Introduction

Vascular endothelial growth factor (VEGF) is one of the most important cytokines that plays an essential role in vascular development and permeability [1]. VEGF and its receptors have been considered fundamental regulators for angiogenesis and vasculogenesis, which describe the processes of new capillary blood vessel formation [2]. VEGF triggers new blood vessel formation through the induction of vascular endothelial cells that can proliferate, migrate, and differentiate in healthy tissues [3, 4].

VEGF165 is the most dominant isoform of the human VEGF (also called VEGF-A) family and has the highest biological activity during angiogenesis [5]. Particularly, it participates in a variety of physiological angiogenic processes, including wound healing [6], female reproduction cycle [7], embryogenesis [8], neural development [9], bone formation [10], and hematopoiesis [11] in healthy adults. However, under certain circumstances, the overexpression of VEGF is well correlated with rheumatoid arthritis [12], Parkinson’s disease [13], psoriasis [14], and several ocular diseases [15, 16], while VEGF activity continues physiologically normal during vascular development. Besides, VEGF has also been widely described as a major tumor angiogenic factor in many tumor types [17, 18]. Particularly, VEGF participates in different stages of tumor, including development, progression, and metastasis, which can also orchestrate tumorigenic angiogenesis in ovarian [19], pancreatic cancer [20], lung (e.g., non-small cell lung cancer) [21], breast [22], oral [23], colorectal cancer [24], and liver cancers [25]. In this regard, elevated VEGF levels have been proposed as an essential biomarker in many types of cancers. In addition to cancer, such diseases as age-related macular degeneration in the eyes result in increased VEGF concentration in patients’ eyes. Therefore, the detection of the VEGF concentration in the patient’s retina is critical for the determination of the amount of anti-VEGF drug for intravitreal injections [26, 27].

Due to the critical importance of VEGF in biomedical research, early diagnosis, and medical applications, the detection and quantification of VEGF is strongly required to trace its levels in biological samples. Until now, a variety of approaches for the detection of VEGF have been reported, including enzyme-linked immunosorbent assay (ELISA) [28] (the standard assay format), fluorescence in situ hybridization (FISH) [29], immunohistochemistry (IHC) [30], and fluorescent spectrometry [31,32,33]. Even though each of these techniques has such superiorities, there is no available technique to overcome such drawbacks, including excessive time requirement, expensiveness, and requirements of complicated instrumentation. Among the analytical approaches in biosensor development, colorimetric techniques can overcome these shortcomings, which are commonly used for field analysis and point-of-care diagnosis for many applications. In the case of antibody-antigen-based reactions, antigen–antibody interaction causes alteration in the color intensity with favorable properties that can fulfill the desired requirements expected from a biosensor [32,33,34]. In recent years, colorimetry based on metal nanoparticles has garnered significant attention and has injected fresh vigor into traditional colorimetric methods, enabling precise and accurate analyzes [35]. Among the metal-based nanoparticles, iron oxide nanoparticles (FeNPs) provide colorimetric changes with many advantages as accuracy, simplicity, broad applicability, and high selectivity, which make them appealing for straightforward colorimetric detection. In addition, FeNPs are cost-effective compared to other metals and do not cause unstable and spontaneous aggregation when they are used as probes in detections [36, 37].

In this study, the synthesized FeNPs were conjugated with VEGF165 antibody for the colorimetric detection of VEGF165. The sensor gave a linear range between 0.5 and 100 ng/mL with 0.29 ng/mL as the limit of detection (LOD), which indicates that the developed immunoassay platform is highly promising for clinical applications. Furthermore, the sensing system gave excellent specificity for the potential interferents such as glucose, urea, insulin, C-reactive protein, and serum amyloid A in human serum.

