Lcd Phage Display

Vol. 98 No. 6 December 2004

ORAL AND MAXILLOFACIAL PATHOLOGY

Editor: Alan R. Gould

Application of laser capture microdissection to phage display peptide library screening Hai Lu, MD,a Dadi Jin, MD,b and Yvonne L. Kapila, DDS, PhD,a,c San Francisco, Calif, Guangzhou, China, and Ann Arbor, Mich UNIVERSITY OF CALIFORNIA SAN FRANCISCO, NANFANG HOSPITAL, AND UNIVERSITY OF MICHIGAN

Objective. When identifying important regulatory genes using methods such as phage display peptide library screening it is critical to select such peptides from cells and tissues in their native state. Here, we report a novel approach to screen tumors using phage display and laser capture microdissection (LCM). Study design. A phage peptide library was screened directly on fresh oral tumor tissue, such that specifically bound peptides were selected from fresh tumor cells in the native tissue state. Tissue processing conditions were modified to ensure the survival of the bacteriophage. Results. Our results demonstrate that live phage-peptide conjugates can be recovered from laser capture microdissected cells in a form suitable for additional cycles of amplification. Conclusion. Thus, LCM will be a valuable adjunct to phage display studies. (Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2004;98:692-7)

Peptides that recognize specific cell types promise to be valuable tools in research and clinical applications. In preclinical animal studies, anticancer drugs conjugated to tumor-targeting agents generally display greater antitumor efficacy than the respective free drugs. In clinical studies, however, the use of monoclonal antibodies or peptides in the targeting of antitumor agents has proven less successful.1 One potential reason for this clinical failure may be that current phage display peptide library screening is usually performed against purified cellsurface markers, whole cells in tissue culture, or human tissues xenografted into animals, instead of directly on human tissues.2 The cellular environment and extracelThis work was supported by the UCSF Comprehensive Cancer Center/ Stewart Trust Research Award Program, ST-04-03 (to Y.L.K.). a Postdoctoral Fellow, Department of Stomatology, School of Dentistry, University of California San Francisco, San Francisco, Calif. b Professor, Nanfang Hospital, Guangzhou, China. c Associate Professor, University of Michigan, School of Dentistry, Department of Periodontics/Prevention/Geriatrics, Ann Arbor, Mich; Assistant Professor, Department of Stomatology, School of Dentistry, University of California San Francisco, San Francisco, Calif. Received for publication Aug 9, 2004; returned for revision Aug 9, 2004; accepted for publication Sep 10, 2004. 1079-2104/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.tripleo.2004.09.004

692

lular matrix (ECM) play important roles in gene expression. Therefore, when methods such as phage display peptide library screening are used to identify important proteins, it is critical to select such peptides from cells in their native state. In addition, selecting peptides that bind specifically to cell-surface proteins responsible for tumor invasion is important for targeted therapy. Comparing cell-surface proteins in the center of the tumor with those in the invasive rim may provide insights into protein expression patterns that are important for this process.3 However, it has been recognized that regardless of tumor type, bulk tissue, even in its highest purity, still has potential contamination from stromal, endothelial, and inflammatory cells.4 Thus, the problem of cellular heterogeneity has been a significant barrier to the molecular analysis of normal and tumor tissue. Here we report a novel approach for screening tumors in their native environment. We used a combination of laser capture microdissection (LCM) and random phage peptide library screening (Fig 1). This novel method enabled us to directly screen a phage peptide library on fresh oral tumor tissue and recover specific cell-binding peptides from tumor cells harvested by LCM. LCM is a rapid and reliable method for procuring pure populations of targeted cells or cell groups from a section of complex, heterogeneous tissue.5,6 LCM is

OOOOE Volume 98, Number 6

Lu, Jin, and Kapila 693

Fig 1. Panning with a phage display peptide library. Phage display describes a selection technique in which a peptide is expressed as a fusion with a coat protein of a bacteriophage, resulting in display of the fused peptide on the exterior surface of the phage virion, while the DNA encoding the fusion resides within the virion. Phage display has been used to create a physical linkage between a vast library of random peptide sequences to the DNA encoding each sequence, allowing rapid idenfication of peptide ligands for a variety of target molecules by an in vitro selection process called ‘‘panning.’’ Panning is carried out by incubating a library of phage-displayed peptides with the target, washing away the unbound phage, and eluting the specifically bound phage. The eluted phage is then amplified and taken through additional cycles of panning and amplification to enrich the pool of phage in favor of the tightest binding sequences. After 3 to 4 rounds, individual clones are characterized by DNA sequencing and ELISA.

based on the adherence of visually selected cells to a transparent thermoplastic film (ethylene vinyl acetate polymer), which prevents the tissue sections from being directly shot by the low-energy infrared laser pulse and protects the phages from the burning laser beam.

