Introduction

Breast cancer (BC) has always been one of the most significant death causes of women, and it becomes the first of cancer patient in the world according to the latest statistics from the World Health Organization [1]. As is well known, early diagnosis and treatment in time would greatly improve the survival rate of patients. Fortunately, detecting cancer biomarkers provides a new future for early diagnosis of BC. Currently, carbohydrate antigen 15–3 (CA15-3) is recognized as the most significant breast tumor markers [2], which was used to evaluate the human normal and diseased states. Usually, the serum CA15-3 levels below 30 U/mL are regarded as healthy states, and above 30 U/mL hints a high risk of BC [3]. Therefore, developing a rapid and sensitive detection technique of CA15-3 is very crucial for early diagnosis and later treatment.

Currently, the detection methods of CA15-3 biomarker have photoelectrochemical (PEC) [4], chemiluminescence [5], fluorescence [6], and enzyme-linked immunosorbent assay (ELISA) [7]. Nevertheless, the above methods suffer from complex operations, expensive equipment, and trained personnel. ELISA needs enzyme to partake, and the activity of enzyme is affected from pH and temperature, resulting in low repeatability. Hence, it is still challenging to develop a method for detecting CA15-3 with simple, sensitive, and high repeatability. Electrochemical immunosensors have been widely applied to cancer biomarkers detection owing to these virtues of sensitivity, high selectivity, quick readout, and easy to miniaturize [8,9,10]. For example, Feng’s group [9] reported a sandwich-type electrochemical immunosensing of CYFRA 21–1 cancer marker.

As is well known, cancer biomarkers in clinical samples are low abundance, in order to detect lower level of cancer biomarkers; signal amplification strategies are often introduced to augment the sensitivity of immunosensors. From the literature, nanomaterials, such as graphene and Au NPs are often employed to enhance sensing sensitivity [11]. However, the sensitivity of immunosensors reported is not high enough and lacks of real sample application. So, it is also challenging to develop higher sensitive electrochemical immunosensors to detect low abundance of serum cancer biomarkers. Recently, graphene aerogel (GA), featuring a 3D hollow network structure and rich porosity, large specific surface area, and ultra-low density [12], has been causing a great deal of attention. The GA not only inherits the good physical and chemical peculiarities of graphene [13], but also stability, ion diffusion, and electron transport rate have been improved clearly [14]. As a result, GA shows good application ability in energy storage and sensing fields [15,16,17].

Metal–organic frameworks (MOFs), a new type of material, have attracted extensive interest owing to their virtues of large specific surface area, adjustable porosity, abundant functional groups, and unsaturated metal sites [18, 19]. Furthermore, different functional materials or signal molecules, catalysts, and signal probes [20] can be loaded on MOFs to improve the sensitivity of sensor. For example, Guan and his coworkers [21] employed Co-MOFs as signal probes and toluidine blue (TB) as reference probes to fabricate a dual-signal ratiometric aptasensor for ochratoxin A detection. Liu and his coworkers [22] employed Ag-MOFs as signal probes to detect cancer biomarker CEA. Inspired by the advantages of MOFs and literature previously, in this work, we developed a sensitive electrochemical immunosensor for detecting the cancer biomarker CA15-3 using CTAB-Co-MOFs materials as signal probes. The designed immunosensor possesses the following advantages. Firstly, GA@Au NPs, 3D nano-network structure, high conductivity, and large specific surface area are used as a sensing matrix to amplify electrochemical signal. Secondly, CTAB-Co-MOFs not only serve as carriers for AuPt NPs and Ab2 immobilization, but also act as signal probes. AuPt NPs have catalysis ability to the system of H2O2-Co2+, resulting in sensing signal amplification clearly. As a result, the immunosensor demonstrated prominent performance, including good sensitivity and selectivity.

Experimental section

Chemicals, reagents, and instruments

Related chemicals, reagents, and instruments are indicated in the Supplementary Materials (Section S1 and Section S2).

