Development of a chemiluminescent enzyme immunoassay to detect cadmium in cereal samples

ABSTRACT

In this study, we develop a sensitive indirect competitive chemiluminescent enzyme immunoassay (ic-CLEIA) to detect cadmium (Cd) in cereals. A luminol-hydrogen peroxide chemiluminescence system catalysed by horseradish peroxidase (HRP) was used as the signal detecting system. The method showed a half-maximal inhibition concentration (IC₅₀) of 0.5 ng/mL and a limit of determination (LOD) of 0.06 ng/mL. The assay indicated an effective linear range from 0.16 to 1.56 ng/mL. There were no obvious cross-reactivities with other metal ions, including Hg²⁺, Pb²⁺, As²⁺, Ag⁺, Fe³⁺, Cu²⁺, Cr²⁺, Au³⁺, Ca²⁺, Mg²⁺, Mn²⁺ and Zn²⁺. Recovery rates ranged from 79.2% to 120.0% for rice, maize, wheat, brown rice and glutinous rice samples. Twenty-five practical cereal samples were analysed, and the comparative analysis of the results showed a strong correlation (r = 0.984) between the CLEIA and the GFAAS methods. The results indicated that the ic-CLEIA developed was acceptable to screen for cadmium in cereal samples.

GRAPHICAL ABSTRACT

KEYWORDS:

• Cadmium
• chemiluminescence immunoassay
• cereals
• analysis

1. Introduction

Cadmium (Cd), with a white and glossy surface, is a toxic heavy metal and an environmental pollutant found around the world (Ran, Wang, Wang, Zhang, & Zhang, 2016). It easily accumulates in the human body and has a biological half-life of longer than 10 years (Jin et al., 2013). In addition, it is known to be a mutagenic metal able to induce apoptosis and oxidative stress (Ambily, Olivier, Karen, Emmy, & Ann, 2013). Initially, it was thought that only high, long-time exposure to Cd would harm human health. However, numerous data show that even low exposure to Cd can be damaging to the health of those exposed (Wallin et al., 2016). Cd exposure is reported to decrease bone density, increase the DNA methylation status of the liver and damage the kidneys and even the whole body (Callan, Devine, Qi, Ng, & Hinwood, 2015; Garcia & Corredor, 2004; Pierron, Baillon, Sow, Gotreau, & Gonzalez, 2014). Therefore, Cd contamination has posed an increasingly greater threat to humans in recent years because of its persistence, bioaccumulation and high toxicity (Singh & Kumar, 2017).

However, the development of technological progress and industrialization has led to the growing use of Cd, resulting in its widespread distribution (Li et al., 2014; Zhu et al., 2016). Such pollution is mainly due to anthropogenic activities, associated with mining activities, landfill leachate or industrial and domestic effluents, resulting in high levels of residual Cd in the surrounding soils, dust and water (Abdullah, 2013; Candeias et al., 2014; Singh et al., 2008). In some mine area, the maximum Cd value in rhizosphere soils, road dust and vegetables could be up to 3.4, 19.40 and 0.4 mg/kg, respectively (Candeias et al., 2014). Plants, especially rice grown on a contaminated field, can accumulate high levels of Cd, and if the level of Cd in the soil is higher, the concentration of Cd in rice can increase significantly (Nuapia, Chimuka, & Cukrowska, 2018). The way that metals accumulate in plants plays an important role in the transfer of Cd into food chains, since the main sources of ingested Cd in a diet are staple foods such as wheat, rice and potatoes, which provide approximately 40–50% of the average total Cd intake (Chen et al., 2018; Huang, Zhou, Sun, & Zhao, 2008; Orisakwe, Nduka, Amadi, Dike, & Bede, 2012). In addition, the daily intake of Cd through the diet varies greatly from differing countries depending on their levels of pollution. The highest intake of Cd is noted in Cd-polluted areas in Thailand, reaching 188–224 μg a day (Satarug et al., 2013), while the lowest is recorded in Sweden, amounting to 10 μg per day (Sand & Becker, 2012). However, according to the world health organization (WHO) and the Food and Agriculture Organization of the United Nations (FAO), the Provisional Tolerable Monthly Intake (PTMI) for Cd intake has been re-established to be 25 μg/kg bw (FAO/WHO, 2010).

