Logo of neoplasiaLink to Publisher's site
Neoplasia. 2006 Sep; 8(9): 772–780.
PMCID: PMC1584300

In Vivo Selection of Phage for the Optical Imaging of PC-3 Human Prostate Carcinoma in Mice1


There is an increasing medical need to detect and spatially localize early and aggressive forms of prostate cancer. Affinity ligands derived from bacteriophage (phage) library screens can be developed to molecularly target prostate cancer with fluorochromes for optical imaging. Toward this goal, we used in vivo phage display and a newly described micropanning assay to select for phage that extravasate and bind human PC-3 prostate carcinoma xenografts in severe combined immune deficiency mice. One resulting phage clone (G1) displaying the peptide sequence IAGLATPGWSHWLAL was fluorescently labeled with the near-infrared fluorophore AlexaFluor 680 and was evaluated both in vitro and in vivo for its ability to bind and target PC-3 prostate carcinomas. The fluorescently labeled phage clone (G1) had a tumor-to-muscle ratio of ~ 30 in experiments. In addition, prostate tumors (PC-3) were readily detectable by optical-imaging methods. These results show proof of principle that diseasespecific library-derived fluorescent probes can be rapidly developed for use in the early detection of cancers by optical means.

Keywords: Phage display, optical imaging, drug development, near-infrared (NIR), tumor-imaging agents


In the United States, the incidence of prostate cancer in men is second only to that of skin cancer and accounts for over 40,000 deaths annually [1]. Although there is evidence that prostate-specific antigen screening can detect early-stage prostate cancer, it is nonspecific and can miss up to 30% of carcinomas [2]. Furthermore, markers for distant prostate cancer metastases are needed. Recent efforts have focused on identifying new prostate cancer-specific cell surface antigens, receptors, and other locally expressed biomarkers. Peptides, due to their ease of synthesis, often ideal pharmacokinetic characteristics, and nonimmunogenic nature, are attractive targeting agents for the development of new prostate cancer imaging or therapeutic agents.

The display of peptide libraries on the surface of bacteriophage (phage) offers a way of searching for peptides with specific binding properties. In vitro phage display has been used to select peptides that target cancer-associated antigens [3,4], whereas cultured carcinoma cells have also been employed to isolate peptides that bind a variety of human carcinoma cell lines [5–7]. In vivo phage display selection procedures offer an advantage over in vitro screening protocols in that phage can be selected based on desired pharmacokinetic properties, including delivery and tumoral accumulation. Recently, in vivo phage display has been explored as a means to identify phage and corresponding peptides with optimal tumor-targeting properties in the context of living animals; however, many of these peptides bind to endothelial cell markers but not directly to tumor cells [8,9]. We believe that in vivo-selected phage can serve as valuable first-line agents to determine if the phage and corresponding synthesized peptides would function as efficacious tumor-targeting and tumor-imaging agents.

Here, we report an in vivo phage display selection protocol and a micropanning assay that allowed the identification of novel prostate cancer-targeting peptides. Identified phage clones were able to extravasate from the endothelium and to specifically target prostate cancer cells in vivo. More precisely, prostate tumor-avid phage was selected from a fUSE5 15-amino-acid peptide library [10] in severe combined immune deficiency (SCID) mice bearing heterotransplanted human PC-3 prostate tumors. An innovative “tumor-to-cell micropanning” assay was used to distinguish phage with high affinity for prostate tumor tissues/heterotransplanted cell lines relative to normal tissues and cell lines. A single-phage clonal population with a high affinity for both prostate tumor tissues and PC-3 carcinoma cell lines was fluorescently labeled with the nearinfrared fluorophore (NIRF) AlexaFluor 680 (AF680), and its ability to bind prostate carcinomas, both in vitro and in vivo, was analyzed. AF680-labeled phage were specific for PC-3 cells in vitro and allowed the successful noninvasive imaging of prostate tumors in SCID mice.

Materials and Methods


Cell culture reagents were purchased from Invitrogen (Carlsbad, CA). All other chemicals were purchased from Sigma Chemical Co. (St. Louis, MO), unless otherwise stated.

Cell Lines

PC-3 cells [11] were grown in Ham's F12K medium, 7% fetal bovine serum (FBS), 2 mM L-glutamine, and 48 µg/ml gentamicin at 37°C in 5% CO2. Human embryonic kidney (HEK) 293 cells were cultured in Dulbecco's modified Eagle's medium (DMEM)-high glucose with 10% heat-inactivated FBS, 2 mM L-glutamine, 1 mM sodium pyruvate, and 48 µg/ml gentamicin at 37°C in 5% CO2. All other cell lines were grown in a general maintenance medium containing RPMI 1640 (custom) with 10% FBS and 48 µg/ml gentamicin. The cell lines were tested for pathogens before injection into mice.