Synthesis of materials

Au NPs [23], Pt NPs [24], CTAB-Co-MOFs [25], GA@Au NPs, and CTAB-Co-MOFs@AuPt NPs materials were synthesized with slight modification according to the reported literature. The related specific synthesis steps can be sought out in the Supplementary Materials (Section S3).

Preparation of labelled-Ab2 bioconjugates

The preparation of labelled-Ab2 bioconjugates is shown in the Supplementary Materials (Section S4).

Fabrication of the electrochemical immunosensor

The building process of the immunosensor is shown in Scheme 1. Simple, the glassy carbon electrode (GCE) was first handled by polishing, washing ultrasonically, and drying in a flow of nitrogen [26]. The following materials were sequentially assembled on the surface of GCE. Firstly, 10 µL of GA@Au NPs (1 mg/mL) suspension solution was dropped on GCE and dried in a 37 °C vacuum oven. Secondly, 6 µL of CA15-3 antibody (Ab1, 20 µg/mL) was coated and incubated at 4 °C overnight to anchor the antibody. At each subsequent step, the modified electrode was washed with PBS (0.01 M, pH = 7.4) solution to avoid analyte adsorption on the electrode surface. Subsequently, the electrode was blocked with 6 µL bovine serum albumin solution (BSA, 1 wt%) and incubated at room temperature for 30 min. After that, 6 µL CA15-3 solution with a series of concentrations was added and incubated at 37 °C for 1 h. Finally, 6 µL of CTAB-Co-MOFs@AuPt NPs-Ab2 solution was added dropwise on the electrode to specifically bind by antibody-antigen interactions to construct a sandwich-type immunosensor. The immunosensor was incubated at 37 °C for 1 h and stored at 4 °C when not in use.

Scheme 1
scheme 1

Schematic diagram of the fabrication process of the immunosensor (A Synthesis of CTAB-Co-MOFs and labelled-Ab2. B The assembling process of immunosensor)

Full size image

Electrochemical measurements

Electrochemical measurements are revealed in the Supplementary Materials (Section S5).

Results and discussion

Design principle of immunosensor

GA@Au NPs were selected for sensing matrix to connect with Ab1 vs Au-NH2 bond. CTAB-Co-MOFs, larger specific surface area, were employed as a carrier for AuPt NPs and Ab2 immobilization. In the presence of CA 15–3, a sandwich-type immune complex was formed on the electrode surface by use of the specific interaction of Ab and Ag. H2O2 can catalyze the Co2+/Co3+ reaction process, meanwhile, AuPt NPs on the surface of CTAB-Co-MOFs can catalyze the reaction of H2O2 oxidizing Co2+ to Co3+. The electrochemical reduction response of Co3+/Co2+ in CTAB-Co-MOFs self was used for the sensing signal. The signal intensity was related to CA15-3 level, so CA15-3 can be quantified via measuring the signal intensity change. The multiple amplification improves the sensitivity of immunosensor. The mechanism is shown as follows:

Characterization of materials

Characterization of GA@AuNPs nanocomposites

Figure 1A and Fig. S1 show photographs and SEM images of GA at different magnifications. It can be seen that GA consists of randomly arranged and fluffy wrinkled graphene sheets and interconnected with each other to present a porous 3D network structure, which improves the stability of the material and enhances the electrical conductivity [14]. Figure 1B shows the SEM of GA @Au NPs; the Au NPs with a size of 16.72 ± 0.91 nm are consistently spread on the surface of GA, indicating that GA@Au NPs nanocomposites were successfully prepared.

