Ultrasensitive bioassaying of HER-2 protein for diagnosis of breast cancer using reduced graphene oxide/chitosan as nanobiocompatible platform

In this label-free bioassay, an electrochemiluminescence (ECL) immunosensor was developed for the quantification of breast cancer using HER-2 protein as a metastatic biomarker. For this purpose, the ECL emitter, [Ru(bpy)3]2+, was embedded into biocompatible chitosan (CS) polymer. The prepared bio-composite offered high ECL reading due to the depletion of human epidermal growth factor receptor 2 (HER-2) protein. Reduced graphene oxide (rGO) was used as substrate to increase signal stability and achieve greater sensitivity. For this, rGO was initially placed electrochemically on the glassy carbon electrode (GCE) surface by cyclic voltammetry (CV) technique. Next, the prepared CS/[Ru(bpy)3]2+ biopolymer solution was coated on a drop of the modified electrode such that the amine groups of CS and the carboxylic groups of rGO could covalently interact. Using EDC/NHS chemistry, monoclonal antibodies (Abs) of HER-2 were linked to CS/[Ru(bpy)3]2+/rGO/GCE via amide bonds between the carboxylic groups of Ab molecules and amine groups of CS. The electrochemical behavior of the electrode was studied using different electrochemical techniques such as electrochemical impedance spectroscopy (EIS), differential pulse voltammetry (DPV) and square wave voltammetry (SWV) and also ECL tests. After passing all optimization steps, the lower limit of detection (LLOQ) and linear dynamic range (LDR) of HER-2 protein were practically obtained as 1 fM and 1 fM to 1 nM, individually. Importantly, the within and between laboratory precisions were performed and the suitable relative standard deviations (RSDs) were recorded as 3.1 and 3.5%, respectively. As a proof of concept, the designed immunosensor was desirably applied for the quantification of HER-2 protein in breast cancer suffering patients. As a result, the designed ECL-based immunosensor has the capability of being used as a conventional test method in biomedical laboratories for early detection of HER-2 protein in biological fluids.

surface-to-volume ratio and desirable conductivity, have made nanomaterials as excellent candidates for modification of biosensing transducers including electrochemical and ECL frameworks (Babaei et al. 2010;Karim-Nezhad et al. 2009;Khalilzadeh et al. 2011;Saghatforoush et al. 2009). Various types of nanomaterials with different shapes and sizes are considerably applied in ECL biosensing of BC such as gold nanoparticles (AuNPs) (Zhu et al. 2012), silver nanoparticles (AgNPs) (Yazdanparast et al. 2018), silver nanoclusters (AgNCs) (Guo et al. 2015), ZnO nanowires (ZnNWs) (Mansor et al. 2014), and CNTs (Nawaz et al. 2016). One of the drawbacks of the mentioned nanostructures is their expensiveness and poor functionality in some cases. Among different nanomaterials, graphene-based nanosheets attracted great interests considering their good conductivity, high loading capacity, good functionality and desirable stability. Also, graphene can simply produce in inexpensive way (Li et al. 2013).
There are several reports based on graphene modified electrodes for electrochemiluminescence determinations using tris(2,2-bipyridyl) ruthenium (II) [Ru(bpy) 3 ] 2+ ) as ECL indicator. For instance, Li et al. developed an electrochemiluminescence biosensing strategy advantaging a triple combination of graphene-nafion-[Ru(bpy) 3 ] 2+ . Another ECL method was introduced by Wang and coworkers using graphene modified electrode. In this probe graphene nanosheets (GNSs) were exploited as desirable supporting substrate to hold ECL indicator. CS was employed as adhesive compound for this purpose (Wang et al. 2012). Employing attachment/adhesive compounds is of high interest in developing of ECL frameworks. This can increase the stability, sensitivity and functionality of the proposed bioassay. In this way, adhesive materials hold luminophores onto the electrode surface and prevent them from releasing (Wang et al. 2020;Bakshi et al. 2020). Nafion is an adhesive material for attachment of nano/biomaterials onto the surface of electrodes. High sensitivity and good signal stability are obtained in the consequence (Li et al. 2009;Bahari et al. 2020;Huang et al. 2020). Chitosan (CS) is another adhesive compound which is an appealing for its great characteristics. Unlike nafion, which is a class of synthetic polymers (Juliandri et al. 2018;Gluschke et al. 2021), CS is a biopolymer which make chitosan as a biocompatible and environmentally friendly material (Bakshi et al. 2020). There are some drawbacks in using of nafion for modification of electrodes. First, nafion is more expensive compared to the CS (Sigmaaldrich 2021a, b), second, in contrast of nafion, CS possess amine functional groups. This help to attach the antibody molecules onto the platform via amide bonds between their carboxylic groups and the amine functions of CS. Considering the fact that Abs are more arranged by attaching from their -COOH groups, CS is more appropriate by its ability to make covalent amide bonds with carboxylic ends of Abs.
To the best of our knowledge, there is no report documenting the combination of electrodeposited graphene nanosheets (GNSs) combined with chitosan and [Ru(bpy) 3 ] 2+ for ECL diagnosis of BC. In this research, we aim to analysis of BC biomarker HER-2 by the ECL methodology enhanced by applying graphene nanomaterials. In this way, GNSs were electrochemically deposited onto the working electrode as supporting substrate. This deposition type led to more uniformity, stable and repeatable grafting of GNSs on the working electrode. Nasrollahpour et al. Cancer Nano (2021) 12:10 Methods