Mouse Strains and Handling

Four- to 6-week-old (approximately 20 g) ICRSC-M SCID outbred mice were obtained from Taconic (Germantown, NY) and maintained in an approved pathogen-free institutional housing. Animal studies were conducted as outlined in the NIH Guidelines for the Care and Use of Laboratory Animals and the Policy and Procedures for Animal Research of the Harry S. Truman Veterans Memorial Hospital and the Massachusetts General Hospital. Solid tumors were established in SCID mice over a period of 4 weeks, resulting in mice with appropriately 1-cm-sized tumors for all experiments. PC-3 prostate carcinoma cells (5 x 106) were injected subcutaneously in one flank of each anesthetized animal. Tumors generally appeared in 30 days. Following in vivo selection or biodistribution, the mice were euthanized, and tumors and organs were excised from the animals and handled as described in later sections.

In Vivo Selection of Tumor-Targeting Phage

A fUSE5 phage display library, which displayed random 15 amino acid peptides on coat protein III, was a gift from Dr. George Smith [10]. Library amplification was conducted as described previously [12]. The phage particle concentration, in virions (V), was determined spectrophotometrically, and the amount of infectious units (transducing units, TU) was determined by titering on Escherichia coli K91 Blue Kan.

About 1012 TU of a 15-amino-acid fUSE5 library suspended in phosphate-buffered saline (PBS) was injected into the tail vein of normal non-tumor-bearing CF-1 mice for 15 minutes to preclear phage that bound to the normal vasculature and other nontumor antigens [13]. The mice were then anesthetized, and blood (~ 3 ml) was obtained. Phage were isolated from the blood by polyethylene glycol (PEG) precipitation, amplified, purified, and dialyzed, as described previously [12]. Next, 1012 TU of the precleared phage library was injected into SCID mice bearing PC-3 human prostate cancer cell xenografts. After 1 hour of in vivo phage circulation, the mice were sacrificed by cervical dislocation. The animals were perfused with 90 ml of PBS to facilitate phage elimination from the vasculature, and organs and tumors were removed, weighed, and quick-frozen in liquid nitrogen. Prostate tumors were ground in DMEM with protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 20 µg/ml aprotinin, and 1 µg/ml leupeptin) and 0.5% bovine serum albumin (BSA) (DMPB) and washed thrice by centrifugation at 6000 rpm. Pelleted lysate was resuspended in DMPB + 2.5% CHAPS to recover extravasated phage, and the recovered phage population was used to infect fresh E. coli K91 Blue Kan host cells. A portion of purified phage preparation (in PBS) was used as the input phage for the next round of in vivo selection. In total, four rounds of selection were performed. Between rounds, foreign DNA inserts and encoded peptide sequences were determined for 10 random phage clones to gauge the selection process and the absence of contamination. The amino acid sequences of displayed peptides were deduced from the DNA sequence. Sequences from selections were compared using Align, PIR, NiceBlast, gapped, and PSI-BLAST search databases [14,15] to identify the presence of consensus motifs or potential tumor targets and to define clones for further analyses.

Micropanning Assay

A newly described multistep “micropanning” assay was devised to quickly and efficiently identify clones that specifically recognized PC-3 tumor tissues and PC-3 human prostate carcinoma cells and, thus, had potential in in vivo PC-3 tumor targeting. In the first tier of micropanning, 1.5 x 107 TU of in vivo-selected phage clones was incubated at 4°C for 2 hours with excised normal muscle tissues or PC-3 tumor tissues. The tissues were then washed five times with DMPB, and bound phage were eluted from the tissues using DMPB + 2.5% CHAPS. The ratio of phage titer in tumor tissues to phage titer in normal tissues was calculated. In vivo-selected phage clones that recognized tumor tissues with a 1.5-fold specificity over normal tissues were assumed to have sufficient affinity and specificity for prostate tumor cells to warrant further investigation. In the second tier of the assay, phage that specifically bound to cultured PC-3 cells versus a nonrelevant control cell line were identified. Nine phage clones identified in the first tier were incubated (similar to the first tier) with either intact PC-3 prostate carcinoma cells or nonrelevant control cells (HEK293). Bound phage were eluted with DMPB + 2.5% CHAPS, and the ratio of phage titer in PC-3 cells to phage titer in HEK293 cells was calculated.

Conjugation of AF680 to Phage

An AF680 carboxylic acid, succinimidyl ester 5-isomer (Invitrogen), was dissolved in dimethyl sulfoxide and added to a suspension of phage in 1 x coupling buffer (0.5 M Na3 citrate, 0.1 M NaHCO3, pH adjusted to 8.5 with NaOH) at a 1:1 molar ratio of AF680/coat protein VIII (1% dimethyl sulfoxide, final concentration). This mixture was then incubated at room temperature for 4 hours in the dark. Subsequently, ethanolamine at 150 mM (pH adjusted to 9.0 with HCl) was added for 2 hours in the dark at room temperature. The samples were then PEG-precipitated twice and dialyzed extensively against 50 mM Tris-HCl and 150 mM NaCl (pH 7.5; TBS) to remove excess hydrolyzed AF680.

