Peptide EBP-35 was conjugated to slides using a 16-well ProPlate tray (0.28 cm2). J82 cells labeled with CellTracker Red were applied to individual wells (384, 192, 96, or 48 cells; N = 4 wells each). After 15 min, unbound cells were washed away, and captured cells were fixed and counted manually by fluorescence microscopy. A no-peptide control was included for the lowest cell input.Cytology preparations
For experiments with cancer cell lines, J82 urothelial cells (7.5 × 104 cells) and anti-coagulated whole human blood (100 ml) were spiked into fresh urine (20 µL). Cells were recovered by centrifugation (500 × g, 10 min), suspended in capture buffer, and either added directly to peptide-conjugated slides (EBP-35) and unmodified controls or fixed with 4% paraformaldehyde in suspension, centrifuged, suspended in capture buffer, and smeared on poly-L-lysine slides. After 15 min, peptide-conjugated slides and unmodified controls were rinsed in wash buffer and fixed, and all slides were stained with hematoxylin and eosin (H&E).
For endogenous cell capture experiments, a voided urine specimen (40 mL; second morning void) was obtained from a healthy female donor, and cells were isolated by centrifugation (500 × g, 10 min). The cell pellet was washed in distilled water, centrifuged, suspended in capture buffer, and smeared on poly- L-lysine slides or deposited on peptide-conjugated slides (15 min). The peptide-conjugated slides were washed 3 times in wash buffer, after which all slides were fixed in 4% paraformaldehyde, dried, and subjected to Papanicolaou or H&E staining. Using bright field microscopy, Papanicolaou stained slides were qualitatively observed (40× objective), while the number of small cells (basal and intermediate) and large cells (squamous and superficial) bound to H&E stained slides were manually counted (10× objective).Immunocytochemistry (ICC)
A suspension of urothelial cells (J82) in capture buffer was applied to peptide-conjugated slides for 15 min, after which slides were washed 3 times in wash buffer to remove unbound cells. As controls, cells were smeared on positively charged slides or poly-L-lysine slides without rinsing. All slides were fixed in 4% paraformaldehyde, dried for 30 min at 37°C, and then washed with distilled water and PBS. For carcinoembryonic antigen (CEA) staining, slides were covered with blocking buffer (1 h at room temperature) then immunostained with rabbit IgG anti-CEA (1:100). For survivin staining, cells were permeabilized with Triton X-100 (0.5% in PBS) for 10 min prior to blocking, then labeled with rabbit monoclonal anti-survivin (1:100). Survivin-labeled cells were also labeled with red fluorescent phalloidin (1:400) to positively identify whole cells. All slides were incubated overnight at 4°C, washed in PBS, labeled with goat anti-rabbit Alexa Fluor 488 (30 min, 37°C), then washed 3 times in PBS. To visualize nuclei, CEA-labeled slides were counterstained with DAPI (200 ng/mL) for 2 min, followed by a final PBS wash. Slides were imaged by fluorescence microscopy (40× objective). For CEA-labeled slides, Image-Pro Plus 7.0 software was used to quantify the CEA expression area in each cell according to published methods (20). Briefly, a color standard was established such that negative controls (no CEA antibody) demonstrated no CEA-positive signal when the standard was applied. The color standard was applied to all slides, and average CEA expression area per cell was calculated by averaging across all analyzed cells for each slide type (8 images per slide, 40× objective).Results and discussion Isolation of urothelial cell binding peptides
Phage display was performed against T24 and J82 urothelial cells with 10 different phage libraries, representing over 20 billion peptide sequences. Cell-surface biopanning revealed a series of phage that bound to T24 and J82 urothelial cells but did not bind to WBCs or RBCs. Synthetic peptides based on six of these sequences were synthesized for the work presented here (see Table 1). Peptide purity, as determined by RP-HPLC, was ≥92%. The synthetic peptides have micromolar affinity for J82 and T24 urothelial cells, and they do not bind to blood cells or platelets (as determined by FACS; data not shown).Conjugation of urothelial cell binding peptides to glass slides
The peptides in Table 1 were conjugated to NHS glass slides through the lysine epsilon amine, and the surface density of peptide on the slides was found to be 29.9 pmol/cm2 (N = 2). Hydrolyzing the NHS groups prior to conjugation eliminated peptide attachment to the slides, confirming covalent attachment. The reproducibility of peptide attachment was assessed by preparing slides on 4 separate days, with an average peptide density of 30.0 ± 4.0 pmol/cm2 (%CV 13.5). Although it was possible to obtain higher or lower peptide densities by varying the peptide input concentration, subsequent experiments revealed that higher densities did not increase cell capture, while lower densities led to decreased cell capture (data not shown). Thus, 200 µM was selected as the optimal peptide input concentration.