The EMT process has a central role in physiological and pathological conditions. During cancer progression, EMT promotes invasion and confers tumor cells with stem cell-like properties27. Recent data from different tumor types, have defined a set of biological networks and pathways modified during EMT.27.x27.De Domenico, S. and Vergara, D. The many-faced program of epithelial-mesenchymal transition: a system biology-based view. Front Oncol. 2017;
Crossref | Scopus (0) | Google ScholarSee all References,28.x28.Guerra, F., Guaragnella, N., Arbini, A.A., Bucci, C., Giannattasio, S., and Moro, L. Mitochondrial dysfunction: a novel potential driver of epithelial-to-mesenchymal transition in Cancer. Front Oncol. 2017;
Crossref | Scopus (0) | Google ScholarSee all References This includes also a long list of biomarkers including proteins, nucleic acids, lipids, that demonstrated a robust clinical validity.3.x3.Vergara, D., Simeone, P., Frank, J., Trerotola, M., Giudetti, A., Capobianco, L. et al. Translating epithelial mesenchymal transition markers into the clinic: novel insights from proteomics. EuPA Open Proteom. 2016;
Crossref | Scopus (2) | Google ScholarSee all References,29.x29.Santamaria, P.G., Moreno-Bueno, G., Portillo, F., and Cano, A. EMT: present and future in clinical oncology. Mol Oncol. 2017;
Crossref | Scopus (12) | Google ScholarSee all References If these biomarkers showed diagnostic or prognostic value when their expression is investigated at the tissue level, it is reasonable to propose these as potential circulating markers if released in blood or other biological fluids.
Based on this assumption, we validated the different expression of selected EMT markers in breast cancer models and investigated their presence in the secretomes of our models. As shown, the classical EMT markers E-cadherin and vimentin well characterized MCF-7 and MDA-231 cells for their epithelial and mesenchymal phenotype, respectively (Fig. 1Fig. 1A). By western blot, we confirmed this difference in the secretome of both cell models (Fig. 1Fig. 1A). Our hypothesis was subsequently verified at pathway level. Alterations in the organization of cell–cell junction proteins orchestrate the activation of EMT, with important consequences on the expression of these markers at cellular level. As we demonstrated here, when we compared western blot data from the secretomes with that collected from cell lysates, epithelial and mesenchymal breast cancer cells showed a different expression of the protein components of the E-cadherin complex (Fig. 1Fig. 1A). Similar findings were described for the secretome of MCF-7 and MCF-7 shEcad cells. Down-regulation of E-cadherin induced changes in the expression of E-cadherin complex proteins at cellular level with a consequent reduction of their secreted levels (Fig. 1Fig. 1B). This confirms that the secretome is functionally linked to the cell phenotype, and that the release of proteins is a time-dependent process, as demonstrated by ELISA (Fig. 1Fig. 1C). With the aim to transfer these biological results in the realization of an SPR–based immunoassay, we decided to focus our attention on E-cadherin. To determine the analytical potential of our system, we determined sEcad level in secretomes and serum samples by ELISA. As shown in Fig. 2Fig. 2A and B, we selected a panel of epithelial breast cancer models and we determined sEcad levels in their secretomes, and E-cadherin levels in their cellular lysates. Data reported in the histograms describe the sEcad levels detected for each concentrated conditioned media samples after 4 h of serum starvation (Supplementary Fig. 1), with a range to less than 50 ng/mL to 200 ng/mL. Serum levels of E-cadherin were significantly higher compared to secretomes, and significantly increased in breast cancer samples (Fig. 2Fig. 2C). Overall, these experiments allowed assessing the concentration range of E-cadherin in secretome and serum samples for the development of QCM-sensor and SPR application. To do this, we assembled an E-cadherin sensing layer using different functionalization steps (consisting of MUA/ProtA/EA/Ab). AFM, WCA and FT-IR studies, shown in Fig. 3Fig. 3B and Supplementary Fig. 2, and the values reported in Table 1Table 1 performed at the different functionalization steps, give a clear indication of the multilayer structures assembled on the top of the sensor.
Subsequently, QCM-Ab experiments were performed to calculate the amount of Ab on the top of the functionalized sensors corresponding to 703 ng/cm2. Considering that from the AFM estimation, the area occupied by a single Ab is 491 nm2, and that the Ab molecular weight is 150 KDa, we calculated a concentration of immobilized Ab of 4.69 × 10−12 mol cm−2. This corresponds to a sensor surface coverage of 91%, which means that a homogeneous distribution of Ab molecules has been achieved using the proposed functionalization method.
