Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES
Article
Adaptable detection strategies in membrane-based immunoassays: calibration-free quantitation with surface-enhanced Raman scattering readout Aleksander Skuratovsky, Robert J Soto, and Marc D. Porter Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b01958 • Publication Date (Web): 24 May 2018 Downloaded from http://pubs.acs.org on May 24, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Adaptable detection strategies in membrane-based immunoassays: calibration-free quantitation with surface-enhanced Raman scattering readout Aleksander Skuratovskyϟ†‡, Robert J. Sotoϟ‡, and Marc D. Porter*†‡⁑ †Department of Chemical Engineering, ‡Nano Institute of Utah, and ⁑Department of Chemistry, The University of Utah, Salt Lake City, UT 84112, United States of America. ϟDenotes equal lead-authorship *Corresponding author
KEYWORDS: SERS, point-of-need, diagnostics, calibration-free, flow-through assays, nitrocellulose ABSTRACT: This paper presents a method for immunometric biomarker quantitation that uses standard flow-through assay reagents and obviates the need for constructing a calibration curve. The approach relies on a nitrocellulose immunoassay substrate with multiple physical addresses for analyte capture, each modified with different amounts of an analyte-specific capture antibody. As such, each address generates a distinctly different readout signal that is proportional to the analyte concentration in the sample. To establish the feasibility of this concept, equations derived from antibodyantigen binding equilibrium were first applied in modeling experiments. Next, nitrocellulose membranes with multiple capture antibody addresses were fabricated for detection of a model analyte, human Immunoglobulin G (hIgG), by a heterogeneous sandwich immunoassay using antibody-modified gold nanoparticles (AuNPs) as the immunolabel. Counting the number of colored capture addresses visible to the unassisted eye enabled semi-quantitative hIgG determination. We then demonstrated that by leveraging the localized surface plasmon resonance of the AuNPs, surface-enhanced Raman spectroscopy (SERS) can be used for quantitative readout. By comparing the SERS signal intensities from each capture address with values predicted using immunoassay equilibrium theory, the concentration of hIgG can be determined (~30% average absolute deviation) without reference to a calibration curve. This work also demonstrates the ability to manipulate the dynamic range of the assay over ~4 orders of magnitude (2 ng mL-1–10 µg mL-1). The potential prospects in applying this concept to point-of-need diagnostics are also discussed.
Introduction Medical diagnoses rely heavily on well-established bioanalytical techniques [e.g., enzyme-linked immunosorbent assays (ELISAs),1-2 polymerase chain reaction (PCR)3-4] for measuring low levels of disease biomarkers and foreign pathogens. Nevertheless, the high level of laboratory infrastructure and user skills required for implementation make it difficult to deploy these techniques in point-of-need (PON) settings where early, low-cost disease screening and diagnosis are critical.5-7 Although research has largely been directed toward medical applications, there is a growing need for PON analytical technologies in environmental testing (e.g., pesticides, heavy metals)8 and the detection of bioterrorism agents.9 Lateral flow assays (LFAs), “dip stick” tests, and flow-through assays (FTAs) are among the most suitable test formats for use in this arena.10-14 LFAs and FTAs use the selective concentration of an analyte on a solid phase membrane for subsequent labeling and readout. In this way, LFAs/FTAs can achieve concentration factors of 1000× or greater, enabling low-level analyte detection while simultaneously meeting the stringent requirements for simplicity and cost effectiveness.15-16 These designs, however, are often realized at the expense of important analytical figures of merit (e.g., sensitivity, limit of detection, reproducibility), compromising quantitative capability and diagnostic accuracy.17-18 As such, the interpretation of results from PON tests is often limited to a binary “yes or
no” determination, and guidelines from the United States Centers for Disease Control (CDC) and other public health agencies advocate for the importance of secondary testing in order to minimize incorrect diagnosis.19 As an extension of our past work on heterogeneous immunoassays20-24 and colorimetric solid-phase extraction,15,25 we and others26-35 have been investigating calibration-free approaches to quantitation in biological assays. We recently reported on the development and concept assessment of a calibration-free electrochemical detection method that exploits the predictable fluxes of an analyte under steady-state hydrodynamics as a means for analyte quantitation.26 In that work, the exact concentration of an analyte was determined using a variable width flow cell to control sample fluid velocities above an array of gold electrodes. In looking to devise concepts aimed at PON heterogeneous immunoassays, we have begun to explore a strategy in which the assay substrate (i.e., membrane) may be rationally designed so that the analyte concentration is determined from known and measurable variables. The approach detailed herein makes use of a membrane containing multiple addresses, each modified with different but predictable amounts of an analyte-specific antibody. Under identical assay conditions, distinctly different amounts of antigen bind to each capture address, leading to a signal profile that can be used for quantitation through binding equilibrium considerations.