MATERIALS AND METHODS Random phage peptide library PhD-C7C phage peptide library CX7C (C = cysteine; X = any amino acid) and empty M13KE vector were purchased from New England Biolabs, Inc (Beverly,

694 Lu, Jin, and Kapila Mass). The titer of the library was ;1012 plaqueforming units/mL (pfu/mL). Purification and amplification of phage particles was performed according to the manufacturer’s protocol. Tissue processing and sectioning Fresh human oral squamous cell carcinoma (SCC: HSC-3 cell) tissue was resected from host nude mice.7 After resection, the tissues were cut into 6 3 6 3 6 mm3 blocks, immediately incubated in a 1-mL 2 3 1011 pfu/ mL phage-peptide library, specific phage-peptides, or M13KE control solution (Dulbecco Modified Eagle’s Minimal Essential Medium containing 200:1 diluted Protease Inhibitor Cocktail [Sigma, St Louis, Mo] and 1% bovine serum albumin), rocked gently for 30 minutes and 1, 2, 5, 7, and 10 hours at 4 8C. The tissues were then washed in phosphate-buffered saline (PBS) 3 times, embedded in Optimal Cutting Temperature (OCT) compound (Sakura Finetek, Torrance, Calif), and frozen in liquid nitrogen. Frozen sections (10 mm thick) of both tumor and normal mucosal tissue were cut on a cryostat (Leica, Milton Keynes, Buckinghamshire, UK), mounted onto clean uncoated glass slides, and stored at ÿ808C until further processing. Institutional Review Board approval was obtained for these studies from the University of California San Francisco Committee on Human Research. Tissue staining and dehydration by freeze drying Frozen tissue sections were thawed at room temperature for 30 seconds, immersed in 70% ethanol for 30 seconds, and stained with hematoxylin and eosin.3,8 Briefly, sections were dipped 3 times in deionized water, incubated in Mayer’s hematoxylin (Sigma) for 1 minute at room temperature, dipped 5 times in TBST (tris (hydroxymethyl) aminomethane-buffered saline with 0.1% Tween-20), incubated in an ammonium hydroxide (Fisher Scientific, Santa Clara, Calif) solution (0.18% in PBS, pH 7.0) for 30 seconds, dipped 5 times in TBST, and incubated in 0.5% aqueous eosin (Sigma) for 30 seconds at room temperature. The sections were rinsed another 5 times with ice-cold TBST to remove nonspecifically or weakly bound phages. After staining, the sections were incubated in a cryoprotectant (PBS with 9% sucrose) for 2 minutes at room temperature. The sections were dried on dry ice for 2 hours under high pressure (200 mtorr) using a mobile freeze-dryer (Virtis Company Inc, Gardiner, NY), then warmed to room temperature. The tissue sections were also dried for 1, 2, 3, 5, or 10 days to determine the effect of drying time on the frozen samples and on live phage recovery. LASER CAPTURE MICRODISSECTION After freeze drying, LCM was performed using CapSur transfer film with the PixCell II LCM system

OOOOE December 2004

(Arcturus Engineering, Inc, Mountain View, Calif). Cells of interest were identified and harvested from the SCC tumor islands and invasive tumor rims. About 600 SCC cells (400 LCM pulses using a 30-mm beam diameter and 1 ms per pulse) were captured from each of the tissue samples. As a negative control and to measure the amount of nonspecific phage contamination when transfer film is overlayed on the tissue sections, another CapSure transfer film was overlayed on sections for the same amount of time, then retrieved but without LCM. Phage recovery from microdissected tissues After microdissection, the cells attached to the CapSure transfer films were placed in 500 mL Eppendorf microcentrifuge tubes (Applied Biosystems, Foster City, Calif) with 100 mL of the host bacterial solution (OD600 = 0.5), vortexed, and incubated at 37 8C for 20 minutes. The pfu number or amount of phage recovered for each sample was determined by serially diluting the sample and subsequently plating these dilutions on Luria-Bertani/isopropylthio-beD-galactoside/5-bromo4-chloro-3-indolyl-beD-galactoside agar plates. Immunohistochemistry For immunohistochemical detection of binding with phage-peptides or M13KE control phages, frozen sections were fixed in 70% ethanol for 30 seconds, then phage proteins were detected in the tissues by immunostaining as described.7 Briefly, nonspecific binding was blocked with a 2% nonfat milk solution in PBS for 15 minutes at room temperature, then the sections were incubated in a 1:500 dilution of a horseradish peroxidase (HRP)-conjugated mouse anti-M13 monoclonal antibody (Amersham Pharmacia Biotech Inc, Piscataway, NJ) for 30 minutes. Staining controls were evaluated with a nonimmune mouse immunoglobulin-G (IgG) antibody. The substrate solution used was 1 mg/mL of diaminobenzidine (DAB) in PBS with 0.1% hydrogen peroxide. Then, the sections were counterstained with 0.5% hematoxylin. RESULTS AND DISCUSSION Fresh human oral SCC tissue and normal mucosal (epithelium, blood vessels, and adjacent connective tissue) tissue sections were incubated and processed in the phage peptide library solution. The sections were rinsed to remove nonspecifically or weakly bound phages and dehydrated by freeze-drying (20 mtorr) for up to 10 days. The phage-bound cells of interest were then selected by LCM, and the specific binding phages were recovered by transfection into host bacteria. The phages eluted from LCM-harvested cells were amplified and used as the input phage sublibrary for the next round of biopanning. This procedure was repeated twice.

