WO2000067771A1

WO2000067771A1 - Antiangiogenic endostatin peptides, endostatin variants and methods of use

Info

Publication number: WO2000067771A1
Application number: PCT/US2000/012063
Authority: WO
Inventors: Kristiina Vuori
Original assignee: The Burnham Institute
Priority date: 1999-05-06
Filing date: 2000-05-02
Publication date: 2000-11-16
Also published as: AU4982200A

Abstract

The invention provides an endostatin peptide having at least about 4-7 endostatin amino acid residues containing substantially the amino acid sequence of RLQD, RAD, DGK/R, or a functional equivalent thereof. The invention also provides an endostatin variant containing the amino acid sequence RGD, or a functional fragment thereof. Methods of inhibiting angiogenesis are also provided.

Description

ANTIANGIOGENIC ENDOSTATIN PEPTIDES , ENDOSTATIN VARIANTS AND METHODS OF USE

BACKGROUND OF THE INVENTION

This invention relates generally to cell survival, proliferation and migration and, more specifically, to the manipulation of these processes to promote or inhibit angiogenesis.

Angiogenesis is the process by which new blood vessels are formed. The cellular constituents of blood vessels include endothelial and smooth muscle cells and the growth of new vessels involves the migration of these cells into vascular tubules. New vessel growth can occur by several processes which include vasculogenesis, resulting in the formation of completely new vessels, branching of pre-existing vessels or vessel enlargement. All of these angiogenic processes are required for the continued supply of blood and nutrients to developing tissues. As such, angiogenesis plays a vital role during αormal neonatal development to maintain and extend the supply of blood to the developing fetus.

In the adult individual, angiogenesis processes are less prevalent in normal, mature tissues because neovascularization is relatively complete following development. However, processes that require regeneration of tissue such as wound healing and the growth cycle of the corpus luteum still require new vessel growth in the normal adult individual. Other than these two normal physiological processes, the occurrence of angiogenesis in the adult individual is generally found in association with a variety of pathological conditions as a means to supply blood and nutrients to aberrant cell growth. Of the number of pathological conditions that require angiogenesis for maintenance or progression of aberrant cell growth, the most notable include cancer and inflammatory disorders. Inhibition of angiogenesis in these and other angiogenesis-dependent diseases has been proposed as a possible therapy for restricting or preventing progression of the unwanted cell growth. For example, the inhibition or prevention of a tumor's blood supply has been proposed as a possible therapy against cancer.

A number of methods have been proposed for inhibiting angiogenesis. Many of these methods involve attempts to interfere with molecules or stimuli suggested to play a role in angiogenesis. For example, aFGF, bFGF and VEGF have been suggested as potential mediators of angiogenesis and attempts have been made to inhibit the activity of these molecules by either directly inhibiting receptor binding or, indirectly by inhibiting their secretion into the extracellular space as well as by inhibiting the signaling, expression or function of their cognate receptors. A variety of other molecules also have been suggested to exert inhibitory effects on angiogenesis and include, for example, collagenase inhibitor, basement membrane turnover inhibitors, angiostatic steroids, platelet factor 4 and thrombospondin to name a few. However, whether the methods involve targeting a proposed mediator of angiogenesis or attempt to exploit potential inhibitors of angiogenesis, the specificity of these methods for angiogenic processes, and therefore the efficacy of such methods, have not been sufficiently realized for therapeutic use.

Angiostatin and endostatin are two endogenous inhibitors of angiogenesis that have been described as specific inhibitors of endothelial cell proliferation. Both have been isolated from tumor bearing mice and as such have been postulated to play a role in balancing the positive effects of angiogenic molecules such as FGF, bFGF and VEGF which are produced at the site of a tumor. Additionally, both of these inhibitors correspond to fragments of larger proteins known to be present in the circulatory system and extracellular matrix. For example, angiostatin is an approximately 38 kiloDalton (kDa) internal fragment of plasminogen whereas endostatin is an approximately 20 kDa carboxyl-terminal fragment of collagen type XVIII. Angiostatin and endostatin are the subject matters of U.S. Patent Nos. 5,639,725 and 5,854,205, respectively. Neither of these patents demonstrate active functional fragments or describe specific sequences that mediate the anti-angiogenic activity of the intact molecules .

As an endogenous inhibitor of endothelial cell proliferation, endostatin has been described as a useful inhibitor of angiogenesis-dependent diseases without exhibiting toxic side effects observed with less specific inhibitors or with treatments directed against less specific targets such as one or more of the various positive effectors of angiogenesis described above. However, the usefulness of this naturally occurring inhibitor is not without drawbacks.

For example, the recombinant production of endostatin has encountered difficulties in expression and purification. Depending on the recombinant system, the amount of endostatin produced is either inadequate for efficient isolation and use in animal studies and, therefore, for efficient use in human patients. Alternatively, extremely high levels of endostatin are produced which unfortunately results in the aggregation of the molecule into an insoluble form. Although the aggregation of endostatin has been described as imparting a beneficial time-release characteristic onto the preparation, it is uncertain what additional, and unwanted, contaminants have co-aggregated with such a preparation. Moreover, the use of such aggregates is primarily limited to administration by intraperitoneal injection or surgical implantation. Such limitations severely hinders the flexibility of using endostatin in alternative modes of treatment and in alternative therapeutic regimes. In addition, there is at least one report describing that an endostatin species which is truncated does not effectively inhibit endothelial cell proliferation in its soluble form. This result indicates that either the optimal therapeutic potential of endostatin may only reside in the aggregated form of the molecule or that only the intact molecule is active .

Finally, the receptor for endostatin has yet to be identified. Although proposed to interact with heparin, the absence of an identified endostatin receptor limits the possibility of obtaining improved versions of endostatin, functional fragments and mimetic compounds. Without knowledge of the receptor identification of such improved versions and compounds can only be accomplished through laborious trial and error experimentation in animal model systems. Moreover, lack of an understanding of the endostatin receptor can lead to erroneous interpretations regarding the therapeutic actions of endostatin antagonists and agonists.

In addition to the endogenous angiogenesis inhibitor endostatin, other reports have described that angiogenesis can be inhibited by inhibiting the α_vβ₃ integrin. Integrins are a group of heterodimeric cell adhesion receptors, which predominantly mediate cell-extracellular matrix interactions, but also cell-cell adhesive events. Members of the integrin family are expressed on a wide variety of cells and tissues including endothelial cells. The a_vb₃ integrin is one marker for proliferating vessels. Antibody and RGD-containing peptide antagonists of the α_vβ₃ integrin have been described which induce apoptosis of the proliferative angiogenic vascular cells while leaving preexisting quiescent blood vessels unaffected. The inhibition of angiogenesis with c. β_j antagonists is the subject matter of U.S. Patent No. 5,753,230.

Although there has been much publicity of the usefulness of specific inhibitors for the _vβ₃ integrin in treating angiogenic-dependent diseases, over the past fifteen years, work in fields such as cell adhesion has resulted in a substantial effort to develop specific inhibitors of this integrin. However, aside from antibody derived inhibitors which have limited therapeutic utility due to eliciting a host immune response, th^ specificity of peptide and peptide mimetics has resulted in limited success. The lack of success has been attributed, in part, to the cross reactivity of many integrin species for the single peptide recognition sequence RGD (Arg-Gly-Asp) . Efforts to develop more specific inhibitors by modification of the RGD sequence itself, the surrounding amino acid sequences, the conformation of the molecule or various combinations of these approaches has improved the selectivity of the inhibitors, but has not yielded inhibitors which are highly selective for a particular integrin, or a particular disease. Nevertheless, there still exists a significant expenditure of time and money toward developing peptide and small molecule inhibitors of various integrins, including the α_vβ₃ integrin. This continuing effort is just one indication that more specific inhibitors are still required to be effective as therapeutics and authenticates the important role of integrins in numerous pathological conditions, including diseases dependent on angiogenesis.

Thus, there exists a need for highly specific inhibitors of angiogenesis which can be efficiently and reliably produced. The present invention satisfies this need and provides related advantages as well.

SUMMARY OF THE INVENTION

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 shows the amino acid sequence of endostatin and sequence alignment of endostatin species from human, mouse, and chicken collagen type XVIII and from human and mouse collagen type XV.

Figure 2 shows SDS-PAGE analysis of purified recombinant endostatin.

Figure 3 shows cell adhesion to endostatin is mediated through α_v and ₅ integrins. Figure 4 shows that immobilized endostatin promotes endothelial cell survival in an _vβ₃-dependent manner while soluble endostatin induces apoptosis.

Figure 5 shows cell adhesion to endostatin derived from collagen type XV is mediated through α_v and α₅ integrins .

Figure 6 shows that immbolized endostatin derived from collagen type XV promotes endothelial cell survival and soluble endostatin induces apoptosis.

Figure 7 shows the binding activity for an endostatin variant and for mutant forms of endostatin.

DETAILED DESCRIPTION OF THE INVENTION

A number of pathological conditions are dependent on angiogenesis and most tissues and organs can support angiogenesis in a diseased state. Most notable of these pathological conditions is the growth of solid tumors. Inhibition of angiogenesis in these pathological conditions reduces the progression of the disease. Inhibition of angiogenesis is therefore a therapy for a wide array of diseases including cancer, diabetic retinopathy, psoriasis and rheumatic disease. In contrast, promotion of angiogenesis has therapeutic effects in pathological conditions where there is an inadequate blood supply to the tissue. Such diseases include limb ischemia, myocardial infarction and peptic ulcers .

The invention is directed to endostatin peptides that exhibit antiangiogenic activity and to methods of inhibiting cell receptors which mediate the antiangiogenic effects of endostatin peptides. The identification of active endostatin peptides and their corresponding family of receptors expands the available modes of administration and treatment regimes for pathological conditions dependent on angiogenesis.

Moreover, the identification of active endostatin peptides and their cognate receptor allows the synthesis and identification of compounds which exhibit improved therapeutic benefits.

In one embodiment, the invention is directed to peptides derived from endostatin that exhibit binding activity and specificity to

integrins. The endostatin peptides contain one or more of the α -integrin binding regions having the amino acid sequence RLQD, RAD or DGK/R. The peptide can be used, for example, to effectively inhibit angiogenesis or treat an angiogenesis-dependent disease. An advantage of the endostatin peptides of the invention is that they provide the specificity of endostatin together with the availability of rapid and efficient methods for recombinant and chemical synthesis of peptides. The endostatin peptides of the invention can therefore be combined with numerous and well known modes of administration for the treatment of a wide variety of diseases dependent on angiogenesis.

In another embodiment, the invention is directed to an endostatin variant that exhibits enhanced α,-mtegrιn binding activity and specificity. The increased α_v integrin binding activity confers a commensurate increase in angiogenesis inhibitory activity and efficacy in treating angiogenesis-dependent diseases. The increased α integrin binding activity of the endostatin variant results from the modification of endostatin to contain the amino acid sequence RGD.

As used herein, the term "endostatin" is intended to mean the approximately 18-22 kDa C-terminal fragment of the non-collagenous 1 (NCI) domain of collagen XVIII known as endostatin. Endostatin corresponds substantially to the 183 amino acid residues which map to amino acids H₁₁₀₅-K₁₂₈₇ of human collagen XVIII or the 184 amino acid residues which map to the corresponding region of collagen XVIII from other known species. These residues also correspond to amino acids 132-316 of the NCI domain of collagen XVIII. The identification and characterization of endostatin has been described in, for example, O'Reilly et al . Cell , 88:277-285 (1987), whereas the nucleotide and deduced amino acid sequence of collagen XVIII has been described by Rehn and Pihlajaniemi, Proc . Na tl . Acad. Sci . USA 91:4234-4238 (1994) and by Oh et al . Proc . Na tl . Acad. Sci . USA 91:4229-4234 (1994), all of which are incorporated herein by reference. The nucleotide and amino acid sequences of endostatin from various species, including, human, mouse, and chicken are set forth as SEQ ID NOS: 1-6, respectively. The sequence identity between these species is about 62% or greater at the nucleotide level and about 58% or greater at the amino acid level. For example, the nucleotide sequence identity between human and chicken endostatin sequences is about 77% and the identity between human and mouse is about 82%. The endostatin amino acid sequence identity is about 86% for human and chicken, and about 95% for human and mouse. The endostatin sequences from these and other species are intended to be encompassed by the term endostatin as used herein. Therefore, the sequence differences between species is included within the meaning of the term. The term "endostatin" is similarly intended to include a peptide or fragment of a related collagen subclass such as the corresponding fragment of collagen XV. The nucleotide and amino acid sequences of endostatin derived from, for example, human and mouse collagen XV are set forth as SEQ ID NOS:7-10, respectively. The sequence identity to endostatin derived from human collagen type XVIII, for example, is about 62%, at the nucleotide level and about 58% at the amino acid level.

Sequence identities were performed as set forth below. Briefly, amino acid sequence alignments were performed using BLASTP version 2.0.8 (Jan-05-1999) and the following parameters: Matrix: 0 BLOSUM62; gap open: 11; gap extension: 1; x_dropoff: 50; expect: 10.0; wordsize: 3; filter: on. Endostatin: endostatin nucleic acid sequence alignments were performed using BLASTN version 2.0.6 (Sept-16-1998) and the following parameters: Match: 1; mismatch: -2; gap open: 5; gap extension: 2; x_dropoff : 50; expect: 10.0; wordsize: 11; filter: off. Type XV collagen endostatin : endostatin nucleic acid sequence alignments were performed as described above, with the exception that the mismatch and gap extension parameters were changed to -1 and 1, respectively.

When referring to amino acid sequences of various endostatin species and subclasses, amino acid positions exhibiting differences between species and subclasses are represented by reciting the alternative amino acids found at the referenced position. The symbol "/" is used to separate the alternative amino acid residues. For example, the tripeptide endostatin sequence DGK found at positions 104-106 of endostatin contains the alternative amino acids K or R in different species at the third position, corresponding to residue 106 of endostatin. Using the above described nomenclature, this position is recited as K/R and the tripeptide amino acid sequence is represented as DGK/R. Such species and subclass differences are understood to represent functionally equivalent amino acids at the particular referenced position.

Therefore, the term "endostatin peptide" or "endostatin fragment" is intended to mean a peptide or fragment encoded by a portion of the nucleotide sequence or having a portion of the amino acid sequence which exhibits substantial sequence identity and similarity to the endostatin sequences as described above and set forth as SEQ ID NOS: 1-10. For example, an endostatin peptide amino acid sequence is about 62% or greater in sequence identity to a portion of the human endostatin sequence shown as SEQ ID NO:2.

As used herein, the term "endostatin variant" is intended to mean endostatin containing at least one RGD amino acid sequence and having increased _v integrin binding activity or angiogenic inhibitory activity compared to authentic or native endostatin. The RGD amino acid sequence can be in any segment of the endostatin molecule so long as it is accessible for binding by an α_v-containing integrin and confers increased _v integrin binding activity or increased angiogenic inhibitory activity. The RGD amino acid sequence also can be at the N- or C- terminus of the endostatin variant. Similarly, the RGD sequence, whether located internally or at either or both termini, can be conformationally constrained, including for example, cyclic structures formed by disulfide bonds of cysteine residues which flank both sides of the RGD amino acid sequence. For example, an RGD sequence can be substituted for any of the endostatin receptor binding sequences RLQD, RAD or DGK/R ammo acid sequences found m native endostatin (SEQ ID NOS: 11-13, respectively) . Alternatively, the RGD sequence can be placed elsewhere in the molecule to confer increased α integrin binding and angiogenic inhibitory activity compared to endostatin or an endostatin peptide. A specific example of an endostatin variant is endostatin havmg the RAD sequence located at am o acid residues 62-64 (194-196 of the NCI) domain replaced by the ammo acid sequence RGD.

It is understood that minor modifications can be made without destroying the α integrin binding activity or antiangiogemc activity of the endostatin peptides, variants or fragments thereof of the invention and that only a portion of the primary structure may be required m order to effect activity. Such modifications are included within the meaning of the terms endostatin peptide, variant and fragment thereof so long as α_v integrin binding activity or angiogenesis inhibitory activity is retained. Further, various molecules can be attached to endostatin peptides, variants and fragments thereof, including for example, other proteins, carbohydrates, lipids and cytotoxic or cytostatic agents. Such modifications are included within the definition of the term.

As used herein, the term "peptide" is intended to mean two or more ammo acids covalently bonded together. A peptide of the invention includes polypeptides havmg a several hundred or more ammo acid residues. A peptide of the invention also includes fragments of the 184 ammo acid endostatin segment of the non-collagenous 1 (NCI) domain of collagen. Usually, the covalent bond between the two or more ammo acid residues is an amide bond. However, the amino acids can be joined together by various other means known to those skilled in the peptide and chemical arts. Therefore, the term "peptide" is intended to include molecules which contain, in whole or in part, non-amide linkages between amino acids, amino acid analogs and mimetics. Similarly, the term also includes cyclic peptides and other conformationally constrained structures.

As used herein, the term "amino acid" is intended to mean both naturally occurring and non-naturally occurring amino acids as well as amino acid analogs and mimetics. Naturally occurring amino acids include the 20 (L) -amino acids utilized during protein biosynthesis as well as others such as 4-hydroxyproline, hydroxylysine, desmosine, isodesmosine, homocysteine, citrulline and ornithine, for example. Non-naturally occurring amino acids include, for example, (D) -amino acids, norleucine, norvaline, p-fluorophenylalanine, ethionine and the like. Amino acid analogs include modified forms of naturally and non-natural ^π y occurring amino acids. Such modifications can incluαe, for example, substitution or replacement of chemical groups and moieties on the amino acid or by derivitization of the amino acid. Amino acid mimetics include, for example, organic structures which exhibit functionally similar properties such as charge and charge spacing characteristic of the reference amino acid. For example, an organic structure which mimics Arginine (Arg or R) would have a positive charge moiety located in similar molecular space and having the same degree of mobility as the ε-amino group of the side chain of the naturally occurring Arg amino acid. Mimetics also include constrained structures so as to maintain optimal spacing and charge interactions of the amino acid or of the amino acid functional groups. Those skilled in the art know or can determine what structures constitute functionally equivalent amino acid analogs and amino acid mimetics.

