Impedimetric Graphene-Based Gold Immunosensor for Colorectal Cancer Biomarker CCSP‑2 Detection

Abstract

Electrochemical immunosensors are emerging as promising tools for cancer detection due to their simplicity, portability, and high sensitivity. Colorectal cancer (CRC), the third most common cancer in the United States, remains challenging to diagnose early, as the standard method, colonoscopy, is invasive and often avoided. To address this gap, a graphene-gold-based immunosensor was developed for the early detection of CRC by targeting colon cancer-secreted protein-2 (CCSP-2), a biomarker overexpressed in CRC patients. The sensor was fabricated by attaching graphene oxide (GO) to a gold (Au) electrode using 4-aminothiophenol (4-ATP) as a linker, followed by immobilization of CCSP-2 antibodies (Anti-CCSP-2) and blocking with bovine serum albumin (BSA). Characterization of the immunosensor using linear sweep voltammetry (LSV), Raman spectroscopy, cyclic voltammetry (CV), and electrochemical impedance spectroscopy (EIS) confirmed successful modifications. The attachment of graphene oxide contributed to an enhancement in current response, likely due to partial reduction and improved electron transfer at the modified surface. The sensor demonstrated a good linear response (R ² = 0.979) to CCSP-2 antigen (CCSP-2) concentrations ranging from 1 ng/μL to 100 ng/μL, with a limit of detection (LOD) of 0.17 ng/μL and a sensitivity of 0.031 (ng/μL)⁻¹. Selectivity was validated using CRC cell extracts (CACO-2) and human kidney extracts (HEK), showing a more significant signal for CACO-2. These findings suggest that the developed immunosensor is a reliable and sensitive platform for CCSP-2 detection, with the potential for adaptation as a point-of-care device for early CRC screening.

Introduction

Electrochemical biosensors are powerful tools for detecting biological interactions such as DNA sequences, enzymatic reactions, and antibody–antigen interactions. Among these, immunosensors have gained significant attention due to their high selectivity and sensitivity. Immunosensors utilize antibodies as bioreceptors, antigens as analytes, and conductive substrates as transducers, which amplify biological signals and facilitate their detection. These sensors are particularly suitable for electrochemical assays of biomarkers, offering precise and sensitive measurements. Immunosensors have found applications in early cancer detection, enabling improved sensitivity for identifying biomarkers associated with cancer. −

Colorectal cancer (CRC) is one of the cancers that can benefit from such advancements. Poor early detection of CRC remains a major challenge in medical diagnosis. CRC originates in the colon or rectum and is the third most common cancer diagnosed annually in the United States. , Early detection is crucial, as patients diagnosed at an early stage have a high survival rate and can receive effective curative treatments, significantly reducing cancer-related mortality. −

CRC progresses through distinct stages, beginning with stage 0, where abnormal cells are confined to the inner lining of the colon or rectum, and advancing to stage 4, where cancer cells spread to distant organs such as the liver or lungs. − Colonoscopy is currently the most widely used diagnostic tool for CRC. Although effective, colonoscopy is invasive, requires unpleasant bowel preparation, and can pose risks for patients with chronic health conditions. Additionally, it often detects CRC at later stages. −

This research aims to address these challenges by developing an electrochemical immunosensor for the early detection of CRC biomarker CCSP-2. The immunosensor fabrication involves Au electrode cleaning, 4-aminothiophenol (4-ATP) immobilization, covalent binding of graphene oxide (GO), Anti-CCSP-2 immobilization, and electrochemical detection of CCSP-2. CCSP-2 is a promising biomarker as it is typically absent in normal cells but is highly expressed in colon cancer cells. The immunosensor can directly detect the biomarker present in cancerous cells. −

4-ATP was selected as a bifunctional linker molecule for electrode surface modification due to its strong anchoring ability and reactive terminal groups. The gold electrode surface forms a stable self-assembled monolayer (SAM) when the thiol (−SH) group bonds strongly through Au–S interactions. The terminal amine (−NH₂) group allows carboxylated GO to attach covalently through carbodiimide coupling chemistry, resulting in efficient and stable GO binding. Moreover, the aromatic structure of 4-ATP facilitates charge transfer across the interface, enhancing the electrochemical signal. Previous studies, such as Rosario-Castro et al., demonstrated the effective use of 4-ATP in lithium sensing platforms involving Pt electrodes and carbon nanotubes, supporting its reliability in nanomaterial-modified systems. Motivated by these results, we employ 4-ATP to functionalize GO surfaces for antibody immobilization, establishing a conductive and stable biointerface for sensitive CCSP-2 detection.

The incorporation of graphene oxide in our immunosensor enhances the sensitivity of the sensor due to graphene oxide’s large surface area and solubility, excellent electronic properties, and high charge mobility. These characteristics allow subtle surface charge changes to be detected, correlating with biochemical interactions on the electrode surface. − Graphene oxide is particularly advantageous for electrochemical biosensors due to its high electron transfer rates, conductivity, robustness, and compatibility with functionalization through its hydroxyl, epoxy, and carboxyl groups. − The development of a gold-graphene-antibody array-based immunosensor offers an innovative diagnostic tool for CRC, providing faster and more sensitive detection of the CRC biomarker CCSP-2. This device has the potential to improve early detection of CRC and ultimately contribute to better patient outcomes.

While numerous studies have explored GO-modified gold electrodes for biosensing, our work distinguishes itself in several key ways. This study presents the first GO-based electrochemical immunosensor specifically targeting CCSP-2, a relatively new and clinically significant biomarker for early-stage colorectal cancer detection. Most existing sensors focus on traditional markers like carcinoembryonic antigen (CEA) or carbohydrate antigen 19-9 (CA19-9); CCSP-2 has not been previously integrated into GO-based sensor platforms. Our approach integrates covalent GO coupling via 4-ATP and provides comprehensive FTIR and Raman validation of surface modifications at each stage. The effective use of DCC/EDC/NHS chemistry and analysis of amide bond formation set this work apart in terms of material characterization depth. In addition to synthetic CCSP-2 standards, our sensor distinguishes between CCSP-2 expression in cancerous (CACO-2) vs noncancerous (HEK) cell samples, demonstrating real-world applicability and biological selectivity.