QCM data reported in Supplementary Fig. 3, permitted to determine a correlation between the levels of Ecad in solution and those adsorbed onto the sensor surface. When the sensor was incubated with a solution of Ecad concentrated 2 μg/mL, 113.02 ± 4.63 ng/cm2 of protein were immobilized. A linear trend was observed with a slope of 51.98 ± 3.61 μL/cm2. If the sensor surfaces were saturated, a concentration of 4,69 × 10−12 mol cm−2 should be present. The weight of Ecad calculated via QCM would be of 281 ng/cm2 corresponding to a solution of 5.4 μg/mL, which represents the highest Ecad concentration that can be measured using the optimized functionalization. This is the range of soluble Ecad concentration described in several studies or public databases (http://www.plasmaproteomedatabase.org/). Therefore, this preliminary study demonstrates that the developed functionalization is efficient enough to allow the detection of E-cad in sera without the risk to saturate the active sites. As further result, FT-IR spectra of Fig. 3Fig. 3D confirm that the mass increase, registered by QCM, is due to the bond of E-cad to the sensing layer. All these experimental evidences were exploited to develop a SPR system, tested for the detection of E-cadherin into DMEM, cells secretome and human sera. Even in this case, FT-IR spectroscopy characterization was used to ensure the functionalization and E-cadherin binding, allowing the identification of the protein marker bands in the assembled multi-layer. Then, the system response to different concentrations of E-cadherin in DMEM solutions was investigated, showing a dynamic response up to 500 ng/mL of the protein. For this reason, a static range was found between 0 and 200 ng/mL and a calibration linear curve ΔAOIvs protein concentration (ng/mL) was obtained and is reported in Fig. 4Fig. 4C (correlation coefficient of 0.9839). The limit of detection (LOD) was calculated from the slope of the calibration linear curve and the standard deviation on multiple measurements of the blank sample (n = 5), as previously described30.x30.Bianco, M., Sonato, A., De Girolamo, A., Pascale, M., Romanato, F., Rinaldi, R. et al. An aptamer-based SPR-polarization platform for high sensitive OTA detection. Sens Actuators B Chem. 2017;
Crossref | Scopus (17) | Google ScholarSee all References,31.x31.Shrivastava, A. and Gupta, V. Methods for the determination of limit of detection and limit of quantitation of the analytical methods. Chronicles Young Sci. 2011;
Crossref | Google ScholarSee all References:
where σ is the standard deviation of the blank sample, F is a factor of 3.3 and b is the slope of the regression line. A LOD of 16 ± 6.5 ng/mL of E-cadherin was found, corresponding to a concentration of about 200 picomolar that is 6 times higher than the LOD of the employed ELISA test (100 ng/mL; ~1.25 nanomolar). The analytical sensitivity was calculated by the ratio of the amount of detected protein (ng/mL) and the ΔAOI (degrees) correspondent, and was found to be 300 (ng/ml)/°. Overall, this highlights the feasibility of our system in the detection of tissue leakage markers whose concentration in plasma ranges from 106 to 102 pg/mL33. The attractiveness of this method relies also in the absence of pre-concentration step as we determined Ecad in serum using only few μl of sample. Sensitivity should be further improved to characterize signal proteins such as cytokines and growth factors and this is a common issue for SPR detection of protein in biological fluids.33.x33.Li, G., Li, X., Yang, M., Chen, M.M., Chen, L.C., and Xiong, X.L. A gold nanoparticles enhanced surface plasmon resonance immunosensor for highly sensitive detection of ischemia-modified albumin. Sensors (Basel). 2013;
Crossref | Scopus (15) | Google ScholarSee all References,34.x34.Huang, C.F., Yao, G.H., Liang, R.P., and Qiu, J.D. Graphene oxide and dextran capped gold nanoparticles based surface plasmon resonance sensor for sensitive detection of concanavalin a. Biosens Bioelectron. 2013;
Crossref | Scopus (56) | Google ScholarSee all References,35.x35.Azzam, E., Abd El-aal, A., Shekhah, O., Arslan, H., and Wöll, C. Fabrication of spr nanosensor using gold nanoparticles and self-assembled monolayer technique for detection of cu2+ in an aqueous solution. J Dispers Sci Technol. 2013;
Google ScholarSee all References In this direction, a possible approach, widely employed in the literature, could be based on the amplification of the signal by introducing colloidal gold nanoparticles36.x36.Kwon, M.J., Lee, J., Wark, A.W., and Lee, H.J. Nanoparticle-enhanced surface plasmon resonance detection of proteins at attomolar concentrations: comparing different nanoparticle shapes and sizes. Anal Chem. 2012;
Crossref | Scopus (102) | Google ScholarSee all References,37.x37.Cao, Y., Griffith, B., Bhomkar, P., Wishart, D.S., and McDermott, M.T. Functionalized gold nanoparticle-enhanced competitive assay for sensitive small-molecule metabolite detection using surface plasmon resonance. Analyst. 2017;
Crossref | Google ScholarSee all References or plasmon metal nanostructures,38.x38.Indutnyi, I., Ushenin, Y., Hegemann, D., Vandenbossche, M., Myn’ko, V., Lukaniuk, M. et al. Enhancing surface Plasmon resonance detection using nanostructured au chips. Nanoscale Res Lett. 2016;
Crossref | Scopus (0) | Google ScholarSee all References,39.x39.Byun, K.M., Yoon, S.J., Kim, D., and Kim, S.J. Experimental study of sensitivity enhancement in surface plasmon resonance biosensors by use of periodic metallic nanowires. Opt Lett. 2007;
Crossref | Scopus (108) | Google ScholarSee all References allowing a detection down to femtomolar amount of proteins.