Fluorescence-expressing viruses allow rapid identification and separation of rare tumor cells in spiked samples of human whole blood

Fluorescence-expressing viruses allow rapid identification and separation of rare tumor cells in spiked samples of human whole blood Sandra M. B. Fong, Melissa K. Lee, Prasad S. Adusumilli, MD, and Kaitlyn J. Kelly, MD, New York, NY

Background. Finding and isolating rare tumor cells in blood allows for diagnosis of disseminated cancer and for molecular profiling to direct the choice of biologic therapy. We explored whether the candidate gene therapy virus NV1066---designed to specifically infect cancer cells and express green fluorescence protein (GFP)---can be used for rapid infection, identification, and isolation of rare circulating tumor cells (CTC) in human whole blood. Methods. Mixtures of human cancer cell lines and human whole blood were exposed to NV1066 or heatinactivated virus, incubated, and then examined for GFP expression by fluorescence microscopy and flow cytometry. Fluorescence-assisted cell sorting (FACS) was used to determine the efficiency of virally assisted tumor cell isolation. Sorted cells were subsequently stained for carcinoembryonic antigen (CEA) to determine if cells isolated in this way would maintain sufficient cellular integrity for molecular characterization. Results. In our study, there was 100% specificity for detection of cancer cells. Detection was consistent even at the highest dilution tested (10 cancer cells in 10 ml whole blood). The processing involved simple incubation without the technical demands of immunohistochemistry. FACS allowed for rapid isolation of GFP-expressing cells. Cells isolated by this method can subsequently undergo molecular characterization. Conclusion. Oncolytic herpes simplex virus mediated green fluorescence in combination with FACS is a novel technique for the identification and isolation of cancer cells in an experimental model of bloodborne metastases. This procedure is a promising method for improving our diagnosis, staging, and molecular profiling of cancer. (Surgery 2009;146:498-505.) From the Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY

THERE IS A CLINICAL NEED FOR INCREASING OUR ABILITY to detect and isolate rare cancer cells in blood. There have long been attempts to identify circulating tumor cells (CTC) during resection of primary cancers even in the absence of clear-cut metastases.1 It has been shown that detection of CTC in patients with breast cancer before initiation of therapy is highly predictive of both progression-free and overall survival.2 In other cancer types, including gastric, colorectal, pancreaticobiliary and lung, Presented at the 3rd Annual Academic Surgical Congress, Huntington Beach, CA, February 13--15, 2008. Supported in part by research grants from the National Institutes of Health and the Flight Attendant Medical Research Institute (FAMRI). Accepted for publication December 5, 2008. Reprint requests: Kaitlyn J. Kelly, MD, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10065. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2009 Mosby, Inc. All rights reserved. doi:10.1016/j.surg.2008.12.007

498 SURGERY

the significance of CTC or markers in peripheral blood is less clear, and so this information is not yet taken into account when making therapeutic decisions. The uncertainty of marker significance in these diseases, however, is often due to limitations in the methods of detection. Current techniques under investigation for detection of CTC include reverse transcriptasepolymerase chain reaction (RT-PCR), flow cytometry, fluorescence in situ hybridization, and, more recently, microfluidics. These techniques all function to identify and isolate CTC based on the interactions of specific antibodies targeted against known cancer antigens.3-8 RT-PCR for tumor markers is limited because it is difficult to determine if positive findings indicate the presence of viable metastatic CTC or only nucleic acids or cellular fragments originating from the primary tumor. Another limitation of antibody-based techniques is that they cannot be used for detection of all cancers, but only those cancers that express the most common and well-characterized markers. There is a need for a more sensitive, specific, and widely

Surgery Volume 146, Number 3

Fig 1. Structure of NV1066 virus. The general genetic structure of wild-type herpes simplex virus type 1 (wtHSV-1) is shown at the top. The genome of the virus is divided into the unique long (UL) and unique short (US) sequences, which are flanked by the terminal repeat long (TRL), internal repeat long (IRL), internal repeat short (IRS), and terminal repeat short (TRS) sequences. In the NV1066 virus, the ICP0, g134.5, and ICP4 genes are within the IRL and IRS sequences; these genes are deleted in favor of the cytomegalovirusenhanced green fluorescent protein (CMV-eGFP) marker gene.