Experimental Section

Materials

The following reagents were used in this study (all purchased from Sigma-Aldrich, MO, USA): sulfuric acid (H₂SO₄, 99.99%), phosphoric acid (H₃PO₄, 85%), graphite powder (<20 μm diameter), anhydrous ethanol (CH₃CH₂OH, 99.5+%), 4-aminothiophenol (4-ATP, 97%), 1,3-dicyclohexyl-carbodiimide (DCC, 99%), hydrogen peroxide (H₂O₂, 35%), anhydrous N,N-dimethylformamide (DMF, 99.8%), potassium ferricyanide (K₃Fe­(CN)₆, 99+%), potassium ferrocyanide trihydrate (K₄Fe­(CN)₆·3H₂O, 98.5%), potassium permanganate (KMnO₄, 99.99%), 1-ethyl-3-(3-(dimethylamino)­propyl)­carbodiimide (EDC, 98%), N-hydroxysuccinimide (NHS, 99%), bovine serum albumin (BSA), potassium hydroxide (KOH, 99.99%), and phosphate-buffered saline (PBS). PBS was prepared by using sodium chloride (NaCl, 99.99%), potassium chloride (KCl, 99%), sodium phosphate monobasic (NaH₂PO₄, 99%), and potassium phosphate dibasic (K₂HPO₄, 99%). The solution was adjusted to pH 7.40 using sodium hydroxide (NaOH, 99.99%). Colon cancer-secreted protein-2 (CCSP-2) antigen and CCSP-2 antibody were diluted in 0.1 M PBS. For cell culture and protein extraction, Dulbecco’s Modified Eagle Medium (DMEM), fetal bovine serum (FBS), ethylenediaminetetraacetic acid (EDTA), trypsin-EDTA, (cholamidopropyl)­dimethylammonio-1-propanesulfonate (CHAPS), Trizma hydrochloride (Tris-HCl), magnesium chloride (MgCl₂), phenylmethanesulfonyl fluoride (PMSF), and glycerol were used. All aqueous solutions were prepared using nanopure water (18 MΩ·cm).

Equipment

Electrochemical measurements were performed using polycrystalline gold electrodes (1.6 mm diameter) as the working electrode, a platinum coil as the counter electrode, and a silver/silver chloride (Ag/AgCl, 3 M NaCl) reference electrode, all obtained from Bioanalytical Systems (BASi, IN, USA). An Autolab M204 multichannel potentiostat was used for the electrochemical measurements.

Raman spectroscopy was performed using a Thermo Scientific DXR3xi Raman Imaging Microscope to analyze clean and GO-modified Au surfaces, as well as pure GO. The measurements were conducted with a 532 nm excitation wavelength and a 50× objective lens for spectral acquisition. For the characterization of the bare Au surface and the formation of the GO monolayer on Au, a Au substrate on silicon was used.

Spectra were collected under identical conditions to ensure consistency in the analysis. Fourier-transform infrared (FTIR) spectroscopy was performed by using a PerkinElmer Spectrum Two FTIR spectrometer equipped with a diamond attenuated total reflectance (ATR) accessory. The technique was used to characterize each modification step of the gold electrode, including Au/4-ATP, Au/4-ATP/GO, Au/4-ATP/GO/Anti-CCSP-2, and the pure precursors. Au substrates were gently rinsed with nanopure water and dried under a nitrogen stream before measurement. Spectra were recorded over the 4000–400 cm^–1 range at a spectral resolution of 4 cm^–1, with 32 scans averaged per sample to improve the signal-to-noise ratio.

Electrochemical Measurements

The Au electrodes were prepared through a combination of mechanical polishing and electrochemical cleaning. Initially, the electrodes were mechanically polished using 0.05 μm alumina (α-Al₂O₃) micropolishing paste (Buehler) for 10 min, or until a mirror-like surface was achieved. Subsequently, the electrodes were sonicated in deionized water for 2 min to remove any residual alumina particles. The program for the electrochemical measurements was Nova 2.1.5. The electrochemical exfoliation was performed with a 0.5 M H₂SO₄ solution with potential cycling between 0.2 and 1.5 V vs Ag/AgCl. The scan rates were systematically decreased from 300 mV/s to 200 mV/s and finally to 100 mV/s, continuing until the cyclic voltammograms displayed reproducible redox peaks characteristic of clean polycrystalline gold surfaces.

For reductive desorption analysis, a 0.1 M ethanolic KOH solution was used in LSV experiments. LSV measurements were performed within a potential window of −0.2 V to −1.3 V vs Ag/AgCl (3 M NaCl) at a scan rate of 10 mV/s. Before each measurement, the solution was purged with N₂ gas for 10 min to remove oxygen. This analysis was performed on electrodes modified with 4-ATP.

All CV and EIS experiments were conducted in a 2 mM K₃[Fe­(CN)₆]/K₄[Fe­(CN)₆] solution containing 0.1 M PBS. For CV measurements, data were collected within a voltage window of −0.2 V to +0.6 V at a scan rate of 10 mV/s over five cycles. EIS measurements were carried out at an open-circuit potential of 0.240 V versus Ag/AgCl (3 M NaCl). The impedance spectra were recorded over a frequency range of 0.1 MHz to 0.1 Hz using 40 frequency points. The amplitude of the AC signal was set to 10 mV. All experiments were conducted at room temperature without purging the solution with nitrogen.

Immunosensor Preparation

Graphene Oxide (GO) Synthesis

The modified method by Marcano et al. and Cunci et al. was utilized for the preparation of graphene oxide. , Briefly, 100 mg of graphite powder was mixed with 40 mL of concentrated H₂SO₄ and H₃PO₄ in a 9:1 ratio under constant stirring. Subsequently, 600 mg of KMnO₄ was slowly added to the mixture, and the solution turned dark green. The reaction mixture was stirred at 50 °C for 18 h.

After the completion of the reaction, the mixture was poured into 400 mL of ice-cold nanopure water, followed by the addition of 6 mL of 35% H₂O₂, which turned the mixture yellow. The yellow solution was heated until boiling occurred. The resulting mixture was filtered by using a nylon membrane with 0.8 μm pores. The brown slurry precipitate obtained was dispersed in 80 mL of a 1:1 mixture of dimethylformamide (DMF) and nanopure water for 30 min.