Materials and methods

Reagents and strains

Recombinant Kluyveromyces lactis GG799 (New England Biolabs, Massachusetts, USA) was prepared by cloning the human VEGF165 cDNA (GenBank accession no. AF486837.1) in XhoI and EcoRI sites of the pKLAC2 vector under the control of the LAC4 promoter of the reported study [38]. The recombinant strain was cultured in YCB plates (3% 1 M Tris–HCl pH 7.4, 1.17% YCB, and 2% agar) containing 5 mM acetamide and maintained at – 80 °C as glycerol stock. In addition, the expression was performed using an unoptimized YPGal medium (1% yeast extract, 2% peptone, and 4% galactose). Ni–NTA agarose resin, 1000 kDa MWCO cellulose acetate dialysis membrane, and Pierce™ BCA Protein Assay Kit were purchased from Qiagen (Hilden, Germany), Spectrum Laboratories (California, USA), and Thermo Fisher Scientific (Waltham, MA, USA), respectively. N′-Ethylcarbodiimide hydrochloride (EDC) and N-hydroxysuccinimide (NHS) were purchased from Sigma-Aldrich (St. Louis, MO, USA) while C-reactive protein (CRP) (1 mg/mL) was obtained from Prospect. Serum amyloid A (SAA) (5.6 mg/mL) was purchased from Merck (Merck Millipore, Darmstadt, Germany). The washing buffer for ELISA was 0.1 M PBS (pH 7.2) containing 0.05% Tween 20 and 0.15 M NaCl. The coating buffer for ELISA was a 50 mM carbonate buffer, pH 9.6. ELISA maxisorp immunoplate (96 flat-bottom wells) was purchased from Nunc (Roskilde, Denmark). Artificial serum sample was prepared using KCl 0.335 g/L, MgCl2 0.152 g/L, CaCl2 0.505 g/L, NaCl 8.470 g/L, NH2CONH2 0.150 g/L, D-glucose 0.450 g/L, and BSA 1.000 g/L as depicted in the literature [39].

Expression and purification of recombinant human VEGF 165

The inoculum was prepared by inoculating a single colony of recombinant K. lactis GG799 strain in 25 mL YPGal medium and incubated at 30 °C at 250 rpm for 72 h. The culture was inoculated with a ratio of 1 to 100 in a 2 L Erlenmeyer shake flask including 300 mL YPGal, which was then followed by, and then, the expression of VEGF165 was performed. After 48 h incubation, the culture suspension was centrifuged at 5000 rpm for 5 min at room temperature. The supernatant containing the secreted recombinant VEGF165 was utilized to purify the protein to use a standard protein for ELISA studies. During the purification process, the supernatant was precipitated with 60% ammonium sulfate solution in an ice bath for 60 min and then centrifuged at 12,000 rpm for 20 min. One hundred millimolars of phosphate buffer (pH 7.4) was used to dissolve the obtained protein pellet. After adding Ni–NTA agarose resin to the polycarbonate column to perform affinity-based purification chromatography, the column was washed with a 100 mM phosphate buffer (pH 7.4). The protein solution was applied to the column, and the target protein bound to the column was obtained through fractionations using the same phosphate buffer containing 300 mM imidazole. The recombinant VEGF165 elution solution was dialyzed using 1000 kDa MWCO cellulose acetate dialysis membrane against HEPES buffer (20 mM pH 7.4) overnight + 4 °C on a magnetic stirrer. The concentration of VEGF165 was calculated using the “Pierce™ BCA Protein Assay Kit.”

Synthesis and characterization of LpAB-FeNP

Synthesis of iron oxide nanoparticles was performed using a modified version of the reported studies [40, 41]. Briefly, 1 M citric acid was mixed with 100 mM of FeCl3, 50 mM of FeSO4, and 10 mM lactose p-aminobenzoic acid (the sugar ligand was provided by Dr. İdris Yazgan of Kastamonu University, Turkey). The mixture was then vortexed at 1200 rpm, and 1 M of NaOH was added to the mixture while vortexing at 600 rpm. The addition of NaOH resulted in the formation of heat, which played a vital role during the iron oxide nanoparticle formation. The colloidal nanoparticles were then transferred to a new sterile 15-mL centrifuge tube. The goal of the addition of the sugar ligand (the synthesis approach can be found elsewhere [42]) was to enhance the stability of the synthesized iron oxide nanoparticles [43] and introduce an amine group. UV–vis spectroscopy, high-resolution transmission electron microscopy-selected area diffraction (HRTEM-SAED), scanning electron microscopy (SEM), and X-ray photoelectron spectroscopy (XPS) were used to characterize LpAB-FeNPs.