Reagents and procedures
HER-2 protein (ab60866) and HER-2 antibody (Ab, ab214275) were purchased from Abcam company. The solution of HER-2 protein was provided by dissolving of the protein (250 mg/mL) in deionized water (DI). All the standard solutions used for calibration curve were provided by dilution of the source protein solution with DI. The 20 µg/mL of HER-2 Ab was used in all experiments. All employed reagents were of high-pure grade and applied without any additional purification. Graphene oxide (GO) solution (2 mg/mL) and [Ru(bpy) 3 ] 2+ were purchased from sigma Aldrich Company. Chitosan (CS) powder, tripropylamine (TPrA), acetic acid, sulfuric acid, nitric acid, K 3 [Fe(CN) 6 ], KCl and EDC/NHS were acquired from Merck Company. HER-2 monoclonal antibody and its recombinant protein were obtained from Abcam Company. Phosphate buffer solutions (PBS) were prepared in room temperature and containing KCl (0.201 g/L), NaCl (8.000 g/L), Na 2 HPO 4 (3.581 g/L), KH 2 PO 4 (0.231 g/L), pH = 7.4. The other pH values were prepared by addition of 3 N solutions of NaOH and HCl. Chitosan solution was prepared by dissolving 0.1 mg chitosan powder in 10 mL acetic acid/water solution (1% v/v). In order to prepare the [Ru(bpy) 3 ] 2+ solution, 10 mg [Ru(bpy) 3 ] 2+ powder was dissolved in 2 mL DI.

Apparatus and electrodes
Electrochemical records were implemented employing a Metrohm Autolab from Nova software. The analysis system included a triple-electrode system with a glassy carbon electrode (as working electrode, 3 mm in diameter), Pt wire electrode (as counter electrode) and an Ag/AgCl (as reference electrode). All the detections, preparation of solutions, incubations, and deposition processes were performed in room temperature. An ultrasonic tool (Transsonic 420) was applied for surface cleaning of electrodes and suspending of solutions. A pH meter (Corning, 120) was used for adjusting the pH of PBS solutions. A magnetic stirrer model (Heidolph) was employed for mixing and preparation of all the solutions utilized in this study. A homemade ECL device was employed for ECL measurements (PMT, Hamamatsu, Japan). The ECL graphs were recorded by a homemade device coupled to autolab. The SEM images and semi-quantitative concentrations of the elements were obtained by TESCAN MIRA3, instrument.

Fabrication of biosensing platform
The working transducers were initially polished by using alumina powder (0.05 MICRON) to obtain mirror-like and clean surface. Then the electrode was ultrasonicated in water/acetone solution to eliminate the remaining materials on the electrode. Then, the electrode was electrochemically swept by immersing in H 2 SO 4 solution (0.5 M) in the range of 1 to − 1 V. After dehumidify with nitrogen stream, the bare electrode was dipped in the GO solution (0.1 mg/mL) followed by scanning with cyclic voltammetry between − 1.5 to 0.2 V (0.08 V/S) for 10 cycles. According to the literature (Ly et al. 2019;Zheng et al. 2015) GO is converted to its reduced form at much negative potentials around − 1 V vs Ag/Ag Cl. Additional file 1: Figure S1A represents the reduction peak at about − 0.7 V vs Ag/AgCl which can be correlated to the reduction of GO into rGO. The rGO modified electrode was placed into the deionized water for 10 min and dried in the incubator in the next. Afterward, the as prepared solutions of CS and [Ru(bpy) 3 ] 2+ mixed (4:6, µL:µL) were drop-cast and incubated on the rGO-GCE for 2 h at ambient conditions (25 °C). Next, 5 µL of EDC/NHS (2:1) was obtained from the previously prepared source solutions and added into the 5µL of Ab solution (18 µg/mL). The obtained solution was let to rest for 1 h to active the carboxylic groups of the Ab by EDC/NHS chemistry. The solution was then incubated onto the CS/[Ru(bpy) 3 ] 2+ /rGO/ GCE for 2 h. This led the Ab molecules to be immobilized onto the CS/[Ru(bpy) 3 ] 2+ / rGO/GCE via amide bonds with amine groups of CS.