applicable technology for detection of rare CTC in blood. Viruses have a natural propensity for the infection and killing of tumors.9 For more than a century, there have been reports of cases of cancer that disappeared after natural viral infections or after treatment with viral vaccines.10 In the 1960s and 1970s, multiple attempts were made at using natural viruses, including West Nile virus and measles virus, as cancer treatments.11-13 During the last 2 decades, genetic engineering has allowed construction of viruses that are even more specific for cancer. A number of viruses that infect, replicate within, and kill only cancer cells while sparing noncancerous cells are now in human clinical trials. One of the most promising classes of these viruses is herpes simplex virus (HSV). This virus, which typically only causes ‘‘cold sores,’’ has been modified to have an improved safety profile, and is currently being tested in humans for killing of brain14 and liver cancers.15 NV1020 is one such genetically engineered, cancer-specific HSV currently in clinical trials for treatment of liver metastases from colon cancer. This virus has been further modified by insertion of the gene for enhanced green fluorescent protein (GFP) into the genome, creating NV1066. We sought to determine whether the inherent tumor specificity of this virus could be harnessed to detect cancer cells using green fluorescence. In addition to the detection of rare cancer cells in peripheral blood, isolation of those cells could be important in directing the choice of therapy. The increasing understanding of the molecular

Fong et al 499

basis of cancer has led to the development of new therapeutic strategies and molecular assays that predict the likelihood that a particular therapy might work.16-18 Having the capability to rapidly and specifically isolate a cancer cell for molecular characterization helps in the decision to use additional systemic therapy and may direct the choice of therapeutic agent.19 Here, we present the first evidence that HSV--mediated detection and isolation of cancer cells in human whole blood may fulfill these potentially important roles. MATERIALS AND METHODS NV1066. NV1066 is a replication-competent, attenuated HSV type 1 based on the NV1020 virus that is currently in human clinical trials as a treatment for cancer.20 This virus has been engineered to be specific for cancer by deletion of 1 copy of each of the viral genes ICP-4, ICP-0, and g134.5. The proteins encoded by these genes are important for virus replication in noncancerous cells, and their deletion renders the virus selective for infection and replication in tumor cells. A gene encoding the marker gene GFP under the control of a constitutive cytomegalovirus promoter was inserted into the internal repeat sequence of the parent construct (Fig 1), resulting in a virus that expresses GFP at high levels within hours of infection of a cell. Virus production. Virus was grown on African green monkey kidney cells (Vero cells; American Type Culture Collection [ATCC], Rockville, MD) by plating at an initial multiplicity of infection of 0.02 at 34oC. After 2 days, cells were subjected to freeze-thaw lysis and sonication. Lysates were clarified by centrifugation (300 g for 10 minutes at 4oC) and viral supernatants stored at --80°C until use. Cell lines. The human gastric adenocarcinoma cell line OCUM-2MD3, human colorectal cancer cell line LS174T, and human breast carcinoma cell line MCF7 were used in these studies. All cell lines were obtained from the ATCC. Cells were maintained in appropriate media and were incubated in a humidified incubator supplied with 5% CO2 at 37°C. Human whole blood. Human whole blood beyond the expiration date for use was obtained from the blood bank. Consultation with the Institutional Review Board determined that the blood is considered medical waste and exempt from the board’s official approval. Fluorescence microscopy. NV1066 infection and GFP expression in blood samples spiked with cancer cells were assessed by fluorescent microscopy