It is necessary to develop analytical methods to monitor Cd in cereals that are consumed by humans daily. Currently, the Cd in some samples is mainly analysed by instrumental methods, such as graphite furnace atomic absorption spectroscopy (GFAAS) (Ahmadi, Khanmohammadi, Jafarpour, & Eftekhari, 2008), atomic fluorescence spectrometry (AFS) (Yang, Chi, Wang, & Zhang, 2015), inductively coupled plasma mass spectrometry (ICP-MS) (Nardi et al., 2009) and inductively coupled plasma atomic emission spectrometry (ICP-AES) (Sun, Zhang, Zhao, Wang, & Li, 2016). Other techniques such as the screen-printed carbon nanotubes electrodes assay (Simpson et al., 2018) and electrochemical impedance spectroscopy (Petovar, Xhanari, & Finšgar, 2018) are also mostly used in the laboratory. These techniques rely on sophisticated instruments and skilled technicians. In addition, the pretreatment of the samples is time-consuming, and the digestion process often requires large amounts of strong acids that can harm the environment. By contrast, immunoassays, primarily relying on the specific reaction between the antibody and the antigen, exhibit some unique characteristics such as being rapid, simple to perform, cost-effective and friendly to the environment, which enable them to satisfy the demands of high-throughput and on-site screens (He et al., 2011; Zhou et al., 2011).

Arai, Nakamura, Yoshida, and Tawarada (2016), Song, Zou, Zhu, Liu, and Kuang (2018) and Xing et al. (2014) developed immunochromatographic strips for semi-quantitative detection of cadmium. Johnson (1999) reported a fluorescence polarization immunoassay for cadmium. Until now, several ELISA methods for Cd determination have been well developed for environment samples. Zhan, Xi, and Zhou (2015) obtained a cadmium-chelate specific polyclonal antibody and developed an indirect competitive ELISA to measure Cd in water. The assay exhibited cross-reactivity with Hg²⁺ (5.8%). He, Liu, Yang, and Sun (2009) and Zhu et al. (2007) developed indirect and direct competitive ELISA respectively, to detect Cd in aqueous samples. There are only a few cases of the application of ELISA to detect Cd in grains (Abe et al., 2011, 2014; Liu et al., 2009) with the limit of detection (LOD) among 0.20 and 5 ng/mL. To date no report on CLEIA for Cd detection is available to the best of our knowledge.

A chemiluminescent immunoassay (CLEIA) using a chemiluminescent substrate instead of traditional chromogenic substrates provides greater sensitivity by 2–3 orders of magnitude (Botchkareva et al., 2003). In addition, due to more rapid colorimetric procedures, the time for the CLEIA can be reduced by 20–30 min compared with ELISA, which makes it more suitable for rapid analysis. In this study, a high-sensitive indirect competitive CLEIA was developed for the trace determination of Cd in cereals. Instead of strong acid digestion, this assay extracts Cd with diluted HNO₃ in the pretreatment step, which makes it safer, more rapid and convenient.

2. Material and methods

2.1. Materials and reagents

Cd, Hg, Pb, As, Ag, Fe, Cu, and Cr (1000 μg/mL in 2% HNO₃) were purchased from Aladdin Industrial Co. (Shanghai, China). Ovalbumin (OVA), 2-[4-(2-Hydroxyethyl)-1-piperazinyl] ethanesulfonic acid (HEPES) and horseradish peroxidase (HRP)-labeled goat anti-mouse IgG were purchased from Sigma-Aldrich (St. Louis, MO, USA). White, opaque, high-binding 96-well polystyrene microtiter plates were purchased from Shenzhen Jincanhua Industrial Co. Ltd. (Shenzhen, China). Anti-Cd-EDTA monoclonal antibody (mAb) was obtained from Wuxi Jieshengjiekang Biotechnology Co. Ltd. (Wuxi, China). Isothiocyanatobenzylethylenediamine-N,N,N′,N′-tetraacetic acid (ITCBE) was purchased from Dojindao Laboratories (Shanghai, China).The Super Signal West Pico CL substrate (stable peroxide buffer, A; luminol/enhancer, B) was obtained from Guangzhou Wanlian Biotechnology Co. Ltd. (Guangzhou, China). All other reagents were purchased from Qixiang Trade Co. Ltd. (Guangzhou, China). Microtiter plates were washed in a DEM-3 microplate washer (Tuopu, Beijing, China). A Spectra Max M5 microplate reader was purchased from Molecular Devices (CA, USA). Atomic absorption spectroscopy equipped with the graphite tube was used for GFAAS verification (ZEEnit 650P, Analytik Jena AG, Germany).