Specific examples of amino acid analogs and mimetics can be found described in, for example, Roberts and Vellaccio, The Peptides: Analysis, Synthesis, Biology, Eds. Gross and Meinhofer, Vol. 5, p. 341, Academic Press, Inc., New York, New York (1983), the entire volume of which is incorporated herein by reference. Other examples include peralkylated amino acids, particularly permethylated amino acids. See, for example, Combinatorial Chemistry, Eds. Wilson and Czarnik, Ch . 11, p. 235, John Wiley & Sons Inc., New York, New York (1997), the entire book of which is incorporated herein by reference. Yet other examples include amino acids whose amide portion (and, therefore, the amide backbone of the resulting peptide) has been replaced, for example, by a sugar ring, steroid, benzodiazepine or carbo cycle. See, for instance, Burger's Medicinal Chemistry and Drug Discovery, Ed. Manfred E. Wolff, Ch . 15, pp. 619-620, John Wiley & Sons Inc., New York, New York (1995), the entire book of which is incorporated herein by reference. Methods for synthesizing peptides, polypeptides, peptidomimetics and proteins are well known in the art (see, for example, U.S. Patent No. 5,420,109; M.

Bodanzsky, Principles of Peptide Synthesis (1st ed. & 2d rev. ed.), Springer-Veriag, New York, New York (1984 & 1993), see Chapter 7; Stewart and Young, Solid Phase Peptide Synthesis, (2d ed.), Pierce Chemical Co., Rockford, Illinois (1984), each of which is incorporated herein by reference) .

As used herein, the term "functional equivalent" when used in reference to a peptide having a specified ammo acid sequence is intended to mean a peptide having minor modifications of the specified ammo acid sequence but which exhibits at least about the same α_v integrin binding activity or angiogenesis inhibitory activity as the referenced peptide. Withm the biological arts, the term "about" when used in reference to a particular activity or measurement is intended to refer to the referenced activity or measurement as being withm a range values encompassing the referenced value and withm accepted standards of a credible assay withm the art, or withm accepted statistical variance of a credible assay withm the art.

Minor modifications of peptides having at least about the same α_v integrin binding activity or angiogenesis inhibitory activity as the referenced peptide include, for example, conservative substitutions of naturally occurring ammo acids and as well as structural alterations which incorporate non-naturally occurring ammo acids, ammo acid analogs and functional mimetics. For example, a Lysine (Lys) is considered to be a conservative substitution for the ammo acid Arg and as such the sequences KLQD, KAD, DGR and RGD are considered to be functional equivalents of RLQD, RAD, DGK and KGD, respectively. A further example of a functional equivalent of the specific sequence RGD is the tπpeptide RYD as it is found in the CDR region of an inhibitory antibody of an RGD binding integrin. Similarly, mimetic structures substituting the positive charged Arg or Lys ammo acids, with organic structures having similar charge and spacial arrangements, or the Ala residue in RAD with a non-ammo ac d linker similarly would be considered functional equivalents of the reference peptide so long as the peptide mimetic exhibits at least about the same activity as the referenced peptide. The nomenclature previously described for indicating species differences at an amino acid position is similarly used, in particular instances, to denote functionally equivalent amino acids at particular positions within the endostatin peptides and variants of the invention.

As used herein, the term "functional fragment" is intended to refer to a portion of endostatin, a portion of an endostatin peptide or a portion of an endostatin variant which retains at least about the same α_v integrin binding activity or angiogenesis inhibitory activity compared to endostatin or the parent endostatin peptide or variant. Such functional fragments can include, for example, the endostatin peptides containing the amino acid sequences RLQD, RAD, DGK/R and RGD.

As used herein, the term "substantially" or

"substantially the same" when used in reference to an amino acid sequence is intended to mean that the amino acid sequence shows a considerable degree, amount or extent of sequence identity when compared to the reference sequence. Such considerable degree, amount or extent of identity is further considered to be significant and meaningful and therefore exhibit characteristics which are definitively recognizable or known as being derived from or related to endostatin. For example, an amino acid sequence which is substantially the same amino acid sequence as an endostatin peptide or an endostatin variant, including fragments thereof, refers to a sequence which exhibits characteristics that are definitively known or recognizable as being sufficiently related to endostatin so as to fall within the classification of endostatin sequences as defined above. Minor modifications thereof are included so long as they are recognizable as an endostatin sequence as defined above. As used herein, the term "α_v integrin" is intended to mean the family of cell surface integrin adhesion receptors which contain the α_v submit complexed with a β subunit. Therefore, the term "α_v integrin" and "α_v-containing integrin" are used synomously herein. The nucleotide and deduced amino acid sequence of the human α_v subunit has been described in Suzuki et al., J. Cell Biol . 262: 14080-14085 (1987), which is incorporated herein by reference. The term includes the α_v-containing integrins α_vβ_lf α_vβ₃, α_vβ₅, _vβ₆ and α_vβ₈ as well as other α_v-containing integrins known to those skilled in the art and having the same or similar binding characteristics as those listed above. The term is similarly intended to include splice variants of the _v subunit and tissue specific- α_v containing integrins .

As used herein, the term "inhibiting" when used in reference to angiogenesis including vasculogenesis is intended to mean effecting a decrease in the extent, amount or rate of neovascularization. Effecting a decrease in the extent, amount or rate of endothelial cell proliferation or migration in a tissue is a specific example of inhibiting angiogenesis.

As used herein, the term "angiogenesis-dependent disease" is intended to mean a disease where the process of angiogenesis or vasculogenesis sustains or augments a pathological condition. Angiogenesis is the formation of new blood vessels from pre-existing capillaries or post-capillary venules. Vasculogenesis results from the formation of new blood vessels arising from angioblasts which are endothelial cell precursors. Both processes result in new blood vessel formation and are included within the meaning of the term angiogenesis-dependent diseases. Similarly, the term "angiogenesis" as used herein is intended to include de novo formation of vessels such as that arising from vasculogenesis as well as those arising from branching and sprouting of existing vessels, capillaries and venules.

Angiogenesis-dependent diseases include, for example, inflammatory disorders such as immune and non-immune inflammation, rheumatoid arthritis, chronic articular rheumatism and psoriasis; disorders associated with inappropriate or inopportune invasion of vessels such as diabetic retinopathy, neovascular glaucoma, retinopathy of prematurity, macular degeneration, corneal graft rejection, neovascular glaucoma, retrolental fibroplasia, rubeosis, capillary proliferation in atherosclerotic plaques and osteoporosis; and cancer associated disorders, including for example, solid tumors, tumor metastases, blood born tumors such as leukemias, angiofibromas, retrolental fibroplasia, Kaposi sarcoma, benign tumors such as hemangiomas, acoustic neuromas, neurofibromas, trachomas, and pyogenic granulomas, as well as other cancers which require neovascularization to support tumor growth. Additional examples of angiogenesis-dependent diseases include, for example, Osier-Webber Syndrome; myocardial angiogenesis; plaque neovascularization; telangiectasia; hemophiliac joints and wound granulation. Other diseases in which angiogenesis plays a role in the maintenance or progression of the pathological state are known to those skilled in the art and are similarly intended to be included within the meaning of the term as used herein.

As used herein, the term "treating" is intended to mean a reduction in severity, or prevention of an angiogenesis-dependent disease. Reduction in severity includes, for example, an arrest or a decrease in clinical symptoms, physiological indicators or biochemical markers. Prevention of the disease includes, for example, precluding the occurrence of the disease or restoring a diseased individual to their predisease state of health.

The invention provides an endostatin peptide having at least about 4-7 endostatin amino acid residues containing one or more of the amino acid sequences RLQD, RAD, DGK/R, or a functional equivalent thereof. The peptide can be used, for example, to effectively inhibit angiogenesis or treat an angiogenesis-dependent disease.

An endostatin peptide of the invention is a peptide consisting of a portion, or a part of the native endostatin which contains one or more of the amino acid sequences RLQD, RAD or DGK/R. These amino acid sequences correspond to residues 53-56, 63-65 and 104-106, respectively, of endostatin as it is shown in SEQ ID NOS:2, 4, and 6 and to residues 49-52, 59-61 and 100-102 in SEQ ID NOS : 8 and 10. Using the nomenclature of Hohenester et al . EMBO J. 17:1656-1664 (1998), which is incorporated herein by reference, the above amino acid sequences correspond to endostatin residues 184-187, 194-196 and 235-237, respectively. Endostatin peptides containing one or more of these amino acid sequences inhibit endothelial cell migration and proliferation as well as inhibit angiogenesis.

The endostatin peptides of the invention consist of at least about 4-7 endostatin amino acid residues . The endostatin residues include at least one of the recited amino acid sequences RLQD, RAD and DGK/R (SEQ ID NOS: 11-13, respectively) . The remainder of the endostatin residues of the peptide correspond to the endostatin residues flanking either or both sides of the above sequences as they are found in endostatin. These flanking endostatin residues provide, for example, additional binding specificity of the peptides for the endostatin receptor compared to the recited amino acid sequences RLQD, RAD and DGK/R alone. Alternatively, the remaining endostatin amino acid sequences can intervene between one or more of the residues within the RLQD, RAD or DGK/R sequence. For example, a RLQD, RAD or DGK/R sequence can be located non-contiguously within the primary endostatin peptide sequence but align sequentially within molecular space of the secondary or tertiary structure of an endostatin peptide. For example, an RLQD, RAD or DGK/R sequence can align on the same side of an α-helical or β-sheet structure and the intervening endostatin sequences corresponding to the remainder of the peptide will serve to configure these residues in close proximity or in an about sequential or linear order.

The invention is described below in reference to endostatin peptides having RLQD, RAD, or DGK/R residues located contiguously along the primary sequence. However, such descriptions and teachings are similarly applicable to endostatin peptides and variants having non-contiguous RLQD, RAD and DGK/R sequences within the primary sequence. For example, the RGAD₂₇_₃₀ sequence within the region

RGIRGADFQCFQQAR₂₄_₃₈ is one example of a non-contiguous RAD sequence of an endostatin peptide (SEQ ID NOS: 14 and 15, respectively) . The corresponding region in endostatin peptides derived from collagen type XV is RADFQCFQQAR₂₄_₃₄ (SEQ ID NO: 16). Another example of a non-contiguous RAD sequence is R₄₇ combined with the residues AD₂₉_₃₀ in the above sequence which are brought together in three-dimensional space within endostatin. Therefore, an endostatin peptide sequence which consists of endostatin residues 24-47 or 27-47 is an endostatin peptide having a non-contiguous RAD sequence (SEQ ID NOS: 17 and 18, respectively) . These endostatin peptides are also examples of a non-contiguous RGD sequence and therefore are described further below in reference to endostatin variants of the invention. Those skilled in the art will know how to make and use endostatin peptides and variants having RLQD, RAD and DGK/R residues proximately located in three dimensional space given the teaching and guidance below in reference to endostatin peptides having RLQD, RAD and DGK/R sequences as contiguous residues.

For example, a 4 amino acid endostatin peptide containing the tetrapeptide RLQD located at positions 53-56 is the tetrapeptide RLQD (RLQD₅₃_₅₅, where the subscripts indicate the amino acid position of the terminal amino acids recited in the sequence) . However, a penta- hexa- or heptapeptide containing the RLQD sequence will contain, respectively, one, two or three endostatin residues flanking this sequence. The flanking sequences can be derived from one side of the RLQD₅₃_₅₆ sequence as it is found in endostatin, or alternatively, the flanking sequences can be derived from a combination of residues on both sides of this sequence. For the specific example of a heptapeptide, the three flanking endostatin residues can be derived from either the amino-terminal (N-terminal) or carboxyl-terminal (C-terminal) ends yielding the heptapeptides LSSRLQD₅₀_₅₆ or RLQDLYS₅₃_₅₉, respectively. Alternatively, a RLQD heptapeptide can be derived from any combination of three residues flanking endostatin positions R₅₃ and D₅₆, such as for example, one N-terminal and two C-terminal endostatin residues which surround the RLQD₅₃_₅₆ sequence. RLQD penta- and hexa- peptides can similarly derive the endostatin residues from either side of the RLQD₅₃_₅₆ sequence and the hexapeptide can derive its endostatin residues from a combination of both N- and C-terminal sides. Therefore, the range of possible endostatin amino acid residue positions for a penta- hexa- or heptapeptide will correspond to endostatin amino acid residues S₅₂-L₅₇, S₅₁-Y₅₈ and L₅₀-S₅₉, respectively, for an RLQD endostatin peptide of about 5-7 endostatin amino acid residues .

Similarly, a 4 amino acid endostatin peptide containing the tripeptide RAD located at position 63-65 (RAD₆₃_₆₅) will include one endostatin amino acid residue on either the N- or C-terminal side of this sequence. A penta- hexa- or heptapeptide containing this sequence will contain, respectively, two, three or four endostatin residues flanking the RAD₆₃.₆₅ sequence in any combination as described above for a RLQD endostatin peptide. Therefore, the range of possible endostatin amino acid residue positions for a tetra- penta- hexa- or heptapeptide will correspond to endostatin amino acid residues R₆₂-R₆₆, V₆₁-A₆₇, I₆₀-A₆₈ and S₅₉-V₆₉, respectively, for a RAD endostatin peptide of about 4-7 endostatin amino acid residues .

As with the above described endostatin peptides, a 4 amino acid endostatin peptide containing the tripeptide DGK/R located at position 104-106 (DGK/R₁₀₄__ιn6) also will include one endostatin amino acid residue on either the N- or C-terminal side of this sequence. A penta- hexa- or heptapeptide containing this sequence will contain, respectively, two, three or four endostation residues flanking the DGK/R₁₀₄_₁₀₆ sequence in any combination as described above for a RLQD or a RAD endostatin peptide. Therefore, the range of possible endostatin amino acid residue positions for a tetra- penta- hexa- or heptapeptide will correspond to endostatin amino acid residues F₁₀₃-D₁₀₇, S₁₀₂-V₁₀₈, F₁₀₁-L₁₀₉ and I₁₀₀-^R _ιιc respectively, for a DGK/R endostatin peptide of about 4-7 endostatin amino acid residues.

Specific examples of endostatin peptides of at least about 4-7 endostatin amino acid residues containing the amino acid sequence RLQD, RAD or DGK/R (SEQ ID NOS: 11-13, respectively) are set forth below in Table 1 (SEQ ID N0S:11 and 25-61). For example, the endostatin peptide species shown in Table 1 include each of the core endostatin peptide sequences RLQD, RAD and DGK/R in context of their corresponding flanking endostatin amino acid sequences. The flanking amino acid sequences encompass the range of positions set forth previously for each of the tetra- penta- hexa- and heptapeptides of each of the core endostatin sequence. Column 1 of Table 1 recites the number of endostatin amino acid residues. Columns 2-4 list the endostatin peptide species for each tetra- penta- hexa- and heptapeptide containing the core endostatin sequences RLQD, RAD and DGK/R (SEQ ID NOS: 11 and 25-33; 34-47 and 48-61, respectively) .

For reference, the recited amino acid sequences correspond to the human endostatin peptide sequences.

Although the human endostatin peptides only differ at two positions compared to mouse or chicken in the recited amino acid sequences of Table 1, namely, residues 67 and 68, flanking the C-terminal end of RAD₆₃_₆₅, it is understood that the corresponding amino acids from these non-human species are similarly included as examples of specific endostatin peptides of the invention. Similarly, the corresponding peptide sequences of collagen type XV derived endostatin differ in two positions flanking each of the RLQD, RAD and DGK/R sequences and are also understood to be included as examples of specific endostatin peptide species of the invention. Given the teachings and guidance described herein, one skilled in the art can compare the sequences of the various endostatin species and subtypes recited as SEQ ID NOS:2, 4, 6, 8, and 10 and understand that the corresponding peptide sequences can be substitutes for those shown in Table 1 as endostatin peptides of the invention.

Moreover, and as described further below, one skilled in the art can similarly confirm that such peptides bind α_v integrins, inhibit endothelial cell migration or proliferation as well as inhibit angiogenesis.

Table 1 : Endostatin Peptide Species

Amino Acids RLQD RAD DGK/R

RLQD RRAD FDGK

RADR DGKD

SRLQD VRRAD SFDGK RLQDL RRADR FDGKD

RADRA DGKDV

SSRLQD IVRRAD FSFDGK SRLQDL VRRADR SFDGKD RLQDLY RRADRA FDGKDV

RADRAA DGKDVL

LSSRLQD SIVRRAD IFSFDGK SSRLQDL IVRRADR FSFDGKD SRLQDLY VRRADRA SFDGKDV RLQDLYS RRADRAA FDGKDVL

RADRAAV DGKDVLR The endostatin peptides of the invention can similarly consist of at least about 4-7 endostatin residues which include functional equivalents of the recited amino acid sequences RLQD, RAD and DGK/R. Functional equivalent endostatin peptides exhibit changes or modifications of the peptide sequence or structure without substantially altering the α_v integrin binding activity or biological activity of the peptide. As described previously, functional equivalents of these sequences can include, for example, conservative amino acid substitutions such as the substitution of amino acids having side chains with similar charge, polarity, hybrophobicity and aromaticity. Similarly, functional equivalent structures of the endostatin peptides can also include the substitution of non-naturally occurring amino acids, amino acid analogs and mimetics having similar size, charge and spacial arrangements of functional groups compared to the original endostatin amino acid. Functional equivalents of the endostatin peptides of the invention similarly includes conservative amino acid substitutions and substitution of non-naturally occurring amino acids, amino acid analogs and mimetic structures of the endostatin residues flanking the RLQD, RAD and DGK/R sequences. Given the sequences and teachings herein, those skilled in the art will know, or can determine, which substitutions, replacements and structural modifications constitute a functional equivalent of an endostation peptide, without substantially altering the α_v integrin binding or biological activity of the peptide.

The size of an endostatin peptide of the invention can vary between about 4-7 amino acids and about 182 amino acids so long as it contains at least about 4-7 endostatin amino acid residues containing substantially the amino acid sequence RLQD, RAD or DGK/R, or functional equivalent thereof, and retains α_v integrin binding activity. The endostatin peptides can include endostatin sequences adjacent to the residues flanking the RLQD, RAD or DGK/R sequences including, for example, substantial portions of endostatin so long as the peptide is less than the entire endostatin molecule and maintains α_v integrin binding activity. For example, the circulating form of endostatin isolated from hemofiltrate lacks 12 N-terminal residues and the C-terminal lysine compared to intact endostatin and as such is less than the entire endostatin molecule. However, this truncated endostatin species lacks endothelial cell proliferation inhibitory activity and therefore, is not encompassed by an endostatin peptide of the invention. The inclusion of additional endostatin residues can, for example, impart further „ integrin binding activity or specificity compared to smaller endostatin peptides of the invention as well as confer functions other than integrin binding.