ACS Paragon Plus Environment
Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Our approach has several parallels to the strategies put forth by Lathwal and Sikes28 and Huynh et al.27 Lathwal and Sikes developed a heteregenous immunoassay for a model analyte [mouse IgG (mIgG)] by immobilizing different amounts of anti-mIgG antibody on activated agarose surfaces.28 Eosin Y, above a critical surface concentration, serves as a photoinitiator for the polymerization of acrylate monomers, generating a stark visual (colorimetric) response to mIgG. This design provides an estimate of mIgG concentration by counting the number of spots that develop a visible color. In a similar approach, Huynh et al. developed a microfluidic system with twelve individual channels, each designed to yield an unambiguous colorimetric response (i.e., either colorless or dark purple) depending on both the analyte (cytostatin C) and antibody concentration in the channel.27 In this case, the binary assay response was achieved using acetylcholinesterase (AChE) and a known AChE inhibitor (syn-(S)TZ2PIQ-A5) that blocks AChE activity at a well-defined concentration, thereby exhibiting a sharp transition from active to inhibited enzyme. In this work, we make notable advancements in building on the strategies developed by Lathwal and Sikes28 and Huynh et al.27 First, the concept of a multi-address, semi-quantitative immunoassay is realized using a traditional LFA and FTA substrate (nitrocellulose) and immunolabel [i.e., a gold nanoparticle (AuNP)]. This approach shares similarities with recent developments in LFAs that examine, for example, the length of the colored region or the number of colored immunocapture lines for semiquantitative detection.32,34-35 To improve upon this design, we leverage the utility of AuNP immunolabels in supporting surface-enhanced Raman spectroscopy (SERS) for the purpose of quantitative assay readout. This detection modality takes advantage of the large enhancement in the incident electric field surrounding the AuNP immunolabel due to the excitement of localized surface plasmons upon laser irradiation.36 In turn, the electric field enhancement amplifies vibrational modes from AuNPimmobilized Raman-active molecules for correlation to analyte concentrations.20-22,37 Note that quantitation by SERS requires the ability to precisely control the size and shape of the nanometric asperities that give rise to plasmonic enhancement.22,38-39 This paper draws on recent work from our group on the use of a handheld, batteryoperated Raman instrument (TacticID from B&W Tek, Inc.) for the low-level detection of disease biomarkers in PON formats.23,40 Herein, we demonstrate the ability to perform quantitation via SERS readout, without constructing a calibration curve, using only common PON test reagents. The success of this approach hinges on careful control over the production of the assay substrates, which enables the concentration of a model analyte to be determined directly from the Raman signal
Page 2 of 10
through the use of heterogeneous immunoassay equilibrium equations. THEORY AND MODELING Equilibrium Considerations The digitized assay readout is predicted from two-step immunoassay equilibrium theory that includes the antigen (Ag) binding reaction and subsequent labeling step. For the antigen (Ag) capture step (Eqn 1), the amount of bound antigen (AbAg) is related to the concentrations of free antigen (Ag) and capture antibody (Abcap) by the equilibrium association constant, Ka1 (Eqn 2). Ab + Ag ↔ AbAg K =
(1) (2)
Eqns 1 and 2 assume a 1:1 stoichiometry for the antibody/antigen binding reaction. Eqn 2 can be recast into a form that reflects a heterogeneous assay, given by Eqn 3. Eqn 3 expresses the quantity of surface-bound antigenantibody complex, ΓAg, in terms of the following parameters: the initial surface concentration of capture antibody, ΓCap,0 (mol cm-2); the initial concentration of antigen, C0 (mol L-1); the volume of the antigenic solution, VAg (L); and the surface area of the membrane modified with capture antibody (i.e., the surface area of the capture address, calculated from the porosity, pore size and thickness of the membrane), A (cm2). K =
, "
(3)
Finally, Eqn 3 can be rearranged to a quadratic form (Eqn 4) to solve for the unknown variable, ΓAg.
