNA 1 and Fe₃O₄@cDNA 2) were prepared by modifying two nucleic acid aptamers (cDNA 1 and cDNA 2) onto the surface of the magnetic core Fe₃O₄. After coupling AuNPs@DTNB@H₁ and AuNPs@4-MBA@H₂ with Fe₃O₄@cDNA 1 and Fe₃O₄@cDNA 2, Fe₃O₄@cDNA-AuNPs@Raman reporter@H was obtained, which generated strong SERS signals. The Fe₃O₄@cDNA-AuNPs@Raman reporter@H and the test solution were added to the inlet area of the pump-free SERS microfluidic chip at the same time. When the test solution contained the target miRNA, the nucleic acid aptamer chains specifically bound to the target miRNA because to their greater affinity and specificity, and the AuNPs carrying Raman reporters were competed away from Fe₃O₄, resulting in a decrease in the SERS signal. Subsequently, we evaluated the specificity, stability, and reproducibility of the pump-free microfluidic chip, and assessed the sensitivity of the pump-free microfluidic chip by measuring the SERS signal intensity of different concentrations of miRNA in serum. Finally, we constructed a mouse liver cancer model, quantitatively detected miR-21 and miR-155 in the serum of mice at various stages, and compared it with qRT-PCR detection. In general, compared to traditional detection methods, the pump-free SERS microfluidic chip proposed in this study is ultra-sensitive, convenient, fast, and highly specific, providing a new idea for the early diagnosis of liver cancer and having great potential.

2. Experimental section

2.1. Materials and reagents

Tetrahydrate chloroauric acid (HAuCl₄·4H₂O), phosphate buffered brine (PBS), 5,5 ‘-dithiobis (2-nitrobenzoic acid) (DTNB), 4-mercaptobenzoic acid (4-MBA), tri (2-carboxyethyl) phosphine (TCEP), anhydrous ethanol, dihydrate, and trisodium citrate (C₆H₅Na₃O₇), polyethylene glycol (PEG), isopropyl alcohol, and polydimethylsiloxane (PDMS) were purchased from Sinopharm Group Chemical Reagents Co. (China), and the purity of the reagents used in the experiment were analytically pure (AR). 100 nm streptavidin-modified Fe₃O₄ nanoparticles, bovine serum protein (BSA), and nucleotides mentioned in Table 1, including miR-21, miR-155, single base mismatch sequence (MT1-1, MT1-2), three base mismatch sequences (MT3-1, MT3-2), random sequence, nucleic acid aptamer sequences (cDNA 1, cDNA 2), and complementary chain sequences (H₁, H₂), etc. were purchased from Shanghai Sangong Biotechnology Co. (China). Quantitative real-time polymerase chain reaction (qRT-PCR) kits were purchased from GeteinBiotech (China). The water used in this experiment was deionized water with a resistivity of 18.2 MΩ. The HepG2-luc cell line used in the experiment was provided by the Cell Bank of the Chinese Academy of Sciences. The nude mice used in the experiment were provided by the Clinical Medicine department of Yangzhou University.

Table 1. Sequences of oligonucleotides used in the experiment.

Name	Sequences
miR-21	UAGCUUAUCAGACUGAUGUUGA
miR-155	UUAAUGCUAAUCGUGAUAGGGGU
MT1-1	UAGCUUAUCAGACUTAUGUUGA
MT1-2	UUAAUGCUAAUCGUGACAGGGGU
MT3-1	UAGAUUAUAAGCCUGAUGUUGA
MT3-2	UUCAUGCAAAUCCUGAUAGGGGU
Random sequence	GGUUUACUUUAGAGACUAGAUC
cDNA 1	Biotin-TCAACATCAGTCTGATAAGCTA
cDNA 2	Biotin-ACCCCUAUCACGAUUAGCAUUAA
H1	SH-TAGCTTATCA
H2	SH-TTAATGCTAA

2.2. Preparation of 20 nm AuNPs

We prepared 20 nm AuNPs by improving the method described by Li et al. [41]. First, 100 mL of HAuCl₄ solution (0.01 wt%) was heated on a magnetic stirrer until boiling, and then 3 mL of 1 wt% sodium citrate solution (TSC) was rapidly added under intense stirring. After reacting for 15 min, the heating was stopped, and the solution was continued to be stirred until it cooled to room temperature. Subsequently, water was added to obtain a total volume of 100 mL, thus preparing the 20 nm AuNPs.