An endostatin peptide of the invention exhibits α_v integrin binding activity. The α_v integrin binding activity includes selective binding activity toward the α_vβ₃ integrin. This integrin is a predominant cell adhesion receptor found on proliferating endothelial cells. In addition to the _vβ₃ integrin binding, an endostatin peptide also exhibits binding activity toward the less predominant α_vβ₅ and _vβ_x integrins present on the surface of endothelial cells.

An endostatin peptide of the invention also exhibits selective binding activity toward non- _v containing integrins which exhibit similar binding specificity as α_v integrins . For example, a subset of the integrin family of adhesion receptors recognize the tripeptide cell adhesion sequence RGD found in a number of adhesion proteins, including for example, extracellular matrix proteins. As such, this subset of integrin adhesion receptors exhibit similar binding specificity due to their common recognition of RGD containing amino acid sequences . Integrins within this subset include the α_v containing integrin family as well as the integrins α_ιIbβ₃ and α₅β_!. An endostatin peptide of the invention therefore exhibits selective binding activity toward α_IIbβ₃ and ₅β_x, and the binding activities of endostatin peptides described herein in reference to α_v integrins are also exhibited toward these integrins. A specific example of an integrin which does not exhibit binding to RGD-containing sequences and therefore does not exhibit similar binding specifically to α_v integrins is the integrin α₁β_1.

Binding activity is conferred by one or more of the RLQD, RAD or DGK/R endostatin amino acid sequences located either contiguously along the primary sequence or non-contiguously but having the above sequences or functional equivalents thereof in close spacial proximity to each other within the secondary or tert_ ^; ry structure of an endostatin peptide. Moreover, differential specificity toward a particular α_v integrin can be influenced or imparted by the flanking endostatin residues which surround the RLQD, RAD and DGK/R sequences. Similarly, modification of these flanking residues, inclusion of additional endostation residues or inclusion of a heterologous domain or sequence can also impart differential specificity of an endostatin peptide for a particular α_v integrin. As such, endostatin peptides can be used to inhibit the binding of one or more _v integrins to their ligands. As stated previously, the inclusion of endostatin residues additional to the RLQD, RAD and DGK/R and their at least about 1-3 flanking residues can confer functions other than integrin binding onto the endostatin peptides of the invention. For example, inclusion of endostatin sequences which span two or more of the RLQD, RAD or DGK/R sequences will, in addition to including two or more α_v integrin binding sequences, also include the intervening sequences which augment the binding energy and stability of an endostatin peptide. Additionally, endostatin exhibits binding activity other than α_v integrin binding activity such as glycosaminoglycan binding activity. Glycosaminoglycan binding domains within endostatin also can be included in the endostatin peptides of the invention. For example, heparin binding domains can be found at amino acid positions R₂₄, R₂₇, R₅₃ and R₁₃₉. Other glycosaminoglycan binding activities exhibited by endostatin or known in the art can be conferred onto the endostatin peptides of the invention. Such glycosaminoglycan binding activities can include, for example, chondroitin sulphate A, chondroitin sulphate B, chondroitin sulphate C, heparin, heparin sulphate, or keratan sulphate binding activity. The inclusion of glycosaminoglycan binding activity with the endostatin peptides of the invention can provide therapeutic advantageous because this binding activity can function to localize the endostatin peptides to glycosaminoglycan rich matrix surrounding angiogenic dependent diseases or to compete for binding of growth factors that bind to glycosaminoglycan which is required for their biological activity.

Further, other functions can be imparted onto the endostatin peptides of the invention by using recombinant or chemical synthesis methods known in the art. The particular heterologous function will depend on the desired use of the peptide. For example, targeting domains, catalytic domains, cytokine, hormone and growth factor domains as well as cytotoxic and cytostatic agents can be incorporated into the endostatin peptides of the invention to produce bifunctional, or multi-functional, molecules exhibiting the function of any of these domains together with α_v integrin binding activity. Additionally, endothelial cell proliferation inhibitory domains can also be incorporated into the endostatin peptides of the invention so as to impart α_v integrin-independent inhibitory activity onto the peptide and thereby increase the anti-angiogenic activity of the endostatin peptide. For example, vascular endothelial growth factor (VEGF) is a major inducer of angiogenesis in normal and pathological conditions, and is essential in embryonic vasculogenesis. The biological effects of VEGF include stimulation of endothelial cell proliferation, survival, migration and tube formation, and regulation of vascular permeability. Incorporation into an endostatin peptide a receptor antagonist domain to one or more of the VEGF receptors, or other receptor which transduces signals through the VEGF signaling pathway, will result in increased anti-angiogenic activity due to the bifunctional activities exhibited by the endostatin peptide. Those skilled in the art will know which heterologous functions are desirable for a particular application and how to make a particular bi- or multi-functional endostatin peptide using methods well known in the art.

An endostatin peptide can include portions of endostatin sequences that are non-adjacent with the flanking residues so as to construct, for example, a segmented peptide containing desired portions of endostatin in a single peptide. A specific example of an endostatin peptide containing non-adjacent endostatin sequences is a peptide containing two or more or the amino acid sequences RLQD, RAD or DGK/R linked together. The linking sequences can be, for example, a flexible peptide linker or altered intervening endostatin sequences. If desired, such linkers can be made to preserve the distance between the core RLQD, RAD or DGK/R peptide sequences as they are found in the native molecule.

Therefore, the invention provides an endostatin peptide that includes two or more of the amino acid sequences RLQD, RAD, DGK/R or a functional equivalent thereof. An endostatin peptide exhibits α_v integrin binding activity, including the α_v-containing integrins α_vβ₃, _vβ₅ and a_v^>₁ . Also provided is an endostatin peptide having glycosaminoglycan binding activity. The glycosaminoglycan binding activity can be chondroitin sulphate A, chondroitin sulphate B, chondroitin sulphate C, heparin, heparin sulphate, or keratan sulphate binding activity. Additionally, the invention provides multi- functional endostatin peptides where a heterologous domain or sequence is fused to an α_v integrin binding endostatin peptide .

The invention also provides an endostatin peptide consisting of substantially the amino acid sequence shown as SEQ ID NO: 20, or a functional fragment thereof .

As will be described further below, there are various methods known in the art which can be employed to obtain functional endostatin peptides having at least about 4-7 endostatin amino acid residues containing substantially the amino acid sequence RLQD, RAD, DGK/R, or functional equivalents thereof. With the identification WO 00/67771 PCT/USOO/l 2063

31 of an endostatin receptor as being an α_v integrin, one method for identifying a functional endostatin peptide is, for example, to fragment endostatin and determine the binding activity of the fragments toward one or more of the α_v integrins α_vβ₃, _vβ_x and _vβ₅. Alternatively, the fragments can be tested for biological function by determining the fragments inhibitory activity against endothelial cell proliferation, migration or against angiogenesis. Chemical and protease fragmentation of endostatin has identified endostatin peptides that maintain α_v integrin binding activity.

For example, cleavage of endostatin with cyanogen bromide (CnBr) produces at least four endostatin fragments identified as SEQ ID NOS: 19-22 in the mouse sequence. The endostatin fragment R₂₄-M₁₃₀ in the mouse sequence contains all three of the core sequences RLQD, RAD and DGK/R (SEQ ID NO: 20) The corresponding fragment in the human sequence is R₂₄-M₁₇₈ and consists of the combined mouse fragments R₂₄-M₁₃₀ and E₁₃₁-M_17B (SEQ ID NOS: 20 and 21, respectively) because the human sequence is substituted by a T residue at position 130. Immbolized CnBr cleavage products of human endostatin promote endothelial cell attachment similar to full length endostatin. Therefore, at least the CnBr fragment shown as SEQ ID NO:20 is an endostatin peptide of the invention. For example, SEQ ID NO:20 contains at least about 4-7 endostatin amino acid residues containing substantially the amino acid sequence RLQD, RAD or DGK/R and exhibits α_v integrin binding activity. Other endostatin peptides include the various CnBr fragments resulting from partial digestion and containing the CnBr fragment shown as SEQ ID NO: 20. WO 00/67771 PCT/USOO/l 2063

32

Cleavage of endostatin with the proteases trypsin and chymotrypsin also results in active endostatin peptides which are fragments of intact endostatin. Trypsin cleaves after the amino acid residues arginine and 5 lysine whereas chymotrypsin cleaves after the residues phenylalanine, tyrosine and tryptophan. Endostatin fragments resulting from complete digestion with trypsin are shown below in Table 2 (SEQ ID NOS: 62-79) whereas fragments resulting from complete digestion with

10 chymotrypsin are shown below in Table 3 (SEQ ID

NOS:80-97) . As with the CnBr cleavage described above, protease digestion with trypsin or chymotrypsin resulted in fragments which inhibited endothelial cell proliferation, migration and α_v integrin binding. Those

15 endostatin fragments resulting from complete or partial protease digestion which exhibit α integrin binding activity and contain at least about 4-7 endostatin amino acid residues containing one or more of the amino acid sequences RLQD, RAD or DGK/R are endostatin peptides of

20 the invention.

Table 2: Endostatin Fragments Resulting From Trypsin Digestion

H₁ - -R₄ D₇₆- -K₉₅

D₅- -R₂₄ P₉₆ R₉₉

Z . G G₂₅5^_ K-2₇7 l i oo ^106

^*peptides are denoted by N- and C-terminal residues and corresponding a ino acid position. Table 3: Endostatin Fragments Resulting From Chymotrypsin Digestion

"ι F₆ S₅₉ F₈₀ Pus W₁₂₀

Qv — F₃₁ Pβι ~W₈₃ H₁₂₁ — Y₁₃₄ Q₃₂~ ^_ F₃₄ E₈₄ - - F₈₇ C₁₃₅- -W₁₃₈

Q35 F ₆ S _{8 8} F₁₀₁ Rl 39 * 169

^47 ^ 9 S l02 F₁₀₃ J- llfl t 178

L₅o^{- _}Y₅₈ D₁₀₄ - -W₁₁₄ M_{17 9}- -K₁₈₃

'peptides denoted by N- and C-terminal residues and corresponding amino acid position .

As with the endostatin peptides described previously, endostatin peptides having α_v integrin binding activity and resulting from endostatin cleavage with CnBr, trypsin, chymotrypsin or other methods known in the art similarly can be of various lengths so long as they are less than the entire endostatin molecule and contain at least about 4-7 endostatin amino acid residues containing substantially the amino acid sequence RLQD, RAD, DGK/R or a functional equivalent thereof. For example, such endostatin peptides can include endostatin or non-endostatin derived sequences which confer glycosaminoglycan binding or other sequences which augment the α_v integrin binding activity of the peptide or confer additional, and different, activities onto the endostatin peptide.

The invention further provides an endostatin variant. The endostatin variant consists of endostatin containing the amino acid sequence RGD, or a functional fragment thereof. Also provided is an endostatin variant containing substantially the amino acid sequence as shown as SEQ ID NO:23, or functional fragment thereof, as well as an endostatin variant containing the amino acid sequence RRGDR. A specific example of an endostatin peptide variant having this sequence is recited in SEQ ID NO:24 as K/RRGDR. An endostatin variant provides for increase angiogenic inhibitory activity or increased α_v integrin binding activity relative to native endostatin. Therefore, an endostatin variant can be advantageously used to inhibit angiogenesis or treat an angiogenesis-dependent disease.

An endostatin peptide exhibits α_v integrin binding activity. Natural ligands of α_v-containing integrins include extracellular matrix proteins containing the cell adhesion recognition sequence RGD. Inclusion of an RGD sequence into or attached to an endostatin peptide will augment the _v integrin binding activity of the endostatin peptide, or alternatively, inclusion of an RGD sequence into or attached to an endostatin amino acid molecule devoid of the sequences RLQD, RAD or DGK/R will confer α_v integrin binding activity onto an otherwise inactive peptide. Moreover, the RGD sequence contained in an endostatin variant can include additional residues which further increase the α_v binding activity of the endostatin variant. A specific example of such modifications would be the inclusion of flaking arginine residues to yield the sequence RRGDR. Therefore, endostatin peptide variants can be produced by incorporating an RGD amino acid sequence into an endostatin peptide, or into a non-α_v integrin binding endostatin molecule so as to augment or confer α_v integrin binding activity onto the peptide or molecule.

An endostatin variant will exhibit increased α_v integrin binding activity compared to endostatin or compared to the corresponding non-α_v integrin binding endostatin molecule. Depending, for example, on the location and number of RGD sequences, an endostatin variant can exhibit between about 3-fold and greater than 100-fold or more of an increase in α_v integrin binding activity. Generally, the increase in _v integrin binding activity is about 3-fold or more, preferably about 5-fold or more, and more preferably about 10-fold or more. The endostatin variant shown as SEQ ID NO: 23 exhibits about 100-fold or more _v integrin binding activity compared to endostatin. The increase m binding activity includes increased activity exhibited collectively against α integrins as well as specific increases against any of the α_vβ₃, α.β^ or _vβ_x integrins, alone or in combination.

One specific example of an endostatin variant is the endostatin peptide shown as SEQ ID NO:23. This variant has the RAD₆₃_₆₅ replaced with the sequence RGD.

The endostatin flanking residues result in the ammo acid sequence RRGDR₅₂_₆₅ at this location. As described in Example IV, the activity of this variant in inhibiting α_v integrin binding is increased about 100-fold or more compared to endostatin.

Endostatin variants can have an RGD sequence substituted for one or more of the endostatin sequences RLQD, RAD and DGK/R as described above. Alternatively, an endostatin variant can have an RGD sequence located in other regions of endostatin or m an endostatin peptide so long as it is accessible by an α_v integrin. The structure of endostatin is known and has been described in Ding et al. Proc . Na tl . Acad. Sci . USA 95: 10443-10448 (1998) and in Hohenester et al . , supra . The crystal structure shows that endostatin has numerous regions which are exposed to solvent. For example, one helical region is between ammo acids 22-40 m the mouse sequence. The RLQD region, the RRADR region, and the WRT region which is located at residues 138-140 also are all exposed to solvent. Such regions are appropriate locations where an RGD sequence can be placed for it to be accessible and, therefore, capable of being bound by an _v integrin. It should be noted that consideration of the three dimensional structure is important only to the extent that the endostatin molecule to be modified, or that the endostatin peptide to be modified, is sufficient in size so as to fold into its corresponding domain structure. Those skilled in the art will know, or can determine by making and testing an endostatin molecule containing RGD, if the RGD sequence binds to an α_v integrin and is therefore an endostatin variant of the invention. Exemplary methods known to those skilled in the art for making and testing such variants are described further below.

The α-helical region described at amino acid residues 22-40 contains the endostatin peptide sequence RGIRGADFQCFQQAR₂₄_₃₈ (SEQ ID NO: 15). As described previously, the RGAD₂₇_₃₀ sequence which is contained within this region is one example of a non-contiguous RGD sequence (SEQ ID NO: 14) . The crystal structure of endostatin for this region shows that the two phenylalanines are exposed to solvent, indicating, for example, a contact surface for protein-protein interactions. Within this helical structure the RGD residues of this sequence are exposed on the same side of the helix. Similarly, residue R₄₇ is proximately located in three-dimentional space and combined with G₂₈ and D₃₀ of this region constitutes a non-contiguous RGD sequence. Therefore, the amino acid sequence

RGADFQCFQQARAVGLAGTFR₂₇_₄₇ or RGIRGADFQCFQQARAVGLAGTFR₂₄_₄₇ are further examples of an endostatin variant having a non-contiguous RGD sequence (SEQ ID NOS: 18 and 17, respectively) .

In addition to the above solvent exposed regions of endostatin, there can be other regions which are cryptic or latent that will allow the incorporation of an RGD sequence without substantially affecting the structure or function of an endostatin molecule or of an endostatin peptide. Additionally, an endostatin variant can be of various lengths, including the entire endostatin molecule, so long as it is an endostatin containing an RGD sequence and binds an _v integrin. Functional fragments of endostatin containing an RGD sequence similarly constitute an endostatin variant so long as α_v integrin binding activity is maintained. As with the endostatin peptides of the invention, an endostatin variant similarly can include endostatin or non-endostatin derived sequences which confer glycosaminoglycan binding or other sequences which augment the α_v integrin binding activity of the peptide or confer additional, and different, activities onto the endostatin peptide. Given the teachings and guidance described herein, those skilled in the art know, or can determine, a peptide length and sequence, including contiguous, non-contiguous and heterologous residues and domains fused to the endostatin variant which constitutes an endostatin variant of the invention.

An endostatin peptide or an endostatin variant of the invention can be prepared or obtained by methods known in the art. For example, an endostatin peptide of the invention can be produced by enzymatic or chemical cleavage of endostatin or of Types XV and XVIII collagen. Methods for enzymatic and chemical cleavage, and for purification of the resultant protein fragments are well known in the art (see, for example, Deutscher, Methods in Enzvmology, Vol. 182, "Guide to Protein Purification," San Diego: Academic Press, Inc. (1990), which is incorporated herein by reference) . One method for purifying an endostatin peptide is described further below in Example I. The choice of a particular method for purifying an endostatin peptide will depend on the particular application of the peptide by the user. For example, higher levels of purity may be desired for in vivo applications compared to in vi tro procedures. Endostatin peptides or variants, including those containing non-contiguous sequences can be chemically synthesized. The synthesized peptides can be additionally synthesized in conformationally constrained structures to facilitate or augment one or more active conformations of the molecule.

An endostatin peptide of the invention can also be recombinantly expressed by appropriate host cells, including bacteria, yeast, avian, insect and mammalian cells, using methods known in the art. Methods for recombinant expression and purification of peptides in various host organisms are described, for example, in Sambrook et al . , Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, New York (1992) and in Ansubel et al . , Current Protocols in Molecular Biology, John Wiley and Sons, Baltimore, MD (1989), both of which are incorporated herein by reference. Similarly, endostatin variants can be generated using recombinant mutagenesis, such as site directed mutagenesis and PCR mutagenesis, and expression of the RGD containing endostatin molecule. Numerous methods for constructing, modifying, expression and purification are known to those skilled in the art. A specific example includes the tagged fusion protein expression system described below in the Examples. The choice of recombinant methods, expression and purification systems will be known by those skilled in the art and will depend on the user and the particular application for the endostatin peptide or variant .

