Skip to main content
NCBI home page
RSC Advances logoLink to RSC Advances
. 2026 Sep 25. Online ahead of print. doi: 10.1039/d6ra07056f

A colorimetric sensor array based on merging silver nanozymes and gold nanoparticles for chronic kidney disease diagnosis and post-transplant monitoring

Mohammad Mahdi Bordbar a,b, Maryam Alborz b, Azarmidokht Sheini c, Raheleh Halabian d, Zohreh Rostami e, Hosein Khoshsafar a, Hasan Bagheri a,✉
PMCID: PMC13613643  PMID: 42798959

Abstract

Herein, a low cost and non-invasive approach is introduced for identifying individuals with chronic kidney disease through the design and fabrication of a colorimetric sensor array. The proposed platform integrates silver nanozymes with gold nanoparticles as sensing elements. The nanozymes, synthesized via both green and conventional chemical routes, catalyze serum metabolites and convert them into species detectable by gold nanoparticles. A total of 120 serum samples were collected from equal numbers of patients with chronic kidney disease and healthy individuals. Upon exposure of the sensor array to serum metabolites, the resulting colorimetric responses were captured using a smartphone equipped with a dedicated sensor array analysis application capable of performing visual exploratory analysis such as principal component analysis-discriminate analysis (PCA-DA). Complete discrimination between patient and healthy samples was achieved with an overall accuracy of 100%. Furthermore, the sensor correctly identified successful kidney transplantation in 87% of patients. The sensor array proved capable of monitoring treatment progression throughout the one-week to one-year post-transplantation follow-up, characterized by response shifts toward either healthy or diseased clusters. Owing to the limitations of conventional follow-up methods, including high cost and poor user friendliness, the proposed strategy offers a promising and practical alternative for chronic kidney disease diagnosis and post-transplant monitoring.

A simple platform based on merging silver nanozymes with gold nanoparticles as sensing elements for chronic kidney disease diagnosis and post-transplant monitoring.graphic file with name d6ra07056f-ga.webp

Introduction

Chronic kidney disease is a progressive disorder marked by impairment of fluid regulation, electrolyte balance, and metabolic waste excretion. It occurs in acute or chronic forms and can advance to end-stage renal disease (ESRD).1,2 At advanced stages, renal replacement therapies are required, with kidney transplantation being the most effective option.3 Transplantation significantly improves patient survival and quality of life. However, its clinical success relies on early detection of renal dysfunction and continuous post-transplant monitoring.4 Although serum creatinine, eGFR, urinalysis, and imaging are routinely used, each has important limitations: serum creatinine is a late and indirect marker influenced by nonrenal factors; eGFR is an estimated rather than directly measured index and may be less reliable during rapid functional changes; urinalysis is often nonspecific and may fail to detect early injury; and imaging primarily reveals structural abnormalities rather than early functional alterations. Collectively, these limitations highlight the need for rapid, sensitive, and minimally invasive diagnostic strategies for renal monitoring.5

An accurate non-invasive strategy relies on comprehensive metabolomic analysis of biological samples from transplant recipients.6 Analytical techniques such as chromatography, mass spectrometry, and NMR spectroscopy are widely used for this purpose.7 These methods provide sensitive and selective identification and quantification of metabolites. They enable assessment of diagnostic trends related to kidney disease, transplantation, and post-transplant therapy. Such analyses also allow effective monitoring of graft-related injuries. Instrument-based studies show clear differences in metabolic profiles among healthy individuals, kidney patients, and transplant recipients.8 These variations arise from the appearance or depletion of specific organic compounds in serum and plasma samples, including trimethylamine N-oxide, pipecolate, 3-methylhistidin, kynurenine, creatinine, urea, hydroxyproline, 3-indoxyl sulfate, trigonelline, saccharate, guanidinoacetate, omithine, creatine, carnitine, lactate, pyruvate, oxoisopentanoate, citrate, glutamate, lysine, arginine, leucine, serine, proline betaine, butyrobetaine, cystine, hypoxanthine, gluconate, succinate, hippurate.9 Despite their high sensitivity and selectivity, instrument-based analytical methods are generally costly and require advanced equipment, highly controlled laboratory conditions, and skilled personnel. In addition, these techniques are often time-consuming and do not allow for real-time or continuous monitoring of a patient's clinical status.10–13

In recent years, various point-of-care platforms, including wearable electrochemical and colorimetric microfluidic paper-based sensors, have emerged for the diagnosis of chronic kidney disease. These sensing systems target key biomarkers of renal function and injury, such as creatinine, urea, cystatin C, albumin, and disease-related metabolites in minimally invasive or non-invasive biological fluids.14–18 Their widespread appeal stems from essential practical advantages: operational simplicity, cost-effectiveness, portability, rapid response times, and suitability for decentralized testing.18 Nevertheless, conventional single-marker diagnostic strategies often face substantial limitations; many target analytes exhibit suboptimal clinical specificity and are prone to confounding factors such as age, sex, diet, muscle mass, medications, and underlying comorbidities. Furthermore, individual markers fail to capture biological heterogeneity, multifaceted pathological mechanisms, or subtle transitions between disease stages, inevitably increasing the risk of false-positive or false-negative outcomes. To overcome these critical shortcomings, developing selective cross-reactive sensor arrays capable of simultaneously recognizing broad biomarker profiles and generating distinctive fingerprint patterns provides a robust strategy for enhancing diagnostic accuracy and distinguishing patients from healthy individuals.19