Determination of the Binding of AF680-Labeled Phage to PC-3 Cells by Modified Enzyme-Linked Immunosorbent Assay (ELISA)

PC-3 carcinoma cells were grown to ~ 90% confluency in a tissue culture-treated 96-well plate (TPP, Trasadingen, Switzerland). The growth medium (previously described) was removed, and a new complete medium with the appropriate concentration (5 x 109 to 1 x 1011 V/ml) of AF680-labeled phage or unlabeled phage was added to the wells. The plate was incubated at 37°C for 1.5 hours then washed with icecold PBS for a number of times. Cells and bound phage were then fixed to the plate by the addition of PBS with 4% formaldehyde. The presence of phage was then probed by the addition of a rabbit polyclonal anti-phage antibody (courtesy of Dr. George Smith). The plate was then incubated at room temperature for 1 hour and extensively washed with TBS. A secondary anti-rabbit antibody conjugated to horseradish peroxidase (Santa Cruz Biotechnology, Santa Cruz, CA) was added and incubated at room temperature for 1 hour. The plate was washed with TBS + 1% Tween, and liquid peroxidase substrate [2,2-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid] was added and incubated for 10 minutes at room temperature. The addition of 1 % sodium dodecyl sulfate was used to stop the reaction. The plate was then read on a µ Quant Universal Microplate Spectrophotometer (Bio-Tek Instruments, Winooski, VT) at an absorbance of 405 nm using an endpoint assay.

Confocal Microscopy for the Determination of Cell Surface Binding

The binding of the AF680-labeled phage to PC-3, PC-3M, LNCaP, DU145, RPWE-1, MDA-MB-435, and HEK293 cells was determined using laser scanning confocal microscopy, as described previously [16]. Cells (1 x 104) were dried on a microscope slide and blocked with 6% BSA in 10 mM Tris. The slides were incubated with wild-type (WT) or G1-phage solutions (1 x 1011 V/ml phage in 10 mM Tris, pH 7.5, 1% BSA) at room temperature for 1 hour in the dark. A comparison of AF680 phage binding and biotinylated peptide binding to PC-3 carcinoma cells and HEK cells was also conducted using laser scanning confocal microscopy. The binding of biotinylated peptide (G1 and scrambled G1) was detected using 10 µg/ml NeutrAvidin-Texas Red (Molecular Probes, Eugene, OR; incubation at room temperature for 30 minutes in the dark). Laser scanning confocal microscopy was performed on a Bio-Rad Laboratories (Hercules, CA) MRC 600 confocal microscope (University of Missouri Molecular Cytology Core Facility).

Biodistribution of AF680-Labeled Phage

AF680-labeled phage (~ 109 TU) were injected intravenously into SCID mice bearing PC-3 prostate cancer cell xenografts. The mice were sacrificed at 5 minutes, 30 minutes, 2 hours, 4 hours, 6 hours, and 24 hours after injection. Mice were perfused with 90 ml of PBS, and organs and tumors were removed and quick-frozen in liquid nitrogen. Organs and tumors were ground in TBS buffer, and the amount of protein in each tissue homogenate was then determined using the Bio-Rad Protein Assay (Bio-Rad Laboratories). Tissue homogenates at 5 mg/ml protein were analyzed for fluorescent activity at appropriate wavelengths, using an IVIS 200 fluorescence reflectance imaging system (Xenogen Corp., Alameda, CA). To account for autofluorescence, organs from SCID mice that had been injected with only PBS were also collected and homogenized, and fluorescence was measured. The autofluorescence of each organ homogenate was then subtracted from experimental organ homogenates.

The presence of the phage particle's coat protein VIII within tissue homogenates was confirmed using immunoblotting. Tissue homogenates at a concentration of 5 mg/ml crude protein were electrophoresed on 16% Tricine gels and transferred to nitrocellulose membranes. Coat protein VIII was detected by incubating the membrane with a rabbit polyclonal anti-coat protein VIII phage antibody at room temperature for 2 hours, and then by extensive washing with TBS + 1% Tween. A goat anti-rabbit antibody conjugated to horseradish peroxidase was added and allowed to incubate at room temperature for 2 hours. The membrane was washed with TBS + 1% Tween, and chemiluminescent peroxidase substrate (SuperSignal West Pico; Pierce, Rockford, IL) was added then exposed to autorad film and developed.