The ability of an endostatin peptide or variant to bind an α_v integrin can be determined using methods known in the art. The field of cell adhesion and, in particular, integrin mediated cell adhesion is well known. Methods for determining integrin mediated ligand binding are therefore also well known and include methods ranging from determination of integrin-mediated ligand binding to the identification of a specific integrin-ligand interactions as well as accurate measurements of integrin-ligand affinities. Such methods include, for example, both in vi tro and in vivo procedures. For a review of integrins, integrin function and methods for determining binding interactions and integrin-mediated functions, see for example, Guan, ed., Signaling Through Cell Adhesion Molecules, CRC Press, Boca Raton, FL (1999); Howlett, ed., Integrin Protocols, Humana Press, Totowa, NJ (1999), which is incorporated herein by reference.

Briefly, integrin binding activity of an endostatin peptide or variant can, in general, be determined using assays which interfere with integrins as a class of cell surface receptors. Such methods include, for example, disruption of integrin binding function through divalent cation chelation and inhibition with extracellular matrix proteins or peptides containing the cell recognition sequence RGD. Specific binding interactions of endostatin peptides and variants can be determined using well known methods in the art such as competitive inhibition using reagents specific to a particular integrin or a panel of reagents specific to different integrins. Reagents can be, for example, monoclonal antibodies and mtegrm selective peptides. Inhibition of mtegrm binding can be determined using a variety of different formats, including, for example, the inhibition of mtegrm function such as cell attachment, migration, proliferation and m the specific example of endothelial cells, tubulogenesis and angiogenesis. Such methods are well known in the art and are described in, for example, Vuoπ, K. and Ruoslahti, E. Science 266:1576-1578 (1994); Ruoslahti et al . Meth Enzymol. 82:803 (1982); Pierschbacher & Ruoslahti PNAS 81:5985 (1984); Bauer et al . J Cell Biol 116:477 (1992); Meredith et al. Mol Biol Cell 4:953 (1993); Zhang et al . PNAS 92:6161 (1995); Wary et al . Cell 87:733 (1996); Gamble et al. J Cell Biol 121:931 (1993); Brooks et al . Cell

79:1157-1164 (1994); Brooks et al. J Clm Invest 96:1815 (1995); and U.S. Patent No. 5,753,230, all of which are incorporated herein by reference. These methods are exemplified further below in reference to α_v mtegrm specificity of endostatin peptides and variants of the invention .

Additionally, binding interactions of an endostatin peptide or variant can be determined by directly measuring the binding activity of such peptides or variants for a particular mtegrm. Such assays include, for example, ELISA based assays using purified mtegrm or using a capture reagent such as an mtegrm specific antibody to specifically attach a particular mtegrm to a solid support. An opposite format can additionally be used where endostatin peptides are absorbed or attached to a solid phase and the binding of integrins is determined m, for example, a purified receptor assay or m biological assays such as cell attachment, migration and survival. Fluorescent activated cell sorting using labeled endostatin peptides and variants also can be used for determining integrin binding specificity. Such methods are well known in the art and are described, for example, in Koivunen et al . J. Biol . Chem . 268:20205-20210 (1993); Barbas et al . Proc . Na tl . Acad. Sci . USA 90:10003-10007 (1993); Pytela et al. Cell 40:191-198 (1985) and Pytela et al . Methods Enzymol . 144:475-489 (1987), all of which are incorporated herein by reference. Moreover, these methods are exemplified further below in reference to α_v integrin specificity of endostatin peptides and variants of the invention.

Therefore, endostatin peptides, endostatin variants, functional fragments and functional equivalents thereof can be made and tested for α_v integrin binding specificity using any of the methods describe above as well as others known to those skilled in the art. Briefly, a peptide or variant derived from endostatin can be produced by recombinant expression or chemical synthesis and tested for α_v integrin binding function by, for example, determining its ability to inhibit cell adhesion. Specificity of the peptide or variant can be assessed using, for example, inhibitory antibodies to α_vβ₃, α_vβ₅ and/or α_vβ_:. Similarly, inhibition of the peptide or variant to bind to any of these _v-containing integrins when attached to a solid support can also be used to assess specificity for an α_v integrin. Those antibodies that block the ability of the peptide or variant to inhibit cell adhesion or specific binding to an _v-containing integrin indicates that the peptide or variant specifically binds an _v integrin. As such the peptide or variant is therefore an endostation peptide or an endostatin variant of the invention. Using such methods, those skilled in the art can also make and test modifications of a particular endostatin peptide, endostatin variant or populations thereof without experimentation. Modifications can be made and tested either systematically or randomly as in screening a library population of modified forms of endostatin fragments. Those modified forms that specifically bind an α_v integrin are considered to be an endostatin peptide of the invention.

In another embodiment, an endostatin peptide or an endostatin variant can have, for example, glycosaminoglycan binding activity. The glycosaminoglycan binding activity can be, for example, chondroitin sulphate A, chondroitin sulphate B, chondroitin sulphate C, heparin, heparan sulphate or keratan sulphate binding activity. Similarly to the assays described above for measuring α_v integrin binding specificity, the glycosaminoglycan binding activity of an endostatin peptide or variant can be determined using methods known in the art .

The invention provides a method of inhibiting angiogenesis by administering an angiogenic inhibitory amount of an endostatin peptide having at least about 4-7 endostatin amino acid residues containing one or more of the amino acid sequences RLQD, RAD, DGK/R, or a functional equivalent thereof. The invention further provides a method of treating an angiogenesis-dependent disease by administering an angiogenic inhibitory amount of an endostatin peptide having at least about 4-7 endostatin amino acid residues containing one or more of the amino acid sequences RLQD, RAD, DGK/R, or a functional equivalent thereof.

The invention also provides a method of inhibiting angiogenesis by administering an angiogenic inhibitory amount of an endostatin variant having the amino acid sequence RGD, or a functional fragment thereof. The invention further provides a method of treating an angiogenesis-dependent disease in a human by administering an angiogenic inhibitory amount of an endostatin variant having the amino acid sequence RGD, or a functional fragment thereof.

Angiogenesis, including vasculogenesis, is an important physiological process, without which embryonic development and wound healing, for example, would not occur. However, angiogenesis is also inappropriately recruited into numerous pathological conditions as a means to provide adequate blood and nutrient supply to the cells within the affected tissue. Many of these pathological conditions involve aberrant cell proliferation or regulation. Such conditions in which angiogenesis is believed to be important are referred to herein as angiogenisis-dependent diseases. However, the methods of the invention also can be used beneficially to inhibit angiogenesis associated with normal physiological processes. For example, the inhibition of angiogenesis associated with the menstrual cycle can be prophylactically used as an effective method of birth control. Therefore, the descriptions below in reference to the treatment of angiogenisis-dependent diseases are also applicable to the inhibition of normal angiogenic responses where there is a prophylactic or therapeutic need or benefit.

Angiogenesis-dependent diseases include, but are not limited to, inflammatory disorders such as immune and non-immune inflammation, chronic articular rheumatism, rheumatoid arthritis, and psoriasis; disorders associated with inappropriate or inopportune invasion of vessels such as diabetic retinopathy, neovascular glaucoma, capillary proliferation in atherosclerotic plaques and osteoporosis; ocular angiogenic diseases, for example, retinopathy of prematurity, macular degeneration, corneal graft rejection, neovascular glaucoma, retrolental fibroplasia, rubeosis; Osier-Webber Syndrome; myocardial angiogenesis; plaque neovascularization; telangiectasia; hemophiliac joints; angiofibroma; wound granulation and cancer associated disorders. Diseases other than those listed above are known to those skilled in the art as requiring angiogenesis for growth, maintenance or survival and as such, are included herein as an angiogenesis-dependent diseases .

Angiogenic-dependent cancers include, but are not limited to, a variety of cancers and neoplastic conditions such as solid tumors, blood born tumors such as leukemias, and tumor metastases; benign tumors, for example hemangiomas, acoustic neuromas, neurofibromas, trachomas, and pyogenic granulomas . Additionally, angiofibromas, retrolental fibroplasia, Kaposi sarcoma and the like are exemplarily cancers which require neovascularization to support tumor growth. Cancers other than those listed above are also known by those skilled in the art as requiring angiogenesis to support tumor growth and are similarly referred to herein as angiogenesis-dependent diseases.

Endostatin peptides and variants of the invention can be selected and prepared by the methods described above as well as other methods known in the art and screened by a variety of in vi tro or in vivo methods known in the art to determine their inhibitory activity against angiogenesis. Endostatin peptides or endostatin variants which inhibit the binding of α_v-containing integrins to their ligands, as determined by, for example, the binding assays or functional assays described previously, will be active in inhibiting angiogenesis. Similarly, antiangiogenic endostatin peptides or endostatin variants are similarly active in treating an angiogenic-dependent disease.

Biological assays known in the art also can be used to measure the effects of the endostatin peptides and variant of the invention on angiogenesis. Such biological assays include, for example, a bovine capillary endothelial cell proliferation assay. Briefly, bovine capillary endothelial cells stimulated with bFGF can be used to determine the efficacy of an endostatin peptide or variant against angiogenesis. The cells are cultured in the presence or absence of an endostatin peptide or variant. The extent of proliferation is measured following an about 72 hour culture period to determine the effect of the endostatin peptide or variant on cell growth and therefore, on angiogenesis. The extent of endothelial cell proliferation compared to control peptide treatments inversely correlates with the activity and/or efficacy of the endostatin peptide or variant. The bovine capillary endothelial cell proliferation assay is well known in the art and is described in, for example, PCT publication WO 97/15666, which is incorporated herein by reference.

A further assay for measuring the activity or efficacy of an endostatin peptide or variant on angiogenesis is the chick chorioallantoic membrane (CAM) assay. Briefly, 3 day old chicken embryos with intact yolks are separated from the egg and placed in a petri dish. After 3 days of incubation a methylcellulose disc containing an endostatin peptide or variant to be tested is applied to the CAM of individual embryos. After 48 hours of incubation, the embryos and CAMs are observed to determine whether endothelial growth has been inhibited. As with the assay described above, the extent of endothelial cell growth compared to control peptide treatments inversely correlates with the activity and/or efficacy of the endostatin peptide or variant. This method is described, for example, in O'Reilly, et al . , Cell 79:315-328 (1994), and in U.S. Patent No. 5,753,230, both of which are incorporated herein by reference.

An additional assay for determining the activity or efficacy of an endostatin peptide or variant on angiogenesis is the rabbit corneal assay. This assay involves implanting a growth factor-containing pellet, along with another pellet containing an endostatin peptide or variant, in the cornea of a rabbit or other laboratory mammal and observing the pattern of capillaries that are elaborated in the cornea. The extent and complexity of capillary formation compared to control peptide treatments inversely correlates with the activity and/or efficacy of the endostatin peptide or variant.

Angiogenic inhibitory activity and efficacy of an endostatin peptide or variant of the invention also can be determined in credible animal models known in the art. For example, animal models for tumor growth and metastasis are applicable for determining the anti-angiogenic effect or the angiogenic-dependent disease inhibitory effect of endostatin peptides or of endostatin variants of the invention. Briefly, tumor growth can be induced in an animal model by, for example, injecting metastatic tumors into the animal and determining the extent of lung colonization or secondary tumor formation in the presence or absence of an endostatin peptide or variant. The extent of lung colonization or secondary tumor function inversely correlates with the activity and/or efficacy of the endostatin peptide or variant. Similar assays can be employed using solid tumors and measuring the size or growth rate of the tumor as an indicator of endostatin peptide activity and/or efficacy. For a description of such animal models see, for example, U.S. Patent No. 5,753,230 and PCT publication WO 97/15666, supra . Other credible animal models are known to those skilled in the art and can similarly be used to determine the effect of an endostatin peptide or variant on inhibiting the extent of tumor growth or metastasis.

As described above, an endostatin peptide or variant can be administered to a tumor bearing animal to determine the inhibiting activity or efficacy of an endostatin peptide or variant on tumor growth, compared to a non-endostatin peptide control. A decrease in the rate or extent of tumor growth, or a disappearance of the tumor correlates with the antiangiogenic activity and efficacy against progression of an angiogenic-dependent disease. For a further description of tumor bearing animal models see, for example, U.S. Patent No. 5,639,725, which is incorporated herein by reference.

In determining the activity and/or efficacy of an endostatin peptide or variant in any of the above methods, the endostatin peptides or variants can be administered within a concentration range known in the art to be indicative of an inhibitor's activity in a particular assay. For example, a concentration of a polypeptide inhibitor which yields indicative results in the bovine capillary endothelial cell assay is generally about 100-1000 ng/ml. Similarly, the concentration of a polypeptide inhibitor which would yield a positive result for a polypeptide inhibitor that is active against angiogenesis in a CAM is generally about, for example, 0.5-20 μg/ml, 10-20 μg/disc, over a range of concentrations of 0.1-100 μg/disc, or 25 μg/disc. An indicative concentration for a polypeptide inhibitor which would be expected to yield positive results in the rabbit corneal assay is generally about 40 μg/hydron pellet. Finally, concentrations for polypeptide inhibitors which would yield indicative results in the tumor metastasis and tumor bearing animal models described above are about 250 μg twice weekly or 10 mg/kg/day for 10 days; and 12.5 μg daily or 1 mg twice a week, respectively. These concentrations are similarly applicable for determining the extent of inhibitory activity or efficacy of an endostatin peptide or variant in each of the respective models described above. Further refinement can be performed by, for example, varying the concentration of an endostatin peptide or an endostatin variant within the active concentration range to determine an optimal concentration or amount for inhibiting angiogenesis. Similarly, the endostatin peptide or variant can be combined with various pharmaceutically acceptable mediums or carriers to determine an appropriate or beneficial composition in which the peptides can be administered in or which may augment or stabilize the endostatin peptide or variant.

Moreover, the above described models, as well as other methods known to those skilled in the art, can similarly be used to determine appropriate dosage regimes in regard to timing of administrations, number of administrations and amount per administration of endostatin peptide or endostatin variant to inhibit angiogenesis-dependent or to treat an angiogenesis-dependent disease. Similarly, the above described methods also can be routinely used to make and identify new, modified or improved endostatin peptides and variants. Given the teachings and guidance described herein, those skilled in the art will know or can determine an endostatin peptide having α_v integrin binding activity and an effective amount of the endostatin peptide or variant to inhibit angiogenesis-dependent or to treat an angiogenesis-dependent disease.

As used herein, the term "angiogenic inhibitory amount" is intended to mean an amount of an endostatin peptide or variant of the invention required to effect a decrease in the extent, amount or rate of neovascularization when administered to an individual. The dosage of an endostatin peptide or variant required to be therapeutically effective will depend, for example, on the angiogenesis-dependent disease to be treated, the route and form of administration, the potency and bio-active half-life of the molecule being administered, the weight and condition of the individual, and previous or concurrent therapies . The appropriate amount considered to be an effective dose for a particular application of the method can be determined by those skilled in the art, using the guidance provided herein. For example, the amount can be extrapolated from in vi tro or in vivo angiogenesis assays described above. One skilled in the art will recognize that the condition of the patient needs to be monitored throughout the course of therapy and that the amount of the composition that is administered can be adjusted accordingly.

For inhibiting angiogenesis or treating an angiogenesis-dependent disease, an angiogenic inhibitory amount of an endostatin peptide of the invention can be, for example, between about 10 μg/kg to 500 mg/kg body weight, for example, between about 0.1 mg/kg to 100 mg/kg, or preferably between about 1 mg/kg to 50 mg/kg, depending on the treatment regimen. For example, if an endostatin peptide is administered from one to several times a day, then a lower dose would be needed than if an endostatin peptide were administered weekly, or monthly or less frequently. Similarly, formulations that allow for timed-release of an endostatin peptide would provide for the continuous release of a smaller amount of an endostatin peptide than would be administered as a single bolus dose. For example, an endostatin peptide can be administered at 4 mg/kg/week.

An endostatin peptide or variant can be delivered systemically, such as intravenously or intraarterially . An endostatin peptide or variant can also be administered locally at a site of angiogenesis. Appropriate sites for administration of an endostatin peptide or variant are known or can be determined by those skilled in the art depending on the clinical indications of the individual being treated. For example, the endostatin peptides and variants, having inhibitory activity described above can be provided as isolated and substantially purified proteins and protein fragments as well as insoluble aggregate in pharmaceutically acceptable formulations using formulation methods known to those of ordinary skill in the art. These formulations can be administered by standard routes, including for example, topical, transdermal, intraperitoneal, intracranial, intracerebroventricular, intracerebral, intravaginal, intrauterine, oral, rectal or parenteral (e.g., intravenous, intraspinal, subcutaneous or intramuscular) routes. In addition, an endostatin peptide or variant can be incorporated into biodegradable polymers allowing for sustained release of the compound, the polymers being implanted in the vicinity of where drug delivery is desired, for example, at the site of a tumor or implanted so that the endostatin is released systemically over time. Osmotic minipumps peptide or variant can also be used to provide controlled delivery of high concentrations of endostatin peptides and variants through cannulae to the site of interest, such as directly into a metastatic growth or into the vascular supply a tumor. The biodegradable polymers and their use are described, for example, in detail in Brem et al . , J. Neurosurg. 74:441-446 (1991), which is incorporated herein by reference .

The invention provides compositions of endostatin peptides and variants together with a pharmaceutically acceptable medium and formulations. Such compositions can be used in a method of the invention to inhibit angiogenesis or treat an angiogenesis-dependent disease. For example, an endostatin peptide or variant can be administered as a solution or suspension together with a pharmaceutically acceptable medium. Such a pharmaceutically acceptable medium can be, for example, water, sodium phosphate buffer, phosphate buffered saline, normal saline or Ringer's solution or other physiologically buffered saline, or other solvent or vehicle such as a glycol, glycerol, an oil such as olive oil or an injectable organic ester.