2.3. Preparation of Fe₃O₄@cDNA-AuNPs@Raman reporter@H

200 µL of 10 mM DTNB solution was added to 10 mL of AuNPs solution and stirred vigorously at room temperature for 2 h. To remove excess DTNB, the mixture was centrifuged at 10,000 rpm for 10 min, and then re-suspend the precipitated DTNB-labeled AuNPs (AuNPs@DTNB) in 10 mL of PBS by using ultrasonic and oscillatory treatment after removing the supernatant. The above steps were repeated twice to ensure complete removal of excess signaling molecules. H₁ (0.1 mM) was activated with 20 µL of 1 M TCEP, and added it to the AuNPs@DTNB solution, and incubated at 37 °C for 2 h. Finally, the mixed solution was added to 100 µL of BSA (1 wt%) solution, and centrifuged at 8000 rpm for 15 min after allowing the reaction to proceed for 2 h. AuNPs@DTNB@H₁ was obtained after removing the supernatant. Another SERS label, AuNPs@4-MBA@H₂, was also prepared using the same method. AuNPs@DTNB@H₁ and AuNPs@4-MBA@H₂ are collectively referred to as AuNPs@H.

Then, cDNA 1 and cDNA 2 were incubated at 95 °C for 5 min, and were gradually cooled to room temperature for 15 min. The streptavidin-modified Fe₃O₄ nanoparticle solution was washed three times with ultrapure water and resuspended in PBS to a final concentration of 1 mg/mL. Subsequently, 10 µL of the solution containing 100 µM of cDNA 1 and cDNA 2, respectively, was added to the Fe₃O₄ nanoparticle solution and oscillated at room temperature for 30 min. After magnetic separation with a magnet, ultrapure water was added and washed twice to remove the excess cDNA 1 and cDNA 2. The cDNA-functionalized Fe₃O₄ (Fe₃O₄@cDNA 1 and Fe₃O₄@cDNA 2) was finally obtained, and they were mixed in a 1:1 ratio by volume to obtain Fe₃O₄@cDNA.

The AuNPs@H was mixed with Fe₃O₄@cDNA at a volume ratio of 2.5:1, incubated at 37 °C for 80 min, and washed twice with PBS to remove excess AuNPs@H, the Fe₃O₄@cDNA-AuNPs@Raman reporter@H complex was constructed.

2.4. Construction of a pump-free SERS microfluidic chip

A polydimethylsiloxane (PDMS) microfluidic mold with two microchannels was designed and fabricated using the AutoCAD software. The PDMS was evenly mixed with the curing agent at a mass ratio of 10:1, and then placed in a vacuum chamber for 30 min to remove bubbles. The mixed solution was then placed on the prepared template and cured on a heating plate at 100 °C for 1 h. After cooling to room temperature, the PDMS layer was removed from the template to obtain the pump-free SERS microfluidic chip. The pump-free SERS microfluidic chip was mainly composed of an inlet area (I), a mixing area (II), an enrichment and detection area (III), and a comb capillary pump (IV). The inlet area consisted of two 3 mm circular chambers, and the mixing area consisted of a 2 mm circular chamber and a meandering channel. The enrichment and detection area consisted of a 1.8 mm rectangular chamber and a 4 mm circular magnet. The comb capillary pump channel was 200 µm wide, whereas the other channels were 300 µm wide and 120 µm high. To ensure that the solution passed through the microchannel smoothly, the pump-free SERS microfluidic chip was treated with hydrophilicity using an immersion method. Briefly, the pump-free SERS microfluidic chip was immersed in 95% ethanol, cleaned for 25 min under ultrasonic conditions, and dried with nitrogen. Then, the pump-free SERS microfluidic chip was then immersed in a PEG solution and heated on a heating plate at 150 °C for 30 min. After heating, the excess PEG was washed off with isopropyl alcohol solution and incubated at 4 °C for 1 h. A pump-free SERS microfluidic chip was also fabricated.

