WO2001009611A2

WO2001009611A2 - An vivo identification of specific binding molecules for cancer detection

Info

Publication number: WO2001009611A2
Application number: PCT/US2000/020273
Authority: WO
Inventors: David N. Krag; Lyn Oligino
Original assignee: University Of Vermont And State Agricultural College
Priority date: 1999-07-29
Filing date: 2000-07-26
Publication date: 2001-02-08
Also published as: AU6608600A; WO2001009611A3

Abstract

Methods of identifying a subject-specific, tissue-specific binding molecule in a human subject are provided. The method involves injecting a library of molecules into a human and identifying molecules which bind to a specific tissue. The binding molecule can be used to direct drugs to the specific tissue of the human to treat various diseases such as cancer. The method can also be used to direct an immunomodulatory molecule to a tissue to stimulate the production of a local immune response at the tissue. The methods are particularly advantageous in developing individualized therapies. The methods can also be used to detect cancer cells in a subject after treatment for cancer and preferably from a subject who has undergone surgery.

Description

IN VIVO METHODS FOR THE IDENTIFICATION OF TARGET SPECIFIC BINDING MOLECULES IN A HUMAN AND THEIR USE IN CANCER DETECTION

Field Of The Invention

The present invention relates to methods of identifying a subject-specific and/or tissue-specific binding molecule in a human subject by injecting a library of molecules into a human and identifying molecules which bind to a specific tissue. The binding molecule can be used to direct drugs to the specific tissue of the human or to other humans. The present invention also relates to methods for the detection of cancer cells during or after treatment, of humans and non-human animals. Background Of The Invention

It has been said that one-third of all people in the United States will develop cancer. Although remarkable progress has been made in understanding the biological basis of cancer and its treatment, cancer remains one of the leading causes of death in the United States. A major problem in developing treatments for cancer has been generally due to the many different types of cancers as well as the heterogeneity of individual malignancies.

It is a well-accepted fact that for most cancers early detection will lead to more favorable outcomes. However early detection of either primary or secondary tumor lesions has proved difficult given the presence of small numbers of cancerous cells and/or small tumor masses which are not readily palpable or visible. Cancer is currently treated using a variety of modalities including surgery, radiation therapy and chemotherapy. The choice of treatment modality will depend upon the type, location and dissemination of the cancer. For example, surgery and radiation therapy may be more appropriate in the case of solid well-defined tumor masses and less practical in the case of non-solid tumor cancers such as leukemia and lymphoma. In these latter cancers and in the case of disseminated disease, chemotherapy is usually required. Commonly, some combination of these modalities is used.

More than 50 chemotherapeutic agents have been developed for the treatment of cancer. Included among chemotherapies for cancer is the use of combination therapy, in which two or more chemotherapeutic agents having different mechanisms of action are given concurrently. The results typically can be additive. Not all tumors, however, respond to chemotherapeutic agents and others although initially responsive to chemotherapeutic agents may develop resistance. As a result, the search for effective anti-cancer drugs and drug combinations has intensified in an effort to find even more effective agents for treating the myriad of cancers.

A major problem in the treatment of cancer is that present therapies lack specificity for tumor cells and are extremely toxic to normal cells. Although many tumor-associated antibodies have been identified, coupled to cytotoxic agents or alone, the performance of antibodies (Abs) or Ab fragments in clinical trials has been disappointing (Doerr RJ et al. (1991), Ann Surg 214: 118-24; Krag D (1992), Applied Radiology 21 : 30-33; Krag DN (1993), JNucl Med 34: 545-8; Krag DN et al. (1993), Arch Surg 128: 819-23; Krag DN et al. (1992), J Surg Oncol 51: 226-30). The major reasons for their failure in this regard are likely due to the unfavorable pharmacokinetics, lack of tumor penetration, and immunogenicity of molecules this large. Because of this disappointing progress in tumor-targeted therapy over past decades, it is clear that dramatically innovative approaches are needed. Exciting progress has been made in the elucidation of key molecules involved in breast cancer; however, effective ways to exploit these tumor-specific targets for therapy have not yet been developed. Single chain Fv (sFv) Ab fragments have been developed with high affinity to the clinically important breast cancer target erbB2 (Schier R et al. (1996), J Moi Biol 255: 28-43; Schier R et al. (1996), J Moi Biol 263: 551-67) and may show promise clinically. However, the vast majority of in vivo effective drugs are of much lower molecular weight than sFvs (25kD) and the discovery of smaller tumor-specific ligands will be extremely valuable. Ligands much smaller than antibody fragments may have important advantages in targeted therapy including improved tumor to non-tumor uptake ratios, better penetration of solid tumors, and non- immunogenicity. Small molecules are also easier to synthesize in the large amounts necessary for clinical use.

Large libraries of small compounds are a rich source of potential small ligands to tumor targets. Several types of these libraries which consist of millions of different molecules have been constructed and used to isolate small ligands to many targets. The construction of such libraries and their use in the identification of specific ligands, known as combinatorial technology, has revolutionized the field of drug discovery.

Summary Of The Invention The invention relates to the treatment of cancer and the detection of cancer cells during the course of treatment using in vitro and in vivo methods. One of the major obstacles to successful cancer therapy at present is the inability to uniquely detect and thus distinguish cancer cells from normal cells of the body. This obstacle is manifest in the inability to specifically target cancer cells for treatment, thus resulting in unnecessary toxicity to normal cells and hemopoietic suppression in the subject. The inability to identify cancer cells specifically within a subject can also preclude early diagnosis. Additionally, when a malignant growth is removed from a subject, such as the surgical removal of a tumor, cancerous cells may still be present in the subject.

The invention, in part, relates to a method for detecting malignant cells during and following the course of therapy. The invention provides a method for identifying a subject in need of aggressive anti-cancer therapy comprising detecting a cancer cell in a subject following treatment for a primary tumor mass. The detection of the cancer cell in the subject indicates the need for aggressive anti-cancer therapy. In important embodiments, the subject has no detectable metastases.

The invention further provides a method for identifying a subject in need of aggressive anti-cancer therapy comprising detecting a cancer cell in a subject within 2 days and two and a half months following treatment for a primary tumor mass. In one aspect, the cancer cell is detected within the second day and the two and half month time point following non-high dose chemotherapy treatment to remove a primary tumor. According to this latter aspect, the invention embraces the detection of a cancer cell on any day within the second day and the two and half month time point after treatment for a primary tumor mass. In one embodiment, the treatment to remove a primary tumor is surgery. In yet another embodiment, the treatment is a combination of treatment modalities including but not limited to surgery, radiation therapy, and non-high dose chemotherapy. In a preferred embodiment, the treatment is surgery. In one embodiment, the cancer cell is detected within 75, 45, 30, or 14 days following treatment. In another embodiment, the cancer cell is detected 10 days following treatment. In still another embodiment, the cancer cell is detected 2 days following treatment. In another aspect, the invention provides a method for identifying a subject in need of aggressive anti-cancer therapy comprising detecting a cancer cell in a subject within 2 to 4 days following high dose chemotherapy treatment for a primary tumor mass. In preferred embodiments, the cancer cell is detected within 2 days following high dose chemotherapy treatment for a primary tumor mass. In another preferred embodiment, the cancer cell is detected within 3 days following high dose chemotherapy treatment for a primary tumor mass. According to some embodiments, cancer cells are detected with a binding molecule. In another embodiment, cancer cells are detected with RT-PCR. In important embodiments, a cancer cell is detected by interaction with a binding molecule. Such interaction includes binding of the cancer cell to the binding molecule. In some embodiments, the binding molecule is a tissue-specific binding molecule. In still other embodiments, the binding molecule is selected from the group consisting of an antibody, an antibody fragment, a ligand for an intracellular or an extracellular receptor, a lectin or a supravital intracellular dye. In yet other embodiments, a combination of a tissue-specific binding molecule and another binding molecule, as described above, may be used to detect the cancer cell. The binding molecules can be conjugated to a label such as a fluorochrome, an enzyme, a biotin molecule, a magnetic compound, a radioactive molecule and the like. In some embodiments, a single binding molecule alone is capable of uniquely identifying a cancer cell from a subject. In other embodiments, a plurality of binding molecules is necessary to uniquely identify a cancer cell. According to one embodiment, the tissue-specific binding molecule used to detect the cancer cell is present in a library of molecules. In another embodiment, the tissue-specific binding molecule is present in a peptide phage display library.

In one embodiment of the invention, the cancer cell contacts the binding molecule in vitro. In yet another embodiment, the cancer cell contacts the binding molecule in vivo. According to still other embodiments, a library containing a tissue-specific binding molecule is administered to the subject. In still further embodiments, a peptide phage display library is administered to the subject. In one embodiment, the binding molecule is administered to the subject via parenteral or oral routes or inhalation. In preferred embodiments, the administration is intravenous. In another embodiment, the cancer cell is harvested following administration of the binding molecule to the subject.

In some embodiments of the invention, the cancer cell derives from a metastasis. In still other embodiments, the cancer cell derives from the primary tumor mass. In preferred embodiments, the cancer cell is present in blood.

In one embodiment, the binding molecule binds specifically to a prostate cancer cell. In preferred embodiments, the binding molecule binds specifically to a breast cancer cell.

The present invention also overcomes the prior art problems by providing methods for the identification of binding molecules which specifically target cancerous cells in vitro and, more importantly, in vivo. The methods of the invention, in some aspects, are aimed at identifying binding molecules which are cancer type as well as subject-specific. These binding molecules are useful both in the detection of cancer cells during treatment, as an indicator of disease progression or abatement, as well as in the delivery of therapeutic agents to the cancer cells themselves.

In one aspect, the invention relates to methods for the identification of target specific binding molecules in a human by administering libraries of potential therapeutic or targeting molecules to a human subject. The in vivo screening methods of the invention offer several potentially critical advantages over in vitro screening or in vivo screening in experimental animals. For example, tumor targets will be in their native conformation with all their human post-translational modifications; only peptides which are stable in vivo will be inherently selected; only targets which are stable in vivo will be inherently targeted; efficient subtraction of library members which bind to normal tissue due to exposure of the injected library to the entire body; purification or even knowledge of targets is not necessary; and potential elucidation of novel tumor or other disease targets is possible.

The binding molecules identified by the methods of the invention can be coupled to cytotoxic agents and used to mediate the specific destruction of tumor or other diseased cells. Small ligands will likely have pharmacokinetics and tumor penetration superior to that of antibodies or antibody fragments, are less immunogenic, and will allow development of more effective targeted therapeutics. Small molecules are also easier to synthesize in the large amounts necessary for clinical use and are less likely to interfere with the effects of conjugated cytotoxic drugs.

In one aspect, the invention is a method of identifying a tissue-specific binding molecule in a human subject. The method includes the steps of administering to a human subject having a target tissue, a library of molecules, isolating a sample of the target tissue, and identifying a tissue-specific binding molecule that interacts with the tissue. Preferably the library of molecules is not a library biased for a NGR (Asn-Gly-Arg), RGD (Arg-Gly-Asp), or GSL (Gly-Ser-Leu) motif.

The library of molecules administered to the human subject may be any type of library available. Libraries of molecules are well known in the art. In a preferred embodiment the library of molecules is selected from the group consisting of a phage random peptide library, a peptides-on-plasmids library, a polysome library, an aptamer library, a synthetic peptide library, and a synthetic small molecule library. In another embodiment a plurality of different libraries of molecules are administered. The method is useful for identifying molecules which bind specifically to a particular tissue. The binding molecules can then be used as active agents if they are functional. For example, the binding molecules may inherently be capable of influencing cell growth and proliferation as well as specific cellular processes. As used herein and depending upon the specific process involved, the term "influencing" embraces inhibiting, enhancing or simply changing the course or rate of a process. According to a one embodiment of the invention, a tissue-specific binding molecule inhibits cell growth, or is a chemotherapeutic agent. Alternatively, the tissue-specific binding molecules may function to deliver an active agent to the tissue site. In this latter embodiment, an active agent such as a chemotherapeutic agent is conjugated to the binding molecule, through chemical bonding and the like. The molecules in one embodiment are useful for treating tumors. In this embodiment, the target tissue is a tissue having a tumor.