The endostatin peptide or variant formulations include those applicable for parenteral administration such as subcutaneous, intraperitoneal, intramuscular, intravenous, intradermal, intracranial, intratracheal, and epidural administration. As well as formulations applicable for oral, rectal, ophthalmic (including intravitreal or intracameral) , nasal, topical (including buccal and sublingual), intrauterine, or vaginal administration. The endostatin peptide or variant formulations can be presented in unit dosage form and can be prepared by pharmaceutical techniques well known to those skilled in the art. Such techniques include the step of bringing into association the active ingredient and the pharmaceutical carrier (s) or excipient(s) .

Formulations suitable for parenteral administration include aqueous and non-aqueous sterile injection solutions such as the pharmaceutically acceptable mediums described above. The solutions can additionally contain, for example, anti-oxidants , buffers, bacteriostats and solutes which render the formulation isotonic with the blood of the intended recipient. Other formulations include, for example, aqueous and non-aqueous sterile suspensions which can include suspending agents and thickening agents. The formulations can be presented in unit-dose or multi-dose containers, for example, sealed ampules and vials, and can be stored in a lyophilized condition requiring, for example, the addition of the sterile liquid carrier, immediately prior to use. Extemporaneous injection solutions and suspensions can be prepared from sterile powders, granules and tablets of the kind previously described.

A pharmaceutically acceptable medium can additionally contain physiologically acceptable compounds that act, for example, to stabilize or increase the absorption of the endostatin peptide. Such physiologically acceptable compounds include, for example, carbohydrates such as glucose, sucrose or dextrans; antioxidants such as ascorbic acid or glutathione; chelating agents such as EDTA, which disrupts microbial membranes; divalent metal ions such as calcium or magnesium; low molecular weight proteins; lipids or liposomes; or other stabilizers or excipients. An endostatin peptide or variant can also be formulated with a pharmaceutically acceptable medium such as a biodegradable polymer.

The endostatin peptides and variants of the invention also can be delivered to an individual for inhibiting angiogenesis or treating an angiogenesis-dependent disease by administering an encoding nucleic acid for the peptide or variant. Therefore, the encoding nucleic acids for the endostatin peptides and variants of the invention are useful in conjunction with a wide variety of gene therapy methods known in the art for delivering an angiogenic inhibitory amount of the peptide or variant. Using the teachings and guidance provided herein, encoding nucleic acids for one or more endostatin peptides, endostatin variants or a combination thereof can be incorporated into a vector or delivery system known in the art and used for delivery and expression of the encoding sequence to achieve an angiogenic inhibitory amount. Applicable vector and delivery systems known in the art include, ror example, retroviral vectors, adenovirus vectors, adenoassociated virus, ligand conjugated particles and nucleic acids for targeting, isolated DNA and RNA, liposomes, polylysine, and cell therapy, including hepatic cell therapy, employing the transplantation of cells modified to express endostatin peptides and variants, as well as various other gene delivery methods and modifications known to those skilled in the art, such as those described in Shea et al., Na ture Biotechnol . 17:551-554 (1999), which is incorporated herein by reference.

Specific examples of methods well known in art are described in, for example, United States Patent No. 5,399,346; United States Patent Nos. 5,580,859; 5,589,466; 5,460,959; 5,656,465; 5,643,578; 5,620,896; 5,460,959; 5,506,125; European Patent Application No. EP 0 779 365 A2; PCT No. WO 97/10343; PCT No. WO 97/09441; PCT No. WO 97/10343, all of which are incorporated herein by reference. Other methods known to those skilled in the art also exist and are similarly applicable for the delivery of an angiogenic inhibitory amount of an endostatin peptide or variant by expressing the encoding nucleic acid sequence.

In one particular method described above, inhibiting angiogenesis or treating an angiogenesis-dependent disease can be accomplished by administering expressible nucleic acids encoding endostatin peptides or variants of the invention. In this method, the encoding nucleic acid can be DNA or RNA and also can be administered substantially free of vector sequences, or alternatively, incorporated into a vector. The nucleic acids can further be administered alone or together with other agents that facilitate the uptake of the nucleic acid into cells. The encoding DNA or RNA can be isolated or synthesized using the methods and teaching described herein.

The endostatin peptide or variant encoding nucleic acid is subsequently administered directly into a tissue of an individual. Preferably, the DNA or RNA-containing the encoding nucleic acid is injected into the skeletal muscle of the individual. For example, a 1.5 cm incision can be made to expose the quadricep muscles of the subject. A 0.1 ml solution containing from 10-100 μg of a DNA or RNA and 5-20% sucrose is injected over 1 minute into the exposed quadriceps muscles about 0.2 cm deep. The skin is thereafter closed. The amount of DNA or RNA can range from 10 to 100 μl of hypotonic, isotonic or hypertonic sucrose solutions or sucrose solutions containing 2 mM CaCl₃. The DNA or RNA containing solutions can also be administered over a longer period of time, for example, 20 minutes, by infusion. The in vivo expression of the desired gene can be tested by determining an increased production of the encoded endostatin peptide or variant by the subject according to methods known in the art or as described, for example, in Wolff et al., Science 247:1465-1468 (1990), which is incorporated herein by reference. As specific example of using this method with intact endostatin is described in Blezinger et al. Na ture Biotechnology 17:343-348 (1999), which is incorporated herein by reference.

The treated cells will respond to the direct injection of DNA or RNA by expressing the encoded peptide for at least about 60 days. Thus, the desired endostatin peptide or variant can be effectively expressed by the cells of the individual as an alternative to administering the peptide or variant to the individual. The above description is exemplary for the expression of nucleic acids by direct injection into the tissue of an individual. Numerous modifications of this procedure are well known to those skilled in the art and are similarly applicable for expression of angiogenic inhibitory amounts of endostatin peptides or variants as described herein.

The present invention also relates to encoding nucleic acids and vectors useful in the gene therapy methods and can be prepared by methods known in the art . Compositions containing such nucleic acids, vectors and pharmaceutically acceptable medium are also provided. The pharmaceutically acceptable medium should not contain elements that would degrade the desired nucleic acids. The methods of using endostatin peptides and variants can employ any of the various species of endostatin peptides and variants previously set forth. For example, angiogenesis can be inhibited and an angiogenic-dependent disease can be treated using an endostatin peptide containing at least about 4-7 endostatin amino acid residues containing substantially the amino acid sequence RLQD, RAD, DGK/R, or a functional fragment thereof. Similarly, angiogenesis can be inhibited and an angiogenesis-dependent disease can be treated using an endostatin variant containing the amino acid sequence RGD, or a functional fragment thereof.

The endostatin peptides can be small, synthetic peptides of about 4-7 residues or they can be large peptides of up to about 182 residues so long as they are not the full length endostatin molecule and retain α_v integrin binding activity. Similarly, the endostatin peptides and variants can be of any size between, or larger than the above boundaries so long as they contain at least about 4-7 endostatin amino acid residues and the sequences RLQD, RAD, DGK/R or RGD. Therefore, the peptides used for inhibiting angiogenesis or treating an angiogenic-dependent disease can contain substantial endostatin amino acid residues and/or non-endostatin amino acid sequences.

Non-endostatin sequences can impart structural or functional characteristics onto the endostatin peptides of the invention. For example, chimeric endostatin peptides or variants can be used to impart a targeting function, such as glycosaminoglycan binding, onto the α_v integrin binding function of the endostatin peptide or to increase the efficacy of the peptide by incorporating multiple endostatin peptides into a single polypeptide chain. Targeting of an endostatin peptide or variant to the site of aberrant angiogenesis using, for example, glycosaminoglycan binding activity confers the additional therapeutic advantage of anchoring the endostatin peptide or variant at the site of the pathological condition. This result therefore sustains a high effective concentration of the peptide diffusible into the angiogenic area over time and essentially allows for a continuous local administration of the endostatin peptide or variant to the site of angiogenesis.

Additionally, two or more endostatin peptides or variants of the invention can be administered in the methods of the invention to inhibit angiogenesis or to treat an angiogenesis-dependent disease. Similarly, one or more endostatin peptides can be administered in combination with one or more endostatin variants to inhibit angiogenesis or treat an angiogenic-dependent disease. Therefore, various combinations and permutations of endostatin peptides, endostatin variants and combinations of endostatin peptides and variants together can be administered in the methods of the invention for the effective inhibition of angiogenesis and treatment of an angiogenesis-dependent disease.

Endostatin peptides, endostatin variants and combinations thereof can also be delivered in an alternating administrations so as to combine their angiogenic inhibiting effects over time. For example, an endostatin variant can be administered in a single bolus dose followed by multiple administrations of one or more endostatin peptides alone or in combination with an endostatin variant. Whether simultaneous or alternating delivery of the endostatin peptide, endostatin variant or combination thereof, the mode of administration can be any of those types of administrations described previously and will depend on the particular therapeutic need and efficacy of the endostatin peptides or variants selected for the purpose. Determining which species of endostatin peptides or variants to combine in a cocktail or to combine in a temporally administered regime, will depend on the angiogenesis-dependent disease and the specific physical characteristics of the individual affected with the disease. Those skilled in the art will know or can determine a specific cocktail or regime of administration which is effective for a particular application using the teachings and guidance provided herein together with diagnostic and clinical criteria known within the field of art of the particular angiogenesis-dependent disease.

The methods of inhibiting angiogenesis or treating angiogenesis-dependent disease by administering endostatin peptides or variants additionally can be practiced in conjunction with other therapies. For example, for inhibiting tumor angiogenesis or treating cancer, a method of the invention can be practiced prior to, during, or subsequent to conventional cancer treatments such as surgery, chemotherapy, radiation or other methods known in the art .

In a further embodiment, an endostatin peptide, endostatin variant or combination thereof can be administered in combination with an effective amount of an angiogenic inhibitor. The angiogenic inhibitor will augment the angiogenic inhibitory effect of the endostatin peptide or variant. As with the endostatin peptide and variant cocktails described previously, administration of an angiogenic inhibitor can be, for example, simultaneous with an endostatin peptide or variant, or delivered in alternative administrations. Simultaneous administration of an endostatin peptide, endostatin variant or combination thereof together with an angiogenesis inhibitor can be, for example, together in the same formulation or in different formulations delivered at about the same time or immediately in sequence. Alternating administrations can be, for example, delivering an endostatin peptide, endostatin variant or combination thereof in temporally separate administrations. As described above in reference to endostatin peptide and variant cocktails, administrations of an endostatin peptide, endostatin variant, or combinations thereof and angiogenic inhibitors can similarly use different modes of delivery and routes.

Angiogenic inhibitors are known in the art and can be prepared by known methods. For example, angiogenic inhibitors include integrin inhibitory compounds such as α_v integrin inhibitory antibodies, cell adhesion proteins or functional fragments thereof which contain a cell adhesion binding sequence. A specific example of an _v inhibitory antibody is the monoclonal antibody LM609,

Cheresh and Spiro. J. Biol . Chem . 262: 17703-17711 (1987). A specific example of inhibitory adhesion protein fragments, including small synthetic peptides, is a non-endostatin variant peptide containing the amino acid sequence RGD. Such inhibitors additionally include, for example, chimeric and humanized antibodies which inhibit α_v-integrins as well as, derivatives and mimetics of cell adhesion proteins and peptides. RGD peptides and their use for inhibition of integrin binding is the subject matter of several patents including, for example, U.S. Patent Nos. 4,517,686, 4,578,079, 4,589,881, 4,614,517, 4,661,111, 4,683,291, 4,792,525, 4,879,237, 4,988,621, 5,041,380, 5,061,693 and 5,753,230, all of which are incorporated herein by reference .

Additional angiogenic inhibitors include, for example, angiostatin, functional fragments of angiostatin, endostatin, fibroblast growth factor (FGF) inhibitors, FGF receptor inhibitors, VEGF inhibitors, VEGF receptor inhibitors, vascular permeability factor (VPF) inhibitors, VPF receptor inhibitors, thrombospondin, platelet factor 4, interferon-alpha, interferon-gamma, interferon-inducible protein 10, interleukin 12, gro-beta, and the 16 kDa N-terminal fragment of prolactin. Angiostatin is the subject matter of U.S. Patent No. 5,639,725, supra . Endostatin is the subject matter of PCT publication WO 97/15666, supra . For a description of the remaining angiogenic inhibitors and targets set forth above, see for example, Chen et al . , Cancer Res . 55:4230-4233 (1995); Good et al., Proc . Na tl . Acad. Sci . USA 87:6624-6628 (1990); O'Reilly et al . , Cell 79:315-328 (1994); Parangi et al . , Proc . Na tl . Acad . Sci . USA 93:2002-2007 (1996); Rastinejad et al., Cell 56:345-355 (1989); Gupta et al., Proc . Na tl . Acad . Sci . USA 92:7799-7803 (1995); Maione et al . , Science 247:77-79 (1990): Angiolillo et al . , J. Exp . Med . 182:155-162 (1995); Strieter et al . , Biochem . Biophys . Res . Comm . 210:51-57 (1995); Voest et al . , J. Na tl . Cancer Inst .

87:581-586 (1995); Cao et al., J. Exp . Med . 182:2069-2077 (1995); Clapp et al . , Endocrinology 133:1292-1299 (1993), respectively. For a description of additional angiogenic inhibitors see, for example, Blood et al., Bioch . Biophys . Acta . , 1032:89-118 (1990); Moses et al . , Science,

248:1408-1410 (1990); Ingber et al . , Lab Inves t . , 59:44-51 (1988) and U.S. Patent Nos. 5,092,885; 5,112,946; 5,192,744 and 5,202,352. All of the above references are incorporated herein by reference.

The invention further provides a method of inducing apoptosis. The method consists of administering to an _v integrin-containing cell an inhibitory amount of an endostatin peptide having at least about 4-7 endostatin amino acid residues containing one or more of the amino acid sequence RLQD, RAD, DGK/R; a functional equivalent of the endostatin peptide comprising one or more of the amino acid sequence RLQD, RAD, DGK/R; an endostatin variant comprising the amino acid sequence RGD, or a functional fragment of the endostatin variant comprising the amino acid sequence RGD.

Apoptosis plays a significant role in numerous pathological conditions in that programmed cell death is either inhibited, resulting in increased cell survival, or enhanced which results in the loss of cell viability. Examples of pathological conditions resulting from increased cell survival include cancers such as lymphomas, carcinomas and hormone dependent tumors. Such hormone dependent tumors include, for example, breast, prostate and ovarian cancer. Autoimmune diseases such as systemic lupus erythematosus and immune-mediated glomerulonephritis as well as viral infections such as herpesvirus, poxvirus and adenovirus also result from increased cell survival or the inhibition of apoptosis. Other apoptotic diseases where enhanced programed cell death is a prevalent cause include, for example, degenerative disorders such as Alzheimer's disease, Parkinson's disease, Amyotrophic lateral sclerosis, Retinitis pigmentosa, and Cerebellar degeneration. Other diseases associated with increased apoptosis include, for example, myelodysplastic syndromes such as aplastic anemia and ischemic injury including myocardial infarction, stroke and reperfusion injury.

As inhibitors of α_v integrins, including for example, α_vβ₃, α_vβ₅ and α_vβχ, as well as integrins which exhibit similar binding specificity such as α_IIβ₃ and oι_{5 1} an endostatin peptide or variant of the invention can be used to inhibit integrin mediated survival signals. Inhibition can be accomplished by, for example, detaching cells from substrates or by blocking the binding of cell survival integrins to their ligands without substantially affecting cell attachment through other integrins or cell adhesion receptors so as to induce apoptosis. For a description of integrin detachment induced apoptosis, see for example, Frisch et al . , J. Cell Biol . 134:793-799 (1996), which is incorporated herein by reference.

Therefore, diseases caused by inappropriate proliferation or abnormal cell regulation, which can be independent of angiogenesis, also can be treated with an endostatin peptide, an endostatin variant or various combinations thereof. The use of an endostatin peptide or variant of the invention for cell detachment induced apoptosis does not preclude simultaneous inhibition of angiogenesis and angiogeneis-dependent diseases.

Cell apoptosis can be accomplished by, for example, inhibiting _v integrin binding to its ligand.

Cell detachment induced apoptosis can be accomplished by, for example, inhibiting α_v integrin binding to its ligand as well as to inhibit the attachment of the cell to non- _v integrin bound extracellular matrix substrates. Therefore, the endostatin peptides and variants can be used directly on cells expressing α_v-containing integrins as their primary repertoire of cell adhesion receptors. Alternatively, in cell types which contain a repertoire of integrins and other cell adhesion receptors, some of which are α_v-containing integrins, endostatin peptides and variants can be used in conjunction with other inhibitors of non-α_v integrin and non-integrin cell adhesion receptors . Substantial inhibition of binding of the repertoire of all cell adhesion receptors will promote the detachment of treated cells from their substrates followed by a concomitant induction of apoptosis.

Methods for inducing cell detachment are similar to those described above for the inhibition of angiogenesis and for treating an angiogenesis-dependent disease. The endostatin peptide, endostatin variant or combination thereof can be devoid of any functional domains that allow incorporation of the endostation peptide or variant into a solid matrix such as the extracellular matrix. This solubility characteristic increases the amount of endostatin peptide or variant in solution for effective inhibition of ligand binding and promotion of cell detachment. Inclusion of a matrix binding domain such as a glycosaminoglycan binding domain can nevertheless be used for recruitment of the endostatin peptide or variant into a localized area for the inhibition of angiogenesis. Subsequent release of endostatin peptide or variant molecules from the insoluble form will also inhibit angiogenesis as well as promote cell detachment. However, the kinetics of release and cell detachment promoting activity can be slower for matrix bound forms compared to administration of endostatin peptides and variants that are refractory to incorporation into the extracellular matrix.

Alternatively, and as described below recruitment of an endostatin peptide or variant into an insoluble form, which is not subsequently released, also can be used to promote cell adhesion as well as to promote neovascularization .