500 Fong et al

in vitro to determine selectivity of viral infection malignant cells in the presence of human whole blood. To enable microscopic examination of cells, human whole blood was diluted 100-fold in Roswell Park Memorial Institute (RPMI) media. Equivalent aliquots of diluted blood were spiked with human cancer cells (OCUM-2MD3, LS174T, and MCF7) at various concentrations ranging from 10 to 1,000,000 cancer cells per aliquot in 6-well plates. Spiked samples were infected with 1 3 106 plaque-forming units (pfu) of NV1066 in 50 ml phosphate-buffered saline (PBS) and incubated for 4--18 hours. Diluted blood aliquots without cancer cells and spiked aliquots infected with PBS only (empty vector) served as controls. After incubation, cells were washed with PBS and cytospun at 1000 rpm 3 4 minutes at room temperature onto silane-coated slides (Electron Microscopy Sciences, Hatfield, PA). For detection of green fluorescence, slides were examined with an inverted stand microscope (Axiovert 200M; Carl Zeiss, Oberkochen, Germany) using a 100 W mercury arc lamp as a light source, and a charge-coupled device digital camera (Retiga EX; Qimaging, Burnaby, Canada) was used to capture images. Selective excitation of GFP was produced through a Chroma 41017 filter set (Chroma Technology Corp., Rockingham, VT). Images were processed and analyzed with the MetaMorph Imaging System (Universal Imaging, Downingtown, PA). Samples were examined with bright field and fluorescence microscopy. To confirm specificity of viral infection to cancer cells, a fluorescent-labeled antibody to carcinoembryonic antigen (CEA) (PE-labeled anti-CD66; BD Pharmingen, Franklin Lakes, NJ) was added to samples containing OCUM-2MD3 and LS174T cells. The samples were then incubated at room temperature for 30 minutes before fluorescent microscopic examination. To evaluate interobserver reliability and ease of GFP detection, blinded observers examined 20 virally treated, cytospun slides: 10 slides from a 1:500,000 tumor/benign cell mixture and 10 controls with only benign cells. Observers interpreted the slides as positive or negative for malignancy based on the presence or absence of green fluorescence. Fluorescence-assisted cell sorting (FACS). To establish a processing method for this assay in the manner it would be performed on actual blood samples from cancer patients, expired whole blood samples were spiked with OCUM-2MD3 cells before layering on Ficoll (Histopaque-1077; Sigma-Aldrich, St. Louis, MO). Next, the samples were centrifuged (10,000 rpm) for 45 minutes.

Surgery September 2009

Nucleated cells in the cushion were then aspirated, washed with PBS, exposed to virus, and incubated for 4 hours in RPMI media before analysis of GFP expression by flow cytometry. Data acquisition analyses were performed on a FACScan flow cytometer (Becton Dickinson Biosciences, San Jose, CA). This instrument allows for the characterization of cells by size as well as by fluorescence at various wavelengths. The CellQuest software program (Becton Dickinson Immunocytometry Systems) was used for data analysis. All the flow cytometry experiments were repeated by 2 independent investigators to ensure reproducibility. Each experiment was repeated in triplicate. These experiments were then repeated with the colorectal cell line LS174T to test whether the size of the cell would affect detection (OCUM-2MD3 cells are relatively large; LS174T cells are relatively small). Limits of detection. To measure the limit of detection of tumor cells in blood, serial dilutions of cancer cells were used to spike whole blood. Accurate counts of harvested cancer cell suspensions were made using trypan blue staining and manual counting using a hemocytometer. Cancer cells (0, 10, 100, 1,000, 10,000, or 100,000) were added to test tubes containing 1 ml of whole blood (approximately 10 million leukocytes and 5 billion erythrocytes). Samples were mixed and processed by density gradient separation. Spiked blood samples were layered using a Ficoll-Hypaque layering technique and centrifuged at 10,000 rpm for 45 minutes. Next, the nucleated cell cushions were aspirated, washed, resuspended in a volume of 300 mL RPMI media, and infected with 1 3 106 pfu of NV1066. Finally, they were incubated for 4 hours in polystyrene round-bottom test tubes (BD Falcon, San Jose, CA) in a humidified incubator, supplied with 5% CO2, and kept at a temperature of 37°C. Specimens incubated without virus served as negative controls. Samples were then analyzed by flow cytometry to determine the presence of GFP expression. These experiments were repeated using an initial volume of 10 ml of whole blood (approximately 100 million leukocytes and 50 billion erythrocytes). Determination of cellular molecular profile. Cells positive (OCUM-2MD3 and LS174T) and negative (MCF7) for CEA expression were mixed with whole blood as described above. FACS was then performed after 4, 6, 8, 12, and 24 hours of incubation with virus. The specimens isolated according to green fluorescence were then counterstained by a

Surgery Volume 146, Number 3

Fong et al 501

Fig 2. Detection of the human gastric adenocarcinoma cell line OCUM-2MD3 cancer cells in human blood by fluorescent microscopy after infection with the NV1066 virus. Human blood (1 3108 benign cells) was mixed with various numbers of OCUM cancer cells. These mixtures were then incubated with the NV1066 virus for 6 hours and inspected under fluorescent microscopy. Tumor cells were easily identified in concentrations as low as the 100 in 5 3 107 benign cell mixture (1:500,000 dilution).