The dispersion was divided into two 50 mL Falcon tubes and centrifuged at 3400 rpm for 8 min. The supernatant was discarded, and the resulting black pellet was washed with deionized nanopure water. The suspension was then filtered using a polycarbodiimide membrane with 0.45 μm pores. The final black-brown solid product was dried under vacuum for 24 h. The process is illustrated in Figure .

1.

Schematic representation of the steps involved in the synthesis of graphene oxide using the modified methods of Marcano et al. and Cunci et al. Each step highlights the key chemical transformations and processing stages, including oxidation, washing, filtration, and drying (image created by the author in Supporting Information https://www.biorender.com).

Au Electrode Modification Steps

The synthesized graphene oxide was utilized for developing the immunosensor. In Figure , the development of the immunosensor is illustrated. Cleaned Au electrodes were first immersed in a 0.1 M ethanolic solution of 4-ATP for 24 h to allow self-assembly. After immobilization, the electrodes were removed from the solution, rinsed thoroughly with nanopure water, and dried gently with nitrogen gas.

2.

Schematic illustration of the stepwise preparation of the immunosensor for CCSP-2 detection with key modification steps.

The Au/4-ATP electrodes were then modified with GO following the procedure previously established by Santiago et al. using single-walled carbon nanotubes (SWCNTs). Briefly, the electrodes were incubated in a solution containing 0.2 mg/mL GO and 1.0 mg/mL N,N′-dicyclohexylcarbodiimide (DCC) dissolved in dimethylformamide (DMF) at 55 °C for 18 h, after which the electrodes were rinsed with nanopure water and gently dried with nitrogen gas.

Subsequently, Anti-CCSP-2 was covalently cross-linked to GO using EDC/NHS carbodiimide chemistry. A solution of 5 mM EDC, 10 mM NHS, and 10 ng/μL Anti-CCSP-2 was prepared. Ten microliters of this solution was added to the modified Au electrode and incubated at 4 °C for 2 h to facilitate covalent binding; afterward, the electrodes were rinsed with nanopure water and dried gently with nitrogen gas.

To block nonspecific binding sites, 10 μL of a 0.1%(w/v) BSA solution was added to the modified electrode and incubated for 30 min at 4 °C, and then, the electrodes were rinsed with nanopure water and gently dried with nitrogen gas. Following this step, the immunosensor was tested for the detection of CCSP-2.

Calibration Curve

To construct the calibration curve, the immunosensor (Au/4-ATP/GO/Anti-CCSP-2/BSA) was tested with 10 μL drops of CCSP-2 at concentrations of 1, 10, 25, 50, and 100 ng/μL. The CCSP-2 solutions were prepared through serial dilution, and the immunosensors were incubated at 4 °C for 1 h for each concentration. For each CCSP-2 concentration, freshly prepared immunosensors were used, and triplicate measurements were performed to ensure reproducibility. The changes in the surface properties of the immunosensor were analyzed using EIS. The charge transfer resistance (R _ct) was determined by fitting the resulting Nyquist plots to the Randles equivalent circuit shown in Figure S4. , The relative charge transfer resistance (ΔR _ct/R _ct(BSA)) for each CCSP-2 concentration was calculated using eq to quantify the immunosensor response:

$$\frac{\mathrm{\Delta }{R}_{\mathrm{c}\mathrm{t}}}{R_{\mathrm{c}\mathrm{t}(\mathrm{B}\mathrm{S}\mathrm{A})}}=\frac{R_{\mathrm{c}\mathrm{t}(\mathrm{C}\mathrm{C}\mathrm{S}\mathrm{P}-2)}-R_{\mathrm{c}\mathrm{t}(\mathrm{B}\mathrm{S}\mathrm{A})}}{R_{\mathrm{c}\mathrm{t}(\mathrm{B}\mathrm{S}\mathrm{A})}}$$ 1

Here, R _ct(BSA) represents the charge transfer resistance of the immunosensor (Au/4-ATP/GO/Anti-CCSP-2/BSA), while R _ct(CCSP‑2) denotes the charge transfer resistance of the immunosensor after binding with CCSP-2. The calculation of ΔR _ct/R _ct(BSA) provided a quantitative measure of the sensor’s response to varying concentrations of CCSP-2, forming the basis for the calibration curve.

Cell Culture and Cell Extract Preparation Procedure

Human colorectal cancer cells (CACO-2, ATCC HTB-37) and kidney cells (HEK-293, ATCC HTB-293) were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). CACO-2 cells were selected as a positive model due to their overexpression of CCSP-2, which is a characteristic feature of colorectal cancer, making CCSP-2 the target biomarker for the developed immunosensor. Conversely, HEK-293 cells, which exhibit low or negligible expression of CCSP-2, were used as a negative model to evaluate the selectivity of the immunosensor for colorectal cancer.

Both cell lines were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) under standard conditions (37 °C and 5% CO₂). To subculture the cells, the medium was aspirated, and the cell cultures were washed with 1× phosphate-buffered saline (PBS) containing 0.25% trypsin. Subsequently, 3 mL of trypsin-EDTA solution was added, and the flasks were incubated at 37 °C until the cells detached. Fresh DMEM was added to the new culture flasks, and the cells were cultured until reaching 80% confluency and were metabolically active, at which point they were suitable for protein extraction. For optimal protein extraction, a cell concentration of 25 × 10⁶ cells/mL was used, although 1 × 10⁶ cells/mL was sufficient for some procedures.

To prepare cell extracts, cells were harvested once they reached the exponential growth phase and washed twice with ice-cold 1× PBS (140 mM NaCl, 2.7 mM KCl, 10 mM NaHPO₄, and 1.8 mM KH₂PO₄). The cells were then resuspended in ice-cold lysis buffer, which was composed of 10 mM Tris-HCl (pH 7.5), 1 mM MgCl₂, 1 mM ethylenediaminetetraacetic acid (EDTA), 0.1 mM phenylmethylsulfonyl fluoride (PMSF), 0.5% CHAPS, and 10% glycerol. The cell suspension in lysis buffer was incubated on ice for 30 min to ensure lysis. Following lysis, the suspension was centrifuged at 15,000 rpm for 20 min at 4 °C. The resulting supernatant, containing the cell extracts, was carefully transferred to a fresh tube and stored at −80 °C until use. These cell extracts served as the source of CCSP-2 for immunosensor evaluation of selectivity.