In yet other embodiments, the methods of the invention are useful for identifying binding molecules which are specific for normal tissue. Such binding molecules can be incorporated into a vaccine to deliver antigen to a specific tissue, for instance (e.g., lymph nodes). Binding molecules specific for a particular organ or cell type within a subject are also embraced by the invention, as is their use in the delivery of other, potentially active, agents to such organs or cell types. The methods of the invention can also be used to identify binding molecules which bind normal tissues ubiquitously, such as might be desirable in cases of systemic infection for example.

In another aspect, the invention provides a method for identifying tissue-specific binding molecules which interact with cancer cells in blood. Circulating cancer cells may be antigenically different from solid tumor cells. Thus binding molecules which recognize cancer cells in the context of a solid tumor may not be able to recognize cancer cells in blood. The invention in this aspect is useful for identifying circulating and non-disseminated cancer cells.

According to another embodiment of the invention a plurality of tissue-specific binding molecules that interact with the tissue are identified. The plurality of tissue-specific binding molecules can be screened to identify a disease-specific binding molecule that interacts with a diseased cell of a tissue but does not interact with a non-diseased cell of the tissue.

In another embodiment the library of molecules can be pre-screened to identify a panel of molecules which bind to the tissue in vitro or in vivo in non-human experimental subjects, and wherein the panel of molecules is administered to the human subject.

The library of molecules may be administered by means known in the art. In a preferred embodiment the library of molecules is directly injected into the tissue. In another embodiment the library of molecules is administered by intravenous injection.

The library of molecules may also be administered to the human subject a plurality of times. Preferably the plurality of times is between two and five times. In an embodiment of the invention the plurality of administrations of the library of molecules is performed within fourteen days. In another embodiment the plurality of administrations of the library of molecules is performed within ten days. In yet another embodiment, the plurality of administrations of the library of molecules is performed within seven days.

In another aspect the invention is a method of treating a human subject having a target tissue in need of treatment. The method involves the steps of administering to a human subject a tissue-specific binding molecule prepared according to the methods described herein conjugated to an active agent. In one embodiment, the tissue-specific binding molecule is one that was identified in the human subject being treated.

In one embodiment the active agent is a medicament. In important embodiments, the alone is capable of inhibiting cell growth or signal transduction. Preferably the medicament is a chemotherapeutic agent. In another embodiment the medicament is an anti-angiogenic factor. In yet another embodiment the active agent is an immunomodulatory agent. In still another embodiment, the active agent is a therapeutic agent. In yet another embodiment, the tissue-specific active agent is itself a therapeutic agent. In another preferred embodiment, the tissue-specific active agent is a chemotherapeutic agent.

A method of treating a human subject having tissue characterized by abnormal cell growth or abnormal cell function is provided according to another aspect of the invention. The method involves the step of administering to the human subject a tissue-specific binding molecule conjugated to an immunomodulatory agent to modulate an immune response at the tissue.

Preferably the immunomodulatory agent is an immune response-inducing compound which induces an immune response at the tissue. In one embodiment, the immune response- inducing compound is a peptide. In another embodiment the immune response-inducing compound is a carbohydrate. In another important embodiment the immunomodulatory agent is an immune response-inhibiting compound which inhibits an immune response at the tissue. In one embodiment the immune response-inhibiting compound is a peptide. In another embodiment the immune response-inhibiting compound is a carbohydrate. As used herein an immune response-inducing or an immune response-inhibiting compound can be a peptide, a carbohydrate or some combination thereof.

According to another aspect the invention is a method for treating a human subject having a target tissue in need of treatment with a tissue-specific binding molecule specific for that subject. The method involves the steps of: administering a library of molecules to a human subject having a target tissue, isolating a sample of the target tissue, identifying a tissue-specific binding molecule that interacts with the tissue, and administering to the human subject the tissue-specific binding molecule conjugated to an active agent.

In another aspect the invention is a method of identifying a tissue-specific active agent in a human subject. The method includes the steps of: administering a library of molecules to a human subject having a target tissue, isolating a sample of the target tissue, selecting at least one binding molecule isolated from the target tissue, and performing a functional assay to determine whether the binding molecule is a tissue-specific active agent.

Each of the limitations of the invention can encompass various embodiments of the invention. It is, therefore, anticipated that each of the limitations of the invention involving any one element or combinations of elements can be included in each aspect of the invention. Detailed Description Of The Invention

The invention, in part, is based on the use of the massive power of library technology to identify ligands which will selectively bind to diseased tissue or cells such as breast cancer cells by in vivo screening in an individual patient. Despite intensive research efforts for many years, present treatments for cancer are too often ineffective and cause a high degree of morbidity in the patients. Present treatments give poor results because they lack specificity for tumor cells. Treatments such as radiation and chemotherapy are toxic and destructive to normal cells as well as tumor cells. In one aspect, the methods of the invention will lead to the identification of drugs which home specifically to tumor cells only, and do not lead to the destruction of normal cells. The methods of the invention are also useful for developing ligands which will be useful for targeting drugs to diseased tissue other than cancer to avoid side effects associated with systemic administration of drugs. In still another aspect, the methods of the invention will allow for early detection of cancer cells following treatment leading to the identification of a patient population requiring further, immediate cancer therapy. The invention also provides methods for identifying binding molecules which bind to normal tissue.

Another problem with current methods of treating diseases such as cancer is the development of resistance to virtually all forms of therapy. The methods of the invention overcome these problems by providing a simple, straightforward method for identifying many different binding molecules which are capable of interacting with a specific target having potential therapeutic value. By utilizing several different therapeutic molecules, resistance to a particular therapy does not occur. Using the methods of the invention, binding molecules which are specific for normal tissue can also be identified. As an example, binding molecules which are specific for lymphoid tissues such as lymph nodes or spleen can be used to deliver antigens to such tissues for the purpose of enhancing an immune response, and thus may be incorporated into a vaccine. Binding molecules specific for a particular organ or cell type within a subject are also embraced by the invention, as is their use in the delivery of other, potentially active, agents to such organs or cell types. Such active agents include drugs, medicaments, growth factors, and the like. For example, a binding molecule specific for a particular cell type, such as for example a lymphocyte, a granulocyte, a macrophage or a hemopoietic stem cell, may be conjugated to a growth factor, such as a cytokine specific for that cell type, and then administered to a subject in order to stimulate, or in yet other cases to inhibit, growth or function of such cells. In still other embodiments, if subjects with a tumor are to be treated, the subject may receive at least two different binding molecules, one which is tumor-specific and one which is specific for a hemopoietic cell, none, one or both of which may be conjugated. In these latter embodiments, the tumor-specific binding molecule may be conjugated to a chemotherapeutic agent and the hemopoietic cell-specific binding molecule may be conjugated to a cytokine which maintains or stimulates hemopoietic function, for instance. The methods of the invention can also be used to identify binding molecules which bind normal tissues ubiquitously, such as would be needed in cases of systemic infection for example. Binding molecules specific for normal tissue are also useful in a variety of preventative medicine therapies. For example, a person at high risk of developing a disease, such as breast cancer, can be treated prophylactically to prevent the growth and development of pre-cancerous cells in mammary tissue. The invention in one aspect relates to a method for identifying a tissue-specific binding molecule in a human subject. The method is accomplished by administering a library of molecules to a human subject having a target tissue, isolating a sample of the target tissue, and identifying a tissue-specific binding molecule that interacts with the tissue. The library of molecules is preferably not a library biased for a NGR, RGD, or GSL motif.

The tissue-specific binding molecule is identified by administering a library of molecules to the human subject. Libraries which consist of millions or even billions of different peptides, oligonucleotides, or synthetic compounds have been constructed and used to isolate small ligands to many targets in vitro. The construction of libraries like these, and their use in the identification of specific ligands, known as combinatorial technology, has revolutionized the field of drug discovery (Gallop MA et al. (1994), J Med Chem 37: 1233- 51). Using a combinatorial approach, random peptide libraries (RPLs) have been constructed which contain millions of different peptides. The power of these libraries lies in their vast size and in the ability to determine the amino acid sequence of even one binding peptide out of millions using currently available technology. The small size of the library particles allows manipulation of millions of different potential binding units in a few micro liters. The methods of the invention utilize this library technology to identify small ligands to in vivo tumor and other disease targets.

A "library of molecules" as used herein is a series of molecules displayed such that the compounds can be identified in a screening assay. The library may be composed of molecules having common structural features which differ in the number or type of group attached to the main structure or may be completely random. Libraries are meant to include but are not limited to, for example, phage display libraries, peptides-on-plasmids libraries, polysome libraries, aptamer libraries, synthetic peptide libraries, synthetic small molecule libraries and chemical libraries. Methods for preparing libraries of molecules are well known in the art and many libraries are commercially available.

One type of library, which is known as a phage display library, includes filamentous bacteriophage which present a library of peptides or proteins on their surface. Phage display libraries can be particularly effective in identifying tissue-specific binding molecules. Briefly. one prepares a phage library (using e.g. ml 3, fd, lambda or T7 phage), displaying inserts from 4 to about 80 amino acid residues using conventional procedures. The inserts may represent, for example, a completely degenerate or biased array. One then can select phage-bearing inserts which bind to the target tissue by administering the library to the human subject and isolating a sample of the tissue. This process can be repeated through several cycles of re- administration and re-selection of phage that bind to the target tissue. Repeated rounds lead to enrichment of phage bearing particular sequences. DNA sequence analysis can be conducted to identify the sequences of the expressed polypeptides. The minimal linear peptide or amino acid sequence that binds to the tissue can be determined. One can repeat the procedure using a biased library containing inserts containing part or all of the minimal linear portion plus one or more additional degenerate residues upstream or downstream thereof. In one embodiment it is preferred that the library of molecules is not a library biased for an NGR, RGD, or GSL motif. As used herein "a library of molecules is not a library biased for an NGR, RGD, or GSL motif refers to a library which is not specifically generated having inserts which include NGR, RGD, or GSL.

In certain other embodiments, the library does not contain more than about 10% of displayed peptide sequences which include a NGR, RGD, or GSL motif. In other embodiments, the library does not contain more than about 50% of displayed peptide sequences having these motifs. And in yet other embodiments, the library does not contain more than about 75% of displayed peptide sequences having these motifs.

Vectors are meant to include, e.g., phage, viruses, plasmids, cosmids, or any other suitable vector known to those skilled in the art. The vector has a gene, native or foreign, the product of which is able to tolerate insertion of a foreign peptide. By gene is meant an intact gene or fragment thereof. In the invention, the expressed gene product contains the foreign peptide expressed from the inserted nucleic acid molecule or DNA.

For certain embodiments of this invention, e.g., where phage display libraries are employed, the preferred vectors are filamentous phage, though other vectors can be used. Filamentous phage are single-stranded DNA phage having coat proteins. Preferably, the gene that the foreign nucleic acid molecule is inserted into is a coat protein gene of the filamentous phage. Preferred coat proteins are gene III or gene VIII coat proteins. Insertion of a foreign nucleic acid molecule or DNA into a coat protein gene results in the display of a foreign peptide on the surface of the phage. Insertion into any other gene product in which the inserted peptide is displayed can also be used in this invention. Examples of filamentous phage vectors which can be used in the invention are fUSE vectors, e.g., fUSEl, fUSE2, fUSE3 and fUSE5, in which the insertion is just downstream of the pill signal peptide. Smith and Scott, (1993) Methods in Enzymology 217:228-257. Yet another phage which can be used for the methods of the invention is the φX174 (Ochs HD et al. (1971), J Clin Invest 50: 2559-68), which has been shown to be non-toxic in humans. Libraries of φX174 have been used extensively for IV injection in humans. φX174 particles contain four different coat proteins, any of which may be useful for the display of foreign peptides.

In other embodiments, e.g., where internal libraries are employed, the preferred vectors are plasmids, though other vectors can be used. The gene that the nucleic acid is inserted into is a gene which also results in display of the inserted peptide sequence.

By recombinant vector it is meant a vector having a nucleic acid sequence which is not normally present in the vector. The foreign nucleic acid molecule or DNA is inserted into a gene present on the vector. Insertion of a foreign nucleic acid into a phage gene is meant to include insertion within the gene or immediately 5' or 3' to, respectively, the beginning or end of the gene, such that when expressed, a fusion gene product is made. The foreign nucleic acid molecule that is inserted includes, e.g., a synthetic nucleic acid molecule or a fragment of another nucleic acid molecule. The nucleic acid molecule encodes a displayed peptide sequence.