(AK )Γ & + −C) K V − Γ,) AK − V "Γ + C) Γ,) K V = 0 (4)
The equations for the labeling step follow are derived similarly in Section 1 of the Supporting Information (SI). In this case, the antibody-conjugated ERL, AbERL, binds to the captured Ag from the previous step to form the sandwich complex that enables detection (Eqn S1 in SI). The expression for the labeling step can be written in a quadratic form (Eqn S3) that relates the surface concentration of bound, labeled Ag (ΓERL) to the volume and initial concentration of the label (VERL and CERL, respectively), and the affinity constant for tagging the captured Ag (Ka2). Note that all of these expressions neglect the impact of antibody/antigen desorption from the membrane during subsequent washing and labeling steps. The binding and capture steps are assumed to be at equilibrium. These two equations (Eqns 4 and S3) are solved sequentially to relate the surface concentration of bound, labeled Ag (ΓERL) to the initial concentration of Ag (C0). Experimentally, the SERS intensity is used as an indirect measure of ΓERL. Equilibrium Simulations
ACS Paragon Plus Environment
Page 3 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
As a starting point, we applied the principles of heterogeneous immunoassay equilibria to the extraction and subsequent labeling of a model antigen to gain insight into the experimental design of the calibration-free assay format. The formation of the sandwich complex, Ab-AgERL with surface concentration ΓERL, was modeled using a range of Ag (C0) and capture Ab (ΓCap,0) concentrations as independent variables (Figure S1). Of note, the surface concentrations of capture antibody (ΓCap,0) used in these modeling experiments were based on the expected amount of antibody deposited from a piezoelectric spotter (see below) in a set of preliminary studies (antibody concentrations of 0.025–1.350 mg mL-1). At a given C0, the labeled Ag surface concentration (ΓERL) increases with ΓCap,0 and levels off at high values for ΓCap,0. At a fixed value for ΓCap,0, there is a predictable increase in ΓERL with larger Ag concentrations. The shape of the ΓERL vs. ΓCap,0 binding curve is dependent on C0, a feature exploited herein and by others27-28,32 for designing semi-quantitative immunoassays. This approach, depicted schematically in Figure 1, relies on the concept of a “visual threshold,” which refers to an empirically determined minimum surface concentration of label needed to generate a colorimetric signal visible to the naked eye. Thus, proper selection of ΓCap,0 would permit semi-quantitation by counting the number of colored spots. By extension, the concentration resolution can be improved by increasing the number of capture addresses, each with a smaller incremental change in ΓCap,0. EXPERIMENTAL Materials
Figure 1. Conceptual layout of the immunoassay with both SERS (A) and visual (B) readout. Binding of AuNPs to the immobilized antigen produces visible spots at the capture addresses with the number of visible spots is dependent on the analyte concentration, as shown by the hypothetical results for two separate assays (B).