2.5. Establishment and serological detection of liver cancer model in nude mice

To establish a liver cancer model in nude mice, 5 × 10⁶ HepG2-luc cells were suspended in 0.1 mL of DMEM culture solution. Eight nude mice were randomly divided into two groups, two in the blank control group and six in the in-situ liver cancer model group. The abdomens of the six nude mice with in-situ liver cancer models, anesthetized with 3% pentobarbital, were dissected. The liver was exposed, and the prepared liver cancer cells were injected into the liver of each nude mouse. The wound was sutured and disinfected, and feeding was continued after woking up. In vivo imaging was performed before the injection and every 10 days thereafter for 30 consecutive days. Fluorescence was observed in the liver of the nude mice, indicating successful modeling. The weights of the mice were measured continuously during the experiment, and blood samples were collected from all nude mice before injection and on days 10, 20 and 30 after injection. The obtained serum was used for SERS detection. After the final blood collection, all nude mice were euthanized, and their livers were removed for hematoxylin and eosin (HE) staining.

2.6. SERS measurement

During the detection process, Fe₃O₄@cDNA-AuNPs@Raman reporter@H and the target miRNAs (miR-21 and miR-155) were added dropwise to the inlet area. The solution automatically flows through the mixing area under the action of the comb capillary pump to react, and the products are enriched by magnets in the enrichment and detection areas. Then, the SERS spectrum of the product was measured with a 50× long working-distance objective lens, and three points at different positions were selected for collection. Their average values were calculated for homogenization processing to ensure the credibility of the results. During the SERS spectrum measurement, the laser wavelength was 785 nm, the exposure time was 10 s, and the power was 5 mW.

2.7. Instruments

Ultraviolet-vis-near infrared (UV-vis-NIR) absorption spectra were measured using a the UNICO 2100 PC UV-Vis spectrophotometer (Japan). Scanning electron microscopy (SEM) figures were obtained using an S-4800II field emission scanning electron microscope. Transmission electron microscopy (TEM) figures were obtained using a Tecnai 12 transmission electron microscope (120 KV). High-resolution figures and element mapping were obtained using a Tecnai G2 F30 field emission transmission electron microscope. All SERS spectra were obtained using an inVia Raman spectrometer (Renishaw, UK).

3. Results and discussions

3.1. Principle of the pump-free SERS microfluidic chip

The preparation process of the pump-free SERS microfluidic chip for the ultra-sensitive and fast analysis of miR-21 and miR-155 is shown in Fig. 1. Figure 1(a) depicts the preparation of the cDNA-functionalized Fe₃O₄: Biotin-modified cDNA 1 and cDNA 2 were annealed and connected to streptavidin modified Fe₃O₄ to obtain Fe₃O₄@cDNA 1 and Fe₃O₄@cDNA 2 (Fe₃O₄@cDNA). Because streptavidin binds biotin with high specificity, biotin-modified cDNA 1 and cDNA 2 can be stably anchored to the surface of streptavidin-modified Fe₃O₄. Figure 1(b) illustrates the preparation process of AuNPs@H: Raman reporters 4-MBA and DTNB can be modified onto the AuNPs surface via Au-S. H₁ and H₂ can also be stably anchored to the AuNPs surface through the modification of sulfhydryl bonds, thus obtaining AuNPs@H (AuNPs@DTNB@H₁ and AuNPs@4-MBA@H₂). Figure 1(c) illustrates the preparation of the Fe₃O₄@cDNA-AuNPs@Raman reporter@H composite structure. Due to partial base complementary pairing, AuNPs@H were able to ligate to Fe₃O₄@cDNA to obtain Fe₃O₄@cDNA-AuNPs@Raman reporter@H. Figure 1(d) shows the detection process for the microfluidic chips. The Fe₃O₄@cDNA-AuNPs@Raman reporter@H solution and the solution to be tested were added to the inlet area (I), and the mixed solution flowed under the driving force of capillary action. The Fe₃O₄@cDNA-AuNPs@Raman reporter@H solution and the solution to be tested react in the mixing area (II). Due to the strong affinity and high specificity between the cDNA and the target, in the presence of the target, the target will bind to the cDNA, resulting in the competitive shedding of AuNPs@H away from Fe₃O₄@cDNA. When the mixture reached the enrichment and detection area (III), Fe₃O₄ was enriched by the magnet and the SERS signal intensity in this region decreased. With an increase in the concentrations of miR-21 and miR-155 in the solution to be tested, the SERS signal decreased more significantly. Based on the linear relationship between the logarithm of the concentration of the target and the SERS signal intensity at 1080 cm⁻¹ and 1330 cm⁻¹, the quantitative detection of miR-21 and miR-155 was realized.