By displayed peptide sequence is meant a peptide sequence that is on the surface of, e.g. a phage or virus, a cell, a spore, or an expressed gene product. It is preferable to have the displayed peptide displayed such that it is able to bind to added target molecules. A displayed peptide sequence can be identical to, or not identical to, a naturally occurring peptide sequence.

The displayed peptide sequence can vary in size. As the size increases, the complexity of the library increases. In certain embodiments, the complexity of the library is at least about 10⁸ to about 10¹¹ Preferably, the complexity is at least about 10⁹. It is preferred that the total size of the displayed peptide sequence (the random amino acids plus any spacer amino acids) should not be greater than about 100 amino acids long, more preferably not greater than about 50 amino acids long, and most preferably not greater than about 25 amino acids long.

Peptide libraries may also be created in plasmids. For instance, DNA encoding the peptides can be inserted into the lac operon to produce a lad fusion protein. Many other types of peptide laboratories are known by those of skill in the art.

Small molecule combinatorial libraries may also be generated. A combinatorial library of small organic compounds is a collection of closely related analogs that differ from each other in one or more points of diversity and are synthesized by organic techniques using multi-step processes. Combinatorial libraries include a vast number of small organic compounds. One type of combinatorial library is prepared by means of parallel synthesis methods to produce a compound array. A "compound array" as used herein is a collection of compounds identifiable by their spatial addresses in Cartesian coordinates and arranged such that each compound has a common molecular core and one or more variable structural diversity elements. The compounds in such a compound array are produced in parallel in separate reaction vessels, with each compound identified and tracked by its spatial address. Examples of parallel synthesis mixtures and parallel synthesis methods are provided in U.S.S.N. 08/177,497, filed January 5, 1994 and its corresponding PCT published patent application W095/18972, published July 13, 1995 and U.S. Patent No. 5,712,171 granted January 27, 1998 and its corresponding PCT published patent application W096/22529, which are hereby incorporated by reference.

Synthetic DNA and RNA in libraries are also commonly used in the art. For instance, Ellington and Szostak describe the use of random polynucleotide libraries to identify novel ligands (Ellington and Szostak, Nature, 346, 818-822 (1990)).

In certain embodiments, the libraries may have at least one constraint imposed upon the displayed peptide sequence. A constraint includes, e.g., a crosslink, a stacking interaction, a positive or negative charge, hydrophobicity, hydrophilicity, a structural motif and combinations thereof. In certain embodiments, more than one constraint is present in each of the displayed peptide sequences of the library. These multiple constraints can be adjacent to, or not adjacent to, each other, and can be identical to, or not identical to, each other.

A crosslink includes, e.g., a disulfide bond. In certain embodiments, the displayed peptide has at least one cysteine residue. A structural motif includes, e.g., a zinc finger formation, a leucine zipper, and a β-turn structure in the peptide. The sequences Asp-Gly or Pro-Gly are likely to induce β-turns, either alone or in combination with, e.g., a disulfide bond.

By using a large panel of libraries which offer a variety of structural contexts for peptide presentation, it is possible according to the invention to isolate peptides which bind to many promising targets such as breast cancer targets. Many of the libraries used in the prior art for in vivo panning methods are biased for integrin binding sequences or short loops. (Pasqualini R. Ruoslahti E (1996) Nature 380: 364-6) The invention also encompasses methods of identifying peptides that bind to tumor cells or are taken up by tumor cells and not just endothelial cells. Binding peptides capable of extravasation will be useful in some aspects of the invention due to their ability to exit from the vasculature and contact extravascular cells such as those present within tissues, organs and solid tumors. The tissue-specific binding molecules identified by the methods of the invention, or a fragment thereof also can be used to screen peptide libraries, including phage display libraries, to identify and select binding partners of the tissue-specific binding molecules. Such binding partners can then be used for screening assays or for purification protocols, etc. Using these binding partners which have targeting moieties that mimic the binding site of the tissue- specific binding molecule identified in the methods of the invention, peptide analogs and nonpeptides or peptidomimetics can be identified. For instance, the binding partner may be used to identify small molecules with the same binding specificity of the tissue-specific binding molecule or other tissue-specific binding molecules. Alternatively, it may be used to purify large batches of the tissue-specific binding molecules identified by the methods of the invention and for other purposes that will be apparent to those of ordinary skill in the art.

Such molecules can be rationally designed based upon the known sequence and/or structure of the tissue-specific binding molecules.

The screening methods of the invention are not likely to cause toxicity in the human subjects as there is an extensive body of literature describing injection of bacteriophage intravenously in humans and even neonates with essentially no side effects. (Ching YC et al. (1966), J Clin Invest 45: 1593-600; Hamblin TJ et al. (1975), Clin Exp Immunol 21: 101-8; Ochs HD et al. (1971), J Clin Invest 50: 2559-68; Peacock DB et al. (1973), Clin Exp Immunol 13: 497-513; Slopek S et al. (1987), Arch Immunol Ther Exp (Warsz) 35: 569-83; Uhr JW, Finkelstein MS (1967), Prog Allergy 10: 37-83). The risk of side effects such as the development of an immune response is minimized by carefully controlling experimental parameters, such as time period over which the libraries are administered. Preferably the libraries are administered at least twice within fourteen days and more preferably at least twice within ten days to minimize the risk of developing an immune response.

In order to increase the specificity of a library for a target the phage may be pre- screened in vitro by, for example, an in vitro procedure such as biopanning. In this method the phage are exposed to a target tissue in vitro or fractions thereof for the incubation period of approximately 3 hours. The unbound phage are then removed and the target tissue material is washed, prior to eluting the phage. The eluted phage can be amplified and then further screened using the same methods. It is preferred that the tissue used for the biopanning procedure is the tissue of the human subject that will be used for the method of the invention. It is not always desirable to use a pre-screen step because the pre-screen step may eliminate molecules from the library which only recognize the tissue in vivo not in vitro. It is possible, however, for the step to be performed.

The libraries also may be pre-screened or post-screened to remove molecules which interact with the diseased tissue's normal counterpart. This can be accomplished by in vitro screening assays which are performed before or after the phage are administered to the human subject. The screening step may also be performed in vivo in an animal or a human subject. For instance, the same human subject may also have a normal sample of tissue removed and tested for the absence or presence of the target molecule, if normal tissue is present.

More than one library can be administered to the human subject at a time. In fact, the more libraries that are administered, the more extensive is the panel of specific target tissue binding molecules ultimately identified. Therefore, it is preferred according to the invention that more than one library is administered at a time. A large panel of random libraries which present a vast number of peptides presented in a variety of structural contexts is likely to yield higher affinity binding molecules than a single library.

The target tissue is any type of target tissue in which it is desirable to deliver an active agent directly to the tissue. The actual type of target tissue will ultimately depend on the disease to be treated. Target tissues include, for example, tumors, tissues deficient in an enzyme or other functional protein, infected tissues, and injured tissues. Tumors are useful targets because it is desirable to deliver a chemotherapeutic agent to the tumor without contacting any other cells in the body. The methods of the invention can be used to identify specific binding molecules that can then be conjugated to an active agent and used to deliver the active agent to the specific tissue.

The library of molecules may be administered by any means known in the art but is preferably administered to the human subject by intravenous injection. Depending on the type of tissue, however, the library of molecules may be administered by other mediums. For example, if the target tissue is the lung, then it is preferred that the library of molecules is administered by aerosol formulation. The library of molecules may also be administered orally, parenterally or locally by direct injection or implantation, such as at the time of surgery.

The library of molecules may be administered a single time, but preferably is administered a plurality of times. The library may be administered a plurality of times in order to more specifically identify target binding molecules. For instance, if the library is administered a single time and those molecules found to be associated with the target tissue are then isolated, purified and re-administered, then the likelihood of identifying target- specific molecules is increased. The plurality of times is preferably between two and five times but may be more than five.

When the library of molecules is administered more than one time the multiple administrations may be performed over any time period. If the library is administered at least twice within a fourteen day period, the likelihood of an immune response developing to the injected library is minimized. Therefore, it is preferred that the multiple administrations be performed within a fourteen day time period and more preferably within a ten day period. Once the libraries have been administered to the human subject, a sample of the target tissue should be obtained. The sample may be obtained, for example, by performing a biopsy. The molecules from the library that bind to the tissue are then isolated from the biopsied tissue and characterized to determine which peptides are expressed on the surface. These peptides are the target tissue-specific binding molecules of the invention. They may be isolated from the tissue by any means known in the art, the particular means depends of course on the type of library that was administered. For instance, peptide/protein phage libraries may be isolated and characterized using an E. coll agar assay. Briefly, this assay involves crushing the biopsied tissue and adding it to an E. coli broth to achieve selective growth of the E. coli. Selective growth can be achieved for example by including a bacterial resistance marker in the bacteriophage. The bacteriophage are then plated on agar following the growth step and allowed to form colonies. A colony is then selected and the sequence of the peptide produced by that bacteriophage clone can be deduced by DNA sequence analysis. The strength of interaction between the target tissue and the binding molecules may also be assessed by various means known in the art.

Identification of the displayed peptide sequence includes, e.g., determining the sequence of amino acids that comprise the peptide. Identification can be accomplished, e.g., by amplifying the recombinant vector which has the nucleic acid sequence which encodes for the displayed peptide sequence which binds to the target, and sequencing the nucleic acid sequence by standard procedures known in the art to determine the displayed peptide sequence which binds to the target. If desired, the peptide thus identified can be synthesized using standard procedures known in the art.

Once binding molecules have been identified, the function of these molecules can be determined. A binding molecule may, for instance, not have any additional function other than the ability to bind to the tissue. Other binding molecules, however, may also function as an active agent. For instance, if the tissue is a tumor, the binding agent may function as a chemotherapeutic agent without further conjugation to other molecules. When the binding molecule has a function, such as the ability to kill or prevent further growth of a cancer cell, the molecule is referred to as a "tissue-specific active agent." The function of the tissue- specific active agent may be assessed in an in vitro assay or even an in vivo functional assay. A "functional assay" as used herein is any in vitro or in vivo assay routinely used in the art that establishes that a molecule is capable of acting on the target tissue to produce a therapeutic result or is capable in conjunction with another therapeutic molecule of producing a therapeutic result. The type of assay performed will depend on the disease and the tissue involved but such assays are routinely used in the art.

The human subject, according to the invention, is a human having a target tissue to which it is desirable to develop binding molecules. This tissue is any tissue which it would be desirable to deliver an active agent directly to the tissue, for instance, a tumor or other diseased tissue. In general, the tissue-specific binding molecules may be developed or identified in any such human subject. These tissue-specific binding molecules can then be used to deliver the active agent to any human subject having the same type of disorder or used directly as an active agent.

In one embodiment of the invention, however, the tissue-specific binding molecule is identified in the same human subject that will eventually be treated with the binding molecule. This aspect of the invention is advantageous because it allows for the production of a panel of binding molecules which are individualized or customized for that particular patient. Many diseases have slightly different etiologies in different patients. As a result a binding molecule which would interact with one individual's tissue may not interact with another patient's tissue. By developing a panel of binding molecules for a specific patient, that patient's diseased tissue can be more accurately targeted.

The method in this aspect of the invention involves the steps of administering to a human subject having a target tissue, a library of molecules, isolating a sample of the target tissue, identifying a tissue-specific binding molecule that interacts with the tissue, and administering to the human subject the tissue-specific binding molecule conjugated to an active agent. Additionally, by performing screens in one human subject, rather than serially as was done with mice, it is possible according to the invention to isolate customized ligands to important targets unique to a particular individual. More generic targets can be identified by serial screens in different patients. Years of cumulative research, however, have indicated that resistance to these molecules will invariably develop through modulation or development of alternative enzyme systems. A system which has only one target has a track record of clinical failure. The screening and the development of customized drugs can be performed within a matter of weeks and repeated as necessary as a method of overcoming drug resistance. The screens may be initially performed on patients with advanced disease but they may also be performed in all patients. It is possible to establish a profile of ligands against all newly diagnosed patients with a disease such as breast cancer immediately prior to definitive surgery. This would allow design of systemic adjuvant therapy to any disease which is most appropriate to each patient.

The tissue-specific binding molecules identified according to the invention can be modified and/or used as a prototype in order to develop other small molecules which will be effective in vivo. Substitution with D-amino acids and non-natural amino acids may confer greater biological half-life to peptides. The NMR structure of peptides can also be used to model peptidomimetics.

