Multivariate-Activated DNA Logic Gate for Predicting Responses to Cisplatin-Based Chemotherapy

Abstract

Abstract Cisplatin-based chemotherapy remains an attractive alternative for controlling advanced cancers. However, a subset of cancer patients exhibits resistance to cisplatin. Accurate and early prediction of individual responses to cisplatin treatment is critical to personalized medicine guidance. The development of an effective method to predict chemotherapy response in cancer patients still faces challenges. Based on extracellular lactate, a key indicator of cisplatin resistance, herein, we report a multivariate-activated DNA logic gate to achieve lactate imaging with high fidelity. This device is assembled by a PNA–peptide–PNA copolymer and a lactate-specific aptamer with a cell membrane-anchoring aptamer module. It operates in response to the matrix metalloproteinase in the tumor microenvironment (TME) after the device anchors to nucleolin; thereby, the liberated aptamer binds to lactate through structure switching, achieving “off–on” fluorescence imaging. Additionally, imaging characteristics in combination with the Logistic Regression algorithm enable noninvasive prediction of cancer responses to cisplatin-based chemotherapy. In 90 cisplatin-resistant tumor-bearing mice, this detection model effectively identifies 77 of the 90 samples, showing a satisfactory accuracy of 85.6%. We further monitor the lactate level decreases in the TME using our method after the drug-resistant tumor-bearing mice treatment by stiripentol and successfully predict the enhancement of the therapeutic effect and the extension of the survival period. This study highlights the combined power of the DNA logic gate and machine learning for the analysis of lactate, paving the way for clinical cisplatin resistance detection and facilitating the formulation of personalized medicine guidance.