Methods of determining the effects of an endostatin peptide or variant of the invention on apoptosis are known in the art. Briefly, apoptosis can be measured using well known methods such as determining the fragmentation chromatin DNA or by measuring the activity of any of a number of proteases, termed cysteine-aspartate specific proteases, or caspases, in response to detachment with an endostatin peptide, an endostatin variant or combination thereof. Methods for determining the apoptotic state of a cell are well known in the art in regard to measurements of DNA fragmentation. Methods for measuring caspase activity are also well known. Other methods for measuring apoptotic activity of a cell also are known to those skilled in the art and can similarly be used to determine the effect of an endostatin peptide, an endostatin variant or combination thereof in promoting cell detachment induced cell death. Moreover, colorimetric peptide substrates which are activated upon cleavage by a caspase and kits for convenient determination are available through commercial sources . Exemplarily commercial sources for caspases, caspase substrates and inhibitors as well as other apoptosis products include, for example, Alexis Corporation, San Diego, California; Biomol Research Lab Inc., Plymouth Meeting, Pennsylvania; Bio-Rad Laboratories, Hercules, California; Calbiochem-Novabiochem International Inc., San Diego, California; Chemicon International Inc., Temecula, California; Kamiya Biomedical Company, Seattle, Washington; MBL International, Corp., Watertown, Massachusetts; Oncolmmunin, Inc., College Park, Maryland. The invention also provides a method of promoting endothelialization or vascularization. The method consists of contacting a tissue having inadequate vascularization with endostatin; an endostatin peptide or functional equivalent thereof; an endostatin variant or a functional fragment thereof.

In tissues having inadequate vascularization, such as wound tissues requiring cellular repair with concomitant vascularization of the newly regenerated area, endostatin peptides and variants as well as endostatin can be used as an endothelial cell adhesion substrate to recruit endothelial cells into the affected area and induce tubule formation. Briefly, an endostatin peptide, endostatin variant or endostatin can be administered to the site of inadequate vascularization in, for example, an insoluble form or in a form that will allow its incorporation into the extracellular matrix. Once administered, cells can adhere to the endostatin derived matrix through α_v integrin dependent interactions resulting in similar effects as occur through α_v integrins and endogenous extracellular matrix proteins .

Endostatin peptides, endostatin variants, endostatin or combinations thereof can be administered alone or in combination with synthetic or biopolymers and matrices to facilitate application and long-term integrity of the endostatin compositions. The administration of cell adhesion peptides and polypeptides for promoting cell migration and attachment, such as that involved in endothelialization, is well known to those skilled in the art. Similarly, the production and use of artificial biomatrices for cell migration and adhesion, such as that involved in endothelialization, also is well know to those skilled in the art. In light of the teachings and guidance provided herein, those skilled in the art can substitute the endostatin peptides, endostatin variants, combinations thereof as well as endostatin in any of various methods known in the art for promoting endothelial cell migration and endothelialization at a site requiring vascularization.

The invention further provides a method of promoting cell adhesion. The method consists of contacting cells with an endostatin peptide having at least about 4-7 endostatin amino acid residues containing one or more of the amino acid sequence RLQD, RAD, DGK/R; a functional equivalent of said endostatin peptide comprising one or more of the amino acid sequence RLQD, RAD, DGK/R; an endostatin variant comprising the amino acid sequence RGD, or a functional fragment of said endostatin variant comprising the amino acid sequence RGD. The method also consists of promoting cell adhesion of non-angiogenic cells by contacting the non-angiogenic cells with endostatin. The method is advantageous, for example, in promoting or facilitating cell attachment and migration in wound healing.

As described above in regard to promoting vascularization, in tissues requiring cellular repair such as in wound healing, endostatin peptides and variants as well as endostatin similarly can be used as a cell adhesion substrate to recruit or promote cell migration into the affected area and facilitate cellular reconstruction of the wounded area. Briefly, an endostatin peptide, endostatin variant or endostatin can be administered to the site of the wound in, for example, an insoluble form or in a form that will allow its incorporation into the extracellular matrix. Once administered, cells can adhere to the endostatin derived matrix in much the same manner as to endogenous extracellular matrix. Endostatin peptides, endostatin variants, endostatin or combinations thereof can be administered alone or in combination with synthetic or biopolymers and matrices to facilitate application and long-term integrity of the endostatin compositions.

As binders of _v integrins and integrins with similar binding specificity such as α_IIbβ₃ and ₅β_1; the endostatin peptides of the invention are applicable in a wide range of methods for treating diseases mediated by these integrins . The endostatin peptides of the invention can therefore be used to promote cell adhesion or inhibit cell adhesion to treat a particular integrin mediated disease. Wound healing as described above is one such example of promoting cell adhesion to treat an integrin mediated disease. Osteoporosis and abberrant platelet aggregation are other examples of integrin mediated diseases. For example, the _vβ₃ and α_IIbβ₃ integrins, respectively, are involved in mediating the detrimental effects of these diseases. Those skilled in the art knows which disease or pathological conditions are influenced by _v integrin binding activity and those with similar binding specificity and are amendable to augmentation or inhibition with the endostatin peptides or variants of the invention.

The invention additionally provides a method of purifying recombinant endostatin which results in improved yields of a soluble, active form of the polypeptide. The method consists of denaturing recombinant endostatin, purifying the denatured endostatin and then renaturing the purified endostatin to result in refolded endostatin. The refolded endostatin is soluble and exhibits binding activity to α_v integrins. The method is applicable for purifying recombinant endostatin as well as endostatin from natural sources. The method also is applicable for purifying endostatin peptides and endostatin variants. The method will be described for exemplary purposes only with reference to recombinant endostatin.

Recombinant endostatin can be expressed in essentially any compatible vector/host system, including for example, both procaryotic and eucaryotic expression systems. Expression systems and corresponding host cell types are well known to those skilled in the art.

Similarly, promoters, enhancers and regulatory elements for augmenting or regulating expression and applicable cell types compatible with such expression elements also are well known to those skilled in the art. Procaryotic expression systems are advantageous due to their ease in manipulation, rapid growth rate and relatively high yields. Eucaryotic systems are advantageous in that they are subject to eucaryotic post-translational modifications, including glycosylation for example. The particular expression system, cell type or modification employed for producing recombinant endostatin will therefore primarily be a choice of the user. An exemplary recombinant expression system is described further below in Example I. However, numerous other systems are similarly applicable for producing recombinant endostatin as well. Such other systems include, for example, all other E . coli vector/host expression systems, yeast expression systems, insect expression systems such as baculovirus, and mammalian systems, including rodent and human vector/host systems.

An advantage of the method for purifying recombinant endostatin described herein is that it does not discriminate against, or lose efficiency, when used in connection with high yield expression systems that result in aggregation or precipitation of the recombinant protein because the method initially denatures all proteins within the cell. Therefore, aggregated, insoluble endostatin is initially solubilized to the same extent as aqueous, soluble endostatin irrespective of their pre-denaturation form. As such the fraction of endostatin protein within the cell that is available for purification is proportional to its amount and not its aqueous solubility characteristics.

The method consists of adding to endostatin expressing cells a denaturant that is capable of lysing the cells and denaturing proteins within the cell. Generally, about lxlO⁶ to 1x10' cells/ ml is a sufficient number of bacterial cells to yield 1.5 mg of soluble endostatin from a liter of culture. Depending on the expression level of the vector/host system, a comparable number of cells from other expression systems will provide similar results. Those skilled in the art will know to scale the cell culture up or down depending on the expression level and desired amount of endostatin to be purified. The cells can be concentrated by, for example, centrifugation or other methods known to those skilled in the art.

The amount of denaturant should be sufficient to substantially solubilize all proteins within the cell and will therefore be proportional to the amount of cells in the starting material. Denaturants also effect lysis and it is therefore unnecessary to include a separate lysis step for the endostatin expressing cells. However, if the selected denaturant is mild or if it is desired increase efficiency of the lysis and denaturation, additional steps can be performed. For example, the cells can be subjected to one or more cycles of freezing and thawing, sonication, or the denaturant can be heated to above room temperature. Denaturants useful lysing the cells and solubilizing proteins include, for example, guanidine-HCl, guanidine-isothiocyanate and urea. Other denaturants well known to those skilled in the art exist as well and can similarly be used in the purification method of the invention.

Following lysis and denaturation, the mixture can be subjected to a fractionation to remove, for example, nucleic acid. Although not necessary, the inclusion of such a fractionation will increase the purification efficiency and therefore the yield.

Purification of endostatin can be accomplished by routine biochemical methods known to those skilled in the art. Such methods include, for example, affinity chromatography, immunoaffinity chromatography, gel filtration as well as anion and cation exchange chromatography. Purification also can be facilitated by fusing endostatin with a purification tag in the expressed molecule. In the expression system described below in Example I, the purification tag is a Histidine tag which can be bound by a nickel affinity resin to efficiently purify endostatin. Other affinity resins and purification tags are known to those skilled in the art and can be employed in the purification methods of the invention. Purification should be compatible with a protein denaturant so as to maintain endostatin in an unaggregated and soluble form. The denaturant can be different from the chemical used for initial lysis and solubilization of the cells. Urea, for example, is one denaturant that is compatible with the biochemical purification methods listed above for use in the purification step of endostatin. Other denaturants are known to those skilled in the art and can similarly be used in the purification step of endostatin.

Renaturation can be performed by various methods known to those skilled in the art. Such methods include the gradual removal of the denaturant to allow the polypeptide to refold into an active conformation. For example, dialysis of purified endostatin in successive buffers containing decreasing concentrations of denaturant can be used to refold denatured endostatin into an active conformation. The dialysis can be performed, for example, using relatively longer dialysis periods combined with greater decreases in denaturant concentrations between successive buffers. Alternatively, shorter dialysis periods can be employed with smaller changes in the denaturant concentration. The final dialysis should be in a physiologically acceptable medium such as physiological buffered saline (PBS) or other appropriate medium that is desired for subsequent use of the purified, aqueous soluble and active endostatin. Reducing agents and the like can be optionally included at various concentrations within the refolding buffer to facilitate renaturation into an active conformation. Other methods for refolding are known to those skilled in the art and also can be used in light of the teaching and guidance provided herein.

Following renaturation into an active conformation, the purified endostatin can optionally be subjected to additionally purification steps or additional procedures. For example, the purified endostatin can be further treated to remove bacterial endotoxin or equilibrated into a different buffer. It is understood that modifications that do not substantially affect the activity of the various embodiments of this invention are also included within the definition of the invention provided herein. Accordingly, the following examples are intended to illustrate but not limit the present invention.

EXAMPLE I

Endostatin Binds to ex.. Integrins and Inhibits α..-Mediated Endothelial Cell Functions

This Example describes the expression, purification and α_v integrin-mediated activity of recombinant, soluble endostatin derived from type XVIII collagen .

Human endostatin (ES) was expressed in __. col i and purified to apparent homogeneity with Ni2π—NTA-agarose and sequential heparin Sepharose chromatography with a typical yield of 1.5 mg of soluble, refolded protein/1 of culture. Briefly, the C-terminal part of the cDNA of collagen XVIII was used to amplify the cDNA of human endostatin which was cloned into the Qiagen pQE-vector. To construct the vector, a 560-bp cDNA fragment of human collagen XVIII that corresponds to mouse endostatin sequences was generated by polymerase chain reaction with primers H18-END2 (5 ' -ATGGTACCCCACAGCCACCGCGACTT-3 ' , having a Hindlll restriction site (SEQ ID NO:98)) and H18-HIS-10 (5'-ATAAGCTTACTTGGAGGCAGTCAT-3' (SEQ ID NO:99), having a Kpnl restriction site) using human cDNA clone HP19.3 as a template (Saarela et al . Ma trix Biol . 16:319-328 (1998)). The fragment was subcloned into the Kpnl/Hindlll-site of the vector pQE-31 (Qiagen, Inc., Santa Clarita, CA) , which can be used to express polypeptides with an N-terminal His-tag. Expression of human endostatin in E . col i strain M15 was induced and carried out according to the manufacturer's protocol (Qiagen).

To purify the recombinant endostatin, bacterial pellets were suspended in 6 M guanidine-HCl, 0.5 M NaCl, 10 mM β-mercaptoethanol, 20 mM Tris-HCl, pH 7.9, snap-frozen in -70°C and lysed by rotation at room temperature for 1 h. The suspension was centrifuged at 12,000g for 20 min, and the supernatant was lightly sonicated, after which it w.as applied to a ProBond column (Invitrogen, San Diego, CA) that had been pre-equilibrated with 8 M urea, 0.5 M NaCl, and 20 mM Tris-HCl, pH 7.9. Bound protein was eluted using a stepwise imidazole gradient from 0 M to 0.5 M in the equilibrium buffer. The eluted fractions were monitored at A280, and verified by SDS-PAGE and Coomassie staining. Fractions containing endostatin were pooled and refolded in vi tro, first by dialyzing overnight at 4°C against 4 M urea, 0.1 M NaCl, 1 mM/0.1 mM reduced/oxidized glutathione, 20 mM Tris-HCl, pH 7.9, and then for 6 h against 1 M urea, 0.1 M NaCl, 0.1 mM/0.01 mM reduced/oxidized glutathione, 20 mM Tris-HCl, pH 7.9. Pooled endostatin was subsequently dialyzed overnight against PBS, pH 6.9 after which the refolded soluble endostatin was separated by centrifuging at 12, OOOg for 20 min.

Soluble endostatin was then applied to a HiTrap

SP cation-exchange column (Amersham Pharmacia Biotech, Piscataway, NJ) and fractions were eluted by an increasing NaCl gradient from 0.1 M to 1.5 M in PBS, pH 6.9. Fractions containing endostatin were pooled and dialyzed against 0.1 M NaCl, 20 mM Tris-HCl, pH 7.4, and applied to a heparin-Sepharose CL-6B column (Amersham Pharmacia Biotech) . Bound endostatin was eluted by an increasing NaCl gradient from 0.1 M to 2 M in 20 mM Tris-HCl, pH 7.4. Endostatin fractions were pooled and dialyzed against PBS, pH 7.4. In order to remove any residual bacterial-derived endotoxin traces, all protein preparations were passed through a Polymyxin agarose column using lxPBS buffer and the purified endostatin was concentrated by ultrafiltration to 0.5-1.5 mg/ml and stored at -20°C until use .

The recombinant endostatin protein spans the 183 C-terminal amino acid residues of human α_: (XVIII) collagen and contains an N-terminal His-tag having the sequence MRGSHHHHHHTDPHASSVP (SEQ ID NO:100). Figure 2 shows an SDS-PAGE analysis of recombinant endostatin purified by the above methods. Briefly, 5μg of recombinant human endostatin was separated under non-reducing conditions (left panel) and reducing conditions (right panel) and stained with Coomassie blue. Molecular weight markers are indicated on each side of the panels. As shown in Figure 2, the purified protein migrated as a discrete band at Mr 18,000 and 24,000 on SDS-PAGE under non-reducing and reducing conditions, respectively. Far-UV circular dichroism (CD) spectroscopy was performed to characterize the isolated, soluble endostatin. Briefly, Far-UV CD spectrum for purified endostatin was recorded on an AVIV Associates (Lakewood, NJ) model 62DS spectrometer equipped with a temperature controller. Protein concentration for CD analysis was determined by using extinction coefficients calculated from amino acid composition. Buffer conditions in the CD analysis were 10 mM potassium phosphate, pH 8.0, and cells of 1 mm path length were used. A five second time constant and a 1.0 nm bandwidth was used during data acquisition over a wavelength range of 184 to 260 nm; three spectra were collected for protein or buffer and were averaged. Buffer spectra were subtracted from the protein spectra. The isolated, soluble endostatin had a spectrum similar to that published earlier for mouse and human endostatin with a characteristic minimum at 210 nm and with secondary structure estimates indicating predominantly α-helix and β-sheet structure (Sasaki et al . EMBO J 17:4249, 1998). Moreover, the recombinant endostatin was found to potently inhibit bFGF-induced proliferation of endothelial cells with an IC50 of approximately 500 ng/ml (O'Reilly et al . Cell 88:277, 1997, Hohenester et al. EMBO J 17:1656, 1998, Nguyen et al, Cancer Res 58:5673, 1998, Blezinger et al . Nat Biotechnol 17:343, 1999, Dhanabal et al . Cancer Res. 59:189, 1999, Ramchandran et al . BBRC 255:735, 1999). These results demonstrate that the purified human endostatin used in the studies described below has very similar biochemical and biological properties as purified endostatin molecules isolated by other methods.

Soluble endostatin was assessed for its ability to bind to cell surface adhesion receptors, and in particular, to the integrin family of adhesion receptors. Integrin-mediated cell attachment on cognate integrin ligands, such as extracellular matrix proteins, results in cell spreading, clustering of integrins at focal adhesions, and the induction of tyrosine phosphorylation of intracellular proteins . When integrin inhibitors such as antibodies and RGD-peptides are immobilized on a substrate, they act as agonists and similarly activate intracellular events downstream of integrins . To examine the binding activity of endostatin to cell adhesion receptors, the integrin agonist activity of immobilized endostatin was assessed.

Briefly, human umbilical vein endothelial cells (Huvec) from pooled donors (Huvec; Clonetics Corp., San Diego, CA) were cultured in EGM medium (Clonetics) supplemented with 12 μg/ml of bovine brain extract, 2 mM L-glutamine, 50 μg/ml streptomycin and 50 U/ml penicillin (Irvine Scientific, Santa Ana, CA) . Experimentation was carried out at cell passage number 4-12. To perform the cell spreading and immunofluorescence analysis, Huvec cells were detached by trypsinization, washed once with M199-medium (Irvine Scientific, Santa Ana, CA) supplemented with 10% FCS and once with M199-medium supplemented with 0.5% BSA, 10 ng/ml bFGF and 10 ng/ml

EGF. The cells were plated for 2 h on coverslips that had been coated with 20 μg/ml of either endostatin or vitronectin, or with 100 μg/ml of polylysine. Cells were then washed, fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100 and stained for vinculin and actin. Staining with the vinculin monoclonal antibody (Sigma) was visualized by FITC-conjugated anti-mouse antibodies (Jackson ImmunoResearch, West Grove, PA). TRITC-phalloidin (Sigma) was used for actin staining according to manufacturer's instructions.

Immunocytochemical analysis was carried out with Nikon inverted microscope, and the degree of cell spreading was analyzed with phase-contrast microscopy.

