Currently, the traditional biomarkers used for human in vitro diagnostic testing are divided into two categories: proteins and nucleic acids. Traditional protein biomarkers include prostate-specific antigen (PSA) for prostate cancer, CA125 for ovarian cancer, CA19-9 for pancreatic cancer, carcinoembryonic antigen (CEA) for colorectal cancer, and alpha-fetoprotein (AFP) for liver cancer. Nucleic acid biomarkers include DNA and RNA. MicroRNA (miRNA) is an endogenous, non-coding, single-stranded, and small RNA, with a length equaling approximately 18-22 nucleotides, and regulates over 50% of human protein-coding genes. miRNA plays a key role in cell differentiation, proliferation, apoptosis, metabolism, and immune response. Its abnormal expression is associated with major diseases such as cardiovascular disease, neurological disease, immune disease, rheumatoid arthritis, and various cancers. Abnormal expression of miRNA can be detected in almost all types of cancer diseases, such as let-7's high expression and miR-155's low expression associated with poor prognosis in non-small cell lung cancer; miR-103 promotes cancer cell migration by increasing vascular permeability in liver cancer; miR-21 serves as upregulation of an oncogene in lung and breast cancers. miR-21 is the most commonly upregulated miRNA in tumor cells and is associated with every aspect of cancer development, including genomic instability and mutation, cell proliferation, inflammation, metabolic abnormalities, evading apoptosis, immune destruction, and growth inhibition.

miRNA is an important biomarker in the detection of major diseases such as cancer. While quantitative real-time polymerase chain reaction (RT-qPCR) suffers from amplification bias due to reverse transcription limitations. Northern blotting has limitations in sensitivity. Next-generation sequencing (NGS) and single-molecule array technology (SiMoA) have advantages in low detection limits and high sensitivity. However, amplification bias or a lack of simplified workflow will hinder the real-time point-of-care medical diagnosis, treatment, and prognosis. In this paper, we presented a miRNA detection method based on the single-molecule detection principle.

To begin with, a sandwich structure was formed by using the Poisson distribution to create a complex. The captured probe-coated magnetic beads bound to half of the base pairs of the miRNA, satisfying the binding rule, while the other half of the miRNA base pairs bound to biotinylated detection probes. The detection probe then bound to streptavidin-poly-HRP, forming a complex. In an H₂O₂ solution, streptavidin-poly-HRP bound to the tyramine-Alexa Fluor 488 molecule in a catalytic deposition manner to amplify the signal. Subsequently, the complex was immobilized by using fibrin hydrogel instead of microfluidic chips. The complex underwent Brownian motion in solution, continuously moving in an irregular pattern and randomly colliding with other suspended complexes. In order to immobilize the complex and facilitate single-molecule counting of miRNA, a simplified method for immobilizing the complex was designed by using fibrin hydrogel instead of microfluidic chips, which solved the design and manufacturing issues required by microfluidic chips. Fibrin hydrogel was generated by the polymerization of fibrinogen under the action of thrombin and catalytic factors. Fibrinogen was a glycoprotein synthesized and secreted from stem cells, and each fibrinogen molecule consisted of three pairs of different peptide chains, namely α, β, and γ, which were arranged symmetrically on both sides. The molecules were connected by disulfide bonds both between and within the molecules, and fibrinogen molecules existed in a polymeric form in solution through intermolecular interactions. Firstly, fibrinogen was converted to fibrin peptides by the action of thrombin, resulting in an unstable soft clot. Then, the inactive fibrin stabilizing factor (Factor XIII) was activated by the action of thrombin and Ca²⁺ to become an active fibrin stabilizing factor (Factor XIIIa). Finally, under the action of Ca²⁺, the fibrin stabilizing factor completed the cross-linking of peptides through transglutaminase activity, forming fibrin hydrogel. As a degradable material, fibrin hydrogel had the characteristics of biocompatibility, a certain degree of transparency, transmission spectra covering ultraviolet to near-infrared, and high transmittance. At last, single-molecule counting processing was performed. After the fibrin hydrogel immobilized the complex, images were acquired in bright and dark fields, and single-molecule counting algorithms were applied. The single-molecule counting algorithm consisted of four steps. First, spot addressing. In the bright field, the complement of the image was calculated to make the magnetic beads bright and the background dark. The uneven illumination distribution effect was removed based on top-hat transformation. In the dark field, contrast enhancement was applied, and the uneven illumination distribution effect was removed based on top-hat transformation. Second, spot screening. In the bright field, single magnetic beads were identified and screened by morphology processing. In the dark field, single bright spots were identified and screened by morphology processing. Third, image overlay. The images obtained in the bright and dark fields were overlaid, aligning the positions of magnetic beads and bright spots. Fourth, information extraction. Magnetic beads with bright signals were recognized and counted as positive spots. Finally, all positive spots were counted to achieve ultra-sensitive quantification of miRNA.

In this paper, human miR-21 was used as the detection target, with a detection limit of 6 fmol/L (Fig. 5). This method has great potential application value for future in vitro diagnosis and detection of miRNA.