A Compact Prism-Based Microscope for Highly Sensitive Measurements in Fluid Biopsy

Abstract

The increasing demand for sensitive, portable, and affordable disease detection methods has spurred the development of advanced biosensors for rapid early-stage diagnosis, population mass screening, and bed-monitoring. Current high-sensitivity devices face hurdles such as high production costs and challenges in multiplexed signal detection. To address these, we developed a prism-based total internal reflection system which, in combination with surface functionalization techniques of gold nanoparticles, enables evanescent wave scattering for highly sensitive and rapid detection of specific analytes in both synthetic and human liquid samples. To validate its efficacy, we conducted scattering experiments in synthetic and human serum samples, exploiting functionalized AuNPs to recognize bacterial lipopolysaccharides as biomarkers for sepsis disease. We demonstrate a remarkable sensitivity in the femtogram per mL concentration range for this specific pathological biomarker. Based on this result we envisage the potential adoption of our technique for liquid biopsy in the clinical scenario.

Keywords: biosensors; compact set‐up; lipopolysaccharides; nanoparticles; prism‐based evanescent wave; scattering.

FIGURE 1

Graphical scheme of the optical coupling between the glass microprism and the glass coverslip through immersion oil, with the same refractive index of glass. A refractive index mismatch occurs at the interface between the glass coverslip and the sample buffer. As a consequence, total internal reflection occurs at this interface, generating an EW, which propagates within the sample. The inset shows the optical path of light rays at the interface between the two media, following the Snell's law.

FIGURE 2

(a) Photos of the compact TIR setup from different views, showing the optomechanical components described in the main text. (b) Scheme of the optical path. L1: Collimating lens (f = 20 mm); M1, 2, 3: Dielectric mirrors. Inset: Geometry and dimensions of the glass microprism.

FIGURE 3

(a) Schematic procedure of gold nanoparticles functionalization with the thiol‐functionalized peptide synthesis for the selective capturing of LPS. Characterization of NPs‐peptide construct: (b) UV–Vis absorption spectra of NSps‐citrate (black line) and NSps‐peptide (red line). Inset shows the region 520–560 nm; (c) autocorrelation curves of NSps‐citrate (black line) and NSps‐peptide (red line) measured through Dynamic Light Scattering analysis. Inset shows size distribution and ζ potential of NSps‐citrate and NSps‐peptide; (d) Raman spectra of NSps‐citrate (black line) and NSps‐peptide (red line). Measurements were performed with a 785 nm laser, 20 mW, time acquisition 60s, and 2 accumulations.

FIGURE 4

Graphical scheme of image processing procedure: First, the image is opened with ImageJ and a threshold value is chosen to remove the background noise. After noise subtraction, the function “Analyse particles” is run on data with a minimum size value for particles of 10 pixels². This value comes from the calculation of the lateral resolution of the image, which is 702 nm (see Supporting Information: Section S7). From this analysis, the number of particles detected, their surface, and their average intensities are given, for each image.

FIGURE 5

(a) Scattering measurements were performed at LPS concentrations spanning from 10⁻⁶ to 1 ng/mL. N ≥ 20 for each LPS concentration. Error bars indicate the standard error. NPs mean density in function of LPS concentration was graphed both in histograms and dot forms, and the shaded area shows the linear dynamic range. Two‐sample student t‐tests have been performed to evaluate the statistical differences between different LPS concentrations. The lowest concentration that can be detected with a statistically significant p value is in the femtogram per milliliter (fg/mL) range. (b) Scattering images at different LPS concentrations were acquired with the compact optical setup. Bright circled spots are scattered AuNPs. Scalebar 10 μm. (c) Histogram showing the reproducibility measurements performed on four different LPS concentrations. Five different LOC‐NPs samples were observed for each concentration (white bars), (N = 24 for each concentration) and the mean value of their density distribution was calculated, along with its standard error. (d) Batch‐to‐batch NPs density variation for the four tested concentrations.

FIGURE 6

(a) Mean NPs density values obtained from experiments in human serum (red bars, N > 60) and in synthetic aqueous samples (blue bars, N > 20) at specified LPS concentrations. (b) Table comparing measured and expected mean densities at specific LPS concentrations and the corresponding calculated recovery rate (RR) %.