Imaging of PC-3 Xenografted Tumors

Male SCID mice implanted with PC-3 cells, as outlined above, were chemically depilated over lesions 24 hours before in vivo imaging to minimize autofluorescence and to maximize signals. A fluorescence reflectance image was obtained before and after (immediately, 1, 4, and 24 hours) the intravenous injection of 109 TU of phage. Fluorescence reflectance imaging was performed using a bonSAI system (Siemens Medical, Erlangen, Germany) with the animals under gas anesthesia. Fluorescence signals were measured using ImageJ software (NIH, Bethesda, MD), choosing consistent exposure times. Tumors were resected for histopathology directly after the last time point.


A 15-amino-acid fUSE5 phage display library precleared of phage clones that bound to organs and vasculature antigens in normal CF-1 mice was amplified and used as input phage for the in vivo selection of PC-3 prostate tumor binding phage. After four rounds of successive in vivo selection, 96 phage clones were sequenced and compared to our laboratory's phage-peptide sequence database (data not shown). When compared to our database, 19 clones were found to be unique and were thus further pursued. Align, PIR, NiceBlast, gapped, and PSI-BLASTsearch algorithms were used to analyze foreign amino acid sequence inserts of the 19 clones; however, no significant similarities to other proteins were found [14,15]. The 19 clones were analyzed for specificity using a two-tier micropanning procedure. In the first tier of the assay, phage were micropanned against PC-3-derived tumor tissues and normal muscle tissues. Nine phage clones generated a tumor-to-muscle specificity ratio of ≥ 1.5 and were carried forward to be analyzed in the next stage of micropanning. The second tier of micropanning was used to select phage clones that bound preferentially to in vitro-cultured PC-3 tumor cells. Of the nine phage clones carried forward from the first tier of micropanning, two (D9 and E8) did not bind at all to cultured PC-3 cells, six had a ratio of < 1.0 (D4, D7, F1, F6, F12, and G8), and only one (G1: IAGLATPGWSHWLAL) had a ratio of > 2 (Table 1). The G1 phage produced a signal-to-noise ratio of > 1.5 in the first tier of micropanning and yielded the highest ratio in the second tier of micropanning, demonstrating the specificity of binding to PC-3-derived tumors and in vitro-cultured prostate carcinoma cells. Therefore, G1 was chosen for further characterization.

Table 1
Two-Step Phage Micropanning Assay to Evaluate the Specificity of Selected Phage Clones.

To directly detect phage and to explore their use in cancer detection, G1 and WT phage were labeled with AF680 at an average of ~ 600 AF680 molecules per phage particle. A modified ELISA was employed to investigate the effect of AF680 labeling on phage affinity for cultured PC-3 prostate carcinoma cells (Figure 1). The presence of AF680 on the surface of phage led to no significant change in the binding of G1 or WT phage to cultured PC-3 prostate carcinoma cells. At phage concentrations between 5.0 x 1010 and 1.0 x 1011 V/ml, AF680 G1 phage exhibited a ~ 3-fold increase in binding to PC-3 carcinoma cells compared to AF680 WT phage (Figure 1).

Figure 1
Binding of AF680-labeled phage and unlabeled phage to PC-3 human prostate carcinoma cells. PC-3 human prostate carcinoma cells were grown to 90% confluency in a 96-well plate. Phage at 5 x 109, 1 x 1010, 5 x 1010, or 1 x 1011 V/ml were added to wells ...

The binding patterns of NIRF-labeled phage were compared to those of free peptides to ensure that the binding properties of G1 phage were due to the presence of the peptide and not to potential inherent nonspecific binding of fd phage particles [17,18]. Laser scanning confocal microscopy was used to investigate and compare the binding of biotinylated free peptides and peptides displayed on AF680-labeled phage particles (Figure 2A). AF680-labeled G1 phage and biotinylated free G1 peptides both displayed high intensity, cell membrane-associated binding to cultured PC-3 cells, and little to no binding to HEK293 cells, demonstrating the specificity of the G1 peptide to PC-3 cells. In contrast, AF680-labeled WT phage and biotinylated free control peptides resulted in no significant binding to either cultured PC-3 or HEK293 cells. The specificity of binding of AF680-labeled G1 phage to other cultured cells was further investigated using confocal microscopy. It was found that AF680 G1 phage bound to the prostate carcinoma cell lines PC-3, PC-3M (highly metastatic subline derived from PC-3), LNCaP (lymph node metastasis), and DU145 (derived from brain metastasis) (Figure 2B). AF680 G1 phage bound with highest intensity to PC-3 cells and with slightly reduced intensity to the other human prostate carcinoma cell lines tested. Little to no binding of AF680 G1 phage to RPWE-1 (a normal human prostate cell line), MB-MDA-435 (a human breast carcinoma cell line), and HEK293 cells was observed. Micropanning data for these seven cell lines were consistent with these results (data not shown). These data demonstrate that the binding of the G1 phage is due to the presence of the G1 peptide and that the peptide is specific for prostate carcinomas.