Colorimetric sensor arrays have emerged as a low-cost and flexible alternative to conventional instrumental techniques for disease diagnosis by translating metabolite interactions into visible colour changes.20 These platforms consist of chemical dyes or functional nanoparticles that exhibit non-selective yet pattern-recognition responses to a broad range of serum metabolites.13 The collective colour variations generate a multidimensional fingerprint that reflects the global metabolic status of the sample.13 The use of nanoparticles as sensing elements is mainly driven by their high surface area and unique physicochemical properties at the nanoscale. These features provide numerous active sites that promote simultaneous interactions with multiple analytes. As a result, the sensitivity of the sensor array is significantly enhanced. In addition, surface functionalization of nanoparticles with diverse chemical groups improves adaptability toward different metabolites and biomarkers.21 Among various nanomaterials, gold nanoparticles have been extensively employed in the fabrication of colorimetric arrays.22 This preference arises from their high chemical stability and favorable biocompatibility. Gold nanoparticles also exhibit unique optical properties associated with localized surface plasmon resonance. The LSPR effect induces distinct and measurable colour changes upon analyte binding. Furthermore, the resistance of gold nanoparticles to oxidation and environmental fluctuations improves the reproducibility and accuracy of sensor responses.23

Direct and selective surface modification of nanoparticles for metabolite recognition is often constrained by practical limitations or extremely low analyte concentrations, which hinder direct detection. Under such conditions, nanozymes serve as an effective alternative by catalytically converting target metabolites into detectable products.24 By mimicking natural enzyme activities such as oxidase, peroxidase, and catalase, nanozymes enable specific reactions that yield distinct colorimetric or signal-based responses.25 This catalytic amplification markedly improves detection sensitivity. At the same time, it reduces dependence on complex and precise nanoparticle surface functionalization strategies. Nanozymes also exhibit high chemical stability and resistance to diverse environmental conditions. Their relatively low cost further supports their practical implementation in sensing platforms.26 Consequently, nanozymes are well suited for integration into smart sensor arrays. Based on chemical composition, nanozymes are classified as monometallic or multimetallic. Bimetallic nanozymes, through synergistic metal interactions, offer enhanced catalytic efficiency and superior sensing performance.27

The aim of this study is to design and fabricate a novel origami-based sensor array for effective detection of metabolites in serum samples, enabling the diagnosis of chronic kidney disease and monitoring of patients after kidney transplantation. The proposed array consists of two distinct functional layers, such that the first layer contains monometallic silver and bimetallic silver–copper nanoparticles which are likely used as nanozyme and are responsible for the selective catalysis of target metabolites and their conversion into detectable species, while the second layer comprises gold nanoparticles synthesized with methionine, trypsin, and mangiferin as capping agents, which are employed for the recognition and generation of measurable responses toward the products of the catalytic reactions. It should be noted that the proposed mechanism of nanozyme is a hypothesis that was not directly tested in this study. The origami platform enables controlled sample fluid transport, enhanced contact between metabolites and active components, and integration of catalysis and detection within a low-cost and portable system.28 In this study, sensor images are captured, stored, and analyzed using a smartphone-embedded application. The resulting data are evaluated using pattern recognition algorithms to assess the capability of the sensor array to discriminate between healthy and diseased samples and to monitor therapeutic outcomes following kidney transplantation.

Results and discussion

The diagnosis of chronic kidney disease and the evaluation of treatment progression after transplantation were assessed using a colorimetric sensor array. Two critical components of the sensor array are responsible for detecting metabolic differences between healthy individuals and patients: monometallic silver and bimetallic silver–copper nanozymes synthesized using E. aureum and ovalbumin. These nanozymes transform amine, hydroxyl, ester, and carboxylic acid functional groups present in serum samples into chemicals that can interact with AuNPs-based probes whose surfaces are modified with methionine, trypsin, and mangiferin, thereby generating colorimetric signals. It is recalled that the proposed mechanism of nanozyme is a hypothesis that was not directly tested in this study. In this section, useful information regarding the characteristics and performance of the sensor is described in detail.

Characteristics of the synthesized nanoparticles

The output of the UV-vis spectrophotometer consists of absorption spectra, which are shown in Fig. S1. As observed, the absorption peaks appeared at 429 nm for E. aureum-AgNPs, 434 nm for OVL-AgNPs, 452 nm for E. aureum-Cu@AgNPs and 468 nm for OVL-Cu@AgNPs. As evident, changing the capping agent from plant extracts to a compound with a defined chemical structure resulted in a shift of the nanoparticle maximum wavelength toward higher values. In addition, a red shift was observed upon changing the nanoparticle structure from monometallic to bimetallic. The absorption bands of the synthesized AuNPs were placed at 540 nm for Met-AuNPs, 573 nm for Try-AuNPs, and 548 nm for Man-AuNPs. As observed, replacing the stabilizing agent from an amino acid with a polyphenolic structure and subsequently with a protein compound resulted in a red-shift of the AuNP absorption peak toward longer wavelengths.

The hydrodynamic diameter of the synthesized nanoparticles was calculated using dynamic light scattering (DLS). The average sizes obtained are shown as a bar chart in Fig. S2. As indicated, the diameters of the prepared nanoparticles were 22 nm for E. aureum-AgNPs, 28 nm for OVL-AgNPs, 24 nm for E. aureum-Cu@AgNPs, 32 nm for OVL-Cu@AgNPs, 22 nm for Met-AuNPs, 26 nm for Try-AuNPs, and 28 nm for Man-AuNPs. For AgNPs, changing the structure of the capping agent from plant extract to a chemical compound with a protein structure led to an increase in nanoparticle diameter. Moreover, a slight increase in nanoparticle size was observed upon changing the nanoparticle core from monometallic to bimetallic.