The tissue-specific binding molecule may be conjugated to an active agent and administered to a subject, or may be administered alone if the binding molecule has activity as discussed above. In addition to developing novel tissue binding molecules, the invention also encompasses a method of effecting therapy once the ligand binds to the tumor or diseased cell. The invention also involves attaching an immunomodulatory agent to the target tissue binding agents of the invention. The immunomodulatory agent can be an immune response inducing compound such as, for example, an immunogenic compound. Preferably, the immunogenic compound is one to which most people have already been immunized against. The binding of a molecule bearing an immunogenic compounds to the tumor cell surface may stimulate the immune system to eliminate the tumor cell. Thus, instead of using dangerous chemicals or radiation, it is possible according to the invention to direct the body's own immune system to more naturally eradicate tumor cells. Immunomodulatory agents also encompass immune response-inhibiting compounds. Such compounds when conjugated to a tissue-specific binding molecule may serve to diminish an inappropriate immune response such as for example, an autoimmune response. Immunomodulatory agents, such as the immune response-inducing and immune response-inhibiting compounds discussed above and embraced by the invention include, but are not limited to, peptides, carbohydrates, peptide mimetics, small molecule glycolipids, as well as combinations thereof. An example of a carbohydrate capable of inducing an immune response is α-Gal,2 which is found on cell surface glycoproteins and glycolipids in non-primate mammals and New World monkeys, but not in humans.

The present invention is also premised, in part, on the observation that cancer cells remained in the blood in a subset of cancer patients at 2 days after the cessation of treatment to remove a tumor mass. The presence of cancer cells in the blood of most cancer patients declined to undetectable levels within 2 days of surgical removal of the tumor mass. The proportion of breast cancer patients with detectable early post-operative cancer cells in blood approximates the proportion of patients expected to succumb eventually to the disease. It has been discovered according to the invention that the early detection of continued presence of cancer cells following treatment is useful for identifying this latter patient population, leading to more directed aggressive anti-cancer therapy. Residual cancer cells following primary treatment reflects continued malignant growth at the site of the primary lesion, perhaps indicative of incomplete resection, or in a lymph node or, in some instances, at secondary, metastatic sites. In these latter cases, the metastatic sites may or may not be apparent using conventional diagnostic methods. According to the methods of the invention, the remainder of patients who do not show signs of residual cancer cells immediately after treatment may then be spared additional aggressive anti-cancer therapy and its related toxicity and hemopoietic suppression. As used herein, primary or initial treatment refers to the treatment to which the subject is initially subjected for the purpose of removing the primary tumor. Also as used herein, secondary, adjunct or aggressive anti-cancer therapy or treatment refers to the treatment administered to the subject following the detection of cancer cells after primary treatment.

Analysis of patients at early time points post primary treatment such as those documented herein has not been previously reported; rather, patients have been more routinely monitored at monthly and, in some cases, yearly periods after the cessation of treatment. Thus, the ability to assess the location, burden, and growth potential of cancer cells outside the primary tumor in patients at early times post-treatment, as provided by the present invention, will greatly facilitate prognosis and subsequent patient management.

Currently despite efforts to increase the accuracy of prognosis of patients having clinically meaningful systemic deposits of cancer, no definitive statement can be made regarding any individual breast cancer patient. The ability to customize a method of cancer cell detection to a particular patient, as provided by the methods of the invention, is most advantageous.

The invention provides a method for identifying a subject in need of aggressive anti-cancer therapy comprising detecting a cancer cell in the subject following treatment for a primary tumor mass. In one aspect, the invention provides a method for identifying a subject in need of aggressive anti-cancer therapy comprising detecting a cancer cell in the subject within 2 days and two and half months following treatment for a primary tumor mass. The preferable time point to monitor subjects for the presence of cancer cells will depend upon the type of treatment administered to the subject in order to treat the primary tumor mass. For subjects who have undergone a non-high dose chemotherapeutic treatment for a primary tumor mass, the method includes detection of cancer cells at any point, or at any combination of points, between 2 days and two and half months after such treatment. In important embodiments, the cancer cell is detected in the subject within 2-14 days of non-high dose chemotherapeutic treatment for a primary tumor. High dose chemotherapy as used herein refers to doses of one or more chemotherapeutic agents capable of eliminating a primary tumor and intended to induce a remission. A non-high dose chemotherapeutic treatment is intended to include, but not be limited to, low dose (e.g., maintenance dose) chemotherapy, surgery, radiation or some combination thereof. Subjects who have received chemotherapy treatment to remove a primary tumor mass should be monitored preferably between 2 and 4 days after such treatment, and even more preferably, between 2 and 3 days after such treatment. The detection of such a cancer cell in the subject indicates the need for aggressive anti-cancer therapy, while the lack of such a cancer cell spares the subject from further anti- cancer therapy.

A cancer cell is a cell that divides and reproduces abnormally due to a loss of normal growth control. Cancer cells almost always arise from at least one genetic mutation. In some instances, it is possible to distinguish cancer cells from their normal counterparts based on profiles of expressed genes and proteins, as well as to the level of their expression. Of the markers currently available, few are able to identify a cancer cell uniquely. More commonly, a panel of markers, with corresponding expression levels, is used in order to identify a cancer cell. Genes commonly affected in cancer cells include oncogenes, such as ras, neu/HER2/erbB, myb, myc and abl, as well as tumor suppressor genes such as p53, Rb, DCC, RET and WT. Cancer-related mutations in some of these genes leads to a decrease in their expression or a complete deletion. In others, mutations cause an increase in expression or the expression of an activated variant of the normal counterpart. Markers which are expressed at the cell surface, rather than intracellularly, are usually more useful in the identification of cancer cells.

As used herein with respect to the aspect of the invention relating to identifying a subject in need of aggressive anti-cancer therapy, a subject includes but is not limited to a human, non-human primate, dog, cat, cow, pig, bird, sheep, goat, horse, rodent and fish. Some aspects of the invention are methods in which a subject is limited to a human subject. In these latter aspects, the subject is always referred to as a human subject.

In some embodiments, the subject has a cancer type characterized by a solid mass tumor. The solid tumor mass, if present, is preferably a primary tumor mass. A primary tumor mass refers to a growth of cancer cells in a tissue resulting from the transformation of a normal cell of that tissue. In most cases, the primary tumor mass is identified as the largest mass of cancer cells detectable in the body, and can be found through visual or palpation methods. However, some primary tumors are not palpable and can be detected only through medical imaging techniques such as X-rays (e.g., mammography), or by needle aspirations. The use of these latter techniques is more common in early detection. Molecular and phenotypic analysis of cancer cells within a tissue will usually confirm if the cancer is endogenous to the tissue or if the lesion is due to metastasis from another site.

Cancers include but are not limited to: biliary tract cancer; brain cancer; breast cancer; cervical cancer; choriocarcinoma; colon cancer; endometrial cancer; esophageal cancer; gastric cancer; intra-epithelial neoplasms; liver cancer; lung cancer (e.g. small cell and non- small cell); melanoma; neuroblastomas; oral cancer; ovarian cancer; pancreas cancer; prostate cancer; rectal cancer; sarcomas; skin cancer; testicular cancer; thyroid cancer; and renal cancer, as well as other carcinomas and sarcomas. In some embodiments, the subject has prostate cancer. In other embodiments, the subject has breast cancer. Either of these types of cancers can be screened for by the presence of a cyst or irregularity in shape, texture or weight of the tissue. A cancer cell as used herein does not include a cell from a hemopoietic malignancy such as a leukemic or lymphoma cell.

In yet other aspects of the invention, the subject has no detectable metastases. A metastasis is a region of cancer cells, distinct from the primary tumor location resulting from the dissemination of cancer cells from the primary tumor to other parts of the body. At the time of diagnosis of the primary tumor mass, the subject may be monitored for the presence of metastases. Metastases are most often detected through the sole or combined use of magnetic resonance imaging (MRI) scans, computed tomography (CT) scans, blood and platelet counts, liver function studies, chest X-rays and bone scans in addition to the monitoring of specific symptoms.

According to the method of the invention, a subject diagnosed with cancer can be monitored during or, more preferably, following treatment. Treatment for a primary tumor mass is any invasive or non-invasive procedure aimed at reducing or eliminating the cancer cell burden at the site of a primary tumor mass, and can include, but is not limited to, surgery, radiation therapy and chemotherapy, alone or in combination. This treatment is commonly administered locally to the primary tumor mass. In preferred embodiments, the subject has recently undergone a surgical procedure to remove a solid tumor. In some embodiments, the subject has prostate cancer and has undergone prostatectomy to remove a prostate tumor mass. In preferred embodiments, the subject has breast cancer subject and has undergone a surgical procedure to remove a breast tumor, such as lumpectomy, partial, radical or segmental mastectomy, quandrantectomy, or wide excision.

A subject who has had a tumor mass surgically removed will be monitored at early time points post-surgery to determine if any cancer cells exist in a particular tissue or in the peripheral blood. As used herein, "early time points following treatment" refers to a two and a half month period following the cessation of treatment. The subject may be monitored 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3 or 2 days following the cessation of treatment. The subject may also be monitored 15, 21, 28, 35, 42, 50, 60, 75 days following the cessation of treatment as well as every day therebetween up to and including the two and a half month time point after initial treatment ends. In preferred embodiments, subjects are monitored for cancer cells as early as 10 days following treatment. Even more preferred is the monitoring of subjects as early as 2 days following treatment. The subject may also be monitored multiple times during the first two and a half month period following the cessation of treatment. For example, the subject may be monitored at 2 days after treatment, and again at 10 days after treatment and again at 14 days after treatment. The subject may also be monitored on each and every day, as well as in any combination of days within the first two and a half month period after primary or initial treatment has ended. For subjects who have received high dose chemotherapy as a primary treatment, early time points after treatment refers to 2-4 days after treatment, and more preferably to 2-3 days after treatment.

The cancer cells to be detected can be harvested from a variety of tissues, including the original tissue in which the primary tumor was located, the immediate area surrounding this tissue (e.g., in cases where the entire tissue may have been removed), a lymph node either proximal or distal to the primary tumor site, and in some instances, a suspected site of metastasis such as for example, bone marrow. Preferably, the peripheral blood is analyzed for the presence of cancer cells. Depending on the subject and disease profile, the detected cancer cell may derive from a primary tumor site which has been incompletely excised, a lymph node, a metastatic lesion or an unknown site. The methods of the invention are directed to the detection of non-hemopoietic lineage cells. Thus the term "cancer cells in blood" refers to non-hemopoietic lineage cells present in blood and usually deriving from a solid tumor. The cancer cells can be detected using the tissue-specific binding molecules of the invention, or other classes of binding molecules, or a combination thereof. Regardless of their nature, the binding molecules to be used in the detection methods provided herein are able to uniquely identify cancer cells, and thereby provide a readout of the efficacy of the prior treatment, and an indication of the potential need for further, possibly immediate, therapy. The invention enables earlier identification and treatment of patients who may have residual, disseminated and/or resistant cancer cells, than has currently been achieved.

Another technique which is useful for the detection of cancer cells in the blood is Reverse-Transcriptase Polymerase Chain Reaction (RT-PCR). RT-PCR is used to determine nucleic acid molecules expressed in a cell. The first step in RT-PCR is reverse transcription of mRNA molecules harvested from one or more cells using, in most cases, a poly-A nonspecific primer. The second step is to amplify the resultant cDNA to detectable levels, using a standard polymerase chain reaction and gene-specific primers. Detailed conditions for RT- PCR will depend upon the primers used and the gene of interest. Transcripts which are useful as markers of non-hemopoietic cancer cells in the blood include epithelial genes such as cytokeratins, mucin-1, carcinoembryonic antigen (CEA), EGFR/erbBl, neu/HER2/erbB2, estrogen receptor, progesterone receptor, prostate specific antigen (PSA) and prostate specific membrane antigen (PSMA). Primers specific for these genes are well-known to one of ordinary skill.

If the detection method requires the use of tissue-specific binding molecules such as those provided in the present invention, then these may be used in an isolated or purified form. The tissue or blood to be harvested may be exposed to the tissue-specific binding molecules in vivo, prior to harvest. Similarly, the other classes of binding molecules described herein may also be exposed to tissue and blood in vivo, prior to harvest.