Fig. 1
figure 1

A SEM image of GA. B SEM images of GA@Au NPs at different magnifications (inset size distribution). C SEM image of CTAB-Co-MOFs (inset size distribution), D TEM image of CTAB-Co-MOFs, E XRD, F XPS spectra of CTAB-Co-MOFs. G High-resolution XPS fitting spectra of Co 2p. H FT-IR patterns of CTAB-Co-MOFs and Co-MOFs

Full size image

Characterization of CTAB-Co-MOFs

SEM and TEM were applied to characterize the morphology of the CTAB-Co-MOFs (Fig. 1C, D). The CTAB-Co-MOFs exhibit a regular polyhedral structure and the average size is about 142.68 ± 12.76 nm. Figure 1E shows the XRD diffraction peak of CTAB-Co-MOFs, the distinctly sharp and intense characteristic peaks at 5° to 35° indicate high crystallinity of CTAB-Co-MOFs [25]. XPS was employed to investigate the surface composition of it. It is clear that the product contains Co, C, N, and O elements (Fig. 1F), which is in accordance with the results of EDS (Fig. S2). Furthermore, the high-resolution Co 2p XPS spectra show that peaks of Co 2p3/2 and Co 2p1/2 appear at 781.18 and 796.68 eV [27], and satellite peaks emerge at 786.58 ~ 789.48 eV and 801.48 ~ 803.98 eV (Fig. 1G). The FT-IR technique was used to verify relevant functional groups in MOFs (Fig. 1H), the stretching vibration characteristic peaks of Co – N at 424.1 cm−1, and C – N and C = N at 1142.3 cm−1 and 1579.1 cm−1 are observed in CTAB-Co-MOFs, respectively [27]. All investigations confirmed the successful synthesis of the CTAB-Co-MOFs.

In addition, the difference between CTAB-Co-MOFs and Co-MOFs was investigated, using SEM (Fig. S4A-B), TEM (Fig. S4C-D), FT-IR (Fig. 1H) and XRD (Fig. S5). No significant difference was observed from these figures above, indicating that CTAB could not change the morphology, functional groups, and crystal structure. What is more, CTAB-Co-MOFs have better stability than Co-MOFs (Fig. S6). This suitable explanation is that the hydrophobic cetyl chain of CTAB molecules improve the water stability of CTAB-Co-MOFs [28].

Characterization of CTAB-Co-MOFs@AuPt NPs nanocomposites

TEM was used to investigate Au NPs and Pt NPs obtained (Fig. 2A, B), and it was found that Au NPs and Pt NPs were uniformly spherical with an average size of 17.5 ± 0.78 nm and 3.90 ± 0.47 nm, respectively. From the SEM and TEM images (Fig. 2C, D), it can be seen that the Au NPs and Pt NPs appeared as small clusters on the surface of the CTAB-Co-MOFs. The CTAB-Co-MOFs@AuPt NPs also exhibit the polyhedral structure similar to CTAB-Co-MOFs. Furthermore, the EDS result confirms the existence of C, N, O, Co, Au, and Pt elements (Fig. 2E). The XRD pattern displays that the CTAB-Co-MOFs@AuPt NPs composites still own an excellent crystalline state (Fig. 2F), and the AuPt NPs exhibit four characteristic peaks from 35 to 80°. Notably, the peak intensity at 38.1° is much higher than that of other planes, which manifests that Pt atoms exist in the Au matrix [29]. These results above confirm the successful preparation of CTAB-Co-MOFs@AuPt NPs.

Fig. 2
figure 2

TEM image with size distribution of A Au NPs and B Pt NPs. C SEM image, D TEM images and E EDS spectra of CTAB-Co-MOFs@AuPt NPs. F XRD patterns of CTAB-Co-MOFs@AuPt NPs and CTAB-Co-MOFs

Full size image

Construction investigation of the immunosensor

Figure 3A demonstrates current density in 1.0 mM [Fe(CN)6]3−/4− solution at the different stepwise process of this immunosensor. Because of the larger specific surface area and bonzer conductivity of GA@Au NPs, the GA@Au NPs/GCE (curve b) exhibited a higher current density in contrast to bare GCE (curve a). As Ab1, BSA, CA15-3 antigen, and labelled-Ab2 were orderly assembled on electrode surface, the current densities were reduced continuously (curves c–f) due to their nonconductivity, indicating the immunosensor was successfully prepared.