Results and Discussion

Preparation of Immunosensor Substrate by Electrochemical Cleaning of Au Electrode

The preparation of the graphene-based immunosensor for CCSP-2 detection began with electrochemical cleaning of the Au electrode substrate. This process was carried out using CV in 0.5 M H₂SO₄. The electrochemical cleaning effectively activated the polycrystalline gold surface, as evidenced by the emergence of three characteristic Au oxidation peaks. These peaks appeared at potentials of 0.210, 0.340, and 0.430 V vs Ag/AgCl (3 M NaCl), confirming the successful cleaning of the Au electrode.

The voltammograms, displaying these characteristic peaks recorded with the Au electrode, are provided in Figure S1. This step ensured a clean and reproducible surface for subsequent biosensor fabrication and functionalization.

Modification of Au Electrode with 4-ATP (Au/4-ATP)

Linear sweep voltammetry (LSV) is an electrochemical technique that can be used to study the desorption of thiol molecules from gold surfaces through a one-electron reduction process, as described by eq :

$$\mathrm{Au}-\mathrm{SR}+{\mathrm{e}}^{-}\to {\mathrm{Au}}^{°}+{\mathrm{RS}}^{-}$$ 2

The surface coverage (Γ) was calculated using eq :

$$\mathrm{\Gamma }=\frac{\mathit{Q}}{\mathit{nFA}}$$ 3

where Γ represents the surface coverage (mol/cm²), Q is the charge corresponding to the reductive peak (C), n is the number of electrons involved in the desorption process (n = 1), F is the Faraday constant (F = 96485 C/mol), and A is the geometric area of the electrode (cm² ). ,

This process is characterized by the appearance of cathodic peaks in the LSV data, with the peak potentials and areas providing valuable information about the molecules adsorbed on the surface. In this study, an LSV experiment was performed on a 4-ATP-modified Au electrode to observe its reductive desorption behavior. The experiment consisted of five scans conducted in 0.1 M KOH within a potential range of −0.2 V to −1.3 V at a scan rate of 10 mV/s.

During the first scan, two distinct reductive peaks were observed, as shown in Figure . These peaks were located at approximately −1030 mV (peak 1) and −880.0 mV (peak 2). Peak 1 is attributed to the reductive desorption of 4-ATP molecules from the Au surface, confirming their successful adsorption. The origin of peak 2 is less well understood; however, it has been suggested by Hayes et al. that it may result from the desorption of a dimer of 4-ATP molecules, which forms through the oxidation of surface-confined 4-ATP to a cation radical that subsequently reacts with a neighboring 4-ATP molecule to form a dimer.

3.

Linear sweep voltammetry (LSV) of a 4-ATP-modified gold electrode in 0.1 M KOH, recorded at a scan rate of 10 mV/s within a potential window of −200 to −1300 mV.

The surface coverage of 4-ATP on the Au surface was calculated to be 8.5 × 10^–4 mol/cm². In subsequent scans (scans 2 to 5), the LSV data exhibited featureless peaks, indicating that the Au surface was devoid of 4-ATP molecules after their complete desorption during the initial scan. This experiment confirms the formation of a 4-ATP monolayer on the Au surface.

Graphene Oxide on 4-ATP-Modified Au Surface (Au/4-ATP/GO)

Raman spectroscopy was employed to characterize the synthesized graphene oxide (GO) and GO on Au/4-ATP. The characterization of the synthesized GO is shown in Figure S2. GO exhibits distinct Raman bands, providing insight into its structure. Figure presents the Raman spectrum of pure GO, showing characteristic bands: the D band at 1,349 cm^–1, corresponding to the A_1g phonon vibrational mode associated with sp³ carbons in defective graphitic regions; the G band at 1590 cm^–1, corresponding to the E_2g phonon vibrational mode linked to sp² carbon atoms; and the 2D band at 2703 cm^–1, indicative of graphene layers. These features are consistent with the expected Raman signatures of GO.

4.

Raman spectra of (A) synthesized pure graphene oxide (GO), (B) clean gold (Au) surface (black line), and Au/4-ATP/GO (red line).

Figure B compares the Raman spectrum of bare gold (black line) with that of the GO-modified Au/4-ATP electrode (red line). The Raman spectrum of clean bare gold is featureless, as expected, due to the nonresonant nature of gold in this spectral region. In contrast, the Raman spectrum of the GO-modified Au/4-ATP electrode displays the same characteristic bands observed in pure GO, with the D, G, and 2D bands shifted to 1339 cm^–1, 1595 cm^–1, and 2712 cm^–1, respectively. These spectra confirm the successful formation of a GO layer on the Au/4-ATP surface.

FTIR Spectroscopic Analysis of Electrode Surface Modification and Antibody Functionalization

Fourier-transform infrared (FTIR) spectroscopy was applied to validate the sequential functionalization of the gold electrode and confirm the successful immobilization of the Anti-CCSP-2 antibody. The FTIR spectra, shown in Figure , illustrate the infrared spectrum of all the Au surface modifications. The spectrum of the Au/4-ATP surface displays distinct peaks characteristic of 4-ATP, including a broad asymmetric band at 3419 cm^–1, symmetric band at 3286 cm^–1 of NH₂ stretching vibrations, and 3193 cm^–1 vNH₂, as well as strong δNH bending vibrations in 1624 cm^–1 and aromatic νCC stretching at 1595 cm^–1, νCC + δNH 1489 cm^–1, νCH_(aliphatic) at 1285 cm^–1, δCH_(aliphatic) at 1178 cm^–1, and δCS_aromatic at 1084 cm^–1. Au/4-ATP shows the same vibrations as the pure 4-ATP spectrum in Figure S3. There is no presence of νSH on the Au-modified surface. These signals confirm that the 4-ATP linker is bonded to the gold electrode surface.

