São Carlos Institute of Physics, University of São Paulo (USP), São Carlos, 13566-590, São Paulo, Brazil; Nanotechnology National Laboratory for Agriculture (LNNA), Embrapa Instrumentação, 13560-970 São Carlos, SP, Brazil.
Nanotechnology National Laboratory for Agriculture (LNNA), Embrapa Instrumentação, 13560-970 São Carlos, SP, Brazil.
Biomater Adv. 2022 Mar;134:112676. doi: 10.1016/j.msec.2022.112676. Epub 2022 Jan 20.
Low-cost sensors to detect cancer biomarkers with high sensitivity and selectivity are essential for early diagnosis. Herein, an immunosensor was developed to detect the cancer biomarker p53 antigen in MCF7 lysates using electrical impedance spectroscopy. Interdigitated electrodes were screen printed on bacterial nanocellulose substrates, then coated with a matrix of layer-by-layer films of chitosan and chondroitin sulfate onto which a layer of anti-p53 antibodies was adsorbed. The immunosensing performance was optimized with a 3-bilayer matrix, with detection of p53 in MCF7 cell lysates at concentrations between 0.01 and 1000 U. mL, and detection limit of 0.16 U mL. The effective buildup of the immunosensor on bacterial nanocellulose was confirmed with polarization-modulated infrared reflection absorption spectroscopy (PM-IRRAS) and surface energy analysis. In spite of the high sensitivity, full selectivity with distinction of the p53-containing cell lysates and possible interferents required treating the data with a supervised machine learning approach based on decision trees. This allowed the creation of a multidimensional calibration space with 11 dimensions (frequencies used to generate decision tree rules), with which the classification of the p53-containing samples can be explained.
低成本传感器对于高灵敏度和选择性地检测癌症生物标志物至关重要，有助于实现癌症的早期诊断。在此，我们开发了一种基于电化学阻抗谱的免疫传感器，用于检测 MCF7 细胞裂解液中的癌症标志物 p53 抗原。叉指电极通过丝网印刷技术被印制在细菌纳米纤维素基底上，然后通过层层自组装技术在基底上涂覆壳聚糖和硫酸软骨素基质，在该基质上吸附了一层抗 p53 抗体。通过使用 3 层基质优化了免疫传感器的性能，在 0.01 至 1000 U·mL 浓度范围内检测 MCF7 细胞裂解液中的 p53，检测限为 0.16 U·mL。通过偏振调制红外反射吸收光谱（PM-IRRAS）和表面能分析证实了在细菌纳米纤维素上有效构建了免疫传感器。尽管具有高灵敏度，但要实现完全选择性，需要区分含有 p53 的细胞裂解液和可能的干扰物，这需要使用基于决策树的监督机器学习方法处理数据。这使得能够创建一个具有 11 个维度的多维校准空间（用于生成决策树规则的频率），通过该空间可以解释含有 p53 的样品的分类。