ECL procedure
In order to recognition of target protein, the prepared electrode surface was firstly treated by PBS solution (pH = 7.4) for eliminating the unbound Abs. Then, different concentrations of HER-2 protein were incubated on the modified electrode. After incubating the modified electrode by various concentrations of HER-2 biomarker at laboratory temperature (20 ℃) for 2 h, the framework was immersed into a mixture of TPrA/PBS (2 mL/8 mL). The ECL evaluations were conducted by scanning in the range of 0.5 to 1.3 V (0.1 V s −1 ) using CV technique and simultaneously recording the intensity of the emitted photons via ECL detector (PMT). During the determination process, the ECL cell containing all three electrodes was placed on the PMT and totally covered to avoid environmental photonic noise. The ECL readouts were recorded at λ = 450 nm, correlating with the concentration of HER-2 protein.

Characterization of produced immunosensor
As a proof of concept, the step-by-step modification and application of the proposed biosensor was characterized by CV, EIS and ECL techniques, and also with imaging methods like SEM and EDX.

ECL investigation of the electrode modification steps
At first, the effect of rGO as a main electrode modifier was studied ( Fig. 1a, b). To this purpose, the ECL signals were measured in the presence and absence of rGO electrodeposition. The results showed very low ECL intensities in the absence of rGO electrodeposition which indicates the intense bleaching of the emitting materials. Furthermore, in the absence of the rGO, the ECL producing composite, chitosan and [Ru(bpy) 3 ] 2+ , was released from the electrode surface and the ECL signals are not stable. Also, the ECL characterization was depicted through obtaining ECL emission for step-by-step preparation of the developed ECL-based immunosensor ( Fig. 2a and b). The results illustrated that the ECL intensity was decreased after incubation of Ab and the target HER-2 protein (1 fM) which is in line with proofs. The decrease in the ECL readouts by Ab and target protein incubation resulted from the steric hindrance, which prevented the TPrA from reaching the electrode surface. This disrupts the reaction that results in ECL, ultimately reducing the intensity of ECL signals.

Morphological
The morphological features of the electrode surface were investigated in different stages using SEM microscope. Figure 3a-f represents the SEM images correlated to the bare electrode (3A-C) and rGO modified electrode (3D-F) at different magnitudes (200, 500 and 1000 nm). According to the SEM images, the rGO was electrodeposited in a proper manner on the bare electrode. The EDX patterns also indicated the presence of carbon and oxygen atoms after electrodeposition of the rGO, which represented the formation of oxygen groups onto the electrode. The quantitative elemental analysis is presented in Additional file 1: Table S1. As shown in Fig. 3d-f, the rGO was arranged is in a good quality.

Proof of principle
Benefiting the high conductivity, stability and desirable loading capacity of rGO along with high loading and adhesion capability of CS, the proposed biosensor provided a highly sensitive and stable biosensing assay. As shown in Additional file 1: Fig. S1B, the potential was screened from 0.5 to 1.3 V resulting an ECL pick current at about 0.75 V. Correspondingly, the ECL emission is only appeared in the presence of TPrA which represent the background signals to be ignored. The mechanism for ECL of [Ru(bpy) 3 ] 2+ could be established as follows Workman and Richter 2000): both [Ru(bpy) 3 ] 2+ and TPrA agents were oxidized on the surface of the electrode. This event appears as two oxidation peaks in the CV voltammograms at about 0.75 and 1.15 V. TPrA is reacted with [Ru(bpy) 3 ] 2+ after losing a proton to form an exited form of Ru(bpy) 3 2+* . The product emits ECL light and relief into the ground state for further reaction cycles.  Looking at the ECL plots in Fig. 5, there is a shift doubling of ECL peaks by implementing mixed sample. This time-dependent shift is related to manual control of the ECL signal processing system. In this ECL instrument, the average signal intensity was applied as quantitative analysis of target protein. As a common rule, the ECL signals corresponded to a cyclic voltammetry technique, which was applied as electrochemical excitation of ECL source. In cyclic voltammetric procedures, the electroactive material in the cell, in this case TPrA, commonly shows a sharp oxidation peak and a lower intensity ECL peak in the reduction process. This doubling of the peaks corresponded to this phenomenon. This phenomenon is presented in some cases like in BSA peaks. Additionally, this process occurred in the other peaks with low intensities.