Fig. 1.

(a) Preparation process of the cDNA-functionalized Fe₃O₄(Fe₃O₄@cDNA 1 and Fe₃O₄@cDNA 2). (b)Preparation process of AuNPs@H. (c) Preparation process of composite structure of Fe₃O₄@cDNA-AuNPs@Raman reporter@H. (d) Flow chart of pump-free SERS microfluidic chip used to detect miR-21 and miR-155.

3.2. Characterization of Fe₃O₄@cDNA-AuNPs@Raman reporter@H

As shown in Fig. 2, SEM and TEM were used to characterize the obtained Fe₃O₄@cDNA-AuNPs@Raman reporter@H in detail. The SEM and TEM figures of the Fe₃O₄ core are shown in Fig. 2(a) and Fig. 2(b). The Fe₃O₄ was spherical in shape, with an average particle size of approximately 100 nm. The Fe₃O₄ was evenly distributed, indicating that it had good dispersion and could better capture target miRNAs. The UV-vis spectrum of Fe₃O₄ is shown in Fig. 2(c), and the absorption peak of Fe₃O₄ is at 413 nm. Figure 2(d) and Fig. 2(e) show the SEM and TEM figures of the AuNPs, respectively. It can be seen from the figures that the AuNPs have a spherical shape with an average particle size of approximately 20 nm, and the absorption peak of the AuNPs is at 522 nm, as shown in Fig. 2(f). This shows that we have successfully synthesized Au nanoparticles with uniform morphologies and sizes. We conducted SEM, TEM, and HRTEM imaging of the prepared Fe₃O₄@cDNA-AuNPs@Raman reporter@H (Fig. 2(g)-Fig. 2(i)): The prepared Fe₃O₄@cDNA-AuNPs@Raman reporter@H, with a diameter of about 140 nm, has good dispersion, and 20 nm AuNPs are evenly distributed on the surface of Fe₃O₄, forming numerous “hot spots”. The lattice figure of the AuNPs shows that the fringe spacing is 0.243 nm, indicating a polycrystalline structure (Fig. 2(j)). In order to further understand the structure and spatial element distribution of Fe₃O₄@cDNA-AuNPs@Raman reporter@H, we have carried out energy-dispersive X-ray (EDX) spectral element mapping. Figure 2 k to Fig. 2(n) are the element mapping figures of Fe₃O₄@cDNA-AuNPs@Raman reporter@H, which show that Fe₃O₄@cDNA-AuNPs@Raman reporter@H is mainly composed of Fe and Au. Fe₃O₄@cDNA-AuNPs@Raman reporter@H had a better SERS enhancement effect, as shown in Fig. S1, the characteristic peak intensity of Fe₃O₄@cDNA-AuNPs@Raman reporter@H at 1330 cm⁻¹ was significantly higher than that of the pure DTNB, and the enhancement factor (EF) was calculated by the formula of EF = (I_SERS/C_SERS)/(I_RS/C_RS), where I_SERS and I_RS represent the signal intensity of Fe₃O₄@cDNA-AuNPs@Raman reporter@H and pure DTNB, respectively, and C_SERS and C_RS represent the concentration of Fe₃O₄@cDNA-AuNPs@Raman reporter@H and pure DTNB, respectively. When C_SERS and C_RS were set to 1 × 10⁻⁸ M and 1 × 10⁻¹ M, respectively, the EF values were calculated to be 8.58 × 10⁷ at 1330 cm⁻¹.

Fig. 2.