Fig. 3
figure 3

A Current density and B EIS curves recorded at different modified electrodes: (a) bare GCE, (b) GA@Au NPs/GCE, (c) Ab1/GA@Au NPs/GCE, (d) BSA/Ab1/GA@Au NPs/GCE, (e) CA15-3/BSA/Ab1/GA@Au NPs/GCE, (f) labelled-Ab2/CA15-3/BSA/Ab1/GA@Au NPs/GCE

Full size image

Figure 3B shows the EIS curves of 5.0 mM [Fe(CN)6]3−/4− at different step of immunosensor preparation. The GA@Au NPs/GCE (curve b) presented a smaller semicircle in contrast to the bare GCE (curve a). After immobilization with Ab1, BSA, and CA15-3 antigen, respectively, the semicircle diameter was increased sequentially (curve b–f). When labelled-Ab2 was captured on the electrode surface, the diameter of the semicircle was further aggrandized due to antigen–antibody complex formation. The EIS results obtained were consistent with that of CV, revealing the successful manufacture of the immunosensor.

Investigation of the signal amplification and feasibility of CA15-3 analysis

In order to verify signal amplification from AuPt NPs and H2O2, DPV response at the same concentration of CA15-3 (100 µU/mL) was recorded for this evaluation at different conditions (seen Fig. 4A). In the absence of the labelled Ab2 (curve a), no any signal response was observed and in the presence of CTAB-Co-MOFs@Au NPs/BSA-Ab2, a clearly DPV signal was obtained (curve b) due to a sandwich structure formation. If the support electrolyte containing H2O2, the increasing DPV signal was obtained clearly (curve c), when AuPt NPs existed (curve d), the DPV signal was further increased, indicating that AuPt NPs had good catalytic ability to the system of H2O2—Co2+.

Fig. 4
figure 4

A DPV responses of immunosensors recorded at different signal probes. (a) No labelled Ab2 probe; (b) CTAB-Co-MOFs@Au NPs-Ab2 probe; (c) CTAB-Co-MOFs@Au NPs-Ab2 probe in the presence of H2O2; (d) CTAB-Co-MOFs@AuPt NPs-Ab2 probe in the presence of H2O2. Condition: The concentration of CA15-3 is 100 µU/mL, incubation time of 60 min. B DPV response of the immunosensor recorded in the absence of CA15-3 antigen (curve a) and in the presence of CA15-3 antigen (curve b). Condition: The concentration of CA15-3 is 10 mU/mL, incubation time of 60 min

Full size image

The feasibility of CA15-3 analysis was investigated, and the results are presented in Fig. 4B. In the absence of CA15-3 (curve a), no obvious response signal was received; however, a larger response signal was obtained when existed CA15-3 (curve b). The signal intensity was related to the concentration of CA15-3. Thus, the concentration of CA15-3 can be quantified based on the peak current intensity changes, indicating the feasibility of immunosensor to detect the serum CA15-3.

Optimization of experimental condition

Herein, experimental conditions (e.g., pH, incubation temperature and time, and the concentration of Ab1) were optimized to improve the analytical performance of the immunosensor. The results obtained are shown in the Supplementary Materials (Section S8). Briefly speaking, the optimized results are listed as follows: pH 7.4; incubation temperature, 37 °C; incubation time, 60 min; the concentration of Ab1, 20 µg/mL; and the concentration of H2O2, 2.5 mM, and the amount of GA@Au NPs, 10 µL.