Figure 2
Binding of AF680-labeled phage and peptides to human carcinoma cells and control normal cells. Slides containing fixed PC-3 or HEK293 cells were incubated with AF680-labeled phage solutions (1 x 1011 V/ml phage in 10 mM Tris, pH7.5, 1% BSA) or biotinylated ...

To determine the in vivo distribution of the G1-phage clone, we injected AF680-labeled G1 phage into SCID mice bearing subcutaneously implanted PC-3 cell-derived tumors (Figure 3A). Both the AF680 G1 phage and the WT phage cleared through the organs of the reticuloendothelial system (RES), such as the liver, lungs, and kidneys (Figure 3, A and B). AF680 phage began to amass in the RES in as early as 5 minutes and continued at 6 hours. Muscle, fat, brain, and spleen exhibited very low levels of fluorescence. The accumulation of AF680 G1 phage in the tumor began as early as 5 minutes and peaked from 4 to 6 hours. The tumor-to-fatuptake ratio for AF680 G1 phage was ~ 10 at 4 hours postinjection, in contrast to ~ 3 for AF680 WT phage. The tumor-to-muscle-uptake ratio for AF680 G1 phage was ~ 30 between 30 minutes and 6 hours postinjection. In contrast, the tumor-to-muscle-uptake ratio for WT phage was ~ 11 at 5 minutes postinjection and diminished to ~ 8 by 6 hours postinjection. Notably, AF680-labeled G1 phage yielded a 73% higher mean in tumor fluorescence over that of AF680-labeled WT phage at 4 hours postinjection. These data provide evidence for the binding of G1 phage to xenografted PC-3 tumors in SCID mice.

Figure 3
Biodistribution of AF680-labeled G1 fUSE5 and WT phage in PC-3 xenografted SCID mice. SCID mice bearing human PC-3 carcinoma tumors were injected with ~ 109 TU of AF680-labeled G1 fUSE5 phage (A) and AF680-labeled WT phage (B) and were allowed ...

The presence of phage particles in tissue homogenates was verified by immunoblotting (Figure 4A). Immunoblots demonstrated the presence of phage in the liver as well as in the tumor, confirming biodistribution results and the fluorescent distribution of AF680 (Figure 4B). These findings support the results of the fluorescent biodistribution studies of AF680 and indicate that the intensity of the AF680 fluorescent signal measured corresponds to the presence of phage in tissues and organs. Furthermore, phage isolated from liver and tumor extracts by PEG precipitation were spectrophotometrically scanned from 600 to 800 nm (for AF680) and from 240 to 320 nm (for phage particles), demonstrating that fluorescence and phage particles coprecipitated (data not shown). Taken together, these data suggest that AF680-labeled G1 phage are stable in vivo and, as such, could easily be used as an in vivo imaging agent.

Figure 4
Biodistribution of AF680-labeled G1 fUSE5 and WTphage in PC-3 xenografted SCID mice. Homogenized tissues from SCID mice bearing human PC-3 carcinoma tumors were probed for the fluorescent activity of AF680 label and for the presence of coat protein VIII. ...

To explore the usefulness of G1 phage as an in vivo imaging agent, we performed surface reflectance imaging with AF680-labeled G1 phage. AF680-labeled phage could be imaged in PC-3-derived tumors in SCID mice as early as 1 hour postinjection (Figure 5). The peak of signal intensity and specificity within the tumor occurred at 4 hours postinjection, consistent with the distribution data presented in Figure 3. At 4 hours postinjection of AF680 G1 phage, a 4.5-fold increase in the fluorescent signal within the tumor compared to that of the normal tissue was observed with a specificity of 2.1-fold (G1-to-WT phage tumoral accumulation). Together, these data confirm that the G1 peptide displayed on the tip of coat protein III has a direct influence on the extravasation, accumulation, and retention of AF680-labeled phage within the tumor and that this sequence is specific for prostate tumors.

Figure 5
In vivo imaging of a PC-3 prostate tumor using AF680-labeled G1 fUSE5 phage. SCID mice bearing PC-3 tumors were injected with AF680-labeled G1 or WTphage. Fluorescence reflectance images of anesthetized mice were obtained 0 minute, 1 hour, 4 hours, and ...


We selected in vivo phage that target human PC-3 prostate carcinomas heterotransplanted in SCID mice. The phage clone G1 displaying the peptide sequence IAGLATPGWSHWLAL bound both cultured PC-3 carcinoma cells in vitro and PC-3 tumors ex vivo. More importantly, fluorescently labeled versions of the phage resulted in increased signal intensity within PC-3 xenografted tumors in SCID mice. These results provide proof of principle that fd filamentous phage can indeed be used for the in vivo selection of extravasating phage that specifically target cancer cells, which can be fluorescently labeled for use as a noninvasive tumor-imaging tool.