According to aspects of the invention which relate to the identification of a binding molecule in a library, such as a phage display library, which is able to interact or bind to a cancer cell within blood, the library in its entirety may be administered to the subject. In this way, it will be possible to identify binding molecules within a library which specifically interact with cancer cells once they are present in the blood. It is possible that such binding molecules will not also interact or bind to such cancer cells in the context of the solid tumor. Thus the panel of binding molecules which interact with blood-borne cancer cells may not be identical to the panel which interact with non-disseminated cancer cells. In this particular aspect of the invention, the subject to whom the library is administered may be a human and a non-human animal.

Other classes of binding molecules, distinct from those of the invention, may be used in the detection methods. These include, but are not limited to, antibodies, antibody fragments, ligands for intracellular or extracellular receptors, lectins, supravital dyes and the like. Examples of antibodies or antibody fragments useful in the detection methods of the invention include, but are not limited to, those directed against the following antigens: for breast cancer, EGFR/erbBl, erbB2/neu HER2, estrogen receptor and progesterone receptor; for prostate cancer, PSA and PSMA. These binding molecules are commercially available from sources such as Sigma, Genentech, Oncogene Sciences, In Vitro Diagnostics and Pharmingen. Binding molecules specific for epithelial markers are useful for detection of cancer cells in the blood (Brandt B. et al, 1998, Int J Cancer, 76:824-8). Examples of epithelial specific markers include cytokeratins (e.g., CK18 or CK19, available from Santa Cruz Biotechnology and Oncogene Research Products), epithelial cell adhesion molecules (e.g., EPCAM), CEA antibodies to which are available from Upstate Biotechnology Incorporated and Oncogene Research Products, MART antigens and mucins (e.g., MUC-1, antibodies to which are available from Santa Cruz Biotechnology). Other binding molecules useful in detecting cancer cells in blood include Panorex® 17-1 A (Centacor), 3622W94 (Glaxo Wellcome), Herceptin (Genentech), C225 (ImClone Systems), BEC2 (ImClone Systems), Ovarex (Altarex), 4B5 (Novopharm Biotech, Inc.), anti-VEGF, RhuMAb (Genentech), MDX-210 (Medarex/Novartis), MDX-220 (Medarex), MDX-447 (Medarex), MDX-260 (Medarex), CYT-424 (Cytogen), Atragen® (Aronex Pharmaceuticals), OV 103

(Cytogen), MELIMMUNE-1 and -2 (IDEC), CEACIDE™ (Immunomedics, Inc.), NovoMAb- G2 (Novopharm Biotech, Inc.), TNT (Techniclone Corporation), Bliomab-H (Novopharm Biotech, Inc.), GNI-250 (Genetics Institute), EMD-72000 (Merck KgaA), Monopharm-C (Novopharm Biotech, Inc.), BABS (Creative BioMolecules), anti-Flk-2 (ImClone Systems), ANA Ab (Procyon Biopharma, Inc.), SMART ABL 364 (Protein Design Labs), ImmunRAIT- CEA (Immunomedics, Inc.). In preferred methods, a cocktail of epithelial specific binding molecules is used to identify cancer cells in blood. Similar to the binding molecules of the invention, the nature of these latter binding molecules may be, but is not limited to, peptide, peptidomimetic, carbohydrate, chemical, organic, nucleic acid, aptamer, or some combination thereof.

Each of the binding molecules used in these methods may in turn be conjugated to a detectable label such as those commonly used in flow cytometry, immunohistochemistry and immunocytochemistry. These labels include fluorochromes such as fluorescein isothiocyanate, fluoroescamine, phycoethythrin, Texas Red®, allophycocyanin, phycocyanin and rhodamine; biotin, avidin or streptavidin; radioactive molecules, chemiluminescent compounds such as luminol, isoluminol, aromatic acridinium esters, imidazoles, and oxalate esters; bioluminescent compounds such as luciferin and luciferase; and enzymes such as peroxidase, alkaline phosphatase, β-galactosidase, glucose-6-dehydrogenase, maleate dehydrogenase and glucose oxidase. As illustrated in the Examples, binding molecules can also be conjugated to magnetic compounds. Magnetic compounds ranging in size from 0.7- 1.5 μm have been described in U.S. Patent Nos. 3,970,518; 4,018,886; 4,230,685; 4,267,234; 4,452,773; 4,554,088; and 4,659,678 and are also commercially available as particles (e.g., BioMags®, Advanced Magnetics, Inc., Cambridge, MA) or beads (e.g., Dynabeads®) or colloids (i.e., nanoparticles suspended permanently in water that act like molecules) such as Ferrofluid (Immunicon, Philadelphia, PA). Separation of cells labeled with magnetic compounds can be effected by the application of a magnet within the vicinity of the cells (e.g., usually on the outside of a culture tube or plate). Commercially available magnets and magnetic separators include quadrupole and hexapole magnetic separators (Immunicon, Philadelphia, PA), MAIA Magnetic Separator (Serono Diagnostics, Norwell, MA), Dynal MPC-1 (Dynal, A.S., Oslo, Norway), BioMag Separator (Advanced Magnetics, Cambridge, MA) and MACS (Miltenyi Biotec GmbH Gladback, West Germany). The cells may then be visualized using detection methods such as, for example, flow cytometry, immunohistochemistry, immunocytochemistry and the like.

According to some embodiments, tissue or blood is harvested from the subject and then exposed to the binding molecules in vitro. Alternatively, the cells can be exposed to at least one of the binding molecules in vivo via administration of the binding molecule to the subject prior to harvest of tissue or blood. The binding molecule may be administered to the subject in a purified or isolated form, or in the form of one or more libraries. In one embodiment, at least one of the administered libraries is a peptide phage display library.

Identification of a small subset of cancer cells may require a combination of positive and negative selection procedures. In negative selection, cells are separated so as to remove or identify extraneous cells within the population. For example, a separation procedure such as cell density or cell size separation can be performed in order to reduce the number of extraneous cells and thus enrich for the population of cancer cells. Cells can also be separated using negative selection based on what they fail to bind. Negative selection is generally followed by a positive selection procedure in which cells of interest are identified by what they do bind.

Magnetic separation using binding molecules conjugated to a magnetic compound can also be useful in the detection methods of the invention (Hildebrandt, M. et al., 1997, Exp. Hematol., 25:57-65; Naume, B. et al, 1997, J. Hematether. 6:103-14). In this technique, the binding molecule, and the cell to which it binds, can be physically manipulated in the presence of a magnetic field. The complex of the magnetically labeled binding molecule and the cell to which it is bound are separated from the cells which do not specifically bind the binding molecule. When used as a positive selection step, this procedure is sensitive enough to recover 75% to 100% of cancer cells added to collected samples of blood. For example, it has been demonstrated that when only 10 cancer cells are added to 10 ml blood, 7 to 10 cancer cells can be recovered.

Aggressive anti-cancer therapy as used herein refers to a secondary treatment or a combination of treatments capable of inflicting high toxicity in a subject, both in tumor and normal cells. Usually, this therapy is administered systemically (i.e., to the entire body) and is most commonly associated with excessive toxic effects, such as for example, hemopoietic suppression. As used herein, the terms adjunct or secondary treatment or therapy are used interchangeably with aggressive anti-cancer therapy to refer to the treatment administered to a subject following the detection of cancer cells subsequent to the initial treatment to remove the primary tumor. Examples of treatments to be used in aggressive anti-cancer therapy include, but are not limited to, radiation therapy, chemotherapy, and therapeutic agent administration, or some combination thereof. In some instances, a surgical procedure may also be used in this adjunct therapy. Chemotherapy administered in adjunct secondary therapy may be high dose chemotherapy, including more than one cytotoxic agent. The nature of the chemotherapeutic agent(s) to be used will depend upon the type of cancer and its grade, and will be known to one of ordinary skill in the art of oncology. Radiation administered at this time may include total body irradiation. Similarly, any further surgical procedure which may be performed at this time may encompass a more drastic excision at the primary tumor site than that which may have been performed previously. For example, the detection of cancer cells after a partial mastectomy may call for a radical mastectomy procedure. Alternatively, the surgical procedure may serve to remove tumorous tissue at other sites in the body.

In the case of breast cancer, aggressive anti-cancer therapy can include treatments such as, inter alia, ovarian ablation, tamoxifen administration, and chemotherapy. Again, as stated earlier, chemotherapy may be administered as a single agent or a combination of agents. For example, potential agent combinations may include cyclophosphamide, methotrexate and 5- fluorouracil (CMF); cyclophosphamide and doxorubicin (CA); and cyclophosphamide. doxorubicin and 5-fluorouracil (CAF). The initiation and timing of aggressive anti-cancer therapy (i.e., secondary or adjunct therapy) will depend upon the type of treatment modalities used and, in some instances, the cancer burden within the blood. Adjunct therapy can be initiated within hours, days or weeks of the initial detection of cancer cells following treatment to remove the primary tumor mass. As an example, such therapy may begin within 12, 24, 36 or 48 hours of detection of cancer cells in the blood post- primary treatment. The therapy can also be initiated within 3, 4, 5, 6, 7, 10, 14, 21, 30 and every day therein between following the detection of cancer cells in the blood. Aggressive anti-cancer therapy can be administered over a period of days, weeks or months, depending on the modality and combination used. Chemotherapy can be administered to a subject for as little as one month to as long as 24 months, with treatment preferably lasting at least 4 months and most preferably 6 months.

When administered, the pharmaceutical preparations are applied in pharmaceutically- acceptable amounts and in pharmaceutically-acceptable compositions. Such preparations may routinely contain salt, buffering agents, preservatives, compatible carriers, and optionally other therapeutic agents. When used in medicine, the salts should be pharmaceutically acceptable, but non-pharmaceutically acceptable salts may conveniently be used to prepare pharmaceutically-acceptable salts thereof and are not excluded from the scope of the invention. Such pharmacologically and pharmaceutically-acceptable salts include, but are not limited to, those prepared from the following acids: hydrochloric, hydrobromic, sulfuric, nitric, phosphoric, maleic, acetic, salicylic, citric, formic, malonic, succinic, and the like. Also, pharmaceutically-acceptable salts can be prepared as alkaline metal or alkaline earth salts, such as sodium, potassium or calcium salts. The compositions used in the methods of the invention may be combined, optionally, with a pharmaceutically-acceptable carrier. The term "pharmaceutically-acceptable carrier" as used herein means one or more compatible solid or liquid filler, diluents or encapsulating substances which are suitable for administration into a human or other animal. The term "carrier" denotes an organic or inorganic ingredient, natural or synthetic, with which the active ingredient is combined to facilitate the application. The components of the pharmaceutical compositions also are capable of being co-mingled with the molecules of the present invention, and with each other, in a manner such that there is no interaction which would substantially impair the desired pharmaceutical efficacy.

The pharmaceutical compositions used in the methods of the invention may contain suitable buffering agents, including: acetic acid in a salt; citric acid in a salt; boric acid in a salt; and phosphoric acid in a salt.

The pharmaceutical compositions also may contain, optionally, suitable preservatives. such as: benzalkonium chloride; chlorobutanol; parabens and thimerosal.

Compositions suitable for parenteral administration conveniently comprise a sterile aqueous preparation of the compositions of the invention, which is preferably isotonic with the blood of the recipient. This aqueous preparation may be formulated according to known methods using suitable dispersing or wetting agents and suspending agents. The sterile injectable preparation also may be a sterile injectable solution or suspension in a non-toxic parenterally-acceptable diluent or solvent, for example, as a solution in 1,3-butane diol. Among the acceptable vehicles and solvents that may be employed are water, Ringer's solution, and isotonic sodium chloride solution. In addition, sterile, fixed oils are conventionally employed as a solvent or suspending medium. For this purpose any bland fixed oil may be employed including synthetic mono- or di-glycerides. In addition, fatty acids such as oleic acid may be used in the preparation of injectables. Carrier formulation suitable for oral, subcutaneous, intravenous, intramuscular, etc. administrations can be found in Remington's Pharmaceutical Sciences, Mack Publishing Co., Easton, PA. A variety of administration routes are available. The particular mode selected will depend of course, upon the particular drug selected, the severity of the condition being treated and the dosage required for therapeutic efficacy. The methods of the invention, generally speaking, may be practiced using any mode of administration that is medically acceptable, meaning any mode that produces effective levels of the active compounds without causing clinically unacceptable adverse effects. Such modes of administration include oral, rectal, topical, nasal, interdermal, or parenteral routes. The term "parenteral" includes subcutaneous, intravenous, intramuscular, or infusion. Intravenous or intramuscular routes are not particularly suitable for long-term therapy and prophylaxis. They could, however, be preferred in emergency situations. Oral administration will be preferred for prophylactic treatment because of the convenience to the patient as well as the dosing schedule.