2.2. Buffers

Carbonate buffer (0.1 mol/L, pH 9.6); phosphate buffered saline (PBS, 0.1 mol/L, pH 7.4 or pH 9.0); PBST (PBS containing 0.05% Tween-20, pH 7.4); HBS buffer (0.1 mol/L HEPES, pH 7.4); HBS-EDTA (HBS containing 0.1 μmol/L EDTA, pH 7.4).

2.3. Preparation of the coating antigen

Two milligrams of ITCBE was dissolved in 200 μL of DMSO; 20 mg of OVA was dissolved in 500 μL of PBS (0.1 mol/L, pH 9.0), and the ITCBE solution was then slowly added dropwise into the OVA solution. After stirring at room temperature for 24 h, the mixture was dialysed against PBS (0.1 mol/L, pH 7.4) for 2 days. After this, 1.10 mL of Cd²⁺ solution (1000 μg/mL) was added to the ITCBE-OVA solution, and the pH of the mixture was adjusted to 7.0 before stirring at room temperature for 3 h. The small unconjugated molecules were removed by buffer dialysing with PBS (0.1 mol/L, pH 7.4) for 2 days (Blake et al., 1996; Song et al., 2018).

2.4. Procedure of CLEIA

The white, opaque 96-well polystyrene microtiter plate was coated with 100 μL/well of Cd-ITCBE-OVA in 0.1 mol/L carbonate buffer (pH 9.6). After incubation at 37°C overnight, the plate was washed with washing buffer (PBS containing 2% Tween-20, pH 7.4) twice and then blocked with 120 μL of 2% BSA in PBS (0.1 mol/L, pH 7.4) each well for 3 h at 37°C. After these steps, 50 μL of the sample or Cd standard solution (HBS-EDTA buffer, containing 0.1 μmol/L of HBS, pH 7.4) was added to each well. For the competition reactions, 50 μL of purified mAb was immediately added to each well, and the plate was incubated at 37°C for 30 min. Then the plate was washed five times, and 100 μL/well HRP-labeled goat anti-mouse IgG was added, and the plate was incubated at 37°C for 30 min. Then 100 μL/well of supersignal chemiluminescence substrate solution (A was mixed with an equivalent volume of B) was added and measured using a Spectra Max M5 microplate reader immediately. The relative luminescent unit (RLU) of each well was determined and analysed individually (Yu et al., 2016).

2.5. Optimization of the assay conditions and standard curve

Several experimental parameters, including the concentrations of the coating antigen and antibody, working buffer systems, concentration of EDTA, reaction time and pH were optimized to improve the CLEIA sensitivity. For each condition, standard curve was established, and the maximum RLU (RLU_max), 50% decrease point in RLU_max (IC₅₀) and RLU_max/IC₅₀ ratio were obtained from the curves, based on which the appropriate value of each parameter was determined. Under the optimized working conditions, calibration curves were fitted using 11 concentration levels of standard solution ranging from 0.005 to 333.33 ng/mL, each measured in triplicate. Standard curves were obtained by plotting the RLU/RLU₀ values against the logarithm of analyte concentration, where RLU and RLU₀ were the values of the wells with and without Cd-EDTA solution, respectively. The results were calculated by using Origin version 8.0 (OriginLab, Northampton, MA, USA).