Selection of extravasating phage that bound tumor cells was accomplished by first preclearing the 15-amino-acid fUSE5 phage display library in normal mice so as to remove unwanted phage that bound organs and normal vasculature components. The preselected library was incubated in tumor-bearing mice for times sufficient for extravasation to occur [13]. From this in vivo selection, the peptide inserts of 96 phage clones were determined and compared to our laboratory's database of phage-displayed peptide sequences. Only 19 clones were unique to this particular phage display selection, supporting the notion that phage display procedures often result in “winning” clones that may be obtained due to inherent undesirable properties such as high infectivity, good growth capabilities, or hydrophobic nonspecific binding properties [18]. A two-tiered micropanning assay was designed to identify which, if any, of the 19 clones had preferential affinity for both PC-3 prostate tumors and PC-3-cultured carcinoma cells. Because fd phage often exhibit nonspecific binding [17,18], the first tier of micropanning was used as an indication of the level of background noise (binding of phage to normal tissues), as well as the ability of the phage to specifically bind tumor tissues. However, a positive finding in such an assay does not guarantee the binding of phage to actual tumor cells because tumor tissues are obviously composed of tumor cells (in addition to connective tissues) and extracellular matrix components. The second tier of micropanning was then used as an indication of whether the selected phage bound to PC-3 cells themselves or bound normal cells such as HEK293 kidney cells. The ratio of the binding of PC-3 cells to HEK293 was calculated, and only one clone (G1) had a ratio of > 2 and was thus further investigated.

In vitro characterization of AF680 G1-phage binding properties, compared to that of AF680 WT, included both modified ELISA and confocal microscopy. These assays provided proof of AF680-labeled G1-phage binding to cultured PC-3 cells in vitro. G1 sequence was not specific for PC-3 cells in that binding to other human prostate cancer cells was observed. Importantly, confocal microscopy indicated that G1 phage bound with much higher affinity to cultured human prostate carcinoma cells than to RPWE-1 normal prostate, MB-MDA-435 breast carcinoma, or HEK293 embryonic kidney cell lines. Thus, the G1 sequence may target a prostate cancer-specific antigen. The very low background of AF680 WT phage observed in confocal microscopy (Figure 2) was likely due to the fact that cultured cells were prefixed with formaldehyde and, therefore, could be washed much more vigorously than live cells used for micropanning and modified ELISA. Both in vitro assays with AF680 G1 phage exhibited a ≤ three-fold increase in signal from AF680 G1 over that of AF680 WT.

In vivo characterization of the AF680 G1 phage included biodistribution studies as well as optical imaging. Biodistribution of AF680-labeled G1 and WT phage in PC-3 xenografted SCID mice exhibited similar distribution and clearance profiles. The muscle-to-tumor ratio found in in vivo distribution was much higher than that found in the micropanning assay. This is most likely due to the inability of large phage particles to extravasate in vivo into muscle tissues, which maintain very tight junctions between endothelial cells. In comparison, the micropanning assay involves the incubation of a large concentration of phage within homogenized tissues, thus allowing the characteristic nonspecific binding of phage. Another reason for the difference of muscle-to-tumor ratios found with micropanning versus in vivo biodistribution was the stringency of the micropanning assay. Micropanning experiments used very stringent washing and detergent elution techniques. Additional micropanning experiments using a more traditional acid elution [12] generated a recovered-phage-from-PC-3-to-recovered-phage-from-HEK293 ratio of 5, compared to that of 2.14 from the more stringent CHAPS detergent elution. Thus, the technique used within micropanning experiments has a direct impact on the number and type of phage recovered from tissues or cell pellets. The high background binding found for WT phage in in vitro assays is not unexpected, as phage are well known to bind to various supports such as plastic [19]. Our goal was not to reduce nonspecific in vitro binding but to find phage that extravasated and targeted tumors in vivo. The G1 phage had a tumor-to-muscle ratio of 30 in vivo, suggesting that high backgrounds in vitro do not necessarily correlate with in vivo tumor targeting.

In view of biodistribution data, we observed the clearance of phage particles through the RES, which is consistent with previous observations from our laboratory [13] and others [20,21]. It was found in the distribution studies that AF680-labeled G1 phage accumulated and bound to human prostate PC-3 tumors in vivo with a 73% higher mean over that of AF680-labeled WT phage. In vivo surface reflectance imaging with AF680 G1 phage resulted in a two-fold increase over the background signal from AF680-labeled WT phage. Both of these findings are in agreement with results from in vitro assays.