5.

FTIR characterization of the stepwise functionalization of the gold electrode surface for CCSP-2 immunosensor construction. (A) Full-range FTIR spectra of Au/4-ATP, Au/4-ATP/GO, and Au/4-ATP/GO/Anti-CCSP-2. (B) In situ view of the amide region (1700–1200 cm^–1) of the Au/4-ATP/GO and Au/4-ATP/GO/Anti-CCSP-2 spectrum, showing the emergence of amide I (∼1650 cm^– 1), amide II (∼1590 cm^– 1), and amide III (∼1252 cm^– 1) bands.

Upon functionalization with graphene oxide (GO), new features emerged in the FTIR spectrum of the Au/4-ATP/GO. A broad and intense band at 3385 cm^–1 representing vOH stretching, and at 1088 cm^–1 assigned to νC–O stretching of epoxy and alkoxy groups in GO, confirmed the successful deposition of GO onto the aminothiophenol-modified surface. In addition to the νC–O vibrations, weak but distinct amide bands between 1650 and 1300 cm^–1 were observed in this spectrum. These amide-related signals likely arise from the formation of an amide bond between the carboxylic acid (COOH) groups of GO and the (NH₂) amines of 4-ATP, activated by dicyclohexylcarbodiimide (DCC) coupling. The appearance of an amide I band around 1653 cm^–1 and a weaker amide II band near 1596 cm^–1 supports this covalent linkage. There is an unknown band at 2088 cm^–1, which may correspond to metal carbonyl formation.

Following the immobilization of Anti-CCSP-2 antibodies using EDC/NHS cross-linking, these amide bands became significantly more pronounced. The spectrum of the final modification (Au/4-ATP/GO/Anti-CCSP-2) showed a stronger and well-defined amide I band at 1659 cm^–1, but the amide II in 1593 cm^–1 and the amide III 1255 cm^–1 bands diminished. The amide I band corresponds to νCO stretching, amide II to δN–H bending and νC–N stretching, and amide III to in-plane δN–H bending coupled with νC–N stretching, all indicative of protein presence. The increased intensity and clarity of these bands confirm the successful immobilization of the antibody and the complete assembly of the biosensing interface necessary for CCSP-2 detection.

Analysis of Immunosensor Fabrication and CCSP-2 Detection Using Cyclic Voltammetry and Electrochemical Impedance Spectroscopy

A self-assembled monolayer (SAM) of 4-ATP was formed on the Au electrode surface to facilitate the covalent immobilization of GO. Subsequently, EDC/NHS carbodiimide chemistry was employed to attach Anti-CCSP-2 onto the Au/4-ATP/GO surface. Figure illustrates the CV responses of the bare Au electrode, Au/4-ATP, and Au/4-ATP/GO surfaces. The voltammogram of the clean bare gold electrode exhibits a characteristic duck-shaped redox peak, indicative of efficient electron transfer of the [Fe­(CN)₆]^3–/4– redox species in the PBS solution and at the electrode surface. Upon immobilization of the 4-ATP SAM, the redox currents decreased due to the blocking effect of the SAM, which impedes electron transfer of the redox species at the electrode surface. In Figure S5, different concentrations of 4-ATP immobilization were tested, and a 100 mM concentration was chosen due to its high amine density, which is critical for efficient and reproducible covalent coupling with GO via EDC/NHS chemistry. Therefore, 100 mM was selected as the optimal concentration.

6.

Electrochemical characterization of the sensor assembly process. (A–C) Cyclic voltammetry (CV) profiles of bare and modified Au electrodes at each surface modification step. (D–F) Corresponding Nyquist plots (EIS) illustrating changes in charge transfer resistance. All measurements were performed in 2 mM [Fe­(CN)₆]­3 ^– /4^– with 0.1 M PBS (pH 7.4) using Ag/AgCl (3 M NaCl) as a reference electrode.

Following the attachment of GO, the redox currents increased significantly, which is attributed to the electron donor/acceptor properties of GO. This enhancement arises from the conjugated system within GO, which facilitates electron transport through the occupied and unoccupied states. The primary charge transport mechanism in GO involves backscattering or electron hopping across defect sites. Also, the thermal steps after stopping the reaction during our GO synthesis induced partial reduction of GO, and the strong π–π stacking or covalent coupling between GO’s aromatic domains and the 4-ATP linker facilitates charge transfer at the interface.

Figure shows a further decrease in the redox current after the immobilization of Anti-CCSP-2 onto the Au/4-ATP/GO surface. This reduction is attributed to electrostatic repulsion between the negatively charged redox species and the negatively charged carboxylic groups (COOH) in the Anti-CSSP-2, which partially block the redox species’ access to the electrode surface. Subsequent addition of BSA resulted in additional current reduction, as BSA blocks unoccupied sites on the electrode surface, ensuring that CCSP-2 only binds to Anti-CCSP-2 rather than to pinholes or uncoated areas on the surface. When CCSP-2 was introduced to the fully functionalized immunosensor (Au/4-ATP/GO/Anti-CCSP-2/BSA), a further significant reduction in current was observed. This decrease confirms the specific binding of CCSP-2 to the immobilized antibodies, with BSA effectively blocking nonspecific interactions. These results demonstrate the successful fabrication of the immunosensor and its ability to detect CCSP-2 efficiently.

The typical impedance behavior of a heterogeneous electron transfer process is represented in Nyquist plots of EIS data, where the imaginary component of impedance (−Z’’) is plotted against the real component (Z’). The high-frequency region of the plot forms a semicircle corresponding to the electron transfer process, and the low-frequency region forms a straight line indicative of diffusion control. The diameter of the semicircle is directly proportional to the charge transfer resistance (R _ct), a key parameter used to analyze electrode surface modifications.

