Novel Decapeptides that Bind Avidly and Deliver Radioisotope to Colon Cancer Cells

Background The rapidly growing field of targeted tumor therapy often utilizes an antibody, sometimes tagged with a tumor-ablating material such as radioisotope, directed against a specific molecule. Methodology/Principal Findings This report describes the discovery of nine novel decapeptides which can be radioactively labeled, bind to, and deliver 32P to colon cancer cells. The decapeptides vary from one another by one to three amino acids and demonstrate vastly different binding abilities. The most avidly binding decapeptide can permanently deliver very high levels of radioisotope to the adenocarcinoma cancer cell lines at an efficiency 35 to 150 times greater than to a variety of other cell types, including cell lines derived from other types of cancer or from normal tissue. Conclusions/Significance This experimental approach represents a new example of a strategy, termed peptide binding therapy, for the potential treatment of colorectal and other adenocarcinomas.


INTRODUCTION
Illness and death due to colorectal and esophageal cancer constitute a monumental health care challenge in the United States and throughout the world [1,2]. Current treatments include radiotherapy, surgery, and chemotherapy. There are a number of immunotherapies approved for use in the treatment of various types of cancers (e.g., Herceptin, Rituxin, Avastin, and others) [3][4][5][6]. All of these immunotherapies utilize a monoclonal antibody directed against a specific cellular molecule [7,8]. Destructive action against tumor cells is thought to involve ADCC (antibodydependent cellular cytotoxicity), cellular lysis via the complement pathway, or the induction of apoptosis [9,10]. Avastin is a monoclonal antibody directed against VEGF (vascular endothelial growth factor) and is approved for treatment of colorectal cancer [11][12][13].
In addition, Non-Hodgkins lymphoma (NHL) is currently treated with two approved radioimmunotherapeutic regimens: Bexxar and Zevalin. Both utilize a monoclonal antibody directed against the B-cell marker CD20 and can deliver either 131 I (Bexxar) or 90 Y (Zevalin) isotopes to target lymphoma cells [14,15]. Beta-particles (electrons) generated by these isotopes can deeply penetrate cells and damage DNA, leading to cell death. However, there are currently no radioimmunotherapies approved for the treatment of patients with colorectal cancer.
The decapeptides described herein bind to and transfer isotope ( 32 P) to cell lines derived from several colorectal carcinomas. Under identical experimental conditions, very little (viz., less than 1% of the colon cancer cell lines' rates) of the most efficient 32 Plabeled decapeptides bind to cell lines established from a variety of other cancers or to normal colon, kidney, or esophageal cells.