Other delivery systems can include time-release, delayed release or sustained release delivery systems. Such systems can avoid repeated administrations of the compositions by methods of the invention described above, increasing convenience to the subject and the physician. Many types of release delivery systems are available and known to those of ordinary skill in the art. They include polymer base systems such as poly(lactide-glycolide), copolyoxalates, polycaprolactones, polyesteramides, polyorthoesters, polyhydroxybutyric acid, and polyanhydrides. Microcapsules of the foregoing polymers containing drugs are described in, for example, U.S. Patent 5,075,109. Delivery systems also include non-polymer systems that are: lipids including sterols such as cholesterol, cholesterol esters and fatty acids or neutral fats such as mono- di- and tri-glycerides; hydrogel release systems; sylastic systems; peptide based systems; wax coatings; compressed tablets using conventional binders and excipients; partially fused implants; and the like. Specific examples include, but are not limited to: (a) erosional systems in which the compositions of the invention is contained in a form within a matrix such as those described in U.S. Patent Nos. 4,452,775; 4,675,189; and 5,736,152, and (b) diffusional systems in which an active component permeates at a controlled rate from a polymer such as described in U.S. Patent Nos. 3,854,480; 5,133,974; and 5,407,686. In addition, pump-based hardware delivery systems can be used, some of which are adapted for implantation.

Use of a long-term sustained release implant may be particularly suitable for treatment of chronic conditions. Long-term release, are used herein, means that the implant is constructed and arranged to delivery therapeutic levels of the active ingredient for at least 30 days, and preferably 60 days. Long-term sustained release implants are well-known to those of ordinary skill in the art and include some of the release systems described above.

Examples Example 1: Construction. Preliminary Screening and Analysis of Phage Displayed RPL. Using the techniques described herein, we have achieved very low backgrounds in both screening and analysis. Additionally, the techniques minimize degradation of the displayed peptides, maximize formation of disulfide bonds within the peptides, and increase the likelihood of a "hit" by employing several novel elution schemes. Techniques for analysis of clones such as DNA sequencing, ELISA, enrichment assays (phage titering), IF A, and spot blotting as employed herein are routine to the ordinary artisan. A novel colony screening assay has been which allows selection of higher affinity clones earlier in screenings, can potentially discriminate between high and low affinity clones, is capable of screening many more clones at once, and is far less labor-intensive than other phage clone assays. This assay will be very useful for discriminating tumor binders from normal tissue binders by using biotinylated protein extracts from both normal and tumor tissues. The peptides identified by these methods can be analyzed using HPLC purification and mass spectroscopy. Peptides can be labeled with biotin using a structureless glycine linker, and the affinity of these biotinylated peptides can be measured and thus compared to that of free unconjugated peptides. Most free peptides identified from RPL screening have binding affinity for target comparable to the original peptide-phage binder. Binding affinities can be measured using biosensors such as that commercially available from Biacore.

ErbB2 bindins molecules: A phage-displayed RPL containing 20 million different nonapeptides which can be constrained by a disulfide loop as the random peptides are flanked with cysteine codons was constructed. The library was screened with the breast cancer target ErbB2 in several forms: live human cells expressing ErbB2, purified native ErbB2, and an ErbB2 ECD-alkaline phosphatase fusion protein, and have resulted in the identification of several strong consensus amino acid sequences. Many inter-screen consensus sequences were detected, sometimes from screens using two different forms of ErbB2. Such inter-screen consensus sequences strongly suggest that the sequences are binding to the only common element in the different presentation systems, ErbB2 ECD. It is especially encouraging that the sequence of the peptide isolated from whole cells overexpressing ErbB2 correlates with one of the major consensus sequences from purified ErbB2. Although the affinity of the putative ErbB2-binding peptides that have been identified are not high enough to give a positive ELISA signal, the consensus sequences are valuable since higher affinity binders can often be identified from libraries biased for the consensus sequence of the initially identified peptides (Cwirla SE et al. (1997), Science 276: 1696-9; Martens CL et al. (1995), J Biol Chem 270: 21129-36; Wrighton NC et al. (1996), Science 273: 458-64; Yanofsky SD et al. (1996), Proc Natl Acad Sci U S A 93: 7381-6).

While ErbB2 is a promising breast cancer target, the methods of the invention can be used to identify ligands to any tumor or other disease target. Binders to several different targets have been identified including several relevant to cancer therapy. The quality of the library was confirmed by screening with streptavidin (SA). We identified the same consensus sequence as the other groups (i.e., HPQF) as well as additional consensus amino acids not previously identified. A completely novel binding sequence was also identified which, by ELISA and spot blot assays, seems to bind with a higher affinity than the HPQF peptides. These results demonstrate that the library is a rich source of potential small ligands, and that its design offers advantages over prior art libraries.

A peptide (Gl) which binds specifically to the SH2 domain of Grb2, an intracellular signal transduction protein has been identified from this library. Identification of Gl took less than two weeks with purified Grb2 target. The discovery of this peptide is particularly exciting in that the nonphosphorylated Gl binds to a site which normally binds only phosphorylated peptides. Highly charged phosphate groups will probably preclude using phosphorylated peptides as therapeutic inhibitor agents, as it will be difficult to deliver them through the cell membrane and intracellular phosphatases would quickly remove the phosphate group which is essential to their binding activity. When added to cell lysates, an analogue of Gl prevented the association of Grb2 and ErbB2. Because it is believed that the overexpression of ErbB2 is part of the pathogenesis of many breast cancers, compounds such as Gl which disrupt the ErbB2 signal transduction pathway may be useful in the treatment of breast cancer.

Construct a larse panel of phage-displayed random peptide libraries A RPL was constructed which displays 9 amino acids in the context of a disulfide loop and which has been used to identify peptide binders to several important tumor targets, as discussed above. To extend our capability to obtain peptide ligands to the variety of tumor-specific targets likely to be present on the tumor cells of breast cancer patients, several new libraries are constructed. Using new combinatorial technology techniques, these libraries are much larger than the original library, incorporate several design improvements, and display peptides in a variety of different structural contexts and display systems. Construction of such a large panel and variety of libraries supplies us with a rich source of potential peptide ligands to a wide variety of potential breast tumor targets.

Peptides displayed in these new RPLs contain cysteine disulfide-constrained loops of 8, 9, 10, 11 and 12 amino acids, flanked by 3-4 random amino acids such as that shown in Table 1. These random peptide loops are presented in gene III phage display (two different systems which offer different structural contexts) and gene VIII phage display (Cwirla SE et al. (1990), Proc Natl Acad Sci U S A 87: 6378-82; Scott JK, Smith GP (1990), Science 249: 386-90; Wrighton NC et al. (1996), Science 273: 458-64). The relative advantages and disadvantages of these systems are described in a recent review article (Scott J (1994), CRC Press, pp 1-27). All of the required degenerate oligonucleotides and vectors have been prepared and purified for construction of the libraries. The cloning techniques are standard and used routinely by those of ordinary skill the art. Table 1

X= random amino acid residues

E. coli infection with phase to generate phagemids Library phage are prepared from E. coli cultures by standard methods, centrifuged twice to remove bacteria, and purified by PEG precipitation and cesium chloride gradients. Filtering the suspension with .45 micron filters to remove bacteria completely helps to reduce background. As well, use a protease inhibitor cocktail during growth of phage and in phage solutions helps to minimize degradation of displayed peptides. The phage suspension is passaged twice through pyrogen-free 0.22 micron filters. The DNA from phage is then analyzed using restriction site mapping, and DNA sequencing.

Establish the safety of intravenous administration of phage RPLs in mice and preliminary screenins in mice. Tests for sterility, endotoxins, mycoplasma, toxicity and dose-response are performed in normal mice in order to test the material for human use according to FDA standards. Although standard sterility tests require a 14 day waiting period, there are other methods, as per FDA advice, that assure sterility for clinical use in a much shorter time. Endotoxins and mycoplasma can be detected in a matter of hours using commercially available kits. Toxicity and dose-response analysis is determined by injecting progressively larger amounts of phage into normal FVB mice. In addition to observation of the animals for any signs of toxicity for a 7 day waiting period, tissue distribution of injected phage is assessed by phage amplification and counts from at least 5 major organs. Histological analyses of tissues from these same organs is used to further assess potential toxicity. In some instances, screenings in humans can be performed using the protocol determined to be most effective in the animal studies. RP(s) are screened in subjects once, twice and preferably three times. In instances in which the animal experiments indicate that in vitro screening is just as effective as in vivo screening, then in vitro screening with resected tumor tissue can be performed.

Naive library phage are injected into normal mice and small tissue biopsies are performed within 10 minutes and within 24 hours. Phage bound and thus harvested with the biopsied tissue are amplified in E. coli, purified, sterilized and re-injected into both the same mouse and a different mouse to assess toxicity using a 7 day observation period, tissue distribution analysis and histological analysis of at least 5 normal tissues. This toxicity testing can be repeated as required.

In addition, transgenic mice with mammary tumors (MMTV-PyV middle T antigen) are useful as screening models for mammary tumor binding molecules. Consensus sequences and tumor-specific binding is determined by phage counts of both putatively specific and nonspecific phage from harvested tissue. Specificity of phage clone binding is assessed by immunohistochemistry using an anti-phage antibody. Peptides thus identified are synthesized with a biotin tag and their tumor to normal tissue homing profile is analyzed by immunohistochemistry with anti-biotin antibodies, which are commercially available.

Establish the safety of intravenous administration of phase RPLs in human patients with breast cancer. There are numerous descriptions in the literature of IV administration of E. coli phage into humans with no toxic reactions of any kind reported (Ching YC et al. (1966), J Clin Invest 45: 1593-600; Hamblin TJ et al. (1975), Clin Exp Immunol 21 : 101-8; Ochs HD et al. (1971), J Clin Invest 50: 2559-68; Peacock DB et al. (1973), Clin Exp Immunol 13: 497- 513; Slopek S et al. (1987), Arch Immunol Ther Exp (Warsz) 35: 569-83; Uhr JW, Finkelstein MS (1967), Prog Allergy 10: 37-83). This safety profile is reestablished using standard methods and tests for sterility and pyrogenicities. Throughout the screening process patients are carefully evaluated for adverse reactions. The level of anti-phage antibodies in each patient is measured by ELISA both prior to and post screening.

Identification of specific breast tumor-binding peptide-phase by in vivo screening and characterization of clones The new libraries constructed above are injected IV into breast cancer patients for in vivo tumor screening. Screen with freshly prepared peptide-phage to minimize degradation of the peptides and keep phage on ice as much as possible. Treatment of phage with air oxidation or DMSO does not affect viability of phage. Phage are recovered by biopsy from a small portion of freshly excised tumor tissue, due to the powerful amplification potential of phage.

After biopsy, phage are eluted from tumor cells and amplified as in in vitro whole-cell screening methods established by us and others (Arap W et al. (1998), Science 279: 377-80; Barry MA et al. (1996), Nat Med 2: 299-305; Fong S et al. (1994), Drug Development

Research 33: 64-70; Pasqualini R, Ruoslahti E (1996), Nature 380: 364-6). Low pH elutions resulted in only 15% viability of phage. For the first screen especially, other less destructive elution methods such as competitive elutions or high pH elutions may be optimal. An interesting and logical elution scheme involves adding the E. coli cells to be infected directly to the phage bound to target. Immediate dilution of eluted phage into cells or eluting with cells directly, as above, may prevent phage from rebinding target after pH neutralization which may decrease its ability to infect E. coli and be amplified, especially for a target as large as ErbB2. For ELISA, phage directly from culture supernatants (no PEG) are used. Ultrafiltration can be used for concentration but has not been necessary with binders of even moderate affinity. Phage amplifications can be minimal (e.g., overnight) provided there is sufficient amplification of specific binders to obtain enrichment. Presenting less displayed peptide to the target after the first few screens will not only decrease background but will select for higher affinity binders.

For both screening and analysis, excess peptide-phage or peptide ligands are washed at least five times in Tween TBS and fresh wash vessels are used whenever possible. Detergent will not be used in buffers to wash harvested tumor tissue before elution of phage. Use of a colony screening assay after the first or second screen can sidestep background problems since one positive colony producing tumor-binding peptide-phage out of thousands can be detected. A colony-screening assay can also identify highly-specific binders which, for unknown reasons, are not well amplified and enriched for during routing screening. Competitive elutions with integrin binding compounds or growth factors may yield useful specific binders.

