Quantum Dot Encoding for In-Solution Single-Molecule Biomarker Counting in Metastatic Prostate Cancer.

Digital assays are in wide development for biomarker quantification at the single-molecule level, but the common use of surface-pulldown steps limits both analytical sensitivity and throughput. Here, we develop surface-free, wash-free, in-solution assays with a sensitivity slope approaching unity for sequence-specific counting of microRNAs (miRs) relevant to metastatic castration-resistant prostate cancer (mCRPC). These assays are enabled by DNA nanoflowers (DNFs) densely encoded with ∼200 fluorescent quantum dots (QDs) that assemble in situ stoichiometrically to miRs. The QD-DNFs are detected as single events in solution by fluorescence microscopy or flow cytometry without washing away unbound labels. A ∼50 aM limit of detection and high agreement with absolute target count (0.95) were achieved by machine learning-guided assay optimization, providing the potential for calibration-free measurements. Multiple miR sequences could be distinguished through ratiometric and colorimetric (5-color) QD signatures with a single excitation source for flexible detection scenarios in static solution or flow streams. The assays were applied for detecting exosomal miRs from small-volume plasma of mCRPC patients and showed strong agreement with RT-qPCR, but with more reliable detection of the trace prognostic biomarker miR-375. Consistent with our prior reports using large volume blood draws, higher plasma levels of miR-375 were associated with poor survival of patients with mCRPC. We anticipate that in-solution absolute counting of clinical biomarkers in plasma will enable robust molecular analysis of trace biomarkers needed for the translation of cancer precision medicine.