(a) SEM figure of Fe₃O₄. (b) TEM figure of Fe₃O₄. (c) UV-vis spectrum of Fe₃O₄. (d) SEM figure of AuNPs. (e) TEM figure of AuNPs. (f) UV-vis spectrum of AuNPs. (g) SEM figure of Fe₃O₄@cDNA-AuNPs@Raman reporter@H. (h) TEM figure of Fe₃O₄@cDNA-AuNPs@Raman reporter@H. (i) HRTEM figure of Fe₃O₄@cDNA-AuNPs@Raman reporter@H. (j) Electron diffraction figure of Fe₃O₄@cDNA-AuNPs@Raman reporter@H. (k-n) Element mapping figure of Fe₃O₄@cDNA-AuNPs@Raman reporter@H.

3.3. Optimization of experimental parameters

In order to find the best detection conditions for the SERS sensor and improve the detection efficiency, the key experimental parameters were optimized. The assembly ratio of Fe₃O₄@cDNA and AuNPs@H is crucial for the establishment of SERS sensors. The Fe₃O₄@cDNA was incubated with the AuNPs@H at different volume ratios (1:1, 1:1.5, 1:2, 1:2.5, 1:3 and 1:3.5) for 1.5 h at 37 °C to allow for complete coupling, and then the mixture was subjected to SERS detection. As shown in Fig. 3(a) and Fig. 3(b), when the ratio of Fe₃O₄@cDNA to AuNPs@H was 1:2.5, the intensity of the characteristic peaks at 1080 cm⁻¹ and 1330 cm⁻¹ reached their highest, and when the volume of AuNPs@H continued to increase, the signal intensity of these two characteristic peaks changed little, indicating that the coupled AuNPs@H on the Fe₃O₄@cDNA has tended towards saturation. This indicates that when the volume ratio of Fe₃O₄@cDNA and AuNPs@H was 1:2.5, it was the best ratio for constructing the SERS sensor. In addition, the incubation time of the Fe₃O₄@cDNA and the AuNPs@H was another key factor affecting the detection efficiency of the SERS sensors. Figure 3(c) and Fig. 3(d) show that the intensities of the characteristic peaks at 1080 cm⁻¹ and 1330 cm⁻¹ increased with the extension of time; however, when the time reached 80 min, the SERS signal intensity did not change significantly, and the increasing trend almost disappeared. Therefore, the optimal incubation times for Fe₃O₄@cDNA and the AuNPs@H were 80 min. Changes in incubation temperature also affect the hybridization efficiency of the complementary nucleic acid chain, and thus, the coupling of Fe₃O₄@cDNA and AuNPs@H. The SERS signal intensities measured at different temperatures are shown in Fig. 3(e) and Fig. 3(f). As the temperature increased, the intensity of the characteristic peaks at 1080 cm⁻¹ and 1330 cm⁻¹ gradually increased and tended to stabilize at 37 °C. Therefore, 37 °C was chosen as the optimal temperature for the experiment.

Fig. 3.

(a) SERS signal intensity when the Fe₃O₄@cDNA and the AuNPs@H were assembled at different ratios. (b) SERS signal intensity histogram when the Fe₃O₄@cDNA and the AuNPs@H were assembled at different ratios. (c) SERS signal intensity when the Fe₃O₄@cDNA and the AuNPs@H were incubated at different times. (d) Histogram of SERS signal intensity at different times when Fe₃O₄@cDNA and AuNPs@H were incubated. (e) SERS signal intensity at different temperatures when the Fe₃O₄@cDNA and the AuNPs@H were incubated. (f) Histogram of SERS signal intensity c temperatures when the Fe₃O₄@cDNA and the AuNPs@H were incubated.