2.6. Determination of cross-reactivity

The cross-reactivity (CR) for several metal ions including Cd²⁺, Hg²⁺, Pb²⁺, As²⁺, Ag⁺, Fe³⁺, Cu²⁺, Cr²⁺, Au³⁺, Ca²⁺, Mg²⁺, Mn²⁺ and Zn²⁺ was tested using CLEIA. Metal ions with different concentrations that ranged from 0.01 to 2000 ng/mL were mixed with 0.1 μmmol/L EDTA. The CR values were calculated as CR (%) = [IC₅₀ (Cd-EDTA) / IC₅₀ (other metal chelate)] × 100% (Cui, Zhang, Huang, Yu, & Peng, 2011).

2.7. Sample preparation

Twenty-five samples were obtained from the local supermarket including rice (n = 5), wheat (n = 5), maize (n = 5), brown rice (n = 5) and glutinous rice (n = 5). The samples were ground and weighed, and 1.0–1.5 g samples were placed in 15 mL polypropylene centrifuge tubes. Then 4 mL of 2 mol/L ultrapure nitric acid was added. After vortexing, the samples were precipitated for 20 min to be completely extracted. After centrifugation at 4000 rpm for 15 min, the upper layer was further centrifuged at 13,000 rpm for 10 min. Then, 2 mL of supernatant was accurately taken, and the pH was adjusted to 7.4 with 4 mol/L NaOH, finally bringing the volume to 4 mL. The solution after pretreatment was analysed using this developed CLEIA and GFAAS at a wavelength of 228.8 nm (Li et al., 2010).

2.8. Recovery

Cd was added to the ground sample solution to produce final concentrations of 37, 200, and 375 ng/g. All of the spiked samples were prepared as described in the “2.7. Sample preparation” section and detected with CLEIA. The recovery (%) was calculated using the following formula: recovery (%) = [(quantity measured)] / (quantity spiked)] × 100%.

3. Results and discussion

3.1. Optimization of CLEIA

3.1.1. Optimization of coating antigen concentration and antibody concentration

For coating antigen concentration and antibody concentration, a preliminary checkerboard assay was performed to assess the optimal coating concentration (ranging from 62.5 to 500 ng/mL) and the antibody concentration (ranging from 39.5 to 400 ng/mL), under which the IC₅₀ was the lowest, and the RLU_max/IC₅₀ was the highest, the result indicated that the pairing of coating concentration (250 ng/mL) and antibody concentration (110 ng/mL) provided the highest sensitivity.

3.1.2. Optimization of working buffer systems

The reaction of antigen and antibody is characterized by weak intermolecular bonds that can be easily affected by reaction systems (Xing et al., 2009). Four buffer systems (PBS, PBST, HBS and Tris-HCl, 0.1 mol/L, pH 7.4) were used for the potential working buffer. The results indicated that there was no significant difference between HBS (1.08 ng/mL) and Tris-HCl (1.27 ng/mL) for IC₅₀, but the higher RLU_max/IC₅₀ was obtained using HBS (Figure 1(a)). Therefore, the HBS buffer system is more suitable for this antibody–antigen interaction. PBS and PBST were considered to not be useful for these experiments because of their high IC₅₀.

Figure 1. Optimized parameters on the CLEIA performance for Cd. (a) The effect of different working buffer for dilution. (b) The effect of different concentrations of EDTA. (c) The effect of different inhibiting reaction time. (d) The effect of different pH of buffer solution.

3.1.3. Optimization of concentration of EDTA

Since the antibody could only specifically recognize Cd-EDTA chelate instead of free Cd, it was necessary to add EDTA to the sample prior to measurement (Sasaki et al., 2008), thus the concentration of EDTA was an important and necessary influencing factor among the metal-CLEIA procedures. The Cd²⁺ was diluted in HBS that contained EDTA at concentrations of 0.0001, 0.001, 0.01, 0.1 and 1 mmol/L. The mixture of Cd²⁺ and EDTA was added to the wells. With the increasing of EDTA concentration, the IC₅₀ increased from 0.26 to 0.97 ng/mL. This effect was probably due to the inhibitory effect of EDTA on the enzyme activity. The highest level of sensitivity was obtained with 0.0001 mmol/L EDTA in HBS. The result indicated that 0.0001 mmol/L EDTA was sufficient enough to chelate a low dose of Cd²⁺ for measurement (Figure 1(b)).