Conjugation of anti-VEGF with LpAB-FeNP

Anti-VEGF conjugation to LpAB-FeNP was performed using the EDC/NHS chemistry method [44]. Anti-VEGF was activated using 50.0 mM EDC and 12.5 mM NHS and then introduced to LpAB-FeNP solution. Anti-VEGF:EDC:NHS (1:2:2; v:v:v) was mixed in PBS and incubated at 250 rpm in an orbital shaker for 15 min. After that, the LpAB-FeNP/anti-VEGF conjugate was washed three times with PBS buffer using 50-kDa centrifugal filters at 4000 rpm. Dynamic light scattering technique (DLS, Malvern Zetasizer Nano ZS model) was used to measure the size distributions and zeta potential of LpAB-FeNP and LpAB-FeNP/anti-VEGF.

Construction of the immunoassay platform

First, the purified recombinant VEGF165 (1 mg/mL) was diluted to 10 different concentrations (0.5, 1.0, 5.0, 10, 25, 50,100, and 250 ng/mL) with coating buffer and added onto the ELISA plate. This plate was then incubated overnight at 4 °C. After that, unbound VEGF165 was removed from the plate by cleaning it with a washing buffer. Next, 2 mg/mL BSA was added to each well and incubated for 1 h at room temperature as a blocking step. The conjugate of LpAB-FeNP/anti-VEGF was introduced to the plate and incubated for 1 h. After the incubation, a washing buffer was used to remove the unreacted components. The UV–Vis absorbance of this system was measured, and the absorbance varied according to the VEGF165 concentration. The proposed immunoassay platform was similar to a previous study [45].

Results and discussion

Characterization of LpAB-FeNP and LpAB-FeNP/anti-VEGF

UV–vis spectroscopy played a significant role in the synthesizing of FeNP, LpAB-FeNP, and its conjugation with anti-VEGF. Fig. S1 shows absorbance at 370 nm due to the metallic iron oxide core. Fig. S2A shows the absorbance at 370 nm due to a metallic iron oxide core that confirms the presence of LpAB-FeNP [46]. The decrease in the absorbance intensity after the conjugation process indicates the successful attachment of anti-VEGF to LpAB-FeNP [47, 48]. In this study, the morphology of LpAB-FeNP and LpAB-FeNP/anti-VEGF was observed by scanning electron microscopy (SEM), as depicted in Fig. S2B and Fig. S2C. As shown in Fig. 1, the LpAB-FeNP nanoparticles exhibited a globular shape. However, after the conjugation of anti-VEGF, the morphology tended to form agglomerations and appeared as cloudy clusters. This change in morphology provides evidence of the attachment of anti-VEGF to LpAB-FeNP.

Fig. 1
figure 1

A HRTEM and B SAED images of LpAB-FeNP

Full size image

The HRTEM characterization reveals that the core size of FeNPs is in 5.5 ± 1.5 nm, where ~ 5 nm spherical ones were in the most abundant form. Selected area diffraction (SAED) patterns reveal that the FeNPs had (012), (222), (200), (024), and (511) hkl indices. The hkl (012) and (024) belong to α-Fe2O3, the hkl (200) belongs to FeO (wustite), and (222) and (511) hkl indices belong to Fe3O4 (magnetite) iron oxide crystal structures [49]. Besides, the d-spacing measured on an individual FeNP (the inset in Fig. 1A) shows that the FeNP had α-Fe2O3 form.

Zeta potential and size distribution of LpAB-FeNP and LpAB-FeNP/anti-VEGF were characterized by DLS. The average size of the LpAB-FeNP was measured to be approximately 33.9 nm, with negative zeta potential − 10.1 ± 2.11. DLS recognizes hydrodynamic size while TEM only recognizes size of the metallic core, so the size difference between the two techniques is expected. After the conjugation, the size of LpAB-FeNP/anti-VEGF increased to 74.4 nm, with a zeta potential of − 13.2 ± 2.01. The increase in hydrodynamic particle size and the difference in zeta potential confirmed the successful covalent modification.