Figure D shows the Nyquist plot of the overlapped Nyquist plots for bare Au, Au/4-ATP/GO, and Au/4-ATP/GO/Anti-CCSP-2, while Figure E displays only Au/4-ATP. Figure F represents, once again, the overlapped Nyquist plots for Au/4-ATP/GO/Anti-CCSP-2/BSA (immunosensor) and Au/4-ATP/GO/Anti-CCSP-2/BSA/CCSP-2. After 4-ATP immobilization, there was an increase in R _ct due to the blocking effect of the SAM, which restricts electron transfer of the [Fe­(CN)₆]^3‑/4– redox species. However, the attachment of GO to the Au/4-ATP surface facilitated electron transfer, as GO introduces pathways for electron hopping, resulting in a reduced diameter of the semicircle.

Also evaluated is the reproducibility of the GO modification on Au electrodes by performing EIS measurements on three independently prepared Au/4-ATP/GO electrodes. The measured charge transfer resistance (R _ct) values for these electrodes were 1125  Ω, 1190  Ω, and 1143  Ω, yielding an average R _ct of 1152 Ω with a standard deviation of 34 Ω. This corresponds to a relative standard deviation (RSD) of approximately 2.95%, demonstrating that our GO functionalization procedure produces highly consistent electrode surfaces. This low RSD (∼3%) indicates excellent reproducibility of the nanomaterial-based surface modification, well within acceptable limits for sensors of this type. R _ct increased further after Anti-CCSP-2 immobilization on Au/4-ATP/GO, reflecting steric and electrostatic repulsion caused by the negatively charged amino and carboxylic groups in the antibody. This hinders electron transfer between the base electrode and the redox probe in solution. The addition of BSA to block nonspecific binding sites further increased R _ct, as BSA effectively covers pinholes and unmodified regions on the electrode surface.

Finally, when the immunosensor (Au/4-ATP/GO/Anti-CCSP-2/BSA) was incubated with CCSP-2, a significant increase in R _ct was observed. This increase confirms the specific binding of CCSP-2 to the immobilized Anti-CCSP-2. The experiment was repeated three times to verify the reproducibility of the impedance signal at the same CCSP-2 concentration, demonstrating consistent and reliable results.

In Figure , the complete immunosensor yielded the highest ΔR _ct/Rct_initial value, indicating a strong increase in charge transfer resistance upon antigen exposure. This is attributed to the specific immunorecognition between CCSP-2 and Anti-CCSP-2 immobilized on the GO-functionalized gold electrode. The formation of the antigen–antibody complex introduces a significant steric and electrostatic barrier at the electrode interface, limiting the access of the negatively charged [Fe­(CN)₆]^3–/^4– redox probe to the gold surface. Furthermore, the incorporation of BSA serves to block unreacted functional sites and prevent nonspecific interactions, resulting in a clean and highly specific sensor response.

7.

Normalized change in charge transfer resistance (ΔR _ct/Rct_initial) measured by EIS for different electrode configurations after incubation with 50 ng/μL CCSP-2 antigen: Au/4-ATP/GO/Anti-CCSP-2/BSA (complete immunosensor), Au/4-ATP/GO/Anti-CCSP-2 (no BSA), and Au/4-ATP/GO/BSA (no antibody).

In contrast, the surface modified with the antibody but lacking BSA blocking exhibited a moderately high ΔR _ct/Rct_initial, albeit with a larger standard deviation. This suggests that while some level of specific antigen binding occurs, the absence of BSA allows nonspecific adsorption of proteins and redox species, contributing to an elevated baseline and reduced reproducibility. This underscores the crucial role of BSA in minimizing background signal and enhancing the reliability of the impedance measurement. The surface containing BSA but no antibody showed the lowest impedance response, indicating minimal interaction with CCSP-2 at 50 ng/μL. This control confirms that CCSP-2 binding is antibody-specific and that the observed signal in the complete sensor is not due to nonspecific adsorption onto the GO or BSA layer. Moreover, this configuration demonstrates that BSA itself does not significantly alter the electrochemical properties of the surface postblocking.

Calibration Curve and Limit of Detection

To assess the reproducibility and sensitivity of the immunosensor, triplicate experiments were conducted with various concentrations of CCSP-2 to confirm its binding to Anti-CCSP-2 immobilized on the Au electrode surface. A calibration curve was generated to establish the system’s detection limit and to determine the concentration range where the immunosensor demonstrates effective signal changes. The calibration curve ensures reproducibility and provides precise and sensitive data.

The CCSP-2 concentrations used to construct the calibration curve were 1 ng/μL, 10 ng/μL, 25 ng/μL, 50 ng/μL, and 100 ng/μL. For each measurement, freshly cleaned and modified electrodes were prepared by functionalizing them to form the biosensor (Au/4-ATP/GO/Anti-CCSP-2/BSA). The relative change in charge transfer resistance, ΔR _ct/R _ct(BSA), was calculated for each concentration. The ΔR _ct/R _ct(BSA) values were plotted against CCSP-2 concentrations to generate the calibration curve (Figure ). Supporting data, including individual ΔR _ct/R _ct(BSA) values for each electrode and concentration, are provided in Table S1.

8.

(A) Nyquist plots showing EIS response of the immunosensor at increasing CCSP-2 concentrations (1–200 ng/μL) in 2 mM [Fe­(CN)₆]­3 ^– /4^– with 0.1 M PBS. (B) Calibration curve of ΔR _ct/R _{ct_BSA} versus CCSP-2 concentration from 1 to 100 ng/μL. The sensor shows a linear response with R² = 0.979 and a calculated LOD of 0.17 ng/μL.

The calibration curve (Figure B exhibits a well-defined trend, demonstrating a good correlation between ΔR _ct/R _ct(BSA) and CCSP-2 concentrations (R ² = 0.979). At lower concentrations (1 ng/μL, 10 ng/μL, and 25 ng/μL), the signal change was minimal, likely reflecting the detection limit of the system. However, these points displayed a low standard deviation, indicating good reproducibility. At higher concentrations (50 ng/μL and 100 ng/μL), significant increases in ΔR _ct/R _ct(BSA) were observed, confirming the sensor’s ability to reliably measure CCSP-2 concentrations of 25 ng/μL and above. The immunosensor exhibited a linear increase in ΔR _ct/R _ct(BSA) between 1 and 100 ng/μL of CCSP-2, as shown in Figure A, confirming a good linear relationship within this range. Although ΔR _ct/R _ct(BSA) values were also recorded for higher concentrations (150 and 200 ng/μL), these points were not included in the regression analysis due to a visibly reduced slope, which suggests the onset of surface saturation, where available antibody binding sites become increasingly occupied. Therefore, the linear dynamic range of the sensor is defined as 1–100 ng/μL.