red fluorescent antibody directed against CEA. Slides were evaluated by an observer blinded to the experimental protocol. RESULTS NV1066 rapidly infects tumor cells. A preliminary experiment was performed to test the speed by which NV1066 infection and GFP expression occur. One million OCUM-2MD3 cells were mixed with an approximately equal number of blood cells, and the mixture was incubated at 38°C for between 4 and 12 hours. Specimens were then analyzed by fluorescent microscopy and flow cytometry for GFP expression. Fluorescence could be detected with confidence within 4 hours of infection. Fluorescent microscopy. NV1066-transduced green fluorescence was expressed exclusively in cancer cells in the background of human whole blood. No fluorescence was observed in any of the negative control samples. Figure 2 shows the appearance of various mixtures of tumor cells and benign blood cells. This Figure illustrates bright field, fluorescent, and overlay images of a diluted blood aliquot spiked with OCUM cells at various concentrations. OCUM-2MD3 cells can be readily distinguished from blood cells by their large size. It is clear that GFP expression is confined to cancer cells. Regardless of the cell line used, fluorescence microscopy showed that all specimens down to this dilution (1 in 500,000) contained malignant cells. Blinded observers accurately and reliably diagnosed malignancy based on GFP detection. At the

1:500,000 dilution, all 10 slides with cancer cells and all 10 control slides were identified correctly by the 3 blinded readers. FACS. The 3 cancer cell lines were mixed with diluted whole blood at ratios from 10 to 1,000,000 cells per 10 ml whole blood. After exposure to NV1066 for 4 hours, specimens were analyzed by fluorescent microscopy or flow cytometry to detect green fluorescence. Figure 3 illustrates a typical histogram. The intensity of the green fluorescence can be observed along the x-axis; the intensity of the orange fluorescence is shown along the y-axis. Normal cells have low green fluorescence and equal orange fluorescence (low-level autofluorescence) and are in the lower left quadrant. Cells with GFP expression are far to the right of the pink line. Specimens exposed to no virus or a heatinactivated virus were not found to contain green fluorescent cells. Similarly, specimens without cancer cells did not exhibit green fluorescence. NV1066-mediated fluorescence detection by flow cytometry was found to be 100% cancer cellspecific in whole blood, given that only specimens with cancer cells exposed to a live virus expressed green fluorescence. By combining density gradient separation with automated flow cytometry, we were able to reliably identify and isolate rare CTC spiked into whole blood with a limit of detection of 10 cancer cells in 10 ml blood (approximately 10 cancer cells in >50 billion normal cells).

502 Fong et al

Fig 3. Histogram from flow cytometry of cancer cell and blood mixture. The intensity of the green fluorescence can be seen along the x-axis and that of the orange fluorescence is along the y-axis. Normal cells with low level autofluorescence (equal green and orange) are in the bottom left quadrant. On the right, beyond the pink lines, are the few GFP--expressing cells. FL1, First fluorescence detector.

Molecular profiling of sorted cells. To determine if cells infected with NV1066 and sorted according to GFP expression maintained sufficient cellular integrity for molecular characterization, tumor cells were mixed at a ratio of 10,000 cells per 10 ml whole blood, separated by a density gradient, and incubated with NV1066 or heatinactivated NV1066. Blood without cancer cells served as additional controls. Both CEA-positive OCUM-2MD3 and CEA-negative MCF7 cells were used. Cells were then sorted by green fluorescence. Sorted OCUM-2MD3 or MCF7 cells were then stained for CEA with a red fluorescent antibody (Fig 4). Green fluorescent cells were only found in specimens with tumor cells that had been incubated with a live virus. All replicates of OCUM2MD3 cells that were incubated with a virus for a period #6 hours reliably expressed CEA as detected by red fluorescence. OCUM-2MD3 replicates incubated with NV1066 for 8 to 12 hours exhibited some red fluorescence with CEA counterstaining, but it was observed to be weak and patchy. OCUM-2MD3 replicates incubated with virus for 24 hours before sorting did not exhibit CEA expression. No CEA was observed in MCF7 cells isolated by GFP expression at any time point.