3.4. Preparation of the pump-free SERS microfluidic chip

During the preparation of pump-free SERS microfluidic chips, hydrophilic treatment was performed to ensure a stable flow of the solution. Therefore, the hydrophilicity and sealing properties of pump-free SERS microfluidic chips were analyzed. Red and blue inks were added to the inlet area respectively, as shown in Fig. 4(a). Under the action of the comb capillary pump, the ink could flow freely in the channel and filled the entire channel within 60 s without any leakage, indicating that the prepared pump-free SERS microfluidic chip had good hydrophilicity and sealing properties. Because the PDMS used to prepare the pump-free SERS microfluidic chip itself had Raman signals, two regions (I and II) on the chip were selected for SERS detection in order to exclude their influence on subsequent detection results, representing the conditions before and after the reaction, respectively (Fig. 4(b)); the corresponding SERS spectrum is shown in Fig. 4(c). The substance in region I is a mixture of miR-21 and miR-155, and the substance in region II is the product of the reaction of Fe₃O₄@cDNA-AuNPs@Raman reporter@H with a mixture of miR-21 and miR-155. Since there were no Raman signaling molecules in region I, the Raman signals in this region were entirely generated by PDMS. The characteristic peak intensities at 1080 cm⁻¹ and 1330 cm⁻¹ in region I were significantly lower than those in region II. Therefore, the microfluidic chip made of PDMS did not affect miR-21 and miR-155 detection. To determine the optimal measurement time for the pump-free SERS microfluidic chip, the SERS signals in the rectangular cavity were measured at different times. The results in Fig. 4(d) show that the SERS intensity decreased with an increase in time, and the SERS intensity hardly changed after 5 min. The optimal measurement time for the prepared pump-free SERS microfluidic chip was 5 min.

Fig. 4.

(a) Inks of different colors flowing freely in the channel of a pump-free SERS microfluidic chip. (b) Micrographs of two regions in a pump-free SERS microfluidic chip. (c) Corresponding SERS spectra of the two regions. (d) SERS signal in a rectangular cavity was Intensity changes of characteristic peaks at 1080 cm⁻¹ and 1330 cm⁻¹.

3.5. Performance assessment of the pump-free SERS microfluidic chip

To evaluate the repeatability of the pump-free SERS microfluidic chips, five different batches of SERS microfluidic chips were prepared and their SERS spectra were measured. As shown in Fig. 5(a), there was no significant difference in the SERS spectra of the five different batches of prepared pump-free SERS microfluidic chips, and their Raman intensities at 1080 cm⁻¹ and 1330 cm⁻¹ are shown in Fig. 5(b). The relative standard deviations (RSD) of the Raman intensities at 1080 cm⁻¹ and 1330 cm⁻¹ were 5.68% and 5.63%, respectively. Figure 5(c) shows the stability of the pump-free SERS microfluidic chips. After storage at room temperature for different durations (1 d,5 d,10 d,15 d,20 d), the SERS signal intensity was measured. Although the SERS signal decreased, the spectral form did not change significantly. Figure 5(d) shows the change in SERS signal intensity at the characteristic peaks of 1080 cm⁻¹ and 1330 cm⁻¹ from 1 to 20 days, and their RSD values were calculated to be 5.64% and 5.72%, respectively. These results show that the prepared pump-free SERS microfluidic chip has good repeatability and stability. Specificity is also an important factor that affects the detection performance of pump-free SERS microfluidic chips. Therefore, we introduced single base mismatch sequences (MT1-1 and MT1-2), three-base mismatch sequences (MT3-1 and MT3-2), and random sequences as interferences in the experiment. Figure 5(e) shows that when the target miR-21 and miR-155 were present, the corresponding characteristic peaks decreased significantly, and the change in SERS signal intensity caused by the target miRNAs was significantly higher than that caused by interferences. According to the corresponding bar chart in Fig. 5(f), it is more obvious that the changes in the characteristic peaks of 1080 cm⁻¹ and 1330 cm⁻¹ represented by the two target miRNAs were stronger than those of the interferers. Thus, the prepared pump-free SERS microfluidic chip was highly sensitive to miR-21 and miR-155 expression. Moreover, miR-21, miR-155, and their interferences can be accurately distinguished with good specificity.

Fig. 5.

(a) SERS spectrum of pump-free SERS microfluidic chips prepared in different batches. (b) Characteristic peak intensities at 1080 cm⁻¹ and 1330 cm⁻¹. (c) SERS spectrum of pump-free SERS microfluidic chips stored at room temperature for different times. (d) Characteristic peak intensities at 1080 cm⁻¹ and 1330 cm⁻¹. (e) Specificity of the pump-free SERS microfluidic chip. (f) Histogram diagram of characteristic peak intensity at 1080 cm⁻¹ and 1330 cm⁻¹.