X-ray photoelectron spectroscopy (XPS) measurements are highly accurate and provide precise details about the chemical composition of the target species. In this study, XPS was used to thoroughly examine the binding between the conjugated molecules. Therefore, LpAB-FeNP samples and LpAB-FeNP/anti-VEGF conjugates were studied using XPS (Fig. 2). Figure 2A and Fig. 2D show Fig. 2C 1 s photoelectron spectra of LpAB-FeNP and LpAB-FeNP/anti-VEGF, respectively. Figure 2A shows three main peaks of C 1 s in LpAB-FeNP spectrum. The components centered at 285.18 eV for C, 286.68 eV for C-N/C-O or C = O, and 288.38 eV for O-C = O resulted from the presence of carboxylic acid of the sugar ligand. The C 1 s spectrum for the LpAB-FeNP/anti-VEGF surface was deconvoluted into four peaks (Fig. 2D). The peaks were observed at 286.18, 286.98, 288.88, and 294.78 eV belong to C–C bond [50], C-N/C-O or C = O bonds [45], amide groups (N–C = O) [51], and π-π* interaction [52].

Fig. 2
figure 2

XPS spectrum of LpAB-FeNP. A C 1 s LpAB-FeNP. B N 1 s LpAB-FeNP. C Fe LpAB-FeNP. D C 1 s LpAB-FeNP/anti-VEGF. E N 1 s LpAB-FeNP/anti-VEGF. F Fe LpAB-FeNP/anti-VEGF

Full size image

Figure 2B shows the N 1 s spectrum, which exhibited three peaks corresponding to N–H/N–H + [53], Fe2+, and Fe3+ interacting N–H/N–H + at 399.98, 403.98, and 406.88 eV, respectively [54]. The N 1 s deconvolution of LpAB-FeNP/anti-VEGF (Fig. 2E) gave two distinct groups of peaks. The peaks at 400.68 and 403.78 eV correspond to N–H, N–H + , N–O [55], and amide groups (N–C = O) [56], respectively, which support the covalent linkage between LpAB-FeNP and anti-VEGF. In Fig. 2C and Fig. 2E, the characteristic peak of Fe 2p3/2 is expected at 710.2 eV, which is indicative of the core level spectra of Fe3O4 nanoparticles [57]. As illustrated, the peaks in the Fe 2p spectrum are situated at 726.68 eV for LpAB-FeNP and 710.88 eV for bivalent iron (Fe2+) in LpAB-FeNP/anti-VEGF. Furthermore, the peaks at 712.48, 714.08, 718.38, and 726.98 eV can be attributed to trivalent iron (Fe3+) [58]. The component at 724.08 eV corresponds to Fe 2p1/2. According to the literature, the characteristic peak of zerovalent Fe (Fe0) is expected at 719.9 eV [56]. As shown in Fig. 2F, a characteristic peak at 719.98 eV for Fe0 is observed. Combining all the XPS results, it can be concluded that the anti-VEGF has been successfully conjugated onto the surface of LpAB-FeNP.

Analytical characterization of VEGF LpAB-FeNP/anti-VEGF

Enzyme-mimicking nanomaterials have emerged as potential substitutes for natural proteins [59]. Similarly, an antibody-based immunoassay was developed for C-reactive protein (CRP) [45] where a metal organic framework (MOF) labeled with anti-CRP was used as a probe. In the presence of CRP, the labeled anti-CRP forms a connection via affinity, and a fluorescence signal is generated depending on the target. In this study, a colorimetric analysis of immunoassay targeting different levels of VEGF165 was performed using colorimetric detection in PBS buffer (pH 7.4). Here, the LpAB-FeNP/anti-VEGF conjugate was employed as the detection antibody, which was captured by VEGF165 coated onto a 96-well plate. LpAB-FeNP served as the source of the colorimetric signal. The schematic presentation of the proposed immunoassay is given in Fig. 3.

Fig. 3
figure 3

Schematic demonstration of colorimetric immunoassay preparation platform

Full size image

The investigated assay parameters for the detection were the concentration of BSA, VEGF165 incubation time, and the concentration of anti-VEGF. The BSA concentration and incubation time for VEGF165 were determined based on previous studies [45, 60]. The calibration curve of the proposed VEGF165 determination assay exhibited a wide linear range between 0.5 and 100 ng/mL with the best fit equation of y = 0.174x − 0.322. Absorbance values of VEGF165 Log (pg/mL) at different concentrations are shown in the Supplementary Materials file (Table S1). Additionally, the correlation coefficient (R2) was calculated as 0.996, indicating a reliable relationship, as shown in Fig. 4B. Figure 4D presents an image depicting the potential impact of interferences on the proposed immunoassay platform.

