Multiplexed cancer biomarker detection using chip-integrated silicon photonic sensor arrays

Abstract

The analysis of disease-specific biomarker panels holds promise for the early detection of a range of diseases, including cancer. Blood-based biomarkers, in particular, are attractive targets for minimally-invasive disease diagnosis. Specifically, a panel of organ-specific biomarkers could find utility as a general disease surveillance tool enabling earlier detection or prognostic monitoring. Using arrays of chip-integrated silicon photonic sensors, we describe the simultaneous detection of eight cancer biomarkers in serum in a relatively rapid (1 hour) and fully automated antibody-based sandwich assay. Biomarkers were chosen for their applicability to a range of organ-specific cancers, including disease of the pancreas, liver, ovary, breast, lung, colorectum, and prostate. Importantly, we demonstrate that selected patient samples reveal biomarker "fingerprints" that may be useful for a personalized cancer diagnosis. More generally, we show that the silicon photonic technology is capable of measuring multiplexed panels of protein biomarkers that may have broad utility in clinical diagnostics.

Figure 1

Photograph of silicon microchip overlaid with four red boxes indicating sub-array regions containing 8 microring resonators, each functionalized with an antibody specific for a different cancer biomarker antigen.

Figure 2

8-plex analysis for cross reactivity and 8-plex serum sensing. (a) Response of 8-plex chip upon sequential addition of 1 μg/mL (or 1 kU/mL) of each antigen in buffer. Each plot shows the responses for the entire array of microring sensors, each functionalized with a different capture antibody, to a single biomarker. (b) Response of the same 8-plex chip upon subsequent addition of 1 μg/mL of the capture antibody. Each plot shows the responses for the entire array exposed to the single listed analyte and corresponding tracer antibody. (c) Schematic showing the layer-by-layer singnal enhancement strategy. (d) A real-time plot showing the resonance wavelength shifts through the entire assay analysing a representative human serum sample. Vertical arrows represent the start of the primary, tracer, and secondary enhancement steps, respectively. The * symbol indicates a buffer rinse. For panels, a, b, and d, responses are color-coded according to the capture antibody on each microring:anti-AFP = black, and-ALCAM = red, anti-CA19-9 = cyan, anti-osteopontin = blue, anti-CA15-3 = purple, anti-CEA = orange, anti-CA125 = yellow, anti-PSA = green.

Figure 3

Calibration curves for each cancer biomarker. Each plot shows the dose-response relationship and fit for each of the targeted antigens as detected by the specific sandwich assay immunoassay.

Figure 4

Real-time response of the 8-plex cancer biomarkers assay in eight evaluated serum samples from healthy patients as well as cancer patients. Each panel shows the real-time shift in resonance shift throughout the entire assay to detect the 8-plex biomarker panel from each commercial, organ-specific serum sample. Although clear differences are observed through the primary antigen binding response, quantification is achieved only from the layer-by-layer secondary enhancement component of the assay.

Figure 5

Relative biomarker “fingerprints” for selected cancer patient serum samples. This “fingerprint” was generated by dividing the sensor response for each biomarker in each cancer serum sample by the response measured in the healthy serum sample. Error bars indicate the relative standard deviation for the four technical replicate sensors on each chip used to measure each antigen. A relative index level of 1.0 indicates (horizondal dashed lines in each plot) that the biomarker was equivalent in both disease and normal sample, with indices greater than 1.0 representing organ-specific biomarker elevation.