The surface charge distribution plots (Fig. S3) indicate that the synthesized nanoparticles possessed a negative surface charge. The surface charge values of each nanoparticle were as follows: −16 mV (E. aureum-AgNPs), −22 mV (OVL-AgNPs), −14 mV (E. aureum-Cu@AgNPs), −18 mV (OVL-Cu@AgNPs), −18 mV (Met-AuNPs), −25 mV (Try-AuNPs), and −12 mV (Man-AuNPs). Although no regular trend was observed for the distributed surface charge, these results indicate that Man-AuNPs exhibited a more positive surface and Try-AuNPs a more negative surface compared with the other NPs.

Optimization of conditions

Various factors, including the concentration of the recognition nanoparticles and nanozymes, pH, buffer concentration, and reaction time, can influence the sensor response. Optimizing these parameters enhances the measurement sensitivity. To determine the optimal conditions, the discrimination ability function (DAF) was employed, which aims to identify the conditions that provide the best separation between healthy and diseased groups. The optimal response corresponds to the maximum DAF value.

In the first step, the concentration of the recognition element (AuNPs) was varied from 0.2 to 1.2 mg mL−1. As shown in Fig. 1a, the discrimination between the two studied groups improved with increasing nanoparticle concentration. The maximum DAF value was obtained at a concentration of 1.0 mg mL−1, beyond which no significant improvement in sensor performance was observed. The poor performance at lower concentrations is attributed to the insufficient number of active surface sites on the nanoparticles for selective recognition of the chemical structures produced by the enzymatic reaction.29,30 Notably, the highest concentration of enzymatic reaction products was obtained when the nanozyme concentration was set to 0.5 mg mL−1 (Fig. 1b).

Fig. 1. The optimized conditions: (a) the concentration of AuNPs, (b) the concentration of nanozymes, (c) pH of media, (d) the concentration of Tris buffer and (e) the time of reaction.

Fig. 1

Open in a new tab

The pH of the reaction medium strongly affects both the catalytic activity of the nanozymes and the interaction between the reaction products and the AuNPs.29 To investigate this effect, the pH of the Tris buffer was varied from 4.0 to 10.0. As illustrated in Fig. 1c, increasing the alkalinity of the reaction medium and adjusting the pH to 8 enhanced the catalytic activity of the nanozymes and promoted stronger interactions between AuNPs and the chemical species. In acidic environments, the high concentration of hydronium ions inhibits the catalytic reaction and blocks the reactive sites on the AuNPs surface. Conversely, under highly alkaline conditions, weaker interactions occur due to interference from excess alkaline and hydroxide ions, resulting in a reduced DAF value.31

In another experiment, the effect of Tris buffer concentration in the range of 0.04 M −0.12 M on sensor performance was evaluated. According to the data shown in Fig. 1d, the optimal response was achieved when the Tris buffer concentration in the buffered nanoparticle solutions was 0.1 M.

The time required for completion of the catalytic reaction and subsequent interaction of the generated metabolites with AuNPs is shown in Fig. 1e. As seen, 4 min after sample injection into the sensor is sufficient for completion of the reaction and stabilization of the colour change in the sensing elements. No significant changes in sensor performance were indicated thereafter; therefore, this time was selected for recording the sensor response.

Sensor responses

The sensor ability strongly depends on the type of silver-copper nanoparticles which probably have the catalytic activity and the selectivity of the stabilizing agents on the recognition nanoparticles. The catalytic activity is governed by the number of active cores in the nanozymes and the functional groups covering their surfaces. The nanozymes catalyze oxidation, dehydrogenation, and hydrolysis reactions. Each gold nanoparticle interacts with all synthesized nanozymes and can therefore interact with all products generated by the catalytic reactions. It is noted that the proposed mechanism of nanozyme is a hypothesis that was not directly tested in this study. The stabilizing agents of the recognition nanoparticles include an amino acid, a protein, and a poly-hydroxyl group. Depending on the type of functional groups, as well as their charge and size, the nanoparticles can interact with metabolites through electrostatic, hydrogen-bonding, and nucleophilic interactions. These interactions reduce the interparticle distance, induce aggregation, and lead to a colour change from red to purple.32

The colorimetric responses of the sensor to serum metabolites from healthy individuals, patients, and kidney-transplant recipients are shown in Fig. 2a. Although the colour changes were discernible with the naked eye, sensor images were captured using a smartphone and uploaded to a dedicated application to better evaluate the discrimination capability of the sensor. In addition to image analysis, the software generated colour-difference maps (Fig. 2b). As observed, each healthy and diseased group exhibited a unique colour pattern, indicating that the sensor effectively discriminated metabolomic variations in the serum samples of the studied groups. According to this figure, nanozymes synthesized with ovalbumin selectively catalyzed characteristic metabolites present in the serum of healthy individuals. Among these, bimetallic nanozymes exhibited higher catalytic activity, and their reaction products preferentially interact with Man-AuNPs and Met-AuNPs. Compared with the chemical structure of methionine, the mangiferin showed a higher capacity for analyte adsorption.

Fig. 2. The sensor operation: (a) the colorimetric response and (b) the colorimetric difference maps. All experiments were performed in optimal conditions expressed in Fig. 1.

Fig. 2

Open in a new tab

In contrast, nanozymes synthesized using E. aureum were capable of reacting both with disease-specific metabolites in patient serum and with chemical species common to both healthy and diseased individuals. However, the products of disease-related catalytic reactions were recognized by Met-AuNPs and Try-AuNPs, whereas the species generated from the catalysis of common metabolites interacted with Man-AuNPs. The relatively large size of trypsin limited its contribution to recognition of the catalytic reaction products.