3.1.4. Optimization of inhibition reaction time

Interaction of antigen and antibody is a dynamic equilibrium reaction and the reaction time is of great importance for the result. Different inhibition reaction times (20, 30, 40 and 50 min) were compared. The 30 min incubation was enough for the antigen–antibody equilibrium interaction, and it had the lowest IC₅₀, which was 0.48 ng/mL (Figure 1(c)).

3.1.5. Optimization of reaction buffer pH

Generally, the methods for extracting Cd involve the use of acid solutions, mostly nitric acid, since Cd only dissolves in acidic solutions. However, extreme acidity or alkalinity would damage the antibody, leading to false positives or negatives. Therefore, it was necessary to examine the pH tolerance of the CLEIA. As shown in Figure 1(d), the IC₅₀ decreased with increasing HBS pH values from the range of 4.4 to 8.4 and then began to increase. At the same time, a gradual increase in RLU_max/IC₅₀ of the assay was observed with increasing pH values from 4.4 to 7.4, with the RLU_max/IC₅₀ reaching its low at pH 8.4. Considering the multiple factors affecting the CLEIA method, pH 7.4 for HBS was chosen to be the optimal pH for the reaction system. Additional studies of the pH indicated that when the pH of the assay was below 4.4 or above 9.4, no effective curve was available. This confirms that it was a critical step to adjust the pH before detection.

3.2. Standard curves

Under the optimized parameters, the standard curve was obtained using 11 concentration levels of standard solution and was plotted in the form of RLU/RLU₀ against the logarithm of the concentrations (Figure 2). The established CLEIA system for Cd detection is highly sensitive with an IC₅₀ of 0.50 ng/mL and a LOD of 0.06 ng/mL.

Figure 2. Standard curve of the CLEIA procedure under the optimized parameters.

3.3. Cross-reactivity and stability of CLEIA

The specificity of the developed CLEIA was examined by testing the cross-reactivity rates of several metal ions (Hg²⁺, Pb²⁺, As²⁺, Ag⁺, Fe³⁺, Cu²⁺, Cr²⁺, Au³⁺, Ca²⁺, Mg²⁺, Mn²⁺ and Zn²⁺). As shown in Table 1, the antibody exhibited negligible cross-reactivity with any of the metals tested (CR% < 0.1%), which demonstrated that the antibody was highly specific to Cd²⁺. In addition, EDTA and the EDTA-based chelates combined with the metal ions mentioned above also showed no reaction with the antibody (results not shown).

Table 1. Cross-reactivity with different metal ions and EDTA.

Metal	IC50 (ng/mL)	Cross-reactivity (%)
Cd	0.5	100
Hg	>2000	<0.1
Pb	>2000	<0.1
As	>2000	<0.1
Ag	>2000	<0.1
Fe	>2000	<0.1
Cu	>2000	<0.1
Cr	>2000	<0.1
Au	>2000	<0.1
Ca	>2000	<0.1
Mg	>2000	<0.1
Mn	>2000	<0.1
Zn	>2000	<0.1
EDTA	>2000	<0.1

3.4. Sample pretreatment and matrix effects elimination

To develop an immunoassay, it was of great importance to deal with the matrix effects. The acid residues would cause non-specific interference with some types of proteins and pigments. In this study, the dilution method was used to eliminate the matrix effects. After the pretreatment of rice samples, the supernatants were diluted 6-, 10-, 20- and 30-fold. As shown in Figure 3, the matrix effects were reduced to acceptable levels when the samples were diluted 30-fold. The diagram shows that the 30-fold dilution supernatant of the rice sample had an IC₅₀ of 0.44 ng/mL, a linear range of 0.26–3.23 ng/mL and an LOD (IC₁₀) of 0.18 ng/mL, similar to the HBS buffer. This trend was also apparent among the maize, wheat, brown rice and glutinous rice samples when the supernatant was diluted 30-fold (Figure 3). Therefore, it was apparent that sample preparation with 30-fold dilution can reduce the matrix effect.

Figure 3. Matrix effects and their elimination of five cereals samples.