OOOOE Volume 98, Number 6

Lu, Jin, and Kapila 695

Fig 2. Selective binding of phage to tumor after 3 rounds of selection. After each round, the selected phages were amplified and used to assess the specificity of phage binding to SCC tumor tissue. After nonspecifically or weakly bound phages were washed away, the whole sections were scraped from the slides, and phages were recovered by transfection into host bacteria. The tumor/ mucosa ratio (number of phages recovered from tumor divided by number of phages recovered from the same amount of mucosa) was about 12:1 after 3 rounds of biopanning. The number of phages (transducing units, TU, y axis) recovered from each section varied to some extent, but the tumor/mucosa was consistent across experiments. The bars show standard error of the mean (SEM) from plating in triplicate.

Fig 3. Graphs illustrate the effects of ethanol (A) and freeze-drying (B) on phage survival. A, The sections, incubated with phages, were fixed in 70% ethanol. The toxic effect of ethanol on phages was time dependent; that is, the longer the exposure to ethanol the more toxic to the phages. For performing LCM, 30 seconds of tissue fixation with ethanol was sufficient and only 29% of phages were killed with this protocol. B, The sections, incubated with phages, were dried for 1, 2, 3, 5, and 10 days. The results demonstrate that freeze-drying did not appreciably alter phage survival for up to 10 days. Approximately, 75% to 85% of phages survived with this protocol. Each value represents the mean of 3 experiments performed in triplicate.

After the second and third rounds of selection, tumor and mucosa tissue sections were washed to remove nonspecifically bound phages, and the binding specificity of phages selected from the previous round was assessed. Analysis of the phage numbers recovered from each whole section showed the phage solutions recovered from the second and third rounds had 8-fold and 12-fold greater affinity, respectively, for SCC tumor tissue than for normal mucosa (Fig 2).

One of the most important steps for successful LCM is the tissue dehydration step. However, we found that prolonged and repeated exposure to ethanol and xylene killed almost all the phages in our tissue samples. Therefore, we tested and used a freeze-drying protocol. Freeze drying has been used to preserve proteins and microorganisms for extended periods.9 Freeze drying did not appreciably alter phage survival for up to 10 days (Fig 3, A). In addition, longer drying times facilitated the

696 Lu, Jin, and Kapila

OOOOE December 2004

Fig 4. Use of LCM to selectively harvest an SCC tumor island (A), the SCC tumor rim (B), and a control lymph node (C). Image magnification 10 3 10. The arrows point to tumor cells that will be targeted for LCM retrieval (Pre-LCM). Tissues attached to the LCM transfer films were removed from the heterogeneous tissues (Post-LCM). The harvested cells were transferred to the transfer film as illustrated (Cap). These 7-mm frozen sections were counterstained with H/E before LCM retrieval to illustrate the detailed histomorphology of the SCC tissue. After H/E staining, the sections were washed extensively in cold TBST, incubated in a sucrose cryoprotectant, then freeze dried for 48 hours under high pressure (20 mtorr). This processing technique likely altered the acid/base characteristics of the stain and also likely washed some of the H/E stain away. LCM was then performed with CapSure transfer film and the PixCell II LCM system (Arcturus Engineering, Mountain View, Calif). About 600 SCC cells (400 1-ms LCM pulses with a 30-mm beam diameter) were captured from each of the tissue samples. Phages recovered from the tumor cells were amplified and reincubated with fresh tissues in 2 consecutive rounds.

dissection and retrieval of cells and tissues by LCM. Furthermore, freeze drying optimally maintains the 3dimensional structure of proteins, thus also allowing for improved protein research via LCM in future studies. Another step that may be toxic to bacteriophages is fixation in 70% ethanol. However, we found that more than 70% of the phages survived a 30-second incubation in 70% ethanol at 48C (Fig 3, B). To avoid additional exposure of the tissue to ethanol, we used an aqueous eosin solution for staining rather than an ethanol solution. These measures allowed successful LCM and retrieval of live phages. The major goal of the present study was to develop a phage peptide library biopanning method suitable for use directly on fresh tissue. To obtain fresh tissue easily and to set up strict parallel controls, we used human tumor tissue xenografted in nude mice rather than fresh