A subtraction of peptide-phage which bind to normal tissue is performed at this step. although the IV injection and whole body screening process is likely to eliminate phage which bind to normal tissues and may be an important advantage to this system. The process is repeated 2-5 times within the same patient as soon as possible to avoid rejection of the peptide-phage ligands by a patient immune response. Screening RPLs and characterization of binders by enrichment analysis, DNA sequencing, ELISA, IFA and/or phage colony immunoblotting is performed using routine methods known to those of skill in the art. For phage ELISA, Nunc Maxisorb plates with "C" wells are optimal. A suitable blocker for phage clone assays is 0.1% Tween, except in the case of ELISA where a casein blocker (Pierce) is better than Tween as a polystyrene blocker. Specificity of peptide-phage clones for tumor- binding is determined relative to their binding of normal breast tissue excised at the same time as the tumor biopsy, as well as by the binding of non-specific phage to tumor. Any consensus sequences identified from phage eluted specifically from the tumor tissue are excellent candidates for tumor-specific peptides. Peptide binders identified by whole body screens, almost by definition, are stable in serum and in general stable in vivo, another major advantage to this technique. These experiments result in the identification of peptides which bind specifically to breast tumor cells or to blood vessels specifically supplying tumor cells in human patients. Whole body in vivo screening experiments will result in the development of methods which may allow identification of novel tumor targets.

Synthesize and characterize free peptides which bind to breast cancer -specific targets

Promising peptides identified as described above are synthesized on a peptide synthesizer, cyclized if necessary, and tested for specific binding to tumor tissue sections both directly (Pennington ME et al. (1996), Moi Divers 2: 19-28 ) and via competition with peptide-phage by methods known in the art (Arap W et al. (1998), Science 279: 377-80) using immunohistochemical staining and IFA. Peptides are tested for binding to the tumor tissue of the original patient as well as to the tumor tissues of other breast cancer patients. High affinity peptides are then coupled to cytotoxic agents such as doxorubicin and tested for their ability to kill tumor cells or treat other diseases in patients. Alternatively, the peptides can be conjugated to an immunogenic compounds, preferably one to which the patient has already been immunized against. The binding of a molecule bearing an immunogenic compound, such as for example, an immunogenic peptide, to the tumor cell surface should stimulate the immune system to eliminate the target cell. Successful completion of this step results in the generation of novel agents which may be used for greatly improved treatment of diseases such as breast cancer.

Example 2: In Vivo Administration of the Libraries

The source of patients will be through the UVM Breast Care Center which handles more than 200 breast cancer patients per year. The patients eligible for this will have advanced breast cancer with multiple superficial cancer nodules amenable to biopsy with minimal trauma. Life expectancy should exceed 4 months. Age range is be 30 to 70 years of age. The method of RPL preparation for human administration are performed according to Good Laboratory Practice (GLP) and all materials are prepared in a facility approved for Good Manufacturing Practice (GMP). In all cases GLP and GMP will be performed and presented to the granting agency for final approval. Standard methods to assure sterility and pyrogenicity are according to standards set by the FDA and consistent with NCI practices. Since human subjects are the primary focus of this research, all activities are completely reviewed by a Human Subjects Protection Committee.

Before phage injection, one sample of tumor tissue from the patient will be biopsied, snap frozen and sixty slides will be prepared for later testing of selected clones for tumor binding affinity. A phage displayed RPL pool, containing peptides displayed in five different size loops, is injected intravenously into a breast cancer patient. The library will be diluted in 250 ml saline and infused intravenously over 10 minutes into a breast cancer patient. Initially 10^9"10 pfu is injected as that amount was found to be completely non-toxic to humans in similar studies (Peacock DB et al. (1973), Clin Exp Immunol 13: 497-513). Higher numbers of phage, up to 10'^4"16 or more can also be used. Within minutes to hours, depending on the experiment, the PI harvests small amounts of both tumor and normal superficial tissue. Phage can be recovered from a small portion of freshly excised tumor tissue, due to the powerful amplification potential of phage. The tissue is rinsed to remove blood, ground and added to E. coli to amplify phage. Phage are eluted from tumor cells and amplified as in in vitro whole-cell screening methods. The presence of harvested phage is detected within hours by ECL spot blot using an anti-phage antibody. Phage is quantified more accurately by titering, with results available within 12 hours. If phage is present, they are absorbed with normal tissue and re-injected as soon as possible. The harvest and amplification is repeated 2- 5 times. Screening is ideally completed in less than 7 days to avoid a patient immune response to phage (Peacock DB et al. (1973), Clin Exp Immunol 13: 497-513). Peptide-phage clones are analyzed for tumor binding specificity by immunohistochemistry with anti-phage Ab on both tumor and normal tissue. Tissues are probed with anti-phage Ab both immediately after harvest and after adding more phage after amplification. The former method shows phage bound in vivo while the latter method is more likely to give a positive signal. Clones eluted from both tumor and normal tissue are subjected to DNA sequencing to look for consensus amino acid sequences of clones specifically isolated from tumor tissue. We also quantify the number of phage eluted per gram of tissue for each screen for both tumor and normal tissue. Binding to normal tissue can be assessed by immunohistology on normal, quick frozen breast tissue excised at the same time as the tumor biopsy and on a large panel of 32 different normal human tissues. Immunohistochemistry with anti-transferrin receptor mAb is used as a positive control to assure tissue and assay reliability. Screening phage-displayed peptide libraries and analysis of peptide-phage ligands is routine. Techniques for analysis of clones such as DNA sequencing, ELISA, enrichment assays, IFA, and spot blotting are routine to those of ordinary skill in the art. In vivo screening can include as discussed above, thorough "subtraction" with normal tissue before injection which will be more efficient by amplifying phage for only a few hours and using a large excess of normal tissue compared to the amount of tumor tissue from which the phage were harvested.

The affinity of free peptides is measured directly by adding a biotin group to the peptide via a glycine linker at the C-terminus for immunohistochemical and IFA analysis (Pennington ME et al. (1996), Moi Divers 2: 19-28). A phage-competition method (Pasqualini R, Ruoslahti E (1996), Nature 380: 364-6) is used in the event that adding the small biotin-linker group destroys the peptide-binding activity. Loss of binding activity after biotin conjugation is not likely since the peptides are originally isolated with a relatively huge phage particle attached at the C-terminus. Peptide binders identified by whole body screens, almost by definition, are likely to be stable in serum and generally stable in vivo, a major advantage to this technique.

Example 3 : Detection of Cancer Cells in the Blood After Surgery Methods

Patients with operable invasive breast cancer and a plan for surgical resection of the primary tumor and regional lymph nodes were eligible for entry to this study. Patients who had a previous excisional biopsy of the primary tumor were excluded. The mean age of the patients was 48 years. Patients were recruited from a variety of practice locations. Blood samples were obtained on two separate occasions before surgery. Blood samples were then obtained post surgery at 2, 4, 8, and 12 hours, and 1, 2, 7, and 14 days. Blood was drawn (15 ml) into a green top sodium, heparinized collection tube at the appropriate time interval and sent unfrozen to the reference laboratory by overnight mail. For enrichment of tumor by positive selection, the mononuclear cell fraction was isolated by Ficoll-Hypaque separation (Pharmacia, Upsala, Sweden) and washed twice in Leibovitz L-15 medium (GIBCO/BRL, Grand Island, NY), supplemented with 10% fetal bovine serum (L-15/FBS; GIBCO/BRL). The mononuclear cells were placed in L-15/FBS at a concentration of 5 x 10⁷ cells/ml. These cells were placed in PBS containing 1% FBS and 0.2% sodium citrate (PBS/FBS) and washed twice at 1000 rpm for 10 minutes. Following washing, all mononuclear cells isolated were incubated in PBS/FBS medium at room temperature for 15 minutes. Cells were incubated with anti-breast cancer epithelium mAbs (9184, 9187, 9189) at a total IgG protein concentration of 2.5 μg at 4°C for 30 minutes (Krag, DN et al., 1999, Breast Journal, in press). The cells were then washed twice with PBS/FBS to remove unbound antibody, and incubated with IgG-coated magnetic beads (Dynal, Inc., Oslo Norway) at a ratio of 1 bead/100 cells, at 4°C for 30 minutes with agitation. The cells were washed twice with PBS-FBS to remove unbound beads and placed in a test tube along with a magnet for two minutes at room temperature to bind cells. Unbound cells were discarded. Bound cells were removed by gentle aspiration, released from the separator, diluted in PBS- FBS medium, washed twice and resuspended in L-15 medium.

All bead:cell conjugates recovered from the magnet were used in the cytopreparations and immunostained. Cytopreparations were made at 1 x 10' - 1 x 10⁶ beads per slide, depending on the number of beads used for the enrichment.

Cytopreparations were fixed in 4% paraformaldehyde fixative, washed thoroughly in Dulbecco's modified phosphate, buffered saline (PBS; GIBCO/BRL) with 1% Triton X, and placed on an automated immunostainer (TechMate; Ventana, Tucson, Arizona). Alkaline phosphatases (AP) immunostaining was then performed as per manufacturer's instructions. Slides were incubated in the following order: Biotinylated anti-cytokeratin mix (anti- cytokeratins 8/18), blocking solution, alkaline phosphatase, chromogen, and finally, hematoxylin. Buffer washes were performed between each step.

Positive control slides consisted of cultured breast cancer cells (e.g. CAMA-1) seeded into normal leukapheresis products or bone marrow and immunostained as above. Negative control slides consisted of the patient's specimen immunostained with normal mouse serum at the same concentration as used for the breast anti-epithelial antibodies.

The total number of tumor cells identified by microscopic evaluation was recorded. Only cells with specific cytoplasmic staining with tumor morphology were considered positive. The tumor cell number was counted for each stained slide. The total number of tumor cells divided by the number of cells used for the enrichment (denominator) generated the tumor concentration for that sample. The number of cancer cells was compared by tumor size (Tl abs vs. T2 + T3), nodal status (N = 0 vs. N > 0) and grad I vs. II vs. III). Since a maximum of two preoperative bloods was available, an average number of cancer cells for these two observations was used. If there were only one observation, that value represented the average. If both were missing, the average was considered as missing. Two-sample t-tests and analysis of variance were used to examine the number of cancer cells at baseline. Since the standard deviations were not equal for most comparisons, nonparametric tests were performed using the Kurskal-Wallis test. A five-percent statistical significance level was used for all tests. The presence or absence of cancer cells in the blood at the preoperative or baseline stage was examined in combination with their presence or absence in the blood at two weeks post surgery using a statistical test for correlated proportions with the p-value being based upon a binomial test of symmetry.

Results

Twenty-four patients registered for this study. Three did not meet the eligibility criteria and 21 patients had blood samples analyzed. Of those patients who were eligible, 19 had ductal and 2 patients had lobular carcinoma. Twelve of the 21 patients were node- negative. Before surgery, the percentage of patients who had at least one sample positive for cancer cells in the blood was 95% (Table 2).

Following removal of the primary breast cancer, the number of patients with cancer cells in the blood decreased. For each of 2 days post surgery, the number of patients with cancer cells in the blood declined. About 30% of patients failed to clear cancer cells from the blood over the 14 days of the study (Table 2). Fifty-three specimens obtained from normal donors or patients with non-epithelial cancer were blindly evaluated with no false-positive results. A total of 18 patients had baseline and two week follow-up blood samples available for analysis. Seventeen (94%) of these 18 patients had cancer cells detected in their baseline blood samples. Sixty-five percent (11/17) of these patients positive at baseline did not have cancer cells detected in their follow-up blood samples. Six of the patients did have detectable cancer cells in the blood following surgery. The one patient without cancer cells in their baseline blood sample remained negative for cancer cells at two weeks. These results indicated that there was a statistically significant clearance of cancer cells (p < 0.001) at two weeks post surgery for those individuals who were positive at baseline.

Discussion

The data presented here demonstrated that 95% patients with operable breast cancer had cancer cells in their blood before removal of the primary tumor. These are profound findings, since many of the patients in this study had what is considered to be early stage disease.

Table 2 Percentage Of Patients With Cancer Cells In The Blood Over Time

* Zero hours equal pre-surgery samples.