3.5. Determination of recovery

Five kinds of negative cereal samples were spiked with Cd at three concentrations and determined using both the CLEIA and the GFAAS to evaluate the recovery rates of the developed CLEIA. The analysis was performed in triplicate for each concentration. The results are shown in Table 2. The recoveries for both methods ranged from 79.2% to 120.0%, and the RSD values were from 6.0% to 15.8%.

Table 2. Recoveries of Cd from spiked cereal samples by CLEIA and GFAAS.

		Cd-CLEIA			Cd-GFAAS
Sample	Spike level (ng/g)	Mean level (ng/g)	Mean recovery (%)	RSD (%)	Mean level (ng/g)	Mean recovery (%)	RSD (%)
Rice	75	64 ± 8	84.7	12.2	71 ± 9	94.1	12.3
	200	233 ± 35	116.7	15.0	234 ± 5	117.2	2.1
	375	353 ± 50	94.2	14.2	370 ± 52	98.5	14.0
Wheat	75	69 ± 10	91.5	14.3	83 ± 3	110.8	4.1
	200	205 ± 14	102.5	6.9	229 ± 14	114.4	6.2
	375	376 ± 43	100.3	11.5	342 ± 51	91.2	14.9
Maize	75	89 ± 14	118.4	15.8	90 ± 14	120.0	15.0
	200	180 ± 26	90.0	14.5	231 ± 27	115.6	11.7
	375	369 ± 53	98.3	14.2	400 ± 42	106.6	10.5
Polished Glutinous Rice	75	82 ± 10	109.3	11.7	88 ± 4	116.8	4.8
	200	172 ± 14	86.2	8.2	214 ± 4	107.2	2.0
	375	297 ± 20	79.2	6.6	360 ± 24	96.0	6.5
Brown Rice	75	86 ± 13	114.0	14.6	75 ± 4	99.4	4.7
	200	180 ± 24	90.0	13.4	240 ± 4	120.0	1.6
	375	404 ± 24	107.6	6.0	399 ± 60	106.4	14.7

3.6. Screening and analysis of practical samples

The commercial cereal samples (n = 25) were analysed using both the CLEIA and the GFAAS. The results were listed in Table 3. The correlation coefficient of the CLEIA and the GFAAS was 0.984, indicating good reliability and accuracy (Figure 4).

Figure 4. Comparison of GFAAS and CLEIA for screening Cd in different food sample.

Table 3. Concentration of Cd in practical cereal samples determined by CLEIA and GFAAS.

Sample	Number	GFAAS (mg/kg)	CLEIA (mg/kg)	Sample	Number	GFAAS (mg/kg)	CLEIA (mg/kg)
Rice	1	0.0305	0.0385	Polished Glutinous Rice	1	0.0048	0.0054
	2	0.1018	0.0874		2	0.0172	0.0246
	3	0.1074	0.1025		3	0.0383	0.0377
	4	0.1058	0.1249		4	0.0137	0.0199
	5	0.1117	0.1805		5	0.0086	0.0091
Wheat	1	0.0303	0.0459	Brown Rice	1	0.0170	0.0411
	2	0.0020	ND		2	0.0089	0.0084
	3	0.0134	0.0479		3	0.1492	0.1424
	4	0.0135	0.0156		4	0.1127	0.1408
	5	0.0136	0.0352		5	0.0410	0.0385
Maize	1	0.0177	0.0186
	2	0.0084	0.006
	3	0.0218	0.0256
	4	0.0592	0.0477
	5	0.0086	0.0076

Notes: ND: Not detected.

4. Conclusion

In this paper, a sensitive and reliable CLEIA was developed to determine the levels of Cd in cereals. The optimized CLEIA provided a linear detection range from0.16 to 1.56 ng/mL, and the IC₅₀ was 0.5 ng/mL. The method showed a detection limit of 0.06 ng/mL without cross-reactivity with 12 other metals or their chelates. The recoveries of spiked cereal samples ranged from 79.2% to 120.0%. There was a good correlation between the CLEIA and the GFAAS method. This developed CLEIA could be a suitable tool to rapidly screen for Cd in cereal samples at the locations where facilities for acid digestion and atomic absorption spectroscopy analysis are not available.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This study was financially supported by Guangdong Science and Technology Department [grant number 2014B070706001].