The calculated slope of the calibration curve (m = 0.0311 ng/μL^–1) highlights the sensor’s high sensitivity. The limit of detection (LOD) was calculated as described in the studies by Armbruster et al. First is calculated the low limit blank (LOB) in eq , and then is calculated the LOD in eq :

$${\mathrm{LOB}=\mathrm{mean}}_{\mathrm{blank}}+1.645({\mathrm{S}}_{\mathrm{blank}})$$ 4

$$\mathrm{LOD}=\mathrm{LoB}+1.645({\mathrm{S}}_{\mathrm{lowest\; concentration\; sample}})$$ 5

The determined LOB was 0.12 ng/μL and the LOD was 0.17 ng/μL, demonstrating the immunosensor’s ability to detect low concentrations of CCSP-2. These results confirm that the functionalized Au electrode (Au/4-ATP/GO/Anti-CCSP-2/BSA) is a reliable and effective tool for CCSP-2 detection and can be used to diagnose colorectal cancer.

Additionally, a theoretical LOD (LOD_Theoretical) was calculated using the IUPAC-recommended regression approach:

$${\mathrm{LOD}}_{\mathrm{Theoretical}}=3{.3\times \mathrm{S}}_{\mathrm{blank}}/\mathrm{Slope}$$ 6

Using the standard deviation of the blank (S_blank = 0.04) and the slope of the calibration curve (0.0311 (ng/μL)⁻¹), this yielded a theoretical LOD of 4.24 ng/μL. The experimentally determined LOD of 0.17 ng/μL was found to be lower than the theoretical LOD of 4.24 ng/μL, which was calculated by using the IUPAC regression-based approach. This difference may be attributed not only to the limited number of replicates used in the experimental determination (n = 3) but also to the high signal-to-noise ratio observed at the lowest tested concentration. The theoretical method provides a conservative estimate based on blank signal variability and does not account for sensor-specific effects such as enhanced charge transfer, efficient surface functionalization, or favorable antigen–antibody interactions. Reporting both values provides a more comprehensive view of the sensor’s analytical performance, capturing both its statistical projection and practical sensitivity.

In our previous study, a label-free impedimetric immunosensor utilizing a bare Au electrode was developed for the detection of CCSP-2, achieving a limit of detection (LoD) of 0.71 ng/μL. In the current work, modification of the Au electrode with GO yielded a significantly lower LOD of 0.17 ng/μL, highlighting the enhanced analytical performance of the GO-functionalized Au sensor platform. While the improvement in LOD may appear moderate, the introduction of GO provided essential advantages, including increased effective surface area and a higher density of reactive oxygen-containing functional groups. These features promote improved antibody immobilization and enhance the stability of the sensing layer, contributing to more consistent and reliable signal responses. Furthermore, GO-modified electrodes exhibited reduced nonspecific binding and better reproducibility in repeated measurements, which are essential parameters for biosensor performance, particularly in complex matrices like cell extracts or blood serum. These findings underscore the importance of surface engineering with nanomaterials, such as GO, for achieving sensitive, robust, and clinically relevant immunosensors.

Table summarizes the EIS-extracted parameters. R _ct increased progressively with higher CCSP-2 concentrations, reflecting antigen–antibody binding at the electrode surface. Concurrently, the constant phase element (CPE) values decreased, consistent with the formation of a more resistive protein layer that reduces capacitive behavior. The extracted χ² values ranged from 0.007 to 0.08. Although the literature suggests that χ² values below 10^–3 indicate highly accurate fits, in the context of complex biosensor systems incorporating nanomaterials and biomolecules, χ² values on the order of 10^–2 are still considered acceptable, especially when the chosen equivalent circuit reflects the physical processes and the residuals are random and unbiased. The fitting error reflects the degree of overlap between the simulated response of the equivalent circuit and the experimental Nyquist plot. In our analysis, the fitting errors were consistently below 2%, confirming that the chosen Randles circuit closely reproduced the impedance spectra. This low deviation supports the good reliability of the measurements.

1. EIS Fitting Data of the Immunosensor Measurements.

Trial	R ct(Ω)	Error(%)	CPE(F)	χ 2
Immunosensor: No 1 ng/μL CCSP-2	1.53 × 105	0.650	8.25 × 10–7	0.062
Immunosensor: 1 ng/μL CCSP-2	2.00 × 105	0.495	8.75 × 10–7	0.073
Immunosensor: No 10 ng/μL CCSP-2	1.86 × 105	1.76	5.75 × 10–7	0.083
Immunosensor: 10 ng/μL CCSP-2	2.73 × 105	1.67	6.56 × 10–7	0.055
Immunosensor: No 25 ng/μL CCSP-2	2.79 × 105	0.511	3.32 × 10–7	0.0074
Immunosensor: 25 ng/μL CCSP-2	5.27 × 105	0.703	3.28 × 10–7	0.011
Immunosensor: No 50 ng/μL CCSP-2	1.74 × 105	0.694	2.85 × 10–7	0.017
Immunosensor: 50 ng/μL CCSP-2	6.54 × 105	0.811	2.72 × 10–7	0.015
Immunosensor: No 100 ng/μL CCSP-2	2.98 × 105	0.748	2.85 × 10–7	0.017
Immunosensor: 100 ng/μL CCSP-2	1.35 × 106	1.418	2.89 × 10–7	0.029
Immunosensor: No blank	2.88 × 105	1.37	1.74 × 10–7	0.024
Immunosensor: blank	2.88 × 105	0.962	2.06 × 10–7	0.032
Immunosensor: No HEK lysate	1.77 × 105	0.793	2.63 × 10–7	0.022
Immunosensor: HEK lysate	3.58 × 105	1.38	3.41 × 10–7	0.049
Immunosensor: No CACO-2 Lysate	1.33 × 105	0.875	1.75 × 10–7	0.033
Immunosensor: CACO-2 lysate	3.52 × 105	0.947	1.84 × 10–7	0.029

Selectivity Test of the Immunosensor Using Cell Extracts

To evaluate the selectivity of the immunosensor, we analyzed cell extracts from colon cancer cells (CACO-2) and human kidney cells (HEK). CACO-2 cell extracts served as the positive control due to their overexpression of the CCSP-2 biomarker, which is associated with colon cancer. HEK cell extracts were used as the negative control, as they are not expected to overexpress CCSP-2, allowing us to confirm that CCSP-2 overexpression is specific to colon cancer cells. PBS buffer was used as the blank sample, and CCSP-2 (100 ng/μL) was used as the standard for comparison.

