Simultaneous Detection of EGFR and VEGF in Colorectal Cancer using Fluorescence-Raman Endoscopy

Fluorescence endomicroscopy provides quick access to molecular targets, while Raman spectroscopy allows the detection of multiple molecular targets. Using a simultaneous fluorescence-Raman endoscopic system (FRES), we herein demonstrate its potential in cancer diagnosis in an orthotopically induced colorectal cancer (CRC) xenograft model. In the model, epidermal growth factor receptor (EGFR) and vascular endothelial growth factor (VEGF) were targeted with antibody-conjugated fluorescence and surface-enhanced Raman scattering (F-SERS) dots. FRES demonstrated fast signal detection and multiplex targeting ability using fluorescence and Raman signals to detect the F-SERS dots. In addition, FRES showed a multiplex targeting ability even on a subcentimeter-sized CRC after spraying with a dose of 50 µg F-SERS dots. In conclusion, molecular characteristics of tumor cells (EGFR in cancer cell membranes) and tumor microenvironments (VEGF in the extracellular matrix) could be simultaneously investigated when performing a colonoscopy.


Immobilization of antibodies on F-SERS dots
To introduce an amino-functional group onto the surface of the F-SERS dots, 1 mg of the latter was treated with a 1 mL APTES solution (5% volume in ethanol) and 10 μL NH 4 OH (27%) for 1 h at room temperature. After washing with ethanol, the amine-3 functionalized F-SERS dots (1 mg) were then reacted with 1.8 mg succinic anhydride and 3 μL DIPEA in 500 μL NMP for 2 h to introduce carboxylic acid groups. Subsequently, the carboxylic acid group-functionalized F-SERS dots were activated with 2 mg NHS, 2.7 μL DIC, and 0.2 mg DMAP for 2 h. To remove excess reagents, the resulting solution was centrifuged and washed with NMP and PBS (pH 7.4). For antibody conjugation, cetuximab (50 μg) or bevacizumab (50 μg) was added either to the NHS-activated F-SERS-A or -B dots, respectively, dispersed in 200 μL of PBS. After incubation for 1 h at room temperature, the antibody-immobilized F-SERS dots were centrifuged and washed consecutively with PBS containing 0.1% (w/v) Tween 20 and then with PBS. Finally, to improve bio-compatibility, the antibody-immobilized F-SERS dots were treated with BSA [1% (w/v) in PBS solution, pH 7.4] for 30 min, washed with PBS solution containing Tween 20, and then with PBS, as described above.

Size and concentration measurements
Size distribution was measured using the nanoparticle tracking analysis (NTA) method (NanoSight NS500; Malvern, Worcestershire, UK). Samples were sufficiently diluted to minimize interference between particles. The apparatus' quick measurement mode was performed to find optimal conditions, then 5 clips of particle motion videos were recorded automatically using the standard measurement mode. Captured videos (five videos per sample) were processed and analyzed. All other conditions were constant. Graphical figures were automatically drawn by the built-in software.
After the membranes had been washed three times with Tris-buffered saline, they were incubated with a secondary anti-rabbit/anti-mouse horseradish peroxidase-conjugated antibody (1:2,000 dilution; Santa Cruz Biotechnology, Santa Cruz, CA, USA) for 2 h, and enhanced chemiluminescence detection reagent (Thermo Scientific, Rockford, IL, USA).
Signal intensities were then measured using an LAS-3000 imaging system (Fuji Film, Tokyo, Japan).

Evaluation of the orthotopic CRC xenograft model
Mice were followed up every two days after orthotopic CRC xenograft modeling to confirm their survival. At the same time, mice were visually examined for signs of anal erosion. For bioluminescence imaging acquisition, an IVIS100 imaging system (Caliper Life Sciences, Hopkinton, MA, USA) was used. D-luciferin potassium salt was diluted to 0.3 mg/mL in PBS before use, and 100 μL of the D-luciferin solution was intraperitoneally injected into mice. The animals were anesthetized with isoflurane and transferred into a lighttight chamber equipped with a charge-coupled device camera operated by Live Image software (Xenogen Corp., Alameda, CA, USA) to obtain bioluminescence images that were serially acquired every 5 min until maximum signals were reached (time: 1 s, binning: medium, f stop: 1). Signal intensities were displayed on a pseudocolor scale. To quantify the emitted light, regions of interest (ROIs) were drawn on each bioluminescent image to quantify the optical flux (p/s/cm 2 /sr: photons per second per cm 2 per steradian) over each tumor area. The same-sized ROI was drawn for comparison over the brain area, and the tumor-to-background ratio (TBR; optical flux ratio between the tumor and the brain area) was calculated. Time points for bioluminescence image acquisition were one and two weeks for the injection of 1 × 10 7 HT29-effluc cells groups, and one week for the injection of 5 × 10 6 HT29-effluc cells group.

