Ultratrace level determination and quantitative analysis of kidney injury biomarkers in patient samples attained by zinc oxide nanorods

Abstract

Determining ultratrace amounts of protein biomarkers in patient samples in a straightforward and quantitative manner is extremely important for early disease diagnosis and treatment. Here, we successfully demonstrate the novel use of zinc oxide nanorods (ZnO NRs) in the ultrasensitive and quantitative detection of two acute kidney injury (AKI)-related protein biomarkers, tumor necrosis factor (TNF)-α and interleukin (IL)-8, directly from patient samples. We first validate the ZnO NRs-based IL-8 results via comparison with those obtained from using a conventional enzyme-linked immunosorbent method in samples from 38 individuals. We further assess the full detection capability of the ZnO NRs-based technique by quantifying TNF-α, whose levels in human urine are often below the detection limits of conventional methods. Using the ZnO NR platforms, we determine the TNF-α concentrations of all 46 patient samples tested, down to the fg per mL level. Subsequently, we screen for TNF-α levels in approximately 50 additional samples collected from different patient groups in order to demonstrate a potential use of the ZnO NRs-based assay in assessing cytokine levels useful for further clinical monitoring. Our research efforts demonstrate that ZnO NRs can be straightforwardly employed in the rapid, ultrasensitive, quantitative, and simultaneous detection of multiple AKI-related biomarkers directly in patient urine samples, providing an unparalleled detection capability beyond those of conventional analysis methods. Additional key advantages of the ZnO NRs-based approach include a fast detection speed, low-volume assay condition, multiplexing ability, and easy automation/integration capability to existing fluorescence instrumentation. Therefore, we anticipate that our ZnO NRs-based detection method will be highly beneficial for overcoming the frequent challenges in early biomarker development and treatment assessment, pertaining to the facile and ultrasensitive quantification of hard-to-trace biomolecules.

Figure 1

(A) The bright field optical microscope image of a ZnO NR square array displays the typical NR platforms used in the chemokine and cytokine assays. The ZnO NRs were vertically grown on a Si substrate prepatterned with microcontact-printed catalysts into a square array. The view frame corresponds to 0.5 × 0.5 mm² in size. (B) Overall schematics of the duplexed detection of IL-8 and TNF-α using the ZnO NR platforms are presented. The simultaneous sandwich assay scheme involved primary as well as labelled secondary antibodies targeting either known concentrations of IL-8 and TNF-α to establish calibration curves or patient urine samples to determine the unknown IL-8 and TNF-α amounts in subject individuals.

Figure 2

(A) The two fluorescence panels are representative emission data collected after performing the sandwich assays on the ZnO NRs against the patient samples and simultaneously measuring the protein concentrations. The left and right panels are the fluorescence signals corresponding to TNF-α and IL-8, respectively, from the patient No 24. (B) The SEM panels display the morphology and dimensions of the square platforms of vertical ZnO NRs used in the fluorescence detection of TNF-α and IL-8. The inset in the left image, indicated as F, shows the fluorescence panel of the as-synthesized ZnO NR array probed with λ_ex = 450–490 nm. ZnO NRs exhibit no background fluorescence emission upon excitation at the wavelength ranges used to detect the proteins. (C) Exemplar fluorescence intensity plots from selected patients are presented to show the different amounts of TNF-α and IL-8 present in the urine. (D) The bar graphs display the variability in the fluorescence signals measured from repeated TNF-α and IL-8 assays of the same sample performed via the ZnO NRs-based method. In order to evaluate the inter- and intra-assay variability, TNF-α and IL-8 assays for the two selected patients (Nos 17 & 19) were carried out on three different ZnO NR arrays (data shown under *) as well as on the same ZnO NR array five consecutive times (data provided with **), respectively.

Figure 3

(A and B) Calibration curves of IL-8 and TNF-α produced by the ELISA method are displayed in panel (A) and (B), respectively. Absorbance readings from IL-8 and TNF-α standards were plotted against its concentration. The solid lines in the graphs are the 4-parameter fitting curves yielding the reported values of adjusted coefficient of determination (R²). (C and D) Calibration curves for IL-8 and TNF-α established on the ZnO NR detection platform are shown in panel (C) and (D), respectively. The standard curves were generated by evaluating fluorescence intensities measured from known amounts of each protein. The solid black line in each graph corresponds to a linear fit through the data points with the indicated R² value. The dotted black lines indicate 95% confidence intervals.

Figure 4

(A) The 3D bar graphs display the IL-8 concentrations in the patients’ samples determined by using the ELISA- and ZnO NRs-based platforms. (B) In order to show the lower range data more clearly, the 3D bar graphs of the patients’ IL-8 values from the two methods are compared by using the upper concentration limit of 400 pg/mL. The truncated bars indicate that the IL-8 levels exceed the upper limit of the 3D graph.

Figure 5

(A) The ELISA and ZnO NR assay results are directly compared from the same set of patients by plotting each patient’s IL-8 reading in pg/mL evaluated by the different methods along the x (ELISA) and y (ZnO NRs) axis. The dashed red line is the linear fit through the data. The ELISA and ZnO NRs assay results fall on or near the black line of y = x, indicating good agreement between the patient IL-8 values established by the two assay techniques. (B) The histogram distributions chart patient counts versus the differences in the evaluated IL-8 concentrations between the two assay techniques for the (ZnO NRs-ELISA) range of less than 40 pg/mL. The majority of the patients’ IL-8 readings fall within the range of ±2.5 pg/mL from the reading of each technique. (C) The IL-8 concentrations in pg/mL determined from the ZnO NR and ELISA assays are compared by plotting the differences between the two techniques, (ZnO NRs-ELISA), on the y-axis and the means of the two techniques, (ZnO NRs + ELISA)/2, on the x-axis. The evaluated data lie close to the black line of y = 0 whose trace corresponds to equivalent readings of IL-8 concentrations in the two assays from the same patients. (D) The scatter plot shown in (C) is rescaled to clearly present the lower range data.

Figure 6

(A) The 3D bar graphs display the TNF-α concentrations in the urine samples measured by the ELISA- and ZnO NRs-based platforms for comparison. The grey regions in the ELISA row correspond to missing concentration data, indicating that the TNF-α levels in the samples were below the ELISA DL of 5.5 pg/mL. In contrast, ZnO NRs were able to measure TNF-α concentrations of all 46 patients. The magnifier signs inserted in the ZnO NRs row correspond to the patients that belong to the grey area of the ELISA-based assay, and the bar graphs of these patients are shown separately in (B and C) for clarity. (B and C) The zoomed-in 3D bar graphs are the missing TNF-α concentrations that were revealed by the ZnO NRs-based assay. The upper limits of the vertical ranges in (B) and (C) are adjusted to 2 pg/mL and 350 fg/mL, respectively, in order to show the variations in the TNF-α concentrations between patients more clearly. The truncated bars indicate that their TNF-α concentrations exceed the upper limit of the 3D graph.

Figure 7

The scatter plots show the distribution patterns of TNF-α concentrations for subjects in the three different categories of no AKI, ICU AKI, and Day 1 AKI. The ZnO NR assay enables highly sensitive TNF-α detection of all patients. In addition, the characteristic distribution of the TNF-α concentration values associated with each AKI category can be identified, as indicated by the green and red circles in the graph.