Optimization of experimental conditions
We optimized various experimental parameters to certify our protocol to be proceed in its highest efficiency. High sensitivity of ECL, bring it to be sensitized against environmental conditions. In this regard, different parameters were optimized in different modification stages. For rGO deposition, the number of electrodeposition cycles was optimized. According to Additional file 1: Fig. S2A and S2B, the best conductivity of the deposited rGO was gained in 10 CV cycles from − 1.5 to 0.2 V (0.08 V s −1 ). It seems that by increasing the thickness of rGO on the electrode decrease the charge transferring capacity in the electrode-solution interface.
One of the important experimental factor in this study, as the ECL source, the CS/ [Ru(bpy) 3 ] 2+ volume ratio was optimized. In this regard, the proposed protocol was implemented in different ratios of CS/Ru (0:5, 2:8, 4:6, 5:5, 6:4, 8:2). These volumes were used from the unique concentrations of CS and [Ru(bpy) 3 ] 2+ solutions. The results indicated the best ECL for same volume ratio (5 µL:5 µL). This ratio was adjusted for subsequent experiments. This optimization step is depicted in Additional file 1: Fig. S3A and S3B.
The temperature of electrode environment has also impact on the signal yields. It could affect the stability of the platforms, their interactions and affinities. In this regard, the platform was incubated in six different temperatures (4, 10, 25, 30, 37 and 45 °C). According to the results shown in Additional file 1: Fig. S4A and S4B, enhancement in the temperature leads to a rise in the ECL readouts up to room temperature (25 °C) and remain constant for further enhancements. So, the next experiments were performed at 25 °C.
As a vital parameter, incubation time of [Ru(bpy) 3 ] 2+ /CS mixture (as a ECL source) was also optimized. The results represented that increasing the incubation times caused the luminescence intensity to rise up to 2 h, but more time had no enhancement effect on the signals (0, 30, 60, 90, 120 and 150 min). The reason could be that increasing the incubation time gives the luminescence integration more time to make more positions inside the rGO network, and also stronger and stable interaction with CS (as supporter). So, at the following experiments incubation time was set as 2 h. The resulted plots are indicated from Additional file 1: Fig. S5A and S5B.
In the following of factors' optimization process, we attempted to adjust the best concentration for coreactant (TPrA). This parameter was important from environmental and economic standpoint. First, the low usage of organic compounds, the reduction of harmful material release will be achieved. Second, the final cost will be decreased as low concentrations obtain as our optimum value. To this purpose, different TPrA/PBS ratios were examined (0.1:9.9, 0.5:9.5, 1:9, 1.5:8.5, 2:8, 3:7). According to the experiments, the best volume ratio of TPrA is find to be at 1.5:8.5. These results are presented in Additional file 1: Fig. S6A and S6B.
As can be seen previous section, H 3 O + is one of the species involved in ECL mechanism. So, optimization of pH range is of great significance for better conditions. In order to obtain the best pH range, 6 experiments were performed in different pH values (5, 7.2, 7.6, 8.2, 8.4, 9). The best result was fixed at moderate alkaline conditions (pH = 8.2). According to the reactions, increasing the concentration of H 3 O + (low pH) lead the TPrA • to be converted into [TPrA•] + and lowered in the concentration which resulted in the lower ECL intensity. Also, it should be mentioned that CS is soluble in low pH ranges (lower than about 5.3) which can be another reason for low intensities in acidic conditions. The results of pH effect on the ECL intensity are illustrated in Additional file 1: Fig. S7A and S7B.

Calibration curve
The calibration curve was drawn measuring the ECL signal readouts for 7 concentrations of the HER-2 protein (1, 0.1, 0.01, 0.001, 0.0001, 0.00001, 0.000001 nM). The LLOQ was practically obtained as (0.000001 nM). According to the analytical definition, sensitivity was calculated as the slope of calibration curve. The more the slope of the curve, the more of sensitivity will be reachable. In this regard, the calibration equation was obtained as Y = − 45.233X + 36.633 and based on this equation, the sensitivity was gained as slope of calibration curve (− 45.233) which means the prepared system is a sensitive protocol for analyzing the HER-2 protein. The calibration curve and related ECL graphs are shown in Fig. 4a and b.