Figure shows the immunosensor signal (ΔR _ct/R _ct(BSA)) for 100 ng/μL of CCSP-2, CACO-2 cell extracts, HEK cell extracts, and PBS. The highest signal was observed for the sample CCSP-2 (100 ng/μL), as it contains only the target biomarker without interference from other cellular components. This confirms the immunosensor’s design for the specific detection of the CCSP-2 biomarker. For cell extract analysis, CACO-2 extracts exhibited a significantly higher signal than HEK extracts, indicating greater binding of CCSP-2 to the Anti-CCSP-2 immobilized on the immunosensor when tested with CACO-2 extracts. This result suggests that CACO-2 cells contain higher levels of CCSP-2 compared to HEK cells, demonstrating the immunosensor’s ability to differentiate between samples based on CCSP-2 expression levels.

9.

Selectivity test of the immunosensor: comparison of (left side) EIS signaling profile and (right side) charge transfer resistance signals (ΔR _ct/R _ct(BSA)) for CCSP-2 (100 ng/μL), CACO-2 cell extracts (positive control), HEK cell extracts (negative control), and PBS buffer (blank).

Cell extracts contain various potential interfering species, such as DNA, ribosomes, and other proteins, that are released during cell lysis. These components can affect the binding interaction of CCSP-2 with the immobilized antibodies on the immunosensor. Despite this, the negative control (HEK cell extracts) exhibited minimal signal compared to the positive control (CACO-2 cell extracts), confirming the overexpression of CCSP-2 in colon cancer cells and the immunosensor’s selectivity. As expected, the PBS buffer showed a negligible signal, consistent with its role as a blank sample. These results highlight the immunosensor’s high selectivity for CCSP-2 detection, even in the presence of complex interfering species from cell extracts, making it a reliable tool for distinguishing between colorectal cancer and other cancerous and noncancerous cell lines based on CCSP-2 expression.

Selectivity was assessed at a concentration of 100 ng/μL to ensure a clear electrochemical response in the presence of potential interfering species. While this concentration lies near the upper end of the sensor’s dynamic range, it allows for distinct signal differentiation under complex matrix conditions. Analytical parameters such as selectivity, reproducibility, and recovery are typically evaluated at intermediate concentrations to provide a more detailed understanding of performance across the linear range. In this context, this work is positioned as a preliminary study focused on sensor design and initial electrochemical characterization. Future research will incorporate comprehensive analytical validation, including assessments of repeatability, reproducibility, and recovery across the concentration range.

Conclusion

In this study, a graphene-gold-based electrochemical immunosensor has been successfully developed for the detection of CCSP-2. The immunosensor leverages GO covalently attached to the Au surface and is used as a transducer, enabling efficient detection of biochemical interactions between Anti-CCSP-2 immobilized on the Au/4-ATP/GO surface and CCSP-2. The integration of GO on the Au surface enhanced surface functionality and conductivity, significantly improving the overall performance of the sensor. Moreover, the incorporation of GO enhanced electron transfer at the electrode surface, likely due to the partial reduction of the GO during synthesis.

The immunosensor exhibited high sensitivity for CCSP-2 detection, achieving a sensitivity of 0.031 ng/μL^–1 and a limit of detection (LOD) of 0.17 ng/μL. A good correlation (R ² = 0.979) between the relative charge transfer resistance (ΔR _ct/R _ct(BSA)) and CCSP-2 concentrations was observed, confirming the sensor’s reproducibility. Furthermore, the immunosensor demonstrated high selectivity, with a larger signal change (ΔR _ct/R _ct(BSA)) for human colorectal cancer cells (CACO-2) compared to human kidney cells (HEK).

This highlights the sensor’s capability to distinguish colorectal cancer cells from other cancerous and noncancerous cells, making it a reliable tool for CCSP-2 detection even in clinical samples. Owing to its high selectivity, sensitivity, portability, and cost-effectiveness, the developed immunosensor presents a promising platform for on-site diagnosis of colorectal cancer. Its performance underscores its potential as a practical and accessible tool for early colorectal cancer detection, addressing a critical need in clinical diagnostics.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c01338.

• Detailed characterization and validation data for the immunosensor, electrochemical cleaning of the gold electrode is confirmed via cyclic voltammetry, showing distinct oxidation and reduction peaks characteristic of clean Au surfaces (Figure S1); GO synthesis is characterized by Raman and FTIR spectroscopy, confirming structural and functional group modifications from graphite (Figure S2); the FTIR spectrum of pure 4-ATP is also included (Figure S3); electrochemical impedance spectroscopy (EIS) data are analyzed using a Randles equivalent circuit, with fitting parameters such as R _ct, CPE, and chi-squared values reported (Figure S4); cyclic voltammetry studies assessing various 4-ATP concentrations are provided to support the selection of 100 mM for optimal SAM formation (Figure S5); scanning electron microscopy (SEM) images illustrate surface modifications at each functionalization step (Figure S6); antibody immobilization optimization is demonstrated using EIS with increasing concentrations of Anti-CCSP-2 (Figure S7); calibration curve data for CCSP-2 detection using ΔR _ct/R _ct(BSA) are tabulated, along with statistical analysis of reproducibility and surface resistance (Tables S1 and S2); all measurements are supported with relevant literature references (PDF)

B.J.S.C.: conceptualization, formal analysis, investigation, methodology; R.P. review and editing; Y.M.-L.: conceptualization, cell culture, review, and editing; C.R.C. and R.D.-A.: conceptualization, review, and editing.

The authors declare no competing financial interest.