By comparing sensor responses to patient serum samples before and after kidney transplantation, the colorimetric pattern is consistent with a decrease in the concentrations of kidney-failure-related metabolites in transplant recipients. As a result, no specific chemical species were available for catalysis by E. aureum-AgNPs, and consequently no colour change was observed for the corresponding AuNPs. Nevertheless, E. aureum-Cu@AgNPs enabled the residual target metabolites to be catalyzed, and the reaction products were detected by Met-AuNPs. On the other hand, kidney transplantation led to an increase in concentration of characteristic metabolites which was shown in healthy serum samples. This was confirmed by the colour changes of Met-AuNPs and Man-AuNPs, which interact with the catalytic reaction products of OVL-Cu@AgNPs and OVL-AgNPs, respectively. An interesting feature in the colour pattern of post-transplant patients was the colour change of trypsin-stabilized AuNPs in response to the catalytic products of E. aureum-AgNPs, suggesting that kidney transplantation may be associated with the release of new metabolites into the serum that were not present in healthy or pre-transplant patient samples; however, as the sensor reports only an aggregate colorimetric pattern rather than the identity of individual analytes, this interpretation remains a hypothesis to be confirmed by a targeted analytical technique such as LC-MS or NMR on a representative subset of samples.

The outputs of visual exploratory analysis

After image analysis, the obtained data were compiled into a matrix with dimensions of 180 × 36. It should be noted that each sample was measured five times and the resulting mean values were incorporated into the data matrix. This matrix was subjected to principal component analysis-discriminant analysis (PCA-DA) to uncover the data distribution patterns hidden within the raw matrix.33 The clustering of samples into three groups (healthy, diseased, and transplant recipients) in the space defined by the two first principal components is shown in Fig. 3. This space retains approximately 86% of the useful information contained in the raw matrix. The statistical parameters derived from this analysis are summarized in Table 1.

Fig. 3. The visual exploratory analysis by PCA-DA algorithm. In this figure, (H), (CKD), and (TP) represent healthy controls, chronic kidney disease controls, and transplant patients, respectively. All experiments were performed in optimal conditions expressed in Fig. 1.

Fig. 3

Open in a new tab

Table 1. The results of the visual exploratory analysis obtained by PCA-DA analysis.

Parameters Patient vs. healthy Patient vs. TP Healthy vs. TP
Sensitivity (%) 100.0 100.0 98.0
Specificity (%) 100.0 87.0 100.0
Accuracy (%) 100.0 93.0 99.0
Error rate (%) 0.0 7.0 1.0
Open in a new tab

As can be seen, the designed sensor successfully discriminated between the healthy and diseased groups. The overall classification accuracy was 100%. Additionally, 87% of serum samples from transplant recipients exhibited responses similar to those of healthy individuals; however, they did not fall exactly within the healthy cluster and instead formed a distinct group. Although these individuals had serum creatinine levels below 1.5 mg dL−1, newly emerged metabolites were present in their serum samples, and a fraction of kidney-failure-related metabolites still remained. Consequently, their distribution patterns differed from those of healthy samples. Based on the sensor data, 12% of transplant recipients had not achieved a fully successful kidney transplantation, as their serum creatinine levels remained above 1.5 mg dL−1.

Evaluation of treatment progress in transplant patients

To monitor patient recovery after kidney transplantation, metabolomic changes in serum samples from 10 transplant recipients were tracked over a 6-month period. The PCA score plot corresponding to each patient is presented in Fig. S4. As observed, the trajectory of serum metabolomics changes during the treatment period may shift toward either the healthy cluster or the diseased group. Among the studied patients, P4, P6, and P9 showed positive responses to the administered medications. In contrast, the levels of CKD-related serum metabolites in patients P5, P8, and P10 increased over the course of treatment.

For the remaining patients, fluctuating metabolomics trends were observed: at certain time points, patients exhibited improvement and their colorimetric patterns moved toward the sensor response of healthy samples, whereas at other times their physical condition worsened and the sensor responses resembled the CKD colour pattern. Nevertheless, samples from patients P1, P2, P3, and P7 showed favourable clinical status during the final two months of monitoring.

In another study, metabolomic changes in serum samples were evaluated one year after transplantation. For this assessment, 20 patients were selected. According to the data presented in Fig. S5, the colorimetric patterns obtained for seven samples were close to the sensor response of CKD patients. For four patients, the response pattern was identical to that of transplant recipients with serum creatinine levels below 1.5 mg dL−1. Based on this figure, the remaining patients successfully completed the one-year treatment period, as their samples were fully clustered within the healthy group.

Evaluation of sensor stability

The colour of the sensing elements used in the sensor array can be affected by environmental physical and chemical conditions.34 Therefore, the colour variations of the sensor array were examined over a 7-month period. The plot shown in Fig. S6 illustrates the sensor response, expressed as Euclidean norm, throughout the analysis period. As observed, the colour of the AuNPs remained unchanged for up to 5 months. Beyond this period, aggregation or surface modification of the nanoparticles occurred over time, leading to an increased sensor response in the absence of analytes.

Reproducibility of sensor response

To evaluate the reproducibility of the colour changes of the sensor array after exposure to analytes, five sensor arrays fabricated under identical conditions were separately immersed in two serum samples obtained from a healthy individual and a patient, respectively. For each sensor, the colour changes of the sensing elements were collected and reported as Euclidean norms, as shown in Fig. S7. The relative standard deviation (RSD) was calculated to be 4.96% for the healthy sample and 5.12% for the patient sample, demonstrating satisfactory reproducibility of the sensor responses for detection of serum metabolites.