Example 4: Detection of Cancer Cells Using Reverse-Transcription Polymerase Chain Reaction (RT-PCR).

Total RNA from blood and tissue specimens is prepared by the guanidinium thiocyanate method. Total RNA from various normal tissues, to be used as controls, is obtained commercially (Clontech Laboratories, Inc. Palo Alto, CA). The mRNA expression of epithelial markers such as CEA, MUC-1, CK18, CK19, as well as tissue specific markers such as PSA, PSMA, erbB2, erbB 1 and estrogen receptor are determined by RT-PCR using RNA from blood and tissue samples from subjects as well as normal controls. The cDNA preparations used in the RT-PCR reactions is synthesized by incubating total RNA template (2 μg), random hexamers (1.66 μg, Boehringer Mannheim, Indianapolis, IN) and MuLV reverse transcriptase (200U, Gibco BRL), in a total reaction volume of 25 μl, at 42°C for 1 hour. MuLV synthesized cDNA (2.5 μl/PCR reaction) is then amplified using gene specific primers (0.2 μM and AmpliTaq Gold DNA polymerase (2.5 U, PE Applied Biosystems, Branchburg, NJ) in 25 μl PCR reactions consisting of 30 cycles at a denaturation temperature of 94°C (1 min/cycle); an annealing temperature of 60°C (1 min cycle); and an extension temperature of 72°C (2 min cycle). As a control for genomic DNA contamination in the RT-PCR reactions, duplicate cDNA templates are prepared as above in the absence of MuLV reverse transcriptase and used in equivalent PCR reactions. Identification of RT-PCR products can be accomplished on the basis of size of the resulting amplified fragment. In addition, selected products are subcloned into pCR2.1 (Invitrogen, San Diego, CA) and the DNA sequence of the resultant clones is deduced using automated DNA sequencing (Cornell University DNA services, Ithaca, NY). Gene specific primers for a panel of useful antigens including MART, cytokeratins, mucins, (e.g., MUC-1), CEA, estrogen receptor, herceptin, progesterone receptor, PSA, PSMA, erbBl and erbB2 are prepared according to published sequences.

References

1. Arap W, et al. (1998), Science 279: 377-80

2. Barinaga M (1998), Science 279: 323-4 3. Barry MA, et al. (1996), Nat Med 2: 299-305

4. Blond-Elguindi S, et al. (1993), Cell 75: 717-28

5. Bonnycastle LL, et al. (1996), J Moi Biol 258: 747-62

6. Bowditch RD, et al. (1994), J Biol Chem 269: 10856-63

7. Ching YC, et al. (1966), J Clin Invest 45: 1593-600 8. Christian RB, et al. (1992), J Moi Biol 227: 711-8

9. Cwirla SE, et al. (1997), Science 276: 1696-9

10. Cwirla SE, et al. (1990), Proc Natl Acad Sci U S A 87: 6378-82 11. Doerr RJ, et al. (1991), Ann Surg 214: 118-24

12. Fong S, et al. (1994), Drug Development Research 33: 64-70

13. Gallop MA, et al. (1994), J Med Chem 37: 1233-51

14. Germino FJ, et al. (1993), Proc Natl Acad Sci U S A 90: 933-7 15. Haaparanta T, Huse WD (1995), Moi Divers 1 : 39-52

16. Hamblin TJ, et al. (1975), Clin Exp Immunol 21 : 101-8

17. Huse WD, et al. (1992) Biotechnology 24: 517-23

18. Krag D (1992) Applied Radiology 21 : 30-33

19. Krag DN (1993) J Nucl Med 34: 545-8 20. Krag DN, et al.(1993) Arch Surg 128: 819-23

21. Krag DN, et al.(1992) J Surg Oncol 51 : 226-30

22. Martens CL, et al. (1995) J Biol Chem 270: 21129-36

23. Ochs HD, et al. (1971), J Clin Invest 50: 2559-68

24. Oligino L, et al. (1997) J Biol Chem 272: 29046-52 25. Pasqualini R, Ruoslahti E (1996), Nature 380: 364-6

26. Peacock DB, et al. (1973), Clin Exp Immunol 13: 497-513

27. Pennington ME, et al. (1996) Moi Divers 2: 19-28

28. Schier R, et al.(1996) J Moi Biol 255: 28-43

29. Schier R, et al. (1996) J Moi Biol 263: 551-67 30. Scott J (1994) Identifying lead peptides from epitope libraries. In: Weiner DaW, WV (ed) Biological Approaches to Rational Drug Design. CRC Press, pp 1-27

31. Scott JK, Smith GP (1990) Science 249: 386-90

32. Skerra A, et al. (1991) Anal Biochem 196: 151-5

33. Slopek S, et al. (1987) Arch Immunol Ther Exp (Warsz) 35: 569-83 34. Sparks A, et al. (1996) Phage Display of Peptides and Proteins:A Laboratory Manual

35. Uhr JW, Finkelstein MS (1967), Prog Allergy 10: 37-83

36. Watkins JD, et al. (1997) Anal Biochem 253: 37-45

37. Watkins JD, et al. (1998) Anal Biochem 256: 169-77

38. Wrighton NC, et al. (1996), Science 273: 458-64 39. Yanofsky SD, et al. (1996) Proc Natl Acad Sci U S A 93: 7381-6

40. Halsted W.7 . AM. S.A. 1907; 25:61.

41. Fisher B. (1977) Cancer 40:574-587. 42. McQuire W.L (1992) N. Eng. J. Med. 326: 1756-1761.

43. Singleton T.P. (1992) Pathol. Ann. 27:165-190.

44. Thor A.D., et al. (1992) J. Natl. Cancer Inst. 84:845-855.

45. Brandt B., et al. (1998) Int. J. Cancer 76:824-8. 46. Eaton M.D., et al. (1997) Biotechniques 1997; 22:100-5.

47. Hildebrandt M., et al. (1997) Exp. Hematol. 25:57-65.

48. Naume B., et al. (1997) J Hematether 6:103-14.

The foregoing written specification is considered to be sufficient to enable one skilled in the art to practice the invention. The present invention is not to be limited in scope by examples provided, since the examples are intended as a single illustration of one aspect of the invention and other functionally equivalent embodiments are within the scope of the invention. Various modifications of the invention in addition to those shown and described herein will become apparent to those skilled in the art from the foregoing description and fall within the scope of the appended claims. The advantages and objects of the invention are not necessarily encompassed by each embodiment of the invention.

All references, patents and patent publications that are recited in this application are incorporated in their entirety herein by reference.

We claim:

Claims

1. A method for identifying a subject in need of aggressive anti-cancer therapy comprising: detecting a cancer cell in a subject within 2 days to two and a half months following treatment for a primary tumor mass, wherein the detection of the cancer cell in the subject indicates the need for aggressive anti-cancer therapy.

2. The method of claim 1 , wherein the cancer cell is detected with a binding molecule.

3. The method of claim 2, wherein the binding molecule is selected from the group consisting of an antibody, an antibody fragment, a ligand, a lectin and a dye.

4. The method of claim 2, wherein the binding molecule is a tissue-specific binding molecule.

5. The method of claim 4, wherein the tissue-specific binding molecule is present in a library of molecules.

6. The method of claim 5, wherein the library of molecules is a peptide phage display library.

7. The method of claim 2, wherein the cancer cell contacts the binding molecule in vivo.

8. The method of claim 2, wherein the cancer cell contacts the binding molecule in vitro.

9. The method of claim 2, wherein the binding molecule is conjugated to a label selected from the group consisting of a radioactive molecule, a biotin molecule, a fluorochrome, a magnetic compound and an enzyme.

10. The method of claim 1, wherein the cancer cell derives from a metastasis.

11. The method of claim 1 , wherein the cancer cell derives from the primary tumor mass.

12. The method of claim 1, wherein the cancer cell is present in blood.

13. The method of claim 2, wherein the binding molecule binds specifically to a breast cancer cell.

14. The method of claim 2, wherein the binding molecule binds specifically to a prostate cancer cell.

15. The method of claim 6, wherein the peptide phage display library is administered to the subject.

16. The method of claim 15 , wherein the administration is intravenous.

17. The method of claim 1, wherein the cancer cell is detected within 75 days following treatment.

18. The method of claim 1 , wherein the cancer cell is detected within 45 days following treatment.

19. The method of claim 1 , wherein the cancer cell is detected within 30 days following treatment.

20. The method of claim 1, wherein the cancer cell is detected within 14 days following treatment.

21. The method of claim 1 , wherein the cancer cell is detected within 10 days following treatment.

22. The method of claim 1, wherein the cancer cell is detected within 2 days following treatment.

23. The method of claim 15, wherein the cancer cell is harvested following administration of the peptide phage display library to the subject.

24. The method of claim 1 , wherein the treatment is surgery.

25. The method of claim 1, wherein the cancer cell is detected with RT-PCR.

26. The method of claim 1 , wherein the cancer cell is detected within 3 days following treatment.

27. The method of claim 26, wherein the treatment is high dose chemotherapy.

28. A method of identifying a tissue-specific binding molecule in a human subject, comprising: administering to a human subject having a target tissue, a library of molecules wherein the library of molecules is not a library biased for a NGR, RGD, or GSL motif; isolating a sample of the target tissue; and identifying a tissue-specific binding molecule that interacts with the tissue.

29. The method of claim 28, wherein the library of molecules is a phage random peptide library.

30. The method of claim 28, wherein the library of molecules is a plurality of different libraries.

31. The method of claim 28, wherein the tissue-specific binding molecule is a chemotherapeutic agent.

32. The method of claim 28, wherein the library of molecules is directly injected into the tissue.

33. The method of claim 28, wherein the library of molecules is administered by intravenous injection.

34. The method of claim 28, wherein the tissue is a tumor.

35. The method of claim 28, further comprising identifying a plurality of tissue-specific binding molecules that interact with the tissue.

36. The method of claim 35, further comprising the step of screening the plurality of tissue-specific binding molecules to identify a disease-specific binding molecule that interacts with a diseased cell of a tissue but does not interact with a non-diseased cell of the tissue.

37. The method of claim 28, wherein the library of molecules is pre-screened to identify a panel of molecules which bind to the tissue in vitro or in vivo in non-human experimental subjects, and wherein the panel of molecules is administered to the human subject.

38. The method of claim 28, further comprising the step of administering the library of molecules to the human subject a plurality of times.

39. The method of claim 38, wherein the library of molecules is administered between two and five times.

40. The method of claim 38, wherein the plurality of administrations of the library of molecules is performed within fourteen days.

41. The method of claim 38, wherein the plurality of administrations of the library of molecules is performed within ten days.

42. A method of treating a human subject having a target tissue in need of treatment, comprising: administering to a human subject a tissue-specific binding molecule of claim 1 conjugated to an active agent.

43. The method of claim 42, wherein the tissue-specific binding molecule is one that was identified in the human subject being treated.

44. The method of claim 42, wherein the active agent is a medicament.

45. The method of claim 44, wherein the medicament is a chemotherapeutic agent.

46. The method of claim 44, wherein the medicament is an anti-angiogenic factor.

47. The method of claim 42, wherein the active agent is an immunomodulatory agent.

48. A method of treating a human subject having tissue characterized by abnormal cell growth or abnormal cell function, comprising: administering to the human subject a tissue-specific binding molecule conjugated to an immunomodulatory agent to modulate an immune response at the tissue.

49. The method of claim 48, wherein the immunomodulatory agent is an immune response-inducing compound and wherein the immune response-inducing compound induces an immune response at the tissue.

50. The method of claim 48, wherein the immunomodulatory agent is an immune response-inhibiting compound and wherein the immune response-inhibiting compound inhibits an immune response at the tissue.

51. A method of treating a human subject having a target tissue in need of treatment with a tissue-specific binding molecule specific for that subject comprising: administering to a human subject having a target tissue, a library of molecules; isolating a sample of the target tissue; identifying a tissue-specific binding molecule that interacts with the tissue; and administering to the human subject the tissue-specific binding molecule conjugated to an active agent.

52. A method of identifying a tissue-specific active agent in a human subject, comprising: administering to a human subject having a target tissue, a library of molecules; isolating a sample of the target tissue; selecting at least one binding molecule isolated from the target tissue; and performing a functional assay to determine whether the binding molecule is a tissue- specific active agent.

53. The method of claim 52, wherein the tissue-specific active agent is a chemotherapeutic agent.