Analytical performance

Under the optimal conditions, the designed immunosensor was employed to detect different concentrations of CA15-3 and recorded their DPV response. As shown in Fig. 5A, the currents are increased gradually at the potential about − 0.85 V (vs SCE) with the increase of CA15-3 concentration, and a good working cure was obtained in the range of 10 µU/mL to 250 U/mL (Fig. 5B) with the limit of detection (LOD) was 1.1 µU/mL (at the 3σ/S, σ is the standard deviation of the blank and S is the slope of the analytical curve). The linear regression equation was y = 4.4500 + 3.9063 lg x (y is the peak current value (µA), x is the concentration of CA15-3 (µU/mL), R2 = 0.999). Compared with other immunosensors reported previously (shown in Table 1), the immunosensor shows a wider linear range and lower LOD. The reasons of good performance can be explained as follows: (1) GA@Au NPs, good conductivity and larger surface area, were employed for sensing matrix to enhance electron transfer rate. (2) CTAB-Co-MOFs, larger specific surface area, were employed as a carrier for improving the amount of AuPt NPs and Ab2. (3) AuPt NPs catalyze the reaction of H2O2 oxidizing Co2+ to Co3+, resulting in the significant amplification of the sensing signal.

Fig. 5
figure 5

A DPV response recorded at various CA15-3 concentrations. B The working curves of peak current (I) vs the logarithmic of CA15-3 concentration. Conditions (U/mL): C1, 0; C2, 0.00001; C3, 0.0001; C4, 0.001; C5, 0.01; C6, 0.1; C7, 1; C8, 10; C9, 100; C10, 250. The datum dot is average value of three times

Full size image
Table 1 Comparison of CA15-3 detection with different analytical techniques
Full size table

Selectivity, reproducibility, and stability

To assess the selectivity of the immunosensor, some interferences (e.g., CA19-9, IgG, PSA, HGB, HSA, Mb, and BSA) were selected for this investigation. Specifically, the immunosensor was employed to detect 0.01 U/mL CA15-3 antigen in the presence of single interfering substance and recorded DPV response corresponding to different conditions. As shown in Fig. 6A, no apparent current change was observed, hinting at the good selectivity of immunosensor.

Fig. 6
figure 6

A Specificity of the immunosensor (condition: 0.01 U/mL CA15-3; interference substances: 1 U/mL CA19-9, 10 µg/mL IgG, 10 µg/mL PSA, 100 µg/mL HGB, 100 µg/mL HSA, 100 µg/mL Mb and 100 µg/mL BSA; n = 3). B Reproducibility of the immunosensor (0.1 U/mL CA15-3; n = 3); C Stability of the immunosensor (100 µU/mL CA15-3; n = 3)

Full size image

Reproducibility was evaluated using six immunosensors to detect CA15-3 (0.1 U/mL) under the same conditions and record their DPV response (shown in Fig. 6B). The relative standard deviation (RSD) is 0.35%. Additionally, the stability of the immunosensor was assessed by depositing it in the refrigerator (4 °C) for 7 days and testing it once a day. The results are presented in Fig. 6C, and it can be seen that the current retained 94.7% of its initial value on the 7th day, indicating the good stability of the immunosensor.

Serum sample analysis

To verify the feasibility of the proposed immunosensor, human serum sample was employed and the recovery test was adopted. According to the method described in this experiment, the sample was parallelly detected five times. The results are shown in Table 2 according to the equation (\(\overline{x}\pm t\frac{s}{\sqrt{n} }\), s is the standard deviation). The recoveries obtained are within 98.33 to 102.08% with RSD values within 2.49 to 2.67%, indicating that the proposed immunosensor has good feasibility in clinical analysis.

Table 2 Analytical results in the human serum sample (P = 0.95; n = 5)
Full size table

Conclusion

In this work, a sensitive sandwich electrochemical CA15-3 immunosensor was developed using GA@Au NPs and CTAB-Co-MOFs@AuPt NPs. The proposed immunosensor shows excellent analytical performance, including wider linear range (10 µU/mL to 250 U/mL) and lower LOD of 1.1 µU/mL. Satisfactory, the immunosensor has potential application in the early clinical diagnosis of BC via clinical serum analysis. Furthermore, the immunosensor also provides a feasible strategy for detecting other cancer biomarkers only adjust corresponding to antibody.