Surgery September 2009

Fig 4. Overlay of red fluorescent carcinoembryonic antigen (CEA) antibody and green fluorescence protein (GFP) expression in OCUM-2MD3 cancer cells. Cells sorted from Ficoll-separated whole blood based on GFP expression were counterstained with red fluorescent antibodies to CEA. Images were then merged to demonstrate membrane expression of CEA in GFPexpressing cells.

DISCUSSION Oncolytic viral therapy harnesses the life cycle of a virus to kill cancer. A virus arrives at and attaches to a susceptible cell and its genetic material is internalized by the cell. Next, the virus hijacks the cellular DNA and protein synthetic machinery to make many copies of itself. The cell is then destroyed by lysis, releasing many more viruses that then attack neighboring cells. During the last decade, many types of viruses have been genetically engineered to be particularly specific in the infection of, replication within, and killing of only cancer cells.21 One very promising candidate virus for human therapy is HSV, which has been shown to specifically infect many types of cancers, such as lung, bladder, head and neck, breast, esophageal, cervical, colorectal, gastric, and mesothelioma,22-27 and is now in clinical trials. We hypothesized that the specificity of this class of viruses for malignant cells can be used to diagnose and isolate rare CTC from whole blood. The data presented in these studies confirm this hypothesis and suggest that designer oncolytic HSV may also be important as diagnostic agents. Pathologists and cytologists are often asked to identify minute numbers of tumor cells in biologic specimens.28,29 The detection of CTC in the bloodstream has been a particular challenge because of technical limitations of identifying very few cancer cells in the background of many normal cells. The detection of CTC in peripheral blood could provide a powerful tool for cancer therapy by aiding

Surgery Volume 146, Number 3

Fig 5. Schematic of a virally assisted fluorescence detection assay. (A) the gold standard assay for detection of circulating tumor cells in which cells are stained and assessed according to size and shape by trained cytopathologists, who may order additional immunohistochemistry. (B) The method proposed in our study in which the blood sample and virus are mixed, incubated, and then undergo automated fluorescence-assisted cell sorting (FACS).

with initial diagnosis, determination of prognosis, detection of recurrence, and assessment of response to ongoing therapy. Isolation and characterization of CTC could potentially direct customized cancer therapy for individual patients. For these reasons, CTC detection has developed into a field of study. Cytopathology in combination with an enrichment technique, such as immunomagnetic or density gradient separation, is currently the gold standard for detection of CTC. If the cytologist or pathologist needs additional confirmation, immunohistochemistry is performed. This technique, however, entails further staining of the microscope slides with antibodies that are more specific for a particular cancer, and adds days to the process.30 Cytologic tests are labor intensive and highly dependent on the skill of the cytopathologist (Fig 5, A). The upper limit of detection of tumor cells using routine examination of microscope slides is thought to be 1 tumor cell in approximately 20,000 normal cells. These limitations have provided the impetus for the development of more sensitive and efficient methods to facilitate the detection of rare CTC. Most investigational methods also involve an enrichment technique combined with a means of cellular identification or profiling, such as immuno-mediated detection, RT-PCR for various tumor markers, or genomic analysis.3-6,31 More recently, microfluidics has been investigated as a method of CTC detection and isolation that does not require blood sample preprocessing or enrichment. Nagrath et al8 described the use of a microfluidic platform chip coated with