Confocal laser scanning microscopy (CLSM)
For the in vitro analysis, the interaction of antibody-conjugated F-SERS dots with colon cancer cells was investigated. HT29-effluc cells (10 4 cells/well) were seeded in an 8well chambered coverglass (Lab-Tek; Thermo Scientific, Rochester, NY, USA) with 300 μL of cell media per well. After 24 h incubation at 37 °C, cells were fixed with 4% paraformaldehyde (Wako, Osaka, Japan) for 20 min and then washed three times with PBS.

6
For the in vivo analysis, tumors were excised, fixed with 4% paraformaldehyde, and sectioned, followed by staining of the nuclei using DAPI. Confocal laser scanning microscopy (CLSM: LSM 510 META; Carl Zeiss, Jena, Germany) was used for fluorescence signal detection. Excitation laser lines for F-SERS dots and DAPI signals were 610 nm and 405 nm, respectively. Data were analyzed using LSM Image Examiner software (Carl Zeiss; Jena, Germany).

Pathologic evaluation
Immunohistochemistry (IHC) for EGFR and VEGF was carried out on formalinfixed, paraffin-embedded serial sections cut at 3 μm and dried at 37 °C overnight.
Parallel sections were stained with hematoxylin and eosin (H&E).

Statistical analysis
The correlation between protein concentration and luciferase activity in the in vitro study was evaluated using Spearman's correlation. Statistical significance of the in vivo bioluminescence studies was determined using a Wilcoxon signed rank test (TBR changes according to HT29-effluc injected cells week) and a Mann-Whitney U test (TBR difference according to HT29-effluc injected cells number). A P-value less than 0.05 was considered significant. All statistical analyses were performed using SPSS software (Version 18.0; SPSS Inc., Chicago, IL, USA) and MedCalc (Version 12.2; MedCalc Inc., Mariakerke, Belgium).  The relative intensities to the initial value of (a) fluorescence (at 625 nm for AF610) and (b) SERS signals (at 1324 cm -1 for FITC and at 1648 cm -1 for RITC). The fluorescence and SERS spectra were obtained from the same single F-SERS dot with continuous laser exposure (300 s). Inset: the individual spectra of fluorescence and SERS signals obtained by 532 nm photoexcitation at 3.6 mW on the sample and 1 s acquisition time. Fluorescence signal (a) and Raman intensity at 1648 cm -1 (b) gradually become definite as the cell number increases.  Figure 5. CLSM for the validation of multiplex targeting ability.
Tumors exposed to F-SERS dots (related to Fig. 5) were excised, fixed, and sectioned (nuclei were stained with DAPI, and CLSM was used for fluorescence signal detection).
Fluorescence signals were observed for both (a) and single (b, c) antibody-conjugated F-SERS dots sprayed on tumors. No signal was found for BSA-F-SERS-A/B dots sprayed on the tumors (d).
Supplementary Figure 6. Dot plot results for the real-time endoscopic system.
All of the colon cancer models showed definite or probable signals, but the normal colon study showed no signals. Mann-Whitney U test showed significant results between colon cancers and normal colons (all P < 0.001 for Raman A and Raman B signals). 14 Supplementary Figure 8. CLSM for sensitivity and lower dose limit detection.
Tumors exposed to antibody-conjugated F-SERS dots (related to Fig. 6) were excised, fixed, and sectioned (nuclei were stained with DAPI, and CLSM was used for fluorescence signal detection). After spraying 100 µg ( anti-EGFR and anti-VEGF antibodies (g, h) revealed no definite fluorescence or Raman signals. Dot plot results (i) demonstrated that no signal was found in all cases when using more than 50 µg of cold antibodies, however, some probable signals were found when using 25 µg of cold antibodies. Immunohistochemistry (IHC) results showed positivity for EGFR (j) and VEGF (k) in colon cancer. In contrast, negative results were found in normal colon. Figure 13. Full-length western blot analysis results of EGFR and VEGF.