of immobilized αhIgG were produced on a membrane by tuning the αhIgG concentration in the spotting solution from 0.025 to 1.350 mg mL-1 (Figure 2B). The capture addresses were printed in an asymmetric pattern to aid visual recognition of the individual addresses upon assay completion. After antibody deposition, the nitrocellulose membranes were dried over desiccant (Drierite) for 14 h. The membranes were then blocked by immersion in a solution of 1% (w/v) BSA in PBS for 15 min, rinsed briefly in PBS, and dried under ambient conditions for 2 h. Flow-Based Immunoassay Procedure Immediately prior to running the assay, the αhIgGmodified nitrocellulose addresses were cut from the surrounding membrane using a 7 mm cork borer. The nitrocellulose disk was secured in a polyether ether ketone (PEEK) Luer-to-MicroTight® fitting (360 µm exit diame-
Amersham Protran Premium nitrocellulose membranes with a nominal pore size and thickness of 0.45 µm and 120 µm, respectively, were from GE Life Sciences (Pittsburgh, PA). Human Immunoglobulin G (hIgG) and goat antiHuman IgG (αhIgG) were from Invitrogen (Carlsbad, CA). Bovine serum albumin (BSA) and acetonitrile (ACN; >99.5%) were from Sigma-Aldrich (St. Louis, MO). Sodium chloride, borate buffer (BB; pH 8.5) packs, and modified Dulbecco’s phosphate-buffed saline (PBS; pH 7.4) packs were from ThermoFisher Scientific (Wilmington, DE). Gold nanoparticles (AuNPs; ~60 nm diameter) were from BBI Solutions (Cardiff, UK). The Raman reporter molecule 5,5'-dithiobis(succinimidyl-2-nitrobenzoate) (DSNB) was synthesized according to a published protocol.24 Water used to prepare all buffers was purified to a final resistivity of 18.2 MΩ-cm using a Barnstead Nanopure water purification system (ThermoFisher Scientific). Fabrication of Capture Membranes The nitrocellulose membranes were modified to capture hIgG by physisorption of the corresponding αhIgG antibody. A piezoelectric liquid dispenser (Scienion sciFLEXArrayer S3; Berlin, Germany) was used to deposit individual droplets (~350 pL) of capture antibody in PBS. The capture antibody droplets were dispensed in a 10×10 grid (50 µm center-to-center spacing between droplets) to yield a single 0.5×0.5 mm2 capture address, as depicted in Figure 2A. Several (6–7) addresses with different amounts
Figure 2. Schematic overview of the capture substrate fabrication and membrane holder for flow-through assay. (A) An illustration of the capture Ab deposition process. Individual droplets of capture Ab solution (~350 pL, depicted as blue circles) are deposited at the positions denoted by the black dots, separated by a center-tocenter distance of 50 µm. (B) A schematic of a typical assay substrate, illustrating the asymmetric layout of the multiple capture addresses. (C) A photograph of the Luer-to-MicroTight® fitting used to house the nitrocellulose disk for the assay. (D) A cross-sectional illustration of the enclosure shown in C.
ACS Paragon Plus Environment
Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ter; Idex Health and Science, Oak Harbor, WA) with the αhIgG addresses facing the Luer connection, as illustrated in Figure 2D. For each assay, 1.0 mL of hIgG in assay buffer (1% w/v BSA in PBS) was pumped through the fitting/membrane at 0.5 mL min-1 using a Harvard Apparatus PHD2000 syringe pump. The membranes were subsequently rinsed with 0.5 mL assay buffer at a flow rate of 0.5 mL min-1, removed from the plastic housing, and allowed to dry for 1 min on the bench top. Bound antigen was labeled by incubating the membranes with a 200 µL solution of antibody-modified AuNPs (i.e., extrinsic Raman labels; ERLs) on a rotator for 45 min. The ERL synthesis procedure has been reported previously.24 A detailed synthesis procedure is provided in Supporting Information (I.2). The membranes were then rinsed with twice with 500 µL of assay buffer (5 min each) to remove loosely adhered ERLs, then dried overnight over Drierite prior to Raman measurements. A schematic illustration of the immunoassay is shown in Figure 3. Raman Instrumentation and Measurements Raman spectra were measured using a ThermoScientific DXR Raman microscope. Radiation from a HeNe laser (632.8 nm) was coupled through a 5x microscope objective to form an elliptical spot (foci axes 16×27 μm) as a means to interrogate the membrane surfaces. Unless stated otherwise, Raman spectra were collected as the average of two 1.0 s integrations at a laser power of 4.9 mW (i.e., power density of 3.6 μW μm-2) at the sample surface. The laser power was measured periodically and was found to vary by