human tumor tissue. However, the findings in this study demonstrate that our method is suitable for screening of a phage peptide library directly on fresh human biopsy samples and rescuing peptide sequences from tumor cells in their native environment or from different cell groups within the same organ (Fig 4). Previous authors10,11 reported that after injection of a phage peptide library solution intravenously in mice, the phage peptide spread outside the blood vessels and into the tumor tissues. In the present study, fresh tissue blocks were incubated in the phage peptide library solution with gentle rocking for 7 hours at 48C. After the frozen sections were washed with TBST 10 times, a strong phage protein signal was still identified. Immunohistochemical analysis showed that the phage protein could penetrate into tumor tissue from both the edge of the tumor and the blood vessels inside the tumor

OOOOE Volume 98, Number 6

Lu, Jin, and Kapila 697

Fig 5. Immunohistochemical detection of phage that has penetrated fresh tumor tissue. Fresh SCC tissue (6 3 6 3 6 mm) was incubated in a 2 3 1011 TU phage-peptide library solution at 48C for 7 hours. After incubation, frozen sections were prepared and stained immunohistochemically with horseradish peroxidaseeconjugated mouse anti-M13 monoclonal antibody (1:500, Amersham Pharmacia, Piscataway, NJ) to identify phage proteins and phage route of entry into tissues. No biopanning was done and sections illustrate the route of entry into the tissues by the phage. Brown staining represents the phage protein signal; blue indicates counterstaining of cell nuclei with hematoxylin. A, Penetration of phage from the edge of the tissue. B, Penetration of phage into the tissue from a blood vessel (arrow). Image magnification 10 3 20.

tissue (Fig 5, A, B). Our findings show that phages selected by the novel method described here penetrate tumor tissues and bind to tumor cells. Thus, peptides selected in this fashion may be clinically useful for delivery of intravenously administered drugs to specifically targeted tumor cells. Although the random peptides selected may not directly identify the target protein, the peptides can be sequenced and this information can be used to scan protein or gene sequence banks and/or used to design probes to scan gene libraries for identification of the target protein or receptor. Once identified in this manner, the target protein or receptor can be further tested for specificity on tumor tissues. In summary, this short communication on proof of concept illustrates that application of laser capture microdissection to phage library screening is possible, however further studies are needed to identify the specific peptides in question. The authors thank Dr Randall Kramer for the SCC tissues, Dr Jeffery S. Cox for the use of his freeze-drying facility, Dr Richard Jordan for critically reading this manuscript, and Stephen Ordway for editorial assistance. REFERENCES 1. Chari RV. Targeted delivery of chemotherapeutics: tumoractivated prodrug therapy. Adv Drug Deliv Rev 1998;31:89-104. 2. Brown KC. New approaches for cell-specific targeting: identification of cell-selective peptides from combinatorial libraries. Curr Opin Chem Biol 2000;4:16-21. 3. Dolter KE, Braman JC. Small-sample total RNA purification: laser capture microdissection and cultured cell applications. Biotechniques 2001;30:1358-61.

4. Rubin MA. Use of laser capture microdissection, cDNA microarrays, and tissue microarrays in advancing our understanding of prostate cancer. J Pathol 2001;195:80-6. 5. Emmert-Buck MR, Bonner RF, Smith PD, Chuaqui RF, Zhuang Z, Goldstein SR, et al. Laser capture microdissection. Science 1996;274:998-1001. 6. Bonner RF, Emmert-Buck M, Cole K, Pohida T, Chuaqui R, Goldstein S, et al. Laser capture microdissection: molecular analysis of tissue. Science 1997;278:1481-3. 7. Ramos DM, Chen BL, Boylen K, Stern M, Kramer RH, Sheppard D, et al. Stromal fibroblasts influence oral squamouscell carcinoma cell interactions with tenascin-C. Int J Cancer 1997;72:369-76. 8. Simone NL, Bonner RF, Gillespie JW, Emmert-Buck MR, Liotta LA. Laser-capture microdissection: opening the microscopic frontier to molecular analysis. Trends Genet 1998;14: 272-6. 9. Gu MB, Choi SH, Kim SW. Some observations in freeze-drying of recombinant bioluminescent Escherichia coli for toxicity monitoring. J Biotechnol 2001;88:95-105. 10. Arap W, Pasqualini R, Ruoslahti E. Cancer treatment by targeted drug delivery to tumor vasculature in a mouse model. Science 1998;279:377-80. 11. Pasqualini R, Ruoslahti E. Organ targeting in vivo using phage display peptide libraries. Nature 1996;380:364-6.

Reprint requests: Yvonne L. Kapila, DDS, PhD University of Michigan School of Dentistry Department of Periodontics/Prevention/Geriatrics 1011 N. University Ave., Room 5213 Ann Arbor, MI 48109-1078 [email protected]