RESULTS
We have identified nine decapeptides, differing from one another by only a few amino acids, that when labeled with 32 P can bind to a number of colorectal carcinoma cell lines. All decapeptides contain a protein kinase A substrate sequence and are designated as MAs (Modified Adjuvant). Figure 1 is a schematic representation of the production of the 32 P-labeled peptides and the experimental design of assays to measure binding of peptides to cell lines. Figure 2 displays the number of 32 P counts per minute (cpm) remaining bound to eighteen different cell lines and blank wells after a two hour incubation with MA5, the most efficient binding decapeptide (see below). The Caco-2 colon adenocarcinoma cell line retained the greatest number of radioactive counts after a twohour incubation and subsequent washes with complete medium, the average value of triplicate wells equaling 298,639 cpm per 10,000 cells. HCT116 colon adenocarcinoma cells retained an average value of 131,998 cpm per 10,000 cells. Blank wells and nonbinding cell lines had mean values of less than 550 cpm; bars representing these values are not visible at the scale used in Figure 2. For example, HeLa S3 cervical cancer cells only retained an average of 534 cpm per 10,000, HT1080 fibrosarcoma cells retained 367 cpm, and the human embryonic kidney cell line 293H retained 429 cpm per 10,000 cells.
Seven of the eighteen cell lines demonstrated very strong retention of radioactivity when incubated with MA5 (Modified Adjuvant radioactive peptide) with five of these being colon adenocarcinoma cell lines (Caco-2, HCT15, HCT116, LoVo, HT29), one being an esophageal adenocarcinoma cell line (SEG1), and one being a Barrett's esophagus cell line (QHTRT). In contrast, the eleven nonbinding cell lines were mostly squamous cell lines derived from carcinomas of the cervix (HeLa S3), colon (RKO), lung (1271, A549), esophagus (KYSE-70), a fibrosacroma (HT1080), or cells cultured from normal kidney (293H), colon (1459), or esophagus (HEEpiC). Nonbinding cell lines included T84, derived from a colon adenocarcinoma metastatic to lung, and SK-BR-3, isolated from a breast adenocarcinoma. The ratio of cpm retained by Caco-2 (298,639) to the average of the eleven nonbinding cell lines (365) was 818:1. Caco-2 cells retained approximately 18% of the total radioactive counts present in the incubation well after two-hour incubation.
Nine MA variants were assayed for adherence to Caco-2 cells after two hours' incubation. The relative binding level and amino acid composition of each MA variant is displayed in Figure 3A. Alteration of only one to three amino acids within the peptide resulted in retention differences as large as 70-fold, e.g., in variant MA2 vs. variant MA5.
To investigate how quickly 32 P isotope could be transferred from the peptide variants and incorporated into cellular proteins, the three most avidly binding MAs (see Figure 3A) were added to replicate wells containing Caco-2 cells, then washed away at varying time intervals and the cells and supernatant assayed. As shown in Figure 3B, substantial percentages of these 32 P-labeled variant decapeptides bound to cells within only a few minutes, with large amounts of radiolabeled cellular proteins appearing at two hours after exposing cells to the labeled peptides. Notably, a parallel experiment in which conditions described in Figure 3 were duplicated, but washed cells were incubated overnight in complete medium (data not shown), still revealed similar levels of 32 P-decapeptide release and retention for all nine MAs, as described for MA5 in Figure 2.
The peptide binding, washing and assay experiment described for Figure 2 was then repeated in the seven most avidly binding cell lines using MA5, except that after three washes of medium, 200 ul of complete medium was added to each well and the cells were incubated overnight at 37uC. Figure 4A shows the cpm  retained by cells or released into the medium after this overnight incubation. Approximately 88% of MA5 radioactive counts initially retained by the colon cancer cell lines was released into the medium, while approximately 12% of initially retained radioactive counts were retained by cells. The two esophageal cell lines that originally retained large amounts of radioactive counts, QH-TRT and SEG1, retained 39% and 37% of their original counts, respectively, after overnight incubation. Caco-2 cells retained the greatest number of counts, averaging 58,305 cpm in triplicate wells containing 10,000 cells each. This figure represents approximately 5.8 cpm, or 348 counts per hour, per cell (i.e., when extrapolated over a potential 2-week exposure period, equivalent to over 87,000 counts per cell). Figure 4B shows the time course of the release of MA5 from the Caco-2 adenocarcinoma cell line over a 48-hour time period. The majority of the total counts released over the 48 hour time period are released by nine hours of incubation. Figures 4C and  4D consist of two autoradiograms showing the locations of the radioactive molecules described in Figure 4B on polyacrylamide-SDS gels. The sizes of the cellular radioactive proteins in the cells are shown in Figure 4C; 32 P-labeled MA5 released into the medium is shown in Figure 4D. There is apparent agreement on the distribution and overall radioactivity levels in comparing Figure 4B and Figures 4C and 4D. As soon as two hours after the introduction of the radioactive peptide, a substantial portion of the isotope appears to have been transferred to higher molecular weight proteins.