Fong et al 503

antibodies against epithelial cell adhesion molecule (EpCAM). They were able to detect CTC in patients with lung, prostate, breast, pancreatic, and colon cancer with high sensitivity and specificity. A potential limitation of this method is that a significant percentage of cancers do not express EpCAM.32 Furthermore, in the process of epithelial--mesenchymal transition that occurs in tumor metastasis, cells that may have expressed EpCAM in the primary tumor lose their epithelial antigens in the conversion to CTC.3,33,34 The clinical application of this assay, therefore, may be limited due to false-negative results in patients with CTC that do not express EpCAM. No antibody exists that is 100% tumor- or tissuespecific.35 This finding is the primary limitation of all current investigational techniques for the detection of CTC and is the reason that no individual assay has been shown to be universally applicable to cancer patients or acceptable for routine clinical use. We introduce a simple and rapid method that is based on the susceptibility of malignant cells to selective infection by an attenuated oncolytic HSV. Oncolytic herpes viral entry into a malignant cell, although not fully understood, is not dependent on the presence of a specific cell membrane protein. Cells become susceptible to infection by oncolytic HSV in the process of malignant transformation. We have demonstrated that NV1066 can productively infect 111 different human cancer cell lines originating from 14 different primary organs and that this infection is selective in the setting of normal cells from the organ of origin.36 The goal of this study was to determine if oncolytic HSV could selectively infect CTC in whole blood. Although our assay requires a preliminary processing step of density gradient separation, this technique is simple to perform. We chose this method because studies have shown that it is more sensitive than other indirect methods of CTC enrichment in patients with cancer.37 Our assay then involves a brief period of incubation with oncolytic HSV. Finally, automated flow cytometry is performed (Fig 5, B), which provides a relatively fast means of detecting and isolating malignant cells present in a blood sample without the need for highly trained personnel. This method of identifying CTC, in turn, could theoretically enable rapid delivery of appropriate therapy. Until recently, administration of medications for treatment of cancer has been a process of trial and error. Patients are treated with chemotherapies and efficacy is determined by radiologic assessment of tumor response. More recently, medications have been designed that specifically

504 Fong et al

target a known cancer cell characteristic.16 Assays are simultaneously being developed that predict which tumors would be susceptible to treatment. For example, investigators have designed radioactive antibodies that bind to cells expressing CEA to treat CEA-positive cancers.38 A method for purifying small amounts of tumor cells and determining cellular characteristics could therefore assist in planning treatment. In this study, we demonstrate that virally enhanced cell sorting can be performed rapidly to produce highly purified cell fractions. These purified fractions can be used to identify cellular profiles to potentially direct therapy. The virus used in these studies has been shown to infect a broad spectrum of cancers with high specificity in the background of various types of normal cells, including cells isolated from murine bronchoalveolar epithelium, lung parenchyma, esophagus, pleura, peritoneum, liver, kidney, urinary bladder, stomach, and gallbladder.22-27,36 The results of these preliminary investigations thus may be applicable to many human cancers. One limitation of this study is our use of expired samples of human whole blood that was spiked with human cancer cell lines. Fresh blood samples from patients with cancer may contain activated leukocytes or other reactive cells that could phagocytize an oncolytic virus, exhibit transgene expression, and produce false-positive results. Furthermore, this assay provides only qualitative information about the presence or absence of cancer cells in whole blood. Although we observed a quantitative trend between the number of cells spiked into our samples and the number isolated from them, these studies were not designed to investigate the use of this assay for quantitative determinations. Finally, our assay is limited by time constraints. Cells can only be incubated with virus for a limited time period (#6 hours) if they are to be used for precise molecular profiling. Expression of cellular antigens is not reliable after longer periods of exposure to virus. This factor is likely due to a change in gene expression from the host cell genes to herpes virus genes that occurs over time after initial virus infection. The experiments described here are proof of principle studies demonstrating that oncolytic HSV can identify human cancer cells spiked into samples of whole blood. We plan to expand upon these preliminary studies by investigating this technology in clinical trials designed to detect rare metastatic cells in the blood and peritoneal fluid of patients with cancer. Furthermore, we hope to explore the question of whether this assay could provide quantitative information about tumor burden.

Surgery September 2009

Ultimately, we plan to determine if the detection of CTC by this assay provides prognostic information for patients with cancer. In addition to showing promise as cancer therapy, these viruses may also be important for cancer diagnosis. We would like to thank Meryl Greenblatt for her outstanding editorial assistance in this work.