Fig. 4
figure 4

A Optimization of anti-VEGF concentration. B Calibration curve of log VEGF165 (pg/mL) constructed by linear fitting. C The result of the proposed FeNP-based ELISA assay for VEGF165 detection over some potentially interfering substances: VEGF165, 10 ng/mL; glucose, 10 mM; urea, 0.1 mg/mL; insulin, 10 μlU/mL; CRP, 10 ng/mL; SAA, 10 ng/mL. D Representative example of interference results obtained in the ELISA plate. Error bars show SD of at least three measurements

Full size image

The limit of detection (LOD) is a crucial parameter for assessing the analytical performance of a method with a certain level of confidence [61]. In the LpAB-FeNP/anti-VEGF immunoassay, the LOD was determined using 10 measurements at the lowest concentration (0.5 ng/mL) on the calibration curve. LOD was calculated using the formula 3SD/m, where SD represents the standard deviation of 10 measurements performed at the lowest VEGF165 concentration and m represents the slope of the VEGF165 calibration curve. The experimental determination of LOD in this study was 0.29 ng/mL.

Table 1 provides an overview of the comparison of some VEGF sensors described in the literature. Based on signal transducer types, immunoassays can be classified as electrochemical, colorimetric, or optical immunoassays. Among these, colorimetric immunoassays for disease biomarker detection have drawn much interest because of their ease of use and high level of effectiveness [62].

Table 1 Comparison of VEGF determination studies summarized in literature
Full size table

In the current literature, the tests are mostly performed using electrochemical and colorimetric methods. They frequently employ aptamer/antibody combinations, offering high specificity and selectivity. Furthermore, the repeatability, stability, and affordability of the test system can be improved by using antibodies and/or aptamers [75]. In our case, the results show that the prepared LpAB-FeNP/anti-VEGF has many advantages over the others, including a better LOD and a broader linear detection range. The repeatability of the immunoassay played a key role in our VEGF165 measurement platform. We investigated the repeatability through conducting eight measurements in the presence of 10 ng/mL VEGF165 protein. The coefficient of variation (CV) of the platform was calculated as 0.152%.

The potential interference effects of glucose, urea, insulin, CRP, and SAA (common biological components found in blood) on the immunoassay response for detecting VEGF165 were examined. The reference levels for the interferences are 0.35–5.0 mM for glucose [76], 0.10–0.16 mg/mL for urea [77], 2–20 mIU/mL for insulin [78], lower than 0.3 to 1.0 mg/mL in healthy individuals for CRP [79], and 0.02–0.05 mg/mL for SAA under normal conditions [80] according to literature. LpAB-FeNP/anti-VEGF/VEGF165 was prepared, and the effect of each substance in the immunoassay response was evaluated using the colorimetric method in the presence of 10 ng/mL VEGF165 (Fig. 4C, D). The selectivity was calculated to be 96.84% for VEGF165 + glucose, 102.22% for VEGF165 + urea, 99.86% for VEGF165 + insulin, 100.54% for VEGF165 + CRP, and 102.67% VEGF165 + SAA as seen in Fig. 4C. The impact of the potential interferences was found to be lower than 5%, indicating that the immunoassay exhibited excellent selectivity. Therefore, glucose, urea, insulin, CRP, and SAA did not interfere with the LpAB-FeNP/anti-VEGF immunoassay platform.

The proposed immunosensor platform was employed to determine VEGF165 in the artificial blood sample. A known amount of VEGF165 (10 ng/mL) was added to artificial serum and determined the recovery % with LpAB-FeNP/anti-VEGF. The recovery % of VEGF165 detection was calculated as 95.16%, and the relative standard deviation (RSS) was calculated as 5.08. Confirming the system’s validity, the obtained result falls within the acceptable recovery range of 95–105% [81].

Conclusion

In this study, we successfully designed a colorimetric measurement system for the detection of VEGF165. The developed LpAB-FeNP/anti-VEGF immunoassay platform offers several advantages, including short analysis time, high accuracy, high specificity, reduced sample volume, and low cost. Besides, the wide linear range and low detection limit, this developed simple ELISA approach reduces the utilization of extended assay time and consumption of chemicals and antibodies. Therefore, the developed platform can be an alternative method in the detection of VEGF, which can be expanded to the detection of different analytes that are critically important in diagnostics.