In this study, PCT was used as a model protein antigen for the comparative study of LFIA with nanospheres and hierarchical structures of nanogold labels (nanopopcorns and nanostars). PCT is an important biomarker that exhibits greater specificity than other pro-inflammatory markers (e.g., cytokines) in identifying patients with sepsis and may be used in the diagnosis of bacterial infections. A serum PCT level higher than 10 ng mL−1 may be indicative of septic shock. PCT concentrations in sepsis reach values between 2 and 10 ng mL−1 PCT values between 0.15 and 2 ng mL−1 do not exclude an infection because localized infections (without systemic signs) may be associated with such low levels.
To detect PCT by LFIA, the sandwich scheme of analysis was applied (Fig. 1). The specific antibodies were immobilized in the test zone of the analytical membrane, and the conjugate pad was impregnated with the paired detection antibodies labeled with gold nanoparticles. The application of the solution containing the antigen to the test strip (sample pad) resulted in the binding of paired antibodies labeled with gold nanoparticles to PCT in the sample to form labeled antigen–antibody complex. This complex moved by capillary forces through the area containing the test zone, wherein the labeled complex attached to the immobilized specific antibodies and formed a sandwich complex. This sandwich complex may be observed as a colored band, and the color intensity of the band was directly proportional to PCT concentration in the samples.
In this study, seven conjugates, namely antibodies labeled with GNPNs, GNSTs, and GNSs of different sizes were prepared and used in LFIA for PCT detection. To compare the effect of different labels on the assay sensitivity, LFIAs based on different gold nanoparticles and silver enhancement procedure were performed.
Characterization of GNSs, GNPNs, and GNSTs
The color of gold nanoparticles strongly depends on their sizes and shapes. Figure 2 displays images of colloidal solutions and TEM. The calculated average diameters of GNSs were 20.0 ± 1, 35.1 ± 2.3, 50.4 ± 1.6, 70.2 ± 1.8, and 100.2 ± 2.0 nm, while those for GNPNs and GNSTs were 100.1 ± 5.7 and 64.3 ± 3.0 nm, respectively. UV–Vis spectra demonstrated maximal absorbance at 520 (20 nm), 530 (35 nm), 537 (50 nm), 547.5 (70 nm), and 582.7 nm (100 nm) wavelengths for GNSs and 683 and 633 nm for GNPNs and GNSTs (Fig. 3), respectively. All the obtained samples exhibited good colloidal stability and homogeneity in terms of composition and size.
Optimization of LFIA Conditions
All gold nanoparticles (GNSs, GNPNs, and GNSTs) were used as labels in LFIA, and the resulting antibody-nanoparticle conjugates were impregnated in the conjugate pad. Optimization of the assay system included the choice of specific antibodies, antibody-gold nanoparticle conjugate, and concentration of immunoreagents in the test and control lines. The combination of immunoreagents and their concentrations were chosen based on low background and minimum PCT detection limit. To choose a pair of antibodies that meets the requirements of LFIA, a few clones of mAb and pAb were used and the calibration curves for all combinations were obtained for LFIA with 20 nm GNSs (data not shown). The highest assay sensitivity with a PCT detection limit of 2 ng mL−1 was observed for the combination of immobilized mAb against calcitonin and labeled pAb against PCT (Fig. 4).
The concentration of antibodies used to prepare a stable immunoprobe with gold nanoparticles was 15 µg mL−1 for GNSs (20 and 35 nm) and GNPNs, 20 µg mL−1 for GNSs (50 nm), and 40 µg mL−1 for GNSs (70 and 100 nm) and GNSTs. Thus, nanospheres of larger size demand higher antibody concentrations during the preparation of stable conjugates and led to the consumption of specific reagents. In contrast, hierarchical 100 nm GNPNs require about half the amount of antibodies than GNSs and multi-branched GNSTs of the same size. In order to choose the optimal dilution of gold nanoparticle–antibody conjugates for LFIA, the optical density of solutions at an appropriate wavelength was varied from 0.5 to 4. Sufficient color intensity of the test line with low background was achieved by the spreading of pAb-GNS with optical density of 2, pAb-GNPN, and pAb-GNST with optical density of 4. The optimal concentration of the immobilized mAb and protein A on the test and control line, respectively, was 0.5 mg mL−1.