REFERENCES 1. Viadana E, Bross IDJ, Pickren JW. Cascade spread of bloodborne metastases in solid and non-solid cancers of humans. In: Weiss L, Gilbert HA, editors. Pulmonary Metastasis. GK Hall: Boston; 1978. p. 142. 2. Elshimali YI, Grody WW. The clinical significance of circulating tumor cells in the peripheral blood. Diagn Mol Pathol 2006;15:187-94. 3. Paterlini-Brechot P, Benali NL. Circulating tumor cells (CTC) detection: clinical impact and future directions. Cancer Lett 2007;253:180-204. 4. Jacob K, Sollier C, Jabado N. Circulating tumor cells: detection, molecular profiling and future prospects. Expert Rev Proteomics 2007;4:741-56. 5. Vona G, Sabile A, Louha M, Sitruk V, Romana S, Schutze K, et al. Isolation by size of epithelial tumor cells: a new method for the immunomorphological and molecular characterization of circulatingtumor cells. Am J Pathol 2000;156:57-63. 6. Smirnov DA, Zweitzig DR, Foulk BW, Miller MC, Doyle GV, Pienta KJ, et al. Global gene expression profiling of circulating tumor cells. Cancer Res 2005;65:4993-7. 7. Mao X, Shaw G, James SY, Purkis P, Kudahetti SC, Tsigani T, et al. Detection of TMPRSS2:ERG fusion gene in circulating prostate cancer cells. Asian J Androl 2008;10:467-73. 8. Nagrath S, Sequist LV, Maheswaran S, Bell DW, Irimia D, Ulkus L, et al. Isolation of rare circulating tumour cells in cancer patients by microchip technology. Nature 2007; 450:1235-9. 9. Hawkins LK, Lemoine NR, Kirn D. Oncolytic biotherapy: a novel therapeutic platform. Lancet Oncol 2002;3:17-26. 10. Dock G. Rabies virus vaccination in a patient with cervical carcinoma. Am J Med Sci 1904;127:563. 11. Smith RR, Huebner RJ, Rowe WP, Schatten WE, Thomas LB. Studies on the use of viruses in the treatment of carcinoma of the cervix. Cancer 1956;9:1211-8. 12. Southam CM, Moore AE. Clinical studies of viruses as antineoplastic agents, with particular reference to Egypt 101 virus. Cancer 1952;5:1025-34. 13. Hansen RM, Libnoch JA. Remission of chronic lymphocytic leukemia after smallpox vaccination. Arch Intern Med 1978; 138:1137-8. 14. Markert JM, Medlock MD, Rabkin SD, Gillespie GY, Todo T, Hunter WD, et al. Conditionally replicating herpes simplex virus mutant, G207 for the treatment of malignant glioma: results of a phase I trial. Gene Ther 2000;7:867-74. 15. Kemeny N, Brown K, Covey A, Kim T, Bhargava A, Brody LA, et al. Phase I, open-label, dose-escalating study of a genetically engineered herpes simplex virus, NV1020, in subjects with metastatic colorectal carcinoma to the liver. Hum Gene Ther 2006;17:1214-24. 16. Mansour JC, Schwarz RE. Molecular mechanisms for individualized cancer care. J Am Coll Surg 2008;207:250-8. 17. Tewes M, Aktas B, Welt A, Mueller S, Hauch S, Kimmig R, et al. Molecular profiling and predictive value of circulating

Surgery Volume 146, Number 3

18.

19.

20.

21. 22.

23.

24.

25.

26.

27.