To investigate tyrosine phosphorylation status of the focal adhesion kinase (FAK), Huvec cells were detached by brief trypsinization followed by washing with soybean trypsin inhibitor. Cells were washed twice with complete EBM medium without serum, suspended to the same medium and either kept in suspension at 37°C for 15 min on a rotator, or plated on dishes coated with endostatin (20 μg/ml), vitronectin (20 μg/ml) or polylysine (100 μg/ml) and incubated at 37°C for 45 min. Cells were washed with PBS, and lysed in a modified RIPA buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, 0.1% SDS, 1% deoxycholate, 50 mM NaF, 0.5 mM Na3V04, 0.1 U/ml aprotinin, 10 μg/ml leupeptin and 4 μg/ml pepstatin A) . Immunoprecipitations with anti-FAK antibody (Transduction Laboratories, Lexington, KY) , as well as immunoblotting with horseradish peroxidase-conjugated anti-phosphotyrosine py20 antibody (Transduction Laboratories) and with the anti-FAK antibody followed by enhanced chemiluminescence detection (Pierce, Rockford, IL) were carried out as in Vuori and Ruoslahti, J. Biol . Chem . 268:21459-21462 (1993).

The results of the above analysis demonstrated that immobilized endostatin promoted cell spreading, focal adhesion formation, actin stress fiber formation and tyrosine phosphorylation of the focal adhesion kinase FAK similar to that induced by the α_v integrin ligand vitronectin. In contrast, endothelial cell attachment to polylysine, to which cells adhere in an integrin-independent manner, did not induce these cell biological events. Specifically, cells on polylysine remained round, and failed to assemble focal adhesions and induce tyrosine phosphorylation of FAK. These results indicate that immobilized endostatin serves as an adhesive substrate for α_v integrin-containing cells and activate post-ligand binding events downstream of integrins.

The α_v integrin binding activity of endostatin was further determined using quantitative cell attachment assays. Cell attachment to immobilized proteins was assayed in microtiter plates as described previously with slight modifications Zhang et al . , J. Cell . Biol . 122:235-242 (1993). Briefly, the plates were coated overnight with 20 μg/ml of either endostatin, vitronectin or fibronectin or 100 μg/ml of polylysine. The wells were then washed with phosphate-buffered saline solution (PBS) and blocked for 30 mm with 0.5% bovine serum albumin (BSA) m PBS; it should be noted that BSA binds in a nonspecific manner to endostatin, and a longer blockage with BSA appears to interfere with the assay.

Exponentially growing Huvec cells were briefly trypsinized, and trypsin was inactivated with serum-containing M199 medium followed by a wash with serum-free M199 medium containing 0.5% BSA. Cells were suspended in 5xl0⁵ cells/ml in the same media with growth factors as described above, and incubated in the presence or absence of 10 mM EDTA, 0.1 or 1.0 mM RGD-peptide or 1.0 mM RGE-peptides or anti-mtegrm antibodies for 30 mm at 4°C. GRGDSP- and GRGESP-peptides and anti-β mtegrm antibody P4C10 were from Gibco-BRL (Grand Island, NY) . The following anti-mtegrm monoclonal antibodies were obtained from Chemicon (Temecula, CA) : FB12

antibody) , P1E6 (anti-α, antibody) , P1B5 (antι-α₃ antibody) , P1H4 (antι-α₄ antibody) , P1D6 (antι-α₅ antibody) , CLB701 (antι-α₆ antibody) , LM609 (anti-α^ antibody) , and P1F6 (antι-αβ₅ antibody) . Antι-α_v antibody L230 was from American Type Culture Collection. Purified o;₁β₁, α₅β₁ and α_vβ₃ integrins were from Chemicon. Cell suspensions (5xl0⁴ cells/well) were added to the wells, and the plates were incubated at 37°C for 50 mm. The plates were washed three times with PBS, and the cells were fixed with 20% methanol or 5% glutaraldehyde for 15 mm at room temperature. Quantitation of cell attachment was achieved by stammg cells with 0.5% crystal violet m 20% methanol, washing cells, and then elutmg the dye with 2% SDS and measuring the absorbance at 590 nm.

Immobilized endostatin supported endothelial cell adhesion in a saturable and concentration-dependent manner (Figure 3A) . A peptide synthesized corresponding to the N-terminal His tag of the recombinant endostatin failed to support cell attachment when immobilized on a solid surface. Also, adhesion to endostatin was blocked by affinity-purified endostatin antibodies. These results confirm that the cell adhesion activity of the isolated, soluble endostatin preparation is attributable to endostatin .

The studies described below and shown in Figure 3B-3D further demonstrate that endothelial cell attachment on immobilized endostatin is integrin-mediated. Specifically, adhesion of cells on endostatin was significantly blocked by the presence of 10 mM EDTA (Figure 3B) , indicating that immobilized endostatin mediates cell adhesion through a divalent cation-dependent cell surface receptor, such as an integrin. Similar inhibition by EDTA was observed on cell attachment to fibronectin and vitronectin, to which Huvec cells adhere primarily through the ₅βj and _v-integrins, respectively (Figure 3B) . Endothelial cell adhesion to the non-integrin substrate polylysine was unaffected by EDTA (Figure 3B) .

Moreover, peptides containing the sequence RGD, which constitutes a recognition sequence for several members of the integrin family including the _v- and α₅-integrins, inhibited endothelial cell attachment on immobilized endostatin and vitronectin, whereas an RGE-containing peptide had little effect. RGD-peptides did not affect cell attachment to collagen I, to which Huvec cells adhere via non-RGD-dependent a₁β_l and of₂β_ι integrins. The results of these adhesion studies are shown in Figure 3C. Inhibitory anti-mtegrm monoclonal antibodies also were used in cell attachment assays to further characterize the integrins that mediate endothelial cell adhesion on immobilized endostatin. Huvec cells were incubated inhibitory antibodies prior to plating on 20 μg/ml endostatin coated plates as described above. As shown in Figure 3D, monoclonal antibody L230 against α -integrins significantly reduced endothelial cell adhesion on endostatin. A slight reduction m the cell attachment on endostatin was observed when antibodies against the α β₃ and _vβ_ς integrins were tested individually (monoclonal antibodies LM609 and P1F6, respectively) . Combination of the two antibodies, however, resulted in a similar inhibition of cell attachment to endostatin as was observed with anti-av antibodies. Monoclonal antibody against the β -mtegrm (antibody P4C10) inhibited cell attachment to endostatin by approximately 30%, indicating that an RGD-dependent βj-integrin, such as α.β and/or α₅β_lr also contributes to cell-endostatin interaction. The binding of endostatin by α₅β_x was confirmed using the inhibitory antibody P1D6 which resulted m reduced endothelial cell attachment on endostatin. Moreover, a combination of antibodies against α- and o^- tegrms abolished cell adhesion to endostatin to the same level as was observed when cells were treated with EDTA or RGD-peptides .

Antibodies against several non-RGD-dependent βl- tegrms, including a._{l r} α,, α₃, a_a and α_e integrins, failed to affect cell adhesion on endostatin when used alone or in combination. Control experiments with appropriate extracellular matrix proteins confirmed that the antibodies utilized in this study blocked endothelial cell adhesion m a specific manner. Therefore, these results demonstrate that endothelial cell adhesion on immobilized endostatin is mediated by the RGD-dependent α_vβ₃, α_vβ₅, and ct_{5 1} integrins.

The above results were corroborated by assessing the binding of soluble endostatin, to _v- and α₅-integrins as well as to α_IIbβ₃ integrin immobilized to a solid phase. Briefly, the solid-phase ligand binding assay was performed in a 96-well flat-bottom microtiter plates (Lindbro/Titertek, ICN Biomedicals Inc., Costa Mesa, CA) that were coated with 5 mg/ml of purified integrins in 150mM NaCl, 50mM Tris-HCl, pH 7.4 , 1 mM

CaC12, ImM MgC12 (= TBS-Ca/Mg) , and 4 mM octyl glucoside overnight at room temperature. The wells were then blocked with 3% BSA in TBS-Ca/Mg at room temperature for 2 h. Integrin ligands were overlaid in TBS-Ca/Mg and incubated with rotation at room temperature for 3 h.

Vitronectin and fibronectin were used at a concentration of 0.2 ng/well, collagen I at a concentration of 10 ng/well, and endostatin at a range from 0-1000 ng/well. Specific antibodies against the matrix proteins and peroxidase-conjugated anti-rabbit, anti-mouse or anti-goat IgG antibodies were used to detect the bound proteins.

Unbound ligands were washed three times with TBS-Ca/Mg, 0.05% Tween 20, and the bound ligands were incubated with antibodies to vitronectin (1/100 dilution of the mouse monoclonal antibody 8E6 ascites, fibronectin (1/1000 dilution of rabbit polyclonal antibody 1653 serum, provided by Dr. Erkki Ruoslahti, The Burnham Institute), collagen type I (0.5 μg/ml of purified goat anti-collagen I antibody, Chemicon) or the N-terminal his-tag of endostatin (0.5 μg/ml of purified RGS-His monoclonal antibody, Qiagen Inc., Valencia, CA) for 1 h at room temperature. After extensive washes with TBS-Ca/Mg, 0.05% Tween 20, the bound antibodies were detected by Ultra-Sensitive ABC Peroxidase Staining reagents (Pierce) according to manufacturer's instructions by using biotinylated goat anti-mouse IgG antibodies (Vector Laboratories, Inc., Burlingame, CA) , goat anti-rabbit IgG antibodies (Pierce) and rabbit anti-goat IgG antibodies (Pierce). TMB liquid substrate (50 μl/sample, Sigma) was used as the enzyme substrate with an incubation time of 15 min, and the reactions were stopped with 50 μl of 0.5M H2S04, whereafter absorbance was measured at 450 nm. The assays were carried out in triplicate.

Soluble endostatin demonstrated a concentration-dependent and saturable binding to purified α_vβ₃ and α₅βχ and α_IIbβ₃ integrins. The specificity of the interaction was further confirmed by inhibition of endostatin binding to the α_vβ₃ integrin with the monoclonal α_v-antibody L230 and to the α₅β_x integrin with the α₅-antibody P1D6. In contrast, endostatin failed to demonstrate any binding to the non-RGD-dependent collagen receptor a_{1 1} integrin. In control experiments, the purified integrins interacted with their known extracellular matrix protein ligands, attesting for proper ligand binding function of the purified integrins. Therefore, purified α_vβ₃ integrin demonstrated binding to vitronectin and fibronectin, α₅β_x integrin bound to fibronectin, and a_{1 1} integrin interacted with collagen I. These cell biological and biochemical evidence demonstrate a specific interaction between the angiogenic inhibitor endostatin and the α_v- and α₅-integrins .

Additional studies were performed to demonstrate that immobilized endostatin promotes whereas soluble endostatin inhibits α_vβ₃ integrin-dependent endothelial cell migration and survival. Briefly, cell motility on immobilized endostatin and gelatin was measured using a modified Boyden chamber (NeuroProbe, Cabin John, MD) as previously described Zhang et al . , supra . The undersurface of 10-μm-pore polycarbonate membrane filter (Poretics, Livermore, CA) was precoated with 10 μg/ml of endostatin or gelatin. Lower chambers were filled with complete EGM culture medium without serum containing 0.1% BSA and 10 μg/ml of endostatin or gelatin. Huvec cells (lxlO⁴) were added to the upper chambers in the same medium except without endostatin or gelatin and migration was scored in the presence or absence of 10 μg/ml of the anti-α_vβ₃ antibody LM609 or the anti-α_:β₁ antibody P1E6.

The effect of soluble endostatin or cell mobility was also determined. Briefly, soluble endostatin was added to the test cells at concentrations of 1,5,10 and 20 μg/ml. The anti-α_vβ₃ antibody LM609 was added at a concentration of 20μg/ml whereas control cells contained no additions. The cells were incubated at 37°C for 2, 4, 6 or 8 h, and the membranes were fixed in methanol and the cells were stained with crystal violet. The cells on the upper surface of the membrane were removed, membranes were mounted on glass slides and the number of cells migrated to the lower surface was counted. Migration results were determined as the average number of cells/high-magnification microscopic field (HMMF) .

To determine the capability of immobilized endostatin to support endothelial cell survival, Huvec cells were plated under serum-free conditions on tissue culture dishes that had been precoated with 20 μg/ml of endostatin or LM609, or with 1% of heat denatured BSA. Soluble LM609 or P1E6 at concentrations of 10 μg/ml were added to cells in some of the studies prior to plating. Apoptotic cell death was monitored 8 h later by measuring DNA fragmentation using the Cell Death Detection ELISA kit according to manufacturer's instructions (Boehringer Mannheim, Indianapolis, IN) . When analyzing the capability of soluble endostatin to induce apoptosis in endothelial cells, Huvec cells cultured in 24-well dishes (lxlO⁵ cells/well) were switched to complete EGM medium without serum, and soluble endostatin and/or inhibitory anti-integrin antibodies at a concentration of 10 μg/ml were added to the wells. Cell apoptosis was monitored as above 12 h later.

Endothelial cell contact with an increased concentration of immobilized α_vβ₃ integrin ligand, such as vitronectin and anti-α_vβ₃ antibodies, can enhance cell migration in a haptotactic migration assay. On the other hand, endothelial cell migration is efficiently prevented by the same α_vβ₃ ligands when administered in solution to the cells (Leavesley et al . , J. Cell Biol . 121:163-170 (1993). Likewise, immobilized agonists of α,,β₃ promote cell survival, while soluble antagonists of this integrin induce apoptosis of endothelial cells (Stromblad et al . , J. Clin . Invest . 98:426-433 (1996); Scatena et al . , J. Cell Biol . , 141:1083-1093 (1998). The results of the studies described above indicate that recombinant, soluble endostatin is similarly capable of modulating endothelial cell function, and it does so in an α_vβ₃ integrin-dependent manner. Specifically, endothelial cells readily migrated in a haptotactic Boyden chamber assay through a microporous membrane toward immobilized endostatin, and about 75% of this migration activity was efficiently blocked by the monoclonal anti-α_vβ₃ antibody LM609, but not by the anti-α^ antibody P1E6. Moreover, soluble endostatin also inhibited endothelial cell migration by greater than 75% on immobilized gelatin and in a concentration-dependent manner. Endothelial cell migration on gelatin was also found to be mediated by the α_vβ₃ integrin, as the α_vβ₃-antibody LM609 efficiently blocked migration on this substrate and was comparable to the inhibition observed with 20 μg/ml endostatin. Therefore, these results demonstrate that immobilized endostatin promotes, and soluble endostatin inhibits endothelial cell migration in an α_vβ₃ integrin-dependent manner .

Similar results also were obtained with respect to endothelial cell survival. Endothelial cells undergo apoptosis upon serum withdrawal if appropriate integrin ligation is denied. As shown in Figure 4A, when Huvec cells were plated on immobilized BSA under serum-free conditions, significant levels of apoptosis were detected after 8 h of incubation. In contrast, plating of Huvec cells on immobilized anti-α_vβ₃ antibody LM609 protected cells from apoptosis. Similarly, negligible levels of apoptosis were detected in cells adherent on immobilized endostatin. Administration of soluble anti-α_vβ₃ antibody LM609, but not of the anti-α₂β₁ antibody P1E6, induced apoptosis on cells plated on endostatin, demonstrating that immobilized endostatin supports cell survival in an α_vβ₃-dependent manner.

Soluble endostatin in turn was found to induce endothelial cell apoptosis, and it did so under conditions in which cell survival was dependent on the α_vβ₃ integrin. The results of these studies are shown in Figure 4B. Briefly, monolayer cultures of Huvec cells undergo apoptosis upon serum withdrawal and concomitant inhibition of the α_vβ₃ and α₅β_x integrins. Huvec cells that were exponentially growing as a monolayer underwent apoptosis following serum withdrawal and simultaneous administration of the anti-α_vβ₃ antibody LM609 and the inhibitory antibody P1D6 against the α₅β_x integrin. Moderate levels of ^"apoptosis were detected when either one of the antibodies was added alone to the cells. Moderate levels of apoptosis were also observed upon addition of endostatin alone to the endothelial cells. Administration of soluble endostatin together with the anti-α^ antibody P1D6 induced apoptosis of endothelial cells to the same extent as when the anti-α_vβ₃ antibody was administered together with the anti-α_sβ_! antibody. Addition of the inhibitory anti-α_:β₁ antibody alone or in combination with the anti-α^ antibody failed to induce apoptosis. Therefore, under conditions where Huvec cell survival is dependent on both the α_vβ₃ and the a ₁ integrins, soluble endostatin preferentially interferes with survival signals mediated by the α_vβ₃ integrin. These results demonstrate that binding of endostatin to integrins is of functional significance, and results in the modulation of endothelial cell functions that are important for angiogenesis, such as migration and survival .

EXAMPLE II

Endostatin Derived from Collagen Type XV Exhibits

Similar Activity as Endostatin Derived from Other Sources

This Example describes the expression, purification and α_v integrin-mediated activity of recombinant, soluble endostatin derived from collagen type XV.

The fragment of human type XV collagen that corresponds to the sequences of the endostatin region in type XVIII collagen was generated by PCR using the clone NH1-2 as a template (Kivirikko et al . J. Biol. Chem. 269:4773-4779, 1994, which is incorporated herein by reference) . Primers used for PCR amplification were END015-1 (5 ' -ATAAAGCTTACTTCCTAGCGTCTGTC-3 ' (SEQ ID NO: 100), where a Kpnl restriction site is underlined) and END015-2 ( 5 ' -ATGGTACCTTCAAGTGCCAATTATGA-3 ' (SEQ ID NO:101), where a Hindlll restriction site is underlined) . The fragment was amplified and subcloned into the Kpnl-Hindlll-site of the vector pQE-31 (Qiagen, Inc., Santa Clarita, CA) as described previously in Example I. Expression of human endostatin derived from collagen XV was also performed in E . col i strain M15 and was induced according to the manufacturer's recommendations (Qiagen) again as described previously in Example I . The purification of endostatin derived from XV collagen was also performed as described previously in Example I for endostatin derived from collagen XVIII.

To demonstrate that endostatin derived from collagen XV interacts with α_v integrins essentially identical to endostatin derived from collagen type XVIII quantitative cell attachment assays were performed. WO 00/67771 PCT/USOO/l 2063

88

Briefly, cell attachment to immobilized endostatin derived from collagen type XV (endostatin XV) was assayed in microtiter plates with slight modifications to the method described by Zhang et al . supra and as described previously in Example I . The plates were coated overnight with the below described concentrations of endostatin XV. The wells were then washed with phosphate-buffered saline solution (PBS) and blocked for 30 min with 0.5% bovine serum albumin (BSA) in PBS. As described previously, BSA binds in a nonspecific manner to endostatin and a longer blocking period with BSA appears to interfere with the assay .