Repeatability, reproducibility and stability investigations
In order to test the within-lab precision of the analytical data, five freshly prepared ECL biosensors (HER-2 concentration: 0.001 nM) was examined in their responses. Recording RSD = 3.1%, the proposed immunosensor had an acceptable repeatability. The reproducibility of the biosensor, as an important factor of reliability, was also examined by preparing six different immunosensors (three electrodes in two concentrations of 0.001 and 0.00001 nM). The obtained RSDs of 3.15 and 3.5% indicated the platform as a reproducible substrate for measuring purposes. The signal stability was also evaluated by calculating the RSD for 8 consecutive signal peaks in 1 fM concentration of HER-2 protein (Fig. 5c). From the obtained RSDs, the framework provided a desirable signal stability during the analysis, consequently, no leakage of ECL source [Ru(bpy) 3 ] 2+ was confirmed. Table 1 represents the analytical performance of the proposed nano-biosensing framework with the other related methods in evaluating the level of HER-2. Notice that the first two methods employed stripping voltammetry which a is more time-consuming methodology. Despite this fact, this work possesses a better LOD. The improved sensitivity could be viewed from several domains: first, the ECL method which is a highly sensitive one in the nature; second, the rGO film for its high surface area and conductivity, provided more and available sites for effectively holding of CS/[Ru(bpy) 3 ] 2+ composites onto the electrode. Using electrodeposition instead of simple incubation procedure lowered the preparation time and raised the signal stability and reproducibility. These are because of the uniform deposition of the rGO on the electrode, which might be the feature of the electrochemical deposition procedures. On the other hand, using CS as a biopolymer not only increases the stability of [Ru(bpy) 3 ] 2+ on the rGO/GCE, but also provides amine functions to covalently attach Ab molecules. This led to further stability of the ECL signals.

Selectivity
To interpret the appealing level of the proposed nano-biosensor toward HER-2, the effect of three possible interferences (BSA, CA-15-3 and CEA) on ECL signals was investigated. A comparative study was performed and the results are illustrated in Fig. 5a and b. It is apparent that there was a low ECL readout on all the three interferences. However, the target protein (HER-2) and the mixture of the interferences within HER-2 (all in 0.1 ng mL) resulted in a high ECL intensity.

Real sample screening
The practicality of the developed immunosensor was interpreted by analyzing the several real samples from different stages of the BC. The corresponding ECL graphs are depicted in Fig. 6a and b. In order to ascertain the credibility of the method, the results were compared to the pathological examinations. Pathological imaging revealed the existence of anaplastic changes indicated by the expression of HER-2 factors in the  (Fig. 6c). Similar to the results obtained by the designed immunosensor, the anaplastic changes were in accordance to the signals intensity. These data showed that by increasing the malignancy and local expansion of cancer cells inside the breast tissue, the serum levels of HER-2 factor were also heightened, showing a close relationship between the anaplastic changes and recorded ECL signals. These data show that our sophisticated modality is also eligible to diagnose the severity and progression of breast cancers, which is critical in cancer candidates. Like the previous studies, it was noticed that the HER-2 factor is released to the systemic circulation. Regarding the notable sensitivity of developed method, one can hypothesize that the current strategy is helpful in an early-stage detection of anaplastic changes soon after the cancer formation. It is important to declare that, the real samples were desirably analyzed without any treatment, in other words, the serum samples were evaluated in unprocessed forms. The signals intensity and the calculated concentrations of HER-2 protein are represented in Table 2. The concentration measures were calculated via the obtained calibration curve

Conclusion
In this research, we designed an ultrasensitive ECL immunosensing framework for determination of HER-2 protein biomarker in serum samples. In this regard, [Ru(bpy) 3 ] 2+ / TPrA was used as anodic ECL system. We employed electrodeposition method for uniform layer of rGO on the electrode. Using CS (as biocompatible polymer), the Ab molecules were strongly attached on the electrode via amide bounds.Consequently, the method indicated a good stability and reproducibility. Using rGO and CS increase the loading capacity for biorecognition elements (Abs) leading to a desirable LLOQ of 1 fM. After proving the high selectivity of the protocol for the target protein, the method was employed for analysis of real samples. Acquired results represented a good detection capability of the immunosensor for applying in clinical domains. Regarding the biocompatibility of the all employed materials in construction of the proposed nano-biosensor, the developed platform can be used in in vivo diagnosis tools.  Nasrollahpour et al. Cancer Nano (2021)