tumor cells in patients with metastatic breast cancer: an option for monitoring response to breast cancer related therapies. Breast Cancer Res Treat 2008. In press. Magne N, Chargari C, Deutsch E, Castadot P, Ghalibafian M, Bourhis J, et al. Molecular profiling of uterine cervix carcinoma: an overview with a special focus on rationally designed target-based anticancer agents. Cancer Metastasis Rev 2008;27:737-50. Wulfing P, Borchard J, Buerger H, Heidl S, Zanker KS, Kiesel L, et al. HER2-positive circulating tumor cells indicate poor clinical outcome in stage I to III breast cancer patients. Clin Cancer Res 2006;12:1715-20. Wong R, Joe J, Kim S, Shah J, Horsburg B, Fong Y. Oncolytic herpesvirus effectively treats murine squamous cell carcinoma and spreads by natural lymphatics to treat sites of lymphatic metastases. Hum Gene Ther 2002;13:1213-23. Woo Y, Adusumilli PS, Fong Y. Advances in oncolytic viral therapy. Curr Opin Investig Drugs 2006;7:549-59. Bennett JJ, Delman KA, Burt BM, Mariotti A, Malhotra S, Zager J, et al. Comparison of safety, delivery, and efficacy of two oncolytic herpes viruses (G207 and NV1020) for peritoneal cancer. Cancer Gene Ther 2002;9:935-45. Bennett JJ, Adusumilli P, Petrowsky H, Burt BM, Roberts G, Delman KA, et al. Up-regulation of GADD34 mediates the synergistic anticancer activity of mitomycin C and a gamma134.5 deleted oncolytic herpes virus (G207). FASEB J 2004;18:1001-3. Stanziale SF, Stiles BM, Bhargava A, Kerns SA, Kalakonda N, Fong Y. Oncolytic herpes simplex virus-1 mutant expressing green fluorescent protein can detect and treat peritoneal cancer. Hum Gene Ther 2004;15:609-18. Stiles B, Adusumilli P, Bhargava A, Stanziale S, Kim T, Chan M, et al. Minimally invasive localization of oncolytic herpes simplex viral therapy of micrometastatic pleural cancer. Cancer Gene Ther 2006;13:53-64. Adusumilli PS, Eisenberg DP, Chun YS, Ryu KW, Ben-Porat L, Hendershott KJ, et al. Virally directed fluorescent imaging improves diagnostic sensitivity in the detection of minimal residual disease after potentially curative cytoreductive surgery. J Gastrointest Surg 2005;9:1138-47. Adusumilli PS, Stiles BM, Chan MK, Eisenberg DP, Yu Z, Stanziale SF, et al. Real-time diagnostic imaging of tumors and metastases by use of a replication-competent herpes vector to facilitate minimally-invasive oncological surgery. FASEB J 2006;20:726-8.

Fong et al 505

28. Flehinger BJ, Melamed MR, Zaman MB, Heelan RT, Perchick WB, Martini N. Early lung cancer detection: results of the initial (prevalence) radiologic and cytologic screening in the Memorial Sloan-Kettering study. Am Rev Respir Dis 1984;130:555-60. 29. Thunnissen FB. Sputum examination for early detection of lung cancer. J Clin Pathol 2003;56:805-10. 30. Trumper L, Menges M, Daus H, Kohler D, Reinhard JO, Sackmann M, et al. Low sensitivity of the ki-ras polymerase chain reaction for diagnosing pancreatic cancer from pancreatic juice and bile: a multicenter prospective trial. J Clin Oncol 2002;20:4331-7. 31. Gervasoni A, Monasterio Munoz RM, Wengler GS, Rizzi A, Zaniboni A, Parolini O. Molecular signature detection of circulating tumor cells using a panel of selected genes. Cancer Lett 2008;263:267-79. 32. Went PT, Lugli A, Meier S, Bundi M, Mirlacher M, Sauter G, et al. Frequent EpCam protein expression in human carcinomas. Hum Pathol 2004;35:122-8. 33. Fehm T, Sagalowsky A, Clifford E, Beitsch P, Saboorian H, Euhus D, et al. Cytogenetic evidence that circulating epithelial cells in patients with carcinoma are malignant. Clin Cancer Res 2002;8:2073-84. 34. Willipinski-Stapelfeldt B, Riethdorf S, Assmann V, Woelfle U, Rau T, Sauter G, et al. Changes in cytoskeletal protein composition indicative of an epithelial-mesenchymal transition in human micrometastatic and primary breast carcinoma cells. Clin Cancer Res 2005;11:8006-14. 35. Goeminne JC, Guillaume T, Symann M. Pitfalls in the detection of disseminated non-hematological tumor cells. Ann Oncol 2000;11:785-92. 36. Adusumilli PS, Chun YS, Mullerad M, Chan M, Yu Z, BenPorat L, et al. Fluorescence assisted cytologic testing (FACT): ex-vivo viral method for enhancing detection of rare cancer cells in body fluids. Personal communication. 37. Sabile A, Louha M, Bonte E, Poussin K, Vona G, Mejean A, et al. Efficiency of Ber-EP4 antibody for isolating circulating epithelial tumor cells before RT-PCR detection. Am J Clin Pathol 1999;112:171-8. 38. Stillwagon GB, Order SE, Haulk T, Herpst J, Ettinger DS, Fishman EK, et al. Variable low dose rate irradiation (131I-anti-CEA) and integrated low dose chemotherapy in the treatment of nonresectable primary intrahepatic cholangiocarcinoma. Int J Radiat Oncol Biol Phys 1991;21: 1601-5.