G1 phage and the corresponding peptide demonstrated preferential binding to human prostate carcinoma cell lines. In an effort to identify the targeted antigen, protein search algorithms were employed; however, no significant homologiestoG1 were found. This result is not surprising in that only a handful of antigens targeted by in vivo-selected peptides have been identified [18,22,23]. The majority of peptides targeted integrins through RGD-like or similar peptide motifs [9,23]. Furthermore, affinity chromatography experiments using the G1 sequence with extracts of cultured prostate cancer cell lines did not yield a candidate protein antigen (data not shown). Taken together, these results suggest that the target of G1 may be a sparse protein antigen or a complex antigen consisting of protein, lipid, and/or carbohydrate components [24]. Our findings further support the idea that in vivo selection of phage can be used to target rare and unique biomarkers on cancer cells, not merely vascular components.

The use of labeled phage as an alternative to labeled peptides for the screening of tumor-targeting agents has several advantages. The first and foremost advantages are affordability and speed, because the synthesis and labeling of multiple peptides are a more costly and lengthy procedure. The screening of phage particles allows for more flexibility than the screening of peptides. The initial screening of phage displaying selected peptides can be accomplished biologically by tracking the phage spectrophotometrically or by titering on host E. coli, whereas peptides must be tagged. The addition of fluorophores, radiochelators, and epitope tags can have negative effects on the binding properties of peptides [25]. Hence, implementation of peptides displayed on phage may be a useful gauge of the ability of synthesized peptides to bind to their target before laborious peptide modifications ensue.

Because phage are organic and nonpathogenic [20,21], they may be thought of as multifunctional self-replicating biologic nanoparticles [26] in that they can be covalently attached to numerous tags or labels while simultaneously expressing multiple copies of foreign peptides. The resulting signal amplification is of enormous benefit to the peptide screening process and may expedite the discovery and characterization of a “dark horse” peptide without the documented side effects of other nanoparticles with metallic cores (such as cadmium-selenium or other metals) [27–30]. Signal amplification is also beneficial to the characterization of phage displaying peptides in in vivo models of diseases such as cancer. It allows for a quick and easy characterization of the distribution profile of peptides. As phage have their own distinct in vivo distribution, the peptide-bearing phage would need to be compared to phage particles with no displayed peptides. In general, the use of phage as selfreplicating bionanoparticles is a very safe, simple, and attractive solution for the development and implementation of new prostate screening modalities.


The authors would like to acknowledge the contributions of Tiffani Shelton, Marie T. Dickerson, Cynthia M. Illy, Linda A. Landon, and Rabi Upadhyay.


AlexaFluor 680
human embryonic kidney
near-infrared fluorophore
polyethylene glycol
reticuloendothelial system
severe combined immune deficiency
transducing units
wild type


1This work was supported, in part, by a Merit Review Award from the Veterans Administration (S.L.D.), Department of Defense DAMD17-03-1-0130 (S.L.D.), National Institutes of Health (NIH) P50 CA103130-01 (S.L.D.), and the University of Missouri's Radiosciences Institute (J.N.) and, in part, by NIH R24 CA92782 (R.W.) and NIH P50 CA86355 (R.W. and K.K.).