Various analytical methods have been developed for the diagnosis of chronic kidney disease through metabolic profile alterations in biological samples, some of which are summarized in Table S1. In these methods, critical metabolites in exhaled breath, serum, saliva, and urine which were primarily responsible for differentiating diseased individuals from healthy controls were qualitatively and quantitatively detected using spectrometric methods and electrochemical platforms modified with conducting polymers, gold nanoparticles, nanocomposites, nanotubes, and graphene. In these studies, separation of the studied groups was achieved using various analysis methods such as ROC plots, SVM, PLS-DA and LDA, resulting in sensitivities ranging from 75% to 100%. Compared to previous studies, the proposed method was based on paper-based platforms in which the sensor response, based on a colour change, was easily observed with the naked eye. In this study, a nanozyme-nanoparticle couple was utilized, allowing the detection of subtle structural differences among metabolites. This has led to achieving excellent sensitivity using PCA-DA. In addition, lower fabrication costs and ease of use were among its advantages; however, since serum samples were used for analysis, this cannot be regarded as a non-invasive method.

Experimental section

Chemicals and solutions

All chemicals used in this study were of analytical grade and used without any further purification. Silver nitrate (AgNO3), copper(ii) nitrate trihydrate (Cu(NO3)2·3H2O), chloroauric acid (HAuCl4), tris(hydroxymethyl)aminomethane (Tris), sodium borohydride (NaBH4), sodium hydroxide (NaOH), and hydrochloric acid (HCl) were purchased from Merck. Ovalbumin (OVL), methionine (Met), trypsin (Try), mangiferin (Man), and sodium citrate were obtained from Sigma-Aldrich. Epipremnum aureum (E. aureum) was purchased from the Tehran local market. NaOH and HCl were used to adjust the pH to desired values.

Instrumentation and software

The localized surface plasmon resonance spectra of the synthesized nanoparticles were recorded using a UV-visible spectrophotometer (Jasco, V-570 model). The hydrodynamic size of the nanoparticles was determined by dynamic light scattering (DLS). In addition, the surface charge (zeta potential) of the nanoparticles was measured using a Zetasizer Nano ZS90. A smartphone (Samsung Galaxy A52) was used for image acquisition. A dedicated smart application installed on the smartphone was employed for image analysis, colour map generation, and classification evaluation. The sensor array pattern was designed using AutoCAD 2016 and printed onto a Whatman grade 2 paper substrate using an HP LaserJet 1320 printer.

Extraction procedure

For extraction, E. aureum leaves were first washed with double-deionized water and then air-dried at room temperature. After complete drying in an oven, the plant material was ground into a fine powder. Subsequently, 10.0 mg of the powdered plant was added to 100.0 mL of water, noting that the water was pre-boiled. The extraction process was carried out on a magnetic stirrer for 1 h. After filtration through filter paper, the extract was centrifuged at 3900 rpm for 20 min to remove any residual species. The supernatant was then repeatedly filtered using a 0.22 µm syringe filter until a colourless solution was obtained. The prepared extract was stored in a clean Falcon tube at 4 °C for subsequent analyses. For nanoparticle synthesis, the obtained extract was diluted 10-fold prior to use.35

Nanoparticles preparation

AgNPs and Cu@AgNPs prepared by E. aureum

To synthesize E. aureum-AgNPs, 50.0 mL of a 1 × 10−3 M AgNO3 solution was prepared and heated to 50 °C under continuous stirring using a magnetic stirrer. Subsequently, 5.0 mL of the purified extract was gradually added to the solution. Stirring was continued until a colour change from colourless to yellow was observed, indicating the initiation of the reaction. The solution was further stirred until a uniform and stable yellow colour was achieved.36 For creating E. aureum-Cu@AgNPs, 25.0 mL of a 1 × 10−3 M AgNO3 solution was mixed with 25.0 mL of a Cu(NO3)2·3H2O solution of the same concentration. The resulting mixture was heated to 50 °C under continuous stirring using a magnetic stirrer. Subsequently, 5.0 mL of the purified extract was added, and stirring was continued until the solution colour changed to yellowish-brown.36

AgNPs and Cu@AgNPs prepared by ovalbumin

A volume of 1.0 mL of a 0.3 M AgNO3 solution was added to 30.0 mL of deionized water containing 30.0 mg of ovalbumin. Subsequently, 1.0 mL of a freshly prepared aqueous NaBH4 solution (0.3 M) was injected into the reaction mixture over a period of 3–5 min, resulting in a colour change from colourless to dark brown. The solution was maintained under stirring for 30 min. Unreacted ovalbumin was removed by centrifugation, and the remaining nanoparticles were re-dispersed in deionized water for subsequent studies.37 The same procedure was employed for the synthesis of bimetallic nanoparticles, except that, in addition to the 0.15 M AgNO3 solution, an equimolar Cu(NO3)2·3H2O solution was also introduced into the reaction system.

AuNPs prepared by methionine, trypsin, and mangiferin

Gold nanoparticles capped with methionine, trypsin, and mangiferin were prepared using the same procedure. Briefly, 1.0 mL of a freshly prepared 1% (w/v) sodium citrate solution was rapidly added to a boiling aqueous solution of HAuCl4 (2.5 × 10−4 M, 100.0 mL) under reflux conditions. After 15 min, a colour change from yellow to red was observed, indicating the formation of AuNPs. The synthesized AuNPs were allowed to cool to room temperature and then mixed with 3.0 mL of a methionine solution (20.0 mg mL−1) under vigorous stirring. The resulting mixture was stirred continuously for 20 h to ensure complete surface functionalization of the NPs. To remove unreacted species, the final solution was centrifuged at 10 000 rpm for 10 min. The modified AuNPs were collected and re-dispersed in deionized water for subsequent experiments.38 The same procedure was followed for the synthesis of AuNPs via the other capping agents, with methionine replaced by trypsin or mangiferin.