3.6. Simultaneous quantitative detection of miR-21 and miR-155

The lower limit of detection (LOD) is an important parameter that influences the performance of pump-free SERS microfluidic chips. A lower LOD can better detect miRNAs at lower concentrations in the blood. After optimizing the detection conditions, we used the prepared pump-free SERS microfluidic chip to detect miR-21 and miR-155 in serum samples dispersed at different concentrations. The concentration of the solutions ranged from 10 pM to 10 mM. As shown in Fig. 6(a), the SERS signal gradually decreased with an increase in the target miRNA concentration. As shown in Fig. 6(b) and Fig. 6(c), the SERS signals at 1080 cm⁻¹ and 1330 cm⁻¹ exhibited a linear relationship with the logarithm of the target miRNA concentration in the serum within the range of 10 pM to 10 mM. The corresponding regression equations were y = -2221.525x-6883.168(R²= 0.996) and y = -2613.809x-8185.615(R²= 0.996). The calculated LODs of miR-21 and miR-155 was 2.39 pM and 2.49 pM, respectively. LOD is defined as the concentration of the analyte at three standard deviations of the blank signal, with the following equation: LOD = 3sd/k, where 3sd is the three standard deviations of the blank signal and k is the slope of the linear equation. Compared with other detection strategies previously reported (Table 2), the detection results obtained using the pump-free SERS microfluidic chip prepared by us were similar, but the detection time was greatly shortened. This pump-free SERS microfluidic chip has good potential for the simultaneous quantitative detection of two miRNAs.

Fig. 6.

(a) SERS spectra of different concentrations of miR-21 and miR-155(10 pM-10 mM) in serum. (b) Calibration curve of characteristic peak intensity and logarithm of miR-155 concentration at 1080 cm⁻¹ and (c) Calibration curve of characteristic peak intensity and logarithm of miR-21 concentration at 1330 cm⁻¹ Thread.

Table 2. Comparison of the proposed microfluidic chip reported in this paper with methods documented in prior research.

Method	Analyte	LOD	Time	Ref.
Electrochemical detection	miR-34a	12.73 pM	60 min	[42]
Fluorescence	miR-21	0.89 fM	40 min	[43]
QCM	miR-21	28 fM	90 min	[44]
LFDA	miR-21	0.25 am	20 min	[45]
SERS	miR-375	10 fM	60 min	[46]
SERS microfluidic	miR-21& miR-155	2.39 pM	5 min	This work
chip		2.49 pM

3.7. Characterization of mice with hepatocellular carcinoma

Nude mice were photographed using a small-animal imaging system to observe and compare the growth of in-situ tumors at different time points, as shown in Fig. 7(a). The fluorescence intensity of in-situ liver cancer cells in nude mice continuously increased over time. HE staining showed that the tumor cells in nude mice with liver cancer were arranged disorderly, with varying sizes, deeply stained nuclei, and abnormal division (Fig. 7(b)). When serum was collected, the body weights of the nude mice were measured simultaneously. As shown in Fig. 7(c), the body weight of nude mice with liver cancer decreased continuously compared with that of the normal control group. All of the above results indicated that the modeling of nude mice with in-situ liver cancer was successful, and the serum could be collected for SERS detection.

Fig. 7.

(a) Fluorescent live imaging of nude mice and (b) liver tissue HE staining. (c) Average body weight of nude mice in each group.

3.8. Real sample analysis

To verify the application value of the detection method in the actual samples, SERS detection was performed on the expression levels of miR-21 and miR-155 in the serum of nude mice at different time periods (0, 10, 20, 30 d), and the detection results of qRT-PCR were compared to ensure the accuracy of the pump-free SERS microfluidic chip detection results. Figure 8(a) and Fig. 8(b) show the average SERS spectrum of nude mice at 0 d (control group) and the SERS intensity at the characteristic peaks of 1080 cm⁻¹ and 1330 cm⁻¹. The SERS intensity of the serum of nude mice in the control group did not change significantly over time. In Fig. 8(c), the average SERS spectrum of nude mice with in-situ liver cancer significantly decreased with tumor growth, and the SERS intensities of the corresponding characteristic peaks at 1080 cm⁻¹ and 1330 cm⁻¹ are shown in Fig. 8(d). The expression levels of miR-21 and miR-155 in nude mice with in-situ liver cancer were higher than those in control mice, and the expression levels increased with tumor development. The comparison between the detection results of miR-21 and miR-155 by the pump-free SERS microfluidic chip and the detection results of qRT-PCR is shown in Table 3, demonstrating a high consistency. This indicates that the pump-free SERS microfluidic chip proposed in this study has a high accuracy rate for the simultaneous and ultra-sensitive detection of miR-21 and miR-155.