Comparison of LFIA Based on GNSs, GNPNs, and GNSTs
According to the chosen optimal conditions, the calibration curves of LFIA based on different nanoparticles were obtained for the accurate comparison of the effects of labels on the sensitivity of analysis (Fig. 5a). It was found that the sensitivity of LFIA improved with an increase in the size of GNSs (up to 40 nm), consistent with the previous data [12]. On the other hand, the use of hierarchical large gold nanoparticles (GNPNs and GNSTs) in LFIA allowed the detection of lower PCT concentrations. Of all seven obtained gold nanoparticles, GNPN-based LFIA exhibited linearity over the range of 0.5–10 ng mL−1 with the best limit of detection at 0.1 ng mL−1 concentration, which was five times better than the sensitivity of the conventional LFIA with 20 nm GNSs. The comparison between large GNS-and GNPN-based LFIA showed that the shape, rather than the particle size, affected the sensitivity of the analysis. The sandwich complex formed after the completion of the assay was seen as a band with color change from red to gray, owing to different sizes and shapes of gold nanoparticles. The results of LFIA based on GNSs with a size of 70 and 100 nm were difficult to interpret because of the pale color of the label in the test zone of the strip (Fig. 5b). On the other hand, the hierarchical gold nanoparticles with a size around 100 nm displayed a contrasting color, which may be visually detected on the test zone of the test strip. This may be attributed to the red shift in the surface plasmon resonance peak for hierarchical-structured gold nanoparticles. For instance, 100 nm GNSs had a surface plasmon resonance peak at 582.7 nm, while 100 nm GNPNs displayed a surface plasmon resonance peak at 683 nm (Fig. 3). Multi-branched GNSTs have found wide applications in photoacoustic imaging [32], photothermal therapy [33], and biosensing [34], given their unique light scattering and absorption properties. The characteristics of GNST-based LFIA were comparable with those of conventional LFIA based on GNSs with a size of less than 50 nm. This result may be associated with the steric hindrance that occurs as a result of many tips and the uneven distribution of antibodies on the surface of GNSTs. As a consequence, the adsorption of the binding complex of antigen-labeled antibodies on the surface of GNSTs may be difficult.
One of the effective approaches for the amplification of line intensities in the test zone of the strips includes the use of silver-enhanced labeling method. The procedure of silver-enhanced LFIA for the detection of PCT was similar to that of the conventional LFIA. After the formation of a colored red band in the test zone of the membrane associated with sandwich complex, the silver-enhancing mixture was applied. The color of the test and control zone changed from red to black. A scanned image of test strips after silver enhancement and typical calibration curves of the silver-enhanced LFIA and 20 nm GNS-based LFIA are presented in Fig. 6. The treatment of the test strip with silver-enhancing mixture resulted in the detection of PCT at 0.05 ng mL−1 concentration, which corresponded to a tenfold increase in the sensitivity. The results obtained are comparable with the recently reported study of silver-enhanced LFIA for prostate specific antigen [35].
Thus, the use of multi-branched GNPNs instead of GNSs as a colored label may improve the sensitivity of LFIA for PCT detection by approximately fivefold. Despite the tenfold increase in the sensitivity with the silver enhancement procedure, this approach requires additional reagents as well as operation steps of analysis and takes more time for completion as compared with the conventional GNS-based LFIA (20 vs. 10 min). In addition, the silver-enhanced method is difficult to apply outside the laboratory. For clinical purpose, the serum of patients is applied to the sample pad of LFIA test strip. The analysis is semi-quantitative, and the result is observed by the presence or absence of a colored line in the test zone of the strip. The analysis time is within 10–15 min, and the test may be performed outside the laboratory, thereby meeting the requirements of point-of-care testing. Preliminary studies have revealed the feasibility of LFIA for PCT detection in serum samples. Further studies will be focused on PCT LFIA validation and application.