Preparation of nanoparticle solutions

The synthesized NPs were converted into powder using a freeze dryer. A defined amount of each powdered nanoparticle was weighed and dispersed in 1.0 mL of Tris buffer to obtain a specified nanoparticle concentration.

Sensor design

The proposed sensor was composed of four main components: a strip for serum sample collection and transport; an injection zone connected to the strip that stored the collected sample and transferred it to subsequent sections; a nanozyme zone serving as the deposition site for the synthesized Ag nanozyme at a defined concentration; and a detection zone containing AuNPs synthesized at a specific concentration. The sensor layout was designed using AutoCAD 2016 (Fig. 4a) and subsequently fabricated by printing onto Whatman grade 2 paper. As shown in Fig. 4a, the sensor comprised two distinct hydrophilic and hydrophobic regions, indicated by white and black colours, respectively. To render the black regions hydrophobic, the printed paper was heated in an oven at 200 °C for 45 min, allowing the printer ink to melt and penetrate into the paper matrix.39 The small square regions corresponding to the nanozyme and detection zones were loaded with 1.0 µL of the buffered solutions of the NPs. The arrangement of AgNPs and AuNPs in the nanozyme and detection zones is illustrated in Fig. 4b. An image of the fabricated sensor is shown in Fig. 4c.

Fig. 4. (a) The design used for preparation of colorimetric sensor array, (b) the list of nanozymes and AuNPs used for fabrication of colorimetric sensor array and (c) the prepared sensor used for analysis of serum metabolites.

Fig. 4

Open in a new tab

Sample selection

The residual serum samples from patients and control individuals were obtained from the clinical laboratory of Baqiyatallah Al-Azam Hospital prior to their disposal. The patient samples belonged to the individuals with chronic kidney disease (CKD). According to national regulations, the use of anonymized leftover biological samples does not require obtaining informed consent, provided that approval from the relevant ethics committee has been secured. All samples were irreversibly anonymized before use and were not subjected to any further treatment. The sample selection also was confirmed by a specialist physician. The serum creatinine level of these patients was equal to or higher than 1.5 mg dL−1. In contrast, healthy control individuals (without kidney disease and with no history of transplantation) were selected from visitors to other hospital departments; all had serum creatinine levels below 1.5 mg dL−1. A total of 60 individuals were included in the patient group and 60 samples in the healthy control group. Table 2 presents the demographic information of the collected samples.

Table 2. Demographic data for the studied population.
Variable Healthy controls Patients
Studied sample 60 60
 
Sex
Male 34 31
Female 26 29
Age (Mean ± SD) 52.43 (±12.77) 53.36 (±11.93)
Open in a new tab

Sensor operation

Fig. 5 illustrates a schematic representation of the sensor fabrication and operating procedure. According to this figure, the designed sensor components (Fig. 5a) were folded in such a manner that the injection, nanozyme, and detection zones were fully aligned and overlapped (Fig. 5b). The detection zone was covered with a transparent polymeric substrate to protect it from environmental effects and to enhance the sensor shelf life (Fig. 5c). The injection zone of the sensor was connected to a strip (Fig. 5d) that could be immersed in the container containing the serum sample (Fig. 5e). After the colour change of the sensor, the response was recorded using a smartphone (Fig. 5f). The smartphone was equipped with a dedicated application through which the sensor images were analyzed (Fig. 5g). This application was capable of generating colour maps based on the obtained numerical data. Subsequently, the data were analyzed using a classification toolbox implementing pattern-recognition algorithms. Image acquisition was performed inside a box with a constant illumination source. In addition, the application includes denoising and preprocessing modules to remove potential image interferences. In this study, each test was repeated five times.

Fig. 5. The schematic for sensor operation: (a) fabricating the sensor array, (b) folding the different layer of sensor, (c) sealing the detection section of the sensor, (d) attaching a strip to the injection section of the sensor, (e) immersing the strip in the serum sample and directing the sample to the injection section of the sensor, (f) collecting sensor response with a smartphone and (g) using an application to analyze the sensor response and evaluate its ability to classify the studied samples.

Fig. 5

Open in a new tab

Statistical analysis

The collected images were uploaded into the application embedded in the smartphone. In the Select Area section, the software allowed the simultaneous selection of the 12 squares located in the detection zone of the sensor (i.e., the sites where the sensing elements were deposited). Subsequently, within each square, the pixel-by-pixel values of the red (R), green (G), and blue (B) colour channels were measured, and the mean value of each colour channel was calculated separately. For each sensing element, the RGB values obtained before interaction were subtracted from the corresponding values achieved after interaction. The resulting 36 numerical values (12 sensing elements × 3 colour channels) were then arranged as a single row in an Excel file. In essence, this numerical vector represents the response of one sensor to one sample.

After analysis of all healthy, diseased, and transplant samples, a matrix with dimensions of 180 × 36 was compiled. This file was uploaded into the Classification Toolbox, which contained the algorithm of principal component analysis-discriminate analysis (PCA-DA) responsible for apparent performance of the proposed sensor array from an exploratory analysis of the studied groups.

In addition, the following equation was used to calculate the overall sensor response:

graphic file with name d6ra07056f-t1.webp 1

where xj represents the elements of the numerical vector obtained for a given sample.40

In this study, the optimal values of the parameters influencing the sensor response were selected using the discrimination ability function (DAF).41 This function calculates the ratio of between-group variation to within-group variation according to the following equation:

graphic file with name d6ra07056f-t2.webp 2

In this equation, n denotes the number of studied groups. The other parameters are defined as the overall sensor response for the j-th sample in i-th group (Xij), the mean response obtained for all samples in i-th group (X̄i), and the mean response received from all collected samples in the patient and healthy groups (X̿).