Fig. 8.

(a) SERS spectrum of serum of nude mice at day 0, 10, 20 and 30 of control group. (b) Characteristic peak intensity of nude mice at 1080 cm⁻¹ and 1330 cm⁻¹ of control group. (c) SERS spectrum of serum of nude mice with in-situ liver cancer at day 0, 10, 20, 30. (d) Characteristic peak intensity of nude mice with in-situ liver cancer at 1080 cm⁻¹ and 1330 cm⁻¹.

Table 3. Results of SERS and qRT-PCR of nude mice with orthotopic hepatocellular carcinoma.

Day	SERS (mol/L)		qRT-PCR (mol/L)		Relative error (%)
	miR-21	miR-155	miR-21	miR-155	miR-21	miR-155
0	(1.08 ± 0.07) × 10−10	(1.12 ± 0.08) × 10−10	(1.01 ± 0.16) × 10−10	(1.09 ± 0.02) × 10−10	6.22	2.60
10	(2.67 ± 0.71) × 10−9	(2.81 ± 0.75) × 10−9	(2.75 ± 0.02) × 10−9	(2.93 ± 0.06) × 10−9	-3.01	-4.05
20	(1.57 ± 0.55) × 10−8	(1.66 ± 0.59) × 10−8	(1.60 ± 0.05) × 10−8	(1.71 ± 0.01) × 10−8	-2.11	-2.88
30	(1.88 ± 0.21) × 10−7	(1.99 ± 0.23) × 10−7	(1.71 ± 0.01) × 10−7	(1.95 ± 0.03) × 10−7	9.70	2.25

4. Conclusions

In summary, we successfully prepared a pump-free SERS microfluidic chip for rapid and ultra-sensitive simultaneous detection of two liver cancer related miRNAs (miR-21 and miR-155), enabling early detection of liver cancer. As a magnetic core, Fe₃O₄ has a good enrichment ability, which enables it to capture more targets and separate them from a complex environment under the action of magnets. AuNPs@H were coupled to Fe₃O₄@cDNA to form rich “hot spots”, which significantly enhanced the SERS signal. We adopted an identification-competitive strategy. The prepared Fe₃O₄@cDNA-AuNPs@Raman reporter@H utilized the strong specificity between the nucleic acid aptamer chains with miR-21 and miR-155, achieving higher sensitivity, which is superior to that of other detection strategies. The SERS microfluidic chip, designed with a capillary pump, eliminated the need for external heavy-duty pumps and could quickly and efficiently complete the detection of miR-21 and miR-155 within 5 min, with LOD values reaching 2.39 pM and 2.49 pM, reflecting the good convenience and rapidity of the pump-free SERS microfluidic chip. The results of the multiple assays showed that the pump-free SERS microfluidic chip had good stability, reproducibility, and specificity. In addition, the SERS detection results for nude mouse serum samples were consistent with those of qRT-PCR, verifying the accuracy and applicability of the detection method. We believe that the method developed in this study for detecting serum miRNA expression levels has good prospects and may be helpful to the early diagnosis of liver cancer.

Funding

Major Programs of Natural Science Foundation of higher education in Jiangsu Province (19KJA480003); Natural Science Foundation of Jiangsu Province10.13039/501100004608 (BK20171290); Administration of Traditional Chinese Medicine Project of Jiangsu Province (MS2021081); Health Project of Lianyungang (202341); Project of Jiangsu Key laboratory of Experimental & Translational Non-coding RNA Research (202274); Lianyungang high-level talent training project (LYG06521202365).

Disclosures

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this study.

Data availability

Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.

Supplemental document

See Supplement 1 ^{(963.7KB, pdf)} for supporting content.