1. Jemal A, Murray T, Ward E, Samuels A, Tiwari RC, Ghafoor A, Feuer EJ, Thun MJ. Cancer statistics, 2005. CA Cancer J Clin. 2005;55:10–30. [PubMed]
2. Centers for Disease Control, author. Fact sheet prostate cancer: The public health perspective. 2003 ( www.cdc.gov/cancer/prostate/prostate.htm)
3. Poul MA, Becerril B, Nielsen UB, Morisson P, Marks JD. Selection of tumor-specific internalizing human antibodies from phage libraries. J Mol Biol. 2000;301:1149–1161. [PubMed]
4. Karasseva N, Glinsky VV, Chen NX, Komatireddy R, Quinn TP. Identification and characterization of peptides that bind human ErbB-2 selected from a bacteriophage display library. J Protein Chem. 2002;21:287–296. [PubMed]
5. Peletskaya EN, Glinsky VV, Glinsky GV, Deutscher SL, Quinn TP. Characterization of peptides that bind the tumor-associated Thomsen-Friedenreich antigen selected from bacteriophage display libraries. J Mol Biol. 1997;270:374–384. [PubMed]
6. Kelly KA, Jones DA. Isolation of a colon tumor specific binding peptide using phage display selection. Neoplasia. 2003;5:437–444. [PMC free article] [PubMed]
7. Romanov VI, Durand DB, Petrenko VA. Phage display selection of peptides that affect prostate carcinoma cells attachment and invasion. Prostate. 2001;47:239–251. [PubMed]
8. Arap W, Pasqualini R, Ruoslahti E. Cancer treatment by targeted drug delivery to tumor vasculature in a mouse model. Science. 1998;279:377–380. [PubMed]
9. Pasqualini R, Ruoslahti E. Organ targeting in vivo using phage display peptide libraries. Nature. 1996;380:364–366. [PubMed]
10. Smith GP. Filamentous fusion phage: novel expression vectors that display cloned antigens on the virion surface. Science. 1985;228:1315–1317. [PubMed]
11. Kaighn ME, Narayan KS, Ohnuki Y, Lechner JF, Jones LW. Establishment and characterization of a human prostatic carcinoma cell line (PC-3) Invest Urol. 1979;17:16–23. [PubMed]
12. Smith GP. G. P. Smith Lab Homepage. 2006 ( www.biosci.missouri.edu/smithGP/)
13. Zou J, Dickerson MT, Owen NK, Landon LA, Deutscher SL. Biodistribution of filamentous phage peptide libraries in mice. Mol Biol Rep. 2004;37:121–129. [PubMed]
14. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 1997;25:3389–3402. [PMC free article] [PubMed]
15. George DG, Barker WC, Mewes HW, Pfeiffer F, Tsugita A. The PIR—International Protein Sequence Database. Nucleic Acids Res. 1996;24:17–20. [PMC free article] [PubMed]
16. Landon LA, Peletshaya EN, Glinsky VV, Karasseva N, Quinn TP, Deutscher SL. Combinatorial evolution of high affinity peptides that bind to the Thomsen-Friedenreich carcinoma antigen. J Protein Chem. 2003;22:193–204. [PubMed]
17. Giordano RJ, Cardo-Vila M, Lahdenranta J, Pasqualini R, Arap W. Biopanning and rapid analysis of selective iterative ligands. Nat Med. 2001;7:1249–1253. [PubMed]
18. Spear MA, Breakefield XO, Beltzer J, Schuback D, Weissleder R, Pardo FS, Ladner R. Isolation, characterization, and recovery of small peptide phage display epitopes selected against viable malignant glioma cells. Cancer Gene Ther. 2001;8:506–511. [PubMed]
19. Adey NB, Mataragnon AH, Rider JE, Carter JM, Kay BK. Characterization of phage that bind plastic from phage-displayed random peptide libraries. Gene. 1995;156:27–31. [PubMed]
20. Merril CR, Scholl D, Adhya SL. The prospect for bacteriophage therapy in western medicine. Nat Rev Drug Discov. 2003;2:489–497. [PubMed]
21. Inchley CJ. The activity of mouse Kupffer cells following intravenous injection of T4 bacteriophage. Clin Exp Immunol. 1969;5:173–187. [PMC free article] [PubMed]
22. Kelly K, Alencar H, Funovies M, Mahmood U, Weissleder R. Detection of invasive colon cancer using a novel, targeted, library-derived fluorescent peptide. Cancer Res. 2004;64:6247–6251. [PubMed]
23. Koivunen E, Wang B, Ruoslahti E. Isolation of a highly specific ligand for the alpha 5 beta 1 integrin from a phage display library. J Cell Biol. 1994;124:373–380. [PMC free article] [PubMed]
24. Burke HB. Proteomics: analysis of spectral data. Cancer Inf. 2005;1:15–24. [PMC free article] [PubMed]
25. Reubi JC. Peptide receptors as molecular targets for cancer diagnosis and therapy. Endocr Rev. 2003;42:389–427. [PubMed]
26. Mao C, Christine FE, Hayhurst A, Sweeney R, Qi J, Georgiou G, Iverson B, Belcher AM. Viral assembly of oriented quantum dot nanowires. Proc Natl Acad Sci USA. 2003;100:6946–6951. [PMC free article] [PubMed]
27. Lovric J, Bazzi HS, Cuie Y, Fortin GR, Winnik FM, Maysinger D. Differences in subcellular distribution and toxicity of green and red emitting CdTe quantum dots. J Mol Med. 2005;83:377–385. [PubMed]
28. Shiohara A, Hoshino A, Hanaki K, Suzuki K, Yamamoto K. On the cyto-toxicity caused by quantum dots. Microbiol Immunol. 2004;48:669–675. [PubMed]
29. Voura EB, Jaiswal JK, Mattoussi H, Simon SM. Tracking metastatic tumor cell extravasation with quantum dot nanocrystals and fluorescence emission-scanning microscopy. Nat Med. 2004;10:993–998. [PubMed]
30. Ballou B, Lagerholm BC, Ernst LA, Bruchez MP, Waggoner AS. Noninvasive imaging of quantum dots in mice. Bioconjug Chem. 2004;15:79–86. [PubMed]

Articles from Neoplasia (New York, N.Y.) are provided here courtesy of Neoplasia Press
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...