To perform the attachment assays, exponentially growing Huvec cells were briefly trypsinized, and trypsin was inactivated with serum-containing M199 medium followed by a wash with serum-free M199 medium containing 0.5% BSA. Cells were suspended in 5xl0³ cells/ml in the same media with growth factors as described previously in Example I and incubated in the presence or absence of RGD- or RGE-peptides or anti-integrin antibodies as indicated in the Figure 5 for 30 min at 4°C. Cell suspensions (5xl0⁴ cells/well) were added to the wells, and the plates were incubated at 37°C for 50 min. The plates were washed three times with PBS, and the cells were fixed with 20% methanol or 5% glutaraldehyde for 15 min at room temperature. Quantitation of cell attachment was achieved by staining cells with 0.5% crystal violet in 20% methanol, washing cells, and then eluting the dye with 2% SDS and measuring the absorbance at 590 nm.

Figure 5A shows the dose dependence of cell adhesion to endostatin XV. Huvec cells were plated onto microtiter wells coated with increasing concentrations of endostatin XV as indicated in the figure and cell attachment was analyzed as described above. The results of this study demonstrate that cells bind to endostatin XV in a concentration dependent manner.

Figure 5B shows the inhibition of cell adhesion to endostatin XV by RGD-peptides . Huvec cells were incubated with or without 0.1 or 1.0 mM GRGDSP or 1.0 mM GRGESP peptides as indicated, prior to plating of the cells to the microtiter wells coated with 20 μg/ml of endostatin XV. Cell attachment assay was carried out as above. The results demonstrate that RGD peptides, which constitutes a recognition sequence for several members of the integrin family including the α_v and α₅ integrins inhibited endothelial cell attachment on immobilized endostatin whereas RGE-containing peptides had little effect.

Figure 5C further demonstrates that cell adhesion to endostatin is α_v and α₅ integrin-dependent . Huvec cells were incubated with the identical pannel of antibodies described previously for Figure 3. Briefly, the anti-integrin antibodies used also are indicated in Figure 5C and were: anti-α_v (L230), anti-α_vβ₃ (LM609), anti-α_vβ₅ (P1F6) , anti-β^_. (P4C10) and anti-α₅β, (P1D6) . Incubation with cells was performed prior to plating on wells coated with 20 μg/ml of endostatin from collagen XV. Cell attachment was analyzed as described above and in Example I. The results of this study mirror those shown in Figure 3 and demonstrate that the integrin binding activity of endostatin derived from collagen type XV is essentially identical to that derived from collagen type XVIII.

To determine the effect of endostatin XV on integrin-mediated endothelial cell functions, the ability of immobilized endostatin XV to support endothelial cell survival and soluble endostatin XV to induce apoptosis was assessed. As described previously in Example I, Huvec cells were plated under serum-free conditions on tissue culture dishes that had been precoated with 20 μg/ml of endostatin XV or with 1% heat denatured BSA. Apoptotic cell death was monitored 8 h later by measuring DNA fragmentation using the Cell Death Detection ELISA kit according to the manufacturer's recomendations (Boehringer Mannheim, Indianapolis, IN) . The data are presented in Figure 6 as the mean +/- SD. In contrast, when analyzing the capability of soluble endostatin XV to induce apoptosis in endothelial cells, Huvec cells cultured in 24-well dishes (1x10^s cells/well) were switched to complete EGM medium without serum, and soluble endostatin XV at a concentration of 10 μg/ml was added to the wells. Cell apoptosis was monitored as above 12 h later.

As shown in Figure 6A, when Huvec cells were plated under serum-free conditions on immobilized endostatin XV the cells were protected from apoptotic cell death. In contrast, significant levels of apoptosis was observed in cells plated on BSA. These results demonstrate that immobilized endostatin XV promotes endothelial cell survival in a similar manner as that observed for endostatin derived from collagen XVIII.

Figure 6B shows the results of soluble endostatin XV on apoptotic cell death. Briefly, compared to control cultures without any additions, significant levels of apoptosis was observed in exponentially growing Huvec cells upon serum withdrawal and simultaneous addition of endostatin XV. As with the above studies, these results demonstrate that endostatin XV results in the modulation of integrin-mediated endothelial cell functions essentially identical to that of endostatin derived from collagen type XVIII.

EXAMPLE III Inhibition of Tubulogenesis by Endostatin

This Example demonstrates that endostatin derived from both collagen types XVIII and XV inhibit tubulogenesis .

To further demonstrate the inhibition of α_v mtegrm-mediated endothelial cell function the inhibition of tubulogenesis was assessed. Briefly, human umbilical vein endothelial cells (HUVECs) were plated onto a gel of collagen I in the presence of 10 ng/ml bFGF and 10 ng/ml EGF. Under these conditions, HUVECs are induced to form a capillary network withm the collagen gel. In duplicate cultures, realignment of the cells, cell invasion into the gel and the beginning of the cell elongation were evident at 8 h after plating of the cells onto the gel. Tube formation was clearly visible by 12 h. In contrast, very little, if any, tube formation occured when the cells under the same culture conditions were treated with 10 μg/ml of endostatin derived from either type XVIII collagen or from type XV collagen. Therefore, processes preceding the tube formation failed to take place efficiently in the presence of endostatin, and the tube number, length and width were significantly reduced by approximately 90%. These results demonstrate that endostatin is a highly efficient inhibitor of endothelial cell tube formation. EXAMPLE IV

Functional Fragments of Endostatin Exhibit ctv

Integrin Binding Activity and Cell Attachment Activity

This Example shows that cleavage products of endostatin exhibit similar activity to intact endostatin.

To determine whether fragments of endostatin maintain integrin binding activity comparable to the intact molecule and to localize binding specificity to subregions of the molecule, endostatin was fragmented by several procedures and assessed for α„ integrin-mediated activity. The ability of the endostatin fragments to support cell adhesion was used as a measure of α_v and α₅ integrin binding activity.

In an initial study, endostatin was fragmented into small peptides by proteolytic digestion with trypsin or chymotrypsin and the fragments were assessed for their ability to maintain cell adhesion activity. The predicted proteolytic fragments for each of these dig stions were presented previously in Tables 2 and 3, respectively (see Detailed Description) . Briefly, purified human endostatin (20 μg) was lyophilized in an eppendorf tube and solubilized in 50 ml of 1 M Urea, 50 m M Tris-Hcl bicarbonate, 1 mM CaCl₂, pH 8.0. Trypsin or chymotrypsin was added at a 1:10 (w/w) ratio, and the mixture was incubated for 24 h at room temperature. The reaction was quenched by diluting the Urea to 0.1 M folllowed by lyophilization . The digest was solubilized in PBS and immobilized on microtiter plates. Undigested endostatin (20 μg) also was immobilized on microtiter plates as a control. Cell attachment assay was carried out as described in Examples I and II, and quantitation of endothelial cell attachment was done by measuring the absorbance at 590 nm.

Endothelial cell attachment to the trypsin and chymotrypsin fragments of endostatin and undigested endostatin was found to be essentially the same, as OD₅₉₀ values of 0.70, 0.72 and 0.8 were obtained, respectively. These measurements are average of three independent experiments. These results demonstrate that integrin binding activity of endostatin is not destroyed by proteolytic digestion and can therefore reside in discrete fragments of the molecule.

In a second study, endostatin was fragmented by chemical cleavage using CnBr. The cleavage products for this reaction have been described previously and are identified as SEQ ID NOS:19-22 (see Detailed Description). Briefly, purified human endostatin (20 μg) was lyophilized in an eppendorf tube and solubilized in 50 μl of 70% formic acid. A crystal of cyanogen bromide (CnBr) was added and swirled to dissolve. The tube was flushed with nitrogen and capped, and kept in the dark at room temperature for 12 h. The reaction was quenched by evaporating the formic acid in a desiccator containing a beaker of NaOH pellets, and the digested endostatin was dried in a Speed-Vac. The digest was solubilized in PBS and immobilized on microtiter plates in parallel with undigested endostatin (20 μg) as a control. Cell attachment assay was carried out as described in Examples I and II and quantitation of endothelial cell attachment was done by measuring the absorbance at 590 nm.

Endothelial cell attachment to the CnBr fragments of endostatin and undigested endostatin was found to be the same, as an OD_59C value of 0.78 was obtained in both cases. These measurements are average of three independent experiments. These results demonstrate that integrin binding activity of endostatin is not destroyed by chemical cleavage at methionine residues within endostatin and that the binding activity can therefore reside in discrete fragments of the molecule.

EXAMPLE V

Endostatin Variants Exhibit Increased α„ Integrin

Binding Activity and α_-Mediated Endothelial Cell Function

This Example shows that inclusion of an RGD sequence into endostatin increases integrin binding activity whereas mutation of endostatin peptide sequences reduce the integrin binding activity compared to endostatin .

To localize and identify endostatin sequences responsible for α_v integrin binding activity within the fragments described in Example IV, putative α_v integrin binding region sequences were targeted by mutational analysis to determine their effect on endostatin function. Three categories of mutations were generated and the corresponding mutant endostatin proteins were studied in cell attachment assays. First, RLQD-, RAD- and DGK/R-sequences in endostatin were mutated so that either nucleotides coding for arginine (R) or lysine (K) were changed to code for alanine (A) , or nucleotides coding for aspartic acid (D) were changed to code for A. As a representative example of these types of mutations, the RAD-sequence of endostatin in residues 63-65 was mutated to AAD ("AAD"-mutant) . In another approach, an RGD sequence was added to the endostatin sequence in a region that is exposed to the solvent. As a representative example of these types of mutations, a variant of endostatin was generated in which RAD₆₃_₆₅ was replaced with the sequence RGD ( "RGD-vaπant" ) . In the third type of approach, mutations were made m the region of endostatin that has a non-contiguous RGD sequence. As a representative example of these types of mutations, nucleotides coding for D at residue 30 were changed to code for threonine (T) ( "D->T"-mutant ) . Mutations were generated by using standard in vi tro site-directed mutagenesis methods, and the variant endostatin molecules were expressed and purified in E . col i as described for the intact or wild-type endostatin (wt-ES) . Endothelial cell attachment was carried out on immobilized wild-type and variant forms of endostatin as described in Examples I, II and IV.

Representative results of the cell attachment activity of each category of mutation is shown in Figure 7. Specifically, the RGD-vaπant form of endostatin showed enhanced mtegrm binding activity compared to wild-type endostam. Conversly, the capability of

AAD-mutants and D->T- mutants of endostatin to support endothelial cell attachment was reduced compared to the wild-type form of endostatin. The residual activity in the mutant versions is likely due to the unmutated RLQD, RAD, . DGK/R or non-contiguous sites remaining in each particular mutant. These results therefore indicate that the mtegrm binding regions of endostatin reside m each of these sequences since a decrease m mtegrm binding activity can be observed with each corresponding mutation of an endostatin peptide sequence to an inactive form. These results also demonstrate that inclusion of an RGD ammo acid sequence into endostatin significantly augments the functional activity of endostatin. Although the invention has been described with reference to the disclosed embodiments, those skilled in the art will readily appreciate that the specific experiments detailed are only illustrative of the invention. It should be understood that various modifications can be made without departing from the spirit of the invention. Accordingly, the invention is limited only by the following claims.

Claims

What is claimed is:

1. An endostatin peptide comprising at least about 4-7 endostatin amino acid residues containing substantially the amino acid sequence RLQD, RAD, DGK/R, (SEQ ID NOS:ll-13) or a functional equivalent thereof.

2. The endostatin peptide of claim 1, wherein said RLQD, RAD, DGK/R amino acid sequence is non-contiguous.

3. The endostatin peptide of claim 2, wherein said non-contiguous sequence comprises the amino acid sequence

RGIRGADFQCFQQAR or RGIRGADFQCFQQARAVGLAGTFR (SEQ ID NOS: 15 or 17) .

4. The endostatin peptide of claim 1, further comprising two or more of the amino acid sequences RLQD, RAD, DGK/R, or a functional equivalent thereof.

5. The endostatin peptide of claim 1, further comprising α_v integrin binding activity.

6. The endostatin peptide of claim 5, wherein said α_v integrin binding activity is α_vβ_lf α_vβ₃ or α_vβ₅ binding activity.

7. The endostatin peptide of claim 5, wherein said α_v integrin binding activity is α_vβ₃ binding activity.

8. The endostatin peptide of claim 1, further comprising glycosaminoglycan binding activity.

9. The endostatin peptide of claim 8, wherem said glycosaminoglycan binding activity further comprises chondroitin sulphate A, chondroitin sulphate B, chondroitin sulphate C, heparin, heparin sulphate, or keratan sulphate binding activity.

10. An endostatin peptide comprising substantially the human endostatin peptide ammo acid sequence corresponding to SEQ ID NOS:20 and 21, SEQ ID NO:20, or a functional fragment thereof.

11. The endostatin peptide of claim 10, further comprising α_v mtegrm binding activity.

12. The endostatin peptide of claim 11, where said α_v mtegrm binding activity is α,β_x, o^β₃ or α_vβ_c binding activity.

13. The endostatin peptide of claim 11, where said α_v mtegrm binding activity is α_vβ₃ binding activity.

14. The endostatin peptide of claim 10, further comprising glycosaminoglycan binding activity.

15. The endostatin peptide of claim 14, wherem said glycosaminoglycan binding activity further comprises chondroitin sulphate A, chondroitin sulphate B, chondroitin sulphate C, heparin, heparin sulphate, or keratan sulphate binding activity.

16. An endostatin variant comprising endostatin containing the ammo acid sequence RGD, or a functional fragment thereof.

17. The endostatin variant of claim 16, further comprising substantially the amino acid sequence as shown as SEQ ID NO:23.

18. The functional fragment of claim 16, wherein said RGD sequence further comprises the amino acid sequence RRGDR (SEQ ID NO:24), or a functional equivalent thereof .

la. The functional fragment of claim 16, further comprising one or more of the amino acid sequences RLQD, RAD, DGK/R, RGD, a non-contiguous sequence thereof, or a functional equivalent thereof.

20. The endostatin peptide of claim 19, wherein said non-contiguous sequence comprises the amino acid sequence RGIRGADFQCFQQAR or RGIRGADFQCFQQARAVGLAGTFR (SEQ ID NOS:15 or 17) .

21. The endostatin variant of claim 16, further comprising increased integrin binding activity compared to endostatin.

22. The endostatin variant of claim 21, wherein said increased integrin binding activity comprises an increase of about 3-fold, preferably about 5-fold, and more preferably about 10-fold or more in integrin binding activity compared to endostatin.

23. The endostatin variant of claim 21, wherein said increased integrin binding activity comprises an increase of about 100-fold or more in integrin binding activity compared to endostatin.

24. The endostatin variant of claim 21, wherein said increased integrin binding activity comprises α_v integrin binding activity.

25. The endostatin variant of claim 24, wherein said increased integrin binding activity comprises α_vβ₁, α_vβ₃ or α_vβ₅ binding activity.

26. The endostatin variant of claim 24, wherein said increased integrin binding activity comprises α_vβ₃ binding activity.

27. The endostatin variant of claim 16, further comprising glycosaminoglycan binding activity.

28. The endostatin variant of claim 27, wherein said glycosaminoglycan binding activity further comprises chondroitin sulphate A, chondroitin sulphate B, chondroitin sulphate C, heparin, heparin sulphate, or keratan sulphate binding activity.

29. A method of inhibiting angiogenesis comprising administering an angiogenic inhibitory amount of an endostatin peptide comprising at least about 4-7 endostatin amino acid residues containing one or more of the amino acid sequence RLQD, RAD, DGK/R (SEQ ID N0S:11-13), a non-contiguous sequence thereof, or a functional equivalent thereof.

30. The method of claim 29, wherein said non-contiguous sequence comprises the amino acid sequence RGIRGADFQCFQQAR or RGIRGADFQCFQQARAVGLAGTFR (SEQ ID NOS: 15 or 17) .

31. The method of claim 29, wherein said endostatin peptide further comprises α_v integrin binding activity.

32. The method of claim 31, wherein said _v integrin binding activity is α_vβ_x, α_vβ₃ or α_vβ₅ binding activity.

33. The method of claim 31, wherein said α_v integrin binding activity is α_vβ₃ binding activity.

34. A method of inhibiting angiogenesis comprising administering an angiogenic inhibitory amount of an endostatin variant comprising endostatin containing the amino acid sequence RGD, or a functional fragment thereof.

35. The method of claim 34, wherein said endostatin variant comprises substantially the amino acid sequence shown as SEQ ID NO:23.

36. The method of claim 34, wherein said RGD sequence in said endostatin variant further comprises the amino acid sequences RRGDR (SEQ ID NO:24), or a functional fragment thereof.

37. The method of claim 34, wherein said functional fragment further comprises one or more of the amino acid sequences RLQD, RAD, DGK/R, RGD, a non-contiguous sequence thereof, or a functional equivalent thereof.

38. The method of claim 37, wherein said non-contiguous sequence further comprises the amino acid sequence RGIRGADFQCFQQAR or RGIRGADFQCFQQARAVGLAGTFR (SEQ ID NOS:15 or 17) .

39. The method of claim 34, wherein said endostatin variant further comprises increased integrin binding activity.

40. The method of claim 39, wherein said increased integrin binding activity comprises an increase of about

3-fold, preferably about 5-fold and more preferably about 10-fold or more in integrin binding activity compared to endostatin .

41. The method of claim 39, wherein said increased integrin binding activity comprises an increase of about

100-fold or more in integrin binding activity compared to endostatin .

42. The method of claim 39, wherein said increased integrin binding activity comprises α_v integrin binding activity.

43. The method of claim 42, wherein said increased integrin binding activity comprises α_vβ_ι, α_vβ₃ or α_vβ₅ binding activity.

44. The method of claim 42, wherein said increased integrin binding activity comprises α_vβ₃ binding activity.

45. A composition comprising one or more of the antiangiogenic peptide or the endostatin variant of claims 1, 10, or 16 and a pharmaceutically acceptable medium.