DISCUSSION
This report describes the discovery of decapeptides that can be labeled with a high energy (1.7 Mev) beta emitter ( 32 P) and can bind avidly to several different adenocarcinoma cell lines, efficiently delivering this potential tumor-ablating material to the cells. The decapeptides, termed MA for Modified Adjuvant, are protein kinase substrates. Previously, it had never been shown or suspected that this substrate, when labeled with a tumor-ablating material such as 32 P, could bind to and transfer the radioisotope to a cell line after one to two hours of incubation. Moreover, we have shown for the first time that transfer of isotope from these decapeptides is restricted to cell types derived from primary colon and esophageal adenocarcinomas. For example, exposure of certain colon cancer cell lines (e.g. Caco-2) to the most avidly binding labeled peptide, MA5, for a two-hour period resulted in the transfer of a radioactive dose of over 29 counts per minute per cell after a two hour incubation, wash, and immediate determination of the retained radioactivity.
The incubation of 32 P-labeled decapeptide with certain cell lines resulted in large amounts of peptide being retained after a twohour incubation, but a substantial proportion of this bound peptide was released after an overnight incubation. For example, after incubation of the labeled variant MA5 with Caco2 cells for two hours, three wash steps, and overnight incubation in medium, 88% of the originally retained 32 P isotope was released. However, the 12% that was retained by cells still represented 5.8 cpm per cell, extrapolating to over 8,300 counts per cell per day. In addition, radioactivity that was still retained by cells after overnight medium incubation was permanently incorporated into a variety of cellular proteins, as demonstrated by polyacrylamide gel electrophoresis of post-exposure cellular lysates Among 18 cell lines assayed for their ability to bind the decapeptides, seven demonstrated very high retention of isotope after two-hour incubation. Although all seven of these lines released from 63% to 88% of this radioactivity after an overnight incubation, the amount of isotope that was retained overnight was still substantial. Of these seven cell lines, five were derived from colorectal adenocarcinomas, one from an esophageal adenocarcinoma, and one from a Barrett's metaplasia specimen. The 11 cell lines that did not bind the radioactively labeled decapeptide MA5 were derived from a variety of tissue origins. These included squamous cell carcinomas of the cervix, lung, breast, and a fibrosarcoma, as well as normal kidney, colon, and esophageal tissues.
The majority of approved immunotherapeutic regimens for cancer involve an antibody directed against a specific cellular molecule [16]. These agents can function by binding to the cell surface and may utilize ADCC, complement activation, or cellular apoptosis. The antibodies may also be coupled to a tumor-ablating agent, such as toxins or radioisotopes [17][18][19][20][21]. The addition of isotope to peptides, and their use for both diagnostic and therapeutic purposes, is an active area of biomedical research [22][23][24][25]. Our work utilizes protein kinase A substrates labeled with 32 P isotope. A high-energy beta-emitting radioisotope results in an electron pathlength range of up to 5 mm, permitting substantial penetration of solid tumors. Due to a predicted ''bystander'' effect, one beta particle will penetrate hundreds or thousands of cells within the tumor, even those not directly binding the decapeptide. Moreover, since the molecular weights of these minuscule decapeptides proteins are far lower than the exclusionary molecular weight limit of the filtering kidneys, these peptides should be rapidly eliminated in the urine, leading to reduced systemic toxicity. Thus, it should be feasible for both a radioactive dose and unbound radioactivity to be eliminated easily and in a relatively short period of time. We anticipate that additional known enzyme substrates may eventually be identified as potential vehicles for the specific delivery of anti-tumor agents to cancer cells and that potential cancer therapeutic regimens employing this peptide or other similar substances might be the newest strategy for peptide binding therapy.   Figure 4B were run on a 16.5% polyacrylamide-SDS gel and exposed to film. Immediately after washing (i.e., at 0 hours), the majority of the counts were visualized as 32 P-peptide. Over the next 48 hours, the peptide counts greatly diminished, with the majority of bound radioactivity incorporated into cellular proteins. (D) Aliquots of medium containing the released (dotted line) 32 P-peptide MA5 were assayed at time intervals after washing, as described in Figure 4B. As time progressed, more of the 32 Ppeptide was released, reaching a plateau by 24 hours after washing. doi:10.1371/journal.pone.0000964.g004 taining these clones were grown overnight in LB (containing 100 mg/ml ampicillin), diluted 1/10 in LB-Amp and grown at 37uC for two hours. IPTG was added to 1 mM and the culture grown at 37uC for five hours. Ten ml of each culture were centrifuged and the cell pellet resuspended in 1 X TBS containing 100 mg/ml lysozyme. After two cycles of freeze-thaw, the lysate was centrifuged and the supernatant was mixed with 100 ml of Sepharose-Glutathione for two hours at RT. Each pellet was washed three times with 1 X TBS, and the bound recombinant fusion proteins were labeled with 32 P using protein kinase A and 32 P-c-ATP according to the manufacturer's instructions (Sigma, St. Louis, Mo.). The pellet was washed four times with 1 X PBS and the labeled decapeptide was cleaved and released into the supernatant with thrombin (GE Healthcare).

MATERIALS AND METHODS
Assay of the binding of 32 P-labeled decapeptides to cell lines: Cell lines were grown in complete medium containing 10% bovine fetal serum (heat inactivated). In each well of a 96-well plate, 10,000 cells from various cell lines were grown overnight in complete medium. Ten ml of the labeled-peptide in 1 X PBS and 90 ml of complete medium were added to each well and incubated at 37uC at various times of up to two hours. The peptide-medium was removed and one ml added to 100 ul gel loading buffer and counted by scintillation counting for the probe control or run on a polyacrylamide-SDS gel (Biorad).The adherent cells were briefly and gently washed with complete medium three times and some wells were assayed immediately by adding 100 ml of gel loading buffer to each well and run on a gel or counted in a scintillation counter. Other wells had 100 ml complete medium added and incubated for a further time period. Samples were either counted in a liquid scintillation counter or run on polyacrylamide-SDS gels, exposed to x-ray film, and the film developed.