Compliance with ethical standards

All procedures involving human samples were conducted in accordance with the Declaration of Helsinki and the National Guidelines of the Iranian Committee for Ethics in Biomedical Research (https://ethics.research.ac.ir). The research ethics committee of Baqiyatallah University of Medical Sciences has approved the project (Approval ID: IR.BMSU.BAQ.REC.1404.085).

Conclusions

In this work, a hybrid colorimetric sensor array based on a combination of nanoparticles was developed for the detection and monitoring of chronic kidney disease. The sensor was capable of converting metabolites present in biological samples into species detectable by gold nanoparticles, enabling discrimination of patients from healthy individuals through the visual exploratory analysis with an overall accuracy of 100%. The sensor responses demonstrated that monometallic and bimetallic nanozymes synthesized via green and chemical methods exhibit distinct catalytic performances toward serum metabolites. Changes in the type and concentration of serum metabolites in patients before and after transplantation, as well as the rapid identification of patient status during the treatment period, together with ease of use, short response time, and acceptable reproducibility, represent key advantages of the proposed sensor. Nevertheless, the sensor response was evaluated on a limited population, the diagnostic comparison relied on a single biomarker (serum creatinine) as ground truth, and the identity of the metabolites underlying the colorimetric response was not chemically confirmed, and validation using a larger, independent cohort of serum samples, together with independent chemical confirmation of the metabolite changes underlying the colorimetric signal, is required before its implementation as an effective clinical diagnostic alternative to conventional methods.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Supplementary Material

RA-OLF-D6RA07056F-s001
RA-OLF-D6RA07056F-s002

Acknowledgments

This work is based upon research funded by Iran National Science Foundation (INSF) under project No. 4023931. The authors extend their gratitude to the Cohort Study of Iranian Chemical Injured (CICI Study) at Baqiyatallah University of Medical Sciences (Ethical approval: IR.BMSU.REC.1399.422) for sharing its support in facilitating sample preparation, and biobanking.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: data on the characterization of the synthesized nanoparticles, PCA score plots, and the evaluation of stability and reproducibility. See DOI: https://doi.org/10.1039/d6ra07056f.

References

  1. Erdem H. and Aydin H. R., Approaches to Common Diseases in Family Medicine Practice with Clinical Practice Guidelines, 2025, pp. 269–278 [Google Scholar]
  2. Abo-Ghneim F. D. F. Mohammed H. J. Al-Koofee D. A. F. World Acad. Sci. J. 2024;6:66. doi: 10.3892/wasj.2024.281. [DOI] [Google Scholar]
  3. Kuttymuratov G. Saliev T. Ainakulov A. Ayaganov A. Oshakbayev K. Zharassov D. Tuleuzhan A. Uderbayev N. Medicina. 2025:1045. doi: 10.3390/medicina61061045. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Krishnan N. Higgins R. Short A. Zehnder D. Pitcher D. Hudson A. Raymond N. T. Am. J. Transplant. 2015;15:2378–2386. doi: 10.1111/ajt.13363. [DOI] [PubMed] [Google Scholar]
  5. Alobaidi S. Diagnostics. 2025;15:1225. doi: 10.3390/diagnostics15101225. [DOI] [Google Scholar]
  6. Kuypers D. R. J. Kamphorst J. J. de Loor H. O'Day E. M. Biomarkers Med. 2024;18:787–794. doi: 10.1080/17520363.2024.2394383. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Pandey S. Nanotheranostics. 2025;9:110–120. doi: 10.7150/ntno.108320. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Abbiss H. Maker G. L. Trengove R. D. Metabolites. 2019:34. doi: 10.3390/metabo9020034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Iwamoto H. Okihara M. Akashi I. Kihara Y. Konno O. Kawachi S. Sunamura M. Sugimoto M. Int. J. Mol. Sci. 2022:13938. doi: 10.3390/ijms232213938. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dehghan K. Joolaei Ahranjani P. Joolaei Ahranjani P. Esfandiari Z. Ferrentino G. Appl. Food Res. 2025:101526. doi: 10.1016/j.afres.2025.101526. [DOI] [Google Scholar]
  11. Wei Y. Gu M. Ren Y. Hu Z. Biomed. Rep. 2026:1–10. doi: 10.3892/br.2026.2120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. McGann P. T. Hoppe C. Blood Cells Mol. Dis. 2017;67:104–113. doi: 10.1016/j.bcmd.2017.08.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Bordbar M. M. Sheini A. Hashemi P. Hajian A. Bagheri H. Biosensors. 2021:316. doi: 10.3390/bios11090316. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Lewińska I. Speichert M. Granica M. Tymecki Ł. Sens. Actuators, B. 2021:129915. doi: 10.1016/j.snb.2021.129915. [DOI] [Google Scholar]
  15. Chen H. Chen J. Liu Y. Li B. Li H. Zhang X. Lv C. Dong H. Langmuir. 2023;39:3420–3430. doi: 10.1021/acs.langmuir.2c03347. [DOI] [PubMed] [Google Scholar]
  16. Xu H. Hua Y. Li H. Wang J. Liu G. Li X. Anal. Biochem. 2025:115824. doi: 10.1016/j.ab.2025.115824. [DOI] [PubMed] [Google Scholar]
  17. Sabarudin A. Fiddaroini S. Fahmi A. L. Roja’i A. M. Salsabila I. E. ’am A. Sakti S. P. Susianti H. Microchem. J. 2025:113071. doi: 10.1016/j.microc.2025.113071. [DOI] [Google Scholar]
  18. Mandal N. Dutta S. Gupta A. Bandyopadhyay D. IEEE Sens. J. 2020;20:9644–9651. [Google Scholar]
  19. Shojaeifard Z. Bordbar M. M. Aseman M. D. Nabavizadeh S. M. Hemmateenejad B. Sens. Actuators, B. 2021:129582. doi: 10.1016/j.snb.2021.129582. [DOI] [Google Scholar]
  20. Shao L. Liu J. Chen X. Xiao W. Anal. 2024:1034–1039. doi: 10.1039/d5an00900f. [DOI] [Google Scholar]
  21. Lu X. Suslick K. S. Li Z. Anal. Sens. 2023;3:e202200050. doi: 10.1002/anse.202200050. [DOI] [Google Scholar]
  22. Sun J. Lu Y. He L. Pang J. Yang F. Liu Y. TrAC, Trends Anal. Chem. 2020:115754. doi: 10.1016/j.trac.2019.115754. [DOI] [Google Scholar]
  23. Aldewachi H. Chalati T. Woodroofe M. N. Bricklebank N. Sharrack B. Gardiner P. Nanoscale. 2018;10:18–33. doi: 10.1039/c7nr06367a. [DOI] [PubMed] [Google Scholar]
  24. Tian Q. Li S. Tang Z. Zhang Z. Du D. Zhang X. Niu X. Lin Y. Adv. Healthcare Mater. 2025;14:2401630. doi: 10.1002/adhm.202401630. [DOI] [PubMed] [Google Scholar]
  25. Das B. Lou Franco J. Logan N. Balasubramanian P. Il Kim M. Cao C. Nano-Micro Lett. 2021;13:193. doi: 10.1007/s40820-021-00717-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Elagamy S. H. Obaydo R. H. Ashkar A. Lotfy H. M. Sens. Acutators Rep. 2025:100394. doi: 10.1016/j.snr.2025.100394. [DOI] [Google Scholar]
  27. He Q. Zhang L. RSC Adv. 2025;15:47648–47665. doi: 10.1039/d5ra08106h. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Sheini A. Aseman M. D. Bordbar M. M. Sci. Rep. 2021;11:3521. doi: 10.1038/s41598-021-83186-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Mirzaei Y. Gholami A. Sheini A. Bordbar M. M. Sci. Rep. 2023;13:7064. doi: 10.1038/s41598-023-34299-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Bordbar M. M. Samadinia H. Sheini A. Halabian R. Parvin S. Ghanei M. Bagheri H. Sens. Actuators, B. 2022;368:132170. doi: 10.1016/j.snb.2022.132170. [DOI] [Google Scholar]
  31. Alborz M. Bordbar M. M. Samadinia H. Bagheri H. Microchem. J. 2025:115069. doi: 10.1016/j.microc.2025.115069. [DOI] [Google Scholar]
  32. Sheini A. Microchim. Acta. 2020;187:167. doi: 10.1007/s00604-020-4147-5. [DOI] [PubMed] [Google Scholar]
  33. Ataide V. N. Pradela Filho L. A. Guinati B. G. S. Moreira N. S. Gonçalves J. D. Ribeiro C. M. G. Grasseschi D. Coltro W. K. T. Salles M. O. Paixão T. R. L. C. TrAC, Trends Anal. Chem. 2024:117451. doi: 10.1016/j.trac.2023.117091. [DOI] [Google Scholar]
  34. Mahdi Bordbar M. Nobakht B. F. Gh. M. Sheini A. Alborz M. Parvin S. Ghanei M. Kulahlioglu N. Mehdi Khoshfetrat S. Samadinia H. Bagheri H. Microchem. J. 2024:111516. doi: 10.1016/j.microc.2024.111516. [DOI] [Google Scholar]
  35. Mirzaei Y. Gholami A. Bordbar M. M. Sens. Actuators, B. 2021:130445. doi: 10.1016/j.snb.2021.130445. [DOI] [Google Scholar]
  36. Bordbar M. M. Tashkhourian J. Hemmateenejad B. ACS Appl. Mater. Interfaces. 2022;14:8333–8342. doi: 10.1021/acsami.1c24194. [DOI] [PubMed] [Google Scholar]
  37. Garza-Ocañas L. Ferrer D. A. Burt J. Diaz-Torres L. A. Cabrera M. R. Rodríguez V. T. Rangel R. L. Romanovicz D. Jose-Yacaman M. Metallomics. 2010;2:204–210. doi: 10.1039/b916107d. [DOI] [PubMed] [Google Scholar]
  38. Fu R. Wang C. Zhuang J. Yang W. Colloids Surf., A. 2014;444:326–329. [Google Scholar]
  39. Taghizadeh-Behbahani M. Hemmateenejad B. Shamsipur M. Tavassoli A. Anal. Chim. Acta. 2019;1080:138–145. doi: 10.1016/j.aca.2019.06.050. [DOI] [PubMed] [Google Scholar]
  40. Bordbar M. M. Samadinia H. Sheini A. Aboonajmi J. Hashemi P. Khoshsafar H. Halabian R. Khanmohammadi A. Nobakht B. F. Gh M. Sharghi H. Ghanei M. Bagheri H. Anal. Chim. Acta. 2022:340286. doi: 10.1016/j.aca.2022.340286. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Brereton R. G. Chemometrics. 2003;8:131–132. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

RA-OLF-D6RA07056F-s001
RA-OLF-D6RA07056F-s001.pdf (2.3MB, pdf)
RA-OLF-D6RA07056F-s002
RA-OLF-D6RA07056F-s002.pdf (3.3MB, pdf)

Data Availability Statement

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: data on the characterization of the synthesized nanoparticles, PCA score plots, and the evaluation of stability and reproducibility. See DOI: https://doi.org/10.1039/d6ra07056f.

Articles from RSC Advances are provided here courtesy of Royal Society of Chemistry

RESOURCES